This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
perl5.git
13 years agoIn struct block change blku_type from U8 to U16, and the "spare" U8 to
Nicholas Clark [Mon, 21 Jan 2008 14:45:56 +0000 (14:45 +0000)]
In struct block change blku_type from U8 to U16, and the "spare" U8 to
U16, with the lockstep changes in struct subst. Eliminate lval from
struct block_sub, and instead store it in the U16 in struct block.

p4raw-id: //depot/perl@33028

13 years agoSqueeze the context type down to 4 bits, and move the private flags to
Nicholas Clark [Mon, 21 Jan 2008 14:09:06 +0000 (14:09 +0000)]
Squeeze the context type down to 4 bits, and move the private flags to
fit within the next 4 bits.

p4raw-id: //depot/perl@33027

13 years agoWhen changing the op_ppaddr of an op, one must keep its op_type
Rafael Garcia-Suarez [Mon, 21 Jan 2008 13:08:11 +0000 (13:08 +0000)]
When changing the op_ppaddr of an op, one must keep its op_type
in sync. That helps writers of alternate runloops.

p4raw-id: //depot/perl@33026

13 years agoGive G_VOID, G_SCALAR and G_ARRAY the same numeric values as
Nicholas Clark [Mon, 21 Jan 2008 11:50:52 +0000 (11:50 +0000)]
Give G_VOID, G_SCALAR and G_ARRAY the same numeric values as
OPf_WANT_VOID, OPf_WANT_SCALAR and OPf_WANT_LIST.

p4raw-id: //depot/perl@33025

13 years agoMore places that could be using G_WANT, not picked up by change 33021.
Nicholas Clark [Mon, 21 Jan 2008 11:49:38 +0000 (11:49 +0000)]
More places that could be using G_WANT, not picked up by change 33021.

p4raw-id: //depot/perl@33024

13 years agoRemove hardcoded cop.h constants from APItest.pm :-(
Nicholas Clark [Mon, 21 Jan 2008 11:48:22 +0000 (11:48 +0000)]
Remove hardcoded cop.h constants from APItest.pm :-(
Add G_WANT. Make call.t use G_WANT.

p4raw-id: //depot/perl@33023

13 years agoChange the wantarray result from caller from IV to bool for the SCALAR/
Nicholas Clark [Mon, 21 Jan 2008 09:04:54 +0000 (09:04 +0000)]
Change the wantarray result from caller from IV to bool for the SCALAR/
ARRAY case. This doesn't contradict the documentation, as there isn't
any. Oops.

p4raw-id: //depot/perl@33022

13 years agoAdd a flag G_WANT, as a mask for the bits G_SCALAR, G_ARRAY and G_VOID.
Nicholas Clark [Mon, 21 Jan 2008 08:54:25 +0000 (08:54 +0000)]
Add a flag G_WANT, as a mask for the bits G_SCALAR, G_ARRAY and G_VOID.

p4raw-id: //depot/perl@33021

13 years agoEncode the G_* to OPf_WANT_* transform, the reverse of OP_GIMME(), in
Nicholas Clark [Mon, 21 Jan 2008 08:23:46 +0000 (08:23 +0000)]
Encode the G_* to OPf_WANT_* transform, the reverse of OP_GIMME(), in
a macro OP_GIMME_REVERSE() [so that it could be changed].

p4raw-id: //depot/perl@33020

13 years agoChange the hasargs return value from caller from IV to bool, as it's
Nicholas Clark [Mon, 21 Jan 2008 07:11:41 +0000 (07:11 +0000)]
Change the hasargs return value from caller from IV to bool, as it's
actually only a boolean, and sv_yes/sv_no are cheaper to "create".
The documentation says 'C<$hasargs> is true if ...' so this is fine.

p4raw-id: //depot/perl@33019

13 years agoEliminate hasargs from structs block_sub and block_format by storing
Nicholas Clark [Sun, 20 Jan 2008 22:20:56 +0000 (22:20 +0000)]
Eliminate hasargs from structs block_sub and block_format by storing
it with a private flag CXp_HASARGS in cx_type. (It's only a boolean.)

p4raw-id: //depot/perl@33018

13 years agoIn struct block_sub and block_format, access the members hasargs and
Nicholas Clark [Sun, 20 Jan 2008 21:50:31 +0000 (21:50 +0000)]
In struct block_sub and block_format, access the members hasargs and
lval via macros CxHASARGS() and CxLVAL(), which will allow the storage
location to be changed.

