This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
perl5.git
18 years agoSince we no longer autovivify stashes (change #26370), we need
Rafael Garcia-Suarez [Mon, 2 Jan 2006 15:19:59 +0000 (15:19 +0000)]
Since we no longer autovivify stashes (change #26370), we need
to turn off strict-refs on them, or we'll have a stricture
error the first time we'd try to access them (when they'll be
actually autovivified).
p4raw-link: @26370 on //depot/perl: adc51b978ed1b2e9d4512c9bfa80386ac917d05a

p4raw-id: //depot/perl@26574

18 years agoDocumentation fix for File::Temp::new
pm@capmon.dk [Mon, 2 Jan 2006 02:48:41 +0000 (18:48 -0800)]
Documentation fix for File::Temp::new

Subject: [perl #38127] Bug in perldoc File::Temp
From: pm@capmon.dk (via RT) <perlbug-followup@perl.org>
Message-ID: <rt-3.0.11-38127-126764.0.578515622359532@perl.org>

p4raw-id: //depot/perl@26573

18 years agoHappy New Year. (It's not just embed.pl).
Nicholas Clark [Mon, 2 Jan 2006 14:43:37 +0000 (14:43 +0000)]
Happy New Year. (It's not just embed.pl).

p4raw-id: //depot/perl@26572

18 years agoNote that the "unlink"ing directories is attempted when running as a
Steve Peters [Mon, 2 Jan 2006 14:24:58 +0000 (14:24 +0000)]
Note that the "unlink"ing directories is attempted when running as a
superuser, and that some operating systems may not support it.

p4raw-id: //depot/perl@26571

18 years agoTypo fix in the description of change 26370.
Nicholas Clark [Mon, 2 Jan 2006 12:19:51 +0000 (12:19 +0000)]
Typo fix in the description of change 26370.
(Also corrected in the repository)

p4raw-id: //depot/perl@26570

18 years agoadd svt_local slot to magic vtable, and fix local $shared
Dave Mitchell [Mon, 2 Jan 2006 12:09:37 +0000 (12:09 +0000)]
add svt_local slot to magic vtable, and fix local $shared

p4raw-id: //depot/perl@26569

18 years agoAdd 2006 to the copyright years in the header generated by embed.pl
Nicholas Clark [Mon, 2 Jan 2006 11:48:28 +0000 (11:48 +0000)]
Add 2006 to the copyright years in the header generated by embed.pl

p4raw-id: //depot/perl@26565

18 years ago[patch] comment grammar and indentation, spurious tab
Jim Cromie [Fri, 30 Dec 2005 22:45:55 +0000 (15:45 -0700)]
[patch] comment grammar and indentation, spurious tab
Message-ID: <43B61B13.3050704@gmail.com>

p4raw-id: //depot/perl@26564

18 years ago[PAUSE] CPAN Upload: A/AN/ANDK/CPAN-1.80_58.tar.gz
Andreas König [Sun, 1 Jan 2006 10:12:46 +0000 (11:12 +0100)]
[PAUSE] CPAN Upload: A/AN/ANDK/CPAN-1.80_58.tar.gz
Message-ID: <87zmmgnwfl.fsf@k75.linux.bogus>

p4raw-id: //depot/perl@26563

18 years agoUpdate copyrights for files modified in 2006
Rafael Garcia-Suarez [Mon, 2 Jan 2006 10:56:05 +0000 (10:56 +0000)]
Update copyrights for files modified in 2006

p4raw-id: //depot/perl@26562

18 years agoGet rid of length argument for the FEATURE_IS_ENABLED() macro.
Gisle Aas [Sun, 1 Jan 2006 20:43:00 +0000 (20:43 +0000)]
Get rid of length argument for the FEATURE_IS_ENABLED() macro.

