ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/libev/ev.html
Revision: 1.74
Committed: Sun Dec 9 19:42:57 2007 UTC (16 years, 5 months ago) by root
Content type: text/html
Branch: MAIN
Changes since 1.73: +20 -10 lines
Log Message:
*** empty log message ***

File Contents

# Content
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">
4 <head>
5 <title>libev</title>
6 <meta name="description" content="Pod documentation for libev" />
7 <meta name="inputfile" content="&lt;standard input&gt;" />
8 <meta name="outputfile" content="&lt;standard output&gt;" />
9 <meta name="created" content="Sun Dec 9 20:30:11 2007" />
10 <meta name="generator" content="Pod::Xhtml 1.57" />
11 <link rel="stylesheet" href="http://res.tst.eu/pod.css"/></head>
12 <body>
13 <div class="pod">
14 <!-- INDEX START -->
15 <h3 id="TOP">Index</h3>
16
17 <ul><li><a href="#NAME">NAME</a></li>
18 <li><a href="#SYNOPSIS">SYNOPSIS</a></li>
19 <li><a href="#EXAMPLE_PROGRAM">EXAMPLE PROGRAM</a></li>
20 <li><a href="#DESCRIPTION">DESCRIPTION</a></li>
21 <li><a href="#FEATURES">FEATURES</a></li>
22 <li><a href="#CONVENTIONS">CONVENTIONS</a></li>
23 <li><a href="#TIME_REPRESENTATION">TIME REPRESENTATION</a></li>
24 <li><a href="#GLOBAL_FUNCTIONS">GLOBAL FUNCTIONS</a></li>
25 <li><a href="#FUNCTIONS_CONTROLLING_THE_EVENT_LOOP">FUNCTIONS CONTROLLING THE EVENT LOOP</a></li>
26 <li><a href="#ANATOMY_OF_A_WATCHER">ANATOMY OF A WATCHER</a>
27 <ul><li><a href="#GENERIC_WATCHER_FUNCTIONS">GENERIC WATCHER FUNCTIONS</a></li>
28 <li><a href="#ASSOCIATING_CUSTOM_DATA_WITH_A_WATCH">ASSOCIATING CUSTOM DATA WITH A WATCHER</a></li>
29 </ul>
30 </li>
31 <li><a href="#WATCHER_TYPES">WATCHER TYPES</a>
32 <ul><li><a href="#code_ev_io_code_is_this_file_descrip"><code>ev_io</code> - is this file descriptor readable or writable?</a></li>
33 <li><a href="#code_ev_timer_code_relative_and_opti"><code>ev_timer</code> - relative and optionally repeating timeouts</a></li>
34 <li><a href="#code_ev_periodic_code_to_cron_or_not"><code>ev_periodic</code> - to cron or not to cron?</a></li>
35 <li><a href="#code_ev_signal_code_signal_me_when_a"><code>ev_signal</code> - signal me when a signal gets signalled!</a></li>
36 <li><a href="#code_ev_child_code_watch_out_for_pro"><code>ev_child</code> - watch out for process status changes</a></li>
37 <li><a href="#code_ev_stat_code_did_the_file_attri"><code>ev_stat</code> - did the file attributes just change?</a></li>
38 <li><a href="#code_ev_idle_code_when_you_ve_got_no"><code>ev_idle</code> - when you've got nothing better to do...</a></li>
39 <li><a href="#code_ev_prepare_code_and_code_ev_che"><code>ev_prepare</code> and <code>ev_check</code> - customise your event loop!</a></li>
40 <li><a href="#code_ev_embed_code_when_one_backend_"><code>ev_embed</code> - when one backend isn't enough...</a></li>
41 <li><a href="#code_ev_fork_code_the_audacity_to_re"><code>ev_fork</code> - the audacity to resume the event loop after a fork</a></li>
42 </ul>
43 </li>
44 <li><a href="#OTHER_FUNCTIONS">OTHER FUNCTIONS</a></li>
45 <li><a href="#LIBEVENT_EMULATION">LIBEVENT EMULATION</a></li>
46 <li><a href="#C_SUPPORT">C++ SUPPORT</a></li>
47 <li><a href="#MACRO_MAGIC">MACRO MAGIC</a></li>
48 <li><a href="#EMBEDDING">EMBEDDING</a>
49 <ul><li><a href="#FILESETS">FILESETS</a>
50 <ul><li><a href="#CORE_EVENT_LOOP">CORE EVENT LOOP</a></li>
51 <li><a href="#LIBEVENT_COMPATIBILITY_API">LIBEVENT COMPATIBILITY API</a></li>
52 <li><a href="#AUTOCONF_SUPPORT">AUTOCONF SUPPORT</a></li>
53 </ul>
54 </li>
55 <li><a href="#PREPROCESSOR_SYMBOLS_MACROS">PREPROCESSOR SYMBOLS/MACROS</a></li>
56 <li><a href="#EXAMPLES">EXAMPLES</a></li>
57 </ul>
58 </li>
59 <li><a href="#COMPLEXITIES">COMPLEXITIES</a></li>
60 <li><a href="#AUTHOR">AUTHOR</a>
61 </li>
62 </ul><hr />
63 <!-- INDEX END -->
64
65 <h1 id="NAME">NAME</h1>
66 <div id="NAME_CONTENT">
67 <p>libev - a high performance full-featured event loop written in C</p>
68
69 </div>
70 <h1 id="SYNOPSIS">SYNOPSIS</h1>
71 <div id="SYNOPSIS_CONTENT">
72 <pre> #include &lt;ev.h&gt;
73
74 </pre>
75
76 </div>
77 <h1 id="EXAMPLE_PROGRAM">EXAMPLE PROGRAM</h1>
78 <div id="EXAMPLE_PROGRAM_CONTENT">
79 <pre> #include &lt;ev.h&gt;
80
81 ev_io stdin_watcher;
82 ev_timer timeout_watcher;
83
84 /* called when data readable on stdin */
85 static void
86 stdin_cb (EV_P_ struct ev_io *w, int revents)
87 {
88 /* puts (&quot;stdin ready&quot;); */
89 ev_io_stop (EV_A_ w); /* just a syntax example */
90 ev_unloop (EV_A_ EVUNLOOP_ALL); /* leave all loop calls */
91 }
92
93 static void
94 timeout_cb (EV_P_ struct ev_timer *w, int revents)
95 {
96 /* puts (&quot;timeout&quot;); */
97 ev_unloop (EV_A_ EVUNLOOP_ONE); /* leave one loop call */
98 }
99
100 int
101 main (void)
102 {
103 struct ev_loop *loop = ev_default_loop (0);
104
105 /* initialise an io watcher, then start it */
106 ev_io_init (&amp;stdin_watcher, stdin_cb, /*STDIN_FILENO*/ 0, EV_READ);
107 ev_io_start (loop, &amp;stdin_watcher);
108
109 /* simple non-repeating 5.5 second timeout */
110 ev_timer_init (&amp;timeout_watcher, timeout_cb, 5.5, 0.);
111 ev_timer_start (loop, &amp;timeout_watcher);
112
113 /* loop till timeout or data ready */
114 ev_loop (loop, 0);
115
116 return 0;
117 }
118
119 </pre>
120
121 </div>
122 <h1 id="DESCRIPTION">DESCRIPTION</h1>
123 <div id="DESCRIPTION_CONTENT">
124 <p>The newest version of this document is also available as a html-formatted
125 web page you might find easier to navigate when reading it for the first
126 time: <a href="http://cvs.schmorp.de/libev/ev.html">http://cvs.schmorp.de/libev/ev.html</a>.</p>
127 <p>Libev is an event loop: you register interest in certain events (such as a
128 file descriptor being readable or a timeout occuring), and it will manage
129 these event sources and provide your program with events.</p>
130 <p>To do this, it must take more or less complete control over your process
131 (or thread) by executing the <i>event loop</i> handler, and will then
132 communicate events via a callback mechanism.</p>
133 <p>You register interest in certain events by registering so-called <i>event
134 watchers</i>, which are relatively small C structures you initialise with the
135 details of the event, and then hand it over to libev by <i>starting</i> the
136 watcher.</p>
137
138 </div>
139 <h1 id="FEATURES">FEATURES</h1>
140 <div id="FEATURES_CONTENT">
141 <p>Libev supports <code>select</code>, <code>poll</code>, the Linux-specific <code>epoll</code>, the
142 BSD-specific <code>kqueue</code> and the Solaris-specific event port mechanisms
143 for file descriptor events (<code>ev_io</code>), the Linux <code>inotify</code> interface
144 (for <code>ev_stat</code>), relative timers (<code>ev_timer</code>), absolute timers
