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

Comparing Coro/Changes (file contents):
Revision 1.351 by root, Sun May 25 01:32:15 2008 UTC vs.
Revision 1.590 by root, Fri Apr 13 10:53:24 2012 UTC

1Revision history for Perl extension Coro. 1Revision history for Perl extension Coro.
2 2
3TODO: should explore PerlIO::via::CoroCede and PerlIO::via::CoroAIO. 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)). 6TODO: __GCC_HAVE_DWARF2_CFI_ASM
7 7
86.08 Fri Apr 13 12:05:47 CEST 2012
9 - be more aggressive about exiting like perl does - formerly,
10 exiting from the non-main thread would not execute END blocks.
11
126.07 Fri Nov 11 21:21:48 CET 2011
13 - work around a bug in PerlIO (setting $SIG{__WARN__} to a PVCV).
14 - update ecb.h.
15
166.06 Mon Aug 8 23:59:48 CEST 2011
17 - cygwin unfortunately patches the stack at runtime, so we use the pthreads
18 backend, which is an order of magnitude slower. unfortunately, cygwins
19 pthread implementation isn't very complete either, so allocate the stack
20 twice just to be sure.
21 (note: cygwin also enables mymalloc, which is NOT THREADSAFE ON WINDOWS,
22 in its ithreaded perl - best recompile cygwin and use the 'w'indows
23 backend for much better performance. also disable ithreads for
24 even better performance...).
25
266.05 Thu Aug 4 21:36:36 CEST 2011
27 - blush, condvar values would not be propagated from send to recv anymore
28 (reported by Chip Salzenberg).
29 - use exponential increase for the readline buffer length in
30 Coro::Handle. also reduce initial allocation to 1020 from 4096 bytes.
31
326.04 Wed Aug 3 17:15:45 CEST 2011
33 - use even more efficient and more compatible condvars for
34 compatibility to AnyEvent 6.x :)
35 - more inconsequential ecb.h updates.
36
376.03 Wed Aug 3 11:41:30 CEST 2011
38 - change how Coro patches AnyEvent condvars for compatibility to
39 AnyEvent 6.x.
40 - update ecb.h, to no longer include <pthread.h> in case WinNT.h
41 hasn't been included.
42
436.02 Wed Jul 13 04:35:19 CEST 2011
44 - "improve portability to Gentoo" - gentoo manages to put perl variables
45 in memory areas that are farther than 2gb apart, which the jit couldn't
46 handle and barfed. now it's just a bit slower on gentoo and similar
47 systems.
48
496.01 Sun Jul 3 12:31:14 CEST 2011
50 - workarounds are good, but the test for whether pthreads are used
51 was not good. thisone should be better.
52 - check differently whether gcc generates cfi instructions itself.
53
546.0 Wed Jun 29 19:43:35 CEST 2011
55 - INCOMPATIBLE CHANGE: unreferenced coro objects will now be
56 destroyed and cleaned up automatically (e.g. async { schedule }).
57 - implement a JIT compiler for part of the thread switch code,
58 which gives a 50% speed improvement on threaded perls, and
59 about 4% on non-thraeded perls (so threaded perls now finally
60 reach about half the speed of non-threaded perls).
61 - slightly modernise Coro::Intro, add section about rouse functions.
62 - avoid DEFSV and ERRSV, giving another 10% improvement
63 in thread switching.
64 - Coro::State->is_destroyed is now called is_zombie.
65 - implement a Coro->safe_cancel method that might fail, but
66 cancels in a "safer" way if it succeeds.
67 - add preliminary support for DEBUGGING perls.
68 - get rid of two hash-accesses when initialising a new Coro - this
69 speeds up coro creation by almost a factor of two.
70 - croak when a coro that is being cancelled tries to block
71 (e.g. while executing a guard block), instead of crashing or
72 deadlocking.
73 - use a more robust and also faster method to identify Coro::State
74 objects - speeds up everything a bit.
