ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/Coro/Changes
Revision: 1.542
Committed: Sat Oct 23 09:28:48 2010 UTC (13 years, 6 months ago) by root
Branch: MAIN
CVS Tags: rel-5_24
Changes since 1.541: +4 -0 lines
Log Message:
5_24

File Contents

# Content
1 Revision history for Perl extension Coro.
2
3 TODO: should explore PerlIO::coroaio (perl leaks like hell).
4 TODO: unready_all
5 TODO: myhttpd header parsing
6 TODO: channel->maxsize(newsize)?
7 TODO: http://www.microsoft.com/msj/archive/s2ce.aspx
8
9 5.24 Sat Oct 23 11:27:12 CEST 2010
10 - port to the EV 4.0 API.
11 - work around bugs in mingw32, making strawberry perl work
12 out of the box.
13 - correctly modify Coro::AIO function prototypes
14 so that they reflect the "no optional parameters" rule.
15 - "ported" libcoro to C++.
16
17 5.23 Mon May 17 18:50:42 CEST 2010
18 - be more resistant to ordering changes when initialising
19 Coro::AnyEvent, Coro::EV and Coro::Event (reported by Matthias
20 Waldorf).
21 - document that perl 5.12 deliberately removed support for cloning.
22
23 5.22 Wed Apr 14 03:55:35 CEST 2010
24 - correctly return udnef on errors in Coro::Handle::read/write
25 (testcase by Marc Mims).
26 - convert Coro::Util into a "perl compatibility wrapper" - the functions
27 are less useful now, but are drop-in replacements for existing
28 functions, listing better alternatives in the documentation. This also
29 fixes a bug in Coro::LWP which naively substituted Socket::inet_aton
30 with Coro::Util::inet_aton.
31 - do not override $Coro::idle unconditionally in Coro.pm, as other
32 modules could have provided their own idle coro already
33 (for exmaple, Coro::AnyEvent).
34 - fix Coro::Util::gethost* functions.
35 - Coro::Timer corretcly exports it's symbols (reported by Hideki Yamamura).
36
37 5.21 Wed Dec 16 07:19:51 CET 2009
38 - automatically load Coro::AnyEvent when AnyEvent and Coro are used
39 together.
40 - add some examples on how to combine other event loops with Coro in
41 Coro::AnyEvent, and how to run it (and not to block). Seems to be
42 the most common source of confusion.
43 - try to catch people naively blocking in an event callback.
44 - work around the perl filehandle bug issue in conjunction with
45 older common::sense (as indirectly pointed out by ZSystem).
46 - clarify the "not from signal handlers" section.
47
48 5.2 Sun Oct 4 14:54:24 CEST 2009
49 - Coro::Storable destroyed the prototypes of the functions it wrapped.
50 - export rouse_cb and rouse_wait by default now.
51 - fix various prototype mismatches in Coro::AnyEvent and Coro::Handle.
52 - new method $state->swap_sv.
53 - added section on "windows process emulation" to the manpage,
54 after a not-so-fruitful (nor-friendly) "discussion" with chip
55 salzenberg (discussion implies arguments, but his only arguments
56 were ad-hominems, one wonders why he started it in the first
57 place). I hope this explains it well enough for him to understand,
58 and maybe well enough for others to understand.
59 - use common::sense everywhere now.
60 - idle callbacks are no longer supported, use idle coros instead.
61 - print a thread listing when a deadlock is detected.
62
63 5.17 Sat Aug 22 23:09:31 CEST 2009
64 - work around a bug in the perl debugger causing crashes
65 when running under the debugger by marking _pool_handler as nodebug.
66 - speed up Coro::async considerably.
67 - try some hacks to get netbsd to work "more often" - their broken
68 setjmp/longjmp, ucontext *and* phtreads are really hard on Coro.
69 - convert Coro to AE 5.0 API.
70
71 5.162 Tue Jul 28 04:04:03 CEST 2009
72 - perl 5.8.2 is now minimum requirement.
73 - skip t/19_handle.t on broken windows perls.
74
75 5.161 Wed Jul 22 04:47:38 CEST 2009
76 - Coro::AnyEvent::poll could have a different prototype when EV was
77 used as backend (analyzed by Tatsuhiko Miyagawa).
78 - Coro::AnyEvent errornously initialised the event loop when loaded,
79 not on demand.
80 - try to workaround rare */t/01_unblock.t failures.
81
82 5.16 Tue Jul 21 01:44:37 CEST 2009
83 - Coro::AnyEvent failed to hook into the event loop
84 when no threads had been readied between detecting
85 the event loop and actually running it.
86 - considerably speed up Coro::Select by taking avdantage
87 of AnyEvent > 4.8 and some other optimisations.
88 - implement paragraph readline mode in Coro::Handle
89 (based on patches by Zsbán Ambrus).
90 - replace WSAEINPROGRESS by WSAEWOULDBLOCK (reported
91 and analyzed by Yasuhiro MATSUMOTO).
92 - clarified libcoro license and copyright.
93 - someone stole my EXTRA_META!!!
94 - implement Coro::Select::patch_pp_sselect and it's brother,
95 for hardcode select overriding.
96
97 5.151 Mon Jul 6 05:41:57 CEST 2009
98 - backport to windows process emulation code again (patch by
99 Yasuhiro MATSUMOTO).
100 - slightly update Coro::MakeMaker.
101
102 5.15 Tue Jun 30 10:28:06 CEST 2009
103 - deprecate Coro::Socket, document how to get ipv6 support via
104 AnyEvent::Socket instead.
105 - implement signal->wait ($cb) interface, similar to semaphores.
106 - work around SvOK not supporting getmagic, so we have to getmagic
107 to test for undef :/ (reported by Matthias Waldorf).
108 - load Coro::AnyEvent in all modules using AnyEvent.
109 - work around perl corrupting our internal data structures,
110 reported by Tokuhiro Matsuno.
111 - enable per-coroutine real and cpu time gathering
112 (Coro::State::enable_times).
113
114 5.14 Wed Jun 24 01:37:48 CEST 2009
115 - provide explicit functions to "cede" to the event loop in Coro::AnyEvent,
116 as this seems to have been a difficult concept (poll, sleep, idle,
117 idle_upto).
118 - add Coro::AnyEvent::readable/writable functions.
119 - clarify Coro::EV/Event/AnyEvent manpages.
120 - free per-thread global scalars in the thread calling ->cancel, to
121 avoid crashes when $_, $@ etc., are magical but some of those
122 had already been freed.
123 - "unexperimentalise" the callback interface for Coro::Semaphore.
124 - speed up ready queue management/context switching by using a linked
125 list instead of an array (~5-10%).
126 - implement "watch" command in Coro::Debug shells.
127 - for fun, implement time-slicing as an example in the manpage.
