This is a live mirror of the Perl 5 development currently hosted at
16 years agoAbstract the specific use of SvCUR in GVs for detecting variables on
Nicholas Clark [Sat, 25 Feb 2006 16:48:46 +0000 (16:48 +0000)]
Abstract the specific use of SvCUR in GVs for detecting variables on
both sides of an assignment using GvASSIGN_GENERATION() and

p4raw-id: //depot/perl@27327

16 years agoxgv_stash can be stored as a union with the NV, reducing the size of
Nicholas Clark [Sat, 25 Feb 2006 15:38:25 +0000 (15:38 +0000)]
xgv_stash can be stored as a union with the NV, reducing the size of
PVGVs by another pointer.

p4raw-id: //depot/perl@27326

16 years agoCuriously none of the macros SvIVX, SvUVX or SvNVX are used on
Nicholas Clark [Sat, 25 Feb 2006 15:11:28 +0000 (15:11 +0000)]
Curiously none of the macros SvIVX, SvUVX or SvNVX are used on

p4raw-id: //depot/perl@27325

16 years agoAmazingly, it seems that none of the other GV specific macros are
Nicholas Clark [Sat, 25 Feb 2006 14:50:38 +0000 (14:50 +0000)]
Amazingly, it seems that none of the other GV specific macros are
accessed unless the GV thinks that it is a GV.

p4raw-id: //depot/perl@27324

16 years agoStore GvGP in the SV head union. For all the common lookups [eg GvCV()]
Nicholas Clark [Sat, 25 Feb 2006 00:39:30 +0000 (00:39 +0000)]
Store GvGP in the SV head union. For all the common lookups [eg GvCV()]
this avoids 1 pointer dereference and the associated risk of a CPU
cache miss. Although this patch looks deceptively small, I fear its
CBV(*) might be rather high.
(* Crack By Volume)

p4raw-id: //depot/perl@27323

16 years agoWith PERL_DONT_CREATE_GVSV, don't need to create a new GvSV in
Nicholas Clark [Fri, 24 Feb 2006 23:30:50 +0000 (23:30 +0000)]
With PERL_DONT_CREATE_GVSV, don't need to create a new GvSV in

p4raw-id: //depot/perl@27322

16 years agoGiven that SvPV on a typeglob is now irrelevant, really we only need
Nicholas Clark [Fri, 24 Feb 2006 22:43:14 +0000 (22:43 +0000)]
Given that SvPV on a typeglob is now irrelevant, really we only need
to store the GvGP.

p4raw-id: //depot/perl@27321

16 years agoGiven that Perl_gp_free() is refcount-dec-and-maybe-free for the glob
Nicholas Clark [Fri, 24 Feb 2006 21:42:33 +0000 (21:42 +0000)]
Given that Perl_gp_free() is refcount-dec-and-maybe-free for the glob
pointer, it's inconsistent that it only sets this GV's GvGP to 0 if
this GV happened to have the last reference. Why should this GV care if
it had the last reference? So always set it to 0. It's free. Gone.

p4raw-id: //depot/perl@27320

16 years agoAvoid special case SvMAGIC macros just to cope with Perl_sv_unmagic.
Nicholas Clark [Fri, 24 Feb 2006 19:32:11 +0000 (19:32 +0000)]
Avoid special case SvMAGIC macros just to cope with Perl_sv_unmagic.

p4raw-id: //depot/perl@27319

16 years agoUse PVMGs rather than PVGVs to store the names of our variables in the
Nicholas Clark [Fri, 24 Feb 2006 14:19:37 +0000 (14:19 +0000)]
Use PVMGs rather than PVGVs to store the names of our variables in the

p4raw-id: //depot/perl@27313

16 years agoStore the stash for our in the magic slot. This will allow us to use
Nicholas Clark [Fri, 24 Feb 2006 13:59:57 +0000 (13:59 +0000)]
Store the stash for our in the magic slot. This will allow us to use
PVMGs in pad names where previously PVGVs were used. In turn, this
gives much greater flexibility for the layout of PVGVs.

