Update emacs-snapshot to 24.0.50.20100903.

Changes:
- Use SMIE's blink-paren for octave-mode.
- Provide blink-matching support to SMIE.
- Use `declare' in defmacros.

(minskim)

Restrict Python versions to match py-gtk2

(joerg)

Sync with filesystem.

(joerg)

Fix pattern to match reality.

(joerg)

-add workaround for bind_textdomain_codeset()-problem like in python25
-disable build of *audiodev and spwd to avoid complexity -- if needed
these can be added as seperate pkgs
ride on recent update

(drochner)

Updated www/nginx to 0.8.50

(joerg)

nginx 0.8.50:
    *) Feature: the "secure_link", "secure_link_md5", and
      "secure_link_expires" directives of the ngx_http_secure_link_module.

    *) Feature: the -q switch.
      Thanks to Gena Makhomed.

    *) Bugfix: worker processes may got caught in an endless loop during
      reconfiguration, if a caching was used; the bug had appeared in
      0.8.48.

    *) Bugfix: in the "gzip_disable" directive.
      Thanks to Derrick Petzold.

    *) Feature: the "image_filter_jpeg_quality" directive supports
      variables.

    *) Bugfix: a segmentation fault might occur in a worker process, if the
      $geoip_region_name variables was used; the bug had appeared in
      0.8.48.

    *) Bugfix: errors intercepted by error_page were cached only for next
      request; the bug had appeared in 0.8.48.

    *) Change: now the "server_name" directive default value is an empty
      name "".
      Thanks to Gena Makhomed.

    *) Change: now the "server_name_in_redirect" directive default value is
      "off".

    *) Feature: the $geoip_dma_code, $geoip_area_code, and
      $geoip_region_name variables.
      Thanks to Christine McGonagle.

    *) Bugfix: the "proxy_pass", "fastcgi_pass", "uwsgi_pass", and
      "scgi_pass" directives were not inherited inside "limit_except"
      blocks.

    *) Bugfix: the "proxy_cache_min_uses", "fastcgi_cache_min_uses"
      "uwsgi_cache_min_uses", and "scgi_cache_min_uses" directives did not
      work; the bug had appeared in 0.8.46.

    *) Bugfix: the "fastcgi_split_path_info" directive used incorrectly
      captures, if only parts of an URI were captured.
      Thanks to Yuriy Taraday and Frank Enderle.

    *) Bugfix: the "rewrite" directive did not escape a ";" character
      during copying from URI to query string.
      Thanks to Daisuke Murase.

    *) Bugfix: the ngx_http_image_filter_module closed a connection, if an
      image was larger than "image_filter_buffer" size.

    *) Bugfix: $request_time variable had invalid values for subrequests.

    *) Bugfix: errors intercepted by error_page could not be cached.

    *) Bugfix: a cache manager process may got caught in an endless loop,
      if max_size parameter was used; the bug had appeared in 0.8.46.

    *) Change: now the "proxy_no_cache", "fastcgi_no_cache",
      "uwsgi_no_cache", and "scgi_no_cache" directives affect on a cached
      response saving only.

    *) Feature: the "proxy_cache_bypass", "fastcgi_cache_bypass",
      "uwsgi_cache_bypass", and "scgi_cache_bypass" directives.

    *) Bugfix: nginx did not free memory in cache keys zones if there was
      an error during working with backend: the memory was freed only
      after inactivity time or on memory low condition.

    *) Feature: ngx_http_xslt_filter improvements.
      Thanks to Laurence Rowe.

    *) Bugfix: SSI response might be truncated after include with
      wait="yes"; the bug had appeared in 0.7.25.
      Thanks to Maxim Dounin.

    *) Bugfix: the "listen" directive did not support the "setfib=0"
      parameter.

    *) Change: now nginx does not cache by default backend responses, if
      they have a "Set-Cookie" header line.

    *) Feature: the "listen" directive supports the "setfib" parameter.
      Thanks to Andrew Filonov.

    *) Bugfix: the "sub_filter" directive might change character case on
      partial match.

    *) Bugfix: compatibility with HP/UX.

    *) Bugfix: compatibility with AIX xcl_r compiler.

    *) Bugfix: nginx treated large SSLv2 packets as plain requests.
      Thanks to Miroslaw Jaworski.

    *) Feature: large geo ranges base loading speed-up.

    *) Bugfix: an error_page redirection to "location /zero {return 204;}"
      without changing status code kept the error body; the bug had
      appeared in 0.8.42.

    *) Bugfix: nginx might close IPv6 listen socket during
      reconfiguration.
      Thanks to Maxim Dounin.

    *) Bugfix: the $uid_set variable may be used at any request processing
      stage.

    *) Change: now nginx tests locations given by regular expressions, if
      request was matched exactly by a location given by a prefix string.
      The previous behavior has been introduced in 0.7.1.

    *) Feature: the ngx_http_scgi_module.
      Thanks to Manlio Perillo.

    *) Feature: a text answer may be added to a "return" directive.

(joerg)

ruby-sequel-core is renamed (or merged) to ruby-sequel.

(taca)

+webmin-1.520,xf86-video-geode-2.11.9.

(taca)

Note update of www/contao29-translations package to 20100905.

(taca)

Update contao29-translations package to 20100905.

Add Kurdish language files.

(taca)

Updated filesystems/perfuse to 20100906nb2

(wiz)

Fix path to perfused. Bump PKGREVISION.

(wiz)

Add patch from KDE SVN to fix build with latest sip; patch found by
drochner. Make pkglint happier while here.

(wiz)

Note update of converters/jcode-perl package to 2.13nb3 and
mail/im package to 150nb2.

(taca)

Stop warning with perl 5.12.

Bump PKGREVISION.

(taca)

Stop warning with perl 5.12.

Bump PKGREVISION.

(taca)

Remove obsolete MESSAGE, bump PKGREVISION.

(wiz)

Make pkglint happier.

(wiz)

Updated filesystems/perfuse to 20100906nb1

(wiz)

Add symlinks for libperfuse. Simplify package. Bump PKGREVISION.

