ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/Coro/Changes
Revision: 1.438
Committed: Wed Nov 19 05:58:48 2008 UTC (15 years, 6 months ago) by root
Branch: MAIN
Changes since 1.437: +1 -0 lines
Log Message:
*** empty log message ***

File Contents

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