| 1 | Revision history for Perl extension AnyEvent. |
1 | Revision history for Perl extension AnyEvent. |
| 2 | |
2 | |
|
|
3 | TODO: implement condvars... |
|
|
4 | - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl. |
|
|
5 | |
|
|
6 | 3.3 Mon Apr 28 09:51:06 CEST 2008 |
|
|
7 | - added AnyEvent::Handle, AnyEvent::Socket and AnyEvent::Util modules. |
|
|
8 | - fix a bug in the pure perl backend that kept watchers alive |
|
|
9 | when multiple watchers were registered for the same fh/poll combo. |
|
|
10 | - add a benchmark section showing AnyEvent overhead and comparing |
|
|
11 | the different event loops with each other. |
|
|
12 | - prefer pure perl over tk when autoprobing, as it's about as fast, |
|
|
13 | but doesn't crash with many watchers. |
|
|
14 | - declare Qt support non-experimental. |
|
|
15 | - clarify the confusing section about the file descriptor being |
|
|
16 | kept alive. |
|
|
17 | - document the race between loading of an event module and |
|
|
18 | child processes exiting. |
|
|
19 | - support POE as "backend" (with some caveats, POE is not generic |
|
|
20 | enough, and darn slow). |
|
|
21 | - support Wx and Prima through POE. |
|
|
22 | - optimise perl backend to use 20% less memory and take advantage |
|
|
23 | of typical timeout behaviour. It can now compete with |
|
|
24 | select/poll-based C event loops in most cases (it is usually |
|
|
25 | faster than Event and Glib :). |
|
|
26 | - roughly cut EV memory use in half and increase its speed by 30%, |
|
|
27 | by removing undocumented functionality. Did something similar |
|
|
28 | to other event loops where possible. |
|
|
29 | |
|
|
30 | 3.2 Thu Apr 24 10:10:40 CEST 2008 |
|
|
31 | - do not die when anyevent watchers are destroyed while running callbacks |
|
|
32 | in the pureperl backend (could only happen when two watchers are registered |
|
|
33 | for the same fh/poll combo). |
|
|
34 | - support autoloading for child watchers, was broken in all versions. |
|
|
35 | - implement PERL_ANYEVENT_MODEL env variable. |
|
|
36 | - (experimental) implement interface to the (very crashy/buggy) Event::Lib module. |
|
|
37 | - (experimental) implement interface to the Qt module (cannot be autoprobed). |
|
|
38 | - this release is IO:AnyEvent-proof. |
|
|
39 | |
|
|
40 | 3.12 Tue Apr 22 07:11:46 CEST 2008 |
|
|
41 | - reinstate AUTHOR section that got lost somehow. |
|
|
42 | - do not hang in the testsuite with badly broken perls (activestate, |
|
|
43 | strawberry...), but instead diagnose the problem and continue. |
|
|
44 | - use INT instead of CHLD in an attempt to support broken windows perls |
|
|
45 | better (this decreases the test quality, unfortunately). |
|
|
46 | - do not send the signal to the process group |
|
|
47 | (no problem for CHLD, bad for INT :). |
|
|
48 | |
|
|
49 | 3.11 Sat Apr 19 06:57:31 CEST 2008 |
|
|
50 | - major documentation rework. |
|
|
51 | - document the fact that child watchers only watch for zombies. |
|
|
52 | - fix the child watcher example. |
|
|
53 | |
|
|
54 | 3.1 Wed Apr 16 17:09:01 CEST 2008 |
|
|
55 | - work around recurring bugs in Tk by dup'ing filehandles, the |
|
|
56 | only method with good success chances on Tk (the bugs apparently |
|
|
57 | don't get fixed anytime soon). |
|
|
58 | - lift the restriction of only one watcher per fh direction |
|
|
59 | (as the Tk bug workaround also lifts it and only Tk imposed |
|
|
60 | such strong limits). |
|
|
61 | - changed probe order to prefer coro adaptors. |
|
|
62 | - explain why recursion into the event loop is not supported |
|
|
63 | unless the backend supports it (only Coro::EV does without |
|
|
64 | any restrictions...). |
|
|
65 | - add simple manpages for all backend modules. |
|
|
66 | |
|
|
67 | 3.0 Mon Apr 7 21:30:23 CEST 2008 |
|
|
68 | - Coro::Signal changed semantics, roll our own, also cleaning |
|
|
69 | up the Coro implementation in general. |
|
|
70 | - rename Coro backend to CoroEvent. |
|
|
