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 | TODO: ev_loop_wakeup |
|
|
4 | |
|
|
5 | TODO: EV_NO_THREADS/NO_SMP and/or EV_USE_PTHREADS or so. also document. |
|
|
6 | 4.10 |
|
|
7 | - fix a race where the workaround against the epoll fork bugs |
|
|
8 | caused signals to not be handled anymore. |
|
|
9 | - correct backend_fudge for most backends, and implement a windows |
|
|
10 | specific workaround to avoid looping because we call both |
|
|
11 | select and Sleep, both with different time resolutions. |
|
|
12 | - document range and guarantees of ev_sleep. |
|
|
13 | - document reasonable ranges for periodics interval and offset. |
|
|
14 | - rename backend_fudge to backend_mintime to avoid future confusion :) |
|
|
15 | - change the default periodic reschedule function to hopefully be more |
|
|
16 | exact and correct even in corner cases or in the far future. |
|
|
17 | - do not rely on -lm anymore: use it when available but use our |
|
|
18 | own floor () if it is missing. This should make it easier to embed, |
|
|
19 | as no external libraries are required. |
|
|
20 | - strategically import macros from libecb and mark rarely-used functions |
|
|
21 | as cache-chold (saving almost 2k code size on typical amd64 setups). |
|
|
22 | |
|
|
23 | 4.04 Wed Feb 16 09:01:51 CET 2011 |
|
|
24 | - fix two problems in the native win32 backend, where reuse of fd's |
|
|
25 | with different underlying handles caused handles not to be removed |
|
|
26 | or added to the select set (analyzed and tested by Bert Belder). |
|
|
27 | - do no rely on ceil() in ev_e?poll.c. |
|
|
28 | - backport libev to HP-UX versions before 11 v3. |
|
|
29 | - configure did not detect nanosleep and clock_gettime properly when |
|
|
30 | they are available in the libc (as opposed to -lrt). |
|
|
31 | |
|
|
32 | 4.03 Tue Jan 11 14:37:25 CET 2011 |
|
|
33 | - officially support polling files with all backends. |
|
|
34 | - support files, /dev/zero etc. the same way as select in the epoll |
|
|
35 | backend, by generating events on our own. |
|
|
36 | - ports backend: work around solaris bug 6874410 and many related ones |
|
|
37 | (EINTR, maybe more), with no performance loss (note that the solaris |
|
|
38 | bug report is actually wrong, reality is far more bizarre and broken |
|
|
39 | than that). |
|
|
40 | - define EV_READ/EV_WRITE as macros in event.h, as some programs use |
|
|
41 | #ifdef to test for them. |
|
|
42 | - new (experimental) function: ev_feed_signal. |
|
|
43 | - new (to become default) EVFLAG_NOSIGMASK flag. |
|
|
44 | - new EVBACKEND_MASK symbol. |
|
|
45 | - updated COMMON IDIOMS SECTION. |
|
|
46 | |
|
|
47 | 4.01 Fri Nov 5 21:51:29 CET 2010 |
|
|
48 | - automake fucked it up, apparently, --add-missing -f is not quite enough |
|
|
49 | to make it update its files, so 4.00 didn't install ev++.h and |
|
|
50 | event.h on make install. grrr. |
|
|
51 | - ev_loop(count|depth) didn't return anything (Robin Haberkorn). |
|
|
52 | - change EV_UNDEF to 0xffffffff to silence some overzealous compilers. |
|
|
53 | - use "(libev) " prefix for all libev error messages now. |
|
|
54 | |
|
|
55 | 4.00 Mon Oct 25 12:32:12 CEST 2010 |
|
|
56 | - "PORTING FROM LIBEV 3.X TO 4.X" (in ev.pod) is recommended reading. |
3 | - ev_embed_stop did not correctly stop the watcher (very good |
57 | - ev_embed_stop did not correctly stop the watcher (very good |
4 | testcase by Vladimir Timofeev). |
58 | testcase by Vladimir Timofeev). |
|
|
59 | - ev_run will now always update the current loop time - it erroneously |
|
|
60 | didn't when idle watchers were active, causing timers not to fire. |
|
|
61 | - fix a bug where a timeout of zero caused the timer not to fire |
|
|
62 | in the libevent emulation (testcase by Péter Szabó). |
5 | - applied win32 fixes by Michael Lenaghan. |
63 | - applied win32 fixes by Michael Lenaghan (also James Mansion). |
|
|
64 | - replace EV_MINIMAL by EV_FEATURES. |
|
|
65 | - prefer EPOLL_CTL_ADD over EPOLL_CTL_MOD in some more cases, as it |
|
|
66 | seems the former is *much* faster than the latter. |
