ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/Coro/Changes
(Generate patch)

Comparing Coro/Changes (file contents):
Revision 1.111 by root, Mon Sep 5 22:13:05 2005 UTC vs.
Revision 1.205 by root, Wed Jan 24 16:22:08 2007 UTC

1Revision history for Perl extension Coro. 1Revision history for Perl extension Coro.
2
3TODO: better (earlier) error messages on broken transfer's
4TODO: destroy gets only called on cancel: fix(?)
5
63.5
7 - do AnyEvent model detection earlier, avoiding problems
8 caused by first using AnyEvent and later Coro::Event.
9 - implement and document Coro::Event event objects.
10 - fix a potential problem in Coro::Event causing crashes.
11
123.41 Mon Jan 22 19:19:49 CET 2007
13 - readline on Coro::Handle did not support undefined $/,
14 nor did it deliver partial lines on EOF or error.
15 - implement malloc fallback for stack allocation because
16 stupid broken idiotic OSX has a stupid broken
17 idiotic fits-the-whole-os mmap "implementation" and
18 my dick feels longer if Coro is portable even to
19 obsolete platforms.
20
213.4 Fri Jan 19 21:52:54 CET 2007
22 - remove t/09_timer.t, as it isn't really testing much
23 but was rather flaky in practise.
24 - async_pool coro would keep arguments and callback alive until
25 it was reused.
26 - cancellation of a coroutine could cause spurious idle calls
27 in cede_notself.
28
293.3 Sat Jan 6 03:45:00 CET 2007
30 - implement $coro->on_destroy.
31 - Coro::Event blocking semantics have been changed,
32 documented and - hopefully - improved.
33 - fix nice adding, not subtracting, from priority.
34 - fix ->prio and api_is_ready (patch by Mark Hinds).
35 - fixed an assert ("... == prev__cctx->idle_te")
36 that could errronously trigger.
37 - fix various large and small memleaks.
38 - use a (hopefully) more stable cancel implementation
39 that immediately frees the coroutine data.
40 - cede/cede_notself return a status now.
41 - added Coro::guard function.
42 - added a global coroutine pool for jobs (on my machine,
43 I can create and execute 48k simple coros/s with async,
44 and 128k coros with async_pool).
45 - Coro::AIO now uses the coroutine priority as io priority.
46
473.2 Fri Dec 22 05:07:09 CET 2006
48 - improve portability to slightly older perls.
49 - use cleaner coroutine destruction.
50 - simplify configuration for users.
51 - optionally (unrecommended) prefer perl functions over
52 their coro replacements.
53
543.11 Tue Dec 5 13:11:24 CET 2006
55 - fixed some bogus assert's, but as perl.h disables assert even
56 without NDEBUG (thank you very much), not too many people should
57 notice (that did include myself). Andreas König noticed, though :)
58 - do not save/restore PL_sortcxix on >= 5.9.x, it doesn't seem to have
59 it. Also noticed by Andreas König :)
60 - save/restore tainted status.
61 - verified to pass the testsuite on my 5.9.5.
62
633.1 Mon Dec 4 23:03:40 CET 2006
64 - INCOMPATIBLE CHANGE: $/ is now per-coroutine (but slow).
65 - incompatible change: transfer flags are now per-state.
66 - give a better error message on deadlock.
67 - document Coro::nready.
68 - enhanced testsuite.
69
703.01 Sun Dec 3 23:47:42 CET 2006
71 - forgot to include Coro::Timer.
72
733.0 Sun Dec 3 22:57:25 CET 2006
74 - the "FINALLY COMPLETELY STABLE" release (coming soon:
75 the "FAMOUS LAST WORDS" release).
76 - implement a new stack sharing algorithm, which uses a stack
77 pool (size currently hardcoded).
78 - make stack sharing mandatory (it no longer uses a heuristic).
79 - eval/die no longer cause weird problems under heavy use.
80 - Coro::Event could cause livelocks if it was used but
81 no Coro::Event watchers were used.
82 - Coro::Event now uses asynccheck as prepare does not
83 check for changed watchers.
84 - Coro::Event allows multiple waiting coros and will wake up one
85 per event.
86 - Coro::Event should be cleaner and more efficient now.
87 - new utility function Coro::unblock_sub.
88 - document the sad fact that Event is no longer reentrant.
89 - putting a coroutine into the ready queue twice could under
