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

Comparing AnyEvent/Changes (file contents):
Revision 1.14 by root, Tue Oct 31 01:33:00 2006 UTC vs.
Revision 1.119 by root, Mon May 26 21:37:43 2008 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2 2
3TODO: testsuite, now that we can be sure that some model 3 - add some recommends to the META.yml.
4TODO: is available at all times 4 - added a a regex read type to AnyEvent::Handle.
5 - always croak in on_error in AnyEvent::Handle.
6 - document how to subclass AnyEvent::Handle.
5 7
62.0 84.05 Mon May 26 19:44:06 CEST 2008
9 - some platforms ignore the rfc and prepend an extra sa_len member to
10 the sockaddr structure, cater for those. also use sockaddr_family,
11 didn't know it was there, but comes in handy for the workaround.
12 - undo the import WIN32 hack.
13
144.04 Mon May 26 08:03:31 CEST 2008
15 - try to work around yet another windows bug: failed connects are reported
16 as if it were out-of-band data. windows users: you suck. Only
17 EV, Glib, Event and the pure perl backend can handle this condition.
18 - optimize write algorithm in AnyEvent::Handle.
19 - properly parse PERL_ANYEVENT_PROTOCOLS and use the info correctly.
20 - tcp_connect now properly iterates through all targets.
21 - check wether IPv6 sockets can actually be created, otherwise, disable ipv6
22 support entirely. also hardcode some AF_INET6 constants for the
23 sake of perl 5.8.
24 - vastly improved nameserver/suffix detection algorithm for windows
25 (still a hack...)
26 - try to use a fallback dns server if no dns servers could be found.
27 - splatter around some local $SIG{__DIE__}'s to work around bad
28 die handlers.
29
304.03 Sun May 25 05:05:57 CEST 2008
31 - work around perl bugs on windows, where perl returns undocumented
32 error codes for sysread, syswrite etc.
33 - fix AnyEvent::Handle::unshift_read to really unshift and not push.
34 - AnyEvent::Handle could sometimes stumble over 0-byte writes.
35 - fix unpack_sockaddr, now works for AF_INET6 sockaddr's.
36 - updated documentation.
37 - lots of workarounds for perl 5.8 quirks.
38 - implement a more flexible type system for both reads and writes
39 in AnyEvent::Handle.
40 - support IPv6 in tcp_server.
41 - pass local hostname and port to the prepare callback in tcp_server.
42 - make types extendable in AnyEvent::Handle.
43 - croak, not die, in AnyEvent::Handle;
44
454.0 Sat May 24 19:58:08 CEST 2008
46 - added (experimental) AnyEvent::DNS module.
47 - added (experimental) AnyEvent::Socket module.
48 - added (experiemntal) support for TLS/SSL in AnyEvent::Handle.
49 - make condvar->end call ->send as documented.
50 - make condvar constructor support the documented cb => argument.
51 - overload condvars so you can use them instead of code references.
52 - support more of the POSIX forms instead of a pure dotted quad for
53 verbatim ip addresses.
54 - AnyEvent::Socket::inet_aton will use AnyEvent::DNS now and not fork anymore.
55 - try very hard not to freeze on broken windows perls.
56 - make on_eof optional in AnyEvent::Handle.
57 - support service names in tcp_connect and tcp_server.
58 - work around netbsd/openbsd bug causing unix sockets not to
59 become writable untill empty.
60 - only make a dummy read for error detection when getpeername
61 returns ENOTCONN.
62 - defend AnyEvent::Util::inet_aton against Coro::LWP brutal
63 override.
64 - new AnyEvent::Guard::cancel.
65
663.5 Sat May 17 23:17:14 CEST 2008
67 - gracefully deal with systems not defining CLOCK_MONOTONIC
68 (as opposed to not supporting the clock itself, which worked in 3.41).
69 - added AnyEvent::Util::tcp_connect and AnyEvent::Util::tcp_server
70 helpers, to replace the ill-designed AnyEvent::Socket class.
71 - added some examples (eg/connect, eg/handle).
72 - added AnyEvent::Util::guard.
73 - AnyEvent::Util now exports most of its functions.
74 - fixed a bug in *_read_chunk, where the callback was called with $self as
75 first and second argument.
