This is a live mirror of the Perl 5 development currently hosted at
9 years agoReorder the members of struct re_save_state to reduce its size on LP64.
Nicholas Clark [Mon, 25 Feb 2013 16:50:35 +0000 (17:50 +0100)]
Reorder the members of struct re_save_state to reduce its size on LP64.

This drops its size by 8 bytes on LP64 platforms. This also makes the
interpreter struct 8 bytes smaller, as it embeds a re_save_state struct.

9 years agoMerge the refactored S_hsplit() and Perl_hv_ksplit() and related cleanups.
Nicholas Clark [Tue, 26 Feb 2013 14:38:27 +0000 (15:38 +0100)]
Merge the refactored S_hsplit() and Perl_hv_ksplit() and related cleanups.

Perl_hv_ksplit() now calls S_hsplit(), avoiding duplicating the code to
split hashes when they grow.

9 years agoIn Perl_hv_common(), call S_clear_placeholders() directly.
Nicholas Clark [Fri, 22 Feb 2013 10:19:17 +0000 (11:19 +0100)]
In Perl_hv_common(), call S_clear_placeholders() directly.

This saves one call to HvPLACEHOLDERS_get().

9 years agoClarify why hv_common() tries to clear placeholders before calling hsplit().
Nicholas Clark [Fri, 22 Feb 2013 09:52:55 +0000 (10:52 +0100)]
Clarify why hv_common() tries to clear placeholders before calling hsplit().

Which makes me realise that if we clear placeholders, we may be able to
avoid the need to split at all.

(In fact, as hv_common() only adds one key to the hash, under the current
definition of DO_HSPLIT() which only considers total number of keys,
clearing any placeholder is going to be enough to drop the total number of
keys, and so no longer trigger the split. But we'll leave the code making a
second check, to avoid a tight coupling with the internals of DO_HSPLIT().)

9 years agoIn S_hsplit(), replace a for with a do/while, as the loop runs at least once.
Nicholas Clark [Thu, 21 Feb 2013 19:54:48 +0000 (20:54 +0100)]
In S_hsplit(), replace a for with a do/while, as the loop runs at least once.

Seems pointless to check the exit condition before any iterations, when we
know that it will always be false the first time.

9 years agoReplace the bulk of Perl_hv_ksplit() with a call to S_hsplit().
Nicholas Clark [Thu, 21 Feb 2013 19:22:32 +0000 (20:22 +0100)]
Replace the bulk of Perl_hv_ksplit() with a call to S_hsplit().

The code duplication that introduced hv_ksplit() as a fork of hsplit() back
with commit 72940dca186befa0 in Sept 1996 is finally healed.

9 years agoTweak S_hsplit() to return early if there are no keys to process.
Nicholas Clark [Thu, 21 Feb 2013 18:58:04 +0000 (19:58 +0100)]
Tweak S_hsplit() to return early if there are no keys to process.

This mimics the behaviour in Perl_hv_ksplit().

Also remove a vestigial comment. The code it relates to was removed in
commit 7dc8663964c66a69 in Nov 2012.

9 years agoPass the current and desired hash sizes to S_hsplit().
Nicholas Clark [Thu, 21 Feb 2013 18:45:38 +0000 (19:45 +0100)]
Pass the current and desired hash sizes to S_hsplit().

Whilst this is slightly more work for its existing two callers, it will
permit Perl_hv_ksplit() to also call it.

Use STRLEN for the parameters, and change a local variable from I32 to
STRLEN to match.

9 years agoMove the code handling allocating a new buffer earlier in Perl_hv_ksplit().
Nicholas Clark [Thu, 21 Feb 2013 18:16:44 +0000 (19:16 +0100)]
Move the code handling allocating a new buffer earlier in Perl_hv_ksplit().

This makes the rest of the code of Perl_hv_ksplit() closer to that of

9 years agoRefactor the loop logic in S_hsplit() and Perl_hv_ksplit() to converge.
Nicholas Clark [Thu, 21 Feb 2013 16:22:23 +0000 (17:22 +0100)]
Refactor the loop logic in S_hsplit() and Perl_hv_ksplit() to converge.

Making the code as similar as possible will make it simpler to merge the two.

9 years agoMove the call to hv_clear_placeholders() from hsplit() to hv_common().
Nicholas Clark [Thu, 21 Feb 2013 16:09:29 +0000 (17:09 +0100)]
Move the call to hv_clear_placeholders() from hsplit() to hv_common().

