ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/Coro/Changes
Revision: 1.446
Committed: Thu Nov 20 03:11:29 2008 UTC (15 years, 6 months ago) by root
Branch: MAIN
Changes since 1.445: +2 -1 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. The overhead for terminating an async_pool, coro over
61 a normal async is now very small.
62 - sped up coroutine creation/destruction by 40%.
63 - forgot to include Coro/libcoro/README in the dist for all these years.
64 - work around a freebsd pthreads bug (manual testcancel is required as
65 pthread_cond_wait isn't a cancellation point on freebsd).
66 - use new rouse functions to speed up and simplify Coro::BDB.
67 - make "prefer perl native functions" work with threaded perls.
68 - condense Coro::Debug ps output, hint at v and w flags.
69 - (libcoro) lots of minor cleanups and portability improvements.
70
71 4.913 Sat Nov 15 07:58:28 CET 2008
72 - async_pool did free a scalar value twice
73 ("Attempt to unreference...").
74
75 4.912 Thu Nov 13 18:31:23 CET 2008
76 - minor cleanups.
77 - use much larger stacks on linux and perl < 5.8.8.
78 - Coro::Debug::new_unix_server did not unlink the socket
79 when destroyed.
80
81 4.911 Tue Nov 11 04:26:17 CET 2008
82 - "port" to threaded perls.
83
84 4.91 Mon Nov 10 05:36:38 CET 2008
85 - the ->throw exception object no longer leaks.
86 - creating a new cctx leaked a scopestack entry (memleak).
87 - new coroutines didn't get created with a zero flags field
88 (unknown impact).
89 - calling ->throw on a not-yet-started coroutine should now work
90 instead of being ignored.
91 - ->throw is now supported on Coro::State objects.
92 - clean up cctx creation code a bit.
93 - entersub is actually a UNOP, not a LOGOP (not a bugfix).
94
95 4.9 Sat Nov 8 17:45:27 CET 2008
96 - (libcoro) did not preserve rbp with CORO_ASM (we are getting there).
97 - (libcoro) no longer leak threads in the experimental pthread backend,
98 also speed it up considerably.
99 - (libcoro) do not rely on makecontext passing void *'s unscathed.
100 - fix compiletime dependencies on libcoro in the Makefile.
101 - cctx_count wasn't always updated properly.
102 - Coro::State::cctx_stacksize wasn't applied correctly.
103 - new function Coro::State::cctx_max_idle.
104 - the default max number of idle C contexts is now 4.
105 - (libcoro) try harder to get _setjmp/_longjmp.
106 - (libcoro) cleanup and extend the libcoro API to officially
107 allow the creation of empty source contexts.
108 - very experimental workaround for the totally broken netbsd platform.
109 - tried to hack around openbsd bugs.
110
111 4.804 Wed Nov 5 16:36:00 CET 2008
112 - Coro::Debug::new_unix_server would not create a non-blocking listening
113 socket, sometimes causing freezes.
114 - (libcoro) fix misaligned stack points for setjmp and assembly
115 methods, which can cause crashes on x86/x86_64 with a sufficiently
116 aggressive compiler.
117 - new function: Coro::Debug::new_tcp_server.
118 - move ->throw into the Coro class because it only works on coro objects.
119
120 4.803 Mon Nov 3 17:16:12 CET 2008
121 - (libcoro) use a global asm statement to become independent of gcc
122 otpimisations for CORO_ASM (thanks to pippijn for the idea).
123 - try to workaround yet another broken bsd, this time dragonfly.
124
125 4.802 Thu Oct 30 10:56:12 CET 2008
126 - support -fno-omit-frame-pointer on x86 with the assembly method.
127 - tune 01_unblock.t tests a bit.
128
129 4.801 Wed Oct 22 18:33:37 CEST 2008
130 - improve readline speed for very long "lines".
131 - backport to 5.8.8.
132
133 4.8 Thu Oct 2 13:34:40 CEST 2008
134 - new function Coro::AIO::aio_wait.
135 - Coro.:AIO and Coro::BDB now "use Coro::AnyEvent".
136 - greatly speed up and reduce memory usage of Coro::AIO requests.
137 - implement some other µ-optimisations.
138
139 4.749 Mon Sep 29 14:40:12 CEST 2008
140 - port to slow and broken pseudo-threaded perls. (courtesy pippijn).
141
142 4.748 Sat Sep 27 14:03:19 CEST 2008
143 - implement, but do not document, PerlIO::cede(granularity).
144 - Coro::Storable forgot to wrap Storable::pstore.
145 - work around the multitude of leaks and memory corruption
146 bugs in PerlIO::via by using our own C-level perliol. As a side
147 effect, Coro::Storable is now much, much, much faster.
148
149 4.747 Tue Sep 23 01:59:41 CEST 2008
150 - fix a per-cv memleak (one empty array was leaked per
151 code reference).
152 - avoid a crash in coro->call|rss when the coroutine was already
153 destroyed (most noticably when using Coro::Debug::ps :)
154 - also protect *Storable::FILE.
155 - push up default storable granularity to 20ms.
