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="Wed Nov 28 12:31:29 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>), the Linux <code>inotify</code> interface
141 (for <code>ev_stat</code>), relative timers (<code>ev_timer</code>), absolute timers
142 with customised rescheduling (<code>ev_periodic</code>), synchronous signals
143 (<code>ev_signal</code>), process status change events (<code>ev_child</code>), and event
144 watchers dealing with the event loop mechanism itself (<code>ev_idle</code>,
145 <code>ev_embed</code>, <code>ev_prepare</code> and <code>ev_check</code> watchers) as well as
146 file watchers (<code>ev_stat</code>) and even limited support for fork events
147 (<code>ev_fork</code>).</p>
148 <p>It also is quite fast (see this
149 <a href="http://libev.schmorp.de/bench.html">benchmark</a> comparing it to libevent
153 <h1 id="CONVENTIONS">CONVENTIONS</h1>
154 <div id="CONVENTIONS_CONTENT">
155 <p>Libev is very configurable. In this manual the default configuration will
156 be described, which supports multiple event loops. For more info about
157 various configuration options please have a look at <strong>EMBED</strong> section in
158 this manual. If libev was configured without support for multiple event
159 loops, then all functions taking an initial argument of name <code>loop</code>
160 (which is always of type <code>struct ev_loop *</code>) will not have this argument.</p>
163 <h1 id="TIME_REPRESENTATION">TIME REPRESENTATION</h1>
164 <div id="TIME_REPRESENTATION_CONTENT">
165 <p>Libev represents time as a single floating point number, representing the
166 (fractional) number of seconds since the (POSIX) epoch (somewhere near
167 the beginning of 1970, details are complicated, don't ask). This type is
168 called <code>ev_tstamp</code>, which is what you should use too. It usually aliases
169 to the <code>double</code> type in C, and when you need to do any calculations on
170 it, you should treat it as such.</p>
173 <h1 id="GLOBAL_FUNCTIONS">GLOBAL FUNCTIONS</h1>
174 <div id="GLOBAL_FUNCTIONS_CONTENT">
175 <p>These functions can be called anytime, even before initialising the
176 library in any way.</p>
178 <dt>ev_tstamp ev_time ()</dt>
180 <p>Returns the current time as libev would use it. Please note that the
181 <code>ev_now</code> function is usually faster and also often returns the timestamp
182 you actually want to know.</p>
184 <dt>int ev_version_major ()</dt>
185 <dt>int ev_version_minor ()</dt>
187 <p>You can find out the major and minor version numbers of the library
188 you linked against by calling the functions <code>ev_version_major</code> and
189 <code>ev_version_minor</code>. If you want, you can compare against the global
190 symbols <code>EV_VERSION_MAJOR</code> and <code>EV_VERSION_MINOR</code>, which specify the
191 version of the library your program was compiled against.</p>
192 <p>Usually, it's a good idea to terminate if the major versions mismatch,
193 as this indicates an incompatible change. Minor versions are usually
194 compatible to older versions, so a larger minor version alone is usually
196 <p>Example: Make sure we haven't accidentally been linked against the wrong
198 <pre> assert (("libev version mismatch",
199 ev_version_major () == EV_VERSION_MAJOR
200 && ev_version_minor () >= EV_VERSION_MINOR));
204 <dt>unsigned int ev_supported_backends ()</dt>
206 <p>Return the set of all backends (i.e. their corresponding <code>EV_BACKEND_*</code>
207 value) compiled into this binary of libev (independent of their
208 availability on the system you are running on). See <code>ev_default_loop</code> for
209 a description of the set values.</p>
210 <p>Example: make sure we have the epoll method, because yeah this is cool and
211 a must have and can we have a torrent of it please!!!11</p>
212 <pre> assert (("sorry, no epoll, no sex",
213 ev_supported_backends () & EVBACKEND_EPOLL));
217 <dt>unsigned int ev_recommended_backends ()</dt>
219 <p>Return the set of all backends compiled into this binary of libev and also
220 recommended for this platform. This set is often smaller than the one
221 returned by <code>ev_supported_backends</code>, as for example kqueue is broken on
222 most BSDs and will not be autodetected unless you explicitly request it
223 (assuming you know what you are doing). This is the set of backends that
224 libev will probe for if you specify no backends explicitly.</p>
226 <dt>unsigned int ev_embeddable_backends ()</dt>
228 <p>Returns the set of backends that are embeddable in other event loops. This
229 is the theoretical, all-platform, value. To find which backends
230 might be supported on the current system, you would need to look at
231 <code>ev_embeddable_backends () & ev_supported_backends ()</code>, likewise for
232 recommended ones.</p>
233 <p>See the description of <code>ev_embed</code> watchers for more info.</p>
235 <dt>ev_set_allocator (void *(*cb)(void *ptr, size_t size))</dt>
237 <p>Sets the allocation function to use (the prototype and semantics are
238 identical to the realloc C function). It is used to allocate and free
239 memory (no surprises here). If it returns zero when memory needs to be
240 allocated, the library might abort or take some potentially destructive
241 action. The default is your system realloc function.</p>
242 <p>You could override this function in high-availability programs to, say,
243 free some memory if it cannot allocate memory, to use a special allocator,
244 or even to sleep a while and retry until some memory is available.</p>
245 <p>Example: Replace the libev allocator with one that waits a bit and then
248 persistent_realloc (void *ptr, size_t size)
252 void *newptr = realloc (ptr, size);
262 ev_set_allocator (persistent_realloc);
266 <dt>ev_set_syserr_cb (void (*cb)(const char *msg));</dt>
268 <p>Set the callback function to call on a retryable syscall error (such
269 as failed select, poll, epoll_wait). The message is a printable string
270 indicating the system call or subsystem causing the problem. If this
271 callback is set, then libev will expect it to remedy the sitution, no
272 matter what, when it returns. That is, libev will generally retry the
273 requested operation, or, if the condition doesn't go away, do bad stuff
275 <p>Example: This is basically the same thing that libev does internally, too.</p>
277 fatal_error (const char *msg)
284 ev_set_syserr_cb (fatal_error);
291 <h1 id="FUNCTIONS_CONTROLLING_THE_EVENT_LOOP">FUNCTIONS CONTROLLING THE EVENT LOOP</h1>
292 <div id="FUNCTIONS_CONTROLLING_THE_EVENT_LOOP-2">
293 <p>An event loop is described by a <code>struct ev_loop *</code>. The library knows two
294 types of such loops, the <i>default</i> loop, which supports signals and child
295 events, and dynamically created loops which do not.</p>
296 <p>If you use threads, a common model is to run the default event loop
297 in your main thread (or in a separate thread) and for each thread you
298 create, you also create another event loop. Libev itself does no locking
299 whatsoever, so if you mix calls to the same event loop in different
300 threads, make sure you lock (this is usually a bad idea, though, even if
301 done correctly, because it's hideous and inefficient).</p>
303 <dt>struct ev_loop *ev_default_loop (unsigned int flags)</dt>
305 <p>This will initialise the default event loop if it hasn't been initialised
306 yet and return it. If the default loop could not be initialised, returns
307 false. If it already was initialised it simply returns it (and ignores the
308 flags. If that is troubling you, check <code>ev_backend ()</code> afterwards).</p>
309 <p>If you don't know what event loop to use, use the one returned from this
311 <p>The flags argument can be used to specify special behaviour or specific
312 backends to use, and is usually specified as <code>0</code> (or <code>EVFLAG_AUTO</code>).</p>
313 <p>The following flags are supported:</p>
316 <dt><code>EVFLAG_AUTO</code></dt>
318 <p>The default flags value. Use this if you have no clue (it's the right
319 thing, believe me).</p>
321 <dt><code>EVFLAG_NOENV</code></dt>
323 <p>If this flag bit is ored into the flag value (or the program runs setuid
324 or setgid) then libev will <i>not</i> look at the environment variable
325 <code>LIBEV_FLAGS</code>. Otherwise (the default), this environment variable will
326 override the flags completely if it is found in the environment. This is
327 useful to try out specific backends to test their performance, or to work
330 <dt><code>EVBACKEND_SELECT</code> (value 1, portable select backend)</dt>
332 <p>This is your standard select(2) backend. Not <i>completely</i> standard, as
333 libev tries to roll its own fd_set with no limits on the number of fds,
334 but if that fails, expect a fairly low limit on the number of fds when
335 using this backend. It doesn't scale too well (O(highest_fd)), but its usually
336 the fastest backend for a low number of fds.</p>
338 <dt><code>EVBACKEND_POLL</code> (value 2, poll backend, available everywhere except on windows)</dt>
340 <p>And this is your standard poll(2) backend. It's more complicated than
341 select, but handles sparse fds better and has no artificial limit on the
342 number of fds you can use (except it will slow down considerably with a
343 lot of inactive fds). It scales similarly to select, i.e. O(total_fds).</p>
345 <dt><code>EVBACKEND_EPOLL</code> (value 4, Linux)</dt>
347 <p>For few fds, this backend is a bit little slower than poll and select,
348 but it scales phenomenally better. While poll and select usually scale like
349 O(total_fds) where n is the total number of fds (or the highest fd), epoll scales
350 either O(1) or O(active_fds).</p>
351 <p>While stopping and starting an I/O watcher in the same iteration will
352 result in some caching, there is still a syscall per such incident
353 (because the fd could point to a different file description now), so its
354 best to avoid that. Also, dup()ed file descriptors might not work very
355 well if you register events for both fds.</p>
356 <p>Please note that epoll sometimes generates spurious notifications, so you
357 need to use non-blocking I/O or other means to avoid blocking when no data
358 (or space) is available.</p>
360 <dt><code>EVBACKEND_KQUEUE</code> (value 8, most BSD clones)</dt>
362 <p>Kqueue deserves special mention, as at the time of this writing, it
363 was broken on all BSDs except NetBSD (usually it doesn't work with
364 anything but sockets and pipes, except on Darwin, where of course its
365 completely useless). For this reason its not being "autodetected"
366 unless you explicitly specify it explicitly in the flags (i.e. using
367 <code>EVBACKEND_KQUEUE</code>).</p>
368 <p>It scales in the same way as the epoll backend, but the interface to the
369 kernel is more efficient (which says nothing about its actual speed, of
370 course). While starting and stopping an I/O watcher does not cause an
371 extra syscall as with epoll, it still adds up to four event changes per
372 incident, so its best to avoid that.</p>
374 <dt><code>EVBACKEND_DEVPOLL</code> (value 16, Solaris 8)</dt>
376 <p>This is not implemented yet (and might never be).</p>
378 <dt><code>EVBACKEND_PORT</code> (value 32, Solaris 10)</dt>
380 <p>This uses the Solaris 10 port mechanism. As with everything on Solaris,
381 it's really slow, but it still scales very well (O(active_fds)).</p>
382 <p>Please note that solaris ports can result in a lot of spurious
383 notifications, so you need to use non-blocking I/O or other means to avoid
384 blocking when no data (or space) is available.</p>
386 <dt><code>EVBACKEND_ALL</code></dt>
388 <p>Try all backends (even potentially broken ones that wouldn't be tried
389 with <code>EVFLAG_AUTO</code>). Since this is a mask, you can do stuff such as
