Jan Vroonhof <vroonhof(a)math.ethz.ch> writes:
Aki Vehtari <Aki.Vehtari(a)hut.fi> writes:
> #4 0x80b4513 in assert_failed (
> file=0x820cd05 "/proj/ave/xemacs/xemacs-21.0-b44/src/event-stream.c",
> line=1212, expr=0x820ce9e "!NILP (rest)")
> at /proj/ave/xemacs/xemacs-21.0-b44/src/emacs.c:2605
Hmm. This can only happen when a timout was returned from X, but it
was not in the list of known id's... Which is "impossible" since
XtAddTimeout is only called in one place...
I can reproduce this with you're last message...
Somehow the fact that Offix has its own mini event loop must have
something to do with it....weird...
This does *not* sound good. Perhaps I could get the event-loop of OffiX
integrated into the event-loop of XEmacs, but...
If a drag is started, OffiX garbs the X pointer (i.e. all events go to
XEmacs). Now it does a loop, and looks for a button-release. If the event is
no button-release, it handles the event via XtDispatchEvent. And this must be
the position where the dump is produced...
The loop is starts in offix.c at line 125 in DndHandleDragging.
I will start picking around in this problem, but it's to late now... tomorrow
I will have time to fix it (I hope I can pinpoint the problem -- I'm crateful
for further hints).
Regards,
Oliver.