This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Karl Williamson [Fri, 24 Feb 2012 16:32:58 +0000 (09:32 -0700)]
mktables: Correct Unicode 6.1 omission
Unicode 6.1 erroneously omitted Takri as a script that uses two
characters, and have voted to publish the correction that this patch
makes. There isn't an official Corrigendum yet.
Ricardo Signes [Fri, 24 Feb 2012 16:08:51 +0000 (11:08 -0500)]
correct a Pod foible that broke blead
Max Maischein [Fri, 24 Feb 2012 13:41:11 +0000 (14:41 +0100)]
Add description of new change process for dev.perl.org
Leo prefers now direct pull requests via Github.
Karl Williamson [Thu, 23 Feb 2012 16:01:54 +0000 (09:01 -0700)]
regcomp.c: Add cast to silence Solaris warning
David Golden [Thu, 23 Feb 2012 14:01:48 +0000 (09:01 -0500)]
Remove ExtUtils-CBuilder's MANIFEST.SKIP
Apparently, 'make distclean' tries to use it in confusing and
counterproducive ways. Easier to remove it than fix the other.
Ricardo Signes [Thu, 23 Feb 2012 00:24:22 +0000 (19:24 -0500)]
work harder to get useful diagnostics on Win32
The pod2html tests can do diffs of big hunks, and so they try to
use /usr/bin/diff to show how things differ. If there is no such
program, we just say "it wasn't good."
Unfortunately, Win32 is where we keep having problems, and where
we are least likely to have diff available. When there is no diff,
we now use Test::More::is, which will output the whole differing file.
Noisy, yes, but now smoke reports with failing Pod::Html tests will
be more useful.
David Golden [Wed, 22 Feb 2012 22:45:06 +0000 (17:45 -0500)]
ExtUtils::CBuilder: add support information
Adds a documentation note about support. Adds a README.patching
file with instructions for bumping versions, Changes, etc.
David Golden [Wed, 22 Feb 2012 21:29:11 +0000 (16:29 -0500)]
Update Maintainers.pl for new ExtUtils::CBuilder
Also updates release instructions with a reminder to
update Maintainers.pl
David Golden [Wed, 22 Feb 2012 21:20:02 +0000 (16:20 -0500)]
add ext/Pod-Functions/.gitignore
David Golden [Wed, 22 Feb 2012 20:47:56 +0000 (15:47 -0500)]
ExtUtils::CBuilder - add dual-life support files
Adds Makefile.PL, LICENSE, etc. and updates Changes for
work in blead from 5.14.0
David Golden [Wed, 22 Feb 2012 20:29:08 +0000 (15:29 -0500)]
ignore all MYMETA files, not just .yml
Ricardo Signes [Tue, 21 Feb 2012 00:35:33 +0000 (19:35 -0500)]
"no feature" now means reset to default
See https://rt.perl.org/rt3/Ticket/Display.html?id=108776
"no feature" now resets to the default feature set. To disable all
features (which is likely to be a pretty special-purpose request, since
it presumably won't match any named set of semantics) you can now
write "no feature ':all'"
Ricardo Signes [Wed, 22 Feb 2012 14:33:45 +0000 (09:33 -0500)]
unixify paths a bit more systematicaly in Pod-Html
Rather than unixifying paths sort of once off, we now do what I
threatened earlier: unixify all the --switch supplied paths during
command line parsing.
There are other bits of File::Spec being sprinkled around, quite
possibly in too-interior sections of the program, but this change
gets *all* tests passing on Win32 as well as GNU/Linux.
I think an audit of filepath normalization "border security" is
still due, but this may tide us over for now.
Ricardo Signes [Wed, 22 Feb 2012 03:54:58 +0000 (22:54 -0500)]
don't assume that filesys starts at / in Pod-Html
The test expectations for the --htmlroot feature wanted to look for
an absolute path, so it concatenated / with the relative-to-root path.
That failed on Win32, where the cwd was C:\etc
I added a new token to the test-expectations-munger for the absolute
cwd. I'm not 100% sure about the features' behavior, but the tests
are now less platform-specific. I look forward to input on whether
the code is doing the right thing now...
Ricardo Signes [Wed, 22 Feb 2012 03:19:56 +0000 (22:19 -0500)]
unixify the --outfile arg to pod2html
I think we probably need a more systematic regimen of of unixify
calls, but for now, this gets tests passing on Win32.