76 - simplified and fixed regex handling in AnyEvent::Handle::*_read_line.
77 - fix manpage condvars examples to use new syntax.
78 - considerably improve inet_aton with EV::ADNS backend to support
79 cname chains (common dns configuration bug, e.g. with www.google.de)
80 and aaaa records if no a record could be found.
81
823.41 Sun May 11 19:53:13 CEST 2008
83 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao.
84
853.4 Sun May 11 00:29:25 CEST 2008
86 - complete re-implementation of condvars with a hopefully much mroe useful
87 API (backward compatible functions exist but are undocumented).
88 - AnyEvent::Handle fully rewritten with a hopefully more powerful API
89 (still subject to change).
90 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl.
91 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which
92 allows module authors to avoid forcing event loop detection.
93 (used by Coro::AIO, Coro::BDB, Coro::AnyEvent for example).
94 - remove coro backends: Coro now provides generic support for AnyEvent
95 via Coro::AnyEvent.
96
973.3 Mon Apr 28 09:51:06 CEST 2008
98 - added AnyEvent::Handle, AnyEvent::Socket and AnyEvent::Util modules.
99 - fix a bug in the pure perl backend that kept watchers alive
100 when multiple watchers were registered for the same fh/poll combo.
101 - add a benchmark section showing AnyEvent overhead and comparing
102 the different event loops with each other.
103 - prefer pure perl over tk when autoprobing, as it's about as fast,
104 but doesn't crash with many watchers.
105 - declare Qt support non-experimental.
106 - clarify the confusing section about the file descriptor being
107 kept alive.
108 - document the race between loading of an event module and
109 child processes exiting.
110 - support POE as "backend" (with some caveats, POE is not generic
111 enough, and darn slow).
112 - support Wx and Prima through POE.
113 - optimise perl backend to use 20% less memory and take advantage
114 of typical timeout behaviour. It can now compete with
115 select/poll-based C event loops in most cases (it is usually
116 faster than Event and Glib :).
117 - roughly cut EV memory use in half and increase its speed by 30%,
118 by removing undocumented functionality. Did something similar
119 to other event loops where possible.
120
1213.2 Thu Apr 24 10:10:40 CEST 2008
122 - do not die when anyevent watchers are destroyed while running callbacks
123 in the pureperl backend (could only happen when two watchers are registered
124 for the same fh/poll combo).
125 - support autoloading for child watchers, was broken in all versions.
126 - implement PERL_ANYEVENT_MODEL env variable.
127 - (experimental) implement interface to the (very crashy/buggy) Event::Lib module.
128 - (experimental) implement interface to the Qt module (cannot be autoprobed).
129 - this release is IO:AnyEvent-proof.
130
1313.12 Tue Apr 22 07:11:46 CEST 2008
132 - reinstate AUTHOR section that got lost somehow.
133 - do not hang in the testsuite with badly broken perls (activestate,
134 strawberry...), but instead diagnose the problem and continue.
135 - use INT instead of CHLD in an attempt to support broken windows perls
136 better (this decreases the test quality, unfortunately).
137 - do not send the signal to the process group
138 (no problem for CHLD, bad for INT :).
139
1403.11 Sat Apr 19 06:57:31 CEST 2008
141 - major documentation rework.
142 - document the fact that child watchers only watch for zombies.
143 - fix the child watcher example.
144
1453.1 Wed Apr 16 17:09:01 CEST 2008
146 - work around recurring bugs in Tk by dup'ing filehandles, the
147 only method with good success chances on Tk (the bugs apparently
148 don't get fixed anytime soon).
149 - lift the restriction of only one watcher per fh direction
150 (as the Tk bug workaround also lifts it and only Tk imposed
151 such strong limits).
152 - changed probe order to prefer coro adaptors.
153 - explain why recursion into the event loop is not supported
154 unless the backend supports it (only Coro::EV does without
155 any restrictions...).
156 - add simple manpages for all backend modules.
157
1583.0 Mon Apr 7 21:30:23 CEST 2008
159 - Coro::Signal changed semantics, roll our own, also cleaning
160 up the Coro implementation in general.
161 - rename Coro backend to CoroEvent.