390 <code>EVBACKEND_ALL & ~EVBACKEND_KQUEUE</code>.</p>
394 <p>If one or more of these are ored into the flags value, then only these
395 backends will be tried (in the reverse order as given here). If none are
396 specified, most compiled-in backend will be tried, usually in reverse
397 order of their flag values :)</p>
398 <p>The most typical usage is like this:</p>
399 <pre> if (!ev_default_loop (0))
400 fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
403 <p>Restrict libev to the select and poll backends, and do not allow
404 environment settings to be taken into account:</p>
405 <pre> ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);
408 <p>Use whatever libev has to offer, but make sure that kqueue is used if
409 available (warning, breaks stuff, best use only with your own private
410 event loop and only if you know the OS supports your types of fds):</p>
411 <pre> ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);
415 <dt>struct ev_loop *ev_loop_new (unsigned int flags)</dt>
417 <p>Similar to <code>ev_default_loop</code>, but always creates a new event loop that is
418 always distinct from the default loop. Unlike the default loop, it cannot
419 handle signal and child watchers, and attempts to do so will be greeted by
420 undefined behaviour (or a failed assertion if assertions are enabled).</p>
421 <p>Example: Try to create a event loop that uses epoll and nothing else.</p>
422 <pre> struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);
424 fatal ("no epoll found here, maybe it hides under your chair");
428 <dt>ev_default_destroy ()</dt>
430 <p>Destroys the default loop again (frees all memory and kernel state
431 etc.). None of the active event watchers will be stopped in the normal
432 sense, so e.g. <code>ev_is_active</code> might still return true. It is your
433 responsibility to either stop all watchers cleanly yoursef <i>before</i>
434 calling this function, or cope with the fact afterwards (which is usually
435 the easiest thing, youc na just ignore the watchers and/or <code>free ()</code> them
438 <dt>ev_loop_destroy (loop)</dt>
440 <p>Like <code>ev_default_destroy</code>, but destroys an event loop created by an
441 earlier call to <code>ev_loop_new</code>.</p>
443 <dt>ev_default_fork ()</dt>
445 <p>This function reinitialises the kernel state for backends that have
446 one. Despite the name, you can call it anytime, but it makes most sense
447 after forking, in either the parent or child process (or both, but that
448 again makes little sense).</p>
449 <p>You <i>must</i> call this function in the child process after forking if and
450 only if you want to use the event library in both processes. If you just
451 fork+exec, you don't have to call it.</p>
452 <p>The function itself is quite fast and it's usually not a problem to call
453 it just in case after a fork. To make this easy, the function will fit in
454 quite nicely into a call to <code>pthread_atfork</code>:</p>
455 <pre> pthread_atfork (0, 0, ev_default_fork);
458 <p>At the moment, <code>EVBACKEND_SELECT</code> and <code>EVBACKEND_POLL</code> are safe to use
459 without calling this function, so if you force one of those backends you
460 do not need to care.</p>
462 <dt>ev_loop_fork (loop)</dt>
464 <p>Like <code>ev_default_fork</code>, but acts on an event loop created by
465 <code>ev_loop_new</code>. Yes, you have to call this on every allocated event loop
466 after fork, and how you do this is entirely your own problem.</p>
468 <dt>unsigned int ev_backend (loop)</dt>
470 <p>Returns one of the <code>EVBACKEND_*</code> flags indicating the event backend in
473 <dt>ev_tstamp ev_now (loop)</dt>
475 <p>Returns the current "event loop time", which is the time the event loop
476 received events and started processing them. This timestamp does not
477 change as long as callbacks are being processed, and this is also the base
478 time used for relative timers. You can treat it as the timestamp of the
479 event occuring (or more correctly, libev finding out about it).</p>
481 <dt>ev_loop (loop, int flags)</dt>
483 <p>Finally, this is it, the event handler. This function usually is called
484 after you initialised all your watchers and you want to start handling
486 <p>If the flags argument is specified as <code>0</code>, it will not return until
487 either no event watchers are active anymore or <code>ev_unloop</code> was called.</p>
488 <p>Please note that an explicit <code>ev_unloop</code> is usually better than
489 relying on all watchers to be stopped when deciding when a program has
490 finished (especially in interactive programs), but having a program that
491 automatically loops as long as it has to and no longer by virtue of
492 relying on its watchers stopping correctly is a thing of beauty.</p>
493 <p>A flags value of <code>EVLOOP_NONBLOCK</code> will look for new events, will handle
494 those events and any outstanding ones, but will not block your process in
495 case there are no events and will return after one iteration of the loop.</p>
496 <p>A flags value of <code>EVLOOP_ONESHOT</code> will look for new events (waiting if
497 neccessary) and will handle those and any outstanding ones. It will block
498 your process until at least one new event arrives, and will return after
499 one iteration of the loop. This is useful if you are waiting for some
500 external event in conjunction with something not expressible using other
501 libev watchers. However, a pair of <code>ev_prepare</code>/<code>ev_check</code> watchers is
502 usually a better approach for this kind of thing.</p>
503 <p>Here are the gory details of what <code>ev_loop</code> does:</p>
504 <pre> * If there are no active watchers (reference count is zero), return.
505 - Queue prepare watchers and then call all outstanding watchers.
506 - If we have been forked, recreate the kernel state.
507 - Update the kernel state with all outstanding changes.
508 - Update the "event loop time".
509 - Calculate for how long to block.
510 - Block the process, waiting for any events.
511 - Queue all outstanding I/O (fd) events.
512 - Update the "event loop time" and do time jump handling.
513 - Queue all outstanding timers.
514 - Queue all outstanding periodics.
515 - If no events are pending now, queue all idle watchers.
516 - Queue all check watchers.
517 - Call all queued watchers in reverse order (i.e. check watchers first).
518 Signals and child watchers are implemented as I/O watchers, and will
519 be handled here by queueing them when their watcher gets executed.
520 - If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK
521 were used, return, otherwise continue with step *.
524 <p>Example: Queue some jobs and then loop until no events are outsanding
526 <pre> ... queue jobs here, make sure they register event watchers as long
527 ... as they still have work to do (even an idle watcher will do..)
528 ev_loop (my_loop, 0);
533 <dt>ev_unloop (loop, how)</dt>
535 <p>Can be used to make a call to <code>ev_loop</code> return early (but only after it
536 has processed all outstanding events). The <code>how</code> argument must be either
537 <code>EVUNLOOP_ONE</code>, which will make the innermost <code>ev_loop</code> call return, or
538 <code>EVUNLOOP_ALL</code>, which will make all nested <code>ev_loop</code> calls return.</p>
540 <dt>ev_ref (loop)</dt>
541 <dt>ev_unref (loop)</dt>
543 <p>Ref/unref can be used to add or remove a reference count on the event
544 loop: Every watcher keeps one reference, and as long as the reference
545 count is nonzero, <code>ev_loop</code> will not return on its own. If you have
546 a watcher you never unregister that should not keep <code>ev_loop</code> from
547 returning, ev_unref() after starting, and ev_ref() before stopping it. For
548 example, libev itself uses this for its internal signal pipe: It is not
549 visible to the libev user and should not keep <code>ev_loop</code> from exiting if
550 no event watchers registered by it are active. It is also an excellent
551 way to do this for generic recurring timers or from within third-party
552 libraries. Just remember to <i>unref after start</i> and <i>ref before stop</i>.</p>
553 <p>Example: Create a signal watcher, but keep it from keeping <code>ev_loop</code>
554 running when nothing else is active.</p>
555 <pre> struct ev_signal exitsig;
556 ev_signal_init (&exitsig, sig_cb, SIGINT);
557 ev_signal_start (loop, &exitsig);
561 <p>Example: For some weird reason, unregister the above signal handler again.</p>
563 ev_signal_stop (loop, &exitsig);
574 <h1 id="ANATOMY_OF_A_WATCHER">ANATOMY OF A WATCHER</h1>
575 <div id="ANATOMY_OF_A_WATCHER_CONTENT">
576 <p>A watcher is a structure that you create and register to record your
577 interest in some event. For instance, if you want to wait for STDIN to
578 become readable, you would create an <code>ev_io</code> watcher for that:</p>
579 <pre> static void my_cb (struct ev_loop *loop, struct ev_io *w, int revents)
582 ev_unloop (loop, EVUNLOOP_ALL);
585 struct ev_loop *loop = ev_default_loop (0);
586 struct ev_io stdin_watcher;
587 ev_init (&stdin_watcher, my_cb);
588 ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);
589 ev_io_start (loop, &stdin_watcher);
593 <p>As you can see, you are responsible for allocating the memory for your
594 watcher structures (and it is usually a bad idea to do this on the stack,
595 although this can sometimes be quite valid).</p>
596 <p>Each watcher structure must be initialised by a call to <code>ev_init
597 (watcher *, callback)</code>, which expects a callback to be provided. This
598 callback gets invoked each time the event occurs (or, in the case of io
599 watchers, each time the event loop detects that the file descriptor given
600 is readable and/or writable).</p>
601 <p>Each watcher type has its own <code>ev_<type>_set (watcher *, ...)</code> macro
602 with arguments specific to this watcher type. There is also a macro
603 to combine initialisation and setting in one call: <code>ev_<type>_init
604 (watcher *, callback, ...)</code>.</p>
605 <p>To make the watcher actually watch out for events, you have to start it
606 with a watcher-specific start function (<code>ev_<type>_start (loop, watcher
607 *)</code>), and you can stop watching for events at any time by calling the
608 corresponding stop function (<code>ev_<type>_stop (loop, watcher *)</code>.</p>
609 <p>As long as your watcher is active (has been started but not stopped) you
610 must not touch the values stored in it. Most specifically you must never
611 reinitialise it or call its <code>set</code> macro.</p>
612 <p>Each and every callback receives the event loop pointer as first, the
613 registered watcher structure as second, and a bitset of received events as
615 <p>The received events usually include a single bit per event type received
616 (you can receive multiple events at the same time). The possible bit masks
619 <dt><code>EV_READ</code></dt>
620 <dt><code>EV_WRITE</code></dt>
622 <p>The file descriptor in the <code>ev_io</code> watcher has become readable and/or
625 <dt><code>EV_TIMEOUT</code></dt>
627 <p>The <code>ev_timer</code> watcher has timed out.</p>
629 <dt><code>EV_PERIODIC</code></dt>
631 <p>The <code>ev_periodic</code> watcher has timed out.</p>
633 <dt><code>EV_SIGNAL</code></dt>
635 <p>The signal specified in the <code>ev_signal</code> watcher has been received by a thread.</p>
637 <dt><code>EV_CHILD</code></dt>
639 <p>The pid specified in the <code>ev_child</code> watcher has received a status change.</p>
641 <dt><code>EV_STAT</code></dt>
643 <p>The path specified in the <code>ev_stat</code> watcher changed its attributes somehow.</p>
645 <dt><code>EV_IDLE</code></dt>
647 <p>The <code>ev_idle</code> watcher has determined that you have nothing better to do.</p>
649 <dt><code>EV_PREPARE</code></dt>