The relevant code calls Perl_hv_clear_placeholders() at split time, if there
are still placeholders left over from a (previously) restricted hash.
There are two callers to S_hsplit(), one from the regular HV code, and one
from the shared string table code. As the shared string table can't contain
placeholders, only the other call site could trigger this condition, so move
the code there. This simplifies S_hsplit(), and will make splitting the
shared string table marginally faster.

9 years agoignore some files from Module-Pluggable
Yves Orton [Tue, 26 Feb 2013 14:52:42 +0000 (15:52 +0100)]
ignore some files from Module-Pluggable

9 years agoUpgrade Module-Pluggable to v4.7
Yves Orton [Tue, 26 Feb 2013 07:27:45 +0000 (08:27 +0100)]
Upgrade Module-Pluggable to v4.7

Preserved old makefile as some of its logic is in Build.PL now which we dont use.

9 years agoNote CPAN release for Data::Dumper
Steffen Mueller [Tue, 26 Feb 2013 06:57:14 +0000 (07:57 +0100)]
Note CPAN release for Data::Dumper

9 years agoUpdate Data::Dumper version and release date
Steffen Mueller [Tue, 26 Feb 2013 06:53:59 +0000 (07:53 +0100)]
Update Data::Dumper version and release date

9 years agoUpdate Data::Dumper changelog
Steffen Mueller [Tue, 26 Feb 2013 06:51:22 +0000 (07:51 +0100)]
Update Data::Dumper changelog

9 years agoFix redeclaration compiler errors on EBCDIC
Nicholas Clark [Tue, 26 Feb 2013 03:53:52 +0000 (20:53 -0700)]
Fix redeclaration compiler errors on EBCDIC

This patch was posted in

9 years agoGet regcomp.c to compile under EBCDIC
John Goodyear [Mon, 25 Feb 2013 21:44:44 +0000 (14:44 -0700)]
Get regcomp.c to compile under EBCDIC

There were a couple of syntax errors here.

The committer modified the patch slightly based on a p5p exchange.

9 years agoAdd, fix comments
Karl Williamson [Thu, 14 Feb 2013 17:54:32 +0000 (10:54 -0700)]
Add, fix comments

9 years agoUnicode/ Fix undef bug
Karl Williamson [Mon, 25 Feb 2013 20:35:34 +0000 (13:35 -0700)]
Unicode/ Fix undef bug

This only happens should Perl be compiled on the very first Unicode
release, which is extremely unlikely, but fix it anyway.

9 years agounwrap #define argument list for REXEC_TRIE_READ_CHAR
Yves Orton [Mon, 25 Feb 2013 12:34:52 +0000 (13:34 +0100)]
unwrap #define argument list for REXEC_TRIE_READ_CHAR

Apparently this breaks the z/OS 1.13 c preprocessor. (Yay, now I can say I have written code for z/OS!)

9 years agomakedepend.SH: Deal with continuation lines
Karl Williamson [Sat, 23 Feb 2013 22:41:28 +0000 (15:41 -0700)]
makedepend.SH: Deal with continuation lines

makedepend was failing to take into account the fact that preprocessor
directives, like any line, may be continued onto succeeding lines by a
trailing backslash.  This led to errors in the C preprocessing, which were
suppressed on all but os390 systems.o

This commit alleviates that, but it is not a complete solution.  It
deletes C comments, but this being sed, it doesn't catch all nuances of
C comments.  We only care, however, about comments on preprocessor
directive lines, and this is good enough to work on the current Perl
without generating any errors, on my system.

Spotted by John Goodyear.

9 years agomakedepend.SH: Preprocessor directives don't have to start at column 1
Karl Williamson [Sat, 23 Feb 2013 22:38:55 +0000 (15:38 -0700)]
makedepend.SH: Preprocessor directives don't have to start at column 1

Like, apparently, the person(s) who originally wrote this script, I
thought all preprocessor directives had to start in column 1.

It turns out that restriction was relaxed in C89; so that they may be

Spotted by John Goodyear.

9 years agomakedepend.SH: Make compiler errors show correct line number
Karl Williamson [Sat, 23 Feb 2013 22:34:05 +0000 (15:34 -0700)]
makedepend.SH: Make compiler errors show correct line number

Instead of inserting "#line 1", insert "#line 2" to account for the
inserted line.

Spotted by John Goodyear.

