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

Comparing Coro/Changes (file contents):
Revision 1.237 by root, Fri Sep 21 02:23:48 2007 UTC vs.
Revision 1.564 by root, Thu May 12 23:55:39 2011 UTC

1Revision history for Perl extension Coro. 1Revision history for Perl extension Coro.
2 2
3TODO: better (earlier) error messages on broken transfer's 3TODO: should explore PerlIO::coroaio (perl leaks like hell).
4TODO: destroy gets only called on cancel: fix(?) 4TODO: channel->maxsize(newsize)?
5TODO: should explore PerlIO::via::CoroCede and PerlIO::via::CoroAIO. 5TODO: http://www.microsoft.com/msj/archive/s2ce.aspx
6TODO: maybe implement a default message channel, very much like Erlang
7 (which is cool in a lot of important aspects (failures!),
8 but very lacking in others).
9 6
103.7 7TODO: tutorial
11 - added Coro::Debug, interactive coroutine debugging. 8 - INCOMPATIBLE CHANGE: unreferenced coro objects will now be
9 destroyed and cleaned up automatically (e.g. async { schedule }).
10 - Coro::State->is_destroyed is now called is_zombie.
11 - croak when a coro that is being cancelled tries to block
12 (e.g. while executing a guard block), instead of crashing or
13 deadlocking.
14 - use a more robust and also faster method to identify Coro::State
15 objects - speeds up everything a bit.
16 - implement a Coro->safe_cancel method that might fail, but
17 cancels in a "safer" way if it succeeds.
18 - implement Coro->cancel in XS for a 20% speed improvement, and to
19 be able to implement mutual cancellation.
20 - speed up context switches by a percent or two by more efficiently
21 allocating context stack entries.
22 - implement Coro->join and Coro->on_destroy in XS for a speedup and
23 a reduction in memory use.
24 - cancelling a coro while it itself is cancelling another coro is
25 now supported and working, instead of triggering an assertion.
26 - be a bit more crash-resistant when calling (buggy) on_destroy
27 callbacks (best effort).
28 - move on_destroy into the slf_frame, to allow extension slf
29 functions to have destructors.
30 - get rid if coro refcounting - simply crash in other interpreter
31 threads by nulling the pointers on clone.
32
335.372 Wed Feb 23 06:14:30 CET 2011
34 - apparently mingw doesn't provide a working gettimeofday, try to
35 work around that by relying on Time::HiRes (indirectly brought to
36 my attention by Max Maischein).
37 - fix some portability issues when Time::HiRes was used.
38
395.371 Mon Feb 21 14:36:08 CET 2011
40 - backport to windows process emulation code again.
41
425.37 Sat Feb 19 07:49:44 CET 2011
43 - add a big "Coro thread life cycle" section to "man Coro".
44 - try a tentative workaround against the breakage that 5.13 has
45 introduced without depreciation period. sigh.
46 - no longer use Time::HiRes if gettimeofday is available, which
47 saves quite a lot of memory.
48
495.36 Sun Feb 13 05:33:41 CET 2011
50 - automatically load Coro::Channel, Coro::RWLock, Coro::Semaphore,
51 Coro::SemaphoreSet, Coro::Signal and Coro::Specific on first "new"
52 method call.
53 - undocument Coro::Timer::sleep and obsolete whole module.
54 - optimise Coro::Timer::timeout memory and cpu usage.
55 - slightly updated Coro::Intro for recent changes.
56 - do not initialise PL_dirty anymore.
57
585.25 Thu Nov 11 01:08:39 CET 2010
59 - try a different approach on netbsd - netbsd 5 finally has marginally
60 working pthreads, but still broken ucontext/sigaltstack.
61 - openbsd 4.8 finally got their act together, Coro works out of the box
62 with asm, setjmp and pthreads (no change, just informational).
63
645.24 Sat Oct 23 11:27:12 CEST 2010
65 - port to the EV 4.0 API.
66 - work around bugs in mingw32, making strawberry perl work
67 out of the box.
68 - correctly modify Coro::AIO function prototypes
69 so that they reflect the "no optional parameters" rule.
70 - "ported" libcoro to C++.
71
725.23 Mon May 17 18:50:42 CEST 2010
73 - be more resistant to ordering changes when initialising
74 Coro::AnyEvent, Coro::EV and Coro::Event (reported by Matthias
75 Waldorf).
76 - document that perl 5.12 deliberately removed support for cloning.