Max Maischein [Tue, 21 Feb 2012 17:59:26 +0000 (18:59 +0100)]
Make corelist-perldelta.pl keep the existing heading
86c08a2ca2546ef08513c65dabf686423cade2f3 changed
Porting/corelist-perldelta.pl to be more accepting in section headings
when reading in an existing perldelta.
caaa1415975f6b8763b186234d77803148a4fd37 changed it to output
the correctly pluralized form of "Pragmata".
corelist-perldelta.pl used the same data structure for both, input
and output. This patch divides up the logic between input and output
in two data structures, and keeps the heading of the input
file when generating a fresh pod section.
Max Maischein [Tue, 21 Feb 2012 17:13:49 +0000 (18:13 +0100)]
Fix typo
Max Maischein [Tue, 21 Feb 2012 17:13:38 +0000 (18:13 +0100)]
Pluralize "Modules and Pragma" to "Modules and Pragmata"
Max Maischein [Tue, 21 Feb 2012 16:47:25 +0000 (17:47 +0100)]
Add perl5158delta.pod
Max Maischein [Tue, 21 Feb 2012 16:16:58 +0000 (17:16 +0100)]
Add epigraph for 5.15.8
Marc Green [Tue, 7 Feb 2012 16:32:54 +0000 (11:32 -0500)]
Fix bug in pod2html crossreferencing
Ricardo Signes [Mon, 20 Feb 2012 22:35:29 +0000 (17:35 -0500)]
makerel should not add -uncommitted to tarballs
For some time now (I suspect since
691ce773f), makerel has always tried
to build releases as perl-5.x.y-uncommitted.tar.gz. This always
confuses and/or panics new release engineers, who see the error before
the "don't worry, use -s ''" advice in the RMG.
This change just makes makerel skip the usually-#ifdef-ed-away line that
is getting that -uncommitted. makerel still picks up other
locally-applied patches, but will skip uncommitted.
Max Maischein [Mon, 20 Feb 2012 23:01:14 +0000 (00:01 +0100)]
Merge branch 'release-5.15.8' into blead
Max Maischein [Mon, 20 Feb 2012 22:08:57 +0000 (23:08 +0100)]
Fix typos in pod/perldelta.pod
Max Maischein [Mon, 20 Feb 2012 20:20:55 +0000 (21:20 +0100)]
Update pod/perldelta.pod from the version numbers
Max Maischein [Mon, 20 Feb 2012 19:47:13 +0000 (20:47 +0100)]
add new release to perlhist
Max Maischein [Mon, 20 Feb 2012 19:36:46 +0000 (20:36 +0100)]
Update Module::CoreList for v5.15.8
Max Maischein [Mon, 20 Feb 2012 19:19:52 +0000 (20:19 +0100)]
Clean out XXX in preparation for Module::CoreList update
Max Maischein [Mon, 20 Feb 2012 18:41:33 +0000 (19:41 +0100)]
Bump versions from 5.15.7 to 5.15.8
Max Maischein [Mon, 20 Feb 2012 18:31:23 +0000 (19:31 +0100)]
Update perldelta.pod for 5.15.8 for review
Pending remains the list of changed/upgraded module versions, to be
done after the version bump.
Karl Williamson [Mon, 20 Feb 2012 01:04:03 +0000 (18:04 -0700)]
handy.h: Silence Solaris compiler warning
Making this an unsigned constant silences the scary and wrong Solaris
warnings about integer overflow
Karl Williamson [Sun, 19 Feb 2012 00:39:27 +0000 (17:39 -0700)]
podcheck.t: Typo in pod
Tony Cook [Sun, 19 Feb 2012 10:55:50 +0000 (21:55 +1100)]
Avoid defining the same global between perl and re.pm
This caused -Uusedl builds to fail due to the duplicate symbol.
A more complex fix can be expected after 5.16.
Max Maischein [Sun, 19 Feb 2012 20:07:02 +0000 (21:07 +0100)]
Make t/porting/authors.t work on Windows
The Windows shell cmd.exe does not know about single quotes. Double
quotes are the only thing available here. This patch selects the kind
of quote to use based on the operating system and constructs
the pipeline accordingly.
