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

Comparing Coro/Changes (file contents):
Revision 1.436 by root, Wed Nov 19 02:56:15 2008 UTC vs.
Revision 1.557 by root, Sat Apr 30 05:17:42 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 - implement Coro->cancel in XS for a 20% speed improvement, and to
9TODO: guard as SLF - to avoid deadlocks. 8 be able to implement mutual cancellation.
10 - avoid deadlock in Coro::Channel if maxsize == 1 (Richard Hundt). 9 - cancelling a coro while it itself is cancelling another coro is
10 now supported and working, instead of triggering an assertion.
11 - be a bit more crash-resistant when calling (buggy) on_destroy
12 callbacks (best effort).
11 13
125.0 145.372 Wed Feb 23 06:14:30 CET 2011
15 - apparently mingw doesn't provide a working gettimeofday, try to
16 work around that by relying on Time::HiRes (indirectly brought to
17 my attention by Max Maischein).
18 - fix some portability issues when Time::HiRes was used.
19
205.371 Mon Feb 21 14:36:08 CET 2011
21 - backport to windows process emulation code again.
22
235.37 Sat Feb 19 07:49:44 CET 2011
24 - add a big "Coro thread life cycle" section to "man Coro".
25 - try a tentative workaround against the breakage that 5.13 has
26 introduced without depreciation period. sigh.
27 - no longer use Time::HiRes if gettimeofday is available, which
28 saves quite a lot of memory.
29
305.36 Sun Feb 13 05:33:41 CET 2011
31 - automatically load Coro::Channel, Coro::RWLock, Coro::Semaphore,
32 Coro::SemaphoreSet, Coro::Signal and Coro::Specific on first "new"
33 method call.
34 - undocument Coro::Timer::sleep and obsolete whole module.
35 - optimise Coro::Timer::timeout memory and cpu usage.
36 - slightly updated Coro::Intro for recent changes.
37 - do not initialise PL_dirty anymore.
38
395.25 Thu Nov 11 01:08:39 CET 2010
40 - try a different approach on netbsd - netbsd 5 finally has marginally
41 working pthreads, but still broken ucontext/sigaltstack.
42 - openbsd 4.8 finally got their act together, Coro works out of the box
43 with asm, setjmp and pthreads (no change, just informational).
44
455.24 Sat Oct 23 11:27:12 CEST 2010
46 - port to the EV 4.0 API.
47 - work around bugs in mingw32, making strawberry perl work
48 out of the box.
49 - correctly modify Coro::AIO function prototypes
50 so that they reflect the "no optional parameters" rule.
51 - "ported" libcoro to C++.
52
535.23 Mon May 17 18:50:42 CEST 2010
54 - be more resistant to ordering changes when initialising
55 Coro::AnyEvent, Coro::EV and Coro::Event (reported by Matthias
56 Waldorf).
57 - document that perl 5.12 deliberately removed support for cloning.
58
595.22 Wed Apr 14 03:55:35 CEST 2010
60 - correctly return udnef on errors in Coro::Handle::read/write
61 (testcase by Marc Mims).
62 - convert Coro::Util into a "perl compatibility wrapper" - the functions
63 are less useful now, but are drop-in replacements for existing
64 functions, listing better alternatives in the documentation. This also
65 fixes a bug in Coro::LWP which naively substituted Socket::inet_aton
66 with Coro::Util::inet_aton.
67 - do not override $Coro::idle unconditionally in Coro.pm, as other
68 modules could have provided their own idle coro already
69 (for exmaple, Coro::AnyEvent).
70 - fix Coro::Util::gethost* functions.
71 - Coro::Timer corretcly exports it's symbols (reported by Hideki Yamamura).
72
735.21 Wed Dec 16 07:19:51 CET 2009
74 - automatically load Coro::AnyEvent when AnyEvent and Coro are used
75 together.
76 - add some examples on how to combine other event loops with Coro in
77 Coro::AnyEvent, and how to run it (and not to block). Seems to be
78 the most common source of confusion.
79 - try to catch people naively blocking in an event callback.
80 - work around the perl filehandle bug issue in conjunction with
81 older common::sense (as indirectly pointed out by ZSystem).
82 - clarify the "not from signal handlers" section.
83
845.2 Sun Oct 4 14:54:24 CEST 2009
85 - Coro::Storable destroyed the prototypes of the functions it wrapped.
86 - export rouse_cb and rouse_wait by default now.
