1 <?xml version="1.0" encoding="UTF-8"?>
2 <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.1//EN" "http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd">
3 <html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en">
6 <meta name="description" content="Pod documentation for libev" />
7 <meta name="inputfile" content="<standard input>" />
8 <meta name="outputfile" content="<standard output>" />
9 <meta name="created" content="Tue Nov 27 21:29:04 2007" />
10 <meta name="generator" content="Pod::Xhtml 1.57" />
11 <link rel="stylesheet" href="http://res.tst.eu/pod.css"/></head>
15 <h3 id="TOP">Index</h3>
17 <ul><li><a href="#NAME">NAME</a></li>
18 <li><a href="#SYNOPSIS">SYNOPSIS</a></li>
19 <li><a href="#EXAMPLE_PROGRAM">EXAMPLE PROGRAM</a></li>
20 <li><a href="#DESCRIPTION">DESCRIPTION</a></li>
21 <li><a href="#FEATURES">FEATURES</a></li>
22 <li><a href="#CONVENTIONS">CONVENTIONS</a></li>
23 <li><a href="#TIME_REPRESENTATION">TIME REPRESENTATION</a></li>
24 <li><a href="#GLOBAL_FUNCTIONS">GLOBAL FUNCTIONS</a></li>
25 <li><a href="#FUNCTIONS_CONTROLLING_THE_EVENT_LOOP">FUNCTIONS CONTROLLING THE EVENT LOOP</a></li>
26 <li><a href="#ANATOMY_OF_A_WATCHER">ANATOMY OF A WATCHER</a>
27 <ul><li><a href="#GENERIC_WATCHER_FUNCTIONS">GENERIC WATCHER FUNCTIONS</a></li>
28 <li><a href="#ASSOCIATING_CUSTOM_DATA_WITH_A_WATCH">ASSOCIATING CUSTOM DATA WITH A WATCHER</a></li>
31 <li><a href="#WATCHER_TYPES">WATCHER TYPES</a>
32 <ul><li><a href="#code_ev_io_code_is_this_file_descrip"><code>ev_io</code> - is this file descriptor readable or writable?</a></li>
33 <li><a href="#code_ev_timer_code_relative_and_opti"><code>ev_timer</code> - relative and optionally repeating timeouts</a></li>
34 <li><a href="#code_ev_periodic_code_to_cron_or_not"><code>ev_periodic</code> - to cron or not to cron?</a></li>
35 <li><a href="#code_ev_signal_code_signal_me_when_a"><code>ev_signal</code> - signal me when a signal gets signalled!</a></li>
36 <li><a href="#code_ev_child_code_watch_out_for_pro"><code>ev_child</code> - watch out for process status changes</a></li>
37 <li><a href="#code_ev_stat_code_did_the_file_attri"><code>ev_stat</code> - did the file attributes just change?</a></li>
38 <li><a href="#code_ev_idle_code_when_you_ve_got_no"><code>ev_idle</code> - when you've got nothing better to do...</a></li>
39 <li><a href="#code_ev_prepare_code_and_code_ev_che"><code>ev_prepare</code> and <code>ev_check</code> - customise your event loop!</a></li>
40 <li><a href="#code_ev_embed_code_when_one_backend_"><code>ev_embed</code> - when one backend isn't enough...</a></li>
41 <li><a href="#code_ev_fork_code_the_audacity_to_re"><code>ev_fork</code> - the audacity to resume the event loop after a fork</a></li>
44 <li><a href="#OTHER_FUNCTIONS">OTHER FUNCTIONS</a></li>
45 <li><a href="#LIBEVENT_EMULATION">LIBEVENT EMULATION</a></li>
46 <li><a href="#C_SUPPORT">C++ SUPPORT</a></li>
47 <li><a href="#MACRO_MAGIC">MACRO MAGIC</a></li>
48 <li><a href="#EMBEDDING">EMBEDDING</a>
49 <ul><li><a href="#FILESETS">FILESETS</a>
50 <ul><li><a href="#CORE_EVENT_LOOP">CORE EVENT LOOP</a></li>
51 <li><a href="#LIBEVENT_COMPATIBILITY_API">LIBEVENT COMPATIBILITY API</a></li>
52 <li><a href="#AUTOCONF_SUPPORT">AUTOCONF SUPPORT</a></li>
55 <li><a href="#PREPROCESSOR_SYMBOLS_MACROS">PREPROCESSOR SYMBOLS/MACROS</a></li>
56 <li><a href="#EXAMPLES">EXAMPLES</a></li>
59 <li><a href="#COMPLEXITIES">COMPLEXITIES</a></li>
60 <li><a href="#AUTHOR">AUTHOR</a>
65 <h1 id="NAME">NAME</h1>
66 <div id="NAME_CONTENT">
67 <p>libev - a high performance full-featured event loop written in C</p>
70 <h1 id="SYNOPSIS">SYNOPSIS</h1>
71 <div id="SYNOPSIS_CONTENT">
72 <pre> #include <ev.h>
77 <h1 id="EXAMPLE_PROGRAM">EXAMPLE PROGRAM</h1>
78 <div id="EXAMPLE_PROGRAM_CONTENT">
79 <pre> #include <ev.h>
82 ev_timer timeout_watcher;
84 /* called when data readable on stdin */
86 stdin_cb (EV_P_ struct ev_io *w, int revents)
88 /* puts ("stdin ready"); */
89 ev_io_stop (EV_A_ w); /* just a syntax example */
90 ev_unloop (EV_A_ EVUNLOOP_ALL); /* leave all loop calls */
94 timeout_cb (EV_P_ struct ev_timer *w, int revents)
96 /* puts ("timeout"); */
97 ev_unloop (EV_A_ EVUNLOOP_ONE); /* leave one loop call */
103 struct ev_loop *loop = ev_default_loop (0);
105 /* initialise an io watcher, then start it */
106 ev_io_init (&stdin_watcher, stdin_cb, /*STDIN_FILENO*/ 0, EV_READ);
107 ev_io_start (loop, &stdin_watcher);
109 /* simple non-repeating 5.5 second timeout */
110 ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);
111 ev_timer_start (loop, &timeout_watcher);
113 /* loop till timeout or data ready */
122 <h1 id="DESCRIPTION">DESCRIPTION</h1>
123 <div id="DESCRIPTION_CONTENT">
124 <p>Libev is an event loop: you register interest in certain events (such as a
125 file descriptor being readable or a timeout occuring), and it will manage
126 these event sources and provide your program with events.</p>
127 <p>To do this, it must take more or less complete control over your process
128 (or thread) by executing the <i>event loop</i> handler, and will then
129 communicate events via a callback mechanism.</p>
130 <p>You register interest in certain events by registering so-called <i>event
131 watchers</i>, which are relatively small C structures you initialise with the
132 details of the event, and then hand it over to libev by <i>starting</i> the
136 <h1 id="FEATURES">FEATURES</h1>
137 <div id="FEATURES_CONTENT">
138 <p>Libev supports <code>select</code>, <code>poll</code>, the linux-specific <code>epoll</code>, the
139 bsd-specific <code>kqueue</code> and the solaris-specific event port mechanisms
140 for file descriptor events (<code>ev_io</code>), relative timers (<code>ev_timer</code>),
141 absolute timers with customised rescheduling (<code>ev_periodic</code>), synchronous
142 signals (<code>ev_signal</code>), process status change events (<code>ev_child</code>), and
143 event watchers dealing with the event loop mechanism itself (<code>ev_idle</code>,
144 <code>ev_embed</code>, <code>ev_prepare</code> and <code>ev_check</code> watchers) as well as
145 file watchers (<code>ev_stat</code>) and even limited support for fork events
146 (<code>ev_fork</code>).</p>
147 <p>It also is quite fast (see this
148 <a href="http://libev.schmorp.de/bench.html">benchmark</a> comparing it to libevent
152 <h1 id="CONVENTIONS">CONVENTIONS</h1>
153 <div id="CONVENTIONS_CONTENT">
154 <p>Libev is very configurable. In this manual the default configuration will
155 be described, which supports multiple event loops. For more info about
156 various configuration options please have a look at <strong>EMBED</strong> section in
157 this manual. If libev was configured without support for multiple event
158 loops, then all functions taking an initial argument of name <code>loop</code>
159 (which is always of type <code>struct ev_loop *</code>) will not have this argument.</p>
162 <h1 id="TIME_REPRESENTATION">TIME REPRESENTATION</h1>
163 <div id="TIME_REPRESENTATION_CONTENT">
164 <p>Libev represents time as a single floating point number, representing the
165 (fractional) number of seconds since the (POSIX) epoch (somewhere near
166 the beginning of 1970, details are complicated, don't ask). This type is
167 called <code>ev_tstamp</code>, which is what you should use too. It usually aliases
168 to the <code>double</code> type in C, and when you need to do any calculations on
169 it, you should treat it as such.</p>
172 <h1 id="GLOBAL_FUNCTIONS">GLOBAL FUNCTIONS</h1>
173 <div id="GLOBAL_FUNCTIONS_CONTENT">
174 <p>These functions can be called anytime, even before initialising the
175 library in any way.</p>
177 <dt>ev_tstamp ev_time ()</dt>
179 <p>Returns the current time as libev would use it. Please note that the
180 <code>ev_now</code> function is usually faster and also often returns the timestamp
181 you actually want to know.</p>
183 <dt>int ev_version_major ()</dt>
184 <dt>int ev_version_minor ()</dt>
186 <p>You can find out the major and minor version numbers of the library
187 you linked against by calling the functions <code>ev_version_major</code> and
188 <code>ev_version_minor</code>. If you want, you can compare against the global
189 symbols <code>EV_VERSION_MAJOR</code> and <code>EV_VERSION_MINOR</code>, which specify the
190 version of the library your program was compiled against.</p>
191 <p>Usually, it's a good idea to terminate if the major versions mismatch,
192 as this indicates an incompatible change. Minor versions are usually
193 compatible to older versions, so a larger minor version alone is usually
195 <p>Example: Make sure we haven't accidentally been linked against the wrong
197 <pre> assert (("libev version mismatch",
198 ev_version_major () == EV_VERSION_MAJOR
199 && ev_version_minor () >= EV_VERSION_MINOR));
203 <dt>unsigned int ev_supported_backends ()</dt>
205 <p>Return the set of all backends (i.e. their corresponding <code>EV_BACKEND_*</code>
206 value) compiled into this binary of libev (independent of their
207 availability on the system you are running on). See <code>ev_default_loop</code> for
208 a description of the set values.</p>
209 <p>Example: make sure we have the epoll method, because yeah this is cool and
210 a must have and can we have a torrent of it please!!!11</p>
211 <pre> assert (("sorry, no epoll, no sex",
212 ev_supported_backends () & EVBACKEND_EPOLL));
216 <dt>unsigned int ev_recommended_backends ()</dt>
218 <p>Return the set of all backends compiled into this binary of libev and also
219 recommended for this platform. This set is often smaller than the one
220 returned by <code>ev_supported_backends</code>, as for example kqueue is broken on
221 most BSDs and will not be autodetected unless you explicitly request it
222 (assuming you know what you are doing). This is the set of backends that
223 libev will probe for if you specify no backends explicitly.</p>
225 <dt>unsigned int ev_embeddable_backends ()</dt>
227 <p>Returns the set of backends that are embeddable in other event loops. This
228 is the theoretical, all-platform, value. To find which backends
229 might be supported on the current system, you would need to look at
230 <code>ev_embeddable_backends () & ev_supported_backends ()</code>, likewise for
231 recommended ones.</p>
232 <p>See the description of <code>ev_embed</code> watchers for more info.</p>
234 <dt>ev_set_allocator (void *(*cb)(void *ptr, size_t size))</dt>
236 <p>Sets the allocation function to use (the prototype and semantics are
237 identical to the realloc C function). It is used to allocate and free
238 memory (no surprises here). If it returns zero when memory needs to be
239 allocated, the library might abort or take some potentially destructive
240 action. The default is your system realloc function.</p>
241 <p>You could override this function in high-availability programs to, say,
242 free some memory if it cannot allocate memory, to use a special allocator,
243 or even to sleep a while and retry until some memory is available.</p>
244 <p>Example: Replace the libev allocator with one that waits a bit and then
247 persistent_realloc (void *ptr, size_t size)
251 void *newptr = realloc (ptr, size);
261 ev_set_allocator (persistent_realloc);
265 <dt>ev_set_syserr_cb (void (*cb)(const char *msg));</dt>
267 <p>Set the callback function to call on a retryable syscall error (such
268 as failed select, poll, epoll_wait). The message is a printable string
269 indicating the system call or subsystem causing the problem. If this