Ricardo Signes [Sun, 19 Feb 2012 16:12:43 +0000 (11:12 -0500)]
Revert updates to compression libraries
Revert "Update Compress-Raw-Bzip2 to CPAN version 2.049"
This reverts commit
db98b5a25ac5537be04f17243c5e929aceb3560b.
Revert "Update Compress-Raw_Zlib to CPAN version 2.049"
This reverts commit
8b782679b6ea1247bd559efc311e7366582feaad.
Revert "Update IO-Compress to CPAN version 2.049"
This reverts commit
785e05e1c22f7d7c6ed38839cc28cc6d13a286de.
Karl Williamson [Sun, 19 Feb 2012 15:57:29 +0000 (08:57 -0700)]
epigraphs.pod: Rmv illegal spaces after | in L<>
Karl Williamson [Sun, 19 Feb 2012 15:56:07 +0000 (08:56 -0700)]
perlintro: Add missing closing ">"
Ricardo Signes [Sun, 19 Feb 2012 14:55:02 +0000 (09:55 -0500)]
...and add unadded file
Ricardo Signes [Sun, 19 Feb 2012 14:29:52 +0000 (09:29 -0500)]
remove MANIFEST entries for pruned files
Max Maischein [Sun, 19 Feb 2012 13:09:55 +0000 (14:09 +0100)]
Refactor the Tk event loop to allow for other event loops
Term::ReadLine supports any event loop, including unpubished ones and
simple IO::Select loops without the need to rewrite existing code for
any particular framework.
Nicholas Clark [Sun, 19 Feb 2012 08:38:52 +0000 (09:38 +0100)]
In S_validate_suid(), move declarations after the first statement.
Otherwise exacting C89 compilers refuse to build the code. And at least
one Win32 compiler is exacting on this.
Nicholas Clark [Sat, 18 Feb 2012 14:01:32 +0000 (15:01 +0100)]
Refactor code in Data::Dumper to assume that >=5.8.0 is the common case.
The 5.8.0 versions of &init_refaddr_format and &format_refaddr are now
compiled by default, and then redefined with their 5.6.x replacements if
it turns out that we're on 5.6.1 or 5.6.2.
Previously both versions were compiled, and the correct pair bound using
typeglob assignment.
Craig A. Berry [Sun, 19 Feb 2012 05:06:16 +0000 (23:06 -0600)]
Fix typo in
985213f2fede57.
Which broke the build on both VMS and Win32.
Ævar Arnfjörð Bjarmason [Sun, 12 Feb 2012 18:56:35 +0000 (18:56 +0000)]
Remove gete?[ug]id caching
Currently we cache the UID/GID and effective UID/GID similarly to how
we used to cache getpid() before v5.14.0-251-g0e21945. Remove this
magical behavior in favor of always calling getpid(), getgid()
etc. This resolves RT #96208.
A minimal testcase for this is the following by Leon Timmermans
attached to RT #96208:
eval { require 'syscall.ph'; 1 } or eval { require 'sys/syscall.ph'; 1 } or die $@;
if (syscall(&SYS_setuid, $ARGV[0] + 0 || 1000) >= 0 or die "$!") {
printf "\$< = %d, getuid = %d\n", $<, syscall(&SYS_getuid);
}
I.e. if we call the sete?[ug]id() functions unbeknownst to perl the
$<, $>, $( and $) variables won't be updated. This results in the same
sort of issues we had with $$ before v5.14.0-251-g0e21945, and
getppid() before my v5.15.7-407-gd7c042c patch.
I'm completely eliminating the PL_egid, PL_euid, PL_gid and PL_uid
variables as part of this patch, this will break some CPAN modules,
but it'll be really easy before the v5.16.0 final to reinstate
them. I'd like to remove them to see what breaks, and how easy it is
to fix it.
These variables are not part of the public API, and the modules using
them could either use the Perl_gete?[ug]id() functions or are working
around the bug I'm fixing with this commit.
The new PL_delaymagic_(egid|euid|gid|uid) variables I'm adding are
*only* intended to be used internally in the interpreter to facilitate
the delaymagic in Perl_pp_sassign. There's probably some way not to
export these to programs that embed perl, but I haven't found out how
to do that.