128 - if AnyEvent detects EV or Event, but we don't have Coro::EV or
129 Coro::Event, use the normal AnyEvent handling instead of dieing
130 (the same is true for Coro::Handle).
131 - properly document Coro::EV::timed_io_once.
132 - avoid unneccessary ->cancel calls in Coro::Handle.
133 - maybe make it work on mingw32 with win32 backend
134 (based on patch by Yasuhiro Matsumoto).
135
136 5.132 Fri May 29 09:00:39 CEST 2009
137 - do not keep a reference to the argument itself in
138 Coro::Semaphore::guard, as it could change later.
139 - support SO_RCVBUF/SO_SNDBUF nonstandard Coro::Socket options,
140 should support a prepare callback.
141
142 5.131 Mon Mar 16 23:20:37 CET 2009
143 - implement and document Coro->suspend, Coro->resume.
144 - fix Coro::Select implementation to not (often) close
145 the passed file descriptors (testcase provided by pippijn).
146
147 5.13 Mon Dec 15 21:51:42 CET 2008
148 - EXPERIMENTAL: implement dynamic winds (on_enter/on_leave).
149 - don't set diehook to C<undef> but instead to NULL, to avoid
150 spurious warnings.
151 - fix a lot of bugs in Coro::SemaphoreSet.
152 - Coro::SemaphoreSet will less often create a semaphore needlessly.
153 - add Coro::SemaphoreSet::count and wait methods.
154 - take advantage of the new Guard module.
155 - deprecate Coro::guard.
156 - try to fix the dreaded 01_unblock tests once more. I hate it when
157 testsuites need more fixing than the code they are supposed to test.
158 - croak in more cases when a required callback isn't resolvable.
159 - fix some minor issues in Coro::State->call/eval.
160 - use current coroutine context instead of a temporary one
161 when temporarily switching to another coroutine.
162 - do not call C-level on_destroy handlers during global destruction,
163 to avoid needless segfaults.
164
165 5.12 Sun Dec 7 13:30:38 CET 2008
166 - add default config for MirOS, which seems to be a bug-to-bug
167 compatible fork of openbsd ("world domination by releasing
168 openbsd cvs before the openbsd folks do it" or so :).
169 - free_padlist did destroy the names pad, not good, but didn't
170 seem to bother perl - this could fix issues such as eval ""
171 inside a function called from multiple coroutines.
172 - use a different method to detect destruction time.
173 - be more careful when freeing padlists just before global
174 destruction.
175 - fixed and expanded the call/cc example.
176 - renamed _terminate to _coro_run.
177 - new method Coro::Channel->shutdown.
178 - try pthreads on openbsd <4.4 (broken sigaltstack, will
179 pthreads fare better?).
180 - be less picky about destroying "a" running coroutine.
181
182 5.11 Tue Nov 25 21:49:05 CET 2008
183 - DEBUGGING in 5.10.0 is a mess: it constantly flags perfectly
184 working code with failed assertions, introducing more bugs than
185 it fixes, requiring elaborate workarounds :(
186
187 5.1 Mon Nov 24 08:54:59 CET 2008
188 - wrote a small introductory tutorial - Coro::Intro.
189 - convert Coro::Timer, Coro::Select and Coro::Util to rouse API.
190 - Coro::Select did errornously dup the file descriptors
191 and didn't work with all AnyEvent backends.
192 - Coro::Select wasn't imported correctly form Coro::LWP, causing blocking
193 LWP data transfers.
194 - disassociate c contexts from coro objects - this is agruably more
195 correct, but mostly allows sharing of cctxs between coro and state
196 objects, for added memory savings and speed increases.
197 - bumped $Coro::POOL_RSS up to 32kb by default.
198 - no longer set the optype to OP_CUSTOM, as B::* understandably
199 doesn't like this very much (and we *are* a type of entersub).
200 - implement state cloning, just to prove that call/cc can be done.
201 - automatically load Coro::AnyEvent in Coro::Handle.
202 - wrap ->cancel calls in eval inside Coro::Handle as EV watchers
203 do not have this method (and don't need it either).
204 - speed up generic anyevent methods in Coro::Handle by using rouse
205 callbacks.
206 - allow coroutines in $Coro::IDLE, speeding up Coro::AnyEvent and
207 others. It also makes the debugger happier, as you can trace
208 through the idle threads now.
209 - add comppad_name* and hints ($^H, %^H) to per-thread variables.
210 - eg/event was pretty broken.
211 - better 5.8.6 compatibility.
212
213 5.0 Thu Nov 20 10:35:05 CET 2008
214 - NEW ARCHITECTURE: use the latest 4.x version if you experience
215 stability issues.
216 - bump API version to 7 - all dependents must be recompiled.
217 - removed timed_* functions - they were not being used anyways
218 and should be replaced by a more generic mechanism -
219 and were annoying to support anyways :)
220 - removed SemaphoreSet's waiter method - use sem method instead.
221 - Coro::Semaphore->adjust didn't correctly wake up enough waiters.
222 - async_pool did free a scalar value twice
223 ("Attempt to unreference...").
224 - fix a longstanding bug where calling terminate on a coro that
225 was waiting for a semaphore that was just becoming available
226 would cause a deadlock (semaphore would get into a state where
227 it was available but waiters were still blocked).
228 - calling throw on a coroutine that is waiting for a semaphore will
229 no longer make it acquire the semaphore (and thus leak a count).
230 - perl's process emulation is now not even theoretically supported
231 anymore.
232 - new functions Coro::rouse_cb and Coro::rouse_wait for easier
233 conversion of callback-style to blocking-style.
234 - new methods $coro->schedule_to and ->cede_to, to specifically
235 schedule or cede to a specific coroutine.
236 - new function Coro::Semaphore::wait.
237 - use named constants in Coro::Channel (Richard Hundt).
238 - directly patch the entersub opcode calling SLF functions (cede,
239 transfer and so on). this does speed up context switching, but
240 more importanly, it frees us from the hardcoded behaviour of
241 entersub, so we might actually be able to return something from
242 those functions and atcually create new ones.
243 - take advantage of __builtin_frame_address on gcc.
244 - expose THX in coroapi (not sure whether this was a wise decision,
245 as "threaded" perls are running at half speed anyways).
246 - implement execute_slf (schedule-like-function) interface that makes
247 it possible to implement schedule-like-functions in XS.
248 - use new SLF interface to massively speed up Coro::EV by roughly a
249 factor of two.
250 - used new SLF interface to massively speed up Coro::Semaphore by a
251 factor of three.
252 - used new SLF interface to speed up Coro::AIO by roughly a factor of
253 four and reduce its memory usage considerably.
254 - implement Coro::SemaphoreSet purely in terms of Coro::Semaphore,
255 for a nice speedup and vastly more correct behaviour. Also implement
256 a new method "sem" to get at the underlying semaphore object.
