poe:poe-test-loops.git
6 years agoVersion bump for release. v1_311
Rocco Caputo [Mon, 2 May 2011 23:11:06 +0000 (19:11 -0400)]
Version bump for release.

6 years agoSwitch to Dist::Zilla.
Rocco Caputo [Wed, 20 Apr 2011 04:07:39 +0000 (00:07 -0400)]
Switch to Dist::Zilla.

6 years agoBump for release.
Rocco Caputo [Tue, 19 Apr 2011 17:10:51 +0000 (13:10 -0400)]
Bump for release.

6 years agoTest for transpose-chars at end of line v1_050
Hinrik Örn Sigurðsson [Mon, 18 Apr 2011 17:12:10 +0000 (17:12 +0000)]
Test for transpose-chars at end of line

Signed-off-by: Apocalypse <perl@0ne.us>
6 years agomake wording more consistent in the skip message
Apocalypse [Mon, 18 Apr 2011 03:54:00 +0000 (20:54 -0700)]
make wording more consistent in the skip message

6 years agoadd missing POE_DANTIC check
Apocalypse [Mon, 18 Apr 2011 03:43:55 +0000 (20:43 -0700)]
add missing POE_DANTIC check

6 years agostylistic tweaks to the recent AF_INET6 patch
Apocalypse [Sun, 17 Apr 2011 02:12:14 +0000 (19:12 -0700)]
stylistic tweaks to the recent AF_INET6 patch

6 years agofinalize the AF_INET6 woes
Apocalypse [Sun, 17 Apr 2011 01:52:05 +0000 (18:52 -0700)]
finalize the AF_INET6 woes

6 years agosquash the AF_INET6 warning for good
Apocalypse [Sat, 16 Apr 2011 22:25:28 +0000 (15:25 -0700)]
squash the AF_INET6 warning for good

6 years agoSilence AF_INET6 warnings, per Apocalypse's research.
Rocco Caputo [Sat, 16 Apr 2011 21:18:43 +0000 (17:18 -0400)]
Silence AF_INET6 warnings, per Apocalypse's research.

6 years agoAdd better support for those times when AF_INET6 can't be found.
Rocco Caputo [Sat, 16 Apr 2011 17:56:39 +0000 (13:56 -0400)]
Add better support for those times when AF_INET6 can't be found.
Resolves an issue discovered by Apocalypse on Perl 5.6.2.

6 years agoFix the way wheel_run_size.pm skips when Term::Size is absent.
Rocco Caputo [Wed, 9 Mar 2011 17:34:50 +0000 (12:34 -0500)]
Fix the way wheel_run_size.pm skips when Term::Size is absent.

6 years agoAdd a test for terminal size changes in Wheel::Run.
Tom Feist [Sun, 20 Feb 2011 18:34:08 +0000 (13:34 -0500)]
Add a test for terminal size changes in Wheel::Run.

6 years agoBump for release. v1_040
Rocco Caputo [Thu, 3 Feb 2011 01:11:23 +0000 (20:11 -0500)]
Bump for release.

6 years agoBump version.
Rocco Caputo [Sat, 29 Jan 2011 02:33:59 +0000 (21:33 -0500)]
Bump version.

6 years agofix non-numeric comparison when using dev release
Apocalypse [Fri, 28 Jan 2011 19:46:19 +0000 (12:46 -0700)]
fix non-numeric comparison when using dev release

6 years agoEliminate Socket6 use from the tests.
Rocco Caputo [Fri, 28 Jan 2011 01:27:48 +0000 (20:27 -0500)]
Eliminate Socket6 use from the tests.

6 years agoResolve a test failure on CentOS, which appears to be a race condition in the test...
Perl Whore [Thu, 27 Jan 2011 23:00:24 +0000 (18:00 -0500)]
Resolve a test failure on CentOS, which appears to be a race condition in the test itself.  Resolves rt.cpan.org #61727.

7 years agoFix alarms tests to match POE's documentation.
Rocco Caputo [Sat, 4 Dec 2010 04:43:40 +0000 (23:43 -0500)]
Fix alarms tests to match POE's documentation.

