ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/Coro/Changes
Revision: 1.421
Committed: Sun Nov 16 10:12:38 2008 UTC (15 years, 6 months ago) by root
Branch: MAIN
Changes since 1.420: +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: aio
10 TODO: Coro::Semaphore
11 TODO: trow disturbs cleanup path inside pp_slf?
12 5.0
13 - NEW ARCHITECTURE: use 4.912 for the stable version.
14 - bump API version to 7 - all dependents must be recompiled.
15 - Coro::Semaphore->adjust didn't correctly wake up enough waiters.
16 - async_pool did free a scalar value twice
17 ("Attempt to unreference...").
18 - fix a longstanding bug where calling terminate on a coro that
19 was waiting for a semaphore that was just becoming available
20 would cause a deadlock (semaphore would get into a state where
21 it was available but waiters were waiting(.
22 - perl's process emulation is now not even theoretically supported
23 anymore.
24 - directly patch the entersub opcode calling SLF functions (cede,
25 transfer and so on). this does speed up context switching, but
26 more importanly, it frees us from the hardcoded behaviour of
27 entersub, so we might actually be able to return something from
28 those functions and atcually create new ones.
29 - take advantage of __builtin_frame_address on gcc.
30 - expose THX in coroapi (not sure whether this was a wise decision,
31 as "threaded" perls are running at half speed anyways).
32 - implement execute_slf (schedule-like-function) interface that makes
33 it possible to implement schedule-like-functions in XS.
34 - use new SLF interface to massively speed up Coro::EV by roughly a factor
35 of two.
36 - removed timed_down and timed_guard functions - they were not being used
37 anyways and should be replaced by a more generic mechanism - and were
38 annoying to support anyways :
39 - used new SLF interface to massively speed up Coro::Semaphore by a factor
40 of three.
41 - forgot to include Coro/libcoro/README in the dist for all these years.
42 - work around a freebsd pthreads bug (manual testcancel is required as
43 pthread_cond_wait isn't a cancellation point on freebsd).
44
45 4.913 Sat Nov 15 07:58:28 CET 2008
46 - async_pool did free a scalar value twice
47 ("Attempt to unreference...").
48
49 4.912 Thu Nov 13 18:31:23 CET 2008
50 - minor cleanups.
51 - use much larger stacks on linux and perl < 5.8.8.
52 - Coro::Debug::new_unix_server did not unlink the socket
53 when destroyed.
54
55 4.911 Tue Nov 11 04:26:17 CET 2008
56 - "port" to threaded perls.
57
58 4.91 Mon Nov 10 05:36:38 CET 2008
59 - the ->throw exception object no longer leaks.
60 - creating a new cctx leaked a scopestack entry (memleak).
61 - new coroutines didn't get created with a zero flags field
62 (unknown impact).
63 - calling ->throw on a not-yet-started coroutine should now work
64 instead of being ignored.
65 - ->throw is now supported on Coro::State objects.
66 - clean up cctx creation code a bit.
67 - entersub is actually a UNOP, not a LOGOP (not a bugfix).
68
69 4.9 Sat Nov 8 17:45:27 CET 2008
70 - (libcoro) did not preserve rbp with CORO_ASM (we are getting there).
71 - (libcoro) no longer leak threads in the experimental pthread backend,
72 also speed it up considerably.
73 - (libcoro) do not rely on makecontext passing void *'s unscathed.
74 - fix compiletime dependencies on libcoro in the Makefile.
75 - cctx_count wasn't always updated properly.
76 - Coro::State::cctx_stacksize wasn't applied correctly.
77 - new function Coro::State::cctx_max_idle.
78 - the default max number of idle C contexts is now 4.
79 - (libcoro) try harder to get _setjmp/_longjmp.
80 - (libcoro) cleanup and extend the libcoro API to officially
81 allow the creation of empty source contexts.
82 - very experimental workaround for the totally broken netbsd platform.
83 - tried to hack around openbsd bugs.
84
85 4.804 Wed Nov 5 16:36:00 CET 2008
86 - Coro::Debug::new_unix_server would not create a non-blocking listening
87 socket, sometimes causing freezes.
88 - (libcoro) fix misaligned stack points for setjmp and assembly
89 methods, which can cause crashes on x86/x86_64 with a sufficiently
90 aggressive compiler.
91 - new function: Coro::Debug::new_tcp_server.
92 - move ->throw into the Coro class because it only works on coro objects.
93
94 4.803 Mon Nov 3 17:16:12 CET 2008
95 - (libcoro) use a global asm statement to become independent of gcc
96 otpimisations for CORO_ASM (thanks to pippijn for the idea).
97 - try to workaround yet another broken bsd, this time dragonfly.
98
99 4.802 Thu Oct 30 10:56:12 CET 2008
100 - support -fno-omit-frame-pointer on x86 with the assembly method.
101 - tune 01_unblock.t tests a bit.
102
103 4.801 Wed Oct 22 18:33:37 CEST 2008
104 - improve readline speed for very long "lines".
105 - backport to 5.8.8.
106
107 4.8 Thu Oct 2 13:34:40 CEST 2008
108 - new function Coro::AIO::aio_wait.
109 - Coro.:AIO and Coro::BDB now "use Coro::AnyEvent".
