This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Peter Martini [Sun, 18 Jan 2015 15:54:01 +0000 (10:54 -0500)]
perly.y changes from Lukas Mai in RT 123069
This moves signatures before attributes in the grammar by
creating separate branches for the prototype and signatures
cases, so that the introduced block and the fact that signatures
do not allow for declarations can be handled properly.
Tests and regen_perly to follow.
James E Keenan [Tue, 13 Jan 2015 00:57:12 +0000 (19:57 -0500)]
Fixes for Pod-HTML tests after Pod-Simple update
Patch supplied by theory.
Signed-off-by: Chris 'BinGOs' Williams <chris@bingosnet.co.uk>
James E Keenan [Sun, 11 Jan 2015 15:20:31 +0000 (10:20 -0500)]
Update Pod-Simple to CPAN version 3.29
[DELTA]
2015-01-13 David E. Wheeler <david@justatheory.org>
* Release 3.29
NOTE: This will be the last release of Pod::Simple to support versions
of Perl earlier than 5.6.
Fixed test failure for unsupported encoding on Perl 5.6.
Removed unnecessary dependency on constant. Reported by Petr Pisar
(RT #85578).
Many documentation and comment typos fixed by David Steinbrunner.
Fixed issue when an encoding is declared via an '=encoding' directive
*after* Pod::Simple has already detected an encoding. Thanks to
Graham Knop for the patch and Patrice Dumas for the tests.
Added '>&STDERR' as an output to better emulate
Pod::Parser::parse_from_file(). Florent Angly.
Tidied the description of the different Pod::Simple interfaces,
with greater emphasis on the four different steps to subclassing.
Florent Angly.
Added the 'top_anchor' attribue to Pod::Simple::HTML, to allow
specification of a custom top anchor format. Gabor Szabo.
Fixed the text formatter to properly set the Text::Wrap overflow policy.
Thanks to Nigel Metheringham for the fix!
Fixed mis-matched quotation marks when adding JavaScript elements
to the XHTML output. Patch from jonathancast.
Fixed issues on EBCDIC. Patch from Karl Williamson (GitHub Pull
Request #58).
Fixed bug where the value passed to Pod::Simple::XHTML's
html_css() method would be output twice. Thanks to Andrew Johnson
for the report (RT #99303).
Documented the 'recurse' attribute of Pod::Simple::Search.
Thanks to Jim Keenan for the report (RT #91886).
An error is no longer thrown when '=encoding' directives are found
on subsequent lines. Instead, we now log a an error message,
"Cannot have multiple =encoding directives" (RT #91757).
The HTML and XHTML formatters now do a better job of preserving
lines in 'begin html' blocks. Fixes a bug that dates from 2002
(Perl RT#9385, CPAN RT #91851).
Collateral damage to 3 tests under ext/Pod-Html/.
James E Keenan [Sun, 18 Jan 2015 03:09:21 +0000 (22:09 -0500)]
perldelta update for 5944609.
Karen Etheridge [Sat, 17 Jan 2015 19:31:47 +0000 (11:31 -0800)]
Module::Metadata 1.000026
James E Keenan [Sat, 17 Jan 2015 22:23:36 +0000 (17:23 -0500)]
One more tiny spelling correction in comment; thanks to alh.
James E Keenan [Sat, 17 Jan 2015 22:02:50 +0000 (17:02 -0500)]
Two small spelling corrections in comment.
David Mitchell [Sat, 17 Jan 2015 19:53:04 +0000 (19:53 +0000)]
avoid $a["foo" eq $x ? 0 : 1] warning
RT #123609
The compiling code that looks for OP_MULTIDEREF candidates was
unconditionally numifying the first constant of an array index expression,
before it had confirmed that the expression consisted solely of a const.
So in something like
$a['foo' eq $x ? 0 : 1]
the 'foo' would be numified and give a spurious warning:
Argument "foo" isn't numeric
This commit fixes it by skipping the OP_MULTIDEREF optimisation if the
array index const isn't SvIOK(). In theory this means that something like
$r->["0"]
no longer gets optimised, but I think we can live with that :-)
It also means that the test for the const being SvROK is no longer
necessary.
