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

Comparing Coro/Changes (file contents):
Revision 1.426 by root, Mon Nov 17 07:19:31 2008 UTC vs.
Revision 1.517 by root, Thu Jul 23 01:35:18 2009 UTC

… …
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: maybe implement a default message channel, very much like Erlang's
5 actor model (which is cool in a lot of important aspects (failures!), 5 actor model (which is cool in a lot of important aspects (failures!),
6 but very lacking in others (higher level ipc)). 6 but very lacking in others (higher level ipc)).
7TODO: unready_all
8TODO: myhttpd header parsing
9TODO: channel->maxsize(newsize)?
7 10
8TODO: testcancel 11 - perl 5.8.2 is now minimum requirement.
9TODO: aio 12
10TODO: Coro::Semaphore 135.161 Wed Jul 22 04:47:38 CEST 2009
11TODO: trow disturbs cleanup path inside pp_slf? 14 - Coro::AnyEvent::poll could have a different prototype when EV was
12TODO: throw special-case in transfer? 15 used as backend (analyzed by Tatsuhiko Miyagawa).
135.0 16 - Coro::AnyEvent errornously initialised the event loop when loaded,
17 not on demand.
18 - try to workaround rare */t/01_unblock.t failures.
19
205.16 Tue Jul 21 01:44:37 CEST 2009
21 - Coro::AnyEvent failed to hook into the event loop
22 when no threads had been readied between detecting
23 the event loop and actually running it.
24 - considerably speed up Coro::Select by taking avdantage
25 of AnyEvent > 4.8 and some other optimisations.
26 - implement paragraph readline mode in Coro::Handle
27 (based on patches by Zsbán Ambrus).
28 - replace WSAEINPROGRESS by WSAEWOULDBLOCK (reported
29 and analyzed by Yasuhiro MATSUMOTO).
30 - clarified libcoro license and copyright.
31 - someone stole my EXTRA_META!!!
32 - implement Coro::Select::patch_pp_sselect and it's brother,
33 for hardcode select overriding.
34
355.151 Mon Jul 6 05:41:57 CEST 2009
36 - backport to windows process emulation code again (patch by
37 Yasuhiro MATSUMOTO).
38 - slightly update Coro::MakeMaker.
39
405.15 Tue Jun 30 10:28:06 CEST 2009
41 - deprecate Coro::Socket, document how to get ipv6 support via
42 AnyEvent::Socket instead.
43 - implement signal->wait ($cb) interface, similar to semaphores.
44 - work around SvOK not supporting getmagic, so we have to getmagic
45 to test for undef :/ (reported by Matthias Waldorf).
46 - load Coro::AnyEvent in all modules using AnyEvent.
47 - work around perl corrupting our internal data structures,
48 reported by Tokuhiro Matsuno.
49 - enable per-coroutine real and cpu time gathering
50 (Coro::State::enable_times).
51
525.14 Wed Jun 24 01:37:48 CEST 2009
53 - provide explicit functions to "cede" to the event loop in Coro::AnyEvent,
54 as this seems to have been a difficult concept (poll, sleep, idle,
55 idle_upto).
56 - add Coro::AnyEvent::readable/writable functions.
57 - clarify Coro::EV/Event/AnyEvent manpages.
58 - free per-thread global scalars in the thread calling ->cancel, to
59 avoid crashes when $_, $@ etc., are magical but some of those
60 had already been freed.
61 - "unexperimentalise" the callback interface for Coro::Semaphore.
62 - speed up ready queue management/context switching by using a linked
63 list instead of an array (~5-10%).
64 - implement "watch" command in Coro::Debug shells.
65 - for fun, implement time-slicing as an example in the manpage.
66 - if AnyEvent detects EV or Event, but we don't have Coro::EV or
67 Coro::Event, use the normal AnyEvent handling instead of dieing
68 (the same is true for Coro::Handle).
69 - properly document Coro::EV::timed_io_once.
70 - avoid unneccessary ->cancel calls in Coro::Handle.
71 - maybe make it work on mingw32 with win32 backend
72 (based on patch by Yasuhiro Matsumoto).
73
745.132 Fri May 29 09:00:39 CEST 2009
75 - do not keep a reference to the argument itself in
76 Coro::Semaphore::guard, as it could change later.
77 - support SO_RCVBUF/SO_SNDBUF nonstandard Coro::Socket options,
78 should support a prepare callback.
79
805.131 Mon Mar 16 23:20:37 CET 2009
81 - implement and document Coro->suspend, Coro->resume.
82 - fix Coro::Select implementation to not (often) close
83 the passed file descriptors (testcase provided by pippijn).
