14 years agoWarn people not to set PERL_CORE in XS
David Mitchell [Mon, 8 Feb 2010 21:07:56 +0000]
Warn people not to set PERL_CORE in XS

14 years agoIn embed.fnc the 'E' flag should imply 'X', but doesn't.
Jan Dubois [Mon, 8 Feb 2010 03:06:50 +0000]
In embed.fnc the 'E' flag should imply 'X', but doesn't.

14 years agorevise perldelta note on strict and lax version rules
David Golden [Mon, 8 Feb 2010 02:25:20 +0000]
revise perldelta note on strict and lax version rules

14 years agoexpand flag descriptions in in embed.fnc
David Mitchell [Sun, 7 Feb 2010 20:12:18 +0000]
expand flag descriptions in in embed.fnc

For each flag, try to list what precise effects that flag has

14 years agoalphabetically sort flag descriptions in embed.fnc
David Mitchell [Sun, 7 Feb 2010 18:44:34 +0000]
alphabetically sort flag descriptions in embed.fnc

14 years ago[perl #72156]: substitute `less -R' for `less' for groff's new usage
Josh ben Jore [Mon, 1 Feb 2010 16:05:13 +0000]
[perl #72156]: substitute `less -R' for `less' for groff's new usage
of ANSI escape codes by setting $Config{less} (and thereby
$Config{pager}).

14 years agoINSTALL doc fix: calling the test suite using -MTestInit changed since it moved to...
Slaven Rezic [Sat, 6 Feb 2010 22:14:56 +0000]
INSTALL doc fix: calling the test suite using -MTestInit changed since it moved to top level (which happened in commit 30b6e59101b252c20a6b50b95203d1e2c6016604)

14 years agochanged "an UTF..." to "a UTF..." in all Pods
Slaven Rezic [Sat, 6 Feb 2010 22:19:39 +0000]
changed "an UTF..." to "a UTF..." in all Pods

14 years agoRemoving invalid port L<> links from Version documentation updates.
Jesse Vincent [Sun, 7 Feb 2010 02:06:56 +0000]
Removing invalid port L<> links from Version documentation updates.

14 years agoMerge branch 'blead' of ssh://perl5.git.perl.org/gitroot/perl into blead
Jesse Vincent [Sun, 7 Feb 2010 02:02:40 +0000]
Merge branch 'blead' of ssh://perl5.git.perl.org/gitroot/perl into blead

* 'blead' of ssh://perl5.git.perl.org/gitroot/perl:
  move version details to version::Internals and other clean up
  document version::is_strict/is_lax
  Document usage of version regexps
  Export and document is_lax and is_strict functions
  note that delete/exists ARRAY_ELEM should be avoided
  Don't try to calculate a time over the conservative failure boundary.

14 years agoImprovements to 31c9a3 - CPAN code did depend on the previous behaviour of blessing...
Nicholas Clark [Sun, 7 Feb 2010 01:55:32 +0000]
Improvements to 31c9a3 - CPAN code did depend on the previous behaviour of blessing filehandles into FileHandle

It turns out that it's not quite as simple as blessing into IO::File.
If you do (just) that, then it breaks any existing code that does
C<require IO::Handle;> to allow it to call methods on file handles,
because they're blessed into IO::File, which isn't loaded. (Note this code
doesn't assume that methods in IO::Seekable are there to be called)

So, it all should work if you also set @IO::File:::ISA correctly?
That way, code that assumes that methods from IO::Handle can be called will
work. However, gv.c now starts complaining (but not failing) if IO::Handle,
IO::Seekable and Exporter aren't present, because it goes looking for
methods in them.

So the solution seems to be to set @IO::File::ISA *and* create (empty)
stashes for the other 3 packages. Patch appended, but not applied.

14 years agomove version details to version::Internals and other clean up
David Golden [Sat, 6 Feb 2010 20:09:20 +0000]
move version details to version::Internals and other clean up

14 years agodocument version::is_strict/is_lax
David Golden [Sat, 6 Feb 2010 19:12:21 +0000]
document version::is_strict/is_lax

14 years agoDocument usage of version regexps
John Peacock [Wed, 27 Jan 2010 01:47:54 +0000]
Document usage of version regexps

Move the discussion of what each regexp coveres to version::Internals
and limit the discussion in the main POD to just include examples.

