bug#33345: 27.0.50; Abort in bidi_cache_iterator_state

[Stefan Monnier]

> But this means that the cache is empty, and we can never scan back
> (scan_dir = -1) when the cache is empty.  So something very strange
> happened there.

I think it's a case of memory corruption and doesn't really have much to
do with bidi code.  But I'm not progressing very fast: it's somewhat
deterministic but it is influenced by elements I'm not aware of, so it's
pretty nasty.

> Instead of me asking you gobs of questions, perhaps you could give me
> an ssh login on that system, and arrange for a GDB session which
> already hit the call to emacs_abort to be in a state where I can
> switch to it after logging in (with 'screen' or 'tmux', I think).
> Then I could look around and see if I find out some clues.

It's a machine that's not reachable from the internet and is off most of
the time.  I haven't yet managed to reproduce it on another machine.
I'll let you know when I'm making progress,
Thanks,


        Stefan