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="Mon Nov 26 11:20:35 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="#GENERIC_WATCHER_FUNCTIONS">GENERIC WATCHER FUNCTIONS</a></li>
27 <li><a href="#ASSOCIATING_CUSTOM_DATA_WITH_A_WATCH">ASSOCIATING CUSTOM DATA WITH A WATCHER</a></li>
30 <li><a href="#WATCHER_TYPES">WATCHER TYPES</a>
31 <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>
32 <li><a href="#code_ev_timer_code_relative_and_opti"><code>ev_timer</code> - relative and optionally repeating timeouts</a></li>
33 <li><a href="#code_ev_periodic_code_to_cron_or_not"><code>ev_periodic</code> - to cron or not to cron?</a></li>
34 <li><a href="#code_ev_signal_code_signal_me_when_a"><code>ev_signal</code> - signal me when a signal gets signalled!</a></li>
35 <li><a href="#code_ev_child_code_watch_out_for_pro"><code>ev_child</code> - watch out for process status changes</a></li>
36 <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>
37 <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>
38 <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="#OTHER_FUNCTIONS">OTHER FUNCTIONS</a></li>
42 <li><a href="#LIBEVENT_EMULATION">LIBEVENT EMULATION</a></li>
43 <li><a href="#C_SUPPORT">C++ SUPPORT</a></li>
44 <li><a href="#EMBEDDING">EMBEDDING</a>
45 <ul><li><a href="#FILESETS">FILESETS</a>
46 <ul><li><a href="#CORE_EVENT_LOOP">CORE EVENT LOOP</a></li>
47 <li><a href="#LIBEVENT_COMPATIBILITY_API">LIBEVENT COMPATIBILITY API</a></li>
48 <li><a href="#AUTOCONF_SUPPORT">AUTOCONF SUPPORT</a></li>
51 <li><a href="#PREPROCESSOR_SYMBOLS_MACROS">PREPROCESSOR SYMBOLS/MACROS</a></li>
52 <li><a href="#EXAMPLES">EXAMPLES</a></li>
55 <li><a href="#COMPLEXITIES">COMPLEXITIES</a></li>
56 <li><a href="#AUTHOR">AUTHOR</a>
61 <h1 id="NAME">NAME</h1><p><a href="#TOP" class="toplink">Top</a></p>
62 <div id="NAME_CONTENT">
63 <p>libev - a high performance full-featured event loop written in C</p>
66 <h1 id="SYNOPSIS">SYNOPSIS</h1><p><a href="#TOP" class="toplink">Top</a></p>
67 <div id="SYNOPSIS_CONTENT">
68 <pre> #include <ev.h>
73 <h1 id="DESCRIPTION">DESCRIPTION</h1><p><a href="#TOP" class="toplink">Top</a></p>
74 <div id="DESCRIPTION_CONTENT">
75 <p>Libev is an event loop: you register interest in certain events (such as a
76 file descriptor being readable or a timeout occuring), and it will manage
77 these event sources and provide your program with events.</p>
78 <p>To do this, it must take more or less complete control over your process
79 (or thread) by executing the <i>event loop</i> handler, and will then
80 communicate events via a callback mechanism.</p>
81 <p>You register interest in certain events by registering so-called <i>event
82 watchers</i>, which are relatively small C structures you initialise with the
83 details of the event, and then hand it over to libev by <i>starting</i> the
87 <h1 id="FEATURES">FEATURES</h1><p><a href="#TOP" class="toplink">Top</a></p>
88 <div id="FEATURES_CONTENT">
89 <p>Libev supports select, poll, the linux-specific epoll and the bsd-specific
90 kqueue mechanisms for file descriptor events, relative timers, absolute
91 timers with customised rescheduling, signal events, process status change
92 events (related to SIGCHLD), and event watchers dealing with the event
93 loop mechanism itself (idle, prepare and check watchers). It also is quite
94 fast (see this <a href="http://libev.schmorp.de/bench.html">benchmark</a> comparing
95 it to libevent for example).</p>
98 <h1 id="CONVENTIONS">CONVENTIONS</h1><p><a href="#TOP" class="toplink">Top</a></p>
99 <div id="CONVENTIONS_CONTENT">
100 <p>Libev is very configurable. In this manual the default configuration
101 will be described, which supports multiple event loops. For more info
102 about various configuration options please have a look at the file
103 <cite>README.embed</cite> in the libev distribution. If libev was configured without
104 support for multiple event loops, then all functions taking an initial
105 argument of name <code>loop</code> (which is always of type <code>struct ev_loop *</code>)
106 will not have this argument.</p>
109 <h1 id="TIME_REPRESENTATION">TIME REPRESENTATION</h1><p><a href="#TOP" class="toplink">Top</a></p>
110 <div id="TIME_REPRESENTATION_CONTENT">
111 <p>Libev represents time as a single floating point number, representing the
112 (fractional) number of seconds since the (POSIX) epoch (somewhere near
113 the beginning of 1970, details are complicated, don't ask). This type is
114 called <code>ev_tstamp</code>, which is what you should use too. It usually aliases
115 to the <code>double</code> type in C, and when you need to do any calculations on
116 it, you should treat it as such.</p>
123 <h1 id="GLOBAL_FUNCTIONS">GLOBAL FUNCTIONS</h1><p><a href="#TOP" class="toplink">Top</a></p>
124 <div id="GLOBAL_FUNCTIONS_CONTENT">
125 <p>These functions can be called anytime, even before initialising the
126 library in any way.</p>
128 <dt>ev_tstamp ev_time ()</dt>
130 <p>Returns the current time as libev would use it. Please note that the
131 <code>ev_now</code> function is usually faster and also often returns the timestamp
132 you actually want to know.</p>
134 <dt>int ev_version_major ()</dt>
135 <dt>int ev_version_minor ()</dt>
137 <p>You can find out the major and minor version numbers of the library
138 you linked against by calling the functions <code>ev_version_major</code> and
139 <code>ev_version_minor</code>. If you want, you can compare against the global
140 symbols <code>EV_VERSION_MAJOR</code> and <code>EV_VERSION_MINOR</code>, which specify the
141 version of the library your program was compiled against.</p>
142 <p>Usually, it's a good idea to terminate if the major versions mismatch,
143 as this indicates an incompatible change. Minor versions are usually
144 compatible to older versions, so a larger minor version alone is usually
146 <p>Example: make sure we haven't accidentally been linked against the wrong
148 <pre> assert (("libev version mismatch",
149 ev_version_major () == EV_VERSION_MAJOR
150 && ev_version_minor () >= EV_VERSION_MINOR));
154 <dt>unsigned int ev_supported_backends ()</dt>
156 <p>Return the set of all backends (i.e. their corresponding <code>EV_BACKEND_*</code>
157 value) compiled into this binary of libev (independent of their
158 availability on the system you are running on). See <code>ev_default_loop</code> for
159 a description of the set values.</p>
160 <p>Example: make sure we have the epoll method, because yeah this is cool and
161 a must have and can we have a torrent of it please!!!11</p>
162 <pre> assert (("sorry, no epoll, no sex",
163 ev_supported_backends () & EVBACKEND_EPOLL));
167 <dt>unsigned int ev_recommended_backends ()</dt>
169 <p>Return the set of all backends compiled into this binary of libev and also
170 recommended for this platform. This set is often smaller than the one
171 returned by <code>ev_supported_backends</code>, as for example kqueue is broken on
172 most BSDs and will not be autodetected unless you explicitly request it
173 (assuming you know what you are doing). This is the set of backends that
174 libev will probe for if you specify no backends explicitly.</p>
176 <dt>unsigned int ev_embeddable_backends ()</dt>
178 <p>Returns the set of backends that are embeddable in other event loops. This
179 is the theoretical, all-platform, value. To find which backends
180 might be supported on the current system, you would need to look at
181 <code>ev_embeddable_backends () & ev_supported_backends ()</code>, likewise for
182 recommended ones.</p>
183 <p>See the description of <code>ev_embed</code> watchers for more info.</p>
185 <dt>ev_set_allocator (void *(*cb)(void *ptr, long size))</dt>
187 <p>Sets the allocation function to use (the prototype is similar to the
188 realloc C function, the semantics are identical). It is used to allocate
189 and free memory (no surprises here). If it returns zero when memory
190 needs to be allocated, the library might abort or take some potentially
191 destructive action. The default is your system realloc function.</p>
192 <p>You could override this function in high-availability programs to, say,
193 free some memory if it cannot allocate memory, to use a special allocator,
194 or even to sleep a while and retry until some memory is available.</p>
195 <p>Example: replace the libev allocator with one that waits a bit and then
196 retries: better than mine).</p>
198 persistent_realloc (void *ptr, long size)
202 void *newptr = realloc (ptr, size);
212 ev_set_allocator (persistent_realloc);
216 <dt>ev_set_syserr_cb (void (*cb)(const char *msg));</dt>
218 <p>Set the callback function to call on a retryable syscall error (such
219 as failed select, poll, epoll_wait). The message is a printable string
220 indicating the system call or subsystem causing the problem. If this
221 callback is set, then libev will expect it to remedy the sitution, no
222 matter what, when it returns. That is, libev will generally retry the
223 requested operation, or, if the condition doesn't go away, do bad stuff
225 <p>Example: do the same thing as libev does internally:</p>
227 fatal_error (const char *msg)
234 ev_set_syserr_cb (fatal_error);
241 <h1 id="FUNCTIONS_CONTROLLING_THE_EVENT_LOOP">FUNCTIONS CONTROLLING THE EVENT LOOP</h1><p><a href="#TOP" class="toplink">Top</a></p>
242 <div id="FUNCTIONS_CONTROLLING_THE_EVENT_LOOP-2">
243 <p>An event loop is described by a <code>struct ev_loop *</code>. The library knows two
244 types of such loops, the <i>default</i> loop, which supports signals and child
245 events, and dynamically created loops which do not.</p>
246 <p>If you use threads, a common model is to run the default event loop
247 in your main thread (or in a separate thread) and for each thread you
248 create, you also create another event loop. Libev itself does no locking
249 whatsoever, so if you mix calls to the same event loop in different
250 threads, make sure you lock (this is usually a bad idea, though, even if
251 done correctly, because it's hideous and inefficient).</p>
253 <dt>struct ev_loop *ev_default_loop (unsigned int flags)</dt>
255 <p>This will initialise the default event loop if it hasn't been initialised
