ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/Coro/Changes
Revision: 1.418
Committed: Sun Nov 16 00:55:41 2008 UTC (15 years, 6 months ago) by root
Branch: MAIN
Changes since 1.417: +2 -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 5.0
9 - NEW ARCHITECTURE: use 4.912 for the stable version.
10 - bump API version to 7 - all dependents must be recompiled.
11 - Coro::Semaphore->adjust didn't correctly wake up enough waiters.
12 - async_pool did free a scalar value twice
13 ("Attempt to unreference...").
14 - perl's process emulation is now not even theoretically supported
15 anymore.
16 - directly patch the entersub opcode calling SLF functions (cede,
17 transfer and so on). this does speed up context switching, but
18 more importanly, it frees us from the hardcoded behaviour of
19 entersub, so we might actually be able to return something from
20 those functions and atcually create new ones.
21 - take advantage of __builtin_frame_address on gcc.
22 - expose THX in coroapi (not sure whether this was a wise decision,
23 as "threaded" perls are running at half speed anyways).
24 - implement execute_slf (schedule-like-function) interface that makes
25 it possible to implement schedule-like-functions in XS.
26 - use new SLF interface to massively speed up Coro::EV by roughly a factor
27 of two.
28 - removed timed_down and timed_guard functions - they were not being used
29 anyways and should be replaced by a more generic mechanism - and were
30 annoying to support anyways :
31 - used new SLF interface to massively speed up Coro::Semaphore by a factor
32 of three.
33 - forgot to include Coro/libcoro/README in the dist for all these years.
34 - work around a freebsd pthreads bug (manual testcancel is required as
35 pthread_cond_wait isn't a cancellation point on freebsd).
36
37 4.913 Sat Nov 15 07:58:28 CET 2008
38 - async_pool did free a scalar value twice
39 ("Attempt to unreference...").
40
41 4.912 Thu Nov 13 18:31:23 CET 2008
42 - minor cleanups.
43 - use much larger stacks on linux and perl < 5.8.8.
44 - Coro::Debug::new_unix_server did not unlink the socket
45 when destroyed.
46
47 4.911 Tue Nov 11 04:26:17 CET 2008
48 - "port" to threaded perls.
49
50 4.91 Mon Nov 10 05:36:38 CET 2008
51 - the ->throw exception object no longer leaks.
52 - creating a new cctx leaked a scopestack entry (memleak).
53 - new coroutines didn't get created with a zero flags field
54 (unknown impact).
55 - calling ->throw on a not-yet-started coroutine should now work
56 instead of being ignored.
57 - ->throw is now supported on Coro::State objects.
58 - clean up cctx creation code a bit.
59 - entersub is actually a UNOP, not a LOGOP (not a bugfix).
60
61 4.9 Sat Nov 8 17:45:27 CET 2008
62 - (libcoro) did not preserve rbp with CORO_ASM (we are getting there).
63 - (libcoro) no longer leak threads in the experimental pthread backend,
64 also speed it up considerably.
65 - (libcoro) do not rely on makecontext passing void *'s unscathed.
66 - fix compiletime dependencies on libcoro in the Makefile.
67 - cctx_count wasn't always updated properly.
68 - Coro::State::cctx_stacksize wasn't applied correctly.
69 - new function Coro::State::cctx_max_idle.
70 - the default max number of idle C contexts is now 4.
71 - (libcoro) try harder to get _setjmp/_longjmp.
72 - (libcoro) cleanup and extend the libcoro API to officially
73 allow the creation of empty source contexts.
74 - very experimental workaround for the totally broken netbsd platform.
75 - tried to hack around openbsd bugs.
76
77 4.804 Wed Nov 5 16:36:00 CET 2008
78 - Coro::Debug::new_unix_server would not create a non-blocking listening
79 socket, sometimes causing freezes.
80 - (libcoro) fix misaligned stack points for setjmp and assembly
81 methods, which can cause crashes on x86/x86_64 with a sufficiently
82 aggressive compiler.
83 - new function: Coro::Debug::new_tcp_server.
84 - move ->throw into the Coro class because it only works on coro objects.
85
86 4.803 Mon Nov 3 17:16:12 CET 2008
87 - (libcoro) use a global asm statement to become independent of gcc
88 otpimisations for CORO_ASM (thanks to pippijn for the idea).
89 - try to workaround yet another broken bsd, this time dragonfly.
90
91 4.802 Thu Oct 30 10:56:12 CET 2008
92 - support -fno-omit-frame-pointer on x86 with the assembly method.
93 - tune 01_unblock.t tests a bit.
94
95 4.801 Wed Oct 22 18:33:37 CEST 2008
96 - improve readline speed for very long "lines".
97 - backport to 5.8.8.
98
99 4.8 Thu Oct 2 13:34:40 CEST 2008
100 - new function Coro::AIO::aio_wait.
101 - Coro.:AIO and Coro::BDB now "use Coro::AnyEvent".
102 - greatly speed up and reduce memory usage of Coro::AIO requests.
103 - implement some other µ-optimisations.
104
105 4.749 Mon Sep 29 14:40:12 CEST 2008
106 - port to slow and broken pseudo-threaded perls. (courtesy pippijn).
