Perl 5 Porters Mailing List Summary: November 16th-22nd

Hey everyone,

Following is the p5p (Perl 5 Porters) mailing list summary for the past week. Enjoy!

November 16th-22nd

This week I've separated all the proposed patches for issues under a different title: Proposed patches. This is because they do not exactly fit under Updates (which I wish to reserve for news) or under Discussion (since I wish to reserve this for conversations that take place).

I hope this works better.

Feedback is always welcome. :)

Updates

Perl 5.23.5 is now available, thanks to Abigail!

Two additional reports from Tony Cook, available here and here. In total for both reports, Tony has spent about 37 hours, worked on or reviewed 24 tickets, and applied 5 patches.

Karen Etheridge provided a patched (which was merged since) to clean up the verbosity of some tests due to passing TODO tests in Module::Metadata.

Bugs

Reported bugs

Perl #126667, reported by Dan Collins, a fuzzer-found assertion failure, triggered by the code:

qr/(?{})@0{()}@0/

Peter Gordon had reported an issue with IO::Select.

Perl #126676, repoted by Todd Rinaldo, mentions a curious case of file handle closing, \shift, and - as Dave Mitchell expanded on - anything (such as open()) that attempts to instantiate an anonymous value into a ref to a typeglob.

Perl #126697, reported by Lukas Mai, mentions that local is not working as expected in embedded code in regexes.

Resolved bugs

Proposed patches

Regarding possible unintended mix of POD and code, Aaron Crane provides a patch to add a warning for such cases.

Ed Avis provided patches for rewording of lookahead vs. look-ahead to be consistent.

Tony Cook provided a patch to resolve Perl #126635.

Bulk88 provided additional patches for parallelness and running all Makefile.PLs in 1 process (here and here).

Discussion

Yves Orton has been investigating an issue with Sereal handling undef SVs from threads::safe. Dave Mitchell chipped in with details about PVLV types.

Karl Williamson found a reported problem with an AIX test to likely be a stack overflow problem.

Todd Rinaldo updates that - given a recent commit on B::C, he sees no reason to keep B::Section in the core. This is part of Perl #126244.

Christian Hansen provided patches to make UTF-X validation about 50% - 300% faster.

Karl Williamson adds more light to the conversation on bitwise string operators.

A long conversation in Perl #124349. took place on arguments to the hostname function call to Sys::Hostname, whether to croak or warn, and at some point, whether to make any change at all.

Ricardo Signes mentions that base has tests which fail on perl 5.6. Ricardo suggests dropping support for 5.6 and require perl 5.8 moving forward.

Leave a comment

About Sawyer X

user-pic Gots to do the bloggingz