Perl 5 Porters Mailing List Summary: May 30th - June 8th
Hey everyone,
Following is the p5p (Perl 5 Porters) mailing list summary for the past week. Enjoy!
May 30th - June 8th
News and updates
Dave Mitchell provides his grant report. The majority of the work was fixing Scope::Upper.
Dave Mitchell also provides more detailed grant reports.
Perl now warns if you accidentally left version control markers! See more in Perl #127993.
Issues
New issues
- Perl #128302:
split
docs don't mention pattern modifiers. - Perl #128295:
Strange bug with
-Dprefix=/usr
and-Dversion
. - Perl #128307: Bleadperl breaks perl-lisp.
- Perl #128311: Typo in perldoc.
- Perl #128313:
Memory leak in perl 5.24.0 when
use re qw[strict]
is used. - Perl #128340:
In-place
sort
incorrectly preserves element lvalue identity.
Resolved issues
- Perl #121734:
Cygwin:
setlocale(LC_ALL, "ja_JP.utf8")
panic. - Perl #127333:
until( assignment )
doesn't warn whenwhile( assignment )
does. - Perl #127915:
$=x~0
segfaults. - Perl #127993: Version control conflict markers.
- Perl #128001: Quadmath builds segfault on repeat with a very large count.
- Perl #128279:
hints/gnu.sh
: Mis-detected prototypes on GNU/Hurd. - Perl #128307: Bleadperl breaks perl-lisp.
- Perl #128316:
CORE::socket()
manglesEMFILE
and turns it intoEBADF
.
Proposed patches
Father Chrysostomos opened
Perl #128294
to skip tests of prereq modules in an internal utility,
bisect-runner.pl
. After discussions, this resulted in a patch
to provide a skip functionality flag, --no-modules-tests
.
Chad Granum provided a patch to update Test-Simple to the latest version, including silencing some noise and test failures. Chad also provided a patch in Perl #128337.
Sullivan Beck provided a patch to bump Locale::Codes from 3.38 to 3.39.
Following a conversation with Zefram and Tony Cook on the IRC channel, Breno G. de Oliveira provided a patch in Perl #128301 to test which operations "eagerly" create a hash entry. Zefram notes there are several other cases which we should cover in tests, while we're addressing this, and provides sample code.
Dominic Hargreaves sent a patch to address an assertion failure.
Discussion
Smylers
notes
that perlpodspec
still promotes
Pod::Parser despite being
cleaned up by Karl Williamson as part of
Perl #119439.
Karl Williamson responds with further explanations on some of the
finer points of replacing those bits in the docs in favor of
Pod::Simple. Seems a bit
difficult.
Smylers created
Perl #128302
mentioning that the documentation for split
do not mention that
it can use pattern modifiers. The
explanation
(by Father Chrysostomos) is that a m//
op is created at compile
time, which itself can have pattern modifiers. Abigail
notes
that some (for example, /g
) produce a warning.
Andy Dougherty
notes
that the change in Configure to make -O
the default behavior
has unintentionally introduced some breakage. Andy analyzes the
problem and asks, was this worth it? You can find more about
the problems raised in
Perl #128295.
ExtUtils::MakeMaker 7.18 has breaking tests on Windows, both in Jenkins and smoking. This was eventually resolved upstream.
An aside from usual business, a user emailed about possible performance degradation in a script using OpenSSL. This was debugged as a benchmark problem, an OpenSSL function call, and a network issue.
Andy Lester provides a patch to fix a typo in perlhacktips
pod
in
Perl #128331.
Karl Williamson asks to take a look at perltodo and update it.
Karl Williamson suggested deprecating the ability to quantify zero-length constructs in regular expressions. Why quantify nothing? Abigail notes that if you're generating code, deprecating this is not desired.
Father Chrysostomos asks for help debugging a program which seldom crashes. Lessons learned? There are various tools: gdb, valgrind, clang with address sanitizer, and rr.
Dan Collins sparked a conversation about the possible problems with in-place sort optimization. You can read additional thoughts in a ticket Zefram opened following this, Perl #128340.
Dan Collins had also been working on testing some old bugs. After verifying Perl #128313, he worked on a solution. You see his progress here.
Leave a comment