9 years agoMove (?[]) doc to perlrecharclass
Karl Williamson [Sun, 24 Feb 2013 18:42:20 +0000 (11:42 -0700)]
Move (?[]) doc to perlrecharclass

This is a form of bracketed character class, and so should be documented
in the same pod.

9 years agoDocument \s change for VT, commit 075b9d7d9a6d4473b240a047655e507c8baa6db3
Karl Williamson [Sun, 24 Feb 2013 18:17:19 +0000 (11:17 -0700)]
Document \s change for VT, commit 075b9d7d9a6d4473b240a047655e507c8baa6db3

9 years agoperlintro: Typo and stop verbatim wrapping
Karl Williamson [Sun, 24 Feb 2013 03:35:04 +0000 (20:35 -0700)]
perlintro: Typo and stop verbatim wrapping

9 years agoperlhacktips: Note existence of regen/
Karl Williamson [Sat, 23 Feb 2013 02:04:25 +0000 (19:04 -0700)]
perlhacktips: Note existence of regen/

9 years agoHandle undef values in magic.t on VMS.
Craig A. Berry [Sat, 23 Feb 2013 18:45:11 +0000 (12:45 -0600)]
Handle undef values in magic.t on VMS.

888a67f6b95603701 added a new test that failed on VMS because it
expected the value of an undefined element in %ENV to be an empty
string, whereas on VMS it's actually a one-byte string containing
a null.  This is made possible by the fact that the logical name
API uses strings of known lengths, and made necessary by the fact
that a zero-length value (equivalence name in VMS parlance) is
not allowed.

So patch up the test implementation to handle this special case.

9 years agoadd array DELETE and EXISTS to perlfunc
David Mitchell [Wed, 30 Jan 2013 17:22:26 +0000 (17:22 +0000)]
add array DELETE and EXISTS to perlfunc

[perl #116579]
These methods weren't listed in the list of methods for tied arrays

9 years agoMark v5.17.9 as done in the release schedule
Chris 'BinGOs' Williams [Sat, 23 Feb 2013 11:36:51 +0000 (11:36 +0000)]
Mark v5.17.9 as done in the release schedule

9 years agoRemove trailing white-space from RMG document
Chris 'BinGOs' Williams [Sat, 23 Feb 2013 11:33:37 +0000 (11:33 +0000)]
Remove trailing white-space from RMG document

9 years agoPATCH: [perl #116899]: jump on uninitialised value
Karl Williamson [Fri, 22 Feb 2013 20:04:05 +0000 (13:04 -0700)]
PATCH: [perl #116899]: jump on uninitialised value

The culprit had nothing to do really with the accuesed commit.
The function S_cl_or() tries to take the union of the code points
matched by its two inputs.  Both of those nodes must be ANYOF-like
(for bracketed character classes and synthetic start classes).  These
come in two flavors, one having extra fields in the struct after the
other one's.  That is used for locale ANYOFs and the synthetic start
class.  The first paramter to cl_or() is always one of these extended
ANYOFS, but the second parameter may be the shorter form

The function was failing to check if the second one was the longer form
before reading data from beyond the short-form's struct.  This could
cause a segfault, but that wasn't the symptom here.  Instead it copied
that data to the other paramter's struct.  valgrind had set that data to
indicate it was uninitialized, so when later it was accessed, we got
this error.

During much of the 5.17 series until the failing commit, more ANYOF
nodes were the larger size.  I presume that is why this commit showed up
the problem.

9 years agoregcomp.c: Rmv comment ref to obsolete regnode
Karl Williamson [Fri, 22 Feb 2013 17:43:11 +0000 (10:43 -0700)]
regcomp.c: Rmv comment ref to obsolete regnode

9 years agoAbolish STRANGE_MALLOC. Now all malloc()s are considered strange :-)
Nicholas Clark [Wed, 20 Feb 2013 20:38:33 +0000 (21:38 +0100)]
Abolish STRANGE_MALLOC. Now all malloc()s are considered strange :-)

STRANGE_MALLOC was added in 5.002 beta 1 (4633a7c4bad06b47) as part of an
work around for typical mallocs, which had a bad interaction with perl's
allocation needs. Specifically, repeatedly extending an array and then
creating SV heads (such as when reading lines of a file into an array)
could end up with each reallocation for the array being unable to extend in
place, needing a fresh chunk of memory, and the released memory not being
suitable for use as more SV heads, so sitting unused. The solution was for
perl to recycle the old array body as SV heads, instead of returning it to
the system, passing the memory from the the AV code to the SV code using
offer_nice_chunk(), PL_nice_chunk and PL_nice_chunk_size.