256 yet and return it. If the default loop could not be initialised, returns
257 false. If it already was initialised it simply returns it (and ignores the
258 flags. If that is troubling you, check <code>ev_backend ()</code> afterwards).</p>
259 <p>If you don't know what event loop to use, use the one returned from this
261 <p>The flags argument can be used to specify special behaviour or specific
262 backends to use, and is usually specified as <code>0</code> (or <code>EVFLAG_AUTO</code>).</p>
263 <p>The following flags are supported:</p>
266 <dt><code>EVFLAG_AUTO</code></dt>
268 <p>The default flags value. Use this if you have no clue (it's the right
269 thing, believe me).</p>
271 <dt><code>EVFLAG_NOENV</code></dt>
273 <p>If this flag bit is ored into the flag value (or the program runs setuid
274 or setgid) then libev will <i>not</i> look at the environment variable
275 <code>LIBEV_FLAGS</code>. Otherwise (the default), this environment variable will
276 override the flags completely if it is found in the environment. This is
277 useful to try out specific backends to test their performance, or to work
280 <dt><code>EVBACKEND_SELECT</code> (value 1, portable select backend)</dt>
282 <p>This is your standard select(2) backend. Not <i>completely</i> standard, as
283 libev tries to roll its own fd_set with no limits on the number of fds,
284 but if that fails, expect a fairly low limit on the number of fds when
285 using this backend. It doesn't scale too well (O(highest_fd)), but its usually
286 the fastest backend for a low number of fds.</p>
288 <dt><code>EVBACKEND_POLL</code> (value 2, poll backend, available everywhere except on windows)</dt>
290 <p>And this is your standard poll(2) backend. It's more complicated than
291 select, but handles sparse fds better and has no artificial limit on the
292 number of fds you can use (except it will slow down considerably with a
293 lot of inactive fds). It scales similarly to select, i.e. O(total_fds).</p>
295 <dt><code>EVBACKEND_EPOLL</code> (value 4, Linux)</dt>
297 <p>For few fds, this backend is a bit little slower than poll and select,
298 but it scales phenomenally better. While poll and select usually scale like
299 O(total_fds) where n is the total number of fds (or the highest fd), epoll scales
300 either O(1) or O(active_fds).</p>
301 <p>While stopping and starting an I/O watcher in the same iteration will
302 result in some caching, there is still a syscall per such incident
303 (because the fd could point to a different file description now), so its
304 best to avoid that. Also, dup()ed file descriptors might not work very
305 well if you register events for both fds.</p>
306 <p>Please note that epoll sometimes generates spurious notifications, so you
307 need to use non-blocking I/O or other means to avoid blocking when no data
308 (or space) is available.</p>
310 <dt><code>EVBACKEND_KQUEUE</code> (value 8, most BSD clones)</dt>
312 <p>Kqueue deserves special mention, as at the time of this writing, it
313 was broken on all BSDs except NetBSD (usually it doesn't work with
314 anything but sockets and pipes, except on Darwin, where of course its
315 completely useless). For this reason its not being "autodetected"
316 unless you explicitly specify it explicitly in the flags (i.e. using
317 <code>EVBACKEND_KQUEUE</code>).</p>
318 <p>It scales in the same way as the epoll backend, but the interface to the
319 kernel is more efficient (which says nothing about its actual speed, of
320 course). While starting and stopping an I/O watcher does not cause an
321 extra syscall as with epoll, it still adds up to four event changes per
322 incident, so its best to avoid that.</p>
324 <dt><code>EVBACKEND_DEVPOLL</code> (value 16, Solaris 8)</dt>
326 <p>This is not implemented yet (and might never be).</p>
328 <dt><code>EVBACKEND_PORT</code> (value 32, Solaris 10)</dt>
330 <p>This uses the Solaris 10 port mechanism. As with everything on Solaris,
331 it's really slow, but it still scales very well (O(active_fds)).</p>
332 <p>Please note that solaris ports can result in a lot of spurious
333 notifications, so you need to use non-blocking I/O or other means to avoid
334 blocking when no data (or space) is available.</p>
336 <dt><code>EVBACKEND_ALL</code></dt>
338 <p>Try all backends (even potentially broken ones that wouldn't be tried
339 with <code>EVFLAG_AUTO</code>). Since this is a mask, you can do stuff such as
340 <code>EVBACKEND_ALL & ~EVBACKEND_KQUEUE</code>.</p>
344 <p>If one or more of these are ored into the flags value, then only these
345 backends will be tried (in the reverse order as given here). If none are
346 specified, most compiled-in backend will be tried, usually in reverse
347 order of their flag values :)</p>
348 <p>The most typical usage is like this:</p>
349 <pre> if (!ev_default_loop (0))
350 fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
353 <p>Restrict libev to the select and poll backends, and do not allow
354 environment settings to be taken into account:</p>
355 <pre> ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);
358 <p>Use whatever libev has to offer, but make sure that kqueue is used if
359 available (warning, breaks stuff, best use only with your own private
360 event loop and only if you know the OS supports your types of fds):</p>
361 <pre> ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);
365 <dt>struct ev_loop *ev_loop_new (unsigned int flags)</dt>
367 <p>Similar to <code>ev_default_loop</code>, but always creates a new event loop that is
368 always distinct from the default loop. Unlike the default loop, it cannot
369 handle signal and child watchers, and attempts to do so will be greeted by
370 undefined behaviour (or a failed assertion if assertions are enabled).</p>
371 <p>Example: try to create a event loop that uses epoll and nothing else.</p>
372 <pre> struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);
374 fatal ("no epoll found here, maybe it hides under your chair");
378 <dt>ev_default_destroy ()</dt>
380 <p>Destroys the default loop again (frees all memory and kernel state
381 etc.). None of the active event watchers will be stopped in the normal
382 sense, so e.g. <code>ev_is_active</code> might still return true. It is your
383 responsibility to either stop all watchers cleanly yoursef <i>before</i>
384 calling this function, or cope with the fact afterwards (which is usually
385 the easiest thing, youc na just ignore the watchers and/or <code>free ()</code> them
388 <dt>ev_loop_destroy (loop)</dt>
390 <p>Like <code>ev_default_destroy</code>, but destroys an event loop created by an
391 earlier call to <code>ev_loop_new</code>.</p>
393 <dt>ev_default_fork ()</dt>
395 <p>This function reinitialises the kernel state for backends that have
396 one. Despite the name, you can call it anytime, but it makes most sense
397 after forking, in either the parent or child process (or both, but that
398 again makes little sense).</p>
399 <p>You <i>must</i> call this function in the child process after forking if and
400 only if you want to use the event library in both processes. If you just
401 fork+exec, you don't have to call it.</p>
402 <p>The function itself is quite fast and it's usually not a problem to call
403 it just in case after a fork. To make this easy, the function will fit in
404 quite nicely into a call to <code>pthread_atfork</code>:</p>
405 <pre> pthread_atfork (0, 0, ev_default_fork);
408 <p>At the moment, <code>EVBACKEND_SELECT</code> and <code>EVBACKEND_POLL</code> are safe to use
409 without calling this function, so if you force one of those backends you
410 do not need to care.</p>
412 <dt>ev_loop_fork (loop)</dt>
414 <p>Like <code>ev_default_fork</code>, but acts on an event loop created by
415 <code>ev_loop_new</code>. Yes, you have to call this on every allocated event loop
416 after fork, and how you do this is entirely your own problem.</p>
418 <dt>unsigned int ev_backend (loop)</dt>
420 <p>Returns one of the <code>EVBACKEND_*</code> flags indicating the event backend in
423 <dt>ev_tstamp ev_now (loop)</dt>
425 <p>Returns the current "event loop time", which is the time the event loop
426 received events and started processing them. This timestamp does not
427 change as long as callbacks are being processed, and this is also the base
428 time used for relative timers. You can treat it as the timestamp of the
429 event occuring (or more correctly, libev finding out about it).</p>
431 <dt>ev_loop (loop, int flags)</dt>
433 <p>Finally, this is it, the event handler. This function usually is called
434 after you initialised all your watchers and you want to start handling
436 <p>If the flags argument is specified as <code>0</code>, it will not return until
437 either no event watchers are active anymore or <code>ev_unloop</code> was called.</p>
438 <p>Please note that an explicit <code>ev_unloop</code> is usually better than
439 relying on all watchers to be stopped when deciding when a program has
440 finished (especially in interactive programs), but having a program that
441 automatically loops as long as it has to and no longer by virtue of
442 relying on its watchers stopping correctly is a thing of beauty.</p>
443 <p>A flags value of <code>EVLOOP_NONBLOCK</code> will look for new events, will handle
444 those events and any outstanding ones, but will not block your process in
445 case there are no events and will return after one iteration of the loop.</p>
446 <p>A flags value of <code>EVLOOP_ONESHOT</code> will look for new events (waiting if
447 neccessary) and will handle those and any outstanding ones. It will block
448 your process until at least one new event arrives, and will return after
449 one iteration of the loop. This is useful if you are waiting for some
450 external event in conjunction with something not expressible using other
451 libev watchers. However, a pair of <code>ev_prepare</code>/<code>ev_check</code> watchers is
452 usually a better approach for this kind of thing.</p>
453 <p>Here are the gory details of what <code>ev_loop</code> does:</p>
454 <pre> * If there are no active watchers (reference count is zero), return.