(wiz)

Updated www/p5-POE-Filter-HTTP-Parser to 1.04

(wiz)

Update to 1.04:

1.04  Wed Jul 28 12:59:47 BST 2010
  - Apply a patch from Dariusz Jackowski [RT#59772] which fixes
    double encoded utf-8 strings.

(wiz)

Updated net/p5-POE-Component-Client-Ident to 1.16

(wiz)

Update to 1.16, set LICENSE.

Changes undocumented.

(wiz)

Updated net/p5-POE-Component-Client-DNS to 1.051

(wiz)

Update to 1.051, set LICENSE .

================================
2009-10-14 03:06:11 -0400 v1_051
================================

    Prepare for release.

    Add a 00_info.t test to help debug platform and dependency based
    issues.

    Administrivia to convert the repository to git.

================================
2009-08-28 06:46:24 +0000 v1_050
================================

    Jonathan Yu found a warning while packaging this distro for Debian.
    I've applied his patch from rt.cpan.org 48335, and I took his
    suggestion to start using Test::NoWarnings.

    Philip Gwyn found a strange case where IPv6 localhost was returned as
    ::1 rather than the long form. Added a check for both forms.

    Bumped the version to 1.050 for release.

    Fix a POD error (line too long, broken link to RT) reported by
    Jonathan Yu, ironically in rt.cpan.org ticket 48336. :)

    Fix a warning for /etc/hosts files that contain blank lines or
    comments. Suggested by Leonid Rashkovsky in e-mail.

===============================
2009-07-28 06:01:07 +0000 v1_04
===============================

    Use latest POE, and prepare for a new release.

    Added a machine-readable repository directory to the distribution.
    Documented the bug tracker, repository, and other resource URLs.

    Resolve (heh) rt.cpan.org ticket #13899. AAAA requests will check for
    IPv6 addresses in /etc/hosts or your favorite operating system's
    equivalent.

    Resolve rt.cpan.org ticket #13492 by S辿bastien Aperghis-Tramoni,
    #14723 by Branislav Gerzo, and umbrella ticket #15009 by cycling
    through the list of resolvers known to Net::DNS if the top one times
    out.

    Applied a patch by Ren辿 Mayorga, resolving rt.cpan.org ticket 41313.
    Ren辿's patch uses Net::DNS to find poe.perl.org's address and use
    that instead of hardcoded values. Thanks also go to Ren辿 for
    submitting the patch through rt.cpan.org. I may never have discovered
    ticket 506915 on bugs.debian.org.

===============================
2009-02-18 04:48:06 +0000 v1_03
===============================

    Fix warnings uncovered by POE's recent global-warning issue.

    Make it safe to have more than one PoCoDNS object.

    Somehow this bug has been no problem for people this entire time. An
    amazing trick really :D

===============================
2009-01-13 20:08:27 +0000 v1_02
===============================

    A host-file test was failing because its address segments were wider
    than single octets.

===============================
2008-10-29 16:18:54 +0000 v1_01
===============================

    New version.

    Accommodate DNS change for poe.perl.org in the tests.

    Apply Martin Ferrari's patch to skip network tests if there's no
    detectable network. Resolves rt.cpan.org ticket 33677.

(wiz)

Updated devel/p5-POE-Component-Cron to 0.021

(wiz)

Update to 0.21, add dependency, set LICENSE.
Same number of test failures as before:
Failed 1/1 test programs. 0/2 subtests failed.

Changes to 0.19 not documented.

(wiz)

Updated devel/p5-POE-Component-Pluggable to 1.26

(wiz)

Update to 1.26, set LICENSE.

1.26  Fri Jun 25 18:35:15 BST 2010
    - Document the 'debug' parameter
    - Croak if the user does not supply the 'types' parameter
    - Fix problems with SYNOPSIS and show example plugin usage
    - Don't warn inside plugin_del
    - Don't complain about invalid eat constants when debug mode is off
    - Add irc_plugin_error event

    ( All the above by Hinrik ).

(wiz)

Updated devel/p5-POE-Component-Schedule to 0.95

(wiz)

Update to 0.95, update dependency patterns, use Module::Build to build.

Changes:

0.95    2010-05-25T22:36+02:00    DOLMEN (Olivier Mengu辿)
        No changes since 0.94_03.
        Only change since 0.94 is fix for RT#56447 (allow POE aliases as target
          session ID for events).

0.94_03 2010-05-25T00:14+02:00    DOLMEN (Olivier Mengu辿)
        No functional change since 0.94_01.
        Fixed constant functions: added "()" prototype to make them inlinable
          by perl. Should have been like that since 0.91_01. That should
          improve performance.
        Added .perlcriticrc to disable Perl::Critic policies that do not match
          our style (filled RT#57638, RT#57639, RT#57640, RT#57807 for issues
          with default policies).
        Made some suggestions from Perl::Critic. Reduced number of lines, so
          improved test coverage ratio.
        Test coverage: 94.5%.
        Author build system: fixed the 'Build tag'.
        Author build system: improved steps to publish 'Changes.rss' in the
          Google Code wiki area.

0.94_02 2010-05-18T01:02+02:00    DOLMEN (Olivier Mengu辿)
        Author build system: Changes.{rss,yml} generation is now automated by
          extending Module::Build (see 'inc/MY/Build.pm' and
          'inc/MY/Build/Changes.pm').
        Added 't/90-critic.t' to 'MANIFEST.SKIP' as it is the main cause of
          CPAN Testers failures.
        Author build system: Drop 'Makefile.PL' generation by Module::Build.
        Author build system: Transformed 'make-svn-tag.pl' into a Build target
          ('tag').

0.94_01 2010-05-09T19:44+02:00    DOLMEN (Olivier Mengu辿)
        RT#56447 fixed, but I expect to unapply the fix once RT#57314
          (allow aliases in $poe_kernel->refcount_{in,de}crement) is
          implemented. Test cases added.
        Bug references are now linked in 'Changes.rss'.
        Test coverage: 91.6%.

