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

Comparing Coro/Changes (file contents):
Revision 1.480 by root, Mon Dec 15 00:28:30 2008 UTC vs.
Revision 1.576 by root, Sun Jul 3 10:51:40 2011 UTC

1Revision history for Perl extension Coro. 1Revision history for Perl extension Coro.
2 2
3TODO: should explore PerlIO::coroaio (perl leaks like hell). 3TODO: should explore PerlIO::coroaio (perl leaks like hell).
4TODO: maybe implement a default message channel, very much like Erlang's 4TODO: channel->maxsize(newsize)?
5 actor model (which is cool in a lot of important aspects (failures!), 5TODO: http://www.microsoft.com/msj/archive/s2ce.aspx
6 but very lacking in others (higher level ipc)).
7 6
8TODO: dynamic wind, test and document 76.01 Sun Jul 3 12:31:14 CEST 2011
95.13 8 - workarounds are good, but the test for whether pthreads are used
9 was not good. thisone should be better.
10 - check differently whether gcc generates cfi instructions itself.
11
126.0 Wed Jun 29 19:43:35 CEST 2011
13 - INCOMPATIBLE CHANGE: unreferenced coro objects will now be
14 destroyed and cleaned up automatically (e.g. async { schedule }).
15 - implement a JIT compiler for part of the thread switch code,
16 which gives a 50% speed improvement on threaded perls, and
17 about 4% on non-thraeded perls (so threaded perls now finally
18 reach about half the speed of non-threaded perls).
19 - slightly modernise Coro::Intro, add section about rouse functions.
20 - avoid DEFSV and ERRSV, giving another 10% improvement
21 in thread switching.
22 - Coro::State->is_destroyed is now called is_zombie.
23 - implement a Coro->safe_cancel method that might fail, but
24 cancels in a "safer" way if it succeeds.
25 - add preliminary support for DEBUGGING perls.
26 - get rid of two hash-accesses when initialising a new Coro - this
27 speeds up coro creation by almost a factor of two.
28 - croak when a coro that is being cancelled tries to block
29 (e.g. while executing a guard block), instead of crashing or
30 deadlocking.
31 - use a more robust and also faster method to identify Coro::State
32 objects - speeds up everything a bit.
33 - implement Coro->cancel in XS for a 20% speed improvement, and to
34 be able to implement mutual cancellation.
35 - speed up context switches by a percent or two by more efficiently
36 allocating context stack entries.
37 - implement Coro->join and Coro->on_destroy in XS for a speedup and
38 a reduction in memory use.
39 - cancelling a coro while it itself is cancelling another coro is
40 now supported and working, instead of triggering an assertion.
41 - be a bit more crash-resistant when calling (buggy) on_destroy
42 callbacks (best effort).
43 - move on_destroy into the slf_frame, to allow extension slf
44 functions to have destructors.
45 - get rid if coro refcounting - simply crash in other interpreter
46 threads by nulling the pointers on clone.
47 - simplify warn/die hook handling when loading Coro - the convoluted
48 logic seems to be no longer neccessary.
49 - use libecb instead of our own home-grown gcc hacks.
50 - document alternatives to Coro::LWP. Please use them :)
51 - work around another mindless idiotic NEEDLESS bug in openbsd/mirbsds
52 sigaltstack. Really. wine suffers from it, erlang suffers from it,
53 and it's known since at least 2006.
54
555.372 Wed Feb 23 06:14:30 CET 2011
56 - apparently mingw doesn't provide a working gettimeofday, try to
57 work around that by relying on Time::HiRes (indirectly brought to
58 my attention by Max Maischein).
59 - fix some portability issues when Time::HiRes was used.
60
615.371 Mon Feb 21 14:36:08 CET 2011
62 - backport to windows process emulation code again.
63
645.37 Sat Feb 19 07:49:44 CET 2011
65 - add a big "Coro thread life cycle" section to "man Coro".
66 - try a tentative workaround against the breakage that 5.13 has
67 introduced without depreciation period. sigh.
68 - no longer use Time::HiRes if gettimeofday is available, which
69 saves quite a lot of memory.
70
715.36 Sun Feb 13 05:33:41 CET 2011
72 - automatically load Coro::Channel, Coro::RWLock, Coro::Semaphore,
73 Coro::SemaphoreSet, Coro::Signal and Coro::Specific on first "new"
74 method call.
75 - undocument Coro::Timer::sleep and obsolete whole module.
76 - optimise Coro::Timer::timeout memory and cpu usage.
77 - slightly updated Coro::Intro for recent changes.
78 - do not initialise PL_dirty anymore.