156
157 4.746 Sun Sep 21 03:22:20 CEST 2008
158 - be more insistent on locking Storable against reentrancy
159 in Coro::Storable.
160 - move swap_def?v and throw to Coro::State, as documented.
161
162 4.745 Thu Jul 24 00:14:38 CEST 2008
163 - remove debugging code related to MgPV_nolen_const, also try to
164 make it compile with perl 5.8.6 (yes, apple apparently loves
165 outdated software). Reported by John S.
166
167 4.744 Tue Jul 8 22:06:35 CEST 2008
168 - correctly provide default DIE/WARN handlers as documented.
169 - also overwrite PL_vtbl_sigelem.svt_clear, even though current
170 implementations inside perl work fine for us.
171
172 4.743 Mon Jun 16 00:21:57 CEST 2008
173 - when using Coro::EV without running EV::loop it could
174 result in busy-waiting for events, this has been fixed.
175 - reduce codesize and improve performance by using EV_DEFAULT_UC.
176
177 4.742 Sat May 31 14:10:21 CEST 2008
178 - implement a workaround for (some) perl <5.8.8 versions.
179 - require EV 3.3+.
180
181 4.741 Fri May 30 23:33:09 CEST 2008
182 - tell netbsd how utterly broken their imitation of an OS is
183 and refuse to build by default if pthreads are in use.
184 - switch to "s" method on all bsd's by default, as their ucontext
185 stuff seems just too broken.
186 - fix a bug in Coro::Select.
187
188 4.74 Thu May 29 20:05:31 CEST 2008
189 - do not test Coro::LWP for lack of dependencies.
190 - do not test Coro::Select for lack of working perls.
191
192 4.73 Thu May 29 2008
193 - fix a bug in Coro::EV which would cause it to block despite
194 there being runnable coroutines.
195 - sprinkle "no warnings" freely over everything, also suppress
196 warnings for some other modules.
197 - fix typo in WSAEWOULDBLOCK.
198
199 4.72 Sun May 25 05:14:36 CEST 2008
200 - tweak META.yaml a bit, unfortunately, there is no documented way
201 to have optional dependencies with CPAN. doh :(
202 - avoid running some tests on windows because they would fail due to
203 perl bug (broken fork, broken pipes...).
204 - work around perl on windows bugs where perl returns undocumented
205 error codes for sysread, syswrite etc. by taking advantage
206 of AnyEvent's workaround for that problem.
207 - use AnyEvent::Util::fh_nonblocking in Coro::Handle to work around
208 a common perl implementation bug on windows.
209 - use unix domain sockets in testsuite to work around
210 common perl implementation bugs on widows (they are emulated by
211 tcp sockets on windows. ugh.)
212
213 4.71 Sat May 24 20:01:27 CEST 2008
214 - fix a bug in Coro::AnyEvent ("Usage: Coro::AnyEvent::_schedule()").
215 - take advantage of async name resolution of AnyEvent::Util.
216 - work around brutal inet_aton override in Coro::LWP.
217 - take advantage of the readyhook in Coro::EV, for smoother
218 scheduling.
219
220 4.7 Sun May 11 00:32:19 CEST 2008
221 - completely reworked the Coro manpage.
222 - added Coro::AnyEvent, generic event loop integration.
223 - implement cancel, ready and kill commands in Coro::Debug.
224 - document find_coro in Coro::Debug.
225 - incompatible change: rename has_stack to has_cctx.
226 - Coro::AIO and Coro::BDB no longer force event model detection,
227 use AnyEvent::AIO and AnyEvent::BDB.
228
229 4.6 Sat Apr 26 10:05:14 CEST 2008
230 - INCOMPATIBLE CHANGE: sub/code attributes are no longer supported
231 by the Coro module. It was a mistake to have it in the first place.
232 - (experimental) support for activestate perl 5.10 (method "w").
233 - (experimental) support for strawberry perl (method "a").
234 - coro_sigelem_set did not return a value although it had to,
235 actual impact unknown.
236
237 4.51 Mon Apr 14 13:28:27 CEST 2008
238 - make it compile again on 5.8.
239
240 4.50 Thu Apr 10 09:43:17 CEST 2008
241 - I did it twice! (see 4.49).
242
243 4.49 Sun Apr 6 21:23:31 CEST 2008
244 - grr, instead of compiling the recent changes on 5.10 only they
245 were compiled on 5.8 only.
246
247 4.48 Sun Apr 6 20:36:46 CEST 2008
248 - allow coroutine switches during eval's under 5.10.x, as it seems
249 the parser is a per-interpreter option now, so this is safe
250 (this might fix the odd crashing bug).
251 - drop support for 5.9.x versions: they are dead, jim.
252
253 4.47 Sun Apr 6 00:37:52 CEST 2008
254 - force cctx allocation on API calls: we know we need to force one
255 and gcc actually manages to confuse our heuristic nowadays,
256 leading to crashes and worse.
257 - document force_cctx.