p4raw-id: //depot/perl@33017

13 years agoMove retop first in structs block_sub, block_format and block_eval.
Nicholas Clark [Sun, 20 Jan 2008 21:05:26 +0000 (21:05 +0000)]
Move retop first in structs block_sub, block_format and block_eval.
This simplifies some code in Perl_deb_stack_all().

p4raw-id: //depot/perl@33016

13 years agoEliminate label from struct block_loop, as it can be accessed via
Nicholas Clark [Sun, 20 Jan 2008 20:40:38 +0000 (20:40 +0000)]
Eliminate label from struct block_loop, as it can be accessed via
oldcop.

p4raw-id: //depot/perl@33015

13 years agoSplit struct block_sub into struct block_sub and struct block_format.
Nicholas Clark [Sun, 20 Jan 2008 19:56:21 +0000 (19:56 +0000)]
Split struct block_sub into struct block_sub and struct block_format.
(CXt_SUB and CXt_FORMAT were using some comon members, but some members
were only for one or the other.)

p4raw-id: //depot/perl@33014

13 years agoIn Perl_cx_dup(), block copy the entire context stack, then walk it
Nicholas Clark [Sun, 20 Jan 2008 18:54:17 +0000 (18:54 +0000)]
In Perl_cx_dup(), block copy the entire context stack, then walk it
to duplicate/fixup only the things that aren't simple binary copies.

p4raw-id: //depot/perl@33013

13 years agoExtend PUSHFORMAT() to take a second parameter to set retop, to save
Nicholas Clark [Sun, 20 Jan 2008 18:03:27 +0000 (18:03 +0000)]
Extend PUSHFORMAT() to take a second parameter to set retop, to save
NULLing it and then reassigning.

p4raw-id: //depot/perl@33012

13 years agoFix (probable) bug-by-inspection - CxREALEVAL(), CxTRYBLOCK(),
Nicholas Clark [Sun, 20 Jan 2008 17:39:22 +0000 (17:39 +0000)]
Fix (probable) bug-by-inspection -  CxREALEVAL(), CxTRYBLOCK(),
CxFOREACH() and CxFOREACHDEF() should all be masking against
CXTYPEMASK rather than their own type number.

p4raw-id: //depot/perl@33011

13 years agoIn struct block_loop access element label via the macro CxLABEL()
Nicholas Clark [Sun, 20 Jan 2008 13:28:10 +0000 (13:28 +0000)]
In struct block_loop access element label via the macro CxLABEL()
(for the places that aren't about to change)

p4raw-id: //depot/perl@33010

13 years ago[perl #49522] state variable not available
Dave Mitchell [Sat, 19 Jan 2008 20:41:03 +0000 (20:41 +0000)]
[perl #49522] state variable not available
Svf_PADSTALE means something different for state vars. Make sure
we always handle it correctly

p4raw-id: //depot/perl@33009

13 years agoAvoid an unused argument in S_sv_2iuv_non_preserve() by using
Nicholas Clark [Fri, 18 Jan 2008 18:00:56 +0000 (18:00 +0000)]
Avoid an unused argument in S_sv_2iuv_non_preserve() by using
conditional compilation to only pass it in if it's needed.

p4raw-id: //depot/perl@33004

13 years agoNick's a muppet - "all" needs to remain the first real target in the
Nicholas Clark [Fri, 18 Jan 2008 16:40:48 +0000 (16:40 +0000)]
Nick's a muppet - "all" needs to remain the first real target in the
Makefile. Do not be distracted by suffix rules above.

p4raw-id: //depot/perl@33003

13 years agoDon't create local variables validarg and fdscript if they aren't going
Nicholas Clark [Fri, 18 Jan 2008 16:11:35 +0000 (16:11 +0000)]
Don't create local variables validarg and fdscript if they aren't going
to be used.

p4raw-id: //depot/perl@33002

13 years agoThe suidscript argument to S_forbid_setid(), S_open_script() and
Nicholas Clark [Fri, 18 Jan 2008 15:28:09 +0000 (15:28 +0000)]
The suidscript argument to S_forbid_setid(), S_open_script() and
S_validate_suid() is actually just a boolean. So make it so.

p4raw-id: //depot/perl@33001

13 years agoEliminate all PERL_UNUSED_ARG()s in S_validate_suid() by changing its
Nicholas Clark [Fri, 18 Jan 2008 14:27:36 +0000 (14:27 +0000)]
Eliminate all PERL_UNUSED_ARG()s in S_validate_suid() by changing its
prototype depending on the compile time options. In turn, this finds
things that are unused in its callers.

