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

Comparing Coro/Changes (file contents):
Revision 1.563 by root, Thu May 12 23:24:27 2011 UTC vs.
Revision 1.682 by root, Tue Feb 18 22:31:43 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: 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?
6 10
7TODO:mg_virtual check 11TODO: enable_times is broken (t_cpu = 4294967287.593... and decreasing)
8TODO:detah on refcnt 12
9TODO:slf_xx_yy=>slf_yy_xx 13 - Coro::EV::time*once functions used ev_once, which is not cancelable.
14 This could cause problems due to I/O watchers not being stopped up
15 on exceptions/cancellation.
16
176.55 Fri Jul 19 17:28:09 CEST 2019
18 - Coro::RWLock woke up only one reader instead of all after a wrlock
19 (analyzed and patch by Maxim Skorobogatov).
20 - slightly improved documentation of Coro::RWLock and RWLock testcase.
21 - quickly (re)-ported to minix 3.3 before minix crashed again.
22
236.54 Sun Dec 16 11:06:07 CET 2018
24 - reinstate Coro::Select::select prototype - it breaks more than it fixes.
25
266.53 Sun Dec 16 10:33:15 CET 2018
27 - the autogenerated prototypes of Coro::AIO now add ";" at the end
28 to avoid generating named unary operators and messing up parsing
29 (mostly affects aio_mlockall 4 | 8 which was parsed as
30 (aio_mlockall 4) | 8).
31 - do not use a prototype for select in Coro::Select, as CORE::select also
32 doesn't have one.
33
346.52 Tue Aug 14 18:49:43 CEST 2018
35 - libcoro update: allow sharing of coro threads using the pthreads
36 backend among different pthreads, as is required by Coro::Multicore.
37 - always allow ->call and ->eval when the coro is the current coroutine. this
38 makes it possible, among other things, to call these on $Coro::main.
39 - minor code cleanups.
40
416.514 Thu Aug 31 18:25:31 CEST 2017
42 - safe_cancel no longer croaks when called on already-destroyed
43 thread, but instead "just works".
44
456.513 Sat Jul 15 01:18:51 CEST 2017
46 - experimental 5.22 comaptibility changes (PadlistNAMES).
47 - removed attributes example and made sure the other examples
48 still work.
49 - fix some statevar types for newer perls, should be of no consequence.
50
516.512 Fri Jul 14 05:03:17 CEST 2017
52 - seems to pass test suite down till 5.8.9, so make this the
53 minimum requirement in Makefile.PL.
54 - experimental 5.24 compatibility changes (SUB_ARGARRAY).
55 - experimental 5.26 compatibility changes (PL_comppad -> PADOFFSET).
56 - disable FORTIFY_SOURCE to avoid broken (but well-intentioned)
57 fortified longjmp variants in some libcs.
58 - re-try JIT allocation without PROT_EXEC, to hopefully improve
59 portability (still doesn't seem to work around broken selinux
60 mmap everywhere).
61 - speling fixes (Xavier Guimard).
62
636.511 Sun Jun 26 23:44:50 CEST 2016
64 - make it compile with threaded perls again.
65 - simplify and speed up __DIE__ and __WARN__ handling.
66
676.51 Sat Jun 25 20:57:20 CEST 2016
68 - 6.5 release didn't compile on most older perls due to
69 a logic error enabling experimental 5.24 code for all versions.
70 - since various broken or deficient patches float around
71 to work around the vtbl API breakage in perl, let's provide
72 our own workaround, which seems to be much less invasive
73 and more compatible than the approaches seen so far. YMMV.
74 - work around assertion failure in perl_destruct on 5.24
75 (analyzed by Dave Mitchell).
76
776.5 Wed Jun 22 22:23:50 CEST 2016
78 - swap_sv swapping order was not symmetrical, causing
79 wrong swaps when swap_sv was used multiple times on the
80 same sv in the same thread.
81 - swap_sv calls can now be undone by calling it again
82 with the same variables.
83 - swap_sv calls will now be undone in async_pool threads.
