ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/Coro/Changes
Revision: 1.445
Committed: Thu Nov 20 03:10:30 2008 UTC (15 years, 6 months ago) by root
Branch: MAIN
Changes since 1.444: +3 -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: async_pool to SLF
9 TODO: _cancel to xs
10 5.0
11 - NEW ARCHITECTURE: use the latest 4.x version if you experience
12 stability issues.
13 - bump API version to 7 - all dependents must be recompiled.
14 - removed timed_* functions - they were not being used anyways
15 and should be replaced by a more generic mechanism -
16 and were annoying to support anyways :)
17 - removed SemaphoreSet's waiter method - use sem instead.
18 - Coro::Semaphore->adjust didn't correctly wake up enough waiters.
19 - async_pool did free a scalar value twice
20 ("Attempt to unreference...").
21 - fix a disastrous bug in the readline optimisation
22 introduced in 4.801.
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 - used new SLF interface and other optimisations to speed up async_pool
59 by a factor of two. It also doesn't rely on perl's exception mechanism
60 to exit anymore.
61 - sped up coroutine creation/destruction by 40%.
62 - forgot to include Coro/libcoro/README in the dist for all these years.
63 - work around a freebsd pthreads bug (manual testcancel is required as
64 pthread_cond_wait isn't a cancellation point on freebsd).
65 - use new rouse functions to speed up and simplify Coro::BDB.
66 - make "prefer perl native functions" work with threaded perls.
67 - condense Coro::Debug ps output, hint at v and w flags.
68 - (libcoro) lots of minor cleanups and portability improvements.
69
70 4.913 Sat Nov 15 07:58:28 CET 2008
71 - async_pool did free a scalar value twice
72 ("Attempt to unreference...").
73
74 4.912 Thu Nov 13 18:31:23 CET 2008
75 - minor cleanups.
76 - use much larger stacks on linux and perl < 5.8.8.
77 - Coro::Debug::new_unix_server did not unlink the socket
78 when destroyed.
79
80 4.911 Tue Nov 11 04:26:17 CET 2008
81 - "port" to threaded perls.
82
83 4.91 Mon Nov 10 05:36:38 CET 2008
84 - the ->throw exception object no longer leaks.
85 - creating a new cctx leaked a scopestack entry (memleak).
86 - new coroutines didn't get created with a zero flags field
87 (unknown impact).
88 - calling ->throw on a not-yet-started coroutine should now work
89 instead of being ignored.
90 - ->throw is now supported on Coro::State objects.
91 - clean up cctx creation code a bit.
92 - entersub is actually a UNOP, not a LOGOP (not a bugfix).
93
94 4.9 Sat Nov 8 17:45:27 CET 2008
95 - (libcoro) did not preserve rbp with CORO_ASM (we are getting there).
96 - (libcoro) no longer leak threads in the experimental pthread backend,
97 also speed it up considerably.
98 - (libcoro) do not rely on makecontext passing void *'s unscathed.
99 - fix compiletime dependencies on libcoro in the Makefile.
100 - cctx_count wasn't always updated properly.
101 - Coro::State::cctx_stacksize wasn't applied correctly.
102 - new function Coro::State::cctx_max_idle.
103 - the default max number of idle C contexts is now 4.
104 - (libcoro) try harder to get _setjmp/_longjmp.
105 - (libcoro) cleanup and extend the libcoro API to officially
106 allow the creation of empty source contexts.
107 - very experimental workaround for the totally broken netbsd platform.
108 - tried to hack around openbsd bugs.
109
110 4.804 Wed Nov 5 16:36:00 CET 2008
111 - Coro::Debug::new_unix_server would not create a non-blocking listening
112 socket, sometimes causing freezes.
113 - (libcoro) fix misaligned stack points for setjmp and assembly
114 methods, which can cause crashes on x86/x86_64 with a sufficiently
115 aggressive compiler.
116 - new function: Coro::Debug::new_tcp_server.
117 - move ->throw into the Coro class because it only works on coro objects.
118
119 4.803 Mon Nov 3 17:16:12 CET 2008
120 - (libcoro) use a global asm statement to become independent of gcc
121 otpimisations for CORO_ASM (thanks to pippijn for the idea).
122 - try to workaround yet another broken bsd, this time dragonfly.
123
124 4.802 Thu Oct 30 10:56:12 CET 2008
125 - support -fno-omit-frame-pointer on x86 with the assembly method.
126 - tune 01_unblock.t tests a bit.
127
128 4.801 Wed Oct 22 18:33:37 CEST 2008
129 - improve readline speed for very long "lines".
130 - backport to 5.8.8.
131
132 4.8 Thu Oct 2 13:34:40 CEST 2008
133 - new function Coro::AIO::aio_wait.
134 - Coro.:AIO and Coro::BDB now "use Coro::AnyEvent".
135 - greatly speed up and reduce memory usage of Coro::AIO requests.
136 - implement some other µ-optimisations.
137
138 4.749 Mon Sep 29 14:40:12 CEST 2008
139 - port to slow and broken pseudo-threaded perls. (courtesy pippijn).
140
141 4.748 Sat Sep 27 14:03:19 CEST 2008
142 - implement, but do not document, PerlIO::cede(granularity).
