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

Comparing libev/Changes (file contents):
Revision 1.184 by root, Thu Oct 14 07:20:12 2010 UTC vs.
Revision 1.286 by sf-exg, Fri Mar 1 11:12:34 2013 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: ABI??? API????? Changes??? 3TODO: ev_loop_wakeup
4TODO: include ev_xyz_start in each example? 4TODO: EV_STANDALONE == NO_HASSEL (do not use clock_gettime in ev_standalone)
5TODO: section watcher states/lifetime 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
9
10TODO: document the special problem of signals around fork.
11TODO: store pid for each signal
12TODO: document file descriptor usage per loop
13TODO: store loop pid_t and compare isndie signal handler,store 1 for same, 2 for differign pid, clean up in loop_fork
14TODO: embed watchers need updating when fd changes
15TODO: document portbaility requirements for atomic pointer access
16 - destroying a non-default loop would stop the global waitpid
17 watcher (Denis Bilenko).
18 - queueing pending watchers of higher priority from a watcher now invokes
19 them in a timely fashion (reported by Denis Bilenko).
20 - add throw() to all libev functions that cannot throw exceptions, for
21 further code size decrease when compiling for C++.
22 - add throw () to callbacks that must not throw exceptions (allocator,
23 syserr, loop acquire/release, periodic reschedule cbs).
24 - fix event_base_loop return code, add event_get_callback, event_base_new,
25 event_base_get_method calls to improve libevent 1.x emulation and add
26 some libevent 2.x functionality (based on a patch by Jeff Davey).
27 - add more memory fences to fix a bug reported by Jeff Davey. Better
28 be overfenced than underprotected.
29 - ev_run now returns a boolean status (true meaning watchers are
30 still active).
31 - ev_once: undef EV_ERROR in ev_kqueue.c, to avoid clashing with
32 libev's EV_ERROR (reported by 191919).
33 - (ecb) add memory fence support for xlC (Darin McBride).
34 - (ecb) add memory fence support for gcc-mips (Anton Kirilov).
35 - (ecb) add memory fence support for gcc-alpha (Christian Weisgerber).
36 - work around some kernels losing file descriptors by leaking
37 the kqueue descriptor in the child.
38 - work around linux inotify not reporting IN_ATTRIB changes for directories
39 in many cases.
40 - include sys/syscall.h instead of plain syscall.h.
41 - check for io watcher loops in ev_verify, check for the most
42 common reported usage bug in ev_io_start.
43 - choose socket vs. WSASocket at compiletime using EV_USE_WSASOCKET.
44 - always use WSASend/WSARecv directly on windows, hoping that this
45 works in all cases (unlike read/write/send/recv...).
46 - try to detect signals around a fork faster (test program by
47 Denis Bilenko).
48 - work around recent glibc versions that leak memory in realloc.
49 - rename ev::embed::set to ev::embed::set_embed to avoid clashing
50 the watcher base set (loop) method.
51 - rewrite the async/signal pipe logic to always keep a valid fd, which
52 simplifies (and hopefully correctifies :) the race checking
53 on fork, at the cost of one extra fd.
54 - add fat, msdos, jffs2, ramfs, ntfs and btrfs to the list of
55 inotify-supporting filesystems.
56 - move orig_CFLAGS assignment to after AC_INIT, as newer autoconf
57 versions ignore it before
58 (https://bugzilla.redhat.com/show_bug.cgi?id=908096).
59 - add some untested android support.
60 - enum expressions must be of type int (reported by Juan Pablo L).
61
624.11 Sat Feb 4 19:52:39 CET 2012
63 - INCOMPATIBLE CHANGE: ev_timer_again now clears the pending status, as
64 was documented already, but not implemented in the repeating case.
65 - new compiletime symbols: EV_NO_SMP and EV_NO_THREADS.
66 - fix a race where the workaround against the epoll fork bugs
67 caused signals to not be handled anymore.
68 - correct backend_fudge for most backends, and implement a windows
69 specific workaround to avoid looping because we call both
70 select and Sleep, both with different time resolutions.
71 - document range and guarantees of ev_sleep.
72 - document reasonable ranges for periodics interval and offset.
73 - rename backend_fudge to backend_mintime to avoid future confusion :)
74 - change the default periodic reschedule function to hopefully be more
75 exact and correct even in corner cases or in the far future.
76 - do not rely on -lm anymore: use it when available but use our
77 own floor () if it is missing. This should make it easier to embed,
78 as no external libraries are required.
79 - strategically import macros from libecb and mark rarely-used functions
80 as cache-cold (saving almost 2k code size on typical amd64 setups).
81 - add Symbols.ev and Symbols.event files, that were missing.