84 - split Coro::Semaphore::up/adjust into separate xs functions
85 for better error reporting, at slight codesize increase.
86 - (libcoro) arm assembly support, please test and report.
87 - adjust to PL_savestack changes in perl 5.24 (adapted from
88 the debian patch, which unfortunately gets it wrong).
89
906.49 Sat Oct 17 01:40:12 CEST 2015
91 - throwing an exception to a thread waiting in
92 Coro::Handle using Coro::EV did not stop the watchers,
93 causing the next call to fail (testcase by Martin Pritchard).
94 - bump minimum perl version to 5.10.
95
966.48 Sun Oct 4 19:03:51 CEST 2015
97 - fix memory leak when Coro::Handle uses Coro::EV internally
98 (testcase by Sten Sten).
99 - update code to libev 4 API (internally, Coro still used the
100 version 3 API calls).
101 - remove unused hv_sig.
102
1036.47 Sat Jul 11 03:58:20 CEST 2015
104 - swap_svs - NVs can be in the SV head beginning with 5.20.
105
1066.46 Tue Jun 30 14:36:30 CEST 2015
107 - restore portability to perls with windows fork emulation
108 (patch by Petr Písař).
109
1106.45 Tue Jun 30 01:40:08 CEST 2015
111 - restore portability to perl 5.16 and below.
112
1136.44 Tue Jun 30 00:41:54 CEST 2015
114 - avoid segfaulting (or worse) when tracing a canceled thread.
115 - the Event module silently broke it's public hook API, causing
116 failures on perls compiled with -Duselongdouble. Requiring
117 the latest version of Event because I don't know which release
118 changed this (it's not mentioned in the ChangeLog).
119 - allow xs level enterleave hooks via CoroAPI.
120
1216.43 Thu Jun 4 15:38:14 CEST 2015
122 - use stability canary.
123 - port to stableperl-5.22.0-1.001.
124 - update libecb.
125
1266.42 Wed Feb 11 20:29:52 CET 2015
127 - Coro::SemaphoreSet->try did not actually work (analyzed by
128 SATO Kentaro).
129 - upgrade libecb to be C11 compliant.
130
1316.41 Sat Sep 6 22:08:46 CEST 2014
132 - restore portability to perl 5.8.x.
133 - give new Coro's a valid GvHV(PL_hintgv) - this is slow and takes
134 up some memory, but fixes "use feature" and similar modules
135 when used inside a Coro.
136 - allow zero as argument to Coro::Channel to mean the same thing
137 as no argument. this works with older versions as well,
138 but wasn't legal until now.
139 - slightly better c header file detection.
140
1416.39 Mon Jun 2 00:00:08 CEST 2014
142 - work around more incompatible changes in 5.20.
143
1446.38 Sun Jun 1 21:54:23 CEST 2014
145 - check that perl slots actually have enough space to hold
146 interpreter variables.
147 - untested port to perl 5.19 (64 bit tmps indices) (reported
148 by Andreas König).
149 - croak when cancel is called without a thread context.
150
1516.37 Tue Mar 4 13:27:33 CET 2014
152 - *sigh*, removed leftover debugging code from debugging a
153 perl bug, of all things.
154
1556.36 Tue Mar 4 07:11:59 CET 2014
156 - semaphores would not clear the destroy hook when interrupted
157 by ->throw, causing segfaults or worse.
158 - ->throw on a thread waiting for a semaphore did not acquire
159 the semaphore, but also didn't wake up other waiters,
160 possibly causing a deadlock.
161 - "FATAL: $Coro::IDLE blocked itself" will now use Carp::confess
162 to report a full stacktrace, which should help find out
163 where the actual call is.
164 - "atomically" destroy data in slf_destroy, because it is
165 the right thing to do, just in case.
166 - disable ecb.h libm dependency, to compile on ancient systems
167 or under adverse conditions.
168
1696.33 Mon Nov 18 11:26:27 CET 2013
170 - do not crash when freeing padlists with holes (in 5.18).
171 - tentative SVt_BIND 5.19 port/fix.
