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

Comparing Coro/Changes (file contents):
Revision 1.456 by root, Fri Nov 21 07:29:04 2008 UTC vs.
Revision 1.586 by root, Mon Aug 8 22:00:15 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
76.06 Mon Aug 8 23:59:48 CEST 2011
8 - cygwin unfortunately patches the stack at runtime, so we use the pthreads
9 backend, which is an order of magnitutde slower. unfortunately, cygwins
10 pthread implementation isn't very complete either, so allocate the stack
11 twice just to be sure.
12 (note: cygwin also enables mymalloc, which is NOT THREADSAFE ON WINDOWS,
13 in its ithreaded perl - best recompile cygwin and use the 'w'indows
14 backend for much better performance. also disable ithreads for
15 even better performance...).
16
176.05 Thu Aug 4 21:36:36 CEST 2011
18 - blush, condvar values would not be propagated from send to recv anymore
19 (reported by Chip Salzenberg).
20 - use exponential increase for the readline buffer length in
21 Coro::Handle. also reduce initial allocation to 1020 from 4096 bytes.
22
236.04 Wed Aug 3 17:15:45 CEST 2011
24 - use even more efficient and more compatible condvars for
25 compatibility to AnyEvent 6.x :)
26 - more inconsequential ecb.h updates.
27
286.03 Wed Aug 3 11:41:30 CEST 2011
29 - change how Coro patches AnyEvent condvars for compatibility to
30 AnyEvent 6.x.
31 - update ecb.h, to no longer include <pthread.h> in case WinNT.h
32 hasn't been included.
33
346.02 Wed Jul 13 04:35:19 CEST 2011
35 - "improve portability to Gentoo" - gentoo manages to put perl variables
36 in memory areas that are farther than 2gb apart, which the jit couldn't
37 handle and barfed. now it's just a bit slower on gentoo and similar
38 systems.
39
406.01 Sun Jul 3 12:31:14 CEST 2011
41 - workarounds are good, but the test for whether pthreads are used
42 was not good. thisone should be better.
43 - check differently whether gcc generates cfi instructions itself.
44
456.0 Wed Jun 29 19:43:35 CEST 2011
46 - INCOMPATIBLE CHANGE: unreferenced coro objects will now be
47 destroyed and cleaned up automatically (e.g. async { schedule }).
48 - implement a JIT compiler for part of the thread switch code,
49 which gives a 50% speed improvement on threaded perls, and
50 about 4% on non-thraeded perls (so threaded perls now finally
51 reach about half the speed of non-threaded perls).
52 - slightly modernise Coro::Intro, add section about rouse functions.
53 - avoid DEFSV and ERRSV, giving another 10% improvement
54 in thread switching.
55 - Coro::State->is_destroyed is now called is_zombie.
56 - implement a Coro->safe_cancel method that might fail, but
57 cancels in a "safer" way if it succeeds.
58 - add preliminary support for DEBUGGING perls.
59 - get rid of two hash-accesses when initialising a new Coro - this
60 speeds up coro creation by almost a factor of two.
61 - croak when a coro that is being cancelled tries to block
62 (e.g. while executing a guard block), instead of crashing or
63 deadlocking.
64 - use a more robust and also faster method to identify Coro::State
65 objects - speeds up everything a bit.
66 - implement Coro->cancel in XS for a 20% speed improvement, and to
67 be able to implement mutual cancellation.
68 - speed up context switches by a percent or two by more efficiently
69 allocating context stack entries.
70 - implement Coro->join and Coro->on_destroy in XS for a speedup and
71 a reduction in memory use.
72 - cancelling a coro while it itself is cancelling another coro is
73 now supported and working, instead of triggering an assertion.
74 - be a bit more crash-resistant when calling (buggy) on_destroy
75 callbacks (best effort).
76 - move on_destroy into the slf_frame, to allow extension slf
77 functions to have destructors.
78 - get rid if coro refcounting - simply crash in other interpreter
79 threads by nulling the pointers on clone.
80 - simplify warn/die hook handling when loading Coro - the convoluted
81 logic seems to be no longer neccessary.
82 - use libecb instead of our own home-grown gcc hacks.
83 - document alternatives to Coro::LWP. Please use them :)
84 - work around another mindless idiotic NEEDLESS bug in openbsd/mirbsds
85 sigaltstack. Really. wine suffers from it, erlang suffers from it,
86 and it's known since at least 2006.