p4raw-id: //depot/perl@26561

18 years agoMaking 0 pointers to NULLs
Andy Lester [Fri, 30 Dec 2005 23:05:58 +0000 (17:05 -0600)]
Making 0 pointers to NULLs
Message-ID: <20051231050558.GA29093@petdance.com>

p4raw-id: //depot/perl@26558

18 years agoConfigure: stdio char (un)signedness
Jarkko Hietaniemi [Sat, 26 Nov 2005 11:25:18 +0000 (13:25 +0200)]
Configure: stdio char (un)signedness
Message-ID: <438829FE.8070202@gmail.com>

p4raw-id: //depot/perl@26557

18 years agoMaking 0 pointers to NULLs
Andy Lester [Fri, 30 Dec 2005 23:05:58 +0000 (17:05 -0600)]
Making 0 pointers to NULLs
Message-ID: <20051231050558.GA29093@petdance.com>

p4raw-id: //depot/perl@26556

18 years agomake threads use MY_CXT API rather than using PL_modglobal
Dave Mitchell [Sat, 31 Dec 2005 23:24:26 +0000 (23:24 +0000)]
make threads use MY_CXT API rather than using PL_modglobal

p4raw-id: //depot/perl@26555

18 years agotidy index arrangements in new MY_CXT code
Dave Mitchell [Sat, 31 Dec 2005 21:36:05 +0000 (21:36 +0000)]
tidy index arrangements in new MY_CXT code

p4raw-id: //depot/perl@26554

18 years agoUpgrade to CPAN-1.80_57
Steve Peters [Sat, 31 Dec 2005 18:45:37 +0000 (18:45 +0000)]
Upgrade to CPAN-1.80_57

p4raw-id: //depot/perl@26553

18 years agoTurn a for loop that's almost a while into an honest-to-goodness while.
Nicholas Clark [Sat, 31 Dec 2005 17:24:34 +0000 (17:24 +0000)]
Turn a for loop that's almost a while into an honest-to-goodness while.

p4raw-id: //depot/perl@26552

18 years agoCalling Perl_hv_clear_placeholders while the hash iterator was active
Nicholas Clark [Sat, 31 Dec 2005 16:57:32 +0000 (16:57 +0000)]
Calling Perl_hv_clear_placeholders while the hash iterator was active
would turn lazy delete on, causing the hash to become corrupted at the
next iterator change.

p4raw-id: //depot/perl@26551

18 years agoRemove unneeded test in Perl_hv_clear_placeholders.
Nicholas Clark [Sat, 31 Dec 2005 16:42:08 +0000 (16:42 +0000)]
Remove unneeded test in Perl_hv_clear_placeholders.
Rejig S_hfreeentries to a double loop, which is clearer and smaller.

p4raw-id: //depot/perl@26550

18 years agoop/pack.t failure on Win64
Jan Dubois [Fri, 30 Dec 2005 21:55:00 +0000 (21:55 +0000)]
op/pack.t failure on Win64
Message-ID: <017a01c60d83$5dc73110$2217a8c0@candy>

p4raw-id: //depot/perl@26549

18 years agoAdd a test for change 26547. (We no longer expect to see errors about
Nicholas Clark [Sat, 31 Dec 2005 15:07:16 +0000 (15:07 +0000)]
Add a test for change 26547. (We no longer expect to see errors about
unbalanced string table reference counts.)

p4raw-id: //depot/perl@26548

18 years agoRe-attempt to clear the hash in S_hfreeentries if anything adds to it.
Nicholas Clark [Sat, 31 Dec 2005 14:57:27 +0000 (14:57 +0000)]
Re-attempt to clear the hash in S_hfreeentries if anything adds to it.
Panic if we seem to be looping forever.

p4raw-id: //depot/perl@26547

18 years agoStop 'sub a { // }; %::=()' panicing because the backreference array
Nicholas Clark [Sat, 31 Dec 2005 13:16:06 +0000 (13:16 +0000)]
Stop 'sub a { // }; %::=()' panicing because the backreference array
became lost.

p4raw-id: //depot/perl@26546

18 years agoMarcus Holland-Moritz suggested that assert should
Nicholas Clark [Fri, 30 Dec 2005 22:44:06 +0000 (22:44 +0000)]
Marcus Holland-Moritz suggested that assert should
Perl_croak_nocontext rather than Perl_croak.

p4raw-id: //depot/perl@26545

18 years agoTweak S_init_main_stash so as allocate PL_curstname as a shared string
Nicholas Clark [Fri, 30 Dec 2005 20:02:56 +0000 (20:02 +0000)]
Tweak S_init_main_stash so as allocate PL_curstname as a shared string
scalar, and hence avoid thrashing the shared string table for "main".

