bug#16234: 24.3.50; emacs_backtrace.txt

[Drew Adams]

> > Aha! the trap worked!  I installed in trunk revision something that
> > will hopefully show us a full backtrace from intervals.c:679, so we
> > could try to find out whodunit.
> 
> emacs-20131225-r115738-bin-i386.zip is on its way to Dropbox.

Thanks.  I picked it up and tried it.  But as I said earlier, I'm
afraid the crash was fortuitous, not something I can easily reproduce.
As I mentioned, I had already used M-k in the minibuffer more than
once, even for the same command invocation.  Only once did the crash
occur when I did that.

So far, I have not gotten a crash.  Will of course let you know.


[However, something else (horrible) must have changed in this build,
because now when I use my (own) command to open a file, it takes
multiple seconds just to calculate the list of candidate files in my
usual directory.  Nothing to do with the problem at hand.  But it
does mean that this build is not something I can use, to do more
than test for this bug and hope for another crash.]