107
108 4.748 Sat Sep 27 14:03:19 CEST 2008
109 - implement, but do not document, PerlIO::cede(granularity).
110 - Coro::Storable forgot to wrap Storable::pstore.
111 - work around the multitude of leaks and memory corruption
112 bugs in PerlIO::via by using our own C-level perliol. As a side
113 effect, Coro::Storable is now much, much, much faster.
114
115 4.747 Tue Sep 23 01:59:41 CEST 2008
116 - fix a per-cv memleak (one empty array was leaked per
117 code reference).
118 - avoid a crash in coro->call|rss when the coroutine was already
119 destroyed (most noticably when using Coro::Debug::ps :)
120 - also protect *Storable::FILE.
121 - push up default storable granularity to 20ms.
122
123 4.746 Sun Sep 21 03:22:20 CEST 2008
124 - be more insistent on locking Storable against reentrancy
125 in Coro::Storable.
126 - move swap_def?v and throw to Coro::State, as documented.
127
128 4.745 Thu Jul 24 00:14:38 CEST 2008
129 - remove debugging code related to MgPV_nolen_const, also try to
130 make it compile with perl 5.8.6 (yes, apple apparently loves
131 outdated software). Reported by John S.
132
133 4.744 Tue Jul 8 22:06:35 CEST 2008
134 - correctly provide default DIE/WARN handlers as documented.
135 - also overwrite PL_vtbl_sigelem.svt_clear, even though current
136 implementations inside perl work fine for us.
137
138 4.743 Mon Jun 16 00:21:57 CEST 2008
139 - when using Coro::EV without running EV::loop it could
140 result in busy-waiting for events, this has been fixed.
141 - reduce codesize and improve performance by using EV_DEFAULT_UC.
142
143 4.742 Sat May 31 14:10:21 CEST 2008
144 - implement a workaround for (some) perl <5.8.8 versions.
145 - require EV 3.3+.
146
147 4.741 Fri May 30 23:33:09 CEST 2008
148 - tell netbsd how utterly broken their imitation of an OS is
149 and refuse to build by default if pthreads are in use.
150 - switch to "s" method on all bsd's by default, as their ucontext
151 stuff seems just too broken.
152 - fix a bug in Coro::Select.
153
154 4.74 Thu May 29 20:05:31 CEST 2008
155 - do not test Coro::LWP for lack of dependencies.
156 - do not test Coro::Select for lack of working perls.
157
158 4.73 Thu May 29 2008
159 - fix a bug in Coro::EV which would cause it to block despite
160 there being runnable coroutines.
161 - sprinkle "no warnings" freely over everything, also suppress
162 warnings for some other modules.
163 - fix typo in WSAEWOULDBLOCK.
164
165 4.72 Sun May 25 05:14:36 CEST 2008
166 - tweak META.yaml a bit, unfortunately, there is no documented way
167 to have optional dependencies with CPAN. doh :(
168 - avoid running some tests on windows because they would fail due to
169 perl bug (broken fork, broken pipes...).
170 - work around perl on windows bugs where perl returns undocumented
171 error codes for sysread, syswrite etc. by taking advantage
172 of AnyEvent's workaround for that problem.
173 - use AnyEvent::Util::fh_nonblocking in Coro::Handle to work around
174 a common perl implementation bug on windows.
175 - use unix domain sockets in testsuite to work around
176 common perl implementation bugs on widows (they are emulated by
177 tcp sockets on windows. ugh.)
178
179 4.71 Sat May 24 20:01:27 CEST 2008
180 - fix a bug in Coro::AnyEvent ("Usage: Coro::AnyEvent::_schedule()").
181 - take advantage of async name resolution of AnyEvent::Util.
182 - work around brutal inet_aton override in Coro::LWP.
183 - take advantage of the readyhook in Coro::EV, for smoother
184 scheduling.
185
186 4.7 Sun May 11 00:32:19 CEST 2008
187 - completely reworked the Coro manpage.
188 - added Coro::AnyEvent, generic event loop integration.
189 - implement cancel, ready and kill commands in Coro::Debug.
190 - document find_coro in Coro::Debug.
191 - incompatible change: rename has_stack to has_cctx.
192 - Coro::AIO and Coro::BDB no longer force event model detection,
193 use AnyEvent::AIO and AnyEvent::BDB.
194
195 4.6 Sat Apr 26 10:05:14 CEST 2008
196 - INCOMPATIBLE CHANGE: sub/code attributes are no longer supported
197 by the Coro module. It was a mistake to have it in the first place.
198 - (experimental) support for activestate perl 5.10 (method "w").
199 - (experimental) support for strawberry perl (method "a").
200 - coro_sigelem_set did not return a value although it had to,
201 actual impact unknown.
202
203 4.51 Mon Apr 14 13:28:27 CEST 2008
204 - make it compile again on 5.8.
205
206 4.50 Thu Apr 10 09:43:17 CEST 2008
207 - I did it twice! (see 4.49).
208
209 4.49 Sun Apr 6 21:23:31 CEST 2008
210 - grr, instead of compiling the recent changes on 5.10 only they
211 were compiled on 5.8 only.