77
785.22 Wed Apr 14 03:55:35 CEST 2010
79 - correctly return udnef on errors in Coro::Handle::read/write
80 (testcase by Marc Mims).
81 - convert Coro::Util into a "perl compatibility wrapper" - the functions
82 are less useful now, but are drop-in replacements for existing
83 functions, listing better alternatives in the documentation. This also
84 fixes a bug in Coro::LWP which naively substituted Socket::inet_aton
85 with Coro::Util::inet_aton.
86 - do not override $Coro::idle unconditionally in Coro.pm, as other
87 modules could have provided their own idle coro already
88 (for exmaple, Coro::AnyEvent).
89 - fix Coro::Util::gethost* functions.
90 - Coro::Timer corretcly exports it's symbols (reported by Hideki Yamamura).
91
925.21 Wed Dec 16 07:19:51 CET 2009
93 - automatically load Coro::AnyEvent when AnyEvent and Coro are used
94 together.
95 - add some examples on how to combine other event loops with Coro in
96 Coro::AnyEvent, and how to run it (and not to block). Seems to be
97 the most common source of confusion.
98 - try to catch people naively blocking in an event callback.
99 - work around the perl filehandle bug issue in conjunction with
100 older common::sense (as indirectly pointed out by ZSystem).
101 - clarify the "not from signal handlers" section.
102
1035.2 Sun Oct 4 14:54:24 CEST 2009
104 - Coro::Storable destroyed the prototypes of the functions it wrapped.
105 - export rouse_cb and rouse_wait by default now.
106 - fix various prototype mismatches in Coro::AnyEvent and Coro::Handle.
107 - new method $state->swap_sv.
108 - added section on "windows process emulation" to the manpage,
109 after a not-so-fruitful (nor-friendly) "discussion" with chip
110 salzenberg (discussion implies arguments, but his only arguments
111 were ad-hominems, one wonders why he started it in the first
112 place). I hope this explains it well enough for him to understand,
113 and maybe well enough for others to understand.
114 - use common::sense everywhere now.
115 - idle callbacks are no longer supported, use idle coros instead.
116 - print a thread listing when a deadlock is detected.
117
1185.17 Sat Aug 22 23:09:31 CEST 2009
119 - work around a bug in the perl debugger causing crashes
120 when running under the debugger by marking _pool_handler as nodebug.
121 - speed up Coro::async considerably.
122 - try some hacks to get netbsd to work "more often" - their broken
123 setjmp/longjmp, ucontext *and* phtreads are really hard on Coro.
124 - convert Coro to AE 5.0 API.
125
1265.162 Tue Jul 28 04:04:03 CEST 2009
127 - perl 5.8.2 is now minimum requirement.
128 - skip t/19_handle.t on broken windows perls.
129
1305.161 Wed Jul 22 04:47:38 CEST 2009
131 - Coro::AnyEvent::poll could have a different prototype when EV was
132 used as backend (analyzed by Tatsuhiko Miyagawa).
133 - Coro::AnyEvent errornously initialised the event loop when loaded,
134 not on demand.
135 - try to workaround rare */t/01_unblock.t failures.
136
1375.16 Tue Jul 21 01:44:37 CEST 2009
138 - Coro::AnyEvent failed to hook into the event loop
139 when no threads had been readied between detecting
140 the event loop and actually running it.
141 - considerably speed up Coro::Select by taking avdantage
142 of AnyEvent > 4.8 and some other optimisations.
143 - implement paragraph readline mode in Coro::Handle
144 (based on patches by Zsbán Ambrus).
145 - replace WSAEINPROGRESS by WSAEWOULDBLOCK (reported
146 and analyzed by Yasuhiro MATSUMOTO).
147 - clarified libcoro license and copyright.
148 - someone stole my EXTRA_META!!!
149 - implement Coro::Select::patch_pp_sselect and it's brother,
150 for hardcode select overriding.
151
1525.151 Mon Jul 6 05:41:57 CEST 2009
153 - backport to windows process emulation code again (patch by
154 Yasuhiro MATSUMOTO).
155 - slightly update Coro::MakeMaker.
156
1575.15 Tue Jun 30 10:28:06 CEST 2009
158 - deprecate Coro::Socket, document how to get ipv6 support via
159 AnyEvent::Socket instead.
160 - implement signal->wait ($cb) interface, similar to semaphores.
161 - work around SvOK not supporting getmagic, so we have to getmagic
162 to test for undef :/ (reported by Matthias Waldorf).