145 with customised rescheduling (<code>ev_periodic</code>), synchronous signals
146 (<code>ev_signal</code>), process status change events (<code>ev_child</code>), and event
147 watchers dealing with the event loop mechanism itself (<code>ev_idle</code>,
148 <code>ev_embed</code>, <code>ev_prepare</code> and <code>ev_check</code> watchers) as well as
149 file watchers (<code>ev_stat</code>) and even limited support for fork events
150 (<code>ev_fork</code>).</p>
151 <p>It also is quite fast (see this
152 <a href="http://libev.schmorp.de/bench.html">benchmark</a> comparing it to libevent
153 for example).</p>
154
155 </div>
156 <h1 id="CONVENTIONS">CONVENTIONS</h1>
157 <div id="CONVENTIONS_CONTENT">
158 <p>Libev is very configurable. In this manual the default configuration will
159 be described, which supports multiple event loops. For more info about
160 various configuration options please have a look at <strong>EMBED</strong> section in
161 this manual. If libev was configured without support for multiple event
162 loops, then all functions taking an initial argument of name <code>loop</code>
163 (which is always of type <code>struct ev_loop *</code>) will not have this argument.</p>
164
165 </div>
166 <h1 id="TIME_REPRESENTATION">TIME REPRESENTATION</h1>
167 <div id="TIME_REPRESENTATION_CONTENT">
168 <p>Libev represents time as a single floating point number, representing the
169 (fractional) number of seconds since the (POSIX) epoch (somewhere near
170 the beginning of 1970, details are complicated, don't ask). This type is
171 called <code>ev_tstamp</code>, which is what you should use too. It usually aliases
172 to the <code>double</code> type in C, and when you need to do any calculations on
173 it, you should treat it as such.</p>
174
175 </div>
176 <h1 id="GLOBAL_FUNCTIONS">GLOBAL FUNCTIONS</h1>
177 <div id="GLOBAL_FUNCTIONS_CONTENT">
178 <p>These functions can be called anytime, even before initialising the
179 library in any way.</p>
180 <dl>
181 <dt>ev_tstamp ev_time ()</dt>
182 <dd>
183 <p>Returns the current time as libev would use it. Please note that the
184 <code>ev_now</code> function is usually faster and also often returns the timestamp
185 you actually want to know.</p>
186 </dd>
187 <dt>int ev_version_major ()</dt>
188 <dt>int ev_version_minor ()</dt>
189 <dd>
190 <p>You can find out the major and minor version numbers of the library
191 you linked against by calling the functions <code>ev_version_major</code> and
192 <code>ev_version_minor</code>. If you want, you can compare against the global
193 symbols <code>EV_VERSION_MAJOR</code> and <code>EV_VERSION_MINOR</code>, which specify the
194 version of the library your program was compiled against.</p>
195 <p>Usually, it's a good idea to terminate if the major versions mismatch,
196 as this indicates an incompatible change. Minor versions are usually
197 compatible to older versions, so a larger minor version alone is usually
198 not a problem.</p>
199 <p>Example: Make sure we haven't accidentally been linked against the wrong
200 version.</p>
201 <pre> assert ((&quot;libev version mismatch&quot;,
202 ev_version_major () == EV_VERSION_MAJOR
203 &amp;&amp; ev_version_minor () &gt;= EV_VERSION_MINOR));
204
205 </pre>
206 </dd>
207 <dt>unsigned int ev_supported_backends ()</dt>
208 <dd>
209 <p>Return the set of all backends (i.e. their corresponding <code>EV_BACKEND_*</code>
210 value) compiled into this binary of libev (independent of their
211 availability on the system you are running on). See <code>ev_default_loop</code> for
212 a description of the set values.</p>
213 <p>Example: make sure we have the epoll method, because yeah this is cool and
214 a must have and can we have a torrent of it please!!!11</p>
215 <pre> assert ((&quot;sorry, no epoll, no sex&quot;,
216 ev_supported_backends () &amp; EVBACKEND_EPOLL));
217
218 </pre>
219 </dd>
220 <dt>unsigned int ev_recommended_backends ()</dt>
221 <dd>
222 <p>Return the set of all backends compiled into this binary of libev and also
223 recommended for this platform. This set is often smaller than the one
224 returned by <code>ev_supported_backends</code>, as for example kqueue is broken on
225 most BSDs and will not be autodetected unless you explicitly request it
226 (assuming you know what you are doing). This is the set of backends that
227 libev will probe for if you specify no backends explicitly.</p>
228 </dd>
229 <dt>unsigned int ev_embeddable_backends ()</dt>
230 <dd>
231 <p>Returns the set of backends that are embeddable in other event loops. This
232 is the theoretical, all-platform, value. To find which backends
233 might be supported on the current system, you would need to look at
234 <code>ev_embeddable_backends () &amp; ev_supported_backends ()</code>, likewise for
235 recommended ones.</p>
236 <p>See the description of <code>ev_embed</code> watchers for more info.</p>
237 </dd>
238 <dt>ev_set_allocator (void *(*cb)(void *ptr, long size))</dt>
239 <dd>
240 <p>Sets the allocation function to use (the prototype is similar - the
241 semantics is identical - to the realloc C function). It is used to
242 allocate and free memory (no surprises here). If it returns zero when
243 memory needs to be allocated, the library might abort or take some
244 potentially destructive action. The default is your system realloc
245 function.</p>
246 <p>You could override this function in high-availability programs to, say,
247 free some memory if it cannot allocate memory, to use a special allocator,
248 or even to sleep a while and retry until some memory is available.</p>
249 <p>Example: Replace the libev allocator with one that waits a bit and then
250 retries).</p>
251 <pre> static void *
252 persistent_realloc (void *ptr, size_t size)
253 {
254 for (;;)
255 {
256 void *newptr = realloc (ptr, size);
257
258 if (newptr)
259 return newptr;
260
261 sleep (60);
262 }
263 }
264
265 ...
266 ev_set_allocator (persistent_realloc);
267
268 </pre>
269 </dd>
270 <dt>ev_set_syserr_cb (void (*cb)(const char *msg));</dt>
271 <dd>
272 <p>Set the callback function to call on a retryable syscall error (such
273 as failed select, poll, epoll_wait). The message is a printable string
274 indicating the system call or subsystem causing the problem. If this
275 callback is set, then libev will expect it to remedy the sitution, no
276 matter what, when it returns. That is, libev will generally retry the
277 requested operation, or, if the condition doesn't go away, do bad stuff
278 (such as abort).</p>
279 <p>Example: This is basically the same thing that libev does internally, too.</p>
280 <pre> static void
281 fatal_error (const char *msg)
282 {
283 perror (msg);
284 abort ();
285 }
286
287 ...