(Finally, I moved the declaration of the iv var down a few scopes as it
wasn't being used elsewhere in the wider scope.)
Matthew Horsfall (alh) [Sat, 17 Jan 2015 16:40:45 +0000 (11:40 -0500)]
Update Porting/Maintainers.pl for Devel::PPPort 3.27->3.28.
Missed this in b127ff5
Chris 'BinGOs' Williams [Sat, 17 Jan 2015 16:12:41 +0000 (16:12 +0000)]
Update Module-Load-Conditional to CPAN version 0.64
[DELTA]
0.64 Sat Jan 17 13:33:57 GMT 2015
* Resolve an edge-case with DEPRECATED
[ https://rt.cpan.org/Public/Bug/Display.html?id=101555 ]
Karl Williamson [Fri, 16 Jan 2015 03:13:45 +0000 (20:13 -0700)]
regcomp.c: White-space only
Indent inside a newly formed block
Karl Williamson [Fri, 16 Jan 2015 03:03:09 +0000 (20:03 -0700)]
regcomp.c: Fix bug in /[A-Z]/i
This also fixes /[a-z]/i.
When not under /i, these two ranges alone in a bracketed character class
can be optimized into qr/[[:upper:]]/a and qr/[[:lower:]]/a respectively.
This optimization saves space in the pattern (as no bitmap is needed),
and I think it executes faster. But this optimization has to be
foregone under /i (unless /a is also present) because otherwise
certain non-ASCII characters such as the \N{KELVIN SIGN} don't match,
and they should.
Karl Williamson [Fri, 16 Jan 2015 02:40:24 +0000 (19:40 -0700)]
regcomp.c: Fix comment
Karl Williamson [Fri, 16 Jan 2015 01:52:35 +0000 (18:52 -0700)]
regcomp.c: Improve generated code for some [:posix:] classes
For regexes compiled not under /l, [:posix:] classes in general require
special handling, as
1) under /d they can match two different sets of code points,
depending on whether or not the target string is in UTF-8.
2) under /a, the matches above the ASCII range need to be filtered
out.
But note that nothing special is needed for /u, and under [:ascii:],
there is nothing to be filtered out, and the UTF-8ness of the target
doesn't matter either. And since [:digit:] and [:xdigit:] don't have
matches in the upper Latin1 range, under /d the UTF-8ness doesn't matter
for them either.
This commit skips the special handling for the above conditions. This
leads to less work during pattern compilation, and can mean faster
run-time code, as well, skipping function calls that otherwise would be
done.
Karl Williamson [Fri, 16 Jan 2015 01:35:40 +0000 (18:35 -0700)]
regcomp.c: Rmv unneeded temporary
Karl Williamson [Thu, 15 Jan 2015 18:32:22 +0000 (11:32 -0700)]
Fix regex pattern dump of ANYOF nodes
An ANYOF node is used to handle bracketed character classes and Unicode
properties. It is implemented in part as a bit vector of code points
that match for smallish ordinals, plus other means to specify larger
code points, or code points that match only under certain conditions.
For example, under /d, whether [\w] matches word characters in the upper
Latin1 range depends on whether the target string is in UTF-8 or not,
so the characters which are conditionally matched are saved outside
the bitmap. For performance, there is a flag that gets set if all code
points too large for the bitmap match. This is a common occurrence for
complemented classes, such as [^a-z]. Prior to this commit, if this
flag was set, those things that are conditionally matched were not
output when dumping the regex pattern (under the command line option
-Dr, for example).
