Bug #81635 for Internals-DumpArenas: Test failures due to hash randomisation in perl 5.17.6

Sun Dec 02 03:27:21 2012 ANDK [...] cpan.org - Ticket created

Subject:

Test failures due to hash randomisation in perl 5.17.6

Since bleadperl v5.17.5-518-g7dc8663 your tests are failing frequently. That commit introduced hash key randomization and it seems at least the test t/10-use.t is hit by that. Find a sample fail report at: http://www.cpantesters.org/cpan/report/567264cc-3196-11e2-be98-8ec2a290f8f5 You can read more about the change at http://perl5.git.perl.org/perl.git/commit/7dc8663964c66a698d31bbdc8e8abed69bddeec3 or at http://www.perlmonks.org/?node_id=1005122 You may have to run the test several times until the randomization causes a fail. HTH&&Thanks&&Regards,

Fri Oct 02 03:14:57 2015 RURBAN [...] cpan.org - Correspondence added

On Sun Dec 02 03:27:21 2012, ANDK wrote: Show quoted text

> Since bleadperl v5.17.5-518-g7dc8663 your tests are failing > frequently. > That commit introduced hash key randomization and it seems at least > the test > > t/10-use.t > > is hit by that. Find a sample fail report at:

I rather think that the SvSTASH of the %version:: is corrupt. It shows up in the debugger as 0x18, which is a wrong SV*. When I omit such a wrong value in dump.c , the test passes. Perl_do_hv_dump: /* %version:: has a corrupt stash */ if (sv && sv>0x100 && (hvname = HvNAME_get(sv))) I'll investigate further. Maybe it is related somehow. I don't see any hash key randomization related code here, but code which dumps all values. Only B::C does something similar. And in B::C I had to add this protection: # need to fix B for SVpad_TYPED without formal STASH my $stash = ( ref($t) eq 'B::PVMG' and ref( $t->SvSTASH ) ne 'B::SPECIAL' ) ? $t->SvSTASH->NAME : ' %version:: has also magic attached. -- Reini Urban

Fri Oct 02 03:14:58 2015 The RT System itself - Status changed from 'new' to 'open'

Fri Oct 02 03:51:05 2015 RURBAN [...] cpan.org - Correspondence added

Confirmed. It is an upstream core bug in the version module. This module just trips over it. Devel::Peek which also justs calls sv_dump fails in the same way. See https://rt.cpan.org/Ticket/Display.html?id=107459

Fri Oct 02 06:24:54 2015 RURBAN [...] cpan.org - Correspondence added

On Fri Oct 02 03:51:05 2015, RURBAN wrote: Show quoted text

> Confirmed. It is an upstream core bug in the version module. > This module just trips over it. Devel::Peek which also justs calls > sv_dump fails in the same way. > > See https://rt.cpan.org/Ticket/Display.html?id=107459

See https://github.com/jbenjore/internals-dumparenas/pull/1 for the fixes. -- Reini Urban