platform/upstream/perl.git
13 years agoIn B's OptreeCheck, implement proper qr// matching for regexps.
Nicholas Clark [Fri, 4 Feb 2011 15:32:28 +0000 (15:32 +0000)]
In B's OptreeCheck, implement proper qr// matching for regexps.

Hence we can now do string matching on strings, rather than treating everything
as a regexp.

13 years agoIn B's OptreeCheck, inline diag_or_fail() into its only caller.
Nicholas Clark [Fri, 4 Feb 2011 15:08:28 +0000 (15:08 +0000)]
In B's OptreeCheck, inline diag_or_fail() into its only caller.

$tc->{goterrs} is not referenced after this function, so no need to re-assign
to it.

13 years agosilence do_curse() compiler warning
David Mitchell [Fri, 4 Feb 2011 15:19:41 +0000 (15:19 +0000)]
silence do_curse() compiler warning

Usage of the static function do_curse() in sv.c has been temporarily
disabled; comment out the function definition too, to avoid a compiler
warning about an unused function.

13 years agoIn Cwd, convert two regexps to explicit ranges, instead of using /i
Nicholas Clark [Fri, 4 Feb 2011 14:10:24 +0000 (14:10 +0000)]
In Cwd, convert two regexps to explicit ranges, instead of using /i

Since change 56ca34cada940c7f moved the folding of ranges from runtime to
compile time, these VMS-only regexps would incur a startup cost penalty on
all platforms.

This also removes build-time warnings when Cwd is invoked before the Unicode
tables have been processed.

13 years agoIn File::Copy, convert two regexps to explicit ranges, instead of using /i
Nicholas Clark [Fri, 4 Feb 2011 13:53:50 +0000 (13:53 +0000)]
In File::Copy, convert two regexps to explicit ranges, instead of using /i

Since change 56ca34cada940c7f moved the folding of ranges from runtime to
compile time, these VMS-only regexps would incur a startup cost penalty on
all platforms.

This also removes build-time warnings when File::Copy is invoked before the
Unicode tables have been processed.

13 years agoFix excess whitespace in pod.
Michael Stevens [Tue, 1 Feb 2011 22:10:23 +0000 (22:10 +0000)]
Fix excess whitespace in pod.

13 years agoIn B's tests, fix calls to like() which weren't being passed a regexp.
Nicholas Clark [Fri, 4 Feb 2011 11:32:55 +0000 (11:32 +0000)]
In B's tests, fix calls to like() which weren't being passed a regexp.

Now that the tests actually *test*, this reveals that one of the tests wasn't
actually correct. Fix that, and in the process make the particular error
message clearer and more consistent.

13 years agoRemove non-working and hence unused features from B's OptreeCheck test code.
Nicholas Clark [Fri, 4 Feb 2011 10:29:53 +0000 (10:29 +0000)]
Remove non-working and hence unused features from B's OptreeCheck test code.

'retry' is no use without 'debug', and 'debug' doesn't work (doesn't enable
regexp debugging output for the retry) because C<use re 'debug'> is lexically
scoped, so can't be applied at runtime after the event to an already compiled
regexp. (And the "obvious" fix of turning it on for compile time isn't working
for some reason, so it's not trivial to fix this unused feature. Version
control will preserve the code if anyone wants to investigate, fix and
resurrect it.)

13 years agoConvert Pod::Html to lexical file handles.
Nicholas Clark [Fri, 4 Feb 2011 09:42:28 +0000 (10:42 +0100)]
Convert Pod::Html to lexical file handles.

This fixes the regression test failures for Module::Build on an NFS filesystem.
What had been happening was that 3 of Module::Build's tests would invoke
Pod::Html::pod2html via ACTION_html(). The invocation is inside an eval, and
is not treated as fatal if it fails. It fails if the install tree doesn't yet
exist, and it doesn't explicitly close the file handle. With package file
handles, this meant that the file was still open when the test attempted to
recursively delete the temporary directory tree, and an NFS file system won't
delete an open file (it will reappear under a different, hidden, filename).
Consequently the directory tree was not empty, cleanup failed, and the test
failed.

Switching to lexical file handles causes the file to automatically be closed
when it goes out of scope due to the thrown exception. This problem is not
going to occur for anyone building Module::Build on NFS against an installed
perl.