82 - fix backend_mintime value for epoll (was 1/1024, is 1/1000 now).
83 - fix #3 "be smart about timeouts" to not "deadlock" when
84 timeout == now, also improve the section overall.
85 - avoid "AVOIDING FINISHING BEFORE RETURNING" idiom.
86 - support new EV_API_STATIC mode to make all libev symbols
87 static.
88 - supply default CFLAGS of -g -O3 with gcc when original CFLAGS
89 were empty.
90
914.04 Wed Feb 16 09:01:51 CET 2011
92 - fix two problems in the native win32 backend, where reuse of fd's
93 with different underlying handles caused handles not to be removed
94 or added to the select set (analyzed and tested by Bert Belder).
95 - do no rely on ceil() in ev_e?poll.c.
96 - backport libev to HP-UX versions before 11 v3.
97 - configure did not detect nanosleep and clock_gettime properly when
98 they are available in the libc (as opposed to -lrt).
99
1004.03 Tue Jan 11 14:37:25 CET 2011
101 - officially support polling files with all backends.
102 - support files, /dev/zero etc. the same way as select in the epoll
103 backend, by generating events on our own.
104 - ports backend: work around solaris bug 6874410 and many related ones
105 (EINTR, maybe more), with no performance loss (note that the solaris
106 bug report is actually wrong, reality is far more bizarre and broken
107 than that).
108 - define EV_READ/EV_WRITE as macros in event.h, as some programs use
109 #ifdef to test for them.
110 - new (experimental) function: ev_feed_signal.
111 - new (to become default) EVFLAG_NOSIGMASK flag.
112 - new EVBACKEND_MASK symbol.
113 - updated COMMON IDIOMS SECTION.
114
1154.01 Fri Nov 5 21:51:29 CET 2010
116 - automake fucked it up, apparently, --add-missing -f is not quite enough
117 to make it update its files, so 4.00 didn't install ev++.h and
118 event.h on make install. grrr.
119 - ev_loop(count|depth) didn't return anything (Robin Haberkorn).
120 - change EV_UNDEF to 0xffffffff to silence some overzealous compilers.
121 - use "(libev) " prefix for all libev error messages now.
122
1234.00 Mon Oct 25 12:32:12 CEST 2010
6 - "PORTING FROM LIBEV 3.X TO 4.X" (in ev.pod) is recommended reading. 124 - "PORTING FROM LIBEV 3.X TO 4.X" (in ev.pod) is recommended reading.
7 - ev_embed_stop did not correctly stop the watcher (very good 125 - ev_embed_stop did not correctly stop the watcher (very good
8 testcase by Vladimir Timofeev). 126 testcase by Vladimir Timofeev).
127 - ev_run will now always update the current loop time - it erroneously
128 didn't when idle watchers were active, causing timers not to fire.
9 - fix a bug where a timeout of zero caused the timer not to fire 129 - fix a bug where a timeout of zero caused the timer not to fire
10 in the libevent emulation (testcase by Péter Szabó). 130 in the libevent emulation (testcase by Péter Szabó).
11 - applied win32 fixes by Michael Lenaghan (also James Mansion). 131 - applied win32 fixes by Michael Lenaghan (also James Mansion).
12 - replace EV_MINIMAL by EV_FEATURES. 132 - replace EV_MINIMAL by EV_FEATURES.
13 - prefer EPOLL_CTL_ADD over EPOLL_CTL_MOD in some more cases, as it 133 - prefer EPOLL_CTL_ADD over EPOLL_CTL_MOD in some more cases, as it
14 seems the former is *much* faster than the latter. 134 seems the former is *much* faster than the latter.
135 - linux kernel version detection (for inotify bug workarounds)
136 did not work properly.
137 - reduce the number of spurious wake-ups with the ports backend.
15 - remove dependency on sys/queue.h on freebsd (patch by Vanilla Hsu). 138 - remove dependency on sys/queue.h on freebsd (patch by Vanilla Hsu).
139 - do async init within ev_async_start, not ev_async_set, which avoids
140 an API quirk where the set function must be called in the C++ API
141 even when there is nothing to set.
16 - add (undocumented) EV_ENABLE when adding events with kqueue, 142 - add (undocumented) EV_ENABLE when adding events with kqueue,
17 this might help with OS X, which seems to need it despite documenting 143 this might help with OS X, which seems to need it despite documenting
18 not to need it (helpfully pointed out by Tilghman Lesher). 144 not to need it (helpfully pointed out by Tilghman Lesher).
19 - do not use poll by default on freebsd, it's broken (what isn't 145 - do not use poll by default on freebsd, it's broken (what isn't
20 on freebsd...). 146 on freebsd...).
147 - allow to embed epoll on kernels >= 2.6.32.
