1 | Revision history for Perl extension AnyEvent. |
1 | Revision history for Perl extension AnyEvent. |
|
|
2 | |
|
|
3 | 3.51 |
|
|
4 | - added (experimental) AnyEvent::DNS module. |
|
|
5 | - make on_eof optional in AnyEvent::Handle. |
|
|
6 | - support service names in tcp_connect and tcp_server. |
|
|
7 | - work around netbsd/openbsd bug causing unix sockets not to |
|
|
8 | become writable untill empty. |
|
|
9 | - only make a dummy read for error detection when getpeername |
|
|
10 | returns ENOTCONN. |
|
|
11 | - defend AnyEvent::Util::inet_aton against Coro::LWP brutal |
|
|
12 | override. |
|
|
13 | |
|
|
14 | 3.5 Sat May 17 23:17:14 CEST 2008 |
|
|
15 | - gracefully deal with systems not defining CLOCK_MONOTONIC |
|
|
16 | (as opposed to not supporting the clock itself, which worked in 3.41). |
|
|
17 | - added AnyEvent::Util::tcp_connect and AnyEvent::Util::tcp_server |
|
|
18 | helpers, to replace the ill-designed AnyEvent::Socket class. |
|
|
19 | - added some examples (eg/connect, eg/handle). |
|
|
20 | - added AnyEvent::Util::guard. |
|
|
21 | - AnyEvent::Util now exports most of its functions. |
|
|
22 | - fixed a bug in *_read_chunk, where the callback was called with $self as |
|
|
23 | first and second argument. |
|
|
24 | - simplified and fixed regex handling in AnyEvent::Handle::*_read_line. |
|
|
25 | - fix manpage condvars examples to use new syntax. |
|
|
26 | - considerably improve inet_aton with EV::ADNS backend to support |
|
|
27 | cname chains (common dns configuration bug, e.g. with www.google.de) |
|
|
28 | and aaaa records if no a record could be found. |
|
|
29 | |
|
|
30 | 3.41 Sun May 11 19:53:13 CEST 2008 |
|
|
31 | - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao. |
|
|
32 | |
|
|
33 | 3.4 Sun May 11 00:29:25 CEST 2008 |
|
|
34 | - complete re-implementation of condvars with a hopefully much mroe useful |
|
|
35 | API (backward compatible functions exist but are undocumented). |
|
|
36 | - AnyEvent::Handle fully rewritten with a hopefully more powerful API |
|
|
37 | (still subject to change). |
|
|
38 | - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl. |
|
|
39 | - provide AnyEvent::post_detect and @AnyEvent::post_detect, which |
|
|
40 | allows module authors to avoid forcing event loop detection. |
|
|
41 | (used by Coro::AIO, Coro::BDB, Coro::AnyEvent for example). |
|
|
42 | - remove coro backends: Coro now provides generic support for AnyEvent |
|
|
43 | via Coro::AnyEvent. |
|
|
44 | |
|
|
45 | 3.3 Mon Apr 28 09:51:06 CEST 2008 |
|
|
46 | - added AnyEvent::Handle, AnyEvent::Socket and AnyEvent::Util modules. |
|
|
47 | - fix a bug in the pure perl backend that kept watchers alive |
|
|
48 | when multiple watchers were registered for the same fh/poll combo. |
|
|
49 | - add a benchmark section showing AnyEvent overhead and comparing |
|
|
50 | the different event loops with each other. |
|
|
51 | - prefer pure perl over tk when autoprobing, as it's about as fast, |
|
|
52 | but doesn't crash with many watchers. |
|
|
53 | - declare Qt support non-experimental. |
|
|
54 | - clarify the confusing section about the file descriptor being |
|
|
55 | kept alive. |
|
|
56 | - document the race between loading of an event module and |
|
|
57 | child processes exiting. |
|
|
58 | - support POE as "backend" (with some caveats, POE is not generic |
|
|
59 | enough, and darn slow). |
|
|
60 | - support Wx and Prima through POE. |
|
|
61 | - optimise perl backend to use 20% less memory and take advantage |
|
|
62 | of typical timeout behaviour. It can now compete with |
|
|
63 | select/poll-based C event loops in most cases (it is usually |
|
|
64 | faster than Event and Glib :). |
|
|
65 | - roughly cut EV memory use in half and increase its speed by 30%, |
|
|
66 | by removing undocumented functionality. Did something similar |
|
|
67 | to other event loops where possible. |
|
|
68 | |