455 - Queue prepare watchers and then call all outstanding watchers.
456 - If we have been forked, recreate the kernel state.
457 - Update the kernel state with all outstanding changes.
458 - Update the "event loop time".
459 - Calculate for how long to block.
460 - Block the process, waiting for any events.
461 - Queue all outstanding I/O (fd) events.
462 - Update the "event loop time" and do time jump handling.
463 - Queue all outstanding timers.
464 - Queue all outstanding periodics.
465 - If no events are pending now, queue all idle watchers.
466 - Queue all check watchers.
467 - Call all queued watchers in reverse order (i.e. check watchers first).
468 Signals and child watchers are implemented as I/O watchers, and will
469 be handled here by queueing them when their watcher gets executed.
470 - If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK
471 were used, return, otherwise continue with step *.
474 <p>Example: queue some jobs and then loop until no events are outsanding
476 <pre> ... queue jobs here, make sure they register event watchers as long
477 ... as they still have work to do (even an idle watcher will do..)
478 ev_loop (my_loop, 0);
483 <dt>ev_unloop (loop, how)</dt>
485 <p>Can be used to make a call to <code>ev_loop</code> return early (but only after it
486 has processed all outstanding events). The <code>how</code> argument must be either
487 <code>EVUNLOOP_ONE</code>, which will make the innermost <code>ev_loop</code> call return, or
488 <code>EVUNLOOP_ALL</code>, which will make all nested <code>ev_loop</code> calls return.</p>
490 <dt>ev_ref (loop)</dt>
491 <dt>ev_unref (loop)</dt>
493 <p>Ref/unref can be used to add or remove a reference count on the event
494 loop: Every watcher keeps one reference, and as long as the reference
495 count is nonzero, <code>ev_loop</code> will not return on its own. If you have
496 a watcher you never unregister that should not keep <code>ev_loop</code> from
497 returning, ev_unref() after starting, and ev_ref() before stopping it. For
498 example, libev itself uses this for its internal signal pipe: It is not
499 visible to the libev user and should not keep <code>ev_loop</code> from exiting if
500 no event watchers registered by it are active. It is also an excellent
501 way to do this for generic recurring timers or from within third-party
502 libraries. Just remember to <i>unref after start</i> and <i>ref before stop</i>.</p>
503 <p>Example: create a signal watcher, but keep it from keeping <code>ev_loop</code>
504 running when nothing else is active.</p>
505 <pre> struct dv_signal exitsig;
506 ev_signal_init (&exitsig, sig_cb, SIGINT);
507 ev_signal_start (myloop, &exitsig);
511 <p>Example: for some weird reason, unregister the above signal handler again.</p>
512 <pre> ev_ref (myloop);
513 ev_signal_stop (myloop, &exitsig);
524 <h1 id="ANATOMY_OF_A_WATCHER">ANATOMY OF A WATCHER</h1><p><a href="#TOP" class="toplink">Top</a></p>
525 <div id="ANATOMY_OF_A_WATCHER_CONTENT">
526 <p>A watcher is a structure that you create and register to record your
527 interest in some event. For instance, if you want to wait for STDIN to
528 become readable, you would create an <code>ev_io</code> watcher for that:</p>
529 <pre> static void my_cb (struct ev_loop *loop, struct ev_io *w, int revents)
532 ev_unloop (loop, EVUNLOOP_ALL);
535 struct ev_loop *loop = ev_default_loop (0);
536 struct ev_io stdin_watcher;
537 ev_init (&stdin_watcher, my_cb);
538 ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);
539 ev_io_start (loop, &stdin_watcher);
543 <p>As you can see, you are responsible for allocating the memory for your
544 watcher structures (and it is usually a bad idea to do this on the stack,
545 although this can sometimes be quite valid).</p>
546 <p>Each watcher structure must be initialised by a call to <code>ev_init
547 (watcher *, callback)</code>, which expects a callback to be provided. This
548 callback gets invoked each time the event occurs (or, in the case of io
549 watchers, each time the event loop detects that the file descriptor given
550 is readable and/or writable).</p>
551 <p>Each watcher type has its own <code>ev_<type>_set (watcher *, ...)</code> macro
552 with arguments specific to this watcher type. There is also a macro
553 to combine initialisation and setting in one call: <code>ev_<type>_init
554 (watcher *, callback, ...)</code>.</p>
555 <p>To make the watcher actually watch out for events, you have to start it
556 with a watcher-specific start function (<code>ev_<type>_start (loop, watcher
557 *)</code>), and you can stop watching for events at any time by calling the
558 corresponding stop function (<code>ev_<type>_stop (loop, watcher *)</code>.</p>
559 <p>As long as your watcher is active (has been started but not stopped) you
560 must not touch the values stored in it. Most specifically you must never
561 reinitialise it or call its <code>set</code> macro.</p>
562 <p>Each and every callback receives the event loop pointer as first, the
563 registered watcher structure as second, and a bitset of received events as
565 <p>The received events usually include a single bit per event type received
566 (you can receive multiple events at the same time). The possible bit masks
569 <dt><code>EV_READ</code></dt>
570 <dt><code>EV_WRITE</code></dt>
572 <p>The file descriptor in the <code>ev_io</code> watcher has become readable and/or
575 <dt><code>EV_TIMEOUT</code></dt>
577 <p>The <code>ev_timer</code> watcher has timed out.</p>
579 <dt><code>EV_PERIODIC</code></dt>
581 <p>The <code>ev_periodic</code> watcher has timed out.</p>
583 <dt><code>EV_SIGNAL</code></dt>
585 <p>The signal specified in the <code>ev_signal</code> watcher has been received by a thread.</p>
587 <dt><code>EV_CHILD</code></dt>
589 <p>The pid specified in the <code>ev_child</code> watcher has received a status change.</p>
591 <dt><code>EV_IDLE</code></dt>
593 <p>The <code>ev_idle</code> watcher has determined that you have nothing better to do.</p>
595 <dt><code>EV_PREPARE</code></dt>
596 <dt><code>EV_CHECK</code></dt>
598 <p>All <code>ev_prepare</code> watchers are invoked just <i>before</i> <code>ev_loop</code> starts
599 to gather new events, and all <code>ev_check</code> watchers are invoked just after
600 <code>ev_loop</code> has gathered them, but before it invokes any callbacks for any
601 received events. Callbacks of both watcher types can start and stop as
602 many watchers as they want, and all of them will be taken into account
603 (for example, a <code>ev_prepare</code> watcher might start an idle watcher to keep
604 <code>ev_loop</code> from blocking).</p>
606 <dt><code>EV_ERROR</code></dt>
608 <p>An unspecified error has occured, the watcher has been stopped. This might
609 happen because the watcher could not be properly started because libev
610 ran out of memory, a file descriptor was found to be closed or any other
611 problem. You best act on it by reporting the problem and somehow coping
612 with the watcher being stopped.</p>
613 <p>Libev will usually signal a few "dummy" events together with an error,
614 for example it might indicate that a fd is readable or writable, and if
615 your callbacks is well-written it can just attempt the operation and cope
616 with the error from read() or write(). This will not work in multithreaded
617 programs, though, so beware.</p>
622 <h2 id="GENERIC_WATCHER_FUNCTIONS">GENERIC WATCHER FUNCTIONS</h2>
623 <div id="GENERIC_WATCHER_FUNCTIONS_CONTENT">
624 <p>In the following description, <code>TYPE</code> stands for the watcher type,
625 e.g. <code>timer</code> for <code>ev_timer</code> watchers and <code>io</code> for <code>ev_io</code> watchers.</p>
627 <dt><code>ev_init</code> (ev_TYPE *watcher, callback)</dt>
629 <p>This macro initialises the generic portion of a watcher. The contents
630 of the watcher object can be arbitrary (so <code>malloc</code> will do). Only
631 the generic parts of the watcher are initialised, you <i>need</i> to call
632 the type-specific <code>ev_TYPE_set</code> macro afterwards to initialise the
633 type-specific parts. For each type there is also a <code>ev_TYPE_init</code> macro
634 which rolls both calls into one.</p>
635 <p>You can reinitialise a watcher at any time as long as it has been stopped
636 (or never started) and there are no pending events outstanding.</p>
637 <p>The callback is always of type <code>void (*)(ev_loop *loop, ev_TYPE *watcher,
638 int revents)</code>.</p>
640 <dt><code>ev_TYPE_set</code> (ev_TYPE *, [args])</dt>
642 <p>This macro initialises the type-specific parts of a watcher. You need to
643 call <code>ev_init</code> at least once before you call this macro, but you can
644 call <code>ev_TYPE_set</code> any number of times. You must not, however, call this
645 macro on a watcher that is active (it can be pending, however, which is a
646 difference to the <code>ev_init</code> macro).</p>
647 <p>Although some watcher types do not have type-specific arguments
648 (e.g. <code>ev_prepare</code>) you still need to call its <code>set</code> macro.</p>
650 <dt><code>ev_TYPE_init</code> (ev_TYPE *watcher, callback, [args])</dt>
652 <p>This convinience macro rolls both <code>ev_init</code> and <code>ev_TYPE_set</code> macro
653 calls into a single call. This is the most convinient method to initialise
654 a watcher. The same limitations apply, of course.</p>
656 <dt><code>ev_TYPE_start</code> (loop *, ev_TYPE *watcher)</dt>
658 <p>Starts (activates) the given watcher. Only active watchers will receive
659 events. If the watcher is already active nothing will happen.</p>
661 <dt><code>ev_TYPE_stop</code> (loop *, ev_TYPE *watcher)</dt>
663 <p>Stops the given watcher again (if active) and clears the pending
664 status. It is possible that stopped watchers are pending (for example,
