#582925 perl: stack should be reference counted

Package:
perl
Source:
perl
Description:
Larry Wall's Practical Extraction and Report Language
Submitter:
Niko Tyni
Date:
2020-11-15 00:21:06 UTC
Severity:
normal
#582925#5
Date:
2010-05-24 14:09:08 UTC
From:
To:
This package fails to build on amd64 with perl 5.12.1-1 from experimental:

  t/multi_warp_value.t ...... ok
  UNIVERSAL->import is deprecated and will be removed in a future perl at /build/niko-libconfig-model-perl_1.202-1-amd64-7pdWZA/libconfig-model-perl-1.202/blib/lib/Config/Model/ObjTreeScanner.pm line 25

  #   Failed test 'fetch_element with unexpected experience'
  #   at t/node.t line 125.
  # expecting: Regexp ((?-xism:Unexpected experience))
  # found: Bizarre copy of ARRAY in sassign at /usr/share/perl/5.12/Carp.pm line 95.
  # Looks like you failed 1 test of 50.
  t/node.t ..................
  Dubious, test returned 1 (wstat 256, 0x100)
  Failed 1/50 subtests
  UNIVERSAL->import is deprecated and will be removed in a future perl at /build/niko-libconfig-model-perl_1.202-1-amd64-7pdWZA/libconfig-model-perl-1.202/blib/lib/Config/Model/ObjTreeScanner.pm line 25
  t/node_get_set.t .......... ok

This is probably a bug in Perl itself, see
http://rt.perl.org/rt3/Public/Bug/Display.html?id=52610
but there doesn't seem to be any easy fix, so a workaround is needed
for Config::Model.

I'll clone a bug against perl when I get the BTS ack for this.

FWIW, this reliably shows on the CPAN tester reports too.
http://matrix.cpantesters.org/?dist=Config-Model+1.202
#582925#22
Date:
2010-05-24 18:38:45 UTC
From:
To:
severity 582925 normal
thanks

Looking at the nasty things Test::Exception does with the help of
Sub::Uplevel to trigger this, I'm lowering the severity of this bug.

IIUC, the root problem is that the stack isn't reference counted,
so modifying things that are on the stack may get them freed
prematurely.
#582925#29
Date:
2011-05-21 00:18:19 UTC
From:
To:
And indeed it was identified as a bug in Test::Exception, fixed in
0.31.

The underlying problem still exists if the upstream report is
to be believed, but it won't manifest in the way this bug report
suggests, so I'm not sure whether it makes sense to leave this open.
#582925#34
Date:
2011-05-21 07:02:44 UTC
From:
To:
retitle 582925 perl: stack should be reference counted
forwarded 582925 http://rt.perl.org/rt3/Public/Bug/Display.html?id=77706
thanks

OK, let's leave it open to track the whole stack ref counting issue.
Changing the forwarded pointer to the upstream meta-bug.
#582925#43
Date:
2017-03-27 09:50:04 UTC
From:
To:
Control: reassign 839218 nama
Control: retitle 839218 nama: FTBFS because of perl's lack of stack reference counting

Thanks, reassigning back to Nama so the workaround can be considered.
The underlying problem in perl is tracked in #582925.

Dominic.