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="Sat Nov 24 05:58:35 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="#DESCRIPTION">DESCRIPTION</a></li>
20 <li><a href="#FEATURES">FEATURES</a></li>
21 <li><a href="#CONVENTIONS">CONVENTIONS</a></li>
22 <li><a href="#TIME_REPRESENTATION">TIME REPRESENTATION</a></li>
23 <li><a href="#GLOBAL_FUNCTIONS">GLOBAL FUNCTIONS</a></li>
24 <li><a href="#FUNCTIONS_CONTROLLING_THE_EVENT_LOOP">FUNCTIONS CONTROLLING THE EVENT LOOP</a></li>
25 <li><a href="#ANATOMY_OF_A_WATCHER">ANATOMY OF A WATCHER</a>
26 <ul><li><a href="#ASSOCIATING_CUSTOM_DATA_WITH_A_WATCH">ASSOCIATING CUSTOM DATA WITH A WATCHER</a></li>
29 <li><a href="#WATCHER_TYPES">WATCHER TYPES</a>
30 <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>
31 <li><a href="#code_ev_timer_code_relative_and_opti"><code>ev_timer</code> - relative and optionally recurring timeouts</a></li>
32 <li><a href="#code_ev_periodic_code_to_cron_or_not"><code>ev_periodic</code> - to cron or not to cron</a></li>
33 <li><a href="#code_ev_signal_code_signal_me_when_a"><code>ev_signal</code> - signal me when a signal gets signalled</a></li>
34 <li><a href="#code_ev_child_code_wait_for_pid_stat"><code>ev_child</code> - wait for pid status changes</a></li>
35 <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>
36 <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>
37 <li><a href="#code_ev_embed_code_when_one_backend_"><code>ev_embed</code> - when one backend isn't enough</a></li>
40 <li><a href="#OTHER_FUNCTIONS">OTHER FUNCTIONS</a></li>
41 <li><a href="#LIBEVENT_EMULATION">LIBEVENT EMULATION</a></li>
42 <li><a href="#C_SUPPORT">C++ SUPPORT</a></li>
43 <li><a href="#AUTHOR">AUTHOR</a>
48 <h1 id="NAME">NAME</h1><p><a href="#TOP" class="toplink">Top</a></p>
49 <div id="NAME_CONTENT">
50 <p>libev - a high performance full-featured event loop written in C</p>
53 <h1 id="SYNOPSIS">SYNOPSIS</h1><p><a href="#TOP" class="toplink">Top</a></p>
54 <div id="SYNOPSIS_CONTENT">
55 <pre> #include <ev.h>
60 <h1 id="DESCRIPTION">DESCRIPTION</h1><p><a href="#TOP" class="toplink">Top</a></p>
61 <div id="DESCRIPTION_CONTENT">
62 <p>Libev is an event loop: you register interest in certain events (such as a
63 file descriptor being readable or a timeout occuring), and it will manage
64 these event sources and provide your program with events.</p>
65 <p>To do this, it must take more or less complete control over your process
66 (or thread) by executing the <i>event loop</i> handler, and will then
67 communicate events via a callback mechanism.</p>
68 <p>You register interest in certain events by registering so-called <i>event
69 watchers</i>, which are relatively small C structures you initialise with the
70 details of the event, and then hand it over to libev by <i>starting</i> the
74 <h1 id="FEATURES">FEATURES</h1><p><a href="#TOP" class="toplink">Top</a></p>
75 <div id="FEATURES_CONTENT">
76 <p>Libev supports select, poll, the linux-specific epoll and the bsd-specific
77 kqueue mechanisms for file descriptor events, relative timers, absolute
78 timers with customised rescheduling, signal events, process status change
79 events (related to SIGCHLD), and event watchers dealing with the event
80 loop mechanism itself (idle, prepare and check watchers). It also is quite
81 fast (see this <a href="http://libev.schmorp.de/bench.html">benchmark</a> comparing
82 it to libevent for example).</p>
85 <h1 id="CONVENTIONS">CONVENTIONS</h1><p><a href="#TOP" class="toplink">Top</a></p>
86 <div id="CONVENTIONS_CONTENT">
87 <p>Libev is very configurable. In this manual the default configuration
88 will be described, which supports multiple event loops. For more info
89 about various configuration options please have a look at the file
90 <cite>README.embed</cite> in the libev distribution. If libev was configured without
91 support for multiple event loops, then all functions taking an initial
92 argument of name <code>loop</code> (which is always of type <code>struct ev_loop *</code>)
93 will not have this argument.</p>
96 <h1 id="TIME_REPRESENTATION">TIME REPRESENTATION</h1><p><a href="#TOP" class="toplink">Top</a></p>
97 <div id="TIME_REPRESENTATION_CONTENT">
98 <p>Libev represents time as a single floating point number, representing the
99 (fractional) number of seconds since the (POSIX) epoch (somewhere near
100 the beginning of 1970, details are complicated, don't ask). This type is
101 called <code>ev_tstamp</code>, which is what you should use too. It usually aliases
102 to the <code>double</code> type in C, and when you need to do any calculations on
103 it, you should treat it as such.</p>
110 <h1 id="GLOBAL_FUNCTIONS">GLOBAL FUNCTIONS</h1><p><a href="#TOP" class="toplink">Top</a></p>
111 <div id="GLOBAL_FUNCTIONS_CONTENT">
112 <p>These functions can be called anytime, even before initialising the
113 library in any way.</p>
115 <dt>ev_tstamp ev_time ()</dt>
117 <p>Returns the current time as libev would use it. Please note that the
118 <code>ev_now</code> function is usually faster and also often returns the timestamp
119 you actually want to know.</p>
121 <dt>int ev_version_major ()</dt>
122 <dt>int ev_version_minor ()</dt>
124 <p>You can find out the major and minor version numbers of the library
125 you linked against by calling the functions <code>ev_version_major</code> and
126 <code>ev_version_minor</code>. If you want, you can compare against the global
127 symbols <code>EV_VERSION_MAJOR</code> and <code>EV_VERSION_MINOR</code>, which specify the
128 version of the library your program was compiled against.</p>
129 <p>Usually, it's a good idea to terminate if the major versions mismatch,
130 as this indicates an incompatible change. Minor versions are usually
131 compatible to older versions, so a larger minor version alone is usually
133 <p>Example: make sure we haven't accidentally been linked against the wrong
135 <pre> assert (("libev version mismatch",
136 ev_version_major () == EV_VERSION_MAJOR
137 && ev_version_minor () >= EV_VERSION_MINOR));
141 <dt>unsigned int ev_supported_backends ()</dt>
143 <p>Return the set of all backends (i.e. their corresponding <code>EV_BACKEND_*</code>
144 value) compiled into this binary of libev (independent of their
145 availability on the system you are running on). See <code>ev_default_loop</code> for
146 a description of the set values.</p>
147 <p>Example: make sure we have the epoll method, because yeah this is cool and
148 a must have and can we have a torrent of it please!!!11</p>
149 <pre> assert (("sorry, no epoll, no sex",
150 ev_supported_backends () & EVBACKEND_EPOLL));
154 <dt>unsigned int ev_recommended_backends ()</dt>
156 <p>Return the set of all backends compiled into this binary of libev and also