21 - configure now prepends -O3, not appends it, so one can still 148 - configure now prepends -O3, not appends it, so one can still
22 override it. 149 override it.
23 - greatly expanded the portability section. 150 - ev.pod: greatly expanded the portability section, added a porting
151 section, a description of watcher states and made lots of minor fixes.
24 - disable poll backend on AIX, the poll header spams the namespace 152 - disable poll backend on AIX, the poll header spams the namespace
25 and it's not worth working around dead platforms (reported 153 and it's not worth working around dead platforms (reported
26 and analyzed by Aivars Kalvans). 154 and analyzed by Aivars Kalvans).
27 - improve header file compatibility of the standalone eventfd code 155 - improve header file compatibility of the standalone eventfd code
28 in an obscure case. 156 in an obscure case.
… …
31 no sscanf dependency). 159 no sscanf dependency).
32 - new EV_CHILD_ENABLE and EV_SIGNAL_ENABLE configurable settings. 160 - new EV_CHILD_ENABLE and EV_SIGNAL_ENABLE configurable settings.
33 - update libev.m4 HAVE_CLOCK_SYSCALL test for newer glibcs. 161 - update libev.m4 HAVE_CLOCK_SYSCALL test for newer glibcs.
34 - add section on accept() problems to the manpage. 162 - add section on accept() problems to the manpage.
35 - rename EV_TIMEOUT to EV_TIMER. 163 - rename EV_TIMEOUT to EV_TIMER.
36 - rename ev_loop_count/depth/verify. 164 - rename ev_loop_count/depth/verify/loop/unloop.
165 - remove ev_default_destroy and ev_default_fork.
37 - switch to two-digit minor version. 166 - switch to two-digit minor version.
38 - work around an apparent gentoo compiler bug. 167 - work around an apparent gentoo compiler bug.
39 - define _DARWIN_UNLIMITED_SELECT. just so. 168 - define _DARWIN_UNLIMITED_SELECT. just so.
169 - use enum instead of #define for most constants.
170 - improve compatibility to older C++ compilers.
171 - (experimental) ev_run/ev_default_loop/ev_break/ev_loop_new have now
172 default arguments when compiled as C++.
173 - enable automake dependency tracking.
174 - ev_loop_new no longer leaks memory when loop creation failed.
175 - new ev_cleanup watcher type.
40 176
413.9 Thu Dec 31 07:59:59 CET 2009 1773.9 Thu Dec 31 07:59:59 CET 2009
42 - signalfd is no longer used by default and has to be requested 178 - signalfd is no longer used by default and has to be requested
43 explicitly - this means that easy to catch bugs become hard to 179 explicitly - this means that easy to catch bugs become hard to
44 catch race conditions, but the users have spoken. 180 catch race conditions, but the users have spoken.
45 - point out the unspecified signal mask in the documentation, and 181 - point out the unspecified signal mask in the documentation, and
46 that this is a race condition regardless of EV_SIGNALFD. 182 that this is a race condition regardless of EV_SIGNALFD.
47 - backport inotify code to C89. 183 - backport inotify code to C89.
48 - inotify file descriptors could leak into child processes. 184 - inotify file descriptors could leak into child processes.
49 - ev_stat watchers could keep an errornous extra ref on the loop, 185 - ev_stat watchers could keep an erroneous extra ref on the loop,
50 preventing exit when unregistering all watchers (testcases 186 preventing exit when unregistering all watchers (testcases
51 provided by ry@tinyclouds.org). 187 provided by ry@tinyclouds.org).
52 - implement EV_WIN32_HANDLE_TO_FD and EV_WIN32_CLOSE_FD configuration 188 - implement EV_WIN32_HANDLE_TO_FD and EV_WIN32_CLOSE_FD configuration
53 symbols to make it easier for apps to do their own fd management. 189 symbols to make it easier for apps to do their own fd management.
54 - support EV_IDLE_ENABLE being disabled in ev++.h 190 - support EV_IDLE_ENABLE being disabled in ev++.h
… …
60 - changed minimum glibc requirement from glibc 2.9 to 2.7, for 196 - changed minimum glibc requirement from glibc 2.9 to 2.7, for
61 signalfd. 197 signalfd.
62 - add missing string.h include (Denis F. Latypoff). 198 - add missing string.h include (Denis F. Latypoff).
63 - only replace ev_stat.prev when we detect an actual difference, 199 - only replace ev_stat.prev when we detect an actual difference,
64 so prev is (almost) always different to attr. this might 200 so prev is (almost) always different to attr. this might
65 have caused the probems with 04_stat.t. 201 have caused the problems with 04_stat.t.
66 - add ev::timer->remaining () method to C++ API. 202 - add ev::timer->remaining () method to C++ API.
