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

Comparing Coro/Changes (file contents):
Revision 1.429 by root, Tue Nov 18 09:59:27 2008 UTC vs.
Revision 1.541 by root, Sun Sep 12 03:06:37 2010 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 9 - correctly modify Coro::AIO function prototypes
9TODO: guard as SLF - to avoid deadlocks. 10 so that they reflect the "no optional parameters" rule.
10 - avoid deadlock in Coro::Channel if maxsize == 1 (Richard Hundt). 11 - "ported" libcoro to C++.
11 12
125.0 135.23 Mon May 17 18:50:42 CEST 2010
14 - be more resistant to ordering changes when initialising
15 Coro::AnyEvent, Coro::EV and Coro::Event (reported by Matthias
16 Waldorf).
17 - document that perl 5.12 deliberately removed support for cloning.
18
195.22 Wed Apr 14 03:55:35 CEST 2010
20 - correctly return udnef on errors in Coro::Handle::read/write
21 (testcase by Marc Mims).
22 - convert Coro::Util into a "perl compatibility wrapper" - the functions
23 are less useful now, but are drop-in replacements for existing
24 functions, listing better alternatives in the documentation. This also
25 fixes a bug in Coro::LWP which naively substituted Socket::inet_aton
26 with Coro::Util::inet_aton.
27 - do not override $Coro::idle unconditionally in Coro.pm, as other
28 modules could have provided their own idle coro already
29 (for exmaple, Coro::AnyEvent).
30 - fix Coro::Util::gethost* functions.
31 - Coro::Timer corretcly exports it's symbols (reported by Hideki Yamamura).
32
335.21 Wed Dec 16 07:19:51 CET 2009
34 - automatically load Coro::AnyEvent when AnyEvent and Coro are used
35 together.
36 - add some examples on how to combine other event loops with Coro in
37 Coro::AnyEvent, and how to run it (and not to block). Seems to be
38 the most common source of confusion.
39 - try to catch people naively blocking in an event callback.
40 - work around the perl filehandle bug issue in conjunction with
41 older common::sense (as indirectly pointed out by ZSystem).
42 - clarify the "not from signal handlers" section.
43
445.2 Sun Oct 4 14:54:24 CEST 2009
45 - Coro::Storable destroyed the prototypes of the functions it wrapped.
46 - export rouse_cb and rouse_wait by default now.
47 - fix various prototype mismatches in Coro::AnyEvent and Coro::Handle.
48 - new method $state->swap_sv.
49 - added section on "windows process emulation" to the manpage,
50 after a not-so-fruitful (nor-friendly) "discussion" with chip
51 salzenberg (discussion implies arguments, but his only arguments
52 were ad-hominems, one wonders why he started it in the first
53 place). I hope this explains it well enough for him to understand,
54 and maybe well enough for others to understand.
55 - use common::sense everywhere now.
56 - idle callbacks are no longer supported, use idle coros instead.
57 - print a thread listing when a deadlock is detected.
58
595.17 Sat Aug 22 23:09:31 CEST 2009
60 - work around a bug in the perl debugger causing crashes
61 when running under the debugger by marking _pool_handler as nodebug.
62 - speed up Coro::async considerably.
63 - try some hacks to get netbsd to work "more often" - their broken
64 setjmp/longjmp, ucontext *and* phtreads are really hard on Coro.
65 - convert Coro to AE 5.0 API.
66
675.162 Tue Jul 28 04:04:03 CEST 2009
68 - perl 5.8.2 is now minimum requirement.
69 - skip t/19_handle.t on broken windows perls.
70
715.161 Wed Jul 22 04:47:38 CEST 2009
72 - Coro::AnyEvent::poll could have a different prototype when EV was
73 used as backend (analyzed by Tatsuhiko Miyagawa).
74 - Coro::AnyEvent errornously initialised the event loop when loaded,
75 not on demand.
76 - try to workaround rare */t/01_unblock.t failures.