143 - Coro::Storable forgot to wrap Storable::pstore.
144 - work around the multitude of leaks and memory corruption
145 bugs in PerlIO::via by using our own C-level perliol. As a side
146 effect, Coro::Storable is now much, much, much faster.
147
148 4.747 Tue Sep 23 01:59:41 CEST 2008
149 - fix a per-cv memleak (one empty array was leaked per
150 code reference).
151 - avoid a crash in coro->call|rss when the coroutine was already
152 destroyed (most noticably when using Coro::Debug::ps :)
153 - also protect *Storable::FILE.
154 - push up default storable granularity to 20ms.
155
156 4.746 Sun Sep 21 03:22:20 CEST 2008
157 - be more insistent on locking Storable against reentrancy
158 in Coro::Storable.
159 - move swap_def?v and throw to Coro::State, as documented.
160
161 4.745 Thu Jul 24 00:14:38 CEST 2008
162 - remove debugging code related to MgPV_nolen_const, also try to
163 make it compile with perl 5.8.6 (yes, apple apparently loves
164 outdated software). Reported by John S.
165
166 4.744 Tue Jul 8 22:06:35 CEST 2008
167 - correctly provide default DIE/WARN handlers as documented.
168 - also overwrite PL_vtbl_sigelem.svt_clear, even though current
169 implementations inside perl work fine for us.
170
171 4.743 Mon Jun 16 00:21:57 CEST 2008
172 - when using Coro::EV without running EV::loop it could
173 result in busy-waiting for events, this has been fixed.
174 - reduce codesize and improve performance by using EV_DEFAULT_UC.
175
176 4.742 Sat May 31 14:10:21 CEST 2008
177 - implement a workaround for (some) perl <5.8.8 versions.
178 - require EV 3.3+.
179
180 4.741 Fri May 30 23:33:09 CEST 2008
181 - tell netbsd how utterly broken their imitation of an OS is
182 and refuse to build by default if pthreads are in use.
183 - switch to "s" method on all bsd's by default, as their ucontext
184 stuff seems just too broken.
185 - fix a bug in Coro::Select.
186
187 4.74 Thu May 29 20:05:31 CEST 2008
188 - do not test Coro::LWP for lack of dependencies.
189 - do not test Coro::Select for lack of working perls.
190
191 4.73 Thu May 29 2008
192 - fix a bug in Coro::EV which would cause it to block despite
193 there being runnable coroutines.
194 - sprinkle "no warnings" freely over everything, also suppress
195 warnings for some other modules.
196 - fix typo in WSAEWOULDBLOCK.
197
198 4.72 Sun May 25 05:14:36 CEST 2008
199 - tweak META.yaml a bit, unfortunately, there is no documented way
200 to have optional dependencies with CPAN. doh :(
201 - avoid running some tests on windows because they would fail due to
202 perl bug (broken fork, broken pipes...).
203 - work around perl on windows bugs where perl returns undocumented
204 error codes for sysread, syswrite etc. by taking advantage
205 of AnyEvent's workaround for that problem.
206 - use AnyEvent::Util::fh_nonblocking in Coro::Handle to work around
207 a common perl implementation bug on windows.
208 - use unix domain sockets in testsuite to work around
209 common perl implementation bugs on widows (they are emulated by
210 tcp sockets on windows. ugh.)
211
212 4.71 Sat May 24 20:01:27 CEST 2008
213 - fix a bug in Coro::AnyEvent ("Usage: Coro::AnyEvent::_schedule()").
214 - take advantage of async name resolution of AnyEvent::Util.
215 - work around brutal inet_aton override in Coro::LWP.
216 - take advantage of the readyhook in Coro::EV, for smoother
217 scheduling.
218
219 4.7 Sun May 11 00:32:19 CEST 2008
220 - completely reworked the Coro manpage.
221 - added Coro::AnyEvent, generic event loop integration.
222 - implement cancel, ready and kill commands in Coro::Debug.
223 - document find_coro in Coro::Debug.
224 - incompatible change: rename has_stack to has_cctx.
225 - Coro::AIO and Coro::BDB no longer force event model detection,
226 use AnyEvent::AIO and AnyEvent::BDB.
227
228 4.6 Sat Apr 26 10:05:14 CEST 2008
229 - INCOMPATIBLE CHANGE: sub/code attributes are no longer supported
230 by the Coro module. It was a mistake to have it in the first place.
231 - (experimental) support for activestate perl 5.10 (method "w").
232 - (experimental) support for strawberry perl (method "a").
233 - coro_sigelem_set did not return a value although it had to,
234 actual impact unknown.
235
236 4.51 Mon Apr 14 13:28:27 CEST 2008
237 - make it compile again on 5.8.
238
239 4.50 Thu Apr 10 09:43:17 CEST 2008
240 - I did it twice! (see 4.49).
241
242 4.49 Sun Apr 6 21:23:31 CEST 2008
243 - grr, instead of compiling the recent changes on 5.10 only they
244 were compiled on 5.8 only.