14 years agoExport and document is_lax and is_strict functions
John Peacock [Tue, 26 Jan 2010 02:49:55 +0000]
Export and document is_lax and is_strict functions

Allow the is_lax and is_strict functions to be optionally
exported to the caller's namespace (without having to use Exporter).

Document the usage of is_lax and is_strict, plus nuke some trailing
spaces.

14 years agonote that delete/exists ARRAY_ELEM should be avoided
Ricardo Signes [Sat, 6 Feb 2010 18:17:01 +0000]
note that delete/exists ARRAY_ELEM should be avoided

addreses final determination of http://rt.perl.org/rt3/Public/Bug/Display.html?id=72064

14 years agoDon't try to calculate a time over the conservative failure boundary.
Michael G. Schwern [Sun, 31 Jan 2010 11:22:08 +0000]
Don't try to calculate a time over the conservative failure boundary.

Otherwise gmtime(2**66) will cause a very, very, very long loop and
DOS Perl.

Add a test that very, very large times don't send gmtime and localtime into a loop

Had to fix some revealed mistakes in op/time.t when warnings were turned on.

Fix Time::gmtime and Time::localtime tests to match the new limits of gm/localtime.

14 years agoUpgrade CPAN to 1.94_55
Steffen Mueller [Thu, 4 Feb 2010 19:18:45 +0000]
Upgrade CPAN to 1.94_55

14 years agoRevert "Update to MIME-Base64 3.09"
Jesse Vincent [Wed, 3 Feb 2010 20:41:13 +0000]
Revert "Update to MIME-Base64 3.09"

I chatted with Gisle and he confirmed that this didn't fix a
release-blocking issue. Since we're frozen for 5.12, he agreed that we
should back it out.

This reverts commit 5e58db16ffcf34442d0ba4b645757884324e35c2.

14 years agoMerge branch 'blead' of ssh://perl5.git.perl.org/gitroot/perl into blead
Jesse Vincent [Wed, 3 Feb 2010 17:44:19 +0000]
Merge branch 'blead' of ssh://perl5.git.perl.org/gitroot/perl into blead

* 'blead' of ssh://perl5.git.perl.org/gitroot/perl:
  Tests for RT #72432

14 years agoAllow arbitrary whitespace between NAME and VERSION in "package NAME VERSION;" statements
Jesse Vincent [Wed, 3 Feb 2010 17:43:08 +0000]
Allow arbitrary whitespace between NAME and VERSION in "package NAME VERSION;" statements

Fixes [perl #72432]

14 years agoTests for RT #72432
Abigail [Wed, 3 Feb 2010 17:25:46 +0000]
Tests for RT #72432

14 years agoadd a watchdog for schwern's time overflow tests.
Jesse Vincent [Wed, 3 Feb 2010 15:23:58 +0000]
add a watchdog for schwern's time overflow tests.

14 years agoAdd sanity checks for far, far distant dates.
Michael G. Schwern [Sat, 30 Jan 2010 09:41:59 +0000]
Add sanity checks for far, far distant dates.

Can't go beyond y2**31 because of the 32 bit year

Adds a test that we can make it to at least +/-2**52.  This is the
missing test for 455f2c6c92e42c1a9e31268cbd491ba661f99882 which kicked this
whole thing off.

Signed-off-by: Michael G. Schwern <schwern@pobox.com>

14 years agoUpdate to MIME-Base64 3.09
Gisle Aas [Tue, 2 Feb 2010 22:22:15 +0000]
Update to MIME-Base64 3.09

Fixes issue where the Quoted-Printable encoder would sometimes output
lines that were 77 characters long.  The max line length should be 76.

14 years agoDiagnostic improvements to better understand why lib/Benchark.t test 13 has been...
Todd Rinaldo [Tue, 2 Feb 2010 17:27:37 +0000]
Diagnostic improvements to better understand why lib/Benchark.t test 13 has been failing smoke randomly.

