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

Comparing Coro/Changes (file contents):
Revision 1.441 by root, Wed Nov 19 11:57:55 2008 UTC vs.
Revision 1.506 by root, Wed Jul 1 07:26:40 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
8TODO: testcancel 11 - backport to windows process emulation code again (patch by
9TODO: guard as SLF - to avoid deadlocks. 12 Yasuhiro MATSUMOTO).
10 - avoid deadlock in Coro::Channel if maxsize == 1 (Richard Hundt).
11 13
125.0 145.15 Tue Jun 30 10:28:06 CEST 2009
15 - deprecate Coro::Socket, document how to get ipv6 support via
16 AnyEvent::Socket instead.
17 - implement signal->wait ($cb) interface, similar to semaphores.
18 - work around SvOK not supporting getmagic, so we have to getmagic
19 to test for undef :/ (reported by Matthias Waldorf).
20 - load Coro::AnyEvent in all modules using AnyEvent.
21 - work around perl corrupting our internal data structures,
22 reported by Tokuhiro Matsuno.
23 - enable per-coroutine real and cpu time gathering
24 (Coro::State::enable_times).
25
265.14 Wed Jun 24 01:37:48 CEST 2009
27 - provide explicit functions to "cede" to the event loop in Coro::AnyEvent,
28 as this seems to have been a difficult concept (poll, sleep, idle,
29 idle_upto).
30 - add Coro::AnyEvent::readable/writable functions.
31 - clarify Coro::EV/Event/AnyEvent manpages.
32 - free per-thread global scalars in the thread calling ->cancel, to
33 avoid crashes when $_, $@ etc., are magical but some of those
34 had already been freed.
35 - "unexperimentalise" the callback interface for Coro::Semaphore.
36 - speed up ready queue management/context switching by using a linked
37 list instead of an array (~5-10%).
38 - implement "watch" command in Coro::Debug shells.
39 - for fun, implement time-slicing as an example in the manpage.
40 - if AnyEvent detects EV or Event, but we don't have Coro::EV or
41 Coro::Event, use the normal AnyEvent handling instead of dieing
42 (the same is true for Coro::Handle).
43 - properly document Coro::EV::timed_io_once.
44 - avoid unneccessary ->cancel calls in Coro::Handle.
45 - maybe make it work on mingw32 with win32 backend
46 (based on patch by Yasuhiro Matsumoto).
47
485.132 Fri May 29 09:00:39 CEST 2009
49 - do not keep a reference to the argument itself in
50 Coro::Semaphore::guard, as it could change later.
51 - support SO_RCVBUF/SO_SNDBUF nonstandard Coro::Socket options,
52 should support a prepare callback.
53
545.131 Mon Mar 16 23:20:37 CET 2009
55 - implement and document Coro->suspend, Coro->resume.
56 - fix Coro::Select implementation to not (often) close
57 the passed file descriptors (testcase provided by pippijn).
58
595.13 Mon Dec 15 21:51:42 CET 2008
60 - EXPERIMENTAL: implement dynamic winds (on_enter/on_leave).
61 - don't set diehook to C<undef> but instead to NULL, to avoid
62 spurious warnings.
63 - fix a lot of bugs in Coro::SemaphoreSet.
64 - Coro::SemaphoreSet will less often create a semaphore needlessly.
65 - add Coro::SemaphoreSet::count and wait methods.
66 - take advantage of the new Guard module.
67 - deprecate Coro::guard.
68 - try to fix the dreaded 01_unblock tests once more. I hate it when
69 testsuites need more fixing than the code they are supposed to test.
70 - croak in more cases when a required callback isn't resolvable.
71 - fix some minor issues in Coro::State->call/eval.
72 - use current coroutine context instead of a temporary one
73 when temporarily switching to another coroutine.
74 - do not call C-level on_destroy handlers during global destruction,
75 to avoid needless segfaults.
76
775.12 Sun Dec 7 13:30:38 CET 2008
78 - add default config for MirOS, which seems to be a bug-to-bug
79 compatible fork of openbsd ("world domination by releasing
80 openbsd cvs before the openbsd folks do it" or so :).
81 - free_padlist did destroy the names pad, not good, but didn't
82 seem to bother perl - this could fix issues such as eval ""
83 inside a function called from multiple coroutines.
84 - use a different method to detect destruction time.
85 - be more careful when freeing padlists just before global
86 destruction.
87 - fixed and expanded the call/cc example.
88 - renamed _terminate to _coro_run.
89 - new method Coro::Channel->shutdown.
90 - try pthreads on openbsd <4.4 (broken sigaltstack, will
91 pthreads fare better?).
92 - be less picky about destroying "a" running coroutine.
93
945.11 Tue Nov 25 21:49:05 CET 2008
95 - DEBUGGING in 5.10.0 is a mess: it constantly flags perfectly
96 working code with failed assertions, introducing more bugs than
97 it fixes, requiring elaborate workarounds :(
98
995.1 Mon Nov 24 08:54:59 CET 2008
100 - wrote a small introductory tutorial - Coro::Intro.
101 - convert Coro::Timer, Coro::Select and Coro::Util to rouse API.
102 - Coro::Select did errornously dup the file descriptors
103 and didn't work with all AnyEvent backends.
104 - Coro::Select wasn't imported correctly form Coro::LWP, causing blocking
105 LWP data transfers.
106 - disassociate c contexts from coro objects - this is agruably more
107 correct, but mostly allows sharing of cctxs between coro and state
108 objects, for added memory savings and speed increases.
109 - bumped $Coro::POOL_RSS up to 32kb by default.
110 - no longer set the optype to OP_CUSTOM, as B::* understandably
111 doesn't like this very much (and we *are* a type of entersub).
112 - implement state cloning, just to prove that call/cc can be done.
113 - automatically load Coro::AnyEvent in Coro::Handle.
114 - wrap ->cancel calls in eval inside Coro::Handle as EV watchers
115 do not have this method (and don't need it either).
116 - speed up generic anyevent methods in Coro::Handle by using rouse