163 - load Coro::AnyEvent in all modules using AnyEvent.
164 - work around perl corrupting our internal data structures,
165 reported by Tokuhiro Matsuno.
166 - enable per-coroutine real and cpu time gathering
167 (Coro::State::enable_times).
168
1695.14 Wed Jun 24 01:37:48 CEST 2009
170 - provide explicit functions to "cede" to the event loop in Coro::AnyEvent,
171 as this seems to have been a difficult concept (poll, sleep, idle,
172 idle_upto).
173 - add Coro::AnyEvent::readable/writable functions.
174 - clarify Coro::EV/Event/AnyEvent manpages.
175 - free per-thread global scalars in the thread calling ->cancel, to
176 avoid crashes when $_, $@ etc., are magical but some of those
177 had already been freed.
178 - "unexperimentalise" the callback interface for Coro::Semaphore.
179 - speed up ready queue management/context switching by using a linked
180 list instead of an array (~5-10%).
181 - implement "watch" command in Coro::Debug shells.
182 - for fun, implement time-slicing as an example in the manpage.
183 - if AnyEvent detects EV or Event, but we don't have Coro::EV or
184 Coro::Event, use the normal AnyEvent handling instead of dieing
185 (the same is true for Coro::Handle).
186 - properly document Coro::EV::timed_io_once.
187 - avoid unneccessary ->cancel calls in Coro::Handle.
188 - maybe make it work on mingw32 with win32 backend
189 (based on patch by Yasuhiro Matsumoto).
190
1915.132 Fri May 29 09:00:39 CEST 2009
192 - do not keep a reference to the argument itself in
193 Coro::Semaphore::guard, as it could change later.
194 - support SO_RCVBUF/SO_SNDBUF nonstandard Coro::Socket options,
195 should support a prepare callback.
196
1975.131 Mon Mar 16 23:20:37 CET 2009
198 - implement and document Coro->suspend, Coro->resume.
199 - fix Coro::Select implementation to not (often) close
200 the passed file descriptors (testcase provided by pippijn).
201
2025.13 Mon Dec 15 21:51:42 CET 2008
203 - EXPERIMENTAL: implement dynamic winds (on_enter/on_leave).
204 - don't set diehook to C<undef> but instead to NULL, to avoid
205 spurious warnings.
206 - fix a lot of bugs in Coro::SemaphoreSet.
207 - Coro::SemaphoreSet will less often create a semaphore needlessly.
208 - add Coro::SemaphoreSet::count and wait methods.
209 - take advantage of the new Guard module.
210 - deprecate Coro::guard.
211 - try to fix the dreaded 01_unblock tests once more. I hate it when
212 testsuites need more fixing than the code they are supposed to test.
213 - croak in more cases when a required callback isn't resolvable.
214 - fix some minor issues in Coro::State->call/eval.
215 - use current coroutine context instead of a temporary one
216 when temporarily switching to another coroutine.
217 - do not call C-level on_destroy handlers during global destruction,
218 to avoid needless segfaults.
219
2205.12 Sun Dec 7 13:30:38 CET 2008
221 - add default config for MirOS, which seems to be a bug-to-bug
222 compatible fork of openbsd ("world domination by releasing
223 openbsd cvs before the openbsd folks do it" or so :).
224 - free_padlist did destroy the names pad, not good, but didn't
225 seem to bother perl - this could fix issues such as eval ""
226 inside a function called from multiple coroutines.
227 - use a different method to detect destruction time.
228 - be more careful when freeing padlists just before global
229 destruction.
230 - fixed and expanded the call/cc example.
231 - renamed _terminate to _coro_run.
232 - new method Coro::Channel->shutdown.
233 - try pthreads on openbsd <4.4 (broken sigaltstack, will
234 pthreads fare better?).
235 - be less picky about destroying "a" running coroutine.
236
2375.11 Tue Nov 25 21:49:05 CET 2008
238 - DEBUGGING in 5.10.0 is a mess: it constantly flags perfectly
239 working code with failed assertions, introducing more bugs than
240 it fixes, requiring elaborate workarounds :(
241
2425.1 Mon Nov 24 08:54:59 CET 2008
243 - wrote a small introductory tutorial - Coro::Intro.
244 - convert Coro::Timer, Coro::Select and Coro::Util to rouse API.
245 - Coro::Select did errornously dup the file descriptors
246 and didn't work with all AnyEvent backends.
