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

Comparing Coro/Changes (file contents):
Revision 1.556 by root, Fri Apr 29 17:38:56 2011 UTC vs.
Revision 1.683 by root, Thu Feb 20 18:42:14 2020 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: channel->maxsize(newsize)? 4TODO: channel->maxsize(newsize)?
5TODO: http://www.microsoft.com/msj/archive/s2ce.aspx 5TODO: __GCC_HAVE_DWARF2_CFI_ASM
6TODO: cancel calling another coro that cancels ourselves => crash (CF_DELAYCANCEL?) -> canceltest 6TODO: swap_sv, maybe add scope_swap_sv?
7TODO: croak when async_pool tries to run canceled thread?
8TODO: how to debug safe_cancel failure (vebrose mode? show how to make backtrace?)
9TODO: scalar context for Coro::AIO wrappers?
7 10
11TODO: enable_times is broken (t_cpu = 4294967287.593... and decreasing)
12
136.56 Thu Feb 20 19:40:07 CET 2020
14 - Coro::EV::time*once functions used ev_once, which is not cancelable.
15 This could cause problems due to I/O watchers not being stopped up
16 on exceptions/cancellation.
17
186.55 Fri Jul 19 17:28:09 CEST 2019
19 - Coro::RWLock woke up only one reader instead of all after a wrlock
20 (analyzed and patch by Maxim Skorobogatov).
21 - slightly improved documentation of Coro::RWLock and RWLock testcase.
22 - quickly (re)-ported to minix 3.3 before minix crashed again.
23
246.54 Sun Dec 16 11:06:07 CET 2018
25 - reinstate Coro::Select::select prototype - it breaks more than it fixes.
26
276.53 Sun Dec 16 10:33:15 CET 2018
28 - the autogenerated prototypes of Coro::AIO now add ";" at the end
29 to avoid generating named unary operators and messing up parsing
30 (mostly affects aio_mlockall 4 | 8 which was parsed as
31 (aio_mlockall 4) | 8).
32 - do not use a prototype for select in Coro::Select, as CORE::select also
33 doesn't have one.
34
356.52 Tue Aug 14 18:49:43 CEST 2018
36 - libcoro update: allow sharing of coro threads using the pthreads
37 backend among different pthreads, as is required by Coro::Multicore.
38 - always allow ->call and ->eval when the coro is the current coroutine. this
39 makes it possible, among other things, to call these on $Coro::main.
40 - minor code cleanups.
41
426.514 Thu Aug 31 18:25:31 CEST 2017
43 - safe_cancel no longer croaks when called on already-destroyed
44 thread, but instead "just works".
45
466.513 Sat Jul 15 01:18:51 CEST 2017
47 - experimental 5.22 comaptibility changes (PadlistNAMES).
48 - removed attributes example and made sure the other examples
49 still work.
50 - fix some statevar types for newer perls, should be of no consequence.
51
526.512 Fri Jul 14 05:03:17 CEST 2017
53 - seems to pass test suite down till 5.8.9, so make this the
54 minimum requirement in Makefile.PL.
55 - experimental 5.24 compatibility changes (SUB_ARGARRAY).
56 - experimental 5.26 compatibility changes (PL_comppad -> PADOFFSET).
57 - disable FORTIFY_SOURCE to avoid broken (but well-intentioned)
58 fortified longjmp variants in some libcs.
59 - re-try JIT allocation without PROT_EXEC, to hopefully improve
60 portability (still doesn't seem to work around broken selinux
61 mmap everywhere).
62 - speling fixes (Xavier Guimard).
63
646.511 Sun Jun 26 23:44:50 CEST 2016
65 - make it compile with threaded perls again.
66 - simplify and speed up __DIE__ and __WARN__ handling.
67
686.51 Sat Jun 25 20:57:20 CEST 2016
69 - 6.5 release didn't compile on most older perls due to
70 a logic error enabling experimental 5.24 code for all versions.
71 - since various broken or deficient patches float around
72 to work around the vtbl API breakage in perl, let's provide
73 our own workaround, which seems to be much less invasive
74 and more compatible than the approaches seen so far. YMMV.