Fix 1: Original code tests for less than but not =. I think that if these values are the same, the test should pass. I don't know the code well enough to be 100% sure. D
Fix 2: convert ok() to cmp_ok() for better diagnostic messages from smoke tests when they happen.
Fix 3: convert print to diag() so it will properly print through harness.

14 years ago[PATCH] Add gcc predefined macros to $Config{cppsymbols} on GNU/Hurd.
H.Merijn Brand [Mon, 1 Feb 2010 11:58:57 +0000]
[PATCH] Add gcc predefined macros to $Config{cppsymbols} on GNU/Hurd.

This is needed at least by h2ph, otherwise the generated .ph files
choke on missing __LONG_MAX__ and similar definitions.

Patch by Samuel Thibault <sthibault@debian.org>.

Signed-off-by: H.Merijn Brand <h.m.brand@xs4all.nl>

14 years agogmtime/localtime are busted around 2**48
Michael G Schwern [Sun, 31 Jan 2010 10:53:43 +0000]
gmtime/localtime are busted around 2**48

Michael G Schwern wrote:
> Anyhow, I'm willing to let that drop, but I want figure out A) why its
> failing so early because it might be masking a bug, this is not
> predicted behavior, and most importantly B) to get it to error, or at
> least warn.
>
> I know B is possible, I did it for Time::y2038.  I'll figure that part
> out and try to dig into A.  I would like to call B a blocker for 5.12.

Found the bug, patch attached.

v_tm_tday is an NV, so casting time to an integer will cause an overflow
2**31 days after 1970 or somewhere between 2**47 and 2**48.  Not sure why the
compiler didn't warn about that.

This makes it work until the year overflows.

As for putting in a warning, there's two options:
1) A few days before the 32 bit year overflows.
2) At some reasonable point before there's too much floating point inaccuracy.

The possibility of long doubles will make #2 difficult to nail down, but I'll
be happy with +/- 2**53.

--
24. Must not tell any officer that I am smarter than they are, especially
     if it's true.
     -- The 213 Things Skippy Is No Longer Allowed To Do In The U.S. Army
            http://skippyslist.com/list/

From 9e940e610ac5b2fbb09a554f505963094a4a0745 Mon Sep 17 00:00:00 2001
From: Michael G. Schwern <schwern@pobox.com>
Date: Sun, 31 Jan 2010 02:24:50 -0800
Subject: [PATCH] Type conversion bug in gmtime64 that was causing it to crap out around 2**48

Signed-off-by: H.Merijn Brand <h.m.brand@xs4all.nl>

14 years agoClarify warning that e.g. \q is reserved
Karl Williamson [Sat, 30 Jan 2010 19:06:05 +0000]
Clarify warning that e.g. \q is reserved

Currently the meaning of e.g. \q is simply 'q', with a warning
generated.  Elsewhere it is documented that this means that \q is
thus a reserved term, available for Perl to co-opt in some future
release.  But the warning doesn't say that.  Clarify it.

14 years agoFix POD typo
Frank Wiegand [Sat, 30 Jan 2010 08:54:32 +0000]
Fix POD typo

14 years agosuggestion to use GET URL
Ricardo Signes [Fri, 29 Jan 2010 00:19:13 +0000]
suggestion to use GET URL

14 years agoFix a broken pod link
Jesse Vincent [Thu, 28 Jan 2010 23:39:42 +0000]
Fix a broken pod link

14 years ago* Fill out the docs on the yada to show it as a satand in for statements, not expressions
brian d foy [Thu, 28 Jan 2010 20:22:43 +0000]
* Fill out the docs on the yada to show it as a satand in for statements, not expressions

14 years agoMerge branch 'blead' of ssh://perl5.git.perl.org/gitroot/perl into blead
Abigail [Thu, 28 Jan 2010 20:16:50 +0000]
Merge branch 'blead' of ssh://perl5.git.perl.org/gitroot/perl into blead

14 years agoAdd a few regression tests, making sure ... is parsed as a range operator
Abigail [Thu, 28 Jan 2010 20:16:33 +0000]
Add a few regression tests, making sure ... is parsed as a range operator

14 years agoClean up 5.12 delta pod concerning regexes and Unicode changes
Karl Williamson [Thu, 28 Jan 2010 16:20:54 +0000]
Clean up 5.12 delta pod concerning regexes and Unicode changes

14 years agoResort MANIFEST.
Nicholas Clark [Thu, 28 Jan 2010 11:37:28 +0000]
Resort MANIFEST.