p4raw-id: //depot/perl@26544

18 years agoConvert t/op/sleep.t to using test.pl
Steve Peters [Fri, 30 Dec 2005 17:47:45 +0000 (17:47 +0000)]
Convert t/op/sleep.t to using test.pl

p4raw-id: //depot/perl@26543

18 years agoDocument why it's a bad plan to move the backreferences array from
Nicholas Clark [Fri, 30 Dec 2005 16:53:39 +0000 (16:53 +0000)]
Document why it's a bad plan to move the backreferences array from
the magic structure to the hv_aux structure during backreference
deletion. Add an optimisation for an empty backreference array.

p4raw-id: //depot/perl@26542

18 years agoNow that the backreference array has no NULL entries, and is correctly
Nicholas Clark [Fri, 30 Dec 2005 15:55:05 +0000 (15:55 +0000)]
Now that the backreference array has no NULL entries, and is correctly
marked as AvREAL_off(), there's no reason not to use sv_dup to clone
it.

p4raw-id: //depot/perl@26541

18 years agoPOSIX now declares missing constants as sub ... ();, so Net::Ping
Nicholas Clark [Fri, 30 Dec 2005 15:39:53 +0000 (15:39 +0000)]
POSIX now declares missing constants as sub ... ();, so Net::Ping
needs to define them with a prototype to match.

p4raw-id: //depot/perl@26540

18 years agoUpdate perltodo. First shot at perl593delta.
Rafael Garcia-Suarez [Fri, 30 Dec 2005 15:10:41 +0000 (15:10 +0000)]
Update perltodo. First shot at perl593delta.

p4raw-id: //depot/perl@26539

18 years agoIn the shared string table, if we add new entries to the head of the
Nicholas Clark [Fri, 30 Dec 2005 14:57:00 +0000 (14:57 +0000)]
In the shared string table, if we add new entries to the head of the
linked list chain, then we don't need as much bookkeeping in the loop.

p4raw-id: //depot/perl@26538

18 years agoDifferences from Perl 6
Robin Houston [Fri, 30 Dec 2005 13:45:34 +0000 (13:45 +0000)]
Differences from Perl 6
Message-ID: <20051230134534.GA18675@rpc142.cs.man.ac.uk>

p4raw-id: //depot/perl@26537

18 years agoRe: [PATCH] Re: INSTALLSCRIPT versus INSTALLDIRS
Rafael Garcia-Suarez [Mon, 12 Dec 2005 19:22:32 +0000 (20:22 +0100)]
Re: [PATCH] Re: INSTALLSCRIPT versus INSTALLDIRS
Message-ID: <20051212192232.39cd0c1e@grubert.mandrakesoft.com>

p4raw-id: //depot/perl@26536

18 years agofix errors in new MY_CTX code
Dave Mitchell [Fri, 30 Dec 2005 12:21:05 +0000 (12:21 +0000)]
fix errors in new MY_CTX code

p4raw-id: //depot/perl@26535

18 years agoReduce size of buffers for identifier names, as suggested in
Tels [Thu, 29 Dec 2005 18:40:49 +0000 (19:40 +0100)]
Reduce size of buffers for identifier names, as suggested in

Subject: Stack usage (in gv_stashpvn and others)
Message-Id: <200512291840.50765@bloodgate.com>

p4raw-id: //depot/perl@26534

18 years agoperlio.c - gv_stashpvn() vs gv_stashpv()
Tels [Thu, 29 Dec 2005 18:57:43 +0000 (19:57 +0100)]
perlio.c - gv_stashpvn() vs gv_stashpv()
Message-Id: <200512291857.59458@bloodgate.com>

p4raw-id: //depot/perl@26533

18 years agoPerl_gv_fetchpv vs. gv_fetchpvn
Tels [Thu, 29 Dec 2005 18:32:10 +0000 (19:32 +0100)]
Perl_gv_fetchpv vs. gv_fetchpvn
Message-Id: <200512291832.20651@bloodgate.com>

p4raw-id: //depot/perl@26532

18 years agoSkip my_cxt_init if PERL_IMPLICIT_CONTEXT isn't defined
Rafael Garcia-Suarez [Fri, 30 Dec 2005 09:53:54 +0000 (09:53 +0000)]
Skip my_cxt_init if PERL_IMPLICIT_CONTEXT isn't defined
(spotted by Jan Dubois)

p4raw-id: //depot/perl@26531

18 years agoRMAGIC on symbol tables is bad, m'kay.
Nicholas Clark [Fri, 30 Dec 2005 01:08:46 +0000 (01:08 +0000)]
RMAGIC on symbol tables is bad, m'kay.
Allow hashes (and therefore all symbol tables) to store the
backreference array in the hv_aux structure, and thereby undo the
performance damage of 24966, which resulted in 60% of all hash lookups
trying to mg_find tiehash magic.