7 years agoConvert metadata and changelogs to Git workflow.
Rocco Caputo [Sat, 4 Dec 2010 04:43:28 +0000 (23:43 -0500)]
Convert metadata and changelogs to Git workflow.

7 years agoNew revision for a new release. v1_036
Rocco Caputo [Fri, 19 Nov 2010 03:37:59 +0000 (22:37 -0500)]
New revision for a new release.

7 years agoFix POE::Wheel::ReadLine test issues on solaris and DragonFly BSD.
Rocco Caputo [Thu, 18 Nov 2010 13:53:16 +0000 (08:53 -0500)]
Fix POE::Wheel::ReadLine test issues on solaris and DragonFly BSD.

Force a known good TERM environment variable on solaris, rather than
whatever one might be set.  Catch the case where POE::Wheel::ReadLine
fails to instantiate, and skip all tests.

7 years agoRemove cvs/svn keywords.
Rocco Caputo [Wed, 8 Sep 2010 02:14:56 +0000 (22:14 -0400)]
Remove cvs/svn keywords.

7 years agoPromote poe-test-loops contents to the top level.
Rocco Caputo [Wed, 8 Sep 2010 02:01:35 +0000 (22:01 -0400)]
Promote poe-test-loops contents to the top level.

7 years agoBump revision for a new release. No significant changes, other than to build a new... v1_035
Rocco Caputo [Sun, 4 Apr 2010 00:01:00 +0000 (20:01 -0400)]
Bump revision for a new release.  No significant changes, other than to build a new dist with the latest EUMM.

7 years agoBump PTL revision for release. v1_034
Rocco Caputo [Sat, 3 Apr 2010 21:46:46 +0000 (17:46 -0400)]
Bump PTL revision for release.

7 years agoReplace z_rt53302_fh_watchers.pm with z_rt53302_fh_watchers_slow.pm.
Rocco Caputo [Tue, 23 Mar 2010 21:27:32 +0000 (17:27 -0400)]
Replace z_rt53302_fh_watchers.pm with z_rt53302_fh_watchers_slow.pm.
The non-slow one was broken since its client shuts down without
waiting for the server's response.

7 years agoSupport legal =for and =begin POD syntax for skip_tests(), in addition
Rocco Caputo [Mon, 22 Mar 2010 10:00:17 +0000 (06:00 -0400)]
Support legal =for and =begin POD syntax for skip_tests(), in addition
to the previous =for/=cut syntax that caused skip_tests() to appear in
everyone's perldoc.  Document the change.  Deprecate the bad syntax,
which won't be a terrible burden since everyone probably wanted to
migrate away from it anyway.

7 years agoAdd optional one-line =for poe_tests syntax to hide it
Larwan Berke [Mon, 22 Mar 2010 06:54:39 +0000 (02:54 -0400)]
Add optional one-line =for poe_tests syntax to hide it
from POD formatters.

Add example from POE::Loop::Glib that uses the new syntax.

7 years agoAdded the 00_info test for better debugging
Larwan Berke [Thu, 11 Mar 2010 11:57:11 +0000 (06:57 -0500)]
Added the 00_info test for better debugging

7 years agoAdd new tcp connection test for RT#53302
Larwan Berke [Thu, 11 Mar 2010 10:28:23 +0000 (05:28 -0500)]
Add new tcp connection test for RT#53302

7 years agoUpdate the skip_tests() examples, and explain what's going on in each.
Rocco Caputo [Thu, 11 Mar 2010 00:14:25 +0000 (19:14 -0500)]
Update the skip_tests() examples, and explain what's going on in each.

7 years agoFix the followtail test to use File::Temp, thanks to http://cpantesters.org/cpan...
Larwan Berke [Tue, 9 Mar 2010 12:59:40 +0000 (07:59 -0500)]
Fix the followtail test to use File::Temp, thanks to cpantesters.org/cpan/report/6877011 and others like it

7 years agoAfter a lot of hacking/digging Cygwin with a TERM=dumb does not work with POE::Wheel...
Larwan Berke [Tue, 9 Mar 2010 12:20:16 +0000 (07:20 -0500)]
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 :)

7 years agoBump for yet another release. v1_033
Rocco Caputo [Mon, 15 Feb 2010 11:27:00 +0000 (06:27 -0500)]
Bump for yet another release.

