================================
2013-08-20 13:23:05 -0400 v1_355
================================
commit e408a83c9a5342887954800c9ec9b109aeb8bc1d
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Tue Aug 20 13:23:05 2013 -0400
Bump version and dependencies for release.
commit 012098ebc019e1cd1c74d23fab7b4e7112a3adec
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Mon Aug 19 18:39:56 2013 -0400
[rt.cpan.org 87931] Redo _data_ev_adjust in terms of peek_items().
We can't add to the POE::Queue API without breaking third-party
extensions. This takes the performance hit to avoid breaking other
CPAN modules.
commit 810cba7763ac815b914af3967b8d33b0a84aca13
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Mon Aug 19 17:32:28 2013 -0400
[rt.cpan.org 87928] POE::Queue::Array didn't need
POE::Resource::Clock?
commit dea8746ea464ea343389d2bf604cc5bd9cf78a0f
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Mon Aug 19 03:51:46 2013 -0400
Skip POD::Coverage checks for POE::Pipe::OneWay and TwoWay.
The documentation has been replaced by deprecations. POD coverage is
moot.
commit deccbf370f3b2a2d89c5331acf1b4d84d972a61d
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Mon Aug 19 03:50:06 2013 -0400
Remove unused code from POE::Resource::Clock.
A number of functions aren't being used. More are only being used
privately within the module. Comment out some of the unused ones
(more dead code checking is needed). Add leading underscores to the
private ones.
commit 9c34ae0c06d339c5de81a538f05847bbf61ff5c5
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Mon Aug 19 03:47:48 2013 -0400
Make a new POE::Queue::Array method private.
It's currently experimental, and it currently breaks other event
queue nimplementations, so it has to go. Making it private will
temporarily reduce temptation to use it, and it helps the POD
coverage tests pass.
commit bcf5d4451637bab12a2c20ccbc028d042323389b
Author: Joseph Huckaby <jhuckaby@gmail.com>
Date: Mon Aug 19 01:04:45 2013 -0400
[rt.cpan.org 87922] Don't crash when ClientPreConnect returns undef.
ClientPreConnect should be allowed to return an undefined value to
abort the client connection. Instead, it caused PoCo::Server::TCP to
crash. This change catches that condition.
commit f42a26db921db2bbbb951a1becd02800d8ceb2d2
Author: Nicko Dehaine <nicko@phone.com>
Date: Mon Aug 19 00:03:29 2013 -0400
[rt.cpan.org 87721] Clear $! in Driver::SysRW::flush().
Nothing prevents the user from calling flush() when there is nothing
to flush. In these cases, $! remained unchanged, which could lead to
detection of false error conditions. This change clears $! to avoid
the problem.
commit 4e7961d6370cef5741d4ac2f9023fb7e2dad309d
Author: Markus Müller <privi@cpan.org>
Date: Sun Aug 18 21:52:36 2013 -0400
[rt.cpan.org 87635] Fix PoCo::Client::TCP "Connected" event
w/watermarks.
If you use watermarks, the Connected Event will not get triggert
because of a wrong return. This change corrects that.
commit 7bce20d1366f2bb3e6014a8d4db19aa93c00e9aa
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Sun Aug 18 04:44:12 2013 -0400
Eliminate an uninitialized value warning in the tests.
commit cb8a944db305c3d9adb1752d2d730bdec192cfd7
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Sun Aug 18 03:35:25 2013 -0400
Add t/10_units/05_filters/51_reference_die.t to MANIFEST.
commit b07d2ccdc10c19f087166125ca91fc8b8468d417
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Sun Aug 18 02:50:22 2013 -0400
Replace POE::Pipe::* guts with IO::Pipely.
This change doesn't change the way POE or its tests use POE::Pipe::*.
All tests pass in my development environment, so I expect no obvious
problems using IO::Pipely.
commit c8974aa92080b684a2976c7574b80ca7cac00c6a
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Sun Aug 18 02:19:15 2013 -0400
Update copyright dates.
commit 890244dd7b20a86301c024b9e994c36d993f9180
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Sun Aug 18 02:08:45 2013 -0400
Move the signal information out of the signal pipe.
The signal pipe included all signal information. This reduces the
number of signals that can be dispatched in a single event loop
iteration. It also limits the amount and/or type of information we
can include in the signal.
This change puts the signal information in a plain Perl list and only
uses the signal pipe to wake up the event loop to dispatch the
signals in a safer place.
commit c88d1d20c151909b823e279af95a1d826368cc46
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Sun Aug 18 01:59:15 2013 -0400
Update sample programs.
commit 78722c75e36ccd9ae8280a3202adb9840a7bedbe
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Sat Aug 17 17:19:24 2013 -0400
Add lib/POE/Resource/Clock.pm to MANIFEST.
