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

Comparing Coro/Changes (file contents):
Revision 1.428 by root, Tue Nov 18 08:59:45 2008 UTC vs.
Revision 1.550 by root, Sat Feb 19 05:42:45 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: testcancel 7 - add a big "Coro thread life cycle" section to "man Coro".
9TODO: guard as SLF - to avoid deadlocks. 8 - try a tentative workaround against the breakage that 5.13 has
10 - avoid deadlock in Coro::Channel if maxsize == 1 (Richard Hundt). 9 introduced without depreciation period. sigh.
11 10
125.0 115.36 Sun Feb 13 05:33:41 CET 2011
12 - automatically load Coro::Channel, Coro::RWLock, Coro::Semaphore,
13 Coro::SemaphoreSet, Coro::Signal and Coro::Specific on first "new"
14 method call.
15 - undocument Coro::Timer::sleep and obsolete whole module.
16 - optimise Coro::Timer::timeout memory and cpu usage.
17 - slightly updated Coro::Intro for recent changes.
18 - do not initialise PL_dirty anymore.
19
205.25 Thu Nov 11 01:08:39 CET 2010
21 - try a different approach on netbsd - netbsd 5 finally has marginally
22 working pthreads, but still broken ucontext/sigaltstack.
23 - openbsd 4.8 finally got their act together, Coro works out of the box
24 with asm, setjmp and pthreads (no change, just informational).
25
265.24 Sat Oct 23 11:27:12 CEST 2010
27 - port to the EV 4.0 API.
28 - work around bugs in mingw32, making strawberry perl work
29 out of the box.
30 - correctly modify Coro::AIO function prototypes
31 so that they reflect the "no optional parameters" rule.
32 - "ported" libcoro to C++.
33
345.23 Mon May 17 18:50:42 CEST 2010
35 - be more resistant to ordering changes when initialising
36 Coro::AnyEvent, Coro::EV and Coro::Event (reported by Matthias
37 Waldorf).
38 - document that perl 5.12 deliberately removed support for cloning.
39
405.22 Wed Apr 14 03:55:35 CEST 2010
41 - correctly return udnef on errors in Coro::Handle::read/write
42 (testcase by Marc Mims).
43 - convert Coro::Util into a "perl compatibility wrapper" - the functions
44 are less useful now, but are drop-in replacements for existing
45 functions, listing better alternatives in the documentation. This also
46 fixes a bug in Coro::LWP which naively substituted Socket::inet_aton
47 with Coro::Util::inet_aton.
48 - do not override $Coro::idle unconditionally in Coro.pm, as other
49 modules could have provided their own idle coro already
50 (for exmaple, Coro::AnyEvent).
51 - fix Coro::Util::gethost* functions.
52 - Coro::Timer corretcly exports it's symbols (reported by Hideki Yamamura).
53
545.21 Wed Dec 16 07:19:51 CET 2009
55 - automatically load Coro::AnyEvent when AnyEvent and Coro are used
56 together.
57 - add some examples on how to combine other event loops with Coro in
58 Coro::AnyEvent, and how to run it (and not to block). Seems to be
59 the most common source of confusion.
60 - try to catch people naively blocking in an event callback.
61 - work around the perl filehandle bug issue in conjunction with
62 older common::sense (as indirectly pointed out by ZSystem).
63 - clarify the "not from signal handlers" section.
64
655.2 Sun Oct 4 14:54:24 CEST 2009
66 - Coro::Storable destroyed the prototypes of the functions it wrapped.
67 - export rouse_cb and rouse_wait by default now.
68 - fix various prototype mismatches in Coro::AnyEvent and Coro::Handle.
69 - new method $state->swap_sv.
70 - added section on "windows process emulation" to the manpage,
71 after a not-so-fruitful (nor-friendly) "discussion" with chip
72 salzenberg (discussion implies arguments, but his only arguments
73 were ad-hominems, one wonders why he started it in the first
74 place). I hope this explains it well enough for him to understand,
75 and maybe well enough for others to understand.
76 - use common::sense everywhere now.
77 - idle callbacks are no longer supported, use idle coros instead.
78 - print a thread listing when a deadlock is detected.
79
805.17 Sat Aug 22 23:09:31 CEST 2009
81 - work around a bug in the perl debugger causing crashes
82 when running under the debugger by marking _pool_handler as nodebug.
83 - speed up Coro::async considerably.