270 callback is set, then libev will expect it to remedy the sitution, no
271 matter what, when it returns. That is, libev will generally retry the
272 requested operation, or, if the condition doesn't go away, do bad stuff
274 <p>Example: This is basically the same thing that libev does internally, too.</p>
276 fatal_error (const char *msg)
283 ev_set_syserr_cb (fatal_error);
290 <h1 id="FUNCTIONS_CONTROLLING_THE_EVENT_LOOP">FUNCTIONS CONTROLLING THE EVENT LOOP</h1>
291 <div id="FUNCTIONS_CONTROLLING_THE_EVENT_LOOP-2">
292 <p>An event loop is described by a <code>struct ev_loop *</code>. The library knows two
293 types of such loops, the <i>default</i> loop, which supports signals and child
294 events, and dynamically created loops which do not.</p>
295 <p>If you use threads, a common model is to run the default event loop
296 in your main thread (or in a separate thread) and for each thread you
297 create, you also create another event loop. Libev itself does no locking
298 whatsoever, so if you mix calls to the same event loop in different
299 threads, make sure you lock (this is usually a bad idea, though, even if
300 done correctly, because it's hideous and inefficient).</p>
302 <dt>struct ev_loop *ev_default_loop (unsigned int flags)</dt>
304 <p>This will initialise the default event loop if it hasn't been initialised
305 yet and return it. If the default loop could not be initialised, returns
306 false. If it already was initialised it simply returns it (and ignores the
307 flags. If that is troubling you, check <code>ev_backend ()</code> afterwards).</p>
308 <p>If you don't know what event loop to use, use the one returned from this
310 <p>The flags argument can be used to specify special behaviour or specific
311 backends to use, and is usually specified as <code>0</code> (or <code>EVFLAG_AUTO</code>).</p>
312 <p>The following flags are supported:</p>
315 <dt><code>EVFLAG_AUTO</code></dt>
317 <p>The default flags value. Use this if you have no clue (it's the right
318 thing, believe me).</p>
320 <dt><code>EVFLAG_NOENV</code></dt>
322 <p>If this flag bit is ored into the flag value (or the program runs setuid
323 or setgid) then libev will <i>not</i> look at the environment variable
324 <code>LIBEV_FLAGS</code>. Otherwise (the default), this environment variable will
325 override the flags completely if it is found in the environment. This is
326 useful to try out specific backends to test their performance, or to work
329 <dt><code>EVBACKEND_SELECT</code> (value 1, portable select backend)</dt>
331 <p>This is your standard select(2) backend. Not <i>completely</i> standard, as
332 libev tries to roll its own fd_set with no limits on the number of fds,
333 but if that fails, expect a fairly low limit on the number of fds when
334 using this backend. It doesn't scale too well (O(highest_fd)), but its usually
335 the fastest backend for a low number of fds.</p>
337 <dt><code>EVBACKEND_POLL</code> (value 2, poll backend, available everywhere except on windows)</dt>
339 <p>And this is your standard poll(2) backend. It's more complicated than
340 select, but handles sparse fds better and has no artificial limit on the
341 number of fds you can use (except it will slow down considerably with a
342 lot of inactive fds). It scales similarly to select, i.e. O(total_fds).</p>
344 <dt><code>EVBACKEND_EPOLL</code> (value 4, Linux)</dt>
346 <p>For few fds, this backend is a bit little slower than poll and select,
347 but it scales phenomenally better. While poll and select usually scale like
348 O(total_fds) where n is the total number of fds (or the highest fd), epoll scales
349 either O(1) or O(active_fds).</p>
350 <p>While stopping and starting an I/O watcher in the same iteration will
351 result in some caching, there is still a syscall per such incident
352 (because the fd could point to a different file description now), so its
353 best to avoid that. Also, dup()ed file descriptors might not work very
354 well if you register events for both fds.</p>
355 <p>Please note that epoll sometimes generates spurious notifications, so you
356 need to use non-blocking I/O or other means to avoid blocking when no data
357 (or space) is available.</p>
359 <dt><code>EVBACKEND_KQUEUE</code> (value 8, most BSD clones)</dt>
361 <p>Kqueue deserves special mention, as at the time of this writing, it
362 was broken on all BSDs except NetBSD (usually it doesn't work with
363 anything but sockets and pipes, except on Darwin, where of course its
364 completely useless). For this reason its not being "autodetected"
365 unless you explicitly specify it explicitly in the flags (i.e. using
366 <code>EVBACKEND_KQUEUE</code>).</p>
367 <p>It scales in the same way as the epoll backend, but the interface to the
368 kernel is more efficient (which says nothing about its actual speed, of
369 course). While starting and stopping an I/O watcher does not cause an
370 extra syscall as with epoll, it still adds up to four event changes per
371 incident, so its best to avoid that.</p>
373 <dt><code>EVBACKEND_DEVPOLL</code> (value 16, Solaris 8)</dt>
375 <p>This is not implemented yet (and might never be).</p>
377 <dt><code>EVBACKEND_PORT</code> (value 32, Solaris 10)</dt>
379 <p>This uses the Solaris 10 port mechanism. As with everything on Solaris,
380 it's really slow, but it still scales very well (O(active_fds)).</p>
381 <p>Please note that solaris ports can result in a lot of spurious
382 notifications, so you need to use non-blocking I/O or other means to avoid
383 blocking when no data (or space) is available.</p>
385 <dt><code>EVBACKEND_ALL</code></dt>
387 <p>Try all backends (even potentially broken ones that wouldn't be tried
388 with <code>EVFLAG_AUTO</code>). Since this is a mask, you can do stuff such as
389 <code>EVBACKEND_ALL & ~EVBACKEND_KQUEUE</code>.</p>
393 <p>If one or more of these are ored into the flags value, then only these
394 backends will be tried (in the reverse order as given here). If none are
395 specified, most compiled-in backend will be tried, usually in reverse
396 order of their flag values :)</p>
397 <p>The most typical usage is like this:</p>
398 <pre> if (!ev_default_loop (0))
399 fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
402 <p>Restrict libev to the select and poll backends, and do not allow
403 environment settings to be taken into account:</p>
404 <pre> ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);
407 <p>Use whatever libev has to offer, but make sure that kqueue is used if
408 available (warning, breaks stuff, best use only with your own private
409 event loop and only if you know the OS supports your types of fds):</p>
410 <pre> ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);
414 <dt>struct ev_loop *ev_loop_new (unsigned int flags)</dt>
416 <p>Similar to <code>ev_default_loop</code>, but always creates a new event loop that is
417 always distinct from the default loop. Unlike the default loop, it cannot
418 handle signal and child watchers, and attempts to do so will be greeted by
419 undefined behaviour (or a failed assertion if assertions are enabled).</p>
420 <p>Example: Try to create a event loop that uses epoll and nothing else.</p>
421 <pre> struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);
423 fatal ("no epoll found here, maybe it hides under your chair");
427 <dt>ev_default_destroy ()</dt>
429 <p>Destroys the default loop again (frees all memory and kernel state
430 etc.). None of the active event watchers will be stopped in the normal
431 sense, so e.g. <code>ev_is_active</code> might still return true. It is your
432 responsibility to either stop all watchers cleanly yoursef <i>before</i>
433 calling this function, or cope with the fact afterwards (which is usually
434 the easiest thing, youc na just ignore the watchers and/or <code>free ()</code> them
437 <dt>ev_loop_destroy (loop)</dt>
439 <p>Like <code>ev_default_destroy</code>, but destroys an event loop created by an
440 earlier call to <code>ev_loop_new</code>.</p>
442 <dt>ev_default_fork ()</dt>
444 <p>This function reinitialises the kernel state for backends that have
445 one. Despite the name, you can call it anytime, but it makes most sense
446 after forking, in either the parent or child process (or both, but that
447 again makes little sense).</p>
448 <p>You <i>must</i> call this function in the child process after forking if and
449 only if you want to use the event library in both processes. If you just
450 fork+exec, you don't have to call it.</p>
451 <p>The function itself is quite fast and it's usually not a problem to call
452 it just in case after a fork. To make this easy, the function will fit in
453 quite nicely into a call to <code>pthread_atfork</code>:</p>
454 <pre> pthread_atfork (0, 0, ev_default_fork);
457 <p>At the moment, <code>EVBACKEND_SELECT</code> and <code>EVBACKEND_POLL</code> are safe to use
458 without calling this function, so if you force one of those backends you
459 do not need to care.</p>
461 <dt>ev_loop_fork (loop)</dt>
463 <p>Like <code>ev_default_fork</code>, but acts on an event loop created by
464 <code>ev_loop_new</code>. Yes, you have to call this on every allocated event loop
465 after fork, and how you do this is entirely your own problem.</p>
467 <dt>unsigned int ev_backend (loop)</dt>
469 <p>Returns one of the <code>EVBACKEND_*</code> flags indicating the event backend in
472 <dt>ev_tstamp ev_now (loop)</dt>
474 <p>Returns the current "event loop time", which is the time the event loop
475 received events and started processing them. This timestamp does not
476 change as long as callbacks are being processed, and this is also the base
477 time used for relative timers. You can treat it as the timestamp of the
478 event occuring (or more correctly, libev finding out about it).</p>
480 <dt>ev_loop (loop, int flags)</dt>
482 <p>Finally, this is it, the event handler. This function usually is called
483 after you initialised all your watchers and you want to start handling
485 <p>If the flags argument is specified as <code>0</code>, it will not return until
486 either no event watchers are active anymore or <code>ev_unloop</code> was called.</p>
487 <p>Please note that an explicit <code>ev_unloop</code> is usually better than
488 relying on all watchers to be stopped when deciding when a program has
489 finished (especially in interactive programs), but having a program that
490 automatically loops as long as it has to and no longer by virtue of
491 relying on its watchers stopping correctly is a thing of beauty.</p>
492 <p>A flags value of <code>EVLOOP_NONBLOCK</code> will look for new events, will handle
493 those events and any outstanding ones, but will not block your process in
494 case there are no events and will return after one iteration of the loop.</p>
495 <p>A flags value of <code>EVLOOP_ONESHOT</code> will look for new events (waiting if
496 neccessary) and will handle those and any outstanding ones. It will block
497 your process until at least one new event arrives, and will return after
498 one iteration of the loop. This is useful if you are waiting for some
499 external event in conjunction with something not expressible using other
500 libev watchers. However, a pair of <code>ev_prepare</code>/<code>ev_check</code> watchers is
501 usually a better approach for this kind of thing.</p>
502 <p>Here are the gory details of what <code>ev_loop</code> does:</p>
503 <pre> * If there are no active watchers (reference count is zero), return.