650 <dt><code>EV_CHECK</code></dt>
652 <p>All <code>ev_prepare</code> watchers are invoked just <i>before</i> <code>ev_loop</code> starts
653 to gather new events, and all <code>ev_check</code> watchers are invoked just after
654 <code>ev_loop</code> has gathered them, but before it invokes any callbacks for any
655 received events. Callbacks of both watcher types can start and stop as
656 many watchers as they want, and all of them will be taken into account
657 (for example, a <code>ev_prepare</code> watcher might start an idle watcher to keep
658 <code>ev_loop</code> from blocking).</p>
660 <dt><code>EV_EMBED</code></dt>
662 <p>The embedded event loop specified in the <code>ev_embed</code> watcher needs attention.</p>
664 <dt><code>EV_FORK</code></dt>
666 <p>The event loop has been resumed in the child process after fork (see
667 <code>ev_fork</code>).</p>
669 <dt><code>EV_ERROR</code></dt>
671 <p>An unspecified error has occured, the watcher has been stopped. This might
672 happen because the watcher could not be properly started because libev
673 ran out of memory, a file descriptor was found to be closed or any other
674 problem. You best act on it by reporting the problem and somehow coping
675 with the watcher being stopped.</p>
676 <p>Libev will usually signal a few "dummy" events together with an error,
677 for example it might indicate that a fd is readable or writable, and if
678 your callbacks is well-written it can just attempt the operation and cope
679 with the error from read() or write(). This will not work in multithreaded
680 programs, though, so beware.</p>
685 <h2 id="GENERIC_WATCHER_FUNCTIONS">GENERIC WATCHER FUNCTIONS</h2>
686 <div id="GENERIC_WATCHER_FUNCTIONS_CONTENT">
687 <p>In the following description, <code>TYPE</code> stands for the watcher type,
688 e.g. <code>timer</code> for <code>ev_timer</code> watchers and <code>io</code> for <code>ev_io</code> watchers.</p>
690 <dt><code>ev_init</code> (ev_TYPE *watcher, callback)</dt>
692 <p>This macro initialises the generic portion of a watcher. The contents
693 of the watcher object can be arbitrary (so <code>malloc</code> will do). Only
694 the generic parts of the watcher are initialised, you <i>need</i> to call
695 the type-specific <code>ev_TYPE_set</code> macro afterwards to initialise the
696 type-specific parts. For each type there is also a <code>ev_TYPE_init</code> macro
697 which rolls both calls into one.</p>
698 <p>You can reinitialise a watcher at any time as long as it has been stopped
699 (or never started) and there are no pending events outstanding.</p>
700 <p>The callback is always of type <code>void (*)(ev_loop *loop, ev_TYPE *watcher,
701 int revents)</code>.</p>
703 <dt><code>ev_TYPE_set</code> (ev_TYPE *, [args])</dt>
705 <p>This macro initialises the type-specific parts of a watcher. You need to
706 call <code>ev_init</code> at least once before you call this macro, but you can
707 call <code>ev_TYPE_set</code> any number of times. You must not, however, call this
708 macro on a watcher that is active (it can be pending, however, which is a
709 difference to the <code>ev_init</code> macro).</p>
710 <p>Although some watcher types do not have type-specific arguments
711 (e.g. <code>ev_prepare</code>) you still need to call its <code>set</code> macro.</p>
713 <dt><code>ev_TYPE_init</code> (ev_TYPE *watcher, callback, [args])</dt>
715 <p>This convinience macro rolls both <code>ev_init</code> and <code>ev_TYPE_set</code> macro
716 calls into a single call. This is the most convinient method to initialise
717 a watcher. The same limitations apply, of course.</p>
719 <dt><code>ev_TYPE_start</code> (loop *, ev_TYPE *watcher)</dt>
721 <p>Starts (activates) the given watcher. Only active watchers will receive
722 events. If the watcher is already active nothing will happen.</p>
724 <dt><code>ev_TYPE_stop</code> (loop *, ev_TYPE *watcher)</dt>
726 <p>Stops the given watcher again (if active) and clears the pending
727 status. It is possible that stopped watchers are pending (for example,
728 non-repeating timers are being stopped when they become pending), but
729 <code>ev_TYPE_stop</code> ensures that the watcher is neither active nor pending. If
730 you want to free or reuse the memory used by the watcher it is therefore a
731 good idea to always call its <code>ev_TYPE_stop</code> function.</p>
733 <dt>bool ev_is_active (ev_TYPE *watcher)</dt>
735 <p>Returns a true value iff the watcher is active (i.e. it has been started
736 and not yet been stopped). As long as a watcher is active you must not modify
739 <dt>bool ev_is_pending (ev_TYPE *watcher)</dt>
741 <p>Returns a true value iff the watcher is pending, (i.e. it has outstanding
742 events but its callback has not yet been invoked). As long as a watcher
743 is pending (but not active) you must not call an init function on it (but
744 <code>ev_TYPE_set</code> is safe) and you must make sure the watcher is available to
745 libev (e.g. you cnanot <code>free ()</code> it).</p>
747 <dt>callback ev_cb (ev_TYPE *watcher)</dt>
749 <p>Returns the callback currently set on the watcher.</p>
751 <dt>ev_cb_set (ev_TYPE *watcher, callback)</dt>
753 <p>Change the callback. You can change the callback at virtually any time
754 (modulo threads).</p>
763 <h2 id="ASSOCIATING_CUSTOM_DATA_WITH_A_WATCH">ASSOCIATING CUSTOM DATA WITH A WATCHER</h2>
764 <div id="ASSOCIATING_CUSTOM_DATA_WITH_A_WATCH-2">
765 <p>Each watcher has, by default, a member <code>void *data</code> that you can change
766 and read at any time, libev will completely ignore it. This can be used
767 to associate arbitrary data with your watcher. If you need more data and
768 don't want to allocate memory and store a pointer to it in that data
769 member, you can also "subclass" the watcher type and provide your own
776 struct whatever *mostinteresting;
780 <p>And since your callback will be called with a pointer to the watcher, you
781 can cast it back to your own type:</p>
782 <pre> static void my_cb (struct ev_loop *loop, struct ev_io *w_, int revents)
784 struct my_io *w = (struct my_io *)w_;
789 <p>More interesting and less C-conformant ways of casting your callback type
790 instead have been omitted.</p>
791 <p>Another common scenario is having some data structure with multiple
793 <pre> struct my_biggy
801 <p>In this case getting the pointer to <code>my_biggy</code> is a bit more complicated,
802 you need to use <code>offsetof</code>:</p>
803 <pre> #include <stddef.h>
806 t1_cb (EV_P_ struct ev_timer *w, int revents)
808 struct my_biggy big = (struct my_biggy *
809 (((char *)w) - offsetof (struct my_biggy, t1));
813 t2_cb (EV_P_ struct ev_timer *w, int revents)
815 struct my_biggy big = (struct my_biggy *
816 (((char *)w) - offsetof (struct my_biggy, t2));
825 <h1 id="WATCHER_TYPES">WATCHER TYPES</h1>
826 <div id="WATCHER_TYPES_CONTENT">
827 <p>This section describes each watcher in detail, but will not repeat
828 information given in the last section. Any initialisation/set macros,
829 functions and members specific to the watcher type are explained.</p>
830 <p>Members are additionally marked with either <i>[read-only]</i>, meaning that,
831 while the watcher is active, you can look at the member and expect some
832 sensible content, but you must not modify it (you can modify it while the
833 watcher is stopped to your hearts content), or <i>[read-write]</i>, which
834 means you can expect it to have some sensible content while the watcher
835 is active, but you can also modify it. Modifying it may not do something
836 sensible or take immediate effect (or do anything at all), but libev will
837 not crash or malfunction in any way.</p>
844 <h2 id="code_ev_io_code_is_this_file_descrip"><code>ev_io</code> - is this file descriptor readable or writable?</h2>
845 <div id="code_ev_io_code_is_this_file_descrip-2">
846 <p>I/O watchers check whether a file descriptor is readable or writable
847 in each iteration of the event loop, or, more precisely, when reading
848 would not block the process and writing would at least be able to write
849 some data. This behaviour is called level-triggering because you keep
850 receiving events as long as the condition persists. Remember you can stop
851 the watcher if you don't want to act on the event and neither want to
852 receive future events.</p>
853 <p>In general you can register as many read and/or write event watchers per
854 fd as you want (as long as you don't confuse yourself). Setting all file
855 descriptors to non-blocking mode is also usually a good idea (but not
856 required if you know what you are doing).</p>
857 <p>You have to be careful with dup'ed file descriptors, though. Some backends
858 (the linux epoll backend is a notable example) cannot handle dup'ed file
859 descriptors correctly if you register interest in two or more fds pointing
860 to the same underlying file/socket/etc. description (that is, they share
861 the same underlying "file open").</p>
862 <p>If you must do this, then force the use of a known-to-be-good backend
863 (at the time of this writing, this includes only <code>EVBACKEND_SELECT</code> and
864 <code>EVBACKEND_POLL</code>).</p>
865 <p>Another thing you have to watch out for is that it is quite easy to
866 receive "spurious" readyness notifications, that is your callback might
867 be called with <code>EV_READ</code> but a subsequent <code>read</code>(2) will actually block
868 because there is no data. Not only are some backends known to create a
869 lot of those (for example solaris ports), it is very easy to get into
870 this situation even with a relatively standard program structure. Thus
871 it is best to always use non-blocking I/O: An extra <code>read</code>(2) returning
872 <code>EAGAIN</code> is far preferable to a program hanging until some data arrives.</p>
873 <p>If you cannot run the fd in non-blocking mode (for example you should not
874 play around with an Xlib connection), then you have to seperately re-test
875 wether a file descriptor is really ready with a known-to-be good interface
876 such as poll (fortunately in our Xlib example, Xlib already does this on
877 its own, so its quite safe to use).</p>
879 <dt>ev_io_init (ev_io *, callback, int fd, int events)</dt>
880 <dt>ev_io_set (ev_io *, int fd, int events)</dt>
882 <p>Configures an <code>ev_io</code> watcher. The <code>fd</code> is the file descriptor to
883 rceeive events for and events is either <code>EV_READ</code>, <code>EV_WRITE</code> or
884 <code>EV_READ | EV_WRITE</code> to receive the given events.</p>
886 <dt>int fd [read-only]</dt>
888 <p>The file descriptor being watched.</p>
890 <dt>int events [read-only]</dt>
892 <p>The events being watched.</p>
895 <p>Example: Call <code>stdin_readable_cb</code> when STDIN_FILENO has become, well
896 readable, but only once. Since it is likely line-buffered, you could
897 attempt to read a whole line in the callback.</p>
899 stdin_readable_cb (struct ev_loop *loop, struct ev_io *w, int revents)
901 ev_io_stop (loop, w);
902 .. read from stdin here (or from w->fd) and haqndle any I/O errors
906 struct ev_loop *loop = ev_default_init (0);
907 struct ev_io stdin_readable;
908 ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
909 ev_io_start (loop, &stdin_readable);
918 <h2 id="code_ev_timer_code_relative_and_opti"><code>ev_timer</code> - relative and optionally repeating timeouts</h2>
919 <div id="code_ev_timer_code_relative_and_opti-2">
920 <p>Timer watchers are simple relative timers that generate an event after a
921 given time, and optionally repeating in regular intervals after that.</p>
922 <p>The timers are based on real time, that is, if you register an event that
923 times out after an hour and you reset your system clock to last years
924 time, it will still time out after (roughly) and hour. "Roughly" because