110 - greatly speed up and reduce memory usage of Coro::AIO requests.
111 - implement some other µ-optimisations.
112
113 4.749 Mon Sep 29 14:40:12 CEST 2008
114 - port to slow and broken pseudo-threaded perls. (courtesy pippijn).
115
116 4.748 Sat Sep 27 14:03:19 CEST 2008
117 - implement, but do not document, PerlIO::cede(granularity).
118 - Coro::Storable forgot to wrap Storable::pstore.
119 - work around the multitude of leaks and memory corruption
120 bugs in PerlIO::via by using our own C-level perliol. As a side
121 effect, Coro::Storable is now much, much, much faster.
122
123 4.747 Tue Sep 23 01:59:41 CEST 2008
124 - fix a per-cv memleak (one empty array was leaked per
125 code reference).
126 - avoid a crash in coro->call|rss when the coroutine was already
127 destroyed (most noticably when using Coro::Debug::ps :)
128 - also protect *Storable::FILE.
129 - push up default storable granularity to 20ms.
130
131 4.746 Sun Sep 21 03:22:20 CEST 2008
132 - be more insistent on locking Storable against reentrancy
133 in Coro::Storable.
134 - move swap_def?v and throw to Coro::State, as documented.
135
136 4.745 Thu Jul 24 00:14:38 CEST 2008
137 - remove debugging code related to MgPV_nolen_const, also try to
138 make it compile with perl 5.8.6 (yes, apple apparently loves
139 outdated software). Reported by John S.
140
141 4.744 Tue Jul 8 22:06:35 CEST 2008
142 - correctly provide default DIE/WARN handlers as documented.
143 - also overwrite PL_vtbl_sigelem.svt_clear, even though current
144 implementations inside perl work fine for us.
145
146 4.743 Mon Jun 16 00:21:57 CEST 2008
147 - when using Coro::EV without running EV::loop it could
148 result in busy-waiting for events, this has been fixed.
149 - reduce codesize and improve performance by using EV_DEFAULT_UC.
150
151 4.742 Sat May 31 14:10:21 CEST 2008
152 - implement a workaround for (some) perl <5.8.8 versions.
153 - require EV 3.3+.
154
155 4.741 Fri May 30 23:33:09 CEST 2008
156 - tell netbsd how utterly broken their imitation of an OS is
157 and refuse to build by default if pthreads are in use.
158 - switch to "s" method on all bsd's by default, as their ucontext
159 stuff seems just too broken.
160 - fix a bug in Coro::Select.
161
162 4.74 Thu May 29 20:05:31 CEST 2008
163 - do not test Coro::LWP for lack of dependencies.
164 - do not test Coro::Select for lack of working perls.
165
166 4.73 Thu May 29 2008
167 - fix a bug in Coro::EV which would cause it to block despite
168 there being runnable coroutines.
169 - sprinkle "no warnings" freely over everything, also suppress
170 warnings for some other modules.
171 - fix typo in WSAEWOULDBLOCK.
172
173 4.72 Sun May 25 05:14:36 CEST 2008
174 - tweak META.yaml a bit, unfortunately, there is no documented way
175 to have optional dependencies with CPAN. doh :(
176 - avoid running some tests on windows because they would fail due to
177 perl bug (broken fork, broken pipes...).
178 - work around perl on windows bugs where perl returns undocumented
179 error codes for sysread, syswrite etc. by taking advantage
180 of AnyEvent's workaround for that problem.
181 - use AnyEvent::Util::fh_nonblocking in Coro::Handle to work around
182 a common perl implementation bug on windows.
183 - use unix domain sockets in testsuite to work around
184 common perl implementation bugs on widows (they are emulated by
185 tcp sockets on windows. ugh.)
186
187 4.71 Sat May 24 20:01:27 CEST 2008
188 - fix a bug in Coro::AnyEvent ("Usage: Coro::AnyEvent::_schedule()").
189 - take advantage of async name resolution of AnyEvent::Util.
190 - work around brutal inet_aton override in Coro::LWP.
191 - take advantage of the readyhook in Coro::EV, for smoother
192 scheduling.
193
194 4.7 Sun May 11 00:32:19 CEST 2008
195 - completely reworked the Coro manpage.
196 - added Coro::AnyEvent, generic event loop integration.
197 - implement cancel, ready and kill commands in Coro::Debug.
198 - document find_coro in Coro::Debug.
199 - incompatible change: rename has_stack to has_cctx.
200 - Coro::AIO and Coro::BDB no longer force event model detection,
201 use AnyEvent::AIO and AnyEvent::BDB.
202
203 4.6 Sat Apr 26 10:05:14 CEST 2008
204 - INCOMPATIBLE CHANGE: sub/code attributes are no longer supported
205 by the Coro module. It was a mistake to have it in the first place.
206 - (experimental) support for activestate perl 5.10 (method "w").
207 - (experimental) support for strawberry perl (method "a").
208 - coro_sigelem_set did not return a value although it had to,
209 actual impact unknown.
210
211 4.51 Mon Apr 14 13:28:27 CEST 2008
212 - make it compile again on 5.8.
213
214 4.50 Thu Apr 10 09:43:17 CEST 2008
215 - I did it twice! (see 4.49).
