From john@rrci.com Thu May 1 06:11:00 1997 From: john@rrci.com (John D. Robertson) Date: Thu, 01 May 1997 06:11:00 -0000 Subject: Cross compilation for MIPS/68000 embedded applications. References: <3366FBCD.B36BA80B@baynetworks.com> Message-ID: I have seen these messages before, but not while cross compiling. I think this happens when you declare a pure virtual function, and then never supply an implementation. class a { public: virtual void myfunc()=0; virtual void impfunc()=0; }; class b : public a { public: void impfunc() {} }; GCC will complain even if the function is never used. I believe that the SGI IRIX compiler will give you specific error messages specifying which function is not implemented, but will say nothing if the function is never accessed. Hope this helps. JDR |======================================================| | John D. Robertson, ADAMS Modeler / Software Engineer | | Robertson & Robertson Consultants, Inc. | | 3637 West Georgia Rd. | | Pelzer, SC 29669 | | | | Phone: (864) 243-2436 | | Fax: (864) 243-3023 | | Email: john@rrci.com | |======================================================| On Wed, 30 Apr 1997, Imerio Ballarini wrote: > I've built the 17.1 distribution of GCC as a cross compiler for the MIPS > processor and have got it to compile all of my embedded app. On linking > however, I'm getting unresolved "virtual table" references for some (and > only some) of my classes and I don't know why. The code I'm compiling > builds fine under different compilers. The specific class methods that > fail have nothing in common and differ little from my other classes that > build and link successfully. > > I'm at a loss. Any ideas? Has anyone successfully xcompiled embedded > code? > > Anyway, the whole job has been a pain and I have to do it all over again > for my 68000 apps. I don't suppose any of you have built a 68000 > xcompiler? > > Thanks, > -Imerio. > > -- > Imerio Ballarini > Bay Networks, Watford, UK. > Tel +44 (0)1923 479819 > - > For help on using this list (especially unsubscribing), send a message to > "gnu-win32-request@cygnus.com" with one line of text: "help". > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From fabio@joplin.colorado.edu Thu May 1 08:20:00 1997 From: fabio@joplin.colorado.edu (Fabio Somenzi) Date: Thu, 01 May 1997 08:20:00 -0000 Subject: random/srandom (was Re: link problems) References: <199704302026.NAA14207@cirdan.cygnus.com> Message-ID: <199705011518.JAA14223@joplin.colorado.edu> >>>>> "GN" == Geoffrey Noer writes: GN> James Paul Morgan wrote: >> As far as I can tell, the gnu-win32 does not have the random and >> srandom functions. It only has the rand and srand functions, >> which are not nearly as random. I have had problems with moving >> software from an AIX system to my Win95 system becuase of this. GN> [...] GN> You are correct. We have not implemented these yet as of the GN> upcoming beta 18 release. GN> If someone can point me to a Berkeley-licensed implementation of GN> them or write them and send me a copyright assignment form, I'll GN> add them... I'm currently away from my Win95 laptop, but I just want to point out that the botched rand function can be identified by looking at the least significant bit: It looks random for a good implementation, but it's a periodic 0-1 sequence for many rand's. On my Linux box at the office, rand and random are the same function and both are "good." Maybe the Linux code is a good starting point for inclusion in cygwin. On a related note, I've given up some time ago relying on system functions to generate random numbers. I use my own, which I derived from the algorithms in "Numerical Recipes in C." This is not to say that cygwin shouldn't provide a decent random/srandom pair. It's just that if you rely on the random number generator for more than an extemporary exercise, and you want reproducibility of results, you should seriously consider using your own random number generator. Fabio -- Fabio Somenzi | Phone: 303-492-3466 University of Colorado | Fax: 303-492-2758 ECE Dept. | Email: Fabio@Colorado.EDU Boulder CO 80309-0425 | WWW: http://vlsi.colorado.edu/~fabio - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Sergio.Valle@studi.epfl.ch Thu May 1 10:17:00 1997 From: Sergio.Valle@studi.epfl.ch (Sergio.Valle@studi.epfl.ch) Date: Thu, 01 May 1997 10:17:00 -0000 Subject: Which gcc libraries I must link with my ODBC application ? Message-ID: <199705011529.RAA06644@disun2.epfl.ch> Hello, I try to compile an ODBC application (to access Microsoft Access for Windows 95) in C++ with the gcc compiler but I don't know wich libraries I must link with my source code, does anybody know it ? Do gcc libraries support ODBC 3.0 ? Sergio Valle, Lausanne 1/5/97 (Switzerland) - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From m_kan@ya2.so-net.or.jp Thu May 1 10:17:00 1997 From: m_kan@ya2.so-net.or.jp (Masahiro Kan, Home) Date: Thu, 01 May 1997 10:17:00 -0000 Subject: G77 and Gnu-win32 References: <33682981.1AAC3224@coent.demon.co.uk> Message-ID: <199705011716.CAA26040@mail.ya2.so-net.or.jp> Dr. David Coe wrote; > Has anyone though managed to merge the G77 add-on with the Cygnus GCC? > Alternatively, are there plans for the extensive gnu-win32 changes to > GCC to be fed back into the GNU distribution? Have you visited the following URL? http://www.xraylith.wisc.edu/~khan/software/gnu-win32 Best regards, ------------------------- Masahiro Kan Surge Arrester Dept., Hamakawasaki-Works, TOSHIBA Corporation, Japan. E-mail:"Masahiro Kan at Toshiba" "Masahiro Kan, Home" WWW(ATP-EMTP-jp): http://club.infopepper.or.jp/~kanmasahiro - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From noer@cygnus.com Thu May 1 12:21:00 1997 From: noer@cygnus.com (Geoffrey Noer) Date: Thu, 01 May 1997 12:21:00 -0000 Subject: random/srandom (was Re: link problems) References: <199704302026.NAA14207@cirdan.cygnus.com> Message-ID: <199705011918.MAA19096@cirdan.cygnus.com> Geoffrey Noer wrote: [...] > If someone can point me to a Berkeley-licensed implementation of them or > write them and send me a copyright assignment form, I'll add them... I found them in the libiberty library available from the standard GNU ftp sites. They are under the Berkeley license so you should be able to include the code in your project or build and link against the libiberty library. -- Geoffrey Noer noer@cygnus.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From greggj@savvis.com Thu May 1 12:21:00 1997 From: greggj@savvis.com (Gregg Jensen) Date: Thu, 01 May 1997 12:21:00 -0000 Subject: waiting on threads Message-ID: <3.0.1.32.19970430112518.00918240@mail.savvis.com> Ok, I have added the use of threads in my program (instead of forking, which I could not get to work, but that is another issue...), and for the most part it works as advertised. I lay out 6 threads and then call WaitForMultipleObjects passing the variable hThread (which is defined as hThread[7]). The handle from each CreateThread is put into an empty spot in the hThread array. When I call WaitForMultipleObjects, I get an error of INVALID_HANDLE_ERROR (or whatever 6 is). I then went through the array one at a time with WaitForSingleObject and found that the fourth and fifth one were invalid, but the rest were OK. This is a little test prototype so there is not much going on, but I am curious to know if anyone, who has been using the native threads, has had any problems (other then the debugging problem from Bruce McLeod), with these functions? Gregg Jensen greggj@savvis.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From khan@xraylith.wisc.edu Thu May 1 13:16:00 1997 From: khan@xraylith.wisc.edu (Mumit Khan) Date: Thu, 01 May 1997 13:16:00 -0000 Subject: G77 and Gnu-win32 References: <33682981.1AAC3224@coent.demon.co.uk> Message-ID: <9705012016.AA00653@modi.xraylith.wisc.edu> David Coe writes: > Hi folks! > > I notice that the John Eaton's crew have completed a fair amount of a > native Octave port via gnu-win32. I think this must have been via f2c > for the Fortran bits rather than G77. > > Has anyone though managed to merge the G77 add-on with the Cygnus GCC? > Alternatively, are there plans for the extensive gnu-win32 changes to > GCC to be fed back into the GNU distribution? > See http://www.xraylith.wisc.edu/~khan/software/gnu-win32 for both g77 0.5.20 binaries/patches and also prebuilt octave 2.0.5. I wouldn't call octave for gnu-win32 a "port", since all it really needed was a "configure; make; make install" + workarounds for gnu-win32 bugs/omissions/misfeatures. Comes with no support whatsoever. Unless gnu-win32 gcc backend changes are merged back into FSF, I know for sure that g77 team would not consider this a viable platform (unlike EMX and DJGPP). Nobody wants to work on gcc development snapshots that by definition are unstable and have ever changing interfaces. Regards, Mumit -- khan@xraylith.wisc.edu http://www.xraylith.wisc.edu/~khan/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From xbao@cs.wisc.edu Thu May 1 15:22:00 1997 From: xbao@cs.wisc.edu (Xuemei Bao) Date: Thu, 01 May 1997 15:22:00 -0000 Subject: linking with Tcl/Tk libraries??? References: <9705011846.AA00594@calvin.dgbt.doc.ca> Message-ID: <199705012140.QAA14584@sol1.cs.wisc.edu> hello, Can anyone give me some hints to build up libtcl.a and libtk.a which can be linked into cygwin application? thanks --- Xuemei - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From BrianB@atl.paysys Thu May 1 15:45:00 1997 From: BrianB@atl.paysys (Brian Beuning) Date: Thu, 01 May 1997 15:45:00 -0000 Subject: beta 17.1 errno not always declared in headers Message-ID: <01BC565F.23F10090@BRIANB> It looks like sys/errno.h only defines a value for errno if we are compiling for reentrant code. There are no ifdef's that result in an extern int errno; getting seen by the compiler. Is this right? Thanks, Brian Beuning - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jp@nuancecom.com Thu May 1 18:00:00 1997 From: jp@nuancecom.com (JP Shipherd) Date: Thu, 01 May 1997 18:00:00 -0000 Subject: Codeview debugging info Message-ID: <3.0.32.19970501175947.00a363e0@awesome.nuance.com> Hello, I'm wondering in Cygnus ever plans to offer an option to generate Codeview debugging info (instead of stabs). Windows NT Task Manager lets you click on any process and debug it, however it only brings up codeview (if it is loaded on your system). There are also a variety of products (BoundChecker, etc), that work on code with Codeview debugging info. The Watcom compiler provides this option and it really lets me choose both a superior compiler and my choice of popular debugging tools. Thanks, --jp - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jp@nuancecom.com Thu May 1 18:00:00 1997 From: jp@nuancecom.com (JP Shipherd) Date: Thu, 01 May 1997 18:00:00 -0000 Subject: Some beta 18 details Message-ID: <3.0.32.19970501175945.00a41240@awesome.nuance.com> At 04:24 PM 4/15/97 -0700, Geoffrey Noer wrote: > >Beta 18: changes in specific tools: >----------------------------------- >Winsock is now initialized upon app startup. What version of winsock is being loaded (in the background presumably)? Thanks, --jp (It's early May...pant, pant, salivate, salivate...) - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Fri May 2 00:04:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Fri, 02 May 1997 00:04:00 -0000 Subject: waiting on threads Message-ID: <01BC56E8.6907C6D0@sos> Gregg Jensen wrote: > > WaitForMultipleObjects passing the variable hThread > (which is defined as hThread[7]). The handle from each > CreateThread is put into an empty spot in the hThread > array. When I call WaitForMultipleObjects, I get an > error of INVALID_HANDLE_ERROR (or whatever 6 is). I then Do You check return value of CreateThread()? -- Sergey Okhapkin Moscow, Russia Looking for a job - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From noer@cygnus.com Fri May 2 01:34:00 1997 From: noer@cygnus.com (Geoffrey Noer) Date: Fri, 02 May 1997 01:34:00 -0000 Subject: Codeview debugging info References: <3.0.32.19970501175947.00a363e0@awesome.nuance.com> Message-ID: <199705020834.BAA14402@rtl.cygnus.com> > I'm wondering in Cygnus ever plans to offer an option to generate > Codeview debugging info (instead of stabs). Windows NT Task Manager lets > you click on any process and debug it, however it only brings up codeview > (if it is loaded on your system). [...] There are not currently any plans to do this... -- Geoffrey Noer noer@cygnus.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From noer@cygnus.com Fri May 2 01:43:00 1997 From: noer@cygnus.com (Geoffrey Noer) Date: Fri, 02 May 1997 01:43:00 -0000 Subject: G77 and Gnu-win32 References: <9705012016.AA00653@modi.xraylith.wisc.edu> Message-ID: <199705020843.BAA14483@rtl.cygnus.com> Mumit Khan wrote: [...] > Unless gnu-win32 gcc backend changes are merged back into FSF, I know > for sure that g77 team would not consider this a viable platform > (unlike EMX and DJGPP). Nobody wants to work on gcc development > snapshots that by definition are unstable and have ever changing > interfaces. Cygnus works closely with the FSF to minimize divergence in gcc and other tools. While the gnu-win32 releases of gcc sometimes contain changes not (yet?) approved by the gcc maintainers, the intention is to merge back Cygwin32 backend changes back into FSF sources. The beta 18 release will only contain a couple very minor gcc changes that haven't been merged in, the rest are all in the current FSF development sources so next time there's a public gcc release they should be there. -- Geoffrey Noer noer@cygnus.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Valeri.Faine@cern.ch Fri May 2 03:17:00 1997 From: Valeri.Faine@cern.ch (Valery Fine) Date: Fri, 02 May 1997 03:17:00 -0000 Subject: waiting on threads Message-ID: <199705020851.KAA30844@mail1.cern.ch> On 30 Apr 97 at 11:25, Gregg Jensen wrote: In my mind one told on this list: "The current libraries are SINGLE thread only !!!" Valery ================================================================= Dr. Valery Fine Telex : 911621 dubna su ----------- LCTA/Joint Inst.for NuclearRes Phone : +7 09621 6 40 80 141980 Dubna, Moscow region Fax : +7 09621 6 51 45 Russia mailto:fine@main1.jinr.dubna.su Dr. Valeri Faine ------------ Phone: +41 22 767 6468 CERN FAX : +41 22 767 7910 CH-1211 Geneva, 23 mailto:fine@mail.cern.ch Switzerland http://nicewww.cern.ch/~fine - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From lvarela@prtc.net Fri May 2 03:17:00 1997 From: lvarela@prtc.net (lvarela) Date: Fri, 02 May 1997 03:17:00 -0000 Subject: Web Page update References: <3.0.32.19970501175945.00a41240@awesome.nuance.com> Message-ID: <3369C020.EDE0F16F@prtc.net> Page has been updated... http://www.geocities.com/SiliconValley/Heights/9069/index.html Updated compiler & documentation. Lorenzo - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From dumser@ti.com Fri May 2 06:04:00 1997 From: dumser@ti.com (James Dumser) Date: Fri, 02 May 1997 06:04:00 -0000 Subject: problem using g++ produced program as a WebSite CGI References: <199704301524.LAA17683@kansas.stk.com> Message-ID: <199705021303.IAA23426@lesol1.dseg.ti.com> On Wed, 30 Apr 1997 11:27:06 -0500, Carl Nicol wrote: >I have successfully compiled a cgi program that I use on a Linux system >using g++ on Win/95 and it runs fine from the command line, however, when I >try to use it as a CGI program, it dies with an exception error. I have >WinSite calling it using the ../cgi-shl/ directory. Does anyone have any >clues as to what the problem may be? I believe your problem may be that due to some initialization limitations in cygwin.dll, only one user at a time can run cygwin32 programs -- though I am not sure how this relates to Win95. On NT, your webserver would probably run as SYSTEM while you're logged in as something else, thus triggering the problem. If your webserver runs as a Win95 "service," try hitting the page from another computer without being logged in on the web computer -- does that work? You alsom might able to use the same user for both the webserver and your console login to test this theory out. Depending on your program, a simple solution may be to use Mingw32 (Minimalist GCC, http://www.geocities.com/Tokyo/Towers/6162/gcc.html ). This avoids use of cygwin.dll so you won't hit the initialization problem, but you'll also lose the Unix-emulation environment cygwin.dll provides. -- James Dumser 972-462-5335 dumser@ti.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From greggj@savvis.com Fri May 2 07:23:00 1997 From: greggj@savvis.com (Gregg Jensen) Date: Fri, 02 May 1997 07:23:00 -0000 Subject: waiting on threads References: <01BC56E8.6907C6D0@sos> Message-ID: <3.0.1.32.19970501080006.00945bb0@mail.savvis.com> At 11:02 AM 5/2/97 +0400, Sergey Okhapkin wrote: >Gregg Jensen wrote: >> >> WaitForMultipleObjects passing the variable hThread ... >> error of INVALID_HANDLE_ERROR (or whatever 6 is). I then > >Do You check return value of CreateThread()? Yes. Well, I called GetLastError after each CreateThread and it returned 0. But, maybe there is another way that I am unaware of. The result of GetLastError was 0. Gregg BTW, this is my first forray into the NT/Windows programming world, so I am just learning the Microsoft way. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Sergio.Valle@studi.epfl.ch Fri May 2 07:23:00 1997 From: Sergio.Valle@studi.epfl.ch (Sergio.Valle@studi.epfl.ch) Date: Fri, 02 May 1997 07:23:00 -0000 Subject: Which gcc libraries I must link with my ODBC application ? Message-ID: <199705021109.NAA18773@disun3.epfl.ch> Hello, I try to compile an ODBC application (to access Microsoft Access for Windows 95) in C++ with the gcc compiler but I don't know wich libraries I must link with my source code, does anybody know it ? Do gcc libraries support ODBC 3.0 ? Sergio Valle, Lausanne 1/5/97 (Switzerland) - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Renovexx@dial.pipex.com Fri May 2 07:23:00 1997 From: Renovexx@dial.pipex.com (Glenn Thompson) Date: Fri, 02 May 1997 07:23:00 -0000 Subject: Getting GCC to work Message-ID: <199705021423.PAA27334@typhoon.dial.pipex.net> Can anybody tell me how to get this win32 thig to work right?? Keep getting the following message: ld: cannot open -lgcc: No such file or directory NO IDEA WHAT IS GOING ON HERE. Thanks in advance, Glenn g.t@dial.pipex.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From bje@air.net.au Fri May 2 07:23:00 1997 From: bje@air.net.au (Ben Elliston) Date: Fri, 02 May 1997 07:23:00 -0000 Subject: System V IPC (shared memory) Message-ID: Hi, I am trying to compile a UNIX utility that uses System V IPC. To be more precise, it uses the shm_* shared memory system calls. On most UNIX systems, these functions are prototyped in /usr/include/sys/ipc.h and the code #includes . There is no such header file in the Cygwin headers, so I'm wondering what is the state of these UNIX primitives being emulated on Win32? According to some API documentation I've read, it should be possible to almost directly map shared memory onto shared memory-mapped files in Win32. Can someone fill me in on what's happening in this area? I'm happy to look into it more closely myself, but there's no point if it's being worked on already. Thanks, Ben - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From uwe@genias.de Fri May 2 07:23:00 1997 From: uwe@genias.de (Uwe Wuerfel) Date: Fri, 02 May 1997 07:23:00 -0000 Subject: Is RPC and XDR supported? Message-ID: Hi all there, does anybody know if RPC and maybe XDR is supported by the Cygnus-Compiler and if not, when will it be supported? Thanks for all answers, uwe 8X--------------------------------------------------------------------X8 If you cut here, you will probably destroy your monitor. Uwe Wuerfel Morgenstund... ...kommt selten allein. uwe@genias.de - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From noer@cygnus.com Fri May 2 14:02:00 1997 From: noer@cygnus.com (Geoffrey Noer) Date: Fri, 02 May 1997 14:02:00 -0000 Subject: signal handling References: Message-ID: <199705022100.OAA23938@cirdan.cygnus.com> Uwe Wuerfel wrote: > > Does this mean, that functions like kill() aren't supportet at the moment, > but will be in b18? Yes. Signals in beta 18 won't be perfect by any means but they'll be much improved with respect to previous releases. -- Geoffrey Noer noer@cygnus.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From b.mcleod@opengroup.org Fri May 2 14:19:00 1997 From: b.mcleod@opengroup.org (Bruce D. McLeod) Date: Fri, 02 May 1997 14:19:00 -0000 Subject: beta 17.1 errno not always declared in headers Message-ID: <2.2.32.19970502140143.00942020@postman.osf.org> What does "compiling for reentrant code" mean? Should one include that _REENT_ONLY define for multithread apps? I don't see a libc_r.a to link against. Bruce At 06:40 PM 5/1/97 -0400, Brian Beuning , com@cygnus.com wrote: >It looks like sys/errno.h only defines a value for errno if >we are compiling for reentrant code. There are no ifdef's >that result in an > extern int errno; > >getting seen by the compiler. > >Is this right? > >Thanks, >Brian Beuning > >- >For help on using this list (especially unsubscribing), send a message to >"gnu-win32-request@cygnus.com" with one line of text: "help". > > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From uwe@genias.de Fri May 2 14:19:00 1997 From: uwe@genias.de (Uwe Wuerfel) Date: Fri, 02 May 1997 14:19:00 -0000 Subject: signal handling References: <199704302016.NAA14175@cirdan.cygnus.com> Message-ID: > Signals are almost not implemented in beta 17.1 so it's not > surprising that you're having troubles here. Once beta 18 comes out > (still shooting for early May), you may well have much better luck since > we've put a good bit of effort in this area in the last couple of months... > Does this mean, that functions like kill() aren't supportet at the moment, but will be in b18? uwe. 8X--------------------------------------------------------------------X8 If you cut here, you will probably destroy your monitor. Uwe Wuerfel Morgenstund... ...kommt selten allein. uwe@genias.de - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From noer@cygnus.com Fri May 2 14:19:00 1997 From: noer@cygnus.com (Geoffrey Noer) Date: Fri, 02 May 1997 14:19:00 -0000 Subject: Unix IPC support? References: <33676058.799D@byu.edu> Message-ID: <199705022117.OAA23971@cirdan.cygnus.com> Kent Anderson wrote: > > Does CYGWIN32 support unix style ipc: message queues, semaphores > and shared memory? Nope, not yet anyhow... -- Geoffrey Noer noer@cygnus.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From dbakin@intrinsa.com Fri May 2 16:00:00 1997 From: dbakin@intrinsa.com (Dave Bakin) Date: Fri, 02 May 1997 16:00:00 -0000 Subject: Codeview debugging info Message-ID: <2.2.32.19970502142515.00acb61c@intrinsa.com> Although the system will automatically bring up one debugger, you get to pick the debugger. On Windows 95 look in win.ini for [aedebug]. On Windows NT look in the registry for an aedebug key. Look at the knowledge base articles at http://www.microsoft.com/kb/articles/q103/8/61.htm and http://www.microsoft.com/kb/articles/q138/7/86.htm . So if you want to always launch gdb, try setting up the system for it and see if it works. (Let us know!) -- Dave At 05:59 PM 5/1/97 -0700, you wrote: >Hello, > > I'm wondering in Cygnus ever plans to offer an option to generate >Codeview debugging info (instead of stabs). Windows NT Task Manager lets >you click on any process and debug it, however it only brings up codeview >(if it is loaded on your system). > > There are also a variety of products (BoundChecker, etc), that work on >code with Codeview debugging info. The Watcom compiler provides this >option and it really lets me choose both a superior compiler and my choice >of popular debugging tools. > >Thanks, >--jp > >- >For help on using this list (especially unsubscribing), send a message to >"gnu-win32-request@cygnus.com" with one line of text: "help". > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From scoleman@sewp.nasa.gov Fri May 2 16:00:00 1997 From: scoleman@sewp.nasa.gov (Steve Coleman) Date: Fri, 02 May 1997 16:00:00 -0000 Subject: NT services and gnu-win32 Message-ID: <199705022019.QAA15288@mushroom.sewp.nasa.gov> Hi, I just downloaded the cygwin suite and ported my first Unix source code to NT yesterday and I was quite impressed with the ease at which I was able to do it. The program (a derivative of the TIS toolkit "smap") works fine from the command line but needs to run as a daemon on the system to be useful. I then tried to use the NT resource kit (srvany.exe) to load the application as an NT service. Everything works fine until someone logs out of the system at which time NT (sends messages? WM_ENDSESSION and CTRL_LOGOFF_EVENT) kills it. If the service is loaded when the machine boots the process still dies when the first person logs out. Even the cygwin nohup.exe does not help. Questions: Is there any way to trap/ignore this action within a cygwin application? Compile options? Runtime signal processing? Special loading instructions for a service? Thanks. Steve Coleman scoleman@sewp.nasa.gov vox: 301.286.7636 fax: 301.286.0317 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From lehman@theonramp.net Fri May 2 16:00:00 1997 From: lehman@theonramp.net (Lehmans) Date: Fri, 02 May 1997 16:00:00 -0000 Subject: directx with gnu-win32 Message-ID: <19970502230101777.AAA201@lehman> i'm sorry if you've already discussed this, but i hadn't downloaded this software yet...could anyone please tell me if it is possible to get DirectX to work with gnu-win32?I've had bad experiences with the import library someone converted...the function CreateSurface won't work. Please respond directly to my email-address since i haven't subscribed to the mailing list. lehman@theonramp.net - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From hemal@superlink.com Fri May 2 16:41:00 1997 From: hemal@superlink.com (Hemal Pandya) Date: Fri, 02 May 1997 16:41:00 -0000 Subject: 'ls /' does not work Message-ID: I am using beta17.1. running on Windows 95 'ls /' gives no output, either from within bash or at command.com prompt. Similarly pressing tab after / also provides no completions. mount says c: is mounted on /, which is what is is in the registry too. If I delete the registry entry, both mount and bash restore the entry. Even if I change the registry entry to c:\ or c:\\, bash changes it to c:. After providing a directory name under root, for example /gnu-win32/ filename completion works, as also ls. Any ideas? TIA, -- Hemal Pandya mailto:hemal@superlink.net phone:(908)340-9455 First He created computers, but they couldn't make mistakes So He created programmers - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Rich.Kucera@ccmail.irs.gov Fri May 2 16:41:00 1997 From: Rich.Kucera@ccmail.irs.gov (Rich Kucera) Date: Fri, 02 May 1997 16:41:00 -0000 Subject: C indentation not working on emacs?? Message-ID: <0009BC51.1950@ccmail.irs.gov> Thank you both. Now I know I can get emacs for NT. (Of course I already knew this from a dream I had...) Now I know to read cc-mode.el when I get it. (Of course I already knew this from a past life centuries ago...) Learn something new every day :) ______________________________ Reply Separator _________________________________ Subject: Re: C indentation not working on emacs?? Author: Jim Balter at Internet Date: 4/30/97 7:23 PM Peter Craft wrote: > > I'm sure that I'm doing something wrong.. > > I'm running "Emacs for Windows NT and Windows 95 Version 19.34.1" > and I'm trying to customize my C indentation environment. > The problem is that Emacs doesn't seem to recognize > the standard C indentation variables. These include > c-indent-level, c-continued-statement-level, c-brace-offset, > etc.. > > If I execute set-variable c-indent-level from a running > emacs, it says "no match". If I try setting it with a setq > in my _emacs it has no effect. > > Oddly, appropos lists these variables despite the fact that > set-variable disagrees. > > As always, thanks for your help. What would help is to limit discussion on this mailing list to topics that are at least vaguely related to GNU-win32. Not only isn't ntemacs a GNU-win32 program, but your problem isn't even specific to ntemacs; you are simply centuries out of date. Read cc-mode.el in your lisp directory. -- - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From smueller@microsoft.com Fri May 2 16:41:00 1997 From: smueller@microsoft.com (Stephan Mueller) Date: Fri, 02 May 1997 16:41:00 -0000 Subject: Codeview debugging info Message-ID: <9106B0327B3ACF11ACEF00805FD47A0B02D8B0AA@RED-67-MSG.dns.microsoft.com> Actually, NT will bring up the debugger registered in the registry. On a 32-bit system, it's unlikely that is ever CodeView. HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows NT\CurrentVersion\AeDebug\Debugger is the debugger-command-with-arguments to launch when you select Debug from the Task Manager (or Just In Time debugging at a GP fault). You can change this value. On my machine, the value is d:\msdev\bin\msdev.exe -p %ld -e %ld I just replaced it temporarily with a batch file that simply prints out the two %ld values. It appears that the presence of the -p %ld -e %ld parameters is ignored, and "-p " is always added, at least when using Debug from Task Manager. So if you hack your debugger to take -p and a process id if it doesn't already, you may be able to get useful results by playing with this registry key. stephan(not speaking in any official capacity, just for myself. Hacking your registry is dangerous. Do so at your own risk.); > -----Original Message----- > From: Geoffrey Noer [SMTP:noer@cygnus.com] > Sent: Friday, May 02, 1997 1:34 AM > To: jp@nuancecom.com > Cc: noer@cygnus.com; gnu-win32@cygnus.com > Subject: Re: Codeview debugging info > > > I'm wondering in Cygnus ever plans to offer an option to generate > > Codeview debugging info (instead of stabs). Windows NT Task Manager > lets > > you click on any process and debug it, however it only brings up > codeview > > (if it is loaded on your system). > [...] > > There are not currently any plans to do this... > > -- > Geoffrey Noer > noer@cygnus.com > - > For help on using this list (especially unsubscribing), send a message > to > "gnu-win32-request@cygnus.com" with one line of text: "help". - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From irio.lavagno@torino.alpcom.it Fri May 2 16:41:00 1997 From: irio.lavagno@torino.alpcom.it (Irio Lavagno) Date: Fri, 02 May 1997 16:41:00 -0000 Subject: building a cross compiler Message-ID: <336A194E.2882315F@torino.alpcom.it> I tried to build a cross compiler version of cygwin32 on my linux box everything works well until it tries to build the utils, at this point it complains as follows configuring in utils running /bin/sh ./configure --host=i386-cygwin32 --with-cross-host=i586-pc-linux-gnu --with-target-subdir=i386-cygwin32 --cache-file=.././config.cache --srcdir=. loading cache .././config.cache checking for gcc... (cached) /usr/local/src/cygwin32/gcc/xgcc -B/usr/local/src/cygwin32/gcc/ -idirafter /usr/local/src/cygwin32/i386-cygwin32/newlib/targ-include -idirafter /usr/local/src/cygwin32/newlib/libc/include -nostdinc checking whether we are using GNU C... (cached) yes configure: error: can not find install-sh or install.sh in . ./.. ./../.. configure: error: ./configure failed for utils make: *** [configure-target-winsup] Error 1 I've been able to make it compile anyway copying the install.sh in one of the dirs it looks in, but I'm not sure this is the best way to solve the problem. The command line I use to build this cross-compiler is (as stated by Geoffrey Noer in one of his answers): configure --target=i386-cygwin32 Configure completes successfully, but make fails after a while with the above message. Thanks for your help Irio Lavagno - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Fri May 2 23:54:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Fri, 02 May 1997 23:54:00 -0000 Subject: Getting GCC to work Message-ID: <01BC57B0.9F360730@sos> Glenn Thompson wrote: > Can anybody tell me how to get this win32 thig to work right?? > > Keep getting the following message: > > ld: cannot open -lgcc: No such file or directory > Did You set GCC_EXEC_PREFIX in environment right? -- Sergey Okhapkin Moscow, Russia Looking for a job - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From bje@air.net.au Sat May 3 03:31:00 1997 From: bje@air.net.au (Ben Elliston) Date: Sat, 03 May 1997 03:31:00 -0000 Subject: Is RPC and XDR supported?5 References: Message-ID: > does anybody know if RPC and maybe XDR is supported by the Cygnus-Compiler > and if not, when will it be supported? Err, can't it be done within a user space library (say, called librpc.a)? Cheers, Ben --- Ben Elliston "For my birthday I got a humidifier and a de-humidifier. I put them in the same room and let them fight it out." -- Steven Wright - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From bje@air.net.au Sat May 3 03:31:00 1997 From: bje@air.net.au (Ben Elliston) Date: Sat, 03 May 1997 03:31:00 -0000 Subject: System V IPC (shared memory) References: <01BC573F.5E56E340@pc.ice.com> Message-ID: On Fri, 2 May 1997, Tim Endres wrote: > Will be provided in B18 in a few weeks. Can someone confirm this please? (I ask only because someone said it wasn't even on the drawing board!) Thanks, Ben --- Ben Elliston "For my birthday I got a humidifier and a de-humidifier. I put them in the same room and let them fight it out." -- Steven Wright - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From bje@air.net.au Sat May 3 03:31:00 1997 From: bje@air.net.au (Ben Elliston) Date: Sat, 03 May 1997 03:31:00 -0000 Subject: getpwuid Message-ID: Does anyone know what the state of the implementation of getpwuid is? Is it possible to have the structure filled in by this function completed using details from the NT user database? I have some code which seems to be getting (null) when doing: getpwuid(geteuid()) Any ideas? Thanks, Ben --- Ben Elliston "For my birthday I got a humidifier and a de-humidifier. I put them in the same room and let them fight it out." -- Steven Wright - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From bje@air.net.au Sat May 3 06:18:00 1997 From: bje@air.net.au (Ben Elliston) Date: Sat, 03 May 1997 06:18:00 -0000 Subject: Performance Message-ID: Someone on another mailing list suggested the applications ported to Win32 using Cygwin are not up to scratch. I don't understand why this might be: the compiler produces native x86 (and presumably very good!) code for the user space code and for UNIX system calls, they would be handled by the DLL with a secondary call into the Windows kernel. Does anyone have any performance comparisons between "Cygwin" apps and native Win32 applications? Is there any real performance penalty for software with a high proportion of system calls? Thanks, Ben --- Ben Elliston "For my birthday I got a humidifier and a de-humidifier. I put them in the same room and let them fight it out." -- Steven Wright - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From fkittred@gwi.net Sat May 3 06:18:00 1997 From: fkittred@gwi.net (Fletcher E Kittredge) Date: Sat, 03 May 1997 06:18:00 -0000 Subject: Is RPC and XDR supported? References: Message-ID: <199705031318.JAA22183@river.gwi.net> On Fri, 2 May 1997 16:04:46 +0100 "Uwe Wuerfel" wrote: > does anybody know if RPC and maybe XDR is supported by the Cygnus-Compiler > and if not, when will it be supported? Neither require compiler support. regards, fletcher - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From m.gregory@rmit.edu.au Sat May 3 06:18:00 1997 From: m.gregory@rmit.edu.au (Mark A Gregory) Date: Sat, 03 May 1997 06:18:00 -0000 Subject: Link problems Message-ID: <01BC57F0.52480370@rem18.co.rmit.edu.au> Hi, my link seems to be getting better, I can link with libiberty.a and this removes the random/srandom problem, however as can be seen below there are still some problems. Does anyone know how to get rid of the multiple reference to impure.c? This seems to be in libcygwin.a(libccrt0.o) and libc.a(impure.c). also in libc.a there is an undefined reference to __progname in getopt.c What do I do about this? My other problems are application specific and I need assistance from the wolfpack. thank you gcc -o emp_server main.o idle.o shutdown.o update.o marketup.o timestamp.o lostitem.o /empire/emp4/lib/libplayer.a /empire/emp4/lib/ libupdate.a /empire/emp4/lib/libcommands.a /empire/emp4/lib/libsubs.a /empire/emp4/lib/libcommon.a /empire/emp4/lib/libgen.a /empire /emp4/lib/libglobal.a /empire/emp4/lib/libas.a /empire/emp4/lib/libempth.a /empire/emp4/lib/liblwp.a -lm -lc -liberty e:\cygnus\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32/cygnus-2.7.2-961023/../../../../i386-cygwin32/lib/libcygwin.a(libccrt0.o)(.data+ 0x0):libccrt0.cc: multiple definition of `_impure_ptr' e:\cygnus\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32/cygnus-2.7.2-961023/../../../../i386-cygwin32/lib/libc.a(impure.o)(.data+0x2ec): impure.c: first defined here /empire/emp4/lib/libcommands.a(info.o)(.text+0x552):info.c: undefined reference to `alphasort' /empire/emp4/lib/libcommands.a(info.o)(.text+0x56c):info.c: undefined reference to `scandir' /empire/emp4/lib/libgen.a(io.o)(.text+0xc33):io.c: undefined reference to `shutdown' /empire/emp4/lib/libgen.a(io.o)(.text+0xc5c):io.c: undefined reference to `shutdown' /empire/emp4/lib/liblwp.a(lwp.o)(.text+0x39b):lwp.c: undefined reference to `sigmask' /empire/emp4/lib/liblwp.a(lwp.o)(.text+0x65e):lwp.c: undefined reference to `lwpInitContext' e:\cygnus\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32/cygnus-2.7.2-961023/../../../../i386-cygwin32/lib/libc.a(getopt.o): In function `getopt': /pizza/mushroom/noer/beta17/src/newlib/libc/stdlib/getopt.c:90: undefined reference to `__progname' /pizza/mushroom/noer/beta17/src/newlib/libc/stdlib/getopt.c:107: undefined reference to `__progname' make: *** [all] Error 1 make: *** [binaries] Error 2 make: *** [win32] Error 2 Mark A Gregory Senior Lecturer Department of Communication and Electronic Engineering Royal Melbourne Institute of Technology - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From root@jacob.remcomp.fr Sat May 3 06:18:00 1997 From: root@jacob.remcomp.fr (root) Date: Sat, 03 May 1997 06:18:00 -0000 Subject: new version of lcc-win32 available Message-ID: There is a new version of lcc-win32 available at http://www.remcomp.com/lcc-win32 This version features: o full support of the mmx instruction set through intrinsics o new higher level optimizer improves speed and code size o many bugs fixes for the utilities/debugger, etc lcc-win32 is another compiler system. It is distributed with source code, and a complete system for programming under windows. There is NO UNIX support though. The zipped archive with everything (compiler+headers+libraries/source code) is around 1,3MB. enjoy! -- Jacob Navia Logiciels/Informatique 41 rue Maurice Ravel Tel 01 48.23.51.44 93430 Villetaneuse Fax 01 48.23.95.39 France - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From lehman@theonramp.net Sat May 3 10:32:00 1997 From: lehman@theonramp.net (Lehmans) Date: Sat, 03 May 1997 10:32:00 -0000 Subject: directx with gnu-win32 Message-ID: <19970503173250640.AAA233@lehman> please help! i really need to now if, and how you can use directx, specifically directdraw with gnu-win32. i've had really bad experiences with an import library, because some functions,such as the direct draw create surface functions, return error messages. Please respond directly to my email address, because i haven't subscribed to the mailing list yet. lehman@theonramp.net - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From voline@geocities.com Sat May 3 12:19:00 1997 From: voline@geocities.com (voline@geocities.com) Date: Sat, 03 May 1997 12:19:00 -0000 Subject: NT services and gnu-win32 Message-ID: <1.5.4.32.19970503192343.008d6710@mail.geocities.com> At 04:19 PM 05/02/97 -0400, scoleman@sewp.nasa.gov (Steve Coleman) wrote: >Hi, > >I just downloaded the cygwin suite and ported my first Unix source code to NT >yesterday and I was quite impressed with the ease at which I was able to do >it. The program (a derivative of the TIS toolkit "smap") works fine from the >command line but needs to run as a daemon on the system to be useful. I then >tried to use the NT resource kit (srvany.exe) to load the application as an NT >service. Everything works fine until someone logs out of the system at which >time NT (sends messages? WM_ENDSESSION and CTRL_LOGOFF_EVENT) kills it. If the >service is loaded when the machine boots the process still dies when the first >person logs out. Even the cygwin nohup.exe does not help. This is a known problem with srvany, not Cygwin. From srvany.wri: " Programming Considerations: For WIN32 graphical applications: when the currently logged-in user is logging-off, all WIN32 top-level windows receive WM_QUERYENDSESSION and WM_ENDSESSION messages. Some WIN32 applications choose to terminate upon receipt of such messages. In order for your WIN32 application to survive logoff, it must not do that: instead, your windows procedure should call the default windows procedure on these messages. For WIN32 Console (i.e. character-mode) applications: when the currently logged-in user is logging-off, all Console applications receive a CTRL_LOGOFF_EVENT event from the Console. If your Console application has registered a Console event handler (via SetConsoleCtrlHandler), it must ignore CTRL_LOGOFF_EVENT in order to survive the logoff. " Either try registering a Console event handler or rewrite the code as a service (not a trivial task). - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From noer@cygnus.com Sat May 3 13:28:00 1997 From: noer@cygnus.com (Geoffrey Noer) Date: Sat, 03 May 1997 13:28:00 -0000 Subject: getpwuid References: Message-ID: <199705032028.NAA22625@rtl.cygnus.com> > Does anyone know what the state of the implementation of getpwuid is? Is > it possible to have the structure filled in by this function completed > using details from the NT user database? In beta 18 we add: Under NT only, chown(), getgrgid(), getgrnam(), endgrent(), getgrent(), setpwend(), getpwent(), endpwent(). Win95 still has these as stubs. You will need to make an /etc/passwd and /etc/group for some calls to work. That said, we provide a mkpasswd and mkgroup that create valid /etc files from the NT user database. It also looks like getpwuid will work in beta 18 so you may be in luck. I expect b18 will be out this coming week, perhaps on Tues/Wed? -- Geoffrey Noer noer@cygnus.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jcassidy@stratos.net Sat May 3 18:05:00 1997 From: jcassidy@stratos.net (James M. J. Cassidy) Date: Sat, 03 May 1997 18:05:00 -0000 Subject: Linux->Win32 Cross Compiler Message-ID: Has anyone built a cross compiler to run on a linux box and build win32 files and put it somewhere where someone like me can download it? - Jim - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From heric@psu.edu Sat May 3 19:44:00 1997 From: heric@psu.edu (eric hilton) Date: Sat, 03 May 1997 19:44:00 -0000 Subject: termcap/curses Message-ID: <336BED9D.4E94@psu.edu> I compiled rogue985 after a few small source edits ($(CC) and termcap dir) it compiled clean except for some signal() warnings. It seems to start fine and shows the intro map but it doesnt seem to accept any kind of keyboard input and/or it isn't updating the screen. I noticed the default (and only) termcap entry is pretty sparse here it is: ansi:all we know about:\ :co#80:li#24:cl=50\E[;H\E[2J:bs:am:cm=\E[%i%d;%dH:\ :nd=\E[C:up=\E[A:ce=\E[K:ho=\E[H:pt: Wondering if anyone happens to know what the problem may be, and also if there is any more functional termcaps. Another thing I changed was to add another entry 'dumb:' rogue seemed to want it. Other info: Compiling on a Win95 machine, with the built-in curses functions that the author of rogue re-wrote for people without curses. Also it doesnt accept ^C to break out. (ie ^C doesnt break it, i have to close it from win95) It may also be a curses incompatibility. Has a working ncurses been ported with cygwin. Sorry if any of this is newbie, I read the faqs, really I did. -Eric - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From ebritten@ea.oac.uci.edu Sat May 3 19:44:00 1997 From: ebritten@ea.oac.uci.edu (Eric Britten) Date: Sat, 03 May 1997 19:44:00 -0000 Subject: Exporting __stdcall Procs without _ and @n References: <0005205E.1893@po.cle.ab.com> Message-ID: On Thu, 1 May 1997, Scott Mintz wrote: > Did you ever get this working? > > -Scott A. Mintz > I finally got this to work with gcc. In your def file include the following for each function that you don't want mangled. unmangled_name=mangled_name@n For my case it wa CPlApplet=CPlApplet@16 I built my dll, renamed to cpl, then ran it and it worked since Windows could find the right name. Thanks to all that helped me. Eric - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From masplund@cs.Helsinki.FI Sun May 4 03:45:00 1997 From: masplund@cs.Helsinki.FI (aspa) Date: Sun, 04 May 1997 03:45:00 -0000 Subject: getpwuid References: Message-ID: On Sat, 3 May 1997, Ben Elliston wrote: > Does anyone know what the state of the implementation of getpwuid is? Is > it possible to have the structure filled in by this function completed > using details from the NT user database? > > I have some code which seems to be getting (null) when doing: > > getpwuid(geteuid()) > > Any ideas? i recall successfully using getpwuid() in b17.1. for the above example to work, you have to have a /etc/passwd entry for the euid running the code. -- aspa - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Rolf.Wester@t-online.de Sun May 4 03:45:00 1997 From: Rolf.Wester@t-online.de (Rolf Wester) Date: Sun, 04 May 1997 03:45:00 -0000 Subject: dll's and LoadLibrary on NT and Win95 Message-ID: Hi, I tried to make a dll and an *.exe file which dynamically loads this dll. I made this according to Ismael Jurado's suggestions on a WindowsNT and a Windows95 system. All works well on the NT system but not on the Win95 system. I tried combinations of the dll and the exe made on the NT and the Win95 system respectively. The WinNT and Win95 exe's work with the NT dll but both exe's do not work with the Win95 dll. The sources, the shell scripts for making the executables and the results are in the file attached to this mail. The problem with dynamically loading dll's not only occured with the I. Jurados example but it also was present when trying to use c/c++ code from a java class via the Java Native Interface. I also want to try using Tcl/Tk and c/c++ together and there the same problem could occur. I would be very appreciative for any help on this problem. Thank you anticipation. With kind regards --Rolf wester@ilt.fhg.de From Ulrich.Lauther@mchp.siemens.de Sun May 4 06:53:00 1997 From: Ulrich.Lauther@mchp.siemens.de (Ulrich Lauther) Date: Sun, 04 May 1997 06:53:00 -0000 Subject: memory.h causes problems Message-ID: <199705041352.PAA10444@fubini.mchp.siemens.de> #include main () { } when I try to compile the above program, stored as test.cc, I get a heap of error messages. Part of the problem is that memory.h includes string.h which gets confused with String.h. This goes away if I use an absolute pathname in memory.h for the inclusion of string.h. The part I do not understand are error messages due to multiple definitions in String.h, which however are legal C++ overloadings. Any hints how to explain and fix the problem? -ulrich gcc -v -c test.cc Reading specs from C:\cygnus\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32/cygnus-2.7.2-961023/specs gcc version cygnus-2.7.2-961023 C:\cygnus\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32/cygnus-2.7.2-961023/cpp.exe -lang-c++ -v -iprefix C:\cygnus\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32/cygnus-2.7.2-961023/ -undef -D__GNUC__=2 -D__GNUG__=2 -D__cplusplus -D__GNUC_MINOR__=7 -Di386 -D_WIN32 -DPOSIX -D__CYGWIN32__ -DWINNT -D_X86_=1 -D__STDC__=1 -D__stdcall=__attribute__((__stdcall__)) -D__cdecl=__attribute__((__cdecl__)) -D__i386__ -D_WIN32 -D__POSIX__ -D__CYGWIN32__ -D__WINNT__ -D_X86_=1 -D__STDC__=1 -D__stdcall=__attribute__((__stdcall__)) -D__cdecl=__attribute__((__cdecl__)) -D__i386 -D__POSIX -D__WINNT -Asystem(winnt) -Acpu(i386) -Amachine(i386) -Di386 -Asystem(unix) -Acpu(i386) -Amachine(i386) -D__i386__ -Asystem(unix) -Acpu(i386) -Amachine(i386) test.cc C:\WINDOWS\TEMP/cc001333.ii GNU CPP version cygnus-2.7.2-961023 (80386, BSD syntax) #include "..." search starts here: #include <...> search starts here: C:\cygnus\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32/cygnus-2.7.2-961023/../../../../../include/g++ C:\cygnus\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32/cygnus-2.7.2-961023/../../../../../include C:\cygnus\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32/cygnus-2.7.2-961023/../../../../i386-cygwin32/include C:\cygnus\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32/cygnus-2.7.2-961023/include End of search list. C:\cygnus\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32/cygnus-2.7.2-961023/cc1plus.exe C:\WINDOWS\TEMP/cc001333.ii -quiet -dumpbase test.cc -version -o C:\WINDOWS\TEMP/cc001333.s GNU C++ version cygnus-2.7.2-961023 (i386-cygwin32) compiled by GNU C version cygnus-2.7.2-961023. In file included from C:\\cygnus\\H-i386-cygwin32\\lib\\gcc-lib\\i386-cygwin32/cygnus-2.7.2-961023/../../../../i386-cygwin32/include/memory.h:4, from test.cc:1: C:\\cygnus\\H-i386-cygwin32\\lib\\gcc-lib\\i386-cygwin32/cygnus-2.7.2-961023/../../../../../include/g++/string.h:45: declaration of C function `struct StrRep * Scat(struct StrRep *, const char *, int, const char *, int, const char *, int)' conflicts with C:\\cygnus\\H-i386-cygwin32\\lib\\gcc-lib\\i386-cygwin32/cygnus-2.7.2-961023/../../../../../include/g++/string.h:44: previous declaration `struct StrRep * Scat(struct StrRep *, const char *, int, const char *, int)' here C:\\cygnus\\H-i386-cygwin32\\lib\\gcc-lib\\i386-cygwin32/cygnus-2.7.2-961023/../../../../../include/g++/string.h:160: declaration of C function `void cat(const class String &, const class SubString &, class String &)' conflicts with C:\\cygnus\\H-i386-cygwin32\\lib\\gcc-lib\\i386-cygwin32/cygnus-2.7.2-961023/../../../../../include/g++/string.h:159: previous declaration `void cat(const class String &, const class String &, class String &)' here C:\\cygnus\\H-i386-cygwin32\\lib\\gcc-lib\\i386-cygwin32/cygnus-2.7.2-961023/../../../../../include/g++/string.h:160: warning: `cat(const String &, const SubString &, String &)' is already a friend of class `String' C:\\cygnus\\H-i386-cygwin32\\lib\\gcc-lib\\i386-cygwin32/cygnus-2.7.2-961023/../../../../../include/g++/string.h:160: warning: previous friend declaration of `cat(const String &, const SubString &, String &)' C:\\cygnus\\H-i386-cygwin32\\lib\\gcc-lib\\i386-cygwin32/cygnus-2.7.2-961023/../../../../../include/g++/string.h:161: declaration of C function `void cat(const class String &, const char *, class String &)' conflicts with C:\\cygnus\\H-i386-cygwin32\\lib\\gcc-lib\\i386-cygwin32/cygnus-2.7.2-961023/../../../../../include/g++/string.h:160: previous declaration `void cat(const class String &, const class SubString &, class String &)' here C:\\cygnus\\H-i386-cygwin32\\lib\\gcc-lib\\i386-cygwin32/cygnus-2.7.2-961023/../../../../../include/g++/string.h:161: warning: `cat(const String &, const char *, String &)' is already a friend of class `String' C:\\cygnus\\H-i386-cygwin32\\lib\\gcc-lib\\i386-cygwin32/cygnus-2.7.2-961023/../../../../../include/g++/string.h:161: warning: previous friend declaration of `cat(const String &, const char *, String &)' C:\\cygnus\\H-i386-cygwin32\\lib\\gcc-lib\\i386-cygwin32/cygnus-2.7.2-961023/../../../../../include/g++/string.h:162: declaration of C function `void cat(const class String &, char, class String &)' conflicts with C:\\cygnus\\H-i386-cygwin32\\lib\\gcc-lib\\i386-cygwin32/cygnus-2.7.2-961023/../../../../../include/g++/string.h:161: previous declaration `void cat(const class String &, const char *, class String &)' here [further error messages deleted] -- -lauther ---------------------------------------------------------------------------- Ulrich Lauther ph: +49 89 636 48834 fx: ... 636 42284 Siemens ZT AN 1 Internet: Ulrich.Lauther@mchp.siemens.de - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From lehman@theonramp.net Sun May 4 10:59:00 1997 From: lehman@theonramp.net (Lehmans) Date: Sun, 04 May 1997 10:59:00 -0000 Subject: directx with gnu-win32 Message-ID: <19970504180013597.AAA195@lehman> please help! does anyone know if and how you can use directx with gnu-win32, specifically direct draw....i've had bad experiences with the import library at cygnus.com....the create_surface function returns an error...please respond directly to my email address since i haven't subscribed to the mailing list yet. lehman@theonramp.net - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From stephenc@wf.net Sun May 4 15:43:00 1997 From: stephenc@wf.net (Stephen Crowley) Date: Sun, 04 May 1997 15:43:00 -0000 Subject: Global Constructors, Objc, and dlls Message-ID: <199705042225.RAA28131@odin.wf.net> Has anyone tried using objective-c with gnuwin32, I downloaded the libobjc.a source and compiled it, linked it as a dll and everything works perfectly, but when i try making a dll with objecitve-c code in it it seems to create an invalid dll, says "Cant start program" or something to that effect. I am thinking it has something to do with global constructors, or relocations or something weird like that, the debugger is of no use because the exe will never start, it page faults in kernel32.dll. Please dont tell me to just build them as a static lib becuase i know its possible to have objc code in a dll, libobjc.a has .m code in it and it works fine, but if i build my own object it creates an invalid dll. Ive been working on this for a week and its really frustrating me, if anyone can help I'd really appreciate it. --Stephen Crowley - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From lehman@theonramp.net Sun May 4 15:43:00 1997 From: lehman@theonramp.net (Lehmans) Date: Sun, 04 May 1997 15:43:00 -0000 Subject: directx with gnu-win32 Message-ID: <19970504224356805.AAA208@lehman> please respond! i really need anybody's help that knows how and if you can use directx with gnu-win32.....please don't direct me to cygnus@dragon....i've already been there and it doesn't work.....directdraw's createsurface screws up. please respond directly to my email address since i haven't subscribed to the mail list yet... lehman@theonramp.net - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From noer@cygnus.com Sun May 4 16:08:00 1997 From: noer@cygnus.com (Geoffrey Noer) Date: Sun, 04 May 1997 16:08:00 -0000 Subject: termcap/curses References: <336BED9D.4E94@psu.edu> Message-ID: <199705042307.QAA13053@rtl.cygnus.com> eric hilton wrote: [...] > Wondering if anyone happens to know what the problem may be, and also if > there is any more functional termcaps. Another thing I changed was to > add another entry 'dumb:' rogue seemed to want it. [...] > It may also be a curses incompatibility. Has a working ncurses been > ported with cygwin. [...] Upcoming beta 18 which should be out in just a few more days will have better console handling in the cygwin.dll. A Linux termcap file is included so you should be able to specify "linux" as your terminal type and have more things work. People have gotten ncurses to work with beta 17.1 and earlier so that's an option as well. -- Geoffrey Noer noer@cygnus.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From noer@cygnus.com Sun May 4 18:27:00 1997 From: noer@cygnus.com (Geoffrey Noer) Date: Sun, 04 May 1997 18:27:00 -0000 Subject: Performance References: Message-ID: <199705050126.SAA13688@rtl.cygnus.com> Ben Elliston wrote: > > Someone on another mailing list suggested the applications ported to Win32 > using Cygwin are not up to scratch. [...] > I don't understand why this might be: the compiler produces native x86 > (and presumably very good!) code for the user space code and for UNIX > system calls, they would be handled by the DLL with a secondary call into > the Windows kernel. I think it really depends on what you're trying to achieve. Let me see if I can explain what I think is going on performance-wise. The cygwin.dll is a porting library that includes libc and libm as well as a subset of standard POSIX and misc other common Unix library calls. The intention is that the Unix semantics will be preserved such that porting Unix apps will be as painless as possible. Some standard Unix API calls (fork(), for example) are painfully slow under Cygwin32 because the Win32 API differs a lot with respect to the Unix API in terms of what's optimized/easily possible. In addition to some specific calls being slow, there is a certain amount of overhead associated with the cygwin.dll because it has to maintain its own process table, file descriptors, etc... In terms of raw execution speed, it is almost certainly better to rewrite Unix software entirely for Win32 than use Cygwin32. However, execution speed isn't the only consideration. You'd lose out on the many advantages of having a single source base and would have to put in a lot of additional time and effort as well. So there are inherent performance problems with using a Unix portability library to port Unix code to Win32. The other factor is that we have been concentrating our efforts on correctness rather than on optimizing performance. I expect performance will become more of an issue as the Cygwin32 library matures. Now what about performance problems when doing plain old native Win32 programming? Unless you use the Minimalist GNU Win32 package from Colin Peters (see http://www.fu.is.saga-u.ac.jp/~colin/gcc.html ), which allows you to use Microsoft libraries instead of the Cygwin32 library, you automatically use libc/libm in the cygwin.dll and inherit lots of unnecessary Unix baggage. This slows things down unnecessarily. I have every reason to believe that if you use the Minimalist GNU Win32 package for native development, you will see similar (or perhaps even better?) runtime performance than with VC++ or another Win32 development tools package. -- Geoffrey Noer noer@cygnus.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From hawkeye@CommerceWave.com Sun May 4 23:48:00 1997 From: hawkeye@CommerceWave.com (Ken Keys) Date: Sun, 04 May 1997 23:48:00 -0000 Subject: getpwuid References: <199705032028.NAA22625@rtl.cygnus.com> Message-ID: <199705050657.XAA04213@dilbert.commercewave.com> Geoffrey Noer wrote: > In beta 18 we add: > > Under NT only, chown(), getgrgid(), getgrnam(), endgrent(), getgrent(), > setpwend(), getpwent(), endpwent(). Win95 still has these as stubs. > > You will need to make an /etc/passwd and /etc/group for some > calls to work. That said, we provide a mkpasswd and mkgroup that > create valid /etc files from the NT user database. Wouldn't it make more sense to have those functions operate directly on the NT user DB, instead of on "unixified" copies of them? The cygwin copy will get out of sync and need to be rebuilt whenever the native copy is changed. The pw API was explicitly designed to *avoid* dependance on a particluar storage format; cygwin should take advantage of that and not require users to maintain artificial copies. One could argue that having a unix-style /etc/passwd would allow the working of programs that read /etc/passwd directly instead of going through the pw API. But, 1) such programs are already nonportable, since not even all unix-like systems use the same format for /etc/passwd, and 2) IMO, supporting badly written software isn't worth the trouble it will cause for users when their user DB's get out of sync. (BTW, I don't use NT myself, just thought I'd point out [what I consider to be] a mistake before it happened). -- Ken Keys (hawkeye@CommerceWave.com) Software Engineering Team Leader CommerceWave, Inc. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From uwe@genias.de Mon May 5 04:03:00 1997 From: uwe@genias.de (Uwe Wuerfel) Date: Mon, 05 May 1997 04:03:00 -0000 Subject: Is RPC and XDR supported?5 References: Message-ID: > > does anybody know if RPC and maybe XDR is supported by the Cygnus-Compiler > > and if not, when will it be supported? > > Err, can't it be done within a user space library (say, called librpc.a)? > I'm sorry, my question was missleading (English isn't my first language). What I want to know is: are there libraries for RPC / XDR, since I couldn't find some in the standard distribution. On Unix they seem to be included with a compiler distribution. So the question should rather be if there (will?) come libraries with the distribution or if I can find some libraries on the net. Thanks, uwe. 8X--------------------------------------------------------------------X8 If you cut here, you will probably destroy your monitor. Uwe Wuerfel Morgenstund... ...kommt selten allein. uwe@genias.de - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From gunther.ebert@ixos-leipzig.de Mon May 5 04:03:00 1997 From: gunther.ebert@ixos-leipzig.de (Gunther Ebert) Date: Mon, 05 May 1997 04:03:00 -0000 Subject: getpwuid References: <199705050657.XAA04213@dilbert.commercewave.com> Message-ID: <336DCCA7.54A8@ixos-leipzig.de> Ken Keys wrote: Geoffrey Noer wrote: > In beta 18 we add: > > Under NT only, chown(), getgrgid(), getgrnam(), endgrent(), getgrent(), > setpwend(), getpwent(), endpwent(). Win95 still has these as stubs. > > You will need to make an /etc/passwd and /etc/group for some > calls to work. That said, we provide a mkpasswd and mkgroup that > create valid /etc files from the NT user database. Wouldn't it make more sense to have those functions operate directly on the NT user DB, instead of on "unixified" copies of them? The cygwin copy will get out of sync and need to be rebuilt whenever the native copy is changed. The pw API was explicitly designed to *avoid* dependance on a particluar storage format; cygwin should take advantage of that and not require users to maintain artificial copies. Basically, you're right. It would be much nicer to use the native API instead of the passwd file operations. However, accessing a /etc/passwd file is *much* faster than browsing through the NT user account database, especially in a (trusted) domain environment. Note that every fstat call results in a call to several pwd functions, and using native APIs instead of /etc/passwd file operations would slow down file operations dramatically. Sure, the implemention of an appropriate cache mechanism could save most of the time. Maybe someone will have the time to rewrite the passwd functions. One could argue that having a unix-style /etc/passwd would allow the working of programs that read /etc/passwd directly instead of going through the pw API. But, 1) such programs are already nonportable, since not even all unix-like systems use the same format for /etc/passwd, No, I think this was not the idea. and 2) IMO, supporting badly written software isn't worth the trouble it will cause for users when their user DB's get out of sync. I think this would not be a big problem. How often does the account database usually change? Once a week? Or once a month? -- Gunther Ebert iXOS Anwendungs-Software GmbH Angerstrasse 40-42 D-04177 Leipzig Phone : +49 341 48503-0 Fax : +49 341 48503-99 E-mail: mailto:gunther.ebert@ixos-leipzig.de www : http://www.ixos-leipzig.de - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From ARichard@stark.cc.oh.us Mon May 5 07:20:00 1997 From: ARichard@stark.cc.oh.us (Richardson, Anthony) Date: Mon, 05 May 1997 07:20:00 -0000 Subject: MFC library support? Message-ID: <199705051420.KAA00073@dialup.oar.net> I just found out about gnu-win32 a few days ago. A big thanks to all involved. I'm primarily interested in porting Unix console applications, but I am also interested in getting started in Windows programming. Windows programming references I've seen recommend programming with MFC reather than directly with the API. Is it possible to use the MFC libraries with gnu-win32? If so, can anyone help me get started, i.e. provide pointers to documentation? Thanks, Tony Richardson arichard@stark.cc.oh.us - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From BrianB@atl.paysys Mon May 5 08:12:00 1997 From: BrianB@atl.paysys (Brian Beuning) Date: Mon, 05 May 1997 08:12:00 -0000 Subject: 'ls /' does not work Message-ID: <01BC5944.87E56FE0@BRIANB> I used to have a similar problem with du. A du under the root directory would give some error (if I remember right it was about not finding a '..' entry). The error has since gone away, but it might be because we are using NTFS instead of FAT now. Brian Beuning ---------- From: Hemal Pandya Sent: Friday, May 02, 1997 10:42 PM To: gnu-win32@cygnus.com Subject: 'ls /' does not work I am using beta17.1. running on Windows 95 'ls /' gives no output, either from within bash or at command.com prompt. Similarly pressing tab after / also provides no completions. mount says c: is mounted on /, which is what is is in the registry too. If I delete the registry entry, both mount and bash restore the entry. Even if I change the registry entry to c:\ or c:\\, bash changes it to c:. After providing a directory name under root, for example /gnu-win32/ filename completion works, as also ls. Any ideas? TIA, -- Hemal Pandya mailto:hemal@superlink.net phone:(908)340-9455 First He created computers, but they couldn't make mistakes So He created programmers - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From JONCKHEERE@d0tng.fnal.gov Mon May 5 09:47:00 1997 From: JONCKHEERE@d0tng.fnal.gov (JONCKHEERE@d0tng.fnal.gov) Date: Mon, 05 May 1997 09:47:00 -0000 Subject: Is RPC and XDR supported?5 Message-ID: <970505114747.2a4000e4@d0tng.fnal.gov> One of my users had exactly this question a couple of months ago. Check out http://set.gmd.de/~mfg/oncrpc.html Or try a web search for +ONC +RPC =========================================================================== Date: Mon, 05 May 1997 13:06:54 +0100 From: Uwe Wuerfel Subject: Re: Is RPC and XDR supported?5 In-reply-to: < Pine.LNX.3.95.970503175610.238A-100000@mailhub.air.net.au > Sender: owner-gnu-win32@cygnus.com To: Ben Elliston , gnu-win32@cygnus.com Reply-to: uwe@genias.de Message-id: < m0wOLZP-00000LC@gimli.genias.de > Organization: Genias Software GmbH / NT-Labor MIME-version: 1.0 X-Mailer: Pegasus Mail for Win32 (v2.53/R1) Content-type: text/plain; charset=US-ASCII Content-transfer-encoding: 7BIT Priority: normal Precedence: bulk References: < m0wNIuu-00000KC@gimli.genias.de > Comments: Authenticated sender is > > does anybody know if RPC and maybe XDR is supported by the Cygnus-Compiler > > and if not, when will it be supported? > > Err, can't it be done within a user space library (say, called librpc.a)? > I'm sorry, my question was missleading (English isn't my first language). What I want to know is: are there libraries for RPC / XDR, since I couldn't find some in the standard distribution. On Unix they seem to be included with a compiler distribution. So the question should rather be if there (will?) come libraries with the distribution or if I can find some libraries on the net. Thanks, uwe. 8X--------------------------------------------------------------------X8 If you cut here, you will probably destroy your monitor. Uwe Wuerfel Morgenstund... ...kommt selten allein. uwe@genias.de - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From bje@air.net.au Mon May 5 15:05:00 1997 From: bje@air.net.au (Ben Elliston) Date: Mon, 05 May 1997 15:05:00 -0000 Subject: Is RPC and XDR supported?5 References: Message-ID: > What I want to know is: are there libraries for RPC / XDR, since I couldn't > find some in the standard distribution. On Unix they seem to be included > with a compiler distribution. Haven't seen one, although I have done a lot of dabbling in the Linux RPC code and I can tell you that it would be fairly trivial to implement RPC using that code as a reference. Also, there is Microsoft RPC in WinNT. Is it compatible with Sun RPC? If it is, that might save you a lot of work -- just link in that native library at compile time. Ben --- Ben Elliston "For my birthday I got a humidifier and a de-humidifier. I put them in the same room and let them fight it out." -- Steven Wright - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From crippel@primenet.com Mon May 5 16:05:00 1997 From: crippel@primenet.com (Christoph Rippel) Date: Mon, 05 May 1997 16:05:00 -0000 Subject: (no subject) Message-ID: <336E67FB.D224DF20@primenet.com> - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From noer@cygnus.com Mon May 5 16:52:00 1997 From: noer@cygnus.com (Geoffrey Noer) Date: Mon, 05 May 1997 16:52:00 -0000 Subject: Cygnus Solutions is hiring... Message-ID: <199705052349.QAA08057@cirdan.cygnus.com> Hello, Like a lot of other companies in the Valley, Cygnus Solutions has quite a few open positions. If you might be interested in any of the following, please email me directly (noer@cygnus.com) with any questions and/or your resume. Job descriptions for some of these are available from: http://www.cygnus.com/jobs/ The current US opening are for: 1 User Interface engineer 1 Test engineer 1 Technical writer 2 Senior gcc engineers 2 Intermediate gcc engineers 2 Junior gcc engineers 1 Senior gdb engineer 1 Intermediate gdb engineer 1 Release engineer - NT The current UK openings are for: 1 Test/release engineer 3 User interface engineers Experience with gnu-win32 is highly desirable for the NT Release Engineer position. Experience with using embedded tools is a big plus for all open positions. Thanks! -- Geoffrey Noer noer@cygnus.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From gpasa@swissonline.ch Mon May 5 16:53:00 1997 From: gpasa@swissonline.ch (Pasa Guglielmo) Date: Mon, 05 May 1997 16:53:00 -0000 Subject: gnat and g77 on cygnus Message-ID: <336E7619.2FCD@swissonline.ch> Hi, I would like to know why we must install gnat and the cygwin32 distribution for win95-nt in two separate directories. Will it soon be possible to merge the two trees together ? Another question is whether it is planed to get g77 working in the cygwin32 ? -- Sincerely, Pasa Guglielmo e-mail address : gpasa@swissonline.ch mailing address: Pasa Guglielmo Rte des Cases 17A CH-1890 St-Maurice (Switzerland) tel.: +41 (0)24 485 50 40 fax: +41 (0)24 485 50 44 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From mikaels@pdc.kth.se Mon May 5 16:53:00 1997 From: mikaels@pdc.kth.se (Michael Strömberg) Date: Mon, 05 May 1997 16:53:00 -0000 Subject: configure solution? Message-ID: <336E612E.3C70BD65@pdc.kth.se> Hi there. I just started using cygwin32 yesterday. After playing around with environment variables for a while, I finally got my little "Hello world" program to compile. So then I decided to jump to a moderately simple GNU application, gzip. My problem is that configure seems to bug out. I am using B17.1 in Windows 95. I have seen people address the problem in the mailing list archives, but no apparent solutions were present in the archives. Is there a solution yet for the configure problem? Example: bash$ sh configure checking for gcc configure: 79605028: No such file or directory configure: 79609284: No such file or directory checking for underline in external names configure: 79609540: No such file or directory configure: grep: command not found checking for assembler configure: grep: command not found checking for install checking for AIX configure: 79600004: No such file or directory checking for minix/config.h configure: 79601060: No such file or directory checking for POSIXized ISC checking for DYNIX/ptx libseq configure: 79600388: No such file or directory <--- cut ---> The grep problem is easily solved, it's the rest that troubles me. BTW Is B18 coming out this week? // Michael - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From bischoff@ise.com Mon May 5 16:53:00 1997 From: bischoff@ise.com (Andreas Bischoff) Date: Mon, 05 May 1997 16:53:00 -0000 Subject: Unix like Tcl/Tk and extensions anyone ? Message-ID: <336E72E6.53FF@ise.com> Hi all, after some sleepless nights I decided to ask this list :-) I'm trying to build Tcl7.6/Tk7.4 from the unix sources. I know that there is a windows distribution available but AFAIK I can't use it: In addition to the core tcl/tk I also need a bunch of extensions: Tix, tclX, BLT, itcl/itk, et There is no windows port of blt (at least I didn't find any). So if I try to use the windows tk together with a nonwindows (X11) blt, they don't cooperate, of course. Therefore my attempt to build everything the unix way. But I'm having problems already with tcl: getsockoptions is missing from cygwin, also mkfifo. I tried to get it from libwsock32.a, but didn't succeed. nm shows a getsockopt@20. Searching through the archives I found an email that it is possible to mix unix sockets and winsock, but it was also mentioned, that it's not so easy. Did anybody build tcl/tk with cugwin32 ? BTW, I managed to build libX11.a, thanks to the patch from Sergei Okhapkin, thanks a lot. Therefore I'm optimistic that if I get tcl running, the rest should be also possible. TIA, Andreas - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From da@maigret.cog.brown.edu Mon May 5 16:53:00 1997 From: da@maigret.cog.brown.edu (David Ascher) Date: Mon, 05 May 1997 16:53:00 -0000 Subject: getpwuid References: <336DCCA7.54A8@ixos-leipzig.de> Message-ID: > Sure, the implemention of an appropriate cache mechanism could save most > of the time. Maybe someone will have the time to rewrite the passwd > functions. > > [...] > > I think this would not be a big problem. How often does the account > database usually change? Once a week? Or once a month? > > [...] > > Gunther Ebert Maybe a compromise solution can be reached -- how about having the pw functions check the mtime on /etc/passwd, and if it has been modified, automatically update it, or at the least print a warning message to stderr to have the sysadmin run the program that does the updating? --david - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jeffdb@netzone.com Mon May 5 20:38:00 1997 From: jeffdb@netzone.com (Mikey) Date: Mon, 05 May 1997 20:38:00 -0000 Subject: Solution! resource compiler for cygwin32/mingw32 Message-ID: <199705060333.UAA20303@nz1.netzone.com> I've noticed a few messages last month asking about a resource compiler for gnuwin32/mingw32, and I seem to have found a setup the works. rc, and a lot of other goodies!!! are available in the Platform SDK from microsoft. (get the big bin package!) also get the oledb SDK 1.1, or the help compiler & some other things won't work (unless you have VC++ I guess.) http://www.microsoft.com/msdn/sdk/default.htm rc has no problem with NL terminated lines, so it works with binary mounts. rsrc from rsxntdj1.zip seems to handle all types of resources correctly, in the last couple of days, I have been able to compile and link all of the sample applications that I've tried from the SDK using these 2 tools. archie for your nearest location, or it's available on ftp://ftp.uni-bielefeld.de/pub/systems/msdos/misc . or any djgpp mirror. BTW the next best solution I was able to come up with was lrc from the lcc package, and res2coff, if you don't want to talk Micros***. also in rsxntdj.zip are the patches for the SDK headers. I did have to do some hand work on them, to get everything to work, so I will post my current diffs, and some additional tools I used, to http://wwwdim.uqac.uquebec.ca/ftp/pub as SDK_for_ming_cygwin32.tar.gz If you're doing developing for win32, I strongly recomend mingw32, it gives you a FREE runtime, faster apps, and you still have cygwin32 for debugging and posix apps. Nice job Colin !!!!!! ;^) - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Tue May 6 00:51:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Tue, 06 May 1997 00:51:00 -0000 Subject: getpwuid Message-ID: <01BC5A13.FA37CA50@gater.krystalbank.msk.ru> Gunther Ebert wrote: > environment. Note that every fstat call results in a call to several pwd > functions, and > using native APIs instead of /etc/passwd file operations would slow down > file operations > dramatically. > Hopefully, You are not right. Fstat calls does not requires acces to user's database, because all neccessary information (uid and gid) is contained in file's security descriptor. -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From micha@genias.de Tue May 6 00:56:00 1997 From: micha@genias.de (Michaela Kurz) Date: Tue, 06 May 1997 00:56:00 -0000 Subject: getpwuid References: <336DCCA7.54A8@ixos-leipzig.de> Message-ID: On Mon, 5 May 1997, Gunther Ebert wrote: > Ken Keys wrote: > > Geoffrey Noer wrote: > > In beta 18 we add: > > > > Under NT only, chown(), getgrgid(), getgrnam(), endgrent(), > getgrent(), > > setpwend(), getpwent(), endpwent(). Win95 still has these as > stubs. > > > > You will need to make an /etc/passwd and /etc/group for some > > calls to work. That said, we provide a mkpasswd and mkgroup that > > create valid /etc files from the NT user database. > > Wouldn't it make more sense to have those functions operate directly > on > the NT user DB, instead of on "unixified" copies of them? The > cygwin > copy will get out of sync and need to be rebuilt whenever the native > > copy is changed. The pw API was explicitly designed to *avoid* > dependance on a particluar storage format; cygwin should take > advantage > of that and not require users to maintain artificial copies. > I don't think an /etc/passwd for NT is an option if you really want to port the software. I have chosen the following workaround: NT has a lot of environment variables. I use HOMEDRIVE and HOMEPATH to assemble the user's home directory and USERNAME to retrieve the user's name. It has worked so far, if I see some day that this is only a limited solution then I'll have to add registry calls. Michaela Kurz ------------------------------------------------------- M. Kurz Genias Software GmbH / Tel: 09401/9200-0 micha@genias.de / Fax: /9200-92 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From nsolsen@dscc.dk Tue May 6 04:22:00 1997 From: nsolsen@dscc.dk (Niels Skou Olsen) Date: Tue, 06 May 1997 04:22:00 -0000 Subject: X11 Message-ID: Hi, I have seen several people on this list mentioning the X11 libraries. How do people use those libraries? * Do they use a commercial X server? * Is there a free X server available from the net? * Or do the X libraries map to native Windows API? Best regards, Niels Skou Olsen (nsolsen@dscc.dk) - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Tue May 6 05:11:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Tue, 06 May 1997 05:11:00 -0000 Subject: configure solution? Message-ID: <01BC5A38.4F389D30@gater.krystalbank.msk.ru> Michael StrCmberg wrote: > So then I decided to jump to a moderately simple GNU application, gzip. > My problem is that configure seems to bug out. I am using B17.1 in > Windows 95. I have seen people address the problem in the mailing list > archives, but no apparent solutions were present in the archives. > > Is there a solution yet for the configure problem? > > Example: > > bash$ sh configure > checking for gcc > configure: 79605028: No such file or directory > configure: 79609284: No such file or directory mkdir /tmp -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Jim_Kramer@zygo.com Tue May 6 06:26:00 1997 From: Jim_Kramer@zygo.com (Jim_Kramer@zygo.com) Date: Tue, 06 May 1997 06:26:00 -0000 Subject: Unix like Tcl/Tk and extensions anyone ? Message-ID: <9704068629.AA862921545@maillink.zygo.com> Try Robin Becker's Home Page at " http://www.jessikat.demon.co.uk/ " for a Win32 port of BLT2.1. Jim Kramer. ______________________________ Reply Separator _________________________________ Subject: Unix like Tcl/Tk and extensions anyone ? Author: Andreas Bischoff at maillink Date: 5/6/97 5:24 AM Hi all, after some sleepless nights I decided to ask this list :-) I'm trying to build Tcl7.6/Tk7.4 from the unix sources. I know that there is a windows distribution available but AFAIK I can't use it: In addition to the core tcl/tk I also need a bunch of extensions: Tix, tclX, BLT, itcl/itk, et There is no windows port of blt (at least I didn't find any). So if I try to use the windows tk together with a nonwindows (X11) blt, they don't cooperate, of course. Therefore my attempt to build everything the unix way. But I'm having problems already with tcl: getsockoptions is missing from cygwin, also mkfifo. I tried to get it from libwsock32.a, but didn't succeed. nm shows a getsockopt@20. Searching through the archives I found an email that it is possible to mix unix sockets and winsock, but it was also mentioned, that it's not so easy. Did anybody build tcl/tk with cugwin32 ? BTW, I managed to build libX11.a, thanks to the patch from Sergei Okhapkin, thanks a lot. Therefore I'm optimistic that if I get tcl running, the rest should be also possible. TIA, Andreas - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From fkittred@gwi.net Tue May 6 08:25:00 1997 From: fkittred@gwi.net (Fletcher E Kittredge) Date: Tue, 06 May 1997 08:25:00 -0000 Subject: Is RPC and XDR supported?5 References: Message-ID: <199705061448.KAA10888@river.gwi.net> On Tue, 6 May 1997 08:04:58 +1000 (EST) Ben Elliston wrote: > Also, there is Microsoft RPC in WinNT. Is it compatible with Sun RPC? If > it is, that might save you a lot of work -- just link in that native > library at compile time. They are wildly different. MS RPC is based on DCE RPC. One of the goals of DCE RPC is to be different from Sun RPC (ONC.) regards, fletcher - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From hubert.fauque@wanadoo.fr Tue May 6 08:25:00 1997 From: hubert.fauque@wanadoo.fr (Hubert FAUQUE) Date: Tue, 06 May 1997 08:25:00 -0000 Subject: Problem with libcygwin Message-ID: <336f3da9.62411533@smtp.wanadoo.fr> I am new to gcc under win95 and I have a problem when linking: gcc -o test ptest.o libtest.a D:\cygnus\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32/cygnus-2.7.2-961023/../../../../i386-cygwin32/lib/libcygwin.a(libcmain.o): In function `main': /pizza/mushroom/noer/beta17/src/winsup/libcmain.cc:30: undefined reference to `WinMain@16' the library libtest.a is built without problems; what is this WinMain? thanks for any help Hubert hubert.fauque@wanadoo.fr - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From loki@dragoncat.net Tue May 6 12:06:00 1997 From: loki@dragoncat.net (Jeremy Blackman) Date: Tue, 06 May 1997 12:06:00 -0000 Subject: configure solution? References: <336E612E.3C70BD65@pdc.kth.se> Message-ID: On Tue, 6 May 1997, Michael [iso-8859-1] Strvmberg wrote: > So then I decided to jump to a moderately simple GNU application, gzip. > My problem is that configure seems to bug out. I am using B17.1 in > Windows 95. I have seen people address the problem in the mailing list > archives, but no apparent solutions were present in the archives. > > Is there a solution yet for the configure problem? mkdir /tmp That'll fix the configure problem. :) +---[ Loki ]------------+---------------------------------------------------+ | Jeremy Blackman | NeonMuck coder/maintainer - Multimedia MU*! | | loki@maison-otaku.net | http://www.maison-otaku.net/neon | +-----------------------+---------------------------------------------------+ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Ulrich.Lauther@mchp.siemens.de Tue May 6 16:52:00 1997 From: Ulrich.Lauther@mchp.siemens.de (Ulrich Lauther) Date: Tue, 06 May 1997 16:52:00 -0000 Subject: problems caused by memory.h Message-ID: <199705061856.UAA13390@fubini.mchp.siemens.de> The following message possibly didn't make it to the list, so I try again: #include main () { } when I try to compile the above program, stored as test.cc, I get a heap of error messages. Part of the problem is that memory.h includes string.h which gets confused with String.h. This goes away if I use an absolute pathname in memory.h for the inclusion of string.h. The part I do not understand are error messages due to multiple definitions in String.h, which however are legal C++ overloadings. Any hints how to explain and fix the problem? -ulrich gcc -v -c test.cc Reading specs from C:\cygnus\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32/cygnus-2.7.2-961023/s pecs gcc version cygnus-2.7.2-961023 C:\cygnus\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32/cygnus-2.7.2-961023/c pp. exe -lang-c++ -v -iprefix C:\cygnus\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32/cygnus-2.7.2-961023/ -undef -D__GNUC__=2 -D__GNUG__=2 -D__cplusplus -D__GNUC_MINOR__=7 -Di386 -D_WIN32 -DPOSIX -D__CYGWIN32__ -DWINNT -D_X86_=1 -D__STDC__=1 -D__stdcall=__attribute__((__stdcall__)) -D__cdecl=__attribute__((__cdecl__)) -D__i386__ -D_WIN32 -D__POSIX__ -D__CYGWIN32__ -D__WINNT__ -D_X86_=1 -D__STDC__=1 -D__stdcall=__attribute__((__stdcall__)) -D__cdecl=__attribute__((__cdecl__)) -D__i386 -D__POSIX -D__WINNT -Asystem(winnt) -Acpu(i386) -Amachine(i386) -Di386 -Asystem(unix) -Acpu(i386) -Amachine(i386) -D__i386__ -Asystem(unix) -Acpu(i386) -Amachine(i386) test.cc C:\WINDOWS\TEMP/cc001333.ii GNU CPP version cygnus-2.7.2-961023 (80386, BSD syntax) #include "..." search starts here: #include <...> search starts here: C:\cygnus\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32/cygnus-2.7.2-961023/. ./. ./../../../include/g++ C:\cygnus\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32/cygnus-2.7.2-961023/. ./. ./../../../include C:\cygnus\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32/cygnus-2.7.2-961023/. ./. ./../../i386-cygwin32/include C:\cygnus\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32/cygnus-2.7.2-961023/i ncl ude End of search list. C:\cygnus\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32/cygnus-2.7.2-961023/c c1p lus.exe C:\WINDOWS\TEMP/cc001333.ii -quiet -dumpbase test.cc -version -o C:\WINDOWS\TEMP/cc001333.s GNU C++ version cygnus-2.7.2-961023 (i386-cygwin32) compiled by GNU C version cygnus-2.7.2-961023. In file included from C:\\cygnus\\H-i386-cygwin32\\lib\\gcc-lib\\i386-cygwin32/cygnus-2.7.2-961 023 /../../../../i386-cygwin32/include/memory.h:4, from test.cc:1: C:\\cygnus\\H-i386-cygwin32\\lib\\gcc-lib\\i386-cygwin32/cygnus-2.7.2-961 023 /../../../../../include/g++/string.h:45: declaration of C function `struct StrRep * Scat(struct StrRep *, const char *, int, const char *, int, const char *, int)' conflicts with C:\\cygnus\\H-i386-cygwin32\\lib\\gcc-lib\\i386-cygwin32/cygnus-2.7.2-961 023 /../../../../../include/g++/string.h:44: previous declaration `struct StrRep * Scat(struct StrRep *, const char *, int, const char *, int)' here C:\\cygnus\\H-i386-cygwin32\\lib\\gcc-lib\\i386-cygwin32/cygnus-2.7.2-961 023 /../../../../../include/g++/string.h:160: declaration of C function `void cat(const class String &, const class SubString &, class String &)' conflicts with C:\\cygnus\\H-i386-cygwin32\\lib\\gcc-lib\\i386-cygwin32/cygnus-2.7.2-961 023 /../../../../../include/g++/string.h:159: previous declaration `void cat(const class String &, const class String &, class String &)' here C:\\cygnus\\H-i386-cygwin32\\lib\\gcc-lib\\i386-cygwin32/cygnus-2.7.2-961 023 /../../../../../include/g++/string.h:160: warning: `cat(const String &, const SubString &, String &)' is already a friend of class `String' C:\\cygnus\\H-i386-cygwin32\\lib\\gcc-lib\\i386-cygwin32/cygnus-2.7.2-961 023 /../../../../../include/g++/string.h:160: warning: previous friend declaration of `cat(const String &, const SubString &, String &)' C:\\cygnus\\H-i386-cygwin32\\lib\\gcc-lib\\i386-cygwin32/cygnus-2.7.2-961 023 /../../../../../include/g++/string.h:161: declaration of C function `void cat(const class String &, const char *, class String &)' conflicts with C:\\cygnus\\H-i386-cygwin32\\lib\\gcc-lib\\i386-cygwin32/cygnus-2.7.2-961 023 /../../../../../include/g++/string.h:160: previous declaration `void cat(const class String &, const class SubString &, class String &)' here C:\\cygnus\\H-i386-cygwin32\\lib\\gcc-lib\\i386-cygwin32/cygnus-2.7.2-961 023 /../../../../../include/g++/string.h:161: warning: `cat(const String &, const char *, String &)' is already a friend of class `String' C:\\cygnus\\H-i386-cygwin32\\lib\\gcc-lib\\i386-cygwin32/cygnus-2.7.2-961 023 /../../../../../include/g++/string.h:161: warning: previous friend declaration of `cat(const String &, const char *, String &)' C:\\cygnus\\H-i386-cygwin32\\lib\\gcc-lib\\i386-cygwin32/cygnus-2.7.2-961 023 /../../../../../include/g++/string.h:162: declaration of C function `void cat(const class String &, char, class String &)' conflicts with C:\\cygnus\\H-i386-cygwin32\\lib\\gcc-lib\\i386-cygwin32/cygnus-2.7.2-961 023 /../../../../../include/g++/string.h:161: previous declaration `void cat(const class String &, const char *, class String &)' here [further error messages deleted] -- -lauther ---------------------------------------------------------------------------- Ulrich Lauther ph: +49 89 636 48834 fx: ... 636 42284 Siemens ZT AN 1 Internet: Ulrich.Lauther@mchp.siemens.de - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jla@arceneaux.com Tue May 6 16:52:00 1997 From: jla@arceneaux.com (Joseph Arceneaux) Date: Tue, 06 May 1997 16:52:00 -0000 Subject: NT Emacs Message-ID: <199705062351.QAA07502@emptiness.arceneaux.com> What happened to the link to NT Emacs on your win32 page? Thanks, Joe - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From weiku@UTKUX1.UTK.EDU Tue May 6 18:29:00 1997 From: weiku@UTKUX1.UTK.EDU (Wei Ku) Date: Tue, 06 May 1997 18:29:00 -0000 Subject: does 'ar' work ? Message-ID: <199705070105.VAA33326@solid.ssd.ornl.gov> Hello: I found that the ar command could not give me the names of the files in the library. Any Help will be appreciated: bash$ ar t libfl.a ar: libg++.a: Malformed archive bash$ Sincerely, Wei Ku *************************************** Department of Physics and Astronomy The University of Tennessee 1408 Circle Drive Knoxville, Tennessee 37996-1200 weiku@utkux.utcc.utk.edu --------------------------------------- Solid State Division Oak Ridge National Laboratory P.O.Box 2008 Oak Ridge, TN 37831-6032 Phone: (423) 574-5795 Fax: (423) 574-4143 weiku@solid.ssd.ornl.gov *************************************** - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From hemal@superlink.com Tue May 6 18:29:00 1997 From: hemal@superlink.com (Hemal Pandya) Date: Tue, 06 May 1997 18:29:00 -0000 Subject: Getting GCC to work References: <01BC57B0.9F360730@sos> Message-ID: Sergey Okhapkin writes: > > Glenn Thompson wrote: > > Can anybody tell me how to get this win32 thig to work right?? > > > > Keep getting the following message: > > > > ld: cannot open -lgcc: No such file or directory > > > > Did You set GCC_EXEC_PREFIX in environment right? I had problems with GCC_EXEC_PREFIX till I was setting it from ~/.bash_profile. Need to do it in autoexec.bat (Windows 95) and system promperties (NT). -- Hemal Pandya mailto:hemal@superlink.net phone:(908)340-9455 First He created computers, but they couldn't make mistakes So He created programmers - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From luke@research.canon.com.au Tue May 6 18:29:00 1997 From: luke@research.canon.com.au (Luke Kendall) Date: Tue, 06 May 1997 18:29:00 -0000 Subject: sed unterminated 's' command Message-ID: <199705070129.LAA23232@tosh.research.canon.com.au> Hi This is related to gnu-win32 beta 17.1 under Windows Nt 3.51, and also the version of August 1996. It looks like sed has had a bug introduced. The true sed version number is unknown, since the old and new versions (as evidenced by executable file size) report the same version number. Here are the details: $ x:/gnu-win32/bin/sed --version GNU sed version 2.05 $ x:/gnu-win32/bin-new/sed --version GNU sed version 2.05 $ d:/user/local/bin/sed --version GNU sed version 2.05 $ ls -l d:/user/local/bin/sed.exe x:/gnu-win32/bin-new/sed.exe x:/gnu-win32/bin/sed.exe -rwxrwxrwa 1 Everyone Everyone 89600 Apr 23 17:15 d:/user/local/bin/sed.exe -rwxrwxrwx 1 PAGE\luke PAGE\Domain+Users 89600 Sep 3 1996 x:/gnu-win32/bin-new/sed.exe -rwxrwxrwx 1 PAGE\greyham PAGE\Domain+Users 52224 Apr 13 1996 x:/gnu-win32/bin/sed.exe $ echo fred | x:/gnu-win32/bin/sed 's/.*/xxx &/' xxx fred $ echo fred | x:/gnu-win32/bin-new/sed 's/.*/xxx &/' //x/gnu-win32/bin-new/sed: Unterminated `s' command $ echo fred | d:/user/local/bin/sed 's/.*/xxx &/' //d/user/local/bin/sed: Unterminated `s' command $ Regards, luke - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From mprobert@udax.com.au Tue May 6 21:59:00 1997 From: mprobert@udax.com.au (Mark Probert) Date: Tue, 06 May 1997 21:59:00 -0000 Subject: configure solution? Message-ID: <01BC5AF1.E6F2EEF0@MARK> Sergey writes: +> +> Is there a solution yet for the configure problem? +> + > Example: +> + +mkdir /tmp + A variation on the theme: I am also having trouble with configure. Here the problem goes: (this is from pdksh) $ ./configure conftest* does not exist confdefs.h does not exist loading cache ./config.cache [lots of checks deleted] creating ./config.status conftest.vals does not exist conftest.tail does not exist creating Makefile creating config.h conftest.frag does not exist conftest.in does not exist conftest.out does not exist conftest.h does not exist config.h is unchanged $ Any ideas? TIA. -mark. ============================================================ Mark Probert phone: +61 2 9954 0407 Udax Pty Ltd fax: +61 2 9956 6404 L11, 157 Walker St mobile: 0411 422 932 Nth Sydney 2060 Australia email: markp@udax.com.au ============================================================ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From bernd@asterix.gecko.de Wed May 7 00:44:00 1997 From: bernd@asterix.gecko.de (Bernd Prager) Date: Wed, 07 May 1997 00:44:00 -0000 Subject: X11 References: Message-ID: > I have seen several people on this list mentioning the X11 libraries. > How do people use those libraries? > > * Do they use a commercial X server? > * Is there a free X server available from the net? I use "X-Win32" a comercial server from starnet, available at " http://www.starnet.com/Products/xwin32.html ". Free full featured demo version, limited with 2 hours (!) running time. After 2 hours (include warning 15 minutes befor stop running) you have to start the server again. (Enaugh for all I do!) Bernd _____________________________________________________________________ Bernd Prager GECKO mbH; Wismarsche Str.3, 18057 Rostock; Germany http://www.gecko.de PGP Key fingerprint = 83 54 6A 3B 7A 9D 6C 0E F3 41 CE 99 11 30 B7 D6 public key by mailto:bpr@gecko.de?Subject=SendPGPKey _____________________________________________________________________ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Imerio_Ballarini@baynetworks.com Wed May 7 03:37:00 1997 From: Imerio_Ballarini@baynetworks.com (Imerio Ballarini) Date: Wed, 07 May 1997 03:37:00 -0000 Subject: Cross compilation for MIPS/68000 embedded applications. References: <3366FBCD.B36BA80B@baynetworks.com> Message-ID: <33705B8F.6ADAEFFC@baynetworks.com> Originally I wrote: > I've built the 17.1 distribution of GCC as a cross compiler for the MIPS > processor and have got it to compile all of my embedded app. On linking > however, I'm getting unresolved "virtual table" references for some (and > only some) of my classes and I don't know why. The code I'm compiling > builds fine under different compilers. The specific class methods that > fail have nothing in common and differ little from my other classes that > build and link successfully. Thanks to all those that replied concerning our problems as outlined above. For those that may be interested we have now built our app with the native Win32 GNU compiler (that we built) and we get exactly the same unresolved references to virtual tables for some of our classes only and we have no idea why. So it has nothing to do with _cross_ compilation and everything to do with the GNU compiler. Me thinks I'll have to get stuck into the compiler - but I'm so scared! > I'm at a loss. Any ideas? Has anyone successfully xcompiled embedded > code? > > Anyway, the whole job has been a pain and I have to do it all over again > for my 68000 apps. I don't suppose any of you have built a 68000 > xcompiler? > > Thanks, > -Imerio. > > -- > Imerio Ballarini > Bay Networks, Watford, UK. > Tel +44 (0)1923 479819 > - > For help on using this list (especially unsubscribing), send a message to > "gnu-win32-request@cygnus.com" with one line of text: "help". -- Imerio Ballarini Bay Networks, Watford, UK. Tel +44 (0)1923 479819 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From weiku@UTKUX1.UTK.EDU Wed May 7 05:03:00 1997 From: weiku@UTKUX1.UTK.EDU (Wei Ku) Date: Wed, 07 May 1997 05:03:00 -0000 Subject: does 'ar' work ? Message-ID: <199705071202.IAA34881@solid.ssd.ornl.gov> Hello: I found that the ar command could not give me the names of the files in the library. Any Help will be appreciated: bash$ ar -t libg++.a ar: libg++.a: Malformed archive bash$ Sincerely, Wei Ku *************************************** Department of Physics and Astronomy The University of Tennessee 1408 Circle Drive Knoxville, Tennessee 37996-1200 weiku@utkux.utcc.utk.edu --------------------------------------- Solid State Division Oak Ridge National Laboratory P.O.Box 2008 Oak Ridge, TN 37831-6032 Phone: (423) 574-5795 Fax: (423) 574-4143 weiku@solid.ssd.ornl.gov *************************************** - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From mgstahl@sonalysts.com Wed May 7 07:17:00 1997 From: mgstahl@sonalysts.com (Matthew G. Stahl) Date: Wed, 07 May 1997 07:17:00 -0000 Subject: GCC on DEC Alpha NT ? Message-ID: <3.0.1.32.19970507101550.006ad024@hatch.sonalysts.com> I've been using gnu-win32 on ix86 (NT/Win95) for a while. Now I'm trying to find *at_least* a GCC binary for DEC Alpha NT so that I can compile the toolset. or, if someone knows of a complete set of binaries, so much the better :) --Matt - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From rafaeldp@caymasa.es Wed May 7 11:06:00 1997 From: rafaeldp@caymasa.es (RAFAELDP) Date: Wed, 07 May 1997 11:06:00 -0000 Subject: Difference with DLL creation Message-ID: <199705071806.LAA09678@cygnus.com> Hi: Please wich is the diference in the creation of DLL, I have the choice of creating no-relocatable dlls or relocatable dlls. Thanks. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Scott.Mintz@po.cle.ab.com Wed May 7 11:06:00 1997 From: Scott.Mintz@po.cle.ab.com (Scott Mintz) Date: Wed, 07 May 1997 11:06:00 -0000 Subject: Machine Description Files Message-ID: <00057FDE.1893@po.cle.ab.com> Does anyone know where I can find documentation on the machine description file formats used by gcc (*.md) ? For those that don't know what they are, they are used in the gcc backend to create assembly code. They are the files in the gcc/config subdirectories. +------------------------------+------------------------------------+ | Scott A. Mintz | voice: (216) 646-4805 | | Allen-Bradley Company | fax: (216) 646-4961 | | 1 Allen-Bradley Drive | email: scott.mintz@po.cle.ab.com | | Mayfield Hts., OH 44124-6118 | CIS: 71461,632 | +------------------------------+------------------------------------+ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From mgstahl@sonalysts.com Wed May 7 14:19:00 1997 From: mgstahl@sonalysts.com (Matthew G. Stahl) Date: Wed, 07 May 1997 14:19:00 -0000 Subject: GCC on DEC Alpha NT ? Message-ID: <3.0.1.32.19970507171805.006a5e2c@hatch.sonalysts.com> I've been using gnu-win32 on ix86 (NT/Win95) for a while. Now I'm trying to find *at_least* a GCC binary for DEC Alpha NT so that I can compile the toolset. or, if someone knows of a complete set of binaries, so much the better :) --Matt - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From scottc@net-community.com Wed May 7 15:08:00 1997 From: scottc@net-community.com (Scott Christley) Date: Wed, 07 May 1997 15:08:00 -0000 Subject: GCC on DEC Alpha NT ? Message-ID: <199705072221.PAA15416@linus.net-community.com> At 10:15 AM 5/7/97 -0400, Matthew G. Stahl wrote: >I've been using gnu-win32 on ix86 (NT/Win95) for a while. > >Now I'm trying to find *at_least* a GCC binary for DEC Alpha NT so that I >can compile the toolset. > >or, if someone knows of a complete set of binaries, so much the better :) I've recently gotten a contract to port ObjC to Alpha/NT. The port has been relatively straight forward; however, the GNU binary utilities do not support Alpha/NT so you need Visual C++. Scott - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jonathan@westwood.com Wed May 7 15:22:00 1997 From: jonathan@westwood.com (Jonathan Lanier) Date: Wed, 07 May 1997 15:22:00 -0000 Subject: I found another compiler bug! Message-ID: Attached is an example of a serious bug in the GCC compiler. The gist of it is, if you use packed structures in GCC, and then attempt to use bitfields in those packed structures, the compiler completely miscalculates the offsets for the data in the structure. This example demonstrates a C++ class containing only 2 members. Each one is initialized separately, and the resulting answers printed for each member should be different, but they are not. Removing the __attribute__((packed)) keyword gets rid of the error, but it means you can't pack your structures, which in my case is something I really need to do! Does anyone know if this bug was previously known, or if it exists on other architectures? - Jonathan jonathan@westwood.com From noer@cygnus.com Wed May 7 18:09:00 1997 From: noer@cygnus.com (Geoffrey Noer) Date: Wed, 07 May 1997 18:09:00 -0000 Subject: GNU-Win32 beta 18 is available Message-ID: <199705080106.SAA22312@cirdan.cygnus.com> Greetings, I'm happy to announce that the gnu-win32 beta 18 release is now available from ftp://ftp.cygnus.com/pub/gnu-win32/latest Following is the readme file for the release. It explains what you'll want to download. I've changed the installation method substantially (InstallShield installer) but there are still a few extra steps that need to be taken so please read it! :-) As requested, "md5sums" in the b18 ftp directory contains md5 checksums for all the various bits. This may be useful for those of you with unreliable ftp connections. Most files are available in chunks and as whole files. I would like to thank the many people who made valuable contributions over the past few months. Sergey Okhapkin made several, ranging from finishing the implementation of select() to increasing the functionality of signals and exception handling. Gunther Ebert made significant improvements in the passwd/security-related code including the addition of a working chown() for NT. Marcus Daniels added directory symlink support. Jeremy Allison added mmap and mprotect, and made quite a lot of other useful changes as well. Enjoy! Geoffrey Noer Cygnus Solutions -------------------------------> snip <-------------------------------- README.txt for gnu-win32 beta 18 release May 7, 1997 You can find a link to the most recent release in ` ftp://ftp.cygnus.com/pub/gnu-win32/latest ' Changes since the last release ============================== This is a major release. The new cygwin.dll is still backwards-compatible with previously linked applications but contains significant changes. We have completely changed the installation process to make use of an InstallShield5-based installer. This should reduce the number of installation problems people have experienced in the past. However, it is still necessary to set environment variables by hand, as explained later. (Future gnu-win32 installers may include the capability to do this automatically). Changes in specific tools: -------------------------- GCC compilation times have been improved by 20-30% by using spawn() instead of fork(). GCC accepts both Win32 and POSIX paths/path lists in its environment variables (COMPILER_PATH, LIBRARY_PATH, C_INCLUDE_PATH, CPLUS_INCLUDE_PATH, OBJC_INCLUDE_PATH) GDB comes with a tcl/tk-based GUI (gdbtk). You can still invoke the command line gdb by invoking it with "gdb -nw". Bash verifies that /tmp exists and is a directory upon startup. It complains if this isn't the case. Running gcc or ld with "-s" used to create invalid executables. The bug in bfd that was responsible for this has been fixed. The conflict between String.h and string.h (and other such pairs of header files) where you include one and get the other has been fixed. The top level install-sh script tries to install foo.exe if asked to install foo when foo's not present. This fixes many installs of Unix software. Dlltool has preliminary support for the IMPORT declaration in .def files when invoked with -I. Feel free to experiment with it but once this functionality is tested more extensively this flag may go away. Time is upgraded to version 1.7. Make is upgraded to version 3.75. Make accepts both Win32 and POSIX path lists in the VPATH variable. Changes in the Cygwin32 API (cygwin.dll): ----------------------------------------- The following is now supported: * UNC paths * Reverse index escapes in console code * Blocking select()s on a combination of sockets/handles * Directory symlinks. * Reparenting of child processes. The following calls have been added: * mmap(), mprotect(), msync(), munmap(). fork() changed to support these. * fsync(), statfs(), fstatfs(). * getprotobynumber() and getservbyport(). * get_osfhandle(), cwait(). * spawnl(), spawnle(), spawnlp(), spawnlpe(), spawnv(), spawnve(), spawnvp(), spawnvpe(). * nice(). * sigpending(), sigsuspend() * Under NT only, chown(), getgrgid(), getgrnam(), endgrent(), getgrent(), setpwend(), getpwent(), endpwent(). Win95 still has these as stubs. Significantly better signals / exception handling support added. The kill signal works much better now (control-C works in bash). Shell scripts now run the shell specified after the #! instead of always defaulting to /bin/sh. Floating point registers are now properly initialized in the crt0.o. Opening non-disk files such as com ports no longer check to see if they are symlinks or executables. The console title now is set to the name of the running process. Winsock is now initialized upon app startup. Moved reent_data from private address space to cygwin.dll. The system() call now invokes spawnvp() instead of fork()/exec(). Support for NT extended attributes has been added but is disabled for now because it slowed things down too much. We want to use them to remember info about symlink and executable status of files. Under NT only, utilities mkpasswd and mkgroup can generate a valid /etc/passwd and /etc/group. Earlier releases stored mount points in the registry under "Cygnus Support". This changed to "Cygnus Solutions" starting with beta 18. Either use a registry editor (regedit under NT) to rename the old entry or just redo your mount points and the cygwin.dll will automatically create the new one for you. Mount points can now be up to MAX_PATH in length instead of 30 characters. Contents ======== The following packages are included in the full release: Cygnus Developer's Kit (CDK): binutils, byacc, diff, flex, gas, gcc, gdb, ld, libg++, libstdc++, make, patch user tools: bash, fileutils, findutils, gawk, grep, gzip, m4, sed, shellutils, tar, textutils, time Full source code is available for these tools. It is split into these two units. Installing the binary release: ============================== First, remove any older versions of the gnu-win32 tools that exist in your $PATH (or delete them). Connect to our ftp server and cd to the following directory: ` ftp://ftp.cygnus.com/pub/gnu-win32/latest '. If you want the development tools... ------------------------------------ If you want the development tools and the programs necessary to run the GNU configure mechanism, you should download the whole release. Download `cdk.exe'. If you have an unreliable connection, download the binary in smaller chunks instead. For the split cdk installer, get the files in the `cdk-split' subdirectory. Once downloaded, combine the split files at the command prompt by doing a: copy /b xaa + xab + xac + ... + xag + xah cdk.exe del xa*.* Run `cdk.exe' and follow the instructions to install the tools. The default install location is `:\gnuwin32\b18'. Future examples will assume this location with being the C drive. Next you must set four gnu-win32-related environment variables or the tools will not function correctly. You may prefer to do this outside of bash so you can use the tools from the command.com as well as from within bash. This means using the `System' control panel in NT or modifying the autoexec.bat under Windows 95. Assuming the `C:\gnuwin32\b18' install location: 1) Set the GCC_EXEC_PREFIX environment variable to `C:\gnuwin32\b18\H-i386-cygwin32\lib\gcc-lib\' The trailing slash is important! 2) Add `C:\gnuwin32\b18\H-i386-cygwin32\bin;C:\gnuwin32\b18\tcl\bin' to your $PATH. This is where the executables live. 3) Set the TCL_LIBRARY environment variable to `C:/gnuwin32/b18/tcl/lib/tcl7.6' You must use forward slashes in this variable. 4) Set the GDBTK_LIBRARY environment variable to `C:/gnuwin32/b18/share/gdbtcl' You must use forward slashes in this variable. Various programs need to be able to find `/bin/sh'. You should create a `/bin' if one doesn't already exist and put a copy of `sh.exe' there. You can use the `mount' utility to select which drive letter is mounted as `/'. See the Frequently Asked Questions (FAQ) file for more information on `mount'. Many programs including bash and byacc need to be able to find a `/tmp'. Verify that `/tmp' exists on the drive mounted as `/', creating one if necessary. And you should be done! If you should ever want to uninstall the tools, you may do so via the "Add/Remove Programs" control panel. If you're not interested in the development tools... ---------------------------------------------------- If you just want bash, fileutils, textutils, and other utilities and don't need the development tools, you do NOT need to install the entire release. Download `usertools.exe'. If you have an unreliable connection, download the binary in smaller chunks instead. For the split user tools installer, get the files in the `usertools-split' subdirectory. Once downloaded, combine the split files at the command prompt by doing a: copy /b xaa + xab + xac + xad + xae usertools.exe del xa*.* Run `usertools.exe' and follow the instructions to install the tools. The default install location is `:\gnuwin32\b18'. Future examples will assume this location with being the C drive. Next you need to add the executables to your $PATH. You may prefer to do this outside of bash so you can use the tools from the command.com as well as from within bash. This means using the `System' control panel in NT or modifying the autoexec.bat under Windows 95. Assuming the `C:\gnuwin32\b18' install location, add `C:\gnuwin32\b18\H-i386-cygwin32\bin' to your $PATH. Various programs need to be able to find `/bin/sh'. You should create a `/bin' if one doesn't already exist and put a copy of `sh.exe' there. You can use the `mount' utility to select which drive letter is mounted as `/'. See the Frequently Asked Questions (FAQ) file for more information on `mount'. Many programs including bash need to be able to find a `/tmp'. Verify that `/tmp' exists on the drive mounted as `/', creating one if necessary. And you should be done! If you should ever want to uninstall the tools, you may do so via the "Add/Remove Programs" control panel. Installing the source code -------------------------- Before downloading the source code corresponding to the release, you should install the latest tools (either the full release or just the user tools). Create the directory that will house the source code. `cd' there. Connect to our ftp server and cd to the following directory: ` ftp://ftp.cygnus.com/pub/gnu-win32/latest '. If you want the user tools source code, `cd' into the `usertools-src' subdirectory. Download the files there. If you want the Cygnus Developer's Kit sources, `cd' into the `cdk-src' subdirectory. Download the files there. Back in the command.com, for the user tools source: copy /b xba + xbb + xbc + xbd + ... + xbk + xbl usersrc.tar.gz del xb*.* gunzip -d usersrc.tar.gz tar xvf usersrc.tar For the compiler tools source: copy /b xca + xcb + xcc + xcd + ... + xck + xcl cdksrc.tar.gz del xc*.* gunzip -d cdksrc.tar.gz tar xvf cdksrc.tar Both will expand into a directory called `src'. And you should be done... Cygnus resources online *********************** The main WWW page for the GNU-Win32 project is: http://www.cygnus.com/misc/gnu-win32/ Tool-specific information can be found here: http://www.cygnus.com/library/ Links to additional documentation are accessible from the main web page. Known/potential problems ************************ Fixed-length arrays in the dll ============================== There are still a lot of places in the dll where fixed-length arrays are used instead of dynamic structures which makes us use more memory than necessary or limits the capabilities of the Cygwin32 layer. Some of these have been fixed since beta 16. Missing POSIX functions ======================= There's still a lot of standard posix functionality that isn't present that could be. Also, some functions aren't implemented fully. For example, sync returns 0, even though it doesn't do much. We're filling in the gaps slowly but surely. Programs can't deal with // pathname scheme in arguments ======================================================== gcc and other tools aren't fully compatible with the current pathname scheme: it can't grok an argument of -I//d/foo which means it is vital that when attempting to self-host, that only normal paths with single slashes are used. DOS special filenames ===================== Files cannot be named com1, lpt1, or aux (to name a few); either as the root filename or as the extension part. If you do, you'll have trouble. Unix programs don't avoid these names which can make things interesting. Eg, the perl distribution has a file called `aux.sh'. The perl configuration tries to make sure that `aux.sh' is there, but an operation on a file with the magic letters 'aux' in it will hang. User defined mallocs. ===================== If you define a function called `malloc' in your own code, and link with the DLL, the DLL *will* call your `malloc'. You'd better make sure that your malloc more or less works. If you run any programs from the DOS command prompt, rather than from in bash, the DLL will try and expand the wildcards on the command line. This process uses `malloc' *before* your main line is started. If you have written your own `malloc' to need some initialization to occur after `main' is called, then this will surely break. Mixed case filenames. ===================== The perl config causes some other problems too; the config.SH scripts generate a file called "Makefile", and use and delete a file called "makefile" on the way. Because of NT's case insensitivity, this will nuke Makefile, leaving you with nothing. You can fix this by mounting with the 'mixed' option. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From fjh@cs.mu.OZ.AU Wed May 7 19:27:00 1997 From: fjh@cs.mu.OZ.AU (Fergus Henderson) Date: Wed, 07 May 1997 19:27:00 -0000 Subject: Machine Description Files References: <00057FDE.1893@po.cle.ab.com> Message-ID: <199705080227.MAA15661@murlibobo.cs.mu.OZ.AU> Scott Mintz, you wrote: > > Does anyone know where I can find documentation on the machine > description file formats used by gcc (*.md) ? See the "Machine Descriptions" chapter of the gcc documentation. Obtaining the GCC documentation. -------------------------------- The GNU CC documentation is contained in the files `*.texi' which come with the gcc source distribution in the file `gcc-.tar.gz' on the usual GNU ftp sites. These are `Texinfo' files, and they can be used to produce either a hypertext online manual using `info', or a printed manual using `TeX'. To produce the `info' online manual, first install the GNU Texinfo package. (It is in the file `texinfo-.tar.gz' on the usual GNU ftp sites.) Then, in the GCC source directory, type the command make install-info This will create the `*.info*' files, and copy them to /usr/local/info. Then type `info gcc' and `info cpp' to read them. You might also consider installing `tkinfo' if you have an X-Windows display. To view the files without installing them, do a `make info' rather than `make install-info', and then type `info -f gcc.info'. If the `*.info' files already exist, you can view them from Emacs without installing the GNU Texinfo package, using the commands M-x info g (./gcc.info) ^^^^^^^^^^ full path inside parentheses To produce the printed manual, install both the GNU Texinfo package and TeX. Then type the command make dvi in the gcc source directory. This will create the files gcc.dvi and cpp.dvi, which can be converted to PostScript using dvi2ps and then printed. >From Lars G. Hecking : > Try and . Look out for > The Cygnus Support Online Library and User Documents/GNU, resp. Another useful > page is . >From Yusuf Goolamabbas : > ftp://primus.com/pub/gnu-ps > This has PS files in US Letter size. For European A4 format. I think > there is a site in switzerland (forgot its address) [If you have any suggestions for how these instructions could be improved, they would be appreciated. These instructions are very old, and some of them may be obsolete.] -- Fergus Henderson | "I have always known that the pursuit WWW: < http://www.cs.mu.oz.au/~fjh > | of excellence is a lethal habit" PGP: finger fjh@128.250.37.3 | -- the last words of T. S. Garp. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From noer@cygnus.com Thu May 8 00:48:00 1997 From: noer@cygnus.com (Geoffrey Noer) Date: Thu, 08 May 1997 00:48:00 -0000 Subject: Cross compilation for MIPS/68000 embedded applications. References: <33705B8F.6ADAEFFC@baynetworks.com> Message-ID: <199705080748.AAA22628@rtl.cygnus.com> Imerio Ballarini wrote: > > > I've built the 17.1 distribution of GCC as a cross compiler for the MIPS > > processor and have got it to compile all of my embedded app. On linking > > however, [...] > Me thinks I'll have to get stuck into the compiler - but I'm so scared! For the record, I'd like to point out that the gnu-win32 betas are made from development snapshots. The only configuration that I test is the i386-cygwin32 native one. Your mileage will vary once you leave that configuration. The sources used for Cygnus customer releases are regression tested on many different hosts against many different targets and are stable for use in commercial environments. The few critical bugs that come up are dealt with promptly. Those of you who are interested in using the GNU tools in the embedded space (who aren't already customers) should at least consider taking out a support contract from Cygnus. If you're interested, send mail to info@cygnus.com. I don't know when i386-cygwin32 hosted cross-toolchains will be available for what targets but someone should get back to you with the answers... -- Geoffrey Noer noer@cygnus.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From kkaempf@progis.de Thu May 8 01:33:00 1997 From: kkaempf@progis.de (Klaus Kaempf) Date: Thu, 08 May 1997 01:33:00 -0000 Subject: GCC on DEC Alpha NT ? References: <199705072221.PAA15416@linus.net-community.com> Message-ID: <9705080812.AA18081@progis.de> > [...] > however, the GNU binary utilities do not > support Alpha/NT so you need Visual C++. > >From my openVMS/Alpha port I've got enough knowledge of bfd, binutils, and gas to make the port. However, there doesn't seem to be enough interest in the NT/Alpha community. Hopefully this will change someday soon. Klaus -- proGIS Software E-Mail: kkaempf@progis.de Dipl.-Inform. Klaus Kdmpf Fax: 0241-47067-29 Jakobstr. 117 Voice: 0241-47067-11 D-52064 Aachen WWW: http://www.progis.de - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From colin@bird.fu.is.saga-u.ac.jp Thu May 8 06:07:00 1997 From: colin@bird.fu.is.saga-u.ac.jp (Colin Peters) Date: Thu, 08 May 1997 06:07:00 -0000 Subject: Difference with DLL creation Message-ID: <01BC5BFC.1185D140@gbird0> RAFAELDP[SMTP:rafaeldp@caymasa.es] wrote: >Please wich is the diference in the creation of DLL, I have the choice of >creating no-relocatable dlls or relocatable dlls. The difference happens when your DLL is loaded by a program. The operating system has to map the code of the DLL into a portion of the address space of the program that loads it. For example, your DLL might contain three 0x100 byte long functions with offsets in the file of 0x0000 0x0100 and 0x200. If the DLL is mapped to address 0x02000000 for example, then the addresses that the program needs to call to use the DLL functions are 0x02000000, 0x02000100 and 0x02000200 respectively (actually the point is not so much the calling program as the DLL's internal calls, see below). DLLs contain a number which tells the operating system where they would like to be mapped. For example the above DLL might contain a number saying that it would like to be mapped at 0x02000000. If the DLL can rely on being mapped to this position then it can include function calls to it's own code (and references to it's static data) and it will know what addresses the code (or data) will be at. But that would require that all processes leave the space between 0x02000000 and 0x02000300 in their address space free. They can't use it for their own code or (more likely) load another DLL into the same address. If two DLLs both want this space and neither can be moved (i.e. they are both not relocatable) then a program that uses both DLLs cannot run. The operating system will simply not load it successfully. A relocatable DLL (or program) still says what address it would like to be loaded at. But it is also marked as relocatable, and contains a section of data indicating what data within the DLL has to be fixed up if the DLL is not loaded in the preferred location (pointers to functions within the DLL code, or static data for example). When loading a relocatable DLL the operating system tries to put it in the requested position, but if the space is occupied it finds a big enough spot and uses the relocation data to fix the DLL (well, the copy in RAM) and load it in the free location. You can't tell what other DLLs might be loaded at the same time as your DLL (unless no one else will ever use your DLL), or what code or data might be occupying the addresses you would like to put your DLL at, so basically DLLs must be relocatable. It also saves you the hassle of trying to juggle base addresses and code+static data sizes to make sure that all DLLs can load successfully. Just leave it to the operating system. Summary: relocatable is best. Colin. PS. This is my understanding based on what I've read and heard. I'm not a certified Microsoft support guy or anything like that. -- Colin Peters - colin@bird.fu.is.saga-u.ac.jp -- Saga University Dept. of Information Science -- http://www.fu.is.saga-u.ac.jp/~colin/index.html -- http://www.geocities.com/Tokyo/Towers/6162/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From PaulMa@t-mi.com Thu May 8 09:41:00 1997 From: PaulMa@t-mi.com (Paul Mapstone) Date: Thu, 08 May 1997 09:41:00 -0000 Subject: wc bug in b18 Message-ID: <199705081640.RAA03502@t-mi.com> Using beta 18 under NT 3.51, I get very strange results with wc: wc usertools.exe 3 14 799 usertools.exe However, wc -c gives the correct solution: 4416422 usertools.exe I get similar premature end of file results with other tools, e.g. cat, split. Is this something to do with the CR/LF thing? I have tried putting an entry in the registry for fmode_binary, but this does not make any difference. The registry entry I have tried is: HKEY_LOCAL_MACHINE SOFTWARE Cygnus Solutions GNU-Win32 b18 fmode_binary:REG_SZ:1 I have only installed the user tools. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From noer@cygnus.com Thu May 8 15:01:00 1997 From: noer@cygnus.com (Geoffrey Noer) Date: Thu, 08 May 1997 15:01:00 -0000 Subject: System V IPC (shared memory) References: Message-ID: <199705082158.OAA00880@cirdan.cygnus.com> Beta 18 contains support for memory mapped files using mmap, mprotect, et al, not System V style shared memory. Just to clear up a little confusion... -- Geoffrey Noer noer@cygnus.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From franl@world.std.omit-this.com Thu May 8 15:50:00 1997 From: franl@world.std.omit-this.com (Francis Litterio) Date: Thu, 08 May 1997 15:50:00 -0000 Subject: [b18] Can't interrupt "sleep 30" Message-ID: <33725462.447294735@world.std.com> Using b18, type sleep 30 to bash, then type ^C to interrupt it. Nothing happens, and after 30 seconds elapses, bash prompts again. -- Francis Litterio PGP Key Fingerprint: franl@world.std.omit-this.com 02 37 DF 6C 66 43 CD 2C http://world.std.omit-this.com/~franl/ 10 C8 B5 8B 57 34 F3 21 Note: email address and WWW URL altered to prevent spam. "They that can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety." -- Ben Franklin, ~1784 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From noer@cygnus.com Thu May 8 20:13:00 1997 From: noer@cygnus.com (Geoffrey Noer) Date: Thu, 08 May 1997 20:13:00 -0000 Subject: String.h vs string.h question References: <199705090113.SAA08234@nz1.netzone.com> Message-ID: <199705090313.UAA14307@rtl.cygnus.com> I was asked: > > So is mount -m reenabled, or what? > > Is that the String.h string.h fix, and if so so we need to reinstall to use > it? > > Confused. The header files fix in b18 has nothing to do with mount -m. I am making use of a gcc internal mechanism that allows you to create a mapping between header file names. String.h is really _string.h in the release. I added a mapping in a file called header.gcc which tells gcc that any references to String.h really mean _string.h. Similar conflicts are handled in the same way. Although this is what's happening inside the compiler, everyone should continue to #include the standard names. (The internal mapping should stay internal). -- Geoffrey Noer noer@cygnus.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From khan@xraylith.wisc.edu Thu May 8 21:40:00 1997 From: khan@xraylith.wisc.edu (Mumit Khan) Date: Thu, 08 May 1997 21:40:00 -0000 Subject: b18 release -- good and not-so-good Message-ID: <9705090439.AA13564@modi.xraylith.wisc.edu> Kudos to Cygnus Support, oops -- Solutions, for bringing out b18! Compiled out of the box under i386-linux both cross and win32 native. Haven't tried building under '95 yet, but hopefully that'll work as well. Thanks for the UNC support. Haven't tried it yet, but it surely will solve some of the awkwardness in using gnu-win32 tools. Tcl/Tk now builds!! Now if I can get BLT/TclX/Tix/Itcl, I'm all set. Thanks for the improved signal support. I like being able to hit ^C :-) Some of the problems I had reported, along with fixes, are still there (some were fixed from the original list). Here're are the remaining one: newlib/libc/stdlib/bsearch.c: This fixes the past the end access bug in bsearch(). winsup/cygwin.din: Zap the export of tgetent() winsup/misc.cc: Zap the definition of tgetent(). This conflicts with -ltermcap. Doesn't belong here anyway. texinfo/libtxi/Makefile.in: Replace hard-code AR option "cq" with AR_FLAGS. Only problem when building under a machine with non-GNU ar of course (I usually build under hppa1.1-hp-hpux9.07 with stock ar). Nice job overall. Regards, Mumit -- khan@xraylith.wisc.edu http://www.xraylith.wisc.edu/~khan/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Scott.Mintz@po.cle.ab.com Fri May 9 03:23:00 1997 From: Scott.Mintz@po.cle.ab.com (Scott Mintz) Date: Fri, 09 May 1997 03:23:00 -0000 Subject: Bug in h8300.c Message-ID: <00059C33.1893@po.cle.ab.com> This bug still exists in the b18 sources... ______________________________ Reply Separator _________________________________ Subject: Re: Bug in h8300.c Author: Geoffrey Noer at Internet Date: 5/8/97 3:01 PM This was already fixed in development sources. Thanks anyway... Geoff Scott Mintz wrote: > > I built a cygwin32 cross compiler for the H8300-hms on a WinNT machine > using the 17.1 CDK. > > The file h8300.c in the gcc/config/h8300 subdirectory has the > following defined: > > /* Output assembly language code for the function prologue. */ > static int push_order[FIRST_PSEUDO_REGISTER] = > {0, 1, 2, 3, 4, 5, 6, -1, -1}; > static int pop_order[FIRST_PSEUDO_REGISTER] = > {6, 5, 4, 3, 2, 1, 0, -1, -1}; > > where FIRST_PSEUDO_REGISTER is defined to be 10. > > The last item in the above arrays gets defined as 0 because there are > only 9 initializers instead of 10. > > This causes the generated interrupt prologue/epilogue code to be > incorrect. > > The arrays should be defined this way: > > /* Output assembly language code for the function prologue. */ > static int push_order[FIRST_PSEUDO_REGISTER] = > {0, 1, 2, 3, 4, 5, 6, -1, -1, -1}; > static int pop_order[FIRST_PSEUDO_REGISTER] = > {6, 5, 4, 3, 2, 1, 0, -1, -1, -1}; > > +------------------------------+------------------------------------+ > | Scott A. Mintz | voice: (216) 646-4805 | > | Allen-Bradley Company | fax: (216) 646-4961 | > | 1 Allen-Bradley Drive | email: scott.mintz@po.cle.ab.com | > | Mayfield Hts., OH 44124-6118 | CIS: 71461,632 | > +------------------------------+------------------------------------+ > > - > For help on using this list (especially unsubscribing), send a message to > "gnu-win32-request@cygnus.com" with one line of text: "help". > -- Geoffrey Noer noer@cygnus.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From gpasa@swissonline.ch Fri May 9 03:23:00 1997 From: gpasa@swissonline.ch (Pasa Guglielmo) Date: Fri, 09 May 1997 03:23:00 -0000 Subject: two entries in HKEY_CURRENT_USER Message-ID: <33730DFB.C28@swissonline.ch> Hi, I noticed that I have two entries in win95 registry under HKEY_CURRENT_USER/Software/Cygnus Support/CYGWIN.DLL setup one named b14.0 and another b15.0 I currently have cygwin32 17.1. Is one of them (b14.0) from the old version, and must I erase it ? -- Sincerely, Pasa Guglielmo e-mail address : gpasa@swissonline.ch mailing address: Pasa Guglielmo Rte des Cases 17A CH-1890 St-Maurice (Switzerland) tel.: +41 (0)24 485 50 40 fax: +41 (0)24 485 50 44 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From csoelle@sghms.ac.uk Fri May 9 03:23:00 1997 From: csoelle@sghms.ac.uk (Christian Soeller) Date: Fri, 09 May 1997 03:23:00 -0000 Subject: X11 References: Message-ID: <50d8r1ot4x.fsf@mbcsg1.sghms.ac.uk> Niels Skou Olsen writes: > > Hi, > > I have seen several people on this list mentioning the X11 libraries. > How do people use those libraries? > > * Do they use a commercial X server? > * Is there a free X server available from the net? > * Or do the X libraries map to native Windows API? > There is a free Xserver for win95/NT. From the README in the binary X11R6.3 distrib for cygwin32 (available at ftp://niteroi.gsfc.nasa.gov/pub/win32/X11R6.3/Win32_X11R6_3.tar.gz ): > And of course, you will need X Windows server software. A freewave X > server can be obtained from MicroImages: > > http://tnt.microimages.com/www/html/freestuf/mix.htm > > This server comes with twm. If you prefer fvwm, from COMMAND.COM enter > > startx > > and it will bring up the Xserver with fvwm. Take a look at the file > /usr/X11R6.3/bin/startx.bat if this fails for some reason. I built fvwm > from sources I had from xlibemu (DJGPP stuff), it does not come from > MicroImages. I have tried the server under win95/NT and it seems to work ok. You just need an installed TCP/IP protocol. Christian -------------------------------------------------------------------- Christian Soeller mailto: csoelle@sghms.ac.uk St. Georges Hospital Medical School Dept. of Pharmacology Cranmer Terrace London SW17 0RE - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From noer@cygnus.com Fri May 9 03:23:00 1997 From: noer@cygnus.com (Geoffrey Noer) Date: Fri, 09 May 1997 03:23:00 -0000 Subject: Bug in h8300.c References: <00059C33.1893@po.cle.ab.com> Message-ID: <199705090726.AAA15569@rtl.cygnus.com> > This bug still exists in the b18 sources... The beta 18 sources were made in mid-April. My up to date development sources do already contain the fix so it probably was addressed at some point since then. -gjn - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From karuottu@freenet.hut.fi Fri May 9 03:23:00 1997 From: karuottu@freenet.hut.fi (Kai Ruottu) Date: Fri, 09 May 1997 03:23:00 -0000 Subject: Has anyone successfully build a powerpc-target xgcc Message-ID: <199705090829.LAA05421@freenet.hut.fi> Geoffrey Noer writes: > You might want to read ftp://ftp.cygnus.com/pub/embedded/crossgcc/FAQ > Perhaps all the answers will be found from the FAQ, but some answers for the listed problems here... > Andy Anh Da Ngo wrote: > > > > 1) float.h-ieee not found during build This comes with some cygnus 'how-to-patch- for-a-cross-gcc' script > > 2) libgcc.a not found after one build that ld was > > looking for (sparc-sun-solaris2.4 hosted) when > > I ran power-eabi-gcc on some simple test.c file Cannot say much, some targets come with an asm file with the needed stuff for libgcc.a routines, sometimes there was a need to compile the libgcc1.a part with a 'native C-compiler'... > > 3) crt0.o not found (sparc-sun-solaris2.4 build, i386-go32 host, > > powerpc-eabi target) Canadian cross The C-startup file, crt0.o, belongs to the C-library, perhaps Cygnus newlib-1.70 could be a good choice for this purpose... But quite sure the crt0.S sources must be fixed for the target board the compiler user uses... and all the I/O-primitives too. The C-library really isn't a 'plug-and-play' component for a embedded target... > > 4) config.h not found during linux-hosted build > > when it was trying to compile hash.c > > (a reference of config.h in runtime.h) during > > objective-c compilation Headers belong to the C-library too - making libobjc.a needs the ppc-target gcc, which needs the proper headers for the target... So, for a cross-development system one usually needs binutils, C/C++ compiler, C-library (newlib), C++ classes (libg++) and a debugger (gdb as a remote debugger ). A simulator for the target, like SPIM for MIPS can also be handy. Perhaps a RTOS, like RTEMS is needed too... Anybody heard about a good free powerpc-simulator? - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From pascal.trouvin@integralis.co.uk Fri May 9 07:39:00 1997 From: pascal.trouvin@integralis.co.uk (pascal.trouvin@integralis.co.uk) Date: Fri, 09 May 1997 07:39:00 -0000 Subject: B18 and dlls Message-ID: <01IIO5NTAOMU003VWL@INTEGD.INTEGRALIS.CO.UK> Until this morning, I was able to build dll with b17, with method described in this mailing list. WITH BETA18, things that compile well this morning, are now broken!!! when starting program: a popup window comes saying "Could not find (garbage characters).DLL" Have you such problems? cheers pascal ____________________________________________________________________ "This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. This communication may contain material protected by attorney-client privilege. If you are not the intended recipient or the person responsible for delivering the email to the intended recipient, be advised that you have received this email in error and that any use, dissemination, forwarding, printing, or copying of this email is strictly prohibited. If you have received this email in error please notify the IT manager by telephone on 44 (0) 118 9306060." info@Integralis.com http://www.integralis.com Innovation, Integration, Integralis _____________________________________________________________________ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Aris.Zakinthinos@cl.cam.ac.uk Fri May 9 07:39:00 1997 From: Aris.Zakinthinos@cl.cam.ac.uk (Aris Zakinthinos) Date: Fri, 09 May 1997 07:39:00 -0000 Subject: bind in bash Message-ID: I have been trying to add a binding so that "Home", "End" etc. do the obvious things for command line editing. I have tried everything I can possibly thing of but I can't get it to work. Is bind broken? I suspect that it has something to do with the escape character since bind "\e[[A":"Funtion Key 1" doesn't work. Am I missing something? Thanks in advance, Aris. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From ARichard@stark.cc.oh.us Fri May 9 07:46:00 1997 From: ARichard@stark.cc.oh.us (Richardson, Anthony) Date: Fri, 09 May 1997 07:46:00 -0000 Subject: Where is install-sh script? Message-ID: <199705091255.IAA11898@dialup.oar.net> The readme for b18 mentions an install-sh script that fixes the problem with install trying to install foo instead of foo.exe. I can't find the script though. Any ideas? Thanks Tony Richardson - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From weiku@UTKUX1.UTK.EDU Fri May 9 07:46:00 1997 From: weiku@UTKUX1.UTK.EDU (Wei Ku) Date: Fri, 09 May 1997 07:46:00 -0000 Subject: g77 with Cygnu.b18 Message-ID: <199705091446.KAA22488@solid.ssd.ornl.gov> Hello Mr Khan: As I understand, the g77 you ported needs the corresponding gcc (b.17.1) in the same tar.Z file. Does that mean that the g77 you ported will not work under b.18 ( because of difference between the gcc's ) ? I am willing to try Cygnu b18. However, the g77 you ported is the only FORTRAN compiler I can use in my home PC. ( My old Microsoft FORTRAN compiler can not handle array of size larger than 64 K bytes). Also, I am porting a FORTRAN package that is originally for UNIX. Please let me know if I can install the Cygnu b18 and kepp g77 working. Sincerely, Wei Ku *************************************** Department of Physics and Astronomy The University of Tennessee 1408 Circle Drive Knoxville, Tennessee 37996-1200 weiku@utkux.utcc.utk.edu --------------------------------------- Solid State Division Oak Ridge National Laboratory P.O.Box 2008 Oak Ridge, TN 37831-6032 Phone: (423) 574-5795 Fax: (423) 574-4143 weiku@solid.ssd.ornl.gov From jbharvey@gte.net Fri May 9 10:49:00 1997 From: jbharvey@gte.net (Justin B. Harvey) Date: Fri, 09 May 1997 10:49:00 -0000 Subject: Standalone executables Message-ID: <199705091513.KAA16116@smtp.gte.net> Can someone refresh my memory as to the location of the instructions on how to create executable that do not need the cygwin library? j -- Justin B. Harvey jbharvey@gte.net http://home1.gte.net/jbharvey - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From blake@edge.net Fri May 9 10:49:00 1997 From: blake@edge.net (Blake McBride) Date: Fri, 09 May 1997 10:49:00 -0000 Subject: Beta 18 an shell pipe (text/binary) mode? Message-ID: <199705091616.LAA21514@edge.edge.net> I had a some problems with the runtime version of 17.1 and was wondering whether or not it was fixed. In 17.1 shell pipes (|) were opened in text (as apposed to binary mode). This killed stuff like gzip | tar. Has this been changed (perhaps a config option added)? Thanks. --blake -- Download source code to my Dynace Object Oriented Extension to C from: http://www.edge.net/algorithms Blake McBride (blake@edge.net) Algorithms Corporation - 615-791-1636 - USA - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From b.mcleod@opengroup.org Fri May 9 10:49:00 1997 From: b.mcleod@opengroup.org (Bruce D. McLeod) Date: Fri, 09 May 1997 10:49:00 -0000 Subject: ncurses on b18 Message-ID: <2.2.32.19970509175023.00940630@postman.osf.org> Does anyone know if the same diffs for termcaps and terminfo are needed to make ncurses run under b18? Has anyone tried? Bruce - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jont@harlequin.co.uk Fri May 9 10:49:00 1997 From: jont@harlequin.co.uk (Jon Thackray) Date: Fri, 09 May 1997 10:49:00 -0000 Subject: Picking up include directories automatically Message-ID: <24774.9705091320@dedekind.cam.harlequin.co.uk> Is there some way to persuade cpp to look in some include directories automatically? The old one used to look at the value of the include environment varaible, the latest (b17.1) appears to ignore this. Putting these directories on the command line is not an option, as they vary from machine to machine. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From gunther.ebert@ixos-leipzig.de Fri May 9 10:58:00 1997 From: gunther.ebert@ixos-leipzig.de (Gunther Ebert) Date: Fri, 09 May 1997 10:58:00 -0000 Subject: bash (b18) and the Alt-Gr key Message-ID: <337316F1.2CCF@ixos-leipzig.de> Hi, I noticed that bash (b18) obviously doesn't like Alt-Gr key combinations. Is there any way to get bash accepting Alt-Gr key combinations (termcap stuff or whatever)? With some international keyboard layouts you need the Alt-Gr key to input some important characters (for example |{}[]). Alt+numpad doesn't work either. It seems to be a problem related only to bash, other utilities (like cat) work fine. Gunther -- Gunther Ebert iXOS Anwendungs-Software GmbH Angerstrasse 40-42 D-04177 Leipzig Phone : +49 341 48503-0 Fax : +49 341 48503-99 E-mail: mailto:gunther.ebert@ixos-leipzig.de www : http://www.ixos-leipzig.de - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From gb@cs.unc.edu Fri May 9 10:58:00 1997 From: gb@cs.unc.edu (Gary Bishop) Date: Fri, 09 May 1997 10:58:00 -0000 Subject: X11R6.3 nolonger works under B18 Message-ID: <3.0.1.32.19970509135623.00f00d64@mail.cs.unc.edu> I get a pop-up dialog saying Error Starting Program "Unable to run d:\USR\X11R6.3\BIN\xwd.exe" Bash says: bash.exe: /USR/X11R6.3/BIN/xwd.exe: (null) Same behavior on all x programs I try. They worked with B17. I have tried rebuilding them with B18 (with great pain) but to no good effect. I'm running on Win95. My PATH is /GNUWIN32/B18/H-I386-CYGWIN32/BIN:/GNUWIN32/B18/TCL/BIN:/USR/X11R6.3/BIN://C /CAFE/BIN://C/CAFE/JAVA/BIN://C/WINDOWS://C/WINDOWS/COMMAND:. thanks gb - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From crvich@raleigh.ibm.com Fri May 9 10:58:00 1997 From: crvich@raleigh.ibm.com (Ernest Crvich) Date: Fri, 09 May 1997 10:58:00 -0000 Subject: environ Message-ID: <9705091311.AA24558@ode1.raleigh.ibm.com> I tried beta 18 out on my sample environ test, and it still fails. This program works fine on OS/2 (EMX 0.9c), AIX (gcc 2.6.0) Solaris (gcc 2.7.2), and HP-UX (native cc)...i.e., the envvar ZZZZZ appears in the list that DumpEnv() prints out. On Win NT 4.0 with cygwin32 beta 17.1 or 18, ZZZZZ does not appear in the env dump. Strangely, if an "execlp( "cmd", "cmd", 0 );" is inserted into main() (instead of or after DumpEnv()), the cmd shell *does* show the ZZZZZ variable. Here it is: -------------------------- #include #include #include DumpEnv() { int i; for (i=0; environ[i]; ++i) printf( "%s\n", environ[i] ); } main() { putenv( "ZZZZZ=value" ); DumpEnv(); } -------------------------- - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From david@coent.demon.co.uk Fri May 9 11:50:00 1997 From: david@coent.demon.co.uk (David Coe) Date: Fri, 09 May 1997 11:50:00 -0000 Subject: X11R6.3 , G77 and Gnu-win32 B18 References: <9705012016.AA00653@modi.xraylith.wisc.edu> Message-ID: <3372FBC8.49E3@coent.demon.co.uk> Many thanks to all those who pointed me to Mumit Khan's web page on G77 - really excellent stuff! I was particularly delighted as the PGplot libraries I use extensively on Linux compiled up, installed and ran without a murmur. Alas, the Gnu-win32 X11R6.3 libraries on ftp://niteroi.gsfc.nasa.gov/pub/win32/X11R6.3/ now seem not to be compatible with Beta 18. Geoffrey's new install brings up the new release flawlessly as far as the command-line utilities are concerned, but - at least for me - all X applications now GPF! Does anyone else see this? -- Dr David Coe \=\ 58 Fairlawn Drive, East Grinstead \=\ Tel +44 1342 326860 West Sussex, RH19 1NT, United Kingdom \=\ Fax +44 1342 316019 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From chs11@columbia.edu Fri May 9 11:50:00 1997 From: chs11@columbia.edu (Carl H. Sayres) Date: Fri, 09 May 1997 11:50:00 -0000 Subject: strange linker messages Message-ID: <199705091849.OAA22393@mailhub2.cc.columbia.edu> I'm getting the following liker messages which I don't understand. (I have no impure.c or _impure_ptr anywhere in my code!) Can soemone explain this to me? Thanks a bunch. - Carl $ cc -O2 -c sblat2.c $ cc sblat2.o ../../blas_alpha.a ../../F2CLIBS/libI77.a \ ../../F2CLIBS/libF77.a -lm -lc -o ../xblat2s C:\usr\GNU\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32\cygnus-2.7.2- 970404\../../../../i386-cygwin32\lib/libcygwin.a (libccrt0.o)(.data+0x0):libccrt0.cc: multiple definition of `_impure_ptr' C:\usr\GNU\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32\cygnus-2.7.2- 970404\../../../../i386-cygwin32\lib/libc.a(impure.o)(.data+0x2ec):impure.c: first defined here cc: Internal compiler error: program ld got fatal signal 1 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From banders@switch.rockwell.com Fri May 9 11:50:00 1997 From: banders@switch.rockwell.com (Bart Anderson) Date: Fri, 09 May 1997 11:50:00 -0000 Subject: connect() failure Message-ID: <33735930.5007@switch.rockwell.com> Has anyone encountered this problem with connect? I checked the faq and the archive, but couldn't find anything that popped out at me. Porting a unix tool I had written on the sun, I had to cast the sockaddr_in sturcture to sockaddr structure which I see has padding for the sockaddr_in. Are these fields not used? If not, how would I get similar functionality? _ /* connect to the host */ _ ret = connect( sd, ( struct sockaddr * ) (struct sockaddr_in * ) &sd_name,_ sizeof( struct sockaddr_in )); _ _ Thanks, Bart. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From dbaker@hobbes.cuckoo.com Fri May 9 11:55:00 1997 From: dbaker@hobbes.cuckoo.com (Daniel Baker) Date: Fri, 09 May 1997 11:55:00 -0000 Subject: [b18] Can't interrupt "sleep 30" References: <33725462.447294735@world.std.com> Message-ID: <199705091854.NAA12887@hobbes.cuckoo.com> I see the same problems. Also, I can use tar or make or other gnu utilities, they just run off with errors and stack traces. Daniel > > Using b18, type > > sleep 30 > > to bash, then type ^C to interrupt it. Nothing happens, and after 30 > seconds elapses, bash prompts again. > -- > Francis Litterio PGP Key Fingerprint: > franl@world.std.omit-this.com 02 37 DF 6C 66 43 CD 2C > http://world.std.omit-this.com/~franl/ 10 C8 B5 8B 57 34 F3 21 > Note: email address and WWW URL altered to prevent spam. > > "They that can give up essential liberty to obtain a little temporary > safety deserve neither liberty nor safety." -- Ben Franklin, ~1784 > - > For help on using this list (especially unsubscribing), send a message to > "gnu-win32-request@cygnus.com" with one line of text: "help". > -- Daniel Baker (dbaker@cuckoo.com) Chief Technology and Executive Officer -- CuckooNet! ( http://www.cuckoo.com ) - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From gb@cs.unc.edu Fri May 9 14:21:00 1997 From: gb@cs.unc.edu (Gary Bishop) Date: Fri, 09 May 1997 14:21:00 -0000 Subject: mail archives not working? Message-ID: <3.0.32.19970509164349.00c9991c@mail.cs.unc.edu> The archives I can see appear to stop in december 96. Am I looking in the right place? thanks gb - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jbharvey@gte.net Fri May 9 14:21:00 1997 From: jbharvey@gte.net (Justin B. Harvey) Date: Fri, 09 May 1997 14:21:00 -0000 Subject: termcap situations... Message-ID: <199705092120.QAA10919@smtp.gte.net> How does compiling programs from bash and running them work with termcap? I've installed ncurses and the program that needs curses still doesn't work....am I missing a linked library somewhere? j -- Justin B. Harvey jbharvey@gte.net http://home1.gte.net/jbharvey - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jdennis@ultranet.com Fri May 9 14:21:00 1997 From: jdennis@ultranet.com (John R. Dennis) Date: Fri, 09 May 1997 14:21:00 -0000 Subject: how do you run .bat files under bash? Message-ID: <199705092038.QAA00074@lucius.ultra.net> I'm stumped. From within a bash shell how do I execute a .bat script? BTW, I don't want to have to do anything verbose like invoke cmd.com with the name of the script, I just want to be able to type the name of the script on the command line and/or be able to invoke it from within another script (or via a "system" call). It seems to me that the convention of #!/path/to/script/interpretor as the first line won't work if the script is invoked from a normal win32 environment. Basically what I'm looking for is the ability to invoke scripts from any environment without regard to the environment I'm in. How do I do this? John - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From hoeckeg@sebb.bel.alcatel.be Fri May 9 16:15:00 1997 From: hoeckeg@sebb.bel.alcatel.be (Guido Van Hoecke) Date: Fri, 09 May 1997 16:15:00 -0000 Subject: bash problem with hexl under ntemacs Message-ID: Triggered by some postings about hexl-mode, I have been doing some experiments with it under "GNU Emacs 19.34.1 (i386-*-nt4.0) of Mon Apr 21 1997 on ESME" on a windows95 portable. I was using Cygnus gnu-win32 v17.1. I use bash as my shell under emacs, and am quite happy with it. I created a new file with C-x C-f newfile, and then I switched the buffer to hexl-mode with M-x hexl-mode. I got following message in the buffer itself: "/cygnus/h-i386-cygwin32/bin/bash.exe: c:/emacs/bin/hexl.exe: No such file or directory". So I tried an existing file, visited it (C-x C-f), and typed M-x hexlify-buffer: buffer gets cleared and same message is displayed at the top of the buffer. Out of curiosity, I switched to MKS toolkit's sh.exe, and bingo, hexlify-buffer does work allright. So I hopped over to www.cygnus.com to see whether I could find anything about this, and saw that there was a new version b18: I downloaded and installed it. But the hexl-mode / hexlify-buffer still causes a problem, allthough the message has changed: "/gnuwin32/b18/h-i386-cygwin32/bin/bash.exe: c:/emacs/bin/hexl.exe: (null)" Follows a transcript of my shell buffer to show the environment as well as the fact that the hexl program is found by bash (note that I have manually split long lines into continued indented lines, and that my shell prompt is configured to show the command nr and working directory on one $ line, followed by the real prompt on a second $ line): $ 1 [/hoeckeg] $ env EMACSDATA=C:/EMACS/etc INFOPATH=c:/hoeckeg/info;/usr/local/info;c:/gnuwin32/b18/info;c:/hoecke\ g/lisp/w3;c:/hoeckeg/lisp/bbdb;c:/hoeckeg/lisp/gnus/texi;C:/EMA CS/info ENV=c:/hoeckeg/.bashrc HISTSIZE=64 LOGNAME=hoeckeg EMACS=t GDBTK_LIBRARY=c:/gnuwin32/b18/share/gdbtcl TEMP=C:/windows/temp/ TMP=C:\WINDOWS\TEMP EMACSPATH=C:/EMACS/bin CMDLINE=WIN TERMCAP=emacs:co#80:tc=unknown: PROMPT=$p$g TERM=emacs __PARENT_PROCESS_ID=-92789 EMACS_DIR=C:/EMACS TCL_LIBRARY=c:/gnuwin32/b18/tcl/lib/tcl7.6 HOSTTYPE=i386 PATH=/GNUWIN32/B18/H-I386-CYGWIN32/BIN:/GNUWIN32/B18/TCL/BIN:/BIN/MKS/M\ KSNT:/BIN:/WINDOWS:/WINDOWS/COMMAND:/TL27:/EMACS/BIN:/ELVIS:/LO\ CAL/RCS/BIN:/BIN/NETSCAPE/PROGRAM:/program files/maestro.95 HOME=c:/hoeckeg INPUTRC=~/.inputrc SHELL=C:/gnuwin32/b18/H-i386-cygwin32/bin/bash.exe WINBOOTDIR=C:\WINDOWS WINDIR=C:\WINDOWS USER=hoeckeg TMPDIR=c:/windows/temp GCC_EXEC_PREFIX=c:\gnuwin32\b18\H-i386-cygwin32\lib\gcc-lib\ EMACSDOC=C:/EMACS/etc EMACSLOADPATH=C:/EMACS/lisp OSTYPE=win32 NNTPSERVER=btmpjg.god.bel.alcatel.be COMSPEC=C:\WINDOWS\COMMAND.COM BLASTER=A220 I5 D1 SHLVL=1 ROOTDIR=c:/bin/mks EMACSLOCKDIR=C:/EMACS/lock _=/GNUWIN32/B18/H-I386-CYGWIN32/BIN/env.exe PID=1001 $ 2 [/hoeckeg] $ type $SHELL C:/gnuwin32/b18/H-i386-cygwin32/bin/bash.exe is C:/gnuwin32/b18/H-i386-\ cygwin32/bin/bash.exe $ 3 [/hoeckeg] $ type bash bash is /GNUWIN32/B18/H-I386-CYGWIN32/BIN/bash.exe $ 4 [/hoeckeg] $ type hexl hexl is /EMACS/BIN/hexl.exe $ 5 [/hoeckeg] $ In my .emacs, I have following statements with respect to shell-mode: (setenv "SHELL" "C:/gnuwin32/b18/H-i386-cygwin32/bin/bash.exe") ;(setenv "SHELL" "c:/bin/mks/mksnt/sh.exe") (setenv "ENV" "c:/hoeckeg/.bashrc") (if (eq system-type 'windows-nt) (progn (setq igrep-expression-quote-char ?' igrep-parenthesis-escape-char ?\\ igrep-expression-quote-char ?' igrep-parenthesis-escape-char ?\\ explicit-sh-args '("-login" "-i") ;for bash ;explicit-sh-args '("-L" "-i") ;for ksh (mks) shell-command-switch "-c" ;For bash etc. /c for cmd. win32-quote-process-args t ;Runs bash subshell. ) (add-hook 'shell-mode-hook '(lambda () (setq comint-file-name-quote-list nil) (setq comint-completion-addsuffix t)) t))) Is this a known problem! Is there a workaround or a known fix? g-) - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From colin@bird.fu.is.saga-u.ac.jp Fri May 9 22:04:00 1997 From: colin@bird.fu.is.saga-u.ac.jp (Colin Peters) Date: Fri, 09 May 1997 22:04:00 -0000 Subject: Standalone executables Message-ID: <01BC5D47.5DD17300@gbird0> Justin B. Harvey[SMTP:jbharvey@gte.net] wrote: >Can someone refresh my memory as to the location of the instructions on how >to create executable that do not need the cygwin library? You can find the Minimalist GNU-Win32 package (version 0.1.3) at http://www.geocities.com/Tokyo/Towers/6162/gcc.html or http://www.fu.is.saga-u.ac.jp/~colin/gcc.html NOTE: Version 0.1.3 was made for beta 17.1. It will work with beta 18 but the installation procedure is somewhat different. You can either figure it out yourself or wait for 0.1.4, which should be out soon. Colin. -- Colin Peters - colin@bird.fu.is.saga-u.ac.jp -- Saga University Dept. of Information Science -- http://www.fu.is.saga-u.ac.jp/~colin/index.html -- http://www.geocities.com/Tokyo/Towers/6162/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jeffdb@netzone.com Fri May 9 22:04:00 1997 From: jeffdb@netzone.com (Mikey) Date: Fri, 09 May 1997 22:04:00 -0000 Subject: ncurses on b18 Message-ID: <199705100251.TAA12808@nz1.netzone.com> No, they moved all of the console handleing routines from fhandler.cc into console.cc, so the diffs no longer work, but I guess somebody rewrote them (much neater job than I did ;^) and included them in the prebeta18 sources, so they are probably in B18 also, the diff's for the termcap/info files should still give you w32ansi support (function keys), but I guess the default termcap that comes with the package, is supposed to be linux, so if you want to use telnet or rlogin, you'll have to upload the termcap in the package to.termcap in your home directory. or change the TERM environment var to vt102 which is linux compatible (mostly). the [J and [0J were broken in p18, so clear.exe won't work. The fixes to Fhandler.cc FakeReadFile, both made it in, so ncurses apps should work with the out of the box .dll, sadly the O_NDELAY fix wasn't there in p18, so info.exe still needs that fix to winsup/ to work. (It's a hack, so I can understand not wanting to use it) ---- From: Bruce D. McLeod To: gnu-win32@cygnus.com Date: Friday, May 09, 1997 5:23 PM Subject: ncurses on b18 >Does anyone know if the same diffs for termcaps and terminfo >are needed to make ncurses run under b18? Has anyone tried? > >Bruce > >- >For help on using this list (especially unsubscribing), send a message to >"gnu-win32-request@cygnus.com" with one line of text: "help". > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From maruzz@matrice.it Fri May 9 22:04:00 1997 From: maruzz@matrice.it (Giovanni Maruzzelli) Date: Fri, 09 May 1997 22:04:00 -0000 Subject: Apache1.2b10 on NT with Cygnus gnu-win32 References: <199705090726.AAA15569@rtl.cygnus.com> Message-ID: Hi, this port is available on: ftp://aaue.matrice.it/pub/apache_nt and on: http://www.matrice.it/risorse/apache-nt The README: =============================================================== may 10 1997 Hi there, this is a porting of apache1.2b10 to cygwin32, the UNIX-like environment for developing and porting software on win32 (windows NT and a fair part on win95). This is by no means well tested, but at least it work. If you like hacking the code, you can apply the apache.diff on a standard 1.2b10 distribution. Then type ./Configure, as usual, then make. Otherwise you can grab the httpd.exe, and the config files. ************************************************* The file apachent.tgz is a tarred and gzipped archive that contains all the other files but not the cygwin.dll. If you grab the binary (httpd.exe) and you have not installed the Cygnus gnu-win32 distribution, you will need to have the cygwin.dll in you path. ************************************************** An annoying thing is that this port run in a console window, will be very better when transformed in a service. Please send e-mail with bug reports (but PATCHES will be gratefully received) to: maruzz@matrice.it ciao, -giovanni maruzzelli *********************************************** * * * * * MATRICE srl * * Information Technology Solutions * * * * Giovanni Maruzzelli * * * * * * * * maruzz@matrice.it * * http://www.matrice.it * *********************************************** - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jeffdb@netzone.com Fri May 9 22:04:00 1997 From: jeffdb@netzone.com (Mikey) Date: Fri, 09 May 1997 22:04:00 -0000 Subject: Picking up include directories automatically Message-ID: <199705100251.TAA12832@nz1.netzone.com> *cpp: -I /some/include/dir H-i386-cygwin32/lib/gcc-lib/i386-cygwin32/version-XXX/specs ---- From: Jon Thackray To: gnu-win32@cygnus.com Date: Friday, May 09, 1997 5:12 PM Subject: Picking up include directories automatically >Is there some way to persuade cpp to look in some include directories >automatically? The old one used to look at the value of the include >environment varaible, the latest (b17.1) appears to ignore this. >Putting these directories on the command line is not an option, as >they vary from machine to machine. >- >For help on using this list (especially unsubscribing), send a message to >"gnu-win32-request@cygnus.com" with one line of text: "help". > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jeffdb@netzone.com Fri May 9 22:04:00 1997 From: jeffdb@netzone.com (Mikey) Date: Fri, 09 May 1997 22:04:00 -0000 Subject: B18 and dlls Message-ID: <199705100252.TAA12869@nz1.netzone.com> I think you forgot about fixup.o maybe install shield deleted it, or you did when you reinstalled. ---- From: pascal.trouvin@integralis.co.uk To: gnu-win32@cygnus.com Date: Friday, May 09, 1997 5:03 PM Subject: B18 and dlls > > Until this morning, I was able to build dll with b17, with method > described in this mailing list. > > > WITH BETA18, things that compile well this morning, are now broken!!! > > when starting program: a popup window comes saying "Could not find > (garbage characters).DLL" > > > Have you such problems? > > > cheers > > pascal >____________________________________________________________________ >"This email and any files transmitted with it are confidential and >intended solely for the use of the individual or entity to whom they >are addressed. This communication may contain material protected by >attorney-client privilege. If you are not the intended recipient or >the person responsible for delivering the email to the intended >recipient, be advised that you have received this email in error and >that any use, dissemination, forwarding, printing, or copying of this >email is strictly prohibited. If you have received this email in error >please notify the IT manager by telephone on 44 (0) 118 9306060." > >info@Integralis.com > http://www.integralis.com > > Innovation, Integration, Integralis >_____________________________________________________________________ >- >For help on using this list (especially unsubscribing), send a message to >"gnu-win32-request@cygnus.com" with one line of text: "help". > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From khan@xraylith.wisc.edu Fri May 9 22:04:00 1997 From: khan@xraylith.wisc.edu (Mumit Khan) Date: Fri, 09 May 1997 22:04:00 -0000 Subject: g77 with Cygnu.b18 References: <199705091446.KAA22488@solid.ssd.ornl.gov> Message-ID: <9705092352.AA23450@modi.xraylith.wisc.edu> "Wei Ku" writes: > > Hello Mr Khan: > As I understand, the g77 you ported needs the corresponding gcc (b.17.1) = > in the same tar.Z file. Does that mean that the g77 you ported will not = > work under b.18 ( because of difference between the gcc's ) ? > I am willing to try Cygnu b18. However, the g77 you ported is the only = > FORTRAN compiler I can use in my home PC. ( My old Microsoft FORTRAN = > compiler can not handle array of size larger than 64 K bytes). Also, I = > am porting a FORTRAN package that is originally for UNIX. > Please let me know if I can install the Cygnu b18 and kepp g77 working. > Sincerely, > Wei Ku > The current plan is to release g77 0.5.19.1, NOT g77 0.5.20, for b18. g77 0.5.20 has some nasty bugs, and even Craig Burley (g77 author) suggests in a posting in comp.lang.fortran using 0.5.19.1 until 0.5.21 i arrives. Even when 0.5.21 arrives, I still won't port it to gnu-win32 until FSF releases 2.8 (and gnu-win32 hopefully will have merged its changes with the FSF release). The back-end differences are too time-consuming, not to mention error-prone, to track for a non-compiler hacker like me. That said, watch for g77-0.5.19.1 source patches and binaries for b18 sometime next week. Will announce here. btw, since b18 still has buggy bsearch, my copy of g77 will use its own fixed one, and functions/subroutines with names like zswap will compile ok. Regards, Mumit -- khan@xraylith.wisc.edu http://www.xraylith.wisc.edu/~khan/ http://www.xraylith.wisc.edu/~khan/software/gnu-win32/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From noer@cygnus.com Fri May 9 22:04:00 1997 From: noer@cygnus.com (Geoffrey Noer) Date: Fri, 09 May 1997 22:04:00 -0000 Subject: Where is install-sh script? References: <199705091255.IAA11898@dialup.oar.net> Message-ID: <199705100504.WAA09972@rtl.cygnus.com> Richardson,Anthony wrote: > > The readme for b18 mentions an install-sh script that fixes the problem > with install trying to install foo instead of foo.exe. I can't find the > script though. Any ideas? It's the top-level install-sh in either of the source tarballs... -- Geoffrey Noer noer@cygnus.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From hemal@superlink.com Fri May 9 22:09:00 1997 From: hemal@superlink.com (Hemal Pandya) Date: Fri, 09 May 1997 22:09:00 -0000 Subject: where is warning: control reaches end of non-void function Message-ID: To my horror the following code compiles without a warning: int Foo() { } on spark, gcc-2.7.2 produces warning: control reaches end of non-void function `Foo()' am i missing something? (like an option maybe?) TIA. -- Hemal Pandya mailto:hemal@superlink.net phone:(908)340-9455 First He created computers, but they couldn't make mistakes So He created programmers - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From colin@bird.fu.is.saga-u.ac.jp Fri May 9 22:23:00 1997 From: colin@bird.fu.is.saga-u.ac.jp (Colin Peters) Date: Fri, 09 May 1997 22:23:00 -0000 Subject: Dlltool not removing temporary files Message-ID: <01BC5D4D.9C9E0D40@gbird0> The version of dlltool that comes with b18 doesn't seem to be removing it's temporary dh.o dst.o and ds0.o ds1.o etc. object files. Do other people have this problem? It is only cosmetic, but somewhat annoying, especially since the old version did it right. NOTE: I am not calling dlltool from bash, but from the command line. Here is a sample command line: dlltool --dllname dll.dll --def dll.def --output-lib libdll.a -k dll.def is this: EXPORTS Add and the above call to dlltool leaves behind these object files: dh.o ds0.o dt.o NOTE: dlltool has a --nodelete option for keeping around temporary files, but I am not using it. Any help or suggestions appreciated. Colin. -- Colin Peters - colin@bird.fu.is.saga-u.ac.jp -- Saga University Dept. of Information Science -- http://www.fu.is.saga-u.ac.jp/~colin/index.html -- http://www.geocities.com/Tokyo/Towers/6162/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From colin@bird.fu.is.saga-u.ac.jp Sat May 10 00:24:00 1997 From: colin@bird.fu.is.saga-u.ac.jp (Colin Peters) Date: Sat, 10 May 1997 00:24:00 -0000 Subject: B18 and dlls Message-ID: <01BC5D4C.350D6820@gbird0> pascal.trouvin@integralis.co.uk wrote: > Until this morning, I was able to build dll with b17, with method > described in this mailing list. > > WITH BETA18, things that compile well this morning, are now broken!!! > > when starting program: a popup window comes saying "Could not find > (garbage characters).DLL" > > Have you such problems? Just reporting that no, I don't have any problems, at least building my simple DLL test with b18. Works fine. The error message you give reminds me of the kind of thing that happens when the import list is not properly terminated, that problem behaved very weirdly (it would go away if the first DLL imported only had one or two references, and come back if there were more) so I wonder if maybe the change in beta 18 has simply made an earlier harmless bug become apparent in your code. Another suggestion is that maybe you didn't recompile all your libraries and object code. If I remember correctly there was a change in the bfd libraries in this release. That the change would have anything to do with your problem is wild speculation however. Colin. -- Colin Peters - colin@bird.fu.is.saga-u.ac.jp -- Saga University Dept. of Information Science -- http://www.fu.is.saga-u.ac.jp/~colin/index.html -- http://www.geocities.com/Tokyo/Towers/6162/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From colin@bird.fu.is.saga-u.ac.jp Sat May 10 00:24:00 1997 From: colin@bird.fu.is.saga-u.ac.jp (Colin Peters) Date: Sat, 10 May 1997 00:24:00 -0000 Subject: strange linker messages Message-ID: <01BC5D5E.8AFBD480@gbird0> Carl H. Sayres[SMTP:chs11@columbia.edu] wrote: >I'm getting the following liker messages which I don't understand. >(I have no impure.c or _impure_ptr anywhere in my code!) >Can soemone explain this to me? Thanks a bunch. > >$ cc -O2 -c sblat2.c >$ cc sblat2.o ../../blas_alpha.a ../../F2CLIBS/libI77.a \ >../../F2CLIBS/libF77.a -lm -lc -o ../xblat2s > >C:\usr\GNU\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32\cygnus-2.7.2- >970404\../../../../i386-cygwin32\lib/libcygwin.a >(libccrt0.o)(.data+0x0):libccrt0.cc: multiple definition of `_impure_ptr' > >C:\usr\GNU\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32\cygnus-2.7.2- >970404\../../../../i386-cygwin32\lib/libc.a(impure.o)(.data+0x2ec):impure.c: >first defined here If I'm not mistaken the problem here is -lc on your link command line. You don't need to link libc directly since it is already part of libcygwin, which gets linked automatically. Just a guess. Colin. PS. Question to GNU-Win32 crowd: Is this also true of libm... or not? -- Colin Peters - colin@bird.fu.is.saga-u.ac.jp -- Saga University Dept. of Information Science -- http://www.fu.is.saga-u.ac.jp/~colin/index.html -- http://www.geocities.com/Tokyo/Towers/6162/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From csoelle@sghms.ac.uk Sat May 10 03:17:00 1997 From: csoelle@sghms.ac.uk (Christian Soeller) Date: Sat, 10 May 1997 03:17:00 -0000 Subject: b18: setpwend instead of setpwent: a typo? Message-ID: <50lo5neiss.fsf@mbcsg1.sghms.ac.uk> In libcygwin.a there is now the function setpwend but my unix manpages only have setpwent and the package I am compiling (perl 5.03_99) wants setpwent as well. Is this a typo that made its way into beta18 ???? Christian -------------------------------------------------------------------- Christian Soeller mailto: csoelle@sghms.ac.uk St. Georges Hospital Medical School Dept. of Pharmacology Cranmer Terrace London SW17 0RE - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From chuck@hysteria.spb.ru Sat May 10 04:05:00 1997 From: chuck@hysteria.spb.ru (Chuck Bogorad) Date: Sat, 10 May 1997 04:05:00 -0000 Subject: bind in bash Message-ID: <3374B81D@hysteria.spb.ru> Aris.Zakinthinos@cl.cam.ac.uk wrote: Ac> I have been trying to add a binding so that "Home", "End" etc. do Ac> the obvious things for command line editing. I have tried Ac> everything I can possibly thing of but I can't get it to work. Is Ac> bind broken? here's how it's done: # home key bind '"\e[1~":beginning-of-line' # del key bind '"\e[3~":delete-char' # end key bind '"\e[4~":end-of-line' # pgup key bind '"\e[5~":history-search-forward' # pgdn key bind '"\e[6~":history-search-backward' - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jeffdb@netzone.com Sat May 10 04:25:00 1997 From: jeffdb@netzone.com (Mikey) Date: Sat, 10 May 1997 04:25:00 -0000 Subject: termcap situations... Message-ID: <199705101119.EAA16862@nz1.netzone.com> When asking for help from the list, it is wise to specify a few details, Platform, environment, program output, the name of the program that you are working on, wheather the ncurses test programs worked etc. from what you have said, it is impossible to tell whats going on, my programs compiled against ncurses from b17.1, all still work ie less elvis man etc. ---- From: Justin B. Harvey To: gnu-win32@cygnus.com Date: Friday, May 09, 1997 9:43 PM Subject: termcap situations... >How does compiling programs from bash and running them work with termcap? >I've installed ncurses and the program that needs curses still doesn't >work....am I missing a linked library somewhere? > >j >-- >Justin B. Harvey >jbharvey@gte.net > http://home1.gte.net/jbharvey >- >For help on using this list (especially unsubscribing), send a message to >"gnu-win32-request@cygnus.com" with one line of text: "help". > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From home@morla.crec.mipt.ru Sat May 10 05:55:00 1997 From: home@morla.crec.mipt.ru (Home user) Date: Sat, 10 May 1997 05:55:00 -0000 Subject: GCC and pipes Message-ID: <199705101254.QAA00385@morla.crec.mipt.ru> Hello! 1st problem: I'm trying to use 'gcc -pipe' on beta 18 to compile sources, and it dies silently. Is this a side-effect of gcc's spawn()ing instead of fork()ing subprocesses? Or am I missing something? PS: Is there any way of getting patches from beta 17.1 to beta 18 source tree? -- Igor V. Kovalenko < mailto:iko@crec.mipt.ru > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jbharvey@gte.net Sat May 10 09:20:00 1997 From: jbharvey@gte.net (Justin Harvey) Date: Sat, 10 May 1997 09:20:00 -0000 Subject: termcap situations... Message-ID: <199705101619.LAA19379@smtp.gte.net> I'm trying to port Tinyfugue to Windows95/NT (a mud client), I can configure it for hardcoded terminal but thought that termcap might be easier. If I compile it for termcap and try to go into it's visual mode (split screen) it says it's not supported. So I grabbed ncurses and compiled it, the test programs didn't work and they said they couldn't open my vt100 screen...I'm wondering if there is any special libraries I need to include while compiling. (i also tried linking ncurses with tinyfugue with termcap and it didn't work). j ---------- > From: Mikey > To: Justin B. Harvey > Cc: cygnus > Subject: Re: termcap situations... > Date: Saturday, May 10, 1997 1:16 AM > > When asking for help from the list, it is wise to specify a few details, > Platform, > environment, program output, the name of the program that you are working > on, wheather the ncurses test programs worked etc. > > from what you have said, it is impossible to tell whats going on, my > programs compiled against ncurses > from b17.1, all still work ie less elvis man etc. > > ---- > From: Justin B. Harvey > To: gnu-win32@cygnus.com > Date: Friday, May 09, 1997 9:43 PM > Subject: termcap situations... > > >How does compiling programs from bash and running them work with termcap? > >I've installed ncurses and the program that needs curses still doesn't > >work....am I missing a linked library somewhere? > > > >j > >-- > >Justin B. Harvey > >jbharvey@gte.net > > http://home1.gte.net/jbharvey > >- > >For help on using this list (especially unsubscribing), send a message to > >"gnu-win32-request@cygnus.com" with one line of text: "help". > > > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From noer@cygnus.com Sat May 10 12:40:00 1997 From: noer@cygnus.com (Geoffrey Noer) Date: Sat, 10 May 1997 12:40:00 -0000 Subject: strange linker messages References: <01BC5D5E.8AFBD480@gbird0> Message-ID: <199705101940.MAA28991@rtl.cygnus.com> Colin Peters wrote: > > PS. Question to GNU-Win32 crowd: Is this also true of libm... or not? Yes. -- Geoffrey Noer noer@cygnus.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From khan@xraylith.wisc.edu Sat May 10 13:29:00 1997 From: khan@xraylith.wisc.edu (Mumit Khan) Date: Sat, 10 May 1997 13:29:00 -0000 Subject: g77 0.5.19.1 for b18 -- 1st cut Message-ID: <9705101932.AA01666@modi.xraylith.wisc.edu> I've uploaded my work in progress to my web page, so if you feel brave enough to try, go for it. http://www.xraylith.wisc.edu/~khan/software/gnu-win32 Please report problems asap so I can fix them before making the next release. Enjoy Mumit - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From pchow@isgtec.com Sat May 10 13:29:00 1997 From: pchow@isgtec.com (Philip Chow) Date: Sat, 10 May 1997 13:29:00 -0000 Subject: gcc -E slower in beta 18? Message-ID: I just recently had some time to move all my tools from beta 14 to beta18 and I am noticing a dramatic performance degradation for C-preprocessing of files with gcc -E. If I have the following command: gcc -D_WIN32 -E -U__GNUC__ -nostdinc -W3 -GX -G5 -DWIN32 -D__STDC__=0 \ -D_NTSDK -D_WINDOWS -I/path1 -I/path2 -I/path3 -I/path4 -I/path5 \ -I/path6 -I/path7 -I/path8 -I/path9 file.c the command executes 4 times slower in beta18 than beta14. The above paths reference mounts to 5 different samba server linked network drives. I have also noticed the same performance degradation in another version of cpp compiled for beta14 and beta18. What has changed between beta14 and beta18 which would cause file/directory access to be slower? I was all set to move to beta18 until I noticed the performance degradation. Everything is better in beta18 except for the above problem - I cannot afford a 60 second gcc -E versus 15 seconds. -- -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= Philip Chow ISG Technologies Inc. Phone: (905) 672-2101 ext 304 Fax: (905) 672-2307 Internet: pchow@isgtec.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From edwards@panasync.canuck.ca Sat May 10 13:29:00 1997 From: edwards@panasync.canuck.ca (Colten Edwards) Date: Sat, 10 May 1997 13:29:00 -0000 Subject: termcap situations... References: <199705092120.QAA10919@smtp.gte.net> Message-ID: On Fri, 9 May 1997, Justin B. Harvey wrote: > How does compiling programs from bash and running them work with termcap? > I've installed ncurses and the program that needs curses still doesn't > work....am I missing a linked library somewhere? > libtermcap from linux will compile rather easily after fixing the termcap.h file slightly. It also works. ncurses on the other hand I couldn't get to work well. Colten Edwards panasync - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From morse@harborcom.net Sat May 10 13:29:00 1997 From: morse@harborcom.net (morse@harborcom.net) Date: Sat, 10 May 1997 13:29:00 -0000 Subject: Picking up include directories automatically Message-ID: <199705101708.NAA26779@ns2.harborcom.net> > Date: Fri, 9 May 1997 14:20:15 +0000 > From: Jon Thackray > To: gnu-win32@cygnus.com > Subject: Picking up include directories automatically > Is there some way to persuade cpp to look in some include directories > automatically? The old one used to look at the value of the include > environment varaible, the latest (b17.1) appears to ignore this. > Putting these directories on the command line is not an option, as > they vary from machine to machine. I've handled this by defining a local makefile 'local.mak' that has default settings for the lib and include dirs among others. App specific make files then include this file (-include local.mak). - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From edwards@panasync.canuck.ca Sat May 10 14:21:00 1997 From: edwards@panasync.canuck.ca (Colten Edwards) Date: Sat, 10 May 1997 14:21:00 -0000 Subject: select() Message-ID: kudos to the new b18 cygnus tools... with it I was able to port my irc client BitchX quiet easily to win95/NT. but the new select has an odd behaviour. when you remove the focus from the window running my client, all input/output in that window stops. Is this a design feature or something else? Also the new console handler is quite good, but still needs some work... I found that I had to hard code the term handler to use ALDL handling instead of CS handling in order to scroll the window properly. tputs_x(tgoto(CS, line2, line1)); /* shouldn't do this each time*/ if (n < 0) { term_move_cursor(0, line1); n = -n; } else term_move_cursor(0, line2); for (i = 0; i < n; i++) tputs_x(thing); tputs_x(tgoto(CS, LI - 1, 0)); /* shouldn't do this each time */ return (0); which is preferable to the following code which I have to use if (n > 0) { term_move_cursor(0, line1); tputs_x(tgoto(DL, n, n)); term_move_cursor(0, line2 - n + 1); tputs_x(tgoto(AL, n, n)); } else if (n < 0) { n = -n; term_move_cursor(0, line2-n+1); tputs_x(tgoto(DL, n, n)); term_move_cursor(0, line1); tputs_x(tgoto(AL, n, n)); } return (0); Colten Edwards - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jbharvey@gte.net Sat May 10 14:21:00 1997 From: jbharvey@gte.net (Justin Harvey) Date: Sat, 10 May 1997 14:21:00 -0000 Subject: termcap situations... Message-ID: <199705101824.NAA14933@smtp.gte.net> Here is some more information, if I can hardcode it into the code it would work, here is what it has now, can anyone suggest how to tweak this? TERMCODE (clear_screen, "\033[H\033[J") TERMCODE (clear_to_eos, "\033[J") TERMCODE (clear_to_eol, "\033[K") TERMCODE (cursor_address, "\033[%d;%dH") /* (in printf format) */ TERMCODE (enter_ca_mode, NULL) /* enable cursor motion mode */ TERMCODE (exit_ca_mode, NULL) /* disable cursor motion mode */ TERMCODE (change_scroll_region, "\033[%d;%dr") /* (in printf format) */ If I compile the program with HARDCODE defined (uses above sequences) the "visual mode" which splits the screen between input and output, when the output comes down the screen is bleeds through the line into the intput area, it should continue to scroll at the line which splits the screen. If I compile with TERMCAP defined and link in libncurses.a, it says it's unable to open vt100. frustated, j ---------- > From: Mikey > To: Justin B. Harvey > Cc: cygnus > Subject: Re: termcap situations... > Date: Saturday, May 10, 1997 1:16 AM > > When asking for help from the list, it is wise to specify a few details, > Platform, > environment, program output, the name of the program that you are working > on, wheather the ncurses test programs worked etc. > > from what you have said, it is impossible to tell whats going on, my > programs compiled against ncurses > from b17.1, all still work ie less elvis man etc. > > ---- > From: Justin B. Harvey > To: gnu-win32@cygnus.com > Date: Friday, May 09, 1997 9:43 PM > Subject: termcap situations... > > >How does compiling programs from bash and running them work with termcap? > >I've installed ncurses and the program that needs curses still doesn't > >work....am I missing a linked library somewhere? > > > >j > >-- > >Justin B. Harvey > >jbharvey@gte.net > > http://home1.gte.net/jbharvey > >- > >For help on using this list (especially unsubscribing), send a message to > >"gnu-win32-request@cygnus.com" with one line of text: "help". > > > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jbharvey@gte.net Sat May 10 14:45:00 1997 From: jbharvey@gte.net (Justin Harvey) Date: Sat, 10 May 1997 14:45:00 -0000 Subject: Solved my problem Message-ID: <199705102143.QAA21363@smtp.gte.net> Ok, I solved my problem with curses/termcap/terminal screen splitting. The solution was: to install GNU Termcap from prep.ai.mit.edu put their termcap in c:\etc link termcap in with my app export TERM=pcansi-mono and it worked fine. Thanks to all who helped. j -- Justin B. Harvey jbharvey@gte.net http://home1.gte.net/jbharvey - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From weiku@UTKUX1.UTK.EDU Sat May 10 18:13:00 1997 From: weiku@UTKUX1.UTK.EDU (Wei Ku) Date: Sat, 10 May 1997 18:13:00 -0000 Subject: bind in bash Message-ID: <199705110113.VAA42677@solid.ssd.ornl.gov> ---- From: Chuck Bogorad Aris.Zakinthinos@cl.cam.ac.uk wrote: here's how it's done: # home key bind '"\e[1~":beginning-of-line' # del key bind '"\e[3~":delete-char' # end key bind '"\e[4~":end-of-line' # pgup key bind '"\e[5~":history-search-forward' # pgdn key bind '"\e[6~":history-search-backward' Hello: I tried to bind key as described above. The home , end, and delete keys work fine. However, pageup and pagedown keys do not work. Any reply is appreciated. Sincerely, Wei Ku *************************************** Department of Physics and Astronomy The University of Tennessee 1408 Circle Drive Knoxville, Tennessee 37996-1200 weiku@utkux.utcc.utk.edu --------------------------------------- Solid State Division Oak Ridge National Laboratory P.O.Box 2008 Oak Ridge, TN 37831-6032 Phone: (423) 574-5795 Fax: (423) 574-4143 weiku@solid.ssd.ornl.gov *************************************** From mikaels@pdc.kth.se Sat May 10 18:13:00 1997 From: mikaels@pdc.kth.se (Michael Strömberg) Date: Sat, 10 May 1997 18:13:00 -0000 Subject: International keyboard support in bash? Changed in 18? Message-ID: <3374E447.61006F83@pdc.kth.se> A few days ago I installed the release 18, and noticed that bash was behaving in a peculiar way. It seems that I cannot get the '$' to come out on my swedish keyboard (which happens to be accessed by pressing the "AltGr" key and the '4' key). After several minutes of curious experimentation, I found that doing a shift '4' would give me a dead key, and if I then did an "AltGr" 4 - I would finally get the '$' symbol. Incidentally shift 4 gives you a circle with 4 lines sticking out in each "corner". To make matters worse, I decided to copy the bash from relase 17.1 and still had the same problem. Have the keyboard handling routines changed in release 18? Testing even further, the problem disappears in other applications such as vim. It seems only to affect the command line (bash). Everything worked great with the keyboard when I was using release 17.1. Anyone else have similar problems? // Michael - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From steveg1@primenet.com Sat May 10 18:13:00 1997 From: steveg1@primenet.com (Steve Gillespie) Date: Sat, 10 May 1997 18:13:00 -0000 Subject: Pdksh compile using gnu-win32 Message-ID: Will pdksh work under windows nt? Is signal support in b18 adequate? I compiled pdksh-5.2.12 using b18 successfully except I get an error message complaining of an invalid switch in sort in siglist.sh (if I remove the +2n +0n in sort everything compiles and I get ksh.exe except when I run it I have problems). Is there a problem with sort in gnu-win32? Here's the offending code from siglist.sh-- sed -n 's/{ QwErTy/{/p' < $out | awk '{print NR, $0}' | sort +2n +0n | sed 's/^[0-9]* //' | awk 'BEGIN { last=0; nsigs=0; } { if ($2 ~ /^[0-9][0-9]*$/ && $3 == ",") { n = $2; if (n > 0 && n != last) { while (++last < n) { printf "\t{ %d , (char *) 0, `Signal %d` } ,\n", last, last; } print; } } }' | tr '`' '"' | grep -v '"DUMMY"' I'd like to use ksh if possible so please give me any pointers. Thanks Steve G - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Rafael.Corvalan@mail.dotcom.fr Sat May 10 18:13:00 1997 From: Rafael.Corvalan@mail.dotcom.fr (Lorena & Rafael) Date: Sat, 10 May 1997 18:13:00 -0000 Subject: Is there another terminal? Message-ID: <01BC5D9C.C2B8E5F0@ppp-197-82.neuilly.club-internet.fr> Is there another terminal to use for console programs? What I don't like on Windows Console is that cut and paste is not really simple. Rafael - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Sun May 11 00:05:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Sun, 11 May 1997 00:05:00 -0000 Subject: Solved my problem Message-ID: <01BC5DFB.86A27720@sos> Justin Harvey wrote: > Ok, I solved my problem with curses/termcap/terminal screen splitting. The > solution was: > > to install GNU Termcap from prep.ai.mit.edu > put their termcap in c:\etc > link termcap in with my app > export TERM=pcansi-mono > Export TERM=linux for proper key bindings. -- Sergey Okhapkin Moscow Russia Looking for a job - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Sun May 11 00:05:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Sun, 11 May 1997 00:05:00 -0000 Subject: bind in bash Message-ID: <01BC5DFB.88F344F0@sos> Wei Ku wrote: > I tried to bind key as described above. The home , end, and delete keys work fine. However, pageup and pagedown keys do not work. > Any reply is appreciated. > export TERM=linux will help you. -- Sergey Okhapkin Moscow Russia Looking for a job - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Sun May 11 00:27:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Sun, 11 May 1997 00:27:00 -0000 Subject: National keyboards and B18 Message-ID: <01BC5DFE.7E6AB510@sos> Hi! Keyboard handler in B18 have linux console behavior - Alt-(alnum symbol) produces ESCsymbol sequence. It makes some problems with some national keyboard layouts. It's easy to change keyboard handler to check only for a left alt key to produce such sequences (this layouts uses right alt?) Any suggestions? -- Sergey Okhapkin Moscow Russia Looking for a job - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Sun May 11 00:49:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Sun, 11 May 1997 00:49:00 -0000 Subject: termcap situations... Message-ID: <01BC5DFB.7FD45E90@sos> Justin Harvey wrote: > I'm trying to port Tinyfugue to Windows95/NT (a mud client), I can > configure it for hardcoded terminal but thought that termcap might be > easier. If I compile it for termcap and try to go into it's visual mode > (split screen) it says it's not supported. So I grabbed ncurses and > compiled it, the test programs didn't work and they said they couldn't open > my vt100 screen...I'm wondering if there is any special libraries I need to > include while compiling. (i also tried linking ncurses with tinyfugue with > termcap and it didn't work). Did you set TERM=linux environment variable? -- Sergey Okhapkin Moscow Russia Looking for a job - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From noer@cygnus.com Sun May 11 00:49:00 1997 From: noer@cygnus.com (Geoffrey Noer) Date: Sun, 11 May 1997 00:49:00 -0000 Subject: gcc -E slower in beta 18? References: Message-ID: <199705110749.AAA03382@rtl.cygnus.com> Philip Chow wrote: > > I just recently had some time to move all my tools from beta 14 to beta18 > and I am noticing a dramatic performance degradation for C-preprocessing > of files with gcc -E. [...] > What has changed between beta14 and beta18 which would cause file/directory > access to be slower? All of the path handling was rewritten after the beta 14 release. A lot of other major changes have happened as well. Speeding up stat calls shouldn't be too hard and is high up on my list of things to try to speed up. Maybe for beta 19? :-) -- Geoffrey Noer noer@cygnus.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Sun May 11 00:49:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Sun, 11 May 1997 00:49:00 -0000 Subject: gcc -E slower in beta 18? Message-ID: <01BC5DFB.835E96C0@sos> Philip Chow wrote: > I just recently had some time to move all my tools from beta 14 to beta18 > and I am noticing a dramatic performance degradation for C-preprocessing > of files with gcc -E. > File open operations in B18 are much slower due to symlinks to directories support. Symlink_check/Path_conv code requires serious modifications :-( -- Sergey Okhapkin Moscow Russia Looking for a job - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Rafael.Corvalan@mail.dotcom.fr Sun May 11 06:38:00 1997 From: Rafael.Corvalan@mail.dotcom.fr (Lorena & Rafael) Date: Sun, 11 May 1997 06:38:00 -0000 Subject: md5 Checksums? Message-ID: <01BC5E21.2326FA60@ppp-197-168.neuilly.club-internet.fr> How can I get the checksum of a file to see if it matches the md5 chescksum appearing on md5sums file? Thanks Rafael - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jmr50@columbia.edu Sun May 11 07:44:00 1997 From: jmr50@columbia.edu (Jacob M. Rosenberg) Date: Sun, 11 May 1997 07:44:00 -0000 Subject: Errno MIA? Message-ID: <3.0.1.32.19970511104256.006da778@pop.columbia.edu> I've been trying to port the program TinyFugue with this package, and actually did it under b14, although with serious reductions in functionality. Anyhow, I just tried again under b18, and ran into a strange problem: the symbol errno (which I always assumed was in the standard c library) couldn't be found. If I simply declared a global integer, it compiled, but now when it crashes -- I get nothing. I also get nothing when exceptions creep up. All I get is: in cygwin32 exception handler and a hang, none of the nice debugging info I used to enjoy. :( :( Jacob Rosenberg .----------------------------------------------------------------------. | Jacob M. Rosenberg / Columbia College \ New York City, NY | | | | "Its not Camelot, but its not Cleveland, either." | | -- Kevin White, mayor of Boston | `----------------------------------------------------------------------` - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From gpasa@swissonline.ch Sun May 11 08:48:00 1997 From: gpasa@swissonline.ch (Pasa Guglielmo) Date: Sun, 11 May 1997 08:48:00 -0000 Subject: registry entries Message-ID: <3375FD11.258B@swissonline.ch> Hi, I noticed that I have two entries in win95 registry under HKEY_CURRENT_USER/Software/Cygnus Support/CYGWIN.DLL setup/ one named .../b14.0 and another .../b15.0 I currently have cygwin32 17.1. Is one of them (b14.0) from the old version, and must I erase it ? Can anybody tell me which is necessary and which not ? -- Sincerely, Pasa Guglielmo ------------------------------------------------------------------ & e-mail : gpasa@swissonline.ch & tel.: +41 (0)24 485 50 40 & & mailing : Pasa Guglielmo & fax : +41 (0)24 485 50 44 & & Rte des Cases 17A & & & CH-1890 St-Maurice & & & (Switzerland) & & ------------------------------------------------------------------ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jdeifik@weasel.com Sun May 11 10:03:00 1997 From: jdeifik@weasel.com (Jeff Deifik) Date: Sun, 11 May 1997 10:03:00 -0000 Subject: b18 problems Message-ID: <3.0.32.19970511100029.008fb900@pop.ni.net> I installed b-18 on my machine running win nt 4.0 server sp2. I have a file called a.c: main() { printf("Hello world\n"); } I set all the environment variables, have a /tmp dir and /bin/sh.exe I start bash, and type gcc a.c 2> gcc-error I get what seems like an infinite loop, with the last two lines of the included excerpt repeating. (unknown) heap_init: unable to allocate heap, win32 error 87 cygwin: terminating (unknown) In cygwin_except_handler (unknown) Exception trapped! (unknown) exception C0000005 at 1001E4B2 (unknown) exception: ax 0 bx 0 cx 201 dx FFFFFFFF (unknown) exception: si 79632D36 di 3833695C bp 241FB30 sp 241FB30 (unknown) exception is: STATUS_ACCESS_VIOLATION (unknown) Stack trace: (unknown) frame 0: sp = 0x241F964, pc = 0x1001282A (unknown) frame 1: sp = 0x241F980, pc = 0x77F94492 (unknown) frame 2: sp = 0x241F9A4, pc = 0x77F88E93 (unknown) frame 3: sp = 0x241FA30, pc = 0x77F76202 (unknown) frame 4: sp = 0x241FB30, pc = 0x10011D4D (unknown) frame 5: sp = 0x241FB44, pc = 0x10018972 (unknown) frame 6: sp = 0x241FB5C, pc = 0x10011615 (unknown) frame 7: sp = 0x241FF94, pc = 0x10011BFF (unknown) frame 8: sp = 0x241FFA0, pc = 0x4118D5 (unknown) frame 9: sp = 0x241FFB0, pc = 0x40103B (unknown) frame 10: sp = 0x241FFC0, pc = 0x77F1B26B (unknown) frame 11: sp = 0x241FFF0, pc = 0x0 (unknown) End of stack trace (unknown) In cygwin_except_handler (unknown) Error while dumping state (probably corrupted stack) (unknown) In cygwin_except_handler (unknown) Error while dumping state (probably corrupted stack) b17.1 used to work fine for me. Any ideas? Jeff Deifik jdeifik@weasel.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From weiku@UTKUX1.UTK.EDU Sun May 11 10:03:00 1997 From: weiku@UTKUX1.UTK.EDU (Wei Ku) Date: Sun, 11 May 1997 10:03:00 -0000 Subject: bind in bash Message-ID: <199705111657.MAA12570@solid.ssd.ornl.gov> ---- From: Sergey Okhapkin > I tried to bind key as described above. The home , end, and delete keys work fine. However, pageup > and pagedown keys do not work. export TERM=linux will help you. -- I tried to export the TERM as described : export TERM=linux The result is still the same: 'pagedown' key took me to the 0 th entry of history. Is it the way it is supposed to be ? By the way, I learned the following binding ( probably with a patch of bash). It seems not working any more. Any one has plan to make it work ? ( it make paste into terminal possible ): bind '"\e[2~":paste-from-clipboard' bind '"\C-v":paste-from-clipboard' ???? Sincerely, Wei Ku *************************************** Department of Physics and Astronomy The University of Tennessee 1408 Circle Drive Knoxville, Tennessee 37996-1200 weiku@utkux.utcc.utk.edu --------------------------------------- Solid State Division Oak Ridge National Laboratory P.O.Box 2008 Oak Ridge, TN 37831-6032 Phone: (423) 574-5795 Fax: (423) 574-4143 weiku@solid.ssd.ornl.gov *************************************** To: gnu-win32@cygnus.com; 'Wei Ku' Date: Sunday, May 11, 1997 3:05 AM Subject: RE: bind in bash Wei Ku wrote: Sergey Okhapkin Moscow Russia Looking for a job From jdeifik@weasel.com Sun May 11 12:02:00 1997 From: jdeifik@weasel.com (Jeff Deifik) Date: Sun, 11 May 1997 12:02:00 -0000 Subject: mount woes Message-ID: <3.0.32.19970511074920.00a0fcc8@pop.ni.net> I have a computer running windows nt 4.0 server, sp2. I installed beta-18. I wish to mount d: / since my NT partition is there, and I installed beta-18 there. I have read the readme.txt, which refers me to faq.txt. I have read the faq.txt, which says: >Where did my mount points go? >============================= > > Earlier releases stored mount points in the registry under "Cygnus >Support". This changed to "Cygnus Solutions" starting with beta 18. >Either use a registry editor (regedit under NT) to rename the old entry >or just redo your mount points and the cygwin.dll will automatically >create the new one for you. So I tried to redo my mount points. I typed 'mount --reset', which produced no errors. I typed 'mount d: /', and I got the message mount failed: Device or resource busy I thrashed around a bit, without being able to mount d: as / When I type mount, I get what I assume the defaults are, which are: Device Directory Type Flags \\.\tape1: /dev/st1 native no-mixed,text!=binary \\.\tape0: /dev/st0 native no-mixed,text!=binary \\.\b: /dev/fd1 native no-mixed,text!=binary \\.\a: /dev/fd0 native no-mixed,text!=binary c: / native no-mixed,text!=binary I would prefer not using the registry editor if possible to fix this. Any ideas? Jeff Deifik jdeifik@weasel.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From bernd@asterix.gecko.de Sun May 11 12:02:00 1997 From: bernd@asterix.gecko.de (Bernd Prager) Date: Sun, 11 May 1997 12:02:00 -0000 Subject: md5 Checksums? References: <01BC5E21.2326FA60@ppp-197-168.neuilly.club-internet.fr> Message-ID: > How can I get the checksum of a file to see if it matches the md5 chescksum appearing on md5sums file? I tried the md5sum.exe that comes with the pgp distribution. Unfortunately it takes different output than in the distribution file: cygnus: md5 sum for cdk.exe: 9fe57ae712c3a92b6a4881957edaffd9 cdk.exe md5sum.exe -b cdk.exe 10f99e5df1da55b5cb60b77eca61d241 *cdk.exe Seems not to be the right way ;-) ??? Bernd _____________________________________________________________________ Bernd Prager GECKO mbH; Wismarsche Str.3, 18057 Rostock; Germany http://www.gecko.de PGP Key fingerprint = 83 54 6A 3B 7A 9D 6C 0E F3 41 CE 99 11 30 B7 D6 public key by mailto:bpr@gecko.de?Subject=SendPGPKey _____________________________________________________________________ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From khan@xraylith.wisc.edu Sun May 11 14:41:00 1997 From: khan@xraylith.wisc.edu (Mumit Khan) Date: Sun, 11 May 1997 14:41:00 -0000 Subject: cygwin32-b18-g77-0.5.19.1: patch and binaries Message-ID: <9705112141.AA07801@modi.xraylith.wisc.edu> I've uploaded the source patch as well as cygwin32-b18-g77-0.5.19.1 (say that 5 times w/out breaking your teeth ;-) binaries to my web page at http://www.xraylith.wisc.edu/~khan/software/gnu-win32 As always, it's barely tested and totally unsupported. Also, please don't send email asking for 0.5.20 for b18 since I simply don't have the time for it. Enjoy, Mumit -- khan@xraylith.wisc.edu http://www.xraylith.wisc.edu/~khan/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From weiku@UTKUX1.UTK.EDU Sun May 11 15:22:00 1997 From: weiku@UTKUX1.UTK.EDU (Wei Ku) Date: Sun, 11 May 1997 15:22:00 -0000 Subject: Pdksh compile using gnu-win32 Message-ID: <199705112222.SAA47468@solid.ssd.ornl.gov> ---- From: Steve Gillespie I'd like to use ksh if possible so please give me any pointers. ---- It seems to me that you can run almost all the script for ksh by bash. However, all the script for csh can not be run by ksh. Unfortunately, I have some csh scripts in hands. As a result, It seems more important to have csh availible in Cygwin32. I tried to compile tcsh I got from ftp://ee.cornell.edu/pub Since I am not familiar with the UNIX shell code, I failed to make it compiled as expected. Any expert that could make csh availble will be appreciated very much. Sincerely, Wei Ku *************************************** Department of Physics and Astronomy The University of Tennessee 1408 Circle Drive Knoxville, Tennessee 37996-1200 weiku@utkux.utcc.utk.edu --------------------------------------- Solid State Division Oak Ridge National Laboratory P.O.Box 2008 Oak Ridge, TN 37831-6032 Phone: (423) 574-5795 Fax: (423) 574-4143 weiku@solid.ssd.ornl.gov *************************************** From maruzz@matrice.it Sun May 11 18:43:00 1997 From: maruzz@matrice.it (Giovanni Maruzzelli) Date: Sun, 11 May 1997 18:43:00 -0000 Subject: Service for NT and linking mixed cygnus-MS libs References: <199705112222.SAA47468@solid.ssd.ornl.gov> Message-ID: Hi there, I've browsed the archives to find hints, so I know that there's interest for this issue. I've succeded in realizing mixed microsoft-cygwin services for NT. If someone is interested, I can post some sample code. -ciao, -giovanni *********************************************** * * * * * MATRICE srl * * Information Technology Solutions * * * * Giovanni Maruzzelli * * * * * * * * maruzz@matrice.it * * http://www.matrice.it * *********************************************** - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From maruzz@matrice.it Sun May 11 18:43:00 1997 From: maruzz@matrice.it (Giovanni Maruzzelli) Date: Sun, 11 May 1997 18:43:00 -0000 Subject: Signals in B18 References: <199705112222.SAA47468@solid.ssd.ornl.gov> Message-ID: Hi there, I'm having lot of troubles with signals. I mean, I would like to make my apps react to signal like SIGTERM, SIGHUP, SIGKILL. I've had no success, so far. Particularly, I would like the parent signaled by the user, after that the parent signaling childs (the normal way to manage server that spawn itself). Any hint wildly appreciated. ciao, -giovanni *********************************************** * * * * * MATRICE srl * * Information Technology Solutions * * * * Giovanni Maruzzelli * * * * * * * * maruzz@matrice.it * * http://www.matrice.it * *********************************************** - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jdeifik@weasel.com Sun May 11 22:42:00 1997 From: jdeifik@weasel.com (Jeff Deifik) Date: Sun, 11 May 1997 22:42:00 -0000 Subject: b18 problems Message-ID: <3.0.32.19970511221553.00ffb9d8@pop.ni.net> At 10:00 AM 5/11/97 -0700, Jeff Deifik wrote: >I installed b-18 on my machine running win nt 4.0 server sp2. > >I have a file called a.c: >main() { printf("Hello world\n"); } > >I set all the environment variables, have a /tmp dir and /bin/sh.exe >I start bash, and type >gcc a.c 2> gcc-error > >I get what seems like an infinite loop, with the last two lines of the >included >excerpt repeating. > >(unknown) heap_init: unable to allocate heap, win32 error 87 >cygwin: terminating I found the problem. I had an old copy of cygwin.dll in d:\winnt\system32 I removed it, and it fixed the problem. Jeff Deifik - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jdeifik@weasel.com Sun May 11 22:47:00 1997 From: jdeifik@weasel.com (Jeff Deifik) Date: Sun, 11 May 1997 22:47:00 -0000 Subject: beta 18 'rint' function with gcc is broken Message-ID: <3.0.32.19970511224439.010096f0@pop.ni.net> I am running windows nt 4.0 server, sp2. I have a library regression test that I run on the compilers I use. It showed a bug with the 'rint' function. This function is supposed to convert a double to an integral double value. It isn't working, as the enclosed test program shows. Compile with no options, i.e. gcc yy.c yy The output should be nothing, but it is: 47.4 47.125000 47.6 47.625000 -47.4 -47.125000 -47.6 -47.625000 Jeff Deifik /* y.c, demonstrates a bug with gnu-w32 beta 18 */ /* Written 11/06/96 by Jeff Deifik */ /* Modified 05/11/97 by Jeff Deifik */ #include #include typedef enum { False = 0, True = 1 } Boolean; float FLOAT_TO_ROUNDED_FLOAT(float); int flo_close(double value,double f,double clo); int main(int argc, char *argv[]) { Boolean fail = False; double d; float f; d = rint(47); if (!flo_close(47.0f,d,0.0001f)) { (void)printf("47 %f\n",d); fail = True; } f = FLOAT_TO_ROUNDED_FLOAT(47.0); if (!flo_close(47.0f,f,0.0001f)) { (void)printf("47 %f\n",f); fail = True; } d = rint(47.4); if (!flo_close(47.0f,d,0.0001f)) { (void)printf("47.4 %f\n",d); fail = True; } d = rint(47.6); if (!flo_close(48.0f,d,0.0001f)) { (void)printf("47.6 %f\n",d); fail = True; } d = rint(-47); if (!flo_close(-47.0f,d,0.0001f)) { (void)printf("-47 %f\n",d); fail = True; } d = rint(-47.4); if (!flo_close(-47.0f,d,0.0001f)) { (void)printf("-47.4 %f\n",d); fail = True; } d = rint(-47.6); if (!flo_close(-48.0f,d,0.0001f)) { (void)printf("-47.6 %f\n",d); fail = True; } return 0; } /* Value is the ideal number, f is what it really is.*/ int flo_close(double value,double f,double clo) { return(((f+clo) >= value) && ((f-clo) <= value)); } float FLOAT_TO_ROUNDED_FLOAT(float x) { return(floor(x+0.5)); } From sos@prospect.com.ru Mon May 12 00:12:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Mon, 12 May 1997 00:12:00 -0000 Subject: registry entries Message-ID: <01BC5EC4.B2193A50@gater.krystalbank.msk.ru> Pasa Guglielmo wrote: > Hi, > I noticed that I have two entries in win95 registry under > > HKEY_CURRENT_USER/Software/Cygnus Support/CYGWIN.DLL setup/ > > one named .../b14.0 > and another .../b15.0 > I currently have cygwin32 17.1. > > Is one of them (b14.0) from the old version, and must I erase it ? Yes, b14 entry may be deleted. -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Mon May 12 00:12:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Mon, 12 May 1997 00:12:00 -0000 Subject: bind in bash Message-ID: <01BC5EC5.84495D20@gater.krystalbank.msk.ru> Wei Ku wrote: > I tried to export the TERM as described : > export TERM=linux > The result is still the same: > 'pagedown' key took me to the 0 th entry of history. Is it the way it is supposed to be ? Check your key bindings in bash startup files and .inputrc. PgDn key generates \E[6~ sequence (kN in termcap). -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Mon May 12 00:26:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Mon, 12 May 1997 00:26:00 -0000 Subject: Signals in B18 Message-ID: <01BC5EC7.80E27980@gater.krystalbank.msk.ru> Giovanni Maruzzelli wrote: > Hi there, > > I'm having lot of troubles with signals. I mean, I would like to make my > apps react to signal like SIGTERM, SIGHUP, SIGKILL. > > I've had no success, so far. > > Particularly, I would like the parent signaled by the user, after that the > parent signaling childs (the normal way to manage server that spawn > itself). > These signals are supported on NT only. -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Mon May 12 00:26:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Mon, 12 May 1997 00:26:00 -0000 Subject: mount woes Message-ID: <01BC5EC6.DC4842B0@gater.krystalbank.msk.ru> Jeff Deifik wrote: > I wish to mount d: / since my NT partition is there, and I > installed beta-18 there. > When I type mount, I get what I assume the defaults are, which are: > Device Directory Type Flags > \\.\tape1: /dev/st1 native no-mixed,text!=binary > \\.\tape0: /dev/st0 native no-mixed,text!=binary > \\.\b: /dev/fd1 native no-mixed,text!=binary > \\.\a: /dev/fd0 native no-mixed,text!=binary > c: / native no-mixed,text!=binary > > I would prefer not using the registry editor if possible to fix this. > No way to remount "/" with mount command:-( Use regedit. -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From ccurtin@trintech.com Mon May 12 04:43:00 1997 From: ccurtin@trintech.com (Colman Curtin) Date: Mon, 12 May 1997 04:43:00 -0000 Subject: X11R6.3 and Gnu-win32 B18 Message-ID: <199705121144.MAA03832@muzak.trintech.ie> >Alas, the Gnu-win32 X11R6.3 libraries on > ftp://niteroi.gsfc.nasa.gov/pub/win32/X11R6.3/ >now seem not to be compatible with Beta 18. Geoffrey's new install >Does anyone else see this? yes me too. I assume the dlls have to be rebuilt with b18. I had previously tried building the X11 libs on 17 but kept getting stuck. maybe I'll try again with 18, if I get time...... ____________________________________________________________ Colman Curtin mailto:ccurtin@trintech.ie Software Engineer ------------------------------------------------------------ Trintech (Manufacturing) Ltd, http://www.trintech.com/ South County Business Park, Leopardstown, Dublin 18. Tel +353-1-2956766 Fax +353-1-2954735 ------------------------------------------------------------ The fact that it works is immaterial. -- L. Ogborn - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From b.mcleod@opengroup.org Mon May 12 06:09:00 1997 From: b.mcleod@opengroup.org (Bruce D. McLeod) Date: Mon, 12 May 1997 06:09:00 -0000 Subject: cp -p Message-ID: <2.2.32.19970512123515.00962504@postman.osf.org> It used to be that cp -p just messed up the access time, but it got worse: C:\temp>cp -p bdmlog bdmlog2 cp: bdmlog2: Invalid argument Running b18, NT Bruce - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From pascal.trouvin@integralis.co.uk Mon May 12 06:53:00 1997 From: pascal.trouvin@integralis.co.uk (pascal.trouvin@integralis.co.uk) Date: Mon, 12 May 1997 06:53:00 -0000 Subject: X11R6.3 nolonger works under B18 Message-ID: <01IIS15GOOGK004RHJ@INTEGD.INTEGRALIS.CO.UK> The behaviour of 'ld' has changed, when you want to build an exe with 'ld' (not through gcc test.c -o test.exe) YOU MUST ADD /usr/lib/crt0.o /usr/lib/libkernel32.a for instance: gcc -c test.c ld test.o /usr/lib/crt0.o /usr/lib/libkernel32.a /usr/lib/libcygwin.a -o test.exe hope this help. pascal ______________________________ Reply Separator _________________________________ Subject: X11R6.3 nolonger works under B18 Author: gb@cs.unc.edu at INTERNET Date: 09/05/1997 17:56 I get a pop-up dialog saying Error Starting Program "Unable to run d:\USR\X11R6.3\BIN\xwd.exe" Bash says: bash.exe: /USR/X11R6.3/BIN/xwd.exe: (null) Same behavior on all x programs I try. They worked with B17. I have tried rebuilding them with B18 (with great pain) but to no good effect. I'm running on Win95. My PATH is /GNUWIN32/B18/H-I386-CYGWIN32/BIN:/GNUWIN32/B18/TCL/BIN:/USR/X11R6.3/BIN://C /CAFE/BIN://C/CAFE/JAVA/BIN://C/WINDOWS://C/WINDOWS/COMMAND:. thanks gb - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". ____________________________________________________________________ "This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. This communication may contain material protected by attorney-client privilege. If you are not the intended recipient or the person responsible for delivering the email to the intended recipient, be advised that you have received this email in error and that any use, dissemination, forwarding, printing, or copying of this email is strictly prohibited. If you have received this email in error please notify the IT manager by telephone on 44 (0) 118 9306060." info@Integralis.com http://www.integralis.com Innovation, Integration, Integralis _____________________________________________________________________ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From weiku@UTKUX1.UTK.EDU Mon May 12 06:53:00 1997 From: weiku@UTKUX1.UTK.EDU (Wei Ku) Date: Mon, 12 May 1997 06:53:00 -0000 Subject: Pdksh compile using gnu-win32 Message-ID: <199705120550.BAA16768@solid.ssd.ornl.gov> ---- From: Mumit Khan Try: http://www.softway.com/OpenNT/tw/ ???? (This may only work under OpenNT whatever that is) or: ftp://ftp.blarg.net/users/amol/tcsh/ ???? (The one with the mount problem I mentioned) ---- Thank you for the hint of the tcsh. However, I can not make the tsch running the csh scripts. Even though I change the name of tcsh.exe to tcsh.exe, the csh.exe always complains that the script file does not exist. I guess that we need a csh.exe that is compatible with the file name scheme of gnuwin32. ? Sincerely, Wei Ku *************************************** Department of Physics and Astronomy The University of Tennessee 1408 Circle Drive Knoxville, Tennessee 37996-1200 weiku@utkux.utcc.utk.edu --------------------------------------- Solid State Division Oak Ridge National Laboratory P.O.Box 2008 Oak Ridge, TN 37831-6032 Phone: (423) 574-5795 Fax: (423) 574-4143 weiku@solid.ssd.ornl.gov *************************************** To: Wei Ku Date: Sunday, May 11, 1997 11:00 PM Subject: Re: Pdksh compile using gnu-win32 "Wei Ku" < weiku@utkux1.utk.edu > writes: > Since I am not familiar with the UNIX shell code, I failed to make it = > compiled as expected. > Any expert that could make csh availble will be appreciated very much. > Sincerely, > Wei Ku > Here's a message from a previous query: -- using template mhl.format -- Date:???????????? Tue, 04 Mar 1997 10:41:36 PST To:???????????????????? Mumit Khan < khan@nuancecom.com > cc:???????????????????? amold@microsoft.com , gnu-win32@cygnus.com From:???????????? JP Shipherd < jp@nuancecom.com > Subject: Re: bash won't hand off to tcsh Return-Path: jp@nuancecom.com X-Sender: jp@awesome.nuance.com X-Mailer: Windows Eudora Pro Version 3.0 (32) Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" content-length: 810 At 10:32 AM 3/4/97 -0600, you wrote: >JP Shipherd < jp@nuancecom.com > writes: >> Hi, >> >>???????? I've got lots of csh shell scripts.???? To my joy today I found a copy of >> tcsh.exe which I put in /bin.???? (I even copied it to csh.exe at one point). >> > >Where did you find it!! I'm in the same situation you are (only about 75 >or so in my case, but enough that I don't want to rewrite it!), and would >very much like to get a copy. I've actually found two.???? I've only tried the second one so far, but it seems to have a hard time understanding file systems that were mounted in a bash shell. > >Thanks >Mumit > Good luck, --jp From banders@switch.rockwell.com Mon May 12 06:53:00 1997 From: banders@switch.rockwell.com (Bart Anderson) Date: Mon, 12 May 1997 06:53:00 -0000 Subject: b18 problems References: <3.0.32.19970511100029.008fb900@pop.ni.net> Message-ID: <33772130.64B7@switch.rockwell.com> Had the same problem myself. Thought I was stupid.. I might still be for that matter.. ;) - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jont@harlequin.co.uk Mon May 12 06:53:00 1997 From: jont@harlequin.co.uk (Jon Thackray) Date: Mon, 12 May 1997 06:53:00 -0000 Subject: Picking up include directories automatically References: <199705100251.TAA12832@nz1.netzone.com> Message-ID: <4543.9705121346@dedekind.cam.harlequin.co.uk> jeffdb@netzone.com writes: > *cpp: > -I /some/include/dir > > H-i386-cygwin32/lib/gcc-lib/i386-cygwin32/version-XXX/specs As I said, putting them on the command line is *NOT* an option. > From: Jon Thackray > To: gnu-win32@cygnus.com > Date: Friday, May 09, 1997 5:12 PM > Subject: Picking up include directories automatically > > >Is there some way to persuade cpp to look in some include directories > >automatically? The old one used to look at the value of the include > >environment varaible, the latest (b17.1) appears to ignore this. > >Putting these directories on the command line is not an option, as > >they vary from machine to machine. > >- > >For help on using this list (especially unsubscribing), send a message to > >"gnu-win32-request@cygnus.com" with one line of text: "help". > > > > - > For help on using this list (especially unsubscribing), send a message to > "gnu-win32-request@cygnus.com" with one line of text: "help". > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From weiku@UTKUX1.UTK.EDU Mon May 12 06:53:00 1997 From: weiku@UTKUX1.UTK.EDU (Wei Ku) Date: Mon, 12 May 1997 06:53:00 -0000 Subject: mount woes Message-ID: <199705120543.BAA41580@solid.ssd.ornl.gov> ---- From: Jeff Deifik So I tried to redo my mount points. I typed 'mount --reset', which produced no errors. I typed 'mount d: /', and I got the message mount failed: Device or resource busy I thrashed around a bit, without being able to mount d: as / I would prefer not using the registry editor if possible to fix this. ---- What you need is not mount -reset. You need umount: bash$ umount / bash$ mount c:\\ / This will do it. ( In my computer, I mounted the root as following: bash$ mount e:\\gnuwin32\\b18 / ) Sincerely, Wei Ku *************************************** Department of Physics and Astronomy The University of Tennessee 1408 Circle Drive Knoxville, Tennessee 37996-1200 weiku@utkux.utcc.utk.edu --------------------------------------- Solid State Division Oak Ridge National Laboratory P.O.Box 2008 Oak Ridge, TN 37831-6032 Phone: (423) 574-5795 Fax: (423) 574-4143 weiku@solid.ssd.ornl.gov *************************************** To: gnu-win32@cygnus.com Date: Sunday, May 11, 1997 7:40 PM Subject: mount woes I have a computer running windows nt 4.0 server, sp2. I installed beta-18. I wish to mount d: / since my NT partition is there, and I installed beta-18 there. I have read the readme.txt, which refers me to faq.txt. I have read the faq.txt, which says: >Where did my mount points go? >============================= > >???????? Earlier releases stored mount points in the registry under "Cygnus >Support".???? This changed to "Cygnus Solutions" starting with beta 18. >Either use a registry editor (regedit under NT) to rename the old entry >or just redo your mount points and the cygwin.dll will automatically >create the new one for you. When I type mount, I get what I assume the defaults are, which are: Device???????????????????????????????????????? Directory???????????????????????????????????????? Type???????????????????????????? Flags \\.\tape1:???????????????????????? /dev/st1???????????????????????????????????????????? native???????????????????? no-mixed,text!=binary \\.\tape0:???????????????????????? /dev/st0???????????????????????????????????????????? native???????????????????? no-mixed,text!=binary \\.\b:???????????????????????????????????????? /dev/fd1???????????????????????????????????????????? native???????????????????? no-mixed,text!=binary \\.\a:???????????????????????????????????????? /dev/fd0???????????????????????????????????????????? native???????????????????? no-mixed,text!=binary c:???????????????????????????????????????????????????????? /???????????????????????????????????????????????????????????????????????? native???????????????????? no-mixed,text!=binary Any ideas? Jeff Deifik jdeifik@weasel.com - For help on using this list (especially unsubscribing), send a message to " gnu-win32-request@cygnus.com " with one line of text: "help". From warisawa@pi.titech.ac.jp Mon May 12 06:53:00 1997 From: warisawa@pi.titech.ac.jp (Shin'ichi Warisawa) Date: Mon, 12 May 1997 06:53:00 -0000 Subject: problems with lgamma() Message-ID: <3376EC01.317B@pi.titech.ac.jp> I'm having a trouble in compiling and liking a program listed as follows. #include #include main() { double l; l = lgamma(1); printf("%lf\n", l); } Compiling environment is on Win NT 4.0 with cygwin32 beta 18. Compiling processes and messages are as follows: bash$ gcc -o lgamma lgamma.c /tmp\cc0015931.o(.text+0x1b):lgamma.c: undefined reference to `lgamma' gcc: Internal compiler error: program ld got fatal signal 1 bash$ gcc -o lgamma lgamma.c -lm C:\gnuwin32\b18\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32\cygnus-2.7.2-970404\../../../../i386-cygwin32\lib/libm.a(w_lgamma.o)(.text+0x7):w_lgamma.c: undefined reference to `signgam' gcc: Internal compiler error: program ld got fatal signal 1 Please give me solutions against the error. Thanks -- Precision and Intelligence Laboratory, Tokyo Institute of Technology Shin'ichi Warisawa (warisawa@pi.titech.ac.jp) - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From frei@fast.cs.utah.edu Mon May 12 06:53:00 1997 From: frei@fast.cs.utah.edu (Kevin Frei) Date: Mon, 12 May 1997 06:53:00 -0000 Subject: mount woes References: <3.0.32.19970511074920.00a0fcc8@pop.ni.net> Message-ID: <33771725.434FA734@fast.cs.utah.edu> I just went in an changed the registry entries - they're in H_KEY_USERS_LOCAL->Software->Cygnus Solutions (note the new name). That fixed it... -- Kevin Frei Center for Software Science - University of Utah "Any sufficiently advanced technology is indistinguishable from magic." - Arthur C. Clarke - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jmaltz@nema.com Mon May 12 07:37:00 1997 From: jmaltz@nema.com (Jon Maltz) Date: Mon, 12 May 1997 07:37:00 -0000 Subject: mount woes References: <01BC5EC6.DC4842B0@gater.krystalbank.msk.ru> Message-ID: <3.0.1.32.19970512103726.00964220@ne.highway1.com> At 11:23 AM 5/12/97 +0400, Sergey Okhapkin wrote something like this: >Jeff Deifik wrote: >> I wish to mount d: / since my NT partition is there, and I >> installed beta-18 there. > >> When I type mount, I get what I assume the defaults are, which are: >> Device Directory Type Flags >> \\.\tape1: /dev/st1 native no-mixed,text!=binary >> \\.\tape0: /dev/st0 native no-mixed,text!=binary >> \\.\b: /dev/fd1 native no-mixed,text!=binary >> \\.\a: /dev/fd0 native no-mixed,text!=binary >> c: / native no-mixed,text!=binary >> >> I would prefer not using the registry editor if possible to fix this. >> > >No way to remount "/" with mount command:-( Use regedit. > >-- >Sergey Okhapkin >Moscow, Russia >Looking for a job. > Ahhh, but you can! Just go to a bash prompt and umount / -- then mount d: / Should work without any problems. If it does give you a bit of grief, try mounting d:\cygnus\whatever\bin /bin from a cmd.exe prompt, then go to a bash prompt and umount / Jon - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From maruzz@matrice.it Mon May 12 07:37:00 1997 From: maruzz@matrice.it (Giovanni Maruzzelli) Date: Mon, 12 May 1997 07:37:00 -0000 Subject: Correction: Service for NT and linking mixed cygnus-MS libs References: Message-ID: Hi there, OK, I've received lot of mail that show the interest. It was late when I posted, so I confused my mind and my e-mail: I've succeded in building services for NT using CYGWIN alone or using mixed objects built with vc++ and gcc, linked with LINK. The problem is that in the second case, when the services built with mixed objects try to access the cygwin functionality they crash. So, they are useless, 'cause they work only if you don't use cygwin specific stuff in the objects compiled with gcc, but only microsoft allowed function (e.g. Sleep not sleep). Probably this can be solved building libcygwin.a without stuff that overlap the microsoft libc.lib or libcmt.lib. But at least you can build services cygwin-only. So, look in: ftp://aaue.matrice.it/pub/cygwin32/services/ or http://www.matrice.it/risorse/cygwin32/services/ Anyway, if you want to have a CYGNUS ONLY SERVICE, without microsoft stuff, the key is to modify a little the service example from microsoft in a way that don't use the special multithreaded function (like special printf, strcmp, etc.). After that, link with -ladvapi. ciao, -giovanni *********************************************** * * * * * MATRICE srl * * Information Technology Solutions * * * * Giovanni Maruzzelli * * * * * * * * maruzz@matrice.it * * http://www.matrice.it * *********************************************** - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From maruzz@matrice.it Mon May 12 07:37:00 1997 From: maruzz@matrice.it (Giovanni Maruzzelli) Date: Mon, 12 May 1997 07:37:00 -0000 Subject: Signals in B18 References: <01BC5EC7.80E27980@gater.krystalbank.msk.ru> Message-ID: On Mon, 12 May 1997, Sergey Okhapkin wrote: > Giovanni Maruzzelli wrote: > > Hi there, > > > > I'm having lot of troubles with signals. I mean, I would like to make my > > apps react to signal like SIGTERM, SIGHUP, SIGKILL. > > > > I've had no success, so far. > > > > Particularly, I would like the parent signaled by the user, after that the > > parent signaling childs (the normal way to manage server that spawn > > itself). > > > > These signals are supported on NT only. > > -- > Sergey Okhapkin > Moscow, Russia > Looking for a job. > > Hi Sergey, I forgot to say that I use NT 4.0 no service pack ciao, -giovanni *********************************************** * * * * * MATRICE srl * * Information Technology Solutions * * * * Giovanni Maruzzelli * * * * * * * * maruzz@matrice.it * * http://www.matrice.it * *********************************************** - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From ARichard@stark.cc.oh.us Mon May 12 08:24:00 1997 From: ARichard@stark.cc.oh.us (Richardson, Anthony) Date: Mon, 12 May 1997 08:24:00 -0000 Subject: Any distributions in the works? Message-ID: <199705121524.LAA14224@dialup.oar.net> Is Gnu-Win32 mature enough for a Linux-type distribution to be feasable? I would certainly be interested in purchasing a complete distribution with the base tools, development system, vi, emacs, X, g77, curses, TeX, groff, octave, etc. , etc., etc., if the price were right. ($20 or less). The Linux redhat rpm tools would be a good choice for creating a distribution. Is anyone else interested? (I'm not considering creating one myself, but thought that this might prompt others into doing the job.) Tony Richardson - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From banders@switch.rockwell.com Mon May 12 08:40:00 1997 From: banders@switch.rockwell.com (Bart Anderson) Date: Mon, 12 May 1997 08:40:00 -0000 Subject: Here's my exception Message-ID: <33773A96.3D80@switch.rockwell.com> I thought (maybe) that because I installed to other than the default directory that there might be a problem. I reinstalled to the default and had the same problem. Here's my output attached: From ray@mgi.com Mon May 12 08:40:00 1997 From: ray@mgi.com (Dave Ray) Date: Mon, 12 May 1997 08:40:00 -0000 Subject: mount woes References: <3.0.32.19970511074920.00a0fcc8@pop.ni.net> Message-ID: <337735DF.25E0@mgi.com> Jeff Deifik wrote: > > I have a computer running windows nt 4.0 server, sp2. > I installed beta-18. > I wish to mount d: / since my NT partition is there, and I > installed beta-18 there. > I have read the readme.txt, which refers me to faq.txt. > I have read the faq.txt, which says: > > >Where did my mount points go? > >============================= > > > > Earlier releases stored mount points in the registry under "Cygnus > >Support". This changed to "Cygnus Solutions" starting with beta 18. > >Either use a registry editor (regedit under NT) to rename the old entry > >or just redo your mount points and the cygwin.dll will automatically > >create the new one for you. > > So I tried to redo my mount points. > I typed 'mount --reset', which produced no errors. > I typed 'mount d: /', and I got the message > mount failed: Device or resource busy > > I thrashed around a bit, without being able to mount d: as / > > When I type mount, I get what I assume the defaults are, which are: > Device Directory Type Flags > \\.\tape1: /dev/st1 native no-mixed,text!=binary > \\.\tape0: /dev/st0 native no-mixed,text!=binary > \\.\b: /dev/fd1 native no-mixed,text!=binary > \\.\a: /dev/fd0 native no-mixed,text!=binary > c: / native no-mixed,text!=binary > > I would prefer not using the registry editor if possible to fix this. > > Any ideas? > > Jeff Deifik jdeifik@weasel.com > > - > For help on using this list (especially unsubscribing), send a message to > "gnu-win32-request@cygnus.com" with one line of text: "help". You need to do a 'umount /' first to free the mount point, then the 'mount d: /' command should work. -Dave Ray ray@mgi.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From andrewi@harlequin.co.uk Mon May 12 09:13:00 1997 From: andrewi@harlequin.co.uk (Andrew Innes) Date: Mon, 12 May 1997 09:13:00 -0000 Subject: bash problem with hexl under ntemacs References: Message-ID: <199705121612.RAA20522@propos.long.harlequin.co.uk> On 10 May 1997 01:12:30 +0200, Guido Van Hoecke said: [ hexl mode not working using Cygnus bash ] >Is this a known problem! Is there a workaround or a known fix? [There is a work-around, see near the end.] Well, problems with hexlify-buffer and friends have been reported for a long time. Most of these reports were from people using Emacs on Windows 95 with command.com as their normal shell, and so I had assumed the difficulties stemmed entirely from known problems with this (which will be fixed by the changes I am making for the upcoming "mini" release of Emacs, BTW). However, I had thought people using Cygnus bash and other Unix-based Win32 shells should not be experiencing these problems (providing Emacs is correctly configured, which your setup is). So I tried this now myself with the beta 17.1 bash. Whether or not the problem occurs seems to depend on which drive cygwin.dll thinks is the "root" drive, and where Emacs is installed (and possibly also where bash is installed and maybe other factors I haven't been able to determine). In your case bash reports that it cannot find "c:/emacs/bin/hexl.exe" - I suspect it is using "c:/" as the location of "/". If you either remove the "c:" or change it to "//c", then bash finds it. I don't understand precisely why that is - from looking at the source code, I believe cygwin.dll should always accept an absolute native filename, so perhaps it is bash itself that is getting confused. For now, a work-around is to set the value of `hexlify-command' in Emacs to "//c/emacs/bin/hexl -hex" (and the corresponding change for `dehexlify-command'). Hopefully Cygnus can shed more light on the correct behaviour of bash. AndrewI - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From hoeckeg@sebb.bel.alcatel.be Mon May 12 13:51:00 1997 From: hoeckeg@sebb.bel.alcatel.be (Guido Van Hoecke) Date: Mon, 12 May 1997 13:51:00 -0000 Subject: bash problem with hexl under ntemacs References: <199705121612.RAA20522@propos.long.harlequin.co.uk> Message-ID: Andrew Innes writes: > However, I had thought people using Cygnus bash and other Unix-based > Win32 shells should not be experiencing these problems (providing Emacs > is correctly configured, which your setup is). So I tried this now > myself with the beta 17.1 bash. > > Whether or not the problem occurs seems to depend on which drive > cygwin.dll thinks is the "root" drive, and where Emacs is installed (and > possibly also where bash is installed and maybe other factors I haven't > been able to determine). > > In your case bash reports that it cannot find "c:/emacs/bin/hexl.exe" - > I suspect it is using "c:/" as the location of "/". If you either > remove the "c:" or change it to "//c", then bash finds it. I don't > understand precisely why that is - from looking at the source code, I > believe cygwin.dll should always accept an absolute native filename, so > perhaps it is bash itself that is getting confused. > > For now, a work-around is to set the value of `hexlify-command' in Emacs > to "//c/emacs/bin/hexl -hex" (and the corresponding change for > `dehexlify-command'). Thanks, this is indeed a simple but effective work-around. g-) - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From kdw3@po.cwru.edu Mon May 12 13:51:00 1997 From: kdw3@po.cwru.edu (Kevin White) Date: Mon, 12 May 1997 13:51:00 -0000 Subject: Man pages Message-ID: <33776400.607DAB28@po.cwru.edu> Is there a port of "man" for NT 4.0? I adore bash but I really would like to be able to use and even build man pages for my NT box. -- __________________________ Kevin White kdw3@po.cwru.edu CWRU Dept of Genetics __________________________ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From arn@haus.haushahn.com Mon May 12 13:51:00 1997 From: arn@haus.haushahn.com (Alex Nelson) Date: Mon, 12 May 1997 13:51:00 -0000 Subject: mount woes Message-ID: <199705122052.QAA25243@hsefw> Use the umount command and unmount "/". Then mount your "D" drive using the mount command. --------- umount / mount d:/ / - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From wstewart@aa.net Mon May 12 14:01:00 1997 From: wstewart@aa.net (Wayne Stewart) Date: Mon, 12 May 1997 14:01:00 -0000 Subject: md5sums for xc? Message-ID: <199705122101.OAA23592@big.aa.net> Here are the checksums I got for the split up cdk-src files. They are different from what's in the 'md5sums' file: e6a114af57fa73ffcde44e5c3e24467d *xca 089d389b493e0c28e5c50c14aad19ae6 *xcb 49f53b3dfedbc82d106fc3f71416d0c5 *xcc 30be604211a99fce17ea1199b6a7e073 *xcd 68ba39d691b89fcfdd2db10aac6e0ba6 *xce 7d91fbaf2dd6b1ec7f4eae5f7729876d *xcf f0cb950bb00d0a139b3ff0e16c86d788 *xcg c0467465afa1a5f00ffc74cdf76b95bc *xch 31fa3204d72a971cc589d2ef4c0b9fcb *xci df253c625b2fd7a2bd28879365dec12d *xcj - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From kunglao@prairienet.org Mon May 12 19:00:00 1997 From: kunglao@prairienet.org (kunglao@prairienet.org) Date: Mon, 12 May 1997 19:00:00 -0000 Subject: Any other shells? Message-ID: <199705130100.UAA01308@bluestem.prairienet.org> Is there any work being done to port some of the other common shells to gnu-win32? It'd be nice to have a bit of choice of shells to use, perhaps tcsh, zsh, or ksh. --- "Oh, what sad times are these when passing ruffians can say 'Ni!' at will to old ladies. There is a pestulance upon this land, nothing is sacred. Even those who arrange and design shrubberies are under considerable economic stress of this period of history." -- Roger the Shrubber - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jim@jimpick.com Mon May 12 19:00:00 1997 From: jim@jimpick.com (Jim Pick) Date: Mon, 12 May 1997 19:00:00 -0000 Subject: Any distributions in the works? References: <199705121524.LAA14224@dialup.oar.net> Message-ID: <199705130200.TAA29290@fleming.jimpick.com> > Is Gnu-Win32 mature enough for a Linux-type distribution to be feasable? > > I would certainly be interested in purchasing a complete distribution > with > the base tools, development system, vi, emacs, X, g77, curses, TeX, > groff, octave, etc. , etc., etc., if the price were right. ($20 or less). > > The Linux redhat rpm tools would be a good choice for creating a > distribution. > > Is anyone else interested? (I'm not considering creating one myself, > but thought that this might prompt others into doing the job.) > > Tony Richardson Klee Dienes and I are planning on attempting to port dpkg (from Debian GNU/Linux) over to Gnu-Win32 fairly soon. It's a little too early for me to say whether or not we're going to have enough success with it to start a full-blown distribution... but we'd sure like to. Cheers, - Jim -- -----BEGIN PGP MESSAGE----- Version: 2.6.3 iQCVAwUBM3fLN+Qz770qyIfJAQGWYgP8CofjuCMxO9itwiuLgWWBND+7Pq6GGuVR adJ9RHPJ6wanEP5s3YdwcBbEhtchlP7MuFv8buXaCECmwueYyd3ZRqoznRQidWCl m9HqqgORlYTydB21DmQBZyXKZ20y+XLkipKhrQbCIbkvxJg4fu8JmVZpmNu9fek5 4xhG0o7+dsU= =jLsL -----END PGP MESSAGE----- From noer@cygnus.com Mon May 12 19:00:00 1997 From: noer@cygnus.com (Geoffrey Noer) Date: Mon, 12 May 1997 19:00:00 -0000 Subject: mount woes References: <3.0.32.19970511074920.00a0fcc8@pop.ni.net> Message-ID: <199705122134.OAA19168@cirdan.cygnus.com> > So I tried to redo my mount points. > I typed 'mount --reset', which produced no errors. 