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

Comparing Coro/Changes (file contents):
Revision 1.540 by root, Tue Jul 27 14:42:34 2010 UTC vs.
Revision 1.669 by root, Fri Jul 14 23:09:51 2017 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
9 - removed attributres example and made sure the other examples
10 still work.
11
126.512 Fri Jul 14 05:03:17 CEST 2017
13 - seems to pass test suite down till 5.8.9, so make this the
14 minimum requirement in Makefile.PL.
15 - experimental 5.24 compatibility changes (SUB_ARGARRAY).
16 - experimental 5.26 compatibility changes (PL_comppad -> PADOFFSET).
17 - disable FORTIFY_SOURCE to avoid broken (but well-intentioned)
18 fortified longjmp variants in some libcs.
19 - re-try JIT allocation without PROT_EXEC, to hopefully improve
20 portability (still doesn't seem to work around broken selinux
21 mmap everywhere).
22 - speling fixes (Xavier Guimard).
23
246.511 Sun Jun 26 23:44:50 CEST 2016
25 - make it compile with threaded perls again.
26 - simplify and speed up __DIE__ and __WARN__ handling.
27
286.51 Sat Jun 25 20:57:20 CEST 2016
29 - 6.5 release didn't compile on most older perls due to
30 a logic error enabling experimental 5.24 code for all versions.
31 - since various broken or deficient patches float around
32 to work around the vtbl API breakage in perl, let's provide
33 our own workaround, which seems to be much less invasive
34 and more compatible than the approaches seen so far. YMMV.
35 - work around assertion failure in perl_destruct on 5.24
36 (analyzed by Dave Mitchell).
37
386.5 Wed Jun 22 22:23:50 CEST 2016
39 - swap_sv swapping order was not symmetrical, causing
40 wrong swaps when swap_sv was used multiple times on the
41 same sv in the same thread.
42 - swap_sv calls can now be undone by calling it again
43 with the same variables.
44 - swap_sv calls will now be undone in async_pool threads.
45 - split Coro::Semaphore::up/adjust into separate xs functions
46 for better error reporting, at slight codesize increase.
47 - (libcoro) arm assembly support, please test and report.
48 - adjust to PL_savestack changes in perl 5.24 (adapted from
49 the debian patch, which unfortunately gets it wrong).
50
516.49 Sat Oct 17 01:40:12 CEST 2015
52 - throwing an exception to a thread waiting in
53 Coro::Handle using Coro::EV did not stop the watchers,
54 causing the next call to fail (testcase by Martin Pritchard).
55 - bump minimum perl version to 5.10.
56
576.48 Sun Oct 4 19:03:51 CEST 2015
58 - fix memory leak when Coro::Handle uses Coro::EV internally
59 (testcase by Sten Sten).
60 - update code to libev 4 API (internally, Coro still used the
61 version 3 API calls).
62 - remove unused hv_sig.
63
646.47 Sat Jul 11 03:58:20 CEST 2015
65 - swap_svs - NVs can be in the SV head beginning with 5.20.
66
676.46 Tue Jun 30 14:36:30 CEST 2015
68 - restore portability to perls with windows fork emulation
69 (patch by Petr Písař).
70
716.45 Tue Jun 30 01:40:08 CEST 2015
72 - restore portability to perl 5.16 and below.
73
746.44 Tue Jun 30 00:41:54 CEST 2015
75 - avoid segfaulting (or worse) when tracing a canceled thread.
76 - the Event module silently broke it's public hook API, causing
77 failures on perls compiled with -Duselongdouble. Requiring
78 the latest version of Event because I don't know which release
79 changed this (it's not mentioned in the ChangeLog).
80 - allow xs level enterleave hooks via CoroAPI.
81
826.43 Thu Jun 4 15:38:14 CEST 2015
83 - use stability canary.
84 - port to stableperl-5.22.0-1.001.
85 - update libecb.
