Nicholas Clark [Tue, 13 Jul 2010 19:20:21 +0000 (20:20 +0100)]
Fix transient pod infelicities spotted by podcheck.t
David Golden [Tue, 13 Jul 2010 16:08:58 +0000 (12:08 -0400)]
Add more recent commits to perldelta
Karl Williamson [Thu, 8 Jul 2010 16:18:58 +0000 (10:18 -0600)]
perlop.pod: Document current \000 \x behavior
Signed-off-by: David Golden <dagolden@cpan.org>
Chris 'BinGOs' Williams [Tue, 13 Jul 2010 11:14:35 +0000 (12:14 +0100)]
Update CPANPLUS to CPAN version 0.9007
[DELTA]
Changes for 0.9007 Tue Jul 13 10:44:30 2010
================================================
* Fix the shell test to skip if test is not being run under a
terminal, reported by Justin Case RT #59254
Changes for 0.9006 Fri Jul 9 13:44:22 2010
================================================
* Finally resolved the issue where a prereq on Config would not be
recognised as a core module
Chris 'BinGOs' Williams [Tue, 13 Jul 2010 11:08:05 +0000 (12:08 +0100)]
Update ExtUtil-ParseXS to CPAN version 2.2206
[DELTA]
2.2206 - Sun Jul 4 15:43:21 EDT 2010
Bug fixes:
- Make xsubpp accept the _ prototype (RT#57157) [Rafael Garcia-Suarez]
- INCLUDE_COMMAND portability fixes for VMS (RT#58181) [Craig Berry]
- INCLUDE_COMMAND fixes to detect non-zero exit codes (RT#52873)
[Steffen Mueller]
Chris 'BinGOs' Williams [Tue, 13 Jul 2010 11:02:16 +0000 (12:02 +0100)]
Update Archive-Tar to CPAN version 1.64
[DELTA]
* important changes in version 1.64 09/07/2010
- Removed the PERL_CORE specific chdir from all the tests
- Apply a patch from David Muir Sharnoff RT #58916,
"skip files via a callback and limit memory use when skipping files"
- Apply a patch from Daphne Pfister RT #59150
"Assumes all references filename are IO::Handle's instead of trying to stringify."
Chris 'BinGOs' Williams [Tue, 13 Jul 2010 10:56:15 +0000 (11:56 +0100)]
Update IPC-Cmd to CPAN version 0.60
[DELTA]
Changes for 0.60 Mon Jul 5 09:04:54 BST 2010
=================================================
* Corrected spelling mistakes in POD, spotted by H.Merijn Brand
* Apply a patch from Burak Gursoy RT #58886, which fixes paths
on MSWin32
* Apply patch from Petya Kohts, RT #50398, which allows more
flexible configuration of run_forked and its children
Josh ben Jore [Tue, 13 Jul 2010 06:56:55 +0000 (23:56 -0700)]
Reset variables being initialized by readline before we do possibly blocking I/O
Tony Cook [Tue, 13 Jul 2010 05:11:54 +0000 (15:11 +1000)]
new is a reserved word in C++, don't use it as a variable name
Fix a cause of black smoke for C++ builds.
Nicholas Clark [Mon, 12 Jul 2010 12:39:19 +0000 (13:39 +0100)]
Create S_assert_uft8_cache_coherent() with one copy of the cache panic code.
Replacing 4 copies of this debugging-only routine with 1 reduces source and
object code size.
Nicholas Clark [Mon, 12 Jul 2010 12:09:28 +0000 (13:09 +0100)]
Perl_sv_len_utf8 can use the UTF-8 offset cache to reduce its linear scan.
Previously, if the scalar's character length wasn't yet known, but an offset
midway was, the offset would be ignored, and the linear scan of UTF-8 was for
the entire length of the scalar.