84 - try some hacks to get netbsd to work "more often" - their broken
85 setjmp/longjmp, ucontext *and* phtreads are really hard on Coro.
86 - convert Coro to AE 5.0 API.
87
885.162 Tue Jul 28 04:04:03 CEST 2009
89 - perl 5.8.2 is now minimum requirement.
90 - skip t/19_handle.t on broken windows perls.
91
925.161 Wed Jul 22 04:47:38 CEST 2009
93 - Coro::AnyEvent::poll could have a different prototype when EV was
94 used as backend (analyzed by Tatsuhiko Miyagawa).
95 - Coro::AnyEvent errornously initialised the event loop when loaded,
96 not on demand.
97 - try to workaround rare */t/01_unblock.t failures.
98
995.16 Tue Jul 21 01:44:37 CEST 2009
100 - Coro::AnyEvent failed to hook into the event loop
101 when no threads had been readied between detecting
102 the event loop and actually running it.
103 - considerably speed up Coro::Select by taking avdantage
104 of AnyEvent > 4.8 and some other optimisations.
105 - implement paragraph readline mode in Coro::Handle
106 (based on patches by Zsbán Ambrus).
107 - replace WSAEINPROGRESS by WSAEWOULDBLOCK (reported
108 and analyzed by Yasuhiro MATSUMOTO).
109 - clarified libcoro license and copyright.
110 - someone stole my EXTRA_META!!!
111 - implement Coro::Select::patch_pp_sselect and it's brother,
112 for hardcode select overriding.
113
1145.151 Mon Jul 6 05:41:57 CEST 2009
115 - backport to windows process emulation code again (patch by
116 Yasuhiro MATSUMOTO).
117 - slightly update Coro::MakeMaker.
118
1195.15 Tue Jun 30 10:28:06 CEST 2009
120 - deprecate Coro::Socket, document how to get ipv6 support via
121 AnyEvent::Socket instead.
122 - implement signal->wait ($cb) interface, similar to semaphores.
123 - work around SvOK not supporting getmagic, so we have to getmagic
124 to test for undef :/ (reported by Matthias Waldorf).
125 - load Coro::AnyEvent in all modules using AnyEvent.
126 - work around perl corrupting our internal data structures,
127 reported by Tokuhiro Matsuno.
128 - enable per-coroutine real and cpu time gathering
129 (Coro::State::enable_times).
130
1315.14 Wed Jun 24 01:37:48 CEST 2009
132 - provide explicit functions to "cede" to the event loop in Coro::AnyEvent,
133 as this seems to have been a difficult concept (poll, sleep, idle,
134 idle_upto).
135 - add Coro::AnyEvent::readable/writable functions.
136 - clarify Coro::EV/Event/AnyEvent manpages.
137 - free per-thread global scalars in the thread calling ->cancel, to
138 avoid crashes when $_, $@ etc., are magical but some of those
139 had already been freed.
140 - "unexperimentalise" the callback interface for Coro::Semaphore.
141 - speed up ready queue management/context switching by using a linked
142 list instead of an array (~5-10%).
143 - implement "watch" command in Coro::Debug shells.
144 - for fun, implement time-slicing as an example in the manpage.
145 - if AnyEvent detects EV or Event, but we don't have Coro::EV or
146 Coro::Event, use the normal AnyEvent handling instead of dieing
147 (the same is true for Coro::Handle).
148 - properly document Coro::EV::timed_io_once.
149 - avoid unneccessary ->cancel calls in Coro::Handle.
150 - maybe make it work on mingw32 with win32 backend
151 (based on patch by Yasuhiro Matsumoto).
152
1535.132 Fri May 29 09:00:39 CEST 2009
154 - do not keep a reference to the argument itself in
155 Coro::Semaphore::guard, as it could change later.
156 - support SO_RCVBUF/SO_SNDBUF nonstandard Coro::Socket options,
157 should support a prepare callback.
158
1595.131 Mon Mar 16 23:20:37 CET 2009
160 - implement and document Coro->suspend, Coro->resume.
161 - fix Coro::Select implementation to not (often) close
162 the passed file descriptors (testcase provided by pippijn).
163
1645.13 Mon Dec 15 21:51:42 CET 2008
165 - EXPERIMENTAL: implement dynamic winds (on_enter/on_leave).
166 - don't set diehook to C<undef> but instead to NULL, to avoid
167 spurious warnings.