86
876.42 Wed Feb 11 20:29:52 CET 2015
88 - Coro::SemaphoreSet->try did not actually work (analyzed by
89 SATO Kentaro).
90 - upgrade libecb to be C11 compliant.
91
926.41 Sat Sep 6 22:08:46 CEST 2014
93 - restore portability to perl 5.8.x.
94 - give new Coro's a valid GvHV(PL_hintgv) - this is slow and takes
95 up some memory, but fixes "use feature" and similar modules
96 when used inside a Coro.
97 - allow zero as argument to Coro::Channel to mean the same thing
98 as no argument. this works with older versions as well,
99 but wasn't legal until now.
100 - slightly better c header file detection.
101
1026.39 Mon Jun 2 00:00:08 CEST 2014
103 - work around more incompatible changes in 5.20.
104
1056.38 Sun Jun 1 21:54:23 CEST 2014
106 - check that perl slots actually have enough space to hold
107 interpreter variables.
108 - untested port to perl 5.19 (64 bit tmps indices) (reported
109 by Andreas König).
110 - croak when cancel is called without a thread context.
111
1126.37 Tue Mar 4 13:27:33 CET 2014
113 - *sigh*, removed leftover debugging code from debugging a
114 perl bug, of all things.
115
1166.36 Tue Mar 4 07:11:59 CET 2014
117 - semaphores would not clear the destroy hook when interrupted
118 by ->throw, causing segfaults or worse.
119 - ->throw on a thread waiting for a semaphore did not acquire
120 the semaphore, but also didn't wake up other waiters,
121 possibly causing a deadlock.
122 - "FATAL: $Coro::IDLE blocked itself" will now use Carp::confess
123 to report a full stacktrace, which should help find out
124 where the actual call is.
125 - "atomically" destroy data in slf_destroy, because it is
126 the right thing to do, just in case.
127 - disable ecb.h libm dependency, to compile on ancient systems
128 or under adverse conditions.
129
1306.33 Mon Nov 18 11:26:27 CET 2013
131 - do not crash when freeing padlists with holes (in 5.18).
132 - tentative SVt_BIND 5.19 port/fix.
133
1346.32 Tue Nov 5 15:35:35 CET 2013
135 - use a new algorithm to derive padlists for perl 5.18. The old
136 one could lead to 0-pointer accesses inside perl (reported
137 by Darin McBride).
138
1396.31 Thu May 9 07:39:48 CEST 2013
140 - Coro::AIO requests would crash if the thread was ready'd
141 while the request was ongoing.
142
1436.29 Wed May 8 02:55:18 CEST 2013
144 - when an on_destroy handler destructs the coro currently being
145 destructed a perl scalar could be accessed after being freed,
146 likely causing a crash.
147
1486.28 Wed Mar 6 06:58:02 CET 2013
149 - clean remnants of existing __DIE__ and __WARN__ handlers so
150 they lose their magic and will not cause segfaults later
151 (testcase by Andrey Sagulin).
152 - improved Coro::State documentation a bit.
153 - Coro::Debug::command now flushes the output.
154 - add hack detection code for x32 abi, because the braindead slugs
155 who designed that made it look exactly like x86_64 without
156 providing proper compile time symbols to test for it. as a result,
157 this detection cannot work reliably.
158 - valgrind stack registering was broken.
159 - do not rely on Time::HiRes anymore in Coro::Debug.
160
1616.23 Fri Dec 7 23:36:37 CET 2012
162 - use experimental fiber implementation on native windows
163 perls.
164 - use sizeof (void *) as multiplication factor for stack sizes,
165 to accomodate the totally braindamaged microsoft 64 bit "os".
166 - changed verifier host from win2k-ap510-32 to win7-sp516-32/64.
167 activeperl 5.16 crashes when PerlIO_define_layer is called due
168 to some bug in the perl dll, strawberry perl at least passes
169 the testsuite.
170 - implement Coro::Handle->peeraddr/host/port, for slightly
171 improved compatibility with LWP.