13 years agoUpdated HTTP::Tiny to CPAN version 0.010
David Golden [Fri, 4 Feb 2011 07:51:14 +0000 (02:51 -0500)]
Updated HTTP::Tiny to CPAN version 0.010

Includes VMS test fixes by Craig Berry

13 years agouse the macro that casts to HV * so threaded C++ builds, build
Tony Cook [Fri, 4 Feb 2011 04:14:49 +0000 (15:14 +1100)]
use the macro that casts to HV * so threaded C++ builds, build

13 years agoConvert B's "no perlio" skips to "skip all" in the BEGIN blocks.
Nicholas Clark [Thu, 3 Feb 2011 14:50:42 +0000 (14:50 +0000)]
Convert B's "no perlio" skips to "skip all" in the BEGIN blocks.

Previously they called the skip function, and needed to know the number of
tests declared by the plan.

13 years agoRemove commented out 'require test.pl' lines in B's tests.
Nicholas Clark [Thu, 3 Feb 2011 14:41:10 +0000 (14:41 +0000)]
Remove commented out 'require test.pl' lines in B's tests.

13 years agoCorrect the "no perlio" skipping in optree_misc.t.
Nicholas Clark [Thu, 3 Feb 2011 14:37:59 +0000 (14:37 +0000)]
Correct the "no perlio" skipping in optree_misc.t.

Move the last test up into the SKIP block, and correct the skip count.

13 years agoUpdated Module::Metadata to CPAN version 1.000004
David Golden [Thu, 3 Feb 2011 07:59:53 +0000 (02:59 -0500)]
Updated Module::Metadata to CPAN version 1.000004

[DELTA]

1.0.4 2011-02-03 07:55:00
  - Fix broken metadata.t when @INC has relative paths (JJORE)

13 years agocode points above 0x7fff_ffff problematic
Karl Williamson [Thu, 3 Feb 2011 00:03:33 +0000 (17:03 -0700)]
code points above 0x7fff_ffff problematic

The analysis by the submitter was correct.

13 years agoMove ANYOF folding from regexec to regcomp
Karl Williamson [Wed, 2 Feb 2011 19:01:34 +0000 (12:01 -0700)]
Move ANYOF folding from regexec to regcomp

This is for security as well as performance.  It allows Unicode properties to
not be matched case sensitively.  As a result the swash inversion hash is
converted from having utf8 keys to numeric, code point, keys.

It also for the first time fixes the bug where /i doesn't work for a code point
not at the end of a range in a bracketed character class has a multi-character
fold

13 years agoregexp_unicode_prop.t: Add tests, refactor
Karl Williamson [Wed, 2 Feb 2011 18:31:41 +0000 (11:31 -0700)]
regexp_unicode_prop.t: Add tests, refactor

Refactor so that can test /i or not.  Add tests for using the new feature to
have user-defined properties work differently under /i.

13 years agopat-advanced.t: Add new test
Karl Williamson [Wed, 2 Feb 2011 18:30:38 +0000 (11:30 -0700)]
pat-advanced.t: Add new test

13 years agoregexp_unicode_prop.t: Comment nits
Karl Williamson [Wed, 2 Feb 2011 18:26:53 +0000 (11:26 -0700)]
regexp_unicode_prop.t: Comment nits

13 years agoAdd synonym \p{Titlecase}
Karl Williamson [Wed, 2 Feb 2011 18:24:08 +0000 (11:24 -0700)]
Add synonym \p{Titlecase}

13 years agoAdd PL_foldable - a list of chars that participate in folds
Karl Williamson [Wed, 2 Feb 2011 00:17:14 +0000 (17:17 -0700)]
Add PL_foldable - a list of chars that participate in folds

13 years agomktables remove obsolete never-used code
Karl Williamson [Tue, 1 Feb 2011 23:31:33 +0000 (16:31 -0700)]
mktables remove obsolete never-used code

13 years agointrpvar.h: Clarify comment
Karl Williamson [Tue, 1 Feb 2011 23:10:59 +0000 (16:10 -0700)]
intrpvar.h: Clarify comment

13 years ago_swash_inversion_hash Rmv X from embed, add const
Karl Williamson [Tue, 1 Feb 2011 23:02:05 +0000 (16:02 -0700)]
_swash_inversion_hash Rmv X from embed, add const

This shouldn't be called from XS code.

