Perl 5 Porters Mailing List Summary: October 16th - Nov 1st

Hey everyone,

Following is the p5p (Perl 5 Porters) mailing list summary for the past two and a half weeks.


October 16th - Nov 1st

News and Updates

Steve Hay released perl 5.27.5!

We are now (C89 requirement) officially only relying on C89.

Dave Mitchell added a new optimization phase.

Also from Dave, a new op for optimization of multiple concatenations: OP_MULTICONCAT.

Dave also made several improvements to

Ryan Voots provided (BBC canary) access to a machine he has to be used as a Blead Breaks CPAN (BBC) canary

Grant Reports


New Issues

Resolved Issues

  • Perl #126707: Race condition in test IO/t/io_unix.t.
  • Perl #131730: exec PROGRAM LIST does not work with empty list.
  • Perl #132160: Use SvLEN_set/SvCUR_set in a few extra locations.
  • Perl #132245: heap-buffer-overflow in S_scan_const.
  • Perl #132298: Provide a way for distro maintainers to skip t/porting/customized.t.
  • Perl #132334: In Perl 5.26 and above, the /o modifier has side effects on split.
  • Perl #132342: stadtx_hash.h: Silence -Wimplicit-fallthrough compilation warnings.
  • Perl #132347: BBC: Blead breaks DBI.
  • Perl #132358: Implicit filehandle close not working when STDERR closed.

Suggested Patches

Tony Cook provided a patch for Perl #132245 (heap-buffer-overflow in S_scan_const). The patch has since been merged.

James Keenan provided a patch for Perl #132342.

Dan Dedrick provided a patch in Perl #132360 to fix race condition in builds on machines with 32 cores and above.

jkahrman provided a patch in Perl #132383 to fix the debugger printing 1 when a reference or object is thrown.


Now that we got rid of the left-curly-brace in regex problem, Karl Williamson has some ideas on reusing it. Think \w{Latin|Greek}, for example.

Karl Williamson has mentioned script runs before and now raised the topic again with more information after we fleshed it out during the core hackathon.

Karl is also working on optimizing the handling of UTF-8.

Philip R. Brenan asked about introducing Unicode characters as operators.

Zefram followed up with thoughts on source encoding.

Lukas Mai worked on using __builtin_add_overflow (and the likes) for overflow checking, with good results.


Re: Zefram's proposal of forcing Perl source to be interpreted as UTF-8, effectively making "use utf8" a no-op.

Many programs today rely on the fact that "substr" and "length" count bytes by default. The proposal would appear to be heading towards a certain amount of breakage...

That proposal would not generally change how functions like substr and length work. It would change how string literals in source code are interpreted, which would cause breakage. But the important part of the proposal there is:

"To get there, first we have to deprecate all source encoding that's incompatible with it. That is, we deprecate the presence of non-ASCII bytes anywhere in a source file other than in the scope of "use utf8"."

In other words, any string literals that would be affected by 'use utf8' would now be considered deprecated if there isn't 'use utf8' present. Whether that deprecation is wishful thinking is another matter.

Leave a comment

About Sawyer X

user-pic Gots to do the bloggingz