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

Comparing Coro/Changes (file contents):
Revision 1.454 by root, Thu Nov 20 15:00:36 2008 UTC vs.
Revision 1.529 by root, Tue Nov 24 06:13:01 2009 UTC

… …
2 2
3TODO: should explore PerlIO::coroaio (perl leaks like hell). 3TODO: should explore PerlIO::coroaio (perl leaks like hell).
4TODO: maybe implement a default message channel, very much like Erlang's 4TODO: maybe implement a default message channel, very much like Erlang's
5 actor model (which is cool in a lot of important aspects (failures!), 5 actor model (which is cool in a lot of important aspects (failures!),
6 but very lacking in others (higher level ipc)). 6 but very lacking in others (higher level ipc)).
7TODO: unready_all
8TODO: myhttpd header parsing
9TODO: channel->maxsize(newsize)?
7 10
11 - clarify the "not from signal handlers" section.
12
135.2 Sun Oct 4 14:54:24 CEST 2009
14 - Coro::Storable destroyed the prototypes of the functions it wrapped.
15 - export rouse_cb and rouse_wait by default now.
16 - fix various prototype mismatches in Coro::AnyEvent and Coro::Handle.
17 - new method $state->swap_sv.
18 - added section on "windows process emulation" to the manpage,
19 after a not-so-fruitful (nor-friendly) "discussion" with chip
20 salzenberg (discussion implies arguments, but his only arguments
21 were ad-hominems, one wonders why he started it in the first
22 place). I hope this explains it well enough for him to understand,
23 and maybe well enough for others to understand.
24 - use common::sense everywhere now.
25 - idle callbacks are no longer supported, use idle coros instead.
26 - print a thread listing when a deadlock is detected.
27
285.17 Sat Aug 22 23:09:31 CEST 2009
29 - work around a bug in the perl debugger causing crashes
30 when running under the debugger by marking _pool_handler as nodebug.
31 - speed up Coro::async considerably.
32 - try some hacks to get netbsd to work "more often" - their broken
33 setjmp/longjmp, ucontext *and* phtreads are really hard on Coro.
34 - convert Coro to AE 5.0 API.
35
365.162 Tue Jul 28 04:04:03 CEST 2009
37 - perl 5.8.2 is now minimum requirement.
38 - skip t/19_handle.t on broken windows perls.
39
405.161 Wed Jul 22 04:47:38 CEST 2009
41 - Coro::AnyEvent::poll could have a different prototype when EV was
42 used as backend (analyzed by Tatsuhiko Miyagawa).
43 - Coro::AnyEvent errornously initialised the event loop when loaded,
44 not on demand.
45 - try to workaround rare */t/01_unblock.t failures.
46
475.16 Tue Jul 21 01:44:37 CEST 2009
48 - Coro::AnyEvent failed to hook into the event loop
49 when no threads had been readied between detecting
50 the event loop and actually running it.
51 - considerably speed up Coro::Select by taking avdantage
52 of AnyEvent > 4.8 and some other optimisations.
53 - implement paragraph readline mode in Coro::Handle
54 (based on patches by Zsbán Ambrus).
55 - replace WSAEINPROGRESS by WSAEWOULDBLOCK (reported
56 and analyzed by Yasuhiro MATSUMOTO).
57 - clarified libcoro license and copyright.
58 - someone stole my EXTRA_META!!!
59 - implement Coro::Select::patch_pp_sselect and it's brother,
60 for hardcode select overriding.
61
625.151 Mon Jul 6 05:41:57 CEST 2009
63 - backport to windows process emulation code again (patch by
64 Yasuhiro MATSUMOTO).
65 - slightly update Coro::MakeMaker.
66
675.15 Tue Jun 30 10:28:06 CEST 2009
68 - deprecate Coro::Socket, document how to get ipv6 support via
69 AnyEvent::Socket instead.
70 - implement signal->wait ($cb) interface, similar to semaphores.
71 - work around SvOK not supporting getmagic, so we have to getmagic
72 to test for undef :/ (reported by Matthias Waldorf).
73 - load Coro::AnyEvent in all modules using AnyEvent.
74 - work around perl corrupting our internal data structures,
75 reported by Tokuhiro Matsuno.
76 - enable per-coroutine real and cpu time gathering
77 (Coro::State::enable_times).
78
795.14 Wed Jun 24 01:37:48 CEST 2009
80 - provide explicit functions to "cede" to the event loop in Coro::AnyEvent,
81 as this seems to have been a difficult concept (poll, sleep, idle,
82 idle_upto).
83 - add Coro::AnyEvent::readable/writable functions.
84 - clarify Coro::EV/Event/AnyEvent manpages.
85 - free per-thread global scalars in the thread calling ->cancel, to
86 avoid crashes when $_, $@ etc., are magical but some of those
87 had already been freed.
88 - "unexperimentalise" the callback interface for Coro::Semaphore.
89 - speed up ready queue management/context switching by using a linked
90 list instead of an array (~5-10%).
91 - implement "watch" command in Coro::Debug shells.
92 - for fun, implement time-slicing as an example in the manpage.
93 - if AnyEvent detects EV or Event, but we don't have Coro::EV or