258
259 4.46 Fri Apr 4 22:05:43 CEST 2008
260 - upgrade libcoro, resulting in pthread-backend (which was only created
261 to fulfill the rules of the programming languages shootout).
262
263 4.45 Thu Mar 13 11:55:36 CET 2008
264 - fix a file leaking bug in eg/mhyttpd that would allow
265 downloading of any file (reported by oesi).
266 - fix deadlock bug in Coro::Channel (reported by Richard Hundt)
267 (also add testcase).
268 - support Broadcast option in Coro::Socket (patch by Richard Hundt,
269 apparently having loads of fun with that).
270 - implement $state->swap_defsv, ->swap_defav and document ->throw.
271
272 4.4 Wed Feb 13 16:44:29 CET 2008
273 - only assume .cfi pseudo isns exist on GNU/Linux.
274 - add get_timed method to Coro::Channels.
275 - fixed Coro::Debug manpage.
276 - perl 5.11 compatbility improvement based on patch by
277 Andreas König.
278
279 4.37 Sun Jan 20 11:25:23 CET 2008
280 - soften the check for an active parser for perl 5.10 (avoiding
281 "Coro::State::transfer called while parsing" in many cases).
282
283 4.36 Sun Jan 13 10:53:56 CET 2008
284 - reset diehook when terminating from an async_pool as to not
285 trigger any __DIE__ handlers.
286
287 4.35 Sun Jan 13 04:14:13 CET 2008
288 - "bt" debug command now displays any exceptions
289 from longmess and also skips the innermost
290 stackframes, giving more useufl output.
291 - allow backtraces in coroutines blocked in Coro::EV,
292 at a <1% speed hit.
293 - handle localising of $SIG{__WARN__} and __DIE__
294 properly (with a proper amount of dirty hacking).
295
296 4.34 Sat Dec 22 17:49:53 CET 2007
297 - upgrade to EV version 2.0 API.
298
299 4.33 Mon Dec 17 08:36:12 CET 2007
300 - make Coro::AIO etc. loadable in the absence of EV.
301
302 4.32 Mon Dec 17 07:46:02 CET 2007
303 - majorly improved Coro::Handle's performance with Coro::EV.
304 - implemented the until now mythical Coro::BDB module.
305 - specialcase EV in Coro::AIO and Coro::BDB for extra speed.
306
307 4.31 Wed Dec 5 12:32:39 CET 2007
308 - remove warn statement form Coro::Util.
309
310 4.3 Tue Dec 4 20:33:14 CET 2007
311 - calls to the idle function could cause stack corruption
312 when the stack changed.
313 - do no longer rely on the presence of EV::DNS if EV is used
314 (because it is gone), but instead take avdantage of EV::ADNS
315 if available.
316 - add ($) prototypes to all functions in Coro::Storable.
317 - use a conventional (safer) idle callback in Coro::EV.
318 - do accept quickly in Coro::Debug to avoid endless loops.
319
320 4.22 Fri Nov 30 16:04:04 CET 2007
321 - really use optimised versions for Event and EV in Coro::Util
322 and Coro::Handle.
323
324 4.21 Sun Nov 25 10:48:59 CET 2007
325 - fix a spurious memory read.
326 - Coro::EV no longer keeps the eventloop "alive".
327
328 4.2 Fri Nov 9 20:47:05 CET 2007
329 - enable/disable tracing from a new coroutine, not a pooled one.
330 - fix a memleak in Coro::Event.
331 - removed killall call from fork_eval.
332 - made sure store_fd is already loaded so that fork_eval
333 does not have to parse autoload in each subprocess.
334 - only use assembly method if -O switch is in $Config{optimize}.
335 - add (optional) Coro::EV module, so far the best event loop module
336 directly supported by Coro.
337 - if the event model is EV, use EV::DNS to resolve
338 stuff in Coro::Util.
339 - don't get confused by multiple event notifications in Coro::Handle.
340 - initial support for EV (libevent interface).
341 - require Event and EV using configure_requires, to force their existance.
342
343 4.13 Wed Oct 24 07:26:45 CEST 2007
344 - add Coro::Storable::blocking_thaw.
345 - use a vastly more complicated technique to localise
346 $SIG{__WARN/DIE__} that also works on perls <= 5.8.8.
347 - use a coroutine for the idle callback Coro::Event,
348 instead of running Event in the current coroutine context.
349 This also catches recursive invocations.
350 - actually report fork errors in gethostbyname and inet_aton.
351
352 4.11 Thu Oct 11 02:40:24 CEST 2007
353 - port to threaded perls.
354
355 4.1 Thu Oct 11 02:38:16 CEST 2007
356 - incompatible change: $SIG{__DIE__} and $SIG{__WARN__} will now
357 be local to each coro (see Coro::State).
358 - incompatible change: for very deep reasons, cede and cede_notself
359 cannot return anything, so nothing will be returned.
360 - possibly bring back 5.10 compatibility (untested).
361 - work around stupid (and wrong) warning on 5.10 :(.
362 - overlay the saved state over the context stack. This saves
363 a few hundred bytes per coroutine on average and also
364 speeds up context switching a bit.