p4raw-id: //depot/perl@27312

16 years agoUpgrade to CPAN-1.86
Steve Peters [Fri, 24 Feb 2006 13:55:23 +0000 (13:55 +0000)]
Upgrade to CPAN-1.86

p4raw-id: //depot/perl@27311

16 years agoSkip unportable test on VMS
Rafael Garcia-Suarez [Fri, 24 Feb 2006 12:52:01 +0000 (12:52 +0000)]
Skip unportable test on VMS

p4raw-id: //depot/perl@27309

16 years agoAssert that GvSTASH is only called on PVGVs and PVLVs.
Nicholas Clark [Fri, 24 Feb 2006 11:52:53 +0000 (11:52 +0000)]
Assert that GvSTASH is only called on PVGVs and PVLVs.

p4raw-id: //depot/perl@27307

16 years agoReplace usage of GvSTASH for storing the stash of C<our> with
Nicholas Clark [Fri, 24 Feb 2006 11:52:28 +0000 (11:52 +0000)]
Replace usage of GvSTASH for storing the stash of C<our> with
OURSTASH. Set the stash with OURSTASH_SET.

p4raw-id: //depot/perl@27306

16 years agoassert that SvMAGIC is never non-NULL for any PVMG used as a pad name.
Nicholas Clark [Fri, 24 Feb 2006 11:21:06 +0000 (11:21 +0000)]
assert that SvMAGIC is never non-NULL for any PVMG used as a pad name.

p4raw-id: //depot/perl@27304

16 years agoRedo previous sv.h change.
Nicholas Clark [Fri, 24 Feb 2006 10:56:00 +0000 (10:56 +0000)]
Redo previous sv.h change.

p4raw-id: //depot/perl@27302

16 years agoPerl_save_aelem can't be const
Andy Lester [Thu, 23 Feb 2006 22:53:40 +0000 (16:53 -0600)]
Perl_save_aelem can't be const
Message-ID: <>

p4raw-id: //depot/perl@27301

16 years agounused context warnings
Andy Lester [Tue, 21 Feb 2006 00:27:11 +0000 (18:27 -0600)]
unused context warnings
Message-ID: <>

p4raw-id: //depot/perl@27300

16 years agoAs SVs can't both be in a pad, and the name of a variable in the pad,
Nicholas Clark [Fri, 24 Feb 2006 10:29:48 +0000 (10:29 +0000)]
As SVs can't both be in a pad, and the name of a variable in the pad,
overlay SVs_PADTMP with SVpad_TYPED and SVs_PADMY with SVpad_OUR.

p4raw-id: //depot/perl@27297

16 years agoRevert change #27295, which I thought fixed builds on Win32.
Steve Peters [Fri, 24 Feb 2006 02:56:02 +0000 (02:56 +0000)]
Revert change #27295, which I thought fixed builds on Win32.
Instead remove &PL_vtbl_glob from globvar.sym.
p4raw-link: @27295 on //depot/perl: 3476b56103cbe13508b1fd6b46ae7b9cb6e0f7ed

p4raw-id: //depot/perl@27296

16 years agoLast instance of PL_vtbl_glob removed. It was missed change #27289.
Steve Peters [Fri, 24 Feb 2006 02:02:14 +0000 (02:02 +0000)]
Last instance of PL_vtbl_glob removed.  It was missed change #27289.
p4raw-link: @27289 on //depot/perl: c0c446747ad6c5bde53bc8415ca16ef77f6320f2

p4raw-id: //depot/perl@27295

16 years agoReplace direct flags tests & manipulations for SVpad_TYPED and
Nicholas Clark [Thu, 23 Feb 2006 23:45:39 +0000 (23:45 +0000)]
Replace direct flags tests & manipulations for SVpad_TYPED and
SVpad_OUR with macros SvPAD_TYPED(), SvPAD_OUR() etc, to abstract
away the flags bits acutally used to store this information.

p4raw-id: //depot/perl@27294

16 years agoDocument a sixth use for SVf_FAKE
Nicholas Clark [Thu, 23 Feb 2006 22:17:07 +0000 (22:17 +0000)]
Document a sixth use for SVf_FAKE

p4raw-id: //depot/perl@27293

16 years agoWe have sufficient spare bits to store the SV type in body_details,
Nicholas Clark [Thu, 23 Feb 2006 20:28:34 +0000 (20:28 +0000)]
We have sufficient spare bits to store the SV type in body_details,
so add a small sanity check to ensure that we have the array in the
correct order.