Nicholas Clark [Wed, 15 Feb 2012 16:26:53 +0000 (17:26 +0100)]
Add t/porting/pending-author.t, fixing a limitation of t/porting/authors.t
t/porting/pending-author.t attempts to avoid the problem of C<make test>
passing 100%, but the subsequent git commit causing F<t/porting/authors.t>
to fail, because it uses a "new" e-mail address.
This test is only run if one is building inside a git checkout, B<and> one
has made local changes. Otherwise it's skipped.
Nicholas Clark [Wed, 15 Feb 2012 15:11:42 +0000 (16:11 +0100)]
t/porting/authors.t only needs to pass the Author to checkAUTHORS.pl
Commit
3ea0c581844689ab didn't go far enough in pruning the input.
When Porting/checkAUTHORS.pl is invoked with --tap it uses
parse_commits_from_stdin_authors() instead of parse_commits_from_stdin(),
which only looks for lines matching /^Author:/
This reduces runtime by a further 8%.
Nicholas Clark [Sat, 18 Feb 2012 22:10:05 +0000 (23:10 +0100)]
Correct the location of t/porting/perlfunc.t in perldelta.pod
It was somewhere completely wrong, thanks to no human double checking of
a git rebase during development.
Craig A. Berry [Sat, 18 Feb 2012 21:10:19 +0000 (15:10 -0600)]
Longer filenames for while_readdir.t.
With filenames varying between only one and five characters in
length, it was fairly easy to get two files differing only in
case ('A' and 'a', for example). Which on non-case-sensitive file
systems could generate a warning at unlink time because it would
unlink 'a' and then check for the existence of 'a' and get true
because 'A' was still there and indistinguishable from 'a'.
So just use longer filenames to make the possibility of collision
vanishingly small.
Father Chrysostomos [Sat, 18 Feb 2012 19:16:31 +0000 (11:16 -0800)]
perlvar: $] is not deprecated
See <
20011022034838.B1676@blackrider> and commit
0c8d858bc.
Chris 'BinGOs' Williams [Sat, 18 Feb 2012 16:24:52 +0000 (16:24 +0000)]
Update IO-Compress to CPAN version 2.049
[DELTA]
2.049 18 February 2012
* IO::Compress::Zip
Error in t/cz-03zlib-v1.t that caused warnings with 5.15
[RT# 110736: warnings from cpan/IO-Compress/t/cz-03zlib-v1.t]
Chris 'BinGOs' Williams [Sat, 18 Feb 2012 16:22:43 +0000 (16:22 +0000)]
Update Compress-Raw_Zlib to CPAN version 2.049
[DELTA]
2.049 30 January 2012
* Include zlib 1.2.6 source.
Chris 'BinGOs' Williams [Sat, 18 Feb 2012 16:18:48 +0000 (16:18 +0000)]
Update Compress-Raw-Bzip2 to CPAN version 2.049
[DELTA]
2.049 18 February 2012
* No Changes
Chris 'BinGOs' Williams [Sat, 18 Feb 2012 15:50:37 +0000 (15:50 +0000)]
Update perlfaq to CPAN version 5.0150039
[DELTA]
5.0150039 Sat 18 Feb 2012 15:33:17 +0100
* Rewrite "How do I create a module?" (ranguard)
* Remove more old questions (ranguard)
* Improve Email::MIME example (madsen)
* Fix to regex (RT #74215)
Ricardo Signes [Thu, 16 Feb 2012 23:45:46 +0000 (18:45 -0500)]
populate the Future Deprecations section
These items aggregate the bullets put forth on the perl5-porters
list, omitting those to which there was an immediate objection.
Nicholas Clark [Sat, 18 Feb 2012 12:23:26 +0000 (13:23 +0100)]
Merge the Pod::Functions refactoring to blead.
Nicholas Clark [Fri, 17 Feb 2012 20:36:31 +0000 (21:36 +0100)]
Describe the update to Pod::Functions in perldelta.
Nicholas Clark [Tue, 14 Feb 2012 12:41:07 +0000 (13:41 +0100)]
Add metadata to perlfunc.pod for which version or feature added a function.
Whilst this isn't used (yet), it makes sense to put it in now, before any
external program decides that parsing the private Pod::Function blocks in
perlfunc.pod is fair game, and then becomes dependent on their format.
