David Mitchell [Wed, 12 Aug 2009 00:01:34 +0000]
ameliorate B::Deparse slowdown
commit
2990415a45 improved the ability to deparse inlined constants,
but at the cost of having to walk all the symbol tables when each new
B::Deparse object is created. Make this scan instead only happen the first
time its needed.
Jan Dubois [Tue, 11 Aug 2009 23:30:32 +0000]
On Windows normalize $^X using GetLongPathName()
If perl.exe is called with a short pathname, then GetModuleFileName()
will return this short name, and $^X will be set to it. This in turn
is used to initialize @INC to privlib, sitelib and vendorlib locations
relative to $^X, so they too will end up with the mangled short names.
Perl will also automatically add versioned Perl directories in the
same tree if their names start with the same major and minor Perl
version numbers. This heuristic can be broken when the pathname
components are using short pathnames. Therefore $^X and @INC should
all be normalized to use the long pathname format.
See also http://rt.cpan.org/Public/Bug/Display.html?id=47890
David Mitchell [Tue, 11 Aug 2009 22:52:38 +0000]
s/PERL5OPTS/PERL5OPTS/ in INSTALL
spotted by Vincent Pit
Ken Williams [Tue, 11 Aug 2009 16:25:07 +0000]
Removed '...' ellipses from docs about '...' flip-flop operator
David Mitchell [Tue, 11 Aug 2009 22:09:01 +0000]
INSTALL: mention that env vars can affect tests
David Mitchell [Tue, 11 Aug 2009 21:51:16 +0000]
release_managers_guide: mention updating cpan.org/src/
David Mitchell [Tue, 11 Aug 2009 20:43:02 +0000]
PERL_DONT_CREATE_GVSV was documented in perl589delta
not perl588delta (although it was introduced in 5.8.8).
Fix up old perldelta files to reflect this fact.
This has been a public service by the Ministry of Truth.
Max Maischein [Tue, 11 Aug 2009 16:22:58 +0000]
Restore t/harness path checks on Win32
Message-ID: <
4A819742.8050109@corion.net>
Jerry D. Hedden [Tue, 11 Aug 2009 14:15:16 +0000]
Reduce thread count in threads::shared stress test
Need to reduce the thread count in t/stress.t for threads::shared.
Signed-off-by: H.Merijn Brand <h.m.brand@xs4all.nl>
Jerry D. Hedden [Mon, 10 Aug 2009 19:26:25 +0000]
Upgrade to threads::shared 1.31
Attached patch changes t/stress.t such that it more robustly handles
thread creation failures due to 'no more processes', 'out of memory',
etc.. Such failures are not indicative of problems with the
'threads::shared' module.
From
2440c97ff633f0a33bf1ca0cceecfd4448cd6167 Mon Sep 17 00:00:00 2001
From: Jerry D. Hedden <jdhedden@cpan.org>
Date: Mon, 10 Aug 2009 15:22:54 -0400
Subject: [PATCH] Upgrade to threads::shared 1.31
Signed-off-by: H.Merijn Brand <h.m.brand@xs4all.nl>
Jerry D. Hedden [Mon, 10 Aug 2009 19:52:34 +0000]
MANIFEST and Porting/Maintainers.pl update
Forgot to add new test to MANIFEST for threads
Forgot to update version for threads::shared in Maintainers.pl
Patches attached.
From
ea5eebe94489df90ab780b07e73a4df875a01705 Mon Sep 17 00:00:00 2001
From: Jerry D. Hedden <jdhedden@cpan.org>
Date: Mon, 10 Aug 2009 15:48:36 -0400
Subject: [PATCH 1/2] Add no_threads.t to MANIFEST
Signed-off-by: H.Merijn Brand <h.m.brand@xs4all.nl>
H.Merijn Brand [Mon, 10 Aug 2009 19:29:11 +0000]
Update threads to 1.74
Jerry D. Hedden [Mon, 10 Aug 2009 19:08:29 +0000]
Upgrade to 'threads' 1.74
Attached patch upgrade 'blead' to 'threads' 1.74:
- Updated DESCRIPTION in POD
- Added 'no_threads' test for non-threaded Perls
The newly added test allows non-threaded Perls to install 'threads'
with a 'PASS' instead of the ambiguous 'UNKNOWN'.
