Bleadperl v5.31.1-139-g2773b4f50f breaks MANWAR/Test-Strict-0.52.tar.gz

[p5pRT]

Migrated from rt.perl.org#134311 (status was 'open')

Searchable as RT134311$

[p5pRT]

From @andk

commit 2773b4f
Author​: Tony Cook <tony@​develop-help.com>
Date​: Thu Jul 4 15​:02​:46 2019 +1000

  Revert "postpone perl_parse() exit(0) bugfix"

Sample fail report​:

http​://www.cpantesters.org/cpan/report/544986ee-aa7f-11e9-889c-32e91573969b

Cheers,
--
andreas
PS​: perl

[p5pRT]

From @andk

Xref​: manwar/Test-Strict#30
--
andreas

[p5pRT]

From @jkeenan

On Thu, 25 Jul 2019 01​:04​:45 GMT, andreas.koenig.7os6VVqR@​franz.ak.mind.de wrote​:

commit 2773b4f
Author​: Tony Cook <tony@​develop-help.com>
Date​: Thu Jul 4 15​:02​:46 2019 +1000

Revert "postpone perl\_parse\(\) exit\(0\) bugfix"

Sample fail report​:

http​://www.cpantesters.org/cpan/report/544986ee-aa7f-11e9-889c-32e91573969b

Cheers,

Beginning with the breaking commit identified by Andreas, the test is stating that the following heredoc's perl syntax is not ok​:

#####
  print $fh2 <<'DUMMY';
#!/usr/bin/perl -vw
use strict;
print "Hello world";

DUMMY
#####

Inline comments in the C-code for the breaking commit do indicate that perlrun is implicated.

Thank you very much.
--
James E Keenan (jkeenan@​cpan.org)

[p5pRT]

The RT System itself - Status changed from 'new' to 'open'

[p5pRT]

From @jkeenan

On Thu, 25 Jul 2019 02​:38​:13 GMT, jkeenan wrote​:

On Thu, 25 Jul 2019 01​:04​:45 GMT,
andreas.koenig.7os6VVqR@​franz.ak.mind.de wrote​:

commit 2773b4f
Author​: Tony Cook <tony@​develop-help.com>
Date​: Thu Jul 4 15​:02​:46 2019 +1000

Revert "postpone perl_parse() exit(0) bugfix"

Sample fail report​:

http​://www.cpantesters.org/cpan/report/544986ee-aa7f-11e9-889c-
32e91573969b

Cheers,

Beginning with the breaking commit identified by Andreas, the test is
stating that the following heredoc's perl syntax is not ok​:

#####
print $fh2 <<'DUMMY';
#!/usr/bin/perl -vw
use strict;
print "Hello world";

DUMMY
#####

And in that heredoc, it's the 'v' switch that the module is judging to be a syntax error. If that shebang line only contains '-w', then the test passes.

Inline comments in the C-code for the breaking commit do indicate that
perlrun is implicated.

Thank you very much.

--
James E Keenan (jkeenan@​cpan.org)

[p5pRT]

From @csjewell

#####
print $fh2 <<'DUMMY';
#!/usr/bin/perl -vw
use strict;
print "Hello world";

DUMMY
#####

1) I want to go 'what the?' at the combination of the v and w switches. I'm not surprised that combination of switches is now giving an error - I would have been expecting it to before this.

2) This desired behavior has been punted for a year (since 5.28.0-rc1 - it was introduced originally in 5.27.7) already.

3) An earlier version of this module was also mentioned at [perl #132577] - so the fact that this commit was coming back and might break this module has been known about for that 18 months, I would think. (It doesn't look like anybody told the module author there, however.)

[p5pRT]

From [Unknown Contact. See original ticket]

#####
print $fh2 <<'DUMMY';
#!/usr/bin/perl -vw
use strict;
print "Hello world";

DUMMY
#####

1) I want to go 'what the?' at the combination of the v and w switches. I'm not surprised that combination of switches is now giving an error - I would have been expecting it to before this.

2) This desired behavior has been punted for a year (since 5.28.0-rc1 - it was introduced originally in 5.27.7) already.

3) An earlier version of this module was also mentioned at [perl #132577] - so the fact that this commit was coming back and might break this module has been known about for that 18 months, I would think. (It doesn't look like anybody told the module author there, however.)

[p5pRT]

From @jkeenan

On Thu, 25 Jul 2019 01​:04​:45 GMT, andreas.koenig.7os6VVqR@​franz.ak.mind.de wrote​:

commit 2773b4f
Author​: Tony Cook <tony@​develop-help.com>
Date​: Thu Jul 4 15​:02​:46 2019 +1000

Revert "postpone perl\_parse\(\) exit\(0\) bugfix"

Sample fail report​:

http​://www.cpantesters.org/cpan/report/544986ee-aa7f-11e9-889c-32e91573969b

Cheers,

Additional CPAN modules broken by this commit​:

asa-1.03
Aspect-1.04
Class-Autouse-2.01
HTTP-Online-0.02
LWP-Online-1.08
Locale-Msgfmt-0.15
ORLite-1.98
Process-0.30
SMS-Send-1.06
Text-FindIndent-0.11

#####
2773b4f is the first bad commit
commit 2773b4f
Author​: Tony Cook <tony@​develop-help.com>
Date​: Thu Jul 4 15​:02​:46 2019 +1000

  Revert "postpone perl_parse() exit(0) bugfix"

  This reverts commit 857320c,
  re-instating the [perl #2754] fix, which was reverted in late
  2017 to allow Module​::Install based distributions to update or
  re-work per [perl #132577].

  # Conflicts​:
  # t/op/blocks.t

perl.c | 27 ++++++++-------------------
t/op/blocks.t | 12 ++++++------
2 files changed, 14 insertions(+), 25 deletions(-)
bisect run success
That took 1361 seconds.
#####

These modules have been discussed on perl5-porters list starting at https://www.nntp.perl.org/group/perl.perl5.porters/2019/07/msg255753.html.  They all use Module​::Install​::DSL. They all PASS against the immediately preceding commit in blead but fail to configure at the commit in question.

I'm inclined to think that these are cases where, though the CPAN failure can be attributed to a specific commit to blead, the commit to blead may be correct and simply exposing dodgy code in the CPAN modules, viz., use of Module​::Install​::DSL.

Thank you very much.

--
James E Keenan (jkeenan@​cpan.org)

[eserte]

Looking at http://matrix.cpantesters.org/?dist=Test-Strict+0.52 it seems that the problem has gone somewhen between 5.31.2 and 5.31.3.

Also, the additional CPAN modules listed in this ticket seem to pass with 5.31.7 (asa has now a new version 1.04 and ORLite does fail for all perls on my systems, unrelated to bleadperl).

[jkeenan]

Looking at http://matrix.cpantesters.org/?dist=Test-Strict+0.52 it seems that the problem has gone somewhen between 5.31.2 and 5.31.3.

Also, the additional CPAN modules listed in this ticket seem to pass with 5.31.7 (asa has now a new version 1.04 and ORLite does fail for all perls on my systems, unrelated to bleadperl).

Confirmed that all CPAN distributions listed in this ticket (except ORLite) now PASS when installed with cpanm against blead. (ORLite is consistenly failing on CPAN-River-3000.)

Bisecting with this invocation:

perl Porting/bisect.pl \
--start=2773b4f50 \
--end=v5.31.3 \
--expect-fail \
--module=Test::Strict

... points to the following commit:

commit 625e8b0bffed3334db9b3897f8713d570fd5385c
Author:     Tony Cook <tony@develop-help.com>
AuthorDate: Tue Aug 6 20:45:52 2019
Commit:     Tony Cook <tony@develop-help.com>
CommitDate: Tue Aug 6 20:45:52 2019

    Revert "Revert "postpone perl_parse() exit(0) bugfix""
    
    This reverts commit 2773b4f50f991900e38d33daace2b9c6a0902c6a.
    
    I haven't made much progress in resolving the problems this produces
    downstream, so rather than leaving it broken, I'll revert it until
    they can be solved.

So we can close this ticket.

Thank you very much.
Jim Keenan