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

Comparing Coro/Changes (file contents):
Revision 1.411 by root, Fri Nov 14 07:22:11 2008 UTC vs.
Revision 1.508 by root, Mon Jul 6 03:42:29 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
85.0 115.151 Mon Jul 6 05:41:57 CEST 2009
9 - NEW ARCHITECTURE: use 4.912 for the stable version. 12 - backport to windows process emulation code again (patch by
13 Yasuhiro MATSUMOTO).
14 - slightly update Coro::MakeMaker.
15
165.15 Tue Jun 30 10:28:06 CEST 2009
17 - deprecate Coro::Socket, document how to get ipv6 support via
18 AnyEvent::Socket instead.
19 - implement signal->wait ($cb) interface, similar to semaphores.
20 - work around SvOK not supporting getmagic, so we have to getmagic
21 to test for undef :/ (reported by Matthias Waldorf).
22 - load Coro::AnyEvent in all modules using AnyEvent.
23 - work around perl corrupting our internal data structures,
24 reported by Tokuhiro Matsuno.
25 - enable per-coroutine real and cpu time gathering
26 (Coro::State::enable_times).
27
285.14 Wed Jun 24 01:37:48 CEST 2009
29 - provide explicit functions to "cede" to the event loop in Coro::AnyEvent,
30 as this seems to have been a difficult concept (poll, sleep, idle,
31 idle_upto).
32 - add Coro::AnyEvent::readable/writable functions.
33 - clarify Coro::EV/Event/AnyEvent manpages.
34 - free per-thread global scalars in the thread calling ->cancel, to
35 avoid crashes when $_, $@ etc., are magical but some of those
36 had already been freed.
37 - "unexperimentalise" the callback interface for Coro::Semaphore.
38 - speed up ready queue management/context switching by using a linked
39 list instead of an array (~5-10%).
40 - implement "watch" command in Coro::Debug shells.
41 - for fun, implement time-slicing as an example in the manpage.
42 - if AnyEvent detects EV or Event, but we don't have Coro::EV or
43 Coro::Event, use the normal AnyEvent handling instead of dieing
44 (the same is true for Coro::Handle).
45 - properly document Coro::EV::timed_io_once.
46 - avoid unneccessary ->cancel calls in Coro::Handle.
47 - maybe make it work on mingw32 with win32 backend
48 (based on patch by Yasuhiro Matsumoto).
49
505.132 Fri May 29 09:00:39 CEST 2009
51 - do not keep a reference to the argument itself in
52 Coro::Semaphore::guard, as it could change later.
53 - support SO_RCVBUF/SO_SNDBUF nonstandard Coro::Socket options,
54 should support a prepare callback.
55
565.131 Mon Mar 16 23:20:37 CET 2009
57 - implement and document Coro->suspend, Coro->resume.
58 - fix Coro::Select implementation to not (often) close
59 the passed file descriptors (testcase provided by pippijn).
60
615.13 Mon Dec 15 21:51:42 CET 2008
62 - EXPERIMENTAL: implement dynamic winds (on_enter/on_leave).
63 - don't set diehook to C<undef> but instead to NULL, to avoid
64 spurious warnings.
65 - fix a lot of bugs in Coro::SemaphoreSet.
66 - Coro::SemaphoreSet will less often create a semaphore needlessly.
67 - add Coro::SemaphoreSet::count and wait methods.
68 - take advantage of the new Guard module.
69 - deprecate Coro::guard.
70 - try to fix the dreaded 01_unblock tests once more. I hate it when
71 testsuites need more fixing than the code they are supposed to test.
72 - croak in more cases when a required callback isn't resolvable.
73 - fix some minor issues in Coro::State->call/eval.
74 - use current coroutine context instead of a temporary one
75 when temporarily switching to another coroutine.
76 - do not call C-level on_destroy handlers during global destruction,
77 to avoid needless segfaults.
78
795.12 Sun Dec 7 13:30:38 CET 2008
80 - add default config for MirOS, which seems to be a bug-to-bug
81 compatible fork of openbsd ("world domination by releasing
82 openbsd cvs before the openbsd folks do it" or so :).
83 - free_padlist did destroy the names pad, not good, but didn't
84 seem to bother perl - this could fix issues such as eval ""
85 inside a function called from multiple coroutines.
86 - use a different method to detect destruction time.
87 - be more careful when freeing padlists just before global
88 destruction.
89 - fixed and expanded the call/cc example.
90 - renamed _terminate to _coro_run.
91 - new method Coro::Channel->shutdown.
92 - try pthreads on openbsd <4.4 (broken sigaltstack, will
93 pthreads fare better?).
94 - be less picky about destroying "a" running coroutine.
95
965.11 Tue Nov 25 21:49:05 CET 2008
97 - DEBUGGING in 5.10.0 is a mess: it constantly flags perfectly
98 working code with failed assertions, introducing more bugs than
99 it fixes, requiring elaborate workarounds :(
100
1015.1 Mon Nov 24 08:54:59 CET 2008
102 - wrote a small introductory tutorial - Coro::Intro.
103 - convert Coro::Timer, Coro::Select and Coro::Util to rouse API.
104 - Coro::Select did errornously dup the file descriptors
105 and didn't work with all AnyEvent backends.
106 - Coro::Select wasn't imported correctly form Coro::LWP, causing blocking
107 LWP data transfers.
108 - disassociate c contexts from coro objects - this is agruably more
109 correct, but mostly allows sharing of cctxs between coro and state
110 objects, for added memory savings and speed increases.
111 - bumped $Coro::POOL_RSS up to 32kb by default.
112 - no longer set the optype to OP_CUSTOM, as B::* understandably
113 doesn't like this very much (and we *are* a type of entersub).
114 - implement state cloning, just to prove that call/cc can be done.