84
855.13 Mon Dec 15 21:51:42 CET 2008
86 - EXPERIMENTAL: implement dynamic winds (on_enter/on_leave).
87 - don't set diehook to C<undef> but instead to NULL, to avoid
88 spurious warnings.
89 - fix a lot of bugs in Coro::SemaphoreSet.
90 - Coro::SemaphoreSet will less often create a semaphore needlessly.
91 - add Coro::SemaphoreSet::count and wait methods.
92 - take advantage of the new Guard module.
93 - deprecate Coro::guard.
94 - try to fix the dreaded 01_unblock tests once more. I hate it when
95 testsuites need more fixing than the code they are supposed to test.
96 - croak in more cases when a required callback isn't resolvable.
97 - fix some minor issues in Coro::State->call/eval.
98 - use current coroutine context instead of a temporary one
99 when temporarily switching to another coroutine.
100 - do not call C-level on_destroy handlers during global destruction,
101 to avoid needless segfaults.
102
1035.12 Sun Dec 7 13:30:38 CET 2008
104 - add default config for MirOS, which seems to be a bug-to-bug
105 compatible fork of openbsd ("world domination by releasing
106 openbsd cvs before the openbsd folks do it" or so :).
107 - free_padlist did destroy the names pad, not good, but didn't
108 seem to bother perl - this could fix issues such as eval ""
109 inside a function called from multiple coroutines.
110 - use a different method to detect destruction time.
111 - be more careful when freeing padlists just before global
112 destruction.
113 - fixed and expanded the call/cc example.
114 - renamed _terminate to _coro_run.
115 - new method Coro::Channel->shutdown.
116 - try pthreads on openbsd <4.4 (broken sigaltstack, will
117 pthreads fare better?).
118 - be less picky about destroying "a" running coroutine.
119
1205.11 Tue Nov 25 21:49:05 CET 2008
121 - DEBUGGING in 5.10.0 is a mess: it constantly flags perfectly
122 working code with failed assertions, introducing more bugs than
123 it fixes, requiring elaborate workarounds :(
124
1255.1 Mon Nov 24 08:54:59 CET 2008
126 - wrote a small introductory tutorial - Coro::Intro.
127 - convert Coro::Timer, Coro::Select and Coro::Util to rouse API.
128 - Coro::Select did errornously dup the file descriptors
129 and didn't work with all AnyEvent backends.
130 - Coro::Select wasn't imported correctly form Coro::LWP, causing blocking
131 LWP data transfers.
132 - disassociate c contexts from coro objects - this is agruably more
133 correct, but mostly allows sharing of cctxs between coro and state
134 objects, for added memory savings and speed increases.
135 - bumped $Coro::POOL_RSS up to 32kb by default.
136 - no longer set the optype to OP_CUSTOM, as B::* understandably
137 doesn't like this very much (and we *are* a type of entersub).
138 - implement state cloning, just to prove that call/cc can be done.
139 - automatically load Coro::AnyEvent in Coro::Handle.
140 - wrap ->cancel calls in eval inside Coro::Handle as EV watchers
141 do not have this method (and don't need it either).
142 - speed up generic anyevent methods in Coro::Handle by using rouse
143 callbacks.
144 - allow coroutines in $Coro::IDLE, speeding up Coro::AnyEvent and
145 others. It also makes the debugger happier, as you can trace
146 through the idle threads now.
147 - add comppad_name* and hints ($^H, %^H) to per-thread variables.
148 - eg/event was pretty broken.
149 - better 5.8.6 compatibility.
150
1515.0 Thu Nov 20 10:35:05 CET 2008
14 - NEW ARCHITECTURE: use the latest 4.x version if you experience 152 - NEW ARCHITECTURE: use the latest 4.x version if you experience
15 stability issues. 153 stability issues.
16 - bump API version to 7 - all dependents must be recompiled. 154 - bump API version to 7 - all dependents must be recompiled.
155 - removed timed_* functions - they were not being used anyways
156 and should be replaced by a more generic mechanism -
157 and were annoying to support anyways :)
158 - removed SemaphoreSet's waiter method - use sem method instead.
17 - Coro::Semaphore->adjust didn't correctly wake up enough waiters. 159 - Coro::Semaphore->adjust didn't correctly wake up enough waiters.
18 - async_pool did free a scalar value twice 160 - async_pool did free a scalar value twice
19 ("Attempt to unreference..."). 161 ("Attempt to unreference...").