665 non-repeating timers are being stopped when they become pending), but
666 <code>ev_TYPE_stop</code> ensures that the watcher is neither active nor pending. If
667 you want to free or reuse the memory used by the watcher it is therefore a
668 good idea to always call its <code>ev_TYPE_stop</code> function.</p>
670 <dt>bool ev_is_active (ev_TYPE *watcher)</dt>
672 <p>Returns a true value iff the watcher is active (i.e. it has been started
673 and not yet been stopped). As long as a watcher is active you must not modify
676 <dt>bool ev_is_pending (ev_TYPE *watcher)</dt>
678 <p>Returns a true value iff the watcher is pending, (i.e. it has outstanding
679 events but its callback has not yet been invoked). As long as a watcher
680 is pending (but not active) you must not call an init function on it (but
681 <code>ev_TYPE_set</code> is safe) and you must make sure the watcher is available to
682 libev (e.g. you cnanot <code>free ()</code> it).</p>
684 <dt>callback = ev_cb (ev_TYPE *watcher)</dt>
686 <p>Returns the callback currently set on the watcher.</p>
688 <dt>ev_cb_set (ev_TYPE *watcher, callback)</dt>
690 <p>Change the callback. You can change the callback at virtually any time
691 (modulo threads).</p>
700 <h2 id="ASSOCIATING_CUSTOM_DATA_WITH_A_WATCH">ASSOCIATING CUSTOM DATA WITH A WATCHER</h2>
701 <div id="ASSOCIATING_CUSTOM_DATA_WITH_A_WATCH-2">
702 <p>Each watcher has, by default, a member <code>void *data</code> that you can change
703 and read at any time, libev will completely ignore it. This can be used
704 to associate arbitrary data with your watcher. If you need more data and
705 don't want to allocate memory and store a pointer to it in that data
706 member, you can also "subclass" the watcher type and provide your own
713 struct whatever *mostinteresting;
717 <p>And since your callback will be called with a pointer to the watcher, you
718 can cast it back to your own type:</p>
719 <pre> static void my_cb (struct ev_loop *loop, struct ev_io *w_, int revents)
721 struct my_io *w = (struct my_io *)w_;
726 <p>More interesting and less C-conformant ways of catsing your callback type
727 have been omitted....</p>
734 <h1 id="WATCHER_TYPES">WATCHER TYPES</h1><p><a href="#TOP" class="toplink">Top</a></p>
735 <div id="WATCHER_TYPES_CONTENT">
736 <p>This section describes each watcher in detail, but will not repeat
737 information given in the last section.</p>
744 <h2 id="code_ev_io_code_is_this_file_descrip"><code>ev_io</code> - is this file descriptor readable or writable?</h2>
745 <div id="code_ev_io_code_is_this_file_descrip-2">
746 <p>I/O watchers check whether a file descriptor is readable or writable
747 in each iteration of the event loop, or, more precisely, when reading
748 would not block the process and writing would at least be able to write
749 some data. This behaviour is called level-triggering because you keep
750 receiving events as long as the condition persists. Remember you can stop
751 the watcher if you don't want to act on the event and neither want to
752 receive future events.</p>
753 <p>In general you can register as many read and/or write event watchers per
754 fd as you want (as long as you don't confuse yourself). Setting all file
755 descriptors to non-blocking mode is also usually a good idea (but not
756 required if you know what you are doing).</p>
757 <p>You have to be careful with dup'ed file descriptors, though. Some backends
758 (the linux epoll backend is a notable example) cannot handle dup'ed file
759 descriptors correctly if you register interest in two or more fds pointing
760 to the same underlying file/socket/etc. description (that is, they share
761 the same underlying "file open").</p>
762 <p>If you must do this, then force the use of a known-to-be-good backend
763 (at the time of this writing, this includes only <code>EVBACKEND_SELECT</code> and
764 <code>EVBACKEND_POLL</code>).</p>
765 <p>Another thing you have to watch out for is that it is quite easy to
766 receive "spurious" readyness notifications, that is your callback might
767 be called with <code>EV_READ</code> but a subsequent <code>read</code>(2) will actually block
768 because there is no data. Not only are some backends known to create a
769 lot of those (for example solaris ports), it is very easy to get into
770 this situation even with a relatively standard program structure. Thus
771 it is best to always use non-blocking I/O: An extra <code>read</code>(2) returning
772 <code>EAGAIN</code> is far preferable to a program hanging until some data arrives.</p>
773 <p>If you cannot run the fd in non-blocking mode (for example you should not
774 play around with an Xlib connection), then you have to seperately re-test
775 wether a file descriptor is really ready with a known-to-be good interface
776 such as poll (fortunately in our Xlib example, Xlib already does this on
777 its own, so its quite safe to use).</p>
779 <dt>ev_io_init (ev_io *, callback, int fd, int events)</dt>
780 <dt>ev_io_set (ev_io *, int fd, int events)</dt>
782 <p>Configures an <code>ev_io</code> watcher. The <code>fd</code> is the file descriptor to
783 rceeive events for and events is either <code>EV_READ</code>, <code>EV_WRITE</code> or
784 <code>EV_READ | EV_WRITE</code> to receive the given events.</p>
787 <p>Example: call <code>stdin_readable_cb</code> when STDIN_FILENO has become, well
788 readable, but only once. Since it is likely line-buffered, you could
789 attempt to read a whole line in the callback:</p>
791 stdin_readable_cb (struct ev_loop *loop, struct ev_io *w, int revents)
793 ev_io_stop (loop, w);
794 .. read from stdin here (or from w->fd) and haqndle any I/O errors
798 struct ev_loop *loop = ev_default_init (0);
799 struct ev_io stdin_readable;
800 ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
801 ev_io_start (loop, &stdin_readable);
810 <h2 id="code_ev_timer_code_relative_and_opti"><code>ev_timer</code> - relative and optionally repeating timeouts</h2>
811 <div id="code_ev_timer_code_relative_and_opti-2">
812 <p>Timer watchers are simple relative timers that generate an event after a
813 given time, and optionally repeating in regular intervals after that.</p>
814 <p>The timers are based on real time, that is, if you register an event that
815 times out after an hour and you reset your system clock to last years
816 time, it will still time out after (roughly) and hour. "Roughly" because
817 detecting time jumps is hard, and some inaccuracies are unavoidable (the
818 monotonic clock option helps a lot here).</p>
819 <p>The relative timeouts are calculated relative to the <code>ev_now ()</code>
820 time. This is usually the right thing as this timestamp refers to the time
821 of the event triggering whatever timeout you are modifying/starting. If
822 you suspect event processing to be delayed and you <i>need</i> to base the timeout
823 on the current time, use something like this to adjust for this:</p>
824 <pre> ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);
827 <p>The callback is guarenteed to be invoked only when its timeout has passed,
828 but if multiple timers become ready during the same loop iteration then
829 order of execution is undefined.</p>
831 <dt>ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)</dt>
832 <dt>ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)</dt>
834 <p>Configure the timer to trigger after <code>after</code> seconds. If <code>repeat</code> is
835 <code>0.</code>, then it will automatically be stopped. If it is positive, then the
836 timer will automatically be configured to trigger again <code>repeat</code> seconds
837 later, again, and again, until stopped manually.</p>
838 <p>The timer itself will do a best-effort at avoiding drift, that is, if you
839 configure a timer to trigger every 10 seconds, then it will trigger at
840 exactly 10 second intervals. If, however, your program cannot keep up with
841 the timer (because it takes longer than those 10 seconds to do stuff) the
842 timer will not fire more than once per event loop iteration.</p>
844 <dt>ev_timer_again (loop)</dt>
846 <p>This will act as if the timer timed out and restart it again if it is
847 repeating. The exact semantics are:</p>
848 <p>If the timer is started but nonrepeating, stop it.</p>
849 <p>If the timer is repeating, either start it if necessary (with the repeat
850 value), or reset the running timer to the repeat value.</p>
851 <p>This sounds a bit complicated, but here is a useful and typical
852 example: Imagine you have a tcp connection and you want a so-called idle
853 timeout, that is, you want to be called when there have been, say, 60
854 seconds of inactivity on the socket. The easiest way to do this is to
855 configure an <code>ev_timer</code> with after=repeat=60 and calling ev_timer_again each
856 time you successfully read or write some data. If you go into an idle
857 state where you do not expect data to travel on the socket, you can stop
858 the timer, and again will automatically restart it if need be.</p>
861 <p>Example: create a timer that fires after 60 seconds.</p>
863 one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
865 .. one minute over, w is actually stopped right here
868 struct ev_timer mytimer;
869 ev_timer_init (&mytimer, one_minute_cb, 60., 0.);
870 ev_timer_start (loop, &mytimer);
873 <p>Example: create a timeout timer that times out after 10 seconds of
876 timeout_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
878 .. ten seconds without any activity
881 struct ev_timer mytimer;
882 ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */
883 ev_timer_again (&mytimer); /* start timer */
886 // and in some piece of code that gets executed on any "activity":
887 // reset the timeout to start ticking again at 10 seconds
888 ev_timer_again (&mytimer);
