| 1 | Revision history for Perl extension AnyEvent. |
1 | Revision history for Perl extension AnyEvent. |
|
|
2 | |
|
|
3 | - add a benchmark section showing AnyEvent overhead and comparing |
|
|
4 | the different event loops with each other. |
|
|
5 | - make Qt and Event::Lib support non-experimental. |
|
|
6 | - support POE as "backend" (with some caveats, POE is not generic |
|
|
7 | enough, and darn slow). |
|
|
8 | - support Wx and Prima through POE. |
|
|
9 | |
|
|
10 | 3.2 Thu Apr 24 10:10:40 CEST 2008 |
|
|
11 | - do not die when anyevent watchers are destroyed while running callbacks |
|
|
12 | in the pureperl backend (could only happen when two watchers are registered |
|
|
13 | for the same fh/poll combo). |
|
|
14 | - support autoloading for child watchers, was broken in all versions. |
|
|
15 | - implement PERL_ANYEVENT_MODEL env variable. |
|
|
16 | - (experimental) implement interface to the (very crashy/buggy) Event::Lib module. |
|
|
17 | - (experimental) implement interface to the Qt module (cannot be autoprobed). |
|
|
18 | - this release is IO:AnyEvent-proof. |
|
|
19 | |
|
|
20 | 3.12 Tue Apr 22 07:11:46 CEST 2008 |
|
|
21 | - reinstate AUTHOR section that got lost somehow. |
|
|
22 | - do not hang in the testsuite with badly broken perls (activestate, |
|
|
23 | strawberry...), but instead diagnose the problem and continue. |
|
|
24 | - use INT instead of CHLD in an attempt to support broken windows perls |
|
|
25 | better (this decreases the test quality, unfortunately). |
|
|
26 | - do not send the signal to the process group |
|
|
27 | (no problem for CHLD, bad for INT :). |
|
|
28 | |
|
|
29 | 3.11 Sat Apr 19 06:57:31 CEST 2008 |
|
|
30 | - major documentation rework. |
|
|
31 | - document the fact that child watchers only watch for zombies. |
|
|
32 | - fix the child watcher example. |
| 2 | |
33 | |
| 3 | 3.1 Wed Apr 16 17:09:01 CEST 2008 |
34 | 3.1 Wed Apr 16 17:09:01 CEST 2008 |
| 4 | - work around recurring bugs in Tk by dup'ing filehandles, the |
35 | - work around recurring bugs in Tk by dup'ing filehandles, the |
| 5 | only method with good success chances on Tk (the bugs apparently |
36 | only method with good success chances on Tk (the bugs apparently |
| 6 | don't get fixed anytime soon). |
37 | don't get fixed anytime soon). |