|
|
67 | - linux kernel version detection (for inotify bug workarounds) |
|
|
68 | did not work properly. |
|
|
69 | - reduce the number of spurious wake-ups with the ports backend. |
6 | - remove dependency on sys/queue.h on freebsd (patch by Vanilla Hsu). |
70 | - remove dependency on sys/queue.h on freebsd (patch by Vanilla Hsu). |
|
|
71 | - do async init within ev_async_start, not ev_async_set, which avoids |
|
|
72 | an API quirk where the set function must be called in the C++ API |
|
|
73 | even when there is nothing to set. |
|
|
74 | - add (undocumented) EV_ENABLE when adding events with kqueue, |
|
|
75 | this might help with OS X, which seems to need it despite documenting |
|
|
76 | not to need it (helpfully pointed out by Tilghman Lesher). |
|
|
77 | - do not use poll by default on freebsd, it's broken (what isn't |
|
|
78 | on freebsd...). |
|
|
79 | - allow to embed epoll on kernels >= 2.6.32. |
7 | - configure now prepends -O3, not appends it, so one can still |
80 | - configure now prepends -O3, not appends it, so one can still |
8 | override it. |
81 | override it. |
|
|
82 | - ev.pod: greatly expanded the portability section, added a porting |
|
|
83 | section, a description of watcher states and made lots of minor fixes. |
9 | - disable poll backend on AIX, the poll header spams the namespace |
84 | - disable poll backend on AIX, the poll header spams the namespace |
10 | and it's not worth working around dead platforms (reported |
85 | and it's not worth working around dead platforms (reported |
11 | and analyzed by Aivars Kalvans). |
86 | and analyzed by Aivars Kalvans). |
12 | - improve header file compatibility of the standalone eventfd code |
87 | - improve header file compatibility of the standalone eventfd code |
13 | in an obscure case. |
88 | in an obscure case. |
14 | - implement EV_AVOID_STDIO option. |
89 | - implement EV_AVOID_STDIO option. |
15 | - do not use sscanf to parse linux version number (smaller, faster, |
90 | - do not use sscanf to parse linux version number (smaller, faster, |
16 | no sscanf dependency). |
91 | no sscanf dependency). |
|
|
92 | - new EV_CHILD_ENABLE and EV_SIGNAL_ENABLE configurable settings. |
|
|
93 | - update libev.m4 HAVE_CLOCK_SYSCALL test for newer glibcs. |
|
|
94 | - add section on accept() problems to the manpage. |
|
|
95 | - rename EV_TIMEOUT to EV_TIMER. |
|
|
96 | - rename ev_loop_count/depth/verify/loop/unloop. |
|
|
97 | - remove ev_default_destroy and ev_default_fork. |
|
|
98 | - switch to two-digit minor version. |
|
|
99 | - work around an apparent gentoo compiler bug. |
|
|
100 | - define _DARWIN_UNLIMITED_SELECT. just so. |
|
|
101 | - use enum instead of #define for most constants. |
|
|
102 | - improve compatibility to older C++ compilers. |
|
|
103 | - (experimental) ev_run/ev_default_loop/ev_break/ev_loop_new have now |
|
|
104 | default arguments when compiled as C++. |
|
|
105 | - enable automake dependency tracking. |
|
|
106 | - ev_loop_new no longer leaks memory when loop creation failed. |
|
|
107 | - new ev_cleanup watcher type. |
17 | |
108 | |
18 | 3.9 Thu Dec 31 07:59:59 CET 2009 |
109 | 3.9 Thu Dec 31 07:59:59 CET 2009 |
19 | - signalfd is no longer used by default and has to be requested |
110 | - signalfd is no longer used by default and has to be requested |
20 | explicitly - this means that easy to catch bugs become hard to |
111 | explicitly - this means that easy to catch bugs become hard to |
21 | catch race conditions, but the users have spoken. |
112 | catch race conditions, but the users have spoken. |
… | |
… | |
37 | - changed minimum glibc requirement from glibc 2.9 to 2.7, for |
128 | - changed minimum glibc requirement from glibc 2.9 to 2.7, for |
38 | signalfd. |
129 | signalfd. |
39 | - add missing string.h include (Denis F. Latypoff). |
130 | - add missing string.h include (Denis F. Latypoff). |
40 | - only replace ev_stat.prev when we detect an actual difference, |
131 | - only replace ev_stat.prev when we detect an actual difference, |
41 | so prev is (almost) always different to attr. this might |