Nicholas Clark [Tue, 14 Feb 2012 12:34:56 +0000 (13:34 +0100)]
Test that Pod::Function's descriptions are stylistically consistent.
Function descriptions should start with a lowercase letter, or the proper
noun SysV.
Nicholas Clark [Mon, 30 Jan 2012 19:05:33 +0000 (20:05 +0100)]
Test that functions in the groups in perlfunc.pod are in sorted order.
This ensures consistency in the documentation, and that the functions in
%Pod::Functions::Kinds are in sorted order.
Nicholas Clark [Mon, 30 Jan 2012 17:41:54 +0000 (18:41 +0100)]
Test that every function in perlfunc.pod has a summary for Pod::Functions.
Historically Pod::Functions has failed to get updated when functions are
added. This should solve that.
Nicholas Clark [Mon, 30 Jan 2012 13:47:13 +0000 (14:47 +0100)]
Add t/porting/perlfunc.t so that porting tests catch problems with perlfunc
Pod::Functions is now generated from pod/perlfunc.pod by
ext/Pod-Functions/Functions_pm.PL
If it can't parse pod/perlfunc.pod, it will abort, which will cause the
build to break. It's really not possible for it to carry on, hence aborting
is the only option. However, innocent-seeming changes to documentation
shouldn't break the build, and we expect everyone to run (at least)
the porting tests, hence this test, to catch such problems before it's too
late. To avoid duplicating the parsing logic, we make Functions_pm.PL take
a --tap option, to test that all is well.
Nicholas Clark [Mon, 30 Jan 2012 11:27:40 +0000 (12:27 +0100)]
In the generated Pod::Functions, store all function data with tab separation.
Previously the data was stored with a comma separated list inside a tab
separated list.
Nicholas Clark [Mon, 30 Jan 2012 11:15:36 +0000 (12:15 +0100)]
Generate Pod::Functions from perlfunc.pod
This avoids it getting out of synchronisation.
Nicholas Clark [Mon, 30 Jan 2012 11:01:08 +0000 (12:01 +0100)]
Add a dependency for ext/Pod-Functions/Functions_pm.PL on pod/perlfunc.pod
This ensures that it gets re-run, and lib/Pod/Functions.pm rebuilt, if
perlfunc.pod is changed.
Nicholas Clark [Mon, 30 Jan 2012 10:41:13 +0000 (11:41 +0100)]
Ensure Pod::Simple is built before Pod::Functions.
This will allow Pod::Functions to use Pod::Simple as part of its build
process.
Nicholas Clark [Sun, 29 Jan 2012 17:21:27 +0000 (17:21 +0000)]
Change ext/Pod-Functions to generate Functions.pm from a Perl script.
For the first step, simply output the current file in one hit.
Nicholas Clark [Sun, 29 Jan 2012 14:56:01 +0000 (14:56 +0000)]
Add annotations to perlfunc.pod for Pod::Functions.
In the list for "Perl Functions by Category" each item is annotated with its
type as used internally by Pod::Functions.
In the "Alphabetical Listing of Perl Functions" each function is annotated
with the summary description returned by %Pod::Functions::Flavor.
Nicholas Clark [Sun, 29 Jan 2012 13:00:41 +0000 (13:00 +0000)]
Move Pod::Functions from lib/ to ext/
Nicholas Clark [Sun, 29 Jan 2012 12:23:31 +0000 (12:23 +0000)]
Bring the joy of strict and warnings to Functions.t
This reveals that use_ok() was not in a BEGIN block, and in turn that the
test count needs to be declared before this BEGIN block runs. Now fixed.
Nicholas Clark [Sun, 29 Jan 2012 12:06:05 +0000 (12:06 +0000)]
Terser code in Pod::Functions to generate $Type_Description and @Type_Order.
Nicholas Clark [Mon, 23 Jan 2012 22:38:43 +0000 (23:38 +0100)]
Teach Pod::Functions that each, keys and values also operate on arrays.
These were added to the section 'Functions for real @ARRAYs' in perlfunc.pod
by commit
a5ce339cb0c533c9 in Sep 2010.
As ever, tweak the golden results in the test to match these changes.