365 - further tune default stack sizes.
366 - (more) correctly calculate stack usage in coro_rss.
367 - Coro::Storable::blocking_* did not properly lock
368 resulting in races between coroutines.
369 - added Coro::Storable::guard.
370 - stopping to trace a coroutine could destroy the cctx of
371 an unrelated coroutine.
372 - explain the relationship between Perl and C coroutines in
373 more detail in Coro::State.
374 - Coro::Util::inet_aton did not short-circuit dotted quad forms,
375 causing a fork per resolve. This also affected Coro::Socket.
376 - switch to a separate stack in $coro->call/eval to avoid
377 invalidating pointers.
378
379 4.03 Sat Oct 6 21:24:00 CEST 2007
380 - added Coro::throw method.
381 - minor code cleanups.
382
383 4.02 Sat Oct 6 02:36:47 CEST 2007
384 - fix a very minor per-coroutine memleak (a single codereference).
385 - fixed a bug where the currently in-use c context would be freed
386 prematurely (can happen only when programs change the stacksize
387 or use tracing).
388 - tracing can no longer keep a coro alive after it terminated.
389 - do static branch prediction in the common path for gcc. gives
390 about 2-5% speed improvement here.
391
392 4.01 Fri Oct 5 22:10:49 CEST 2007
393 - instead of recreating *a* standard output handle we simply
394 use STDOUT, which is faster and hopefully more robust.
395
396 4.0 Fri Oct 5 12:56:00 CEST 2007
397 - incompatibly changed Coro::Storable::freeze.
398 - major new feature: added Coro::Debug, for interactive coroutine
399 debugging, tracing and much more.
400 - major bug fix: unbelievable, but true: $_, $/ and many other
401 "saved" variables actually weren't being saved. This has been fixed,
402 of course, while increasing performance while losing all the save
403 flags.
404 - save flags are gone, and all the api functions dealing with them.
405 - added Coro::Semaphore::adjust.
406 - added Coro::Util::fork_eval.
407 - added Coro::Storable::{nfreeze,blocking_{freeze,nfreeze}}.
408 - added Coro::killall.
409 - reduce initial stack sizes to allow for "micro-coroutines".
410 - better async_pool resource management, moved parts of async_pool
411 handling to XS (major speed improvement).
412 - actually croak before modifying important data structures.
413 - refuse to transfer while compiling.
414 - possibly support eval EXPR better now.
415 - enable assembly per default on linux+bsd x86+amd64.
416 - all internal members were renamed _something for easier subclassing.
417 - many minor tweaks.
418
419 3.63 Wed May 16 14:10:06 CEST 2007
420 - implement handcoded assembly for x86/amd64 SVR ABI.
421
422 3.62 Fri Apr 27 21:36:06 CEST 2007
423 - upgrade libcoro (which might set unwind info correctly).
424 - change default on linux to setjmp/longjmp.
425
426 3.61 Thu Apr 19 12:36:18 CEST 2007
427 - Coro::Storable caused an endless loop when thawing invalid
428 pst's sometimes.
429 - use a Semaphore in Coro::Storable, as Storable doesn't
430 seem to be reentrant (although it is documented to
431 be threadsafe...).
432 - fix Coro::Signal to bring back the original unreliable
433 but stateful semantics.
434 - fixed a lot of typos in Coro.pm (patch submitted by David
435 Steinbrunner, which applied flawlessly).
436
437 3.6 Sat Apr 14 17:13:31 CEST 2007
438 - added some bugfixes to get eg/myhttpd working again.
439 - added Coro::Storable for often-cede'ing freeze/thaw.
440 - try to do a clean exit when a coroutine calls exit
441 (EXPERIMENTAL).
442 - got rid of indirect call through _coro_init.
443 - updated the partly antique examples in eg/ to
444 work again and be a bit less magic, too.
445 - fixed Coro::Signal semantics to work as documented again.
446
447 3.55 Sun Mar 25 01:20:47 CET 2007
448 - add SAVE_DEFFH to save the default filehandle and enable
449 it by default.
450 - finally move socket-operations from Coro::Socket to Coro::Handle
451 to be able to unblock foreign sockets.
452 - add Coro::State::save_also and guarded_save.
453 - add count accessor to Coro::Semaphore.
454 - add Coro::State::cctx_stacksize.
455 - just for the fun of it, do not rely on implicit context,
456 which can dramatically improve performance, but people
457 using windows-process-emulation perls do not care much about
458 performance.
459
460 3.51 Sun Mar 4 14:18:04 CET 2007
461 - fixed a problem when you weakened references to Coro::State's
462 (patch by Yuval Kogman).
463
464 3.501 Wed Feb 28 12:44:07 CET 2007
465 - rename some global symbols as macosx from hell redefines
466 them without asking.
467
468 3.5 Tue Feb 13 20:22:53 CET 2007
469 - do AnyEvent model detection earlier, avoiding problems
470 caused by first using AnyEvent and later Coro::Event.
471 - implement and document Coro::Event event objects.
472 - fix a potential problem in Coro::Event causing crashes.