925 detecting time jumps is hard, and some inaccuracies are unavoidable (the
926 monotonic clock option helps a lot here).</p>
927 <p>The relative timeouts are calculated relative to the <code>ev_now ()</code>
928 time. This is usually the right thing as this timestamp refers to the time
929 of the event triggering whatever timeout you are modifying/starting. If
930 you suspect event processing to be delayed and you <i>need</i> to base the timeout
931 on the current time, use something like this to adjust for this:</p>
932 <pre> ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);
935 <p>The callback is guarenteed to be invoked only when its timeout has passed,
936 but if multiple timers become ready during the same loop iteration then
937 order of execution is undefined.</p>
939 <dt>ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)</dt>
940 <dt>ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)</dt>
942 <p>Configure the timer to trigger after <code>after</code> seconds. If <code>repeat</code> is
943 <code>0.</code>, then it will automatically be stopped. If it is positive, then the
944 timer will automatically be configured to trigger again <code>repeat</code> seconds
945 later, again, and again, until stopped manually.</p>
946 <p>The timer itself will do a best-effort at avoiding drift, that is, if you
947 configure a timer to trigger every 10 seconds, then it will trigger at
948 exactly 10 second intervals. If, however, your program cannot keep up with
949 the timer (because it takes longer than those 10 seconds to do stuff) the
950 timer will not fire more than once per event loop iteration.</p>
952 <dt>ev_timer_again (loop)</dt>
954 <p>This will act as if the timer timed out and restart it again if it is
955 repeating. The exact semantics are:</p>
956 <p>If the timer is started but nonrepeating, stop it.</p>
957 <p>If the timer is repeating, either start it if necessary (with the repeat
958 value), or reset the running timer to the repeat value.</p>
959 <p>This sounds a bit complicated, but here is a useful and typical
960 example: Imagine you have a tcp connection and you want a so-called
961 idle timeout, that is, you want to be called when there have been,
962 say, 60 seconds of inactivity on the socket. The easiest way to do
963 this is to configure an <code>ev_timer</code> with <code>after</code>=<code>repeat</code>=<code>60</code> and calling
964 <code>ev_timer_again</code> each time you successfully read or write some data. If
965 you go into an idle state where you do not expect data to travel on the
966 socket, you can stop the timer, and again will automatically restart it if
968 <p>You can also ignore the <code>after</code> value and <code>ev_timer_start</code> altogether
969 and only ever use the <code>repeat</code> value:</p>
970 <pre> ev_timer_init (timer, callback, 0., 5.);
971 ev_timer_again (loop, timer);
973 timer->again = 17.;
974 ev_timer_again (loop, timer);
976 timer->again = 10.;
977 ev_timer_again (loop, timer);
980 <p>This is more efficient then stopping/starting the timer eahc time you want
981 to modify its timeout value.</p>
983 <dt>ev_tstamp repeat [read-write]</dt>
985 <p>The current <code>repeat</code> value. Will be used each time the watcher times out
986 or <code>ev_timer_again</code> is called and determines the next timeout (if any),
987 which is also when any modifications are taken into account.</p>
990 <p>Example: Create a timer that fires after 60 seconds.</p>
992 one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
994 .. one minute over, w is actually stopped right here
997 struct ev_timer mytimer;
998 ev_timer_init (&mytimer, one_minute_cb, 60., 0.);
999 ev_timer_start (loop, &mytimer);
1002 <p>Example: Create a timeout timer that times out after 10 seconds of
1005 timeout_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
1007 .. ten seconds without any activity
1010 struct ev_timer mytimer;
1011 ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */
1012 ev_timer_again (&mytimer); /* start timer */
1015 // and in some piece of code that gets executed on any "activity":
1016 // reset the timeout to start ticking again at 10 seconds
1017 ev_timer_again (&mytimer);
1025 <h2 id="code_ev_periodic_code_to_cron_or_not"><code>ev_periodic</code> - to cron or not to cron?</h2>
1026 <div id="code_ev_periodic_code_to_cron_or_not-2">
1027 <p>Periodic watchers are also timers of a kind, but they are very versatile
1028 (and unfortunately a bit complex).</p>
1029 <p>Unlike <code>ev_timer</code>'s, they are not based on real time (or relative time)
1030 but on wallclock time (absolute time). You can tell a periodic watcher
1031 to trigger "at" some specific point in time. For example, if you tell a
1032 periodic watcher to trigger in 10 seconds (by specifiying e.g. <code>ev_now ()
1033 + 10.</code>) and then reset your system clock to the last year, then it will
1034 take a year to trigger the event (unlike an <code>ev_timer</code>, which would trigger
1035 roughly 10 seconds later and of course not if you reset your system time
1037 <p>They can also be used to implement vastly more complex timers, such as
1038 triggering an event on eahc midnight, local time.</p>
1039 <p>As with timers, the callback is guarenteed to be invoked only when the
1040 time (<code>at</code>) has been passed, but if multiple periodic timers become ready
1041 during the same loop iteration then order of execution is undefined.</p>
1043 <dt>ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)</dt>
1044 <dt>ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)</dt>
1046 <p>Lots of arguments, lets sort it out... There are basically three modes of
1047 operation, and we will explain them from simplest to complex:</p>
1050 <dt>* absolute timer (interval = reschedule_cb = 0)</dt>
1052 <p>In this configuration the watcher triggers an event at the wallclock time
1053 <code>at</code> and doesn't repeat. It will not adjust when a time jump occurs,
1054 that is, if it is to be run at January 1st 2011 then it will run when the
1055 system time reaches or surpasses this time.</p>
1057 <dt>* non-repeating interval timer (interval > 0, reschedule_cb = 0)</dt>
1059 <p>In this mode the watcher will always be scheduled to time out at the next
1060 <code>at + N * interval</code> time (for some integer N) and then repeat, regardless
1061 of any time jumps.</p>
1062 <p>This can be used to create timers that do not drift with respect to system
1064 <pre> ev_periodic_set (&periodic, 0., 3600., 0);
1067 <p>This doesn't mean there will always be 3600 seconds in between triggers,
1068 but only that the the callback will be called when the system time shows a
1069 full hour (UTC), or more correctly, when the system time is evenly divisible
1071 <p>Another way to think about it (for the mathematically inclined) is that
1072 <code>ev_periodic</code> will try to run the callback in this mode at the next possible
1073 time where <code>time = at (mod interval)</code>, regardless of any time jumps.</p>
1075 <dt>* manual reschedule mode (reschedule_cb = callback)</dt>
1077 <p>In this mode the values for <code>interval</code> and <code>at</code> are both being
1078 ignored. Instead, each time the periodic watcher gets scheduled, the
1079 reschedule callback will be called with the watcher as first, and the
1080 current time as second argument.</p>
1081 <p>NOTE: <i>This callback MUST NOT stop or destroy any periodic watcher,
1082 ever, or make any event loop modifications</i>. If you need to stop it,
1083 return <code>now + 1e30</code> (or so, fudge fudge) and stop it afterwards (e.g. by
1084 starting a prepare watcher).</p>
1085 <p>Its prototype is <code>ev_tstamp (*reschedule_cb)(struct ev_periodic *w,
1086 ev_tstamp now)</code>, e.g.:</p>
1087 <pre> static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)
1093 <p>It must return the next time to trigger, based on the passed time value
1094 (that is, the lowest time value larger than to the second argument). It
1095 will usually be called just before the callback will be triggered, but
1096 might be called at other times, too.</p>
1097 <p>NOTE: <i>This callback must always return a time that is later than the
1098 passed <code>now</code> value</i>. Not even <code>now</code> itself will do, it <i>must</i> be larger.</p>
1099 <p>This can be used to create very complex timers, such as a timer that
1100 triggers on each midnight, local time. To do this, you would calculate the
1101 next midnight after <code>now</code> and return the timestamp value for this. How
1102 you do this is, again, up to you (but it is not trivial, which is the main
1103 reason I omitted it as an example).</p>
1108 <dt>ev_periodic_again (loop, ev_periodic *)</dt>
1110 <p>Simply stops and restarts the periodic watcher again. This is only useful
1111 when you changed some parameters or the reschedule callback would return
1112 a different time than the last time it was called (e.g. in a crond like
1113 program when the crontabs have changed).</p>
1115 <dt>ev_tstamp interval [read-write]</dt>
1117 <p>The current interval value. Can be modified any time, but changes only
1118 take effect when the periodic timer fires or <code>ev_periodic_again</code> is being
1121 <dt>ev_tstamp (*reschedule_cb)(struct ev_periodic *w, ev_tstamp now) [read-write]</dt>
1123 <p>The current reschedule callback, or <code>0</code>, if this functionality is
1124 switched off. Can be changed any time, but changes only take effect when
1125 the periodic timer fires or <code>ev_periodic_again</code> is being called.</p>
1128 <p>Example: Call a callback every hour, or, more precisely, whenever the
1129 system clock is divisible by 3600. The callback invocation times have
1130 potentially a lot of jittering, but good long-term stability.</p>
1132 clock_cb (struct ev_loop *loop, struct ev_io *w, int revents)
1134 ... its now a full hour (UTC, or TAI or whatever your clock follows)
1137 struct ev_periodic hourly_tick;
1138 ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);
1139 ev_periodic_start (loop, &hourly_tick);
1142 <p>Example: The same as above, but use a reschedule callback to do it:</p>
1143 <pre> #include <math.h>
1146 my_scheduler_cb (struct ev_periodic *w, ev_tstamp now)
1148 return fmod (now, 3600.) + 3600.;
1151 ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
1154 <p>Example: Call a callback every hour, starting now:</p>
1155 <pre> struct ev_periodic hourly_tick;
1156 ev_periodic_init (&hourly_tick, clock_cb,
1157 fmod (ev_now (loop), 3600.), 3600., 0);
1158 ev_periodic_start (loop, &hourly_tick);
1166 <h2 id="code_ev_signal_code_signal_me_when_a"><code>ev_signal</code> - signal me when a signal gets signalled!</h2>
1167 <div id="code_ev_signal_code_signal_me_when_a-2">
1168 <p>Signal watchers will trigger an event when the process receives a specific
1169 signal one or more times. Even though signals are very asynchronous, libev
1170 will try it's best to deliver signals synchronously, i.e. as part of the
1171 normal event processing, like any other event.</p>
1172 <p>You can configure as many watchers as you like per signal. Only when the
1173 first watcher gets started will libev actually register a signal watcher
1174 with the kernel (thus it coexists with your own signal handlers as long
1175 as you don't register any with libev). Similarly, when the last signal
1176 watcher for a signal is stopped libev will reset the signal handler to
1177 SIG_DFL (regardless of what it was set to before).</p>
1179 <dt>ev_signal_init (ev_signal *, callback, int signum)</dt>
1180 <dt>ev_signal_set (ev_signal *, int signum)</dt>
1182 <p>Configures the watcher to trigger on the given signal number (usually one
1183 of the <code>SIGxxx</code> constants).</p>
1185 <dt>int signum [read-only]</dt>
