emacs-pretest-bug
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: seg fault on mouse move, run on Linux, display on Cygwin


From: garynot
Subject: Re: seg fault on mouse move, run on Linux, display on Cygwin
Date: Wed, 25 Aug 4 22:41:20 PDT

> > Symptoms:
> >
> > I'm running 21.3.50.1 taken directly from savannah.  This version of 
> > Emacs
> > seg faults.  The version that ships with SuSE is 21.3.1 and it gets a
> > BadWindow (invalid Window parameter) X protocol error rather than a 
> > seg fault.
> 
> There was a bug in the CVS version, you should now get a BadWindow there
> also.

Right.  After updating tonight the BadWindow is back.
I've been trying to run this under xscope but I haven't been able to get
xscope to work on my SuSE machine.  No idea why.  It works fine on Solaris
and HPUX and others.

> > I'm running SuSE 9.1 Linux on my desktop.  My laptop is running 
> > Windows XP.
> > I have Cygwin (www.cygwin.org) installed on the laptop.  I run Cygwin 
> > and
> > start up the XFree86 X display.  Then I run Emacs on the desktop 
> > machine
> > with $DISPLAY set to point to my laptop.
> 
> This crash is strange, the BadWindow is the default root window, which
> should never be bad.  Are you running the Cygwin display in some 
> root-less
> mode?  Not that it should make any difference.

I'm running it in the "traditional" mode which means I have one big window
that is the X server's root window.  I'm using twm which means I don't have
any virtual roots.

> It is also possible that there is some bad interaction with SuSE 
> libraries
> that corrupts memory.  There are other reports that seems to indicate
> that.

I've also got cygwin on another desktop machine.  Emacs doesn't get the
BadWindow when talking to that machine.  The desktop cygwin machine has the
motherboard Intel graphics chip and the laptop cygwin machine has a Mobile
Radeon 9000 video chip set.  So they are using the same base X server but
different engines down at the bottommost draw-the-graphics level.  Could be
an XFree86 bug I suppose.  If I can ever get xscope to work I might be able
to tell at least whether or not the protocol is in fact bad or whether the
BadWindow error is bogus.

   Gary




reply via email to

[Prev in Thread] Current Thread [Next in Thread]