90 some circumstances lead to stack corruption.
91 - minor incompatibility: subclassing Coro::State is supported
92 directly now without going through a _coro_state member.
93 - state/coro switching is much faster now.
94 - very minor optimisations and code/documentation cleanup.
95 - avoid problems due to compiler inlining.
96 - removed timers from Coro::Timer -> use AnyEvent instead.
97 - replaced Coro::idle coroutine by (cleaner) idle handler.
98 - updated to newest libcoro.
99 - implement enhanced support for valgrind.
100 - implement is_ready and return value for ready.
101 - removed Coro::Cont, it was a misnomer (it's generators),
102 and it was somewhat annoying to get it right. Will come back
103 if somebody asks for it :)
104 - many ->wait methods and Coro::Event could return
105 spuriously without the event having happened.
106
1072.5 Tue Nov 7 12:22:33 CET 2006
108 - made Coro::Util, Coro::Select, Coro::Handle and Coro::Socket
109 use AnyEvent, moved them to Coro/.
110 - added Coro::LWP which contains all the uglyness required to
111 make LWP non-blocking.
112 - should work with perl 5.9.x now (Andreas König made me do it).
113 - fixed another bug in Coro::Select when the timeout was undef.
114 - reuse PL_start_env for all coros, saves some memory per coroutine.
115 - manage PL_top_env differently, hopefully to avoid panic: top_env.
116 - timeout argument was not properly used in Coro::Socket.
117 - allow limited forms of subclassing in Coro::Handle/Coro::Socket.
118 - emulate undocumented(!) functionality of IO::Socket required
119 by LWP(!!).
120 - updated eg/lwp to work with newer lwp's.
121 - remove "FATAL: uncaught exception" prefix. Coroutines that die
122 kill the whole process, just as exceptions in the main "coroutine"
123 did already.
124
1252.1 Wed Nov 1 23:01:13 CET 2006
126 - fix a long-standing bug in Coro::Select where select with
127 zero timeout would instead change the current default filehandle.
128 - use a simpler and hopefully more robust way to clone padlists
129 (uses less memory and a perl function instead of our own).
130 - coro can now create a stack guard area on many architectures.
131 - Coro::AIO properly reexports additional functions from IO::AIO.
132 - updated libcoro with a workaround for OS X,
133 pach and testing by Michael Schwern.
134
1352.0 Tue Oct 24 05:47:17 CEST 2006
136 - support additional aio requests in Coro::AIO.
137
1381.9 (never properly released due to a glitch)
139
1401.8 Thu Feb 2 00:59:06 CET 2006
141 - applied suggested patch by SAMV to avoid problems during stupid
142 mark & sweep gc run.
143 - applied patch by Scott Walters for 5.9.3 compatibility.
144
1451.7 Tue Dec 27 01:41:58 CET 2005
146 - added Coro::AIO, a thin wrapper around IO::AIO.
147 - improved Makefile.PL explanations.
148
1491.6 Mon Dec 12 21:30:05 CET 2005
150 - additionally save PM_curpm might fix as-of-yet
151 unseen problems with regex matches being attributed to the
152 wrong package.
153 - add t/10_bugs.t, which currently checks against imho broken
154 perls that use 0.26MB of stack space per Perl_magic_get
155 invocation (newer linuxes) as opposed to the 0.0002MB perl
156 normally uses.
157 - make stacksize configurable for the ultimate debian experience.
158
1591.51 Mon Dec 12 18:48:36 CET 2005
160 - remove debugging warn() accidentally left in Coro::Select.
161
1621.5 Tue Nov 29 13:32:44 CET 2005
163 - use Coro::Event inside Coro::Select to avoid spurious deadlocks.
164 - fix Coro::Select 'select'.
165 - strict'ify some modules.
166 - libcoro errornously restored the SIGUSR2 handler to SIGUSR1.
167 - use XSLoader in selected modules.
168 - remove some 5.6 compatibility cruft.
2 169
31.4 Tue Sep 6 00:11:05 CEST 2005 1701.4 Tue Sep 6 00:11:05 CEST 2005
4 - libcoro did not take into account the trampoline on amd64, 171 - libcoro did not take into account the trampoline on amd64,
5 when 'l'inux method was used. 172 when 'l'inux method was used.
6 173

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines