New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Bleadperl v5.15.5-306-gb50b205 breaks RIBASUSHI/namespace-clean-0.21.tar.gz #11807
Comments
From @andkgit bisect commit b50b205 Implement new ‘use 5.xxx' plan diagnostics You need to install in this order: (1) B::Hooks::EndOfScope Otherwise the test will get skipped. The test output looks like perl -V Summary of my perl5 (revision 5 version 15 subversion 5) configuration: Characteristics of this binary (from libperl): -- |
From @nwc10On Wed, Dec 14, 2011 at 11:12:02PM -0800, Andreas J. Koenig via RT wrote:
# Failed test 'Exit 139 from: /home/sand/src/perl/repoperls/installed-perls/perl/v5.15.5-366-ga877535/9980/bin/perl -MDevel::Hide=B::Hooks::EndOfScope /tmp/loop_over_bdir-UX21EL/namespace-clean-0.21-jqdYWh/t/02-inheritance.t' Not looked at this further than spotting the above. But to me (and to my surprise, not Wikipedia) 139 is "SEGV with coredump". Nicholas Clark |
The RT System itself - Status changed from 'new' to 'open' |
From @cpansproutOn Wed Dec 14 23:12:02 2011, andreas.koenig.7os6VVqR@franz.ak.mind.de wrote:
I don’t know that I’ll have time to look at this today, but here are C perl5.15.5 -Mblib -MDevel::Hide=B::Hooks::EndOfScope t/02-inheritance.t where perl5.15.5 is actually v5.15.5-425-g5dca8ed: #0 Perl_sv_magic (sv=0x0, obj=0x0, how=104, name=0x8a9340 "???", blah blah blah at (eval { use warnings; use base \'Tie::ExtraHash\'; { use warnings; sub arm { bless [ $_[1] ] } sub DESTROY { $_[0]->[0]->() } sub on_scope_end (&) { if( my $stack = tied( %^H ) ) {
|
From @cpansproutOn Sun Dec 18 13:24:34 2011, sprout wrote:
Notice in particular that the sv passed to Perl_sv_magic is null:
-- Father Chrysostomos |
From @cpansproutThis is what I have reduced it to so far: #!/usr/bin/env perl package namespace::clean::_TieHintHash; sub TIEHASH { my $p = shift; bless [{}, @_], $p } package main; use strict; sub bar { foo() } -- Father Chrysostomos |
From @cpansproutOn Sun Dec 18 22:29:46 2011, sprout wrote:
This started crashing because strict.pm started using the hint hash, and I can produce the same crash with 5.10 like this: use Devel::Hide 'B::Hooks::EndOfScope'; Just using sort.pm instead of strict.pm will do it. And here is an example that tickles the underlying bug without any modules: #!/usr/bin/env perl package namespace::clean::_TieHintHash; sub TIEHASH { bless[] } package main; BEGIN { $^H{foo} = "bar" } The crash was caused by this change: 5b9c067 is the first bad commit Automatically set HINT_LOCALIZE_HH whenever %^H is modified. In trying to debug this, I’ve noticed so far that this code in while ((entry = hv_iternext_flags(ohv, 0))) { But I’m stuck, because I don’t understand magic well enough to what is If anyone could shed light on that, that would be much appreciated. -- Father Chrysostomos |
From @cpansproutThis is now about as fixed as it can get, with this commit: commit 95cf236 [perl #106282] Don’t crash cloning tied %^H -- Father Chrysostomos |
From [Unknown Contact. See original ticket]This is now about as fixed as it can get, with this commit: commit 95cf236 [perl #106282] Don’t crash cloning tied %^H -- Father Chrysostomos |
@cpansprout - Status changed from 'open' to 'resolved' |
From @andk
> According to our records, your request regarding I now get OOM: t/08-const-sub.t ......... ok # Failed test 'Exit 256 from: /home/sand/src/perl/repoperls/installed-perls/perl/v5.15.6-10-g95cf236/127e/bin/perl -MDevel::Hide=B::Hooks::EndOfScope /tmp/tmp.2V6nCmp6FH/namespace-clean-0.21-BFmxvt/t/02-inheritance.t' Test Summary Report t/10-pure-perl.t (Wstat: 256 Tests: 9 Failed: 1) perl -V Summary of my perl5 (revision 5 version 15 subversion 6) configuration: Characteristics of this binary (from libperl): -- |
From @cpansproutOn Tue Dec 20 21:06:46 2011, andreas.koenig.7os6VVqR@franz.ak.mind.de wrote:
I believe I have fixed that with commit 7ef9d42. -- Father Chrysostomos |
From @andk
> I now get OOM: > t/08-const-sub.t ......... ok The above was with 5.15.6-10 with -Duselongdouble. Better results with t/06-other-types.t ....... ok Here all tests passed. -- |
From @ribasushiAndreas J. Koenig wrote:
Please try with the new namespace::clean 0.21_01 I shipped an hour ago. |
Migrated from rt.perl.org#106282 (status was 'resolved')
Searchable as RT106282$
The text was updated successfully, but these errors were encountered: