Perl 5 Porters Mailing List Summary: January 25th - February 1st
Hey everyone,
Following is the p5p (Perl 5 Porters) mailing list summary for the past week. Enjoy!
January 25th - February 1st
Corrections
The previous summary accidentally included the wrong ticket number for a Storable bug and blamed JSON::XS and Cpanel::JSON::XS. Those had been fixed in the published blog post and in the repo. My apologies. Thanks, Ben Bullock, for the correction!
News and updates
Encode 2.80 released! You can read more here.
Dagfinn Ilmari Mannsåker merged his branch that exposes more
siginfo_t
fields to the sounds of appreciation from fellow
developers.
Craig A. Berry has integrated podlators into core.
podlators 4.06 released!
The 12th grant report from Tony Cook's 6th grant in which approximately 9 tickets were reviewed or worked on, and 3 patches were applied in roughly 17 hours.
Tony also provides a summary of the month of December. Roughly 50 hours in which approximately 28 tickets were reviewed, and 5 patches were applied.
Bugs
Reported bugs
- Perl #127379: Breakpoint has no effect.
- Perl #127384: Use Importer to enhance Exporter in bleadperl.
- Perl #127377:
caller
does not return right info. - Perl #127405:
Remove
dump
from the core. - Perl #127435:
Outdated information in documentation
perlmodlib
. - Perl #127392: Module constant uses +1000 kB in 5.22 compared with 5.14.
Resolved bugs
- Perl #127351: Bleadperl breaks Test::Stream.
- Perl #126045:
Fix a Win32 Visual C 2003 DEBUGGING build failure in compiling
regexec.obj
. - Perl #119667: Smartmatch example in perlop is broken.
- Perl #127371:
Fix link failure of
APItest.dll
on VC 6. - Perl #127381: Fix a race condition in parallel builds with Visual C.
- Perl #127372:
Fix
op/infnan.t
test fails with NAN conversion on VC 6. - Perl #127426: Commit led to buffer overflow.
Rejected bugs
- Perl #127349:
Segfault in
Perl_newSVpv
. - Perl #127436: Fail to find a match with a regular expression.
Proposed patches
Another proposed patch by Tony Cook in Perl #126410 which does not break on debugging/threaded builds.
Discussion
Following Chad Granum's release of Importer, Aristotle commented on the list not favoring this suggestion while Kent Fredric commented on the benefit of it in comparison with the current exporting approach.
The discussion of the topic Karl Williamson raised with two different implementations of Unicode sentence boundary continues. It is still unclear what should be supported and how.
Ben Bullock pinned the problem in Perl #127232 to a Storable breaking the encapsulation of objects.
Bulk88 covered several ways of storing C resources in Perl. This is a worthy read.
Karl Williamson provided a review of a patch provided by Niko Tyni in Perl #127288.
Ricardo Signes is
pinging
the list on
Perl #125833
and suggesting simply forbidding any leading colons in require
or use
statements.
Another ping from Ricardo on Perl #125569, regarding a memory saving patch by Bulk88.
And one more ping from Ricardo on Perl #116965 which garnished some interest and discussion.
Ricardo also
commented
on Perl #124368
with regards to handling literal //
and /$null/
.
Dennis Kaarsemaker and Tony Cook fixed the Win32 Jenkins build and Dennis took the time to share with the list how the build script was fixed.
James E. Keenan started testing blead on Darwin/PPC and found two
failing tests on older Darwins for the new siginfo_t
fields that
Dagfinn Ilmari Mannsåker exposed. Ilmari and Lukas provided
patches with a fix and James is running a smoke test with them.
Ed Avis opened
Perl #127405
on removing the core function dump
since it serves little to no
value. Several comments added information on its purpose, problems,
and lack of current usefulness.
Felipe Gasper opened
Perl #127386
regarding setting the proper value for $!
. This led to an
interesting
talk
on the list regarding how Perl handles exit codes.
Jarkko Hietaniemi sent a Git hook he wrote that enforces a smoke test before a commit push, which he uses frequently, along with explanations on how it works.
Perl #127391 does not seem like a bug, but led to a discussion on associative subtleties.
Did you know that in the old days you could start a shell script with a colon? More explanations from Zefram here.
Leave a comment