Nicholas Clark [Mon, 23 Jan 2012 22:16:56 +0000 (23:16 +0100)]
Add all missing functions to Pod::Functions.
evalbytes was added to perlfunc.pod by commit
7289c5e6ca773d7c in Nov 2011.
fc was added to perlfunc.pod by commit
628253b8ba8b9cbe in Jan 2012.
say was added by commit
0d863452f5cac863 in Dec 2005.
state was added.pod by commit
36fb85f3330d45ee in Jul 2006.
__FILE__, __LINE__ and __PACKAGE__ were added by commit
cfa52385fa426b5e in
Aug 2011, and __SUB__ by commit
84ed01088568ffe9 in Nov 2011.
Again, tweak the golden results in the test to match these changes.
Nicholas Clark [Mon, 23 Jan 2012 21:44:52 +0000 (22:44 +0100)]
Teach Pod::Functions about 'Keywords related to the switch feature'.
Commit
0d863452f5cac863 in Dec 2005 added the switch feature, along with
documentation in perlfunc.pod, but did not update Pod::Functions.
Again, tweak the golden results in the test to match these changes.
Nicholas Clark [Mon, 23 Jan 2012 20:47:16 +0000 (21:47 +0100)]
Update Pod::Functions with changes from perlfunc.pod
Updated description of Binary from commit
5dac7880bdc47787 in Feb 2011.
Updated description of Flow from commit
cf2649810f00335b in Jul 2005, and
added the "the" which has always been missing from Pod::Function's version.
Updated description of Modules from commit
3b10bc60979cfe9a in Jan 2010.
Updated description of Objects from commit
353c650532037e40 in Oct 2007.
The description of Namespaces had always differed from that in perlfunc.pod.
Remove stray tabs from the descriptions of gets and sprintf.
Commit
19799a22062ef658 (May 1999) added lock to perlfunc.pod without a
it is the only function in "Threads", move it to "Misc", instead of creating
a category just for it.
use always had two entries with different descriptions in the __DATA__
section. This isn't actually sensible, as the code that builds the exported
data structures ends up taking Types from both, and using the last
description that it sees. So merge the two together to reflect this.
Drop the CHANGES section from the Pod, which is both incomplete and
redundant, given that version control does this job much better.
Tweak the golden results in the test to match these changes.
Nicholas Clark [Mon, 23 Jan 2012 16:51:48 +0000 (17:51 +0100)]
In perlfunc.pod, sort the descriptions more consistently.
The order is neither strictly lexical, nor strictly dictionary, but now for
all cases __FOO__ follows foo, and qx// is after qw//.
In the "Regular expressions and pattern matching" summary, put qr// in the
correct position alphabetically.
Nicholas Clark [Mon, 23 Jan 2012 16:32:17 +0000 (17:32 +0100)]
Add 4 functions missing from perlfunc.pod's 'Perl Functions by Category'.
C<readline> added to 'Input and output functions'
Documentation on readline was added by commit
8490252049bf42d3 in Aug 1997.
That commit also added readpipe without adding it to a category group, an
omission that was fixed by commit
4319b00c03e6a517 in Nov 2011.
C<sysseek> added to 'Functions for fixed-length data or records'
Documentation on sysseek was added by commit
137443ea0a858c43 in Apr 1997,
which implemented sysseek.
C<prototype> added to both 'Keywords related to the control flow of your Perl
program' and 'Miscellaneous functions', as Pod::Functions places it in both.
Commit
da0045b73af6f504 in Jul 1996, included 'Add documentation for new
"prototype" operator', but only added the main documentation, not an entry
in a category group.
C<lock> added to 'Miscellaneous functions'.
Documentation on lock was added by commit
19799a22062ef658 in May 1999.
Zefram [Sat, 18 Feb 2012 11:31:48 +0000 (11:31 +0000)]
correctly clone eval context frames
When cloning stacks (only used for Win32 fork emulation, not for ordinary
threads), the CV referenced by an eval context frame wasn't being cloned.
This led to crashes when Win32 forked inside an eval [perl #109718].
Dominic Hargreaves [Fri, 3 Feb 2012 19:40:30 +0000 (19:40 +0000)]
Revert "Increase the fallback value of MAXPATHLEN"
This reverts commit
ffa23acf6bf9670bd1d5fdc9a958c918b6cf3d06.
This change was made without realisation that the fallback value for
MAXPATHLEN on POSIX systems comes (in perl.h) from _POSIX_MAX_PATH,
so the fallback value here was not used.
