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

Comparing Coro/Changes (file contents):
Revision 1.549 by root, Sat Feb 19 04:34:58 2011 UTC vs.
Revision 1.656 by root, Fri Jun 17 04:59:27 2016 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: unready_all
5TODO: myhttpd header parsing
6TODO: channel->maxsize(newsize)? 4TODO: channel->maxsize(newsize)?
7TODO: 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?
8 8
96.5
10 - swap_sv swapping order was not symmetrical, causing
11 wrong swaps when swap_sv was used multiple times on the
12 same sv in the same thread.
13 - swap_sv calls can now be undone by calling it again
14 with the same variables.
15 - swap_sv calls will now be undone in async_pool threads.
16 - split Coro::Semaphore::up/adjust into separate xs functions
17 for better error reporting, at slight codesize increase.
18 - (libcoro) arm assembly support, please test and report.
19
206.49 Sat Oct 17 01:40:12 CEST 2015
21 - throwing an exception to a thread waiting in
22 Coro::Handle using Coro::EV did not stop the watchers,
23 causing the next call to fail (testcase by Martin Pritchard).
24 - bump minimum perl version to 5.10.
25
266.48 Sun Oct 4 19:03:51 CEST 2015
27 - fix memory leak when Coro::Handle uses Coro::EV internally
28 (testcase by Sten Sten).
29 - update code to libev 4 API (internally, Coro still used the
30 version 3 API calls).
31 - remove unused hv_sig.
32
336.47 Sat Jul 11 03:58:20 CEST 2015
34 - swap_svs - NVs can be in the SV head beginning with 5.20.
35
366.46 Tue Jun 30 14:36:30 CEST 2015
37 - restore portability to perls with windows fork emulation
38 (patch by Petr Písař).
39
406.45 Tue Jun 30 01:40:08 CEST 2015
41 - restore portability to perl 5.16 and below.
42
436.44 Tue Jun 30 00:41:54 CEST 2015
44 - avoid segfaulting (or worse) when tracing a canceled thread.
45 - the Event module silently broke it's public hook API, causing
46 failures on perls compiled with -Duselongdouble. Requiring
47 the latest version of Event because I don't know which release
48 changed this (it's not mentioned in the ChangeLog).
49 - allow xs level enterleave hooks via CoroAPI.
50
516.43 Thu Jun 4 15:38:14 CEST 2015
52 - use stability canary.
53 - port to stableperl-5.22.0-1.001.
54 - update libecb.
55
566.42 Wed Feb 11 20:29:52 CET 2015
57 - Coro::SemaphoreSet->try did not actually work (analyzed by
58 SATO Kentaro).
59 - upgrade libecb to be C11 compliant.
60
616.41 Sat Sep 6 22:08:46 CEST 2014
62 - restore portability to perl 5.8.x.
63 - give new Coro's a valid GvHV(PL_hintgv) - this is slow and takes
64 up some memory, but fixes "use feature" and similar modules
65 when used inside a Coro.
66 - allow zero as argument to Coro::Channel to mean the same thing
67 as no argument. this works with older versions as well,
68 but wasn't legal until now.
69 - slightly better c header file detection.
70
716.39 Mon Jun 2 00:00:08 CEST 2014
72 - work around more incompatible changes in 5.20.
73
746.38 Sun Jun 1 21:54:23 CEST 2014
75 - check that perl slots actually have enough space to hold
76 interpreter variables.
77 - untested port to perl 5.19 (64 bit tmps indices) (reported
78 by Andreas König).
79 - croak when cancel is called without a thread context.
80
816.37 Tue Mar 4 13:27:33 CET 2014
82 - *sigh*, removed leftover debugging code from debugging a
83 perl bug, of all things.
84
856.36 Tue Mar 4 07:11:59 CET 2014
86 - semaphores would not clear the destroy hook when interrupted
87 by ->throw, causing segfaults or worse.
88 - ->throw on a thread waiting for a semaphore did not acquire
89 the semaphore, but also didn't wake up other waiters,
90 possibly causing a deadlock.
91 - "FATAL: $Coro::IDLE blocked itself" will now use Carp::confess
92 to report a full stacktrace, which should help find out
93 where the actual call is.
94 - "atomically" destroy data in slf_destroy, because it is
95 the right thing to do, just in case.
