The Perl Toolchain Summit needs more sponsors. If your company depends on Perl, please support this very important event.
JMASON / IPC-DirQueue-0.07 / CHANGES 
------------------------------------------------------------------------
r2354 | jmason | 2006-01-02 18:07:51 -0800 (Mon, 02 Jan 2006) | 1 line

make the 600-second active-lock timeout configurable; document it better; and reconcile code behaviour with what the documentation says it does re behaviour when a stale lock occurs on a task supposedly active on a remote system (doco wins)
------------------------------------------------------------------------
r2353 | jmason | 2006-01-02 17:38:35 -0800 (Mon, 02 Jan 2006) | 1 line

patch from Anton Berezin <tobez.at.tobez.org>: return reference of sorted files list internally, and use built-in sort instead of a { $a cmp $b } callback.  Both are good for speed of ordered dequeuing
------------------------------------------------------------------------
r2303 | jmason | 2005-12-08 17:32:51 -0800 (Thu, 08 Dec 2005) | 1 line

rename test
------------------------------------------------------------------------
r2301 | jmason | 2005-12-08 17:31:15 -0800 (Thu, 08 Dec 2005) | 1 line

back_compat_0_05 setting not necessary; removed
------------------------------------------------------------------------
r2300 | jmason | 2005-12-08 17:30:43 -0800 (Thu, 08 Dec 2005) | 1 line

add 'dq-indexd', an index daemon for use on heavily congested NFS networks, to offload indexing tasks away from NFS
------------------------------------------------------------------------
r2299 | jmason | 2005-12-08 17:25:03 -0800 (Thu, 08 Dec 2005) | 1 line

increase timeouts so tests still work across very slow NFS links
------------------------------------------------------------------------
r2295 | jmason | 2005-12-02 17:50:42 -0800 (Fri, 02 Dec 2005) | 1 line

add dq-indexd, an optional, experimental central server to maintain DQ indexes, instead of using readdir() filesystem APIs; fix race condition in wait_for_queued_job() that can result in missing jobs enqueued inside the same 1-second window; add code to test scripts to ensure that queue directories are empty at end-of-test, which is how this bug was caught
------------------------------------------------------------------------
r2294 | jmason | 2005-12-01 17:26:52 -0800 (Thu, 01 Dec 2005) | 1 line

refactor queue-fanout code out of the mainline a little more cleanly
------------------------------------------------------------------------
r2293 | jmason | 2005-12-01 16:16:43 -0800 (Thu, 01 Dec 2005) | 1 line

remove 'lastenq' flag file; turns out it's entirely unnecessary to maintain a separate file for the purpose, when the 'queue' directory itself works just as well, and will be in the buffer cache already
------------------------------------------------------------------------
r2290 | jmason | 2005-11-29 19:42:19 -0800 (Tue, 29 Nov 2005) | 1 line

add tests
------------------------------------------------------------------------
r2289 | jmason | 2005-11-29 19:24:07 -0800 (Tue, 29 Nov 2005) | 1 line

add 'queue_fanout' option, to work around NFS scalability issues with massive queue directories
------------------------------------------------------------------------
r2285 | jmason | 2005-11-28 13:32:53 -0800 (Mon, 28 Nov 2005) | 1 line

add test scripts for that bug - namely ordering issues when starting from a fresh slate
------------------------------------------------------------------------
r2284 | jmason | 2005-11-28 11:44:47 -0800 (Mon, 28 Nov 2005) | 1 line

bug in lastenq flag-file usage; the flag file may not exist, even if the clients support it, if the queue dir has not been enqueued-to yet
------------------------------------------------------------------------
r2259 | jmason | 2005-11-23 16:25:26 -0800 (Wed, 23 Nov 2005) | 1 line

add test script for enqueue_fh()
------------------------------------------------------------------------
r2258 | jmason | 2005-11-23 16:23:32 -0800 (Wed, 23 Nov 2005) | 1 line

API CHANGE: invalid metadata now causes die() to be called, instead of silent ignore.  also note that enqueue_fh() closes the FH
------------------------------------------------------------------------
r2257 | jmason | 2005-11-22 18:58:38 -0800 (Tue, 22 Nov 2005) | 1 line

fixed a long-running bug; left-over control files in 'active'.   Turns out these were a side effect of qproc A completing a task before qproc B even started creating a lockfile; in that case, qproc B would get a lock on the now-complete task, then find that the queue control file was nonexistent and give up, attempting to remove the 'active' file, but instead unlinking the now-already-unlinked temporary filename, and leaving the real 'active' file behind.  Totally harmless, fixed anyway.
------------------------------------------------------------------------
r2256 | jmason | 2005-11-22 13:43:21 -0800 (Tue, 22 Nov 2005) | 1 line

