FAQ

[P5P] 5.10.1-RC2 is released

Dave Mitchell
Aug 19, 2009 at 12:01 am
I've just uploaded 5.10.1-RC2 to CPAN.

Once it's hit the mirrors it should be available as

http://www.cpan.org/authors/id/D/DA/DAPM/perl-5.10.1-RC2.tar.bz2

but until then, its available as

http://www.iabyn.com/tmp/perl-5.10.1-RC2.tar.bz2

but please don't advertise that outside of this list.

The .gz variant is also available.

Rafael, once its hit the mirrors, can you announce it on use.perl.org?

So please everyone, test away!

I'm not planning to change anything between now and final release unless
it's a real showstopper: I'm past caring about the occasional test
failure on some random platform.

Significant fixes since RC1:

NDBM_File sometimes wasn't built
Many library files were installed into the wrong directories
Deparsing was slow (fixed by removing const deparsing)
Safe could emit a spurious "used once" warning; fixed with a non-CPAN
version bump to 2.17_01
"cpan -r" didn't work. Fixed with a local (non-CPAN) fix.
allow Perl_die_nocontext() to be called with a NULL format
some Win build fixes that I didn't understand...
perlbug reported spurious patch lists
Configure added duplicate config args in config.sh
readline(BAREHANDLE) was broken
Module::Build version bumped to 0.340201 (having a devel version
number broke things)
perlfaq*.pod updated



--
A walk of a thousand miles begins with a single step...
then continues for another 1,999,999 or so.
reply

Search Discussions

14 responses

  • Rainer Tammer at Aug 19, 2009 at 6:00 am
    Hello,
    please could you add this to bleed and maint-5.10 ?
    It's only a documentation fix and a formating cleanup... this should not
    break anything :-))

    Thanks very much.

    Bye
    Rainer

    Dave Mitchell wrote:
    I've just uploaded 5.10.1-RC2 to CPAN.

    Once it's hit the mirrors it should be available as

    http://www.cpan.org/authors/id/D/DA/DAPM/perl-5.10.1-RC2.tar.bz2

    but until then, its available as

    http://www.iabyn.com/tmp/perl-5.10.1-RC2.tar.bz2

    but please don't advertise that outside of this list.

    The .gz variant is also available.

    Rafael, once its hit the mirrors, can you announce it on use.perl.org?

    So please everyone, test away!

    I'm not planning to change anything between now and final release unless
    it's a real showstopper: I'm past caring about the occasional test
    failure on some random platform.

    Significant fixes since RC1:

    NDBM_File sometimes wasn't built
    Many library files were installed into the wrong directories
    Deparsing was slow (fixed by removing const deparsing)
    Safe could emit a spurious "used once" warning; fixed with a non-CPAN
    version bump to 2.17_01
    "cpan -r" didn't work. Fixed with a local (non-CPAN) fix.
    allow Perl_die_nocontext() to be called with a NULL format
    some Win build fixes that I didn't understand...
    perlbug reported spurious patch lists
    Configure added duplicate config args in config.sh
    readline(BAREHANDLE) was broken
    Module::Build version bumped to 0.340201 (having a devel version
    number broke things)
    perlfaq*.pod updated


  • H.Merijn Brand at Aug 19, 2009 at 6:22 am

    On Wed, 19 Aug 2009 08:00:07 +0200, Rainer Tammer wrote:

    please could you add this to bleed and maint-5.10 ?
    It's only a documentation fix and a formating cleanup... this should not
    break anything :-))
    Applying: bleed and 5.10.1 - documentation fix - README.aix add info the lib gdbm is automatically removed + remove trailing blanks
    Thanks, patch successfully applied as e9d2e818e09bc1bed6b2109f74fcc75ceb03eb76 to blead

    For the rest, the non-whitespace diff was:

    diff --git a/README.aix b/README.aix
    index 12ca6ae..62d574e 100644
    --- a/README.aix
    +++ b/README.aix
    @@ -62,7 +62,8 @@ development team.

    If the AIX Toolbox version of lib gdbm 1.8.x is installed on the
    system then Perl will not work. This library contains a defect version
    -of the dbm_store() function.
    +of the dbm_store() function. The lib gdbm will be automatically removed
    +from the wanted libraries.

