ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/Coro/Changes
(Generate patch)

Comparing Coro/Changes (file contents):
Revision 1.498 by root, Tue Jun 23 00:56:14 2009 UTC vs.
Revision 1.618 by root, Sat Nov 2 19:48:50 2013 UTC

1Revision history for Perl extension Coro. 1Revision history for Perl extension Coro.
2 2
3TODO: should explore PerlIO::coroaio (perl leaks like hell). 3TODO: should explore PerlIO::coroaio (perl leaks like hell).
4TODO: 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)).
7TODO: unready_all
8TODO: myhttpd header parsing
9TODO: channel->maxsize(newsize)? 4TODO: channel->maxsize(newsize)?
5TODO: __GCC_HAVE_DWARF2_CFI_ASM
6TODO: swap_sv, maybe add scope_swap_sv, andallow to unswap by repeat?
7TODO: swap_sv not undone in pool
10 8
9 - use a new algorithm to derive padlists for perl 5.18. The old
10 one could lead to 0-pointer accesses inside perl.
11
126.31 Thu May 9 07:39:48 CEST 2013
13 - Coro::AIO requests would crash if the thread was ready'd
14 while the request was ongoing.
15
166.29 Wed May 8 02:55:18 CEST 2013
17 - when an on_destroy handler destructs the coro currently being
18 destructed a perl scalar could be accessed after being freed,
19 likely causing a crash.
20
216.28 Wed Mar 6 06:58:02 CET 2013
22 - clean remnants of existing __DIE__ and __WARN__ handlers so
23 they lose their magic and will not cause segfaults later
24 (testcase by Andrey Sagulin).
25 - improved Coro::State documentation a bit.
26 - Coro::Debug::command now flushes the output.
27 - add hack detection code for x32 abi, because the braindead slugs
28 who designed that made it look exactly like x86_64 without
29 providing proper compile time symbols to test for it. as a result,
30 this detection cannot work reliably.
31 - valgrind stack registering was broken.
32 - do not rely on Time::HiRes anymore in Coro::Debug.
33
346.23 Fri Dec 7 23:36:37 CET 2012
35 - use experimental fiber implementation on native windows
36 perls.
37 - use sizeof (void *) as multiplication factor for stack sizes,
38 to accomodate the totally braindamaged microsoft 64 bit "os".
39 - changed verifier host from win2k-ap510-32 to win7-sp516-32/64.
40 activeperl 5.16 crashes when PerlIO_define_layer is called due
41 to some bug in the perl dll, strawberry perl at least passes
42 the testsuite.
43 - implement Coro::Handle->peeraddr/host/port, for slightly
44 improved compatibility with LWP.
45 - implement 5.17 compatibility by almost blindly applying a
46 good-looking patch by Father Chrysostomos.
47 - move stack management functions into libcoro 3.
48 - libcoro version 3 "released".
49 - support magic values as timed_io_once args.
50 - recommend AnyEvent 7+ or EV 4+, also require EV
51 version 4 or newer for Coro::EV.
52
536.10 Tue Oct 9 01:14:27 CEST 2012
54 - updated ecb.h, it had a typo that caused it to not compile on many
55 big endian systems (reported by many people).
56 - disable memory fences in ecb.h to improve portability.
57
586.09 Sat Oct 6 23:25:02 CEST 2012
59 - Coro::EV I/O watchers were not interruptible by exceptions
60 (Coro::State::throw) (testcase by sten).
61 - ->throw now puts threads into the ready queue, as this seems to
62 be expected by existing code, and code that doesn't cope with spurious
63 wakeups needs fixing anyway.
64 - use fd -1 in mmap.
65 - cast I32 to int in error message printf.
66 - warn about broken so-called "hardened" kernels.
67
686.08 Fri Apr 13 12:05:47 CEST 2012
69 - be more aggressive about exiting like perl does - formerly,
70 exiting from the non-main thread would not execute END blocks.
71
726.07 Fri Nov 11 21:21:48 CET 2011
73 - work around a bug in PerlIO (setting $SIG{__WARN__} to a PVCV).
74 - update ecb.h.
