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

Comparing Coro/Changes (file contents):
Revision 1.457 by root, Sat Nov 22 02:09:54 2008 UTC vs.
Revision 1.544 by root, Thu Nov 11 00:09:08 2010 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: maybe implement a default message channel, very much like Erlang's 4TODO: unready_all
5 actor model (which is cool in a lot of important aspects (failures!), 5TODO: myhttpd header parsing
6 but very lacking in others (higher level ipc)). 6TODO: channel->maxsize(newsize)?
7TODO: http://www.microsoft.com/msj/archive/s2ce.aspx
7 8
95.25 Thu Nov 11 01:08:39 CET 2010
10 - try a different approach on netbsd - netbsd 5 finally has marginally
11 working pthreads, but still broken ucontext/sigaltstack.
12 - openbsd 4.8 finally got their act together, Coro works out of the box
13 with asm, setjmp and pthreads (no change, just informational).
14
155.24 Sat Oct 23 11:27:12 CEST 2010
16 - port to the EV 4.0 API.
17 - work around bugs in mingw32, making strawberry perl work
18 out of the box.
19 - correctly modify Coro::AIO function prototypes
20 so that they reflect the "no optional parameters" rule.
21 - "ported" libcoro to C++.
22
235.23 Mon May 17 18:50:42 CEST 2010
24 - be more resistant to ordering changes when initialising
25 Coro::AnyEvent, Coro::EV and Coro::Event (reported by Matthias
26 Waldorf).
27 - document that perl 5.12 deliberately removed support for cloning.
28
295.22 Wed Apr 14 03:55:35 CEST 2010
30 - correctly return udnef on errors in Coro::Handle::read/write
31 (testcase by Marc Mims).
32 - convert Coro::Util into a "perl compatibility wrapper" - the functions
33 are less useful now, but are drop-in replacements for existing
34 functions, listing better alternatives in the documentation. This also
35 fixes a bug in Coro::LWP which naively substituted Socket::inet_aton
36 with Coro::Util::inet_aton.
37 - do not override $Coro::idle unconditionally in Coro.pm, as other
38 modules could have provided their own idle coro already
39 (for exmaple, Coro::AnyEvent).
40 - fix Coro::Util::gethost* functions.
41 - Coro::Timer corretcly exports it's symbols (reported by Hideki Yamamura).
42
435.21 Wed Dec 16 07:19:51 CET 2009
44 - automatically load Coro::AnyEvent when AnyEvent and Coro are used
45 together.
46 - add some examples on how to combine other event loops with Coro in
47 Coro::AnyEvent, and how to run it (and not to block). Seems to be
48 the most common source of confusion.
49 - try to catch people naively blocking in an event callback.
50 - work around the perl filehandle bug issue in conjunction with
51 older common::sense (as indirectly pointed out by ZSystem).
52 - clarify the "not from signal handlers" section.
53
545.2 Sun Oct 4 14:54:24 CEST 2009
55 - Coro::Storable destroyed the prototypes of the functions it wrapped.
56 - export rouse_cb and rouse_wait by default now.
57 - fix various prototype mismatches in Coro::AnyEvent and Coro::Handle.
58 - new method $state->swap_sv.
59 - added section on "windows process emulation" to the manpage,
60 after a not-so-fruitful (nor-friendly) "discussion" with chip
61 salzenberg (discussion implies arguments, but his only arguments
62 were ad-hominems, one wonders why he started it in the first
63 place). I hope this explains it well enough for him to understand,
64 and maybe well enough for others to understand.
65 - use common::sense everywhere now.
66 - idle callbacks are no longer supported, use idle coros instead.
67 - print a thread listing when a deadlock is detected.
68
695.17 Sat Aug 22 23:09:31 CEST 2009
70 - work around a bug in the perl debugger causing crashes
71 when running under the debugger by marking _pool_handler as nodebug.
72 - speed up Coro::async considerably.
73 - try some hacks to get netbsd to work "more often" - their broken
74 setjmp/longjmp, ucontext *and* phtreads are really hard on Coro.
75 - convert Coro to AE 5.0 API.
