ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/Coro/Changes
Revision: 1.448
Committed: Thu Nov 20 07:02:43 2008 UTC (15 years, 6 months ago) by root
Branch: MAIN
Changes since 1.447: +0 -2 lines
Log Message:
*** empty log message ***

File Contents

# Content
1 Revision history for Perl extension Coro.
2
3 TODO: should explore PerlIO::coroaio (perl leaks like hell).
4 TODO: maybe implement a default message channel, very much like Erlang's
5 actor model (which is cool in a lot of important aspects (failures!),
6 but very lacking in others (higher level ipc)).
7
8 5.0
9 - NEW ARCHITECTURE: use the latest 4.x version if you experience
10 stability issues.
11 - bump API version to 7 - all dependents must be recompiled.
12 - removed timed_* functions - they were not being used anyways
13 and should be replaced by a more generic mechanism -
14 and were annoying to support anyways :)
15 - removed SemaphoreSet's waiter method - use sem instead.
16 - Coro::Semaphore->adjust didn't correctly wake up enough waiters.
17 - async_pool did free a scalar value twice
18 ("Attempt to unreference...").
19 - fix a disastrous bug in the readline optimisation
20 introduced in 4.801.
21 - fix a longstanding bug where calling terminate on a coro that
22 was waiting for a semaphore that was just becoming available
23 would cause a deadlock (semaphore would get into a state where
24 it was available but waiters were still blocked).
25 - calling throw on a coroutine that is waiting for a semaphore will
26 no longer make it acquire the semaphore (and thus leak a count).
27 - perl's process emulation is now not even theoretically supported
28 anymore.
29 - new functions Coro::rouse_cb and Coro::rouse_wait for easier
30 conversion of callback-style to blocking-style.
31 - new methods $coro->schedule_to and ->cede_to, to specifically
32 schedule or cede to a specific coroutine.
33 - new function Coro::Semaphore::wait.
34 - use named constants in Coro::Channel (Richard Hundt).
35 - directly patch the entersub opcode calling SLF functions (cede,
36 transfer and so on). this does speed up context switching, but
37 more importanly, it frees us from the hardcoded behaviour of
38 entersub, so we might actually be able to return something from
39 those functions and atcually create new ones.
40 - take advantage of __builtin_frame_address on gcc.
41 - expose THX in coroapi (not sure whether this was a wise decision,
42 as "threaded" perls are running at half speed anyways).
43 - implement execute_slf (schedule-like-function) interface that makes
44 it possible to implement schedule-like-functions in XS.
45 - use new SLF interface to massively speed up Coro::EV by roughly a
46 factor of two.
47 - used new SLF interface to massively speed up Coro::Semaphore by a
48 factor of three.
49 - used new SLF interface to speed up Coro::AIO by roughly a factor of
50 four and reduce its memory usage considerably.
51 - implement Coro::SemaphoreSet purely in terms of Coro::Semaphore,
52 for a nice speedup and vastly more correct behaviour. Also implement
53 a new method "sem" to get at the underlying semaphore object.
54 - implement Coro::Channel in terms of Coro::Semaphore, for a moderate
55 (in comparison) 20-40% speedup.
56 - used new SLF interface to reimplement Coro::Signal gaining
57 some unknown (because I was too lazy), but certain, speedup, and also
58 making signals reliable for the first time.
59 - used new SLF interface and other optimisations to speed up async_pool
60 by a factor of two. It also doesn't rely on perl's exception mechanism
61 to exit anymore. The overhead for terminating an async_pool, coro over
62 a normal async is now very small.
63 - sped up coroutine creation/destruction by 40%.
64 - forgot to include Coro/libcoro/README in the dist for all these years.
65 - work around a freebsd pthreads bug (manual testcancel is required as
66 pthread_cond_wait isn't a cancellation point on freebsd).
67 - use new rouse functions to speed up and simplify Coro::BDB.
68 - make "prefer perl native functions" work with threaded perls.
69 - condense Coro::Debug ps output, hint at v and w flags.
70 - (libcoro) lots of minor cleanups and portability improvements.
71
72 4.913 Sat Nov 15 07:58:28 CET 2008
73 - async_pool did free a scalar value twice
74 ("Attempt to unreference...").
75
76 4.912 Thu Nov 13 18:31:23 CET 2008
77 - minor cleanups.
78 - use much larger stacks on linux and perl < 5.8.8.
79 - Coro::Debug::new_unix_server did not unlink the socket
80 when destroyed.
81
82 4.911 Tue Nov 11 04:26:17 CET 2008
83 - "port" to threaded perls.
84
85 4.91 Mon Nov 10 05:36:38 CET 2008
86 - the ->throw exception object no longer leaks.
87 - creating a new cctx leaked a scopestack entry (memleak).
88 - new coroutines didn't get created with a zero flags field
89 (unknown impact).
90 - calling ->throw on a not-yet-started coroutine should now work
91 instead of being ignored.
92 - ->throw is now supported on Coro::State objects.
93 - clean up cctx creation code a bit.
94 - entersub is actually a UNOP, not a LOGOP (not a bugfix).
95
96 4.9 Sat Nov 8 17:45:27 CET 2008
97 - (libcoro) did not preserve rbp with CORO_ASM (we are getting there).
98 - (libcoro) no longer leak threads in the experimental pthread backend,
99 also speed it up considerably.
100 - (libcoro) do not rely on makecontext passing void *'s unscathed.
101 - fix compiletime dependencies on libcoro in the Makefile.