p4raw-id: //depot/perl@33000

13 years agoA target for sperl.i
Nicholas Clark [Fri, 18 Jan 2008 14:13:30 +0000 (14:13 +0000)]
A target for sperl.i

p4raw-id: //depot/perl@32999

13 years agoChange 32997 missed one conditionally unused argument.
Nicholas Clark [Fri, 18 Jan 2008 11:01:15 +0000 (11:01 +0000)]
Change 32997 missed one conditionally unused argument.

p4raw-id: //depot/perl@32998

13 years agoCollect all the S_validate_suid() unused argument logic in one place.
Nicholas Clark [Fri, 18 Jan 2008 09:40:02 +0000 (09:40 +0000)]
Collect all the S_validate_suid() unused argument logic in one place.

p4raw-id: //depot/perl@32997

13 years agoRemove commented out declaration from S_validate_suid()
Nicholas Clark [Fri, 18 Jan 2008 09:27:13 +0000 (09:27 +0000)]
Remove commented out declaration from S_validate_suid()

p4raw-id: //depot/perl@32996

13 years agoNew tests for new ops. Most of them are TODO
Rafael Garcia-Suarez [Fri, 18 Jan 2008 09:18:58 +0000 (09:18 +0000)]
New tests for new ops. Most of them are TODO

p4raw-id: //depot/perl@32995

13 years agoRemove commented out code from S_validate_suid()
Nicholas Clark [Fri, 18 Jan 2008 09:09:56 +0000 (09:09 +0000)]
Remove commented out code from S_validate_suid()

p4raw-id: //depot/perl@32994

13 years agoAdd Weed out needless PERL_UNUSED_ARG to perltodo. It's a good
Nicholas Clark [Fri, 18 Jan 2008 08:20:13 +0000 (08:20 +0000)]
Add Weed out needless PERL_UNUSED_ARG to perltodo. It's a good
"cage cleaner" task.

p4raw-id: //depot/perl@32993

13 years agowarn if ++ or -- are unable to change the value because it's beyond
Nicholas Clark [Thu, 17 Jan 2008 14:23:48 +0000 (14:23 +0000)]
warn if ++ or -- are unable to change the value because it's beyond
the limit of representation in NVs, using a new warnings category
"imprecision".

p4raw-id: //depot/perl@32990

13 years agoUpgrade to Math-Complex-1.47
Steve Peters [Thu, 17 Jan 2008 12:24:21 +0000 (12:24 +0000)]
Upgrade to Math-Complex-1.47

p4raw-id: //depot/perl@32989

13 years agoGenerate the warnings masks programatically.
Nicholas Clark [Thu, 17 Jan 2008 11:31:38 +0000 (11:31 +0000)]
Generate the warnings masks programatically.
Get the correct line number when reporting errors from &check_bits.

p4raw-id: //depot/perl@32988

13 years agoAdd a diagram for the 5.11 SV class structure, including B::REGEXP.
Nicholas Clark [Thu, 17 Jan 2008 08:17:11 +0000 (08:17 +0000)]
Add a diagram for the 5.11 SV class structure, including B::REGEXP.

p4raw-id: //depot/perl@32987

13 years agoDon't ask about a module when using -ok
Abigail [Wed, 16 Jan 2008 18:46:17 +0000 (19:46 +0100)]
Don't ask about a module when using -ok
Message-ID: <20080116174617.GA30446@abigail.be>
Date: Wed, 16 Jan 2008 18:46:17 +0100

p4raw-id: //depot/perl@32986

13 years agoA couple of POD fixes by Steven Schubiger
Rafael Garcia-Suarez [Wed, 16 Jan 2008 12:12:10 +0000 (12:12 +0000)]
A couple of POD fixes by Steven Schubiger

p4raw-id: //depot/perl@32982

13 years agoThanks to Sébstien Aperghis-Tramoni's suggestion, updated version
Steve Peters [Tue, 15 Jan 2008 19:20:53 +0000 (19:20 +0000)]
Thanks to Sébstien Aperghis-Tramoni's suggestion,  updated version
numbers in a few extra files and regenerated uconfig.h.

p4raw-id: //depot/perl@32981

13 years agoBoolean priority bug, found by Mashrab Kuvatov:
Mashrab Kuvatov [Tue, 15 Jan 2008 15:17:42 +0000 (16:17 +0100)]
Boolean priority bug, found by Mashrab Kuvatov:

