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

Comparing Coro/Changes (file contents):
Revision 1.616 by root, Wed May 8 00:55:40 2013 UTC vs.
Revision 1.690 by root, Wed Jul 29 13:09:19 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: __GCC_HAVE_DWARF2_CFI_ASM 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?
10
11TODO: State.xs could short-circuit some macros such as CORO_EXECUTE_SLF
12TODO: use frame->destroy to improve Coro::EV
13
146.57 Wed Jul 29 15:08:44 CEST 2020
15 - override AnyEvent condvar methods "properly", as AnyEvent
16 has support for pluggable condvars since 2011. This also
17 avoids problems with recursive waits by avoiding to mix
18 impl methods with coro code.
19 - allow rouse_wait to be called in a different thread.
20 - slightly clarify rouse_cb/rouse_wait documentation.
21 - slightly improve rouse callback performance and memory usage.
22 - try to report correct calller when throwing exception in
23 on $condvar->recv, not a useless line in Coro::AnyEvent.
24
256.56 Thu Feb 20 19:40:07 CET 2020
26 - Coro::EV::time*once functions used ev_once, which is not cancelable.
27 This could cause problems due to I/O watchers not being stopped up
28 on exceptions/cancellation.
29
306.55 Fri Jul 19 17:28:09 CEST 2019
31 - Coro::RWLock woke up only one reader instead of all after a wrlock
32 (analyzed and patch by Maxim Skorobogatov).
33 - slightly improved documentation of Coro::RWLock and RWLock testcase.
34 - quickly (re)-ported to minix 3.3 before minix crashed again.
35
366.54 Sun Dec 16 11:06:07 CET 2018
37 - reinstate Coro::Select::select prototype - it breaks more than it fixes.
38
396.53 Sun Dec 16 10:33:15 CET 2018
40 - the autogenerated prototypes of Coro::AIO now add ";" at the end
41 to avoid generating named unary operators and messing up parsing
42 (mostly affects aio_mlockall 4 | 8 which was parsed as
43 (aio_mlockall 4) | 8).
44 - do not use a prototype for select in Coro::Select, as CORE::select also
45 doesn't have one.
46
476.52 Tue Aug 14 18:49:43 CEST 2018
48 - libcoro update: allow sharing of coro threads using the pthreads
49 backend among different pthreads, as is required by Coro::Multicore.
50 - always allow ->call and ->eval when the coro is the current coroutine. this
51 makes it possible, among other things, to call these on $Coro::main.
52 - minor code cleanups.
53
546.514 Thu Aug 31 18:25:31 CEST 2017
55 - safe_cancel no longer croaks when called on already-destroyed
56 thread, but instead "just works".
57
586.513 Sat Jul 15 01:18:51 CEST 2017
59 - experimental 5.22 comaptibility changes (PadlistNAMES).
60 - removed attributes example and made sure the other examples
61 still work.
62 - fix some statevar types for newer perls, should be of no consequence.
63
646.512 Fri Jul 14 05:03:17 CEST 2017
65 - seems to pass test suite down till 5.8.9, so make this the
66 minimum requirement in Makefile.PL.
67 - experimental 5.24 compatibility changes (SUB_ARGARRAY).
68 - experimental 5.26 compatibility changes (PL_comppad -> PADOFFSET).
69 - disable FORTIFY_SOURCE to avoid broken (but well-intentioned)
70 fortified longjmp variants in some libcs.
71 - re-try JIT allocation without PROT_EXEC, to hopefully improve
72 portability (still doesn't seem to work around broken selinux
73 mmap everywhere).
74 - speling fixes (Xavier Guimard).
75
766.511 Sun Jun 26 23:44:50 CEST 2016
77 - make it compile with threaded perls again.
78 - simplify and speed up __DIE__ and __WARN__ handling.
79
806.51 Sat Jun 25 20:57:20 CEST 2016
81 - 6.5 release didn't compile on most older perls due to
82 a logic error enabling experimental 5.24 code for all versions.
83 - since various broken or deficient patches float around
84 to work around the vtbl API breakage in perl, let's provide
85 our own workaround, which seems to be much less invasive
86 and more compatible than the approaches seen so far. YMMV.
87 - work around assertion failure in perl_destruct on 5.24
88 (analyzed by Dave Mitchell).
89
906.5 Wed Jun 22 22:23:50 CEST 2016
91 - swap_sv swapping order was not symmetrical, causing
92 wrong swaps when swap_sv was used multiple times on the
93 same sv in the same thread.
94 - swap_sv calls can now be undone by calling it again
95 with the same variables.
96 - swap_sv calls will now be undone in async_pool threads.
97 - split Coro::Semaphore::up/adjust into separate xs functions
98 for better error reporting, at slight codesize increase.
99 - (libcoro) arm assembly support, please test and report.