473 - initialise PL_comppad when creating a new coroutine,
474 avoids crashes on early coro destruction.
475
476 3.41 Mon Jan 22 19:19:49 CET 2007
477 - readline on Coro::Handle did not support undefined $/,
478 nor did it deliver partial lines on EOF or error.
479 - implement malloc fallback for stack allocation because
480 stupid broken idiotic OSX has a stupid broken
481 idiotic fits-the-whole-os mmap "implementation" and
482 my dick feels longer if Coro is portable even to
483 obsolete platforms.
484
485 3.4 Fri Jan 19 21:52:54 CET 2007
486 - remove t/09_timer.t, as it isn't really testing much
487 but was rather flaky in practise.
488 - async_pool coro would keep arguments and callback alive until
489 it was reused.
490 - cancellation of a coroutine could cause spurious idle calls
491 in cede_notself.
492
493 3.3 Sat Jan 6 03:45:00 CET 2007
494 - implement $coro->on_destroy.
495 - Coro::Event blocking semantics have been changed,
496 documented and - hopefully - improved.
497 - fix nice adding, not subtracting, from priority.
498 - fix ->prio and api_is_ready (patch by Mark Hinds).
499 - fixed an assert ("... == prev__cctx->idle_te")
500 that could errronously trigger.
501 - fix various large and small memleaks.
502 - use a (hopefully) more stable cancel implementation
503 that immediately frees the coroutine data.
504 - cede/cede_notself return a status now.
505 - added Coro::guard function.
506 - added a global coroutine pool for jobs (on my machine,
507 I can create and execute 48k simple coros/s with async,
508 and 128k coros with async_pool).
509 - Coro::AIO now uses the coroutine priority as io priority.
510
511 3.2 Fri Dec 22 05:07:09 CET 2006
512 - improve portability to slightly older perls.
513 - use cleaner coroutine destruction.
514 - simplify configuration for users.
515 - optionally (unrecommended) prefer perl functions over
516 their coro replacements.
517
518 3.11 Tue Dec 5 13:11:24 CET 2006
519 - fixed some bogus assert's, but as perl.h disables assert even
520 without NDEBUG (thank you very much), not too many people should
521 notice (that did include myself). Andreas König noticed, though :)
522 - do not save/restore PL_sortcxix on >= 5.9.x, it doesn't seem to have
523 it. Also noticed by Andreas König :)
524 - save/restore tainted status.
525 - verified to pass the testsuite on my 5.9.5.
526
527 3.1 Mon Dec 4 23:03:40 CET 2006
528 - INCOMPATIBLE CHANGE: $/ is now per-coroutine (but slow).
529 - incompatible change: transfer flags are now per-state.
530 - give a better error message on deadlock.
531 - document Coro::nready.
532 - enhanced testsuite.
533
534 3.01 Sun Dec 3 23:47:42 CET 2006
535 - forgot to include Coro::Timer.
536
537 3.0 Sun Dec 3 22:57:25 CET 2006
538 - the "FINALLY COMPLETELY STABLE" release (coming soon:
539 the "FAMOUS LAST WORDS" release).
540 - implement a new stack sharing algorithm, which uses a stack
541 pool (size currently hardcoded).
542 - make stack sharing mandatory (it no longer uses a heuristic).
543 - eval/die no longer cause weird problems under heavy use.
544 - Coro::Event could cause livelocks if it was used but
545 no Coro::Event watchers were used.
546 - Coro::Event now uses asynccheck as prepare does not
547 check for changed watchers.
548 - Coro::Event allows multiple waiting coros and will wake up one
549 per event.
550 - Coro::Event should be cleaner and more efficient now.
551 - new utility function Coro::unblock_sub.
552 - document the sad fact that Event is no longer reentrant.
553 - putting a coroutine into the ready queue twice could under
554 some circumstances lead to stack corruption.
555 - minor incompatibility: subclassing Coro::State is supported
556 directly now without going through a _coro_state member.
557 - state/coro switching is much faster now.
558 - very minor optimisations and code/documentation cleanup.
559 - avoid problems due to compiler inlining.
560 - removed timers from Coro::Timer -> use AnyEvent instead.
561 - replaced Coro::idle coroutine by (cleaner) idle handler.
562 - updated to newest libcoro.
563 - implement enhanced support for valgrind.
564 - implement is_ready and return value for ready.
565 - removed Coro::Cont, it was a misnomer (it's generators),
566 and it was somewhat annoying to get it right. Will come back
567 if somebody asks for it :)
568 - many ->wait methods and Coro::Event could return
569 spuriously without the event having happened.
570
571 2.5 Tue Nov 7 12:22:33 CET 2006
572 - made Coro::Util, Coro::Select, Coro::Handle and Coro::Socket
573 use AnyEvent, moved them to Coro/.
574 - added Coro::LWP which contains all the uglyness required to
575 make LWP non-blocking.
576 - should work with perl 5.9.x now (Andreas König made me do it).
577 - fixed another bug in Coro::Select when the timeout was undef.
578 - reuse PL_start_env for all coros, saves some memory per coroutine.