75 - implement Coro->cancel in XS for a 20% speed improvement, and to
76 be able to implement mutual cancellation.
77 - speed up context switches by a percent or two by more efficiently
78 allocating context stack entries.
79 - implement Coro->join and Coro->on_destroy in XS for a speedup and
80 a reduction in memory use.
81 - cancelling a coro while it itself is cancelling another coro is
82 now supported and working, instead of triggering an assertion.
83 - be a bit more crash-resistant when calling (buggy) on_destroy
84 callbacks (best effort).
85 - move on_destroy into the slf_frame, to allow extension slf
86 functions to have destructors.
87 - get rid if coro refcounting - simply crash in other interpreter
88 threads by nulling the pointers on clone.
89 - simplify warn/die hook handling when loading Coro - the convoluted
90 logic seems to be no longer neccessary.
91 - use libecb instead of our own home-grown gcc hacks.
92 - document alternatives to Coro::LWP. Please use them :)
93 - work around another mindless idiotic NEEDLESS bug in openbsd/mirbsds
94 sigaltstack. Really. wine suffers from it, erlang suffers from it,
95 and it's known since at least 2006.
96
975.372 Wed Feb 23 06:14:30 CET 2011
98 - apparently mingw doesn't provide a working gettimeofday, try to
99 work around that by relying on Time::HiRes (indirectly brought to
100 my attention by Max Maischein).
101 - fix some portability issues when Time::HiRes was used.
102
1035.371 Mon Feb 21 14:36:08 CET 2011
104 - backport to windows process emulation code again.
105
1065.37 Sat Feb 19 07:49:44 CET 2011
107 - add a big "Coro thread life cycle" section to "man Coro".
108 - try a tentative workaround against the breakage that 5.13 has
109 introduced without depreciation period. sigh.
110 - no longer use Time::HiRes if gettimeofday is available, which
111 saves quite a lot of memory.
112
1135.36 Sun Feb 13 05:33:41 CET 2011
114 - automatically load Coro::Channel, Coro::RWLock, Coro::Semaphore,
115 Coro::SemaphoreSet, Coro::Signal and Coro::Specific on first "new"
116 method call.
117 - undocument Coro::Timer::sleep and obsolete whole module.
118 - optimise Coro::Timer::timeout memory and cpu usage.
119 - slightly updated Coro::Intro for recent changes.
120 - do not initialise PL_dirty anymore.
121
1225.25 Thu Nov 11 01:08:39 CET 2010
123 - try a different approach on netbsd - netbsd 5 finally has marginally
124 working pthreads, but still broken ucontext/sigaltstack.
125 - openbsd 4.8 finally got their act together, Coro works out of the box
126 with asm, setjmp and pthreads (no change, just informational).
127
1285.24 Sat Oct 23 11:27:12 CEST 2010
129 - port to the EV 4.0 API.
130 - work around bugs in mingw32, making strawberry perl work
131 out of the box.
132 - correctly modify Coro::AIO function prototypes
133 so that they reflect the "no optional parameters" rule.
134 - "ported" libcoro to C++.
135
1365.23 Mon May 17 18:50:42 CEST 2010
137 - be more resistant to ordering changes when initialising
138 Coro::AnyEvent, Coro::EV and Coro::Event (reported by Matthias
139 Waldorf).
140 - document that perl 5.12 deliberately removed support for cloning.
141
1425.22 Wed Apr 14 03:55:35 CEST 2010
143 - correctly return udnef on errors in Coro::Handle::read/write
144 (testcase by Marc Mims).
145 - convert Coro::Util into a "perl compatibility wrapper" - the functions
146 are less useful now, but are drop-in replacements for existing
147 functions, listing better alternatives in the documentation. This also
148 fixes a bug in Coro::LWP which naively substituted Socket::inet_aton
149 with Coro::Util::inet_aton.
150 - do not override $Coro::idle unconditionally in Coro.pm, as other
151 modules could have provided their own idle coro already
152 (for exmaple, Coro::AnyEvent).
