Received: from mail.netbsd.org (mail.netbsd.org [204.152.190.11]) by www.NetBSD.org (Postfix) with ESMTP id 55FBA63BC49 for ; Mon, 6 Sep 2010 11:45:13 +0000 (UTC) Received: by mail.netbsd.org (Postfix, from userid 0) id 000CF63B101; Mon, 6 Sep 2010 11:45:12 +0000 (UTC) Received: from cvs.netbsd.org (cvs.NetBSD.org [IPv6:2001:4f8:3:7:2e0:81ff:fe30:95bd]) by mail.netbsd.org (Postfix) with ESMTP id D039A63B100 for ; Mon, 6 Sep 2010 11:45:09 +0000 (UTC) Received: by cvs.netbsd.org (Postfix, from userid 500) id BF6B9175DD; Mon, 6 Sep 2010 11:45:09 +0000 (UTC) MIME-Version: 1.0 Content-Disposition: inline Content-Transfer-Encoding: 8bit Content-Type: text/plain Date: Mon, 6 Sep 2010 11:45:09 +0000 From: Thomas Klausner Subject: CVS commit: pkgsrc/devel/p5-POE To: pkgsrc-changes@NetBSD.org Reply-To: wiz@netbsd.org X-Mailer: log_accum Message-Id: <20100906114509.BF6B9175DD@cvs.netbsd.org> Sender: pkgsrc-changes-owner@NetBSD.org List-Id: pkgsrc-changes.NetBSD.org Precedence: list Module Name: pkgsrc Committed By: wiz Date: Mon Sep 6 11:45:09 UTC 2010 Modified Files: pkgsrc/devel/p5-POE: Makefile distinfo Log Message: 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. To generate a diff of this commit: cvs rdiff -u -r1.17 -r1.18 pkgsrc/devel/p5-POE/Makefile cvs rdiff -u -r1.8 -r1.9 pkgsrc/devel/p5-POE/distinfo Please note that diffs are not public domain; they are subject to the copyright notices on the relevant files.