75 - work around assertion failure in perl_destruct on 5.24
76 (analyzed by Dave Mitchell).
77
786.5 Wed Jun 22 22:23:50 CEST 2016
79 - swap_sv swapping order was not symmetrical, causing
80 wrong swaps when swap_sv was used multiple times on the
81 same sv in the same thread.
82 - swap_sv calls can now be undone by calling it again
83 with the same variables.
84 - swap_sv calls will now be undone in async_pool threads.
85 - split Coro::Semaphore::up/adjust into separate xs functions
86 for better error reporting, at slight codesize increase.
87 - (libcoro) arm assembly support, please test and report.
88 - adjust to PL_savestack changes in perl 5.24 (adapted from
89 the debian patch, which unfortunately gets it wrong).
90
916.49 Sat Oct 17 01:40:12 CEST 2015
92 - throwing an exception to a thread waiting in
93 Coro::Handle using Coro::EV did not stop the watchers,
94 causing the next call to fail (testcase by Martin Pritchard).
95 - bump minimum perl version to 5.10.
96
976.48 Sun Oct 4 19:03:51 CEST 2015
98 - fix memory leak when Coro::Handle uses Coro::EV internally
99 (testcase by Sten Sten).
100 - update code to libev 4 API (internally, Coro still used the
101 version 3 API calls).
102 - remove unused hv_sig.
103
1046.47 Sat Jul 11 03:58:20 CEST 2015
105 - swap_svs - NVs can be in the SV head beginning with 5.20.
106
1076.46 Tue Jun 30 14:36:30 CEST 2015
108 - restore portability to perls with windows fork emulation
109 (patch by Petr Písař).
110
1116.45 Tue Jun 30 01:40:08 CEST 2015
112 - restore portability to perl 5.16 and below.
113
1146.44 Tue Jun 30 00:41:54 CEST 2015
115 - avoid segfaulting (or worse) when tracing a canceled thread.
116 - the Event module silently broke it's public hook API, causing
117 failures on perls compiled with -Duselongdouble. Requiring
118 the latest version of Event because I don't know which release
119 changed this (it's not mentioned in the ChangeLog).
120 - allow xs level enterleave hooks via CoroAPI.
121
1226.43 Thu Jun 4 15:38:14 CEST 2015
123 - use stability canary.
124 - port to stableperl-5.22.0-1.001.
125 - update libecb.
126
1276.42 Wed Feb 11 20:29:52 CET 2015
128 - Coro::SemaphoreSet->try did not actually work (analyzed by
129 SATO Kentaro).
130 - upgrade libecb to be C11 compliant.
131
1326.41 Sat Sep 6 22:08:46 CEST 2014
133 - restore portability to perl 5.8.x.
134 - give new Coro's a valid GvHV(PL_hintgv) - this is slow and takes
135 up some memory, but fixes "use feature" and similar modules
136 when used inside a Coro.
137 - allow zero as argument to Coro::Channel to mean the same thing
138 as no argument. this works with older versions as well,
139 but wasn't legal until now.
140 - slightly better c header file detection.
141
1426.39 Mon Jun 2 00:00:08 CEST 2014
143 - work around more incompatible changes in 5.20.
144
1456.38 Sun Jun 1 21:54:23 CEST 2014
146 - check that perl slots actually have enough space to hold
147 interpreter variables.
148 - untested port to perl 5.19 (64 bit tmps indices) (reported
149 by Andreas König).
150 - croak when cancel is called without a thread context.
151
1526.37 Tue Mar 4 13:27:33 CET 2014
153 - *sigh*, removed leftover debugging code from debugging a
154 perl bug, of all things.
155
1566.36 Tue Mar 4 07:11:59 CET 2014
157 - semaphores would not clear the destroy hook when interrupted
158 by ->throw, causing segfaults or worse.
159 - ->throw on a thread waiting for a semaphore did not acquire
160 the semaphore, but also didn't wake up other waiters,
161 possibly causing a deadlock.
162 - "FATAL: $Coro::IDLE blocked itself" will now use Carp::confess
163 to report a full stacktrace, which should help find out
164 where the actual call is.