157 recommended for this platform. This set is often smaller than the one
158 returned by <code>ev_supported_backends</code>, as for example kqueue is broken on
159 most BSDs and will not be autodetected unless you explicitly request it
160 (assuming you know what you are doing). This is the set of backends that
161 libev will probe for if you specify no backends explicitly.</p>
163 <dt>unsigned int ev_embeddable_backends ()</dt>
165 <p>Returns the set of backends that are embeddable in other event loops. This
166 is the theoretical, all-platform, value. To find which backends
167 might be supported on the current system, you would need to look at
168 <code>ev_embeddable_backends () & ev_supported_backends ()</code>, likewise for
169 recommended ones.</p>
170 <p>See the description of <code>ev_embed</code> watchers for more info.</p>
172 <dt>ev_set_allocator (void *(*cb)(void *ptr, long size))</dt>
174 <p>Sets the allocation function to use (the prototype is similar to the
175 realloc C function, the semantics are identical). It is used to allocate
176 and free memory (no surprises here). If it returns zero when memory
177 needs to be allocated, the library might abort or take some potentially
178 destructive action. The default is your system realloc function.</p>
179 <p>You could override this function in high-availability programs to, say,
180 free some memory if it cannot allocate memory, to use a special allocator,
181 or even to sleep a while and retry until some memory is available.</p>
182 <p>Example: replace the libev allocator with one that waits a bit and then
183 retries: better than mine).</p>
185 persistent_realloc (void *ptr, long size)
189 void *newptr = realloc (ptr, size);
199 ev_set_allocator (persistent_realloc);
203 <dt>ev_set_syserr_cb (void (*cb)(const char *msg));</dt>
205 <p>Set the callback function to call on a retryable syscall error (such
206 as failed select, poll, epoll_wait). The message is a printable string
207 indicating the system call or subsystem causing the problem. If this
208 callback is set, then libev will expect it to remedy the sitution, no
209 matter what, when it returns. That is, libev will generally retry the
210 requested operation, or, if the condition doesn't go away, do bad stuff
212 <p>Example: do the same thing as libev does internally:</p>
214 fatal_error (const char *msg)
221 ev_set_syserr_cb (fatal_error);
228 <h1 id="FUNCTIONS_CONTROLLING_THE_EVENT_LOOP">FUNCTIONS CONTROLLING THE EVENT LOOP</h1><p><a href="#TOP" class="toplink">Top</a></p>
229 <div id="FUNCTIONS_CONTROLLING_THE_EVENT_LOOP-2">
230 <p>An event loop is described by a <code>struct ev_loop *</code>. The library knows two
231 types of such loops, the <i>default</i> loop, which supports signals and child
232 events, and dynamically created loops which do not.</p>
233 <p>If you use threads, a common model is to run the default event loop
234 in your main thread (or in a separate thread) and for each thread you
235 create, you also create another event loop. Libev itself does no locking
236 whatsoever, so if you mix calls to the same event loop in different
237 threads, make sure you lock (this is usually a bad idea, though, even if
238 done correctly, because it's hideous and inefficient).</p>
240 <dt>struct ev_loop *ev_default_loop (unsigned int flags)</dt>
242 <p>This will initialise the default event loop if it hasn't been initialised
243 yet and return it. If the default loop could not be initialised, returns
244 false. If it already was initialised it simply returns it (and ignores the
245 flags. If that is troubling you, check <code>ev_backend ()</code> afterwards).</p>
246 <p>If you don't know what event loop to use, use the one returned from this
248 <p>The flags argument can be used to specify special behaviour or specific
249 backends to use, and is usually specified as <code>0</code> (or <code>EVFLAG_AUTO</code>).</p>
250 <p>The following flags are supported:</p>
253 <dt><code>EVFLAG_AUTO</code></dt>
255 <p>The default flags value. Use this if you have no clue (it's the right
256 thing, believe me).</p>
258 <dt><code>EVFLAG_NOENV</code></dt>
260 <p>If this flag bit is ored into the flag value (or the program runs setuid
261 or setgid) then libev will <i>not</i> look at the environment variable
262 <code>LIBEV_FLAGS</code>. Otherwise (the default), this environment variable will
263 override the flags completely if it is found in the environment. This is
264 useful to try out specific backends to test their performance, or to work
267 <dt><code>EVBACKEND_SELECT</code> (value 1, portable select backend)</dt>
269 <p>This is your standard select(2) backend. Not <i>completely</i> standard, as
270 libev tries to roll its own fd_set with no limits on the number of fds,
271 but if that fails, expect a fairly low limit on the number of fds when
272 using this backend. It doesn't scale too well (O(highest_fd)), but its usually
273 the fastest backend for a low number of fds.</p>
275 <dt><code>EVBACKEND_POLL</code> (value 2, poll backend, available everywhere except on windows)</dt>
277 <p>And this is your standard poll(2) backend. It's more complicated than
278 select, but handles sparse fds better and has no artificial limit on the
279 number of fds you can use (except it will slow down considerably with a
280 lot of inactive fds). It scales similarly to select, i.e. O(total_fds).</p>
282 <dt><code>EVBACKEND_EPOLL</code> (value 4, Linux)</dt>
284 <p>For few fds, this backend is a bit little slower than poll and select,
285 but it scales phenomenally better. While poll and select usually scale like
286 O(total_fds) where n is the total number of fds (or the highest fd), epoll scales
287 either O(1) or O(active_fds).</p>
288 <p>While stopping and starting an I/O watcher in the same iteration will
289 result in some caching, there is still a syscall per such incident
290 (because the fd could point to a different file description now), so its
291 best to avoid that. Also, dup()ed file descriptors might not work very
292 well if you register events for both fds.</p>
293 <p>Please note that epoll sometimes generates spurious notifications, so you
294 need to use non-blocking I/O or other means to avoid blocking when no data
295 (or space) is available.</p>
297 <dt><code>EVBACKEND_KQUEUE</code> (value 8, most BSD clones)</dt>
299 <p>Kqueue deserves special mention, as at the time of this writing, it
300 was broken on all BSDs except NetBSD (usually it doesn't work with
301 anything but sockets and pipes, except on Darwin, where of course its
302 completely useless). For this reason its not being "autodetected"
303 unless you explicitly specify it explicitly in the flags (i.e. using
304 <code>EVBACKEND_KQUEUE</code>).</p>