p4raw-id: //depot/perl@27292

16 years agoShrink struct body_details and hence sv.o slightly further by using
Nicholas Clark [Thu, 23 Feb 2006 19:40:23 +0000 (19:40 +0000)]
Shrink struct body_details and hence sv.o slightly further by using

p4raw-id: //depot/perl@27291

16 years agoUsing U8 rather than size type shrinks the body_details table.
Nicholas Clark [Thu, 23 Feb 2006 19:15:48 +0000 (19:15 +0000)]
Using U8 rather than size type shrinks the body_details table.

p4raw-id: //depot/perl@27290

16 years agoRemove set magic from typeglobs. Remove typeglob magic entirely.
Nicholas Clark [Thu, 23 Feb 2006 18:00:19 +0000 (18:00 +0000)]
Remove set magic from typeglobs. Remove typeglob magic entirely.
Typeglobs now never access the SvPVX, SvIVX or SvNVX when holding a
valid GvGP().

p4raw-id: //depot/perl@27289

16 years agoIn XS_attributes__guess_stash, attempting to call Gv* on a PVMG is
Nicholas Clark [Thu, 23 Feb 2006 17:45:55 +0000 (17:45 +0000)]
In XS_attributes__guess_stash, attempting to call Gv* on a PVMG is
bad and wrong.

p4raw-id: //depot/perl@27288

16 years agoAdd test for RT #2166: foreach spuriously autovivifies
Steve Peters [Thu, 23 Feb 2006 17:42:09 +0000 (17:42 +0000)]
Add test for RT #2166: foreach spuriously autovivifies

p4raw-id: //depot/perl@27287

16 years agoConverted t/op/pos.t to use and added TODO test for
Steve Peters [Thu, 23 Feb 2006 16:24:19 +0000 (16:24 +0000)]
Converted t/op/pos.t to use and added TODO test for
RT #1716 - search position reset after 'local' save/restore

p4raw-id: //depot/perl@27286

16 years agoImplement ${^WIN32_SLOPPY_STAT}
Jan Dubois [Tue, 21 Feb 2006 12:36:31 +0000 (04:36 -0800)]
Implement ${^WIN32_SLOPPY_STAT}
From: "Jan Dubois" <>
Message-ID: <019601c63726$7fcca200$6062a8c0@candy>

p4raw-id: //depot/perl@27283

16 years agoRecalculate dstr in Perl_sv_setsv_flags, as dstr may have been upgraded.
Nicholas Clark [Thu, 23 Feb 2006 13:42:29 +0000 (13:42 +0000)]
Recalculate dstr in Perl_sv_setsv_flags, as dstr may have been upgraded.

p4raw-id: //depot/perl@27282

16 years agoRemove un-needed case in Perl_sv_setsv_flags (sv_upgrade will cover
Nicholas Clark [Thu, 23 Feb 2006 13:07:54 +0000 (13:07 +0000)]
Remove un-needed case in Perl_sv_setsv_flags (sv_upgrade will cover
this one for us).

p4raw-id: //depot/perl@27279

16 years agoRemove get magic from typeglobs. This means that PVGVs holding
Nicholas Clark [Thu, 23 Feb 2006 11:11:12 +0000 (11:11 +0000)]
Remove get magic from typeglobs. This means that PVGVs holding
typeglobs never need to use SvPVX. This comes at price - typeglobs
were using magic get for their stringificiation, and to pass SvOK(),
so need to make typeglobs SvOK by default (by sucking SVp_SCREAM into
SVf_OK - it's the only flag left), tweak SvSCREAM() to also check
SVp_POK, and teach sv_2[inpu]v how to convert globs.
However, it should free up SvPVX for the next part of the plan to
pointer indirections, and therefore CPU cache pressure.

p4raw-id: //depot/perl@27278

16 years agoFix the copy sizes for PVFM and PVCV back to sanity.
Nicholas Clark [Thu, 23 Feb 2006 00:29:04 +0000 (00:29 +0000)]
Fix the copy sizes for PVFM and PVCV back to sanity.

p4raw-id: //depot/perl@27276

16 years agoFix typos and a missing bracket.
Nicholas Clark [Wed, 22 Feb 2006 23:43:10 +0000 (23:43 +0000)]
Fix typos and a missing bracket.

p4raw-id: //depot/perl@27274

16 years agoAvoid C<study>ing any strings that might change underneath us, such
Nicholas Clark [Wed, 22 Feb 2006 22:30:19 +0000 (22:30 +0000)]
Avoid C<study>ing any strings that might change underneath us, such
as tied scalars and scalars with overloaded stringification.