504 - Queue prepare watchers and then call all outstanding watchers.
505 - If we have been forked, recreate the kernel state.
506 - Update the kernel state with all outstanding changes.
507 - Update the "event loop time".
508 - Calculate for how long to block.
509 - Block the process, waiting for any events.
510 - Queue all outstanding I/O (fd) events.
511 - Update the "event loop time" and do time jump handling.
512 - Queue all outstanding timers.
513 - Queue all outstanding periodics.
514 - If no events are pending now, queue all idle watchers.
515 - Queue all check watchers.
516 - Call all queued watchers in reverse order (i.e. check watchers first).
517 Signals and child watchers are implemented as I/O watchers, and will
518 be handled here by queueing them when their watcher gets executed.
519 - If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK
520 were used, return, otherwise continue with step *.
523 <p>Example: Queue some jobs and then loop until no events are outsanding
525 <pre> ... queue jobs here, make sure they register event watchers as long
526 ... as they still have work to do (even an idle watcher will do..)
527 ev_loop (my_loop, 0);
532 <dt>ev_unloop (loop, how)</dt>
534 <p>Can be used to make a call to <code>ev_loop</code> return early (but only after it
535 has processed all outstanding events). The <code>how</code> argument must be either
536 <code>EVUNLOOP_ONE</code>, which will make the innermost <code>ev_loop</code> call return, or
537 <code>EVUNLOOP_ALL</code>, which will make all nested <code>ev_loop</code> calls return.</p>
539 <dt>ev_ref (loop)</dt>
540 <dt>ev_unref (loop)</dt>
542 <p>Ref/unref can be used to add or remove a reference count on the event
543 loop: Every watcher keeps one reference, and as long as the reference
544 count is nonzero, <code>ev_loop</code> will not return on its own. If you have
545 a watcher you never unregister that should not keep <code>ev_loop</code> from
546 returning, ev_unref() after starting, and ev_ref() before stopping it. For
547 example, libev itself uses this for its internal signal pipe: It is not
548 visible to the libev user and should not keep <code>ev_loop</code> from exiting if
549 no event watchers registered by it are active. It is also an excellent
550 way to do this for generic recurring timers or from within third-party
551 libraries. Just remember to <i>unref after start</i> and <i>ref before stop</i>.</p>
552 <p>Example: Create a signal watcher, but keep it from keeping <code>ev_loop</code>
553 running when nothing else is active.</p>
554 <pre> struct ev_signal exitsig;
555 ev_signal_init (&exitsig, sig_cb, SIGINT);
556 ev_signal_start (loop, &exitsig);
560 <p>Example: For some weird reason, unregister the above signal handler again.</p>
562 ev_signal_stop (loop, &exitsig);
573 <h1 id="ANATOMY_OF_A_WATCHER">ANATOMY OF A WATCHER</h1>
574 <div id="ANATOMY_OF_A_WATCHER_CONTENT">
575 <p>A watcher is a structure that you create and register to record your
576 interest in some event. For instance, if you want to wait for STDIN to
577 become readable, you would create an <code>ev_io</code> watcher for that:</p>
578 <pre> static void my_cb (struct ev_loop *loop, struct ev_io *w, int revents)
581 ev_unloop (loop, EVUNLOOP_ALL);
584 struct ev_loop *loop = ev_default_loop (0);
585 struct ev_io stdin_watcher;
586 ev_init (&stdin_watcher, my_cb);
587 ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);
588 ev_io_start (loop, &stdin_watcher);
592 <p>As you can see, you are responsible for allocating the memory for your
593 watcher structures (and it is usually a bad idea to do this on the stack,
594 although this can sometimes be quite valid).</p>
595 <p>Each watcher structure must be initialised by a call to <code>ev_init
596 (watcher *, callback)</code>, which expects a callback to be provided. This
597 callback gets invoked each time the event occurs (or, in the case of io
598 watchers, each time the event loop detects that the file descriptor given
599 is readable and/or writable).</p>
600 <p>Each watcher type has its own <code>ev_<type>_set (watcher *, ...)</code> macro
601 with arguments specific to this watcher type. There is also a macro
602 to combine initialisation and setting in one call: <code>ev_<type>_init
603 (watcher *, callback, ...)</code>.</p>
604 <p>To make the watcher actually watch out for events, you have to start it
605 with a watcher-specific start function (<code>ev_<type>_start (loop, watcher
606 *)</code>), and you can stop watching for events at any time by calling the
607 corresponding stop function (<code>ev_<type>_stop (loop, watcher *)</code>.</p>
608 <p>As long as your watcher is active (has been started but not stopped) you
609 must not touch the values stored in it. Most specifically you must never
610 reinitialise it or call its <code>set</code> macro.</p>
611 <p>Each and every callback receives the event loop pointer as first, the
612 registered watcher structure as second, and a bitset of received events as
614 <p>The received events usually include a single bit per event type received
615 (you can receive multiple events at the same time). The possible bit masks
618 <dt><code>EV_READ</code></dt>
619 <dt><code>EV_WRITE</code></dt>
621 <p>The file descriptor in the <code>ev_io</code> watcher has become readable and/or
624 <dt><code>EV_TIMEOUT</code></dt>
626 <p>The <code>ev_timer</code> watcher has timed out.</p>
628 <dt><code>EV_PERIODIC</code></dt>
630 <p>The <code>ev_periodic</code> watcher has timed out.</p>
632 <dt><code>EV_SIGNAL</code></dt>
634 <p>The signal specified in the <code>ev_signal</code> watcher has been received by a thread.</p>
636 <dt><code>EV_CHILD</code></dt>
638 <p>The pid specified in the <code>ev_child</code> watcher has received a status change.</p>
640 <dt><code>EV_STAT</code></dt>
642 <p>The path specified in the <code>ev_stat</code> watcher changed its attributes somehow.</p>
644 <dt><code>EV_IDLE</code></dt>
646 <p>The <code>ev_idle</code> watcher has determined that you have nothing better to do.</p>
648 <dt><code>EV_PREPARE</code></dt>
649 <dt><code>EV_CHECK</code></dt>
651 <p>All <code>ev_prepare</code> watchers are invoked just <i>before</i> <code>ev_loop</code> starts
652 to gather new events, and all <code>ev_check</code> watchers are invoked just after
653 <code>ev_loop</code> has gathered them, but before it invokes any callbacks for any
654 received events. Callbacks of both watcher types can start and stop as
655 many watchers as they want, and all of them will be taken into account
656 (for example, a <code>ev_prepare</code> watcher might start an idle watcher to keep
657 <code>ev_loop</code> from blocking).</p>
659 <dt><code>EV_EMBED</code></dt>
661 <p>The embedded event loop specified in the <code>ev_embed</code> watcher needs attention.</p>
663 <dt><code>EV_FORK</code></dt>
665 <p>The event loop has been resumed in the child process after fork (see
666 <code>ev_fork</code>).</p>
668 <dt><code>EV_ERROR</code></dt>
670 <p>An unspecified error has occured, the watcher has been stopped. This might
671 happen because the watcher could not be properly started because libev
672 ran out of memory, a file descriptor was found to be closed or any other
673 problem. You best act on it by reporting the problem and somehow coping
674 with the watcher being stopped.</p>
675 <p>Libev will usually signal a few "dummy" events together with an error,
676 for example it might indicate that a fd is readable or writable, and if
677 your callbacks is well-written it can just attempt the operation and cope
678 with the error from read() or write(). This will not work in multithreaded
679 programs, though, so beware.</p>
684 <h2 id="GENERIC_WATCHER_FUNCTIONS">GENERIC WATCHER FUNCTIONS</h2>
685 <div id="GENERIC_WATCHER_FUNCTIONS_CONTENT">
686 <p>In the following description, <code>TYPE</code> stands for the watcher type,
687 e.g. <code>timer</code> for <code>ev_timer</code> watchers and <code>io</code> for <code>ev_io</code> watchers.</p>
689 <dt><code>ev_init</code> (ev_TYPE *watcher, callback)</dt>
691 <p>This macro initialises the generic portion of a watcher. The contents
692 of the watcher object can be arbitrary (so <code>malloc</code> will do). Only
693 the generic parts of the watcher are initialised, you <i>need</i> to call
694 the type-specific <code>ev_TYPE_set</code> macro afterwards to initialise the
695 type-specific parts. For each type there is also a <code>ev_TYPE_init</code> macro
696 which rolls both calls into one.</p>
697 <p>You can reinitialise a watcher at any time as long as it has been stopped
698 (or never started) and there are no pending events outstanding.</p>
699 <p>The callback is always of type <code>void (*)(ev_loop *loop, ev_TYPE *watcher,
700 int revents)</code>.</p>
702 <dt><code>ev_TYPE_set</code> (ev_TYPE *, [args])</dt>
704 <p>This macro initialises the type-specific parts of a watcher. You need to
705 call <code>ev_init</code> at least once before you call this macro, but you can
706 call <code>ev_TYPE_set</code> any number of times. You must not, however, call this
707 macro on a watcher that is active (it can be pending, however, which is a
708 difference to the <code>ev_init</code> macro).</p>
709 <p>Although some watcher types do not have type-specific arguments
710 (e.g. <code>ev_prepare</code>) you still need to call its <code>set</code> macro.</p>
712 <dt><code>ev_TYPE_init</code> (ev_TYPE *watcher, callback, [args])</dt>
714 <p>This convinience macro rolls both <code>ev_init</code> and <code>ev_TYPE_set</code> macro
715 calls into a single call. This is the most convinient method to initialise
716 a watcher. The same limitations apply, of course.</p>
718 <dt><code>ev_TYPE_start</code> (loop *, ev_TYPE *watcher)</dt>
720 <p>Starts (activates) the given watcher. Only active watchers will receive
721 events. If the watcher is already active nothing will happen.</p>
723 <dt><code>ev_TYPE_stop</code> (loop *, ev_TYPE *watcher)</dt>
725 <p>Stops the given watcher again (if active) and clears the pending
726 status. It is possible that stopped watchers are pending (for example,
727 non-repeating timers are being stopped when they become pending), but
728 <code>ev_TYPE_stop</code> ensures that the watcher is neither active nor pending. If
729 you want to free or reuse the memory used by the watcher it is therefore a
730 good idea to always call its <code>ev_TYPE_stop</code> function.</p>
732 <dt>bool ev_is_active (ev_TYPE *watcher)</dt>
734 <p>Returns a true value iff the watcher is active (i.e. it has been started
735 and not yet been stopped). As long as a watcher is active you must not modify
738 <dt>bool ev_is_pending (ev_TYPE *watcher)</dt>
740 <p>Returns a true value iff the watcher is pending, (i.e. it has outstanding
741 events but its callback has not yet been invoked). As long as a watcher
742 is pending (but not active) you must not call an init function on it (but
743 <code>ev_TYPE_set</code> is safe) and you must make sure the watcher is available to
744 libev (e.g. you cnanot <code>free ()</code> it).</p>
746 <dt>callback = ev_cb (ev_TYPE *watcher)</dt>
748 <p>Returns the callback currently set on the watcher.</p>
750 <dt>ev_cb_set (ev_TYPE *watcher, callback)</dt>
752 <p>Change the callback. You can change the callback at virtually any time
753 (modulo threads).</p>
762 <h2 id="ASSOCIATING_CUSTOM_DATA_WITH_A_WATCH">ASSOCIATING CUSTOM DATA WITH A WATCHER</h2>
763 <div id="ASSOCIATING_CUSTOM_DATA_WITH_A_WATCH-2">
764 <p>Each watcher has, by default, a member <code>void *data</code> that you can change
765 and read at any time, libev will completely ignore it. This can be used
766 to associate arbitrary data with your watcher. If you need more data and
767 don't want to allocate memory and store a pointer to it in that data
768 member, you can also "subclass" the watcher type and provide your own
775 struct whatever *mostinteresting;
779 <p>And since your callback will be called with a pointer to the watcher, you
780 can cast it back to your own type:</p>
781 <pre> static void my_cb (struct ev_loop *loop, struct ev_io *w_, int revents)
783 struct my_io *w = (struct my_io *)w_;
788 <p>More interesting and less C-conformant ways of catsing your callback type