Nicholas Clark [Mon, 12 Jul 2010 10:56:59 +0000 (11:56 +0100)]
Perl_sv_pos_b2u now calls utf8_mg_len_cache_update for the string end offset.
Previously it would not take special action if the offset requested happened to
be the end of the string, meaning that the (fixed size) UTF-8 offset cache
would be used for a value which could (and should) be stored elsewhere.
Nicholas Clark [Mon, 12 Jul 2010 10:38:31 +0000 (11:38 +0100)]
S_sv_pos_u2b_cached now updates the UTF-8 length cache if at the end of string.
Pass in a boolean to S_sv_pos_u2b_forwards, which sets it to true if it
discovers that the UTF-8 offset is at (or after) the end of the string.
This can only happen if we don't already know the SV's length (in Unicode
characters), because if we know it, we always call S_sv_pos_u2b_midway().
Nicholas Clark [Mon, 12 Jul 2010 10:16:41 +0000 (11:16 +0100)]
Break S_utf8_mg_len_cache_update() out from Perl_sv_len_utf8().
George Greer [Sat, 10 Jul 2010 23:18:12 +0000 (19:18 -0400)]
Fix for RT#2353: catch yyparse() exceptions in (?{...})
Ticket is entitled:
"panic: top_env when localizing a lexical variable in qr{(?{...})}"
but a better description would be:
"(?{...}) does not catch yyparse() exceptions in regex compilation"
Changes Perl_sv_compile_2op() to tell doeval() that it must use
S_try_yyparse() to catch any exception that yyparse() may throw.
Ben Morrow [Wed, 9 Dec 2009 10:37:14 +0000 (10:37 +0000)]
Update MANIFEST.
Ben Morrow [Wed, 9 Dec 2009 10:32:23 +0000 (10:32 +0000)]
Document the blockhook functions and macros.
Ben Morrow [Wed, 9 Dec 2009 10:24:33 +0000 (10:24 +0000)]
Teach autodoc.pl about 'o' functions.
That is, functions with no #define foo Perl_foo. I'm not certain this is
the right way to do it, as I don't really understand which flags autodoc
honours from which places; currently, it's necessary to put the 'o' flag
on the =for apidoc line or it will be ignored.
Ben Morrow [Mon, 7 Dec 2009 19:00:04 +0000 (19:00 +0000)]
Systematic tests for the block hooks.
I've left the dummy implementation of @{^C_S_C} in, as it's actually
useful for some of the other tests. (Something simpler would work just
as well, of course.)
Ben Morrow [Mon, 7 Dec 2009 12:55:57 +0000 (12:55 +0000)]
Wrap PL_blockhooks in an API function.
This should help prevent people from thinking they can get cute with the
contents.
Ben Morrow [Mon, 7 Dec 2009 11:52:23 +0000 (11:52 +0000)]
Macroify the block_hooks structure.
Add a flags member, so it can be extended later if necessary. Add a
bhk_eval member, called from doeval to catch requires and string evals.
Ben Morrow [Thu, 26 Nov 2009 17:22:22 +0000 (17:22 +0000)]
Initial very basic tests for PL_blockhooks.
This is taken directly from rafl's @{^COMPILE_SCOPE_CONTAINER}
implementation posted on p5p.
Ben Morrow [Thu, 26 Nov 2009 17:18:29 +0000 (17:18 +0000)]
Generic hooks into Perl_block_{start,end}.
These take the form of a vtable pushed onto the new PL_blockhooks array.
This could probably do with a API around it later. Separate pre_end and
post_end hooks are needed to capture globals before the stack is unwound
(like needblockscope in the existing code). The intention is that once
a vtable is installed it never gets removed, so where necessary
extensions using this will need to use a hinthv element to determine
whether to do anything or not.
Nicholas Clark [Sun, 11 Jul 2010 19:11:10 +0000 (20:11 +0100)]
Avoid UTF-8 cache panics with offsets beyond the string. Fixes RT #75898.