87
885.372 Wed Feb 23 06:14:30 CET 2011
89 - apparently mingw doesn't provide a working gettimeofday, try to
90 work around that by relying on Time::HiRes (indirectly brought to
91 my attention by Max Maischein).
92 - fix some portability issues when Time::HiRes was used.
93
945.371 Mon Feb 21 14:36:08 CET 2011
95 - backport to windows process emulation code again.
96
975.37 Sat Feb 19 07:49:44 CET 2011
98 - add a big "Coro thread life cycle" section to "man Coro".
99 - try a tentative workaround against the breakage that 5.13 has
100 introduced without depreciation period. sigh.
101 - no longer use Time::HiRes if gettimeofday is available, which
102 saves quite a lot of memory.
103
1045.36 Sun Feb 13 05:33:41 CET 2011
105 - automatically load Coro::Channel, Coro::RWLock, Coro::Semaphore,
106 Coro::SemaphoreSet, Coro::Signal and Coro::Specific on first "new"
107 method call.
108 - undocument Coro::Timer::sleep and obsolete whole module.
109 - optimise Coro::Timer::timeout memory and cpu usage.
110 - slightly updated Coro::Intro for recent changes.
111 - do not initialise PL_dirty anymore.
112
1135.25 Thu Nov 11 01:08:39 CET 2010
114 - try a different approach on netbsd - netbsd 5 finally has marginally
115 working pthreads, but still broken ucontext/sigaltstack.
116 - openbsd 4.8 finally got their act together, Coro works out of the box
117 with asm, setjmp and pthreads (no change, just informational).
118
1195.24 Sat Oct 23 11:27:12 CEST 2010
120 - port to the EV 4.0 API.
121 - work around bugs in mingw32, making strawberry perl work
122 out of the box.
123 - correctly modify Coro::AIO function prototypes
124 so that they reflect the "no optional parameters" rule.
125 - "ported" libcoro to C++.
126
1275.23 Mon May 17 18:50:42 CEST 2010
128 - be more resistant to ordering changes when initialising
129 Coro::AnyEvent, Coro::EV and Coro::Event (reported by Matthias
130 Waldorf).
131 - document that perl 5.12 deliberately removed support for cloning.
132
1335.22 Wed Apr 14 03:55:35 CEST 2010
134 - correctly return udnef on errors in Coro::Handle::read/write
135 (testcase by Marc Mims).
136 - convert Coro::Util into a "perl compatibility wrapper" - the functions
137 are less useful now, but are drop-in replacements for existing
138 functions, listing better alternatives in the documentation. This also
139 fixes a bug in Coro::LWP which naively substituted Socket::inet_aton
140 with Coro::Util::inet_aton.
141 - do not override $Coro::idle unconditionally in Coro.pm, as other
142 modules could have provided their own idle coro already
143 (for exmaple, Coro::AnyEvent).
144 - fix Coro::Util::gethost* functions.
145 - Coro::Timer corretcly exports it's symbols (reported by Hideki Yamamura).
146
1475.21 Wed Dec 16 07:19:51 CET 2009
148 - automatically load Coro::AnyEvent when AnyEvent and Coro are used
149 together.
150 - add some examples on how to combine other event loops with Coro in
151 Coro::AnyEvent, and how to run it (and not to block). Seems to be
152 the most common source of confusion.
153 - try to catch people naively blocking in an event callback.
154 - work around the perl filehandle bug issue in conjunction with
155 older common::sense (as indirectly pointed out by ZSystem).
156 - clarify the "not from signal handlers" section.
157
1585.2 Sun Oct 4 14:54:24 CEST 2009
159 - Coro::Storable destroyed the prototypes of the functions it wrapped.
160 - export rouse_cb and rouse_wait by default now.
161 - fix various prototype mismatches in Coro::AnyEvent and Coro::Handle.
162 - new method $state->swap_sv.
163 - added section on "windows process emulation" to the manpage,
164 after a not-so-fruitful (nor-friendly) "discussion" with chip
165 salzenberg (discussion implies arguments, but his only arguments
166 were ad-hominems, one wonders why he started it in the first
167 place). I hope this explains it well enough for him to understand,
168 and maybe well enough for others to understand.
169 - use common::sense everywhere now.
170 - idle callbacks are no longer supported, use idle coros instead.
171 - print a thread listing when a deadlock is detected.