14 years agoRemove duplicate entries from MANIFEST
Steve Hay [Wed, 27 Jan 2010 22:54:16 +0000]
Remove duplicate entries from MANIFEST

14 years agoAdd Porting/check-cpan-pollution to MANIFEST
Steve Hay [Wed, 27 Jan 2010 22:46:23 +0000]
Add Porting/check-cpan-pollution to MANIFEST

14 years ago* Note that the flip-flop operator maintains state across subroutine calls
brian d foy [Wed, 27 Jan 2010 03:54:32 +0000]
* Note that the flip-flop operator maintains state across subroutine calls

14 years agoCheck for Maintainers.pl updates
Steffen Mueller [Tue, 26 Jan 2010 16:31:52 +0000]
Check for Maintainers.pl updates

Make Porting/check-cpan-pollution test whether a given, presumably safe
commit to dual-lived modules also updated Porting/Maintainers.pl.

14 years agoPorting/ tool: Check for bad commits against cpan/
Steffen Mueller [Tue, 26 Jan 2010 16:23:38 +0000]
Porting/ tool: Check for bad commits against cpan/

Porting/check-cpan-pollution runs a series of tests to find potentially
unsafe commits that change dual-lived modules and prints a summary.

14 years agoA small typo in perlrun, spotted by Olivier Raginel.
Vincent Pit [Tue, 26 Jan 2010 09:47:54 +0000]
A small typo in perlrun, spotted by Olivier Raginel.

14 years agocomment grammar, indentation
Jim Cromie [Sat, 23 Jan 2010 16:35:37 +0000]
comment grammar, indentation

14 years agorevise package NAME VERSION entry in perl5120delta
David Golden [Mon, 25 Jan 2010 20:55:52 +0000]
revise package NAME VERSION entry in perl5120delta

14 years agocorrect a typo
Ricardo Signes [Mon, 25 Jan 2010 17:00:28 +0000]
correct a typo

11:58 <@Zefram> "deference" is a typo
11:58 <@Zefram> (in that context)
11:58 <@rjbs> ...and it should be what?
11:58 <@Zefram> should be "dereference"

14 years agoapply a second set of eyes to perl5120delta
Ricardo Signes [Mon, 25 Jan 2010 16:31:22 +0000]
apply a second set of eyes to perl5120delta

* fix some grammatical and spelling errors
* adjust some usage nits
* remove irrelevant data
* reorganize some lists that became interspersed

14 years agoMerge branch 'blead' of ssh://perl5.git.perl.org/gitroot/perl into blead
Jesse Vincent [Mon, 25 Jan 2010 12:54:34 +0000]
Merge branch 'blead' of ssh://perl5.git.perl.org/gitroot/perl into blead

* 'blead' of ssh://perl5.git.perl.org/gitroot/perl:
  Update CPANPLUS::Dist::Build to cpan version 0.46
  Change from a hard coded temporary file name in lib/AnyDBM_File.t.

14 years agoAdd perl5120delta to pod.lst and MANIFEST
Jesse Vincent [Mon, 25 Jan 2010 12:54:15 +0000]
Add perl5120delta to pod.lst and MANIFEST

14 years agoUpdate CPANPLUS::Dist::Build to cpan version 0.46
Chris Williams [Mon, 25 Jan 2010 10:51:13 +0000]
Update CPANPLUS::Dist::Build to cpan version 0.46

  Extract from Changes file
  =========================

  0.46 Sat Jan 23 07:52:19 GMT 2010
    - actually make the prereq resolving fallback to _build/ querying
      if the 'prereq_data' action fails. *facepalm*

Resolves an issue in prereq resolution where the failure of using
'Build prereq_data' would not fallback to querying _build/ data.