212
213 4.48 Sun Apr 6 20:36:46 CEST 2008
214 - allow coroutine switches during eval's under 5.10.x, as it seems
215 the parser is a per-interpreter option now, so this is safe
216 (this might fix the odd crashing bug).
217 - drop support for 5.9.x versions: they are dead, jim.
218
219 4.47 Sun Apr 6 00:37:52 CEST 2008
220 - force cctx allocation on API calls: we know we need to force one
221 and gcc actually manages to confuse our heuristic nowadays,
222 leading to crashes and worse.
223 - document force_cctx.
224
225 4.46 Fri Apr 4 22:05:43 CEST 2008
226 - upgrade libcoro, resulting in pthread-backend (which was only created
227 to fulfill the rules of the programming languages shootout).
228
229 4.45 Thu Mar 13 11:55:36 CET 2008
230 - fix a file leaking bug in eg/mhyttpd that would allow
231 downloading of any file (reported by oesi).
232 - fix deadlock bug in Coro::Channel (reported by Richard Hundt)
233 (also add testcase).
234 - support Broadcast option in Coro::Socket (patch by Richard Hundt,
235 apparently having loads of fun with that).
236 - implement $state->swap_defsv, ->swap_defav and document ->throw.
237
238 4.4 Wed Feb 13 16:44:29 CET 2008
239 - only assume .cfi pseudo isns exist on GNU/Linux.
240 - add get_timed method to Coro::Channels.
241 - fixed Coro::Debug manpage.
242 - perl 5.11 compatbility improvement based on patch by
243 Andreas König.
244
245 4.37 Sun Jan 20 11:25:23 CET 2008
246 - soften the check for an active parser for perl 5.10 (avoiding
247 "Coro::State::transfer called while parsing" in many cases).
248
249 4.36 Sun Jan 13 10:53:56 CET 2008
250 - reset diehook when terminating from an async_pool as to not
251 trigger any __DIE__ handlers.
252
253 4.35 Sun Jan 13 04:14:13 CET 2008
254 - "bt" debug command now displays any exceptions
255 from longmess and also skips the innermost
256 stackframes, giving more useufl output.
257 - allow backtraces in coroutines blocked in Coro::EV,
258 at a <1% speed hit.
259 - handle localising of $SIG{__WARN__} and __DIE__
260 properly (with a proper amount of dirty hacking).
261
262 4.34 Sat Dec 22 17:49:53 CET 2007
263 - upgrade to EV version 2.0 API.
264
265 4.33 Mon Dec 17 08:36:12 CET 2007
266 - make Coro::AIO etc. loadable in the absence of EV.
267
268 4.32 Mon Dec 17 07:46:02 CET 2007
269 - majorly improved Coro::Handle's performance with Coro::EV.
270 - implemented the until now mythical Coro::BDB module.
271 - specialcase EV in Coro::AIO and Coro::BDB for extra speed.
272
273 4.31 Wed Dec 5 12:32:39 CET 2007
274 - remove warn statement form Coro::Util.
275
276 4.3 Tue Dec 4 20:33:14 CET 2007
277 - calls to the idle function could cause stack corruption
278 when the stack changed.
279 - do no longer rely on the presence of EV::DNS if EV is used
280 (because it is gone), but instead take avdantage of EV::ADNS
281 if available.
282 - add ($) prototypes to all functions in Coro::Storable.
283 - use a conventional (safer) idle callback in Coro::EV.
284 - do accept quickly in Coro::Debug to avoid endless loops.
285
286 4.22 Fri Nov 30 16:04:04 CET 2007
287 - really use optimised versions for Event and EV in Coro::Util
288 and Coro::Handle.
289
290 4.21 Sun Nov 25 10:48:59 CET 2007
291 - fix a spurious memory read.
292 - Coro::EV no longer keeps the eventloop "alive".
293
294 4.2 Fri Nov 9 20:47:05 CET 2007
295 - enable/disable tracing from a new coroutine, not a pooled one.
296 - fix a memleak in Coro::Event.
297 - removed killall call from fork_eval.
298 - made sure store_fd is already loaded so that fork_eval
299 does not have to parse autoload in each subprocess.
300 - only use assembly method if -O switch is in $Config{optimize}.
301 - add (optional) Coro::EV module, so far the best event loop module
302 directly supported by Coro.
303 - if the event model is EV, use EV::DNS to resolve
304 stuff in Coro::Util.
305 - don't get confused by multiple event notifications in Coro::Handle.
306 - initial support for EV (libevent interface).
307 - require Event and EV using configure_requires, to force their existance.
308
309 4.13 Wed Oct 24 07:26:45 CEST 2007
310 - add Coro::Storable::blocking_thaw.
311 - use a vastly more complicated technique to localise
312 $SIG{__WARN/DIE__} that also works on perls <= 5.8.8.
313 - use a coroutine for the idle callback Coro::Event,
314 instead of running Event in the current coroutine context.
315 This also catches recursive invocations.
316 - actually report fork errors in gethostbyname and inet_aton.
317
318 4.11 Thu Oct 11 02:40:24 CEST 2007
319 - port to threaded perls.
320
321 4.1 Thu Oct 11 02:38:16 CEST 2007
322 - incompatible change: $SIG{__DIE__} and $SIG{__WARN__} will now
323 be local to each coro (see Coro::State).
324 - incompatible change: for very deep reasons, cede and cede_notself
325 cannot return anything, so nothing will be returned.
326 - possibly bring back 5.10 compatibility (untested).
327 - work around stupid (and wrong) warning on 5.10 :(.
328 - overlay the saved state over the context stack. This saves
329 a few hundred bytes per coroutine on average and also
330 speeds up context switching a bit.
331 - further tune default stack sizes.
332 - (more) correctly calculate stack usage in coro_rss.
333 - Coro::Storable::blocking_* did not properly lock
334 resulting in races between coroutines.
335 - added Coro::Storable::guard.
336 - stopping to trace a coroutine could destroy the cctx of
337 an unrelated coroutine.
338 - explain the relationship between Perl and C coroutines in
339 more detail in Coro::State.
340 - Coro::Util::inet_aton did not short-circuit dotted quad forms,
341 causing a fork per resolve. This also affected Coro::Socket.
342 - switch to a separate stack in $coro->call/eval to avoid
343 invalidating pointers.
344
345 4.03 Sat Oct 6 21:24:00 CEST 2007
346 - added Coro::throw method.
347 - minor code cleanups.
348
349 4.02 Sat Oct 6 02:36:47 CEST 2007
350 - fix a very minor per-coroutine memleak (a single codereference).
351 - fixed a bug where the currently in-use c context would be freed
352 prematurely (can happen only when programs change the stacksize
353 or use tracing).
354 - tracing can no longer keep a coro alive after it terminated.
355 - do static branch prediction in the common path for gcc. gives
356 about 2-5% speed improvement here.
357
358 4.01 Fri Oct 5 22:10:49 CEST 2007
359 - instead of recreating *a* standard output handle we simply
360 use STDOUT, which is faster and hopefully more robust.
361
362 4.0 Fri Oct 5 12:56:00 CEST 2007
363 - incompatibly changed Coro::Storable::freeze.
364 - major new feature: added Coro::Debug, for interactive coroutine
365 debugging, tracing and much more.
366 - major bug fix: unbelievable, but true: $_, $/ and many other
367 "saved" variables actually weren't being saved. This has been fixed,
368 of course, while increasing performance while losing all the save
369 flags.
370 - save flags are gone, and all the api functions dealing with them.
371 - added Coro::Semaphore::adjust.
372 - added Coro::Util::fork_eval.
373 - added Coro::Storable::{nfreeze,blocking_{freeze,nfreeze}}.
374 - added Coro::killall.
375 - reduce initial stack sizes to allow for "micro-coroutines".
376 - better async_pool resource management, moved parts of async_pool
377 handling to XS (major speed improvement).
378 - actually croak before modifying important data structures.
379 - refuse to transfer while compiling.
380 - possibly support eval EXPR better now.
381 - enable assembly per default on linux+bsd x86+amd64.
382 - all internal members were renamed _something for easier subclassing.
383 - many minor tweaks.
384
385 3.63 Wed May 16 14:10:06 CEST 2007
386 - implement handcoded assembly for x86/amd64 SVR ABI.
387
388 3.62 Fri Apr 27 21:36:06 CEST 2007
389 - upgrade libcoro (which might set unwind info correctly).
390 - change default on linux to setjmp/longjmp.
391
392 3.61 Thu Apr 19 12:36:18 CEST 2007
393 - Coro::Storable caused an endless loop when thawing invalid
394 pst's sometimes.
395 - use a Semaphore in Coro::Storable, as Storable doesn't
396 seem to be reentrant (although it is documented to
397 be threadsafe...).
398 - fix Coro::Signal to bring back the original unreliable
399 but stateful semantics.
400 - fixed a lot of typos in Coro.pm (patch submitted by David
401 Steinbrunner, which applied flawlessly).
402
403 3.6 Sat Apr 14 17:13:31 CEST 2007
404 - added some bugfixes to get eg/myhttpd working again.
405 - added Coro::Storable for often-cede'ing freeze/thaw.
406 - try to do a clean exit when a coroutine calls exit
407 (EXPERIMENTAL).
408 - got rid of indirect call through _coro_init.
409 - updated the partly antique examples in eg/ to
410 work again and be a bit less magic, too.
411 - fixed Coro::Signal semantics to work as documented again.
412
413 3.55 Sun Mar 25 01:20:47 CET 2007
414 - add SAVE_DEFFH to save the default filehandle and enable
415 it by default.
416 - finally move socket-operations from Coro::Socket to Coro::Handle
417 to be able to unblock foreign sockets.
418 - add Coro::State::save_also and guarded_save.
419 - add count accessor to Coro::Semaphore.
420 - add Coro::State::cctx_stacksize.
421 - just for the fun of it, do not rely on implicit context,
422 which can dramatically improve performance, but people
423 using windows-process-emulation perls do not care much about
424 performance.
425
426 3.51 Sun Mar 4 14:18:04 CET 2007
427 - fixed a problem when you weakened references to Coro::State's
428 (patch by Yuval Kogman).
