| 1 | Revision history for Perl extension EV |
1 | Revision history for Perl extension EV |
| 2 | |
2 | |
|
|
3 | - (libev) fix some issues in the select backend when in fd_set mode. |
|
|
4 | - (libev) due to a thinko, instead of disabling everything but |
|
|
5 | select on the borked OS X platform, everything but select was |
|
|
6 | allowed (reported by Emanuele Giaquinta). |
|
|
7 | - (libev) actually verify that local and remote port are matching in |
|
|
8 | libev's socketpair emulation, which makes denial-of-service |
|
|
9 | attacks harder (but not impossible - it's windows). Make sure |
|
|
10 | it even works under vista, which thinks that getpeer/sockname |
|
|
11 | should return fantasy port numbers. |
|
|
12 | |
|
|
13 | 3.51 Wed Dec 24 23:01:59 CET 2008 |
|
|
14 | - do not cache the arguments passed to callbacks if the refcount |
|
|
15 | indicates that the callback has stolen them. |
|
|
16 | - (libev) try to avoid librt on GNU/Linux. |
|
|
17 | - (libev) check port_getn return value dfferently, might potentially |
|
|
18 | avoid problems. |
|
|
19 | - (libev) fix a bug with stat watchers possibly causing freezes. |
|
|
20 | - (libev) work around OS X 10.5 breaking poll, now select is the |
|
|
21 | only viable choice left on that pile of garbage. |
|
|
22 | - play tester whore: disable some tests that typically fail only |
|
|
23 | on cpan tester machines. |
|
|
24 | |
|
|
25 | 3.49 Wed Nov 19 11:26:53 CET 2008 |
|
|
26 | - fix typos in manpage (Alex Efros). |
|
|
27 | - increase timing even further, for the ever-overloaded cpan-tester |
|
|
28 | machines, and to the detriment of everybody else who wants a fast |
|
|
29 | make test. |
|
|
30 | - possible 5.6 compatibility. |
|
|
31 | - (libev) use inotify even on buggy kernels, but don't rely on it. |
|
|
32 | - (libev) use inotify only as an added hint on network filesystems. |
|
|
33 | |
|
|
34 | 3.48 Thu Oct 30 09:09:48 CET 2008 |
| 3 | - (libev) use a generation counter to detetc spurious epoll events |
35 | - (libev) use a generation counter to detect spurious epoll events |
| 4 | and recreate the kernel in such a case (sorry, it's slow, but |
36 | and recreate the kernel in such a case (sorry, it's slow, but |
| 5 | I didn't design epoll...). |
37 | I didn't design epoll...). |
| 6 | - (libev) optimise away an EPOLL_CTL_ADD/MOD combo in the epoll |
38 | - (libev) optimise away an EPOLL_CTL_ADD/MOD combo in the epoll |
| 7 | backend in some cases. |
39 | backend in some cases. |
| 8 | - (libev) use memset to initialise most arrays now and do away |
40 | - (libev) use memset to initialise most arrays now and do away |
| 9 | with the init functions. |
41 | with the init functions. |
| 10 | - speed up event callback invocation further (~15%) by |
42 | - speed up event callback invocation further (~15%), by |
| 11 | not creating the object reference each time. |
43 | not creating the object reference each time. |
|
|
44 | - EV::sleep was documented, but not implemented. fun. |
| 12 | |
45 | |
| 13 | 3.45 Tue Oct 21 22:20:39 CEST 2008 |
46 | 3.45 Tue Oct 21 22:20:39 CEST 2008 |
| 14 | - (libev) disable inotify usage on linux <2.6.25 (kernel bug). |
47 | - (libev) disable inotify usage on linux <2.6.25 (kernel bug). |
| 15 | - (libev) ev_embed will now automatically follow fork. |
48 | - (libev) ev_embed will now automatically follow fork. |
| 16 | - (libev) ev_once will now pass both io and timeout events to |
49 | - (libev) ev_once will now pass both io and timeout events to |