789 have been omitted....</p>
796 <h1 id="WATCHER_TYPES">WATCHER TYPES</h1>
797 <div id="WATCHER_TYPES_CONTENT">
798 <p>This section describes each watcher in detail, but will not repeat
799 information given in the last section. Any initialisation/set macros,
800 functions and members specific to the watcher type are explained.</p>
801 <p>Members are additionally marked with either <i>[read-only]</i>, meaning that,
802 while the watcher is active, you can look at the member and expect some
803 sensible content, but you must not modify it (you can modify it while the
804 watcher is stopped to your hearts content), or <i>[read-write]</i>, which
805 means you can expect it to have some sensible content while the watcher
806 is active, but you can also modify it. Modifying it may not do something
807 sensible or take immediate effect (or do anything at all), but libev will
808 not crash or malfunction in any way.</p>
815 <h2 id="code_ev_io_code_is_this_file_descrip"><code>ev_io</code> - is this file descriptor readable or writable?</h2>
816 <div id="code_ev_io_code_is_this_file_descrip-2">
817 <p>I/O watchers check whether a file descriptor is readable or writable
818 in each iteration of the event loop, or, more precisely, when reading
819 would not block the process and writing would at least be able to write
820 some data. This behaviour is called level-triggering because you keep
821 receiving events as long as the condition persists. Remember you can stop
822 the watcher if you don't want to act on the event and neither want to
823 receive future events.</p>
824 <p>In general you can register as many read and/or write event watchers per
825 fd as you want (as long as you don't confuse yourself). Setting all file
826 descriptors to non-blocking mode is also usually a good idea (but not
827 required if you know what you are doing).</p>
828 <p>You have to be careful with dup'ed file descriptors, though. Some backends
829 (the linux epoll backend is a notable example) cannot handle dup'ed file
830 descriptors correctly if you register interest in two or more fds pointing
831 to the same underlying file/socket/etc. description (that is, they share
832 the same underlying "file open").</p>
833 <p>If you must do this, then force the use of a known-to-be-good backend
834 (at the time of this writing, this includes only <code>EVBACKEND_SELECT</code> and
835 <code>EVBACKEND_POLL</code>).</p>
836 <p>Another thing you have to watch out for is that it is quite easy to
837 receive "spurious" readyness notifications, that is your callback might
838 be called with <code>EV_READ</code> but a subsequent <code>read</code>(2) will actually block
839 because there is no data. Not only are some backends known to create a
840 lot of those (for example solaris ports), it is very easy to get into
841 this situation even with a relatively standard program structure. Thus
842 it is best to always use non-blocking I/O: An extra <code>read</code>(2) returning
843 <code>EAGAIN</code> is far preferable to a program hanging until some data arrives.</p>
844 <p>If you cannot run the fd in non-blocking mode (for example you should not
845 play around with an Xlib connection), then you have to seperately re-test
846 wether a file descriptor is really ready with a known-to-be good interface
847 such as poll (fortunately in our Xlib example, Xlib already does this on
848 its own, so its quite safe to use).</p>
850 <dt>ev_io_init (ev_io *, callback, int fd, int events)</dt>
851 <dt>ev_io_set (ev_io *, int fd, int events)</dt>
853 <p>Configures an <code>ev_io</code> watcher. The <code>fd</code> is the file descriptor to
854 rceeive events for and events is either <code>EV_READ</code>, <code>EV_WRITE</code> or
855 <code>EV_READ | EV_WRITE</code> to receive the given events.</p>
857 <dt>int fd [read-only]</dt>
859 <p>The file descriptor being watched.</p>
861 <dt>int events [read-only]</dt>
863 <p>The events being watched.</p>
866 <p>Example: Call <code>stdin_readable_cb</code> when STDIN_FILENO has become, well
867 readable, but only once. Since it is likely line-buffered, you could
868 attempt to read a whole line in the callback.</p>
870 stdin_readable_cb (struct ev_loop *loop, struct ev_io *w, int revents)
872 ev_io_stop (loop, w);
873 .. read from stdin here (or from w->fd) and haqndle any I/O errors
877 struct ev_loop *loop = ev_default_init (0);
878 struct ev_io stdin_readable;
879 ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
880 ev_io_start (loop, &stdin_readable);
889 <h2 id="code_ev_timer_code_relative_and_opti"><code>ev_timer</code> - relative and optionally repeating timeouts</h2>
890 <div id="code_ev_timer_code_relative_and_opti-2">
891 <p>Timer watchers are simple relative timers that generate an event after a
892 given time, and optionally repeating in regular intervals after that.</p>
893 <p>The timers are based on real time, that is, if you register an event that
894 times out after an hour and you reset your system clock to last years
895 time, it will still time out after (roughly) and hour. "Roughly" because
896 detecting time jumps is hard, and some inaccuracies are unavoidable (the
897 monotonic clock option helps a lot here).</p>
898 <p>The relative timeouts are calculated relative to the <code>ev_now ()</code>
899 time. This is usually the right thing as this timestamp refers to the time
900 of the event triggering whatever timeout you are modifying/starting. If
901 you suspect event processing to be delayed and you <i>need</i> to base the timeout
902 on the current time, use something like this to adjust for this:</p>
903 <pre> ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);
906 <p>The callback is guarenteed to be invoked only when its timeout has passed,
907 but if multiple timers become ready during the same loop iteration then
908 order of execution is undefined.</p>
910 <dt>ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)</dt>
911 <dt>ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)</dt>
913 <p>Configure the timer to trigger after <code>after</code> seconds. If <code>repeat</code> is
914 <code>0.</code>, then it will automatically be stopped. If it is positive, then the
915 timer will automatically be configured to trigger again <code>repeat</code> seconds
916 later, again, and again, until stopped manually.</p>
917 <p>The timer itself will do a best-effort at avoiding drift, that is, if you
918 configure a timer to trigger every 10 seconds, then it will trigger at
919 exactly 10 second intervals. If, however, your program cannot keep up with
920 the timer (because it takes longer than those 10 seconds to do stuff) the
921 timer will not fire more than once per event loop iteration.</p>
923 <dt>ev_timer_again (loop)</dt>
925 <p>This will act as if the timer timed out and restart it again if it is
926 repeating. The exact semantics are:</p>
927 <p>If the timer is started but nonrepeating, stop it.</p>
928 <p>If the timer is repeating, either start it if necessary (with the repeat
929 value), or reset the running timer to the repeat value.</p>
930 <p>This sounds a bit complicated, but here is a useful and typical
931 example: Imagine you have a tcp connection and you want a so-called
932 idle timeout, that is, you want to be called when there have been,
933 say, 60 seconds of inactivity on the socket. The easiest way to do
934 this is to configure an <code>ev_timer</code> with <code>after</code>=<code>repeat</code>=<code>60</code> and calling
935 <code>ev_timer_again</code> each time you successfully read or write some data. If
936 you go into an idle state where you do not expect data to travel on the
937 socket, you can stop the timer, and again will automatically restart it if
939 <p>You can also ignore the <code>after</code> value and <code>ev_timer_start</code> altogether
940 and only ever use the <code>repeat</code> value:</p>
941 <pre> ev_timer_init (timer, callback, 0., 5.);
942 ev_timer_again (loop, timer);
944 timer->again = 17.;
945 ev_timer_again (loop, timer);
947 timer->again = 10.;
948 ev_timer_again (loop, timer);
951 <p>This is more efficient then stopping/starting the timer eahc time you want
952 to modify its timeout value.</p>
954 <dt>ev_tstamp repeat [read-write]</dt>
956 <p>The current <code>repeat</code> value. Will be used each time the watcher times out
957 or <code>ev_timer_again</code> is called and determines the next timeout (if any),
958 which is also when any modifications are taken into account.</p>
961 <p>Example: Create a timer that fires after 60 seconds.</p>
963 one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
965 .. one minute over, w is actually stopped right here
968 struct ev_timer mytimer;
969 ev_timer_init (&mytimer, one_minute_cb, 60., 0.);
970 ev_timer_start (loop, &mytimer);
973 <p>Example: Create a timeout timer that times out after 10 seconds of
976 timeout_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
978 .. ten seconds without any activity
981 struct ev_timer mytimer;
982 ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */
983 ev_timer_again (&mytimer); /* start timer */
986 // and in some piece of code that gets executed on any "activity":
987 // reset the timeout to start ticking again at 10 seconds
988 ev_timer_again (&mytimer);
996 <h2 id="code_ev_periodic_code_to_cron_or_not"><code>ev_periodic</code> - to cron or not to cron?</h2>
997 <div id="code_ev_periodic_code_to_cron_or_not-2">
998 <p>Periodic watchers are also timers of a kind, but they are very versatile
999 (and unfortunately a bit complex).</p>
1000 <p>Unlike <code>ev_timer</code>'s, they are not based on real time (or relative time)
1001 but on wallclock time (absolute time). You can tell a periodic watcher
1002 to trigger "at" some specific point in time. For example, if you tell a
1003 periodic watcher to trigger in 10 seconds (by specifiying e.g. <code>ev_now ()
1004 + 10.</code>) and then reset your system clock to the last year, then it will
1005 take a year to trigger the event (unlike an <code>ev_timer</code>, which would trigger
1006 roughly 10 seconds later and of course not if you reset your system time
1008 <p>They can also be used to implement vastly more complex timers, such as
1009 triggering an event on eahc midnight, local time.</p>
1010 <p>As with timers, the callback is guarenteed to be invoked only when the
1011 time (<code>at</code>) has been passed, but if multiple periodic timers become ready
1012 during the same loop iteration then order of execution is undefined.</p>
1014 <dt>ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)</dt>
1015 <dt>ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)</dt>
1017 <p>Lots of arguments, lets sort it out... There are basically three modes of
1018 operation, and we will explain them from simplest to complex:</p>
1021 <dt>* absolute timer (interval = reschedule_cb = 0)</dt>
1023 <p>In this configuration the watcher triggers an event at the wallclock time
1024 <code>at</code> and doesn't repeat. It will not adjust when a time jump occurs,
1025 that is, if it is to be run at January 1st 2011 then it will run when the
1026 system time reaches or surpasses this time.</p>
1028 <dt>* non-repeating interval timer (interval > 0, reschedule_cb = 0)</dt>
1030 <p>In this mode the watcher will always be scheduled to time out at the next
1031 <code>at + N * interval</code> time (for some integer N) and then repeat, regardless
1032 of any time jumps.</p>
1033 <p>This can be used to create timers that do not drift with respect to system
1035 <pre> ev_periodic_set (&periodic, 0., 3600., 0);
1038 <p>This doesn't mean there will always be 3600 seconds in between triggers,
1039 but only that the the callback will be called when the system time shows a
1040 full hour (UTC), or more correctly, when the system time is evenly divisible
1042 <p>Another way to think about it (for the mathematically inclined) is that
1043 <code>ev_periodic</code> will try to run the callback in this mode at the next possible
1044 time where <code>time = at (mod interval)</code>, regardless of any time jumps.</p>
1046 <dt>* manual reschedule mode (reschedule_cb = callback)</dt>