13 years agoAdd initial inversion list object
Karl Williamson [Tue, 1 Feb 2011 17:43:49 +0000 (10:43 -0700)]
Add initial inversion list object

Going forward the intent is to convert from swashes to the better-suited
inversion list data structure.  This adds rudimentary inversion lists that have
only the functionality needed for 5.14.  As a result, they are as much as
possible static to one file.

What's necessary for 5.14 is enough to allow folding of ANYOF nodes to be moved
from regexec to regcomp.  Why they are needed for that is to generate as
compact as possible class definitions; otherwise, very long linear lists might
be generated.  (They still may be, but that's inherent in the problem domain;
this generates as compact as possible, combining overlapping ranges, etc.)

The only two non-trivial methods in this object are from published algorithms.

13 years agoregcomp.c: Generate different property for /i matching
Karl Williamson [Tue, 1 Feb 2011 05:50:00 +0000 (22:50 -0700)]
regcomp.c: Generate different property for /i matching

This patch causes regcomp.c to generate a different property name under /i than
not.  utf8_heavy.pl will later resolve whether this is to match the same under
/i or not, based on the data structure generated by mktables.

This is part of moving non-locale folding into regcomp from regexec.  The
reasons are primarily security, but this has been planned to do at some point
anyway for performance.  It was not until a 5.13.X build that fixed the regexec
code that the case-insensitive matching mostly worked.  With that change, things like

    /\p{ASCII_Hex_Digit}+/i

would match non-ASCII characters, such as LATIN SMALL LIGATURE FF, and almost
certainly that would not be the expectation of the coder.  The Unicode
Standard is silent on the matter, but as of this writing, it appears that they
will act to recommend against caseless matching of properties; I get the sense
that they would never have thought someone would think to do it, but Perl has.
I ran some experiments, and actually very few properties have differences under
caseless matching anyway.  have submitted a proposal to them that says that,
but suggests that certain properties can be grandfathered-in.  Perl users have
come to expect that /\p{Uppercase}/i would match lower case letters, and have
written bug reports that they don't, until 5.13.X fixed them, but in addition
added the unintended wrinkle from the example above.

The design is for mktables to generate tables for /i matching for the few
properties that have differences, and to create a hash mapping the standard
table to the /i table, which is read by utf8_heavy.pl.  regcomp.c munges the
names of all properties under /i to be __foo_i.  The two initial underscores
make sure there is no conflict with existing single underscore initial tables.
utf8_heavy strips these off, and computes the table as normal from the
remaining unmunged name.  At the last moment, it looks up that name in the list
of those that have /i tables, and substitutes if found.

This completely hides all this from the swash mechanism and regexec.c.

This can't be completely hidden from user-defined properties.  Now, a boolean
will be passed to those subroutines indicating if /i is in effect or not.  They
are free to ignore it, but they can return a different set of code points
depending on its value.  They will be called once for each type, and the
results cached by the normal swash mechanism, which thinks these are two
different properties.

13 years agoutf8_heavy: Use new mktables caseless feature
Karl Williamson [Tue, 1 Feb 2011 05:40:24 +0000 (22:40 -0700)]
utf8_heavy: Use new mktables caseless feature

This patch causes utf8_heavy.pl to know about the new data structure that
mktables now generates to indicate what substitute table to use for one that
has different results under /i matching.

Note that regcomp.c, as of this commit, does not generate the names that would
exercise this code.

13 years agomktables: Set caseless equivalent tables
Karl Williamson [Tue, 1 Feb 2011 05:35:48 +0000 (22:35 -0700)]
mktables: Set caseless equivalent tables

Only a few tables will be affected by /i matching.  This hard-codes them all.

Note that at the point of this commit, the rest of the code doesn't implement
this, so the pod changes will delayed until the rest of the infrastructure is
in place.

13 years agomktables: Add concept of caseless equiv table
Karl Williamson [Tue, 1 Feb 2011 05:32:23 +0000 (22:32 -0700)]
mktables: Add concept of caseless equiv table

This adds a new field to the table object giving what table it is equivalent to
under /i matching.