245
246 4.48 Sun Apr 6 20:36:46 CEST 2008
247 - allow coroutine switches during eval's under 5.10.x, as it seems
248 the parser is a per-interpreter option now, so this is safe
249 (this might fix the odd crashing bug).
250 - drop support for 5.9.x versions: they are dead, jim.
251
252 4.47 Sun Apr 6 00:37:52 CEST 2008
253 - force cctx allocation on API calls: we know we need to force one
254 and gcc actually manages to confuse our heuristic nowadays,
255 leading to crashes and worse.
256 - document force_cctx.
257
258 4.46 Fri Apr 4 22:05:43 CEST 2008
259 - upgrade libcoro, resulting in pthread-backend (which was only created
260 to fulfill the rules of the programming languages shootout).
261
262 4.45 Thu Mar 13 11:55:36 CET 2008
263 - fix a file leaking bug in eg/mhyttpd that would allow
264 downloading of any file (reported by oesi).
265 - fix deadlock bug in Coro::Channel (reported by Richard Hundt)
266 (also add testcase).
267 - support Broadcast option in Coro::Socket (patch by Richard Hundt,
268 apparently having loads of fun with that).
269 - implement $state->swap_defsv, ->swap_defav and document ->throw.
270
271 4.4 Wed Feb 13 16:44:29 CET 2008
272 - only assume .cfi pseudo isns exist on GNU/Linux.
273 - add get_timed method to Coro::Channels.
274 - fixed Coro::Debug manpage.
275 - perl 5.11 compatbility improvement based on patch by
276 Andreas König.
277
278 4.37 Sun Jan 20 11:25:23 CET 2008
279 - soften the check for an active parser for perl 5.10 (avoiding
280 "Coro::State::transfer called while parsing" in many cases).
281
282 4.36 Sun Jan 13 10:53:56 CET 2008
283 - reset diehook when terminating from an async_pool as to not
284 trigger any __DIE__ handlers.
285
286 4.35 Sun Jan 13 04:14:13 CET 2008
287 - "bt" debug command now displays any exceptions
288 from longmess and also skips the innermost
289 stackframes, giving more useufl output.
290 - allow backtraces in coroutines blocked in Coro::EV,
291 at a <1% speed hit.
292 - handle localising of $SIG{__WARN__} and __DIE__
293 properly (with a proper amount of dirty hacking).
294
295 4.34 Sat Dec 22 17:49:53 CET 2007
296 - upgrade to EV version 2.0 API.
297
298 4.33 Mon Dec 17 08:36:12 CET 2007
299 - make Coro::AIO etc. loadable in the absence of EV.
300
301 4.32 Mon Dec 17 07:46:02 CET 2007
302 - majorly improved Coro::Handle's performance with Coro::EV.
303 - implemented the until now mythical Coro::BDB module.
304 - specialcase EV in Coro::AIO and Coro::BDB for extra speed.
305
306 4.31 Wed Dec 5 12:32:39 CET 2007
307 - remove warn statement form Coro::Util.
308
309 4.3 Tue Dec 4 20:33:14 CET 2007
310 - calls to the idle function could cause stack corruption
311 when the stack changed.
312 - do no longer rely on the presence of EV::DNS if EV is used
313 (because it is gone), but instead take avdantage of EV::ADNS
314 if available.
315 - add ($) prototypes to all functions in Coro::Storable.
316 - use a conventional (safer) idle callback in Coro::EV.
317 - do accept quickly in Coro::Debug to avoid endless loops.
318
319 4.22 Fri Nov 30 16:04:04 CET 2007
320 - really use optimised versions for Event and EV in Coro::Util
321 and Coro::Handle.
322
323 4.21 Sun Nov 25 10:48:59 CET 2007
324 - fix a spurious memory read.
325 - Coro::EV no longer keeps the eventloop "alive".
326
327 4.2 Fri Nov 9 20:47:05 CET 2007
328 - enable/disable tracing from a new coroutine, not a pooled one.
329 - fix a memleak in Coro::Event.
330 - removed killall call from fork_eval.
331 - made sure store_fd is already loaded so that fork_eval
332 does not have to parse autoload in each subprocess.
333 - only use assembly method if -O switch is in $Config{optimize}.
334 - add (optional) Coro::EV module, so far the best event loop module
335 directly supported by Coro.
336 - if the event model is EV, use EV::DNS to resolve
337 stuff in Coro::Util.
338 - don't get confused by multiple event notifications in Coro::Handle.
339 - initial support for EV (libevent interface).
340 - require Event and EV using configure_requires, to force their existance.
341
342 4.13 Wed Oct 24 07:26:45 CEST 2007
343 - add Coro::Storable::blocking_thaw.
344 - use a vastly more complicated technique to localise
345 $SIG{__WARN/DIE__} that also works on perls <= 5.8.8.
346 - use a coroutine for the idle callback Coro::Event,
347 instead of running Event in the current coroutine context.
348 This also catches recursive invocations.
349 - actually report fork errors in gethostbyname and inet_aton.
350
351 4.11 Thu Oct 11 02:40:24 CEST 2007
352 - port to threaded perls.
353
354 4.1 Thu Oct 11 02:38:16 CEST 2007
355 - incompatible change: $SIG{__DIE__} and $SIG{__WARN__} will now
356 be local to each coro (see Coro::State).