165 - "atomically" destroy data in slf_destroy, because it is
166 the right thing to do, just in case.
167 - disable ecb.h libm dependency, to compile on ancient systems
168 or under adverse conditions.
169
1706.33 Mon Nov 18 11:26:27 CET 2013
171 - do not crash when freeing padlists with holes (in 5.18).
172 - tentative SVt_BIND 5.19 port/fix.
173
1746.32 Tue Nov 5 15:35:35 CET 2013
175 - use a new algorithm to derive padlists for perl 5.18. The old
176 one could lead to 0-pointer accesses inside perl (reported
177 by Darin McBride).
178
1796.31 Thu May 9 07:39:48 CEST 2013
180 - Coro::AIO requests would crash if the thread was ready'd
181 while the request was ongoing.
182
1836.29 Wed May 8 02:55:18 CEST 2013
184 - when an on_destroy handler destructs the coro currently being
185 destructed a perl scalar could be accessed after being freed,
186 likely causing a crash.
187
1886.28 Wed Mar 6 06:58:02 CET 2013
189 - clean remnants of existing __DIE__ and __WARN__ handlers so
190 they lose their magic and will not cause segfaults later
191 (testcase by Andrey Sagulin).
192 - improved Coro::State documentation a bit.
193 - Coro::Debug::command now flushes the output.
194 - add hack detection code for x32 abi, because the braindead slugs
195 who designed that made it look exactly like x86_64 without
196 providing proper compile time symbols to test for it. as a result,
197 this detection cannot work reliably.
198 - valgrind stack registering was broken.
199 - do not rely on Time::HiRes anymore in Coro::Debug.
200
2016.23 Fri Dec 7 23:36:37 CET 2012
202 - use experimental fiber implementation on native windows
203 perls.
204 - use sizeof (void *) as multiplication factor for stack sizes,
205 to accomodate the totally braindamaged microsoft 64 bit "os".
206 - changed verifier host from win2k-ap510-32 to win7-sp516-32/64.
207 activeperl 5.16 crashes when PerlIO_define_layer is called due
208 to some bug in the perl dll, strawberry perl at least passes
209 the testsuite.
210 - implement Coro::Handle->peeraddr/host/port, for slightly
211 improved compatibility with LWP.
212 - implement 5.17 compatibility by almost blindly applying a
213 good-looking patch by Father Chrysostomos.
214 - move stack management functions into libcoro 3.
215 - libcoro version 3 "released".
216 - support magic values as timed_io_once args.
217 - recommend AnyEvent 7+ or EV 4+, also require EV
218 version 4 or newer for Coro::EV.
219
2206.10 Tue Oct 9 01:14:27 CEST 2012
221 - updated ecb.h, it had a typo that caused it to not compile on many
222 big endian systems (reported by many people).
223 - disable memory fences in ecb.h to improve portability.
224
2256.09 Sat Oct 6 23:25:02 CEST 2012
226 - Coro::EV I/O watchers were not interruptible by exceptions
227 (Coro::State::throw) (testcase by sten).
228 - ->throw now puts threads into the ready queue, as this seems to
229 be expected by existing code, and code that doesn't cope with spurious
230 wakeups needs fixing anyway.
231 - use fd -1 in mmap.
232 - cast I32 to int in error message printf.
233 - warn about broken so-called "hardened" kernels.
234
2356.08 Fri Apr 13 12:05:47 CEST 2012
236 - be more aggressive about exiting like perl does - formerly,
237 exiting from the non-main thread would not execute END blocks.
238
2396.07 Fri Nov 11 21:21:48 CET 2011
240 - work around a bug in PerlIO (setting $SIG{__WARN__} to a PVCV).
241 - update ecb.h.
242
2436.06 Mon Aug 8 23:59:48 CEST 2011
244 - cygwin unfortunately patches the stack at runtime, so we use the pthreads
245 backend, which is an order of magnitude slower. unfortunately, cygwins
246 pthread implementation isn't very complete either, so allocate the stack
247 twice just to be sure.