13 years agomktables: factor out sub that duplicates utf8_heavy
Karl Williamson [Tue, 1 Feb 2011 05:14:03 +0000 (22:14 -0700)]
mktables: factor out sub that duplicates utf8_heavy

This code may eventually be placed so there's no duplication

13 years agoregcomp.c: cl_and() fix
Karl Williamson [Tue, 1 Feb 2011 04:05:26 +0000 (21:05 -0700)]
regcomp.c: cl_and() fix

If ANDing two nodes together and they both have UNICODE_ALL set, the result
should also.  I don't have a test case for this, but the bug is exposed by some
commits soon to come in a test case in pat_advanced.t for cl_and.

13 years agoupdate the manicheck step in the RMG to keep up with the times
Jesse Vincent [Wed, 2 Feb 2011 16:15:07 +0000 (11:15 -0500)]
update the manicheck step in the RMG to keep up with the times

13 years agoremove cmpVERSION step from the RMG since abigail automated it
Jesse Vincent [Wed, 2 Feb 2011 16:09:27 +0000 (11:09 -0500)]
remove cmpVERSION step from the RMG since abigail automated it

13 years agomade abigail's new porting/cmp_version.t not scream about blead-upstream
Jesse Vincent [Wed, 2 Feb 2011 16:07:37 +0000 (11:07 -0500)]
made abigail's new porting/cmp_version.t not scream about blead-upstream
module versions since we don't control them

13 years agoBump Attribute::Handlers version since it changed since the last cpan
Jesse Vincent [Wed, 2 Feb 2011 16:04:41 +0000 (11:04 -0500)]
Bump Attribute::Handlers version since it changed since the last cpan
release

13 years agoBump I18N::LangTgs::Detect since it changed since the last CPAN release
Jesse Vincent [Wed, 2 Feb 2011 16:04:15 +0000 (11:04 -0500)]
Bump I18N::LangTgs::Detect since it changed since the last CPAN release

13 years agoBump IO::Select's version since it changed since the last cpan release
Jesse Vincent [Wed, 2 Feb 2011 16:03:54 +0000 (11:03 -0500)]
Bump IO::Select's version since it changed since the last cpan release

13 years agoBump Locale::Maketext's version
Jesse Vincent [Wed, 2 Feb 2011 16:03:40 +0000 (11:03 -0500)]
Bump Locale::Maketext's version

13 years agoMake it so that it acts as if cmpVERSION.pl had not been passed the -x flag.
Abigail [Wed, 2 Feb 2011 12:46:32 +0000 (13:46 +0100)]
Make it so that it acts as if cmpVERSION.pl had not been passed the -x flag.

We'll check non-dual lived files first, then the dual lived ones.

13 years agoBump OS2::Process version since the docs changed triggering a test fail
Jesse Vincent [Tue, 1 Feb 2011 18:54:06 +0000 (13:54 -0500)]
Bump OS2::Process version since the docs changed triggering a test fail

13 years agoRemove whitespace for pod fix.
Michael Stevens [Tue, 18 Jan 2011 11:33:47 +0000 (11:33 +0000)]
Remove whitespace for pod fix.

13 years agoFix NAME section to be valid.
Michael Stevens [Mon, 24 Jan 2011 14:39:44 +0000 (14:39 +0000)]
Fix NAME section to be valid.

13 years agoMake podchecker a little happier
Michael Stevens [Mon, 31 Jan 2011 21:44:53 +0000 (21:44 +0000)]
Make podchecker a little happier

Fix =item ambiguity and remove whitespace on empty line.

13 years agoPod cleanup for F<> and L<>
Michael Stevens [Mon, 31 Jan 2011 21:55:12 +0000 (21:55 +0000)]
Pod cleanup for F<> and L<>

Add more L<> around links and F<> around filenames.
Fix one case where F<> had been incorrectly used instead of L<>.

13 years agoupdate the release manager's guide to remove a now-autmated version
Jesse Vincent [Tue, 1 Feb 2011 18:05:16 +0000 (13:05 -0500)]
update the release manager's guide to remove a now-autmated version
checking step.