75
766.06 Mon Aug 8 23:59:48 CEST 2011
77 - cygwin unfortunately patches the stack at runtime, so we use the pthreads
78 backend, which is an order of magnitude slower. unfortunately, cygwins
79 pthread implementation isn't very complete either, so allocate the stack
80 twice just to be sure.
81 (note: cygwin also enables mymalloc, which is NOT THREADSAFE ON WINDOWS,
82 in its ithreaded perl - best recompile cygwin and use the 'w'indows
83 backend for much better performance. also disable ithreads for
84 even better performance...).
85
866.05 Thu Aug 4 21:36:36 CEST 2011
87 - blush, condvar values would not be propagated from send to recv anymore
88 (reported by Chip Salzenberg).
89 - use exponential increase for the readline buffer length in
90 Coro::Handle. also reduce initial allocation to 1020 from 4096 bytes.
91
926.04 Wed Aug 3 17:15:45 CEST 2011
93 - use even more efficient and more compatible condvars for
94 compatibility to AnyEvent 6.x :)
95 - more inconsequential ecb.h updates.
96
976.03 Wed Aug 3 11:41:30 CEST 2011
98 - change how Coro patches AnyEvent condvars for compatibility to
99 AnyEvent 6.x.
100 - update ecb.h, to no longer include <pthread.h> in case WinNT.h
101 hasn't been included.
102
1036.02 Wed Jul 13 04:35:19 CEST 2011
104 - "improve portability to Gentoo" - gentoo manages to put perl variables
105 in memory areas that are farther than 2gb apart, which the jit couldn't
106 handle and barfed. now it's just a bit slower on gentoo and similar
107 systems.
108
1096.01 Sun Jul 3 12:31:14 CEST 2011
110 - workarounds are good, but the test for whether pthreads are used
111 was not good. thisone should be better.
112 - check differently whether gcc generates cfi instructions itself.
113
1146.0 Wed Jun 29 19:43:35 CEST 2011
115 - INCOMPATIBLE CHANGE: unreferenced coro objects will now be
116 destroyed and cleaned up automatically (e.g. async { schedule }).
117 - implement a JIT compiler for part of the thread switch code,
118 which gives a 50% speed improvement on threaded perls, and
119 about 4% on non-thraeded perls (so threaded perls now finally
120 reach about half the speed of non-threaded perls).
121 - slightly modernise Coro::Intro, add section about rouse functions.
122 - avoid DEFSV and ERRSV, giving another 10% improvement
123 in thread switching.
124 - Coro::State->is_destroyed is now called is_zombie.
125 - implement a Coro->safe_cancel method that might fail, but
126 cancels in a "safer" way if it succeeds.
127 - add preliminary support for DEBUGGING perls.
128 - get rid of two hash-accesses when initialising a new Coro - this
129 speeds up coro creation by almost a factor of two.
130 - croak when a coro that is being cancelled tries to block
131 (e.g. while executing a guard block), instead of crashing or
132 deadlocking.
133 - use a more robust and also faster method to identify Coro::State
134 objects - speeds up everything a bit.
135 - implement Coro->cancel in XS for a 20% speed improvement, and to
136 be able to implement mutual cancellation.
137 - speed up context switches by a percent or two by more efficiently
138 allocating context stack entries.
139 - implement Coro->join and Coro->on_destroy in XS for a speedup and
140 a reduction in memory use.
141 - cancelling a coro while it itself is cancelling another coro is
142 now supported and working, instead of triggering an assertion.
143 - be a bit more crash-resistant when calling (buggy) on_destroy
144 callbacks (best effort).
145 - move on_destroy into the slf_frame, to allow extension slf
146 functions to have destructors.
147 - get rid if coro refcounting - simply crash in other interpreter
148 threads by nulling the pointers on clone.
149 - simplify warn/die hook handling when loading Coro - the convoluted
150 logic seems to be no longer neccessary.
151 - use libecb instead of our own home-grown gcc hacks.
152 - document alternatives to Coro::LWP. Please use them :)
153 - work around another mindless idiotic NEEDLESS bug in openbsd/mirbsds
154 sigaltstack. Really. wine suffers from it, erlang suffers from it,
155 and it's known since at least 2006.