305 <p>It scales in the same way as the epoll backend, but the interface to the
306 kernel is more efficient (which says nothing about its actual speed, of
307 course). While starting and stopping an I/O watcher does not cause an
308 extra syscall as with epoll, it still adds up to four event changes per
309 incident, so its best to avoid that.</p>
311 <dt><code>EVBACKEND_DEVPOLL</code> (value 16, Solaris 8)</dt>
313 <p>This is not implemented yet (and might never be).</p>
315 <dt><code>EVBACKEND_PORT</code> (value 32, Solaris 10)</dt>
317 <p>This uses the Solaris 10 port mechanism. As with everything on Solaris,
318 it's really slow, but it still scales very well (O(active_fds)).</p>
319 <p>Please note that solaris ports can result in a lot of spurious
320 notifications, so you need to use non-blocking I/O or other means to avoid
321 blocking when no data (or space) is available.</p>
323 <dt><code>EVBACKEND_ALL</code></dt>
325 <p>Try all backends (even potentially broken ones that wouldn't be tried
326 with <code>EVFLAG_AUTO</code>). Since this is a mask, you can do stuff such as
327 <code>EVBACKEND_ALL & ~EVBACKEND_KQUEUE</code>.</p>
331 <p>If one or more of these are ored into the flags value, then only these
332 backends will be tried (in the reverse order as given here). If none are
333 specified, most compiled-in backend will be tried, usually in reverse
334 order of their flag values :)</p>
335 <p>The most typical usage is like this:</p>
336 <pre> if (!ev_default_loop (0))
337 fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
340 <p>Restrict libev to the select and poll backends, and do not allow
341 environment settings to be taken into account:</p>
342 <pre> ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);
345 <p>Use whatever libev has to offer, but make sure that kqueue is used if
346 available (warning, breaks stuff, best use only with your own private
347 event loop and only if you know the OS supports your types of fds):</p>
348 <pre> ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);
352 <dt>struct ev_loop *ev_loop_new (unsigned int flags)</dt>
354 <p>Similar to <code>ev_default_loop</code>, but always creates a new event loop that is
355 always distinct from the default loop. Unlike the default loop, it cannot
356 handle signal and child watchers, and attempts to do so will be greeted by
357 undefined behaviour (or a failed assertion if assertions are enabled).</p>
358 <p>Example: try to create a event loop that uses epoll and nothing else.</p>
359 <pre> struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);
361 fatal ("no epoll found here, maybe it hides under your chair");
365 <dt>ev_default_destroy ()</dt>
367 <p>Destroys the default loop again (frees all memory and kernel state
368 etc.). This stops all registered event watchers (by not touching them in
369 any way whatsoever, although you cannot rely on this :).</p>
371 <dt>ev_loop_destroy (loop)</dt>
373 <p>Like <code>ev_default_destroy</code>, but destroys an event loop created by an
374 earlier call to <code>ev_loop_new</code>.</p>
376 <dt>ev_default_fork ()</dt>
378 <p>This function reinitialises the kernel state for backends that have
379 one. Despite the name, you can call it anytime, but it makes most sense
380 after forking, in either the parent or child process (or both, but that
381 again makes little sense).</p>
382 <p>You <i>must</i> call this function in the child process after forking if and
383 only if you want to use the event library in both processes. If you just
384 fork+exec, you don't have to call it.</p>
385 <p>The function itself is quite fast and it's usually not a problem to call
386 it just in case after a fork. To make this easy, the function will fit in
387 quite nicely into a call to <code>pthread_atfork</code>:</p>
388 <pre> pthread_atfork (0, 0, ev_default_fork);
391 <p>At the moment, <code>EVBACKEND_SELECT</code> and <code>EVBACKEND_POLL</code> are safe to use
392 without calling this function, so if you force one of those backends you
393 do not need to care.</p>
395 <dt>ev_loop_fork (loop)</dt>
397 <p>Like <code>ev_default_fork</code>, but acts on an event loop created by
398 <code>ev_loop_new</code>. Yes, you have to call this on every allocated event loop
399 after fork, and how you do this is entirely your own problem.</p>
401 <dt>unsigned int ev_backend (loop)</dt>
403 <p>Returns one of the <code>EVBACKEND_*</code> flags indicating the event backend in
406 <dt>ev_tstamp ev_now (loop)</dt>
408 <p>Returns the current "event loop time", which is the time the event loop
409 received events and started processing them. This timestamp does not
410 change as long as callbacks are being processed, and this is also the base
411 time used for relative timers. You can treat it as the timestamp of the
412 event occuring (or more correctly, libev finding out about it).</p>
414 <dt>ev_loop (loop, int flags)</dt>
416 <p>Finally, this is it, the event handler. This function usually is called
417 after you initialised all your watchers and you want to start handling
419 <p>If the flags argument is specified as <code>0</code>, it will not return until
420 either no event watchers are active anymore or <code>ev_unloop</code> was called.</p>
421 <p>Please note that an explicit <code>ev_unloop</code> is usually better than
422 relying on all watchers to be stopped when deciding when a program has
423 finished (especially in interactive programs), but having a program that
424 automatically loops as long as it has to and no longer by virtue of
425 relying on its watchers stopping correctly is a thing of beauty.</p>
426 <p>A flags value of <code>EVLOOP_NONBLOCK</code> will look for new events, will handle
427 those events and any outstanding ones, but will not block your process in
428 case there are no events and will return after one iteration of the loop.</p>
429 <p>A flags value of <code>EVLOOP_ONESHOT</code> will look for new events (waiting if
430 neccessary) and will handle those and any outstanding ones. It will block
431 your process until at least one new event arrives, and will return after
432 one iteration of the loop. This is useful if you are waiting for some
433 external event in conjunction with something not expressible using other
434 libev watchers. However, a pair of <code>ev_prepare</code>/<code>ev_check</code> watchers is
435 usually a better approach for this kind of thing.</p>
436 <p>Here are the gory details of what <code>ev_loop</code> does:</p>
437 <pre> * If there are no active watchers (reference count is zero), return.
438 - Queue prepare watchers and then call all outstanding watchers.
439 - If we have been forked, recreate the kernel state.
440 - Update the kernel state with all outstanding changes.
441 - Update the "event loop time".