79
805.25 Thu Nov 11 01:08:39 CET 2010
81 - try a different approach on netbsd - netbsd 5 finally has marginally
82 working pthreads, but still broken ucontext/sigaltstack.
83 - openbsd 4.8 finally got their act together, Coro works out of the box
84 with asm, setjmp and pthreads (no change, just informational).
85
865.24 Sat Oct 23 11:27:12 CEST 2010
87 - port to the EV 4.0 API.
88 - work around bugs in mingw32, making strawberry perl work
89 out of the box.
90 - correctly modify Coro::AIO function prototypes
91 so that they reflect the "no optional parameters" rule.
92 - "ported" libcoro to C++.
93
945.23 Mon May 17 18:50:42 CEST 2010
95 - be more resistant to ordering changes when initialising
96 Coro::AnyEvent, Coro::EV and Coro::Event (reported by Matthias
97 Waldorf).
98 - document that perl 5.12 deliberately removed support for cloning.
99
1005.22 Wed Apr 14 03:55:35 CEST 2010
101 - correctly return udnef on errors in Coro::Handle::read/write
102 (testcase by Marc Mims).
103 - convert Coro::Util into a "perl compatibility wrapper" - the functions
104 are less useful now, but are drop-in replacements for existing
105 functions, listing better alternatives in the documentation. This also
106 fixes a bug in Coro::LWP which naively substituted Socket::inet_aton
107 with Coro::Util::inet_aton.
108 - do not override $Coro::idle unconditionally in Coro.pm, as other
109 modules could have provided their own idle coro already
110 (for exmaple, Coro::AnyEvent).
111 - fix Coro::Util::gethost* functions.
112 - Coro::Timer corretcly exports it's symbols (reported by Hideki Yamamura).
113
1145.21 Wed Dec 16 07:19:51 CET 2009
115 - automatically load Coro::AnyEvent when AnyEvent and Coro are used
116 together.
117 - add some examples on how to combine other event loops with Coro in
118 Coro::AnyEvent, and how to run it (and not to block). Seems to be
119 the most common source of confusion.
120 - try to catch people naively blocking in an event callback.
121 - work around the perl filehandle bug issue in conjunction with
122 older common::sense (as indirectly pointed out by ZSystem).
123 - clarify the "not from signal handlers" section.
124
1255.2 Sun Oct 4 14:54:24 CEST 2009
126 - Coro::Storable destroyed the prototypes of the functions it wrapped.
127 - export rouse_cb and rouse_wait by default now.
128 - fix various prototype mismatches in Coro::AnyEvent and Coro::Handle.
129 - new method $state->swap_sv.
130 - added section on "windows process emulation" to the manpage,
131 after a not-so-fruitful (nor-friendly) "discussion" with chip
132 salzenberg (discussion implies arguments, but his only arguments
133 were ad-hominems, one wonders why he started it in the first
134 place). I hope this explains it well enough for him to understand,
135 and maybe well enough for others to understand.
136 - use common::sense everywhere now.
137 - idle callbacks are no longer supported, use idle coros instead.
138 - print a thread listing when a deadlock is detected.
139
1405.17 Sat Aug 22 23:09:31 CEST 2009
141 - work around a bug in the perl debugger causing crashes
142 when running under the debugger by marking _pool_handler as nodebug.
143 - speed up Coro::async considerably.
144 - try some hacks to get netbsd to work "more often" - their broken
145 setjmp/longjmp, ucontext *and* phtreads are really hard on Coro.
146 - convert Coro to AE 5.0 API.
147
1485.162 Tue Jul 28 04:04:03 CEST 2009
149 - perl 5.8.2 is now minimum requirement.
150 - skip t/19_handle.t on broken windows perls.
151
1525.161 Wed Jul 22 04:47:38 CEST 2009
153 - Coro::AnyEvent::poll could have a different prototype when EV was
154 used as backend (analyzed by Tatsuhiko Miyagawa).
155 - Coro::AnyEvent errornously initialised the event loop when loaded,
156 not on demand.
157 - try to workaround rare */t/01_unblock.t failures.
158
1595.16 Tue Jul 21 01:44:37 CEST 2009
160 - Coro::AnyEvent failed to hook into the event loop
161 when no threads had been readied between detecting
162 the event loop and actually running it.
163 - considerably speed up Coro::Select by taking avdantage
164 of AnyEvent > 4.8 and some other optimisations.
165 - implement paragraph readline mode in Coro::Handle
166 (based on patches by Zsbán Ambrus).
167 - replace WSAEINPROGRESS by WSAEWOULDBLOCK (reported
168 and analyzed by Yasuhiro MATSUMOTO).
169 - clarified libcoro license and copyright.
170 - someone stole my EXTRA_META!!!