357 - incompatible change: for very deep reasons, cede and cede_notself
358 cannot return anything, so nothing will be returned.
359 - possibly bring back 5.10 compatibility (untested).
360 - work around stupid (and wrong) warning on 5.10 :(.
361 - overlay the saved state over the context stack. This saves
362 a few hundred bytes per coroutine on average and also
363 speeds up context switching a bit.
364 - further tune default stack sizes.
365 - (more) correctly calculate stack usage in coro_rss.
366 - Coro::Storable::blocking_* did not properly lock
367 resulting in races between coroutines.
368 - added Coro::Storable::guard.
369 - stopping to trace a coroutine could destroy the cctx of
370 an unrelated coroutine.
371 - explain the relationship between Perl and C coroutines in
372 more detail in Coro::State.
373 - Coro::Util::inet_aton did not short-circuit dotted quad forms,
374 causing a fork per resolve. This also affected Coro::Socket.
375 - switch to a separate stack in $coro->call/eval to avoid
376 invalidating pointers.
377
378 4.03 Sat Oct 6 21:24:00 CEST 2007
379 - added Coro::throw method.
380 - minor code cleanups.
381
382 4.02 Sat Oct 6 02:36:47 CEST 2007
383 - fix a very minor per-coroutine memleak (a single codereference).
384 - fixed a bug where the currently in-use c context would be freed
385 prematurely (can happen only when programs change the stacksize
386 or use tracing).
387 - tracing can no longer keep a coro alive after it terminated.
388 - do static branch prediction in the common path for gcc. gives
389 about 2-5% speed improvement here.
390
391 4.01 Fri Oct 5 22:10:49 CEST 2007
392 - instead of recreating *a* standard output handle we simply
393 use STDOUT, which is faster and hopefully more robust.
394
395 4.0 Fri Oct 5 12:56:00 CEST 2007
396 - incompatibly changed Coro::Storable::freeze.
397 - major new feature: added Coro::Debug, for interactive coroutine
398 debugging, tracing and much more.
399 - major bug fix: unbelievable, but true: $_, $/ and many other
400 "saved" variables actually weren't being saved. This has been fixed,
401 of course, while increasing performance while losing all the save
402 flags.
403 - save flags are gone, and all the api functions dealing with them.
404 - added Coro::Semaphore::adjust.
405 - added Coro::Util::fork_eval.
406 - added Coro::Storable::{nfreeze,blocking_{freeze,nfreeze}}.
407 - added Coro::killall.
408 - reduce initial stack sizes to allow for "micro-coroutines".
409 - better async_pool resource management, moved parts of async_pool
410 handling to XS (major speed improvement).
411 - actually croak before modifying important data structures.
412 - refuse to transfer while compiling.
413 - possibly support eval EXPR better now.
414 - enable assembly per default on linux+bsd x86+amd64.
415 - all internal members were renamed _something for easier subclassing.
416 - many minor tweaks.
417
418 3.63 Wed May 16 14:10:06 CEST 2007
419 - implement handcoded assembly for x86/amd64 SVR ABI.
420
421 3.62 Fri Apr 27 21:36:06 CEST 2007
422 - upgrade libcoro (which might set unwind info correctly).
423 - change default on linux to setjmp/longjmp.
424
425 3.61 Thu Apr 19 12:36:18 CEST 2007
426 - Coro::Storable caused an endless loop when thawing invalid
427 pst's sometimes.
428 - use a Semaphore in Coro::Storable, as Storable doesn't
429 seem to be reentrant (although it is documented to
430 be threadsafe...).
431 - fix Coro::Signal to bring back the original unreliable
432 but stateful semantics.
433 - fixed a lot of typos in Coro.pm (patch submitted by David
434 Steinbrunner, which applied flawlessly).
435
436 3.6 Sat Apr 14 17:13:31 CEST 2007
437 - added some bugfixes to get eg/myhttpd working again.
438 - added Coro::Storable for often-cede'ing freeze/thaw.
439 - try to do a clean exit when a coroutine calls exit
440 (EXPERIMENTAL).
441 - got rid of indirect call through _coro_init.
442 - updated the partly antique examples in eg/ to
443 work again and be a bit less magic, too.
444 - fixed Coro::Signal semantics to work as documented again.
445
446 3.55 Sun Mar 25 01:20:47 CET 2007
447 - add SAVE_DEFFH to save the default filehandle and enable
448 it by default.
449 - finally move socket-operations from Coro::Socket to Coro::Handle
450 to be able to unblock foreign sockets.
451 - add Coro::State::save_also and guarded_save.
452 - add count accessor to Coro::Semaphore.
453 - add Coro::State::cctx_stacksize.
454 - just for the fun of it, do not rely on implicit context,
455 which can dramatically improve performance, but people
456 using windows-process-emulation perls do not care much about
457 performance.
458
459 3.51 Sun Mar 4 14:18:04 CET 2007
460 - fixed a problem when you weakened references to Coro::State's
461 (patch by Yuval Kogman).
462
463 3.501 Wed Feb 28 12:44:07 CET 2007
464 - rename some global symbols as macosx from hell redefines
465 them without asking.
