| 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 | |
|
|
3 | TODO: ev_loop_wakeup |
|
|
4 | TODO: EV_STANDALONE == NO_HASSEL (do not use clock_gettime in ev_standalone) |
|
|
5 | |
|
|
6 | TODO: document WSA_EV_USE_SOCKET in win32 part |
|
|
7 | TODO: ^ OR use WSASend/WSARecv on the handle, which always works |
|
|
8 | TODO: assert on fd watcher linked list pointing to itself |
|
|
9 | - add throw() to all libev functions that cannot throw exceptions, for |
|
|
10 | further code size decrease when compiling for C++. |
|
|
11 | - add throw () to callbacks that must not throw exceptions (allocator, |
|
|
12 | syserr, loop acquire/release, periodic reschedule cbs). |
|
|
13 | - fix event_base_loop return code, add event_get_callback, event_base_new, |
|
|
14 | event_base_get_method calls to improve libevent 1.x emulation and add |
|
|
15 | some libevent 2.x funcitonality (based on a patch by Jeff Davey). |
|
|
16 | - ev_run now returns a boolean status (true meaning watchers are |
|
|
17 | still active). |
|
|
18 | - ev_once: undef EV_ERROR in ev_kqueue.c, to avoid clashing with |
|
|
19 | libev's EV_ERROR (reported by 191919). |
|
|
20 | - (ecb) add memory fence support for xlC (Darin McBride). |
|
|
21 | - (ecb) add memory fence support for gcc-mips (Anton Kirilov). |
|
|
22 | - (ecb) add memory fence support for gcc-alpha (Christian Weisgerber). |
|
|
23 | - work around some kernels losing file descriptors by leaking |
|
|
24 | the kqueue descriptor in the child. |
|
|
25 | |
|
|
26 | 4.11 Sat Feb 4 19:52:39 CET 2012 |
|
|
27 | - INCOMPATIBLE CHANGE: ev_timer_again now clears the pending status, as |
|
|
28 | was documented already, but not implemented in the repeating case. |
|
|
29 | - new compiletime symbols: EV_NO_SMP and EV_NO_THREADS. |
|
|
30 | - fix a race where the workaround against the epoll fork bugs |
|
|
31 | caused signals to not be handled anymore. |
|
|
32 | - correct backend_fudge for most backends, and implement a windows |
|
|
33 | specific workaround to avoid looping because we call both |
|
|
34 | select and Sleep, both with different time resolutions. |
|
|
35 | - document range and guarantees of ev_sleep. |
|
|
36 | - document reasonable ranges for periodics interval and offset. |
|
|
37 | - rename backend_fudge to backend_mintime to avoid future confusion :) |
|
|
38 | - change the default periodic reschedule function to hopefully be more |
|
|
39 | exact and correct even in corner cases or in the far future. |
|
|
40 | - do not rely on -lm anymore: use it when available but use our |
|
|
41 | own floor () if it is missing. This should make it easier to embed, |
|
|
42 | as no external libraries are required. |
|
|
43 | - strategically import macros from libecb and mark rarely-used functions |
|
|
44 | as cache-cold (saving almost 2k code size on typical amd64 setups). |
|
|
45 | - add Symbols.ev and Symbols.event files, that were missing. |
|
|
46 | - fix backend_mintime value for epoll (was 1/1024, is 1/1000 now). |
|
|
47 | - fix #3 "be smart about timeouts" to not "deadlock" when |
|
|
48 | timeout == now, also improve the section overall. |
|
|
49 | - avoid "AVOIDING FINISHING BEFORE RETURNING" idiom. |
|
|
50 | - support new EV_API_STATIC mode to make all libev symbols |
|
|
51 | static. |
|
|
52 | - supply default CFLAGS of -g -O3 with gcc when original CFLAGS |
|
|
53 | were empty. |
|
|
54 | |
|
|
55 | 4.04 Wed Feb 16 09:01:51 CET 2011 |
|
|
56 | - fix two problems in the native win32 backend, where reuse of fd's |
|
|
57 | with different underlying handles caused handles not to be removed |
|
|
58 | or added to the select set (analyzed and tested by Bert Belder). |
|
|
59 | - do no rely on ceil() in ev_e?poll.c. |
|
|
60 | - backport libev to HP-UX versions before 11 v3. |
|
|
61 | - configure did not detect nanosleep and clock_gettime properly when |
