ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/Coro/Changes
Revision: 1.598
Committed: Mon Oct 8 23:16:15 2012 UTC (11 years, 7 months ago) by root
Branch: MAIN
CVS Tags: rel-6_10
Changes since 1.597: +4 -2 lines
Log Message:
6.10

File Contents

# Content
1 Revision history for Perl extension Coro.
2
3 TODO: should explore PerlIO::coroaio (perl leaks like hell).
4 TODO: channel->maxsize(newsize)?
5 TODO: http://www.microsoft.com/msj/archive/s2ce.aspx
6 TODO: __GCC_HAVE_DWARF2_CFI_ASM
7
8 6.10 Tue Oct 9 01:14:27 CEST 2012
9 - updated ecb.h, it had a typo that caused it to not compile on many
10 big endian systems (reported by many people).
11 - disable memory fences in ecb.h to improve portability.
12
13 6.09 Sat Oct 6 23:25:02 CEST 2012
14 - Coro::EV I/O watchers were not interruptible by exceptions
15 (Coro::State::throw) (testcase by sten).
16 - ->throw now puts threads into the ready queue, as this seems to
17 be expected by existing code, and code that doesn't cope with spurious
18 wakeups needs fixing anyway.
19 - use fd -1 in mmap.
20 - cast I32 to int in error message printf.
21 - warn about broken so-called "hardened" kernels.
22
23 6.08 Fri Apr 13 12:05:47 CEST 2012
24 - be more aggressive about exiting like perl does - formerly,
25 exiting from the non-main thread would not execute END blocks.
26
27 6.07 Fri Nov 11 21:21:48 CET 2011
28 - work around a bug in PerlIO (setting $SIG{__WARN__} to a PVCV).
29 - update ecb.h.
30
31 6.06 Mon Aug 8 23:59:48 CEST 2011
32 - cygwin unfortunately patches the stack at runtime, so we use the pthreads
33 backend, which is an order of magnitude slower. unfortunately, cygwins
34 pthread implementation isn't very complete either, so allocate the stack
35 twice just to be sure.
36 (note: cygwin also enables mymalloc, which is NOT THREADSAFE ON WINDOWS,
37 in its ithreaded perl - best recompile cygwin and use the 'w'indows
38 backend for much better performance. also disable ithreads for
39 even better performance...).
40
41 6.05 Thu Aug 4 21:36:36 CEST 2011
42 - blush, condvar values would not be propagated from send to recv anymore
43 (reported by Chip Salzenberg).
44 - use exponential increase for the readline buffer length in
45 Coro::Handle. also reduce initial allocation to 1020 from 4096 bytes.
46
47 6.04 Wed Aug 3 17:15:45 CEST 2011
48 - use even more efficient and more compatible condvars for
49 compatibility to AnyEvent 6.x :)
50 - more inconsequential ecb.h updates.
51
52 6.03 Wed Aug 3 11:41:30 CEST 2011
53 - change how Coro patches AnyEvent condvars for compatibility to
54 AnyEvent 6.x.
55 - update ecb.h, to no longer include <pthread.h> in case WinNT.h
56 hasn't been included.
57
58 6.02 Wed Jul 13 04:35:19 CEST 2011
59 - "improve portability to Gentoo" - gentoo manages to put perl variables
60 in memory areas that are farther than 2gb apart, which the jit couldn't
61 handle and barfed. now it's just a bit slower on gentoo and similar
62 systems.
63
64 6.01 Sun Jul 3 12:31:14 CEST 2011
65 - workarounds are good, but the test for whether pthreads are used
66 was not good. thisone should be better.
67 - check differently whether gcc generates cfi instructions itself.
68
69 6.0 Wed Jun 29 19:43:35 CEST 2011
70 - INCOMPATIBLE CHANGE: unreferenced coro objects will now be
71 destroyed and cleaned up automatically (e.g. async { schedule }).
72 - implement a JIT compiler for part of the thread switch code,
73 which gives a 50% speed improvement on threaded perls, and
74 about 4% on non-thraeded perls (so threaded perls now finally
75 reach about half the speed of non-threaded perls).
76 - slightly modernise Coro::Intro, add section about rouse functions.
77 - avoid DEFSV and ERRSV, giving another 10% improvement
78 in thread switching.
79 - Coro::State->is_destroyed is now called is_zombie.
80 - implement a Coro->safe_cancel method that might fail, but
81 cancels in a "safer" way if it succeeds.
82 - add preliminary support for DEBUGGING perls.
83 - get rid of two hash-accesses when initialising a new Coro - this
84 speeds up coro creation by almost a factor of two.
85 - croak when a coro that is being cancelled tries to block
86 (e.g. while executing a guard block), instead of crashing or
87 deadlocking.
88 - use a more robust and also faster method to identify Coro::State
89 objects - speeds up everything a bit.
90 - implement Coro->cancel in XS for a 20% speed improvement, and to
91 be able to implement mutual cancellation.
92 - speed up context switches by a percent or two by more efficiently
93 allocating context stack entries.
94 - implement Coro->join and Coro->on_destroy in XS for a speedup and
95 a reduction in memory use.
96 - cancelling a coro while it itself is cancelling another coro is
97 now supported and working, instead of triggering an assertion.
98 - be a bit more crash-resistant when calling (buggy) on_destroy
99 callbacks (best effort).
100 - move on_destroy into the slf_frame, to allow extension slf
101 functions to have destructors.
102 - get rid if coro refcounting - simply crash in other interpreter
103 threads by nulling the pointers on clone.
104 - simplify warn/die hook handling when loading Coro - the convoluted
105 logic seems to be no longer neccessary.
106 - use libecb instead of our own home-grown gcc hacks.
107 - document alternatives to Coro::LWP. Please use them :)
108 - work around another mindless idiotic NEEDLESS bug in openbsd/mirbsds
109 sigaltstack. Really. wine suffers from it, erlang suffers from it,
110 and it's known since at least 2006.
111
112 5.372 Wed Feb 23 06:14:30 CET 2011
113 - apparently mingw doesn't provide a working gettimeofday, try to
114 work around that by relying on Time::HiRes (indirectly brought to
115 my attention by Max Maischein).
116 - fix some portability issues when Time::HiRes was used.
117
118 5.371 Mon Feb 21 14:36:08 CET 2011
119 - backport to windows process emulation code again.
120
121 5.37 Sat Feb 19 07:49:44 CET 2011
122 - add a big "Coro thread life cycle" section to "man Coro".
123 - try a tentative workaround against the breakage that 5.13 has
124 introduced without depreciation period. sigh.
125 - no longer use Time::HiRes if gettimeofday is available, which
126 saves quite a lot of memory.
127
128 5.36 Sun Feb 13 05:33:41 CET 2011
129 - automatically load Coro::Channel, Coro::RWLock, Coro::Semaphore,
130 Coro::SemaphoreSet, Coro::Signal and Coro::Specific on first "new"
131 method call.
132 - undocument Coro::Timer::sleep and obsolete whole module.
133 - optimise Coro::Timer::timeout memory and cpu usage.
134 - slightly updated Coro::Intro for recent changes.
135 - do not initialise PL_dirty anymore.
136
137 5.25 Thu Nov 11 01:08:39 CET 2010
138 - try a different approach on netbsd - netbsd 5 finally has marginally
139 working pthreads, but still broken ucontext/sigaltstack.
140 - openbsd 4.8 finally got their act together, Coro works out of the box
141 with asm, setjmp and pthreads (no change, just informational).
142
143 5.24 Sat Oct 23 11:27:12 CEST 2010
144 - port to the EV 4.0 API.
145 - work around bugs in mingw32, making strawberry perl work
146 out of the box.
