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

Perl 5.20.0 Blockers, 2014-04-21

Thread Next
From:
Ricardo Signes
Date:
April 21, 2014 14:51
Subject:
Perl 5.20.0 Blockers, 2014-04-21
Message ID:
20140421145052.GA12869@cancer.codesimply.com
Today is April 21, 2014. There are approximately 29 days until the expected
release of Perl 5.20.0.

My plan is roughly as follows:

* finish editing Porting/perl5200delta.pod
* bump version to 5.20.0 RC in a few days
* continue to post blocker emails
* focus on test reports with builds versioned 5.20.0
* once it's soup, serve it
* tropical vacation

Below, I list only the bugs that are actually potential blockers:

1.  VC6 has missing ')'s on macro expansion, cBOOL and U8
    https://rt.perl.org/Ticket/Display.html?id=121554

    bulk88 has provided a suggested fix.  Meantime, looks like blead won't
    compile on VS6, which we still propose to support.  Are there reasons not
    to fix this as suggested?

    Steve H. has written to Dan Kogai about the patch, and I will follow up on
    that.  Without a reply, I think we will have to patch directly in blead.

2.  IO::Socket getsockopt subroutine on AIX
    https://rt.perl.org/Ticket/Display.html?id=120835

    Without testing or fixing, this will get called a Known Problem for 5.20.0.
    I think that's fine.

3.  Test failure in t/run/switchM.t with -DPERL_DISABLE_PMC
    https://rt.perl.org/Ticket/Display.html?id=121662

    and

    Avoid double quotes in switchM.t runperl args.
    https://rt.perl.org/Ticket/Display.html?id=121672

    We've got patches, and Tony and I both think they look good to apply.
    These blockers should be pretty short-lived!  (I leave it for Tony to
    apply, as he thinks they will need some tweaking.)

4.  Can't build perl with gcc4.9 due to UB in sv.c
    https://rt.perl.org/Ticket/Display.html?id=121505

    We (seem to) have patches for this, to address the problem by disabling the
    optimization, which seems like a simple solution for the time being.  Reini
    is concerned that it's papering around the actual problem of relying on
    undefined behavior, and suggests that we address that problem instead.
    Zefram's initial take, as I recall, was that the patch was insufficient to
    actually solve the problem.  It seems to me like the path forward is
    "disable optimization for 5.20.0, debate more extensive action after that."

-- 
rjbs

Thread Next


nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About