257 - implement Coro::Channel in terms of Coro::Semaphore, for a moderate
258 (in comparison) 20-40% speedup.
259 - used new SLF interface to reimplement Coro::Signal gaining
260 some unknown (because I was too lazy), but certain, speedup, and also
261 making signals reliable for the first time.
262 - used new SLF interface and other optimisations to speed up async_pool
263 by a factor of two. It also doesn't rely on perl's exception mechanism
264 to exit anymore. The overhead for terminating an async_pool, coro over
265 a normal async is now very small.
266 - sped up coroutine creation/destruction by 40%.
267 - forgot to include Coro/libcoro/README in the dist for all these years.
268 - work around a freebsd pthreads bug (manual testcancel is required as
269 pthread_cond_wait isn't a cancellation point on freebsd).
270 - use new rouse functions to speed up and simplify Coro::BDB.
271 - make "prefer perl native functions" work with threaded perls.
272 - condense Coro::Debug ps output, hint at v and w flags.
273 - (libcoro) lots of minor cleanups and portability improvements.
274
275 4.914 Wed Nov 19 12:54:18 CET 2008
276 - fix a disastrous bug in the readline optimisation
277 introduced in 4.801.
278
279 4.913 Sat Nov 15 07:58:28 CET 2008
280 - async_pool did free a scalar value twice
281 ("Attempt to unreference...").
282
283 4.912 Thu Nov 13 18:31:23 CET 2008
284 - minor cleanups.
285 - use much larger stacks on linux and perl < 5.8.8.
286 - Coro::Debug::new_unix_server did not unlink the socket
287 when destroyed.
288
289 4.911 Tue Nov 11 04:26:17 CET 2008
290 - "port" to threaded perls.
291
292 4.91 Mon Nov 10 05:36:38 CET 2008
293 - the ->throw exception object no longer leaks.
294 - creating a new cctx leaked a scopestack entry (memleak).
295 - new coroutines didn't get created with a zero flags field
296 (unknown impact).
297 - calling ->throw on a not-yet-started coroutine should now work
298 instead of being ignored.
299 - ->throw is now supported on Coro::State objects.
300 - clean up cctx creation code a bit.
301 - entersub is actually a UNOP, not a LOGOP (not a bugfix).
302
303 4.9 Sat Nov 8 17:45:27 CET 2008
304 - (libcoro) did not preserve rbp with CORO_ASM (we are getting there).
305 - (libcoro) no longer leak threads in the experimental pthread backend,
306 also speed it up considerably.
307 - (libcoro) do not rely on makecontext passing void *'s unscathed.
308 - fix compiletime dependencies on libcoro in the Makefile.
309 - cctx_count wasn't always updated properly.
310 - Coro::State::cctx_stacksize wasn't applied correctly.
311 - new function Coro::State::cctx_max_idle.
312 - the default max number of idle C contexts is now 4.
313 - (libcoro) try harder to get _setjmp/_longjmp.
314 - (libcoro) cleanup and extend the libcoro API to officially
315 allow the creation of empty source contexts.
316 - very experimental workaround for the totally broken netbsd platform.
317 - tried to hack around openbsd bugs.
318
319 4.804 Wed Nov 5 16:36:00 CET 2008
320 - Coro::Debug::new_unix_server would not create a non-blocking listening
321 socket, sometimes causing freezes.
322 - (libcoro) fix misaligned stack points for setjmp and assembly
323 methods, which can cause crashes on x86/x86_64 with a sufficiently
324 aggressive compiler.
325 - new function: Coro::Debug::new_tcp_server.
326 - move ->throw into the Coro class because it only works on coro objects.
327
328 4.803 Mon Nov 3 17:16:12 CET 2008
329 - (libcoro) use a global asm statement to become independent of gcc
330 otpimisations for CORO_ASM (thanks to pippijn for the idea).
331 - try to workaround yet another broken bsd, this time dragonfly.
332
333 4.802 Thu Oct 30 10:56:12 CET 2008
334 - support -fno-omit-frame-pointer on x86 with the assembly method.
335 - tune 01_unblock.t tests a bit.
336
337 4.801 Wed Oct 22 18:33:37 CEST 2008
338 - improve readline speed for very long "lines".
339 - backport to 5.8.8.
340
341 4.8 Thu Oct 2 13:34:40 CEST 2008
342 - new function Coro::AIO::aio_wait.
343 - Coro.:AIO and Coro::BDB now "use Coro::AnyEvent".
344 - greatly speed up and reduce memory usage of Coro::AIO requests.
345 - implement some other µ-optimisations.
346
347 4.749 Mon Sep 29 14:40:12 CEST 2008
348 - port to slow and broken pseudo-threaded perls. (courtesy pippijn).
349
350 4.748 Sat Sep 27 14:03:19 CEST 2008
351 - implement, but do not document, PerlIO::cede(granularity).
352 - Coro::Storable forgot to wrap Storable::pstore.
353 - work around the multitude of leaks and memory corruption
354 bugs in PerlIO::via by using our own C-level perliol. As a side
355 effect, Coro::Storable is now much, much, much faster.
356
357 4.747 Tue Sep 23 01:59:41 CEST 2008
358 - fix a per-cv memleak (one empty array was leaked per
359 code reference).
360 - avoid a crash in coro->call|rss when the coroutine was already
361 destroyed (most noticably when using Coro::Debug::ps :)
362 - also protect *Storable::FILE.
363 - push up default storable granularity to 20ms.
364
365 4.746 Sun Sep 21 03:22:20 CEST 2008
366 - be more insistent on locking Storable against reentrancy
367 in Coro::Storable.
368 - move swap_def?v and throw to Coro::State, as documented.
369
370 4.745 Thu Jul 24 00:14:38 CEST 2008
371 - remove debugging code related to MgPV_nolen_const, also try to
372 make it compile with perl 5.8.6 (yes, apple apparently loves
373 outdated software). Reported by John S.
374
375 4.744 Tue Jul 8 22:06:35 CEST 2008
376 - correctly provide default DIE/WARN handlers as documented.
377 - also overwrite PL_vtbl_sigelem.svt_clear, even though current
378 implementations inside perl work fine for us.
379
380 4.743 Mon Jun 16 00:21:57 CEST 2008
381 - when using Coro::EV without running EV::loop it could
382 result in busy-waiting for events, this has been fixed.
383 - reduce codesize and improve performance by using EV_DEFAULT_UC.
384
385 4.742 Sat May 31 14:10:21 CEST 2008
386 - implement a workaround for (some) perl <5.8.8 versions.
387 - require EV 3.3+.
388
389 4.741 Fri May 30 23:33:09 CEST 2008
390 - tell netbsd how utterly broken their imitation of an OS is
391 and refuse to build by default if pthreads are in use.
392 - switch to "s" method on all bsd's by default, as their ucontext
393 stuff seems just too broken.
