develooper Front page | perl.perl5.porters | Postings from October 2011

[perl #101430] Bleadperl v5.15.3-244-gaa33328 breaks ASG/List-Gen-0.973.tar.gz

Thread Previous | Thread Next
From:
Father Chrysostomos via RT
Date:
October 16, 2011 14:42
Subject:
[perl #101430] Bleadperl v5.15.3-244-gaa33328 breaks ASG/List-Gen-0.973.tar.gz
Message ID:
rt-3.6.HEAD-31297-1318801368-360.101430-15-0@perl.org
On Sun Oct 16 13:11:35 2011, sprout wrote:
> On Sun Oct 16 12:49:06 2011, sprout wrote:
> > On Sun Oct 16 00:56:56 2011, andreas.koenig.7os6VVqR@franz.ak.mind.de
> wrote:
> > > git bisect
> > > ----------
> > > commit aa33328e822d3a36d7448488403f30a991f148b2
> > > Author: Brian Fraser <fraserbn@gmail.com>
> > > Date:   Sat Jul 23 18:48:51 2011 -0300
> > > 
> > >     util.c for threads: stashpv_hvname_match UTF8 cleanup.
> > 
> > In the diff shown below, the null checks (stashpv && name) are removed.
> >  I should have noticed that when I applied it, so it’s partly my fault.
> > 
> > Before I go and fix this, though, I suspect there is a bug elsewhere, as
> > PL_curstash is set to an undefined *scalar* at the time this is called.
> 
> List::Generator contains this code:
> 
> package List::Generator;
> BEGIN {require List::Gen}
> 
> List::Gen does *List::Generator:: = *List::Gen::;
> 
> So when ‘require List::Gen’ finishes, PL_curstash points to an SV that
> has been freed and re-used.  That explains why name can be null in
> stashpv_hvname_match.  But that’s a separate bug.  stashpv_hvname_match
> is an API function, so it should continue to check for nulls.  I’ll have
> it fixed soon....

Now fixed with commit 6852b00.


Thread Previous | Thread Next


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