p4raw-id: //depot/perl@27273

16 years agoMark some new study tests as TODO
Rafael Garcia-Suarez [Wed, 22 Feb 2006 22:09:41 +0000 (22:09 +0000)]
Mark some new study tests as TODO

p4raw-id: //depot/perl@27272

16 years agoAdd a test for study() on tied scalars, by Andy Lester after
Rafael Garcia-Suarez [Wed, 22 Feb 2006 21:04:33 +0000 (21:04 +0000)]
Add a test for study() on tied scalars, by Andy Lester after
Rick Delaney

p4raw-id: //depot/perl@27271

16 years agoTurn on match string copying when /e flag is set on a substitution.
Jarkko Hietaniemi [Wed, 22 Feb 2006 20:16:00 +0000 (20:16 +0000)]
Turn on match string copying when /e flag is set on a substitution.

Subject: [PATCH] dodge a valgrind error (for maint or blead)
Message-ID: <>

p4raw-id: //depot/perl@27270

16 years agosv_dump should report the PV for PVGVs, as it can get set.
Nicholas Clark [Wed, 22 Feb 2006 18:21:30 +0000 (18:21 +0000)]
sv_dump should report the PV for PVGVs, as it can get set.
(Typeglobs stringify via a call to GET magic, which uses the PV slot.)

p4raw-id: //depot/perl@27269

16 years agoTest warnings for converting globs to other forms.
Nicholas Clark [Wed, 22 Feb 2006 16:50:25 +0000 (16:50 +0000)]
Test warnings for converting globs to other forms.

p4raw-id: //depot/perl@27268

16 years agoChange 24643 made the mistake of assuming that CvCONST can only be true
Nicholas Clark [Wed, 22 Feb 2006 00:23:49 +0000 (00:23 +0000)]
Change 24643 made the mistake of assuming that CvCONST can only be true
on XSUBs. Somehow it can also end up on perl subs. Bug spotted by and
test case from Marcus Holland-Moritz.

p4raw-id: //depot/perl@27267

16 years agoCan merge the two arms of Perl_magic_getglob to save space.
Nicholas Clark [Tue, 21 Feb 2006 20:09:25 +0000 (20:09 +0000)]
Can merge the two arms of Perl_magic_getglob to save space.

p4raw-id: //depot/perl@27265

16 years agoReorder the SV flags so that SVp_?OK and SVf_?OK occupy 8 contiguous
Nicholas Clark [Tue, 21 Feb 2006 19:04:39 +0000 (19:04 +0000)]
Reorder the SV flags so that SVp_?OK and SVf_?OK occupy 8 contiguous
bits (0x0000XX00). This makes perl 3K smaller with -Os on x86 FreeBSD,
and might also help on other architectures (eg with 8 contiguous bits,
the SvOK() test on ARM won't need an intermediate constant).

p4raw-id: //depot/perl@27264

16 years agoAdditional hints needed to build threaded Perls on NetBSD.
Steve Peters [Tue, 21 Feb 2006 18:44:21 +0000 (18:44 +0000)]
Additional hints needed to build threaded Perls on NetBSD.

p4raw-id: //depot/perl@27263

16 years agoPL_body_arenas should be set to NULL when the interpreter is cleared.
Nicholas Clark [Tue, 21 Feb 2006 18:00:57 +0000 (18:00 +0000)]
PL_body_arenas should be set to NULL when the interpreter is cleared.

p4raw-id: //depot/perl@27262

16 years agoTeach B about CVf_ISXSUB
Nicholas Clark [Tue, 21 Feb 2006 16:28:02 +0000 (16:28 +0000)]
Teach B about CVf_ISXSUB

p4raw-id: //depot/perl@27261

16 years agoGoodbye PERL_XSUB_OLDSTYLE.
Nicholas Clark [Tue, 21 Feb 2006 16:12:37 +0000 (16:12 +0000)]

p4raw-id: //depot/perl@27260

16 years agoRe-order the definitions of the private bits in SvFLAGS by value, so
Nicholas Clark [Tue, 21 Feb 2006 15:40:04 +0000 (15:40 +0000)]
Re-order the definitions of the private bits in SvFLAGS by value, so
that we can see where we're already double-booked.