172
1735.17 Sat Aug 22 23:09:31 CEST 2009
174 - work around a bug in the perl debugger causing crashes
175 when running under the debugger by marking _pool_handler as nodebug.
176 - speed up Coro::async considerably.
177 - try some hacks to get netbsd to work "more often" - their broken
178 setjmp/longjmp, ucontext *and* phtreads are really hard on Coro.
179 - convert Coro to AE 5.0 API.
180
1815.162 Tue Jul 28 04:04:03 CEST 2009
182 - perl 5.8.2 is now minimum requirement.
183 - skip t/19_handle.t on broken windows perls.
184
1855.161 Wed Jul 22 04:47:38 CEST 2009
186 - Coro::AnyEvent::poll could have a different prototype when EV was
187 used as backend (analyzed by Tatsuhiko Miyagawa).
188 - Coro::AnyEvent errornously initialised the event loop when loaded,
189 not on demand.
190 - try to workaround rare */t/01_unblock.t failures.
191
1925.16 Tue Jul 21 01:44:37 CEST 2009
193 - Coro::AnyEvent failed to hook into the event loop
194 when no threads had been readied between detecting
195 the event loop and actually running it.
196 - considerably speed up Coro::Select by taking avdantage
197 of AnyEvent > 4.8 and some other optimisations.
198 - implement paragraph readline mode in Coro::Handle
199 (based on patches by Zsbán Ambrus).
200 - replace WSAEINPROGRESS by WSAEWOULDBLOCK (reported
201 and analyzed by Yasuhiro MATSUMOTO).
202 - clarified libcoro license and copyright.
203 - someone stole my EXTRA_META!!!
204 - implement Coro::Select::patch_pp_sselect and it's brother,
205 for hardcode select overriding.
206
2075.151 Mon Jul 6 05:41:57 CEST 2009
208 - backport to windows process emulation code again (patch by
209 Yasuhiro MATSUMOTO).
210 - slightly update Coro::MakeMaker.
211
2125.15 Tue Jun 30 10:28:06 CEST 2009
213 - deprecate Coro::Socket, document how to get ipv6 support via
214 AnyEvent::Socket instead.
215 - implement signal->wait ($cb) interface, similar to semaphores.
216 - work around SvOK not supporting getmagic, so we have to getmagic
217 to test for undef :/ (reported by Matthias Waldorf).
218 - load Coro::AnyEvent in all modules using AnyEvent.
219 - work around perl corrupting our internal data structures,
220 reported by Tokuhiro Matsuno.
221 - enable per-coroutine real and cpu time gathering
222 (Coro::State::enable_times).
223
2245.14 Wed Jun 24 01:37:48 CEST 2009
225 - provide explicit functions to "cede" to the event loop in Coro::AnyEvent,
226 as this seems to have been a difficult concept (poll, sleep, idle,
227 idle_upto).
228 - add Coro::AnyEvent::readable/writable functions.
229 - clarify Coro::EV/Event/AnyEvent manpages.
230 - free per-thread global scalars in the thread calling ->cancel, to
231 avoid crashes when $_, $@ etc., are magical but some of those
232 had already been freed.
233 - "unexperimentalise" the callback interface for Coro::Semaphore.
234 - speed up ready queue management/context switching by using a linked
235 list instead of an array (~5-10%).
236 - implement "watch" command in Coro::Debug shells.
237 - for fun, implement time-slicing as an example in the manpage.
238 - if AnyEvent detects EV or Event, but we don't have Coro::EV or
239 Coro::Event, use the normal AnyEvent handling instead of dieing
240 (the same is true for Coro::Handle).
241 - properly document Coro::EV::timed_io_once.
242 - avoid unneccessary ->cancel calls in Coro::Handle.
243 - maybe make it work on mingw32 with win32 backend
244 (based on patch by Yasuhiro Matsumoto).
245
2465.132 Fri May 29 09:00:39 CEST 2009
247 - do not keep a reference to the argument itself in
248 Coro::Semaphore::guard, as it could change later.
249 - support SO_RCVBUF/SO_SNDBUF nonstandard Coro::Socket options,
250 should support a prepare callback.
251
2525.131 Mon Mar 16 23:20:37 CET 2009
253 - implement and document Coro->suspend, Coro->resume.
254 - fix Coro::Select implementation to not (often) close
255 the passed file descriptors (testcase provided by pippijn).
