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

Comparing Coro/Changes (file contents):
Revision 1.249 by root, Tue Oct 2 23:16:23 2007 UTC vs.
Revision 1.535 by root, Wed Apr 14 01:56:02 2010 UTC

1Revision history for Perl extension Coro. 1Revision history for Perl extension Coro.
2 2
3TODO: destroy gets only called on cancel: fix(?) 3TODO: should explore PerlIO::coroaio (perl leaks like hell).
4TODO: should explore PerlIO::via::CoroCede and PerlIO::via::CoroAIO. 4TODO: unready_all
5TODO: maybe implement a default message channel, very much like Erlang 5TODO: myhttpd header parsing
6 (which is cool in a lot of important aspects (failures!), 6TODO: channel->maxsize(newsize)?
7 but very lacking in others (higher level ipc)). 7TODO: http://www.microsoft.com/msj/archive/s2ce.aspx
8 8
93.8 95.22 Wed Apr 14 03:55:35 CEST 2010
10 - correctly return udnef on errors in Coro::Handle::read/write
11 (testcase by Marc Wims).
12 - convert Coro::Util into a "perl compatibility wrapper" - the functions
13 are less useful now, but are drop-in replacements for existing
14 functions, listing better alternatives in the documentation. This also
15 fixes a bug in Coro::LWP which naively substituted Socket::inet_aton
16 with Coro::Util::inet_aton.
17 - do not override $Coro::idle unconditionally in Coro.pm, as other
18 modules could have provided their own idle coro already
19 (for exmaple, Coro::AnyEvent).
20 - fix Coro::Util::gethost* functions.
21 - Coro::Timer corretcly exports it's symbols (reported by Hideki Yamamura).
22
235.21 Wed Dec 16 07:19:51 CET 2009
24 - automatically load Coro::AnyEvent when AnyEvent and Coro are used
25 together.
26 - add some examples on how to combine other event loops with Coro in
27 Coro::AnyEvent, and how to run it (and not to block). Seems to be
28 the most common source of confusion.
29 - try to catch people naively blocking in an event callback.
30 - work around the perl filehandle bug issue in conjunction with
31 older common::sense (as indirectly pointed out by ZSystem).
32 - clarify the "not from signal handlers" section.
33
345.2 Sun Oct 4 14:54:24 CEST 2009
35 - Coro::Storable destroyed the prototypes of the functions it wrapped.
36 - export rouse_cb and rouse_wait by default now.
37 - fix various prototype mismatches in Coro::AnyEvent and Coro::Handle.
38 - new method $state->swap_sv.
39 - added section on "windows process emulation" to the manpage,
40 after a not-so-fruitful (nor-friendly) "discussion" with chip
41 salzenberg (discussion implies arguments, but his only arguments
42 were ad-hominems, one wonders why he started it in the first
43 place). I hope this explains it well enough for him to understand,
44 and maybe well enough for others to understand.
45 - use common::sense everywhere now.
46 - idle callbacks are no longer supported, use idle coros instead.
47 - print a thread listing when a deadlock is detected.
48
495.17 Sat Aug 22 23:09:31 CEST 2009
50 - work around a bug in the perl debugger causing crashes
51 when running under the debugger by marking _pool_handler as nodebug.
52 - speed up Coro::async considerably.
53 - try some hacks to get netbsd to work "more often" - their broken
54 setjmp/longjmp, ucontext *and* phtreads are really hard on Coro.
55 - convert Coro to AE 5.0 API.
56
575.162 Tue Jul 28 04:04:03 CEST 2009
58 - perl 5.8.2 is now minimum requirement.
59 - skip t/19_handle.t on broken windows perls.
60
615.161 Wed Jul 22 04:47:38 CEST 2009
62 - Coro::AnyEvent::poll could have a different prototype when EV was
63 used as backend (analyzed by Tatsuhiko Miyagawa).
64 - Coro::AnyEvent errornously initialised the event loop when loaded,
65 not on demand.
66 - try to workaround rare */t/01_unblock.t failures.
