1 | Revision history for Perl extension Coro. |
1 | Revision history for Perl extension Coro. |
2 | |
2 | |
3 | TODO: should explore PerlIO::coroaio (perl leaks like hell). |
3 | TODO: should explore PerlIO::coroaio (perl leaks like hell). |
4 | TODO: maybe implement a default message channel, very much like Erlang's |
4 | TODO: unready_all |
5 | actor model (which is cool in a lot of important aspects (failures!), |
5 | TODO: myhttpd header parsing |
6 | but very lacking in others (higher level ipc)). |
6 | TODO: channel->maxsize(newsize)? |
7 | |
7 | |
8 | TODO: testcancel |
8 | - convert Coro::Util into a "perl compatibility wrapper" - the functions |
9 | TODO: aio |
9 | are less useful now, but are drop-in replacements for existing |
10 | TODO: Coro::Semaphore |
10 | functions, listing better alternatives in the documentation. This also |
11 | TODO: trow disturbs cleanup path inside pp_slf? |
11 | fixes a bug in Coro::LWP which naively substituted Socket::inet_aton |
12 | TODO: throw special-case in transfer? |
12 | with Coro::Util::inet_aton. |
13 | 5.0 |
13 | - do not override $Coro::idle unconditionally in Coro.pm, as other |
|
|
14 | modules could have provided their own idle coro already |
|
|
15 | (for exmaple, Coro::AnyEvent). |
|
|
16 | - fix Coro::Util::gethost* functions. |
|
|
17 | |
|
|
18 | 5.21 Wed Dec 16 07:19:51 CET 2009 |
|
|
19 | - automatically load Coro::AnyEvent when AnyEvent and Coro are used |
|
|
20 | together. |
|
|
21 | - add some examples on how to combine other event loops with Coro in |
|
|
22 | Coro::AnyEvent, and how to run it (and not to block). Seems to be |
|
|
23 | the most common source of confusion. |
|
|
24 | - try to catch people naively blocking in an event callback. |
|
|
25 | - work around the perl filehandle bug issue in conjunction with |
|
|
26 | older common::sense (as indirectly pointed out by ZSystem). |
|
|
27 | - clarify the "not from signal handlers" section. |
|
|
28 | |
|
|
29 | 5.2 Sun Oct 4 14:54:24 CEST 2009 |
|
|
30 | - Coro::Storable destroyed the prototypes of the functions it wrapped. |
|
|
31 | - export rouse_cb and rouse_wait by default now. |
|
|
32 | - fix various prototype mismatches in Coro::AnyEvent and Coro::Handle. |
|
|
33 | - new method $state->swap_sv. |
|
|
34 | - added section on "windows process emulation" to the manpage, |
|
|
35 | after a not-so-fruitful (nor-friendly) "discussion" with chip |
|
|
36 | salzenberg (discussion implies arguments, but his only arguments |
|
|
37 | were ad-hominems, one wonders why he started it in the first |
|
|
38 | place). I hope this explains it well enough for him to understand, |
|
|
39 | and maybe well enough for others to understand. |
|
|
40 | - use common::sense everywhere now. |
|
|
41 | - idle callbacks are no longer supported, use idle coros instead. |
|
|
42 | - print a thread listing when a deadlock is detected. |
|
|
43 | |
|
|
44 | 5.17 Sat Aug 22 23:09:31 CEST 2009 |
|
|
45 | - work around a bug in the perl debugger causing crashes |
|
|
46 | when running under the debugger by marking _pool_handler as nodebug. |
|
|
47 | - speed up Coro::async considerably. |
|
|
48 | - try some hacks to get netbsd to work "more often" - their broken |
|
|
49 | setjmp/longjmp, ucontext *and* phtreads are really hard on Coro. |
|
|
50 | - convert Coro to AE 5.0 API. |
|
|
51 | |
|
|
52 | 5.162 Tue Jul 28 04:04:03 CEST 2009 |
|
|
53 | - perl 5.8.2 is now minimum requirement. |
|
|
54 | - skip t/19_handle.t on broken windows perls. |