13 years agoBump Carp.pm version since it now differs from 5.13.9
Jesse Vincent [Tue, 1 Feb 2011 18:00:11 +0000 (13:00 -0500)]
Bump Carp.pm version since it now differs from 5.13.9

13 years agoPerltidy Carp.pm and Carp.t
Dave Rolsky [Sun, 30 Jan 2011 15:19:34 +0000 (09:19 -0600)]
Perltidy Carp.pm and Carp.t

[committer's note: while we tend to avoid "perltidy this" commits which
 enforce a new author's style upon existing code, Dave's working to
 pick up an existing codebase with a horribly inconsistent style
 and modernize and (hopefully) dual-life it eventually]

13 years agoMake Carp.pm strict and warnings safe.
Dave Rolsky [Tue, 1 Feb 2011 17:38:57 +0000 (12:38 -0500)]
Make Carp.pm strict and warnings safe.

13 years agoBump versions of modules changed since the last blead release (found by
Jesse Vincent [Tue, 1 Feb 2011 16:41:28 +0000 (11:41 -0500)]
Bump versions of modules changed since the last blead release (found by
abigail's new failing test)

13 years agoGet the tag from git describe instead of $]
Abigail [Tue, 1 Feb 2011 12:26:22 +0000 (13:26 +0100)]
Get the tag from git describe instead of $]

13 years agoMerge branch 'blead' of ssh://perl5.git.perl.org/gitroot/perl into blead
Abigail [Tue, 1 Feb 2011 07:23:47 +0000 (08:23 +0100)]
Merge branch 'blead' of ssh://perl5.git.perl.org/gitroot/perl into blead

13 years agoUpdate IPC-Cmd to CPAN version 0.70
Chris 'BinGOs' Williams [Tue, 1 Feb 2011 05:30:40 +0000 (05:30 +0000)]
Update IPC-Cmd to CPAN version 0.70

  [DELTA]

  Changes for 0.70    Mon Jan 31 20:34:09 GMT 2011
  =================================================
  * Apply a patch from Petya Kohts, RT #65276, with
    changes for run_forked:

    1) fix for the typo in the name of the signal
    2) changed default for clean_up_children (which
       seems to be the behavior expected by the majority of the users)
    3) added detection (and forwarding to the caller) of the case
       when run program is killed by signal
    4) kill_gently is now used in cases when parent died
       and when the executed program times out
    5) added options which allow to execute some user code
       in the beginning and at the end of the child

13 years agoTurn exec flag off
Abigail [Mon, 31 Jan 2011 23:47:41 +0000 (00:47 +0100)]
Turn exec flag off

13 years agoMerge branch 'blead' of ssh://perl5.git.perl.org/gitroot/perl into blead
Abigail [Mon, 31 Jan 2011 23:01:52 +0000 (00:01 +0100)]
Merge branch 'blead' of ssh://perl5.git.perl.org/gitroot/perl into blead

13 years agoTurn Porting/cmpVERSION.pl into a test
Abigail [Mon, 31 Jan 2011 22:57:25 +0000 (23:57 +0100)]
Turn Porting/cmpVERSION.pl into a test

13 years agoPartially revert 4155e4fe
Father Chrysostomos [Mon, 31 Jan 2011 02:57:35 +0000 (18:57 -0800)]
Partially revert 4155e4fe

This disables the fix for [perl #36347], which made sure that unrefer-
enced objects were DESTROYed, marking the tests as to-do.

This bug fix broke three CPAN modules. I will probably not have time
to fix them before 5.14, so disabling the fix until after 5.14 seems
the safest option.

This resolves [perl #82542] and other related tickets.

13 years agoDon't refer to specific line numbers in test code regex
Dave Rolsky [Sun, 30 Jan 2011 15:18:42 +0000 (09:18 -0600)]
Don't refer to specific line numbers in test code regex

13 years ago[perl #83018] TEST PATCH: commit 95ea86d5 also fixes sprintf("%.2f")
Zsbán Ambrus [Sun, 30 Jan 2011 22:39:33 +0000 (14:39 -0800)]
[perl #83018] TEST PATCH: commit 95ea86d5 also fixes sprintf("%.2f")

This adds some more test cases about the bug that commit
95ea86d5 fixes.

