Hey everyone,

Following is the p5p (Perl 5 Porters) mailing list summary for two weeks I've missed. Sorry about that. Enjoy!

April 14th-27th

Correction: Previous summary stated that Perl #121734 was resolved. It was not. Thanks, Tony, for the correction!

Edit: Vincent Pit notes that there was no discussion about extracting parts of Scope::Upper (as suggested in the summary), but instead there was a single comment on writing a subset of it. This post was corrected. Thank you, Vincent.

News and updates

A lot has happened during these two weeks since the previous summaries went out.

Ricardo Signes stepped down from the role of the Perl Pumpking. Feel free to offer words of praise and thanks.

Sawyer X is the next pumpking.

Ricardo Signes released Perl 5.24.0-RC3. You can read more in the release announcement.

Dave Mitchell provides grant #2 reports #123 and #124 available here. Most of his time was spent on getting Scope::Upper to work on Perl 5.23.8 and above.

Tony Cook provides grant 7's 3rd and 4th.

More from Tony Cook, a summary of the March grant work.

Issues

New issues

• Perl #127914: Module::CoreList 5.20160410 not on CPAN.
• Perl #127915: Segfault triggered by "ab" x ~0.
• Perl #127921: Possible inconsistency in perlop documentation on precedence of || and //.
• Perl #127922: 5.24.0-RC1 Windows build failure with mingw.org gcc-3.4.5.
• Perl #127934: Assertion failure triggered by $z *= *z=0.
• Perl #127952: Assertion failure triggered by 0,!n||!\r=0.
• Perl #127953: stat.t build tests failing on some Cygwin.
• Perl #127956: Null pointer dereference.
• Perl #127976: each($scalar) error message anomalies.
• Perl #126993: Version control conflict markers.
• Perl #127999: Slowdown in split + list assign.
• Perl #128001: Quadmath builds segfault on repeat with a very large count.

Resolved issues

• Perl #113644: Panic error in perl5db.pl.
• Perl #125584: Mysterious taint issue in Bugzilla4Intranet.
• Perl #127709: Documentation problem with links and perlpod, podchecker.
• Perl #127894: -DDEBUGGING -Dusequadmath -Dusethreads builds crash early.
• Perl #127899: Extra slash in perldelta example in 5.22.2-RC1 and 5.24.0-RC1 confusing.
• Perl #127936: sprintf typo in 5.24 perldelta.

Proposed patches

Jim Keenan provides a patch for Perl #127391 in order to move forward with the documentation issue.

John Lightsey provided a patch in Perl #127923 to add blacklist and whitelist functionality to Locale::Maketext.

Jerry D. Hedden provided patches to upgrade threads to 2.06, Thread::Queue to 3.08, and threads::shared to 1.51.

Yves Orton provided a patch for Perl #123562, a problem with regular expressions possibly hanging on CPU 100%, which is considered a security problem.

Aaron Crane provided a patch for RT #100183, but since 5.24 is already at RC releases, it is frozen and the patch will get in on version 5.25. You can read Aaron's comment here.

Matthew Horsfall provided a patch relating to Perl #126579, warnings about newlines in open.

Matthew Horsfall also provided a patch for Perl #124050, t/harness.t can mistakenly run tests outside the perl source tree.

Aristotle Pagaltzis provided a patch to clean up Module::CoreList.

Discussion

Todd Rinaldo raised RT#127810 to provide a -Dfortify_inc Configure option to control the current directory appearing in @INC. The conversation around it continued further.

Zefram provides a detailed explaination about an observation made by Slaven Rezić in Perl #127909.

Sisyphus raised a confusing bit of documentation, which was fixed and backported to 5.22.

Sisyphus also asks about the binary name expected for make.exe. Bulk88 explains that it is easier to have it called gmake.exe to know what options it supports, and Leon Timmermans suggests it is possible to address it.

Dave Mitchell discussed his work on Scope::Upper. It seems Dave was able to get most of it working, but due to how the module works, Dave does not believe any sensible API can be shield the module from breakage. One comment surfaced the idea of writing a subset of its functionality.

Maxwell Carey asks what could cause a problem described on Stack Overflow with a failure to print.

In an email to the list, Sisyphus asks about the current state of ExtUtils::MakeMaker with relation to the current version in blead vs. CPAN.

The conversation about changing how the signatures feature worked with relation to @_ started by Dave Mitchell continues. Ricardo Signes provided a summary of Zefram's position and his conclusions.

Karl Williamson suggests POSIX::set_locale refusing to switch to a locale we know will cause a libc crash.

Smylers asks in Perl #122551 whether Term::ReadLine should not use Term::ReadLine::Perl as the default.

The conversation in Perl #127232 continues. You can read more here.

Renée Baecker mentions that perllol still reflects autoderef, which was removed, and should be updated.

Ed Avis suggested in Perl #127993 to add version control conflict markers, so Perl could warn you correctly when you forgot merge conflict markers in your code.