0.94    2010-03-13T16:37+01:00    DOLMEN (Olivier Mengu辿)
        Only minor POD changes in Schedule.pm since 0.93_02.
        Increased dependencies versions.
        Small tweak to the 'Changes' file format to make it more parseable
          and be able to generate 'Changes.yml' and 'Changes.rss'.
        Added 'Changes.rss' and 'Changes.yml', generated from 'Changes'
          using either 'make-Changes-rss.pl' or 'make-Changes-rss-2.pl' (see
          SVN repo).
        Test coverage: 92.1%.

0.93_02 2010-02-13T16:30+01:00    DOLMEN (Olivier Mengu辿)
        More test coverage, but still only 92.1%. Help still welcome for
          RT#53925.
        Improved argument check for ->add() (and now, with tests too).
        POE >= 1.281 is recommended as it fixes the _child(lose) issue.
        Updated POD: removed obsolete warnings, updated copyright.
        Updated README: should be more helpful for distribution packagers.

0.93_01 2010-02-13T01:55+01:00    DOLMEN (Olivier Mengu辿)
        Applied fix for RT#53925 (Patch by Zeph Gillen).
        Added t/50-subclass.t to try to catch subclassing problems which may
          affect POE::Component::Cron. Unfortunately it does not catch
          RT#53925. Help needed.
        Added strictures in tests.
        Updated 'MANIFEST.SKIP' to match new Module::Build: ignore MYMETA.yml.

(wiz)

Updated devel/p5-POE to 1.293

(wiz)

Update to 1.293, same number of test failures as 1.280:
Failed 8/152 test programs. 6/3913 subtests failed.

==================================
2010-09-04T19:20:41.964969Z v1_293
==================================

    Bump version for release.

    Add user continuation parameters to sig() and sig_child().

    Roll back most of the lexical FIFO event arrays. Going through a
    POE::Queue for this will be a bit slower, but POE::XS::Queue::Array
    may be fast enough.

    Continue to guarantee FIFO event ordering despite the rollback to
    dumping them in a time-ordered queue. Detect minor time regressions,
    and use the last time value plus a small number when needed.

    Fixed a mistake in the alias_list() example code in POE::Kernel
    documentation, spotted by Paul G Webster

    Chris Benedict reported that Mac OS X returns ESPIPE when accept() is
    aborted, rather than ECONNABORTED everywhere else. Add ESPIPE to the
    list of harmless accept() errors.

    Chris Benedict reported that nmap will crash Wheel::SocketFactory
    based things. nmap disconnects before accept() can be called, so
    accept() returns ECONNABORTED. Technically applications can handle
    this themselves and not destroy the wheel, but the wheel maybe
    shouldn't report harmless errors.

    Apply Hinrik's patch to increase test failure verbosity.

==================================
2010-08-02T06:01:19.221768Z v1_292
==================================

    Bump revision for release.

    Resolve rt.cpan.org 59925, wherein sig_child() use would increase
    some events' latency to at 1sec. First reported by Getty, then Avar
    opened the ticket. Fixed with Hinrik's copious valuable assistance.

    Fix a method name typo in TRACE_REFCNT code.

==================================
2010-07-27T06:54:51.380682Z v1_291
==================================

    New version to spackle over failed one.

    Version bump for release.

    Nick Perez pointed out that we don't absolutely need HighEvent, since
    put() returns the high-water status much more conveniently. Resolve
    his rt.cpan.org ticket 56432. Oh, and fix an unrelated test plan.

    Clarify the refcount_{inc,dec}rement() return values per Olivier
    'dolmen' Mengu

    Remove some logic that stopped being used. Adjust idle-kernel checks
    into a slightly more optimal order.

    Finish a sentence. Incomplete docs noted by Hinrik.

    Move non-alarms out of POE::Queue::Array.

    Reduce calls to get_item_count() by one per _data_ev_dispatch_due().
    Reduce calls to get_next_priority() by approx. one per
    _data_ev_dispatch_due(). Micro-optimize get_item_count().

    Fix tests related to recent POE::Resource::Events optimizations.

    At long last, apply Ton Hospel's micro-optimization for
    get_next_priority().

    Small optimizations. Avoid _data_stat_add() calls when not needed.
    Avoid redundant session refcounts when the source and destination are
    the same. Breaks some pedantic tests, but we can fix those.

    Apply Hinrik's test changes from http://gist.github.com/449231

    Apply Hinrik's patch to catch die() and his test case for proper
    shutdown if a child coderef process dies.

    Fix a method typo at Hinrik
    stdio closing on exec() errors, just in case.

    Add TODO test for double POE::Loop loading

    Add support for POE_CATCH_EXCEPTIONS environment variable. Apply
    Justin Hunter's doc patch for POE::XS::Loop::EPoll.

    Tweak the test so it uses RELEASE_TESTING

    Fix invalid POD links

    Remove the obnoxious warning about the win32 leak in Wheel::Run

    Add proper exitval handling to POE::Wheel::Run, the previous patch
    munged it :(

    Add deprecation warning for 'Winsize' to POE::Wheel::Run->new

    Patch POE::Wheel::Run for RT#56417 and add a test+more docs

==================================
2010-04-03T20:03:35.441534Z v1_289
==================================

    Bump revision for a new release. No significant changes, other than
    to build a new dist with the latest EUMM.

==================================
2010-04-03T18:17:45.177433Z v1_288
==================================

    Fix dependencies and bump POE revision for release.

    Set svn:ignore for: META.yml CHANGES run_network_tests PoLo-Gtk
    forgot to clean run_network_tests

    Back out of my previous commit to auto-clean META.yml and CHANGES, it
    will blow up if an end-user did it on their box! /me is stupid :(

    Automatically clean META.yml and CHANGES files on "make distclean"
    because they are auto-generated

    Fix a doc typo in ReadWrite in get_output_handle()

    Fix a documentation typo in PoCo-Client-TCP

    Fix a doc typo in my previous commit :(

    Update the Wheel::ReadLine docs for a better explanation of the TERM
    issue and workarounds, thanks BiNGOs!