The bug was that in the perl 5.12 branch, sprintf("%.2f") returns ""
whereas it should return "0.00" -- but this only applies in the very
special cases, not eg. when the template has extra characters. The
log message and the tests of that commit only talks about "%.0f", but
I think it would be worth to also test the "%.2f".

13 years agoPerl_do_sv_dump: dump REGEXP body
David Mitchell [Sun, 30 Jan 2011 23:56:01 +0000 (23:56 +0000)]
Perl_do_sv_dump: dump REGEXP body

Now that a regexp is a first-class SV, dump all the fields
of the 'struct regexp': it's just another SV body type

13 years agoPerl_do_sv_dump: move stub REGEXP code
David Mitchell [Sun, 30 Jan 2011 15:00:27 +0000 (15:00 +0000)]
Perl_do_sv_dump: move stub REGEXP code

SVt_REGEXP was in the 'general SV fields' section; move it to the
'type-specific SV fields' section. Not that it does anything yet.

13 years agoPerl_do_sv_dump: add some blank lines and headers
David Mitchell [Sun, 30 Jan 2011 14:55:41 +0000 (14:55 +0000)]
Perl_do_sv_dump: add some blank lines and headers

Make this long function easier on the eye by separating "paragraphs"
with a blank lines, and with a comment at the start of each major
section.

13 years agoRemove vestigial ORANGE references
David Mitchell [Sun, 30 Jan 2011 14:38:29 +0000 (14:38 +0000)]
Remove vestigial ORANGE references

This was a temporary type used while regexes were in the process
of being promoted to first-class SVs

13 years agoSkip the right number of tests in DosGlob.t.
Craig A. Berry [Sun, 30 Jan 2011 22:47:32 +0000 (16:47 -0600)]
Skip the right number of tests in DosGlob.t.

Three times three is nine, last I checked.  Follow-up to:

dd2d1456ae6cdb1b6f476dd4923e18c36a45b6ea

13 years agofix a confusing open scope in pp_subst
David Mitchell [Fri, 28 Jan 2011 15:08:34 +0000 (15:08 +0000)]
fix a confusing open scope in pp_subst

just a white space change

13 years agoTidy up perlport.pod with more L<> around URLs.
Michael Stevens [Sun, 30 Jan 2011 10:21:14 +0000 (10:21 +0000)]
Tidy up perlport.pod with more L<> around URLs.

13 years agoTidy up pod.
Michael Stevens [Sun, 30 Jan 2011 10:14:54 +0000 (10:14 +0000)]
Tidy up pod.

Add more L<> for http links and F<> for filenames.

13 years agoembed.fnc: Add E flag to regcurly entry
Karl Williamson [Sun, 30 Jan 2011 06:09:25 +0000 (23:09 -0700)]
embed.fnc: Add E flag to regcurly entry

On C++ builds, it wasn't getting seen in extensions.

13 years agoS_regcurly needs thread context.
Craig A. Berry [Sun, 30 Jan 2011 04:29:53 +0000 (22:29 -0600)]
S_regcurly needs thread context.

Actually I can't see any reason it actually *needs* thread context,
but 881ffab65cdbee2f146ada660e5593bad2e71472 added thread context
to the prototype without adding it to the function definition, thus
breaking builds with -Dusethreads.

13 years agoAnother address for Schubiger
Father Chrysostomos [Sat, 29 Jan 2011 19:53:56 +0000 (11:53 -0800)]
Another address for Schubiger

13 years ago[perl #82996] Use the user's from address as return-path in perlbug
David Leadbeater [Sat, 29 Jan 2011 19:24:49 +0000 (11:24 -0800)]
[perl #82996] Use the user's from address as return-path in perlbug

Many systems thesedays don't have a valid internet domain name and
perlbug@perl.org does not accept email with a return-path that does not
resolve. Therefore pass the user's address to sendmail so it's less
likely to get stuck in a mail queue somewhere.

13 years ago[perl #82988] Skip hanging taint.t test on MirBSD 10
Steven Schubiger [Sat, 29 Jan 2011 19:23:48 +0000 (11:23 -0800)]
[perl #82988] Skip hanging taint.t test on MirBSD 10

This patch allows for make test to complete when perl is configured
with ithreads support (verified for MirBSD 10 on eurynome.mirbsd.org
where also the Test::Smoke core smoker runs).