67
685.16 Tue Jul 21 01:44:37 CEST 2009
69 - Coro::AnyEvent failed to hook into the event loop
70 when no threads had been readied between detecting
71 the event loop and actually running it.
72 - considerably speed up Coro::Select by taking avdantage
73 of AnyEvent > 4.8 and some other optimisations.
74 - implement paragraph readline mode in Coro::Handle
75 (based on patches by Zsbán Ambrus).
76 - replace WSAEINPROGRESS by WSAEWOULDBLOCK (reported
77 and analyzed by Yasuhiro MATSUMOTO).
78 - clarified libcoro license and copyright.
79 - someone stole my EXTRA_META!!!
80 - implement Coro::Select::patch_pp_sselect and it's brother,
81 for hardcode select overriding.
82
835.151 Mon Jul 6 05:41:57 CEST 2009
84 - backport to windows process emulation code again (patch by
85 Yasuhiro MATSUMOTO).
86 - slightly update Coro::MakeMaker.
87
885.15 Tue Jun 30 10:28:06 CEST 2009
89 - deprecate Coro::Socket, document how to get ipv6 support via
90 AnyEvent::Socket instead.
91 - implement signal->wait ($cb) interface, similar to semaphores.
92 - work around SvOK not supporting getmagic, so we have to getmagic
93 to test for undef :/ (reported by Matthias Waldorf).
94 - load Coro::AnyEvent in all modules using AnyEvent.
95 - work around perl corrupting our internal data structures,
96 reported by Tokuhiro Matsuno.
97 - enable per-coroutine real and cpu time gathering
98 (Coro::State::enable_times).
99
1005.14 Wed Jun 24 01:37:48 CEST 2009
101 - provide explicit functions to "cede" to the event loop in Coro::AnyEvent,
102 as this seems to have been a difficult concept (poll, sleep, idle,
103 idle_upto).
104 - add Coro::AnyEvent::readable/writable functions.
105 - clarify Coro::EV/Event/AnyEvent manpages.
106 - free per-thread global scalars in the thread calling ->cancel, to
107 avoid crashes when $_, $@ etc., are magical but some of those
108 had already been freed.
109 - "unexperimentalise" the callback interface for Coro::Semaphore.
110 - speed up ready queue management/context switching by using a linked
111 list instead of an array (~5-10%).
112 - implement "watch" command in Coro::Debug shells.
113 - for fun, implement time-slicing as an example in the manpage.
114 - if AnyEvent detects EV or Event, but we don't have Coro::EV or
115 Coro::Event, use the normal AnyEvent handling instead of dieing
116 (the same is true for Coro::Handle).
117 - properly document Coro::EV::timed_io_once.
118 - avoid unneccessary ->cancel calls in Coro::Handle.
119 - maybe make it work on mingw32 with win32 backend
120 (based on patch by Yasuhiro Matsumoto).
121
1225.132 Fri May 29 09:00:39 CEST 2009
123 - do not keep a reference to the argument itself in
124 Coro::Semaphore::guard, as it could change later.
125 - support SO_RCVBUF/SO_SNDBUF nonstandard Coro::Socket options,
126 should support a prepare callback.
127
1285.131 Mon Mar 16 23:20:37 CET 2009
129 - implement and document Coro->suspend, Coro->resume.
130 - fix Coro::Select implementation to not (often) close
131 the passed file descriptors (testcase provided by pippijn).
132
1335.13 Mon Dec 15 21:51:42 CET 2008
134 - EXPERIMENTAL: implement dynamic winds (on_enter/on_leave).
135 - don't set diehook to C<undef> but instead to NULL, to avoid
136 spurious warnings.
137 - fix a lot of bugs in Coro::SemaphoreSet.
138 - Coro::SemaphoreSet will less often create a semaphore needlessly.
139 - add Coro::SemaphoreSet::count and wait methods.
140 - take advantage of the new Guard module.
141 - deprecate Coro::guard.
142 - try to fix the dreaded 01_unblock tests once more. I hate it when
143 testsuites need more fixing than the code they are supposed to test.