    Avoid undefined value warnings when trying to tab-complete
    nothingness. dreamer on irc.perl.org #poe reported the problem, with
    a useful test case.

    After a lot of hacking/digging Cygwin with a TERM=dumb does not work
    with POE::Wheel::ReadLine. This fix skips the test, and makes
    Wheel::ReadLine die if it detects this situation. Added some docs for
    our Cygwin users :)

    Document how POE::Wheel::ReadWrite reports EOF.

    Added a lot of loop-loading tests, so we can start work on the
    "naive" loop load hacks

    Tweak the @modules logic, thanks to TonyC ( it might not be perfect
    but good enough for now )

    Remove any XS modules from @modules, as usual TonyC spotted it :)

    Fix loading of XS loops via POE.pm, thanks TonyC! Better document the
    various loop loader shenanigans we have :)

==================================
2010-02-19T06:11:52.979284Z v1_287
==================================

    Bump revision for new release.

    Fix a sig_child() race condition discovered by Tom Lanyon and
    reported (with the help of Jay Hannah) in irc.perl.org #poe. I
    haven't been able to reproduce the problem in a non-MooseX::Workers
    test case, but this change resolves Tom's more complex test case and
    passes tests.

    Fix an infinite file rollover when open() succeeds but stat() fails.

    Jay Hannah discovered a bug when he passed in a filename with
    trailing whitespace. Two-parameter open() treats trailing whitespace
    as insignificant, so it passed. On the other hand, trailing
    whitespace is significant in stat(), so it always failed, triggering
    an unrelenting stream of file rollover events. Switched to
    three-parameter open(), which considers trailing whitespace as
    significant. Now both open() and stat() agree the file doesn't exist.

    Thanks to Jay and the University of Nebraska at Omaha for securing me
    some debugging time on their machines.

==================================
2010-02-15T06:36:51.553032Z v1_286
==================================

    Bump for yet another release.

    POE::Pipe and its subclasses work best using Cygwin's pipe(), not
    IO::Socket. Previous Cygwin versions work better with sockets. If
    01_sysrw.t hangs for you, upgrade your Cygwin and its Perl.

==================================
2010-02-14T21:37:14.937566Z v1_285
==================================

    Bump the revision number for a new release.

    Some code cleanup in POE::NFA.

    Revision 2803 is mistaken. Rob Hildred reported the problem and
    provided the test case that was instrumental in helping me resolve
    it. And the bug report was rt.cpan.org 53517. I mistakenly assumed
    the currently open tab was the same ticket I was debugging.

    Fix state bleed between POE::NFA instances.

    Resolve rt.cpan.org 53682 reported by Serguei Trouchelle. Thank you,
    Serguei, for a very useful test case.

    Replace the detach-during-start test with an enhanced test that
    Olivier Mengu

    Applied Andrew Feren's patch to avoid a POE::Wheel::Run memory leak
    when running under Windows. Should resolve rt.cpan.org #50068.

    Remove dependency on Test::Exception (it appears to be unused). Fix a
    spot of tab damage while I'm in there.

    Rely upon the latest POE::Test::Loops.

    Resolve rt.cpan.org #54319. In polling mode, the wheel wasn't
    iterating get_one() for the entire input from get_one_start(). As a
    result, only one log line was returned per PollInterval. This change
    iterates through get_one() until all lines are returned. Many thanks
    to Oleg A. Mamontov for opening the ticket, and to Bazerka on
    irc.perl.org #poe for supplying the basis for a test case.

    Fix a documentation bug, thanks zephg! ( reported by arcanez@irc )

    Added RELEASE_TESTING env var checks Added test for 404 links in POD
    Cleanup old stuff and fix MANIFEST.SKIP

    Correct a type reported by Hinrik
    #poe.

    Fix a cpantesters.org link that gcola reported as 404 in irc.perl.org
    #poe.

    Add more to what I've done.

    [rt.cpan.org 53757] Apply G. Allen Morris III's spelling patch but
    not his test based on Test::Spelling. That module doesn't seem to
    work for me, so I couldn't use it.

    Define a _stop handler to mollify assertions.

    [rt.cpan.org 53777] Restore _start & _stop return values. Mikko Ehto
    reported that _start and _stop don't return values as documented. He
    also provided a testcase, which was very helpful. I added his test
    case and fixed the underlying bug during the Perl Oasis hackathon.

    Correct a documentation typo reported by jussive in IRC.

    Fix a typo in a trace message, reported by "jussive" in irc.perl.org
    #poe.

==================================
2010-01-14T07:20:50.660659Z v1_284
==================================

    Bump the version for release.

    [rt.cpan.org 53519] The issue was a disagreement in the
    interpretation of "\n" on Windows (DOS newline semantics). Changed
    the newline charater to "!", because I can.

==================================
2010-01-08T00:13:11.003173Z v1_283
==================================

    Bump revision for release.

    Fix POE::Wheel::FollowTail, at least as far as Hinrik's test and the
    rest of the test suite are concerned.

    Hinrik
    code doesn't flush out the file on startup. Applied his changes to an
    existing regression test for this wheel.

==================================
2010-01-06T20:55:14.534499Z v1_282
==================================

    [rt.cpan.org 53352] Correct spelling mistakes found by Jonathan Yu.
    Simplifies Debian package maintenance since that's one less
    downstream patch to worrry about.

    Bump the version for a new release.

    POE::Wheel::ReadLine appears to hang on AIX when STDIN is set
    blocking(1). Reported (and graciously debugged) by Criveti Mihai from
    the Ministry of Economy and Finance in Romania.

    Several doc changes to fix issues reported by garu in irc.perl.org
    #poe. Thanks, garu!

    Clarify what's going on in two Session References examples. Reported
    by "garu" in irc.perl.org #poe.

==================================
2010-01-01T09:15:20.168519Z v1_281
==================================

    Bump up the version for release.