288 ev_set_syserr_cb (fatal_error);
289
290 </pre>
291 </dd>
292 </dl>
293
294 </div>
295 <h1 id="FUNCTIONS_CONTROLLING_THE_EVENT_LOOP">FUNCTIONS CONTROLLING THE EVENT LOOP</h1>
296 <div id="FUNCTIONS_CONTROLLING_THE_EVENT_LOOP-2">
297 <p>An event loop is described by a <code>struct ev_loop *</code>. The library knows two
298 types of such loops, the <i>default</i> loop, which supports signals and child
299 events, and dynamically created loops which do not.</p>
300 <p>If you use threads, a common model is to run the default event loop
301 in your main thread (or in a separate thread) and for each thread you
302 create, you also create another event loop. Libev itself does no locking
303 whatsoever, so if you mix calls to the same event loop in different
304 threads, make sure you lock (this is usually a bad idea, though, even if
305 done correctly, because it's hideous and inefficient).</p>
306 <dl>
307 <dt>struct ev_loop *ev_default_loop (unsigned int flags)</dt>
308 <dd>
309 <p>This will initialise the default event loop if it hasn't been initialised
310 yet and return it. If the default loop could not be initialised, returns
311 false. If it already was initialised it simply returns it (and ignores the
312 flags. If that is troubling you, check <code>ev_backend ()</code> afterwards).</p>
313 <p>If you don't know what event loop to use, use the one returned from this
314 function.</p>
315 <p>The flags argument can be used to specify special behaviour or specific
316 backends to use, and is usually specified as <code>0</code> (or <code>EVFLAG_AUTO</code>).</p>
317 <p>The following flags are supported:</p>
318 <p>
319 <dl>
320 <dt><code>EVFLAG_AUTO</code></dt>
321 <dd>
322 <p>The default flags value. Use this if you have no clue (it's the right
323 thing, believe me).</p>
324 </dd>
325 <dt><code>EVFLAG_NOENV</code></dt>
326 <dd>
327 <p>If this flag bit is ored into the flag value (or the program runs setuid
328 or setgid) then libev will <i>not</i> look at the environment variable
329 <code>LIBEV_FLAGS</code>. Otherwise (the default), this environment variable will
330 override the flags completely if it is found in the environment. This is
331 useful to try out specific backends to test their performance, or to work
332 around bugs.</p>
333 </dd>
334 <dt><code>EVFLAG_FORKCHECK</code></dt>
335 <dd>
336 <p>Instead of calling <code>ev_default_fork</code> or <code>ev_loop_fork</code> manually after
337 a fork, you can also make libev check for a fork in each iteration by
338 enabling this flag.</p>
339 <p>This works by calling <code>getpid ()</code> on every iteration of the loop,
340 and thus this might slow down your event loop if you do a lot of loop
341 iterations and little real work, but is usually not noticeable (on my
342 Linux system for example, <code>getpid</code> is actually a simple 5-insn sequence
343 without a syscall and thus <i>very</i> fast, but my Linux system also has
344 <code>pthread_atfork</code> which is even faster).</p>
345 <p>The big advantage of this flag is that you can forget about fork (and
346 forget about forgetting to tell libev about forking) when you use this
347 flag.</p>
348 <p>This flag setting cannot be overriden or specified in the <code>LIBEV_FLAGS</code>
349 environment variable.</p>
350 </dd>
351 <dt><code>EVBACKEND_SELECT</code> (value 1, portable select backend)</dt>
352 <dd>
353 <p>This is your standard select(2) backend. Not <i>completely</i> standard, as
354 libev tries to roll its own fd_set with no limits on the number of fds,
355 but if that fails, expect a fairly low limit on the number of fds when
356 using this backend. It doesn't scale too well (O(highest_fd)), but its usually
357 the fastest backend for a low number of fds.</p>
358 </dd>
359 <dt><code>EVBACKEND_POLL</code> (value 2, poll backend, available everywhere except on windows)</dt>
360 <dd>
361 <p>And this is your standard poll(2) backend. It's more complicated than
362 select, but handles sparse fds better and has no artificial limit on the
363 number of fds you can use (except it will slow down considerably with a
364 lot of inactive fds). It scales similarly to select, i.e. O(total_fds).</p>
365 </dd>
366 <dt><code>EVBACKEND_EPOLL</code> (value 4, Linux)</dt>
367 <dd>
368 <p>For few fds, this backend is a bit little slower than poll and select,
369 but it scales phenomenally better. While poll and select usually scale like
370 O(total_fds) where n is the total number of fds (or the highest fd), epoll scales
371 either O(1) or O(active_fds).</p>
372 <p>While stopping and starting an I/O watcher in the same iteration will
373 result in some caching, there is still a syscall per such incident
374 (because the fd could point to a different file description now), so its
375 best to avoid that. Also, dup()ed file descriptors might not work very
376 well if you register events for both fds.</p>
377 <p>Please note that epoll sometimes generates spurious notifications, so you
378 need to use non-blocking I/O or other means to avoid blocking when no data
379 (or space) is available.</p>
380 </dd>
381 <dt><code>EVBACKEND_KQUEUE</code> (value 8, most BSD clones)</dt>
382 <dd>
383 <p>Kqueue deserves special mention, as at the time of this writing, it
384 was broken on all BSDs except NetBSD (usually it doesn't work with
385 anything but sockets and pipes, except on Darwin, where of course its
386 completely useless). For this reason its not being &quot;autodetected&quot;
387 unless you explicitly specify it explicitly in the flags (i.e. using
388 <code>EVBACKEND_KQUEUE</code>).</p>
389 <p>It scales in the same way as the epoll backend, but the interface to the
390 kernel is more efficient (which says nothing about its actual speed, of
391 course). While starting and stopping an I/O watcher does not cause an
392 extra syscall as with epoll, it still adds up to four event changes per
393 incident, so its best to avoid that.</p>
394 </dd>
395 <dt><code>EVBACKEND_DEVPOLL</code> (value 16, Solaris 8)</dt>
396 <dd>
397 <p>This is not implemented yet (and might never be).</p>
398 </dd>
399 <dt><code>EVBACKEND_PORT</code> (value 32, Solaris 10)</dt>
400 <dd>
401 <p>This uses the Solaris 10 port mechanism. As with everything on Solaris,
402 it's really slow, but it still scales very well (O(active_fds)).</p>
403 <p>Please note that solaris ports can result in a lot of spurious
404 notifications, so you need to use non-blocking I/O or other means to avoid
405 blocking when no data (or space) is available.</p>
406 </dd>
407 <dt><code>EVBACKEND_ALL</code></dt>
408 <dd>
409 <p>Try all backends (even potentially broken ones that wouldn't be tried
410 with <code>EVFLAG_AUTO</code>). Since this is a mask, you can do stuff such as
411 <code>EVBACKEND_ALL &amp; ~EVBACKEND_KQUEUE</code>.</p>
412 </dd>
413 </dl>
414 </p>
415 <p>If one or more of these are ored into the flags value, then only these
416 backends will be tried (in the reverse order as given here). If none are
417 specified, most compiled-in backend will be tried, usually in reverse
418 order of their flag values :)</p>
419 <p>The most typical usage is like this:</p>
420 <pre> if (!ev_default_loop (0))
421 fatal (&quot;could not initialise libev, bad $LIBEV_FLAGS in environment?&quot;);
422
423 </pre>
424 <p>Restrict libev to the select and poll backends, and do not allow
425 environment settings to be taken into account:</p>
426 <pre> ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);
427
428 </pre>
429 <p>Use whatever libev has to offer, but make sure that kqueue is used if
430 available (warning, breaks stuff, best use only with your own private
431 event loop and only if you know the OS supports your types of fds):</p>
432 <pre> ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);
433
434 </pre>
435 </dd>
436 <dt>struct ev_loop *ev_loop_new (unsigned int flags)</dt>
437 <dd>
438 <p>Similar to <code>ev_default_loop</code>, but always creates a new event loop that is
439 always distinct from the default loop. Unlike the default loop, it cannot
440 handle signal and child watchers, and attempts to do so will be greeted by
441 undefined behaviour (or a failed assertion if assertions are enabled).</p>
442 <p>Example: Try to create a event loop that uses epoll and nothing else.</p>
443 <pre> struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);
444 if (!epoller)
445 fatal (&quot;no epoll found here, maybe it hides under your chair&quot;);
446
447 </pre>
448 </dd>
449 <dt>ev_default_destroy ()</dt>
450 <dd>
451 <p>Destroys the default loop again (frees all memory and kernel state
