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

Comparing libev/Changes (file contents):
Revision 1.270 by root, Sun May 6 19:47:38 2012 UTC vs.
Revision 1.350 by root, Sat Dec 21 16:11:51 2019 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_wakeup 3TODO: revisit 59.x timer in the light of modern powersaving
4TODO: EV_STANDALONE == NO_HASSEL (do not use clock_gettime in ev_standalone) 4TODO: document EV_TSTAMP_T
5TODO: faq, process a thing in each iteration
6TODO: dbeugging tips, ev_verify, ev_init twice
7 5
64.31 Fri Dec 20 21:58:29 CET 2019
7 - handle backends with minimum wait time a bit better by not
8 waiting in the presence of already-expired timers
9 (behaviour reported by Felipe Gasper).
10 - new feature: use timerfd to detect timejumps quickly,
11 can be disabled with the new EVFLAG_NOTIMERFD loop flag.
12 - document EV_USE_SIGNALFD feature macro.
13
144.30 (EV only)
15 - change non-autoconf test for __kernel_rwf_t by testing
16 LINUX_VERSION_CODE, the most direct test I could find.
17 - fix a bug in the io_uring backend that polled the wrong
18 backend fd, causing it to not work in many cases.
19
204.29 (EV only)
21 - add io uring autoconf and non-autoconf detection.
22 - disable io_uring when some header files are too old.
23
244.28 (EV only)
25 - linuxaio backend resulted in random memory corruption
26 when loop is forked.
27 - linuxaio backend might have tried to cancel an iocb
28 multiple times (was unable to trigger this).
29 - linuxaio backend now employs a generation counter to
30 avoid handling spurious events from cancelled requests.
31 - io_cancel can return EINTR, deal with it. also, assume
32 io_submit also returns EINTR.
33 - fix some other minor bugs in linuxaio backend.
34 - ev_tstamp type can now be overriden by defining EV_TSTAMP_T.
35 - cleanup: replace expect_true/false and noinline by their
36 libecb counterparts.
37 - move syscall infrastructure from ev_linuxaio.c to ev.c.
38 - prepare io_uring integration.
39 - tweak ev_floor.
40 - epoll, poll, win32 Sleep and other places that use millisecond
41 reslution now all try to round up times.
42 - solaris port backend didn't compile.
43 - abstract time constants into their macros, for more flexibility.
44
454.27 Thu Jun 27 22:43:44 CEST 2019
46 - linux aio backend almost completely rewritten to work around its
47 limitations.
48 - linux aio backend now requires linux 4.19+.
49 - epoll backend now mandatory for linux aio backend.
50 - fail assertions more aggressively on invalid fd's detected
51 in the event loop, do not just silently fd_kill in case of
52 user error.
53 - ev_io_start/ev_io_stop now verify the watcher fd using
54 a syscall when EV_VERIFY is 2 or higher.
55
564.26 (EV only)
57 - update to libecb 0x00010006.
58 - new experimental linux aio backend (linux 4.18+).
59 - removed redundant 0-ptr check in ev_once.
60 - updated/extended ev_set_allocator documentation.
61 - replaced EMPTY2 macro by array_needsize_noinit.
62 - minor code cleanups.
63 - epoll backend now uses epoll_create1 also after fork.
64
654.25 Fri Dec 21 07:49:20 CET 2018
66 - INCOMPATIBLE CHANGE: EV_THROW was renamed to EV_NOEXCEPT
67 (EV_THROW still provided) and now uses noexcept on C++11 or newer.
68 - move the darwin select workaround higher in ev.c, as newer versions of
69 darwin managed to break their broken select even more.
70 - ANDROID => __ANDROID__ (reported by enh@google.com).
71 - disable epoll_create1 on android because it has broken header files
72 and google is unwilling to fix them (reported by enh@google.com).
73 - avoid a minor compilation warning on win32.
74 - c++: remove deprecated dynamic throw() specifications.
75 - c++: improve the (unsupported) bad_loop exception class.
76 - backport perl ev_periodic example to C, untested.
77 - update libecb, biggets change is to include a memory fence
78 in ECB_MEMORY_FENCE_RELEASE on x86/amd64.
79 - minor autoconf/automake modernisation.