172 - implement 5.17 compatibility by almost blindly applying a
173 good-looking patch by Father Chrysostomos.
174 - move stack management functions into libcoro 3.
175 - libcoro version 3 "released".
176 - support magic values as timed_io_once args.
177 - recommend AnyEvent 7+ or EV 4+, also require EV
178 version 4 or newer for Coro::EV.
179
1806.10 Tue Oct 9 01:14:27 CEST 2012
181 - updated ecb.h, it had a typo that caused it to not compile on many
182 big endian systems (reported by many people).
183 - disable memory fences in ecb.h to improve portability.
184
1856.09 Sat Oct 6 23:25:02 CEST 2012
186 - Coro::EV I/O watchers were not interruptible by exceptions
187 (Coro::State::throw) (testcase by sten).
188 - ->throw now puts threads into the ready queue, as this seems to
189 be expected by existing code, and code that doesn't cope with spurious
190 wakeups needs fixing anyway.
191 - use fd -1 in mmap.
192 - cast I32 to int in error message printf.
193 - warn about broken so-called "hardened" kernels.
194
1956.08 Fri Apr 13 12:05:47 CEST 2012
196 - be more aggressive about exiting like perl does - formerly,
197 exiting from the non-main thread would not execute END blocks.
198
1996.07 Fri Nov 11 21:21:48 CET 2011
200 - work around a bug in PerlIO (setting $SIG{__WARN__} to a PVCV).
201 - update ecb.h.
202
2036.06 Mon Aug 8 23:59:48 CEST 2011
204 - cygwin unfortunately patches the stack at runtime, so we use the pthreads
205 backend, which is an order of magnitude slower. unfortunately, cygwins
206 pthread implementation isn't very complete either, so allocate the stack
207 twice just to be sure.
208 (note: cygwin also enables mymalloc, which is NOT THREADSAFE ON WINDOWS,
209 in its ithreaded perl - best recompile cygwin and use the 'w'indows
210 backend for much better performance. also disable ithreads for
211 even better performance...).
212
2136.05 Thu Aug 4 21:36:36 CEST 2011
214 - blush, condvar values would not be propagated from send to recv anymore
215 (reported by Chip Salzenberg).
216 - use exponential increase for the readline buffer length in
217 Coro::Handle. also reduce initial allocation to 1020 from 4096 bytes.
218
2196.04 Wed Aug 3 17:15:45 CEST 2011
220 - use even more efficient and more compatible condvars for
221 compatibility to AnyEvent 6.x :)
222 - more inconsequential ecb.h updates.
223
2246.03 Wed Aug 3 11:41:30 CEST 2011
225 - change how Coro patches AnyEvent condvars for compatibility to
226 AnyEvent 6.x.
227 - update ecb.h, to no longer include <pthread.h> in case WinNT.h
228 hasn't been included.
229
2306.02 Wed Jul 13 04:35:19 CEST 2011
231 - "improve portability to Gentoo" - gentoo manages to put perl variables
232 in memory areas that are farther than 2gb apart, which the jit couldn't
233 handle and barfed. now it's just a bit slower on gentoo and similar
234 systems.
235
2366.01 Sun Jul 3 12:31:14 CEST 2011
237 - workarounds are good, but the test for whether pthreads are used
238 was not good. this one should be better.
239 - check differently whether gcc generates cfi instructions itself.
240
2416.0 Wed Jun 29 19:43:35 CEST 2011
242 - INCOMPATIBLE CHANGE: unreferenced coro objects will now be
243 destroyed and cleaned up automatically (e.g. async { schedule }).
244 - implement a JIT compiler for part of the thread switch code,
245 which gives a 50% speed improvement on threaded perls, and
246 about 4% on non-threaded perls (so threaded perls now finally
247 reach about half the speed of non-threaded perls).
248 - slightly modernise Coro::Intro, add section about rouse functions.
249 - avoid DEFSV and ERRSV, giving another 10% improvement
250 in thread switching.