466
467 3.5 Tue Feb 13 20:22:53 CET 2007
468 - do AnyEvent model detection earlier, avoiding problems
469 caused by first using AnyEvent and later Coro::Event.
470 - implement and document Coro::Event event objects.
471 - fix a potential problem in Coro::Event causing crashes.
472 - initialise PL_comppad when creating a new coroutine,
473 avoids crashes on early coro destruction.
474
475 3.41 Mon Jan 22 19:19:49 CET 2007
476 - readline on Coro::Handle did not support undefined $/,
477 nor did it deliver partial lines on EOF or error.
478 - implement malloc fallback for stack allocation because
479 stupid broken idiotic OSX has a stupid broken
480 idiotic fits-the-whole-os mmap "implementation" and
481 my dick feels longer if Coro is portable even to
482 obsolete platforms.
483
484 3.4 Fri Jan 19 21:52:54 CET 2007
485 - remove t/09_timer.t, as it isn't really testing much
486 but was rather flaky in practise.
487 - async_pool coro would keep arguments and callback alive until
488 it was reused.
489 - cancellation of a coroutine could cause spurious idle calls
490 in cede_notself.
491
492 3.3 Sat Jan 6 03:45:00 CET 2007
493 - implement $coro->on_destroy.
494 - Coro::Event blocking semantics have been changed,
495 documented and - hopefully - improved.
496 - fix nice adding, not subtracting, from priority.
497 - fix ->prio and api_is_ready (patch by Mark Hinds).
498 - fixed an assert ("... == prev__cctx->idle_te")
499 that could errronously trigger.
500 - fix various large and small memleaks.
501 - use a (hopefully) more stable cancel implementation
502 that immediately frees the coroutine data.
503 - cede/cede_notself return a status now.
504 - added Coro::guard function.
505 - added a global coroutine pool for jobs (on my machine,
506 I can create and execute 48k simple coros/s with async,
507 and 128k coros with async_pool).
508 - Coro::AIO now uses the coroutine priority as io priority.
509
510 3.2 Fri Dec 22 05:07:09 CET 2006
511 - improve portability to slightly older perls.
512 - use cleaner coroutine destruction.
513 - simplify configuration for users.
514 - optionally (unrecommended) prefer perl functions over
515 their coro replacements.
516
517 3.11 Tue Dec 5 13:11:24 CET 2006
518 - fixed some bogus assert's, but as perl.h disables assert even
519 without NDEBUG (thank you very much), not too many people should
520 notice (that did include myself). Andreas König noticed, though :)
521 - do not save/restore PL_sortcxix on >= 5.9.x, it doesn't seem to have
522 it. Also noticed by Andreas König :)
523 - save/restore tainted status.
524 - verified to pass the testsuite on my 5.9.5.
525
526 3.1 Mon Dec 4 23:03:40 CET 2006
527 - INCOMPATIBLE CHANGE: $/ is now per-coroutine (but slow).
528 - incompatible change: transfer flags are now per-state.
529 - give a better error message on deadlock.
530 - document Coro::nready.
531 - enhanced testsuite.
532
533 3.01 Sun Dec 3 23:47:42 CET 2006
534 - forgot to include Coro::Timer.
535
536 3.0 Sun Dec 3 22:57:25 CET 2006
537 - the "FINALLY COMPLETELY STABLE" release (coming soon:
538 the "FAMOUS LAST WORDS" release).
539 - implement a new stack sharing algorithm, which uses a stack
540 pool (size currently hardcoded).
541 - make stack sharing mandatory (it no longer uses a heuristic).
542 - eval/die no longer cause weird problems under heavy use.
543 - Coro::Event could cause livelocks if it was used but
544 no Coro::Event watchers were used.
545 - Coro::Event now uses asynccheck as prepare does not
546 check for changed watchers.
547 - Coro::Event allows multiple waiting coros and will wake up one
548 per event.
549 - Coro::Event should be cleaner and more efficient now.
550 - new utility function Coro::unblock_sub.
551 - document the sad fact that Event is no longer reentrant.
552 - putting a coroutine into the ready queue twice could under
553 some circumstances lead to stack corruption.
554 - minor incompatibility: subclassing Coro::State is supported
555 directly now without going through a _coro_state member.
556 - state/coro switching is much faster now.
557 - very minor optimisations and code/documentation cleanup.
558 - avoid problems due to compiler inlining.
559 - removed timers from Coro::Timer -> use AnyEvent instead.
560 - replaced Coro::idle coroutine by (cleaner) idle handler.
561 - updated to newest libcoro.
562 - implement enhanced support for valgrind.
563 - implement is_ready and return value for ready.
564 - removed Coro::Cont, it was a misnomer (it's generators),
565 and it was somewhat annoying to get it right. Will come back
566 if somebody asks for it :)
567 - many ->wait methods and Coro::Event could return
568 spuriously without the event having happened.
569
570 2.5 Tue Nov 7 12:22:33 CET 2006
571 - made Coro::Util, Coro::Select, Coro::Handle and Coro::Socket
572 use AnyEvent, moved them to Coro/.