247 - Coro::Select wasn't imported correctly form Coro::LWP, causing blocking
248 LWP data transfers.
249 - disassociate c contexts from coro objects - this is agruably more
250 correct, but mostly allows sharing of cctxs between coro and state
251 objects, for added memory savings and speed increases.
252 - bumped $Coro::POOL_RSS up to 32kb by default.
253 - no longer set the optype to OP_CUSTOM, as B::* understandably
254 doesn't like this very much (and we *are* a type of entersub).
255 - implement state cloning, just to prove that call/cc can be done.
256 - automatically load Coro::AnyEvent in Coro::Handle.
257 - wrap ->cancel calls in eval inside Coro::Handle as EV watchers
258 do not have this method (and don't need it either).
259 - speed up generic anyevent methods in Coro::Handle by using rouse
260 callbacks.
261 - allow coroutines in $Coro::IDLE, speeding up Coro::AnyEvent and
262 others. It also makes the debugger happier, as you can trace
263 through the idle threads now.
264 - add comppad_name* and hints ($^H, %^H) to per-thread variables.
265 - eg/event was pretty broken.
266 - better 5.8.6 compatibility.
267
2685.0 Thu Nov 20 10:35:05 CET 2008
269 - NEW ARCHITECTURE: use the latest 4.x version if you experience
270 stability issues.
271 - bump API version to 7 - all dependents must be recompiled.
272 - removed timed_* functions - they were not being used anyways
273 and should be replaced by a more generic mechanism -
274 and were annoying to support anyways :)
275 - removed SemaphoreSet's waiter method - use sem method instead.
276 - Coro::Semaphore->adjust didn't correctly wake up enough waiters.
277 - async_pool did free a scalar value twice
278 ("Attempt to unreference...").
279 - fix a longstanding bug where calling terminate on a coro that
280 was waiting for a semaphore that was just becoming available
281 would cause a deadlock (semaphore would get into a state where
282 it was available but waiters were still blocked).
283 - calling throw on a coroutine that is waiting for a semaphore will
284 no longer make it acquire the semaphore (and thus leak a count).
285 - perl's process emulation is now not even theoretically supported
286 anymore.
287 - new functions Coro::rouse_cb and Coro::rouse_wait for easier
288 conversion of callback-style to blocking-style.
289 - new methods $coro->schedule_to and ->cede_to, to specifically
290 schedule or cede to a specific coroutine.
291 - new function Coro::Semaphore::wait.
292 - use named constants in Coro::Channel (Richard Hundt).
293 - directly patch the entersub opcode calling SLF functions (cede,
294 transfer and so on). this does speed up context switching, but
295 more importanly, it frees us from the hardcoded behaviour of
296 entersub, so we might actually be able to return something from
297 those functions and atcually create new ones.
298 - take advantage of __builtin_frame_address on gcc.
299 - expose THX in coroapi (not sure whether this was a wise decision,
300 as "threaded" perls are running at half speed anyways).
301 - implement execute_slf (schedule-like-function) interface that makes
302 it possible to implement schedule-like-functions in XS.
303 - use new SLF interface to massively speed up Coro::EV by roughly a
304 factor of two.
305 - used new SLF interface to massively speed up Coro::Semaphore by a
306 factor of three.
307 - used new SLF interface to speed up Coro::AIO by roughly a factor of
308 four and reduce its memory usage considerably.
309 - implement Coro::SemaphoreSet purely in terms of Coro::Semaphore,
310 for a nice speedup and vastly more correct behaviour. Also implement
311 a new method "sem" to get at the underlying semaphore object.
312 - implement Coro::Channel in terms of Coro::Semaphore, for a moderate
313 (in comparison) 20-40% speedup.
314 - used new SLF interface to reimplement Coro::Signal gaining
315 some unknown (because I was too lazy), but certain, speedup, and also
316 making signals reliable for the first time.
317 - used new SLF interface and other optimisations to speed up async_pool
318 by a factor of two. It also doesn't rely on perl's exception mechanism
319 to exit anymore. The overhead for terminating an async_pool, coro over
320 a normal async is now very small.
321 - sped up coroutine creation/destruction by 40%.
322 - forgot to include Coro/libcoro/README in the dist for all these years.
323 - work around a freebsd pthreads bug (manual testcancel is required as
324 pthread_cond_wait isn't a cancellation point on freebsd).
325 - use new rouse functions to speed up and simplify Coro::BDB.
326 - make "prefer perl native functions" work with threaded perls.
