Re: EMACS problems with Cygwin 1.5.7.1

[Thomas L Roche <tlroche@us.ibm.com>]

On Tue, 17 Feb 2004 andy<dot>glew<at>amd<dot>com wrote:
>> I have been seeing the same frequent crashes of EMACS on CYGWIN
>> 1.5.7.1 that "Ehud Karni" <ehud at unix dot mvs dot co dot il> and
>> "Charles Plager" <cplager+cygwin at physics dot ucla dot edu> have
>> already reported to this list.

>> Grepping further:
>>    Thomas L Roche <tlroche <AT> us.ibm.com>
>>    Dr.. Volkmer Zell

Igor Pechtchanski Tue, 17 Feb 2004 13:39:26 -0500 (EST)
> FWIW, "Ehud Karni" <ehud at unix dot mvs dot co dot il> reported in
> <http://cygwin.com/ml/cygwin/2004-02/msg00822.html> that the latest
> snapshot *does* fix his problem.

For another datapoint: emacs life _has_ improved with the 2004021{4, 5,
6} snapshots, so I advise installing

http://cygwin.com/snapshots/cygwin1-20040216.dll.bz2

(or whatever is the latest snapshot @

http://cygwin.com/snapshots/

when you get this. Note however that snapshots do not monotonically
improve--20040213 was nasty--so if you get a snapshot > 0216 that
seems bad, try dropping back to 0216, which Works For Me (YMMV).)

That being said, emacs is still not nearly as stable now as it was on
1.5.5-1--in fact, I got an exit 128 (haven't seen one of those before,
usually it's just SEGVs) while composing this reply. So while I don't
believe Plager's response (dropping back to 1.5.5-1) was optimal, I
can see why one might choose that route. Instead I hope to debug emacs
after the unzip problem (which more directly harms my productivity) is
fixed ... and use that to {test, refine} an expanded debugging-Cygwin
howto, which hopefully will encourage others to debug it as well.
(Though it's not rocket science, it would be a lot easier if better
documented.)