147 - correctly modify Coro::AIO function prototypes
148 so that they reflect the "no optional parameters" rule.
149 - "ported" libcoro to C++.
150
151 5.23 Mon May 17 18:50:42 CEST 2010
152 - be more resistant to ordering changes when initialising
153 Coro::AnyEvent, Coro::EV and Coro::Event (reported by Matthias
154 Waldorf).
155 - document that perl 5.12 deliberately removed support for cloning.
156
157 5.22 Wed Apr 14 03:55:35 CEST 2010
158 - correctly return udnef on errors in Coro::Handle::read/write
159 (testcase by Marc Mims).
160 - convert Coro::Util into a "perl compatibility wrapper" - the functions
161 are less useful now, but are drop-in replacements for existing
162 functions, listing better alternatives in the documentation. This also
163 fixes a bug in Coro::LWP which naively substituted Socket::inet_aton
164 with Coro::Util::inet_aton.
165 - do not override $Coro::idle unconditionally in Coro.pm, as other
166 modules could have provided their own idle coro already
167 (for exmaple, Coro::AnyEvent).
168 - fix Coro::Util::gethost* functions.
169 - Coro::Timer corretcly exports it's symbols (reported by Hideki Yamamura).
170
171 5.21 Wed Dec 16 07:19:51 CET 2009
172 - automatically load Coro::AnyEvent when AnyEvent and Coro are used
173 together.
174 - add some examples on how to combine other event loops with Coro in
175 Coro::AnyEvent, and how to run it (and not to block). Seems to be
176 the most common source of confusion.
177 - try to catch people naively blocking in an event callback.
178 - work around the perl filehandle bug issue in conjunction with
179 older common::sense (as indirectly pointed out by ZSystem).
180 - clarify the "not from signal handlers" section.
181
182 5.2 Sun Oct 4 14:54:24 CEST 2009
183 - Coro::Storable destroyed the prototypes of the functions it wrapped.
184 - export rouse_cb and rouse_wait by default now.
185 - fix various prototype mismatches in Coro::AnyEvent and Coro::Handle.
186 - new method $state->swap_sv.
187 - added section on "windows process emulation" to the manpage,
188 after a not-so-fruitful (nor-friendly) "discussion" with chip
189 salzenberg (discussion implies arguments, but his only arguments
190 were ad-hominems, one wonders why he started it in the first
191 place). I hope this explains it well enough for him to understand,
192 and maybe well enough for others to understand.
193 - use common::sense everywhere now.
194 - idle callbacks are no longer supported, use idle coros instead.
195 - print a thread listing when a deadlock is detected.
196
197 5.17 Sat Aug 22 23:09:31 CEST 2009
198 - work around a bug in the perl debugger causing crashes
199 when running under the debugger by marking _pool_handler as nodebug.
200 - speed up Coro::async considerably.
201 - try some hacks to get netbsd to work "more often" - their broken
202 setjmp/longjmp, ucontext *and* phtreads are really hard on Coro.
203 - convert Coro to AE 5.0 API.
204
205 5.162 Tue Jul 28 04:04:03 CEST 2009
206 - perl 5.8.2 is now minimum requirement.
207 - skip t/19_handle.t on broken windows perls.
208
209 5.161 Wed Jul 22 04:47:38 CEST 2009
210 - Coro::AnyEvent::poll could have a different prototype when EV was
211 used as backend (analyzed by Tatsuhiko Miyagawa).
212 - Coro::AnyEvent errornously initialised the event loop when loaded,
213 not on demand.
214 - try to workaround rare */t/01_unblock.t failures.
215
216 5.16 Tue Jul 21 01:44:37 CEST 2009
217 - Coro::AnyEvent failed to hook into the event loop
218 when no threads had been readied between detecting
219 the event loop and actually running it.
220 - considerably speed up Coro::Select by taking avdantage
221 of AnyEvent > 4.8 and some other optimisations.
222 - implement paragraph readline mode in Coro::Handle
223 (based on patches by Zsbán Ambrus).
224 - replace WSAEINPROGRESS by WSAEWOULDBLOCK (reported
225 and analyzed by Yasuhiro MATSUMOTO).
226 - clarified libcoro license and copyright.
227 - someone stole my EXTRA_META!!!
228 - implement Coro::Select::patch_pp_sselect and it's brother,
229 for hardcode select overriding.
230
231 5.151 Mon Jul 6 05:41:57 CEST 2009
232 - backport to windows process emulation code again (patch by
233 Yasuhiro MATSUMOTO).
234 - slightly update Coro::MakeMaker.
235
236 5.15 Tue Jun 30 10:28:06 CEST 2009
237 - deprecate Coro::Socket, document how to get ipv6 support via
238 AnyEvent::Socket instead.
239 - implement signal->wait ($cb) interface, similar to semaphores.
240 - work around SvOK not supporting getmagic, so we have to getmagic
241 to test for undef :/ (reported by Matthias Waldorf).
242 - load Coro::AnyEvent in all modules using AnyEvent.
243 - work around perl corrupting our internal data structures,
244 reported by Tokuhiro Matsuno.
245 - enable per-coroutine real and cpu time gathering
246 (Coro::State::enable_times).
247
248 5.14 Wed Jun 24 01:37:48 CEST 2009
249 - provide explicit functions to "cede" to the event loop in Coro::AnyEvent,
250 as this seems to have been a difficult concept (poll, sleep, idle,
251 idle_upto).
252 - add Coro::AnyEvent::readable/writable functions.
253 - clarify Coro::EV/Event/AnyEvent manpages.
254 - free per-thread global scalars in the thread calling ->cancel, to
255 avoid crashes when $_, $@ etc., are magical but some of those
256 had already been freed.
257 - "unexperimentalise" the callback interface for Coro::Semaphore.
258 - speed up ready queue management/context switching by using a linked
259 list instead of an array (~5-10%).
260 - implement "watch" command in Coro::Debug shells.
261 - for fun, implement time-slicing as an example in the manpage.
262 - if AnyEvent detects EV or Event, but we don't have Coro::EV or
263 Coro::Event, use the normal AnyEvent handling instead of dieing
264 (the same is true for Coro::Handle).
265 - properly document Coro::EV::timed_io_once.
266 - avoid unneccessary ->cancel calls in Coro::Handle.
267 - maybe make it work on mingw32 with win32 backend
268 (based on patch by Yasuhiro Matsumoto).
269
270 5.132 Fri May 29 09:00:39 CEST 2009
271 - do not keep a reference to the argument itself in
272 Coro::Semaphore::guard, as it could change later.
273 - support SO_RCVBUF/SO_SNDBUF nonstandard Coro::Socket options,
274 should support a prepare callback.
275
276 5.131 Mon Mar 16 23:20:37 CET 2009
277 - implement and document Coro->suspend, Coro->resume.
278 - fix Coro::Select implementation to not (often) close
279 the passed file descriptors (testcase provided by pippijn).
280
281 5.13 Mon Dec 15 21:51:42 CET 2008
282 - EXPERIMENTAL: implement dynamic winds (on_enter/on_leave).
283 - don't set diehook to C<undef> but instead to NULL, to avoid
284 spurious warnings.
285 - fix a lot of bugs in Coro::SemaphoreSet.
286 - Coro::SemaphoreSet will less often create a semaphore needlessly.
287 - add Coro::SemaphoreSet::count and wait methods.
288 - take advantage of the new Guard module.
289 - deprecate Coro::guard.
290 - try to fix the dreaded 01_unblock tests once more. I hate it when
291 testsuites need more fixing than the code they are supposed to test.
292 - croak in more cases when a required callback isn't resolvable.
293 - fix some minor issues in Coro::State->call/eval.