'mount --reset' resets the mount points to their default values with c: mounted as /. > I typed 'mount d: /', and I got the message > mount failed: Device or resource busy C: is already mounted as /. You need to: umount / mount d: / And that should do the trick! -- Geoffrey Noer noer@cygnus.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From tg@ict-gmbh.de Mon May 12 20:25:00 1997 From: tg@ict-gmbh.de (Thomas Göttgens) Date: Mon, 12 May 1997 20:25:00 -0000 Subject: Linker problem Message-ID: bsd.o: In function `check_init': /develop/binmake/bsd.c:649: undefined reference to `cygwin32_inet_netof' What can cause this message? Do i have to include anything more than the standard libs? (this is the ONLY linker error i get, the rest of the code makes heavy use of arpa/inet.h code but does not complain.) I grep'd the lib*.a files for the symbol but it came up empty. (???) What am I missing here? Thanks, Thomas - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From mike@mbsun.mlb.org Mon May 12 20:25:00 1997 From: mike@mbsun.mlb.org (Mike Bernson) Date: Mon, 12 May 1997 20:25:00 -0000 Subject: Problem rebuild cygwin.dll on Beta 18 Message-ID: <3377D96F.59E2B600@mbsun.mlb.org> I have unpacked the Beta 18 source and then Tryed to rebuild the Whole system for cross compiling. When it got time to build the cgywin.dll I got the following: /src1/gnu/beta18/cdk/ld/ld.new --base-file=cygwin.base --dll -o new-cygwin.dll setjmp.o longjmp.o ../librx/rx.o console.o dcrt0.o delqueue.o dirsearch.o exceptions.o exec.o fcntl.o fhandler.o fork.o grp.o heap.o hinfo.o init.o ioctl.o key.o malloc.o misc.o mmap.o net.o ntea.o passwd.o path.o pinfo.o pipe.o registry.o resource.o security.o select.o shared.o signal.o smallprint.o spawn.o strace.o strerror.o strsep.o stubs.o syscalls.o sysconf.o syslog.o termios.o times.o uinfo.o uname.o wait.o libkernel32.a libuser32.a libgdi32.a libadvapi32.a libcomdlg32.a libwsock32.a libnetapi32.a glob/libglob.a ../newlib/libc.a ../newlib/libm.a ../newlib/libm.a -e _dll_entry@12 `/src1/gnu/beta18/cdk/gcc/xgcc -B/src1/gnu/beta18/cdk/gcc/ -idirafter /src1/gnu/beta18/cdk/i386-cygwin32/newlib/targ-include -idirafter /src1/gnu/beta18/cdk/newlib/libc/include -nostdinc -print-libgcc-file-name` dcrt0.o: In function `globify(int *, char ***)': /src1/gnu/beta18/cdk/i386-cygwin32/winsup/dcrt0.cc:197: undefined reference to `glob' make: *** [new-cygwin.dll] Error 1 -print-libgcc-file-name` dcrt0.o: In function `globify(int *, char ***)': /src1/gnu/beta18/cdk/i386-cygwin32/winsup/dcrt0.cc:197: undefined reference to `glob' make: *** [new-cygwin.dll] Error 1 The glob/libglob.a is built. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Tue May 13 00:49:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Tue, 13 May 1997 00:49:00 -0000 Subject: Signals in B18 Message-ID: <01BC5F93.FDBC0440@gater.krystalbank.msk.ru> Giovanni Maruzzelli wrote: > > > I'm having lot of troubles with signals. I mean, I would like to make my > > > apps react to signal like SIGTERM, SIGHUP, SIGKILL. > > These signals are supported on NT only. > > I forgot to say that I use NT 4.0 no service pack > What troubles with signals do you have? The simpliest way to check signal support is to run bash, enter trap "echo SIGINT" SIGINT and press Ctrl-C. Bash must complain about receiving SIGINT signal. This should work on both W95 and NT. Or run "kill -1 bash's PID" from within another window. The second case works on NT only. -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Tue May 13 00:53:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Tue, 13 May 1997 00:53:00 -0000 Subject: Here's my exception Message-ID: <01BC5F94.85D5A160@gater.krystalbank.msk.ru> Bart Anderson wrote: > I thought (maybe) that because I installed to other than the default > directory that there might be a problem. > I reinstalled to the default and had the same problem. > Here's my output attached: > (unknown) heap_init: unable to allocate heap, win32 error 8 > cygwin: terminating > (unknown) In cygwin_except_handler > (unknown) Exception trapped! > (unknown) exception C0000005 at 1001E4B2 > (unknown) exception: ax 0 bx 0 cx C60395E4 dx BFFBF9E0 > (unknown) exception: si 8162F6C0 di 8162D360 bp 254F9A8 sp 254F9A8 > Seems like you have old cygwin.dll anywhere in your PATH. In %systemroot%\system32? -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From bernd@asterix.gecko.de Tue May 13 00:57:00 1997 From: bernd@asterix.gecko.de (Bernd Prager) Date: Tue, 13 May 1997 00:57:00 -0000 Subject: Pls help: can't rebuild terminfo after installing ncurses References: <199703211009.AA43886@alijku04.edvz.uni-linz.ac.at> Message-ID: Hi, maybe this is not a cygwin problem, but during installing the ncurses package I get a lot of errors while rebuilding the terminfo database: ... "/tmp/3811", line 10145, terminal 'tek4125': resolution of use=vt100 failed ... "/tmp/3811", line 12655, terminal 'env230': resolution of use=vt100 failed "/tmp/3811", line 12664, terminal 'ep40': resolution of use=ep4080 failed ? tic could not build /usr/local/share/terminfo make: *** [install.data] Error 1 make: *** [install] Error 2 I remember that someone had successfull done this!? Thanx for help, Bernd _____________________________________________________________________ Bernd Prager GECKO mbH; Wismarsche Str.3, 18057 Rostock; Germany http://www.gecko.de PGP Key fingerprint = 83 54 6A 3B 7A 9D 6C 0E F3 41 CE 99 11 30 B7 D6 public key by mailto:bpr@gecko.de?Subject=SendPGPKey _____________________________________________________________________ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jman@lx.net Tue May 13 00:57:00 1997 From: jman@lx.net (jman) Date: Tue, 13 May 1997 00:57:00 -0000 Subject: Question Message-ID: <3377E64C.4D76@lx.net> Howdy, I've been on this list for a while an have never posted. I've had great fun learning with gnuwin32 since beta 14. I have wrote a few interesting programs with them. My question is this. If a .c file calls for a header of tcp.h or udp.h or the others that are found on *nix but not in the gnuwin32 distribution how do you work these files. I mean you can't just take the say linux tcp.h an use it? Sorry if this is a stupid question but hey I'm no programer or anything just someone learning. Thanks cygnus its great I've compiled a multitude of great Unix programs for both my NT an win95 boxes.... Jason L. Esman - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jeffdb@netzone.com Tue May 13 04:04:00 1997 From: jeffdb@netzone.com (Mikey) Date: Tue, 13 May 1997 04:04:00 -0000 Subject: Man pages Message-ID: <199705131058.DAA00281@nz1.netzone.com> Make sure groff, and less are installed, and do export CC="gcc -DNOGETOPT -DNONLS -DLC_MESSAGES=6";sh configure -d make make install in the man-1.4h root dir. ---- From: Kevin White To: gnu-win32@cygnus.com Date: Monday, May 12, 1997 11:09 PM Subject: Man pages >Is there a port of "man" for NT 4.0? I adore bash but I really would >like to be able to use and even build man pages for my NT box. >-- >__________________________ >Kevin White >kdw3@po.cwru.edu >CWRU Dept of Genetics >__________________________ >- >For help on using this list (especially unsubscribing), send a message to >"gnu-win32-request@cygnus.com" with one line of text: "help". > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From maruzz@matrice.it Tue May 13 04:48:00 1997 From: maruzz@matrice.it (Giovanni Maruzzelli) Date: Tue, 13 May 1997 04:48:00 -0000 Subject: Apache for NT, the REVENGE References: <199705122134.OAA19168@cirdan.cygnus.com> Message-ID: Hi there, read the manual, read the manual, read the manual, read the manual, read the manual, read the manual, read the manual. After a lot of tries, I've succeded in building Apache as a Service. And it work very well, when debugging it in console mode. Fast, reliable, also found how to substitute the signaling stuff that i'm not able to make works. Really a big satisfaction..... Just to discover that the services under NT have no right to fork children. So now there's an interesting case of software that is much more powerful in debugging mode (where as a console app can make how many children it want) than in production mode (where it is allowed one only process, the controlling one, that don't do nothing apart from starting and killing the others). Anyway, in the making of this, I've found that you can use thread (at least as a mean to receive events instead of using signals. Signals don't work for me). You can do that using win32 calls like CreateThread, OpenEvent, ExitProcess, etc.) Each process spawned can have a thread that wait for events .... So, any hints, apart from rewriting apache in multithreading mode? ciao, -giovanni *********************************************** * * * * * MATRICE srl * * Information Technology Solutions * * * * Giovanni Maruzzelli * * * * * * * * maruzz@matrice.it * * http://www.matrice.it * *********************************************** - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Tue May 13 04:48:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Tue, 13 May 1997 04:48:00 -0000 Subject: Man pages Message-ID: <01BC5FAA.42BC0A70@gater.krystalbank.msk.ru> Kevin White wrote: > Is there a port of "man" for NT 4.0? I adore bash but I really would > like to be able to use and even build man pages for my NT box. Man (I've tried man-1.4d), groff-1.10 and less-321 (requred by man) compiles with cygnus tools without problems. Just try :-) -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Tue May 13 04:48:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Tue, 13 May 1997 04:48:00 -0000 Subject: [b18] Can't interrupt "sleep 30" Message-ID: <01BC5FB5.34ED0880@gater.krystalbank.msk.ru> Francis Litterio wrote: > Using b18, type > > sleep 30 > > to bash, then type ^C to interrupt it. Nothing happens, and after 30 > seconds elapses, bash prompts again. Ok, it's my fault :-) While modifying signal support for B18 I forgot about sleep() and usleep() syscalls. Thank you for this bug report. The functions sleep() and usleep() in winsup/signal.cc must be the following one: extern "C" unsigned int sleep (unsigned int seconds) { syscall_printf ("sleep (%d);\n", seconds); WaitForSingleObject (u->signal_arrived, seconds * 1000); syscall_printf ("0 = sleep (%d)\n", seconds); return 0; } extern "C" unsigned int usleep (unsigned int useconds) { syscall_printf ("usleep (%d)\n", useconds); WaitForSingleObject (u->signal_arrived, (useconds + 500) / 1000); syscall_printf ("0 = usleep (%d)\n", useconds); return 0; } -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jeffdb@netzone.com Tue May 13 04:48:00 1997 From: jeffdb@netzone.com (Mikey) Date: Tue, 13 May 1997 04:48:00 -0000 Subject: Picking up include directories automatically Message-ID: <199705131054.DAA29962@nz1.netzone.com> I think you misunderstood the thrust of my message, if you are willing to alter your specs file you can have any include directories you want, just make sure that the specs file you want to use is in the first directory searched by gcc ie. export LIBRARY_PATH=/usr./lib will use the specs file from /usr/lib if there is one. just make sure you DON'T have one in H-i386-cygwin32/lib/gcc-lib/etc... cause that will get used first so again in your specs file right under the line that says *cpp: put -I/your/include/dir DON"T insert or delete any lines, just type the dirs you want on that line. if you want to exclude all of the system dirs add -nostdinc, then gcc cpp and cc1 will only search in the dirs you specify. see the gcc man page or info page preprocessor options for more info. >jeffdb@netzone.com writes: > > *cpp: > > -I /some/include/dir > > > > H-i386-cygwin32/lib/gcc-lib/i386-cygwin32/version-XXX/specs > >As I said, putting them on the command line is *NOT* an option. > > > From: Jon Thackray > > To: gnu-win32@cygnus.com > > Date: Friday, May 09, 1997 5:12 PM > > Subject: Picking up include directories automatically > > > > >Is there some way to persuade cpp to look in some include directories > > >automatically? The old one used to look at the value of the include > > >environment varaible, the latest (b17.1) appears to ignore this. > > >Putting these directories on the command line is not an option, as > > >they vary from machine to machine. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From maruzz@matrice.it Tue May 13 04:48:00 1997 From: maruzz@matrice.it (Giovanni Maruzzelli) Date: Tue, 13 May 1997 04:48:00 -0000 Subject: Signals in B18 References: <01BC5F93.FDBC0440@gater.krystalbank.msk.ru> Message-ID: On Tue, 13 May 1997, Sergey Okhapkin wrote: > > What troubles with signals do you have? The simpliest way to check signal > support is to run bash, enter > > trap "echo SIGINT" SIGINT > > and press Ctrl-C. Bash must complain about receiving SIGINT signal. This > should work on both W95 and NT. Or run "kill -1 bash's PID" from within > another window. The second case works on NT only. > > -- > Sergey Okhapkin > Moscow, Russia > Looking for a job. > > Hi, I was trying something more: a server fork some children, and I was needing a way to signaling all them (like killpg, or kill -getpid()). Also, how about the other signals? (HUP, TERM, etc?) But now I have less need of them. ciao, -giovanni *********************************************** * * * * * MATRICE srl * * Information Technology Solutions * * * * Giovanni Maruzzelli * * * * * * * * maruzz@matrice.it * * http://www.matrice.it * *********************************************** - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From khan@xraylith.wisc.edu Tue May 13 04:48:00 1997 From: khan@xraylith.wisc.edu (Mumit Khan) Date: Tue, 13 May 1997 04:48:00 -0000 Subject: Missing signgam definition in libm [Re: problems with lgamma()] References: <3376EC01.317B@pi.titech.ac.jp> Message-ID: <9705122346.AA10143@modi.xraylith.wisc.edu> "Shin'ichi Warisawa" writes: > I'm having a trouble in compiling and liking a program listed as > follows. > > #include > #include > > main() > { > double l; > l = lgamma(1); > printf("%lf\n", l); > } > > Compiling environment is on Win NT 4.0 with cygwin32 beta 18. > > Compiling processes and messages are as follows: > > bash$ gcc -o lgamma lgamma.c > /tmp\cc0015931.o(.text+0x1b):lgamma.c: undefined reference to `lgamma' > gcc: Internal compiler error: program ld got fatal signal 1 > > bash$ gcc -o lgamma lgamma.c -lm > C:\gnuwin32\b18\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32\cygnus-2.7.2-970404 > \../../../../i386-cygwin32\lib/libm.a(w_lgamma.o)(.text+0x7):w_lgamma.c: > undefined reference to `signgam' > gcc: Internal compiler error: program ld got fatal signal 1 > For some reason, the file s_signgam.c, where signgam is defined, wasn't included in the 'obj' variable in the Makefile (it is there in the src one) and hence left out of the library. You can fix it one of 2 ways: 1. Get the file cdk/newlib/libm/math/s_signgam.c, compile it and put in libm, or 2. temporarily, include the following in your program somewhere in a file scope: #ifdef __CYGWIN32__ int signgam = 0; #endif and rebuild. In any case, the compiler shouldn't be dying with internal compiler error, and that's a bug. The Linux cross-compiler, btw, doesn't die, and simply reports the linker error. Mumit Cc: "Shin'ichi Warisawa" , Gnu-Win32 Mailing List - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From gb@cs.unc.edu Tue May 13 04:48:00 1997 From: gb@cs.unc.edu (Gary Bishop) Date: Tue, 13 May 1997 04:48:00 -0000 Subject: X11R6.3 and Gnu-win32 B18 Message-ID: <3.0.32.19970512193819.00ba4a60@mail.cs.unc.edu> I have rebuilt both the dll's and the clients. They still don't work. gb At 12:43 PM 5/12/97 +0100, Colman Curtin wrote: > > >>Alas, the Gnu-win32 X11R6.3 libraries on >> ftp://niteroi.gsfc.nasa.gov/pub/win32/X11R6.3/ >>now seem not to be compatible with Beta 18. Geoffrey's new install > >>Does anyone else see this? > >yes me too. I assume the dlls have to be rebuilt with b18. >I had previously tried building the X11 libs on 17 but kept getting stuck. >maybe I'll try again with 18, if I get time...... > >____________________________________________________________ >Colman Curtin mailto:ccurtin@trintech.ie >Software Engineer >------------------------------------------------------------ >Trintech (Manufacturing) Ltd, http://www.trintech.com/ >South County Business Park, >Leopardstown, >Dublin 18. >Tel +353-1-2956766 Fax +353-1-2954735 >------------------------------------------------------------ > >The fact that it works is immaterial. >-- L. Ogborn > > > > >- >For help on using this list (especially unsubscribing), send a message to >"gnu-win32-request@cygnus.com" with one line of text: "help". > > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Tue May 13 04:48:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Tue, 13 May 1997 04:48:00 -0000 Subject: cp -p Message-ID: <01BC5F8C.F0997420@gater.krystalbank.msk.ru> Bruce D. McLeod wrote: > It used to be that cp -p just messed up the access time, but > it got worse: > > C:\temp>cp -p bdmlog bdmlog2 > cp: bdmlog2: Invalid argument > > Running b18, NT The same config. 1:/usr/src$ ls -l t.c -rw-r--r-- 1 Administ Domain U 865 Apr 24 12:06 t.c 2:/usr/src$ cp -p t.c ttt.c 3:/usr/src$ ls -l ttt.c -rw-r--r-- 1 sos Domain U 865 Apr 24 12:06 ttt.c 4:/usr/src$ I have no idea why this doesn't work for you... -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Tue May 13 04:53:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Tue, 13 May 1997 04:53:00 -0000 Subject: Signals in B18 Message-ID: <01BC5FB6.10AA15C0@gater.krystalbank.msk.ru> Giovanni Maruzzelli wrote: > I was trying something more: a server fork some children, and I was > needing a way to signaling all them (like killpg, or kill -getpid()). > Process groups are not currently implemented. I'm planning to add this support in a few days. > Also, how about the other signals? (HUP, TERM, etc?) All POSIX signals are supported. -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From david@coent.demon.co.uk Tue May 13 04:53:00 1997 From: david@coent.demon.co.uk (David Coe) Date: Tue, 13 May 1997 04:53:00 -0000 Subject: B18 & Configure Scripts Message-ID: <3377FE4A.2E70@coent.demon.co.uk> Folks, I'm noticing a couple of new but consistent error messages during the configure phase while passing a slew of GNU stuff through B18. cygwin: Cygwin32: Catastrophic fail - unable to create pinfo-mutex Not updating unwritable cache .././config.cache The first doesn't seem to affect the outcome of the build and leaves no trace on config.log. The second has the obvious affect in having to reevaluate the cache contents as one passes through each of the build directories - my guess is that config.cache is not being closed (or is not yet closed) by the preceding call. Any obvious reasons? B17.5 didn't do this! -- Dr David Coe \=\ 58 Fairlawn Drive, East Grinstead \=\ Tel +44 1342 326860 West Sussex, RH19 1NT, United Kingdom \=\ Fax +44 1342 316019 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Tue May 13 04:53:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Tue, 13 May 1997 04:53:00 -0000 Subject: bind in bash Message-ID: <01BC5FAA.35DD8400@gater.krystalbank.msk.ru> Aris Zakinthinos wrote: > I have been trying to add a binding so that "Home", "End" etc. do the > obvious things for command line editing. I have tried everything I can > possibly thing of but I can't get it to work. Is bind broken? > > I suspect that it has something to do with the escape character since > > bind "\e[[A":"Funtion Key 1" > > doesn't work. Am I missing something? > export TERM=linux cp .../H-i386-cygwin32/etc/termcap /etc -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jont@harlequin.co.uk Tue May 13 06:39:00 1997 From: jont@harlequin.co.uk (Jon Thackray) Date: Tue, 13 May 1997 06:39:00 -0000 Subject: Picking up include directories automatically References: <199705131054.DAA29962@nz1.netzone.com> Message-ID: <22755.9705131326@dedekind.cam.harlequin.co.uk> jeffdb@netzone.com writes: > I think you misunderstood the thrust of my message, if you > are willing to alter your specs file > you can have any include directories you want, just make sure > that the specs file you want to use is in the first directory searched > by gcc ie. export LIBRARY_PATH=/usr./lib will use the specs file from > /usr/lib > if there is one. > just make sure you DON'T have one in H-i386-cygwin32/lib/gcc-lib/etc... > cause that will get used first > so again in your specs file right under the line that says > *cpp: > put -I/your/include/dir > DON"T insert or delete any lines, just type the dirs you want on that > line. > if you want to exclude all of the system dirs add -nostdinc, then gcc cpp > and cc1 will > only search in the dirs you specify. > see the gcc man page or info page preprocessor options for more info. Hmm, the gcc man page doesn't seem to tell me anything about how to edit this file, and info doesn't have anything on gcc. Can somebody tell me exactly how to edit this file? The current contents (presumably the default) are:- .... *cpp: %(cpp_cpu) %[cpp_cpu] %{posix:-D_POSIX_SOURCE} *cc1 .... - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Tue May 13 06:39:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Tue, 13 May 1997 06:39:00 -0000 Subject: X11R6.3 and Gnu-win32 B18 Message-ID: <01BC5F8B.44307310@gater.krystalbank.msk.ru> Colman Curtin wrote: > > > >Alas, the Gnu-win32 X11R6.3 libraries on > > ftp://niteroi.gsfc.nasa.gov/pub/win32/X11R6.3/ > >now seem not to be compatible with Beta 18. Geoffrey's new install > > >Does anyone else see this? > > yes me too. I assume the dlls have to be rebuilt with b18. > I had previously tried building the X11 libs on 17 but kept getting stuck. > maybe I'll try again with 18, if I get time...... > I didn't try to rebuild X11R6.3 with B18 yet, but due to moving _reent_data to cygwin.dll, rule for junk.c in config/cf/cygwin32.rules must begin with junk.c: @@\ echo "#include " >junk.c @@\ echo "#include " >>junk.c @@\ echo "int _fmode;" >>junk.c @@\ echo "extern struct _reent *_impure_ptr, *__imp_reent_data;" >>junk.c @@\ echo "int WINAPI dll_main(HANDLE a, DWORD reason, void *q)" >>junk.c @@\ echo "{ switch (reason){" >>junk.c @@\ echo " case DLL_PROCESS_ATTACH: break;" >>junk.c @@\ echo " case DLL_PROCESS_DETACH: break;" >>junk.c @@\ echo " case DLL_THREAD_ATTACH: break;" >>junk.c @@\ echo " case DLL_THREAD_DETACH: break;" >>junk.c @@\ echo " };" >>junk.c @@\ echo " _impure_ptr=__imp_reent_data;" >>junk.c @@\ echo " return 1;" >>junk.c @@\ echo "}" >>junk.c @@\ echo "main(){}" >>junk.c @@\ Concat3(lib,libname,.dll): solist junk.o libname.def @@\ RemoveFile(LibraryTargetName(libname)) @@\ ...... -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From bernd@asterix.gecko.de Tue May 13 06:39:00 1997 From: bernd@asterix.gecko.de (Bernd Prager) Date: Tue, 13 May 1997 06:39:00 -0000 Subject: Here's my exception References: <33773A96.3D80@switch.rockwell.com> Message-ID: > I thought (maybe) that because I installed to other than the default > directory that there might be a problem. > I reinstalled to the default and had the same problem. > Here's my output attached: > > (unknown) heap_init: unable to allocate heap, win32 error 8 > cygwin: terminating I had the same problem with leaving an old cygwin.dll in Win95\System directory. (The old dll size is about 2M, the new one is less then 500K) Delete it! Bernd _____________________________________________________________________ Bernd Prager GECKO mbH; Wismarsche Str.3, 18057 Rostock; Germany http://www.gecko.de PGP Key fingerprint = 83 54 6A 3B 7A 9D 6C 0E F3 41 CE 99 11 30 B7 D6 public key by mailto:bpr@gecko.de?Subject=SendPGPKey _____________________________________________________________________ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jeffdb@netzone.com Tue May 13 06:39:00 1997 From: jeffdb@netzone.com (Mikey) Date: Tue, 13 May 1997 06:39:00 -0000 Subject: Problems w/B18 & suggestions for future Message-ID: <199705130914.CAA10700@nz1.netzone.com> Platform win95, ServicePack1 w/fixes gnu-win32 Beta18 Problems w/B18:: in .../H-i386-cygwin32/i386-cygwin32/include/header.gcc forgot Mapping Complex.h _complex.h .(or dosen't this work because Complex and complex are in the same dir? mention in readme or FAQ) Large file lists will still cause stack faults in kernel32, try cp include/* in the microsoft SDK dir. [J [0J broken (no effect) termcap/info dosen't seem to work very well in general, extra "\r" inserted at physical eol of wrapped sentences by both termcap fallback support, and terminfo support in ncurses when using linux, or ansi full screen terminals. ie. elvis less info (I am working on a fix for these, and will post it asap) TEMP FIX change TERM=linux to TERM=pcansi in your .bashrc (it's not as pretty or smooth, but at least it works mostly) Mount still won't allow you to remount / as binary. without using regedit. (how about mount -b as the default, least surprise everything works OOTB, and let the masochists unlike me deal with mount -t. ;^) updatedb needs --outputfile=.../var/locatedb --local-paths'/d /e /etc.. all mounted drives' locate needs LOCATE_PATH=.../var/locatedb (mention in readme) readme still talks about mount -m make install still installs into prefix/i386-pc-cygwin32 instead of prefix/i386-cygwin32 (except when configured from linux ;^) Sugestions:: How about using IPROP.DLL for extended attributes? 1. it's Platform independant (only one code base for 95/NT) 2. it's fast (the first registry call loads it, after that it's all memory) 3 it's freely redistributable (I don't know that this would work, but it looks good from the help file in the sp2sdk.exe file from Micros***) Not everyone, not even all your customers have a T1, how about a regular tarball of the distribution 10.8M instead of 19M (usr+cdk) for those of us with 28.8 access next time Please? And maybe a patch to upgrade? the diffs for user-src were only 215k. instead of 7.5M, would have saved a lot of bandwidth. Also splitting the source distribution into smaller logical parts would IMHO encourage more net participation, posibly base for the files in cdk/ + cdk/include/ +cdk/libiberty libs for newlib winsup libio librx libstdc++ etc... compiler for bfd binutils gas gcc ld opcodes etc.... cdk-tools for make byacc etc... tcltk-debug for gdb gprof tcl tk dejagnu etc...... 27M is a pretty large chunk of bandwidth/time to invest just to make one or two changes in something you would like to make work better, (except for masochists like me ;^) Nice Job on the Installer, very Professional, thanks VERY much for the speed increase, and the increased support for sockets. could have used more testing on changes. Questions comments to jeffdb@netzone.com Mikey Flames&>/dev/null so don't bother. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From iko@morla.crec.mipt.ru Tue May 13 06:39:00 1997 From: iko@morla.crec.mipt.ru (Igor V. Kovalenko) Date: Tue, 13 May 1997 06:39:00 -0000 Subject: gcc -pipe Message-ID: <199705130947.NAA00557@morla.crec.mipt.ru> Hello! Sorry for repost, it seems like I'm heving trouble with e-mail. The problem is when I'm trying to use 'gcc -pipe' on beta 18 to compile sources, gcc dies silently. Is this a side-effect of gcc's spawn()ing instead of fork()ing subprocesses? Or am I missing something important? - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From david@coent.demon.co.uk Tue May 13 07:47:00 1997 From: david@coent.demon.co.uk (David Coe) Date: Tue, 13 May 1997 07:47:00 -0000 Subject: Man pages References: <33776400.607DAB28@po.cwru.edu> Message-ID: <33783346.5628@coent.demon.co.uk> Kevin White wrote: > > Is there a port of "man" for NT 4.0? I adore bash but I really would > like to be able to use and even build man pages for my NT box. Man-1.4h (from the Linux distribution) configures and compiles under B17.5 without change except for deletion in man.c of setlocale(LC_MESSAGES, ""); You will need less and groff. Roll your own under gnu-win32 or download VCC versions from: http://www.itribe.net/virtunix/ If you want to automatically generate preformatted and compressed man pages you also need a good version of gzip: http://sunsite.doc.ac.uk/packages/zip/WIN32/ I'm afraid the gnu-win32 (B17.5) version of gunzip gives unpacking problems for me. Good luck, -- Dr David Coe \=\ 58 Fairlawn Drive, East Grinstead \=\ Tel +44 1342 326860 West Sussex, RH19 1NT, United Kingdom \=\ Fax +44 1342 316019 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jeffdb@netzone.com Tue May 13 07:47:00 1997 From: jeffdb@netzone.com (Mikey) Date: Tue, 13 May 1997 07:47:00 -0000 Subject: Picking up include directories automatically Message-ID: <199705131423.HAA12797@nz1.netzone.com> Here is an example, it's for mingw32, just make sure you don't insert or delete any lines. and you'll be fine, BTW there's gcc info in the gcc source dir, it just dosen't get installed by default. >jeffdb@netzone.com writes: > > I think you misunderstood the thrust of my message, if you > > are willing to alter your specs file > > you can have any include directories you want, just make sure > > that the specs file you want to use is in the first directory searched > > by gcc ie. export LIBRARY_PATH=/usr./lib will use the specs file from > > /usr/lib > > if there is one. > > just make sure you DON'T have one in H-i386-cygwin32/lib/gcc-lib/etc... > > cause that will get used first > > so again in your specs file right under the line that says > > *cpp: > > put -I/your/include/dir > > DON"T insert or delete any lines, just type the dirs you want on that > > line. > > if you want to exclude all of the system dirs add -nostdinc, then gcc cpp > > and cc1 will > > only search in the dirs you specify. > > see the gcc man page or info page preprocessor options for more info. > >Hmm, the gcc man page doesn't seem to tell me anything about how to >edit this file, and info doesn't have anything on gcc. Can somebody >tell me exactly how to edit this file? The current contents >(presumably the default) are:- >.... >*cpp: >%(cpp_cpu) %[cpp_cpu] %{posix:-D_POSIX_SOURCE} > >*cc1 >.... > -------------- next part -------------- *asm: *asm_final: *cpp: %{posix:-D_POSIX_SOURCE} -iprefix /usr/w/include/ -I/usr/w/include %{windows:-I/usr/w/include/nonansi -I/usr/w/include/win32} %{console:-I/usr/w/include/nonansi -I/usr/w/include/win32} *cc1: %{!mcpu*: %{m386:-mcpu=i386 -march=i386} %{mno-486:-mcpu=i386 -march=i386} %{m486:-mcpu=i486 -march=i486} %{mno-386:-mcpu=i486 -march=i486} %{mno-pentium:-mcpu=i486 -march=i486} %{mpentium:-mcpu=pentium} %{mno-pentiumpro:-mcpu=pentium} %{mpentiumpro:-mcpu=pentiumpro}} *cc1plus: *endfile: *link: %{windows:--subsystem windows} %{dll:--dll -e _DllMainCRTStartup@12} *lib: %{windows:-luser32 -lgdi32 -lcomdlg32} -lkernel32 *libgcc: -lmingw32 -lgcc -lmoldname -lcrtdll *startfile: %{dll:dllcrt0%O%s} %{!dll:crt0%O%s} *switches_need_spaces: *signed_char: %{funsigned-char:-D__CHAR_UNSIGNED__} *predefines: -Di386 -D_WIN32 -D__WIN32__ -DWINNT -D__MINGW32__ -D_X86_=1 -D__STDC__=1 -D__stdcall=__attribute__((__stdcall__)) -D__cdecl=__attribute__((__cdecl__)) -Asystem(winnt) -Acpu(i386) -Amachine(i386) *cross_compile: 1 *version: cygnus-2.7.2-961023 *multilib: . ; *multilib_defaults: *multilib_extra: *multilib_matches: From jont@harlequin.co.uk Tue May 13 08:13:00 1997 From: jont@harlequin.co.uk (Jon Thackray) Date: Tue, 13 May 1997 08:13:00 -0000 Subject: Picking up include directories automatically References: <199705131423.HAA12797@nz1.netzone.com> Message-ID: <25439.9705131512@dedekind.cam.harlequin.co.uk> jeffdb@netzone.com writes: I can't get this to work. My specs file has the following:- *cpp: %(cpp_cpu) %[cpp_cpu] %{posix:-D_POSIX_SOURCE} -I//d/msdev/include -I//d/msdev/mfc/include -iprefix//d/msdev -iwithprefix include -iwithprefix mfc/include plus the other guff that was there already. My source file contains #include MMSYSTEM.H is in d:\msdev\include LIBRARY_PATH is set to c:\usr\local\lib in the control panel, and also to /usr/local/lib in sh. What am I doing wrong? gcc still tells me no include path in which to find mmsystem.h - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From voyager@eskimo.com Tue May 13 09:15:00 1997 From: voyager@eskimo.com (Nick Rossi) Date: Tue, 13 May 1997 09:15:00 -0000 Subject: csh under win32 Message-ID: <33789180.1A4A@eskimo.com> Hi, Has someone out there successfully built the C shell (csh) under Gnu-win32? I've made attempts myself but get stuck at "glob.c". Thanks, -- | Nick Rossi ........... http://www.eskimo.com/~voyager/ | | The Exit Light Show .. http://www.eskimo.com/~voyager/exitlight/ | | Novaterm ............. http://www.eskimo.com/~voyager/novaterm/ | | Freedom writings ..... http://www.eskimo.com/~voyager/freedom/ | - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Lassi.Tuura@cern.ch Tue May 13 09:15:00 1997 From: Lassi.Tuura@cern.ch (Lassi A. Tuura) Date: Tue, 13 May 1997 09:15:00 -0000 Subject: bash bug: variables with lowercase names will not `export' Message-ID: <33789304.2934@cern.ch> I just installed the gnuwin32/b18 stuff on my NT 4.0 Pentium Pro machine, and in general everything seems to work fine. Nice work, indeed! Trying to port some of my stuff from unix-world broke in the following scenario: ---- foo.sh ---- #! /bin/sh foo=bar export foo echo foo=$foo ./bar.sh ---- bar.sh ---- #! /bin/sh echo foo=$foo ---- end ------- Executing `foo.sh' prints the lines: foo=bar foo= If I replace `foo' with `FOO' everything works as expected. Is this a bug or something normal? On unixen all shells I know do pass lowercase variable names to subshells -- bash included -- so I'd assume this is a bug. Any quick/easy way to fix it? Cheers, //lat -- Lassi.Tuura@cern.ch There's no sunrise without a night - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Tue May 13 12:54:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Tue, 13 May 1997 12:54:00 -0000 Subject: Pls help: can't rebuild terminfo after installing ncurses Message-ID: <01BC5FF9.3140B170@sos> Bernd Prager wrote: > maybe this is not a cygwin problem, but during installing > the ncurses package I get a lot of errors while rebuilding the > terminfo database: > .. > "/tmp/3811", line 10145, terminal 'tek4125': resolution of use=vt100 > failed > .. > "/tmp/3811", line 12655, terminal 'env230': resolution of use=vt100 > failed "/tmp/3811", line 12664, terminal 'ep40': resolution of > use=ep4080 failed > ? tic could not build /usr/local/share/terminfo > make: *** [install.data] Error 1 > make: *** [install] Error 2 > Looks like text/binary problem... I've compiled ncurses without _any_ troubles using only binary mode mounts. -- Sergey Okhapkin Moscow, Russia Looking for a job - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From hoeckeg@sebb.bel.alcatel.be Tue May 13 21:28:00 1997 From: hoeckeg@sebb.bel.alcatel.be (Guido Van Hoecke) Date: Tue, 13 May 1997 21:28:00 -0000 Subject: Browsing UN*X manual pages References: <3377ED99.20B36A1B@mathworks.com> <33787723.A5E@Maths.QMW.ac.uk> Message-ID: "Dr Francis J. Wright" writes: > A preliminary version of an Emacs Lisp library currently called > `woman.el' is available in the directory > > http://www.maths.qmw.ac.uk/~fjw/public_emacs/ > > If you make any serious use of this code then please let me know what > you think of it, especially which bits don't work! > > Have fun. Francis It seems to work quite well with several man pages that came with gnuwin32/b18. I've spliced in the C-h keymap as (global-set-key [(control h) ?u] 'woman) Up to now, I had no tool to read these man pages. This seems odd though, cygnus provides all these pages but no means to read them. I guess somebody will point me to the man page viewer I didn't know of. Anyhow, I'm glad to be able to read these pages now. I must confess that I do not know enough of the input format to be able to define whether we're dealing with "... idiosyncratic files that `break the rules' or use ?roff requests directly... " rather than a problem with woman.el g-; - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jpmorgan@unm.edu Tue May 13 21:28:00 1997 From: jpmorgan@unm.edu (James Paul Morgan) Date: Tue, 13 May 1997 21:28:00 -0000 Subject: link problems References: <199704301641.MAA30060@solid.ssd.ornl.gov> Message-ID: On Wed, 30 Apr 1997, Wei Ku wrote: > Is it true that srand and rand are not nearly as random ? I had suspicion > on the randomness of the result of them. However, it is not verified yet. > Actually, I am taking that statment from the documentation for either the g++ or xlC compiler that is installed on our school system. (I didn't look to see which compiler's lib that document was specificallly from.) There were specific numbers quoted with in the documentation. Basically, it boiled down to the fact that the random() has a much longer sequence of numbers before repeating, therefore having more variety. (Or something similar.) It was also mentioned in one of our C++ text books, though no specifics were given there. See ya around the Mulberry bush. --James :) - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From tomaz.mertelj@ijs.si Tue May 13 23:53:00 1997 From: tomaz.mertelj@ijs.si (Tomaz Mertelj) Date: Tue, 13 May 1997 23:53:00 -0000 Subject: X11R6.3 and Gnu-win32 B18 References: <3.0.32.19970512193819.00ba4a60@mail.cs.unc.edu> Message-ID: <3.0.1.32.19970514085123.006b3ca8@cathy.ijs.si> At 07:40 PM 5/12/97 -0400, you wrote: >I have rebuilt both the dll's and the clients. They still don't work. >gb > It works if you build static libraries and rebuild clients. It seems that there is something wrong when linking dll's with Gnu-win32 B18. Regards, Tomaz Mertelj ********************************************************************* Dr. Tomaz Mertelj E-mail: tomaz.mertelj@ijs.si Jozef Stefan Institute Tel.: +386 (61) 17 73 900 P.O. Box. 3000 Fax: +386 (61) 12 35 400 1001 Ljubljana Home page: http://optlab.ijs.si/tmertelj Slovenia ********************************************************************* - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sam@inf.enst.fr Wed May 14 01:08:00 1997 From: sam@inf.enst.fr (Samuel Tardieu) Date: Wed, 14 May 1997 01:08:00 -0000 Subject: Apache for NT, the REVENGE References: Message-ID: >>>>> "Giovanni" == Giovanni Maruzzelli writes: Giovanni> So, any hints, apart from rewriting apache in multithreading Giovanni> mode? I think that RST used to have a multi-threaded version of Apache, you should ask him on the apache mailing-list (see http://www.apache.org/ for the right address). Sam -- Samuel Tardieu -- sam@inf.enst.fr - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Ramzi.Guetari@w3.org Wed May 14 06:58:00 1997 From: Ramzi.Guetari@w3.org (Ramzi Guetari) Date: Wed, 14 May 1997 06:58:00 -0000 Subject: Problem with gcc Message-ID: <3379971B.29279E6D@w3.org> I'm trying to compile a sotware using a configure file. After the generation of the Makefile, I run the gnu make and I have the following error message: ..../libgcc.a In function 'mprotect': .... Undefined reference to 'VirtualProtect'. I'm linking with the libkernel32.a --lkernel32). Any idea ?? Thank you very much in advance. -- Ramzi GUETARI | Tel: (33) 4 76 61 52 38 World Wide Web Consortium / INRIA | Fax: (33) 4 76 61 52 52 655 Avenue de l'Europe | e-mail: Ramzi.Guetari@w3.org 38330 Montbonnot Saint Martin - France | http://www.w3.org - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From vassilii@optimedia.co.il Wed May 14 06:58:00 1997 From: vassilii@optimedia.co.il (Vassilii Khachaturov) Date: Wed, 14 May 1997 06:58:00 -0000 Subject: RCS 5.7 on gnuwin32 b. 18 Message-ID: <3379D29D@gandalf.optimedia.co.il> I have also posted my q. to rcs-bugs@cs.purdue.edu. -- vassilii@optimedia.co.il Vassilii Khachaturov Skribu al mi per Esperanto! - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From pascal.trouvin@integralis.co.uk Wed May 14 06:58:00 1997 From: pascal.trouvin@integralis.co.uk (pascal.trouvin@integralis.co.uk) Date: Wed, 14 May 1997 06:58:00 -0000 Subject: Does any of you build successfully a DLL with B18? Message-ID: <01IIUMR80B4I004ZOU@INTEGD.INTEGRALIS.CO.UK> Hi folks, I've some problems with DLL creation since I've downloaded B18. I want to be sure that I've not missed somethng in the installation process. So in case that some of you have success with B18 and DLL creation, could you please send me, a little sample script? Thanks al lot pascal here my files: ------------------------------------------------ fixup.s ------------------------------------------------ .section .idata$3 .long 0,0,0,0, 0,0,0,0 ------------------------------------------------ init.c ------------------------------------------------ #include extern int WINAPI dll_entry (HANDLE h, DWORD reason, void *ptr); int WINAPI dll_entry (HANDLE h, DWORD reason, void *ptr) { switch (reason) { case DLL_PROCESS_ATTACH: break; case DLL_PROCESS_DETACH: break; case DLL_THREAD_ATTACH: break; case DLL_THREAD_DETACH: break; } return 1; } ------------------------------------------------ main.c ------------------------------------------------ /* Main file to try linking with a DLL under gnuwin32 */ int main() { printf("doit(5) returns %d\n", doit(5)); return(0); } ------------------------------------------------ foo.c ------------------------------------------------ /* Test file to check out building DLLs with gnuwin32 This uses printf from the std lib */ #include int doit (i) int i; { printf("from foo.c\n"); return( i+10 ); } ------------------------------------------------ loadDll.c ------------------------------------------------ #include #include typedef int (*fooFunc)(int i); int WINAPI WinMain (HINSTANCE hInstance, HINSTANCE hPrevInstance,PSTR szCmdLine, int iCmdShow) { HANDLE dllh; fooFunc doit; char buf[126]; dllh=LoadLibrary("foo.dll"); if ((UINT)dllh <= HINSTANCE_ERROR){ sprintf(buf,"Error loading dll: %d", dllh); MessageBox(NULL, buf, "loadDll", MB_OK); return(1); } else { doit=(fooFunc)GetProcAddress(dllh, "doit"); if (doit==NULL) MessageBox(NULL, "doit == null", "loadDll", MB_OK); else printf("doit(10) returns: %d\n",doit(10)); } if ((UINT)dllh > HINSTANCE_ERROR) FreeLibrary(dllh); return(0); } ------------------------------------------------ Makefile ------------------------------------------------ OBJS=init.o foo.o fixup.o CYGLIB=-lcygwin LIBS=$(CYGLIB) ENTRY=_dll_entry@12 RELOCATABLE=yes DLL=foo LD_OPT=-o $(DLL).dll $(OBJS) $(LIBS) -e $(ENTRY) .PHONY : all test clean all: $(DLL).dll $(DLL).a main.exe test test: $(DLL).dll test4.exe loadDll.exe echo "test4: validate functions call" ./test4.exe echo "loadDll: try to load dll" ./loadDll.exe clean: rm -f *.exp* *.def *.o *.base *.a %.def: %.o echo EXPORTS > $@ nm $^ | egrep "^........ [T] _" | sed "s/[^_]*_//" >> $@ $(DLL).dll: $(OBJS) $(DLL).def ifdef RELOCATABLE ld --dll --base-file $(DLL).base $(LD_OPT) dlltool --as=as --dllname $(DLL).dll --def $(DLL).def --base-file\ $(DLL).base --output-exp $(DLL).exp ld --dll --base-file $(DLL).base $(DLL).exp $(LD_OPT) dlltool --as=as --dllname $(DLL).dll --def $(DLL).def --base-file\ $(DLL).base --output-exp $(DLL).exp ld $(DLL).exp --dll $(LD_OPT) else ld --dll --base-file $(DLL).base $(LD_OPT) dlltool --as=as --dllname $(DLL).dll --def $(DLL).def --base-file\ $(DLL).base --output-exp $(DLL).exp ld --dll $(DLL).exp $(LD_OPT) endif $(DLL).a: $(DLL).def $(DLL).dll dlltool --as=as --dllname $(DLL).dll --def $< --output-lib $@ main.exe: main.c $(DLL).a gcc $^ -o $@ test4.exe: main.c foo.o gcc $^ -o $@ loadDll.exe: loadDll.c gcc $^ -o $@ -lcygwin -luser32 -lkernel32 ____________________________________________________________________ "This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. This communication may contain material protected by attorney-client privilege. If you are not the intended recipient or the person responsible for delivering the email to the intended recipient, be advised that you have received this email in error and that any use, dissemination, forwarding, printing, or copying of this email is strictly prohibited. If you have received this email in error please notify the IT manager by telephone on 44 (0) 118 9306060." info@Integralis.com http://www.integralis.com Innovation, Integration, Integralis _____________________________________________________________________ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jman@lx.net Wed May 14 07:20:00 1997 From: jman@lx.net (jman) Date: Wed, 14 May 1997 07:20:00 -0000 Subject: Any other shells? References: <199705130100.UAA01308@bluestem.prairienet.org> Message-ID: <3379D122.4025@lx.net> kunglao@prairienet.org wrote: > > Is there any work being done to port some of the other common shells > to gnu-win32? It'd be nice to have a bit of choice of shells to use, > perhaps tcsh, zsh, or ksh. > > --- > "Oh, what sad times are these when passing ruffians can say 'Ni!' at will to old > ladies. There is a pestulance upon this land, nothing is sacred. Even those > who arrange and design shrubberies are under considerable economic stress of > this period of history." -- Roger the Shrubber > - > For help on using this list (especially unsubscribing), send a message to > "gnu-win32-request@cygnus.com" with one line of text: "help". Try ftp.blarg.net/pub/amol he's done ksh an tsch I belive.... - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From rosnbrg@watson.ibm.com Wed May 14 07:20:00 1997 From: rosnbrg@watson.ibm.com (Bryan Rosenburg) Date: Wed, 14 May 1997 07:20:00 -0000 Subject: "install -d" and "cp -p" fail [b18 on Win NT 4.0] Message-ID: <85256497.00462F5D.00@watngi01.watson.ibm.com> I'm new to this mailing list, so I apologize if this subject has already been covered. I just upgraded from b17.1 to b18 on a Windows NT 4.0 system. The commands "install -d foo" and "cp -p bar foo", which worked fine before the upgrade, started to fail in similar ways. In both cases the error message is "foo: Invalid argument". I think the problem is that I did not have /etc/passwd and /etc/group files on my system. When I created these files, "install -d" and "cp -p" started to work properly. The b18 README.txt file mentions the programs "mkpasswd" and "mkgroup" that are useful for creating /etc/passwd and /etc/group files on Windows NT systems. I think it should also say that these files are now required for the correct operation of certain commands. Bryan Rosenburg IBM T.J. Watson Research Center Yorktown Heights, NY 10598 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From j-cerney1@ti.com Wed May 14 07:20:00 1997 From: j-cerney1@ti.com (John Cerney) Date: Wed, 14 May 1997 07:20:00 -0000 Subject: DLL Example no longer works under b18 Message-ID: The simple DLL example on the cygnus web page: http://www.cygnus.com/misc/gnu-win32/building-reloc-dlls.txt no longer works with the beta 18 release. When running main.exe a window pops up that says: "Unable to run c:\temp\main.exe". This happens both for the main.exe built under b17.1 and the main.exe built using b18. Running main.exe under the debugger (gdb) reports "unable to create process" errors, and error number 11. Error number 11 in include/Windows32/Errors.h is defined as "Bad Format". I tried this on a Windows 95 machine. I have not tried it on a Win NT machine yet. -John - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From irio.lavagno@torino.alpcom.it Wed May 14 07:20:00 1997 From: irio.lavagno@torino.alpcom.it (Irio Lavagno) Date: Wed, 14 May 1997 07:20:00 -0000 Subject: non blocking sockets Message-ID: <33796FBB.7F141B06@torino.alpcom.it> Hello, It seems that sockets can't be made non-blocking in cygwin32 b17 under winNT 4.0, is it just me or is it true ? I have a program which makes an accept on a non blocking socket, on my linux machine the accept exits with errno set to EWOULDBLOCK, on my NT machine it just blocks until it receives a connection request. If needed I can post the code, the socket is made non-blocking with a fcntl(sockid,F_SETFL,O_NONBLOCK), so could it be that this fcntl operation is not supported on cygwin32 ? The fcntl doesn't fail anyway. I have not tried it under b18 yet. Thanks for the help Irio Lavagno - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jon@heber.co.uk Wed May 14 07:20:00 1997 From: jon@heber.co.uk (Jon McGugan) Date: Wed, 14 May 1997 07:20:00 -0000 Subject: cross compiling to m68k no floating point References: <3376EC01.317B@pi.titech.ac.jp> <9705122346.AA10143@modi.xraylith.wisc.edu> Message-ID: <06391549900115@DOMAIN_H> Am trying to cross compile the gnu gcc compiler b17 version to target m68k with no floating point. Gas and Ld configured. The make in gcc fails in the libgcc2 when it tries to extract _floatdidf with errors of form: Error: Invalid instruction for this architecture; needs 68040 or 6060 -- statement `fdmove.l .....` ignored Lots of this. It appears to be trying to use floating point opcodes to implement functions/macros. Help appreciated in what I've done wrong. configured with --target=m68k-motorola-coff --nfp Regards Jon - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From gb@cs.unc.edu Wed May 14 07:20:00 1997 From: gb@cs.unc.edu (Gary Bishop) Date: Wed, 14 May 1997 07:20:00 -0000 Subject: X11R6.3 nolonger works under B18 Message-ID: <3.0.1.32.19970514101824.0069d284@mail.cs.unc.edu> If I use statically linked versions of the X libraries made with the usual ar r libX11static.a *.o and then link X apps with those, they work! Of course apps are 1/2 meg or bigger but they do work. I think this indicates that the difficulties with X11R6.3 and B18 are dll differences. I tried Sergey's suggested changes to junk.c but it still produces the same error for the dll version. If some of you dll wizards have a clue what is wrong, I'll be happy to try an experiment. After the initial pain of setting this X11 source tree up, changes are pretty easy. gb - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From vassilii@optimedia.co.il Wed May 14 07:20:00 1997 From: vassilii@optimedia.co.il (Vassilii Khachaturov) Date: Wed, 14 May 1997 07:20:00 -0000 Subject: RCS 5.7 on gnuwin32 b. 18 Message-ID: <3379CCFB@gandalf.optimedia.co.il> Has someone built RCS 5.7 on gnuwin32 b. 18 or another recent build? I get the following diagnostics when I run make (after configure succeeds): ================================================================== cd man && make all make: Nothing to be done for `all'. cd src && make all rm -f a.* ALL_CFLAGS=' -Dhas_conf_h -I. -I. -g -O' \ CC='gcc' \ COMPAT2='0' \ DIFF3='/gnuwin32/b18/H-i386-cygwin32/bin/diff3' DIFF3_BIN='1' \ DIFF='/gnuwin32/b18/H-i386-cygwin32/bin/diff' DIFFFLAGS='-an' DIFF_L='1' \ DIFF_SUCCESS='0' \ DIFF_FAILURE='1' \ DIFF_TROUBLE='2' \ ED='ed' \ LDFLAGS='' LIBS='' \ RCSPREFIX='/usr/local/bin/' \ SENDMAIL='' \ /bin/sh -x ./conf.sh 3>&1 >a.h 2>conf.err ./conf.sh: testing permissions ... OK ./conf.sh: testing compiler for plausibility ... OK ./conf.sh: configuring exitmain ... OK ./conf.sh: configuring _POSIX_C_SOURCE ... OK ./conf.sh: configuring _POSIX_SOURCE ... OK ./conf.sh: configuring how to check for syntax errors ... gcc -Dhas_conf_h -I. -I. -g -O -S ./conf.sh: configuring #include ... OK ./conf.sh: configuring #include ... OK ./conf.sh: configuring #include ... OK ./conf.sh: configuring #include ... OK ./conf.sh: configuring #include ... OK ./conf.sh: configuring #include ... commenting it out ./conf.sh: configuring #include ... commenting it out ./conf.sh: configuring #include ... OK ./conf.sh: configuring #include ... commenting it out ./conf.sh: configuring #include ... OK ./conf.sh: configuring #include ... OK ./conf.sh: configuring #include ... OK ./conf.sh: configuring #include ... OK ./conf.sh: configuring #include ... OK ./conf.sh: configuring #include ... commenting it out ./conf.sh: configuring #include ... OK ./conf.sh: configuring #include ... OK ./conf.sh: configuring #include ... commenting it out ./conf.sh: configuring has_sys_param_h ... OK ./conf.sh: configuring errno ... OK ./conf.sh: configuring has_readlink, readlink_isreg_errno ... 1, EINVAL ./conf.sh: configuring mode_t ... OK ./conf.sh: configuring off_t ... OK ./conf.sh: configuring pid_t ... OK ./conf.sh: configuring sig_atomic_t ... OK ./conf.sh: configuring size_t ... OK ./conf.sh: configuring ssize_t ... OK ./conf.sh: configuring time_t ... OK ./conf.sh: configuring uid_t ... OK ./conf.sh: configuring const ... OK ./conf.sh: configuring volatile ... OK ./conf.sh: configuring has_prototypes, has_stdarg, has_varargs, va_start_args ... 1, 1, -1, 2 ./conf.sh: configuring bad_chmod_close ... OK ./conf.sh: configuring bad_creat0 ... OK ./conf.sh: configuring bad_fopen_wplus ... OK ./conf.sh: configuring has_attribute_noreturn ... OK ./conf.sh: configuring has_dirent, void_closedir ... OK ./conf.sh: configuring has_fchmod ... does not work ./conf.sh: configuring has_fflush_input ... OK ./conf.sh: configuring has_fputs ... OK ./conf.sh: configuring has_ftruncate ... does not work ./conf.sh: configuring has_getuid ... OK ./conf.sh: configuring has_getpwuid ... does not work ./conf.sh: configuring has_kill ... OK ./conf.sh: configuring has_memcmp ... OK ./conf.sh: configuring has_memcpy ... OK ./conf.sh: configuring has_memmove ... OK ./conf.sh: configuring has_map_fd, has_mmap, has_madvise, mmap_signal ... make: *** [conf.h] Error 1 make: *** [all] Error 2 ================================================================== Thanks a lot in advance, vassilii -- vassilii@optimedia.co.il Vassilii Khachaturov Skribu al mi per Esperanto! - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Wed May 14 07:20:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Wed, 14 May 1997 07:20:00 -0000 Subject: B18 & Configure Scripts Message-ID: <01BC6061.E0F666B0@gater.krystalbank.msk.ru> David Coe wrote: > cygwin: Cygwin32: Catastrophic fail - unable to create pinfo-mutex > > Not updating unwritable cache .././config.cache > > The first doesn't seem to affect the outcome of the build and leaves no > trace on config.log. The second has the obvious affect in having to > reevaluate the cache contents as one passes through each of the build > directories - my guess is that config.cache is not being closed (or is > not yet closed) by the preceding call. > You are using windows NT and you are member of "Administrators" group. It's NT's (mis)feature - in this case the owner of file is "Administrators", but not you! This fools bash to check file's write permissions properly (getuid() returns yours, but not "Administrators" uid). Chown config.cache "you login name". Or remove youself from "administrators" group. -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jeffdb@netzone.com Wed May 14 07:20:00 1997 From: jeffdb@netzone.com (Mikey) Date: Wed, 14 May 1997 07:20:00 -0000 Subject: Picking up include directories automatically Message-ID: <199705141149.EAA03572@nz1.netzone.com> see readme or maybe it's FAQ asbout using // with gcc cpp etc. mount -b d:\\ /d and eliminate // from the -I lines, and you should be fine. ---- From: Jon Thackray To: cygnus Date: Wednesday, May 14, 1997 1:42 AM Subject: Re: Picking up include directories automatically >jeffdb@netzone.com writes: >I can't get this to work. My specs file has the following:- >*cpp: >%(cpp_cpu) %[cpp_cpu] %{posix:-D_POSIX_SOURCE} -I//d/msdev/include -I//d/msdev/mfc/include -iprefix//d/msdev -iwithprefix include -iwithprefix mfc/include > >plus the other guff that was there already. > >My source file contains >#include > >MMSYSTEM.H is in d:\msdev\include >LIBRARY_PATH is set to c:\usr\local\lib in the control panel, and also >to /usr/local/lib in sh. > >What am I doing wrong? gcc still tells me no include path in which to >find mmsystem.h >- >For help on using this list (especially unsubscribing), send a message to >"gnu-win32-request@cygnus.com" with one line of text: "help". > > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From lce@dhi.dk Wed May 14 07:43:00 1997 From: lce@dhi.dk (Lars Ekebjerg) Date: Wed, 14 May 1997 07:43:00 -0000 Subject: Console problems with bash 18 Message-ID: <97May14.143727gmt.27785-1@gateway.dhi.dk> Hi I have a problem with Bash from new Beta 18 release. Being danish I naturally use a danish keyboard what bash doesn't seem to like. It won't let me use our national characters and even worse neither characters like $ and | (which without Unix is less usefull). '$', '|' and also other characters are typed together with thr tight 'Alt' key. When I try bash simply keeps on beeping. Beta 17.1 didn't have this problem. Are there anyone who can help me with my missing characters. Thanks Lars PS Anyone out there having a 'more' or 'less' port. -- \!!!/ o|o ---------------------------------------------------------------------- Lars Chresten Ekebjfrg Danish Hydraulic Institute E-mail: lce@dhi.dk Telephone: +45 45179246 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From vassilii@optimedia.co.il Wed May 14 07:48:00 1997 From: vassilii@optimedia.co.il (Vassilii Khachaturov) Date: Wed, 14 May 1997 07:48:00 -0000 Subject: termcap-1.3 on gnuwin32 b. 18 Message-ID: <3379DE70@gandalf.optimedia.co.il> I tried to build termcap-1.3 I did `configure' successfully and then `make' gave me the following: gcc -c -DHAVE_STRING_H=1 -DHAVE_UNISTD_H=1 -DSTDC_HEADERS=1 -DTERMCAP_FILE=\"/ etc/termcap\" -I. -I. -g termcap.c gcc -c -DHAVE_STRING_H=1 -DHAVE_UNISTD_H=1 -DSTDC_HEADERS=1 -DTERMCAP_FILE=\"/ etc/termcap\" -I. -I. -g tparam.c In file included from tparam.c:29: e:\gnuwin32\b18\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32\cygnus-2.7.2-97040 4\.. /../../../i386-cygwin32/include/string.h:55: syntax error before `(' e:\gnuwin32\b18\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32\cygnus-2.7.2-97040 4\.. /../../../i386-cygwin32/include/string.h:55: warning: conflicting types for buil t-in function `memcpy' e:\gnuwin32\b18\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32\cygnus-2.7.2-97040 4\.. /../../../i386-cygwin32/include/string.h:55: syntax error before `const' gcc: Internal compiler error: program cc1 got fatal signal 33 make: *** [tparam.o] Error 1 Does someone have an applicable patch? BTW, the mailing list archives at Cygnus were last updated last December (!). Does anyone know where I could find an updated archive in order not to bother the developers with obviously FAQs like the one I asked in this letter? -- Vassilii Khachaturov vassilii@optimedia.co.il Skribu al mi per Esperanto! - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From wjones@TC.FLUKE.COM Wed May 14 09:13:00 1997 From: wjones@TC.FLUKE.COM (Warren Jones) Date: Wed, 14 May 1997 09:13:00 -0000 Subject: cd problems with //x paths Message-ID: <97May14.091002pdt.35713-1@gateway.fluke.com> Has anyone else noticed that bash won't let you "cd .." when you're in a directory like "//x/subdir" ? //d/work [65] pwd //d/work //d/work [66] cd .. //d/work [67] pwd //d/work //d/work [68] cd //d //d/work [69] pwd //d/work //d/work [70] cd d:/ //d [71] pwd //d I'm using b18. Thanks to anyone who has insight into this situation. -------------------------------------------------------------------- Warren Jones | To keep every cog and wheel is the first Fluke Corporation | precaution of intelligent tinkering. Everett, Washington, USA | -- Aldo Leopold - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sganly@lehman.com Wed May 14 09:37:00 1997 From: sganly@lehman.com (Steve Ganly) Date: Wed, 14 May 1997 09:37:00 -0000 Subject: Problem extracting beta18 on NT4.0 with compressed drive Message-ID: <9705141736.ZM27217@lonadmin226.lonadmin.lehman.com> I'm getting the following error message when trying to install beta 18 on an NT 4.0 machine. I run cdk.exe, once the extraction reaches 98% I get the following message: InstallShield Self-extracting EXE The decompression of %s failed. There may not be enough free disk space available in the TEMP directory. The windows explorer shows 131Mb free on the drive. I've tried it with 192Mb free on the drive but get the same error. I'm a bit doubtful that a 13.7Mb file would need more than 100Mb to decompress which makes me wonder if it's the drive compression causing problems. I have beta 17 installed already. I've tried renaming the base directory but haven't tried removing it. Has anyone else installed beta 18 on a compressed drive? Does anyone know how much disk space is required? Is there a tar file available to use instead? Steve Ganly. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sup@cri.dk Wed May 14 09:37:00 1997 From: sup@cri.dk (Steen Ulrik Palm) Date: Wed, 14 May 1997 09:37:00 -0000 Subject: Bash Problems Message-ID: <33796F53.4F7C@cri.dk> I have downloaded and installed the 18b version of the gnu-win32 user tools under Windows 95. Bash has now been defined as the default shell in Emacs and bash runs smoothly within a *shell* buffer. However, I have problems when running bash from a DOS prompt window. Firstly, the `Alt Char' key does not work. This means that I'm not able to get access to the characters `$', `|', ``' and others which hampers the usefulness of bash considerably. I have tried to define some keybindings for these characters in ~/.inputrc, but for unknown reasons it does not work. Another problem is related to copying/pasting text to and from the DOS prompt window. When bash is started trying to paste something into the DOS prompt window (using the paste button on the tool bar) will make the window hang, and the only possible operation left is closing the window. I would be very grateful for help on these problems. Steen -- \\\|/// \\ - - // ( @ @ ) +-----------------------oOOo-(_)-oOOo--------------------------------+ Steen Palm, CRI A/S, Bregnerxdvej 144, DK-3460 Birkerxd, Denmark Tel: +45 45 94 96 65, Fax: +45 45 94 96 45, E-mail: sup@cri.dk +-------------------------------Oooo---------------------------------+ oooO ( ) ( ) ) / \ ( (_/ \_) - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jeffdb@netzone.com Wed May 14 09:37:00 1997 From: jeffdb@netzone.com (Mikey) Date: Wed, 14 May 1997 09:37:00 -0000 Subject: Problem rebuild cygwin.dll on Beta 18 Message-ID: <199705140727.AAA17377@nz1.netzone.com> Possibly an interupt durring build of glob/libglob.a. try deleting glob/libglob.a and running make in winsup/glob/ and then relinking in winsup/, the clean target for winsup is broken so libglob.a dosen't get deleted with make clean. ---- From: Mike Bernson To: gnu-win32@cygnus.com Date: Tuesday, May 13, 1997 11:11 PM Subject: Problem rebuild cygwin.dll on Beta 18 >I have unpacked the Beta 18 source and then Tryed to rebuild the >Whole system for cross compiling. > >When it got time to build the cgywin.dll I got the following: > >/src1/gnu/beta18/cdk/ld/ld.new --base-file=cygwin.base --dll -o >new-cygwin.dll setjmp.o longjmp.o ../librx/rx.o console.o dcrt0.o >delqueue.o dirsearch.o exceptions.o exec.o fcntl.o fhandler.o fork.o >grp.o heap.o hinfo.o init.o ioctl.o key.o malloc.o misc.o mmap.o net.o >ntea.o passwd.o path.o pinfo.o pipe.o registry.o resource.o security.o >select.o shared.o signal.o smallprint.o spawn.o strace.o strerror.o >strsep.o stubs.o syscalls.o sysconf.o syslog.o termios.o times.o uinfo.o >uname.o wait.o libkernel32.a libuser32.a libgdi32.a libadvapi32.a >libcomdlg32.a libwsock32.a libnetapi32.a glob/libglob.a ../newlib/libc.a >../newlib/libm.a ../newlib/libm.a -e _dll_entry@12 >`/src1/gnu/beta18/cdk/gcc/xgcc -B/src1/gnu/beta18/cdk/gcc/ -idirafter >/src1/gnu/beta18/cdk/i386-cygwin32/newlib/targ-include -idirafter >/src1/gnu/beta18/cdk/newlib/libc/include -nostdinc >-print-libgcc-file-name` >dcrt0.o: In function `globify(int *, char ***)': >/src1/gnu/beta18/cdk/i386-cygwin32/winsup/dcrt0.cc:197: undefined >reference to `glob' >make: *** [new-cygwin.dll] Error 1 > -print-libgcc-file-name` >dcrt0.o: In function `globify(int *, char ***)': >/src1/gnu/beta18/cdk/i386-cygwin32/winsup/dcrt0.cc:197: undefined >reference to `glob' >make: *** [new-cygwin.dll] Error 1 > >The glob/libglob.a is built. >- >For help on using this list (especially unsubscribing), send a message to >"gnu-win32-request@cygnus.com" with one line of text: "help". > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From latexduck@geocities.com Wed May 14 10:25:00 1997 From: latexduck@geocities.com (Justin) Date: Wed, 14 May 1997 10:25:00 -0000 Subject: curious ... Message-ID: <199705141724.KAA07404@omail.eee.org> Hello, i was just curious, IF, anyone was considering making something like a linux distrobution of Cygwin's, im not a programmer, but i would glady help in any way possible for this to happen, i've noticed, OpenNT, and Portage, cost from 900$ to 2000$ just for a NT based unix util distro. If anyone is interested, please let me know, i think it would be fun ;) - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From franl@world.std.omit-this.com Wed May 14 11:06:00 1997 From: franl@world.std.omit-this.com (Francis Litterio) Date: Wed, 14 May 1997 11:06:00 -0000 Subject: UWIN vs. GNU-Win32: comparison wanted. Message-ID: <337cfeeb.8496877@world.std.com> Does anyone have any info about how UWIN compares with GNU-Win32? I have never used UWIN. In particular: How good/bad is UWIN's signal handling? Does UWIN support process groups? How does UWIN's fork implementation differ from that of GNU-Win32? How does UWIN deal with the perennial issue of text vs. binary data in files and pipes? -- Francis Litterio PGP Key Fingerprint: franl@world.std.omit-this.com 02 37 DF 6C 66 43 CD 2C http://world.std.omit-this.com/~franl/ 10 C8 B5 8B 57 34 F3 21 Note: email address and WWW URL altered to prevent spam. "They that can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety." -- Ben Franklin, ~1784 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Wed May 14 11:44:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Wed, 14 May 1997 11:44:00 -0000 Subject: cp -p Message-ID: <01BC60B8.9AF99C70@sos> JP Shipherd wrote: > I also have the same problem as Bruce. > > /temp> ls -l tab-test > -rw-r--r-- 1 544 everyone 105 Apr 24 10:13 tab-test > /temp> cp -p tab-test tab-test-new > cp: tab-test-new: Invalid argument > > Sergey, did you have the problem where cp -p gave a Null timestamp (Dec 31, > 1969) to the target file in Beta 17? Yes, I had the same problem. B18 fixes it. > > (BTW, how did you manage to get real user ID's associated with the files > and not user 500?) Remove yourself from "Administrators" group. -- Sergey Okhapkin Moscow, Russia Looking for a job - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From kunglao@prairienet.org Wed May 14 13:58:00 1997 From: kunglao@prairienet.org (kunglao@prairienet.org) Date: Wed, 14 May 1997 13:58:00 -0000 Subject: Any other shells? References: <3379D122.4025@lx.net> Message-ID: <199705142053.PAA27874@bluestem.prairienet.org> > kunglao@prairienet.org wrote: > > > > Is there any work being done to port some of the other common > > shells to gnu-win32? It'd be nice to have a bit of choice of > > shells to use, perhaps tcsh, zsh, or ksh. > > Try ftp.blarg.net/pub/amol > he's done ksh an tsch I belive.... Yes, but since their not compiled with gnu-win32 gcc, all the nifty features, like the mounted dirs, we enjoy aren't there. I haven't gotten around to trying to compile another shell yet, but I'd imagine it'd be fairly easy to get a simple shell made. Perhaps strip out the signaling and stuff (I could live without hitting Ctrl-C for a while longer, though it is nice to have it back in bash). --- "Oh, what sad times are these when passing ruffians can say 'Ni!' at will to old ladies. There is a pestulance upon this land, nothing is sacred. Even those who arrange and design shrubberies are under considerable economic stress of this period of history." -- Roger the Shrubber - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From kunglao@prairienet.org Wed May 14 13:58:00 1997 From: kunglao@prairienet.org (kunglao@prairienet.org) Date: Wed, 14 May 1997 13:58:00 -0000 Subject: b18 - rm still has the ol' same dir name bug? Message-ID: <199705142053.PAA27879@bluestem.prairienet.org> I was rebuilding groff/man this morning, and when I went to delete groff with the usual rm -rf, much to my horror, dirs with a subdir by the same name give a warning and won't be deleted. bash$ md a bash$ md a/b bash$ md a/b/b bash$ rm -rf a rm: WARNING: Circular directory structure. This almost certainly means that you have a corrupted file system. NOTIFY YOUR SYSTEM MANAGER. Cycle detected: a/b/b is the same file as a bash$ That's basically how a couple groff dirs are set up. You have to rm a/* first, then rm -rf a to get around. Also, tar has been spewing a ton of messages about not being able to change access and modification times on the dirs it extracts, on my Win95 at least. --- "Oh, what sad times are these when passing ruffians can say 'Ni!' at will to old ladies. There is a pestulance upon this land, nothing is sacred. Even those who arrange and design shrubberies are under considerable economic stress of this period of history." -- Roger the Shrubber - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jp@nuancecom.com Wed May 14 13:58:00 1997 From: jp@nuancecom.com (JP Shipherd) Date: Wed, 14 May 1997 13:58:00 -0000 Subject: cp -p Message-ID: <3.0.32.19970514090608.00a7ace0@awesome.nuance.com> I also have the same problem as Bruce. /temp> ls -l tab-test -rw-r--r-- 1 544 everyone 105 Apr 24 10:13 tab-test /temp> cp -p tab-test tab-test-new cp: tab-test-new: Invalid argument Sergey, did you have the problem where cp -p gave a Null timestamp (Dec 31, 1969) to the target file in Beta 17? (BTW, how did you manage to get real user ID's associated with the files and not user 500?) --jp At 10:53 AM 5/13/97 +0400, Sergey Okhapkin wrote: >Bruce D. McLeod wrote: >> It used to be that cp -p just messed up the access time, but >> it got worse: >> >> C:\temp>cp -p bdmlog bdmlog2 >> cp: bdmlog2: Invalid argument >> >> Running b18, NT > >The same config. > >1:/usr/src$ ls -l t.c >-rw-r--r-- 1 Administ Domain U 865 Apr 24 12:06 t.c >2:/usr/src$ cp -p t.c ttt.c >3:/usr/src$ ls -l ttt.c >-rw-r--r-- 1 sos Domain U 865 Apr 24 12:06 ttt.c >4:/usr/src$ > >I have no idea why this doesn't work for you... > >-- >Sergey Okhapkin >Moscow, Russia >Looking for a job. > > >- >For help on using this list (especially unsubscribing), send a message to >"gnu-win32-request@cygnus.com" with one line of text: "help". > > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From stefan.bittner@erlangen.netsurf.de Wed May 14 13:58:00 1997 From: stefan.bittner@erlangen.netsurf.de (Stefan Bittner) Date: Wed, 14 May 1997 13:58:00 -0000 Subject: German Keyboard in b18 - no ~,@,|,\ Message-ID: <01BC5C00.00C81800@user3-2.erlangen.netsurf.de> I installed the Beta18 usertools.exe on a German Vesrion of Win95. I use a german keyboard, where some characters (~,|,\,@) are to be typed using the left Alt-Key (It ist named "Alt Gr" on the keyboard) It works in a DosBox, It worked in b17.1 - but not in b18. When I do a cat - > testfile and enter - say "Alt Gr"+Q ( that's where @ is) i see a @ appearing. When I use vi - lienedit-mode (set -o vi) and enter Ctrl+V "Alt Gr"+q I see a ^[@ appearing. That made me think the combination "Alt Gr"+q sends a ESC+@ sequence. I redefined ESC+@ to @ in .inputrc : "\e@": "@" Then I got @ where i wanted it. But this does'nt harmony with vi - style linediting. I can't use ESC+k to get the previous history -line anymore. Any help would be appreciated. Since I did subscribe only the announcment-Mailing List, Iwould prefer a direct Email answer, if possible. The only thing i changed, when upgrading to b18 was - all mounts are binary (including C: to /). Stefan Bittner stefan.bittner@pobox.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From kdw3@po.cwru.edu Wed May 14 13:58:00 1997 From: kdw3@po.cwru.edu (Kevin White) Date: Wed, 14 May 1997 13:58:00 -0000 Subject: Telnetd Message-ID: <337A0AE9.B4FDCA5A@po.cwru.edu> Anyone know of a decent telnet daemon for NT 4.0 that will be have nicely with the win32 port of bash? -- __________________________ Kevin White kdw3@po.cwru.edu CWRU Dept of Genetics __________________________ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From kdw3@po.cwru.edu Wed May 14 13:58:00 1997 From: kdw3@po.cwru.edu (Kevin White) Date: Wed, 14 May 1997 13:58:00 -0000 Subject: CPP Where are you Message-ID: <337A20D5.58812ED2@po.cwru.edu> does anyone have a nice version of cpp or cccp that compiles under beta 18 on a pentium? -- __________________________ Kevin White kdw3@po.cwru.edu CWRU Dept of Genetics __________________________ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From anazawa@hic-soft.co.jp Wed May 14 13:58:00 1997 From: anazawa@hic-soft.co.jp (ANAZAWA Akio) Date: Wed, 14 May 1997 13:58:00 -0000 Subject: cp -p References: <2.2.32.19970512123515.00962504@postman.osf.org> Message-ID: <199705141550.AAA16177@mail.hic-soft.co.jp> In message "cp -p" on Mon, 12 May 1997 08:35:15 -0400, "Bruce D. McLeod" writes: Bruce> It used to be that cp -p just messed up the access time, Bruce> but it got worse: Bruce> C:\temp>cp -p bdmlog bdmlog2 cp: bdmlog2: Invalid argument Bruce> Running b18, NT I'm facing same problem. But at least in my case, copy operation and timestamp preservation seems to work fine. > bash$ touch tmp1 > bash$ ls -l tmp1 > -rw-r--r-- 1 544 everyone 0 May 15 00:34 tmp1 > bash$ date > Thu May 15 00:35:20 GMT--9:00 1997 > bash$ cp -p tmp1 tmp2 > cp: tmp2: Invalid argument > bash$ ls -l tmp? > -rw-r--r-- 1 544 everyone 0 May 15 00:34 tmp1 > -rw-r--r-- 1 544 everyone 0 May 15 00:34 tmp2 > bash$ My configuration may same with Bruce. gnuwin32 b18 (from cdk.exe) Windows NT 4.0 + SP2 (Japanese edition, file system is NTFS) Pentium Pro 200MHz (ya, this may no relation with this problem) p.s. As you can see at the citation in above, output from date is still incorrect. I'm setting TZ=JST-9 for our standard time. The output should be like this; Thu May 15 00:35:20 JST 1997 (This is same as output from date of FreeBSD and Solaris. Did GNU change default output format?) Thank you. -- ANAZAWA, Akio - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From b.mcleod@opengroup.org Wed May 14 13:58:00 1997 From: b.mcleod@opengroup.org (Bruce D. McLeod) Date: Wed, 14 May 1997 13:58:00 -0000 Subject: gcc -E slower in beta 18? Message-ID: <2.2.32.19970514202956.01579918@postman.osf.org> At 10:56 AM 5/11/97 +0400, Sergey Okhapkin wrote: >Philip Chow wrote: >> I just recently had some time to move all my tools from beta 14 to beta18 >> and I am noticing a dramatic performance degradation for C-preprocessing >> of files with gcc -E. >> > >File open operations in B18 are much slower due to symlinks to directories support. Symlink_check/Path_conv code requires serious modifications :-( My compiles have slowed down dramatically also. Is there a workaround? perhaps to disable symlinks if we are not using them? Bruce - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From mdt@wwa.com Wed May 14 16:18:00 1997 From: mdt@wwa.com (Mark Terner) Date: Wed, 14 May 1997 16:18:00 -0000 Subject: Using 'mount' for tape drives ... Message-ID: <337A9CA6.52AA@wwa.com> Hi, Is there currently any way to mount a tape drive using the 'mount' command in Cygwin32? Thanks. -Mark ------- Mark Terner mdt@wwa.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From time@ice.com Wed May 14 19:06:00 1997 From: time@ice.com (Tim Endres) Date: Wed, 14 May 1997 19:06:00 -0000 Subject: Any other shells? Message-ID: <01BC60B3.78DE0540@pc.ice.com> >From: jman >Subject: Re: Any other shells? >Try ftp.blarg.net/pub/amol >he's done ksh an tsch I belive.... The correct URL is FTP://ftp.blarg.net/users/amol Has zsh, tcsh, and irc. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From colin@bird.fu.is.saga-u.ac.jp Wed May 14 19:25:00 1997 From: colin@bird.fu.is.saga-u.ac.jp (Colin Peters) Date: Wed, 14 May 1997 19:25:00 -0000 Subject: Does any of you build successfully a DLL with B18? Message-ID: <01BC6122.721CBE80@gbird0> pascal.trouvin@integralis.co.uk wrote: > I've some problems with DLL creation since I've downloaded B18. I want > to be sure that I've not missed somethng in the installation process. > > So in case that some of you have success with B18 and DLL creation, > could you please send me, a little sample script? This is for Mingw32 (and using the soon to be released 0.1.4 files, you need your own DllMain in 0.1.3) so that may be a significant difference. In any case here's the source: --- dll.c --- #include int Add (int x, int y) { printf ("In add!\nx = %d\ny = %d\n", x, y); return (x + y); } --- dll.def --- EXPORTS Add --- exe.c --- #include int Add (int x, int y); int main() { int i, j, k; i = 10; j = 13; k = Add(i, j); printf ("i %d, j %d, k %d\n", i, j, k); return 0; } ------ I use Jam to build my programs, but here are the commands it runs to build the dll and executable from the above source: # compile exe.c gcc -c -O -o exe.o exe.c # create import library for dll.dll dlltool --dllname dll.dll --def dll.def --output-lib libdll.a -k # compile dll.c gcc -c -O -o dll.o dll.c # link dll.dll with relocation info (note the -dll option is defined in the # mingw32 specs file to pass -dll to the linker and set the dll entry point # appropriately, as well as link a different startup file dllcrt0.o into the # dll). gcc -dll -o junk.tmp -Wl,--base-file,base.tmp dll.o rm junk.tmp dlltool --dllname dll.dll --base-file base.tmp --output-exp _temp.exp --def dll .def rm base.tmp gcc -dll -o dll.dll dll.o -Wl,_temp.exp rm _temp.exp # link exe.exe from the source and using the import library. Also relocatable. gcc -L. -o junk.tmp -Wl,--base-file,base.tmp exe.o libdll.a rm junk.tmp dlltool --dllname exe.exe --base-file base.tmp --output-exp _temp.exp rm base.tmp gcc -L. -o exe.exe exe.o libdll.a -Wl,_temp.exp rm _temp.exp ------ Running the resulting exe.exe produces the following (expected) output: In add! x = 10 y = 13 i 10, j 13, k 23 I've only tested on Win95. The only thing I can notice in the source that you give is that your init file will not invoke any constructors/destructors for global C++ objects in the DLL. This shouldn't be a problem in your case though... Good luck, Colin. -- Colin Peters - colin@bird.fu.is.saga-u.ac.jp -- Saga University Dept. of Information Science -- http://www.fu.is.saga-u.ac.jp/~colin/index.html -- http://www.geocities.com/Tokyo/Towers/6162/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Markus.Maeder@siemens.ch Wed May 14 22:52:00 1997 From: Markus.Maeder@siemens.ch (Markus Maeder) Date: Wed, 14 May 1997 22:52:00 -0000 Subject: Opening COM2 Message-ID: <3.0.1.32.19970515074957.0094c100@dilbert> Bob Wilson posted a message some time ago with some C source lines: if ((comfd = open("com2", O_RDWR | O_TEXT)) <= 0) { fprintf(stderr, "Cmd: Unable to open terminal\n"); exit(1); } He stated, this doesnt function under b17. open hangs... It seems, that this doesn't work under b18 either. Is there a patch? Regards, Markus - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jpmorgan@unm.edu Wed May 14 22:52:00 1997 From: jpmorgan@unm.edu (James Paul Morgan) Date: Wed, 14 May 1997 22:52:00 -0000 Subject: My apologies... References: Message-ID: For some reason, the UNM mail system is resending old email from weeks past, including email to this mailing list. My apologies to those seeing threads from weeks ago pop back up, but I have no idea what is wrong here at UNM that this is happening. Feel free to ignore these re-posts. See ya around the Mulberry bush. --James :) - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jimen@adtech-inc.com Wed May 14 22:52:00 1997 From: jimen@adtech-inc.com (Jimen Ching) Date: Wed, 14 May 1997 22:52:00 -0000 Subject: SIGALRM doesn't work? Message-ID: <01BC6083.7589D200@ilima010.lava.net> Hi, could someone verify that the following program does not cause an alarm signal to be generated? Does SIGALRM work at all with b18? That's the version I have on my NT workstation. The NT system id is Microsoft Windows NT 4.00.1381. Does this help identify my system? I have 64 meg ram, 1 gig harddrive, and connected to the my lan through an ne2000 clone. --jc ------------------------------------------ begin --------------------------------------------- /* * socket.c */ #include #include #include #include #include #include #include #include #include #include #include long jiffies = 0xffffff00L; int cbreak(struct termios *saved) { struct termios tio; if (tcgetattr(STDIN_FILENO, saved) < 0) { fprintf(stderr, "tcgetattr failed.\n"); return -1; } tio = *saved; tio.c_lflag &= ~(ICANON); tio.c_cc[VMIN] = 0; tio.c_cc[VTIME] = 0; if (tcsetattr(STDIN_FILENO, TCSAFLUSH, &tio) < 0) { fprintf(stderr, "tcsetattr failed.\n"); return -1; } return 0; } int server_wait(int port) { int lfd, cfd; /* Listen file descriptor. */ struct sockaddr_in saddr, caddr; /* Server/client address structure. */ int clen; /* Get socket file descriptor. */ if ((lfd = socket(AF_INET, SOCK_STREAM, 0)) < 0) { perror("socket"); return -1; } /* Bind to address. */ memset(&saddr, 0, sizeof(saddr)); saddr.sin_family = AF_INET; saddr.sin_addr.s_addr = htonl(INADDR_ANY); saddr.sin_port = htons(port); if (bind(lfd, (struct sockaddr *) &saddr, sizeof(saddr)) < 0) { perror("bind"); return -1; } /* Listen on socket. */ if (listen(lfd, 1) < 0) { perror("bind"); return -1; } /* Accept new connections, block until one is found. */ if ((cfd = accept(lfd, (struct sockaddr *) &caddr, &clen)) < 0) { perror("accept"); return -1; } fprintf(stderr, "Connected with %s.\n", inet_ntoa(caddr.sin_addr)); close(lfd); /* Close listen file descriptor. */ return cfd; } int client_wait(const char *ip_addr, int port) { int cfd; struct sockaddr_in saddr; /* Get socket file descriptor. */ if ((cfd = socket(AF_INET, SOCK_STREAM, 0)) < 0) { perror("socket"); return -1; } /* Connect to server. */ memset(&saddr, 0, sizeof(saddr)); saddr.sin_family = AF_INET; saddr.sin_addr.s_addr = inet_addr(ip_addr); saddr.sin_port = htons(port); if (connect(cfd, (struct sockaddr *) &saddr, sizeof(saddr)) < 0) { perror("connect"); return -1; } return cfd; } int socket_handler(int fd) { int len; char buf[256]; if ((len = read(fd, buf, sizeof(buf))) < 0) { fprintf(stderr, "Read failed with %d.\n", len); return -1; } printf("%s", buf); fflush(stdout); if (buf[0] == 'q') return -1; return 0; } int terminal_handler(int fd) { int len; char buf[256]; if ((len = read(STDIN_FILENO, buf, sizeof(buf))) < 0) { fprintf(stderr, "Read failed with %d.\n", len); return -1; } write(fd, buf, len); if (buf[0] == 'q') return -1; return 0; } void timer_handler() { fprintf(stderr, "This program has been in the main loop for %ld seconds.\n", jiffies); ++jiffies; alarm(1); } int main(int argc, char *argv[]) { struct termios saved_tio; struct sigaction sact; int sockfd = -1; if (argc > 1) { if (argv[1][1] == 'c' && argc == 4) { fprintf(stderr, "Waiting for server connection.\n"); sockfd = client_wait(argv[2], atoi(argv[3])); } else if (argv[1][1] == 's' && argc == 3) { fprintf(stderr, "Waiting for client connection.\n"); sockfd = server_wait(atoi(argv[2])); } else goto error_handler; } else goto error_handler; if (sockfd < 0) { fprintf(stderr, "Link layer connection failed.\n"); goto error_handler; } if (cbreak(&saved_tio) < 0) goto error_handler; sact.sa_handler = timer_handler; sigemptyset(&sact.sa_mask); sact.sa_flags = 0; if (sigaction(SIGALRM, &sact, 0) < 0) { fprintf(stderr, "Can't register signal handler for alarm.\n"); goto error_handler; } alarm(1); fprintf(stderr, "Entering main loop.\n"); for (;;) { struct timeval tv = { 0, 1000 }; fd_set rdset; int fds; FD_ZERO(&rdset); FD_SET(STDIN_FILENO, &rdset); FD_SET(sockfd, &rdset); fds = select(sockfd+1, &rdset, 0, 0, 0); if (fds < 0) { if (errno == EINTR) { fprintf(stderr, "Interrupted, continuing.\n"); continue; } else break; } if (fds > 0 && FD_ISSET(STDIN_FILENO, &rdset)) { --fds; if (terminal_handler(sockfd) < 0) break; } if (fds > 0 && FD_ISSET(sockfd, &rdset)) { --fds; if (socket_handler(sockfd) < 0) break; } } tcsetattr(STDIN_FILENO, TCSAFLUSH, &saved_tio); close(sockfd); return 0; error_handler: fprintf(stderr, "syntax: %s [-c | -s ip_addr] port\n", argv[0]); return 1; } ------------------------------------------ end ----------------------------------------------- - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From kunglao@prairienet.org Wed May 14 22:52:00 1997 From: kunglao@prairienet.org (kunglao@prairienet.org) Date: Wed, 14 May 1997 22:52:00 -0000 Subject: termcap-1.3 on gnuwin32 b. 18 References: <3379DE70@gandalf.optimedia.co.il> Message-ID: <199705150410.XAA26640@bluestem.prairienet.org> > > I tried to build termcap-1.3 > I did `configure' successfully and then `make' gave me the following: > > gcc -c -DHAVE_STRING_H=1 -DHAVE_UNISTD_H=1 -DSTDC_HEADERS=1 > -DTERMCAP_FILE=\"/ > etc/termcap\" -I. -I. -g termcap.c > gcc -c -DHAVE_STRING_H=1 -DHAVE_UNISTD_H=1 -DSTDC_HEADERS=1 > -DTERMCAP_FILE=\"/ > etc/termcap\" -I. -I. -g tparam.c > In file included from tparam.c:29: > e:\gnuwin32\b18\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32\cygnus-2.7.2-97040 > 4\.. > /../../../i386-cygwin32/include/string.h:55: syntax error before `(' > e:\gnuwin32\b18\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32\cygnus-2.7.2-97040 > 4\.. > /../../../i386-cygwin32/include/string.h:55: warning: conflicting types for > buil > t-in function `memcpy' > e:\gnuwin32\b18\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32\cygnus-2.7.2-97040 > 4\.. > /../../../i386-cygwin32/include/string.h:55: syntax error before `const' > gcc: Internal compiler error: program cc1 got fatal signal 33 > make: *** [tparam.o] Error 1 > > Does someone have an applicable patch? comment out that line in string.h (search for bcopy). b17 had the same problem, and that seemed to fix it. how smart it is is up to debate, but I had no problems ever compiling anything that used string.h. --- "Oh, what sad times are these when passing ruffians can say 'Ni!' at will to old ladies. There is a pestulance upon this land, nothing is sacred. Even those who arrange and design shrubberies are under considerable economic stress of this period of history." -- Roger the Shrubber - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Vardhan@bitsmart.com Thu May 15 00:03:00 1997 From: Vardhan@bitsmart.com (Vardhan Varma) Date: Thu, 15 May 1997 00:03:00 -0000 Subject: Any other shells? References: <199705130100.UAA01308@bluestem.prairienet.org> <3379D122.4025@lx.net> Message-ID: <199705151103.LAA03287@linux> >>>>> "jman" == jman writes: jman> kunglao@prairienet.org wrote: >> >> Is there any work being done to port some of the other common shells >> to gnu-win32? It'd be nice to have a bit of choice of shells to use, >> perhaps tcsh, zsh, or ksh. jman> Try ftp.blarg.net/pub/amol jman> he's done ksh an tsch I belive.... zsh too... -- #===# Vardhan Varma | An addicted NetHacker | |PAR| Parsec Technologies (India) | An addicted Advisor | |SEC| Fax and Voice processing systems +------------------------+ #===# No problem can withstand the assault of sustained thinking | - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jman@lx.net Thu May 15 00:27:00 1997 From: jman@lx.net (jman) Date: Thu, 15 May 1997 00:27:00 -0000 Subject: Sorry had to... Message-ID: <337AF31D.6F31@lx.net> Well I was working on a program an it compiled sorta... any way I went an poped it in a hex editor just for the hell of it an what do I see but /pizza/mushroom/noer/ So it seems Mr. Noer is a Mushroom Pizza person :) hehe Jason L. Esman - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Thu May 15 01:32:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Thu, 15 May 1997 01:32:00 -0000 Subject: Problem extracting beta18 on NT4.0 with compressed drive Message-ID: <01BC612A.31BFB7E0@gater.krystalbank.msk.ru> Steve Ganly wrote: > > I'm getting the following error message when trying to install beta 18 > on an NT 4.0 machine. I run cdk.exe, once the extraction reaches > 98% I get the following message: > > InstallShield Self-extracting EXE > > The decompression of %s failed. There may not be enough free > disk space available in the TEMP directory. > I couldn't run cdk.exe and usertools.exe from NTFS partition at all! But they runs from FAT partition... -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Thu May 15 01:32:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Thu, 15 May 1997 01:32:00 -0000 Subject: termcap-1.3 on gnuwin32 b. 18 Message-ID: <01BC6122.E9703AC0@gater.krystalbank.msk.ru> Vassilii Khachaturov wrote: > > I tried to build termcap-1.3 > I did `configure' successfully and then `make' gave me the following: > > gcc -c -DHAVE_STRING_H=1 -DHAVE_UNISTD_H=1 -DSTDC_HEADERS=1 > -DTERMCAP_FILE=\"/ > etc/termcap\" -I. -I. -g termcap.c > gcc -c -DHAVE_STRING_H=1 -DHAVE_UNISTD_H=1 -DSTDC_HEADERS=1 > -DTERMCAP_FILE=\"/ > etc/termcap\" -I. -I. -g tparam.c > In file included from tparam.c:29: > e:\gnuwin32\b18\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32\cygnus-2.7.2-97040 > 4\.. > /../../../i386-cygwin32/include/string.h:55: syntax error before `(' > e:\gnuwin32\b18\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32\cygnus-2.7.2-97040 > 4\.. > /../../../i386-cygwin32/include/string.h:55: warning: conflicting types for > buil > t-in function `memcpy' > e:\gnuwin32\b18\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32\cygnus-2.7.2-97040 > 4\.. > /../../../i386-cygwin32/include/string.h:55: syntax error before `const' > gcc: Internal compiler error: program cc1 got fatal signal 33 > make: *** [tparam.o] Error 1 > > Does someone have an applicable patch? > Do not use text-mode mounts at all! Only binaries one. This will avoid a lot of problems. -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From christian.schenk@detewe.de Thu May 15 01:32:00 1997 From: christian.schenk@detewe.de (Christian Schenk) Date: Thu, 15 May 1997 01:32:00 -0000 Subject: RCS 5.7 on gnuwin32 b. 18 References: <3379CCFB@gandalf.optimedia.co.il> Message-ID: >>>>> "Vassilii" == Vassilii Khachaturov writes: Vassilii> has_memcmp ... OK ./conf.sh: configuring has_memcpy Vassilii> ... OK ./conf.sh: configuring has_memmove ... OK Vassilii> ./conf.sh: configuring has_map_fd, has_mmap, Vassilii> has_madvise, mmap_signal ... make: *** [conf.h] Error 1 Vassilii> make: *** [all] Error 2 *** src/conf.sh.~1~ Fri Jun 16 08:19:24 1995 --- src/conf.sh Sun May 11 11:38:02 1997 *************** case `(uname -s -r -v) 2>/dev/null` in *** 1025,1030 **** --- 1025,1032 ---- # Early editions of SunOS 5.4 are reported to have problems with mmap # that generate NUL bytes in RCS files with a Solaris 2.2 NFS server. # This has been reported to be fixed as of patch 101945-10. + 'CYGWIN32/NT 4.0 17.5') ;; + # *) $PREPARE_CC || exit if ($CL -DTRY_MAP_FD=1 a.c $L && $aout &2 -- Christian Schenk Cordless Technology A/S ZN Berlin - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From csoelle@sghms.ac.uk Thu May 15 01:32:00 1997 From: csoelle@sghms.ac.uk (Christian Soeller) Date: Thu, 15 May 1997 01:32:00 -0000 Subject: cd problems with //x paths References: <97May14.091002pdt.35713-1@gateway.fluke.com> Message-ID: <5067wlrvf8.fsf@mbcsg1.sghms.ac.uk> Warren Jones writes: > > Has anyone else noticed that bash won't let you "cd .." when you're > in a directory like "//x/subdir" ? > > //d/work [65] pwd > //d/work B17.1 (and apparently b18 as well) seemed to have a problem to read the root directory on any disk (related to a bug in one of the readdir, etc routines ?). This seems to be related to the fact that you can't change up to the root directory (//d, //c, /) with a 'cd ..'. Christian - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From vassilii@optimedia.co.il Thu May 15 14:19:00 1997 From: vassilii@optimedia.co.il (Vassilii Khachaturov) Date: Thu, 15 May 1997 14:19:00 -0000 Subject: RCS 5.7 on gnuwin32 b. 18 Message-ID: <337B690B@gandalf.optimedia.co.il> Thanks, it works! I am passing your patch to rcs-bugs@cs.purdue.edu so that they can patch the original distribution. -- Christian Schenk Cordless Technology A/S ZN Berlin - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From garyl@padd.press.net Thu May 15 14:19:00 1997 From: garyl@padd.press.net (Gary Liddle) Date: Thu, 15 May 1997 14:19:00 -0000 Subject: Opening COM2 Message-ID: <8803.9705151818@sage.compnews.co.uk> the way to do it is to fp = fopen("com2", "whatever"); comfd = fileno(fp); > > Bob Wilson posted a message some time ago with some C source lines: > > if ((comfd = open("com2", O_RDWR | O_TEXT)) <= 0) { > fprintf(stderr, "Cmd: Unable to open terminal\n"); > exit(1); > } > > He stated, this doesnt function under b17. open hangs... > > It seems, that this doesn't work under b18 either. Is there a patch? > > Regards, > Markus > > > > - > For help on using this list (especially unsubscribing), send a message to > "gnu-win32-request@cygnus.com" with one line of text: "help". > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Thu May 15 14:19:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Thu, 15 May 1997 14:19:00 -0000 Subject: X11R6.3 and beta 18. Message-ID: <01BC6127.AFC58EB0@gater.krystalbank.msk.ru> Hi! I succesfully recompiled X11R6.3 with beta 18. Binaries runs on both Win95 and NT with DLL versions of libraries. The following patch to X11R6.3+fix01 is almost the same as for beta 17. The only two differences in config/cf/cygwin32.rules - GCCLIB rule and junk.cc rule differs from patch for B17. I'm using only binary-mode mounts. See compilation instructions in my X11R6.3 patch for B17. If you have sources and b17 patch already installed, modify cygwin32.rules according to this patch (or just replace it). -- Sergey Okhapkin Moscow, Russia Looking for a job. begin 600 X11R6.3-cygwin-b18.diff.gz M'XL("-U8>3,``U@Q,5(V+C,M8WEG=VEN+6(Q."YD:69F`,0\:W?B.+*?,[]" MF\V<@6Y(,"0DS6SF- &2L N$P:1#W]/G,,(6X(FQ&3_RF,=_OR7)#\F62=*W M9R\G"595J52E*E65;"NFM5RBJN&-4'6&G@P3FH?DR;#A^M#UK-61X3I+^K4\ MZF_P/3DTEH!20+][]^[=CDY[.@Z03K:H?H;JC9;6:,&%]N%#\[MJM:KFN'=' M3#3$'JJ=(*U&NS2:M,LI'4K\L*&/M>/*L?8!,0#ER0#U)H+KZG<(_1,1![1# M1^_0?([M[1JC=T<4GH!MRPF?./"[]^B?UM(D2Z#M?+ZZZX\:]?F<0A$`+8> M7(;G]AW##DUR:=D$_.SEG:RRUE8GS=Y"_B&Z"O0K-=B3X&?R>V@ M=]D?]'1TCKY F[ND%]K$A_;[U)(Q"*&A&UA+H7V7P<\T+6FI)&HTP7U3B1JG ME>.&(-%T.![D)0HV6UL6*((@\*J%A[WG%, %3)HZ\1Z(%[=5$IV<59J-5*)F MK=)L"A)U+O/R0'"0I&%MA-H;ERQPW.I>W<[BZTN/D N]RYMO\#!Q!,D?Q#6F M]C*!@OG9O[&#:AJJ-5JU&OR SYS6\GXF]0KI$GE&6@-I9ZV3T\@[U9Y6$[RL MHL% ; ;?HUGG9CA$6QL'2]?;M! ZF'5]P JA*2KHNS[S_J-:_:AVQL;\ MT&IH:&79)L'H@*Y[%K L#@1K68.0+D/>*CI/OIH-$V"BL3F$X!< M3V+$05E&G>>5$_IJ)D/\J^M](IYON8[$2D2DK(X+N%A.`1 M55RRB+BK0'01PF1?ND[@[^U][NE9#%]->WNC&]D8MX[UU-EN!7(!HACF&OM7 MAI&WE3PDIZJ_CNK2]3I;._3IKX(*2&Y]XE]A?QX)X%R&H[>)2K^ M;FVI&W3*RV;3,/%&1CX_XS2=NV[1JU"8%"P?'IZ I%1X28-- $,%[@ M!L];HO(*2G49.D8`^HF$>:H1\0,FO, L3_4)0S):^1)9GNK.,DE&K,R,D6 : M/%]Y;BA[F434>]K"; ^Q0\.>GR$2Z*9KCV"3F+-HV0JNP%$Z7I+VN"]C!TYG M8Z+\QW90U1?H.H::;F48HI_X:BHLLAJ8!4.:HO'O34-66*6%AYRFK]E(D&WO4$PSP M>IDOG]9J3R1U_0"JVJZ524R_1+35K6)3^T\*[&6?C(XE5SZX=?9+NXFP5<7=IX)3A:54J:V\#:6+_#HNN2 M1;A*2*LW=53=')\UI8 `CHL]L\N:*<=J=SZG\D"57^W.YH.;3GO0H]#A5.@] MPX]][21JL&'C8M&C.PG55&6L4T6=[H>36[W=FT/Z2J77;G>F^J]_^G=UX_:4JK M&P(^#;N9_C QTYX^[6F;$]T/I@5W1\JHER0/Z $@J%K ?9FH/M_ZQ24'!#X]7"ZM M)WD3&@&YC-J3JD("(EA$F6[BLCHH#=LC_>9VTNF-V]/KLJ;B`JM;*8((!UX- MI0A M,E)D0#54C14C.CN52F&A$#H1"E&1)21(X4JY3@IJ.FFQ-L8>!MMR"#0 M.4;H^6Q%L"%Y:!%N(+!MXD^\!$BP3TO8*ITUZ5:=%QM?LUOB&U#UQH;A7MHS M,:*OV39%'>6=TTFKT7CMSNFL+N^<"K9+7,.C#;:<(ZT>;9NTHQJ,<]PZT5K: M&;K'-EG$NZ:C=_ 'O:-[&,8`^1#]("W8D;4XOU)W,$ /?(/@EUD/L?#B%2UT MXS:VH$>^N $([T6835B$"0H0^C ; MTV-,O],K0LW(4U"(LTB1#%8!8@RU09'8OE+P&]NWX0Z@[ATM=16#%+6#!W.5Y#:=Q$-&"4&# MWLCYZ461!B[F&:M(L(0@)UZ\4F0]8$JG4-F3`%71BC@$YI9$BR9PD0%E-+2C M-64",5IRC>2<5<"SU@>%8L?[3:AB!OT+= [)R%\3D)D8:W>/?^*L M3>_([54'O\ ^@-6]7WZ!:A=%57%2#I_'5?"?O"K^^"7Z?#SZN/KARR^_[+V> M#W?*8E9O$NJ>> ZQP<.+V7%F-N-@\\'A(NZ(&(9.+$3QZWE9!B?)"\ MOQF=4Q4.2GW8AESVK]B]]W)9/4H\@>?,T5#5W>%G?%+0KZ%S?^C"Y$]Z/]_V M)STZN@[F)F@.9'-:3WV$>NJ@Q"-./'!BT/.O]>G$G1*CIHY.31V;G(_WK3Z< M&U/::'TSKC+W/1J+T7Y:PC_R&^Z'ZY_VP50_\=$3#\]2^X%IN3%M3"S10JA$ M\^4&2I(?*5$!%=1L$ B1'WBA`?2PBX!N[^;69AMZ9+X-O JT:).CYC3%_;B/ M8F8H-^1=?]0>]U'L%J7K]J@[Z"%<0=V[FTD70<[W7:>"Z"U0].ZWLL!+9O8' M\A^MP%BC$N]2_J-85YA5@_H&=;CQY*;3T_5Y>SIM=ZY;: '=[V$*7M^SVWMK MS^GUI-?N)D.^O6<\Y(L]_^+&+,(+ACO/FZVPF\=N"B-MEZ?\M6M@9NKR'VJ: MPB7?RL078>VGL7U"-NX#JV=+4O"F]V_C&%D6DX'0H3A&"_0%F5T,F7L#R[EG MCW(B$=2<<3E-FH9-L--J96+OWZ^-0,]GM5PT/K=167HT8H@ M5L NR_P+\DNFDN!;<:E;:*>)0.OB8=^X);&\`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`]6.:7+7M%9=>>%YIW^">3U<.^K+_O?]E_XL#WBXO[3@>++!Q'VYA5<&26.#['%WH M@/GN(S(`_2,V`3,;I5@:MNN3TM*T'-8E78UT)4/NH"7A)EF^$2?T+U0KJP7G M76"-TI=60'Q1^*2_2A(6.:@<2>!(M/.DD/&J^4H&2E1BQ\6B>8_CZG]UWJ,W M-DU48CJ7T9]_IB!Q,97_WPWS-E'_;LLI5]!QH](\_I"L(!IY_J(7[,]R:8?^ M&F1RPR"K4^E6O^KKUZ).I;G>^_FV-YIR?8SUQC6%N?6#0S^8TT?:D50\U\6? M):>GM(X;C9GIQ'PODCCVO6\E\6[;?'M=7@S.L*\]@MJ7_L8OCM#(J0!+(5F! M3[9T],BMUM).6PTY&JOZY /Q67$@KI]6&O4T0=)F.G1\B.F')&W>>%9"8 MH.<$;&.%YH"9>MCQKTAP[?H!#2RT@P@?$^)!!>JIX",2/+K>?=^,D3W/<[T+ M;+*3DZ_SL)FFT=^LAV7 .0_+X%DE1O.]]@'5:O2F@:;E/"S;1_2PXU;MK'6L M[3C47='J\COC\;V&>+,DU5$(16]8T;/926']/H&"/#$BR@?1FU@Q1?QF%F!Z ML_'-9*H7#X7F,Y.>_YEOSZ,KU+D9Z=/V:,JQ'59_#\. /,V7#B.301EZ>G): MHA8!&5KJW1*M",C05J&50AGYENPMLSRD%(Q;:*:4PAVC6A@7U`(KI!GUS M1<X MQHYI\^'FLQX.V%NZO$&7Q4[S7=*LT0\D72+H5:>#C35)#3.%711OP3)D(ZX#( MZ:?M,2"5R;Y>@[D[%2:S`8#&!V$R;QV/K&A,]2!2!?%LIM"^0U\XP[8S M3_UNX'ZR_!#;.U4*V.'+N07S( $>L45'DWSI?M%?01PCJ8=P* W>\;1PR,\A M\9YW.](IJ'K:$/[EP>EQO7+:/$YT!S8Z";KDP3+(11@$KM-.LP@@IR[H''BN M+1+#4E]!UDC>NE?OZ.8SGP3LZ17A3>98_-(V-HO ?33BUJ,1N)M%BO,ITD^Q M/D5';;'G9@&E!K\46<@,Y.[0,K;/R;4C-@1FOK'9"D0;^7@[HJ^RRS=ZQ702 M'=8YYH=UM/H1Q&VMV3IIP ^RR=KU?'1 WWA_7;+"C_0WEZQD<#Y9R?@7[XRK M^KRI'))/-VG)_X40LP]^!$!1TGDO)!V@;/N&9=%@FFFV_>@,20)T0\\@G35V M5L3,P6GTHN?2HP"CTR MA)*>8R#ON;:]P!X]S;X.-Q$3W:)/VH?$"6&VKU@6;2=UK(2G]]D\ED63.E'" MTPR3Q4[)4P`!$4,(F! S21@I1@GH8(\$*;3G%/2'$?N.3SSV[P))>NAYF:;N M%L2V9!:W,*#$>AXAVC0AJA#1JZ4.F=(I%!"=-3'N$Z=+H,S$CMDA-)FFYHGQ MO2=B0#UVNXTF:RY,*COA++HD$XM'S68P(RP([D=1$F[PL4`U,,P0NRVP4=R)>1W#LY7L9 ML+@F:1P4P(/HR+P`!1]Q[8>,2#2$X(5(R:.[ZT!5K71^AF>U80Z676@,K!LX MUSVW!".HTF8T)6#1YL.+J&%`/C?YJZMJY'<6C_W_ KG)0L2 M5@!#=\.\&4/33C!F7*9AJ#'AR\M_/ODG:4E4QD$PFDPM= M^K:DTF5KWR2K_AJOZFHZS:2-G2J%FU0A\-@KU>"#E?Q7V0I_'M10F^VW8M,O M:3(W8YG-9G/1K,-XFF9T48FF,9\X)<=H5(N2C/%QT&7T2V&M7SF&$.&/_\YP M53KM]2U-GE<9IG;]&-Y_.&6CG#:5* D;NQ:93^ MBP[8Z):9+4XU!# Z.&! Z!J!'X/TA)T'A;!SH0!LDVJGI>"D9)-U0<9)H>$` M<2Z+@@HI;:&;R466/:Y7GA4QCL6LQBRV/S?I=(:+H'RCJ0H<1'\JQFI(J)JK M88YPZ_0HW$UI+N6;K,?-)NO1\JD^?2W^1A!F7=-F.=C'!_5N1B+]5M=C,4:[&-[ M\TLV1[&^*IB"%B=XC7+Q' `<`T!,IWD)21G:,+<9L.C@WJ7Y[ )?42=OGTZ[ MN$\CBOLK+$HF\/*LI_'QJWH:"F/I!EK^5@KBVRGR9J@:NTI?`/F29XN3M'0T M4N[8AU'66TZ3%Q(3@'>Y<R;BR MD4^XB641?@J5E$DX.T+[`8;.:\'I#]P9DW("G9(U;3[]@4?2!T%VPGRNT^(! MC5IZ:U%+@E%+YH:4@W.7S3*7&J+1E$Q!Z$/[9W-I%NCG>0UE?'LV35$CD>$L M@U-H`LSK_J<`.&@'P%&0CM(9^&N5]/LOW=B=!<794 MIJL5WD-''&Y>`$9T44L"GG**3[@0WZ/ PJQ/6%H)+Q<3BS"H+"=CGP&#TWB MV<[909=P:+6.BJ\R5"L'.E<77C*/[]*))R@V.7"6CYW#>SQI\C6YRY-GC?RY M%]WNM??V-?:7>!GK] 4Z5/L5I%U!#BK(H49<1,4A=2L;#^>#=]:+!OK-W6RU ML=!%]ISD'.<)*->KE:.@0\"#*A=75# 4>6[V.'8_N-<"&MCS(5Z)$239ZIO] M5-J)!*Z$AZC,\J1*U!+6P529[84>C\ODN8'"_G4_7GU-,4Y/&,?*&U3NS;(+PK$25^.>-I,//>W564OE6(ZGLK%8@&=ES-*9&R;HS`+MT M6R<]FE$N^PY#EXB?QU K/IHM*DA_3<1F<4FI0U(XU5QCDL_P7"+&70IN(=A1 M2A_H=&$;9A )U_@@9_L:%SX@O0SSB0FF7^E#A?31:Y)V9$MSZ"UH M)V8`F0ABJDI G&&#@KI9?7@'YF'B2V)UD)"P-6,-'DI"Z'R7UA< P@ M;5:)GU\&.4"PWT$UK]$W"D3U=1Z7'RHB!E"9Y.\K(=#48N#&@EOCF(!](.$IF,2KM M(5_F65SZ2+;T@>J[G,5EH0B<5;_8-;C%:+WBQ9V&8'2,\:@TICN@,=4'IVOI MZF(0(#927M:J6XV>`M,;3ST@X$QG^<89.#[9.A<:-E:'A^G!M* :3XU%23U, M-Z*$%&4J:%AT)RCCLM*87O6UWDP:$)1F&:/CMZEVR)M5`[XVL78C_=9LUC"% M^;B@(Q3YTHQR,XT%;;A "F]]>1&["9E?LHYS\N8+L@8=QH5[RXFU(L^TA,B8LB<;Y;L5& `$*6J@Y)?3#-9W2C],S,/OOOV0JL MZ2"SRA4&\TRSQC7-&)ORJ-,'=_^C(>\]XYH7B<]M$X+7%AK MX/,A^HV#P^6J:EUS! O:Y73"[)YXM3';%9GANCN6"N.#FAYHXOGWL!"7./V*T9Z1)_BZ8 \OC//P_BY M6G2UZH.Q>)%E;#K0YC/EZL?%HUW">,2.T"JBG5/9+ZN\1<%HK55)=4W#0[P$ MRI+A``3!:"N#@ERL?)+7$Q/QPK"AV' .M#L$'%(5UF/#O=(0!8>M9Y(=9$[6 M=0;9$#$S= *[?26OF1PK42-K'+N;&V6/BA;%-HTRF2Z Z3W!SED9=/@[:P9S MMUREB'EO&2RQ>TXH1AA^P4#J?=_ 'C3"2+B=N@L.(Q"O=#:P?9)T) M6WJBL0:^7L[@[2X239M7"$!!G(_NBZ$`6-6R)B*^#L= M.H-^'F6KB^0IF7L8&BT8[M&@O]F(R$W?2[DETU?C)L_*XKU>+CRJ;#XD MYUVMD"4%(RA/XEZ==V5HNWB'F8T<=.G'4NL\\1KQF)PGV2(I\\T0#\T7#N?M M4P\%)E>%:7?-E)8&%-8L\@`3`K2F7D+[S:))FUHAIY2RY5^23;%9&,S-KXE? M635YGM1BT?KNR6SL$F"W>1_8+E/&I"#H;SH7FCSF$SRQ;(,S^% ,[K55PF) M9V/3F3A@E'DV8*_X:E]DWA'84*Q8OUC?6<7JRB KVD=J=;#H(9/;TZ"XEQ\O M9VNP>U&"\HR 6_>85@1'G5G'R_4:!+=LXSD3GA-\%J@PG[/!(RK,>O8U3M]C M3* SF20KNYE^JT;<&,.60W8L&51GE MO32T",OA`3F37M]-;./,)F;)<5'%[U>Q.8M12H7^8L#C.5=Q^6!V5F22U*/7 MJ6@-IJ![+VJ-L%(U1/RH*D!=@5M@IO'B#(QUAP;WC'#A9&K$N1E'6'IW69RS MZ%?OC0*2O(X/I7$5;EM&5*+** K<5.%L##X;(Q,563&%R>Y#.I^&Z=R,7H(2 M=07^W$.RU/.,LM0O[2&F/&U6L?KA-0MNM3C/D$+#8^"(-UG^:!<0"P=5F-*Z M/ %>%82$M6#\W T$`!AZ#(#E5-G2,H@NBU_#95:FDT0UQ*]/5U:IB469ZQ2G M55UR*Y[J%".J4Q02O<^3X@%GQAJ_-GQ*M0L_W,0I'@/#KYV4#YPU\X2P2AFG M'TT#=^:8DL!/^IEXRP)T]-,:8#._- .FO$VY&AAR==C%13>TD)DR6'J^.FMI M1?=87VI2G"^"7@Z8\;SXSWQ4C)\B_/7K#"_C)T8?ZPRFJK.7R7Q=&"EFT,ML M6054SIQ-(S/R.FF'WG/??=7$^C97F(NY!\J9C5FGS/GRMF0Y25VYU0;W%6W( MR !6T@HVREP6TAT]X.Z7&!J]QNQ$66@5'MZ6V]U'S@-7/EN>@ MUPX+%,IY4B4T'!0(IT6JF?[7ZN=A5[G*$ MHJLB=^S"C6RH0](9_#O)ILY;,Y5A67_#C2U.XT"*A6K]2:.Y-3#A`PLFB#+J M>]%7"GFI9BE-M2Z3"SJ0I >YC_&;ZP* 4_RZ9TP_V*G0S%D,H?D[;>#NS/&, M:Y:[@5!TM0_BZ%?H4-66)$I-&78ZJ;?>5HT=F@9?3[VD\ CV`Z2-.8(='+)R M0*5[/$D(F3M1\K\*,AMF)GWKZH&.K&+57IBH*[1XE%%#1_80!2OH6R\_3"8=IZC6)-;@J.^6(OT,N)(1CPQDT_0^ MQ9_)FG.BB/ZPCO-''U*EC:(6PF:5J*R]IOX1WXJN'-\VM5T8#5H;C@&LUO4\ MSJD:@Y"C8+*=;%0SQ&>HU&\94X46`<.-R"1X(?AN-6TDD8N_\UX)VI@OCO4:\*Y9\N8[N!^K[;NUTL@J=@P:BH^:V0:R>608ZY!MM,?O'5E%_\\YWR553.U(%ZCPG9B'^IA:0U$\X4;GU3[A#Y\/Z"N%S5>-[>T$<<7?^S1=]/: W&W>%99_1_@_%H>'W9A7^'IS=#^.>D M=]D9_K@#;?GTJ7)O#55#-\69>B0;77E /?@S0US-ZJ4[GMJ'*ASJ_1[C=Q MS0O9F+NW](]CEP"N\$E MZ<2V\?(((?'W^_1Y88?-8,TE;$G&0_W7KG_Y&"O MO0-_C@QWV'N=\,A _2-'R$%GE]_J4&_1EO%#*C;MI9-^J_6CO? M.>[2@?7UI4I9W.V$G^BVHO;W[_!Y;^?@LQ7F MT/-B`^[1HMAI%1E^?G+36L2;UEUB1F3Z)\I6,P>X<&AXFQ>,S.X_LV:*27;0 M/I19,C?,XF>MOFU-4OJ$6)H]T9_)LMS^,!IV+J,M_/W M^VZ#%@/W;TJ_3>_ M(?RM]_]GV/>7&9A:1CDZWCDX_F@8)9"7V.M@"6\WK6'NB0QA\RC^`NQQ,^R- MSEX;!KKP[&<81/HM#/(;[?<;V>)?'9 WB<&?Z ]>G>H$H,8JHD\3G2'Z&7]$ M_/'C]X>?*S\B]@I4Q=TK][,=?MKYV+8\W#*?5UG?DJKK'YY?7W6""N1:_YO'XH'W2&^'TDM5+G3(, M^FMT*FB-SV__SN[^+ /*IUE (_*7#79OZ1+7W;@L\_1N729B/;\EG\>H;RG@ M?93@,_ZTNRT?)1#F?5,E[V+H_?;!P0[\L7>CXG_D2XL3/B6!GL@_:@1.-#P? MGW2BLU-O`N@&:%^-%>NE/R'1M^'A=J#J$/(R54I=1Q?;VTH8I<5SL8HGR=9D MN[65%C'>`(7/4"8MF%)F!N62EH')H+1=5S=<_X:[_IJ,_B7&Y/UK0S[TNLO) MQ93E]1NRO;XRJODI,D&F;!LOVFY_#"_>?$L=[UH7!_MXU^N!7A4U\_PW0[*3 MV>D/SDXZOI=WM&=MMK^%=?5[E[U;(?W.)XDH$PUL/M3>[YR?C4]^')V-!\/3 MLZ&:^Y#TX2(Z^9+F1<;DE MG%%"<';^$0@W=CJTUTORJ:)C3?F_FX<"?U@\:4T>H.UWZ_O_.KG^$O7^^M_V M1452/@'>VBK**;1@!S/MM,:]P<7)EYT6YY8XG.F5O7G]5^O5:\SY;^GNVSGV M)4U627Z_N\#=ETW KB&QGE?#7$'$]^ 8>-6+^#:7K')I^Y5K=]I[.T?J"P:L M9;=:^_QQ$:=B==#7^J*=R].MULM.:],"I0;_S4$S;VVW=EM;6T@;]SNWK3^V M]A%I_VEO&W+]`7)M6M #>/XC_E7!U/=QQ7B<%??&4#;=,&S9V(U7;L1O5-V_ ,C6[^/ZOS](E-FP`` ` end - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From maruzz@matrice.it Thu May 15 14:19:00 1997 From: maruzz@matrice.it (Giovanni Maruzzelli) Date: Thu, 15 May 1997 14:19:00 -0000 Subject: BAD NEWS - Apache 1.2b10 on NT Message-ID: Hi there, There's some bad news that maybe is final, at least for me. I've built another version, that is multithreaded. But, also if is very fast, you: 1) can make a lot of connection per second but from one only machine(I've done almost 15 conn/sec) 2) if you connect from two machine at the same time it crash I think the problem is that all threads share the same memory, so, if they try to modify at the same time, they crash. I'vealso utilized a lot of Mutex, but this don't help. There's no mean to have memory storage space reserved for a thread. 3) Put apart that I can't fork from a thread (cause the library return an error), so there's no cgi support. Probably this can be done through a thread that execute the cgi, but at this point, better use msvc++. I think that I'll put out this new (and last?) version today or tomorrow. ciao, -giovanni *********************************************** * * * * * MATRICE srl * * Information Technology Solutions * * * * Giovanni Maruzzelli * * * * * * * * maruzz@matrice.it * * http://www.matrice.it * *********************************************** - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From bernd@asterix.gecko.de Thu May 15 14:19:00 1997 From: bernd@asterix.gecko.de (Bernd Prager) Date: Thu, 15 May 1997 14:19:00 -0000 Subject: Pls help: can't rebuild terminfo after installing ncurses References: <01BC5FF9.3140B170@sos> Message-ID: Hi, after my problems with building ncurses, I got some hints to remount my drive as binary. I done this, but now I get some more problems. I can't even simple programms compile now: ---- snip ------------- [3]: mount Device Directory Type Flags \\.\tape1: /dev/st1 native no-mixed,text!=binary \\.\tape0: /dev/st0 native no-mixed,text!=binary \\.\b: /dev/fd1 native no-mixed,text!=binary \\.\a: /dev/fd0 native no-mixed,text!=binary c:\gnuwin32\b18 / native no-mixed,text=binary [4]: gcc tst.c tst.c:1: syntax error before `<' gcc: Internal compiler error: program cc1 got fatal signal 33 [5]: cat tst.c [5] cat tst.c main() {puts ("hello world!"); } [6]: ---- snip ------------- Did I forgot something? Is it really better to mount the root drive as binary? Thanx for help, Bernd _____________________________________________________________________ Bernd Prager GECKO mbH; Wismarsche Str.3, 18057 Rostock; Germany http://www.gecko.de PGP Key fingerprint = 83 54 6A 3B 7A 9D 6C 0E F3 41 CE 99 11 30 B7 D6 public key by mailto:bpr@gecko.de?Subject=SendPGPKey _____________________________________________________________________ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Rafael.Corvalan@mail.dotcom.fr Thu May 15 14:19:00 1997 From: Rafael.Corvalan@mail.dotcom.fr (Lorena & Rafael) Date: Thu, 15 May 1997 14:19:00 -0000 Subject: Raw device ? Message-ID: <01BC6186.4A7AB940@ppp-197-245.neuilly.club-internet.fr> I'm using b18, and I can't tar directly to my diskette : [rafael] /Home/Rafael >tar cf /dev/fd0 DOMICILE.DOC tar: Cannot open /dev/fd0: No such file or directory Anyone know what's wrong? I got the command line from the FAQ (How can I access other drives?) Thanks Rafael - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Lassi.Tuura@cern.ch Thu May 15 16:10:00 1997 From: Lassi.Tuura@cern.ch (Lassi A. Tuura) Date: Thu, 15 May 1997 16:10:00 -0000 Subject: RCS 5.7 on gnuwin32 b. 18 References: <3379CCFB@gandalf.optimedia.co.il> Message-ID: <337AF62F.3CFB@cern.ch> Vassilii Khachaturov wrote: > Has someone built RCS 5.7 on gnuwin32 b. 18 or another recent build? > I get the following diagnostics when I run make (after configure succeeds): [snip] > ================================================================== > ./conf.sh: configuring has_map_fd, has_mmap, has_madvise, mmap_signal ... Applying the following patch will allow the stuff to compile, and I hope the changes are correct. I'd appreciate if someone could verify the correctness. Cheers, //lat ----- % diff -ru /asis/src/GNU.LANG/rcs-5.7 rcs-5.7 diff -ru /asis/src/GNU.LANG/rcs-5.7/src/conf.sh rcs-5.7/src/conf.sh --- /asis/src/GNU.LANG/rcs-5.7/src/conf.sh Fri Jun 16 08:19:24 1995 +++ rcs-5.7/src/conf.sh Wed May 14 10:32:19 1997 @@ -1025,6 +1025,10 @@ # Early editions of SunOS 5.4 are reported to have problems with mmap # that generate NUL bytes in RCS files with a Solaris 2.2 NFS server. # This has been reported to be fixed as of patch 101945-10. +CYGWIN32/NT*) + # GNU win32 does not permit mmapped file to be opened (returns with + # "not owner"). So which is signal rcs gets is immaterial. + has_map_fd=0 has_mmap=1 has_madvise=0 mmap_signal=SIGBUS ;; *) $PREPARE_CC || exit if ($CL -DTRY_MAP_FD=1 a.c $L && $aout &2 diff -ru /asis/src/GNU.LANG/rcs-5.7/src/rcslex.c rcs-5.7/src/rcslex.c --- /asis/src/GNU.LANG/rcs-5.7/src/rcslex.c Fri Jun 16 08:19:24 1995 +++ rcs-5.7/src/rcslex.c Wed May 14 10:39:40 1997 @@ -1113,7 +1113,7 @@ /* Open NAME for reading, yield its descriptor, and set *STATUS. */ { int fd = fdSafer(open(name, O_RDONLY -# if OPEN_O_BINARY +# if ! (large_memory && maps_memory) && OPEN_O_BINARY | (strchr(type,'b') ? OPEN_O_BINARY : 0) # endif - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From lethin@equator.com Thu May 15 16:10:00 1997 From: lethin@equator.com (Richard A. Lethin) Date: Thu, 15 May 1997 16:10:00 -0000 Subject: PDB files Message-ID: <199705151016.GAA02466@grape-nuts.ai.mit.edu> VC++4.2 doesn't produce STABS symbols describing types and doesn't label variables with type information, even when directed to produce a COFF file. Is there a way to use gdb on programs produced with VC++4.2? Looks like most of the information goes into a proprietary-format .PDB file. Is it necessary to reverse engineer this format? Know of anyone who's done so? Rich - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jbharvey@gte.net Thu May 15 16:10:00 1997 From: jbharvey@gte.net (Justin Harvey) Date: Thu, 15 May 1997 16:10:00 -0000 Subject: termcap-1.3 on gnuwin32 b. 18 Message-ID: <199705152309.SAA10762@smtp.gte.net> I took my working termcap-1.3 dir and tarred it up: http://home1.gte.net/jbharvey/termcapwin32.tar.gz have at it, it's got the object files, Makefile and libtermcap.a in it. j ---------- > From: kunglao@prairienet.org > To: Vassilii Khachaturov ; gnu-win32@cygnus.com > Subject: Re: termcap-1.3 on gnuwin32 b. 18 > Date: Thursday, May 15, 1997 6:07 AM > > > > > I tried to build termcap-1.3 > > I did `configure' successfully and then `make' gave me the following: > > > > gcc -c -DHAVE_STRING_H=1 -DHAVE_UNISTD_H=1 -DSTDC_HEADERS=1 > > -DTERMCAP_FILE=\"/ > > etc/termcap\" -I. -I. -g termcap.c > > gcc -c -DHAVE_STRING_H=1 -DHAVE_UNISTD_H=1 -DSTDC_HEADERS=1 > > -DTERMCAP_FILE=\"/ > > etc/termcap\" -I. -I. -g tparam.c > > In file included from tparam.c:29: > > e:\gnuwin32\b18\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32\cygnus-2.7.2-97040 > > 4\.. > > /../../../i386-cygwin32/include/string.h:55: syntax error before `(' > > e:\gnuwin32\b18\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32\cygnus-2.7.2-97040 > > 4\.. > > /../../../i386-cygwin32/include/string.h:55: warning: conflicting types for > > buil > > t-in function `memcpy' > > e:\gnuwin32\b18\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32\cygnus-2.7.2-97040 > > 4\.. > > /../../../i386-cygwin32/include/string.h:55: syntax error before `const' > > gcc: Internal compiler error: program cc1 got fatal signal 33 > > make: *** [tparam.o] Error 1 > > > > Does someone have an applicable patch? > > comment out that line in string.h (search for bcopy). b17 had the > same problem, and that seemed to fix it. how smart it is is up to > debate, but I had no problems ever compiling anything that used > string.h. > > --- > "Oh, what sad times are these when passing ruffians can say 'Ni!' at will to old > ladies. There is a pestulance upon this land, nothing is sacred. Even those > who arrange and design shrubberies are under considerable economic stress of > this period of history." -- Roger the Shrubber > - > For help on using this list (especially unsubscribing), send a message to > "gnu-win32-request@cygnus.com" with one line of text: "help". - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From marc@watson.ibm.com Thu May 15 16:10:00 1997 From: marc@watson.ibm.com (Marc Auslander) Date: Thu, 15 May 1997 16:10:00 -0000 Subject: B18 title line behavior Message-ID: <9705151558.AA14656@marc.watson.ibm.com> I don't completely like the new title line behavior. In particular, by using full path names, all useful information in the task bar is lost, since only a short prefix of the title appear. Looking for a mimimized program is much harder. Marc Auslander 914 945-4346 (Tieline 862 Fax x4426) - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jeffdb@netzone.com Thu May 15 16:10:00 1997 From: jeffdb@netzone.com (Mikey) Date: Thu, 15 May 1997 16:10:00 -0000 Subject: Pls help: can't rebuild terminfo after installing ncurses Message-ID: <199705151902.MAA11694@nz1.netzone.com> did you get rid of the \r\n's in tst.c? you now have a un*x system, any "text" that you send to gcc or cpp must have LF only. appended is a small program that I rewrote from djgpp source for linux, and gnu-win32 to convert, it will do disk files if given them on the command line, ie ./dtou windows.h gives you a windows.h with no \r's, ./dtou * converts all files in the directory and if called with no arguments stdin|dtou|stdout. or ./dtou< windows.h > windows.new. it will not work across mount points, unless they are on the same filesystem. if you link it to ./utod, it will convert the other way, \n to \r\n ln -s dtou.exe utod, at least under win95, this works. I didn't bother with ^Z since you almost never see it any more as an eof except that micros*** aparently uses it to protect notes at the end of some of their header files.Well what can you expect from a company that is still selling DOS in 1997 disguised as a GUI? ---- From: Bernd Prager To: Mikey Date: Thursday, May 15, 1997 6:31 AM Subject: Re: Pls help: can't rebuild terminfo after installing ncurses > >> Nucrses recognizes systems that define O_BINARY as needing text mode >> reads. >> >> Solution, remount your filesystems as binary. >> ie for drv in d e f g;do umount $drv:\\;mount -b $drv:\\ /$drv;done >> then use registry editor to change fbinary to 1 for /, >> and reinstall user-tools.exe, and cdk.exe, to fix the header files. >> you will find that many things work much better. >> ---- > >I done it exactly. Now I can't compile simple programs: >-------- snip ------------------- >[6]: gcc tst.c >tst.c:1: syntax error before `<' >gcc: Internal compiler error: program cc1 got fatal signal 33 >[7]: cat tst.c > > >main() >{puts ("hello world!"); } > >[9]: mount >Device Directory Type Flags >\\.\tape1: /dev/st1 native no-mixed,text!=binary >\\.\tape0: /dev/st0 native no-mixed,text!=binary >\\.