STRANGE_MALLOC was actually a signal that the malloc() didn't need
protecting from itself, and to disable the work around.

offer_nice_chunk(), PL_nice_chunk and PL_nice_chunk_size were removed by
commit 9a87bd09eea1d037 in Nov 2010, without any ill effects, hence the
code used when STRANGE_MALLOC was *not* defined is essentially doing extra
work for no benefits.

From the lack of problems reported, one can assume that in the intervening
15 years malloc technology has got significantly improved, and it is probably
better to be honest with it, rather than trying to second guess it.

Hence remove all the non-STRANGE_MALLOC code, and leave everyone using the
much simpler code. See also

9 years agoperlfunc: clarify that "do $file" isn't really like "eval `cat $file`"
Aaron Crane [Tue, 19 Feb 2013 14:17:59 +0000 (14:17 +0000)]
perlfunc: clarify that "do $file" isn't really like "eval `cat $file`"

Closes RT#116713.

9 years agoUpdate for Safe 2.35
Rafael Garcia-Suarez [Thu, 21 Feb 2013 13:08:18 +0000 (14:08 +0100)]
Update for Safe 2.35

9 years agoUpgrade to Safe 2.35 from CPAN
Rafael Garcia-Suarez [Thu, 21 Feb 2013 07:35:38 +0000 (08:35 +0100)]
Upgrade to Safe 2.35 from CPAN

9 years agoSimplify the defintiion of optimization flags in Windows makefiles
Steve Hay [Thu, 21 Feb 2013 08:56:10 +0000 (08:56 +0000)]
Simplify the defintiion of optimization flags in Windows makefiles

The setting of -O1 for release mode builds can be factored out of the
x86 and x64 cases following commit 688ea162f4.

9 years agoPATCH: [perl #116787] -Ox on Win64 Visual C 64 bit is flawed, switch to -O1
Daniel Dragan [Thu, 21 Feb 2013 08:42:13 +0000 (08:42 +0000)]
PATCH: [perl #116787] -Ox on Win64 Visual C 64 bit is flawed, switch to -O1

-Ox is missing -GF and -Gy and -Gs

So the remaining question is, use -O1 or use -O2 on VC 64? VC 32 bit uses
-O1, and that is also slightly faster on x64 too.

9 years agoadd a missing 5 to the release announcement template
Ricardo Signes [Thu, 21 Feb 2013 00:17:06 +0000 (19:17 -0500)]
add a missing 5 to the release announcement template

9 years agoAmend 'new perldelta' advice for Win32
Chris 'BinGOs' Williams [Wed, 20 Feb 2013 23:27:30 +0000 (23:27 +0000)]
Amend 'new perldelta' advice for Win32

test_porting is supported as target for both gcc and VC++ builds

win32/Makefile since 247c55c00b852ec1da004aff2c0e2d4b07af55ab

win32/ since 106cf76e392294bc0fe10a3efa306d011f3b2e3c

9 years agoAlso skip io/eintr.t on MidnightBSD
Chris 'BinGOs' Williams [Wed, 20 Feb 2013 21:40:15 +0000 (21:40 +0000)]
Also skip io/eintr.t on MidnightBSD

9 years agoBump to v5.17.10
Chris 'BinGOs' Williams [Wed, 20 Feb 2013 23:00:21 +0000 (23:00 +0000)]
Bump to v5.17.10

9 years agoNew perldelta template for v5.17.10
Chris 'BinGOs' Williams [Wed, 20 Feb 2013 22:54:02 +0000 (22:54 +0000)]
New perldelta template for v5.17.10

9 years agoUpdate epigraphs with the one for v5.17.9
Chris 'BinGOs' Williams [Wed, 20 Feb 2013 22:42:24 +0000 (22:42 +0000)]
Update epigraphs with the one for v5.17.9

9 years agoFinalise perldelta with updated module information v5.17.9
Chris 'BinGOs' Williams [Wed, 20 Feb 2013 19:15:46 +0000 (19:15 +0000)]
Finalise perldelta with updated module information

9 years agoAdd new release to perlhist
Chris 'BinGOs' Williams [Wed, 20 Feb 2013 18:44:20 +0000 (18:44 +0000)]
Add new release to perlhist

