This is the mail archive of the
cygwin-developers@cygwin.com
mailing list for the Cygwin project.
Re: YANDQ
- To: cygwin-developers at cygwin dot com
- Subject: Re: YANDQ
- From: Christopher Faylor <cgf at redhat dot com>
- Date: Tue, 20 Mar 2001 09:14:40 -0500
- References: <006501c0b120$094af0f0$0200a8c0@lifelesswks>
- Reply-To: cygwin-developers at cygwin dot com
On Tue, Mar 20, 2001 at 08:27:45PM +1100, Robert Collins wrote:
>From: "Robert Collins" <robert.collins@itdomain.com.au>
>To: <cygwin-developers@cygwin.com>
>Subject: YANDQ
>Date: Tue, 20 Mar 2001 20:27:45 +1100
>
>YetAnotherNewDeveloperQuestion (couldn't see anything obvious in the
>archives either)
>
>The posix specs I'm reading http://www.opengroup.org/onlinepubs/7908799/
>indicate that most of the types I'm working with (sem_t pthread_*_t)
>should be defined in <sys/types.h> with only the prototypes and #defines
>in the respective header files.
>
>Are these specs way of base? If not, how do I go about altering
>sys/types.h (which I note appears to be part of newlib).
You submit a patch to the newlib mailing list which is conditionalized
#ifdef __CYGWIN__
#endif
I can approve patches like this as long as they don't make changes to
any non-cygwin part of newlib.
cgf
>The question arises because I have an external typedef which is
>different from the internal typedef. While I could type cast everything
>to make it match up (class pthread ** to void ** for instance) it reads
>a lot more easily (which helps debugging) with no casting involved.
>
>FYI the external typedefs are of the form
>typedef void * foo_t
>and the internal ones
>typedef class foo * foo_t.
>
>At the moment I've just duplicated the #defines across the files, but
>I'd really rather have the types separated out from the prototypes and
>#defines...
>
>Rob
--
cgf@cygnus.com Red Hat, Inc.
http://sources.redhat.com/ http://www.redhat.com/