87 - fix various prototype mismatches in Coro::AnyEvent and Coro::Handle.
88 - new method $state->swap_sv.
89 - added section on "windows process emulation" to the manpage,
90 after a not-so-fruitful (nor-friendly) "discussion" with chip
91 salzenberg (discussion implies arguments, but his only arguments
92 were ad-hominems, one wonders why he started it in the first
93 place). I hope this explains it well enough for him to understand,
94 and maybe well enough for others to understand.
95 - use common::sense everywhere now.
96 - idle callbacks are no longer supported, use idle coros instead.
97 - print a thread listing when a deadlock is detected.
98
995.17 Sat Aug 22 23:09:31 CEST 2009
100 - work around a bug in the perl debugger causing crashes
101 when running under the debugger by marking _pool_handler as nodebug.
102 - speed up Coro::async considerably.
103 - try some hacks to get netbsd to work "more often" - their broken
104 setjmp/longjmp, ucontext *and* phtreads are really hard on Coro.
105 - convert Coro to AE 5.0 API.
106
1075.162 Tue Jul 28 04:04:03 CEST 2009
108 - perl 5.8.2 is now minimum requirement.
109 - skip t/19_handle.t on broken windows perls.
110
1115.161 Wed Jul 22 04:47:38 CEST 2009
112 - Coro::AnyEvent::poll could have a different prototype when EV was
113 used as backend (analyzed by Tatsuhiko Miyagawa).
114 - Coro::AnyEvent errornously initialised the event loop when loaded,
115 not on demand.
116 - try to workaround rare */t/01_unblock.t failures.
117
1185.16 Tue Jul 21 01:44:37 CEST 2009
119 - Coro::AnyEvent failed to hook into the event loop
120 when no threads had been readied between detecting
121 the event loop and actually running it.
122 - considerably speed up Coro::Select by taking avdantage
123 of AnyEvent > 4.8 and some other optimisations.
124 - implement paragraph readline mode in Coro::Handle
125 (based on patches by Zsbán Ambrus).
126 - replace WSAEINPROGRESS by WSAEWOULDBLOCK (reported
127 and analyzed by Yasuhiro MATSUMOTO).
128 - clarified libcoro license and copyright.
129 - someone stole my EXTRA_META!!!
130 - implement Coro::Select::patch_pp_sselect and it's brother,
131 for hardcode select overriding.
132
1335.151 Mon Jul 6 05:41:57 CEST 2009
134 - backport to windows process emulation code again (patch by
135 Yasuhiro MATSUMOTO).
136 - slightly update Coro::MakeMaker.
137
1385.15 Tue Jun 30 10:28:06 CEST 2009
139 - deprecate Coro::Socket, document how to get ipv6 support via
140 AnyEvent::Socket instead.
141 - implement signal->wait ($cb) interface, similar to semaphores.
142 - work around SvOK not supporting getmagic, so we have to getmagic
143 to test for undef :/ (reported by Matthias Waldorf).
144 - load Coro::AnyEvent in all modules using AnyEvent.
145 - work around perl corrupting our internal data structures,
146 reported by Tokuhiro Matsuno.
147 - enable per-coroutine real and cpu time gathering
148 (Coro::State::enable_times).
149
1505.14 Wed Jun 24 01:37:48 CEST 2009
151 - provide explicit functions to "cede" to the event loop in Coro::AnyEvent,
152 as this seems to have been a difficult concept (poll, sleep, idle,
153 idle_upto).
154 - add Coro::AnyEvent::readable/writable functions.
155 - clarify Coro::EV/Event/AnyEvent manpages.
156 - free per-thread global scalars in the thread calling ->cancel, to
157 avoid crashes when $_, $@ etc., are magical but some of those
158 had already been freed.
159 - "unexperimentalise" the callback interface for Coro::Semaphore.
160 - speed up ready queue management/context switching by using a linked
161 list instead of an array (~5-10%).
162 - implement "watch" command in Coro::Debug shells.
163 - for fun, implement time-slicing as an example in the manpage.
164 - if AnyEvent detects EV or Event, but we don't have Coro::EV or
165 Coro::Event, use the normal AnyEvent handling instead of dieing
166 (the same is true for Coro::Handle).
167 - properly document Coro::EV::timed_io_once.
168 - avoid unneccessary ->cancel calls in Coro::Handle.
169 - maybe make it work on mingw32 with win32 backend
170 (based on patch by Yasuhiro Matsumoto).