Highlighted by EekBoek distribution which bundled an older version of
Module::Build.

14 years agoChange from a hard coded temporary file name in lib/AnyDBM_File.t.
Nicholas Clark [Mon, 25 Jan 2010 10:27:38 +0000]
Change from a hard coded temporary file name in lib/AnyDBM_File.t.

Previously lib/AnyDBM_File.t and lib/DBM_Filter/t/*t shared the same hard coded
filename for their test database, causing a race condition during parallel
testing, and some spurious test failures.

14 years agoMinor typo fix
Jesse Vincent [Sun, 24 Jan 2010 19:52:32 +0000]
Minor typo fix

14 years agoFirst full pass at perl5120delta
Jesse Vincent [Sun, 24 Jan 2010 19:52:05 +0000]
First full pass at perl5120delta

14 years agoAdded a stub of the Perl 5.12.0 delta which will be a rollup of the various perl...
Jesse Vincent [Sun, 24 Jan 2010 18:23:06 +0000]
Added a stub of the Perl 5.12.0 delta which will be a rollup of the various perl 5.11.x deltas

14 years agoUpdate to podlators 2.3.0
Rafael Garcia-Suarez [Fri, 22 Jan 2010 16:19:17 +0000]
Update to podlators 2.3.0

14 years agoClarify in INSTALL what an extension name looks like:
David Mitchell [Fri, 22 Jan 2010 13:16:27 +0000]
Clarify in INSTALL what an extension name looks like:

is it IPC::SysV or IPC/SysV, or... ?

14 years agofix SEGV in /\N{...}/
David Mitchell [Fri, 22 Jan 2010 12:57:27 +0000]
fix SEGV in /\N{...}/

A simple program like the following could coredump:

    use charnames ':full';
    /\N{LATIN SMALL LETTER E}/;

The moral being, make sure sp is synced on return from call_sv()
*before* using the stack!

(Was a regression since 5.10)

14 years agoupdate Mark Jason Dominus's name in AUTHORS
Ricardo Signes [Thu, 21 Jan 2010 20:44:56 +0000]
update Mark Jason Dominus's name in AUTHORS

14 years agoinitialize undefined version objects with zero, not 'undef'
David Golden [Thu, 21 Jan 2010 17:18:11 +0000]
initialize undefined version objects with zero, not 'undef'

14 years agoAdd USE_PERL_ATOF to the list of -V's compile-time options.
Nicholas Clark [Thu, 21 Jan 2010 16:07:59 +0000]
Add USE_PERL_ATOF to the list of -V's compile-time options.

14 years agoRevert "remove defunct "Changes" file step from RMG"
Ricardo Signes [Thu, 21 Jan 2010 13:43:24 +0000]
Revert "remove defunct "Changes" file step from RMG"

This reverts commit 6642915e130e2410e5cc7d1e00a39aa7b8774d05.

14 years agoremove defunct "Changes" file step from RMG
Ricardo Signes [Thu, 21 Jan 2010 01:23:16 +0000]
remove defunct "Changes" file step from RMG

14 years agoRemove CPAN's Makefile.PL; let make_ext.pl write one for us.
Nicholas Clark [Wed, 20 Jan 2010 17:22:32 +0000]
Remove CPAN's Makefile.PL; let make_ext.pl write one for us.

Nothing in CPAN's Makefile.PL was pertinent to building as part of the core
distribution, but code within it caused `make` after `make clean` to fail.
Resolves RT #72218

14 years agoMake perl's magic variables in gv.c grep-able
Ævar Arnfjörð Bjarmason [Wed, 20 Jan 2010 03:48:15 +0000]
Make perl's magic variables in gv.c grep-able

It can be hard especially for those unfamiliar with the source to find
where variables like $` and $[ are set up, now they're just a ack -Q
away.