172
1736.32 Tue Nov 5 15:35:35 CET 2013
174 - use a new algorithm to derive padlists for perl 5.18. The old
175 one could lead to 0-pointer accesses inside perl (reported
176 by Darin McBride).
177
1786.31 Thu May 9 07:39:48 CEST 2013
179 - Coro::AIO requests would crash if the thread was ready'd
180 while the request was ongoing.
181
1826.29 Wed May 8 02:55:18 CEST 2013
183 - when an on_destroy handler destructs the coro currently being
184 destructed a perl scalar could be accessed after being freed,
185 likely causing a crash.
186
1876.28 Wed Mar 6 06:58:02 CET 2013
188 - clean remnants of existing __DIE__ and __WARN__ handlers so
189 they lose their magic and will not cause segfaults later
190 (testcase by Andrey Sagulin).
191 - improved Coro::State documentation a bit.
192 - Coro::Debug::command now flushes the output.
193 - add hack detection code for x32 abi, because the braindead slugs
194 who designed that made it look exactly like x86_64 without
195 providing proper compile time symbols to test for it. as a result,
196 this detection cannot work reliably.
197 - valgrind stack registering was broken.
198 - do not rely on Time::HiRes anymore in Coro::Debug.
199
2006.23 Fri Dec 7 23:36:37 CET 2012
201 - use experimental fiber implementation on native windows
202 perls.
203 - use sizeof (void *) as multiplication factor for stack sizes,
204 to accomodate the totally braindamaged microsoft 64 bit "os".
205 - changed verifier host from win2k-ap510-32 to win7-sp516-32/64.
206 activeperl 5.16 crashes when PerlIO_define_layer is called due
207 to some bug in the perl dll, strawberry perl at least passes
208 the testsuite.
209 - implement Coro::Handle->peeraddr/host/port, for slightly
210 improved compatibility with LWP.
211 - implement 5.17 compatibility by almost blindly applying a
212 good-looking patch by Father Chrysostomos.
213 - move stack management functions into libcoro 3.
214 - libcoro version 3 "released".
215 - support magic values as timed_io_once args.
216 - recommend AnyEvent 7+ or EV 4+, also require EV
217 version 4 or newer for Coro::EV.
218
2196.10 Tue Oct 9 01:14:27 CEST 2012
220 - updated ecb.h, it had a typo that caused it to not compile on many
221 big endian systems (reported by many people).
222 - disable memory fences in ecb.h to improve portability.
223
2246.09 Sat Oct 6 23:25:02 CEST 2012
225 - Coro::EV I/O watchers were not interruptible by exceptions
226 (Coro::State::throw) (testcase by sten).
227 - ->throw now puts threads into the ready queue, as this seems to
228 be expected by existing code, and code that doesn't cope with spurious
229 wakeups needs fixing anyway.
230 - use fd -1 in mmap.
231 - cast I32 to int in error message printf.
232 - warn about broken so-called "hardened" kernels.
233
2346.08 Fri Apr 13 12:05:47 CEST 2012
235 - be more aggressive about exiting like perl does - formerly,
236 exiting from the non-main thread would not execute END blocks.
237
2386.07 Fri Nov 11 21:21:48 CET 2011
239 - work around a bug in PerlIO (setting $SIG{__WARN__} to a PVCV).
240 - update ecb.h.
241
2426.06 Mon Aug 8 23:59:48 CEST 2011
243 - cygwin unfortunately patches the stack at runtime, so we use the pthreads
244 backend, which is an order of magnitude slower. unfortunately, cygwins
245 pthread implementation isn't very complete either, so allocate the stack
246 twice just to be sure.
247 (note: cygwin also enables mymalloc, which is NOT THREADSAFE ON WINDOWS,
248 in its ithreaded perl - best recompile cygwin and use the 'w'indows
249 backend for much better performance. also disable ithreads for
250 even better performance...).
251
2526.05 Thu Aug 4 21:36:36 CEST 2011
253 - blush, condvar values would not be propagated from send to recv anymore
254 (reported by Chip Salzenberg).