|
|
55 | |
|
|
56 | 5.161 Wed Jul 22 04:47:38 CEST 2009 |
|
|
57 | - Coro::AnyEvent::poll could have a different prototype when EV was |
|
|
58 | used as backend (analyzed by Tatsuhiko Miyagawa). |
|
|
59 | - Coro::AnyEvent errornously initialised the event loop when loaded, |
|
|
60 | not on demand. |
|
|
61 | - try to workaround rare */t/01_unblock.t failures. |
|
|
62 | |
|
|
63 | 5.16 Tue Jul 21 01:44:37 CEST 2009 |
|
|
64 | - Coro::AnyEvent failed to hook into the event loop |
|
|
65 | when no threads had been readied between detecting |
|
|
66 | the event loop and actually running it. |
|
|
67 | - considerably speed up Coro::Select by taking avdantage |
|
|
68 | of AnyEvent > 4.8 and some other optimisations. |
|
|
69 | - implement paragraph readline mode in Coro::Handle |
|
|
70 | (based on patches by Zsbán Ambrus). |
|
|
71 | - replace WSAEINPROGRESS by WSAEWOULDBLOCK (reported |
|
|
72 | and analyzed by Yasuhiro MATSUMOTO). |
|
|
73 | - clarified libcoro license and copyright. |
|
|
74 | - someone stole my EXTRA_META!!! |
|
|
75 | - implement Coro::Select::patch_pp_sselect and it's brother, |
|
|
76 | for hardcode select overriding. |
|
|
77 | |
|
|
78 | 5.151 Mon Jul 6 05:41:57 CEST 2009 |
|
|
79 | - backport to windows process emulation code again (patch by |
|
|
80 | Yasuhiro MATSUMOTO). |
|
|
81 | - slightly update Coro::MakeMaker. |
|
|
82 | |
|
|
83 | 5.15 Tue Jun 30 10:28:06 CEST 2009 |
|
|
84 | - deprecate Coro::Socket, document how to get ipv6 support via |
|
|
85 | AnyEvent::Socket instead. |
|
|
86 | - implement signal->wait ($cb) interface, similar to semaphores. |
|
|
87 | - work around SvOK not supporting getmagic, so we have to getmagic |
|
|
88 | to test for undef :/ (reported by Matthias Waldorf). |
|
|
89 | - load Coro::AnyEvent in all modules using AnyEvent. |
|
|
90 | - work around perl corrupting our internal data structures, |
|
|
91 | reported by Tokuhiro Matsuno. |
|
|
92 | - enable per-coroutine real and cpu time gathering |
|
|
93 | (Coro::State::enable_times). |
|
|
94 | |
|
|
95 | 5.14 Wed Jun 24 01:37:48 CEST 2009 |
|
|
96 | - provide explicit functions to "cede" to the event loop in Coro::AnyEvent, |
|
|
97 | as this seems to have been a difficult concept (poll, sleep, idle, |
|
|
98 | idle_upto). |
|
|
99 | - add Coro::AnyEvent::readable/writable functions. |
|
|
100 | - clarify Coro::EV/Event/AnyEvent manpages. |
|
|
101 | - free per-thread global scalars in the thread calling ->cancel, to |
|
|
102 | avoid crashes when $_, $@ etc., are magical but some of those |
|
|
103 | had already been freed. |
|
|
104 | - "unexperimentalise" the callback interface for Coro::Semaphore. |
|
|
105 | - speed up ready queue management/context switching by using a linked |
|
|
106 | list instead of an array (~5-10%). |
|
|
107 | - implement "watch" command in Coro::Debug shells. |
|
|
108 | - for fun, implement time-slicing as an example in the manpage. |
|
|
109 | - if AnyEvent detects EV or Event, but we don't have Coro::EV or |
|
|
110 | Coro::Event, use the normal AnyEvent handling instead of dieing |
|
|
111 | (the same is true for Coro::Handle). |
|
|
112 | - properly document Coro::EV::timed_io_once. |
|
|
113 | - avoid unneccessary ->cancel calls in Coro::Handle. |
|
|
114 | - maybe make it work on mingw32 with win32 backend |
|
|
115 | (based on patch by Yasuhiro Matsumoto). |
|
|
116 | |
|
|
117 | 5.132 Fri May 29 09:00:39 CEST 2009 |
|
|