153 - fix Coro::Util::gethost* functions.
154 - Coro::Timer corretcly exports it's symbols (reported by Hideki Yamamura).
155
1565.21 Wed Dec 16 07:19:51 CET 2009
157 - automatically load Coro::AnyEvent when AnyEvent and Coro are used
158 together.
159 - add some examples on how to combine other event loops with Coro in
160 Coro::AnyEvent, and how to run it (and not to block). Seems to be
161 the most common source of confusion.
162 - try to catch people naively blocking in an event callback.
163 - work around the perl filehandle bug issue in conjunction with
164 older common::sense (as indirectly pointed out by ZSystem).
165 - clarify the "not from signal handlers" section.
166
1675.2 Sun Oct 4 14:54:24 CEST 2009
168 - Coro::Storable destroyed the prototypes of the functions it wrapped.
169 - export rouse_cb and rouse_wait by default now.
170 - fix various prototype mismatches in Coro::AnyEvent and Coro::Handle.
171 - new method $state->swap_sv.
172 - added section on "windows process emulation" to the manpage,
173 after a not-so-fruitful (nor-friendly) "discussion" with chip
174 salzenberg (discussion implies arguments, but his only arguments
175 were ad-hominems, one wonders why he started it in the first
176 place). I hope this explains it well enough for him to understand,
177 and maybe well enough for others to understand.
178 - use common::sense everywhere now.
179 - idle callbacks are no longer supported, use idle coros instead.
180 - print a thread listing when a deadlock is detected.
181
1825.17 Sat Aug 22 23:09:31 CEST 2009
183 - work around a bug in the perl debugger causing crashes
184 when running under the debugger by marking _pool_handler as nodebug.
185 - speed up Coro::async considerably.
186 - try some hacks to get netbsd to work "more often" - their broken
187 setjmp/longjmp, ucontext *and* phtreads are really hard on Coro.
188 - convert Coro to AE 5.0 API.
189
1905.162 Tue Jul 28 04:04:03 CEST 2009
191 - perl 5.8.2 is now minimum requirement.
192 - skip t/19_handle.t on broken windows perls.
193
1945.161 Wed Jul 22 04:47:38 CEST 2009
195 - Coro::AnyEvent::poll could have a different prototype when EV was
196 used as backend (analyzed by Tatsuhiko Miyagawa).
197 - Coro::AnyEvent errornously initialised the event loop when loaded,
198 not on demand.
199 - try to workaround rare */t/01_unblock.t failures.
200
2015.16 Tue Jul 21 01:44:37 CEST 2009
202 - Coro::AnyEvent failed to hook into the event loop
203 when no threads had been readied between detecting
204 the event loop and actually running it.
205 - considerably speed up Coro::Select by taking avdantage
206 of AnyEvent > 4.8 and some other optimisations.
207 - implement paragraph readline mode in Coro::Handle
208 (based on patches by Zsbán Ambrus).
209 - replace WSAEINPROGRESS by WSAEWOULDBLOCK (reported
210 and analyzed by Yasuhiro MATSUMOTO).
211 - clarified libcoro license and copyright.
212 - someone stole my EXTRA_META!!!
213 - implement Coro::Select::patch_pp_sselect and it's brother,
214 for hardcode select overriding.
215
2165.151 Mon Jul 6 05:41:57 CEST 2009
217 - backport to windows process emulation code again (patch by
218 Yasuhiro MATSUMOTO).
219 - slightly update Coro::MakeMaker.
220
2215.15 Tue Jun 30 10:28:06 CEST 2009
222 - deprecate Coro::Socket, document how to get ipv6 support via
223 AnyEvent::Socket instead.
224 - implement signal->wait ($cb) interface, similar to semaphores.
225 - work around SvOK not supporting getmagic, so we have to getmagic
226 to test for undef :/ (reported by Matthias Waldorf).
227 - load Coro::AnyEvent in all modules using AnyEvent.
228 - work around perl corrupting our internal data structures,
229 reported by Tokuhiro Matsuno.