p4raw-id: //depot/perl@27259

16 years agoDocument the many uses of SVf_FAKE.
Nicholas Clark [Tue, 21 Feb 2006 15:23:21 +0000 (15:23 +0000)]
Document the many uses of SVf_FAKE.

p4raw-id: //depot/perl@27258

16 years agoAdd tests for the previously untested Hash::Util::all_keys().
Steve Peters [Tue, 21 Feb 2006 15:18:58 +0000 (15:18 +0000)]
Add tests for the previously untested Hash::Util::all_keys().

p4raw-id: //depot/perl@27257

16 years ago$ExtUtils::MM_Unix::VERSION needs to stay numeric to avoid test
Steve Peters [Tue, 21 Feb 2006 00:43:43 +0000 (00:43 +0000)]
$ExtUtils::MM_Unix::VERSION needs to stay numeric to avoid test

p4raw-id: //depot/perl@27256

16 years agoPatches: B, CGI, ExtUtils::MM_Unix
Joshua ben Jore [Sun, 19 Feb 2006 02:58:10 +0000 (20:58 -0600)]
Patches: B, CGI, ExtUtils::MM_Unix
From: "Joshua ben Jore" <>
Message-ID: <>

p4raw-id: //depot/perl@27255

16 years agoRe: Patches: B, CGI, ExtUtils::MM_Unix
Joshua ben Jore [Mon, 20 Feb 2006 09:50:15 +0000 (03:50 -0600)]
Re: Patches: B, CGI, ExtUtils::MM_Unix
From: "Joshua ben Jore" <>
Message-ID: <>

Includes changes to increment test count in ext/B/t/concise-xs.t and
uses "no warnings 'once';" in ext/B/t/b.t rather than kludging to
avoid the warning.

p4raw-id: //depot/perl@27254

16 years agoop/magic failure on cygwin after 1.5.19-4
Yitzchak Scott-Thoennes [Mon, 20 Feb 2006 11:09:33 +0000 (03:09 -0800)]
op/magic failure on cygwin after 1.5.19-4
Message-ID: <>

p4raw-id: //depot/perl@27252

16 years agoSteal code from maint, and use PERL_ARENA_ROOTS_SIZE to size the arena
Nicholas Clark [Mon, 20 Feb 2006 20:32:09 +0000 (20:32 +0000)]
Steal code from maint, and use PERL_ARENA_ROOTS_SIZE to size the arena
arrays, rather than SVt_LAST, so that SVt_LAST can truthfully remain
the number of genuine SV types, unclouded by implementation details.

p4raw-id: //depot/perl@27251

16 years agoMake SDBM_File work with -Duse64bitall on Darwin (Mac OS X)
Dominic Dunlop [Mon, 20 Feb 2006 13:46:00 +0000 (13:46 +0000)]
Make SDBM_File work with -Duse64bitall on Darwin (Mac OS X)
Message-Id: <>

p4raw-id: //depot/perl@27250

16 years agoPVCVs don't need XNVs either.
Nicholas Clark [Mon, 20 Feb 2006 19:00:22 +0000 (19:00 +0000)]
PVCVs don't need XNVs either.
(And actually remove xcv_depth)
(And fix the copy lengths in bodies_by_type)

p4raw-id: //depot/perl@27249

16 years agoTrouble with $ENV{CDPATH} after change #27236
Dominic Dunlop [Mon, 20 Feb 2006 14:49:00 +0000 (14:49 +0000)]
Trouble with $ENV{CDPATH} after change #27236
Message-Id: <>

p4raw-id: //depot/perl@27248

16 years agoPVFMs don't need CvDEPTH, and PVCVs don't use SvIVX, so moving
Nicholas Clark [Mon, 20 Feb 2006 17:48:21 +0000 (17:48 +0000)]
PVFMs don't need CvDEPTH, and PVCVs don't use SvIVX, so moving
xcv_depth into the IV union saves 4(ish) bytes per CV and format.
"ish" because it was a long, but has been changed to I32 (along with
the corresponding field in struct block_sub) so as not to enlarge the
IV union on platforms where sizeof(long) > sizeof(IV), or struct
block_sub where sizeof(long) > sizeof(I32)

p4raw-id: //depot/perl@27247

16 years agoFix sv_dump to dump formats without the (non-existent) NVX.
Nicholas Clark [Mon, 20 Feb 2006 14:01:59 +0000 (14:01 +0000)]
Fix sv_dump to dump formats without the (non-existent) NVX.