168 - fix a lot of bugs in Coro::SemaphoreSet.
169 - Coro::SemaphoreSet will less often create a semaphore needlessly.
170 - add Coro::SemaphoreSet::count and wait methods.
171 - take advantage of the new Guard module.
172 - deprecate Coro::guard.
173 - try to fix the dreaded 01_unblock tests once more. I hate it when
174 testsuites need more fixing than the code they are supposed to test.
175 - croak in more cases when a required callback isn't resolvable.
176 - fix some minor issues in Coro::State->call/eval.
177 - use current coroutine context instead of a temporary one
178 when temporarily switching to another coroutine.
179 - do not call C-level on_destroy handlers during global destruction,
180 to avoid needless segfaults.
181
1825.12 Sun Dec 7 13:30:38 CET 2008
183 - add default config for MirOS, which seems to be a bug-to-bug
184 compatible fork of openbsd ("world domination by releasing
185 openbsd cvs before the openbsd folks do it" or so :).
186 - free_padlist did destroy the names pad, not good, but didn't
187 seem to bother perl - this could fix issues such as eval ""
188 inside a function called from multiple coroutines.
189 - use a different method to detect destruction time.
190 - be more careful when freeing padlists just before global
191 destruction.
192 - fixed and expanded the call/cc example.
193 - renamed _terminate to _coro_run.
194 - new method Coro::Channel->shutdown.
195 - try pthreads on openbsd <4.4 (broken sigaltstack, will
196 pthreads fare better?).
197 - be less picky about destroying "a" running coroutine.
198
1995.11 Tue Nov 25 21:49:05 CET 2008
200 - DEBUGGING in 5.10.0 is a mess: it constantly flags perfectly
201 working code with failed assertions, introducing more bugs than
202 it fixes, requiring elaborate workarounds :(
203
2045.1 Mon Nov 24 08:54:59 CET 2008
205 - wrote a small introductory tutorial - Coro::Intro.
206 - convert Coro::Timer, Coro::Select and Coro::Util to rouse API.
207 - Coro::Select did errornously dup the file descriptors
208 and didn't work with all AnyEvent backends.
209 - Coro::Select wasn't imported correctly form Coro::LWP, causing blocking
210 LWP data transfers.
211 - disassociate c contexts from coro objects - this is agruably more
212 correct, but mostly allows sharing of cctxs between coro and state
213 objects, for added memory savings and speed increases.
214 - bumped $Coro::POOL_RSS up to 32kb by default.
215 - no longer set the optype to OP_CUSTOM, as B::* understandably
216 doesn't like this very much (and we *are* a type of entersub).
217 - implement state cloning, just to prove that call/cc can be done.
218 - automatically load Coro::AnyEvent in Coro::Handle.
219 - wrap ->cancel calls in eval inside Coro::Handle as EV watchers
220 do not have this method (and don't need it either).
221 - speed up generic anyevent methods in Coro::Handle by using rouse
222 callbacks.
223 - allow coroutines in $Coro::IDLE, speeding up Coro::AnyEvent and
224 others. It also makes the debugger happier, as you can trace
225 through the idle threads now.
226 - add comppad_name* and hints ($^H, %^H) to per-thread variables.
227 - eg/event was pretty broken.
228 - better 5.8.6 compatibility.
229
2305.0 Thu Nov 20 10:35:05 CET 2008
13 - NEW ARCHITECTURE: use the latest 4.x version if you experience 231 - NEW ARCHITECTURE: use the latest 4.x version if you experience
14 stability issues. 232 stability issues.
15 - bump API version to 7 - all dependents must be recompiled. 233 - bump API version to 7 - all dependents must be recompiled.
234 - removed timed_* functions - they were not being used anyways
235 and should be replaced by a more generic mechanism -
236 and were annoying to support anyways :)
237 - removed SemaphoreSet's waiter method - use sem method instead.
16 - Coro::Semaphore->adjust didn't correctly wake up enough waiters. 238 - Coro::Semaphore->adjust didn't correctly wake up enough waiters.
17 - async_pool did free a scalar value twice 239 - async_pool did free a scalar value twice
18 ("Attempt to unreference..."). 240 ("Attempt to unreference...").