Change S_sv_pos_u2b_forwards() to take a point to the (requested) UTF-8 offset,
and return the actual UTF-8 offset for the byte position returned. This ensures
that the cache is consistent with reality.
Nicholas Clark [Sun, 11 Jul 2010 16:17:37 +0000 (17:17 +0100)]
In Perl_sv_pos_u2b_flags and S_sv_pos_u2b_cached, return early for offset 0.
0 Unicode characters are always 0 octets long. Returning early ensures that
any offsets we calculate later will always be non-zero.
Nicholas Clark [Sun, 11 Jul 2010 15:49:29 +0000 (16:49 +0100)]
In S_sv_pos_u2b_midway, inline the call to S_sv_pos_u2b_forwards.
Vincent Pit [Sun, 11 Jul 2010 11:34:28 +0000 (13:34 +0200)]
Sort and complete the op flags lists for op_dump()
Tony Cook [Sun, 11 Jul 2010 11:23:30 +0000 (21:23 +1000)]
RT #73520: POSIX::strftime memory leak
Tested before/after with valgrind.
Josh ben Jore [Sat, 10 Jul 2010 14:54:55 +0000 (07:54 -0700)]
Die with $@ instead of empty message
Josh ben Jore [Sat, 10 Jul 2010 14:54:32 +0000 (07:54 -0700)]
Remove extra/useless $@ check after eval { require PadWalker } (which is still checked)
Josh ben Jore [Sat, 10 Jul 2010 14:53:35 +0000 (07:53 -0700)]
Promote eval( "require ..." ) to eval { require ... }
Josh ben Jore [Sat, 10 Jul 2010 14:52:06 +0000 (07:52 -0700)]
Promote eval { require( ... )} || die to mere require( ... )
Josh ben Jore [Sat, 10 Jul 2010 14:50:41 +0000 (07:50 -0700)]
Remove indirect object notation from debugger
Josh ben Jore [Sat, 10 Jul 2010 14:38:31 +0000 (07:38 -0700)]
Document that @{$main::{'_<'.$filename}} lines are dualvar to (COP*).
Nicholas Clark [Fri, 9 Jul 2010 09:42:41 +0000 (10:42 +0100)]
In t/harness, clear PERL5LIB, PERLLIB, PERL5OPT as t/TEST does.
In fact, as t/harness requires t/TEST, simply get t/TEST to do it for
t/harness too.
Jerry D. Hedden [Thu, 8 Jul 2010 15:20:35 +0000 (11:20 -0400)]
Silence compiler warning
Signed-off-by: David Golden <dagolden@cpan.org>
karl williamson (via RT) [Tue, 6 Jul 2010 15:27:34 +0000 (08:27 -0700)]
PATCH: t/lib/common.pl localized $/ setting; allow input 'no_plan'
# New Ticket Created by karl williamson
# Please include the string: [perl #76398]
# in the subject line of all future correspondence about this issue.
# <URL: http://rt.perl.org/rt3/Ticket/Display.html?id=76398 >
See attached patches
>From
275169a8df5851331669da2dd791483379c3fc5e Mon Sep 17 00:00:00 2001
From: Karl Williamson <khw@khw-desktop.(none)>
Date: Tue, 6 Jul 2010 09:20:42 -0600
Subject: [PATCH] t/lib/common.pl: localize changing $/
$/ changes should not affect callers.
Signed-off-by: H.Merijn Brand <h.m.brand@xs4all.nl>
Abigail [Thu, 8 Jul 2010 07:21:22 +0000 (09:21 +0200)]
Fix typo
Jerry D. Hedden [Wed, 7 Jul 2010 16:22:09 +0000 (12:22 -0400)]
Reorganize dist/threads-shared
Move dist/threads-shared/shared.pm to dist/threads-shared/lib/threads/shared.pm
and remove its Makefile.PL.