\b: /dev/fd1 native no-mixed,text!=binary >\\.\a: /dev/fd0 native no-mixed,text!=binary >c:\gnuwin32\b18 / native no-mixed,text=binary >-------- snip ------------------- >Did I forgot something? >Bernd >_____________________________________________________________________ >Bernd Prager >GECKO mbH; Wismarsche Str.3, 18057 Rostock; Germany > http://www.gecko.de > >PGP Key fingerprint = 83 54 6A 3B 7A 9D 6C 0E F3 41 CE 99 11 30 B7 D6 >public key by mailto:bpr@gecko.de?Subject=SendPGPKey >_____________________________________________________________________ > -------------- next part -------------- A non-text attachment was scrubbed... Name: dtou.c Type: text/x-c Size: 2880 bytes Desc: not available URL: From jeffdb@netzone.com Thu May 15 16:10:00 1997 From: jeffdb@netzone.com (Mikey) Date: Thu, 15 May 1997 16:10:00 -0000 Subject: Picking up include directories automatically Message-ID: <199705152232.PAA00504@nz1.netzone.com> You don't change a multi platform developement environment, to accomadate a temporary kludge that will probably go away in 1 or 2 revisions. Since the dll and gcc now support dos style naming. and the dll supports UNC's the need for // isn't really apparent, at least not to me . I'm kind of surprised that Geoff left the support in the dll, (plus \\\\c is a pretty awkward way to refer to c) but he was rushed getting out this release. Also not all gnu tools work as of yet with the dll, posix support is only about 90-95% done, so some programs eg. emacs won't yet compile and run, under gnu-win32. untill that fatefull day, there will remain kludges, live with it. personally, I would dump support for // and dos style path's, (neither is portable, and will only screw up the code base) require all dirs to be mounted. and let the users worry about it if they want to use someone elses tools, so be glad I'm not running the show, (You don't like the way my software works DELETE IT AND DON'T EVER CALL ME AGAIN). >At 08:31 AM 5/15/97 -0700, you wrote: >>gcc, and cpp understand both, // isn't posix, it's a kludge for cygwin32. > >That's my point. Since we need to use this // cygwin32 klugde and since the last two messages I've sent you explained that you don't need // for anything. >gcc and cpp require cygwin32, why not make gcc and ccp understand the >pathname conventions of the dll which they require to operate? > >--jp > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From rminnich@sarnoff.com Thu May 15 16:10:00 1997 From: rminnich@sarnoff.com (Ron G. Minnich) Date: Thu, 15 May 1997 16:10:00 -0000 Subject: UWIN vs. GNU-Win32: comparison wanted. References: <337cfeeb.8496877@world.std.com> Message-ID: the only thing that really mattered to me in uwin was the rule on source: you don't get it. rm -r * ron Ron Minnich |"I would point them out but ... rminnich@sarnoff.com | I have no hands." -- Coconut Monkey (609)-734-3120 | (see CM at www.pcgamer.com/coconut.html) ftp://ftp.sarnoff.com/pub/mnfs/www/docs/cluster.html - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From vassilii@optimedia.co.il Thu May 15 16:10:00 1997 From: vassilii@optimedia.co.il (Vassilii Khachaturov) Date: Thu, 15 May 1997 16:10:00 -0000 Subject: termcap-1.3 on gnuwin32 b. 18 Message-ID: <337B4FCE@gandalf.optimedia.co.il> Thanks, it works. vassilii@optimedia.co.il >>>kunglao wrote:<<< > comment out that line in string.h (search for bcopy). b17 had the > same problem, and that seemed to fix it. how smart it is is up to > debate, but I had no problems ever compiling anything that used > string.h. --- "Oh, what sad times are these when passing ruffians can say 'Ni!' at will to old ladies. There is a pestulance upon this land, nothing is sacred. Even those who arrange and design shrubberies are under considerable economic stress of this period of history." -- Roger the Shrubber - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Thu May 15 16:52:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Thu, 15 May 1997 16:52:00 -0000 Subject: UWIN vs. GNU-Win32: comparison wanted. Message-ID: <01BC6143.5CA17700@gater.krystalbank.msk.ru> Francis Litterio wrote: > Does anyone have any info about how UWIN compares with GNU-Win32? I > have never used UWIN. In particular: I have tried to use UWIN. Generally speaking, mr. Korn made very good work. But... The package does not contain cc compiler, only wrapper for MSVC. So, you need MSVC installed to compile anything. I have MSVC, but I could not compile with UWIN anything more complex than "hello, world" :-( Probably, it's a problem with my head :-) UWIN sources are not available. UWIN have very good tty/pty handling, but only when all tools are used from ksh. Tty system does not initialize properly when tools are running from NT's command prompt. For a lot of operations UWIN starts posix tools from NT resource kit. > > How good/bad is UWIN's signal handling? Almost the same as in cygnus B18. > > Does UWIN support process groups? Yes. > > How does UWIN's fork implementation differ from that of GNU-Win32? > No differences. The same method like in cygwin. But most of UWIN tools are rewritten to use spawn() instead of fork/exec, so they runs a bit faster. > How does UWIN deal with the perennial issue of text vs. binary data in > files and pipes? All files/pipes/etc are handled in binary mode only. -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From D.J.Fiddes@hw.ac.uk Thu May 15 16:52:00 1997 From: D.J.Fiddes@hw.ac.uk (David Fiddes) Date: Thu, 15 May 1997 16:52:00 -0000 Subject: Problems building b18 under win95 Message-ID: <199705151008.LAA00154@ma.hw.ac.uk> Hi, I'm having quite a few problems rebuilding a "plain vanilla" version of gcc for win32. I got the CDK and installed it on a "virgin" machine in exactly the same directory as suggested by the installer except on my F: drive rather than C:. I then created a batch file to set all the path settings and other environment vars. I ran the batch file and did a "umount /" and a "mount f: /" as suggested here a few days ago to make sure that the root, /, directory was connected to the right drive. I then ran bash, create a directory and a simple "Hello World!" app to prove that the compiler was working and stepped through that with gdb to check that GDB and TCL were also working. I then did a "cd /" and a "tar xvf //h/cdk-src.tar" to untar the source for the compiler. Problem number 1: When the tar process was mostly finished it seems to set time stamps on directories and produces the following error:(short version) :-) tar: Could not change access and modification times of cdk/bfd/hosts: Permission denied tar: Could not change access and modification times of cdk/bfd/doc: Permission d enied tar: Could not change access and modification times of cdk/bfd: Permission denie d tar: Could not change access and modification times of cdk: Permission denied ,etc,etc,etc. I thought that this wasn't too critical and I took it to be something to do with the differences between the win32 API on 95 and NT. I then created a /build/native directory and in it ran the "/cdk/configure --prefix=/install -v > configure.log 2>&1" which seemed to work OK although I didn't sift through its log file too closely. I then ran "make > make.log 2>&1" which bombed out after about 30mins with the following error. <...snip snip...> gcc -O2 -c -DDLLTOOL_I386 -D_GNU_SOURCE -I. -I/cdk/binutils -I../bfd -I/cdk/binutils/../bfd -I/cdk/binutils/../include -O2 /cdk/binutils/dlltool .c /build/native/byacc/byacc -d /cdk/binutils/defparse.y mv -f y.tab.c defparse.c mv -f y.tab.h defparse.h gcc -O2 -c -D_GNU_SOURCE -I. -I/cdk/binutils -I../bfd -I/cdk/binutils/../bfd -I/cdk/binutils/../include -O2 defparse.c /build/native/flex/flex /cdk/binutils/deflex.l mv lex.yy.c deflex.c gcc -O2 -c -D_GNU_SOURCE -I. -I/cdk/binutils -I../bfd -I/cdk/binutils/../bfd -I/cdk/binutils/../include -O2 deflex.c gcc -O2 -O2 -o dlltool dlltool.o defparse.o deflex.o bucomm.o version.o filemode.o -L../bfd -lbfd ../libiberty/libiberty.a make: DLLTOOL@: Command not found make: *** [libtcl7.6.a] Error 127 make: *** [all] Error 2 make: *** [all-tcl] Error 2 I'm not sure what to do. Is this a problem with my setup or does it just not build under 95 at all. I'm using Win95 build 950a with British codepages on a P133 with 32MB RAM on a 500MB partition with 150MB free after configure and 132MB free or there abouts after make bombs. Sorry for the long boring email . Dave Dave Fiddes, CALM Software Production Officer Department of Mathematics, Heriot-Watt University, Edinburgh email D.J.Fiddes@hw.ac.uk - Tel: 0131-451-3251 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From vassilii@optimedia.co.il Thu May 15 16:52:00 1997 From: vassilii@optimedia.co.il (Vassilii Khachaturov) Date: Thu, 15 May 1997 16:52:00 -0000 Subject: termcap-1.3 on gnuwin32 b. 18 Message-ID: <337B5082@gandalf.optimedia.co.il> My mounts *are* binary. To accomplish this, after I had run 'cdk.exe' (which by default creates text mounts /BTW I think this should be configurable at the install wizard level, as well as the root mount and the mixed-case mount policies/) I went to the registry key and changed it manually there, after what I made sure with mount that all my mounts are binary. (And only then I unpacked the termcap distribution). Please correct me if I did something wrong. I succeeded to compile termcap-1.3 applying the fix suggested by kunglao (just commenting out the line #55 from the string.h). -- BACbKA vassilii@optimedia.co.il Skribu al mi per Esperanto! ---------- From: sos To: 'Vassilii Khachaturov'; gnu-win32 Subject: RE: termcap-1.3 on gnuwin32 b. 18 Date: Thursday, May 15, 1997 11:27AM Do not use text-mode mounts at all! Only binaries one. This will avoid a lot of problems. -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From crvich@raleigh.ibm.com Thu May 15 16:52:00 1997 From: crvich@raleigh.ibm.com (Ernest Crvich) Date: Thu, 15 May 1997 16:52:00 -0000 Subject: getcwd() and sysdrive Message-ID: <9705151616.AA44472@ode1.raleigh.ibm.com> I just noticed that getcwd() doesn't return the drive in the path for the system drive (e.g., if C: is the sysdrive and you're in \DIR, getcwd() returns "/DIR"). Furthermore, if you're in the root directory of the system drive, it returns the empty string (rather than "/" or "//C/"). For other [network] drives, it works as expected, and returns the form "//W/DIR" and "//W/" for the above tests. This behavior is identical on b17.1 and b18 (on WinNT 4.0), so maybe I'm just expecting too much out of getcwd(). 8-) Fortunately, the Windows function GetCurrentDirectory[AW]() always returns the drive and path correctly, and so is a viable alternative [at least for my purposes]. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From wjones@TC.FLUKE.COM Thu May 15 16:52:00 1997 From: wjones@TC.FLUKE.COM (Warren Jones) Date: Thu, 15 May 1997 16:52:00 -0000 Subject: Segmentation violations with b18 (pdcurses) Message-ID: <97May15.164940pdt.35725-2@gateway.fluke.com> I first built pdcurses 2.3-beta with the b17.1 tools. The distribution includes a gcc/win32 makefile, and after a couple minor tweaks everything compiled without difficulty, and all the example programs ran more or less correctly. I just rebuilt pdcurses with the b18 tools. Everything compiled and linked OK, just as before, but when I went to run the example programs, they all died with a segmentation violation. The stack trace from gdb is not very illuminating (at least not to me): Program received signal SIGSEGV, Segmentation fault. 0x77f6cb4e in ?? () (gdb) bt #0 0x77f6cb4e in ?? () #1 0x77f674c3 in ?? () #2 0x2a92333 in _size_of_stack_reserve__ () #3 0x2a92467 in _size_of_stack_reserve__ () #4 0x2a95717 in _size_of_stack_reserve__ () #5 0x77f6b42b in ?? () #6 0x77f63a9e in ?? () #7 0x77f1372d in ?? () #8 0x77f13445 in ?? () #9 0x77e786a0 in _libwsock32_a_iname () #10 0x77f634c9 in ?? () #11 0x77f61c3f in ?? () #12 0x77f61187 in ?? () Has anyone else had similar problems? I'd be grateful for any insights. -------------------------------------------------------------------- Warren Jones | To keep every cog and wheel is the first Fluke Corporation | precaution of intelligent tinkering. Everett, Washington, USA | -- Aldo Leopold - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From b.mcleod@opengroup.org Thu May 15 16:52:00 1997 From: b.mcleod@opengroup.org (Bruce D. McLeod) Date: Thu, 15 May 1997 16:52:00 -0000 Subject: cp -p Message-ID: <2.2.32.19970515171103.00976258@postman.osf.org> My probem was fixed when I created valid passwd and group files in /etc. Bruce At 12:50 AM 5/15/97 +0900, ANAZAWA Akio wrote: > >In message "cp -p" > on Mon, 12 May 1997 08:35:15 -0400, > "Bruce D. McLeod" writes: > > Bruce> It used to be that cp -p just messed up the access time, > Bruce> but it got worse: > > Bruce> C:\temp>cp -p bdmlog bdmlog2 cp: bdmlog2: Invalid argument > > Bruce> Running b18, NT > >I'm facing same problem. But at least in my case, copy operation and >timestamp preservation seems to work fine. > >> bash$ touch tmp1 >> bash$ ls -l tmp1 >> -rw-r--r-- 1 544 everyone 0 May 15 00:34 tmp1 >> bash$ date >> Thu May 15 00:35:20 GMT--9:00 1997 >> bash$ cp -p tmp1 tmp2 >> cp: tmp2: Invalid argument >> bash$ ls -l tmp? >> -rw-r--r-- 1 544 everyone 0 May 15 00:34 tmp1 >> -rw-r--r-- 1 544 everyone 0 May 15 00:34 tmp2 >> bash$ > >My configuration may same with Bruce. > > gnuwin32 b18 (from cdk.exe) > Windows NT 4.0 + SP2 (Japanese edition, file system is NTFS) > Pentium Pro 200MHz (ya, this may no relation with this problem) > > >p.s. As you can see at the citation in above, output from date is >still incorrect. I'm setting TZ=JST-9 for our standard time. The >output should be like this; > > Thu May 15 00:35:20 JST 1997 > >(This is same as output from date of FreeBSD and Solaris. Did GNU >change default output format?) > > >Thank you. > >-- >ANAZAWA, Akio > >- >For help on using this list (especially unsubscribing), send a message to >"gnu-win32-request@cygnus.com" with one line of text: "help". > > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From riemer@fossi.uni-weimar.de Thu May 15 16:52:00 1997 From: riemer@fossi.uni-weimar.de (Stefan Riemer) Date: Thu, 15 May 1997 16:52:00 -0000 Subject: socketpair?? Message-ID: <199705151153.NAA20296@fossi.uni-weimar.de> Hi, for some reason I need for a project the fun socketpair. Are there any solutions for this in cygwin32? Any suggestions?, Stefan - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jp@awesome.nuancecom.com Thu May 15 16:52:00 1997 From: jp@awesome.nuancecom.com (J. P. Shipherd) Date: Thu, 15 May 1997 16:52:00 -0000 Subject: Picking up include directories automatically Message-ID: <199705151511.IAA09580@nuancecom.com> Unfortunatly the mount option fools lots of other programs and I've found makes things a bit more unstable. Specifically, if I have csh or perl scripts (I'm using the tcsh from the Virtual Unix page) that are in directories on mounted disks they seem to get lost. If I unmount everything and keep the posix paths in my $PATH everything works ok. Couldn't gcc (or cpp) be modified to understand posix paths? (Or better yet, DOS drive letters) --jp > see readme or maybe it's FAQ asbout using // with gcc cpp etc. > mount -b d:\\ /d > and eliminate // from the -I lines, and you should be fine. > ---- > From: Jon Thackray > To: cygnus > Date: Wednesday, May 14, 1997 1:42 AM > Subject: Re: Picking up include directories automatically > > >jeffdb@netzone.com writes: > >I can't get this to work. My specs file has the following:- > >*cpp: > >%(cpp_cpu) %[cpp_cpu] %{posix:-D_POSIX_SOURCE} -I//d/msdev/include > -I//d/msdev/mfc/include -iprefix//d/msdev -iwithprefix include -iwithprefix > mfc/include > > > >plus the other guff that was there already. > > > >My source file contains > >#include > > > >MMSYSTEM.H is in d:\msdev\include > >LIBRARY_PATH is set to c:\usr\local\lib in the control panel, and also > >to /usr/local/lib in sh. > > > >What am I doing wrong? gcc still tells me no include path in which to > >find mmsystem.h > >- > >For help on using this list (especially unsubscribing), send a message to > >"gnu-win32-request@cygnus.com" with one line of text: "help". > > > > > > - > For help on using this list (especially unsubscribing), send a message to > "gnu-win32-request@cygnus.com" with one line of text: "help". > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From p.dalgaard@kubism.ku.dk Thu May 15 22:55:00 1997 From: p.dalgaard@kubism.ku.dk (Peter Dalgaard BSA) Date: Thu, 15 May 1997 22:55:00 -0000 Subject: Bash Problems References: <33796F53.4F7C@cri.dk> Message-ID: Steen Ulrik Palm writes: > > I have downloaded and installed the 18b version of the gnu-win32 user > tools under Windows 95. Bash has now been defined as the default shell > in Emacs and bash runs smoothly within a *shell* buffer. Heyyy!! What a neat trick! Would you mind putting a small document together on how you did that? (Which Emacs version, etc.) > However, I have problems when running bash from a DOS prompt > window. Firstly, the `Alt Char' key does not work. This means that I'm > not able to get access to the characters `$', `|', ``' and others > which hampers the usefulness of bash considerably. I have tried to > define some keybindings for these characters in ~/.inputrc, but for > unknown reasons it does not work. Most of this didn't happen in 17.1, so it should be fixable. However, 17.1 had trouble with deadkeys, which lost us ~, ^, and `, so restoring the old behavior is not good enough. This is a known bug in the DOS box itself under W95/Win32 (it's on Microsofts bug list), and there's no workaround. We need a better terminal window altogether! -- O__ ---- Peter Dalgaard Blegdamsvej 3 c/ /'_ --- Dept. of Biostatistics 2200 Cph. N (*) \(*) -- University of Copenhagen Denmark Ph: (+45) 35327918 ~~~~~~~~~~ - (p.dalgaard@biostat.ku.dk) FAX: (+45) 35327907 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From colin@bird.fu.is.saga-u.ac.jp Thu May 15 22:55:00 1997 From: colin@bird.fu.is.saga-u.ac.jp (Colin Peters) Date: Thu, 15 May 1997 22:55:00 -0000 Subject: BAD NEWS - Apache 1.2b10 on NT Message-ID: <01BC6208.5EF05380@gbird0> Giovanni Maruzzelli[SMTP:maruzz@matrice.it] wrote: >2) if you connect from two machine at the same time it crash > >I think the problem is that all threads share the same memory, so, if they >try to modify at the same time, they crash. >I'vealso utilized a lot of Mutex, but this don't help. > >There's no mean to have memory storage space reserved for a thread. It probably won't make your problems go away but there *are* ways of reserving memory space on a per-thread basis using the Win32 API. Unfortunately the use of these functions will probably require massive changes to the overall form of the server threads... which probably won't be worth it to you... Anyway you can look in the book Advanced Windows by Jeffery Richter (Microsoft Press), chapter 12 "Thread Local Storage" or look at the Win32 API documents for TlsAlloc, TlsFree, TlsSetValue and TlsGetValue. Basically the main program allocates a thread local index using TlsAlloc. Then the multi thread code sets the value associated with this index using TlsSetValue (for example to point to a block of memory holding thread local data), probably right at the start of the new thread. Whenever any function uses TlsGetValue to obtain the value associated with the index (put it in a static variable so that all threads can see it) it sees the value associated with the current thread, and each thread gets its own value. If GCC supported the __declspec(thread) syntax (or some equivalent) then you could simply mark all your thread local "static" or "global" variables with that and everything would be simpler, but I don't think that is likely to work now or anytime soon. If the problem is caused by data that really is supposed to be shared then I suggest you keep looking at your mutexes and critical sections, because it should be possible to make things work that way. Note that the default process heap is serialized, so it should be safe to call malloc or, failing that, HeapAlloc, from within threads. Anyway, just wanted to clear that up (or is that make it even muddier? :) ). Colin. -- Colin Peters - colin@bird.fu.is.saga-u.ac.jp -- Saga University Dept. of Information Science -- http://www.fu.is.saga-u.ac.jp/~colin/index.html -- http://www.geocities.com/Tokyo/Towers/6162/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Thu May 15 23:48:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Thu, 15 May 1997 23:48:00 -0000 Subject: Telnetd Message-ID: <01BC615D.8E51A760@gater.krystalbank.msk.ru> Kevin White wrote: > Anyone know of a decent telnet daemon for NT 4.0 that will be have > nicely with the win32 port of bash? Try ftp://kgb.quarta.msk.ru/NTutil/cs32demo.zip -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Markus.Maeder@siemens.ch Thu May 15 23:48:00 1997 From: Markus.Maeder@siemens.ch (Markus Maeder) Date: Thu, 15 May 1997 23:48:00 -0000 Subject: Opening COM2 Message-ID: <3.0.1.32.19970516084827.00977bd0@dilbert> Sorry, it seems to work on another Computer. I have to check configurations.... Bye, Markus Maeder - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From curtis@pdi.com Thu May 15 23:48:00 1997 From: curtis@pdi.com (Curtis Galloway) Date: Thu, 15 May 1997 23:48:00 -0000 Subject: Some bugs in b18 Message-ID: <337B1D2D.36F6BADD@pdi.com> Here are some problems I have come across so far in b18 that I haven't seen reported elsewhere. 1. Problems with file permissions. In winsup/winsup.h, STD_WBITS is defined as only S_IWUSR. This means that files that are not marked read-only in the filesystem can never have 'group' or 'other' write permission. In my case, this ended up causing "test -w" to report that a file was not writable when in fact it was. (See below.) Here's the diff to make 'other' and 'group' write permissions follow the 'owner' permission, as it did in b17.1: *** winsup.h.00 Fri May 09 13:55:27 1997 --- winsup.h Fri May 09 13:55:39 1997 *************** *** 388,394 **** #define alloca(x) __builtin_alloca (x) #define STD_RBITS S_IRUSR | S_IRGRP | S_IROTH ! #define STD_WBITS S_IWUSR #define STD_XBITS S_IXUSR | S_IXGRP | S_IXOTH #define O_NOSYMLINK 0x80000 --- 388,394 ---- #define alloca(x) __builtin_alloca (x) #define STD_RBITS S_IRUSR | S_IRGRP | S_IROTH ! #define STD_WBITS S_IWUSR | S_IWGRP | S_IWOTH #define STD_XBITS S_IXUSR | S_IXGRP | S_IXOTH #define O_NOSYMLINK 0x80000 2. Problems with user ids. If I don't have an /etc/passwd file, I get very strange results. For example: bash> whoami whoami: cannot find username for UID 500 bash> touch foo bash> ls -l foo -rw-r--r-- 1 544 everyone 0 May 15 07:11 foo So am I userid 500 or 544? This ambiguity causes "test -w" to fail above, because I'm not the "owner" of the file, and I don't have write permission. 3. Environment variables are uppercased. Windows does the same thing with environment variable names that it does with file names: it preserves case, but it is not case sensitive. In winsup/dcrt0.cc, environment variable names are converted to uppercase to "prevent confusion." Unfortunately, this breaks many Win32 programs that expect to retrieve environment variables by the name under which they are stored in the registry. The right approach, I think, is to make retrieval of environment variables ignore case, as Windows does. The second best approach is not to uppercase environment variables retrieved from Win32. Here's the diff to implement that: *** dcrt0.cc.00 Wed May 14 11:33:44 1997 --- dcrt0.cc Thu May 15 07:20:35 1997 *************** *** 445,450 **** --- 445,451 ---- } (*u->envptr)[n] = 0; + #ifdef NEVER /* Amazingly, NT has a case sensitive environment name list, only sometimes. eg, it's normal to have NT set your "Path" to something. *************** *** 460,465 **** --- 461,467 ---- if (islower (*p)) *p = toupper (*p); } + #endif /* If we're using posix paths, we do a special thing for the PATH [and potentially others]. They might be in native format so we turn 4. Terminal support. The ncurses library does work, but not if you change the default console window settings. If you have a different background and foreground color set, they are ignored, and if you have the scrollback buffer turned on, you get very strange behavior. --Curtis Galloway - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Bertrand.Guiheneuf@inria.fr Fri May 16 06:22:00 1997 From: Bertrand.Guiheneuf@inria.fr (Bertrand.Guiheneuf@inria.fr) Date: Fri, 16 May 1997 06:22:00 -0000 Subject: sh in /bin does not work Message-ID: <199705161320.PAA03242@petibonum.inria.fr> Hi everybody, I have the following problem. I cleaned my old version of gnuwin32 and just installed b18. Now, if I execute the "sh" which is in /gnuwin32/b18/H-i386-cygwin32 all goes well. But if I copy the same sh.exe in /bin (as needed for make) I get a : "can't allocate heap" and tons of errors line. I guess it fails to find a library or has a problem with the mount of / I looked in The register editor and saw there are two entries for Cygnus: One in HKEY_CURRENT_USER/Software/Cygnus Solutions/b15.0 (with the mount entries) and the other one in HKEY_LOCAL_MACHINE/SOFTWARE/Cygnus Solutions/GNU-Win32/b18 Could this have something to do with my problem (I mean the mounts entries are in b15 not in b18) Thanks in advance. Bertrand Guiheneuf - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Peter.Eggimann@mgb.ch Fri May 16 06:22:00 1997 From: Peter.Eggimann@mgb.ch (Peter.Eggimann@mgb.ch) Date: Fri, 16 May 1997 06:22:00 -0000 Subject: select & pipe does not work! Message-ID: Hi, I encountered a problem while porting some software. The select call (waiting for a pipe) returns immediate even thougt there is no data available from the pipe. Then I tried a small test program (see below) and encountered the same problem. pipetest.c ------------------------------------------------------------------------ ------------------------ #include #include #include #include #include main () { int result; int pipeFd [2]; struct fd_set rset; if (pipe( pipeFd) < 0) { printf ("Error %d creating pipe\n", errno); exit (-1); } FD_ZERO (&rset); FD_SET (pipeFd[0], &rset); result = select (pipeFd[0]+1, &rset, (fd_set*)0, (fd_set*)0, (struct timeval*)0); switch (result) { case 0: printf ("Select: received 0 ?!?\n"); break; case -1: printf ("Select: Error occured\n"); break; default: printf ("normal event\n"); break; } return (0); } ------------------------------------------------------------------------ ------------------------ compiled whith: gcc -g -o pipetest pipetest.c I've installes b18 on NT4.0. The same code on a UN*X box works fine. Has anyone any ideas? - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From msanderson@bni.ca Fri May 16 06:29:00 1997 From: msanderson@bni.ca (Michael Sanderson) Date: Fri, 16 May 1997 06:29:00 -0000 Subject: Problem extracting beta18 on NT4.0 with compressed drive Message-ID: <01BC61D3.5008D0C0@aspen.broadbandnet.com> Sergey Okhapkin wrote: >I couldn't run cdk.exe and usertools.exe from NTFS partition at all! But >they runs from FAT partition... I've been able to install and run *most* of the tools on an NTFS partition under NT4.0, so that's probably not it. But ... I have been having some difficulties that I can't solve. Installed fine on NT4.0 with NTFS (cdk.exe, 14058516 bytes) And most utilities run as expected, but ... Can't for the life of me get gcc to work. It dies when the preprocessor runs, just goes nuts reporting: In cygwin_except_handler Error while dumping state forever. I'm executing gcc in a command shell, so as the errors fly by, more and more memory gets tagged against the process. I have to break it with ctrl-c, at which point Dr. Watson generates an app. error log (actually 3 of them): Unable to locate the open procedure "NwPfOpenODIPerformanceData" in DLL "Nwpfctrs.dll" for the "OdiLoad" service. Performance data for this service will not be available. Error Status is data DWORD 0. I dunno, has anybody else had this kind of trouble? I sure don't know what I'm up against. Cheers, Michael - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From m_kan@ya2.so-net.or.jp Fri May 16 07:32:00 1997 From: m_kan@ya2.so-net.or.jp (Masahiro Kan, Home) Date: Fri, 16 May 1997 07:32:00 -0000 Subject: cygwin32-b18-g77-0.5.19.1: patch and binaries References: <9705112141.AA07801@modi.xraylith.wisc.edu> Message-ID: <199705161430.XAA13844@mail.ya2.so-net.or.jp> Dear Dr. Mumit Khan, > I've uploaded the source patch as well as cygwin32-b18-g77-0.5.19.1 > (say that 5 times w/out breaking your teeth ;-) binaries to my web > page at http://www.xraylith.wisc.edu/~khan/software/gnu-win32 Thanks for your effort. I have downloaded g77-0.5.19.1 binary for Cygwin32-b18 from your WWW page successfully. I have tested briefly and it worked fine with some problems. Test1: Electric field calculation program Source code size : 160kB 5.4klines Executable file size: 178kB Note : Confirmed that bsearch bug is fixed. Subroutine zwei was successfully compiled. The execution speed is comparable with that compiled by g77-0.5.20 binary for Cygwin32-b17.1. Test2: Alternative Transients Program Source code size :3.3MB 110klines Executable size :2.6MB Note : The execution speed is about 20% slower than that compiled by g77-0.5.20 binary for Cygwin32-b17.1, and has some problems when executed. Best regards, ------------------------- Masahiro Kan Surge Arrester Dept., Hamakawasaki-Works, TOSHIBA Corporation, Japan. E-mail:"Masahiro Kan at Toshiba" "Masahiro Kan, Home" WWW(ATP-EMTP-jp): http://www02.so-net.or.jp/~m_kan/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From chabe04@cai.com Fri May 16 07:32:00 1997 From: chabe04@cai.com (Benoit Champagne) Date: Fri, 16 May 1997 07:32:00 -0000 Subject: Any other shells? References: <199705130100.UAA01308@bluestem.prairienet.org> <3379D122.4025@lx.net> <199705151103.LAA03287@linux> Message-ID: <9705161215.AA24506@calasu13.cai.com> -----BEGIN PGP SIGNED MESSAGE----- ===>Vardhan Varma writes: >>>>>> "jman" == jman writes: jman> kunglao@prairienet.org wrote: >>> Is there any work being done to port some of the other common shells >>> to gnu-win32? It'd be nice to have a bit of choice of shells to use, >>> perhaps tcsh, zsh, or ksh. jman> Try ftp.blarg.net/pub/amol he's done ksh an tsch I belive.... Vardhan> zsh too... Sorry, but that address does not resolve. /pub does not contain a directory amol. Do you know if the address changed? Do you have another address mirroring his ports (mainly zsh)? - -- - --------------------------------------------------------------------- Continuum all the way COMPUTER ASSOCIATES Benoit Champagne ift.a | Main Line : (514) 978-4899 Tel : (514) 978-4913 | 3090 Boul. Le Carrefour FAX : (514) 978-4909 | Suite 300 Email: chabe04@cai.com | Laval, PQ PLAN : Live Long And Prosper | H7T 2J7 - --------------------------------------------------------------------- - application/pgp-keys -------------- next part -------------- - -----BEGIN PGP PUBLIC KEY BLOCK----- Version: 2.6.3ia Comment: Processed by Mailcrypt 3.4, an Emacs/PGP interface mQCNAzGbL8kAAAEEAMlajekEt3h7gLcFmOov2CdhcUn+53Ux2+IwAaf0kTGPsvXo 0WsQp6yjbRp4rpruxKiduIlRUP6nqixGP7q7LqCEbGCQQHAH59NXefEp7A4muPmI VUhr1zLEyuq5VlWUtyHr0ACQvyeIoumY13ZVYjk2VUeVkD47TXdV9JrD+yYxAAUR tCJCZW5vaXQgQ2hhbXBhZ25lIDxjaGFiZTA0QGNhaS5jb20+iQCVAwUQMZsvyndV 9JrD+yYxAQGaFAP/a+vooLpbv3bXr6hs0ZfhS8x5mgH1x1Z5ImdR7XBCNGZrQNWu s0wEUTylNjE4e9M1Dpv9kVvyHD0OauvE+xNaIFgV2ABKzGzVzyyg0EoTNRBjY9cN iub++OT2QA+56bLF3gy5/vD+z3HzunsfrcgYhsEj30iA0rbQecrfGrVbD2Q= =4XrH - -----END PGP PUBLIC KEY BLOCK----- -----BEGIN PGP SIGNATURE----- Version: 2.6.3ia Charset: noconv Comment: Processed by Mailcrypt 3.4, an Emacs/PGP interface iQCVAwUBM3xPuHdV9JrD+yYxAQHyvgP/aaR97bfL6VkHhdrMehcQvAFfeP1Or7pz BTXh1NQ5CM/y7SRH26evTkKRSxRuh/b4PeYkDZ2LZ66jlsCZrpjIFXBFK+s3RgWn LL8WBZjgtdWtbmGXly2cQgGgpOR9+DwGIQCGEPVn3PXF5C3pZArC6GUDJqV44qIQ UkI6sSCTp80= =q5/T -----END PGP SIGNATURE----- From schulte@acm.org Fri May 16 09:02:00 1997 From: schulte@acm.org (Richard Schulte) Date: Fri, 16 May 1997 09:02:00 -0000 Subject: cp -p References: <2.2.32.19970515171103.00976258@postman.osf.org> Message-ID: <199705161346.GAA13835@cygnus.com> Could you post an example of a valid passwd and group file? > My probem was fixed when I created valid passwd and group files > in /etc. > > Bruce > Richard Schulte Idea Development Incorporated http://ideadev.com/Richard - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From khan@xraylith.wisc.edu Fri May 16 09:02:00 1997 From: khan@xraylith.wisc.edu (Mumit Khan) Date: Fri, 16 May 1997 09:02:00 -0000 Subject: cygwin32-b18-g77-0.5.19.1: patch and binaries References: <199705161430.XAA13844@mail.ya2.so-net.or.jp> Message-ID: <9705161602.AA19643@modi.xraylith.wisc.edu> "Masahiro Kan, Home" writes: > > Thanks for your effort. I have downloaded g77-0.5.19.1 binary for > Cygwin32-b18 from your WWW page successfully. I have tested briefly and > it worked fine with some problems. > > Test1: Electric field calculation program > Source code size : 160kB 5.4klines > Executable file size: 178kB > Note : Confirmed that bsearch bug is fixed. Subroutine zwei > was successfully compiled. The execution speed is > comparable with that compiled by g77-0.5.20 binary for > Cygwin32-b17.1. Well, not quite fixed. I just use a fixed version compiled into the binary instead of using the buggy one from libc/cygwin.dll. > > Test2: Alternative Transients Program > Source code size :3.3MB 110klines > Executable size :2.6MB > Note : The execution speed is about 20% slower than that > compiled by g77-0.5.20 binary for Cygwin32-b17.1, > and has some problems when executed. > Could you please send me the bug report on the exact problem you're having? You can run it under gdb and send me the backtrace, which will be a good starting point. As for performance, g77-0.5.20 is quite a bit better, but a the cost of the following: - new bugs! Complex support is broken with -finit-local-zero and many others. If you end up using the backend complex (fno-emulate- complex), then the performance goes down again. And there is the question of aliasing in 0.5.20 that is new from 0.5.19.1, which requires quite a bit of backend mods. - backend changes which I'm not comfortable with, since I've been out of touch with gcc backend for a few years and too much has changed since. I'll probably wait for 0.5.21 before I do anything further. btw, you should check out a few of the code-generation options to speed up the code (-fast-math and a few others). Regards, Mumit -- khan@xraylith.wisc.edu http://www.xraylith.wisc.edu/~khan/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From bas@healthcare.com Fri May 16 09:02:00 1997 From: bas@healthcare.com (Bret A. Schuhmacher) Date: Fri, 16 May 1997 09:02:00 -0000 Subject: newbie dll question Message-ID: <199705161320.JAA12675@alt.healthcare.com> Hi, I've been using gnu-win32 since back in the v16 days, but never seriously until now. I downloaded v18 recently and attempted to make a relocatable dll following all the instructions I could find (thanks to the people who forged the path :-)). All the files below were taken from the instructions at http://www.cygnus.com/misc/gnu-win32/building-reloc-dlls.txt . My problem is I can't resolve one symbol that appears to be related to the dll creation. Here's fixup.c : asm(".section .idata$3\n" ".long 0,0,0,0, 0,0,0,0"); Here's init.cc #include extern "C" { int WINAPI dll_entry (HANDLE h, DWORD reason, void *ptr); }; int WINAPI dll_entry (HANDLE , DWORD reason, void *) { switch (reason) { case DLL_PROCESS_ATTACH: break; case DLL_PROCESS_DETACH: break; case DLL_THREAD_ATTACH: break; case DLL_THREAD_DETACH: break; } return 1; } Here's my "makedll" script: #! /bin/sh LIBPATH=//d/gnuwin32/b18/H-i386-cygwin32/i386-cygwin32/lib #tclMtherr.o \ #make a .def file echo EXPORTS > tclvc.def nm \ regexp.o tclAppInit.o tclAsync.o tclBasic.o tclCkalloc.o \ tclClock.o tclCmdAH.o tclCmdIL.o tclCmdMZ.o tclDate.o tclEnv.o \ tclEvent.o tclExpr.o tclFCmd.o tclFHandle.o tclFileName.o tclGet.o \ tclHash.o tclHistory.o tclIO.o tclIOCmd.o tclIOSock.o tclIOUtil.o \ tclInterp.o tclLink.o tclLoad.o tclLoadNone.o tclMain.o \ tclNotify.o tclParse.o tclPkg.o tclPosixStr.o tclPreserve.o tclProc.o \ tclUnixChan.o tclUnixFCmd.o tclUnixFile.o tclUnixInit.o tclUnixNotfy.o \ tclUnixPipe.o tclUnixSock.o tclUnixTime.o tclUtil.o tclVar.o panic.o \ init.o fixup.o | grep '^........ [T] _' | sed 's/[^_]*_//' >> tclvc.def #link DLL ld --base-file tclvc.base --dll -o tcl76.dll \ regexp.o tclAppInit.o tclAsync.o tclBasic.o tclCkalloc.o \ tclClock.o tclCmdAH.o tclCmdIL.o tclCmdMZ.o tclDate.o tclEnv.o \ tclEvent.o tclExpr.o tclFCmd.o tclFHandle.o tclFileName.o tclGet.o \ tclHash.o tclHistory.o tclIO.o tclIOCmd.o tclIOSock.o tclIOUtil.o \ tclInterp.o tclLink.o tclLoad.o tclLoadNone.o tclMain.o \ tclNotify.o tclParse.o tclPkg.o tclPosixStr.o tclPreserve.o tclProc.o \ tclUnixChan.o tclUnixFCmd.o tclUnixFile.o tclUnixInit.o tclUnixNotfy.o \ tclUnixPipe.o tclUnixSock.o tclUnixTime.o tclUtil.o tclVar.o panic.o \ init.o fixup.o $LIBPATH/libcygwin.a -e _dll_entry@12 dlltool --as=as --dllname tcl76.dll --def tclvc.def --base-file tclvc.base \ --output-exp tclvc.exp ld --base-file tclvc.base tclvc.exp --dll -o tcl76.dll \ regexp.o tclAppInit.o tclAsync.o tclBasic.o tclCkalloc.o \ tclClock.o tclCmdAH.o tclCmdIL.o tclCmdMZ.o tclDate.o tclEnv.o \ tclEvent.o tclExpr.o tclFCmd.o tclFHandle.o tclFileName.o tclGet.o \ tclHash.o tclHistory.o tclIO.o tclIOCmd.o tclIOSock.o tclIOUtil.o \ tclInterp.o tclLink.o tclLoad.o tclLoadNone.o tclMain.o \ tclNotify.o tclParse.o tclPkg.o tclPosixStr.o tclPreserve.o tclProc.o \ tclUnixChan.o tclUnixFCmd.o tclUnixFile.o tclUnixInit.o tclUnixNotfy.o \ tclUnixPipe.o tclUnixSock.o tclUnixTime.o tclUtil.o tclVar.o panic.o \ init.o fixup.o $LIBPATH/libcygwin.a -e _dll_entry@12 dlltool --as=as --dllname tcl76.dll --def tclvc.def --base-file tclvc.base \ --output-lib tcl76.a ld tclvc.exp --dll -o tcl76.dll \ regexp.o tclAppInit.o tclAsync.o tclBasic.o tclCkalloc.o \ tclClock.o tclCmdAH.o tclCmdIL.o tclCmdMZ.o tclDate.o tclEnv.o \ tclEvent.o tclExpr.o tclFCmd.o tclFHandle.o tclFileName.o tclGet.o \ tclHash.o tclHistory.o tclIO.o tclIOCmd.o tclIOSock.o tclIOUtil.o \ tclInterp.o tclLink.o tclLoad.o tclLoadNone.o tclMain.o \ tclNotify.o tclParse.o tclPkg.o tclPosixStr.o tclPreserve.o tclProc.o \ tclUnixChan.o tclUnixFCmd.o tclUnixFile.o tclUnixInit.o tclUnixNotfy.o \ tclUnixPipe.o tclUnixSock.o tclUnixTime.o tclUtil.o tclVar.o panic.o \ init.o fixup.o $LIBPATH/libcygwin.a -e _dll_entry@12 #build the tclvc.a lib to link to dlltool --as=as --dllname tcl76.dll --def tclvc.def --output-lib tcl76.a #link with main gcc tclAppInit.c tcl76.a -o tclsh76.exe Here's the output when I run it: //d/gnuwin32/b18/H-i386-cygwin32/i386-cygwin32/lib/libcygwin.a(libccrt0.o)(.text+0x8a):libccrt0.cc: undefined reference to `GetModuleHandleA@4' //d/gnuwin32/b18/H-i386-cygwin32/i386-cygwin32/lib/libcygwin.a(libccrt0.o)(.text+0x8a):libccrt0.cc: undefined reference to `GetModuleHandleA@4' //d/gnuwin32/b18/H-i386-cygwin32/i386-cygwin32/lib/libcygwin.a(libccrt0.o)(.text+0x8a):libccrt0.cc: undefined reference to `GetModuleHandleA@4' Is libcygwin.a missing this symbol or is there another lib I should link with? Have I missed something obvious? I'll keep poking at my installation to see if I've hosed something. FWIW, my platform is NT4.0. Thanks for your thoughts! Rgds, Bret -- Bret A. Schuhmacher - Software Engineer bas@healthcare.com These opinions are no one's Healthcare Communications, Inc. fault by my own. I stopped to think and forgot how to start again. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Fri May 16 14:55:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Fri, 16 May 1997 14:55:00 -0000 Subject: SIGALRM doesn't work? Message-ID: <01BC61F5.4E5DE770@gater.krystalbank.msk.ru> Jimen Ching wrote: > Hi, could someone verify that the following program does not cause an alarm signal to be generated? Does SIGALRM work at all with b18? That's the version I have on my NT SIGALRM and other timer stuff are not supported in b18. -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Fri May 16 14:55:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Fri, 16 May 1997 14:55:00 -0000 Subject: getcwd() and sysdrive Message-ID: <01BC61F7.FE236480@gater.krystalbank.msk.ru> Ernest Crvich wrote: > I just noticed that getcwd() doesn't return the drive in the path > for the system drive (e.g., if C: is the sysdrive and you're in \DIR, getcwd() > returns "/DIR"). Furthermore, if you're in the root directory of the system Getcwd() returns cwd in "posix" notation. -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Fri May 16 14:55:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Fri, 16 May 1997 14:55:00 -0000 Subject: B18 title line behavior Message-ID: <01BC61F6.A2DD9150@gater.krystalbank.msk.ru> Marc Auslander wrote: > I don't completely like the new title line behavior. > Comment out SetConsoleTitle() calls in winsup/dcrt0.cc -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Fri May 16 14:55:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Fri, 16 May 1997 14:55:00 -0000 Subject: Some bugs in b18 Message-ID: <01BC6265.8E3CF5F0@sos> Curtis Galloway wrote: > 2. Problems with user ids. If I don't have an /etc/passwd file, I get > very strange results. For example: > > bash> whoami > whoami: cannot find username for UID 500 > bash> touch foo > bash> ls -l foo > -rw-r--r-- 1 544 everyone 0 May 15 07:11 foo > > So am I userid 500 or 544? This ambiguity causes "test -w" to fail It's a strange NT's behavior. Your userid is 500 - "Administrator", but all files you created have owner id 544 (local group "Administrators"). The only solution I know for a now - do not work on NT box as a member of "Administrators" group :-( > 4. Terminal support. The ncurses library does work, but not if you > change the default console window settings. If you have a different > background and foreground color set, they are ignored, and if you have Ncurses uses standard ANSI color mapping (according to terminfo terminal description). > the scrollback buffer turned on, you get very strange behavior. The same. Ncurses takes number of screen lines from terminfo database. It's a Unix world :-) -- Sergey Okhapkin Moscow, Russia Looking for a job - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Fri May 16 15:15:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Fri, 16 May 1997 15:15:00 -0000 Subject: select & pipe does not work! Message-ID: <01BC6268.440B66D0@sos> Peter.Eggimann@mgb.ch wrote: > Hi, > I encountered a problem while porting some software. The select call > (waiting for a pipe) returns immediate even thougt there is no data > available from the pipe. Looks like a cygwin bug. Pipes are handled with fhandler_base:fhandler_disk_file class, whose member function always_read_ready() returns 1 (fhandler.h). We need a separate fhandler_pipe class... -- Sergey Okhapkin Moscow, Russia Looking for a job - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jk@soas.ac.uk Fri May 16 15:15:00 1997 From: jk@soas.ac.uk (Jonathan Kaye) Date: Fri, 16 May 1997 15:15:00 -0000 Subject: Telnetd References: <337A0AE9.B4FDCA5A@po.cwru.edu> Message-ID: <3.0.1.32.19970516095028.008ed0b0@soas-hub.soas.ac.uk> At 14:56 14/05/97 -0400, Kevin White wrote: >Anyone know of a decent telnet daemon for NT 4.0 that will be have >nicely with the win32 port of bash? I use Ataman's telnetd and it supports gnuwin bash ok (unlike M$'s telnetd which comes with the NT4.0 reskit) http://www.ataman.com Jonathan Kaye Department of Linguistics SOAS Thornhaugh Street London WC1H 0XG UK Tel. +44-171-323-6362 Fax. +44-171-323-6362 WWW. http://jk.soas.ac.uk - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Alec@tca.co.uk Fri May 16 16:26:00 1997 From: Alec@tca.co.uk (Alec Clews) Date: Fri, 16 May 1997 16:26:00 -0000 Subject: Handy script for MS Internet Explorer Users Message-ID: Here is something I wrote to export the bookmark URL's out of MS Internet Explorer so that you can load them into another browser. Enjoy, let me know if you find any problems with it Regards, Alec -- Alec Clews, , PGP keyid:48FA EB81 TCA Consulting Ltd 44-(0)171-415-8159 Fax:44-(0)171-556-0022 New CIty Court, 20 St Thomas Street, London, Britain, SE1 9SD ====================== Usual Disclaimers Apply ===================== [Personal and PGP key http://www.earth.demon.co.uk/alec ] #! /bin/sh # traverse a diretcory of MS Internet Explorer URL files and # print out an html list of urls to insert in a bookmark file # $1 - location of URL files - default is /windows/Favorites echo '' - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jp@nuancecom.com Fri May 16 16:26:00 1997 From: jp@nuancecom.com (JP Shipherd) Date: Fri, 16 May 1997 16:26:00 -0000 Subject: Telnetd Message-ID: <3.0.32.19970516162526.00aae740@awesome.nuance.com> At 02:56 PM 5/14/97 -0400, you wrote: >Anyone know of a decent telnet daemon for NT 4.0 that will be have >nicely with the win32 port of bash? >-- If you don't mind spending 150$ try: http://www.pragmasys.com/Telnetd.html (They've got free trial period.) --jp - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Scott.Mintz@po.cle.ab.com Fri May 16 16:26:00 1997 From: Scott.Mintz@po.cle.ab.com (Scott Mintz) Date: Fri, 16 May 1997 16:26:00 -0000 Subject: Problems building b18 under win95 Message-ID: <00061639.1893@po.cle.ab.com> I ran into this problem also. It seems that the filtering to remove the @stuff@ values in the Makefile doesn't modify @DLLTOOL@ so it gets left in the Makefile just like that. This is a Configure bug but I'm not sure which file needs to get modified to fix it. The Makefiles produced for both Tcl *and* Tk both contain the following define: DLLTOOL = @DLLTOOL@ If you modify the Makefiles manually, and change the values to: DLLTOOL = dlltool The problem goes away. Configure was supposed to do this. +------------------------------+------------------------------------+ | Scott A. Mintz | voice: (216) 646-4805 | | Allen-Bradley Company | fax: (216) 646-4961 | | 1 Allen-Bradley Drive | email: scott.mintz@po.cle.ab.com | | Mayfield Hts., OH 44124-6118 | CIS: 71461,632 | +------------------------------+------------------------------------+ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sup@cri.dk Fri May 16 16:26:00 1997 From: sup@cri.dk (Steen Ulrik Palm) Date: Fri, 16 May 1997 16:26:00 -0000 Subject: Bash Problems References: <33796F53.4F7C@cri.dk> Message-ID: <337C272C.173C@cri.dk> Peter Dalgaard BSA wrote: > > Steen Ulrik Palm writes: > > > > > I have downloaded and installed the 18b version of the gnu-win32 user > > tools under Windows 95. Bash has now been defined as the default shell > > in Emacs and bash runs smoothly within a *shell* buffer. > > Heyyy!! What a neat trick! Would you mind putting a small document > together on how you did that? (Which Emacs version, etc.) > I use NT Emacs ( http://www.cs.washington.edu/homes/voelker/ntemacs.html ) which seems to be a fairly nice port of GNU Emacs 19.34 to NT/Win95 (I have only used it for a couple of weeks so this is not an expert speaking). It is possible to copy/paste between Emacs and any other Win95 application (including the DOS box). In the Emacs window the copying/pasting is done solely by means of the mouse. To make bash your default shell in Emacs place the definition `set SHELL=\bash.exe' in autoexec.bat and add the below lines to your .emacs file: (setq win32-quote-process-args t) (setq shell-command-switch "-c") and you should be in business! (This is also described in ntemacs.html which is a pretty good README kind of file.) > > However, I have problems when running bash from a DOS prompt > > window. Firstly, the `Alt Char' key does not work. This means that I'm > > not able to get access to the characters `$', `|', ``' and others > > which hampers the usefulness of bash considerably. I have tried to > > define some keybindings for these characters in ~/.inputrc, but for > > unknown reasons it does not work. > > Most of this didn't happen in 17.1, so it should be fixable. However, > 17.1 had trouble with deadkeys, which lost us ~, ^, and `, so > restoring the old behavior is not good enough. This is a known bug in > the DOS box itself under W95/Win32 (it's on Microsofts bug list), and > there's no workaround. We need a better terminal window altogether! > Ok, we have to look/wait for a port of xterm to Win95. Thanks for your reply. Steen -- \\\|/// \\ - - // ( @ @ ) +-----------------------oOOo-(_)-oOOo--------------------------------+ Steen Palm, CRI A/S, Bregnerxdvej 144, DK-3460 Birkerxd, Denmark Tel: +45 45 94 96 65, Fax: +45 45 94 96 45, E-mail: sup@cri.dk +-------------------------------Oooo---------------------------------+ oooO ( ) ( ) ) / \ ( (_/ \_) - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From js@mapsitna.lorien.slab.de Fri May 16 16:26:00 1997 From: js@mapsitna.lorien.slab.de (Jan) Date: Fri, 16 May 1997 16:26:00 -0000 Subject: beta 18 cdk installation temp space problem under win95 Message-ID: <337C9E6D.4C26@mapsitna.lorien.slab.de> the cdk.exe refuses to run, showing the following MessageBox: "InstallShield Self-extracting EXE Could nopt find enough disk space for extracting files." situation: tried to run cdk from explorer or command prompt. german win95 (NOT win95a, win95b) machine has 32 megabytes ram. TEMP,TMP,TMPDIR point to a network drive with about 700 megabytes free space. local drives c:, WINDRIVE (d:), e: and g: have almost 100 free megabytes each. It's quite unusual if that's not enough, so please state the requirements somewhere one would usually search for them (FAQ); but if this machine should meet the requirements the installer appears to have problems. -- To answer, if you happen to find 'mapsitna.' in the reply address, please remove it. I apologize for the inconvenience. Legal Warning: Anyone sending me unsolicited/commercial/junk/spam e-mail WILL be charged a $500 proof-reading fee. Do NOT send unsolicited advertisements and do NOT add my e-mail address to your list(s): "By US Code Title 47, Sec.227(a)(2)(B), a computer/modem/printer meets the definition of a telephone fax machine. By Sec.227(b)(1)(C), it is unlawful to send any unsolicited advertisement to such equipment. By Sec. 227(b)(3)(C), a violation of the aforementioned Section is punishable by action to recover actual monetary loss, or $500, whichever is greater, for each violation." - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jp@nuancecom.com Fri May 16 16:26:00 1997 From: jp@nuancecom.com (JP Shipherd) Date: Fri, 16 May 1997 16:26:00 -0000 Subject: cp -p Message-ID: <3.0.32.19970516162108.00a997f0@awesome.nuance.com> At 10:45 PM 5/14/97 +0400, Sergey Okhapkin wrote: >JP Shipherd wrote: >> (BTW, how did you manage to get real user ID's associated with the files >> and not user 500?) > >Remove yourself from "Administrators" group. > Can I do this in the /etc/passwd file? I'd like to stay an NT administrator, but I'm happy to let bash think I'm not in the Administrators group. Here's my passwd file: Administrator::500:513:::/bin/sh jp::1003:513:JP Shipherd::/bin/sh Administrators::544:0::: Backup Operators::551:0::: Guests::546:0::: Power Users::547:0::: Replicator::552:0::: Users::545:0::: And my group file: None::513: Everyone::0: Thanks, --jp - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From fjh@cs.mu.OZ.AU Fri May 16 16:26:00 1997 From: fjh@cs.mu.OZ.AU (Fergus Henderson) Date: Fri, 16 May 1997 16:26:00 -0000 Subject: thread-local storage (was: BAD NEWS - Apache ...) References: <01BC6208.5EF05380@gbird0> Message-ID: <199705161526.BAA02259@mundook.cs.mu.OZ.AU> Colin Peters, you wrote: > > If GCC supported the __declspec(thread) syntax (or some equivalent) > then you could simply mark all your thread local "static" or "global" > variables with that and everything would be simpler, but I don't > think that is likely to work now or anytime soon. What would be necessary to get this to work? Is #define THREADLOCAL __attribute__((section(".tls"))) int foo THREADLOCAL = 42; enough, or does the compiler need to generate special code for accessing thread-local storage? -- Fergus Henderson | "I have always known that the pursuit WWW: < http://www.cs.mu.oz.au/~fjh > | of excellence is a lethal habit" PGP: finger fjh@128.250.37.3 | -- the last words of T. S. Garp. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sganly@lehman.com Fri May 16 18:21:00 1997 From: sganly@lehman.com (Steve Ganly) Date: Fri, 16 May 1997 18:21:00 -0000 Subject: Problem extracting beta18 on NT4.0 with compressed drive References: <199705152227.RAA03797@bluestem.prairienet.org> Message-ID: <9705161236.ZM2907@lonadmin226.lonadmin.lehman.com> Thanks to kunglao's advice, I downloaded cdk.exe again and it has installed fine. So it looks like NTFS and compressed drives were false concerns. Interestingly, sum(1) and cmp(1) but think the two files are identical, but as they only take a fraction of a second to run I don't think they're reading the whole 14Mb file. Looks like it's time to learn about MD5 checksums. Thanks to kunglao, Sergey and Ian Reinhart for their suggestions, Steve Ganly. On May 16, 5:30am, kunglao@prairienet.org wrote: > Subject: Re: Problem extracting beta18 on NT4.0 with compressed drive > > > > I'm getting the following error message when trying to install beta 18 > > on an NT 4.0 machine. I run cdk.exe, once the extraction reaches > > 98% I get the following message: > > > > InstallShield Self-extracting EXE > > > > The decompression of %s failed. There may not be enough free > > disk space available in the TEMP directory. > > > > I had that too. I just re-downloaded the cdk.exe again and it > worked ok. > > >-- End of excerpt from kunglao@prairienet.org On May 15, 12:19pm, Sergey Okhapkin wrote: > > I couldn't run cdk.exe and usertools.exe from NTFS partition at all! > But they runs from FAT partition... > > -- > Sergey Okhapkin > Moscow, Russia > Looking for a job. > > >-- End of excerpt from Sergey Okhapkin - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From gb@cs.unc.edu Fri May 16 18:21:00 1997 From: gb@cs.unc.edu (Gary Bishop) Date: Fri, 16 May 1997 18:21:00 -0000 Subject: compile fails with binary mount Message-ID: <3.0.1.32.19970516164708.00cd1c44@mail.cs.unc.edu> I found the problem. I had edited specs and emacs had inserted ^M characters all in it. Anyone know how to make emacs19.34.1 not put ^M on lines? gb >Date: Fri, 16 May 1997 16:28:35 -0400 >To: gnu-win32@cygnus.com >From: Gary Bishop >Subject: compile fails with binary mount > >I'm trying to take the advice I've been seeing to use binary mounts only. But I find that I can't even compile a simple program like dtou.c with a binary mount. I get a message like: > >bash$ gcc -o dtou dtou.c >dtou.c: In function `cvt': >dtou.c:57: warning: assignment from incompatible pointer type >dtou.c:58: warning: assignment from incompatible pointer type >Assembler messages: >Error: Can't open for reading. > : Invalid argument >gcc: Internal compiler error: program as got fatal signal 1 > >my mounts were > >bash$ mount >Device Directory Type Flags >\\.\tape1: /dev/st1 native no-mixed,text!=binary >\\.\tape0: /dev/st0 native no-mixed,text!=binary >\\.\b: /dev/fd1 native no-mixed,text!=binary >\\.\a: /dev/fd0 native no-mixed,text!=binary >d: / native no-mixed,text=binary > >What am I doing wrong? This is driving me crazy! >gb > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Charles_Boatwright@cisnc.canon.com Fri May 16 18:21:00 1997 From: Charles_Boatwright@cisnc.canon.com (Boatwright, Charles) Date: Fri, 16 May 1997 18:21:00 -0000 Subject: Cross compiling for a mips 4600 Message-ID: I know that this has been discussed several times, however, I don't recall seeing results -- success or fail. I need to target a mips 4600. Should I attempt this on NT, or will I end up pulling out all of my hair... I have the FreeBSD build and host working, but it would be nice to have a NT host compiler. To those who tried... Should I attempt a host=nt and build=un*x first? -chuck - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jp@nuancecom.com Fri May 16 18:21:00 1997 From: jp@nuancecom.com (JP Shipherd) Date: Fri, 16 May 1997 18:21:00 -0000 Subject: sh in /bin does not work Message-ID: <3.0.32.19970516175414.00abf100@awesome.nuance.com> Do you still have any b17 exectuables on your path? I had this problem until I deleted my bin directory and repopulated it exclusively with exes from the b18 bin directory. --jp At 03:20 PM 5/16/97 +0200, you wrote: >Hi everybody, > >I have the following problem. I cleaned my old version of >gnuwin32 and just installed b18. >Now, if I execute the "sh" which is in >/gnuwin32/b18/H-i386-cygwin32 >all goes well. >But if I copy the same sh.exe in /bin (as needed for make) >I get a : "can't allocate heap" and tons of errors line. > >I guess it fails to find a library or has a problem >with the mount of / > >I looked in The register editor and saw there are >two entries for Cygnus: >One in >HKEY_CURRENT_USER/Software/Cygnus Solutions/b15.0 >(with the mount entries) > >and the other one in >HKEY_LOCAL_MACHINE/SOFTWARE/Cygnus Solutions/GNU-Win32/b18 > > >Could this have something to do with my problem >(I mean the mounts entries are in b15 not in b18) > > > >Thanks in advance. > > Bertrand Guiheneuf > >- >For help on using this list (especially unsubscribing), send a message to >"gnu-win32-request@cygnus.com" with one line of text: "help". > > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From gb@cs.unc.edu Fri May 16 18:22:00 1997 From: gb@cs.unc.edu (Gary Bishop) Date: Fri, 16 May 1997 18:22:00 -0000 Subject: b18 "patch" assertion fails Message-ID: <3.0.1.32.19970516152217.00cda650@mail.cs.unc.edu> patch dies with assertion "hunk" failed: file "/home/noer/src/beta18/cdk/patch/patch.c", line 339 when I try to apply Sergey's patches for X11R6.3. It has successfully applied several patches before failing. Anyone else seen this? gb - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From m.hadfield@niwa.cri.nz Fri May 16 18:22:00 1997 From: m.hadfield@niwa.cri.nz (Mark Hadfield) Date: Fri, 16 May 1997 18:22:00 -0000 Subject: Gnu-Win32 g++ can't compile ... Message-ID: <199705170122.NAA11534@clam.niwa.cri.nz> I downloaded Robert Pierce's numerical array extensions to libg++ (g++array-0.9.2.tar.gz) and tried to compile a test program. It works OK on a DEC alpha box running Digital Unix (gcc version 2.7.2.1) and on my PC using the EMX+RSXNT port of the Gnu tools (also reports gcc version 2.7.2.1) but under Gnu-Win32 beta18 or beta17.1 I get: In file included from ../include/mArray.h:22, from marr_test.C:19: ../include/array_store.h:145: syntax error before `{' ../include/array_store.h:148: syntax error before `public' g++: Internal compiler error: program cc1plus got fatal signal 33 array_store.h is full of template definitions (which don't mean much to me yet) and the place where the "syntax error" is detected is about the 8th such definition: template // line 143 class value_mask_pointer : public mask_pointer // line 144 { // line 145 private: // line 146 size_t s_; // line 147 public: // line 148 value_mask_pointer() : s_(0) {} // line 149 I checked this has nothing to do with binary vs text line endings etc. Gnu-Win32 works fine on a variety of other programs. Any ideas? ============================================================== Mark Hadfield NIWA (Taihoro Nukurangi) PO Box 14-901 m.hadfield@niwa.cri.nz Wellington, New Zealand - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From oliverf@secnet.com Fri May 16 18:22:00 1997 From: oliverf@secnet.com (Oliver Friedrichs) Date: Fri, 16 May 1997 18:22:00 -0000 Subject: non blocking sockets References: <33796FBB.7F141B06@torino.alpcom.it> Message-ID: On Wed, 14 May 1997, Irio Lavagno wrote: > It seems that sockets can't be made non-blocking in cygwin32 b17 under > winNT 4.0, is it just me or is it true ? We've been having the same problems as well, just to confirm your suspicions. It appears that the socket simply blocks all the time, and a fcntl(x, F_SETFL, O_NONBLOCK) has no effect. - Oliver - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From matiss@outwest.net Fri May 16 18:22:00 1997 From: matiss@outwest.net (M. KARKLINS) Date: Fri, 16 May 1997 18:22:00 -0000 Subject: Split on large files Message-ID: <199705161920.MAA23339@outwest1.outwest.net> I haven't been able to get split to work on large files, specifically the cdk.exe file. The split files available are to large to fit onto 3.5" floppies formatted for DMF, so I was trying to resize the split. There appears to be a size limitation, where it is I don't know. Split does appear to work on smaller files. I am running Win95 on a 100 MHz Pentium with a 500 Meg SCSI HD on a Novell network. Anyone else experience something similar? Matiss email: matiss@outwest.net - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From voyager@eskimo.com Fri May 16 22:21:00 1997 From: voyager@eskimo.com (Nick Rossi) Date: Fri, 16 May 1997 22:21:00 -0000 Subject: Linking VC++ objects with GCC objects Message-ID: <337CB8B0.27F1@eskimo.com> Hello all, I am new to the Cygnus library, so I apologize if this is an old question. I have exhausted the FAQ and mailing list archives looking for an answer, with no luck. I am trying to figure out how to link Visual C++ 5.0 object files with GCC object files. The FAQ says that this "supposedly" works, but beyond saying to use MS's LINK.EXE, doesn't provide any details. LINK.EXE does not seem to find any of the functions in a GCC object file. Is there something specific I need to do - such as append characters to GCC function names or use a GCC flag for the purpose? If so, perhaps it would be worthwhile to update the FAQ with a more specific example. Thanks a ton for any help... -- | Nick Rossi ........... http://www.eskimo.com/~voyager/ | | The Exit Light Show .. http://www.eskimo.com/~voyager/exitlight/ | | Novaterm ............. http://www.eskimo.com/~voyager/novaterm/ | | Freedom writings ..... http://www.eskimo.com/~voyager/freedom/ | - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jazz@softway.com Fri May 16 22:21:00 1997 From: jazz@softway.com (Jason Zions) Date: Fri, 16 May 1997 22:21:00 -0000 Subject: Picking up include directories automatically References: <199705152232.PAA00504@nz1.netzone.com> Message-ID: <337CB6C4.4B88@NoSpam.com> Actually, using a doubled-slash is indeed POSIX; a leading double-slash in a pathname is explicitly called out as implementation-defined, and a portable application should (a) deal with it without blowing up, and (b) not ascribe any particular meaning to it. If anyone is interested, I can get a specific chapter-and-verse citation. The OpenNT POSIX-conforming environment for NT uses /// consistently, with very few (if any) source changes required to publicly-available code. Once the tmpdir()/tmpnam()/tempnam() functions are modified to "do the right thing", most apps simply need slight changes to environment variables or compile-time #defines to build correctly. Apache, for instance, came up quite easily under OpenNT 2.0alpha and 2.0beta. Disclosure: I work for Softway Systems, maker of OpenNT. Jason Zions http://www.softway.com (email address diddled to prevent spamming) - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From ogre@aa.net Fri May 16 22:21:00 1997 From: ogre@aa.net (Joe Rumsey) Date: Fri, 16 May 1997 22:21:00 -0000 Subject: compiling and running ssh-1.2.20 with b18 Message-ID: <337D4051.C9B04DB@aa.net> I've actually got ssh working, albeit with some warnings from cygwin.dll about various things, but I did have to work around a few things in the ssh code: 1) gcc -pipe doesn't work (I am using Win95), but it doesn't return an error either. The configure script for ssh tries to see if it works... result, the first time I tried to make it, it looked like everything was compiling without errors. Until it got to linking and not one object file had been produced. Fixed by editing the configure script to not check -pipe. 2) One file is looking for sys/stream.h and stropts.h for queue_t and strfdinsert respectively. Easy to fix, it didn't actually need any of the code in this file, I #ifdef'd it all out. I'm not sure if this is a problem with the configure script not excluding this file from the makefile or what. 3) There is a getrusage(RUSAGE_CHILDREN, &cru); that causes a SEGFAULT. Again easy to fix, it was only using this to seed a random number generator (one of many things it blends together) 4) It tries to read the password upon connecting from /dev/tty, but never reads anything. Luckily it includes code to read from stdin instead, which works just fine. There are probably some other quirks, I've barely tested it, but the 4 above are what I had to fix to get a basic connection up and running. The version of ssh used is 1.2.20. The ssh home page is http://www.cs.hut.fi/ssh Joe - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From gb@cs.unc.edu Fri May 16 23:35:00 1997 From: gb@cs.unc.edu (Gary Bishop) Date: Fri, 16 May 1997 23:35:00 -0000 Subject: compile fails with binary mount Message-ID: <3.0.1.32.19970516162835.00788c28@mail.cs.unc.edu> I'm trying to take the advice I've been seeing to use binary mounts only. But I find that I can't even compile a simple program like dtou.c with a binary mount. I get a message like: bash$ gcc -o dtou dtou.c dtou.c: In function `cvt': dtou.c:57: warning: assignment from incompatible pointer type dtou.c:58: warning: assignment from incompatible pointer type Assembler messages: Error: Can't open for reading. : Invalid argument gcc: Internal compiler error: program as got fatal signal 1 my mounts were bash$ mount Device Directory Type Flags \\.\tape1: /dev/st1 native no-mixed,text!=binary \\.\tape0: /dev/st0 native no-mixed,text!=binary \\.\b: /dev/fd1 native no-mixed,text!=binary \\.\a: /dev/fd0 native no-mixed,text!=binary d: / native no-mixed,text=binary What am I doing wrong? This is driving me crazy! gb - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From cjeffare@cs.iupui.edu Fri May 16 23:35:00 1997 From: cjeffare@cs.iupui.edu (Chris Jeffares) Date: Fri, 16 May 1997 23:35:00 -0000 Subject: Telnetd References: <3.0.1.32.19970516095028.008ed0b0@soas-hub.soas.ac.uk> Message-ID: I have used and prefer the telnetd by Ataman www.ataman.com Chris J =+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+= Chris Jeffares Graduate Research Assistant Purdue University at Indianapolis office ==> SL041, (317) 274-3089 email ==> cjeffare@cs.iupui.edu www ==> http://klingon.iupui.edu/~cjeffare/home.html The Nebraska Cornhuskers, Gotta luv 'em Go Big Red!!...Huckin' Fuskers in 1997 "Experience...the name everyone gives to their mistakes." =+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+= - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Sat May 17 01:55:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Sat, 17 May 1997 01:55:00 -0000 Subject: b18 "patch" assertion fails Message-ID: <01BC62C1.C6BE1F10@sos> Gary Bishop wrote: > patch dies with > > assertion "hunk" failed: file "/home/noer/src/beta18/cdk/patch/patch.c", > line 339 > > when I try to apply Sergey's patches for X11R6.3. It has successfully > applied several patches before failing. Looks like your xc source tree is unpacked in "text", but not "binary" mode... -- Sergey Okhapkin Moscow, Russia Looking for a job - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Sat May 17 01:55:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Sat, 17 May 1997 01:55:00 -0000 Subject: newbie dll question Message-ID: <01BC62BF.13F1CAA0@sos> Bret A. Schuhmacher wrote: > //d/gnuwin32/b18/H-i386-cygwin32/i386-cygwin32/lib/libcygwin.a(libccrt0. o)(.text+0x8a):libccrt0.cc: undefined reference to `GetModuleHandleA@4' > > > Is libcygwin.a missing this symbol or is there another lib I should > link with? Have I missed something obvious? I'll keep poking at my Add libkernel32.a to link list. -- Sergey Okhapkin Moscow, Russia Looking for a job - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jeffdb@netzone.com Sat May 17 03:17:00 1997 From: jeffdb@netzone.com (Mikey) Date: Sat, 17 May 1997 03:17:00 -0000 Subject: Picking up include directories automatically Message-ID: <199705170929.CAA29400@nz1.netzone.com> Well, since you obviously know what you're talking about, would you mind taking a look at the cdk/winsup/ dir, and giving us the benefit of your advice on fixing THIS implementation to do "The Right Thing :^)"? PS. look out for default text/binary handleing. ---- From: Jason Zions To: Mikey Cc: cygnus Date: Saturday, May 17, 1997 2:10 AM Subject: Re: Picking up include directories automatically >Actually, using a doubled-slash is indeed POSIX; a leading double-slash >in a pathname is explicitly called out as implementation-defined, and a >portable application should (a) deal with it without blowing up, and (b) >not ascribe any particular meaning to it. If anyone is interested, I can >get a specific chapter-and-verse citation. > >The OpenNT POSIX-conforming environment for NT uses >/// consistently, with very few (if any) source >changes required to publicly-available code. Once the >tmpdir()/tmpnam()/tempnam() functions are modified to "do the right >thing", most apps simply need slight changes to environment variables or >compile-time #defines to build correctly. Apache, for instance, came up >quite easily under OpenNT 2.0alpha and 2.0beta. > >Disclosure: I work for Softway Systems, maker of OpenNT. > >Jason Zions > http://www.softway.com >(email address diddled to prevent spamming) > >- >For help on using this list (especially unsubscribing), send a message to >"gnu-win32-request@cygnus.com" with one line of text: "help". > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Sat May 17 03:17:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Sat, 17 May 1997 03:17:00 -0000 Subject: cp -p Message-ID: <01BC62BE.1E8A0000@sos> JP Shipherd wrote: > >> (BTW, how did you manage to get real user ID's associated with the files > >> and not user 500?) > > > >Remove yourself from "Administrators" group. > > > Can I do this in the /etc/passwd file? I'd like to stay an NT No, it will not help. > administrator, but I'm happy to let bash think I'm not in the > Administrators group. NT Resource Kit contains very useful for this case utility "su". Also, I found on the Net freeware su utility for NT. Just run bash in security context of another user. -- Sergey Okhapkin Moscow, Russia Looking for a job begin 600 SU.ZIP M4$L#!!0````(`&A0Y!XVE]"[BP0``*P(```*````4D5!1$U%+E185&U536_; M1A ]RX#_PT"72HA%1W:-(CY%L)-4K>TZ^F@+)#FLR"&YT'*7V5F*5@_][9U9 M2K+<&) _1,Z\>?/>S"XU25K7IR>]Y2X:&/\%\LW]+0QNG"5G<"@-39I0.D]"]59M=,;]E@8&+?]^3_P& M$XN!(WOS@'F.%G[WBEN `:WC/^^U994J%?0Z*9O1"CU33#(![YV>S'"C23L+ MOW(WSF^YS-O+T6_+N]'XW;NK!*96!ZT,U,W*Z!0\&E2$PNL621=6:$T/*FE; M@.H$9,AG8Z0QQ9)X1*" -=3>I4BQJ?&H=^<*#C_6/O7(FG/.G#7S.FQ!9VB# MSC5ZSKD8]3Y8Q8RD7BI(DB0`^YKSP&P8E+K'%ENS!>.*`C.A)G6B:\J"MFR" M2D,"BY)Y\L>B0"J_!=RPH#J/&"\\ES"!$P;.R@1+0?C8O9_0#\!SG)",`>-6HI>0 MF^/4BG^$GQL7U%%;(M/"0>$5=\ME"(]>G@$A1J932T$9TQG(LLG?)&JUB#FA M=<]J3AZGD'(T0OC/Z'I=-6IN\RN1J?P:K1)H/QVZM?$I@OH=7& M,"3W"ZWSZVXI47G#QL &O6P%=66G@G(FW'S<;>L@X_;3`*+#1AEQS^4@Y]1/ M!+XE/WKB#^3<&M3H*TVT;V*&%+R.35'4ZTY7NIO6N \C$-E8VT%_*/R/#Y)X MN,AL".M:$7&+J?-"Q6R3F+PX(GD8T!QRQ50RD$EJXL+DC6&3E4S"!/(F-)Q5 MJ2U79,#:>;:=\0!:5CF.<+=('S;2ZYTKDGA\'%MUW1G$&I+F'J%E-,XB-&+A M!MF8@_.[L^!HX@_3SBER?(52$P/6WG%2Q0U=1RO&H]X\*-^!Q"O@7EGNPL/@ M?CF?W7^:#;OSX69_@EALN:YK:AA@4B309^\ED?K#;FLG6?8ZI5W+N^2-5O#H MC$ZYN:^Q\DP79:!AO$1Z_4D:9(%J(>>Z$\.Q]=V.T98"5GWVY^4:=JG[?8'O MLC#4!WXX^O\>=:$SK(U*\6AI#;*TS'6-MM^EO5A)29.?AS\6'ZYE.G[L4S:A M6YP5@EP'*X-RZ&U=`VF)Z1H$H#\O70N3;*-LRC8=*=#?CVFFE7$%K-R33,?/ MS])F2-%;.54.NAX9(]H*X-='+YH%D?B3O!KNUEX3[">A99%35]5QGAN;,8T_ M-37*P,V;-W"1C..6?70>*B>;L+_]Y)AP^ M3CZSL64(]?7Y>=NVR>&N/?]7[M[SQNJG"QN2,E3F].0_4$L#!!0````(`!E0 MY!Y]5F@W>0P``*8?```&````4U4N0U!0Q5G_<]K(%?^9F[G_X86;R8D$8SMI MIG-.!DX2<=LDP\C2`GL66FYW@?C:W-_>]W970M@XN;8_-./8H'W[ MOGS>UUT=/_GVFX9:=]+5"C\T)L/@;TI6"Z3(D,^!?OXD=:W"R8M9:,BXPKT M@@%]!RW@&I\L5TPJ422:91WHLUFRSC71=;,E+[C2,M%"=@R;NHR2E7OFN+%/ M+$7MLS9LN5[ *I$H5:. /=:]RSYX/5$HD;.6L:J[U@LAE=&VGVQXAF8O&R9[_B8SXD0JB M,-'P_*C1(\4M#LY(%$7?:?]!8XQ_.?I/XC/MSHPJ!1PP!,WS^!)V>*U6B4+Y32(T^MQ*:,#!]1-\*<02"PK MAH:*C2U!5]WC'+&NHG! QDO"C-2+6%"0%Q7[>2UT4C/,0!4+F,L$#49!BM56 MVZ 8,[H&A=))GELW(G3TMV/QBLTFO14[2+OC`%(D5[!%)!$0=#?NO\[Y;P@? M+R@,GW=>G+;A>LWS#$Y/7ORY`]$$MCS/B2?:#%LA;VR:LD3FZ![8,$DIHIS@ M@/BT23UITKT0D"$$J0;"8I/DY$,Q`ZI?WRN06R6//N$/S- Z6#&YY$I5=H1, M:<$4<5U+@KB5:=69]YV;W@"8Y'S% W\8&2'?+[0JF6; M3*/9335EU(KT$[:*B!63-NG4K=)LV41'[>>EVULF$/Q*&:2:T##_CN[FEB,/ MV2I/4E9+Y9PAR*CQ#2N:;N->HIJ-YM=P%/MG%"[W#:;\L.ETS8 :QG7.J"#> MBC6D"Y;>@.'0C!9B"]ULDQ0I.JV&1;.,W(PGN9C#M?ADPN5/.Y0SIHRKJ>!4 M$-><1# 3QP]C2>AI0OLM+;7*>L 5E(&Q1;Q3L5R9&%\7&2KRCJMUDD/OZ5-X MUCFUR7%;JST,N<9#XYIM_?\2+-UQF#5UO;JSJ+'QN4 M7 #U12Z41A\O@4^T M%TF1Y90/@W$4AX#[EVJ.T/T3U4_Q&[QZ!4U#>09-^F(IS...90$]D3&W^I;I M0:*T>>ZU##WR?_GM-Y])[)O1: `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`PQ\Y7RU88<**KG5($ZI))IX77"MX%?KQ)!S^V(9>>#0XMW<_VFR@ M*V@E(,.=VBBWX,N:3F7-_@ID?]#;.\@D4XSN-=F&B[7:2X^#`$*CEPO%+"NH M))OC3P,L"+L+XZ1>14OMJ_M/L)7,#.D641B,WHZH]9J_TV 8^V&W%P?O?#N! M-1HEP3@*\&PPFH7]'D2LIBGD5J'^(7;=GAM&^/PQ\9]?^=&XO`R@T M,U9P5C:D&K=Z2[JS>1=8.7GY_M[/#^5DUGR5<@4DQN6@3WY/TSP[ &*]+I.<7*?@9OI<;%YQ8M( M)R\+BEN?N6ML$Z@+&FYW8>I. MR(=>V%0G.F)[TJ[/-FTX,3_F*&S6>Z'?C?WIT+^:]D;#:#3PX5]0>UB>.-^& MH\D82I;X\[@6^]G1KOWDX]6.(>+$K5<./>6CIW'"A- M4'?C5>+Z\,Z1].93+%POU"#0!>-*RM[O_!E!+`P04````" `I4.0>R>RF*BTX```` M:@``!@```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`IU9@IT9 M'H]T9E,<:7?H_F0'#!]#I,_3$>W$!-H*B@;S)9FW:)' _P[!U2+N$Q.X MLRFN>_Q>X+ $.[L=!UB''I>GN@0WK+G,BUT`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`7O8 M(?Q$6H7L83H`SPKS*&+<98A-ER#6Y0HO)Q `#.HI=7)0UM8!`G1>;YV,R"V' M\-A7HK@/HYV]9*?')O80S<^JC,$0\/()W@Z8H?TD+CP`T1=5&-L-F4ELJ TF MHCMM)L".T1+[A3881D,"/'H$%LFSP^KY MG#,0\?U9: MWX$5>U_))>U>(K4[AG7BBU:"1P0C+"A*?HT>)'C0FH:0G9J@54=RGVQ3>DOF MZ_@^&$'])(_P!;3YBPZ,C7\IQ88^@*9SOSP+8-^-E_K'&>D?$ 7&TY7\>5@? M:Q-PM=!T=*$@8O6I_,!*J$(+7Z]/[AQE:_DX,C_RE,)^I!(6V?F:O:7Z-/,A M=G'G9ZRB2YG!QRV+X*P#'//GVUX+94(^'/8FE0S:\MJ)-WGH:M]M2H9-XN/$ MY:I(MOF03;91%5)BJ?:0"@/D!0I0`1/*`64&PP?7!Y<3<0G3YZ :1_K)N,D6 MQVV)1SG1XD29$K9!W?9U7=+B75*+8UC'O'B(X-T#4ZN?Q(6G(/KB_1C[`60F ML:'[8P;UL)0\+L$,64-#/3N-_E#[1^C-MK^$3[[!*T_)1EX]2=JX;E'!8^^CU2* MU_R):JQ!&1@.@K9QI"14@#%!6):$NTD'D-HS")--^+D:W[/QO00SO#X%D#Z$ M(=JC1WKI\'#TI)%G*CZ%]K^B1J=V"']"E7,>;QOF&9[QV\Z(W+!#D"=0<<$# MP]P9X1VJ_@$'9/]C&]E8LP=Y(>*U7YAW%8-"12D47D75;Y'L=1U0$/9A)6YU M<6G4.0-EVJ!M&.T3Y7Z;X+<-!FR#02A3%&^#!<(AW)Q,A! _(I:=V3Z%MPT. M`"';( A[VR#HUL)B4H *&AT+=L+0A6'+J2'EWZ@J%\IQ"IE,9+3&[4-^0W-' M[8E6.BZGBRU:8=YL24V1$:#ON+BAOP*)J:N,(F]7\X>\'RF\Y^5;N*GY-QJO-T6WJ%`RG<0"BHS?W<5,!_TLMJ]5_Y[>I0O'>E6G> -))=Y8>% M2%(\8";!EER2LL#*>9?QR]9,^-M-W4<8E50EI M= N-LF#.LM LJ.G53\K9JZH$Q40"%_-;+,T7U)D)B9;?\OV0FY7SKZ(6W"_< M`:E>"TIOAHT9FCH'))ZQ-RC+A0'J+U+ZB]3^(ACX.G^1WE]D"!;A-I )%J6C M[C=0M!#@:7Q?OY-F\.U(8T!/2N17*D&[C 7A*Q.U4@V#=I*5Z'J(=QD6#@&_ M'?LDUKM%]65(Z5VI4D86W)\"CW2EC7<*\1,Z.:224=5-NXMHD9_.)OL(T-/6 M`D-!>S:6[;<;)/4Q*%M+^,P&/OFC;)R@A>10C'!21A3-4[,)AU6^'0:1/\\N M`&EY#)1,<9L!-@--NHCRRFI]%T5V2G"Y@?1D$$7=1JK%7G,-DG^NUI?J<.%U.1G0DIU_#.--ZN,Q]O3H']A@J$ M#V!T)1 U'2N!1I8Z&0PS@S PDPPS&*CL(*7^`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`[(QDKT7A7(5;NH/D%%)=X#"^VC0!3F[V&1Y_Q$KTY98IW(( M!^@X#"7RB<=W6YDMO++(-XB\SB1%D(S3(6/HU0DE3+;QBD9\"!MQ)HR4B>W[ M@Q>CM^^/3FS?(^V%D*'?`A+.EA?HWF_*^V(8NLE#T6+LK>$3\;U]AQ))"DBSW(2B!@KI1_&#$4:<23\*X]1.^0AJI.<0F M)8*5I!W\I E""61)2UD'6CH,W-\(#Y!F49H2V&3?076/&RO>JXY!N/G@-H&7 M9_01D7CMP7Y8]OD4U.Q1=1RM].8Q[$QOGH65V?D+4:9XQ0E(X3X-*%] 9G OX\_)1KYR&3:6ST$`F$0#9>.. M-RA7@+544%,N^3!(J(0ZE5#_.9V#-!L/UM. \L>.*J>0.I&UQ&_!S=#&/;@9 MVB!-Q>0*F'G)=!^ 9N9*$"VI3) XO8-R,C^G`$:J#]*4?KFW8!YK$U'HQ71/ M[]):NZ?;-7LOC-ZY`[I1R:JEE#E=J5:[]X*.C>\XR<9[+\1SKW>['@8Z4AD8H$VE!!5&&;A3A* $O9(5GR5 MV8VBMX"A#JONZ5AH#!:HL(?.$O@`->JD7EH/B.I$^4M8/^-GWH)U7)S=!@%[ M'FE>._8J)G S8'U0!YPR;]X\3LG_%92ITSY1-DL<^_#$9_Q;96#^3J*LK+/R M;XDR`'!_(76Q8Z<@*Q9@AZ_7(-K& M+;M X?2$`K8SZ#?+)1XVF:;S*LA$RC?V1CCPM9Z1:>[X&%X"97_*M?V)7=!= M).OH9>=I]B9I]O::#WJTWF[2VF!GE2TU'VP>\?7)O2:&>X?(\XZ_81/X4WJD!T">7M)B<:34"9,KUM[8_VV,&\+*PX&\E]]!K)X"QEV4?XE%ENU=9AA*\'EPW/#3T!YOK!U9\P^K"@` M#J#6#W9)_A#UQU+?Z"58`74G&8:5PR@?;>,'3D&6`*L%-ZSQ]'Y\X2WY6$Z! MR*W9-PK\M5]$**@RRT2T3J"U$:P3#NI&]<>BSI0UX,.2VQWB._-%ZJ+FXKT\ MPJPXKA6L'[F.2L#MVK%-?$GB;\A=((\Z$4I%@G1BBHUA-P1*.KQ.T9OX O2G/P1:D. M(5]"`F70C=-92*/FC9XG=T_8[;$7)TQ#(,B4L-C#A#(`1I4IB57[^E36]O.8 MP3/2-Z@P[<(4;@&_*WLWFE;17L!?<$2MR"MZL*_+5\.XJ4==@YJ<8%W35E7R M;5JT6-VNZ?@CE/BD5=/Y(88R=E;$1P:I6DW'81(%&Y9WB_HBP(R?@69O5^O_ M`@\)<\RN5D/=N>O:+Z)GD%6U7T0O()=8);PX75(9.=AT5M,70"?6M.62-+K7%T+31^_$L:MC$^OD=4R5<',22>G&4P`K M55VRT'2R8N'A@0I'E; [DM$<`@L3&.%":')*)C,DE*#&O1<1`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`%X=`1OL<4>G M218?3<U^N")M;PSM/#)4J:)V%NV(B-O@LV'F#@2%C-*ABDR/$#^2> M^(N_;'@H33'1BP!#2 ) >I[?+1TPXL/ '!5J@1WBT&>DYJ<2L&F.6$,N4M1< M6A1?=F:T'S3>+7Y52.Y715KS-S0+[LE&^[ULEO>8/'0#M)Z*C)_7XTAP2BLE M7S5@&S7 /"0V2=OKRO(NV]AHT,M>1TH:4$KM&/97CH>V2IR$/@`ZUV-D+S+P MG=%^*V*SR<*=.$ .]-(L/\$\)J$98;]&6&A>!#-)6#]!F:!-$THG`)(:"+CL M=>C'OTVW,];7&F::4T&_XLO&QT[Y3D 7OC\&`U3F+QL,J2D<0/ &S',JK^M] MT><:9' $M<9#Y(G8'N*# A!+(/I?)N:H,0B M"KN*MYWZR6J_[7E_Y3Z1>U[D!D7N?>$`NJ*X20#HJ,"V4(("NS5L0CXJ!]MB M^;+797#BP3;<7186RX:[5*%_";0^E=$Z+-J&1X-PS$Z+PM8V[,TU<"J8$YVH MF"\'3( -,(:AL[CF)X+4N059)57PV;0JD7L9E "?`!&!1S\GG2S^LJ<#94_[ M*Y_1[*\,K^:_=,!!Q9%'!^(_T_&I2QL3!$;P@EC_;3 M: 8<'NJ7LU,@JH,>)%8VV_&,8(;MS=%^M=]VBIM&]PG3-@ M&W4(OZ"PT7X5>_5HOS(*WR"LIVF^UD,,NP(L@C;HC, JD8PMZ,OW`V4C./U! M1H#9_(.IDEQH^VZ@[$RY$\XQC90+-U(2?!FIUF'N<"B)% `;H&%A4:1LI::# M`^MF%*L?$X5(.[0!P*MRK$Q\MI59#+<7[EZX7X=["L /PCT(]X+%]';"?<\2 M*_,)A"& %T,\>S$-'7!O@[L+[FJ '87P';@S\ZQP+ERFE6LG+P4^E-J8R'ML MG$JETL9C-&&*6JM-U$[5:C5X<@#,IB7^I@`FJ;+"23QCZ#1H MA1DF$-WEFS-03V<+F_M:[0[F@D ,>8G%:\&E?CJN)<&<# M"1T.Q=-V\]O.H4SG7+[6B&%DTYQ 4]RREL28PX0GL0 MQ,J6^P9TYI&63_@@F"UQF#BA(NM]@I(W@),7O#QP1B5&0&LR>?W*>UCF7ZGT M61;"/DYDY_:D06@J4;9C&XB7 M6M4N>:E5Q$N=["]*!8=TQ%V=[B]:&"RB+DCBHT9T]/L*&U"LM,'^#_8Z6O1V M.(1??43V>^ OV*%EN(58"[1#T9,_;!)XOK&!H!C%02@(_&A%U(_6ASK]03P' M!6=!R:P#CSW.-S8;'?>^?U8Q7!)_8YKI1EU+?'=LETQ P@H,/#/\[]!C9U##' M=7S_C5!>&U^F1N]G"_">)H)3`.XT$#82+G\$L&\$`7LJM 8S@:6M%--*=� M3084L&TB+TMX+IVW+5;!V00\ MGW"T#M;Y:,^L'5W(G)Y?N1 TUVO!MY=Q:$6YP^H0*E G7IF.6TL0.K +34,6 M0M/)4[NB&T\YK$SOD@&.9K\LXBNJ_BK*!;O@PR@7+ PJ/.F60PQZL[Y1&(+S M'LYTQ/@NRKF9=7)!'R\9O.%(22J-F\/L'*WBKIF,Q.2=>J)JLUM!\LF:,@'"6 MF/%J.A(HOUY3`J=4*T-R_9=@&>HBR?R7:+/L!VEL/J7QG48S>ES^3$W'`8CE M)[*JY5V*H9^3+>VY%>9C+6F!53(D,P5,E@.*A*&[,.D\X* [:CFI4*!,E%K% M`:<`+J_;A/0/$^E?ZFL=85C[*.&E!!P\K2*L3T YE(-K5+=-["*-%;K>M%[- M39E,AIE3!]U;0+U[" =K.>0PMY[3W&["_ITFU=_*@8>65C9CO=J_'HVT]'B+ MM!K\?>-Q!4RH9'*2--4I/! ;&75=TL<>XKAY_Q59^(JL**^_ M#@.NRLG#">%6%5^FA8.?:GY T6<.:VY',\!8$%VAW&K^A.]C!3&VBOH''#8K MLZL+'AUPLW!OLJ'/.':+/Z[;+N]:#O96U8:-A_MW49CTMHCO!1D=>=OE.Z]N MN99_9^P]()SQ&M_'OP86W$.^P4@IG5@*Y$&[;B3_5XJSBO,22QL)2W#DU?<7 M980Q`;)<9P<-!6XMW&'"EUV^Q;]N M0@[G7\F?[\\*WU\FN,0B=WW^>"1?3].C^/$&:IY*Z=@O+C&X>+6+S'%1;$%U M:@_QJ"I@J4CU_7'<^%F5,"AMI/&K,FN5T!<[<1XJN4KHCVSHN;?)EQZA8[# M3#?=\/INM%=7B2D8H0@O@/I1`FLZ40VDU3XCZ LKV9G@A#-(!W][05TIB6@+ MX)RW`QZ>9ICJ)[H`>??UDF\YI&$Y01(/"$\J$<$6G?"+(?IQ`44(E)J40-(O MXZ?[M?Y8Z?@6L2+C8>:J2EC/DJ!]=,()R9%QE)[)U4'-M17"C#'IT[6AO>@? MOSC19CCW?0,*_I_'?9^.\\N1A7Y?>!R4G[NAI>$T&>&JGQ?@RYMD\XBG",V3 M3C!/MC#2V2EVD9-X,?QD!4$59#4?YC_G?T>="+^K($NO\&-JG0[E\*N4?)]> MD2_C4'7RKU*"DS^Y>::S2A (ZU>THQ]/P4XNU42+=^+Z6,5G8ILBJ7SX^$!F M)GH6Z%NTN6/Y-B6,`UJ#H$S.$QN]O\"?0ZSRD;SL*[CV;=Q@DL-)':0-.P5> M"SL@[ W)^HWV^D"1K,*"._AM4WJ(=>[7RI)01??O^11JJ^L?+5X*IKJD?6AI MZ5H?AF]BF*ZR\^!0PS6AJPP`Y]&V7781'I7A!^`&1E#AYU*0<=QZ9VN'50K)@S^"@N!G=1&1E+OM+04*'[G8BH MQ;J]3LZ1\+;D*;U<"@A(4/.-)\?.\J?0&W*6?P,V`8?X]5J^$K=!)X9.G#[Q MIQ.GIASFSHU]SH^,N9+O!0O)V*>*$QZUN5*]+,!'/"TZ%I8VI&PTEP/.+D!*>Y\24 M7>2+U& L64FB=(++%BG:D5)9TR)EQ?-*4M;?R$C&M?'D/M7/\#PV)IE4[%SP M<"VUH&3UG/6=O\ISC>^\?MM,2@./T82R>.5;:LBQ!R$A??MYM/![IE$,W%RT MGT>S>4L"A4RG!VP.D#Y&X09N%RM7`*=_\YQ"XQM41PR#\X0:`O'0\;$O)(6Y M-8YD`Q>'/8WJP[?\E1[0MJO)WA)7U;IIPN:3U+59-T-8*T4KC+VX%)"YWWY1 M#T_6080FG'>6<6:^@-^@]/YAG*]2\J^!P,DXXAM2FF'_H/$]AD;7CQ5>.-R3 MS&O\4WGBIO-?[2>"`@YE+:;TN&?XE%0(S2E8!EL;\(!,E^B[^#Q^D>^3<91J M[T(Q&;_WA91F$,G-5@EC&9!.X&<3TM=0TF8+$52O^ E!/H%?C,)W!0KZC*,2 M>QXY?Y#:.09"8_PT6C;7$0I5_(8T,#F+&J. V3^?@:-LQL/[A6K<]((?:L,X M.)W3?1= ^3>UA3DY+.D#ZJ+VBT(-=/6'G2=WICI6"S/IHA5*`C/7?5*<;G>( M3:8'7:?.JL S(O@^VS_%(DS$B=H2I^C=@XFX^)K'FN&4#+I4=0P]\T+=J>QB M6)6T5?R1"CCVY&\?1*_JL\A?L/U,-1V7PNR)57-:A?!;9E*5HD>Y0N'("G^K MJ.;OQ&S\/4@BXQ#_,R2H.&4JU:=NV\G?8'B8',%,;7\5F2!<7<66/]U%$PPISWW/?142HE6QRP&J+J_0L*-N]&^!\3C' J\2W]W20 M+.PJ^#!U&HA@]]./XN=7T 7@_6)-9!':@V31F;X>IC_L^S@P2J!L!SU!'WV# M(1FI#;T`>=S//$KI=U'ZLR+T>PE]0CV.+@:/`F.W3Q8CX[8;3QM'HG#[OT33T4%6_A_((E*!#Y?"YX]#+Y-@'?+P[7NY&X@'EU:N= MA)HPG>:F9BY_#HK*H0>(EA!V]"!&B4/HI[,'E\$8P8BG?J%QJ,Y+"P+:&8<4 M[P!7_A3,XPC&8C%.+#RT%!4.D5 &`Z13>)R12D1(:"$1S+:P23/P0,P<'0O_3=*XIDERB.TG0$6 M#!(0=5(&9SD('O=[28B1(WAP] &-#C]4DL_C@MA);!D4231@.9>"^BEHGXH@ M[!/T%_$\@C)T`P)A!5%\B';((^QLF+UI8-#.A;V#7^;?"*O917"/0\2?Q[^S MT9\5V4YH-?\!\LBF]X];=;L,5)QX!EZ2M"WT"V1QD'B67- M"EMH9V5^4&1E7BH$_YC5RNR'O,L`Y@.8 MLFBR7"3QR/67O@>E]UC\)\.?2E+$,K%RK=)FL\7@E6-@'.3'DM)S'6OC5,P2 M@Z'*`99:G8XL#*I-Y *&FK;#I0(,)HX!OQERKJ+LX_4Y\/$:\-8!/-867E97X6)UENKT-%'BA> M47U]X^9JUJ4KA:!>5^;:VNC9(:5M]K@PIK:=3A[XRY MJUM<.M:-I10U-]?>TN#PU&ZM]NS0`3\5KA4-2*/95 M1,M'`LU0&6#L5K:Q">)21:/2JFH;:AJW5;#5;&UCPR0&'-*J;0"H85VQ:TLU M5\\RE0W5-]<#+XVZ^L9;&AN8HLW(M:ZV65?C:JAUU2#_59[&AEMT3=7-S=L: M/36+_G9]UFR^>8)C77HU=$1U,^3UL+K&+;I5%0MTT#)2,M9GLGR/"QEO\KA: M:ANY9MWFQH;F1DC!'V'#-FQ@H6LC/&#O1^=E.4\#T-^BLVUV-TK]< M0@VI[ M7 TLC!A*K(P0FTQW5S?4`!@XJUA3O&*E]!-N>&W>6L,4U6RM;:AM9CW5M/.; M.6#R!UPM<*JK9G68N)7;JENY1M?B\C1C0QL7Y63J;N9JZVMTF8:<)8M(GFVU M]?7D9^4:&^JAAQNDSFF&C(NBZH#\K*J(D$(>YM=SS'_E]:.95OH;=C63H4P6 MA:"],@_J^M$H7YY52_,&H M.'YR%%D/[XZ*WQ,5WQ45OSH&'[SP:)1=?:YB4A?B/2M^>]/;IOH-MS\=+PA?]- M?/SJY-AC9QZ;R7SRF/#835?@MTMXUC::HQP&7OE[&][3,1O?JWLOZTK\-TY] M, AG?3/;:$XK\P_Z7D*\>%59T8J5DYQ5K%BY/)K/-:6K*BJB M`/@`'8NMW>JB4!W50ZP/4]N"^V'JK+@?0L3OA1!'QAD(DW"4_(N5_+RL%<*K M\!MA"/&#QUT0SH#P%0BW0S@,(>K1XQ#^$/=]NZT,"K,T"''QTNVFXS"]@:NO M7\ P3\-[.M/ <$P]_"V@ZYV$\QZ$[Y&X[HJ_](F_*].N_"N!NORMOXZO^;L4 MHSWJ3_9W_D7GOSWJ3_YW_D7R0E=]NX;B_A9U\$>L$_IUP&']VEM&?V9Z0M=6 M2*&2FCM0?R).-;0]QTLZC5HR?Z@EF%K2Y'TODQZ=TAY5DOP"BFLDNA_7RJ[7H)OE?9?33B?T/TG\1HKX7LE MGMNE\'8IW2>E=TAT0-GP55OS%I44X_L,\M=; 7K@35ES8XF5Q'%N016]9\0C+JJI; $?[1NEI7) MFFV=2)L_W\KT54R^>W36OY_P__"KYIG)N.-7,"Z?!7DYAN)7/278\B)\$&J_ ;87X10@%N/?!;8#W.7 O`MS?P#T` M]P^?A[4%;AW$GP0<.X1KX=X"MP?NN^%^'.Y_AUL)>'?#_2;$/X7['-Q:>-?# M;8![*=QKGZ=\;(!R6N#N_M7_/%X_PI%?6MW,$D,1S%,9,9/9/2X7TT;CQ(#& M, 7RJNI:UM[HJ8!94N^BECLFB/F74=/-&J+*K*C!&;ZLOK'954),-LQ',K#G M268=:M7Y#]%14[BQPK(N7=S-A TY)R,,>Q/6R3 M"B@51BW+L/H47+'/#)%<96V,F,O$DC9I6\"IJ*E ARWU[/ M-;LQOI3;L@4J@&>>5T-'D60FC&,$HPY4\Y'GI?*JVAK7,G>U9TUC&6K*2W>P M+C#15UPR3A21WBVJJ?$0&VYHLJX4"6KZ(]D$A36-$;+,M]=_X:4%F0?*C]R@ M,B09KC;<:' :&@W;#7<:=AN>,#QO>,OP@>&/A@6969F.S,K,G9GMF<.9%5GW M906S/L\2LV8;KS,N-:XVUAH;C'<8`\8?&Y\Q'C8.&_]JU&:G9:=G?S<[.WM= M-I?MS[X[^Q?9SV:_FOU:]M0<7<["G,RG'\DC5+:I?L6O+X MDE>6O+UDR'3.),]+R$O*FYUW?5YFGCFO(F]#WN:\ICPN[U_S7LKKSWL][ZT\ M(>\O>>?R9.98?R_\R/\Z2 M9)EEF6_Q6(8L?[;$%,PIT!6C"^QY#Z4.Y8K,R69KC;--LTW+3)EF^RF M2I/;],^FNTT]IN=-0=-QTP>F"Z;8O*EY<_*NRRO,:\F[)^\JJ,U<\_7F+/-= MYOO,GYAOR-=8TBVW6'YHZ;#<:_D7RQ\L(Y8O+!8W= K36;6O-V\T^PU=YB[H$Z[S/>;'S(_E_]2_M'\W^=_ M"CT4;TFVS+'8+1LMKUF.6T8MXY:4`D-!?L'W"QH*MD/M'BEXMN O!6,%286S M"J\K-!1N+*1&(OP49D;6W*R,K+HL-NNG62]D_1;&Y$B6W#C=>+WQ!N,2&)S/YKMCKGQAQ'SIJQ08 MK_,,BPV;#*SAHR3F?] M-4MI/&0\:3QE_-1XUOB%49D].]L`LY1N!HIQWV)(->@-Z89-2@RE1G:C.3,U,STS)UF?K,],R%F89,:B N0=:,:XV; MC*=R4O/2\K+S\_.+\TOSU^2ORZ_)K\]G\W?F=^3?E7]__N[\Q_.?SS^4?SS_ MS?PS!2,%YPK"!>,%3*&RD'F489*!SHCAG"%L4&:JH"SFV^L??OTO4$L!`A0` M% ````@`:%#D'C:7T+N+! ``K @```H``````````0`@`````````%)%041- M12Y46%102P$"% `4````" `94.0>?59H-WD,``"F'P``!@`````````!`" ` M``"S! ``4U4N0U!04$L!`A0`% ````@`*5#D'LGLIBHM. ```&H```8````` I```````@````4!$``%-5+D5815!+!08``````P`#`* ```"A20`````` ` end - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From paul.toseland@virgin.net Sat May 17 03:17:00 1997 From: paul.toseland@virgin.net (Paul Toseland) Date: Sat, 17 May 1997 03:17:00 -0000 Subject: B18&17.1 problem reading large files (more than 28k!) with read() or fread() Message-ID: <337D6B74.705A@virgin.net> I stat a file (a Net$cape FAT.DB), using its DOS filename, and it tells me it is 692224 bytes long. I open it, fstat it, with the same result. I read from it, it reads 28720 bytes and gives end of file. The same result happens whether I use low-level I/O or streams, or whether I use the DOS filename or the UNIX filename. It didn't work with b17.1 either. Code: #include "stdio.h" #include "_String.h" #include "sys\stat.h" #include "fcntl.h" #include "unistd.h" void main(int argc, char **argv) { if(argc < 2) { printf("Invocation: import netscape_dir\nE.G. import c:\\netscape\\cache\n"); return 1; // error! }; String s1(argv[1]); s1 = argv[1]; // netscape cache dir s1 += "\\FAT.DB"; cout << s1 << "\n"; struct stat st; if(stat(s1, &st)) { perror(NULL); return 2; }; printf("stat file size: %i\n", st.st_size); char* pChar = new char[st.st_size]; cout << "Successfully allocated " << st.st_size << " bytes for file.\n"; int handle = open(s1, O_RDONLY); cout << "Handle: " << handle << ".\n"; //FILE* pFile = fopen(s1, "r"); // Get file size from handle fstat(handle, &st); printf("fstat file size: %i\n", st.st_size); int iRead; int iTotal = 0; char* pThisChar = pChar; while(iRead = read(handle, pThisChar, st.st_size)) { pThisChar+=iRead; iTotal += iRead; cout << iRead << " bytes read for a total of " << iTotal << ".\n"; }; cout << iRead << ".\n"; close(handle); cout << "File closed.\n"; /* while(iRead = fread(pChar, 1, st.st_size, pFile)) { cout << "Read " << iRead << ".\n"; }; cout << "EOF: " << feof(pFile) << ".\n"; cout << "Read " << iRead << ".\n"; cout << "EOF: " << feof(pFile) << ".\n"; fclose(pFile);*/ }; Results: H:\Matthew\CPP\CACHIMPT>a \c\netscape\cache \c\netscape\cache\FAT.DB stat file size: 692224 Successfully allocated 692224 bytes for file. Handle: 3. fstat file size: 692224 28720 bytes read for a total of 28720. 0. File closed. I have tried this for other FAT.DB's also: it returns different numbers with the same pattern - file size several hundred k, read less than 32k, do EOF. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Sat May 17 03:17:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Sat, 17 May 1997 03:17:00 -0000 Subject: compile fails with binary mount Message-ID: <01BC62C1.40C761F0@sos> Gary Bishop wrote: > I found the problem. I had edited specs and emacs had inserted ^M > characters all in it. > > Anyone know how to make emacs19.34.1 not put ^M on lines? >From just mentioned here NTEmacs FAQ: How can I control CR/LF translation (e.g., to access UNIX files via NFS)? There are currently two methods by which you can control automatic CR/LF translation in Emacs. The first method is by file extension using the alist file-name-buffer-file-type-alist. This alist contains pairs of regular expressions and booleans that determine whether a file is opened in binary (untranslated) or text (translated) mode. The alist already contains a number of pairs that you can use as examples for creating your own; do a "C-h v" on the variable name to see the help text and the default pairs. You could add your own pairs to the alist, or change the alist entirely. For example, if you wanted to open every file in binary mode except for batch files, you could add the following to your startup file: (setq file-name-buffer-file-type-alist '(("\\.bat$" . nil) (".*" . t))) The first pair says anything ending with the .bat extension should be opened in text mode, and the second pair opens everything else that doesn't match the first pair in binary mode. A second method for controlling translation is with the functions at the end of untranslate.el. These functions enable you to turn on and off translation on a filesystem basis (e.g., open anything on C: in translated mode, and anything on U: in untranslated mode). If you want to use these functions, download and place untranslate.el in your load-path and have it loaded from your startup file. Then you can place calls to add-untranslated-filesystem to selectively access filesystems in binary mode. -- Sergey Okhapkin Moscow, Russia Looking for a job - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From chuck@hazbro.com Sat May 17 09:37:00 1997 From: chuck@hazbro.com (TC Hazzard) Date: Sat, 17 May 1997 09:37:00 -0000 Subject: Access to NT 4.0 from UNIX system running X Message-ID: <337DDED3.AA0F06E0@hazbro.com> Hello: I would like to be able to access applications running on an NT 4.0 server from a BSD UNIX system running X. I would like to know if there is a good X Client for NT or something similar which might work. I had seen mention of a product called XWindows Connection by ConnectSoft but have not found any details or even if this product was ever released. Any other suggestions are gratefully welcome. Thanks. -- TC Hazzard chuck@hazbro.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From wfire@pimpz.darkillustrated.org Sat May 17 09:37:00 1997 From: wfire@pimpz.darkillustrated.org (Holister) Date: Sat, 17 May 1997 09:37:00 -0000 Subject: Segmentation violations with b18 (pdcurses) References: <97May15.164940pdt.35725-2@gateway.fluke.com> Message-ID: On Thu, 15 May 1997, Warren Jones wrote: > Program received signal SIGSEGV, Segmentation fault. > > 0x77f6cb4e in ?? () > (gdb) bt > #0 0x77f6cb4e in ?? () > #1 0x77f674c3 in ?? () > #2 0x2a92333 in _size_of_stack_reserve__ () > #3 0x2a92467 in _size_of_stack_reserve__ () > #4 0x2a95717 in _size_of_stack_reserve__ () > #5 0x77f6b42b in ?? () > #6 0x77f63a9e in ?? () > #7 0x77f1372d in ?? () > #8 0x77f13445 in ?? () > #9 0x77e786a0 in _libwsock32_a_iname () > #10 0x77f634c9 in ?? () > #11 0x77f61c3f in ?? () > #12 0x77f61187 in ?? () > just from glancing, and from my unix experience (I haven't used gnu-win32 much) it seems you need to compile with the libs statically...which I dunno if that's possible in gnu-win32 since DLL's are dynamic-only...try gcc -static...see if it works? - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From wfire@pimpz.darkillustrated.org Sat May 17 09:37:00 1997 From: wfire@pimpz.darkillustrated.org (Holister) Date: Sat, 17 May 1997 09:37:00 -0000 Subject: a few probs Message-ID: first thing, in gnu-win32 there is about 200 defines missing from errno.h, why? second, is there a way to write absolute memory addresses in gnu-win32, i.e. like movedata in DJGPP? - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From mike@mbsun.mlb.org Sat May 17 13:58:00 1997 From: mike@mbsun.mlb.org (Mike Bernson) Date: Sat, 17 May 1997 13:58:00 -0000 Subject: Non blocking I/O Message-ID: <337E1C01.1CFBAE39@mbsun.mlb.org> I am trying to set the the non-block read or write for a tcp-socket. I do the following fcntl(fd, SETFL, O_NONBLOCK) which does not return any errors. Then a do a read on the fd and it blocks. I need a way to do none blocking reads or find out how much data is pending so I can do the read for just that amount of data. The same would be nice for writes but not needed. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From wfire@pimpz.darkillustrated.org Sat May 17 19:22:00 1997 From: wfire@pimpz.darkillustrated.org (Holister) Date: Sat, 17 May 1997 19:22:00 -0000 Subject: Non blocking I/O References: <337E1C01.1CFBAE39@mbsun.mlb.org> Message-ID: On Sat, 17 May 1997, Mike Bernson wrote: > I am trying to set the the non-block read or write for a > tcp-socket. > > I do the following fcntl(fd, SETFL, O_NONBLOCK) which does not > return any errors. Then a do a read on the fd and it blocks. > > I need a way to do none blocking reads or find out how much data > is pending so I can do the read for just that amount of data. > > The same would be nice for writes but not needed. select() will tell you IF data is pending, which is good... fctl() doesn't work in linux for setting non_blocking on tcp sockets either...so don't feel bad :) - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From peterham@peterham.com Sat May 17 21:04:00 1997 From: peterham@peterham.com (Peter Ham) Date: Sat, 17 May 1997 21:04:00 -0000 Subject: kaffe 0.9 and gnuwin b18 don't mix Message-ID: <337E8BEC.77B5@peterham.com> The memory manager in kaffe-0.9 requires that the operating system memory allocator return "page" aligned pointers. Unfortunately, gnuwin b18 doesn't support memalign(), or valloc(), and malloc doesn't return data aligned to kaffe's preference. This causes a crash on gnuwin b18 immediately upon running HelloWorld with a stack trace similar to the one reported on AIX by Kevin B. Hendricks (kbhend@dogwood.tyler.wm.edu. Anybody have any suggestions on how to work around this? I can't think of any easy way. For now, I think that I will try kaffe-0.8.x if it's still available. Any hints would be appreciated. Thanks, Peter Ham - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From peterham@peterham.com Sat May 17 23:15:00 1997 From: peterham@peterham.com (Peter Ham) Date: Sat, 17 May 1997 23:15:00 -0000 Subject: ls starts VERY SLOWLY w/ b18 on Windows NT 4.0 Message-ID: <337E8D83.7A5C@peterham.com> I have noticed that program startup is MUCH SLOWER than the previous release that I downloaded in November or so. It takes 16 seconds on an empty directory on a lightly loaded Pentium 133mhz. Could it be that the cygwin.dll keeps getting loaded and unloaded for each program execution? Is there any way to lock it in? It's frustrating to use "ls" (or any other frequently used program) in this configuration. I also noticed that gcc takes a long time to compile programs, but that the performance monitor indicates that the machine is only lightly loaded in terms of cpu AND io. What's going on? It seems that all of the gnuwin programs are waiting (unnecessarily) on some sychronization or timing event to occur. Has anyone else noticed this? Peter Ham - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From fjh@cs.mu.OZ.AU Sun May 18 00:55:00 1997 From: fjh@cs.mu.OZ.AU (Fergus Henderson) Date: Sun, 18 May 1997 00:55:00 -0000 Subject: kaffe 0.9 and gnuwin b18 don't mix References: <337E8BEC.77B5@peterham.com> Message-ID: <199705180755.RAA21968@mundook.cs.mu.OZ.AU> Peter Ham, you wrote: > > The memory manager in kaffe-0.9 requires that the operating system > memory allocator return "page" aligned pointers. > > Unfortunately, gnuwin b18 doesn't support memalign(), or valloc(), and > malloc doesn't return data aligned to kaffe's > preference. Well, if you don't mind wasting a large amount of memory, try the following (msotly untested) code... /* ** round_up(amount, align) returns `amount' rounded up to the nearest ** alignment boundary. `align' must be a power of 2. */ #define round_up(amount, align) ((((amount) - 1) | ((align) - 1)) + 1) #define PAGESIZE 4096 #define memalign(size) round_up(malloc((size) + PAGESIZE), PAGESIZE) -- Fergus Henderson | "I have always known that the pursuit WWW: < http://www.cs.mu.oz.au/~fjh > | of excellence is a lethal habit" PGP: finger fjh@128.250.37.3 | -- the last words of T. S. Garp. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Sun May 18 01:26:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Sun, 18 May 1997 01:26:00 -0000 Subject: B18&17.1 problem reading large files (more than 28k!) with read() or fread() Message-ID: <01BC6385.D63578C0@sos> Paul Toseland wrote: > I stat a file (a Net$cape FAT.DB), using its DOS filename, and it tells > me it is 692224 bytes long. I open it, fstat it, with the same result. > I read from it, it reads 28720 bytes and gives end of file. The same > int handle = open(s1, O_RDONLY); int handle = open(s1, O_RDONLY | O_BINARY); -- Sergey Okhapkin Moscow, Russia Looking for a job - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From chuck@hazbro.com Sun May 18 04:18:00 1997 From: chuck@hazbro.com (TC Hazzard) Date: Sun, 18 May 1997 04:18:00 -0000 Subject: Access to NT 4.0 from UNIX system running X References: <199705180343.UAA26129@nz1.netzone.com> Message-ID: <337EE576.B24BBCC1@hazbro.com> Mikey wrote: > > http://tnt.microimages.com/www/html/freestuf/mix.htm I had seen the MI/X product before. Perhaps I am missing something, but is this not focused on accessing network applications FROM the Win95/NT machine, ie. an X Server for the Win95/NT environment? I am looking for the opposite, access TO the Win NT machine from a network computer, ie. an X Client on the Win NT machine. The intent would be to allow Windows apps to run where they belong...on a Windows OS. I would like access to them from time to time and am not really thrilled with the implementation of WABI. Thanks. ---- > From: TC Hazzard > To: gnu-win32 list > Date: Saturday, May 17, 1997 12:18 PM > Subject: Access to NT 4.0 from UNIX system running X > > Hello: I would like to be able to access applications running on an NT > 4.0 server from a BSD UNIX system running X. I would like to know if > there is a good X Client for NT or something similar which might work. > I > had seen mention of a product called XWindows Connection by > ConnectSoft > but have not found any details or even if this product was ever > released. Any other suggestions are gratefully welcome. Thanks. -- TC Hazzard chuck@hazbro.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From home@morla.crec.mipt.ru Sun May 18 06:18:00 1997 From: home@morla.crec.mipt.ru (Home user) Date: Sun, 18 May 1997 06:18:00 -0000 Subject: compiling and running ssh-1.2.20 with b18 References: <337D4051.C9B04DB@aa.net> Message-ID: <199705181309.RAA00304@morla.crec.mipt.ru> Joe Rumsey wrote: > I've actually got ssh working, albeit with some warnings from cygwin.dll > about various things, but I did have to work around a few things in the > ssh code: [xxx] > 4) It tries to read the password upon connecting from /dev/tty, but > never reads anything. Luckily it includes code to read from stdin > instead, which works just fine. Same problem. But I managed to trick it to work. The changes are in tio.c_lflag settings : one need to add tio.c_lflag |= ICANON | ISIG; before calling tcsetattr() for non-echoed input, otherwise it would hang. Works for me (for 3 days). Any ideas why it is not needed for Linux-based compilations, anyone? > There are probably some other quirks, I've barely tested it, but the 4 > above are what I had to fix to get a basic connection up and running. > > The version of ssh used is 1.2.20. The ssh home page is > http://www.cs.hut.fi/ssh > > Joe -- Igor V. Kovalenko mailto:iko@crec.mipt.ru - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From kdh5j@weyl.math.virginia.edu Sun May 18 07:57:00 1997 From: kdh5j@weyl.math.virginia.edu (Kirk Hilliard) Date: Sun, 18 May 1997 07:57:00 -0000 Subject: Bison for NT Message-ID: <199705181456.KAA29831@weyl.math.Virginia.EDU> A parser generated by the cygwin-32 byacc behaves differently than ones generated by bison or the various vendor supplied yaccs on 5 UNIX platforms (IBM AIX, HP HP-UX, SGI IRIX, DEC OSF1, and SUN SunOS). With my grammar, the byacc generated parser reduces a rule later than the other parsers, reporting a parse error instead of exiting via a YYACCEPT action. Has anyone here ported bison to the NT? If so, I would appreciate the details. When successful, I will post instructions and also put them up on a web site. While I would be happy to use the cygwin-32 package to build bison, I need to be able to compile the parsers it generates using MSVC. Kirk Hilliard - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - The discussion that follows is supplied for those from Missouri (geographically or philosophically) who, having read my message, would like a concrete example. My grammar is for a program which intercepts statements from an enhanced SQL language being sent to a data server, passes the standard SQL statements through, and acts on the enhanced statements. Here is a simplified grammar: ------------------------------ 8< ------------------------------ sql: set_command | update_command ; set_command: 'S' { YYACCEPT; } | 'S' 'A' 'X' ; update_command: 'U' | 'U' 'T' 'S' 'G' ; ------------------------------ 8< ------------------------------ (I have included an entire .y file using this grammar at the end of this message. In it, each rule has an additional action which prints out a message when it is reduced.) Here, my desire is to act on the special statement "S A X", but to pass through all other statements that start with 'S' and are followed by any token other than an 'A'. When the parser is generated by bison I get the desired behavior: SAX token: S token: A token: X reduced: set_command <-> 'S' 'A' 'X' reduced: sql <-> set_command yyparse returned 0 SAY token: S token: A token: Y parse error yyparse returned 1 SQ token: S token: Q reduced: set_command <-> 'S' YYACCEPT yyparse returned 0 When the parser is generated by the cygwin-32 byacc, however, I get this behavior: SQ token: S token: Q parse error yyparse returned 1 The yacc generated parser is checking if the 'Q' token can possibly fit anywhere in the grammar before reducing the 'S'. This could be considered a feature since it will catch some syntax errors before reducing rules and taking actions where the bison generated parser will reduce the rules and take actions before deciding that there is a syntax error. First the bison generated parser: UQ token: U token: Q reduced: update_command <-> 'U' reduced: sql <-> update_command parse error yyparse returned 1 And then the byacc generated parser: UQ token: U token: Q parse error yyparse returned 1 Note that the behavior of the bison generated parser is identical to those generated by the vendor supplied yacc on the 5 UNIX platforms mentioned above. (The only other NT yacc I have tried is PCYACC on a demo disk from ABRAXAS, and it generates a parser from my simplified grammar which behaves the same as bison, but it chokes on my actual grammar (too many ) and there is no indication that as a demo it is crippled in this respect. Even if it could handle my complete grammar, I can't see spending US$995 for a yacc!) Not a great deal is said about the use of YYACCEPT in the books on yacc that I have read. ORA's "lex and yacc" says, "Since the parser may have a one-token lookahead, the rule action containing the YYACCEPT may not be reduced until the parser has read another token." This makes sense (even bison does it) but I don't want the parser jumping to conclusions about a parse error before I get a chance for an early return. One solution would be for me to add rules for all 'S' commands from the standard grammar, but I would rather not have to do this. I realize that I don't really need to add the entire grammar for these commands, just the token following the 'S' such as: set_command: 'S' pass_through_set_clause { YYACCEPT; } | 'S' 'A' 'X' ; pass_through_set_clause: 'Q' | 'W' | 'E' ; Now both "S Q" and "S W J" will work, even on the byacc generated parser, but since bison and the other yaccs give me what I want, I would rather simply get bison running on the NT. Here is a complete .y file containing my simple grammar ------------------------------ 8< ------------------------------ %{ #include int yylex(); int yyerror( char *s ); %} %% sql: set_command { printf("reduced: sql <-> set_command\n"); } | update_command { printf("reduced: sql <-> update_command\n"); } ; set_command: 'S' { printf("reduced: set_command <-> 'S'\n"); printf(" YYACCEPT\n"); YYACCEPT; } | 'S' 'A' 'X' { printf("reduced: set_command <-> 'S' 'A' 'X'\n"); } ; update_command: 'U' { printf("reduced: update_command <-> 'U'\n"); } | 'U' 'T' 'S' 'G' { printf("reduced: update_command <-> 'U' 'T' 'S' 'G'\n"); } ; %% int yylex () { int c; while ( ( c = getchar() ) == ' ' || c == '\t' ) ; if ( c == '\n' ) { ungetc( c, stdin ); /* so main knows we made it to end of line */ return 0; } if ( c == EOF ) return 0; printf( "token: %c\n", c ); return c; } /* yylex */ int yyerror( char *s ) { printf ("%s\n", s); return 0; } /* yyerror */ int main() { int c; while (1) { if( ( c = getchar() ) == EOF ) break; ungetc( c, stdin ); printf( "yyparse returned %d\n", yyparse() ); while ( ( c = getchar() ) != '\n' && c != EOF ) ; /* in case yyparse() returned midline */ } return 0; } /* main */ ------------------------------ 8< ------------------------------ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Keetnet@wilmington.net Sun May 18 08:44:00 1997 From: Keetnet@wilmington.net (Keet) Date: Sun, 18 May 1997 08:44:00 -0000 Subject: Fb5.55 - Permission Denied Message-ID: <2.2.32.19970518154329.0068d7c0@wilmington.net> I've succesfully compiled TinyMUCK FuzzBall 5.55 under Win95 using GNU B18. Thanks to Cygnus for all the wonderful work they did on the compiler. I've noted one difference in moving from B17.1, to B18 though. When running the program, it dumps a BINARY mode database every so often (for which i had to mount the dir to binary, and not text.. i'm hoping to fix it so that the mount isn't needed, but i'm just beginning to work with C, and am still a newbie). The difference is that now I get a - '/data/minimal.db: Permission Denied' when trying to dump the database for the second time. This never happened in B17.1, and I was wondering if there was something I missed. System summary is as follows: P90 Running Win95, 24 MB of ram, around 400 mb of free hd space, and B18. Any help would be greatly appreciated. - Greg Neujahr (Keet) - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From RoryORice@worldnet.att.net Sun May 18 11:59:00 1997 From: RoryORice@worldnet.att.net (Rory Rice) Date: Sun, 18 May 1997 11:59:00 -0000 Subject: How do i get the c compiler to work? Message-ID: <337F502F.4E82@worldnet.att.net> I can figure out how to get the c compiler to work, its the only reason i downloaded this, the pages say it has it, but i can't find it, can anyone help me out? -Shawm Rice - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From yruan@midway.uchicago.edu Sun May 18 12:43:00 1997 From: yruan@midway.uchicago.edu (Yuxin Ruan) Date: Sun, 18 May 1997 12:43:00 -0000 Subject: Access to NT 4.0 from UNIX system running X References: <337EE576.B24BBCC1@hazbro.com> Message-ID: What you are discribing is WinFrame. They use a modified version of NT(I believe the 4.0 version just came out of the door) over ICA protocal. It won't talk with X directly, but clients are widely available. (I think Sun had a version). If you just want to stick with X, then there are a whole bunch of products, win-center, win-dd, etc that translate GDI to X calls. Yuxin On Sun, 18 May 1997, TC Hazzard wrote: > Mikey wrote: > > > > http://tnt.microimages.com/www/html/freestuf/mix.htm > > I had seen the MI/X product before. Perhaps I am missing something, but > is this not focused on accessing network applications FROM the Win95/NT > machine, ie. an X Server for the Win95/NT environment? I am looking for > the opposite, access TO the Win NT machine from a network computer, ie. > an X Client on the Win NT machine. The intent would be to allow Windows > apps to run where they belong...on a Windows OS. I would like access to > them from time to time and am not really thrilled with the > implementation of WABI. Thanks. > > ---- > > From: TC Hazzard > > To: gnu-win32 list > > Date: Saturday, May 17, 1997 12:18 PM > > Subject: Access to NT 4.0 from UNIX system running X > > > > Hello: I would like to be able to access applications running on an NT > > 4.0 server from a BSD UNIX system running X. I would like to know if > > there is a good X Client for NT or something similar which might work. > > I > > had seen mention of a product called XWindows Connection by > > ConnectSoft > > but have not found any details or even if this product was ever > > released. Any other suggestions are gratefully welcome. Thanks. > > -- > TC Hazzard > chuck@hazbro.com > - > For help on using this list (especially unsubscribing), send a message to > "gnu-win32-request@cygnus.com" with one line of text: "help". > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From RoryORice@worldnet.att.net Sun May 18 13:27:00 1997 From: RoryORice@worldnet.att.net (Rory Rice) Date: Sun, 18 May 1997 13:27:00 -0000 Subject: How do i get teh C Compiler to work? Message-ID: <337F50BD.5369@worldnet.att.net> I can figure out how to get the c compiler to work, its the only reason i downloaded this, the pages say it has it, but i can't find it, can anyone help me out? -Shawm Rice - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jp@awesome.nuancecom.com Sun May 18 13:27:00 1997 From: jp@awesome.nuancecom.com (J. P. Shipherd) Date: Sun, 18 May 1997 13:27:00 -0000 Subject: cp -p Message-ID: <199705182025.NAA04463@nuancecom.com> > Just run bash in security context of another user. Is it possible to be logged into NT as user a, but to start a bash shell as user b? How is the current user communictated to bash. (Or is the only way to do this to su as another user once in the shell)? Thanks, --jp - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jp@awesome.nuancecom.com Sun May 18 13:27:00 1997 From: jp@awesome.nuancecom.com (J. P. Shipherd) Date: Sun, 18 May 1997 13:27:00 -0000 Subject: cp -p Message-ID: <199705182025.NAA04454@nuancecom.com> > Just run bash in security context of another user. Is it possible to be logged into NT as user a, but to start a bash shell as user b? How is the current user communictated to bash. (Or is the only way to do this to su as another user once in the shell)? Thanks, --jp - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From latexduck@geocities.com Sun May 18 13:59:00 1997 From: latexduck@geocities.com (Justin) Date: Sun, 18 May 1997 13:59:00 -0000 Subject: Errors ; Message-ID: <199705181818.LAA08609@omail.eee.org> i was trying to port a program.... and.. i get this. gcc -g -O2 -D_VERSION_=\"BitchX\" -I../include -c alias.c In file included from ../include/irc.h:122, from alias.c:11: ../include/struct.h:524: syntax error before `HANDLE' gcc: Internal compiler error: program cc1 got fatal signal 33 So... naturally.... i look under line 524... to see whats up... and all it is....is.. HANDLE hStdin,hStdout; that... so what could possible be wrong? .x justin. & suicide@dazedandconfused.com x. .x[ * lick * ]x. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From gpasa@swissonline.ch Sun May 18 13:59:00 1997 From: gpasa@swissonline.ch (Pasa Guglielmo) Date: Sun, 18 May 1997 13:59:00 -0000 Subject: malloc.h Message-ID: <337F7C62.356B@swissonline.ch> Hi, While trying to compile some code (sesh of the HieroTeX pack) I've got some compaints about malloc.h. What is the procedure to follow to get things compiled, since there is no malloc.h in cygwin32 ? -- Sincerely, Pasa Guglielmo ------------------------------------------------------------------ & e-mail : gpasa@swissonline.ch & tel.: +41 (0)24 485 50 40 & & mailing : Pasa Guglielmo & fax : +41 (0)24 485 50 44 & & Rte des Cases 17A & & & CH-1890 St-Maurice & & & (Switzerland) & & ------------------------------------------------------------------ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From stephenc@wf.net Sun May 18 13:59:00 1997 From: stephenc@wf.net (Stephen Crowley) Date: Sun, 18 May 1997 13:59:00 -0000 Subject: what went wrong? Message-ID: <199705182114.QAA13995@odin.wf.net> Just wondering but what did you guys do to screw up b18 so bad? The configure scripts that worked perfeclty with 17.1 wont work at all with b18, sometimes the script is supposed to wait for use input but it just keeps going and then it wont find files it creates in the /tmp dir, e.g. it creates a temp file to pass to gcc but gcc cant find the file, and also it goes into infinite recursion for some reason, keeps saying stack error or something. --Stephen Crowley - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From tage.westlund@stockholm.mail.telia.com Sun May 18 14:57:00 1997 From: tage.westlund@stockholm.mail.telia.com (Tage Westlund) Date: Sun, 18 May 1997 14:57:00 -0000 Subject: NATIONAL CHARACTERS IN b18 Message-ID: <337F2955.2421@stockholm.mail.telia.com> To Geoffrey Noer! In the mailing list I read that b18 was to be released and I was happy. Now is time to widen the experience, to import the source, not only the binaries! But, recently I have found that swedish, danish and german users have complained about the national charac- ters, typed by using the Alt Gr shift key. They simply do not work, they say. Do we have to wait for a soon coming b18.1 version or for a small national characters' patch? For us this is a real obstacle because those letters are as frequent in our languages as are the characters of the english language. Thank you for your important development! Tage - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From kdh5j@weyl.math.virginia.edu Sun May 18 15:33:00 1997 From: kdh5j@weyl.math.virginia.edu (Kirk Hilliard) Date: Sun, 18 May 1997 15:33:00 -0000 Subject: Bison for NT Message-ID: <199705182233.SAA65118@weyl.math.Virginia.EDU> Earlier today I wrote: > Has anyone here ported bison to the NT? If so, I would appreciate > the details. When successful, I will post instructions and also put > them up on a web site. > > While I would be happy to use the cygwin-32 package to build bison, > I need to be able to compile the parsers it generates using MSVC. I reinstalled gnu-win32 b18 and this time bison configured and compiled right out of the box. I ran `configure' from bash (and ignored `configure.bat' this time) and the ran `make'. The parser code that the resulting bison produced compiles without a hitch under gcc, but I had a little trouble compiling it under MSVC until I realized that I needed to #define MSDOS myself. The initial C code of my yacc file now contains: #ifdef _MSC_VER #define MSDOS 1 /* NT bison needs this */ #endif The resulting parser treats my grammer in the manner that I desired. Thanks to all who replied, Kirk Hilliard P.S. I would still love to hear from anyone who has an opinion on whether my grammer is simply taking advantage of a loophole in bison and the 5 UNIX yaccs. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From alex!@bigfoot.com Sun May 18 15:56:00 1997 From: alex!@bigfoot.com (Hmmmz) Date: Sun, 18 May 1997 15:56:00 -0000 Subject: Beta 18 quirks Message-ID: <199705182256.PAA25706@mom.hooked.net> Well I finally cleared up a 160mb partition on my hard drive for beta 18. I downloaded cdk.exe and installed fine. However if i type uname -a i get results like the following: C:\home\alex\Desktop>uname -a CYGWIN32/95 MEGA.HOOKED.NET 4.0 17.5 i6364486 Is that normal? I haven't found any b17.1 dlls around either. *shrug* BTW, what's an i6364486? I then compiled ncurses1.9.9e (From ftp.probe.net) and it installed, however unlike advertised install.exe did _not_ find the clear.exe tic.exe, etc, etc. Again is this a bug? Disregarding that I renamed the exes, and manually ran make install in each of the subdirs, which then proceded to install without a hitch. After setting my paths correctly and renaming the executables, I ran clear.exe. However this program doesn't seem to work correctly in b18, whereas in b17.1 it cleared the sreen if my TERM variable was set to pcansi, mono, etc, etc. Any ideas? Ls, and almost everything else seem _very_ slow compared to b17.1. Oh well. A few suggestions I've noticed that if i have a directory say "/" (C:\unix) and then mount C:\blah as "/local" /local won't show up at all under ls's and tab key stuff. I think it'd be really nice if that could be fixed. However another idea might be to add a /dev/* directory, so that /dev/lpt0 or /dev/sio[0,1] could be accessed. And perhaps /dev/tty* could be mapped to the standard console. Also what about using the method that is used for case sensitive filenames for reserved filenames? so that if perl creates a file called aux its written to the partition as _aux or similar? P.S. yes I'm using binary mounts, and non-posix paths crashed tar (i.e. tar -xzkvf c:\cdk-src.tar.gz). WinNT and *nix rule, so learn from the past: Hiroshima '45, Vietnam '64, Tjernobyl '86, MacOS '96 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From martin@ibcinc.com Sun May 18 20:27:00 1997 From: martin@ibcinc.com (Martin Lira) Date: Sun, 18 May 1997 20:27:00 -0000 Subject: problem running X11R6.3 client programs Message-ID: <25448.864012404@eeyore.ibcinc.com> Could someone tell me what am I missing, I compiled X11R6.3 with b18. The programs that compiled cannot be run. I keep getting the following message. Error starting program Can not run c:\usr\X11R6.3\bin\bitmap.exe. Thanks in advance. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From colin@bird.fu.is.saga-u.ac.jp Sun May 18 20:51:00 1997 From: colin@bird.fu.is.saga-u.ac.jp (Colin Peters) Date: Sun, 18 May 1997 20:51:00 -0000 Subject: b18 "patch" assertion fails Message-ID: <01BC6453.137F3C20@gbird0> Gary Bishop[SMTP:gb@cs.unc.edu] wrote: >patch dies with > >assertion "hunk" failed: file "/home/noer/src/beta18/cdk/patch/patch.c", >line 339 > >when I try to apply Sergey's patches for X11R6.3. It has successfully >applied several patches before failing. > >Anyone else seen this? I saw this error, among others, when trying to get a diff I had made to properly patch some files I have. I also saw patch become strangely confused on a couple of hunks and output a lot of garbage into the "patched" file. I heard that patch was broken in b17.1, and it seems to me that it is still broken in b18... Maybe you could try rebuilding patch from the GNU sources... Colin. -- Colin Peters - colin@bird.fu.is.saga-u.ac.jp -- Saga University Dept. of Information Science -- http://www.fu.is.saga-u.ac.jp/~colin/index.html -- http://www.geocities.com/Tokyo/Towers/6162/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From dbaker@hobbes.cuckoo.com Sun May 18 21:39:00 1997 From: dbaker@hobbes.cuckoo.com (Daniel Baker) Date: Sun, 18 May 1997 21:39:00 -0000 Subject: SSH 1.2.20 Message-ID: <199705190438.XAA15491@hobbes.cuckoo.com> I saw a couple days ago someone email the modifications required to compile ssh 1.2.20, but I've lost it. I also notice that the Mailing List Archive with hypermail on www.cygnus.com dosen't seem to be working anymore... Can someone email me the diffs or patch or just text for explaning what needs to be modified to get it compiled? Thanks Daniel -- Daniel Baker (dbaker@cuckoo.com) Chief Technology and Executive Officer -- CuckooNet! ( http://www.cuckoo.com ) - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From weiku@UTKUX1.UTK.EDU Sun May 18 21:51:00 1997 From: weiku@UTKUX1.UTK.EDU (Wei Ku) Date: Sun, 18 May 1997 21:51:00 -0000 Subject: How do i get the c compiler to work? Message-ID: <199705190451.AAA45401@solid.ssd.ornl.gov> Did you install the CDK.exe ? If so, it will be in //[drive]/gnuwin32/b18/H-i386-cygwin32/bin. Sincerely, Wei Ku *************************************** Department of Physics and Astronomy The University of Tennessee 1408 Circle Drive Knoxville, Tennessee 37996-1200 weiku@utkux.utcc.utk.edu --------------------------------------- Solid State Division Oak Ridge National Laboratory P.O.Box 2008 Oak Ridge, TN 37831-6032 Phone: (423) 574-5795 Fax: (423) 574-4143 weiku@solid.ssd.ornl.gov *************************************** ---- From: Rory Rice To: Cygnus GNU-win32 mailing list Cc: NOYB@cygnus.com Date: Sunday, May 18, 1997 7:09 PM Subject: How do i get the c compiler to work? I can figure out how to get the c compiler to work, its the only reason i downloaded this, the pages say it has it, but i can't find it, can anyone help me out? -Shawm Rice - For help on using this list (especially unsubscribing), send a message to " gnu-win32-request@cygnus.com " with one line of text: "help". From colin@bird.fu.is.saga-u.ac.jp Sun May 18 22:07:00 1997 From: colin@bird.fu.is.saga-u.ac.jp (Colin Peters) Date: Sun, 18 May 1997 22:07:00 -0000 Subject: thread-local storage (was: BAD NEWS - Apache ...) Message-ID: <01BC645D.053F4560@gbird0> Fergus Henderson[SMTP:fjh@cs.mu.oz.au] wrote: >Is > > #define THREADLOCAL __attribute__((section(".tls"))) > int foo THREADLOCAL = 42; > >enough, or does the compiler need to generate special code for accessing >thread-local storage? >From the section of "Advanced Windows" (Jeffery Richter, Microsoft Press) on static thread local storage: ---Quote--- To actually make static TLS work, the operating system needs to get involved. When your application is loaded into memory, the system looks for the .tls section in your EXE file and dynamically allocates a block of memory large enough to hold all the static TLS variables. Every time the code in your application refers to one of these variables, the reference resolves to a memory location contained in the allocated block of memory. As a result, the compiler must generate additional code in order to reference the static TLS variables, which makes your application both larger in size and slower to execute. On an x86 CPU, three additional machine instructions are generated for every reference to a static TLS variable. ---Endquote--- Which sounds to me like the compiler has to do some special stuff (and that makes sense to me). I really don't know exactly what that stuff might be... presumably accessing some special thingamajig that makes sure the reference is made to the appropriate block given the current thread (perhaps part of an internal thread context structure?). Anyway, like I said, I don't think that is up and running in the GCC compiler yet (though I would love to be shown wrong). Colin. -- Colin Peters - colin@bird.fu.is.saga-u.ac.jp -- Saga University Dept. of Information Science -- http://www.fu.is.saga-u.ac.jp/~colin/index.html -- http://www.geocities.com/Tokyo/Towers/6162/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Mon May 19 00:47:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Mon, 19 May 1997 00:47:00 -0000 Subject: problem running X11R6.3 client programs Message-ID: <01BC644A.B09F2AF0@gater.krystalbank.msk.ru> Martin Lira wrote: > Could someone tell me what am I missing, I compiled X11R6.3 with > b18. The programs that compiled cannot be run. I keep getting the following > message. > > Error starting program > Can not run c:\usr\X11R6.3\bin\bitmap.exe. > Is c:\usr\X11R6.3\bin in your PATH? -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From chris@netman.se Mon May 19 04:38:00 1997 From: chris@netman.se (Christopher Arnold) Date: Mon, 19 May 1997 04:38:00 -0000 Subject: NATIONAL CHARACTERS IN b18 References: <337F2955.2421@stockholm.mail.telia.com> Message-ID: <97May19.111458gmt+0100.17026@ariadne.netman.se> On Sun, 18 May 1997, Tage Westlund wrote: > But, recently I have found that swedish, danish and > german users have complained about the national charac- > ters, typed by using the Alt Gr shift key. They simply > do not work, they say. > For us this is a real obstacle because those letters > are as frequent in our languages as are the characters > of the english language. Tage missed to spell out the main obstackle with national characters... It's not only a matter that we can't write "rdksmvrges", shrimp sandwich that is... But more serious is that we can't type in "|$" or even "{[]}"! (That is pipe, dollar, curly and hard braces) So us international users are quite grounded... /chris@netman.se - unix without pipe is like "rdksmvrges" without shrimp! - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From rafaeldp@caymasa.es Mon May 19 04:38:00 1997 From: rafaeldp@caymasa.es (Rafael Delgado Peña) Date: Mon, 19 May 1997 04:38:00 -0000 Subject: Library question Message-ID: <199705191138.EAA02821@cygnus.com> I'have got a .LIB library, Can I use it with the b18? What should I do now? I don't have the source code of the library. Thanks. Rafael Delgado - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From root@jacob.remcomp.fr Mon May 19 06:22:00 1997 From: root@jacob.remcomp.fr (root) Date: Mon, 19 May 1997 06:22:00 -0000 Subject: new interfaces for lcc-win32 useful for gcc users Message-ID: New interfaces have been added to lcc-win32: o spooler support (winspool.lib, import library, and the corresponding header file) o direct draw support. (ddraw.lib, and ddraw.h) Since lcc is a C compiler only the C support for the OLE C++ classes is provided. o OpenGL support. You will need the header files of openGL, available either from Microsoft or from Silicon Graphics. The library glaux.lib from Microsoft is necessary too, you can download it from Microsoft's web site. o inet.dll support. This very high level internet interface allows you to retrieve any file from the internet with a couple of function calls. It supports ftp/http/gopher protocols. Header files and import libraries are provided. o metafile support (enhanced metafiles) has been revised and many problems therein corrected. For GNU-CYGNUS users: You can surely use the header files. I do not know if the import libraries will work with cygnus-gcc, since I do not have that system and I haven't tested it. If it doesn't work you will have to build your own libraries according to the most current theory about dll's held by cygnus... For GNU_MINGW32 users, things should be easier. If something doesn't work just drop me a message and I will try to arrange with Colins so that we fix the problems as they appear. lcc-win32 is available at http://www.remcomp.com/lcc-win32 The new version should be available tuesday morning. My site is updated once a day, not in real time. Remember: NO ROYALTIES (expressed or implied) are due for your use of lcc-win32. This is the spirit of free software. -- Jacob Navia Logiciels/Informatique 41 rue Maurice Ravel Tel 01 48.23.51.44 93430 Villetaneuse Fax 01 48.23.95.39 France - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From gb@cs.unc.edu Mon May 19 07:10:00 1997 From: gb@cs.unc.edu (Gary Bishop) Date: Mon, 19 May 1997 07:10:00 -0000 Subject: X11R6.3 nolonger works under B18 Message-ID: <3.0.1.32.19970519100433.00785b18@mail.cs.unc.edu> I still can't get the client programs to work, I'm sure there is some simple secret here that I am overlooking. They work for Sergey so it is known to be possible, but there is some subtle difference in his environment that is making the difference. I will carefully describe my setup. Please see if you can see what I am doing wrong. I'm setup "binary only" as Sergey suggested: bash$ mount Device Directory Type Flags \\.\tape1: /dev/st1 native no-mixed,text!=binary \\.\tape0: /dev/st0 native no-mixed,text!=binary \\.\b: /dev/fd1 native no-mixed,text!=binary \\.\a: /dev/fd0 native no-mixed,text!=binary d: / native no-mixed,text=binary I unpacked and patched the X11R6.3 xc-1.c distribution in binary mode into /usr/X11R6.3/src/xc. I added fixup.o to my "specs" file as the *endfile for dlls. The line suggested by Mikey that I am using is: *endfile: %(dll:/lib/fixup.o) The source to /lib/fixup.o is: #ifdef __CYGWIN32__ asm (".section .idata$3\n" ".long 0,0,0,0,0,0,0,0"); #endif Everything builds just fine. But when I try to run one of the clients, I get a popup from windows95 saying "Error starting program" unable to start /usr/X11R6.3/bin/xrefresh.exe. And in bash I see: bash$ xrefresh sh.exe: /USR/X11R6.3/BIN/xrefresh.exe: (null) bash$ I've been very careful to make sure everything is binary mode (this was the source of some early problems). I even used "grep" to search for carriage returns in the source files. If I make static libraries and link with them, the clients work, so the problem has to be in the dll startup stuff. It appears to me (with my very limited understanding so far) that all the dll stuff that Sergey is doing is correct. And it works for him! Anyone see what I am doing wrong? thanks gb - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From mansur@tm.net.my Mon May 19 09:06:00 1997 From: mansur@tm.net.my (Mansuriatus Shahrir Amir) Date: Mon, 19 May 1997 09:06:00 -0000 Subject: Building GNAT-3.09 on NT using GCc-Win32 b18 Message-ID: <3.0.1.32.19970519154716.0092c5b0@mail.tm.net.my> As of now, the binary built of GNAT NT-3.09 on cs.nyu.edu is incompatible with the binary releases of GCC-Win32 of Cygnus i.e. I haven't got these two to coexist peacefully. GNAT 3.09 on NT expect to be installed on [drive]:/usr on whatever drive you install it on. Whereas GCCWin32 can be installed on other directory mounted as / using the mount command and the registry. And even as I made the GCCWin32 directory mimicks that of GNAT distribution, I still couldn't get it to work. Whereas Binary distribution of GNAT on LinuxElf can be installed on top of the distributed gcc 2.7.2.1 by only making 2 soft links. So I'm trying to rebuilt GNAT 3.09 on NT using GCC-Win32 b18 (the latest). BTW 3.09 GNAT binary was built on 17.1. I have downloaded cdk-src for b18 and gnat-3.09-src. However the problem is on the GNAT src distribution there are only readme's for OS/2 and Linux and No readme and instruction on how to do it on NT. Furthermore there are diff files for gcc-2.7.2 in the GNAT distribution. Is it wise to apply that to GCC-Win32 cdk-src distributions? (My hunch is NO) BTW I have lost the GCC-Win32 mailing list messages on how to built gcc. If ne1 can resend me on instructions how to do it, will be much appreciated. So ne1 can offer instructions and pointers if this is possible? I would really be indebted. Also if ne1 on the GNAT chat reading this, is it possible to build dll's using GNAT on NT? Thank you. ------------------------------ Mansuriatus Shahrir Amir mansur@tm.net.my mansur@pc.jaring.my - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From huott@pinebush.com Mon May 19 09:06:00 1997 From: huott@pinebush.com (Ed Huott) Date: Mon, 19 May 1997 09:06:00 -0000 Subject: Access to NT 4.0 from UNIX system running X References: <337DDED3.AA0F06E0@hazbro.com> Message-ID: <33807A0E.473F28B9@pinebush.com> TC Hazzard wrote: > > Hello: I would like to be able to access applications running on an NT > 4.0 server from a BSD UNIX system running X. I would like to know if > there is a good X Client for NT or something similar which might work. I > had seen mention of a product called XWindows Connection by ConnectSoft > but have not found any details or even if this product was ever > released. Any other suggestions are gratefully welcome. Thanks. There is a fairly nice product called "NTrigue" from Insignia Solutions that might give you the kind of functionality you're looking for. NTrigue runs on an NT server and provides access to separate, virtualized Windows sessions to any X (or ICA (Citrix)) clients that connect to it. The only real drawback I saw to the product was that it's a little pricey. I minimum setup with a 5 user license was going to cost us in the neighborhood of $500 per user. For more information, check out http://www.insignia.com/NTRIGUE/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Mon May 19 09:06:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Mon, 19 May 1997 09:06:00 -0000 Subject: NATIONAL CHARACTERS IN b18 Message-ID: <01BC6449.2121BA60@gater.krystalbank.msk.ru> Tage Westlund wrote: > In the mailing list I read that b18 was to be released > and I was happy. Now is time to widen the experience, > to import the source, not only the binaries! > But, recently I have found that swedish, danish and > german users have complained about the national charac- > ters, typed by using the Alt Gr shift key. They simply > do not work, they say. > Do we have to wait for a soon coming b18.1 version or > for a small national characters' patch? Here is a fragment of winsup/console.cc: else /* keep it */ { if (input_rec.Event.KeyEvent.dwControlKeyState & > (LEFT_ALT_PRESSED | RIGHT_ALT_PRESSED)) { char tmp[2]; tmp[0] = tolower(input_rec.Event.KeyEvent.AsciiChar); tmp[1] = 0; buf[copied_chars++] = '\033'; undo_input(hndl, tmp); } else Modify marked line according to your needs. -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From noer@cygnus.com Mon May 19 10:52:00 1997 From: noer@cygnus.com (Geoffrey Noer) Date: Mon, 19 May 1997 10:52:00 -0000 Subject: what went wrong? References: <199705182114.QAA13995@odin.wf.net> Message-ID: <199705191749.KAA28822@cirdan.cygnus.com> Stephen Crowley wrote: > > Just wondering but what did you guys do to screw up b18 so bad? The > configure scripts that worked perfeclty with 17.1 wont work at all with > b18, sometimes the script is supposed to wait for use input but it just > keeps going and then it wont find files it creates in the /tmp dir, e.g. it > creates a temp file to pass to gcc but gcc cant find the file, and also it > goes into infinite recursion for some reason, keeps saying stack error or > something. I haven't seen problems like this when I've run configure scripts using b18. It is true that if you run out of stack space, gcc becomes rather unhappy. :-( I also haven't seen a slowing in starting programs or in compiling executables. In fact, I see a speed-up in compile times over beta 17... -- Geoffrey Noer noer@cygnus.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From condict@opengroup.org Mon May 19 11:35:00 1997 From: condict@opengroup.org (Michael Condict) Date: Mon, 19 May 1997 11:35:00 -0000 Subject: Fix to readline for case-insensitive file-name completion Message-ID: <9705191749.AA22998@hare.osf.org> Geoffrey, I found the case-sensitive file-name completion in bash unuseable in Windows, because one often doesn't know the case of names in advance. So I changed it to case-insensitive by adding a few lines to bash/lib/readline/complete.c I'd like to propose that this feature be integrated into the next release of Cygnus Win32. I suppose I should have put the change inside an "#ifdef MSDO" or something, but you can do that when integrating it. The changes are very small so I've appended it as a context diff to the end of this message. Please let me know if you want this contribution in a different format. Thanks. Michael Condict m.condict@opengroup.org The Open Group Research Inst. (617) 621-7349 11 Cambridge Center Cambridge, MA 02142 --------- changes to src/user-tools/bash/lib/readline/complete.c ---------- *** complete.c Mon May 19 13:34:16 1997 --- complete.c.ORIG Mon May 19 13:32:04 1997 *************** *** 1077,1083 **** /* **************************************************************** */ /* Non-zero means that case is not significant in completion. */ ! int completion_case_fold = 1; /* Return an array of (char *) which is a list of completions for TEXT. If there are no completions, return a NULL pointer. --- 1077,1083 ---- /* **************************************************************** */ /* Non-zero means that case is not significant in completion. */ ! int completion_case_fold = 0; /* Return an array of (char *) which is a list of completions for TEXT. If there are no completions, return a NULL pointer. *************** *** 1265,1293 **** { /* Otherwise, if these match up to the length of filename, then it is a match. */ ! if (completion_case_fold) ! { ! if (((int)D_NAMLEN (entry)) >= filename_len) ! { ! register int c1, c2, si; ! ! for (si = 0; si < filename_len; si++) ! { ! c1 = to_lower(entry->d_name[si]); ! c2 = to_lower(filename[si]); ! if (c1 != c2) break; ! } ! if (si == filename_len) ! break; ! } ! } ! else ! { ! if ((entry->d_name[0] == filename[0]) && ! (((int)D_NAMLEN (entry)) >= filename_len) && ! (strncmp (filename, entry->d_name, filename_len) == 0)) ! break; ! } } } --- 1265,1274 ---- { /* Otherwise, if these match up to the length of filename, then it is a match. */ ! if ((entry->d_name[0] == filename[0]) && ! (((int)D_NAMLEN (entry)) >= filename_len) && ! (strncmp (filename, entry->d_name, filename_len) == 0)) ! break; } } - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From banders@switch.rockwell.com Mon May 19 11:35:00 1997 From: banders@switch.rockwell.com (Bart Anderson) Date: Mon, 19 May 1997 11:35:00 -0000 Subject: Gdk95.lib Message-ID: <33809E15.6832@switch.rockwell.com> Does anyone know if gnuwin32 can link with Gdk95.lib. I tried putting it with the other libs and specifying -lGdk95 and -lgdk95 with no luck. Bart - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From cro@nca.asu.edu Mon May 19 15:01:00 1997 From: cro@nca.asu.edu (C. R. Oldham) Date: Mon, 19 May 1997 15:01:00 -0000 Subject: what went wrong? References: <199705191749.KAA28822@cirdan.cygnus.com> Message-ID: <87u3jzf7za.fsf@socrates.nca.asu.edu> Stephen Crowley wrote: >> Just wondering but what did you guys do to screw up b18 so bad? Well, I don't think that was really deserved. I thought the "b" in "b18" was for "beta"--thus some problems are to be expected. Most everything still works for me, and lots of stuff that didn't work under b17 does now. >>>>> "Geoffrey" == Geoffrey Noer writes: Geoffrey> I also haven't seen a slowing in starting programs or in Geoffrey> compiling executables. In fact, I see a speed-up in compile Geoffrey> times over beta 17... Hm. I don't see that. In general things, including compiles, seem much slower now. -- | Charles R. (C. R.) Oldham | NCA Commission on Schools | | cro@nca.asu.edu | Arizona St. Univ., PO Box 873011 | | V:602/965-8700 F:602/965-9423 |________ Tempe, AZ 85287-3011_ | | "I like it!"--Citizen G'Kar, Babylon 5 | #include X_>| - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From kunglao@prairienet.org Mon May 19 15:01:00 1997 From: kunglao@prairienet.org (kunglao@prairienet.org) Date: Mon, 19 May 1997 15:01:00 -0000 Subject: ncurses/termcap, just won't clear the screen? Message-ID: <199705192156.QAA29407@bluestem.prairienet.org> I've just rebuilt termcap & ncurses, that went fine. But now I'm trying to run some of the test programs that ncurses has. And it appears that it just cannot erase the screen. The curses is set to the top of the window, but all the text is still there. I've tried a few different things for my TERM env variable, linux, ansi, pc-ansi, vt100, etc. They all behaved the same. Might I have forgotten a step somewhere? --- "Oh, what sad times are these when passing ruffians can say 'Ni!' at will to old ladies. There is a pestulance upon this land, nothing is sacred. Even those who arrange and design shrubberies are under considerable economic stress of this period of history." -- Roger the Shrubber - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From mka@redes.int.com.mx Mon May 19 18:39:00 1997 From: mka@redes.int.com.mx (Michael Anderson) Date: Mon, 19 May 1997 18:39:00 -0000 Subject: Trouble downloading b18 file Message-ID: <338100BE.5AF5163@redes.int.com.mx> Ive made several attempts to ftp this file: ftp.cygnus.com:/pub/gnu-win32/gnu-win32-b18/cdk-split/xah and have never been able to get the entire file. The ftp hangs each time at 1664436 bytes but the file size should be 1668516. If I combine the xah with the others to make a cdk.exe, it will not execute and complains of a corrupted file. Could xah be corrupted? I was able to download all the other files in the same directory without a problem. Thanks, Mike Anderson mka@redes.int.com.mx - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From noer@cygnus.com Mon May 19 19:23:00 1997 From: noer@cygnus.com (Geoffrey Noer) Date: Mon, 19 May 1997 19:23:00 -0000 Subject: Problems building b18 under win95 References: <199705151008.LAA00154@ma.hw.ac.uk> Message-ID: <199705200222.TAA23133@rtl.cygnus.com> David Fiddes wrote: > > I'm having quite a few problems rebuilding a "plain vanilla" version of gcc > for win32. [...] > tar: Could not change access and modification times of cdk: Permission > denied Should be pretty harmless I think. [...] > gcc -O2 -c -D_GNU_SOURCE -I. -I/cdk/binutils -I../bfd -I/cdk/binutils/../bfd > -I/cdk/binutils/../include -O2 deflex.c > gcc -O2 -O2 -o dlltool dlltool.o defparse.o deflex.o bucomm.o version.o > filemode.o -L../bfd -lbfd ../libiberty/libiberty.a > make: DLLTOOL@: Command not found > make: *** [libtcl7.6.a] Error 127 > make: *** [all] Error 2 > make: *** [all-tcl] Error 2 > > I'm not sure what to do. Is this a problem with my setup or does it just > not build under 95 at all. There is a known problem with building the distribution under both Win 95 and NT -- I think for configuration reasons tcl, tk, and gdbtk cannot be rebuilt under NT and Win95, only cross-built under the Unix of your choice. (This part of b18 is more work-in-progress than the rest). If you configure the sources adding a "--disable-gdbtk", it will build the command-line version of gdb without the GUI and that should work better. If you only want to rebuild gcc, you might want to invoke make with "make all-gcc" and then "make install-gcc" to avoid build failures in tcl, tk, and/or gdb. Alternatively, configure with "--disable-gdbtk" and then build with "make -k" and install with "make -k install" which tells make to try to kontinue past any build problems... -- Geoffrey Noer noer@cygnus.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From bill@dstc.edu.au Tue May 20 00:17:00 1997 From: bill@dstc.edu.au (Bill Segall) Date: Tue, 20 May 1997 00:17:00 -0000 Subject: C pthreads wrappers help Message-ID: <199705200717.RAA14686@piglet.dstc.edu.au> Has anyone got a pthreads layer for cygwin32? I'm reasonably desperate :-) Bill. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Tue May 20 00:33:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Tue, 20 May 1997 00:33:00 -0000 Subject: X11R6.3 nolonger works under B18 Message-ID: <01BC6505.2A42BD30@gater.krystalbank.msk.ru> Gary Bishop wrote: > I still can't get the client programs to work, I'm sure there is some > simple secret here that I am overlooking. They work for Sergey so it is > known to be possible, but there is some subtle difference in his > environment that is making the difference. > > I added fixup.o to my "specs" file as the *endfile for dlls. The line > suggested by Mikey that I am using is: > *endfile: > %(dll:/lib/fixup.o) > I didn't added this fixup. -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From ogre@aa.net Tue May 20 00:33:00 1997 From: ogre@aa.net (Joe Rumsey) Date: Tue, 20 May 1997 00:33:00 -0000 Subject: Diffs for ssh-1.2.20 Message-ID: <3381539F.9EDBB5E7@aa.net> Since several people have asked, I've made the diffs for compiling ssh-1.2.20 available. It's at http://www.aa.net/~ogre/ssh-1.2.20.diff sshd compiles too, but it doesn't work. Might under NT, I dunno. No guarantees as to security or anything else. Have fun. Joe - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From noer@cygnus.com Tue May 20 06:03:00 1997 From: noer@cygnus.com (Geoffrey Noer) Date: Tue, 20 May 1997 06:03:00 -0000 Subject: ls starts VERY SLOWLY w/ b18 on Windows NT 4.0 References: <337E8D83.7A5C@peterham.com> Message-ID: <199705200754.AAA26910@rtl.cygnus.com> Peter Ham wrote: > > I have noticed that program startup is MUCH SLOWER than > the previous release that I downloaded in November or so. > It takes 16 seconds on an empty directory on a lightly loaded > Pentium 133mhz. > > Could it be that the cygwin.dll keeps getting loaded and unloaded > for each program execution? Is there any way to lock it in? > It's frustrating to use "ls" (or any other frequently used program) > in this configuration. That's very odd. My main test machine is a fairly normal Micron 166Mhz Pentium with NT 4.0 server and Windows 95 installed. It has 32 MB of RAM which isn't a completely obscene amount these days. I see a definite speed-up in compile times and usage of the various tools feels snappier. I should actually measure the time to do configures under b17 and 18 at some point but it certainly isn't taking 16 seconds to do an 'ls' of a fairly empty directory for me. The cygwin.dll in beta 18 does use more RAM, so it could be that the dll is being swapped in and out of memory. At the risk of getting hundreds of replies :-), those of you who are seeing the slow-down: 1) Are you running NT or Win 95? 2) How much RAM? 3) Are you using memory for other open applications when you see the poor performance? If so, does cygwin performance increase when you quit those apps? Please reply to me directly. I'll post a summary on the list. Thanks, -- Geoffrey Noer noer@cygnus.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From noer@cygnus.com Tue May 20 06:03:00 1997 From: noer@cygnus.com (Geoffrey Noer) Date: Tue, 20 May 1997 06:03:00 -0000 Subject: Moving the default cygwin.dll load location Message-ID: <199705200759.AAA26928@rtl.cygnus.com> Another explanation for the slowness some folks are seeing with b18 is that another dll might be trying to load in the same memory space as the cygwin.dll, causing them to get swapped in and out of memory. The default load location appears to be 0x10001000. Some time ago, a few people suggested that this be changed in favor of a more favorable place. What should it be? How does one pick a sensible value? -- Geoffrey Noer noer@cygnus.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From nicolas@bnp-eng.remcomp.com Tue May 20 06:03:00 1997 From: nicolas@bnp-eng.remcomp.com (Eric Nicolas) Date: Tue, 20 May 1997 06:03:00 -0000 Subject: Compiling PERL5.003 with B18 Message-ID: <338194D2.15FB@bnp-eng.remcomp.com> Hi, I just installed BETA18. Everything seems to work just fine. Even the 'ls -l' is a lot faster than before. "Bravo" ! I'm now trying to compile Perl 5.003 with B18. I read a lot of mails about perl mixing its 'configure' and 'Configure' files, etc. So : * I have a /src directory, mounted with the '-m' option, so that file names are case sensitives ; * I have the perl5003.tar file in the /src directory. * I try a 'tar xvf perl5003.tar' and I've got the message : creating 'Configure' : access denied. And finally I have only the 'configure' script, not the 'Configure' one !! I thought 'tar' should have created both files, as my directory is mounted with '-b'. I am wrong ?? Thanks. * Eric. -- ----------------------------------------------------------------------------- Eric NICOLAS Tel : 01.40.14.01.39 BNP-BFI-IM Fax : 01.40.14.99.47 13, rue LaFayette EMail: nicolas@bnp-eng.remcomp.com 75009 PARIS ----------------------------------------------------------------------------- - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Tue May 20 06:03:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Tue, 20 May 1997 06:03:00 -0000 Subject: ncurses/termcap, just won't clear the screen? Message-ID: <01BC6509.875CACC0@gater.krystalbank.msk.ru> kunglao@prairienet.org wrote: > > I've just rebuilt termcap & ncurses, that went fine. But now I'm > trying to run some of the test programs that ncurses has. And it > appears that it just cannot erase the screen. The curses is set to > the top of the window, but all the text is still there. I've tried a > few different things for my TERM env variable, linux, ansi, pc-ansi, > vt100, etc. They all behaved the same. > Is your /etc/termcap right? -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Tue May 20 06:03:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Tue, 20 May 1997 06:03:00 -0000 Subject: Fix to readline for case-insensitive file-name completion Message-ID: <01BC6509.83F2A6C0@gater.krystalbank.msk.ru> Michael Condict wrote: > > Geoffrey, > > I found the case-sensitive file-name completion in bash unuseable in > Windows, because one often doesn't know the case of names in advance. > So I changed it to case-insensitive by adding a few lines to > > bash/lib/readline/complete.c > > I'd like to propose that this feature be integrated into the next > release of Cygnus Win32. Sometimes I'm using case-sensitive file names with cygnus tools (just by adding FILE_FLAG_POSIX_SEMANTICS to some CreateFile calls in cygwin.dll). Huh?-) Moreover, if I have two directories "X11R6.3" and "xfiles" I prefer different completions for X and x. -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From chabe04@cai.com Tue May 20 06:03:00 1997 From: chabe04@cai.com (Benoit Champagne) Date: Tue, 20 May 1997 06:03:00 -0000 Subject: Split on large files References: <199705161920.MAA23339@outwest1.outwest.net> Message-ID: <9705201235.AA28452@calasu13.cai.com> ===>M KARKLINS writes: Matiss> I haven't been able to get split to work on large files, Matiss> specifically the cdk.exe file. The split files available are to Matiss> large to fit onto 3.5" floppies formatted for DMF, so I was trying Matiss> to resize the split. There appears to be a size limitation, where Matiss> it is I don't know. Split does appear to work on smaller files. Matiss> I am running Win95 on a 100 MHz Pentium with a 500 Meg SCSI HD on a Matiss> Novell network. Matiss> Anyone else experience something similar? Matiss> Matiss email: matiss@outwest.net The zipsplit exec has an option '-n' used to limit the size of the resulting splits. Just give the option along with the value 1350000 and the splits should fit easily on a 3.5" floppy. If you don't have the '-n' option then we're not using the same version of the tool. Hope that helps a little. -- --------------------------------------------------------------------- Continuum all the way COMPUTER ASSOCIATES Benoit Champagne ift.a | Main Line : (514) 978-4899 Tel : (514) 978-4913 | 3090 Boul. Le Carrefour FAX : (514) 978-4909 | Suite 300 Email: chabe04@cai.com | Laval, PQ PLAN : Live Long And Prosper | H7T 2J7 --------------------------------------------------------------------- -- -----BEGIN PGP PUBLIC KEY BLOCK----- Version: 2.6.3ia Comment: Processed by Mailcrypt 3.4, an Emacs/PGP interface mQCNAzGbL8kAAAEEAMlajekEt3h7gLcFmOov2CdhcUn+53Ux2+IwAaf0kTGPsvXo 0WsQp6yjbRp4rpruxKiduIlRUP6nqixGP7q7LqCEbGCQQHAH59NXefEp7A4muPmI VUhr1zLEyuq5VlWUtyHr0ACQvyeIoumY13ZVYjk2VUeVkD47TXdV9JrD+yYxAAUR tCJCZW5vaXQgQ2hhbXBhZ25lIDxjaGFiZTA0QGNhaS5jb20+iQCVAwUQMZsvyndV 9JrD+yYxAQGaFAP/a+vooLpbv3bXr6hs0ZfhS8x5mgH1x1Z5ImdR7XBCNGZrQNWu s0wEUTylNjE4e9M1Dpv9kVvyHD0OauvE+xNaIFgV2ABKzGzVzyyg0EoTNRBjY9cN iub++OT2QA+56bLF3gy5/vD+z3HzunsfrcgYhsEj30iA0rbQecrfGrVbD2Q= =4XrH -----END PGP PUBLIC KEY BLOCK----- From uwe@genias.de Tue May 20 06:03:00 1997 From: uwe@genias.de (Uwe Wuerfel) Date: Tue, 20 May 1997 06:03:00 -0000 Subject: Access to NT 4.0 from UNIX system running X References: <337EE576.B24BBCC1@hazbro.com> Message-ID: > Mikey wrote: > > > > http://tnt.microimages.com/www/html/freestuf/mix.htm > > I had seen the MI/X product before. Perhaps I am missing something, but is > this not focused on accessing network applications FROM the Win95/NT > machine, ie. an X Server for the Win95/NT environment? I am looking for the > opposite, access TO the Win NT machine from a network computer, ie. an X > Client on the Win NT machine. The intent would be to allow Windows apps to > run where they belong...on a Windows OS. I would like access to them from > time to time and am not really thrilled with the implementation of WABI. > Thanks. Try WinCenter Pro from NCD. ciao, uwe 8X--------------------------------------------------------------------X8 If you cut here, you will probably destroy your monitor. Uwe Wuerfel Morgenstund... ...kommt selten allein. uwe@genias.de - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From JoopvandeWege@mail.mococo.nl Tue May 20 06:03:00 1997 From: JoopvandeWege@mail.mococo.nl (Joop van de Wege) Date: Tue, 20 May 1997 06:03:00 -0000 Subject: GNU-Win32 m68K cross compiler References: <335A2C3D.4FC9.cygnus.gnu-win32@datasalg.no> <5js886$hcf$1@majipoor.cygnus.com> Message-ID: <11480028000003@mococo.nl> On 26 Apr 1997 06:44:22 GMT dje@cygnus.com (Doug Evans) wrote: > Jan Ove Etterlid (joe@datasalg.no) wrote: > > Is there anyone who can give me a list of the steps I have to go to > > build a GNU-Win32 gcc cross compiler, assembler and linker which can > > produce m68k coff format executable files, running on Win NT 3.51 > > pentium PC? > You may wish to check out the crossgcc FAQ. I did that but didn't find the answers needed for a cross-compiler like this: --host=i386-cygwin32 --target=m68k-amiga{d}os The binutils compile fine and it has only minor problems trying to install the documentation (guide/info), that is no problem since i have it on my Amiga. The problem is with compiling gcc itself. The cygwin32 distribution doesn't know about m68k-amiga{d}os and the Amiga distribution doesn't know about cygwin32 (files involved are: config.sub and config.guess). Simply exchanging them doesn't work. So I have directions from both environments on how to make a crosscompiler but with an dependency that is like a chicken and egg problem. Anyone make a m68k target with cygwin32 as a host? Can anyone help me a bit in the right direction on solving this? Thanks for any help given, Joop - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From kunglao@prairienet.org Tue May 20 06:03:00 1997 From: kunglao@prairienet.org (kunglao@prairienet.org) Date: Tue, 20 May 1997 06:03:00 -0000 Subject: ncurses/termcap, just won't clear the screen? References: <01BC6509.875CACC0@gater.krystalbank.msk.ru> Message-ID: <199705201132.GAA10230@bluestem.prairienet.org> > kunglao@prairienet.org wrote: > > > > I've just rebuilt termcap & ncurses, that went fine. But now I'm > > trying to run some of the test programs that ncurses has. And it > > appears that it just cannot erase the screen. The curses is set to > > the top of the window, but all the text is still there. I've tried a > > few different things for my TERM env variable, linux, ansi, pc-ansi, > > vt100, etc. They all behaved the same. > > > > Is your /etc/termcap right? copied from /usr/etc after the install. I have now found that unixpc and vt200 can clear the screen, but then they cannot do things like color, inverse, etc. in the ncurses tests. I'll have to skim through the termcap and see what it's doing different with these. --- "Oh, what sad times are these when passing ruffians can say 'Ni!' at will to old ladies. There is a pestulance upon this land, nothing is sacred. Even those who arrange and design shrubberies are under considerable economic stress of this period of history." -- Roger the Shrubber - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From nicolas@bnp-eng.remcomp.com Tue May 20 06:03:00 1997 From: nicolas@bnp-eng.remcomp.com (Eric Nicolas) Date: Tue, 20 May 1997 06:03:00 -0000 Subject: Problem with '@' character in B18 Message-ID: <33819516.59E2@bnp-eng.remcomp.com> Hi, I have a problem with BASH, in B18. I am not able to type any '@' character in the command line !!!!! I am very embarased, as I use CVS and CVSROOT variables contain '@' !! Thanks for any help. Eric. -- ----------------------------------------------------------------------------- Eric NICOLAS Tel : 01.40.14.01.39 BNP-BFI-IM Fax : 01.40.14.99.47 13, rue LaFayette EMail: nicolas@bnp-eng.remcomp.com 75009 PARIS ----------------------------------------------------------------------------- - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From noer@cygnus.com Tue May 20 18:08:00 1997 From: noer@cygnus.com (Geoffrey Noer) Date: Tue, 20 May 1997 18:08:00 -0000 Subject: b18: setpwend instead of setpwent: a typo? References: <50lo5neiss.fsf@mbcsg1.sghms.ac.uk> Message-ID: <199705202123.OAA08553@cirdan.cygnus.com> Christian Soeller wrote: > > In libcygwin.a there is now the function setpwend but my unix manpages > only have setpwent and the package I am compiling (perl 5.03_99) wants > setpwent as well. Is this a typo that made its way into beta18 ???? Looks like it. :-) I've made the change in current development sources. -- Geoffrey Noer noer@cygnus.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From billtut@microsoft.com Tue May 20 18:08:00 1997 From: billtut@microsoft.com (Bill Tutt) Date: Tue, 20 May 1997 18:08:00 -0000 Subject: Moving the default cygwin.dll load location Message-ID: There really isn't a sensible value AFAIK. The best bet is to make use of bind.exe, and rebase.exe that come with the Win32SDK. Rebase: Reads a series of DLL's and make sure the base addresses of all the DLL's don't conflict. If necessary rebasing the DLL's so they don't have conflicting base addresses. Bind: Bind is a command-line tool that you can use to compute addresses of exported DLL functions. The linker stores addresses of exported DLL functions in a table that is separate from the table that contains the names of the exported DLL functions. Calls to exported DLL functions are made using an indirect jump through the table that contains the addresses. You can minimize load time by using Bind to bypass this lookup. Bind your executable with all of your DLLs, plus the system DLLs, by running Bind once you have installed the application on the computer where you will be running it. The idea behind these 2 programs is something like this: After building the program: Run rebase.exe on the DLL's associated with your program. While installing the program: Run rebase on the DLL's associated with your program. (Also including stuff that's in winnt\system32 woudn't be a bad idea either.) Run bind on your executable images. Unfortunately those 2 files don't appear to be redistributable as far as I can tell.. :( On the up side however.. The functionality in rebase.exe and bind.exe both come from the IMGHLP.DLL. In particular: ReBaseImage, and BindImageEx. Hope this helps, Bill Tutt billtut@microsoft.com Not speaking for Microsoft, just myself. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From rrankin@primenet.com Tue May 20 19:30:00 1997 From: rrankin@primenet.com (Rick Rankin) Date: Tue, 20 May 1997 19:30:00 -0000 Subject: Fix to readline for case-insensitive file-name completion References: <01BC6509.83F2A6C0@gater.krystalbank.msk.ru> Message-ID: <3.0.1.32.19970520190749.006a6194@pop.primenet.com> I too generally prefer case-sensitive completion. It's the same behaviour as Unix and, perhaps I'm weird (no, make that "different"), but I almost always know the case of my filenames. Even if I don't remember, I can hit before I type any characters and bash shows me the possible completions. Rick Rankin rrankin@primenet.com At 10:28 AM 5/20/97 +0400, Sergey Okhapkin wrote: >Michael Condict wrote: >> >> Geoffrey, >> >> I found the case-sensitive file-name completion in bash unuseable in >> Windows, because one often doesn't know the case of names in advance. >> So I changed it to case-insensitive by adding a few lines to >> >> bash/lib/readline/complete.c >> >> I'd like to propose that this feature be integrated into the next >> release of Cygnus Win32. > >Sometimes I'm using case-sensitive file names with cygnus tools (just by >adding FILE_FLAG_POSIX_SEMANTICS to some CreateFile calls in cygwin.dll). >Huh?-) Moreover, if I have two directories "X11R6.3" and "xfiles" I prefer >different completions for X and x. > >-- >Sergey Okhapkin >Moscow, Russia >Looking for a job. > > >- >For help on using this list (especially unsubscribing), send a message to >"gnu-win32-request@cygnus.com" with one line of text: "help". > > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From home@morla.crec.mipt.ru Tue May 20 19:30:00 1997 From: home@morla.crec.mipt.ru (Igor V. Kovalenko) Date: Tue, 20 May 1997 19:30:00 -0000 Subject: Diffs for ssh-1.2.20 References: <3381539F.9EDBB5E7@aa.net> Message-ID: <199705210230.GAA00353@morla.crec.mipt.ru> > Since several people have asked, I've made the diffs for compiling > ssh-1.2.20 available. > It's at http://www.aa.net/~ogre/ssh-1.2.20.diff > > sshd compiles too, but it doesn't work. Might under NT, I dunno. -- No, mostly because of missing pseudo-terminal support code. Also (cosmetic) note: still missing cygwin32_setsockopt in cygwin.din (beta18 source). setXXXent/endXXXent (XXX=host,net,pw...) pairs might as well be defined to zero code in appropriate header files. -- Good luck! PS: I have compiled telnet and finger clients from NetKit-B-0.06 (plus RedHat patches) with minor source changes under beta18 environment. It even works! under WinNT4.0(+sp2) so I replaced original telnet with a cygwin-compiled one. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From banders@switch.rockwell.com Tue May 20 19:30:00 1997 From: banders@switch.rockwell.com (Bart Anderson) Date: Tue, 20 May 1997 19:30:00 -0000 Subject: Gimp on gnuwin32 Message-ID: <3381B3E8.5B5B@switch.rockwell.com> Has anyone gotten Gimp (Linux image processing) to work with gnuwin32? If so were there any tricks/patches required? Any advice? I'm pretty early into trying to get it to work, but it takes a day for my email to be posted, so I play the field. Thanks, Bart - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From *jeffdb@netzone.nospam.com Tue May 20 19:30:00 1997 From: *jeffdb@netzone.nospam.com (Mikey) Date: Tue, 20 May 1997 19:30:00 -0000 Subject: X11R6.3 nolonger works under B18 Message-ID: <199705210042.RAA21927@nz1.netzone.com> Then how are you terminating the list of dll Exports? (*jeffdb@netzone.com) the return address for this message is anti spammed remove * from the above address to reply. Mikey >Gary Bishop wrote: >> I still can't get the client programs to work, I'm sure there is some >> simple secret here that I am overlooking. They work for Sergey so it is >> known to be possible, but there is some subtle difference in his >> environment that is making the difference. >> >> I added fixup.o to my "specs" file as the *endfile for dlls. The line >> suggested by Mikey that I am using is: >> *endfile: >> %(dll:/lib/fixup.o) >> > >I didn't added this fixup. > >-- >Sergey Okhapkin >Moscow, Russia >Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From lhall@rfk.com Tue May 20 21:40:00 1997 From: lhall@rfk.com (Larry Hall) Date: Tue, 20 May 1997 21:40:00 -0000 Subject: Fix to readline for case-insensitive file-name completion Message-ID: <2.2.32.19970521001522.0092b1b8@ma.ultranet.com> At 10:28 AM 5/20/97 +0400, Sergey Okhapkin wrote: >Michael Condict wrote: >> >> Geoffrey, >> >> I found the case-sensitive file-name completion in bash unuseable in >> Windows, because one often doesn't know the case of names in advance. >> So I changed it to case-insensitive by adding a few lines to >> >> bash/lib/readline/complete.c >> >> I'd like to propose that this feature be integrated into the next >> release of Cygnus Win32. > >Sometimes I'm using case-sensitive file names with cygnus tools (just by >adding FILE_FLAG_POSIX_SEMANTICS to some CreateFile calls in cygwin.dll). >Huh?-) Moreover, if I have two directories "X11R6.3" and "xfiles" I prefer >different completions for X and x. > >-- >Sergey Okhapkin >Moscow, Russia >Looking for a job. > > If we're voting, I'm with Sergey. If there's going to be a change, I'd like it to be configurable. Larry Hall lhall@rfk.com RFK Partners, Inc. (617) 239-1053 8 Grove Street (617) 239-1655 - FAX Wellesley, MA 02181 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jwilkins@secnet.com Tue May 20 21:40:00 1997 From: jwilkins@secnet.com (Jonathan Wilkins) Date: Tue, 20 May 1997 21:40:00 -0000 Subject: Beta 18 breaks gettimeofday Message-ID: <3.0.32.19970520174814.02732ca8@silence.secnet.com> I have just upgraded to beta 18 and was rebuilding an application and had some troubles.. after tracking things down it seems to be a problem with the gettimeofday() call.. sample code follows: #include #include #include #include #include int main() { struct timeval *tvp; struct timeval tv; tvp = &tv; if(gettimeofday(tvp, NULL) != 0){ printf("\n failed \n error was : %s\n", strerror(errno)); } else { printf("\n worked \n tv_sec is : %i\n tv_usec is : %i", tv.tv_sec, tv.tv_usec); } return(0); } this code works on freebsd, and other gettimeofday calls used to work on B17.1 any ideas? thanks Jonathan -=-=-=-=-=-=-=- Jonathan Wilkins | Futuaris | If only they had used their jwilkins@secnet.com | Non Irresus | terminals for niceness instead http://www.secnet.com | Ridebus | of evil ...-Maxwell Smart - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From *jeffdb@netzone.nospam.com Tue May 20 21:40:00 1997 From: *jeffdb@netzone.nospam.com (Mikey) Date: Tue, 20 May 1997 21:40:00 -0000 Subject: Picking up include directories automatically Message-ID: <199705210042.RAA21924@nz1.netzone.com> How about putting export /c/dir because you mounted c:\\ to /c which was my point in the first place YOU DON'T NEED // The few programs that I've worked with that did their own path handleing under linux "handled" // by assuming that they were dupes, and stripping them out whenever they found them, begining middle or end, and since those programs were being compiled on a platform that "aims" for posix compliance, I ASS U MEd that // isn't posix, but under gnu-win32, it dosen't work correctly for all programs, so DON'T USE IT. if you want to write non portable code, that can only ever be compiled on 1 platform, then go out and buy vc++, or use mingw32. This mailing list isn't win32, it's gnu-win32. posix unix emulation on a win32 platform. the more closely you emulate the unix environment, the better your programs will work. Any one who prefers command.com or cmd.exe to bash, csh, tcsh, zsh, or even ash has a large hole in their ummm computer educations. So drop the stuff from the virtual unix page, and help port your favorite shell it really isn't that hard. (*jeffdb@netzone.com) the return address for this message is anti spammed remove * from the above address to reply. Mikey >At 10:39 AM 5/15/97 -0700, Mikey wrote: >>Since the dll and gcc now support dos style naming. >>and the dll supports UNC's the need for // isn't really apparent, >>at least not to me . >>I'm kind of surprised that Geoff left the support in the dll, >>(plus \\\\c is a pretty awkward way to refer to c) >> > >Here's a reason for keeping the //c/dir rather than using the dos c:/dir > >export MYPATH=c:/dir Why would you ever do this under bash anyway? > >Now put this in your Makefile and watch gnumake choke: > >$(MYPATH)/target.o: target.c > $(BUILD_OBJECT) > >Whoops now you've got a colon in your target. That just don't work. > >--jp > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From kunglao@prairienet.org Tue May 20 21:40:00 1997 From: kunglao@prairienet.org (kunglao@prairienet.org) Date: Tue, 20 May 1997 21:40:00 -0000 Subject: A call to find mounted dirs? Message-ID: <199705202137.QAA19342@bluestem.prairienet.org> Is there any gnu-win32 functions that I can use to get the list of mounted dirs, and their DOS paths? I suppose it's possible to go through Win32 and check the registry, but before I have to dig into that, I just wanted to check for anything easier. Thanks - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From ebritten@ea.oac.uci.edu Tue May 20 21:40:00 1997 From: ebritten@ea.oac.uci.edu (Eric Britten) Date: Tue, 20 May 1997 21:40:00 -0000 Subject: newbie dll question References: <199705161320.JAA12675@alt.healthcare.com> Message-ID: On Fri, 16 May 1997, Bret A. Schuhmacher wrote: > Hi, > > I've been using gnu-win32 since back in the v16 days, but never > seriously until now. I downloaded v18 recently and attempted to make > a relocatable dll following all the instructions I could find (thanks > to the people who forged the path :-)). All the files below were > taken from the instructions at > http://www.cygnus.com/misc/gnu-win32/building-reloc-dlls.txt . > > [snip] To solve the problem of undefined references in your dll, you need to include the import library that containes GetModuleHandleA@4. In this case you need to include libkernel32.a since GetModuleHandle is in kernel32.dll Change you link command to... #link DLL ld --base-file tclvc.base --dll -o tcl76.dll \ regexp.o tclAppInit.o tclAsync.o tclBasic.o tclCkalloc.o \ tclClock.o tclCmdAH.o tclCmdIL.o tclCmdMZ.o tclDate.o tclEnv.o \ tclEvent.o tclExpr.o tclFCmd.o tclFHandle.o tclFileName.o tclGet.o \ tclHash.o tclHistory.o tclIO.o tclIOCmd.o tclIOSock.o tclIOUtil.o \ tclInterp.o tclLink.o tclLoad.o tclLoadNone.o tclMain.o \ tclNotify.o tclParse.o tclPkg.o tclPosixStr.o tclPreserve.o tclProc.o \ tclUnixChan.o tclUnixFCmd.o tclUnixFile.o tclUnixInit.o tclUnixNotfy.o \ tclUnixPipe.o tclUnixSock.o tclUnixTime.o tclUtil.o tclVar.o panic.o \ init.o fixup.o $LIBPATH/libcygwin.a -e _dll_entry@12 \ $LIBPATH/libkernel32.a | | Here is the addition that will solve your problem. Eric Britten - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From cgf@bbc.com Tue May 20 21:40:00 1997 From: cgf@bbc.com (Chris Faylor) Date: Tue, 20 May 1997 21:40:00 -0000 Subject: Compiling PERL5.003 with B18 References: <338194D2.15FB@bnp-eng.remcomp.com> Message-ID: In article < 338194D2.15FB@bnp-eng.remcomp.com >, Eric Nicolas wrote: >Hi, > >I just installed BETA18. Everything seems to work just fine. Even >the 'ls -l' is a lot faster than before. "Bravo" ! > >I'm now trying to compile Perl 5.003 with B18. I read a lot of >mails about perl mixing its 'configure' and 'Configure' files, >etc. I've just finished building perl 5.004 for B18. It has a cygwin32 hint file that gets you most of the way there. I had to patch two files to get it working correctly. Environment variable clearing was screwed up and loading of .dll's from mounted directories wasn't working correctly. If you're interested in perl 5.004, I can send you the patches. -- http://www.bbc.com/ cgf@bbc.com "Strange how unreal VMS=>UNIX Solutions Boston Business Computing the real can be." - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From glier@stud.uni-frankfurt.de Wed May 21 02:43:00 1997 From: glier@stud.uni-frankfurt.de (Oliver Glier) Date: Wed, 21 May 1997 02:43:00 -0000 Subject: problem with inner classes in template functions Message-ID: <3382BFCB.89F0591C@stud.uni-frankfurt.de> Dear peoply from the GNU-WIN32 project, my name is Oliver Glier and I'm porting a C++ class library to Windows. The code is already compilable by Borland C++, Sun Sparc C++, and GCC 2.7 (EMX , Linux, and Sun). My problem is that I would like to use your compiler (beta 1.8) , but it doesn't accept the following construct: template class Container { public: class Iterator { public: void setnull(); }; }; template void createIterator(const C&) { C::Iterator it; it.setnull(); } void test() { Container numbers; createIterator(numbers); } The compiler always yields a syntax error at the line with "C::Iterator it;" in function createIterator(). The funny thing is that the compiler rejects the code before createIterator( Container) is generated, that is, while createIterator() is nothing more than a raw template. Still, I think that code as above is correct C++. At least it works with a lot of other C++ compilers and seems to be important in the use of class libraries such as STL. Last but not least I must say that I'm very grateful for your efforts of porting GNU C++ to Windows. You project is great and I hope the best for its future. With best regards, Oliver. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From noer@cygnus.com Wed May 21 02:43:00 1997 From: noer@cygnus.com (Geoffrey Noer) Date: Wed, 21 May 1997 02:43:00 -0000 Subject: two entries in HKEY_CURRENT_USER References: <33730DFB.C28@swissonline.ch> Message-ID: <199705202125.OAA08566@cirdan.cygnus.com> > I noticed that I have two entries in win95 registry under > HKEY_CURRENT_USER/Software/Cygnus Support/CYGWIN.DLL setup > one named b14.0 > and another b15.0 > I currently have cygwin32 17.1. > Is one of them (b14.0) from the old version, and must I erase it ? And beta 18 has HKEY_CURRENT_USER/Software/Cygnus Solutions/CYGWIN.DLL setup with one named b15.0. Old entries shouldn't affect the operation of the newer tools. You may delete them if you like. -- Geoffrey Noer noer@cygnus.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From morse@harborcom.net Wed May 21 02:43:00 1997 From: morse@harborcom.net (Kevin Dahlhausen) Date: Wed, 21 May 1997 02:43:00 -0000 Subject: v18 vs v17 compile times.... Message-ID: <199705201902.PAA09618@ns2.harborcom.net> v17 local build start Tue May 20 14:32:25 GMT-5:00 1997 end Tue May 20 14:34:05 GMT-5:00 1997 v18 local build start Tue May 20 14:36:46 GMT-5:00 1997 end Tue May 20 14:42:52 GMT-5:00 1997 >From this (statistically irrelevant) test, version 18 appears about 27% slower than version 17.1. Somebody else already asked, but I did not see the answer to the question, "Is it (reasonably) easy to build without or turn off symbolic links?" ~~~~~~~~~~~~~~~~~~~ Kevin Dahlhausen ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Morse Controls ap096@po.cwru.edu "Do or do not. There is no 'Try.' " Yoda - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From bischoff@ise.com Wed May 21 02:43:00 1997 From: bischoff@ise.com (Andreas Bischoff) Date: Wed, 21 May 1997 02:43:00 -0000 Subject: How to build gygwin on SUN/Solaris ? Message-ID: <199705201926.MAA28142@ws4.ise.com> Hi all, Geoffrey mentioned that tcl/tk can only be build under Unix. I tried it on my Solaris box, but it couldn't build wish: gcc -O2 tkAppInit.o -L/smallhome/cygnusb18/cdk/tk/unix -ltk4.2 -L/smallhome/cygnusb18/cdk/tcl/unix -ltcl7.6 -lsocket -lnsl -Wl,-Bstatic -L/usr/openwin/lib -lX11 -lXext -lX11 -Wl,-Bdynamic -ldl -lm -lc \ -o wish Undefined first referenced symbol in file wslen /usr/openwin/lib/libX11.a(XSunWcTProp.o) (/usr/lib/libw.so.1?) ld: fatal: Symbol referencing errors. No output written to wish I used the following configure in the top level cdk directory: ./configure --build=sparc-sun-solaris2.4 --host=sparc-sun-solaris2.4 --target=i486-pc-cygwin32 Anything wrong with it ? I admit I never build a cross compiler before :-) TIA, Andreas - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From noer@cygnus.com Wed May 21 02:43:00 1997 From: noer@cygnus.com (Geoffrey Noer) Date: Wed, 21 May 1997 02:43:00 -0000 Subject: ncurses/termcap, just won't clear the screen? References: <01BC6509.875CACC0@gater.krystalbank.msk.ru> Message-ID: <199705202236.PAA09101@cirdan.cygnus.com> Sergey Okhapkin wrote: > > kunglao@prairienet.org wrote: > > > > I've just rebuilt termcap & ncurses, that went fine. But now I'm > > trying to run some of the test programs that ncurses has. And it > > appears that it just cannot erase the screen. The curses is set to > > the top of the window, but all the text is still there. I've tried a > > few different things for my TERM env variable, linux, ansi, pc-ansi, > > vt100, etc. They all behaved the same. > > Is your /etc/termcap right? With the new /etc/termcap from b18, I do an "export TERM=linux" in bash, and then hit a control-l to clear. Under Windows NT, the right thing happens. Under Windows 95, however I see the behavior kunglao@prairienet.org reports where the prompt is moved but the old text doesn't get erased. -- Geoffrey Noer noer@cygnus.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From kh@wg.icl.co.uk Wed May 21 02:43:00 1997 From: kh@wg.icl.co.uk (Kevin Hughes) Date: Wed, 21 May 1997 02:43:00 -0000 Subject: SCCS Message-ID: <9705201533.AA06898@ss38.wg.icl.co.uk> I realise this is slightly off track BUT I have alot of s/w on Unix under SCCS control and would like to port it all to run under win9x. I really would like the same dev.env and control but I cannot find an sccs port (I realise rcs is better but.....) If anyone knows of a port or the source ;-) I would be most interested. Many thanks Kevin - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Wed May 21 02:43:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Wed, 21 May 1997 02:43:00 -0000 Subject: ncurses/termcap, just won't clear the screen? Message-ID: <01BC65ED.28765730@gater.krystalbank.msk.ru> Geoffrey Noer wrote: > With the new /etc/termcap from b18, I do an "export TERM=linux" > in bash, and then hit a control-l to clear. > > Under Windows NT, the right thing happens. Under Windows 95, however > I see the behavior kunglao@prairienet.org reports where the > prompt is moved but the old text doesn't get erased. > ScrollConsoleScreenBuffer() on Windows95 fails to scroll more than half of screen... I'm trying to find a workaround :-( -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From vassilii@optimedia.co.il Wed May 21 02:43:00 1997 From: vassilii@optimedia.co.il (Vassilii Khachaturov) Date: Wed, 21 May 1997 02:43:00 -0000 Subject: gnu-win32-specific bug in RCS 5.7 Message-ID: <33822080@gandalf.optimedia.co.il> Having successfully compiled and used RCS 5.7 for a couple of days on gnuwin32 b.18 with the help of the following patch by Christian Schenk: *** src/conf.sh.~1~ Fri Jun 16 08:19:24 1995 --- src/conf.sh Sun May 11 11:38:02 1997 *************** case `(uname -s -r -v) 2>/dev/null` in *** 1025,1030 **** --- 1025,1032 ---- # Early editions of SunOS 5.4 are reported to have problems with mmap # that generate NUL bytes in RCS files with a Solaris 2.2 NFS server. # This has been reported to be fixed as of patch 101945-10. + 'CYGWIN32/NT 4.0 17.5') ;; + # *) $PREPARE_CC || exit if ($CL -DTRY_MAP_FD=1 a.c $L && $aout &2 I have run into an obvious bug. I have tested it on RCS 5.7 on RS6000, AIX4 and it works there OK. Attached in the end is the uuencoded zip file containing the following files: RCS/cps_ipdf.dtd,v cps_ipdf.dtd 2.log PLEASE NOTE. THESE FILES ARE PROBABLY COPYRIGHTED MATTER. YOU MAY NOT USE THEM FOR REASONS OTHER THAN RCS DEBUGGING. YOU MAY REDISTRIBUTE THEM UNDER THE SAME CONDITIONS PROVIDED THIS NOTICE IS PRESERVED. There is rev. 1.1 checked into the RCS arch., and it's locked. The file `cps_ipdf.dtd' is the second revision to be checked in. Now, as I do ci cps_ipdf.dtd < 2.log and later co cps_ipdf.dtd I get the file incomplete and the RCS archive is corrupted then! Has anyone else observed it on b.18? What about b. 17.1 with the same patch on RCS sources? Does anyone have another/additional patch that doesn't show this problem up? -- Vassilii Khachaturov vassilii@optimedia.co.il Skribu al mi per Esperanto! [[ rcs57bug.ZIP : 42 in rcs57bug.ZIP ]] The following binary file has been uuencoded to ensure successful transmission. Use UUDECODE to extract. begin 600 rcs57bug.ZIP M4$L#!!0``@```"^YM"(````````````````$````4D-3+U!+`P04``(`"`!' MM[0B15N&ZY`%```G%0``$@```%)#4R]C<'-?:7!D9BYD=&0L=JU8_W/:-A3_ M&?\5*CUVD`36M.MV*;T"!;?+4J@'Y);>;K>3+1FT^-M),@D]__%[,@8,-@)V M4W()V!]]WGN?]_0D>TXQJ5RWKML&=APJ1-L02]\./?C@AOU*_7[]J9U_:9=P;&< MAQRMB=H5(0&)S.>H;=@'O7'SYHV!FLT/AO'^Q>!K?_K-,M$<9$%_JBN*OS@^8L$<-`B= M6*F`3(^J_V+#8XZFM]-OJ(:4FVI"M?[2Z@]ZTUZ"&Q?5(B&82:&72.4!J-05 MXT.>RHF$F/D>FGR;3,TAJJJH)Y^'7UJ`J"JLXJEEJ'9^9M2B?BN:HRJ3B9W8 M+.%^(IRH6N($]:,YQ%9NWL?$W3$_[*U$S\S7,LR>\>_?4P7<0R>BMBFNYH>3,4P)-0K41XB1BOE:\J%=JGL^W;71HOAIP*NZ-91 M'$ID-.=84.0KQ\J('I_(+E$"X1;(8.(C73Z%G.PP'1:B)V%M`B%%`^HRM6+" MH$07W"*A(Y>H.AXEUBB9C!)SE`Q&2<]*+"L96\G$2DPK&<#?0?+P4%5V84:, MY#*B)<)@%8_BLWK);V;2OTON[\J5WHOJ()^'@QGH0X/$Y0FAJ1.'Z10ZQC.Z MPZ>1R?QB#L'8!)TQFDTCM54R!EABU`\#J?K`*(0.LRO[RAH`TY:BN."G_@\C M5\9![\,.#OBQN`T*?80OD?MG:@9K<^LJ(O$AJ:1MM%_4A.60&TJ2A/C&MJS77 M16,_NC5;671IU\NBN[]+H]NI2!IM',DIEP/$-MKT@TTX\+VN=@.]6I_&7X%C M"(Z:XU,;FJ:'S;S0QEZ)[JZ?JQ')9CHE"24=]`.*N.@T=.T%ZUCF0MM\ZCZV ME1'_\:G3.*+1E$F/HL\\C*.2R""8%!6N(M,U6T]>%,I,,K2=#LF+=$4F;-DI ME-::H:RTU):655:V<';:F+=;XN?;7C,=^-(0@:4 M,`>.WN1'BU,!S9^2DL1`&>66A3H3%Q@_F4]Y&XA;KA6UE.VB5T#U.#9 M1I9VU\'8_(1.2*IS5AP^ZQ0WV1AM&=S`5^&)P"_@X-865Z>::_94)&?(E["S0-$J/AK!; M;#DY5:7&%^HO/$.GVT1)?7**COH*TX]B^.(XC\_0N7W*%I)C1Y;V*'M;X=@6 MQ28$%[?FHO7]7-O/[I^_Y40(G=?K[U;/>*5GN<>G+=?='X/#9[G5<_+_=99; MLYUXECL86C_DL'U%(1Q8`X3,[%OIBX,R+O4.(WV'`Q3K`V/Z M94\R>G-"=:,A.=N:`4&AB]2+ M,H2Y9`X<"LGF_8B6;I6'OSX87>-?4$L#!!0``@`(`!JXM"+*.B8Q204``+P4 M```,````8W!S7VEP9&8N9'1DK5AM<]HX$/[.KU!IN($DD#9M[RY)IPP%M\S5N[/S5^C\U>6[MY=O?D<+',?,90P9 M3R$Z0NWVAUKM_8MO:/AY_&5BH&]PTF[7CFZ">07RJ-8WIS^&YN!3Y^+M.6HR M/TPX@GLME(]^$D74YRCFF"GO^_?4)P2N&!]Q MS&PT".S$DTB&2^7_N%9$:HQGP]E7U$"?1S<=L*A+6XG3R*VNRC/##O4Z MX3VJ,RXL83$1>2*VP[HB".J%]Y";VKV'B;/F?M3+6,[=-W*;#><_?Z8,.*+A MM*V$/_$K!0'288@C/(\P1!HG%EWG=&LI1@&AKH)YR!AES#=D"%=*VC?]%DZK MZ:<$9G!%%ML*&=Y'.*;(DX&I@!X>R3J0@'0K8##Q@3X_!A%90]I.1(_SB`$@ M10/J,)]Q%O@*7G"'!#9_1O7)6)AC,1T+8RP&8]$SA6F*B2FFIC!,,8#?@;B[ MJTN_,"-!_#FD"F*PS$?BF3WQIR'ZU^+V6LWT1E9;\5SLSX$?Z@LG$H2F06R' MD]8)GM,U/`U-QHTQ`F=3=,"`7I/Z4HP!YACU`Y_+/C`.H*6LTYYY`\-[BDF* M!7_-?Q@YK6UU%\>$=37WL7;V@A'=Y+3ZFOL>X[KICJ>Y:6/2;>E5.J<^C;`+ M?9D`8\QA-(I/D9]X%ART%+P!55D-4("*)MK*-;,TDHSM-,+[("W*1HO`U63+ MXEC'5*AC.0RKG@-W=WC@<[=1^(SV,`KW,,I:169DC$QX0%.#WFQV,YR6#>18 M'L!H-O(^ED,?4*?8`F,/-6S`YI5?DA)PL*O@W?%*&N%LKF.24-)%OZ`PBKLM77O!.I3[ M6-M\FAZVI!/OX;';VL'1C'&7HL]1D(2*S""9U"K(,M,U6YFZN3H];)\#^23.(G$K><$W54U:% M+0SDB/*%+.VN@XGQ">U15/N@/#S6K2ZR"5HA.+XGTXM]KV('MU9V37HAEOE7 MA%:V+%=[O0"XI,:=!<".4Q!^4DW!<0XIYI?Q:%E+G5U_]G4?N_YX:98&D2E! M*H*1JB)*NQ6U(DHJ*!V"(M:];E5$L(7]-=7D341KXS\?^,S^P6(>1,^PLD#3 M4&X-8;5884942BU:R%]XATZ7"84^(XIVQ@K3=]I$B]TX'D.']BDKYA&VN;)' M62N%8RNN-B&XN'(7%O=+;3^_?_B2$R)T6*^_SM[QE'NYA\<5UO7?@^U[N>P] M^?_:RQ5H>^[EMJ;6#R)8OL(`-JR^31$F!$Y5B=J8'%A],_UPH,*2WS#2;S@` M46P8TY,-RNC%'NI&)1ML:_=T6'A8)+BE>$1V.UI-S^W*;QE;2?@O0[.W-7R" M`@?)#V4(1YS9L"DDR^\C6KBL#O\"4$L#!!0``@`(`**VM"*IIU?+E@```+,` M```%````,BYL;V9>88#S3/VX*T%'A#453:M+B#R8JW(Y-"Y=ZKWRK2F*6J@/@)[T/>\ M*B5XAQ$.KR4$=#QN:>@C`KEOPO1[;C#A460W=-WT]ZR#'Q$LI4@>JW MM")%6X;KD`4``"<5```2````````````(0```"(```!20U,O8W!S7VEP9&8N M9'1D+'902P$"%``4``(`"``:N+0BRCHF,4D%``"\%```#````````````"`` M``#B!0``8W!S7VEP9&8N9'1D4$L!`A0`%``"``@`HK:T(JFG5\N6````LP`` M``4````````````@````50L``#(N;&]G4$L%!@`````$``0`WP````X,```` !``4` ` end - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From *jeffdb@netzone.nospam.com Wed May 21 02:43:00 1997 From: *jeffdb@netzone.nospam.com (Mikey) Date: Wed, 21 May 1997 02:43:00 -0000 Subject: ncurses/termcap, just won't clear the screen? Message-ID: <199705210042.RAA21919@nz1.netzone.com> Yeah, here it is. but use ansi, or w32ansi for less elvis info etc, the linux termcap/info is (at least on my system) broken for full screen access. It keeps putting ^M at the end of every wrapped screen line so an extra empty line is inserted, and the cursor shows up on the wrong line, and I can't figure out where it is coming from. In fact if you could confirm for me that it's working the same way for you I would appreciate it ;^) This will also allow texinfo, and the info reader to compile and work with the above TERM's, it does break using "|lesspipe.sh" though, don't yet know why. If you havn't downloaded cdk.tgz yet mail me and I'll send you a copy of the dll, it's 320k, so I don't want to include it for the list. The ATT6whatever below is actually stripdll in outlook express disguise, it automates stripping dll's, I thought you might find it usefull ;^) (*jeffdb@netzone.com) the return address for this message is anti spammed remove * from the above address to reply. Mikey>> You must have missed my post the other day, \E[J & \E0[J are broken. >> clear.exe uses this to clear the screen. >> >ah, ok. That makes things kinda difficult. If you build any >programs, they won't be clearing the screen... Do you happen to know >if anyone is working on a fix for this? > >Thanks > -------------- next part -------------- A non-text attachment was scrubbed... Name: bin00000.bin Type: text/x-shellscript Size: 1557 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: current.dif Type: text/x-diff Size: 11946 bytes Desc: not available URL: From royk@sco.com Wed May 21 06:15:00 1997 From: royk@sco.com (Rohinton Kazak) Date: Wed, 21 May 1997 06:15:00 -0000 Subject: BL18 Installation problems. Message-ID: <3382C62A.D53@sco.com> Hi, I have tried to install cdk.exe bl18 on both W95 and NT3.51 without success. The archive extracts itself and around the 98% mark fails with a message reporting that there is not enough free space in TEMP. I have 150Mb of free disk space, and the archive is 14Mb compressed. I don't have access to the WIN-NT tools used to create this type of archive so I cannot investigate it further... Any help appericated, Roy. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jbharvey@gte.net Wed May 21 06:15:00 1997 From: jbharvey@gte.net (Justin B. Harvey) Date: Wed, 21 May 1997 06:15:00 -0000 Subject: Diffs for ssh-1.2.20 Message-ID: <199705211313.IAA18309@smtp.gte.net> Yeah I can get ssh working barely with Joe's patches, although RSA authentication doesn't work because my uid doesn't match the uid of the key's I've generated. I have no idea why this is... ---------- > From: Igor V. Kovalenko > To: Joe Rumsey > Cc: gnu-win32@cygnus.com > Subject: Re: Diffs for ssh-1.2.20 > Date: Tuesday, May 20, 1997 9:30 PM > > > Since several people have asked, I've made the diffs for compiling > > ssh-1.2.20 available. > > It's at http://www.aa.net/~ogre/ssh-1.2.20.diff > > > > sshd compiles too, but it doesn't work. Might under NT, I dunno. > > -- No, mostly because of missing pseudo-terminal support code. > > Also (cosmetic) note: still missing cygwin32_setsockopt in cygwin.din > (beta18 source). > > setXXXent/endXXXent (XXX=host,net,pw...) pairs might as well > be defined to zero code in appropriate header files. > > -- Good luck! > > PS: I have compiled telnet and finger clients from NetKit-B-0.06 > (plus RedHat patches) with minor source changes under beta18 > environment. It even works! under WinNT4.0(+sp2) so I replaced > original telnet with a cygwin-compiled one. > - > For help on using this list (especially unsubscribing), send a message to > "gnu-win32-request@cygnus.com" with one line of text: "help". - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From fjh@cs.mu.OZ.AU Wed May 21 06:49:00 1997 From: fjh@cs.mu.OZ.AU (Fergus Henderson) Date: Wed, 21 May 1997 06:49:00 -0000 Subject: A call to find mounted dirs? References: <199705202137.QAA19342@bluestem.prairienet.org> Message-ID: <199705211349.XAA03911@mundook.cs.mu.OZ.AU> kunglao@prairienet.org, you wrote: > > Is there any gnu-win32 functions that I can use to get the list of > mounted dirs, and their DOS paths? Yes, setmntent() and getmntent(). See the show_mount() function in winsup/utils/mount.cc for an example. Or alternatively, you could just use system("mount"). -- Fergus Henderson | "I have always known that the pursuit WWW: < http://www.cs.mu.oz.au/~fjh > | of excellence is a lethal habit" PGP: finger fjh@128.250.37.3 | -- the last words of T. S. Garp. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From skachman@unl.edu Wed May 21 07:27:00 1997 From: skachman@unl.edu (Steve Kachman) Date: Wed, 21 May 1997 07:27:00 -0000 Subject: Bash is very slow to respond to keyboard input Message-ID: <3.0.1.32.19970521092652.007951f0@unlinfo.unl.edu> After switching to B18, bash frequently ignores keyboard input when I am using the cursor keys to edit a command line. I will be using the cursor keys to backup quickly and the cursor will pause for awhile and then after a few seconds it will spit out some ANSI escape codes where my cursor was. Possibly related to this is when man fires up less. Less will also tend to get confused. It will ignore some of the keystrokes. After a bit of struggle it will appear to disappear only to have the less prompt appear a bit later. This is all happening on a WIN95 machine. I haven't seen these problem on the NT machine. Any suggestions, pointers, or fixes will be greatly appreciated. Steve Kachman skachman@unl.edu - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From *jeffdb@netzone.nospam.com Wed May 21 07:27:00 1997 From: *jeffdb@netzone.nospam.com (Mikey) Date: Wed, 21 May 1997 07:27:00 -0000 Subject: Strip bug(feature?) + a Preasent for all installers Message-ID: <199705211341.GAA19037@nz1.netzone.com> typing strip libncurses.a produced a libncurses.a file with no exported functions whatsoever. This is not the way strip behaves under linux, typing strip --strip-unneeded still gives a usable library though. I know what the man page says, but don't you think defaulting to --strip-unneeded would give a smaller surpries factor, expecially since it's always, AFAIK, been that way? ;^) To use the included install-sh, cp it to .../cdk/install-sh and to /bin/install and ln -s install.exe install-real somewhere in your path. it's about 70% faster, and the installed files retain the .exe extension (*jeffdb@netzone.nospam.com) the return address for this message is anti spammed remove *&nospam from the above address to reply. Mikey - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From gpasa@swissonline.ch Wed May 21 07:27:00 1997 From: gpasa@swissonline.ch (Pasa Guglielmo) Date: Wed, 21 May 1997 07:27:00 -0000 Subject: error compiling groff-1.10 under b17.1 Message-ID: <3381C42B.69D7@swissonline.ch> > g++ -I. -I. -I../include -I./../include -DHAVE_UNISTD_H=1 -DHAVE_DIRENT_H=1 -DH > VE_LIMITS_H=1 -DHAVE_SYS_DIR_H=1 -DHAVE_STDLIB_H=1 -DSTDLIB_H_DECLARES_PUTENV=1 > -DSTDIO_H_DECLARES_POPEN=1 -DSTDIO_H_DECLARES_PCLOSE=1 -DHAVE_CC_OSFCN_H=1 -DHA > E_CC_LIMITS_H=1 -DRETSIGTYPE=void -DHAVE_STRUCT_EXCEPTION=1 -DHAVE_RENAME=1 -DH > VE_MKSTEMP=1 -DWCOREFLAG=0200 -g -O -c index.cc > index.cc: In method `int index_search_item::load(int)': > index.cc:213: warning: implicit declaration of function `int strchr(...)' > index.cc:213: warning: assignment to `const char *' from `int' lacks a cast > index.cc: In function `class search_item * make_index_search_item(const char *, > int)': > index.cc:276: warning: implicit declaration of function `int strcat(...)' > index.cc: In method `const char * index_search_item::munge_filename(const char > )': > index.cc:429: invalid types `int[int]' for array subscript > index.cc: In method `void index_search_item::read_common_words_file()': > index.cc:548: no matching function for call to `index_search_item::munge_filena > e (int)' > index.cc:425: candidates are: index_search_item::munge_filename(const char *) > index.cc: In method `void index_search_item::check_files()': > index.cc:608: warning: initialization to `const char *' from `int' lacks a cast > index.cc:610: warning: assignment to `const char *' from `int' lacks a cast > make: *** [index.o] Error 1 > make: *** [libbib] Error 2 > make: *** [all] Error 2 This is what I get after launching 'make'. I'm using b17.1. I use the groff-1.10 distribution of Linux on which I did ./configure. I've already installed grep-2.0 succesfully and also termcap (which I got also in Linux/Slackware). and finally for less-332 : > gcc -I. -c -g -O output.c > output.c:288: warning: static declaration for `iprintf' follows non-static > output.c: In function `iprintf': > output.c:290: argument `fmt' doesn't match prototype > D:\cygnus\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32/cygnus-2.7.2-961023/../../.. > /../i386-cygwin32/include/stdio.h:175: prototype declaration > output.c:290: number of arguments doesn't match prototype > D:\cygnus\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32/cygnus-2.7.2-961023/../../.. > /../i386-cygwin32/include/stdio.h:175: prototype declaration > make: *** [output.o] Error 1 -- Sincerely, Pasa Guglielmo ------------------------------------------------------------------ & e-mail : gpasa@swissonline.ch & tel.: +41 (0)24 485 50 40 & & mailing : Pasa Guglielmo & fax : +41 (0)24 485 50 44 & & Rte des Cases 17A & & & CH-1890 St-Maurice & & & (Switzerland) & & ------------------------------------------------------------------ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From *jeffdb@netzone.nospam.com Wed May 21 07:27:00 1997 From: *jeffdb@netzone.nospam.com (Mikey) Date: Wed, 21 May 1997 07:27:00 -0000 Subject: Strip bug(feature?) + a Present for all installers. Message-ID: <199705211348.GAA19651@nz1.netzone.com> Sorry, I forgot to save before send. typing strip libncurses.a produced a libncurses.a file with no exported functions whatsoever. This is not the way strip behaves under linux, typing strip --strip-unneeded still gives a usable library though. I know what the man page says, but don't you think defaulting to --strip-unneeded would give a smaller surpries factor, expecially since it's always, AFAIK, been that way? ;^) To use the included install-sh, cp it to .../cdk/install-sh and to /bin/install and ln -s install.exe install-real somewhere in your path. it's about 70% faster, and the installed files retain the .exe extension works with ncurese less man winsup and groff(all I have tested it with so far ;^) (*jeffdb@netzone.nospam.com) the return address for this message is anti spammed remove *&nospam from the above address to reply. Mikey -------------- next part -------------- A non-text attachment was scrubbed... Name: bin00000.bin Type: text/x-shellscript Size: 6263 bytes Desc: not available URL: From Carl.E.Lawrence-Slater@ae.ge.com Wed May 21 07:27:00 1997 From: Carl.E.Lawrence-Slater@ae.ge.com (Carl Lawrence-Slater) Date: Wed, 21 May 1997 07:27:00 -0000 Subject: Split file downloads Message-ID: <3381C9BF.5767@h0013.ae.ge.com> I too have had problems getting the "cat"ed file to run install shield. There is a document that give me the MD5 checksums for each file, but where do I get a win95 version of MD5? Also has anybody loaded all of gnu-win32 on a Zip disk? How bad is the performance over the parallel port? Regards Carl ======================================================================= | Carl Lawrence-Slater | Phone : (513) 243-7531 | | Mechanical Engineer | Dial Comm : 8*332-7531 | | Analytical Aero/Cooling Technology| Fax : (513) 243-1343 | | GE Aircraft Engines | | | Mail Drop A405 4 | E-Mail : | | 1 Neumann Way | Carl.E.Lawrence-Slater@ae.ge.com| | Cincinnati,Oh. 45215 | | ======================================================================= - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From condict@opengroup.org Wed May 21 07:27:00 1997 From: condict@opengroup.org (Michael Condict) Date: Wed, 21 May 1997 07:27:00 -0000 Subject: ncurses/termcap, just won't clear the screen? References: <199705192156.QAA29407@bluestem.prairienet.org> Message-ID: <199705201611.MAA03601@postman.osf.org> In message < 199705192156.QAA29407@bluestem.prairienet.org >, you write: > I've just rebuilt termcap & ncurses, that went fine. But now I'm > trying to run some of the test programs that ncurses has. And it > appears that it just cannot erase the screen. The curses is set to > the top of the window, but all the text is still there. I've tried a > few different things for my TERM env variable, linux, ansi, pc-ansi, > vt100, etc. They all behaved the same. > > Might I have forgotten a step somewhere? I rebuilt termcap and the "less" program, which configured itself to use libtermcap. When it comes up, it says that my tty is not fully functional, and then it treats it like a dumb scrolling teletype. It doesn't even know how to scroll backwards one line -- instead, it redraws the whole screen. I also observe that visual-history editing in bash (which uses termcap) has the same defect. Unlike the Unix version, where it displays commands that are longer than one line on multiple screen lines, allowing you to move back and forth among any of the lines, the Win32 version uses the fallback method of scrolling sideways and showing you "<" or ">" characters to indicate that some of the line is off the screen. This requires only space and CR to implement, i.e. it works on a completely dumb tty. So it looks like termcap (and ncurses too?) are almost completely non-functional. Michael Condict m.condict@opengroup.org The Open Group Research Inst. (617) 621-7349 11 Cambridge Center Cambridge, MA 02142 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From crash@cygnus.co.jp Wed May 21 07:27:00 1997 From: crash@cygnus.co.jp (Jason Molenda) Date: Wed, 21 May 1997 07:27:00 -0000 Subject: Quick test of gnu-win32 web archives Message-ID: <199705211100.UAA07467@cygnus.co.jp> I think I have things sorted out now. This note is just to make sure I didn't make a typeo any place critical. http://www.cygnus.com/ml/gnu-win32/ The software we use to make the archives (hypermail) is an amazingly weak program when you feed it improperly formed mail notes. When it fails on an incoming message, it ceases to work again until a human goes in and cleans up after it. LAME. I'm now copying a setup I saw the Gnus people using: the archives are now put in subdirectories, one directory for each month. The result of this is that if hypermail chokes on a mail note, the average outage of the archives should be ~15 days (i.e. until a new month rolls around, unless a human fixes it by hand first). Jason Nihon Cygnus If you're north or south of me, good night. If you're east of me, have a nice lunch. If you're west of me, good morning. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From ggp@informix.com Wed May 21 07:51:00 1997 From: ggp@informix.com (Guy Gascoigne - Piggford) Date: Wed, 21 May 1997 07:51:00 -0000 Subject: Fix to readline for case-insensitive file-name completion Message-ID: <3.0.32.19970521074928.009bca20@pop.pdx.informix.com> At 07:07 PM 5/20/97 -0700, Rick Rankin wrote: >I too generally prefer case-sensitive completion. It's the same behaviour >as Unix and, perhaps I'm weird (no, make that "different"), but I almost >always know the case of my filenames. Even if I don't remember, I can hit > before I type any characters and bash shows me the possible completions. Well just to show that this discussion isn't completely one sided :-) I'd have to say that I'm quite in favour of case-insensitive file name completion. Having moved to NT quite a while ago, I've got used to the case-ignorance of the file system and as I use the cygwin tools to enhance NT rather than simply mimic UNIX I have absolutely no problem if there are features here that are only really going to be of interest under NT. Perhaps this behaviour could be dependant upon the value of a shell variable or setting? Guy -- Guy Gascoigne - Piggford (ggp@informix.com) Software Engineer, Informix Software, Inc. (Portland, Oregon) (503) 225-4574, Fax (503) 221-2633 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From bas@healthcare.com Wed May 21 07:51:00 1997 From: bas@healthcare.com (Bret A. Schuhmacher) Date: Wed, 21 May 1997 07:51:00 -0000 Subject: Problems building b18 under win95 References: <199705151008.LAA00154@ma.hw.ac.uk> <199705200222.TAA23133@rtl.cygnus.com> Message-ID: <199705201635.MAA26013@alt.healthcare.com> >>>>> "Geoffrey" == Geoffrey Noer writes: > There is a known problem with building the distribution under both > Win 95 and NT -- I think for configuration reasons tcl, tk, and > gdbtk cannot be rebuilt under NT and Win95, only cross-built under > the Unix of your choice. (This part of b18 is more work-in-progress > than the rest). What problem are you referring to? I just took a version of Tcl I originally compiled under NuTCRACKER, did a make clean, removed the config.cache, and ran the configure script successfully. I then was able to compile Tcl7.6 and came very close to creating a dll out of it (I was missing a lib Sergey helped me find). I'm using b18. Can you elaborate on the problem so it doesn't sneak up on me? Thanks, Bret -- Bret A. Schuhmacher - Software Engineer bas@healthcare.com These opinions are no one's Healthcare Communications, Inc. fault by my own. I stopped to think and forgot how to start again. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From noer@cygnus.com Wed May 21 09:56:00 1997 From: noer@cygnus.com (Geoffrey Noer) Date: Wed, 21 May 1997 09:56:00 -0000 Subject: A call to find mounted dirs? References: <199705202137.QAA19342@bluestem.prairienet.org> Message-ID: <199705211654.JAA13836@cirdan.cygnus.com> kunglao@prairienet.org wrote: > > Is there any gnu-win32 functions that I can use to get the list of > mounted dirs, and their DOS paths? But of course -- just type "mount" with no arguments... -- Geoffrey Noer noer@cygnus.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From curtis@pdi.com Wed May 21 09:56:00 1997 From: curtis@pdi.com (Curtis Galloway) Date: Wed, 21 May 1997 09:56:00 -0000 Subject: Fix to readline for case-insensitive file-name completion References: <3.0.1.32.19970520190749.006a6194@pop.primenet.com> Message-ID: <33831AD6.DD559635@pdi.com> You can argue this several ways, but my opinion is that the function of the shell should follow the function of the OS. On UNIX, I want my shell to be case-sensitive; on Win32, I want it to be case-insensitive. But I can see why you would want the filename completion to pay attention to case. I think it's even possible to have a case-sensitive filesystem on a Win32 machine with an NFS client. Here are diffs to do case-insensitive globbing and filename completion. They are set to be case-insensitive by default on Win32, but can be turned off by unsetting the special variables "glob_case_fold" and "completion_case_fold". --Curtis Galloway *** subst.c.00 Tue May 20 16:25:54 1997 --- subst.c Wed May 21 07:28:22 1997 *************** *** 74,79 **** --- 74,83 ---- extern int no_line_editing; extern int hostname_list_initialized; #endif + #if defined(__GO32__) || defined(__CYGWIN32__) + extern int glob_case_fold; + extern int completion_case_fold; + #endif #if !defined (USE_POSIX_GLOB_LIBRARY) extern int glob_dot_filenames, noglob_dot_filenames; *************** *** 4477,4482 **** --- 4481,4491 ---- sv_glob_dot_filenames (), sv_nolinks (), sv_noclobber (), sv_allow_null_glob_expansion (), sv_strict_posix (); + #if defined(__GO32__) || defined(__CYGWIN32__) + void sv_glob_case_fold (); + void sv_completion_case_fold (); + #endif + #if defined (READLINE) void sv_terminal (), sv_hostname_completion_file (); #endif *************** *** 4549,4554 **** --- 4558,4567 ---- { "allow_null_glob_expansion", sv_allow_null_glob_expansion }, { "noclobber", sv_noclobber }, { "nolinks", sv_nolinks }, + #if defined(__GO32__) || defined(__CYGWIN32__) + { "glob_case_fold", sv_glob_case_fold }, + { "completion_case_fold", sv_completion_case_fold }, + #endif /* __GO32__ || __CYGWIN32__ */ { (char *)0x00, (VFunction *)0x00 } }; *************** *** 4865,4867 **** --- 4878,4896 ---- posix_readline_initialize (posixly_correct); #endif /* READLINE */ } + + #if defined(__GO32__) || defined(__CYGWIN32__) + void + sv_glob_case_fold (name) + char *name; + { + SET_INT_VAR (name, glob_case_fold); + } + + void + sv_completion_case_fold (name) + char *name; + { + SET_INT_VAR (name, completion_case_fold); + } + #endif /* __GO32__ || __CYGWIN32__ */ *** variables.c.00 Wed May 21 07:17:27 1997 --- variables.c Wed May 21 07:28:42 1997 *************** *** 44,49 **** --- 44,53 ---- extern char *primary_prompt, *secondary_prompt; extern Function *this_shell_builtin; extern time_t shell_start_time; + #if defined(__GO32__) || defined(__CYGWIN32__) + extern int glob_case_fold; + extern int completion_case_fold; + #endif /* The list of shell variables that the user has created, or that came from the environment. */ *************** *** 249,254 **** --- 253,263 ---- bind_variable ("OPTIND", "1"); bind_variable ("OPTERR", "1"); #endif /* GETOPTS_BUILTIN */ + + #if defined(__GO32__) || defined(__CYGWIN32__) + bind_variable ("glob_case_fold", itos(glob_case_fold)); + bind_variable ("completion_case_fold", itos(completion_case_fold)); + #endif /* Get the full pathname to THIS shell, and set the BASH variable to it. */ *** lib/glob/fnmatch.c.00 Fri May 09 15:10:00 1997 --- lib/glob/fnmatch.c Wed May 21 07:26:41 1997 *************** *** 25,30 **** --- 25,36 ---- # endif /* !errno */ #endif + #if defined(__GO32__) || defined (__CYGWIN32__) + int glob_case_fold = 1; + #else + int glob_case_fold = 0; + #endif + /* Match STRING against the filename pattern PATTERN, returning zero if it matches, FNM_NOMATCH if not. */ int *************** *** 59,66 **** case '\\': if (!(flags & FNM_NOESCAPE)) c = *p++; ! if (*n != c) ! return (FNM_NOMATCH); break; case '*': --- 65,77 ---- case '\\': if (!(flags & FNM_NOESCAPE)) c = *p++; ! if (glob_case_fold) { ! if (tolower(c) != tolower(*n)) ! return (FNM_NOMATCH); ! } else { ! if (c != *n) ! return (FNM_NOMATCH); ! } break; case '*': *************** *** 175,182 **** break; default: ! if (c != *n) ! return (FNM_NOMATCH); } ++n; --- 186,198 ---- break; default: ! if (glob_case_fold) { ! if (tolower(c) != tolower(*n)) ! return (FNM_NOMATCH); ! } else { ! if (c != *n) ! return (FNM_NOMATCH); ! } } ++n; *** lib/readline/complete.c.00 Fri May 09 15:10:10 1997 --- lib/readline/complete.c Tue May 20 08:49:58 1997 *************** *** 1077,1083 **** --- 1077,1087 ---- /* **************************************************************** */ /* Non-zero means that case is not significant in completion. */ + #if defined (__GO32__) || defined (__CYGWIN32__) + int completion_case_fold = 1; + #else int completion_case_fold = 0; + #endif /* Return an array of (char *) which is a list of completions for TEXT. If there are no completions, return a NULL pointer. *************** *** 1265,1274 **** --- 1269,1285 ---- { /* Otherwise, if these match up to the length of filename, then it is a match. */ + if (completion_case_fold) { + if ((tolower(entry->d_name[0]) == tolower(filename[0])) && + (((int)D_NAMLEN (entry)) >= filename_len) && + (strncasecmp (filename, entry->d_name, filename_len) == 0)) + break; + } else { if ((entry->d_name[0] == filename[0]) && (((int)D_NAMLEN (entry)) >= filename_len) && (strncmp (filename, entry->d_name, filename_len) == 0)) break; + } } } - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jp@nuancecom.com Wed May 21 09:56:00 1997 From: jp@nuancecom.com (JP Shipherd) Date: Wed, 21 May 1997 09:56:00 -0000 Subject: Picking up include directories automatically Message-ID: <3.0.32.19970520103100.00a15490@awesome.nuance.com> At 10:39 AM 5/15/97 -0700, Mikey wrote: >Since the dll and gcc now support dos style naming. >and the dll supports UNC's the need for // isn't really apparent, >at least not to me . >I'm kind of surprised that Geoff left the support in the dll, >(plus \\\\c is a pretty awkward way to refer to c) > Here's a reason for keeping the //c/dir rather than using the dos c:/dir export MYPATH=c:/dir Now put this in your Makefile and watch gnumake choke: $(MYPATH)/target.o: target.c $(BUILD_OBJECT) Whoops now you've got a colon in your target. That just don't work. --jp - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From pfitz@pratique.fr Wed May 21 09:56:00 1997 From: pfitz@pratique.fr (Tony FITZPATRICK) Date: Wed, 21 May 1997 09:56:00 -0000 Subject: beta 18 cdk installation temp space problem under win95 Message-ID: <199705201824.UAA05839@prat.iway.fr> At 19:50 16/05/97 +0200, you wrote: >the cdk.exe refuses to run, showing the following MessageBox: >"InstallShield Self-extracting EXE >Could not find enough disk space for extracting files." >-- Sorry to be late replying and if you haven't already found a solution you might find the following intresting : >It often happens that the Stirling Install Shield wizzard fails to >complete with a stupid Error message about insufficient disk space. The >release Note indicates that it will try to replace Uninst16.exe and/or >uninst.exe in the Windows directory and if these are write protected >(which they are on my system, by an anti virus) then this error is generated. tony - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jbharvey@gte.net Wed May 21 09:56:00 1997 From: jbharvey@gte.net (Justin B. Harvey) Date: Wed, 21 May 1997 09:56:00 -0000 Subject: Standalone linking with minimalist Message-ID: <199705211602.LAA21663@smtp.gte.net> What the heck am I doing wrong? When I compile programs and want them to run by themselves without cygwin.dll I do a: gcc hello.c -o hello.exe -L/src/ming/lib -lcrtdll -lmingw32 -dll Then I can export this program to other machines and run it, but....when I try to compile more complex programs I get errors like this: (c:\src\tf-35b4\src\tf.exe 1736) In cygwin_except_handler (c:\src\tf-35b4\src\tf.exe 1736) Exception trapped! (c:\src\tf-35b4\src\tf.exe 1736) exception C0000005 at 776415CE (c:\src\tf-35b4\src\tf.exe 1736) exception: ax FCC4FCA8 bx 0 cx 1D dx FFFFFFFF (c:\src\tf-35b4\src\tf.exe 1736) exception: si FCC4FCA8 di 6A78ECE0 bp 242F3AC sp242F3A4 (c:\src\tf-35b4\src\tf.exe 1736) exception is: STATUS_ACCESS_VIOLATION (c:\src\tf-35b4\src\tf.exe 1736) Stack trace: (c:\src\tf-35b4\src\tf.exe 1736) frame 0: sp = 0x242F1D8, pc = 0x1000CEC2 (c:\src\tf-35b4\src\tf.exe 1736) frame 1: sp = 0x242F1F4, pc = 0x77F94512 (c:\src\tf-35b4\src\tf.exe 1736) frame 2: sp = 0x242F218, pc = 0x77F88EEB (c:\src\tf-35b4\src\tf.exe 1736) frame 3: sp = 0x242F2A4, pc = 0x77F76266 (c:\src\tf-35b4\src\tf.exe 1736) frame 4: sp = 0x242F3AC, pc = 0x7763C15C (c:\src\tf-35b4\src\tf.exe 1736) frame 5: sp = 0x242F3C8, pc = 0x419112 (c:\src\tf-35b4\src\tf.exe 1736) frame 6: sp = 0x242F3E0, pc = 0x40D442 (c:\src\tf-35b4\src\tf.exe 1736) frame 7: sp = 0x242F3F4, pc = 0x40D2EC (c:\src\tf-35b4\src\tf.exe 1736) frame 8: sp = 0x242F43C, pc = 0x1000C102 (c:\src\tf-35b4\src\tf.exe 1736) frame 9: sp = 0x242FF94, pc = 0x1000C113 (c:\src\tf-35b4\src\tf.exe 1736) frame 10: sp = 0x242FFA0, pc = 0x41C40D (c:\src\tf-35b4\src\tf.exe 1736) frame 11: sp = 0x242FFB0, pc = 0x40103B (c:\src\tf-35b4\src\tf.exe 1736) frame 12: sp = 0x242FFC0, pc = 0x77F1B304 (c:\src\tf-35b4\src\tf.exe 1736) frame 13: sp = 0x242FFF0, pc = 0x0 (c:\src\tf-35b4\src\tf.exe 1736) End of stack trace bash$ What is the correct syntax on a linkline for building standalone executables without cygwin.dll? Frustrated, Justin -- Justin B. Harvey jbharvey@gte.net http://home1.gte.net/jbharvey - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From nordwick@scam.XCF.Berkeley.EDU Wed May 21 10:17:00 1997 From: nordwick@scam.XCF.Berkeley.EDU (Jay Nordwick) Date: Wed, 21 May 1997 10:17:00 -0000 Subject: Gimp on gnuwin32 References: <3381B3E8.5B5B@switch.rockwell.com> Message-ID: <199705211717.KAA19153@cygnus.com> "Has anyone gotten Gimp (Linux image processing) to work with gnuwin32? First, it is not Linux image processing, it is general unix based. In the lab we have it running on FreeBSD, Solaris, HP-sUX, Digital Linux, etc.... Pete and Spencer (the main GIMP programmers) were thinking of doing a native win32 version (actually, having someone else do it). The thought is that a porting library will be too slow, since there are quite a few fork/vfork calls (at least one for each plugin opperation) and alot of the X/shm is very platform specific. "If so were there any tricks/patches required? Any advice? "I'm pretty early into trying to get it to work, but it takes a day for "my email to be posted, so I play the field. Yeah, my advice: wait for the 1.0 release that should be less than three months away. There is no need porting it until then. Also a native port is almost a necessity, since the gimp is very performace based: it is, after all, a photoshop emulator (read: wanna-be) and you wouldn't want photoshop to be built using a porting library would you ? That would be too painfully slow. jay - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From mh@mike.franken.de Wed May 21 10:32:00 1997 From: mh@mike.franken.de (Michael Hirmke) Date: Wed, 21 May 1997 10:32:00 -0000 Subject: problems with b18 Message-ID: <6XGQQHMKpfB@mike.franken.de> Hi *, I donwloaded Cygnus' CDK b18 yesterday and ran into a few problems: 1.) I use a German keyboard and am not able to use |, @ or \ in bash, because the AltGr key combinations won't work. This problem didn't exist in b17.1 ! 2.) find does not work as expected ! Example: E:\>dir Datentraeger in Laufwerk E: ist home Datentraegernummer: 387F-2345 Verzeichnis von E:\ 21.04.97 01:09 DB 13.04.97 17:50 DFUE 10.05.97 17:24 home 17.05.97 17:39 17 Q3.DIR 4 Datei(en) 17 Bytes E:\>ls -l total 1 drwxr-xr-x 12 544 everyone 0 Apr 21 01:09 DB/ drwxr-xr-x 6 544 everyone 0 Apr 13 17:50 DFUE/ -rw-r--r-- 1 544 everyone 17 May 17 17:39 Q3.DIR drwxr-xr-x 5 544 everyone 0 Apr 21 17:53 RECYCLER/ drwxr-xr-x 10 544 everyone 0 May 10 17:24 home/ *but* E:\>find . -name "test" find: ./DFUE: No such file or directory find: ./home: No such file or directory find: ./Q3.DIR: No such file or directory find: ./RECYCLER: No such file or directory What's going on here ? This problem already existed in b17.1 ! Can anybody throw any light on this ? There are some minor glitches, too - but these two are really big problems. Thx and bye. Michael. -- Michael Hirmke | Telefon +49 (911) 557999 Georg-Strobel-Strasse 81 | FAX +49 (911) 557664 90489 Nuernberg | E-Mail mh@mike.franken.de | WWW http://minimike.franken.de/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From noer@cygnus.com Wed May 21 22:21:00 1997 From: noer@cygnus.com (Geoffrey Noer) Date: Wed, 21 May 1997 22:21:00 -0000 Subject: v18 vs v17 compile times.... References: <199705201902.PAA09618@ns2.harborcom.net> Message-ID: <199705220521.WAA26461@rtl.cygnus.com> Kevin Dahlhausen wrote: [...] > Somebody else already asked, but I did not see the answer to the > question, "Is it (reasonably) easy to build without or turn off symbolic > links?" The better question is "is it (reasonably) easy to change the cygwin.dll such that symbolic links aren't a performance problem?". The answer is, we are working on it. -- Geoffrey Noer noer@cygnus.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From fjh@cs.mu.OZ.AU Wed May 21 22:21:00 1997 From: fjh@cs.mu.OZ.AU (Fergus Henderson) Date: Wed, 21 May 1997 22:21:00 -0000 Subject: Fix to readline for case-insensitive file-name completion References: <33831AD6.DD559635@pdi.com> Message-ID: <199705220513.PAA20548@mundook.cs.mu.OZ.AU> Curtis Galloway, you wrote: > > Here are diffs to do case-insensitive globbing and filename completion. > They are set to be case-insensitive by default on Win32, but can be > turned off by unsetting the special variables "glob_case_fold" and > "completion_case_fold". Two suggestions: 1. I think "glob_ignore_case" and "completion_ignore_case" would be better names. 2. For non-interactive scripts at least, the default for globbing should be case-sensitive even on Win32, because scripts may rely on it. -- Fergus Henderson | "I have always known that the pursuit WWW: < http://www.cs.mu.oz.au/~fjh > | of excellence is a lethal habit" PGP: finger fjh@128.250.37.3 | -- the last words of T. S. Garp. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Wed May 21 23:43:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Wed, 21 May 1997 23:43:00 -0000 Subject: ncurses/termcap, just won't clear the screen? Message-ID: <01BC6696.97BB3FD0@gater.krystalbank.msk.ru> Mikey wrote: > <><> > Yeah, here it is. but use ansi, or w32ansi for less elvis info etc, the linux termcap/info is (at least on my system) broken for > full screen access. This diff contains some bugs fixed in b18 :-) -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From cuma@cuma.SNMG.net Wed May 21 23:43:00 1997 From: cuma@cuma.SNMG.net (cuma@cuma.SNMG.net) Date: Wed, 21 May 1997 23:43:00 -0000 Subject: How about tcsh? Message-ID: <199705220638.OAA29808@solaris.net.ncu.edu.tw> Hello , I am a new user of cygnuwin32 is there any body who make tcsh successfully? I try to make tcsh 6.07 ..but seem configure and config.guess can't fingure out system "CYGNUWIN32" (under wni95) any suggestion? Cuma - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From rvaughn@trcinc.com Wed May 21 23:43:00 1997 From: rvaughn@trcinc.com (Roger Vaughn) Date: Wed, 21 May 1997 23:43:00 -0000 Subject: Building Objective-C? Message-ID: <01BC65FA.4AFD12A0@DALIUS.trcinc.com> The Cygnus distribution does not include the pre-built Objective-C support, and does not build it by itself. Can anyone give me pointers on building libobjc.a with Cygnus? Thanks! - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Wed May 21 23:56:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Wed, 21 May 1997 23:56:00 -0000 Subject: Bash is very slow to respond to keyboard input Message-ID: <01BC669E.FF049E90@gater.krystalbank.msk.ru> Steve Kachman wrote: > This is all happening on a WIN95 machine. I haven't seen these problem on > the NT machine. > > Any suggestions, pointers, or fixes will be greatly appreciated. > I saw an effect of random "freezing" non-cygwin console applications on startup/execution on win95 many times. Looks like win95 bug. -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Thu May 22 00:44:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Thu, 22 May 1997 00:44:00 -0000 Subject: v18 vs v17 compile times.... Message-ID: <01BC669C.7C655BC0@gater.krystalbank.msk.ru> Kevin Dahlhausen wrote: > >From this (statistically irrelevant) test, version 18 appears about 27% slower > than version 17.1. > > Somebody else already asked, but I did not see the answer to the question, "Is > it (reasonably) easy to build without or turn off symbolic links?" I've made some improvements in symlink checking and sent a patch to mr. Noer. My patch checks "system" file attribute symlinks must to have. It doesn't affect non-cygwin programs any way - they knows nothing about symlinks. -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From banders@switch.rockwell.com Thu May 22 00:44:00 1997 From: banders@switch.rockwell.com (Bart Anderson) Date: Thu, 22 May 1997 00:44:00 -0000 Subject: serial access Message-ID: <3383486E.6781@switch.rockwell.com> With this mixed bag of unix/dos/win calls, how should/may the serial ports be accessed. I've tried /dev/cua0 and com1 and com1:. I've tried using mount with bash to get com1: seen as /dev/cua0. Any help would be appreciated. If there are 2 methods with different library calls, I'd appreciate both just to see how it's done. Thanks, Bart - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Thu May 22 00:44:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Thu, 22 May 1997 00:44:00 -0000 Subject: problem with inner classes in template functions Message-ID: <01BC6697.5CAE1A60@gater.krystalbank.msk.ru> Oliver Glier wrote: > Dear peoply from the GNU-WIN32 project, > my name is Oliver Glier and I'm porting a C++ class library to > Windows. The code is already compilable by Borland C++, > Sun Sparc C++, and GCC 2.7 (EMX , Linux, and Sun). > My problem is that I would like to use your compiler (beta 1.8) , > but it doesn't accept the following construct: > It's not gnu-win32 specific bug, it's a bug of gcc 2.8.0 snapshot used in b18. You should send your bug report to gcc maintainers! -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From mh@mike.franken.de Thu May 22 00:44:00 1997 From: mh@mike.franken.de (Michael Hirmke) Date: Thu, 22 May 1997 00:44:00 -0000 Subject: more problems with b18 Message-ID: <6XGQQt$4pfB@mike.franken.de> Hi *, some more problems with Cygnus' CDK b18: 1.) chmod no lomger supports switches like s (system), h (hidden) and a (archive) since b18 ! Is this the behaviour we have to live with in the future or is it a bug ? 2.) If I try to backup with "tar -cvzf ...", I always get lots of messages "setgrent called, and it's not done yet" and also "././@LongLink" - what do they mean ? Thx and bye. Michael. -- Michael Hirmke | Telefon +49 (911) 557999 Georg-Strobel-Strasse 81 | FAX +49 (911) 557664 90489 Nuernberg | E-Mail mh@mike.franken.de | WWW http://minimike.franken.de/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From root@jacob.remcomp.fr Thu May 22 00:44:00 1997 From: root@jacob.remcomp.fr (root) Date: Thu, 22 May 1997 00:44:00 -0000 Subject: Justin's problem: what does the 1736 means? References: <199705211602.LAA21663@smtp.gte.net> Message-ID: > > What the heck am I doing wrong? When I compile programs and want them to > run by themselves without cygwin.dll I do a: > > gcc hello.c -o hello.exe -L/src/ming/lib -lcrtdll -lmingw32 -dll > > Then I can export this program to other machines and run it, but....when I > try to compile more complex programs I get errors like this: > > (c:\src\tf-35b4\src\tf.exe 1736) In cygwin_except_handler ^^^^ Maybe this 1736 is a line number? At least it looks like... Check that Am I correct Noer? -- Jacob Navia Logiciels/Informatique 41 rue Maurice Ravel Tel 01 48.23.51.44 93430 Villetaneuse Fax 01 48.23.95.39 France - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Thu May 22 00:44:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Thu, 22 May 1997 00:44:00 -0000 Subject: ncurses/termcap, just won't clear the screen? Message-ID: <01BC669A.52905A40@gater.krystalbank.msk.ru> Michael Condict wrote: > There's another problem, this one on NT. When it clears the screen it > sets the background to black, whether or not you've got that color as > your chosen background (a DOS window in NT can have many different colors > of text and background). Minor, but annoying. > I'll try to fix it. It is possible to read default console color attributes on first cygwin.dll startup and store it in a shared_info class. -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From ian@parti.inforum.org Thu May 22 00:44:00 1997 From: ian@parti.inforum.org (Ian) Date: Thu, 22 May 1997 00:44:00 -0000 Subject: X11R6.3 nolonger works under B18 References: <3.0.1.32.19970519100433.00785b18@mail.cs.unc.edu> Message-ID: <3.0.1.32.19970521225846.00826210@192.160.61.2> Is there somewhere where one could ftp an already working X11R6.3? I tried to build it myself (with Sergey Okhapkin's patch), but I don't have imake, and am probably missing something else needed to compile it myself :( (If not, the location of a working imake and other tools needed to compile X would of course be welcome) -- Ian Reinhart ian@parti.inforum.org - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From bill@dstc.edu.au Thu May 22 00:44:00 1997 From: bill@dstc.edu.au (Bill Segall) Date: Thu, 22 May 1997 00:44:00 -0000 Subject: Beta 18 breaks gettimeofday References: <3.0.32.19970520174814.02732ca8@silence.secnet.com> Message-ID: <199705220140.LAA12418@piglet.dstc.edu.au> I just hit this too although I'm running b17. It seems that it works but get the return code wrong. Bill. ---------------------------------------------------------------------- Bill Segall Senior Research Scientist phone : +61 7 3365 4310 DSTC Pty Ltd fax : +61 7 3365 4311 Level 7, Gehrmann Laboratories email : bill@dstc.edu.au The University of Queensland www : http://www.dstc.edu.au Queensland, 4072, Australia ---------------------------------------------------------------------- - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From *jeffdb@netzone.nospam.com Thu May 22 00:57:00 1997 From: *jeffdb@netzone.nospam.com (Mikey) Date: Thu, 22 May 1997 00:57:00 -0000 Subject: wcslen in string.h without windows.h Message-ID: <199705220424.VAA14164@nz1.netzone.com> Warning Could not process message with given Content-Type: multipart/mixed;boundary="----=_NextPart_000_01BC6539.D5276000" From home@morla.crec.mipt.ru Thu May 22 00:57:00 1997 From: home@morla.crec.mipt.ru (Igor V. Kovalenko) Date: Thu, 22 May 1997 00:57:00 -0000 Subject: Telnet and finger from NetKit-B-0.06 [was Diffs for ssh-1.2.20] References: <199705211244.FAA15147@nz1.netzone.com> Message-ID: <199705211800.WAA01196@morla.crec.mipt.ru> Mikey wrote: > > Do you have finger, and telnet posted somewhere? > > I tried the telnet client from b14 net.tar.gz, but I couldn't get it to handle passwords, > so if you could post them, or give a pointer, I'm sure everyone would appreciate it. > I put telnet and finger on URL http://www.crec.mipt.ru/~iko (That is a tiny home page of mine). Binaries works for me, so why don't you all try them? :) I have also put ncftp-2.4.2 binary that I have compiled under beta 17.1 with pdcurses. Sorry, I hadn't tried it on b18 yet. Ssh-1.2.20 client which is able to read passwords from /dev/tty is also at that URL. With best wishes, -- Iko Igor V. Kovalenko mailto:iko@crec.mipt.ru - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Thu May 22 00:57:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Thu, 22 May 1997 00:57:00 -0000 Subject: ncurses/termcap, just won't clear the screen? Message-ID: <01BC66A7.82F531D0@gater.krystalbank.msk.ru> Mikey wrote: > <> > The diff that I sent out before causes random crashes during compiles under win95, > apparently micros*** lied about the output chars stacking up see remarks below. I've already fixed clear screen problem on W95 almost the same way :-) What EOT character in output stream shold to do? Is output CP support really necessary? And why do you use ndelay_set variable instead of VMIN/VTIME fields in termios structure? I can upload patched cygwin.dll to some ftp site. Any ideas? -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From wester@ilt.fhg.de Thu May 22 01:47:00 1997 From: wester@ilt.fhg.de (Rolf Wester) Date: Thu, 22 May 1997 01:47:00 -0000 Subject: gcc does not works at all Message-ID: <3.0.1.32.19970522104554.005c3d40@pop.ilt.fhg.de> Hi, after installing gnuwin32 b18 gcc does not work at all. When trying to compile the small test programm: main() { int i=1,j=1,k; k=i+j; } with gcc -c test.c I get the error messages: heap_init: unable to allocate heap, win32 error 87 cygwin: terminating In cygwin_except_handler Exception trapped! exception C0000005 at 1001E4B2 exception: ax 0 bx cx 6 dx 7 exception: si 322E372E di 322D7375bp 241FB60 sp 241FB60 exception is: STATUS_ACCESS_VIOLATION Stack trace: .. on so on I would be very appreciative for any help on this. Thanks in advance Rolf - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From hgeleff@spidernet.com.cy Thu May 22 01:49:00 1997 From: hgeleff@spidernet.com.cy (Henrik Geleff) Date: Thu, 22 May 1997 01:49:00 -0000 Subject: SCCS References: <9705201533.AA06898@ss38.wg.icl.co.uk> Message-ID: <3384082F.72CF@spidernet.com.cy> Kevin Hughes wrote: > > I realise this is slightly off track BUT I have alot of s/w on Unix under SCCS > control and would like to port it all to run under win9x. I really would like > the same dev.env and control but I cannot find an sccs port (I realise rcs is > better but.....) If anyone knows of a port or the source ;-) I would be most > interested. > > Many thanks > Kevin > - > For help on using this list (especially unsubscribing), send a message to > "gnu-win32-request@cygnus.com" with one line of text: "help". SCCS is part of SVR4 -- and I think SVR3 as well -- so unless you want to license the source from AT&T :-(, I do not think you will find it. RCS which you mention has the option to import the SCCS files, so you can continue with out any inconvinience. And then you can use RCS or maybe even CVS on your UN*X environment as well, I would advise that -- unless somebody above you hinders it, that is. Regards Henrik Geleff - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From chin.cheekai@singapore.sterling.com Thu May 22 01:49:00 1997 From: chin.cheekai@singapore.sterling.com (Chin Chee-Kai) Date: Thu, 22 May 1997 01:49:00 -0000 Subject: A call to find mounted dirs? References: <199705211654.JAA13836@singapore.sterling.com> Message-ID: A question related to mount: I'm using b18 and have noticed that the mounts seem to survive across reboots. Ie., after mounting a directory to a mount point, shutting down and booting up again, "mount" shows the old *mounted* configuration instead of plain configuration. While this is convenient in my case, it's not the right Unix semantics. So is this a bug or feature? :) Chin Chee-Kai Internet Email: cheekai@singapore.sterling.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From ebritten@ea.oac.uci.edu Thu May 22 01:49:00 1997 From: ebritten@ea.oac.uci.edu (Eric Britten) Date: Thu, 22 May 1997 01:49:00 -0000 Subject: Standalone linking with minimalist References: <199705211602.LAA21663@smtp.gte.net> Message-ID: On Wed, 21 May 1997, Justin B. Harvey wrote: > What the heck am I doing wrong? When I compile programs and want them to > run by themselves without cygwin.dll I do a: > > gcc hello.c -o hello.exe -L/src/ming/lib -lcrtdll -lmingw32 -dll > If you are building an exe(hello.exe) you don't need the '-dll' option. Otherwise everything else looks fine. > Then I can export this program to other machines and run it, but....when I > try to compile more complex programs I get errors like this: > > (c:\src\tf-35b4\src\tf.exe 1736) In cygwin_except_handler > (c:\src\tf-35b4\src\tf.exe 1736) Exception trapped! > (c:\src\tf-35b4\src\tf.exe 1736) exception C0000005 at 776415CE > (c:\src\tf-35b4\src\tf.exe 1736) exception: ax FCC4FCA8 bx 0 cx 1D dx > FFFFFFFF > (c:\src\tf-35b4\src\tf.exe 1736) exception: si FCC4FCA8 di 6A78ECE0 bp > 242F3AC sp242F3A4 > (c:\src\tf-35b4\src\tf.exe 1736) exception is: STATUS_ACCESS_VIOLATION > (c:\src\tf-35b4\src\tf.exe 1736) Stack trace: > (c:\src\tf-35b4\src\tf.exe 1736) frame 0: sp = 0x242F1D8, pc = 0x1000CEC2 > (c:\src\tf-35b4\src\tf.exe 1736) frame 1: sp = 0x242F1F4, pc = 0x77F94512 > (c:\src\tf-35b4\src\tf.exe 1736) frame 2: sp = 0x242F218, pc = 0x77F88EEB > (c:\src\tf-35b4\src\tf.exe 1736) frame 3: sp = 0x242F2A4, pc = 0x77F76266 > (c:\src\tf-35b4\src\tf.exe 1736) frame 4: sp = 0x242F3AC, pc = 0x7763C15C > (c:\src\tf-35b4\src\tf.exe 1736) frame 5: sp = 0x242F3C8, pc = 0x419112 > (c:\src\tf-35b4\src\tf.exe 1736) frame 6: sp = 0x242F3E0, pc = 0x40D442 > (c:\src\tf-35b4\src\tf.exe 1736) frame 7: sp = 0x242F3F4, pc = 0x40D2EC > (c:\src\tf-35b4\src\tf.exe 1736) frame 8: sp = 0x242F43C, pc = 0x1000C102 > (c:\src\tf-35b4\src\tf.exe 1736) frame 9: sp = 0x242FF94, pc = 0x1000C113 > (c:\src\tf-35b4\src\tf.exe 1736) frame 10: sp = 0x242FFA0, pc = 0x41C40D > (c:\src\tf-35b4\src\tf.exe 1736) frame 11: sp = 0x242FFB0, pc = 0x40103B > (c:\src\tf-35b4\src\tf.exe 1736) frame 12: sp = 0x242FFC0, pc = 0x77F1B304 > (c:\src\tf-35b4\src\tf.exe 1736) frame 13: sp = 0x242FFF0, pc = 0x0 > (c:\src\tf-35b4\src\tf.exe 1736) End of stack trace > bash$ > > What is the correct syntax on a linkline for building standalone > executables without cygwin.dll? > Your syntax above was correct except for the -dll option that you added. The syntax is the same as building with cygwin32 except you need to specify the path to the mingw32 librarys in the environment variable LIBRARY_PATH or on the command line with -L. When using mingw32 make sure that gcc does not look in the directory that containes libcygwin32.a. One way to do this is locate the directory with a find utility, then remove the path to that directory from the LIBRARY_PATH environment variable. This way gcc and ld do not link the cygwin32.dll library to your application. In my environment when I use gcc and mingw32 I have my LIBRARY_PATH set to /cygnus/mingw32/lib:/cygnus/win32/lib where /cygnus/mingw32 contains all the Minimalist GNU Win32 files. I also placed all the win32 files in the win32 directory. This way it is easier to keep the cygwin and the stand alone libraries separate. It is obvious from your exception that the cygwin32 DLL is linked to your program. Hope this helps. Eric Britten > Frustrated, > > Justin > -- > Justin B. Harvey > jbharvey@gte.net > http://home1.gte.net/jbharvey > - > For help on using this list (especially unsubscribing), send a message to > "gnu-win32-request@cygnus.com" with one line of text: "help". > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sganly@lehman.com Thu May 22 01:49:00 1997 From: sganly@lehman.com (Steve Ganly) Date: Thu, 22 May 1997 01:49:00 -0000 Subject: BL18 Installation problems. References: <3382C62A.D53@sco.com> Message-ID: <9705220836.ZM12265@lonadmin226.lonadmin.lehman.com> Roy, I had *exactly* the same problem. The cdk.exe file appears to have been corrupt. Pulling down the file again made it work for me. Steve Ganly. On May 21, 10:53am, Rohinton Kazak wrote: > Subject: BL18 Installation problems. > Hi, > I have tried to install cdk.exe bl18 on both W95 and NT3.51 without > success. The archive extracts itself and around the 98% mark fails > with a message reporting that there is not enough free space in TEMP. > > I have 150Mb of free disk space, and the archive is 14Mb compressed. > I don't have access to the WIN-NT tools used to create this type of > archive so I cannot investigate it further... > > Any help appericated, > > Roy. > - > For help on using this list (especially unsubscribing), send a message to > "gnu-win32-request@cygnus.com" with one line of text: "help". >-- End of excerpt from Rohinton Kazak - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From esh124@email.psu.edu Thu May 22 01:49:00 1997 From: esh124@email.psu.edu (eric hilton) Date: Thu, 22 May 1997 01:49:00 -0000 Subject: ncurses/termcap, just won't clear the screen? References: <199705192156.QAA29407@bluestem.prairienet.org> Message-ID: <199705220753.DAA25252@r02n05.cac.psu.edu> > From: kunglao@prairienet.org > I've just rebuilt termcap & ncurses, that went fine. But now I'm > trying to run some of the test programs that ncurses has. And it > appears that it just cannot erase the screen. The curses is set to > the top of the window, but all the text is still there. I've tried a > few different things for my TERM env variable, linux, ansi, pc-ansi, > vt100, etc. They all behaved the same. Check your /etc/termcap file to see if you have the updated one. Basic term abilities seem to work now. I got rogue compiled and working just fine. (Although I didn't use ncurses, it comes with a toned down curses implementation) Try this: #include int main() { printf("\033[2J"); } Compile and run it, if it clears the screen your term has the properties of a proper terminal. BTW my $TERM is ansi. I set it in autoexec.bat. I seem to recall when I used linux term it didnt clear my screen either. -e - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From alex!