216
217 4.49 Sun Apr 6 21:23:31 CEST 2008
218 - grr, instead of compiling the recent changes on 5.10 only they
219 were compiled on 5.8 only.
220
221 4.48 Sun Apr 6 20:36:46 CEST 2008
222 - allow coroutine switches during eval's under 5.10.x, as it seems
223 the parser is a per-interpreter option now, so this is safe
224 (this might fix the odd crashing bug).
225 - drop support for 5.9.x versions: they are dead, jim.
226
227 4.47 Sun Apr 6 00:37:52 CEST 2008
228 - force cctx allocation on API calls: we know we need to force one
229 and gcc actually manages to confuse our heuristic nowadays,
230 leading to crashes and worse.
231 - document force_cctx.
232
233 4.46 Fri Apr 4 22:05:43 CEST 2008
234 - upgrade libcoro, resulting in pthread-backend (which was only created
235 to fulfill the rules of the programming languages shootout).
236
237 4.45 Thu Mar 13 11:55:36 CET 2008
238 - fix a file leaking bug in eg/mhyttpd that would allow
239 downloading of any file (reported by oesi).
240 - fix deadlock bug in Coro::Channel (reported by Richard Hundt)
241 (also add testcase).
242 - support Broadcast option in Coro::Socket (patch by Richard Hundt,
243 apparently having loads of fun with that).
244 - implement $state->swap_defsv, ->swap_defav and document ->throw.
245
246 4.4 Wed Feb 13 16:44:29 CET 2008
247 - only assume .cfi pseudo isns exist on GNU/Linux.
248 - add get_timed method to Coro::Channels.
249 - fixed Coro::Debug manpage.
250 - perl 5.11 compatbility improvement based on patch by
251 Andreas König.
252
253 4.37 Sun Jan 20 11:25:23 CET 2008
254 - soften the check for an active parser for perl 5.10 (avoiding
255 "Coro::State::transfer called while parsing" in many cases).
256
257 4.36 Sun Jan 13 10:53:56 CET 2008
258 - reset diehook when terminating from an async_pool as to not
259 trigger any __DIE__ handlers.
260
261 4.35 Sun Jan 13 04:14:13 CET 2008
262 - "bt" debug command now displays any exceptions
263 from longmess and also skips the innermost
264 stackframes, giving more useufl output.
265 - allow backtraces in coroutines blocked in Coro::EV,
266 at a <1% speed hit.
267 - handle localising of $SIG{__WARN__} and __DIE__
268 properly (with a proper amount of dirty hacking).
269
270 4.34 Sat Dec 22 17:49:53 CET 2007
271 - upgrade to EV version 2.0 API.
272
273 4.33 Mon Dec 17 08:36:12 CET 2007
274 - make Coro::AIO etc. loadable in the absence of EV.
275
276 4.32 Mon Dec 17 07:46:02 CET 2007
277 - majorly improved Coro::Handle's performance with Coro::EV.
278 - implemented the until now mythical Coro::BDB module.
279 - specialcase EV in Coro::AIO and Coro::BDB for extra speed.
280
281 4.31 Wed Dec 5 12:32:39 CET 2007
282 - remove warn statement form Coro::Util.
283
284 4.3 Tue Dec 4 20:33:14 CET 2007
285 - calls to the idle function could cause stack corruption
286 when the stack changed.
287 - do no longer rely on the presence of EV::DNS if EV is used
288 (because it is gone), but instead take avdantage of EV::ADNS
289 if available.
290 - add ($) prototypes to all functions in Coro::Storable.
291 - use a conventional (safer) idle callback in Coro::EV.
292 - do accept quickly in Coro::Debug to avoid endless loops.
293
294 4.22 Fri Nov 30 16:04:04 CET 2007
295 - really use optimised versions for Event and EV in Coro::Util
296 and Coro::Handle.
297
298 4.21 Sun Nov 25 10:48:59 CET 2007
299 - fix a spurious memory read.
300 - Coro::EV no longer keeps the eventloop "alive".
301
302 4.2 Fri Nov 9 20:47:05 CET 2007
303 - enable/disable tracing from a new coroutine, not a pooled one.
304 - fix a memleak in Coro::Event.
305 - removed killall call from fork_eval.
306 - made sure store_fd is already loaded so that fork_eval
307 does not have to parse autoload in each subprocess.
308 - only use assembly method if -O switch is in $Config{optimize}.
309 - add (optional) Coro::EV module, so far the best event loop module
310 directly supported by Coro.
311 - if the event model is EV, use EV::DNS to resolve
312 stuff in Coro::Util.
313 - don't get confused by multiple event notifications in Coro::Handle.
314 - initial support for EV (libevent interface).
315 - require Event and EV using configure_requires, to force their existance.
316
317 4.13 Wed Oct 24 07:26:45 CEST 2007
318 - add Coro::Storable::blocking_thaw.
319 - use a vastly more complicated technique to localise
320 $SIG{__WARN/DIE__} that also works on perls <= 5.8.8.
321 - use a coroutine for the idle callback Coro::Event,
322 instead of running Event in the current coroutine context.
323 This also catches recursive invocations.
324 - actually report fork errors in gethostbyname and inet_aton.