118 | - do not keep a reference to the argument itself in |
|
|
119 | Coro::Semaphore::guard, as it could change later. |
|
|
120 | - support SO_RCVBUF/SO_SNDBUF nonstandard Coro::Socket options, |
|
|
121 | should support a prepare callback. |
|
|
122 | |
|
|
123 | 5.131 Mon Mar 16 23:20:37 CET 2009 |
|
|
124 | - implement and document Coro->suspend, Coro->resume. |
|
|
125 | - fix Coro::Select implementation to not (often) close |
|
|
126 | the passed file descriptors (testcase provided by pippijn). |
|
|
127 | |
|
|
128 | 5.13 Mon Dec 15 21:51:42 CET 2008 |
|
|
129 | - EXPERIMENTAL: implement dynamic winds (on_enter/on_leave). |
|
|
130 | - don't set diehook to C<undef> but instead to NULL, to avoid |
|
|
131 | spurious warnings. |
|
|
132 | - fix a lot of bugs in Coro::SemaphoreSet. |
|
|
133 | - Coro::SemaphoreSet will less often create a semaphore needlessly. |
|
|
134 | - add Coro::SemaphoreSet::count and wait methods. |
|
|
135 | - take advantage of the new Guard module. |
|
|
136 | - deprecate Coro::guard. |
|
|
137 | - try to fix the dreaded 01_unblock tests once more. I hate it when |
|
|
138 | testsuites need more fixing than the code they are supposed to test. |
|
|
139 | - croak in more cases when a required callback isn't resolvable. |
|
|
140 | - fix some minor issues in Coro::State->call/eval. |
|
|
141 | - use current coroutine context instead of a temporary one |
|
|
142 | when temporarily switching to another coroutine. |
|
|
143 | - do not call C-level on_destroy handlers during global destruction, |
|
|
144 | to avoid needless segfaults. |
|
|
145 | |
|
|
146 | 5.12 Sun Dec 7 13:30:38 CET 2008 |
|
|
147 | - add default config for MirOS, which seems to be a bug-to-bug |
|
|
148 | compatible fork of openbsd ("world domination by releasing |
|
|
149 | openbsd cvs before the openbsd folks do it" or so :). |
|
|
150 | - free_padlist did destroy the names pad, not good, but didn't |
|
|
151 | seem to bother perl - this could fix issues such as eval "" |
|
|
152 | inside a function called from multiple coroutines. |
|
|
153 | - use a different method to detect destruction time. |
|
|
154 | - be more careful when freeing padlists just before global |
|
|
155 | destruction. |
|
|
156 | - fixed and expanded the call/cc example. |
|
|
157 | - renamed _terminate to _coro_run. |
|
|
158 | - new method Coro::Channel->shutdown. |
|
|
159 | - try pthreads on openbsd <4.4 (broken sigaltstack, will |
|
|
160 | pthreads fare better?). |
|
|
161 | - be less picky about destroying "a" running coroutine. |
|
|
162 | |
|
|
163 | 5.11 Tue Nov 25 21:49:05 CET 2008 |
|
|
164 | - DEBUGGING in 5.10.0 is a mess: it constantly flags perfectly |
|
|
165 | working code with failed assertions, introducing more bugs than |
|
|
166 | it fixes, requiring elaborate workarounds :( |
|
|
167 | |
|
|
168 | 5.1 Mon Nov 24 08:54:59 CET 2008 |
|
|
169 | - wrote a small introductory tutorial - Coro::Intro. |
|
|
170 | - convert Coro::Timer, Coro::Select and Coro::Util to rouse API. |
|
|
171 | - Coro::Select did errornously dup the file descriptors |
|
|
172 | and didn't work with all AnyEvent backends. |
|
|
173 | - Coro::Select wasn't imported correctly form Coro::LWP, causing blocking |
|
|
174 | LWP data transfers. |
|
|
175 | - disassociate c contexts from coro objects - this is agruably more |
|
|
176 | correct, but mostly allows sharing of cctxs between coro and state |
|
|
177 | objects, for added memory savings and speed increases. |
|
|
