Hey everyone,

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

August 15th-21st

News and highlights

Perl 5.25.4 is now available!

There's a suggestion to dual-life Hash::Util. This would allow moving a function there and deprecating it in the Internals:: namespace.

Dave Mitchell suggested to improving the output of high hints bits displayed in B::Concise. That has been done following the suggestion.

Dave Mitchell also found a specific check in the internal av_fetch function which slows down all array fetches and doesn't even cover more than one particular way in which a problem could be triggered. Dave suggested removing it, which makes sense. This had also been done following the suggestion.

Ricardo Signes pinged the topic of being able to write state(@a) in Perl #114932.

A surge of new tickets provided by Brian Carpenter who is diligently fuzzing perl. Brian also wrote a blog post about it.

Grant reports

• Tony Cook TPF Grant 7 report 18.
• Tony Cook TPF Grant 7 report 19.
• Tony Cook TPF Grant 7 report 20.
• Tony Cook TPF Grant July 2016.
• Dave Mitchell TPF Grant 2 report 136.

Issues

New issues

• Perl #128951: heap-buffer-overflow in Perl_sv_vcatpvfn_flags.
• Perl #128952: (possible) stack-buffer-overflow in S_missingterm.
• Perl #128954: Small improvements for Hurd's hints.
• Perl #128966: Bleadperl v5.25.3-232-g10f9b9b breaks Const::Fast.
• Perl #128967: Inconsistency between doc and code for chown using negative argument -1.
• Perl #128972: Build failure on OS X 10.12 Sierra.
• Perl #128989: Bleadperl v5.25.3-266-g1d7e644 breaks VPIT/Variable-Magic-0.59.tar.gz.
• Perl #128995: Environment variables are not always propagated to sub-process.
• Perl #129027: null pointer deref Perl_mess_sv.
• Perl #129029: SIGBUS Perl_sv_peek.
• Perl #129036: Assertion Failure: S_parse_ident.
• Perl #129037: Assertion Failure: Perl_yylex.
• Perl #129038: Assertion Failure: *Perl_pp_subst.

Resolved issues

• Perl #78330: Stringifying NaN and Inf depends on the underlying C library.
• Perl #89142: On Win32 setting any encoding on STD(IN|OUT|ERR) with binmode causes subsequent fork calls to crash perl.
• Perl #95368: perl dump core.
• Perl #103566: blead h2xs.t failure on HP-UX 11.23 IA-64.
• Perl #126482: Assert fail w/o other symptoms - Perl_cv_const_sv_or_av.
• Perl #128775: [PATCH] ignore PERL_*_BACKEND variables during core builds.
• Perl #128951: heap-buffer-overflow in Perl_sv_vcatpvfn_flags.
• Perl #128952: (possible) stack-buffer-overflow in S_missingterm.
• Perl #128954: Small improvements for Hurd's hints.
• Perl #128967: Inconsistency between doc and code for chown using negative argument -1.
• Perl #128972: Build failure on OS X 10.12 Sierra.

Rejected issues

• Perl #94654: push doesn't work on blessed references.
• Perl #128950: Possible bug in perl \u string processing.
• Perl #129009: [Win32] UV_MAX is wrong on use 64bit int builds.

Suggested patches

Bulk88 provides an alternative patch in Perl #128930 (automatically detect GCC vs VC and set VC ver number on Win32) which was merged.

James E. Keenan provided a new patch for Perl #128856: DirHandle: Improve test coverage and provide descriptions., based on Tony Cook's feedback.

Karen Etheridge provided a patch in Perl #128987: to upgrade CPAN::Meta to 2.150010.

To address HP-UX problems, H. Merijn Brand (Tux) suggested (Getting HP-UX smokes green again) a patch.

Discussion

Jarkko Hietaniemi continues running Coverty on the codebase and found another group of errors relating to utf8 which are a bit odd. Aristotle Pagaltzis provided the background for the mentioned code.

If you're interested in how the lexing and parsing code of perl, Dave Mitchell provided an analysis of Perl #128951 (heap-buffer-overflow in Perl_sv_vcatpvfn_flags). Father Chrysostomos interjected to look into fixing it (which he did). I must include Dave's comment to Father C.:

Isn't it handy that people have such differing concepts of what constitutes 'fun'? Yes, please do take over :)

Perl #128950 (possible bug in perl \u string processing) showcases some of the complexities in Unicode, which I found very interesting. You might as well.

In an email forwarded from the perl-unicode mailing list, an issue is raised on the interaction between Encode's utf8 warnings and flags and the warning pragma's utf8 warnings.

James E. Keenan asked (lib/perldebug.t: puzzling test failure) about a puzzling test failure.