Jan Dubois [Wed, 7 Jul 2010 22:22:44 +0000 (15:22 -0700)]
Get rid of PERL_DECL_PROT
It was added for PERL_OBJECT support in commit 0cb9638,
which has been removed again with commit acfe0ab.
Jerry D. Hedden [Wed, 30 Jun 2010 15:15:50 +0000 (11:15 -0400)]
Reorganize dist/threads
Move threads.pm to lib/threads.pm and remove Makefile.PL
Ricardo Signes [Tue, 6 Jul 2010 15:41:53 +0000 (11:41 -0400)]
cautionary note about "no VERSION"
Signed-off-by: David Golden <dagolden@cpan.org>
Bo Lindbergh [Tue, 6 Jul 2010 15:13:12 +0000 (17:13 +0200)]
Change of address for Bo Lindbergh
Signed-off-by: David Golden <dagolden@cpan.org>
David Golden [Tue, 6 Jul 2010 12:27:13 +0000 (08:27 -0400)]
Add Bo Lindburgh alternate email to checkAUTHORS.pl
Bo Lindbergh [Sun, 4 Jul 2010 23:31:01 +0000 (01:31 +0200)]
Code for allowing uppercase X/B in hexadecimal/binary numbers (#76296).
Signed-off-by: David Golden <dagolden@cpan.org>
Bo Lindbergh [Sun, 4 Jul 2010 22:04:46 +0000 (00:04 +0200)]
Tests for allowing uppercase X/B in hexadecimal/binary numbers (#76296).
Signed-off-by: David Golden <dagolden@cpan.org>
David Golden [Tue, 6 Jul 2010 05:38:12 +0000 (01:38 -0400)]
Clean up new Locale-Maketest test for use in core
Todd Rinaldo [Tue, 6 Jul 2010 05:28:00 +0000 (01:28 -0400)]
Locale::Maketext external cache support
This patch with tests provides RO support for lexicon hashes in
Locale::Maketext. This allows you to have GDBM language files owned by
root which can be accessed by non-root, but not altered.
If your lexicon is a tied hash the simple act of caching the compiled
value can be fatal.
For example a GDBM_File GDBM_READER tied hash will die with something
like:
gdbm store returned -1, errno 2, key "..." at ...
All you need to do is turn on caching outside of the lexicon hash itself
like so:
sub init {
my ($lh) = @_;
...
$lh->{'use_external_lex_cache'} = 1;
...
}
And then instead of storing the compiled value in the lexicon hash it
will store it in $lh->{'_external_lex_cache'}
I've verified that blead is the authoritative location for
Locale::Maketext source.
Signed-off-by: David Golden <dagolden@cpan.org>
David Golden [Tue, 6 Jul 2010 04:42:53 +0000 (00:42 -0400)]
replace spaces with tabs in MANIFEST
David Golden [Tue, 6 Jul 2010 04:33:31 +0000 (00:33 -0400)]
Don't set strict for 'no 6;'
Commit
faee19b51573e81abe8811f1256a1d27777d6d04 was incomplete
and only stopped features from being enabled under 'no'.
This patch merges all 'use N.NN' type logic into a single if clause
and then checks version numbers in separate if statements within it.
David Golden [Tue, 6 Jul 2010 04:10:03 +0000 (00:10 -0400)]
Ignore generated files from Time-HiRes
Jerry D. Hedden [Mon, 5 Jul 2010 18:50:04 +0000 (14:50 -0400)]
Ensure new arg to watchdog() in t/test.pl gets initialized
Jesse Vincent [Mon, 5 Jul 2010 11:46:48 +0000 (12:46 +0100)]
added a "list" command to list open tickets that might have patches
Jesse Vincent [Mon, 5 Jul 2010 11:41:28 +0000 (12:41 +0100)]
we can't depend on patches having filenames. Thanks git!
