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

Comparing Coro/Changes (file contents):
Revision 1.259 by root, Fri Oct 5 23:38:39 2007 UTC vs.
Revision 1.348 by root, Fri May 23 00:27:06 2008 UTC

1Revision history for Perl extension Coro. 1Revision history for Perl extension Coro.
2 2
3TODO: should explore PerlIO::via::CoroCede and PerlIO::via::CoroAIO. 3TODO: should explore PerlIO::via::CoroCede and PerlIO::via::CoroAIO.
4TODO: maybe implement a default message channel, very much like Erlang 4TODO: maybe implement a default message channel, very much like Erlang's
5 (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)).
7 7
84.02 8 - fix a bug in Coro::AnyEvent ("Usage: Coro::AnyEvent::_schedule()").
9 - take advantage of async name resolution of AnyEvent::Util.
10 - work around brutal inet_aton override in Coro::LWP.
11 - take advantage of the readyhook in Coro::EV, for smoother
12 scheduling.
13
144.7 Sun May 11 00:32:19 CEST 2008
15 - completely reworked the Coro manpage.
16 - added Coro::AnyEvent, generic event loop integration.
17 - implement cancel, ready and kill commands in Coro::Debug.
18 - document find_coro in Coro::Debug.
19 - incompatible change: rename has_stack to has_cctx.
20 - Coro::AIO and Coro::BDB no longer force event model detection,
21 use AnyEvent::AIO and AnyEvent::BDB.
22
234.6 Sat Apr 26 10:05:14 CEST 2008
24 - INCOMPATIBLE CHANGE: sub/code attributes are no longer supported
25 by the Coro module. It was a mistake to have it in the first place.
26 - (experimental) support for activestate perl 5.10 (method "w").
27 - (experimental) support for strawberry perl (method "a").
28 - coro_sigelem_set did not return a value although it had to,
29 actual impact unknown.
30
314.51 Mon Apr 14 13:28:27 CEST 2008
32 - make it compile again on 5.8.
33
344.50 Thu Apr 10 09:43:17 CEST 2008
35 - I did it twice! (see 4.49).
36
374.49 Sun Apr 6 21:23:31 CEST 2008
38 - grr, instead of compiling the recent changes on 5.10 only they
39 were compiled on 5.8 only.
40
414.48 Sun Apr 6 20:36:46 CEST 2008
42 - allow coroutine switches during eval's under 5.10.x, as it seems
43 the parser is a per-interpreter option now, so this is safe
44 (this might fix the odd crashing bug).
45 - drop support for 5.9.x versions: they are dead, jim.
46
474.47 Sun Apr 6 00:37:52 CEST 2008
48 - force cctx allocation on API calls: we know we need to force one
49 and gcc actually manages to confuse our heuristic nowadays,
50 leading to crashes and worse.
51 - document force_cctx.
52
534.46 Fri Apr 4 22:05:43 CEST 2008
54 - upgrade libcoro, resulting in pthread-backend (which was only created
55 to fulfill the rules of the programming languages shootout).
56
574.45 Thu Mar 13 11:55:36 CET 2008
58 - fix a file leaking bug in eg/mhyttpd that would allow
59 downloading of any file (reported by oesi).
60 - fix deadlock bug in Coro::Channel (reported by Richard Hundt)
61 (also add testcase).
62 - support Broadcast option in Coro::Socket (patch by Richard Hundt,
63 apparently having loads of fun with that).
64 - implement $state->swap_defsv, ->swap_defav and document ->throw.
65
664.4 Wed Feb 13 16:44:29 CET 2008
67 - only assume .cfi pseudo isns exist on GNU/Linux.
68 - add get_timed method to Coro::Channels.
69 - fixed Coro::Debug manpage.
70 - perl 5.11 compatbility improvement based on patch by
71 Andreas König.
72
734.37 Sun Jan 20 11:25:23 CET 2008
74 - soften the check for an active parser for perl 5.10 (avoiding
75 "Coro::State::transfer called while parsing" in many cases).
76
774.36 Sun Jan 13 10:53:56 CET 2008
78 - reset diehook when terminating from an async_pool as to not
79 trigger any __DIE__ handlers.
80
814.35 Sun Jan 13 04:14:13 CET 2008
82 - "bt" debug command now displays any exceptions
83 from longmess and also skips the innermost
84 stackframes, giving more useufl output.
85 - allow backtraces in coroutines blocked in Coro::EV,
86 at a <1% speed hit.
87 - handle localising of $SIG{__WARN__} and __DIE__
88 properly (with a proper amount of dirty hacking).
89
904.34 Sat Dec 22 17:49:53 CET 2007
91 - upgrade to EV version 2.0 API.
92
934.33 Mon Dec 17 08:36:12 CET 2007
94 - make Coro::AIO etc. loadable in the absence of EV.
95
964.32 Mon Dec 17 07:46:02 CET 2007
97 - majorly improved Coro::Handle's performance with Coro::EV.
98 - implemented the until now mythical Coro::BDB module.
99 - specialcase EV in Coro::AIO and Coro::BDB for extra speed.
100
1014.31 Wed Dec 5 12:32:39 CET 2007
