| 1 | Revision history for Perl extension AnyEvent. |
1 | Revision history for Perl extension AnyEvent. |
|
|
2 | |
|
|
3 | 4.351 Sat Apr 11 07:56:14 CEST 2009 |
|
|
4 | - actually make the signal pipe work under win32. |
|
|
5 | - localise $! in signal handler to avoid changing $!, although |
|
|
6 | perl itself does not seem to save/restore errno either. |
|
|
7 | - set the cloexec flag on the signal pipe (normally set by perl too). |
|
|
8 | |
|
|
9 | 4.35 Fri Mar 27 11:48:20 CET 2009 |
|
|
10 | - event models relying on AnyEvent's signal watcher emulation did |
|
|
11 | invoke the callback asynchronously, contradicting documentation |
|
|
12 | and causing signals to get lost (this includes AnyEvent's own |
|
|
13 | event loop). AnyEvent now uses the standard pipe trick to make |
|
|
14 | callback execution synchronous to the event loop. |
|
|
15 | - AnyEvent::Handle didn't free TLS context data on DESTROY |
|
|
16 | (patch by Pavel Shaydo). |
|
|
17 | - work around the ever-incompatibly-changing API of MakeMaker. |
|
|
18 | - document that changing global variables without restoring them |
|
|
19 | is a bad idea in Perl (noted by Adam Rosenstein). |
|
|
20 | - AnyEvent::Strict now barfs if ->io is passed a file. |
| 2 | |
21 | |
| 3 | 4.34 Thu Feb 12 18:32:45 CET 2009 |
22 | 4.34 Thu Feb 12 18:32:45 CET 2009 |
| 4 | - separately buffer TLS read data, as otherwise the read queue could |
23 | - separately buffer TLS read data, as otherwise the read queue could |
| 5 | deadlock as receiving data is not expected while draining the |
24 | deadlock as receiving data is not expected while draining the |
| 6 | read queue (which cna only happen with TLS). |
25 | read queue (which cna only happen with TLS). |