325
326 4.11 Thu Oct 11 02:40:24 CEST 2007
327 - port to threaded perls.
328
329 4.1 Thu Oct 11 02:38:16 CEST 2007
330 - incompatible change: $SIG{__DIE__} and $SIG{__WARN__} will now
331 be local to each coro (see Coro::State).
332 - incompatible change: for very deep reasons, cede and cede_notself
333 cannot return anything, so nothing will be returned.
334 - possibly bring back 5.10 compatibility (untested).
335 - work around stupid (and wrong) warning on 5.10 :(.
336 - overlay the saved state over the context stack. This saves
337 a few hundred bytes per coroutine on average and also
338 speeds up context switching a bit.
339 - further tune default stack sizes.
340 - (more) correctly calculate stack usage in coro_rss.
341 - Coro::Storable::blocking_* did not properly lock
342 resulting in races between coroutines.
343 - added Coro::Storable::guard.
344 - stopping to trace a coroutine could destroy the cctx of
345 an unrelated coroutine.
346 - explain the relationship between Perl and C coroutines in
347 more detail in Coro::State.
348 - Coro::Util::inet_aton did not short-circuit dotted quad forms,
349 causing a fork per resolve. This also affected Coro::Socket.
350 - switch to a separate stack in $coro->call/eval to avoid
351 invalidating pointers.
352
353 4.03 Sat Oct 6 21:24:00 CEST 2007
354 - added Coro::throw method.
355 - minor code cleanups.
356
357 4.02 Sat Oct 6 02:36:47 CEST 2007
358 - fix a very minor per-coroutine memleak (a single codereference).
359 - fixed a bug where the currently in-use c context would be freed
360 prematurely (can happen only when programs change the stacksize
361 or use tracing).
362 - tracing can no longer keep a coro alive after it terminated.
363 - do static branch prediction in the common path for gcc. gives
364 about 2-5% speed improvement here.
365
366 4.01 Fri Oct 5 22:10:49 CEST 2007
367 - instead of recreating *a* standard output handle we simply
368 use STDOUT, which is faster and hopefully more robust.
369
370 4.0 Fri Oct 5 12:56:00 CEST 2007
371 - incompatibly changed Coro::Storable::freeze.
372 - major new feature: added Coro::Debug, for interactive coroutine
373 debugging, tracing and much more.
374 - major bug fix: unbelievable, but true: $_, $/ and many other
375 "saved" variables actually weren't being saved. This has been fixed,
376 of course, while increasing performance while losing all the save
377 flags.
378 - save flags are gone, and all the api functions dealing with them.
379 - added Coro::Semaphore::adjust.
380 - added Coro::Util::fork_eval.
381 - added Coro::Storable::{nfreeze,blocking_{freeze,nfreeze}}.
382 - added Coro::killall.
383 - reduce initial stack sizes to allow for "micro-coroutines".
384 - better async_pool resource management, moved parts of async_pool
385 handling to XS (major speed improvement).
386 - actually croak before modifying important data structures.
387 - refuse to transfer while compiling.
388 - possibly support eval EXPR better now.
389 - enable assembly per default on linux+bsd x86+amd64.
390 - all internal members were renamed _something for easier subclassing.
391 - many minor tweaks.
392
393 3.63 Wed May 16 14:10:06 CEST 2007
394 - implement handcoded assembly for x86/amd64 SVR ABI.
395
396 3.62 Fri Apr 27 21:36:06 CEST 2007
397 - upgrade libcoro (which might set unwind info correctly).
398 - change default on linux to setjmp/longjmp.
399
400 3.61 Thu Apr 19 12:36:18 CEST 2007
401 - Coro::Storable caused an endless loop when thawing invalid
402 pst's sometimes.
403 - use a Semaphore in Coro::Storable, as Storable doesn't
404 seem to be reentrant (although it is documented to
405 be threadsafe...).
406 - fix Coro::Signal to bring back the original unreliable
407 but stateful semantics.
408 - fixed a lot of typos in Coro.pm (patch submitted by David
409 Steinbrunner, which applied flawlessly).
410
411 3.6 Sat Apr 14 17:13:31 CEST 2007
412 - added some bugfixes to get eg/myhttpd working again.
413 - added Coro::Storable for often-cede'ing freeze/thaw.
414 - try to do a clean exit when a coroutine calls exit
415 (EXPERIMENTAL).
416 - got rid of indirect call through _coro_init.
417 - updated the partly antique examples in eg/ to
418 work again and be a bit less magic, too.
419 - fixed Coro::Signal semantics to work as documented again.
420
421 3.55 Sun Mar 25 01:20:47 CET 2007
422 - add SAVE_DEFFH to save the default filehandle and enable
423 it by default.
424 - finally move socket-operations from Coro::Socket to Coro::Handle
425 to be able to unblock foreign sockets.
426 - add Coro::State::save_also and guarded_save.
427 - add count accessor to Coro::Semaphore.
428 - add Coro::State::cctx_stacksize.
429 - just for the fun of it, do not rely on implicit context,
430 which can dramatically improve performance, but people
431 using windows-process-emulation perls do not care much about
432 performance.