67 203
683.8 Sun Aug 9 14:30:45 CEST 2009 2043.8 Sun Aug 9 14:30:45 CEST 2009
69 - incompatible change: do not necessarily reset signal handler 205 - incompatible change: do not necessarily reset signal handler
70 to SIG_DFL when a sighandler is stopped. 206 to SIG_DFL when a sighandler is stopped.
71 - ev_default_destroy did not properly free or zero some members, 207 - ev_default_destroy did not properly free or zero some members,
72 potentially causing crashes and memory corruption on repated 208 potentially causing crashes and memory corruption on repeated
73 ev_default_destroy/ev_default_loop calls. 209 ev_default_destroy/ev_default_loop calls.
74 - take advantage of signalfd on GNU/Linux systems. 210 - take advantage of signalfd on GNU/Linux systems.
75 - document that the signal mask might be in an unspecified 211 - document that the signal mask might be in an unspecified
76 state when using libev's signal handling. 212 state when using libev's signal handling.
77 - take advantage of some GNU/Linux calls to set cloexec/nonblock 213 - take advantage of some GNU/Linux calls to set cloexec/nonblock
… …
112 - correctly zero the idx values for arrays, so destroying and 248 - correctly zero the idx values for arrays, so destroying and
113 reinitialising the default loop actually works (patch by 249 reinitialising the default loop actually works (patch by
114 Malek Hadj-Ali). 250 Malek Hadj-Ali).
115 - implement ev_suspend and ev_resume. 251 - implement ev_suspend and ev_resume.
116 - new EV_CUSTOM revents flag for use by applications. 252 - new EV_CUSTOM revents flag for use by applications.
117 - add documentation section about priorites. 253 - add documentation section about priorities.
118 - add a glossary to the dcoumentation. 254 - add a glossary to the documentation.
119 - extend the ev_fork description slightly. 255 - extend the ev_fork description slightly.
120 - optimize a jump out of call_pending. 256 - optimize a jump out of call_pending.
121 257
1223.53 Sun Feb 15 02:38:20 CET 2009 2583.53 Sun Feb 15 02:38:20 CET 2009
123 - fix a bug in event pipe creation on win32 that would cause a 259 - fix a bug in event pipe creation on win32 that would cause a
… …
199 do similar changes for event.c. 335 do similar changes for event.c.
200 336
2013.45 Tue Oct 21 21:59:26 CEST 2008 3373.45 Tue Oct 21 21:59:26 CEST 2008
202 - disable inotify usage on linux <2.6.25, as it is broken 338 - disable inotify usage on linux <2.6.25, as it is broken
203 (reported by Yoann Vandoorselaere). 339 (reported by Yoann Vandoorselaere).
204 - ev_stat errornously would try to add inotify watchers 340 - ev_stat erroneously would try to add inotify watchers
205 even when inotify wasn't available (this should only 341 even when inotify wasn't available (this should only
206 have a performance impact). 342 have a performance impact).
207 - ev_once now passes both timeout and io to the callback if both 343 - ev_once now passes both timeout and io to the callback if both
208 occur concurrently, instead of giving timeouts precedence. 344 occur concurrently, instead of giving timeouts precedence.
209 - disable EV_USE_INOTIFY when sys/inotify.h is too old. 345 - disable EV_USE_INOTIFY when sys/inotify.h is too old.
… …
277 - enable inlining with gcc versions 2 and 3. 413 - enable inlining with gcc versions 2 and 3.
278 - work around broken poll implementations potentially 414 - work around broken poll implementations potentially
279 not clearing revents field in ev_poll (Brandon Black) 415 not clearing revents field in ev_poll (Brandon Black)
280 (no such systems are known at this time). 416 (no such systems are known at this time).
281 - work around a bug in realloc on openbsd and darwin, 417 - work around a bug in realloc on openbsd and darwin,
282 also makes the errornous valgrind complaints 418 also makes the erroneous valgrind complaints
283 go away (noted by various people). 419 go away (noted by various people).
284 - fix ev_async_pending, add c++ wrapper for ev_async 420 - fix ev_async_pending, add c++ wrapper for ev_async
285 (based on patch sent by Johannes Deisenhofer. 421 (based on patch sent by Johannes Deisenhofer).
286 - add sensible set method to ev::embed. 422 - add sensible set method to ev::embed.
287 - made integer constants type int in ev.h. 423 - made integer constants type int in ev.h.
288 424
2893.2 Wed Apr 2 17:11:19 CEST 2008 4253.2 Wed Apr 2 17:11:19 CEST 2008
290 - fix a 64 bit overflow issue in the select backend, 426 - fix a 64 bit overflow issue in the select backend,

Diff Legend

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