Karl Williamson [Wed, 26 Jun 2013 23:44:33 +0000 (17:44 -0600)]
t/re/reg_mesg.t: Don't need to skip on EBCDIC
It now works
Karl Williamson [Fri, 16 Jan 2015 02:44:19 +0000 (19:44 -0700)]
perlrecharclass: Add some missing documentation
Karl Williamson [Fri, 16 Jan 2015 01:42:10 +0000 (18:42 -0700)]
Regen podcheck db because of recent fixes to epigraphs.pod
Karl Williamson [Fri, 16 Jan 2015 04:20:54 +0000 (21:20 -0700)]
t/run/locale.t: Fix typos
Matthew Horsfall [Fri, 16 Jan 2015 12:53:16 +0000 (07:53 -0500)]
Update Devel-PPPort to CPAN version 3.28
[DELTA]
3.28 - 2015-01-16
* Update MANIFEST for regened base/todo files
Chris 'BinGOs' Williams [Fri, 16 Jan 2015 10:18:41 +0000 (10:18 +0000)]
Update Unicode-Collate to CPAN version 1.10
[DELTA]
1.10 Thu Jan 15 21:37:58 2015
- XS: const &c [rt.cpan.org #101170] [PATCH]
- Makefile.PL: [rt.cpan.org #101500] [PATCH]
Chris 'BinGOs' Williams [Fri, 16 Jan 2015 10:17:25 +0000 (10:17 +0000)]
Update Test-Harness to CPAN version 3.35
[DELTA]
3.35 2015-01-14
- Fix prove --version to actually print the version (Leon Timmermans, #101216)
- Add --version to usage message (Leon Timmermans, #101215)
Steve Hay [Thu, 15 Jan 2015 13:43:35 +0000 (13:43 +0000)]
perl5216delta correction - IO::Socket has connected(), not isconnected()
Steve Hay [Thu, 15 Jan 2015 09:01:38 +0000 (09:01 +0000)]
epigraphs - Fix a broken I<>
Steve Hay [Thu, 15 Jan 2015 08:57:10 +0000 (08:57 +0000)]
epigraphs - Consistent -RCn style
Steve Hay [Thu, 15 Jan 2015 08:56:33 +0000 (08:56 +0000)]
epigraphs - Sort by Perl version number
This was originally the case, then resorted chronologically by commit
2831a86cee, then back to version order by commit
c7bed2601e, since when
some inconsistencies have crept back in (entries inserted in chronological
order again).
I really don't care what order it is in, just so long as there is
consistency in whatever order is chosen.
Steve Hay [Thu, 15 Jan 2015 08:31:06 +0000 (08:31 +0000)]
epigraphs - Remove trailing whitespace
Steve Hay [Thu, 15 Jan 2015 08:28:08 +0000 (08:28 +0000)]
epigraphs - Consistent quoting of titles
Steve Hay [Wed, 14 Jan 2015 18:43:12 +0000 (18:43 +0000)]
epigraphs - No need to repeat authors/titles after the epigraphs
Steve Hay [Wed, 14 Jan 2015 18:33:36 +0000 (18:33 +0000)]
epigraphs - Standardize two-space indent for poetry etc; no indent for prose
Steve Hay [Wed, 14 Jan 2015 18:07:36 +0000 (18:07 +0000)]
epigraphs - Remove use of =over/=back in six cases since no others do that
Steve Hay [Wed, 14 Jan 2015 18:03:17 +0000 (18:03 +0000)]
epigraphs - No need to break up long L<> lines
Steve Hay [Wed, 14 Jan 2015 17:59:58 +0000 (17:59 +0000)]
epigraphs - Prefer nntp.perl.org/ links, and fill in a missing one
Also correct a couple more dates in the process, and fill in a missing
epigraph.
Steve Hay [Wed, 14 Jan 2015 13:40:53 +0000 (13:40 +0000)]
epigraphs - Correct some dates and links and fill in a couple of missing names
Karl Williamson [Wed, 14 Jan 2015 20:45:40 +0000 (13:45 -0700)]
Add text about EBCDIC to pods: perlhack* perlport
Karl Williamson [Wed, 14 Jan 2015 18:16:50 +0000 (11:16 -0700)]
perlhacktips: Fix typo
Karl Williamson [Wed, 14 Jan 2015 18:01:21 +0000 (11:01 -0700)]
t/charset_tools.pl: Improve function names
'latin1_to_native' and 'native_to_latin1' actually operate on all code
points, not just the ranges implied by their names. This commit changes
the names to reflect that, to 'uni_to_native' and 'native_to_uni'.
The old names have only been available since 5.21, so there is no
backcompat issue.