442 - Calculate for how long to block.
443 - Block the process, waiting for any events.
444 - Queue all outstanding I/O (fd) events.
445 - Update the "event loop time" and do time jump handling.
446 - Queue all outstanding timers.
447 - Queue all outstanding periodics.
448 - If no events are pending now, queue all idle watchers.
449 - Queue all check watchers.
450 - Call all queued watchers in reverse order (i.e. check watchers first).
451 Signals and child watchers are implemented as I/O watchers, and will
452 be handled here by queueing them when their watcher gets executed.
453 - If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK
454 were used, return, otherwise continue with step *.
457 <p>Example: queue some jobs and then loop until no events are outsanding
459 <pre> ... queue jobs here, make sure they register event watchers as long
460 ... as they still have work to do (even an idle watcher will do..)
461 ev_loop (my_loop, 0);
466 <dt>ev_unloop (loop, how)</dt>
468 <p>Can be used to make a call to <code>ev_loop</code> return early (but only after it
469 has processed all outstanding events). The <code>how</code> argument must be either
470 <code>EVUNLOOP_ONE</code>, which will make the innermost <code>ev_loop</code> call return, or
471 <code>EVUNLOOP_ALL</code>, which will make all nested <code>ev_loop</code> calls return.</p>
473 <dt>ev_ref (loop)</dt>
474 <dt>ev_unref (loop)</dt>
476 <p>Ref/unref can be used to add or remove a reference count on the event
477 loop: Every watcher keeps one reference, and as long as the reference
478 count is nonzero, <code>ev_loop</code> will not return on its own. If you have
479 a watcher you never unregister that should not keep <code>ev_loop</code> from
480 returning, ev_unref() after starting, and ev_ref() before stopping it. For
481 example, libev itself uses this for its internal signal pipe: It is not
482 visible to the libev user and should not keep <code>ev_loop</code> from exiting if
483 no event watchers registered by it are active. It is also an excellent
484 way to do this for generic recurring timers or from within third-party
485 libraries. Just remember to <i>unref after start</i> and <i>ref before stop</i>.</p>
486 <p>Example: create a signal watcher, but keep it from keeping <code>ev_loop</code>
487 running when nothing else is active.</p>
488 <pre> struct dv_signal exitsig;
489 ev_signal_init (&exitsig, sig_cb, SIGINT);
490 ev_signal_start (myloop, &exitsig);
494 <p>Example: for some weird reason, unregister the above signal handler again.</p>
495 <pre> ev_ref (myloop);
496 ev_signal_stop (myloop, &exitsig);
503 <h1 id="ANATOMY_OF_A_WATCHER">ANATOMY OF A WATCHER</h1><p><a href="#TOP" class="toplink">Top</a></p>
504 <div id="ANATOMY_OF_A_WATCHER_CONTENT">
505 <p>A watcher is a structure that you create and register to record your
506 interest in some event. For instance, if you want to wait for STDIN to
507 become readable, you would create an <code>ev_io</code> watcher for that:</p>
508 <pre> static void my_cb (struct ev_loop *loop, struct ev_io *w, int revents)
511 ev_unloop (loop, EVUNLOOP_ALL);
514 struct ev_loop *loop = ev_default_loop (0);
515 struct ev_io stdin_watcher;
516 ev_init (&stdin_watcher, my_cb);
517 ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);
518 ev_io_start (loop, &stdin_watcher);
522 <p>As you can see, you are responsible for allocating the memory for your
523 watcher structures (and it is usually a bad idea to do this on the stack,
524 although this can sometimes be quite valid).</p>
525 <p>Each watcher structure must be initialised by a call to <code>ev_init
526 (watcher *, callback)</code>, which expects a callback to be provided. This
527 callback gets invoked each time the event occurs (or, in the case of io
528 watchers, each time the event loop detects that the file descriptor given
529 is readable and/or writable).</p>
530 <p>Each watcher type has its own <code>ev_<type>_set (watcher *, ...)</code> macro
531 with arguments specific to this watcher type. There is also a macro
532 to combine initialisation and setting in one call: <code>ev_<type>_init
533 (watcher *, callback, ...)</code>.</p>
534 <p>To make the watcher actually watch out for events, you have to start it
535 with a watcher-specific start function (<code>ev_<type>_start (loop, watcher
536 *)</code>), and you can stop watching for events at any time by calling the
537 corresponding stop function (<code>ev_<type>_stop (loop, watcher *)</code>.</p>
538 <p>As long as your watcher is active (has been started but not stopped) you
539 must not touch the values stored in it. Most specifically you must never
540 reinitialise it or call its set macro.</p>
541 <p>You can check whether an event is active by calling the <code>ev_is_active
542 (watcher *)</code> macro. To see whether an event is outstanding (but the
543 callback for it has not been called yet) you can use the <code>ev_is_pending
544 (watcher *)</code> macro.</p>
545 <p>Each and every callback receives the event loop pointer as first, the
546 registered watcher structure as second, and a bitset of received events as
548 <p>The received events usually include a single bit per event type received
549 (you can receive multiple events at the same time). The possible bit masks
552 <dt><code>EV_READ</code></dt>
553 <dt><code>EV_WRITE</code></dt>
555 <p>The file descriptor in the <code>ev_io</code> watcher has become readable and/or
558 <dt><code>EV_TIMEOUT</code></dt>
560 <p>The <code>ev_timer</code> watcher has timed out.</p>
562 <dt><code>EV_PERIODIC</code></dt>
564 <p>The <code>ev_periodic</code> watcher has timed out.</p>
566 <dt><code>EV_SIGNAL</code></dt>
568 <p>The signal specified in the <code>ev_signal</code> watcher has been received by a thread.</p>
570 <dt><code>EV_CHILD</code></dt>
572 <p>The pid specified in the <code>ev_child</code> watcher has received a status change.</p>
574 <dt><code>EV_IDLE</code></dt>
576 <p>The <code>ev_idle</code> watcher has determined that you have nothing better to do.</p>
578 <dt><code>EV_PREPARE</code></dt>
579 <dt><code>EV_CHECK</code></dt>
581 <p>All <code>ev_prepare</code> watchers are invoked just <i>before</i> <code>ev_loop</code> starts
582 to gather new events, and all <code>ev_check</code> watchers are invoked just after
583 <code>ev_loop</code> has gathered them, but before it invokes any callbacks for any
584 received events. Callbacks of both watcher types can start and stop as
585 many watchers as they want, and all of them will be taken into account
586 (for example, a <code>ev_prepare</code> watcher might start an idle watcher to keep
587 <code>ev_loop</code> from blocking).</p>
589 <dt><code>EV_ERROR</code></dt>
591 <p>An unspecified error has occured, the watcher has been stopped. This might
592 happen because the watcher could not be properly started because libev