896 <h2 id="code_ev_periodic_code_to_cron_or_not"><code>ev_periodic</code> - to cron or not to cron?</h2>
897 <div id="code_ev_periodic_code_to_cron_or_not-2">
898 <p>Periodic watchers are also timers of a kind, but they are very versatile
899 (and unfortunately a bit complex).</p>
900 <p>Unlike <code>ev_timer</code>'s, they are not based on real time (or relative time)
901 but on wallclock time (absolute time). You can tell a periodic watcher
902 to trigger "at" some specific point in time. For example, if you tell a
903 periodic watcher to trigger in 10 seconds (by specifiying e.g. <code>ev_now ()
904 + 10.</code>) and then reset your system clock to the last year, then it will
905 take a year to trigger the event (unlike an <code>ev_timer</code>, which would trigger
906 roughly 10 seconds later and of course not if you reset your system time
908 <p>They can also be used to implement vastly more complex timers, such as
909 triggering an event on eahc midnight, local time.</p>
910 <p>As with timers, the callback is guarenteed to be invoked only when the
911 time (<code>at</code>) has been passed, but if multiple periodic timers become ready
912 during the same loop iteration then order of execution is undefined.</p>
914 <dt>ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)</dt>
915 <dt>ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)</dt>
917 <p>Lots of arguments, lets sort it out... There are basically three modes of
918 operation, and we will explain them from simplest to complex:</p>
921 <dt>* absolute timer (interval = reschedule_cb = 0)</dt>
923 <p>In this configuration the watcher triggers an event at the wallclock time
924 <code>at</code> and doesn't repeat. It will not adjust when a time jump occurs,
925 that is, if it is to be run at January 1st 2011 then it will run when the
926 system time reaches or surpasses this time.</p>
928 <dt>* non-repeating interval timer (interval > 0, reschedule_cb = 0)</dt>
930 <p>In this mode the watcher will always be scheduled to time out at the next
931 <code>at + N * interval</code> time (for some integer N) and then repeat, regardless
932 of any time jumps.</p>
933 <p>This can be used to create timers that do not drift with respect to system
935 <pre> ev_periodic_set (&periodic, 0., 3600., 0);
938 <p>This doesn't mean there will always be 3600 seconds in between triggers,
939 but only that the the callback will be called when the system time shows a
940 full hour (UTC), or more correctly, when the system time is evenly divisible
942 <p>Another way to think about it (for the mathematically inclined) is that
943 <code>ev_periodic</code> will try to run the callback in this mode at the next possible
944 time where <code>time = at (mod interval)</code>, regardless of any time jumps.</p>
946 <dt>* manual reschedule mode (reschedule_cb = callback)</dt>
948 <p>In this mode the values for <code>interval</code> and <code>at</code> are both being
949 ignored. Instead, each time the periodic watcher gets scheduled, the
950 reschedule callback will be called with the watcher as first, and the
951 current time as second argument.</p>
952 <p>NOTE: <i>This callback MUST NOT stop or destroy any periodic watcher,
953 ever, or make any event loop modifications</i>. If you need to stop it,
954 return <code>now + 1e30</code> (or so, fudge fudge) and stop it afterwards (e.g. by
955 starting a prepare watcher).</p>
956 <p>Its prototype is <code>ev_tstamp (*reschedule_cb)(struct ev_periodic *w,
957 ev_tstamp now)</code>, e.g.:</p>
958 <pre> static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)
964 <p>It must return the next time to trigger, based on the passed time value
965 (that is, the lowest time value larger than to the second argument). It
966 will usually be called just before the callback will be triggered, but
967 might be called at other times, too.</p>
968 <p>NOTE: <i>This callback must always return a time that is later than the
969 passed <code>now</code> value</i>. Not even <code>now</code> itself will do, it <i>must</i> be larger.</p>
970 <p>This can be used to create very complex timers, such as a timer that
971 triggers on each midnight, local time. To do this, you would calculate the
972 next midnight after <code>now</code> and return the timestamp value for this. How
973 you do this is, again, up to you (but it is not trivial, which is the main
974 reason I omitted it as an example).</p>
979 <dt>ev_periodic_again (loop, ev_periodic *)</dt>
981 <p>Simply stops and restarts the periodic watcher again. This is only useful
982 when you changed some parameters or the reschedule callback would return
983 a different time than the last time it was called (e.g. in a crond like
984 program when the crontabs have changed).</p>
987 <p>Example: call a callback every hour, or, more precisely, whenever the
988 system clock is divisible by 3600. The callback invocation times have
989 potentially a lot of jittering, but good long-term stability.</p>
991 clock_cb (struct ev_loop *loop, struct ev_io *w, int revents)
993 ... its now a full hour (UTC, or TAI or whatever your clock follows)
996 struct ev_periodic hourly_tick;
997 ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);
998 ev_periodic_start (loop, &hourly_tick);
1001 <p>Example: the same as above, but use a reschedule callback to do it:</p>
1002 <pre> #include <math.h>
1005 my_scheduler_cb (struct ev_periodic *w, ev_tstamp now)
1007 return fmod (now, 3600.) + 3600.;
1010 ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
1013 <p>Example: call a callback every hour, starting now:</p>
1014 <pre> struct ev_periodic hourly_tick;
1015 ev_periodic_init (&hourly_tick, clock_cb,
1016 fmod (ev_now (loop), 3600.), 3600., 0);
1017 ev_periodic_start (loop, &hourly_tick);
1025 <h2 id="code_ev_signal_code_signal_me_when_a"><code>ev_signal</code> - signal me when a signal gets signalled!</h2>
1026 <div id="code_ev_signal_code_signal_me_when_a-2">
1027 <p>Signal watchers will trigger an event when the process receives a specific
1028 signal one or more times. Even though signals are very asynchronous, libev
1029 will try it's best to deliver signals synchronously, i.e. as part of the
1030 normal event processing, like any other event.</p>
1031 <p>You can configure as many watchers as you like per signal. Only when the
1032 first watcher gets started will libev actually register a signal watcher
1033 with the kernel (thus it coexists with your own signal handlers as long
1034 as you don't register any with libev). Similarly, when the last signal
1035 watcher for a signal is stopped libev will reset the signal handler to
1036 SIG_DFL (regardless of what it was set to before).</p>
1038 <dt>ev_signal_init (ev_signal *, callback, int signum)</dt>
1039 <dt>ev_signal_set (ev_signal *, int signum)</dt>
1041 <p>Configures the watcher to trigger on the given signal number (usually one
1042 of the <code>SIGxxx</code> constants).</p>
1051 <h2 id="code_ev_child_code_watch_out_for_pro"><code>ev_child</code> - watch out for process status changes</h2>
1052 <div id="code_ev_child_code_watch_out_for_pro-2">
1053 <p>Child watchers trigger when your process receives a SIGCHLD in response to
1054 some child status changes (most typically when a child of yours dies).</p>
1056 <dt>ev_child_init (ev_child *, callback, int pid)</dt>
1057 <dt>ev_child_set (ev_child *, int pid)</dt>
1059 <p>Configures the watcher to wait for status changes of process <code>pid</code> (or
1060 <i>any</i> process if <code>pid</code> is specified as <code>0</code>). The callback can look
1061 at the <code>rstatus</code> member of the <code>ev_child</code> watcher structure to see
1062 the status word (use the macros from <code>sys/wait.h</code> and see your systems
1063 <code>waitpid</code> documentation). The <code>rpid</code> member contains the pid of the
1064 process causing the status change.</p>
1067 <p>Example: try to exit cleanly on SIGINT and SIGTERM.</p>
1069 sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)
1071 ev_unloop (loop, EVUNLOOP_ALL);
1074 struct ev_signal signal_watcher;
1075 ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
1076 ev_signal_start (loop, &sigint_cb);
1084 <h2 id="code_ev_idle_code_when_you_ve_got_no"><code>ev_idle</code> - when you've got nothing better to do...</h2>
1085 <div id="code_ev_idle_code_when_you_ve_got_no-2">
1086 <p>Idle watchers trigger events when there are no other events are pending
1087 (prepare, check and other idle watchers do not count). That is, as long
1088 as your process is busy handling sockets or timeouts (or even signals,
1089 imagine) it will not be triggered. But when your process is idle all idle
1090 watchers are being called again and again, once per event loop iteration -
1091 until stopped, that is, or your process receives more events and becomes
1093 <p>The most noteworthy effect is that as long as any idle watchers are
1094 active, the process will not block when waiting for new events.</p>
1095 <p>Apart from keeping your process non-blocking (which is a useful
1096 effect on its own sometimes), idle watchers are a good place to do
1097 "pseudo-background processing", or delay processing stuff to after the
1098 event loop has handled all outstanding events.</p>
1100 <dt>ev_idle_init (ev_signal *, callback)</dt>
1102 <p>Initialises and configures the idle watcher - it has no parameters of any
1103 kind. There is a <code>ev_idle_set</code> macro, but using it is utterly pointless,
1107 <p>Example: dynamically allocate an <code>ev_idle</code>, start it, and in the
1108 callback, free it. Alos, use no error checking, as usual.</p>
1110 idle_cb (struct ev_loop *loop, struct ev_idle *w, int revents)
1113 // now do something you wanted to do when the program has
1114 // no longer asnything immediate to do.