1048 <p>In this mode the values for <code>interval</code> and <code>at</code> are both being
1049 ignored. Instead, each time the periodic watcher gets scheduled, the
1050 reschedule callback will be called with the watcher as first, and the
1051 current time as second argument.</p>
1052 <p>NOTE: <i>This callback MUST NOT stop or destroy any periodic watcher,
1053 ever, or make any event loop modifications</i>. If you need to stop it,
1054 return <code>now + 1e30</code> (or so, fudge fudge) and stop it afterwards (e.g. by
1055 starting a prepare watcher).</p>
1056 <p>Its prototype is <code>ev_tstamp (*reschedule_cb)(struct ev_periodic *w,
1057 ev_tstamp now)</code>, e.g.:</p>
1058 <pre> static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)
1064 <p>It must return the next time to trigger, based on the passed time value
1065 (that is, the lowest time value larger than to the second argument). It
1066 will usually be called just before the callback will be triggered, but
1067 might be called at other times, too.</p>
1068 <p>NOTE: <i>This callback must always return a time that is later than the
1069 passed <code>now</code> value</i>. Not even <code>now</code> itself will do, it <i>must</i> be larger.</p>
1070 <p>This can be used to create very complex timers, such as a timer that
1071 triggers on each midnight, local time. To do this, you would calculate the
1072 next midnight after <code>now</code> and return the timestamp value for this. How
1073 you do this is, again, up to you (but it is not trivial, which is the main
1074 reason I omitted it as an example).</p>
1079 <dt>ev_periodic_again (loop, ev_periodic *)</dt>
1081 <p>Simply stops and restarts the periodic watcher again. This is only useful
1082 when you changed some parameters or the reschedule callback would return
1083 a different time than the last time it was called (e.g. in a crond like
1084 program when the crontabs have changed).</p>
1086 <dt>ev_tstamp interval [read-write]</dt>
1088 <p>The current interval value. Can be modified any time, but changes only
1089 take effect when the periodic timer fires or <code>ev_periodic_again</code> is being
1092 <dt>ev_tstamp (*reschedule_cb)(struct ev_periodic *w, ev_tstamp now) [read-write]</dt>
1094 <p>The current reschedule callback, or <code>0</code>, if this functionality is
1095 switched off. Can be changed any time, but changes only take effect when
1096 the periodic timer fires or <code>ev_periodic_again</code> is being called.</p>
1099 <p>Example: Call a callback every hour, or, more precisely, whenever the
1100 system clock is divisible by 3600. The callback invocation times have
1101 potentially a lot of jittering, but good long-term stability.</p>
1103 clock_cb (struct ev_loop *loop, struct ev_io *w, int revents)
1105 ... its now a full hour (UTC, or TAI or whatever your clock follows)
1108 struct ev_periodic hourly_tick;
1109 ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);
1110 ev_periodic_start (loop, &hourly_tick);
1113 <p>Example: The same as above, but use a reschedule callback to do it:</p>
1114 <pre> #include <math.h>
1117 my_scheduler_cb (struct ev_periodic *w, ev_tstamp now)
1119 return fmod (now, 3600.) + 3600.;
1122 ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
1125 <p>Example: Call a callback every hour, starting now:</p>
1126 <pre> struct ev_periodic hourly_tick;
1127 ev_periodic_init (&hourly_tick, clock_cb,
1128 fmod (ev_now (loop), 3600.), 3600., 0);
1129 ev_periodic_start (loop, &hourly_tick);
1137 <h2 id="code_ev_signal_code_signal_me_when_a"><code>ev_signal</code> - signal me when a signal gets signalled!</h2>
1138 <div id="code_ev_signal_code_signal_me_when_a-2">
1139 <p>Signal watchers will trigger an event when the process receives a specific
1140 signal one or more times. Even though signals are very asynchronous, libev
1141 will try it's best to deliver signals synchronously, i.e. as part of the
1142 normal event processing, like any other event.</p>
1143 <p>You can configure as many watchers as you like per signal. Only when the
1144 first watcher gets started will libev actually register a signal watcher
1145 with the kernel (thus it coexists with your own signal handlers as long
1146 as you don't register any with libev). Similarly, when the last signal
1147 watcher for a signal is stopped libev will reset the signal handler to
1148 SIG_DFL (regardless of what it was set to before).</p>
1150 <dt>ev_signal_init (ev_signal *, callback, int signum)</dt>
1151 <dt>ev_signal_set (ev_signal *, int signum)</dt>
1153 <p>Configures the watcher to trigger on the given signal number (usually one
1154 of the <code>SIGxxx</code> constants).</p>
1156 <dt>int signum [read-only]</dt>
1158 <p>The signal the watcher watches out for.</p>
1167 <h2 id="code_ev_child_code_watch_out_for_pro"><code>ev_child</code> - watch out for process status changes</h2>
1168 <div id="code_ev_child_code_watch_out_for_pro-2">
1169 <p>Child watchers trigger when your process receives a SIGCHLD in response to
1170 some child status changes (most typically when a child of yours dies).</p>
1172 <dt>ev_child_init (ev_child *, callback, int pid)</dt>
1173 <dt>ev_child_set (ev_child *, int pid)</dt>
1175 <p>Configures the watcher to wait for status changes of process <code>pid</code> (or
1176 <i>any</i> process if <code>pid</code> is specified as <code>0</code>). The callback can look
1177 at the <code>rstatus</code> member of the <code>ev_child</code> watcher structure to see
1178 the status word (use the macros from <code>sys/wait.h</code> and see your systems
1179 <code>waitpid</code> documentation). The <code>rpid</code> member contains the pid of the
1180 process causing the status change.</p>
1182 <dt>int pid [read-only]</dt>
1184 <p>The process id this watcher watches out for, or <code>0</code>, meaning any process id.</p>
1186 <dt>int rpid [read-write]</dt>
1188 <p>The process id that detected a status change.</p>
1190 <dt>int rstatus [read-write]</dt>
1192 <p>The process exit/trace status caused by <code>rpid</code> (see your systems
1193 <code>waitpid</code> and <code>sys/wait.h</code> documentation for details).</p>
1196 <p>Example: Try to exit cleanly on SIGINT and SIGTERM.</p>
1198 sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)
1200 ev_unloop (loop, EVUNLOOP_ALL);
1203 struct ev_signal signal_watcher;
1204 ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
1205 ev_signal_start (loop, &sigint_cb);
1213 <h2 id="code_ev_stat_code_did_the_file_attri"><code>ev_stat</code> - did the file attributes just change?</h2>
1214 <div id="code_ev_stat_code_did_the_file_attri-2">
1215 <p>This watches a filesystem path for attribute changes. That is, it calls
1216 <code>stat</code> regularly (or when the OS says it changed) and sees if it changed
1217 compared to the last time, invoking the callback if it did.</p>
1218 <p>The path does not need to exist: changing from "path exists" to "path does
1219 not exist" is a status change like any other. The condition "path does
1220 not exist" is signified by the <code>st_nlink</code> field being zero (which is
1221 otherwise always forced to be at least one) and all the other fields of
1222 the stat buffer having unspecified contents.</p>
1223 <p>Since there is no standard to do this, the portable implementation simply
1224 calls <code>stat (2)</code> regulalry on the path to see if it changed somehow. You
1225 can specify a recommended polling interval for this case. If you specify
1226 a polling interval of <code>0</code> (highly recommended!) then a <i>suitable,
1227 unspecified default</i> value will be used (which you can expect to be around
1228 five seconds, although this might change dynamically). Libev will also
1229 impose a minimum interval which is currently around <code>0.1</code>, but thats
1230 usually overkill.</p>
1231 <p>This watcher type is not meant for massive numbers of stat watchers,
1232 as even with OS-supported change notifications, this can be
1233 resource-intensive.</p>
1234 <p>At the time of this writing, no specific OS backends are implemented, but
1235 if demand increases, at least a kqueue and inotify backend will be added.</p>
1237 <dt>ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)</dt>
1238 <dt>ev_stat_set (ev_stat *, const char *path, ev_tstamp interval)</dt>
1240 <p>Configures the watcher to wait for status changes of the given
1241 <code>path</code>. The <code>interval</code> is a hint on how quickly a change is expected to
1242 be detected and should normally be specified as <code>0</code> to let libev choose
1243 a suitable value. The memory pointed to by <code>path</code> must point to the same
1244 path for as long as the watcher is active.</p>
1245 <p>The callback will be receive <code>EV_STAT</code> when a change was detected,
1246 relative to the attributes at the time the watcher was started (or the
1247 last change was detected).</p>
1249 <dt>ev_stat_stat (ev_stat *)</dt>
1251 <p>Updates the stat buffer immediately with new values. If you change the
1252 watched path in your callback, you could call this fucntion to avoid
1253 detecting this change (while introducing a race condition). Can also be
1254 useful simply to find out the new values.</p>
1256 <dt>ev_statdata attr [read-only]</dt>
1258 <p>The most-recently detected attributes of the file. Although the type is of
1259 <code>ev_statdata</code>, this is usually the (or one of the) <code>struct stat</code> types
1260 suitable for your system. If the <code>st_nlink</code> member is <code>0</code>, then there
1261 was some error while <code>stat</code>ing the file.</p>
1263 <dt>ev_statdata prev [read-only]</dt>
1265 <p>The previous attributes of the file. The callback gets invoked whenever
1266 <code>prev</code> != <code>attr</code>.</p>
1268 <dt>ev_tstamp interval [read-only]</dt>
1270 <p>The specified interval.</p>
1272 <dt>const char *path [read-only]</dt>
1274 <p>The filesystem path that is being watched.</p>
1277 <p>Example: Watch <code>/etc/passwd</code> for attribute changes.</p>
1279 passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)
1281 /* /etc/passwd changed in some way */
1282 if (w->attr.st_nlink)
1284 printf ("passwd current size %ld\n", (long)w->attr.st_size);
1285 printf ("passwd current atime %ld\n", (long)w->attr.st_mtime);
1286 printf ("passwd current mtime %ld\n", (long)w->attr.st_mtime);
1289 /* you shalt not abuse printf for puts */
1290 puts ("wow, /etc/passwd is not there, expect problems. "
1291 "if this is windows, they already arrived\n");
1297 ev_stat_init (&passwd, passwd_cb, "/etc/passwd");
1298 ev_stat_start (loop, &passwd);
1306 <h2 id="code_ev_idle_code_when_you_ve_got_no"><code>ev_idle</code> - when you've got nothing better to do...</h2>
1307 <div id="code_ev_idle_code_when_you_ve_got_no-2">
1308 <p>Idle watchers trigger events when there are no other events are pending
1309 (prepare, check and other idle watchers do not count). That is, as long
1310 as your process is busy handling sockets or timeouts (or even signals,
1311 imagine) it will not be triggered. But when your process is idle all idle
1312 watchers are being called again and again, once per event loop iteration -
1313 until stopped, that is, or your process receives more events and becomes
1315 <p>The most noteworthy effect is that as long as any idle watchers are
1316 active, the process will not block when waiting for new events.</p>
1317 <p>Apart from keeping your process non-blocking (which is a useful
1318 effect on its own sometimes), idle watchers are a good place to do
1319 "pseudo-background processing", or delay processing stuff to after the
1320 event loop has handled all outstanding events.</p>
1322 <dt>ev_idle_init (ev_signal *, callback)</dt>
1324 <p>Initialises and configures the idle watcher - it has no parameters of any
1325 kind. There is a <code>ev_idle_set</code> macro, but using it is utterly pointless,
1329 <p>Example: Dynamically allocate an <code>ev_idle</code> watcher, start it, and in the
1330 callback, free it. Also, use no error checking, as usual.</p>
1332 idle_cb (struct ev_loop *loop, struct ev_idle *w, int revents)
1335 // now do something you wanted to do when the program has
1336 // no longer asnything immediate to do.