80
814.24 Wed Dec 28 05:19:55 CET 2016
82 - bump version to 4.24, as the release tarball inexplicably
83 didn't have the right version in ev.h, even though the cvs-tagged
84 version did have the right one (reported by Ales Teska).
85
864.23 Wed Nov 16 18:23:41 CET 2016
87 - move some declarations at the beginning to help certain retarded
88 microsoft compilers, even though their documentation claims
89 otherwise (reported by Ruslan Osmanov).
90
914.22 Sun Dec 20 22:11:50 CET 2015
92 - when epoll detects unremovable fds in the fd set, rebuild
93 only the epoll descriptor, not the signal pipe, to avoid
94 SIGPIPE in ev_async_send. This doesn't solve it on fork,
95 so document what needs to be done in ev_loop_fork
96 (analyzed by Benjamin Mahler).
97 - remove superfluous sys/timeb.h include on win32
98 (analyzed by Jason Madden).
99 - updated libecb.
100
1014.20 Sat Jun 20 13:01:43 CEST 2015
102 - prefer noexcept over throw () with C++ 11.
103 - update ecb.h due to incompatibilities with c11.
104 - fix a potential aliasing issue when reading and writing
105 watcher callbacks.
106
1074.19 Thu Sep 25 08:18:25 CEST 2014
108 - ev.h wasn't valid C++ anymore, which tripped compilers other than
109 clang, msvc or gcc (analyzed by Raphael 'kena' Poss). Unfortunately,
110 C++ doesn't support typedefs for function pointers fully, so the affected
111 declarations have to spell out the types each time.
112 - when not using autoconf, tighten the check for clock_gettime and related
113 functionality.
114
1154.18 Fri Sep 5 17:55:26 CEST 2014
116 - events on files were not always generated properly with the
117 epoll backend (testcase by Assaf Inbal).
118 - mark event pipe fd as cloexec after a fork (analyzed by Sami Farin).
119 - (ecb) support m68k, m88k and sh (patch by Miod Vallat).
120 - use a reasonable fallback for EV_NSIG instead of erroring out
121 when we can't detect the signal set size.
122 - in the absence of autoconf, do not use the clock syscall
123 on glibc >= 2.17 (avoids the syscall AND -lrt on systems
124 doing clock_gettime in userspace).
125 - ensure extern "C" function pointers are used for externally-visible
126 loop callbacks (not watcher callbacks yet).
127 - (ecb) work around memory barriers and volatile apparently both being
128 broken in visual studio 2008 and later (analysed and patch by Nicolas Noble).
129
1304.15 Fri Mar 1 12:04:50 CET 2013
131 - destroying a non-default loop would stop the global waitpid
132 watcher (Denis Bilenko).
8 - queueing pending watchers of higher priority from a watcher now invokes 133 - queueing pending watchers of higher priority from a watcher now invokes
9 them in a timely fashion (reported by Denis Bilenko). 134 them in a timely fashion (reported by Denis Bilenko).
10 - add throw() to all libev functions that cannot throw exceptions, for 135 - add throw() to all libev functions that cannot throw exceptions, for
11 further code size decrease when compiling for C++. 136 further code size decrease when compiling for C++.
12 - add throw () to callbacks that must not throw exceptions (allocator, 137 - add throw () to callbacks that must not throw exceptions (allocator,
13 syserr, loop acquire/release, periodic reschedule cbs). 138 syserr, loop acquire/release, periodic reschedule cbs).
14 - fix event_base_loop return code, add event_get_callback, event_base_new, 139 - fix event_base_loop return code, add event_get_callback, event_base_new,
15 event_base_get_method calls to improve libevent 1.x emulation and add 140 event_base_get_method calls to improve libevent 1.x emulation and add
16 some libevent 2.x funcitonality (based on a patch by Jeff Davey). 141 some libevent 2.x functionality (based on a patch by Jeff Davey).
17 - add more memory fences to fix a bug reported by Jeff Davey. Better 142 - add more memory fences to fix a bug reported by Jeff Davey. Better
18 be overfenced than underprotected. 143 be overfenced than underprotected.
19 - ev_run now returns a boolean status (true meaning watchers are 144 - ev_run now returns a boolean status (true meaning watchers are
20 still active). 145 still active).