452 etc.). None of the active event watchers will be stopped in the normal
453 sense, so e.g. <code>ev_is_active</code> might still return true. It is your
454 responsibility to either stop all watchers cleanly yoursef <i>before</i>
455 calling this function, or cope with the fact afterwards (which is usually
456 the easiest thing, youc na just ignore the watchers and/or <code>free ()</code> them
457 for example).</p>
458 </dd>
459 <dt>ev_loop_destroy (loop)</dt>
460 <dd>
461 <p>Like <code>ev_default_destroy</code>, but destroys an event loop created by an
462 earlier call to <code>ev_loop_new</code>.</p>
463 </dd>
464 <dt>ev_default_fork ()</dt>
465 <dd>
466 <p>This function reinitialises the kernel state for backends that have
467 one. Despite the name, you can call it anytime, but it makes most sense
468 after forking, in either the parent or child process (or both, but that
469 again makes little sense).</p>
470 <p>You <i>must</i> call this function in the child process after forking if and
471 only if you want to use the event library in both processes. If you just
472 fork+exec, you don't have to call it.</p>
473 <p>The function itself is quite fast and it's usually not a problem to call
474 it just in case after a fork. To make this easy, the function will fit in
475 quite nicely into a call to <code>pthread_atfork</code>:</p>
476 <pre> pthread_atfork (0, 0, ev_default_fork);
477
478 </pre>
479 <p>At the moment, <code>EVBACKEND_SELECT</code> and <code>EVBACKEND_POLL</code> are safe to use
480 without calling this function, so if you force one of those backends you
481 do not need to care.</p>
482 </dd>
483 <dt>ev_loop_fork (loop)</dt>
484 <dd>
485 <p>Like <code>ev_default_fork</code>, but acts on an event loop created by
486 <code>ev_loop_new</code>. Yes, you have to call this on every allocated event loop
487 after fork, and how you do this is entirely your own problem.</p>
488 </dd>
489 <dt>unsigned int ev_loop_count (loop)</dt>
490 <dd>
491 <p>Returns the count of loop iterations for the loop, which is identical to
492 the number of times libev did poll for new events. It starts at <code>0</code> and
493 happily wraps around with enough iterations.</p>
494 <p>This value can sometimes be useful as a generation counter of sorts (it
495 &quot;ticks&quot; the number of loop iterations), as it roughly corresponds with
496 <code>ev_prepare</code> and <code>ev_check</code> calls.</p>
497 </dd>
498 <dt>unsigned int ev_backend (loop)</dt>
499 <dd>
500 <p>Returns one of the <code>EVBACKEND_*</code> flags indicating the event backend in
501 use.</p>
502 </dd>
503 <dt>ev_tstamp ev_now (loop)</dt>
504 <dd>
505 <p>Returns the current &quot;event loop time&quot;, which is the time the event loop
506 received events and started processing them. This timestamp does not
507 change as long as callbacks are being processed, and this is also the base
508 time used for relative timers. You can treat it as the timestamp of the
509 event occuring (or more correctly, libev finding out about it).</p>
510 </dd>
511 <dt>ev_loop (loop, int flags)</dt>
512 <dd>
513 <p>Finally, this is it, the event handler. This function usually is called
514 after you initialised all your watchers and you want to start handling
515 events.</p>
516 <p>If the flags argument is specified as <code>0</code>, it will not return until
517 either no event watchers are active anymore or <code>ev_unloop</code> was called.</p>
518 <p>Please note that an explicit <code>ev_unloop</code> is usually better than
519 relying on all watchers to be stopped when deciding when a program has
520 finished (especially in interactive programs), but having a program that
521 automatically loops as long as it has to and no longer by virtue of
522 relying on its watchers stopping correctly is a thing of beauty.</p>
523 <p>A flags value of <code>EVLOOP_NONBLOCK</code> will look for new events, will handle
524 those events and any outstanding ones, but will not block your process in
525 case there are no events and will return after one iteration of the loop.</p>
526 <p>A flags value of <code>EVLOOP_ONESHOT</code> will look for new events (waiting if
527 neccessary) and will handle those and any outstanding ones. It will block
528 your process until at least one new event arrives, and will return after
529 one iteration of the loop. This is useful if you are waiting for some
530 external event in conjunction with something not expressible using other
531 libev watchers. However, a pair of <code>ev_prepare</code>/<code>ev_check</code> watchers is
532 usually a better approach for this kind of thing.</p>
533 <p>Here are the gory details of what <code>ev_loop</code> does:</p>
534 <pre> - Before the first iteration, call any pending watchers.
535 * If there are no active watchers (reference count is zero), return.
536 - Queue all prepare watchers and then call all outstanding watchers.
537 - If we have been forked, recreate the kernel state.
538 - Update the kernel state with all outstanding changes.
539 - Update the &quot;event loop time&quot;.
540 - Calculate for how long to block.
541 - Block the process, waiting for any events.
542 - Queue all outstanding I/O (fd) events.
543 - Update the &quot;event loop time&quot; and do time jump handling.
544 - Queue all outstanding timers.
545 - Queue all outstanding periodics.
546 - If no events are pending now, queue all idle watchers.
547 - Queue all check watchers.
548 - Call all queued watchers in reverse order (i.e. check watchers first).
549 Signals and child watchers are implemented as I/O watchers, and will
550 be handled here by queueing them when their watcher gets executed.
551 - If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK
552 were used, return, otherwise continue with step *.
553
554 </pre>
555 <p>Example: Queue some jobs and then loop until no events are outsanding
556 anymore.</p>
557 <pre> ... queue jobs here, make sure they register event watchers as long
558 ... as they still have work to do (even an idle watcher will do..)
559 ev_loop (my_loop, 0);
560 ... jobs done. yeah!
561
562 </pre>
563 </dd>
564 <dt>ev_unloop (loop, how)</dt>
565 <dd>
566 <p>Can be used to make a call to <code>ev_loop</code> return early (but only after it
567 has processed all outstanding events). The <code>how</code> argument must be either
568 <code>EVUNLOOP_ONE</code>, which will make the innermost <code>ev_loop</code> call return, or
569 <code>EVUNLOOP_ALL</code>, which will make all nested <code>ev_loop</code> calls return.</p>
570 </dd>
571 <dt>ev_ref (loop)</dt>
572 <dt>ev_unref (loop)</dt>
573 <dd>
574 <p>Ref/unref can be used to add or remove a reference count on the event
575 loop: Every watcher keeps one reference, and as long as the reference
576 count is nonzero, <code>ev_loop</code> will not return on its own. If you have
577 a watcher you never unregister that should not keep <code>ev_loop</code> from
578 returning, ev_unref() after starting, and ev_ref() before stopping it. For
579 example, libev itself uses this for its internal signal pipe: It is not
580 visible to the libev user and should not keep <code>ev_loop</code> from exiting if
581 no event watchers registered by it are active. It is also an excellent
582 way to do this for generic recurring timers or from within third-party
583 libraries. Just remember to <i>unref after start</i> and <i>ref before stop</i>.</p>
584 <p>Example: Create a signal watcher, but keep it from keeping <code>ev_loop</code>
585 running when nothing else is active.</p>
586 <pre> struct ev_signal exitsig;
587 ev_signal_init (&amp;exitsig, sig_cb, SIGINT);
588 ev_signal_start (loop, &amp;exitsig);
589 evf_unref (loop);
590
591 </pre>
592 <p>Example: For some weird reason, unregister the above signal handler again.</p>
593 <pre> ev_ref (loop);
594 ev_signal_stop (loop, &amp;exitsig);
595
596 </pre>
597 </dd>
598 </dl>
599
600
601
602
603
604 </div>
605 <h1 id="ANATOMY_OF_A_WATCHER">ANATOMY OF A WATCHER</h1>
606 <div id="ANATOMY_OF_A_WATCHER_CONTENT">
607 <p>A watcher is a structure that you create and register to record your
608 interest in some event. For instance, if you want to wait for STDIN to
609 become readable, you would create an <code>ev_io</code> watcher for that:</p>
610 <pre> static void my_cb (struct ev_loop *loop, struct ev_io *w, int revents)
611 {
612 ev_io_stop (w);
613 ev_unloop (loop, EVUNLOOP_ALL);
614 }
615
616 struct ev_loop *loop = ev_default_loop (0);
617 struct ev_io stdin_watcher;
618 ev_init (&amp;stdin_watcher, my_cb);
