bug#74547: 31.0.50; igc: assertion failed in buffer.c

[Gerd Möllmann]

Óscar Fuentes <oscarfv@telefonica.net> writes:

> Gerd Möllmann <gerd.moellmann@gmail.com> writes:
>
>> Hm, reading Oscar's report again - he was using LSP. Maybe it's
>> something in the parser.
>
> The three times that I caugth the crash with the debugger mini-echo [1]
> was in the Elisp backtrace. Maybe lsp-mode does something that upsets
> the garbage collector and it surfaces when mini-echo executes. Just now
> tried for some minutes with mini-echo active and without lsp-mode and no
> crash, but maybe I was unlucky. lsp-mode is heavy on GC and it is
> difficult to replicate that.

Yeah, It's pretty difficult to find the exact time when something goes
wrong because the allocation points have some memory reserved, so that
it's not exactly predictable when a GC step happens.

But I find Pip's change very promising. It's fixing something that
cleqrly cannot work with MPS. Could you please try it (it doesn't seem
to be pushed yet.)