248 (note: cygwin also enables mymalloc, which is NOT THREADSAFE ON WINDOWS,
249 in its ithreaded perl - best recompile cygwin and use the 'w'indows
250 backend for much better performance. also disable ithreads for
251 even better performance...).
252
2536.05 Thu Aug 4 21:36:36 CEST 2011
254 - blush, condvar values would not be propagated from send to recv anymore
255 (reported by Chip Salzenberg).
256 - use exponential increase for the readline buffer length in
257 Coro::Handle. also reduce initial allocation to 1020 from 4096 bytes.
258
2596.04 Wed Aug 3 17:15:45 CEST 2011
260 - use even more efficient and more compatible condvars for
261 compatibility to AnyEvent 6.x :)
262 - more inconsequential ecb.h updates.
263
2646.03 Wed Aug 3 11:41:30 CEST 2011
265 - change how Coro patches AnyEvent condvars for compatibility to
266 AnyEvent 6.x.
267 - update ecb.h, to no longer include <pthread.h> in case WinNT.h
268 hasn't been included.
269
2706.02 Wed Jul 13 04:35:19 CEST 2011
271 - "improve portability to Gentoo" - gentoo manages to put perl variables
272 in memory areas that are farther than 2gb apart, which the jit couldn't
273 handle and barfed. now it's just a bit slower on gentoo and similar
274 systems.
275
2766.01 Sun Jul 3 12:31:14 CEST 2011
277 - workarounds are good, but the test for whether pthreads are used
278 was not good. this one should be better.
279 - check differently whether gcc generates cfi instructions itself.
280
2816.0 Wed Jun 29 19:43:35 CEST 2011
282 - INCOMPATIBLE CHANGE: unreferenced coro objects will now be
283 destroyed and cleaned up automatically (e.g. async { schedule }).
284 - implement a JIT compiler for part of the thread switch code,
285 which gives a 50% speed improvement on threaded perls, and
286 about 4% on non-threaded perls (so threaded perls now finally
287 reach about half the speed of non-threaded perls).
288 - slightly modernise Coro::Intro, add section about rouse functions.
289 - avoid DEFSV and ERRSV, giving another 10% improvement
290 in thread switching.
291 - Coro::State->is_destroyed is now called is_zombie.
292 - implement a Coro->safe_cancel method that might fail, but
293 cancels in a "safer" way if it succeeds.
294 - add preliminary support for DEBUGGING perls.
295 - get rid of two hash-accesses when initialising a new Coro - this
296 speeds up coro creation by almost a factor of two.
297 - croak when a coro that is being cancelled tries to block
298 (e.g. while executing a guard block), instead of crashing or
299 deadlocking.
300 - use a more robust and also faster method to identify Coro::State
301 objects - speeds up everything a bit.
8 - implement Coro->cancel in XS for a 20% speed improvement, and to 302 - implement Coro->cancel in XS for a 20% speed improvement, and to
9 be able to implement mutual cancellation. 303 be able to implement mutual cancellation.
304 - speed up context switches by a percent or two by more efficiently
305 allocating context stack entries.
306 - implement Coro->join and Coro->on_destroy in XS for a speedup and
307 a reduction in memory use.
10 - cancelling a coro while it itself is cancelling another coro is 308 - cancelling a coro while it itself is cancelling another coro is
11 now supported and working, instead of triggering an assertion. 309 now supported and working, instead of triggering an assertion.
12 - be a bit more crash-resistant when calling (buggy) on_destroy 310 - be a bit more crash-resistant when calling (buggy) on_destroy
13 callbacks (best effort). 311 callbacks (best effort).
312 - move on_destroy into the slf_frame, to allow extension slf
313 functions to have destructors.
314 - get rid if coro refcounting - simply crash in other interpreter
315 threads by nulling the pointers on clone.
316 - simplify warn/die hook handling when loading Coro - the convoluted
317 logic seems to be no longer neccessary.
318 - use libecb instead of our own home-grown gcc hacks.
319 - document alternatives to Coro::LWP. Please use them :)
320 - work around another mindless idiotic NEEDLESS bug in openbsd/mirbsds
321 sigaltstack. Really. wine suffers from it, erlang suffers from it,
322 and it's known since at least 2006.