433
434 3.51 Sun Mar 4 14:18:04 CET 2007
435 - fixed a problem when you weakened references to Coro::State's
436 (patch by Yuval Kogman).
437
438 3.501 Wed Feb 28 12:44:07 CET 2007
439 - rename some global symbols as macosx from hell redefines
440 them without asking.
441
442 3.5 Tue Feb 13 20:22:53 CET 2007
443 - do AnyEvent model detection earlier, avoiding problems
444 caused by first using AnyEvent and later Coro::Event.
445 - implement and document Coro::Event event objects.
446 - fix a potential problem in Coro::Event causing crashes.
447 - initialise PL_comppad when creating a new coroutine,
448 avoids crashes on early coro destruction.
449
450 3.41 Mon Jan 22 19:19:49 CET 2007
451 - readline on Coro::Handle did not support undefined $/,
452 nor did it deliver partial lines on EOF or error.
453 - implement malloc fallback for stack allocation because
454 stupid broken idiotic OSX has a stupid broken
455 idiotic fits-the-whole-os mmap "implementation" and
456 my dick feels longer if Coro is portable even to
457 obsolete platforms.
458
459 3.4 Fri Jan 19 21:52:54 CET 2007
460 - remove t/09_timer.t, as it isn't really testing much
461 but was rather flaky in practise.
462 - async_pool coro would keep arguments and callback alive until
463 it was reused.
464 - cancellation of a coroutine could cause spurious idle calls
465 in cede_notself.
466
467 3.3 Sat Jan 6 03:45:00 CET 2007
468 - implement $coro->on_destroy.
469 - Coro::Event blocking semantics have been changed,
470 documented and - hopefully - improved.
471 - fix nice adding, not subtracting, from priority.
472 - fix ->prio and api_is_ready (patch by Mark Hinds).
473 - fixed an assert ("... == prev__cctx->idle_te")
474 that could errronously trigger.
475 - fix various large and small memleaks.
476 - use a (hopefully) more stable cancel implementation
477 that immediately frees the coroutine data.
478 - cede/cede_notself return a status now.
479 - added Coro::guard function.
480 - added a global coroutine pool for jobs (on my machine,
481 I can create and execute 48k simple coros/s with async,
482 and 128k coros with async_pool).
483 - Coro::AIO now uses the coroutine priority as io priority.
484
485 3.2 Fri Dec 22 05:07:09 CET 2006
486 - improve portability to slightly older perls.
487 - use cleaner coroutine destruction.
488 - simplify configuration for users.
489 - optionally (unrecommended) prefer perl functions over
490 their coro replacements.
491
492 3.11 Tue Dec 5 13:11:24 CET 2006
493 - fixed some bogus assert's, but as perl.h disables assert even
494 without NDEBUG (thank you very much), not too many people should
495 notice (that did include myself). Andreas König noticed, though :)
496 - do not save/restore PL_sortcxix on >= 5.9.x, it doesn't seem to have
497 it. Also noticed by Andreas König :)
498 - save/restore tainted status.
499 - verified to pass the testsuite on my 5.9.5.
500
501 3.1 Mon Dec 4 23:03:40 CET 2006
502 - INCOMPATIBLE CHANGE: $/ is now per-coroutine (but slow).
503 - incompatible change: transfer flags are now per-state.
504 - give a better error message on deadlock.
505 - document Coro::nready.
506 - enhanced testsuite.
507
508 3.01 Sun Dec 3 23:47:42 CET 2006
509 - forgot to include Coro::Timer.
510
511 3.0 Sun Dec 3 22:57:25 CET 2006
512 - the "FINALLY COMPLETELY STABLE" release (coming soon:
513 the "FAMOUS LAST WORDS" release).
514 - implement a new stack sharing algorithm, which uses a stack
515 pool (size currently hardcoded).
516 - make stack sharing mandatory (it no longer uses a heuristic).
517 - eval/die no longer cause weird problems under heavy use.
518 - Coro::Event could cause livelocks if it was used but
519 no Coro::Event watchers were used.
520 - Coro::Event now uses asynccheck as prepare does not
521 check for changed watchers.
522 - Coro::Event allows multiple waiting coros and will wake up one
523 per event.
524 - Coro::Event should be cleaner and more efficient now.
525 - new utility function Coro::unblock_sub.
526 - document the sad fact that Event is no longer reentrant.
527 - putting a coroutine into the ready queue twice could under
528 some circumstances lead to stack corruption.
529 - minor incompatibility: subclassing Coro::State is supported
530 directly now without going through a _coro_state member.
531 - state/coro switching is much faster now.
532 - very minor optimisations and code/documentation cleanup.
533 - avoid problems due to compiler inlining.
534 - removed timers from Coro::Timer -> use AnyEvent instead.
535 - replaced Coro::idle coroutine by (cleaner) idle handler.
536 - updated to newest libcoro.
537 - implement enhanced support for valgrind.
538 - implement is_ready and return value for ready.
539 - removed Coro::Cont, it was a misnomer (it's generators),
540 and it was somewhat annoying to get it right. Will come back
541 if somebody asks for it :)
542 - many ->wait methods and Coro::Event could return
543 spuriously without the event having happened.
