Pete Ware <ware(a)cis.ohio-state.edu> writes:
Didier is right, I misstated that the phase 2 happens when all
events
are processed. It happens when the current event is finished being
dispatched (i.e. XtDispatchEvent() is about to return).
> OK, I'm about ready to give about!
What was the problem again? :-)
over-reaction on my part :-)
When I had sent the patch yesterday I discovered that I could reliably
crash XEmacs by closing a newly created frame. As I was in a hurry I
panicked and declared my patch flawed :-)
More careful examination showed that this crash was not related to the
patch (apparently it was caused by a (badly) miscompiled libdb.a
(don't ask -- I will LART our admin right now). This lib totally
fsck'ed up the XEmacs binary (dbx even reported an internal error when
reading it), but it still worked, kind of. After a clean recompile
everything is great now.
Without the patch I get the X error and with the patch it goes
away. No crashes -- everything is great now!!!
Steve, if you are still reading this, please consider this patch
(maybe even for 21.0?).
Sorry for the confusion.
Gunnar