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

Comparing Coro/Changes (file contents):
Revision 1.422 by root, Sun Nov 16 11:12:57 2008 UTC vs.
Revision 1.547 by root, Sun Feb 13 04:39:14 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: unready_all
5 actor model (which is cool in a lot of important aspects (failures!), 5TODO: myhttpd header parsing
6 but very lacking in others (higher level ipc)). 6TODO: channel->maxsize(newsize)?
7TODO: http://www.microsoft.com/msj/archive/s2ce.aspx
7 8
8TODO: testcancel 95.36 Sun Feb 13 05:33:41 CET 2011
9TODO: aio 10 - automatically load Coro::Channel, Coro::RWLock, Coro::Semaphore,
10TODO: Coro::Semaphore 11 Coro::SemaphoreSet, Coro::Signal and Coro::Specific on first "new"
11TODO: trow disturbs cleanup path inside pp_slf? 12 method call.
12TODO: throw special-case in transfer? 13 - undocument Coro::Timer::sleep and obsolete whole module.
135.0 14 - optimise Coro::Timer::timeout memory and cpu usage.
14 - NEW ARCHITECTURE: use 4.912 for the stable version. 15 - slightly updated Coro::Intro for recent changes.
16 - do not initialise PL_dirty anymore.
17
185.25 Thu Nov 11 01:08:39 CET 2010
19 - try a different approach on netbsd - netbsd 5 finally has marginally
20 working pthreads, but still broken ucontext/sigaltstack.
21 - openbsd 4.8 finally got their act together, Coro works out of the box
22 with asm, setjmp and pthreads (no change, just informational).
23
245.24 Sat Oct 23 11:27:12 CEST 2010
25 - port to the EV 4.0 API.
26 - work around bugs in mingw32, making strawberry perl work
27 out of the box.
28 - correctly modify Coro::AIO function prototypes
29 so that they reflect the "no optional parameters" rule.
30 - "ported" libcoro to C++.
31
325.23 Mon May 17 18:50:42 CEST 2010
33 - be more resistant to ordering changes when initialising
34 Coro::AnyEvent, Coro::EV and Coro::Event (reported by Matthias
35 Waldorf).
36 - document that perl 5.12 deliberately removed support for cloning.
37
385.22 Wed Apr 14 03:55:35 CEST 2010
39 - correctly return udnef on errors in Coro::Handle::read/write
40 (testcase by Marc Mims).
41 - convert Coro::Util into a "perl compatibility wrapper" - the functions
42 are less useful now, but are drop-in replacements for existing
43 functions, listing better alternatives in the documentation. This also
44 fixes a bug in Coro::LWP which naively substituted Socket::inet_aton
45 with Coro::Util::inet_aton.
46 - do not override $Coro::idle unconditionally in Coro.pm, as other
47 modules could have provided their own idle coro already
48 (for exmaple, Coro::AnyEvent).
49 - fix Coro::Util::gethost* functions.
50 - Coro::Timer corretcly exports it's symbols (reported by Hideki Yamamura).
51
525.21 Wed Dec 16 07:19:51 CET 2009
53 - automatically load Coro::AnyEvent when AnyEvent and Coro are used
54 together.
55 - add some examples on how to combine other event loops with Coro in
56 Coro::AnyEvent, and how to run it (and not to block). Seems to be
57 the most common source of confusion.
58 - try to catch people naively blocking in an event callback.
59 - work around the perl filehandle bug issue in conjunction with
60 older common::sense (as indirectly pointed out by ZSystem).
61 - clarify the "not from signal handlers" section.
62
635.2 Sun Oct 4 14:54:24 CEST 2009
64 - Coro::Storable destroyed the prototypes of the functions it wrapped.
65 - export rouse_cb and rouse_wait by default now.
66 - fix various prototype mismatches in Coro::AnyEvent and Coro::Handle.
67 - new method $state->swap_sv.
68 - added section on "windows process emulation" to the manpage,
69 after a not-so-fruitful (nor-friendly) "discussion" with chip
70 salzenberg (discussion implies arguments, but his only arguments
71 were ad-hominems, one wonders why he started it in the first
72 place). I hope this explains it well enough for him to understand,
73 and maybe well enough for others to understand.
74 - use common::sense everywhere now.
75 - idle callbacks are no longer supported, use idle coros instead.
76 - print a thread listing when a deadlock is detected.