Subject: Re: [perl #49646] perlbug AutoReply: open ':locale' does not work under locale with the modifier
Message-Id: <200801151517.46296.kmashrab@uni-bremen.de>

p4raw-id: //depot/perl@32980

13 years agoFix range operator
Jerry D. Hedden [Mon, 14 Jan 2008 19:56:48 +0000 (14:56 -0500)]
Fix range operator
From: "Jerry D. Hedden" <jdhedden@cpan.org>
Message-ID: <1ff86f510801141656i325ac69ev8a0af47f9fe72a1e@mail.gmail.com>

p4raw-id: //depot/perl@32979

13 years agoconsting IO.xs
Robin Barker [Sat, 22 Dec 2007 00:56:18 +0000 (00:56 +0000)]
consting IO.xs
From: "Robin Barker" <Robin.Barker@npl.co.uk>
Message-ID: <46A0F33545E63740BC7563DE59CA9C6D0939A1@exchsvr2.npl.ad.local>

p4raw-id: //depot/perl@32978

13 years agoWhen parsing LC_ALL or LANG to get the locale's encoding, ignore
Rafael Garcia-Suarez [Mon, 14 Jan 2008 22:48:46 +0000 (22:48 +0000)]
When parsing LC_ALL or LANG to get the locale's encoding, ignore
whatever is after the @, since that's a modifier, not an encoding.

p4raw-id: //depot/perl@32977

13 years agoconsting
Robin Barker [Mon, 14 Jan 2008 20:39:35 +0000 (20:39 +0000)]
consting
From: "Robin Barker" <Robin.Barker@npl.co.uk>
Message-ID: <46A0F33545E63740BC7563DE59CA9C6D0939CA@exchsvr2.npl.ad.local>

p4raw-id: //depot/perl@32976

13 years agoIntermediate update/regen in cleanup process, dtrace support
H.Merijn Brand [Mon, 14 Jan 2008 16:37:08 +0000 (16:37 +0000)]
Intermediate update/regen in cleanup process, dtrace support
is now backported

p4raw-id: //depot/perl@32974

13 years agoRe-order so that the !SvOK() case is last (which should be rare)
Nicholas Clark [Sun, 13 Jan 2008 20:58:56 +0000 (20:58 +0000)]
Re-order so that the !SvOK() case is last (which should be rare)
Remove the FIXME comment as I had already fixed it.

p4raw-id: //depot/perl@32971

13 years agoUpgrade to Math-Complex-1.44
Steve Peters [Sun, 13 Jan 2008 15:28:30 +0000 (15:28 +0000)]
Upgrade to Math-Complex-1.44

p4raw-id: //depot/perl@32970

13 years agoFor 5.12: saner behaviour for `length`
Nicholas Clark [Sat, 12 Jan 2008 22:20:39 +0000 (22:20 +0000)]
For 5.12: saner behaviour for `length`
(Make C<length undef> return undef).
Patch mostly by Rafael, with some fine tuning by me.

p4raw-id: //depot/perl@32969

13 years agoFix bug whereby length on a tied scalar that returned a UTF-8 value
Nicholas Clark [Sat, 12 Jan 2008 21:57:06 +0000 (21:57 +0000)]
Fix bug whereby length on a tied scalar that returned a UTF-8 value
would not be correct the first time. (And for the more pathological
case, would be incorrect if the UTF-8-ness of the returned value
changed.)

p4raw-id: //depot/perl@32968

13 years agoRe: [PATCH] Stop DTrace config option looping in non-interactive mode
Andy Armstrong [Sat, 12 Jan 2008 20:27:46 +0000 (20:27 +0000)]
Re: [PATCH] Stop DTrace config option looping in non-interactive mode
Message-Id: <4C6BD298-5CD3-49EA-8FED-4A6DA24BFA7B@hexten.net>
Date: Sat, 12 Jan 2008 20:27:46 +0000

p4raw-id: //depot/perl@32967

13 years agotru64: poison stack, write protect string constants
Jarkko Hietaniemi [Sat, 12 Jan 2008 03:27:10 +0000 (05:27 +0200)]
tru64: poison stack, write protect string constants
Message-Id: <200801120127.m0C1RAZ6088857@kosh.hut.fi>

p4raw-id: //depot/perl@32966

13 years agoNit disabling dtrace on VMS (2 single quotes inside double quotes
Craig A. Berry [Sat, 12 Jan 2008 01:07:54 +0000 (01:07 +0000)]
Nit disabling dtrace on VMS (2 single quotes inside double quotes
causes symbol interpolation in DCL).