593 ran out of memory, a file descriptor was found to be closed or any other
594 problem. You best act on it by reporting the problem and somehow coping
595 with the watcher being stopped.</p>
596 <p>Libev will usually signal a few "dummy" events together with an error,
597 for example it might indicate that a fd is readable or writable, and if
598 your callbacks is well-written it can just attempt the operation and cope
599 with the error from read() or write(). This will not work in multithreaded
600 programs, though, so beware.</p>
605 <h2 id="ASSOCIATING_CUSTOM_DATA_WITH_A_WATCH">ASSOCIATING CUSTOM DATA WITH A WATCHER</h2>
606 <div id="ASSOCIATING_CUSTOM_DATA_WITH_A_WATCH-2">
607 <p>Each watcher has, by default, a member <code>void *data</code> that you can change
608 and read at any time, libev will completely ignore it. This can be used
609 to associate arbitrary data with your watcher. If you need more data and
610 don't want to allocate memory and store a pointer to it in that data
611 member, you can also "subclass" the watcher type and provide your own
618 struct whatever *mostinteresting;
622 <p>And since your callback will be called with a pointer to the watcher, you
623 can cast it back to your own type:</p>
624 <pre> static void my_cb (struct ev_loop *loop, struct ev_io *w_, int revents)
626 struct my_io *w = (struct my_io *)w_;
631 <p>More interesting and less C-conformant ways of catsing your callback type
632 have been omitted....</p>
639 <h1 id="WATCHER_TYPES">WATCHER TYPES</h1><p><a href="#TOP" class="toplink">Top</a></p>
640 <div id="WATCHER_TYPES_CONTENT">
641 <p>This section describes each watcher in detail, but will not repeat
642 information given in the last section.</p>
649 <h2 id="code_ev_io_code_is_this_file_descrip"><code>ev_io</code> - is this file descriptor readable or writable</h2>
650 <div id="code_ev_io_code_is_this_file_descrip-2">
651 <p>I/O watchers check whether a file descriptor is readable or writable
652 in each iteration of the event loop (This behaviour is called
653 level-triggering because you keep receiving events as long as the
654 condition persists. Remember you can stop the watcher if you don't want to
655 act on the event and neither want to receive future events).</p>
656 <p>In general you can register as many read and/or write event watchers per
657 fd as you want (as long as you don't confuse yourself). Setting all file
658 descriptors to non-blocking mode is also usually a good idea (but not
659 required if you know what you are doing).</p>
660 <p>You have to be careful with dup'ed file descriptors, though. Some backends
661 (the linux epoll backend is a notable example) cannot handle dup'ed file
662 descriptors correctly if you register interest in two or more fds pointing
663 to the same underlying file/socket etc. description (that is, they share
664 the same underlying "file open").</p>
665 <p>If you must do this, then force the use of a known-to-be-good backend
666 (at the time of this writing, this includes only <code>EVBACKEND_SELECT</code> and
667 <code>EVBACKEND_POLL</code>).</p>
669 <dt>ev_io_init (ev_io *, callback, int fd, int events)</dt>
670 <dt>ev_io_set (ev_io *, int fd, int events)</dt>
672 <p>Configures an <code>ev_io</code> watcher. The fd is the file descriptor to rceeive
673 events for and events is either <code>EV_READ</code>, <code>EV_WRITE</code> or <code>EV_READ |
674 EV_WRITE</code> to receive the given events.</p>
675 <p>Please note that most of the more scalable backend mechanisms (for example
676 epoll and solaris ports) can result in spurious readyness notifications
677 for file descriptors, so you practically need to use non-blocking I/O (and
678 treat callback invocation as hint only), or retest separately with a safe
679 interface before doing I/O (XLib can do this), or force the use of either
680 <code>EVBACKEND_SELECT</code> or <code>EVBACKEND_POLL</code>, which don't suffer from this
681 problem. Also note that it is quite easy to have your callback invoked
682 when the readyness condition is no longer valid even when employing
683 typical ways of handling events, so its a good idea to use non-blocking
684 I/O unconditionally.</p>
687 <p>Example: call <code>stdin_readable_cb</code> when STDIN_FILENO has become, well
688 readable, but only once. Since it is likely line-buffered, you could
689 attempt to read a whole line in the callback:</p>
691 stdin_readable_cb (struct ev_loop *loop, struct ev_io *w, int revents)
693 ev_io_stop (loop, w);
694 .. read from stdin here (or from w->fd) and haqndle any I/O errors
698 struct ev_loop *loop = ev_default_init (0);
699 struct ev_io stdin_readable;
700 ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
701 ev_io_start (loop, &stdin_readable);
710 <h2 id="code_ev_timer_code_relative_and_opti"><code>ev_timer</code> - relative and optionally recurring timeouts</h2>
711 <div id="code_ev_timer_code_relative_and_opti-2">
712 <p>Timer watchers are simple relative timers that generate an event after a
713 given time, and optionally repeating in regular intervals after that.</p>
714 <p>The timers are based on real time, that is, if you register an event that
715 times out after an hour and you reset your system clock to last years
716 time, it will still time out after (roughly) and hour. "Roughly" because
717 detecting time jumps is hard, and some inaccuracies are unavoidable (the
718 monotonic clock option helps a lot here).</p>
719 <p>The relative timeouts are calculated relative to the <code>ev_now ()</code>
720 time. This is usually the right thing as this timestamp refers to the time
721 of the event triggering whatever timeout you are modifying/starting. If
722 you suspect event processing to be delayed and you <i>need</i> to base the timeout
723 on the current time, use something like this to adjust for this:</p>
724 <pre> ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);
727 <p>The callback is guarenteed to be invoked only when its timeout has passed,
728 but if multiple timers become ready during the same loop iteration then
729 order of execution is undefined.</p>
731 <dt>ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)</dt>
732 <dt>ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)</dt>
734 <p>Configure the timer to trigger after <code>after</code> seconds. If <code>repeat</code> is
735 <code>0.</code>, then it will automatically be stopped. If it is positive, then the
736 timer will automatically be configured to trigger again <code>repeat</code> seconds
737 later, again, and again, until stopped manually.</p>
738 <p>The timer itself will do a best-effort at avoiding drift, that is, if you
739 configure a timer to trigger every 10 seconds, then it will trigger at
740 exactly 10 second intervals. If, however, your program cannot keep up with
741 the timer (because it takes longer than those 10 seconds to do stuff) the