9 years agoUpdated Module::CoreList for v5.17.9
Chris 'BinGOs' Williams [Wed, 20 Feb 2013 18:42:29 +0000 (18:42 +0000)]
Updated Module::CoreList for v5.17.9

9 years agoUpdated ACKNOWLEDGEMENTS in perldelta
Chris 'BinGOs' Williams [Wed, 20 Feb 2013 18:26:11 +0000 (18:26 +0000)]
Updated ACKNOWLEDGEMENTS in perldelta

9 years agosuppress "experimental" warnings for my $_
Ricardo Signes [Wed, 20 Feb 2013 17:00:26 +0000 (12:00 -0500)]
suppress "experimental" warnings for my $_

9 years agoput an experimental warning on lexical topic
Ricardo Signes [Wed, 20 Feb 2013 16:40:45 +0000 (11:40 -0500)]
put an experimental warning on lexical topic

9 years agoupdate release schedule for a volunteer
Ricardo Signes [Wed, 20 Feb 2013 16:34:06 +0000 (11:34 -0500)]
update release schedule for a volunteer

9 years agoRemove more perldelta boilerplate
Chris 'BinGOs' Williams [Wed, 20 Feb 2013 16:46:34 +0000 (16:46 +0000)]
Remove more perldelta boilerplate

9 years agoPerldelta note for undeprecating my $_
Rafael Garcia-Suarez [Wed, 20 Feb 2013 15:58:55 +0000 (16:58 +0100)]
Perldelta note for undeprecating my $_

9 years agoUn-deprecate the support for lexical $_
Rafael Garcia-Suarez [Tue, 19 Feb 2013 14:53:10 +0000 (15:53 +0100)]
Un-deprecate the support for lexical $_

9 years agoRemove boilerplate from perldelta
Chris 'BinGOs' Williams [Wed, 20 Feb 2013 15:32:21 +0000 (15:32 +0000)]
Remove boilerplate from perldelta

9 years ago[perl #72892] Perl should support PERLmmnnLIB environment variable
Andy Dougherty [Wed, 20 Feb 2013 13:58:49 +0000 (08:58 -0500)]
[perl #72892] Perl should support PERLmmnnLIB environment variable

This patch to perlrun.pod expands the documentation of the directory
hierarchy searched under the locations specified by PERL5LIB.  It now
explicitly mentions the version- and architecture-specific directories
added, and also notes that $Config{inc_version_list} directories are
added.  These additions were suggested by Damon Atkins.

9 years agoSilence irrelevant stderr output from op/magic.t
Steve Hay [Wed, 20 Feb 2013 13:28:46 +0000 (13:28 +0000)]
Silence irrelevant stderr output from op/magic.t

Some 'set' commands in this test are intended to verify that an environment
variable is not set, which causes the command to send a warning message to

9 years agoFix failing op/magic.t test 157
Steve Hay [Wed, 20 Feb 2013 13:25:41 +0000 (13:25 +0000)]
Fix failing op/magic.t test 157

Change 888a67f6b9 added manipulation of __NoNeSuCh2, but it gets
accidentally picked up by the 'set' command looking for __NoNeSuCh in test
157 because 'set' will display the value of *all* variables whose prefix
matches the name given to the command.

9 years agoEliminate 'swap' from struct regexp_internal.
Nicholas Clark [Tue, 19 Feb 2013 13:35:49 +0000 (14:35 +0100)]
Eliminate 'swap' from struct regexp_internal.

It's been unused since commit e9105d30edfbaa7f in July 2009.

9 years agoRemove the check for SVt_BIND from SvOK().
Nicholas Clark [Tue, 19 Feb 2013 10:43:08 +0000 (11:43 +0100)]
Remove the check for SVt_BIND from SvOK().

The original plan to use SVt_BIND to implement read-only aliases to
read-write values is unlikely to happen. More importantly, it's not going
to happen within a maint branch, so there's no reason to have the code
"just in case" it does. The code can easily be re-instated in blead if it
is needed in future. Nothing on CPAN is relying on this code. (Almost no
code on CPAN even references SVt_BIND.)

This effectively reverts part of commit 1cb9cd5016282146 from Dec 2006.

9 years ago[perl #116865] Upgrade to Thread::Queue 3.02
Jerry D. Hedden [Tue, 19 Feb 2013 20:23:46 +0000 (12:23 -0800)]
[perl #116865] Upgrade to Thread::Queue 3.02

Attached patch updates Thread::Queue to v3.0.2 in preparation for a
CPAN release.