Karl Williamson [Wed, 14 Jan 2015 17:54:31 +0000 (10:54 -0700)]
Use more universal function in tests where makes sense
By using a different function that is just as applicable, or more so, in
these cases, we can avoid having to require a file.
The functions that translate code points between character sets are
always available and just as appropriate for the uses in the three tests
t/re/fold_grind.t, t/re/reg_fold.t, and lib/unicore/TestProp.pl. Change
to use it instead of the functions that take string inputs. This is
slightly cleaner, as the file that contains it no longer has to be
'required'. And in some of the cases the code point function is a
better choice, as there had to be translation between chr and ord
before.
Karl Williamson [Wed, 14 Jan 2015 17:51:09 +0000 (10:51 -0700)]
t/test.pl: Rmv obsolete comment
The referred-to functions have been moved to a separate file, and are no
longer contained in this one.
Chris 'BinGOs' Williams [Wed, 14 Jan 2015 16:49:58 +0000 (16:49 +0000)]
Update Devel-PPPort to CPAN version 3.27
[DELTA]
3.27 - 2015-01-13
* Regen base and todo files.
* Fix isASCII and isCNTRL for <5.6.0
* Other minor fixups.
(Thanks to mhx for all of the work on this release)
Chris 'BinGOs' Williams [Wed, 14 Jan 2015 16:07:44 +0000 (16:07 +0000)]
Update Getopt-Long to CPAN version 2.43
[DELTA]
Changes in version 2.43
-----------------------
* Fix bug https://rt.cpan.org/Ticket/Display.html?id=92462
* Implement enhancement
https://rt.cpan.org/Public/Bug/Display.html?id=101537
* Fix the Getopt::Long part of bug
https://rt.cpan.org/Ticket/Display.html?id=100335
Chris 'BinGOs' Williams [Wed, 14 Jan 2015 14:12:22 +0000 (14:12 +0000)]
uname -r works on MINIX as well
Karl Williamson [Tue, 13 Jan 2015 19:53:34 +0000 (12:53 -0700)]
locale.c: Add comment; move #if
A better comment is added. The #if is moved so that the rare
compilation that doesn't use LC_CTYPE, no unused variable warning would
be generated.
Karl Williamson [Tue, 13 Jan 2015 05:31:07 +0000 (22:31 -0700)]
Move unlikely executed macro to function
The bulk of this macro is extremely rarely executed, so it makes sense
to optimize for space, as it is called from a fair number of places, and
move as much as possible to a single function.
For whatever it's worth, on my system with my typical compilation
options, including -O0, the savings was 19640 bytes in regexec.o, 4528
in utf8.o, at a cost of 1488 in locale.o.
Karl Williamson [Tue, 13 Jan 2015 05:01:53 +0000 (22:01 -0700)]
locale.c: Fix memory leak.
I spotted this in code review. I didn't add a test for it, because to
expose the much more serious bug fixed by the previous commit, I had to
temporarily change the C code to force these extremely
unlikely-to-be-taken branches to execute.
Karl Williamson [Tue, 13 Jan 2015 05:06:34 +0000 (22:06 -0700)]
sv.c: Fix cloning of interp variable
This should have been cloned by sv_dup_inc. I don't know why it didn't
fail outside of Win32, but fail there it did, spectacularly, but only
when I temporarily change some C code to force execution of the affected
code, which only gets executed for problematic locales, which may very
well not exist on most machines.
Karl Williamson [Mon, 12 Jan 2015 22:21:51 +0000 (15:21 -0700)]
t/re/reg_mesg.t: White-space only
Indent, and wrap lines
Karl Williamson [Mon, 12 Jan 2015 22:13:02 +0000 (15:13 -0700)]
regcomp.c: Fix comment typos
Karl Williamson [Mon, 12 Jan 2015 22:12:24 +0000 (15:12 -0700)]
dquote_static.c: Rmv obsolete comment
Karl Williamson [Mon, 5 Jan 2015 20:17:58 +0000 (13:17 -0700)]
Add 'strict' subpragma to 'use re'
This subpragma is to allow p5p to add warnings/errors for regex patterns
without having to worry about backwards compatibility. And it allows
users who want to have the latest checks on their code to do so. An
experimental warning is raised by default when it is used, not because
the subpragma might go away, but because what it catches is subject to
change from release-to-release, and so the user is acknowledging that
they waive the right to backwards compatibility. I will be working in
the near term to make some changes to what is detected by this.