102 - cctx_count wasn't always updated properly.
103 - Coro::State::cctx_stacksize wasn't applied correctly.
104 - new function Coro::State::cctx_max_idle.
105 - the default max number of idle C contexts is now 4.
106 - (libcoro) try harder to get _setjmp/_longjmp.
107 - (libcoro) cleanup and extend the libcoro API to officially
108 allow the creation of empty source contexts.
109 - very experimental workaround for the totally broken netbsd platform.
110 - tried to hack around openbsd bugs.
111
112 4.804 Wed Nov 5 16:36:00 CET 2008
113 - Coro::Debug::new_unix_server would not create a non-blocking listening
114 socket, sometimes causing freezes.
115 - (libcoro) fix misaligned stack points for setjmp and assembly
116 methods, which can cause crashes on x86/x86_64 with a sufficiently
117 aggressive compiler.
118 - new function: Coro::Debug::new_tcp_server.
119 - move ->throw into the Coro class because it only works on coro objects.
120
121 4.803 Mon Nov 3 17:16:12 CET 2008
122 - (libcoro) use a global asm statement to become independent of gcc
123 otpimisations for CORO_ASM (thanks to pippijn for the idea).
124 - try to workaround yet another broken bsd, this time dragonfly.
125
126 4.802 Thu Oct 30 10:56:12 CET 2008
127 - support -fno-omit-frame-pointer on x86 with the assembly method.
128 - tune 01_unblock.t tests a bit.
129
130 4.801 Wed Oct 22 18:33:37 CEST 2008
131 - improve readline speed for very long "lines".
132 - backport to 5.8.8.
133
134 4.8 Thu Oct 2 13:34:40 CEST 2008
135 - new function Coro::AIO::aio_wait.
136 - Coro.:AIO and Coro::BDB now "use Coro::AnyEvent".
137 - greatly speed up and reduce memory usage of Coro::AIO requests.
138 - implement some other µ-optimisations.
139
140 4.749 Mon Sep 29 14:40:12 CEST 2008
141 - port to slow and broken pseudo-threaded perls. (courtesy pippijn).
142
143 4.748 Sat Sep 27 14:03:19 CEST 2008
144 - implement, but do not document, PerlIO::cede(granularity).
145 - Coro::Storable forgot to wrap Storable::pstore.
146 - work around the multitude of leaks and memory corruption
147 bugs in PerlIO::via by using our own C-level perliol. As a side
148 effect, Coro::Storable is now much, much, much faster.
149
150 4.747 Tue Sep 23 01:59:41 CEST 2008
151 - fix a per-cv memleak (one empty array was leaked per
152 code reference).
153 - avoid a crash in coro->call|rss when the coroutine was already
154 destroyed (most noticably when using Coro::Debug::ps :)
155 - also protect *Storable::FILE.
156 - push up default storable granularity to 20ms.
157
158 4.746 Sun Sep 21 03:22:20 CEST 2008
159 - be more insistent on locking Storable against reentrancy
160 in Coro::Storable.
161 - move swap_def?v and throw to Coro::State, as documented.
162
163 4.745 Thu Jul 24 00:14:38 CEST 2008
164 - remove debugging code related to MgPV_nolen_const, also try to
165 make it compile with perl 5.8.6 (yes, apple apparently loves
166 outdated software). Reported by John S.
167
168 4.744 Tue Jul 8 22:06:35 CEST 2008
169 - correctly provide default DIE/WARN handlers as documented.
170 - also overwrite PL_vtbl_sigelem.svt_clear, even though current
171 implementations inside perl work fine for us.
172
173 4.743 Mon Jun 16 00:21:57 CEST 2008
174 - when using Coro::EV without running EV::loop it could
175 result in busy-waiting for events, this has been fixed.
176 - reduce codesize and improve performance by using EV_DEFAULT_UC.
177
178 4.742 Sat May 31 14:10:21 CEST 2008
179 - implement a workaround for (some) perl <5.8.8 versions.
180 - require EV 3.3+.
181
182 4.741 Fri May 30 23:33:09 CEST 2008
183 - tell netbsd how utterly broken their imitation of an OS is
184 and refuse to build by default if pthreads are in use.
185 - switch to "s" method on all bsd's by default, as their ucontext
186 stuff seems just too broken.
187 - fix a bug in Coro::Select.
188
189 4.74 Thu May 29 20:05:31 CEST 2008
190 - do not test Coro::LWP for lack of dependencies.
191 - do not test Coro::Select for lack of working perls.
192
193 4.73 Thu May 29 2008
194 - fix a bug in Coro::EV which would cause it to block despite
195 there being runnable coroutines.
196 - sprinkle "no warnings" freely over everything, also suppress
197 warnings for some other modules.
198 - fix typo in WSAEWOULDBLOCK.
199
200 4.72 Sun May 25 05:14:36 CEST 2008
201 - tweak META.yaml a bit, unfortunately, there is no documented way
202 to have optional dependencies with CPAN. doh :(
203 - avoid running some tests on windows because they would fail due to
204 perl bug (broken fork, broken pipes...).
205 - work around perl on windows bugs where perl returns undocumented
206 error codes for sysread, syswrite etc. by taking advantage
207 of AnyEvent's workaround for that problem.
208 - use AnyEvent::Util::fh_nonblocking in Coro::Handle to work around
209 a common perl implementation bug on windows.
