ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/libev/Changes
(Generate patch)

Comparing libev/Changes (file contents):
Revision 1.226 by root, Sun Feb 20 02:56:55 2011 UTC vs.
Revision 1.306 by sf-exg, Thu Jun 18 10:14:30 2015 UTC

1Revision history for libev, a high-performance and full-featured event loop. 1Revision history for libev, a high-performance and full-featured event loop.
2 2
3TODO: ev_loop_prefork hint? 3TODO: ev_loop_wakeup
4TODO: EV_STANDALONE == NO_HASSEL (do not use clock_gettime in ev_standalone)
5TODO: faq, process a thing in each iteration
6TODO: dbeugging tips, ev_verify, ev_init twice
7TODO: ev_break for immediate exit (EVBREAK_NOW?)
8TODO: ev_feed_child_event
9TODO: document the special problem of signals around fork.
10TODO: store pid for each signal
11TODO: document file descriptor usage per loop
12TODO: store loop pid_t and compare isndie signal handler,store 1 for same, 2 for differign pid, clean up in loop_fork
13TODO: embed watchers need updating when fd changes
14TODO: document portability requirements for atomic pointer access
15TODO: document requirements for function pointers and calling conventions.
4 16
17 - prefer noexcept over throw () with C++ 11.
18 - update ecb.h due to incompatibilities with c11.
19 - fix a potential aliasing issue when reading and writing
20 watcher callbacks.
21
224.19 Thu Sep 25 08:18:25 CEST 2014
23 - ev.h wasn't valid C++ anymore, which tripped compilers other than
24 clang, msvc or gcc (analyzed by Raphael 'kena' Poss). Unfortunately,
25 C++ doesn't support typedefs for function pointers fully, so the affected
26 declarations have to spell out the types each time.
27 - when not using autoconf, tighten the check for clock_gettime and related
28 functionality.
29
304.18 Fri Sep 5 17:55:26 CEST 2014
31 - events on files were not always generated properly with the
32 epoll backend (testcase by Assaf Inbal).
33 - mark event pipe fd as cloexec after a fork (analyzed by Sami Farin).
34 - (ecb) support m68k, m88k and sh (patch by Miod Vallat).
35 - use a reasonable fallback for EV_NSIG instead of erroring out
36 when we can't detect the signal set size.
37 - in the absence of autoconf, do not use the clock syscall
38 on glibc >= 2.17 (avoids the syscall AND -lrt on systems
39 doing clock_gettime in userspace).
40 - ensure extern "C" function pointers are used for externally-visible
41 loop callbacks (not watcher callbacks yet).
42 - (ecb) work around memory barriers and volatile apparently both being
43 broken in visual studio 2008 and later (analysed and patch by Nicolas Noble).
44
454.15 Fri Mar 1 12:04:50 CET 2013
46 - destroying a non-default loop would stop the global waitpid
47 watcher (Denis Bilenko).
48 - queueing pending watchers of higher priority from a watcher now invokes
49 them in a timely fashion (reported by Denis Bilenko).
50 - add throw() to all libev functions that cannot throw exceptions, for
51 further code size decrease when compiling for C++.
52 - add throw () to callbacks that must not throw exceptions (allocator,
53 syserr, loop acquire/release, periodic reschedule cbs).
54 - fix event_base_loop return code, add event_get_callback, event_base_new,
55 event_base_get_method calls to improve libevent 1.x emulation and add
56 some libevent 2.x functionality (based on a patch by Jeff Davey).
57 - add more memory fences to fix a bug reported by Jeff Davey. Better
58 be overfenced than underprotected.
59 - ev_run now returns a boolean status (true meaning watchers are
60 still active).
61 - ev_once: undef EV_ERROR in ev_kqueue.c, to avoid clashing with
62 libev's EV_ERROR (reported by 191919).
63 - (ecb) add memory fence support for xlC (Darin McBride).
64 - (ecb) add memory fence support for gcc-mips (Anton Kirilov).
65 - (ecb) add memory fence support for gcc-alpha (Christian Weisgerber).
66 - work around some kernels losing file descriptors by leaking
67 the kqueue descriptor in the child.
68 - work around linux inotify not reporting IN_ATTRIB changes for directories
69 in many cases.
70 - include sys/syscall.h instead of plain syscall.h.
71 - check for io watcher loops in ev_verify, check for the most
72 common reported usage bug in ev_io_start.
73 - choose socket vs. WSASocket at compiletime using EV_USE_WSASOCKET.
74 - always use WSASend/WSARecv directly on windows, hoping that this
75 works in all cases (unlike read/write/send/recv...).
76 - try to detect signals around a fork faster (test program by
77 Denis Bilenko).
78 - work around recent glibc versions that leak memory in realloc.
79 - rename ev::embed::set to ev::embed::set_embed to avoid clashing
80 the watcher base set (loop) method.
81 - rewrite the async/signal pipe logic to always keep a valid fd, which
82 simplifies (and hopefully correctifies :) the race checking
83 on fork, at the cost of one extra fd.
84 - add fat, msdos, jffs2, ramfs, ntfs and btrfs to the list of
85 inotify-supporting filesystems.
86 - move orig_CFLAGS assignment to after AC_INIT, as newer autoconf
87 versions ignore it before
88 (https://bugzilla.redhat.com/show_bug.cgi?id=908096).