171
1725.132 Fri May 29 09:00:39 CEST 2009
173 - do not keep a reference to the argument itself in
174 Coro::Semaphore::guard, as it could change later.
175 - support SO_RCVBUF/SO_SNDBUF nonstandard Coro::Socket options,
176 should support a prepare callback.
177
1785.131 Mon Mar 16 23:20:37 CET 2009
179 - implement and document Coro->suspend, Coro->resume.
180 - fix Coro::Select implementation to not (often) close
181 the passed file descriptors (testcase provided by pippijn).
182
1835.13 Mon Dec 15 21:51:42 CET 2008
184 - EXPERIMENTAL: implement dynamic winds (on_enter/on_leave).
185 - don't set diehook to C<undef> but instead to NULL, to avoid
186 spurious warnings.
187 - fix a lot of bugs in Coro::SemaphoreSet.
188 - Coro::SemaphoreSet will less often create a semaphore needlessly.
189 - add Coro::SemaphoreSet::count and wait methods.
190 - take advantage of the new Guard module.
191 - deprecate Coro::guard.
192 - try to fix the dreaded 01_unblock tests once more. I hate it when
193 testsuites need more fixing than the code they are supposed to test.
194 - croak in more cases when a required callback isn't resolvable.
195 - fix some minor issues in Coro::State->call/eval.
196 - use current coroutine context instead of a temporary one
197 when temporarily switching to another coroutine.
198 - do not call C-level on_destroy handlers during global destruction,
199 to avoid needless segfaults.
200
2015.12 Sun Dec 7 13:30:38 CET 2008
202 - add default config for MirOS, which seems to be a bug-to-bug
203 compatible fork of openbsd ("world domination by releasing
204 openbsd cvs before the openbsd folks do it" or so :).
205 - free_padlist did destroy the names pad, not good, but didn't
206 seem to bother perl - this could fix issues such as eval ""
207 inside a function called from multiple coroutines.
208 - use a different method to detect destruction time.
209 - be more careful when freeing padlists just before global
210 destruction.
211 - fixed and expanded the call/cc example.
212 - renamed _terminate to _coro_run.
213 - new method Coro::Channel->shutdown.
214 - try pthreads on openbsd <4.4 (broken sigaltstack, will
215 pthreads fare better?).
216 - be less picky about destroying "a" running coroutine.
217
2185.11 Tue Nov 25 21:49:05 CET 2008
219 - DEBUGGING in 5.10.0 is a mess: it constantly flags perfectly
220 working code with failed assertions, introducing more bugs than
221 it fixes, requiring elaborate workarounds :(
222
2235.1 Mon Nov 24 08:54:59 CET 2008
224 - wrote a small introductory tutorial - Coro::Intro.
225 - convert Coro::Timer, Coro::Select and Coro::Util to rouse API.
226 - Coro::Select did errornously dup the file descriptors
227 and didn't work with all AnyEvent backends.
228 - Coro::Select wasn't imported correctly form Coro::LWP, causing blocking
229 LWP data transfers.
230 - disassociate c contexts from coro objects - this is agruably more
231 correct, but mostly allows sharing of cctxs between coro and state
232 objects, for added memory savings and speed increases.
233 - bumped $Coro::POOL_RSS up to 32kb by default.
234 - no longer set the optype to OP_CUSTOM, as B::* understandably
235 doesn't like this very much (and we *are* a type of entersub).
236 - implement state cloning, just to prove that call/cc can be done.
237 - automatically load Coro::AnyEvent in Coro::Handle.
238 - wrap ->cancel calls in eval inside Coro::Handle as EV watchers
239 do not have this method (and don't need it either).
240 - speed up generic anyevent methods in Coro::Handle by using rouse
241 callbacks.
242 - allow coroutines in $Coro::IDLE, speeding up Coro::AnyEvent and
243 others. It also makes the debugger happier, as you can trace
244 through the idle threads now.
245 - add comppad_name* and hints ($^H, %^H) to per-thread variables.
246 - eg/event was pretty broken.
247 - better 5.8.6 compatibility.
248
2495.0 Thu Nov 20 10:35:05 CET 2008
13 - NEW ARCHITECTURE: use the latest 4.x version if you experience 250 - NEW ARCHITECTURE: use the latest 4.x version if you experience
14 stability issues. 251 stability issues.
15 - bump API version to 7 - all dependents must be recompiled. 252 - bump API version to 7 - all dependents must be recompiled.