Note that there is no indication in the pattern stringification that it
was compiled under this. This means I didn't have to figure out how to
stringify it. It is fine because using this doesn't affect what the
pattern gets compiled into, if successful. And interpolating the
stringified pattern under either strict or non-strict should both just
work.
Karl Williamson [Mon, 5 Jan 2015 18:59:04 +0000 (11:59 -0700)]
regcomp.c: Add 'strict' parameter to S_regclass()
This function has the capability to do strict checking, with the
variable 'strict', but it is initialized based on another parameter's
value. This commit causes 'strict' to be passed in, so it is
independent of other parameters.
Karl Williamson [Mon, 5 Jan 2015 04:29:10 +0000 (21:29 -0700)]
Reserve a bit for 'the re strict subpragma.
This is another step in the process
Karl Williamson [Sun, 4 Jan 2015 16:55:35 +0000 (09:55 -0700)]
Add new warnings category for "use re 'strict'"
This is a step in the process of adding that subpragma.
Karl Williamson [Sun, 11 Jan 2015 17:06:01 +0000 (10:06 -0700)]
Output warning in qr// only once
This warning is being output in both passes of pattern compilation
Steve Hay [Tue, 13 Jan 2015 17:45:05 +0000 (17:45 +0000)]
Consistent output from core-cpan-diff
It uses 'Perl:', not 'perl:', a few lines later.
(We currently have "found more than one CPAN candidate" for libnet, which
I have addressed by scheduling the removal of old versions from CPAN: 1.29
was the last version containing some packages, so it remained in
02packages.details.txt because of that.)
David Mitchell [Tue, 13 Jan 2015 14:58:22 +0000 (14:58 +0000)]
update AUTHORS
Rostislav Skudnov [Wed, 24 Dec 2014 06:12:52 +0000 (08:12 +0200)]
make re_intuit_string() return correct string
Fix #123469 - Bug in split function, with utf8 strings
Each regex has two SV pointers, check_substr and check_utf8,
which hold a constant string (if any) corresponding to the longest
constant string in the regexp.
When the regex is first compiled, only one pointer is set, depending on
whether the pattern is utf8 or not; but subsequent usage of the regex can
instantiate the other pointer too. So which of the two strings
re_intuit_string() should return should be based on the UTF8ness of the
pattern, not whether check_substr is set.
Chris 'BinGOs' Williams [Tue, 13 Jan 2015 11:11:37 +0000 (11:11 +0000)]
Update libnet to CPAN version 3.05
[DELTA]
3.05 2015-01-12
- Fixed infinite loop in Net::SMTP::auth(). [CPAN RT#100235]
Tony Cook [Tue, 13 Jan 2015 06:05:11 +0000 (17:05 +1100)]
perldelta for
ac0e53cd2e68
Tony Cook [Tue, 13 Jan 2015 05:28:15 +0000 (16:28 +1100)]
report which files have duplicates in porting/manifest.t
Ed J [Mon, 29 Dec 2014 11:59:40 +0000 (11:59 +0000)]
Simplify ext/SDBM_File layout and build
Tony: fix MANIFEST
Chris 'BinGOs' Williams [Tue, 13 Jan 2015 00:39:46 +0000 (00:39 +0000)]
Update MIME-Base64 to CPAN version 3.15
[DELTA]
2015-01-12 Gisle Aas <gisle@ActiveState.com>
Release 3.15
Avoid compilation warnings on Windows.