327 - condense Coro::Debug ps output, hint at v and w flags.
328 - (libcoro) lots of minor cleanups and portability improvements.
329
3304.914 Wed Nov 19 12:54:18 CET 2008
331 - fix a disastrous bug in the readline optimisation
332 introduced in 4.801.
333
3344.913 Sat Nov 15 07:58:28 CET 2008
335 - async_pool did free a scalar value twice
336 ("Attempt to unreference...").
337
3384.912 Thu Nov 13 18:31:23 CET 2008
339 - minor cleanups.
340 - use much larger stacks on linux and perl < 5.8.8.
341 - Coro::Debug::new_unix_server did not unlink the socket
342 when destroyed.
343
3444.911 Tue Nov 11 04:26:17 CET 2008
345 - "port" to threaded perls.
346
3474.91 Mon Nov 10 05:36:38 CET 2008
348 - the ->throw exception object no longer leaks.
349 - creating a new cctx leaked a scopestack entry (memleak).
350 - new coroutines didn't get created with a zero flags field
351 (unknown impact).
352 - calling ->throw on a not-yet-started coroutine should now work
353 instead of being ignored.
354 - ->throw is now supported on Coro::State objects.
355 - clean up cctx creation code a bit.
356 - entersub is actually a UNOP, not a LOGOP (not a bugfix).
357
3584.9 Sat Nov 8 17:45:27 CET 2008
359 - (libcoro) did not preserve rbp with CORO_ASM (we are getting there).
360 - (libcoro) no longer leak threads in the experimental pthread backend,
361 also speed it up considerably.
362 - (libcoro) do not rely on makecontext passing void *'s unscathed.
363 - fix compiletime dependencies on libcoro in the Makefile.
364 - cctx_count wasn't always updated properly.
365 - Coro::State::cctx_stacksize wasn't applied correctly.
366 - new function Coro::State::cctx_max_idle.
367 - the default max number of idle C contexts is now 4.
368 - (libcoro) try harder to get _setjmp/_longjmp.
369 - (libcoro) cleanup and extend the libcoro API to officially
370 allow the creation of empty source contexts.
371 - very experimental workaround for the totally broken netbsd platform.
372 - tried to hack around openbsd bugs.
373
3744.804 Wed Nov 5 16:36:00 CET 2008
375 - Coro::Debug::new_unix_server would not create a non-blocking listening
376 socket, sometimes causing freezes.
377 - (libcoro) fix misaligned stack points for setjmp and assembly
378 methods, which can cause crashes on x86/x86_64 with a sufficiently
379 aggressive compiler.
380 - new function: Coro::Debug::new_tcp_server.
381 - move ->throw into the Coro class because it only works on coro objects.
382
3834.803 Mon Nov 3 17:16:12 CET 2008
384 - (libcoro) use a global asm statement to become independent of gcc
385 otpimisations for CORO_ASM (thanks to pippijn for the idea).
386 - try to workaround yet another broken bsd, this time dragonfly.
387
3884.802 Thu Oct 30 10:56:12 CET 2008
389 - support -fno-omit-frame-pointer on x86 with the assembly method.
390 - tune 01_unblock.t tests a bit.
391
3924.801 Wed Oct 22 18:33:37 CEST 2008
393 - improve readline speed for very long "lines".
394 - backport to 5.8.8.
395
3964.8 Thu Oct 2 13:34:40 CEST 2008
397 - new function Coro::AIO::aio_wait.
398 - Coro.:AIO and Coro::BDB now "use Coro::AnyEvent".
399 - greatly speed up and reduce memory usage of Coro::AIO requests.
400 - implement some other µ-optimisations.
401
4024.749 Mon Sep 29 14:40:12 CEST 2008
403 - port to slow and broken pseudo-threaded perls. (courtesy pippijn).
404
4054.748 Sat Sep 27 14:03:19 CEST 2008
406 - implement, but do not document, PerlIO::cede(granularity).
407 - Coro::Storable forgot to wrap Storable::pstore.
408 - work around the multitude of leaks and memory corruption
409 bugs in PerlIO::via by using our own C-level perliol. As a side
410 effect, Coro::Storable is now much, much, much faster.
411
4124.747 Tue Sep 23 01:59:41 CEST 2008
413 - fix a per-cv memleak (one empty array was leaked per
414 code reference).
415 - avoid a crash in coro->call|rss when the coroutine was already
416 destroyed (most noticably when using Coro::Debug::ps :)
417 - also protect *Storable::FILE.