@bigfoot.com Thu May 22 09:38:00 1997 From: alex!@bigfoot.com (Hmmmz) Date: Thu, 22 May 1997 09:38:00 -0000 Subject: ncurses/termcap, just won't clear the screen? Message-ID: <199705220232.TAA25993@mom.hooked.net> > With the new /etc/termcap from b18, I do an "export TERM=linux" > in bash, and then hit a control-l to clear. > > Under Windows NT, the right thing happens. Under Windows 95, however > I see the behavior kunglao@prairienet.org reports where the > prompt is moved but the old text doesn't get erased. Umm, this is great if you're just in bash, but what about programs that try to clear the screen like say NcFTP? And why did this work in B17.x? > -- > Geoffrey Noer > noer@cygnus.com WinNT and *nix rule, so learn from the past: Hiroshima '45, Vietnam '64, Tjernobyl '86, MacOS '96 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jpmorgan@unm.edu Thu May 22 09:38:00 1997 From: jpmorgan@unm.edu (James Paul Morgan) Date: Thu, 22 May 1997 09:38:00 -0000 Subject: problem with inner classes in template functions References: <3382BFCB.89F0591C@stud.uni-frankfurt.de> Message-ID: On Wed, 21 May 1997, Oliver Glier wrote: [Code SNIP of a class in the public area of a template class.] > > The compiler always yields a syntax error at the line with... [SNIP] > Still, I think that > code as above is correct C++. At least it works with a lot > of other C++ compilers and seems to be important in the > use of class libraries such as STL. > I've had similar problems with the g++ compiler under Cygwin32 17.1 with a class inside of a normal class. It was for a school project and compiled successfully on the g++ compiler on UNM's AIX machine. For some reason, Cygwin's g++ does not like a class within another class. My situation was that the the sub-class was declared inside the private portion of the main class. The sub-class was only used in the constructor of the main class and nowhere else. However, the compiler on Cygwin32 complained that the sub-class was private when it was accessed. (Yes, I know that declaring in a class in private that is only used in the constructor is very odd, but that's the kind of code this professor wrote and wanted us to use.) I have created other classes within classes in other school projects and seen it in text books, so I figure it's valid C++, but Cygwin doesn't always accept it. See ya around the Mulberry bush. --James :) - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From fjh@cs.mu.OZ.AU Thu May 22 09:38:00 1997 From: fjh@cs.mu.OZ.AU (Fergus Henderson) Date: Thu, 22 May 1997 09:38:00 -0000 Subject: problem with inner classes in template functions References: <3382BFCB.89F0591C@stud.uni-frankfurt.de> Message-ID: <199705220324.NAA29943@mundook.cs.mu.OZ.AU> Oliver Glier, you wrote: > > template > void createIterator(const C&) > { > C::Iterator it; > it.setnull(); > } ... > The compiler always yields a syntax error at the line with > "C::Iterator it;" in function createIterator(). The compiler is correct. The above is not valid C++ syntax (no matter what all those other compilers do). I think the correct syntax is typename C::Iterator it; ^^^^^^^^ This syntax is a fairly recent addition to C++, and I have no idea whether the cygwin g++ supports it yet (and no intention of rebooting my Linux box in W95 mode to find out ;-). -- Fergus Henderson | "I have always known that the pursuit WWW: < http://www.cs.mu.oz.au/~fjh > | of excellence is a lethal habit" PGP: finger fjh@128.250.37.3 | -- the last words of T. S. Garp. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Keetnet@wilmington.net Thu May 22 09:38:00 1997 From: Keetnet@wilmington.net (Keet) Date: Thu, 22 May 1997 09:38:00 -0000 Subject: rename() in b17.1 vs. b18 Message-ID: <2.2.32.19970522003654.00689090@wilmington.net> I reported earlier a problem with TinyMuck FuzzBall 5.55. It would generate an 'access denied' error when dumping a database using the '@dump' command. Well, i've narrowed it down to the rename() command in game.c of the Fb5.55 source (not that that would matter any). The following lines generate the error: if (rename(tmpfile, dumpfile) < 0) perror(tmpfile); the error context is './data/minimal.new.#2#: Access Denied' The only effect this error has on the program, is a failed database dump, which are definately important to this program. In b17.1 this problem never happened (it was a CR/LF problem - which i since fixed).. it seems i've traded one error for another. Any help on this matter would be appreciated. System Summary: Win95 P90 - 24 MB Ram 400+MB of HD Space - Greg Neujahr ( Keet - Foxbird) keetnet@wilmington.net - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jazz@softway.com Thu May 22 09:38:00 1997 From: jazz@softway.com (Jason Zions) Date: Thu, 22 May 1997 09:38:00 -0000 Subject: SCCS References: <9705201533.AA06898@ss38.wg.icl.co.uk> Message-ID: <33847403.55D2@softway.com> There is no public-domain SCCS code that I'm aware of; it's all under AT&T copyright. If someone has an SCCS clone source code base that's truly free or copylefted, I'd love to hear about it as well. Jason - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From colin@bird.fu.is.saga-u.ac.jp Thu May 22 10:22:00 1997 From: colin@bird.fu.is.saga-u.ac.jp (Colin Peters) Date: Thu, 22 May 1997 10:22:00 -0000 Subject: Standalone linking with minimalist Message-ID: <01BC66B5.61736F00@gbird0> Justin B. Harvey[SMTP:jbharvey@gte.net] wrote: >What the heck am I doing wrong? When I compile programs and want them to >run by themselves without cygwin.dll I do a: > >gcc hello.c -o hello.exe -L/src/ming/lib -lcrtdll -lmingw32 -dll Are you sure you meant to put -dll there? If you were using the mingw32 specs file (version 0.1.3 or higher) that would attempt to link your program as a DLL... which doesn't seem right. Aside from that, you also shouldn't have to explicitly link crtdll and mingw32 libraries on your link line, and doing it might lead to similar errors as compiling with kernel32 explicitly on your link line (well, crtdll might have that problem... mingw32 should be ok). This might be what's causing your trouble below. Check that the specs file from /src/ming/lib (? a guess based on your command line above) is being used by gcc. Add a -v to your command line and look for the "using specs from ..." line. >Then I can export this program to other machines and run it, but....when I >try to compile more complex programs I get errors like this: Can you run it on machines that have no cygwin.dll at all? From your below trace I think you might still have some cygwin references. >(c:\src\tf-35b4\src\tf.exe 1736) In cygwin_except_handler >(c:\src\tf-35b4\src\tf.exe 1736) Exception trapped! >(c:\src\tf-35b4\src\tf.exe 1736) exception C0000005 at 776415CE [snip: etc...] Note how it reports running into cygwin_except_handler... This stuff is all in cygwin.dll if I remember correctly, as is the code for setting up that exception handler in the first place. I suspect you are still linking Cygwin's crt0.o (which would happen if you were using the Cygwin specs instead of the Mingw32 specs), and that is still linking up to cygwin.dll. Try gcc -v to find out what specs file you're using. There are instructions in the readme for Mingw32 about moving the Cygwin specs file so that gcc won't find it unless you explicitly set up for Cygwin compiles (it is possible to have both setups live together, but you have to change environment variables in between). If you don't do this, and if you don't change your LIBRARY_PATH so that gcc finds the Mingw32 specs file, then it will be very difficult to get Mingw32 compiles to work properly. Colin. -- Colin Peters - colin@bird.fu.is.saga-u.ac.jp -- Saga University Dept. of Information Science -- http://www.fu.is.saga-u.ac.jp/~colin/index.html -- http://www.geocities.com/Tokyo/Towers/6162/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From *jeffdb@netzone.nospam.com Thu May 22 10:22:00 1997 From: *jeffdb@netzone.nospam.com (Mikey) Date: Thu, 22 May 1997 10:22:00 -0000 Subject: Bash is very slow to respond to keyboard input Message-ID: <199705212347.QAA17847@nz1.netzone.com> Hey give the poor thing a break, DOS, (whoops I mean win95) has been around since 197*, don't you think it's a lot to ask to be able to accept keyboard input, and do disk io at the same time? ;^) But seriously I have had the same problem since B14, I just figured it was 95's well known bad console handling If you didn't have the problem under b17.1, then you were very lucky, or you were on a very fast machine, I'm on a P100, and it's stuttering even as I type. (*jeffdb@netzone.com) the return address for this message is anti spammed remove * from the above address to reply. Mikey>After switching to B18, bash frequently ignores keyboard input when I am >using the cursor keys to edit a command line. I will be using the cursor >keys to backup quickly and the cursor will pause for awhile and then after >a few seconds it will spit out some ANSI escape codes where my cursor was. > >Possibly related to this is when man fires up less. Less will also tend to >get confused. It will ignore some of the keystrokes. After a bit of >struggle it will appear to disappear only to have the less prompt appear a >bit later. > >This is all happening on a WIN95 machine. I haven't seen these problem on >the NT machine. > >Any suggestions, pointers, or fixes will be greatly appreciated. > > >Steve Kachman >skachman@unl.edu - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From noer@cygnus.com Thu May 22 10:22:00 1997 From: noer@cygnus.com (Geoffrey Noer) Date: Thu, 22 May 1997 10:22:00 -0000 Subject: Justin's problem: what does the 1736 means? References: Message-ID: <199705221720.KAA18906@cirdan.cygnus.com> root wrote: > > > (c:\src\tf-35b4\src\tf.exe 1736) In cygwin_except_handler > ^^^^ > Maybe this 1736 is a line number? At least it looks like... Check that > Am I correct Noer? No, that's the cygwin PID of the process. -- Geoffrey Noer noer@cygnus.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From *jeffdb@netzone.nospam.com Thu May 22 18:33:00 1997 From: *jeffdb@netzone.nospam.com (Mikey) Date: Thu, 22 May 1997 18:33:00 -0000 Subject: A call to find mounted dirs? Message-ID: <199705230126.SAA25375@nz1.netzone.com> Can you say fstab? (*jeffdb@netzone.com) the return address for this message is anti spammed remove * from the above address to reply. Mikey ---- From: Chin Chee-Kai To: gnu-win32@cygnus.com Date: Thursday, May 22, 1997 5:08 PM Subject: Re: A call to find mounted dirs? >A question related to mount: I'm using b18 and have noticed >that the mounts seem to survive across reboots. Ie., after >mounting a directory to a mount point, shutting down and >booting up again, "mount" shows the old *mounted* configuration >instead of plain configuration. While this is convenient in >my case, it's not the right Unix semantics. So is this a >bug or feature? :) > > >Chin Chee-Kai >Internet Email: cheekai@singapore.sterling.com > >- >For help on using this list (especially unsubscribing), send a message to >"gnu-win32-request@cygnus.com" with one line of text: "help". > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jazz@softway.com Thu May 22 18:33:00 1997 From: jazz@softway.com (Jason Zions) Date: Thu, 22 May 1997 18:33:00 -0000 Subject: Fix to readline for case-insensitive file-name completion References: <3.0.1.32.19970520190749.006a6194@pop.primenet.com> <33831AD6.DD559635@pdi.com> Message-ID: <3384755E.7D47@softway.com> > I think it's even possible to have a case-sensitive > filesystem on a Win32 machine with an NFS client. NTFS is by nature case-sensitive; it's the Win32 API to it that is case-insensitive. The NT POSIX subsystem running on top of the same NTFS filesystem is case-sensitive. NT FAT filesystems are a little funnier. Although the filenames are stored in case-sensitive fashion, I'm not sure if all of the filesystem implementation code is careful to handle it correctly. Jason - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From azure@a29c.otanner.ton.tut.fi Thu May 22 18:33:00 1997 From: azure@a29c.otanner.ton.tut.fi (Hannu Koivisto) Date: Thu, 22 May 1997 18:33:00 -0000 Subject: ssh 1.2.20 diff and patch problems Message-ID: Greetings, I tried to apply the announced patch for ssh 1.2.20 to compile it with gnuwin32. However, patching (patch < ssh-1.2.20.diff) didn't succeed: Hmm... The next patch looks like a new-style context diff to me... The text leading up to this was: -------------------------- |E_GETRUSAGE) && defined(XXXJR) | { | struct rusage ru, cru; | getrusage(RUSAGE_SELF, &ru); |diff -cr ssh-1.2.20/servconf.c ssh-1.2.20.new/servconf.c |*** ssh-1.2.20/servconf.c Tue Apr 22 17:40:07 1997 -------------------------- Patching file ssh-1.2.20/servconf.c using Plan A... assertion "hunk" failed: file "/home/noer/src/beta18/cdk/patch/patch.c", line 33 9 I decided to try patching it in Linux. I moved the ssh archive and the diff file to our Linux server and ran patch < ssh-1.2.20.diff. The result was even more errors and failed hunks. Actually, almost all hunks failed with different error messages. Here's just one example: Patching file ssh-1.2.20/socketpair.c using Plan A... (Fascinating--this is really a new-style context diff but without the telltale extra asterisks on the *** line that usually indicate the new style...) Hunk #1 failed at 9. Hunk #2 succeeded at 67 with fuzz 2. What am I doing wrong? Thanks in advance. --- Hannu Koivisto | What you see is all you get. Prime jee | - Brian Kernighan ------------------------------------------------------------- - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jpmorgan@unm.edu Thu May 22 18:33:00 1997 From: jpmorgan@unm.edu (James Paul Morgan) Date: Thu, 22 May 1997 18:33:00 -0000 Subject: Split file downloads References: <3381C9BF.5767@h0013.ae.ge.com> Message-ID: On Tue, 20 May 1997, Carl Lawrence-Slater wrote: > Also has anybody loaded all of gnu-win32 on a Zip disk? How > bad is the performance over the parallel port? > I run 17.1 for Win95 on my Zip drive because I just don't have the hard drive space right now. The performance is not great. There is a noticible difference between running speed on any program that has to access the drive, plus the initial lag loading up. If you configure programs that create temp files, such as a full compile with g++, to use a temporary directory on your hard drive rather than the Zip drive, it helps. I have never clocked it specifically, so I can't give you any statistics. See ya around the Mulberry bush. --James :) - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From tphan@asl.dl.nec.com Thu May 22 18:33:00 1997 From: tphan@asl.dl.nec.com (Timothy Phan) Date: Thu, 22 May 1997 18:33:00 -0000 Subject: ncurses, curses, termcap, etc. Message-ID: <199705221713.MAA12289@aslws156.asl.dl.nec.com> Hi, I'd like to know where is ncurses, curses, termcap, etc? I was building the vile7-1 and at link time, gcc gave me 'undefined reference to `killpg', please help. BTW, what TERM should be set to? Thanks! make: [vile] Error 1 (ignored) gcc -g -O2 -o vile tcap.o main.o basic.o bind.o buffer.o crypt.o csrch.o displa y.o dumbterm.o eval.o exec.o externs.o fences.o file.o filec.o fileio.o finderr. o glob.o globals.o history.o input.o insert.o isearch.o itbuff.o lckfiles.o line .o map.o modes.o msgs.o npopen.o oneliner.o opers.o path.o random.o regexp.o reg ion.o search.o select.o spawn.o tags.o tbuff.o termio.o undo.o version.o window. o word.o wordmov.o -lcurses spawn.o: In function `bktoshell': /home/root/build/vile-7.1/spawn.c:132: undefined reference to `killpg' gcc: Internal compiler error: program ld got fatal signal 1 make: *** [vile] Error 1 -- -------------------- Timothy C. Phan (tphan@asl.dl.nec.com) -------------------- NEC America, Inc. ASL -------------------- 1525 Walnut Hill Ln. Irving, TX 75038 -------------------- tel: (214)-518-3437 fax: (214)-518-3499 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From *jeffdb@netzone.nospam.com Thu May 22 18:33:00 1997 From: *jeffdb@netzone.nospam.com (Mikey) Date: Thu, 22 May 1997 18:33:00 -0000 Subject: gcc does not works at all Message-ID: <199705230126.SAA25386@nz1.netzone.com> If you are doing this from command.com, It looks like you have an old copy of cygwin.dll somewhere in your path. If you are doing this from bash, I have no Idea what could be wrong, since you didn't include your platform, environment, or mount table From vassilii@optimedia.co.il Thu May 22 18:33:00 1997 From: vassilii@optimedia.co.il (Vassilii Khachaturov) Date: Thu, 22 May 1997 18:33:00 -0000 Subject: gnu-win32-specific bug in RCS 5.7 Message-ID: <3384906B@gandalf.optimedia.co.il> Thank you. It helped on NT, too. Consider passing the patch to the RCS bugs mailing list (rcs-bugs@cs.purdue.edu). --- conf.sh.orig Mon Mar 17 18:15:23 1997 +++ conf.sh Mon Mar 17 18:22:37 1997 @@ -1794,8 +1794,8 @@ EOF $PREPARE_CC || exit h=0 p=0 -if ($CL a.c $L && sh -c 'pid=$$; (sleep 3; kill $pid)& exec '$aout) >&2 -then +if ($CL a.c $L && sh -c 'pid=$$; (sleep 3; kill $pid)& sleep 4;exec '$aout) >&2 +then #extra sleep for fork bug h=1 $PREPARE_CC || exit $CS -Dhas_attribute_format_printf=1 a.c >&2 && $CS_OK && p=1 @@ -1890,10 +1890,11 @@ echo "$a#undef EXIT_FAILURE $z/* Uncomment this if EXIT_FAILURE is broken. */" : configuring large_memory -case "$has_map_fd$has_mmap" in -*1*) l=1;; -*) l=0 -esac +#case "$has_map_fd$has_mmap" in +#*1*) l=1;; +#*) l=0 +#esac +l=1 #force large_memory to overcome win95 short read bug echo "#define large_memory $l /* Can main memory hold entire RCS files? */" $ech >&3 "$0: configuring LONG_MAX $dots" --- rcstest.orig Mon Mar 17 19:14:23 1997 +++ rcstest Mon Mar 17 19:14:51 1997 @@ -173,7 +173,7 @@ case $USER in ?*) me=$USER;; *) - me=`who am i` || exit 2 + me=`whoami` || exit 2 me=`echo "$me" | sed -e 's/ .*//' -e 's/.*!//'` case $me in '') echo >&2 "$0: cannot deduce user name"; exit 2 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jp@awesome.nuancecom.com Thu May 22 18:33:00 1997 From: jp@awesome.nuancecom.com (J. P. Shipherd) Date: Thu, 22 May 1997 18:33:00 -0000 Subject: gcc does not works at all Message-ID: <199705230016.RAA22675@nuancecom.com> Rolf, This output generally means you're mixing b17 and b18 stuff somehow. Make sure the the b17 cygwin.dll is gone from your system and that none of the old b17 binaries are in your path. --jp > Hi, > > I get the error messages: > > heap_init: unable to allocate heap, win32 error 87 > cygwin: terminating > In cygwin_except_handler > Exception trapped! > exception C0000005 at 1001E4B2 > exception: ax 0 bx cx 6 dx 7 > exception: si 322E372E di 322D7375bp 241FB60 sp 241FB60 > exception is: STATUS_ACCESS_VIOLATION > Stack trace: > .. on so on > > I would be very appreciative for any help on this. > > Thanks in advance > > Rolf - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jazz@softway.com Thu May 22 18:33:00 1997 From: jazz@softway.com (Jason Zions) Date: Thu, 22 May 1997 18:33:00 -0000 Subject: Picking up include directories automatically References: <199705210042.RAA21924@nz1.netzone.com> Message-ID: <33847305.7DCD@softway.com> Mikey wrote: > The few programs that I've worked with that did their own path handleing > under linux "handled" // by assuming that they were dupes, and stripping > them out whenever they found them, begining middle or end, and since those > programs were being compiled on a platform that "aims" for posix compliance, > I ASS U MEd that // isn't posix, but under gnu-win32, it dosen't work > correctly for all programs, so DON'T USE IT. Nope; just bad programming. // is explicitly reserve in POSIX *precisely* to allow it to be used for things like drive letters (//C/whatever) or hostnames (Apollo Computer's DomainOS used //hostname/path to treat an entire network of systems like a single giant filesystem). Conforming applications should leave a leading // alone; 1003.1 is very clear on this. Three or more slashes can be collapsed to a single slash. The only question that cygwin needs to decide is if it wants to expose the DOS drive letter forest-of-trees *at all*. Might be smartest to simply stick with a single tree whose root is some directory on some drive; with a solid mount model and symbolic links, you can glue the thing together however you'd like. Personally, I would mount the root directories of the individual DOS drives under something like /drives and drop symbolic links elsewhere. (Perhaps I'm making a rash assumption; does cygwin support symlinks?) Jason - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From khan@xraylith.wisc.edu Thu May 22 19:00:00 1997 From: khan@xraylith.wisc.edu (Mumit Khan) Date: Thu, 22 May 1997 19:00:00 -0000 Subject: b18 C++ bugs [Re: problem with inner classes in template functions] References: <01BC6697.5CAE1A60@gater.krystalbank.msk.ru> Message-ID: <9705230129.AA10593@modi.xraylith.wisc.edu> Sergey Okhapkin writes: > Oliver Glier wrote: > > Dear peoply from the GNU-WIN32 project, > > my name is Oliver Glier and I'm porting a C++ class library to > > Windows. The code is already compilable by Borland C++, > > Sun Sparc C++, and GCC 2.7 (EMX , Linux, and Sun). > > My problem is that I would like to use your compiler (beta 1.8) , > > but it doesn't accept the following construct: > > > > It's not gnu-win32 specific bug, it's a bug of gcc 2.8.0 snapshot used in b18 > . You should send your bug report to gcc maintainers! > Not only this, this particular snapshot (and quite a few around it) also miscompiles stream code (when used in complex ways), making quite a bit of my code and octave (which I use to test various FSF snapshots) useless. I suspect (after many hours of tearing my hair out) that it has something to do with virtual base classes, but I've never been able to create a managable sized standalone program to demonstrate the bug, so I could be far off about the reason. The trick is to use the c++ front-end from b17.1 and use it with the rest of b18. Since I *only* compile on a Linux right now, that's no problem; if you're using '95/NT, then that would be a big problem (cygwin.dll version incompatibility). For those who may want to use cc1plus from b17.1, you will also have to add "-D__extension__=" to the specs file (or supply it every time). I supposed this is to be expected when basing a release on a snapshot, and especially when that "release" is a beta release itself ;-) Regards, Mumit -- khan@xraylith.wisc.edu http://www.xraylith.wisc.edu/~khan/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From chin.cheekai@singapore.sterling.com Thu May 22 19:00:00 1997 From: chin.cheekai@singapore.sterling.com (Chin Chee-Kai) Date: Thu, 22 May 1997 19:00:00 -0000 Subject: A call to find mounted dirs? References: <199705230126.SAA25375@singapore.sterling.com> Message-ID: On Wed, 21 May 1997, Mikey wrote: > Can you say fstab? Hi Mikey, not sure what you mean by "say fstab". If what you mean is that I had created /etc/fstab file that resulted in the remnant mounts, I have no such file created actually. C:\>dir \etc\fstab File not found C:\>mount Device Directory Type Flags \\.\tape1: /dev/st1 native no-mixed,text!=binary \\.\tape0: /dev/st0 native no-mixed,text!=binary \\.\b: /dev/fd1 native no-mixed,text!=binary \\.\a: /dev/fd0 native no-mixed,text!=binary c: / native no-mixed,text!=binary C:\>mount /usr/local/ /local <-- create /usr/local and /local before this command C:\>mount Device Directory Type Flags \usr\local /local native no-mixed,text!=binary \\.\tape1: /dev/st1 native no-mixed,text!=binary \\.\tape0: /dev/st0 native no-mixed,text!=binary \\.\b: /dev/fd1 native no-mixed,text!=binary \\.\a: /dev/fd0 native no-mixed,text!=binary c: / native no-mixed,text!=binary C:\> Now reboot and immediately pull out a DOS prompt to type mount again and you'll see the same display as the last list. (I did all commands outside bash, but don't suspect the result to be different if they were run inside bash). In b17.1, I didnt' see such "long-lived mounts" behaving in this manner though (which prompted me to put in b17.1 an alias "setup" to do all my required mounts the very first time I run bash after booting up. This alias now (in b18) returns "already mounted" messages when I run "setup", which means the "mount" semantics had changed). I am also not sure if this is due to b17.1 and b18 file interferences as the same behavior happened on another PC on which I installed b18 with a clean start without b17.1. Any ideas? Chin Chee-Kai Internet Email: cheekai@singapore.sterling.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From dahms@ifk20.mach.uni-karlsruhe.de Thu May 22 19:25:00 1997 From: dahms@ifk20.mach.uni-karlsruhe.de (dahms@ifk20.mach.uni-karlsruhe.de) Date: Thu, 22 May 1997 19:25:00 -0000 Subject: Justin's problem: what does the 1736 means? Message-ID: <009B4A4C.B749F900.15746@ifk20.mach.uni-karlsruhe.de> Hi Jacob, you wrote: : > (c:\src\tf-35b4\src\tf.exe 1736) In cygwin_except_handler : ^^^^ : Maybe this 1736 is a line number? At least it looks like... Check that No, it's the process id (Important for errors coming from forked childs). Bye, Heribert (dahms@ifk20.mach.uni-karlsruhe.de) - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From condict@opengroup.org Thu May 22 19:25:00 1997 From: condict@opengroup.org (Michael Condict) Date: Thu, 22 May 1997 19:25:00 -0000 Subject: Fix to readline for case-insensitive file-name completion References: <3.0.32.19970521074928.009bca20@pop.pdx.informix.com> Message-ID: <199705221509.LAA30863@postman.opengroup.org> In message < 3.0.32.19970521074928.009bca20@pop.pdx.informix.com >, you write: > Well just to show that this discussion isn't completely one sided :-) I'd > have to say that I'm quite in favour of case-insensitive file name completion. Thanks, I was beginning to feel that I'm the only one who wants it. I've now modified the shell file globbing as well as the filename completion, so they're both case insensitive. It's really nice to be able to say: ls make* and not worry about what case the file was stored in when you unzipped it or ftp'd it or are accessing it via a distributed file server. All of these ways of getting files tend to do unpredictable things to the upper/lower case. I also discovered a quaint fact about GNU ls and the layout of the NT file system: try "ls -U". It causes ls not to sort files, but print them in the order that they're stored in the directory. Well it turns out that NT file systems (but not DOS FAT file systems) are stored in case-insensitive alphabetic order! So "ls -U" prints the files in the this order, which is exactly what I want. I have some questions for those of you who like case sensitivity: Do you get irritated or confused by the index in the back of books, which is printed in case-insensitive order? Or how about when you look up words in the dictionary? Are "House" (at the beginning of a sentence) and "house" (in the middle of a sentence) two different words? I ask this as a UNIX lover and systems programmer for more years than most of you. It's one of the few things that UNIX and C got wrong. > Having moved to NT quite a while ago, I've got used to the case-ignorance > of the file system and as I use the cygwin tools to enhance NT rather than > simply mimic UNIX I have absolutely no problem if there are features here > that are only really going to be of interest under NT. I concur on this. I guess this is why you and I want case-insensitivity. We are interested in using native NT (not covering it up with a complete UNIX emulation layer) and merely want some better non-GUI tools for it, like a decent find / grep and an intelligent shell that is more user-friendly then the NT/95 command processor (which can't even handle "cd //host/sharename/...") > Perhaps this behaviour could be dependant upon the value of a shell > variable or setting? It looks like an environment variable is the way to go. I admit that when I tried out my changes on a networked UNIX file system, in which I had two side-by-side directories named "jade" and "JADE", it was suboptimal. When I typed "cd ja", my version of file completion modified the name to "JADE" and beeped at me because there was more than one matching name. This is fine if it picks the right version of two matching files, but it doesn't prefer to match the one that is in the same case that you typed -- it just picks the first file that matches what you typed so far, and modifies your case to that name. So, while I could fix this by making the code always try a case-sensitive match first, then revert to case-insensitive only if nothing matches, if might be better to be able to dynamically turn on case-sensitivity when perusing particular file systems or directories. Michael Condict m.condict@opengroup.org The Open Group Research Inst. (617) 621-7349 11 Cambridge Center Cambridge, MA 02142 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From *jeffdb@netzone.nospam.com Thu May 22 19:25:00 1997 From: *jeffdb@netzone.nospam.com (Mikey) Date: Thu, 22 May 1997 19:25:00 -0000 Subject: Missing proto for wcslen in installed headers + automaking dll's corrected. Message-ID: <199705212309.QAA14035@nz1.netzone.com> Title pretty much says it all. I did find a reference in .../include/g++/std/straits.h and it is in cygwin.din, and misc.cc did I miss something? diff file for .../H-i386-cygwin32/i386-cygwin32/include/string.h and .../cdk/newlib/libc/include/string.h attached if anyone needs it. also the advice I put in fixup.c was incorrect, I hadn't actuall make the change yet, fixup.o can't be used as an endfile it must go on the *startfile line. in the specs file. see attached. (*jeffdb@netzone.com) the return address for this message is anti spammed remove * from the above address to reply. Mikey -------------- next part -------------- A non-text attachment was scrubbed... Name: fixup.c Type: text/x-c Size: 719 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: specsB18.dif Type: text/x-diff Size: 345 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: string.h.dif Type: text/x-diff Size: 329 bytes Desc: not available URL: From ggp@informix.com Thu May 22 19:25:00 1997 From: ggp@informix.com (Guy Gascoigne - Piggford) Date: Thu, 22 May 1997 19:25:00 -0000 Subject: Fix to readline for case-insensitive file-name completion Message-ID: <3.0.32.19970522083432.00904290@pop.pdx.informix.com> At 11:08 AM 5/22/97 -0400, Michael Condict wrote: >In message < 3.0.32.19970521074928.009bca20@pop.pdx.informix.com >, you write: >> Well just to show that this discussion isn't completely one sided :-) I'd >> have to say that I'm quite in favour of case-insensitive file name completion. > >Thanks, I was beginning to feel that I'm the only one who wants it. I've >now modified the shell file globbing as well as the filename completion, >so they're both case insensitive. It's really nice to be able to say: > > ls make* > >and not worry about what case the file was stored in when you unzipped >it or ftp'd it or are accessing it via a distributed file server. >All of these ways of getting files tend to do unpredictable things to >the upper/lower case. Being fairly new to this group I have a question. Is it expected that changes like this will get folded back into the main source distribution or do they remain idividual efforts? Will you have to apply the same patches on the next release etc.? How are decisions like this made? Is it simply a consensus and then cygnus takes the best ideas/suggestions and then merges them in when they get the opportunity? >I also discovered a quaint fact about GNU ls and the layout of the NT >file system: try "ls -U". It causes ls not to sort files, but print >them in the order that they're stored in the directory. Well it turns >out that NT file systems (but not DOS FAT file systems) are stored in >case-insensitive alphabetic order! So "ls -U" prints the files in the >this order, which is exactly what I want. I like it - now to change my aliases :-) >> Perhaps this behaviour could be dependant upon the value of a shell >> variable or setting? > >It looks like an environment variable is the way to go. I admit that >when I tried out my changes on a networked UNIX file system, in which >I had two side-by-side directories named "jade" and "JADE", it was >suboptimal. When I typed "cd ja", my version of file completion >modified the name to "JADE" and beeped at me because there was more than >one matching name. This is fine if it picks the right version of two >matching files, but it doesn't prefer to match the one that is in the >same case that you typed -- it just picks the first file that matches >what you typed so far, and modifies your case to that name. So, while I >could fix this by making the code always try a case-sensitive match >first, then revert to case-insensitive only if nothing matches, if might >be better to be able to dynamically turn on case-sensitivity when >perusing particular file systems or directories. True, also as you said, if there are two possible matches differentiated by case then it should go for the case matching one first. That'll cut down on the number of times that the option will need to be toggled. Thanks for the follow up - Guy -- Guy Gascoigne - Piggford (ggp@informix.com) Software Engineer, Informix Software, Inc. (Portland, Oregon) (503) 225-4574, Fax (503) 221-2633 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From *jeffdb@netzone.nospam.com Thu May 22 19:25:00 1997 From: *jeffdb@netzone.nospam.com (Mikey) Date: Thu, 22 May 1997 19:25:00 -0000 Subject: ncurses/termcap, just won't clear the screen? Message-ID: <199705220304.UAA06282@nz1.netzone.com> The diff that I sent out before causes random crashes during compiles under win95, apparently micros*** lied about the output chars stacking up see remarks below. The corrected diff is attached.The WriteConsoleOutputCharacter function copies a number of characters to consecutive cells of a console screen buffer, beginning at a specified location. BOOL WriteConsoleOutputCharacter( HANDLE hConsoleOutput, // handle to a console screen buffer LPCTSTR lpCharacter, // pointer to buffer to write characters from DWORD nLength, // number of character cells to write to COORD dwWriteCoord, // coordinates of first cell to write to LPDWORD lpNumberOfCharsWritten // pointer to number of cells written to ); Parameters (b) hConsoleOutput Identifies the screen buffer. The handle must have GENERIC_WRITE access. (c) lpCharacter Points to a buffer that contains the characters to write to the screen buffer. (d) nLength Specifies the number of screen buffer character cells to write to. (e) dwWriteCoord Specifies the column and row coordinates of the first cell in the screen buffer to write to. (f) lpNumberOfCharsWritten Points to a 32-bit variable that receives the number of characters actually written. Return Values If the function succeeds, the return value is nonzero. If the function fails, the return value is zero. To get extended error information, call GetLastError. Remarks If the number of characters to be written to extends beyond the end of the specified row in the screen buffer, characters are written to the next row. If the number of characters to be written to extends beyond the end of the screen buffer, characters are written up to the end of the screen buffer. The attribute values at the positions written to are not changed. Windows NT: This function uses either Unicode characters or 8-bit characters from the console's current codepage. The console's codepage defaults initially to the system's OEM codepage. To change the console's codepage, use the SetConsoleCP or SetConsoleOutputCP functions, or use the chcp or mode con cp select= commands. (*jeffdb@netzone.com) the return address for this message is anti spammed remove * from the above address to reply. Mikey -------------- next part -------------- A non-text attachment was scrubbed... Name: current.dif Type: text/x-diff Size: 13189 bytes Desc: not available URL: From *jeffdb@netzone.nospam.com Thu May 22 19:25:00 1997 From: *jeffdb@netzone.nospam.com (Mikey) Date: Thu, 22 May 1997 19:25:00 -0000 Subject: ncurses/termcap, just won't clear the screen? Message-ID: <199705221636.JAA06612@nz1.netzone.com> Sergey exactly which bugs are you talking about? Fixing the Makefile so that it actually cleans the subdirectories? and lets you install just the dll instead of 20 minutes for headers and all? Invisible text text that shows up on your screen. (duh ;^) The delete key should IMHO delete forward by default, at the bash prompt not require the user to know that bind '"\e[3~":delete-char' is necessary to get his keyboard to work properly, or that he has to use D. Something similar to this should probably be done for the home and end keys too. (maybe this is a case of programers preference, but I don't like support calls that waste my time with why dosen't XXX key work call me funny that way ;^) The linux termcap/info is broken for full screen access on win95, so untill someone can tell me how to get rid of the extra ^M in wrapped lines, w32ansi will continue to be needed, I happen to like elvis. \[J, and \[0J are broken on win95, \E[1J is missing for linux support, it may not be documented in ncurses, but it sure is there in the linux kernel, at least through 2.0.29, so if you're going to tell people they can use a linux termcap don't you think you ought to support it? Code page support may not be top priorty, but the escape chars should at least be eaten , not just give BAD ESCAPE messages. BAD ESCAPE by itself isn't a very informative error message, why not tell people the most likely reason? info dosen't work, without some O_NDELAY implementation, so untill sombody wants to take a crack at FILE_FLAG_OVERLAPPED, this is the best that I, working from Tony Fitzpatrick's example could do, and since the reads for info.exe are done by FakeReadFile something similar to this is probably going to be necessary anyway. When you redirect stdout to a disk file, you WANT all the ctrl chars, at least I do. ;^) Geoff just told us that setpwent is the correct name for setpwend! tgetent() dosen't belong in the dll, unless you are going to provide a full termcap implmentation. and the fix for Sleep was from YOU!!! So what am I fixing that dosen't need fixing? (*jeffdb@netzone.com) the return address for this message is anti spammed remove * from the above address to reply. Mikey>Mikey wrote: >> <><> >> Yeah, here it is. but use ansi, or w32ansi for less elvis info etc, the linux termcap/info is (at least on my system) broken for >> full screen access. > >This diff contains some bugs fixed in b18 :-) > >-- >Sergey Okhapkin >Moscow, Russia >Looking for a job. > > >- >For help on using this list (especially unsubscribing), send a message to >"gnu-win32-request@cygnus.com" with one line of text: "help". > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From *jeffdb@netzone.nospam.com Thu May 22 19:25:00 1997 From: *jeffdb@netzone.nospam.com (Mikey) Date: Thu, 22 May 1997 19:25:00 -0000 Subject: ncurses/termcap, just won't clear the screen? Message-ID: <199705230028.RAA20394@nz1.netzone.com> > >Mikey wrote: >> Sergey exactly which bugs are you talking about? > >Linefeed code (case DWN:) in write_normal(), for examle. Granted rearanging DWN wasn't necessary, but I was trying to fix the extra ^M from linux termcap, and I forgot to take out the change before I made the diff, it dosen't hurt anything, kind of like the yeast in beer :^). > >> >> Fixing the Makefile so that it actually cleans the subdirectories? >> and lets you install just the dll instead of 20 minutes for headers and >all? > >I just copy/rename new-cygwin.dll to the target directory :-) Of course, >this fix is useful, but I am not winsup maintainer :-) > >> >> Invisible text text that shows up on your screen. (duh ;^) >> > >I added these lines to console.cc code and sent a patch to GN. > >> The delete key >> should IMHO delete forward by default, at the bash prompt >> not require the user to know that bind '"\e[3~":delete-char' >> is necessary to get his keyboard to work properly, or that he has to use >D. > >Ctrl-D is the end-of-file on unix systems. Should we change this behavior? Yes. (this was the most annoying thing for me when I first started using linux, I hit the delete key, like I've been doing forever, and the stupid console sits there and beeps at me, and puts a ~ on the screen besides ;^) I'm sorry to disagree with you on this, but ^d is the settable console EOF, a proper stty implementation, would allow you to set this as anything you want (barring conflicts), I think it would be more usefull given the underlying platform , and the fact that some (uninformed) people are going to want to use text mode mounts to have the dll, and stty set ^Z as default EOF, since anyone in text mode is going to have it interpreted that way anyway. > >> Something similar to this should probably be done for the home and end >keys too. >> (maybe this is a case of programers preference, >> but I don't like support calls that waste my time with >> why dosen't XXX key work call me funny that way ;^) > >I think, gnu-win32 distribution must have proper inputrc file and installer >must to set up INPUTRC environment variable properly to avoid beginner's >problems. How about .profile .bashrc .netrc etc................... in other words gnu-win32 needs an /etc/skel/ directory and a login that would read the users, and groups from the regristry and create apropriate $HOME/. dirs/files if they don't already exist, copy the skel/.files and probably an adduser utility that would add them to /etc/passwd /etc/group and to the regristy? That's a LOT of work and maintaining Two complete sets of startup and admin files, for every user is about as close to a sysadmins worst nightmare as you can come. $HOME could possibly be interpreted as the users Profile/ directory, IF user profiles are enabled, but I think it would be much simpler in the long run to have $HOME and the /etc/ dir end up virtual, like /dev/tty and the /etc/files .profile .bashrc even termcap are going to have to end up in somewhere the registry. If you want any kind of acceptance for this by the people that really count all of this kind of configuration stuff is going to have to happen inside the dll completly transparently, otherwise they are going to say sorry it's too complicated, I don't want it on my disk. (remember these are windows user's you are trying to attract ;^) Even now, there should be a login utility in the dll, that would disallow startup if the user dosen't have an authorized windows password in the registry. > >> >> The linux termcap/info is broken for full screen access on win95, so >untill someone can tell me how to >> >> get rid of the extra ^M in wrapped lines, w32ansi will continue to be >needed, I happen to like elvis. Sorry, I should have said dll linux termcap/info support. BTW did you know you can get a real nice termcap with head -n 1535 .../ncurses/misc/terminfo.src|grep -v "^#" > terminfo.src.new tic -C -RBSD terminfo.src.new > termcap and it's a lot less work? > >Termcap is not broken. Some support routines in console.cc fails on Win95. >I've fixed it today. BTW, what do you think about national keyboard support >bug? Is AltGr left or right Alt key on a keyboard? It's easy to fix this >bug, but I don't want to loose Alt-key to esc-key translation... I don't think it really matters which alt key you use for altGr, as long as it's documented somewhere, but since (I think) the majority of the keyboards in the world have the keypad on the right, I would suggest the using left alt > >> >> >> \[J, and \[0J are broken on win95, >> Maybe it's fixed to you, and Geoff, but untill it gets posted, It ain't fixed for nobody else, which is why I posted it. > >Fixed. > >> \E[1J is missing for linux support, it may not be documented in ncurses, >but it sure is there >> in the linux kernel, at least through 2.0.29, so if you're going to tell >people >> they can use a linux termcap don't you think you ought to support it? >> > >Added. > >> Code page support may not be top priorty, but the escape chars should at >least be eaten >> , >> not just give BAD ESCAPE messages. > >May be, but not for a now. We need now a full diagnostic of all >unimlemented but required features. > >> info dosen't work, without some O_NDELAY implementation, so untill >sombody >> wants to take a crack at FILE_FLAG_OVERLAPPED, this is the best that I, >working from >> Tony Fitzpatrick's example >> could do, and since the reads for info.exe are done by FakeReadFile >something >> similar to this is probably going to be necessary anyway. >> > >It's a useful, but a temporary O_NDELAY implementation. I want to implement >full posix tty/pty support including job control, process and session >groups, control tty, etc in the nearest future :-) That will be REALLY nice, but but untill it actually happens, new users that see 6 Megabytes of installed man pages, and 4 Megabytes of info files, are probably wondering why they are there since they can't read them don't you think? > >> Geoff just told us that setpwent is the correct name for setpwend! >> > >He is not right. Looks like setpassent must be named setpwent. I forgot to >write him about it :-( > >> tgetent() dosen't belong in the dll, unless you are going to provide a >full termcap >> implmentation. > >Yes, you are right. > >> >> and the fix for Sleep was from YOU!!! >> > >:-) The diff to B18 sources I sent to Geoff today is more than 40K in size >:-) It contains some new features and symlinks speed up. I asked him to >make patched winsup sources and precompiled cygwin.dll publically available >as a beta 18.1. > >-- >Sergey Okhapkin >Moscow, Russia >Looking for a job (*jeffdb@netzone.com) the return address for this message is anti spammed remove * from the above address to reply. Mikey - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From nfeller@ccc.isd.csc.com Thu May 22 19:25:00 1997 From: nfeller@ccc.isd.csc.com (Neal Feller) Date: Thu, 22 May 1997 19:25:00 -0000 Subject: (no subject) Message-ID: <3383593D.6D4A@ccc.isd.csc.com> I downloaded cdk.exe from the cygnus site and got freecell. Not what I expected. What happened? nfeller@ccc.isd.csc.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Thu May 22 19:25:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Thu, 22 May 1997 19:25:00 -0000 Subject: X11R6.3 nolonger works under B18 Message-ID: <01BC66D5.9E43C540@gater.krystalbank.msk.ru> Arlindo da Silva wrote: > Sergey, > > I installed your new cygwin32.rules, compiled everything, but > I am having the sample problem as GB: pop window with: Unable to run... > Which clients did you use to test your installation? > Bitmap.exe, for example. I can't understand the cause of your (and the others) problem :-( I'll try to install win95 and cygnus tools from the scratch and to compile/run X11... BTW, are you using original Windows 95 (like I am) or OSR2? -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From ggp@informix.com Thu May 22 19:25:00 1997 From: ggp@informix.com (Guy Gascoigne - Piggford) Date: Thu, 22 May 1997 19:25:00 -0000 Subject: More on case insensitivity Message-ID: <3.0.32.19970522084806.0096be10@pop.pdx.informix.com> After more thinking about this there are quite a few places where the case insensitivity of the file system rears it's ugly head. What's worse they can crop up fairly rarely, which means that I can forget to deal with them wich leads to getting unexpected results every now and again. e.g Wildcard expansion in the shell would need to be case insensitive, having to do things like: ls *.[Zz][Ii][Pp] simply to get this list becomes a pain. Likewise programs such as find that do their own wildcard expansion (perhaps they actually use the same library - I'm not sure, I don't have the code here) also need to be made aware of this. Again, this is a pain. $ find -name "*.[Zz][iI][Pp]" -maxdepth 1 ./Memoryan.zip ./Perf.zip ./SNIP9510.ZIP ./TABJAZZ.ZIP ./TABS.ZIP ./TCSAMP.ZIP ./TZ30.ZIP I suspect that there are others as well. Guy -- Guy Gascoigne - Piggford (ggp@informix.com) Software Engineer, Informix Software, Inc. (Portland, Oregon) (503) 225-4574, Fax (503) 221-2633 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From chin.cheekai@singapore.sterling.com Thu May 22 19:25:00 1997 From: chin.cheekai@singapore.sterling.com (Chin Chee-Kai) Date: Thu, 22 May 1997 19:25:00 -0000 Subject: A call to find mounted dirs? Message-ID: Sorry if this is a double-post -- I got a rebounced mail so trying to repost again. Chin Chee-Kai Internet Email: cheekai@singapore.sterling.com Sterling Software Asia Pacific Group ------------------------------------------------------------ On Wed, 21 May 1997, Mikey wrote: > Can you say fstab? Hi Mikey, not sure what you mean by "say fstab". If what you mean is that I had created /etc/fstab file that resulted in the remnant mounts, I have no such file created actually. C:\>dir \etc\fstab File not found C:\>mount Device Directory Type Flags \\.\tape1: /dev/st1 native no-mixed,text!=binary \\.\tape0: /dev/st0 native no-mixed,text!=binary \\.\b: /dev/fd1 native no-mixed,text!=binary \\.\a: /dev/fd0 native no-mixed,text!=binary c: / native no-mixed,text!=binary C:\>mount /usr/local/ /local <-- create /usr/local and /local before this command C:\>mount Device Directory Type Flags \usr\local /local native no-mixed,text!=binary \\.\tape1: /dev/st1 native no-mixed,text!=binary \\.\tape0: /dev/st0 native no-mixed,text!=binary \\.\b: /dev/fd1 native no-mixed,text!=binary \\.\a: /dev/fd0 native no-mixed,text!=binary c: / native no-mixed,text!=binary C:\> Now reboot and immediately pull out a DOS prompt to type mount again and you'll see the same display as the last list. (I did all commands outside bash, but don't suspect the result to be different if they were run inside bash). In b17.1, I didnt' see such "long-lived mounts" behaving in this manner though (which prompted me to put in b17.1 an alias "setup" to do all my required mounts the very first time I run bash after booting up. This alias now (in b18) returns "already mounted" messages when I run "setup", which means the "mount" semantics had changed). I am also not sure if this is due to b17.1 and b18 file interferences as the same behavior happened on another PC on which I installed b18 with a clean start without b17.1. Any ideas? Chin Chee-Kai Internet Email: cheekai@singapore.sterling.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Valeri.Faine@cern.ch Fri May 23 01:36:00 1997 From: Valeri.Faine@cern.ch (Valery Fine) Date: Fri, 23 May 1997 01:36:00 -0000 Subject: serial access References: <3383486E.6781@switch.rockwell.com> Message-ID: <199705221301.PAA99020@sp063.cern.ch> On 21 May 97 at 14:09, Bart Anderson wrote: > With this mixed bag of unix/dos/win calls, how should/may the serial > ports be accessed. I've tried /dev/cua0 and com1 and com1:. I've > tried using mount with bash to get com1: seen as /dev/cua0. Any help > would be appreciated. If there are 2 methods with different library > calls, I'd appreciate both just to see how it's done. At least for WIN32 API it is done as follows: 8.4. Work with serial lines. 8.4.1. Initialzation. #ifndef WIN32 serial_port = open("/dev/tty00", O_RDWR); /* ------------------- Set paprameters for serial port /dev/tty00 -------- */ rc = system("stty -f /dev/tty00 tab0 bs0 vt0 ff0 -isig -icanon -iexten"); rc = system("stty -f /dev/tty00 -echo -echoe -echok -echonl -noflsh"); #else /* ------------------- Set paprameters for serial port com2 -------- */ { DCB Com2Dcb; COMMTIMEOUTS TimeOut; int exterr; serial_port = CreateFile("com2:", // pointer to name of the file GENERIC_READ | GENERIC_WRITE, // access (read-write) mode 0, // share mode NULL, // pointer to security descriptor OPEN_EXISTING, // how to create FILE_FLAG_NO_BUFFERING, // file attributes NULL // handle to file with attributes to copy ); if ( serial_port == INVALID_HANDLE_VALUE) { exterr = GetLastError(); printf(" Can not open com port . err code= %x \n", exterr); return exterr; } GetCommState(serial_port, &Com2Dcb); if (Com2Dcb.BaudRate != 9600) Com2Dcb.BaudRate = 9600; // current baud rate if (Com2Dcb.Parity) Com2Dcb.Parity = NOPARITY; // 0-4=no,odd,even,mark,space if (Com2Dcb.StopBits != 1) Com2Dcb.StopBits = ONESTOPBIT; // 0,1,2 = 1, 1.5, 2 if (Com2Dcb.ByteSize != 8) Com2Dcb.ByteSize = 8; // number of bits/byte, 4-8 SetCommState(serial_port, &Com2Dcb); // Timeout = (MULTIPLIER * number_of_bytes) + CONSTANT TimeOut.ReadIntervalTimeout = 0; TimeOut.ReadTotalTimeoutMultiplier = 0; TimeOut.ReadTotalTimeoutConstant = 0; TimeOut.WriteTotalTimeoutMultiplier = 0; TimeOut.WriteTotalTimeoutConstant = 0; SetCommTimeouts(serial_port, &TimeOut); } rc = 0; #endif 8.4.2. Error handling If ReadFile/WriteFile WIN32 API return the code TRUE and the nBytesToBeSent (3d par) is more the nByteSent (4th par)this means the Time Out event has been occured. ZERO TimeOut value means "No Time Out Events" ================================================================= Dr. Valery Fine Telex : 911621 dubna su ----------- LCTA/Joint Inst.for NuclearRes Phone : +7 09621 6 40 80 141980 Dubna, Moscow region Fax : +7 09621 6 51 45 Russia mailto:fine@main1.jinr.dubna.su Dr. Valeri Faine ------------ Phone: +41 22 767 6468 CERN FAX : +41 22 767 7910 CH-1211 Geneva, 23 mailto:fine@mail.cern.ch Switzerland http://nicewww.cern.ch/~fine - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From gpasa@swissonline.ch Fri May 23 01:36:00 1997 From: gpasa@swissonline.ch (Pasa Guglielmo) Date: Fri, 23 May 1997 01:36:00 -0000 Subject: is patch working Message-ID: <33856A23.7B76@swissonline.ch> Hi, I'using patch version 2.1 (under cygwin32 17.1)but applying a patch to less with patch < less.diff (patch -p0 ...) give the same result I get modifications in output.c as it should but which are completely disorganised and with a lot of @. what's wrong with that ? -- Sincerely, Pasa Guglielmo ------------------------------------------------------------------ & e-mail : gpasa@swissonline.ch & tel.: +41 (0)24 485 50 40 & & mailing : Pasa Guglielmo & fax : +41 (0)24 485 50 44 & & Rte des Cases 17A & & & CH-1890 St-Maurice & & & (Switzerland) & & ------------------------------------------------------------------ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Keetnet@wilmington.net Fri May 23 01:36:00 1997 From: Keetnet@wilmington.net (Keet) Date: Fri, 23 May 1997 01:36:00 -0000 Subject: PROBLEM: rename() in b17.1 vs. b18 Message-ID: <2.2.32.19970522231746.00673320@wilmington.net> I reported earlier a problem with TinyMuck FuzzBall 5.55. It would generate an 'access denied' error when dumping a database using the '@dump' command. Well, i've narrowed it down to the rename() command in game.c of the Fb5.55 source (not that that would matter any). The following lines generate the error: if (rename(tmpfile, dumpfile) < 0) perror(tmpfile); the error context is './data/minimal.new.#2#: Access Denied' The only effect this error has on the program, is a failed database dump, which are definately important to this program. In b17.1 this problem never happened (it was a CR/LF problem - which i since fixed).. it seems i've traded one error for another. Any help on this matter would be appreciated. System Summary: Win95 P90 - 24 MB Ram 400+MB of HD Space - Greg Neujahr ( Keet - Foxbird ) keetnet@wilmington.net - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From dmg@arteque.com Fri May 23 01:36:00 1997 From: dmg@arteque.com (David Goodine) Date: Fri, 23 May 1997 01:36:00 -0000 Subject: g++ Compiler wedges on simple source error (B18) Message-ID: <3384A041.25CE@arteque.com> The g++ compiler seems to get wedged when it encounters this error. It's really an error in my code, but I have to kill the compiler process manually... it looks like cclplus.exe is what's wedging. I'm using Beta 18, here's the output... g++ -c -I/OSpace/include -g HashTable.cpp HashTable.cpp:66: `default_size' was not declared in this scope (C:\gnuwin32\b18\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32\cygnus-2.7.2-970404\cc1plus.exe 1060) In cygwin_except_handler David Goodine dmg@arteque.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From condict@opengroup.org Fri May 23 01:36:00 1997 From: condict@opengroup.org (Michael Condict) Date: Fri, 23 May 1997 01:36:00 -0000 Subject: Picking up include directories automatically References: <199705210042.RAA21924@nz1.netzone.com> Message-ID: <199705212123.RAA13147@postman.opengroup.org> In message < 199705210042.RAA21924@nz1.netzone.com >, you write: > How about putting export /c/dir because you mounted c:\\ to /c > which was my point in the first place YOU DON'T NEED // Exactly. Or you could just use "export /dir", if you've mounted c: under /, which is the default, after all. But the mount command doesn't mind that you mount c: twice, once under / and again under /c. I've done this for consistency. Also, whenever I want to refer to a networked drive, I mount it somewhere first. Usually I just mount the drive letter I've mapped it to: mount g: /g In any case, I agree 100% with you that "//" is dangerous because any UNIX program that notices it will think it's equivalent to "/" and may feel free to delete one of the slashes. For similar reasons "c:/..." paths are dangerous. Unix programs will assume that this is a relative pathname, not an absolute one, and may even create files in the wrong place. The other reason I never use "//" or "c:/" paths is that bash can't do file-name completion on such names. This is a fixable bug, I guess, but why bother. Michael Condict m.condict@opengroup.org The Open Group Research Inst. (617) 621-7349 11 Cambridge Center Cambridge, MA 02142 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From JoopvandeWege@mail.mococo.nl Fri May 23 01:36:00 1997 From: JoopvandeWege@mail.mococo.nl (Joop van de Wege) Date: Fri, 23 May 1997 01:36:00 -0000 Subject: GNU-Win32 m68K cross compiler References: <5js886$hcf$1@majipoor.cygnus.com> Message-ID: <13130469400236@mococo.nl> On Tue, 20 May 1997 13:50:41 +0100 Joop van de Wege wrote: > > > On 26 Apr 1997 06:44:22 GMT > dje@cygnus.com (Doug Evans) wrote: > > > Jan Ove Etterlid (joe@datasalg.no) wrote: > > > Is there anyone who can give me a list of the steps I have to go to > > > build a GNU-Win32 gcc cross compiler, assembler and linker which can > > > produce m68k coff format executable files, running on Win NT 3.51 > > > pentium PC? > > You may wish to check out the crossgcc FAQ. > > I did that but didn't find the answers needed for a cross-compiler like > this: > --host=i386-cygwin32 --target=m68k-amiga{d}os > > The binutils compile fine and it has only minor problems trying to > install the documentation (guide/info), that is no problem since i have > it on my Amiga. > The problem is with compiling gcc itself. > The cygwin32 distribution doesn't know about m68k-amiga{d}os and the > Amiga distribution doesn't know about cygwin32 (files involved are: > config.sub and config.guess). Simply exchanging them doesn't work. > So I have directions from both environments on how to make a crosscompiler > but with an dependency that is like a chicken and egg problem. I solved this one by running diff on makefile.in and configure of both systems and modifying cygnwin32 environment to include the Amiga specific stuff. > > Anyone make a m68k target with cygwin32 as a host? > > Can anyone help me a bit in the right direction on solving this? OK, I'm now sofar that the cross-compiler is being used in the makefile. (xgcc -D... -o dummy.o -c dummy.c). It tries to compile dummy.c but it fails because it calls the assembler like this: 'as.exe -mc68010 .....'. This is the native assembler (i386) and not the cross-assembler. I made sure I specified correct paths while configuring the compiler. Is this a fault in the configure script? xgcc --print-search-dirs reveals that my path is there together with a bunch of others. Telling xgcc to compile to assembler and feeding the result file to m68k -amigaos-as.exe will work fine. Again, any help appreciated, Joop - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From marc@watson.ibm.com Fri May 23 01:36:00 1997 From: marc@watson.ibm.com (Marc Auslander) Date: Fri, 23 May 1997 01:36:00 -0000 Subject: GUD Message-ID: <9705222054.AA34870@marc.watson.ibm.com> I've had no luck running gdb under emacs GUD. It hangs - requireing that I blow away gdb to get emacs to respond. This is on NT4 workstation. I am using the target remote option over tcpip. Marc Auslander 914 945-4346 (Tieline 862 Fax x4426) - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From pat@po.cwru.edu Fri May 23 01:36:00 1997 From: pat@po.cwru.edu (Paul A. Thompson) Date: Fri, 23 May 1997 01:36:00 -0000 Subject: Nothing happens Message-ID: <3.0.1.32.19970522153408.006acec0@pop.cwru.edu> I recently downloaded Ver 18 of gnu-win18 tools. I installed them in d:\gnuw\h-i386-cygwin32. I then used mount to mount d:\gnuw\h-i386-cygwin32 as /usr. Nothing happens when I use the tools. My windows directory is c:\nwin; I put a copy of cygwin.dll in c:\nwin\system, although the instructions say nothing about this. Any omissions in my setup? - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From gesteld@se.bel.alcatel.be Fri May 23 01:36:00 1997 From: gesteld@se.bel.alcatel.be (Dirk Vangestel) Date: Fri, 23 May 1997 01:36:00 -0000 Subject: X11R6.3 nolonger works under B18 References: <3.0.1.32.19970519100433.00785b18@mail.cs.unc.edu> <3.0.1.32.19970521225846.00826210@192.160.61.2> Message-ID: <3.0.1.16.19970523080517.47d7e5a6@btmpf3.se.bel.alcatel.be> At 10:58 PM 5/21/97 -0400, Ian wrote: >Is there somewhere where one could ftp an already working X11R6.3? I >tried to build it myself (with Sergey Okhapkin's patch), but I don't >have imake, and am probably missing something else needed to compile >it myself :( > >(If not, the location of a working imake and other tools needed to >compile X would of course be welcome) imake is part of the X distribution. Check the output of your make command, and you'll probably see it's trying to build imake using cc instead of gcc. I'm not sure about the correct way to solve this, but copying gcc.exe to cc.exe works for me. I'm still having problems compiling X though (it hangs somewhere) but I'm getting closer :) Dirk - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From condict@opengroup.org Fri May 23 01:36:00 1997 From: condict@opengroup.org (Michael Condict) Date: Fri, 23 May 1997 01:36:00 -0000 Subject: ncurses/termcap, just won't clear the screen? References: <01BC65ED.28765730@gater.krystalbank.msk.ru> Message-ID: <199705212147.RAA15201@postman.opengroup.org> In message < 01BC65ED.28765730@gater.krystalbank.msk.ru >, you write: > Geoffrey Noer wrote: > > > With the new /etc/termcap from b18, I do an "export TERM=linux" > > in bash, and then hit a control-l to clear. > > > > Under Windows NT, the right thing happens. Under Windows 95, however > > I see the behavior kunglao@prairienet.org reports where the > > prompt is moved but the old text doesn't get erased. > > > > ScrollConsoleScreenBuffer() on Windows95 fails to scroll more than half of sc > reen... I'm trying to find a workaround :-( There's another problem, this one on NT. When it clears the screen it sets the background to black, whether or not you've got that color as your chosen background (a DOS window in NT can have many different colors of text and background). Minor, but annoying. Michael Condict - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Scott.Mintz@po.cle.ab.com Fri May 23 01:36:00 1997 From: Scott.Mintz@po.cle.ab.com (Scott Mintz) Date: Fri, 23 May 1997 01:36:00 -0000 Subject: Fix to readline for case-insensitive file-name comple Message-ID: <000670F0.1893@po.cle.ab.com> I use 4DOS/NT as my primary command shell. It is far and away superior to CMD, much easier to use than bash, and better integrated with the Windows NT world. 4DOS/NT has case-insensitive filename completion. However, by continuing to press the key, 4DOS will cycle through all the names. For instance, if I have a directory named foo and another named Fred, Typing F first brings up foo and pressing it again brings up Fred. +------------------------------+------------------------------------+ | Scott A. Mintz | voice: (216) 646-4805 | | Allen-Bradley Company | fax: (216) 646-4961 | | 1 Allen-Bradley Drive | email: scott.mintz@po.cle.ab.com | | Mayfield Hts., OH 44124-6118 | CIS: 71461,632 | +------------------------------+------------------------------------+ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From lhall@rfk.com Fri May 23 01:36:00 1997 From: lhall@rfk.com (Larry Hall) Date: Fri, 23 May 1997 01:36:00 -0000 Subject: error compiling groff-1.10 under b17.1 Message-ID: <2.2.32.19970522130714.0094c4a4@ma.ultranet.com> At 04:07 PM 5/20/97 -0025, Pasa Guglielmo wrote: >> g++ -I. -I. -I../include -I./../include -DHAVE_UNISTD_H=1 -DHAVE_DIRENT_H=1 -DH >> VE_LIMITS_H=1 -DHAVE_SYS_DIR_H=1 -DHAVE_STDLIB_H=1 -DSTDLIB_H_DECLARES_PUTENV=1 >> -DSTDIO_H_DECLARES_POPEN=1 -DSTDIO_H_DECLARES_PCLOSE=1 -DHAVE_CC_OSFCN_H=1 -DHA >> E_CC_LIMITS_H=1 -DRETSIGTYPE=void -DHAVE_STRUCT_EXCEPTION=1 -DHAVE_RENAME=1 -DH >> VE_MKSTEMP=1 -DWCOREFLAG=0200 -g -O -c index.cc >> index.cc: In method `int index_search_item::load(int)': >> index.cc:213: warning: implicit declaration of function `int strchr(...)' >> index.cc:213: warning: assignment to `const char *' from `int' lacks a cast >> index.cc: In function `class search_item * make_index_search_item(const char *, >> int)': >> index.cc:276: warning: implicit declaration of function `int strcat(...)' >> index.cc: In method `const char * index_search_item::munge_filename(const char >> )': >> index.cc:429: invalid types `int[int]' for array subscript >> index.cc: In method `void index_search_item::read_common_words_file()': >> index.cc:548: no matching function for call to `index_search_item::munge_filena >> e (int)' >> index.cc:425: candidates are: index_search_item::munge_filename(const char *) >> index.cc: In method `void index_search_item::check_files()': >> index.cc:608: warning: initialization to `const char *' from `int' lacks a cast >> index.cc:610: warning: assignment to `const char *' from `int' lacks a cast >> make: *** [index.o] Error 1 >> make: *** [libbib] Error 2 >> make: *** [all] Error 2 > >This is what I get after launching 'make'. I'm using b17.1. >I use the groff-1.10 distribution of Linux on which I did ./configure. >I've already installed grep-2.0 succesfully and also termcap (which I >got also in Linux/Slackware). > >and finally for less-332 : > >> gcc -I. -c -g -O output.c >> output.c:288: warning: static declaration for `iprintf' follows non-static >> output.c: In function `iprintf': >> output.c:290: argument `fmt' doesn't match prototype >> D:\cygnus\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32/cygnus-2.7.2-961023/../../.. >> /../i386-cygwin32/include/stdio.h:175: prototype declaration >> output.c:290: number of arguments doesn't match prototype >> D:\cygnus\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32/cygnus-2.7.2-961023/../../.. >> /../i386-cygwin32/include/stdio.h:175: prototype declaration >> make: *** [output.o] Error 1 > > I don't know about your groff problem, but your problem with less is caused by the fact that output.c defines a function iprintf() statically which is prototyped in stdio.h. I don't know what's the best way to fix this but I did something easy which was rename the static function and change all the references (about 3 if I remember correctly). This worked fine for me. Larry Hall lhall@rfk.com RFK Partners, Inc. (617) 239-1053 8 Grove Street (617) 239-1655 - FAX Wellesley, MA 02181 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From bernd@asterix.gecko.de Fri May 23 01:36:00 1997 From: bernd@asterix.gecko.de (Bernd Prager) Date: Fri, 23 May 1997 01:36:00 -0000 Subject: gcc does not works at all References: <3.0.1.32.19970522104554.005c3d40@pop.ilt.fhg.de> Message-ID: > I get the error messages: > > heap_init: unable to allocate heap, win32 error 87 > cygwin: terminating Seems that u have an old cygwin32.dll in your path (win system dir?). The old one has about 2MB the new 496K size. Remove the old dll. That solved this problem for me. Bernd _____________________________________________________________________ Bernd Prager GECKO mbH; Wismarsche Str.3, 18057 Rostock; Germany http://www.gecko.de PGP Key fingerprint = 83 54 6A 3B 7A 9D 6C 0E F3 41 CE 99 11 30 B7 D6 public key by mailto:bpr@gecko.de?Subject=SendPGPKey _____________________________________________________________________ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From kunglao@prairienet.org Fri May 23 21:55:00 1997 From: kunglao@prairienet.org (kunglao@prairienet.org) Date: Fri, 23 May 1997 21:55:00 -0000 Subject: Finally, a decent df command! Message-ID: <199705240451.XAA21138@bluestem.prairienet.org> Ok, I've finally put together something I've wished for, for quite a while now. All I've seen in a df command, was one from virtunix, and didn't have any real flexibility, nor the usual gnu-df look I've been used to for years. So, I made one. This isn't a port of GNU's df. I just too the man page, and implemented the same functions and look of it. The only things it doesn't do (yet) are 1) df /tmp - check the space on one particular partition, and 2) it's limited to partitions less than 2 gigs. I'll have to see about converting Win95 OSR2's kernel.dll into libkernel32.a to get the updated GetFreeDiskSpaceEx() function - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From kunglao@prairienet.org Fri May 23 21:55:00 1997 From: kunglao@prairienet.org (kunglao@prairienet.org) Date: Fri, 23 May 1997 21:55:00 -0000 Subject: error compiling groff-1.10 under b17.1 References: <3381C42B.69D7@swissonline.ch> Message-ID: <199705222100.QAA14942@bluestem.prairienet.org> > This is what I get after launching 'make'. I'm using b17.1. > I use the groff-1.10 distribution of Linux on which I did ./configure. > I've already installed grep-2.0 succesfully and also termcap (which I > got also in Linux/Slackware). I have a patch to groff-1.10 (and man) on my ftp. ftp.ice.net/pub/home/python install it, and it should compile completely. --- "Oh, what sad times are these when passing ruffians can say 'Ni!' at will to old ladies. There is a pestulance upon this land, nothing is sacred. Even those who arrange and design shrubberies are under considerable economic stress of this period of history." -- Roger the Shrubber - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From kunglao@prairienet.org Fri May 23 21:55:00 1997 From: kunglao@prairienet.org (kunglao@prairienet.org) Date: Fri, 23 May 1997 21:55:00 -0000 Subject: ncurses just won't clear the screen? References: <199705192156.QAA29407@bluestem.prairienet.org> <199705220753.DAA25252@r02n05.cac.psu.edu> Message-ID: <199705231134.GAA05152@bluestem.prairienet.org> > Check your /etc/termcap file to see if you have the updated one. > Basic term abilities seem to work now. I got rogue compiled and > working just fine. (Although I didn't use ncurses, it comes with a > toned down curses implementation) I copied it straight from /usr/etc. > Try this: > > #include > int > main() > { > printf("\033[2J"); > } > > Compile and run it, if it clears the screen your term has the > properties of a proper terminal. > > BTW my $TERM is ansi. I set it in autoexec.bat. I seem to recall when > I used linux term it didnt clear my screen either. That did work perfectly with TERM=linux. So perhaps it's just ncurses that's not doing the screen clears. That's where the clear.exe came from, and many of the test programs aren't working. BTW, I found that vt200 does work for screen clears. But then you lose things like color. --- "Oh, what sad times are these when passing ruffians can say 'Ni!' at will to old ladies. There is a pestulance upon this land, nothing is sacred. Even those who arrange and design shrubberies are under considerable economic stress of this period of history." -- Roger the Shrubber - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Sat May 24 05:03:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Sat, 24 May 1997 05:03:00 -0000 Subject: ssh 1.2.20 diff and patch problems Message-ID: <01BC685C.490F2320@gater.krystalbank.msk.ru> Hannu Koivisto wrote: > Greetings, > > I tried to apply the announced patch for ssh 1.2.20 to compile it with > gnuwin32. However, patching (patch < ssh-1.2.20.diff) didn't succeed: > Remove all CRs from the diff. -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jman@lx.net Sat May 24 06:31:00 1997 From: jman@lx.net (jman) Date: Sat, 24 May 1997 06:31:00 -0000 Subject: stdio.h Message-ID: <338725E0.5726@lx.net> Howdy, Has anyone had a problem with stdio.h Most the programs I have compiled work but ANY the use stdio.h crash or just won't compile no mater what... Just wondering if anyone else has had any similar problems with it.. -- Jason L. Esman aka _Jman System Admin. Network Consultant Irc Admin of irc.lx.net(newnet) Co. Owner of Den Internet Services http://www.lx.net && http://www.deninc.com PGP Keys finger jman@lx.net - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From mh@mike.franken.de Sat May 24 07:21:00 1997 From: mh@mike.franken.de (Michael Hirmke) Date: Sat, 24 May 1997 07:21:00 -0000 Subject: ncurses/termcap, just won't clear the screen? References: <199705220753.DAA25252@r02n05.cac.psu.edu> Message-ID: <6XSv1y54pfB@mike.franken.de> Hi eric, [...] > Try this: > > #include > int > main() > { > printf("\033[2J"); > } > > Compile and run it, if it clears the screen your term has the > properties of a proper terminal. > > BTW my $TERM is ansi. I set it in autoexec.bat. I seem to recall when > I used linux term it didnt clear my screen either. I use NT 4 with SP3 and CDK b18 and it works properly with TERM=ansi *and* TERM=linux ! Bye. Michael. -- Michael Hirmke | Telefon +49 (911) 557999 Georg-Strobel-Strasse 81 | FAX +49 (911) 557664 90489 Nuernberg | E-Mail mh@mike.franken.de | WWW http://minimike.franken.de/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jimen@adtech-inc.com Sat May 24 07:21:00 1997 From: jimen@adtech-inc.com (Jimen Ching) Date: Sat, 24 May 1997 07:21:00 -0000 Subject: Bug with fwrite? Message-ID: <01BC67A6.01B718E0@ilima038.lava.net> The following program does not generate the correct output file. I ran the program on a file with just "11" in the contents. The chksum is valid in the calc_chksum function. But when I view the output file, it is not the same. I'm using cygwin32 b18. --jc INTSTK.C -------------- next part -------------- A non-text attachment was scrubbed... Name: INTSTK.C Type: text/x-c Size: 5603 bytes Desc: not available URL: From kdw3@po.cwru.edu Sat May 24 08:27:00 1997 From: kdw3@po.cwru.edu (Kevin White) Date: Sat, 24 May 1997 08:27:00 -0000 Subject: GNU tools Message-ID: <338597B8.A950740@po.cwru.edu> Has anyone gotten the GNU internet tools to compile under b18? Bash 2.0? Fileutils? On all of them the configure script runs with no problems but the make always dies. -- __________________________ Kevin White kdw3@po.cwru.edu CWRU Dept of Genetics __________________________ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From gpasa@swissonline.ch Sat May 24 08:27:00 1997 From: gpasa@swissonline.ch (Pasa Guglielmo) Date: Sat, 24 May 1997 08:27:00 -0000 Subject: binary mount Message-ID: <33871BF1.19B3@swissonline.ch> Hi, After all that time I heard of mounting in binary I've taken the decision to ask what does it mean ? Is it to be done in the registries of win95 and then how ? or I must do it manually with 'mount -b' and then how and should I do it every time I use cygwin ? -- Sincerely, Pasa Guglielmo ------------------------------------------------------------------ & e-mail : gpasa@swissonline.ch & tel.: +41 (0)24 485 50 40 & & mailing : Pasa Guglielmo & fax : +41 (0)24 485 50 44 & & Rte des Cases 17A & & & CH-1890 St-Maurice & & & (Switzerland) & & ------------------------------------------------------------------ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jman@lx.net Sat May 24 09:30:00 1997 From: jman@lx.net (jman) Date: Sat, 24 May 1997 09:30:00 -0000 Subject: Registery Message-ID: <33874FEE.22BC@lx.net> Just wondering how many registery entries does cygnus have in win95... I opened the regestry to find 4 entries with some saying b15.. the only one that says b18 has nothing in it... -- Jason L. Esman aka _Jman System Admin. Network Consultant Irc Admin of irc.lx.net(newnet) Co. Owner of Den Internet Services http://www.lx.net && http://www.deninc.com PGP Keys finger jman@lx.net - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From *jeffdb@netzone.nospam.com Sat May 24 09:30:00 1997 From: *jeffdb@netzone.nospam.com (Mikey) Date: Sat, 24 May 1997 09:30:00 -0000 Subject: stdio.h Message-ID: <199705241617.JAA17375@nz1.netzone.com> Did you fix your mounts after switching to B18? Should all say text=binary, not text!=binary. (*jeffdb@netzone.com) the return address for this message is anti spammed remove * from the above address to reply. Mikey ---- From: jman To: gnu-win32@cygnus.com Date: Saturday, May 24, 1997 9:06 AM Subject: stdio.h >Howdy, > Has anyone had a problem with stdio.h >Most the programs I have compiled work but ANY the use stdio.h crash or >just won't compile no mater what... > Just wondering if anyone else has had any similar problems with it.. >-- >Jason L. Esman aka _Jman >System Admin. Network Consultant >Irc Admin of irc.lx.net(newnet) >Co. Owner of Den Internet Services > http://www.lx.net && http://www.deninc.com >PGP Keys finger jman@lx.net >- >For help on using this list (especially unsubscribing), send a message to >"gnu-win32-request@cygnus.com" with one line of text: "help". > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jman@lx.net Sat May 24 09:53:00 1997 From: jman@lx.net (jman) Date: Sat, 24 May 1997 09:53:00 -0000 Subject: stdio.h References: <199705241617.JAA17375@nz1.netzone.com> Message-ID: <33875539.309C@lx.net> Mikey wrote: > > Did you fix your mounts after switching to B18? > Should all say text=binary, not text!=binary. > > (*jeffdb@netzone.com) > the return address for this message is anti spammed > remove * from the above address to reply. > Mikey > ---- > From: jman > To: gnu-win32@cygnus.com > Date: Saturday, May 24, 1997 9:06 AM > Subject: stdio.h > > >Howdy, > > Has anyone had a problem with stdio.h > >Most the programs I have compiled work but ANY the use stdio.h crash or > >just won't compile no mater what... > > Just wondering if anyone else has had any similar problems with it.. > >-- > >Jason L. Esman aka _Jman > >System Admin. Network Consultant > >Irc Admin of irc.lx.net(newnet) > >Co. Owner of Den Internet Services > > http://www.lx.net && http://www.deninc.com > >PGP Keys finger jman@lx.net > >- > >For help on using this list (especially unsubscribing), send a message to > >"gnu-win32-request@cygnus.com" with one line of text: "help". > > Ok I unmounted an remounted unmounted text!=binary to text=binary just like ya said... and this simple easy program wouldn't compile at all.. I know it don't call for stdio.h but it didn't an wouldn't compile. #include int STDCALL WinMain (HINSTANCE hInst, HINSTANCE hPrev, LPSTR lpCmd, int nShow) { MessageBox (NULL, "Hello World!", "Test Compiler works", MB_OK); return 0; } Compiled with gcc -o test.exe test.c -luser32 -Wl,--subsystem,windows I remounted the old way with text!=binary and tada it compiled no problems.. I guess I will just have to play around I never had this problem till I went to b18 tho. I do see that b18 is faster tho at least it is for me.. -- Jason L. Esman aka _Jman System Admin. Network Consultant Irc Admin of irc.lx.net(newnet) Co. Owner of Den Internet Services http://www.lx.net && http://www.deninc.com PGP Keys finger jman@lx.net - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From *jeffdb@netzone.nospam.com Sat May 24 09:53:00 1997 From: *jeffdb@netzone.nospam.com (Mikey) Date: Sat, 24 May 1997 09:53:00 -0000 Subject: Diffs for ssh-1.2.20 Message-ID: <199705232058.NAA26223@nz1.netzone.com> IT'S ALIVE!!!!!!!! I'm on win95, so the console handling is probably even worse than NT. I had the same problem with this telnet as the one from B14 net.tar.gz. THE SECRET. I finaly tried repeating the first character of the password twice and it worked!! I was able to login, and got a bash prompt, and everything! I tried out ssh in insecure mode by copying telnet to //D/WINNT/system/rsh.exe and I was able to login that way too. finger worked fine, at least I was able to finger myself, (look MA no HANDS ;^). The really wierd part is that after I copied termcap to .termcap in my home dir (my ISP runs BSDI which never heard of linux) the linux terminal worked perfectly as long as I was using telnet. still the extra ^M's at the command prompt though. not only that, but the PgDn key worked for the first time ever in w32ansi and the cursor didn't show up on the first character of the next line, when I pressed the down arrow. go figure. (*jeffdb@netzone.com) the return address for this message is anti spammed remove * from the above address to reply. Mikey ---- From: Igor V. Kovalenko To: Joe Rumsey Cc: gnu-win32@cygnus.com Date: Tuesday, May 20, 1997 10:43 PM Subject: Re: Diffs for ssh-1.2.20 >> Since several people have asked, I've made the diffs for compiling >> ssh-1.2.20 available. >> It's at http://www.aa.net/~ogre/ssh-1.2.20.diff >> >> sshd compiles too, but it doesn't work. Might under NT, I dunno. > > -- No, mostly because of missing pseudo-terminal support code. > > Also (cosmetic) note: still missing cygwin32_setsockopt in cygwin.din >(beta18 source). > >setXXXent/endXXXent (XXX=host,net,pw...) pairs might as well > be defined to zero code in appropriate header files. > >-- Good luck! > >PS: I have compiled telnet and finger clients from NetKit-B-0.06 > (plus RedHat patches) with minor source changes under beta18 > environment. It even works! under WinNT4.0(+sp2) so I replaced > original telnet with a cygwin-compiled one. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Sat May 24 11:57:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Sat, 24 May 1997 11:57:00 -0000 Subject: ncurses/termcap, just won't clear the screen? Message-ID: <01BC6896.105EE990@sos> Michael Hirmke wrote: > I use NT 4 with SP3 and CDK b18 and it works properly with TERM=ansi > *and* TERM=linux ! > It's better to set TERM=linux, because hard-coded in cygwin.dll functional key translation is based on linux terminal description. Screen outputs should works fine with both terminal names. -- Sergey Okhapkin Moscow, Russia Looking for a job - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From rgr@rio.sci.ccny.cuny.edu Sat May 24 11:57:00 1997 From: rgr@rio.sci.ccny.cuny.edu (Roger Kuhlman) Date: Sat, 24 May 1997 11:57:00 -0000 Subject: Re NTEMACS and ^M References: <9705230129.AA10593@modi.xraylith.wisc.edu> Message-ID: <199705241821.OAA05853@rio.sci.ccny.cuny.edu> Greetings: To everyone who is mounting datadirectories(source) as binary, and has fixed NTEMACS to stop supplying \n\r on saves. The fastest way to fix files(especially makefiles) go to first ^M in text, do a query-replace, and copy ^M from the text file to the query-replace: command in the minibuffer. Then press enter to the with: You can then ! the command and remove all \r's from the file. Regards, Roger Kuhlman e-mail: rgr@rio.sci.ccny.cuny.edu - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Sat May 24 13:16:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Sat, 24 May 1997 13:16:00 -0000 Subject: X11 and B18 Message-ID: <01BC67C8.82162220@sos> X11 DLLs built on my NT box works on both NT and W95. The same built on W95 works on NT only. Today I'll try those builds on my home machine. Any ideas? -- Sergey Okhapkin Moscow, Russia Looking for a job - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From mh@mike.franken.de Sat May 24 13:16:00 1997 From: mh@mike.franken.de (Michael Hirmke) Date: Sat, 24 May 1997 13:16:00 -0000 Subject: ncurses just won't clear the screen? References: <199705231134.GAA05152@bluestem.prairienet.org> Message-ID: <6XSxXXN4pfB@mike.franken.de> Hi kunglao, [...] > That did work perfectly with TERM=linux. So perhaps it's just > ncurses that's not doing the screen clears. That's where the > clear.exe came from, and many of the test programs aren't working. Even ncurses and its clear work with TERM=linux under NT4 - with one exception: they always set the background to black and the foreground to white. Bye. Michael. -- Michael Hirmke | Telefon +49 (911) 557999 Georg-Strobel-Strasse 81 | FAX +49 (911) 557664 90489 Nuernberg | E-Mail mh@mike.franken.de | WWW http://minimike.franken.de/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Henri.Valeins@rmsb.u-bordeaux2.fr Sat May 24 13:16:00 1997 From: Henri.Valeins@rmsb.u-bordeaux2.fr (Henri.Valeins@rmsb.u-bordeaux2.fr) Date: Sat, 24 May 1997 13:16:00 -0000 Subject: Windows print fonctions and gnu-win32 Message-ID: <199705231345.PAA08125@icis.rmsb.u-bordeaux2.fr> Hi, I'm a new user of cygnus B18. When I try to compile a program whith a W95 print fonction I have a error message like this : >gcc -o beeper1.exe beeper.c -lcygwin -lkernel32 -lgdi -lwinspool >... >c:\windows\tempcc0010691.o(.text+0x5ea):beeper.c: undefined reference to 'OpenPrinter' I try to compile with win32spl library but it's the same thing. Do you have solution for this problem ? Thanks ,., (o o) +-----------oOO-|(_)|-OOo--+---------------------------------------------+ |Henri Valeins |Risonance Magnitique des Systhmes Biologiques| | |CNRS UMR 5536, Case 93 | |Tel : 05 57 57 17 85 |Universite Victor Sigalen Bordeaux 2 | |Fax : 05 56 96 13 41 |146, rue Leo Saignat | | |33076 Bordeaux Cedex FRANCE | +--------------------------+---------------------------------------------+ | E-mail : henri.valeins@rmsb.u-bordeaux2.fr | +------------------------------------------------------------------------+ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From condict@opengroup.org Sat May 24 17:35:00 1997 From: condict@opengroup.org (Michael Condict) Date: Sat, 24 May 1997 17:35:00 -0000 Subject: v18 vs v17 compile times.... References: <01BC669C.7C655BC0@gater.krystalbank.msk.ru> Message-ID: <199705231513.LAA06332@postman.opengroup.org> In message < 01BC669C.7C655BC0@gater.krystalbank.msk.ru >, you write: > > Somebody else already asked, but I did not see the answer to the question, "Is > > it (reasonably) easy to build without or turn off symbolic links?" > > I've made some improvements in symlink checking and sent a patch to mr. > Noer. My patch checks "system" file attribute symlinks must to have. It > doesn't affect non-cygwin programs any way - they knows nothing about > symlinks. Isn't the performance problem with symlinks inherent in the current design? They are files without a suffix, therefore without an identifiable Windows type. Therefore you have to look at the contents of the file to see if it is a symlink. And checking the system attribute is not a complete solution to the performance problem. It just means that you don't have to check the contents of the file unless it has the system attribute. But you still have to always perform an extra operation to check the system attribute. And for the many non- symlink files that have the system attribute, you still have to read the file to see if it's a symlink. I'd like to step back and point out that Windows 95 / NT 4 already *have* symlinks. They are ".lnk" files for folders, data files and Win32 programs, and ".pif" files for 16-bit programs. They have all the power of UNIX symlinks in what information is stored in them, and more. Now it's true that Windows makes very poor use of these symlinks, allowing them to have meaning only if you double-click on them in an Explorer window or feed them to the "Run" menu or the "start" command. You can't "cd" through them or print the contents of the file to which they point using the "type" command. But this is no reason that Cygnus Win32 can't endow them with the full power of UNIX symlinks. This would be a solution that does not require reading a symlink file. It's also much better integrated with Windows. Wouldn't it be nice if you could navigate through Cygnus Win32 symlinks using Explorer? When you've made c:/bin be a symlink/shortcut to the Cygnus godawfully named ".../H-i386-cygwin32/bin" directory, wouldn't it be nice if you could type e.g. "/bin/bash" to the Start / Run window, and have it start up bash? Here are some details of how such links could work: 1) Readdir, when returning the names of files in a directory, should delete the suffix ".lnk" or ".pif", if present. 2) In the open syscall, if an attempt to open a file fails with a "not found" error, open should suffix ".lnk" and try again. If that fails, it should suffix ".pif" and try again. If either of these succeeds, open should then read the shortcut file, find the path name of the target, and do a Win32 open of the target file. This will slow down the processing of sym links a bit (because a failed open has to take place first), but allows non-symlink files to be processed without any overhead added by the symlink code. 3) The other syscalls that take a pathname and are supposed to follow symlinks (e.g. chdir, stat), should select the target file or folder similarly to the algorithm used by open. Comments? Michael Condict m.condict@opengroup.org The Open Group Research Inst. (617) 621-7349 11 Cambridge Center Cambridge, MA 02142 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From banders@switch.rockwell.com Sat May 24 17:35:00 1997 From: banders@switch.rockwell.com (Bart Anderson) Date: Sat, 24 May 1997 17:35:00 -0000 Subject: Gdk95.lib References: <33809E15.6832@switch.rockwell.com> Message-ID: <3385CFBA.4888@switch.rockwell.com> Bart Anderson wrote: > > Does anyone know if gnuwin32 can link with Gdk95.lib. > I tried putting it with the other libs and specifying -lGdk95 and > -lgdk95 with no luck. > > Bart Well Bart, I took a look at your problem and here's what I found ;) Gdk95.lib has object files that can be used after (manually editing a couple of the filenames... change the nasty .\Debug\Dd.obj to Dd.obj and then using ar -x Dd.obj) and renaming to .o file. I found Object.obj Sprite.Obj Dd.obj and Proc.obj. Now I'm only left with 2 references that I guess exist in standard .dll's (ie. Ddraw.dll). Hope I've helped.. but I don't know how to reference the functions in dll files. Good luck, Bart > - > For help on using this list (especially unsubscribing), send a message to > "gnu-win32-request@cygnus.com" with one line of text: "help". - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Scott.Mintz@po.cle.ab.com Sat May 24 17:35:00 1997 From: Scott.Mintz@po.cle.ab.com (Scott Mintz) Date: Sat, 24 May 1997 17:35:00 -0000 Subject: A call to find mounted dirs? Message-ID: <00068C71.1893@po.cle.ab.com> The mount history is not stored in a disk file with gnu-win32, it is stored in the registry. The registry key name changed between b17.1 and b18. It was "Cygnus Support" and is now "Cygnus Solutions". I'm not sure why you lost your "old" mounts in b17.1. My mounted drives behave exactly the same under both. -Scott A. Mintz ______________________________ Reply Separator _________________________________ Subject: Re: A call to find mounted dirs? Author: Chin Chee-Kai at Internet Date: 5/23/97 9:59 AM On Wed, 21 May 1997, Mikey wrote: > Can you say fstab? Hi Mikey, not sure what you mean by "say fstab". If what you mean is that I had created /etc/fstab file that resulted in the remnant mounts, I have no such file created actually. C:\>dir \etc\fstab File not found C:\>mount Device Directory Type Flags \\.\tape1: /dev/st1 native no-mixed,text!=binary \\.\tape0: /dev/st0 native no-mixed,text!=binary \\.\b: /dev/fd1 native no-mixed,text!=binary \\.\a: /dev/fd0 native no-mixed,text!=binary c: / native no-mixed,text!=binary C:\>mount /usr/local/ /local <-- create /usr/local and /local before this command C:\>mount Device Directory Type Flags \usr\local /local native no-mixed,text!=binary \\.\tape1: /dev/st1 native no-mixed,text!=binary \\.\tape0: /dev/st0 native no-mixed,text!=binary \\.\b: /dev/fd1 native no-mixed,text!=binary \\.\a: /dev/fd0 native no-mixed,text!=binary c: / native no-mixed,text!=binary C:\> Now reboot and immediately pull out a DOS prompt to type mount again and you'll see the same display as the last list. (I did all commands outside bash, but don't suspect the result to be different if they were run inside bash). In b17.1, I didnt' see such "long-lived mounts" behaving in this manner though (which prompted me to put in b17.1 an alias "setup" to do all my required mounts the very first time I run bash after booting up. This alias now (in b18) returns "already mounted" messages when I run "setup", which means the "mount" semantics had changed). I am also not sure if this is due to b17.1 and b18 file interferences as the same behavior happened on another PC on which I installed b18 with a clean start without b17.1. Any ideas? Chin Chee-Kai Internet Email: cheekai@singapore.sterling.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From *jeffdb@netzone.nospam.com Sat May 24 17:35:00 1997 From: *jeffdb@netzone.nospam.com (Mikey) Date: Sat, 24 May 1997 17:35:00 -0000 Subject: gnu-win32-specific bug in RCS 5.7 Message-ID: <199705231839.LAA13715@nz1.netzone.com> I would do that Vassilii, but by the time the patch were integrated, and released it would (I hope), be useless, because (I hope), fixing the dll as time goes on will fix the problems. I don't like fixing applications to work with development "kernels" because generally speaking you are wasting your time. You are better off fixing the "kernel" in the first place or in this case the dll. Sadly I couldn't figure out a "kernel" fix for this one, and I need rcs cause I use it under linux, djgpp, and OS/2 so I "fixed" it. (*jeffdb@netzone.com) the return address for this message is anti spammed remove * from the above address to reply. Mikey ---- From: Vassilii Khachaturov To: 'smtp:jeffdb@netzone.com' Cc: gnu-win32 Date: Friday, May 23, 1997 10:31 AM Subject: Re: gnu-win32-specific bug in RCS 5.7 > >Thank you. It helped on NT, too. >Consider passing the patch to the RCS bugs mailing list >(rcs-bugs@cs.purdue.edu). > > --- conf.sh.orig Mon Mar 17 18:15:23 1997 >+++ conf.sh Mon Mar 17 18:22:37 1997 >@@ -1794,8 +1794,8 @@ > EOF > $PREPARE_CC || exit > h=0 p=0 > -if ($CL a.c $L && sh -c 'pid=$$; (sleep 3; kill $pid)& exec '$aout) >&2 > -then >+if ($CL a.c $L && sh -c 'pid=$$; (sleep 3; kill $pid)& sleep 4;exec '$aout) >>&2 >+then #extra sleep for fork bug > h=1 > $PREPARE_CC || exit > $CS -Dhas_attribute_format_printf=1 a.c >&2 && $CS_OK && p=1 >@@ -1890,10 +1890,11 @@ > echo "$a#undef EXIT_FAILURE $z/* Uncomment this if EXIT_FAILURE is broken. >*/" > > : configuring large_memory > -case "$has_map_fd$has_mmap" in > -*1*) l=1;; > -*) l=0 > -esac >+#case "$has_map_fd$has_mmap" in >+#*1*) l=1;; >+#*) l=0 >+#esac >+l=1 #force large_memory to overcome win95 short read bug > echo "#define large_memory $l /* Can main memory hold entire RCS files? > */" > > $ech >&3 "$0: configuring LONG_MAX $dots" > --- rcstest.orig Mon Mar 17 19:14:23 1997 >+++ rcstest Mon Mar 17 19:14:51 1997 >@@ -173,7 +173,7 @@ > case $USER in > ?*) me=$USER;; > *) > - me=`who am i` || exit 2 >+ me=`whoami` || exit 2 > me=`echo "$me" | sed -e 's/ .*//' -e 's/.*!//'` > case $me in > '') echo >&2 "$0: cannot deduce user name"; exit 2 >- >For help on using this list (especially unsubscribing), send a message to >"gnu-win32-request@cygnus.com" with one line of text: "help". > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From alex!@bigfoot.com Sat May 24 17:35:00 1997 From: alex!@bigfoot.com (Hmmmz) Date: Sat, 24 May 1997 17:35:00 -0000 Subject: libtcl.a Message-ID: <199705250035.RAA06540@mom.hooked.net> I see that the tcl libs that are included with B18 are made from a tcl dll built with BC4.5. However when I exit a program that uses the tcl library (like gdb) it crashes. Will this be fixed in b19? WinNT and *nix rule, so learn from the past: Hiroshima '45, Vietnam '64, Tjernobyl '86, MacOS '96 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From thunder1@fast.net Sat May 24 17:35:00 1997 From: thunder1@fast.net (Thunder) Date: Sat, 24 May 1997 17:35:00 -0000 Subject: How to get into bash? Message-ID: <3385E00F.3675@fast.net> Well this question may seem a little lame for this list seeing how most of the mail i read from here is about the actual source and such, but my question is a lot more general. I downloaded and installed the gnu.. edited my autoexec.bat like it specifies and i run the wish42.exe program and it opens up the console and some smaller window. Now my question is from there what is it do I have to do to setup this program so that i can enter bash. My whole objective is to get into bash so that i can compile a MUD lib that i currently have and i was recommended to download and use the gnu-win32 program to compile it. If someone could mail me back directly or could possibly point me to some docs somewhere I would be greatly appreciative. And I'll apologize in advance if these general type questions aren't what this mail list is for. Thanks in advance.. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From *jeffdb@netzone.nospam.com Sat May 24 17:35:00 1997 From: *jeffdb@netzone.nospam.com (Mikey) Date: Sat, 24 May 1997 17:35:00 -0000 Subject: ssh 1.2.20 diff and patch problems Message-ID: <199705231840.LAA13730@nz1.netzone.com> You have your filesystems mounted in text mode. if you absoulutly can't stand working with LF only files, do patch -p0 --ignore-whitespace < patchfile then use utod from djgpp to get back all the "\r"'s otherwise remount your filesystems -b reinstall cdk&user untar the stuff again, and everything will work much better. (*jeffdb@netzone.com) the return address for this message is anti spammed remove * from the above address to reply. Mikey ---- From: Hannu Koivisto To: gnu-win32@cygnus.com Date: Friday, May 23, 1997 9:36 AM Subject: ssh 1.2.20 diff and patch problems >Greetings, > >I tried to apply the announced patch for ssh 1.2.20 to compile it with >gnuwin32. However, patching (patch < ssh-1.2.20.diff) didn't succeed: > >Hmm... The next patch looks like a new-style context diff to me... >The text leading up to this was: >-------------------------- >|E_GETRUSAGE) && defined(XXXJR) >| { >| struct rusage ru, cru; >| getrusage(RUSAGE_SELF, &ru); >|diff -cr ssh-1.2.20/servconf.c ssh-1.2.20.new/servconf.c >|*** ssh-1.2.20/servconf.c Tue Apr 22 17:40:07 1997 >-------------------------- >Patching file ssh-1.2.20/servconf.c using Plan A... >assertion "hunk" failed: file "/home/noer/src/beta18/cdk/patch/patch.c", line 33 >9 > >I decided to try patching it in Linux. I moved the ssh archive and the diff >file to our Linux server and ran patch < ssh-1.2.20.diff. The result was even >more errors and failed hunks. Actually, almost all hunks failed with >different error messages. Here's just one example: > >Patching file ssh-1.2.20/socketpair.c using Plan A... >(Fascinating--this is really a new-style context diff but without >the telltale extra asterisks on the *** line that usually indicate >the new style...) >Hunk #1 failed at 9. >Hunk #2 succeeded at 67 with fuzz 2. > >What am I doing wrong? Thanks in advance. > >--- >Hannu Koivisto | What you see is all you get. >Prime jee | - Brian Kernighan >------------------------------------------------------------- > >- >For help on using this list (especially unsubscribing), send a message to >"gnu-win32-request@cygnus.com" with one line of text: "help". > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From allan@interport.net Sat May 24 20:49:00 1997 From: allan@interport.net (Allan Peda) Date: Sat, 24 May 1997 20:49:00 -0000 Subject: Does the install scriprt create directory /dev/i386-cygwin32 ? References: <199705190451.AAA45401@solid.ssd.ornl.gov> Message-ID: <3388C22E.6A39@interport.net> I just upgraded to build 18 of cygwin32 (installed under /gnuwin32). I found some redundant binaries under /dev/i38-cywin32/bin (ld.exe for one)... okay here: C:\>dir ld.exe /s Volume in drive C has no label. Volume Serial Number is CCA5-6764 Directory of C:\dev\i386-cygwin32\bin 05/08/97 10:52a 265,216 ld.exe 1 File(s) 265,216 bytes Directory of C:\gnuwin32\b18\H-i386-cygwin32\bin 05/08/97 10:52a 265,216 ld.exe 1 File(s) 265,216 bytes Question: Is this normal, or did I drag-n-drop something in error. I'd like to delete this stuff. Thanks - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From weiku@UTKUX1.UTK.EDU Sun May 25 00:04:00 1997 From: weiku@UTKUX1.UTK.EDU (Wei Ku) Date: Sun, 25 May 1997 00:04:00 -0000 Subject: v18 vs v17 compile times.... Message-ID: <199705250704.DAA47324@solid.ssd.ornl.gov> Sincerely, Wei Ku *************************************** Department of Physics and Astronomy The University of Tennessee 1408 Circle Drive Knoxville, Tennessee 37996-1200 weiku@utkux.utcc.utk.edu --------------------------------------- Solid State Division Oak Ridge National Laboratory P.O.Box 2008 Oak Ridge, TN 37831-6032 Phone: (423) 574-5795 Fax: (423) 574-4143 weiku@solid.ssd.ornl.gov *************************************** ???? ---- From: Michael Condict To: Sergey Okhapkin Cc: gnu-win32@cygnus.com; 'ap096@po.cwru.edu'; condict@opengroup.org Date: Sunday, May 25, 1997 2:55 AM Subject: Re: v18 vs v17 compile times.... In message < 01BC669C.7C655BC0@gater.krystalbank.msk.ru >, you write: > > Somebody else already asked, but I did not see the answer to the question, "Is > > it (reasonably) easy to build without or turn off symbolic links?" > > I've made some improvements in symlink checking and sent a patch to mr. > Noer. My patch checks "system" file attribute symlinks must to have. It > doesn't affect non-cygwin programs any way - they knows nothing about > symlinks. Isn't the performance problem with symlinks inherent in the current design????? They are files without a suffix, therefore without an identifiable Windows type.???? Therefore you have to look at the contents of the file to see if it is a symlink.???? And checking the system attribute is not a complete solution to the performance problem.???? It just means that you don't have to check the contents of the file unless it has the system attribute.???? But you still have to always perform an extra operation to check the system attribute.???? And for the many non- symlink files that have the system attribute, you still have to read the file to see if it's a symlink. I'd like to step back and point out that Windows 95 / NT 4 already *have* symlinks.???? They are ".lnk" files for folders, data files and Win32 programs, and ".pif" files for 16-bit programs.???? They have all the power of UNIX symlinks in what information is stored in them, and more. Now it's true that Windows makes very poor use of these symlinks, allowing them to have meaning only if you double-click on them in an Explorer window or feed them to the "Run" menu or the "start" command. You can't "cd" through them or print the contents of the file to which they point using the "type" command.???? But this is no reason that Cygnus Win32 can't endow them with the full power of UNIX symlinks.???? This would be a solution that does not require reading a symlink file. It's also much better integrated with Windows.???? Wouldn't it be nice if you could navigate through Cygnus Win32 symlinks using Explorer????? When you've made c:/bin be a symlink/shortcut to the Cygnus godawfully named ".../H-i386-cygwin32/bin" directory, wouldn't it be nice if you could type e.g. "/bin/bash" to the Start / Run window, and have it start up bash? Here are some details of how such links could work: 1) Readdir, when returning the names of files in a directory, ???????? should delete the suffix ".lnk" or ".pif", if present. 2) In the open syscall, if an attempt to open a file fails with ???????? a "not found" error, open should suffix ".lnk" and try ???????? again.???? If that fails, it should suffix ".pif" and try ???????? again.???? If either of these succeeds, open should then read ???????? the shortcut file, find the path name of the target, and do a ???????? Win32 open of the target file.???? This will slow down the ???????? processing of sym links a bit (because a failed open has to ???????? take place first), but allows non-symlink files to be ???????? processed without any overhead added by the symlink code. 3) The other syscalls that take a pathname and are supposed to ???????? follow symlinks (e.g. chdir, stat), should select the target ???????? file or folder similarly to the algorithm used by open. Comments? Michael Condict m.condict@opengroup.org The Open Group Research Inst. (617) 621-7349 11 Cambridge Center Cambridge, MA 02142 - For help on using this list (especially unsubscribing), send a message to " gnu-win32-request@cygnus.com " with one line of text: "help". From weiku@UTKUX1.UTK.EDU Sun May 25 02:01:00 1997 From: weiku@UTKUX1.UTK.EDU (Wei Ku) Date: Sun, 25 May 1997 02:01:00 -0000 Subject: Does the install scriprt create directory /dev/i386-cygwin32 ? Message-ID: <199705250854.EAA42095@solid.ssd.ornl.gov> Apparently, it is not the default installation. In my Windows 95 machine, there is ever no c:\dev. You can delete the whole c:\dev\bin directory. Sincerely, Wei Ku *************************************** Department of Physics and Astronomy The University of Tennessee 1408 Circle Drive Knoxville, Tennessee 37996-1200 weiku@utkux.utcc.utk.edu --------------------------------------- Solid State Division Oak Ridge National Laboratory P.O.Box 2008 Oak Ridge, TN 37831-6032 Phone: (423) 574-5795 Fax: (423) 574-4143 weiku@solid.ssd.ornl.gov *************************************** ---- From: Allan Peda To: gnu-win32@cygnus.com Date: Sunday, May 25, 1997 4:35 AM Subject: Does the install scriprt create directory /dev/i386-cygwin32 ? I just upgraded to build 18 of cygwin32 (installed under /gnuwin32). I found some redundant binaries under /dev/i38-cywin32/bin (ld.exe for one)... okay here: C:\>dir ld.exe /s Volume in drive C has no label. Volume Serial Number is CCA5-6764 Directory of C:\dev\i386-cygwin32\bin 05/08/97???? 10:52a???????????????????????????????????????????????????????? 265,216 ld.exe ???????????????????????????????????????????????????????? 1 File(s)???????????????????????????? 265,216 bytes Directory of C:\gnuwin32\b18\H-i386-cygwin32\bin 05/08/97???? 10:52a???????????????????????????????????????????????????????? 265,216 ld.exe ???????????????????????????????????????????????????????? 1 File(s)???????????????????????????? 265,216 bytes Question:???? Is this normal, or did I drag-n-drop something in error. I'd like to delete this stuff. Thanks - For help on using this list (especially unsubscribing), send a message to " gnu-win32-request@cygnus.com " with one line of text: "help". From weiku@UTKUX1.UTK.EDU Sun May 25 02:01:00 1997 From: weiku@UTKUX1.UTK.EDU (Wei Ku) Date: Sun, 25 May 1997 02:01:00 -0000 Subject: two question: ar and libf2c.a Message-ID: <199705250900.FAA27769@solid.ssd.ornl.gov> I have two stupid question: 1. Is it safe to use ar to make my own library like libmyown.a if I use text mode mounting ? 2. I found the libf2c.a and libf2c.h in cygwin32 directories. However, there is no f2c.exe. Are we supposed to use precompiled f2c.exe or f2cx.exe dor DOS from the netlib ? Sincerely, Wei Ku *************************************** Department of Physics and Astronomy The University of Tennessee 1408 Circle Drive Knoxville, Tennessee 37996-1200 weiku@utkux.utcc.utk.edu --------------------------------------- Solid State Division Oak Ridge National Laboratory P.O.Box 2008 Oak Ridge, TN 37831-6032 Phone: (423) 574-5795 Fax: (423) 574-4143 weiku@solid.ssd.ornl.gov *************************************** From weiku@UTKUX1.UTK.EDU Sun May 25 02:01:00 1997 From: weiku@UTKUX1.UTK.EDU (Wei Ku) Date: Sun, 25 May 1997 02:01:00 -0000 Subject: v18 vs v17 compile times.... Message-ID: <199705250705.DAA37345@solid.ssd.ornl.gov> I like your idea very much. Sincerely, Wei Ku *************************************** Department of Physics and Astronomy The University of Tennessee 1408 Circle Drive Knoxville, Tennessee 37996-1200 weiku@utkux.utcc.utk.edu --------------------------------------- Solid State Division Oak Ridge National Laboratory P.O.Box 2008 Oak Ridge, TN 37831-6032 Phone: (423) 574-5795 Fax: (423) 574-4143 weiku@solid.ssd.ornl.gov *************************************** ---- From: Michael Condict To: Sergey Okhapkin Cc: gnu-win32@cygnus.com; 'ap096@po.cwru.edu'; condict@opengroup.org Date: Sunday, May 25, 1997 2:55 AM Subject: Re: v18 vs v17 compile times.... In message < 01BC669C.7C655BC0@gater.krystalbank.msk.ru >, you write: > > Somebody else already asked, but I did not see the answer to the question, "Is > > it (reasonably) easy to build without or turn off symbolic links?" > > I've made some improvements in symlink checking and sent a patch to mr. > Noer. My patch checks "system" file attribute symlinks must to have. It > doesn't affect non-cygwin programs any way - they knows nothing about > symlinks. Isn't the performance problem with symlinks inherent in the current design????? They are files without a suffix, therefore without an identifiable Windows type.???? Therefore you have to look at the contents of the file to see if it is a symlink.???? And checking the system attribute is not a complete solution to the performance problem.???? It just means that you don't have to check the contents of the file unless it has the system attribute.???? But you still have to always perform an extra operation to check the system attribute.???? And for the many non- symlink files that have the system attribute, you still have to read the file to see if it's a symlink. I'd like to step back and point out that Windows 95 / NT 4 already *have* symlinks.???? They are ".lnk" files for folders, data files and Win32 programs, and ".pif" files for 16-bit programs.???? They have all the power of UNIX symlinks in what information is stored in them, and more. Now it's true that Windows makes very poor use of these symlinks, allowing them to have meaning only if you double-click on them in an Explorer window or feed them to the "Run" menu or the "start" command. You can't "cd" through them or print the contents of the file to which they point using the "type" command.???? But this is no reason that Cygnus Win32 can't endow them with the full power of UNIX symlinks.???? This would be a solution that does not require reading a symlink file. It's also much better integrated with Windows.???? Wouldn't it be nice if you could navigate through Cygnus Win32 symlinks using Explorer????? When you've made c:/bin be a symlink/shortcut to the Cygnus godawfully named ".../H-i386-cygwin32/bin" directory, wouldn't it be nice if you could type e.g. "/bin/bash" to the Start / Run window, and have it start up bash? Here are some details of how such links could work: 1) Readdir, when returning the names of files in a directory, ???????? should delete the suffix ".lnk" or ".pif", if present. 