    =head2 Perl 5.10 was successfully compiled and tested on:

    --
    H.Merijn Brand http://tux.nl Perl Monger http://amsterdam.pm.org/
    using & porting perl 5.6.2, 5.8.x, 5.10.x, 5.11.x on HP-UX 10.20, 11.00,
    11.11, 11.23, and 11.31, OpenSuSE 10.3, 11.0, and 11.1, AIX 5.2 and 5.3.
    http://mirrors.develooper.com/hpux/ http://www.test-smoke.org/
    http://qa.perl.org http://www.goldmark.org/jeff/stupid-disclaimers/
  • Rafael Garcia-Suarez at Aug 19, 2009 at 7:34 am

    2009/8/19 Dave Mitchell <davem@iabyn.com>:
    Rafael, once its hit the mirrors, can you announce it on use.perl.org?
    Done now!
    http://use.perl.org/article.pl?sid=09/08/19/0731241
    Significant fixes since RC1: [...]
    Safe could emit a spurious "used once" warning; fixed with a non-CPAN
    version bump to 2.17_01
    You can bump that to 2.18, I'll release a CPAN version as soon as
    possible.
  • Dave Mitchell at Aug 19, 2009 at 10:08 am

    On Wed, Aug 19, 2009 at 09:33:59AM +0200, Rafael Garcia-Suarez wrote:
    You can bump that to 2.18, I'll release a CPAN version as soon as
    possible.
    Now bumped.

    --
    My Dad used to say 'always fight fire with fire', which is probably why
    he got thrown out of the fire brigade.
  • Curtis Jewell at Aug 19, 2009 at 11:55 pm
    The Strawberry Perl 5.10.1-RC2-0.beta-1 packages are now uploaded at
    http://strawberryperl.com/beta/ - only took 22 hours or so that time. :)

    --Curtis
    --
    Curtis Jewell
    swordsman@csjewell.fastmail.us

    %DCL-E-MEM-BAD, bad memory
    -VMS-F-PDGERS, pudding between the ears

    [I use PC-Alpine, which deliberately does not display colors and pictures in HTML mail]
  • Merijn Broeren at Aug 20, 2009 at 10:20 am

    Quoting Dave Mitchell (davem@iabyn.com):
    I've just uploaded 5.10.1-RC2 to CPAN.
    I get one test failure :

    Failed 1 test out of 1663, 99.94% okay.
    ../lib/parent/t/parent-pmc.t

    which is most likely due to the fact that I compiled it with
    -DPERL_DISABLE_PMC.

    Cheers,
    --
    Merijn Broeren | We take risks, we know we take them. Therefore, when things
    come out against us, we have no cause for complaint.
    - Scott, last journal entry, march 1912
  • Nicholas Clark at Aug 20, 2009 at 12:22 pm

    On Thu, Aug 20, 2009 at 12:19:54PM +0200, Merijn Broeren wrote:
    Quoting Dave Mitchell (davem@iabyn.com):
    I've just uploaded 5.10.1-RC2 to CPAN.
    I get one test failure :

    Failed 1 test out of 1663, 99.94% okay.
    ../lib/parent/t/parent-pmc.t

    which is most likely due to the fact that I compiled it with
    -DPERL_DISABLE_PMC.
    Yes, I discovered this yesterday, and I've mailed a proposed patch for
    that test to Max Maischein. Hence why there was also this patch:

    http://perl5.git.perl.org/perl.git/commit/f89a1e08e6fb41e93df685b7f838ff9bf916

    diff --git a/perl.c b/perl.c
    index 34fb282..9ef9cd7 100644
    --- a/perl.c
    +++ b/perl.c
    @@ -1770,6 +1770,9 @@ S_parse_body(pTHX_ char **env, XSINIT_t xsinit)
    # ifdef NO_MATHOMS
    " NO_MATHOMS"
    # endif
    +# ifdef PERL_DISABLE_PMC
    + " PERL_DISABLE_PMC"
    +# endif
    # ifdef PERL_DONT_CREATE_GVSV
    " PERL_DONT_CREATE_GVSV"
    # endif

    Nicholas Clark
  • Dave Mitchell at Aug 20, 2009 at 1:11 pm

    On Thu, Aug 20, 2009 at 01:22:29PM +0100, Nicholas Clark wrote:
    +# ifdef PERL_DISABLE_PMC
    Sigh. An undocumented compiler switch whose action is to disable an
    undocumented function....