p4raw-id: //depot/perl@27245

16 years agoxcv_root and xcv_xsub can also be merged into a union, providing a new
Nicholas Clark [Mon, 20 Feb 2006 13:42:47 +0000 (13:42 +0000)]
xcv_root and xcv_xsub can also be merged into a union, providing a new
flag is added to denote whether the PVCV is perl or XSUB.

p4raw-id: //depot/perl@27244

16 years agoxcv_start and xcv_xsubany can be merged into a union, as they are never
Nicholas Clark [Mon, 20 Feb 2006 11:54:38 +0000 (11:54 +0000)]
xcv_start and xcv_xsubany can be merged into a union, as they are never
both needed.

p4raw-id: //depot/perl@27243

16 years agoNeed to clear CvXSUBANY() too to turn a constant sub into a prototype.
Nicholas Clark [Mon, 20 Feb 2006 11:54:03 +0000 (11:54 +0000)]
Need to clear CvXSUBANY() too to turn a constant sub into a prototype.

p4raw-id: //depot/perl@27242

16 years agoAdd a new CvISXSUB() macro, for abstracting the test as to whether a
Nicholas Clark [Mon, 20 Feb 2006 10:40:59 +0000 (10:40 +0000)]
Add a new CvISXSUB() macro, for abstracting the test as to whether a
PVCV is perl or XS.

p4raw-id: //depot/perl@27241

16 years agoAbolish BROKEN_UNION_INIT in B::C, as it works around problems in
Nicholas Clark [Mon, 20 Feb 2006 10:10:11 +0000 (10:10 +0000)]
Abolish BROKEN_UNION_INIT in B::C, as it works around problems in
pre-ANSI C compilers, but pays in code duplication.

p4raw-id: //depot/perl@27240

16 years agopatch@27236 vms glob/readdir/chdir EFS/long filename support
John E. Malmberg [Mon, 20 Feb 2006 03:43:00 +0000 (03:43 +0000)]
patch@27236 vms glob/readdir/chdir EFS/long filename support
Message-ID: <>

p4raw-id: //depot/perl@27239

16 years agoMore NullXXX macro removal from Andy Lester
Rafael Garcia-Suarez [Mon, 20 Feb 2006 09:26:15 +0000 (09:26 +0000)]
More NullXXX macro removal from Andy Lester

p4raw-id: //depot/perl@27238

16 years agoYitzchak points out that the perldiag entry for "Integer overflow in
Rafael Garcia-Suarez [Mon, 20 Feb 2006 08:54:22 +0000 (08:54 +0000)]
Yitzchak points out that the perldiag entry for "Integer overflow in
division" is no longer useful.

p4raw-id: //depot/perl@27237

16 years agoutftaint.t won't be able to run tests with -T if you have
Nicholas Clark [Mon, 20 Feb 2006 00:36:34 +0000 (00:36 +0000)]
utftaint.t won't be able to run tests with -T if you have
Insecure directory in $ENV{PATH}
so skip them. (Probably this means you have . in your PATH)

p4raw-id: //depot/perl@27236

16 years agoRemove the last (U16) cast for CV depths, missed by change 17835.
Nicholas Clark [Mon, 20 Feb 2006 00:21:09 +0000 (00:21 +0000)]
Remove the last (U16) cast for CV depths, missed by change 17835.

p4raw-id: //depot/perl@27235

16 years agoRe-order CV flags to bring the 4 CVf_BUILTIN_ATTRS into adjacent bits,
Nicholas Clark [Mon, 20 Feb 2006 00:02:53 +0000 (00:02 +0000)]
Re-order CV flags to bring the 4 CVf_BUILTIN_ATTRS into adjacent bits,
and make other flag bits that are paired in the code adjacent.
Will produce tighter code on ARM; might help on other platforms too.

p4raw-id: //depot/perl@27234

16 years agoMinistry of Truth removes typo in description of Change 26166.
Nicholas Clark [Sun, 19 Feb 2006 23:59:52 +0000 (23:59 +0000)]
Ministry of Truth removes typo in description of Change 26166.
Change 26166 was correct from the start.
Nothing to see. Move along.
We thank Big Brother for raising the chocolate ration to 20g per week.