    [rt.cpan.org 24956] Set STDIN blocking. Large amounts of STDOUT data
    will overrun terminal buffers if STDIN is non-blocking. Weird!

    [rt.cpan.org 52452] Attempt to make the destruction order in the test
    more deterministic.

    [rt.cpan.org 51171] Olivier Mengu
    completely different kind of value on MSWin32. Use his fix.

    Upgrade the detachment test to use Test::More.

    [rt.cpan.org 51772] Don't fire EN_CHILD,CHILD_LOSE when
    detach_myself() is called from _start, since the parent session never
    gets EN_CHILD,CHILD_CREATE to begin with. Fixes a bug reported by
    Olivier Mengu

    Change the timing of file reset detection to match previous logic.

    Fix regression in the new file rotation code.

    Make this test line-ending neutral.

    Fix filename-based reset detection in Windows.

    Detangle filehandle vs. filename timer callbacks. Simplified each at
    the expense of some redundant code between them.

    Improve reset detection on Windows. Reduce polling when a file is
    idle.

    Fix a typo in a code example reported by Hinrik
    IRC.

    Clarify session multitasking documentation per IRC feedback from a
    person going by the nickname "skor".

    Don't bother propagating wantarray() and tracking return values if
    we're not dispatching ET_CALL.

    Eliminate some temporary variables.

    mousemask() is very not portable. Remove it.

    Update POE::Test::Loops dependency to latest CPAN release.

(wiz)

Updated devel/p5-Log-Log4perl to 1.30

(wiz)

Update to 1.30, improve COMMENT.

1.30 (2010/08/30)
  *    (ms) [RT 60665] HUP handlers are stacked on top of each other now,
            to make sure that multiple file appenders recreate multiple
            files and not just one (patch provided by Karen Etheridge).
  *    (ms) [RT 60197] Fixed uninitialized value warnings with
            the multiline appender and provided a test case (patch provided
            by Karen Etheridge)
  *    (ms) [rt.cpan.org #59617] Fixed system-wide threshold without appender
            thresholds. Bug reported by Dmitry Bigunyak.
  *    (ms) [rt.cpan.org #24884] Using require() instead of incomplete
            logic in L4p::Util::module_available(). local __DIE__
            handler takes care of user-defined __DIE__ handlers
            ignoring $^S (suggested by Eric Wilhelm and others).
  *    (ms) [rt.cpan.org #60386] Fixed init_and_watch() which
            double-bumped the caller_level and led to uninitialized
            values in the pattern layout. Thanks to Mitja Bartsch for
            the report.
  *    (ms) Applied patch by Karsten Silkenb辰umer to add an optional
            $log_dispatch_level to create_custom_level(). Updated
            documentation.

1.29 (2010/06/16)
  *    (ms) Added documentation on how to use Log4perl's :easy macros
            with Catalyst in Log::Log4perl::Catalyst.
  *    (ms) wrapper_register() now deals with caller_depth automatically.
            Backwards compatibility with old wrapper classes using
            caller_depth directly is provided. Documentation has been
            updated.
  *    (ms) Felix Antonius Wilhelm Ostmann reported Resurrector.pm
            crashes, fixed as suggested by setting the %INC value to
            the module path.
  *    (ms) Another caller_depth fix in Log::Log4perl::Catalyst.
  *    (ms) Fixed logdie() caller_depth bug reported by Rob Retter.
  *    (ms) [RT 56145] Saving errstr in DBI appender to survive ping()
  *    (ms) Added INTERNAL_DEBUG env variable to test suite triggering
            all _INTERNAL_DEBUG statements to be printed for better
            error diagnosis on misbehaving systems.

(wiz)

Updated graphics/p5-RRDTool-OO to 0.31

(wiz)

Update to 0.31:

0.31  (05/26/2010)
    (ms) [rt.cpan.org #54870] Fixed typos/pod errors reported by
        Salvatore Bonaccorso.
    (ms) Fixed overwriting input parameters in create(), bug reported
        by Erik Wasser.

(wiz)

Updated graphics/p5-Image-Size to 3.230

(wiz)

Update to 3.230/3.2.3.0:

3.230 Sunday August 29, 2010, 04:00:00 PM -0700

* imgsize
* lib/Image/Size.pm
perlcritic clean-ups from new rules.

* lib/Image/Size.pm
* t/Test_emf_small.emf (added)
* t/all.t
RT #59995: Added support for Windows Enhanced Metafile Format
(EMF).

* t/00_load.t (deleted)
* t/01_pod.t (deleted)
* t/02_pod_coverage.t (deleted)
* t/03_meta.t (deleted)
* t/04_minimumversion.t (deleted)
* t/05_critic.t (deleted)
* xt/00_load.t (added)
* xt/01_pod.t (added)
* xt/02_pod_coverage.t (added)
* xt/03_meta.t (added)
* xt/04_minimumversion.t (added)
* xt/05_critic.t (added)
Move the author/distro-sanity tests to an "xt" directory.

3.221 Sunday April 25, 2010, 02:10:00 PM -0700

* imgsize
* lib/Image/Size.pm
Large-scale code and documentation clean-up based on perlcritic
and visual review.

(wiz)

Updated graphics/p5-Image-Imlib2 to 2.03

(wiz)

Update to 2.03, set LICENSE. Same number of test failures as in 2.02:
Failed 5/7 test programs. 5/7 subtests failed.

2.03 Tue Nov 24 21:40:49 GMT 2009
    - add a new_transparent method
    - add a create_rotated_image method
    - add a human-readable license

(wiz)

Updated graphics/p5-GD to 2.45

(wiz)

Update to 2.45, same number of test failures as before:
Failed 1/2 test programs. 1/13 subtests failed.

2.45 Clarified the GD license. There is now a formal LICENSE file in the package.

(wiz)

Updated x11/p5-gtk2 to 1.222

(wiz)

Update to 1.222:

Overview of changes in Gtk2 1.222
=================================

* Make Gtk2::ListStore->remove()'s return value consistent on gtk+ 2.0
* Fix some build and test failures.