429
430 3.501 Wed Feb 28 12:44:07 CET 2007
431 - rename some global symbols as macosx from hell redefines
432 them without asking.
433
434 3.5 Tue Feb 13 20:22:53 CET 2007
435 - do AnyEvent model detection earlier, avoiding problems
436 caused by first using AnyEvent and later Coro::Event.
437 - implement and document Coro::Event event objects.
438 - fix a potential problem in Coro::Event causing crashes.
439 - initialise PL_comppad when creating a new coroutine,
440 avoids crashes on early coro destruction.
441
442 3.41 Mon Jan 22 19:19:49 CET 2007
443 - readline on Coro::Handle did not support undefined $/,
444 nor did it deliver partial lines on EOF or error.
445 - implement malloc fallback for stack allocation because
446 stupid broken idiotic OSX has a stupid broken
447 idiotic fits-the-whole-os mmap "implementation" and
448 my dick feels longer if Coro is portable even to
449 obsolete platforms.
450
451 3.4 Fri Jan 19 21:52:54 CET 2007
452 - remove t/09_timer.t, as it isn't really testing much
453 but was rather flaky in practise.
454 - async_pool coro would keep arguments and callback alive until
455 it was reused.
456 - cancellation of a coroutine could cause spurious idle calls
457 in cede_notself.
458
459 3.3 Sat Jan 6 03:45:00 CET 2007
460 - implement $coro->on_destroy.
461 - Coro::Event blocking semantics have been changed,
462 documented and - hopefully - improved.
463 - fix nice adding, not subtracting, from priority.
464 - fix ->prio and api_is_ready (patch by Mark Hinds).
465 - fixed an assert ("... == prev__cctx->idle_te")
466 that could errronously trigger.
467 - fix various large and small memleaks.
468 - use a (hopefully) more stable cancel implementation
469 that immediately frees the coroutine data.
470 - cede/cede_notself return a status now.
471 - added Coro::guard function.
472 - added a global coroutine pool for jobs (on my machine,
473 I can create and execute 48k simple coros/s with async,
474 and 128k coros with async_pool).
475 - Coro::AIO now uses the coroutine priority as io priority.
476
477 3.2 Fri Dec 22 05:07:09 CET 2006
478 - improve portability to slightly older perls.
479 - use cleaner coroutine destruction.
480 - simplify configuration for users.
481 - optionally (unrecommended) prefer perl functions over
482 their coro replacements.
483
484 3.11 Tue Dec 5 13:11:24 CET 2006
485 - fixed some bogus assert's, but as perl.h disables assert even
486 without NDEBUG (thank you very much), not too many people should
487 notice (that did include myself). Andreas König noticed, though :)
488 - do not save/restore PL_sortcxix on >= 5.9.x, it doesn't seem to have
489 it. Also noticed by Andreas König :)
490 - save/restore tainted status.
491 - verified to pass the testsuite on my 5.9.5.
492
493 3.1 Mon Dec 4 23:03:40 CET 2006
494 - INCOMPATIBLE CHANGE: $/ is now per-coroutine (but slow).
495 - incompatible change: transfer flags are now per-state.
496 - give a better error message on deadlock.
497 - document Coro::nready.
498 - enhanced testsuite.
499
500 3.01 Sun Dec 3 23:47:42 CET 2006
501 - forgot to include Coro::Timer.
502
503 3.0 Sun Dec 3 22:57:25 CET 2006
504 - the "FINALLY COMPLETELY STABLE" release (coming soon:
505 the "FAMOUS LAST WORDS" release).
506 - implement a new stack sharing algorithm, which uses a stack
507 pool (size currently hardcoded).
508 - make stack sharing mandatory (it no longer uses a heuristic).
509 - eval/die no longer cause weird problems under heavy use.
510 - Coro::Event could cause livelocks if it was used but
511 no Coro::Event watchers were used.
512 - Coro::Event now uses asynccheck as prepare does not
513 check for changed watchers.
514 - Coro::Event allows multiple waiting coros and will wake up one
515 per event.
516 - Coro::Event should be cleaner and more efficient now.
517 - new utility function Coro::unblock_sub.
518 - document the sad fact that Event is no longer reentrant.
519 - putting a coroutine into the ready queue twice could under
520 some circumstances lead to stack corruption.
521 - minor incompatibility: subclassing Coro::State is supported
522 directly now without going through a _coro_state member.
523 - state/coro switching is much faster now.
524 - very minor optimisations and code/documentation cleanup.
525 - avoid problems due to compiler inlining.
526 - removed timers from Coro::Timer -> use AnyEvent instead.
527 - replaced Coro::idle coroutine by (cleaner) idle handler.
528 - updated to newest libcoro.
529 - implement enhanced support for valgrind.
530 - implement is_ready and return value for ready.
531 - removed Coro::Cont, it was a misnomer (it's generators),
532 and it was somewhat annoying to get it right. Will come back
533 if somebody asks for it :)
534 - many ->wait methods and Coro::Event could return
535 spuriously without the event having happened.
536
537 2.5 Tue Nov 7 12:22:33 CET 2006
538 - made Coro::Util, Coro::Select, Coro::Handle and Coro::Socket
539 use AnyEvent, moved them to Coro/.