544
545 2.5 Tue Nov 7 12:22:33 CET 2006
546 - made Coro::Util, Coro::Select, Coro::Handle and Coro::Socket
547 use AnyEvent, moved them to Coro/.
548 - added Coro::LWP which contains all the uglyness required to
549 make LWP non-blocking.
550 - should work with perl 5.9.x now (Andreas König made me do it).
551 - fixed another bug in Coro::Select when the timeout was undef.
552 - reuse PL_start_env for all coros, saves some memory per coroutine.
553 - manage PL_top_env differently, hopefully to avoid panic: top_env.
554 - timeout argument was not properly used in Coro::Socket.
555 - allow limited forms of subclassing in Coro::Handle/Coro::Socket.
556 - emulate undocumented(!) functionality of IO::Socket required
557 by LWP(!!).
558 - updated eg/lwp to work with newer lwp's.
559 - remove "FATAL: uncaught exception" prefix. Coroutines that die
560 kill the whole process, just as exceptions in the main "coroutine"
561 did already.
562
563 2.1 Wed Nov 1 23:01:13 CET 2006
564 - fix a long-standing bug in Coro::Select where select with
565 zero timeout would instead change the current default filehandle.
566 - use a simpler and hopefully more robust way to clone padlists
567 (uses less memory and a perl function instead of our own).
568 - coro can now create a stack guard area on many architectures.
569 - Coro::AIO properly reexports additional functions from IO::AIO.
570 - updated libcoro with a workaround for OS X,
571 pach and testing by Michael Schwern.
572
573 2.0 Tue Oct 24 05:47:17 CEST 2006
574 - support additional aio requests in Coro::AIO.
575
576 1.9 (never properly released due to a glitch)
577
578 1.8 Thu Feb 2 00:59:06 CET 2006
579 - applied suggested patch by SAMV to avoid problems during stupid
580 mark & sweep gc run.
581 - applied patch by Scott Walters for 5.9.3 compatibility.
582
583 1.7 Tue Dec 27 01:41:58 CET 2005
584 - added Coro::AIO, a thin wrapper around IO::AIO.
585 - improved Makefile.PL explanations.
586
587 1.6 Mon Dec 12 21:30:05 CET 2005
588 - additionally save PM_curpm might fix as-of-yet
589 unseen problems with regex matches being attributed to the
590 wrong package.
591 - add t/10_bugs.t, which currently checks against imho broken
592 perls that use 0.26MB of stack space per Perl_magic_get
593 invocation (newer linuxes) as opposed to the 0.0002MB perl
594 normally uses.
595 - make stacksize configurable for the ultimate debian experience.
596
597 1.51 Mon Dec 12 18:48:36 CET 2005
598 - remove debugging warn() accidentally left in Coro::Select.
599
600 1.5 Tue Nov 29 13:32:44 CET 2005
601 - use Coro::Event inside Coro::Select to avoid spurious deadlocks.
602 - fix Coro::Select 'select'.
603 - strict'ify some modules.
604 - libcoro errornously restored the SIGUSR2 handler to SIGUSR1.
605 - use XSLoader in selected modules.
606 - remove some 5.6 compatibility cruft.
607
608 1.4 Tue Sep 6 00:11:05 CEST 2005
609 - libcoro did not take into account the trampoline on amd64,
610 when 'l'inux method was used.
611
612 1.31 Tue Aug 30 23:31:33 CEST 2005
613 - some portability fixes/workarounds.
614
615 1.3 Sat Aug 20 03:08:56 CEST 2005
616 - no code changes. module seems to work fine.
617 - improve Coro::State docs, remove reference to nonexisting
618 Coro::State::flush method.
619 - no longer autodetect windows, present it as an option instead.
620
621 1.2 Mon May 16 02:00:55 CEST 2005
622 - included libcoro.c earlier in State.xs, might improve portability.
623 - use a faster and possibly more stable Coro::Cont implementation.
624 - accept x86_64 in addition to amd64 for optimized linux-amd64 support.
625 - fix bugs Coro::SemaphoreSet that could cause locks never to be freed.
626 - fix bugs in CoroAPI.h and document it in Coro::MakeMaker.
627
628 1.11 Thu Mar 3 18:00:52 CET 2005
629 - change of contact address.
630
631 1.1 Tue Feb 22 20:51:16 CET 2005
632 - support [l]inux method on amd64.
633 - allow some context switches while compiling/eval'ing.
634
635 1.0 Tue Aug 10 03:47:27 CEST 2004
636 - Event 0.89 fixes the bug again.
637 - fix the remaining(??) memleak. I didn't free the cache-AV for
638 padlists, so one AV per created anon-sub leaked.
639
640 0.97 Fri May 14 15:23:32 CEST 2004
641 - remove stupid gcc extension that was an experiment only anyway.
642
643 0.96 Thu May 13 18:09:29 CEST 2004
644 - tell people if they have yet another broken Event variant.
645 - made Coro to work _better_ with threads (it seems that, as long
646 as you use Coros in one thread only everything will work fine.
647 Looking at the perl source the problem might be missing locking
648 between Coro and some perl internal routines. That's not easily
649 fixable).
650 - (hopefully) got rid of the static xnext inside transfer.