Jesse Vincent [Mon, 5 Jul 2010 11:26:52 +0000 (12:26 +0100)]
First pass at a tool to help porters apply RT patches from the
commandline
Nicholas Clark [Mon, 5 Jul 2010 11:45:19 +0000 (12:45 +0100)]
In Perl_lex_start(), use newSVpvn_flags() to reduce source and object size.
Nicholas Clark [Mon, 5 Jul 2010 09:32:38 +0000 (10:32 +0100)]
In pp_regcomp and pp_entereval, use newSVpvn_flags() to simplify code.
Karl Williamson [Sat, 3 Jul 2010 18:31:54 +0000 (12:31 -0600)]
Speed up viacode
Capturing parentheses greatly slow down regexes, at least here.
On my machine, viacode took 27 seconds for the 22K Unicode names without
capturing parens; 45s with.
Karl Williamson [Sat, 3 Jul 2010 17:55:57 +0000 (11:55 -0600)]
charnames.t update because of rebase
Use of t/lib/common.pl caused some glitches; some behaviors of the
underlying is() functions changed, so revised .t to work under this
scheme.
Karl Williamson [Sat, 3 Jul 2010 17:55:06 +0000 (11:55 -0600)]
Correct perldelta for these changes
Karl Williamson [Sat, 3 Jul 2010 16:22:59 +0000 (10:22 -0600)]
Add vi hint for non-std format of charnames.pm
Karl Williamson [Sat, 3 Jul 2010 16:22:04 +0000 (10:22 -0600)]
More charnames pod updates
Karl Williamson [Sat, 3 Jul 2010 16:12:33 +0000 (10:12 -0600)]
charnames: check for use bytes in vianame; efficiency
When vianame returns a chr, it now verifies that it is legal under 'use
bytes'. Update .t
An instance of taking of a substr of a huge string is needed only in an
error leg. Move it to that leg for performance.
And make the message a subroutine so will be identical whenever raised.
Karl Williamson [Thu, 1 Jul 2010 22:25:08 +0000 (16:25 -0600)]
Clean up charnames pod, including new changes
This patch brings the charnames pod up-to-date, and rewords it to
hopefully be more clear.
Karl Williamson [Thu, 1 Jul 2010 22:22:14 +0000 (16:22 -0600)]
Clean up viacode, accept large aliases
This changes viacode to accept aliases that the user has defined beyond
the Unicode range.
Karl Williamson [Thu, 1 Jul 2010 22:06:51 +0000 (16:06 -0600)]
Extend \N{} enhancements to vianame()
This patch refactors charnames so that vianame and \N call the same
common subroutine so that they have as identical behavior as possible.
Karl Williamson [Wed, 30 Jun 2010 21:27:39 +0000 (15:27 -0600)]
Bump version; some pod cleanup
Karl Williamson [Wed, 30 Jun 2010 21:19:39 +0000 (15:19 -0600)]
charnames: add CORE:: to hex()
Other programs do this; I don't know why just hex() needs to be
protected from user override, but I'm just copying prior art.
Karl Williamson [Wed, 30 Jun 2010 20:55:16 +0000 (14:55 -0600)]
Abandon plans to change viacode's return of unassigned
The BUGS section of the charnames pod said that it was a bug to return
undef for unassigned characters, whereas the real Unicode name is the
empty string. demerphq noted that undef stringifies to the empty
string, so we are in fact in compliance with the standard. This
clarifies the pod wording, removing the text from the BUGS section.
Karl Williamson [Wed, 30 Jun 2010 20:42:59 +0000 (14:42 -0600)]
Allow defining custom charnames to ordinals
This adds the ability of a user to create a custom alias that maps to a
numeric ordinal value, instead of an official Unicode name.
The number of hashes went up so that is better to refer to them by a
name than a number, so I renamed them.
Also, viacode will return any defined user's alias for an otherwise
unamed code point.
This change is principally so that private use characters can be named
so it is more convenient to use them in Perl.