Dominic Hargreaves [Fri, 3 Feb 2012 19:35:36 +0000 (19:35 +0000)]
[perl #109262] Don't use _POSIX_PATH_MAX as a fallback PATH_MAX
_POSIX_PATH_MAX is required to be defined for POSIX systems as
256, which is too small for a fallback PATH_MAX (some systems, such as
GNU/Hurd, do not have an inherent limit on path length).
Reini Urban [Sat, 18 Feb 2012 05:09:23 +0000 (23:09 -0600)]
Filter-1.40 released to CPAN
Also clarify cpan as upstream
Craig A. Berry [Sat, 18 Feb 2012 03:31:26 +0000 (21:31 -0600)]
Bump Cwd, File::Spec version.
Craig A. Berry [Fri, 17 Feb 2012 23:31:59 +0000 (17:31 -0600)]
Rethink EFS in File::Spec::VMS.
ae5a807c7dcf made extensive changes to File::Spec on VMS, nominally
to support Extended Filename Syntax (EFS). The idea behind the
changes was that with EFS in effect, the File::Spec functions can
guess at whether incoming file specifications are in Unix format
or VMS format and provide output in the same format as the input.
This principle is in some ways desireable in that round-trip
conversions do have the potential to lose information and it
sounds like a nice bit of DWIMmery. However, in practice it leads
to its being a crap shoot what format you'll get back, and in some
really important cases (such as MakeMaker's assembly of paths for
external build utilities) you really have to know which syntax you
have and you may really have to have native syntax, which has
always been the default.
It's also impossible to guess in some of the more common use cases,
such as a splitdir followed by catdir, where catdir has no way of
knowing what delimiters were discarded by splitdir, and thus no
hints about what syntax the original path was in.
Plus there were numerous problems with the implementation, which
broke the build when EFS was in effect and didn't pass very many
of its own tests (which it had inadvertently disabled, since fixed
in
06ecd9c7d6f).
Plus EFS isn't primarily about Unix syntax versus VMS syntax but
about allowing additional characters, such as multiple dots in
filenames or any dots in directory names. The only real difference
for the File::Spec functions should be that they now need to avoid
splitting on traditional delimiters if those characters are escaped
with the caret (^) character.
So revert most of
ae5a807c7dcf, and sprinkle negative look-behind
assertions liberally so that we correctly recognize when
traditional delimiters have been escaped and are not being used
as delimiters.
The partial support for Unix reporting mode (where we explicitly
request that output file specifications are in Unix format
regardless of input format) is left in place. It's somewhat less
partial than it was, but still incomplete.
Florian Ragwitz [Fri, 17 Feb 2012 20:15:17 +0000 (21:15 +0100)]
There is no G_LIST, only G_ARRAY
Zefram [Fri, 17 Feb 2012 21:24:13 +0000 (21:24 +0000)]
prune faulty dead logic in pp_flock
pp_flock had code to implicitly use the "last read" I/O handle if invoked
with no arguments. Actually both of its arguments are mandatory, so this
could never be reached. It was presumably cargo-culted from another op
such as pp_tell.
Father Chrysostomos [Thu, 16 Feb 2012 21:42:56 +0000 (13:42 -0800)]
make regen
Reini Urban [Wed, 8 Feb 2012 19:22:11 +0000 (13:22 -0600)]
fix -Uuserelocatableinc
-Uuserelocatableinc resulted in PERL_RELOCATABLE_INC defined as "undef" which is defined
Nicholas Clark [Fri, 17 Feb 2012 17:36:55 +0000 (18:36 +0100)]
bisect.pl now reports a meaningful error for certain "can't start" scenarios.
If --end is specified but not --start, then don't probe for start revisions
that are more recent than the --end. Previously bisect.pl would test all
stable revisions for a start point, and if it happened to find a (seemingly)
valid start point, it would continue on to run git bisect, which would fail
with the error:
Some good revs are not ancestor of the bad rev.
git bisect cannot work properly in this case.
Maybe you mistake good and bad revs?
which doesn't make clear what the problem actually is. Now the error is:
Can't find a suitable start revision to default to.
Tried perl-5.002 perl-5.003 perl-5.004 perl-5.005 perl-5.6.0 at ...