14 323
155.372 Wed Feb 23 06:14:30 CET 2011 3245.372 Wed Feb 23 06:14:30 CET 2011
16 - apparently mingw doesn't provide a working gettimeofday, try to 325 - apparently mingw doesn't provide a working gettimeofday, try to
17 work around that by relying on Time::HiRes (indirectly brought to 326 work around that by relying on Time::HiRes (indirectly brought to
18 my attention by Max Maischein). 327 my attention by Max Maischein).
… …
336 - ->throw is now supported on Coro::State objects. 645 - ->throw is now supported on Coro::State objects.
337 - clean up cctx creation code a bit. 646 - clean up cctx creation code a bit.
338 - entersub is actually a UNOP, not a LOGOP (not a bugfix). 647 - entersub is actually a UNOP, not a LOGOP (not a bugfix).
339 648
3404.9 Sat Nov 8 17:45:27 CET 2008 6494.9 Sat Nov 8 17:45:27 CET 2008
341 - (libcoro) did not preserve rbp with CORO_ASM (we are getting there). 650 - (libcoro) did not preserve rbp with CORO_ASM (we are getting there).
342 - (libcoro) no longer leak threads in the experimental pthread backend, 651 - (libcoro) no longer leak threads in the experimental pthread backend,
343 also speed it up considerably. 652 also speed it up considerably.
344 - (libcoro) do not rely on makecontext passing void *'s unscathed. 653 - (libcoro) do not rely on makecontext passing void *'s unscathed.
345 - fix compiletime dependencies on libcoro in the Makefile. 654 - fix compiletime dependencies on libcoro in the Makefile.
346 - cctx_count wasn't always updated properly. 655 - cctx_count wasn't always updated properly.
… …
408 - remove debugging code related to MgPV_nolen_const, also try to 717 - remove debugging code related to MgPV_nolen_const, also try to
409 make it compile with perl 5.8.6 (yes, apple apparently loves 718 make it compile with perl 5.8.6 (yes, apple apparently loves
410 outdated software). Reported by John S. 719 outdated software). Reported by John S.
411 720
4124.744 Tue Jul 8 22:06:35 CEST 2008 7214.744 Tue Jul 8 22:06:35 CEST 2008
413 - correctly provide default DIE/WARN handlers as documented. 722 - correctly provide default DIE/WARN handlers as documented.
414 - also overwrite PL_vtbl_sigelem.svt_clear, even though current 723 - also overwrite PL_vtbl_sigelem.svt_clear, even though current
415 implementations inside perl work fine for us. 724 implementations inside perl work fine for us.
416 725
4174.743 Mon Jun 16 00:21:57 CEST 2008 7264.743 Mon Jun 16 00:21:57 CEST 2008
418 - when using Coro::EV without running EV::loop it could 727 - when using Coro::EV without running EV::loop it could
… …
440 - sprinkle "no warnings" freely over everything, also suppress 749 - sprinkle "no warnings" freely over everything, also suppress
441 warnings for some other modules. 750 warnings for some other modules.
442 - fix typo in WSAEWOULDBLOCK. 751 - fix typo in WSAEWOULDBLOCK.
443 752
4444.72 Sun May 25 05:14:36 CEST 2008 7534.72 Sun May 25 05:14:36 CEST 2008
445 - tweak META.yaml a bit, unfortunately, there is no documented way 754 - tweak META.yaml a bit, unfortunately, there is no documented way
446 to have optional dependencies with CPAN. doh :( 755 to have optional dependencies with CPAN. doh :(
447 - avoid running some tests on windows because they would fail due to 756 - avoid running some tests on windows because they would fail due to
448 perl bug (broken fork, broken pipes...). 757 perl bug (broken fork, broken pipes...).
449 - work around perl on windows bugs where perl returns undocumented 758 - work around perl on windows bugs where perl returns undocumented
450 error codes for sysread, syswrite etc. by taking advantage 759 error codes for sysread, syswrite etc. by taking advantage
… …
454 - use unix domain sockets in testsuite to work around 763 - use unix domain sockets in testsuite to work around
455 common perl implementation bugs on widows (they are emulated by 764 common perl implementation bugs on widows (they are emulated by
456 tcp sockets on windows. ugh.) 765 tcp sockets on windows. ugh.)