19 - fix a longstanding bug where calling terminate on a coro that 241 - fix a longstanding bug where calling terminate on a coro that
20 was waiting for a semaphore that was just becoming available 242 was waiting for a semaphore that was just becoming available
21 would cause a deadlock (semaphore would get into a state where 243 would cause a deadlock (semaphore would get into a state where
22 it was available but waiters were waiting(. 244 it was available but waiters were still blocked).
245 - calling throw on a coroutine that is waiting for a semaphore will
246 no longer make it acquire the semaphore (and thus leak a count).
23 - perl's process emulation is now not even theoretically supported 247 - perl's process emulation is now not even theoretically supported
24 anymore. 248 anymore.
249 - new functions Coro::rouse_cb and Coro::rouse_wait for easier
250 conversion of callback-style to blocking-style.
251 - new methods $coro->schedule_to and ->cede_to, to specifically
252 schedule or cede to a specific coroutine.
253 - new function Coro::Semaphore::wait.
25 - use named constants in Coro::Channel (Richard Hundt). 254 - use named constants in Coro::Channel (Richard Hundt).
26 - directly patch the entersub opcode calling SLF functions (cede, 255 - directly patch the entersub opcode calling SLF functions (cede,
27 transfer and so on). this does speed up context switching, but 256 transfer and so on). this does speed up context switching, but
28 more importanly, it frees us from the hardcoded behaviour of 257 more importanly, it frees us from the hardcoded behaviour of
29 entersub, so we might actually be able to return something from 258 entersub, so we might actually be able to return something from
… …
33 as "threaded" perls are running at half speed anyways). 262 as "threaded" perls are running at half speed anyways).
34 - implement execute_slf (schedule-like-function) interface that makes 263 - implement execute_slf (schedule-like-function) interface that makes
35 it possible to implement schedule-like-functions in XS. 264 it possible to implement schedule-like-functions in XS.
36 - use new SLF interface to massively speed up Coro::EV by roughly a 265 - use new SLF interface to massively speed up Coro::EV by roughly a
37 factor of two. 266 factor of two.
38 - removed timed_down and timed_guard functions - they were not being
39 used anyways and should be replaced by a more generic mechanism -
40 and were annoying to support anyways :)
41 - used new SLF interface to massively speed up Coro::Semaphore by a 267 - used new SLF interface to massively speed up Coro::Semaphore by a
42 factor of three. 268 factor of three.
43 - used new SLF interface to speed up Coro::AIO by roughly a factor of 269 - used new SLF interface to speed up Coro::AIO by roughly a factor of
44 four and reduce its memory usage considerably. 270 four and reduce its memory usage considerably.
271 - implement Coro::SemaphoreSet purely in terms of Coro::Semaphore,
272 for a nice speedup and vastly more correct behaviour. Also implement
273 a new method "sem" to get at the underlying semaphore object.
274 - implement Coro::Channel in terms of Coro::Semaphore, for a moderate
275 (in comparison) 20-40% speedup.
276 - used new SLF interface to reimplement Coro::Signal gaining
277 some unknown (because I was too lazy), but certain, speedup, and also
278 making signals reliable for the first time.
279 - used new SLF interface and other optimisations to speed up async_pool
280 by a factor of two. It also doesn't rely on perl's exception mechanism
281 to exit anymore. The overhead for terminating an async_pool, coro over
282 a normal async is now very small.
283 - sped up coroutine creation/destruction by 40%.
45 - forgot to include Coro/libcoro/README in the dist for all these years. 284 - forgot to include Coro/libcoro/README in the dist for all these years.
46 - work around a freebsd pthreads bug (manual testcancel is required as 285 - work around a freebsd pthreads bug (manual testcancel is required as
47 pthread_cond_wait isn't a cancellation point on freebsd). 286 pthread_cond_wait isn't a cancellation point on freebsd).
287 - use new rouse functions to speed up and simplify Coro::BDB.
288 - make "prefer perl native functions" work with threaded perls.
289 - condense Coro::Debug ps output, hint at v and w flags.
290 - (libcoro) lots of minor cleanups and portability improvements.
291
2924.914 Wed Nov 19 12:54:18 CET 2008
293 - fix a disastrous bug in the readline optimisation
294 introduced in 4.801.
48 295
494.913 Sat Nov 15 07:58:28 CET 2008 2964.913 Sat Nov 15 07:58:28 CET 2008
50 - async_pool did free a scalar value twice 297 - async_pool did free a scalar value twice
51 ("Attempt to unreference..."). 298 ("Attempt to unreference...").
52 299

Diff Legend

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