573 - added Coro::LWP which contains all the uglyness required to
574 make LWP non-blocking.
575 - should work with perl 5.9.x now (Andreas König made me do it).
576 - fixed another bug in Coro::Select when the timeout was undef.
577 - reuse PL_start_env for all coros, saves some memory per coroutine.
578 - manage PL_top_env differently, hopefully to avoid panic: top_env.
579 - timeout argument was not properly used in Coro::Socket.
580 - allow limited forms of subclassing in Coro::Handle/Coro::Socket.
581 - emulate undocumented(!) functionality of IO::Socket required
582 by LWP(!!).
583 - updated eg/lwp to work with newer lwp's.
584 - remove "FATAL: uncaught exception" prefix. Coroutines that die
585 kill the whole process, just as exceptions in the main "coroutine"
586 did already.
587
588 2.1 Wed Nov 1 23:01:13 CET 2006
589 - fix a long-standing bug in Coro::Select where select with
590 zero timeout would instead change the current default filehandle.
591 - use a simpler and hopefully more robust way to clone padlists
592 (uses less memory and a perl function instead of our own).
593 - coro can now create a stack guard area on many architectures.
594 - Coro::AIO properly reexports additional functions from IO::AIO.
595 - updated libcoro with a workaround for OS X,
596 pach and testing by Michael Schwern.
597
598 2.0 Tue Oct 24 05:47:17 CEST 2006
599 - support additional aio requests in Coro::AIO.
600
601 1.9 (never properly released due to a glitch)
602
603 1.8 Thu Feb 2 00:59:06 CET 2006
604 - applied suggested patch by SAMV to avoid problems during stupid
605 mark & sweep gc run.
606 - applied patch by Scott Walters for 5.9.3 compatibility.
607
608 1.7 Tue Dec 27 01:41:58 CET 2005
609 - added Coro::AIO, a thin wrapper around IO::AIO.
610 - improved Makefile.PL explanations.
611
612 1.6 Mon Dec 12 21:30:05 CET 2005
613 - additionally save PM_curpm might fix as-of-yet
614 unseen problems with regex matches being attributed to the
615 wrong package.
616 - add t/10_bugs.t, which currently checks against imho broken
617 perls that use 0.26MB of stack space per Perl_magic_get
618 invocation (newer linuxes) as opposed to the 0.0002MB perl
619 normally uses.
620 - make stacksize configurable for the ultimate debian experience.
621
622 1.51 Mon Dec 12 18:48:36 CET 2005
623 - remove debugging warn() accidentally left in Coro::Select.
624
625 1.5 Tue Nov 29 13:32:44 CET 2005
626 - use Coro::Event inside Coro::Select to avoid spurious deadlocks.
627 - fix Coro::Select 'select'.
628 - strict'ify some modules.
629 - libcoro errornously restored the SIGUSR2 handler to SIGUSR1.
630 - use XSLoader in selected modules.
631 - remove some 5.6 compatibility cruft.
632
633 1.4 Tue Sep 6 00:11:05 CEST 2005
634 - libcoro did not take into account the trampoline on amd64,
635 when 'l'inux method was used.
636
637 1.31 Tue Aug 30 23:31:33 CEST 2005
638 - some portability fixes/workarounds.
639
640 1.3 Sat Aug 20 03:08:56 CEST 2005
641 - no code changes. module seems to work fine.
642 - improve Coro::State docs, remove reference to nonexisting
643 Coro::State::flush method.
644 - no longer autodetect windows, present it as an option instead.
645
646 1.2 Mon May 16 02:00:55 CEST 2005
647 - included libcoro.c earlier in State.xs, might improve portability.
648 - use a faster and possibly more stable Coro::Cont implementation.
649 - accept x86_64 in addition to amd64 for optimized linux-amd64 support.
650 - fix bugs Coro::SemaphoreSet that could cause locks never to be freed.
651 - fix bugs in CoroAPI.h and document it in Coro::MakeMaker.
652
653 1.11 Thu Mar 3 18:00:52 CET 2005
654 - change of contact address.
655
656 1.1 Tue Feb 22 20:51:16 CET 2005
657 - support [l]inux method on amd64.
658 - allow some context switches while compiling/eval'ing.
659
660 1.0 Tue Aug 10 03:47:27 CEST 2004
661 - Event 0.89 fixes the bug again.
662 - fix the remaining(??) memleak. I didn't free the cache-AV for
663 padlists, so one AV per created anon-sub leaked.
664
665 0.97 Fri May 14 15:23:32 CEST 2004
666 - remove stupid gcc extension that was an experiment only anyway.
667
668 0.96 Thu May 13 18:09:29 CEST 2004
669 - tell people if they have yet another broken Event variant.
670 - made Coro to work _better_ with threads (it seems that, as long
671 as you use Coros in one thread only everything will work fine.
672 Looking at the perl source the problem might be missing locking
673 between Coro and some perl internal routines. That's not easily
674 fixable).
675 - (hopefully) got rid of the static xnext inside transfer.
676
677 0.95 Sun Feb 15 03:20:28 CET 2004
678 - removed Coro::State::flush, implemented a new and much
679 faster caching mechanism by attaching magic to the CV.