251 - Coro::State->is_destroyed is now called is_zombie.
252 - implement a Coro->safe_cancel method that might fail, but
253 cancels in a "safer" way if it succeeds.
254 - add preliminary support for DEBUGGING perls.
255 - get rid of two hash-accesses when initialising a new Coro - this
256 speeds up coro creation by almost a factor of two.
257 - croak when a coro that is being cancelled tries to block
258 (e.g. while executing a guard block), instead of crashing or
259 deadlocking.
260 - use a more robust and also faster method to identify Coro::State
261 objects - speeds up everything a bit.
262 - implement Coro->cancel in XS for a 20% speed improvement, and to
263 be able to implement mutual cancellation.
264 - speed up context switches by a percent or two by more efficiently
265 allocating context stack entries.
266 - implement Coro->join and Coro->on_destroy in XS for a speedup and
267 a reduction in memory use.
268 - cancelling a coro while it itself is cancelling another coro is
269 now supported and working, instead of triggering an assertion.
270 - be a bit more crash-resistant when calling (buggy) on_destroy
271 callbacks (best effort).
272 - move on_destroy into the slf_frame, to allow extension slf
273 functions to have destructors.
274 - get rid if coro refcounting - simply crash in other interpreter
275 threads by nulling the pointers on clone.
276 - simplify warn/die hook handling when loading Coro - the convoluted
277 logic seems to be no longer neccessary.
278 - use libecb instead of our own home-grown gcc hacks.
279 - document alternatives to Coro::LWP. Please use them :)
280 - work around another mindless idiotic NEEDLESS bug in openbsd/mirbsds
281 sigaltstack. Really. wine suffers from it, erlang suffers from it,
282 and it's known since at least 2006.
283
2845.372 Wed Feb 23 06:14:30 CET 2011
285 - apparently mingw doesn't provide a working gettimeofday, try to
286 work around that by relying on Time::HiRes (indirectly brought to
287 my attention by Max Maischein).
288 - fix some portability issues when Time::HiRes was used.
289
2905.371 Mon Feb 21 14:36:08 CET 2011
291 - backport to windows process emulation code again.
292
2935.37 Sat Feb 19 07:49:44 CET 2011
294 - add a big "Coro thread life cycle" section to "man Coro".
295 - try a tentative workaround against the breakage that 5.13 has
296 introduced without depreciation period. sigh.
297 - no longer use Time::HiRes if gettimeofday is available, which
298 saves quite a lot of memory.
299
3005.36 Sun Feb 13 05:33:41 CET 2011
301 - automatically load Coro::Channel, Coro::RWLock, Coro::Semaphore,
302 Coro::SemaphoreSet, Coro::Signal and Coro::Specific on first "new"
303 method call.
304 - undocument Coro::Timer::sleep and obsolete whole module.
305 - optimise Coro::Timer::timeout memory and cpu usage.
306 - slightly updated Coro::Intro for recent changes.
307 - do not initialise PL_dirty anymore.
308
3095.25 Thu Nov 11 01:08:39 CET 2010
310 - try a different approach on netbsd - netbsd 5 finally has marginally
311 working pthreads, but still broken ucontext/sigaltstack.
312 - openbsd 4.8 finally got their act together, Coro works out of the box
313 with asm, setjmp and pthreads (no change, just informational).
314
3155.24 Sat Oct 23 11:27:12 CEST 2010
316 - port to the EV 4.0 API.
317 - work around bugs in mingw32, making strawberry perl work
318 out of the box.
319 - correctly modify Coro::AIO function prototypes
320 so that they reflect the "no optional parameters" rule.
9 - "ported" libcoro to C++. 321 - "ported" libcoro to C++.
10 322
115.23 Mon May 17 18:50:42 CEST 2010 3235.23 Mon May 17 18:50:42 CEST 2010
12 - be more resistant to ordering changes when initialising 324 - be more resistant to ordering changes when initialising
13 Coro::AnyEvent, Coro::EV and Coro::Event (reported by Matthias 325 Coro::AnyEvent, Coro::EV and Coro::Event (reported by Matthias
… …
293 - ->throw is now supported on Coro::State objects. 605 - ->throw is now supported on Coro::State objects.