p4raw-id: //depot/perl@26530

18 years agochange #26523 didn't make my_cxt_init public
Dave Mitchell [Fri, 30 Dec 2005 01:06:05 +0000 (01:06 +0000)]
change #26523 didn't make my_cxt_init public
p4raw-link: @26523 on //depot/perl: f16dd614412ea67a8eb64bb09a88fccdbd9db6b6

p4raw-id: //depot/perl@26529

18 years ago$a = \$a oddity
Tels [Thu, 29 Dec 2005 18:01:42 +0000 (19:01 +0100)]
$a = \$a oddity
Message-Id: <200512291802.00742@bloodgate.com>

plus trim some whitespace

p4raw-id: //depot/perl@26528

18 years agoActual real consting of parms
Andy Lester [Thu, 29 Dec 2005 01:15:09 +0000 (19:15 -0600)]
Actual real consting of parms
Message-ID: <20051229071509.GA32653@petdance.com>

p4raw-id: //depot/perl@26527

18 years agoav.c cleanup, plus!
Andy Lester [Wed, 28 Dec 2005 15:44:14 +0000 (09:44 -0600)]
av.c cleanup, plus!
Message-ID: <20051228214414.GD26033@petdance.com>

p4raw-id: //depot/perl@26526

18 years agoFix test for perl version number in Makefile.PL
Rafael Garcia-Suarez [Thu, 29 Dec 2005 14:32:09 +0000 (14:32 +0000)]
Fix test for perl version number in Makefile.PL

p4raw-id: //depot/perl@26525

18 years agoFixing the SVREFCNT macro
Andy Lester [Wed, 28 Dec 2005 13:30:51 +0000 (07:30 -0600)]
Fixing the SVREFCNT macro
Message-ID: <20051228193051.GD23207@petdance.com>

p4raw-id: //depot/perl@26524

18 years agore-implement MY_CXT API more efficiently, and add explicit
Dave Mitchell [Thu, 29 Dec 2005 12:00:29 +0000 (12:00 +0000)]
re-implement MY_CXT API more efficiently, and add explicit
interpeter arg variant

p4raw-id: //depot/perl@26523

18 years agoadd tests for MY_CXT API and improve its documentation
Dave Mitchell [Thu, 29 Dec 2005 11:35:04 +0000 (11:35 +0000)]
add tests for MY_CXT API and improve its documentation

p4raw-id: //depot/perl@26522

18 years agoTest that constant overloading is propagated into eval
Robin Houston [Mon, 26 Dec 2005 17:56:48 +0000 (17:56 +0000)]
Test that constant overloading is propagated into eval
Message-ID: <20051226175648.GA28402@rpc142.cs.man.ac.uk>

p4raw-id: //depot/perl@26521

18 years ago[DOC PATCH] rindex clarification
Yitzchak Scott-Thoennes [Mon, 26 Dec 2005 18:42:40 +0000 (10:42 -0800)]
[DOC PATCH] rindex clarification
Message-ID: <20051227024240.GA2940@efn.org>

plus another formatting nit

p4raw-id: //depot/perl@26520

18 years agofortify Pod::Simple::Search against non-case-preserving filesystems
Craig A. Berry [Tue, 27 Dec 2005 17:59:35 +0000 (11:59 -0600)]
fortify Pod::Simple::Search against non-case-preserving filesystems
From: "Craig A. Berry" <craigberry@mac.com>
Message-id: <43B1D567.9080504@mac.com>

p4raw-id: //depot/perl@26519

18 years agoA GVs stash can be NULL, so don't call macros that assume otherwise
Nicholas Clark [Thu, 29 Dec 2005 00:19:01 +0000 (00:19 +0000)]
A GVs stash can be NULL, so don't call macros that assume otherwise
without checking.

