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

Comparing Coro/Changes (file contents):
Revision 1.546 by root, Sun Feb 6 00:27:43 2011 UTC vs.
Revision 1.638 by root, Wed Feb 11 19:30:15 2015 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, andallow to unswap by repeat?
7TODO: swap_sv not undone in pool
8TODO: croak when async_pool triesd to run canceled thread?
8 9
106.42 Wed Feb 11 20:29:52 CET 2015
11 - Coro::SemaphoreSet->try did not actually work (analyzed by
12 SATO Kentaro).
13 - upgrade libecb to be C11 compliant.
14
156.41 Sat Sep 6 22:08:46 CEST 2014
16 - restore portability to perl 5.8.x.
17 - give new Coro's a valid GvHV(PL_hintgv) - this is slow and takes
18 up some memory, but fixes "use feature" and similar modules
19 when used inside a Coro.
20 - allow zero as argument to Coro::Channel to mean the same thing
21 as no argument. this works with older versions as well,
22 but wasn't legal until now.
23 - slightly better c header file detection.
24
256.39 Mon Jun 2 00:00:08 CEST 2014
26 - work around more incompatible changes in 5.20.
27
286.38 Sun Jun 1 21:54:23 CEST 2014
29 - check that perl slots actually have enough space to hold
30 interpreter variables.
31 - untested port to perl 5.19 (64 bit tmps indices) (reported
32 by Andreas König).
33 - croak when cancel is called without a thread context.
34
356.37 Tue Mar 4 13:27:33 CET 2014
36 - *sigh*, removed leftover debugging code from debugging a
37 perl bug, of all things.
38
396.36 Tue Mar 4 07:11:59 CET 2014
40 - semaphores would not clear the destroy hook when interrupted
41 by ->throw, causing segfaults or worse.
42 - ->throw on a thread waiting for a semaphore did not acquire
43 the semaphore, but also didn't wake up other waiters,
44 possibly causing a deadlock.
45 - "FATAL: $Coro::IDLE blocked itself" will now use Carp::confess
46 to report a full stacktrace, which should help find out
47 where the actual call is.
48 - "atomically" destroy data in slf_destroy, because it is
49 the right thing to do, just in case.
50 - disable ecb.h libm dependency, to compile on ancient systems
51 or under adverse conditions.
52
536.33 Mon Nov 18 11:26:27 CET 2013
54 - do not crash when freeing padlists with holes (in 5.18).
55 - tentative SVt_BIND 5.19 port/fix.
56
576.32 Tue Nov 5 15:35:35 CET 2013
58 - use a new algorithm to derive padlists for perl 5.18. The old
59 one could lead to 0-pointer accesses inside perl (reported
60 by Darin McBride).
61
626.31 Thu May 9 07:39:48 CEST 2013
63 - Coro::AIO requests would crash if the thread was ready'd
64 while the request was ongoing.
65
666.29 Wed May 8 02:55:18 CEST 2013
67 - when an on_destroy handler destructs the coro currently being
68 destructed a perl scalar could be accessed after being freed,
69 likely causing a crash.
70
716.28 Wed Mar 6 06:58:02 CET 2013
72 - clean remnants of existing __DIE__ and __WARN__ handlers so
73 they lose their magic and will not cause segfaults later
74 (testcase by Andrey Sagulin).
75 - improved Coro::State documentation a bit.
76 - Coro::Debug::command now flushes the output.
77 - add hack detection code for x32 abi, because the braindead slugs
78 who designed that made it look exactly like x86_64 without
79 providing proper compile time symbols to test for it. as a result,
80 this detection cannot work reliably.
81 - valgrind stack registering was broken.
82 - do not rely on Time::HiRes anymore in Coro::Debug.
83
846.23 Fri Dec 7 23:36:37 CET 2012
85 - use experimental fiber implementation on native windows
86 perls.
87 - use sizeof (void *) as multiplication factor for stack sizes,
88 to accomodate the totally braindamaged microsoft 64 bit "os".