2) In the open syscall, if an attempt to open a file fails with ???????? a "not found" error, open should suffix ".lnk" and try ???????? again.???? If that fails, it should suffix ".pif" and try ???????? again.???? If either of these succeeds, open should then read ???????? the shortcut file, find the path name of the target, and do a ???????? Win32 open of the target file.???? This will slow down the ???????? processing of sym links a bit (because a failed open has to ???????? take place first), but allows non-symlink files to be ???????? processed without any overhead added by the symlink code. 3) The other syscalls that take a pathname and are supposed to ???????? follow symlinks (e.g. chdir, stat), should select the target ???????? file or folder similarly to the algorithm used by open. Comments? Michael Condict m.condict@opengroup.org The Open Group Research Inst. (617) 621-7349 11 Cambridge Center Cambridge, MA 02142 - For help on using this list (especially unsubscribing), send a message to " gnu-win32-request@cygnus.com " with one line of text: "help". From hgeleff@spidernet.com.cy Sun May 25 06:34:00 1997 From: hgeleff@spidernet.com.cy (Henrik Geleff) Date: Sun, 25 May 1997 06:34:00 -0000 Subject: SCCS Message-ID: <199705251333.QAA10478@mail0.spidernet.net> There is a problem with a free SCCS, I tried to make one but I could not find a description of the SCCS file format which did not come from AT&T or if it looked as a independent source, they contained one or more paragraphs which was word to word the same as the AT&T description. It is in my opinion impossible to write a copylefted SCCS as it will always deal with the format of the SCCS file which is copyrighted by AT&T. The task of writing a clone of SCCS should be pretty easy, but it will be a clone and not white-roomed-code and therefore not copy-left-able. FSF by the way is about developping tools that are equivalent or better than the comercial tools; for version control the tool is called RCS, and it is better than SCCS and is even used as an engine in CVS and MKS Source Integrity to mention two. I do not think the world needs a SCCS clone as free software, but if anybody disagrees I could provide my design document, which never was coded as I came across RCS and did not see the need. Regards Henrik Geleff ---------- > From: Jason Zions > To: Kevin Hughes > Cc: GNU-Win32@cygnus.com > Subject: Re: SCCS > Date: 22 May 1997 19:27 > > There is no public-domain SCCS code that I'm aware of; it's all under > AT&T copyright. If someone has an SCCS clone source code base that's > truly free or copylefted, I'd love to hear about it as well. > > Jason > > > - > For help on using this list (especially unsubscribing), send a message to > "gnu-win32-request@cygnus.com" with one line of text: "help". - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From mh@mike.franken.de Sun May 25 06:34:00 1997 From: mh@mike.franken.de (Michael Hirmke) Date: Sun, 25 May 1997 06:34:00 -0000 Subject: X11 and B18 References: <01BC67C8.82162220@sos> Message-ID: <6XW-DBh$pfB@mike.franken.de> Hi Sergey, > X11 DLLs built on my NT box works on both NT and W95. The same built on W95 > works on NT only. Today I'll try those builds on my home machine. Any > ideas? Sorry, this probably is a stupid question, but ... I read some postings about compiling X11 for Windows NT/95 - why would one want to do this ? Do I have a complete X11 after that ? Including X-Server and window manager ? Or do I only get all the X client programs ? If I'd get them all, I'd like to try that, too - but without the X server it seems quite useless to do this. > > -- > Sergey Okhapkin bye. Michael. -- Michael Hirmke | Telefon +49 (911) 557999 Georg-Strobel-Strasse 81 | FAX +49 (911) 557664 90489 Nuernberg | E-Mail mh@mike.franken.de | WWW http://minimike.franken.de/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Sun May 25 08:52:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Sun, 25 May 1997 08:52:00 -0000 Subject: X11 and B18 Message-ID: <01BC6943.2AC4D800@sos> Hi! Looking into body of X11 dlls compiled on NT and W95, I see one difference - arrays of zeroes in NT compiled dll and random bytes in the same place in W95 compiled dlls... Is it the cause of troubles with X11 dlls compiled on 95? It seems to me it's because of difference in handling file seeks on NT and W95: #include #include main() { int i=1; int fd; fd=open("test", O_WRONLY | O_CREAT | O_BINARY, 0666); write(fd, &i, 1); lseek(fd, 0x2b7, SEEK_SET); i++; write(fd, &i, 1); lseek(fd, 0x400, SEEK_SET); i++; write(fd, &i, 1); } When running this program on NT, output file contains all zeroes and bytes 1, 2 and 3 in desired file positions. But when running on W95, file contains garbage and desired bytes in desired positions. Geoffrey, probably there is a bug in bfd or ld - they make a seek to fill a portion of a file with zeroes, this seeks results in zeroes on unixes and NT, but in random contents on W95! -- Sergey Okhapkin Moscow, Russia Looking for a job - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From root@jacob.remcomp.fr Sun May 25 08:52:00 1997 From: root@jacob.remcomp.fr (root) Date: Sun, 25 May 1997 08:52:00 -0000 Subject: Interfacing OCX/Ole controls Message-ID: I would like to interface OCX/OLE controls with a C application. Has anyone information about how this can be done? The information I need is the interface definition at the assembly language level (or C language level). Thanks in advance -- Jacob Navia Logiciels/Informatique 41 rue Maurice Ravel Tel 01 48.23.51.44 93430 Villetaneuse Fax 01 48.23.95.39 France - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Paul@chocolat.foobar.co.uk Sun May 25 12:44:00 1997 From: Paul@chocolat.foobar.co.uk (Paul Shirley) Date: Sun, 25 May 1997 12:44:00 -0000 Subject: v18 vs v17 compile times.... References: <199705231513.LAA06332@postman.opengroup.org> Message-ID: In message < 199705231513.LAA06332@postman.opengroup.org >, Michael Condict writes > 1) Readdir, when returning the names of files in a directory, > should delete the suffix ".lnk" or ".pif", if present. > >Comments? Fairly trivial... .lnk is a common type for linker script files in the DOS/Windoze world. With Win95 this is already annoying in explorer, but if they get filtered out by renaming this way it will cause some of us problems. Perhaps you should also check the system flag before stripping the suffix. -- Paul Shirley - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From rminnich@sarnoff.com Sun May 25 12:44:00 1997 From: rminnich@sarnoff.com (Ron G. Minnich) Date: Sun, 25 May 1997 12:44:00 -0000 Subject: More on case insensitivity References: <3.0.32.19970522084806.0096be10@pop.pdx.informix.com> Message-ID: My one thought on all this is that the way emacs handles the case problems is about as good as things can get, IMHO. ron Ron Minnich |Java: an operating-system-independent, rminnich@sarnoff.com |architecture-independent programming language (609)-734-3120 |for Windows/95 and Windows/NT on the Pentium ftp://ftp.sarnoff.com/pub/mnfs/www/docs/cluster.html - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From weiku@UTKUX1.UTK.EDU Sun May 25 13:20:00 1997 From: weiku@UTKUX1.UTK.EDU (Wei Ku) Date: Sun, 25 May 1997 13:20:00 -0000 Subject: X11 and B18 Message-ID: <199705252020.QAA22306@solid.ssd.ornl.gov> There are some X server available: some o fthem are commercial. The following two sites should provide you a taste of the X server on Windows 32 system: 1. http://tnt.microimages.com/www/html/freestuf/mix.htm : TNT free X server, using twm or fvwm. 2. http://www.starnet.com/ : Starnet X demo using window management of Windows itself. ( two-hour limit for demo version ) Sincerely, Wei Ku *************************************** Department of Physics and Astronomy The University of Tennessee 1408 Circle Drive Knoxville, Tennessee 37996-1200 weiku@utkux.utcc.utk.edu --------------------------------------- Solid State Division Oak Ridge National Laboratory P.O.Box 2008 Oak Ridge, TN 37831-6032 Phone: (423) 574-5795 Fax: (423) 574-4143 weiku@solid.ssd.ornl.gov *************************************** ---- From: Michael Hirmke To: gnu-win32@cygnus.com Date: Sunday, May 25, 1997 2:51 PM Subject: Re: X11 and B18 Hi Sergey, > X11 DLLs built on my NT box works on both NT and W95. The same built on W95 > works on NT only. Today I'll try those builds on my home machine. Any > ideas? Sorry, this probably is a stupid question, but ... I read some postings about compiling X11 for Windows NT/95 - why would one want to do this ? Do I have a complete X11 after that ? Including X-Server and window manager ? Or do I only get all the X client programs ? If I'd get them all, I'd like to try that, too - but without the X server it seems quite useless to do this. > > -- > Sergey Okhapkin bye. Michael. -- Michael Hirmke???????????????????????????????????????? | Telefon +49 (911) 557999 Georg-Strobel-Strasse 81 | FAX???????????????? +49 (911) 557664 90489 Nuernberg???????????????????????????????????? | E-Mail???? mh@mike.franken.de ???????????????????????????????????????????????????????????????????????????????????????????????? | WWW???????????????? http://minimike.franken.de/ - For help on using this list (especially unsubscribing), send a message to " gnu-win32-request@cygnus.com " with one line of text: "help". From hawkeye@CommerceWave.com Sun May 25 17:09:00 1997 From: hawkeye@CommerceWave.com (Ken Keys) Date: Sun, 25 May 1997 17:09:00 -0000 Subject: Shortcuts as symlinks (was: v18 vs v17 compile times....) References: <199705231513.LAA06332@postman.opengroup.org> Message-ID: <199705252310.QAA18601@dilbert.commercewave.com> Michael Condict (m.condict@opengroup.org) wrote: > I'd like to step back and point out that Windows 95 / NT 4 already *have* > symlinks. They are ".lnk" files for folders, data files and Win32 programs, > and ".pif" files for 16-bit programs. They have all the power of UNIX > symlinks in what information is stored in them, and more. > > Now it's true that Windows makes very poor use of these symlinks, > allowing them to have meaning only if you double-click on them in an > Explorer window or feed them to the "Run" menu or the "start" command. > You can't "cd" through them or print the contents of the file to which > they point using the "type" command. But this is no reason that Cygnus > Win32 can't endow them with the full power of UNIX symlinks. This would > be a solution that does not require reading a symlink file. > > It's also much better integrated with Windows. Wouldn't it be nice if > you could navigate through Cygnus Win32 symlinks using Explorer? When > you've made c:/bin be a symlink/shortcut to the Cygnus godawfully named > ".../H-i386-cygwin32/bin" directory, wouldn't it be nice if you could > type e.g. "/bin/bash" to the Start / Run window, and have it start up > bash? > > Here are some details of how such links could work: > > 1) Readdir, when returning the names of files in a directory, > should delete the suffix ".lnk" or ".pif", if present. > > 2) In the open syscall, if an attempt to open a file fails with > a "not found" error, open should suffix ".lnk" and try > again. If that fails, it should suffix ".pif" and try > again. If either of these succeeds, open should then read > the shortcut file, find the path name of the target, and do a > Win32 open of the target file. This will slow down the > processing of sym links a bit (because a failed open has to > take place first), but allows non-symlink files to be > processed without any overhead added by the symlink code. > > 3) The other syscalls that take a pathname and are supposed to > follow symlinks (e.g. chdir, stat), should select the target > file or folder similarly to the algorithm used by open. I agree, this would be very useful. It's so obvious that I wonder why it wasn't done in the first place. The only reason I didn't propose it myself several weeks ago is that I didn't know enough about how .lnk and .pif files really worked. There is one minor problem with your proposal: readdir can't just delete the suffix, since the prefix may not be unique (e.g., files "foo" and "foo.lnk" would both look like "foo"). The solution is simply to have readdir() not remove the suffix. This would parallel the way ".exe" and ".bat" files are already handled. Having symlinks be shortcuts would also allow symlinks to executables to work, instead of failing to find the dll as they do currently. (I guess the cygwin exec() would have to expand its search to include files ending in ".lnk" and ".pif", but it wouldn't have to do anything special with them, since the windows calls would take care of the rest.) -- Ken Keys (hawkeye@CommerceWave.com) Software Engineering Team Leader CommerceWave, Inc. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From pete@horus.cix.co.uk Sun May 25 17:09:00 1997 From: pete@horus.cix.co.uk (Pete Jordan) Date: Sun, 25 May 1997 17:09:00 -0000 Subject: Process table Message-ID: Beta 18 doesn't clear out its process table after programs I launch with spawnvp() terminate; the ps list just keeps getting longer until I shut down all instances of bash. Pete Jordan = = = = = = = = = = = = = = = = = = = = = = = Horus Communications http://www.horus.cix.co.uk/ = = = = = = = = = = = = = = = = = = = = = = = "Is a polar bear a rectangular bear after a coordinate transform?" - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From allan@interport.net Sun May 25 17:09:00 1997 From: allan@interport.net (Allan Peda) Date: Sun, 25 May 1997 17:09:00 -0000 Subject: Embarassing question on linking References: Message-ID: <3389D542.523C@interport.net> I am learning C++ using cygwin32, so I am not too well versed in these unix type utilities (previous experience with IDE based products). I'll cut to the chase.. hello world based C++ program yields the following: The code: bash$ cat hellopp.cc #include main() { cout << "Hello there GNU++ world!\n"; } bash$ The error: bash$ make -f hellopp.mak /gnuwin32/b18/H-i386-cygwin32/bin/gcc.exe -pedantic -o hellopp.exe hellopp.cc C:\TEMP\cc0010051.o(.text+0x2a):hellopp.cc: undefined reference to `cout' C:\TEMP\cc0010051.o(.text+0x2f):hellopp.cc: undefined reference to `ostream::ope rator<<(char const *)' gcc.exe: Internal compiler error: program ld got fatal signal 1 make: *** [hellopp.exe] Error 1 Now under hello.c, I have no errors, I'm guessing that I need to do something to link the correct libraries which are needed under (G)C++, and were included by default under GCC (ansi). Thanks Allan (Yes, I am reading the notes at cygnus, and am looking into prep.mit.edu) - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From pete@horus.cix.co.uk Sun May 25 17:09:00 1997 From: pete@horus.cix.co.uk (Pete Jordan) Date: Sun, 25 May 1997 17:09:00 -0000 Subject: Mount table Message-ID: Is there any way of getting at the mount table from a C program (other than by looking at the registry or actually running 'mount')? I ask as I have a little program to launch Windows apps from bash that munges its parameters from cygwin32 format to absolute DOS paths so I can run, say, W4W on a doc from bash. I use a symlink to my program for each app so I can identify from argv[0] (via a table) where the app actually lives and what sort of parameters it takes. ATM, I have my mount table hardcoded into the thing, but that's hardly satisfactory. A minor bitch also: One of the non-cygwin32 apps I use my program for is perl. All works fine if I type "perl scriptname", but if I just type "scriptname", argv[0] in my program is set to the actual program name (/usr/bin/debash.exe), not the symlink name (/usr/bin/perl). I get around this by assuming perl in this situation. A nasty kludge. Pete Jordan = = = = = = = = = = = = = = = = = = = = = = = Horus Communications http://www.horus.cix.co.uk/ = = = = = = = = = = = = = = = = = = = = = = = "Is a polar bear a rectangular bear after a coordinate transform?" - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Mon May 26 00:34:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Mon, 26 May 1997 00:34:00 -0000 Subject: X11 and B18 Message-ID: <01BC69C8.E6B8D5E0@gater.krystalbank.msk.ru> Wei Ku wrote: > <> > There are some X server available: some o fthem are commercial. The following two sites should provide you a taste of the X server on Windows 32 system: > 1. http://tnt.microimages.com/www/html/freestuf/mix.htm : TNT free X server, using twm or fvwm. > 2. http://www.starnet.com/ : Starnet X demo using window management of Windows itself. ( two-hour limit for demo version ) > Sincerely, > Wei Ku 3. Http://www.windows.digital.com - eXcursin 2.0 - the best win32 X server I ever see. -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From bartlee@geocities.com Mon May 26 00:39:00 1997 From: bartlee@geocities.com (Bartlee A. Anderson) Date: Mon, 26 May 1997 00:39:00 -0000 Subject: patch X11R6.3 fails with fix-01 and the b18 patch Message-ID: <33893AA2.203F@geocities.com> I installed the source directory from xc-1_tar.gz per the instructions. Got the fix-01 from ftp.x.org got Sergey's patch tried mounting with -b and without -b flag tried patching with the b18 patch and then reinstalled and tried Sergeys patch after fix-01 and all I get are rejected patch attempts. bash$ patch -p -s Paul Shirley wrote: > In message < 199705231513.LAA06332@postman.opengroup.org >, Michael > Condict writes > > 1) Readdir, when returning the names of files in a directory, > > should delete the suffix ".lnk" or ".pif", if present. > > > >Comments? > > Fairly trivial... .lnk is a common type for linker script files in the > DOS/Windoze world. With Win95 this is already annoying in explorer, but > if they get filtered out by renaming this way it will cause some of us > problems. Perhaps you should also check the system flag before stripping > the suffix. > And one more comment: the format of explorer's .lnk files is undocumented... -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From weiku@UTKUX1.UTK.EDU Mon May 26 00:59:00 1997 From: weiku@UTKUX1.UTK.EDU (Wei Ku) Date: Mon, 26 May 1997 00:59:00 -0000 Subject: Embarassing question on linking Message-ID: <199705260756.DAA37171@solid.ssd.ornl.gov> ---- From: Allan Peda Now under hello.c, I have no errors,???? I'm guessing that I need to do something to link the correct libraries which are needed under (G)C++, and were included by default under GCC (ansi). ---- 1. use g++ instead of gcc 2. use void if the function does not return anything. ( the default return type is int if not specified) The following is the similar code compiled and run in my Windows 95: //g/tmp/tmp ) cat hello.cpp #include ???? void main() { cout << "Hello there GNU++ world!" << endl; } //g/tmp/tmp ) g++ -o hello.exe hello.cpp //g/tmp/tmp ) hello Hello there GNU++ world! ???? Sincerely, Wei Ku *************************************** Department of Physics and Astronomy The University of Tennessee 1408 Circle Drive Knoxville, Tennessee 37996-1200 weiku@utkux.utcc.utk.edu --------------------------------------- Solid State Division Oak Ridge National Laboratory P.O.Box 2008 Oak Ridge, TN 37831-6032 Phone: (423) 574-5795 Fax: (423) 574-4143 weiku@solid.ssd.ornl.gov *************************************** To: gnu-win32@cygnus.com Date: Monday, May 26, 1997 1:01 AM Subject: Embarassing question on linking I am learning C++ using cygwin32, so I am not too well versed in these unix type utilities (previous experience with IDE based products). I'll cut to the chase..???? hello world based C++ program yields the following: The code: bash$ cat hellopp.cc #include main() { ???????????????????????????? cout << "Hello there GNU++ world!\n"; } bash$ The error: bash$ make -f hellopp.mak /gnuwin32/b18/H-i386-cygwin32/bin/gcc.exe -pedantic -o hellopp.exe hellopp.cc C:\TEMP\cc0010051.o(.text+0x2a):hellopp.cc: undefined reference to `cout' C:\TEMP\cc0010051.o(.text+0x2f):hellopp.cc: undefined reference to `ostream::ope rator<<(char const *)' gcc.exe: Internal compiler error: program ld got fatal signal 1 make: *** [hellopp.exe] Error 1 Thanks Allan (Yes, I am reading the notes at cygnus, and am looking into prep.mit.edu) - For help on using this list (especially unsubscribing), send a message to " gnu-win32-request@cygnus.com " with one line of text: "help". From micha@genias.de Mon May 26 00:59:00 1997 From: micha@genias.de (Michaela Kurz) Date: Mon, 26 May 1997 00:59:00 -0000 Subject: Beta 18 breaks gettimeofday References: <3.0.32.19970520174814.02732ca8@silence.secnet.com> Message-ID: On Tue, 20 May 1997, Jonathan Wilkins wrote: > I have just upgraded to beta 18 and was rebuilding an application and had > some troubles.. after tracking things down it seems to be a problem with > the gettimeofday() call.. sample code follows: > > #include > #include > #include > #include > #include > > int main() > { > struct timeval *tvp; > struct timeval tv; > > tvp = &tv; > > if(gettimeofday(tvp, NULL) != 0){ > printf("\n failed \n error was : %s\n", strerror(errno)); > } else { > printf("\n worked \n tv_sec is : %i\n tv_usec is : %i", > tv.tv_sec, tv.tv_usec); > } > return(0); > } > > this code works on freebsd, and other gettimeofday calls used to work on B17.1 > > any ideas? sorry no, but another problem with this function: Where is the prototype? --------------------------------------------------- M. Kurz Genias Software GmbH / Tel: 09401/9200-0 micha@genias.de / Fax: /9200-92 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From fjh@cs.mu.OZ.AU Mon May 26 01:17:00 1997 From: fjh@cs.mu.OZ.AU (Fergus Henderson) Date: Mon, 26 May 1997 01:17:00 -0000 Subject: Mount table References: Message-ID: <199705260816.SAA09641@murlibobo.cs.mu.OZ.AU> Pete Jordan, you wrote: > > Is there any way of getting at the mount table from a C program (other than > by looking at the registry or actually running 'mount')? Yes. See the source code for mount. It is in winsup/util/mount.cc in the source code distribution. -- Fergus Henderson | "I have always known that the pursuit WWW: < http://www.cs.mu.oz.au/~fjh > | of excellence is a lethal habit" PGP: finger fjh@128.250.37.3 | -- the last words of T. S. Garp. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Mon May 26 03:49:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Mon, 26 May 1997 03:49:00 -0000 Subject: My patches Message-ID: <01BC69E4.10E4EEB0@gater.krystalbank.msk.ru> Hi! Here is the changes I've made in winsup code after B18 release. - Fixed clear screen problem on Win95, fill color now set to default console color, - Added support for invisible screen attribute, - Changed AltGr-key behavior - left alt-key produces ESC-key sequence like in B18, but right alt (AltGr)-key should produce characters according to national keyboard layout (not tested). - Interprocess signals works on Win95 too, - Implemented timer stuff - alarm(), set(get)itimer() (ITIMER_REAL only, I have no ideas how to implement ITIMER_VIRTUAL and ITIMER_PROF on win32), - Fixed problem with environment variables uppercasing. "EXPORT foo=bar" in bash produces now "foo" environment variable but not "FOO", - Added mode parameter support for open() syscall, umask() syscall implemented (NT only). - Emulation of unix rwx and "symlink" attributes with EA on NT. - Symlinks are now marked with "system" file attribute, fstat() speed improved. - sleep()/usleep() syscalls are interruptable now, - fixed return code of cwait() call, and some bugs in spawn() code, - fixed stat() of directories (incorrect number of links). To compile modified winsup sources replace file sys/time.h in newlib and cygwin installation with this one: -- Sergey Okhapkin Moscow, Russia Looking for a job. begin 600 time.h M+RH@5&AI2@I('-Y6EN9RYD:B(L(&%V86EL86)L92!F2!O M9B B8V]P>6EN9RYD:B(@2!T:&ES(&9I;&4[(&EF M(&YO="P@82!C;W!Y"BHJ('-H;W5L9"!B92!A=F%I;&%B;&4@9G)O;2!W:&5R M92!T:&ES(&9I;&4@=V%S(&]B=&%I;F5D+B @5&AI2!N M;W0@8F4@9&ES=')I8G5T960@=VET:&]U="!A('9E2!O9B!-15)#2$%.5$%"24Q)5%D@;W(@1DE43D534R!& M3U(@02!005)424-53$%2(%!54E!/4T4N"BHO"@HC:69N9&5F(%]365-?5$E- M15](7PHC:69D968@7U]C<&QUPH@(&QO;F<@ M='9?F]N92![ M"B @:6YT('1Z7VUI;G5T97-W97-T.PH@(&EN="!T>E]D2AS=')U8W0@=&EM979A;" J='8L('-TF]N92 J='HI.PII;G0@'0@=F%L=64@*B\*(" @('-T8'^XH.')>6'4^4JT6_K:^OYU5?.;LV$Y5]> '__D9N*3D?.6Z? M^*. ^+<><:Z'?A 1USD/[&!"(I]9=$ZL#IA$+X*/G^ M-\?KN:,^)?\()^%F-!D"H2_?I%[TL'2Z,(SZCH^%&^E"X/%4U5<"X\M77#XV MU\G!X9=O3,2]@+?=:V 7H#4D6WR M3=<(_"L9^/_'%D*[&C>[\9U^W #^4.H5WNT<[Q]UR*5&'!#1L<'_3L3?L2F> MQ=^Q)YX]E+-O"!;9.SGNGAQUK.[>::=S;.U^.CCHG%J'QP/(O^:=X8":!@&\$=. MT4?QAS%F<&E[?9<&EI"*=CO-JE^00: @+7\4#4>1Q9$K%'\6RS*GF-G2C*HN MIQA#K/P%NA][;[;U/^!/.VLD5NTK5)OP:I.I:E]Y-21;>>_2#LH[8<]Q\!.T M>4U>;S'"L=<[410XYZ.(AO#JX.2T\_;TY-/QOK5[]*E#OJLE;Y%[Z:+3SC[O MJLO0SF #GR-A@+]4TF@,!RA%(!@O?G"])H7PW XI68?>9\72'U*/%/ 1E!XB MM8YVB E>(G7=SMZGT\.SK];.V=GIX>ZGLTZ7A#8G3&B7O2/J7427@'0(%/.1 M=3; P296!11?U9 3ZV=SI4\']LB-`'_7#^;2.UODGX\#&KGV^]1Z-"CB1\[7*74JER-J3-,MB M;7&;<)4-Q2N>CRZL80"$&Q1>S1":(*%I'^VTX_T=.OJ]K^$##,^?_NV]TB3B MR0^"Z'AI"#/ 9FS/<;^;X.U48\W$&8F*W:6>Y%3!&;CVA54D&W&!*PNF@!+O M4^.K#=%N\0(FWS,-[> B5.M[O&"+BRRRK' ^&I#M[41O?^-N6JT*?IH>2\A+ M120I2(G.' 0S.=TT0'RN7-/KWG""`VDPN]F\ M+JYQ"+4I@$6+=%^2*QMYG:6PT[)Q%NW37>>0!;D(E<#+#4MOY/2)HDF[[3H> M^,/0&D$0X K5`*VR%8S9JFD574^(5]$K6L70%R#>#-T8`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`QKI`=E"_=P/<60^1H+XO8;DIR#&G.2/I:YZI&)Q2F<) M=L.H`N6GW!3N*A+A*^HI7S'M*$H>97 (0_:SOME4K')Z&F:H`4DE.<8#>I:\ MG>:LF^;LE!F2+"8JCU/;;>K\RUOE)G(W4:;?PMWH:C=2I.:03[)_:N?6![S/ M[;Y%PYX]I%S;-ZH5K5&;VN9>MC2LI"F;1]BTY"L4?*S,/&A\/37^ZKW&?PIQ MRX)JV9*0J2=JH*&;*3W!<'Q]]#K67SM'+$(=TB BYZ[M73&:ARJ:L0AA,$Q\ M_$-^:F/@:Y84LY18V&+BTX84WV20%(TX.A\D.ON(3I^Z-*(_&1$)^HP)OP.1 M_Y&('.XE?,$@^<]!)*4"QPI*<_#XF#!D+V'(3\1#!3U/&6?@T95X= \`#WZP M"IP`I#;#T MA4K',\R6?TER[-J]*Q+9YT &J>GOPC)E$K@ST*QH32/E#+P0E?\K:O=?49'_ M@CK[Q:KG%ZF)?SFE^P#]^K-5:98WW3),K96DO*S$6Z@K?!.6'0Q>#X,>'L,$ MI55D8)S2&QK RT.O3\XL9H<8\O/@`>V5.S?4B\KOZ81_Z-_N^5X$ M)(62+L)"U@C?^R\<=0[.K)VC,^OC::?;[>R3[^3T\.V[5%FQR '^%L/-3OQ$ MU\,_S;\D+OC(5$[DN_XM#?*AD4D1<3:&T:P"OBU==7B>#=_-]0R,US>?$.>< MA,C)!3R6^TF6;E(RFQ I7ZT;\6TPJ'K1-;0YDD>\@&J6%8($N^*! A\JSQ: MUL''O1[]#8\B`Y6LOAW9OVT0F$(.IMD&\/E"?LXF6[\71+J211H_SY L?L$( MMC,,,"54K[2KM79E-JE95O[@B_S1.C',-B8UF_GDJC1UK=)4]AIY@>+(HR_I M`3ZVZ_R'LK3?8>#W:!BBE0,S`40()F422_L0SXM8#K00856VBY?NY-+I]ZDG MF=K% M.X)9%X"I]T?(3O+JX\[9NU<8@@O]:\I35G&*PUAGETY(;!>JAU -I.95.>EF M=Q2A\L9N@#($C"FFN=K$L_&0'@@']6X(\$CT#_5>@Y>&<3S/9Z.4E?2.)*+7 M]3'Y>C0&')(=LHX(U&$*#POJH] 8P&$9!\4_GKRVH0%;! M0FV_)FMK6)L,>43]`)!^ M#<@!-D#^H1\Z8X+3--20&HRNX9#V0#Y$4C$"BI1!XB9D_!.Y,/2!]"A*[H3X MF' <_L5X22?DFB50GH/5]8!/[#3E`.U>!+S'@3#I0)PX-4%RIH+>_X\E)^GE M,R,WS"O?`_K%9!XX01A)C0#4`W[ )^P`=]Z',Y*',@$2%%)W4'HS=/KHWB3/ M4) (X3PAE?Z.*JBBV:*B>H>PWBFN+UI@LP^7-+5ZI:(<+DF?M(GSNE(\6IWA M$8?RRG'=0NJ-1KJ';_?>'>V+E!^AU86O&M%XJZWG^B&UT*RB/@_E%ISL3+:0 MKT+@,NU;Z1?9%C8^`JD86:5HQLXJ[YBS@::VHN-E#15P-O094ZO6CZVM`4:V MT@:#6[W#.:D88%P-Q4D6Z4SGDXB&UFW@1!$F@0FA6D6!.L&DUG 21O0:)B], M=S&VO%.#[U,E-R_L,_Y[%/"PR?]BVNMK7$B/KD&4_A!F@ M8&&)WI\`Y+1W!83 .19.KEW'NPJ718 XI9C]>0@U,@7$K&J5BLR]YW(_"7N* M.ZHDY*J0_@X`Z>/?QT5,OB4<5LXH>2B!Y6Z9! MX =):J?2U#@\838K\/4;,Q+<&^#;KB4TZ'X'Y#2/K MW/5[5R%!.8V+4(;(!NE:NT?ONX?_ZL#(9#-YW$JU'X'C@#_;+,F?L=>_]="# MRF%M,=W^8K8]=RGO:+^JM$3L MW19W#+W7$;E /PYM(//).;8TZG&C+4$4ELP:6KD#6X<'1\?O.>8L4VGG;>L"R0%'=,>KO\%+5:R>?"%\WGNL!\//W9RA^V> M[+W/ZN-'KGG30?H&&$VR7^1_Q_.$IYGS\*Y,Y&D9V->?2'@7+BF9G .D%+- MD7NJ.K($=76H_:62_1%U&":Z'/5X1R5F =MK!:5[<4[>I3J<8X$R;C& MK)K,XUVB*N\'BE29>:!(M9G3WP\IZ2]=C6;N-C2JFJ%79#[(_-NJ^DYXQ;36 M4M?8V 29; V3E!8LP_D2EY5$V,H(&N'F+M&'Q">=ZZ%(,TF!0T$5Y MU^*P+BQY`BJ#^.D5E>QX=E7%]+N$<"ILJM[V(.^=$MJ5A4UBZJ*0L>WC-8X; M?. DF3INC*.Q;K9Q01+/>'6#?%J(?D%(W_@2XF$W+E1Q"Q)("O,ML M-=E25T.U\>YVT!M.2$$*/>]R:EG)S$8651.)62.KG"P,X_QN4Y:(W,$)W/"7 MC!CXKNO?QIQ0>LUF`4K:>P20[Y'Y-$3K2L<.*!@PZBS,C"R R3:TPQ#FGEB1 M3^='(([\Y$')0!R9<11ADB)&@SO')QV8G=^_SR[N^6U@Q:FCZ"] #E2 5+VX M.B,?TA;Q@+Y'<7/##B8:QN]O[,#Q1R$/R/,8/DOX"2D[LX&*D\UQW!2Y".QK MOC6Y.CT?F>@(+L;G!V)-P^:?E1AM1:"2(!@W`5,W8*F!-$5<% 4<8\^]*"%K MUI +ON(GP!L.#P_MM/"\@V[*C:_E&KH'QE+O8_6>V;[]7(7_>$T^J\7%S'WB M6:O.V&4K[$=HU:5HU&5HTQE-^HS\6,6PY&J&YGQZ;?FL"E()+3]$36:>BZOA M?>4U0X97OVW=0W/2&\L;N6Z6XIRZ3SA,%JWBTO(BV>_\T\*I;O$"ZY\[1Y\Z MZ:LFD2CB(%M-KP" N@S:+P7 V>MV'P5H-F4;30"\65V4L@/7'PXG"ZV^,'G! MO;4G(;OY%H25]\'U`)-HH:D%U)G2I@@:S.__QA,\9D/#`&P:C479<&]L9EGR MM%AE\ZR%6+8:B_(LLH?T<1S#'G(Q>Q##T"&J)0= EXS*/=BU3)3F'*.XG#GY M<)E_B.(R=5RQVJZUVD8S_PS%I?P*%M/$`XN59AN_BR7W?"?,>KV:[%:-O-"Y MP-NH71_\IGA?#/?[P+9^B\DA]P*W4N;JHWK Y@9,,C^8>.,$T0A=M?AK2\JQ M49)OTDR>+YEGXOW> M=5_CWYBP7HQ==D8I&?'_-2B56@ILZ\]-N.P+WFN:T3!BF1N.SEVGUTZO4!)= M/'/#)/I*B5?R* GZ(7R3G@O.9H9>$;"E.Q=WTG,/LPE#+9U !/ MIM&:RR;L]CF8A&E8$;3'G#!TQ_D:`F.8R \![#Q^/"&P=W$C%_9,PC?KFI'< MI)V-2[P>>4K*I^XFS=+'W($ZGI;](()$)?FTF6_&&RNX5UB?B#F9EC*E;6]3V`77SQ^#?/X'7QL:'4!E'I%WDK.$; MD9:XQ+"0O)7L@S5.1JFTNYSF#>44U(-HOI%]I;6"N%]UG3AA*_+8F' #"V"IV))(:\V_UX.G7U1AWJ$[[D1SM*&SX%_YH%G.N4X'A$/)\MNW)%P_L*T;T^?[)4^+]`I<"%;.I52JFH@]&3A\@G#[[))4* M(1?I"N)PDZRPRKY7DWJP`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`<-YV;:_YF-->0SNG M;=T3CW7BD-PZAVB W4PU69SI8"H2/Z!$G6'8/75=O\SGNNT.=.5%MWZN,NK0 MO+M&M.?^OACZ(ET#\- G;B_0W%SI1#:7=M2;]J:/^\8K3>KC+?XR4X>X0?.- M,\+_V2F4*^:@FQS7T9_337_C_OC"TRE?*GGYXUS4(?D3EX34[D:D8JU()0YK M2I.;B@]LON1[0&/4)_A7I^]IW;R%\.2=F*P*PIU8/\9V8IUN=U<_3V'?CRM^ MW)+8YM4QW[,8ZKU2S#VR$^.BXY=\=]#)-6L8YD_8D UQH>SR>0M=R_(YZPJZ M&?'\T/4&M(*__E&_K+9^$D:4\/5K)\!G*%3""0`.51\,OE8#D[IG5U<_U&L8 MG*BAJ9/LS&U M"&N.G)<,L^9SH>OP> 0G>MD\33@,H*#B?B!'2F@T"GW5=H"/KIK I-F28`.K;'2UD$/R%X5'-C;XH-&1[GT)KIFI(M6-+EC% QV@DLRXJM@U&+SR3)U+$ M%@WH",4CORW/V%?!,[XTLQJ39T>J/::^4-/-"F\![@+TIX(:='U*)( MF&"U?VIW:A(89B-X@0MLN79UK"7*PW:HN MJ>^\%>DHU%^%5B./5=<=V3M_15T-_ PO=) M;.LB*;:UBW?Q@W8AOKW+_-)1")RT"Y5"=L<67RJ'D .^OX>$DY(;,5QW`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`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`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``-#C0/2K^-;^Q$>\6?]0&$D!WS!3H'HXVA6M9@C8ZWGRI7- MC6?.&:$I#EZIG#K FKNF3MTS:(LD`S?JSP;!'71*(1=_,P8FB7N+MXZ"DAP9 M)_CB13C3>;V%OFX-FDG_OA:_M:M5!BV5$;14+KN8F__"@?@O'81D;!0`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`OHWD#4A?52'/:PW(U.T M#>IWYH.-N=VBU[#)&>O$\*,S\HT23[]98J!YM.L/K2!%&U:_702SS26[8DAX M2Z.A7]LQ$SU+^BW3MY0Y_>*^GXECO!GJ;J^3SWHS1($LS:C($>&[Y0CO:;,^ M`:8W(\NX"=>:VPZ@1CVQV_O< M9=Q73^MNT]MM7!1YONBM;#*?+R(39H3@<8R:8H 2+Z:X2D36-HHDC[M&I_6^ M%N>8\_$*1+C^2)8165F4Q&;#ZR->AT[MJH-_20IS3/KMGB^@0@?B;IF1/]98 M(LI";!>,;W3.5\Q4&HUF840P5XCCC^M*[:4YKR9JM^1582 M[^IOWX%0(SB%P+9I7Z0X6![=!DF%YB,Y( +&8$>E'H\^'Q.-7\ZPJ/]$I+CAJ_#M)-]"6*C!H-'5)KBGH5VB*Y M(A!;5_U)`3*SRT;L"29BCTB=+JT/R7R/:;>&\97D_"T\G\C7<(!E`?8O8>XO MSD/[":R:WML]A^B3P]=1_A*S(S'/)I8I*E,_N$SLCBM"6)QC+E7RV;R.91R^ M$HO4CO;!XTOGRRU-A.?G8;['6-$Q=3E(,@T*35WL.3S*H;/@?NWAE=V:*Z7_ M6>Y."U:$C12S92S/XJ='\<;J.D^3\[B#:5Y'#JN[,[A%K+;0R@VSITYY.QFC M^A&#,:.8$1H@>DY2)-&+B-,!5)#&K1)CF7,%-"TK9BO97:"'LEV+% C M$&PLF))Q<[=M0,C3]:2P>U M#Q+=OKFEX=N696\\P5@OP"E'R:YCLCDOEQ6N8YBF5:#!7@RE1Z@5`NY-AP"[ M+7I+'5;&=3$Z6FKHWFDJY:8;T7A/@@+)D16L8)]6UF# TX<`?!X9'.!T?@>T M##84O0VM>07:RYK/G%I/XD2/.A.6_J MLK];"3]BF->J(D__4QV5R"FYLI?+B]F9R^#9;!:SY=;`(S%G.\5 Sergey Okhapkin wrote: > Here is the changes I've made in winsup code after B18 release. > Precompiled cygwin.dll and libcygwin.a (cygwin-sos.052697.tar.gz) and X11 DLLs (X11-dll.tar.gz) uploaded to ftp://dimcom.uqac.uquebec.ca/incoming . -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From chris@actrix.gen.nz Mon May 26 04:33:00 1997 From: chris@actrix.gen.nz (Chris Double) Date: Mon, 26 May 1997 04:33:00 -0000 Subject: Embarassing question on linking Message-ID: <199705261133.XAA28921@atlantis.actrix.gen.nz> >gcc.exe: Internal compiler error: program ld got fatal signal 1 >make: *** [hellopp.exe] Error 1 > >Now under hello.c, I have no errors, I'm guessing that I need to do >something to link the correct libraries which are needed under (G)C++, >and were included by default under GCC (ansi). You need to link against the stdc++ library. I used your source code and saved it as hello.cpp. The following line compiled and created the executable a.exe: gcc hello.cpp -lstdc++ Regards, Chris Double. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From pete@horus.cix.co.uk Mon May 26 05:59:00 1997 From: pete@horus.cix.co.uk (Pete Jordan) Date: Mon, 26 May 1997 05:59:00 -0000 Subject: Mount table References: <199705260816.SAA09641@murlibobo.cs.mu.OZ.AU> Message-ID: Fergus Henderson wrote: > Yes. See the source code for mount. > It is in winsup/util/mount.cc in the source code distribution. [FX: hits self violently about the head] That's the first place I should have thought of looking... Thanks. I don't have any of the sources here ATM, but what's another 7MB down my dialup link anyhow :) Pete Jordan = = = = = = = = = = = = = = = = = = = = = = = Horus Communications http://www.horus.cix.co.uk/ = = = = = = = = = = = = = = = = = = = = = = = "As soon as we have defined something, we have ipso facto created three things: whatever is included in our definition, whatever is not included in our definition, and the relationship between them." - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From *jeffdb@netzone.nospam.com Mon May 26 07:53:00 1997 From: *jeffdb@netzone.nospam.com (Mikey) Date: Mon, 26 May 1997 07:53:00 -0000 Subject: Shortcuts as symlinks (was: v18 vs v17 compile times....) Message-ID: <199705261443.HAA12756@nz1.netzone.com> Ken symlinks to executables do work. (at least on win95 with binary mounts) type in cd / ln -s /usr/H-i386-cygwin32/bin/bash.exe /bin/sh and see what happens. note do not link to sh.exe link to sh. (*jeffdb@netzone.com) the return address for this message is anti spammed remove * from the above address to reply. Mikey ---- From: Ken Keys >Michael Condict (m.condict@opengroup.org) wrote: > >> I'd like to step back and point out that Windows 95 / NT 4 already *have* >> symlinks. They are ".lnk" files for folders, data files and Win32 programs, >> and ".pif" files for 16-bit programs. They have all the power of UNIX >> symlinks in what information is stored in them, and more. >> >> Now it's true that Windows makes very poor use of these symlinks, >> allowing them to have meaning only if you double-click on them in an >> Explorer window or feed them to the "Run" menu or the "start" command. >> You can't "cd" through them or print the contents of the file to which >> they point using the "type" command. But this is no reason that Cygnus >> Win32 can't endow them with the full power of UNIX symlinks. This would >> be a solution that does not require reading a symlink file. >> >> It's also much better integrated with Windows. Wouldn't it be nice if >> you could navigate through Cygnus Win32 symlinks using Explorer? When >> you've made c:/bin be a symlink/shortcut to the Cygnus godawfully named >> ".../H-i386-cygwin32/bin" directory, wouldn't it be nice if you could >> type e.g. "/bin/bash" to the Start / Run window, and have it start up >> bash? >> >> Here are some details of how such links could work: >> >> 1) Readdir, when returning the names of files in a directory, >> should delete the suffix ".lnk" or ".pif", if present. >> >> 2) In the open syscall, if an attempt to open a file fails with >> a "not found" error, open should suffix ".lnk" and try >> again. If that fails, it should suffix ".pif" and try >> again. If either of these succeeds, open should then read >> the shortcut file, find the path name of the target, and do a >> Win32 open of the target file. This will slow down the >> processing of sym links a bit (because a failed open has to >> take place first), but allows non-symlink files to be >> processed without any overhead added by the symlink code. >> >> 3) The other syscalls that take a pathname and are supposed to >> follow symlinks (e.g. chdir, stat), should select the target >> file or folder similarly to the algorithm used by open. > >I agree, this would be very useful. It's so obvious that I wonder why >it wasn't done in the first place. The only reason I didn't propose it >myself several weeks ago is that I didn't know enough about how .lnk and >.pif files really worked. > >There is one minor problem with your proposal: readdir can't just delete >the suffix, since the prefix may not be unique (e.g., files "foo" and >"foo.lnk" would both look like "foo"). The solution is simply to have >readdir() not remove the suffix. This would parallel the way ".exe" and >".bat" files are already handled. > >Having symlinks be shortcuts would also allow symlinks to executables to >work, instead of failing to find the dll as they do currently. (I guess >the cygwin exec() would have to expand its search to include files >ending in ".lnk" and ".pif", but it wouldn't have to do anything >special with them, since the windows calls would take care of the rest.) - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From papresco@calum.csclub.uwaterloo.ca Mon May 26 08:47:00 1997 From: papresco@calum.csclub.uwaterloo.ca (Paul Prescod) Date: Mon, 26 May 1997 08:47:00 -0000 Subject: Embarassing question on linking References: <3389D542.523C@interport.net> Message-ID: <199705261547.LAA19774@calum.csclub.uwaterloo.ca> Change your makefile to use g++ rather than gcc. g++ links in the c++ libraries. Paul Prescod - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From stratos@dnn.gr Mon May 26 09:48:00 1997 From: stratos@dnn.gr (Stratos Margaritis - System Admin) Date: Mon, 26 May 1997 09:48:00 -0000 Subject: man under NT ??? Message-ID: <97052619472618800@athena.dnn.gr> Hello to everybody, Does anyone knows if there exists a unix man command that runs under winNT 4.0 ?? Stratos :-) System Administrator & Software Engineer at DNNet ---------------------------------------------------------------------------- Name -> Stratos Margaritis e-mail -> stratos@dnn.gr, webmaster@dnn.gr, hostmaster@dnn.gr, postmaster@dnn.gr WWW -> http://www.dnn.gr/~stratos/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From mh@mike.franken.de Mon May 26 14:51:00 1997 From: mh@mike.franken.de (Michael Hirmke) Date: Mon, 26 May 1997 14:51:00 -0000 Subject: X11 and B18 References: <199705252020.QAA22306@solid.ssd.ornl.gov> Message-ID: <6Xa6J9s4pfB@mike.franken.de> Hi Wei, [...] > There are some X server available: some o fthem are commercial. The = > following two sites should provide you a taste of the X server on = > Windows 32 system: [...] Thx a lot for your suggestions - I already use an X-Server for NT (Xoftware now part of the Netmanage Chameleon suite). I just have to decide, what to do for an NT installation with about 100000 workstations and 6000 servers. If I could compile and use some free X-Server - can you imagine, how much money that would save us ? > Sincerely, > Wei Ku Bye. Michael. -- Michael Hirmke | Telefon +49 (911) 557999 Georg-Strobel-Strasse 81 | FAX +49 (911) 557664 90489 Nuernberg | E-Mail mh@mike.franken.de | WWW http://minimike.franken.de/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From hgeleff@spidernet.com.cy Mon May 26 14:51:00 1997 From: hgeleff@spidernet.com.cy (Henrik Geleff) Date: Mon, 26 May 1997 14:51:00 -0000 Subject: SCCS Message-ID: <199705261607.TAA06848@mail0.spidernet.net> NO!!! I am saying that if you have a need for a "free" SCCS you can clone the comercial one for personal or internal use, and as long as the change of hands of the code does not involve payment or the like, AT&T will have a ultra thin case against what is going on. Secondly I am saying that I do not see the need for a free SCCS because RCS is in existance and I still have to find a platform where it does not run. Additionally there exist a tool that translates SCCS files to RCS files and RCS files are binary compatible on all platforms finally any project maintained using RCS is not restricted to AT&T supported platforms. Regards Henrik Geleff > Are you saying that you don't think a person could sit with a purchased > copy of sccs and reverse engineer sccs (which I think they could), or that > the process of doing so would cause the code to not be copyrightable by > the author (which I am not sure of)? If so, what part of the process do you > believe to cause the result to not be copyrightable? > thanks, tim. ---------- >> From: Henrik Geleff >> Subject: Re: SCCS >> >> There is a problem with a free SCCS, I tried to make one but I could not >> find a description of the SCCS file format which did not come from AT&T or >> if it looked as a independent source, they contained one or more paragraphs >> which was word to word the same as the AT&T description. It is in my >> opinion impossible to write a copylefted SCCS as it will always deal with >> the format of the SCCS file which is copyrighted by AT&T. >> >> The task of writing a clone of SCCS should be pretty easy, but it will be a >> clone and not white-roomed-code and therefore not copy-left-able. >> >> FSF by the way is about developping tools that are equivalent or better >> than the comercial tools; for version control the tool is called RCS, and >> it is better than SCCS and is even used as an engine in CVS and MKS Source >> Integrity to mention two. >> >> I do not think the world needs a SCCS clone as free software, but if >> anybody disagrees I could provide my design document, which never was coded >> as I came across RCS and did not see the need. >> >> Regards >> >> Henrik Geleff >> - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From mh@mike.franken.de Mon May 26 14:51:00 1997 From: mh@mike.franken.de (Michael Hirmke) Date: Mon, 26 May 1997 14:51:00 -0000 Subject: Beta 18 breaks gettimeofday References: <3.0.32.19970520174814.02732ca8@silence.secnet.com> Message-ID: <6Xa7XCippfB@mike.franken.de> Jonathan Wilkins [jwilkins@secnet.com] wrote: [...] > this code works on freebsd, and other gettimeofday calls used to work on > B17.1 > > any ideas? gettimeofday gave return code 1 == EPERM ! Does this mean, there is a wrong permission check somewhere ? Or is this just accidential ? > > thanks > Jonathan Bye. Michael. -- Michael Hirmke | Telefon +49 (911) 557999 Georg-Strobel-Strasse 81 | FAX +49 (911) 557664 90489 Nuernberg | E-Mail mh@mike.franken.de | WWW http://minimike.franken.de/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jwilkins@secnet.com Mon May 26 14:53:00 1997 From: jwilkins@secnet.com (Jonathan Wilkins) Date: Mon, 26 May 1997 14:53:00 -0000 Subject: Beta 18 breaks gettimeofday Message-ID: <3.0.32.19970526122602.027fd614@silence.secnet.com> At 10:00 AM 5/26/97 +0200, Michaela Kurz wrote: >On Tue, 20 May 1997, Jonathan Wilkins wrote: > >> I have just upgraded to beta 18 and was rebuilding an application and had >> some troubles.. after tracking things down it seems to be a problem with >> the gettimeofday() call.. sample code follows: >> if(gettimeofday(tvp, NULL) != 0){ >> printf("\n failed \n error was : %s\n", strerror(errno)); >> } else { >> printf("\n worked \n tv_sec is : %i\n tv_usec is : %i", >> tv.tv_sec, tv.tv_usec); >> } in regards to this problem.. it was simply a problem with the returned result from gettimeofday().. all other *NIX's seem to return 0 on success, and the cygwin returns non-0 on success.. perhaps this should be changed.. Jonathan -=-=-=-=-=-=-=- Jonathan Wilkins | Futuaris | If only they had used their jwilkins@secnet.com | Non Irresus | terminals for niceness instead http://www.secnet.com | Ridebus | of evil ...-Maxwell Smart - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jim@jimpick.com Mon May 26 14:53:00 1997 From: jim@jimpick.com (Jim Pick) Date: Mon, 26 May 1997 14:53:00 -0000 Subject: My patches References: <01BC69E9.56340F00@gater.krystalbank.msk.ru> Message-ID: <199705262152.OAA31615@fleming.jimpick.com> > Sergey Okhapkin wrote: > > > Here is the changes I've made in winsup code after B18 release. > > > > Precompiled cygwin.dll and libcygwin.a (cygwin-sos.052697.tar.gz) and X11 > DLLs (X11-dll.tar.gz) uploaded to ftp://dimcom.uqac.uquebec.ca/incoming . How do I get the files from here? I just get "User anonymous access denied." Or do I just need to be patient? :-) Cheers, - Jim -- -----BEGIN PGP MESSAGE----- Version: 2.6.3a iQCVAwUBM4oGIeQz770qyIfJAQFTXgP8D+KQuVEG4Mdb0fbn/s8RHU4YRfMkdfqp TDziiL4Bjg6KiWUairqaJkdxu5zjLlkLrIizHpLDwe1xRB6Bw26hcWQmFzoIYt/m YORQpBeXZxaGiEWb4CyeCux4ZeAx3OUcK1hf0roM2k8wcgY7wn6PO93c5FSGnRhB rPxPoz8Qp4c= =8PQ4 -----END PGP MESSAGE----- From jcassidy@stratos.net Mon May 26 14:53:00 1997 From: jcassidy@stratos.net (James M. J. Cassidy) Date: Mon, 26 May 1997 14:53:00 -0000 Subject: Embarassing question on linking References: <3389D542.523C@interport.net> Message-ID: On Mon, 26 May 1997, Allan Peda wrote: > I am learning C++ using cygwin32, so I am not too well versed in these > unix type utilities (previous experience with IDE based products). > > I'll cut to the chase.. hello world based C++ program yields the > following: > > The code: > > bash$ cat hellopp.cc > #include > > main() > { > cout << "Hello there GNU++ world!\n"; > } > bash$ > > The error: > > bash$ make -f hellopp.mak > /gnuwin32/b18/H-i386-cygwin32/bin/gcc.exe -pedantic -o hellopp.exe > hellopp.cc Add '-lg++' to the line above > C:\TEMP\cc0010051.o(.text+0x2a):hellopp.cc: undefined reference to > `cout' > C:\TEMP\cc0010051.o(.text+0x2f):hellopp.cc: undefined reference to > `ostream::ope > rator<<(char const *)' > gcc.exe: Internal compiler error: program ld got fatal signal 1 > make: *** [hellopp.exe] Error 1 > > Now under hello.c, I have no errors, I'm guessing that I need to do > something to link the correct libraries which are needed under (G)C++, > and were included by default under GCC (ansi). > > Thanks > Allan > > (Yes, I am reading the notes at cygnus, and am looking into > prep.mit.edu) > - > For help on using this list (especially unsubscribing), send a message to > "gnu-win32-request@cygnus.com" with one line of text: "help". > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From bill@dstc.edu.au Mon May 26 20:06:00 1997 From: bill@dstc.edu.au (Bill Segall) Date: Mon, 26 May 1997 20:06:00 -0000 Subject: man under NT ??? References: <97052619472618800@athena.dnn.gr> Message-ID: <199705270306.NAA20626@piglet.dstc.edu.au> "Stratos" == Stratos Margaritis <- System Admin > writes: Stratos> Hello to everybody, Does anyone knows if there exists a unix man Stratos> command that runs under winNT 4.0 ?? I've taken to using vim on .man .1 files etc. It does a lovely job. Bill. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From bartlee@geocities.com Mon May 26 21:48:00 1997 From: bartlee@geocities.com (Bartlee A. Anderson) Date: Mon, 26 May 1997 21:48:00 -0000 Subject: patch X11R6.3 fails with fix-01 and the b18 Message-ID: <338A670C.4A10@geocities.com> Found out what the problem is with the patches rejecting... Things move along and there are about 11 or more patches now against the version of x11R6.3 that I downloaded. I'll try getting the patches and reverse them so that the patches will apply to what they should have been. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Mon May 26 23:55:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Mon, 26 May 1997 23:55:00 -0000 Subject: Mount table Message-ID: <01BC6A8C.A7AE1260@gater.krystalbank.msk.ru> Pete Jordan wrote: > I don't have any of the sources here ATM, but what's another 7MB down my > dialup link anyhow :) > #include #include main() { FILE *m; struct mntent *mnt; if(m=setmntent(MOUNTED, "r")) { while(mnt=getmntent(m)) printf("%-30s\t%s\n", mnt->mnt_fsname, mnt->mnt_dir); endmntent(m); } } -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Tue May 27 00:54:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Tue, 27 May 1997 00:54:00 -0000 Subject: My patches Message-ID: <01BC6A94.E93F8710@gater.krystalbank.msk.ru> Jim Pick wrote: > > Precompiled cygwin.dll and libcygwin.a (cygwin-sos.052697.tar.gz) and X11 > > DLLs (X11-dll.tar.gz) uploaded to ftp://dimcom.uqac.uquebec.ca/incoming . > > How do I get the files from here? I just get "User anonymous access denied." > Let's wait for ftp site maintainer will move these files to /pub/gnu-win32/bin... -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From x4278303@turing.ugr.es Tue May 27 00:54:00 1997 From: x4278303@turing.ugr.es (Pelos) Date: Tue, 27 May 1997 00:54:00 -0000 Subject: Spanish keyboard under Win95 Message-ID: I have installed b18 under my Winslow95 with a Spanish keyboard. I have compiled some programs without problems, but when I try to use bash I can't type '\' '|' and some other control characters. Is it a problem of W95 or a problem of cygnus.dll? Suggestions are welcome... (my english is pathetic :( Si tu problema tiene solucion, de que te preocupas? Si tu problema no tiene solucion, de que te preocupas? ___ _ | _ \___| |___ ___ Miguel Hernandez Martos | _/ -_) / _ (_-< e-mail: x4278303@turing.ugr.es |_| \___|_\___/__/ Granada (Spain) - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Me@myhost.com Tue May 27 02:25:00 1997 From: Me@myhost.com (Mikey) Date: Tue, 27 May 1997 02:25:00 -0000 Subject: still ^M under win95 Message-ID: <3389104b.5888722@smtp.netzone.com> Sergey I used your patch against beta 18 sources. Everything works fine, except for I am still getting extra ^M's at the physical EOL in lines longer than 80 chars using linux termcap/info , I tried to download your compiled dll, but I guess the school computers were busy, couldn't connect. Update the files haven't yet been moved out of /incoming, anonymous ftp gives access denied. Appended are the test.txt file, that I have been using to check this behavior, just two extremely long "sentences" of garbage chars, my startup batch file, usually invoked as "u.bat W32 usr" from a shortcut, and my .bashrc my .termcap file is the one from B18, and the behavior is the same whether I use .terminfo, or .termcap fallback support, obviously I can't check termcap, cause bash isn't full screen ;^) One strange thing, when I connect to my ISP using telnet, with this same termcap as .termcap in my ~/ everything works fine, no extra ^M, ctrl keys work, scrolling is smooth, and appears to work like it should? any Ideas? (*jeffdb@netzone.nospam.com) Remove *&.nospam from the above to reply Mikey -------------- next part -------------- A non-text attachment was scrubbed... Name: u.bat Type: text/x-msdos-batch Size: 9506 bytes Desc: not available URL: From pete@horus.cix.co.uk Tue May 27 04:08:00 1997 From: pete@horus.cix.co.uk (Pete Jordan) Date: Tue, 27 May 1997 04:08:00 -0000 Subject: Mount table References: <01BC6A8C.A7AE1260@gater.krystalbank.msk.ru> Message-ID: Many thanks Sergey :) Pete Jordan = = = = = = = = = = = = = = = = = = = = = = = Horus Communications http://www.horus.cix.co.uk/ = = = = = = = = = = = = = = = = = = = = = = = "Is a polar bear a rectangular bear after a coordinate transform?" - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Tue May 27 05:25:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Tue, 27 May 1997 05:25:00 -0000 Subject: still ^M under win95 Message-ID: <01BC6ABA.DAE89C30@gater.krystalbank.msk.ru> Mikey wrote: > <><> > Sergey > > I used your patch against beta 18 sources. > Everything works fine, except for I am still getting extra ^M's at the > physical EOL in lines longer than 80 chars using linux termcap/info , > Appended are the test.txt file, that I have been using to > check this behavior, just two extremely long "sentences" of garbage Please, be more specific! What is the extra ^M you have and what is test.txt file you mentioned? > One strange thing, when I connect to my ISP using telnet, with > this same termcap as .termcap in my ~/ everything works fine, no extra > ^M, ctrl keys work, scrolling is smooth, and appears to work like it > should? any Ideas? Looks like your local bug:-) -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From m.gregory@rmit.edu.au Tue May 27 05:35:00 1997 From: m.gregory@rmit.edu.au (Mark A Gregory) Date: Tue, 27 May 1997 05:35:00 -0000 Subject: Need help with some missing functions Message-ID: <01BC6AEE.9126C890@rem18.co.rmit.edu.au> Hi, I'm trying to compile a program and I have got it down to these missing functions: /empire/emp4/lib/libcommands.a(info.o)(.text+0x37e):info.c: undefined reference to `errno' /empire/emp4/lib/libcommands.a(info.o)(.text+0x552):info.c: undefined reference to `alphasort' /empire/emp4/lib/libcommands.a(info.o)(.text+0x56c):info.c: undefined reference to `scandir' /empire/emp4/lib/libcommon.a(hours.o)(.text+0xaf):hours.c: undefined reference to `errno' /empire/emp4/lib/libcommon.a(wantupd.o)(.text+0x96):wantupd.c: undefined reference to `errno' /empire/emp4/lib/libgen.a(io.o)(.text+0xc33):io.c: undefined reference to `shutdown' /empire/emp4/lib/libgen.a(io.o)(.text+0xc5c):io.c: undefined reference to `shutdown' /empire/emp4/lib/liblwp.a(lwp.o)(.text+0x39b):lwp.c: undefined reference to `sigmask' /empire/emp4/lib/liblwp.a(lwp.o)(.text+0x65e):lwp.c: undefined reference to `lwpInitContext' would someone please tell me if the gnu-win32 has alternatives or a library that I need to include. please email, thank you Mark A Gregory Senior Lecturer Department of Communication and Electronic Engineering Royal Melbourne Institute of Technology - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From banders@switch.rockwell.com Tue May 27 08:11:00 1997 From: banders@switch.rockwell.com (Bart Anderson) Date: Tue, 27 May 1997 08:11:00 -0000 Subject: X11R6.3 nolonger works under B18 References: <3.0.1.32.19970519100433.00785b18@mail.cs.unc.edu> <3.0.1.16.19970523080517.47d7e5a6@btmpf3.se.bel.alcatel.be> Message-ID: <338AFA26.13A4@switch.rockwell.com> Yes, Since imake is looking for cc you have to give it one. Tried to use alias cc=gcc and it didn't like it. It might be how imake is trying to exec it, but I didn't want to get to deep, I just typed "ln /usr/bin/cc /bin/gcc" /usr/bin is a place I keep the non-distrib exe's that I have to get, so I'll have a list of what I need for the next Beta (19.x?, 18.2?). Bart - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From kemp@eease.com Tue May 27 08:11:00 1997 From: kemp@eease.com (John Kemp) Date: Tue, 27 May 1997 08:11:00 -0000 Subject: Libcrypt Message-ID: <338AE01B.5AB8@eease.com> Is there one for gnu-win32? If not, is there one in progress? Thanks, John -- ............................ John Kemp Employease Inc. http://www.employease.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From dahms@ifk20.mach.uni-karlsruhe.de Tue May 27 12:18:00 1997 From: dahms@ifk20.mach.uni-karlsruhe.de (dahms@ifk20.mach.uni-karlsruhe.de) Date: Tue, 27 May 1997 12:18:00 -0000 Subject: SCCS Message-ID: <009B4E63.ECA4BDE0.16030@ifk20.mach.uni-karlsruhe.de> Hi Henrik, you wrote: : Additionally there exist a tool that translates SCCS files to RCS files and : RCS files are binary compatible on all platforms finally any project : maintained using RCS is not restricted to AT&T supported platforms. Does RCS open the RCS file in binary mode to avoid to CR/LF issue for real binary portability? Bye, Heribert (dahms@ifk20.mach.uni-karlsruhe.de) - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jman@lx.net Tue May 27 14:55:00 1997 From: jman@lx.net (jman) Date: Tue, 27 May 1997 14:55:00 -0000 Subject: Missing files stdio.h problems still...... Message-ID: <338B908B.5D0B@lx.net> I don't know if this is a bug. I have installed gnuwin32 correct I'm not new to this ;) but I am kinda new to unix. I don't consider myself a Unix guru for anything but I am learning. I worked on all the help I could get. I converted now my setup is text=binary not text!=binary I cleaned my registery boy did it need it. I also went converting files using these to commands to remove \r's from my files... :$ cat filename|tr -d \r >newfile;mv newfile filename or to convert a whole directory you can do :$ for file in *;do cat $file|tr -d \r>tmpname;mv tmpname;$file The first one works to remove 1 file at a time but the second don't... it just goes to > in bash an stay there... Then I ran another program I don't know if this is a bug in stdio.h but this file is missing :( its included in stdio.h but it is not in sys\ /* * defines __sFILE, _fpos_t. * They must be defined thee because stuct _eent needs them (and we don't * want eent.h to include this file. */ #include any help would be appreciated 1. how to convert my whole file system instead of 1 file at a time? and 2. Is this file missing on everyone's system... hehe did I accidently deleat it... -- Jason L. Esman aka _Jman System Admin. Network Consultant Irc Admin of irc.lx.net(newnet) Co. Owner of Den Internet Services http://www.lx.net && http://www.deninc.com PGP Keys finger jman@lx.net - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From noer@cygnus.com Tue May 27 14:55:00 1997 From: noer@cygnus.com (Geoffrey Noer) Date: Tue, 27 May 1997 14:55:00 -0000 Subject: Beta 18 breaks gettimeofday References: <3.0.32.19970526122602.027fd614@silence.secnet.com> Message-ID: <199705271948.MAA23262@cirdan.cygnus.com> Jonathan Wilkins wrote: > > in regards to this problem.. it was simply a problem with the returned > result from gettimeofday().. all other *NIX's seem to return 0 on success, > and the cygwin returns non-0 on success.. perhaps this should be changed.. Thanks. The code now reads instead of always returning 1: int gettimeofday (struct timeval *p, struct timezone *z) { /* FIXME: should return zero on success, -1 on failure */ SYSTEMTIME t; FILETIME f; TIME_ZONE_INFORMATION tz; GetTimeZoneInformation (&tz); GetSystemTime (&t); /* FIXME: ignores tz */ SystemTimeToFileTime (&t, &f); if (p) { totimeval (p, &f, tz.Bias * 60); } if (z) { z->tz_minuteswest = tz.Bias; z->tz_dsttime = tz.StandardBias != tz.Bias; } return 0; } More return code setting/checking needs to be added to this and other functions in winsup/times.cc (and other files in winsup). Hopefully returning zero will be a bit better than what was there previously though... -- Geoffrey Noer noer@cygnus.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From dmg@arteque.com Tue May 27 14:55:00 1997 From: dmg@arteque.com (David Goodine) Date: Tue, 27 May 1997 14:55:00 -0000 Subject: Problem running "g++ --help" Message-ID: <338B2777.60FD@arteque.com> This might be a reasonable error for this version of g++, but it looks like I might noth have things installed right. I'm quite sure I installed the B18 package exactly according to the instructions, and it works fine otherwise when I use it to compile files. When I try "g++ --help" I get the following message: C:\gnuwin32\b18\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32\cygnus-2.7.2-970404\../../../../i386-cygwin32\lib/libcygwin.a(libcmain.o)(.text+0x1e):libcmain.cc: undefined reference to `WinMain@16' g++: Internal compiler error: program ld got fatal signal 1 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From mka@redes.int.com.mx Tue May 27 15:48:00 1997 From: mka@redes.int.com.mx (Michael Anderson) Date: Tue, 27 May 1997 15:48:00 -0000 Subject: tty programming for modems and fax References: <01BC69E4.10E4EEB0@gater.krystalbank.msk.ru> Message-ID: <338B5CC5.5AED293D@redes.int.com.mx> I installed B18 and discovered that most of the tty (serial port) stuff doesn't seem to work. I therefore downloaded the B18 source and discovered that the features were not implemented! Do I have the wrong source or is no one else trying to do tty programming with cygwin? In any case, I've spent the last coupla' days implementing what I needfor my projects. In particular, I have the following functionality coded and awaiting through testing: open() supports O_NDELAY and O_NONBLOCK in the 2nd argument. I also added the third argument, mode, for future implementation. fcntl() supports toggling of blocking and non-blocking I/O modes with O_NDELAY and O_NONBLOCK. tcgetattr() previously handled CBAUD (all speeds), CSIZE, ONLCR, and VTIME and VMIN for c_cc[]. New support is for CSTOPB, PARENB, PARODD, INPCK, IGNPAR, IXON, IXOFF, CRTSXOFF, CRTSCTS, and VSTART and VSTOP for c_cc[]. tcsetattr() previously handled CBAUD, CSIZE, PARENB, PARODD, IGNCR, ONLCR, and VTIME and VMIN for c_cc[]. New support is for CSTOPB, INPCK, IGNPAR, IXON, IXOFF, CRTSXOFF, CRTSCTS, and VSTART and VSTOP for c_cc[]. TCSANOW, TCSADRAIN, and TCSAFLUSH arguments are implemented. tcsendbreak() is implemented. tcdrain() is implemented. tcflow() is implemented. I'm still working on CREAD and CLOCAL for tc[gs]etattr(). Many other flags such as HUPCL may have no reasonable implementation. If anybody else cares about this stuff and has code which could be used to test it, that would be great. I'm only using a subset for my current project and as usual, am way behind, and don't have time to do testing or much more development. Regards, Mike Anderson mka@redes.int.com.mx - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From dmg@arteque.com Tue May 27 15:48:00 1997 From: dmg@arteque.com (David Goodine) Date: Tue, 27 May 1997 15:48:00 -0000 Subject: Simple questions about windows programs Message-ID: <338B26B1.2F2A@arteque.com> I've downloaded and intalled B18 of the Gnu-Win32 developers kit and have gotten simple programs running via bash. When I try to compile windows programs, however, it seems that g++ is not linking against the correct windows libraries, and I can't find ANY examples or documnetation that says what libraries I need to link against. If someone can point me to some specific documentation I would greatly appreciate it. Also, I'm especially interested in sample programs. (BTW, is Cygnus might want to create a respository of example programs for use with GNU-WIN32 toolkit.) Another question, given that I have a random DLL from someone else (w/o source code), how do I link it into a programing using g++? Thanks. David Goodine Arteque, Inc. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From giac@dalim.de Tue May 27 15:48:00 1997 From: giac@dalim.de (Philippe GIACINTI) Date: Tue, 27 May 1997 15:48:00 -0000 Subject: problem with make in b18 Message-ID: <9705271652.ZM6570@pion.kehl.dalim.de> I have a problem with the .EXPORT_ALL_VARIABLES rule in make of b18. In fact my variables aren't exported on Windows NT, but they are on Unix. If I remove the rule .EXPORT_ALL_VARIABLES and then export explicitly a variable, it works on Unix but not on Windows NT. Can someone help me ? -- ========================================================================== | Philippe GIACINTI | | | | DALiM GmbH R&D email: giac@dalim.de | | Daimler Strasse 2, tel: +49.7851.9196-28 | | D-77694 Kehl-Sundheim Germany fax: +49.7851.73576 | | | ========================================================================== - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From mh@mike.franken.de Tue May 27 15:48:00 1997 From: mh@mike.franken.de (Michael Hirmke) Date: Tue, 27 May 1997 15:48:00 -0000 Subject: Spanish keyboard under Win95 References: Message-ID: <6XiEt7I$pfB@mike.franken.de> Hi Pelos, > > I have installed b18 under my Winslow95 with a Spanish keyboard. I > have compiled some programs without problems, but when I try to use bash > I can't type '\' '|' and some other control characters. Is it a problem > of W95 or a problem of cygnus.dll? Suggestions are welcome... Its a problem with cygwin32.dll - wait a few days, there will be a patch available by Sergey Okhapkin. > > > (my english is pathetic :( Don't worry about that :) Bye. Michael. -- Ich bin vom 1. Juni bis 3. Juli in Urlaub, Ihr muesst Euch also leider bis dahin mit einer Antwort auf Euere Postings und Mails gedulden :)) -- I'm on vacation from June, 1st to Juli, 3rd - so please be patient to get an answer to your postings and mails until I'll be back - thx :)) -- Michael Hirmke | Telefon +49 (911) 557999 Georg-Strobel-Strasse 81 | FAX +49 (911) 557664 90489 Nuernberg | E-Mail mh@mike.franken.de | WWW http://minimike.franken.de/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Alec@tca.co.uk Tue May 27 20:13:00 1997 From: Alec@tca.co.uk (Alec Clews) Date: Tue, 27 May 1997 20:13:00 -0000 Subject: tar problems (b18) Message-ID: I'm getting problems running tar as in : bash$ tar -cvf //a * tar: Cannot open //a: Permission denied User id's etc. are bit meaningless on winge95 so I'm not sure what to do next. If this is an RTFM please point me in the right direction. (What I'm trying to do is produce a multi volume floppy archive to read on my Linux box so other approaches are welcome). Thanks Regards, Alec -- Alec Clews, , PGP keyid:48FA EB81 TCA Consulting Ltd 44-(0)171-415-8159 Fax:44-(0)171-556-0022 New CIty Court, 20 St Thomas Street, London, Britain, SE1 9SD ====================== Usual Disclaimers Apply ===================== [Personal and PGP key http://www.earth.demon.co.uk/alec ] - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From pat@po.cwru.edu Tue May 27 20:13:00 1997 From: pat@po.cwru.edu (Paul A. Thompson) Date: Tue, 27 May 1997 20:13:00 -0000 Subject: man under NT ??? References: <97052619472618800@athena.dnn.gr> <199705270306.NAA20626@piglet.dstc.edu.au> Message-ID: <3.0.1.32.19970527122059.006ed870@pop.cwru.edu> At 01:06 PM 5/27/97 +1000, Bill Segall wrote: >"Stratos" == Stratos Margaritis <- System Admin > writes: > >Stratos> Hello to everybody, Does anyone knows if there exists a unix man >Stratos> command that runs under winNT 4.0 ?? > >I've taken to using vim on .man .1 files etc. It does a lovely job. > >Bill. >- >For help on using this list (especially unsubscribing), send a message to >"gnu-win32-request@cygnus.com" with one line of text: "help". > > If you would care to hint where this is located at, the rest of us might be interested as well. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From leisner@sdsp.mc.xerox.com Tue May 27 20:13:00 1997 From: leisner@sdsp.mc.xerox.com (Marty Leisner) Date: Tue, 27 May 1997 20:13:00 -0000 Subject: SIGALRM doesn't work? References: <01BC61F5.4E5DE770@gater.krystalbank.msk.ru> Message-ID: <9705271739.AA14892@gnu.sdsp.mc.xerox.com> > Jimen Ching wrote: > > Hi, could someone verify that the following program does not cause an > alarm signal to be generated? Does SIGALRM work at all with b18? That's > the version I have on my NT > > SIGALRM and other timer stuff are not supported in b18. > And its not supported in djgpp. Its incredibly useful. Is there a way to generate a signal in a win32 program using the native api? -- marty leisner@sdsp.mc.xerox.com The Feynman problem solving Algorithm 1) Write down the problem 2) Think real hard 3) Write down the answer Murray Gel-mann in the NY Times - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From mh@mike.franken.de Tue May 27 20:13:00 1997 From: mh@mike.franken.de (Michael Hirmke) Date: Tue, 27 May 1997 20:13:00 -0000 Subject: man under NT ??? References: <97052619472618800@athena.dnn.gr> Message-ID: <6XeCGh6KpfB@mike.franken.de> Hi Stratos, > Hello to everybody, > > Does anyone knows if there exists a unix man command that runs under winNT > 4.0 ?? I have one, but you also need working packages of groff, less and ncurses. > > > > Stratos :-) Bye. Michael. -- Michael Hirmke | Telefon +49 (911) 557999 Georg-Strobel-Strasse 81 | FAX +49 (911) 557664 90489 Nuernberg | E-Mail mh@mike.franken.de | WWW http://minimike.franken.de/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From kunglao@prairienet.org Tue May 27 20:13:00 1997 From: kunglao@prairienet.org (kunglao@prairienet.org) Date: Tue, 27 May 1997 20:13:00 -0000 Subject: Oops, that df command again Message-ID: <199705272028.PAA23713@bluestem.prairienet.org> Looks like my earlier message about this didn't attach the df.c file like I had hoped. So, here it is again. Enjoy --- "Oh, what sad times are these when passing ruffians can say 'Ni!' at will to old ladies. There is a pestulance upon this land, nothing is sacred. Even those who arrange and design shrubberies are under considerable economic stress of this period of history." -- Roger the Shrubber The following section of this message contains a file attachment prepared for transmission using the Internet MIME message format. If you are using Pegasus Mail, or any another MIME-compliant system, you should be able to save it or view it from within your mailer. If you cannot, please ask your system administrator for assistance. ---- File information ----------- File: df.c Date: 23 May 1997, 21:54 Size: 8585 bytes. Type: Program-source -------------- next part -------------- A non-text attachment was scrubbed... Name: df.c Type: text/x-c Size: 8585 bytes Desc: not available URL: From jpmorgan@unm.edu Tue May 27 20:13:00 1997 From: jpmorgan@unm.edu (James Paul Morgan) Date: Tue, 27 May 1997 20:13:00 -0000 Subject: Problem running "g++ --help" References: <338B2777.60FD@arteque.com> Message-ID: On Tue, 27 May 1997, David Goodine wrote: > When I try "g++ --help" I get the following message: > > C:\gnuwin32\b18\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32\cygnus-2.7.2-970404\../../../../i386-cygwin32\lib/libcygwin.a(libcmain.o)(.text+0x1e):libcmain.cc: > undefined reference to `WinMain@16' > g++: Internal compiler error: program ld got fatal signal 1 > Under 17.1, I get this error with this rather amusing path: bash$ g++ --help /usr/H-i386-cygwin32/lib/gcc-lib/i386-cygwin32/cygnus-2.7.2-961023/.. /../../../i386-cygwin32/lib/libcygwin.a(libcmain.o): In function `main': /pizza/mushroom/noer/beta17/src/winsup/libcmain.cc:30: undefined reference to `WinMain@16' See ya around the Mulberry bush. --James :) - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From mh@mike.franken.de Tue May 27 21:19:00 1997 From: mh@mike.franken.de (Michael Hirmke) Date: Tue, 27 May 1997 21:19:00 -0000 Subject: My patches References: <01BC69E9.56340F00@gater.krystalbank.msk.ru> Message-ID: <6XeCGoN4pfB@mike.franken.de> Hi Sergey, > Sergey Okhapkin wrote: > > > Here is the changes I've made in winsup code after B18 release. > > > > Precompiled cygwin.dll and libcygwin.a (cygwin-sos.052697.tar.gz) and X11 > DLLs (X11-dll.tar.gz) uploaded to ftp://dimcom.uqac.uquebec.ca/incoming . Sorry, this server does not allow anonymous login 8-// > > -- > Sergey Okhapkin Bye. Michael. -- Michael Hirmke | Telefon +49 (911) 557999 Georg-Strobel-Strasse 81 | FAX +49 (911) 557664 90489 Nuernberg | E-Mail mh@mike.franken.de | WWW http://minimike.franken.de/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From dbe@wgn.net Tue May 27 21:19:00 1997 From: dbe@wgn.net ($Bill Luebkert) Date: Tue, 27 May 1997 21:19:00 -0000 Subject: Missing files stdio.h problems still...... References: <338B908B.5D0B@lx.net> Message-ID: <338BB214.566A@wgn.net> jman wrote: > > I don't know if this is a bug. I have installed gnuwin32 correct I'm not > new to this ;) but I am kinda new to unix. I don't consider myself a > Unix guru for anything but I am learning. I worked on all the help I > could get. I converted now my setup is text=binary not text!=binary I > cleaned my registery boy did it need it. I also went converting files > using these to commands to remove \r's from my files... > :$ cat filename|tr -d \r >newfile;mv newfile filename > > or to convert a whole directory you can do > > :$ for file in *;do cat $file|tr -d \r>tmpname;mv tmpname;$file I think like this is good: for file in *; do cat $file | tr -d \r >tmpname; mv tmpname $file; done > The first one works to remove 1 file at a time but the second don't... > it just goes to > > > in bash an stay there... > > Then I ran another program I don't know if this is a bug in stdio.h but > this file is missing :( > its included in stdio.h but it is not in sys\ > > /* > * defines __sFILE, _fpos_t. ^ reent.h > * They must be defined thee because stuct _eent needs them (and we > don't > * want eent.h to include this file. > */ > > #include ^ reent.h > > any help would be appreciated 1. how to convert my whole file system > instead of 1 file at a time? > and 2. Is this file missing on everyone's system... hehe did I > accidently deleat it... Looks like your /r got changed to / in stdio.h. A little mis-type maybe? :) You shouldn't need to convert the distribution files, just unpack them properly. I used Winzip with no problems there. As far as existing files you have a script above that should work. You could precede it with a grep so only the bad files would be modified. A Perl script would probably be better. Walk the tree finding all files with ^M in them and make a dir array, then run a tr on just those files. -- ,-/- __ _ _ $Bill Luebkert (_/ / ) // // DBE Collectibles / ) /--< o // // http://www.wgn.net/~dbe/ -/-' /___/_<_ Message-ID: <199705280531.WAA19450@rtl.cygnus.com> > Is there one for gnu-win32? If not, is there one in progress? No. Not that I know of. -- Geoffrey Noer noer@cygnus.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From bfishman@es.com Tue May 27 22:31:00 1997 From: bfishman@es.com (Barry Fishman) Date: Tue, 27 May 1997 22:31:00 -0000 Subject: Beta 18 breaks gettimeofday References: <3.0.32.19970526122602.027fd614@silence.secnet.com> Message-ID: >From: Jonathan Wilkins > in regards to this problem.. it was simply a problem with the returned > result from gettimeofday().. all other *NIX's seem to return 0 on success, > and the cygwin returns non-0 on success.. perhaps this should be changed.. According to the SunOS 5.5 man page: NAME gettimeofday, settimeofday - get or set the date and time ... RETURN VALUES A -1 return value indicates that an error occurred and errno has been set. Ultrix (4.3) and IRIX (5.3) man pages state that -1 indicates error and 0 indicates success. The portable thing to do seems to be checking the return value against -1. This will work with the cygnus implementation. if (gettimeofday(tvp, NULL) == -1) { printf("failed\n error was : %s\n", strerror(errno)); } ... Barry - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From stratos@dnn.gr Tue May 27 23:20:00 1997 From: stratos@dnn.gr (Stratos Margaritis - System Admin) Date: Tue, 27 May 1997 23:20:00 -0000 Subject: man under NT ??? Message-ID: <97052809190933800@athena.dnn.gr> > At 01:06 PM 5/27/97 +1000, Bill Segall wrote: > >"Stratos" == Stratos Margaritis <- System Admin > writes: > > > >Stratos> Hello to everybody, Does anyone knows if there exists a unix man > >Stratos> command that runs under winNT 4.0 ?? > > > >I've taken to using vim on .man .1 files etc. It does a lovely job. > > > >Bill. > >- > >For help on using this list (especially unsubscribing), send a message to > >"gnu-win32-request@cygnus.com" with one line of text: "help". > > > > > > If you would care to hint where this is located at, the rest of us might > be interested as well. > > > You will find out that vim does read correctly only .man files something that both less and emacs can do. But if you still want to get vim try our site at: ftp://ftp.dnn.gr/pub/more/vim46w32.zip (Command line VIM) ftp://ftp.dnn.gr/pub/more/wvim32.zip (Windows application) or you can try at the original site at: http://www.math.fu-berlin.de/~guckes/vim/ Stratos :-) System Administrator & Software Engineer at DNNet ---------------------------------------------------------------------------- Name -> Stratos Margaritis e-mail -> stratos@dnn.gr, webmaster@dnn.gr, hostmaster@dnn.gr, postmaster@dnn.gr WWW -> http://www.dnn.gr/~stratos/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From ccurtin@trintech.com Wed May 28 02:09:00 1997 From: ccurtin@trintech.com (Colman Curtin) Date: Wed, 28 May 1997 02:09:00 -0000 Subject: My patches Message-ID: <199705280910.KAA31892@muzak.trintech.ie> >> >> Precompiled cygwin.dll and libcygwin.a (cygwin-sos.052697.tar.gz) and X11 >> DLLs (X11-dll.tar.gz) uploaded to ftp://dimcom.uqac.uquebec.ca/incoming . >Sorry, this server does not allow anonymous login 8-// You can only upload to /incoming when the admin moves the files to /pub/gnuwin32/bin/ download it from there. Coley. ____________________________________________________________ Colman Curtin mailto:ccurtin@trintech.ie Software Engineer ------------------------------------------------------------ Trintech (Manufacturing) Ltd, http://www.trintech.com/ South County Business Park, Leopardstown, Dublin 18. Tel +353-1-2956766 Fax +353-1-2954735 ------------------------------------------------------------ He who Laughs, Lasts. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From allan@interport.net Wed May 28 02:35:00 1997 From: allan@interport.net (Allan Peda) Date: Wed, 28 May 1997 02:35:00 -0000 Subject: Where to put .bash (and .ksh)? Message-ID: <338D4E22.4DA3@interport.net> I defined the variable $HOME as : bash$ echo $home bash$ echo $HOME C:\WINNT\PROFILES\ADMINISTRATOR bash$ and I put .emacs there. Why won't It be read? Thanks Allan - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From RPraetorius@AspenRes.Com Wed May 28 02:35:00 1997 From: RPraetorius@AspenRes.Com (Robert Praetorius) Date: Wed, 28 May 1997 02:35:00 -0000 Subject: CMD, case-insensitive file-name completion References: <000670F0.1893@po.cle.ab.com> Message-ID: CMD also has case-insensitive filename completion. I believe it's undocumented, although this may have changed. . . Here's what you need in your registry (under HKEY_CURRENT_USER) to turn it on: Key Name: Software\Microsoft\Command Processor Name: CompletionChar Type: REG_SZ Data: 9 (grepping the web for CompletionChar may turn up more info). -------p--a--s--s--i--o--n-----n--e--e--d--s-----a-----f--a--c--e------- "oncology recapitulates philately" --Mark Maxson Robert M. Praetorius "balance, not symmetry" --Mark Stanley work: RPraetorius@AspenRes.Com (attribution by Stigler) fun & recreation: rmp@PopJ1.MA.UltraNet.Com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From stratos@dnn.gr Wed May 28 02:35:00 1997 From: stratos@dnn.gr (Stratos Margaritis - System Admin) Date: Wed, 28 May 1997 02:35:00 -0000 Subject: man under NT ??? Message-ID: <97052809074833600@athena.dnn.gr> > > Does anyone knows if there exists a unix man command that runs under winNT > > 4.0 ?? > > I use the following script for 'man': > ---------- cut here ------------------- > #!/bin/sh > typeset -x PATH="$PATH:/usr/local/bin/groff" > MANDIR="/usr/local/man /usr/man /d/Apps/gnuwin32/b18/man" > for md in $MANDIR; do > files=`echo $md/man*/$*.*` > if [ -e "$files" ] ; then > nroff -man $files | less 2>/dev/null > else > echo $md: $*: not found > fi > done > ---------- cut here ------------------- > Richard Schulte Idea Development Incorporated > http://ideadev.com/Richard Thank you Richard for that script but my problem is that I don't have groff or nroff working for NT4.0. I've compiled without problems groff-1.10 but I could not install it under NT4.0 and I also had problems when running it when the application looks for devices since gnu-win32 b18 didn't create any devices when I was installing it. Do you know where I can find a precompiled version of groff ??? Thank you in advanced.. Stratos :-) System Administrator & Software Engineer at DNNet ---------------------------------------------------------------------------- Name -> Stratos Margaritis e-mail -> stratos@dnn.gr, webmaster@dnn.gr, hostmaster@dnn.gr, postmaster@dnn.gr WWW -> http://www.dnn.gr/~stratos/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Scott.Mintz@po.cle.ab.com Wed May 28 02:35:00 1997 From: Scott.Mintz@po.cle.ab.com (Scott Mintz) Date: Wed, 28 May 1997 02:35:00 -0000 Subject: CMD, case-insensitive file-name completion Message-ID: <0006A9FE.1893@po.cle.ab.com> That's true. The data value below is the character. You could theoretically change it to something else. The actual key is: HKEY_CURRENT_USER\Software\Microsoft\Command Processor Note the space between Command and Processor is necessary. -Scott A. Mintz ______________________________ Reply Separator _________________________________ Subject: Re: CMD, case-insensitive file-name completion Author: RPraetorius@AspenRes.Com at Internet Date: 5/27/97 12:08 PM CMD also has case-insensitive filename completion. I believe it's undocumented, although this may have changed. . . Here's what you need in your registry (under HKEY_CURRENT_USER) to turn it on: Key Name: Software\Microsoft\Command Processor Name: CompletionChar Type: REG_SZ Data: 9 (grepping the web for CompletionChar may turn up more info). -------p--a--s--s--i--o--n-----n--e--e--d--s-----a-----f--a--c--e------- "oncology recapitulates philately" --Mark Maxson Robert M. Praetorius "balance, not symmetry" --Mark Stanley work: RPraetorius@AspenRes.Com (attribution by Stigler) fun & recreation: rmp@PopJ1.MA.UltraNet.Com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From promero@eecs.wsu.edu Wed May 28 02:35:00 1997 From: promero@eecs.wsu.edu (Pedro Romero - EECS) Date: Wed, 28 May 1997 02:35:00 -0000 Subject: Spanish keyboard under Win95 References: Message-ID: <199705271653.JAA29594@zmaj.eecs.wsu.edu> > > I have installed b18 under my Winslow95 with a Spanish keyboard. I > have compiled some programs without problems, but when I try to use bash > I can't type '\' '|' and some other control characters. Is it a problem > of W95 or a problem of cygnus.dll? Suggestions are welcome... > > Miguel, I too use a spanish keyboard in Win95. I can't type some characters either, but that problem is related to the keyboard map. I have installed both keyboard maps (US English, which I use most of the time, and "latin american", which I use when writing in Spanish). In Win95 you just press "ctrl+shift" to toggle between both keyboards, this allows you to use the most convenient for each task (of course, I use the english keyboard when working in bash). To rapidly check which keyboard you are using, look at the tray on the taskbar: You will see "En" when using the english keyboard and "Es" when using a spanish one. Another neat Win95 feature: the keyboard mapping is program-dependent, so, you can have, say, both Word and Corel Draw open at the same time and set the spanish keyboard for Word and the english one for Corel Draw (or your bash console, for that matter). I hope this helps you. Saludos ********************************************* Pedro R. Romero promero@eecs.wsu.edu Electrical Engineering & Computer Science Washington State University ********************************************* - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Wed May 28 02:35:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Wed, 28 May 1997 02:35:00 -0000 Subject: SIGALRM doesn't work? Message-ID: <01BC6B6A.0784CB60@gater.krystalbank.msk.ru> Marty Leisner wrote: > > SIGALRM and other timer stuff are not supported in b18. > > > > And its not supported in djgpp. > > Its incredibly useful. Is there a way to generate a signal > in a win32 program using the native api? > My patch I post on monday, fixes this problem. SIGALRM works now. -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From *jeffdb@netzone.nospam.com Wed May 28 02:35:00 1997 From: *jeffdb@netzone.nospam.com (Mikey) Date: Wed, 28 May 1997 02:35:00 -0000 Subject: make-3.75 export bug under gnu-win32/win95 Message-ID: <338c1e0b.35997859@smtp.netzone.com> make 3.75 apparently has a problem with exported variables, invoking make -n with the following makefile causes actual compilation in the subdirectories. I had to add the following line to get the expected behavior MAKE += $(MFLAGS) So apparently MAKEFLAGS is not being interpreted correctly by sub-makes, also placing "export" in a makefile also dosen't seem to have any effect. I have seen this problem reported , or at least discussed once before, but never saw an answer as to whether this is a win95, gnu-win32, or make-3.75 issue, anyone know? does make-3.75 do this under NT? Or am I reading the make.info wrong? ################################################################## UNMAKE = Makefile makefile makefile.cl readme.txt cabview dropext fileview \ mandel DIRLIST = $(filter-out $(UNMAKE),$(wildcard *)) .PHONY: all clean distclean linkclean unix all clean distclean linkclean unix: for file in $(DIRLIST) ; do \ echo making in $$file ;\ $(MAKE) -C $$file $@ || exit 1;\ done - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Wed May 28 02:35:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Wed, 28 May 1997 02:35:00 -0000 Subject: still ^M under win95 Message-ID: <01BC6B69.FFEB7390@gater.krystalbank.msk.ru> Mikey wrote: > Everything works fine, except for I am still getting extra ^M's at the > physical EOL in lines longer than 80 chars using linux termcap/info , Remove "xn" capability from linux terminal description in termcap. The name of corresponding entry in terminfo database is "xenl". Remove it too. -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Wed May 28 02:35:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Wed, 28 May 1997 02:35:00 -0000 Subject: Need help with some missing functions Message-ID: <01BC6B6A.1A33A970@gater.krystalbank.msk.ru> Mark A Gregory wrote: > Hi, > > I'm trying to compile a program and I have got it down to these missing functions: > > /empire/emp4/lib/libcommands.a(info.o)(.text+0x37e):info.c: undefined reference to `errno' > /empire/emp4/lib/libcommands.a(info.o)(.text+0x552):info.c: undefined reference to `alphasort' > /empire/emp4/lib/libcommands.a(info.o)(.text+0x56c):info.c: undefined reference to `scandir' > /empire/emp4/lib/libcommon.a(hours.o)(.text+0xaf):hours.c: undefined reference to `errno' > /empire/emp4/lib/libcommon.a(wantupd.o)(.text+0x96):wantupd.c: undefined reference to `errno' > /empire/emp4/lib/libgen.a(io.o)(.text+0xc33):io.c: undefined reference to `shutdown' > /empire/emp4/lib/libgen.a(io.o)(.text+0xc5c):io.c: undefined reference to `shutdown' > /empire/emp4/lib/liblwp.a(lwp.o)(.text+0x39b):lwp.c: undefined reference to `sigmask' > /empire/emp4/lib/liblwp.a(lwp.o)(.text+0x65e):lwp.c: undefined reference to `lwpInitContext' > > would someone please tell me if the gnu-win32 has alternatives or a library that I > need to include. Add #include to info.c, hours.c and wantupd.c, use POSIX sigprocmask() instead of BSD sigmask(). Looks like scandir() is something about POSIX opendir/readdir. Nothing about the other missed calls... -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jman@lx.net Wed May 28 03:16:00 1997 From: jman@lx.net (jman) Date: Wed, 28 May 1997 03:16:00 -0000 Subject: Missing files stdio.h problems still...... References: <338B908B.5D0B@lx.net> <338BB214.566A@wgn.net> Message-ID: <338C3E37.73D1@lx.net> howdy, Yea I see what I did. I went an was trying to change all the ^m's at once an hit the gnuwin32 dir also :( Oh well I allready installed an it works. Your idea about the perl script would be great is anyone good enough with perl to pull this off for windows. I have perl installed just only use it for some small things ;). Seeing I'm sure or at least I hope I ain't the only one with ^m's in files it would be a good thing for alot of people... $Bill Luebkert wrote: > > jman wrote: > > > > I don't know if this is a bug. I have installed gnuwin32 correct I'm not > > new to this ;) but I am kinda new to unix. I don't consider myself a > > Unix guru for anything but I am learning. I worked on all the help I > > could get. I converted now my setup is text=binary not text!=binary I > > cleaned my registery boy did it need it. I also went converting files > > using these to commands to remove \r's from my files... > > :$ cat filename|tr -d \r >newfile;mv newfile filename > > > > or to convert a whole directory you can do > > > > :$ for file in *;do cat $file|tr -d \r>tmpname;mv tmpname;$file > > I think like this is good: > > for file in *; do cat $file | tr -d \r >tmpname; mv tmpname $file; done > > > The first one works to remove 1 file at a time but the second don't... > > it just goes to > > > > > in bash an stay there... > > > > Then I ran another program I don't know if this is a bug in stdio.h but > > this file is missing :( > > its included in stdio.h but it is not in sys\ > > > > /* > > * defines __sFILE, _fpos_t. > ^ > reent.h > > * They must be defined thee because stuct _eent needs them (and we > > don't > > * want eent.h to include this file. > > */ > > > > #include > ^ > reent.h > > > > any help would be appreciated 1. how to convert my whole file system > > instead of 1 file at a time? > > and 2. Is this file missing on everyone's system... hehe did I > > accidently deleat it... > > Looks like your /r got changed to / in stdio.h. > > A little mis-type maybe? :) > > You shouldn't need to convert the distribution files, just > unpack them properly. I used Winzip with no problems there. > > As far as existing files you have a script above that should work. > You could precede it with a grep so only the bad files would be > modified. > > A Perl script would probably be better. Walk the tree finding > all files with ^M in them and make a dir array, then run a tr on > just those files. > > -- > ,-/- __ _ _ $Bill Luebkert > (_/ / ) // // DBE Collectibles > / ) /--< o // // http://www.wgn.net/~dbe/ > -/-' /___/_<_ Your stuff won't compile with binary mounts until you remove the CR's from your makefile, source, headers, etc. CR's are generally what causes strange behavior in some cases when trying to build things, with or without binary mounts. Larry At 03:53 PM 5/24/97 -0500, jman wrote: >Mikey wrote: >> >> Did you fix your mounts after switching to B18? >> Should all say text=binary, not text!=binary. >> >> (*jeffdb@netzone.com) >> the return address for this message is anti spammed >> remove * from the above address to reply. >> Mikey >> ---- >> From: jman >> To: gnu-win32@cygnus.com >> Date: Saturday, May 24, 1997 9:06 AM >> Subject: stdio.h >> >> >Howdy, >> > Has anyone had a problem with stdio.h >> >Most the programs I have compiled work but ANY the use stdio.h crash or >> >just won't compile no mater what... >> > Just wondering if anyone else has had any similar problems with it.. >> >-- >> >Jason L. Esman aka _Jman >> >System Admin. Network Consultant >> >Irc Admin of irc.lx.net(newnet) >> >Co. Owner of Den Internet Services >> > http://www.lx.net && http://www.deninc.com >> >PGP Keys finger jman@lx.net >> >- >> >For help on using this list (especially unsubscribing), send a message to >> >"gnu-win32-request@cygnus.com" with one line of text: "help". >> > >Ok I unmounted an remounted >unmounted text!=binary to text=binary just like ya said... and this >simple easy program wouldn't compile at all.. I know it don't call for >stdio.h but it didn't an wouldn't compile. > >#include > >int STDCALL >WinMain (HINSTANCE hInst, HINSTANCE hPrev, LPSTR lpCmd, int nShow) >{ > MessageBox (NULL, "Hello World!", "Test Compiler works", MB_OK); > return 0; >} > >Compiled with gcc -o test.exe test.c -luser32 -Wl,--subsystem,windows > >I remounted the old way with text!=binary and tada it compiled no >problems.. > >I guess I will just have to play around I never had this problem till I >went to b18 tho. I do see that b18 is faster tho at least it is for me.. > >-- >Jason L. Esman aka _Jman >System Admin. Network Consultant >Irc Admin of irc.lx.net(newnet) >Co. Owner of Den Internet Services > http://www.lx.net && http://www.deninc.com >PGP Keys finger jman@lx.net >- >For help on using this list (especially unsubscribing), send a message to >"gnu-win32-request@cygnus.com" with one line of text: "help". > > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From hrabri@kai.com Wed May 28 03:16:00 1997 From: hrabri@kai.com (Hrabri Rajic) Date: Wed, 28 May 1997 03:16:00 -0000 Subject: My patches References: <01BC6A94.E93F8710@gater.krystalbank.msk.ru> Message-ID: <9705271521.AA02683@kai.com> ++ ++ > > Precompiled cygwin.dll and libcygwin.a (cygwin-sos.052697.tar.gz) and X11 ++ > > DLLs (X11-dll.tar.gz) uploaded to ftp://dimcom.uqac.uquebec.ca/incoming . ++ > ++ > How do I get the files from here? I just get "User anonymous access denied." ++ > ++ ++ Let's wait for ftp site maintainer will move these files to /pub/gnu-win32/bin... Access ftp site with this address: http://wwwdim.uqac.uquebec.ca/ftp/ Hrabri - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From m.gregory@rmit.edu.au Wed May 28 05:00:00 1997 From: m.gregory@rmit.edu.au (Mark A Gregory) Date: Wed, 28 May 1997 05:00:00 -0000 Subject: Need help with some missing functions Message-ID: <01BC6BB2.E2619D10@rem18.co.rmit.edu.au> Hi Sergey, I have sorted out many problems thanks to your help. I have one more question that I hope you can answer. I have attached a file called arch.c and I need to choose between one of the many options for lwpInitContext. It seems they are architecture dependent. Which would you choose to work with the gnu-win32 tools? thank you Mark /* * arch.c * * architecture-dependent process context code * */ #ifndef AIX32 #include "lwp.h" #include "lwpint.h" #if defined(hpc) static struct lwpProc *tempcontext; struct lwpProc *initcontext=NULL; int startpoint; startcontext() { int space[10000]; int x; startpoint = (void *) &x; if (!setjmp(initcontext->context)) longjmp(tempcontext->context,1); if (!setjmp(tempcontext->context)) longjmp(LwpCurrent->context,1); lwpEntryPoint(); } void lwpInitContext(newp, sp) struct lwpProc *newp; void *sp; { struct lwpProc holder; int endpoint; if (initcontext == NULL) { initcontext = (struct lwpProc *) malloc (sizeof(struct lwpProc)); tempcontext = &holder; if (!setjmp(tempcontext->context)) startcontext(); } tempcontext = newp; endpoint = &endpoint; if (endpoint < startpoint) { if (!setjmp(LwpCurrent->context)) longjmp(initcontext->context,1); } else { LwpCurrent->size = endpoint - startpoint; LwpCurrent->sbtm = realloc(LwpCurrent->sbtm, LwpCurrent->size); memcpy(LwpCurrent->sbtm, startpoint, LwpCurrent->size); if (!setjmp(LwpCurrent->context)) longjmp(initcontext->context,1); memcpy(startpoint, LwpCurrent->sbtm, LwpCurrent->size); } } #elif defined(hpux) void lwpInitContext(newp, sp) volatile struct lwpProc *volatile newp; void *sp; { static jmp_buf *cpp; extern struct lwpProc *LwpCurrent; if (!lwpSave(LwpCurrent->context)) { cpp = (jmp_buf *)&newp->context; asm volatile ("ldw %0, %%sp": : "o" (sp)); if (!lwpSave(*cpp)) lwpRestore(LwpCurrent->context); lwpEntryPoint(); } } int lwpSave(jb) jmp_buf jb; { /* save stack pointer and return program counter */ asm ("stw %sp, 4(%arg0)"); asm ("stw %rp, 8(%arg0)"); /* save "callee save" registers */ asm ("stw %r3, 12(%arg0)"); asm ("stw %r4, 16(%arg0)"); asm ("stw %r5, 20(%arg0)"); asm ("stw %r6, 24(%arg0)"); asm ("stw %r7, 28(%arg0)"); asm ("stw %r8, 32(%arg0)"); asm ("stw %r9, 36(%arg0)"); asm ("stw %r10, 40(%arg0)"); asm ("stw %r11, 44(%arg0)"); asm ("stw %r12, 48(%arg0)"); asm ("stw %r13, 52(%arg0)"); asm ("stw %r14, 56(%arg0)"); asm ("stw %r15, 60(%arg0)"); asm ("stw %r16, 64(%arg0)"); asm ("stw %r17, 68(%arg0)"); asm ("stw %r18, 72(%arg0)"); /* save "callee save" space register */ asm volatile ("mfsp %sr3, %r1"); asm ("stw %r1, 0(%arg0)"); /* indicate "true return" from saved() */ asm ("ldi 0, %ret0"); asm (".LABEL _comefrom_"); } void lwpRestore(jb) jmp_buf jb; { /* restore stack pointer and program counter */ asm volatile ("ldw 4(%arg0), %sp"); asm volatile ("ldw 8(%arg0), %rp"); /* restore "callee save" space register */ asm volatile ("ldw 0(%arg0), %r1"); asm volatile ("mtsp %r1, %sr3"); /* restore "callee save" registers */ asm volatile ("ldw 12(%arg0), %r3"); asm volatile ("ldw 16(%arg0), %r4"); asm volatile ("ldw 20(%arg0), %r5"); asm volatile ("ldw 24(%arg0), %r6"); asm volatile ("ldw 28(%arg0), %r7"); asm volatile ("ldw 32(%arg0), %r8"); asm volatile ("ldw 36(%arg0), %r9"); asm volatile ("ldw 40(%arg0), %r10"); asm volatile ("ldw 44(%arg0), %r11"); asm volatile ("ldw 48(%arg0), %r12"); asm volatile ("ldw 52(%arg0), %r13"); asm volatile ("ldw 56(%arg0), %r14"); asm volatile ("ldw 60(%arg0), %r15"); asm volatile ("ldw 64(%arg0), %r16"); asm volatile ("ldw 68(%arg0), %r17"); asm volatile ("ldw 72(%arg0), %r18"); /* warp to saved() to unwind the frame correctly */ asm volatile ("bl _comefrom_, %r0"); asm volatile ("ldi 1, %ret0"); } #elif defined(BSD386) void lwpInitContext(newp, sp) struct lwpProc *newp; void *sp; { newp->context[2] = (int)sp; newp->context[0] = (int)lwpEntryPoint; } #elif defined(FBSD) void lwpInitContext(newp, sp) struct lwpProc *newp; void *sp; { setjmp (newp->context); newp->context->_jb[2] = (int)sp; newp->context->_jb[3] = (int)sp; newp->context->_jb[0] = (int)lwpEntryPoint; } #elif defined(__linux__) void lwpInitContext(newp, sp) struct lwpProc *newp; void *sp; { newp->context->__sp = sp; newp->context->__bp = sp; newp->context->__pc = (void *)lwpEntryPoint; } #elif defined(SUN3) void lwpInitContext(newp, sp) struct lwpProc *newp; void *sp; { newp->context[2] = (int)sp; newp->context[3] = (int)lwpEntryPoint; } #elif defined(SUN4) void lwpInitContext(newp, sp) struct lwpProc *newp; void *sp; { static jmp_buf *cpp; extern struct lwpProc *LwpCurrent; bzero(newp->context, sizeof(newp->context)); newp->context[0] = (int)sp; /* preserve cpp for new context */ cpp = (jmp_buf *)&newp->context; if (!_setjmp(LwpCurrent->context)) { /* create new context */ /* flush registers */ asm ("ta 0x03"); /* %o0 <- newp */ asm ("ld [%fp+0x44], %o0"); /* %o1 <- newp->context[0] */ asm ("ld [%o0], %o1"); /* create min frame on new stack */ asm ("save %o1,-96, %sp"); if (!_setjmp(*cpp)) _longjmp(LwpCurrent->context, 1); lwpEntryPoint(); } } #elif defined(__USLC__) && defined(i386) /* USL/Unixware on an Intel 386/486/... processor. * Tested on Unixware v1.1.2, based on SYSV R4.2 */ /* As per normal empire documentation, there is none. * * But, what we are attempting to do here is set up a longjump * context buffer so that the lwpEntryPoint is called when * the thread starts. * * I.E., what a setjmp/longjmp call set would do. * * How to figure this out? Well, without the setjmp code, you * need to reverse engineer it by printing out the context buffer * and the processor registers, and mapping which ones need * to be set. * * Alternatively, you can single instruction step through the longjmp * function, and figure out the offsets that it uses. * * Using offsets in bytes, * context + 0x04 [1] -> esi (general purpose reg) * context + 0x08 [2] -> edi (general purpose reg) * context + 0x0C [3] -> ebp (general purpose or parameter passing) * context + 0x10 [4] -> esp (stack) * context + 0x14 [5] -> jump location for return */ void lwpInitContext(newp, sp) struct lwpProc *newp; void *sp; { newp->context[4] = (int)sp; newp->context[5] = (int)lwpEntryPoint; } #elif defined UCONTEXT /* * Alternate aproach using setcontext en getcontext in stead of setjmp and * longjump. This should work on any SVr4 machine independant of * architecture. Unfortunaltely some changes are still nessesary in lwp.c. * Tested on IRIX 5.3 */ void lwpInitContext(newp, spp) struct lwpProc *newp; stack_t *spp; { getcontext (&(newp->context)); newp->context.uc_stack.ss_sp = spp->ss_sp; newp->context.uc_stack.ss_size = spp->ss_size; makecontext (&(newp->context), lwpEntryPoint, 0); } #elif defined(ALPHA) #include void lwpInitContext(newp, sp) struct lwpProc *newp; void *sp; { extern long *_gp; /* register values obtained from setjmp.h */ _setjmp(newp->context); newp->context[2] = (long)lwpEntryPoint; /* program counter */ newp->context[30] = (long)lwpEntryPoint; /* return address */ newp->context[31] = (long)lwpEntryPoint; /* fake program value (!) */ newp->context[34] = (long)sp; /* stack pointer */ } int lwpSave(jb) jmp_buf jb; { return _setjmp(jb); } void lwpRestore(jb) jmp_buf jb; { /* resume, but get the pv from the jmp_buf */ asm("ldq %pv, 248(%a0)"); asm("stq %a0, 16(%sp)"); /* generates a warning, but functions just fine */ asm("bsr %ra, __longjump_resume"); } #endif #endif From sos@prospect.com.ru Wed May 28 10:40:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Wed, 28 May 1997 10:40:00 -0000 Subject: Where to put .bash (and .ksh)? Message-ID: <01BC6BAF.F9221320@sos> Allan Peda wrote: > I defined the variable $HOME as : > > bash$ echo $home > > bash$ echo $HOME > C:\WINNT\PROFILES\ADMINISTRATOR > bash$ > > and I put .emacs there. Why won't It be read? Where did you define this variable? You should do it in environment settings on NT, or in autoexec.bat on win95. -- Sergey Okhapkin Moscow, Russia Looking for a job - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From karmic@hotmail.com Wed May 28 10:40:00 1997 From: karmic@hotmail.com (Mr. Karmic) Date: Wed, 28 May 1997 10:40:00 -0000 Subject: Passing $home to other prog's Message-ID: <199705281625.JAA09118@f36.hotmail.com> i'm trying to run the win95 port of ncftp under bash, but it tells me it cant find the $HOME env. i know $HOME is set, and the ncftp dir is in my paths, but it still wont run under bash. is there a way to force bash to pass on the $HOME env ? karmic@hotmail.com --------------------------------------------------------- Get Your *Web-Based* Free Email at http://www.hotmail.com --------------------------------------------------------- - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From ggp@informix.com Wed May 28 10:40:00 1997 From: ggp@informix.com (Guy Gascoigne - Piggford) Date: Wed, 28 May 1997 10:40:00 -0000 Subject: man under NT ??? Message-ID: <3.0.32.19970528072914.0095f7b0@pop.pdx.informix.com> At 07:47 PM 5/26/97 +0300, you wrote: >Hello to everybody, > >Does anyone knows if there exists a unix man command that runs under winNT >4.0 ?? This may or may not be exactly what you are after but if you are an Emacs user there is a new man package for emacs being developed calles woman (With Out MAN), here is the original message: >Date: Mon, 19 May 1997 13:51:25 +0100 >From: "Dr Francis J. Wright" >Reply-To: F.J.Wright@qmw.ac.uk >Organization: Mathematical Sciences, Queen Mary - University of London >To: ntemacs-users@cs.washington.edu >Subject: Woman UN*X manual browser >References: <199705152207.SAA08270@I1.pilgrim.com> >Content-Length: 708 > >An updated version (0.11) is now available (via the URL below). It >understands enough ?roff to format bash.1 readably (although slowly). >For the next version (0.2) I am re-writing the code for handling font >changes to do something sensible with two-character font names, among >other things, and I hope to be able to handle user-defined strings (.ds >etc.). > >Francis > >-- > >Dr Francis J. Wright (Senior Tutor for Mathematics) >School of Mathematical Sciences | Email: F.J.Wright@QMW.ac.uk >Queen Mary & Westfield College | Telephone: +44 171-975 5453 (direct) >Mile End Road, London E1 4NS, UK | Fax: +44 181-981 9587 (dept.) > World Wide Web Home Page URL: http://www.maths.qmw.ac.uk/~fjw/ > It's not perfect but it really is quite good. Guy -- Guy Gascoigne - Piggford (ggp@informix.com) Software Engineer, Informix Software, Inc. (Portland, Oregon) (503) 225-4574, Fax (503) 221-2633 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From kemp@eease.com Wed May 28 13:30:00 1997 From: kemp@eease.com (John Kemp) Date: Wed, 28 May 1997 13:30:00 -0000 Subject: Netscape Enterprise dll Message-ID: <338C92D5.2F16@eease.com> I'm trying to write dlls that export functions for Netscape Enterprise Server, portable between unix and NT, but I'm having some trouble getting them into a valid dll - Netscape Server fails to load the dlls as it says they are not valid Windows dlls. Anyone know what that's all about? My compile is shown below :- winmain.c contains the function dllEntry as spec'd, and fix.c contains the line asm(".section .idata$3\n" ".long 0,0,0,0,0") ; #!/bin/sh ld --dll --base-file lw_buffer.base -o junk lw_buffer.o winmain.o fix.o -e _dllEntry@12 /usr/i386-cygwin32/lib/libcygwin.a dlltool --dllname lw_buffer.dll --def lw_buffer.def --base-file lw_buffer.base --output-exp lw_buffer.exp --output-lib lw_buffer.a ld lw_buffer.exp --dll -o lw_buffer.dll lw_buffer.o winmain.o fix.o $LIBPATH/libcygwin.a -e _dllEntry@12 Any thoughts? Thanks John -- ............................ John Kemp Employease Inc. http://www.employease.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From allan@interport.net Wed May 28 13:30:00 1997 From: allan@interport.net (Allan Peda) Date: Wed, 28 May 1997 13:30:00 -0000 Subject: Where to put .bash (and .ksh)? References: <01BC6BAF.F9221320@sos> Message-ID: <338DE7BE.49CE@interport.net> Sergey Okhapkin wrote: > > Allan Peda wrote: > > I defined the variable $HOME as : > > > > bash$ echo $home > > > > bash$ echo $HOME > > C:\WINNT\PROFILES\ADMINISTRATOR > > bash$ > > > > and I put .emacs there. Why won't It be read? > > Where did you define this variable? You should do it in environment settings on NT, or in autoexec.bat on win95. > I defined it under: My Computer \ Control Panel \ System \ Environment \ User variables That's how I sert up the other GNU envir. variables. Does case matter? - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From weiku@solid.ssd.ornl.gov Wed May 28 14:16:00 1997 From: weiku@solid.ssd.ornl.gov (Ku Wei) Date: Wed, 28 May 1997 14:16:00 -0000 Subject: Where to put .bash (and .ksh)? References: <338D4E22.4DA3@interport.net> Message-ID: about .bash: 1. HOME must be defined before bash run. ex: put the following line in the batch file you use to initial the cygwin32: HOME=C:\WINNT\PROFILES\ADMINISTRATOR 2. It make life easier ( to me) to redefine HOME in .bash like the following: export HOME=//C/WINNT/PROFILES/ADMINISTRATOR Hope this will help. Sincerely, Wei Ku *************************************** Department of Physics and Astronomy The University of Tennessee 1408 Circle Drive Knoxville, Tennessee 37996-1200 weiku@utkux.utcc.utk.edu --------------------------------------- Solid State Division Oak Ridge National Laboratory P.O.Box 2008 Oak Ridge, TN 37831-6032 Phone: (423) 574-5795 Fax: (423) 574-4143 weiku@solid.ssd.ornl.gov *************************************** On Thu, 29 May 1997, Allan Peda wrote: > I defined the variable $HOME as : > > bash$ echo $home > > bash$ echo $HOME > C:\WINNT\PROFILES\ADMINISTRATOR > bash$ > > and I put .emacs there. Why won't It be read? > > Thanks > Allan > - > For help on using this list (especially unsubscribing), send a message to > "gnu-win32-request@cygnus.com" with one line of text: "help". > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From johnm@artisan.com Wed May 28 14:16:00 1997 From: johnm@artisan.com (John Gerard Malecki) Date: Wed, 28 May 1997 14:16:00 -0000 Subject: Completely Windowless Bashing Message-ID: <199705281852.LAA08337@carp.vlibs.com> I have Java program that wants to exec a long bash script like "sh -c script > /tmp/msg 2>&1" and I would prefer if the execution of the shell script proceeded without any window creation. I rebuilt bash using '-Wl,--subsystem,windows' and now the top-level process (bash) does not create a window but each subprocess does. (Incredibly this seems to run faster than the original bash where each subprocess updated the top-level title bar.) An obvious but unpleasant solution is to compile all the user-tools using -Wl,--subsystem,windows. Can anyone recommend a better tasting solution? Maybe there is some run-time mechanism whereby I can substitute /dev/null for the console? -thanks - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From bernd@asterix.gecko.de Wed May 28 14:31:00 1997 From: bernd@asterix.gecko.de (Bernd Prager) Date: Wed, 28 May 1997 14:31:00 -0000 Subject: missing rexec in building tar-1.12 References: <199703211009.AA43886@alijku04.edvz.uni-linz.ac.at> Message-ID: While building tar-1.12 I got the error message: rtapelib.c:249: undefined reference to `rexec' Is rexec not supported in B18 or have I to bind a separate library to do this? Bernd _____________________________________________________________________ Bernd Prager GECKO mbH; Wismarsche Str.3, 18057 Rostock; Germany http://www.gecko.de PGP Key fingerprint = 83 54 6A 3B 7A 9D 6C 0E F3 41 CE 99 11 30 B7 D6 public key by mailto:bpr@gecko.de?Subject=SendPGPKey _____________________________________________________________________ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From allan@interport.net Wed May 28 17:27:00 1997 From: allan@interport.net (Allan Peda) Date: Wed, 28 May 1997 17:27:00 -0000 Subject: Where to put .bash (and .ksh)? References: <338D4E22.4DA3@interport.net> Message-ID: <338E1F32.54F0@interport.net> Allan Peda wrote by mistake .emacs , I _meant to type .bash > > I defined the variable $HOME as : > > bash$ echo $home > > bash$ echo $HOME > C:\WINNT\PROFILES\ADMINISTRATOR > bash$ > > and I put .emacs there. Why won't It be read? ^^^^^^oops .bash > > Thanks > Allan > - > For help on using this list (especially unsubscribing), send a message to > "gnu-win32-request@cygnus.com" with one line of text: "help". - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From csoelle@sghms.ac.uk Wed May 28 18:27:00 1997 From: csoelle@sghms.ac.uk (Christian Soeller) Date: Wed, 28 May 1997 18:27:00 -0000 Subject: My patches References: <01BC6A94.E93F8710@gater.krystalbank.msk.ru> Message-ID: <50wwojs9j6.fsf@mbcsg1.sghms.ac.uk> Sergey Okhapkin writes: > > Jim Pick wrote: > > > Precompiled cygwin.dll and libcygwin.a (cygwin-sos.052697.tar.gz) and X11 > > > DLLs (X11-dll.tar.gz) uploaded to ftp://dimcom.uqac.uquebec.ca/incoming . > > > > How do I get the files from here? I just get "User anonymous access denied." > > > > Let's wait for ftp site maintainer will move these files to /pub/gnu-win32/bin... > > -- > Sergey Okhapkin I think Jim had a point here. The server doesn't seem to allow any anonymous logins at all. So we won't be able to see anything in any pub area there, whenever it is being moved there or not! Christian - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From vassilii@optimedia.co.il Wed May 28 18:27:00 1997 From: vassilii@optimedia.co.il (Vassilii Khachaturov) Date: Wed, 28 May 1997 18:27:00 -0000 Subject: bash path search bug & workarounds Message-ID: <338C99FC@gandalf.optimedia.co.il> I have both BC5 and gnu-win32 installed on my NT4. The gnu-win32 executables come before BC5 in my path. I wanted to create such an environment in which in CMD -- GREP.COM and MAKE.EXE from BC in bash -- gnu grep & make would start up when I simply type "grep" and "make". Solution for Grep (works): I moved the gnu grep.exe to ggrep.exe and created a symbolic link 'grep' to the ggrep.exe. The link was on my search path, before BC5. Thus, bash started up the gnu grep, and NT was stupid enough to ignore the symlink and proceed further along the path to the BC5's GREP.COM Solutions for Make (all didn't work!) I moved the gnu make.exe to gmake.exe and created a symbolic link 'make' to it, just the same way as with grep. Then... both bash and cmd invoked the BC5 MAKE.EXE when I typed 'make', ALTHOUGH MAKE.EXE WAS ON MY PATH *after* the link to GMAKE.EXE !!! I also tried to replace the link with a shell script called `make ' that simply passed its params to gmake. It also didn't work! (as understood, I tested in the both cases that as long as I rename the link/script from `make' to, say, `m', then 'm' does get found on the path! In my case, there are two working solutions: either remove the BC5 dir. from the path in .bashrc, or alias make=gmake. Nevertheless, it seems like that's a bash/cygwin.dll bug. Your BACbKA -- Vassilii Khachaturov vassilii@optimedia.co.il Skribu al mi per Esperanto! - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From allan@interport.net Wed May 28 18:27:00 1997 From: allan@interport.net (Allan Peda) Date: Wed, 28 May 1997 18:27:00 -0000 Subject: chmod broke? Or perhaps ' ls -l ' is inaccurate here References: <0009386C.1950@ccmail.irs.gov> <3362107E.321B940D@smva.com> Message-ID: <338E2D6C.3644@interport.net> While playing with the deleting the .exe suffix, I employed chmod 777 as illustrated to a file formerly known as input.exe, but now renamed to 'input.' It executes, but ls -l reveals: bash$ chmod -v 777 input. mode of input. changed to 0777 (rwxrwxrwx) bash$ ls -l total 186 ... .... > stuff deleted < -rw-r--r-- 1 544 everyone 111586 May 29 21:12 input .... ... So what gives here? As I mentioned, the executable executes, but 'ls -l' doesn't hint that it can. hmmm? Allan - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Keetnet@wilmington.net Wed May 28 18:47:00 1997 From: Keetnet@wilmington.net (Keet) Date: Wed, 28 May 1997 18:47:00 -0000 Subject: rename() broken?? Message-ID: <2.2.32.19970529014620.006afc98@wilmington.net> This is a restatement of what you might have read earlier, except more facts have been gathered in relation to the problem. The problem is rename() in B18 produces a 'Access Denied' message, where as in LINUX and B17.1, this problem never occured. The following segment of code is the original source, that is causing the problem: if (rename(tmpfile, dumpfile) < 0) perror(tmpfile); the file called 'tmpfile' is named './data/minimal.new.#2#' and 'dumpfile' is named './data/minimal.new' when trying to rename tmpfile to dumpfile, it produces: './data/minimal.new.#2#: Access Denied'. 'dumpfile' already exists from a previous dump, and is never removed or modified. (hence the timestamp not changing) the following section is what the code was changed to in an attempt to fix the problem, by removing 'dumpfile' first and then renaming 'tmpfile' to 'dumpfile', hopefully eliminating the problem. if ((temp = fopen(dumpfile, "r")) != NULL) { fclose(temp); (void) unlink(dumpfile); if (rename(tmpfile, dumpfile) < 0) perror(tmpfile); } else { rename(tmpfile, dumpfile); } this section of code returned the following errors: './data/minimal.new.#2#: Permission Denied' './data/minimal.new: No such file or directory' .. i'm not sure why the Permission denied comes before the access denied message.. but it does.. appearently, something isn't right here. but when it hits this loop a second time.. it freezes with a CYGWIN_EXCEPT_HANDLER error, and that's all she wrote.. consider the following, before telling me to check my sources and mounts: 1. It works fine as the oringal source under LINUX 2. It worked fine as the original source under B17.1 3. The mounts are all set the same, and are original as set by the installation 4. a /tmp dir is on each drive of my system and is mounted 5. the GCC_EXEC_PREFIX is set correctly 6. the two paths into the GnuWin32 binaries are set in my path if you can offer any help at all.. please let me know asap... thanks in advance.. - Greg Neujahr Foxbird - Keet keetnet@wilmington.net - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From johnm@artisan.com Wed May 28 19:29:00 1997 From: johnm@artisan.com (John Gerard Malecki) Date: Wed, 28 May 1997 19:29:00 -0000 Subject: Windowless Bashing Message-ID: <199705290147.SAA10616@carp.vlibs.com> I would like to have a java program exec off a bash script. When I do this the bash script likes to open it's own console window. This was annoying so I built a version of bash using the gcc -Wl,--subsystem,windows switch. Sure enough the top-level (bash) window has now disappeared but each and every sub-process brings up its own window. Can anyone tell me the "proper" things to do? Do I need to build the entire tool-set without the console window linked in? Is there a trick I can do to force the console window to select /dev/null at run time? PS - Sorry if I sent this once already today. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From grapeape@scott.net Wed May 28 20:56:00 1997 From: grapeape@scott.net (Chris Green) Date: Wed, 28 May 1997 20:56:00 -0000 Subject: Passing $home to other prog's Message-ID: <3.0.32.19970528225603.00911740@scott.net> In NT under enviroment I put in a HOME variable and set it to /home/grapeape. I believe that since ncftp doesn't check the enviroments of Bash but instead checks the win32 style variables. Hope this helps. Chris At 09:25 AM 5/28/97 PDT, Mr. Karmic wrote: >i'm trying to run the win95 port of ncftp under bash, but it tells me it cant >find the $HOME env. > >i know $HOME is set, and the ncftp dir is in my paths, but it still wont run >under bash. > >is there a way to force bash to pass on the $HOME env ? > >karmic@hotmail.com > > >--------------------------------------------------------- >Get Your *Web-Based* Free Email at http://www.hotmail.com >--------------------------------------------------------- >- >For help on using this list (especially unsubscribing), send a message to >"gnu-win32-request@cygnus.com" with one line of text: "help". > > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From dbe@wgn.net Wed May 28 22:02:00 1997 From: dbe@wgn.net ($Bill Luebkert) Date: Wed, 28 May 1997 22:02:00 -0000 Subject: bash won't accept the letter 'k', etc Message-ID: <338D0DAB.63AF@wgn.net> I've managed to get most of the stuff in b18 running (still have a lot of questions and playing to do) on Win95. When i type 'k' on the command line I get a gong on Win95 and no echo of the 'k'. All other keyboard chars are ok. Anyone know what's up? Is there some further info on setting up your environment somewhere? In particular, directories /etc and /bin (or /usr/bin), mounts, termcaps, config files in general, what gets called when, and all of that kind of stuff. Maybe a nice detailed FAQ on setup of both bash and b18 in general (mounts and such). Sorry if I missed a pointer somewhere, but I can't find much in the doc area online. Someone mentioned using vim for man pages. I use vim all the time, how do you get it to convert nroff in vim? Is there a macro or set stmt you need to issue? -- ,-/- __ _ _ $Bill Luebkert (_/ / ) // // DBE Collectibles / ) /--< o // // http://www.wgn.net/~dbe/ -/-' /___/_<_ <3362107E.321B940D@smva.com> <338E2D6C.3644@interport.net> Message-ID: <338D46CD.1941@cern.ch> Allan Peda wrote: > bash$ chmod -v 777 input. [snip] > -rw-r--r-- 1 544 everyone 111586 May 29 21:12 input Am I misunderstanding something? You show two files, `input.' and `input'. Perhaps you are just looking at the wrong file? Cheers, //lat -- Lassi.Tuura@cern.ch There's no sunrise without a night - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Thu May 29 03:15:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Thu, 29 May 1997 03:15:00 -0000 Subject: Netscape Enterprise dll Message-ID: <01BC6C3B.115625E0@gater.krystalbank.msk.ru> John Kemp wrote: > I'm trying to write dlls that export functions for Netscape Enterprise > Server, portable between unix and NT, but I'm having some trouble > getting them into a valid dll - Netscape Server fails to load the dlls > as it says they are not valid Windows dlls. Did you compile on NT or on DosShell'95? -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Thu May 29 03:15:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Thu, 29 May 1997 03:15:00 -0000 Subject: Where to put .bash (and .ksh)? Message-ID: <01BC6C3A.FB296BB0@gater.krystalbank.msk.ru> Allan Peda wrote: > I defined it under: > > My Computer > \ Control Panel > \ System > \ Environment > \ User variables > > That's how I sert up the other GNU envir. variables. Does case matter? > Yes, I think. Windows have case insensitive environment vars. My definition of HOME is in upper case and emacs finds its startup file. -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From mlof@penti.sit.fi Thu May 29 08:26:00 1997 From: mlof@penti.sit.fi (Mika Löf) Date: Thu, 29 May 1997 08:26:00 -0000 Subject: bash keys Message-ID: <338d9f18.5097559@mail.sit.fi> I cant get the keys to work in bash. I'm lacking all the keys behind the AltGr key, that is @#${[]}\. Is this a problem of readline or of termcap? Which term should I use anyway in NT? BTW, GNU-Win32 is really great! "The only stupid question is the unasked one" - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From cclim@gintic.gov.sg Thu May 29 08:26:00 1997 From: cclim@gintic.gov.sg (Lim Chu Cheow) Date: Thu, 29 May 1997 08:26:00 -0000 Subject: GNU-Win32 : gettimeofday Message-ID: <338D5CA4.40E1@gintic.gov.sg> Hi, I just installed b18, and compiled a simple program (given below). Compilation went through, but the execution received some exception and died. Am I missing something? Thanks in advance for any advice. ============ #include #include void main(void) { struct timeval tp; fprintf(stdout,"Hello world!!