742 timer will not fire more than once per event loop iteration.</p>
744 <dt>ev_timer_again (loop)</dt>
746 <p>This will act as if the timer timed out and restart it again if it is
747 repeating. The exact semantics are:</p>
748 <p>If the timer is started but nonrepeating, stop it.</p>
749 <p>If the timer is repeating, either start it if necessary (with the repeat
750 value), or reset the running timer to the repeat value.</p>
751 <p>This sounds a bit complicated, but here is a useful and typical
752 example: Imagine you have a tcp connection and you want a so-called idle
753 timeout, that is, you want to be called when there have been, say, 60
754 seconds of inactivity on the socket. The easiest way to do this is to
755 configure an <code>ev_timer</code> with after=repeat=60 and calling ev_timer_again each
756 time you successfully read or write some data. If you go into an idle
757 state where you do not expect data to travel on the socket, you can stop
758 the timer, and again will automatically restart it if need be.</p>
761 <p>Example: create a timer that fires after 60 seconds.</p>
763 one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
765 .. one minute over, w is actually stopped right here
768 struct ev_timer mytimer;
769 ev_timer_init (&mytimer, one_minute_cb, 60., 0.);
770 ev_timer_start (loop, &mytimer);
773 <p>Example: create a timeout timer that times out after 10 seconds of
776 timeout_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
778 .. ten seconds without any activity
781 struct ev_timer mytimer;
782 ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */
783 ev_timer_again (&mytimer); /* start timer */
786 // and in some piece of code that gets executed on any "activity":
787 // reset the timeout to start ticking again at 10 seconds
788 ev_timer_again (&mytimer);
796 <h2 id="code_ev_periodic_code_to_cron_or_not"><code>ev_periodic</code> - to cron or not to cron</h2>
797 <div id="code_ev_periodic_code_to_cron_or_not-2">
798 <p>Periodic watchers are also timers of a kind, but they are very versatile
799 (and unfortunately a bit complex).</p>
800 <p>Unlike <code>ev_timer</code>'s, they are not based on real time (or relative time)
801 but on wallclock time (absolute time). You can tell a periodic watcher
802 to trigger "at" some specific point in time. For example, if you tell a
803 periodic watcher to trigger in 10 seconds (by specifiying e.g. c<ev_now ()
804 + 10.>) and then reset your system clock to the last year, then it will
805 take a year to trigger the event (unlike an <code>ev_timer</code>, which would trigger
806 roughly 10 seconds later and of course not if you reset your system time
808 <p>They can also be used to implement vastly more complex timers, such as
809 triggering an event on eahc midnight, local time.</p>
810 <p>As with timers, the callback is guarenteed to be invoked only when the
811 time (<code>at</code>) has been passed, but if multiple periodic timers become ready
812 during the same loop iteration then order of execution is undefined.</p>
814 <dt>ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)</dt>
815 <dt>ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)</dt>
817 <p>Lots of arguments, lets sort it out... There are basically three modes of
818 operation, and we will explain them from simplest to complex:</p>
821 <dt>* absolute timer (interval = reschedule_cb = 0)</dt>
823 <p>In this configuration the watcher triggers an event at the wallclock time
824 <code>at</code> and doesn't repeat. It will not adjust when a time jump occurs,
825 that is, if it is to be run at January 1st 2011 then it will run when the
826 system time reaches or surpasses this time.</p>
828 <dt>* non-repeating interval timer (interval > 0, reschedule_cb = 0)</dt>
830 <p>In this mode the watcher will always be scheduled to time out at the next
831 <code>at + N * interval</code> time (for some integer N) and then repeat, regardless
832 of any time jumps.</p>
833 <p>This can be used to create timers that do not drift with respect to system
835 <pre> ev_periodic_set (&periodic, 0., 3600., 0);
838 <p>This doesn't mean there will always be 3600 seconds in between triggers,
839 but only that the the callback will be called when the system time shows a
840 full hour (UTC), or more correctly, when the system time is evenly divisible
842 <p>Another way to think about it (for the mathematically inclined) is that
843 <code>ev_periodic</code> will try to run the callback in this mode at the next possible
844 time where <code>time = at (mod interval)</code>, regardless of any time jumps.</p>
846 <dt>* manual reschedule mode (reschedule_cb = callback)</dt>
848 <p>In this mode the values for <code>interval</code> and <code>at</code> are both being
849 ignored. Instead, each time the periodic watcher gets scheduled, the
850 reschedule callback will be called with the watcher as first, and the
851 current time as second argument.</p>
852 <p>NOTE: <i>This callback MUST NOT stop or destroy any periodic watcher,
853 ever, or make any event loop modifications</i>. If you need to stop it,
854 return <code>now + 1e30</code> (or so, fudge fudge) and stop it afterwards (e.g. by
855 starting a prepare watcher).</p>
856 <p>Its prototype is <code>ev_tstamp (*reschedule_cb)(struct ev_periodic *w,
857 ev_tstamp now)</code>, e.g.:</p>
858 <pre> static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)
864 <p>It must return the next time to trigger, based on the passed time value
865 (that is, the lowest time value larger than to the second argument). It
866 will usually be called just before the callback will be triggered, but
867 might be called at other times, too.</p>
868 <p>NOTE: <i>This callback must always return a time that is later than the
869 passed <code>now</code> value</i>. Not even <code>now</code> itself will do, it <i>must</i> be larger.</p>
870 <p>This can be used to create very complex timers, such as a timer that
871 triggers on each midnight, local time. To do this, you would calculate the
872 next midnight after <code>now</code> and return the timestamp value for this. How
873 you do this is, again, up to you (but it is not trivial, which is the main
874 reason I omitted it as an example).</p>
879 <dt>ev_periodic_again (loop, ev_periodic *)</dt>
881 <p>Simply stops and restarts the periodic watcher again. This is only useful
882 when you changed some parameters or the reschedule callback would return
883 a different time than the last time it was called (e.g. in a crond like
884 program when the crontabs have changed).</p>
887 <p>Example: call a callback every hour, or, more precisely, whenever the
888 system clock is divisible by 3600. The callback invocation times have
889 potentially a lot of jittering, but good long-term stability.</p>
891 clock_cb (struct ev_loop *loop, struct ev_io *w, int revents)