71 | - add some decision helping paragraph to the manpage that should |
|
|
72 | help people to decide whether AnyEvent is the right thing for them. |
|
|
73 | |
|
|
74 | 2.9 Mon Jan 28 13:31:54 CET 2008 |
|
|
75 | - update for EV 3.0 API changes. |
|
|
76 | |
|
|
77 | 2.8 Sun Nov 25 15:06:03 CET 2007 |
|
|
78 | - waitpid can validly return 0. accept this fact of life |
|
|
79 | instead of reporting it to any watchers. |
|
|
80 | |
|
|
81 | 2.7 Fri Nov 23 11:41:14 CET 2007 |
|
|
82 | - force use of AnyEvent::Impl::Perl in testsuite, there is |
|
|
83 | too much breakage outside AnyEvent. |
|
|
84 | - deliver signals synchronously in AnyEvent::Impl::Perl |
|
|
85 | or any other event loop that relies on AnyEvents child |
|
|
86 | watcher emulation. *Could* help with hanging testsuite |
|
|
87 | (except when Event or EV are installed). |
|
|
88 | |
|
|
89 | 2.6 Fri Nov 9 20:36:35 CET 2007 |
|
|
90 | - fix bug in testsuite. |
|
|
91 | - move EV adaptor modules to AnyEvent. |
|
|
92 | - add Coro+EV adaptor module. |
|
|
93 | |
|
|
94 | 2.55 Tue Nov 6 17:41:32 CET 2007 |
|
|
95 | - add EV to the list of supported event models. |
|
|
96 | - do not auto-reset pid watchers, pass pid and status to them. |
|
|
97 | - allow a pid of zero to watch for all children in child watchers. |
|
|
98 | |
|
|
99 | 2.54 Wed Jul 18 17:36:23 CEST 2007 |
|
|
100 | - work around a perl bug that results in |
|
|
101 | BEGIN not safe after errors--compilation aborted without |
|
|
102 | any discernible reason or error message by once |
|
|
103 | again not use'ing strict. |
|
|
104 | |
|
|
105 | 2.53 Sun Jul 8 10:51:53 CEST 2007 |
|
|
106 | - make Glib interface work again (spotted by elmex). |
|
|
107 | |
|
|
108 | 2.52 Wed Mar 7 18:36:16 CET 2007 |
|
|
109 | - child watchers stopped working permanently when |
|
|
110 | all current child watchers were destroyed. |
|
|
111 | |
|
|
112 | 2.51 Mon Dec 11 21:33:24 CET 2006 |
|
|
113 | - work around bugs in perl where eval "require Module" returns |
|
|
114 | true even if the module couldn't be loaded :/. |
|
|
115 | |
|
|
116 | 2.5 Mon Dec 11 02:15:28 CET 2006 |
| 3 | - avoid $AUTOLOAD because many perls corrupt it. |
117 | - avoid $AUTOLOAD because many perls corrupt it. |
|
|
118 | - AnyEvent::detect forces autodetection. |
|
|
119 | - implement signal watchers (experimental) |
|
|
120 | (Tk does not support async signals (<= 804.027 at least), |
|
|
121 | so they might get delayed indefinitely on Tk). |
|
|
122 | - implement child watchers (experimental). |
|
|
123 | - moved default condvar implementation into base module, simplifying |
|
|
124 | most implementation modules. |
| 4 | |
125 | |
| 5 | 2.1 Fri Nov 24 15:50:48 CET 2006 |
126 | 2.1 Fri Nov 24 15:50:48 CET 2006 |
| 6 | - better docs. |
127 | - better docs. |
| 7 | - simple testsuite. |
128 | - simple testsuite. |
| 8 | - added AnyEvent->one_event method for special purposes. |
129 | - added AnyEvent->one_event method for special purposes. |
| … | |
… | |
| 42 | - fixed documentation. |
163 | - fixed documentation. |
| 43 | - clarified some corner cases. |
164 | - clarified some corner cases. |
| 44 | - vastly improved documentation with a long example. |
165 | - vastly improved documentation with a long example. |
| 45 | |
166 | |
| 46 | 0.2 Thu Dec 1 22:18:49 CET 2005 |
167 | 0.2 Thu Dec 1 22:18:49 CET 2005 |
| 47 | - check wether a valid method was called, to avoid endless loops. |
168 | - check whether a valid method was called, to avoid endless loops. |
| 48 | - move condvars into their own class (cleaner, safer). |
169 | - move condvars into their own class (cleaner, safer). |
| 49 | - don't die when some modules are not available. |
170 | - don't die when some modules are not available. |
| 50 | - Tk is too broken to allow for words. Will work once Tk works. |
171 | - Tk is too broken to allow for words. Will work once Tk works. |
| 51 | |
172 | |
| 52 | 0.1 Thu Dec 1 19:44:57 CET 2005 |
173 | 0.1 Thu Dec 1 19:44:57 CET 2005 |