    --
    SCO - a train crash in slow motion
  • Merijn Broeren at Aug 20, 2009 at 1:24 pm

    Quoting Nicholas Clark (nick@ccl4.org):
    On Thu, Aug 20, 2009 at 12:19:54PM +0200, Merijn Broeren wrote:
    Quoting Dave Mitchell (davem@iabyn.com):
    I've just uploaded 5.10.1-RC2 to CPAN.
    I get one test failure :

    Failed 1 test out of 1663, 99.94% okay.
    ../lib/parent/t/parent-pmc.t

    which is most likely due to the fact that I compiled it with
    -DPERL_DISABLE_PMC.
    Yes, I discovered this yesterday, and I've mailed a proposed patch for
    that test to Max Maischein.
    I submitted a patch through perlbug just now. I looked at the best way
    to skip this test but there doesn't seem to be a nice flag in Config.
    Checking other tests I found only one related to this flag (ther e might
    be more of course), t/comp/require.t. I adapted the code from that test
    to skip the parent-pmc tests in the patch that follows.

    I'm not sure skipping is the best solution, but it seems reasonable to
    do so for this test if pmc files are disabled. We disable them to gain a
    small performance advantage.

    My patch:

    --- /var/tmp/build.merijnb/parent-pmc.t 2009-08-20 08:37:43.000000000 -0400
    +++ ./lib/parent/t/parent-pmc.t 2009-08-20 08:36:20.000000000 -0400
    @@ -1,10 +1,18 @@
    #!/usr/bin/perl -w
    +
    +my $ccflags;
    BEGIN {
    if( $ENV{PERL_CORE} ) {
    chdir 't' if -d 't';
    chdir '../lib/parent';
    @INC = '..';
    }
    + # Right. We really really need Config here.
    + require Config;
    + die "Failed to load Config for some reason"
    + unless $Config::Config{version};
    + $ccflags = $Config::Config{ccflags};
    + die "Failed to get ccflags for some reason" unless defined $ccflags;
    }

    use strict;
    @@ -12,6 +20,7 @@
    use lib 't/lib';

    plan skip_all => ".pmc are only available with 5.6 and later" if $] < 5.006;
    +plan skip_all => ".pmc files are ignored in this build" if ($ccflags =~ /(?:^|\s)-DPERL_DISABLE_PMC\b/);
    plan tests => 3;

    use vars qw($got_here);

    Hence why there was also this patch:

    http://perl5.git.perl.org/perl.git/commit/f89a1e08e6fb41e93df685b7f838ff9bf916

    diff --git a/perl.c b/perl.c
    index 34fb282..9ef9cd7 100644
    --- a/perl.c
    +++ b/perl.c
    @@ -1770,6 +1770,9 @@ S_parse_body(pTHX_ char **env, XSINIT_t xsinit)
    # ifdef NO_MATHOMS
    " NO_MATHOMS"
    # endif
    +# ifdef PERL_DISABLE_PMC
    + " PERL_DISABLE_PMC"
    +# endif
    # ifdef PERL_DONT_CREATE_GVSV
    " PERL_DONT_CREATE_GVSV"
    # endif
    Yes, but that doesn't help with testing from inside a perl
    program, correct me if I'm wrong...

    Cheers,
    --
    Merijn Broeren | We take risks, we know we take them. Therefore, when things
    come out against us, we have no cause for complaint.
    - Scott, last journal entry, march 1912
  • H.Merijn Brand at Aug 20, 2009 at 1:49 pm

    On Thu, 20 Aug 2009 13:22:29 +0100, Nicholas Clark wrote:
    On Thu, Aug 20, 2009 at 12:19:54PM +0200, Merijn Broeren wrote:
    Quoting Dave Mitchell (davem@iabyn.com):
    I've just uploaded 5.10.1-RC2 to CPAN.
    I get one test failure :