7 years agoMake Bazerka's regression test a little more forgiving about elapsed run time.
Rocco Caputo [Mon, 15 Feb 2010 06:05:37 +0000 (01:05 -0500)]
Make Bazerka's regression test a little more forgiving about elapsed run time.

7 years agoSkip pty tests if IO::Pty isn't available. Also some style fixes.
Rocco Caputo [Mon, 15 Feb 2010 04:55:56 +0000 (23:55 -0500)]
Skip pty tests if IO::Pty isn't available.  Also some style fixes.

7 years agoMake the test a little more forgiving for slower systems, such as virtualized
Rocco Caputo [Mon, 15 Feb 2010 04:55:18 +0000 (23:55 -0500)]
Make the test a little more forgiving for slower systems, such as virtualized
Windows systems and other forms of emulation.

7 years agoAdd a test case from rt.cpan.org #54319. Verify Po::Wh::FollowTail v1_032
Rocco Caputo [Sun, 14 Feb 2010 07:44:47 +0000 (02:44 -0500)]
Add a test case from rt.cpan.org #54319.  Verify Po::Wh::FollowTail
actually tails files.  Thanks to Bazerka on irc.perl.org for
submitting the basis of this test case.

7 years agoTry to resolve some CPAN indexer issues.
Rocco Caputo [Sat, 2 Jan 2010 07:57:35 +0000 (02:57 -0500)]
Try to resolve some CPAN indexer issues.

Eliminate a $VERSION string being used to test a version (not define
one).  Eliminate a package statement used to define a constant (not a
class).  Change a test package name to avoid indexer conflict with
another distribution.

7 years agoPrepare for release. v1_031
Rocco Caputo [Fri, 1 Jan 2010 13:53:54 +0000 (08:53 -0500)]
Prepare for release.

7 years agoCheck for initial reset on delayed open.
Rocco Caputo [Sun, 27 Dec 2009 14:56:05 +0000 (09:56 -0500)]
Check for initial reset on delayed open.

7 years agoClarify FollowTail debugging messages.
Rocco Caputo [Sun, 27 Dec 2009 03:32:43 +0000 (22:32 -0500)]
Clarify FollowTail debugging messages.

7 years agoUnfortunately we can't test connect errors on Windows. Behavior seems to vacillate...
Rocco Caputo [Sat, 26 Dec 2009 14:19:55 +0000 (09:19 -0500)]
Unfortunately we can't test connect errors on Windows.  Behavior seems to vacillate from one version of ActivePerl to another.

7 years agoResolve a line ending issue that caused false failures in Windows.
Rocco Caputo [Sat, 26 Dec 2009 13:50:19 +0000 (08:50 -0500)]
Resolve a line ending issue that caused false failures in Windows.

8 years agoNew version for a new release. v1_030
Rocco Caputo [Wed, 14 Oct 2009 08:24:18 +0000 (04:24 -0400)]
New version for a new release.

8 years agoExplicitly bind to loopback address for the concurrent TCP tests
Chris Williams [Tue, 13 Oct 2009 16:43:42 +0000 (12:43 -0400)]
Explicitly bind to loopback address for the concurrent TCP tests

8 years agoFix Marco's name. Oops!
Rocco Caputo [Sun, 11 Oct 2009 22:53:43 +0000 (18:53 -0400)]
Fix Marco's name. Oops!

8 years agoSet version for dev release.
Rocco Caputo [Wed, 7 Oct 2009 05:50:42 +0000 (01:50 -0400)]
Set version for dev release.

8 years agoMake the tests generally ASSERT_DEFAULT clean. POE_TRACE_FILENAME hid a multitude...
Rocco Caputo [Tue, 6 Oct 2009 07:50:33 +0000 (03:50 -0400)]
Make the tests generally ASSERT_DEFAULT clean.  POE_TRACE_FILENAME hid a multitude of sins in the test output file.  Update the vim options lines while I'm in there.