p4raw-id: //depot/perl@26518

18 years agoAs the backreferences AV doesn't hold references on its contents,
Nicholas Clark [Wed, 28 Dec 2005 19:40:26 +0000 (19:40 +0000)]
As the backreferences AV doesn't hold references on its contents,
surely it should have AvREAL turned off?

p4raw-id: //depot/perl@26517

18 years agoPL_amagic_generation is always non-zero, so remove the test for this.
Nicholas Clark [Wed, 28 Dec 2005 13:35:03 +0000 (13:35 +0000)]
PL_amagic_generation is always non-zero, so remove the test for this.
(Since change 17990 added version object overloading)

p4raw-id: //depot/perl@26516

18 years agoUpgrade to Sys-Syslog-0.11
Steve Peters [Wed, 28 Dec 2005 13:31:02 +0000 (13:31 +0000)]
Upgrade to Sys-Syslog-0.11

p4raw-id: //depot/perl@26515

18 years agoCleanup; replace a croak("panic: ...") with assert().
Gisle Aas [Wed, 28 Dec 2005 13:25:30 +0000 (13:25 +0000)]
Cleanup; replace a croak("panic: ...") with assert().

p4raw-id: //depot/perl@26514

18 years agoassert(av); Cleanup the way that the av_*() functions check
Gisle Aas [Wed, 28 Dec 2005 13:12:30 +0000 (13:12 +0000)]
assert(av);  Cleanup the way that the av_*() functions check
that they are really passed an av argument.
Should improve the test coverage a bit.

p4raw-id: //depot/perl@26513

18 years agoUse memchr() instead of ninstr() to locate the end-of-line.
Gisle Aas [Wed, 28 Dec 2005 12:38:03 +0000 (12:38 +0000)]
Use memchr() instead of ninstr() to locate the end-of-line.
The call to memchr() will faster.

p4raw-id: //depot/perl@26512

18 years agoLeaner ninstr().
Gisle Aas [Wed, 28 Dec 2005 12:18:58 +0000 (12:18 +0000)]
Leaner ninstr().
The compiled function ended up 37% smaller on Linux/gcc.
I ought to be faster as well, but I did not try to measure that.

p4raw-id: //depot/perl@26511

18 years agoRemove test for NUL ending in r?instr().
Gisle Aas [Wed, 28 Dec 2005 11:01:46 +0000 (11:01 +0000)]
Remove test for NUL ending in r?instr().
This test is actually harmful because we will search
for any garbage found past the empty search string.
Ref change 26509.

p4raw-id: //depot/perl@26510

18 years agoFix 2 off-by-one errors in the call to ninstr().
Gisle Aas [Wed, 28 Dec 2005 10:57:46 +0000 (10:57 +0000)]
Fix 2 off-by-one errors in the call to ninstr().
This code managed to still work because of a bug
in how ninstr() treats empty search strings.

p4raw-id: //depot/perl@26509

18 years agoUpgrade to PathTools-3.15
Steve Peters [Wed, 28 Dec 2005 03:27:13 +0000 (03:27 +0000)]
Upgrade to PathTools-3.15

p4raw-id: //depot/perl@26508

18 years agos/%/%%/ in the package name so that it can be used safely in the
Nicholas Clark [Tue, 27 Dec 2005 23:39:42 +0000 (23:39 +0000)]
s/%/%%/ in the package name so that it can be used safely in the
sprintf format string (and hence not need runtime %s interpolation)

p4raw-id: //depot/perl@26507

18 years agoPass in the full arguments to ExtUtils::Constant::ProxySubs
Nicholas Clark [Tue, 27 Dec 2005 22:19:28 +0000 (22:19 +0000)]
Pass in the full arguments to ExtUtils::Constant::ProxySubs
Prototype all the missing constants, so that parsing doesn't depend on
the phase of the moon. (Well, the system headers)
Add an option for making missing constants generate errors at read time
(which can be during constant folding). This isn't the default.

p4raw-id: //depot/perl@26506

18 years agopre-likely cleanup
Andy Lester [Tue, 27 Dec 2005 14:39:39 +0000 (08:39 -0600)]
pre-likely cleanup
Message-ID: <20051227203939.GC1781@petdance.com>