619 ev_io_set (&amp;stdin_watcher, STDIN_FILENO, EV_READ);
620 ev_io_start (loop, &amp;stdin_watcher);
621 ev_loop (loop, 0);
622
623 </pre>
624 <p>As you can see, you are responsible for allocating the memory for your
625 watcher structures (and it is usually a bad idea to do this on the stack,
626 although this can sometimes be quite valid).</p>
627 <p>Each watcher structure must be initialised by a call to <code>ev_init
628 (watcher *, callback)</code>, which expects a callback to be provided. This
629 callback gets invoked each time the event occurs (or, in the case of io
630 watchers, each time the event loop detects that the file descriptor given
631 is readable and/or writable).</p>
632 <p>Each watcher type has its own <code>ev_&lt;type&gt;_set (watcher *, ...)</code> macro
633 with arguments specific to this watcher type. There is also a macro
634 to combine initialisation and setting in one call: <code>ev_&lt;type&gt;_init
635 (watcher *, callback, ...)</code>.</p>
636 <p>To make the watcher actually watch out for events, you have to start it
637 with a watcher-specific start function (<code>ev_&lt;type&gt;_start (loop, watcher
638 *)</code>), and you can stop watching for events at any time by calling the
639 corresponding stop function (<code>ev_&lt;type&gt;_stop (loop, watcher *)</code>.</p>
640 <p>As long as your watcher is active (has been started but not stopped) you
641 must not touch the values stored in it. Most specifically you must never
642 reinitialise it or call its <code>set</code> macro.</p>
643 <p>Each and every callback receives the event loop pointer as first, the
644 registered watcher structure as second, and a bitset of received events as
645 third argument.</p>
646 <p>The received events usually include a single bit per event type received
647 (you can receive multiple events at the same time). The possible bit masks
648 are:</p>
649 <dl>
650 <dt><code>EV_READ</code></dt>
651 <dt><code>EV_WRITE</code></dt>
652 <dd>
653 <p>The file descriptor in the <code>ev_io</code> watcher has become readable and/or
654 writable.</p>
655 </dd>
656 <dt><code>EV_TIMEOUT</code></dt>
657 <dd>
658 <p>The <code>ev_timer</code> watcher has timed out.</p>
659 </dd>
660 <dt><code>EV_PERIODIC</code></dt>
661 <dd>
662 <p>The <code>ev_periodic</code> watcher has timed out.</p>
663 </dd>
664 <dt><code>EV_SIGNAL</code></dt>
665 <dd>
666 <p>The signal specified in the <code>ev_signal</code> watcher has been received by a thread.</p>
667 </dd>
668 <dt><code>EV_CHILD</code></dt>
669 <dd>
670 <p>The pid specified in the <code>ev_child</code> watcher has received a status change.</p>
671 </dd>
672 <dt><code>EV_STAT</code></dt>
673 <dd>
674 <p>The path specified in the <code>ev_stat</code> watcher changed its attributes somehow.</p>
675 </dd>
676 <dt><code>EV_IDLE</code></dt>
677 <dd>
678 <p>The <code>ev_idle</code> watcher has determined that you have nothing better to do.</p>
679 </dd>
680 <dt><code>EV_PREPARE</code></dt>
681 <dt><code>EV_CHECK</code></dt>
682 <dd>
683 <p>All <code>ev_prepare</code> watchers are invoked just <i>before</i> <code>ev_loop</code> starts
684 to gather new events, and all <code>ev_check</code> watchers are invoked just after
685 <code>ev_loop</code> has gathered them, but before it invokes any callbacks for any
686 received events. Callbacks of both watcher types can start and stop as
687 many watchers as they want, and all of them will be taken into account
688 (for example, a <code>ev_prepare</code> watcher might start an idle watcher to keep
689 <code>ev_loop</code> from blocking).</p>
690 </dd>
691 <dt><code>EV_EMBED</code></dt>
692 <dd>
693 <p>The embedded event loop specified in the <code>ev_embed</code> watcher needs attention.</p>
694 </dd>
695 <dt><code>EV_FORK</code></dt>
696 <dd>
697 <p>The event loop has been resumed in the child process after fork (see
698 <code>ev_fork</code>).</p>
699 </dd>
700 <dt><code>EV_ERROR</code></dt>
701 <dd>
702 <p>An unspecified error has occured, the watcher has been stopped. This might
703 happen because the watcher could not be properly started because libev
704 ran out of memory, a file descriptor was found to be closed or any other
705 problem. You best act on it by reporting the problem and somehow coping
706 with the watcher being stopped.</p>
707 <p>Libev will usually signal a few &quot;dummy&quot; events together with an error,
708 for example it might indicate that a fd is readable or writable, and if
709 your callbacks is well-written it can just attempt the operation and cope
710 with the error from read() or write(). This will not work in multithreaded
711 programs, though, so beware.</p>
712 </dd>
713 </dl>
714
715 </div>
716 <h2 id="GENERIC_WATCHER_FUNCTIONS">GENERIC WATCHER FUNCTIONS</h2>
717 <div id="GENERIC_WATCHER_FUNCTIONS_CONTENT">
718 <p>In the following description, <code>TYPE</code> stands for the watcher type,
719 e.g. <code>timer</code> for <code>ev_timer</code> watchers and <code>io</code> for <code>ev_io</code> watchers.</p>
720 <dl>
721 <dt><code>ev_init</code> (ev_TYPE *watcher, callback)</dt>
722 <dd>
723 <p>This macro initialises the generic portion of a watcher. The contents
724 of the watcher object can be arbitrary (so <code>malloc</code> will do). Only
725 the generic parts of the watcher are initialised, you <i>need</i> to call
726 the type-specific <code>ev_TYPE_set</code> macro afterwards to initialise the
727 type-specific parts. For each type there is also a <code>ev_TYPE_init</code> macro
728 which rolls both calls into one.</p>
729 <p>You can reinitialise a watcher at any time as long as it has been stopped
730 (or never started) and there are no pending events outstanding.</p>
731 <p>The callback is always of type <code>void (*)(ev_loop *loop, ev_TYPE *watcher,
732 int revents)</code>.</p>
733 </dd>
734 <dt><code>ev_TYPE_set</code> (ev_TYPE *, [args])</dt>
735 <dd>
736 <p>This macro initialises the type-specific parts of a watcher. You need to
737 call <code>ev_init</code> at least once before you call this macro, but you can
738 call <code>ev_TYPE_set</code> any number of times. You must not, however, call this
739 macro on a watcher that is active (it can be pending, however, which is a
740 difference to the <code>ev_init</code> macro).</p>
741 <p>Although some watcher types do not have type-specific arguments
742 (e.g. <code>ev_prepare</code>) you still need to call its <code>set</code> macro.</p>
743 </dd>
744 <dt><code>ev_TYPE_init</code> (ev_TYPE *watcher, callback, [args])</dt>
745 <dd>
746 <p>This convinience macro rolls both <code>ev_init</code> and <code>ev_TYPE_set</code> macro
747 calls into a single call. This is the most convinient method to initialise
748 a watcher. The same limitations apply, of course.</p>
749 </dd>
750 <dt><code>ev_TYPE_start</code> (loop *, ev_TYPE *watcher)</dt>
751 <dd>
752 <p>Starts (activates) the given watcher. Only active watchers will receive
753 events. If the watcher is already active nothing will happen.</p>
754 </dd>
755 <dt><code>ev_TYPE_stop</code> (loop *, ev_TYPE *watcher)</dt>
756 <dd>
757 <p>Stops the given watcher again (if active) and clears the pending
758 status. It is possible that stopped watchers are pending (for example,
759 non-repeating timers are being stopped when they become pending), but
760 <code>ev_TYPE_stop</code> ensures that the watcher is neither active nor pending. If
761 you want to free or reuse the memory used by the watcher it is therefore a
762 good idea to always call its <code>ev_TYPE_stop</code> function.</p>
763 </dd>
764 <dt>bool ev_is_active (ev_TYPE *watcher)</dt>
765 <dd>
766 <p>Returns a true value iff the watcher is active (i.e. it has been started
767 and not yet been stopped). As long as a watcher is active you must not modify
768 it.</p>
769 </dd>
770 <dt>bool ev_is_pending (ev_TYPE *watcher)</dt>
771 <dd>
772 <p>Returns a true value iff the watcher is pending, (i.e. it has outstanding
773 events but its callback has not yet been invoked). As long as a watcher
774 is pending (but not active) you must not call an init function on it (but
775 <code>ev_TYPE_set</code> is safe), you must not change its priority, and you must
776 make sure the watcher is available to libev (e.g. you cannot <code>free ()</code>
777 it).</p>
778 </dd>
779 <dt>callback ev_cb (ev_TYPE *watcher)</dt>
780 <dd>
781 <p>Returns the callback currently set on the watcher.</p>
782 </dd>
783 <dt>ev_cb_set (ev_TYPE *watcher, callback)</dt>
784 <dd>
785 <p>Change the callback. You can change the callback at virtually any time
786 (modulo threads).</p>