144 - croak in more cases when a required callback isn't resolvable.
145 - fix some minor issues in Coro::State->call/eval.
146 - use current coroutine context instead of a temporary one
147 when temporarily switching to another coroutine.
148 - do not call C-level on_destroy handlers during global destruction,
149 to avoid needless segfaults.
150
1515.12 Sun Dec 7 13:30:38 CET 2008
152 - add default config for MirOS, which seems to be a bug-to-bug
153 compatible fork of openbsd ("world domination by releasing
154 openbsd cvs before the openbsd folks do it" or so :).
155 - free_padlist did destroy the names pad, not good, but didn't
156 seem to bother perl - this could fix issues such as eval ""
157 inside a function called from multiple coroutines.
158 - use a different method to detect destruction time.
159 - be more careful when freeing padlists just before global
160 destruction.
161 - fixed and expanded the call/cc example.
162 - renamed _terminate to _coro_run.
163 - new method Coro::Channel->shutdown.
164 - try pthreads on openbsd <4.4 (broken sigaltstack, will
165 pthreads fare better?).
166 - be less picky about destroying "a" running coroutine.
167
1685.11 Tue Nov 25 21:49:05 CET 2008
169 - DEBUGGING in 5.10.0 is a mess: it constantly flags perfectly
170 working code with failed assertions, introducing more bugs than
171 it fixes, requiring elaborate workarounds :(
172
1735.1 Mon Nov 24 08:54:59 CET 2008
174 - wrote a small introductory tutorial - Coro::Intro.
175 - convert Coro::Timer, Coro::Select and Coro::Util to rouse API.
176 - Coro::Select did errornously dup the file descriptors
177 and didn't work with all AnyEvent backends.
178 - Coro::Select wasn't imported correctly form Coro::LWP, causing blocking
179 LWP data transfers.
180 - disassociate c contexts from coro objects - this is agruably more
181 correct, but mostly allows sharing of cctxs between coro and state
182 objects, for added memory savings and speed increases.
183 - bumped $Coro::POOL_RSS up to 32kb by default.
184 - no longer set the optype to OP_CUSTOM, as B::* understandably
185 doesn't like this very much (and we *are* a type of entersub).
186 - implement state cloning, just to prove that call/cc can be done.
187 - automatically load Coro::AnyEvent in Coro::Handle.
188 - wrap ->cancel calls in eval inside Coro::Handle as EV watchers
189 do not have this method (and don't need it either).
190 - speed up generic anyevent methods in Coro::Handle by using rouse
191 callbacks.
192 - allow coroutines in $Coro::IDLE, speeding up Coro::AnyEvent and
193 others. It also makes the debugger happier, as you can trace
194 through the idle threads now.
195 - add comppad_name* and hints ($^H, %^H) to per-thread variables.
196 - eg/event was pretty broken.
197 - better 5.8.6 compatibility.
198
1995.0 Thu Nov 20 10:35:05 CET 2008
200 - NEW ARCHITECTURE: use the latest 4.x version if you experience
201 stability issues.
202 - bump API version to 7 - all dependents must be recompiled.
203 - removed timed_* functions - they were not being used anyways
204 and should be replaced by a more generic mechanism -
205 and were annoying to support anyways :)
206 - removed SemaphoreSet's waiter method - use sem method instead.
207 - Coro::Semaphore->adjust didn't correctly wake up enough waiters.
208 - async_pool did free a scalar value twice
209 ("Attempt to unreference...").
210 - fix a longstanding bug where calling terminate on a coro that
211 was waiting for a semaphore that was just becoming available
212 would cause a deadlock (semaphore would get into a state where
213 it was available but waiters were still blocked).
214 - calling throw on a coroutine that is waiting for a semaphore will
215 no longer make it acquire the semaphore (and thus leak a count).
216 - perl's process emulation is now not even theoretically supported
217 anymore.
218 - new functions Coro::rouse_cb and Coro::rouse_wait for easier
219 conversion of callback-style to blocking-style.
220 - new methods $coro->schedule_to and ->cede_to, to specifically
221 schedule or cede to a specific coroutine.
