This is the mail archive of the
mailing list for the Cygwin project.
RE: cygwin-xemacs sometimes hangs in compile with latest cygwin
- From: "Andy Piper" <andy at xemacs dot org>
- To: <lipp at Danet dot de>,<cygwin at cygwin dot com>,<xemacs-beta at xemacs dot org>
- Date: Tue, 20 Aug 2002 09:28:48 -0700
- Subject: RE: cygwin-xemacs sometimes hangs in compile with latest cygwin
> I'm using the latest cygwin and the latest xemacs-cygwin32.
> One time out of about five, when I start a compilation in cygwin-xemacs
> (M-x compile) the subprocess is started, but no output from the
> subprocess appears in the compilation buffer (the subprocess is a shell
> script (ant) that starts a dos executable (java)). I observe the same
> behaviour sometimes when doign a "M-x grep". The subprocess runs,
> though very slowly because xemacs uses nearly 100% of the CPU time
> (polling for results?). This behaviour is experienced by all collegues
> that use xemacs-cygwin as well.
> The bug appeared when I upgraded cygwin in mid-December 2001. I send
> this bug report then, but never got a response. Not being experienced
> with the windows API, I cannot debug this further (something I usually
> try before reporting bugs). However, obviously this bug is "genuine" as
> it does not disappear as bug fixes are made to cygwin/xemacs. So I
> decided to report it again.
I rather think this is a cygwin bug. You would need to check with the
cygwin folks how to confirm whether this is the case or not. On a similar
note it used to be the case that you had to have CYGWIN=tty set in your
environment to get proper sub-process control. However, I have found that
setting CYGWIN=tty now breaks a lot of other things (e.g. C-c no longer
and also that sub-process support seems to work pretty well without it.
So my guess is that CYGWIN=tty is deprecated in some way. Having said that
you might want to play around with switching it on/off to see if that makes
I am not on the cygwin mailing list so please respond to me directly.
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html