77
785.16 Tue Jul 21 01:44:37 CEST 2009
79 - Coro::AnyEvent failed to hook into the event loop
80 when no threads had been readied between detecting
81 the event loop and actually running it.
82 - considerably speed up Coro::Select by taking avdantage
83 of AnyEvent > 4.8 and some other optimisations.
84 - implement paragraph readline mode in Coro::Handle
85 (based on patches by Zsbán Ambrus).
86 - replace WSAEINPROGRESS by WSAEWOULDBLOCK (reported
87 and analyzed by Yasuhiro MATSUMOTO).
88 - clarified libcoro license and copyright.
89 - someone stole my EXTRA_META!!!
90 - implement Coro::Select::patch_pp_sselect and it's brother,
91 for hardcode select overriding.
92
935.151 Mon Jul 6 05:41:57 CEST 2009
94 - backport to windows process emulation code again (patch by
95 Yasuhiro MATSUMOTO).
96 - slightly update Coro::MakeMaker.
97
985.15 Tue Jun 30 10:28:06 CEST 2009
99 - deprecate Coro::Socket, document how to get ipv6 support via
100 AnyEvent::Socket instead.
101 - implement signal->wait ($cb) interface, similar to semaphores.
102 - work around SvOK not supporting getmagic, so we have to getmagic
103 to test for undef :/ (reported by Matthias Waldorf).
104 - load Coro::AnyEvent in all modules using AnyEvent.
105 - work around perl corrupting our internal data structures,
106 reported by Tokuhiro Matsuno.
107 - enable per-coroutine real and cpu time gathering
108 (Coro::State::enable_times).
109
1105.14 Wed Jun 24 01:37:48 CEST 2009
111 - provide explicit functions to "cede" to the event loop in Coro::AnyEvent,
112 as this seems to have been a difficult concept (poll, sleep, idle,
113 idle_upto).
114 - add Coro::AnyEvent::readable/writable functions.
115 - clarify Coro::EV/Event/AnyEvent manpages.
116 - free per-thread global scalars in the thread calling ->cancel, to
117 avoid crashes when $_, $@ etc., are magical but some of those
118 had already been freed.
119 - "unexperimentalise" the callback interface for Coro::Semaphore.
120 - speed up ready queue management/context switching by using a linked
121 list instead of an array (~5-10%).
122 - implement "watch" command in Coro::Debug shells.
123 - for fun, implement time-slicing as an example in the manpage.
124 - if AnyEvent detects EV or Event, but we don't have Coro::EV or
125 Coro::Event, use the normal AnyEvent handling instead of dieing
126 (the same is true for Coro::Handle).
127 - properly document Coro::EV::timed_io_once.
128 - avoid unneccessary ->cancel calls in Coro::Handle.
129 - maybe make it work on mingw32 with win32 backend
130 (based on patch by Yasuhiro Matsumoto).
131
1325.132 Fri May 29 09:00:39 CEST 2009
133 - do not keep a reference to the argument itself in
134 Coro::Semaphore::guard, as it could change later.
135 - support SO_RCVBUF/SO_SNDBUF nonstandard Coro::Socket options,
136 should support a prepare callback.
137
1385.131 Mon Mar 16 23:20:37 CET 2009
139 - implement and document Coro->suspend, Coro->resume.
140 - fix Coro::Select implementation to not (often) close
141 the passed file descriptors (testcase provided by pippijn).
142
1435.13 Mon Dec 15 21:51:42 CET 2008
144 - EXPERIMENTAL: implement dynamic winds (on_enter/on_leave).
145 - don't set diehook to C<undef> but instead to NULL, to avoid
146 spurious warnings.
147 - fix a lot of bugs in Coro::SemaphoreSet.
148 - Coro::SemaphoreSet will less often create a semaphore needlessly.
149 - add Coro::SemaphoreSet::count and wait methods.
150 - take advantage of the new Guard module.
151 - deprecate Coro::guard.
152 - try to fix the dreaded 01_unblock tests once more. I hate it when
153 testsuites need more fixing than the code they are supposed to test.