222 - new function Coro::Semaphore::wait.
223 - use named constants in Coro::Channel (Richard Hundt).
224 - directly patch the entersub opcode calling SLF functions (cede,
225 transfer and so on). this does speed up context switching, but
226 more importanly, it frees us from the hardcoded behaviour of
227 entersub, so we might actually be able to return something from
228 those functions and atcually create new ones.
229 - take advantage of __builtin_frame_address on gcc.
230 - expose THX in coroapi (not sure whether this was a wise decision,
231 as "threaded" perls are running at half speed anyways).
232 - implement execute_slf (schedule-like-function) interface that makes
233 it possible to implement schedule-like-functions in XS.
234 - use new SLF interface to massively speed up Coro::EV by roughly a
235 factor of two.
236 - used new SLF interface to massively speed up Coro::Semaphore by a
237 factor of three.
238 - used new SLF interface to speed up Coro::AIO by roughly a factor of
239 four and reduce its memory usage considerably.
240 - implement Coro::SemaphoreSet purely in terms of Coro::Semaphore,
241 for a nice speedup and vastly more correct behaviour. Also implement
242 a new method "sem" to get at the underlying semaphore object.
243 - implement Coro::Channel in terms of Coro::Semaphore, for a moderate
244 (in comparison) 20-40% speedup.
245 - used new SLF interface to reimplement Coro::Signal gaining
246 some unknown (because I was too lazy), but certain, speedup, and also
247 making signals reliable for the first time.
248 - used new SLF interface and other optimisations to speed up async_pool
249 by a factor of two. It also doesn't rely on perl's exception mechanism
250 to exit anymore. The overhead for terminating an async_pool, coro over
251 a normal async is now very small.
252 - sped up coroutine creation/destruction by 40%.
253 - forgot to include Coro/libcoro/README in the dist for all these years.
254 - work around a freebsd pthreads bug (manual testcancel is required as
255 pthread_cond_wait isn't a cancellation point on freebsd).
256 - use new rouse functions to speed up and simplify Coro::BDB.
257 - make "prefer perl native functions" work with threaded perls.
258 - condense Coro::Debug ps output, hint at v and w flags.
259 - (libcoro) lots of minor cleanups and portability improvements.
260
2614.914 Wed Nov 19 12:54:18 CET 2008
262 - fix a disastrous bug in the readline optimisation
263 introduced in 4.801.
264
2654.913 Sat Nov 15 07:58:28 CET 2008
266 - async_pool did free a scalar value twice
267 ("Attempt to unreference...").
268
2694.912 Thu Nov 13 18:31:23 CET 2008
270 - minor cleanups.
271 - use much larger stacks on linux and perl < 5.8.8.
272 - Coro::Debug::new_unix_server did not unlink the socket
273 when destroyed.
274
2754.911 Tue Nov 11 04:26:17 CET 2008
276 - "port" to threaded perls.
277
2784.91 Mon Nov 10 05:36:38 CET 2008
279 - the ->throw exception object no longer leaks.
280 - creating a new cctx leaked a scopestack entry (memleak).
281 - new coroutines didn't get created with a zero flags field
282 (unknown impact).
283 - calling ->throw on a not-yet-started coroutine should now work
284 instead of being ignored.
285 - ->throw is now supported on Coro::State objects.
286 - clean up cctx creation code a bit.
287 - entersub is actually a UNOP, not a LOGOP (not a bugfix).
288
2894.9 Sat Nov 8 17:45:27 CET 2008
290 - (libcoro) did not preserve rbp with CORO_ASM (we are getting there).
291 - (libcoro) no longer leak threads in the experimental pthread backend,
292 also speed it up considerably.
293 - (libcoro) do not rely on makecontext passing void *'s unscathed.
294 - fix compiletime dependencies on libcoro in the Makefile.
295 - cctx_count wasn't always updated properly.
296 - Coro::State::cctx_stacksize wasn't applied correctly.
297 - new function Coro::State::cctx_max_idle.
298 - the default max number of idle C contexts is now 4.