294 - use current coroutine context instead of a temporary one
295 when temporarily switching to another coroutine.
296 - do not call C-level on_destroy handlers during global destruction,
297 to avoid needless segfaults.
298
299 5.12 Sun Dec 7 13:30:38 CET 2008
300 - add default config for MirOS, which seems to be a bug-to-bug
301 compatible fork of openbsd ("world domination by releasing
302 openbsd cvs before the openbsd folks do it" or so :).
303 - free_padlist did destroy the names pad, not good, but didn't
304 seem to bother perl - this could fix issues such as eval ""
305 inside a function called from multiple coroutines.
306 - use a different method to detect destruction time.
307 - be more careful when freeing padlists just before global
308 destruction.
309 - fixed and expanded the call/cc example.
310 - renamed _terminate to _coro_run.
311 - new method Coro::Channel->shutdown.
312 - try pthreads on openbsd <4.4 (broken sigaltstack, will
313 pthreads fare better?).
314 - be less picky about destroying "a" running coroutine.
315
316 5.11 Tue Nov 25 21:49:05 CET 2008
317 - DEBUGGING in 5.10.0 is a mess: it constantly flags perfectly
318 working code with failed assertions, introducing more bugs than
319 it fixes, requiring elaborate workarounds :(
320
321 5.1 Mon Nov 24 08:54:59 CET 2008
322 - wrote a small introductory tutorial - Coro::Intro.
323 - convert Coro::Timer, Coro::Select and Coro::Util to rouse API.
324 - Coro::Select did errornously dup the file descriptors
325 and didn't work with all AnyEvent backends.
326 - Coro::Select wasn't imported correctly form Coro::LWP, causing blocking
327 LWP data transfers.
328 - disassociate c contexts from coro objects - this is agruably more
329 correct, but mostly allows sharing of cctxs between coro and state
330 objects, for added memory savings and speed increases.
331 - bumped $Coro::POOL_RSS up to 32kb by default.
332 - no longer set the optype to OP_CUSTOM, as B::* understandably
333 doesn't like this very much (and we *are* a type of entersub).
334 - implement state cloning, just to prove that call/cc can be done.
335 - automatically load Coro::AnyEvent in Coro::Handle.
336 - wrap ->cancel calls in eval inside Coro::Handle as EV watchers
337 do not have this method (and don't need it either).
338 - speed up generic anyevent methods in Coro::Handle by using rouse
339 callbacks.
340 - allow coroutines in $Coro::IDLE, speeding up Coro::AnyEvent and
341 others. It also makes the debugger happier, as you can trace
342 through the idle threads now.
343 - add comppad_name* and hints ($^H, %^H) to per-thread variables.
344 - eg/event was pretty broken.
345 - better 5.8.6 compatibility.
346
347 5.0 Thu Nov 20 10:35:05 CET 2008
348 - NEW ARCHITECTURE: use the latest 4.x version if you experience
349 stability issues.
350 - bump API version to 7 - all dependents must be recompiled.
351 - removed timed_* functions - they were not being used anyways
352 and should be replaced by a more generic mechanism -
353 and were annoying to support anyways :)
354 - removed SemaphoreSet's waiter method - use sem method instead.
355 - Coro::Semaphore->adjust didn't correctly wake up enough waiters.
356 - async_pool did free a scalar value twice
357 ("Attempt to unreference...").
358 - fix a longstanding bug where calling terminate on a coro that
359 was waiting for a semaphore that was just becoming available
360 would cause a deadlock (semaphore would get into a state where
361 it was available but waiters were still blocked).
362 - calling throw on a coroutine that is waiting for a semaphore will
363 no longer make it acquire the semaphore (and thus leak a count).
364 - perl's process emulation is now not even theoretically supported
365 anymore.
366 - new functions Coro::rouse_cb and Coro::rouse_wait for easier
367 conversion of callback-style to blocking-style.
368 - new methods $coro->schedule_to and ->cede_to, to specifically
369 schedule or cede to a specific coroutine.
370 - new function Coro::Semaphore::wait.
371 - use named constants in Coro::Channel (Richard Hundt).
372 - directly patch the entersub opcode calling SLF functions (cede,
373 transfer and so on). this does speed up context switching, but
374 more importanly, it frees us from the hardcoded behaviour of
375 entersub, so we might actually be able to return something from
376 those functions and atcually create new ones.
377 - take advantage of __builtin_frame_address on gcc.
378 - expose THX in coroapi (not sure whether this was a wise decision,
379 as "threaded" perls are running at half speed anyways).
380 - implement execute_slf (schedule-like-function) interface that makes
381 it possible to implement schedule-like-functions in XS.
382 - use new SLF interface to massively speed up Coro::EV by roughly a
383 factor of two.
384 - used new SLF interface to massively speed up Coro::Semaphore by a
385 factor of three.
386 - used new SLF interface to speed up Coro::AIO by roughly a factor of
387 four and reduce its memory usage considerably.
388 - implement Coro::SemaphoreSet purely in terms of Coro::Semaphore,
389 for a nice speedup and vastly more correct behaviour. Also implement
390 a new method "sem" to get at the underlying semaphore object.
391 - implement Coro::Channel in terms of Coro::Semaphore, for a moderate
392 (in comparison) 20-40% speedup.
393 - used new SLF interface to reimplement Coro::Signal gaining
394 some unknown (because I was too lazy), but certain, speedup, and also
395 making signals reliable for the first time.
396 - used new SLF interface and other optimisations to speed up async_pool
397 by a factor of two. It also doesn't rely on perl's exception mechanism
398 to exit anymore. The overhead for terminating an async_pool, coro over
399 a normal async is now very small.
400 - sped up coroutine creation/destruction by 40%.
401 - forgot to include Coro/libcoro/README in the dist for all these years.
402 - work around a freebsd pthreads bug (manual testcancel is required as
403 pthread_cond_wait isn't a cancellation point on freebsd).
404 - use new rouse functions to speed up and simplify Coro::BDB.
405 - make "prefer perl native functions" work with threaded perls.
406 - condense Coro::Debug ps output, hint at v and w flags.
407 - (libcoro) lots of minor cleanups and portability improvements.
408
409 4.914 Wed Nov 19 12:54:18 CET 2008
410 - fix a disastrous bug in the readline optimisation
411 introduced in 4.801.
412
413 4.913 Sat Nov 15 07:58:28 CET 2008
414 - async_pool did free a scalar value twice
415 ("Attempt to unreference...").
416
417 4.912 Thu Nov 13 18:31:23 CET 2008
418 - minor cleanups.
419 - use much larger stacks on linux and perl < 5.8.8.
420 - Coro::Debug::new_unix_server did not unlink the socket
421 when destroyed.
422
423 4.911 Tue Nov 11 04:26:17 CET 2008
424 - "port" to threaded perls.
425
426 4.91 Mon Nov 10 05:36:38 CET 2008
427 - the ->throw exception object no longer leaks.
428 - creating a new cctx leaked a scopestack entry (memleak).
429 - new coroutines didn't get created with a zero flags field
430 (unknown impact).
431 - calling ->throw on a not-yet-started coroutine should now work
432 instead of being ignored.
433 - ->throw is now supported on Coro::State objects.
434 - clean up cctx creation code a bit.
435 - entersub is actually a UNOP, not a LOGOP (not a bugfix).
436
437 4.9 Sat Nov 8 17:45:27 CET 2008
438 - (libcoro) did not preserve rbp with CORO_ASM (we are getting there).
439 - (libcoro) no longer leak threads in the experimental pthread backend,
440 also speed it up considerably.
441 - (libcoro) do not rely on makecontext passing void *'s unscathed.
442 - fix compiletime dependencies on libcoro in the Makefile.