p4raw-id: //depot/perl@32965

13 years agoDon't quote a whitespace-containing ABSPERL when the whitespace
Craig A. Berry [Sat, 12 Jan 2008 01:01:45 +0000 (01:01 +0000)]
Don't quote a whitespace-containing ABSPERL when the whitespace
means it is a command with parameter(s).

p4raw-id: //depot/perl@32964

13 years agosh doesn't know about !
Andy Armstrong [Fri, 11 Jan 2008 19:34:25 +0000 (19:34 +0000)]
sh doesn't know about !
Message-Id: <7EC45604-54BE-4738-90B2-7AC5C6D17026@hexten.net>
Date: Fri, 11 Jan 2008 19:34:25 +0000

p4raw-id: //depot/perl@32963

13 years agoWell, I know *something* passed make test from a clean build before
Nicholas Clark [Fri, 11 Jan 2008 20:27:23 +0000 (20:27 +0000)]
Well, I know *something* passed make test from a clean build before
change 32961, and I thought that it was the right thing, but I guess
not. It should have read like this.

p4raw-id: //depot/perl@32962

13 years agoassert that these are the regexps you were looking for.
Nicholas Clark [Fri, 11 Jan 2008 19:01:39 +0000 (19:01 +0000)]
assert that these are the regexps you were looking for.
(at least for the most commonly used macros).
Remove the duplicate definition of RX_SUBBEG(), which I was sure I'd
done earlier.

p4raw-id: //depot/perl@32961

13 years ago__DATE__ and __TIME__ are both string literals, so we can concatentate
Nicholas Clark [Fri, 11 Jan 2008 18:03:18 +0000 (18:03 +0000)]
__DATE__ and __TIME__ are both string literals, so we can concatentate
them, rather than formating them via %s.

p4raw-id: //depot/perl@32960

13 years agoPack the recycled pad offsets into an SV at PL_regex_pad[0]. This will
Nicholas Clark [Fri, 11 Jan 2008 17:00:59 +0000 (17:00 +0000)]
Pack the recycled pad offsets into an SV at PL_regex_pad[0]. This will
use less memory than an AV.

p4raw-id: //depot/perl@32959

13 years agoRemove C variables (and a parameter!) now unused since -P bit the dust.
Nicholas Clark [Fri, 11 Jan 2008 16:42:11 +0000 (16:42 +0000)]
Remove C variables (and a parameter!) now unused since -P bit the dust.

p4raw-id: //depot/perl@32958

13 years agoNote the U8 sized space created by removing -P, and check that it is
Nicholas Clark [Fri, 11 Jan 2008 15:04:51 +0000 (15:04 +0000)]
Note the U8 sized space created by removing -P, and check that it is
now an illegal command line flag.

p4raw-id: //depot/perl@32956

13 years agoRename PERL_MAX_SUB_DEPTH to PERL_SUB_DEPTH_WARN, per Tim Bunce's
Rafael Garcia-Suarez [Fri, 11 Jan 2008 13:59:32 +0000 (13:59 +0000)]
Rename PERL_MAX_SUB_DEPTH to PERL_SUB_DEPTH_WARN, per Tim Bunce's
suggestion

p4raw-id: //depot/perl@32955

13 years agoRemove the -P switch
Rafael Garcia-Suarez [Fri, 11 Jan 2008 13:55:07 +0000 (13:55 +0000)]
Remove the -P switch

p4raw-id: //depot/perl@32954

13 years agoAdd dtrace support
Andy Armstrong [Thu, 10 Jan 2008 22:20:52 +0000 (22:20 +0000)]
Add dtrace support
Message-Id: <F4AC553F-7C7F-49C3-98C2-E04681E1004F@hexten.net>

with fixups as discussed on list, plus adding usedtrace to Glossary,
plus propagating all the new config variables everywhere.
(Was there an automatic way to do that? I did it with emacs macros)

p4raw-id: //depot/perl@32953

13 years agoRemove PM_GETRE_SAFE and PM_SETRE_SAFE as nothing uses them.
Nicholas Clark [Fri, 11 Jan 2008 13:22:35 +0000 (13:22 +0000)]
Remove PM_GETRE_SAFE and PM_SETRE_SAFE as nothing uses them.

