I think it's reasonable to do both, but maybe it makes more sense to me to
start with the latter. I'll need some help from John though. We'll keep you
posted.

2017-07-19 14:40 GMT-04:00 Eli Zaretskii <eliz@gnu.org>:

> > From: John Wiegley <jwiegley@gmail.com>
> > Cc: denis.redozubov@gmail.com,  27761@debbugs.gnu.org
> > Date: Wed, 19 Jul 2017 14:34:15 -0400
> >
> > >>>>> Eli Zaretskii <eliz@gnu.org> writes:
> >
> > >> This may be hard, but we'll try. It's 100% reproducible in both 25.1
> and
> > >> 25.2, but involves company-coq, Proof General, and some particular
> source
> > >> files.
> >
> > > Aren't those available for reproducing on another machine?
> >
> > Sure, if you're willing to install them, we can send the source files
> and the
> > exact steps.
>
> Please go ahead.
>
> > > Alternatively, I could guide you through debugging the recursion, but
> I'll
> > > need you to use GDB, not the macOS debugger.
> >
> > Sadly, only lldb is possible here. :(
>
> Then I cannot help you with specific commands, but if you could step
> through that code in get_next_display_element which recurses, and tell
> me why it does so, what is the character where this happens, and what
> is the value of it->method there, it could give some hints.
>