394 - fix a bug in Coro::Select.
395
396 4.74 Thu May 29 20:05:31 CEST 2008
397 - do not test Coro::LWP for lack of dependencies.
398 - do not test Coro::Select for lack of working perls.
399
400 4.73 Thu May 29 2008
401 - fix a bug in Coro::EV which would cause it to block despite
402 there being runnable coroutines.
403 - sprinkle "no warnings" freely over everything, also suppress
404 warnings for some other modules.
405 - fix typo in WSAEWOULDBLOCK.
406
407 4.72 Sun May 25 05:14:36 CEST 2008
408 - tweak META.yaml a bit, unfortunately, there is no documented way
409 to have optional dependencies with CPAN. doh :(
410 - avoid running some tests on windows because they would fail due to
411 perl bug (broken fork, broken pipes...).
412 - work around perl on windows bugs where perl returns undocumented
413 error codes for sysread, syswrite etc. by taking advantage
414 of AnyEvent's workaround for that problem.
415 - use AnyEvent::Util::fh_nonblocking in Coro::Handle to work around
416 a common perl implementation bug on windows.
417 - use unix domain sockets in testsuite to work around
418 common perl implementation bugs on widows (they are emulated by
419 tcp sockets on windows. ugh.)
420
421 4.71 Sat May 24 20:01:27 CEST 2008
422 - fix a bug in Coro::AnyEvent ("Usage: Coro::AnyEvent::_schedule()").
423 - take advantage of async name resolution of AnyEvent::Util.
424 - work around brutal inet_aton override in Coro::LWP.
425 - take advantage of the readyhook in Coro::EV, for smoother
426 scheduling.
427
428 4.7 Sun May 11 00:32:19 CEST 2008
429 - completely reworked the Coro manpage.
430 - added Coro::AnyEvent, generic event loop integration.
431 - implement cancel, ready and kill commands in Coro::Debug.
432 - document find_coro in Coro::Debug.
433 - incompatible change: rename has_stack to has_cctx.
434 - Coro::AIO and Coro::BDB no longer force event model detection,
435 use AnyEvent::AIO and AnyEvent::BDB.
436
437 4.6 Sat Apr 26 10:05:14 CEST 2008
438 - INCOMPATIBLE CHANGE: sub/code attributes are no longer supported
439 by the Coro module. It was a mistake to have it in the first place.
440 - (experimental) support for activestate perl 5.10 (method "w").
441 - (experimental) support for strawberry perl (method "a").
442 - coro_sigelem_set did not return a value although it had to,
443 actual impact unknown.
444
445 4.51 Mon Apr 14 13:28:27 CEST 2008
446 - make it compile again on 5.8.
447
448 4.50 Thu Apr 10 09:43:17 CEST 2008
449 - I did it twice! (see 4.49).
450
451 4.49 Sun Apr 6 21:23:31 CEST 2008
452 - grr, instead of compiling the recent changes on 5.10 only they
453 were compiled on 5.8 only.
454
455 4.48 Sun Apr 6 20:36:46 CEST 2008
456 - allow coroutine switches during eval's under 5.10.x, as it seems
457 the parser is a per-interpreter option now, so this is safe
458 (this might fix the odd crashing bug).
459 - drop support for 5.9.x versions: they are dead, jim.
460
461 4.47 Sun Apr 6 00:37:52 CEST 2008
462 - force cctx allocation on API calls: we know we need to force one
463 and gcc actually manages to confuse our heuristic nowadays,
464 leading to crashes and worse.
465 - document force_cctx.
466
467 4.46 Fri Apr 4 22:05:43 CEST 2008
468 - upgrade libcoro, resulting in pthread-backend (which was only created
469 to fulfill the rules of the programming languages shootout).
470
471 4.45 Thu Mar 13 11:55:36 CET 2008
472 - fix a file leaking bug in eg/mhyttpd that would allow
473 downloading of any file (reported by oesi).
474 - fix deadlock bug in Coro::Channel (reported by Richard Hundt)
475 (also add testcase).
476 - support Broadcast option in Coro::Socket (patch by Richard Hundt,
477 apparently having loads of fun with that).
478 - implement $state->swap_defsv, ->swap_defav and document ->throw.
479
480 4.4 Wed Feb 13 16:44:29 CET 2008
481 - only assume .cfi pseudo isns exist on GNU/Linux.
482 - add get_timed method to Coro::Channels.
483 - fixed Coro::Debug manpage.
484 - perl 5.11 compatbility improvement based on patch by
485 Andreas König.
486
487 4.37 Sun Jan 20 11:25:23 CET 2008
488 - soften the check for an active parser for perl 5.10 (avoiding
489 "Coro::State::transfer called while parsing" in many cases).
490
491 4.36 Sun Jan 13 10:53:56 CET 2008
492 - reset diehook when terminating from an async_pool as to not
493 trigger any __DIE__ handlers.
494
495 4.35 Sun Jan 13 04:14:13 CET 2008
496 - "bt" debug command now displays any exceptions
497 from longmess and also skips the innermost
498 stackframes, giving more useufl output.
499 - allow backtraces in coroutines blocked in Coro::EV,
500 at a <1% speed hit.
501 - handle localising of $SIG{__WARN__} and __DIE__
502 properly (with a proper amount of dirty hacking).
503
504 4.34 Sat Dec 22 17:49:53 CET 2007
505 - upgrade to EV version 2.0 API.
506
507 4.33 Mon Dec 17 08:36:12 CET 2007
508 - make Coro::AIO etc. loadable in the absence of EV.
509
510 4.32 Mon Dec 17 07:46:02 CET 2007
511 - majorly improved Coro::Handle's performance with Coro::EV.
512 - implemented the until now mythical Coro::BDB module.
513 - specialcase EV in Coro::AIO and Coro::BDB for extra speed.
514
515 4.31 Wed Dec 5 12:32:39 CET 2007
516 - remove warn statement form Coro::Util.
517
518 4.3 Tue Dec 4 20:33:14 CET 2007
519 - calls to the idle function could cause stack corruption
520 when the stack changed.
521 - do no longer rely on the presence of EV::DNS if EV is used
522 (because it is gone), but instead take avdantage of EV::ADNS
523 if available.
524 - add ($) prototypes to all functions in Coro::Storable.
525 - use a conventional (safer) idle callback in Coro::EV.
526 - do accept quickly in Coro::Debug to avoid endless loops.
527
528 4.22 Fri Nov 30 16:04:04 CET 2007
529 - really use optimised versions for Event and EV in Coro::Util
530 and Coro::Handle.
531
532 4.21 Sun Nov 25 10:48:59 CET 2007
533 - fix a spurious memory read.
534 - Coro::EV no longer keeps the eventloop "alive".
