1 | Revision history for Perl extension EV |
1 | Revision history for Perl extension EV |
2 | |
2 | |
|
|
3 | - do not cache the arguments passed to callbacks if the refcount |
|
|
4 | indicates that the callback has stolen them. |
|
|
5 | - (libev) try to avoid librt on GNU/Linux. |
|
|
6 | - (libev) check port_getn return value dfferently, might potentially |
|
|
7 | avoid problems. |
|
|
8 | - play tester whore: disable some tests that typically fail only |
|
|
9 | on cpan tester machines. |
|
|
10 | |
|
|
11 | 3.49 Wed Nov 19 11:26:53 CET 2008 |
3 | - fix typos in manpage (Alex Efros). |
12 | - fix typos in manpage (Alex Efros). |
4 | - increase timing even further, for the ever-overloaded cpan-tester |
13 | - increase timing even further, for the ever-overloaded cpan-tester |
5 | machines, and to the detriment of everybody else who wants a fast |
14 | machines, and to the detriment of everybody else who wants a fast |
6 | make test. |
15 | make test. |
|
|
16 | - possible 5.6 compatibility. |
|
|
17 | - (libev) use inotify even on buggy kernels, but don't rely on it. |
|
|
18 | - (libev) use inotify only as an added hint on network filesystems. |
7 | |
19 | |
8 | 3.48 Thu Oct 30 09:09:48 CET 2008 |
20 | 3.48 Thu Oct 30 09:09:48 CET 2008 |
9 | - (libev) use a generation counter to detect spurious epoll events |
21 | - (libev) use a generation counter to detect spurious epoll events |
10 | and recreate the kernel in such a case (sorry, it's slow, but |
22 | and recreate the kernel in such a case (sorry, it's slow, but |
11 | I didn't design epoll...). |
23 | I didn't design epoll...). |