1187 <p>The signal the watcher watches out for.</p>
1196 <h2 id="code_ev_child_code_watch_out_for_pro"><code>ev_child</code> - watch out for process status changes</h2>
1197 <div id="code_ev_child_code_watch_out_for_pro-2">
1198 <p>Child watchers trigger when your process receives a SIGCHLD in response to
1199 some child status changes (most typically when a child of yours dies).</p>
1201 <dt>ev_child_init (ev_child *, callback, int pid)</dt>
1202 <dt>ev_child_set (ev_child *, int pid)</dt>
1204 <p>Configures the watcher to wait for status changes of process <code>pid</code> (or
1205 <i>any</i> process if <code>pid</code> is specified as <code>0</code>). The callback can look
1206 at the <code>rstatus</code> member of the <code>ev_child</code> watcher structure to see
1207 the status word (use the macros from <code>sys/wait.h</code> and see your systems
1208 <code>waitpid</code> documentation). The <code>rpid</code> member contains the pid of the
1209 process causing the status change.</p>
1211 <dt>int pid [read-only]</dt>
1213 <p>The process id this watcher watches out for, or <code>0</code>, meaning any process id.</p>
1215 <dt>int rpid [read-write]</dt>
1217 <p>The process id that detected a status change.</p>
1219 <dt>int rstatus [read-write]</dt>
1221 <p>The process exit/trace status caused by <code>rpid</code> (see your systems
1222 <code>waitpid</code> and <code>sys/wait.h</code> documentation for details).</p>
1225 <p>Example: Try to exit cleanly on SIGINT and SIGTERM.</p>
1227 sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)
1229 ev_unloop (loop, EVUNLOOP_ALL);
1232 struct ev_signal signal_watcher;
1233 ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
1234 ev_signal_start (loop, &sigint_cb);
1242 <h2 id="code_ev_stat_code_did_the_file_attri"><code>ev_stat</code> - did the file attributes just change?</h2>
1243 <div id="code_ev_stat_code_did_the_file_attri-2">
1244 <p>This watches a filesystem path for attribute changes. That is, it calls
1245 <code>stat</code> regularly (or when the OS says it changed) and sees if it changed
1246 compared to the last time, invoking the callback if it did.</p>
1247 <p>The path does not need to exist: changing from "path exists" to "path does
1248 not exist" is a status change like any other. The condition "path does
1249 not exist" is signified by the <code>st_nlink</code> field being zero (which is
1250 otherwise always forced to be at least one) and all the other fields of
1251 the stat buffer having unspecified contents.</p>
1252 <p>Since there is no standard to do this, the portable implementation simply
1253 calls <code>stat (2)</code> regularly on the path to see if it changed somehow. You
1254 can specify a recommended polling interval for this case. If you specify
1255 a polling interval of <code>0</code> (highly recommended!) then a <i>suitable,
1256 unspecified default</i> value will be used (which you can expect to be around
1257 five seconds, although this might change dynamically). Libev will also
1258 impose a minimum interval which is currently around <code>0.1</code>, but thats
1259 usually overkill.</p>
1260 <p>This watcher type is not meant for massive numbers of stat watchers,
1261 as even with OS-supported change notifications, this can be
1262 resource-intensive.</p>
1263 <p>At the time of this writing, only the Linux inotify interface is
1264 implemented (implementing kqueue support is left as an exercise for the
1265 reader). Inotify will be used to give hints only and should not change the
1266 semantics of <code>ev_stat</code> watchers, which means that libev sometimes needs
1267 to fall back to regular polling again even with inotify, but changes are
1268 usually detected immediately, and if the file exists there will be no
1271 <dt>ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)</dt>
1272 <dt>ev_stat_set (ev_stat *, const char *path, ev_tstamp interval)</dt>
1274 <p>Configures the watcher to wait for status changes of the given
1275 <code>path</code>. The <code>interval</code> is a hint on how quickly a change is expected to
1276 be detected and should normally be specified as <code>0</code> to let libev choose
1277 a suitable value. The memory pointed to by <code>path</code> must point to the same
1278 path for as long as the watcher is active.</p>
1279 <p>The callback will be receive <code>EV_STAT</code> when a change was detected,
1280 relative to the attributes at the time the watcher was started (or the
1281 last change was detected).</p>
1283 <dt>ev_stat_stat (ev_stat *)</dt>
1285 <p>Updates the stat buffer immediately with new values. If you change the
1286 watched path in your callback, you could call this fucntion to avoid
1287 detecting this change (while introducing a race condition). Can also be
1288 useful simply to find out the new values.</p>
1290 <dt>ev_statdata attr [read-only]</dt>
1292 <p>The most-recently detected attributes of the file. Although the type is of
1293 <code>ev_statdata</code>, this is usually the (or one of the) <code>struct stat</code> types
1294 suitable for your system. If the <code>st_nlink</code> member is <code>0</code>, then there
1295 was some error while <code>stat</code>ing the file.</p>
1297 <dt>ev_statdata prev [read-only]</dt>
1299 <p>The previous attributes of the file. The callback gets invoked whenever
1300 <code>prev</code> != <code>attr</code>.</p>
1302 <dt>ev_tstamp interval [read-only]</dt>
1304 <p>The specified interval.</p>
1306 <dt>const char *path [read-only]</dt>
1308 <p>The filesystem path that is being watched.</p>
1311 <p>Example: Watch <code>/etc/passwd</code> for attribute changes.</p>
1313 passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)
1315 /* /etc/passwd changed in some way */
1316 if (w->attr.st_nlink)
1318 printf ("passwd current size %ld\n", (long)w->attr.st_size);
1319 printf ("passwd current atime %ld\n", (long)w->attr.st_mtime);
1320 printf ("passwd current mtime %ld\n", (long)w->attr.st_mtime);
1323 /* you shalt not abuse printf for puts */
1324 puts ("wow, /etc/passwd is not there, expect problems. "
1325 "if this is windows, they already arrived\n");
1331 ev_stat_init (&passwd, passwd_cb, "/etc/passwd");
1332 ev_stat_start (loop, &passwd);
1340 <h2 id="code_ev_idle_code_when_you_ve_got_no"><code>ev_idle</code> - when you've got nothing better to do...</h2>
1341 <div id="code_ev_idle_code_when_you_ve_got_no-2">
1342 <p>Idle watchers trigger events when there are no other events are pending
1343 (prepare, check and other idle watchers do not count). That is, as long
1344 as your process is busy handling sockets or timeouts (or even signals,
1345 imagine) it will not be triggered. But when your process is idle all idle
1346 watchers are being called again and again, once per event loop iteration -
1347 until stopped, that is, or your process receives more events and becomes
1349 <p>The most noteworthy effect is that as long as any idle watchers are
1350 active, the process will not block when waiting for new events.</p>
1351 <p>Apart from keeping your process non-blocking (which is a useful
1352 effect on its own sometimes), idle watchers are a good place to do
1353 "pseudo-background processing", or delay processing stuff to after the
1354 event loop has handled all outstanding events.</p>
1356 <dt>ev_idle_init (ev_signal *, callback)</dt>
1358 <p>Initialises and configures the idle watcher - it has no parameters of any
1359 kind. There is a <code>ev_idle_set</code> macro, but using it is utterly pointless,
1363 <p>Example: Dynamically allocate an <code>ev_idle</code> watcher, start it, and in the
1364 callback, free it. Also, use no error checking, as usual.</p>
1366 idle_cb (struct ev_loop *loop, struct ev_idle *w, int revents)
1369 // now do something you wanted to do when the program has
1370 // no longer asnything immediate to do.
1373 struct ev_idle *idle_watcher = malloc (sizeof (struct ev_idle));
1374 ev_idle_init (idle_watcher, idle_cb);
1375 ev_idle_start (loop, idle_cb);
1383 <h2 id="code_ev_prepare_code_and_code_ev_che"><code>ev_prepare</code> and <code>ev_check</code> - customise your event loop!</h2>
1384 <div id="code_ev_prepare_code_and_code_ev_che-2">
1385 <p>Prepare and check watchers are usually (but not always) used in tandem:
1386 prepare watchers get invoked before the process blocks and check watchers
1388 <p>You <i>must not</i> call <code>ev_loop</code> or similar functions that enter
1389 the current event loop from either <code>ev_prepare</code> or <code>ev_check</code>
1390 watchers. Other loops than the current one are fine, however. The
1391 rationale behind this is that you do not need to check for recursion in
1392 those watchers, i.e. the sequence will always be <code>ev_prepare</code>, blocking,
1393 <code>ev_check</code> so if you have one watcher of each kind they will always be
1394 called in pairs bracketing the blocking call.</p>
1395 <p>Their main purpose is to integrate other event mechanisms into libev and
1396 their use is somewhat advanced. This could be used, for example, to track
1397 variable changes, implement your own watchers, integrate net-snmp or a
1398 coroutine library and lots more. They are also occasionally useful if
1399 you cache some data and want to flush it before blocking (for example,
1400 in X programs you might want to do an <code>XFlush ()</code> in an <code>ev_prepare</code>
1402 <p>This is done by examining in each prepare call which file descriptors need
1403 to be watched by the other library, registering <code>ev_io</code> watchers for
1404 them and starting an <code>ev_timer</code> watcher for any timeouts (many libraries
1405 provide just this functionality). Then, in the check watcher you check for
1406 any events that occured (by checking the pending status of all watchers
1407 and stopping them) and call back into the library. The I/O and timer
1408 callbacks will never actually be called (but must be valid nevertheless,
1409 because you never know, you know?).</p>
1410 <p>As another example, the Perl Coro module uses these hooks to integrate
1411 coroutines into libev programs, by yielding to other active coroutines
1412 during each prepare and only letting the process block if no coroutines
1413 are ready to run (it's actually more complicated: it only runs coroutines
1414 with priority higher than or equal to the event loop and one coroutine
1415 of lower priority, but only once, using idle watchers to keep the event
1416 loop from blocking if lower-priority coroutines are active, thus mapping
1417 low-priority coroutines to idle/background tasks).</p>
1419 <dt>ev_prepare_init (ev_prepare *, callback)</dt>
1420 <dt>ev_check_init (ev_check *, callback)</dt>
1422 <p>Initialises and configures the prepare or check watcher - they have no
1423 parameters of any kind. There are <code>ev_prepare_set</code> and <code>ev_check_set</code>
1424 macros, but using them is utterly, utterly and completely pointless.</p>
1427 <p>Example: To include a library such as adns, you would add IO watchers
1428 and a timeout watcher in a prepare handler, as required by libadns, and
1429 in a check watcher, destroy them and call into libadns. What follows is
1430 pseudo-code only of course:</p>
1431 <pre> static ev_io iow [nfd];
1435 io_cb (ev_loop *loop, ev_io *w, int revents)
1437 // set the relevant poll flags
1438 // could also call adns_processreadable etc. here
1439 struct pollfd *fd = (struct pollfd *)w->data;
1440 if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;
1441 if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;
1444 // create io watchers for each fd and a timer before blocking
1446 adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents)
1448 int timeout = 3600000;truct pollfd fds [nfd];
1449 // actual code will need to loop here and realloc etc.