8 years agoRedirect STDERR while testing some errors. ASSERT_DEFAULT is on and making noise...
Rocco Caputo [Tue, 6 Oct 2009 07:39:18 +0000 (03:39 -0400)]
Redirect STDERR while testing some errors.  ASSERT_DEFAULT is on and making noise when things are called wrong, even when we do it on purpose.

8 years agoAdd an option to enable all the broken MSWin32 tests.
Rocco Caputo [Mon, 5 Oct 2009 09:32:30 +0000 (05:32 -0400)]
Add an option to enable all the broken MSWin32 tests.

Set POE_DANTIC=1 before running "make test" to watch the pretty fireworks.  Oooh!  Ahhhhhh!

8 years ago[rt.cpan.org 48715] Apply Craig's patch to skip the IPv6 tests if IPv6
Rocco Caputo [Fri, 2 Oct 2009 07:47:17 +0000 (03:47 -0400)]
[rt.cpan.org 48715] Apply Craig's patch to skip the IPv6 tests if IPv6
isn't actually enabled.  The port autodetect code to didn't handle
failure due to IPv6 unavailability.

8 years ago!!! Adjust tests for mark-and-sweep session GC changes.
Rocco Caputo [Tue, 29 Sep 2009 19:13:53 +0000 (15:13 -0400)]
!!! Adjust tests for mark-and-sweep session GC changes.

Some of the tests are sensitive to session GC timing.

8 years agoTest that POE::Kernel->run() without sessions returns right away.
Rocco Caputo [Wed, 16 Sep 2009 19:11:30 +0000 (15:11 -0400)]
Test that POE::Kernel->run() without sessions returns right away.

8 years agoBump up the version, and prevent perl-reversion from seeing test v1_022
Rocco Caputo [Mon, 7 Sep 2009 23:13:43 +0000 (19:13 -0400)]
Bump up the version, and prevent perl-reversion from seeing test
version strings (that are commented out, no less).

8 years ago[rt 36648] Add tests for the new Client and Server TCP component
Rocco Caputo [Sun, 6 Sep 2009 07:32:57 +0000 (03:32 -0400)]
[rt 36648] Add tests for the new Client and Server TCP component
capabilities.

8 years ago[rt 49384] Explicitly bind to 127.0.0.1 in tests, reported anonymously.
Rocco Caputo [Sat, 5 Sep 2009 21:56:52 +0000 (17:56 -0400)]
[rt 49384] Explicitly bind to 127.0.0.1 in tests, reported anonymously.
[cpan testers 5241203] Bind to INPORT_ANY to avoid "address already in use" failures, reported by Oliver Paukstadt.

8 years agoMake write-combining test optional. The timing of some event loops
Rocco Caputo [Fri, 4 Sep 2009 11:49:18 +0000 (07:49 -0400)]
Make write-combining test optional.  The timing of some event loops
triggers write-combining, and others don't.  Resolves rt.cpan.org
28804 (finally!), reported by Martijn van Beers.

8 years agoTk and fork() don't get along well. Skip some fork() based tests when
Rocco Caputo [Fri, 4 Sep 2009 08:28:55 +0000 (04:28 -0400)]
Tk and fork() don't get along well.  Skip some fork() based tests when
Tk is loaded, to avoid XIO errors that have shown up in FreeBSD.

8 years agoSkip pty tests in z_leolo_wheel_run.pm on MSWin32.
Rocco Caputo [Thu, 3 Sep 2009 20:08:06 +0000 (16:08 -0400)]
Skip pty tests in z_leolo_wheel_run.pm on MSWin32.
Change some skip messages.

8 years agoSkip a known issue regarding pseudy-ttys, Event, and Mac OS X. On OS
Rocco Caputo [Sat, 8 Aug 2009 12:25:22 +0000 (08:25 -0400)]
Skip a known issue regarding pseudy-ttys, Event, and Mac OS X.  On OS
X, poll() returns POLLNVAL for pseudo-ttys.  Event cancels its
watchers in this case without a callback.

8 years agoMove Philip Gwyn's signals test to POE::Test::Loops, so it can
Rocco Caputo [Sat, 8 Aug 2009 08:56:28 +0000 (04:56 -0400)]
Move Philip Gwyn's signals test to POE::Test::Loops, so it can
exercise everyone's event loops.