Signed-off-by: H.Merijn Brand <h.m.brand@xs4all.nl>
David Mitchell [Mon, 10 Aug 2009 08:35:05 +0000]
bump Module::Build in Maintainers.pl
David Golden [Mon, 10 Aug 2009 02:13:50 +0000]
Bump Module::Build version to 0.340201
Module::Build automatically sets configure_requires to its
current $VERSION, which causes problems if $VERSION is a dev
release not on CPAN. To stabilize 5.10.1, this patch bumps
Module::Build 0.34_02 (currently in 5.10.1-RC1) to 0.340201.
No other code changes are included. A matching version will
be released to CPAN.
Abigail [Sun, 9 Aug 2009 21:27:35 +0000]
Add 5.10.1-RC1 release to perlhist
Shlomi Fish [Sun, 9 Aug 2009 16:44:19 +0000]
Add a regression test for RT #68182.
readdir($x=".") segfaults on Linux.
Nicholas Clark [Sun, 9 Aug 2009 09:44:28 +0000]
Run the main tests "slowest first" by ordering the rules correctly.
(It seems that the current TAP::Harness implementation sequences based on the
parallel/serial rules, not the order of the list of test names presented to it.)
Nicholas Clark [Sun, 9 Aug 2009 08:16:44 +0000]
Inline _seq_dir_rules(), as it is now used in only one place.
Nicholas Clark [Sat, 8 Aug 2009 14:28:31 +0000]
Merge gv_IOadd() into gv_add_by_type().
Nicholas Clark [Sat, 8 Aug 2009 14:01:48 +0000]
Perl_newIO() can become a mathom by making newIO() a wrapper around newSV_type()
and tweaking Perl_sv_upgrade().
Nicholas Clark [Sat, 8 Aug 2009 11:05:40 +0000]
Merge gv_AVadd(), gv_HVadd() and gv_SVadd() into gv_add_by_type().
The "short" names become macro wrappers, and the Perl_* versions become mathoms.
Nicholas Clark [Sat, 8 Aug 2009 09:20:40 +0000]
Move the "types are equal" early return ahead of the COW-removal.
Steffen Mueller [Sat, 8 Aug 2009 11:01:37 +0000]
Synchronize with CPAN's Data::Dumper 2.125
No significant code changes. Improvements to the CPAN distribution.
Alexandr Ciornii [Sat, 8 Aug 2009 11:00:30 +0000]
Move variable declaration to where it's used
Alexandr Ciornii [Sat, 8 Aug 2009 10:58:10 +0000]
Spellcheck
Steffen Mueller [Sat, 8 Aug 2009 10:56:23 +0000]
Synchronize with CPAN's Attribute::Handlers 0.86
No functional changes, only CPAN distribution upgrades.
Steffen Mueller [Sat, 8 Aug 2009 10:55:18 +0000]
Synchronize AutoLoader with CPAN's 5.69
No functional changes, only CPAN distribution upgrades.
Steffen Mueller [Sat, 8 Aug 2009 10:54:24 +0000]
Update some distributions in Porting/Maintainers.pl
New (hardly changed) versions of Data::Dumper, AutoLoader,
Attribute::Handlers.
Nicholas Clark [Sat, 8 Aug 2009 08:40:42 +0000]
In pp_caller, remove a GvMULTI_on() by adding GV_ADDMULTI to gv_fetchpvs().
Nicholas Clark [Fri, 7 Aug 2009 20:44:00 +0000]
Remove GvREFCNT_inc(), which is deprecated and unused.
Rainer Tammer [Fri, 7 Aug 2009 12:29:07 +0000]
5.10.1-RC1 is released
Hello,
I have attached a small fix for README.aix (5.10.1 / blead).
Please could you commit this?
Fixes in this patch:
* correct some typos
* adjust some dates
Bye
Rainer
Dave Mitchell wrote:
> I've just uploaded 5.10.1-RC1 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-RC1.tar.bz2
>
> but until then, its available as
>
> http://www.iabyn.com/tmp/perl-5.10.1-RC1.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.
>
> Also note that I'm away this weekend.
