develooper Front page | perl.perl5.porters | Postings from September 2014

Re: [perl #122747] Assertion failed in Perl_reg_numbered_buff_fetch,file regcomp.c, line 7459

Thread Next
From:
Mark Martinec
Date:
September 12, 2014 15:46
Subject:
Re: [perl #122747] Assertion failed in Perl_reg_numbered_buff_fetch,file regcomp.c, line 7459
Message ID:
dc1c775db677a834274fe66924c83cdd@mailbox.ijs.si
>>>> Seeing that this is a crashing bug, it meets the policy.  It gets my
>>>> vote.  (2c1f00b90, that is, which alone is sufficient for maint.)
>>> 
>>> On the other hand, its a long-standing (and clearly rare) issue, and
>>> squeezing it in at the very last gasp into RC3 when it's had no time 
>>> to
>>> settle or be BBCed seems like a really good way to inadvertently 
>>> break
>>> 5.20.1. There's always 5.20.2.
>>> 
>> What-Dave-said++
> 
> FWIW, I think delaying for a minor release is a good idea. If Mark 
> really
> needs this he can cherry-pick the patch and build a custom perl.
> 
> Sorry for any inconvenience Mark, but this patch has characteristics, 
> (such
> as being almost *too* easy), which make me a think a bit of time to 
> cook in
> blead is a good idea.

Sure, understood. It is indeed uncomfortably close to a 5.20.1 release.

> On the other hand, its a long-standing (and clearly rare) issue

If the issue is otherwise not harmful (like causing memory corruption)
and the assert failure can be safely ignored, perhaps there can just be
a warning in perldelta that -DDEBUGGING must not be used in regular use.

The event is not so rare (e.g. one such case per several days of mail
filtering), but goes by unnoticed as the system-installed perl is
usually not built with debugging. I had debugging enabled because of
trying out a fresh release candidate.

Thread Next


nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About