156
1575.372 Wed Feb 23 06:14:30 CET 2011
158 - apparently mingw doesn't provide a working gettimeofday, try to
159 work around that by relying on Time::HiRes (indirectly brought to
160 my attention by Max Maischein).
161 - fix some portability issues when Time::HiRes was used.
162
1635.371 Mon Feb 21 14:36:08 CET 2011
164 - backport to windows process emulation code again.
165
1665.37 Sat Feb 19 07:49:44 CET 2011
167 - add a big "Coro thread life cycle" section to "man Coro".
168 - try a tentative workaround against the breakage that 5.13 has
169 introduced without depreciation period. sigh.
170 - no longer use Time::HiRes if gettimeofday is available, which
171 saves quite a lot of memory.
172
1735.36 Sun Feb 13 05:33:41 CET 2011
174 - automatically load Coro::Channel, Coro::RWLock, Coro::Semaphore,
175 Coro::SemaphoreSet, Coro::Signal and Coro::Specific on first "new"
176 method call.
177 - undocument Coro::Timer::sleep and obsolete whole module.
178 - optimise Coro::Timer::timeout memory and cpu usage.
179 - slightly updated Coro::Intro for recent changes.
180 - do not initialise PL_dirty anymore.
181
1825.25 Thu Nov 11 01:08:39 CET 2010
183 - try a different approach on netbsd - netbsd 5 finally has marginally
184 working pthreads, but still broken ucontext/sigaltstack.
185 - openbsd 4.8 finally got their act together, Coro works out of the box
186 with asm, setjmp and pthreads (no change, just informational).
187
1885.24 Sat Oct 23 11:27:12 CEST 2010
189 - port to the EV 4.0 API.
190 - work around bugs in mingw32, making strawberry perl work
191 out of the box.
192 - correctly modify Coro::AIO function prototypes
193 so that they reflect the "no optional parameters" rule.
194 - "ported" libcoro to C++.
195
1965.23 Mon May 17 18:50:42 CEST 2010
197 - be more resistant to ordering changes when initialising
198 Coro::AnyEvent, Coro::EV and Coro::Event (reported by Matthias
199 Waldorf).
200 - document that perl 5.12 deliberately removed support for cloning.
201
2025.22 Wed Apr 14 03:55:35 CEST 2010
203 - correctly return udnef on errors in Coro::Handle::read/write
204 (testcase by Marc Mims).
205 - convert Coro::Util into a "perl compatibility wrapper" - the functions
206 are less useful now, but are drop-in replacements for existing
207 functions, listing better alternatives in the documentation. This also
208 fixes a bug in Coro::LWP which naively substituted Socket::inet_aton
209 with Coro::Util::inet_aton.
210 - do not override $Coro::idle unconditionally in Coro.pm, as other
211 modules could have provided their own idle coro already
212 (for exmaple, Coro::AnyEvent).
213 - fix Coro::Util::gethost* functions.
214 - Coro::Timer corretcly exports it's symbols (reported by Hideki Yamamura).
215
2165.21 Wed Dec 16 07:19:51 CET 2009
217 - automatically load Coro::AnyEvent when AnyEvent and Coro are used
218 together.
219 - add some examples on how to combine other event loops with Coro in
220 Coro::AnyEvent, and how to run it (and not to block). Seems to be
221 the most common source of confusion.
222 - try to catch people naively blocking in an event callback.
223 - work around the perl filehandle bug issue in conjunction with
224 older common::sense (as indirectly pointed out by ZSystem).
225 - clarify the "not from signal handlers" section.
226
2275.2 Sun Oct 4 14:54:24 CEST 2009
228 - Coro::Storable destroyed the prototypes of the functions it wrapped.
229 - export rouse_cb and rouse_wait by default now.
230 - fix various prototype mismatches in Coro::AnyEvent and Coro::Handle.
231 - new method $state->swap_sv.
232 - added section on "windows process emulation" to the manpage,
233 after a not-so-fruitful (nor-friendly) "discussion" with chip
234 salzenberg (discussion implies arguments, but his only arguments
235 were ad-hominems, one wonders why he started it in the first
236 place). I hope this explains it well enough for him to understand,
237 and maybe well enough for others to understand.