787 </dd>
788 <dt>ev_set_priority (ev_TYPE *watcher, priority)</dt>
789 <dt>int ev_priority (ev_TYPE *watcher)</dt>
790 <dd>
791 <p>Set and query the priority of the watcher. The priority is a small
792 integer between <code>EV_MAXPRI</code> (default: <code>2</code>) and <code>EV_MINPRI</code>
793 (default: <code>-2</code>). Pending watchers with higher priority will be invoked
794 before watchers with lower priority, but priority will not keep watchers
795 from being executed (except for <code>ev_idle</code> watchers).</p>
796 <p>This means that priorities are <i>only</i> used for ordering callback
797 invocation after new events have been received. This is useful, for
798 example, to reduce latency after idling, or more often, to bind two
799 watchers on the same event and make sure one is called first.</p>
800 <p>If you need to suppress invocation when higher priority events are pending
801 you need to look at <code>ev_idle</code> watchers, which provide this functionality.</p>
802 <p>You <i>must not</i> change the priority of a watcher as long as it is active or
803 pending.</p>
804 <p>The default priority used by watchers when no priority has been set is
805 always <code>0</code>, which is supposed to not be too high and not be too low :).</p>
806 <p>Setting a priority outside the range of <code>EV_MINPRI</code> to <code>EV_MAXPRI</code> is
807 fine, as long as you do not mind that the priority value you query might
808 or might not have been adjusted to be within valid range.</p>
809 </dd>
810 <dt>ev_invoke (loop, ev_TYPE *watcher, int revents)</dt>
811 <dd>
812 <p>Invoke the <code>watcher</code> with the given <code>loop</code> and <code>revents</code>. Neither
813 <code>loop</code> nor <code>revents</code> need to be valid as long as the watcher callback
814 can deal with that fact.</p>
815 </dd>
816 <dt>int ev_clear_pending (loop, ev_TYPE *watcher)</dt>
817 <dd>
818 <p>If the watcher is pending, this function returns clears its pending status
819 and returns its <code>revents</code> bitset (as if its callback was invoked). If the
820 watcher isn't pending it does nothing and returns <code>0</code>.</p>
821 </dd>
822 </dl>
823
824
825
826
827
828 </div>
829 <h2 id="ASSOCIATING_CUSTOM_DATA_WITH_A_WATCH">ASSOCIATING CUSTOM DATA WITH A WATCHER</h2>
830 <div id="ASSOCIATING_CUSTOM_DATA_WITH_A_WATCH-2">
831 <p>Each watcher has, by default, a member <code>void *data</code> that you can change
832 and read at any time, libev will completely ignore it. This can be used
833 to associate arbitrary data with your watcher. If you need more data and
834 don't want to allocate memory and store a pointer to it in that data
835 member, you can also &quot;subclass&quot; the watcher type and provide your own
836 data:</p>
837 <pre> struct my_io
838 {
839 struct ev_io io;
840 int otherfd;
841 void *somedata;
842 struct whatever *mostinteresting;
843 }
844
845 </pre>
846 <p>And since your callback will be called with a pointer to the watcher, you
847 can cast it back to your own type:</p>
848 <pre> static void my_cb (struct ev_loop *loop, struct ev_io *w_, int revents)
849 {
850 struct my_io *w = (struct my_io *)w_;
851 ...
852 }
853
854 </pre>
855 <p>More interesting and less C-conformant ways of casting your callback type
856 instead have been omitted.</p>
857 <p>Another common scenario is having some data structure with multiple
858 watchers:</p>
859 <pre> struct my_biggy
860 {
861 int some_data;
862 ev_timer t1;
863 ev_timer t2;
864 }
865
866 </pre>
867 <p>In this case getting the pointer to <code>my_biggy</code> is a bit more complicated,
868 you need to use <code>offsetof</code>:</p>
869 <pre> #include &lt;stddef.h&gt;
870
871 static void
872 t1_cb (EV_P_ struct ev_timer *w, int revents)
873 {
874 struct my_biggy big = (struct my_biggy *
875 (((char *)w) - offsetof (struct my_biggy, t1));
876 }
877
878 static void
879 t2_cb (EV_P_ struct ev_timer *w, int revents)
880 {
881 struct my_biggy big = (struct my_biggy *
882 (((char *)w) - offsetof (struct my_biggy, t2));
883 }
884
885
886
887
888 </pre>
889
890 </div>
891 <h1 id="WATCHER_TYPES">WATCHER TYPES</h1>
892 <div id="WATCHER_TYPES_CONTENT">
893 <p>This section describes each watcher in detail, but will not repeat
894 information given in the last section. Any initialisation/set macros,
895 functions and members specific to the watcher type are explained.</p>
896 <p>Members are additionally marked with either <i>[read-only]</i>, meaning that,
897 while the watcher is active, you can look at the member and expect some
898 sensible content, but you must not modify it (you can modify it while the
899 watcher is stopped to your hearts content), or <i>[read-write]</i>, which
900 means you can expect it to have some sensible content while the watcher
901 is active, but you can also modify it. Modifying it may not do something
902 sensible or take immediate effect (or do anything at all), but libev will
903 not crash or malfunction in any way.</p>
904
905
906
907
908
909 </div>
910 <h2 id="code_ev_io_code_is_this_file_descrip"><code>ev_io</code> - is this file descriptor readable or writable?</h2>
911 <div id="code_ev_io_code_is_this_file_descrip-2">
912 <p>I/O watchers check whether a file descriptor is readable or writable
913 in each iteration of the event loop, or, more precisely, when reading
914 would not block the process and writing would at least be able to write
915 some data. This behaviour is called level-triggering because you keep
916 receiving events as long as the condition persists. Remember you can stop
917 the watcher if you don't want to act on the event and neither want to
918 receive future events.</p>
919 <p>In general you can register as many read and/or write event watchers per
920 fd as you want (as long as you don't confuse yourself). Setting all file
921 descriptors to non-blocking mode is also usually a good idea (but not
922 required if you know what you are doing).</p>
923 <p>You have to be careful with dup'ed file descriptors, though. Some backends
924 (the linux epoll backend is a notable example) cannot handle dup'ed file
925 descriptors correctly if you register interest in two or more fds pointing
926 to the same underlying file/socket/etc. description (that is, they share
927 the same underlying &quot;file open&quot;).</p>
928 <p>If you must do this, then force the use of a known-to-be-good backend
929 (at the time of this writing, this includes only <code>EVBACKEND_SELECT</code> and
930 <code>EVBACKEND_POLL</code>).</p>
931 <p>Another thing you have to watch out for is that it is quite easy to
932 receive &quot;spurious&quot; readyness notifications, that is your callback might
933 be called with <code>EV_READ</code> but a subsequent <code>read</code>(2) will actually block
934 because there is no data. Not only are some backends known to create a
935 lot of those (for example solaris ports), it is very easy to get into
936 this situation even with a relatively standard program structure. Thus
937 it is best to always use non-blocking I/O: An extra <code>read</code>(2) returning
938 <code>EAGAIN</code> is far preferable to a program hanging until some data arrives.</p>
939 <p>If you cannot run the fd in non-blocking mode (for example you should not
940 play around with an Xlib connection), then you have to seperately re-test
941 whether a file descriptor is really ready with a known-to-be good interface
942 such as poll (fortunately in our Xlib example, Xlib already does this on
943 its own, so its quite safe to use).</p>
944 <dl>
945 <dt>ev_io_init (ev_io *, callback, int fd, int events)</dt>
946 <dt>ev_io_set (ev_io *, int fd, int events)</dt>
947 <dd>
948 <p>Configures an <code>ev_io</code> watcher. The <code>fd</code> is the file descriptor to
949 rceeive events for and events is either <code>EV_READ</code>, <code>EV_WRITE</code> or
950 <code>EV_READ | EV_WRITE</code> to receive the given events.</p>
951 </dd>
952 <dt>int fd [read-only]</dt>
953 <dd>
954 <p>The file descriptor being watched.</p>
955 </dd>
956 <dt>int events [read-only]</dt>
957 <dd>
958 <p>The events being watched.</p>
959 </dd>
960 </dl>
961 <p>Example: Call <code>stdin_readable_cb</code> when STDIN_FILENO has become, well
962 readable, but only once. Since it is likely line-buffered, you could
963 attempt to read a whole line in the callback.</p>
964 <pre> static void
965 stdin_readable_cb (struct ev_loop *loop, struct ev_io *w, int revents)
966 {
967 ev_io_stop (loop, w);
968 .. read from stdin here (or from w-&gt;fd) and haqndle any I/O errors
969 }
970
971 ...