14 years agoUpdate META.yml for Module::CoreList
Rafael Garcia-Suarez [Wed, 20 Jan 2010 17:09:54 +0000]
Update META.yml for Module::CoreList

14 years agocreate perl5115delta and update other versions
Ricardo Signes [Wed, 20 Jan 2010 17:04:05 +0000]
create perl5115delta and update other versions

14 years agomake clean deletes runtests, so add a rule and dependency to regenerate it.
Nicholas Clark [Wed, 20 Jan 2010 16:43:26 +0000]
make clean deletes runtests, so add a rule and dependency to regenerate it.

14 years agoadd -u to core-cpan-diffs to filter modules
Ricardo Signes [Wed, 20 Jan 2010 14:52:31 +0000]
add -u to core-cpan-diffs to filter modules

14 years agoperlfunc tweak for package NAME VERSION
David Golden [Wed, 20 Jan 2010 12:13:11 +0000]
perlfunc tweak for package NAME VERSION

14 years agoremove remaining boilerplate from perl5114delta
Ricardo Signes [Wed, 20 Jan 2010 05:52:45 +0000]
remove remaining boilerplate from perl5114delta

14 years agoperldelta5114 updated prior to shipping
Ricardo Signes [Wed, 20 Jan 2010 05:42:49 +0000]
perldelta5114 updated prior to shipping

14 years agoadd 5.11.4 (and myself) to perlhist
Ricardo Signes [Wed, 20 Jan 2010 04:38:22 +0000]
add 5.11.4 (and myself) to perlhist

14 years agoregen_perly prepping for 5.11.4
Ricardo Signes [Wed, 20 Jan 2010 04:13:44 +0000]
regen_perly prepping for 5.11.4

14 years agoupdate Module::CoreList for pending 5.11.4
Ricardo Signes [Wed, 20 Jan 2010 03:19:20 +0000]
update Module::CoreList for pending 5.11.4

14 years agoUpdate Module::Build to 0.3603
David Golden [Wed, 20 Jan 2010 02:33:35 +0000]
Update Module::Build to 0.3603

0.3603 - Mon Jan 18 22:28:59 EST 2010

 Bug fixes:

 - Module::Build::Compat would croak on distibutions that set requires
   'perl' to a dotted decimal like '5.6.2'.  We now skip that key
   since it doesn't go into PREREQ_PM and we numify it properly for
   'use 5.006002' in the generated Makefile.PL (RT#53409)
   [David Golden, adapted from patch by G. Allen Morris III]

0.3602 - Mon Jan 18 22:09:54 EST 2010

 Bug fixes:

 - Fix failures in t/properties/needs_compiler.t when $ENV{CC} is set
   (RT#53296) [David Golden, adapted from patch by Jens Rehsack]

0.3601 - Mon Dec 21 14:39:33 EST 2009

 Bug fixes:

 - When the currently running Module::Build is not the same as the one
   that created the Build file, there is now a warning rather than a fatal
   error.  This helps installation of dependency chains where a dependency
   might configure_requires a new Module::Build after Build.PL was already
   run for an earlier distribution. [David Golden, on advice of Matt Trout]

 Other:

 - t/bundle_inc.t fails in odd ways.  This test of an experimental feature
   should not prevent users from installing Module::Build, so this test
   now skips unless $ENV{MB_TEST_EXPERIMENTAL} is true

14 years agoreminder: update Module::CoreList when bumping version number
David Golden [Wed, 20 Jan 2010 01:19:24 +0000]
reminder: update Module::CoreList when bumping version number

14 years agobump versions for core libs changed since 5.11.3
Ricardo Signes [Wed, 20 Jan 2010 01:16:55 +0000]
bump versions for core libs changed since 5.11.3

14 years agoregen warnings.pl for new version number
Ricardo Signes [Tue, 19 Jan 2010 23:53:21 +0000]
regen warnings.pl for new version number

14 years agobump version for 5.11.4
Ricardo Signes [Tue, 19 Jan 2010 23:43:57 +0000]
bump version for 5.11.4

14 years agofix qr// and get-magic problems
Father Chrysostomos [Tue, 19 Jan 2010 20:35:04 +0000]
fix qr// and get-magic problems

[N.B. I converted package name separators from q{'} to q{::} in
the test files as suggested by demerphq. -- dagolden]

14 years agoFix for #71506: work around possible gcc bug
Tony Cook [Tue, 19 Jan 2010 14:07:21 +0000]
Fix for #71506: work around possible gcc bug

When memcpy() is used on a long double pointer with gcc 4.4 in some
cases it seems to treat it as a long double assignment, copying only
the first 12 bytes.  Use unions so the types we're copying into the
pack output or from the unpack input are unsigned char[], to avoid the
apparent bug.