132 | so prev is (almost) always different to attr. this might |
42 | have caused the probems with 04_stat.t. |
133 | have caused the problems with 04_stat.t. |
43 | - add ev::timer->remaining () method to C++ API. |
134 | - add ev::timer->remaining () method to C++ API. |
44 | |
135 | |
45 | 3.8 Sun Aug 9 14:30:45 CEST 2009 |
136 | 3.8 Sun Aug 9 14:30:45 CEST 2009 |
46 | - incompatible change: do not necessarily reset signal handler |
137 | - incompatible change: do not necessarily reset signal handler |
47 | to SIG_DFL when a sighandler is stopped. |
138 | to SIG_DFL when a sighandler is stopped. |
48 | - ev_default_destroy did not properly free or zero some members, |
139 | - ev_default_destroy did not properly free or zero some members, |
49 | potentially causing crashes and memory corruption on repated |
140 | potentially causing crashes and memory corruption on repeated |
50 | ev_default_destroy/ev_default_loop calls. |
141 | ev_default_destroy/ev_default_loop calls. |
51 | - take advantage of signalfd on GNU/Linux systems. |
142 | - take advantage of signalfd on GNU/Linux systems. |
52 | - document that the signal mask might be in an unspecified |
143 | - document that the signal mask might be in an unspecified |
53 | state when using libev's signal handling. |
144 | state when using libev's signal handling. |
54 | - take advantage of some GNU/Linux calls to set cloexec/nonblock |
145 | - take advantage of some GNU/Linux calls to set cloexec/nonblock |
… | |
… | |
176 | do similar changes for event.c. |
267 | do similar changes for event.c. |
177 | |
268 | |
178 | 3.45 Tue Oct 21 21:59:26 CEST 2008 |
269 | 3.45 Tue Oct 21 21:59:26 CEST 2008 |
179 | - disable inotify usage on linux <2.6.25, as it is broken |
270 | - disable inotify usage on linux <2.6.25, as it is broken |
180 | (reported by Yoann Vandoorselaere). |
271 | (reported by Yoann Vandoorselaere). |
181 | - ev_stat errornously would try to add inotify watchers |
272 | - ev_stat erroneously would try to add inotify watchers |
182 | even when inotify wasn't available (this should only |
273 | even when inotify wasn't available (this should only |
183 | have a performance impact). |
274 | have a performance impact). |
184 | - ev_once now passes both timeout and io to the callback if both |
275 | - ev_once now passes both timeout and io to the callback if both |
185 | occur concurrently, instead of giving timeouts precedence. |
276 | occur concurrently, instead of giving timeouts precedence. |
186 | - disable EV_USE_INOTIFY when sys/inotify.h is too old. |
277 | - disable EV_USE_INOTIFY when sys/inotify.h is too old. |
… | |
… | |
254 | - enable inlining with gcc versions 2 and 3. |
345 | - enable inlining with gcc versions 2 and 3. |
255 | - work around broken poll implementations potentially |
346 | - work around broken poll implementations potentially |
256 | not clearing revents field in ev_poll (Brandon Black) |
347 | not clearing revents field in ev_poll (Brandon Black) |
257 | (no such systems are known at this time). |
348 | (no such systems are known at this time). |
258 | - work around a bug in realloc on openbsd and darwin, |
349 | - work around a bug in realloc on openbsd and darwin, |
259 | also makes the errornous valgrind complaints |
350 | also makes the erroneous valgrind complaints |
260 | go away (noted by various people). |
351 | go away (noted by various people). |
261 | - fix ev_async_pending, add c++ wrapper for ev_async |
352 | - fix ev_async_pending, add c++ wrapper for ev_async |
262 | (based on patch sent by Johannes Deisenhofer. |
353 | (based on patch sent by Johannes Deisenhofer). |
263 | - add sensible set method to ev::embed. |
354 | - add sensible set method to ev::embed. |
264 | - made integer constants type int in ev.h. |
355 | - made integer constants type int in ev.h. |
265 | |
356 | |
266 | 3.2 Wed Apr 2 17:11:19 CEST 2008 |
357 | 3.2 Wed Apr 2 17:11:19 CEST 2008 |
267 | - fix a 64 bit overflow issue in the select backend, |
358 | - fix a 64 bit overflow issue in the select backend, |