p4raw-id: //depot/perl@32952

13 years agoRemove code that protected pp_qr against REGEXPs going away during
Nicholas Clark [Fri, 11 Jan 2008 12:47:57 +0000 (12:47 +0000)]
Remove code that protected pp_qr against REGEXPs going away during
global destruction whilst they were stored via true references in
PL_regex_padav. It won't happen again (unless someone manages to get
a reference to an OP's regexp, which is also a risk in 5.10 and 5.8)

p4raw-id: //depot/perl@32951

13 years agoREGEXPs are now stored directly in PL_regex_padav, rather than
Nicholas Clark [Fri, 11 Jan 2008 12:14:04 +0000 (12:14 +0000)]
REGEXPs are now stored directly in PL_regex_padav, rather than
indirectly via RVs. This saves memory, and removes 1 level of pointer
indirection.

p4raw-id: //depot/perl@32950

13 years agoIntroduce macro PERL_MAX_SUB_DEPTH
Rafael Garcia-Suarez [Fri, 11 Jan 2008 11:31:34 +0000 (11:31 +0000)]
Introduce macro PERL_MAX_SUB_DEPTH

p4raw-id: //depot/perl@32949

13 years agoRe: [perl #49564] Re: MRO and av_clear
Rick Delaney [Wed, 9 Jan 2008 13:36:55 +0000 (08:36 -0500)]
Re: [perl #49564] Re: MRO and av_clear
Message-ID: <20080109183655.GB11282@bort.ca>

p4raw-id: //depot/perl@32948

13 years agoFrom: "Robin Barker" <Robin.Barker@npl.co.uk>
Robin Barker [Thu, 10 Jan 2008 18:50:12 +0000 (18:50 +0000)]
From: "Robin Barker" <Robin.Barker@npl.co.uk>
Message-ID: <46A0F33545E63740BC7563DE59CA9C6D0939C3@exchsvr2.npl.ad.local>

More consting.

p4raw-id: //depot/perl@32947

13 years agoFrom: "Robin Barker" <Robin.Barker@npl.co.uk>
Robin Barker [Thu, 10 Jan 2008 18:48:04 +0000 (18:48 +0000)]
From: "Robin Barker" <Robin.Barker@npl.co.uk>
Message-ID: <46A0F33545E63740BC7563DE59CA9C6D0939C2@exchsvr2.npl.ad.local>

p4raw-id: //depot/perl@32946

13 years ago[DOC PATCH] was Re: things deprecated for removal in 5.12
Paul Johnson [Thu, 10 Jan 2008 23:44:20 +0000 (00:44 +0100)]
[DOC PATCH] was Re: things deprecated for removal in 5.12
Message-ID: <20080110224420.GB29292@pjcj.net>

p4raw-id: //depot/perl@32945

13 years agoBetter descriptions for PL_regex_pad and PL_regex_padav.
Nicholas Clark [Thu, 10 Jan 2008 23:55:40 +0000 (23:55 +0000)]
Better descriptions for PL_regex_pad and PL_regex_padav.

p4raw-id: //depot/perl@32944

13 years agoReal regexps are stored via RVs, IVs can only mean spare slots. Hence
Nicholas Clark [Thu, 10 Jan 2008 23:54:22 +0000 (23:54 +0000)]
Real regexps are stored via RVs, IVs can only mean spare slots. Hence
it's safe to make the scalar undefined, rather than calling out to
write a 0 that would also be interpreted as a NULL pointer.

p4raw-id: //depot/perl@32943

13 years agoChange 32899 missed undoing the reference count increase when the SV
Nicholas Clark [Thu, 10 Jan 2008 23:32:27 +0000 (23:32 +0000)]
Change 32899 missed undoing the reference count increase when the SV
is popped off the AV.
"There's Something Wrong with our Bloody Leak Checking Today", as
Beattie didn't put it. It seems that we really can't check for leaking
scalars in perl_destruct, because we do our damndest to free them
brute force, rather than by undefining the symbol table and seeing
what sticks around.

p4raw-id: //depot/perl@32942

13 years agoReverse the TODOs on t/op/regexp_qr_embed_thr.t - they all pass now.
Nicholas Clark [Thu, 10 Jan 2008 22:38:10 +0000 (22:38 +0000)]
Reverse the TODOs on t/op/regexp_qr_embed_thr.t - they all pass now.

p4raw-id: //depot/perl@32941

13 years agoUn-TODO (and un-SKIP) tests in t/op/pat_thr.t that now pass.
Nicholas Clark [Thu, 10 Jan 2008 22:37:20 +0000 (22:37 +0000)]
Un-TODO (and un-SKIP) tests in t/op/pat_thr.t that now pass.
Some TODOs still remain. Bounties?