972 struct ev_loop *loop = ev_default_init (0);
973 struct ev_io stdin_readable;
974 ev_io_init (&amp;stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
975 ev_io_start (loop, &amp;stdin_readable);
976 ev_loop (loop, 0);
977
978
979
980
981 </pre>
982
983 </div>
984 <h2 id="code_ev_timer_code_relative_and_opti"><code>ev_timer</code> - relative and optionally repeating timeouts</h2>
985 <div id="code_ev_timer_code_relative_and_opti-2">
986 <p>Timer watchers are simple relative timers that generate an event after a
987 given time, and optionally repeating in regular intervals after that.</p>
988 <p>The timers are based on real time, that is, if you register an event that
989 times out after an hour and you reset your system clock to last years
990 time, it will still time out after (roughly) and hour. &quot;Roughly&quot; because
991 detecting time jumps is hard, and some inaccuracies are unavoidable (the
992 monotonic clock option helps a lot here).</p>
993 <p>The relative timeouts are calculated relative to the <code>ev_now ()</code>
994 time. This is usually the right thing as this timestamp refers to the time
995 of the event triggering whatever timeout you are modifying/starting. If
996 you suspect event processing to be delayed and you <i>need</i> to base the timeout
997 on the current time, use something like this to adjust for this:</p>
998 <pre> ev_timer_set (&amp;timer, after + ev_now () - ev_time (), 0.);
999
1000 </pre>
1001 <p>The callback is guarenteed to be invoked only when its timeout has passed,
1002 but if multiple timers become ready during the same loop iteration then
1003 order of execution is undefined.</p>
1004 <dl>
1005 <dt>ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)</dt>
1006 <dt>ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)</dt>
1007 <dd>
1008 <p>Configure the timer to trigger after <code>after</code> seconds. If <code>repeat</code> is
1009 <code>0.</code>, then it will automatically be stopped. If it is positive, then the
1010 timer will automatically be configured to trigger again <code>repeat</code> seconds
1011 later, again, and again, until stopped manually.</p>
1012 <p>The timer itself will do a best-effort at avoiding drift, that is, if you
1013 configure a timer to trigger every 10 seconds, then it will trigger at
1014 exactly 10 second intervals. If, however, your program cannot keep up with
1015 the timer (because it takes longer than those 10 seconds to do stuff) the
1016 timer will not fire more than once per event loop iteration.</p>
1017 </dd>
1018 <dt>ev_timer_again (loop)</dt>
1019 <dd>
1020 <p>This will act as if the timer timed out and restart it again if it is
1021 repeating. The exact semantics are:</p>
1022 <p>If the timer is pending, its pending status is cleared.</p>
1023 <p>If the timer is started but nonrepeating, stop it (as if it timed out).</p>
1024 <p>If the timer is repeating, either start it if necessary (with the
1025 <code>repeat</code> value), or reset the running timer to the <code>repeat</code> value.</p>
1026 <p>This sounds a bit complicated, but here is a useful and typical
1027 example: Imagine you have a tcp connection and you want a so-called idle
1028 timeout, that is, you want to be called when there have been, say, 60
1029 seconds of inactivity on the socket. The easiest way to do this is to
1030 configure an <code>ev_timer</code> with a <code>repeat</code> value of <code>60</code> and then call
1031 <code>ev_timer_again</code> each time you successfully read or write some data. If
1032 you go into an idle state where you do not expect data to travel on the
1033 socket, you can <code>ev_timer_stop</code> the timer, and <code>ev_timer_again</code> will
1034 automatically restart it if need be.</p>
1035 <p>That means you can ignore the <code>after</code> value and <code>ev_timer_start</code>
1036 altogether and only ever use the <code>repeat</code> value and <code>ev_timer_again</code>:</p>
1037 <pre> ev_timer_init (timer, callback, 0., 5.);
1038 ev_timer_again (loop, timer);
1039 ...
1040 timer-&gt;again = 17.;
1041 ev_timer_again (loop, timer);
1042 ...
1043 timer-&gt;again = 10.;
1044 ev_timer_again (loop, timer);
1045
1046 </pre>
1047 <p>This is more slightly efficient then stopping/starting the timer each time
1048 you want to modify its timeout value.</p>
1049 </dd>
1050 <dt>ev_tstamp repeat [read-write]</dt>
1051 <dd>
1052 <p>The current <code>repeat</code> value. Will be used each time the watcher times out
1053 or <code>ev_timer_again</code> is called and determines the next timeout (if any),
1054 which is also when any modifications are taken into account.</p>
1055 </dd>
1056 </dl>
1057 <p>Example: Create a timer that fires after 60 seconds.</p>
1058 <pre> static void
1059 one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
1060 {
1061 .. one minute over, w is actually stopped right here
1062 }
1063
1064 struct ev_timer mytimer;
1065 ev_timer_init (&amp;mytimer, one_minute_cb, 60., 0.);
1066 ev_timer_start (loop, &amp;mytimer);
1067
1068 </pre>
1069 <p>Example: Create a timeout timer that times out after 10 seconds of
1070 inactivity.</p>
1071 <pre> static void
1072 timeout_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
1073 {
1074 .. ten seconds without any activity
1075 }
1076
1077 struct ev_timer mytimer;
1078 ev_timer_init (&amp;mytimer, timeout_cb, 0., 10.); /* note, only repeat used */
1079 ev_timer_again (&amp;mytimer); /* start timer */
1080 ev_loop (loop, 0);
1081
1082 // and in some piece of code that gets executed on any &quot;activity&quot;:
1083 // reset the timeout to start ticking again at 10 seconds
1084 ev_timer_again (&amp;mytimer);
1085
1086
1087
1088
1089 </pre>
1090
1091 </div>
1092 <h2 id="code_ev_periodic_code_to_cron_or_not"><code>ev_periodic</code> - to cron or not to cron?</h2>
1093 <div id="code_ev_periodic_code_to_cron_or_not-2">
1094 <p>Periodic watchers are also timers of a kind, but they are very versatile
1095 (and unfortunately a bit complex).</p>
1096 <p>Unlike <code>ev_timer</code>'s, they are not based on real time (or relative time)
1097 but on wallclock time (absolute time). You can tell a periodic watcher
1098 to trigger &quot;at&quot; some specific point in time. For example, if you tell a
1099 periodic watcher to trigger in 10 seconds (by specifiying e.g. <code>ev_now ()
1100 + 10.</code>) and then reset your system clock to the last year, then it will
1101 take a year to trigger the event (unlike an <code>ev_timer</code>, which would trigger
1102 roughly 10 seconds later).</p>
1103 <p>They can also be used to implement vastly more complex timers, such as
1104 triggering an event on each midnight, local time or other, complicated,
1105 rules.</p>
1106 <p>As with timers, the callback is guarenteed to be invoked only when the
1107 time (<code>at</code>) has been passed, but if multiple periodic timers become ready
1108 during the same loop iteration then order of execution is undefined.</p>
1109 <dl>
1110 <dt>ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)</dt>
1111 <dt>ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)</dt>
1112 <dd>
1113 <p>Lots of arguments, lets sort it out... There are basically three modes of