117 callbacks.
118 - allow coroutines in $Coro::IDLE, speeding up Coro::AnyEvent and
119 others. It also makes the debugger happier, as you can trace
120 through the idle threads now.
121 - add comppad_name* and hints ($^H, %^H) to per-thread variables.
122 - eg/event was pretty broken.
123 - better 5.8.6 compatibility.
124
1255.0 Thu Nov 20 10:35:05 CET 2008
13 - NEW ARCHITECTURE: use the latest 4.x version if you experience 126 - NEW ARCHITECTURE: use the latest 4.x version if you experience
14 stability issues. 127 stability issues.
15 - bump API version to 7 - all dependents must be recompiled. 128 - bump API version to 7 - all dependents must be recompiled.
16 - removed timed_* functions - they were not being used anyways 129 - removed timed_* functions - they were not being used anyways
17 and should be replaced by a more generic mechanism - 130 and should be replaced by a more generic mechanism -
18 and were annoying to support anyways :) 131 and were annoying to support anyways :)
19 - removed SemaphoreSet's waiter method - use sem instead. 132 - removed SemaphoreSet's waiter method - use sem method instead.
20 - Coro::Semaphore->adjust didn't correctly wake up enough waiters. 133 - Coro::Semaphore->adjust didn't correctly wake up enough waiters.
21 - async_pool did free a scalar value twice 134 - async_pool did free a scalar value twice
22 ("Attempt to unreference..."). 135 ("Attempt to unreference...").
23 - fix a disastrous bug in the readline optimisation
24 introduced in 4.801.
25 - fix a longstanding bug where calling terminate on a coro that 136 - fix a longstanding bug where calling terminate on a coro that
26 was waiting for a semaphore that was just becoming available 137 was waiting for a semaphore that was just becoming available
27 would cause a deadlock (semaphore would get into a state where 138 would cause a deadlock (semaphore would get into a state where
28 it was available but waiters were still blocked). 139 it was available but waiters were still blocked).
29 - calling throw on a coroutine that is waiting for a semaphore will 140 - calling throw on a coroutine that is waiting for a semaphore will
30 no longer make it acquire the semaphore (and thus leak a count). 141 no longer make it acquire the semaphore (and thus leak a count).
31 - perl's process emulation is now not even theoretically supported 142 - perl's process emulation is now not even theoretically supported
32 anymore. 143 anymore.
33 - new functions Coro::rouse_cb and Coro::rouse_wait. 144 - new functions Coro::rouse_cb and Coro::rouse_wait for easier
145 conversion of callback-style to blocking-style.
146 - new methods $coro->schedule_to and ->cede_to, to specifically
147 schedule or cede to a specific coroutine.
34 - new function Coro::Semaphore::wait. 148 - new function Coro::Semaphore::wait.
35 - use named constants in Coro::Channel (Richard Hundt). 149 - use named constants in Coro::Channel (Richard Hundt).
36 - directly patch the entersub opcode calling SLF functions (cede, 150 - directly patch the entersub opcode calling SLF functions (cede,
37 transfer and so on). this does speed up context switching, but 151 transfer and so on). this does speed up context switching, but
38 more importanly, it frees us from the hardcoded behaviour of 152 more importanly, it frees us from the hardcoded behaviour of
… …
55 - implement Coro::Channel in terms of Coro::Semaphore, for a moderate 169 - implement Coro::Channel in terms of Coro::Semaphore, for a moderate
56 (in comparison) 20-40% speedup. 170 (in comparison) 20-40% speedup.
57 - used new SLF interface to reimplement Coro::Signal gaining 171 - used new SLF interface to reimplement Coro::Signal gaining
58 some unknown (because I was too lazy), but certain, speedup, and also 172 some unknown (because I was too lazy), but certain, speedup, and also
59 making signals reliable for the first time. 173 making signals reliable for the first time.
174 - used new SLF interface and other optimisations to speed up async_pool
175 by a factor of two. It also doesn't rely on perl's exception mechanism
176 to exit anymore. The overhead for terminating an async_pool, coro over
177 a normal async is now very small.
178 - sped up coroutine creation/destruction by 40%.
60 - forgot to include Coro/libcoro/README in the dist for all these years. 179 - forgot to include Coro/libcoro/README in the dist for all these years.
61 - work around a freebsd pthreads bug (manual testcancel is required as 180 - work around a freebsd pthreads bug (manual testcancel is required as
62 pthread_cond_wait isn't a cancellation point on freebsd). 181 pthread_cond_wait isn't a cancellation point on freebsd).
63 - use new rouse functions to speed up and simplify Coro::BDB. 182 - use new rouse functions to speed up and simplify Coro::BDB.
64 - make "prefer perl native functions" work with threaded perls. 183 - make "prefer perl native functions" work with threaded perls.
65 - condense Coro::Debug ps output, hint at v and w flags. 184 - condense Coro::Debug ps output, hint at v and w flags.
66 - (libcoro) lots of minor cleanups and portability improvements. 185 - (libcoro) lots of minor cleanups and portability improvements.
186
1874.914 Wed Nov 19 12:54:18 CET 2008
188 - fix a disastrous bug in the readline optimisation
189 introduced in 4.801.
67 190
684.913 Sat Nov 15 07:58:28 CET 2008 1914.913 Sat Nov 15 07:58:28 CET 2008
69 - async_pool did free a scalar value twice 192 - async_pool did free a scalar value twice
70 ("Attempt to unreference..."). 193 ("Attempt to unreference...").
71 194

Diff Legend

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