178 | - bumped $Coro::POOL_RSS up to 32kb by default. |
|
|
179 | - no longer set the optype to OP_CUSTOM, as B::* understandably |
|
|
180 | doesn't like this very much (and we *are* a type of entersub). |
|
|
181 | - implement state cloning, just to prove that call/cc can be done. |
|
|
182 | - automatically load Coro::AnyEvent in Coro::Handle. |
|
|
183 | - wrap ->cancel calls in eval inside Coro::Handle as EV watchers |
|
|
184 | do not have this method (and don't need it either). |
|
|
185 | - speed up generic anyevent methods in Coro::Handle by using rouse |
|
|
186 | callbacks. |
|
|
187 | - allow coroutines in $Coro::IDLE, speeding up Coro::AnyEvent and |
|
|
188 | others. It also makes the debugger happier, as you can trace |
|
|
189 | through the idle threads now. |
|
|
190 | - add comppad_name* and hints ($^H, %^H) to per-thread variables. |
|
|
191 | - eg/event was pretty broken. |
|
|
192 | - better 5.8.6 compatibility. |
|
|
193 | |
|
|
194 | 5.0 Thu Nov 20 10:35:05 CET 2008 |
14 | - NEW ARCHITECTURE: use the latest 4.x version if you experience |
195 | - NEW ARCHITECTURE: use the latest 4.x version if you experience |
15 | stability issues. |
196 | stability issues. |
16 | - bump API version to 7 - all dependents must be recompiled. |
197 | - bump API version to 7 - all dependents must be recompiled. |
|
|
198 | - removed timed_* functions - they were not being used anyways |
|
|
199 | and should be replaced by a more generic mechanism - |
|
|
200 | and were annoying to support anyways :) |
|
|
201 | - removed SemaphoreSet's waiter method - use sem method instead. |
17 | - Coro::Semaphore->adjust didn't correctly wake up enough waiters. |
202 | - Coro::Semaphore->adjust didn't correctly wake up enough waiters. |
18 | - async_pool did free a scalar value twice |
203 | - async_pool did free a scalar value twice |
19 | ("Attempt to unreference..."). |
204 | ("Attempt to unreference..."). |
20 | - fix a longstanding bug where calling terminate on a coro that |
205 | - fix a longstanding bug where calling terminate on a coro that |
21 | was waiting for a semaphore that was just becoming available |
206 | was waiting for a semaphore that was just becoming available |
22 | would cause a deadlock (semaphore would get into a state where |
207 | would cause a deadlock (semaphore would get into a state where |
23 | it was available but waiters were waiting(. |
208 | it was available but waiters were still blocked). |
24 | - avoid deadlock in Coro::Channel if maxsize == 1 (Richard Hundt). |
209 | - calling throw on a coroutine that is waiting for a semaphore will |
|
|
210 | no longer make it acquire the semaphore (and thus leak a count). |
25 | - perl's process emulation is now not even theoretically supported |
211 | - perl's process emulation is now not even theoretically supported |
26 | anymore. |
212 | anymore. |
|
|
213 | - new functions Coro::rouse_cb and Coro::rouse_wait for easier |
|
|
214 | conversion of callback-style to blocking-style. |
|
|
215 | - new methods $coro->schedule_to and ->cede_to, to specifically |
|
|
216 | schedule or cede to a specific coroutine. |
|
|
217 | - new function Coro::Semaphore::wait. |
|
|
218 | - use named constants in Coro::Channel (Richard Hundt). |
27 | - directly patch the entersub opcode calling SLF functions (cede, |
219 | - directly patch the entersub opcode calling SLF functions (cede, |
28 | transfer and so on). this does speed up context switching, but |