443 - cctx_count wasn't always updated properly.
444 - Coro::State::cctx_stacksize wasn't applied correctly.
445 - new function Coro::State::cctx_max_idle.
446 - the default max number of idle C contexts is now 4.
447 - (libcoro) try harder to get _setjmp/_longjmp.
448 - (libcoro) cleanup and extend the libcoro API to officially
449 allow the creation of empty source contexts.
450 - very experimental workaround for the totally broken netbsd platform.
451 - tried to hack around openbsd bugs.
452
453 4.804 Wed Nov 5 16:36:00 CET 2008
454 - Coro::Debug::new_unix_server would not create a non-blocking listening
455 socket, sometimes causing freezes.
456 - (libcoro) fix misaligned stack points for setjmp and assembly
457 methods, which can cause crashes on x86/x86_64 with a sufficiently
458 aggressive compiler.
459 - new function: Coro::Debug::new_tcp_server.
460 - move ->throw into the Coro class because it only works on coro objects.
461
462 4.803 Mon Nov 3 17:16:12 CET 2008
463 - (libcoro) use a global asm statement to become independent of gcc
464 otpimisations for CORO_ASM (thanks to pippijn for the idea).
465 - try to workaround yet another broken bsd, this time dragonfly.
466
467 4.802 Thu Oct 30 10:56:12 CET 2008
468 - support -fno-omit-frame-pointer on x86 with the assembly method.
469 - tune 01_unblock.t tests a bit.
470
471 4.801 Wed Oct 22 18:33:37 CEST 2008
472 - improve readline speed for very long "lines".
473 - backport to 5.8.8.
474
475 4.8 Thu Oct 2 13:34:40 CEST 2008
476 - new function Coro::AIO::aio_wait.
477 - Coro.:AIO and Coro::BDB now "use Coro::AnyEvent".
478 - greatly speed up and reduce memory usage of Coro::AIO requests.
479 - implement some other µ-optimisations.
480
481 4.749 Mon Sep 29 14:40:12 CEST 2008
482 - port to slow and broken pseudo-threaded perls. (courtesy pippijn).
483
484 4.748 Sat Sep 27 14:03:19 CEST 2008
485 - implement, but do not document, PerlIO::cede(granularity).
486 - Coro::Storable forgot to wrap Storable::pstore.
487 - work around the multitude of leaks and memory corruption
488 bugs in PerlIO::via by using our own C-level perliol. As a side
489 effect, Coro::Storable is now much, much, much faster.
490
491 4.747 Tue Sep 23 01:59:41 CEST 2008
492 - fix a per-cv memleak (one empty array was leaked per
493 code reference).
494 - avoid a crash in coro->call|rss when the coroutine was already
495 destroyed (most noticably when using Coro::Debug::ps :)
496 - also protect *Storable::FILE.
497 - push up default storable granularity to 20ms.
498
499 4.746 Sun Sep 21 03:22:20 CEST 2008
500 - be more insistent on locking Storable against reentrancy
501 in Coro::Storable.
502 - move swap_def?v and throw to Coro::State, as documented.
503
504 4.745 Thu Jul 24 00:14:38 CEST 2008
505 - remove debugging code related to MgPV_nolen_const, also try to
506 make it compile with perl 5.8.6 (yes, apple apparently loves
507 outdated software). Reported by John S.
508
509 4.744 Tue Jul 8 22:06:35 CEST 2008
510 - correctly provide default DIE/WARN handlers as documented.
511 - also overwrite PL_vtbl_sigelem.svt_clear, even though current
512 implementations inside perl work fine for us.
513
514 4.743 Mon Jun 16 00:21:57 CEST 2008
515 - when using Coro::EV without running EV::loop it could
516 result in busy-waiting for events, this has been fixed.
517 - reduce codesize and improve performance by using EV_DEFAULT_UC.
518
519 4.742 Sat May 31 14:10:21 CEST 2008
520 - implement a workaround for (some) perl <5.8.8 versions.
521 - require EV 3.3+.
522
523 4.741 Fri May 30 23:33:09 CEST 2008
524 - tell netbsd how utterly broken their imitation of an OS is
525 and refuse to build by default if pthreads are in use.
526 - switch to "s" method on all bsd's by default, as their ucontext
527 stuff seems just too broken.
528 - fix a bug in Coro::Select.
529
530 4.74 Thu May 29 20:05:31 CEST 2008
531 - do not test Coro::LWP for lack of dependencies.
532 - do not test Coro::Select for lack of working perls.
533
534 4.73 Thu May 29 2008
535 - fix a bug in Coro::EV which would cause it to block despite
536 there being runnable coroutines.
537 - sprinkle "no warnings" freely over everything, also suppress
538 warnings for some other modules.
539 - fix typo in WSAEWOULDBLOCK.
540
541 4.72 Sun May 25 05:14:36 CEST 2008
542 - tweak META.yaml a bit, unfortunately, there is no documented way
543 to have optional dependencies with CPAN. doh :(
544 - avoid running some tests on windows because they would fail due to
545 perl bug (broken fork, broken pipes...).
546 - work around perl on windows bugs where perl returns undocumented
547 error codes for sysread, syswrite etc. by taking advantage
548 of AnyEvent's workaround for that problem.
549 - use AnyEvent::Util::fh_nonblocking in Coro::Handle to work around
550 a common perl implementation bug on windows.
551 - use unix domain sockets in testsuite to work around
552 common perl implementation bugs on widows (they are emulated by
553 tcp sockets on windows. ugh.)
554
555 4.71 Sat May 24 20:01:27 CEST 2008
556 - fix a bug in Coro::AnyEvent ("Usage: Coro::AnyEvent::_schedule()").
557 - take advantage of async name resolution of AnyEvent::Util.
558 - work around brutal inet_aton override in Coro::LWP.
559 - take advantage of the readyhook in Coro::EV, for smoother
560 scheduling.
561
562 4.7 Sun May 11 00:32:19 CEST 2008
563 - completely reworked the Coro manpage.
564 - added Coro::AnyEvent, generic event loop integration.
565 - implement cancel, ready and kill commands in Coro::Debug.
566 - document find_coro in Coro::Debug.
567 - incompatible change: rename has_stack to has_cctx.
568 - Coro::AIO and Coro::BDB no longer force event model detection,
569 use AnyEvent::AIO and AnyEvent::BDB.
570
571 4.6 Sat Apr 26 10:05:14 CEST 2008
572 - INCOMPATIBLE CHANGE: sub/code attributes are no longer supported
573 by the Coro module. It was a mistake to have it in the first place.
574 - (experimental) support for activestate perl 5.10 (method "w").
575 - (experimental) support for strawberry perl (method "a").
576 - coro_sigelem_set did not return a value although it had to,
577 actual impact unknown.
578
579 4.51 Mon Apr 14 13:28:27 CEST 2008
580 - make it compile again on 5.8.
581
582 4.50 Thu Apr 10 09:43:17 CEST 2008
583 - I did it twice! (see 4.49).
584
585 4.49 Sun Apr 6 21:23:31 CEST 2008
586 - grr, instead of compiling the recent changes on 5.10 only they
587 were compiled on 5.8 only.
588
589 4.48 Sun Apr 6 20:36:46 CEST 2008
590 - allow coroutine switches during eval's under 5.10.x, as it seems
591 the parser is a per-interpreter option now, so this is safe
592 (this might fix the odd crashing bug).
593 - drop support for 5.9.x versions: they are dead, jim.
594
595 4.47 Sun Apr 6 00:37:52 CEST 2008
596 - force cctx allocation on API calls: we know we need to force one
597 and gcc actually manages to confuse our heuristic nowadays,
598 leading to crashes and worse.
599 - document force_cctx.