Help wanted: Migrate build/release process to Dist::Zilla, or advise
on same. It's complicated right now. Probably too complicated.
commit 3fb1edba213e10ef85559eabfbbf3e9790841564
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Tue Jul 16 00:55:03 2013 -0400
[rt.cpan.org 81550] Untaint %ENV as it's examined for TRACE/ASSERTs.
commit 055ccf3a4eb5645272c3a076641f9ca89e5d88fa
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Mon Jul 15 22:39:15 2013 -0400
Fix a typo introduced in an earlier commit.
commit 0f36f724df6c77d0c53d248fc74e71e7cbdf8057
Author: Mike <pru-mike@mail.ru>
Date: Sat Jul 6 15:10:38 2013 -0400
Add a test case for the MSWin32 stat() bug fixed a few commits ago.
commit 6ecb93be7fc24fadf5efd1d351e69070d37a2cab
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Sat Jul 6 15:09:47 2013 -0400
Turn off debugging.
commit bf9514fbddfc5e954b3c2f5e54724ee2da3868fe
Merge: 223a8cc 940189d
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Sat Jun 8 21:54:52 2013 -0400
Merge remote-tracking branch 'remotes/sf/master'
commit 940189dfb8177e3685bc46fcf256bf6752024f32
Author: Philip Gwyn <gwyn@cpan.org>
Date: Sat Jun 8 12:36:31 2013 -0400
Added POE::Resource::Clock
commit 223a8cca126e70aed5f15979cb3ea9fc661909b1
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Fri Jun 7 17:08:20 2013 -0400
Consistently stat file names or handles.
stat() on Windows reports different device IDs for a file depending
whether it was stat()ed via name or opened handle. If used
inconsistently, stat() will always report differences. Discovered by
"pru-mike" at
blogs.perl.org/users/pru-mike/2013/06/creepy-perl-stat-functions-on-w
indows.html
commit 0aa938883c65605c057f7d234b2f2fd01cafad74
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Thu May 2 22:10:22 2013 -0400
Fix iThread-safety issue in POE::Component docs.
Stringified object references aren't iThread-safe. Spawn a new
interpreter, and copies of object have different memory addresses.
Previously stringified versions are not fixed to match. Thanks to
Paul Evans for pointing it out.
commit 219f5e1f4e526905d452d6112924224a2568e5da
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Sun Apr 28 16:29:27 2013 -0400
Note how to get ClosedEvent when killing processes in Windows.
Rafael Kitover provided a test case where kill() would prevent
ClosedEvent from firing on Winodws. We worked out a work-around,
which he'll be using in MooseX::Workers. I've documented it here for
others.
commit 42a5f9cae2d6cc29339d022153d44926a8ac4711
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Sat Apr 20 22:50:37 2013 -0400
Replace Z<comment> with =for comment, per perlpod.
commit b7eec09b0f052c210925999b09a921a165b24f0f
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Tue Dec 18 01:37:52 2012 -0500
[rt.cpan.org 77907] Apply Alexander Becker's documentation
suggestion.
commit 032420f716604cac2914d692e3c5d3c8a7dbeefa
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Tue Dec 18 01:34:44 2012 -0500
[rt.cpan.org 77233] Apply Lukas Mai's suggested POD fix.
commit 814486ce85ac3846267a4e3aeb85c5f62c4c0e5c
Author: Philip Gwyn <gwyn@cpan.org>
Date: Tue Dec 4 17:13:32 2012 -0500
Added POE::Resource::Clock which exports monotime() and walltime()
The Queue uses monotime() as a priority, but the public API uses
walltime() Fixed EV_TIME to EV_FROMSTATE EV_SEQ is set again
commit 513999b229c682a25773be4199f93685223957ee
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Mon Oct 29 18:27:52 2012 -0400
Don't bother with IO::Tty on Cygwin.
average_drifter on freenode #perl-cats pointed out that IO::Tty fails
to install on contemporary Cygwin. This change skips that requirement
when $^O eq 'cygwin'.
After investigating, I found and reported two problems. These must be
resolved before undoing this work-around:
IO::Tty compile failure:
https://rt.cpan.org/Ticket/Display.html?id=80477
App::cpanminus conrigure timeout is too short:
https://rt.cpan.org/Ticket/Display.html?id=80478
commit ec43fcdf40ff511de6cd08785298670181a6b7e6
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Thu Oct 11 09:26:16 2012 -0400
Don't confess on the use of AF_INET6 or PF_INET6 in SocketFactory.
SocketFactory uses those constants unconditionally during
initialization, so things that don't even need them fail.
http://www.cpantesters.org/cpan/report/9c1ad42e-12dc-11e2-81a5-a3382b
22a531 is a Perl 5.6.2 test failure, but as far as I can tell all the
errors are due to missing IPv6 support on that platform.
commit d8eec5291a94ad006e4f98d43db11a7d1821627b
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Thu Oct 11 01:31:26 2012 -0400
[rt.cpan.org 79379] Document repositories and online change log
browsers.
Greg Sabino Mullane pointed out the difficulty of finding a complete,
up to date change log. The CHANGES file distributed with POE only
goes back a year, and he needed to go farther back.
commit 616bb1050e8f01150c6db1a02570d6dff9c0320c
Author: Philip Gwyn <gwyn@cpan.org>
Date: Thu Sep 27 15:33:33 2012 -0400
[rt.cpan.org 79886] Fix SIGDIE handler save/restore for Perl 5.8.8.
Applied a modified version of the patch and test case included in the
RT ticket.
commit cb4f65f0aefdf7ed7b22d5db8ea8f0773f47d41b
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Fri Aug 31 11:09:35 2012 -0400
Optionally export FIRST_UNUSED so subclasses know where in @$self to
start.
commit f7f761cd596cafac6d86e3d69e74965f9747bf9b
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Thu Jun 28 00:17:40 2012 -0400
Add more signal handling diagnostics when asserting data at shutdown.
commit f27fa471d95ac270c1a927a27438587cc92bccf8
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Thu Jun 28 00:13:31 2012 -0400
Docment that waitpid() calls interfere with sig_child().
Programs that reap their own child processes AND use sig_child() to
watch for those processes MUST call sig_child($pid, undef) to clear
the watchers. sig_child() only notifies programs after some event
loop reaps the process. If an application reaps it first, POE may
never know the child went away.
==============
End of Excerpt
==============