20 - fix a longstanding bug where calling terminate on a coro that 162 - fix a longstanding bug where calling terminate on a coro that
21 was waiting for a semaphore that was just becoming available 163 was waiting for a semaphore that was just becoming available
22 would cause a deadlock (semaphore would get into a state where 164 would cause a deadlock (semaphore would get into a state where
23 it was available but waiters were waiting(. 165 it was available but waiters were still blocked).
24 - avoid deadlock in Coro::Channel if maxsize == 1 (Richard Hundt). 166 - calling throw on a coroutine that is waiting for a semaphore will
167 no longer make it acquire the semaphore (and thus leak a count).
25 - perl's process emulation is now not even theoretically supported 168 - perl's process emulation is now not even theoretically supported
26 anymore. 169 anymore.
170 - new functions Coro::rouse_cb and Coro::rouse_wait for easier
171 conversion of callback-style to blocking-style.
172 - new methods $coro->schedule_to and ->cede_to, to specifically
173 schedule or cede to a specific coroutine.
174 - new function Coro::Semaphore::wait.
175 - use named constants in Coro::Channel (Richard Hundt).
27 - directly patch the entersub opcode calling SLF functions (cede, 176 - directly patch the entersub opcode calling SLF functions (cede,
28 transfer and so on). this does speed up context switching, but 177 transfer and so on). this does speed up context switching, but
29 more importanly, it frees us from the hardcoded behaviour of 178 more importanly, it frees us from the hardcoded behaviour of
30 entersub, so we might actually be able to return something from 179 entersub, so we might actually be able to return something from
31 those functions and atcually create new ones. 180 those functions and atcually create new ones.
… …
34 as "threaded" perls are running at half speed anyways). 183 as "threaded" perls are running at half speed anyways).
35 - implement execute_slf (schedule-like-function) interface that makes 184 - implement execute_slf (schedule-like-function) interface that makes
36 it possible to implement schedule-like-functions in XS. 185 it possible to implement schedule-like-functions in XS.
37 - use new SLF interface to massively speed up Coro::EV by roughly a 186 - use new SLF interface to massively speed up Coro::EV by roughly a
38 factor of two. 187 factor of two.
39 - removed timed_down and timed_guard functions - they were not being
40 used anyways and should be replaced by a more generic mechanism -
41 and were annoying to support anyways :)
42 - used new SLF interface to massively speed up Coro::Semaphore by a 188 - used new SLF interface to massively speed up Coro::Semaphore by a
43 factor of three. 189 factor of three.
44 - used new SLF interface to speed up Coro::AIO by roughly a factor of 190 - used new SLF interface to speed up Coro::AIO by roughly a factor of
45 four and reduce its memory usage considerably. 191 four and reduce its memory usage considerably.
192 - implement Coro::SemaphoreSet purely in terms of Coro::Semaphore,
193 for a nice speedup and vastly more correct behaviour. Also implement
194 a new method "sem" to get at the underlying semaphore object.
195 - implement Coro::Channel in terms of Coro::Semaphore, for a moderate
196 (in comparison) 20-40% speedup.
197 - used new SLF interface to reimplement Coro::Signal gaining
198 some unknown (because I was too lazy), but certain, speedup, and also
199 making signals reliable for the first time.
200 - used new SLF interface and other optimisations to speed up async_pool
201 by a factor of two. It also doesn't rely on perl's exception mechanism
202 to exit anymore. The overhead for terminating an async_pool, coro over
203 a normal async is now very small.
204 - sped up coroutine creation/destruction by 40%.
46 - forgot to include Coro/libcoro/README in the dist for all these years. 205 - forgot to include Coro/libcoro/README in the dist for all these years.
47 - work around a freebsd pthreads bug (manual testcancel is required as 206 - work around a freebsd pthreads bug (manual testcancel is required as
48 pthread_cond_wait isn't a cancellation point on freebsd). 207 pthread_cond_wait isn't a cancellation point on freebsd).
208 - use new rouse functions to speed up and simplify Coro::BDB.
209 - make "prefer perl native functions" work with threaded perls.
210 - condense Coro::Debug ps output, hint at v and w flags.
211 - (libcoro) lots of minor cleanups and portability improvements.
212
2134.914 Wed Nov 19 12:54:18 CET 2008
214 - fix a disastrous bug in the readline optimisation
215 introduced in 4.801.
49 216
504.913 Sat Nov 15 07:58:28 CET 2008 2174.913 Sat Nov 15 07:58:28 CET 2008
51 - async_pool did free a scalar value twice 218 - async_pool did free a scalar value twice
52 ("Attempt to unreference..."). 219 ("Attempt to unreference...").
53 220

Diff Legend

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