600
601 4.46 Fri Apr 4 22:05:43 CEST 2008
602 - upgrade libcoro, resulting in pthread-backend (which was only created
603 to fulfill the rules of the programming languages shootout).
604
605 4.45 Thu Mar 13 11:55:36 CET 2008
606 - fix a file leaking bug in eg/mhyttpd that would allow
607 downloading of any file (reported by oesi).
608 - fix deadlock bug in Coro::Channel (reported by Richard Hundt)
609 (also add testcase).
610 - support Broadcast option in Coro::Socket (patch by Richard Hundt,
611 apparently having loads of fun with that).
612 - implement $state->swap_defsv, ->swap_defav and document ->throw.
613
614 4.4 Wed Feb 13 16:44:29 CET 2008
615 - only assume .cfi pseudo isns exist on GNU/Linux.
616 - add get_timed method to Coro::Channels.
617 - fixed Coro::Debug manpage.
618 - perl 5.11 compatbility improvement based on patch by
619 Andreas König.
620
621 4.37 Sun Jan 20 11:25:23 CET 2008
622 - soften the check for an active parser for perl 5.10 (avoiding
623 "Coro::State::transfer called while parsing" in many cases).
624
625 4.36 Sun Jan 13 10:53:56 CET 2008
626 - reset diehook when terminating from an async_pool as to not
627 trigger any __DIE__ handlers.
628
629 4.35 Sun Jan 13 04:14:13 CET 2008
630 - "bt" debug command now displays any exceptions
631 from longmess and also skips the innermost
632 stackframes, giving more useufl output.
633 - allow backtraces in coroutines blocked in Coro::EV,
634 at a <1% speed hit.
635 - handle localising of $SIG{__WARN__} and __DIE__
636 properly (with a proper amount of dirty hacking).
637
638 4.34 Sat Dec 22 17:49:53 CET 2007
639 - upgrade to EV version 2.0 API.
640
641 4.33 Mon Dec 17 08:36:12 CET 2007
642 - make Coro::AIO etc. loadable in the absence of EV.
643
644 4.32 Mon Dec 17 07:46:02 CET 2007
645 - majorly improved Coro::Handle's performance with Coro::EV.
646 - implemented the until now mythical Coro::BDB module.
647 - specialcase EV in Coro::AIO and Coro::BDB for extra speed.
648
649 4.31 Wed Dec 5 12:32:39 CET 2007
650 - remove warn statement form Coro::Util.
651
652 4.3 Tue Dec 4 20:33:14 CET 2007
653 - calls to the idle function could cause stack corruption
654 when the stack changed.
655 - do no longer rely on the presence of EV::DNS if EV is used
656 (because it is gone), but instead take avdantage of EV::ADNS
657 if available.
658 - add ($) prototypes to all functions in Coro::Storable.
659 - use a conventional (safer) idle callback in Coro::EV.
660 - do accept quickly in Coro::Debug to avoid endless loops.
661
662 4.22 Fri Nov 30 16:04:04 CET 2007
663 - really use optimised versions for Event and EV in Coro::Util
664 and Coro::Handle.
665
666 4.21 Sun Nov 25 10:48:59 CET 2007
667 - fix a spurious memory read.
668 - Coro::EV no longer keeps the eventloop "alive".
669
670 4.2 Fri Nov 9 20:47:05 CET 2007
671 - enable/disable tracing from a new coroutine, not a pooled one.
672 - fix a memleak in Coro::Event.
673 - removed killall call from fork_eval.
674 - made sure store_fd is already loaded so that fork_eval
675 does not have to parse autoload in each subprocess.
676 - only use assembly method if -O switch is in $Config{optimize}.
677 - add (optional) Coro::EV module, so far the best event loop module
678 directly supported by Coro.
679 - if the event model is EV, use EV::DNS to resolve
680 stuff in Coro::Util.
681 - don't get confused by multiple event notifications in Coro::Handle.
682 - initial support for EV (libevent interface).
683 - require Event and EV using configure_requires, to force their existance.
684
685 4.13 Wed Oct 24 07:26:45 CEST 2007
686 - add Coro::Storable::blocking_thaw.
687 - use a vastly more complicated technique to localise
688 $SIG{__WARN/DIE__} that also works on perls <= 5.8.8.
689 - use a coroutine for the idle callback Coro::Event,
690 instead of running Event in the current coroutine context.
691 This also catches recursive invocations.
692 - actually report fork errors in gethostbyname and inet_aton.
693
694 4.11 Thu Oct 11 02:40:24 CEST 2007
695 - port to threaded perls.
696
697 4.1 Thu Oct 11 02:38:16 CEST 2007
698 - incompatible change: $SIG{__DIE__} and $SIG{__WARN__} will now
699 be local to each coro (see Coro::State).
700 - incompatible change: for very deep reasons, cede and cede_notself
701 cannot return anything, so nothing will be returned.
702 - possibly bring back 5.10 compatibility (untested).
703 - work around stupid (and wrong) warning on 5.10 :(.
704 - overlay the saved state over the context stack. This saves
705 a few hundred bytes per coroutine on average and also
706 speeds up context switching a bit.
707 - further tune default stack sizes.
708 - (more) correctly calculate stack usage in coro_rss.
709 - Coro::Storable::blocking_* did not properly lock
710 resulting in races between coroutines.
711 - added Coro::Storable::guard.
712 - stopping to trace a coroutine could destroy the cctx of
713 an unrelated coroutine.
714 - explain the relationship between Perl and C coroutines in
715 more detail in Coro::State.
716 - Coro::Util::inet_aton did not short-circuit dotted quad forms,
717 causing a fork per resolve. This also affected Coro::Socket.
718 - switch to a separate stack in $coro->call/eval to avoid
719 invalidating pointers.
720
721 4.03 Sat Oct 6 21:24:00 CEST 2007
722 - added Coro::throw method.
723 - minor code cleanups.
724
725 4.02 Sat Oct 6 02:36:47 CEST 2007
726 - fix a very minor per-coroutine memleak (a single codereference).
727 - fixed a bug where the currently in-use c context would be freed
728 prematurely (can happen only when programs change the stacksize
729 or use tracing).
730 - tracing can no longer keep a coro alive after it terminated.
731 - do static branch prediction in the common path for gcc. gives
732 about 2-5% speed improvement here.
733
734 4.01 Fri Oct 5 22:10:49 CEST 2007
735 - instead of recreating *a* standard output handle we simply
736 use STDOUT, which is faster and hopefully more robust.
737
738 4.0 Fri Oct 5 12:56:00 CEST 2007
739 - incompatibly changed Coro::Storable::freeze.
740 - major new feature: added Coro::Debug, for interactive coroutine
741 debugging, tracing and much more.
742 - major bug fix: unbelievable, but true: $_, $/ and many other
743 "saved" variables actually weren't being saved. This has been fixed,
744 of course, while increasing performance while losing all the save
745 flags.
746 - save flags are gone, and all the api functions dealing with them.
747 - added Coro::Semaphore::adjust.
748 - added Coro::Util::fork_eval.
749 - added Coro::Storable::{nfreeze,blocking_{freeze,nfreeze}}.
750 - added Coro::killall.
751 - reduce initial stack sizes to allow for "micro-coroutines".
752 - better async_pool resource management, moved parts of async_pool
753 handling to XS (major speed improvement).
754 - actually croak before modifying important data structures.
755 - refuse to transfer while compiling.
756 - possibly support eval EXPR better now.
757 - enable assembly per default on linux+bsd x86+amd64.
758 - all internal members were renamed _something for easier subclassing.
759 - many minor tweaks.
760
761 3.63 Wed May 16 14:10:06 CEST 2007
762 - implement handcoded assembly for x86/amd64 SVR ABI.
763
764 3.62 Fri Apr 27 21:36:06 CEST 2007
765 - upgrade libcoro (which might set unwind info correctly).
766 - change default on linux to setjmp/longjmp.
767
768 3.61 Thu Apr 19 12:36:18 CEST 2007
769 - Coro::Storable caused an endless loop when thawing invalid
770 pst's sometimes.
771 - use a Semaphore in Coro::Storable, as Storable doesn't
772 seem to be reentrant (although it is documented to
773 be threadsafe...).
774 - fix Coro::Signal to bring back the original unreliable
775 but stateful semantics.
776 - fixed a lot of typos in Coro.pm (patch submitted by David
777 Steinbrunner, which applied flawlessly).
778
779 3.6 Sat Apr 14 17:13:31 CEST 2007
780 - added some bugfixes to get eg/myhttpd working again.
781 - added Coro::Storable for often-cede'ing freeze/thaw.
782 - try to do a clean exit when a coroutine calls exit
783 (EXPERIMENTAL).
784 - got rid of indirect call through _coro_init.
785 - updated the partly antique examples in eg/ to
786 work again and be a bit less magic, too.
787 - fixed Coro::Signal semantics to work as documented again.
788
789 3.55 Sun Mar 25 01:20:47 CET 2007
790 - add SAVE_DEFFH to save the default filehandle and enable
791 it by default.
792 - finally move socket-operations from Coro::Socket to Coro::Handle
793 to be able to unblock foreign sockets.
794 - add Coro::State::save_also and guarded_save.
795 - add count accessor to Coro::Semaphore.
796 - add Coro::State::cctx_stacksize.
797 - just for the fun of it, do not rely on implicit context,
798 which can dramatically improve performance, but people
799 using windows-process-emulation perls do not care much about
800 performance.
801
802 3.51 Sun Mar 4 14:18:04 CET 2007
803 - fixed a problem when you weakened references to Coro::State's
804 (patch by Yuval Kogman).
805
806 3.501 Wed Feb 28 12:44:07 CET 2007
807 - rename some global symbols as macosx from hell redefines
808 them without asking.
809
810 3.5 Tue Feb 13 20:22:53 CET 2007
811 - do AnyEvent model detection earlier, avoiding problems
812 caused by first using AnyEvent and later Coro::Event.
813 - implement and document Coro::Event event objects.
814 - fix a potential problem in Coro::Event causing crashes.
815 - initialise PL_comppad when creating a new coroutine,
816 avoids crashes on early coro destruction.
817
818 3.41 Mon Jan 22 19:19:49 CET 2007
819 - readline on Coro::Handle did not support undefined $/,
820 nor did it deliver partial lines on EOF or error.
821 - implement malloc fallback for stack allocation because
822 stupid broken idiotic OSX has a stupid broken
823 idiotic fits-the-whole-os mmap "implementation" and
824 my dick feels longer if Coro is portable even to
825 obsolete platforms.
826
827 3.4 Fri Jan 19 21:52:54 CET 2007
828 - remove t/09_timer.t, as it isn't really testing much
829 but was rather flaky in practise.
830 - async_pool coro would keep arguments and callback alive until
831 it was reused.
832 - cancellation of a coroutine could cause spurious idle calls
833 in cede_notself.
834
835 3.3 Sat Jan 6 03:45:00 CET 2007
836 - implement $coro->on_destroy.
837 - Coro::Event blocking semantics have been changed,
838 documented and - hopefully - improved.
839 - fix nice adding, not subtracting, from priority.
840 - fix ->prio and api_is_ready (patch by Mark Hinds).
841 - fixed an assert ("... == prev__cctx->idle_te")
842 that could errronously trigger.
843 - fix various large and small memleaks.
844 - use a (hopefully) more stable cancel implementation
845 that immediately frees the coroutine data.
846 - cede/cede_notself return a status now.
847 - added Coro::guard function.
848 - added a global coroutine pool for jobs (on my machine,
849 I can create and execute 48k simple coros/s with async,
850 and 128k coros with async_pool).
851 - Coro::AIO now uses the coroutine priority as io priority.
852
853 3.2 Fri Dec 22 05:07:09 CET 2006
854 - improve portability to slightly older perls.
855 - use cleaner coroutine destruction.
856 - simplify configuration for users.
857 - optionally (unrecommended) prefer perl functions over
858 their coro replacements.
859
860 3.11 Tue Dec 5 13:11:24 CET 2006
861 - fixed some bogus assert's, but as perl.h disables assert even
862 without NDEBUG (thank you very much), not too many people should
863 notice (that did include myself). Andreas König noticed, though :)
864 - do not save/restore PL_sortcxix on >= 5.9.x, it doesn't seem to have
865 it. Also noticed by Andreas König :)
866 - save/restore tainted status.
867 - verified to pass the testsuite on my 5.9.5.
868
869 3.1 Mon Dec 4 23:03:40 CET 2006
870 - INCOMPATIBLE CHANGE: $/ is now per-coroutine (but slow).
871 - incompatible change: transfer flags are now per-state.
872 - give a better error message on deadlock.
873 - document Coro::nready.
874 - enhanced testsuite.
875
876 3.01 Sun Dec 3 23:47:42 CET 2006
877 - forgot to include Coro::Timer.
878
879 3.0 Sun Dec 3 22:57:25 CET 2006
880 - the "FINALLY COMPLETELY STABLE" release (coming soon:
881 the "FAMOUS LAST WORDS" release).
882 - implement a new stack sharing algorithm, which uses a stack
883 pool (size currently hardcoded).
884 - make stack sharing mandatory (it no longer uses a heuristic).
885 - eval/die no longer cause weird problems under heavy use.
886 - Coro::Event could cause livelocks if it was used but
887 no Coro::Event watchers were used.
888 - Coro::Event now uses asynccheck as prepare does not
889 check for changed watchers.
890 - Coro::Event allows multiple waiting coros and will wake up one
891 per event.
892 - Coro::Event should be cleaner and more efficient now.
893 - new utility function Coro::unblock_sub.
894 - document the sad fact that Event is no longer reentrant.
895 - putting a coroutine into the ready queue twice could under
896 some circumstances lead to stack corruption.
897 - minor incompatibility: subclassing Coro::State is supported
898 directly now without going through a _coro_state member.
899 - state/coro switching is much faster now.
900 - very minor optimisations and code/documentation cleanup.
901 - avoid problems due to compiler inlining.
902 - removed timers from Coro::Timer -> use AnyEvent instead.
903 - replaced Coro::idle coroutine by (cleaner) idle handler.
904 - updated to newest libcoro.
905 - implement enhanced support for valgrind.
906 - implement is_ready and return value for ready.
907 - removed Coro::Cont, it was a misnomer (it's generators),
908 and it was somewhat annoying to get it right. Will come back
909 if somebody asks for it :)
910 - many ->wait methods and Coro::Event could return
911 spuriously without the event having happened.
912
913 2.5 Tue Nov 7 12:22:33 CET 2006
914 - made Coro::Util, Coro::Select, Coro::Handle and Coro::Socket
915 use AnyEvent, moved them to Coro/.
916 - added Coro::LWP which contains all the uglyness required to
917 make LWP non-blocking.
918 - should work with perl 5.9.x now (Andreas König made me do it).
919 - fixed another bug in Coro::Select when the timeout was undef.
920 - reuse PL_start_env for all coros, saves some memory per coroutine.
921 - manage PL_top_env differently, hopefully to avoid panic: top_env.
922 - timeout argument was not properly used in Coro::Socket.
923 - allow limited forms of subclassing in Coro::Handle/Coro::Socket.
924 - emulate undocumented(!) functionality of IO::Socket required
925 by LWP(!!).
926 - updated eg/lwp to work with newer lwp's.
927 - remove "FATAL: uncaught exception" prefix. Coroutines that die
928 kill the whole process, just as exceptions in the main "coroutine"
929 did already.
930
931 2.1 Wed Nov 1 23:01:13 CET 2006
932 - fix a long-standing bug in Coro::Select where select with
933 zero timeout would instead change the current default filehandle.
934 - use a simpler and hopefully more robust way to clone padlists
935 (uses less memory and a perl function instead of our own).
936 - coro can now create a stack guard area on many architectures.
937 - Coro::AIO properly reexports additional functions from IO::AIO.
938 - updated libcoro with a workaround for OS X,
939 pach and testing by Michael Schwern.
940
941 2.0 Tue Oct 24 05:47:17 CEST 2006
942 - support additional aio requests in Coro::AIO.
943
944 1.9 (never properly released due to a glitch)
945
946 1.8 Thu Feb 2 00:59:06 CET 2006
947 - applied suggested patch by SAMV to avoid problems during stupid
948 mark & sweep gc run.
949 - applied patch by Scott Walters for 5.9.3 compatibility.
950
951 1.7 Tue Dec 27 01:41:58 CET 2005
952 - added Coro::AIO, a thin wrapper around IO::AIO.
953 - improved Makefile.PL explanations.
954
955 1.6 Mon Dec 12 21:30:05 CET 2005
956 - additionally save PM_curpm might fix as-of-yet
957 unseen problems with regex matches being attributed to the
958 wrong package.
959 - add t/10_bugs.t, which currently checks against imho broken
960 perls that use 0.26MB of stack space per Perl_magic_get
961 invocation (newer linuxes) as opposed to the 0.0002MB perl
962 normally uses.
963 - make stacksize configurable for the ultimate debian experience.
964
965 1.51 Mon Dec 12 18:48:36 CET 2005
966 - remove debugging warn() accidentally left in Coro::Select.
967
968 1.5 Tue Nov 29 13:32:44 CET 2005
969 - use Coro::Event inside Coro::Select to avoid spurious deadlocks.
970 - fix Coro::Select 'select'.
971 - strict'ify some modules.
972 - libcoro errornously restored the SIGUSR2 handler to SIGUSR1.
973 - use XSLoader in selected modules.
974 - remove some 5.6 compatibility cruft.
975
976 1.4 Tue Sep 6 00:11:05 CEST 2005
977 - libcoro did not take into account the trampoline on amd64,
978 when 'l'inux method was used.
979
980 1.31 Tue Aug 30 23:31:33 CEST 2005
981 - some portability fixes/workarounds.
982
983 1.3 Sat Aug 20 03:08:56 CEST 2005
984 - no code changes. module seems to work fine.
985 - improve Coro::State docs, remove reference to nonexisting
986 Coro::State::flush method.
987 - no longer autodetect windows, present it as an option instead.
988
989 1.2 Mon May 16 02:00:55 CEST 2005
990 - included libcoro.c earlier in State.xs, might improve portability.
991 - use a faster and possibly more stable Coro::Cont implementation.
992 - accept x86_64 in addition to amd64 for optimized linux-amd64 support.
993 - fix bugs Coro::SemaphoreSet that could cause locks never to be freed.
994 - fix bugs in CoroAPI.h and document it in Coro::MakeMaker.
995
996 1.11 Thu Mar 3 18:00:52 CET 2005
997 - change of contact address.
998
999 1.1 Tue Feb 22 20:51:16 CET 2005
1000 - support [l]inux method on amd64.
1001 - allow some context switches while compiling/eval'ing.
1002
1003 1.0 Tue Aug 10 03:47:27 CEST 2004
1004 - Event 0.89 fixes the bug again.
1005 - fix the remaining(??) memleak. I didn't free the cache-AV for
1006 padlists, so one AV per created anon-sub leaked.
1007
1008 0.97 Fri May 14 15:23:32 CEST 2004
1009 - remove stupid gcc extension that was an experiment only anyway.
1010
1011 0.96 Thu May 13 18:09:29 CEST 2004
1012 - tell people if they have yet another broken Event variant.
1013 - made Coro to work _better_ with threads (it seems that, as long
1014 as you use Coros in one thread only everything will work fine.
1015 Looking at the perl source the problem might be missing locking
1016 between Coro and some perl internal routines. That's not easily
1017 fixable).
1018 - (hopefully) got rid of the static xnext inside transfer.
1019
1020 0.95 Sun Feb 15 03:20:28 CET 2004
1021 - removed Coro::State::flush, implemented a new and much
1022 faster caching mechanism by attaching magic to the CV.
1023 - a memleak with (real) closures remains.
1024 - dramatically reduced size of stacks to 65536 (linux/x86), after
1025 observing that even myhttpd never had more than a kilobyte
1026 of stack in use (stack is still only physically allocated in
1027 pagesize increments for systems suporting mmap).
1028 - fix for a "cygwin-64" architecture added, proposed by Stefan Traby.
1029 - fix for threaded perl.
1030
1031 0.9 Sun Nov 23 04:32:39 CET 2003
1032 - fix a few very hard to track down but (of course) stupid
1033 memory leaks.
1034 - stringify version number to avoid locale problems :(.
1035 - perl 5.9.0 does not have SvPADBUSY, reported by Scott Walters,
1036 hopefully ignoring it "just works".
1037
1038 0.8 Wed Nov 5 19:38:40 CET 2003
1039 - port to perl5.005_03, but only 5.8.x+ is supported!
1040 - honor LocalAddr even without LocalPort in Coro::Socket.
1041
1042 0.7 Tue May 27 03:12:38 CEST 2003
1043 - the version jump indicates some level of testing, not gobs
1044 of new features.
1045 - uh... I found the prompt function in ExtUtils::MakeMaker.
1046 Highly correct stuff, that is...
1047 - fixed(?) a bug with die's in coroutines causing "panic: top_env".
1048 the fix is not well understood by the author, so beware :(.
1049
1050 0.652 Thu May 8 02:54:46 CEST 2003
1051 - Applied patch by Slaven Rezic to set default to "s" on FreeBSD,
1052 cause version 4 doesn't have ucontext.
1053 - Benjamin Reed reported that setjmp works fine on darwin,
1054 so preselect it.
1055
1056 0.651 Sat Mar 29 15:00:23 CET 2003
1057 - fix a bug in Coro::Handle where some bytes could get lost
1058 on reading, reported by jason@nichego.net.
1059
1060 0.65 Sun Mar 23 00:08:26 CET 2003
1061 - added a README.linux-glibc.
1062 - new module Coro::Select.
1063 - also save/restore PL_comppad, fixes segfaults with 5.8.1.
1064 (I never claimed that I know what I am doing ;).
1065 - set default to setjmp/longjmp on non-x86-linux.
1066
1067 0.6 Thu Nov 21 11:09:06 CET 2002
1068 - made lazy context switching the default.
1069 - fixed the bug where SemaphoreSet::guard
1070 was the same as timed_guard and timed_guard was missing.
1071 - fix a memleak in Coro::Timer.
1072 - reclassified context sharing as not experimental.
1073
1074 0.534 Sun Apr 14 03:05:12 CEST 2002
1075 - fixed backspace => tab in header parsing (myhttpd).
1076 - added eg/readline, for elmex.
1077 - Coro::Event now calls ->start on first call instead of
1078 ->again, which created wrong timeouts for the first ->next.
1079 - fixed a bug where Coro::Socket returned a closed socket instead
1080 of nothing on ECONNREFUSED and similar error conditions.
1081
1082 0.533 Mon Feb 18 18:49:40 CET 2002
1083 - adapt to finally working Event-0.85.