220 | transfer and so on). this does speed up context switching, but |
29 | more importanly, it frees us from the hardcoded behaviour of |
221 | more importanly, it frees us from the hardcoded behaviour of |
30 | entersub, so we might actually be able to return something from |
222 | entersub, so we might actually be able to return something from |
31 | those functions and atcually create new ones. |
223 | those functions and atcually create new ones. |
… | |
… | |
34 | as "threaded" perls are running at half speed anyways). |
226 | as "threaded" perls are running at half speed anyways). |
35 | - implement execute_slf (schedule-like-function) interface that makes |
227 | - implement execute_slf (schedule-like-function) interface that makes |
36 | it possible to implement schedule-like-functions in XS. |
228 | it possible to implement schedule-like-functions in XS. |
37 | - use new SLF interface to massively speed up Coro::EV by roughly a |
229 | - use new SLF interface to massively speed up Coro::EV by roughly a |
38 | factor of two. |
230 | 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 |
231 | - used new SLF interface to massively speed up Coro::Semaphore by a |
43 | factor of three. |
232 | factor of three. |
44 | - used new SLF interface to speed up Coro::AIO by roughly 400% and |
233 | - used new SLF interface to speed up Coro::AIO by roughly a factor of |
45 | reduce its memory usage considerably. |
234 | four and reduce its memory usage considerably. |
|
|
235 | - implement Coro::SemaphoreSet purely in terms of Coro::Semaphore, |
|
|
236 | for a nice speedup and vastly more correct behaviour. Also implement |
|
|
237 | a new method "sem" to get at the underlying semaphore object. |
|
|
238 | - implement Coro::Channel in terms of Coro::Semaphore, for a moderate |
|
|
239 | (in comparison) 20-40% speedup. |
|
|
240 | - used new SLF interface to reimplement Coro::Signal gaining |
|
|
241 | some unknown (because I was too lazy), but certain, speedup, and also |
|
|
242 | making signals reliable for the first time. |
|
|
243 | - used new SLF interface and other optimisations to speed up async_pool |
|
|
244 | by a factor of two. It also doesn't rely on perl's exception mechanism |
|
|
245 | to exit anymore. The overhead for terminating an async_pool, coro over |
|
|
246 | a normal async is now very small. |
|
|
247 | - sped up coroutine creation/destruction by 40%. |
46 | - forgot to include Coro/libcoro/README in the dist for all these years. |
248 | - 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 |
249 | - work around a freebsd pthreads bug (manual testcancel is required as |
48 | pthread_cond_wait isn't a cancellation point on freebsd). |
250 | pthread_cond_wait isn't a cancellation point on freebsd). |
|
|
251 | - use new rouse functions to speed up and simplify Coro::BDB. |
|
|
252 | - make "prefer perl native functions" work with threaded perls. |
|
|
253 | - condense Coro::Debug ps output, hint at v and w flags. |
|
|
254 | - (libcoro) lots of minor cleanups and portability improvements. |
|
|
255 | |
|
|
256 | 4.914 Wed Nov 19 12:54:18 CET 2008 |
|
|
257 | - fix a disastrous bug in the readline optimisation |
|
|
258 | introduced in 4.801. |
49 | |
259 | |
50 | 4.913 Sat Nov 15 07:58:28 CET 2008 |
260 | 4.913 Sat Nov 15 07:58:28 CET 2008 |
51 | - async_pool did free a scalar value twice |
261 | - async_pool did free a scalar value twice |
52 | ("Attempt to unreference..."). |
262 | ("Attempt to unreference..."). |
53 | |
263 | |