13 years ago[perl #82980] Typo fix in Attribute::Handlers
David Leadbeater [Sat, 29 Jan 2011 19:22:57 +0000 (11:22 -0800)]
[perl #82980] Typo fix in Attribute::Handlers

13 years agoReindent recent IPv6/getaddrinfo code consistently with the rest of the files
Paul "LeoNerd" Evans [Mon, 24 Jan 2011 18:55:38 +0000 (18:55 +0000)]
Reindent recent IPv6/getaddrinfo code consistently with the rest of the files

13 years agoAdd Larwan Berke (Apocalypse) to AUTHORS
Jesse Vincent [Sat, 29 Jan 2011 19:12:53 +0000 (14:12 -0500)]
Add Larwan Berke (Apocalypse) to AUTHORS

13 years agoMove IN6ADDR_ANY and IN6ADDR_LOOPBACK to @EXPORT_OK in Socket.pm, requested by Nicholas
Larwan Berke [Thu, 27 Jan 2011 19:54:14 +0000 (12:54 -0700)]
Move IN6ADDR_ANY and IN6ADDR_LOOPBACK to @EXPORT_OK in Socket.pm, requested by Nicholas

13 years agoUse embed.pl inline capabilities for regcurly
Karl Williamson [Sat, 29 Jan 2011 17:11:39 +0000 (10:11 -0700)]
Use embed.pl inline capabilities for regcurly

Change the regcurly definition to use the new inline abilities.

13 years agoembed.fnc: Add inline function capability
Karl Williamson [Sat, 29 Jan 2011 16:53:44 +0000 (09:53 -0700)]
embed.fnc: Add inline function capability

This patch adds to embed.pl the capability to generate static inline functions.

13 years agomktables: simplify a little code
Karl Williamson [Fri, 28 Jan 2011 16:43:11 +0000 (09:43 -0700)]
mktables: simplify a little code

13 years agomktables: Rmv surrogates caution
Karl Williamson [Fri, 28 Jan 2011 16:39:16 +0000 (09:39 -0700)]
mktables: Rmv surrogates caution

As isolated surrogates are now accepted in the Perl core

13 years agoperlapi: Clarify statement about av_push()
Karl Williamson [Sat, 22 Jan 2011 16:22:08 +0000 (09:22 -0700)]
perlapi: Clarify statement about av_push()

Contrary to the previous text, av_store() requires the user to deal with the
reference count of the scalar being added.

13 years agoperldiag.pod: Expand \p in locale description
Karl Williamson [Fri, 28 Jan 2011 16:01:05 +0000 (09:01 -0700)]
perldiag.pod: Expand \p in locale description

13 years agoPOD typo
Jay Hannah [Sat, 29 Jan 2011 02:07:16 +0000 (18:07 -0800)]
POD typo

13 years agoAdd another address for Jay Hannah to checkAUTHORS.pl
Father Chrysostomos [Sat, 29 Jan 2011 04:30:27 +0000 (20:30 -0800)]
Add another address for Jay Hannah to checkAUTHORS.pl

13 years agobmodpow() fails when GMP library is used.
Peter John Acklam [Sat, 29 Jan 2011 02:00:30 +0000 (18:00 -0800)]
bmodpow() fails when GMP library is used.

This bug is not seen when the libraries Calc and FastCalc are used,
because their _modpow() method modifies the first argument. However,
the GMP library's _modpow() does not modify the first argument,
causing a so far undetected bug to show up and fail tests.

Using the Calc library prints the correct "-4":

    use Math::BigInt lib => Calc;
    $x = Math::BigInt->new(8);
    $y = Math::BigInt->new(8);
    $z = Math::BigInt->new(-5);
    print $x -> bmodpow($y, $z), "\n";

Using the GMP library prints the incorrect "--":

    use Math::BigInt lib => GMP;
    $x = Math::BigInt->new(8);
    $y = Math::BigInt->new(8);
    $z = Math::BigInt->new(-5);
    print $x -> bmodpow($y, $z), "\n";

13 years agoFix endless loop in Porting/makemeta.
Craig A. Berry [Fri, 28 Jan 2011 15:20:17 +0000 (09:20 -0600)]
Fix endless loop in Porting/makemeta.