893 ... its now a full hour (UTC, or TAI or whatever your clock follows)
896 struct ev_periodic hourly_tick;
897 ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);
898 ev_periodic_start (loop, &hourly_tick);
901 <p>Example: the same as above, but use a reschedule callback to do it:</p>
902 <pre> #include <math.h>
905 my_scheduler_cb (struct ev_periodic *w, ev_tstamp now)
907 return fmod (now, 3600.) + 3600.;
910 ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
913 <p>Example: call a callback every hour, starting now:</p>
914 <pre> struct ev_periodic hourly_tick;
915 ev_periodic_init (&hourly_tick, clock_cb,
916 fmod (ev_now (loop), 3600.), 3600., 0);
917 ev_periodic_start (loop, &hourly_tick);
925 <h2 id="code_ev_signal_code_signal_me_when_a"><code>ev_signal</code> - signal me when a signal gets signalled</h2>
926 <div id="code_ev_signal_code_signal_me_when_a-2">
927 <p>Signal watchers will trigger an event when the process receives a specific
928 signal one or more times. Even though signals are very asynchronous, libev
929 will try it's best to deliver signals synchronously, i.e. as part of the
930 normal event processing, like any other event.</p>
931 <p>You can configure as many watchers as you like per signal. Only when the
932 first watcher gets started will libev actually register a signal watcher
933 with the kernel (thus it coexists with your own signal handlers as long
934 as you don't register any with libev). Similarly, when the last signal
935 watcher for a signal is stopped libev will reset the signal handler to
936 SIG_DFL (regardless of what it was set to before).</p>
938 <dt>ev_signal_init (ev_signal *, callback, int signum)</dt>
939 <dt>ev_signal_set (ev_signal *, int signum)</dt>
941 <p>Configures the watcher to trigger on the given signal number (usually one
942 of the <code>SIGxxx</code> constants).</p>
951 <h2 id="code_ev_child_code_wait_for_pid_stat"><code>ev_child</code> - wait for pid status changes</h2>
952 <div id="code_ev_child_code_wait_for_pid_stat-2">
953 <p>Child watchers trigger when your process receives a SIGCHLD in response to
954 some child status changes (most typically when a child of yours dies).</p>
956 <dt>ev_child_init (ev_child *, callback, int pid)</dt>
957 <dt>ev_child_set (ev_child *, int pid)</dt>
959 <p>Configures the watcher to wait for status changes of process <code>pid</code> (or
960 <i>any</i> process if <code>pid</code> is specified as <code>0</code>). The callback can look
961 at the <code>rstatus</code> member of the <code>ev_child</code> watcher structure to see
962 the status word (use the macros from <code>sys/wait.h</code> and see your systems
963 <code>waitpid</code> documentation). The <code>rpid</code> member contains the pid of the
964 process causing the status change.</p>
967 <p>Example: try to exit cleanly on SIGINT and SIGTERM.</p>
969 sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)
971 ev_unloop (loop, EVUNLOOP_ALL);
974 struct ev_signal signal_watcher;
975 ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
976 ev_signal_start (loop, &sigint_cb);
984 <h2 id="code_ev_idle_code_when_you_ve_got_no"><code>ev_idle</code> - when you've got nothing better to do</h2>
985 <div id="code_ev_idle_code_when_you_ve_got_no-2">
986 <p>Idle watchers trigger events when there are no other events are pending
987 (prepare, check and other idle watchers do not count). That is, as long
988 as your process is busy handling sockets or timeouts (or even signals,
989 imagine) it will not be triggered. But when your process is idle all idle
990 watchers are being called again and again, once per event loop iteration -
991 until stopped, that is, or your process receives more events and becomes
993 <p>The most noteworthy effect is that as long as any idle watchers are
994 active, the process will not block when waiting for new events.</p>
995 <p>Apart from keeping your process non-blocking (which is a useful
996 effect on its own sometimes), idle watchers are a good place to do
997 "pseudo-background processing", or delay processing stuff to after the
998 event loop has handled all outstanding events.</p>
1000 <dt>ev_idle_init (ev_signal *, callback)</dt>
1002 <p>Initialises and configures the idle watcher - it has no parameters of any
1003 kind. There is a <code>ev_idle_set</code> macro, but using it is utterly pointless,
1007 <p>Example: dynamically allocate an <code>ev_idle</code>, start it, and in the
1008 callback, free it. Alos, use no error checking, as usual.</p>
1010 idle_cb (struct ev_loop *loop, struct ev_idle *w, int revents)
1013 // now do something you wanted to do when the program has
1014 // no longer asnything immediate to do.
1017 struct ev_idle *idle_watcher = malloc (sizeof (struct ev_idle));
1018 ev_idle_init (idle_watcher, idle_cb);
1019 ev_idle_start (loop, idle_cb);
1027 <h2 id="code_ev_prepare_code_and_code_ev_che"><code>ev_prepare</code> and <code>ev_check</code> - customise your event loop</h2>
1028 <div id="code_ev_prepare_code_and_code_ev_che-2">
1029 <p>Prepare and check watchers are usually (but not always) used in tandem:
1030 prepare watchers get invoked before the process blocks and check watchers
1032 <p>Their main purpose is to integrate other event mechanisms into libev and
1033 their use is somewhat advanced. This could be used, for example, to track
1034 variable changes, implement your own watchers, integrate net-snmp or a
1035 coroutine library and lots more.</p>
1036 <p>This is done by examining in each prepare call which file descriptors need
1037 to be watched by the other library, registering <code>ev_io</code> watchers for
1038 them and starting an <code>ev_timer</code> watcher for any timeouts (many libraries
1039 provide just this functionality). Then, in the check watcher you check for
1040 any events that occured (by checking the pending status of all watchers
1041 and stopping them) and call back into the library. The I/O and timer
1042 callbacks will never actually be called (but must be valid nevertheless,
1043 because you never know, you know?).</p>
1044 <p>As another example, the Perl Coro module uses these hooks to integrate
1045 coroutines into libev programs, by yielding to other active coroutines
1046 during each prepare and only letting the process block if no coroutines
1047 are ready to run (it's actually more complicated: it only runs coroutines
1048 with priority higher than or equal to the event loop and one coroutine
1049 of lower priority, but only once, using idle watchers to keep the event
1050 loop from blocking if lower-priority coroutines are active, thus mapping
1051 low-priority coroutines to idle/background tasks).</p>
1053 <dt>ev_prepare_init (ev_prepare *, callback)</dt>
1054 <dt>ev_check_init (ev_check *, callback)</dt>
1056 <p>Initialises and configures the prepare or check watcher - they have no
