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

Comparing Coro/Changes (file contents):
Revision 1.418 by root, Sun Nov 16 00:55:41 2008 UTC vs.
Revision 1.521 by root, Sat Aug 22 22:36:23 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
85.0 115.17 Sat Aug 22 23:09:31 CEST 2009
9 - NEW ARCHITECTURE: use 4.912 for the stable version. 12 - work around a bug in the perl debugger causing crashes
13 when running under the debugger by marking _pool_handler as nodebug.
14 - speed up Coro::async considerably.
15 - try some hacks to get netbsd to work "more often" - their broken
16 setjmp/longjmp, ucontext *and* phtreads are really hard on Coro.
17 - convert Coro to AE 5.0 API.
18
195.162 Tue Jul 28 04:04:03 CEST 2009
20 - perl 5.8.2 is now minimum requirement.
21 - skip t/19_handle.t on broken windows perls.
22
235.161 Wed Jul 22 04:47:38 CEST 2009
24 - Coro::AnyEvent::poll could have a different prototype when EV was
25 used as backend (analyzed by Tatsuhiko Miyagawa).
26 - Coro::AnyEvent errornously initialised the event loop when loaded,
27 not on demand.
28 - try to workaround rare */t/01_unblock.t failures.
29
305.16 Tue Jul 21 01:44:37 CEST 2009
31 - Coro::AnyEvent failed to hook into the event loop
32 when no threads had been readied between detecting
33 the event loop and actually running it.
34 - considerably speed up Coro::Select by taking avdantage
35 of AnyEvent > 4.8 and some other optimisations.
36 - implement paragraph readline mode in Coro::Handle
37 (based on patches by Zsbán Ambrus).
38 - replace WSAEINPROGRESS by WSAEWOULDBLOCK (reported
39 and analyzed by Yasuhiro MATSUMOTO).
40 - clarified libcoro license and copyright.
41 - someone stole my EXTRA_META!!!
42 - implement Coro::Select::patch_pp_sselect and it's brother,
43 for hardcode select overriding.
44
455.151 Mon Jul 6 05:41:57 CEST 2009
46 - backport to windows process emulation code again (patch by
47 Yasuhiro MATSUMOTO).
48 - slightly update Coro::MakeMaker.
49
505.15 Tue Jun 30 10:28:06 CEST 2009
51 - deprecate Coro::Socket, document how to get ipv6 support via
52 AnyEvent::Socket instead.
53 - implement signal->wait ($cb) interface, similar to semaphores.
54 - work around SvOK not supporting getmagic, so we have to getmagic
55 to test for undef :/ (reported by Matthias Waldorf).
56 - load Coro::AnyEvent in all modules using AnyEvent.
57 - work around perl corrupting our internal data structures,
58 reported by Tokuhiro Matsuno.
59 - enable per-coroutine real and cpu time gathering
60 (Coro::State::enable_times).
61
625.14 Wed Jun 24 01:37:48 CEST 2009
63 - provide explicit functions to "cede" to the event loop in Coro::AnyEvent,
64 as this seems to have been a difficult concept (poll, sleep, idle,
65 idle_upto).
66 - add Coro::AnyEvent::readable/writable functions.
67 - clarify Coro::EV/Event/AnyEvent manpages.
68 - free per-thread global scalars in the thread calling ->cancel, to
69 avoid crashes when $_, $@ etc., are magical but some of those
70 had already been freed.
71 - "unexperimentalise" the callback interface for Coro::Semaphore.
72 - speed up ready queue management/context switching by using a linked
73 list instead of an array (~5-10%).
74 - implement "watch" command in Coro::Debug shells.
75 - for fun, implement time-slicing as an example in the manpage.
76 - if AnyEvent detects EV or Event, but we don't have Coro::EV or
77 Coro::Event, use the normal AnyEvent handling instead of dieing
78 (the same is true for Coro::Handle).
79 - properly document Coro::EV::timed_io_once.
80 - avoid unneccessary ->cancel calls in Coro::Handle.
81 - maybe make it work on mingw32 with win32 backend
82 (based on patch by Yasuhiro Matsumoto).
83
845.132 Fri May 29 09:00:39 CEST 2009
85 - do not keep a reference to the argument itself in
86 Coro::Semaphore::guard, as it could change later.
87 - support SO_RCVBUF/SO_SNDBUF nonstandard Coro::Socket options,
88 should support a prepare callback.
89
905.131 Mon Mar 16 23:20:37 CET 2009
91 - implement and document Coro->suspend, Coro->resume.
92 - fix Coro::Select implementation to not (often) close
93 the passed file descriptors (testcase provided by pippijn).
94
955.13 Mon Dec 15 21:51:42 CET 2008
96 - EXPERIMENTAL: implement dynamic winds (on_enter/on_leave).
97 - don't set diehook to C<undef> but instead to NULL, to avoid
98 spurious warnings.
99 - fix a lot of bugs in Coro::SemaphoreSet.
100 - Coro::SemaphoreSet will less often create a semaphore needlessly.
101 - add Coro::SemaphoreSet::count and wait methods.
102 - take advantage of the new Guard module.
103 - deprecate Coro::guard.
104 - try to fix the dreaded 01_unblock tests once more. I hate it when
105 testsuites need more fixing than the code they are supposed to test.
106 - croak in more cases when a required callback isn't resolvable.
107 - fix some minor issues in Coro::State->call/eval.
108 - use current coroutine context instead of a temporary one
109 when temporarily switching to another coroutine.
110 - do not call C-level on_destroy handlers during global destruction,
111 to avoid needless segfaults.
112
1135.12 Sun Dec 7 13:30:38 CET 2008
114 - add default config for MirOS, which seems to be a bug-to-bug
115 compatible fork of openbsd ("world domination by releasing
116 openbsd cvs before the openbsd folks do it" or so :).
117 - free_padlist did destroy the names pad, not good, but didn't
118 seem to bother perl - this could fix issues such as eval ""
119 inside a function called from multiple coroutines.
120 - use a different method to detect destruction time.
121 - be more careful when freeing padlists just before global
122 destruction.
123 - fixed and expanded the call/cc example.
124 - renamed _terminate to _coro_run.
125 - new method Coro::Channel->shutdown.
126 - try pthreads on openbsd <4.4 (broken sigaltstack, will
127 pthreads fare better?).
128 - be less picky about destroying "a" running coroutine.
129
1305.11 Tue Nov 25 21:49:05 CET 2008
131 - DEBUGGING in 5.10.0 is a mess: it constantly flags perfectly
132 working code with failed assertions, introducing more bugs than
133 it fixes, requiring elaborate workarounds :(
134
1355.1 Mon Nov 24 08:54:59 CET 2008
136 - wrote a small introductory tutorial - Coro::Intro.
137 - convert Coro::Timer, Coro::Select and Coro::Util to rouse API.
138 - Coro::Select did errornously dup the file descriptors
139 and didn't work with all AnyEvent backends.
140 - Coro::Select wasn't imported correctly form Coro::LWP, causing blocking
141 LWP data transfers.
142 - disassociate c contexts from coro objects - this is agruably more
143 correct, but mostly allows sharing of cctxs between coro and state
144 objects, for added memory savings and speed increases.
145 - bumped $Coro::POOL_RSS up to 32kb by default.
146 - no longer set the optype to OP_CUSTOM, as B::* understandably
147 doesn't like this very much (and we *are* a type of entersub).
148 - implement state cloning, just to prove that call/cc can be done.
149 - automatically load Coro::AnyEvent in Coro::Handle.
150 - wrap ->cancel calls in eval inside Coro::Handle as EV watchers
151 do not have this method (and don't need it either).
152 - speed up generic anyevent methods in Coro::Handle by using rouse
153 callbacks.
154 - allow coroutines in $Coro::IDLE, speeding up Coro::AnyEvent and
155 others. It also makes the debugger happier, as you can trace
156 through the idle threads now.
157 - add comppad_name* and hints ($^H, %^H) to per-thread variables.
158 - eg/event was pretty broken.
159 - better 5.8.6 compatibility.
160
1615.0 Thu Nov 20 10:35:05 CET 2008
162 - NEW ARCHITECTURE: use the latest 4.x version if you experience
163 stability issues.
10 - bump API version to 7 - all dependents must be recompiled. 164 - bump API version to 7 - all dependents must be recompiled.
165 - removed timed_* functions - they were not being used anyways
166 and should be replaced by a more generic mechanism -
167 and were annoying to support anyways :)
168 - removed SemaphoreSet's waiter method - use sem method instead.
11 - Coro::Semaphore->adjust didn't correctly wake up enough waiters. 169 - Coro::Semaphore->adjust didn't correctly wake up enough waiters.
12 - async_pool did free a scalar value twice 170 - async_pool did free a scalar value twice
13 ("Attempt to unreference..."). 171 ("Attempt to unreference...").