>
>
> (*) This is my first ever use of PAUSE
>
>
From
4d658a15475a7866f56ff938d851e6df719c056e Mon Sep 17 00:00:00 2001
From: Rainer Tammer <tammer@tammer.net>
Date: Fri, 7 Aug 2009 14:25:29 +0200
Subject: [PATCH] Small fix for READEM.aix
Signed-off-by: H.Merijn Brand <h.m.brand@xs4all.nl>
Rafael Garcia-Suarez [Fri, 7 Aug 2009 11:25:21 +0000]
perltodo: Propagate compilation hints to the debugger
Suggested by Ben Morrow
Jerry D. Hedden [Fri, 7 Aug 2009 09:28:15 +0000]
t/safeuniversal.t failure under 5.8.9
[RT CPAN #48505]
The "no warnings 'redefined'" directives in the compartments
in t/safeuniversal.t cause test failures under 5.8.9 because
warnings.pm has a 'use Carp' directive and Carp.pm uses
'eval' to 'require Carp::Heavy'.
The attached patch replaces the "no warnings 'redefined'"
directive with a $SIG{__WARN__} trap for 5.8.9. While the
redefine warning for UNIVERSAL::can is suppressed by the
trap, the warning UNIVERSAL::isa still shows up. However,
the failing test will now pass under 5.8.9.
Father Chrysostomos [Fri, 7 Aug 2009 08:10:31 +0000]
[perl #68108] : also fix if/else constant folding
Sam Vilain [Fri, 31 Jul 2009 03:32:14 +0000]
pod/perlipc.pod: add some hints on avoiding pipe deadlocks
Tracking down deadlocks when using pipes for IPC can be hard, so put
even more notes about gotchas in this section of perlipc.
Rafael Garcia-Suarez [Thu, 6 Aug 2009 20:12:47 +0000]
Make gv_IOadd accept a NULL argument
This fixes bug #68182
Jesse Vincent [Thu, 6 Aug 2009 13:16:58 +0000]
broke out checkAUTHORS aliasing for our two "merijnb" porters
Hopefully I got it right
Jesse Vincent [Thu, 6 Aug 2009 13:13:42 +0000]
Minor typo and formatting fixes to the release manager guide
David Mitchell [Thu, 6 Aug 2009 02:06:48 +0000]
more release_managers_guide.pod tweaking
David Mitchell [Thu, 6 Aug 2009 00:52:44 +0000]
Porting/makerel: handle more sha1 variants,
and be quieter on failure
David Mitchell [Thu, 6 Aug 2009 00:01:07 +0000]
release_managers_guide: tweak ordering of 'on-the-day' stuff
and lots of little fixes therein.
David Mitchell [Wed, 5 Aug 2009 23:34:28 +0000]
Updated Module::CoreList for the 5.10.1 release
David Mitchell [Wed, 5 Aug 2009 20:40:55 +0000]
fix for EU:MM 6.55_02 failing test,
as suggested by Schwern
David Mitchell [Wed, 5 Aug 2009 19:46:32 +0000]
ExtUtils::MakeMaker 6.55_02
David Mitchell [Wed, 5 Aug 2009 16:24:58 +0000]
The current CPAN is very noisy on download failure;
it's showing a full perl stack trace. temporary fix for now.
David Mitchell [Wed, 5 Aug 2009 14:25:58 +0000]
Revert "add CLONEf_KEEP_PTR_TABLE to win fork emulation."
This reverts commit
240bfeb992996de28603a7f9e3ee34ecc6e3f286.
(its now fixed in a better way in
136e7e597c0d85149921561591710bc54123b221)
David Mitchell [Wed, 5 Aug 2009 14:23:41 +0000]
in perl_clone_using, free the prt_table
*after* we have called any CLONE methods
(better fix for Variable::Magic than
240bfeb992)
Jesse Vincent [Wed, 5 Aug 2009 13:31:10 +0000]
Now down to 17 unknown addresses which "should" be in the authors file.
Extensive googling and other sleuthing helped resolve the names of many
people previously listed as "unknown"
David Mitchell [Wed, 5 Aug 2009 13:20:29 +0000]
add CLONEf_KEEP_PTR_TABLE to win fork emulation.
This keeps the ptr-table table around after the thread has been created.
This has been on for user threads for a while now, but not for
Windows forks. Turns out that Variable::Magic relies on the table still
being there, so crashes under windows forks. This increases memory usage
per fork, but stops things crashing. Sigh.