540 - added Coro::LWP which contains all the uglyness required to
541 make LWP non-blocking.
542 - should work with perl 5.9.x now (Andreas König made me do it).
543 - fixed another bug in Coro::Select when the timeout was undef.
544 - reuse PL_start_env for all coros, saves some memory per coroutine.
545 - manage PL_top_env differently, hopefully to avoid panic: top_env.
546 - timeout argument was not properly used in Coro::Socket.
547 - allow limited forms of subclassing in Coro::Handle/Coro::Socket.
548 - emulate undocumented(!) functionality of IO::Socket required
549 by LWP(!!).
550 - updated eg/lwp to work with newer lwp's.
551 - remove "FATAL: uncaught exception" prefix. Coroutines that die
552 kill the whole process, just as exceptions in the main "coroutine"
553 did already.
554
555 2.1 Wed Nov 1 23:01:13 CET 2006
556 - fix a long-standing bug in Coro::Select where select with
557 zero timeout would instead change the current default filehandle.
558 - use a simpler and hopefully more robust way to clone padlists
559 (uses less memory and a perl function instead of our own).
560 - coro can now create a stack guard area on many architectures.
561 - Coro::AIO properly reexports additional functions from IO::AIO.
562 - updated libcoro with a workaround for OS X,
563 pach and testing by Michael Schwern.
564
565 2.0 Tue Oct 24 05:47:17 CEST 2006
566 - support additional aio requests in Coro::AIO.
567
568 1.9 (never properly released due to a glitch)
569
570 1.8 Thu Feb 2 00:59:06 CET 2006
571 - applied suggested patch by SAMV to avoid problems during stupid
572 mark & sweep gc run.
573 - applied patch by Scott Walters for 5.9.3 compatibility.
574
575 1.7 Tue Dec 27 01:41:58 CET 2005
576 - added Coro::AIO, a thin wrapper around IO::AIO.
577 - improved Makefile.PL explanations.
578
579 1.6 Mon Dec 12 21:30:05 CET 2005
580 - additionally save PM_curpm might fix as-of-yet
581 unseen problems with regex matches being attributed to the
582 wrong package.
583 - add t/10_bugs.t, which currently checks against imho broken
584 perls that use 0.26MB of stack space per Perl_magic_get
585 invocation (newer linuxes) as opposed to the 0.0002MB perl
586 normally uses.
587 - make stacksize configurable for the ultimate debian experience.
588
589 1.51 Mon Dec 12 18:48:36 CET 2005
590 - remove debugging warn() accidentally left in Coro::Select.
591
592 1.5 Tue Nov 29 13:32:44 CET 2005
593 - use Coro::Event inside Coro::Select to avoid spurious deadlocks.
594 - fix Coro::Select 'select'.
595 - strict'ify some modules.
596 - libcoro errornously restored the SIGUSR2 handler to SIGUSR1.
597 - use XSLoader in selected modules.
598 - remove some 5.6 compatibility cruft.
599
600 1.4 Tue Sep 6 00:11:05 CEST 2005
601 - libcoro did not take into account the trampoline on amd64,
602 when 'l'inux method was used.
603
604 1.31 Tue Aug 30 23:31:33 CEST 2005
605 - some portability fixes/workarounds.
606
607 1.3 Sat Aug 20 03:08:56 CEST 2005
608 - no code changes. module seems to work fine.
609 - improve Coro::State docs, remove reference to nonexisting
610 Coro::State::flush method.
611 - no longer autodetect windows, present it as an option instead.
612
613 1.2 Mon May 16 02:00:55 CEST 2005
614 - included libcoro.c earlier in State.xs, might improve portability.
615 - use a faster and possibly more stable Coro::Cont implementation.
616 - accept x86_64 in addition to amd64 for optimized linux-amd64 support.
617 - fix bugs Coro::SemaphoreSet that could cause locks never to be freed.
618 - fix bugs in CoroAPI.h and document it in Coro::MakeMaker.
619
620 1.11 Thu Mar 3 18:00:52 CET 2005
621 - change of contact address.
622
623 1.1 Tue Feb 22 20:51:16 CET 2005
624 - support [l]inux method on amd64.
625 - allow some context switches while compiling/eval'ing.
626
627 1.0 Tue Aug 10 03:47:27 CEST 2004
628 - Event 0.89 fixes the bug again.
629 - fix the remaining(??) memleak. I didn't free the cache-AV for
630 padlists, so one AV per created anon-sub leaked.
631
632 0.97 Fri May 14 15:23:32 CEST 2004
633 - remove stupid gcc extension that was an experiment only anyway.
634
635 0.96 Thu May 13 18:09:29 CEST 2004
636 - tell people if they have yet another broken Event variant.
637 - made Coro to work _better_ with threads (it seems that, as long
638 as you use Coros in one thread only everything will work fine.
639 Looking at the perl source the problem might be missing locking
640 between Coro and some perl internal routines. That's not easily
641 fixable).
642 - (hopefully) got rid of the static xnext inside transfer.
643
644 0.95 Sun Feb 15 03:20:28 CET 2004
645 - removed Coro::State::flush, implemented a new and much
646 faster caching mechanism by attaching magic to the CV.