100 - adjust to PL_savestack changes in perl 5.24 (adapted from
101 the debian patch, which unfortunately gets it wrong).
102
1036.49 Sat Oct 17 01:40:12 CEST 2015
104 - throwing an exception to a thread waiting in
105 Coro::Handle using Coro::EV did not stop the watchers,
106 causing the next call to fail (testcase by Martin Pritchard).
107 - bump minimum perl version to 5.10.
108
1096.48 Sun Oct 4 19:03:51 CEST 2015
110 - fix memory leak when Coro::Handle uses Coro::EV internally
111 (testcase by Sten Sten).
112 - update code to libev 4 API (internally, Coro still used the
113 version 3 API calls).
114 - remove unused hv_sig.
115
1166.47 Sat Jul 11 03:58:20 CEST 2015
117 - swap_svs - NVs can be in the SV head beginning with 5.20.
118
1196.46 Tue Jun 30 14:36:30 CEST 2015
120 - restore portability to perls with windows fork emulation
121 (patch by Petr Písař).
122
1236.45 Tue Jun 30 01:40:08 CEST 2015
124 - restore portability to perl 5.16 and below.
125
1266.44 Tue Jun 30 00:41:54 CEST 2015
127 - avoid segfaulting (or worse) when tracing a canceled thread.
128 - the Event module silently broke it's public hook API, causing
129 failures on perls compiled with -Duselongdouble. Requiring
130 the latest version of Event because I don't know which release
131 changed this (it's not mentioned in the ChangeLog).
132 - allow xs level enterleave hooks via CoroAPI.
133
1346.43 Thu Jun 4 15:38:14 CEST 2015
135 - use stability canary.
136 - port to stableperl-5.22.0-1.001.
137 - update libecb.
138
1396.42 Wed Feb 11 20:29:52 CET 2015
140 - Coro::SemaphoreSet->try did not actually work (analyzed by
141 SATO Kentaro).
142 - upgrade libecb to be C11 compliant.
143
1446.41 Sat Sep 6 22:08:46 CEST 2014
145 - restore portability to perl 5.8.x.
146 - give new Coro's a valid GvHV(PL_hintgv) - this is slow and takes
147 up some memory, but fixes "use feature" and similar modules
148 when used inside a Coro.
149 - allow zero as argument to Coro::Channel to mean the same thing
150 as no argument. this works with older versions as well,
151 but wasn't legal until now.
152 - slightly better c header file detection.
153
1546.39 Mon Jun 2 00:00:08 CEST 2014
155 - work around more incompatible changes in 5.20.
156
1576.38 Sun Jun 1 21:54:23 CEST 2014
158 - check that perl slots actually have enough space to hold
159 interpreter variables.
160 - untested port to perl 5.19 (64 bit tmps indices) (reported
161 by Andreas König).
162 - croak when cancel is called without a thread context.
163
1646.37 Tue Mar 4 13:27:33 CET 2014
165 - *sigh*, removed leftover debugging code from debugging a
166 perl bug, of all things.
167
1686.36 Tue Mar 4 07:11:59 CET 2014
169 - semaphores would not clear the destroy hook when interrupted
170 by ->throw, causing segfaults or worse.
171 - ->throw on a thread waiting for a semaphore did not acquire
172 the semaphore, but also didn't wake up other waiters,
173 possibly causing a deadlock.
174 - "FATAL: $Coro::IDLE blocked itself" will now use Carp::confess
175 to report a full stacktrace, which should help find out
176 where the actual call is.
177 - "atomically" destroy data in slf_destroy, because it is
178 the right thing to do, just in case.
179 - disable ecb.h libm dependency, to compile on ancient systems
180 or under adverse conditions.
181
1826.33 Mon Nov 18 11:26:27 CET 2013
183 - do not crash when freeing padlists with holes (in 5.18).
184 - tentative SVt_BIND 5.19 port/fix.
185
1866.32 Tue Nov 5 15:35:35 CET 2013
187 - use a new algorithm to derive padlists for perl 5.18. The old
188 one could lead to 0-pointer accesses inside perl (reported
189 by Darin McBride).
190
1916.31 Thu May 9 07:39:48 CEST 2013
192 - Coro::AIO requests would crash if the thread was ready'd
193 while the request was ongoing.
6 194
76.29 Wed May 8 02:55:18 CEST 2013 1956.29 Wed May 8 02:55:18 CEST 2013
8 - when an on_destroy handler destructs the coro currently being 196 - when an on_destroy handler destructs the coro currently being
9 destructed a perl scalar could be accessed after being freed, 197 destructed a perl scalar could be accessed after being freed,
10 likely causing a crash. 198 likely causing a crash.
… …
97 handle and barfed. now it's just a bit slower on gentoo and similar 285 handle and barfed. now it's just a bit slower on gentoo and similar
98 systems. 286 systems.
99 287
1006.01 Sun Jul 3 12:31:14 CEST 2011 2886.01 Sun Jul 3 12:31:14 CEST 2011
101 - workarounds are good, but the test for whether pthreads are used 289 - workarounds are good, but the test for whether pthreads are used
102 was not good. thisone should be better. 290 was not good. this one should be better.
103 - check differently whether gcc generates cfi instructions itself. 291 - check differently whether gcc generates cfi instructions itself.
104 292
1056.0 Wed Jun 29 19:43:35 CEST 2011 2936.0 Wed Jun 29 19:43:35 CEST 2011
106 - INCOMPATIBLE CHANGE: unreferenced coro objects will now be 294 - INCOMPATIBLE CHANGE: unreferenced coro objects will now be
107 destroyed and cleaned up automatically (e.g. async { schedule }). 295 destroyed and cleaned up automatically (e.g. async { schedule }).
108 - implement a JIT compiler for part of the thread switch code, 296 - implement a JIT compiler for part of the thread switch code,
109 which gives a 50% speed improvement on threaded perls, and 297 which gives a 50% speed improvement on threaded perls, and
110 about 4% on non-thraeded perls (so threaded perls now finally 298 about 4% on non-threaded perls (so threaded perls now finally
111 reach about half the speed of non-threaded perls). 299 reach about half the speed of non-threaded perls).
112 - slightly modernise Coro::Intro, add section about rouse functions. 300 - slightly modernise Coro::Intro, add section about rouse functions.
113 - avoid DEFSV and ERRSV, giving another 10% improvement 301 - avoid DEFSV and ERRSV, giving another 10% improvement
114 in thread switching. 302 in thread switching.
115 - Coro::State->is_destroyed is now called is_zombie. 303 - Coro::State->is_destroyed is now called is_zombie.
… …
469 - ->throw is now supported on Coro::State objects. 657 - ->throw is now supported on Coro::State objects.
470 - clean up cctx creation code a bit. 658 - clean up cctx creation code a bit.
471 - entersub is actually a UNOP, not a LOGOP (not a bugfix). 659 - entersub is actually a UNOP, not a LOGOP (not a bugfix).
472 660
4734.9 Sat Nov 8 17:45:27 CET 2008 6614.9 Sat Nov 8 17:45:27 CET 2008
474 - (libcoro) did not preserve rbp with CORO_ASM (we are getting there). 662 - (libcoro) did not preserve rbp with CORO_ASM (we are getting there).
475 - (libcoro) no longer leak threads in the experimental pthread backend, 663 - (libcoro) no longer leak threads in the experimental pthread backend,
476 also speed it up considerably. 664 also speed it up considerably.
477 - (libcoro) do not rely on makecontext passing void *'s unscathed. 665 - (libcoro) do not rely on makecontext passing void *'s unscathed.
478 - fix compiletime dependencies on libcoro in the Makefile. 666 - fix compiletime dependencies on libcoro in the Makefile.
479 - cctx_count wasn't always updated properly. 667 - cctx_count wasn't always updated properly.
… …
541 - remove debugging code related to MgPV_nolen_const, also try to 729 - remove debugging code related to MgPV_nolen_const, also try to
542 make it compile with perl 5.8.6 (yes, apple apparently loves 730 make it compile with perl 5.8.6 (yes, apple apparently loves
543 outdated software). Reported by John S. 731 outdated software). Reported by John S.
544 732
5454.744 Tue Jul 8 22:06:35 CEST 2008 7334.744 Tue Jul 8 22:06:35 CEST 2008
546 - correctly provide default DIE/WARN handlers as documented. 734 - correctly provide default DIE/WARN handlers as documented.
547 - also overwrite PL_vtbl_sigelem.svt_clear, even though current 735 - also overwrite PL_vtbl_sigelem.svt_clear, even though current
548 implementations inside perl work fine for us. 736 implementations inside perl work fine for us.
549 737
5504.743 Mon Jun 16 00:21:57 CEST 2008 7384.743 Mon Jun 16 00:21:57 CEST 2008
551 - when using Coro::EV without running EV::loop it could 739 - when using Coro::EV without running EV::loop it could
… …
573 - sprinkle "no warnings" freely over everything, also suppress 761 - sprinkle "no warnings" freely over everything, also suppress
574 warnings for some other modules. 762 warnings for some other modules.
575 - fix typo in WSAEWOULDBLOCK. 763 - fix typo in WSAEWOULDBLOCK.