1117 struct ev_idle *idle_watcher = malloc (sizeof (struct ev_idle));
1118 ev_idle_init (idle_watcher, idle_cb);
1119 ev_idle_start (loop, idle_cb);
1127 <h2 id="code_ev_prepare_code_and_code_ev_che"><code>ev_prepare</code> and <code>ev_check</code> - customise your event loop!</h2>
1128 <div id="code_ev_prepare_code_and_code_ev_che-2">
1129 <p>Prepare and check watchers are usually (but not always) used in tandem:
1130 prepare watchers get invoked before the process blocks and check watchers
1132 <p>You <i>must not</i> call <code>ev_loop</code> or similar functions that enter
1133 the current event loop from either <code>ev_prepare</code> or <code>ev_check</code>
1134 watchers. Other loops than the current one are fine, however. The
1135 rationale behind this is that you do not need to check for recursion in
1136 those watchers, i.e. the sequence will always be <code>ev_prepare</code>, blocking,
1137 <code>ev_check</code> so if you have one watcher of each kind they will always be
1138 called in pairs bracketing the blocking call.</p>
1139 <p>Their main purpose is to integrate other event mechanisms into libev and
1140 their use is somewhat advanced. This could be used, for example, to track
1141 variable changes, implement your own watchers, integrate net-snmp or a
1142 coroutine library and lots more. They are also occasionally useful if
1143 you cache some data and want to flush it before blocking (for example,
1144 in X programs you might want to do an <code>XFlush ()</code> in an <code>ev_prepare</code>
1146 <p>This is done by examining in each prepare call which file descriptors need
1147 to be watched by the other library, registering <code>ev_io</code> watchers for
1148 them and starting an <code>ev_timer</code> watcher for any timeouts (many libraries
1149 provide just this functionality). Then, in the check watcher you check for
1150 any events that occured (by checking the pending status of all watchers
1151 and stopping them) and call back into the library. The I/O and timer
1152 callbacks will never actually be called (but must be valid nevertheless,
1153 because you never know, you know?).</p>
1154 <p>As another example, the Perl Coro module uses these hooks to integrate
1155 coroutines into libev programs, by yielding to other active coroutines
1156 during each prepare and only letting the process block if no coroutines
1157 are ready to run (it's actually more complicated: it only runs coroutines
1158 with priority higher than or equal to the event loop and one coroutine
1159 of lower priority, but only once, using idle watchers to keep the event
1160 loop from blocking if lower-priority coroutines are active, thus mapping
1161 low-priority coroutines to idle/background tasks).</p>
1163 <dt>ev_prepare_init (ev_prepare *, callback)</dt>
1164 <dt>ev_check_init (ev_check *, callback)</dt>
1166 <p>Initialises and configures the prepare or check watcher - they have no
1167 parameters of any kind. There are <code>ev_prepare_set</code> and <code>ev_check_set</code>
1168 macros, but using them is utterly, utterly and completely pointless.</p>
1171 <p>Example: To include a library such as adns, you would add IO watchers
1172 and a timeout watcher in a prepare handler, as required by libadns, and
1173 in a check watcher, destroy them and call into libadns. What follows is
1174 pseudo-code only of course:</p>
1175 <pre> static ev_io iow [nfd];
1179 io_cb (ev_loop *loop, ev_io *w, int revents)
1181 // set the relevant poll flags
1182 // could also call adns_processreadable etc. here
1183 struct pollfd *fd = (struct pollfd *)w->data;
1184 if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;
1185 if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;
1188 // create io watchers for each fd and a timer before blocking
1190 adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents)
1192 int timeout = 3600000;truct pollfd fds [nfd];
1193 // actual code will need to loop here and realloc etc.
1194 adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));
1196 /* the callback is illegal, but won't be called as we stop during check */
1197 ev_timer_init (&tw, 0, timeout * 1e-3);
1198 ev_timer_start (loop, &tw);
1200 // create on ev_io per pollfd
1201 for (int i = 0; i < nfd; ++i)
1203 ev_io_init (iow + i, io_cb, fds [i].fd,
1204 ((fds [i].events & POLLIN ? EV_READ : 0)
1205 | (fds [i].events & POLLOUT ? EV_WRITE : 0)));
1207 fds [i].revents = 0;
1208 iow [i].data = fds + i;
1209 ev_io_start (loop, iow + i);
1213 // stop all watchers after blocking
1215 adns_check_cb (ev_loop *loop, ev_check *w, int revents)
1217 ev_timer_stop (loop, &tw);
1219 for (int i = 0; i < nfd; ++i)
1220 ev_io_stop (loop, iow + i);
1222 adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));
1231 <h2 id="code_ev_embed_code_when_one_backend_"><code>ev_embed</code> - when one backend isn't enough...</h2>
1232 <div id="code_ev_embed_code_when_one_backend_-2">
1233 <p>This is a rather advanced watcher type that lets you embed one event loop
1234 into another (currently only <code>ev_io</code> events are supported in the embedded
1235 loop, other types of watchers might be handled in a delayed or incorrect
1236 fashion and must not be used).</p>
1237 <p>There are primarily two reasons you would want that: work around bugs and
1239 <p>As an example for a bug workaround, the kqueue backend might only support
1240 sockets on some platform, so it is unusable as generic backend, but you
1241 still want to make use of it because you have many sockets and it scales
1242 so nicely. In this case, you would create a kqueue-based loop and embed it
1243 into your default loop (which might use e.g. poll). Overall operation will
1244 be a bit slower because first libev has to poll and then call kevent, but
1245 at least you can use both at what they are best.</p>
1246 <p>As for prioritising I/O: rarely you have the case where some fds have
1247 to be watched and handled very quickly (with low latency), and even
1248 priorities and idle watchers might have too much overhead. In this case
1249 you would put all the high priority stuff in one loop and all the rest in
1250 a second one, and embed the second one in the first.</p>
1251 <p>As long as the watcher is active, the callback will be invoked every time
1252 there might be events pending in the embedded loop. The callback must then
1253 call <code>ev_embed_sweep (mainloop, watcher)</code> to make a single sweep and invoke
1254 their callbacks (you could also start an idle watcher to give the embedded
1255 loop strictly lower priority for example). You can also set the callback
1256 to <code>0</code>, in which case the embed watcher will automatically execute the
1257 embedded loop sweep.</p>
1258 <p>As long as the watcher is started it will automatically handle events. The
1259 callback will be invoked whenever some events have been handled. You can
1260 set the callback to <code>0</code> to avoid having to specify one if you are not
1261 interested in that.</p>
1262 <p>Also, there have not currently been made special provisions for forking:
1263 when you fork, you not only have to call <code>ev_loop_fork</code> on both loops,
1264 but you will also have to stop and restart any <code>ev_embed</code> watchers
1266 <p>Unfortunately, not all backends are embeddable, only the ones returned by
1267 <code>ev_embeddable_backends</code> are, which, unfortunately, does not include any
1269 <p>So when you want to use this feature you will always have to be prepared
1270 that you cannot get an embeddable loop. The recommended way to get around
1271 this is to have a separate variables for your embeddable loop, try to
1272 create it, and if that fails, use the normal loop for everything:</p>
1273 <pre> struct ev_loop *loop_hi = ev_default_init (0);
1274 struct ev_loop *loop_lo = 0;
1275 struct ev_embed embed;
1277 // see if there is a chance of getting one that works
1278 // (remember that a flags value of 0 means autodetection)
1279 loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
1280 ? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
1283 // if we got one, then embed it, otherwise default to loop_hi
1286 ev_embed_init (&embed, 0, loop_lo);
1287 ev_embed_start (loop_hi, &embed);
1294 <dt>ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)</dt>
1295 <dt>ev_embed_set (ev_embed *, callback, struct ev_loop *embedded_loop)</dt>
1297 <p>Configures the watcher to embed the given loop, which must be
1298 embeddable. If the callback is <code>0</code>, then <code>ev_embed_sweep</code> will be
1299 invoked automatically, otherwise it is the responsibility of the callback
1300 to invoke it (it will continue to be called until the sweep has been done,
1301 if you do not want thta, you need to temporarily stop the embed watcher).</p>
1303 <dt>ev_embed_sweep (loop, ev_embed *)</dt>
1305 <p>Make a single, non-blocking sweep over the embedded loop. This works
1306 similarly to <code>ev_loop (embedded_loop, EVLOOP_NONBLOCK)</code>, but in the most
1307 apropriate way for embedded loops.</p>
1316 <h1 id="OTHER_FUNCTIONS">OTHER FUNCTIONS</h1><p><a href="#TOP" class="toplink">Top</a></p>
1317 <div id="OTHER_FUNCTIONS_CONTENT">
1318 <p>There are some other functions of possible interest. Described. Here. Now.</p>
1320 <dt>ev_once (loop, int fd, int events, ev_tstamp timeout, callback)</dt>
1322 <p>This function combines a simple timer and an I/O watcher, calls your
1323 callback on whichever event happens first and automatically stop both
1324 watchers. This is useful if you want to wait for a single event on an fd
1325 or timeout without having to allocate/configure/start/stop/free one or
1326 more watchers yourself.</p>
1327 <p>If <code>fd</code> is less than 0, then no I/O watcher will be started and events
1328 is being ignored. Otherwise, an <code>ev_io</code> watcher for the given <code>fd</code> and
1329 <code>events</code> set will be craeted and started.</p>
1330 <p>If <code>timeout</code> is less than 0, then no timeout watcher will be
1331 started. Otherwise an <code>ev_timer</code> watcher with after = <code>timeout</code> (and
1332 repeat = 0) will be started. While <code>0</code> is a valid timeout, it is of
1334 <p>The callback has the type <code>void (*cb)(int revents, void *arg)</code> and gets
1335 passed an <code>revents</code> set like normal event callbacks (a combination of
1336 <code>EV_ERROR</code>, <code>EV_READ</code>, <code>EV_WRITE</code> or <code>EV_TIMEOUT</code>) and the <code>arg</code>
1337 value passed to <code>ev_once</code>:</p>
1338 <pre> static void stdin_ready (int revents, void *arg)
1340 if (revents & EV_TIMEOUT)
1341 /* doh, nothing entered */;
1342 else if (revents & EV_READ)
1343 /* stdin might have data for us, joy! */;
1346 ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
1350 <dt>ev_feed_event (ev_loop *, watcher *, int revents)</dt>
1352 <p>Feeds the given event set into the event loop, as if the specified event
1353 had happened for the specified watcher (which must be a pointer to an
1354 initialised but not necessarily started event watcher).</p>
1356 <dt>ev_feed_fd_event (ev_loop *, int fd, int revents)</dt>
1358 <p>Feed an event on the given fd, as if a file descriptor backend detected
1359 the given events it.</p>
1361 <dt>ev_feed_signal_event (ev_loop *loop, int signum)</dt>
1363 <p>Feed an event as if the given signal occured (<code>loop</code> must be the default
1373 <h1 id="LIBEVENT_EMULATION">LIBEVENT EMULATION</h1><p><a href="#TOP" class="toplink">Top</a></p>
1374 <div id="LIBEVENT_EMULATION_CONTENT">
1375 <p>Libev offers a compatibility emulation layer for libevent. It cannot
1376 emulate the internals of libevent, so here are some usage hints:</p>
1378 <dt>* Use it by including <event.h>, as usual.</dt>
1379 <dt>* The following members are fully supported: ev_base, ev_callback,
1380 ev_arg, ev_fd, ev_res, ev_events.</dt>
1381 <dt>* Avoid using ev_flags and the EVLIST_*-macros, while it is
1382 maintained by libev, it does not work exactly the same way as in libevent (consider
1383 it a private API).</dt>
1384 <dt>* Priorities are not currently supported. Initialising priorities
1385 will fail and all watchers will have the same priority, even though there
1386 is an ev_pri field.</dt>
1387 <dt>* Other members are not supported.</dt>
1388 <dt>* The libev emulation is <i>not</i> ABI compatible to libevent, you need
1389 to use the libev header file and library.</dt>
1393 <h1 id="C_SUPPORT">C++ SUPPORT</h1><p><a href="#TOP" class="toplink">Top</a></p>
1394 <div id="C_SUPPORT_CONTENT">
1395 <p>Libev comes with some simplistic wrapper classes for C++ that mainly allow
1396 you to use some convinience methods to start/stop watchers and also change
1397 the callback model to a model using method callbacks on objects.</p>
1399 <pre> #include <ev++.h>
1402 <p>(it is not installed by default). This automatically includes <cite>ev.h</cite>
1403 and puts all of its definitions (many of them macros) into the global
1404 namespace. All C++ specific things are put into the <code>ev</code> namespace.</p>
1405 <p>It should support all the same embedding options as <cite>ev.h</cite>, most notably
1406 <code>EV_MULTIPLICITY</code>.</p>
1407 <p>Here is a list of things available in the <code>ev</code> namespace:</p>
1409 <dt><code>ev::READ</code>, <code>ev::WRITE</code> etc.</dt>
1411 <p>These are just enum values with the same values as the <code>EV_READ</code> etc.