avoid silly misleading test failures caused by accidental sharing of DQ dirs, by removing the test-suite output dir at the start of each run
------------------------------------------------------------------------
r2255 | jmason | 2005-11-22 13:41:14 -0800 (Tue, 22 Nov 2005) | 1 line

add flags file dir, and lastenq flags file, to support Reiserfs DQ dirs, which do not update directory mtimes when a file is creating within them.  Also rename enqueue_backend() to _enqueue_backend(), to emphasise its private API nature; and a couple of minor code readability improvements
------------------------------------------------------------------------
r2161 | jmason | 2005-10-21 15:25:26 -0700 (Fri, 21 Oct 2005) | 1 line

explain what the 'active' dir is for; possible workaround for the lidnr errors, with more data-ful messages for working out what's going on
------------------------------------------------------------------------
r1014 | jmason | 2005-04-22 13:24:44 -0700 (Fri, 22 Apr 2005) | 1 line

0.05 RELEASED
------------------------------------------------------------------------
r929 | jmason | 2005-03-15 15:12:28 -0800 (Tue, 15 Mar 2005) | 1 line

more paranoid link_into_dir_no_retry implementation, hopefully will fix a bug that leaves tmp active files lying around
------------------------------------------------------------------------
r863 | jmason | 2005-02-08 16:58:22 -0800 (Tue, 08 Feb 2005) | 1 line

redo the enqueue_sub() API; the read() style of string/empty-str/undef produced really messy, complex calling code, so just go with a more perlish string/undef/die() API instead
------------------------------------------------------------------------
r862 | jmason | 2005-02-08 16:16:50 -0800 (Tue, 08 Feb 2005) | 1 line

bump version number; not yet ready for a release, but worth doing this before I forget
------------------------------------------------------------------------
r861 | jmason | 2005-02-08 16:12:08 -0800 (Tue, 08 Feb 2005) | 1 line

added enqueue_sub() API, to enqueue data from a closure
------------------------------------------------------------------------
r857 | jmason | 2005-02-06 20:31:14 -0800 (Sun, 06 Feb 2005) | 1 line

skip fork()-requiring tests on win32; failure log at http://www.nntp.perl.org/group/perl.cpan.testers/179567
------------------------------------------------------------------------
r663 | jmason | 2005-01-18 17:09:00 -0800 (Tue, 18 Jan 2005) | 1 line

RELEASING 0.04
------------------------------------------------------------------------
r662 | jmason | 2005-01-12 18:26:40 -0800 (Wed, 12 Jan 2005) | 1 line

add note about mailing list
------------------------------------------------------------------------
r661 | jmason | 2005-01-12 18:22:44 -0800 (Wed, 12 Jan 2005) | 2 lines

bug: dqserver should run forever when --njobs is 0, or unset; instead, it was exiting immediately.  Patch from Jeffrey Wescott <jeff . binaryfeed . org>.

------------------------------------------------------------------------
r461 | jmason | 2004-09-17 04:06:57 -0700 (Fri, 17 Sep 2004) | 1 line

added debugging; also, tracked down a mysterious condition that resulted in unused files being left in the active dir.
------------------------------------------------------------------------
r460 | jmason | 2004-09-17 04:05:35 -0700 (Fri, 17 Sep 2004) | 1 line

MANIFEST fixes
------------------------------------------------------------------------
r453 | jmason | 2004-09-09 18:14:30 -0700 (Thu, 09 Sep 2004) | 1 line

updated documentation on the directory structure used; readying release
------------------------------------------------------------------------
r449 | jmason | 2004-08-30 14:26:49 -0700 (Mon, 30 Aug 2004) | 1 line

bump version to 0.03
------------------------------------------------------------------------
r448 | jmason | 2004-08-26 21:22:07 -0700 (Thu, 26 Aug 2004) | 1 line

now up to the latest. added hashing of the data dir; long tests to measure speed effects of code changes; filenames are now significantly shortened using a hash instead of the raw string for the (hostname,pid) part
------------------------------------------------------------------------
r447 | jmason | 2004-08-26 21:19:25 -0700 (Thu, 26 Aug 2004) | 1 line

version 0.02, for the record
------------------------------------------------------------------------
r446 | jmason | 2004-08-26 21:18:41 -0700 (Thu, 26 Aug 2004) | 1 line

version 0.01, for the record
------------------------------------------------------------------------
r445 | robert | 2004-08-26 20:11:46 -0700 (Thu, 26 Aug 2004) | 1 line

[perl #31356] new svn for IPC::DirQueue
------------------------------------------------------------------------