89 - changed verifier host from win2k-ap510-32 to win7-sp516-32/64.
90 activeperl 5.16 crashes when PerlIO_define_layer is called due
91 to some bug in the perl dll, strawberry perl at least passes
92 the testsuite.
93 - implement Coro::Handle->peeraddr/host/port, for slightly
94 improved compatibility with LWP.
95 - implement 5.17 compatibility by almost blindly applying a
96 good-looking patch by Father Chrysostomos.
97 - move stack management functions into libcoro 3.
98 - libcoro version 3 "released".
99 - support magic values as timed_io_once args.
100 - recommend AnyEvent 7+ or EV 4+, also require EV
101 version 4 or newer for Coro::EV.
102
1036.10 Tue Oct 9 01:14:27 CEST 2012
104 - updated ecb.h, it had a typo that caused it to not compile on many
105 big endian systems (reported by many people).
106 - disable memory fences in ecb.h to improve portability.
107
1086.09 Sat Oct 6 23:25:02 CEST 2012
109 - Coro::EV I/O watchers were not interruptible by exceptions
110 (Coro::State::throw) (testcase by sten).
111 - ->throw now puts threads into the ready queue, as this seems to
112 be expected by existing code, and code that doesn't cope with spurious
113 wakeups needs fixing anyway.
114 - use fd -1 in mmap.
115 - cast I32 to int in error message printf.
116 - warn about broken so-called "hardened" kernels.
117
1186.08 Fri Apr 13 12:05:47 CEST 2012
119 - be more aggressive about exiting like perl does - formerly,
120 exiting from the non-main thread would not execute END blocks.
121
1226.07 Fri Nov 11 21:21:48 CET 2011
123 - work around a bug in PerlIO (setting $SIG{__WARN__} to a PVCV).
124 - update ecb.h.
125
1266.06 Mon Aug 8 23:59:48 CEST 2011
127 - cygwin unfortunately patches the stack at runtime, so we use the pthreads
128 backend, which is an order of magnitude slower. unfortunately, cygwins
129 pthread implementation isn't very complete either, so allocate the stack
130 twice just to be sure.
131 (note: cygwin also enables mymalloc, which is NOT THREADSAFE ON WINDOWS,
132 in its ithreaded perl - best recompile cygwin and use the 'w'indows
133 backend for much better performance. also disable ithreads for
134 even better performance...).
135
1366.05 Thu Aug 4 21:36:36 CEST 2011
137 - blush, condvar values would not be propagated from send to recv anymore
138 (reported by Chip Salzenberg).
139 - use exponential increase for the readline buffer length in
140 Coro::Handle. also reduce initial allocation to 1020 from 4096 bytes.
141
1426.04 Wed Aug 3 17:15:45 CEST 2011
143 - use even more efficient and more compatible condvars for
144 compatibility to AnyEvent 6.x :)
145 - more inconsequential ecb.h updates.
146
1476.03 Wed Aug 3 11:41:30 CEST 2011
148 - change how Coro patches AnyEvent condvars for compatibility to
149 AnyEvent 6.x.
150 - update ecb.h, to no longer include <pthread.h> in case WinNT.h
151 hasn't been included.
152
1536.02 Wed Jul 13 04:35:19 CEST 2011
154 - "improve portability to Gentoo" - gentoo manages to put perl variables
155 in memory areas that are farther than 2gb apart, which the jit couldn't
156 handle and barfed. now it's just a bit slower on gentoo and similar
157 systems.
158
1596.01 Sun Jul 3 12:31:14 CEST 2011
160 - workarounds are good, but the test for whether pthreads are used
161 was not good. this one should be better.
162 - check differently whether gcc generates cfi instructions itself.
163
1646.0 Wed Jun 29 19:43:35 CEST 2011
165 - INCOMPATIBLE CHANGE: unreferenced coro objects will now be
166 destroyed and cleaned up automatically (e.g. async { schedule }).