(wiz)

Updated www/p5-WWW-Curl to 4.13

(wiz)

Update to 4.13:

4.13 Wed Sep 01 2010: - Balint Szilakszi <szbalint at cpan.org>

    - Fixed WWW::Curl::Form (again, formadd and formaddfile working now).
    - Made constant constant handling more robust and added tests [Fuji, Goro].
    - Modernized *.pm and AUTOLOAD now throws an error on unknown method calls [Fuji, Goro].
    - Fixed code depending on CURLINFO_SLIST to be optional [Fuji, Goro].

(wiz)

Updated misc/p5-Locale-Maketext-Lexicon to 0.82

(wiz)

Update to 0.82:

[Changes for 0.82 - 2010-04-12]
* Locale::Maketext::Extract::Plugin::Base
* Locale::Maketext::Extract::Plugin::TT2
* Locale::Maketext::Extract
    - fixed some spelling errors - thanks to Ansgar Burchardt for the patch

[Changes for 0.81 - 2010-04-09]
* Locale::Maketext::Extract::PLugin::FormFu
    - the plugin now understands YAML of the form:
        mesage_loc: ['Max files: [_1]', 10]

      Thanks to Charlie Garrison for the patch

[Changes for 0.80 - 2010-04-08]
* Locale::Maketext::Lexicon
    - Added the _preload option, which forces parsing of the lexicon
      as soon as it is loaded. Improves memory usage and performance
      in a prefork environment

      Thanks to Ruslan Zakirov for the patch

(wiz)

Updated misc/p5-Locale-Codes to 3.13

(wiz)

Update to 3.13:

3.13  2010-06-04 sbeck
        * NEW CODE(s)

3.12  2010-04-06 sbeck
        * NEW CODE(s)
        * Renamed test.pl to testfunc.pl to avoid causing an error
          when built as part of perl.

(wiz)

Updated misc/p5-File-MMagic-XS to 0.09006

(wiz)

Update to 0.09006, set LICENSE.

0.09006 - 19 May 2010
  - Should fix errors in t/003_error.t
    http://cpantesters.org/cpan/report/07280906-b19f-3f77-b713-d32bba55d77f

0.09005 - 17 May 2010
  - Should fixe problems with t/005_clone.t
  - Minor XS tweaks

0.09004 - 11 May 2010
  - Disable MSWin32 (patches welcome)

0.09003_03 - 10 May 2010
  - Explicitly test add_file_ext
  - Add thread safety test

0.09003_02 - 26 Apr 2010
  - Ah, typemaps! I guess I didn't know about it when I first wrote this
  - Add many tests
  - Fix a memory leak under clone()
  - Fix a memory leak under fhmagic()
  - Update function declarations to a more modern style
    * move all functions out of XS section, and make use of
      automatic linkage by xsubpp
  - Add a memory leak test
    * only runnable when running under author mode
    * set TEST_MEMLEAK=1
    * ignore any leaks that does not directly show a Perl_*
      namespace.

0.09003_01 - 07 Apr 2010
  - Fix memory corruption (Martin Blapp)
  - Fix rt #28072. Changing $/ doesn't change the return value (Martin Blapp)
  - Fix how SV returned by error() is handled.
  - Add a clone() method.

(wiz)

Updated devel/p5-glib2 to 1.223

(wiz)

Update to 1.223:

Overview of changes in Glib 1.223
=================================

* Fix a few test failures.

(wiz)

Updated textproc/p5-Text-vCard to 2.09

(wiz)

Update to 2.09:

2.09 Thu 19 Aug 2010 16:08:20 BST
    - Add set_encoding() for export - Stefan

2.08 Tue Jul 13 08:47:58 BST 2010
    - Doc fix (Blair Christensen)

(wiz)

Updated textproc/p5-Text-vFile-asData to 0.07

(wiz)

Update to 0.07, please pkglint:

0.07    Wednesday 25th August, 2010
        Switch to ExtUtils::MakeMaker as Module::Build got
        uppity and started generating bad Makefile.PL files

0.06 Friday 20th August, 2010
Fix case with ' in entities (Richard Russo)

(wiz)

Updated math/p5-Set-Crontab to 1.03

(wiz)

Update to 1.03, set LICENSE.

1.03    2010-05-08

    * Return list elements in sorted order.

(wiz)

Updated multimedia/p5-GStreamer to 0.15

(wiz)

Update to 0.15, less test failures than 0.14 (2/31, 4/429 instead
of 3/31, 5/425). Set LICENSE.

2009-02-22  Torsten Schoenfeld  <kaffeetisch@gmx.de>

* lib/GStreamer.pm
* NEWS: Release 0.15.

2009-02-22  Torsten Schoenfeld  <kaffeetisch@gmx.de>

* Makefile.PL
* MANIFEST
* GStreamer.pm (removed)
* lib/GStreamer.pm (added): Move GStreamer.pm into the new lib
directory.

2009-02-22  Torsten Schoenfeld  <kaffeetisch@gmx.de>

* t/GstPad.t
* xs/GstPad.xs: Wrap gst_pad_is_blocking, gst_pad_peer_query,
gst_pad_iterate_internal_links, and
gst_pad_iterate_internal_links_default.

2009-02-22  Torsten Schoenfeld  <kaffeetisch@gmx.de>

* xs/GstIterator.xs (newSVGstIterator): Convert NULL iterators to
undef.

2009-02-21  Torsten Schoenfeld  <kaffeetisch@gmx.de>

* xs/GstFormat.xs (SvGstFormat): Throw an exception if
gst_format_get_by_nick failed to convert the string to a format.
This means that unregistered format strings now cause errors
instead of being silently turned into GST_FORMAT_UNDEFINED.

2009-02-21  Torsten Schoenfeld  <kaffeetisch@gmx.de>

* t/GstClock.t: Make the clock resolution and master clock tests
more robust.

2009-02-21  Torsten Schoenfeld  <kaffeetisch@gmx.de>

* t/GstIndex.t: Make the GstIndexEntry tests more robust and correct.

