Hey everyone,

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

February 8th-14th

Corrections

The previous summary contained an incorrect link in the last paragraph which led to a 404. Thank you, Tony Cook, Ben Bullock, and Choroba, for spotting it.

News and updates

Grant reports from Dave Mitchell available. Due to a vacation, few hours were used this week.

Dave also provides his monthly grant report.

Additional grant reports from Tony Cook here, here, and here.

In total, about 52 hours and approximately 38 tickets were reviewed or worked on, and 7 patches were applied.

Tony also provides his monthly grant report for January 2016. Approximately 43 tickets were reviews, and 11 patches were applied.

Ricardo Signes announced that Time::HiRes is now upstream-blead.

Ricardo Signes is hoping to find more release managers for perl 5.25.x.

Karl Williamson updated that the Unicode Consortium has accepted 5 new emoji characters for Unicode 10.0.

Ricardo Signes has announced an upcoming code freeze. February 20th will be another development release and the user-visible code freeze. All additional changes, apart from pending bug fixes, should go through the list.

In preparations for perl 5.24, Ricardo started a review of all 5.24 blocking tickets.

Issues

New issues

• Perl #127511: Commit broke t/op/threads-dirh.t on Solaris threaded builds.
• Perl #127514: Minor comment in hv.h and scope.h.
• Perl #127517: h2ph should emit code that at least warns if a sizeof() entry is missing.
• Perl #127521: Supporting apostrophe format flag for builtin commifying in printf.
• Perl #127531: Permit \\ escape on right hand side of my.
• Perl #127537: /\W/ regression in 5.20 with UTF-8 and a synthetic start class.

Resolved issues

• Perl #125807: Move Time::HiRes to dist/.
• Perl #127435: Outdated information in documentation perlmodlib.
• Perl #127334: Segfault in S_incline at toke.c.
• Perl #126871: Perl_cv_const_sv_or_av assertion failure.
• Perl #8904: Missing warning for [[:digit:]]. (14 year old ticket!)
• Perl #122136: Bleadperl breaks PDF::Burst.
• Perl #126472: Bleadperl breaks Tie::Simple. (This includes Imager, Text::Fuzzy, Gtk2::Ex::ListModelConcat, and Gapp.)
• Perl #30608: $1 not set properly on backtrack. This was already resolved in 2006 by Dave Mitchell. This was verified and resolved by Karl Williamson.
• Perl #127497: Time::HiRes has moved from cpan to dist.
• Perl #127494: Bleadperl breaks Module::Info.
• Perl #127508: Add shortcut around syscalls when file not found in win32_stat.
• Perl #127337: Bleadperl breaks re::engine::GNU.
• Perl #84680: POSIX's isdigit() returns 1 for empty string.

Proposed patches

Tony Cook submitted an updated patch and worked with Felipe Gasper on the best phrasing for documentation in Perl #127386.

Tony also provided a patch for Perl #125833 (leading semicolons in require).

James E Keenan provided a patch for the documentation of pos in Perl #127518.

Bulk88 provides a patch in Perl #127532 to fix Win32 gmake with Win64 VC with 32 bit GCC in PATH build failure.

Tom Hukins provides a patch in Perl #127533 to fix tests in a FreeBSD jail.

Discussion

Ricardo Signes has asked for more clarifications about the suggested change in Perl #127384, making Exporter to use Importer for additional features.

Chad Granum adds more information on it.

Jim Avera adds more information on how important fatal warnings are, relating to Perl #127323, asking to have fatal warnings collect all fatal warnings during compilation and not halt on the first one.

Perl #126414 (perl rounds inode in PP stat) is an interesting scenario of "this shouldn't work but does". Bulk88 and Aristotle Pagaltzis have a subthread with more details.

Kent Fredric requests anyone that objects to tainted hash keys in core to provide proof that it is unacceptable.

James E. Keenan seeks for more information relating to gethostabyname() in Perl #127519. Paul Evans adds information on how the resolving interface works.

Ricardo Signes opened Perl #127531 to discuss a nicer, possibly more obvious syntax for variable aliasing by reference:

my \@y = $x;

(It is currently \my @y = $x.)

Vadim Konovalov asked about the behavior of strict regarding a change made to verify it was loaded as strict.pm (or strict.pmc), in order to prevent use Strict; silently being a no-op on case insensitive filesystems. This caused a failure in a hand-crafted fatpacker Vadim has written because of the value of __FILE__ (being used in the verification code of strict).

The solution includes using the line tracking abilities of Perl, such as #line 1 "strict.pm" to set __FILE__ in evaled code.