1412 macros from <cite>ev.h</cite>.</p>
1414 <dt><code>ev::tstamp</code>, <code>ev::now</code></dt>
1416 <p>Aliases to the same types/functions as with the <code>ev_</code> prefix.</p>
1418 <dt><code>ev::io</code>, <code>ev::timer</code>, <code>ev::periodic</code>, <code>ev::idle</code>, <code>ev::sig</code> etc.</dt>
1420 <p>For each <code>ev_TYPE</code> watcher in <cite>ev.h</cite> there is a corresponding class of
1421 the same name in the <code>ev</code> namespace, with the exception of <code>ev_signal</code>
1422 which is called <code>ev::sig</code> to avoid clashes with the <code>signal</code> macro
1423 defines by many implementations.</p>
1424 <p>All of those classes have these methods:</p>
1427 <dt>ev::TYPE::TYPE (object *, object::method *)</dt>
1428 <dt>ev::TYPE::TYPE (object *, object::method *, struct ev_loop *)</dt>
1429 <dt>ev::TYPE::~TYPE</dt>
1431 <p>The constructor takes a pointer to an object and a method pointer to
1432 the event handler callback to call in this class. The constructor calls
1433 <code>ev_init</code> for you, which means you have to call the <code>set</code> method
1434 before starting it. If you do not specify a loop then the constructor
1435 automatically associates the default loop with this watcher.</p>
1436 <p>The destructor automatically stops the watcher if it is active.</p>
1438 <dt>w->set (struct ev_loop *)</dt>
1440 <p>Associates a different <code>struct ev_loop</code> with this watcher. You can only
1441 do this when the watcher is inactive (and not pending either).</p>
1443 <dt>w->set ([args])</dt>
1445 <p>Basically the same as <code>ev_TYPE_set</code>, with the same args. Must be
1446 called at least once. Unlike the C counterpart, an active watcher gets
1447 automatically stopped and restarted.</p>
1449 <dt>w->start ()</dt>
1451 <p>Starts the watcher. Note that there is no <code>loop</code> argument as the
1452 constructor already takes the loop.</p>
1454 <dt>w->stop ()</dt>
1456 <p>Stops the watcher if it is active. Again, no <code>loop</code> argument.</p>
1458 <dt>w->again () <code>ev::timer</code>, <code>ev::periodic</code> only</dt>
1460 <p>For <code>ev::timer</code> and <code>ev::periodic</code>, this invokes the corresponding
1461 <code>ev_TYPE_again</code> function.</p>
1463 <dt>w->sweep () <code>ev::embed</code> only</dt>
1465 <p>Invokes <code>ev_embed_sweep</code>.</p>
1471 <p>Example: Define a class with an IO and idle watcher, start one of them in
1472 the constructor.</p>
1475 ev_io io; void io_cb (ev::io &w, int revents);
1476 ev_idle idle void idle_cb (ev::idle &w, int revents);
1481 myclass::myclass (int fd)
1482 : io (this, &myclass::io_cb),
1483 idle (this, &myclass::idle_cb)
1485 io.start (fd, ev::READ);
1491 <h1 id="EMBEDDING">EMBEDDING</h1><p><a href="#TOP" class="toplink">Top</a></p>
1492 <div id="EMBEDDING_CONTENT">
1493 <p>Libev can (and often is) directly embedded into host
1494 applications. Examples of applications that embed it include the Deliantra
1495 Game Server, the EV perl module, the GNU Virtual Private Ethernet (gvpe)
1496 and rxvt-unicode.</p>
1497 <p>The goal is to enable you to just copy the neecssary files into your
1498 source directory without having to change even a single line in them, so
1499 you can easily upgrade by simply copying (or having a checked-out copy of
1500 libev somewhere in your source tree).</p>
1503 <h2 id="FILESETS">FILESETS</h2>
1504 <div id="FILESETS_CONTENT">
1505 <p>Depending on what features you need you need to include one or more sets of files
1509 <h3 id="CORE_EVENT_LOOP">CORE EVENT LOOP</h3>
1510 <div id="CORE_EVENT_LOOP_CONTENT">
1511 <p>To include only the libev core (all the <code>ev_*</code> functions), with manual
1512 configuration (no autoconf):</p>
1513 <pre> #define EV_STANDALONE 1
1514 #include "ev.c"
1517 <p>This will automatically include <cite>ev.h</cite>, too, and should be done in a
1518 single C source file only to provide the function implementations. To use
1519 it, do the same for <cite>ev.h</cite> in all files wishing to use this API (best
1520 done by writing a wrapper around <cite>ev.h</cite> that you can include instead and
1521 where you can put other configuration options):</p>
1522 <pre> #define EV_STANDALONE 1
1523 #include "ev.h"
1526 <p>Both header files and implementation files can be compiled with a C++
1527 compiler (at least, thats a stated goal, and breakage will be treated
1529 <p>You need the following files in your source tree, or in a directory
1530 in your include path (e.g. in libev/ when using -Ilibev):</p>
1536 ev_win32.c required on win32 platforms only
1538 ev_select.c only when select backend is enabled (which is by default)
1539 ev_poll.c only when poll backend is enabled (disabled by default)
1540 ev_epoll.c only when the epoll backend is enabled (disabled by default)
1541 ev_kqueue.c only when the kqueue backend is enabled (disabled by default)
1542 ev_port.c only when the solaris port backend is enabled (disabled by default)
1545 <p><cite>ev.c</cite> includes the backend files directly when enabled, so you only need
1546 to compile this single file.</p>
1549 <h3 id="LIBEVENT_COMPATIBILITY_API">LIBEVENT COMPATIBILITY API</h3>
1550 <div id="LIBEVENT_COMPATIBILITY_API_CONTENT">
1551 <p>To include the libevent compatibility API, also include:</p>
1552 <pre> #include "event.c"
1555 <p>in the file including <cite>ev.c</cite>, and:</p>
1556 <pre> #include "event.h"
1559 <p>in the files that want to use the libevent API. This also includes <cite>ev.h</cite>.</p>
1560 <p>You need the following additional files for this:</p>
1567 <h3 id="AUTOCONF_SUPPORT">AUTOCONF SUPPORT</h3>
1568 <div id="AUTOCONF_SUPPORT_CONTENT">
1569 <p>Instead of using <code>EV_STANDALONE=1</code> and providing your config in
1570 whatever way you want, you can also <code>m4_include([libev.m4])</code> in your
1571 <cite>configure.ac</cite> and leave <code>EV_STANDALONE</code> undefined. <cite>ev.c</cite> will then
1572 include <cite>config.h</cite> and configure itself accordingly.</p>
1573 <p>For this of course you need the m4 file:</p>
1579 <h2 id="PREPROCESSOR_SYMBOLS_MACROS">PREPROCESSOR SYMBOLS/MACROS</h2>
1580 <div id="PREPROCESSOR_SYMBOLS_MACROS_CONTENT">
1581 <p>Libev can be configured via a variety of preprocessor symbols you have to define
1582 before including any of its files. The default is not to build for multiplicity
1583 and only include the select backend.</p>
1585 <dt>EV_STANDALONE</dt>
1587 <p>Must always be <code>1</code> if you do not use autoconf configuration, which
1588 keeps libev from including <cite>config.h</cite>, and it also defines dummy
1589 implementations for some libevent functions (such as logging, which is not
1590 supported). It will also not define any of the structs usually found in
1591 <cite>event.h</cite> that are not directly supported by the libev core alone.</p>
1593 <dt>EV_USE_MONOTONIC</dt>
1595 <p>If defined to be <code>1</code>, libev will try to detect the availability of the
1596 monotonic clock option at both compiletime and runtime. Otherwise no use
1597 of the monotonic clock option will be attempted. If you enable this, you
1598 usually have to link against librt or something similar. Enabling it when
1599 the functionality isn't available is safe, though, althoguh you have
1600 to make sure you link against any libraries where the <code>clock_gettime</code>
1601 function is hiding in (often <cite>-lrt</cite>).</p>
1603 <dt>EV_USE_REALTIME</dt>
1605 <p>If defined to be <code>1</code>, libev will try to detect the availability of the
1606 realtime clock option at compiletime (and assume its availability at
1607 runtime if successful). Otherwise no use of the realtime clock option will
1608 be attempted. This effectively replaces <code>gettimeofday</code> by <code>clock_get
1609 (CLOCK_REALTIME, ...)</code> and will not normally affect correctness. See tzhe note about libraries
1610 in the description of <code>EV_USE_MONOTONIC</code>, though.</p>