2009-02-21  Torsten Schoenfeld  <kaffeetisch@gmx.de>

* xs/GstIndex.xs: We don't own the GstIndexEntry objects returned
by the various GstIndex methods, so don't use the _own typemap
variant.

(wiz)

Updated geography/p5-Geo-Coordinates-UTM to 0.09

(wiz)

Update to 0.09, set LICENSE, fix typo in COMMENT.

0.09  Mon May 17 16:30:32
- Patched mgrs_to_utm to handle short MGRS strings

(wiz)

Updated converters/p5-MIME-Base32 to 1.02a

(wiz)

Update to 1.02a, set LICENSE.

Version 1.02 provides minor changes only:
- added COPYRIGHT according to FEDORA & DEBIAN community requests
- test.pl - extended test string (just for sure ;)

(wiz)

Updated comms/p5-Device-Gsm to 1.55

(wiz)

Update to 1.55:

1.55  Sun Jun 27 18:07:11 CEST 2010
- Fixed RT #58869, incorrect decoding of text7 messages.
  Thanks to Alexander Onokhov.

(wiz)

Updated comms/p5-Device-Modem to 1.53

(wiz)

Update to 1.53:

1.53  Thu Apr 01 13:49:00 CET  2010
    - ***CHANGED*** default log file position
      from /var/log/modem.log to /tmp/modem.log.
      Too many failed tests and user reports made me
      reconsider my poor default choice.
    - Added voice dialing. Just dialing though.
      You can't perform real voice calls through Device::Modem (yet :)
      Thanks to Marek Jaros.
    - Added ';' (voice dialing) and 'p' (pause) as valid values
      for dial() number.

1.52  Sun Mar 28 15:50:00 CET  2010
- Added automatic port reconnection in the port() method.
  This should improve connection reliability and reduce risk
  of "Can't call method XXXXXX on undefined value YYYYYY" errors.

(wiz)

Set LICENSE.

(wiz)

Updated chat/p5-POE-Component-IRC to 6.39

(wiz)

Update to 6.39:

6.39  Sat Sep 4 02:16:28 UTC 2010
    - AutoJoin.pm: Allow channel keys to be undefined
    - Make the 'nickserv' command do the right thing on ratbox ircds
    - Add a server_version() method

6.38  Fri Sep 3 18:33:50 UTC 2010
    - Only process the first CTCP chunk we find in a message. This prevents
      someone from flooding our outgoing queue by having us e.g. reply to 20
      VERSION requests at a time.
    - CTCP.pm: Reply to VERSION with "dev-git" when no version is available.

6.37  Tue Aug 17 22:53:22 GMT 2010
    - Make all warnings fatal
    - Use real temp files in tests instead of littering the dist directory

6.36  Mon Jul 26 03:53:50 GMT 2010
    - Added a logged_in() method to see if we're logged into IRC

6.35  Sun Jun 27 09:32:22 GMT 2010
    - Disconnecting.pod: Mention when it is appropriate to use
      C<< $irc->yield('shutdown') >>.
    - Connector.pm: Clear the reconnect timer when the plugin is deleted so
      that we can actually shut down the IRC component.
    - Depend on POE::Component::Pluggable 1.26 for irc_plugin_error

6.34  Fri Jun 25 18:16:40 GMT 2010
    - CTCP.pm: Do "use POE::Component::IRC;" to avoid weird failures when
      this plugin is compiled by code which hasn't done the same.

6.33  Mon Jun 21 20:27:42 GMT 2010
    - BotCommand.pm: Allow user to choose how help messages are delivered
    - BotCommand.pm: Require the command prefix in private messages
    - BotCommand.pm: Make the help messages more context-sensitive
    - BotCommand.pm: Add support for custom auth checks
    - BotCommand.pm: If Eat == 1, we eat everything that looks like a command
    - Cookbook: Add Gtk2 example by Damian Kaczmarek
    - Logger.pm: Support a hook for custom log storage
    - IRC.pm: Remove redundant version() method
    - Convert distribution over to Dist::Zilla

(wiz)

Updated audio/p5-CDDB to 1.220

(wiz)

Update to 1.220:

================================
2010-03-08 01:18:09 -0500 v1_220
================================
    Prepare for git based distribution toolchain.
    Rearrange repository for git.
    Create an OSX example program for Max Wassermann.

(wiz)

Updated audio/p5-Audio-Wav to 0.12

(wiz)

Update to 0.12:

0.12  Sun May 30 08:45:03 GMT 2010
  - don't use values > MAXINT. +2147483648 is not a valid unsigned
    int32 (it's MAXINT+1). while we're at it, put all constants in
    the XS in hex format for clarity. possibly fixes:
      http://www.cpantesters.org/cpan/report/6945385
  - fall back to pureperl when Inline::C is installed but not functional
    as reported by Michel Lanne.

0.11  Fri Mar 12 04:12:44 GMT 2010
  - fix two bugs triggered with inline on win32 - one related to assuming
    C99 support and one related to lack of stdint.h or equivalent. tested
    with strawberry perl, should work with msvc too. bug report here:
      http://www.cpantesters.org/cpan/report/6920668
    same smoker reports tests with 0.11 here:
      http://www.cpantesters.org/cpan/report/6930825

0.10  Sun Feb 14 04:09:00 GMT 2010
  - add COPYRIGHT information for debian folks
  - include xt/

0.09  Thu Feb 11 14:58:37 GMT 2010
  - Support 24 and 32 bit wav reading w/o Inline::C (from Wolfram humann)
    ( see https://rt.cpan.org/Public/Bug/Display.html?id=36452 )
  - add a META.yml and some other easy kwalitee tweaks
  - Audio::Wav::Read::_has_inline moved to Audio::Wav::_has_inline so
    it can be queried before instantiating a reader (and later be used
    internally for Audio::Wav::Write)

(wiz)

Updated audio/p5-Audio-Scan to 0.84

(wiz)

Update to 0.84:

0.84    2010-08-27 19:10:00
        - Added an optional audio_md5 info value, this is the MD5 checksum of the
          first N audio bytes of the file.  This is useful in determining if a file
          is the same even when tags have been changed.  See the documentation for details.
        - Added audio_size info value to all formats.
        - Added image offset values in AUDIO_SCAN_NO_ARTWORK mode for some formats, so
          image data may be extracted directly from the file if needed.  See the
          documentation for specific details.
        - AAC: Fixed samplerate values for AACplus files. The samplerate in the header
          needs to be doubled to obtain the actual playback samplerate.
        - ID3: Bug 16452, Don't add null strings as array entries for T* frames.