p4raw-id: //depot/perl@32940

13 years agoIt seems that you don't need to reference count PL_reg_curpm without
Nicholas Clark [Thu, 10 Jan 2008 22:06:53 +0000 (22:06 +0000)]
It seems that you don't need to reference count PL_reg_curpm without
ithreads, so don't waste time doing it there.

p4raw-id: //depot/perl@32939

13 years agoThe correct solution is to reference count the regexp in PL_reg_curpm,
Nicholas Clark [Thu, 10 Jan 2008 21:16:01 +0000 (21:16 +0000)]
The correct solution is to reference count the regexp in PL_reg_curpm,
rather than put in lots of hacks to work round not reference counting
it.

p4raw-id: //depot/perl@32938

13 years agoEnsure DEBUG_LEAKING_SCALARS_ABORT can't be circumvented by fatal
Nicholas Clark [Thu, 10 Jan 2008 21:15:02 +0000 (21:15 +0000)]
Ensure DEBUG_LEAKING_SCALARS_ABORT can't be circumvented by fatal
warnings. Add an abort() if you try to dup a freed scalar.

p4raw-id: //depot/perl@32937

13 years agoFix bug picked up by printf format warnings - a cast is needed where
Nicholas Clark [Thu, 10 Jan 2008 17:27:39 +0000 (17:27 +0000)]
Fix bug picked up by printf format warnings - a cast is needed where
UV is larger than U32 for dumping leaking scalars.

p4raw-id: //depot/perl@32935

13 years agoMove the reg_stringify logic to Perl_sv_2pv_flags
Ævar Arnfjörð Bjarmason [Wed, 9 Jan 2008 21:05:15 +0000 (21:05 +0000)]
Move the reg_stringify logic to Perl_sv_2pv_flags
Message-ID: <86zlveaewk.fsf@cpan.org>

with two corrections.
Plus remove reg_stringify from embed.fnc and regen.

p4raw-id: //depot/perl@32934

13 years agoseen_eval regex field wasn't getting cloned
Dave Mitchell [Wed, 9 Jan 2008 23:57:36 +0000 (23:57 +0000)]
seen_eval regex field wasn't getting cloned

p4raw-id: //depot/perl@32933

13 years agoFix the bug introduced by the bug fix of change 30755.
Nicholas Clark [Wed, 9 Jan 2008 23:26:14 +0000 (23:26 +0000)]
Fix the bug introduced by the bug fix of change 30755.
(Certain regexps could SEGV if cloned).

p4raw-id: //depot/perl@32932

13 years agoVariants of several regression tests that run the actul tests inside
Nicholas Clark [Wed, 9 Jan 2008 22:42:04 +0000 (22:42 +0000)]
Variants of several regression tests that run the actul tests inside
a new thread, to test ithread's cloning, particularly of regexps.

p4raw-id: //depot/perl@32931

13 years agogrep the MANIFEST to avoid having hard coded numbers in readdir.t
Nicholas Clark [Wed, 9 Jan 2008 22:36:21 +0000 (22:36 +0000)]
grep the MANIFEST to avoid having hard coded numbers in readdir.t

p4raw-id: //depot/perl@32930

13 years agoUpgrade to Math-Complex-1.43
Steve Peters [Wed, 9 Jan 2008 21:50:25 +0000 (21:50 +0000)]
Upgrade to Math-Complex-1.43

p4raw-id: //depot/perl@32929

13 years agoOnly code paths in hv_store either XSRETURN_NO or XSRETURN_YES.
Steve Peters [Wed, 9 Jan 2008 21:18:39 +0000 (21:18 +0000)]
Only code paths in hv_store either XSRETURN_NO or XSRETURN_YES.
Change the prototype to return void, remove the output to RETVAL
(which isn't set anyway), and fix a typo in the code.

p4raw-id: //depot/perl@32928

13 years agoIntermediate re-gen for AndyA's dtrace work. More to come
H.Merijn Brand [Wed, 9 Jan 2008 21:05:38 +0000 (21:05 +0000)]
Intermediate re-gen for AndyA's dtrace work. More to come

p4raw-id: //depot/perl@32927

13 years agoRe: [PATCH] Exclude .svn and _darcs from IPC::SysV's Makefile.PL libscan
Marcus Holland-Moritz [Wed, 9 Jan 2008 19:58:11 +0000 (20:58 +0100)]
Re: [PATCH] Exclude .svn and _darcs from IPC::SysV's Makefile.PL libscan
Message-ID: <20080109195811.2c5353e5@r2d2>