94 Coro::Event, use the normal AnyEvent handling instead of dieing
95 (the same is true for Coro::Handle).
96 - properly document Coro::EV::timed_io_once.
97 - avoid unneccessary ->cancel calls in Coro::Handle.
98 - maybe make it work on mingw32 with win32 backend
99 (based on patch by Yasuhiro Matsumoto).
100
1015.132 Fri May 29 09:00:39 CEST 2009
102 - do not keep a reference to the argument itself in
103 Coro::Semaphore::guard, as it could change later.
104 - support SO_RCVBUF/SO_SNDBUF nonstandard Coro::Socket options,
105 should support a prepare callback.
106
1075.131 Mon Mar 16 23:20:37 CET 2009
108 - implement and document Coro->suspend, Coro->resume.
109 - fix Coro::Select implementation to not (often) close
110 the passed file descriptors (testcase provided by pippijn).
111
1125.13 Mon Dec 15 21:51:42 CET 2008
113 - EXPERIMENTAL: implement dynamic winds (on_enter/on_leave).
114 - don't set diehook to C<undef> but instead to NULL, to avoid
115 spurious warnings.
116 - fix a lot of bugs in Coro::SemaphoreSet.
117 - Coro::SemaphoreSet will less often create a semaphore needlessly.
118 - add Coro::SemaphoreSet::count and wait methods.
119 - take advantage of the new Guard module.
120 - deprecate Coro::guard.
121 - try to fix the dreaded 01_unblock tests once more. I hate it when
122 testsuites need more fixing than the code they are supposed to test.
123 - croak in more cases when a required callback isn't resolvable.
124 - fix some minor issues in Coro::State->call/eval.
125 - use current coroutine context instead of a temporary one
126 when temporarily switching to another coroutine.
127 - do not call C-level on_destroy handlers during global destruction,
128 to avoid needless segfaults.
129
1305.12 Sun Dec 7 13:30:38 CET 2008
131 - add default config for MirOS, which seems to be a bug-to-bug
132 compatible fork of openbsd ("world domination by releasing
133 openbsd cvs before the openbsd folks do it" or so :).
134 - free_padlist did destroy the names pad, not good, but didn't
135 seem to bother perl - this could fix issues such as eval ""
136 inside a function called from multiple coroutines.
137 - use a different method to detect destruction time.
138 - be more careful when freeing padlists just before global
139 destruction.
140 - fixed and expanded the call/cc example.
141 - renamed _terminate to _coro_run.
142 - new method Coro::Channel->shutdown.
143 - try pthreads on openbsd <4.4 (broken sigaltstack, will
144 pthreads fare better?).
145 - be less picky about destroying "a" running coroutine.
146
1475.11 Tue Nov 25 21:49:05 CET 2008
148 - DEBUGGING in 5.10.0 is a mess: it constantly flags perfectly
149 working code with failed assertions, introducing more bugs than
150 it fixes, requiring elaborate workarounds :(
151
1525.1 Mon Nov 24 08:54:59 CET 2008
153 - wrote a small introductory tutorial - Coro::Intro.
8 - convert Coro::Timer, Coro::Select and Coro::Util to rouse API. 154 - convert Coro::Timer, Coro::Select and Coro::Util to rouse API.
9 - Coro::Select did errornously dup the file descriptors 155 - Coro::Select did errornously dup the file descriptors
10 and didn't work with all AnyEvent backends. 156 and didn't work with all AnyEvent backends.
157 - Coro::Select wasn't imported correctly form Coro::LWP, causing blocking
158 LWP data transfers.
159 - disassociate c contexts from coro objects - this is agruably more
160 correct, but mostly allows sharing of cctxs between coro and state
161 objects, for added memory savings and speed increases.
162 - bumped $Coro::POOL_RSS up to 32kb by default.
163 - no longer set the optype to OP_CUSTOM, as B::* understandably
164 doesn't like this very much (and we *are* a type of entersub).
165 - implement state cloning, just to prove that call/cc can be done.
166 - automatically load Coro::AnyEvent in Coro::Handle.
167 - wrap ->cancel calls in eval inside Coro::Handle as EV watchers
168 do not have this method (and don't need it either).
169 - speed up generic anyevent methods in Coro::Handle by using rouse
170 callbacks.
171 - allow coroutines in $Coro::IDLE, speeding up Coro::AnyEvent and
172 others. It also makes the debugger happier, as you can trace
173 through the idle threads now.
174 - add comppad_name* and hints ($^H, %^H) to per-thread variables.
175 - eg/event was pretty broken.
176 - better 5.8.6 compatibility.
11 177
125.0 Thu Nov 20 10:35:05 CET 2008 1785.0 Thu Nov 20 10:35:05 CET 2008
13 - NEW ARCHITECTURE: use the latest 4.x version if you experience 179 - NEW ARCHITECTURE: use the latest 4.x version if you experience
14 stability issues. 180 stability issues.
15 - bump API version to 7 - all dependents must be recompiled. 181 - bump API version to 7 - all dependents must be recompiled.

Diff Legend

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