535
536 4.2 Fri Nov 9 20:47:05 CET 2007
537 - enable/disable tracing from a new coroutine, not a pooled one.
538 - fix a memleak in Coro::Event.
539 - removed killall call from fork_eval.
540 - made sure store_fd is already loaded so that fork_eval
541 does not have to parse autoload in each subprocess.
542 - only use assembly method if -O switch is in $Config{optimize}.
543 - add (optional) Coro::EV module, so far the best event loop module
544 directly supported by Coro.
545 - if the event model is EV, use EV::DNS to resolve
546 stuff in Coro::Util.
547 - don't get confused by multiple event notifications in Coro::Handle.
548 - initial support for EV (libevent interface).
549 - require Event and EV using configure_requires, to force their existance.
550
551 4.13 Wed Oct 24 07:26:45 CEST 2007
552 - add Coro::Storable::blocking_thaw.
553 - use a vastly more complicated technique to localise
554 $SIG{__WARN/DIE__} that also works on perls <= 5.8.8.
555 - use a coroutine for the idle callback Coro::Event,
556 instead of running Event in the current coroutine context.
557 This also catches recursive invocations.
558 - actually report fork errors in gethostbyname and inet_aton.
559
560 4.11 Thu Oct 11 02:40:24 CEST 2007
561 - port to threaded perls.
562
563 4.1 Thu Oct 11 02:38:16 CEST 2007
564 - incompatible change: $SIG{__DIE__} and $SIG{__WARN__} will now
565 be local to each coro (see Coro::State).
566 - incompatible change: for very deep reasons, cede and cede_notself
567 cannot return anything, so nothing will be returned.
568 - possibly bring back 5.10 compatibility (untested).
569 - work around stupid (and wrong) warning on 5.10 :(.
570 - overlay the saved state over the context stack. This saves
571 a few hundred bytes per coroutine on average and also
572 speeds up context switching a bit.
573 - further tune default stack sizes.
574 - (more) correctly calculate stack usage in coro_rss.
575 - Coro::Storable::blocking_* did not properly lock
576 resulting in races between coroutines.
577 - added Coro::Storable::guard.
578 - stopping to trace a coroutine could destroy the cctx of
579 an unrelated coroutine.
580 - explain the relationship between Perl and C coroutines in
581 more detail in Coro::State.
582 - Coro::Util::inet_aton did not short-circuit dotted quad forms,
583 causing a fork per resolve. This also affected Coro::Socket.
584 - switch to a separate stack in $coro->call/eval to avoid
585 invalidating pointers.
586
587 4.03 Sat Oct 6 21:24:00 CEST 2007
588 - added Coro::throw method.
589 - minor code cleanups.
590
591 4.02 Sat Oct 6 02:36:47 CEST 2007
592 - fix a very minor per-coroutine memleak (a single codereference).
593 - fixed a bug where the currently in-use c context would be freed
594 prematurely (can happen only when programs change the stacksize
595 or use tracing).
596 - tracing can no longer keep a coro alive after it terminated.
597 - do static branch prediction in the common path for gcc. gives
598 about 2-5% speed improvement here.
599
600 4.01 Fri Oct 5 22:10:49 CEST 2007
601 - instead of recreating *a* standard output handle we simply
602 use STDOUT, which is faster and hopefully more robust.
603
604 4.0 Fri Oct 5 12:56:00 CEST 2007
605 - incompatibly changed Coro::Storable::freeze.
606 - major new feature: added Coro::Debug, for interactive coroutine
607 debugging, tracing and much more.
608 - major bug fix: unbelievable, but true: $_, $/ and many other
609 "saved" variables actually weren't being saved. This has been fixed,
610 of course, while increasing performance while losing all the save
611 flags.
612 - save flags are gone, and all the api functions dealing with them.
613 - added Coro::Semaphore::adjust.
614 - added Coro::Util::fork_eval.
615 - added Coro::Storable::{nfreeze,blocking_{freeze,nfreeze}}.
616 - added Coro::killall.
617 - reduce initial stack sizes to allow for "micro-coroutines".
618 - better async_pool resource management, moved parts of async_pool
619 handling to XS (major speed improvement).
620 - actually croak before modifying important data structures.
621 - refuse to transfer while compiling.
622 - possibly support eval EXPR better now.
623 - enable assembly per default on linux+bsd x86+amd64.
624 - all internal members were renamed _something for easier subclassing.
625 - many minor tweaks.
626
627 3.63 Wed May 16 14:10:06 CEST 2007
628 - implement handcoded assembly for x86/amd64 SVR ABI.
629
630 3.62 Fri Apr 27 21:36:06 CEST 2007
631 - upgrade libcoro (which might set unwind info correctly).
632 - change default on linux to setjmp/longjmp.
633
634 3.61 Thu Apr 19 12:36:18 CEST 2007
635 - Coro::Storable caused an endless loop when thawing invalid
636 pst's sometimes.
637 - use a Semaphore in Coro::Storable, as Storable doesn't
638 seem to be reentrant (although it is documented to
639 be threadsafe...).
640 - fix Coro::Signal to bring back the original unreliable
641 but stateful semantics.
642 - fixed a lot of typos in Coro.pm (patch submitted by David
643 Steinbrunner, which applied flawlessly).
644
645 3.6 Sat Apr 14 17:13:31 CEST 2007
646 - added some bugfixes to get eg/myhttpd working again.
647 - added Coro::Storable for often-cede'ing freeze/thaw.
648 - try to do a clean exit when a coroutine calls exit
649 (EXPERIMENTAL).
650 - got rid of indirect call through _coro_init.
651 - updated the partly antique examples in eg/ to
652 work again and be a bit less magic, too.
653 - fixed Coro::Signal semantics to work as documented again.
654
655 3.55 Sun Mar 25 01:20:47 CET 2007
656 - add SAVE_DEFFH to save the default filehandle and enable
657 it by default.
658 - finally move socket-operations from Coro::Socket to Coro::Handle
659 to be able to unblock foreign sockets.
660 - add Coro::State::save_also and guarded_save.
661 - add count accessor to Coro::Semaphore.
662 - add Coro::State::cctx_stacksize.
663 - just for the fun of it, do not rely on implicit context,
664 which can dramatically improve performance, but people
665 using windows-process-emulation perls do not care much about
666 performance.
667
668 3.51 Sun Mar 4 14:18:04 CET 2007
669 - fixed a problem when you weakened references to Coro::State's
670 (patch by Yuval Kogman).
671
672 3.501 Wed Feb 28 12:44:07 CET 2007
673 - rename some global symbols as macosx from hell redefines
674 them without asking.
675
676 3.5 Tue Feb 13 20:22:53 CET 2007
677 - do AnyEvent model detection earlier, avoiding problems
678 caused by first using AnyEvent and later Coro::Event.
679 - implement and document Coro::Event event objects.