Make libscan() in IPC::SysV's Makefile.PL inherit EU::MM's
libscan() behaviour.

p4raw-id: //depot/perl@32926

13 years agoFix prototype in regexp code following #32851, and regen
Steve Hay [Wed, 9 Jan 2008 17:59:11 +0000 (17:59 +0000)]
Fix prototype in regexp code following #32851, and regen

p4raw-id: //depot/perl@32925

13 years agoRevert change #32920, since it was solved another way by #32890.
Rafael Garcia-Suarez [Wed, 9 Jan 2008 16:52:36 +0000 (16:52 +0000)]
Revert change #32920, since it was solved another way by #32890.
p4raw-link: @32920 on //depot/perl: e3dd4663a7a9c4d106a591d5b1511f7c9d1bfa1e

p4raw-id: //depot/perl@32922

13 years agorefactor PL_opargs generation in opcode.pl and fix helem
Marcus Holland-Moritz [Thu, 3 Jan 2008 01:29:35 +0000 (02:29 +0100)]
refactor PL_opargs generation in opcode.pl and fix helem
Message-ID: <20080103012935.759bda90@r2d2>

p4raw-id: //depot/perl@32921

13 years agoRE: compile warnings w/ PM_SETRE and ReREFCNT_inc
Robin Barker [Mon, 7 Jan 2008 14:10:30 +0000 (14:10 +0000)]
RE: compile warnings w/ PM_SETRE and ReREFCNT_inc
From: "Robin Barker" <Robin.Barker@npl.co.uk>
Message-ID: <46A0F33545E63740BC7563DE59CA9C6D0939AE@exchsvr2.npl.ad.local>

p4raw-id: //depot/perl@32920

13 years agoAdd DEBUG_LEAKING_SCALARS_ABORT, to call abort() if any scalars leak,
Nicholas Clark [Wed, 9 Jan 2008 15:34:42 +0000 (15:34 +0000)]
Add DEBUG_LEAKING_SCALARS_ABORT, to call abort() if any scalars leak,
or are freed too many times.

p4raw-id: //depot/perl@32918

13 years agoAdd a few =cuts (so configpm is displayed mostly correctly
Rafael Garcia-Suarez [Wed, 9 Jan 2008 13:16:34 +0000 (13:16 +0000)]
Add a few =cuts (so configpm is displayed mostly correctly
by perldoc or by search.cpan.org)

p4raw-id: //depot/perl@32916

13 years agoFix leak introduced by change #32873, thanks to Nicholas
Rafael Garcia-Suarez [Wed, 9 Jan 2008 12:33:27 +0000 (12:33 +0000)]
Fix leak introduced by change #32873, thanks to Nicholas
p4raw-link: @32873 on //depot/perl: 3a28f3fb1bfd44e4e3dfe6842af867c8c1c9de28

p4raw-id: //depot/perl@32915

13 years ago1e4 isn't large enough with 16 byte long doubles (at least on x86_64).
Nicholas Clark [Wed, 9 Jan 2008 11:47:12 +0000 (11:47 +0000)]
1e4 isn't large enough with 16 byte long doubles (at least on x86_64).
However, 1e5 does take us "to infinity and beyond"
(Plus use cmp_ok, for better diagnostics)

p4raw-id: //depot/perl@32914

13 years agoChange 32899 missed the other double-reference count.
Nicholas Clark [Wed, 9 Jan 2008 11:40:12 +0000 (11:40 +0000)]
Change 32899 missed the other double-reference count.

p4raw-id: //depot/perl@32913

13 years ago newSV() + sv_upgrade() => newSV_type()
Nicholas Clark [Wed, 9 Jan 2008 11:34:41 +0000 (11:34 +0000)]
 newSV() + sv_upgrade() => newSV_type()

p4raw-id: //depot/perl@32912

13 years agoMove re::regexp_pattern to universal.c
Jerry D. Hedden [Tue, 8 Jan 2008 15:01:02 +0000 (10:01 -0500)]
Move re::regexp_pattern to universal.c
From: "Jerry D. Hedden" <jdhedden@cpan.org>
Message-ID: <1ff86f510801081201q5c36f055re6165ebfe8876c2e@mail.gmail.com>

p4raw-id: //depot/perl@32911