\n"); gettimeofday(&tp); fprintf(stdout,"Current time = %lf s\n", ((double)tp.tv_sec+tp.tv_usec*1.0e-6)); } ============== - Chu-Cheow - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From root@jacob.remcomp.fr Thu May 29 08:26:00 1997 From: root@jacob.remcomp.fr (root) Date: Thu, 29 May 1997 08:26:00 -0000 Subject: changes in lcc-win32 useful for gnu users Message-ID: I have updated lcc-win32 with wincrypt.h Header file for the cryptography API Doing this I noticed that kernel32.dll has 79 more entry points than the last version I had. I updated that, and advapi32.lib, that now has more than 140 new entry points... Probably the libraries of cygnus need an update... You can use the header files in your projects, but you will need to rebuild advapi32.a and kernel32.a (gnu libraries are called xxx.a, lcc-win32 are called xxx.lib) I would like to know if my libraries work with gnu. Has anyone tried this? I would be glad to know, and if it didn't work, what the problems are. -- Jacob Navia Logiciels/Informatique 41 rue Maurice Ravel Tel 01 48.23.51.44 93430 Villetaneuse Fax 01 48.23.95.39 France - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From john@jpsc.co.uk Thu May 29 08:26:00 1997 From: john@jpsc.co.uk (John Cooper) Date: Thu, 29 May 1997 08:26:00 -0000 Subject: Out of queue slots!! Message-ID: Every time I try this (v18) I get the message: rm -f *.o *.syx *.rfx Out of queue slots!! Out of queue slots!! Out of queue slots!! Out of queue slots!! Out of queue slots!! Out of queue slots!! Out of queue slots!! Out of queue slots!! Out of queue slots!! . . . -- John Cooper - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From condict@opengroup.org Thu May 29 09:16:00 1997 From: condict@opengroup.org (Michael Condict) Date: Thu, 29 May 1997 09:16:00 -0000 Subject: CMD, case-insensitive file-name completion References: Message-ID: <199705291423.KAA23215@postman.opengroup.org> In message < B0000119176@mx1.aspenres.com >, you write: > CMD also has case-insensitive filename completion. I'd also like to note, if we're using other Windows command processors as a precedent, that all of the following command processors have case-insensitive filename completion, and also have case-insensitive file-name pattern matching (i.e. "a*" matches both "abc" and "Alpha"): - NT command processor - 4DOS for NT - MKS UNIX Toolkit Korn Shell Moreover, all of these command processors correct the case as you type the completion character, just like my proposed changes to bash. So bash currently stands alone in not being case-insensitive. I'd say that's more surprising behavior, on a Windows system, than the fact that UNIX users expect case-sensitivity. And while were on the topic of file globbing, can someone please tell me why there are at least three slightly different copies of glob.c spread throughout the source tree? I've modified the one in the bash source tree to do case-insensitivity, but this won't help the find command, for example, and whatever other programs have built-in file- name matching. Michael Condict m.condict@opengroup.org The Open Group Research Inst. (617) 621-7349 11 Cambridge Center Cambridge, MA 02142 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From dasilva@woodmore.gsfc.nasa.gov Thu May 29 09:16:00 1997 From: dasilva@woodmore.gsfc.nasa.gov (Arlindo da Silva) Date: Thu, 29 May 1997 09:16:00 -0000 Subject: Sergey's patches Message-ID: <199705291305.JAA05961@woodmore.gsfc.nasa.gov> > > > > Let's wait for ftp site maintainer will move these files to /pub/gnu-win32/bin... > > > > -- > > Sergey Okhapkin > > I think Jim had a point here. The server doesn't seem to allow any > anonymous logins at all. So we won't be able to see anything in > any pub area there, whenever it is being moved there or not! > I was able to get the libraries this morning. Do not use ftp, go thru the URL: http://wwwdim.uqac.uquebec.ca/ftp/pub/gnuwin32/bin However, I tried replacing my win95 DLL's with Sergey's without much luck. Same error: unable to run. One may need the corresponding *.a as well, I am not sure. Geoffrey: any clue why the X11R6.3 are failing to be built on win95? Arlindo. -- *-----------------------------------------------------------------* | | | ^|^ ARLINDO DA SILVA | | | 212 Lexington Drive | | _|_ Silver Spring, MD 20901 | | @ ___|___|__ | | _|_|__________|_______/ mailto://dasilva@alum.mit.edu | | \ o o o o o o o / | | \_______________/ Home: (301) 754-1121 | | ^^^^^^^^^^^^^^^^^^^^^^^^ FAX: (301) 805-7959 | | | *-----------------------------------------------------------------* - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From allan@interport.net Thu May 29 09:16:00 1997 From: allan@interport.net (Allan Peda) Date: Thu, 29 May 1997 09:16:00 -0000 Subject: chmod broke? Or perhaps ' ls -l ' is inaccurate here References: <0009386C.1950@ccmail.irs.gov> <3362107E.321B940D@smva.com> <338E2D6C.3644@interport.net> <338D46CD.1941@cern.ch> Message-ID: <338EAF1C.53EC@interport.net> Lassi A. Tuura wrote: > > Allan Peda wrote: > > bash$ chmod -v 777 input. > [snip] > > -rw-r--r-- 1 544 everyone 111586 May 29 21:12 input > > Am I misunderstanding something? You show two files, `input.' and > `input'. Perhaps you are just looking at the wrong file? Okay... double check here... (I try both spellings below): ash$ cd /usr/work ash$ ls -l i* rw-r--r-- 1 544 everyone 111586 May 29 21:12 input rw-r--r-- 1 544 everyone 227 May 29 21:10 input.cpp rw-r--r-- 1 544 everyone 203 May 29 21:06 input.mak ash$ chmod -v 777 input. ode of input. changed to 0777 (rwxrwxrwx) ash$ ls -l i* rw-r--r-- 1 544 everyone 111586 May 29 21:12 input rw-r--r-- 1 544 everyone 227 May 29 21:10 input.cpp rw-r--r-- 1 544 everyone 203 May 29 21:06 input.mak ash$ chmod -v 777 input ode of input changed to 0777 (rwxrwxrwx) ash$ ls -l i* rw-r--r-- 1 544 everyone 111586 May 29 21:12 input rw-r--r-- 1 544 everyone 227 May 29 21:10 input.cpp rw-r--r-- 1 544 everyone 203 May 29 21:06 input.mak ash$ Hmmm, still peculiar. Thanks for the reply Allan - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From lhall@rfk.com Thu May 29 09:16:00 1997 From: lhall@rfk.com (Larry Hall) Date: Thu, 29 May 1997 09:16:00 -0000 Subject: My patches Message-ID: <2.2.32.19970529125158.00944e9c@ma.ultranet.com> At 08:04 PM 5/28/97 +0100, Christian Soeller wrote: >Sergey Okhapkin writes: > > >> >> Jim Pick wrote: >> > > Precompiled cygwin.dll and libcygwin.a (cygwin-sos.052697.tar.gz) and X11 >> > > DLLs (X11-dll.tar.gz) uploaded to ftp://dimcom.uqac.uquebec.ca/incoming . >> > >> > How do I get the files from here? I just get "User anonymous access denied." >> > >> >> Let's wait for ftp site maintainer will move these files to /pub/gnu-win32/bin... >> >> -- >> Sergey Okhapkin > >I think Jim had a point here. The server doesn't seem to allow any >anonymous logins at all. So we won't be able to see anything in >any pub area there, whenever it is being moved there or not! > > Christian I'm there now and have no problems downloading the files. Larry - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From time@ice.com Thu May 29 09:16:00 1997 From: time@ice.com (Tim Endres) Date: Thu, 29 May 1997 09:16:00 -0000 Subject: My patches Message-ID: <01BC6C2A.88151300@pc> I logged in as anonymous without any problems. ---------- From: Christian Soeller Subject: Re: My patches Sergey Okhapkin writes: > Jim Pick wrote: > > > Precompiled cygwin.dll and libcygwin.a (cygwin-sos.052697.tar.gz) and X11 > > > DLLs (X11-dll.tar.gz) uploaded to ftp://dimcom.uqac.uquebec.ca/incoming . > > > > How do I get the files from here? I just get "User anonymous access denied." > > > > Let's wait for ftp site maintainer will move these files to /pub/gnu-win32/bin... > > -- > Sergey Okhapkin I think Jim had a point here. The server doesn't seem to allow any anonymous logins at all. So we won't be able to see anything in any pub area there, whenever it is being moved there or not! Christian - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From rsvagera@atpibm6000.tuwien.acdot Thu May 29 09:16:00 1997 From: rsvagera@atpibm6000.tuwien.acdot (rsvagera@atpibm6000.tuwien.acdot) Date: Thu, 29 May 1997 09:16:00 -0000 Subject: Problem running "g++ --help" Message-ID: At 09:13 PM 27/5/97, James Paul Morgan wrote: > >Under 17.1, I get this error with this rather amusing path: > >bash$ g++ --help >/usr/H-i386-cygwin32/lib/gcc-lib/i386-cygwin32/cygnus-2.7.2-961023/.. >/../../../i386-cygwin32/lib/libcygwin.a(libcmain.o): In function `main': >/pizza/mushroom/noer/beta17/src/winsup/libcmain.cc:30: undefined >reference to `WinMain@16' > I have the same problem with 'g++ --help' or simply 'g++' under B18/Win95. On the other side I can compile small programs like #include main() { cout << "Hello there GNU++ world!\n"; } with the command line g++ -o testpp.exe testpp.cpp when doing this on a local drive. If testpp.cpp is stored on a network drive, I again receive D:\gnuwin32\b18\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32\cygnus-2.7.2-97040 4\../../../../i386-cygwin32\lib/libcygwin.a(libcmain.o)(.text+0x1e):libcmain .cc: undefined reference to `WinMain@16' g++: Internal compiler error: program ld got fatal signal 1 Any hint? Thank you in advance, Robert - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From mh@mike.franken.de Thu May 29 09:17:00 1997 From: mh@mike.franken.de (Michael Hirmke) Date: Thu, 29 May 1997 09:17:00 -0000 Subject: bash path search bug & workarounds References: <338C99FC@gandalf.optimedia.co.il> Message-ID: <6XmQi4S4pfB@mike.franken.de> Hi Vassilii, > > I have both BC5 and gnu-win32 installed on my NT4. > The gnu-win32 executables come before BC5 in my path. > > I wanted to create such an environment in which > in CMD -- GREP.COM and MAKE.EXE from BC > in bash -- gnu grep & make > would start up when I simply type "grep" and "make". Why don't you just create a .bashrc in your $HOME directory and change the $PATH for that bash session in it ? You could even do the same thing for cmd.exe with a small DOS batch. Just create a shortcut with the following command line: C:\WINDOWS\system32\cmd.exe /e:2048 /k :\\dosrc.bat Then dosrc.bat gets executed and afterwards you get a command prompt. If you change/set any environment variables in dosrc.bat, they exist for the actual session. [...] > > Your > BACbKA Bye. Michael. -- Ich bin vom 1. Juni bis 3. Juli in Urlaub, Ihr muesst Euch also leider bis dahin mit einer Antwort auf Euere Postings und Mails gedulden :)) -- I'm on vacation from June, 1st to Juli, 3rd - so please be patient to get an answer to your postings and mails until I'll be back - thx :)) -- Michael Hirmke | Telefon +49 (911) 557999 Georg-Strobel-Strasse 81 | FAX +49 (911) 557664 90489 Nuernberg | E-Mail mh@mike.franken.de | WWW http://minimike.franken.de/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From noer@cygnus.com Thu May 29 16:18:00 1997 From: noer@cygnus.com (Geoffrey Noer) Date: Thu, 29 May 1997 16:18:00 -0000 Subject: Uname -m and arch References: <199612160553.WAA10796@joplin.colorado.edu> Message-ID: <199705292242.PAA00741@cirdan.cygnus.com> Fabio Somenzi wrote: [...] > When I type "uname -m" on my Pentium 120 running Win95, I get: > i6395286 > Under Linux, I would get i586. Is this as it should be? If so, what is > the meaning of 6395286? Either you have an Intel i6395286 chip in your machine or you're running into a cygwin.dll bug. I have a hunch as to which is more likely. :-) What's happening is that uname() isn't setting the processor level correctly. uname() gets some of its info using the SYSTEM_INFO struct. This structure has various members many of which aren't supported under both Windows 95 and NT. :-( I've fixed the development sources so uname will behave better under Windows 95 in future releases. -- Geoffrey Noer noer@cygnus.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From ian@parti.inforum.org Thu May 29 16:18:00 1997 From: ian@parti.inforum.org (Ian) Date: Thu, 29 May 1997 16:18:00 -0000 Subject: problems setting TERM Message-ID: <3.0.1.32.19970529191756.0083c760@192.160.61.2> In my .bashrc, I have the line: export TERM=linux and echo $TERM does indeed show the TERM as linux, but ^L doesn't clear the screen, long lines don't wrap, etc. If I type exactly the same thing from the command line, however, ^L works, lines wrap, and so on. Is there some reason why bash is ignoring the setting from my .bashrc? System details: win95 (4.00.950), B18 (with Sergey's patched cygwin.dll (the one the whole "My patches" thread is about) -- Ian Reinhart ian@parti.inforum.org - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jp@nuancecom.com Thu May 29 16:18:00 1997 From: jp@nuancecom.com (JP Shipherd) Date: Thu, 29 May 1997 16:18:00 -0000 Subject: More //F/ style woes Message-ID: <3.0.32.19970529151121.00989d20@awesome.nuance.com> Currently gcc does not allow switches like: -I//f/include/path It has been argued here that gcc can understand dos style paths like: -If:/include/path This leads to unfortunate problems when using generated dependency lists. The colon in the path confuses gnumake. I usually have gcc generate a dependency file for me by using the -MM switch. Something like: gcc -IF:/include/path -MM test-nuance-config.c >> .Dependencies The problem is that this puts things like F:/include/path/header.h into my Dependency file and gnumake chokes complains: /nuance/nuance-config/src> make .Dependencies.win32:16: *** multiple target patterns. Stop. If I get rid of the drive letter colon things, gnumake works along quite happily. So does anyone have a solution on how to address drive/path combinations in such a way that will work with both gcc and gnumake. Currently I'm parsing the .Dependencies file and replacing instances of F: with //F, but frankly it seems like we need to find a method of naming drive/paths that work for all the tools. Thanks, --jp - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From kunglao@prairienet.org Thu May 29 16:18:00 1997 From: kunglao@prairienet.org (Seymore Buttz) Date: Thu, 29 May 1997 16:18:00 -0000 Subject: less & searching - never mind Message-ID: <338DF364.E3BB9FCF@prairienet.org> I guess less's configure wasn't putting the regexp.o file in Makefile. Added that to the OBJS= line and works now. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From tphan@asl.dl.nec.com Thu May 29 17:27:00 1997 From: tphan@asl.dl.nec.com (Timothy Phan) Date: Thu, 29 May 1997 17:27:00 -0000 Subject: mount binary vs. text Message-ID: <199705292116.QAA02009@aslws156.asl.dl.nec.com> Hi, I'd like to know what should I set to the 'mount' (bin or text). I've download the mpack/uudecode/pgp on to NT and build with little or no problem. I, however, have problem to run this program under any partition mount as 'text'. I've tried to mount every in 'binary' mode, but 'bash' seems to have problem reading the .bashrc file. I guessed that it could not find the end of line marker since the .bashrc was save with non-gnu editors. secondly, I'd like to know who has successfully built any Unix programs (i.e. vi) that use curses/ncurses/termcap/etc libraries and which of this shall I use. And what TERM type should I set to? Many thanks! -- -------------------- Timothy C. Phan (tphan@asl.dl.nec.com) -------------------- NEC America, Inc. ASL -------------------- 1525 Walnut Hill Ln. Irving, TX 75038 -------------------- tel: (214)-518-3437 fax: (214)-518-3499 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From kunglao@prairienet.org Thu May 29 17:27:00 1997 From: kunglao@prairienet.org (Seymore Buttz) Date: Thu, 29 May 1997 17:27:00 -0000 Subject: Caps lock problem and/or less can't search? Message-ID: <338DE336.19CB7AFB@prairienet.org> I've noticed that when you put the caps lock on, every key applies the shift, not just letters. I think it was back in 17.1 as well. I finally got around to compiling less last night. All went well, just had to rename the iprintf(). Compiles easily. But you can't search with the / or ? commands. I haven't had a chance to investigate it further, but I figured less is a fairly popular program, maybe another kind person has already found/fixed it. So, has anyone? Many thank - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Don.Sharp@dddandr.octacon.co.uk Thu May 29 17:27:00 1997 From: Don.Sharp@dddandr.octacon.co.uk (Don Sharp) Date: Thu, 29 May 1997 17:27:00 -0000 Subject: CMD, case-insensitive file-name completion References: <0006A9FE.1893@po.cle.ab.com> Message-ID: <338DF38E.5E96@dddandr.octacon.co.uk> Scott Mintz wrote: > > That's true. The data value below is the character. You could > theoretically change it to something else. > > The actual key is: > HKEY_CURRENT_USER\Software\Microsoft\Command Processor > > Note the space between Command and Processor is necessary. > I am running NT 3.51 SP5 and I haven't got any completion going after making the entry in my registry. What should I have to do apart from typing TAB during command line entry? Should it be a binary rather than string value? Cheers Don Sharp P.S. I have logged off and logged back in. > -Scott A. Mintz > > ______________________________ Reply Separator _________________________________ > Subject: Re: CMD, case-insensitive file-name completion > Author: RPraetorius@AspenRes.Com at Internet > Date: 5/27/97 12:08 PM > > CMD also has case-insensitive filename completion. I believe it's > undocumented, although this may have changed. . . Here's what you need > in your registry (under HKEY_CURRENT_USER) to turn it on: > > Key Name: Software\Microsoft\Command Processor > Name: CompletionChar > Type: REG_SZ > Data: 9 > > (grepping the web for CompletionChar may turn up more info). > > -------p--a--s--s--i--o--n-----n--e--e--d--s-----a-----f--a--c--e------- > "oncology recapitulates philately" --Mark Maxson Robert M. Praetorius > "balance, not symmetry" --Mark Stanley work: RPraetorius@AspenRes.Com > (attribution by Stigler) fun & recreation: rmp@PopJ1.MA.UltraNet.Com > > - > For help on using this list (especially unsubscribing), send a message to > "gnu-win32-request@cygnus.com" with one line of text: "help". - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From *jeffdb@netzone.nospam.com Thu May 29 17:27:00 1997 From: *jeffdb@netzone.nospam.com (Mikey) Date: Thu, 29 May 1997 17:27:00 -0000 Subject: Problems w/your patch on win95 Message-ID: <338f1e64.211016@smtp.netzone.com> Hi Sergey Thanks VERY much for the solution to the ^M problem, it was driving me CRAZY, I wonder why it didn't mess up on NT? there are some other caps that need to be deleted for TERM=linux to work correctly on win95, xon=xo ccc=cc xenl=xn initc=Ic. corrects no down arrow, left arrow. Where are the lib*.a files to go with the X dll's? (more useful that way no? ;^) Guess you are probably getting tired of hearing from me about problems with win95 right? ;^) First time I ran bash with your dll that I downloaded, it ate my registry. this was after running for a while with my patched version, so I figured ok, version conflict, no problem, ERD fix me right up. With your patched dll, run the attached shell script install.ncurses.txt in the .../ncurses/ directory, on a win95 box, like this ./install.ncurses.txt 2>&1|tee /tmp/errors.txt The first time I did this with your dll that I downloaded, I got a exception 0E in VCACHE.VMM during the make. LOCKUP!! Once you see "checking for ranlib", you can ^C out, try it at least twice from the same console. With the dll that I built with your winsup.diff file ;^), you will get a lockup from user.exe. No Idea why. Tried the same thing with your patched dll from ftp. same result. after ^C. It ate my registry again, I couldn't send this the first time. This quickly gets old. I think everyone appreciates the amount of work you have put in on improving gnu-win32, but just like with prebeta18, you or someone needs to test this stuff on all the target platforms, BEFORE it gets released, OK? -------------- next part -------------- A non-text attachment was scrubbed... Name: filesize Type: text/x-shellscript Size: 102 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: makepkg Type: text/x-shellscript Size: 5291 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: terminfo.dif Type: text/x-diff Size: 978 bytes Desc: not available URL: From david@coent.demon.co.uk Thu May 29 22:51:00 1997 From: david@coent.demon.co.uk (David Coe) Date: Thu, 29 May 1997 22:51:00 -0000 Subject: My patches References: <01BC69E9.56340F00@gater.krystalbank.msk.ru> Message-ID: <338E67B1.5436@coent.demon.co.uk> Splendid stuff again, Sergey! The ftp site is inacessible but I got your patches from http://wwwdim.uqac.uquebec.ca/ftp/pub/gnuwin32/bin/ X now works on B18 at last (I was about to backgrade to B17). A few peculiarities in my use of the G77 have automagically disappeared. The X driver from the Pgplot libraries no longer opens incorrectly twice instead of reusing the old window; repeated G77 system calls to "ls > file" no longer prevoke file permission messages. I have notice that using -lc now gives duplication of _impure_ptr; I assume libc.a needs fixing to match your libcygwin.a? Many thanks -- Dr David Coe \=\ 58 Fairlawn Drive, East Grinstead \=\ Tel +44 1342 326860 West Sussex, RH19 1NT, United Kingdom \=\ Fax +44 1342 316019 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From tomaz.mertelj@ijs.si Fri May 30 00:12:00 1997 From: tomaz.mertelj@ijs.si (Tomaz Mertelj) Date: Fri, 30 May 1997 00:12:00 -0000 Subject: Sergey's patches References: <199705291305.JAA05961@woodmore.gsfc.nasa.gov> Message-ID: <3.0.1.32.19970530085650.006a9c14@cathy.ijs.si> At 09:05 AM 5/29/97 -0400, you wrote: >I was able to get the libraries this morning. Do not use ftp, go >thru the URL: > > http://wwwdim.uqac.uquebec.ca/ftp/pub/gnuwin32/bin > >However, I tried replacing my win95 DLL's with Sergey's without much >luck. Same error: unable to run. One may need the corresponding *.a >as well, I am not sure. > >Geoffrey: any clue why the X11R6.3 are failing to be built on win95? > > Arlindo. > It works both on WIN95 and NT4.0 if you replace also B18 cygwin.dll with Sergey's patched one. But unfortunately, imake and some other binutils stop to work with the new cygwin.dll. With the new cygwin.dll I was also able to succesfuly build und run the exemplary reloctable dll at http://www.cygnus.com/misc/gnu-win32/building-reloc-dlls.txt which does not run with the original one. It seems there is a bug in the binary distribution of B18 cygwin.dll. Regards, Tomaz Mertelj ********************************************************************* Dr. Tomaz Mertelj E-mail: tomaz.mertelj@ijs.si Jozef Stefan Institute Tel.: +386 (61) 17 73 900 P.O. Box. 3000 Fax: +386 (61) 12 35 400 1001 Ljubljana Home page: http://optlab.ijs.si/tmertelj Slovenia ********************************************************************* - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From mlof@penti.sit.fi Fri May 30 00:12:00 1997 From: mlof@penti.sit.fi (Mika Löf) Date: Fri, 30 May 1997 00:12:00 -0000 Subject: BASh again... Message-ID: <338ee285.22358359@mail.sit.fi> Isn't there any documentation on how CygWin bash differs from un*x bash? I can't seem to configure bash to read anything else but ~/.bashrc on startup. It completely ignores ~/.bash_profile, ~/.bash_login, ~/.bash_logout, /etc/profile. Neither does readline respond to changes in ~/.inputrc and currently I lack a lot of important keys, all those behind AltGr and I can't remap them in ~/.inputrc or using bind. Also i lack my language specific umlaut-characters, edv EDV. Is this normal or is my configuration screwed? I only downloaded the user tools, not the development version, do I get more info from that on bash/readline details? A quite funny detail I find in the fact that i got a lot of manual pages, but no man or less :-) ...otherwise bash is working great, much better than 4DOS for NT, 4NT. "The only stupid question is the unasked one" - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From dbe@wgn.net Fri May 30 00:12:00 1997 From: dbe@wgn.net ($Bill Luebkert) Date: Fri, 30 May 1997 00:12:00 -0000 Subject: GNU-Win32 : gettimeofday References: <338D5CA4.40E1@gintic.gov.sg> Message-ID: <338E3967.209E@wgn.net> Lim Chu Cheow wrote: > > Hi, > I just installed b18, and compiled a simple program > (given below). Compilation went through, but the > execution received some exception and died. Am I > missing something? Thanks in advance for any advice. > > ============ > #include > #include > > void main(void) > { > struct timeval tp; add: struct timezone tz; > > fprintf(stdout,"Hello world!!\n"); > gettimeofday(&tp); replace: gettimeofday(&tp,&tz); > fprintf(stdout,"Current time = %lf s\n", > ((double)tp.tv_sec+tp.tv_usec*1.0e-6)); > } -- ,-/- __ _ _ $Bill Luebkert (_/ / ) // // DBE Collectibles / ) /--< o // // http://www.wgn.net/~dbe/ -/-' /___/_<_ Mikey wrote: > Tried the same thing with your patched dll from ftp. same result. > after ^C. > > It ate my registry again, I couldn't send this the first time. > This quickly gets old. > I think everyone appreciates the amount of work you have put in on > improving gnu-win32, but just like with prebeta18, you or someone > needs to test this stuff on all the target platforms, BEFORE it gets > released, OK? Before posting those diff I've made several rebuilds of X11 and winsup on two Win95 machines. I spent many time to locate some win95-specific bugs in window code compiling on Win95. It's difficult for me to count how many times I've pressed ^C :-) And I didn't have such a problems... Don't tell me that your MD'95 never crashes on system startup or while starting notepad :-) -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From chin.cheekai@singapore.sterling.com Fri May 30 00:12:00 1997 From: chin.cheekai@singapore.sterling.com (Chin Chee-Kai) Date: Fri, 30 May 1997 00:12:00 -0000 Subject: GNU-Win32 : gettimeofday References: <338D5CA4.40E1@singapore.sterling.com> Message-ID: Just to add to this, I compiled the exact same program on b17.1 on Win95 and got a smooth compiled binary that run : C:\tmp>gcc test.c test.c: In function `main': test.c:5: warning: return type of `main' is not `int' C:\tmp>a Hello world!! Current time = 864984734.359000 s (I had to eventually revert back to b17.1 as gcc ran way too slowly on my project and the resulting binary doesn't even run.) Chin Chee-Kai Internet Email: cheekai@singapore.sterling.com Sterling Software Asia Pacific Group Tel: (65)-732-9996 Fax: (65)-733-4358 -----BEGIN PGP PUBLIC KEY BLOCK----- Version: 2.6.3i mQCNAzNQdfYAAAEEAOYCoROPNxjmfykpxfzCSzdZ9lHAiIIu2pwFjAtOxqXsgvq1 ZtBDjzDHl4ttFUSAjYHqRYaJ0P123ST0Tzt7YpKWjXf+tetcSGv45LX8JJDn25je WWFI8dM02dkfPaM8rWxyGcNKr6O6Nkois6Kjs02P/d3pSQ+aRESCVVXqJEytAAUX tCFDaGluIENoZWUtS2FpIDxjaGVla2FpQGdlbi5jby5qcD6JAJUDBRAzUHX2RIJV VeokTK0BAdEHBACSFXwEksBagVvmJp+wo6K5qrmCcTJUKaWqSnkjnHTcvNIy8VZ9 is50VHuCLcexul2YjqzAKFgLmCnnXNxKw4FPphgwLTxVlp5rQ+AzxxGSv/GInlCm rEzuQqIfkrmPTv2jIRSW3BO8DT7J+MpwJtflw/phCmJZH//cW0MbZ0/0NA== =cpCR -----END PGP PUBLIC KEY BLOCK----- On Thu, 29 May 1997, Lim Chu Cheow wrote: > Hi, > I just installed b18, and compiled a simple program > (given below). Compilation went through, but the > execution received some exception and died. Am I > missing something? Thanks in advance for any advice. > > ============ > #include > #include > > void main(void) > { > struct timeval tp; > > fprintf(stdout,"Hello world!!\n"); > gettimeofday(&tp); > fprintf(stdout,"Current time = %lf s\n", > ((double)tp.tv_sec+tp.tv_usec*1.0e-6)); > } > > ============== > - Chu-Cheow > - > For help on using this list (especially unsubscribing), send a message to > "gnu-win32-request@cygnus.com" with one line of text: "help". > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From loki@dragoncat.net Fri May 30 00:12:00 1997 From: loki@dragoncat.net (Jeremy Blackman) Date: Fri, 30 May 1997 00:12:00 -0000 Subject: Uname -m and arch References: <199705292242.PAA00741@cirdan.cygnus.com> Message-ID: On Thu, 29 May 1997, Geoffrey Noer wrote: > > Under Linux, I would get i586. Is this as it should be? If so, what is > > the meaning of 6395286? > Either you have an Intel i6395286 chip in your machine or you're running > into a cygwin.dll bug. I have a hunch as to which is more likely. :-) Indeed. Unless Intel's CPU design group has been -really- busy... :) I noticed this problem under b17 as well, and reported it, but I suspect it got lost in all the rush. My solution for the moment was to write my own 'arch' (for autoconf) and ignore the uname problem since it didn't -break- anything. > What's happening is that uname() isn't setting the processor level > correctly. uname() gets some of its info using the SYSTEM_INFO struct. > This structure has various members many of which aren't supported under > both Windows 95 and NT. :-( I've fixed the development sources so uname > will behave better under Windows 95 in future releases. Perhaps a way to set this if you have a specific need, and default to i486 or something otherwise? Unless you have a way to determine the chipset accurately. :) Only reason I suggest a way to set it is if you have something which depends on 386, 486, or 586 compiler optimizations... - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From cclim@gintic.gov.sg Fri May 30 00:12:00 1997 From: cclim@gintic.gov.sg (Lim Chu Cheow, Dr) Date: Fri, 30 May 1997 00:12:00 -0000 Subject: GNU-Win32 : gettimeofday Message-ID: Thanks! Since I don't follow the Unix standard changes very closely, I'm a little confused now. On a SVR4 (Solaris 2.4), from the man page, I get: int gettimeofday(struct timeval *tp); Whereas for a SunOS 4.1.3, I get from the man page: int gettimeofday(tp, tzp) struct timeval *tp; struct timezone *tzp; So GNU-Win32 is following an "older" Unix standard (whichever it is)? - Chu-Cheow >---------- >From: $Bill Luebkert[SMTP:dbe@wgn.net] >Sent: Friday, May 30, 1997 10:20 AM >To: Lim Chu Cheow, Dr >Cc: gnu-win32@cygnus.com >Subject: Re: GNU-Win32 : gettimeofday > >Lim Chu Cheow wrote: >> >> Hi, >> I just installed b18, and compiled a simple program >> (given below). Compilation went through, but the >> execution received some exception and died. Am I >> missing something? Thanks in advance for any advice. >> >> ============ >> #include >> #include >> >> void main(void) >> { >> struct timeval tp; > >add: > struct timezone tz; > >> >> fprintf(stdout,"Hello world!!\n"); >> gettimeofday(&tp); > >replace: > gettimeofday(&tp,&tz); > >> fprintf(stdout,"Current time = %lf s\n", >> ((double)tp.tv_sec+tp.tv_usec*1.0e-6)); >> } > >-- > ,-/- __ _ _ $Bill Luebkert > (_/ / ) // // DBE Collectibles > / ) /--< o // // http://www.wgn.net/~dbe/ >-/-' /___/_<_ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From steve_dum@mentorg.com Fri May 30 00:12:00 1997 From: steve_dum@mentorg.com (Steve Dum) Date: Fri, 30 May 1997 00:12:00 -0000 Subject: Problem running "g++ --help" References: Message-ID: <199705300618.XAA03414@sdum.wv.mentorg.com> the problem is that --help isn't an option that g++ recognizes. So when you type in g++ --help, the g++ driver just calls 'ld' to link the 'application'. ld looks at the list of object modules finds a main in libcygwin.a (in libcmain.o), but then complains because winmain is not defined - which is true, since there are no object modules included. The closest to a g++ --help is g++ -v which only prints out the version of g++. steve In message <"mr.tuwien..194:29.05.97.14.05.01"@mr.tuwien.ac.at>you write: >At 09:13 PM 27/5/97, James Paul Morgan wrote: >> >>Under 17.1, I get this error with this rather amusing path: >> >>bash$ g++ --help >>/usr/H-i386-cygwin32/lib/gcc-lib/i386-cygwin32/cygnus-2.7.2-961023/.. >>/../../../i386-cygwin32/lib/libcygwin.a(libcmain.o): In function `main': >>/pizza/mushroom/noer/beta17/src/winsup/libcmain.cc:30: undefined >>reference to `WinMain@16' >> >I have the same problem with 'g++ --help' or simply 'g++' under B18/Win95. >On the other side I can compile small programs like > > #include > > main() > { > cout << "Hello there GNU++ world!\n"; > } > >with the command line > > g++ -o testpp.exe testpp.cpp > >when doing this on a local drive. If testpp.cpp is stored on a >network drive, I again receive > >D:\gnuwin32\b18\H-i386-cygwin32\lib\gcc-lib\i386-cygwin32\cygnus-2.7.2-97040 >4\../../../../i386-cygwin32\lib/libcygwin.a(libcmain.o)(.text+0x1e):libcmain >.cc: undefined reference to `WinMain@16' >g++: Internal compiler error: program ld got fatal signal 1 > >Any hint? > >Thank you in advance, >Robert > >- >For help on using this list (especially unsubscribing), send a message to >"gnu-win32-request@cygnus.com" with one line of text: "help". - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jbanes@midplains.net Fri May 30 00:12:00 1997 From: jbanes@midplains.net (theace) Date: Fri, 30 May 1997 00:12:00 -0000 Subject: Simple windows program won't compile Message-ID: <338DE7DB.1F7E8FA2@midplains.net> Hello. I have a little problem that I was hoping someone could help with. I can not figure out why the following program won't compile without errors. If you don't see anything wrong with it, could you try compiling it and see if you have problems? Thank you. Begin Program: From sos@prospect.com.ru Fri May 30 00:12:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Fri, 30 May 1997 00:12:00 -0000 Subject: GNU-Win32 : gettimeofday Message-ID: <01BC6CEA.AD4B1FA0@gater.krystalbank.msk.ru> Lim Chu Cheow wrote: > I just installed b18, and compiled a simple program > (given below). Compilation went through, but the > execution received some exception and died. Am I > missing something? Thanks in advance for any advice. > > ============ > #include > #include > > void main(void) > { > struct timeval tp; > > fprintf(stdout,"Hello world!!\n"); > gettimeofday(&tp); > fprintf(stdout,"Current time = %lf s\n", > ((double)tp.tv_sec+tp.tv_usec*1.0e-6)); > } > Gettimeofday() must have two arguments. -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From godfrey@io.nosc.mil Fri May 30 00:12:00 1997 From: godfrey@io.nosc.mil (Michael Godfrey) Date: Fri, 30 May 1997 00:12:00 -0000 Subject: Problems with make Message-ID: <199705300129.SAA05781@io.nosc.mil> I have encountered two problems with 'make'. I'm running NT v4.0, Cygnus B18. All mounts are binary, /tmp is in place, I'm the only user on the system, etc. The first problem is that the make binary that came with the B18 distribution outputs the following, no matter what I'm trying to make: $ make (c:\usr\gnuwin32\b18\H-i386-cygwin32\bin\make.exe 3684) In cygwin_except_handler (c:\usr\gnuwin32\b18\H-i386-cygwin32\bin\make.exe 3684) Exception trapped! (c:\usr\gnuwin32\b18\H-i386-cygwin32\bin\make.exe 3684) exception C0000005 at 10029ACE (c:\usr\gnuwin32\b18\H-i386-cygwin32\bin\make.exe 3684) exception: ax 241F457 bx 241E812 cx 45535500 dx 241F457 (c:\usr\gnuwin32\b18\H-i386-cygwin32\bin\make.exe 3684) exception: si 241E838 di 241F0EB bp 241E800 sp 241E7FC (c:\usr\gnuwin32\b18\H-i386-cygwin32\bin\make.exe 3684) exception is: STATUS_ACCESS_VIOLATION (c:\usr\gnuwin32\b18\H-i386-cygwin32\bin\make.exe 3684) Stack trace: (c:\usr\gnuwin32\b18\H-i386-cygwin32\bin\make.exe 3684) frame 0: sp = 0x241E630, pc = 0x1000CEC2 (c:\usr\gnuwin32\b18\H-i386-cygwin32\bin\make.exe 3684) frame 1: sp = 0x241E64C, pc = 0x77F94072 (c:\usr\gnuwin32\b18\H-i386-cygwin32\bin\make.exe 3684) frame 2: sp = 0x241E670, pc = 0x77F88A53 (c:\usr\gnuwin32\b18\H-i386-cygwin32\bin\make.exe 3684) frame 3: sp = 0x241E6FC, pc = 0x77F75E82 (c:\usr\gnuwin32\b18\H-i386-cygwin32\bin\make.exe 3684) frame 4: sp = 0x241E800, pc = 0x40817F (c:\usr\gnuwin32\b18\H-i386-cygwin32\bin\make.exe 3684) frame 5: sp = 0x241E878, pc = 0x406CD9 (c:\usr\gnuwin32\b18\H-i386-cygwin32\bin\make.exe 3684) frame 6: sp = 0x241F1D4, pc = 0x1000C102 (c:\usr\gnuwin32\b18\H-i386-cygwin32\bin\make.exe 3684) frame 7: sp = 0x241FF94, pc = 0x1000C113 (c:\usr\gnuwin32\b18\H-i386-cygwin32\bin\make.exe 3684) frame 8: sp = 0x241FFA0, pc = 0x416309 (c:\usr\gnuwin32\b18\H-i386-cygwin32\bin\make.exe 3684) frame 9: sp = 0x241FFB0, pc = 0x40103B (c:\usr\gnuwin32\b18\H-i386-cygwin32\bin\make.exe 3684) frame 10: sp = 0x241FFC0, pc = 0x77F1AFC1 (c:\usr\gnuwin32\b18\H-i386-cygwin32\bin\make.exe 3684) frame 11: sp = 0x241FFF0, pc = 0x0 (c:\usr\gnuwin32\b18\H-i386-cygwin32\bin\make.exe 3684) End of stack trace So, I got and built GNU make v3.75 (using MSVC++ and nmake). This version of make has successfully made a variety of good executables, although sometimes I get the following error when compiling: g++: Internal compiler error: program cc1plus got fatal signal 33 After which make stops. However, the code compiles fine if I issue the same g++ command interactively (e.g., from the bash command line instead of from within a Makefile). Has anyone else encountered these errors? If so, have you found a solution or a workaround? Any help will be greatly appreciated. Thanks. -Mike Godfrey - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Fri May 30 01:07:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Fri, 30 May 1997 01:07:00 -0000 Subject: My patches Message-ID: <01BC6CED.A80637C0@gater.krystalbank.msk.ru> David Coe wrote: > X now works on B18 at last (I was about to backgrade to B17). A few On NT or 95? Please, always report os type you are using!!! > I have notice that using -lc now gives duplication of _impure_ptr; I > assume libc.a needs fixing to match your libcygwin.a? Do not include -lc at all! Libcygwin.a contains all libc stuff and is linked automatically. -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From nicolas@bnp-eng.remcomp.com Fri May 30 01:07:00 1997 From: nicolas@bnp-eng.remcomp.com (Eric Nicolas) Date: Fri, 30 May 1997 01:07:00 -0000 Subject: DirectX with gnu-win32 ?? Message-ID: <338E8430.446B@bnp-eng.remcomp.com> Hi ! Has someone succeded in using DirectX microsoft library with gnu-win32 ? I suppose there are many changes to do in include files in order to remove all msoft specifics features. Moreover there are copyright problems. In fact, I don't mind if I have to buy MSVC++ and DirectX, but I need to use DirectX under GCC to mix full screen access under NT and gas-style assembly code. In fact, in DirectX in only need to be able to open a screen a gain access to the frame buffer. I don't need all the 3d... features. Thanks for any help. Eric NICOLAS. PS: Thanks to people who helped me in compiling Perl5.004 ! -- ----------------------------------------------------------------------------- Eric NICOLAS Tel : 01.40.14.01.39 BNP-BFI-IM Fax : 01.40.14.99.47 13, rue LaFayette EMail: nicolas@bnp-eng.remcomp.com 75009 PARIS ----------------------------------------------------------------------------- - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Fri May 30 01:07:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Fri, 30 May 1997 01:07:00 -0000 Subject: More //F/ style woes Message-ID: <01BC6CEF.4D85D010@gater.krystalbank.msk.ru> JP Shipherd wrote: > Currently gcc does not allow switches like: > -I//f/include/path > > It has been argued here that gcc can understand dos style paths like: > -If:/include/path > > So does anyone have a solution on how to address drive/path combinations in > such a way that will work with both gcc and gnumake. Currently I'm parsing > the .Dependencies file and replacing instances of F: with //F, but frankly > it seems like we need to find a method of naming drive/paths that work for > all the tools. mount f:/ /f gcc -I/f/include/path .... -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Fri May 30 01:07:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Fri, 30 May 1997 01:07:00 -0000 Subject: less & searching - never mind Message-ID: <01BC6CEE.D3083D00@gater.krystalbank.msk.ru> Seymore Buttz wrote: > I guess less's configure wasn't putting the regexp.o file in Makefile. > Added that to the OBJS= line and works now. Cdk install does not copies librx/rx.h to /usr/include so less's configure fails to check regexp type. It's a bug in cdk installation script. Also, while compiling newlib gcc fails to find "glue.h" include file. I had to create empty file by hands. -- Sergey Okhapkin Moscow, Russia Looking for a job. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From gunther.ebert@ixos-leipzig.de Fri May 30 01:47:00 1997 From: gunther.ebert@ixos-leipzig.de (Gunther Ebert) Date: Fri, 30 May 1997 01:47:00 -0000 Subject: CMD, case-insensitive file-name completion References: <0006A9FE.1893@po.cle.ab.com> <338DF38E.5E96@dddandr.octacon.co.uk> Message-ID: <338EB027.30D2@ixos-leipzig.de> Don Sharp wrote: > > I am running NT 3.51 SP5 and I haven't got any completion going after > making the entry in my registry. What should I have to do apart from > typing TAB during command line entry? Should it be a binary rather than > string value? > The completion feature does only work with NT 4.0. Gunther -- Gunther Ebert iXOS Anwendungs-Software GmbH Angerstrasse 40-42 D-04177 Leipzig Phone : +49 341 48503-0 Fax : +49 341 48503-99 E-mail: mailto:gunther.ebert@ixos-leipzig.de www : http://www.ixos-leipzig.de - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From uknt@brwlserv.micro.lucent.com Fri May 30 02:25:00 1997 From: uknt@brwlserv.micro.lucent.com (Nick Thompson) Date: Fri, 30 May 1997 02:25:00 -0000 Subject: b18 source patches for X11R6.3? Message-ID: <9705301024.ZM293@brwlserv.micro.lucent.com> Hi Looking in the archives I see X11 is now working on b18 and I have downloaded the dlls'. I'd like to build the X11R6.3 distribution under win95 so as to get access to all the bin stuff. I have applied all the b17.? patches but have had a lot of problems building under b18. I have a pure /gnuwin32/b18/... distribution and during the X11 build all references to the distribution path (libs, includes) have all occurances of 'i386' pre-processed to '1'. Can I just copy include and lib stuff into /usr/include and /usr/lib? How do I tell gnuwin where they are now? (spec file modification? Anything else?) Some of the build is looking for un.h. Where is that from? Do I need it? Xterm fails to build because ws_ and wy_pixels (?) are not defined. Any fixes? Is there a 'jumbo patch' anywhere that fixes all these things for X11R6.3 under b18? (If not and I get it going anyway, maybe I'll make one.) And why did patch fail to apply the b17.x patches correctly? Many Thanks, -- _ _______ __ __ Nick Thompson | Lucent Technologies / \/ / / __\/ // /\ Lucent Technologies | (Bell Labs Innovations) / / / /__/ _ / / Microelectronics House | nick.thompson@lucent.com /_/\_/_/____/_/\\_\/ Broad Lane, Bracknell | Tel: +44 1344 865951 \_\/\\_\____\_\/ \_\ Berkshire RG12 9GX | Fax: +44 1344 865959 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From mh@mike.franken.de Fri May 30 03:07:00 1997 From: mh@mike.franken.de (Michael Hirmke) Date: Fri, 30 May 1997 03:07:00 -0000 Subject: Sergey's patches References: <199705291305.JAA05961@woodmore.gsfc.nasa.gov> Message-ID: <6XqWDha4pfB@mike.franken.de> Hi Arlindo, [...] > However, I tried replacing my win95 DLL's with Sergey's without much > luck. Same error: unable to run. One may need the corresponding *.a > as well, I am not sure. Sorry, but this archive already contains a libcygwin.a !!! > > Geoffrey: any clue why the X11R6.3 are failing to be built on win95? > > Arlindo. Bye. Michael. -- Ich bin vom 1. Juni bis 3. Juli in Urlaub, Ihr muesst Euch also leider bis dahin mit einer Antwort auf Euere Postings und Mails gedulden :)) -- I'm on vacation from June, 1st to Juli, 3rd - so please be patient to get an answer to your postings and mails until I'll be back - thx :)) -- Michael Hirmke | Telefon +49 (911) 557999 Georg-Strobel-Strasse 81 | FAX +49 (911) 557664 90489 Nuernberg | E-Mail mh@mike.franken.de | WWW http://minimike.franken.de/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From mh@mike.franken.de Fri May 30 03:07:00 1997 From: mh@mike.franken.de (Michael Hirmke) Date: Fri, 30 May 1997 03:07:00 -0000 Subject: My patches References: <2.2.32.19970529125158.00944e9c@ma.ultranet.com> Message-ID: <6XqWDmt4pfB@mike.franken.de> Hi Larry, [...] > >> Let's wait for ftp site maintainer will move these files to > /pub/gnu-win32/bin... > >> [...] > > I'm there now and have no problems downloading the files. I can't get in via ftp, because anonymous login isn't allowed. I can't download the files via http from /incoming, because I get "Forbidden - Your client is not allowed to access the requested object." and the files are not available in /ftp/pub/gnuwin32/patches/ ! But ... :)) I finally got them today from * http://wwwdim.uqac.uquebec.ca/ftp/pub/gnuwin32/bin* !!!! And - Sergey you're a genius :)) International keyboard problem is fixed and also the background/foreground problem ! > > Larry Bye. Michael. -- Ich bin vom 1. Juni bis 3. Juli in Urlaub, Ihr muesst Euch also leider bis dahin mit einer Antwort auf Euere Postings und Mails gedulden :)) -- I'm on vacation from June, 1st to Juli, 3rd - so please be patient to get an answer to your postings and mails until I'll be back - thx :)) -- Michael Hirmke | Telefon +49 (911) 557999 Georg-Strobel-Strasse 81 | FAX +49 (911) 557664 90489 Nuernberg | E-Mail mh@mike.franken.de | WWW http://minimike.franken.de/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From ARichard@stark.cc.oh.us Fri May 30 09:19:00 1997 From: ARichard@stark.cc.oh.us (Richardson, Anthony) Date: Fri, 30 May 1997 09:19:00 -0000 Subject: Extra newline with bash read? Message-ID: <199705301343.JAA06726@dialup.oar.net> My shell scripts that use "read" to get user input aren't working. Here's a simple example: #!/bin/sh echo -n "input a1: " read a1 echo -n "input a2: " read a2 echo "a1 = " $a1 echo "a2 = " $a2 It doesn't wait for input at "read a2" and a2 doesn't get set. Is the "Enter" key being read as a pair? Is there a fix? Thanks, Tony Richardson arichard@stark.cc.oh.us - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Fri May 30 09:19:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Fri, 30 May 1997 09:19:00 -0000 Subject: Select on pipes Message-ID: <01BC6D36.F109E3B0@sos> Hi! WaitForMultipleObjects() does not works with anonymous pipes... Any ideas about select() on pipes implementation? The more code for win32 I write, the less I want to do it... -- Sergey Okhapkin Moscow, Russia Looking for a job - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From franl@world.std.omit-this.com Fri May 30 09:37:00 1997 From: franl@world.std.omit-this.com (Francis Litterio) Date: Fri, 30 May 1997 09:37:00 -0000 Subject: Controlling the title bar. Message-ID: <3396015f.96568558@world.std.com> Here's a way to get the string "bash" in the title bar of a console when it is minimized, running bash, and sitting at a shell prompt: 1. Put this in your .bashrc file: PROMPT_COMMAND="title bash" 2. Build this program, name it title.exe, and put it in a directory named in your PATH: #include void main(int argc, char * argv[]) { SetConsoleTitle(argv[1]); } -- Francis Litterio PGP Key Fingerprint: franl@world.std.omit-this.com 02 37 DF 6C 66 43 CD 2C http://world.std.omit-this.com/~franl/ 10 C8 B5 8B 57 34 F3 21 Note: email address and WWW URL altered to prevent spam. "They that can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety." -- Ben Franklin, ~1784 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From ggp@informix.com Fri May 30 09:37:00 1997 From: ggp@informix.com (Guy Gascoigne - Piggford) Date: Fri, 30 May 1997 09:37:00 -0000 Subject: CMD, case-insensitive file-name completion Message-ID: <3.0.32.19970530093131.009bda90@pop.pdx.informix.com> At 10:22 PM 5/29/97 +0100, Don Sharp wrote: >Scott Mintz wrote: >> >> That's true. The data value below is the character. You could >> theoretically change it to something else. >> >> The actual key is: >> HKEY_CURRENT_USER\Software\Microsoft\Command Processor >> >> Note the space between Command and Processor is necessary. >> > >I am running NT 3.51 SP5 and I haven't got any completion going after >making the entry in my registry. What should I have to do apart from >typing TAB during command line entry? Should it be a binary rather than >string value? > >Cheers > >Don Sharp > >P.S. I have logged off and logged back in. This worked for me imediately under NT 4.0 so my guess is that it isn't supported under 3.51 :-( I don't have that system here to check. Guy - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From cgf@bbc.com Fri May 30 09:37:00 1997 From: cgf@bbc.com (Chris Faylor) Date: Fri, 30 May 1997 09:37:00 -0000 Subject: Problems w/your patch on win95 References: <01BC6CEA.3BB3D6C0@gater.krystalbank.msk.ru> Message-ID: In article < 01BC6CEA.3BB3D6C0@gater.krystalbank.msk.ru >, Sergey Okhapkin wrote: >Mikey wrote: >> Tried the same thing with your patched dll from ftp. same result. >> after ^C. >> >> It ate my registry again, I couldn't send this the first time. >> This quickly gets old. >> I think everyone appreciates the amount of work you have put in on >> improving gnu-win32, but just like with prebeta18, you or someone >> needs to test this stuff on all the target platforms, BEFORE it gets >> released, OK? > >Before posting those diff I've made several rebuilds of X11 and winsup on >two Win95 machines. I spent many time to locate some win95-specific bugs in >window code compiling on Win95. It's difficult for me to count how many >times I've pressed ^C :-) And I didn't have such a problems... > >Don't tell me that your MD'95 never crashes on system startup or while >starting notepad :-) That happens to me all the time. How did you manage to crash my system before I ever installed your software? :-) (Thanks for all your efforts, btw). -- http://www.bbc.com/ cgf@bbc.com "Strange how unreal VMS=>UNIX Solutions Boston Business Computing the real can be." - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From dumser@ti.com Fri May 30 09:37:00 1997 From: dumser@ti.com (James Dumser) Date: Fri, 30 May 1997 09:37:00 -0000 Subject: More //F/ style woes References: <3.0.32.19970529151121.00989d20@awesome.nuance.com> Message-ID: <199705301525.KAA05598@lesol1.dseg.ti.com> On Thu, 29 May 1997 15:11:21 -0700, JP Shipherd wrote: >So does anyone have a solution on how to address drive/path combinations in >such a way that will work with both gcc and gnumake. Currently I'm parsing >the .Dependencies file and replacing instances of F: with //F, but frankly >it seems like we need to find a method of naming drive/paths that work for >all the tools. Mount F: as /f or /dos/f or /fdrive or /dev/hdf or ... :) Then use gcc -I/f/include/path -MM test-nuance-config.c >> .Dependencies -- James Dumser 972-462-5335 dumser@ti.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From vassilii@optimedia.co.il Fri May 30 09:37:00 1997 From: vassilii@optimedia.co.il (Vassilii Khachaturov) Date: Fri, 30 May 1997 09:37:00 -0000 Subject: bash path search bug & workarounds Message-ID: <338F01D6@gandalf.optimedia.co.il> Yes, it's the same workaround I proposed later on in the same posting -- but the bug's still out there... ---------- From: gnu-win32-owner To: gnu-win32 Subject: Re: bash path search bug & workarounds Date: Thursday, May 29, 1997 6:12PM Hi Vassilii, Why don't you just create a .bashrc in your $HOME directory and change the $PATH for that bash session in it ? You could even do the same thing for cmd.exe with a small DOS batch. Just create a shortcut with the following command line: C:\WINDOWS\system32\cmd.exe /e:2048 /k :\\dosrc.bat 90489 Nuernberg | E-Mail mh@mike.franken.de | WWW http://minimike.franken.de/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From tony@axnet.it Fri May 30 09:37:00 1997 From: tony@axnet.it (Tony Bom) Date: Fri, 30 May 1997 09:37:00 -0000 Subject: Caps lock problem and/or less can't Message-ID: <199705301550.AA01800@axnet.it> > Hi, Ciao! > I'd like to know what should I set to the 'mount' (bin or text). > I've download the mpack/uudecode/pgp on to NT and build with little > or no problem. I, however, have problem to run this program under > any partition mount as 'text'. I've tried to mount every in 'binary' > mode, but 'bash' seems to have problem reading the .bashrc file. > I guessed that it could not find the end of line marker since the > .bashrc was save with non-gnu editors. > > secondly, I'd like to know who has successfully built any Unix > programs (i.e. vi) that use curses/ncurses/termcap/etc > libraries and which of this shall I use. And what TERM type should > I set to? I compiled correctly ncurses (1.9.9e), without to much modifications, and even compiled programs that use it. However, the thing I had to change was the /etc/termcap file. Basicly the clear didn't work (setting TERM to linux). I changed the cl=\E[H\EJ to cl=\E[H\E2J to make it work. Furthermore I removed all other entries, because curses complained about multiple definitions. Apart from this, it works just fine! Note that you must copy your_cygnus_dir_here/etc/termcap to /etc/termcap > Many thanks! np :) Tone. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From psksvp@ccs.neu.edu Fri May 30 09:37:00 1997 From: psksvp@ccs.neu.edu (psksvp) Date: Fri, 30 May 1997 09:37:00 -0000 Subject: GNAT with B18 Message-ID: <01BC6CE7.09792230@RISA> Hi Does anyone try install GNAT for win32 with B-18? GNAT comes with its own version of GNU-WIN32, but I don't know what version. How can I install GNU-WIN32 b18 and GNAT together? thanks psksvp@ccs.neu.edu - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From franl@world.std.com Fri May 30 09:37:00 1997 From: franl@world.std.com (Francis Litterio) Date: Fri, 30 May 1997 09:37:00 -0000 Subject: [b18] Are pipes always in text mode? Message-ID: <3394ed84.91485068@world.std.com> I checked the FAQ and the list archives, but found nothing about this... In the b18 release, I'm using binary mounts, but I'm seeing text-mode-style CRNL->NL conversions on data sent down a pipe: $ od -c foo 0000000 h e l l o \r \n 0000007 $ od -c < foo 0000000 h e l l o \r \n 0000007 $ cat foo > bar; od -c bar 0000000 h e l l o \r \n 0000007 $ cat foo | od -c 0000000 h e l l o \n 0000006 First, the inconsistent text/binary mode for redirection versus pipes can probably be considered a bug. File descriptor 1 is sometimes text-mode and sometimes binary mode depending on how the application was spawned by the shell! Second, is there any way to make all data traversing pipes be treated as binary (without recompiling anything)? If not, perhaps there can be a mount option that controls the text/binary distinction for redirected and piped file descriptors (for those of us who wish to live in an all-binary-all-the-time GNU-Win32 world). -- Francis Litterio PGP Key Fingerprint: franl@world.std.com 02 37 DF 6C 66 43 CD 2C http://world.std.com/~franl/ 10 C8 B5 8B 57 34 F3 21 "They that can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety." -- Ben Franklin, ~1784 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From jp@nuancecom.com Fri May 30 10:25:00 1997 From: jp@nuancecom.com (JP Shipherd) Date: Fri, 30 May 1997 10:25:00 -0000 Subject: More //F/ style woes Message-ID: <3.0.32.19970530102347.00971e10@awesome.nuance.com> This is wonderful if you only ever use tools that use the cygwin.dll but abysmal if you use anything else. I'm beginning to come to the conclusion that to use the gnu-win32 tools in conjuction with other win32 apps, gnumake will need to change. It needs to have a behaviour where if it finds: [Space or Beginning of line][Alpha]:/ it knows not to inerpret that colon as the seperator between a target and a dependency list. Using mount and the //F/ drive naming is handy but if you're tyring to build without depending on cygwin.dll (ie minimalist gcc), it's better to keep this out of your build environment since you can't have it in your test envionment as well. Just my $.02 --jp At 10:38 AM 5/30/97 -0500, Bart Anderson wrote: >I think I saw something like this in the list before, but don't think of >multiple drives, think of a root file system "/" and mount your drives >there some place. Treat them like partitions, or mount them under >/drives. Organization is your choice. Standard directories for standard >stuff will help with the portability issues. If you don't think of >drives as drives but deal with file systems it will be much easier. You >should be carefull to avoid cross-mounting stuff, I'm not sure how that >might affect performance or other tools if you have a somehow recursive >structure. > >Opinions my own, in case my .signature didn't make it in. >Bart > > - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From cjeffare@cs.iupui.edu Fri May 30 17:16:00 1997 From: cjeffare@cs.iupui.edu (Chris Jeffares) Date: Fri, 30 May 1997 17:16:00 -0000 Subject: Korn Shell Message-ID: Is there a Korn shell for the b18 ?? If not is there a way to run a Korn Shell on an NT platform ?? Thanks for the time... Chris J =+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+= Chris Jeffares Graduate Research Assistant Purdue University at Indianapolis office ==> SL041, (317) 274-3089 email ==> cjeffare@cs.iupui.edu www ==> http://klingon.iupui.edu/~cjeffare/home.html The Nebraska Cornhuskers, Gotta luv 'em Go Big Red!!...Huckin' Fuskers in 1997 "Experience...the name everyone gives to their mistakes." =+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+= - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From david@coent.demon.co.uk Fri May 30 20:46:00 1997 From: david@coent.demon.co.uk (David Coe) Date: Fri, 30 May 1997 20:46:00 -0000 Subject: My patches References: <01BC6CED.A80637C0@gater.krystalbank.msk.ru> Message-ID: <338F20FF.7A38@coent.demon.co.uk> Further elaboration on experience with Sergey's B18 patch with NT 4: o both the new cygwin.dll and libX??.dll's are needed for X11R6.3, neither on their own is enough. o previously compiled (B17.1) X clients now work fine. o previously supplied libX??.a stubs work fine for rebuilding clients. o NT extended attributes are a mixed blessing ... copying NTFS files generated under SOS cygwin.dll across to VFAT gives an ever-growing "EA DATA. SF" file that needs CHKDSK to repair. -- Dr David Coe \=\ 58 Fairlawn Drive, East Grinstead \=\ Tel +44 1342 326860 West Sussex, RH19 1NT, United Kingdom \=\ Fax +44 1342 316019 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From dasilva@woodmore.gsfc.nasa.gov Fri May 30 20:46:00 1997 From: dasilva@woodmore.gsfc.nasa.gov (Arlindo da Silva) Date: Fri, 30 May 1997 20:46:00 -0000 Subject: X11R6.3 now works on win95!!! Message-ID: <199705310132.VAA06974@woodmore.gsfc.nasa.gov> Yes, I also got X11R6.3 working with DLLs on win95! You need both the X11 DLLs Sergey built on NT, as well as his cygwin.dll patched dll. (Previously I had only Sergey's X11 DLLs and it didn't work.) I'll finish building some clients and I will make the binaries available. I've been also able to build lesstif, the GNU clone of Motif, basicaly out of the tar file; static libs for now. I'll post these as well after a bit of testing. Many thanks again to Sergey for all his work on the X11R6.3 port. Arlindo. -- *-----------------------------------------------------------------* | | | ^|^ ARLINDO DA SILVA | | | 212 Lexington Drive | | _|_ Silver Spring, MD 20901 | | @ ___|___|__ | | _|_|__________|_______/ mailto://dasilva@alum.mit.edu | | \ o o o o o o o / | | \_______________/ Home: (301) 754-1121 | | ^^^^^^^^^^^^^^^^^^^^^^^^ FAX: (301) 805-7959 | | | *-----------------------------------------------------------------* - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Fabio@Colorado.EDU Fri May 30 20:46:00 1997 From: Fabio@Colorado.EDU (Fabio Somenzi) Date: Fri, 30 May 1997 20:46:00 -0000 Subject: Extra newline with bash read? References: <199705301343.JAA06726@dialup.oar.net> Message-ID: <199705310346.VAA03898@boulder.Colorado.EDU> >>>>> "R" == Richardson,Anthony writes: R> My shell scripts that use "read" to get user input aren't working. R> Here's a simple example: R> #!/bin/sh echo -n "input a1: " read a1 echo -n "input a2: " read R> a2 echo "a1 = " $a1 echo "a2 = " $a2 R> It doesn't wait for input at "read a2" and a2 doesn't get set. Is R> the "Enter" key being read as a pair? Is there a fix? You may want to try "C-j" instead of "Enter." It's a workaround rather than a fix. Fabio -- Fabio Somenzi | Phone: 303-492-3466 University of Colorado | Fax: 303-492-2758 ECE Dept. | Email: Fabio@Colorado.EDU Boulder CO 80309-0425 | WWW: http://vlsi.colorado.edu/~fabio - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From mh@mike.franken.de Fri May 30 20:46:00 1997 From: mh@mike.franken.de (Michael Hirmke) Date: Fri, 30 May 1997 20:46:00 -0000 Subject: Caps lock problem and/or less can't search? References: <338DE336.19CB7AFB@prairienet.org> Message-ID: <6XqZ1gi4pfB@mike.franken.de> Hi Seymore, > I've noticed that when you put the caps lock on, every key applies the > shift, not just letters. I think it was back in 17.1 as well. > > I finally got around to compiling less last night. All went well, just > had to rename the iprintf(). Compiles easily. But you can't search > with the / or ? commands. I haven't had a chance to investigate it > further, but I figured less is a fairly popular program, maybe another > kind person has already found/fixed it. So, has anyone? Only thing, I can say - I came across the same problem and no solution yet. > > Many thank Bye. Michael. -- Ich bin vom 1. Juni bis 3. Juli in Urlaub, Ihr muesst Euch also leider bis dahin mit einer Antwort auf Euere Postings und Mails gedulden :)) -- I'm on vacation from June, 1st to Juli, 3rd - so please be patient to get an answer to your postings and mails until I'll be back - thx :)) -- Michael Hirmke | Telefon +49 (911) 557999 Georg-Strobel-Strasse 81 | FAX +49 (911) 557664 90489 Nuernberg | E-Mail mh@mike.franken.de | WWW http://minimike.franken.de/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From mka@redes.int.com.mx Fri May 30 20:46:00 1997 From: mka@redes.int.com.mx (Michael Anderson) Date: Fri, 30 May 1997 20:46:00 -0000 Subject: tty programming for modems and fax References: <01BC69E4.10E4EEB0@gater.krystalbank.msk.ru> <338B5CC5.5AED293D@redes.int.com.mx> <338F1B9A.5D2F@switch.rockwell.com> Message-ID: <338F3E1F.2266F20C@redes.int.com.mx> Bart Anderson wrote: > > How's that serial code coming? > -- Hi Bart: After I sent the notice about the changes for tty programming, I switched to Unix to program the next tty project (the first was an outgoing fax driver - the second is an incoming data modem driver). This driver has different tty demands (termio flags) and thus will need additional changes to the NT dll. I'm just today (or tommorrow) porting the new driver to NT and will post the first set of changes as soon as I figure out which files were changed. I don't recomend that anyboby take the changes seriously until they have personally tested them against their own code. My fax driver works but only uses a small piece of the changed code. I plan to assign the code to Cygnus for inclusion in the base product as soon as I understand what those copyright issues are (if I agree to them.) It would be great to have several people validate and correct the code before it makes it into a release. Sorry for the delay, but I only have one usable programming machine - its either in NT or Unix and Unix has had the last few days. Mike Anderson mka@redes.int.com.mx - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From tphan@asl.dl.nec.com Fri May 30 22:15:00 1997 From: tphan@asl.dl.nec.com (Timothy Phan) Date: Fri, 30 May 1997 22:15:00 -0000 Subject: vim/vile on B18 Message-ID: <199705310515.AAA05479@aslws156.asl.dl.nec.com> Hi, Has anyone built vim4.6 or vile7.1 using B18 and worked? I've built both editors with ncurses 1.9.9 successfully, but unfortunately, NEITHER editors could run/work as exected. Anyone has any idea? Thanks in advance! -- -------------------- Timothy C. Phan (tphan@asl.dl.nec.com) -------------------- NEC America, Inc. ASL -------------------- 1525 Walnut Hill Ln. Irving, TX 75038 -------------------- tel: (214)-518-3437 fax: (214)-518-3499 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From dumser@ti.com Fri May 30 22:15:00 1997 From: dumser@ti.com (James Dumser) Date: Fri, 30 May 1997 22:15:00 -0000 Subject: Sergey's patches References: <3.0.1.32.19970530085650.006a9c14@cathy.ijs.si> Message-ID: <199705301844.NAA25730@lesol1.dseg.ti.com> On Fri, 30 May 1997 08:56:50 +0200, Tomaz Mertelj wrote: >It works both on WIN95 and NT4.0 if you replace also B18 cygwin.dll with >Sergey's patched one. But unfortunately, imake and some other binutils stop >to work with the new cygwin.dll. > >With the new cygwin.dll I was also able to succesfuly build und run the >exemplary reloctable dll at > http://www.cygnus.com/misc/gnu-win32/building-reloc-dlls.txt which does not >run with the original one. One thing you need to be aware of is that switching to Sergey's cygwin.dll changes the way symlinks are identified. If you had symlinks prior to installing the rebuilt cygwin.dll, they don't disappear; but they are no longer recognized as symlinks until you set their system bit (requiring the use of attrib, not chmod -- or changing properties one by one). As Sergey has explained, the reason for doing this was to spead up fstats. Previously, each file had to be opened and read to determine if it was a symbolic link. With Sergey's modifications, the open/read test is only done if the system bit is set -- if it's not set, cygwin.dll assumes it's not a symlink. At one point, I believe Sergey indicated he had a utility to search for and update these "non-conforming" symlink files; but it was not included in the files he uploaded for distribution. I found my symlinks using Start->Find->Files or Folders->Advanced to search for "!". (For more than one partition/drive, you have to set the "Look in:" field under Name & Location.) Save the list to a file. If you set the system bit on extra files, it's not catastrophic; but you lose some of the performance speedup this change is meant to provide so it would be best if you verify your list to make sure they are actually symlinks. (For example, cygwin.dll will be found because it contains "!", but it is definitely not a symlink.) Real symlinks will contain only 1 line ("!" & the path pointed to) so they should be very small -- less than 100 bytes unless you've got REALLY long directory paths. Once verified, you can do something like for f in `cat links.list`; do attrib +s $f done (Realize that attrib expects DOS-style [C:\bin\link] filenames.) -- James Dumser 972-462-5335 dumser@ti.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From bdoyle@usgs.gov Fri May 30 22:15:00 1997 From: bdoyle@usgs.gov (Brian S Doyle, Computer Clk, Lawrence, KS ) Date: Fri, 30 May 1997 22:15:00 -0000 Subject: Mail from NCSA X Mosaic 2.7b5 Message-ID: <199705301900.TAA05676@ais7dkslwr.cr.usgs.gov> I am running NT4 and WincCenter, I downloaded cdk.exe from latest and when I try to run cdk, i get an error. "The CMOS file cmos.ram could no be updated." Then i have to terminate, Any help? - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From mike@mbsun.mlb.org Fri May 30 22:15:00 1997 From: mike@mbsun.mlb.org (Mike Bernson) Date: Fri, 30 May 1997 22:15:00 -0000 Subject: Serial/TTY code Message-ID: I am having problems getting the serial code to work under beta18. Write seem to work but read hang. I have not taken a close look at ffhandle.cc yet. Are there any patches dealing with accessing serial ports. Has anyone gotten serial code to work. -- Mike Bernson (mike@mbsun.mlb.org) - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From noer@cygnus.com Fri May 30 22:15:00 1997 From: noer@cygnus.com (Geoffrey Noer) Date: Fri, 30 May 1997 22:15:00 -0000 Subject: BASh again... References: <338ee285.22358359@mail.sit.fi> Message-ID: <199705301946.MAA21423@cirdan.cygnus.com> > Isn't there any documentation on how CygWin bash differs from un*x > bash? Well, it is the Unix bash. :-) In the bash sources distrubuted with recent gnu-win32 beta releases, you will find a file called ChangeLog.Cygnus which discribes the modifications made. I'd be happy to include changes to make this version of bash faster or behave better. I'd be even happier if someone got bash 2.0 working, including doing the configury work so you can build it natively and with cross-tools from Unix. > I only downloaded the user tools, not the development version, do I > get more info from that on bash/readline details? Nope. There are a couple of bash setup pages pointed to from the project web page: http://www.cygnus.com/misc/gnu-win32 I haven't had time yet to work on including ncurses, less, and man yet although all three would be good to include in the gnu-win32 distributions. I would have to do work to not include the man pages but getting man and less takes some engineering effort and this has been fairly low priority for me. Last time I tried, all three programs need configury work before they can be built with a cross-compiler which would be highly desirable. -- Geoffrey Noer noer@cygnus.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From kuku@gilberto.physik.rwth-aachen.de Sat May 31 02:56:00 1997 From: kuku@gilberto.physik.rwth-aachen.de (Christoph Kukulies) Date: Sat, 31 May 1997 02:56:00 -0000 Subject: Korn Shell References: Message-ID: <19970531115935.51563@gil.physik.rwth-aachen.de> On Fri, May 30, 1997 at 07:12:00PM -0500, Chris Jeffares wrote: > > Is there a Korn shell for the b18 ?? > > If not is there a way to run a Korn Shell on an NT platform ?? UWIN (www.att.com) (written by David Korn - no sources, btw) would give you ksh. > > Thanks for the time... > > Chris J > > =+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+= > > Chris Jeffares > Graduate Research Assistant Purdue University at Indianapolis > > office ==> SL041, (317) 274-3089 > email ==> cjeffare@cs.iupui.edu > www ==> http://klingon.iupui.edu/~cjeffare/home.html > > The Nebraska Cornhuskers, Gotta luv 'em > Go Big Red!!...Huckin' Fuskers in 1997 > > "Experience...the name everyone gives to their mistakes." > > =+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+= > > - > For help on using this list (especially unsubscribing), send a message to > "gnu-win32-request@cygnus.com" with one line of text: "help". -- -- Chris Christoph P. U. Kukulies kuku@gil.physik.rwth-aachen.de - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From root@jacob.remcomp.fr Sat May 31 02:56:00 1997 From: root@jacob.remcomp.fr (root) Date: Sat, 31 May 1997 02:56:00 -0000 Subject: Simple windows program won't compile References: <338DE7DB.1F7E8FA2@midplains.net> Message-ID: > > Hello. I have a little problem that I was hoping someone could help > with. I can not figure out why the following program won't compile > without errors. If you don't see anything wrong with it, could you try > compiling it and see if you have problems? Thank you. > > Begin Program: > > > #include > > LRESULT CALLBACK MyFunc(HWND, UNIT, WPARAM, LPARAM); ^^^^ 'UNIT' doesn't exist... UINT does. > [snip snip] > > } > > LRESULT CALLBACK MyFunc(HWND hwnd,UNIT message, WPARAM wParam, LPARAM lParam) ^^^^ Same as above The problem with brains, is that they see what they want to see. Machines are different. I have done this same kind of mistake many times, and can look hundreds of times the same syntax error without seeing anything wrong with it! -- Jacob Navia Logiciels/Informatique 41 rue Maurice Ravel Tel 01 48.23.51.44 93430 Villetaneuse Fax 01 48.23.95.39 France - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From kunglao@prairienet.org Sat May 31 02:56:00 1997 From: kunglao@prairienet.org (Seymore Buttz) Date: Sat, 31 May 1997 02:56:00 -0000 Subject: More //F/ style woes Message-ID: <338F597A.418A52E0@prairienet.org> Sergey Okhapkin wrote: > > mount f:/ /f > gcc -I/f/include/path .... Rather than cluttering up the mount table, making a symlink to your dos drives also works. On my setup, I made a dir /dos. Then put symlinks to all my drives in there. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From noer@cygnus.com Sat May 31 05:36:00 1997 From: noer@cygnus.com (Geoffrey Noer) Date: Sat, 31 May 1997 05:36:00 -0000 Subject: Uname -m and arch References: Message-ID: <199705302340.QAA21886@cirdan.cygnus.com> [...] > Perhaps a way to set this if you have a specific need, and default to i486 > or something otherwise? [...] Following is a replacement winsup/uname.cc for those who care. -- Geoffrey Noer noer@cygnus.com ------------------------------> snip <----------------------------------- /* uname.cc: uname for WIN32. Copyright 1996, 1997 Cygnus Solutions Written by Steve Chamberlain of Cygnus Support Rewritten by Geoffrey Noer of Cygnus Solutions This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */ #include #include #include #include "winsup.h" extern "C" int uname (struct utsname *name) { DWORD len, version; SYSTEM_INFO sysinfo; GetSystemInfo (&sysinfo); /* name of computer */ memset (name, 0, sizeof (*name)); len = sizeof (name->nodename) - 1; GetComputerNameA (name->nodename, &len); /* operating system type */ if (windows_95 ()) strcpy (name->sysname, "CYGWIN32/95"); else strcpy (name->sysname, "CYGWIN32/NT"); /* operating system revision */ version = GetVersion (); sprintf (name->release, "%d.%d", version & 255, (version >> 8) & 255); /* Cygwin32 dll revision */ sprintf (name->version, "%d.%d", CYGWIN_DLL_VERSION_MAJOR, CYGWIN_DLL_VERSION_MINOR); /* CPU type */ switch (sysinfo.u.s.wProcessorArchitecture) { case PROCESSOR_ARCHITECTURE_INTEL: /* But which of the x86 chips are we? */ if (windows_95 ()) { /* dwProcessorType only valid in Windows 95 */ if ((sysinfo.dwProcessorType == PROCESSOR_INTEL_386) || (sysinfo.dwProcessorType == PROCESSOR_INTEL_486) || (sysinfo.dwProcessorType == PROCESSOR_INTEL_PENTIUM)) sprintf (name->machine, "i%d", sysinfo.dwProcessorType); else strcpy (name->machine, "unknown"); } else { /* wProcessorLevel only valid in Windows NT */ sprintf (name->machine, "i%d86", sysinfo.wProcessorLevel); } break; case PROCESSOR_ARCHITECTURE_PPC: strcpy (name->machine, "ppc"); break; case PROCESSOR_ARCHITECTURE_ALPHA: strcpy (name->machine, "alpha"); break; case PROCESSOR_ARCHITECTURE_MIPS: strcpy (name->machine, "mips"); break; default: sprintf (name->machine, "unknown"); break; } return 0; } - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From dumser@ti.com Sat May 31 05:36:00 1997 From: dumser@ti.com (James Dumser) Date: Sat, 31 May 1997 05:36:00 -0000 Subject: GNU-Win32 : gettimeofday References: Message-ID: <199705301859.NAA28301@lesol1.dseg.ti.com> On Fri, 30 May 1997 10:43:28 +0800, Lim Chu Cheow, Dr wrote: >Thanks! Since I don't follow the Unix standard changes very >closely, I'm a little confused now. On a SVR4 (Solaris 2.4), >from the man page, I get: > > int gettimeofday(struct timeval *tp); > >Whereas for a SunOS 4.1.3, I get from the man page: > > int gettimeofday(tp, tzp) > struct timeval *tp; > struct timezone *tzp; > >So GNU-Win32 is following an "older" Unix standard (whichever it is)? It's more of a System V (Solaris) versus BSD (SunOS) thing. Cygwin currently follows BSD closer than System V because there's lots of BSD source code available that they can grab for free (System V source would have to be licensed from AT&T or developed completely independently). -- James Dumser 972-462-5335 dumser@ti.com - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Sat May 31 05:36:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Sat, 31 May 1997 05:36:00 -0000 Subject: BASh again... Message-ID: <01BC6DE0.FFEDD5D0@sos> Geoffrey Noer wrote: > I'd be even happier if someone got bash 2.0 working, including > doing the configury work so you can build it natively and with Run configure, add "#define _POSIX_VERSION" to config.h.top, run make and enjoy :-) -- Sergey Okhapkin Moscow, Russia Looking for a job - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From stratos@dnn.gr Sat May 31 05:44:00 1997 From: stratos@dnn.gr (Stratos Margaritis - System Admin) Date: Sat, 31 May 1997 05:44:00 -0000 Subject: Controlling the title bar. Message-ID: <97053111094542500@athena.dnn.gr> > Here's a way to get the string "bash" in the title bar of a console > when it is minimized, running bash, and sitting at a shell prompt: > > 1. Put this in your .bashrc file: > > PROMPT_COMMAND="title bash" > > 2. Build this program, name it title.exe, and put it in a directory > named in your PATH: > > #include > void main(int argc, char * argv[]) > { > SetConsoleTitle(argv[1]); > } > -- > Francis Litterio PGP Key Fingerprint: > franl@world.std.omit-this.com 02 37 DF 6C 66 43 CD 2C > http://world.std.omit-this.com/~franl/ 10 C8 B5 8B 57 34 F3 21 > Note: email address and WWW URL altered to prevent spam. > > "They that can give up essential liberty to obtain a little temporary > safety deserve neither liberty nor safety." -- Ben Franklin, ~1784 > - > For help on using this list (especially unsubscribing), send a message to > "gnu-win32-request@cygnus.com" with one line of text: "help". I've used that on my WIN NT4.0 and it want change the title bar, any suggestions ??? Stratos :-) System Administrator & Software Engineer at DNNet ---------------------------------------------------------------------------- Name -> Stratos Margaritis e-mail -> stratos@dnn.gr, webmaster@dnn.gr, hostmaster@dnn.gr, postmaster@dnn.gr WWW -> http://www.dnn.gr/~stratos/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From allan@interport.net Sat May 31 05:44:00 1997 From: allan@interport.net (Allan Peda) Date: Sat, 31 May 1997 05:44:00 -0000 Subject: Korn Shell References: <19970531115935.51563@gil.physik.rwth-aachen.de> Message-ID: <33916EC8.3A64@interport.net> Christoph Kukulies wrote: > > On Fri, May 30, 1997 at 07:12:00PM -0500, Chris Jeffares wrote: > > > > Is there a Korn shell for the b18 ?? > > > > If not is there a way to run a Korn Shell on an NT platform ?? > > UWIN (www.att.com) (written by David Korn - no sources, btw) > would give you ksh. Look at http://www.research.att.com/sw/tools/uwin/ Ciao Allan - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From drs@cs.wisc.edu Sat May 31 08:49:00 1997 From: drs@cs.wisc.edu (dave) Date: Sat, 31 May 1997 08:49:00 -0000 Subject: ioctl() --where? Message-ID: <339047A7.1716@cs.wisc.edu> Greetings, I need ioctl(), which on solaris is in unistd.h, but I can't find it in the cygwin32 distribution. Anyone know where it is or if it is supported? Thanks, please respond directly to drs@cs.wisc.edu, as I am not currently on the mailing list. Thanks again! dave - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From mka@redes.int.com.mx Sat May 31 08:49:00 1997 From: mka@redes.int.com.mx (Michael Anderson) Date: Sat, 31 May 1997 08:49:00 -0000 Subject: tty programming code changes Message-ID: <338F5809.54C5D82B@redes.int.com.mx> Attached is a tar/gzip file containing the changes I made to support tty programming on NT. The files are: ./winsup/console.cc ./winsup/fcntl.cc ./winsup/fhandler.cc ./winsup/fhandler.h ./winsup/include/fcntl.h ./winsup/include/sys/termios.h ./winsup/path.cc ./winsup/syscalls.cc ./winsup/syscalls.h ./winsup/termios.cc These are the entire files, not diffs (I didn't keep the originals). Be sure to save copies of the originals before overwritng with these. Most of the code is untested so please satisfy yourself that it works correctly before depending on it. I am still making changes to the code to support other features that I need in my own project. Corrections, suggestions, and criticisms are welcome. I know almost nothing about Win32 API, so there may be better approaches to the tty problem than the ones in the code. In particular, I made no attempt to accomodate NT vs. Win95 differences. If your mailer doesn't support attachments or you can't otherwise get the tar/gzip file, mail me and I will post the code as text in a mail message. Good luck, mike Anderson mka@redes.int.com.mx -------------- next part -------------- A non-text attachment was scrubbed... Name: changes.gz Type: application/x-gzip Size: 51035 bytes Desc: not available URL: From jman@lx.net Sat May 31 09:02:00 1997 From: jman@lx.net (jman) Date: Sat, 31 May 1997 09:02:00 -0000 Subject: Korn Shell References: <19970531115935.51563@gil.physik.rwth-aachen.de> Message-ID: <339083A5.3522@lx.net> Yes but I could never get it to work with Gnuwin32. Cat,make,gcc nothing would work.... Christoph Kukulies wrote: > > On Fri, May 30, 1997 at 07:12:00PM -0500, Chris Jeffares wrote: > > > > Is there a Korn shell for the b18 ?? > > > > If not is there a way to run a Korn Shell on an NT platform ?? > > UWIN (www.att.com) (written by David Korn - no sources, btw) > would give you ksh. > > > > > Thanks for the time... > > > > Chris J > > > > =+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+= > > > > Chris Jeffares > > Graduate Research Assistant Purdue University at Indianapolis > > > > office ==> SL041, (317) 274-3089 > > email ==> cjeffare@cs.iupui.edu > > www ==> http://klingon.iupui.edu/~cjeffare/home.html > > > > The Nebraska Cornhuskers, Gotta luv 'em > > Go Big Red!!...Huckin' Fuskers in 1997 > > > > "Experience...the name everyone gives to their mistakes." > > > > =+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+= > > > > - > > For help on using this list (especially unsubscribing), send a message to > > "gnu-win32-request@cygnus.com" with one line of text: "help". > > -- > -- > Chris Christoph P. U. Kukulies kuku@gil.physik.rwth-aachen.de > - > For help on using this list (especially unsubscribing), send a message to > "gnu-win32-request@cygnus.com" with one line of text: "help". -- Jason L. Esman aka _Jman System Admin. Network Consultant Irc Admin of irc.lx.net(newnet) Co. Owner of Den Internet Services http://www.lx.net && http://www.deninc.com PGP Keys finger jman@lx.net - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From mh@mike.franken.de Sat May 31 11:02:00 1997 From: mh@mike.franken.de (Michael Hirmke) Date: Sat, 31 May 1997 11:02:00 -0000 Subject: Simple windows program won't compile References: <338DE7DB.1F7E8FA2@midplains.net> Message-ID: <6XubbvEppfB@mike.franken.de> Hi theace, [...] > Hello. I have a little problem that I was hoping someone could help > with. I can not figure out why the following program won't compile > without errors. If you don't see anything wrong with it, could you try > compiling it and see if you have problems? Thank you. UNIT is unknown - that's the only problem I could find. Bye. Michael. -- Ich bin vom 1. Juni bis 3. Juli in Urlaub, Ihr muesst Euch also leider bis dahin mit einer Antwort auf Euere Postings und Mails gedulden :)) -- I'm on vacation from June, 1st to Juli, 3rd - so please be patient to get an answer to your postings and mails until I'll be back - thx :)) -- Michael Hirmke | Telefon +49 (911) 557999 Georg-Strobel-Strasse 81 | FAX +49 (911) 557664 90489 Nuernberg | E-Mail mh@mike.franken.de | WWW http://minimike.franken.de/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From mh@mike.franken.de Sat May 31 11:02:00 1997 From: mh@mike.franken.de (Michael Hirmke) Date: Sat, 31 May 1997 11:02:00 -0000 Subject: Extra newline with bash read? References: <199705301343.JAA06726@dialup.oar.net> Message-ID: <6XubcMW$pfB@mike.franken.de> Hi, [...] > It doesn't wait for input at "read a2" and a2 doesn't get set. Is the > "Enter" > key being read as a pair? Is there a fix? Hm, works perfectly for me with Cygnus B18. > > Thanks, > Tony Richardson Bye. Michael. -- Ich bin vom 1. Juni bis 3. Juli in Urlaub, Ihr muesst Euch also leider bis dahin mit einer Antwort auf Euere Postings und Mails gedulden :)) -- I'm on vacation from June, 1st to Juli, 3rd - so please be patient to get an answer to your postings and mails until I'll be back - thx :)) -- Michael Hirmke | Telefon +49 (911) 557999 Georg-Strobel-Strasse 81 | FAX +49 (911) 557664 90489 Nuernberg | E-Mail mh@mike.franken.de | WWW http://minimike.franken.de/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From cjeffare@cs.iupui.edu Sat May 31 11:02:00 1997 From: cjeffare@cs.iupui.edu (Chris Jeffares) Date: Sat, 31 May 1997 11:02:00 -0000 Subject: Using arrays with the Bash Shell Message-ID: Is it possible to use arrays with the Bash Shell, such as this var[$lineno]=$line if ["${var[$lineno]}"= x] then echo found fi lineno=`expr $lineno + 1` This is possible in other Shells, but is it possible in the Bash Shell, if how do I go about it. Thanks again Chris J =+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+= Chris Jeffares Graduate Research Assistant Purdue University at Indianapolis office ==> SL041, (317) 274-3089 email ==> cjeffare@cs.iupui.edu www ==> http://klingon.iupui.edu/~cjeffare/home.html The Nebraska Cornhuskers, Gotta luv 'em Go Big Red!!...Huckin' Fuskers in 1997 "Experience...the name everyone gives to their mistakes." =+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+= - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From *jeffdb@netzone.nospam.com Sat May 31 11:02:00 1997 From: *jeffdb@netzone.nospam.com (Mikey) Date: Sat, 31 May 1997 11:02:00 -0000 Subject: Caps lock problem and/or less can't search? References: <6XqZ1gi4pfB@mike.franken.de> Message-ID: <339056c1.408050@smtp.netzone.com> less, and other programs that use regex need the rx.h file from librx directory in cdksrc.tar.gz, i've appended it here, just copy it to .../H-i386-cygwin32/i386-cygwin32/include/ , and less builds with the included (trivial) diff. No solution (TIKO) for the caps lock problem. On 30 May 1997 21:32:00 +0200, you wrote: >Hi Seymore, > >> I've noticed that when you put the caps lock on, every key applies the >> shift, not just letters. I think it was back in 17.1 as well. >> >> I finally got around to compiling less last night. All went well, just >> had to rename the iprintf(). Compiles easily. But you can't search >> with the / or ? commands. I haven't had a chance to investigate it >> further, but I figured less is a fairly popular program, maybe another >> kind person has already found/fixed it. So, has anyone? > >Only thing, I can say - I came across the same problem and no solution >yet. > >> >> Many thank > >Bye. >Michael. >-- >Ich bin vom 1. Juni bis 3. Juli in Urlaub, Ihr muesst Euch also leider >bis dahin mit einer Antwort auf Euere Postings und Mails gedulden :)) >-- >I'm on vacation from June, 1st to Juli, 3rd - so please be patient to >get an answer to your postings and mails until I'll be back - thx :)) >-- >Michael Hirmke | Telefon +49 (911) 557999 >Georg-Strobel-Strasse 81 | FAX +49 (911) 557664 >90489 Nuernberg | E-Mail mh@mike.franken.de > | WWW http://minimike.franken.de/ >- >For help on using this list (especially unsubscribing), send a message to >"gnu-win32-request@cygnus.com" with one line of text: "help". > -------------- next part -------------- A non-text attachment was scrubbed... Name: less.dif.gz Type: application/x-gzip Size: 396 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: rx.h Type: text/x-c Size: 116190 bytes Desc: not available URL: From mh@mike.franken.de Sat May 31 11:02:00 1997 From: mh@mike.franken.de (Michael Hirmke) Date: Sat, 31 May 1997 11:02:00 -0000 Subject: CMD, case-insensitive file-name completion References: <0006A9FE.1893@po.cle.ab.com> Message-ID: <6Xubc3f$pfB@mike.franken.de> Gunther Ebert [gunther.ebert@ixos-leipzig.de] wrote: > Don Sharp wrote: > > > > I am running NT 3.51 SP5 and I haven't got any completion going after > > making the entry in my registry. What should I have to do apart from > > typing TAB during command line entry? Should it be a binary rather than > > string value? > > > > The completion feature does only work with NT 4.0. Hm, don't know about your NT4, but it works perfectly for mine :)) > > Gunther Bye. Michael. -- Ich bin vom 1. Juni bis 3. Juli in Urlaub, Ihr muesst Euch also leider bis dahin mit einer Antwort auf Euere Postings und Mails gedulden :)) -- I'm on vacation from June, 1st to Juli, 3rd - so please be patient to get an answer to your postings and mails until I'll be back - thx :)) -- Michael Hirmke | Telefon +49 (911) 557999 Georg-Strobel-Strasse 81 | FAX +49 (911) 557664 90489 Nuernberg | E-Mail mh@mike.franken.de | WWW http://minimike.franken.de/ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From alex!@bigfoot.com Sat May 31 14:48:00 1997 From: alex!@bigfoot.com (Hmmmz) Date: Sat, 31 May 1997 14:48:00 -0000 Subject: bash-2.0 Message-ID: <199705312148.OAA17675@mom.hooked.net> I've finally compiled bash 2.0 and it seems to work fine, thanks to Sergey's help, however I'm not sure how it will handle programs trying to dump corefiles. I always seem to run into the problem however that the function setpwent isn't defined anywhere afaik. However according to the man page on setpwent (from a BSD/OS system) "The setpwent() function is identical to setpassent() with an argument of zero." WinNT and *nix rule, so learn from the past: Hiroshima '45, Vietnam '64, Tjernobyl '86, MacOS '96 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From mka@redes.int.com.mx Sat May 31 14:48:00 1997 From: mka@redes.int.com.mx (Michael Anderson) Date: Sat, 31 May 1997 14:48:00 -0000 Subject: Don't use tty code changes I posted yet ... Message-ID: <33907976.7A5818B7@redes.int.com.mx> The code I posted yesterday to implement termio/ioctl features for tty programming should not be used. I had not yet integrated the patch from Sergey which changes some of the same files. I'm working on integrating the two changes so there is no conflict If you hav already used the files I posted, restore the saved copies to get back to the previous state. I'll post the new code as diffs shortly, which will require that Sergey's patch be applied first. Sorry for the confusion. Mike Anderson mka@redes.int.com.mx Guanajuato, GTO, Mexico - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From Don.Sharp@dddandr.octacon.co.uk Sat May 31 14:48:00 1997 From: Don.Sharp@dddandr.octacon.co.uk (Don Sharp) Date: Sat, 31 May 1997 14:48:00 -0000 Subject: CMD, case-insensitive file-name completion References: <0006A9FE.1893@po.cle.ab.com> <338DF38E.5E96@dddandr.octacon.co.uk> <338EB027.30D2@ixos-leipzig.de> Message-ID: <339092AB.6CBE@dddandr.octacon.co.uk> My thanks to everyone who were kind enough to enlighten me regarding filename completion. I'll have to possess myself of patience! Ta Don Sharp Gunther Ebert wrote: > > Don Sharp wrote: > > > > I am running NT 3.51 SP5 and I haven't got any completion going after > > making the entry in my registry. What should I have to do apart from > > typing TAB during command line entry? Should it be a binary rather than > > string value? > > > > The completion feature does only work with NT 4.0. > > Gunther > -- > > Gunther Ebert > iXOS Anwendungs-Software GmbH > Angerstrasse 40-42 > D-04177 Leipzig > > Phone : +49 341 48503-0 > Fax : +49 341 48503-99 > E-mail: mailto:gunther.ebert@ixos-leipzig.de > www : http://www.ixos-leipzig.de > - > For help on using this list (especially unsubscribing), send a message to > "gnu-win32-request@cygnus.com" with one line of text: "help". - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From franl@world.std.com Sat May 31 16:55:00 1997 From: franl@world.std.com (Francis Litterio) Date: Sat, 31 May 1997 16:55:00 -0000 Subject: [b18] Bash does not expand ~ and $HOME the same Message-ID: <199705312355.AA10752@world.std.com> I just noticed this bash oddity (with c:/ mounted as /): bash$ echo $HOME /franl bash $ echo ~ c:/franl Isn't ~ supposed to expand to the current value of the variable HOME? Can anyone see how this might break an existing shell script? -- Francis Litterio PGP Key Fingerprint: franl@world.std.omit-this.com 02 37 DF 6C 66 43 CD 2C http://world.std.omit-this.com/~franl/ 10 C8 B5 8B 57 34 F3 21 Note: email address and WWW URL altered to prevent spam. "They that can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety." -- Ben Franklin, ~1784 - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From da@maigret.cog.brown.edu Sat May 31 16:55:00 1997 From: da@maigret.cog.brown.edu (David Ascher) Date: Sat, 31 May 1997 16:55:00 -0000 Subject: List of supported/unsupported function calls... Message-ID: I think I've asked this before, but if so I didn't get an answer (or I can't remember it). Is there a master list of the calls (from POSIX/BSD/SYSV) and their 'support' level? (e.g. "working always", "not implemented", "implemented and works but special cases exist". If there isn't such a list, is there a technical reason for this not being the case? Or is it just lack of time? The reason I ask is that I keep coming up with projects which I feel should work under gnu-win32, and as I'm working on them someone mentions in passing that, oh, right, function X doesn't work yet (e.g. select() on unnamed pipes), and that blows my project out of the water. Frustrating. I understand that those who know are better off spending time fixing the holes than maintaining a list thereof -- so I volunteer to be the maintainer of that list, if that'll help. Now, to start this list, it'd help to have a list of the target function set -- is POSIX.1 available on the net? Is there a better 'master list' to start from? --david - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From garnett@bsquare.com Sat May 31 18:37:00 1997 From: garnett@bsquare.com (John Garnett) Date: Sat, 31 May 1997 18:37:00 -0000 Subject: problems with patch.exe Message-ID: <01BC6DF1.89575840@paladin.bsquare.com> I'm using GNU-WIN32 b18 on Windows NT 4.0 with service pack 3 installed. I'm trying to use the 'diff' and 'patch' that come with b18 to create a diff and apply it via the patch.exe program. I can get patch.exe to work for simple diffs and small files but it dies for larger files. I saw the same problem with b17. Here is my test case: download http://www.cygnus.com/misc/gnu-win32/index.html to a file. copy index.html index2.html delete line number 17 from index2.html (or any line for that matter) D:\TEMP>copy index.html index2.html 1 file(s) copied. D:\TEMP>vi index2.html D:\TEMP>where diff.exe d:\cygnus\H-i386-cygwin32\bin\diff.exe D:\TEMP>diff -cr index.html index2.html > diff D:\TEMP>type diff *** index.html Sat May 31 18:28:56 1997 --- index2.html Sat May 31 18:29:12 1997 *************** *** 14,20 **** -

The GNU-Win32 Project Page

The GNU-Win32 tools are ports of the popular GNU development tools to Windows NT/95 for the x86 and PowerPC processors. --- 14,19 ---- D:\TEMP>patch < diff Hmm... Looks like a new-style context diff to me... The text leading up to this was: -------------------------- |*** index.html Sat May 31 18:28:56 1997 |--- index2.html Sat May 31 18:29:12 1997 -------------------------- Patching file index.html using Plan B... Hunk #1 failed at 14. 1 out of 1 hunks failed--saving rejects to index.html.rej Hmm... Ignoring the trailing garbage. done D:\TEMP>diff index.html index2.html Binary files index.html and index2.html differ ......... Not only does the patch fail but index.html is garbled and diff considers it a binary file even though diff worked fine on index.html before 'patch' was used. Has anyone else seen this problem? Does someone have a solution? This is quite frustrating. Thanks, John Garnett garnett@bsquare.com Here is some version information and checksum info: D:\TEMP>where sum.exe d:\cygnus\H-i386-cygwin32\bin\sum.exe D:\TEMP>sum d:\cygnus\h-i386-cygwin32\bin\patch.exe 08609 13 D:\TEMP>sum d:\cygnus\h-i386-cygwin32\bin\diff.exe 08623 2 D:\TEMP>diff -v diff - GNU diffutils version 2.7 D:\TEMP>patch -v Patch version 2.1 D:\TEMP>where diff.exe d:\cygnus\H-i386-cygwin32\bin\diff.exe D:\TEMP>where patch.exe d:\cygnus\H-i386-cygwin32\bin\patch.exe - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From dave@lumos.com Sat May 31 20:41:00 1997 From: dave@lumos.com (David Spencer) Date: Sat, 31 May 1997 20:41:00 -0000 Subject: getrusage(RUSAGE_CHILDREN, ...) core dumps Message-ID: <3390EF54.D55A7E50@lumos.com> getrusage(RUSAGE_SELF) seems to work ( at least for finding out cpu and system time) but getrusage(RUSAGE_CHILDREN) "core dumps" on my system. The stack trace is shown below. What I really want is a version of "time" that works - the one included w/ the latest b18 only reports elapsed time correctly. Here's the stack trace when I call getrusage(RUSAGE_CHILDREN): (d:\cdev\cpu-time\dtime.exe 1419) In cygwin_except_handler (d:\cdev\cpu-time\dtime.exe 1419) Exception trapped! (d:\cdev\cpu-time\dtime.exe 1419) exception C0000005 at 10019360 (d:\cdev\cpu-time\dtime.exe 1419) exception: ax 0 bx 0 cx 26160 dx 64 (d:\cdev\cpu-time\dtime.exe 1419) exception: si 5 di 2E73004 bp 240F110 sp 240F0E0 (d:\cdev\cpu-time\dtime.exe 1419) exception is: STATUS_ACCESS_VIOLATION (d:\cdev\cpu-time\dtime.exe 1419) Stack trace: (d:\cdev\cpu-time\dtime.exe 1419) frame 0: sp = 0x240EF14, pc = 0x1000CEC2 (d:\cdev\cpu-time\dtime.exe 1419) frame 1: sp = 0x240EF30, pc = 0x77F94072 (d:\cdev\cpu-time\dtime.exe 1419) frame 2: sp = 0x240EF54, pc = 0x77F88A53 (d:\cdev\cpu-time\dtime.exe 1419) frame 3: sp = 0x240EFE0, pc = 0x77F75E82 (d:\cdev\cpu-time\dtime.exe 1419) frame 4: sp = 0x240F110, pc = 0x401248 (d:\cdev\cpu-time\dtime.exe 1419) frame 5: sp = 0x240F124, pc = 0x4010B8 (d:\cdev\cpu-time\dtime.exe 1419) frame 6: sp = 0x240F1E0, pc = 0x1000C102 (d:\cdev\cpu-time\dtime.exe 1419) frame 7: sp = 0x240FF94, pc = 0x1000C113 (d:\cdev\cpu-time\dtime.exe 1419) frame 8: sp = 0x240FFA0, pc = 0x401501 (d:\cdev\cpu-time\dtime.exe 1419) frame 9: sp = 0x240FFB0, pc = 0x40103B (d:\cdev\cpu-time\dtime.exe 1419) frame 10: sp = 0x240FFC0, pc = 0x77F1AFC1 (d:\cdev\cpu-time\dtime.exe 1419) frame 11: sp = 0x240FFF0, pc = 0x0 (d:\cdev\cpu-time\dtime.exe 1419) End of stack trace bash$ - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From sos@prospect.com.ru Sat May 31 23:18:00 1997 From: sos@prospect.com.ru (Sergey Okhapkin) Date: Sat, 31 May 1997 23:18:00 -0000 Subject: List of supported/unsupported function calls... Message-ID: <01BC6E75.5997A270@sos> David Ascher wrote: > Is there a master list of the calls (from POSIX/BSD/SYSV) and their > 'support' level? (e.g. "working always", "not implemented", "implemented > and works but special cases exist". If there isn't such a list, is there The whole list of cygwin's calls is in cygwin.din :-) Most of unsupported calls are located in stubs.cc. Select() on pipes does not works, itimer stuff supports ITIMER_REAL only. -- Sergey Okhapkin Moscow, Russia Looking for a job - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help". From fjh@cs.mu.OZ.AU Sat May 31 23:20:00 1997 From: fjh@cs.mu.OZ.AU (Fergus Henderson) Date: Sat, 31 May 1997 23:20:00 -0000 Subject: List of supported/unsupported function calls... References: Message-ID: <199706010619.QAA10158@mundook.cs.mu.OZ.AU> David Ascher, you wrote: > > I understand that those who know are better off spending time fixing the > holes than maintaining a list thereof -- so I volunteer to be the > maintainer of that list, if that'll help. Now, to start this list, it'd > help to have a list of the target function set -- is POSIX.1 available on > the net? Is there a better 'master list' to start from? The following was posted on comp.std.c a few years ago. ANSI C and POSIX.1 Symbol/Header Cross Reference ================================================ The following list provides a cross reference table between ANSI/ISO Standard C and POSIX P1003.1 symbols The table includes the following fields (in order): Std - which standard defines the symbol ANSI/ISO Standard C, or POSIX P1003.1 Primary Header - the header where the symbol is declared or defined. Symbols that are defined in multiple headers have multiple entries in the cross-reference table. Symbol Type - macro, function, variable, struct, or typedef. Note that any function can also (optionally) exist as a macro, and that some symbols can be either macros or functions depending on the implemention. Symbol - actual identifier. Functions and function-like macros are shown with parens Prerequisite Headers - other headers that should be included before the primary header in code using this function Primary Symbol Prerequisite Std. Header Type Symbol Headers ==== ======= ====== ====== ============ ANSI macro NDEBUG ANSI function isalnum() POSIX function readdir() POSIX struct dirent POSIX typedef DIR POSIX macro EINTR ANSI macro NDEBUG ANSI macro assert() ANSI function isalnum() ANSI function isalpha() ANSI function iscntrl() ANSI function isdigit() ANSI function isgraph() ANSI function islower() ANSI function isprint() ANSI function ispunct() ANSI function isspace() ANSI function isupper() ANSI function isxdigit() ANSI function tolower() ANSI function toupper() POSIX function closedir() POSIX function opendir() POSIX function readdir() POSIX function rewinddir() POSIX struct dirent POSIX typedef DIR POSIX macro E2BIG POSIX macro EACCES POSIX macro EAGAIN POSIX macro EBADF POSIX macro EBUSY POSIX macro ECHILD POSIX macro EDEADLK ANSI macro EDOM POSIX macro EEXIST POSIX macro EFAULT POSIX macro EFBIG POSIX macro EINTR POSIX macro EINVAL POSIX macro EIO POSIX macro EISDIR POSIX macro EMFILE POSIX macro EMLINK POSIX macro ENAMETOOLONG POSIX macro ENFILE POSIX macro ENODEV POSIX macro ENOENT POSIX macro ENOEXEC POSIX macro ENOLCK POSIX macro ENOMEM POSIX macro ENOSPC POSIX macro ENOSYS POSIX macro ENOTDIR POSIX macro ENOTEMPTY POSIX macro ENOTTY POSIX macro ENXIO POSIX macro EPERM POSIX macro EPIPE ANSI macro ERANGE POSIX macro EROFS POSIX macro ESPIPE POSIX macro ESRCH POSIX macro EXDEV ANSI variable errno POSIX function creat() POSIX function fcntl() POSIX function open() POSIX macro FD_CLOEXEC POSIX macro F_DUPFD POSIX macro F_GETFD POSIX macro F_GETFL POSIX macro F_GETLK POSIX macro F_RDLCK POSIX macro F_SETFD POSIX macro F_SETFL POSIX macro F_SETLK POSIX macro F_SETLKW POSIX macro F_UNLCK POSIX macro F_WRLCK POSIX macro O_ACCMODE POSIX macro O_APPEND POSIX macro O_CREAT POSIX macro O_EXCL POSIX macro O_NOCTTY POSIX macro O_NONBLOCK POSIX macro O_RDONLY POSIX macro O_RDWR POSIX macro O_TRUNC POSIX macro O_WRONLY POSIX struct flock ANSI macro DBL_DIG ANSI macro DBL_EPSILON ANSI macro DBL_MANT_DIG ANSI macro DBL_MAX ANSI macro DBL_MAX_10_EXP ANSI macro DBL_MAX_EXP ANSI macro DBL_MIN ANSI macro DBL_MIN_10_EXP ANSI macro DBL_MIN_EXP ANSI macro FLT_DIG ANSI macro FLT_EPSILON ANSI macro FLT_MANT_DIG ANSI macro FLT_MAX ANSI macro FLT_MAX_10_EXP ANSI macro FLT_MAX_EXP ANSI macro FLT_MIN ANSI macro FLT_MIN_10_EXP ANSI macro FLT_MIN_EXP ANSI macro FLT_RADIX ANSI macro FLT_ROUNDS ANSI macro LDBL_DIG ANSI macro LDBL_EPSILON ANSI macro LDBL_MANT_DIG ANSI macro LDBL_MAX ANSI macro LDBL_MAX_10_EXP ANSI macro LDBL_MAX_EXP ANSI macro LDBL_MIN ANSI macro LDBL_MIN_10_EXP ANSI macro LDBL_MIN_EXP POSIX function getgrgid() POSIX function getgrnam() POSIX struct group POSIX macro ARG_MAX ANSI macro CHAR_BIT ANSI macro CHAR_MAX ANSI macro CHAR_MIN POSIX macro CHILD_MAX ANSI macro INT_MAX ANSI macro INT_MIN POSIX macro LINK_MAX ANSI macro LONG_MAX ANSI macro LONG_MIN POSIX macro MAX_CANON POSIX macro MAX_INPUT ANSI macro MB_LEN_MAX POSIX macro NAME_MAX POSIX macro NGROUPS_MAX POSIX macro OPEN_MAX POSIX macro PATH_MAX POSIX macro PIPE_BUF ANSI macro SCHAR_MAX ANSI macro SCHAR_MIN ANSI macro SHRT_MAX ANSI macro SHRT_MIN POSIX macro SSIZE_MAX POSIX macro STREAM_MAX POSIX macro TZNAME_MAX ANSI macro UCHAR_MAX ANSI macro UINT_MAX ANSI macro ULONG_MAX ANSI macro USHRT_MAX POSIX macro _POSIX_ARG_MAX POSIX macro _POSIX_CHILD_MAX POSIX macro _POSIX_LINK_MAX POSIX macro _POSIX_MAX_CANON POSIX macro _POSIX_MAX_INPUT POSIX macro _POSIX_NAME_MAX POSIX macro _POSIX_NGROUPS_MAX POSIX macro _POSIX_OPEN_MAX POSIX macro _POSIX_PATH_MAX POSIX macro _POSIX_PIPE_BUF POSIX macro _POSIX_SSIZE_MAX POSIX macro _POSIX_STREAM_MAX POSIX macro _POSIX_TZNAME_MAX ANSI function localeconv() ANSI function setlocale() ANSI macro LC_ALL ANSI macro LC_COLLATE ANSI macro LC_CTYPE ANSI macro LC_MONETARY ANSI macro LC_NUMERIC ANSI macro LC_TIME ANSI macro NULL ANSI struct lconv ANSI function acos() ANSI function asin() ANSI function atan() ANSI function atan2() ANSI function ceil() ANSI function cos() ANSI function cosh() ANSI function exp() ANSI function fabs() ANSI function floor() ANSI function fmod() ANSI function frexp() ANSI function ldexp() ANSI function log() ANSI function log10() ANSI function modf() ANSI function pow() ANSI function sin() ANSI function sinh() ANSI function sqrt() ANSI function tan() ANSI function tanh() ANSI macro HUGE_VAL POSIX function getpwnam() POSIX function getpwuid() POSIX struct passwd ANSI function longjmp() ANSI function setjmp() POSIX function siglongjmp() POSIX function sigsetjmp() ANSI typedef jmp_buf POSIX typedef sigjmp_buf POSIX function kill() ANSI function raise() POSIX function sigaction() POSIX function sigaddset() POSIX function sigdelset() POSIX function sigemptyset() POSIX function sigfillset() POSIX function sigismember() ANSI function signal() POSIX function sigpending() POSIX function sigprocmask() POSIX function sugsuspend() POSIX macro SA_NOCLDSTOP POSIX macro SIGABRT ANSI macro SIGABRT POSIX macro SIGALRM POSIX macro SIGCHLD POSIX macro SIGCONT POSIX macro SIGFPE ANSI macro SIGFPE POSIX macro SIGHUP POSIX macro SIGILL ANSI macro SIGILL POSIX macro SIGINT ANSI macro SIGINT POSIX macro SIGKILL POSIX macro SIGPIPE POSIX macro SIGQUIT POSIX macro SIGSEGV ANSI macro SIGSEGV POSIX macro SIGSTOP POSIX macro SIGTERM ANSI macro SIGTERM POSIX macro SIGTSTP POSIX macro SIGTTIN POSIX macro SIGTTOU POSIX macro SIGUSR1 POSIX macro SIGUSR2 POSIX macro SIG_BLOCK ANSI macro SIG_DFL POSIX macro SIG_DLF POSIX macro SIG_ERR ANSI macro SIG_ERR ANSI macro SIG_IGN POSIX macro SIG_SETMASK POSIX macro SIG_UNBLOCK POSIX struct sigaction ANSI typedef sig_atomic_t POSIX typeset sigset_t ANSI macro va_arg() ANSI macro va_end() ANSI macro va_start() ANSI typedef va_list ANSI macro NULL ANSI macro offsetof() ANSI typedef ptrdiff_t ANSI typedef size_t ANSI typedef wchar_t ANSI function clearerr() ANSI function fclose() POSIX function fdopen() ANSI function feof() ANSI function ferror() ANSI function fflush() ANSI function fgetc() ANSI function fgetpos() ANSI function fgets() POSIX function fileno() ANSI function fopen() ANSI function fprintf() ANSI function fputc() ANSI function fputs() ANSI function fread() ANSI function freopen() ANSI function fscanf() ANSI function fseek() ANSI function fsetpos() ANSI function ftell() ANSI function fwrite() ANSI function getc() ANSI function getchar() ANSI function gets() ANSI function perror() ANSI function printf() ANSI function putc() ANSI function putchar() ANSI function puts() ANSI function remove() ANSI function rename() ANSI function rewind() ANSI function scanf() ANSI function setbuf() ANSI function setvbuf() ANSI function sprintf() ANSI function sscanf() ANSI function tmpfile() ANSI function tmpnam() ANSI function ungetc() ANSI function vfprintf() ANSI function vprintf() ANSI function vsprintf() ANSI macro BUFSIZ ANSI macro EOF ANSI macro FILENAME_MAX ANSI macro FOPEN_MAX POSIX macro L_ctermid ANSI macro L_tmpnam ANSI macro NULL ANSI macro SEEK_CUR ANSI macro SEEK_END ANSI macro SEEK_SET POSIX macro STREAM_MAX ANSI macro TMP_MAX ANSI macro _IOFBF ANSI macro _IOLBF ANSI macro _IONBF ANSI typedef FILE ANSI typedef fpos_t ANSI typedef size_t ANSI variable stderr ANSI variable stdin ANSI variable stdout ANSI function abort() ANSI function abs() ANSI function atexit() ANSI function atof() ANSI function atoi() ANSI function atol() ANSI function bsearch() ANSI function calloc() ANSI function div() ANSI function exit() ANSI function free() ANSI function getenv() ANSI function labs() ANSI function ldiv() ANSI function malloc() ANSI function mblen() ANSI function mbstowcs() ANSI function mbtowc() ANSI function qsort() ANSI function rand() ANSI function realloc() ANSI function srand() ANSI function strtod() ANSI function strtol() ANSI function strtoul() ANSI function system() ANSI function wcstombs() ANSI function wctomb() ANSI macro EXIT_FAILURE ANSI macro EXIT_SUCCESS ANSI macro MB_CUR_MAX ANSI macro NULL ANSI macro RAND_MAX ANSI typedef div_t ANSI typedef ldiv_t ANSI typedef size_t ANSI typedef wchar_t ANSI function memchr() ANSI function memcmp() ANSI function memcpy() ANSI function memmove() ANSI function memset() ANSI function strcat() ANSI function strchr() ANSI function strcmp() ANSI function strcoll() ANSI function strcpy() ANSI function strcspn() ANSI function strerror() ANSI function strlen() ANSI function strncat() ANSI function strncmp() ANSI function strncpy() ANSI function strpbrk() ANSI function strrchr() ANSI function strspn() ANSI function strstr() ANSI function strtok() ANSI function strxfrm() ANSI macro NULL ANSI typedef size_t POSIX function chmod() POSIX function fstat() POSIX function mkdir() POSIX function mkfifo() POSIX function stat() POSIX function umask() POSIX macro S_IRGRP POSIX macro S_IROTH POSIX macro S_IRUSR POSIX macro S_IRWXG POSIX macro S_IRWXO POSIX macro S_IRWXU POSIX macro S_ISBLK() POSIX macro S_ISCHR() POSIX macro S_ISDIR() POSIX macro S_ISFIFO() POSIX macro S_ISGID POSIX macro S_ISREG() POSIX macro S_ISUID POSIX macro S_IWGRP POSIX macro S_IWOTH POSIX macro S_IWUSR POSIX macro S_IXGRP POSIX macro S_IXOTH POSIX macro S_IXUSR POSIX struct stat POSIX function times() POSIX struct tms POSIX typedef clock_t POSIX typedef dev_t POSIX typedef gid_t POSIX typedef ino_t POSIX typedef mode_t POSIX typedef nlink_t POSIX typedef off_t POSIX typedef pid_t POSIX typedef size_t POSIX typedef ssize_t POSIX typedef uid_t POSIX function uname() POSIX struct utsname POSIX function wait() POSIX function waitpid() POSIX macro WEXITSTATUS POSIX macro WIFEXITED POSIX macro WIFSIGNALED POSIX macro WIFSTOPPED POSIX macro WNOHANG POSIX macro WSTOPSIG POSIX macro WTERMSIG POSIX macro WUNTRACED POSIX function cfgetispeed() POSIX function cfgetospeed() POSIX function cfsetispeed() POSIX function cfsetospeed() POSIX function tcdrain() POSIX function tcflow() POSIX function tcflush() POSIX function tcgetattr() POSIX function tcsendbreak() POSIX function tcsetattr() POSIX macro B0 POSIX macro B110 POSIX macro B1200 POSIX macro B134 POSIX macro B150 POSIX macro B1800 POSIX macro B19200 POSIX macro B200 POSIX macro B2400 POSIX macro B300 POSIX macro B38400 POSIX macro B4800 POSIX macro B50 POSIX macro B600 POSIX macro B75 POSIX macro B9600 POSIX macro BRKINT POSIX macro CLOCAL POSIX macro CREAD POSIX macro CS5 POSIX macro CS6 POSIX macro CS7 POSIX macro CS8 POSIX macro CSIZE POSIX macro CSTOPB POSIX macro ECHO POSIX macro ECHOE POSIX macro ECHOK POSIX macro ECHONL POSIX macro HUPCL POSIX macro ICANON POSIX macro ICRNL POSIX macro IEXTEN POSIX macro IGNBRK POSIX macro IGNCR POSIX macro IGNPAR POSIX macro INLCR POSIX macro INPCK POSIX macro ISIG POSIX macro ISTRIP POSIX macro IXOFF POSIX macro IXON POSIX macro NCCS POSIX macro NOFLSH POSIX macro OPOST POSIX macro PARENB POSIX macro PARMRK POSIX macro PARODD POSIX macro TCIFLUSH POSIX macro TCIOFF POSIX macro TCIOFLUSH POSIX macro TCION POSIX macro TCOFLUSH POSIX macro TCOOFF POSIX macro TCOON POSIX macro TCSADRAIN POSIX macro TCSAFLUSH POSIX macro TCSANOW POSIX macro TOSTOP POSIX macro VEOF POSIX macro VEOL POSIX macro VERASE POSIX macro VINTR POSIX macro VKILL POSIX macro VMIN POSIX macro VQUIT POSIX macro VSTART POSIX macro VSTOP POSIX macro VSUSP POSIX macro VTIME POSIX struct termios POSIX typedef cc_t POSIX typedef speed_t POSIX typedef tcflag_t ANSI function asctime() ANSI function clock() ANSI function ctime() ANSI function difftime() ANSI function gmtime() ANSI function localtime() ANSI function mktime() ANSI function strftime() ANSI function time() POSIX function tzset() ANSI macro CLOCKS_PER_SEC POSIX macro CLOCK_TCK ANSI macro NULL ANSI struct tm ANSI typedef clock_t ANSI typedef size_t ANSI typedef time_t ANSI variable tzname POSIX function _exit() POSIX function access() POSIX function alarm() POSIX function chdir() POSIX function chown() POSIX function close() POSIX function ctermid() POSIX function dup() POSIX function dup2() POSIX function execl() POSIX function execle() POSIX function execlp() POSIX function execv() POSIX function execve() POSIX function execvp() POSIX function fork() POSIX function fpathconf() POSIX function getcwd() POSIX function getegid() POSIX function geteuid() POSIX function getgid() POSIX function getgroups() POSIX function getlogin() POSIX function getpgrp() POSIX function getpid() POSIX function getppid() POSIX function getuid() POSIX function isatty() POSIX function link() POSIX function lseek() POSIX function pathconf() POSIX function pause() POSIX function pipe() POSIX function read() POSIX function rmdir() POSIX function setgid() POSIX function setpgid() POSIX function setsid() POSIX function setuid() POSIX function sleep() POSIX function sysconf() POSIX function tcgetpgrp() POSIX function tcsetpgrp() POSIX function ttyname() POSIX function unlink() POSIX function write() POSIX macro F_OK POSIX macro NULL POSIX macro R_OK POSIX macro SEEK_CUR POSIX macro SEEK_END POSIX macro SEEK_SET POSIX macro STDERR_FILENO POSIX macro STDIN_FILENO POSIX macro STDOUT_FILENO POSIX macro W_OK POSIX macro X_OK POSIX macro _PC_CHOWN_RESTRICTED POSIX macro _PC_LINK_MAX POSIX macro _PC_MAX_CANON POSIX macro _PC_MAX_INPUT POSIX macro _PC_NAME_MAX POSIX macro _PC_NO_TRUNC POSIX macro _PC_PATH_MAX POSIX macro _PC_PIPE_BUF POSIX macro _PC_VDISABLE POSIX macro _POSIX_CHOWN_RESTRICTED POSIX macro _POSIX_JOB_CONTROL POSIX macro _POSIX_NO_TRUNC POSIX macro _POSIX_SAVED_IDS POSIX macro _POSIX_VDISABLE POSIX macro _POSIX_VERSION POSIX macro _SC_ARG_MAX POSIX macro _SC_CHILD_MAX POSIX macro _SC_CLK_TCK POSIX macro _SC_JOB_CONTROL POSIX macro _SC_NGROUPS_MAX POSIX macro _SC_OPEN_MAX POSIX macro _SC_SAVED_IDS POSIX macro _SC_STREAM_MAX POSIX macro _SC_TZNAME_MAX POSIX macro _SC_VERSION POSIX typedef size_t POSIX typedef ssize_t POSIX function utime() POSIX struct utimbuf -- Fergus Henderson | "I have always known that the pursuit WWW: < http://www.cs.mu.oz.au/~fjh > | of excellence is a lethal habit" PGP: finger fjh@128.250.37.3 | -- the last words of T. S. Garp. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request@cygnus.com" with one line of text: "help".