develooper Front page | perl.perl5.porters | Postings from March 2014

Perl 5.20.0 Blockers, 2014-03-31

Thread Next
Ricardo Signes
March 31, 2014 15:02
Perl 5.20.0 Blockers, 2014-03-31
Message ID:
Today is March 31, 2014. There are approximately 50 days until the expected
release of Perl 5.20.0.  A summary of blockers follows.

## Critical Behavior Changes or Bugs

1.  $!, UTF8, and locales
2.  Gconvert() obeys LC_NUMERIC without "use locale"

    I believe both these changes should be reverted, restoring the 5.18
    behavior.  Karl has suggested more substantial changes to make in early
    5.21, which I believe will be the right thing.

3.  "Make perlbug Unicode-aware" broke perlbug on Win32

    I'd rather not revert the change, but we need this to work on Win32.  There
    has been some discussion about the fix, but so far it's not fixed.

## Other Bugs

4.  IO::Socket getsockopt subroutine on AIX

    We have a patch.  It needs review, application, and testing.  Still.

5.  p /x in debugger yields memory fault

    This real bug is still a real bug, in need of debugger debugging.

## BBC Reports

6.  PERLER/JavaScript-Dumper-0.011.tar.gz

    This is part of locale fixes for stringifying numbers.  My gut instinct is
    to think that this is bound to be something for PERLER to fix downstream.

    Anyone else?

7.  YVES/Data-Dump-Streamer-2.37.tar.gz

    YVES said he'd follow up.  Hey, Yves, what's the word? :)

8.  RYBSKEJ/forks-0.34.tar.gz

    I have no idea what's up here.  Broken by "Make set-magic handle vstrings
    properly."  This needs eyes on it.

9. does not set new filehandles into ReadLine after pager

    Just patches to be applied!

10. .pmc not loaded if @INC has a trailing slash

    Matthew Horsfall has supplied a patch to fix this, awaiting review and

11. Can't build perl with gcc4.9 due to UB in sv.c

    sv.c is hitting undefined behavior.  Patch supplied!  Needs review.

## Minor Issues

12. speed up miniperl @INC, buildsitecustomize

    Without definitive review, this seems bound for 5.21 instead.

13. core tests fail when built on FAT drive with Win32 OS on NTFS

    I think the actual course of action for 5.20.0 is to add a Known Problems
    entry.  Thoughts?

## To Observe Only

These don't block 5.20.0, but are collected as things to keep an eye on.

13. New copy-on-write breaks CPAN

    I think that this has become an informational ticket.  There are still some
    things broken, but "that's their problem."  We would like to see them
    fixed, and will help if we can, but this isn't really *blocking* the next

14. Our Asian documentation should be updated

    This is informational for me.  I am going to make a renewed call for the
    un-updated documentation *today*.  No matter what, I will close this ticket
    (possibly replacing it with more specific tickets) at or around 5.20.0.


Thread Next Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at | Group listing | About