Karl Williamson [Wed, 30 Jun 2010 19:48:33 +0000 (13:48 -0600)]
Add clarifying comments to charnames.t
Karl Williamson [Wed, 30 Jun 2010 19:46:39 +0000 (13:46 -0600)]
Reword feedback request.
It's not clear to me what should be done about the problem of vianame
being bipolar.
Karl Williamson [Tue, 29 Jun 2010 18:54:33 +0000 (12:54 -0600)]
Add a number of abbrs and variants to \N{}
This patch adds the standard abbreviations for the control characters
(such as ACK, BEL, etc) to the repertoire that \N{} knows about. It
also adds a few common variants of their full names, and the old names
for the 4 controls that Unicode has chosen not to have any names at all
for.
The patch also adds all the abbreviations that Unicode lists in 5.2 for
longer characters, such as NBSP, SHY, LRE, ...
To preserve complete backward compatibilty for these and future changes,
user-defined aliases are now checked first, before these are.
As a performance enhancement, these aliases are mapped to their actual
code values instead of their full names which then had to be looked up
in the large table. Now that is avoided, and the table is not loaded
at all until a name is encountered that is not one of these aliases.
The pod and .t are updated.
Jesse Vincent [Sun, 4 Jul 2010 15:47:34 +0000 (11:47 -0400)]
Remove a release_manager_guide step since it's now been automated away
Jesse Vincent [Sun, 4 Jul 2010 15:44:03 +0000 (11:44 -0400)]
Update Makefile to regen META.yml as it regens other files in regen_all
Jesse Vincent [Sun, 4 Jul 2010 15:40:33 +0000 (11:40 -0400)]
Make the META.yml regen code use regen_lib.pl per the request in release_mananger_guide,
Jesse Vincent [Sun, 4 Jul 2010 15:17:20 +0000 (11:17 -0400)]
Missing AUTHORS will now generate failing tests. Removing the manual
checkAUTHORS step from the release engineering guide.
Jesse Vincent [Sun, 4 Jul 2010 15:15:35 +0000 (11:15 -0400)]
Make checkAUTHORS a test to further reduce the release burden
Jesse Vincent [Sun, 4 Jul 2010 15:14:55 +0000 (11:14 -0400)]
Further refactoring of checkAUTHORS
Jesse Vincent [Sun, 4 Jul 2010 14:38:50 +0000 (10:38 -0400)]
refactoring checkAUTHORS in advance of testifcation
Jesse Vincent [Sun, 4 Jul 2010 14:25:02 +0000 (10:25 -0400)]
Bring AUTHORS up to date
Chris 'BinGOs' Williams [Sun, 4 Jul 2010 19:02:28 +0000 (20:02 +0100)]
Fix Digest-MD5 tests to run properly under core
Chris 'BinGOs' Williams [Sun, 4 Jul 2010 15:46:01 +0000 (16:46 +0100)]
Update Digest-MD5 to CPAN version 2.40
[DELTA]
2010-07-03 Gisle Aas <gisle@ActiveState.com>
Release 2.40
Marc Pignat (1):
Safer alignment test [RT#35823]
Robin Barker (1):
consting in new ext/
Gisle Aas (1):
Remove the MacOS branch of this test
David Golden [Sun, 4 Jul 2010 18:56:21 +0000 (14:56 -0400)]
remove trailing space in cmpVERSION
George Greer [Sun, 4 Jul 2010 17:09:56 +0000 (13:09 -0400)]
Make Test::Harness source_handler.t use the build perl (%PERL_CORE%) on Win32.
Craig A. Berry [Sun, 4 Jul 2010 14:49:55 +0000 (09:49 -0500)]
Add Matt Johnson to AUTHORS.