Includes a small fix to the changes in tryAMAGICbinW_var() in pp.h.

p4raw-id: //depot/perl@26505

18 years agoIt'd be nice if I hadn't broken corecpan.pl. Its fixed now, though.
Steve Peters [Tue, 27 Dec 2005 18:36:19 +0000 (18:36 +0000)]
It'd be nice if I hadn't broken corecpan.pl.  Its fixed now, though.

p4raw-id: //depot/perl@26504

18 years agogv_fetch_flags in newATTRSUB can actually be const.
Nicholas Clark [Tue, 27 Dec 2005 17:33:13 +0000 (17:33 +0000)]
gv_fetch_flags in newATTRSUB can actually be const.
Comment the logic behind the terms in the ternary that chooses the
flags value.

p4raw-id: //depot/perl@26503

18 years agoAs what we're doing is generating subroutines, which are methods, we
Nicholas Clark [Tue, 27 Dec 2005 17:24:11 +0000 (17:24 +0000)]
As what we're doing is generating subroutines, which are methods, we
need to invalidate cached methods. For constant.pm we need to create
&Internals::inc_sub_generation

p4raw-id: //depot/perl@26502

18 years ago[patch] ext/b/t/lint.t fails on win32
Yves Orton [Fri, 23 Dec 2005 20:06:52 +0000 (21:06 +0100)]
[patch] ext/b/t/lint.t fails on win32
Message-ID: <9b18b3110512231106k7ad395bfp8a0c05cd0f87fc61@mail.gmail.com>

p4raw-id: //depot/perl@26501

18 years agoRefactor the require failure message generation to use 1 less temporary
Nicholas Clark [Tue, 27 Dec 2005 12:02:57 +0000 (12:02 +0000)]
Refactor the require failure message generation to use 1 less temporary
SV, and to build as much text as possible with newSVpvf

p4raw-id: //depot/perl@26500

18 years agoRemoving MAKE_JOBS_FIFO from %ENV causes FreeBSD make to forget about
Nicholas Clark [Tue, 27 Dec 2005 00:29:33 +0000 (00:29 +0000)]
Removing MAKE_JOBS_FIFO from %ENV causes FreeBSD make to forget about
any -j flags. (And their implied disabling of backwards compatibility,
which is the real cause of the make disttest failure).

p4raw-id: //depot/perl@26499

18 years agotabs, not spaces, in MANIFEST. No, I can't remember why this is.
Nicholas Clark [Mon, 26 Dec 2005 23:52:44 +0000 (23:52 +0000)]
tabs, not spaces, in MANIFEST. No, I can't remember why this is.
Or whether it could be relaxed.

p4raw-id: //depot/perl@26498

18 years agoRemove the hard wired test numbers in the generated test script for
Nicholas Clark [Mon, 26 Dec 2005 23:50:14 +0000 (23:50 +0000)]
Remove the hard wired test numbers in the generated test script for
simple tests. Use this script twice, firstly for testing the old style
autoloaded constants, then for testing the ProxySubs.
Make testing the dogfood/regeneration code optional, as the ProxySubs
output doesn't contain it.

p4raw-id: //depot/perl@26497

18 years agoSupport type UNDEF, and correct a bug in the macro generation by using
Nicholas Clark [Mon, 26 Dec 2005 23:48:00 +0000 (23:48 +0000)]
Support type UNDEF, and correct a bug in the macro generation by using
macro_from_item.

p4raw-id: //depot/perl@26496

18 years agoAdd method macro_from_item to encapsulate the entire logic for getting
Nicholas Clark [Mon, 26 Dec 2005 23:47:01 +0000 (23:47 +0000)]
Add method macro_from_item to encapsulate the entire logic for getting
the correct macro direct from the item hash.

p4raw-id: //depot/perl@26495

18 years agoDon't try to store PL_sv_undef in pads. (For now)
Nicholas Clark [Mon, 26 Dec 2005 23:10:53 +0000 (23:10 +0000)]
Don't try to store PL_sv_undef in pads. (For now)

p4raw-id: //depot/perl@26494

18 years agoUpgrade to CPAN-1.80_56
Steve Peters [Mon, 26 Dec 2005 22:01:49 +0000 (22:01 +0000)]
Upgrade to CPAN-1.80_56

p4raw-id: //depot/perl@26493

18 years agoIn this, the last tale of the NGROUPS saga, a former pumpking prods
Steve Peters [Mon, 26 Dec 2005 17:29:13 +0000 (17:29 +0000)]
In this, the last tale of the NGROUPS saga, a former pumpking prods
a mere committer to remove the last of the NGROUPS-sized arrays...