230 - enable per-coroutine real and cpu time gathering
231 (Coro::State::enable_times).
232
2335.14 Wed Jun 24 01:37:48 CEST 2009
234 - provide explicit functions to "cede" to the event loop in Coro::AnyEvent,
235 as this seems to have been a difficult concept (poll, sleep, idle,
236 idle_upto).
237 - add Coro::AnyEvent::readable/writable functions.
238 - clarify Coro::EV/Event/AnyEvent manpages.
239 - free per-thread global scalars in the thread calling ->cancel, to
240 avoid crashes when $_, $@ etc., are magical but some of those
241 had already been freed.
242 - "unexperimentalise" the callback interface for Coro::Semaphore.
243 - speed up ready queue management/context switching by using a linked
244 list instead of an array (~5-10%).
245 - implement "watch" command in Coro::Debug shells.
246 - for fun, implement time-slicing as an example in the manpage.
247 - if AnyEvent detects EV or Event, but we don't have Coro::EV or
248 Coro::Event, use the normal AnyEvent handling instead of dieing
249 (the same is true for Coro::Handle).
250 - properly document Coro::EV::timed_io_once.
251 - avoid unneccessary ->cancel calls in Coro::Handle.
252 - maybe make it work on mingw32 with win32 backend
253 (based on patch by Yasuhiro Matsumoto).
254
2555.132 Fri May 29 09:00:39 CEST 2009
256 - do not keep a reference to the argument itself in
257 Coro::Semaphore::guard, as it could change later.
258 - support SO_RCVBUF/SO_SNDBUF nonstandard Coro::Socket options,
259 should support a prepare callback.
260
2615.131 Mon Mar 16 23:20:37 CET 2009
262 - implement and document Coro->suspend, Coro->resume.
263 - fix Coro::Select implementation to not (often) close
264 the passed file descriptors (testcase provided by pippijn).
265
2665.13 Mon Dec 15 21:51:42 CET 2008
267 - EXPERIMENTAL: implement dynamic winds (on_enter/on_leave).
268 - don't set diehook to C<undef> but instead to NULL, to avoid
269 spurious warnings.
270 - fix a lot of bugs in Coro::SemaphoreSet.
271 - Coro::SemaphoreSet will less often create a semaphore needlessly.
272 - add Coro::SemaphoreSet::count and wait methods.
273 - take advantage of the new Guard module.
274 - deprecate Coro::guard.
275 - try to fix the dreaded 01_unblock tests once more. I hate it when
276 testsuites need more fixing than the code they are supposed to test.
277 - croak in more cases when a required callback isn't resolvable.
278 - fix some minor issues in Coro::State->call/eval.
279 - use current coroutine context instead of a temporary one
280 when temporarily switching to another coroutine.
281 - do not call C-level on_destroy handlers during global destruction,
282 to avoid needless segfaults.
283
2845.12 Sun Dec 7 13:30:38 CET 2008
285 - add default config for MirOS, which seems to be a bug-to-bug
286 compatible fork of openbsd ("world domination by releasing
287 openbsd cvs before the openbsd folks do it" or so :).
288 - free_padlist did destroy the names pad, not good, but didn't
289 seem to bother perl - this could fix issues such as eval ""
290 inside a function called from multiple coroutines.
291 - use a different method to detect destruction time.
292 - be more careful when freeing padlists just before global
293 destruction.
294 - fixed and expanded the call/cc example.
295 - renamed _terminate to _coro_run.
296 - new method Coro::Channel->shutdown.
297 - try pthreads on openbsd <4.4 (broken sigaltstack, will
298 pthreads fare better?).
299 - be less picky about destroying "a" running coroutine.
300
3015.11 Tue Nov 25 21:49:05 CET 2008
302 - DEBUGGING in 5.10.0 is a mess: it constantly flags perfectly
303 working code with failed assertions, introducing more bugs than
304 it fixes, requiring elaborate workarounds :(
305
3065.1 Mon Nov 24 08:54:59 CET 2008
307 - wrote a small introductory tutorial - Coro::Intro.