256
2575.13 Mon Dec 15 21:51:42 CET 2008
258 - EXPERIMENTAL: implement dynamic winds (on_enter/on_leave).
259 - don't set diehook to C<undef> but instead to NULL, to avoid
260 spurious warnings.
261 - fix a lot of bugs in Coro::SemaphoreSet.
262 - Coro::SemaphoreSet will less often create a semaphore needlessly.
263 - add Coro::SemaphoreSet::count and wait methods.
264 - take advantage of the new Guard module.
265 - deprecate Coro::guard.
266 - try to fix the dreaded 01_unblock tests once more. I hate it when
267 testsuites need more fixing than the code they are supposed to test.
268 - croak in more cases when a required callback isn't resolvable.
269 - fix some minor issues in Coro::State->call/eval.
270 - use current coroutine context instead of a temporary one
271 when temporarily switching to another coroutine.
272 - do not call C-level on_destroy handlers during global destruction,
273 to avoid needless segfaults.
274
2755.12 Sun Dec 7 13:30:38 CET 2008
276 - add default config for MirOS, which seems to be a bug-to-bug
277 compatible fork of openbsd ("world domination by releasing
278 openbsd cvs before the openbsd folks do it" or so :).
279 - free_padlist did destroy the names pad, not good, but didn't
280 seem to bother perl - this could fix issues such as eval ""
281 inside a function called from multiple coroutines.
282 - use a different method to detect destruction time.
283 - be more careful when freeing padlists just before global
284 destruction.
285 - fixed and expanded the call/cc example.
286 - renamed _terminate to _coro_run.
287 - new method Coro::Channel->shutdown.
288 - try pthreads on openbsd <4.4 (broken sigaltstack, will
289 pthreads fare better?).
290 - be less picky about destroying "a" running coroutine.
291
2925.11 Tue Nov 25 21:49:05 CET 2008
293 - DEBUGGING in 5.10.0 is a mess: it constantly flags perfectly
294 working code with failed assertions, introducing more bugs than
295 it fixes, requiring elaborate workarounds :(
296
2975.1 Mon Nov 24 08:54:59 CET 2008
298 - wrote a small introductory tutorial - Coro::Intro.
8 - convert Coro::Timer, Coro::Select and Coro::Util to rouse API. 299 - convert Coro::Timer, Coro::Select and Coro::Util to rouse API.
9 - Coro::Select did errornously dup the file descriptors 300 - Coro::Select did errornously dup the file descriptors
10 and didn't work with all AnyEvent backends. 301 and didn't work with all AnyEvent backends.
302 - Coro::Select wasn't imported correctly form Coro::LWP, causing blocking
303 LWP data transfers.
304 - disassociate c contexts from coro objects - this is agruably more
305 correct, but mostly allows sharing of cctxs between coro and state
306 objects, for added memory savings and speed increases.
11 - bumped $Coro::POOL_RSS up to 32kb by default. 307 - bumped $Coro::POOL_RSS up to 32kb by default.
12 - no longer set the optype to OP_CUSTOM, as B::* understandably 308 - no longer set the optype to OP_CUSTOM, as B::* understandably
13 doesn't like this very much (and we *are* a type of entersub). 309 doesn't like this very much (and we *are* a type of entersub).
310 - implement state cloning, just to prove that call/cc can be done.
311 - automatically load Coro::AnyEvent in Coro::Handle.
312 - wrap ->cancel calls in eval inside Coro::Handle as EV watchers
313 do not have this method (and don't need it either).
314 - speed up generic anyevent methods in Coro::Handle by using rouse
315 callbacks.
316 - allow coroutines in $Coro::IDLE, speeding up Coro::AnyEvent and
317 others. It also makes the debugger happier, as you can trace
318 through the idle threads now.
319 - add comppad_name* and hints ($^H, %^H) to per-thread variables.
320 - eg/event was pretty broken.
321 - better 5.8.6 compatibility.
14 322
155.0 Thu Nov 20 10:35:05 CET 2008 3235.0 Thu Nov 20 10:35:05 CET 2008
16 - NEW ARCHITECTURE: use the latest 4.x version if you experience 324 - NEW ARCHITECTURE: use the latest 4.x version if you experience
17 stability issues. 325 stability issues.
18 - bump API version to 7 - all dependents must be recompiled. 326 - bump API version to 7 - all dependents must be recompiled.

Diff Legend

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