238 - use common::sense everywhere now.
239 - idle callbacks are no longer supported, use idle coros instead.
240 - print a thread listing when a deadlock is detected.
241
2425.17 Sat Aug 22 23:09:31 CEST 2009
243 - work around a bug in the perl debugger causing crashes
244 when running under the debugger by marking _pool_handler as nodebug.
245 - speed up Coro::async considerably.
246 - try some hacks to get netbsd to work "more often" - their broken
247 setjmp/longjmp, ucontext *and* phtreads are really hard on Coro.
248 - convert Coro to AE 5.0 API.
249
2505.162 Tue Jul 28 04:04:03 CEST 2009
251 - perl 5.8.2 is now minimum requirement.
252 - skip t/19_handle.t on broken windows perls.
253
2545.161 Wed Jul 22 04:47:38 CEST 2009
255 - Coro::AnyEvent::poll could have a different prototype when EV was
256 used as backend (analyzed by Tatsuhiko Miyagawa).
257 - Coro::AnyEvent errornously initialised the event loop when loaded,
258 not on demand.
259 - try to workaround rare */t/01_unblock.t failures.
260
2615.16 Tue Jul 21 01:44:37 CEST 2009
262 - Coro::AnyEvent failed to hook into the event loop
263 when no threads had been readied between detecting
264 the event loop and actually running it.
265 - considerably speed up Coro::Select by taking avdantage
266 of AnyEvent > 4.8 and some other optimisations.
267 - implement paragraph readline mode in Coro::Handle
268 (based on patches by Zsbán Ambrus).
269 - replace WSAEINPROGRESS by WSAEWOULDBLOCK (reported
270 and analyzed by Yasuhiro MATSUMOTO).
271 - clarified libcoro license and copyright.
272 - someone stole my EXTRA_META!!!
273 - implement Coro::Select::patch_pp_sselect and it's brother,
274 for hardcode select overriding.
275
2765.151 Mon Jul 6 05:41:57 CEST 2009
277 - backport to windows process emulation code again (patch by
278 Yasuhiro MATSUMOTO).
279 - slightly update Coro::MakeMaker.
280
2815.15 Tue Jun 30 10:28:06 CEST 2009
282 - deprecate Coro::Socket, document how to get ipv6 support via
283 AnyEvent::Socket instead.
284 - implement signal->wait ($cb) interface, similar to semaphores.
285 - work around SvOK not supporting getmagic, so we have to getmagic
286 to test for undef :/ (reported by Matthias Waldorf).
287 - load Coro::AnyEvent in all modules using AnyEvent.
288 - work around perl corrupting our internal data structures,
289 reported by Tokuhiro Matsuno.
290 - enable per-coroutine real and cpu time gathering
291 (Coro::State::enable_times).
292
2935.14 Wed Jun 24 01:37:48 CEST 2009
11 - provide explicit functions to "cede" to the event loop in Coro::AnyEvent, 294 - provide explicit functions to "cede" to the event loop in Coro::AnyEvent,
12 as this seems to have been a difficult concept (poll, sleep, idle, 295 as this seems to have been a difficult concept (poll, sleep, idle,
13 idle_upto). 296 idle_upto).
14 - add Coro::AnyEvent::readable/writable functions. 297 - add Coro::AnyEvent::readable/writable functions.
15 - clarify Coro::EV/Event/AnyEvent manpages. 298 - clarify Coro::EV/Event/AnyEvent manpages.
16 - free per-thread global scalars in the thread calling ->cancel, to 299 - free per-thread global scalars in the thread calling ->cancel, to
17 avoid crashes when $_, $@ etc., are magical but some of those 300 avoid crashes when $_, $@ etc., are magical but some of those
18 had already been freed. 301 had already been freed.
19 - "unexperimentalize" the callback interface for Coro::Semaphore. 302 - "unexperimentalise" the callback interface for Coro::Semaphore.
20 - speed up ready queue management/context switching by using a linked 303 - speed up ready queue management/context switching by using a linked
21 list instead of an array (~5-10%). 304 list instead of an array (~5-10%).