576 764
5774.72 Sun May 25 05:14:36 CEST 2008 7654.72 Sun May 25 05:14:36 CEST 2008
578 - tweak META.yaml a bit, unfortunately, there is no documented way 766 - tweak META.yaml a bit, unfortunately, there is no documented way
579 to have optional dependencies with CPAN. doh :( 767 to have optional dependencies with CPAN. doh :(
580 - avoid running some tests on windows because they would fail due to 768 - avoid running some tests on windows because they would fail due to
581 perl bug (broken fork, broken pipes...). 769 perl bug (broken fork, broken pipes...).
582 - work around perl on windows bugs where perl returns undocumented 770 - work around perl on windows bugs where perl returns undocumented
583 error codes for sysread, syswrite etc. by taking advantage 771 error codes for sysread, syswrite etc. by taking advantage
… …
587 - use unix domain sockets in testsuite to work around 775 - use unix domain sockets in testsuite to work around
588 common perl implementation bugs on widows (they are emulated by 776 common perl implementation bugs on widows (they are emulated by
589 tcp sockets on windows. ugh.) 777 tcp sockets on windows. ugh.)
590 778
5914.71 Sat May 24 20:01:27 CEST 2008 7794.71 Sat May 24 20:01:27 CEST 2008
592 - fix a bug in Coro::AnyEvent ("Usage: Coro::AnyEvent::_schedule()"). 780 - fix a bug in Coro::AnyEvent ("Usage: Coro::AnyEvent::_schedule()").
593 - take advantage of async name resolution of AnyEvent::Util. 781 - take advantage of async name resolution of AnyEvent::Util.
594 - work around brutal inet_aton override in Coro::LWP. 782 - work around brutal inet_aton override in Coro::LWP.
595 - take advantage of the readyhook in Coro::EV, for smoother 783 - take advantage of the readyhook in Coro::EV, for smoother
596 scheduling. 784 scheduling.
597 785
5984.7 Sun May 11 00:32:19 CEST 2008 7864.7 Sun May 11 00:32:19 CEST 2008
599 - completely reworked the Coro manpage. 787 - completely reworked the Coro manpage.
600 - added Coro::AnyEvent, generic event loop integration. 788 - added Coro::AnyEvent, generic event loop integration.
601 - implement cancel, ready and kill commands in Coro::Debug. 789 - implement cancel, ready and kill commands in Coro::Debug.
602 - document find_coro in Coro::Debug. 790 - document find_coro in Coro::Debug.
603 - incompatible change: rename has_stack to has_cctx. 791 - incompatible change: rename has_stack to has_cctx.
604 - Coro::AIO and Coro::BDB no longer force event model detection, 792 - Coro::AIO and Coro::BDB no longer force event model detection,
605 use AnyEvent::AIO and AnyEvent::BDB. 793 use AnyEvent::AIO and AnyEvent::BDB.
… …
729 917
7304.11 Thu Oct 11 02:40:24 CEST 2007 9184.11 Thu Oct 11 02:40:24 CEST 2007
731 - port to threaded perls. 919 - port to threaded perls.
732 920
7334.1 Thu Oct 11 02:38:16 CEST 2007 9214.1 Thu Oct 11 02:38:16 CEST 2007
734 - incompatible change: $SIG{__DIE__} and $SIG{__WARN__} will now 922 - incompatible change: $SIG{__DIE__} and $SIG{__WARN__} will now
735 be local to each coro (see Coro::State). 923 be local to each coro (see Coro::State).
736 - incompatible change: for very deep reasons, cede and cede_notself 924 - incompatible change: for very deep reasons, cede and cede_notself
737 cannot return anything, so nothing will be returned. 925 cannot return anything, so nothing will be returned.
738 - possibly bring back 5.10 compatibility (untested). 926 - possibly bring back 5.10 compatibility (untested).
739 - work around stupid (and wrong) warning on 5.10 :(. 927 - work around stupid (and wrong) warning on 5.10 :(.
… …
811 but stateful semantics. 999 but stateful semantics.
812 - fixed a lot of typos in Coro.pm (patch submitted by David 1000 - fixed a lot of typos in Coro.pm (patch submitted by David
813 Steinbrunner, which applied flawlessly). 1001 Steinbrunner, which applied flawlessly).
814 1002
8153.6 Sat Apr 14 17:13:31 CEST 2007 10033.6 Sat Apr 14 17:13:31 CEST 2007
816 - added some bugfixes to get eg/myhttpd working again. 1004 - added some bugfixes to get eg/myhttpd working again.
817 - added Coro::Storable for often-cede'ing freeze/thaw. 1005 - added Coro::Storable for often-cede'ing freeze/thaw.
818 - try to do a clean exit when a coroutine calls exit 1006 - try to do a clean exit when a coroutine calls exit
819 (EXPERIMENTAL). 1007 (EXPERIMENTAL).
820 - got rid of indirect call through _coro_init. 1008 - got rid of indirect call through _coro_init.
821 - updated the partly antique examples in eg/ to 1009 - updated the partly antique examples in eg/ to

Diff Legend

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