1084 - fixed buggy event initialization (trapped under windows, but how
1085 could it possibly have worked anywhere?)
1086
1087 0.532 Wed Jan 16 02:45:32 CET 2002
1088 - removed the reference to pp_entersub, might work on
1089 (some) win32 perls now (testsuite works under cygwin,
1090 without Event, which freezes).
1091
1092 0.531 Mon Dec 10 22:18:44 CET 2001
1093 - Fixed a bug in SemaphoreSet::timed_down.
1094
1095 0.53 Tue Nov 27 21:11:13 CET 2001
1096 - some tests for Event.
1097 - slightly different internal architecture to get refcnt into the object
1098 and not the reference. might fix a few bugs, certainly did introduce
1099 new ones.
1100 - Coro::Timer (independent of Event).
1101 - new timed_wait functions for Coro::Signal, Semaphore, SemaphoreSet.
1102
1103 0.52 Tue Nov 6 21:36:18 CET 2001
1104 - ported to cygwin (trivial).
1105 patch by Gerrit P. Haase <gp@familiehaase.de>.
1106 - small setjmp code fix by Sullivan.DanielJ@epamail.epa.gov.
1107
1108 0.51 Thu Nov 1 20:39:01 CET 2001
1109 - terminate/cancel now work properly (otherwise termination
1110 could cause a "next coroutine is not and contains not..."
1111 error.
1112 - added Coro::Socket::shutdown.
1113 - Coro::Event::loop no is the same as Event::loop.
1114 - implemented terminate with args + join.
1115
1116 0.5 Fri Sep 28 16:15:35 CEST 2001
1117 - fixed "print" on a Coro::Handle. The print method worked.
1118 - small tweaks (seem to reduce memory consumption a lot)
1119 in various modules.
1120 - splendid use of "no warnings" scattered throughout.
1121 - added Coro::Handle::rbuf, fixed Coro::Handle::fh.
1122
1123 0.49 Sun Sep 16 02:42:45 CEST 2001
1124 - changed some method calls to function calls for speed
1125 inside Coro::Handle.
1126 - make Coro::Handle use an array instead of a hash for
1127 speed reasons.
1128 - IRIX mystery solved: it's SGI's NT, after all: "standard, huh?".
1129 sjlj and ucontext should now work.
1130 - IRIX-specific port for libcoro.
1131 - swapped order of accept results to match IO::Socket.
1132 - changed getsock/peername to sock/peername to match IO::Socket.
1133 - fixed a bug that caused segfault when returning to main task
1134 under some circumstances.
1135 - other bugfixes.
1136
1137 0.45 Sun Sep 2 02:54:01 CEST 2001
1138 - new method Coro::Handle::timeout.
1139 - corrected speling of set...name to get...name in Coro::Socket.
1140 - Coro::Socket::accept now returns a Coro::Socket, not a Coro::Handle.
1141 - Coro::Handle now supports fileno and FILENO.
1142 - added eg/myhttpd, a web server.
1143 - Coro::Socket now accepts numeric ports as well as "port(number)"
1144 syntax.
1145 - moved some scheduling primitives into xs code. more to come.
1146 - new simple priority system.
1147 - implemented Coro::Event using XS, almost four times faster!
1148 - small memory corruption problem fixed (boy that was difficult).
1149
1150 0.13 Wed Aug 8 16:53:07 CEST 2001
1151 - new method Coro::Handle::readline.
1152 - added eg/lwp and Coro::Handle::autoflush, to make LWP work.
1153 (see eg/lwp on how to make LWP non-blocking).
1154 - renamed Coro::Socket::new_inet to new.
1155 - added Coro::Util, some utility functions.
1156 - die/eval now works better (still get annoying (but true ;)
1157 "Callback called exit" messages).
1158 - Coro::Handle now supports timeout, and Coro::Socket Timeout.
1159 - much better Coro::Socket support.
1160
1161 0.12 Fri Jul 27 04:19:01 CEST 2001
1162 - do not use mmap if neither MAP_ANON nor MAP_ANONYMOUS is defined.
1163 - very experimental stack sharing algorithm. not 100% safe but
1164 should work well in practise ;)
1165 - added Coro::RWLock.
1166 - Coro::Specific now works.
1167
1168 0.11 Tue Jul 24 22:49:21 CEST 2001
1169 - added specialized hack for newer and older linux versions (fast).
1170 - renamed Coro::Event::IO to Coro::Handle.
1171 - new module Coro::Socket.
1172
1173 0.1 Tue Jul 24 01:47:53 CEST 2001
1174 - release candidate 3. A rather new internal structure :(
1175 - the great renaming: Damian Conway gave me a suitable replacement
1176 for yield.
1177 - added Coro/libcoro, a portable coroutine implementation for C
1178 (not even perl-dependent!!!), which can be used by Coro::State.
1179 - renamed SAVE_DEFSV (on xs-level only) to avoid symbol clash in
1180 perl-5.6.
1181 - new function Coro::Event::idle.
1182 - the idle process is now overriden
1183 by default in Coro::Event.
1184 - Coro::Channel now enforces the size.
1185 - canceling events now works in all cases.
1186 - Coro state now include $_ and $@.
1187 - yet another bug workaround that I do not really understand :(
1188 - new module Coro::Event::IO (very undocumented) to do
1189 non-blocking i/o.
1190 - performance tuning.
1191
1192 0.08 Thu Jul 19 06:13:25 CEST 2001
1193 - release candidate 2 (new functionality).
1194 - add Coro::State::flush function.
1195 - transfer now supports a flags value (mostly for speed ATM).
1196 - might compile and work in the presence of threads now.
1197 - continuations have a saner syntax.
1198 - no more memleaks.
1199
1200 0.07 Tue Jul 17 17:40:18 CEST 2001
1201 - release candidate 1 ;)
1202 - slightly nicer code.
1203 - fixed a scheduling bug in Coro::Event.
1204
1205 0.06 Tue Jul 17 04:23:24 CEST 2001
1206 - ok, I found the showstopper - the same sub must not be
1207 re-entered in two different coroutines, otherwise => crash. I
1208 see no easy solution to this problem, except by walking the call
1209 chaing and saving/restoring the cv's, which is what I do now.
1210 - memory leaks still latent, especially at thread termination.
1211 - Coro::Event now works (haha).
1212
1213 0.05 Sun Jul 15 17:32:20 CEST 2001
1214 - fixed a few issues in Coro::Event.
1215 - I forgot to include Coro::Event in 0.04 :(:(:(
1216
1217 0.04 Sun Jul 15 05:24:59 CEST 2001
1218 - @_ is now properly localized.
1219 - Coro::State is now easier subclassable.
1220 - Coro::Cont now coroutine-aware.
1221 - Coro::Specific is a low-overhead module to create
1222 coroutine-specific vars.
1223 - Coro::Event provides a simple interface to Event.
1224
1225 0.03 Fri Jul 13 14:51:52 CEST 2001
1226 - transfer() now implemented in XS (beware).
1227 - new module Coro::Cont for really faked continuations.
1228 - big internal architecture changes: Coro::State is now
1229 really low-level and can thus be used to implement other
1230 interesting things, While "Coro::" implements a process-like
1231 model. Still crude and subject to change.
1232 - $_ and $@ are no longer being localized.
1233
1234 0.02 Tue Jul 10 01:38:17 CEST 2001
1235 - implemented "async" attribute.
1236 - $_ and $@ are now localized.
1237 - added Coro::Channel.
1238 - more testcases, still no docs.
1239
1240 0.01 Tue Jul 3 02:18:41 CEST 2001
1241 - original version; copied from Convert::Scalar.
1242