89 - add some untested android support.
90 - enum expressions must be of type int (reported by Juan Pablo L).
91
924.11 Sat Feb 4 19:52:39 CET 2012
93 - INCOMPATIBLE CHANGE: ev_timer_again now clears the pending status, as
94 was documented already, but not implemented in the repeating case.
95 - new compiletime symbols: EV_NO_SMP and EV_NO_THREADS.
96 - fix a race where the workaround against the epoll fork bugs
97 caused signals to not be handled anymore.
98 - correct backend_fudge for most backends, and implement a windows
99 specific workaround to avoid looping because we call both
100 select and Sleep, both with different time resolutions.
101 - document range and guarantees of ev_sleep.
102 - document reasonable ranges for periodics interval and offset.
103 - rename backend_fudge to backend_mintime to avoid future confusion :)
5 - change the default periodic reschedule function to hopefully be more 104 - change the default periodic reschedule function to hopefully be more
6 exact and correct even in corner cases or in the far future. 105 exact and correct even in corner cases or in the far future.
7 - document reasonable ranges for interval and offset.
8 - do not rely on -lm anymore: use it when available but use our 106 - do not rely on -lm anymore: use it when available but use our
9 own floor () if it is missing. This should make it easier to embed. 107 own floor () if it is missing. This should make it easier to embed,
108 as no external libraries are required.
109 - strategically import macros from libecb and mark rarely-used functions
110 as cache-cold (saving almost 2k code size on typical amd64 setups).
111 - add Symbols.ev and Symbols.event files, that were missing.
112 - fix backend_mintime value for epoll (was 1/1024, is 1/1000 now).
113 - fix #3 "be smart about timeouts" to not "deadlock" when
114 timeout == now, also improve the section overall.
115 - avoid "AVOIDING FINISHING BEFORE RETURNING" idiom.
116 - support new EV_API_STATIC mode to make all libev symbols
117 static.
118 - supply default CFLAGS of -g -O3 with gcc when original CFLAGS
119 were empty.
10 120
114.04 Wed Feb 16 09:01:51 CET 2011 1214.04 Wed Feb 16 09:01:51 CET 2011
12 - fix two problems in the native win32 backend, where reuse of fd's 122 - fix two problems in the native win32 backend, where reuse of fd's
13 with different underlying handles caused handles not to be removed 123 with different underlying handles caused handles not to be removed
14 or added to the select set (analyzed and tested by Bert Belder). 124 or added to the select set (analyzed and tested by Bert Belder).
… …
100 catch race conditions, but the users have spoken. 210 catch race conditions, but the users have spoken.
101 - point out the unspecified signal mask in the documentation, and 211 - point out the unspecified signal mask in the documentation, and
102 that this is a race condition regardless of EV_SIGNALFD. 212 that this is a race condition regardless of EV_SIGNALFD.
103 - backport inotify code to C89. 213 - backport inotify code to C89.
104 - inotify file descriptors could leak into child processes. 214 - inotify file descriptors could leak into child processes.
105 - ev_stat watchers could keep an errornous extra ref on the loop, 215 - ev_stat watchers could keep an erroneous extra ref on the loop,
106 preventing exit when unregistering all watchers (testcases 216 preventing exit when unregistering all watchers (testcases
107 provided by ry@tinyclouds.org). 217 provided by ry@tinyclouds.org).
108 - implement EV_WIN32_HANDLE_TO_FD and EV_WIN32_CLOSE_FD configuration 218 - implement EV_WIN32_HANDLE_TO_FD and EV_WIN32_CLOSE_FD configuration
109 symbols to make it easier for apps to do their own fd management. 219 symbols to make it easier for apps to do their own fd management.
110 - support EV_IDLE_ENABLE being disabled in ev++.h 220 - support EV_IDLE_ENABLE being disabled in ev++.h
… …
168 - correctly zero the idx values for arrays, so destroying and 278 - correctly zero the idx values for arrays, so destroying and
169 reinitialising the default loop actually works (patch by 279 reinitialising the default loop actually works (patch by
170 Malek Hadj-Ali). 280 Malek Hadj-Ali).
171 - implement ev_suspend and ev_resume. 281 - implement ev_suspend and ev_resume.
172 - new EV_CUSTOM revents flag for use by applications. 282 - new EV_CUSTOM revents flag for use by applications.
173 - add documentation section about priorites. 283 - add documentation section about priorities.
174 - add a glossary to the dcoumentation. 284 - add a glossary to the documentation.
175 - extend the ev_fork description slightly. 285 - extend the ev_fork description slightly.
176 - optimize a jump out of call_pending. 286 - optimize a jump out of call_pending.
177 287
1783.53 Sun Feb 15 02:38:20 CET 2009 2883.53 Sun Feb 15 02:38:20 CET 2009
179 - fix a bug in event pipe creation on win32 that would cause a 289 - fix a bug in event pipe creation on win32 that would cause a

Diff Legend

– Removed lines
+ Added lines
< Changed lines
> Changed lines