167 - implement a JIT compiler for part of the thread switch code,
168 which gives a 50% speed improvement on threaded perls, and
169 about 4% on non-threaded perls (so threaded perls now finally
170 reach about half the speed of non-threaded perls).
171 - slightly modernise Coro::Intro, add section about rouse functions.
172 - avoid DEFSV and ERRSV, giving another 10% improvement
173 in thread switching.
174 - Coro::State->is_destroyed is now called is_zombie.
175 - implement a Coro->safe_cancel method that might fail, but
176 cancels in a "safer" way if it succeeds.
177 - add preliminary support for DEBUGGING perls.
178 - get rid of two hash-accesses when initialising a new Coro - this
179 speeds up coro creation by almost a factor of two.
180 - croak when a coro that is being cancelled tries to block
181 (e.g. while executing a guard block), instead of crashing or
182 deadlocking.
183 - use a more robust and also faster method to identify Coro::State
184 objects - speeds up everything a bit.
185 - implement Coro->cancel in XS for a 20% speed improvement, and to
186 be able to implement mutual cancellation.
187 - speed up context switches by a percent or two by more efficiently
188 allocating context stack entries.
189 - implement Coro->join and Coro->on_destroy in XS for a speedup and
190 a reduction in memory use.
191 - cancelling a coro while it itself is cancelling another coro is
192 now supported and working, instead of triggering an assertion.
193 - be a bit more crash-resistant when calling (buggy) on_destroy
194 callbacks (best effort).
195 - move on_destroy into the slf_frame, to allow extension slf
196 functions to have destructors.
197 - get rid if coro refcounting - simply crash in other interpreter
198 threads by nulling the pointers on clone.
199 - simplify warn/die hook handling when loading Coro - the convoluted
200 logic seems to be no longer neccessary.
201 - use libecb instead of our own home-grown gcc hacks.
202 - document alternatives to Coro::LWP. Please use them :)
203 - work around another mindless idiotic NEEDLESS bug in openbsd/mirbsds
204 sigaltstack. Really. wine suffers from it, erlang suffers from it,
205 and it's known since at least 2006.
206
2075.372 Wed Feb 23 06:14:30 CET 2011
208 - apparently mingw doesn't provide a working gettimeofday, try to
209 work around that by relying on Time::HiRes (indirectly brought to
210 my attention by Max Maischein).
211 - fix some portability issues when Time::HiRes was used.
212
2135.371 Mon Feb 21 14:36:08 CET 2011
214 - backport to windows process emulation code again.
215
2165.37 Sat Feb 19 07:49:44 CET 2011
217 - add a big "Coro thread life cycle" section to "man Coro".
218 - try a tentative workaround against the breakage that 5.13 has
219 introduced without depreciation period. sigh.
220 - no longer use Time::HiRes if gettimeofday is available, which
221 saves quite a lot of memory.
222
2235.36 Sun Feb 13 05:33:41 CET 2011
9 - automatically load Coro::Channel, Coro::RWLock, Coro::Semaphore, 224 - automatically load Coro::Channel, Coro::RWLock, Coro::Semaphore,
10 Coro::SemaphoreSet, Coro::Signal and Coro::Specific on first "new" 225 Coro::SemaphoreSet, Coro::Signal and Coro::Specific on first "new"
11 method call. 226 method call.
12 - undocument Coro::Timer::sleep and obsolete whole module. 227 - undocument Coro::Timer::sleep and obsolete whole module.
13 - optimise Coro::Timer::timeout memory and cpu usage. 228 - optimise Coro::Timer::timeout memory and cpu usage.
14 - slightly updated Coro::Intro for recent changes. 229 - slightly updated Coro::Intro for recent changes.
230 - do not initialise PL_dirty anymore.
15 231
165.25 Thu Nov 11 01:08:39 CET 2010 2325.25 Thu Nov 11 01:08:39 CET 2010
17 - try a different approach on netbsd - netbsd 5 finally has marginally 233 - try a different approach on netbsd - netbsd 5 finally has marginally
18 working pthreads, but still broken ucontext/sigaltstack. 234 working pthreads, but still broken ucontext/sigaltstack.