154 - croak in more cases when a required callback isn't resolvable.
155 - fix some minor issues in Coro::State->call/eval.
156 - use current coroutine context instead of a temporary one
157 when temporarily switching to another coroutine.
158 - do not call C-level on_destroy handlers during global destruction,
159 to avoid needless segfaults.
160
1615.12 Sun Dec 7 13:30:38 CET 2008
162 - add default config for MirOS, which seems to be a bug-to-bug
163 compatible fork of openbsd ("world domination by releasing
164 openbsd cvs before the openbsd folks do it" or so :).
165 - free_padlist did destroy the names pad, not good, but didn't
166 seem to bother perl - this could fix issues such as eval ""
167 inside a function called from multiple coroutines.
168 - use a different method to detect destruction time.
169 - be more careful when freeing padlists just before global
170 destruction.
171 - fixed and expanded the call/cc example.
172 - renamed _terminate to _coro_run.
173 - new method Coro::Channel->shutdown.
174 - try pthreads on openbsd <4.4 (broken sigaltstack, will
175 pthreads fare better?).
176 - be less picky about destroying "a" running coroutine.
177
1785.11 Tue Nov 25 21:49:05 CET 2008
179 - DEBUGGING in 5.10.0 is a mess: it constantly flags perfectly
180 working code with failed assertions, introducing more bugs than
181 it fixes, requiring elaborate workarounds :(
182
1835.1 Mon Nov 24 08:54:59 CET 2008
184 - wrote a small introductory tutorial - Coro::Intro.
185 - convert Coro::Timer, Coro::Select and Coro::Util to rouse API.
186 - Coro::Select did errornously dup the file descriptors
187 and didn't work with all AnyEvent backends.
188 - Coro::Select wasn't imported correctly form Coro::LWP, causing blocking
189 LWP data transfers.
190 - disassociate c contexts from coro objects - this is agruably more
191 correct, but mostly allows sharing of cctxs between coro and state
192 objects, for added memory savings and speed increases.
193 - bumped $Coro::POOL_RSS up to 32kb by default.
194 - no longer set the optype to OP_CUSTOM, as B::* understandably
195 doesn't like this very much (and we *are* a type of entersub).
196 - implement state cloning, just to prove that call/cc can be done.
197 - automatically load Coro::AnyEvent in Coro::Handle.
198 - wrap ->cancel calls in eval inside Coro::Handle as EV watchers
199 do not have this method (and don't need it either).
200 - speed up generic anyevent methods in Coro::Handle by using rouse
201 callbacks.
202 - allow coroutines in $Coro::IDLE, speeding up Coro::AnyEvent and
203 others. It also makes the debugger happier, as you can trace
204 through the idle threads now.
205 - add comppad_name* and hints ($^H, %^H) to per-thread variables.
206 - eg/event was pretty broken.
207 - better 5.8.6 compatibility.
208
2095.0 Thu Nov 20 10:35:05 CET 2008
13 - NEW ARCHITECTURE: use the latest 4.x version if you experience 210 - NEW ARCHITECTURE: use the latest 4.x version if you experience
14 stability issues. 211 stability issues.
15 - bump API version to 7 - all dependents must be recompiled. 212 - bump API version to 7 - all dependents must be recompiled.
16 - removed timed_down and timed_guard functions - they were not being 213 - removed timed_* functions - they were not being used anyways
17 used anyways and should be replaced by a more generic mechanism - 214 and should be replaced by a more generic mechanism -
18 and were annoying to support anyways :) 215 and were annoying to support anyways :)
19 - removed SemaphoreSet's waiter method - use sem instead. 216 - removed SemaphoreSet's waiter method - use sem method instead.
20 - Coro::Semaphore->adjust didn't correctly wake up enough waiters. 217 - Coro::Semaphore->adjust didn't correctly wake up enough waiters.
21 - async_pool did free a scalar value twice 218 - async_pool did free a scalar value twice
22 ("Attempt to unreference..."). 219 ("Attempt to unreference...").