1339 struct ev_idle *idle_watcher = malloc (sizeof (struct ev_idle));
1340 ev_idle_init (idle_watcher, idle_cb);
1341 ev_idle_start (loop, idle_cb);
1349 <h2 id="code_ev_prepare_code_and_code_ev_che"><code>ev_prepare</code> and <code>ev_check</code> - customise your event loop!</h2>
1350 <div id="code_ev_prepare_code_and_code_ev_che-2">
1351 <p>Prepare and check watchers are usually (but not always) used in tandem:
1352 prepare watchers get invoked before the process blocks and check watchers
1354 <p>You <i>must not</i> call <code>ev_loop</code> or similar functions that enter
1355 the current event loop from either <code>ev_prepare</code> or <code>ev_check</code>
1356 watchers. Other loops than the current one are fine, however. The
1357 rationale behind this is that you do not need to check for recursion in
1358 those watchers, i.e. the sequence will always be <code>ev_prepare</code>, blocking,
1359 <code>ev_check</code> so if you have one watcher of each kind they will always be
1360 called in pairs bracketing the blocking call.</p>
1361 <p>Their main purpose is to integrate other event mechanisms into libev and
1362 their use is somewhat advanced. This could be used, for example, to track
1363 variable changes, implement your own watchers, integrate net-snmp or a
1364 coroutine library and lots more. They are also occasionally useful if
1365 you cache some data and want to flush it before blocking (for example,
1366 in X programs you might want to do an <code>XFlush ()</code> in an <code>ev_prepare</code>
1368 <p>This is done by examining in each prepare call which file descriptors need
1369 to be watched by the other library, registering <code>ev_io</code> watchers for
1370 them and starting an <code>ev_timer</code> watcher for any timeouts (many libraries
1371 provide just this functionality). Then, in the check watcher you check for
1372 any events that occured (by checking the pending status of all watchers
1373 and stopping them) and call back into the library. The I/O and timer
1374 callbacks will never actually be called (but must be valid nevertheless,
1375 because you never know, you know?).</p>
1376 <p>As another example, the Perl Coro module uses these hooks to integrate
1377 coroutines into libev programs, by yielding to other active coroutines
1378 during each prepare and only letting the process block if no coroutines
1379 are ready to run (it's actually more complicated: it only runs coroutines
1380 with priority higher than or equal to the event loop and one coroutine
1381 of lower priority, but only once, using idle watchers to keep the event
1382 loop from blocking if lower-priority coroutines are active, thus mapping
1383 low-priority coroutines to idle/background tasks).</p>
1385 <dt>ev_prepare_init (ev_prepare *, callback)</dt>
1386 <dt>ev_check_init (ev_check *, callback)</dt>
1388 <p>Initialises and configures the prepare or check watcher - they have no
1389 parameters of any kind. There are <code>ev_prepare_set</code> and <code>ev_check_set</code>
1390 macros, but using them is utterly, utterly and completely pointless.</p>
1393 <p>Example: To include a library such as adns, you would add IO watchers
1394 and a timeout watcher in a prepare handler, as required by libadns, and
1395 in a check watcher, destroy them and call into libadns. What follows is
1396 pseudo-code only of course:</p>
1397 <pre> static ev_io iow [nfd];
1401 io_cb (ev_loop *loop, ev_io *w, int revents)
1403 // set the relevant poll flags
1404 // could also call adns_processreadable etc. here
1405 struct pollfd *fd = (struct pollfd *)w->data;
1406 if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;
1407 if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;
1410 // create io watchers for each fd and a timer before blocking
1412 adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents)
1414 int timeout = 3600000;truct pollfd fds [nfd];
1415 // actual code will need to loop here and realloc etc.
1416 adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));
1418 /* the callback is illegal, but won't be called as we stop during check */
1419 ev_timer_init (&tw, 0, timeout * 1e-3);
1420 ev_timer_start (loop, &tw);
1422 // create on ev_io per pollfd
1423 for (int i = 0; i < nfd; ++i)
1425 ev_io_init (iow + i, io_cb, fds [i].fd,
1426 ((fds [i].events & POLLIN ? EV_READ : 0)
1427 | (fds [i].events & POLLOUT ? EV_WRITE : 0)));
1429 fds [i].revents = 0;
1430 iow [i].data = fds + i;
1431 ev_io_start (loop, iow + i);
1435 // stop all watchers after blocking
1437 adns_check_cb (ev_loop *loop, ev_check *w, int revents)
1439 ev_timer_stop (loop, &tw);
1441 for (int i = 0; i < nfd; ++i)
1442 ev_io_stop (loop, iow + i);
1444 adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));
1453 <h2 id="code_ev_embed_code_when_one_backend_"><code>ev_embed</code> - when one backend isn't enough...</h2>
1454 <div id="code_ev_embed_code_when_one_backend_-2">
1455 <p>This is a rather advanced watcher type that lets you embed one event loop
1456 into another (currently only <code>ev_io</code> events are supported in the embedded
1457 loop, other types of watchers might be handled in a delayed or incorrect
1458 fashion and must not be used).</p>
1459 <p>There are primarily two reasons you would want that: work around bugs and
1461 <p>As an example for a bug workaround, the kqueue backend might only support
1462 sockets on some platform, so it is unusable as generic backend, but you
1463 still want to make use of it because you have many sockets and it scales
1464 so nicely. In this case, you would create a kqueue-based loop and embed it
1465 into your default loop (which might use e.g. poll). Overall operation will
1466 be a bit slower because first libev has to poll and then call kevent, but
1467 at least you can use both at what they are best.</p>
1468 <p>As for prioritising I/O: rarely you have the case where some fds have
1469 to be watched and handled very quickly (with low latency), and even
1470 priorities and idle watchers might have too much overhead. In this case
1471 you would put all the high priority stuff in one loop and all the rest in
1472 a second one, and embed the second one in the first.</p>
1473 <p>As long as the watcher is active, the callback will be invoked every time
1474 there might be events pending in the embedded loop. The callback must then
1475 call <code>ev_embed_sweep (mainloop, watcher)</code> to make a single sweep and invoke
1476 their callbacks (you could also start an idle watcher to give the embedded
1477 loop strictly lower priority for example). You can also set the callback
1478 to <code>0</code>, in which case the embed watcher will automatically execute the
1479 embedded loop sweep.</p>
1480 <p>As long as the watcher is started it will automatically handle events. The
1481 callback will be invoked whenever some events have been handled. You can
1482 set the callback to <code>0</code> to avoid having to specify one if you are not
1483 interested in that.</p>
1484 <p>Also, there have not currently been made special provisions for forking:
1485 when you fork, you not only have to call <code>ev_loop_fork</code> on both loops,
1486 but you will also have to stop and restart any <code>ev_embed</code> watchers
1488 <p>Unfortunately, not all backends are embeddable, only the ones returned by
1489 <code>ev_embeddable_backends</code> are, which, unfortunately, does not include any
1491 <p>So when you want to use this feature you will always have to be prepared
1492 that you cannot get an embeddable loop. The recommended way to get around
1493 this is to have a separate variables for your embeddable loop, try to
1494 create it, and if that fails, use the normal loop for everything:</p>
1495 <pre> struct ev_loop *loop_hi = ev_default_init (0);
1496 struct ev_loop *loop_lo = 0;
1497 struct ev_embed embed;
1499 // see if there is a chance of getting one that works
1500 // (remember that a flags value of 0 means autodetection)
1501 loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
1502 ? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
1505 // if we got one, then embed it, otherwise default to loop_hi
1508 ev_embed_init (&embed, 0, loop_lo);
1509 ev_embed_start (loop_hi, &embed);
1516 <dt>ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)</dt>
1517 <dt>ev_embed_set (ev_embed *, callback, struct ev_loop *embedded_loop)</dt>
1519 <p>Configures the watcher to embed the given loop, which must be
1520 embeddable. If the callback is <code>0</code>, then <code>ev_embed_sweep</code> will be
1521 invoked automatically, otherwise it is the responsibility of the callback
1522 to invoke it (it will continue to be called until the sweep has been done,
1523 if you do not want thta, you need to temporarily stop the embed watcher).</p>
1525 <dt>ev_embed_sweep (loop, ev_embed *)</dt>
1527 <p>Make a single, non-blocking sweep over the embedded loop. This works
1528 similarly to <code>ev_loop (embedded_loop, EVLOOP_NONBLOCK)</code>, but in the most
1529 apropriate way for embedded loops.</p>
1531 <dt>struct ev_loop *loop [read-only]</dt>
1533 <p>The embedded event loop.</p>
1542 <h2 id="code_ev_fork_code_the_audacity_to_re"><code>ev_fork</code> - the audacity to resume the event loop after a fork</h2>
1543 <div id="code_ev_fork_code_the_audacity_to_re-2">
1544 <p>Fork watchers are called when a <code>fork ()</code> was detected (usually because
1545 whoever is a good citizen cared to tell libev about it by calling
1546 <code>ev_default_fork</code> or <code>ev_loop_fork</code>). The invocation is done before the
1547 event loop blocks next and before <code>ev_check</code> watchers are being called,
1548 and only in the child after the fork. If whoever good citizen calling
1549 <code>ev_default_fork</code> cheats and calls it in the wrong process, the fork
1550 handlers will be invoked, too, of course.</p>
1552 <dt>ev_fork_init (ev_signal *, callback)</dt>
1554 <p>Initialises and configures the fork watcher - it has no parameters of any
1555 kind. There is a <code>ev_fork_set</code> macro, but using it is utterly pointless,
1565 <h1 id="OTHER_FUNCTIONS">OTHER FUNCTIONS</h1>
1566 <div id="OTHER_FUNCTIONS_CONTENT">
1567 <p>There are some other functions of possible interest. Described. Here. Now.</p>
1569 <dt>ev_once (loop, int fd, int events, ev_tstamp timeout, callback)</dt>
1571 <p>This function combines a simple timer and an I/O watcher, calls your
1572 callback on whichever event happens first and automatically stop both
1573 watchers. This is useful if you want to wait for a single event on an fd
1574 or timeout without having to allocate/configure/start/stop/free one or
1575 more watchers yourself.</p>
1576 <p>If <code>fd</code> is less than 0, then no I/O watcher will be started and events
1577 is being ignored. Otherwise, an <code>ev_io</code> watcher for the given <code>fd</code> and
1578 <code>events</code> set will be craeted and started.</p>
1579 <p>If <code>timeout</code> is less than 0, then no timeout watcher will be
1580 started. Otherwise an <code>ev_timer</code> watcher with after = <code>timeout</code> (and
1581 repeat = 0) will be started. While <code>0</code> is a valid timeout, it is of
1583 <p>The callback has the type <code>void (*cb)(int revents, void *arg)</code> and gets
1584 passed an <code>revents</code> set like normal event callbacks (a combination of
1585 <code>EV_ERROR</code>, <code>EV_READ</code>, <code>EV_WRITE</code> or <code>EV_TIMEOUT</code>) and the <code>arg</code>
1586 value passed to <code>ev_once</code>:</p>
1587 <pre> static void stdin_ready (int revents, void *arg)
1589 if (revents & EV_TIMEOUT)
1590 /* doh, nothing entered */;
1591 else if (revents & EV_READ)
1592 /* stdin might have data for us, joy! */;
1595 ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
1599 <dt>ev_feed_event (ev_loop *, watcher *, int revents)</dt>
1601 <p>Feeds the given event set into the event loop, as if the specified event
1602 had happened for the specified watcher (which must be a pointer to an
1603 initialised but not necessarily started event watcher).</p>
1605 <dt>ev_feed_fd_event (ev_loop *, int fd, int revents)</dt>
1607 <p>Feed an event on the given fd, as if a file descriptor backend detected
1608 the given events it.</p>
1610 <dt>ev_feed_signal_event (ev_loop *loop, int signum)</dt>
1612 <p>Feed an event as if the given signal occured (<code>loop</code> must be the default
1622 <h1 id="LIBEVENT_EMULATION">LIBEVENT EMULATION</h1>
1623 <div id="LIBEVENT_EMULATION_CONTENT">
1624 <p>Libev offers a compatibility emulation layer for libevent. It cannot
1625 emulate the internals of libevent, so here are some usage hints:</p>
1627 <dt>* Use it by including <event.h>, as usual.</dt>
1628 <dt>* The following members are fully supported: ev_base, ev_callback,
1629 ev_arg, ev_fd, ev_res, ev_events.</dt>
1630 <dt>* Avoid using ev_flags and the EVLIST_*-macros, while it is
1631 maintained by libev, it does not work exactly the same way as in libevent (consider
1632 it a private API).</dt>
1633 <dt>* Priorities are not currently supported. Initialising priorities
1634 will fail and all watchers will have the same priority, even though there
1635 is an ev_pri field.</dt>
1636 <dt>* Other members are not supported.</dt>
1637 <dt>* The libev emulation is <i>not</i> ABI compatible to libevent, you need
1638 to use the libev header file and library.</dt>
1642 <h1 id="C_SUPPORT">C++ SUPPORT</h1>
1643 <div id="C_SUPPORT_CONTENT">
1644 <p>Libev comes with some simplistic wrapper classes for C++ that mainly allow
1645 you to use some convinience methods to start/stop watchers and also change
1646 the callback model to a model using method callbacks on objects.</p>
1648 <pre> #include <ev++.h>
1651 <p>(it is not installed by default). This automatically includes <cite>ev.h</cite>
1652 and puts all of its definitions (many of them macros) into the global
1653 namespace. All C++ specific things are put into the <code>ev</code> namespace.</p>
1654 <p>It should support all the same embedding options as <cite>ev.h</cite>, most notably
1655 <code>EV_MULTIPLICITY</code>.</p>
1656 <p>Here is a list of things available in the <code>ev</code> namespace:</p>
1658 <dt><code>ev::READ</code>, <code>ev::WRITE</code> etc.</dt>
1660 <p>These are just enum values with the same values as the <code>EV_READ</code> etc.