p4raw-id: //depot/perl@27233

16 years agoShave sizeof(NV) bytes from formats, by using the same offset
Nicholas Clark [Sun, 19 Feb 2006 22:35:24 +0000 (22:35 +0000)]
Shave sizeof(NV) bytes from formats, by using the same offset
manoeuvre as PVs, PVIVs, PVAVs and PVHVs.

p4raw-id: //depot/perl@27231

16 years agoAVs and HVs don't have IVXs or NVXs, so assert this too.
Nicholas Clark [Sun, 19 Feb 2006 20:30:24 +0000 (20:30 +0000)]
AVs and HVs don't have IVXs or NVXs, so assert this too.

p4raw-id: //depot/perl@27230

16 years agoUnder -DDEBUGGING, assert that SvIVX, SvUVX and SvNVX aren't being
Nicholas Clark [Sun, 19 Feb 2006 20:10:34 +0000 (20:10 +0000)]
Under -DDEBUGGING, assert that SvIVX, SvUVX and SvNVX aren't being
used on scalars that don't have the memory allocated. Correct
SvSTASH and SvMAGIC to only evaluate the sv argument once.

p4raw-id: //depot/perl@27229

16 years agoTo make arithmetic on tainted dualvars work properly requires that
Nicholas Clark [Sun, 19 Feb 2006 18:57:35 +0000 (18:57 +0000)]
To make arithmetic on tainted dualvars work properly requires that
sv_2nv uses SvIVX in preference to SvPVX, if SVp_IOK is true.

p4raw-id: //depot/perl@27228

16 years agoSvROK(sv) will never be true when SvIOKp() or SvNOKp() is true, so the
Nicholas Clark [Sun, 19 Feb 2006 12:14:15 +0000 (12:14 +0000)]
SvROK(sv) will never be true when SvIOKp() or SvNOKp() is true, so the
code inside the if() test in sv_2pv_flags is dead and can be removed.

p4raw-id: //depot/perl@27227

16 years agoAs Anno Siegel points out, mtime/ctime test should be skipped
Rafael Garcia-Suarez [Sat, 18 Feb 2006 20:13:51 +0000 (20:13 +0000)]
As Anno Siegel points out, mtime/ctime test should be skipped
on OS X on HFS+.

p4raw-id: //depot/perl@27223

16 years agosave_re_context() and even errsv_save = newSVsv(ERRSV); can turn
Nicholas Clark [Sat, 18 Feb 2006 19:01:48 +0000 (19:01 +0000)]
save_re_context() and even errsv_save = newSVsv(ERRSV); can turn
PL_tainted back on, so defer turning if off as late as possible.
This gets lib/locale.t working once more.

p4raw-id: //depot/perl@27222

16 years agoTurn of tainting locally inside swash_init(), as the internal
Nicholas Clark [Sat, 18 Feb 2006 18:23:43 +0000 (18:23 +0000)]
Turn of tainting locally inside swash_init(), as the internal
implementation of the regexp engine has nothing to do with user data,
so when and where it happens to call require is agnostic of the current
expression's taintedness.

p4raw-id: //depot/perl@27221

16 years agoAssume that if runperl is called under tainting, that the caller really
Nicholas Clark [Sat, 18 Feb 2006 18:22:13 +0000 (18:22 +0000)]
Assume that if runperl is called under tainting, that the caller really
really wanted to run perl, so brute force untaint everything.

p4raw-id: //depot/perl@27220

16 years agoConvert utftaint.t to (from and provide it with
Nicholas Clark [Sat, 18 Feb 2006 17:38:38 +0000 (17:38 +0000)]
Convert utftaint.t to (from and provide it with
tainted() from taint.t to remove the dependency on Scalar::Util.
(So it will now work with miniperl)

p4raw-id: //depot/perl@27219

16 years agoUpgrade to Digest-SHA-5.34.
Steve Peters [Sat, 18 Feb 2006 12:58:49 +0000 (12:58 +0000)]
Upgrade to Digest-SHA-5.34.

p4raw-id: //depot/perl@27218

16 years agoarena-rework : consolidated patch
Jim Cromie [Mon, 13 Feb 2006 14:12:41 +0000 (07:12 -0700)]
arena-rework : consolidated patch
Message-ID: <>