[perl #66158]
David Mitchell [Wed, 5 Aug 2009 13:18:37 +0000]
win32/perlhost.h: use symbolic constant
CLONEf_COPY_STACKS rather than 1
David Mitchell [Tue, 4 Aug 2009 22:08:59 +0000]
release_managers_guide: add patchlevel.h stuff
David Mitchell [Tue, 4 Aug 2009 21:15:27 +0000]
minor tweaks to release_managers_guide.pod
Jesse Vincent [Tue, 4 Aug 2009 17:50:48 +0000]
The AUTHORS file is now short only about thirty 'real names'
David Mitchell [Tue, 4 Aug 2009 16:49:11 +0000]
release_managers_guide: make into two logical sections:
those to be done in the week or two before a release,
and those to be done on the day
Vincent Pit [Tue, 4 Aug 2009 14:13:28 +0000]
Promote blocks resulting from constant folding to first-class do { } blocks
This solves [perl #68108]: no retval from sub { if(1){ ... } }
Reini Urban [Tue, 4 Aug 2009 11:12:24 +0000]
WIFEXITED() cygwin regression
t/run/exit.t:58: eval { POSIX::WIFEXITED() };
fails on the new cygwin-1.7 with
5 [main] perl 3500 fork: child -1 - died waiting for longjmp
before initialization, retry 0, exit code 0x
C0000005, errno 11
...
until I kill perl, because you must use an argument for WIFEXITED.
It didn't fail before, so I assume it's a cygwin regression, but it
must be fixed in the perl testsuite,
since an empty argument cries for trouble.
Use 0 or 1 as argument, ${^CHILD_ERROR_NATIVE} preferred.
Jesse Vincent [Tue, 4 Aug 2009 10:21:29 +0000]
Removed bald-faced lies I inserted about the mergelog from the release_managers_guide
Jesse Vincent [Tue, 4 Aug 2009 08:43:18 +0000]
Improve the broken order of the release_managers_guide
From: Dave Mitchell <davem@iabyn.com>
To: jesse@fsck.com
Cc: perl5-porters@perl.org
Subject: recent changes to Porting/release_managers_guide.pod
Message-ID: <
20090803225903.GA4204@iabyn.com>
Jesse Vincent [Tue, 4 Aug 2009 08:38:30 +0000]
Fix corelist generation section of the release manager documentation.
From: Dave Mitchell <davem@iabyn.com>
Subject: recent changes to Porting/release_managers_guide.pod
Message-ID: <
20090803225903.GA4204@iabyn.com>
Jesse Vincent [Mon, 3 Aug 2009 22:15:31 +0000]
Now down to 50 addresses in the git logs who aren't in the AUTHORS file,
All have only one patch. Many of them are junk addresses.
Jesse Vincent [Mon, 3 Aug 2009 21:52:26 +0000]
Continued work to get Porting/checkAUTHORS.pl to a point of sanity
David Mitchell [Mon, 3 Aug 2009 20:10:50 +0000]
bump-perl-version: over-keen false-positive removal
commit
5bd03515d294382b9bededef5a3bb4cd415656c3 got a bit carried
away; these files may well need version bumping
David Mitchell [Mon, 3 Aug 2009 19:33:17 +0000]
Maintainers.pl: update for ExtUtils::Install
(folloup to
16f0d0fcc43b16d1afa4a05a4599655310733b2c)
Jesse Vincent [Mon, 3 Aug 2009 15:21:26 +0000]
Further work to cut down the unattributed authors in the AUTHORS file.
We're now down to about 100 unattributed authors
Jesse Vincent [Mon, 3 Aug 2009 15:06:30 +0000]
Remove the README for the now-gone machten port
Vincent Pit [Mon, 3 Aug 2009 13:51:03 +0000]
Minor spelling nit in last perlfunc updated
Jesse Vincent [Mon, 3 Aug 2009 12:37:03 +0000]
Yuval Kogman's address was special cased for non-inclusion in authors for no reason
Yuval confirmed for me that he would like his email address included
in AUTHORS
Jesse Vincent [Sat, 1 Aug 2009 21:09:41 +0000]
Removed vestigal support for Tenon Intersystems MachTen Unix layer for MacOS Classic.
Perl hasn't supported MachTen since 5.6.