1661 macros from <cite>ev.h</cite>.</p>
1663 <dt><code>ev::tstamp</code>, <code>ev::now</code></dt>
1665 <p>Aliases to the same types/functions as with the <code>ev_</code> prefix.</p>
1667 <dt><code>ev::io</code>, <code>ev::timer</code>, <code>ev::periodic</code>, <code>ev::idle</code>, <code>ev::sig</code> etc.</dt>
1669 <p>For each <code>ev_TYPE</code> watcher in <cite>ev.h</cite> there is a corresponding class of
1670 the same name in the <code>ev</code> namespace, with the exception of <code>ev_signal</code>
1671 which is called <code>ev::sig</code> to avoid clashes with the <code>signal</code> macro
1672 defines by many implementations.</p>
1673 <p>All of those classes have these methods:</p>
1676 <dt>ev::TYPE::TYPE (object *, object::method *)</dt>
1677 <dt>ev::TYPE::TYPE (object *, object::method *, struct ev_loop *)</dt>
1678 <dt>ev::TYPE::~TYPE</dt>
1680 <p>The constructor takes a pointer to an object and a method pointer to
1681 the event handler callback to call in this class. The constructor calls
1682 <code>ev_init</code> for you, which means you have to call the <code>set</code> method
1683 before starting it. If you do not specify a loop then the constructor
1684 automatically associates the default loop with this watcher.</p>
1685 <p>The destructor automatically stops the watcher if it is active.</p>
1687 <dt>w->set (struct ev_loop *)</dt>
1689 <p>Associates a different <code>struct ev_loop</code> with this watcher. You can only
1690 do this when the watcher is inactive (and not pending either).</p>
1692 <dt>w->set ([args])</dt>
1694 <p>Basically the same as <code>ev_TYPE_set</code>, with the same args. Must be
1695 called at least once. Unlike the C counterpart, an active watcher gets
1696 automatically stopped and restarted.</p>
1698 <dt>w->start ()</dt>
1700 <p>Starts the watcher. Note that there is no <code>loop</code> argument as the
1701 constructor already takes the loop.</p>
1703 <dt>w->stop ()</dt>
1705 <p>Stops the watcher if it is active. Again, no <code>loop</code> argument.</p>
1707 <dt>w->again () <code>ev::timer</code>, <code>ev::periodic</code> only</dt>
1709 <p>For <code>ev::timer</code> and <code>ev::periodic</code>, this invokes the corresponding
1710 <code>ev_TYPE_again</code> function.</p>
1712 <dt>w->sweep () <code>ev::embed</code> only</dt>
1714 <p>Invokes <code>ev_embed_sweep</code>.</p>
1716 <dt>w->update () <code>ev::stat</code> only</dt>
1718 <p>Invokes <code>ev_stat_stat</code>.</p>
1724 <p>Example: Define a class with an IO and idle watcher, start one of them in
1725 the constructor.</p>
1728 ev_io io; void io_cb (ev::io &w, int revents);
1729 ev_idle idle void idle_cb (ev::idle &w, int revents);
1734 myclass::myclass (int fd)
1735 : io (this, &myclass::io_cb),
1736 idle (this, &myclass::idle_cb)
1738 io.start (fd, ev::READ);
1747 <h1 id="MACRO_MAGIC">MACRO MAGIC</h1>
1748 <div id="MACRO_MAGIC_CONTENT">
1749 <p>Libev can be compiled with a variety of options, the most fundemantal is
1750 <code>EV_MULTIPLICITY</code>. This option determines wether (most) functions and
1751 callbacks have an initial <code>struct ev_loop *</code> argument.</p>
1752 <p>To make it easier to write programs that cope with either variant, the
1753 following macros are defined:</p>
1755 <dt><code>EV_A</code>, <code>EV_A_</code></dt>
1757 <p>This provides the loop <i>argument</i> for functions, if one is required ("ev
1758 loop argument"). The <code>EV_A</code> form is used when this is the sole argument,
1759 <code>EV_A_</code> is used when other arguments are following. Example:</p>
1760 <pre> ev_unref (EV_A);
1761 ev_timer_add (EV_A_ watcher);
1765 <p>It assumes the variable <code>loop</code> of type <code>struct ev_loop *</code> is in scope,
1766 which is often provided by the following macro.</p>
1768 <dt><code>EV_P</code>, <code>EV_P_</code></dt>
1770 <p>This provides the loop <i>parameter</i> for functions, if one is required ("ev
1771 loop parameter"). The <code>EV_P</code> form is used when this is the sole parameter,
1772 <code>EV_P_</code> is used when other parameters are following. Example:</p>
1773 <pre> // this is how ev_unref is being declared
1774 static void ev_unref (EV_P);
1776 // this is how you can declare your typical callback
1777 static void cb (EV_P_ ev_timer *w, int revents)
1780 <p>It declares a parameter <code>loop</code> of type <code>struct ev_loop *</code>, quite
1781 suitable for use with <code>EV_A</code>.</p>
1783 <dt><code>EV_DEFAULT</code>, <code>EV_DEFAULT_</code></dt>
1785 <p>Similar to the other two macros, this gives you the value of the default
1786 loop, if multiple loops are supported ("ev loop default").</p>
1789 <p>Example: Declare and initialise a check watcher, working regardless of
1790 wether multiple loops are supported or not.</p>
1792 check_cb (EV_P_ ev_timer *w, int revents)
1794 ev_check_stop (EV_A_ w);
1798 ev_check_init (&check, check_cb);
1799 ev_check_start (EV_DEFAULT_ &check);
1800 ev_loop (EV_DEFAULT_ 0);
1808 <h1 id="EMBEDDING">EMBEDDING</h1>
1809 <div id="EMBEDDING_CONTENT">
1810 <p>Libev can (and often is) directly embedded into host
1811 applications. Examples of applications that embed it include the Deliantra
1812 Game Server, the EV perl module, the GNU Virtual Private Ethernet (gvpe)
1813 and rxvt-unicode.</p>
1814 <p>The goal is to enable you to just copy the neecssary files into your
1815 source directory without having to change even a single line in them, so
1816 you can easily upgrade by simply copying (or having a checked-out copy of
1817 libev somewhere in your source tree).</p>
1820 <h2 id="FILESETS">FILESETS</h2>
1821 <div id="FILESETS_CONTENT">
1822 <p>Depending on what features you need you need to include one or more sets of files
1826 <h3 id="CORE_EVENT_LOOP">CORE EVENT LOOP</h3>
1827 <div id="CORE_EVENT_LOOP_CONTENT">
1828 <p>To include only the libev core (all the <code>ev_*</code> functions), with manual
1829 configuration (no autoconf):</p>
1830 <pre> #define EV_STANDALONE 1
1831 #include "ev.c"
1834 <p>This will automatically include <cite>ev.h</cite>, too, and should be done in a
1835 single C source file only to provide the function implementations. To use
1836 it, do the same for <cite>ev.h</cite> in all files wishing to use this API (best
1837 done by writing a wrapper around <cite>ev.h</cite> that you can include instead and
1838 where you can put other configuration options):</p>
1839 <pre> #define EV_STANDALONE 1
1840 #include "ev.h"
1843 <p>Both header files and implementation files can be compiled with a C++
1844 compiler (at least, thats a stated goal, and breakage will be treated
1846 <p>You need the following files in your source tree, or in a directory
1847 in your include path (e.g. in libev/ when using -Ilibev):</p>
1853 ev_win32.c required on win32 platforms only
1855 ev_select.c only when select backend is enabled (which is by default)
1856 ev_poll.c only when poll backend is enabled (disabled by default)
1857 ev_epoll.c only when the epoll backend is enabled (disabled by default)
1858 ev_kqueue.c only when the kqueue backend is enabled (disabled by default)
1859 ev_port.c only when the solaris port backend is enabled (disabled by default)
1862 <p><cite>ev.c</cite> includes the backend files directly when enabled, so you only need
1863 to compile this single file.</p>
1866 <h3 id="LIBEVENT_COMPATIBILITY_API">LIBEVENT COMPATIBILITY API</h3>
1867 <div id="LIBEVENT_COMPATIBILITY_API_CONTENT">
1868 <p>To include the libevent compatibility API, also include:</p>
1869 <pre> #include "event.c"
1872 <p>in the file including <cite>ev.c</cite>, and:</p>
1873 <pre> #include "event.h"
1876 <p>in the files that want to use the libevent API. This also includes <cite>ev.h</cite>.</p>
1877 <p>You need the following additional files for this:</p>
1884 <h3 id="AUTOCONF_SUPPORT">AUTOCONF SUPPORT</h3>
1885 <div id="AUTOCONF_SUPPORT_CONTENT">
1886 <p>Instead of using <code>EV_STANDALONE=1</code> and providing your config in
1887 whatever way you want, you can also <code>m4_include([libev.m4])</code> in your
1888 <cite>configure.ac</cite> and leave <code>EV_STANDALONE</code> undefined. <cite>ev.c</cite> will then
1889 include <cite>config.h</cite> and configure itself accordingly.</p>
1890 <p>For this of course you need the m4 file:</p>
1896 <h2 id="PREPROCESSOR_SYMBOLS_MACROS">PREPROCESSOR SYMBOLS/MACROS</h2>
1897 <div id="PREPROCESSOR_SYMBOLS_MACROS_CONTENT">
1898 <p>Libev can be configured via a variety of preprocessor symbols you have to define
1899 before including any of its files. The default is not to build for multiplicity
1900 and only include the select backend.</p>
1902 <dt>EV_STANDALONE</dt>
1904 <p>Must always be <code>1</code> if you do not use autoconf configuration, which
1905 keeps libev from including <cite>config.h</cite>, and it also defines dummy
1906 implementations for some libevent functions (such as logging, which is not
1907 supported). It will also not define any of the structs usually found in
1908 <cite>event.h</cite> that are not directly supported by the libev core alone.</p>
1910 <dt>EV_USE_MONOTONIC</dt>
1912 <p>If defined to be <code>1</code>, libev will try to detect the availability of the
1913 monotonic clock option at both compiletime and runtime. Otherwise no use
1914 of the monotonic clock option will be attempted. If you enable this, you
1915 usually have to link against librt or something similar. Enabling it when
1916 the functionality isn't available is safe, though, althoguh you have
1917 to make sure you link against any libraries where the <code>clock_gettime</code>