16 - removed timed_* functions - they were not being used anyways 253 - removed timed_* functions - they were not being used anyways
17 and should be replaced by a more generic mechanism - 254 and should be replaced by a more generic mechanism -
18 and were annoying to support anyways :) 255 and were annoying to support anyways :)
19 - removed SemaphoreSet's waiter method - use sem instead. 256 - removed SemaphoreSet's waiter method - use sem method instead.
20 - Coro::Semaphore->adjust didn't correctly wake up enough waiters. 257 - Coro::Semaphore->adjust didn't correctly wake up enough waiters.
21 - async_pool did free a scalar value twice 258 - async_pool did free a scalar value twice
22 ("Attempt to unreference..."). 259 ("Attempt to unreference...").
23 - fix a longstanding bug where calling terminate on a coro that 260 - fix a longstanding bug where calling terminate on a coro that
24 was waiting for a semaphore that was just becoming available 261 was waiting for a semaphore that was just becoming available
… …
26 it was available but waiters were still blocked). 263 it was available but waiters were still blocked).
27 - calling throw on a coroutine that is waiting for a semaphore will 264 - calling throw on a coroutine that is waiting for a semaphore will
28 no longer make it acquire the semaphore (and thus leak a count). 265 no longer make it acquire the semaphore (and thus leak a count).
29 - perl's process emulation is now not even theoretically supported 266 - perl's process emulation is now not even theoretically supported
30 anymore. 267 anymore.
268 - new functions Coro::rouse_cb and Coro::rouse_wait for easier
269 conversion of callback-style to blocking-style.
270 - new methods $coro->schedule_to and ->cede_to, to specifically
271 schedule or cede to a specific coroutine.
31 - new function Coro::Semaphore::wait. 272 - new function Coro::Semaphore::wait.
32 - use named constants in Coro::Channel (Richard Hundt). 273 - use named constants in Coro::Channel (Richard Hundt).
33 - directly patch the entersub opcode calling SLF functions (cede, 274 - directly patch the entersub opcode calling SLF functions (cede,
34 transfer and so on). this does speed up context switching, but 275 transfer and so on). this does speed up context switching, but
35 more importanly, it frees us from the hardcoded behaviour of 276 more importanly, it frees us from the hardcoded behaviour of
… …
52 - implement Coro::Channel in terms of Coro::Semaphore, for a moderate 293 - implement Coro::Channel in terms of Coro::Semaphore, for a moderate
53 (in comparison) 20-40% speedup. 294 (in comparison) 20-40% speedup.
54 - used new SLF interface to reimplement Coro::Signal gaining 295 - used new SLF interface to reimplement Coro::Signal gaining
55 some unknown (because I was too lazy), but certain, speedup, and also 296 some unknown (because I was too lazy), but certain, speedup, and also
56 making signals reliable for the first time. 297 making signals reliable for the first time.
298 - used new SLF interface and other optimisations to speed up async_pool
299 by a factor of two. It also doesn't rely on perl's exception mechanism
300 to exit anymore. The overhead for terminating an async_pool, coro over
301 a normal async is now very small.
302 - sped up coroutine creation/destruction by 40%.
57 - forgot to include Coro/libcoro/README in the dist for all these years. 303 - forgot to include Coro/libcoro/README in the dist for all these years.
58 - work around a freebsd pthreads bug (manual testcancel is required as 304 - work around a freebsd pthreads bug (manual testcancel is required as
59 pthread_cond_wait isn't a cancellation point on freebsd). 305 pthread_cond_wait isn't a cancellation point on freebsd).
306 - use new rouse functions to speed up and simplify Coro::BDB.
307 - make "prefer perl native functions" work with threaded perls.
308 - condense Coro::Debug ps output, hint at v and w flags.
60 - (libcoro) lots of minor cleanups and portability improvements. 309 - (libcoro) lots of minor cleanups and portability improvements.
310
3114.914 Wed Nov 19 12:54:18 CET 2008
312 - fix a disastrous bug in the readline optimisation
313 introduced in 4.801.
61 314
624.913 Sat Nov 15 07:58:28 CET 2008 3154.913 Sat Nov 15 07:58:28 CET 2008
63 - async_pool did free a scalar value twice 316 - async_pool did free a scalar value twice
64 ("Attempt to unreference..."). 317 ("Attempt to unreference...").
65 318

Diff Legend

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