579 - manage PL_top_env differently, hopefully to avoid panic: top_env.
580 - timeout argument was not properly used in Coro::Socket.
581 - allow limited forms of subclassing in Coro::Handle/Coro::Socket.
582 - emulate undocumented(!) functionality of IO::Socket required
583 by LWP(!!).
584 - updated eg/lwp to work with newer lwp's.
585 - remove "FATAL: uncaught exception" prefix. Coroutines that die
586 kill the whole process, just as exceptions in the main "coroutine"
587 did already.
588
589 2.1 Wed Nov 1 23:01:13 CET 2006
590 - fix a long-standing bug in Coro::Select where select with
591 zero timeout would instead change the current default filehandle.
592 - use a simpler and hopefully more robust way to clone padlists
593 (uses less memory and a perl function instead of our own).
594 - coro can now create a stack guard area on many architectures.
595 - Coro::AIO properly reexports additional functions from IO::AIO.
596 - updated libcoro with a workaround for OS X,
597 pach and testing by Michael Schwern.
598
599 2.0 Tue Oct 24 05:47:17 CEST 2006
600 - support additional aio requests in Coro::AIO.
601
602 1.9 (never properly released due to a glitch)
603
604 1.8 Thu Feb 2 00:59:06 CET 2006
605 - applied suggested patch by SAMV to avoid problems during stupid
606 mark & sweep gc run.
607 - applied patch by Scott Walters for 5.9.3 compatibility.
608
609 1.7 Tue Dec 27 01:41:58 CET 2005
610 - added Coro::AIO, a thin wrapper around IO::AIO.
611 - improved Makefile.PL explanations.
612
613 1.6 Mon Dec 12 21:30:05 CET 2005
614 - additionally save PM_curpm might fix as-of-yet
615 unseen problems with regex matches being attributed to the
616 wrong package.
617 - add t/10_bugs.t, which currently checks against imho broken
618 perls that use 0.26MB of stack space per Perl_magic_get
619 invocation (newer linuxes) as opposed to the 0.0002MB perl
620 normally uses.
621 - make stacksize configurable for the ultimate debian experience.
622
623 1.51 Mon Dec 12 18:48:36 CET 2005
624 - remove debugging warn() accidentally left in Coro::Select.
625
626 1.5 Tue Nov 29 13:32:44 CET 2005
627 - use Coro::Event inside Coro::Select to avoid spurious deadlocks.
628 - fix Coro::Select 'select'.
629 - strict'ify some modules.
630 - libcoro errornously restored the SIGUSR2 handler to SIGUSR1.
631 - use XSLoader in selected modules.
632 - remove some 5.6 compatibility cruft.
633
634 1.4 Tue Sep 6 00:11:05 CEST 2005
635 - libcoro did not take into account the trampoline on amd64,
636 when 'l'inux method was used.
637
638 1.31 Tue Aug 30 23:31:33 CEST 2005
639 - some portability fixes/workarounds.
640
641 1.3 Sat Aug 20 03:08:56 CEST 2005
642 - no code changes. module seems to work fine.
643 - improve Coro::State docs, remove reference to nonexisting
644 Coro::State::flush method.
645 - no longer autodetect windows, present it as an option instead.
646
647 1.2 Mon May 16 02:00:55 CEST 2005
648 - included libcoro.c earlier in State.xs, might improve portability.
649 - use a faster and possibly more stable Coro::Cont implementation.
650 - accept x86_64 in addition to amd64 for optimized linux-amd64 support.
651 - fix bugs Coro::SemaphoreSet that could cause locks never to be freed.
652 - fix bugs in CoroAPI.h and document it in Coro::MakeMaker.
653
654 1.11 Thu Mar 3 18:00:52 CET 2005
655 - change of contact address.
656
657 1.1 Tue Feb 22 20:51:16 CET 2005
658 - support [l]inux method on amd64.
659 - allow some context switches while compiling/eval'ing.
660
661 1.0 Tue Aug 10 03:47:27 CEST 2004
662 - Event 0.89 fixes the bug again.
663 - fix the remaining(??) memleak. I didn't free the cache-AV for
664 padlists, so one AV per created anon-sub leaked.
665
666 0.97 Fri May 14 15:23:32 CEST 2004
667 - remove stupid gcc extension that was an experiment only anyway.
668
669 0.96 Thu May 13 18:09:29 CEST 2004
670 - tell people if they have yet another broken Event variant.
671 - made Coro to work _better_ with threads (it seems that, as long
672 as you use Coros in one thread only everything will work fine.
673 Looking at the perl source the problem might be missing locking
674 between Coro and some perl internal routines. That's not easily
675 fixable).
676 - (hopefully) got rid of the static xnext inside transfer.
677
678 0.95 Sun Feb 15 03:20:28 CET 2004
679 - removed Coro::State::flush, implemented a new and much
680 faster caching mechanism by attaching magic to the CV.
681 - a memleak with (real) closures remains.
682 - dramatically reduced size of stacks to 65536 (linux/x86), after
683 observing that even myhttpd never had more than a kilobyte
684 of stack in use (stack is still only physically allocated in
685 pagesize increments for systems suporting mmap).