8 years agoAdded test to make sure we get the 3 "completion" events from
Philip Gwyn [Fri, 7 Aug 2009 11:29:13 +0000 (07:29 -0400)]
Added test to make sure we get the 3 "completion" events from
POE::Wheel::Run.  Multiplied by the 4 conduit types.

8 years agoRoll back previous patch. We need to call ->close_slave() on the _read
Philip Gwyn [Thu, 6 Aug 2009 14:55:58 +0000 (10:55 -0400)]
Roll back previous patch.  We need to call ->close_slave() on the _read
handles.  Do that instead.

8 years agoFix case where POE::Wheel::Run CloseEvent doesn't fire on Linux.
Rocco Caputo [Thu, 6 Aug 2009 11:21:19 +0000 (07:21 -0400)]
Fix case where POE::Wheel::Run CloseEvent doesn't fire on Linux.
Thanks to Philip Gwyn's investigation and test cases, we found out
that close() wasn't sufficient to close and destruct slave ptys, so
the parent processes were keeping them open.  Replaced close() with
$io_pty->close_slave() as needed.

8 years agoSmall comment tweak re: top/bottom handlers
Philip Gwyn [Tue, 4 Aug 2009 08:39:59 +0000 (04:39 -0400)]
Small comment tweak re: top/bottom handlers

8 years agoUpdated agaran-filter-httpd.t to bind server poco specifically to the loopback as...
Chris Williams [Fri, 31 Jul 2009 13:09:21 +0000 (09:09 -0400)]
Updated agaran-filter-httpd.t to bind server poco specifically to the loopback as per RT #48341

8 years ago!!! This change breaks backward compatibility on a relatively unused
Rocco Caputo [Wed, 29 Jul 2009 07:53:01 +0000 (03:53 -0400)]
!!! This change breaks backward compatibility on a relatively unused
!!! feature.  You are affected if you use ARG0 or ARG1 in a
!!! POE::Component::Server::TCP ClientConnected callback.

ClientArgs promised more than it could deliver, and people finally
noticed.  This change backs off supplying the socket in $_[ARG0], and
it expands ClientArgs' arrayref into @_[ARG0..$#_].  Thanks to Michael
Fowler for rt.cpan.org #47855 (which this resolves), and POE's mailing
list for advice on which way this change should go.

8 years agoInitialize the output masks for sigprocmask. Undefined output masks
Rocco Caputo [Wed, 29 Jul 2009 06:42:33 +0000 (02:42 -0400)]
Initialize the output masks for sigprocmask.  Undefined output masks
may trigger warnings in some versions of Perl.  At least, I hope
http://www.nntp.perl.org/group/perl.cpan.testers/2009/07/msg4809122.html
warnings will be resolved by this change.  I wasn't able to reproduce
them myself.

8 years agoSome people may be using POE without actually reading the docs.
Rocco Caputo [Tue, 28 Jul 2009 09:36:12 +0000 (05:36 -0400)]
Some people may be using POE without actually reading the docs.
Furthermore, they may go on to rant about the side effects in obscure
places.  Recommend reading the documentation in one of POE's most
"infamous" warnings.

8 years agoRely on POE::Test::Loops 1.021, which should be everywhere by now.
Rocco Caputo [Mon, 27 Jul 2009 22:51:38 +0000 (18:51 -0400)]
Rely on POE::Test::Loops 1.021, which should be everywhere by now.
Update links in documentation.

8 years agoReorganize documentation links.
Rocco Caputo [Mon, 27 Jul 2009 21:07:20 +0000 (17:07 -0400)]
Reorganize documentation links.

8 years agoAdded 'homepage' meta to the META.yml generation code.
Chris Williams [Mon, 27 Jul 2009 11:57:24 +0000 (07:57 -0400)]
Added 'homepage' meta to the META.yml generation code.

8 years agoDocumented the bug tracker, repository, and other resource URLs.
Rocco Caputo [Mon, 27 Jul 2009 09:12:48 +0000 (05:12 -0400)]
Documented the bug tracker, repository, and other resource URLs.