308 - convert Coro::Timer, Coro::Select and Coro::Util to rouse API.
309 - Coro::Select did errornously dup the file descriptors
310 and didn't work with all AnyEvent backends.
311 - Coro::Select wasn't imported correctly form Coro::LWP, causing blocking
312 LWP data transfers.
313 - disassociate c contexts from coro objects - this is agruably more
314 correct, but mostly allows sharing of cctxs between coro and state
315 objects, for added memory savings and speed increases.
316 - bumped $Coro::POOL_RSS up to 32kb by default.
317 - no longer set the optype to OP_CUSTOM, as B::* understandably
318 doesn't like this very much (and we *are* a type of entersub).
319 - implement state cloning, just to prove that call/cc can be done.
320 - automatically load Coro::AnyEvent in Coro::Handle.
321 - wrap ->cancel calls in eval inside Coro::Handle as EV watchers
322 do not have this method (and don't need it either).
323 - speed up generic anyevent methods in Coro::Handle by using rouse
324 callbacks.
325 - allow coroutines in $Coro::IDLE, speeding up Coro::AnyEvent and
326 others. It also makes the debugger happier, as you can trace
327 through the idle threads now.
328 - add comppad_name* and hints ($^H, %^H) to per-thread variables.
329 - eg/event was pretty broken.
330 - better 5.8.6 compatibility.
331
3325.0 Thu Nov 20 10:35:05 CET 2008
333 - NEW ARCHITECTURE: use the latest 4.x version if you experience
334 stability issues.
335 - bump API version to 7 - all dependents must be recompiled.
336 - removed timed_* functions - they were not being used anyways
337 and should be replaced by a more generic mechanism -
338 and were annoying to support anyways :)
339 - removed SemaphoreSet's waiter method - use sem method instead.
340 - Coro::Semaphore->adjust didn't correctly wake up enough waiters.
341 - async_pool did free a scalar value twice
342 ("Attempt to unreference...").
343 - fix a longstanding bug where calling terminate on a coro that
344 was waiting for a semaphore that was just becoming available
345 would cause a deadlock (semaphore would get into a state where
346 it was available but waiters were still blocked).
347 - calling throw on a coroutine that is waiting for a semaphore will
348 no longer make it acquire the semaphore (and thus leak a count).
349 - perl's process emulation is now not even theoretically supported
350 anymore.
351 - new functions Coro::rouse_cb and Coro::rouse_wait for easier
352 conversion of callback-style to blocking-style.
353 - new methods $coro->schedule_to and ->cede_to, to specifically
354 schedule or cede to a specific coroutine.
355 - new function Coro::Semaphore::wait.
356 - use named constants in Coro::Channel (Richard Hundt).
357 - directly patch the entersub opcode calling SLF functions (cede,
358 transfer and so on). this does speed up context switching, but
359 more importanly, it frees us from the hardcoded behaviour of
360 entersub, so we might actually be able to return something from
361 those functions and atcually create new ones.
362 - take advantage of __builtin_frame_address on gcc.
363 - expose THX in coroapi (not sure whether this was a wise decision,
364 as "threaded" perls are running at half speed anyways).
365 - implement execute_slf (schedule-like-function) interface that makes
366 it possible to implement schedule-like-functions in XS.
367 - use new SLF interface to massively speed up Coro::EV by roughly a
368 factor of two.
369 - used new SLF interface to massively speed up Coro::Semaphore by a
370 factor of three.
371 - used new SLF interface to speed up Coro::AIO by roughly a factor of
372 four and reduce its memory usage considerably.
373 - implement Coro::SemaphoreSet purely in terms of Coro::Semaphore,
374 for a nice speedup and vastly more correct behaviour. Also implement
375 a new method "sem" to get at the underlying semaphore object.
376 - implement Coro::Channel in terms of Coro::Semaphore, for a moderate
377 (in comparison) 20-40% speedup.
