================================
2012-05-06 18:29:51 -0400 v1_353
================================
commit e7a90bebb55dca81d2373cef0a39e0d80fdf59c2
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Sun May 6 18:29:51 2012 -0400
Bump the version for release.
commit 2b27ce2c0edecb73af625bf921e7e4758a0adfc1
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Sun May 6 16:41:14 2012 -0400
Add Paul Webster's test for edge-case YAML input.
commit 23a38aa8a5a201f8974f7c806934311ce936cb1e
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Sun May 6 13:51:48 2012 -0400
Allow POE::Filter::Reference to not die on bad wire data.
Paul Webster discovered that YAML from other languages can cause the
YAML module to die. For robustness, POE::Filter::Reference needs to
catch and return those errors gracefully. This commit adds thaw()
exception handling per Paul's specifications.
commit 5f22fcada8b4493dcbcd83c5716c929dc56949f6
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Sun Apr 29 18:44:46 2012 -0400
[rt.cpan.org 76314] Untaint externally supplied domains.
Resolves part of the ticket reported by Dylan Doxey. Thank you for
the test case, Dylan!
commit 71b16471861a87cbf0fe519b18c421d936392884
Author: David Huebner <david.huebner@bolderthinking.com>
Date: Tue Apr 24 11:06:02 2012 -0500
Moved documentation change to a less illogical spot.
commit ffbd225a86a8ce5b919afa576a610c8cd0af33fc
Author: David Huebner <david.huebner@bolderthinking.com>
Date: Tue Apr 24 10:29:06 2012 -0500
Added a bit of documentation about the new immediate option as well
as how POE::NFA operates by default.
commit 7f09336ff35176f8702b20ce12b2dcd145711ba8
Author: David Huebner <david.huebner@bolderthinking.com>
Date: Thu Apr 19 15:06:51 2012 -0500
Added 'immediate' option for changing states immediately. [call
rather than post]
commit 3444b1668e6c1c49562dc9d1a267da368ef29861
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Sun Apr 8 09:54:14 2012 -0400
Include the unknown or illegal alias in error messages about them.
commit dae42ae88e2ddbe261bb1cc217ede562397dfaf0
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Sat Apr 7 23:24:56 2012 -0400
Remove test reliance on deterministic _stop order.
Stephan Jauernick discovered that the order of _stop calls is not
necessarily deterministic, and it depends on the speed of the machine
under test. I relaxed the timing requirements of the neyuki_detach
regression test to account for the variation.
================================
2012-03-22 22:26:33 -0400 v1_352
================================
commit cfce866c4c08967bb0c7cd198a02eef342f4d21d
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Thu Mar 22 22:26:33 2012 -0400
Bump version for release. *sigh*
commit 6f719e9175695d8a456c37bf907a08d32f730d06
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Thu Mar 22 22:15:30 2012 -0400
Very large select() timeouts can trigger EINVAL on Mac OSX.
Sérgio Bernardino reported that POE::Component::Server::IRC's
synopsis would go into a busy loop when a client tried to connect. It
turns out that POE::Loop::Select would fail when the component set a
timeout for a very long time in the future. Cap select() timeouts to
an hour to avoid the problem.
================================
2012-03-11 14:40:57 -0400 v1_351
================================
commit 476c7e5d79b6935218d8ec048a191ddd974b4c24
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Sun Mar 11 14:40:57 2012 -0400
Bump versions for release.
commit 8a0894ef9c2be24d9681c166d2fbfb1970a02fbb
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Sun Mar 11 11:32:46 2012 -0400
[rt.cpan.org 67680] Use IO::Handle::blocking() rather than
$fh->blocking().
Some file handles, such as my $fh = do { no warnings; local *FH };
can't be used as IO::Handle objects. IO::Handle::blocking($fh) seems
to work, though.
This was caught by Troy, who was nice enough to open the bug. Thanks!
commit 61da9a35079aa61cdcbcf2ed3345d91d4baac218
Author: Mark Nunberg <mnunberg@haskalah.org>
Date: Mon Dec 19 00:18:34 2011 -0800
STDIN stays open by default
commit b21285bb449f39dec8c507b595bda0a389cd485d
Author: Mark Nunberg <mnunberg@haskalah.org>
Date: Thu Dec 15 13:29:34 2011 -0800
Allow ::Run to discard stdio file descriptors
- Std{in|out|err}Event no longer required - Stdio file descriptor
discarded for each stdio
which lacks a corresponding event (except read) - Added option to not
open stdin - Added option to redirect stdio to/from other filenames
or filehandles
- Corresponding fix needed for test-loops
commit d78e3d2fdcd999375fbb496fd206e64c649ebe55
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Tue Jan 24 23:28:44 2012 -0500
Work around a consistent timeout issue in Hinrik's POE::Wheel::Run
test.
There is a mysterious 5sec delay when starting processes on some
Windows machines. The consistent, roundish number of seconds makes me
suspect something environmental, like an antivirus program, but
that's just a wild guess.