Perl_magic_set() was using the last of these arrays to do the
lvalue work on $).  Instead of an array, a pointer is used and
re-sized as needed.

p4raw-id: //depot/perl@26492

18 years agoAdd a couple of newlines for tidier regen output, consistent with the
Nicholas Clark [Mon, 26 Dec 2005 17:01:35 +0000 (17:01 +0000)]
Add a couple of newlines for tidier regen output, consistent with the
code generated by WriteConstants()

p4raw-id: //depot/perl@26491

18 years agoAdd C_FH and XS_FH arguments to ExtUtils::Constant::WriteConstants,
Nicholas Clark [Mon, 26 Dec 2005 16:13:57 +0000 (16:13 +0000)]
Add C_FH and XS_FH arguments to ExtUtils::Constant::WriteConstants,
to allow the caller to pass in file handles. Use this in Contant.t
with tied file handles to capture the output, rather than calling
the lower level routines directly.

p4raw-id: //depot/perl@26490

18 years agochg 26486 removes last use of NGROUPS, so remove the macro-def
Jim Cromie [Mon, 26 Dec 2005 07:57:51 +0000 (00:57 -0700)]
chg 26486 removes last use of NGROUPS, so remove the macro-def
Message-ID: <43B004EF.8060500@gmail.com>

p4raw-id: //depot/perl@26489

18 years agoThe values for proxy constant subroutines need to be read only.
Nicholas Clark [Mon, 26 Dec 2005 12:48:10 +0000 (12:48 +0000)]
The values for proxy constant subroutines need to be read only.

p4raw-id: //depot/perl@26488

18 years agoRework constant.pm to take advantage of the space savings of proxy
Nicholas Clark [Mon, 26 Dec 2005 12:31:01 +0000 (12:31 +0000)]
Rework constant.pm to take advantage of the space savings of proxy
constant subroutines whenever it can.

p4raw-id: //depot/perl@26487

18 years agoObviously, Perl_ingroup() is also using 256k of stack memory on Linux.
Steve Peters [Mon, 26 Dec 2005 03:51:24 +0000 (03:51 +0000)]
Obviously, Perl_ingroup() is also using 256k of stack memory on Linux.
Adapt change #26480 to reduce memory usage here as well.
p4raw-link: @26480 on //depot/perl: 57d7c65eded7a5f963c5ce38ee196978a06e35df

p4raw-id: //depot/perl@26486

18 years agoThe value from caller doesn't change, so my $pkg = caller;
Nicholas Clark [Sat, 24 Dec 2005 16:27:28 +0000 (16:27 +0000)]
The value from caller doesn't change, so my $pkg = caller;
should be moved out of the loop.

p4raw-id: //depot/perl@26485

18 years agoTests for creating constants where prototypes or other symbols of the
Nicholas Clark [Sat, 24 Dec 2005 16:06:10 +0000 (16:06 +0000)]
Tests for creating constants where prototypes or other symbols of the
same name already exist.

p4raw-id: //depot/perl@26484

18 years agoUsing a hash reference rather than a hash can save a hash copy.
Nicholas Clark [Sat, 24 Dec 2005 15:09:28 +0000 (15:09 +0000)]
Using a hash reference rather than a hash can save a hash copy.

p4raw-id: //depot/perl@26483

18 years agoUnless the peephole optimiser already knows that we're in void context,
Nicholas Clark [Sat, 24 Dec 2005 14:48:42 +0000 (14:48 +0000)]
Unless the peephole optimiser already knows that we're in void context,
avoid even attempting the whole proxy constant sub copying
optimisation.

