------- You are receiving this mail because: -------
You are on the CC list for the bug.
http://bugs.exim.org/show_bug.cgi?id=1615
--- Comment #8 from Zoltan Herczeg <hzmester@???> 2015-04-17 13:43:42 ---
> which as I understand it does do the UTF-8 check, and wouldn't suffer from jit
> assumptions of valid UTF-8, would it?
Yes it should.
It is difficult to figure out this bug without knowing the input. I can only do
wild guesses. I know that \b does a backward character scan. It searches for a
valid utf start byte. That can underflow for invalid UTF sequences. Or maybe
there is a bug, which incorrectly processes this search. Or there is a bug in
the UTF validation. Could you share the first few bytes of the input (as hex
bytes) and the start offset?
--
Configure bugmail:
http://bugs.exim.org/userprefs.cgi?tab=email