210 - use unix domain sockets in testsuite to work around
211 common perl implementation bugs on widows (they are emulated by
212 tcp sockets on windows. ugh.)
213
214 4.71 Sat May 24 20:01:27 CEST 2008
215 - fix a bug in Coro::AnyEvent ("Usage: Coro::AnyEvent::_schedule()").
216 - take advantage of async name resolution of AnyEvent::Util.
217 - work around brutal inet_aton override in Coro::LWP.
218 - take advantage of the readyhook in Coro::EV, for smoother
219 scheduling.
220
221 4.7 Sun May 11 00:32:19 CEST 2008
222 - completely reworked the Coro manpage.
223 - added Coro::AnyEvent, generic event loop integration.
224 - implement cancel, ready and kill commands in Coro::Debug.
225 - document find_coro in Coro::Debug.
226 - incompatible change: rename has_stack to has_cctx.
227 - Coro::AIO and Coro::BDB no longer force event model detection,
228 use AnyEvent::AIO and AnyEvent::BDB.
229
230 4.6 Sat Apr 26 10:05:14 CEST 2008
231 - INCOMPATIBLE CHANGE: sub/code attributes are no longer supported
232 by the Coro module. It was a mistake to have it in the first place.
233 - (experimental) support for activestate perl 5.10 (method "w").
234 - (experimental) support for strawberry perl (method "a").
235 - coro_sigelem_set did not return a value although it had to,
236 actual impact unknown.
237
238 4.51 Mon Apr 14 13:28:27 CEST 2008
239 - make it compile again on 5.8.
240
241 4.50 Thu Apr 10 09:43:17 CEST 2008
242 - I did it twice! (see 4.49).
243
244 4.49 Sun Apr 6 21:23:31 CEST 2008
245 - grr, instead of compiling the recent changes on 5.10 only they
246 were compiled on 5.8 only.
247
248 4.48 Sun Apr 6 20:36:46 CEST 2008
249 - allow coroutine switches during eval's under 5.10.x, as it seems
250 the parser is a per-interpreter option now, so this is safe
251 (this might fix the odd crashing bug).
252 - drop support for 5.9.x versions: they are dead, jim.
253
254 4.47 Sun Apr 6 00:37:52 CEST 2008
255 - force cctx allocation on API calls: we know we need to force one
256 and gcc actually manages to confuse our heuristic nowadays,
257 leading to crashes and worse.
258 - document force_cctx.
259
260 4.46 Fri Apr 4 22:05:43 CEST 2008
261 - upgrade libcoro, resulting in pthread-backend (which was only created
262 to fulfill the rules of the programming languages shootout).
263
264 4.45 Thu Mar 13 11:55:36 CET 2008
265 - fix a file leaking bug in eg/mhyttpd that would allow
266 downloading of any file (reported by oesi).
267 - fix deadlock bug in Coro::Channel (reported by Richard Hundt)
268 (also add testcase).
269 - support Broadcast option in Coro::Socket (patch by Richard Hundt,
270 apparently having loads of fun with that).
271 - implement $state->swap_defsv, ->swap_defav and document ->throw.
272
273 4.4 Wed Feb 13 16:44:29 CET 2008
274 - only assume .cfi pseudo isns exist on GNU/Linux.
275 - add get_timed method to Coro::Channels.
276 - fixed Coro::Debug manpage.
277 - perl 5.11 compatbility improvement based on patch by
278 Andreas König.
279
280 4.37 Sun Jan 20 11:25:23 CET 2008
281 - soften the check for an active parser for perl 5.10 (avoiding
282 "Coro::State::transfer called while parsing" in many cases).
283
284 4.36 Sun Jan 13 10:53:56 CET 2008
285 - reset diehook when terminating from an async_pool as to not
286 trigger any __DIE__ handlers.
287
288 4.35 Sun Jan 13 04:14:13 CET 2008
289 - "bt" debug command now displays any exceptions
290 from longmess and also skips the innermost
291 stackframes, giving more useufl output.
292 - allow backtraces in coroutines blocked in Coro::EV,
293 at a <1% speed hit.
294 - handle localising of $SIG{__WARN__} and __DIE__
295 properly (with a proper amount of dirty hacking).
296
297 4.34 Sat Dec 22 17:49:53 CET 2007
298 - upgrade to EV version 2.0 API.
299
300 4.33 Mon Dec 17 08:36:12 CET 2007
301 - make Coro::AIO etc. loadable in the absence of EV.
302
303 4.32 Mon Dec 17 07:46:02 CET 2007
304 - majorly improved Coro::Handle's performance with Coro::EV.
305 - implemented the until now mythical Coro::BDB module.
306 - specialcase EV in Coro::AIO and Coro::BDB for extra speed.
307
308 4.31 Wed Dec 5 12:32:39 CET 2007
309 - remove warn statement form Coro::Util.
310
311 4.3 Tue Dec 4 20:33:14 CET 2007
312 - calls to the idle function could cause stack corruption
313 when the stack changed.
314 - do no longer rely on the presence of EV::DNS if EV is used
315 (because it is gone), but instead take avdantage of EV::ADNS
316 if available.
317 - add ($) prototypes to all functions in Coro::Storable.
318 - use a conventional (safer) idle callback in Coro::EV.
319 - do accept quickly in Coro::Debug to avoid endless loops.
320
321 4.22 Fri Nov 30 16:04:04 CET 2007
322 - really use optimised versions for Event and EV in Coro::Util
323 and Coro::Handle.
324
325 4.21 Sun Nov 25 10:48:59 CET 2007
326 - fix a spurious memory read.
327 - Coro::EV no longer keeps the eventloop "alive".
328
329 4.2 Fri Nov 9 20:47:05 CET 2007
330 - enable/disable tracing from a new coroutine, not a pooled one.
331 - fix a memleak in Coro::Event.
332 - removed killall call from fork_eval.
333 - made sure store_fd is already loaded so that fork_eval
334 does not have to parse autoload in each subprocess.
335 - only use assembly method if -O switch is in $Config{optimize}.
336 - add (optional) Coro::EV module, so far the best event loop module
337 directly supported by Coro.
338 - if the event model is EV, use EV::DNS to resolve
339 stuff in Coro::Util.
340 - don't get confused by multiple event notifications in Coro::Handle.
341 - initial support for EV (libevent interface).
342 - require Event and EV using configure_requires, to force their existance.
343
344 4.13 Wed Oct 24 07:26:45 CEST 2007
345 - add Coro::Storable::blocking_thaw.
346 - use a vastly more complicated technique to localise
347 $SIG{__WARN/DIE__} that also works on perls <= 5.8.8.
348 - use a coroutine for the idle callback Coro::Event,
349 instead of running Event in the current coroutine context.
350 This also catches recursive invocations.
351 - actually report fork errors in gethostbyname and inet_aton.
352
353 4.11 Thu Oct 11 02:40:24 CEST 2007
354 - port to threaded perls.
355
356 4.1 Thu Oct 11 02:38:16 CEST 2007
357 - incompatible change: $SIG{__DIE__} and $SIG{__WARN__} will now
358 be local to each coro (see Coro::State).
359 - incompatible change: for very deep reasons, cede and cede_notself
360 cannot return anything, so nothing will be returned.
361 - possibly bring back 5.10 compatibility (untested).
362 - work around stupid (and wrong) warning on 5.10 :(.
363 - overlay the saved state over the context stack. This saves
364 a few hundred bytes per coroutine on average and also
365 speeds up context switching a bit.
366 - further tune default stack sizes.
367 - (more) correctly calculate stack usage in coro_rss.
368 - Coro::Storable::blocking_* did not properly lock
369 resulting in races between coroutines.
370 - added Coro::Storable::guard.
371 - stopping to trace a coroutine could destroy the cctx of
372 an unrelated coroutine.
373 - explain the relationship between Perl and C coroutines in
374 more detail in Coro::State.
375 - Coro::Util::inet_aton did not short-circuit dotted quad forms,
376 causing a fork per resolve. This also affected Coro::Socket.
377 - switch to a separate stack in $coro->call/eval to avoid
378 invalidating pointers.
379
380 4.03 Sat Oct 6 21:24:00 CEST 2007
381 - added Coro::throw method.
382 - minor code cleanups.
383
384 4.02 Sat Oct 6 02:36:47 CEST 2007
385 - fix a very minor per-coroutine memleak (a single codereference).
386 - fixed a bug where the currently in-use c context would be freed
387 prematurely (can happen only when programs change the stacksize
388 or use tracing).
389 - tracing can no longer keep a coro alive after it terminated.
390 - do static branch prediction in the common path for gcc. gives
391 about 2-5% speed improvement here.
392
393 4.01 Fri Oct 5 22:10:49 CEST 2007
394 - instead of recreating *a* standard output handle we simply
395 use STDOUT, which is faster and hopefully more robust.
396
397 4.0 Fri Oct 5 12:56:00 CEST 2007
398 - incompatibly changed Coro::Storable::freeze.
399 - major new feature: added Coro::Debug, for interactive coroutine
400 debugging, tracing and much more.
401 - major bug fix: unbelievable, but true: $_, $/ and many other
402 "saved" variables actually weren't being saved. This has been fixed,
403 of course, while increasing performance while losing all the save
404 flags.
405 - save flags are gone, and all the api functions dealing with them.
406 - added Coro::Semaphore::adjust.
407 - added Coro::Util::fork_eval.
408 - added Coro::Storable::{nfreeze,blocking_{freeze,nfreeze}}.
409 - added Coro::killall.
410 - reduce initial stack sizes to allow for "micro-coroutines".
411 - better async_pool resource management, moved parts of async_pool
412 handling to XS (major speed improvement).
413 - actually croak before modifying important data structures.
414 - refuse to transfer while compiling.
415 - possibly support eval EXPR better now.
416 - enable assembly per default on linux+bsd x86+amd64.
417 - all internal members were renamed _something for easier subclassing.
418 - many minor tweaks.
419
420 3.63 Wed May 16 14:10:06 CEST 2007
421 - implement handcoded assembly for x86/amd64 SVR ABI.
422
423 3.62 Fri Apr 27 21:36:06 CEST 2007
424 - upgrade libcoro (which might set unwind info correctly).
425 - change default on linux to setjmp/longjmp.
426
427 3.61 Thu Apr 19 12:36:18 CEST 2007
428 - Coro::Storable caused an endless loop when thawing invalid
429 pst's sometimes.
430 - use a Semaphore in Coro::Storable, as Storable doesn't
431 seem to be reentrant (although it is documented to
432 be threadsafe...).
433 - fix Coro::Signal to bring back the original unreliable
434 but stateful semantics.
435 - fixed a lot of typos in Coro.pm (patch submitted by David
436 Steinbrunner, which applied flawlessly).
437
438 3.6 Sat Apr 14 17:13:31 CEST 2007
439 - added some bugfixes to get eg/myhttpd working again.
440 - added Coro::Storable for often-cede'ing freeze/thaw.
441 - try to do a clean exit when a coroutine calls exit
442 (EXPERIMENTAL).
443 - got rid of indirect call through _coro_init.
444 - updated the partly antique examples in eg/ to
445 work again and be a bit less magic, too.
446 - fixed Coro::Signal semantics to work as documented again.
447
448 3.55 Sun Mar 25 01:20:47 CET 2007
449 - add SAVE_DEFFH to save the default filehandle and enable
450 it by default.
451 - finally move socket-operations from Coro::Socket to Coro::Handle
452 to be able to unblock foreign sockets.
453 - add Coro::State::save_also and guarded_save.
454 - add count accessor to Coro::Semaphore.
455 - add Coro::State::cctx_stacksize.
456 - just for the fun of it, do not rely on implicit context,
457 which can dramatically improve performance, but people
458 using windows-process-emulation perls do not care much about
459 performance.
460
461 3.51 Sun Mar 4 14:18:04 CET 2007
462 - fixed a problem when you weakened references to Coro::State's
463 (patch by Yuval Kogman).
464
465 3.501 Wed Feb 28 12:44:07 CET 2007
466 - rename some global symbols as macosx from hell redefines
467 them without asking.
468
469 3.5 Tue Feb 13 20:22:53 CET 2007
470 - do AnyEvent model detection earlier, avoiding problems
471 caused by first using AnyEvent and later Coro::Event.
472 - implement and document Coro::Event event objects.
473 - fix a potential problem in Coro::Event causing crashes.
474 - initialise PL_comppad when creating a new coroutine,
475 avoids crashes on early coro destruction.
476
477 3.41 Mon Jan 22 19:19:49 CET 2007
478 - readline on Coro::Handle did not support undefined $/,
479 nor did it deliver partial lines on EOF or error.
480 - implement malloc fallback for stack allocation because
481 stupid broken idiotic OSX has a stupid broken
482 idiotic fits-the-whole-os mmap "implementation" and
483 my dick feels longer if Coro is portable even to
484 obsolete platforms.
485
486 3.4 Fri Jan 19 21:52:54 CET 2007
487 - remove t/09_timer.t, as it isn't really testing much
488 but was rather flaky in practise.
489 - async_pool coro would keep arguments and callback alive until
490 it was reused.
491 - cancellation of a coroutine could cause spurious idle calls
492 in cede_notself.
493
494 3.3 Sat Jan 6 03:45:00 CET 2007
495 - implement $coro->on_destroy.
496 - Coro::Event blocking semantics have been changed,
497 documented and - hopefully - improved.
498 - fix nice adding, not subtracting, from priority.
499 - fix ->prio and api_is_ready (patch by Mark Hinds).
500 - fixed an assert ("... == prev__cctx->idle_te")
501 that could errronously trigger.
502 - fix various large and small memleaks.
503 - use a (hopefully) more stable cancel implementation
504 that immediately frees the coroutine data.
505 - cede/cede_notself return a status now.
506 - added Coro::guard function.
507 - added a global coroutine pool for jobs (on my machine,
508 I can create and execute 48k simple coros/s with async,
509 and 128k coros with async_pool).
510 - Coro::AIO now uses the coroutine priority as io priority.
511
512 3.2 Fri Dec 22 05:07:09 CET 2006
513 - improve portability to slightly older perls.
514 - use cleaner coroutine destruction.
515 - simplify configuration for users.
516 - optionally (unrecommended) prefer perl functions over
517 their coro replacements.
518
519 3.11 Tue Dec 5 13:11:24 CET 2006
520 - fixed some bogus assert's, but as perl.h disables assert even
521 without NDEBUG (thank you very much), not too many people should
522 notice (that did include myself). Andreas König noticed, though :)
523 - do not save/restore PL_sortcxix on >= 5.9.x, it doesn't seem to have
524 it. Also noticed by Andreas König :)
525 - save/restore tainted status.
526 - verified to pass the testsuite on my 5.9.5.
527
528 3.1 Mon Dec 4 23:03:40 CET 2006
529 - INCOMPATIBLE CHANGE: $/ is now per-coroutine (but slow).
530 - incompatible change: transfer flags are now per-state.
531 - give a better error message on deadlock.
532 - document Coro::nready.
533 - enhanced testsuite.
534
535 3.01 Sun Dec 3 23:47:42 CET 2006
536 - forgot to include Coro::Timer.
537
538 3.0 Sun Dec 3 22:57:25 CET 2006
539 - the "FINALLY COMPLETELY STABLE" release (coming soon:
540 the "FAMOUS LAST WORDS" release).
541 - implement a new stack sharing algorithm, which uses a stack
542 pool (size currently hardcoded).
543 - make stack sharing mandatory (it no longer uses a heuristic).
544 - eval/die no longer cause weird problems under heavy use.
545 - Coro::Event could cause livelocks if it was used but
546 no Coro::Event watchers were used.
547 - Coro::Event now uses asynccheck as prepare does not
548 check for changed watchers.
549 - Coro::Event allows multiple waiting coros and will wake up one
550 per event.
551 - Coro::Event should be cleaner and more efficient now.
552 - new utility function Coro::unblock_sub.
553 - document the sad fact that Event is no longer reentrant.
554 - putting a coroutine into the ready queue twice could under
555 some circumstances lead to stack corruption.
556 - minor incompatibility: subclassing Coro::State is supported
557 directly now without going through a _coro_state member.
558 - state/coro switching is much faster now.
559 - very minor optimisations and code/documentation cleanup.
560 - avoid problems due to compiler inlining.
561 - removed timers from Coro::Timer -> use AnyEvent instead.
562 - replaced Coro::idle coroutine by (cleaner) idle handler.
563 - updated to newest libcoro.
564 - implement enhanced support for valgrind.
565 - implement is_ready and return value for ready.
566 - removed Coro::Cont, it was a misnomer (it's generators),
567 and it was somewhat annoying to get it right. Will come back
568 if somebody asks for it :)
569 - many ->wait methods and Coro::Event could return
570 spuriously without the event having happened.
571
572 2.5 Tue Nov 7 12:22:33 CET 2006
573 - made Coro::Util, Coro::Select, Coro::Handle and Coro::Socket
574 use AnyEvent, moved them to Coro/.
575 - added Coro::LWP which contains all the uglyness required to
576 make LWP non-blocking.
577 - should work with perl 5.9.x now (Andreas König made me do it).
578 - fixed another bug in Coro::Select when the timeout was undef.
579 - reuse PL_start_env for all coros, saves some memory per coroutine.
580 - manage PL_top_env differently, hopefully to avoid panic: top_env.
581 - timeout argument was not properly used in Coro::Socket.
582 - allow limited forms of subclassing in Coro::Handle/Coro::Socket.
583 - emulate undocumented(!) functionality of IO::Socket required
584 by LWP(!!).
585 - updated eg/lwp to work with newer lwp's.
586 - remove "FATAL: uncaught exception" prefix. Coroutines that die
587 kill the whole process, just as exceptions in the main "coroutine"
588 did already.
589
590 2.1 Wed Nov 1 23:01:13 CET 2006
591 - fix a long-standing bug in Coro::Select where select with
592 zero timeout would instead change the current default filehandle.
593 - use a simpler and hopefully more robust way to clone padlists
594 (uses less memory and a perl function instead of our own).
595 - coro can now create a stack guard area on many architectures.
596 - Coro::AIO properly reexports additional functions from IO::AIO.
597 - updated libcoro with a workaround for OS X,
598 pach and testing by Michael Schwern.
599
600 2.0 Tue Oct 24 05:47:17 CEST 2006
601 - support additional aio requests in Coro::AIO.
602
603 1.9 (never properly released due to a glitch)
604
605 1.8 Thu Feb 2 00:59:06 CET 2006
606 - applied suggested patch by SAMV to avoid problems during stupid
607 mark & sweep gc run.
608 - applied patch by Scott Walters for 5.9.3 compatibility.
609
610 1.7 Tue Dec 27 01:41:58 CET 2005
611 - added Coro::AIO, a thin wrapper around IO::AIO.
612 - improved Makefile.PL explanations.
613
614 1.6 Mon Dec 12 21:30:05 CET 2005
615 - additionally save PM_curpm might fix as-of-yet
616 unseen problems with regex matches being attributed to the
617 wrong package.
618 - add t/10_bugs.t, which currently checks against imho broken
619 perls that use 0.26MB of stack space per Perl_magic_get
620 invocation (newer linuxes) as opposed to the 0.0002MB perl
621 normally uses.
622 - make stacksize configurable for the ultimate debian experience.
623
624 1.51 Mon Dec 12 18:48:36 CET 2005
625 - remove debugging warn() accidentally left in Coro::Select.
626
627 1.5 Tue Nov 29 13:32:44 CET 2005
628 - use Coro::Event inside Coro::Select to avoid spurious deadlocks.
629 - fix Coro::Select 'select'.
630 - strict'ify some modules.
631 - libcoro errornously restored the SIGUSR2 handler to SIGUSR1.
632 - use XSLoader in selected modules.
633 - remove some 5.6 compatibility cruft.
634
635 1.4 Tue Sep 6 00:11:05 CEST 2005
636 - libcoro did not take into account the trampoline on amd64,
637 when 'l'inux method was used.
638
639 1.31 Tue Aug 30 23:31:33 CEST 2005
640 - some portability fixes/workarounds.
641
642 1.3 Sat Aug 20 03:08:56 CEST 2005
643 - no code changes. module seems to work fine.
644 - improve Coro::State docs, remove reference to nonexisting
645 Coro::State::flush method.
646 - no longer autodetect windows, present it as an option instead.
647
648 1.2 Mon May 16 02:00:55 CEST 2005
649 - included libcoro.c earlier in State.xs, might improve portability.
650 - use a faster and possibly more stable Coro::Cont implementation.
651 - accept x86_64 in addition to amd64 for optimized linux-amd64 support.
652 - fix bugs Coro::SemaphoreSet that could cause locks never to be freed.
653 - fix bugs in CoroAPI.h and document it in Coro::MakeMaker.
654
655 1.11 Thu Mar 3 18:00:52 CET 2005
656 - change of contact address.
657
658 1.1 Tue Feb 22 20:51:16 CET 2005
659 - support [l]inux method on amd64.
660 - allow some context switches while compiling/eval'ing.
661
662 1.0 Tue Aug 10 03:47:27 CEST 2004
663 - Event 0.89 fixes the bug again.
664 - fix the remaining(??) memleak. I didn't free the cache-AV for
665 padlists, so one AV per created anon-sub leaked.
666
667 0.97 Fri May 14 15:23:32 CEST 2004
668 - remove stupid gcc extension that was an experiment only anyway.
669
670 0.96 Thu May 13 18:09:29 CEST 2004
671 - tell people if they have yet another broken Event variant.
672 - made Coro to work _better_ with threads (it seems that, as long
673 as you use Coros in one thread only everything will work fine.
674 Looking at the perl source the problem might be missing locking
675 between Coro and some perl internal routines. That's not easily
676 fixable).
677 - (hopefully) got rid of the static xnext inside transfer.
678
679 0.95 Sun Feb 15 03:20:28 CET 2004
680 - removed Coro::State::flush, implemented a new and much
681 faster caching mechanism by attaching magic to the CV.
682 - a memleak with (real) closures remains.
683 - dramatically reduced size of stacks to 65536 (linux/x86), after
684 observing that even myhttpd never had more than a kilobyte
685 of stack in use (stack is still only physically allocated in
686 pagesize increments for systems suporting mmap).
687 - fix for a "cygwin-64" architecture added, proposed by Stefan Traby.
688 - fix for threaded perl.
689
690 0.9 Sun Nov 23 04:32:39 CET 2003
691 - fix a few very hard to track down but (of course) stupid
692 memory leaks.
693 - stringify version number to avoid locale problems :(.
694 - perl 5.9.0 does not have SvPADBUSY, reported by Scott Walters,
695 hopefully ignoring it "just works".
696
697 0.8 Wed Nov 5 19:38:40 CET 2003
698 - port to perl5.005_03, but only 5.8.x+ is supported!
699 - honor LocalAddr even without LocalPort in Coro::Socket.
700
701 0.7 Tue May 27 03:12:38 CEST 2003
702 - the version jump indicates some level of testing, not gobs
703 of new features.
704 - uh... I found the prompt function in ExtUtils::MakeMaker.
705 Highly correct stuff, that is...
706 - fixed(?) a bug with die's in coroutines causing "panic: top_env".
707 the fix is not well understood by the author, so beware :(.
708
709 0.652 Thu May 8 02:54:46 CEST 2003
710 - Applied patch by Slaven Rezic to set default to "s" on FreeBSD,
711 cause version 4 doesn't have ucontext.
712 - Benjamin Reed reported that setjmp works fine on darwin,
713 so preselect it.
714
715 0.651 Sat Mar 29 15:00:23 CET 2003
716 - fix a bug in Coro::Handle where some bytes could get lost
717 on reading, reported by jason@nichego.net.
718
719 0.65 Sun Mar 23 00:08:26 CET 2003
720 - added a README.linux-glibc.
721 - new module Coro::Select.
722 - also save/restore PL_comppad, fixes segfaults with 5.8.1.
723 (I never claimed that I know what I am doing ;).
724 - set default to setjmp/longjmp on non-x86-linux.
725
726 0.6 Thu Nov 21 11:09:06 CET 2002
727 - made lazy context switching the default.
728 - fixed the bug where SemaphoreSet::guard
729 was the same as timed_guard and timed_guard was missing.
730 - fix a memleak in Coro::Timer.
731 - reclassified context sharing as not experimental.
732
733 0.534 Sun Apr 14 03:05:12 CEST 2002
734 - fixed backspace => tab in header parsing (myhttpd).
735 - added eg/readline, for elmex.
736 - Coro::Event now calls ->start on first call instead of
737 ->again, which created wrong timeouts for the first ->next.
738 - fixed a bug where Coro::Socket returned a closed socket instead
739 of nothing on ECONNREFUSED and similar error conditions.
740
741 0.533 Mon Feb 18 18:49:40 CET 2002
742 - adapt to finally working Event-0.85.
743 - fixed buggy event initialization (trapped under windows, but how
744 could it possibly have worked anywhere?)
745
746 0.532 Wed Jan 16 02:45:32 CET 2002
747 - removed the reference to pp_entersub, might work on
748 (some) win32 perls now (testsuite works under cygwin,
749 without Event, which freezes).
750
751 0.531 Mon Dec 10 22:18:44 CET 2001
752 - Fixed a bug in SemaphoreSet::timed_down.
753
754 0.53 Tue Nov 27 21:11:13 CET 2001
755 - some tests for Event.
756 - slightly different internal architecture to get refcnt into the object
757 and not the reference. might fix a few bugs, certainly did introduce
758 new ones.
759 - Coro::Timer (independent of Event).
760 - new timed_wait functions for Coro::Signal, Semaphore, SemaphoreSet.
761
762 0.52 Tue Nov 6 21:36:18 CET 2001
763 - ported to cygwin (trivial).
764 patch by Gerrit P. Haase <gp@familiehaase.de>.
765 - small setjmp code fix by Sullivan.DanielJ@epamail.epa.gov.
766
767 0.51 Thu Nov 1 20:39:01 CET 2001
768 - terminate/cancel now work properly (otherwise termination
769 could cause a "next coroutine is not and contains not..."
770 error.
771 - added Coro::Socket::shutdown.
772 - Coro::Event::loop no is the same as Event::loop.
773 - implemented terminate with args + join.
774
775 0.5 Fri Sep 28 16:15:35 CEST 2001
776 - fixed "print" on a Coro::Handle. The print method worked.
777 - small tweaks (seem to reduce memory consumption a lot)
778 in various modules.
779 - splendid use of "no warnings" scattered throughout.
780 - added Coro::Handle::rbuf, fixed Coro::Handle::fh.
781
782 0.49 Sun Sep 16 02:42:45 CEST 2001
783 - changed some method calls to function calls for speed
784 inside Coro::Handle.
785 - make Coro::Handle use an array instead of a hash for
786 speed reasons.
787 - IRIX mystery solved: it's SGI's NT, after all: "standard, huh?".
788 sjlj and ucontext should now work.
789 - IRIX-specific port for libcoro.
790 - swapped order of accept results to match IO::Socket.
791 - changed getsock/peername to sock/peername to match IO::Socket.
792 - fixed a bug that caused segfault when returning to main task
793 under some circumstances.
794 - other bugfixes.
795
796 0.45 Sun Sep 2 02:54:01 CEST 2001
797 - new method Coro::Handle::timeout.
798 - corrected speling of set...name to get...name in Coro::Socket.
799 - Coro::Socket::accept now returns a Coro::Socket, not a Coro::Handle.
800 - Coro::Handle now supports fileno and FILENO.
801 - added eg/myhttpd, a web server.
802 - Coro::Socket now accepts numeric ports as well as "port(number)"
803 syntax.
804 - moved some scheduling primitives into xs code. more to come.
805 - new simple priority system.
806 - implemented Coro::Event using XS, almost four times faster!
807 - small memory corruption problem fixed (boy that was difficult).
808
809 0.13 Wed Aug 8 16:53:07 CEST 2001
810 - new method Coro::Handle::readline.
811 - added eg/lwp and Coro::Handle::autoflush, to make LWP work.
812 (see eg/lwp on how to make LWP non-blocking).
813 - renamed Coro::Socket::new_inet to new.
814 - added Coro::Util, some utility functions.
815 - die/eval now works better (still get annoying (but true ;)
816 "Callback called exit" messages).
817 - Coro::Handle now supports timeout, and Coro::Socket Timeout.
818 - much better Coro::Socket support.
819
820 0.12 Fri Jul 27 04:19:01 CEST 2001
821 - do not use mmap if neither MAP_ANON nor MAP_ANONYMOUS is defined.
822 - very experimental stack sharing algorithm. not 100% safe but
823 should work well in practise ;)
824 - added Coro::RWLock.
825 - Coro::Specific now works.
826
827 0.11 Tue Jul 24 22:49:21 CEST 2001
828 - added specialized hack for newer and older linux versions (fast).
829 - renamed Coro::Event::IO to Coro::Handle.
830 - new module Coro::Socket.
831
832 0.1 Tue Jul 24 01:47:53 CEST 2001
833 - release candidate 3. A rather new internal structure :(
834 - the great renaming: Damian Conway gave me a suitable replacement
835 for yield.
836 - added Coro/libcoro, a portable coroutine implementation for C
837 (not even perl-dependent!!!), which can be used by Coro::State.
838 - renamed SAVE_DEFSV (on xs-level only) to avoid symbol clash in
839 perl-5.6.
840 - new function Coro::Event::idle.
841 - the idle process is now overriden
842 by default in Coro::Event.
843 - Coro::Channel now enforces the size.
844 - canceling events now works in all cases.
845 - Coro state now include $_ and $@.
846 - yet another bug workaround that I do not really understand :(
847 - new module Coro::Event::IO (very undocumented) to do
848 non-blocking i/o.
849 - performance tuning.
850
851 0.08 Thu Jul 19 06:13:25 CEST 2001
852 - release candidate 2 (new functionality).
853 - add Coro::State::flush function.
854 - transfer now supports a flags value (mostly for speed ATM).
855 - might compile and work in the presence of threads now.
856 - continuations have a saner syntax.
857 - no more memleaks.
858
859 0.07 Tue Jul 17 17:40:18 CEST 2001
860 - release candidate 1 ;)
861 - slightly nicer code.
862 - fixed a scheduling bug in Coro::Event.
863
864 0.06 Tue Jul 17 04:23:24 CEST 2001
865 - ok, I found the showstopper - the same sub must not be
866 re-entered in two different coroutines, otherwise => crash. I
867 see no easy solution to this problem, except by walking the call
868 chaing and saving/restoring the cv's, which is what I do now.
869 - memory leaks still latent, especially at thread termination.
870 - Coro::Event now works (haha).
871
872 0.05 Sun Jul 15 17:32:20 CEST 2001
873 - fixed a few issues in Coro::Event.
874 - I forgot to include Coro::Event in 0.04 :(:(:(
875
876 0.04 Sun Jul 15 05:24:59 CEST 2001
877 - @_ is now properly localized.
878 - Coro::State is now easier subclassable.
879 - Coro::Cont now coroutine-aware.
880 - Coro::Specific is a low-overhead module to create
881 coroutine-specific vars.
882 - Coro::Event provides a simple interface to Event.
883
884 0.03 Fri Jul 13 14:51:52 CEST 2001
885 - transfer() now implemented in XS (beware).
886 - new module Coro::Cont for really faked continuations.
887 - big internal architecture changes: Coro::State is now
888 really low-level and can thus be used to implement other
889 interesting things, While "Coro::" implements a process-like
890 model. Still crude and subject to change.
891 - $_ and $@ are no longer being localized.
892
893 0.02 Tue Jul 10 01:38:17 CEST 2001
894 - implemented "async" attribute.
895 - $_ and $@ are now localized.
896 - added Coro::Channel.
897 - more testcases, still no docs.
898
899 0.01 Tue Jul 3 02:18:41 CEST 2001
900 - original version; copied from Convert::Scalar.
901