96 - disable ecb.h libm dependency, to compile on ancient systems
97 or under adverse conditions.
98
996.33 Mon Nov 18 11:26:27 CET 2013
100 - do not crash when freeing padlists with holes (in 5.18).
101 - tentative SVt_BIND 5.19 port/fix.
102
1036.32 Tue Nov 5 15:35:35 CET 2013
104 - use a new algorithm to derive padlists for perl 5.18. The old
105 one could lead to 0-pointer accesses inside perl (reported
106 by Darin McBride).
107
1086.31 Thu May 9 07:39:48 CEST 2013
109 - Coro::AIO requests would crash if the thread was ready'd
110 while the request was ongoing.
111
1126.29 Wed May 8 02:55:18 CEST 2013
113 - when an on_destroy handler destructs the coro currently being
114 destructed a perl scalar could be accessed after being freed,
115 likely causing a crash.
116
1176.28 Wed Mar 6 06:58:02 CET 2013
118 - clean remnants of existing __DIE__ and __WARN__ handlers so
119 they lose their magic and will not cause segfaults later
120 (testcase by Andrey Sagulin).
121 - improved Coro::State documentation a bit.
122 - Coro::Debug::command now flushes the output.
123 - add hack detection code for x32 abi, because the braindead slugs
124 who designed that made it look exactly like x86_64 without
125 providing proper compile time symbols to test for it. as a result,
126 this detection cannot work reliably.
127 - valgrind stack registering was broken.
128 - do not rely on Time::HiRes anymore in Coro::Debug.
129
1306.23 Fri Dec 7 23:36:37 CET 2012
131 - use experimental fiber implementation on native windows
132 perls.
133 - use sizeof (void *) as multiplication factor for stack sizes,
134 to accomodate the totally braindamaged microsoft 64 bit "os".
135 - changed verifier host from win2k-ap510-32 to win7-sp516-32/64.
136 activeperl 5.16 crashes when PerlIO_define_layer is called due
137 to some bug in the perl dll, strawberry perl at least passes
138 the testsuite.
139 - implement Coro::Handle->peeraddr/host/port, for slightly
140 improved compatibility with LWP.
141 - implement 5.17 compatibility by almost blindly applying a
142 good-looking patch by Father Chrysostomos.
143 - move stack management functions into libcoro 3.
144 - libcoro version 3 "released".
145 - support magic values as timed_io_once args.
146 - recommend AnyEvent 7+ or EV 4+, also require EV
147 version 4 or newer for Coro::EV.
148
1496.10 Tue Oct 9 01:14:27 CEST 2012
150 - updated ecb.h, it had a typo that caused it to not compile on many
151 big endian systems (reported by many people).
152 - disable memory fences in ecb.h to improve portability.
153
1546.09 Sat Oct 6 23:25:02 CEST 2012
155 - Coro::EV I/O watchers were not interruptible by exceptions
156 (Coro::State::throw) (testcase by sten).
157 - ->throw now puts threads into the ready queue, as this seems to
158 be expected by existing code, and code that doesn't cope with spurious
159 wakeups needs fixing anyway.
160 - use fd -1 in mmap.
161 - cast I32 to int in error message printf.
162 - warn about broken so-called "hardened" kernels.
163
1646.08 Fri Apr 13 12:05:47 CEST 2012
165 - be more aggressive about exiting like perl does - formerly,
166 exiting from the non-main thread would not execute END blocks.
167
1686.07 Fri Nov 11 21:21:48 CET 2011
169 - work around a bug in PerlIO (setting $SIG{__WARN__} to a PVCV).
170 - update ecb.h.
171
1726.06 Mon Aug 8 23:59:48 CEST 2011
173 - cygwin unfortunately patches the stack at runtime, so we use the pthreads
174 backend, which is an order of magnitude slower. unfortunately, cygwins
175 pthread implementation isn't very complete either, so allocate the stack
176 twice just to be sure.
177 (note: cygwin also enables mymalloc, which is NOT THREADSAFE ON WINDOWS,
178 in its ithreaded perl - best recompile cygwin and use the 'w'indows
179 backend for much better performance. also disable ithreads for
180 even better performance...).
181
1826.05 Thu Aug 4 21:36:36 CEST 2011
183 - blush, condvar values would not be propagated from send to recv anymore
184 (reported by Chip Salzenberg).
185 - use exponential increase for the readline buffer length in
186 Coro::Handle. also reduce initial allocation to 1020 from 4096 bytes.
187
1886.04 Wed Aug 3 17:15:45 CEST 2011
189 - use even more efficient and more compatible condvars for
190 compatibility to AnyEvent 6.x :)
191 - more inconsequential ecb.h updates.
192
1936.03 Wed Aug 3 11:41:30 CEST 2011
194 - change how Coro patches AnyEvent condvars for compatibility to
195 AnyEvent 6.x.
196 - update ecb.h, to no longer include <pthread.h> in case WinNT.h
197 hasn't been included.
198
1996.02 Wed Jul 13 04:35:19 CEST 2011
200 - "improve portability to Gentoo" - gentoo manages to put perl variables
201 in memory areas that are farther than 2gb apart, which the jit couldn't
202 handle and barfed. now it's just a bit slower on gentoo and similar
203 systems.
204
2056.01 Sun Jul 3 12:31:14 CEST 2011
206 - workarounds are good, but the test for whether pthreads are used
207 was not good. this one should be better.
208 - check differently whether gcc generates cfi instructions itself.
209
2106.0 Wed Jun 29 19:43:35 CEST 2011
211 - INCOMPATIBLE CHANGE: unreferenced coro objects will now be
212 destroyed and cleaned up automatically (e.g. async { schedule }).
213 - implement a JIT compiler for part of the thread switch code,
214 which gives a 50% speed improvement on threaded perls, and
215 about 4% on non-threaded perls (so threaded perls now finally
216 reach about half the speed of non-threaded perls).
217 - slightly modernise Coro::Intro, add section about rouse functions.
218 - avoid DEFSV and ERRSV, giving another 10% improvement
219 in thread switching.
220 - Coro::State->is_destroyed is now called is_zombie.
221 - implement a Coro->safe_cancel method that might fail, but
222 cancels in a "safer" way if it succeeds.
223 - add preliminary support for DEBUGGING perls.
224 - get rid of two hash-accesses when initialising a new Coro - this
225 speeds up coro creation by almost a factor of two.
226 - croak when a coro that is being cancelled tries to block
227 (e.g. while executing a guard block), instead of crashing or
228 deadlocking.
229 - use a more robust and also faster method to identify Coro::State
230 objects - speeds up everything a bit.
231 - implement Coro->cancel in XS for a 20% speed improvement, and to
232 be able to implement mutual cancellation.
233 - speed up context switches by a percent or two by more efficiently
234 allocating context stack entries.
235 - implement Coro->join and Coro->on_destroy in XS for a speedup and
236 a reduction in memory use.
237 - cancelling a coro while it itself is cancelling another coro is
238 now supported and working, instead of triggering an assertion.
239 - be a bit more crash-resistant when calling (buggy) on_destroy
240 callbacks (best effort).
241 - move on_destroy into the slf_frame, to allow extension slf
242 functions to have destructors.
243 - get rid if coro refcounting - simply crash in other interpreter
244 threads by nulling the pointers on clone.
245 - simplify warn/die hook handling when loading Coro - the convoluted
246 logic seems to be no longer neccessary.
247 - use libecb instead of our own home-grown gcc hacks.
248 - document alternatives to Coro::LWP. Please use them :)
249 - work around another mindless idiotic NEEDLESS bug in openbsd/mirbsds
250 sigaltstack. Really. wine suffers from it, erlang suffers from it,
251 and it's known since at least 2006.
252
2535.372 Wed Feb 23 06:14:30 CET 2011
254 - apparently mingw doesn't provide a working gettimeofday, try to
255 work around that by relying on Time::HiRes (indirectly brought to
256 my attention by Max Maischein).
257 - fix some portability issues when Time::HiRes was used.
258
2595.371 Mon Feb 21 14:36:08 CET 2011
260 - backport to windows process emulation code again.
261
2625.37 Sat Feb 19 07:49:44 CET 2011
9 - add a big "Coro thread life cycle" section to "man Coro". 263 - add a big "Coro thread life cycle" section to "man Coro".
10 - try a tentative workaround against the breakage that 5.13 has 264 - try a tentative workaround against the breakage that 5.13 has
11 introduced without depreciation period. sigh. 265 introduced without depreciation period. sigh.
266 - no longer use Time::HiRes if gettimeofday is available, which
267 saves quite a lot of memory.
12 268
135.36 Sun Feb 13 05:33:41 CET 2011 2695.36 Sun Feb 13 05:33:41 CET 2011
14 - automatically load Coro::Channel, Coro::RWLock, Coro::Semaphore, 270 - automatically load Coro::Channel, Coro::RWLock, Coro::Semaphore,
15 Coro::SemaphoreSet, Coro::Signal and Coro::Specific on first "new" 271 Coro::SemaphoreSet, Coro::Signal and Coro::Specific on first "new"
16 method call. 272 method call.
… …
318 - ->throw is now supported on Coro::State objects. 574 - ->throw is now supported on Coro::State objects.
319 - clean up cctx creation code a bit. 575 - clean up cctx creation code a bit.
320 - entersub is actually a UNOP, not a LOGOP (not a bugfix). 576 - entersub is actually a UNOP, not a LOGOP (not a bugfix).
321 577
3224.9 Sat Nov 8 17:45:27 CET 2008 5784.9 Sat Nov 8 17:45:27 CET 2008
323 - (libcoro) did not preserve rbp with CORO_ASM (we are getting there). 579 - (libcoro) did not preserve rbp with CORO_ASM (we are getting there).
324 - (libcoro) no longer leak threads in the experimental pthread backend, 580 - (libcoro) no longer leak threads in the experimental pthread backend,
325 also speed it up considerably. 581 also speed it up considerably.
326 - (libcoro) do not rely on makecontext passing void *'s unscathed. 582 - (libcoro) do not rely on makecontext passing void *'s unscathed.
327 - fix compiletime dependencies on libcoro in the Makefile. 583 - fix compiletime dependencies on libcoro in the Makefile.
328 - cctx_count wasn't always updated properly. 584 - cctx_count wasn't always updated properly.
… …
390 - remove debugging code related to MgPV_nolen_const, also try to 646 - remove debugging code related to MgPV_nolen_const, also try to
391 make it compile with perl 5.8.6 (yes, apple apparently loves 647 make it compile with perl 5.8.6 (yes, apple apparently loves
392 outdated software). Reported by John S. 648 outdated software). Reported by John S.
393 649
3944.744 Tue Jul 8 22:06:35 CEST 2008 6504.744 Tue Jul 8 22:06:35 CEST 2008
395 - correctly provide default DIE/WARN handlers as documented. 651 - correctly provide default DIE/WARN handlers as documented.
396 - also overwrite PL_vtbl_sigelem.svt_clear, even though current 652 - also overwrite PL_vtbl_sigelem.svt_clear, even though current
397 implementations inside perl work fine for us. 653 implementations inside perl work fine for us.
398 654
3994.743 Mon Jun 16 00:21:57 CEST 2008 6554.743 Mon Jun 16 00:21:57 CEST 2008
400 - when using Coro::EV without running EV::loop it could 656 - when using Coro::EV without running EV::loop it could
… …
422 - sprinkle "no warnings" freely over everything, also suppress 678 - sprinkle "no warnings" freely over everything, also suppress
423 warnings for some other modules. 679 warnings for some other modules.
424 - fix typo in WSAEWOULDBLOCK. 680 - fix typo in WSAEWOULDBLOCK.
425 681
4264.72 Sun May 25 05:14:36 CEST 2008 6824.72 Sun May 25 05:14:36 CEST 2008
427 - tweak META.yaml a bit, unfortunately, there is no documented way 683 - tweak META.yaml a bit, unfortunately, there is no documented way
428 to have optional dependencies with CPAN. doh :( 684 to have optional dependencies with CPAN. doh :(
429 - avoid running some tests on windows because they would fail due to 685 - avoid running some tests on windows because they would fail due to
430 perl bug (broken fork, broken pipes...). 686 perl bug (broken fork, broken pipes...).