162 - add some decision helping paragraph to the manpage that should
163 help people to decide whether AnyEvent is the right thing for them.
164
1652.9 Mon Jan 28 13:31:54 CET 2008
166 - update for EV 3.0 API changes.
167
1682.8 Sun Nov 25 15:06:03 CET 2007
169 - waitpid can validly return 0. accept this fact of life
170 instead of reporting it to any watchers.
171
1722.7 Fri Nov 23 11:41:14 CET 2007
173 - force use of AnyEvent::Impl::Perl in testsuite, there is
174 too much breakage outside AnyEvent.
175 - deliver signals synchronously in AnyEvent::Impl::Perl
176 or any other event loop that relies on AnyEvents child
177 watcher emulation. *Could* help with hanging testsuite
178 (except when Event or EV are installed).
179
1802.6 Fri Nov 9 20:36:35 CET 2007
181 - fix bug in testsuite.
182 - move EV adaptor modules to AnyEvent.
183 - add Coro+EV adaptor module.
184
1852.55 Tue Nov 6 17:41:32 CET 2007
186 - add EV to the list of supported event models.
187 - do not auto-reset pid watchers, pass pid and status to them.
188 - allow a pid of zero to watch for all children in child watchers.
189
1902.54 Wed Jul 18 17:36:23 CEST 2007
191 - work around a perl bug that results in
192 BEGIN not safe after errors--compilation aborted without
193 any discernible reason or error message by once
194 again not use'ing strict.
195
1962.53 Sun Jul 8 10:51:53 CEST 2007
197 - make Glib interface work again (spotted by elmex).
198
1992.52 Wed Mar 7 18:36:16 CET 2007
200 - child watchers stopped working permanently when
201 all current child watchers were destroyed.
202
2032.51 Mon Dec 11 21:33:24 CET 2006
204 - work around bugs in perl where eval "require Module" returns
205 true even if the module couldn't be loaded :/.
206
2072.5 Mon Dec 11 02:15:28 CET 2006
208 - avoid $AUTOLOAD because many perls corrupt it.
209 - AnyEvent::detect forces autodetection.
210 - implement signal watchers (experimental)
211 (Tk does not support async signals (<= 804.027 at least),
212 so they might get delayed indefinitely on Tk).
213 - implement child watchers (experimental).
214 - moved default condvar implementation into base module, simplifying
215 most implementation modules.
216
2172.1 Fri Nov 24 15:50:48 CET 2006
218 - better docs.
219 - simple testsuite.
220 - added AnyEvent->one_event method for special purposes.
221
2222.0 Wed Nov 1 02:21:30 CET 2006
7 - INCOMPATIBLE CHANGE: poll can now either be "r" or "w" but not both. 223 - INCOMPATIBLE CHANGE: poll can now either be "r" or "w" but not both.
8 - INCOMPATIBLE CHANGE: io watcher callbacks have no arguments anymore. 224 - INCOMPATIBLE CHANGE: io watcher callbacks have no arguments anymore.
9 - EXPERIMENTAL: 225 - EXPERIMENTAL:
10 new pure-perl model added, AnyEvent now always finds a suitable 226 new pure-perl model added, AnyEvent now always finds a suitable
11 event model. 227 event model.
… …
38 - fixed documentation. 254 - fixed documentation.
39 - clarified some corner cases. 255 - clarified some corner cases.
40 - vastly improved documentation with a long example. 256 - vastly improved documentation with a long example.
41 257
420.2 Thu Dec 1 22:18:49 CET 2005 2580.2 Thu Dec 1 22:18:49 CET 2005
43 - check wether a valid method was called, to avoid endless loops. 259 - check whether a valid method was called, to avoid endless loops.
44 - move condvars into their own class (cleaner, safer). 260 - move condvars into their own class (cleaner, safer).
45 - don't die when some modules are not available. 261 - don't die when some modules are not available.
46 - Tk is too broken to allow for words. Will work once Tk works. 262 - Tk is too broken to allow for words. Will work once Tk works.
47 263
480.1 Thu Dec 1 19:44:57 CET 2005 2640.1 Thu Dec 1 19:44:57 CET 2005

Diff Legend

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