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

Comparing Coro/Changes (file contents):
Revision 1.136 by root, Fri Nov 3 00:56:02 2006 UTC vs.
Revision 1.217 by root, Wed Apr 11 03:05:15 2007 UTC

1Revision history for Perl extension Coro. 1Revision history for Perl extension Coro.
2 2
3TODO: possibly convert some of the Event modules into AnyEvent users. 3TODO: better (earlier) error messages on broken transfer's
4TODO: destroy gets only called on cancel: fix(?)
4 5
5 - should compile with perl 5.9.x now. 6 - added some bugfixes to get eg/myhttpd working again.
7
83.55 Sun Mar 25 01:20:47 CET 2007
9 - add SAVE_DEFFH to save the default filehandle and enable
10 it by default.
11 - finally move socket-operations from Coro::Socket to Coro::Handle
12 to be able to unblock foreign sockets.
13 - add Coro::State::save_also and guarded_save.
14 - add count accessor to Coro::Semaphore.
15 - add Coro::State::cctx_stacksize.
16 - just for the fun of it, do not rely on implicit context,
17 which can dramatically improve performance, but people
18 using windows-process-emulation perls do not care much about
19 performance.
20
213.51 Sun Mar 4 14:18:04 CET 2007
22 - fixed a problem when you weakened references to Coro::State's
23 (patch by Yuval Kogman).
24
253.501 Wed Feb 28 12:44:07 CET 2007
26 - rename some global symbols as macosx from hell redefines
27 them without asking.
28
293.5 Tue Feb 13 20:22:53 CET 2007
30 - do AnyEvent model detection earlier, avoiding problems
31 caused by first using AnyEvent and later Coro::Event.
32 - implement and document Coro::Event event objects.
33 - fix a potential problem in Coro::Event causing crashes.
34 - initialise PL_comppad when creating a new coroutine,
35 avoids crashes on early coro destruction.
36
373.41 Mon Jan 22 19:19:49 CET 2007
38 - readline on Coro::Handle did not support undefined $/,
39 nor did it deliver partial lines on EOF or error.
40 - implement malloc fallback for stack allocation because
41 stupid broken idiotic OSX has a stupid broken
42 idiotic fits-the-whole-os mmap "implementation" and
43 my dick feels longer if Coro is portable even to
44 obsolete platforms.
45
463.4 Fri Jan 19 21:52:54 CET 2007
47 - remove t/09_timer.t, as it isn't really testing much
48 but was rather flaky in practise.
49 - async_pool coro would keep arguments and callback alive until
50 it was reused.
51 - cancellation of a coroutine could cause spurious idle calls
52 in cede_notself.
53
543.3 Sat Jan 6 03:45:00 CET 2007
55 - implement $coro->on_destroy.
56 - Coro::Event blocking semantics have been changed,
57 documented and - hopefully - improved.
58 - fix nice adding, not subtracting, from priority.
59 - fix ->prio and api_is_ready (patch by Mark Hinds).
60 - fixed an assert ("... == prev__cctx->idle_te")
61 that could errronously trigger.
62 - fix various large and small memleaks.
63 - use a (hopefully) more stable cancel implementation
64 that immediately frees the coroutine data.
65 - cede/cede_notself return a status now.
66 - added Coro::guard function.
67 - added a global coroutine pool for jobs (on my machine,
68 I can create and execute 48k simple coros/s with async,
69 and 128k coros with async_pool).
70 - Coro::AIO now uses the coroutine priority as io priority.
71
723.2 Fri Dec 22 05:07:09 CET 2006
73 - improve portability to slightly older perls.
74 - use cleaner coroutine destruction.
75 - simplify configuration for users.
76 - optionally (unrecommended) prefer perl functions over
77 their coro replacements.
78
793.11 Tue Dec 5 13:11:24 CET 2006
80 - fixed some bogus assert's, but as perl.h disables assert even
81 without NDEBUG (thank you very much), not too many people should
82 notice (that did include myself). Andreas König noticed, though :)
83 - do not save/restore PL_sortcxix on >= 5.9.x, it doesn't seem to have
84 it. Also noticed by Andreas König :)
85 - save/restore tainted status.