22 - implement "watch" command in Coro::Debug shells. 305 - implement "watch" command in Coro::Debug shells.
23 - for fun, implement time-slicing as an example in the manpage. 306 - for fun, implement time-slicing as an example in the manpage.
24 - if AnyEvent detects EV or Event, but we don't have Coro::EV or 307 - if AnyEvent detects EV or Event, but we don't have Coro::EV or
… …
195 - ->throw is now supported on Coro::State objects. 478 - ->throw is now supported on Coro::State objects.
196 - clean up cctx creation code a bit. 479 - clean up cctx creation code a bit.
197 - entersub is actually a UNOP, not a LOGOP (not a bugfix). 480 - entersub is actually a UNOP, not a LOGOP (not a bugfix).
198 481
1994.9 Sat Nov 8 17:45:27 CET 2008 4824.9 Sat Nov 8 17:45:27 CET 2008
200 - (libcoro) did not preserve rbp with CORO_ASM (we are getting there). 483 - (libcoro) did not preserve rbp with CORO_ASM (we are getting there).
201 - (libcoro) no longer leak threads in the experimental pthread backend, 484 - (libcoro) no longer leak threads in the experimental pthread backend,
202 also speed it up considerably. 485 also speed it up considerably.
203 - (libcoro) do not rely on makecontext passing void *'s unscathed. 486 - (libcoro) do not rely on makecontext passing void *'s unscathed.
204 - fix compiletime dependencies on libcoro in the Makefile. 487 - fix compiletime dependencies on libcoro in the Makefile.
205 - cctx_count wasn't always updated properly. 488 - cctx_count wasn't always updated properly.
… …
267 - remove debugging code related to MgPV_nolen_const, also try to 550 - remove debugging code related to MgPV_nolen_const, also try to
268 make it compile with perl 5.8.6 (yes, apple apparently loves 551 make it compile with perl 5.8.6 (yes, apple apparently loves
269 outdated software). Reported by John S. 552 outdated software). Reported by John S.
270 553
2714.744 Tue Jul 8 22:06:35 CEST 2008 5544.744 Tue Jul 8 22:06:35 CEST 2008
272 - correctly provide default DIE/WARN handlers as documented. 555 - correctly provide default DIE/WARN handlers as documented.
273 - also overwrite PL_vtbl_sigelem.svt_clear, even though current 556 - also overwrite PL_vtbl_sigelem.svt_clear, even though current
274 implementations inside perl work fine for us. 557 implementations inside perl work fine for us.
275 558
2764.743 Mon Jun 16 00:21:57 CEST 2008 5594.743 Mon Jun 16 00:21:57 CEST 2008
277 - when using Coro::EV without running EV::loop it could 560 - when using Coro::EV without running EV::loop it could
… …
299 - sprinkle "no warnings" freely over everything, also suppress 582 - sprinkle "no warnings" freely over everything, also suppress
300 warnings for some other modules. 583 warnings for some other modules.
301 - fix typo in WSAEWOULDBLOCK. 584 - fix typo in WSAEWOULDBLOCK.
302 585
3034.72 Sun May 25 05:14:36 CEST 2008 5864.72 Sun May 25 05:14:36 CEST 2008
304 - tweak META.yaml a bit, unfortunately, there is no documented way 587 - tweak META.yaml a bit, unfortunately, there is no documented way
305 to have optional dependencies with CPAN. doh :( 588 to have optional dependencies with CPAN. doh :(
306 - avoid running some tests on windows because they would fail due to 589 - avoid running some tests on windows because they would fail due to
307 perl bug (broken fork, broken pipes...). 590 perl bug (broken fork, broken pipes...).
308 - work around perl on windows bugs where perl returns undocumented 591 - work around perl on windows bugs where perl returns undocumented
309 error codes for sysread, syswrite etc. by taking advantage 592 error codes for sysread, syswrite etc. by taking advantage
… …
313 - use unix domain sockets in testsuite to work around 596 - use unix domain sockets in testsuite to work around
314 common perl implementation bugs on widows (they are emulated by 597 common perl implementation bugs on widows (they are emulated by
315 tcp sockets on windows. ugh.) 598 tcp sockets on windows. ugh.)