680 - a memleak with (real) closures remains.
681 - dramatically reduced size of stacks to 65536 (linux/x86), after
682 observing that even myhttpd never had more than a kilobyte
683 of stack in use (stack is still only physically allocated in
684 pagesize increments for systems suporting mmap).
685 - fix for a "cygwin-64" architecture added, proposed by Stefan Traby.
686 - fix for threaded perl.
687
688 0.9 Sun Nov 23 04:32:39 CET 2003
689 - fix a few very hard to track down but (of course) stupid
690 memory leaks.
691 - stringify version number to avoid locale problems :(.
692 - perl 5.9.0 does not have SvPADBUSY, reported by Scott Walters,
693 hopefully ignoring it "just works".
694
695 0.8 Wed Nov 5 19:38:40 CET 2003
696 - port to perl5.005_03, but only 5.8.x+ is supported!
697 - honor LocalAddr even without LocalPort in Coro::Socket.
698
699 0.7 Tue May 27 03:12:38 CEST 2003
700 - the version jump indicates some level of testing, not gobs
701 of new features.
702 - uh... I found the prompt function in ExtUtils::MakeMaker.
703 Highly correct stuff, that is...
704 - fixed(?) a bug with die's in coroutines causing "panic: top_env".
705 the fix is not well understood by the author, so beware :(.
706
707 0.652 Thu May 8 02:54:46 CEST 2003
708 - Applied patch by Slaven Rezic to set default to "s" on FreeBSD,
709 cause version 4 doesn't have ucontext.
710 - Benjamin Reed reported that setjmp works fine on darwin,
711 so preselect it.
712
713 0.651 Sat Mar 29 15:00:23 CET 2003
714 - fix a bug in Coro::Handle where some bytes could get lost
715 on reading, reported by jason@nichego.net.
716
717 0.65 Sun Mar 23 00:08:26 CET 2003
718 - added a README.linux-glibc.
719 - new module Coro::Select.
720 - also save/restore PL_comppad, fixes segfaults with 5.8.1.
721 (I never claimed that I know what I am doing ;).
722 - set default to setjmp/longjmp on non-x86-linux.
723
724 0.6 Thu Nov 21 11:09:06 CET 2002
725 - made lazy context switching the default.
726 - fixed the bug where SemaphoreSet::guard
727 was the same as timed_guard and timed_guard was missing.
728 - fix a memleak in Coro::Timer.
729 - reclassified context sharing as not experimental.
730
731 0.534 Sun Apr 14 03:05:12 CEST 2002
732 - fixed backspace => tab in header parsing (myhttpd).
733 - added eg/readline, for elmex.
734 - Coro::Event now calls ->start on first call instead of
735 ->again, which created wrong timeouts for the first ->next.
736 - fixed a bug where Coro::Socket returned a closed socket instead
737 of nothing on ECONNREFUSED and similar error conditions.
738
739 0.533 Mon Feb 18 18:49:40 CET 2002
740 - adapt to finally working Event-0.85.
741 - fixed buggy event initialization (trapped under windows, but how
742 could it possibly have worked anywhere?)
743
744 0.532 Wed Jan 16 02:45:32 CET 2002
745 - removed the reference to pp_entersub, might work on
746 (some) win32 perls now (testsuite works under cygwin,
747 without Event, which freezes).
748
749 0.531 Mon Dec 10 22:18:44 CET 2001
750 - Fixed a bug in SemaphoreSet::timed_down.
751
752 0.53 Tue Nov 27 21:11:13 CET 2001
753 - some tests for Event.
754 - slightly different internal architecture to get refcnt into the object
755 and not the reference. might fix a few bugs, certainly did introduce
756 new ones.
757 - Coro::Timer (independent of Event).
758 - new timed_wait functions for Coro::Signal, Semaphore, SemaphoreSet.
759
760 0.52 Tue Nov 6 21:36:18 CET 2001
761 - ported to cygwin (trivial).
762 patch by Gerrit P. Haase <gp@familiehaase.de>.
763 - small setjmp code fix by Sullivan.DanielJ@epamail.epa.gov.
764
765 0.51 Thu Nov 1 20:39:01 CET 2001
766 - terminate/cancel now work properly (otherwise termination
767 could cause a "next coroutine is not and contains not..."
768 error.
769 - added Coro::Socket::shutdown.
770 - Coro::Event::loop no is the same as Event::loop.
771 - implemented terminate with args + join.
772
773 0.5 Fri Sep 28 16:15:35 CEST 2001
774 - fixed "print" on a Coro::Handle. The print method worked.
775 - small tweaks (seem to reduce memory consumption a lot)
776 in various modules.
777 - splendid use of "no warnings" scattered throughout.
778 - added Coro::Handle::rbuf, fixed Coro::Handle::fh.
779
780 0.49 Sun Sep 16 02:42:45 CEST 2001
781 - changed some method calls to function calls for speed
782 inside Coro::Handle.
783 - make Coro::Handle use an array instead of a hash for
784 speed reasons.
785 - IRIX mystery solved: it's SGI's NT, after all: "standard, huh?".
786 sjlj and ucontext should now work.