255 - use exponential increase for the readline buffer length in
256 Coro::Handle. also reduce initial allocation to 1020 from 4096 bytes.
257
2586.04 Wed Aug 3 17:15:45 CEST 2011
259 - use even more efficient and more compatible condvars for
260 compatibility to AnyEvent 6.x :)
261 - more inconsequential ecb.h updates.
262
2636.03 Wed Aug 3 11:41:30 CEST 2011
264 - change how Coro patches AnyEvent condvars for compatibility to
265 AnyEvent 6.x.
266 - update ecb.h, to no longer include <pthread.h> in case WinNT.h
267 hasn't been included.
268
2696.02 Wed Jul 13 04:35:19 CEST 2011
270 - "improve portability to Gentoo" - gentoo manages to put perl variables
271 in memory areas that are farther than 2gb apart, which the jit couldn't
272 handle and barfed. now it's just a bit slower on gentoo and similar
273 systems.
274
2756.01 Sun Jul 3 12:31:14 CEST 2011
276 - workarounds are good, but the test for whether pthreads are used
277 was not good. this one should be better.
278 - check differently whether gcc generates cfi instructions itself.
279
2806.0 Wed Jun 29 19:43:35 CEST 2011
281 - INCOMPATIBLE CHANGE: unreferenced coro objects will now be
282 destroyed and cleaned up automatically (e.g. async { schedule }).
283 - implement a JIT compiler for part of the thread switch code,
284 which gives a 50% speed improvement on threaded perls, and
285 about 4% on non-threaded perls (so threaded perls now finally
286 reach about half the speed of non-threaded perls).
287 - slightly modernise Coro::Intro, add section about rouse functions.
288 - avoid DEFSV and ERRSV, giving another 10% improvement
289 in thread switching.
290 - Coro::State->is_destroyed is now called is_zombie.
291 - implement a Coro->safe_cancel method that might fail, but
292 cancels in a "safer" way if it succeeds.
293 - add preliminary support for DEBUGGING perls.
294 - get rid of two hash-accesses when initialising a new Coro - this
295 speeds up coro creation by almost a factor of two.
10 - croak when a coro that is being cancelled tries to block 296 - croak when a coro that is being cancelled tries to block
11 (e.g. while executing a guard block), instead of crashing or 297 (e.g. while executing a guard block), instead of crashing or
12 deadlocking. 298 deadlocking.
13 - use a more robust and also faster method to identify Coro::State 299 - use a more robust and also faster method to identify Coro::State
14 objects - speeds up everything a bit. 300 objects - speeds up everything a bit.
15 - implement a Coro->safe_cancel method that might fail, but
16 cancels in a "safer" way if it succeeds.
17 - implement Coro->cancel in XS for a 20% speed improvement, and to 301 - implement Coro->cancel in XS for a 20% speed improvement, and to
18 be able to implement mutual cancellation. 302 be able to implement mutual cancellation.
19 - speed up context switches by a percent or two by more efficiently 303 - speed up context switches by a percent or two by more efficiently
20 allocating context stack entries. 304 allocating context stack entries.
21 - implement Coro->join and Coro->on_destroy in XS for a speedup and 305 - implement Coro->join and Coro->on_destroy in XS for a speedup and
… …
26 callbacks (best effort). 310 callbacks (best effort).
27 - move on_destroy into the slf_frame, to allow extension slf 311 - move on_destroy into the slf_frame, to allow extension slf
28 functions to have destructors. 312 functions to have destructors.
29 - get rid if coro refcounting - simply crash in other interpreter 313 - get rid if coro refcounting - simply crash in other interpreter
30 threads by nulling the pointers on clone. 314 threads by nulling the pointers on clone.
315 - simplify warn/die hook handling when loading Coro - the convoluted
316 logic seems to be no longer neccessary.
317 - use libecb instead of our own home-grown gcc hacks.
318 - document alternatives to Coro::LWP. Please use them :)
319 - work around another mindless idiotic NEEDLESS bug in openbsd/mirbsds
320 sigaltstack. Really. wine suffers from it, erlang suffers from it,
321 and it's known since at least 2006.