294 - clean up cctx creation code a bit. 606 - clean up cctx creation code a bit.
295 - entersub is actually a UNOP, not a LOGOP (not a bugfix). 607 - entersub is actually a UNOP, not a LOGOP (not a bugfix).
296 608
2974.9 Sat Nov 8 17:45:27 CET 2008 6094.9 Sat Nov 8 17:45:27 CET 2008
298 - (libcoro) did not preserve rbp with CORO_ASM (we are getting there). 610 - (libcoro) did not preserve rbp with CORO_ASM (we are getting there).
299 - (libcoro) no longer leak threads in the experimental pthread backend, 611 - (libcoro) no longer leak threads in the experimental pthread backend,
300 also speed it up considerably. 612 also speed it up considerably.
301 - (libcoro) do not rely on makecontext passing void *'s unscathed. 613 - (libcoro) do not rely on makecontext passing void *'s unscathed.
302 - fix compiletime dependencies on libcoro in the Makefile. 614 - fix compiletime dependencies on libcoro in the Makefile.
303 - cctx_count wasn't always updated properly. 615 - cctx_count wasn't always updated properly.
… …
365 - remove debugging code related to MgPV_nolen_const, also try to 677 - remove debugging code related to MgPV_nolen_const, also try to
366 make it compile with perl 5.8.6 (yes, apple apparently loves 678 make it compile with perl 5.8.6 (yes, apple apparently loves
367 outdated software). Reported by John S. 679 outdated software). Reported by John S.
368 680
3694.744 Tue Jul 8 22:06:35 CEST 2008 6814.744 Tue Jul 8 22:06:35 CEST 2008
370 - correctly provide default DIE/WARN handlers as documented. 682 - correctly provide default DIE/WARN handlers as documented.
371 - also overwrite PL_vtbl_sigelem.svt_clear, even though current 683 - also overwrite PL_vtbl_sigelem.svt_clear, even though current
372 implementations inside perl work fine for us. 684 implementations inside perl work fine for us.
373 685
3744.743 Mon Jun 16 00:21:57 CEST 2008 6864.743 Mon Jun 16 00:21:57 CEST 2008
375 - when using Coro::EV without running EV::loop it could 687 - when using Coro::EV without running EV::loop it could
… …
397 - sprinkle "no warnings" freely over everything, also suppress 709 - sprinkle "no warnings" freely over everything, also suppress
398 warnings for some other modules. 710 warnings for some other modules.
399 - fix typo in WSAEWOULDBLOCK. 711 - fix typo in WSAEWOULDBLOCK.
400 712
4014.72 Sun May 25 05:14:36 CEST 2008 7134.72 Sun May 25 05:14:36 CEST 2008
402 - tweak META.yaml a bit, unfortunately, there is no documented way 714 - tweak META.yaml a bit, unfortunately, there is no documented way
403 to have optional dependencies with CPAN. doh :( 715 to have optional dependencies with CPAN. doh :(
404 - avoid running some tests on windows because they would fail due to 716 - avoid running some tests on windows because they would fail due to
405 perl bug (broken fork, broken pipes...). 717 perl bug (broken fork, broken pipes...).
406 - work around perl on windows bugs where perl returns undocumented 718 - work around perl on windows bugs where perl returns undocumented
407 error codes for sysread, syswrite etc. by taking advantage 719 error codes for sysread, syswrite etc. by taking advantage
… …
411 - use unix domain sockets in testsuite to work around 723 - use unix domain sockets in testsuite to work around
412 common perl implementation bugs on widows (they are emulated by 724 common perl implementation bugs on widows (they are emulated by
413 tcp sockets on windows. ugh.) 725 tcp sockets on windows. ugh.)