457 766
4584.71 Sat May 24 20:01:27 CEST 2008 7674.71 Sat May 24 20:01:27 CEST 2008
459 - fix a bug in Coro::AnyEvent ("Usage: Coro::AnyEvent::_schedule()"). 768 - fix a bug in Coro::AnyEvent ("Usage: Coro::AnyEvent::_schedule()").
460 - take advantage of async name resolution of AnyEvent::Util. 769 - take advantage of async name resolution of AnyEvent::Util.
461 - work around brutal inet_aton override in Coro::LWP. 770 - work around brutal inet_aton override in Coro::LWP.
462 - take advantage of the readyhook in Coro::EV, for smoother 771 - take advantage of the readyhook in Coro::EV, for smoother
463 scheduling. 772 scheduling.
464 773
4654.7 Sun May 11 00:32:19 CEST 2008 7744.7 Sun May 11 00:32:19 CEST 2008
466 - completely reworked the Coro manpage. 775 - completely reworked the Coro manpage.
467 - added Coro::AnyEvent, generic event loop integration. 776 - added Coro::AnyEvent, generic event loop integration.
468 - implement cancel, ready and kill commands in Coro::Debug. 777 - implement cancel, ready and kill commands in Coro::Debug.
469 - document find_coro in Coro::Debug. 778 - document find_coro in Coro::Debug.
470 - incompatible change: rename has_stack to has_cctx. 779 - incompatible change: rename has_stack to has_cctx.
471 - Coro::AIO and Coro::BDB no longer force event model detection, 780 - Coro::AIO and Coro::BDB no longer force event model detection,
472 use AnyEvent::AIO and AnyEvent::BDB. 781 use AnyEvent::AIO and AnyEvent::BDB.
… …
596 905
5974.11 Thu Oct 11 02:40:24 CEST 2007 9064.11 Thu Oct 11 02:40:24 CEST 2007
598 - port to threaded perls. 907 - port to threaded perls.
599 908
6004.1 Thu Oct 11 02:38:16 CEST 2007 9094.1 Thu Oct 11 02:38:16 CEST 2007
601 - incompatible change: $SIG{__DIE__} and $SIG{__WARN__} will now 910 - incompatible change: $SIG{__DIE__} and $SIG{__WARN__} will now
602 be local to each coro (see Coro::State). 911 be local to each coro (see Coro::State).
603 - incompatible change: for very deep reasons, cede and cede_notself 912 - incompatible change: for very deep reasons, cede and cede_notself
604 cannot return anything, so nothing will be returned. 913 cannot return anything, so nothing will be returned.
605 - possibly bring back 5.10 compatibility (untested). 914 - possibly bring back 5.10 compatibility (untested).
606 - work around stupid (and wrong) warning on 5.10 :(. 915 - work around stupid (and wrong) warning on 5.10 :(.
… …
678 but stateful semantics. 987 but stateful semantics.
679 - fixed a lot of typos in Coro.pm (patch submitted by David 988 - fixed a lot of typos in Coro.pm (patch submitted by David
680 Steinbrunner, which applied flawlessly). 989 Steinbrunner, which applied flawlessly).
681 990
6823.6 Sat Apr 14 17:13:31 CEST 2007 9913.6 Sat Apr 14 17:13:31 CEST 2007
683 - added some bugfixes to get eg/myhttpd working again. 992 - added some bugfixes to get eg/myhttpd working again.
684 - added Coro::Storable for often-cede'ing freeze/thaw. 993 - added Coro::Storable for often-cede'ing freeze/thaw.
685 - try to do a clean exit when a coroutine calls exit 994 - try to do a clean exit when a coroutine calls exit
686 (EXPERIMENTAL). 995 (EXPERIMENTAL).
687 - got rid of indirect call through _coro_init. 996 - got rid of indirect call through _coro_init.
688 - updated the partly antique examples in eg/ to 997 - updated the partly antique examples in eg/ to

Diff Legend

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