686 - fix for a "cygwin-64" architecture added, proposed by Stefan Traby.
687 - fix for threaded perl.
688
689 0.9 Sun Nov 23 04:32:39 CET 2003
690 - fix a few very hard to track down but (of course) stupid
691 memory leaks.
692 - stringify version number to avoid locale problems :(.
693 - perl 5.9.0 does not have SvPADBUSY, reported by Scott Walters,
694 hopefully ignoring it "just works".
695
696 0.8 Wed Nov 5 19:38:40 CET 2003
697 - port to perl5.005_03, but only 5.8.x+ is supported!
698 - honor LocalAddr even without LocalPort in Coro::Socket.
699
700 0.7 Tue May 27 03:12:38 CEST 2003
701 - the version jump indicates some level of testing, not gobs
702 of new features.
703 - uh... I found the prompt function in ExtUtils::MakeMaker.
704 Highly correct stuff, that is...
705 - fixed(?) a bug with die's in coroutines causing "panic: top_env".
706 the fix is not well understood by the author, so beware :(.
707
708 0.652 Thu May 8 02:54:46 CEST 2003
709 - Applied patch by Slaven Rezic to set default to "s" on FreeBSD,
710 cause version 4 doesn't have ucontext.
711 - Benjamin Reed reported that setjmp works fine on darwin,
712 so preselect it.
713
714 0.651 Sat Mar 29 15:00:23 CET 2003
715 - fix a bug in Coro::Handle where some bytes could get lost
716 on reading, reported by jason@nichego.net.
717
718 0.65 Sun Mar 23 00:08:26 CET 2003
719 - added a README.linux-glibc.
720 - new module Coro::Select.
721 - also save/restore PL_comppad, fixes segfaults with 5.8.1.
722 (I never claimed that I know what I am doing ;).
723 - set default to setjmp/longjmp on non-x86-linux.
724
725 0.6 Thu Nov 21 11:09:06 CET 2002
726 - made lazy context switching the default.
727 - fixed the bug where SemaphoreSet::guard
728 was the same as timed_guard and timed_guard was missing.
729 - fix a memleak in Coro::Timer.
730 - reclassified context sharing as not experimental.
731
732 0.534 Sun Apr 14 03:05:12 CEST 2002
733 - fixed backspace => tab in header parsing (myhttpd).
734 - added eg/readline, for elmex.
735 - Coro::Event now calls ->start on first call instead of
736 ->again, which created wrong timeouts for the first ->next.
737 - fixed a bug where Coro::Socket returned a closed socket instead
738 of nothing on ECONNREFUSED and similar error conditions.
739
740 0.533 Mon Feb 18 18:49:40 CET 2002
741 - adapt to finally working Event-0.85.
742 - fixed buggy event initialization (trapped under windows, but how
743 could it possibly have worked anywhere?)
744
745 0.532 Wed Jan 16 02:45:32 CET 2002
746 - removed the reference to pp_entersub, might work on
747 (some) win32 perls now (testsuite works under cygwin,
748 without Event, which freezes).
749
750 0.531 Mon Dec 10 22:18:44 CET 2001
751 - Fixed a bug in SemaphoreSet::timed_down.
752
753 0.53 Tue Nov 27 21:11:13 CET 2001
754 - some tests for Event.
755 - slightly different internal architecture to get refcnt into the object
756 and not the reference. might fix a few bugs, certainly did introduce
757 new ones.
758 - Coro::Timer (independent of Event).
759 - new timed_wait functions for Coro::Signal, Semaphore, SemaphoreSet.
760
761 0.52 Tue Nov 6 21:36:18 CET 2001
762 - ported to cygwin (trivial).
763 patch by Gerrit P. Haase <gp@familiehaase.de>.
764 - small setjmp code fix by Sullivan.DanielJ@epamail.epa.gov.
765
766 0.51 Thu Nov 1 20:39:01 CET 2001
767 - terminate/cancel now work properly (otherwise termination
768 could cause a "next coroutine is not and contains not..."
769 error.
770 - added Coro::Socket::shutdown.
771 - Coro::Event::loop no is the same as Event::loop.
772 - implemented terminate with args + join.
773
774 0.5 Fri Sep 28 16:15:35 CEST 2001
775 - fixed "print" on a Coro::Handle. The print method worked.
776 - small tweaks (seem to reduce memory consumption a lot)
777 in various modules.
778 - splendid use of "no warnings" scattered throughout.
779 - added Coro::Handle::rbuf, fixed Coro::Handle::fh.
780
781 0.49 Sun Sep 16 02:42:45 CEST 2001
782 - changed some method calls to function calls for speed
783 inside Coro::Handle.
784 - make Coro::Handle use an array instead of a hash for
785 speed reasons.
786 - IRIX mystery solved: it's SGI's NT, after all: "standard, huh?".
787 sjlj and ucontext should now work.
788 - IRIX-specific port for libcoro.
789 - swapped order of accept results to match IO::Socket.