Adds a new method to Thread::Queue to dequeue items with a timeout
feature.  This addition was suggested by Andreas Huber.

Signed-off-by: Chris 'BinGOs' Williams <>
9 years agore/charset.t, re/fold_grind.t: Note C locale can match 128-255
Karl Williamson [Tue, 19 Feb 2013 20:56:20 +0000 (13:56 -0700)]
re/charset.t, re/fold_grind.t: Note C locale can match 128-255

Craig Berry pointed out that an implementation can have the C locale
give meanings to the above-ASCII code points.  (I had thought this was
an error.)  Change label name and comments in these .t files

9 years agore/charset.t: Make sure locale is loaded at compile time
Karl Williamson [Sun, 17 Feb 2013 19:27:03 +0000 (12:27 -0700)]
re/charset.t: Make sure locale is loaded at compile time

Commit 569f7fc5d4ec06501b46a72075ff434fe1bf4332 caused this test file
to start failing on VMS.  The reason is that the test for locale became
run-time.  This patch restores the test to compile time, by using the
'if' module.

The reason 'if' wasn't used in the original patch was to reduce reliance
on possibly untested modules, but it makes things easier.  We hope to
get a resolution on p5p on which test groups in the suite can use things
like 'if', but since 5.17.9 is due out momentarily, I'm using 'if' for
now so that VMS will pass its tests again.

9 years agoUpdate Win32 to CPAN version 0.46
Jan Dubois [Tue, 19 Feb 2013 21:06:24 +0000 (13:06 -0800)]
Update Win32 to CPAN version 0.46


    0.46    [2013-02-19]
            - add Win2012/Win8 detection (thanks to Michiel Beijen) [rt#82572]

9 years agoAdd some text on the interoperability of Moose & Moo to perlootut
Dave Rolsky [Tue, 19 Feb 2013 20:56:35 +0000 (14:56 -0600)]
Add some text on the interoperability of Moose & Moo to perlootut

9 years agoAdd a note to the future that they should read the latest perlootut
Dave Rolsky [Tue, 19 Feb 2013 20:51:08 +0000 (14:51 -0600)]
Add a note to the future that they should read the latest perlootut

9 years agoRun podtidy on perlootut
Dave Rolsky [Tue, 19 Feb 2013 20:50:30 +0000 (14:50 -0600)]
Run podtidy on perlootut

9 years agoRename PL_interp_size_5_16_0 to PL_interp_size_5_18_0.
Nicholas Clark [Tue, 19 Feb 2013 18:47:14 +0000 (19:47 +0100)]
Rename PL_interp_size_5_16_0 to PL_interp_size_5_18_0.

9 years agoRe-order intrpvar.h to minimise holes in the interpreter struct.
Nicholas Clark [Tue, 19 Feb 2013 15:49:34 +0000 (16:49 +0100)]
Re-order intrpvar.h to minimise holes in the interpreter struct.

Holes were created by commit f59909ab8dad6ceb (April 2012) which removed
PL_reginterp_cnt, commit 7dc8663964c66a69 (Nov 2012) which removed
PL_rehash_seed_set, and commit 8936b48a49448f4e (Dec 2012) which removed

There is still an unavoidable U16 sized hole on the default threaded
configuration on x86_64. (U8 if PERL_SAWAMPERSAND is defined).

9 years agoFix [perl #116523] Configure is confused by '=' in output from uname -a
Andy Dougherty [Mon, 18 Feb 2013 14:33:09 +0000 (09:33 -0500)]
Fix [perl #116523] Configure is confused by '=' in output from uname -a

Previously, Configure used to try to post-process the output of uname -a.
This was to support Xenix, which gave output like this:
    release=2.3.2 .. etc...

Configure attempted to parse that output and build up a $myuname variable
that resembled those generated on traditional Unix systems.
More recently, a user reported uname -a output on a CVS checkout
of NetBSD that also included information such as

       root=username@hostname:/dirname module=src branch=HEAD

The Configure code misparsed that and ended up throwing most of it away.
This patch takes the simplest approach and just skips the attempt to
process the Xenix output.

This is based on commit 05a827803ab3908a795df94f6ed3890e24a6df2 in the
metaconfig repository.