115 - automatically load Coro::AnyEvent in Coro::Handle.
116 - wrap ->cancel calls in eval inside Coro::Handle as EV watchers
117 do not have this method (and don't need it either).
118 - speed up generic anyevent methods in Coro::Handle by using rouse
119 callbacks.
120 - allow coroutines in $Coro::IDLE, speeding up Coro::AnyEvent and
121 others. It also makes the debugger happier, as you can trace
122 through the idle threads now.
123 - add comppad_name* and hints ($^H, %^H) to per-thread variables.
124 - eg/event was pretty broken.
125 - better 5.8.6 compatibility.
126
1275.0 Thu Nov 20 10:35:05 CET 2008
128 - NEW ARCHITECTURE: use the latest 4.x version if you experience
129 stability issues.
10 - bump API version to 7 - all dependents must be recompiled. 130 - bump API version to 7 - all dependents must be recompiled.
131 - removed timed_* functions - they were not being used anyways
132 and should be replaced by a more generic mechanism -
133 and were annoying to support anyways :)
134 - removed SemaphoreSet's waiter method - use sem method instead.
135 - Coro::Semaphore->adjust didn't correctly wake up enough waiters.
136 - async_pool did free a scalar value twice
137 ("Attempt to unreference...").
138 - fix a longstanding bug where calling terminate on a coro that
139 was waiting for a semaphore that was just becoming available
140 would cause a deadlock (semaphore would get into a state where
141 it was available but waiters were still blocked).
142 - calling throw on a coroutine that is waiting for a semaphore will
143 no longer make it acquire the semaphore (and thus leak a count).
144 - perl's process emulation is now not even theoretically supported
145 anymore.
146 - new functions Coro::rouse_cb and Coro::rouse_wait for easier
147 conversion of callback-style to blocking-style.
148 - new methods $coro->schedule_to and ->cede_to, to specifically
149 schedule or cede to a specific coroutine.
150 - new function Coro::Semaphore::wait.
151 - use named constants in Coro::Channel (Richard Hundt).
11 - directly patch the entersub opcode calling ssl functions (cede, 152 - directly patch the entersub opcode calling SLF functions (cede,
12 transfer and so on). this does speed up context switching by 153 transfer and so on). this does speed up context switching, but
13 up to 10%. More importanly, it frees us from the hardcoded 154 more importanly, it frees us from the hardcoded behaviour of
14 behaviour of entersub, so we might actually be able to return 155 entersub, so we might actually be able to return something from
15 something from those functions. 156 those functions and atcually create new ones.
16 - take advanatge of __builtin_frame_address on gcc. 157 - take advantage of __builtin_frame_address on gcc.
17 - expose THX in coroapi (not sure whether this was a wise decision). 158 - expose THX in coroapi (not sure whether this was a wise decision,
159 as "threaded" perls are running at half speed anyways).
160 - implement execute_slf (schedule-like-function) interface that makes
161 it possible to implement schedule-like-functions in XS.
162 - use new SLF interface to massively speed up Coro::EV by roughly a
163 factor of two.
164 - used new SLF interface to massively speed up Coro::Semaphore by a
165 factor of three.
166 - used new SLF interface to speed up Coro::AIO by roughly a factor of
167 four and reduce its memory usage considerably.
168 - implement Coro::SemaphoreSet purely in terms of Coro::Semaphore,
169 for a nice speedup and vastly more correct behaviour. Also implement
170 a new method "sem" to get at the underlying semaphore object.
171 - implement Coro::Channel in terms of Coro::Semaphore, for a moderate
172 (in comparison) 20-40% speedup.
173 - used new SLF interface to reimplement Coro::Signal gaining
174 some unknown (because I was too lazy), but certain, speedup, and also
175 making signals reliable for the first time.
176 - used new SLF interface and other optimisations to speed up async_pool
177 by a factor of two. It also doesn't rely on perl's exception mechanism
178 to exit anymore. The overhead for terminating an async_pool, coro over
179 a normal async is now very small.
180 - sped up coroutine creation/destruction by 40%.
181 - forgot to include Coro/libcoro/README in the dist for all these years.
182 - work around a freebsd pthreads bug (manual testcancel is required as
183 pthread_cond_wait isn't a cancellation point on freebsd).
184 - use new rouse functions to speed up and simplify Coro::BDB.
185 - make "prefer perl native functions" work with threaded perls.
186 - condense Coro::Debug ps output, hint at v and w flags.
187 - (libcoro) lots of minor cleanups and portability improvements.
188
1894.914 Wed Nov 19 12:54:18 CET 2008
190 - fix a disastrous bug in the readline optimisation
191 introduced in 4.801.
192
1934.913 Sat Nov 15 07:58:28 CET 2008
194 - async_pool did free a scalar value twice
195 ("Attempt to unreference...").
18 196
194.912 Thu Nov 13 18:31:23 CET 2008 1974.912 Thu Nov 13 18:31:23 CET 2008
20 - minor cleanups. 198 - minor cleanups.
21 - use much larger stacks on linux and perl < 5.8.8. 199 - use much larger stacks on linux and perl < 5.8.8.
22 - Coro::Debug::new_unix_server did not unlink the socket 200 - Coro::Debug::new_unix_server did not unlink the socket

Diff Legend

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