    Failed 1 test out of 1663, 99.94% okay.
    ../lib/parent/t/parent-pmc.t

    which is most likely due to the fact that I compiled it with
    -DPERL_DISABLE_PMC.
    Yes, I discovered this yesterday, and I've mailed a proposed patch for
    that test to Max Maischein. Hence why there was also this patch:

    http://perl5.git.perl.org/perl.git/commit/f89a1e08e6fb41e93df685b7f838ff9bf916
    FWIW, I already included this in Config::Perl::V, so it will be in the
    next release
    diff --git a/perl.c b/perl.c
    index 34fb282..9ef9cd7 100644
    --- a/perl.c
    +++ b/perl.c
    @@ -1770,6 +1770,9 @@ S_parse_body(pTHX_ char **env, XSINIT_t xsinit)
    # ifdef NO_MATHOMS
    " NO_MATHOMS"
    # endif
    +# ifdef PERL_DISABLE_PMC
    + " PERL_DISABLE_PMC"
    +# endif
    # ifdef PERL_DONT_CREATE_GVSV
    " PERL_DONT_CREATE_GVSV"
    # endif

    Nicholas Clark

    --
    H.Merijn Brand http://tux.nl Perl Monger http://amsterdam.pm.org/
    using & porting perl 5.6.2, 5.8.x, 5.10.x, 5.11.x on HP-UX 10.20, 11.00,
    11.11, 11.23, and 11.31, OpenSuSE 10.3, 11.0, and 11.1, AIX 5.2 and 5.3.
    http://mirrors.develooper.com/hpux/ http://www.test-smoke.org/
    http://qa.perl.org http://www.goldmark.org/jeff/stupid-disclaimers/
  • Tom Hukins at Aug 20, 2009 at 1:49 pm

    On Wed, Aug 19, 2009 at 01:01:32AM +0100, Dave Mitchell wrote:
    I've just uploaded 5.10.1-RC2 to CPAN.
    Reading through its perldelta on search.cpan, I noticed a strange
    link:
    http://search.cpan.org/~dapm/perl-5.10.1-RC2/pod/perl5101delta.pod#DTrace_support

    The 'See "DTrace support" in INSTALL' link ends up at:
    http://search.cpan.org/~capttofu/DBD-drizzle-0.300/lib/DBD/drizzle/INSTALL.pod#DTrace_support

    Arguably, this quirk needs fixing on search.cpan itself, but many
    people read Perl's documentation this way.

    Given that INSTALL only ships with Perl's source, the link won't go
    anywhere meaningful for most people. I suggest keeping the text as-is
    but removing the link. Or maybe the link could go somewhere more
    suitable.

    Tom
  • Dave Mitchell at Aug 22, 2009 at 7:56 pm

    On Thu, Aug 20, 2009 at 02:49:31PM +0100, Tom Hukins wrote:
    On Wed, Aug 19, 2009 at 01:01:32AM +0100, Dave Mitchell wrote:
    I've just uploaded 5.10.1-RC2 to CPAN.
    Reading through its perldelta on search.cpan, I noticed a strange
    link:
    http://search.cpan.org/~dapm/perl-5.10.1-RC2/pod/perl5101delta.pod#DTrace_support

    The 'See "DTrace support" in INSTALL' link ends up at:
    http://search.cpan.org/~capttofu/DBD-drizzle-0.300/lib/DBD/drizzle/INSTALL.pod#DTrace_support

    Arguably, this quirk needs fixing on search.cpan itself, but many
    people read Perl's documentation this way.

    Given that INSTALL only ships with Perl's source, the link won't go
    anywhere meaningful for most people. I suggest keeping the text as-is
    but removing the link. Or maybe the link could go somewhere more
    suitable.
    Thanks for spotting this. I've removed the link as the easiest workaround.

    --
    This is a great day for France!
    -- Nixon at Charles De Gaulle's funeral
  • Craig A. Berry at Aug 21, 2009 at 2:21 am

    On Tue, Aug 18, 2009 at 7:01 PM, Dave Mitchellwrote:
    I've just uploaded 5.10.1-RC2 to CPAN.
    Totally uneventful on VMS. Default configuration with current
    compiler and OS version gives

    Failed 1 test out of 1554, 99.94% okay.
    [-.lib.Module.Build.t]debug.t


    which a test nit already fixed upstream and not worth worrying about.
  • Nicholas Clark at Aug 22, 2009 at 8:02 pm

    On Wed, Aug 19, 2009 at 01:01:32AM +0100, Dave Mitchell wrote:

    So please everyone, test away!
    All work's code works. (Which pulls in endless amounts of CPAN).

    This isn't a fair comparison, but it seems to be about 10% faster than the
    5.8.8 we're running on currently. It's not fair because I took the
    opportunity of this complete build from scratch to build without threads,
    without mathoms, and with -DPERL_DISABLE_PMC.

    Nicholas Clark

Related Discussions