1450 adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));
1452 /* the callback is illegal, but won't be called as we stop during check */
1453 ev_timer_init (&tw, 0, timeout * 1e-3);
1454 ev_timer_start (loop, &tw);
1456 // create on ev_io per pollfd
1457 for (int i = 0; i < nfd; ++i)
1459 ev_io_init (iow + i, io_cb, fds [i].fd,
1460 ((fds [i].events & POLLIN ? EV_READ : 0)
1461 | (fds [i].events & POLLOUT ? EV_WRITE : 0)));
1463 fds [i].revents = 0;
1464 iow [i].data = fds + i;
1465 ev_io_start (loop, iow + i);
1469 // stop all watchers after blocking
1471 adns_check_cb (ev_loop *loop, ev_check *w, int revents)
1473 ev_timer_stop (loop, &tw);
1475 for (int i = 0; i < nfd; ++i)
1476 ev_io_stop (loop, iow + i);
1478 adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));
1487 <h2 id="code_ev_embed_code_when_one_backend_"><code>ev_embed</code> - when one backend isn't enough...</h2>
1488 <div id="code_ev_embed_code_when_one_backend_-2">
1489 <p>This is a rather advanced watcher type that lets you embed one event loop
1490 into another (currently only <code>ev_io</code> events are supported in the embedded
1491 loop, other types of watchers might be handled in a delayed or incorrect
1492 fashion and must not be used).</p>
1493 <p>There are primarily two reasons you would want that: work around bugs and
1495 <p>As an example for a bug workaround, the kqueue backend might only support
1496 sockets on some platform, so it is unusable as generic backend, but you
1497 still want to make use of it because you have many sockets and it scales
1498 so nicely. In this case, you would create a kqueue-based loop and embed it
1499 into your default loop (which might use e.g. poll). Overall operation will
1500 be a bit slower because first libev has to poll and then call kevent, but
1501 at least you can use both at what they are best.</p>
1502 <p>As for prioritising I/O: rarely you have the case where some fds have
1503 to be watched and handled very quickly (with low latency), and even
1504 priorities and idle watchers might have too much overhead. In this case
1505 you would put all the high priority stuff in one loop and all the rest in
1506 a second one, and embed the second one in the first.</p>
1507 <p>As long as the watcher is active, the callback will be invoked every time
1508 there might be events pending in the embedded loop. The callback must then
1509 call <code>ev_embed_sweep (mainloop, watcher)</code> to make a single sweep and invoke
1510 their callbacks (you could also start an idle watcher to give the embedded
1511 loop strictly lower priority for example). You can also set the callback
1512 to <code>0</code>, in which case the embed watcher will automatically execute the
1513 embedded loop sweep.</p>
1514 <p>As long as the watcher is started it will automatically handle events. The
1515 callback will be invoked whenever some events have been handled. You can
1516 set the callback to <code>0</code> to avoid having to specify one if you are not
1517 interested in that.</p>
1518 <p>Also, there have not currently been made special provisions for forking:
1519 when you fork, you not only have to call <code>ev_loop_fork</code> on both loops,
1520 but you will also have to stop and restart any <code>ev_embed</code> watchers
1522 <p>Unfortunately, not all backends are embeddable, only the ones returned by
1523 <code>ev_embeddable_backends</code> are, which, unfortunately, does not include any
1525 <p>So when you want to use this feature you will always have to be prepared
1526 that you cannot get an embeddable loop. The recommended way to get around
1527 this is to have a separate variables for your embeddable loop, try to
1528 create it, and if that fails, use the normal loop for everything:</p>
1529 <pre> struct ev_loop *loop_hi = ev_default_init (0);
1530 struct ev_loop *loop_lo = 0;
1531 struct ev_embed embed;
1533 // see if there is a chance of getting one that works
1534 // (remember that a flags value of 0 means autodetection)
1535 loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
1536 ? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
1539 // if we got one, then embed it, otherwise default to loop_hi
1542 ev_embed_init (&embed, 0, loop_lo);
1543 ev_embed_start (loop_hi, &embed);
1550 <dt>ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)</dt>
1551 <dt>ev_embed_set (ev_embed *, callback, struct ev_loop *embedded_loop)</dt>
1553 <p>Configures the watcher to embed the given loop, which must be
1554 embeddable. If the callback is <code>0</code>, then <code>ev_embed_sweep</code> will be
1555 invoked automatically, otherwise it is the responsibility of the callback
1556 to invoke it (it will continue to be called until the sweep has been done,
1557 if you do not want thta, you need to temporarily stop the embed watcher).</p>
1559 <dt>ev_embed_sweep (loop, ev_embed *)</dt>
1561 <p>Make a single, non-blocking sweep over the embedded loop. This works
1562 similarly to <code>ev_loop (embedded_loop, EVLOOP_NONBLOCK)</code>, but in the most
1563 apropriate way for embedded loops.</p>
1565 <dt>struct ev_loop *loop [read-only]</dt>
1567 <p>The embedded event loop.</p>
1576 <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>
1577 <div id="code_ev_fork_code_the_audacity_to_re-2">
1578 <p>Fork watchers are called when a <code>fork ()</code> was detected (usually because
1579 whoever is a good citizen cared to tell libev about it by calling
1580 <code>ev_default_fork</code> or <code>ev_loop_fork</code>). The invocation is done before the
1581 event loop blocks next and before <code>ev_check</code> watchers are being called,
1582 and only in the child after the fork. If whoever good citizen calling
1583 <code>ev_default_fork</code> cheats and calls it in the wrong process, the fork
1584 handlers will be invoked, too, of course.</p>
1586 <dt>ev_fork_init (ev_signal *, callback)</dt>
1588 <p>Initialises and configures the fork watcher - it has no parameters of any
1589 kind. There is a <code>ev_fork_set</code> macro, but using it is utterly pointless,
1599 <h1 id="OTHER_FUNCTIONS">OTHER FUNCTIONS</h1>
1600 <div id="OTHER_FUNCTIONS_CONTENT">
1601 <p>There are some other functions of possible interest. Described. Here. Now.</p>
1603 <dt>ev_once (loop, int fd, int events, ev_tstamp timeout, callback)</dt>
1605 <p>This function combines a simple timer and an I/O watcher, calls your
1606 callback on whichever event happens first and automatically stop both
1607 watchers. This is useful if you want to wait for a single event on an fd
1608 or timeout without having to allocate/configure/start/stop/free one or
1609 more watchers yourself.</p>
1610 <p>If <code>fd</code> is less than 0, then no I/O watcher will be started and events
1611 is being ignored. Otherwise, an <code>ev_io</code> watcher for the given <code>fd</code> and
1612 <code>events</code> set will be craeted and started.</p>
1613 <p>If <code>timeout</code> is less than 0, then no timeout watcher will be
1614 started. Otherwise an <code>ev_timer</code> watcher with after = <code>timeout</code> (and
1615 repeat = 0) will be started. While <code>0</code> is a valid timeout, it is of
1617 <p>The callback has the type <code>void (*cb)(int revents, void *arg)</code> and gets
1618 passed an <code>revents</code> set like normal event callbacks (a combination of
1619 <code>EV_ERROR</code>, <code>EV_READ</code>, <code>EV_WRITE</code> or <code>EV_TIMEOUT</code>) and the <code>arg</code>
1620 value passed to <code>ev_once</code>:</p>
1621 <pre> static void stdin_ready (int revents, void *arg)
1623 if (revents & EV_TIMEOUT)
1624 /* doh, nothing entered */;
1625 else if (revents & EV_READ)
1626 /* stdin might have data for us, joy! */;
1629 ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
1633 <dt>ev_feed_event (ev_loop *, watcher *, int revents)</dt>
1635 <p>Feeds the given event set into the event loop, as if the specified event
1636 had happened for the specified watcher (which must be a pointer to an
1637 initialised but not necessarily started event watcher).</p>
1639 <dt>ev_feed_fd_event (ev_loop *, int fd, int revents)</dt>
1641 <p>Feed an event on the given fd, as if a file descriptor backend detected
1642 the given events it.</p>
1644 <dt>ev_feed_signal_event (ev_loop *loop, int signum)</dt>
1646 <p>Feed an event as if the given signal occured (<code>loop</code> must be the default
1656 <h1 id="LIBEVENT_EMULATION">LIBEVENT EMULATION</h1>
1657 <div id="LIBEVENT_EMULATION_CONTENT">
1658 <p>Libev offers a compatibility emulation layer for libevent. It cannot
1659 emulate the internals of libevent, so here are some usage hints:</p>
1661 <dt>* Use it by including <event.h>, as usual.</dt>
1662 <dt>* The following members are fully supported: ev_base, ev_callback,
1663 ev_arg, ev_fd, ev_res, ev_events.</dt>
1664 <dt>* Avoid using ev_flags and the EVLIST_*-macros, while it is
1665 maintained by libev, it does not work exactly the same way as in libevent (consider
1666 it a private API).</dt>
1667 <dt>* Priorities are not currently supported. Initialising priorities
1668 will fail and all watchers will have the same priority, even though there
1669 is an ev_pri field.</dt>
1670 <dt>* Other members are not supported.</dt>
1671 <dt>* The libev emulation is <i>not</i> ABI compatible to libevent, you need
1672 to use the libev header file and library.</dt>
1676 <h1 id="C_SUPPORT">C++ SUPPORT</h1>
1677 <div id="C_SUPPORT_CONTENT">
1678 <p>Libev comes with some simplistic wrapper classes for C++ that mainly allow
1679 you to use some convinience methods to start/stop watchers and also change
1680 the callback model to a model using method callbacks on objects.</p>
1682 <pre> #include <ev++.h>
1685 <p>(it is not installed by default). This automatically includes <cite>ev.h</cite>
1686 and puts all of its definitions (many of them macros) into the global
1687 namespace. All C++ specific things are put into the <code>ev</code> namespace.</p>
1688 <p>It should support all the same embedding options as <cite>ev.h</cite>, most notably
1689 <code>EV_MULTIPLICITY</code>.</p>
1690 <p>Here is a list of things available in the <code>ev</code> namespace:</p>
1692 <dt><code>ev::READ</code>, <code>ev::WRITE</code> etc.</dt>
1694 <p>These are just enum values with the same values as the <code>EV_READ</code> etc.
