Re: apparent infloop in re_match_2_internal

[Stefan Monnier]

>> In emacs.git (again from ~3 days ago, debian unstable),
>> using GNUS, when I try to reply to a certain message, I
>> get into what looks like an unbreakable infinite loop.
> What's the regexp?

Apparent inf-looping in regexp matching is common because we use
a backtracking search (hence with really bad worst case behavior).
This is a problem in itself, but we've lived with it for a long time
now.  The problem in his report is that additionally C-g failed to work.
Can you check the value of Vinhibit_quit and show us the xbacktrace to
see if the reason C-g didn't work is because we're in an inhibit-quit
context and if so which one?  If the inhibit-quit is really needed
there, then we'll indeed have to fix the regexp to not trigger the
worst-case behavior.


        Stefan