378 - used new SLF interface to reimplement Coro::Signal gaining
379 some unknown (because I was too lazy), but certain, speedup, and also
380 making signals reliable for the first time.
381 - used new SLF interface and other optimisations to speed up async_pool
382 by a factor of two. It also doesn't rely on perl's exception mechanism
383 to exit anymore. The overhead for terminating an async_pool, coro over
384 a normal async is now very small.
385 - sped up coroutine creation/destruction by 40%.
386 - forgot to include Coro/libcoro/README in the dist for all these years.
387 - work around a freebsd pthreads bug (manual testcancel is required as
388 pthread_cond_wait isn't a cancellation point on freebsd).
389 - use new rouse functions to speed up and simplify Coro::BDB.
390 - make "prefer perl native functions" work with threaded perls.
391 - condense Coro::Debug ps output, hint at v and w flags.
392 - (libcoro) lots of minor cleanups and portability improvements.
393
3944.914 Wed Nov 19 12:54:18 CET 2008
395 - fix a disastrous bug in the readline optimisation
396 introduced in 4.801.
397
3984.913 Sat Nov 15 07:58:28 CET 2008
399 - async_pool did free a scalar value twice
400 ("Attempt to unreference...").
401
4024.912 Thu Nov 13 18:31:23 CET 2008
403 - minor cleanups.
404 - use much larger stacks on linux and perl < 5.8.8.
405 - Coro::Debug::new_unix_server did not unlink the socket
406 when destroyed.
407
4084.911 Tue Nov 11 04:26:17 CET 2008
409 - "port" to threaded perls.
410
4114.91 Mon Nov 10 05:36:38 CET 2008
412 - the ->throw exception object no longer leaks.
413 - creating a new cctx leaked a scopestack entry (memleak).
414 - new coroutines didn't get created with a zero flags field
415 (unknown impact).
416 - calling ->throw on a not-yet-started coroutine should now work
417 instead of being ignored.
418 - ->throw is now supported on Coro::State objects.
419 - clean up cctx creation code a bit.
420 - entersub is actually a UNOP, not a LOGOP (not a bugfix).
421
4224.9 Sat Nov 8 17:45:27 CET 2008
423 - (libcoro) did not preserve rbp with CORO_ASM (we are getting there).
424 - (libcoro) no longer leak threads in the experimental pthread backend,
425 also speed it up considerably.
426 - (libcoro) do not rely on makecontext passing void *'s unscathed.
427 - fix compiletime dependencies on libcoro in the Makefile.
428 - cctx_count wasn't always updated properly.
429 - Coro::State::cctx_stacksize wasn't applied correctly.
430 - new function Coro::State::cctx_max_idle.
431 - the default max number of idle C contexts is now 4.
432 - (libcoro) try harder to get _setjmp/_longjmp.
433 - (libcoro) cleanup and extend the libcoro API to officially
434 allow the creation of empty source contexts.
435 - very experimental workaround for the totally broken netbsd platform.
436 - tried to hack around openbsd bugs.
437
4384.804 Wed Nov 5 16:36:00 CET 2008
439 - Coro::Debug::new_unix_server would not create a non-blocking listening
440 socket, sometimes causing freezes.
441 - (libcoro) fix misaligned stack points for setjmp and assembly
442 methods, which can cause crashes on x86/x86_64 with a sufficiently
443 aggressive compiler.
444 - new function: Coro::Debug::new_tcp_server.
445 - move ->throw into the Coro class because it only works on coro objects.
446
4474.803 Mon Nov 3 17:16:12 CET 2008
448 - (libcoro) use a global asm statement to become independent of gcc
449 otpimisations for CORO_ASM (thanks to pippijn for the idea).
450 - try to workaround yet another broken bsd, this time dragonfly.
451
4524.802 Thu Oct 30 10:56:12 CET 2008
453 - support -fno-omit-frame-pointer on x86 with the assembly method.
454 - tune 01_unblock.t tests a bit.