418 - push up default storable granularity to 20ms.
419
4204.746 Sun Sep 21 03:22:20 CEST 2008
421 - be more insistent on locking Storable against reentrancy
422 in Coro::Storable.
423 - move swap_def?v and throw to Coro::State, as documented.
424
4254.745 Thu Jul 24 00:14:38 CEST 2008
426 - remove debugging code related to MgPV_nolen_const, also try to
427 make it compile with perl 5.8.6 (yes, apple apparently loves
428 outdated software). Reported by John S.
429
4304.744 Tue Jul 8 22:06:35 CEST 2008
431 - correctly provide default DIE/WARN handlers as documented.
432 - also overwrite PL_vtbl_sigelem.svt_clear, even though current
433 implementations inside perl work fine for us.
434
4354.743 Mon Jun 16 00:21:57 CEST 2008
436 - when using Coro::EV without running EV::loop it could
437 result in busy-waiting for events, this has been fixed.
438 - reduce codesize and improve performance by using EV_DEFAULT_UC.
439
4404.742 Sat May 31 14:10:21 CEST 2008
441 - implement a workaround for (some) perl <5.8.8 versions.
442 - require EV 3.3+.
443
4444.741 Fri May 30 23:33:09 CEST 2008
445 - tell netbsd how utterly broken their imitation of an OS is
446 and refuse to build by default if pthreads are in use.
447 - switch to "s" method on all bsd's by default, as their ucontext
448 stuff seems just too broken.
449 - fix a bug in Coro::Select.
450
4514.74 Thu May 29 20:05:31 CEST 2008
452 - do not test Coro::LWP for lack of dependencies.
453 - do not test Coro::Select for lack of working perls.
454
4554.73 Thu May 29 2008
456 - fix a bug in Coro::EV which would cause it to block despite
457 there being runnable coroutines.
458 - sprinkle "no warnings" freely over everything, also suppress
459 warnings for some other modules.
460 - fix typo in WSAEWOULDBLOCK.
461
4624.72 Sun May 25 05:14:36 CEST 2008
463 - tweak META.yaml a bit, unfortunately, there is no documented way
464 to have optional dependencies with CPAN. doh :(
465 - avoid running some tests on windows because they would fail due to
466 perl bug (broken fork, broken pipes...).
467 - work around perl on windows bugs where perl returns undocumented
468 error codes for sysread, syswrite etc. by taking advantage
469 of AnyEvent's workaround for that problem.
470 - use AnyEvent::Util::fh_nonblocking in Coro::Handle to work around
471 a common perl implementation bug on windows.
472 - use unix domain sockets in testsuite to work around
473 common perl implementation bugs on widows (they are emulated by
474 tcp sockets on windows. ugh.)
475
4764.71 Sat May 24 20:01:27 CEST 2008
477 - fix a bug in Coro::AnyEvent ("Usage: Coro::AnyEvent::_schedule()").
478 - take advantage of async name resolution of AnyEvent::Util.
479 - work around brutal inet_aton override in Coro::LWP.
480 - take advantage of the readyhook in Coro::EV, for smoother
481 scheduling.
482
4834.7 Sun May 11 00:32:19 CEST 2008
484 - completely reworked the Coro manpage.
485 - added Coro::AnyEvent, generic event loop integration.
486 - implement cancel, ready and kill commands in Coro::Debug.
487 - document find_coro in Coro::Debug.
488 - incompatible change: rename has_stack to has_cctx.
489 - Coro::AIO and Coro::BDB no longer force event model detection,
490 use AnyEvent::AIO and AnyEvent::BDB.
491
4924.6 Sat Apr 26 10:05:14 CEST 2008
493 - INCOMPATIBLE CHANGE: sub/code attributes are no longer supported
494 by the Coro module. It was a mistake to have it in the first place.
495 - (experimental) support for activestate perl 5.10 (method "w").
496 - (experimental) support for strawberry perl (method "a").
497 - coro_sigelem_set did not return a value although it had to,
498 actual impact unknown.
499
5004.51 Mon Apr 14 13:28:27 CEST 2008
501 - make it compile again on 5.8.
502
5034.50 Thu Apr 10 09:43:17 CEST 2008
504 - I did it twice! (see 4.49).
505
5064.49 Sun Apr 6 21:23:31 CEST 2008
507 - grr, instead of compiling the recent changes on 5.10 only they
508 were compiled on 5.8 only.