Restore compatibility with perl-5.4
Chris 'BinGOs' Williams [Tue, 13 Jan 2015 00:23:00 +0000 (00:23 +0000)]
Update Digest-MD5 to CPAN version 2.54
[DELTA]
2015-01-12 Gisle Aas <gisle@ActiveState.com>
Release 2.54
David Mitchell: silence some compiler warnings
Jonathan Hall: Add ->context() feature
Steve Hay: Sync with blead
bulk88: const the vtable
zefram: 5.6 threads test fix
David Mitchell [Mon, 12 Jan 2015 16:36:51 +0000 (16:36 +0000)]
perl.c: silence compiler warning:
perl.c:3629:75: warning: value computed is not used [-Wunused-value]
gv_SVadd(PL_errgv);
Jarkko Hietaniemi [Mon, 12 Jan 2015 03:57:37 +0000 (22:57 -0500)]
Test for nan range ends.
Jarkko Hietaniemi [Mon, 12 Jan 2015 03:58:35 +0000 (22:58 -0500)]
Detect infnan range ends.
Jarkko Hietaniemi [Mon, 12 Jan 2015 03:58:15 +0000 (22:58 -0500)]
Separate bad range end detection.
Jarkko Hietaniemi [Mon, 12 Jan 2015 03:27:45 +0000 (22:27 -0500)]
Test for inf range ends.
Jarkko Hietaniemi [Mon, 12 Jan 2015 03:18:00 +0000 (22:18 -0500)]
Explicitly test x infnan producing empty string.
Chris 'BinGOs' Williams [Mon, 12 Jan 2015 11:02:38 +0000 (11:02 +0000)]
Update IO-Socket-IP to CPAN version 0.36
[DELTA]
0.36 2014/01/10 20:54:58
[BUGFIXES]
* Always bind() on MSWin32 because Windows doesn't like it if we
listen() having not done so (RT101396)
Chris 'BinGOs' Williams [Mon, 12 Jan 2015 11:01:41 +0000 (11:01 +0000)]
Update Digest-SHA to CPAN version 5.95
[DELTA]
5.95 Sat Jan 10 12:15:36 MST 2015
- modified the bit-ordering test (ref. t/bitorder.t)
-- supplied directory-change preamble for CORE builds
Father Chrysostomos [Mon, 12 Jan 2015 06:28:18 +0000 (22:28 -0800)]
Father Chrysostomos [Mon, 12 Jan 2015 06:25:51 +0000 (22:25 -0800)]
perldelta for
08b999a9d / #123452
Father Chrysostomos [Mon, 12 Jan 2015 06:21:56 +0000 (22:21 -0800)]
perldelta for
2384afee9 / #123553
Father Chrysostomos [Mon, 12 Jan 2015 06:18:12 +0000 (22:18 -0800)]
Father Chrysostomos [Mon, 12 Jan 2015 06:14:00 +0000 (22:14 -0800)]
Father Chrysostomos [Mon, 12 Jan 2015 06:11:56 +0000 (22:11 -0800)]
perldelta for
eb4ec35b34 / #123514
Father Chrysostomos [Mon, 12 Jan 2015 06:09:20 +0000 (22:09 -0800)]
Father Chrysostomos [Mon, 12 Jan 2015 06:05:33 +0000 (22:05 -0800)]
perldelta for
d8bd3d828 / #123495
Father Chrysostomos [Mon, 12 Jan 2015 06:03:21 +0000 (22:03 -0800)]
Father Chrysostomos [Mon, 12 Jan 2015 05:47:34 +0000 (21:47 -0800)]
perldelta for
3c6ef0a5 / #103260 / #123071
Father Chrysostomos [Mon, 12 Jan 2015 05:44:53 +0000 (21:44 -0800)]
Father Chrysostomos [Mon, 12 Jan 2015 05:44:13 +0000 (21:44 -0800)]
perldelta for
62ae7cfb674910
Father Chrysostomos [Mon, 12 Jan 2015 05:38:07 +0000 (21:38 -0800)]
Father Chrysostomos [Mon, 12 Jan 2015 05:34:06 +0000 (21:34 -0800)]
Correct perl ver in Safe Changes
Daniel Dragan [Thu, 1 Jan 2015 04:09:28 +0000 (23:09 -0500)]
refactor op.c S_bad_type_*v
-flags arg of both funcs is unused in all callers. Move the 0 to the funcs.