316 599
3174.71 Sat May 24 20:01:27 CEST 2008 6004.71 Sat May 24 20:01:27 CEST 2008
318 - fix a bug in Coro::AnyEvent ("Usage: Coro::AnyEvent::_schedule()"). 601 - fix a bug in Coro::AnyEvent ("Usage: Coro::AnyEvent::_schedule()").
319 - take advantage of async name resolution of AnyEvent::Util. 602 - take advantage of async name resolution of AnyEvent::Util.
320 - work around brutal inet_aton override in Coro::LWP. 603 - work around brutal inet_aton override in Coro::LWP.
321 - take advantage of the readyhook in Coro::EV, for smoother 604 - take advantage of the readyhook in Coro::EV, for smoother
322 scheduling. 605 scheduling.
323 606
3244.7 Sun May 11 00:32:19 CEST 2008 6074.7 Sun May 11 00:32:19 CEST 2008
325 - completely reworked the Coro manpage. 608 - completely reworked the Coro manpage.
326 - added Coro::AnyEvent, generic event loop integration. 609 - added Coro::AnyEvent, generic event loop integration.
327 - implement cancel, ready and kill commands in Coro::Debug. 610 - implement cancel, ready and kill commands in Coro::Debug.
328 - document find_coro in Coro::Debug. 611 - document find_coro in Coro::Debug.
329 - incompatible change: rename has_stack to has_cctx. 612 - incompatible change: rename has_stack to has_cctx.
330 - Coro::AIO and Coro::BDB no longer force event model detection, 613 - Coro::AIO and Coro::BDB no longer force event model detection,
331 use AnyEvent::AIO and AnyEvent::BDB. 614 use AnyEvent::AIO and AnyEvent::BDB.
… …
455 738
4564.11 Thu Oct 11 02:40:24 CEST 2007 7394.11 Thu Oct 11 02:40:24 CEST 2007
457 - port to threaded perls. 740 - port to threaded perls.
458 741
4594.1 Thu Oct 11 02:38:16 CEST 2007 7424.1 Thu Oct 11 02:38:16 CEST 2007
460 - incompatible change: $SIG{__DIE__} and $SIG{__WARN__} will now 743 - incompatible change: $SIG{__DIE__} and $SIG{__WARN__} will now
461 be local to each coro (see Coro::State). 744 be local to each coro (see Coro::State).
462 - incompatible change: for very deep reasons, cede and cede_notself 745 - incompatible change: for very deep reasons, cede and cede_notself
463 cannot return anything, so nothing will be returned. 746 cannot return anything, so nothing will be returned.
464 - possibly bring back 5.10 compatibility (untested). 747 - possibly bring back 5.10 compatibility (untested).
465 - work around stupid (and wrong) warning on 5.10 :(. 748 - work around stupid (and wrong) warning on 5.10 :(.
… …
537 but stateful semantics. 820 but stateful semantics.
538 - fixed a lot of typos in Coro.pm (patch submitted by David 821 - fixed a lot of typos in Coro.pm (patch submitted by David
539 Steinbrunner, which applied flawlessly). 822 Steinbrunner, which applied flawlessly).
540 823
5413.6 Sat Apr 14 17:13:31 CEST 2007 8243.6 Sat Apr 14 17:13:31 CEST 2007
542 - added some bugfixes to get eg/myhttpd working again. 825 - added some bugfixes to get eg/myhttpd working again.
543 - added Coro::Storable for often-cede'ing freeze/thaw. 826 - added Coro::Storable for often-cede'ing freeze/thaw.
544 - try to do a clean exit when a coroutine calls exit 827 - try to do a clean exit when a coroutine calls exit
545 (EXPERIMENTAL). 828 (EXPERIMENTAL).
546 - got rid of indirect call through _coro_init. 829 - got rid of indirect call through _coro_init.
547 - updated the partly antique examples in eg/ to 830 - updated the partly antique examples in eg/ to

Diff Legend

– Removed lines
+ Added lines
< Changed lines
> Changed lines