8 years agoAdded a machine-readable repository directory to the distribution.
Rocco Caputo [Mon, 27 Jul 2009 09:10:57 +0000 (05:10 -0400)]
Added a machine-readable repository directory to the distribution.
Documented the bug tracker, repository, and other resource URLs.

8 years agoRoll the version back slightly so CPAN can test before an official release.
Rocco Caputo [Sun, 26 Jul 2009 02:35:58 +0000 (22:35 -0400)]
Roll the version back slightly so CPAN can test before an official release.

8 years agoSet the POE::Test::Loops dependency to 1.020, and bump the version to
Rocco Caputo [Sun, 26 Jul 2009 02:28:49 +0000 (22:28 -0400)]
Set the POE::Test::Loops dependency to 1.020, and bump the version to
1.007 for release.

8 years agoCan't sleep! HP-UX will get me! Resolves Dolmen's bug report at v1_021
Rocco Caputo [Sat, 25 Jul 2009 21:35:34 +0000 (17:35 -0400)]
Can't sleep! HP-UX will get me!  Resolves Dolmen's bug report at
rt.cpan.org 45158 by using select(undef,undef,undef,$time) instead of
sleep().  I hope that actually works.

8 years agoNew version for a new release. v1_020
Rocco Caputo [Fri, 24 Jul 2009 08:33:31 +0000 (04:33 -0400)]
New version for a new release.

8 years agoPlan for a specific test count. Sometimes the no_plan end-run plan
Rocco Caputo [Thu, 23 Jul 2009 00:24:33 +0000 (20:24 -0400)]
Plan for a specific test count.  Sometimes the no_plan end-run plan
doesn't get output, breaking the test despite all subtests passing.

8 years agoExplain the likely cause and solution for the "Child process (pid)
Rocco Caputo [Wed, 22 Jul 2009 23:03:28 +0000 (19:03 -0400)]
Explain the likely cause and solution for the "Child process (pid)
reaped" messages, and why it's important for programs to avoid leaking
processes.

8 years agoAdd error number to signal pipe errors. Windows doesn't always have an error string...
Rocco Caputo [Wed, 22 Jul 2009 19:17:40 +0000 (15:17 -0400)]
Add error number to signal pipe errors.  Windows doesn't always have an error string for some of winsock's wacky error numbers.

8 years agoAdd rt47966-sigchld.t to the MANIFEST.
Rocco Caputo [Wed, 22 Jul 2009 11:45:31 +0000 (07:45 -0400)]
Add rt47966-sigchld.t to the MANIFEST.

8 years agoWait for longer for all children to exit
Philip Gwyn [Wed, 22 Jul 2009 08:34:06 +0000 (04:34 -0400)]
Wait for longer for all children to exit

8 years agoTry to speed up the tests by triggering the children to exit upon
Rocco Caputo [Wed, 22 Jul 2009 06:56:35 +0000 (02:56 -0400)]
Try to speed up the tests by triggering the children to exit upon
demand rather than after a certain amount of time.  Doesn't seem
significant on my G4 laptop, but it might be on faster machines.

8 years agoRun perltidy & convert all tabs to spaces. No actual code changed
Rocco Caputo [Wed, 22 Jul 2009 06:41:27 +0000 (02:41 -0400)]
Run perltidy & convert all tabs to spaces.  No actual code changed
this time around.  It's bad to mix semantic changes with stylistic
ones.  The semantic changes get lost.

8 years agoAdded rt47966-sigchld.t which tortures the signal-pipe with CHLD, for testing
Philip Gwyn [Wed, 22 Jul 2009 00:27:13 +0000 (20:27 -0400)]
Added rt47966-sigchld.t which tortures the signal-pipe with CHLD, for testing
  on win32.  Maybe this test should be moved to POE::Test::Loops

8 years agoWin32 fixes :
Philip Gwyn [Tue, 21 Jul 2009 23:24:24 +0000 (19:24 -0400)]
Win32 fixes :
- Assign fake numbers (>=128) to signals that don't have a SIG* constant
- Don't call sigprocmask() and related

8 years agoMake the somni-poco-server-tcp test explicitly bind to the loopback address
Chris Williams [Tue, 21 Jul 2009 15:19:37 +0000 (11:19 -0400)]
Make the somni-poco-server-tcp test explicitly bind to the loopback address
It is hit and miss if the OS supports binding client connections to 0.0.0.0
otherwise.