19 - openbsd 4.8 finally got their act together, Coro works out of the box 235 - openbsd 4.8 finally got their act together, Coro works out of the box
… …
312 - ->throw is now supported on Coro::State objects. 528 - ->throw is now supported on Coro::State objects.
313 - clean up cctx creation code a bit. 529 - clean up cctx creation code a bit.
314 - entersub is actually a UNOP, not a LOGOP (not a bugfix). 530 - entersub is actually a UNOP, not a LOGOP (not a bugfix).
315 531
3164.9 Sat Nov 8 17:45:27 CET 2008 5324.9 Sat Nov 8 17:45:27 CET 2008
317 - (libcoro) did not preserve rbp with CORO_ASM (we are getting there). 533 - (libcoro) did not preserve rbp with CORO_ASM (we are getting there).
318 - (libcoro) no longer leak threads in the experimental pthread backend, 534 - (libcoro) no longer leak threads in the experimental pthread backend,
319 also speed it up considerably. 535 also speed it up considerably.
320 - (libcoro) do not rely on makecontext passing void *'s unscathed. 536 - (libcoro) do not rely on makecontext passing void *'s unscathed.
321 - fix compiletime dependencies on libcoro in the Makefile. 537 - fix compiletime dependencies on libcoro in the Makefile.
322 - cctx_count wasn't always updated properly. 538 - cctx_count wasn't always updated properly.
… …
384 - remove debugging code related to MgPV_nolen_const, also try to 600 - remove debugging code related to MgPV_nolen_const, also try to
385 make it compile with perl 5.8.6 (yes, apple apparently loves 601 make it compile with perl 5.8.6 (yes, apple apparently loves
386 outdated software). Reported by John S. 602 outdated software). Reported by John S.
387 603
3884.744 Tue Jul 8 22:06:35 CEST 2008 6044.744 Tue Jul 8 22:06:35 CEST 2008
389 - correctly provide default DIE/WARN handlers as documented. 605 - correctly provide default DIE/WARN handlers as documented.
390 - also overwrite PL_vtbl_sigelem.svt_clear, even though current 606 - also overwrite PL_vtbl_sigelem.svt_clear, even though current
391 implementations inside perl work fine for us. 607 implementations inside perl work fine for us.
392 608
3934.743 Mon Jun 16 00:21:57 CEST 2008 6094.743 Mon Jun 16 00:21:57 CEST 2008
394 - when using Coro::EV without running EV::loop it could 610 - when using Coro::EV without running EV::loop it could
… …
416 - sprinkle "no warnings" freely over everything, also suppress 632 - sprinkle "no warnings" freely over everything, also suppress
417 warnings for some other modules. 633 warnings for some other modules.
418 - fix typo in WSAEWOULDBLOCK. 634 - fix typo in WSAEWOULDBLOCK.
419 635
4204.72 Sun May 25 05:14:36 CEST 2008 6364.72 Sun May 25 05:14:36 CEST 2008
421 - tweak META.yaml a bit, unfortunately, there is no documented way 637 - tweak META.yaml a bit, unfortunately, there is no documented way
422 to have optional dependencies with CPAN. doh :( 638 to have optional dependencies with CPAN. doh :(
423 - avoid running some tests on windows because they would fail due to 639 - avoid running some tests on windows because they would fail due to
424 perl bug (broken fork, broken pipes...). 640 perl bug (broken fork, broken pipes...).
425 - work around perl on windows bugs where perl returns undocumented 641 - work around perl on windows bugs where perl returns undocumented
426 error codes for sysread, syswrite etc. by taking advantage 642 error codes for sysread, syswrite etc. by taking advantage
… …
430 - use unix domain sockets in testsuite to work around 646 - use unix domain sockets in testsuite to work around
431 common perl implementation bugs on widows (they are emulated by 647 common perl implementation bugs on widows (they are emulated by
432 tcp sockets on windows. ugh.) 648 tcp sockets on windows. ugh.)
