| 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? |
|
|
4 | 3.45 Tue Oct 21 21:59:26 CEST 2008 |
|
|
5 | - disable inotify usage on linux <2.6.25, as it is broken |
|
|
6 | (reported by Yoann Vandoorselaere). |
|
|
7 | - ev_stat errornously would try to add inotify watchers |
|
|
8 | even when inotify wasn't available (this should only |
|
|
9 | have a performance impact). |
|
|
10 | - ev_once now passes both timeout and io to the callback if both |
|
|
11 | occur concurrently, instead of giving timeouts precedence. |
|
|
12 | - disable EV_USE_INOTIFY when sys/inotify.h is too old. |
|
|
13 | |
|
|
14 | 3.44 Mon Sep 29 05:18:39 CEST 2008 |
|
|
15 | - embed watchers now automatically invoke ev_loop_fork on the |
|
|
16 | embedded loop when the parent loop forks. |
|
|
17 | - new function: ev_now_update (loop). |
|
|
18 | - verify_watcher was not marked static. |
|
|
19 | - improve the "associating..." manpage section. |
|
|
20 | - documentation tweaks here and there. |
|
|
21 | |
|
|
22 | 3.43 Sun Jul 6 05:34:41 CEST 2008 |
|
|
23 | - include more include files on windows to get struct _stati64 |
|
|
24 | (reported by Chris Hulbert, but doesn't quite fix his issue). |
|
|
25 | - add missing #include <io.h> in ev.c on windows (reported by |
|
|
26 | Matt Tolton). |
|
|
27 | |
|
|
28 | 3.42 Tue Jun 17 12:12:07 CEST 2008 |
|
|
29 | - work around yet another windows bug: FD_SET actually adds fd's |
|
|
30 | multiple times to the fd_*SET*, despite official MSN docs claiming |
|
|
31 | otherwise. Reported and well-analysed by Matt Tolton. |
|
|
32 | - define NFDBITS to 0 when EV_SELECT_IS_WINSOCKET to make it compile |
|
|
33 | (reported any analysed by Chris Hulbert). |
|
|
34 | - fix a bug in ev_ebadf (this function is only used to catch |
|
|
35 | programming errors in the libev user). reported by Matt Tolton. |
|
|
36 | - fix a bug in fd_intern on win32 (could lead to compile errors |
|
|
37 | under some circumstances, but would work correctly if it compiles). |
|
|
38 | reported by Matt Tolton. |
|
|
39 | - (try to) work around missing lstat on windows. |
|
|
40 | - pass in the write fd set as except fd set under windows. windows |
|
|
41 | is so uncontrollably lame that it requires this. this means that |
|
|
42 | switching off oobinline is not supported (but tcp/ip doesn't |
|
|
43 | have oob, so that would be stupid anyways. |
|
|
44 | - use posix module symbol to auto-detect monotonic clock presence |
|
|
45 | and some other default values. |
|
|
46 | |
|
|
47 | 3.41 Fri May 23 18:42:54 CEST 2008 |
|
|
48 | - work around an obscure bug in winsocket select: if you |
|
|
49 | provide only empty fd sets then select returns WSAEINVAL. how sucky. |
|
|
50 | - improve timer scheduling stability and reduce use of time_epsilon. |
|
|
51 | - use 1-based 2-heap for EV_MINIMAL, simplifies code, reduces |
|
|
52 | codesize and makes for better cache-efficiency. |
|
|
53 | - use 3-based 4-heap for !EV_MINIMAL. this makes better use |
|
|
54 | of cpu cache lines and gives better growth behaviour than |
|
|
55 | 2-based heaps. |
|
|
56 | - cache timestamp within heap for !EV_MINIMAL, to avoid random |
|
|
57 | memory accesses. |
|
|
58 | - document/add EV_USE_4HEAP and EV_HEAP_CACHE_AT. |
|
|
59 | - fix a potential aliasing issue in ev_timer_again. |
|
|
60 | - add/document ev_periodic_at, retract direct access to ->at. |
|
|
61 | - improve ev_stat docs. |
|
|
62 | - add portability requirements section. |
|
|
63 | - fix manpage headers etc. |
|
|
64 | - normalise WSA error codes to lower range on windows. |
|
|
65 | - add consistency check code that can be called automatically |
|
|
66 | or on demand to check for internal structures (ev_loop_verify). |
|
|
67 | |
|
|
68 | 3.31 Wed Apr 16 20:45:04 CEST 2008 |
|
|
69 | - added last minute fix for ev_poll.c by Brandon Black. |
|
|
70 | |
|
|
71 | 3.3 Wed Apr 16 19:04:10 CEST 2008 |
| 3 | - event_base_loopexit should return 0 on success |
72 | - event_base_loopexit should return 0 on success |
| 4 | (W.C.A. Wijngaards). |
73 | (W.C.A. Wijngaards). |
| 5 | - work around poll() potentially not clearing revents |
|
|
| 6 | field in ev_poll (Brandon Black). |
|
|
| 7 | - added linux eventfd support. |
74 | - added linux eventfd support. |
| 8 | - try to autodetect epoll and inotify support |
75 | - try to autodetect epoll and inotify support |
| 9 | by libc header version if not using autoconf. |
76 | by libc header version if not using autoconf. |
| 10 | - new symbols: EV_DEFAULT_UC and EV_DEFAULT_UC_. |
77 | - new symbols: EV_DEFAULT_UC and EV_DEFAULT_UC_. |
| 11 | - declare functions defined in ev.h as inline if |
78 | - declare functions defined in ev.h as inline if |
| 12 | C99 or gcc are available. |
79 | C99 or gcc are available. |
| 13 | - enable inlining with gcc versions 2 and 3. |
80 | - enable inlining with gcc versions 2 and 3. |
|
|
81 | - work around broken poll implementations potentially |
|
|
82 | not clearing revents field in ev_poll (Brandon Black) |
|
|
83 | (no such systems are known at this time). |
| 14 | - work around a bug in realloc on openbsd and darwin, |
84 | - work around a bug in realloc on openbsd and darwin, |
| 15 | also makes the errornous valgrind complaints |
85 | also makes the errornous valgrind complaints |
| 16 | go away (noted by various people). |
86 | go away (noted by various people). |
|
|
87 | - fix ev_async_pending, add c++ wrapper for ev_async |
|
|
88 | (based on patch sent by Johannes Deisenhofer. |
|
|
89 | - add sensible set method to ev::embed. |
|
|
90 | - made integer constants type int in ev.h. |
| 17 | |
91 | |
| 18 | 3.2 Wed Apr 2 17:11:19 CEST 2008 |
92 | 3.2 Wed Apr 2 17:11:19 CEST 2008 |
| 19 | - fix a 64 bit overflow issue in the select backend, |
93 | - fix a 64 bit overflow issue in the select backend, |
| 20 | by using fd_mask instead of int for the mask. |
94 | by using fd_mask instead of int for the mask. |
| 21 | - rename internal sighandler to avoid clash with very old perls. |
95 | - rename internal sighandler to avoid clash with very old perls. |