31 322
325.372 Wed Feb 23 06:14:30 CET 2011 3235.372 Wed Feb 23 06:14:30 CET 2011
33 - apparently mingw doesn't provide a working gettimeofday, try to 324 - apparently mingw doesn't provide a working gettimeofday, try to
34 work around that by relying on Time::HiRes (indirectly brought to 325 work around that by relying on Time::HiRes (indirectly brought to
35 my attention by Max Maischein). 326 my attention by Max Maischein).
… …
353 - ->throw is now supported on Coro::State objects. 644 - ->throw is now supported on Coro::State objects.
354 - clean up cctx creation code a bit. 645 - clean up cctx creation code a bit.
355 - entersub is actually a UNOP, not a LOGOP (not a bugfix). 646 - entersub is actually a UNOP, not a LOGOP (not a bugfix).
356 647
3574.9 Sat Nov 8 17:45:27 CET 2008 6484.9 Sat Nov 8 17:45:27 CET 2008
358 - (libcoro) did not preserve rbp with CORO_ASM (we are getting there). 649 - (libcoro) did not preserve rbp with CORO_ASM (we are getting there).
359 - (libcoro) no longer leak threads in the experimental pthread backend, 650 - (libcoro) no longer leak threads in the experimental pthread backend,
360 also speed it up considerably. 651 also speed it up considerably.
361 - (libcoro) do not rely on makecontext passing void *'s unscathed. 652 - (libcoro) do not rely on makecontext passing void *'s unscathed.
362 - fix compiletime dependencies on libcoro in the Makefile. 653 - fix compiletime dependencies on libcoro in the Makefile.
363 - cctx_count wasn't always updated properly. 654 - cctx_count wasn't always updated properly.
… …
425 - remove debugging code related to MgPV_nolen_const, also try to 716 - remove debugging code related to MgPV_nolen_const, also try to
426 make it compile with perl 5.8.6 (yes, apple apparently loves 717 make it compile with perl 5.8.6 (yes, apple apparently loves
427 outdated software). Reported by John S. 718 outdated software). Reported by John S.
428 719
4294.744 Tue Jul 8 22:06:35 CEST 2008 7204.744 Tue Jul 8 22:06:35 CEST 2008
430 - correctly provide default DIE/WARN handlers as documented. 721 - correctly provide default DIE/WARN handlers as documented.
431 - also overwrite PL_vtbl_sigelem.svt_clear, even though current 722 - also overwrite PL_vtbl_sigelem.svt_clear, even though current
432 implementations inside perl work fine for us. 723 implementations inside perl work fine for us.
433 724
4344.743 Mon Jun 16 00:21:57 CEST 2008 7254.743 Mon Jun 16 00:21:57 CEST 2008
435 - when using Coro::EV without running EV::loop it could 726 - when using Coro::EV without running EV::loop it could
… …
457 - sprinkle "no warnings" freely over everything, also suppress 748 - sprinkle "no warnings" freely over everything, also suppress
458 warnings for some other modules. 749 warnings for some other modules.
459 - fix typo in WSAEWOULDBLOCK. 750 - fix typo in WSAEWOULDBLOCK.
460 751
4614.72 Sun May 25 05:14:36 CEST 2008 7524.72 Sun May 25 05:14:36 CEST 2008
462 - tweak META.yaml a bit, unfortunately, there is no documented way 753 - tweak META.yaml a bit, unfortunately, there is no documented way
463 to have optional dependencies with CPAN. doh :( 754 to have optional dependencies with CPAN. doh :(
464 - avoid running some tests on windows because they would fail due to 755 - avoid running some tests on windows because they would fail due to
465 perl bug (broken fork, broken pipes...). 756 perl bug (broken fork, broken pipes...).
466 - work around perl on windows bugs where perl returns undocumented 757 - work around perl on windows bugs where perl returns undocumented
467 error codes for sysread, syswrite etc. by taking advantage 758 error codes for sysread, syswrite etc. by taking advantage
… …
471 - use unix domain sockets in testsuite to work around 762 - use unix domain sockets in testsuite to work around
472 common perl implementation bugs on widows (they are emulated by 763 common perl implementation bugs on widows (they are emulated by
473 tcp sockets on windows. ugh.) 764 tcp sockets on windows. ugh.)