431 - work around perl on windows bugs where perl returns undocumented 687 - work around perl on windows bugs where perl returns undocumented
432 error codes for sysread, syswrite etc. by taking advantage 688 error codes for sysread, syswrite etc. by taking advantage
… …
436 - use unix domain sockets in testsuite to work around 692 - use unix domain sockets in testsuite to work around
437 common perl implementation bugs on widows (they are emulated by 693 common perl implementation bugs on widows (they are emulated by
438 tcp sockets on windows. ugh.) 694 tcp sockets on windows. ugh.)
439 695
4404.71 Sat May 24 20:01:27 CEST 2008 6964.71 Sat May 24 20:01:27 CEST 2008
441 - fix a bug in Coro::AnyEvent ("Usage: Coro::AnyEvent::_schedule()"). 697 - fix a bug in Coro::AnyEvent ("Usage: Coro::AnyEvent::_schedule()").
442 - take advantage of async name resolution of AnyEvent::Util. 698 - take advantage of async name resolution of AnyEvent::Util.
443 - work around brutal inet_aton override in Coro::LWP. 699 - work around brutal inet_aton override in Coro::LWP.
444 - take advantage of the readyhook in Coro::EV, for smoother 700 - take advantage of the readyhook in Coro::EV, for smoother
445 scheduling. 701 scheduling.
446 702
4474.7 Sun May 11 00:32:19 CEST 2008 7034.7 Sun May 11 00:32:19 CEST 2008
448 - completely reworked the Coro manpage. 704 - completely reworked the Coro manpage.
449 - added Coro::AnyEvent, generic event loop integration. 705 - added Coro::AnyEvent, generic event loop integration.
450 - implement cancel, ready and kill commands in Coro::Debug. 706 - implement cancel, ready and kill commands in Coro::Debug.
451 - document find_coro in Coro::Debug. 707 - document find_coro in Coro::Debug.
452 - incompatible change: rename has_stack to has_cctx. 708 - incompatible change: rename has_stack to has_cctx.
453 - Coro::AIO and Coro::BDB no longer force event model detection, 709 - Coro::AIO and Coro::BDB no longer force event model detection,
454 use AnyEvent::AIO and AnyEvent::BDB. 710 use AnyEvent::AIO and AnyEvent::BDB.
… …
578 834
5794.11 Thu Oct 11 02:40:24 CEST 2007 8354.11 Thu Oct 11 02:40:24 CEST 2007
580 - port to threaded perls. 836 - port to threaded perls.
581 837
5824.1 Thu Oct 11 02:38:16 CEST 2007 8384.1 Thu Oct 11 02:38:16 CEST 2007
583 - incompatible change: $SIG{__DIE__} and $SIG{__WARN__} will now 839 - incompatible change: $SIG{__DIE__} and $SIG{__WARN__} will now
584 be local to each coro (see Coro::State). 840 be local to each coro (see Coro::State).
585 - incompatible change: for very deep reasons, cede and cede_notself 841 - incompatible change: for very deep reasons, cede and cede_notself
586 cannot return anything, so nothing will be returned. 842 cannot return anything, so nothing will be returned.
587 - possibly bring back 5.10 compatibility (untested). 843 - possibly bring back 5.10 compatibility (untested).
588 - work around stupid (and wrong) warning on 5.10 :(. 844 - work around stupid (and wrong) warning on 5.10 :(.
… …
660 but stateful semantics. 916 but stateful semantics.
661 - fixed a lot of typos in Coro.pm (patch submitted by David 917 - fixed a lot of typos in Coro.pm (patch submitted by David
662 Steinbrunner, which applied flawlessly). 918 Steinbrunner, which applied flawlessly).
663 919
6643.6 Sat Apr 14 17:13:31 CEST 2007 9203.6 Sat Apr 14 17:13:31 CEST 2007
665 - added some bugfixes to get eg/myhttpd working again. 921 - added some bugfixes to get eg/myhttpd working again.
666 - added Coro::Storable for often-cede'ing freeze/thaw. 922 - added Coro::Storable for often-cede'ing freeze/thaw.
667 - try to do a clean exit when a coroutine calls exit 923 - try to do a clean exit when a coroutine calls exit
668 (EXPERIMENTAL). 924 (EXPERIMENTAL).
669 - got rid of indirect call through _coro_init. 925 - got rid of indirect call through _coro_init.
670 - updated the partly antique examples in eg/ to 926 - updated the partly antique examples in eg/ to

Diff Legend

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