77
785.17 Sat Aug 22 23:09:31 CEST 2009
79 - work around a bug in the perl debugger causing crashes
80 when running under the debugger by marking _pool_handler as nodebug.
81 - speed up Coro::async considerably.
82 - try some hacks to get netbsd to work "more often" - their broken
83 setjmp/longjmp, ucontext *and* phtreads are really hard on Coro.
84 - convert Coro to AE 5.0 API.
85
865.162 Tue Jul 28 04:04:03 CEST 2009
87 - perl 5.8.2 is now minimum requirement.
88 - skip t/19_handle.t on broken windows perls.
89
905.161 Wed Jul 22 04:47:38 CEST 2009
91 - Coro::AnyEvent::poll could have a different prototype when EV was
92 used as backend (analyzed by Tatsuhiko Miyagawa).
93 - Coro::AnyEvent errornously initialised the event loop when loaded,
94 not on demand.
95 - try to workaround rare */t/01_unblock.t failures.
96
975.16 Tue Jul 21 01:44:37 CEST 2009
98 - Coro::AnyEvent failed to hook into the event loop
99 when no threads had been readied between detecting
100 the event loop and actually running it.
101 - considerably speed up Coro::Select by taking avdantage
102 of AnyEvent > 4.8 and some other optimisations.
103 - implement paragraph readline mode in Coro::Handle
104 (based on patches by Zsbán Ambrus).
105 - replace WSAEINPROGRESS by WSAEWOULDBLOCK (reported
106 and analyzed by Yasuhiro MATSUMOTO).
107 - clarified libcoro license and copyright.
108 - someone stole my EXTRA_META!!!
109 - implement Coro::Select::patch_pp_sselect and it's brother,
110 for hardcode select overriding.
111
1125.151 Mon Jul 6 05:41:57 CEST 2009
113 - backport to windows process emulation code again (patch by
114 Yasuhiro MATSUMOTO).
115 - slightly update Coro::MakeMaker.
116
1175.15 Tue Jun 30 10:28:06 CEST 2009
118 - deprecate Coro::Socket, document how to get ipv6 support via
119 AnyEvent::Socket instead.
120 - implement signal->wait ($cb) interface, similar to semaphores.
121 - work around SvOK not supporting getmagic, so we have to getmagic
122 to test for undef :/ (reported by Matthias Waldorf).
123 - load Coro::AnyEvent in all modules using AnyEvent.
124 - work around perl corrupting our internal data structures,
125 reported by Tokuhiro Matsuno.
126 - enable per-coroutine real and cpu time gathering
127 (Coro::State::enable_times).
128
1295.14 Wed Jun 24 01:37:48 CEST 2009
130 - provide explicit functions to "cede" to the event loop in Coro::AnyEvent,
131 as this seems to have been a difficult concept (poll, sleep, idle,
132 idle_upto).
133 - add Coro::AnyEvent::readable/writable functions.
134 - clarify Coro::EV/Event/AnyEvent manpages.
135 - free per-thread global scalars in the thread calling ->cancel, to
136 avoid crashes when $_, $@ etc., are magical but some of those
137 had already been freed.
138 - "unexperimentalise" the callback interface for Coro::Semaphore.
139 - speed up ready queue management/context switching by using a linked
140 list instead of an array (~5-10%).
141 - implement "watch" command in Coro::Debug shells.
142 - for fun, implement time-slicing as an example in the manpage.
143 - if AnyEvent detects EV or Event, but we don't have Coro::EV or
144 Coro::Event, use the normal AnyEvent handling instead of dieing
145 (the same is true for Coro::Handle).
146 - properly document Coro::EV::timed_io_once.
147 - avoid unneccessary ->cancel calls in Coro::Handle.
148 - maybe make it work on mingw32 with win32 backend
149 (based on patch by Yasuhiro Matsumoto).
150
1515.132 Fri May 29 09:00:39 CEST 2009
152 - do not keep a reference to the argument itself in
153 Coro::Semaphore::guard, as it could change later.
154 - support SO_RCVBUF/SO_SNDBUF nonstandard Coro::Socket options,
155 should support a prepare callback.
156
1575.131 Mon Mar 16 23:20:37 CET 2009
158 - implement and document Coro->suspend, Coro->resume.
159 - fix Coro::Select implementation to not (often) close
160 the passed file descriptors (testcase provided by pippijn).
161
1625.13 Mon Dec 15 21:51:42 CET 2008
163 - EXPERIMENTAL: implement dynamic winds (on_enter/on_leave).