23 - fix a longstanding bug where calling terminate on a coro that 220 - fix a longstanding bug where calling terminate on a coro that
24 was waiting for a semaphore that was just becoming available 221 was waiting for a semaphore that was just becoming available
25 would cause a deadlock (semaphore would get into a state where 222 would cause a deadlock (semaphore would get into a state where
26 it was available but waiters were waiting(. 223 it was available but waiters were still blocked).
224 - calling throw on a coroutine that is waiting for a semaphore will
225 no longer make it acquire the semaphore (and thus leak a count).
27 - perl's process emulation is now not even theoretically supported 226 - perl's process emulation is now not even theoretically supported
28 anymore. 227 anymore.
228 - new functions Coro::rouse_cb and Coro::rouse_wait for easier
229 conversion of callback-style to blocking-style.
230 - new methods $coro->schedule_to and ->cede_to, to specifically
231 schedule or cede to a specific coroutine.
232 - new function Coro::Semaphore::wait.
29 - use named constants in Coro::Channel (Richard Hundt). 233 - use named constants in Coro::Channel (Richard Hundt).
30 - directly patch the entersub opcode calling SLF functions (cede, 234 - directly patch the entersub opcode calling SLF functions (cede,
31 transfer and so on). this does speed up context switching, but 235 transfer and so on). this does speed up context switching, but
32 more importanly, it frees us from the hardcoded behaviour of 236 more importanly, it frees us from the hardcoded behaviour of
33 entersub, so we might actually be able to return something from 237 entersub, so we might actually be able to return something from
… …
44 - used new SLF interface to speed up Coro::AIO by roughly a factor of 248 - used new SLF interface to speed up Coro::AIO by roughly a factor of
45 four and reduce its memory usage considerably. 249 four and reduce its memory usage considerably.
46 - implement Coro::SemaphoreSet purely in terms of Coro::Semaphore, 250 - implement Coro::SemaphoreSet purely in terms of Coro::Semaphore,
47 for a nice speedup and vastly more correct behaviour. Also implement 251 for a nice speedup and vastly more correct behaviour. Also implement
48 a new method "sem" to get at the underlying semaphore object. 252 a new method "sem" to get at the underlying semaphore object.
253 - implement Coro::Channel in terms of Coro::Semaphore, for a moderate
254 (in comparison) 20-40% speedup.
255 - used new SLF interface to reimplement Coro::Signal gaining
256 some unknown (because I was too lazy), but certain, speedup, and also
257 making signals reliable for the first time.
258 - used new SLF interface and other optimisations to speed up async_pool
259 by a factor of two. It also doesn't rely on perl's exception mechanism
260 to exit anymore. The overhead for terminating an async_pool, coro over
261 a normal async is now very small.
262 - sped up coroutine creation/destruction by 40%.
49 - forgot to include Coro/libcoro/README in the dist for all these years. 263 - forgot to include Coro/libcoro/README in the dist for all these years.
50 - work around a freebsd pthreads bug (manual testcancel is required as 264 - work around a freebsd pthreads bug (manual testcancel is required as
51 pthread_cond_wait isn't a cancellation point on freebsd). 265 pthread_cond_wait isn't a cancellation point on freebsd).
266 - use new rouse functions to speed up and simplify Coro::BDB.
267 - make "prefer perl native functions" work with threaded perls.
268 - condense Coro::Debug ps output, hint at v and w flags.
269 - (libcoro) lots of minor cleanups and portability improvements.
270
2714.914 Wed Nov 19 12:54:18 CET 2008
272 - fix a disastrous bug in the readline optimisation
273 introduced in 4.801.
52 274
534.913 Sat Nov 15 07:58:28 CET 2008 2754.913 Sat Nov 15 07:58:28 CET 2008
54 - async_pool did free a scalar value twice 276 - async_pool did free a scalar value twice
55 ("Attempt to unreference..."). 277 ("Attempt to unreference...").
56 278

Diff Legend

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