509
5104.48 Sun Apr 6 20:36:46 CEST 2008
511 - allow coroutine switches during eval's under 5.10.x, as it seems
512 the parser is a per-interpreter option now, so this is safe
513 (this might fix the odd crashing bug).
514 - drop support for 5.9.x versions: they are dead, jim.
515
5164.47 Sun Apr 6 00:37:52 CEST 2008
517 - force cctx allocation on API calls: we know we need to force one
518 and gcc actually manages to confuse our heuristic nowadays,
519 leading to crashes and worse.
520 - document force_cctx.
521
5224.46 Fri Apr 4 22:05:43 CEST 2008
523 - upgrade libcoro, resulting in pthread-backend (which was only created
524 to fulfill the rules of the programming languages shootout).
525
5264.45 Thu Mar 13 11:55:36 CET 2008
527 - fix a file leaking bug in eg/mhyttpd that would allow
528 downloading of any file (reported by oesi).
529 - fix deadlock bug in Coro::Channel (reported by Richard Hundt)
530 (also add testcase).
531 - support Broadcast option in Coro::Socket (patch by Richard Hundt,
532 apparently having loads of fun with that).
533 - implement $state->swap_defsv, ->swap_defav and document ->throw.
534
5354.4 Wed Feb 13 16:44:29 CET 2008
536 - only assume .cfi pseudo isns exist on GNU/Linux.
537 - add get_timed method to Coro::Channels.
538 - fixed Coro::Debug manpage.
539 - perl 5.11 compatbility improvement based on patch by
540 Andreas König.
541
5424.37 Sun Jan 20 11:25:23 CET 2008
543 - soften the check for an active parser for perl 5.10 (avoiding
544 "Coro::State::transfer called while parsing" in many cases).
545
5464.36 Sun Jan 13 10:53:56 CET 2008
547 - reset diehook when terminating from an async_pool as to not
548 trigger any __DIE__ handlers.
549
5504.35 Sun Jan 13 04:14:13 CET 2008
551 - "bt" debug command now displays any exceptions
552 from longmess and also skips the innermost
553 stackframes, giving more useufl output.
554 - allow backtraces in coroutines blocked in Coro::EV,
555 at a <1% speed hit.
556 - handle localising of $SIG{__WARN__} and __DIE__
557 properly (with a proper amount of dirty hacking).
558
5594.34 Sat Dec 22 17:49:53 CET 2007
560 - upgrade to EV version 2.0 API.
561
5624.33 Mon Dec 17 08:36:12 CET 2007
563 - make Coro::AIO etc. loadable in the absence of EV.
564
5654.32 Mon Dec 17 07:46:02 CET 2007
566 - majorly improved Coro::Handle's performance with Coro::EV.
567 - implemented the until now mythical Coro::BDB module.
568 - specialcase EV in Coro::AIO and Coro::BDB for extra speed.
569
5704.31 Wed Dec 5 12:32:39 CET 2007
571 - remove warn statement form Coro::Util.
572
5734.3 Tue Dec 4 20:33:14 CET 2007
574 - calls to the idle function could cause stack corruption
575 when the stack changed.
576 - do no longer rely on the presence of EV::DNS if EV is used
577 (because it is gone), but instead take avdantage of EV::ADNS
578 if available.
579 - add ($) prototypes to all functions in Coro::Storable.
580 - use a conventional (safer) idle callback in Coro::EV.
581 - do accept quickly in Coro::Debug to avoid endless loops.
582
5834.22 Fri Nov 30 16:04:04 CET 2007
584 - really use optimised versions for Event and EV in Coro::Util
585 and Coro::Handle.
586
5874.21 Sun Nov 25 10:48:59 CET 2007
588 - fix a spurious memory read.
589 - Coro::EV no longer keeps the eventloop "alive".
590
5914.2 Fri Nov 9 20:47:05 CET 2007
592 - enable/disable tracing from a new coroutine, not a pooled one.
593 - fix a memleak in Coro::Event.
594 - removed killall call from fork_eval.
595 - made sure store_fd is already loaded so that fork_eval
596 does not have to parse autoload in each subprocess.
597 - only use assembly method if -O switch is in $Config{optimize}.
598 - add (optional) Coro::EV module, so far the best event loop module
599 directly supported by Coro.
600 - if the event model is EV, use EV::DNS to resolve
601 stuff in Coro::Util.
602 - don't get confused by multiple event notifications in Coro::Handle.