164 - don't set diehook to C<undef> but instead to NULL, to avoid
165 spurious warnings.
166 - fix a lot of bugs in Coro::SemaphoreSet.
167 - Coro::SemaphoreSet will less often create a semaphore needlessly.
168 - add Coro::SemaphoreSet::count and wait methods.
169 - take advantage of the new Guard module.
170 - deprecate Coro::guard.
171 - try to fix the dreaded 01_unblock tests once more. I hate it when
172 testsuites need more fixing than the code they are supposed to test.
173 - croak in more cases when a required callback isn't resolvable.
174 - fix some minor issues in Coro::State->call/eval.
175 - use current coroutine context instead of a temporary one
176 when temporarily switching to another coroutine.
177 - do not call C-level on_destroy handlers during global destruction,
178 to avoid needless segfaults.
179
1805.12 Sun Dec 7 13:30:38 CET 2008
181 - add default config for MirOS, which seems to be a bug-to-bug
182 compatible fork of openbsd ("world domination by releasing
183 openbsd cvs before the openbsd folks do it" or so :).
184 - free_padlist did destroy the names pad, not good, but didn't
185 seem to bother perl - this could fix issues such as eval ""
186 inside a function called from multiple coroutines.
187 - use a different method to detect destruction time.
188 - be more careful when freeing padlists just before global
189 destruction.
190 - fixed and expanded the call/cc example.
191 - renamed _terminate to _coro_run.
192 - new method Coro::Channel->shutdown.
193 - try pthreads on openbsd <4.4 (broken sigaltstack, will
194 pthreads fare better?).
195 - be less picky about destroying "a" running coroutine.
196
1975.11 Tue Nov 25 21:49:05 CET 2008
198 - DEBUGGING in 5.10.0 is a mess: it constantly flags perfectly
199 working code with failed assertions, introducing more bugs than
200 it fixes, requiring elaborate workarounds :(
201
2025.1 Mon Nov 24 08:54:59 CET 2008
203 - wrote a small introductory tutorial - Coro::Intro.
204 - convert Coro::Timer, Coro::Select and Coro::Util to rouse API.
205 - Coro::Select did errornously dup the file descriptors
206 and didn't work with all AnyEvent backends.
207 - Coro::Select wasn't imported correctly form Coro::LWP, causing blocking
208 LWP data transfers.
209 - disassociate c contexts from coro objects - this is agruably more
210 correct, but mostly allows sharing of cctxs between coro and state
211 objects, for added memory savings and speed increases.
212 - bumped $Coro::POOL_RSS up to 32kb by default.
213 - no longer set the optype to OP_CUSTOM, as B::* understandably
214 doesn't like this very much (and we *are* a type of entersub).
215 - implement state cloning, just to prove that call/cc can be done.
216 - automatically load Coro::AnyEvent in Coro::Handle.
217 - wrap ->cancel calls in eval inside Coro::Handle as EV watchers
218 do not have this method (and don't need it either).
219 - speed up generic anyevent methods in Coro::Handle by using rouse
220 callbacks.
221 - allow coroutines in $Coro::IDLE, speeding up Coro::AnyEvent and
222 others. It also makes the debugger happier, as you can trace
223 through the idle threads now.
224 - add comppad_name* and hints ($^H, %^H) to per-thread variables.
225 - eg/event was pretty broken.
226 - better 5.8.6 compatibility.
227
2285.0 Thu Nov 20 10:35:05 CET 2008
229 - NEW ARCHITECTURE: use the latest 4.x version if you experience
230 stability issues.
15 - bump API version to 7 - all dependents must be recompiled. 231 - bump API version to 7 - all dependents must be recompiled.
232 - removed timed_* functions - they were not being used anyways
233 and should be replaced by a more generic mechanism -
234 and were annoying to support anyways :)
235 - removed SemaphoreSet's waiter method - use sem method instead.
16 - Coro::Semaphore->adjust didn't correctly wake up enough waiters. 236 - Coro::Semaphore->adjust didn't correctly wake up enough waiters.
17 - async_pool did free a scalar value twice 237 - async_pool did free a scalar value twice
18 ("Attempt to unreference..."). 238 ("Attempt to unreference...").
19 - fix a longstanding bug where calling terminate on a coro that 239 - fix a longstanding bug where calling terminate on a coro that
20 was waiting for a semaphore that was just becoming available 240 was waiting for a semaphore that was just becoming available
21 would cause a deadlock (semaphore would get into a state where 241 would cause a deadlock (semaphore would get into a state where
22 it was available but waiters were waiting(. 242 it was available but waiters were still blocked).