86 - verified to pass the testsuite on my 5.9.5.
87
883.1 Mon Dec 4 23:03:40 CET 2006
89 - INCOMPATIBLE CHANGE: $/ is now per-coroutine (but slow).
90 - incompatible change: transfer flags are now per-state.
91 - give a better error message on deadlock.
92 - document Coro::nready.
93 - enhanced testsuite.
94
953.01 Sun Dec 3 23:47:42 CET 2006
96 - forgot to include Coro::Timer.
97
983.0 Sun Dec 3 22:57:25 CET 2006
99 - the "FINALLY COMPLETELY STABLE" release (coming soon:
100 the "FAMOUS LAST WORDS" release).
101 - implement a new stack sharing algorithm, which uses a stack
102 pool (size currently hardcoded).
103 - make stack sharing mandatory (it no longer uses a heuristic).
104 - eval/die no longer cause weird problems under heavy use.
105 - Coro::Event could cause livelocks if it was used but
106 no Coro::Event watchers were used.
107 - Coro::Event now uses asynccheck as prepare does not
108 check for changed watchers.
109 - Coro::Event allows multiple waiting coros and will wake up one
110 per event.
111 - Coro::Event should be cleaner and more efficient now.
112 - new utility function Coro::unblock_sub.
113 - document the sad fact that Event is no longer reentrant.
114 - putting a coroutine into the ready queue twice could under
115 some circumstances lead to stack corruption.
116 - minor incompatibility: subclassing Coro::State is supported
117 directly now without going through a _coro_state member.
118 - state/coro switching is much faster now.
119 - very minor optimisations and code/documentation cleanup.
120 - avoid problems due to compiler inlining.
121 - removed timers from Coro::Timer -> use AnyEvent instead.
122 - replaced Coro::idle coroutine by (cleaner) idle handler.
123 - updated to newest libcoro.
124 - implement enhanced support for valgrind.
125 - implement is_ready and return value for ready.
126 - removed Coro::Cont, it was a misnomer (it's generators),
127 and it was somewhat annoying to get it right. Will come back
128 if somebody asks for it :)
129 - many ->wait methods and Coro::Event could return
130 spuriously without the event having happened.
131
1322.5 Tue Nov 7 12:22:33 CET 2006
133 - made Coro::Util, Coro::Select, Coro::Handle and Coro::Socket
134 use AnyEvent, moved them to Coro/.
135 - added Coro::LWP which contains all the uglyness required to
136 make LWP non-blocking.
137 - should work with perl 5.9.x now (Andreas König made me do it).
138 - fixed another bug in Coro::Select when the timeout was undef.
6 - reuse PL_start_env for all coros, saves some memory per coroutine. 139 - reuse PL_start_env for all coros, saves some memory per coroutine.
7 - manage PL_top_env differently, hopefully to avoid panic: top_env. 140 - manage PL_top_env differently, hopefully to avoid panic: top_env.
141 - timeout argument was not properly used in Coro::Socket.
142 - allow limited forms of subclassing in Coro::Handle/Coro::Socket.
143 - emulate undocumented(!) functionality of IO::Socket required
144 by LWP(!!).
145 - updated eg/lwp to work with newer lwp's.
8 - remove "FATAL: uncaught exception" prefix: coroutines that die 146 - remove "FATAL: uncaught exception" prefix. Coroutines that die
9 kill the whole process, just as exceptions in the main coroutine 147 kill the whole process, just as exceptions in the main "coroutine"
10 do. 148 did already.
11 149
122.1 Wed Nov 1 23:01:13 CET 2006 1502.1 Wed Nov 1 23:01:13 CET 2006
13 - fix a long-standing bug in Coro::Select where select with 151 - fix a long-standing bug in Coro::Select where select with
14 zero timeout would instead change the current default filehandle. 152 zero timeout would instead change the current default filehandle.
15 - use a simpler and hopefully more robust way to clone padlists 153 - use a simpler and hopefully more robust way to clone padlists

Diff Legend

– Removed lines
+ Added lines
< Changed lines
> Changed lines