76
775.162 Tue Jul 28 04:04:03 CEST 2009
78 - perl 5.8.2 is now minimum requirement.
79 - skip t/19_handle.t on broken windows perls.
80
815.161 Wed Jul 22 04:47:38 CEST 2009
82 - Coro::AnyEvent::poll could have a different prototype when EV was
83 used as backend (analyzed by Tatsuhiko Miyagawa).
84 - Coro::AnyEvent errornously initialised the event loop when loaded,
85 not on demand.
86 - try to workaround rare */t/01_unblock.t failures.
87
885.16 Tue Jul 21 01:44:37 CEST 2009
89 - Coro::AnyEvent failed to hook into the event loop
90 when no threads had been readied between detecting
91 the event loop and actually running it.
92 - considerably speed up Coro::Select by taking avdantage
93 of AnyEvent > 4.8 and some other optimisations.
94 - implement paragraph readline mode in Coro::Handle
95 (based on patches by Zsbán Ambrus).
96 - replace WSAEINPROGRESS by WSAEWOULDBLOCK (reported
97 and analyzed by Yasuhiro MATSUMOTO).
98 - clarified libcoro license and copyright.
99 - someone stole my EXTRA_META!!!
100 - implement Coro::Select::patch_pp_sselect and it's brother,
101 for hardcode select overriding.
102
1035.151 Mon Jul 6 05:41:57 CEST 2009
104 - backport to windows process emulation code again (patch by
105 Yasuhiro MATSUMOTO).
106 - slightly update Coro::MakeMaker.
107
1085.15 Tue Jun 30 10:28:06 CEST 2009
109 - deprecate Coro::Socket, document how to get ipv6 support via
110 AnyEvent::Socket instead.
111 - implement signal->wait ($cb) interface, similar to semaphores.
112 - work around SvOK not supporting getmagic, so we have to getmagic
113 to test for undef :/ (reported by Matthias Waldorf).
114 - load Coro::AnyEvent in all modules using AnyEvent.
115 - work around perl corrupting our internal data structures,
116 reported by Tokuhiro Matsuno.
117 - enable per-coroutine real and cpu time gathering
118 (Coro::State::enable_times).
119
1205.14 Wed Jun 24 01:37:48 CEST 2009
121 - provide explicit functions to "cede" to the event loop in Coro::AnyEvent,
122 as this seems to have been a difficult concept (poll, sleep, idle,
123 idle_upto).
124 - add Coro::AnyEvent::readable/writable functions.
125 - clarify Coro::EV/Event/AnyEvent manpages.
126 - free per-thread global scalars in the thread calling ->cancel, to
127 avoid crashes when $_, $@ etc., are magical but some of those
128 had already been freed.
129 - "unexperimentalise" the callback interface for Coro::Semaphore.
130 - speed up ready queue management/context switching by using a linked
131 list instead of an array (~5-10%).
132 - implement "watch" command in Coro::Debug shells.
133 - for fun, implement time-slicing as an example in the manpage.
134 - if AnyEvent detects EV or Event, but we don't have Coro::EV or
135 Coro::Event, use the normal AnyEvent handling instead of dieing
136 (the same is true for Coro::Handle).
137 - properly document Coro::EV::timed_io_once.
138 - avoid unneccessary ->cancel calls in Coro::Handle.
139 - maybe make it work on mingw32 with win32 backend
140 (based on patch by Yasuhiro Matsumoto).
141
1425.132 Fri May 29 09:00:39 CEST 2009
143 - do not keep a reference to the argument itself in
144 Coro::Semaphore::guard, as it could change later.
145 - support SO_RCVBUF/SO_SNDBUF nonstandard Coro::Socket options,
146 should support a prepare callback.
147
1485.131 Mon Mar 16 23:20:37 CET 2009
149 - implement and document Coro->suspend, Coro->resume.
150 - fix Coro::Select implementation to not (often) close
151 the passed file descriptors (testcase provided by pippijn).
152
1535.13 Mon Dec 15 21:51:42 CET 2008
154 - EXPERIMENTAL: implement dynamic winds (on_enter/on_leave).
