Hey everyone,

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

September 12th-28th

News and highlights

Perl 5.25.5 released by Stevan Little!

The epigraph reads:

We hypostatize information into objects. Rearrangement of objects is change in the content of the information; the message has changed. This is a language which we have lost the ability to read. We ourselves are a part of this language; changes in us are changes in the content of the information. We ourselves are information-rich; information enters us, is processed and is then projected outward once more, now in an altered form. We are not aware that we are doing this, that in fact this is all we are doing

-- Philip K. Dick, VALIS

Karl Williamson updated that the Unicode 9.0 core specification is now available in paperback form.

Jarkko Hietaniemi provided an update about the -fsanitize=undefined in gcc and clang.

podlators 4.08 has been released.

Grant reports

• Dave Mitchell TPF Grant #2 reports #140 and #141.
• Dave Mitchell TPF Grant #2 report #142.

Issues

New issues

• Perl #129267: Possible string overrun with invalid len in gv.c.
• Perl #129283: Compile-time check for $1, $2 on RHS of s/// operator.
• Perl #129285: Perl_gv_fullname4 assertion failure.
• Perl #129286: Perl_hv_common assertion failure.
• Perl #129287: Perl_re_op_compile assertion failure.
• Perl #129288: Null ptr deref, segfault in S_rv2gv.
• Perl #129297: perldoc perlsub continues to mention split in scalar context clobbers @_ (a feature removed in Perl 5.12).
• Perl #129300: Assertion Failure: Perl_sv_grow.
• Perl #129301: Null ptr deref, segfault in Perl_mro_isa_changed_in.
• Perl #129302: Assertion failure in Perl_hv_ename_add.
• Perl #129315: DB::sub is not called for eval.
• Perl #129317: DB::DB is not called.
• Perl #129320: Assertion failed.
• Perl #129322: S_invlist_clear assertion failure.
• Perl #129333: Warning if regexp used as LHS of =~ instead of RHS.
• Perl #129336: #!perl -i whatever is interpreted as #!perl -i -whatever.
• Perl #129347: null pointer deref S_ft_return_false.
• Perl #129348: File::Find issue with link counts on Bash on Ubuntu on Windows.
• Perl #129350: null pointer deref Perl_re_intuit_start.
• Perl #129354: Default configuration on x86_64-linux-gnu cannot build binary extension modules.
• Perl #129361: Perl 5.25.5 / 5.24.0 do not build on SunOS4.
• Perl #129371: Documentation of binary | and & operators.

Resolved issues

• Perl #113934: eval 'BEGIN { goto ... }'.
• Perl #125603: failed to compile perl5 when path includes the string "blib".
• Perl #128629: locale.t on Win32 failing "Verify that strings with embedded NUL collate" and "Verify that strings with embedded NUL and extra trailing NUL collate".
• Perl #129267: Possible string overrun with invalid len in gv.c.
• Perl #129287: Perl_re_op_compile assertion failure.
• Perl #129297: perldoc perlsub continues to mention split in scalar context clobbers @_ (a feature removed in Perl 5.12).

Rejected issues

• Perl #114378: cond_signal does not wake up a thread.

Suggested patches

Kevin Goess provided a patch in Perl #129277: to rephrase a part in perlrecharclass.

Pali provided a patch in Perl #129298 to update documentation about UTF-8.

Dan Collins provided a patch to fix the -i argument parsing raised in Perl #129336.

Discussion

Lukas Mai (mauke) asked (Weird XXX comment in pp_ctl.c) about a comment in the code from 1999.

Yves Orton noted a problem in List::Util not triggering magic.

Eric Brine asked why COW (Copy On Write) is not used in a specific piece of code. Explanation provided (Re: Why isn't COW used here?) by Dave Mitchell.

H. Merijn Brand (Tux) is concerned about memory leaks in threaded perls.

Hanno Böck provided more context for Perl #126599 (Unnecessary preload wrapper in Makefile).

There were additional comments made in relation to Perl #123665 ('Useless use of a constant in void context' is compile-time but not syntax warning), explaining the observed behavior.

Craig A. Berry asked (Re: [perl #128972] Build failure on OS X 10.12 Sierra) whether Time::HiRes uses the system clock_gettime() on OS X 10.12 Sierra. Does anyone have Sierra and can help verify what Craig is asking?

Further discussion on Perl #126310 (no "Malformed UTF-8 character" warning on single-quoted strings under use utf8) about whether a malformed UTF-8 error should die or not.

A discussion started from Perl #129283 (Compile-time check for $1, $2 on RHS of s/// operator) on whether we should be catching situation in regular expressions where a capture variable is used while there was no capture there. The start of the read is here (Re: [perl #129283] Compile-time check for $1, $2 om RHS of s///operator) but the thread continues further.

A. Sinan Unur asked (I am confused about the future of Const::Fast) about the future of Const::Fast. The conversation continued here.

In a conversation about parallel builds, it was observed that MANIFEST is sorted as part of the build process and that can cause issues.

The discussion around a deprecation Pod document (tentatively referred to as perldeprecated.pod continues.

Michiel Beijen raised an interesting problem when running Ubuntu on Windows, which eventually ended up in opening Perl #129348 (File::Find issue with link counts on Bash on Ubuntu on Windows).

Ævar Arnfjörð Bjarmason provided a branch which removes the ability to call use UNIVERSAL;. Calling use UNIVERSAL qw(can isa) has been deprecated and warning since 5.12 and has been a fatal error since 5.22.

Dave Mitchell is working on cleaning up output to STDERR when building perl. He asked about a warning that appears when you build perl when it exists in additional system directories. Should we still warn about this? If so, should we still output this on STDERR or move it to STDOUT?

Andy Lester started working on config.SH to clear up what -Weverything will include when using clang, since it doesn't map directly to gcc's -Wall.

Dave Mitchell provided a code review of a patch to bench.pl.