172 - fix a longstanding bug where calling terminate on a coro that
173 was waiting for a semaphore that was just becoming available
174 would cause a deadlock (semaphore would get into a state where
175 it was available but waiters were still blocked).
176 - calling throw on a coroutine that is waiting for a semaphore will
177 no longer make it acquire the semaphore (and thus leak a count).
14 - perl's process emulation is now not even theoretically supported 178 - perl's process emulation is now not even theoretically supported
15 anymore. 179 anymore.
180 - new functions Coro::rouse_cb and Coro::rouse_wait for easier
181 conversion of callback-style to blocking-style.
182 - new methods $coro->schedule_to and ->cede_to, to specifically
183 schedule or cede to a specific coroutine.
184 - new function Coro::Semaphore::wait.
185 - use named constants in Coro::Channel (Richard Hundt).
16 - directly patch the entersub opcode calling SLF functions (cede, 186 - directly patch the entersub opcode calling SLF functions (cede,
17 transfer and so on). this does speed up context switching, but 187 transfer and so on). this does speed up context switching, but
18 more importanly, it frees us from the hardcoded behaviour of 188 more importanly, it frees us from the hardcoded behaviour of
19 entersub, so we might actually be able to return something from 189 entersub, so we might actually be able to return something from
20 those functions and atcually create new ones. 190 those functions and atcually create new ones.
21 - take advantage of __builtin_frame_address on gcc. 191 - take advantage of __builtin_frame_address on gcc.
22 - expose THX in coroapi (not sure whether this was a wise decision, 192 - expose THX in coroapi (not sure whether this was a wise decision,
23 as "threaded" perls are running at half speed anyways). 193 as "threaded" perls are running at half speed anyways).
24 - implement execute_slf (schedule-like-function) interface that makes 194 - implement execute_slf (schedule-like-function) interface that makes
25 it possible to implement schedule-like-functions in XS. 195 it possible to implement schedule-like-functions in XS.
26 - use new SLF interface to massively speed up Coro::EV by roughly a factor 196 - use new SLF interface to massively speed up Coro::EV by roughly a
27 of two. 197 factor of two.
28 - removed timed_down and timed_guard functions - they were not being used
29 anyways and should be replaced by a more generic mechanism - and were
30 annoying to support anyways :
31 - used new SLF interface to massively speed up Coro::Semaphore by a factor 198 - used new SLF interface to massively speed up Coro::Semaphore by a
32 of three. 199 factor of three.
200 - used new SLF interface to speed up Coro::AIO by roughly a factor of
201 four and reduce its memory usage considerably.
202 - implement Coro::SemaphoreSet purely in terms of Coro::Semaphore,
203 for a nice speedup and vastly more correct behaviour. Also implement
204 a new method "sem" to get at the underlying semaphore object.
205 - implement Coro::Channel in terms of Coro::Semaphore, for a moderate
206 (in comparison) 20-40% speedup.
207 - used new SLF interface to reimplement Coro::Signal gaining
208 some unknown (because I was too lazy), but certain, speedup, and also
209 making signals reliable for the first time.
210 - used new SLF interface and other optimisations to speed up async_pool
211 by a factor of two. It also doesn't rely on perl's exception mechanism
212 to exit anymore. The overhead for terminating an async_pool, coro over
213 a normal async is now very small.
214 - sped up coroutine creation/destruction by 40%.
33 - forgot to include Coro/libcoro/README in the dist for all these years. 215 - forgot to include Coro/libcoro/README in the dist for all these years.
34 - work around a freebsd pthreads bug (manual testcancel is required as 216 - work around a freebsd pthreads bug (manual testcancel is required as
35 pthread_cond_wait isn't a cancellation point on freebsd). 217 pthread_cond_wait isn't a cancellation point on freebsd).
218 - use new rouse functions to speed up and simplify Coro::BDB.
219 - make "prefer perl native functions" work with threaded perls.
220 - condense Coro::Debug ps output, hint at v and w flags.
221 - (libcoro) lots of minor cleanups and portability improvements.
222
2234.914 Wed Nov 19 12:54:18 CET 2008
224 - fix a disastrous bug in the readline optimisation
225 introduced in 4.801.
36 226
374.913 Sat Nov 15 07:58:28 CET 2008 2274.913 Sat Nov 15 07:58:28 CET 2008
38 - async_pool did free a scalar value twice 228 - async_pool did free a scalar value twice
39 ("Attempt to unreference..."). 229 ("Attempt to unreference...").
40 230

Diff Legend

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