This is the mail archive of the
mailing list for the glibc project.
Re: [PATCH] regexec: Fix off-by-one bug in weight comparison [BZ #23036]
* Adhemerval Zanella:
> On 09/07/2018 15:00, Carlos O'Donell wrote:
>> On 07/09/2018 01:20 PM, Florian Weimer wrote:
>>> 2018-07-09 Florian Weimer <firstname.lastname@example.org>
>>> [BZ #23036]
>>> * posix/regexec.c (check_node_accept_bytes): When comparing
>>> weights, do not compare an extra byte after the end of the
>> I assume that this fixes the issue and that you have minimally passed
>> the regression testing for x86_64. I would like a test for this, but it
>> can wait. This regression is a blocker for the release, and I appreciate
>> your root cause analysis of the issue.
> I still think we should add a regression test for it, the bug report
> have one quite simple which we can add as an independent test (which
> should not invalidate testsuite validation already done).
I'm working on a test which constructs all matching single-character
multi-byte strings matching a given pattern in a given locale, so we
can check that we get the expected result.