680 - fix a potential problem in Coro::Event causing crashes.
681 - initialise PL_comppad when creating a new coroutine,
682 avoids crashes on early coro destruction.
683
684 3.41 Mon Jan 22 19:19:49 CET 2007
685 - readline on Coro::Handle did not support undefined $/,
686 nor did it deliver partial lines on EOF or error.
687 - implement malloc fallback for stack allocation because
688 stupid broken idiotic OSX has a stupid broken
689 idiotic fits-the-whole-os mmap "implementation" and
690 my dick feels longer if Coro is portable even to
691 obsolete platforms.
692
693 3.4 Fri Jan 19 21:52:54 CET 2007
694 - remove t/09_timer.t, as it isn't really testing much
695 but was rather flaky in practise.
696 - async_pool coro would keep arguments and callback alive until
697 it was reused.
698 - cancellation of a coroutine could cause spurious idle calls
699 in cede_notself.
700
701 3.3 Sat Jan 6 03:45:00 CET 2007
702 - implement $coro->on_destroy.
703 - Coro::Event blocking semantics have been changed,
704 documented and - hopefully - improved.
705 - fix nice adding, not subtracting, from priority.
706 - fix ->prio and api_is_ready (patch by Mark Hinds).
707 - fixed an assert ("... == prev__cctx->idle_te")
708 that could errronously trigger.
709 - fix various large and small memleaks.
710 - use a (hopefully) more stable cancel implementation
711 that immediately frees the coroutine data.
712 - cede/cede_notself return a status now.
713 - added Coro::guard function.
714 - added a global coroutine pool for jobs (on my machine,
715 I can create and execute 48k simple coros/s with async,
716 and 128k coros with async_pool).
717 - Coro::AIO now uses the coroutine priority as io priority.
718
719 3.2 Fri Dec 22 05:07:09 CET 2006
720 - improve portability to slightly older perls.
721 - use cleaner coroutine destruction.
722 - simplify configuration for users.
723 - optionally (unrecommended) prefer perl functions over
724 their coro replacements.
725
726 3.11 Tue Dec 5 13:11:24 CET 2006
727 - fixed some bogus assert's, but as perl.h disables assert even
728 without NDEBUG (thank you very much), not too many people should
729 notice (that did include myself). Andreas König noticed, though :)
730 - do not save/restore PL_sortcxix on >= 5.9.x, it doesn't seem to have
731 it. Also noticed by Andreas König :)
732 - save/restore tainted status.
733 - verified to pass the testsuite on my 5.9.5.
734
735 3.1 Mon Dec 4 23:03:40 CET 2006
736 - INCOMPATIBLE CHANGE: $/ is now per-coroutine (but slow).
737 - incompatible change: transfer flags are now per-state.
738 - give a better error message on deadlock.
739 - document Coro::nready.
740 - enhanced testsuite.
741
742 3.01 Sun Dec 3 23:47:42 CET 2006
743 - forgot to include Coro::Timer.
744
745 3.0 Sun Dec 3 22:57:25 CET 2006
746 - the "FINALLY COMPLETELY STABLE" release (coming soon:
747 the "FAMOUS LAST WORDS" release).
748 - implement a new stack sharing algorithm, which uses a stack
749 pool (size currently hardcoded).
750 - make stack sharing mandatory (it no longer uses a heuristic).
751 - eval/die no longer cause weird problems under heavy use.
752 - Coro::Event could cause livelocks if it was used but
753 no Coro::Event watchers were used.
754 - Coro::Event now uses asynccheck as prepare does not
755 check for changed watchers.
756 - Coro::Event allows multiple waiting coros and will wake up one
757 per event.
758 - Coro::Event should be cleaner and more efficient now.
759 - new utility function Coro::unblock_sub.
760 - document the sad fact that Event is no longer reentrant.
761 - putting a coroutine into the ready queue twice could under
762 some circumstances lead to stack corruption.
763 - minor incompatibility: subclassing Coro::State is supported
764 directly now without going through a _coro_state member.
765 - state/coro switching is much faster now.
766 - very minor optimisations and code/documentation cleanup.
767 - avoid problems due to compiler inlining.
768 - removed timers from Coro::Timer -> use AnyEvent instead.
769 - replaced Coro::idle coroutine by (cleaner) idle handler.
770 - updated to newest libcoro.
771 - implement enhanced support for valgrind.
772 - implement is_ready and return value for ready.
773 - removed Coro::Cont, it was a misnomer (it's generators),
774 and it was somewhat annoying to get it right. Will come back
775 if somebody asks for it :)
776 - many ->wait methods and Coro::Event could return
777 spuriously without the event having happened.
778
779 2.5 Tue Nov 7 12:22:33 CET 2006
780 - made Coro::Util, Coro::Select, Coro::Handle and Coro::Socket
781 use AnyEvent, moved them to Coro/.
782 - added Coro::LWP which contains all the uglyness required to
783 make LWP non-blocking.
784 - should work with perl 5.9.x now (Andreas König made me do it).
785 - fixed another bug in Coro::Select when the timeout was undef.
786 - reuse PL_start_env for all coros, saves some memory per coroutine.
787 - manage PL_top_env differently, hopefully to avoid panic: top_env.
788 - timeout argument was not properly used in Coro::Socket.
789 - allow limited forms of subclassing in Coro::Handle/Coro::Socket.
790 - emulate undocumented(!) functionality of IO::Socket required
791 by LWP(!!).
792 - updated eg/lwp to work with newer lwp's.
793 - remove "FATAL: uncaught exception" prefix. Coroutines that die
794 kill the whole process, just as exceptions in the main "coroutine"
795 did already.
796
797 2.1 Wed Nov 1 23:01:13 CET 2006
798 - fix a long-standing bug in Coro::Select where select with
799 zero timeout would instead change the current default filehandle.
800 - use a simpler and hopefully more robust way to clone padlists
801 (uses less memory and a perl function instead of our own).
802 - coro can now create a stack guard area on many architectures.
803 - Coro::AIO properly reexports additional functions from IO::AIO.
804 - updated libcoro with a workaround for OS X,
805 pach and testing by Michael Schwern.
806
807 2.0 Tue Oct 24 05:47:17 CEST 2006
808 - support additional aio requests in Coro::AIO.
809
810 1.9 (never properly released due to a glitch)
811
812 1.8 Thu Feb 2 00:59:06 CET 2006
813 - applied suggested patch by SAMV to avoid problems during stupid
814 mark & sweep gc run.
815 - applied patch by Scott Walters for 5.9.3 compatibility.
816
817 1.7 Tue Dec 27 01:41:58 CET 2005
818 - added Coro::AIO, a thin wrapper around IO::AIO.
819 - improved Makefile.PL explanations.
820
821 1.6 Mon Dec 12 21:30:05 CET 2005
822 - additionally save PM_curpm might fix as-of-yet
823 unseen problems with regex matches being attributed to the
824 wrong package.
