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 1.342 by root, Fri May 9 22:04:37 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 TODO: Coro::AnyEvent
9 - completely reworked the Coro manpage.
10 - added Coro::AnyEvent, generic event loop integration.
11 - implement cancel, ready and kill commands in Coro::Debug.
12 - document find_coro in Coro::Debug.
13
144.6 Sat Apr 26 10:05:14 CEST 2008
15 - INCOMPATIBLE CHANGE: sub/code attributes are no longer supported
16 by the Coro module. It was a mistake to have it in the first place.
17 - (experimental) support for activestate perl 5.10 (method "w").
18 - (experimental) support for strawberry perl (method "a").
19 - coro_sigelem_set did not return a value although it had to,
20 actual impact unknown.
21
224.51 Mon Apr 14 13:28:27 CEST 2008
23 - make it compile again on 5.8.
24
254.50 Thu Apr 10 09:43:17 CEST 2008
26 - I did it twice! (see 4.49).
27
284.49 Sun Apr 6 21:23:31 CEST 2008
29 - grr, instead of compiling the recent changes on 5.10 only they
30 were compiled on 5.8 only.
31
324.48 Sun Apr 6 20:36:46 CEST 2008
33 - allow coroutine switches during eval's under 5.10.x, as it seems
34 the parser is a per-interpreter option now, so this is safe
35 (this might fix the odd crashing bug).
36 - drop support for 5.9.x versions: they are dead, jim.
37
384.47 Sun Apr 6 00:37:52 CEST 2008
39 - force cctx allocation on API calls: we know we need to force one
40 and gcc actually manages to confuse our heuristic nowadays,
41 leading to crashes and worse.
42 - document force_cctx.
43
444.46 Fri Apr 4 22:05:43 CEST 2008
45 - upgrade libcoro, resulting in pthread-backend (which was only created
46 to fulfill the rules of the programming languages shootout).
47
484.45 Thu Mar 13 11:55:36 CET 2008
49 - fix a file leaking bug in eg/mhyttpd that would allow
50 downloading of any file (reported by oesi).
51 - fix deadlock bug in Coro::Channel (reported by Richard Hundt)
52 (also add testcase).
53 - support Broadcast option in Coro::Socket (patch by Richard Hundt,
54 apparently having loads of fun with that).
55 - implement $state->swap_defsv, ->swap_defav and document ->throw.
56
574.4 Wed Feb 13 16:44:29 CET 2008
58 - only assume .cfi pseudo isns exist on GNU/Linux.
59 - add get_timed method to Coro::Channels.
60 - fixed Coro::Debug manpage.
61 - perl 5.11 compatbility improvement based on patch by
62 Andreas König.
63
644.37 Sun Jan 20 11:25:23 CET 2008
65 - soften the check for an active parser for perl 5.10 (avoiding
66 "Coro::State::transfer called while parsing" in many cases).
67
684.36 Sun Jan 13 10:53:56 CET 2008
69 - reset diehook when terminating from an async_pool as to not
70 trigger any __DIE__ handlers.
71
724.35 Sun Jan 13 04:14:13 CET 2008
73 - "bt" debug command now displays any exceptions
74 from longmess and also skips the innermost
75 stackframes, giving more useufl output.
76 - allow backtraces in coroutines blocked in Coro::EV,
77 at a <1% speed hit.
78 - handle localising of $SIG{__WARN__} and __DIE__
79 properly (with a proper amount of dirty hacking).
80
814.34 Sat Dec 22 17:49:53 CET 2007
82 - upgrade to EV version 2.0 API.
83
844.33 Mon Dec 17 08:36:12 CET 2007
85 - make Coro::AIO etc. loadable in the absence of EV.
86
874.32 Mon Dec 17 07:46:02 CET 2007
88 - majorly improved Coro::Handle's performance with Coro::EV.
89 - implemented the until now mythical Coro::BDB module.
90 - specialcase EV in Coro::AIO and Coro::BDB for extra speed.
91
924.31 Wed Dec 5 12:32:39 CET 2007
93 - remove warn statement form Coro::Util.
94
954.3 Tue Dec 4 20:33:14 CET 2007
96 - calls to the idle function could cause stack corruption
97 when the stack changed.
98 - do no longer rely on the presence of EV::DNS if EV is used
99 (because it is gone), but instead take avdantage of EV::ADNS
100 if available.