647 - a memleak with (real) closures remains.
648 - dramatically reduced size of stacks to 65536 (linux/x86), after
649 observing that even myhttpd never had more than a kilobyte
650 of stack in use (stack is still only physically allocated in
651 pagesize increments for systems suporting mmap).
652 - fix for a "cygwin-64" architecture added, proposed by Stefan Traby.
653 - fix for threaded perl.
654
655 0.9 Sun Nov 23 04:32:39 CET 2003
656 - fix a few very hard to track down but (of course) stupid
657 memory leaks.
658 - stringify version number to avoid locale problems :(.
659 - perl 5.9.0 does not have SvPADBUSY, reported by Scott Walters,
660 hopefully ignoring it "just works".
661
662 0.8 Wed Nov 5 19:38:40 CET 2003
663 - port to perl5.005_03, but only 5.8.x+ is supported!
664 - honor LocalAddr even without LocalPort in Coro::Socket.
665
666 0.7 Tue May 27 03:12:38 CEST 2003
667 - the version jump indicates some level of testing, not gobs
668 of new features.
669 - uh... I found the prompt function in ExtUtils::MakeMaker.
670 Highly correct stuff, that is...
671 - fixed(?) a bug with die's in coroutines causing "panic: top_env".
672 the fix is not well understood by the author, so beware :(.
673
674 0.652 Thu May 8 02:54:46 CEST 2003
675 - Applied patch by Slaven Rezic to set default to "s" on FreeBSD,
676 cause version 4 doesn't have ucontext.
677 - Benjamin Reed reported that setjmp works fine on darwin,
678 so preselect it.
679
680 0.651 Sat Mar 29 15:00:23 CET 2003
681 - fix a bug in Coro::Handle where some bytes could get lost
682 on reading, reported by jason@nichego.net.
683
684 0.65 Sun Mar 23 00:08:26 CET 2003
685 - added a README.linux-glibc.
686 - new module Coro::Select.
687 - also save/restore PL_comppad, fixes segfaults with 5.8.1.
688 (I never claimed that I know what I am doing ;).
689 - set default to setjmp/longjmp on non-x86-linux.
690
691 0.6 Thu Nov 21 11:09:06 CET 2002
692 - made lazy context switching the default.
693 - fixed the bug where SemaphoreSet::guard
694 was the same as timed_guard and timed_guard was missing.
695 - fix a memleak in Coro::Timer.
696 - reclassified context sharing as not experimental.
697
698 0.534 Sun Apr 14 03:05:12 CEST 2002
699 - fixed backspace => tab in header parsing (myhttpd).
700 - added eg/readline, for elmex.
701 - Coro::Event now calls ->start on first call instead of
702 ->again, which created wrong timeouts for the first ->next.
703 - fixed a bug where Coro::Socket returned a closed socket instead
704 of nothing on ECONNREFUSED and similar error conditions.
705
706 0.533 Mon Feb 18 18:49:40 CET 2002
707 - adapt to finally working Event-0.85.
708 - fixed buggy event initialization (trapped under windows, but how
709 could it possibly have worked anywhere?)
710
711 0.532 Wed Jan 16 02:45:32 CET 2002
712 - removed the reference to pp_entersub, might work on
713 (some) win32 perls now (testsuite works under cygwin,
714 without Event, which freezes).
715
716 0.531 Mon Dec 10 22:18:44 CET 2001
717 - Fixed a bug in SemaphoreSet::timed_down.
718
719 0.53 Tue Nov 27 21:11:13 CET 2001
720 - some tests for Event.
721 - slightly different internal architecture to get refcnt into the object
722 and not the reference. might fix a few bugs, certainly did introduce
723 new ones.
724 - Coro::Timer (independent of Event).
725 - new timed_wait functions for Coro::Signal, Semaphore, SemaphoreSet.
726
727 0.52 Tue Nov 6 21:36:18 CET 2001
728 - ported to cygwin (trivial).
729 patch by Gerrit P. Haase <gp@familiehaase.de>.
730 - small setjmp code fix by Sullivan.DanielJ@epamail.epa.gov.
731
732 0.51 Thu Nov 1 20:39:01 CET 2001
733 - terminate/cancel now work properly (otherwise termination
734 could cause a "next coroutine is not and contains not..."
735 error.
736 - added Coro::Socket::shutdown.
737 - Coro::Event::loop no is the same as Event::loop.
738 - implemented terminate with args + join.
739
740 0.5 Fri Sep 28 16:15:35 CEST 2001
741 - fixed "print" on a Coro::Handle. The print method worked.
742 - small tweaks (seem to reduce memory consumption a lot)
743 in various modules.
744 - splendid use of "no warnings" scattered throughout.
745 - added Coro::Handle::rbuf, fixed Coro::Handle::fh.
746
747 0.49 Sun Sep 16 02:42:45 CEST 2001
748 - changed some method calls to function calls for speed
749 inside Coro::Handle.
750 - make Coro::Handle use an array instead of a hash for
751 speed reasons.
752 - IRIX mystery solved: it's SGI's NT, after all: "standard, huh?".
753 sjlj and ucontext should now work.