825 - add t/10_bugs.t, which currently checks against imho broken
826 perls that use 0.26MB of stack space per Perl_magic_get
827 invocation (newer linuxes) as opposed to the 0.0002MB perl
828 normally uses.
829 - make stacksize configurable for the ultimate debian experience.
830
831 1.51 Mon Dec 12 18:48:36 CET 2005
832 - remove debugging warn() accidentally left in Coro::Select.
833
834 1.5 Tue Nov 29 13:32:44 CET 2005
835 - use Coro::Event inside Coro::Select to avoid spurious deadlocks.
836 - fix Coro::Select 'select'.
837 - strict'ify some modules.
838 - libcoro errornously restored the SIGUSR2 handler to SIGUSR1.
839 - use XSLoader in selected modules.
840 - remove some 5.6 compatibility cruft.
841
842 1.4 Tue Sep 6 00:11:05 CEST 2005
843 - libcoro did not take into account the trampoline on amd64,
844 when 'l'inux method was used.
845
846 1.31 Tue Aug 30 23:31:33 CEST 2005
847 - some portability fixes/workarounds.
848
849 1.3 Sat Aug 20 03:08:56 CEST 2005
850 - no code changes. module seems to work fine.
851 - improve Coro::State docs, remove reference to nonexisting
852 Coro::State::flush method.
853 - no longer autodetect windows, present it as an option instead.
854
855 1.2 Mon May 16 02:00:55 CEST 2005
856 - included libcoro.c earlier in State.xs, might improve portability.
857 - use a faster and possibly more stable Coro::Cont implementation.
858 - accept x86_64 in addition to amd64 for optimized linux-amd64 support.
859 - fix bugs Coro::SemaphoreSet that could cause locks never to be freed.
860 - fix bugs in CoroAPI.h and document it in Coro::MakeMaker.
861
862 1.11 Thu Mar 3 18:00:52 CET 2005
863 - change of contact address.
864
865 1.1 Tue Feb 22 20:51:16 CET 2005
866 - support [l]inux method on amd64.
867 - allow some context switches while compiling/eval'ing.
868
869 1.0 Tue Aug 10 03:47:27 CEST 2004
870 - Event 0.89 fixes the bug again.
871 - fix the remaining(??) memleak. I didn't free the cache-AV for
872 padlists, so one AV per created anon-sub leaked.
873
874 0.97 Fri May 14 15:23:32 CEST 2004
875 - remove stupid gcc extension that was an experiment only anyway.
876
877 0.96 Thu May 13 18:09:29 CEST 2004
878 - tell people if they have yet another broken Event variant.
879 - made Coro to work _better_ with threads (it seems that, as long
880 as you use Coros in one thread only everything will work fine.
881 Looking at the perl source the problem might be missing locking
882 between Coro and some perl internal routines. That's not easily
883 fixable).
884 - (hopefully) got rid of the static xnext inside transfer.
885
886 0.95 Sun Feb 15 03:20:28 CET 2004
887 - removed Coro::State::flush, implemented a new and much
888 faster caching mechanism by attaching magic to the CV.
889 - a memleak with (real) closures remains.
890 - dramatically reduced size of stacks to 65536 (linux/x86), after
891 observing that even myhttpd never had more than a kilobyte
892 of stack in use (stack is still only physically allocated in
893 pagesize increments for systems suporting mmap).
894 - fix for a "cygwin-64" architecture added, proposed by Stefan Traby.
895 - fix for threaded perl.
896
897 0.9 Sun Nov 23 04:32:39 CET 2003
898 - fix a few very hard to track down but (of course) stupid
899 memory leaks.
900 - stringify version number to avoid locale problems :(.
901 - perl 5.9.0 does not have SvPADBUSY, reported by Scott Walters,
902 hopefully ignoring it "just works".
903
904 0.8 Wed Nov 5 19:38:40 CET 2003
905 - port to perl5.005_03, but only 5.8.x+ is supported!
906 - honor LocalAddr even without LocalPort in Coro::Socket.
907
908 0.7 Tue May 27 03:12:38 CEST 2003
909 - the version jump indicates some level of testing, not gobs
910 of new features.
911 - uh... I found the prompt function in ExtUtils::MakeMaker.
912 Highly correct stuff, that is...
913 - fixed(?) a bug with die's in coroutines causing "panic: top_env".
914 the fix is not well understood by the author, so beware :(.
915
916 0.652 Thu May 8 02:54:46 CEST 2003
917 - Applied patch by Slaven Rezic to set default to "s" on FreeBSD,
918 cause version 4 doesn't have ucontext.
919 - Benjamin Reed reported that setjmp works fine on darwin,
920 so preselect it.
921
922 0.651 Sat Mar 29 15:00:23 CET 2003
923 - fix a bug in Coro::Handle where some bytes could get lost
924 on reading, reported by jason@nichego.net.
925
926 0.65 Sun Mar 23 00:08:26 CET 2003
927 - added a README.linux-glibc.
928 - new module Coro::Select.
929 - also save/restore PL_comppad, fixes segfaults with 5.8.1.
930 (I never claimed that I know what I am doing ;).
931 - set default to setjmp/longjmp on non-x86-linux.
932
933 0.6 Thu Nov 21 11:09:06 CET 2002
934 - made lazy context switching the default.
935 - fixed the bug where SemaphoreSet::guard
936 was the same as timed_guard and timed_guard was missing.
937 - fix a memleak in Coro::Timer.
938 - reclassified context sharing as not experimental.
939
940 0.534 Sun Apr 14 03:05:12 CEST 2002
941 - fixed backspace => tab in header parsing (myhttpd).
942 - added eg/readline, for elmex.
943 - Coro::Event now calls ->start on first call instead of
944 ->again, which created wrong timeouts for the first ->next.
945 - fixed a bug where Coro::Socket returned a closed socket instead
946 of nothing on ECONNREFUSED and similar error conditions.
947
948 0.533 Mon Feb 18 18:49:40 CET 2002
949 - adapt to finally working Event-0.85.
950 - fixed buggy event initialization (trapped under windows, but how
951 could it possibly have worked anywhere?)
952
953 0.532 Wed Jan 16 02:45:32 CET 2002
954 - removed the reference to pp_entersub, might work on
955 (some) win32 perls now (testsuite works under cygwin,
956 without Event, which freezes).
957
958 0.531 Mon Dec 10 22:18:44 CET 2001
959 - Fixed a bug in SemaphoreSet::timed_down.
960
961 0.53 Tue Nov 27 21:11:13 CET 2001
962 - some tests for Event.
963 - slightly different internal architecture to get refcnt into the object
964 and not the reference. might fix a few bugs, certainly did introduce
965 new ones.
966 - Coro::Timer (independent of Event).