155 - don't set diehook to C<undef> but instead to NULL, to avoid
156 spurious warnings.
157 - fix a lot of bugs in Coro::SemaphoreSet.
158 - Coro::SemaphoreSet will less often create a semaphore needlessly.
159 - add Coro::SemaphoreSet::count and wait methods.
160 - take advantage of the new Guard module.
161 - deprecate Coro::guard.
162 - try to fix the dreaded 01_unblock tests once more. I hate it when
163 testsuites need more fixing than the code they are supposed to test.
164 - croak in more cases when a required callback isn't resolvable.
165 - fix some minor issues in Coro::State->call/eval.
166 - use current coroutine context instead of a temporary one
167 when temporarily switching to another coroutine.
168 - do not call C-level on_destroy handlers during global destruction,
169 to avoid needless segfaults.
170
1715.12 Sun Dec 7 13:30:38 CET 2008
172 - add default config for MirOS, which seems to be a bug-to-bug
173 compatible fork of openbsd ("world domination by releasing
174 openbsd cvs before the openbsd folks do it" or so :).
175 - free_padlist did destroy the names pad, not good, but didn't
176 seem to bother perl - this could fix issues such as eval ""
177 inside a function called from multiple coroutines.
178 - use a different method to detect destruction time.
179 - be more careful when freeing padlists just before global
180 destruction.
181 - fixed and expanded the call/cc example.
182 - renamed _terminate to _coro_run.
183 - new method Coro::Channel->shutdown.
184 - try pthreads on openbsd <4.4 (broken sigaltstack, will
185 pthreads fare better?).
186 - be less picky about destroying "a" running coroutine.
187
1885.11 Tue Nov 25 21:49:05 CET 2008
189 - DEBUGGING in 5.10.0 is a mess: it constantly flags perfectly
190 working code with failed assertions, introducing more bugs than
191 it fixes, requiring elaborate workarounds :(
192
1935.1 Mon Nov 24 08:54:59 CET 2008
194 - wrote a small introductory tutorial - Coro::Intro.
8 - convert Coro::Timer, Coro::Select and Coro::Util to rouse API. 195 - convert Coro::Timer, Coro::Select and Coro::Util to rouse API.
9 - Coro::Select did errornously dup the file descriptors 196 - Coro::Select did errornously dup the file descriptors
10 and didn't work with all AnyEvent backends. 197 and didn't work with all AnyEvent backends.
198 - Coro::Select wasn't imported correctly form Coro::LWP, causing blocking
199 LWP data transfers.
200 - disassociate c contexts from coro objects - this is agruably more
201 correct, but mostly allows sharing of cctxs between coro and state
202 objects, for added memory savings and speed increases.
11 - bumped $Coro::POOL_RSS up to 32kb by default. 203 - bumped $Coro::POOL_RSS up to 32kb by default.
12 - no longer set the optype to OP_CUSTOM, as B::* understandably 204 - no longer set the optype to OP_CUSTOM, as B::* understandably
13 doesn't like this very much (and we *are* a type of entersub). 205 doesn't like this very much (and we *are* a type of entersub).
206 - implement state cloning, just to prove that call/cc can be done.
207 - automatically load Coro::AnyEvent in Coro::Handle.
208 - wrap ->cancel calls in eval inside Coro::Handle as EV watchers
209 do not have this method (and don't need it either).
210 - speed up generic anyevent methods in Coro::Handle by using rouse
211 callbacks.
212 - allow coroutines in $Coro::IDLE, speeding up Coro::AnyEvent and
213 others. It also makes the debugger happier, as you can trace
214 through the idle threads now.
215 - add comppad_name* and hints ($^H, %^H) to per-thread variables.
216 - eg/event was pretty broken.
14 - better 5.8.6 compatibility. 217 - better 5.8.6 compatibility.
15 218
165.0 Thu Nov 20 10:35:05 CET 2008 2195.0 Thu Nov 20 10:35:05 CET 2008
17 - NEW ARCHITECTURE: use the latest 4.x version if you experience 220 - NEW ARCHITECTURE: use the latest 4.x version if you experience
18 stability issues. 221 stability issues.

Diff Legend

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