p4raw-id: //depot/perl@26482

18 years agoGiven that IO::Socket is documented as exporting all of Socket's
Nicholas Clark [Sat, 24 Dec 2005 14:09:56 +0000 (14:09 +0000)]
Given that IO::Socket is documented as exporting all of Socket's
exports, there's no need for IO::Socket::INET or IO::Socket::UNIX to
use both. (And re-exporting removes some of the efficiency of the new
constants implementation)

p4raw-id: //depot/perl@26481

18 years agoRe: [perl #37907] Perl_magic_get reqires ridiculous amounts of stackmemory
Jim Cromie [Fri, 23 Dec 2005 13:24:29 +0000 (06:24 -0700)]
Re: [perl #37907] Perl_magic_get reqires ridiculous amounts of stackmemory
Message-ID: <43AC5CFD.3060200@gmail.com>

Includes a change in variable names from "j" to "num_groups".

p4raw-id: //depot/perl@26480

18 years agoLooks like the macro CLK_TCK is not actually a constant on Solaris.
Nicholas Clark [Sat, 24 Dec 2005 00:01:50 +0000 (00:01 +0000)]
Looks like the macro CLK_TCK is not actually a constant on Solaris.

p4raw-id: //depot/perl@26479

18 years agoFor PERL_TRACK_MEMPOOL with PERL_POISON, Poison the end of any block
Nicholas Clark [Fri, 23 Dec 2005 17:30:19 +0000 (17:30 +0000)]
For PERL_TRACK_MEMPOOL with PERL_POISON, Poison the end of any block
being shrunk in realloc()

p4raw-id: //depot/perl@26478

18 years agoFor the rare case of EMFILE during require, save object code space
Nicholas Clark [Fri, 23 Dec 2005 17:24:44 +0000 (17:24 +0000)]
For the rare case of EMFILE during require, save object code space
by using newSVpvf
(Would it be worth using it for all require failures?)

p4raw-id: //depot/perl@26477

18 years agoIf PERL_TRACK_MEMPOOL and PERL_POISON are in use, then scribble all
Nicholas Clark [Fri, 23 Dec 2005 16:55:35 +0000 (16:55 +0000)]
If PERL_TRACK_MEMPOOL and PERL_POISON are in use, then scribble all
over memory to invalidate it just before free()ing it.

p4raw-id: //depot/perl@26476

18 years agoProvide support for types PVN and UNDEF in
Nicholas Clark [Fri, 23 Dec 2005 16:20:53 +0000 (16:20 +0000)]
Provide support for types PVN and UNDEF in
 ExtUtils::Constant::ProxySubs

p4raw-id: //depot/perl@26475

18 years agoCreate a struct to use as the header with PERL_TRACK_MEMPOOL, so that
Nicholas Clark [Fri, 23 Dec 2005 15:27:10 +0000 (15:27 +0000)]
Create a struct to use as the header with PERL_TRACK_MEMPOOL, so that
other information can be stored in it.

p4raw-id: //depot/perl@26474

18 years agoExtUtils::ParseXS needs to stack $filepathname to make #line directives
Nicholas Clark [Fri, 23 Dec 2005 13:54:37 +0000 (13:54 +0000)]
ExtUtils::ParseXS needs to stack $filepathname to make #line directives
in #INCLUDEs work.

p4raw-id: //depot/perl@26473

18 years agoAdd support for PV to ExtUtils::Constant::ProxySubs, and enable its
Nicholas Clark [Fri, 23 Dec 2005 12:21:19 +0000 (12:21 +0000)]
Add support for PV to ExtUtils::Constant::ProxySubs, and enable its
use in Sys::Syslog

p4raw-id: //depot/perl@26472

18 years agoinlining autoloaded constants is TODOne.
Nicholas Clark [Fri, 23 Dec 2005 12:01:29 +0000 (12:01 +0000)]
inlining autoloaded constants is TODOne.

p4raw-id: //depot/perl@26471

18 years agoNow that proxy subroutines can be unproxied in 2 places without
Nicholas Clark [Fri, 23 Dec 2005 11:36:43 +0000 (11:36 +0000)]
Now that proxy subroutines can be unproxied in 2 places without
becoming the same newCONSTSUB, need an explicit check to avoid a
warning about subroutines being redefined, as there has never been
a warning when you assing the same subroutine to a glob's GvCV()

p4raw-id: //depot/perl@26470