603 - initial support for EV (libevent interface).
604 - require Event and EV using configure_requires, to force their existance.
605
6064.13 Wed Oct 24 07:26:45 CEST 2007
607 - add Coro::Storable::blocking_thaw.
608 - use a vastly more complicated technique to localise
609 $SIG{__WARN/DIE__} that also works on perls <= 5.8.8.
610 - use a coroutine for the idle callback Coro::Event,
611 instead of running Event in the current coroutine context.
612 This also catches recursive invocations.
613 - actually report fork errors in gethostbyname and inet_aton.
614
6154.11 Thu Oct 11 02:40:24 CEST 2007
616 - port to threaded perls.
617
6184.1 Thu Oct 11 02:38:16 CEST 2007
619 - incompatible change: $SIG{__DIE__} and $SIG{__WARN__} will now
620 be local to each coro (see Coro::State).
621 - incompatible change: for very deep reasons, cede and cede_notself
622 cannot return anything, so nothing will be returned.
623 - possibly bring back 5.10 compatibility (untested).
624 - work around stupid (and wrong) warning on 5.10 :(.
625 - overlay the saved state over the context stack. This saves
626 a few hundred bytes per coroutine on average and also
627 speeds up context switching a bit.
628 - further tune default stack sizes.
629 - (more) correctly calculate stack usage in coro_rss.
630 - Coro::Storable::blocking_* did not properly lock
631 resulting in races between coroutines.
632 - added Coro::Storable::guard.
633 - stopping to trace a coroutine could destroy the cctx of
634 an unrelated coroutine.
635 - explain the relationship between Perl and C coroutines in
636 more detail in Coro::State.
637 - Coro::Util::inet_aton did not short-circuit dotted quad forms,
638 causing a fork per resolve. This also affected Coro::Socket.
639 - switch to a separate stack in $coro->call/eval to avoid
640 invalidating pointers.
641
6424.03 Sat Oct 6 21:24:00 CEST 2007
643 - added Coro::throw method.
644 - minor code cleanups.
645
6464.02 Sat Oct 6 02:36:47 CEST 2007
647 - fix a very minor per-coroutine memleak (a single codereference).
648 - fixed a bug where the currently in-use c context would be freed
649 prematurely (can happen only when programs change the stacksize
650 or use tracing).
651 - tracing can no longer keep a coro alive after it terminated.
652 - do static branch prediction in the common path for gcc. gives
653 about 2-5% speed improvement here.
654
6554.01 Fri Oct 5 22:10:49 CEST 2007
656 - instead of recreating *a* standard output handle we simply
657 use STDOUT, which is faster and hopefully more robust.
658
6594.0 Fri Oct 5 12:56:00 CEST 2007
660 - incompatibly changed Coro::Storable::freeze.
661 - major new feature: added Coro::Debug, for interactive coroutine
662 debugging, tracing and much more.
663 - major bug fix: unbelievable, but true: $_, $/ and many other
664 "saved" variables actually weren't being saved. This has been fixed,
665 of course, while increasing performance while losing all the save
666 flags.
667 - save flags are gone, and all the api functions dealing with them.
12 - added Coro::Semaphore::adjust. 668 - added Coro::Semaphore::adjust.
13 - added Coro::Util::fork_eval. 669 - added Coro::Util::fork_eval.
14 - incompatibly changed Coro::Storable::freeze.
15 - added Coro::Storable::{nfreeze,blocking_{freeze,nfreeze}}. 670 - added Coro::Storable::{nfreeze,blocking_{freeze,nfreeze}}.
671 - added Coro::killall.
16 - reduce initial stack sizes to allow for "micro-coroutines". 672 - reduce initial stack sizes to allow for "micro-coroutines".
673 - better async_pool resource management, moved parts of async_pool
674 handling to XS (major speed improvement).
675 - actually croak before modifying important data structures.
676 - refuse to transfer while compiling.
677 - possibly support eval EXPR better now.
678 - enable assembly per default on linux+bsd x86+amd64.
679 - all internal members were renamed _something for easier subclassing.
17 - many minor tweaks. 680 - many minor tweaks.
18 681
193.63 Wed May 16 14:10:06 CEST 2007 6823.63 Wed May 16 14:10:06 CEST 2007
20 - implement handcoded assembly for x86/amd64 SVR ABI. 683 - implement handcoded assembly for x86/amd64 SVR ABI.
21 684

Diff Legend

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