|
|
62 | they are available in the libc (as opposed to -lrt). |
|
|
63 | |
|
|
64 | 4.03 Tue Jan 11 14:37:25 CET 2011 |
|
|
65 | - officially support polling files with all backends. |
|
|
66 | - support files, /dev/zero etc. the same way as select in the epoll |
|
|
67 | backend, by generating events on our own. |
|
|
68 | - ports backend: work around solaris bug 6874410 and many related ones |
|
|
69 | (EINTR, maybe more), with no performance loss (note that the solaris |
|
|
70 | bug report is actually wrong, reality is far more bizarre and broken |
|
|
71 | than that). |
|
|
72 | - define EV_READ/EV_WRITE as macros in event.h, as some programs use |
|
|
73 | #ifdef to test for them. |
|
|
74 | - new (experimental) function: ev_feed_signal. |
|
|
75 | - new (to become default) EVFLAG_NOSIGMASK flag. |
|
|
76 | - new EVBACKEND_MASK symbol. |
|
|
77 | - updated COMMON IDIOMS SECTION. |
| 2 | |
78 | |
| 3 | 4.01 Fri Nov 5 21:51:29 CET 2010 |
79 | 4.01 Fri Nov 5 21:51:29 CET 2010 |
| 4 | - automake fucked it up, apparently, --add-missing -f is not quite enough |
80 | - automake fucked it up, apparently, --add-missing -f is not quite enough |
| 5 | to make it update its files, so 4.00 didn't install ev++.h and |
81 | to make it update its files, so 4.00 didn't install ev++.h and |
| 6 | event.h on make install. grrr. |
82 | event.h on make install. grrr. |
| … | |
… | |
| 68 | catch race conditions, but the users have spoken. |
144 | catch race conditions, but the users have spoken. |
| 69 | - point out the unspecified signal mask in the documentation, and |
145 | - point out the unspecified signal mask in the documentation, and |
| 70 | that this is a race condition regardless of EV_SIGNALFD. |
146 | that this is a race condition regardless of EV_SIGNALFD. |
| 71 | - backport inotify code to C89. |
147 | - backport inotify code to C89. |
| 72 | - inotify file descriptors could leak into child processes. |
148 | - inotify file descriptors could leak into child processes. |
| 73 | - ev_stat watchers could keep an errornous extra ref on the loop, |
149 | - ev_stat watchers could keep an erroneous extra ref on the loop, |
| 74 | preventing exit when unregistering all watchers (testcases |
150 | preventing exit when unregistering all watchers (testcases |
| 75 | provided by ry@tinyclouds.org). |
151 | provided by ry@tinyclouds.org). |
| 76 | - implement EV_WIN32_HANDLE_TO_FD and EV_WIN32_CLOSE_FD configuration |
152 | - implement EV_WIN32_HANDLE_TO_FD and EV_WIN32_CLOSE_FD configuration |
| 77 | symbols to make it easier for apps to do their own fd management. |
153 | symbols to make it easier for apps to do their own fd management. |
| 78 | - support EV_IDLE_ENABLE being disabled in ev++.h |
154 | - support EV_IDLE_ENABLE being disabled in ev++.h |
| … | |
… | |
| 136 | - correctly zero the idx values for arrays, so destroying and |
212 | - correctly zero the idx values for arrays, so destroying and |
| 137 | reinitialising the default loop actually works (patch by |
213 | reinitialising the default loop actually works (patch by |
| 138 | Malek Hadj-Ali). |
214 | Malek Hadj-Ali). |
| 139 | - implement ev_suspend and ev_resume. |
215 | - implement ev_suspend and ev_resume. |
| 140 | - new EV_CUSTOM revents flag for use by applications. |
216 | - new EV_CUSTOM revents flag for use by applications. |
| 141 | - add documentation section about priorites. |
217 | - add documentation section about priorities. |
| 142 | - add a glossary to the dcoumentation. |
218 | - add a glossary to the dcoumentation. |
| 143 | - extend the ev_fork description slightly. |
219 | - extend the ev_fork description slightly. |
| 144 | - optimize a jump out of call_pending. |
220 | - optimize a jump out of call_pending. |
| 145 | |
221 | |
| 146 | 3.53 Sun Feb 15 02:38:20 CET 2009 |
222 | 3.53 Sun Feb 15 02:38:20 CET 2009 |