Thanks to ultramantis on irc.perl.org #poe or letting me access their
machine with teamviewer. I couldn't reproduce the problem on my own
systems, so it was vital to see it failing somewhere.
commit 054da731ea9eb83ff7c21343c2538ab9db29ad13
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Sun Dec 18 15:40:57 2011 -0500
Remove Module::Build usage.
It adds itself to META.yml, but its presence is misleading for POE's
purposes. POE doesn't need it to be built, configured, or installed.
================================
2011-12-15 15:53:01 -0500 v1_350
================================
commit ba53ec4bdd9cacbb18cd3cb82c944ad40739240f
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Thu Dec 15 15:53:01 2011 -0500
Bump revision to 1.350.
commit eae7a261fc6cf5b2950b5cd92fd8d394f078688c
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Wed Dec 7 10:06:29 2011 -0500
Don't block on active child processes at endtime.
It's important to note that child processes exist, but we don't need
to wait around for them. Reap the ones that have already exited, and
alert the user if at least one active process remains after that.
Return immediately; don't wait for it/them to exit.
commit ecd5a778a807dad3d820f3537967d972877f560c
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Wed Aug 17 15:10:11 2011 -0400
Avoid trying to dispatch SIGDIE to sessions in a POE::Kernel that has
been made defunct by the stop() method. This resolves at least one
"please report this stacktrace" bug.
commit 8e782eba985c85f6e074ca59b17e773c687f097e
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Fri Aug 5 14:32:50 2011 -0400
Allow event handlers to persistently alter $SIG{__DIE__}.
Prior to this change, modifications to $SIG{__DIE__} in event
handlers would be lost due to that signal entry being localized. This
brings up an interesting question: Should the rest of %SIG be
localized?
================================
2011-07-28 22:34:13 -0400 v1_312
================================
commit 6ec943edbcfea9d299bc3da8e453ae4edc76627f
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Thu Jul 28 22:34:13 2011 -0400
Bump the version.
commit a4b961a547bb32d4bd9057fb9e0501aa138f602a
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Thu Jul 28 01:52:35 2011 -0400
Apply Stephen Oberholtzer's patch to fix a test freeze.
IO::Handle->blocking() is broken on MSWin32, locking up
t/10_units/04_drivers/01_sysrw.t in some cases. Stephen provided a
patch to Perl and to POE to work around the issue. It's portable
enough that things shouldn't immediately break after Perl is fixed.
Resolves rt.cpan.org ticket #69769.
commit 2cd1404a6fc0d6ed0e6361d75e1e9d67837889db
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Wed Jul 27 02:27:29 2011 -0400
getnameinfo() returns host and port in the opposite order of
unpack_sockaddr_in(). Swap the return values so the right ones appear
in ARG1 and ARG2. Thanks to Luiz Eduardo Roncato Cordeiro at CERT
Brazil for spotting the problem and suggesting the fix. This should
resolve his rt.cpan.org ticket, #68959.
commit b17a3d894fa6f11deef4e7187e87fd5061058cb9
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Wed Jul 27 01:50:25 2011 -0400
Add stack traces to _data_handle_remove() failures.
commit 9bcbb9098bf12139b4f80ef96650eafba3e1d95b
Author: Hinrik Örn Sigurðsson <hinrik.sig@gmail.com>
Date: Tue Jun 21 21:07:37 2011 +0000
Add note about Perl 5.6 issues with '>&' open mode
commit 937f52e99ce487326a176c0742fb20fc71b32bf1
Author: Hinrik Örn Sigurðsson <hinrik.sig@gmail.com>
Date: Sun May 22 06:37:34 2011 +0000
Document using getsockname to verify a listening socket
commit f0eff5d176cd9e5522bd8d2d2461396f704f2429
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Wed May 18 14:07:20 2011 -0400
Localize $SIG{__DIE__} when rethrowing userspace exceptions from a
safe point. Avoids shenanigans at a distance when reporting on
userspace errors. May thwart someone's exception handling, but I
don't know a good way to satisfy everyone here. Suggestions?
commit 9a8c4a3871456ef70ecf82085b4af2f6bc0e1509
Author: Rocco Caputo <rcaputo@cpan.org>
Date: Sat May 14 17:23:41 2011 -0400
Don't fail t/00_info.t if POE::Test::Loops is unavailable.
This test dependency is intended to be optional, and the loop tests
should be skipped if the user doesn't want them. Problem reported by
skbz in irc.perl.org #poe, related to cpanm installing
POE::Test:Loops in a user-local directory but not looking there when
installing POE.
commit e6887bddb925a775960a168e4dabcf700b2c376c
Author: Hinrik Örn Sigurðsson <hinrik.sig@gmail.com>
Date: Tue May 3 00:04:26 2011 +0000
Fix mangled stdio checks. This one passes all tests.
==============
End of Excerpt
==============