| 1 | Revision history for libev, a high-performance and full-featured event loop. |
1 | Revision history for libev, a high-performance and full-featured event loop. |
| 2 | |
2 | |
| 3 | WISH? monotonic clocks times/GetTickCount for coarse corrections? |
3 | - actually verify that local and remote port are matching in |
|
|
4 | libev's socketpair emulation, which makes denial-of-service |
|
|
5 | attacks harder (but not impossible - it's windows). Make sure |
|
|
6 | it even works under vista, which thinks that getpeername should |
|
|
7 | return some fantasy port number. |
| 4 | |
8 | |
|
|
9 | 3.51 Wed Dec 24 23:00:11 CET 2008 |
|
|
10 | - fix a bug where an inotify watcher was added twice, causing |
|
|
11 | freezes on hash collisions (reported and analysed by Graham Leggett). |
|
|
12 | - new config symbol, EV_USE_CLOCK_SYSCALL, to make libev use |
|
|
13 | a direct syscall - slower, but no dependency on librt et al. |
|
|
14 | - assume negative return values != -1 signals success of port_getn |
|
|
15 | (http://cvs.epicsol.org/cgi/viewcvs.cgi/epic5/source/newio.c?rev=1.52) |
|
|
16 | (no known failure reports, but it doesn't hurt). |
|
|
17 | - fork detection in ev_embed now stops and restarts the watcher |
|
|
18 | automatically. |
|
|
19 | - EXPERIMENTAL: default the method to operator () in ev++.h, |
|
|
20 | to make it nicer to use functors (requested by Benedek László). |
|
|
21 | - fixed const object callbacks in ev++.h. |
|
|
22 | - replaced loop_ref argument of watcher.set (loop) by a direct |
|
|
23 | ev_loop * in ev++.h, to avoid clashes with functor patch. |
|
|
24 | - do not try to watch the empty string via inotify. |
|
|
25 | - inotify watchers could be leaked under certain circumstances. |
|
|
26 | - OS X 10.5 is actually even more broken than earlier versions, |
|
|
27 | so fall back to select on that piece of garbage. |
|
|
28 | - fixed some weirdness in the ev_embed documentation. |
|
|
29 | |
|
|
30 | 3.49 Wed Nov 19 11:26:53 CET 2008 |
|
|
31 | - ev_stat watchers will now use inotify as a mere hint on |
|
|
32 | kernels <2.6.25, or if the filesystem is not in the |
|
|
33 | "known to be good" list. |
|
|
34 | - better mingw32 compatibility (it's not as borked as native win32) |
|
|
35 | (analysed by Roger Pack). |
|
|
36 | - include stdio.h in the example program, as too many people are |
|
|
37 | confused by the weird C language otherwise. I guess the next thing |
|
|
38 | I get told is that the "..." ellipses in the examples don't compile |
|
|
39 | with their C compiler. |
|
|
40 | |
|
|
41 | 3.48 Thu Oct 30 09:02:37 CET 2008 |
| 5 | - further optimise away the EPOLL_CTL_ADD/MOD combo in the epoll |
42 | - further optimise away the EPOLL_CTL_ADD/MOD combo in the epoll |
| 6 | backend by assuming the kernel event mask hasn't changed if |
43 | backend by assuming the kernel event mask hasn't changed if |
| 7 | ADD fails with EEXIST. |
44 | ADD fails with EEXIST. |
| 8 | - work around spurious event notification bugs in epoll by using |
45 | - work around spurious event notification bugs in epoll by using |
| 9 | a 32-bit generation counter. recreate kernel state if we receive |
46 | a 32-bit generation counter. recreate kernel state if we receive |