102 - remove warn statement form Coro::Util.
103
1044.3 Tue Dec 4 20:33:14 CET 2007
105 - calls to the idle function could cause stack corruption
106 when the stack changed.
107 - do no longer rely on the presence of EV::DNS if EV is used
108 (because it is gone), but instead take avdantage of EV::ADNS
109 if available.
110 - add ($) prototypes to all functions in Coro::Storable.
111 - use a conventional (safer) idle callback in Coro::EV.
112 - do accept quickly in Coro::Debug to avoid endless loops.
113
1144.22 Fri Nov 30 16:04:04 CET 2007
115 - really use optimised versions for Event and EV in Coro::Util
116 and Coro::Handle.
117
1184.21 Sun Nov 25 10:48:59 CET 2007
119 - fix a spurious memory read.
120 - Coro::EV no longer keeps the eventloop "alive".
121
1224.2 Fri Nov 9 20:47:05 CET 2007
123 - enable/disable tracing from a new coroutine, not a pooled one.
124 - fix a memleak in Coro::Event.
125 - removed killall call from fork_eval.
126 - made sure store_fd is already loaded so that fork_eval
127 does not have to parse autoload in each subprocess.
128 - only use assembly method if -O switch is in $Config{optimize}.
129 - add (optional) Coro::EV module, so far the best event loop module
130 directly supported by Coro.
131 - if the event model is EV, use EV::DNS to resolve
132 stuff in Coro::Util.
133 - don't get confused by multiple event notifications in Coro::Handle.
134 - initial support for EV (libevent interface).
135 - require Event and EV using configure_requires, to force their existance.
136
1374.13 Wed Oct 24 07:26:45 CEST 2007
138 - add Coro::Storable::blocking_thaw.
139 - use a vastly more complicated technique to localise
140 $SIG{__WARN/DIE__} that also works on perls <= 5.8.8.
141 - use a coroutine for the idle callback Coro::Event,
142 instead of running Event in the current coroutine context.
143 This also catches recursive invocations.
144 - actually report fork errors in gethostbyname and inet_aton.
145
1464.11 Thu Oct 11 02:40:24 CEST 2007
147 - port to threaded perls.
148
1494.1 Thu Oct 11 02:38:16 CEST 2007
150 - incompatible change: $SIG{__DIE__} and $SIG{__WARN__} will now
151 be local to each coro (see Coro::State).
152 - incompatible change: for very deep reasons, cede and cede_notself
153 cannot return anything, so nothing will be returned.
154 - possibly bring back 5.10 compatibility (untested).
155 - work around stupid (and wrong) warning on 5.10 :(.
156 - overlay the saved state over the context stack. This saves
157 a few hundred bytes per coroutine on average and also
158 speeds up context switching a bit.
159 - further tune default stack sizes.
160 - (more) correctly calculate stack usage in coro_rss.
161 - Coro::Storable::blocking_* did not properly lock
162 resulting in races between coroutines.
163 - added Coro::Storable::guard.
164 - stopping to trace a coroutine could destroy the cctx of
165 an unrelated coroutine.
166 - explain the relationship between Perl and C coroutines in
167 more detail in Coro::State.
168 - Coro::Util::inet_aton did not short-circuit dotted quad forms,
169 causing a fork per resolve. This also affected Coro::Socket.
170 - switch to a separate stack in $coro->call/eval to avoid
171 invalidating pointers.
172
1734.03 Sat Oct 6 21:24:00 CEST 2007
174 - added Coro::throw method.
175 - minor code cleanups.
176
1774.02 Sat Oct 6 02:36:47 CEST 2007
9 - fix a very minor per-coroutine memleak (a single codereference). 178 - fix a very minor per-coroutine memleak (a single codereference).
10 - fixed a bug where the currently in-use c context would be freed 179 - fixed a bug where the currently in-use c context would be freed
11 prematurely (can happen only when programs change the stacksize 180 prematurely (can happen only when programs change the stacksize
12 or use tracing). 181 or use tracing).
13 - tracing can no longer keep a coro alive after it terminated. 182 - tracing can no longer keep a coro alive after it terminated.
183 - do static branch prediction in the common path for gcc. gives
184 about 2-5% speed improvement here.
14 185
154.01 Fri Oct 5 22:10:49 CEST 2007 1864.01 Fri Oct 5 22:10:49 CEST 2007
16 - instead of recreating *a* standard output handle we simply 187 - instead of recreating *a* standard output handle we simply
17 use STDOUT, which is faster and hopefully more robust. 188 use STDOUT, which is faster and hopefully more robust.
18 189

Diff Legend

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