101 - add ($) prototypes to all functions in Coro::Storable.
102 - use a conventional (safer) idle callback in Coro::EV.
103 - do accept quickly in Coro::Debug to avoid endless loops.
104
1054.22 Fri Nov 30 16:04:04 CET 2007
106 - really use optimised versions for Event and EV in Coro::Util
107 and Coro::Handle.
108
1094.21 Sun Nov 25 10:48:59 CET 2007
110 - fix a spurious memory read.
111 - Coro::EV no longer keeps the eventloop "alive".
112
1134.2 Fri Nov 9 20:47:05 CET 2007
114 - enable/disable tracing from a new coroutine, not a pooled one.
115 - fix a memleak in Coro::Event.
116 - removed killall call from fork_eval.
117 - made sure store_fd is already loaded so that fork_eval
118 does not have to parse autoload in each subprocess.
119 - only use assembly method if -O switch is in $Config{optimize}.
120 - add (optional) Coro::EV module, so far the best event loop module
121 directly supported by Coro.
122 - if the event model is EV, use EV::DNS to resolve
123 stuff in Coro::Util.
124 - don't get confused by multiple event notifications in Coro::Handle.
125 - initial support for EV (libevent interface).
126 - require Event and EV using configure_requires, to force their existance.
127
1284.13 Wed Oct 24 07:26:45 CEST 2007
129 - add Coro::Storable::blocking_thaw.
130 - use a vastly more complicated technique to localise
131 $SIG{__WARN/DIE__} that also works on perls <= 5.8.8.
132 - use a coroutine for the idle callback Coro::Event,
133 instead of running Event in the current coroutine context.
134 This also catches recursive invocations.
135 - actually report fork errors in gethostbyname and inet_aton.
136
1374.11 Thu Oct 11 02:40:24 CEST 2007
138 - port to threaded perls.
139
1404.1 Thu Oct 11 02:38:16 CEST 2007
141 - incompatible change: $SIG{__DIE__} and $SIG{__WARN__} will now
142 be local to each coro (see Coro::State).
143 - incompatible change: for very deep reasons, cede and cede_notself
144 cannot return anything, so nothing will be returned.
145 - possibly bring back 5.10 compatibility (untested).
146 - work around stupid (and wrong) warning on 5.10 :(.
147 - overlay the saved state over the context stack. This saves
148 a few hundred bytes per coroutine on average and also
149 speeds up context switching a bit.
150 - further tune default stack sizes.
151 - (more) correctly calculate stack usage in coro_rss.
152 - Coro::Storable::blocking_* did not properly lock
153 resulting in races between coroutines.
154 - added Coro::Storable::guard.
155 - stopping to trace a coroutine could destroy the cctx of
156 an unrelated coroutine.
157 - explain the relationship between Perl and C coroutines in
158 more detail in Coro::State.
159 - Coro::Util::inet_aton did not short-circuit dotted quad forms,
160 causing a fork per resolve. This also affected Coro::Socket.
161 - switch to a separate stack in $coro->call/eval to avoid
162 invalidating pointers.
163
1644.03 Sat Oct 6 21:24:00 CEST 2007
165 - added Coro::throw method.
166 - minor code cleanups.
167
1684.02 Sat Oct 6 02:36:47 CEST 2007
9 - fix a very minor per-coroutine memleak (a single codereference). 169 - fix a very minor per-coroutine memleak (a single codereference).
10 - fixed a bug where the currently in-use c context would be freed 170 - fixed a bug where the currently in-use c context would be freed
11 prematurely (can happen only when programs change the stacksize 171 prematurely (can happen only when programs change the stacksize
12 or use tracing). 172 or use tracing).
13 - tracing can no longer keep a coro alive after it terminated. 173 - tracing can no longer keep a coro alive after it terminated.
174 - do static branch prediction in the common path for gcc. gives
175 about 2-5% speed improvement here.
14 176
154.01 Fri Oct 5 22:10:49 CEST 2007 1774.01 Fri Oct 5 22:10:49 CEST 2007
16 - instead of recreating *a* standard output handle we simply 178 - instead of recreating *a* standard output handle we simply
17 use STDOUT, which is faster and hopefully more robust. 179 use STDOUT, which is faster and hopefully more robust.
18 180

Diff Legend

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