Re: Mixing X and GNUstep

[Raffael Herzog]

On 2003-07-11 19:43:11 +0200 Adam Fedor <fedor@doc.com> wrote:

> Well, the last log message shows that the backend is getting an event that we 
> don't handle. Perhaps you could debug it and find out which event it is (or 
> make the log statement more informative ---see 
> back/Source/x11/XGServerEvent.m: 1269)

New message:

2003-07-12 11:34:11.453 GSXine[11991] Received an untrapped event 94

Looks like event 30 (SelectionRequest) with some bit 6 masked in... I get it 
each time, the window arrangement changes. It doesn't seem like this really 
has something to do with the segfault, for now, I think I can savely ignore 
this message.


> Then continue debugging and get a backtrace after the segfault...

There seems to be a bug in gdb (or whatever): The post mortem backtrace is 
obviously wrong. I'm now trying to get a correct backtrace by running it from 
within gdb. Dooh, this takes *ages*... Because it took so long, the 
application wasn't active any more when I got there - and see, no crash until 
I activated it. This time, the backtrace looks realistic (reproduced twice):

#0  0x405ed799 in objc_msg_lookup ()
  from /usr/GNUstep/System/Library/Libraries/ix86/linux-gnu/libobjc.so.1
#1  0x402109d2 in _c_NSDocumentController__isDocumentBasedApplication ()
  from 
/usr/GNUstep/System/Library/Libraries/ix86/linux-gnu/gnu-gnu-gnu/libgnustep-gui.so.0
#2  0x401cac7f in _i_NSApplication__targetForAction_ ()
  from 
/usr/GNUstep/System/Library/Libraries/ix86/linux-gnu/gnu-gnu-gnu/libgnustep-gui.so.0
#3  0x402421a3 in _i_NSMenu__update ()
  from 
/usr/GNUstep/System/Library/Libraries/ix86/linux-gnu/gnu-gnu-gnu/libgnustep-gui.so.0
#4  0x4024226b in _i_NSMenu__update ()
  from 
/usr/GNUstep/System/Library/Libraries/ix86/linux-gnu/gnu-gnu-gnu/libgnustep-gui.so.0
#5  0x401c9371 in _i_NSApplication__run ()
  from 
/usr/GNUstep/System/Library/Libraries/ix86/linux-gnu/gnu-gnu-gnu/libgnustep-gui.so.0
#6  0x401c1617 in NSApplicationMain ()
  from 
/usr/GNUstep/System/Library/Libraries/ix86/linux-gnu/gnu-gnu-gnu/libgnustep-gui.so.0
#7  0x08049468 in main (argc=1, argv=0xbffff184) at main.m:111

Then I tried to reproduce it to see whether the backtrace is two times the 
same. However, I couldn't, because this time, there was an exception before 
the segfault:

/home/rherzog/projects/GNUstep/GSXine/GSXine.debug/ix86/linux-gnu/gnu-gnu-gnu/GSXine: 
Uncaught exception NSInvalidArgumentException, reason: GSCString(instance) 
does not recognize pathForResource:ofType:

(Would be nice, if exceptions could provide a Java-like backtrace)


Well, I'll now first clean up the code a bit... (branched off from a tutorial)


Thanks,
  Raffi

--
 The difference between theory and practice is that in theory, there is
                no difference, but in practice, there is.

Raffael Herzog - herzog@raffael.ch - http://www.raffael.ch - ICQ #67961355