1695 macros from <cite>ev.h</cite>.</p>
1697 <dt><code>ev::tstamp</code>, <code>ev::now</code></dt>
1699 <p>Aliases to the same types/functions as with the <code>ev_</code> prefix.</p>
1701 <dt><code>ev::io</code>, <code>ev::timer</code>, <code>ev::periodic</code>, <code>ev::idle</code>, <code>ev::sig</code> etc.</dt>
1703 <p>For each <code>ev_TYPE</code> watcher in <cite>ev.h</cite> there is a corresponding class of
1704 the same name in the <code>ev</code> namespace, with the exception of <code>ev_signal</code>
1705 which is called <code>ev::sig</code> to avoid clashes with the <code>signal</code> macro
1706 defines by many implementations.</p>
1707 <p>All of those classes have these methods:</p>
1710 <dt>ev::TYPE::TYPE (object *, object::method *)</dt>
1711 <dt>ev::TYPE::TYPE (object *, object::method *, struct ev_loop *)</dt>
1712 <dt>ev::TYPE::~TYPE</dt>
1714 <p>The constructor takes a pointer to an object and a method pointer to
1715 the event handler callback to call in this class. The constructor calls
1716 <code>ev_init</code> for you, which means you have to call the <code>set</code> method
1717 before starting it. If you do not specify a loop then the constructor
1718 automatically associates the default loop with this watcher.</p>
1719 <p>The destructor automatically stops the watcher if it is active.</p>
1721 <dt>w->set (struct ev_loop *)</dt>
1723 <p>Associates a different <code>struct ev_loop</code> with this watcher. You can only
1724 do this when the watcher is inactive (and not pending either).</p>
1726 <dt>w->set ([args])</dt>
1728 <p>Basically the same as <code>ev_TYPE_set</code>, with the same args. Must be
1729 called at least once. Unlike the C counterpart, an active watcher gets
1730 automatically stopped and restarted.</p>
1732 <dt>w->start ()</dt>
1734 <p>Starts the watcher. Note that there is no <code>loop</code> argument as the
1735 constructor already takes the loop.</p>
1737 <dt>w->stop ()</dt>
1739 <p>Stops the watcher if it is active. Again, no <code>loop</code> argument.</p>
1741 <dt>w->again () <code>ev::timer</code>, <code>ev::periodic</code> only</dt>
1743 <p>For <code>ev::timer</code> and <code>ev::periodic</code>, this invokes the corresponding
1744 <code>ev_TYPE_again</code> function.</p>
1746 <dt>w->sweep () <code>ev::embed</code> only</dt>
1748 <p>Invokes <code>ev_embed_sweep</code>.</p>
1750 <dt>w->update () <code>ev::stat</code> only</dt>
1752 <p>Invokes <code>ev_stat_stat</code>.</p>
1758 <p>Example: Define a class with an IO and idle watcher, start one of them in
1759 the constructor.</p>
1762 ev_io io; void io_cb (ev::io &w, int revents);
1763 ev_idle idle void idle_cb (ev::idle &w, int revents);
1768 myclass::myclass (int fd)
1769 : io (this, &myclass::io_cb),
1770 idle (this, &myclass::idle_cb)
1772 io.start (fd, ev::READ);
1781 <h1 id="MACRO_MAGIC">MACRO MAGIC</h1>
1782 <div id="MACRO_MAGIC_CONTENT">
1783 <p>Libev can be compiled with a variety of options, the most fundemantal is
1784 <code>EV_MULTIPLICITY</code>. This option determines wether (most) functions and
1785 callbacks have an initial <code>struct ev_loop *</code> argument.</p>
1786 <p>To make it easier to write programs that cope with either variant, the
1787 following macros are defined:</p>
1789 <dt><code>EV_A</code>, <code>EV_A_</code></dt>
1791 <p>This provides the loop <i>argument</i> for functions, if one is required ("ev
1792 loop argument"). The <code>EV_A</code> form is used when this is the sole argument,
1793 <code>EV_A_</code> is used when other arguments are following. Example:</p>
1794 <pre> ev_unref (EV_A);
1795 ev_timer_add (EV_A_ watcher);
1799 <p>It assumes the variable <code>loop</code> of type <code>struct ev_loop *</code> is in scope,
1800 which is often provided by the following macro.</p>
1802 <dt><code>EV_P</code>, <code>EV_P_</code></dt>
1804 <p>This provides the loop <i>parameter</i> for functions, if one is required ("ev
1805 loop parameter"). The <code>EV_P</code> form is used when this is the sole parameter,
1806 <code>EV_P_</code> is used when other parameters are following. Example:</p>
1807 <pre> // this is how ev_unref is being declared
1808 static void ev_unref (EV_P);
1810 // this is how you can declare your typical callback
1811 static void cb (EV_P_ ev_timer *w, int revents)
1814 <p>It declares a parameter <code>loop</code> of type <code>struct ev_loop *</code>, quite
1815 suitable for use with <code>EV_A</code>.</p>
1817 <dt><code>EV_DEFAULT</code>, <code>EV_DEFAULT_</code></dt>
1819 <p>Similar to the other two macros, this gives you the value of the default
1820 loop, if multiple loops are supported ("ev loop default").</p>
1823 <p>Example: Declare and initialise a check watcher, working regardless of
1824 wether multiple loops are supported or not.</p>
1826 check_cb (EV_P_ ev_timer *w, int revents)
1828 ev_check_stop (EV_A_ w);
1832 ev_check_init (&check, check_cb);
1833 ev_check_start (EV_DEFAULT_ &check);
1834 ev_loop (EV_DEFAULT_ 0);
1842 <h1 id="EMBEDDING">EMBEDDING</h1>
1843 <div id="EMBEDDING_CONTENT">
1844 <p>Libev can (and often is) directly embedded into host
1845 applications. Examples of applications that embed it include the Deliantra
1846 Game Server, the EV perl module, the GNU Virtual Private Ethernet (gvpe)
1847 and rxvt-unicode.</p>
1848 <p>The goal is to enable you to just copy the neecssary files into your
1849 source directory without having to change even a single line in them, so
1850 you can easily upgrade by simply copying (or having a checked-out copy of
1851 libev somewhere in your source tree).</p>
1854 <h2 id="FILESETS">FILESETS</h2>
1855 <div id="FILESETS_CONTENT">
1856 <p>Depending on what features you need you need to include one or more sets of files
1860 <h3 id="CORE_EVENT_LOOP">CORE EVENT LOOP</h3>
1861 <div id="CORE_EVENT_LOOP_CONTENT">
1862 <p>To include only the libev core (all the <code>ev_*</code> functions), with manual
1863 configuration (no autoconf):</p>
1864 <pre> #define EV_STANDALONE 1
1865 #include "ev.c"
1868 <p>This will automatically include <cite>ev.h</cite>, too, and should be done in a
1869 single C source file only to provide the function implementations. To use
1870 it, do the same for <cite>ev.h</cite> in all files wishing to use this API (best
1871 done by writing a wrapper around <cite>ev.h</cite> that you can include instead and
1872 where you can put other configuration options):</p>
1873 <pre> #define EV_STANDALONE 1
1874 #include "ev.h"
1877 <p>Both header files and implementation files can be compiled with a C++
1878 compiler (at least, thats a stated goal, and breakage will be treated
1880 <p>You need the following files in your source tree, or in a directory
1881 in your include path (e.g. in libev/ when using -Ilibev):</p>
1887 ev_win32.c required on win32 platforms only
1889 ev_select.c only when select backend is enabled (which is by default)
1890 ev_poll.c only when poll backend is enabled (disabled by default)
1891 ev_epoll.c only when the epoll backend is enabled (disabled by default)
1892 ev_kqueue.c only when the kqueue backend is enabled (disabled by default)
1893 ev_port.c only when the solaris port backend is enabled (disabled by default)
1896 <p><cite>ev.c</cite> includes the backend files directly when enabled, so you only need
1897 to compile this single file.</p>
1900 <h3 id="LIBEVENT_COMPATIBILITY_API">LIBEVENT COMPATIBILITY API</h3>
1901 <div id="LIBEVENT_COMPATIBILITY_API_CONTENT">
1902 <p>To include the libevent compatibility API, also include:</p>
1903 <pre> #include "event.c"
1906 <p>in the file including <cite>ev.c</cite>, and:</p>
1907 <pre> #include "event.h"
1910 <p>in the files that want to use the libevent API. This also includes <cite>ev.h</cite>.</p>
1911 <p>You need the following additional files for this:</p>
1918 <h3 id="AUTOCONF_SUPPORT">AUTOCONF SUPPORT</h3>
1919 <div id="AUTOCONF_SUPPORT_CONTENT">
1920 <p>Instead of using <code>EV_STANDALONE=1</code> and providing your config in
1921 whatever way you want, you can also <code>m4_include([libev.m4])</code> in your
1922 <cite>configure.ac</cite> and leave <code>EV_STANDALONE</code> undefined. <cite>ev.c</cite> will then
1923 include <cite>config.h</cite> and configure itself accordingly.</p>
1924 <p>For this of course you need the m4 file:</p>
1930 <h2 id="PREPROCESSOR_SYMBOLS_MACROS">PREPROCESSOR SYMBOLS/MACROS</h2>
1931 <div id="PREPROCESSOR_SYMBOLS_MACROS_CONTENT">
1932 <p>Libev can be configured via a variety of preprocessor symbols you have to define
1933 before including any of its files. The default is not to build for multiplicity
1934 and only include the select backend.</p>
1936 <dt>EV_STANDALONE</dt>
1938 <p>Must always be <code>1</code> if you do not use autoconf configuration, which
1939 keeps libev from including <cite>config.h</cite>, and it also defines dummy
1940 implementations for some libevent functions (such as logging, which is not
1941 supported). It will also not define any of the structs usually found in
1942 <cite>event.h</cite> that are not directly supported by the libev core alone.</p>
1944 <dt>EV_USE_MONOTONIC</dt>
1946 <p>If defined to be <code>1</code>, libev will try to detect the availability of the
1947 monotonic clock option at both compiletime and runtime. Otherwise no use
1948 of the monotonic clock option will be attempted. If you enable this, you
1949 usually have to link against librt or something similar. Enabling it when
1950 the functionality isn't available is safe, though, althoguh you have
1951 to make sure you link against any libraries where the <code>clock_gettime</code>
1952 function is hiding in (often <cite>-lrt</cite>).</p>
1954 <dt>EV_USE_REALTIME</dt>
1956 <p>If defined to be <code>1</code>, libev will try to detect the availability of the
1957 realtime clock option at compiletime (and assume its availability at
1958 runtime if successful). Otherwise no use of the realtime clock option will