21 - ev_once: undef EV_ERROR in ev_kqueue.c, to avoid clashing with 146 - ev_once: undef EV_ERROR in ev_kqueue.c, to avoid clashing with
… …
23 - (ecb) add memory fence support for xlC (Darin McBride). 148 - (ecb) add memory fence support for xlC (Darin McBride).
24 - (ecb) add memory fence support for gcc-mips (Anton Kirilov). 149 - (ecb) add memory fence support for gcc-mips (Anton Kirilov).
25 - (ecb) add memory fence support for gcc-alpha (Christian Weisgerber). 150 - (ecb) add memory fence support for gcc-alpha (Christian Weisgerber).
26 - work around some kernels losing file descriptors by leaking 151 - work around some kernels losing file descriptors by leaking
27 the kqueue descriptor in the child. 152 the kqueue descriptor in the child.
153 - work around linux inotify not reporting IN_ATTRIB changes for directories
154 in many cases.
28 - include sys/syscall.h instead of plain syscall.h. 155 - include sys/syscall.h instead of plain syscall.h.
29 - check for io watcher loops in ev_verify, check for the most 156 - check for io watcher loops in ev_verify, check for the most
30 common reported usage bug in ev_io_start. 157 common reported usage bug in ev_io_start.
31 - chose socket vs. WSASocket at compiletime using EV_USE_WSASOCKET. 158 - choose socket vs. WSASocket at compiletime using EV_USE_WSASOCKET.
32 - always use WSASend/WSARecv directly on windows, hoping that this 159 - always use WSASend/WSARecv directly on windows, hoping that this
33 works in all cases (unlike read/write/send/recv...). 160 works in all cases (unlike read/write/send/recv...).
161 - try to detect signals around a fork faster (test program by
162 Denis Bilenko).
163 - work around recent glibc versions that leak memory in realloc.
164 - rename ev::embed::set to ev::embed::set_embed to avoid clashing
165 the watcher base set (loop) method.
166 - rewrite the async/signal pipe logic to always keep a valid fd, which
167 simplifies (and hopefully correctifies :) the race checking
168 on fork, at the cost of one extra fd.
169 - add fat, msdos, jffs2, ramfs, ntfs and btrfs to the list of
170 inotify-supporting filesystems.
171 - move orig_CFLAGS assignment to after AC_INIT, as newer autoconf
172 versions ignore it before
173 (https://bugzilla.redhat.com/show_bug.cgi?id=908096).
174 - add some untested android support.
175 - enum expressions must be of type int (reported by Juan Pablo L).
34 176
354.11 Sat Feb 4 19:52:39 CET 2012 1774.11 Sat Feb 4 19:52:39 CET 2012
36 - INCOMPATIBLE CHANGE: ev_timer_again now clears the pending status, as 178 - INCOMPATIBLE CHANGE: ev_timer_again now clears the pending status, as
37 was documented already, but not implemented in the repeating case. 179 was documented already, but not implemented in the repeating case.
38 - new compiletime symbols: EV_NO_SMP and EV_NO_THREADS. 180 - new compiletime symbols: EV_NO_SMP and EV_NO_THREADS.
… …
222 reinitialising the default loop actually works (patch by 364 reinitialising the default loop actually works (patch by
223 Malek Hadj-Ali). 365 Malek Hadj-Ali).
224 - implement ev_suspend and ev_resume. 366 - implement ev_suspend and ev_resume.
225 - new EV_CUSTOM revents flag for use by applications. 367 - new EV_CUSTOM revents flag for use by applications.
226 - add documentation section about priorities. 368 - add documentation section about priorities.
227 - add a glossary to the dcoumentation. 369 - add a glossary to the documentation.
228 - extend the ev_fork description slightly. 370 - extend the ev_fork description slightly.
229 - optimize a jump out of call_pending. 371 - optimize a jump out of call_pending.
230 372
2313.53 Sun Feb 15 02:38:20 CET 2009 3733.53 Sun Feb 15 02:38:20 CET 2009
232 - fix a bug in event pipe creation on win32 that would cause a 374 - fix a bug in event pipe creation on win32 that would cause a

Diff Legend

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