8 years agoConditionally call has_forked() if it's available. Should prevent
Rocco Caputo [Tue, 21 Jul 2009 06:22:58 +0000 (02:22 -0400)]
Conditionally call has_forked() if it's available.  Should prevent
issues where the tests are released before the feature.  Also, moved
has_forked() earlier because it was called too early for new
PID-change checks contributed by Philip Gwyn.

8 years agoApply Philip Gwyn's second patch from rt.cpan.org #47966. This one
Rocco Caputo [Tue, 21 Jul 2009 06:22:15 +0000 (02:22 -0400)]
Apply Philip Gwyn's second patch from rt.cpan.org #47966.  This one
builds upon his earlier patch on the same ticket.  It reinitializes
POE's signal handlers in child processes after has_forked() is called.
It documents has_forked() and USE_SIGNAL_PIPE.  It renames the _top
methods to _bottom to follow interrupt handler semantics.  It avoids a
race condition by blocking signals for the duration of pipe rebuilding.

8 years agoGcola on IRC discovered that the internal signals socket wasn't
Rocco Caputo [Mon, 20 Jul 2009 20:27:13 +0000 (16:27 -0400)]
Gcola on IRC discovered that the internal signals socket wasn't
compatible with select() on Windows.  This patch makes it a plain
socket when POE is running on Windows.

8 years agoAdd a bit of debugging. Add a delay before POE is loaded, in case the race is betwee...
Rocco Caputo [Mon, 20 Jul 2009 12:40:10 +0000 (08:40 -0400)]
Add a bit of debugging.  Add a delay before POE is loaded, in case the race is between the test framework's Tk detection and POE initializing Tk at start time.  To avoid unnecessary delays, make them dependent on Tk being loaded.

8 years agoThis test causes Tk or X to throw an error if it's run too soon after
Rocco Caputo [Mon, 20 Jul 2009 11:34:01 +0000 (07:34 -0400)]
This test causes Tk or X to throw an error if it's run too soon after
the previous test.  Add a 2sec delay before the test to give the GUI
time to settle down.

8 years agoThese tests have gone into POE::Test::Loops, where they may be useful
Rocco Caputo [Mon, 20 Jul 2009 10:21:48 +0000 (06:21 -0400)]
These tests have gone into POE::Test::Loops, where they may be useful
for other POE::Loop writers.

8 years agoPromote some of the signal-related regression tests into the event
Rocco Caputo [Mon, 20 Jul 2009 10:19:54 +0000 (06:19 -0400)]
Promote some of the signal-related regression tests into the event
loop tests.  Event loops handle signals in a variety of ways, so they
should all try these tests.  Except the one commented out of the
MANIFEST, as it has a tendency to segfault for people.

8 years agoApply Philip Gwyn's patch to deliver signals via pipe rather than
Rocco Caputo [Sun, 19 Jul 2009 11:38:58 +0000 (07:38 -0400)]
Apply Philip Gwyn's patch to deliver signals via pipe rather than
directly into the queue.  This avoids queue re-entrancy during find or
modification operations, which has previously caused inconsistencies
in the queue, event enqueuing and event dispatch.

Philip's patch may be found along with a great problem description at
http://rt.cpan.org/Public/Bug/Display.html?id=47966

8 years agoRT #47892: added regression test for PoCo-Server-TCP from Michael Fowler.
Chris Williams [Wed, 15 Jul 2009 14:47:05 +0000 (10:47 -0400)]
RT #47892: added regression test for PoCo-Server-TCP from Michael Fowler.

8 years agoRT #47854: Fix and document the default setting for Port option, by Michael Fowler
Chris Williams [Wed, 15 Jul 2009 01:40:50 +0000 (21:40 -0400)]
RT #47854: Fix and document the default setting for Port option, by Michael Fowler