243 - calling throw on a coroutine that is waiting for a semaphore will
244 no longer make it acquire the semaphore (and thus leak a count).
23 - perl's process emulation is now not even theoretically supported 245 - perl's process emulation is now not even theoretically supported
24 anymore. 246 anymore.
247 - new functions Coro::rouse_cb and Coro::rouse_wait for easier
248 conversion of callback-style to blocking-style.
249 - new methods $coro->schedule_to and ->cede_to, to specifically
250 schedule or cede to a specific coroutine.
251 - new function Coro::Semaphore::wait.
252 - use named constants in Coro::Channel (Richard Hundt).
25 - directly patch the entersub opcode calling SLF functions (cede, 253 - directly patch the entersub opcode calling SLF functions (cede,
26 transfer and so on). this does speed up context switching, but 254 transfer and so on). this does speed up context switching, but
27 more importanly, it frees us from the hardcoded behaviour of 255 more importanly, it frees us from the hardcoded behaviour of
28 entersub, so we might actually be able to return something from 256 entersub, so we might actually be able to return something from
29 those functions and atcually create new ones. 257 those functions and atcually create new ones.
30 - take advantage of __builtin_frame_address on gcc. 258 - take advantage of __builtin_frame_address on gcc.
31 - expose THX in coroapi (not sure whether this was a wise decision, 259 - expose THX in coroapi (not sure whether this was a wise decision,
32 as "threaded" perls are running at half speed anyways). 260 as "threaded" perls are running at half speed anyways).
33 - implement execute_slf (schedule-like-function) interface that makes 261 - implement execute_slf (schedule-like-function) interface that makes
34 it possible to implement schedule-like-functions in XS. 262 it possible to implement schedule-like-functions in XS.
35 - use new SLF interface to massively speed up Coro::EV by roughly a factor 263 - use new SLF interface to massively speed up Coro::EV by roughly a
36 of two. 264 factor of two.
37 - removed timed_down and timed_guard functions - they were not being used
38 anyways and should be replaced by a more generic mechanism - and were
39 annoying to support anyways :
40 - used new SLF interface to massively speed up Coro::Semaphore by a factor 265 - used new SLF interface to massively speed up Coro::Semaphore by a
41 of three. 266 factor of three.
267 - used new SLF interface to speed up Coro::AIO by roughly a factor of
268 four and reduce its memory usage considerably.
269 - implement Coro::SemaphoreSet purely in terms of Coro::Semaphore,
270 for a nice speedup and vastly more correct behaviour. Also implement
271 a new method "sem" to get at the underlying semaphore object.
272 - implement Coro::Channel in terms of Coro::Semaphore, for a moderate
273 (in comparison) 20-40% speedup.
274 - used new SLF interface to reimplement Coro::Signal gaining
275 some unknown (because I was too lazy), but certain, speedup, and also
276 making signals reliable for the first time.
277 - used new SLF interface and other optimisations to speed up async_pool
278 by a factor of two. It also doesn't rely on perl's exception mechanism
279 to exit anymore. The overhead for terminating an async_pool, coro over
280 a normal async is now very small.
281 - sped up coroutine creation/destruction by 40%.
42 - forgot to include Coro/libcoro/README in the dist for all these years. 282 - forgot to include Coro/libcoro/README in the dist for all these years.
43 - work around a freebsd pthreads bug (manual testcancel is required as 283 - work around a freebsd pthreads bug (manual testcancel is required as
44 pthread_cond_wait isn't a cancellation point on freebsd). 284 pthread_cond_wait isn't a cancellation point on freebsd).
285 - use new rouse functions to speed up and simplify Coro::BDB.
286 - make "prefer perl native functions" work with threaded perls.
287 - condense Coro::Debug ps output, hint at v and w flags.
288 - (libcoro) lots of minor cleanups and portability improvements.
289
2904.914 Wed Nov 19 12:54:18 CET 2008
291 - fix a disastrous bug in the readline optimisation
292 introduced in 4.801.
45 293
464.913 Sat Nov 15 07:58:28 CET 2008 2944.913 Sat Nov 15 07:58:28 CET 2008
47 - async_pool did free a scalar value twice 295 - async_pool did free a scalar value twice
48 ("Attempt to unreference..."). 296 ("Attempt to unreference...").
49 297

Diff Legend

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