433 649
4344.71 Sat May 24 20:01:27 CEST 2008 6504.71 Sat May 24 20:01:27 CEST 2008
435 - fix a bug in Coro::AnyEvent ("Usage: Coro::AnyEvent::_schedule()"). 651 - fix a bug in Coro::AnyEvent ("Usage: Coro::AnyEvent::_schedule()").
436 - take advantage of async name resolution of AnyEvent::Util. 652 - take advantage of async name resolution of AnyEvent::Util.
437 - work around brutal inet_aton override in Coro::LWP. 653 - work around brutal inet_aton override in Coro::LWP.
438 - take advantage of the readyhook in Coro::EV, for smoother 654 - take advantage of the readyhook in Coro::EV, for smoother
439 scheduling. 655 scheduling.
440 656
4414.7 Sun May 11 00:32:19 CEST 2008 6574.7 Sun May 11 00:32:19 CEST 2008
442 - completely reworked the Coro manpage. 658 - completely reworked the Coro manpage.
443 - added Coro::AnyEvent, generic event loop integration. 659 - added Coro::AnyEvent, generic event loop integration.
444 - implement cancel, ready and kill commands in Coro::Debug. 660 - implement cancel, ready and kill commands in Coro::Debug.
445 - document find_coro in Coro::Debug. 661 - document find_coro in Coro::Debug.
446 - incompatible change: rename has_stack to has_cctx. 662 - incompatible change: rename has_stack to has_cctx.
447 - Coro::AIO and Coro::BDB no longer force event model detection, 663 - Coro::AIO and Coro::BDB no longer force event model detection,
448 use AnyEvent::AIO and AnyEvent::BDB. 664 use AnyEvent::AIO and AnyEvent::BDB.
… …
572 788
5734.11 Thu Oct 11 02:40:24 CEST 2007 7894.11 Thu Oct 11 02:40:24 CEST 2007
574 - port to threaded perls. 790 - port to threaded perls.
575 791
5764.1 Thu Oct 11 02:38:16 CEST 2007 7924.1 Thu Oct 11 02:38:16 CEST 2007
577 - incompatible change: $SIG{__DIE__} and $SIG{__WARN__} will now 793 - incompatible change: $SIG{__DIE__} and $SIG{__WARN__} will now
578 be local to each coro (see Coro::State). 794 be local to each coro (see Coro::State).
579 - incompatible change: for very deep reasons, cede and cede_notself 795 - incompatible change: for very deep reasons, cede and cede_notself
580 cannot return anything, so nothing will be returned. 796 cannot return anything, so nothing will be returned.
581 - possibly bring back 5.10 compatibility (untested). 797 - possibly bring back 5.10 compatibility (untested).
582 - work around stupid (and wrong) warning on 5.10 :(. 798 - work around stupid (and wrong) warning on 5.10 :(.
… …
654 but stateful semantics. 870 but stateful semantics.
655 - fixed a lot of typos in Coro.pm (patch submitted by David 871 - fixed a lot of typos in Coro.pm (patch submitted by David
656 Steinbrunner, which applied flawlessly). 872 Steinbrunner, which applied flawlessly).
657 873
6583.6 Sat Apr 14 17:13:31 CEST 2007 8743.6 Sat Apr 14 17:13:31 CEST 2007
659 - added some bugfixes to get eg/myhttpd working again. 875 - added some bugfixes to get eg/myhttpd working again.
660 - added Coro::Storable for often-cede'ing freeze/thaw. 876 - added Coro::Storable for often-cede'ing freeze/thaw.
661 - try to do a clean exit when a coroutine calls exit 877 - try to do a clean exit when a coroutine calls exit
662 (EXPERIMENTAL). 878 (EXPERIMENTAL).
663 - got rid of indirect call through _coro_init. 879 - got rid of indirect call through _coro_init.
664 - updated the partly antique examples in eg/ to 880 - updated the partly antique examples in eg/ to

Diff Legend

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