14 years agonote strict/lax version requirements in documentation
David Golden [Tue, 19 Jan 2010 15:14:32 +0000]
note strict/lax version requirements in documentation

14 years agoadd version number format note to perldelta
David Golden [Tue, 19 Jan 2010 15:03:06 +0000]
add version number format note to perldelta

14 years agoFix a NULL pointer dereference when looking for a DESTROY method
Niko Tyni [Tue, 19 Jan 2010 10:03:08 +0000]
Fix a NULL pointer dereference when looking for a DESTROY method

The empty DESTROY method optimization introduced by commit
fbb3ee5af3d would crash the interpreter if a DESTROY method
was declared but not actually defined.

This is seen in the real world with AutoLoader / AutoSplit,
where the crash defeats autoloading a DESTROY method.

14 years agoFix for #71254: SEGV in Data::Dumper
Father Chrysostomos [Mon, 18 Jan 2010 20:57:01 +0000]
Fix for #71254: SEGV in Data::Dumper

This was caused by change 27323/f7877b281b4, which changes the way
globs are stored in SVs. This patch teaches Perl_magic_setmglob (which
resets the match position after an assignment) about globs. What was
happening was that the globness was turned off (with the type still as
PVGV), which essentially turned the variable into a strange empty
string. Data::Dumper, seeing a PVGV, assumes the string form is at
least 1 char (which should always be the case), and ends up reading
past the end of the string if it is blank.

14 years agoFixed an editing error in perldoc about commit messages.
Jesse Vincent [Mon, 18 Jan 2010 19:24:36 +0000]
Fixed an editing error in perldoc about commit messages.

(Thanks to Aristotle Pagaltzis)

14 years agoMinor updates to the git tutorial and our commit message policies.
Jesse Vincent [Mon, 18 Jan 2010 19:00:26 +0000]
Minor updates to the git tutorial and our commit message policies.

14 years ago[perl #71676] GDBM_File support on Win32
kmx [Mon, 18 Jan 2010 17:39:22 +0000]
[perl #71676] GDBM_File support on Win32

Date: Thu, 07 Jan 2010 00:36:25 -0800
Message-ID: <rt-3.6.HEAD-1505-1262853384-1207.71676-15-0@perl.org>

Tweaked as suggested in:

From: "H.Merijn Brand" <h.m.brand@xs4all.nl>
Date: Fri, 8 Jan 2010 13:35:17 +0100
Message-ID: <20100108133517.1f2916d3@pc09.procura.nl>

14 years agoPerl blead on AIX - fail: t/porting/test_bootstrap
Rainer Tammer [Mon, 18 Jan 2010 10:43:07 +0000]
Perl blead on AIX - fail: t/porting/test_bootstrap

Hello,

Nicholas Clark wrote:
> On Mon, Jan 18, 2010 at 11:33:07AM +0100, Rainer Tammer wrote:
>
>> Hello,
>> the 32 bit blead build looks quite good on AIX:
>>
>> A threaded, shared 32 bit build only produced one error:
>>
> Thanks for continuing to look out for AIX.
>
>
No problem ...
>> Is this a know error ??
>>
> Yes, but I believe that I committed a change to fix it about 30 minutes ago.
>
>
OK,
I try this fix.

I have also a fix for the GDBM problem.
Now GDBM is only disabled if the "problem" headers are installed,
e.g. if the gdbm rpm < 1.8.3-5 is installed.

I have attached the patch for this.