Jesse Vincent [Sat, 1 Aug 2009 18:42:50 +0000]
Remove the port to MiNT. It's a dead platform that hasn't had any love since 5.005
Jesse Vincent [Mon, 3 Aug 2009 11:06:00 +0000]
Update the description of the 'system' function to point users at autodie
pjf++ for complaining about the POD for system during his Klingon
Programming talk at YAPC::EU 2009
Jesse Vincent [Mon, 3 Aug 2009 10:39:49 +0000]
MacPerl is gone. Its test bits should be now removed (per nicholas)
Jesse Vincent [Mon, 3 Aug 2009 10:39:07 +0000]
Document what needs to happen when you add a new test directory
Jesse Vincent [Mon, 3 Aug 2009 10:33:23 +0000]
Add the new "porting" test directory to the test harness
(per advice from rgs)
Jesse Vincent [Sun, 2 Aug 2009 15:31:09 +0000]
Move content of Todo.micro into perltodo
Jesse Vincent [Sat, 1 Aug 2009 21:03:24 +0000]
Y2038 doc pointers in the little helper C tools
Jesse Vincent [Sat, 1 Aug 2009 20:57:53 +0000]
Minor typo fix in a comment
Jesse Vincent [Sat, 1 Aug 2009 18:39:24 +0000]
Be a bit clearer about always needing to regen perl.pod and friends at release time
Jesse Vincent [Sat, 1 Aug 2009 17:41:40 +0000]
Flipped checkcase.pl from a porting tool to an actual test file
Jesse Vincent [Sat, 1 Aug 2009 20:29:26 +0000]
buildtoc should be executable
Jesse Vincent [Sat, 1 Aug 2009 17:22:32 +0000]
Reoved the suggestion that we might want to consider integrating MacPerl from pumpkin.pod
Jesse Vincent [Mon, 3 Aug 2009 08:55:14 +0000]
Updated AUTHORS to include all contributors with more than two patches.
Now that we have lots and lots of pre-p4 history, we know about many
more porters whose identities had previously been lost in the mists
of time.
Patches to porting/checkAUTHORS.pl and AUTHORS to add the remaining
missing porters or mappings would be most welcome.
Jesse Vincent [Sun, 2 Aug 2009 15:25:12 +0000]
Updated Ingy's entry in Authors to match his legal name change
Jesse Vincent [Mon, 3 Aug 2009 08:54:45 +0000]
Updated the release_managers_guide with information about how to run checkAUTHORS
Jesse Vincent [Mon, 3 Aug 2009 08:51:48 +0000]
Udpdated checkAUTHORS.pl to expect the output of git log rather than p4; added a bunch of author email equivalences
Jesse Vincent [Sun, 2 Aug 2009 15:17:08 +0000]
Clean up the new maintainers.t to be a bit more sensible about paths
The previous version of the invocation ignore the fact that we chdir,
more libs might be required.
Josh ben Jore [Sat, 1 Aug 2009 05:36:17 +0000]
Pre-resolve symlink for File::Stat test under -Dmksymlinks
Jesse Vincent [Sat, 1 Aug 2009 02:36:35 +0000]
Removed a release-engineering step. (It's been replaced with automated tests)
Jesse Vincent [Fri, 31 Jul 2009 17:17:57 +0000]
made the manicheck actions (previously releng) into tests that always run
Jesse Vincent [Fri, 31 Jul 2009 15:44:24 +0000]
Fixed maintainership for lib/ExtUtils/t/InstallWithMM.t
per http://www.nntp.perl.org/group/perl.perl5.porters/2009/07/msg149219.html
Jesse Vincent [Fri, 31 Jul 2009 15:28:36 +0000]
Linearized the release-manager's guide to make it less of a choose-your-own-adventure novel
Jesse Vincent [Fri, 31 Jul 2009 14:34:17 +0000]
Copyediting the release_managers_guide.
No functional changes
Jesse Vincent [Fri, 31 Jul 2009 13:58:50 +0000]
Start refactoring out release management steps that aren't actually tied to building a release
Craig A. Berry [Fri, 31 Jul 2009 10:42:47 +0000]
Clean up the .exists files on VMS.
Steffen Mueller [Fri, 31 Jul 2009 09:14:26 +0000]
Mention in apidocs that av_push takes ownership of a refcount
Nicholas Clark [Fri, 31 Jul 2009 08:12:00 +0000]
Add Porting/GitUtils.pm to MANIFEST
Nicholas Clark [Fri, 31 Jul 2009 08:04:37 +0000]
Add lib/version/Internals.pod to MANIFEST and Porting/Maintainers.pl
Rafael Garcia-Suarez [Thu, 30 Jul 2009 21:19:13 +0000]
Add new error "Can't use keyword '%s' as a label"
Nicholas Clark [Thu, 30 Jul 2009 20:41:14 +0000]
Try longst prefixes first, else lib/CPAN may match lib/CPANPLUS/... and similar.
Jesse Vincent [Thu, 30 Jul 2009 05:10:52 +0000]
'git st' is not a standard git command.
Spotted by Tux++