474 765
4754.71 Sat May 24 20:01:27 CEST 2008 7664.71 Sat May 24 20:01:27 CEST 2008
476 - fix a bug in Coro::AnyEvent ("Usage: Coro::AnyEvent::_schedule()"). 767 - fix a bug in Coro::AnyEvent ("Usage: Coro::AnyEvent::_schedule()").
477 - take advantage of async name resolution of AnyEvent::Util. 768 - take advantage of async name resolution of AnyEvent::Util.
478 - work around brutal inet_aton override in Coro::LWP. 769 - work around brutal inet_aton override in Coro::LWP.
479 - take advantage of the readyhook in Coro::EV, for smoother 770 - take advantage of the readyhook in Coro::EV, for smoother
480 scheduling. 771 scheduling.
481 772
4824.7 Sun May 11 00:32:19 CEST 2008 7734.7 Sun May 11 00:32:19 CEST 2008
483 - completely reworked the Coro manpage. 774 - completely reworked the Coro manpage.
484 - added Coro::AnyEvent, generic event loop integration. 775 - added Coro::AnyEvent, generic event loop integration.
485 - implement cancel, ready and kill commands in Coro::Debug. 776 - implement cancel, ready and kill commands in Coro::Debug.
486 - document find_coro in Coro::Debug. 777 - document find_coro in Coro::Debug.
487 - incompatible change: rename has_stack to has_cctx. 778 - incompatible change: rename has_stack to has_cctx.
488 - Coro::AIO and Coro::BDB no longer force event model detection, 779 - Coro::AIO and Coro::BDB no longer force event model detection,
489 use AnyEvent::AIO and AnyEvent::BDB. 780 use AnyEvent::AIO and AnyEvent::BDB.
… …
613 904
6144.11 Thu Oct 11 02:40:24 CEST 2007 9054.11 Thu Oct 11 02:40:24 CEST 2007
615 - port to threaded perls. 906 - port to threaded perls.
616 907
6174.1 Thu Oct 11 02:38:16 CEST 2007 9084.1 Thu Oct 11 02:38:16 CEST 2007
618 - incompatible change: $SIG{__DIE__} and $SIG{__WARN__} will now 909 - incompatible change: $SIG{__DIE__} and $SIG{__WARN__} will now
619 be local to each coro (see Coro::State). 910 be local to each coro (see Coro::State).
620 - incompatible change: for very deep reasons, cede and cede_notself 911 - incompatible change: for very deep reasons, cede and cede_notself
621 cannot return anything, so nothing will be returned. 912 cannot return anything, so nothing will be returned.
622 - possibly bring back 5.10 compatibility (untested). 913 - possibly bring back 5.10 compatibility (untested).
623 - work around stupid (and wrong) warning on 5.10 :(. 914 - work around stupid (and wrong) warning on 5.10 :(.
… …
695 but stateful semantics. 986 but stateful semantics.
696 - fixed a lot of typos in Coro.pm (patch submitted by David 987 - fixed a lot of typos in Coro.pm (patch submitted by David
697 Steinbrunner, which applied flawlessly). 988 Steinbrunner, which applied flawlessly).
698 989
6993.6 Sat Apr 14 17:13:31 CEST 2007 9903.6 Sat Apr 14 17:13:31 CEST 2007
700 - added some bugfixes to get eg/myhttpd working again. 991 - added some bugfixes to get eg/myhttpd working again.
701 - added Coro::Storable for often-cede'ing freeze/thaw. 992 - added Coro::Storable for often-cede'ing freeze/thaw.
702 - try to do a clean exit when a coroutine calls exit 993 - try to do a clean exit when a coroutine calls exit
703 (EXPERIMENTAL). 994 (EXPERIMENTAL).
704 - got rid of indirect call through _coro_init. 995 - got rid of indirect call through _coro_init.
705 - updated the partly antique examples in eg/ to 996 - updated the partly antique examples in eg/ to

Diff Legend

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