171 - implement Coro::Select::patch_pp_sselect and it's brother,
172 for hardcode select overriding.
173
1745.151 Mon Jul 6 05:41:57 CEST 2009
175 - backport to windows process emulation code again (patch by
176 Yasuhiro MATSUMOTO).
177 - slightly update Coro::MakeMaker.
178
1795.15 Tue Jun 30 10:28:06 CEST 2009
180 - deprecate Coro::Socket, document how to get ipv6 support via
181 AnyEvent::Socket instead.
182 - implement signal->wait ($cb) interface, similar to semaphores.
183 - work around SvOK not supporting getmagic, so we have to getmagic
184 to test for undef :/ (reported by Matthias Waldorf).
185 - load Coro::AnyEvent in all modules using AnyEvent.
186 - work around perl corrupting our internal data structures,
187 reported by Tokuhiro Matsuno.
188 - enable per-coroutine real and cpu time gathering
189 (Coro::State::enable_times).
190
1915.14 Wed Jun 24 01:37:48 CEST 2009
192 - provide explicit functions to "cede" to the event loop in Coro::AnyEvent,
193 as this seems to have been a difficult concept (poll, sleep, idle,
194 idle_upto).
195 - add Coro::AnyEvent::readable/writable functions.
196 - clarify Coro::EV/Event/AnyEvent manpages.
197 - free per-thread global scalars in the thread calling ->cancel, to
198 avoid crashes when $_, $@ etc., are magical but some of those
199 had already been freed.
200 - "unexperimentalise" the callback interface for Coro::Semaphore.
201 - speed up ready queue management/context switching by using a linked
202 list instead of an array (~5-10%).
203 - implement "watch" command in Coro::Debug shells.
204 - for fun, implement time-slicing as an example in the manpage.
205 - if AnyEvent detects EV or Event, but we don't have Coro::EV or
206 Coro::Event, use the normal AnyEvent handling instead of dieing
207 (the same is true for Coro::Handle).
208 - properly document Coro::EV::timed_io_once.
209 - avoid unneccessary ->cancel calls in Coro::Handle.
210 - maybe make it work on mingw32 with win32 backend
211 (based on patch by Yasuhiro Matsumoto).
212
2135.132 Fri May 29 09:00:39 CEST 2009
214 - do not keep a reference to the argument itself in
215 Coro::Semaphore::guard, as it could change later.
216 - support SO_RCVBUF/SO_SNDBUF nonstandard Coro::Socket options,
217 should support a prepare callback.
218
2195.131 Mon Mar 16 23:20:37 CET 2009
220 - implement and document Coro->suspend, Coro->resume.
221 - fix Coro::Select implementation to not (often) close
222 the passed file descriptors (testcase provided by pippijn).
223
2245.13 Mon Dec 15 21:51:42 CET 2008
225 - EXPERIMENTAL: implement dynamic winds (on_enter/on_leave).
10 - don't set diehook to C<undef> but instead to NULL, to avoid 226 - don't set diehook to C<undef> but instead to NULL, to avoid
11 spurious warnings. 227 spurious warnings.
12 - fix a lot of bugs in Coro::SemaphoreSet. 228 - fix a lot of bugs in Coro::SemaphoreSet.
13 - Coro::SemaphoreSet will less often create a semaphore needlessly. 229 - Coro::SemaphoreSet will less often create a semaphore needlessly.
14 - add Coro::SemaphoreSet::count and wait methods. 230 - add Coro::SemaphoreSet::count and wait methods.
… …
18 testsuites need more fixing than the code they are supposed to test. 234 testsuites need more fixing than the code they are supposed to test.
19 - croak in more cases when a required callback isn't resolvable. 235 - croak in more cases when a required callback isn't resolvable.
20 - fix some minor issues in Coro::State->call/eval. 236 - fix some minor issues in Coro::State->call/eval.
21 - use current coroutine context instead of a temporary one 237 - use current coroutine context instead of a temporary one
22 when temporarily switching to another coroutine. 238 when temporarily switching to another coroutine.
239 - do not call C-level on_destroy handlers during global destruction,
240 to avoid needless segfaults.
23 241
245.12 Sun Dec 7 13:30:38 CET 2008 2425.12 Sun Dec 7 13:30:38 CET 2008
25 - add default config for MirOS, which seems to be a bug-to-bug 243 - add default config for MirOS, which seems to be a bug-to-bug
26 compatible fork of openbsd ("world domination by releasing 244 compatible fork of openbsd ("world domination by releasing
27 openbsd cvs before the openbsd folks do it" or so :). 245 openbsd cvs before the openbsd folks do it" or so :).

Diff Legend

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