> Nicholas Clark
>
>
>
Bye
  Rainer

From 4755db7f86c38d3c2b69c668c5747368e98c3eae Mon Sep 17 00:00:00 2001
From: Rainer Tammer <tammer@tammer.net>
Date: Mon, 18 Jan 2010 11:40:27 +0100
Subject: [PATCH] Enabld GDBM on AIX if possible

Signed-off-by: H.Merijn Brand <h.m.brand@xs4all.nl>

14 years agoMove packagev.t from t/comp to t/op, as it fouls the test bootstrapping policy.
Nicholas Clark [Mon, 18 Jan 2010 10:01:32 +0000]
Move packagev.t from t/comp to t/op, as it fouls the test bootstrapping policy.

The package VERSION syntax isn't strictly an op, but it needs to use several
features that aren't yet tested at the time that tests in t/comp are run.

14 years agoRevert "[perl #62646] Maximum string length with substr"
Rafael Garcia-Suarez [Mon, 18 Jan 2010 06:29:50 +0000]
Revert "[perl #62646] Maximum string length with substr"

This reverts commit b6d1426f94a845fb8fece8b6ad0b7d9f35f2d62e.

Conflicts:

pp.c

14 years agoupdate AUTHORS for new commits, spelling
Ricardo Signes [Sun, 17 Jan 2010 20:19:08 +0000]
update AUTHORS for new commits, spelling

14 years agoAdd t/comp/packagev.t to MANIFEST
Vincent Pit [Sun, 17 Jan 2010 18:48:39 +0000]
Add t/comp/packagev.t to MANIFEST

14 years agoUpgrade to CPAN.pm 1.94_54
Rafael Garcia-Suarez [Sun, 17 Jan 2010 15:11:42 +0000]
Upgrade to CPAN.pm 1.94_54

14 years agoClean up mktables intro comments; remove trailing white space
Karl Williamson [Sun, 27 Dec 2009 17:48:56 +0000]
Clean up mktables intro comments; remove trailing white space

14 years agoCorrect VMS-specific handling of $! in Perl_magic_get.
Craig A. Berry [Sat, 16 Jan 2010 22:31:01 +0000]
Correct VMS-specific handling of $! in Perl_magic_get.

0097b436152452e403cc71b4f1a1cfd30ec0ba1a had introduced a test failure
on VMS because it only set the POK flag in the non-VMS code path,
which was an easy mistake to make because accumulated patches had
over time made the #ifdef jungle way more tangled than it needed to
be.  There is really only one line that needs to be VMS-specific.

14 years agoA fig leaf for calling sv_pos_u2b with IV* where it expects I32*.
Craig A. Berry [Sat, 16 Jan 2010 22:15:47 +0000]
A fig leaf for calling sv_pos_u2b with IV* where it expects I32*.

Following v5.11.3-103-gb6d1426, any compiler paying attention
whines about the pointer mismatch (which on VMS breaks the
build).

Zefram reports a further patch in progress:
  Message-ID: <20100116022631.GA10264@fysh.org>

but this gets the warning out of the way so other work can proceed.

14 years agoParse 'use NAME VERSION' with C locale
David Golden [Sat, 16 Jan 2010 23:03:52 +0000]
Parse 'use NAME VERSION' with C locale

14 years agowrap a link in L<> to avoid bad autolinking
Ricardo Signes [Sat, 16 Jan 2010 19:58:36 +0000]
wrap a link in L<> to avoid bad autolinking

14 years ago[perl #72098] File::Copy stripping 06000 perms on cp for root
Todd Rinaldo [Sat, 16 Jan 2010 18:19:23 +0000]
[perl #72098] File::Copy stripping 06000 perms on cp for root

The problem is with the use of cp when perms & 06000.

There is logic that checks to see if the target file is owned by the
user ($>) running the copy and/or if $> is a member of the group that
owns the target file. If this is not the case, then the 06000 bits are
masked out before the chmod is called after the copy.

This is mostly good logic except when root is executing this, in which
case root should get to do whatever it wants to do.

Looking closer at the code to test for group membership, I think it can
be more easily and more cheaply be written using $). I've added this
change to the patch.

This will also fix the problem where someone has a group membership
based on /etc/passwd and is not mentioned in /etc/group.