1057 parameters of any kind. There are <code>ev_prepare_set</code> and <code>ev_check_set</code>
1058 macros, but using them is utterly, utterly and completely pointless.</p>
1061 <p>Example: *TODO*.</p>
1068 <h2 id="code_ev_embed_code_when_one_backend_"><code>ev_embed</code> - when one backend isn't enough</h2>
1069 <div id="code_ev_embed_code_when_one_backend_-2">
1070 <p>This is a rather advanced watcher type that lets you embed one event loop
1072 <p>There are primarily two reasons you would want that: work around bugs and
1074 <p>As an example for a bug workaround, the kqueue backend might only support
1075 sockets on some platform, so it is unusable as generic backend, but you
1076 still want to make use of it because you have many sockets and it scales
1077 so nicely. In this case, you would create a kqueue-based loop and embed it
1078 into your default loop (which might use e.g. poll). Overall operation will
1079 be a bit slower because first libev has to poll and then call kevent, but
1080 at least you can use both at what they are best.</p>
1081 <p>As for prioritising I/O: rarely you have the case where some fds have
1082 to be watched and handled very quickly (with low latency), and even
1083 priorities and idle watchers might have too much overhead. In this case
1084 you would put all the high priority stuff in one loop and all the rest in
1085 a second one, and embed the second one in the first.</p>
1086 <p>As long as the watcher is started it will automatically handle events. The
1087 callback will be invoked whenever some events have been handled. You can
1088 set the callback to <code>0</code> to avoid having to specify one if you are not
1089 interested in that.</p>
1090 <p>Also, there have not currently been made special provisions for forking:
1091 when you fork, you not only have to call <code>ev_loop_fork</code> on both loops,
1092 but you will also have to stop and restart any <code>ev_embed</code> watchers
1094 <p>Unfortunately, not all backends are embeddable, only the ones returned by
1095 <code>ev_embeddable_backends</code> are, which, unfortunately, does not include any
1097 <p>So when you want to use this feature you will always have to be prepared
1098 that you cannot get an embeddable loop. The recommended way to get around
1099 this is to have a separate variables for your embeddable loop, try to
1100 create it, and if that fails, use the normal loop for everything:</p>
1101 <pre> struct ev_loop *loop_hi = ev_default_init (0);
1102 struct ev_loop *loop_lo = 0;
1103 struct ev_embed embed;
1105 // see if there is a chance of getting one that works
1106 // (remember that a flags value of 0 means autodetection)
1107 loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
1108 ? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
1111 // if we got one, then embed it, otherwise default to loop_hi
1114 ev_embed_init (&embed, 0, loop_lo);
1115 ev_embed_start (loop_hi, &embed);
1122 <dt>ev_embed_init (ev_embed *, callback, struct ev_loop *loop)</dt>
1123 <dt>ev_embed_set (ev_embed *, callback, struct ev_loop *loop)</dt>
1125 <p>Configures the watcher to embed the given loop, which must be embeddable.</p>
1134 <h1 id="OTHER_FUNCTIONS">OTHER FUNCTIONS</h1><p><a href="#TOP" class="toplink">Top</a></p>
1135 <div id="OTHER_FUNCTIONS_CONTENT">
1136 <p>There are some other functions of possible interest. Described. Here. Now.</p>
1138 <dt>ev_once (loop, int fd, int events, ev_tstamp timeout, callback)</dt>
1140 <p>This function combines a simple timer and an I/O watcher, calls your
1141 callback on whichever event happens first and automatically stop both
1142 watchers. This is useful if you want to wait for a single event on an fd
1143 or timeout without having to allocate/configure/start/stop/free one or
1144 more watchers yourself.</p>
1145 <p>If <code>fd</code> is less than 0, then no I/O watcher will be started and events
1146 is being ignored. Otherwise, an <code>ev_io</code> watcher for the given <code>fd</code> and
1147 <code>events</code> set will be craeted and started.</p>
1148 <p>If <code>timeout</code> is less than 0, then no timeout watcher will be
1149 started. Otherwise an <code>ev_timer</code> watcher with after = <code>timeout</code> (and
1150 repeat = 0) will be started. While <code>0</code> is a valid timeout, it is of
1152 <p>The callback has the type <code>void (*cb)(int revents, void *arg)</code> and gets
1153 passed an <code>revents</code> set like normal event callbacks (a combination of
1154 <code>EV_ERROR</code>, <code>EV_READ</code>, <code>EV_WRITE</code> or <code>EV_TIMEOUT</code>) and the <code>arg</code>
1155 value passed to <code>ev_once</code>:</p>
1156 <pre> static void stdin_ready (int revents, void *arg)
1158 if (revents & EV_TIMEOUT)
1159 /* doh, nothing entered */;
1160 else if (revents & EV_READ)
1161 /* stdin might have data for us, joy! */;
1164 ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
1168 <dt>ev_feed_event (loop, watcher, int events)</dt>
1170 <p>Feeds the given event set into the event loop, as if the specified event
1171 had happened for the specified watcher (which must be a pointer to an
1172 initialised but not necessarily started event watcher).</p>
1174 <dt>ev_feed_fd_event (loop, int fd, int revents)</dt>
1176 <p>Feed an event on the given fd, as if a file descriptor backend detected
1177 the given events it.</p>
1179 <dt>ev_feed_signal_event (loop, int signum)</dt>
1181 <p>Feed an event as if the given signal occured (loop must be the default loop!).</p>
1190 <h1 id="LIBEVENT_EMULATION">LIBEVENT EMULATION</h1><p><a href="#TOP" class="toplink">Top</a></p>
1191 <div id="LIBEVENT_EMULATION_CONTENT">
1192 <p>Libev offers a compatibility emulation layer for libevent. It cannot
1193 emulate the internals of libevent, so here are some usage hints:</p>
1195 <dt>* Use it by including <event.h>, as usual.</dt>
1196 <dt>* The following members are fully supported: ev_base, ev_callback,
1197 ev_arg, ev_fd, ev_res, ev_events.</dt>
1198 <dt>* Avoid using ev_flags and the EVLIST_*-macros, while it is
1199 maintained by libev, it does not work exactly the same way as in libevent (consider
1200 it a private API).</dt>
1201 <dt>* Priorities are not currently supported. Initialising priorities
1202 will fail and all watchers will have the same priority, even though there
1203 is an ev_pri field.</dt>
1204 <dt>* Other members are not supported.</dt>
1205 <dt>* The libev emulation is <i>not</i> ABI compatible to libevent, you need
1206 to use the libev header file and library.</dt>
1210 <h1 id="C_SUPPORT">C++ SUPPORT</h1><p><a href="#TOP" class="toplink">Top</a></p>
1211 <div id="C_SUPPORT_CONTENT">
1215 <h1 id="AUTHOR">AUTHOR</h1><p><a href="#TOP" class="toplink">Top</a></p>
1216 <div id="AUTHOR_CONTENT">
1217 <p>Marc Lehmann <libev@schmorp.de>.</p>
projects / software / libev.git / blob