754 - IRIX-specific port for libcoro.
755 - swapped order of accept results to match IO::Socket.
756 - changed getsock/peername to sock/peername to match IO::Socket.
757 - fixed a bug that caused segfault when returning to main task
758 under some circumstances.
759 - other bugfixes.
760
761 0.45 Sun Sep 2 02:54:01 CEST 2001
762 - new method Coro::Handle::timeout.
763 - corrected speling of set...name to get...name in Coro::Socket.
764 - Coro::Socket::accept now returns a Coro::Socket, not a Coro::Handle.
765 - Coro::Handle now supports fileno and FILENO.
766 - added eg/myhttpd, a web server.
767 - Coro::Socket now accepts numeric ports as well as "port(number)"
768 syntax.
769 - moved some scheduling primitives into xs code. more to come.
770 - new simple priority system.
771 - implemented Coro::Event using XS, almost four times faster!
772 - small memory corruption problem fixed (boy that was difficult).
773
774 0.13 Wed Aug 8 16:53:07 CEST 2001
775 - new method Coro::Handle::readline.
776 - added eg/lwp and Coro::Handle::autoflush, to make LWP work.
777 (see eg/lwp on how to make LWP non-blocking).
778 - renamed Coro::Socket::new_inet to new.
779 - added Coro::Util, some utility functions.
780 - die/eval now works better (still get annoying (but true ;)
781 "Callback called exit" messages).
782 - Coro::Handle now supports timeout, and Coro::Socket Timeout.
783 - much better Coro::Socket support.
784
785 0.12 Fri Jul 27 04:19:01 CEST 2001
786 - do not use mmap if neither MAP_ANON nor MAP_ANONYMOUS is defined.
787 - very experimental stack sharing algorithm. not 100% safe but
788 should work well in practise ;)
789 - added Coro::RWLock.
790 - Coro::Specific now works.
791
792 0.11 Tue Jul 24 22:49:21 CEST 2001
793 - added specialized hack for newer and older linux versions (fast).
794 - renamed Coro::Event::IO to Coro::Handle.
795 - new module Coro::Socket.
796
797 0.1 Tue Jul 24 01:47:53 CEST 2001
798 - release candidate 3. A rather new internal structure :(
799 - the great renaming: Damian Conway gave me a suitable replacement
800 for yield.
801 - added Coro/libcoro, a portable coroutine implementation for C
802 (not even perl-dependent!!!), which can be used by Coro::State.
803 - renamed SAVE_DEFSV (on xs-level only) to avoid symbol clash in
804 perl-5.6.
805 - new function Coro::Event::idle.
806 - the idle process is now overriden
807 by default in Coro::Event.
808 - Coro::Channel now enforces the size.
809 - canceling events now works in all cases.
810 - Coro state now include $_ and $@.
811 - yet another bug workaround that I do not really understand :(
812 - new module Coro::Event::IO (very undocumented) to do
813 non-blocking i/o.
814 - performance tuning.
815
816 0.08 Thu Jul 19 06:13:25 CEST 2001
817 - release candidate 2 (new functionality).
818 - add Coro::State::flush function.
819 - transfer now supports a flags value (mostly for speed ATM).
820 - might compile and work in the presence of threads now.
821 - continuations have a saner syntax.
822 - no more memleaks.
823
824 0.07 Tue Jul 17 17:40:18 CEST 2001
825 - release candidate 1 ;)
826 - slightly nicer code.
827 - fixed a scheduling bug in Coro::Event.
828
829 0.06 Tue Jul 17 04:23:24 CEST 2001
830 - ok, I found the showstopper - the same sub must not be
831 re-entered in two different coroutines, otherwise => crash. I
832 see no easy solution to this problem, except by walking the call
833 chaing and saving/restoring the cv's, which is what I do now.
834 - memory leaks still latent, especially at thread termination.
835 - Coro::Event now works (haha).
836
837 0.05 Sun Jul 15 17:32:20 CEST 2001
838 - fixed a few issues in Coro::Event.
839 - I forgot to include Coro::Event in 0.04 :(:(:(
840
841 0.04 Sun Jul 15 05:24:59 CEST 2001
842 - @_ is now properly localized.
843 - Coro::State is now easier subclassable.
844 - Coro::Cont now coroutine-aware.
845 - Coro::Specific is a low-overhead module to create
846 coroutine-specific vars.
847 - Coro::Event provides a simple interface to Event.
848
849 0.03 Fri Jul 13 14:51:52 CEST 2001
850 - transfer() now implemented in XS (beware).
851 - new module Coro::Cont for really faked continuations.
852 - big internal architecture changes: Coro::State is now
853 really low-level and can thus be used to implement other
854 interesting things, While "Coro::" implements a process-like
855 model. Still crude and subject to change.
856 - $_ and $@ are no longer being localized.
857
858 0.02 Tue Jul 10 01:38:17 CEST 2001
859 - implemented "async" attribute.
860 - $_ and $@ are now localized.
861 - added Coro::Channel.
862 - more testcases, still no docs.
863
864 0.01 Tue Jul 3 02:18:41 CEST 2001
865 - original version; copied from Convert::Scalar.
866