9 years agoSilence deprecation warnings in the core test suite
Rafael Garcia-Suarez [Tue, 19 Feb 2013 08:59:09 +0000 (09:59 +0100)]
Silence deprecation warnings in the core test suite

9 years agoUpdated Changes for Module-CoreList
Chris 'BinGOs' Williams [Tue, 19 Feb 2013 08:50:09 +0000 (08:50 +0000)]
Updated Changes for Module-CoreList

9 years agoperldelta for #116322 and #112244
Karl Williamson [Tue, 19 Feb 2013 05:02:33 +0000 (22:02 -0700)]
perldelta for #116322 and #112244

9 years agodist/IO/IO.xs: Silence compiler warning
Karl Williamson [Sat, 2 Feb 2013 15:41:39 +0000 (08:41 -0700)]
dist/IO/IO.xs: Silence compiler warning

This variable is unused, doesn't need to be declared.

9 years agoTests for [perl #116322]
Karl Williamson [Sun, 17 Feb 2013 21:59:39 +0000 (14:59 -0700)]
Tests for [perl #116322]

9 years agoPATCH: [perl #116322]: getc() and ungetc() with unicode failure
Christian Hansen [Sun, 17 Feb 2013 21:50:30 +0000 (14:50 -0700)]
PATCH: [perl #116322]: getc() and ungetc() with unicode failure

ungetc() had no knowledge of UTF-8.  This patch adds it.

The committer fleshed out the author's code to make a patch, making
a few small changes.

9 years agoSome tests for [perl #112244]
Karl Williamson [Sun, 17 Feb 2013 21:46:29 +0000 (14:46 -0700)]
Some tests for [perl #112244]

Leon Timmermans is short of tuits right now, so I added these tests for
his patch.  Since I don't really know what I'm doing here, other tests
should eventually be added by someone who does know.

9 years agoPATCH [perl #112244]
Leon Timmermans [Thu, 5 Jul 2012 14:17:44 +0000 (16:17 +0200)]
PATCH [perl #112244]

:crlf currently doesn't fall back on :pending the way :perlio
does when the unread data doesn't fit into its own buffer. Instead it
just rejects them. This patch resolves that.  Tests are coming in a
future commit

The committer added a cast to get it to compile on Win32, and silence a
gcc warning on Linux

9 years agostart making entries for 5.19 releases
Ricardo Signes [Fri, 15 Feb 2013 20:37:05 +0000 (15:37 -0500)]
start making entries for 5.19 releases

9 years agoDocument d8a821fe in perldelta
Chris 'BinGOs' Williams [Mon, 18 Feb 2013 20:11:05 +0000 (20:11 +0000)]
Document d8a821fe in perldelta

9 years ago[perl #116809] Upstream status in corelist
Alexandr Ciornii [Mon, 18 Feb 2013 20:03:14 +0000 (20:03 +0000)]
[perl #116809] Upstream status in corelist

First patch explains in corelist output that deprecation means that
module will be CPAN-only.
Second patch adds new option, -u (upstream information). Shows if
module is primarily maintained in perl core or on CPAN and bug tracker
URL. Module::CoreList already contains this data, only output will be

Signed-off-by: Chris 'BinGOs' Williams <>
9 years agoDocument 721fe537 in perldelta
Chris 'BinGOs' Williams [Mon, 18 Feb 2013 19:04:21 +0000 (19:04 +0000)]
Document 721fe537 in perldelta

9 years agoDocument 0f87678 in perldelta
Chris 'BinGOs' Williams [Mon, 18 Feb 2013 18:57:51 +0000 (18:57 +0000)]
Document 0f87678 in perldelta

9 years agoDocument encoding deprecation in perldelta
Chris 'BinGOs' Williams [Mon, 18 Feb 2013 18:39:15 +0000 (18:39 +0000)]
Document encoding deprecation in perldelta

9 years agoUpdate Encode to CPAN version 2.48
Chris 'BinGOs' Williams [Mon, 18 Feb 2013 08:08:25 +0000 (08:08 +0000)]
Update Encode to CPAN version 2.48


  $Revision: 2.48 $ $Date: 2013/02/18 02:23:56 $
    t/ t/enc_data.t t/enc_eucjp.t t/enc_module.t t/enc_utf8.t
    t/encoding.t t/jperl.t
    [PATCH] Deprecate
  ! Encode/Supported.pod
    Fixed: Pod errors
  ! t/Encode.t
    [PATCH] Fix for shared hash key scalars
    Fixed: Uninitialized value warning from Encode->encodings()
  ! Makefile.PL
    Install to 'site' instead of 'perl' when perl version is 5.11+
  ! Encode/Makefile_PL.e2x
    find enc2xs.bat if it works on windows.
  ! t/piconv.t
    Fix finding piconv in t/piconv.t