651
652 0.95 Sun Feb 15 03:20:28 CET 2004
653 - removed Coro::State::flush, implemented a new and much
654 faster caching mechanism by attaching magic to the CV.
655 - a memleak with (real) closures remains.
656 - dramatically reduced size of stacks to 65536 (linux/x86), after
657 observing that even myhttpd never had more than a kilobyte
658 of stack in use (stack is still only physically allocated in
659 pagesize increments for systems suporting mmap).
660 - fix for a "cygwin-64" architecture added, proposed by Stefan Traby.
661 - fix for threaded perl.
662
663 0.9 Sun Nov 23 04:32:39 CET 2003
664 - fix a few very hard to track down but (of course) stupid
665 memory leaks.
666 - stringify version number to avoid locale problems :(.
667 - perl 5.9.0 does not have SvPADBUSY, reported by Scott Walters,
668 hopefully ignoring it "just works".
669
670 0.8 Wed Nov 5 19:38:40 CET 2003
671 - port to perl5.005_03, but only 5.8.x+ is supported!
672 - honor LocalAddr even without LocalPort in Coro::Socket.
673
674 0.7 Tue May 27 03:12:38 CEST 2003
675 - the version jump indicates some level of testing, not gobs
676 of new features.
677 - uh... I found the prompt function in ExtUtils::MakeMaker.
678 Highly correct stuff, that is...
679 - fixed(?) a bug with die's in coroutines causing "panic: top_env".
680 the fix is not well understood by the author, so beware :(.
681
682 0.652 Thu May 8 02:54:46 CEST 2003
683 - Applied patch by Slaven Rezic to set default to "s" on FreeBSD,
684 cause version 4 doesn't have ucontext.
685 - Benjamin Reed reported that setjmp works fine on darwin,
686 so preselect it.
687
688 0.651 Sat Mar 29 15:00:23 CET 2003
689 - fix a bug in Coro::Handle where some bytes could get lost
690 on reading, reported by jason@nichego.net.
691
692 0.65 Sun Mar 23 00:08:26 CET 2003
693 - added a README.linux-glibc.
694 - new module Coro::Select.
695 - also save/restore PL_comppad, fixes segfaults with 5.8.1.
696 (I never claimed that I know what I am doing ;).
697 - set default to setjmp/longjmp on non-x86-linux.
698
699 0.6 Thu Nov 21 11:09:06 CET 2002
700 - made lazy context switching the default.
701 - fixed the bug where SemaphoreSet::guard
702 was the same as timed_guard and timed_guard was missing.
703 - fix a memleak in Coro::Timer.
704 - reclassified context sharing as not experimental.
705
706 0.534 Sun Apr 14 03:05:12 CEST 2002
707 - fixed backspace => tab in header parsing (myhttpd).
708 - added eg/readline, for elmex.
709 - Coro::Event now calls ->start on first call instead of
710 ->again, which created wrong timeouts for the first ->next.
711 - fixed a bug where Coro::Socket returned a closed socket instead
712 of nothing on ECONNREFUSED and similar error conditions.
713
714 0.533 Mon Feb 18 18:49:40 CET 2002
715 - adapt to finally working Event-0.85.
716 - fixed buggy event initialization (trapped under windows, but how
717 could it possibly have worked anywhere?)
718
719 0.532 Wed Jan 16 02:45:32 CET 2002
720 - removed the reference to pp_entersub, might work on
721 (some) win32 perls now (testsuite works under cygwin,
722 without Event, which freezes).
723
724 0.531 Mon Dec 10 22:18:44 CET 2001
725 - Fixed a bug in SemaphoreSet::timed_down.
726
727 0.53 Tue Nov 27 21:11:13 CET 2001
728 - some tests for Event.
729 - slightly different internal architecture to get refcnt into the object
730 and not the reference. might fix a few bugs, certainly did introduce
731 new ones.
732 - Coro::Timer (independent of Event).
733 - new timed_wait functions for Coro::Signal, Semaphore, SemaphoreSet.
734
735 0.52 Tue Nov 6 21:36:18 CET 2001
736 - ported to cygwin (trivial).
737 patch by Gerrit P. Haase <gp@familiehaase.de>.
738 - small setjmp code fix by Sullivan.DanielJ@epamail.epa.gov.
739
740 0.51 Thu Nov 1 20:39:01 CET 2001
741 - terminate/cancel now work properly (otherwise termination
742 could cause a "next coroutine is not and contains not..."
743 error.
744 - added Coro::Socket::shutdown.
745 - Coro::Event::loop no is the same as Event::loop.
746 - implemented terminate with args + join.
747
748 0.5 Fri Sep 28 16:15:35 CEST 2001
749 - fixed "print" on a Coro::Handle. The print method worked.
750 - small tweaks (seem to reduce memory consumption a lot)
751 in various modules.
752 - splendid use of "no warnings" scattered throughout.
753 - added Coro::Handle::rbuf, fixed Coro::Handle::fh.
754
755 0.49 Sun Sep 16 02:42:45 CEST 2001
756 - changed some method calls to function calls for speed
757 inside Coro::Handle.
758 - make Coro::Handle use an array instead of a hash for
759 speed reasons.
760 - IRIX mystery solved: it's SGI's NT, after all: "standard, huh?".