flags arg is from commit
ce16c625ec in 2012
-all bad_type_gv calls are right before the end of the switch, the pushing
of 1st 3 args and call asm ops can be merged together, leaving the 1
string constant push as the only unique op between the 7 src code
callers of bad_type_gv, this requires reordering the args so the only
unique one is the last/right most one, reordering can't be done to
bad_type_pv because each following execution point after each bad_type_pv
is different, bad_type_pv's caller/s are not a switch statement
- commit
53e06cf030 probably overlooked the 2 PL_op_desc[type] places,
OP_DESC is a fancier superset of PL_op_desc[type], since calling
bad_type_pv only happens during a PP syntax error, that is not
performance critical, so replace PL_op_desc[type] with OP_DESC and
factor out OP to description string lookup, plus custom ops are very rare
so this shouldn't impact the error message seen by the user
VC2003 .text section of perl521.dll before 0xc9543 after 0xC9523
Father Chrysostomos [Mon, 12 Jan 2015 04:39:03 +0000 (20:39 -0800)]
Confused cloning of nested state subs
use feature 'lexical_subs','state';
no warnings 'experimental';
my $sub = sub{
state sub sb4;
state sub a {
state $x = 42;
sub sb4 { $x; }
}
a();
print sb4(), "\n";
};
$sub->();
The output:
Bizarre copy of CODE in subroutine exit at - line 10.
The sb4 sub was trying to close over the wrong pad; namely, the one
belonging to the anonymous sub.
Father Chrysostomos [Mon, 12 Jan 2015 04:08:12 +0000 (20:08 -0800)]
perldiag grammo
Jarkko Hietaniemi [Mon, 12 Jan 2015 01:05:34 +0000 (20:05 -0500)]
Catch infnan repeat counts.
Not entirely convinced this is worth the extra code but getting
"Negative repeat count" warning for NaN repeat count is too much.
Even before this patch, "a" x $Inf didn't go all Genghis over your
virtual memory. This is all about the right warning.
Daniel Dragan [Thu, 8 Jan 2015 05:10:13 +0000 (00:10 -0500)]
Win32: stat() only after a failed open() on a module
See RT ticket for this patch for details.
Steve Hay [Sat, 10 Jan 2015 16:17:19 +0000 (16:17 +0000)]
Add missing =item * in perldelta (now perl5217delta) from commit
070733dfda
Steve Hay [Sat, 10 Jan 2015 16:09:36 +0000 (16:09 +0000)]
Correct a bug number in perldelta
Steve Hay [Sat, 10 Jan 2015 15:39:08 +0000 (15:39 +0000)]
Steve Hay [Sat, 10 Jan 2015 15:35:27 +0000 (15:35 +0000)]
Bump $VERSION following previous commit
The code is currently identical to 0.23, but when the next change is made
the $VERSION must not be bumped to 0.24 since that has already been
released. For the same reason, we can't use 0.24 now, so bump to 0.25.
Steve Hay [Sat, 10 Jan 2015 15:29:26 +0000 (15:29 +0000)]
Revert "fix attributes memcmp without len<6 (asan catch)"
This reverts commit
ff5314c5aa54b172122507017cac034acf1c4bf7.
As noted by Andy Dougherty in #122629 the problem that this fixes
was already fixed more succinctly in
dd369969a/#122701.
Karl Williamson [Thu, 8 Jan 2015 21:36:00 +0000 (14:36 -0700)]
perldelta for [perl #123530]
Karl Williamson [Thu, 8 Jan 2015 20:49:56 +0000 (13:49 -0700)]
POSIX.pod: Various nits, typos, clarifications
Karl Williamson [Thu, 8 Jan 2015 19:22:21 +0000 (12:22 -0700)]
perlpodspec: Corrections/adds to detecting =encoding
C0 and C1 are not legal UTF-8 start bytes. utf8::decode() is a more
accurate way of determining UTF-8.
Karl Williamson [Thu, 8 Jan 2015 19:16:50 +0000 (12:16 -0700)]
toke.c: White-space only
Outdent a block that was indented by 2 tab widths.