Tweaked somewhat to split the arena boolean from the arena_size,
and with the PTE still doubling-up with one of the SV types in the

p4raw-id: //depot/perl@27215

16 years agoRe: Smoke [5.9.4] 27195 FAIL(m) irix 6.2 (IP22/1 cpu)
Abe Timmerman [Fri, 17 Feb 2006 23:58:23 +0000 (00:58 +0100)]
Re: Smoke [5.9.4] 27195 FAIL(m) irix 6.2 (IP22/1 cpu)
Message-Id: <>
Date: Fri, 17 Feb 2006 23:58:23 +0100

p4raw-id: //depot/perl@27214

16 years agoThe two whitespace changes from
Jim Cromie [Mon, 13 Feb 2006 14:12:41 +0000 (07:12 -0700)]
The two whitespace changes from

Subject: arena-rework : consolidated patch
Message-ID: <>
Date: Mon, 13 Feb 2006 14:12:41 -0700

p4raw-id: //depot/perl@27213

16 years agoDon't add -fPIC if already present.
Brendan O'Dea [Sat, 11 Feb 2006 00:37:15 +0000 (11:37 +1100)]
Don't add -fPIC if already present.

Subject: [PATCH] 5.8.8: Debian patches
From: "Brendan O'Dea" <>
Message-ID: <>

p4raw-id: //depot/perl@27212

16 years agoFix precedence.
Brendan O'Dea [Sat, 11 Feb 2006 00:37:15 +0000 (11:37 +1100)]
Fix precedence.

Subject: [PATCH] 5.8.8: Debian patches
From: "Brendan O'Dea" <>
Message-ID: <>

p4raw-id: //depot/perl@27211

16 years agoFix quote typo.
Brendan O'Dea [Sat, 11 Feb 2006 00:37:15 +0000 (11:37 +1100)]
Fix quote typo.

Subject: [PATCH] 5.8.8: Debian patches
From: "Brendan O'Dea" <>
Message-ID: <>

p4raw-id: //depot/perl@27210

16 years agoIn TODO: make lexical ?
Rafael Garcia-Suarez [Fri, 17 Feb 2006 15:48:59 +0000 (15:48 +0000)]
In TODO: make lexical ?

p4raw-id: //depot/perl@27209

16 years agocleaning up perly.c
Andy Lester [Sun, 12 Feb 2006 23:04:56 +0000 (17:04 -0600)]
cleaning up perly.c
Message-ID: <>

p4raw-id: //depot/perl@27208

16 years agoRe: Perl_sighandler voodoo (was Re: Smoke [5.8.7] 26127 FAIL(XF) MSWin32 WinXP/.Net...
Robin Houston [Sun, 27 Nov 2005 00:43:13 +0000 (00:43 +0000)]
Re: Perl_sighandler voodoo (was Re: Smoke [5.8.7] 26127 FAIL(XF) MSWin32 WinXP/.Net SP2 (x86/2 cpu))
Message-ID: <>

Simplification of the code in my_exit_jump() that unwinds context

p4raw-id: //depot/perl@27207

16 years agoRe: [perl #38485] use integer; 0x80000000/-1; # coredump
Yitzchak Scott-Thoennes [Mon, 13 Feb 2006 01:39:25 +0000 (17:39 -0800)]
Re: [perl #38485] use integer; 0x80000000/-1; # coredump
Message-ID: <>

p4raw-id: //depot/perl@27205

16 years agoRe: how to build with -DPERL_MEM_LOG ?
Andy Dougherty [Fri, 3 Feb 2006 11:55:58 +0000 (06:55 -0500)]
Re: how to build with -DPERL_MEM_LOG ?
Message-ID: <>

p4raw-id: //depot/perl@27204

16 years agoProblem compiling swigged c++ code with 5.8.8
Merijn Broeren [Wed, 15 Feb 2006 13:51:49 +0000 (14:51 +0100)]
Problem compiling swigged c++ code with 5.8.8
Message-ID: <>

p4raw-id: //depot/perl@27203

16 years agoUpgrade to CGI-3.16, with version bump on for documentation
Steve Peters [Thu, 16 Feb 2006 12:54:08 +0000 (12:54 +0000)]
Upgrade to CGI-3.16, with version bump on for documentation
fixes not yet integrated.

p4raw-id: //depot/perl@27202

16 years agoRemove lib/Hash directory when doing make distclean
Rafael Garcia-Suarez [Thu, 16 Feb 2006 12:47:13 +0000 (12:47 +0000)]
Remove lib/Hash directory when doing make distclean

p4raw-id: //depot/perl@27201