ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/Coro/Changes
Revision: 1.466
Committed: Wed Nov 26 09:30:01 2008 UTC (15 years, 5 months ago) by root
Branch: MAIN
Changes since 1.465: +2 -0 lines
Log Message:
*** empty log message ***

File Contents

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