1959 be attempted. This effectively replaces <code>gettimeofday</code> by <code>clock_get
1960 (CLOCK_REALTIME, ...)</code> and will not normally affect correctness. See tzhe note about libraries
1961 in the description of <code>EV_USE_MONOTONIC</code>, though.</p>
1963 <dt>EV_USE_SELECT</dt>
1965 <p>If undefined or defined to be <code>1</code>, libev will compile in support for the
1966 <code>select</code>(2) backend. No attempt at autodetection will be done: if no
1967 other method takes over, select will be it. Otherwise the select backend
1968 will not be compiled in.</p>
1970 <dt>EV_SELECT_USE_FD_SET</dt>
1972 <p>If defined to <code>1</code>, then the select backend will use the system <code>fd_set</code>
1973 structure. This is useful if libev doesn't compile due to a missing
1974 <code>NFDBITS</code> or <code>fd_mask</code> definition or it misguesses the bitset layout on
1975 exotic systems. This usually limits the range of file descriptors to some
1976 low limit such as 1024 or might have other limitations (winsocket only
1977 allows 64 sockets). The <code>FD_SETSIZE</code> macro, set before compilation, might
1978 influence the size of the <code>fd_set</code> used.</p>
1980 <dt>EV_SELECT_IS_WINSOCKET</dt>
1982 <p>When defined to <code>1</code>, the select backend will assume that
1983 select/socket/connect etc. don't understand file descriptors but
1984 wants osf handles on win32 (this is the case when the select to
1985 be used is the winsock select). This means that it will call
1986 <code>_get_osfhandle</code> on the fd to convert it to an OS handle. Otherwise,
1987 it is assumed that all these functions actually work on fds, even
1988 on win32. Should not be defined on non-win32 platforms.</p>
1990 <dt>EV_USE_POLL</dt>
1992 <p>If defined to be <code>1</code>, libev will compile in support for the <code>poll</code>(2)
1993 backend. Otherwise it will be enabled on non-win32 platforms. It
1994 takes precedence over select.</p>
1996 <dt>EV_USE_EPOLL</dt>
1998 <p>If defined to be <code>1</code>, libev will compile in support for the Linux
1999 <code>epoll</code>(7) backend. Its availability will be detected at runtime,
2000 otherwise another method will be used as fallback. This is the
2001 preferred backend for GNU/Linux systems.</p>
2003 <dt>EV_USE_KQUEUE</dt>
2005 <p>If defined to be <code>1</code>, libev will compile in support for the BSD style
2006 <code>kqueue</code>(2) backend. Its actual availability will be detected at runtime,
2007 otherwise another method will be used as fallback. This is the preferred
2008 backend for BSD and BSD-like systems, although on most BSDs kqueue only
2009 supports some types of fds correctly (the only platform we found that
2010 supports ptys for example was NetBSD), so kqueue might be compiled in, but
2011 not be used unless explicitly requested. The best way to use it is to find
2012 out whether kqueue supports your type of fd properly and use an embedded
2015 <dt>EV_USE_PORT</dt>
2017 <p>If defined to be <code>1</code>, libev will compile in support for the Solaris
2018 10 port style backend. Its availability will be detected at runtime,
2019 otherwise another method will be used as fallback. This is the preferred
2020 backend for Solaris 10 systems.</p>
2022 <dt>EV_USE_DEVPOLL</dt>
2024 <p>reserved for future expansion, works like the USE symbols above.</p>
2026 <dt>EV_USE_INOTIFY</dt>
2028 <p>If defined to be <code>1</code>, libev will compile in support for the Linux inotify
2029 interface to speed up <code>ev_stat</code> watchers. Its actual availability will
2030 be detected at runtime.</p>
2034 <p>The name of the <cite>ev.h</cite> header file used to include it. The default if
2035 undefined is <code><ev.h></code> in <cite>event.h</cite> and <code>"ev.h"</code> in <cite>ev.c</cite>. This
2036 can be used to virtually rename the <cite>ev.h</cite> header file in case of conflicts.</p>
2038 <dt>EV_CONFIG_H</dt>
2040 <p>If <code>EV_STANDALONE</code> isn't <code>1</code>, this variable can be used to override
2041 <cite>ev.c</cite>'s idea of where to find the <cite>config.h</cite> file, similarly to
2042 <code>EV_H</code>, above.</p>
2046 <p>Similarly to <code>EV_H</code>, this macro can be used to override <cite>event.c</cite>'s idea
2047 of how the <cite>event.h</cite> header can be found.</p>
2049 <dt>EV_PROTOTYPES</dt>
2051 <p>If defined to be <code>0</code>, then <cite>ev.h</cite> will not define any function
2052 prototypes, but still define all the structs and other symbols. This is
2053 occasionally useful if you want to provide your own wrapper functions
2054 around libev functions.</p>
2056 <dt>EV_MULTIPLICITY</dt>
2058 <p>If undefined or defined to <code>1</code>, then all event-loop-specific functions
2059 will have the <code>struct ev_loop *</code> as first argument, and you can create
2060 additional independent event loops. Otherwise there will be no support
2061 for multiple event loops and there is no first event loop pointer
2062 argument. Instead, all functions act on the single default loop.</p>
2064 <dt>EV_PERIODIC_ENABLE</dt>
2066 <p>If undefined or defined to be <code>1</code>, then periodic timers are supported. If
2067 defined to be <code>0</code>, then they are not. Disabling them saves a few kB of
2070 <dt>EV_EMBED_ENABLE</dt>
2072 <p>If undefined or defined to be <code>1</code>, then embed watchers are supported. If
2073 defined to be <code>0</code>, then they are not.</p>
2075 <dt>EV_STAT_ENABLE</dt>
2077 <p>If undefined or defined to be <code>1</code>, then stat watchers are supported. If
2078 defined to be <code>0</code>, then they are not.</p>
2080 <dt>EV_FORK_ENABLE</dt>
2082 <p>If undefined or defined to be <code>1</code>, then fork watchers are supported. If
2083 defined to be <code>0</code>, then they are not.</p>
2087 <p>If you need to shave off some kilobytes of code at the expense of some
2088 speed, define this symbol to <code>1</code>. Currently only used for gcc to override
2089 some inlining decisions, saves roughly 30% codesize of amd64.</p>
2091 <dt>EV_PID_HASHSIZE</dt>
2093 <p><code>ev_child</code> watchers use a small hash table to distribute workload by
2094 pid. The default size is <code>16</code> (or <code>1</code> with <code>EV_MINIMAL</code>), usually more
2095 than enough. If you need to manage thousands of children you might want to
2096 increase this value (<i>must</i> be a power of two).</p>
2098 <dt>EV_INOTIFY_HASHSIZE</dt>
2100 <p><code>ev_staz</code> watchers use a small hash table to distribute workload by
2101 inotify watch id. The default size is <code>16</code> (or <code>1</code> with <code>EV_MINIMAL</code>),
2102 usually more than enough. If you need to manage thousands of <code>ev_stat</code>
2103 watchers you might want to increase this value (<i>must</i> be a power of
2108 <p>By default, all watchers have a <code>void *data</code> member. By redefining
2109 this macro to a something else you can include more and other types of
2110 members. You have to define it each time you include one of the files,
2111 though, and it must be identical each time.</p>
2112 <p>For example, the perl EV module uses something like this:</p>
2113 <pre> #define EV_COMMON \
2114 SV *self; /* contains this struct */ \
2115 SV *cb_sv, *fh /* note no trailing ";" */
2119 <dt>EV_CB_DECLARE (type)</dt>
2120 <dt>EV_CB_INVOKE (watcher, revents)</dt>
2121 <dt>ev_set_cb (ev, cb)</dt>
2123 <p>Can be used to change the callback member declaration in each watcher,
2124 and the way callbacks are invoked and set. Must expand to a struct member
2125 definition and a statement, respectively. See the <cite>ev.v</cite> header file for
2126 their default definitions. One possible use for overriding these is to
2127 avoid the <code>struct ev_loop *</code> as first argument in all cases, or to use
2128 method calls instead of plain function calls in C++.</p>
2131 <h2 id="EXAMPLES">EXAMPLES</h2>
2132 <div id="EXAMPLES_CONTENT">
2133 <p>For a real-world example of a program the includes libev
2134 verbatim, you can have a look at the EV perl module
2135 (<a href="http://software.schmorp.de/pkg/EV.html">http://software.schmorp.de/pkg/EV.html</a>). It has the libev files in
2136 the <cite>libev/</cite> subdirectory and includes them in the <cite>EV/EVAPI.h</cite> (public
2137 interface) and <cite>EV.xs</cite> (implementation) files. Only the <cite>EV.xs</cite> file
2138 will be compiled. It is pretty complex because it provides its own header
2140 <p>The usage in rxvt-unicode is simpler. It has a <cite>ev_cpp.h</cite> header file
2141 that everybody includes and which overrides some autoconf choices:</p>
2142 <pre> #define EV_USE_POLL 0
2143 #define EV_MULTIPLICITY 0
2144 #define EV_PERIODICS 0
2145 #define EV_CONFIG_H <config.h>
2147 #include "ev++.h"
2150 <p>And a <cite>ev_cpp.C</cite> implementation file that contains libev proper and is compiled:</p>
2151 <pre> #include "ev_cpp.h"
2152 #include "ev.c"
2160 <h1 id="COMPLEXITIES">COMPLEXITIES</h1>
2161 <div id="COMPLEXITIES_CONTENT">
2162 <p>In this section the complexities of (many of) the algorithms used inside
2163 libev will be explained. For complexity discussions about backends see the
2164 documentation for <code>ev_default_init</code>.</p>
2167 <dt>Starting and stopping timer/periodic watchers: O(log skipped_other_timers)</dt>
2168 <dt>Changing timer/periodic watchers (by autorepeat, again): O(log skipped_other_timers)</dt>
2169 <dt>Starting io/check/prepare/idle/signal/child watchers: O(1)</dt>
2170 <dt>Stopping check/prepare/idle watchers: O(1)</dt>
2171 <dt>Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))</dt>
2172 <dt>Finding the next timer per loop iteration: O(1)</dt>
2173 <dt>Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)</dt>
2174 <dt>Activating one watcher: O(1)</dt>
2183 <h1 id="AUTHOR">AUTHOR</h1>
2184 <div id="AUTHOR_CONTENT">
2185 <p>Marc Lehmann <libev@schmorp.de>.</p>
projects / software / libev.git / blob