please test ptys and ttys in the latest snapshot

[Volker Quetschke]

Christopher Faylor wrote:
> I've rewritten some of the pty/tty code in the latest snapshot to work
> around problems that Dave Korn raised.  Thanks to Dave for finding the
> problem and for his patch which would have made things work better.
> 
> If you read the cygwin-patches mailing list, you will see that, after
> thinking about Dave's patch, I realized that Cygwin was fairly badly
> broken wrt handling of ptys.  So, I reimplemented the pty/tty master
> component similar to a rewrite of the pty/tty slave stuff that was done
> a while ago.
> 
> What all of this means, of course, is that I probably broke something
> in a subtle way and, so, I'd like to hear if the current snapshot
> works for you.  Please respond to this thread, if so.

Ok, starting from snapshot from 20060602 I get problems. This only happens
on my Opteron system (Windows XP SP2), I'll attach a cygcheck created with
snapshot 20060601. On another note I don't see any of the problems reported
below on a different system (Celeron M Notebook).

When using 20060602 and I start my rxvt shortcut
  (c:\cygwin\bin\rxvt.exe -sl 1000 -tn xterm -e /bin/bash --login -i)
I got (once) an "Windows - Application Error" alert box that says:

The instruction at "0x6100365f" referenced memory at "0x0067ca5c". The
memory could not be "written". ...

Later the console window (not the rxvt window) just flashes briefly
and no trace of the program (Windows Task Manager) is left.

Another bit of information: I open a cmd window and cd into c:\cygwin\bin:

When starting "cygcheck.exe -svr" from there an alert box with:
"cygcheck.exe has encountered a problem and needs to close. ..."
problem report box comes up.

Also cmd.exe in c:\cygwin\bin: I start "uname -a" and the prompt just returns,
no output, no error. -- Ooops, I have to correct myself. Several tries late
uname hangs.

The same strange behaviour also happens with the 20060603 snapshot.
I even could reproduce the same

The instruction at "0x6100365f" referenced memory at "0x0067ca5c". The
memory could not be "written". ...

Old snapshots, (as far as they were used) up to 20060601 worked fine
on that machine. This is in fact my OOo building machine, so cygwin was
stable enough to build OOo on that machine so far. And it is moderately
fast, a 2 GHz opteron (old 940 pin socket with 2GB dual channel memory).

As said before, no problems on my "slow" notebook.

    Volker

-- 
PGP/GPG key  (ID: 0x9F8A785D)  available  from  wwwkeys.de.pgp.net
key-fingerprint 550D F17E B082 A3E9 F913  9E53 3D35 C9BA 9F8A 785D
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: cygcheck.log
URL: <http://cygwin.com/pipermail/cygwin/attachments/20060603/ea8ba273/attachment.log>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 254 bytes
Desc: OpenPGP digital signature
URL: <http://cygwin.com/pipermail/cygwin/attachments/20060603/ea8ba273/attachment.sig>