299 - (libcoro) try harder to get _setjmp/_longjmp.
300 - (libcoro) cleanup and extend the libcoro API to officially
301 allow the creation of empty source contexts.
302 - very experimental workaround for the totally broken netbsd platform.
303 - tried to hack around openbsd bugs.
304
3054.804 Wed Nov 5 16:36:00 CET 2008
306 - Coro::Debug::new_unix_server would not create a non-blocking listening
307 socket, sometimes causing freezes.
308 - (libcoro) fix misaligned stack points for setjmp and assembly
309 methods, which can cause crashes on x86/x86_64 with a sufficiently
310 aggressive compiler.
311 - new function: Coro::Debug::new_tcp_server.
312 - move ->throw into the Coro class because it only works on coro objects.
313
3144.803 Mon Nov 3 17:16:12 CET 2008
315 - (libcoro) use a global asm statement to become independent of gcc
316 otpimisations for CORO_ASM (thanks to pippijn for the idea).
317 - try to workaround yet another broken bsd, this time dragonfly.
318
3194.802 Thu Oct 30 10:56:12 CET 2008
320 - support -fno-omit-frame-pointer on x86 with the assembly method.
321 - tune 01_unblock.t tests a bit.
322
3234.801 Wed Oct 22 18:33:37 CEST 2008
324 - improve readline speed for very long "lines".
325 - backport to 5.8.8.
326
3274.8 Thu Oct 2 13:34:40 CEST 2008
328 - new function Coro::AIO::aio_wait.
329 - Coro.:AIO and Coro::BDB now "use Coro::AnyEvent".
330 - greatly speed up and reduce memory usage of Coro::AIO requests.
331 - implement some other µ-optimisations.
332
3334.749 Mon Sep 29 14:40:12 CEST 2008
334 - port to slow and broken pseudo-threaded perls. (courtesy pippijn).
335
3364.748 Sat Sep 27 14:03:19 CEST 2008
337 - implement, but do not document, PerlIO::cede(granularity).
338 - Coro::Storable forgot to wrap Storable::pstore.
339 - work around the multitude of leaks and memory corruption
340 bugs in PerlIO::via by using our own C-level perliol. As a side
341 effect, Coro::Storable is now much, much, much faster.
342
3434.747 Tue Sep 23 01:59:41 CEST 2008
344 - fix a per-cv memleak (one empty array was leaked per
345 code reference).
346 - avoid a crash in coro->call|rss when the coroutine was already
347 destroyed (most noticably when using Coro::Debug::ps :)
348 - also protect *Storable::FILE.
349 - push up default storable granularity to 20ms.
350
3514.746 Sun Sep 21 03:22:20 CEST 2008
352 - be more insistent on locking Storable against reentrancy
353 in Coro::Storable.
354 - move swap_def?v and throw to Coro::State, as documented.
355
3564.745 Thu Jul 24 00:14:38 CEST 2008
357 - remove debugging code related to MgPV_nolen_const, also try to
358 make it compile with perl 5.8.6 (yes, apple apparently loves
359 outdated software). Reported by John S.
360
3614.744 Tue Jul 8 22:06:35 CEST 2008
362 - correctly provide default DIE/WARN handlers as documented.
363 - also overwrite PL_vtbl_sigelem.svt_clear, even though current
364 implementations inside perl work fine for us.
365
3664.743 Mon Jun 16 00:21:57 CEST 2008
367 - when using Coro::EV without running EV::loop it could
368 result in busy-waiting for events, this has been fixed.
369 - reduce codesize and improve performance by using EV_DEFAULT_UC.
370
3714.742 Sat May 31 14:10:21 CEST 2008
372 - implement a workaround for (some) perl <5.8.8 versions.
373 - require EV 3.3+.
374
3754.741 Fri May 30 23:33:09 CEST 2008
376 - tell netbsd how utterly broken their imitation of an OS is
377 and refuse to build by default if pthreads are in use.
378 - switch to "s" method on all bsd's by default, as their ucontext
379 stuff seems just too broken.
380 - fix a bug in Coro::Select.
381
3824.74 Thu May 29 20:05:31 CEST 2008
383 - do not test Coro::LWP for lack of dependencies.
384 - do not test Coro::Select for lack of working perls.
385
3864.73 Thu May 29 2008
387 - fix a bug in Coro::EV which would cause it to block despite
388 there being runnable coroutines.
389 - sprinkle "no warnings" freely over everything, also suppress
390 warnings for some other modules.
391 - fix typo in WSAEWOULDBLOCK.
392
3934.72 Sun May 25 05:14:36 CEST 2008
394 - tweak META.yaml a bit, unfortunately, there is no documented way
395 to have optional dependencies with CPAN. doh :(
396 - avoid running some tests on windows because they would fail due to
397 perl bug (broken fork, broken pipes...).
398 - work around perl on windows bugs where perl returns undocumented
399 error codes for sysread, syswrite etc. by taking advantage
400 of AnyEvent's workaround for that problem.
401 - use AnyEvent::Util::fh_nonblocking in Coro::Handle to work around
402 a common perl implementation bug on windows.
403 - use unix domain sockets in testsuite to work around
404 common perl implementation bugs on widows (they are emulated by
405 tcp sockets on windows. ugh.)
406
4074.71 Sat May 24 20:01:27 CEST 2008
408 - fix a bug in Coro::AnyEvent ("Usage: Coro::AnyEvent::_schedule()").
409 - take advantage of async name resolution of AnyEvent::Util.
410 - work around brutal inet_aton override in Coro::LWP.
411 - take advantage of the readyhook in Coro::EV, for smoother
412 scheduling.
413
4144.7 Sun May 11 00:32:19 CEST 2008
415 - completely reworked the Coro manpage.
416 - added Coro::AnyEvent, generic event loop integration.
417 - implement cancel, ready and kill commands in Coro::Debug.
418 - document find_coro in Coro::Debug.
419 - incompatible change: rename has_stack to has_cctx.
420 - Coro::AIO and Coro::BDB no longer force event model detection,
421 use AnyEvent::AIO and AnyEvent::BDB.
422
4234.6 Sat Apr 26 10:05:14 CEST 2008
424 - INCOMPATIBLE CHANGE: sub/code attributes are no longer supported
425 by the Coro module. It was a mistake to have it in the first place.
426 - (experimental) support for activestate perl 5.10 (method "w").
427 - (experimental) support for strawberry perl (method "a").
428 - coro_sigelem_set did not return a value although it had to,
429 actual impact unknown.
430
4314.51 Mon Apr 14 13:28:27 CEST 2008
432 - make it compile again on 5.8.
433
4344.50 Thu Apr 10 09:43:17 CEST 2008
435 - I did it twice! (see 4.49).
436
4374.49 Sun Apr 6 21:23:31 CEST 2008
438 - grr, instead of compiling the recent changes on 5.10 only they
439 were compiled on 5.8 only.
440
4414.48 Sun Apr 6 20:36:46 CEST 2008
442 - allow coroutine switches during eval's under 5.10.x, as it seems
443 the parser is a per-interpreter option now, so this is safe
444 (this might fix the odd crashing bug).
445 - drop support for 5.9.x versions: they are dead, jim.
446
4474.47 Sun Apr 6 00:37:52 CEST 2008
448 - force cctx allocation on API calls: we know we need to force one
449 and gcc actually manages to confuse our heuristic nowadays,
450 leading to crashes and worse.
451 - document force_cctx.
452
4534.46 Fri Apr 4 22:05:43 CEST 2008
454 - upgrade libcoro, resulting in pthread-backend (which was only created
455 to fulfill the rules of the programming languages shootout).
456
4574.45 Thu Mar 13 11:55:36 CET 2008
458 - fix a file leaking bug in eg/mhyttpd that would allow
459 downloading of any file (reported by oesi).
460 - fix deadlock bug in Coro::Channel (reported by Richard Hundt)
461 (also add testcase).
462 - support Broadcast option in Coro::Socket (patch by Richard Hundt,
463 apparently having loads of fun with that).
464 - implement $state->swap_defsv, ->swap_defav and document ->throw.
465
4664.4 Wed Feb 13 16:44:29 CET 2008
467 - only assume .cfi pseudo isns exist on GNU/Linux.
468 - add get_timed method to Coro::Channels.
469 - fixed Coro::Debug manpage.
470 - perl 5.11 compatbility improvement based on patch by
471 Andreas König.
472
4734.37 Sun Jan 20 11:25:23 CET 2008
474 - soften the check for an active parser for perl 5.10 (avoiding
475 "Coro::State::transfer called while parsing" in many cases).
476
4774.36 Sun Jan 13 10:53:56 CET 2008
478 - reset diehook when terminating from an async_pool as to not
479 trigger any __DIE__ handlers.
480
4814.35 Sun Jan 13 04:14:13 CET 2008
482 - "bt" debug command now displays any exceptions
483 from longmess and also skips the innermost
484 stackframes, giving more useufl output.
485 - allow backtraces in coroutines blocked in Coro::EV,
486 at a <1% speed hit.
487 - handle localising of $SIG{__WARN__} and __DIE__
488 properly (with a proper amount of dirty hacking).
489
4904.34 Sat Dec 22 17:49:53 CET 2007
491 - upgrade to EV version 2.0 API.
492
4934.33 Mon Dec 17 08:36:12 CET 2007
494 - make Coro::AIO etc. loadable in the absence of EV.
495
4964.32 Mon Dec 17 07:46:02 CET 2007
497 - majorly improved Coro::Handle's performance with Coro::EV.
498 - implemented the until now mythical Coro::BDB module.
499 - specialcase EV in Coro::AIO and Coro::BDB for extra speed.
500
5014.31 Wed Dec 5 12:32:39 CET 2007
502 - remove warn statement form Coro::Util.
503
5044.3 Tue Dec 4 20:33:14 CET 2007
505 - calls to the idle function could cause stack corruption
506 when the stack changed.
507 - do no longer rely on the presence of EV::DNS if EV is used
508 (because it is gone), but instead take avdantage of EV::ADNS
509 if available.
510 - add ($) prototypes to all functions in Coro::Storable.
511 - use a conventional (safer) idle callback in Coro::EV.
512 - do accept quickly in Coro::Debug to avoid endless loops.
513
5144.22 Fri Nov 30 16:04:04 CET 2007
515 - really use optimised versions for Event and EV in Coro::Util
516 and Coro::Handle.
517
5184.21 Sun Nov 25 10:48:59 CET 2007
519 - fix a spurious memory read.
520 - Coro::EV no longer keeps the eventloop "alive".
521
5224.2 Fri Nov 9 20:47:05 CET 2007
523 - enable/disable tracing from a new coroutine, not a pooled one.
524 - fix a memleak in Coro::Event.
525 - removed killall call from fork_eval.
526 - made sure store_fd is already loaded so that fork_eval
527 does not have to parse autoload in each subprocess.
528 - only use assembly method if -O switch is in $Config{optimize}.
529 - add (optional) Coro::EV module, so far the best event loop module
530 directly supported by Coro.
531 - if the event model is EV, use EV::DNS to resolve
532 stuff in Coro::Util.
533 - don't get confused by multiple event notifications in Coro::Handle.
534 - initial support for EV (libevent interface).
535 - require Event and EV using configure_requires, to force their existance.
536
5374.13 Wed Oct 24 07:26:45 CEST 2007
538 - add Coro::Storable::blocking_thaw.
539 - use a vastly more complicated technique to localise
540 $SIG{__WARN/DIE__} that also works on perls <= 5.8.8.
541 - use a coroutine for the idle callback Coro::Event,
542 instead of running Event in the current coroutine context.
543 This also catches recursive invocations.
544 - actually report fork errors in gethostbyname and inet_aton.
545
5464.11 Thu Oct 11 02:40:24 CEST 2007
547 - port to threaded perls.
548
5494.1 Thu Oct 11 02:38:16 CEST 2007
550 - incompatible change: $SIG{__DIE__} and $SIG{__WARN__} will now
551 be local to each coro (see Coro::State).
552 - incompatible change: for very deep reasons, cede and cede_notself
553 cannot return anything, so nothing will be returned.
554 - possibly bring back 5.10 compatibility (untested).
555 - work around stupid (and wrong) warning on 5.10 :(.
556 - overlay the saved state over the context stack. This saves
557 a few hundred bytes per coroutine on average and also
558 speeds up context switching a bit.
559 - further tune default stack sizes.
560 - (more) correctly calculate stack usage in coro_rss.
561 - Coro::Storable::blocking_* did not properly lock
562 resulting in races between coroutines.
563 - added Coro::Storable::guard.
564 - stopping to trace a coroutine could destroy the cctx of
565 an unrelated coroutine.
566 - explain the relationship between Perl and C coroutines in
567 more detail in Coro::State.
568 - Coro::Util::inet_aton did not short-circuit dotted quad forms,
569 causing a fork per resolve. This also affected Coro::Socket.
570 - switch to a separate stack in $coro->call/eval to avoid
571 invalidating pointers.
572
5734.03 Sat Oct 6 21:24:00 CEST 2007
574 - added Coro::throw method.
575 - minor code cleanups.
576
5774.02 Sat Oct 6 02:36:47 CEST 2007
578 - fix a very minor per-coroutine memleak (a single codereference).
579 - fixed a bug where the currently in-use c context would be freed
580 prematurely (can happen only when programs change the stacksize
581 or use tracing).
582 - tracing can no longer keep a coro alive after it terminated.
583 - do static branch prediction in the common path for gcc. gives
584 about 2-5% speed improvement here.
585
5864.01 Fri Oct 5 22:10:49 CEST 2007
587 - instead of recreating *a* standard output handle we simply
588 use STDOUT, which is faster and hopefully more robust.
589
5904.0 Fri Oct 5 12:56:00 CEST 2007
10 - incompatibly changed Coro::Storable::freeze. 591 - incompatibly changed Coro::Storable::freeze.
11 - major new feature: added Coro::Debug, for interactive coroutine 592 - major new feature: added Coro::Debug, for interactive coroutine
12 debugging, tracing and much more. 593 debugging, tracing and much more.
594 - major bug fix: unbelievable, but true: $_, $/ and many other
595 "saved" variables actually weren't being saved. This has been fixed,
596 of course, while increasing performance while losing all the save
597 flags.
598 - save flags are gone, and all the api functions dealing with them.
13 - added Coro::Semaphore::adjust. 599 - added Coro::Semaphore::adjust.
14 - added Coro::Util::fork_eval. 600 - added Coro::Util::fork_eval.
15 - added Coro::Storable::{nfreeze,blocking_{freeze,nfreeze}}. 601 - added Coro::Storable::{nfreeze,blocking_{freeze,nfreeze}}.
16 - added Coro::killall. 602 - added Coro::killall.
17 - reduce initial stack sizes to allow for "micro-coroutines". 603 - reduce initial stack sizes to allow for "micro-coroutines".
… …
19 handling to XS (major speed improvement). 605 handling to XS (major speed improvement).
20 - actually croak before modifying important data structures. 606 - actually croak before modifying important data structures.
21 - refuse to transfer while compiling. 607 - refuse to transfer while compiling.
22 - possibly support eval EXPR better now. 608 - possibly support eval EXPR better now.
23 - enable assembly per default on linux+bsd x86+amd64. 609 - enable assembly per default on linux+bsd x86+amd64.
24 - all internal members are named _something for easier subclassing. 610 - all internal members were renamed _something for easier subclassing.
25 - many minor tweaks. 611 - many minor tweaks.
26 612
273.63 Wed May 16 14:10:06 CEST 2007 6133.63 Wed May 16 14:10:06 CEST 2007
28 - implement handcoded assembly for x86/amd64 SVR ABI. 614 - implement handcoded assembly for x86/amd64 SVR ABI.
29 615

Diff Legend

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