967 - new timed_wait functions for Coro::Signal, Semaphore, SemaphoreSet.
968
969 0.52 Tue Nov 6 21:36:18 CET 2001
970 - ported to cygwin (trivial).
971 patch by Gerrit P. Haase <gp@familiehaase.de>.
972 - small setjmp code fix by Sullivan.DanielJ@epamail.epa.gov.
973
974 0.51 Thu Nov 1 20:39:01 CET 2001
975 - terminate/cancel now work properly (otherwise termination
976 could cause a "next coroutine is not and contains not..."
977 error.
978 - added Coro::Socket::shutdown.
979 - Coro::Event::loop no is the same as Event::loop.
980 - implemented terminate with args + join.
981
982 0.5 Fri Sep 28 16:15:35 CEST 2001
983 - fixed "print" on a Coro::Handle. The print method worked.
984 - small tweaks (seem to reduce memory consumption a lot)
985 in various modules.
986 - splendid use of "no warnings" scattered throughout.
987 - added Coro::Handle::rbuf, fixed Coro::Handle::fh.
988
989 0.49 Sun Sep 16 02:42:45 CEST 2001
990 - changed some method calls to function calls for speed
991 inside Coro::Handle.
992 - make Coro::Handle use an array instead of a hash for
993 speed reasons.
994 - IRIX mystery solved: it's SGI's NT, after all: "standard, huh?".
995 sjlj and ucontext should now work.
996 - IRIX-specific port for libcoro.
997 - swapped order of accept results to match IO::Socket.
998 - changed getsock/peername to sock/peername to match IO::Socket.
999 - fixed a bug that caused segfault when returning to main task
1000 under some circumstances.
1001 - other bugfixes.
1002
1003 0.45 Sun Sep 2 02:54:01 CEST 2001
1004 - new method Coro::Handle::timeout.
1005 - corrected speling of set...name to get...name in Coro::Socket.
1006 - Coro::Socket::accept now returns a Coro::Socket, not a Coro::Handle.
1007 - Coro::Handle now supports fileno and FILENO.
1008 - added eg/myhttpd, a web server.
1009 - Coro::Socket now accepts numeric ports as well as "port(number)"
1010 syntax.
1011 - moved some scheduling primitives into xs code. more to come.
1012 - new simple priority system.
1013 - implemented Coro::Event using XS, almost four times faster!
1014 - small memory corruption problem fixed (boy that was difficult).
1015
1016 0.13 Wed Aug 8 16:53:07 CEST 2001
1017 - new method Coro::Handle::readline.
1018 - added eg/lwp and Coro::Handle::autoflush, to make LWP work.
1019 (see eg/lwp on how to make LWP non-blocking).
1020 - renamed Coro::Socket::new_inet to new.
1021 - added Coro::Util, some utility functions.
1022 - die/eval now works better (still get annoying (but true ;)
1023 "Callback called exit" messages).
1024 - Coro::Handle now supports timeout, and Coro::Socket Timeout.
1025 - much better Coro::Socket support.
1026
1027 0.12 Fri Jul 27 04:19:01 CEST 2001
1028 - do not use mmap if neither MAP_ANON nor MAP_ANONYMOUS is defined.
1029 - very experimental stack sharing algorithm. not 100% safe but
1030 should work well in practise ;)
1031 - added Coro::RWLock.
1032 - Coro::Specific now works.
1033
1034 0.11 Tue Jul 24 22:49:21 CEST 2001
1035 - added specialized hack for newer and older linux versions (fast).
1036 - renamed Coro::Event::IO to Coro::Handle.
1037 - new module Coro::Socket.
1038
1039 0.1 Tue Jul 24 01:47:53 CEST 2001
1040 - release candidate 3. A rather new internal structure :(
1041 - the great renaming: Damian Conway gave me a suitable replacement
1042 for yield.
1043 - added Coro/libcoro, a portable coroutine implementation for C
1044 (not even perl-dependent!!!), which can be used by Coro::State.
1045 - renamed SAVE_DEFSV (on xs-level only) to avoid symbol clash in
1046 perl-5.6.
1047 - new function Coro::Event::idle.
1048 - the idle process is now overriden
1049 by default in Coro::Event.
1050 - Coro::Channel now enforces the size.
1051 - canceling events now works in all cases.
1052 - Coro state now include $_ and $@.
1053 - yet another bug workaround that I do not really understand :(
1054 - new module Coro::Event::IO (very undocumented) to do
1055 non-blocking i/o.
1056 - performance tuning.
1057
1058 0.08 Thu Jul 19 06:13:25 CEST 2001
1059 - release candidate 2 (new functionality).
1060 - add Coro::State::flush function.
1061 - transfer now supports a flags value (mostly for speed ATM).
1062 - might compile and work in the presence of threads now.
1063 - continuations have a saner syntax.
1064 - no more memleaks.
1065
1066 0.07 Tue Jul 17 17:40:18 CEST 2001
1067 - release candidate 1 ;)
1068 - slightly nicer code.
1069 - fixed a scheduling bug in Coro::Event.
1070
1071 0.06 Tue Jul 17 04:23:24 CEST 2001
1072 - ok, I found the showstopper - the same sub must not be
1073 re-entered in two different coroutines, otherwise => crash. I
1074 see no easy solution to this problem, except by walking the call
1075 chaing and saving/restoring the cv's, which is what I do now.
1076 - memory leaks still latent, especially at thread termination.
1077 - Coro::Event now works (haha).
1078
1079 0.05 Sun Jul 15 17:32:20 CEST 2001
1080 - fixed a few issues in Coro::Event.
1081 - I forgot to include Coro::Event in 0.04 :(:(:(
1082
1083 0.04 Sun Jul 15 05:24:59 CEST 2001
1084 - @_ is now properly localized.
1085 - Coro::State is now easier subclassable.
1086 - Coro::Cont now coroutine-aware.
1087 - Coro::Specific is a low-overhead module to create
1088 coroutine-specific vars.
1089 - Coro::Event provides a simple interface to Event.
1090
1091 0.03 Fri Jul 13 14:51:52 CEST 2001
1092 - transfer() now implemented in XS (beware).
1093 - new module Coro::Cont for really faked continuations.
1094 - big internal architecture changes: Coro::State is now
1095 really low-level and can thus be used to implement other
1096 interesting things, While "Coro::" implements a process-like
1097 model. Still crude and subject to change.
1098 - $_ and $@ are no longer being localized.
1099
1100 0.02 Tue Jul 10 01:38:17 CEST 2001
1101 - implemented "async" attribute.
1102 - $_ and $@ are now localized.
1103 - added Coro::Channel.
1104 - more testcases, still no docs.
1105
1106 0.01 Tue Jul 3 02:18:41 CEST 2001
1107 - original version; copied from Convert::Scalar.
1108