790 - changed getsock/peername to sock/peername to match IO::Socket.
791 - fixed a bug that caused segfault when returning to main task
792 under some circumstances.
793 - other bugfixes.
794
795 0.45 Sun Sep 2 02:54:01 CEST 2001
796 - new method Coro::Handle::timeout.
797 - corrected speling of set...name to get...name in Coro::Socket.
798 - Coro::Socket::accept now returns a Coro::Socket, not a Coro::Handle.
799 - Coro::Handle now supports fileno and FILENO.
800 - added eg/myhttpd, a web server.
801 - Coro::Socket now accepts numeric ports as well as "port(number)"
802 syntax.
803 - moved some scheduling primitives into xs code. more to come.
804 - new simple priority system.
805 - implemented Coro::Event using XS, almost four times faster!
806 - small memory corruption problem fixed (boy that was difficult).
807
808 0.13 Wed Aug 8 16:53:07 CEST 2001
809 - new method Coro::Handle::readline.
810 - added eg/lwp and Coro::Handle::autoflush, to make LWP work.
811 (see eg/lwp on how to make LWP non-blocking).
812 - renamed Coro::Socket::new_inet to new.
813 - added Coro::Util, some utility functions.
814 - die/eval now works better (still get annoying (but true ;)
815 "Callback called exit" messages).
816 - Coro::Handle now supports timeout, and Coro::Socket Timeout.
817 - much better Coro::Socket support.
818
819 0.12 Fri Jul 27 04:19:01 CEST 2001
820 - do not use mmap if neither MAP_ANON nor MAP_ANONYMOUS is defined.
821 - very experimental stack sharing algorithm. not 100% safe but
822 should work well in practise ;)
823 - added Coro::RWLock.
824 - Coro::Specific now works.
825
826 0.11 Tue Jul 24 22:49:21 CEST 2001
827 - added specialized hack for newer and older linux versions (fast).
828 - renamed Coro::Event::IO to Coro::Handle.
829 - new module Coro::Socket.
830
831 0.1 Tue Jul 24 01:47:53 CEST 2001
832 - release candidate 3. A rather new internal structure :(
833 - the great renaming: Damian Conway gave me a suitable replacement
834 for yield.
835 - added Coro/libcoro, a portable coroutine implementation for C
836 (not even perl-dependent!!!), which can be used by Coro::State.
837 - renamed SAVE_DEFSV (on xs-level only) to avoid symbol clash in
838 perl-5.6.
839 - new function Coro::Event::idle.
840 - the idle process is now overriden
841 by default in Coro::Event.
842 - Coro::Channel now enforces the size.
843 - canceling events now works in all cases.
844 - Coro state now include $_ and $@.
845 - yet another bug workaround that I do not really understand :(
846 - new module Coro::Event::IO (very undocumented) to do
847 non-blocking i/o.
848 - performance tuning.
849
850 0.08 Thu Jul 19 06:13:25 CEST 2001
851 - release candidate 2 (new functionality).
852 - add Coro::State::flush function.
853 - transfer now supports a flags value (mostly for speed ATM).
854 - might compile and work in the presence of threads now.
855 - continuations have a saner syntax.
856 - no more memleaks.
857
858 0.07 Tue Jul 17 17:40:18 CEST 2001
859 - release candidate 1 ;)
860 - slightly nicer code.
861 - fixed a scheduling bug in Coro::Event.
862
863 0.06 Tue Jul 17 04:23:24 CEST 2001
864 - ok, I found the showstopper - the same sub must not be
865 re-entered in two different coroutines, otherwise => crash. I
866 see no easy solution to this problem, except by walking the call
867 chaing and saving/restoring the cv's, which is what I do now.
868 - memory leaks still latent, especially at thread termination.
869 - Coro::Event now works (haha).
870
871 0.05 Sun Jul 15 17:32:20 CEST 2001
872 - fixed a few issues in Coro::Event.
873 - I forgot to include Coro::Event in 0.04 :(:(:(
874
875 0.04 Sun Jul 15 05:24:59 CEST 2001
876 - @_ is now properly localized.
877 - Coro::State is now easier subclassable.
878 - Coro::Cont now coroutine-aware.
879 - Coro::Specific is a low-overhead module to create
880 coroutine-specific vars.
881 - Coro::Event provides a simple interface to Event.
882
883 0.03 Fri Jul 13 14:51:52 CEST 2001
884 - transfer() now implemented in XS (beware).
885 - new module Coro::Cont for really faked continuations.
886 - big internal architecture changes: Coro::State is now
887 really low-level and can thus be used to implement other
888 interesting things, While "Coro::" implements a process-like
889 model. Still crude and subject to change.
890 - $_ and $@ are no longer being localized.
891
892 0.02 Tue Jul 10 01:38:17 CEST 2001
893 - implemented "async" attribute.
894 - $_ and $@ are now localized.
895 - added Coro::Channel.
896 - more testcases, still no docs.
897
898 0.01 Tue Jul 3 02:18:41 CEST 2001
899 - original version; copied from Convert::Scalar.
900