414 726
4154.71 Sat May 24 20:01:27 CEST 2008 7274.71 Sat May 24 20:01:27 CEST 2008
416 - fix a bug in Coro::AnyEvent ("Usage: Coro::AnyEvent::_schedule()"). 728 - fix a bug in Coro::AnyEvent ("Usage: Coro::AnyEvent::_schedule()").
417 - take advantage of async name resolution of AnyEvent::Util. 729 - take advantage of async name resolution of AnyEvent::Util.
418 - work around brutal inet_aton override in Coro::LWP. 730 - work around brutal inet_aton override in Coro::LWP.
419 - take advantage of the readyhook in Coro::EV, for smoother 731 - take advantage of the readyhook in Coro::EV, for smoother
420 scheduling. 732 scheduling.
421 733
4224.7 Sun May 11 00:32:19 CEST 2008 7344.7 Sun May 11 00:32:19 CEST 2008
423 - completely reworked the Coro manpage. 735 - completely reworked the Coro manpage.
424 - added Coro::AnyEvent, generic event loop integration. 736 - added Coro::AnyEvent, generic event loop integration.
425 - implement cancel, ready and kill commands in Coro::Debug. 737 - implement cancel, ready and kill commands in Coro::Debug.
426 - document find_coro in Coro::Debug. 738 - document find_coro in Coro::Debug.
427 - incompatible change: rename has_stack to has_cctx. 739 - incompatible change: rename has_stack to has_cctx.
428 - Coro::AIO and Coro::BDB no longer force event model detection, 740 - Coro::AIO and Coro::BDB no longer force event model detection,
429 use AnyEvent::AIO and AnyEvent::BDB. 741 use AnyEvent::AIO and AnyEvent::BDB.
… …
553 865
5544.11 Thu Oct 11 02:40:24 CEST 2007 8664.11 Thu Oct 11 02:40:24 CEST 2007
555 - port to threaded perls. 867 - port to threaded perls.
556 868
5574.1 Thu Oct 11 02:38:16 CEST 2007 8694.1 Thu Oct 11 02:38:16 CEST 2007
558 - incompatible change: $SIG{__DIE__} and $SIG{__WARN__} will now 870 - incompatible change: $SIG{__DIE__} and $SIG{__WARN__} will now
559 be local to each coro (see Coro::State). 871 be local to each coro (see Coro::State).
560 - incompatible change: for very deep reasons, cede and cede_notself 872 - incompatible change: for very deep reasons, cede and cede_notself
561 cannot return anything, so nothing will be returned. 873 cannot return anything, so nothing will be returned.
562 - possibly bring back 5.10 compatibility (untested). 874 - possibly bring back 5.10 compatibility (untested).
563 - work around stupid (and wrong) warning on 5.10 :(. 875 - work around stupid (and wrong) warning on 5.10 :(.
… …
635 but stateful semantics. 947 but stateful semantics.
636 - fixed a lot of typos in Coro.pm (patch submitted by David 948 - fixed a lot of typos in Coro.pm (patch submitted by David
637 Steinbrunner, which applied flawlessly). 949 Steinbrunner, which applied flawlessly).
638 950
6393.6 Sat Apr 14 17:13:31 CEST 2007 9513.6 Sat Apr 14 17:13:31 CEST 2007
640 - added some bugfixes to get eg/myhttpd working again. 952 - added some bugfixes to get eg/myhttpd working again.
641 - added Coro::Storable for often-cede'ing freeze/thaw. 953 - added Coro::Storable for often-cede'ing freeze/thaw.
642 - try to do a clean exit when a coroutine calls exit 954 - try to do a clean exit when a coroutine calls exit
643 (EXPERIMENTAL). 955 (EXPERIMENTAL).
644 - got rid of indirect call through _coro_init. 956 - got rid of indirect call through _coro_init.
645 - updated the partly antique examples in eg/ to 957 - updated the partly antique examples in eg/ to

Diff Legend

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