1918 function is hiding in (often <cite>-lrt</cite>).</p>
1920 <dt>EV_USE_REALTIME</dt>
1922 <p>If defined to be <code>1</code>, libev will try to detect the availability of the
1923 realtime clock option at compiletime (and assume its availability at
1924 runtime if successful). Otherwise no use of the realtime clock option will
1925 be attempted. This effectively replaces <code>gettimeofday</code> by <code>clock_get
1926 (CLOCK_REALTIME, ...)</code> and will not normally affect correctness. See tzhe note about libraries
1927 in the description of <code>EV_USE_MONOTONIC</code>, though.</p>
1929 <dt>EV_USE_SELECT</dt>
1931 <p>If undefined or defined to be <code>1</code>, libev will compile in support for the
1932 <code>select</code>(2) backend. No attempt at autodetection will be done: if no
1933 other method takes over, select will be it. Otherwise the select backend
1934 will not be compiled in.</p>
1936 <dt>EV_SELECT_USE_FD_SET</dt>
1938 <p>If defined to <code>1</code>, then the select backend will use the system <code>fd_set</code>
1939 structure. This is useful if libev doesn't compile due to a missing
1940 <code>NFDBITS</code> or <code>fd_mask</code> definition or it misguesses the bitset layout on
1941 exotic systems. This usually limits the range of file descriptors to some
1942 low limit such as 1024 or might have other limitations (winsocket only
1943 allows 64 sockets). The <code>FD_SETSIZE</code> macro, set before compilation, might
1944 influence the size of the <code>fd_set</code> used.</p>
1946 <dt>EV_SELECT_IS_WINSOCKET</dt>
1948 <p>When defined to <code>1</code>, the select backend will assume that
1949 select/socket/connect etc. don't understand file descriptors but
1950 wants osf handles on win32 (this is the case when the select to
1951 be used is the winsock select). This means that it will call
1952 <code>_get_osfhandle</code> on the fd to convert it to an OS handle. Otherwise,
1953 it is assumed that all these functions actually work on fds, even
1954 on win32. Should not be defined on non-win32 platforms.</p>
1956 <dt>EV_USE_POLL</dt>
1958 <p>If defined to be <code>1</code>, libev will compile in support for the <code>poll</code>(2)
1959 backend. Otherwise it will be enabled on non-win32 platforms. It
1960 takes precedence over select.</p>
1962 <dt>EV_USE_EPOLL</dt>
1964 <p>If defined to be <code>1</code>, libev will compile in support for the Linux
1965 <code>epoll</code>(7) backend. Its availability will be detected at runtime,
1966 otherwise another method will be used as fallback. This is the
1967 preferred backend for GNU/Linux systems.</p>
1969 <dt>EV_USE_KQUEUE</dt>
1971 <p>If defined to be <code>1</code>, libev will compile in support for the BSD style
1972 <code>kqueue</code>(2) backend. Its actual availability will be detected at runtime,
1973 otherwise another method will be used as fallback. This is the preferred
1974 backend for BSD and BSD-like systems, although on most BSDs kqueue only
1975 supports some types of fds correctly (the only platform we found that
1976 supports ptys for example was NetBSD), so kqueue might be compiled in, but
1977 not be used unless explicitly requested. The best way to use it is to find
1978 out whether kqueue supports your type of fd properly and use an embedded
1981 <dt>EV_USE_PORT</dt>
1983 <p>If defined to be <code>1</code>, libev will compile in support for the Solaris
1984 10 port style backend. Its availability will be detected at runtime,
1985 otherwise another method will be used as fallback. This is the preferred
1986 backend for Solaris 10 systems.</p>
1988 <dt>EV_USE_DEVPOLL</dt>
1990 <p>reserved for future expansion, works like the USE symbols above.</p>
1994 <p>The name of the <cite>ev.h</cite> header file used to include it. The default if
1995 undefined is <code><ev.h></code> in <cite>event.h</cite> and <code>"ev.h"</code> in <cite>ev.c</cite>. This
1996 can be used to virtually rename the <cite>ev.h</cite> header file in case of conflicts.</p>
1998 <dt>EV_CONFIG_H</dt>
2000 <p>If <code>EV_STANDALONE</code> isn't <code>1</code>, this variable can be used to override
2001 <cite>ev.c</cite>'s idea of where to find the <cite>config.h</cite> file, similarly to
2002 <code>EV_H</code>, above.</p>
2006 <p>Similarly to <code>EV_H</code>, this macro can be used to override <cite>event.c</cite>'s idea
2007 of how the <cite>event.h</cite> header can be found.</p>
2009 <dt>EV_PROTOTYPES</dt>
2011 <p>If defined to be <code>0</code>, then <cite>ev.h</cite> will not define any function
2012 prototypes, but still define all the structs and other symbols. This is
2013 occasionally useful if you want to provide your own wrapper functions
2014 around libev functions.</p>
2016 <dt>EV_MULTIPLICITY</dt>
2018 <p>If undefined or defined to <code>1</code>, then all event-loop-specific functions
2019 will have the <code>struct ev_loop *</code> as first argument, and you can create
2020 additional independent event loops. Otherwise there will be no support
2021 for multiple event loops and there is no first event loop pointer
2022 argument. Instead, all functions act on the single default loop.</p>
2024 <dt>EV_PERIODIC_ENABLE</dt>
2026 <p>If undefined or defined to be <code>1</code>, then periodic timers are supported. If
2027 defined to be <code>0</code>, then they are not. Disabling them saves a few kB of
2030 <dt>EV_EMBED_ENABLE</dt>
2032 <p>If undefined or defined to be <code>1</code>, then embed watchers are supported. If
2033 defined to be <code>0</code>, then they are not.</p>
2035 <dt>EV_STAT_ENABLE</dt>
2037 <p>If undefined or defined to be <code>1</code>, then stat watchers are supported. If
2038 defined to be <code>0</code>, then they are not.</p>
2040 <dt>EV_FORK_ENABLE</dt>
2042 <p>If undefined or defined to be <code>1</code>, then fork watchers are supported. If
2043 defined to be <code>0</code>, then they are not.</p>
2047 <p>If you need to shave off some kilobytes of code at the expense of some
2048 speed, define this symbol to <code>1</code>. Currently only used for gcc to override
2049 some inlining decisions, saves roughly 30% codesize of amd64.</p>
2051 <dt>EV_PID_HASHSIZE</dt>
2053 <p><code>ev_child</code> watchers use a small hash table to distribute workload by
2054 pid. The default size is <code>16</code> (or <code>1</code> with <code>EV_MINIMAL</code>), usually more
2055 than enough. If you need to manage thousands of children you might want to
2056 increase this value.</p>
2060 <p>By default, all watchers have a <code>void *data</code> member. By redefining
2061 this macro to a something else you can include more and other types of
2062 members. You have to define it each time you include one of the files,
2063 though, and it must be identical each time.</p>
2064 <p>For example, the perl EV module uses something like this:</p>
2065 <pre> #define EV_COMMON \
2066 SV *self; /* contains this struct */ \
2067 SV *cb_sv, *fh /* note no trailing ";" */
2071 <dt>EV_CB_DECLARE (type)</dt>
2072 <dt>EV_CB_INVOKE (watcher, revents)</dt>
2073 <dt>ev_set_cb (ev, cb)</dt>
2075 <p>Can be used to change the callback member declaration in each watcher,
2076 and the way callbacks are invoked and set. Must expand to a struct member
2077 definition and a statement, respectively. See the <cite>ev.v</cite> header file for
2078 their default definitions. One possible use for overriding these is to
2079 avoid the <code>struct ev_loop *</code> as first argument in all cases, or to use
2080 method calls instead of plain function calls in C++.</p>
2083 <h2 id="EXAMPLES">EXAMPLES</h2>
2084 <div id="EXAMPLES_CONTENT">
2085 <p>For a real-world example of a program the includes libev
2086 verbatim, you can have a look at the EV perl module
2087 (<a href="http://software.schmorp.de/pkg/EV.html">http://software.schmorp.de/pkg/EV.html</a>). It has the libev files in
2088 the <cite>libev/</cite> subdirectory and includes them in the <cite>EV/EVAPI.h</cite> (public
2089 interface) and <cite>EV.xs</cite> (implementation) files. Only the <cite>EV.xs</cite> file
2090 will be compiled. It is pretty complex because it provides its own header
2092 <p>The usage in rxvt-unicode is simpler. It has a <cite>ev_cpp.h</cite> header file
2093 that everybody includes and which overrides some autoconf choices:</p>
2094 <pre> #define EV_USE_POLL 0
2095 #define EV_MULTIPLICITY 0
2096 #define EV_PERIODICS 0
2097 #define EV_CONFIG_H <config.h>
2099 #include "ev++.h"
2102 <p>And a <cite>ev_cpp.C</cite> implementation file that contains libev proper and is compiled:</p>
2103 <pre> #include "ev_cpp.h"
2104 #include "ev.c"
2112 <h1 id="COMPLEXITIES">COMPLEXITIES</h1>
2113 <div id="COMPLEXITIES_CONTENT">
2114 <p>In this section the complexities of (many of) the algorithms used inside
2115 libev will be explained. For complexity discussions about backends see the
2116 documentation for <code>ev_default_init</code>.</p>
2119 <dt>Starting and stopping timer/periodic watchers: O(log skipped_other_timers)</dt>
2120 <dt>Changing timer/periodic watchers (by autorepeat, again): O(log skipped_other_timers)</dt>
2121 <dt>Starting io/check/prepare/idle/signal/child watchers: O(1)</dt>
2122 <dt>Stopping check/prepare/idle watchers: O(1)</dt>
2123 <dt>Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % 16))</dt>
2124 <dt>Finding the next timer per loop iteration: O(1)</dt>
2125 <dt>Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)</dt>
2126 <dt>Activating one watcher: O(1)</dt>
2135 <h1 id="AUTHOR">AUTHOR</h1>
2136 <div id="AUTHOR_CONTENT">
2137 <p>Marc Lehmann <libev@schmorp.de>.</p>
projects / software / libev.git / blob