Matt Johnson [Sat, 3 Jul 2010 18:08:15 +0000 (19:08 +0100)]
docs: updates for git-based cmpVERSION.pl
Update the release_managers_guide.pod with the new command usage,
and remove the todo item
Matt Johnson [Sat, 3 Jul 2010 18:01:58 +0000 (19:01 +0100)]
Update Porting/cmpVERSION.pl to use git
Update Porting/cmpVERSION.pl to take a directory and a git tag,
rather than two directories, as suggested in perltodo.
Rafael Garcia-Suarez [Sun, 4 Jul 2010 11:31:57 +0000 (13:31 +0200)]
Make previous test fix conditional on the platform being Windows
George Greer [Sun, 4 Jul 2010 05:20:53 +0000 (01:20 -0400)]
Force t/io/openpid.t to use the alarm() watchdog strategy.
* The default watchdog strategy on Win32 doesn't work because the watchdog
is executed via subshell ("cmd /c perl -e '...'"). When the test finishes
and tries to kill the watchdog all it manages to do is whack cmd.exe and
the watchdog still fires, potentially killing an innocent process.
* Trying to use the fork() strategy instead causes openpid.t's test #9's
"ok" to be lost occasionally. The message itself is generated by the 4th
child of the test and should go directly to stdout, but sometimes the
output vanishes for unexplained reasons. This doesn't appear to happen
without the watchdog enabled.
* Using the alarm() watchdog strategy seems to not cause any badness.
David Mitchell [Sat, 3 Jul 2010 16:01:33 +0000 (17:01 +0100)]
add some comments to pp_concat
make it clearer what type of concat each code branch handles
David Mitchell [Sat, 3 Jul 2010 15:47:05 +0000 (16:47 +0100)]
make_patchnum.pl: handle not on a branch properly
David Mitchell [Sat, 3 Jul 2010 14:41:34 +0000 (15:41 +0100)]
avoid multiple FETCH/stringify on filetest ops
some of the filetest operators could call mg_get and/or overload fallback
stringify multiple times
David Mitchell [Sat, 3 Jul 2010 13:24:11 +0000 (14:24 +0100)]
add my_[l]stat_flags(); make my_[l]stat() mathoms
my_stat() and my_lstat() call get magic on the stack arg, so create _flags()
variants that allow us to control this. (I can't just change the signature
or the mg_get() behaviour since my_[l]stat() are listed as being in the
public API, even though they're undocumented.)
David Mitchell [Sat, 3 Jul 2010 12:36:59 +0000 (13:36 +0100)]
PL_amagic_generation doesn't show overload loaded
PL_amagic_generation is non-zero even without the presence of
'use overload', so don't bother using it as a short-cut test of
whether we can skip AMAGIC processing
David Mitchell [Sat, 3 Jul 2010 12:24:08 +0000 (13:24 +0100)]
fix bad indentation in pp_regcomp
David Mitchell [Sat, 3 Jul 2010 12:17:40 +0000 (13:17 +0100)]
avoid extra FETCHes on overloaded qr stringify
/$tied/ called FETCH too many times if the FETCH returned an overloaded
object with no qr method, but with stringify fallback
David Mitchell [Fri, 2 Jul 2010 21:06:49 +0000 (22:06 +0100)]
overload.t: clarify concat #FETCH expected
It turns out that the number of FETCHes for the fallback ($tied_ovld . foo)
just needed explaining, not fixing.
David Mitchell [Fri, 2 Jul 2010 20:33:01 +0000 (21:33 +0100)]
remove double stringify-overload from $ovld .= foo
There was a piece of code in pp_concat who's job it was to determine the
UT8ness of the LHS, and it did it in a heavy-handed way to cope with the
special case of a regexp (which is an RV pointing to REGEXP which might
be UTF8)
David Mitchell [Wed, 23 Jun 2010 23:02:39 +0000 (00:02 +0100)]
eval: handle taint of overloaded/tied arg
string eval would check its arg for taint before processing magic, overload
etc. If the magic or overload returned a tainted value, it wouldn't
be detected. Fixes part of #75716.