787 - IRIX-specific port for libcoro.
788 - swapped order of accept results to match IO::Socket.
789 - changed getsock/peername to sock/peername to match IO::Socket.
790 - fixed a bug that caused segfault when returning to main task
791 under some circumstances.
792 - other bugfixes.
793
794 0.45 Sun Sep 2 02:54:01 CEST 2001
795 - new method Coro::Handle::timeout.
796 - corrected speling of set...name to get...name in Coro::Socket.
797 - Coro::Socket::accept now returns a Coro::Socket, not a Coro::Handle.
798 - Coro::Handle now supports fileno and FILENO.
799 - added eg/myhttpd, a web server.
800 - Coro::Socket now accepts numeric ports as well as "port(number)"
801 syntax.
802 - moved some scheduling primitives into xs code. more to come.
803 - new simple priority system.
804 - implemented Coro::Event using XS, almost four times faster!
805 - small memory corruption problem fixed (boy that was difficult).
806
807 0.13 Wed Aug 8 16:53:07 CEST 2001
808 - new method Coro::Handle::readline.
809 - added eg/lwp and Coro::Handle::autoflush, to make LWP work.
810 (see eg/lwp on how to make LWP non-blocking).
811 - renamed Coro::Socket::new_inet to new.
812 - added Coro::Util, some utility functions.
813 - die/eval now works better (still get annoying (but true ;)
814 "Callback called exit" messages).
815 - Coro::Handle now supports timeout, and Coro::Socket Timeout.
816 - much better Coro::Socket support.
817
818 0.12 Fri Jul 27 04:19:01 CEST 2001
819 - do not use mmap if neither MAP_ANON nor MAP_ANONYMOUS is defined.
820 - very experimental stack sharing algorithm. not 100% safe but
821 should work well in practise ;)
822 - added Coro::RWLock.
823 - Coro::Specific now works.
824
825 0.11 Tue Jul 24 22:49:21 CEST 2001
826 - added specialized hack for newer and older linux versions (fast).
827 - renamed Coro::Event::IO to Coro::Handle.
828 - new module Coro::Socket.
829
830 0.1 Tue Jul 24 01:47:53 CEST 2001
831 - release candidate 3. A rather new internal structure :(
832 - the great renaming: Damian Conway gave me a suitable replacement
833 for yield.
834 - added Coro/libcoro, a portable coroutine implementation for C
835 (not even perl-dependent!!!), which can be used by Coro::State.
836 - renamed SAVE_DEFSV (on xs-level only) to avoid symbol clash in
837 perl-5.6.
838 - new function Coro::Event::idle.
839 - the idle process is now overriden
840 by default in Coro::Event.
841 - Coro::Channel now enforces the size.
842 - canceling events now works in all cases.
843 - Coro state now include $_ and $@.
844 - yet another bug workaround that I do not really understand :(
845 - new module Coro::Event::IO (very undocumented) to do
846 non-blocking i/o.
847 - performance tuning.
848
849 0.08 Thu Jul 19 06:13:25 CEST 2001
850 - release candidate 2 (new functionality).
851 - add Coro::State::flush function.
852 - transfer now supports a flags value (mostly for speed ATM).
853 - might compile and work in the presence of threads now.
854 - continuations have a saner syntax.
855 - no more memleaks.
856
857 0.07 Tue Jul 17 17:40:18 CEST 2001
858 - release candidate 1 ;)
859 - slightly nicer code.
860 - fixed a scheduling bug in Coro::Event.
861
862 0.06 Tue Jul 17 04:23:24 CEST 2001
863 - ok, I found the showstopper - the same sub must not be
864 re-entered in two different coroutines, otherwise => crash. I
865 see no easy solution to this problem, except by walking the call
866 chaing and saving/restoring the cv's, which is what I do now.
867 - memory leaks still latent, especially at thread termination.
868 - Coro::Event now works (haha).
869
870 0.05 Sun Jul 15 17:32:20 CEST 2001
871 - fixed a few issues in Coro::Event.
872 - I forgot to include Coro::Event in 0.04 :(:(:(
873
874 0.04 Sun Jul 15 05:24:59 CEST 2001
875 - @_ is now properly localized.
876 - Coro::State is now easier subclassable.
877 - Coro::Cont now coroutine-aware.
878 - Coro::Specific is a low-overhead module to create
879 coroutine-specific vars.
880 - Coro::Event provides a simple interface to Event.
881
882 0.03 Fri Jul 13 14:51:52 CEST 2001
883 - transfer() now implemented in XS (beware).
884 - new module Coro::Cont for really faked continuations.
885 - big internal architecture changes: Coro::State is now
886 really low-level and can thus be used to implement other
887 interesting things, While "Coro::" implements a process-like
888 model. Still crude and subject to change.
889 - $_ and $@ are no longer being localized.
890
891 0.02 Tue Jul 10 01:38:17 CEST 2001
892 - implemented "async" attribute.
893 - $_ and $@ are now localized.
894 - added Coro::Channel.
895 - more testcases, still no docs.
896
897 0.01 Tue Jul 3 02:18:41 CEST 2001
898 - original version; copied from Convert::Scalar.
899