9 years agoRemove two spaces after periods to keep the document consistent
Dave Rolsky [Mon, 18 Feb 2013 01:43:42 +0000 (19:43 -0600)]
Remove two spaces after periods to keep the document consistent

9 years agoRemove two superfluous commas in perlobj
Dave Rolsky [Mon, 18 Feb 2013 01:38:34 +0000 (19:38 -0600)]
Remove two superfluous commas in perlobj

9 years agoVery small tweak to perlobj in dereferencing method call example
Dave Rolsky [Mon, 18 Feb 2013 01:37:24 +0000 (19:37 -0600)]
Very small tweak to perlobj in dereferencing method call example

9 years agoReplace Mouse with Moo in perlootut
Dave Rolsky [Sun, 17 Feb 2013 23:55:16 +0000 (17:55 -0600)]
Replace Mouse with Moo in perlootut

9 years Add 'use re "/aa"'
Karl Williamson [Sun, 17 Feb 2013 21:17:27 +0000 (14:17 -0700)] Add 'use re "/aa"'

This would speed up compilation and execution by a trivial amount, but
makes sure that no regular expression even tries to compile in Unicode,
which could cause recursive invocation of this.  Before this construct
was invented, special care had to be taken to keep that from happening;
those measures remain in effect, but this gives added insurance.

9 years agoAdd Kent Fredric to AUTHORS
Father Chrysostomos [Sun, 17 Feb 2013 22:43:18 +0000 (14:43 -0800)]
Add Kent Fredric to AUTHORS

9 years agomg.c : revert ENV{x} = undef behaviour to be empty string, not key deletion
Kent Fredric [Sat, 16 Feb 2013 17:07:59 +0000 (06:07 +1300)]
mg.c : revert ENV{x} = undef behaviour to be empty string, not key deletion

pod/perldelta.pod: document reversion of ENV{foo} = undef behaviour  in delta
t/op/magic.t: add a test for ENV{foo} = undef

9 years agoAdded Lucas Holt to AUTHORS
Chris 'BinGOs' Williams [Sun, 17 Feb 2013 14:10:55 +0000 (14:10 +0000)]
Added Lucas Holt to AUTHORS

9 years ago[perl #116785] libc_r was removed from recent versions of MidnightBSD and older versi...
Lucas Holt [Sat, 16 Feb 2013 21:18:58 +0000 (13:18 -0800)]
[perl #116785] libc_r was removed from recent versions of MidnightBSD and older versions work better with -pthread

This corrects build errors with threading enabled on MidnightBSD 0.4-CURRENT

Signed-off-by: Chris 'BinGOs' Williams <>
9 years ago[perl #116587] Data::Dumper intermittently fails tests on 5.8.*
Steffen Mueller [Sun, 17 Feb 2013 12:02:23 +0000 (13:02 +0100)]
[perl #116587] Data::Dumper intermittently fails tests on 5.8.*

In a nutshell, we don't care enough about 5.8 to try really hard to fix
(obscure) vstring problems there. But we care enough not to make DD fail
its tests because of such a corner case. We don't want to screw all
users that are trapped in the last decade.

Thus, skip certain tests pre-5.10.

9 years agoIn, delete a hash slice, instead of using a loop.
Nicholas Clark [Sun, 17 Feb 2013 08:16:01 +0000 (09:16 +0100)]
In, delete a hash slice, instead of using a loop.

Deleting a hash slice compiles 5 fewer ops, and executes 21 fewer than
looping over the keys to delete each in turn. Whilst this is arguably a
micro-optimisation, it does not increase obfuscation and is in code loaded
by nearly every Perl program, so feels worthwhile.

9 years agoperldelta for 012528a99df1906d6fe.
Craig A. Berry [Sat, 16 Feb 2013 23:43:59 +0000 (17:43 -0600)]
perldelta for 012528a99df1906d6fe.

9 years agoImplement useversionedarchname in
Craig A. Berry [Sat, 16 Feb 2013 21:33:50 +0000 (15:33 -0600)]
Implement useversionedarchname in

If this ever becomes the default, we should really remove the
versioned subdirectory from archlib that we've always had.
As of now, using this feature on VMS gives us redundancies such as: