ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/Coro/Changes
Revision: 1.473
Committed: Thu Dec 4 17:29:40 2008 UTC (15 years, 5 months ago) by root
Branch: MAIN
Changes since 1.472: +5 -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 5.12
9 - add default config for MirOS, which seems to be a bug-to-bug
10 compatible fork of openbsd ("world domination by releasing
11 openbsd cvs before the openbsd folks do it" or so :).
12 - free_padlist did destroy the names pad, not good, but didn't
13 seem to bother perl - this could fix issues such as eval ""
14 inside a function called form multiple coroutines.
15 - use a different method to detect destruction time.
16 - be more careful when freeing padlists just before global
17 destruction.
18 - fixed and expanded the call/cc example.
19 - renamed _terminate to _coro_run.
20 - new method Coro::Channel->shutdown.
21 - try pthreads on openbsd <4.4 (broken sigaltstack, will
22 pthreads fare better?).
23 - be less picky about destroying "a" running coroutine.
24
25 5.11 Tue Nov 25 21:49:05 CET 2008
26 - DEBUGGING in 5.10.0 is a mess: it constantly flags perfectly
27 working code with failed assertions, introducing more bugs than
28 it fixes, requiring elaborate workarounds :(
29
30 5.1 Mon Nov 24 08:54:59 CET 2008
31 - wrote a small introductory tutorial - Coro::Intro.
32 - convert Coro::Timer, Coro::Select and Coro::Util to rouse API.
33 - Coro::Select did errornously dup the file descriptors
34 and didn't work with all AnyEvent backends.
35 - Coro::Select wasn't imported correctly form Coro::LWP, causing blocking
36 LWP data transfers.
37 - disassociate c contexts from coro objects - this is agruably more
38 correct, but mostly allows sharing of cctxs between coro and state
39 objects, for added memory savings and speed increases.
40 - bumped $Coro::POOL_RSS up to 32kb by default.
41 - no longer set the optype to OP_CUSTOM, as B::* understandably
42 doesn't like this very much (and we *are* a type of entersub).
43 - implement state cloning, just to prove that call/cc can be done.
44 - automatically load Coro::AnyEvent in Coro::Handle.
45 - wrap ->cancel calls in eval inside Coro::Handle as EV watchers
46 do not have this method (and don't need it either).
47 - speed up generic anyevent methods in Coro::Handle by using rouse
48 callbacks.
49 - allow coroutines in $Coro::IDLE, speeding up Coro::AnyEvent and
50 others. It also makes the debugger happier, as you can trace
51 through the idle threads now.
52 - add comppad_name* and hints ($^H, %^H) to per-thread variables.
53 - eg/event was pretty broken.
54 - better 5.8.6 compatibility.
55
56 5.0 Thu Nov 20 10:35:05 CET 2008
57 - NEW ARCHITECTURE: use the latest 4.x version if you experience
58 stability issues.
59 - bump API version to 7 - all dependents must be recompiled.
60 - removed timed_* functions - they were not being used anyways
61 and should be replaced by a more generic mechanism -
62 and were annoying to support anyways :)
63 - removed SemaphoreSet's waiter method - use sem method instead.
64 - Coro::Semaphore->adjust didn't correctly wake up enough waiters.
65 - async_pool did free a scalar value twice
66 ("Attempt to unreference...").
67 - fix a longstanding bug where calling terminate on a coro that
68 was waiting for a semaphore that was just becoming available
69 would cause a deadlock (semaphore would get into a state where
70 it was available but waiters were still blocked).
71 - calling throw on a coroutine that is waiting for a semaphore will
72 no longer make it acquire the semaphore (and thus leak a count).
73 - perl's process emulation is now not even theoretically supported
74 anymore.
75 - new functions Coro::rouse_cb and Coro::rouse_wait for easier
76 conversion of callback-style to blocking-style.
77 - new methods $coro->schedule_to and ->cede_to, to specifically
78 schedule or cede to a specific coroutine.
79 - new function Coro::Semaphore::wait.
80 - use named constants in Coro::Channel (Richard Hundt).
81 - directly patch the entersub opcode calling SLF functions (cede,
82 transfer and so on). this does speed up context switching, but
83 more importanly, it frees us from the hardcoded behaviour of
84 entersub, so we might actually be able to return something from
85 those functions and atcually create new ones.
86 - take advantage of __builtin_frame_address on gcc.
87 - expose THX in coroapi (not sure whether this was a wise decision,
88 as "threaded" perls are running at half speed anyways).
89 - implement execute_slf (schedule-like-function) interface that makes
90 it possible to implement schedule-like-functions in XS.
91 - use new SLF interface to massively speed up Coro::EV by roughly a
92 factor of two.
93 - used new SLF interface to massively speed up Coro::Semaphore by a
94 factor of three.
95 - used new SLF interface to speed up Coro::AIO by roughly a factor of
96 four and reduce its memory usage considerably.
97 - implement Coro::SemaphoreSet purely in terms of Coro::Semaphore,
98 for a nice speedup and vastly more correct behaviour. Also implement
99 a new method "sem" to get at the underlying semaphore object.
100 - implement Coro::Channel in terms of Coro::Semaphore, for a moderate
101 (in comparison) 20-40% speedup.
102 - used new SLF interface to reimplement Coro::Signal gaining
103 some unknown (because I was too lazy), but certain, speedup, and also
104 making signals reliable for the first time.
105 - used new SLF interface and other optimisations to speed up async_pool
106 by a factor of two. It also doesn't rely on perl's exception mechanism
107 to exit anymore. The overhead for terminating an async_pool, coro over
108 a normal async is now very small.
109 - sped up coroutine creation/destruction by 40%.
110 - forgot to include Coro/libcoro/README in the dist for all these years.
111 - work around a freebsd pthreads bug (manual testcancel is required as
112 pthread_cond_wait isn't a cancellation point on freebsd).
113 - use new rouse functions to speed up and simplify Coro::BDB.
114 - make "prefer perl native functions" work with threaded perls.
115 - condense Coro::Debug ps output, hint at v and w flags.
116 - (libcoro) lots of minor cleanups and portability improvements.
117
118 4.914 Wed Nov 19 12:54:18 CET 2008
119 - fix a disastrous bug in the readline optimisation
120 introduced in 4.801.
121
122 4.913 Sat Nov 15 07:58:28 CET 2008
123 - async_pool did free a scalar value twice
124 ("Attempt to unreference...").
125
126 4.912 Thu Nov 13 18:31:23 CET 2008
127 - minor cleanups.
128 - use much larger stacks on linux and perl < 5.8.8.
129 - Coro::Debug::new_unix_server did not unlink the socket
130 when destroyed.
131
132 4.911 Tue Nov 11 04:26:17 CET 2008
133 - "port" to threaded perls.
134
135 4.91 Mon Nov 10 05:36:38 CET 2008
136 - the ->throw exception object no longer leaks.
137 - creating a new cctx leaked a scopestack entry (memleak).
138 - new coroutines didn't get created with a zero flags field
139 (unknown impact).
140 - calling ->throw on a not-yet-started coroutine should now work
141 instead of being ignored.
142 - ->throw is now supported on Coro::State objects.
143 - clean up cctx creation code a bit.
144 - entersub is actually a UNOP, not a LOGOP (not a bugfix).
145
146 4.9 Sat Nov 8 17:45:27 CET 2008
147 - (libcoro) did not preserve rbp with CORO_ASM (we are getting there).
148 - (libcoro) no longer leak threads in the experimental pthread backend,
149 also speed it up considerably.
150 - (libcoro) do not rely on makecontext passing void *'s unscathed.
151 - fix compiletime dependencies on libcoro in the Makefile.
152 - cctx_count wasn't always updated properly.
153 - Coro::State::cctx_stacksize wasn't applied correctly.
154 - new function Coro::State::cctx_max_idle.
155 - the default max number of idle C contexts is now 4.
156 - (libcoro) try harder to get _setjmp/_longjmp.
157 - (libcoro) cleanup and extend the libcoro API to officially
158 allow the creation of empty source contexts.
159 - very experimental workaround for the totally broken netbsd platform.
160 - tried to hack around openbsd bugs.
161
162 4.804 Wed Nov 5 16:36:00 CET 2008
163 - Coro::Debug::new_unix_server would not create a non-blocking listening
164 socket, sometimes causing freezes.
165 - (libcoro) fix misaligned stack points for setjmp and assembly
166 methods, which can cause crashes on x86/x86_64 with a sufficiently
167 aggressive compiler.
168 - new function: Coro::Debug::new_tcp_server.
169 - move ->throw into the Coro class because it only works on coro objects.
170
171 4.803 Mon Nov 3 17:16:12 CET 2008
172 - (libcoro) use a global asm statement to become independent of gcc
173 otpimisations for CORO_ASM (thanks to pippijn for the idea).
174 - try to workaround yet another broken bsd, this time dragonfly.
175
176 4.802 Thu Oct 30 10:56:12 CET 2008
177 - support -fno-omit-frame-pointer on x86 with the assembly method.
178 - tune 01_unblock.t tests a bit.
179
180 4.801 Wed Oct 22 18:33:37 CEST 2008
181 - improve readline speed for very long "lines".
182 - backport to 5.8.8.
183
184 4.8 Thu Oct 2 13:34:40 CEST 2008
185 - new function Coro::AIO::aio_wait.
186 - Coro.:AIO and Coro::BDB now "use Coro::AnyEvent".
187 - greatly speed up and reduce memory usage of Coro::AIO requests.
188 - implement some other ยต-optimisations.
189
190 4.749 Mon Sep 29 14:40:12 CEST 2008
191 - port to slow and broken pseudo-threaded perls. (courtesy pippijn).
192
193 4.748 Sat Sep 27 14:03:19 CEST 2008
194 - implement, but do not document, PerlIO::cede(granularity).
195 - Coro::Storable forgot to wrap Storable::pstore.
196 - work around the multitude of leaks and memory corruption
197 bugs in PerlIO::via by using our own C-level perliol. As a side
198 effect, Coro::Storable is now much, much, much faster.
199
200 4.747 Tue Sep 23 01:59:41 CEST 2008
201 - fix a per-cv memleak (one empty array was leaked per
202 code reference).
203 - avoid a crash in coro->call|rss when the coroutine was already
204 destroyed (most noticably when using Coro::Debug::ps :)
205 - also protect *Storable::FILE.
206 - push up default storable granularity to 20ms.
207
208 4.746 Sun Sep 21 03:22:20 CEST 2008
209 - be more insistent on locking Storable against reentrancy
210 in Coro::Storable.
211 - move swap_def?v and throw to Coro::State, as documented.
212
213 4.745 Thu Jul 24 00:14:38 CEST 2008
214 - remove debugging code related to MgPV_nolen_const, also try to
215 make it compile with perl 5.8.6 (yes, apple apparently loves
216 outdated software). Reported by John S.
217
218 4.744 Tue Jul 8 22:06:35 CEST 2008
219 - correctly provide default DIE/WARN handlers as documented.
220 - also overwrite PL_vtbl_sigelem.svt_clear, even though current
221 implementations inside perl work fine for us.
222
223 4.743 Mon Jun 16 00:21:57 CEST 2008
224 - when using Coro::EV without running EV::loop it could
225 result in busy-waiting for events, this has been fixed.
226 - reduce codesize and improve performance by using EV_DEFAULT_UC.
227
228 4.742 Sat May 31 14:10:21 CEST 2008
229 - implement a workaround for (some) perl <5.8.8 versions.
230 - require EV 3.3+.
231
232 4.741 Fri May 30 23:33:09 CEST 2008
233 - tell netbsd how utterly broken their imitation of an OS is
234 and refuse to build by default if pthreads are in use.
235 - switch to "s" method on all bsd's by default, as their ucontext
236 stuff seems just too broken.
237 - fix a bug in Coro::Select.
238
239 4.74 Thu May 29 20:05:31 CEST 2008
240 - do not test Coro::LWP for lack of dependencies.
241 - do not test Coro::Select for lack of working perls.
242
243 4.73 Thu May 29 2008
244 - fix a bug in Coro::EV which would cause it to block despite
245 there being runnable coroutines.
246 - sprinkle "no warnings" freely over everything, also suppress
247 warnings for some other modules.
248 - fix typo in WSAEWOULDBLOCK.
249
250 4.72 Sun May 25 05:14:36 CEST 2008
251 - tweak META.yaml a bit, unfortunately, there is no documented way
252 to have optional dependencies with CPAN. doh :(
253 - avoid running some tests on windows because they would fail due to
254 perl bug (broken fork, broken pipes...).
255 - work around perl on windows bugs where perl returns undocumented
256 error codes for sysread, syswrite etc. by taking advantage
257 of AnyEvent's workaround for that problem.
258 - use AnyEvent::Util::fh_nonblocking in Coro::Handle to work around
259 a common perl implementation bug on windows.
260 - use unix domain sockets in testsuite to work around
261 common perl implementation bugs on widows (they are emulated by
262 tcp sockets on windows. ugh.)
263
264 4.71 Sat May 24 20:01:27 CEST 2008
265 - fix a bug in Coro::AnyEvent ("Usage: Coro::AnyEvent::_schedule()").
266 - take advantage of async name resolution of AnyEvent::Util.
267 - work around brutal inet_aton override in Coro::LWP.
268 - take advantage of the readyhook in Coro::EV, for smoother
269 scheduling.
270
271 4.7 Sun May 11 00:32:19 CEST 2008
272 - completely reworked the Coro manpage.
273 - added Coro::AnyEvent, generic event loop integration.
274 - implement cancel, ready and kill commands in Coro::Debug.
275 - document find_coro in Coro::Debug.
276 - incompatible change: rename has_stack to has_cctx.
277 - Coro::AIO and Coro::BDB no longer force event model detection,
278 use AnyEvent::AIO and AnyEvent::BDB.
279
280 4.6 Sat Apr 26 10:05:14 CEST 2008
281 - INCOMPATIBLE CHANGE: sub/code attributes are no longer supported
282 by the Coro module. It was a mistake to have it in the first place.
283 - (experimental) support for activestate perl 5.10 (method "w").
284 - (experimental) support for strawberry perl (method "a").
285 - coro_sigelem_set did not return a value although it had to,
286 actual impact unknown.
287
288 4.51 Mon Apr 14 13:28:27 CEST 2008
289 - make it compile again on 5.8.
290
291 4.50 Thu Apr 10 09:43:17 CEST 2008
292 - I did it twice! (see 4.49).
293
294 4.49 Sun Apr 6 21:23:31 CEST 2008
295 - grr, instead of compiling the recent changes on 5.10 only they
296 were compiled on 5.8 only.
297
298 4.48 Sun Apr 6 20:36:46 CEST 2008
299 - allow coroutine switches during eval's under 5.10.x, as it seems
300 the parser is a per-interpreter option now, so this is safe
301 (this might fix the odd crashing bug).
302 - drop support for 5.9.x versions: they are dead, jim.
303
304 4.47 Sun Apr 6 00:37:52 CEST 2008
305 - force cctx allocation on API calls: we know we need to force one
306 and gcc actually manages to confuse our heuristic nowadays,
307 leading to crashes and worse.
308 - document force_cctx.
309
310 4.46 Fri Apr 4 22:05:43 CEST 2008
311 - upgrade libcoro, resulting in pthread-backend (which was only created
312 to fulfill the rules of the programming languages shootout).
313
314 4.45 Thu Mar 13 11:55:36 CET 2008
315 - fix a file leaking bug in eg/mhyttpd that would allow
316 downloading of any file (reported by oesi).
317 - fix deadlock bug in Coro::Channel (reported by Richard Hundt)
318 (also add testcase).
319 - support Broadcast option in Coro::Socket (patch by Richard Hundt,
320 apparently having loads of fun with that).
321 - implement $state->swap_defsv, ->swap_defav and document ->throw.
322
323 4.4 Wed Feb 13 16:44:29 CET 2008
324 - only assume .cfi pseudo isns exist on GNU/Linux.
325 - add get_timed method to Coro::Channels.
326 - fixed Coro::Debug manpage.
327 - perl 5.11 compatbility improvement based on patch by
328 Andreas Kรถnig.
329
330 4.37 Sun Jan 20 11:25:23 CET 2008
331 - soften the check for an active parser for perl 5.10 (avoiding
332 "Coro::State::transfer called while parsing" in many cases).
333
334 4.36 Sun Jan 13 10:53:56 CET 2008
335 - reset diehook when terminating from an async_pool as to not
336 trigger any __DIE__ handlers.
337
338 4.35 Sun Jan 13 04:14:13 CET 2008
339 - "bt" debug command now displays any exceptions
340 from longmess and also skips the innermost
341 stackframes, giving more useufl output.
342 - allow backtraces in coroutines blocked in Coro::EV,
343 at a <1% speed hit.
344 - handle localising of $SIG{__WARN__} and __DIE__
345 properly (with a proper amount of dirty hacking).
346
347 4.34 Sat Dec 22 17:49:53 CET 2007
348 - upgrade to EV version 2.0 API.
349
350 4.33 Mon Dec 17 08:36:12 CET 2007
351 - make Coro::AIO etc. loadable in the absence of EV.
352
353 4.32 Mon Dec 17 07:46:02 CET 2007
354 - majorly improved Coro::Handle's performance with Coro::EV.
355 - implemented the until now mythical Coro::BDB module.
356 - specialcase EV in Coro::AIO and Coro::BDB for extra speed.
357
358 4.31 Wed Dec 5 12:32:39 CET 2007
359 - remove warn statement form Coro::Util.
360
361 4.3 Tue Dec 4 20:33:14 CET 2007
362 - calls to the idle function could cause stack corruption
363 when the stack changed.
364 - do no longer rely on the presence of EV::DNS if EV is used
365 (because it is gone), but instead take avdantage of EV::ADNS
366 if available.
367 - add ($) prototypes to all functions in Coro::Storable.
368 - use a conventional (safer) idle callback in Coro::EV.
369 - do accept quickly in Coro::Debug to avoid endless loops.
370
371 4.22 Fri Nov 30 16:04:04 CET 2007
372 - really use optimised versions for Event and EV in Coro::Util
373 and Coro::Handle.
374
375 4.21 Sun Nov 25 10:48:59 CET 2007
376 - fix a spurious memory read.
377 - Coro::EV no longer keeps the eventloop "alive".
378
379 4.2 Fri Nov 9 20:47:05 CET 2007
380 - enable/disable tracing from a new coroutine, not a pooled one.
381 - fix a memleak in Coro::Event.
382 - removed killall call from fork_eval.
383 - made sure store_fd is already loaded so that fork_eval
384 does not have to parse autoload in each subprocess.
385 - only use assembly method if -O switch is in $Config{optimize}.
386 - add (optional) Coro::EV module, so far the best event loop module
387 directly supported by Coro.
388 - if the event model is EV, use EV::DNS to resolve
389 stuff in Coro::Util.
390 - don't get confused by multiple event notifications in Coro::Handle.
391 - initial support for EV (libevent interface).
392 - require Event and EV using configure_requires, to force their existance.
393
394 4.13 Wed Oct 24 07:26:45 CEST 2007
395 - add Coro::Storable::blocking_thaw.
396 - use a vastly more complicated technique to localise
397 $SIG{__WARN/DIE__} that also works on perls <= 5.8.8.
398 - use a coroutine for the idle callback Coro::Event,
399 instead of running Event in the current coroutine context.
400 This also catches recursive invocations.
401 - actually report fork errors in gethostbyname and inet_aton.
402
403 4.11 Thu Oct 11 02:40:24 CEST 2007
404 - port to threaded perls.
405
406 4.1 Thu Oct 11 02:38:16 CEST 2007
407 - incompatible change: $SIG{__DIE__} and $SIG{__WARN__} will now
408 be local to each coro (see Coro::State).
409 - incompatible change: for very deep reasons, cede and cede_notself
410 cannot return anything, so nothing will be returned.
411 - possibly bring back 5.10 compatibility (untested).
412 - work around stupid (and wrong) warning on 5.10 :(.
413 - overlay the saved state over the context stack. This saves
414 a few hundred bytes per coroutine on average and also
415 speeds up context switching a bit.
416 - further tune default stack sizes.
417 - (more) correctly calculate stack usage in coro_rss.
418 - Coro::Storable::blocking_* did not properly lock
419 resulting in races between coroutines.
420 - added Coro::Storable::guard.
421 - stopping to trace a coroutine could destroy the cctx of
422 an unrelated coroutine.
423 - explain the relationship between Perl and C coroutines in
424 more detail in Coro::State.
425 - Coro::Util::inet_aton did not short-circuit dotted quad forms,
426 causing a fork per resolve. This also affected Coro::Socket.
427 - switch to a separate stack in $coro->call/eval to avoid
428 invalidating pointers.
429
430 4.03 Sat Oct 6 21:24:00 CEST 2007
431 - added Coro::throw method.
432 - minor code cleanups.
433
434 4.02 Sat Oct 6 02:36:47 CEST 2007
435 - fix a very minor per-coroutine memleak (a single codereference).
436 - fixed a bug where the currently in-use c context would be freed
437 prematurely (can happen only when programs change the stacksize
438 or use tracing).
439 - tracing can no longer keep a coro alive after it terminated.
440 - do static branch prediction in the common path for gcc. gives
441 about 2-5% speed improvement here.
442
443 4.01 Fri Oct 5 22:10:49 CEST 2007
444 - instead of recreating *a* standard output handle we simply
445 use STDOUT, which is faster and hopefully more robust.
446
447 4.0 Fri Oct 5 12:56:00 CEST 2007
448 - incompatibly changed Coro::Storable::freeze.
449 - major new feature: added Coro::Debug, for interactive coroutine
450 debugging, tracing and much more.
451 - major bug fix: unbelievable, but true: $_, $/ and many other
452 "saved" variables actually weren't being saved. This has been fixed,
453 of course, while increasing performance while losing all the save
454 flags.
455 - save flags are gone, and all the api functions dealing with them.
456 - added Coro::Semaphore::adjust.
457 - added Coro::Util::fork_eval.
458 - added Coro::Storable::{nfreeze,blocking_{freeze,nfreeze}}.
459 - added Coro::killall.
460 - reduce initial stack sizes to allow for "micro-coroutines".
461 - better async_pool resource management, moved parts of async_pool
462 handling to XS (major speed improvement).
463 - actually croak before modifying important data structures.
464 - refuse to transfer while compiling.
465 - possibly support eval EXPR better now.
466 - enable assembly per default on linux+bsd x86+amd64.
467 - all internal members were renamed _something for easier subclassing.
468 - many minor tweaks.
469
470 3.63 Wed May 16 14:10:06 CEST 2007
471 - implement handcoded assembly for x86/amd64 SVR ABI.
472
473 3.62 Fri Apr 27 21:36:06 CEST 2007
474 - upgrade libcoro (which might set unwind info correctly).
475 - change default on linux to setjmp/longjmp.
476
477 3.61 Thu Apr 19 12:36:18 CEST 2007
478 - Coro::Storable caused an endless loop when thawing invalid
479 pst's sometimes.
480 - use a Semaphore in Coro::Storable, as Storable doesn't
481 seem to be reentrant (although it is documented to
482 be threadsafe...).
483 - fix Coro::Signal to bring back the original unreliable
484 but stateful semantics.
485 - fixed a lot of typos in Coro.pm (patch submitted by David
486 Steinbrunner, which applied flawlessly).
487
488 3.6 Sat Apr 14 17:13:31 CEST 2007
489 - added some bugfixes to get eg/myhttpd working again.
490 - added Coro::Storable for often-cede'ing freeze/thaw.
491 - try to do a clean exit when a coroutine calls exit
492 (EXPERIMENTAL).
493 - got rid of indirect call through _coro_init.
494 - updated the partly antique examples in eg/ to
495 work again and be a bit less magic, too.
496 - fixed Coro::Signal semantics to work as documented again.
497
498 3.55 Sun Mar 25 01:20:47 CET 2007
499 - add SAVE_DEFFH to save the default filehandle and enable
500 it by default.
501 - finally move socket-operations from Coro::Socket to Coro::Handle
502 to be able to unblock foreign sockets.
503 - add Coro::State::save_also and guarded_save.
504 - add count accessor to Coro::Semaphore.
505 - add Coro::State::cctx_stacksize.
506 - just for the fun of it, do not rely on implicit context,
507 which can dramatically improve performance, but people
508 using windows-process-emulation perls do not care much about
509 performance.
510
511 3.51 Sun Mar 4 14:18:04 CET 2007
512 - fixed a problem when you weakened references to Coro::State's
513 (patch by Yuval Kogman).
514
515 3.501 Wed Feb 28 12:44:07 CET 2007
516 - rename some global symbols as macosx from hell redefines
517 them without asking.
518
519 3.5 Tue Feb 13 20:22:53 CET 2007
520 - do AnyEvent model detection earlier, avoiding problems
521 caused by first using AnyEvent and later Coro::Event.
522 - implement and document Coro::Event event objects.
523 - fix a potential problem in Coro::Event causing crashes.
524 - initialise PL_comppad when creating a new coroutine,
525 avoids crashes on early coro destruction.
526
527 3.41 Mon Jan 22 19:19:49 CET 2007
528 - readline on Coro::Handle did not support undefined $/,
529 nor did it deliver partial lines on EOF or error.
530 - implement malloc fallback for stack allocation because
531 stupid broken idiotic OSX has a stupid broken
532 idiotic fits-the-whole-os mmap "implementation" and
533 my dick feels longer if Coro is portable even to
534 obsolete platforms.
535
536 3.4 Fri Jan 19 21:52:54 CET 2007
537 - remove t/09_timer.t, as it isn't really testing much
538 but was rather flaky in practise.
539 - async_pool coro would keep arguments and callback alive until
540 it was reused.
541 - cancellation of a coroutine could cause spurious idle calls
542 in cede_notself.
543
544 3.3 Sat Jan 6 03:45:00 CET 2007
545 - implement $coro->on_destroy.
546 - Coro::Event blocking semantics have been changed,
547 documented and - hopefully - improved.
548 - fix nice adding, not subtracting, from priority.
549 - fix ->prio and api_is_ready (patch by Mark Hinds).
550 - fixed an assert ("... == prev__cctx->idle_te")
551 that could errronously trigger.
552 - fix various large and small memleaks.
553 - use a (hopefully) more stable cancel implementation
554 that immediately frees the coroutine data.
555 - cede/cede_notself return a status now.
556 - added Coro::guard function.
557 - added a global coroutine pool for jobs (on my machine,
558 I can create and execute 48k simple coros/s with async,
559 and 128k coros with async_pool).
560 - Coro::AIO now uses the coroutine priority as io priority.
561
562 3.2 Fri Dec 22 05:07:09 CET 2006
563 - improve portability to slightly older perls.
564 - use cleaner coroutine destruction.
565 - simplify configuration for users.
566 - optionally (unrecommended) prefer perl functions over
567 their coro replacements.
568
569 3.11 Tue Dec 5 13:11:24 CET 2006
570 - fixed some bogus assert's, but as perl.h disables assert even
571 without NDEBUG (thank you very much), not too many people should
572 notice (that did include myself). Andreas Kรถnig noticed, though :)
573 - do not save/restore PL_sortcxix on >= 5.9.x, it doesn't seem to have
574 it. Also noticed by Andreas Kรถnig :)
575 - save/restore tainted status.
576 - verified to pass the testsuite on my 5.9.5.
577
578 3.1 Mon Dec 4 23:03:40 CET 2006
579 - INCOMPATIBLE CHANGE: $/ is now per-coroutine (but slow).
580 - incompatible change: transfer flags are now per-state.
581 - give a better error message on deadlock.
582 - document Coro::nready.
583 - enhanced testsuite.
584
585 3.01 Sun Dec 3 23:47:42 CET 2006
586 - forgot to include Coro::Timer.
587
588 3.0 Sun Dec 3 22:57:25 CET 2006
589 - the "FINALLY COMPLETELY STABLE" release (coming soon:
590 the "FAMOUS LAST WORDS" release).
591 - implement a new stack sharing algorithm, which uses a stack
592 pool (size currently hardcoded).
593 - make stack sharing mandatory (it no longer uses a heuristic).
594 - eval/die no longer cause weird problems under heavy use.
595 - Coro::Event could cause livelocks if it was used but
596 no Coro::Event watchers were used.
597 - Coro::Event now uses asynccheck as prepare does not
598 check for changed watchers.
599 - Coro::Event allows multiple waiting coros and will wake up one
600 per event.
601 - Coro::Event should be cleaner and more efficient now.
602 - new utility function Coro::unblock_sub.
603 - document the sad fact that Event is no longer reentrant.
604 - putting a coroutine into the ready queue twice could under
605 some circumstances lead to stack corruption.
606 - minor incompatibility: subclassing Coro::State is supported
607 directly now without going through a _coro_state member.
608 - state/coro switching is much faster now.
609 - very minor optimisations and code/documentation cleanup.
610 - avoid problems due to compiler inlining.
611 - removed timers from Coro::Timer -> use AnyEvent instead.
612 - replaced Coro::idle coroutine by (cleaner) idle handler.
613 - updated to newest libcoro.
614 - implement enhanced support for valgrind.
615 - implement is_ready and return value for ready.
616 - removed Coro::Cont, it was a misnomer (it's generators),
617 and it was somewhat annoying to get it right. Will come back
618 if somebody asks for it :)
619 - many ->wait methods and Coro::Event could return
620 spuriously without the event having happened.
621
622 2.5 Tue Nov 7 12:22:33 CET 2006
623 - made Coro::Util, Coro::Select, Coro::Handle and Coro::Socket
624 use AnyEvent, moved them to Coro/.
625 - added Coro::LWP which contains all the uglyness required to
626 make LWP non-blocking.
627 - should work with perl 5.9.x now (Andreas Kรถnig made me do it).
628 - fixed another bug in Coro::Select when the timeout was undef.
629 - reuse PL_start_env for all coros, saves some memory per coroutine.
630 - manage PL_top_env differently, hopefully to avoid panic: top_env.
631 - timeout argument was not properly used in Coro::Socket.
632 - allow limited forms of subclassing in Coro::Handle/Coro::Socket.
633 - emulate undocumented(!) functionality of IO::Socket required
634 by LWP(!!).
635 - updated eg/lwp to work with newer lwp's.
636 - remove "FATAL: uncaught exception" prefix. Coroutines that die
637 kill the whole process, just as exceptions in the main "coroutine"
638 did already.
639
640 2.1 Wed Nov 1 23:01:13 CET 2006
641 - fix a long-standing bug in Coro::Select where select with
642 zero timeout would instead change the current default filehandle.
643 - use a simpler and hopefully more robust way to clone padlists
644 (uses less memory and a perl function instead of our own).
645 - coro can now create a stack guard area on many architectures.
646 - Coro::AIO properly reexports additional functions from IO::AIO.
647 - updated libcoro with a workaround for OS X,
648 pach and testing by Michael Schwern.
649
650 2.0 Tue Oct 24 05:47:17 CEST 2006
651 - support additional aio requests in Coro::AIO.
652
653 1.9 (never properly released due to a glitch)
654
655 1.8 Thu Feb 2 00:59:06 CET 2006
656 - applied suggested patch by SAMV to avoid problems during stupid
657 mark & sweep gc run.
658 - applied patch by Scott Walters for 5.9.3 compatibility.
659
660 1.7 Tue Dec 27 01:41:58 CET 2005
661 - added Coro::AIO, a thin wrapper around IO::AIO.
662 - improved Makefile.PL explanations.
663
664 1.6 Mon Dec 12 21:30:05 CET 2005
665 - additionally save PM_curpm might fix as-of-yet
666 unseen problems with regex matches being attributed to the
667 wrong package.
668 - add t/10_bugs.t, which currently checks against imho broken
669 perls that use 0.26MB of stack space per Perl_magic_get
670 invocation (newer linuxes) as opposed to the 0.0002MB perl
671 normally uses.
672 - make stacksize configurable for the ultimate debian experience.
673
674 1.51 Mon Dec 12 18:48:36 CET 2005
675 - remove debugging warn() accidentally left in Coro::Select.
676
677 1.5 Tue Nov 29 13:32:44 CET 2005
678 - use Coro::Event inside Coro::Select to avoid spurious deadlocks.
679 - fix Coro::Select 'select'.
680 - strict'ify some modules.
681 - libcoro errornously restored the SIGUSR2 handler to SIGUSR1.
682 - use XSLoader in selected modules.
683 - remove some 5.6 compatibility cruft.
684
685 1.4 Tue Sep 6 00:11:05 CEST 2005
686 - libcoro did not take into account the trampoline on amd64,
687 when 'l'inux method was used.
688
689 1.31 Tue Aug 30 23:31:33 CEST 2005
690 - some portability fixes/workarounds.
691
692 1.3 Sat Aug 20 03:08:56 CEST 2005
693 - no code changes. module seems to work fine.
694 - improve Coro::State docs, remove reference to nonexisting
695 Coro::State::flush method.
696 - no longer autodetect windows, present it as an option instead.
697
698 1.2 Mon May 16 02:00:55 CEST 2005
699 - included libcoro.c earlier in State.xs, might improve portability.
700 - use a faster and possibly more stable Coro::Cont implementation.
701 - accept x86_64 in addition to amd64 for optimized linux-amd64 support.
702 - fix bugs Coro::SemaphoreSet that could cause locks never to be freed.
703 - fix bugs in CoroAPI.h and document it in Coro::MakeMaker.
704
705 1.11 Thu Mar 3 18:00:52 CET 2005
706 - change of contact address.
707
708 1.1 Tue Feb 22 20:51:16 CET 2005
709 - support [l]inux method on amd64.
710 - allow some context switches while compiling/eval'ing.
711
712 1.0 Tue Aug 10 03:47:27 CEST 2004
713 - Event 0.89 fixes the bug again.
714 - fix the remaining(??) memleak. I didn't free the cache-AV for
715 padlists, so one AV per created anon-sub leaked.
716
717 0.97 Fri May 14 15:23:32 CEST 2004
718 - remove stupid gcc extension that was an experiment only anyway.
719
720 0.96 Thu May 13 18:09:29 CEST 2004
721 - tell people if they have yet another broken Event variant.
722 - made Coro to work _better_ with threads (it seems that, as long
723 as you use Coros in one thread only everything will work fine.
724 Looking at the perl source the problem might be missing locking
725 between Coro and some perl internal routines. That's not easily
726 fixable).
727 - (hopefully) got rid of the static xnext inside transfer.
728
729 0.95 Sun Feb 15 03:20:28 CET 2004
730 - removed Coro::State::flush, implemented a new and much
731 faster caching mechanism by attaching magic to the CV.
732 - a memleak with (real) closures remains.
733 - dramatically reduced size of stacks to 65536 (linux/x86), after
734 observing that even myhttpd never had more than a kilobyte
735 of stack in use (stack is still only physically allocated in
736 pagesize increments for systems suporting mmap).
737 - fix for a "cygwin-64" architecture added, proposed by Stefan Traby.
738 - fix for threaded perl.
739
740 0.9 Sun Nov 23 04:32:39 CET 2003
741 - fix a few very hard to track down but (of course) stupid
742 memory leaks.
743 - stringify version number to avoid locale problems :(.
744 - perl 5.9.0 does not have SvPADBUSY, reported by Scott Walters,
745 hopefully ignoring it "just works".
746
747 0.8 Wed Nov 5 19:38:40 CET 2003
748 - port to perl5.005_03, but only 5.8.x+ is supported!
749 - honor LocalAddr even without LocalPort in Coro::Socket.
750
751 0.7 Tue May 27 03:12:38 CEST 2003
752 - the version jump indicates some level of testing, not gobs
753 of new features.
754 - uh... I found the prompt function in ExtUtils::MakeMaker.
755 Highly correct stuff, that is...
756 - fixed(?) a bug with die's in coroutines causing "panic: top_env".
757 the fix is not well understood by the author, so beware :(.
758
759 0.652 Thu May 8 02:54:46 CEST 2003
760 - Applied patch by Slaven Rezic to set default to "s" on FreeBSD,
761 cause version 4 doesn't have ucontext.
762 - Benjamin Reed reported that setjmp works fine on darwin,
763 so preselect it.
764
765 0.651 Sat Mar 29 15:00:23 CET 2003
766 - fix a bug in Coro::Handle where some bytes could get lost
767 on reading, reported by jason@nichego.net.
768
769 0.65 Sun Mar 23 00:08:26 CET 2003
770 - added a README.linux-glibc.
771 - new module Coro::Select.
772 - also save/restore PL_comppad, fixes segfaults with 5.8.1.
773 (I never claimed that I know what I am doing ;).
774 - set default to setjmp/longjmp on non-x86-linux.
775
776 0.6 Thu Nov 21 11:09:06 CET 2002
777 - made lazy context switching the default.
778 - fixed the bug where SemaphoreSet::guard
779 was the same as timed_guard and timed_guard was missing.
780 - fix a memleak in Coro::Timer.
781 - reclassified context sharing as not experimental.
782
783 0.534 Sun Apr 14 03:05:12 CEST 2002
784 - fixed backspace => tab in header parsing (myhttpd).
785 - added eg/readline, for elmex.
786 - Coro::Event now calls ->start on first call instead of
787 ->again, which created wrong timeouts for the first ->next.
788 - fixed a bug where Coro::Socket returned a closed socket instead
789 of nothing on ECONNREFUSED and similar error conditions.
790
791 0.533 Mon Feb 18 18:49:40 CET 2002
792 - adapt to finally working Event-0.85.
793 - fixed buggy event initialization (trapped under windows, but how
794 could it possibly have worked anywhere?)
795
796 0.532 Wed Jan 16 02:45:32 CET 2002
797 - removed the reference to pp_entersub, might work on
798 (some) win32 perls now (testsuite works under cygwin,
799 without Event, which freezes).
800
801 0.531 Mon Dec 10 22:18:44 CET 2001
802 - Fixed a bug in SemaphoreSet::timed_down.
803
804 0.53 Tue Nov 27 21:11:13 CET 2001
805 - some tests for Event.
806 - slightly different internal architecture to get refcnt into the object
807 and not the reference. might fix a few bugs, certainly did introduce
808 new ones.
809 - Coro::Timer (independent of Event).
810 - new timed_wait functions for Coro::Signal, Semaphore, SemaphoreSet.
811
812 0.52 Tue Nov 6 21:36:18 CET 2001
813 - ported to cygwin (trivial).
814 patch by Gerrit P. Haase <gp@familiehaase.de>.
815 - small setjmp code fix by Sullivan.DanielJ@epamail.epa.gov.
816
817 0.51 Thu Nov 1 20:39:01 CET 2001
818 - terminate/cancel now work properly (otherwise termination
819 could cause a "next coroutine is not and contains not..."
820 error.
821 - added Coro::Socket::shutdown.
822 - Coro::Event::loop no is the same as Event::loop.
823 - implemented terminate with args + join.
824
825 0.5 Fri Sep 28 16:15:35 CEST 2001
826 - fixed "print" on a Coro::Handle. The print method worked.
827 - small tweaks (seem to reduce memory consumption a lot)
828 in various modules.
829 - splendid use of "no warnings" scattered throughout.
830 - added Coro::Handle::rbuf, fixed Coro::Handle::fh.
831
832 0.49 Sun Sep 16 02:42:45 CEST 2001
833 - changed some method calls to function calls for speed
834 inside Coro::Handle.
835 - make Coro::Handle use an array instead of a hash for
836 speed reasons.
837 - IRIX mystery solved: it's SGI's NT, after all: "standard, huh?".
838 sjlj and ucontext should now work.
839 - IRIX-specific port for libcoro.
840 - swapped order of accept results to match IO::Socket.
841 - changed getsock/peername to sock/peername to match IO::Socket.
842 - fixed a bug that caused segfault when returning to main task
843 under some circumstances.
844 - other bugfixes.
845
846 0.45 Sun Sep 2 02:54:01 CEST 2001
847 - new method Coro::Handle::timeout.
848 - corrected speling of set...name to get...name in Coro::Socket.
849 - Coro::Socket::accept now returns a Coro::Socket, not a Coro::Handle.
850 - Coro::Handle now supports fileno and FILENO.
851 - added eg/myhttpd, a web server.
852 - Coro::Socket now accepts numeric ports as well as "port(number)"
853 syntax.
854 - moved some scheduling primitives into xs code. more to come.
855 - new simple priority system.
856 - implemented Coro::Event using XS, almost four times faster!
857 - small memory corruption problem fixed (boy that was difficult).
858
859 0.13 Wed Aug 8 16:53:07 CEST 2001
860 - new method Coro::Handle::readline.
861 - added eg/lwp and Coro::Handle::autoflush, to make LWP work.
862 (see eg/lwp on how to make LWP non-blocking).
863 - renamed Coro::Socket::new_inet to new.
864 - added Coro::Util, some utility functions.
865 - die/eval now works better (still get annoying (but true ;)
866 "Callback called exit" messages).
867 - Coro::Handle now supports timeout, and Coro::Socket Timeout.
868 - much better Coro::Socket support.
869
870 0.12 Fri Jul 27 04:19:01 CEST 2001
871 - do not use mmap if neither MAP_ANON nor MAP_ANONYMOUS is defined.
872 - very experimental stack sharing algorithm. not 100% safe but
873 should work well in practise ;)
874 - added Coro::RWLock.
875 - Coro::Specific now works.
876
877 0.11 Tue Jul 24 22:49:21 CEST 2001
878 - added specialized hack for newer and older linux versions (fast).
879 - renamed Coro::Event::IO to Coro::Handle.
880 - new module Coro::Socket.
881
882 0.1 Tue Jul 24 01:47:53 CEST 2001
883 - release candidate 3. A rather new internal structure :(
884 - the great renaming: Damian Conway gave me a suitable replacement
885 for yield.
886 - added Coro/libcoro, a portable coroutine implementation for C
887 (not even perl-dependent!!!), which can be used by Coro::State.
888 - renamed SAVE_DEFSV (on xs-level only) to avoid symbol clash in
889 perl-5.6.
890 - new function Coro::Event::idle.
891 - the idle process is now overriden
892 by default in Coro::Event.
893 - Coro::Channel now enforces the size.
894 - canceling events now works in all cases.
895 - Coro state now include $_ and $@.
896 - yet another bug workaround that I do not really understand :(
897 - new module Coro::Event::IO (very undocumented) to do
898 non-blocking i/o.
899 - performance tuning.
900
901 0.08 Thu Jul 19 06:13:25 CEST 2001
902 - release candidate 2 (new functionality).
903 - add Coro::State::flush function.
904 - transfer now supports a flags value (mostly for speed ATM).
905 - might compile and work in the presence of threads now.
906 - continuations have a saner syntax.
907 - no more memleaks.
908
909 0.07 Tue Jul 17 17:40:18 CEST 2001
910 - release candidate 1 ;)
911 - slightly nicer code.
912 - fixed a scheduling bug in Coro::Event.
913
914 0.06 Tue Jul 17 04:23:24 CEST 2001
915 - ok, I found the showstopper - the same sub must not be
916 re-entered in two different coroutines, otherwise => crash. I
917 see no easy solution to this problem, except by walking the call
918 chaing and saving/restoring the cv's, which is what I do now.
919 - memory leaks still latent, especially at thread termination.
920 - Coro::Event now works (haha).
921
922 0.05 Sun Jul 15 17:32:20 CEST 2001
923 - fixed a few issues in Coro::Event.
924 - I forgot to include Coro::Event in 0.04 :(:(:(
925
926 0.04 Sun Jul 15 05:24:59 CEST 2001
927 - @_ is now properly localized.
928 - Coro::State is now easier subclassable.
929 - Coro::Cont now coroutine-aware.
930 - Coro::Specific is a low-overhead module to create
931 coroutine-specific vars.
932 - Coro::Event provides a simple interface to Event.
933
934 0.03 Fri Jul 13 14:51:52 CEST 2001
935 - transfer() now implemented in XS (beware).
936 - new module Coro::Cont for really faked continuations.
937 - big internal architecture changes: Coro::State is now
938 really low-level and can thus be used to implement other
939 interesting things, While "Coro::" implements a process-like
940 model. Still crude and subject to change.
941 - $_ and $@ are no longer being localized.
942
943 0.02 Tue Jul 10 01:38:17 CEST 2001
944 - implemented "async" attribute.
945 - $_ and $@ are now localized.
946 - added Coro::Channel.
947 - more testcases, still no docs.
948
949 0.01 Tue Jul 3 02:18:41 CEST 2001
950 - original version; copied from Convert::Scalar.
951