Craig A. Berry [Fri, 17 Feb 2012 18:18:58 +0000 (12:18 -0600)]
VMS does have sockaddr_in6 and ipv6_mreq.
Or at least has had for 10 years or more. If someone shows up
wanting to build on an older system, we can add real probes.
Plus we can't write to config.sh before it's been opened, as
18126d98dd7 had us doing.
Nicholas Clark [Fri, 17 Feb 2012 15:13:05 +0000 (16:13 +0100)]
For make test, move the 'porting' tests before 'lib', 'ext', 'dist' & 'cpan'
This makes the order more consistent with test_harness, and moves the
"interesting" tests earlier. "interesting", in that these are more likely
to spot unexpected problems with the tested changes.
Nicholas Clark [Thu, 16 Feb 2012 22:20:53 +0000 (23:20 +0100)]
In Perl_sv_del_backref(), don't panic if tsv is already freed.
During global destruction it's possible for tsv, the target of this weak
reference, to already be freed. This isn't a bug, and hence the interpreter
should not panic.
Chris 'BinGOs' Williams [Thu, 16 Feb 2012 16:28:35 +0000 (16:28 +0000)]
Update Socket to CPAN version 1.98
[DELTA]
1.98 CHANGES:
* Detect presence of sockaddr_in6 and ipv6_mreq; conditionally build
pack/unpack functions on this
* Back-compatibility improvements for older perls, back as far as
5.6.1 (thanks Zefram)
* Fix for picky compilers or platforms on which size_t doesn't
printf() correctly by %d
* Suppress some harmless compile-time warnings about unused variables
Zefram [Fri, 17 Feb 2012 09:24:17 +0000 (09:24 +0000)]
Carp-1.25 released to CPAN
Zefram [Fri, 17 Feb 2012 09:20:49 +0000 (09:20 +0000)]
in Carp, cope with "_" in $warnings::VERSION
The check for very old versions of warnings was generating a numeric
warning on some versions of warnings where $warnings::VERSION isn't a
proper number. Oh, the irony. Specifically affected the version of
warnings that was bundled with Perl 5.8.9.
H.Merijn Brand [Thu, 16 Feb 2012 16:17:18 +0000 (17:17 +0100)]
Add new probes for IPv6 (LeoNerd)
Ricardo Signes [Thu, 16 Feb 2012 03:49:36 +0000 (22:49 -0500)]
begin filling the 5.16.0 delta from 5.15.7
This is largely a copy and paste job. Once I copy and paste most
things in, I will then start condensing them.
This does *not* include the following sections from perl5157delta:
* module updates
* configuration and compilation changes
Ricardo Signes [Thu, 16 Feb 2012 03:39:47 +0000 (22:39 -0500)]
begin filling the 5.16.0 delta from 5.15.6
This is largely a copy and paste job. Once I copy and paste most
things in, I will then start condensing them.
This does *not* include the following sections from perl5156delta:
* module updates
* internals changes
Ricardo Signes [Thu, 16 Feb 2012 03:20:31 +0000 (22:20 -0500)]
begin filling the 5.16.0 delta from 5.15.5
This is largely a copy and paste job. Once I copy and paste most
things in, I will then start condensing them.
This does *not* include the following sections from perl5155delta:
* module updates
* configuration and compilation changes
* internals changes
Ricardo Signes [Thu, 16 Feb 2012 03:10:09 +0000 (22:10 -0500)]
begin filling the 5.16.0 delta from 5.15.4
This is largely a copy and paste job. Once I copy and paste most
things in, I will then start condensing them.
This does *not* include the following sections from perl5154delta:
* module updates
* configuration and compilation changes
* internals changes
Ricardo Signes [Thu, 16 Feb 2012 02:55:56 +0000 (21:55 -0500)]
begin filling the 5.16.0 delta from 5.15.3
This is largely a copy and paste job. Once I copy and paste most
things in, I will then start condensing them.
This does *not* include the following sections from perl5153delta:
* module updates
Craig A. Berry [Thu, 16 Feb 2012 02:39:16 +0000 (20:39 -0600)]
Simplify VMS symlink override.
The implementation assumed that the CRTL's requirement to specify
the target name in Unix syntax is somehow related to EFS (Extended
Filename Syntax). It isn't, so remove that assumption and simplify
the implementation. Bug introduced by
4148925f.