|
|
69 | 3.2 Thu Apr 24 10:10:40 CEST 2008 |
|
|
70 | - do not die when anyevent watchers are destroyed while running callbacks |
|
|
71 | in the pureperl backend (could only happen when two watchers are registered |
|
|
72 | for the same fh/poll combo). |
|
|
73 | - support autoloading for child watchers, was broken in all versions. |
|
|
74 | - implement PERL_ANYEVENT_MODEL env variable. |
|
|
75 | - (experimental) implement interface to the (very crashy/buggy) Event::Lib module. |
|
|
76 | - (experimental) implement interface to the Qt module (cannot be autoprobed). |
|
|
77 | - this release is IO:AnyEvent-proof. |
|
|
78 | |
|
|
79 | 3.12 Tue Apr 22 07:11:46 CEST 2008 |
|
|
80 | - reinstate AUTHOR section that got lost somehow. |
|
|
81 | - do not hang in the testsuite with badly broken perls (activestate, |
|
|
82 | strawberry...), but instead diagnose the problem and continue. |
|
|
83 | - use INT instead of CHLD in an attempt to support broken windows perls |
|
|
84 | better (this decreases the test quality, unfortunately). |
|
|
85 | - do not send the signal to the process group |
|
|
86 | (no problem for CHLD, bad for INT :). |
|
|
87 | |
|
|
88 | 3.11 Sat Apr 19 06:57:31 CEST 2008 |
|
|
89 | - major documentation rework. |
|
|
90 | - document the fact that child watchers only watch for zombies. |
|
|
91 | - fix the child watcher example. |
|
|
92 | |
|
|
93 | 3.1 Wed Apr 16 17:09:01 CEST 2008 |
|
|
94 | - work around recurring bugs in Tk by dup'ing filehandles, the |
|
|
95 | only method with good success chances on Tk (the bugs apparently |
|
|
96 | don't get fixed anytime soon). |
|
|
97 | - lift the restriction of only one watcher per fh direction |
|
|
98 | (as the Tk bug workaround also lifts it and only Tk imposed |
|
|
99 | such strong limits). |
|
|
100 | - changed probe order to prefer coro adaptors. |
|
|
101 | - explain why recursion into the event loop is not supported |
|
|
102 | unless the backend supports it (only Coro::EV does without |
|
|
103 | any restrictions...). |
|
|
104 | - add simple manpages for all backend modules. |
|
|
105 | |
|
|
106 | 3.0 Mon Apr 7 21:30:23 CEST 2008 |
|
|
107 | - Coro::Signal changed semantics, roll our own, also cleaning |
|
|
108 | up the Coro implementation in general. |
|
|
109 | - rename Coro backend to CoroEvent. |
|
|
110 | - add some decision helping paragraph to the manpage that should |
|
|
111 | help people to decide whether AnyEvent is the right thing for them. |
|
|
112 | |
|
|
113 | 2.9 Mon Jan 28 13:31:54 CET 2008 |
|
|
114 | - update for EV 3.0 API changes. |
2 | |
115 | |
3 | 2.8 Sun Nov 25 15:06:03 CET 2007 |
116 | 2.8 Sun Nov 25 15:06:03 CET 2007 |
4 | - waitpid can validly return 0. accept this fact of life |
117 | - waitpid can validly return 0. accept this fact of life |
5 | instead of reporting it to any watchers. |
118 | instead of reporting it to any watchers. |
6 | |
119 | |
… | |
… | |
89 | - fixed documentation. |
202 | - fixed documentation. |
90 | - clarified some corner cases. |
203 | - clarified some corner cases. |
91 | - vastly improved documentation with a long example. |
204 | - vastly improved documentation with a long example. |
92 | |
205 | |
93 | 0.2 Thu Dec 1 22:18:49 CET 2005 |
206 | 0.2 Thu Dec 1 22:18:49 CET 2005 |
94 | - check wether a valid method was called, to avoid endless loops. |
207 | - check whether a valid method was called, to avoid endless loops. |
95 | - move condvars into their own class (cleaner, safer). |
208 | - move condvars into their own class (cleaner, safer). |
96 | - don't die when some modules are not available. |
209 | - don't die when some modules are not available. |
97 | - Tk is too broken to allow for words. Will work once Tk works. |
210 | - Tk is too broken to allow for words. Will work once Tk works. |
98 | |
211 | |
99 | 0.1 Thu Dec 1 19:44:57 CET 2005 |
212 | 0.1 Thu Dec 1 19:44:57 CET 2005 |