1612 <dt>EV_USE_SELECT</dt>
1614 <p>If undefined or defined to be <code>1</code>, libev will compile in support for the
1615 <code>select</code>(2) backend. No attempt at autodetection will be done: if no
1616 other method takes over, select will be it. Otherwise the select backend
1617 will not be compiled in.</p>
1619 <dt>EV_SELECT_USE_FD_SET</dt>
1621 <p>If defined to <code>1</code>, then the select backend will use the system <code>fd_set</code>
1622 structure. This is useful if libev doesn't compile due to a missing
1623 <code>NFDBITS</code> or <code>fd_mask</code> definition or it misguesses the bitset layout on
1624 exotic systems. This usually limits the range of file descriptors to some
1625 low limit such as 1024 or might have other limitations (winsocket only
1626 allows 64 sockets). The <code>FD_SETSIZE</code> macro, set before compilation, might
1627 influence the size of the <code>fd_set</code> used.</p>
1629 <dt>EV_SELECT_IS_WINSOCKET</dt>
1631 <p>When defined to <code>1</code>, the select backend will assume that
1632 select/socket/connect etc. don't understand file descriptors but
1633 wants osf handles on win32 (this is the case when the select to
1634 be used is the winsock select). This means that it will call
1635 <code>_get_osfhandle</code> on the fd to convert it to an OS handle. Otherwise,
1636 it is assumed that all these functions actually work on fds, even
1637 on win32. Should not be defined on non-win32 platforms.</p>
1639 <dt>EV_USE_POLL</dt>
1641 <p>If defined to be <code>1</code>, libev will compile in support for the <code>poll</code>(2)
1642 backend. Otherwise it will be enabled on non-win32 platforms. It
1643 takes precedence over select.</p>
1645 <dt>EV_USE_EPOLL</dt>
1647 <p>If defined to be <code>1</code>, libev will compile in support for the Linux
1648 <code>epoll</code>(7) backend. Its availability will be detected at runtime,
1649 otherwise another method will be used as fallback. This is the
1650 preferred backend for GNU/Linux systems.</p>
1652 <dt>EV_USE_KQUEUE</dt>
1654 <p>If defined to be <code>1</code>, libev will compile in support for the BSD style
1655 <code>kqueue</code>(2) backend. Its actual availability will be detected at runtime,
1656 otherwise another method will be used as fallback. This is the preferred
1657 backend for BSD and BSD-like systems, although on most BSDs kqueue only
1658 supports some types of fds correctly (the only platform we found that
1659 supports ptys for example was NetBSD), so kqueue might be compiled in, but
1660 not be used unless explicitly requested. The best way to use it is to find
1661 out whether kqueue supports your type of fd properly and use an embedded
1664 <dt>EV_USE_PORT</dt>
1666 <p>If defined to be <code>1</code>, libev will compile in support for the Solaris
1667 10 port style backend. Its availability will be detected at runtime,
1668 otherwise another method will be used as fallback. This is the preferred
1669 backend for Solaris 10 systems.</p>
1671 <dt>EV_USE_DEVPOLL</dt>
1673 <p>reserved for future expansion, works like the USE symbols above.</p>
1677 <p>The name of the <cite>ev.h</cite> header file used to include it. The default if
1678 undefined is <code><ev.h></code> in <cite>event.h</cite> and <code>"ev.h"</code> in <cite>ev.c</cite>. This
1679 can be used to virtually rename the <cite>ev.h</cite> header file in case of conflicts.</p>
1681 <dt>EV_CONFIG_H</dt>
1683 <p>If <code>EV_STANDALONE</code> isn't <code>1</code>, this variable can be used to override
1684 <cite>ev.c</cite>'s idea of where to find the <cite>config.h</cite> file, similarly to
1685 <code>EV_H</code>, above.</p>
1689 <p>Similarly to <code>EV_H</code>, this macro can be used to override <cite>event.c</cite>'s idea
1690 of how the <cite>event.h</cite> header can be found.</p>
1692 <dt>EV_PROTOTYPES</dt>
1694 <p>If defined to be <code>0</code>, then <cite>ev.h</cite> will not define any function
1695 prototypes, but still define all the structs and other symbols. This is
1696 occasionally useful if you want to provide your own wrapper functions
1697 around libev functions.</p>
1699 <dt>EV_MULTIPLICITY</dt>
1701 <p>If undefined or defined to <code>1</code>, then all event-loop-specific functions
1702 will have the <code>struct ev_loop *</code> as first argument, and you can create
1703 additional independent event loops. Otherwise there will be no support
1704 for multiple event loops and there is no first event loop pointer
1705 argument. Instead, all functions act on the single default loop.</p>
1707 <dt>EV_PERIODICS</dt>
1709 <p>If undefined or defined to be <code>1</code>, then periodic timers are supported,
1710 otherwise not. This saves a few kb of code.</p>
1714 <p>By default, all watchers have a <code>void *data</code> member. By redefining
1715 this macro to a something else you can include more and other types of
1716 members. You have to define it each time you include one of the files,
1717 though, and it must be identical each time.</p>
1718 <p>For example, the perl EV module uses something like this:</p>
1719 <pre> #define EV_COMMON \
1720 SV *self; /* contains this struct */ \
1721 SV *cb_sv, *fh /* note no trailing ";" */
1725 <dt>EV_CB_DECLARE (type)</dt>
1726 <dt>EV_CB_INVOKE (watcher, revents)</dt>
1727 <dt>ev_set_cb (ev, cb)</dt>
1729 <p>Can be used to change the callback member declaration in each watcher,
1730 and the way callbacks are invoked and set. Must expand to a struct member
1731 definition and a statement, respectively. See the <cite>ev.v</cite> header file for
1732 their default definitions. One possible use for overriding these is to
1733 avoid the <code>struct ev_loop *</code> as first argument in all cases, or to use
1734 method calls instead of plain function calls in C++.</p>
1737 <h2 id="EXAMPLES">EXAMPLES</h2>
1738 <div id="EXAMPLES_CONTENT">
1739 <p>For a real-world example of a program the includes libev
1740 verbatim, you can have a look at the EV perl module
1741 (<a href="http://software.schmorp.de/pkg/EV.html">http://software.schmorp.de/pkg/EV.html</a>). It has the libev files in
1742 the <cite>libev/</cite> subdirectory and includes them in the <cite>EV/EVAPI.h</cite> (public
1743 interface) and <cite>EV.xs</cite> (implementation) files. Only the <cite>EV.xs</cite> file
1744 will be compiled. It is pretty complex because it provides its own header
1746 <p>The usage in rxvt-unicode is simpler. It has a <cite>ev_cpp.h</cite> header file
1747 that everybody includes and which overrides some autoconf choices:</p>
1748 <pre> #define EV_USE_POLL 0
1749 #define EV_MULTIPLICITY 0
1750 #define EV_PERIODICS 0
1751 #define EV_CONFIG_H <config.h>
1753 #include "ev++.h"
1756 <p>And a <cite>ev_cpp.C</cite> implementation file that contains libev proper and is compiled:</p>
1757 <pre> #include "ev_cpp.h"
1758 #include "ev.c"
1766 <h1 id="COMPLEXITIES">COMPLEXITIES</h1><p><a href="#TOP" class="toplink">Top</a></p>
1767 <div id="COMPLEXITIES_CONTENT">
1768 <p>In this section the complexities of (many of) the algorithms used inside
1769 libev will be explained. For complexity discussions about backends see the
1770 documentation for <code>ev_default_init</code>.</p>
1773 <dt>Starting and stopping timer/periodic watchers: O(log skipped_other_timers)</dt>
1774 <dt>Changing timer/periodic watchers (by autorepeat, again): O(log skipped_other_timers)</dt>
1775 <dt>Starting io/check/prepare/idle/signal/child watchers: O(1)</dt>
1776 <dt>Stopping check/prepare/idle watchers: O(1)</dt>
1777 <dt>Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % 16))</dt>
1778 <dt>Finding the next timer per loop iteration: O(1)</dt>
1779 <dt>Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)</dt>
1780 <dt>Activating one watcher: O(1)</dt>
1789 <h1 id="AUTHOR">AUTHOR</h1><p><a href="#TOP" class="toplink">Top</a></p>
1790 <div id="AUTHOR_CONTENT">
1791 <p>Marc Lehmann <libev@schmorp.de>.</p>
projects / software / libev.git / blob