761 sjlj and ucontext should now work.
762 - IRIX-specific port for libcoro.
763 - swapped order of accept results to match IO::Socket.
764 - changed getsock/peername to sock/peername to match IO::Socket.
765 - fixed a bug that caused segfault when returning to main task
766 under some circumstances.
767 - other bugfixes.
768
769 0.45 Sun Sep 2 02:54:01 CEST 2001
770 - new method Coro::Handle::timeout.
771 - corrected speling of set...name to get...name in Coro::Socket.
772 - Coro::Socket::accept now returns a Coro::Socket, not a Coro::Handle.
773 - Coro::Handle now supports fileno and FILENO.
774 - added eg/myhttpd, a web server.
775 - Coro::Socket now accepts numeric ports as well as "port(number)"
776 syntax.
777 - moved some scheduling primitives into xs code. more to come.
778 - new simple priority system.
779 - implemented Coro::Event using XS, almost four times faster!
780 - small memory corruption problem fixed (boy that was difficult).
781
782 0.13 Wed Aug 8 16:53:07 CEST 2001
783 - new method Coro::Handle::readline.
784 - added eg/lwp and Coro::Handle::autoflush, to make LWP work.
785 (see eg/lwp on how to make LWP non-blocking).
786 - renamed Coro::Socket::new_inet to new.
787 - added Coro::Util, some utility functions.
788 - die/eval now works better (still get annoying (but true ;)
789 "Callback called exit" messages).
790 - Coro::Handle now supports timeout, and Coro::Socket Timeout.
791 - much better Coro::Socket support.
792
793 0.12 Fri Jul 27 04:19:01 CEST 2001
794 - do not use mmap if neither MAP_ANON nor MAP_ANONYMOUS is defined.
795 - very experimental stack sharing algorithm. not 100% safe but
796 should work well in practise ;)
797 - added Coro::RWLock.
798 - Coro::Specific now works.
799
800 0.11 Tue Jul 24 22:49:21 CEST 2001
801 - added specialized hack for newer and older linux versions (fast).
802 - renamed Coro::Event::IO to Coro::Handle.
803 - new module Coro::Socket.
804
805 0.1 Tue Jul 24 01:47:53 CEST 2001
806 - release candidate 3. A rather new internal structure :(
807 - the great renaming: Damian Conway gave me a suitable replacement
808 for yield.
809 - added Coro/libcoro, a portable coroutine implementation for C
810 (not even perl-dependent!!!), which can be used by Coro::State.
811 - renamed SAVE_DEFSV (on xs-level only) to avoid symbol clash in
812 perl-5.6.
813 - new function Coro::Event::idle.
814 - the idle process is now overriden
815 by default in Coro::Event.
816 - Coro::Channel now enforces the size.
817 - canceling events now works in all cases.
818 - Coro state now include $_ and $@.
819 - yet another bug workaround that I do not really understand :(
820 - new module Coro::Event::IO (very undocumented) to do
821 non-blocking i/o.
822 - performance tuning.
823
824 0.08 Thu Jul 19 06:13:25 CEST 2001
825 - release candidate 2 (new functionality).
826 - add Coro::State::flush function.
827 - transfer now supports a flags value (mostly for speed ATM).
828 - might compile and work in the presence of threads now.
829 - continuations have a saner syntax.
830 - no more memleaks.
831
832 0.07 Tue Jul 17 17:40:18 CEST 2001
833 - release candidate 1 ;)
834 - slightly nicer code.
835 - fixed a scheduling bug in Coro::Event.
836
837 0.06 Tue Jul 17 04:23:24 CEST 2001
838 - ok, I found the showstopper - the same sub must not be
839 re-entered in two different coroutines, otherwise => crash. I
840 see no easy solution to this problem, except by walking the call
841 chaing and saving/restoring the cv's, which is what I do now.
842 - memory leaks still latent, especially at thread termination.
843 - Coro::Event now works (haha).
844
845 0.05 Sun Jul 15 17:32:20 CEST 2001
846 - fixed a few issues in Coro::Event.
847 - I forgot to include Coro::Event in 0.04 :(:(:(
848
849 0.04 Sun Jul 15 05:24:59 CEST 2001
850 - @_ is now properly localized.
851 - Coro::State is now easier subclassable.
852 - Coro::Cont now coroutine-aware.
853 - Coro::Specific is a low-overhead module to create
854 coroutine-specific vars.
855 - Coro::Event provides a simple interface to Event.
856
857 0.03 Fri Jul 13 14:51:52 CEST 2001
858 - transfer() now implemented in XS (beware).
859 - new module Coro::Cont for really faked continuations.
860 - big internal architecture changes: Coro::State is now
861 really low-level and can thus be used to implement other
862 interesting things, While "Coro::" implements a process-like
863 model. Still crude and subject to change.
864 - $_ and $@ are no longer being localized.
865
866 0.02 Tue Jul 10 01:38:17 CEST 2001
867 - implemented "async" attribute.
868 - $_ and $@ are now localized.
869 - added Coro::Channel.
870 - more testcases, still no docs.
871
872 0.01 Tue Jul 3 02:18:41 CEST 2001
873 - original version; copied from Convert::Scalar.
874