0.83    2010-06-10 14:15:00
        - APE: Tag could be ignored when ID3v1 tag is present.
        - Fixed error handling for invalid RVAD tags.
        - Win32: Added binmode for all open files and filehandles.

(wiz)

On NetBSD, pull in perfuse.

(wiz)

Remove -Werror from compiler flags until code is clean enough.

(wiz)

Missing commit for checksum update

(manu)

Updated inputmethod/ibus-pinyin to 1.3.11

(obache)

Update ibus-pinyin to 1.3.11.

    Fix make distcheck errors.

    Tiny modification to m4/boost.m4.

    Pass $CXXFLAGS to $ac_cpp to make cross-compilation of ibus-pinyin easy.

    Remove sqlite3 in create_db.py

    Remove duplicated EXTRA_DIST value in lua/Makefile.am

    Add PY prefix of all file names to avoid name conflict on some file
    system than is case insensitive.

    change the default fuzzy settings

(obache)

Updated devel/p5-Test-WWW-Selenium to 1.23

(sno)

Updating devel/p5-Test-WWW-Selenium from 1.21nb1 to 1.23

Upstream changes:
1.23 - So Sep  5 2010
  - 3 Patches from Tom Hukins++ to clean up inc/, pod and the license.
  - Suggestion from So Matsui++ to restore HTTP POST from the 1.20 release.
  - http_method now defaults to POST but can be set to GET.

1.22 - Fri Apr  2 10:51:58 PDT 2010
  - Update URI::Encode dependency so we can depend on a consistent
    encoding of *firefox to %2Afirefox.
  - No code changes, just test fixes.

(sno)

LP64 fixes

(manu)

Remove unused variable pointing to non-existent dir.
pkglint is now much happier.

(wiz)

Fix build on Solaris, from Aleksey Cheusov in PR 43841.

(wiz)

Add VDPAU support, from Aleksey Cheusov.

(wiz)

Add RCS Id.

(wiz)

Add RCS Id.

(wiz)

pkglint cleanup.

(wiz)

Updated net/openvpn to 2.1.3

(adam)

Changes 2.1.3:
* Fixed potential local privilege escalation vulnerability in
  Windows service.
* Added Python-based based alternative build system for Windows using
  Visual Studio 2008 (in win directory).
* When aborting in a non-graceful way, try to execute do_close_tun in
  init.c prior to daemon exit to ensure that the tun/tap interface is
  closed and any added routes are deleted.
* Fixed an issue where AUTH_FAILED was not being properly delivered
  to the client when a bad password is given for mid-session reauth,
  causing the connection to fail without an error indication.
* Don't advance to the next connection profile on AUTH_FAILED errors.
* Fixed an issue in the Management Interface that could cause
  a process hang with 100% CPU utilization in --management-client
  mode if the management interface client disconnected at the
  point where credentials are queried.
* Fixed an issue where if reneg-sec was set to 0 on the client,
  so that the server-side value would take precedence,
  the auth_deferred_expire_window function would incorrectly
  return a window period of 0 seconds.  In this case, the
  correct window period should be the handshake window period.
* Modified ">PASSWORD:Verification Failed" management interface
  notification to include a client reason string:
    >PASSWORD:Verification Failed: 'AUTH_TYPE' ['REASON_STRING']
* Enable exponential backoff in reliability layer retransmits.
* Set socket buffers (SO_SNDBUF and SO_RCVBUF) immediately after
  socket is created rather than waiting until after connect/listen.
* Management interface performance optimizations:
  1. Added env-filter MI command to perform filtering on env vars
    passed through as a part of --management-client-auth
  2. man_write will now try to aggregate output into larger blocks
    (up to 1024 bytes) for more efficient i/o
* Fixed minor issue in Windows TAP driver DEBUG builds
  where non-null-terminated unicode strings were being
  printed incorrectly.
* Fixed issue on Windows with MSVC compiler, where TCP_NODELAY support
  was not being compiled in.
* Proxy improvements:
* Implemented http-proxy-override and http-proxy-fallback directives to make it
  easier for OpenVPN client UIs to start a pre-existing client config file with
  proxy options, or to adaptively fall back to a proxy connection if a direct
  connection fails.
* Implemented a key/value auth channel from client to server.
* Fixed issue where bad creds provided by the management interface
  for HTTP Proxy Basic Authentication would go into an infinite
  retry-fail loop instead of requerying the management interface for
  new creds.

(adam)

+ bftpd-3.0, cvsd-1.0.20, eric4-4.4.8, gdb-7.2, kdevelop-4.0.2,
  mkvtoolnix-4.3.0.

(wiz)

+ ImageMagick-6.6.4.0, docbook-xsl-1.76.0, gnumeric110-1.10.10,
  goffice0.8-0.8.10, gst-plugins0.10-bad-0.10.20,
  gst-plugins0.10-ugly-0.10.16, libproxy-0.4.6, libv4l-0.8.1,
  mc-4.7.0.8, p5-Module-Signature-0.65, p5-PerlMagick-6.6.4.0,
  p5-WWW-Curl-4.13, scmgit-base-1.7.2.3, scmgit-docs-1.7.2.3,
  vala-0.9.8, x264-devel-20100904.

(wiz)

Require at least version 3.1.3 of the "rpm2pkg" package to make sure we
support RPM package files which use LZMA compression for the data section.

(tron)

Note update of the "rpm2pkg" package to version 3.1.3.

(tron)