455
4564.801 Wed Oct 22 18:33:37 CEST 2008
457 - improve readline speed for very long "lines".
458 - backport to 5.8.8.
459
4604.8 Thu Oct 2 13:34:40 CEST 2008
461 - new function Coro::AIO::aio_wait.
462 - Coro.:AIO and Coro::BDB now "use Coro::AnyEvent".
463 - greatly speed up and reduce memory usage of Coro::AIO requests.
464 - implement some other µ-optimisations.
465
4664.749 Mon Sep 29 14:40:12 CEST 2008
467 - port to slow and broken pseudo-threaded perls. (courtesy pippijn).
468
4694.748 Sat Sep 27 14:03:19 CEST 2008
470 - implement, but do not document, PerlIO::cede(granularity).
471 - Coro::Storable forgot to wrap Storable::pstore.
472 - work around the multitude of leaks and memory corruption
473 bugs in PerlIO::via by using our own C-level perliol. As a side
474 effect, Coro::Storable is now much, much, much faster.
475
4764.747 Tue Sep 23 01:59:41 CEST 2008
477 - fix a per-cv memleak (one empty array was leaked per
478 code reference).
479 - avoid a crash in coro->call|rss when the coroutine was already
480 destroyed (most noticably when using Coro::Debug::ps :)
481 - also protect *Storable::FILE.
482 - push up default storable granularity to 20ms.
483
4844.746 Sun Sep 21 03:22:20 CEST 2008
485 - be more insistent on locking Storable against reentrancy
486 in Coro::Storable.
487 - move swap_def?v and throw to Coro::State, as documented.
488
4894.745 Thu Jul 24 00:14:38 CEST 2008
490 - remove debugging code related to MgPV_nolen_const, also try to
491 make it compile with perl 5.8.6 (yes, apple apparently loves
492 outdated software). Reported by John S.
493
4944.744 Tue Jul 8 22:06:35 CEST 2008
495 - correctly provide default DIE/WARN handlers as documented.
496 - also overwrite PL_vtbl_sigelem.svt_clear, even though current
497 implementations inside perl work fine for us.
498
4994.743 Mon Jun 16 00:21:57 CEST 2008
500 - when using Coro::EV without running EV::loop it could
501 result in busy-waiting for events, this has been fixed.
502 - reduce codesize and improve performance by using EV_DEFAULT_UC.
503
5044.742 Sat May 31 14:10:21 CEST 2008
505 - implement a workaround for (some) perl <5.8.8 versions.
506 - require EV 3.3+.
507
5084.741 Fri May 30 23:33:09 CEST 2008
509 - tell netbsd how utterly broken their imitation of an OS is
510 and refuse to build by default if pthreads are in use.
511 - switch to "s" method on all bsd's by default, as their ucontext
512 stuff seems just too broken.
513 - fix a bug in Coro::Select.
514
5154.74 Thu May 29 20:05:31 CEST 2008
516 - do not test Coro::LWP for lack of dependencies.
517 - do not test Coro::Select for lack of working perls.
518
5194.73 Thu May 29 2008
520 - fix a bug in Coro::EV which would cause it to block despite
521 there being runnable coroutines.
522 - sprinkle "no warnings" freely over everything, also suppress
523 warnings for some other modules.
524 - fix typo in WSAEWOULDBLOCK.
525
5264.72 Sun May 25 05:14:36 CEST 2008
8 - tweak META.yaml a bit, unfortunately, there is no documented way 527 - tweak META.yaml a bit, unfortunately, there is no documented way
9 to have optional dependencies with CPAN. doh :( 528 to have optional dependencies with CPAN. doh :(
10 - avoid running some tests on windows because they would fail due to 529 - avoid running some tests on windows because they would fail due to
11 perl bug (broken fork, broken pipes...). 530 perl bug (broken fork, broken pipes...).
12 - work around perl on windows bugs where perl returns undocumented 531 - work around perl on windows bugs where perl returns undocumented

Diff Legend

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