Crash in xserv

[Joe Baptista <baptista@dot-god.com>]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: TEXT/PLAIN; charset=X-UNKNOWN, Size: 3532 bytes --]


I hope someday I can actually use this cygwin-xfree server.

I have tried several of the different versions and they always crash when
I attempt to run a display like

netscape -display hostname:0.0

or

xpdf -display hostname:0.0

will crash my xserver

but if i run

xterm -display hostname:0.0

it runs - no problem.  seems the problem is with graphic intensive
xsessions.  Here is the windows error report.

XWIN caused an invalid page fault in
module CYGWIN1.DLL at 0167:61093b43.
Registers:
EAX=00000000 CS=0167 EIP=61093b43 EFLGS=00010246
EBX=00001db8 SS=016f ESP=0181ee10 EBP=0181ee28
ECX=1016e168 DS=016f ESI=1016e170 FS=5c8f
EDX=00333366 ES=016f EDI=00000000 GS=0000
Bytes at CS:EIP:
89 50 0c 01 fb 89 42 08 c7 41 0c c4 83 12 61 a1
Stack dump:
00000000 bff813f8 00000000 1016e170 1016e170 1038a220 0181ee58 6103eb65
181283b0 00310000 61128390 00000001 00000032 00000001 0181ee88 00000018

I reported this problem back on Dec 13 - which was caused by a prior
version of xserv

http://cygwin.com/ml/cygwin-xfree/2003-12/msg00240.html

you'll notice it's the same error - so it would be nice if the next
version of xserv can address this problem.

thanks
joe baptista

On Tue, 13 Jan 2004, [ISO-8859-1] Øyvind Harboe wrote:

> >What is "crashing"? XWin.exe? Does your entire X11 session bomb? Or, is just
> > Evolution crashing? If Evolution is crashing, then it is likely due to bugs
> > in Evolution; granted, some of these bugs may be brought out by XWin.exe,
> >but they would still be bugs nonetheless.
>
> Some thread in the XWin.exe process causes an access violation.
>
> >>- I can use the Evolution interface after the "Do you want to debug" requester has popped
> >>up. I interpret this as some "non-essential thread" having been stopped while the rest of
> >>XWin threads continue "unperturbed"
> >
> >That message is coming from Evolution, not from XWin.exe. If one of XWin.exe's
> > thread's crashed, that thread would just exit silently (usually with a message
> >saying so in XWin.log).
>
> The message I'm referring to is the Windows Dr Watson message that allows me to launch
> the Visual C++ debugger which will then show me the instruction that caused the access violation
> in the XWin.exe process.
>
> Depending on the configuration of the system, this requester could be anything. I presume it is even
> possible to launch GDB at this point.
>
> >>- Copy & paste works after the crash(from Evolution to Windows)
> >
> >So, XWin.exe is not crashing,
>
> Crashing as you pointed out is a terribly vague description, but I would have to insist
> that XWin.exe is indeed crashing. :-)
>
> I interpret the situation as a thread within the XWin.exe process having been halted after causing
> an access violation. The remaining threads in XWin.exe are allowed to continue.
> The requester that is put up is Windows allowing me to debug it.
>
> If I press "no I don't want to debug", then Xwin.exe is immediately terminated.
>
> >Are you sure XWin.exe did this?
>
> Yes, I've even looked at the XWin.exe in the Visual C++ debugger, but without symbol information
> its kinda pointless.
>
> >This sounds like it is happening in Evolution. You
> > are running Evolution on a separate Linux/*nix box, right?
>
> Yes.
>
> > You haven't done something silly like compiled it for Cygwin, have you?
>
> Nope.
>
> I don't have enough knowledge about xfree86 to do anything particularly harmfull :-)
>
> Øyvind
>
>