File::Basename::dirname() returns the current device and directory
on VMS when there is no directory information in the path (i.e.,
you ran out of directories), and the current device and directory
are never ".".  Now that we run makemeta in the test suite, the
expectation that you always get "." was causing the test suite to
hang in a CPU-bound loop.

13 years agoSwitch ExtUtils::CBuilder to use Perl::OSType
David Golden [Fri, 28 Jan 2011 19:15:56 +0000 (14:15 -0500)]
Switch ExtUtils::CBuilder to use Perl::OSType

13 years agoRevert "pod/perlpolicy.pod: fix grammar in 70a565f"
Ricardo Signes [Fri, 28 Jan 2011 16:08:34 +0000 (11:08 -0500)]
Revert "pod/perlpolicy.pod: fix grammar in 70a565f"

This reverts commit 5ccb3a28fc30a358c7870c0f979a11d432b97664.

13 years agoremove some stray apostrophes from perlpolicy
Ricardo Signes [Fri, 28 Jan 2011 15:49:29 +0000 (10:49 -0500)]
remove some stray apostrophes from perlpolicy

13 years agoperldelta: 3bef022 depend on files needed for porting/buildtoc.t
Tony Cook [Fri, 28 Jan 2011 11:01:32 +0000 (22:01 +1100)]
perldelta: 3bef022 depend on files needed for porting/buildtoc.t

13 years agoregcomp: Add warning if tries to use \p in locale.
Karl Williamson [Fri, 28 Jan 2011 05:29:51 +0000 (22:29 -0700)]
regcomp: Add warning if tries to use \p in locale.

\p implies Unicode matching rules, which are likely going to be different than
the locale's.

13 years agotoke.c: Add comment
Karl Williamson [Fri, 28 Jan 2011 05:13:33 +0000 (22:13 -0700)]
toke.c: Add comment

13 years ago[perl #38456] binmode FH, ":crlf" only modifies top crlf layer
Leon Timmermans [Thu, 20 Jan 2011 22:32:28 +0000 (23:32 +0100)]
[perl #38456] binmode FH, ":crlf" only modifies top crlf layer

When pushed on top of the stack, crlf will no longer enable crlf layers
lower in the stack. This will prevent unexpected results.

13 years agoSuppress "Name used only once" warnings.
David E. Wheeler [Fri, 28 Jan 2011 04:33:44 +0000 (20:33 -0800)]
Suppress "Name used only once" warnings.

From Locale::Maketext and I18N::LangTags::Detect. To replicate, run
    perl -Idist/Locale-Maketext/lib -Icpan/IPC-Cmd/lib -Idist/I18N-LangTags/lib -we "use IPC::Cmd"

13 years agoFix test count introduced by previous commit
Karl Williamson [Thu, 27 Jan 2011 23:24:24 +0000 (16:24 -0700)]
Fix test count introduced by previous commit

Commit 99ca48e139a839de3f512cb8cae93f1fe51062ab

13 years agoregex: \p{} in pattern implies Unicode semantics
Karl Williamson [Thu, 27 Jan 2011 23:03:11 +0000 (16:03 -0700)]
regex: \p{} in pattern implies Unicode semantics

Now, a Unicode property match specified in the pattern will indicate that the pattern is meant for matching according to Unicode rules

13 years agoperlre.pod: Fix omission of /a
Karl Williamson [Thu, 27 Jan 2011 21:48:01 +0000 (14:48 -0700)]
perlre.pod: Fix omission of /a

13 years agoAdd Franz Fasching to AUTHORS
Father Chrysostomos [Thu, 27 Jan 2011 22:52:27 +0000 (14:52 -0800)]
Add Franz Fasching to AUTHORS

13 years agoIO::Select: allow removal of IO::Handle objects without fileno
Franz Fasching [Thu, 27 Jan 2011 16:46:59 +0000 (08:46 -0800)]
IO::Select: allow removal of IO::Handle objects without fileno

13 years agopod/perlpolicy.pod: fix grammar in 70a565f
Ævar Arnfjörð Bjarmason [Thu, 27 Jan 2011 22:09:21 +0000 (22:09 +0000)]
pod/perlpolicy.pod: fix grammar in 70a565f

Change the "was within" wording added in 70a565f to "was released
within".