1114 operation, and we will explain them from simplest to complex:</p>
1115 <p>
1116 <dl>
1117 <dt>* absolute timer (at = time, interval = reschedule_cb = 0)</dt>
1118 <dd>
1119 <p>In this configuration the watcher triggers an event at the wallclock time
1120 <code>at</code> and doesn't repeat. It will not adjust when a time jump occurs,
1121 that is, if it is to be run at January 1st 2011 then it will run when the
1122 system time reaches or surpasses this time.</p>
1123 </dd>
1124 <dt>* non-repeating interval timer (at = offset, interval &gt; 0, reschedule_cb = 0)</dt>
1125 <dd>
1126 <p>In this mode the watcher will always be scheduled to time out at the next
1127 <code>at + N * interval</code> time (for some integer N, which can also be negative)
1128 and then repeat, regardless of any time jumps.</p>
1129 <p>This can be used to create timers that do not drift with respect to system
1130 time:</p>
1131 <pre> ev_periodic_set (&amp;periodic, 0., 3600., 0);
1132
1133 </pre>
1134 <p>This doesn't mean there will always be 3600 seconds in between triggers,
1135 but only that the the callback will be called when the system time shows a
1136 full hour (UTC), or more correctly, when the system time is evenly divisible
1137 by 3600.</p>
1138 <p>Another way to think about it (for the mathematically inclined) is that
1139 <code>ev_periodic</code> will try to run the callback in this mode at the next possible
1140 time where <code>time = at (mod interval)</code>, regardless of any time jumps.</p>
1141 <p>For numerical stability it is preferable that the <code>at</code> value is near
1142 <code>ev_now ()</code> (the current time), but there is no range requirement for
1143 this value.</p>
1144 </dd>
1145 <dt>* manual reschedule mode (at and interval ignored, reschedule_cb = callback)</dt>
1146 <dd>
1147 <p>In this mode the values for <code>interval</code> and <code>at</code> are both being
1148 ignored. Instead, each time the periodic watcher gets scheduled, the
1149 reschedule callback will be called with the watcher as first, and the
1150 current time as second argument.</p>
1151 <p>NOTE: <i>This callback MUST NOT stop or destroy any periodic watcher,
1152 ever, or make any event loop modifications</i>. If you need to stop it,
1153 return <code>now + 1e30</code> (or so, fudge fudge) and stop it afterwards (e.g. by
1154 starting an <code>ev_prepare</code> watcher, which is legal).</p>
1155 <p>Its prototype is <code>ev_tstamp (*reschedule_cb)(struct ev_periodic *w,
1156 ev_tstamp now)</code>, e.g.:</p>
1157 <pre> static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)
1158 {
1159 return now + 60.;
1160 }
1161
1162 </pre>
1163 <p>It must return the next time to trigger, based on the passed time value
1164 (that is, the lowest time value larger than to the second argument). It
1165 will usually be called just before the callback will be triggered, but
1166 might be called at other times, too.</p>
1167 <p>NOTE: <i>This callback must always return a time that is later than the
1168 passed <code>now</code> value</i>. Not even <code>now</code> itself will do, it <i>must</i> be larger.</p>
1169 <p>This can be used to create very complex timers, such as a timer that
1170 triggers on each midnight, local time. To do this, you would calculate the
1171 next midnight after <code>now</code> and return the timestamp value for this. How
1172 you do this is, again, up to you (but it is not trivial, which is the main
1173 reason I omitted it as an example).</p>
1174 </dd>
1175 </dl>
1176 </p>
1177 </dd>
1178 <dt>ev_periodic_again (loop, ev_periodic *)</dt>
1179 <dd>
1180 <p>Simply stops and restarts the periodic watcher again. This is only useful
1181 when you changed some parameters or the reschedule callback would return
1182 a different time than the last time it was called (e.g. in a crond like
1183 program when the crontabs have changed).</p>
1184 </dd>
1185 <dt>ev_tstamp offset [read-write]</dt>
1186 <dd>
1187 <p>When repeating, this contains the offset value, otherwise this is the
1188 absolute point in time (the <code>at</code> value passed to <code>ev_periodic_set</code>).</p>
1189 <p>Can be modified any time, but changes only take effect when the periodic
1190 timer fires or <code>ev_periodic_again</code> is being called.</p>
1191 </dd>
1192 <dt>ev_tstamp interval [read-write]</dt>
1193 <dd>
1194 <p>The current interval value. Can be modified any time, but changes only
1195 take effect when the periodic timer fires or <code>ev_periodic_again</code> is being
1196 called.</p>
1197 </dd>
1198 <dt>ev_tstamp (*reschedule_cb)(struct ev_periodic *w, ev_tstamp now) [read-write]</dt>
1199 <dd>
1200 <p>The current reschedule callback, or <code>0</code>, if this functionality is
1201 switched off. Can be changed any time, but changes only take effect when
1202 the periodic timer fires or <code>ev_periodic_again</code> is being called.</p>
1203 </dd>
1204 </dl>
1205 <p>Example: Call a callback every hour, or, more precisely, whenever the
1206 system clock is divisible by 3600. The callback invocation times have
1207 potentially a lot of jittering, but good long-term stability.</p>
1208 <pre> static void
1209 clock_cb (struct ev_loop *loop, struct ev_io *w, int revents)
1210 {
1211 ... its now a full hour (UTC, or TAI or whatever your clock follows)
1212 }
1213
1214 struct ev_periodic hourly_tick;
1215 ev_periodic_init (&amp;hourly_tick, clock_cb, 0., 3600., 0);
1216 ev_periodic_start (loop, &amp;hourly_tick);
1217
1218 </pre>
1219 <p>Example: The same as above, but use a reschedule callback to do it:</p>
1220 <pre> #include &lt;math.h&gt;
1221
1222 static ev_tstamp
1223 my_scheduler_cb (struct ev_periodic *w, ev_tstamp now)
1224 {
1225 return fmod (now, 3600.) + 3600.;
1226 }
1227
1228 ev_periodic_init (&amp;hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
1229
1230 </pre>
1231 <p>Example: Call a callback every hour, starting now:</p>
1232 <pre> struct ev_periodic hourly_tick;
1233 ev_periodic_init (&amp;hourly_tick, clock_cb,
1234 fmod (ev_now (loop), 3600.), 3600., 0);
1235 ev_periodic_start (loop, &amp;hourly_tick);
1236
1237
1238
1239
1240 </pre>
1241
1242 </div>
1243 <h2 id="code_ev_signal_code_signal_me_when_a"><code>ev_signal</code> - signal me when a signal gets signalled!</h2>
1244 <div id="code_ev_signal_code_signal_me_when_a-2">
1245 <p>Signal watchers will trigger an event when the process receives a specific
1246 signal one or more times. Even though signals are very asynchronous, libev
1247 will try it's best to deliver signals synchronously, i.e. as part of the
1248 normal event processing, like any other event.</p>
1249 <p>You can configure as many watchers as you like per signal. Only when the
1250 first watcher gets started will libev actually register a signal watcher
1251 with the kernel (thus it coexists with your own signal handlers as long
1252 as you don't register any with libev). Similarly, when the last signal
1253 watcher for a signal is stopped libev will reset the signal handler to
1254 SIG_DFL (regardless of what it was set to before).</p>
1255 <dl>
1256 <dt>ev_signal_init (ev_signal *, callback, int signum)</dt>
1257 <dt>ev_signal_set (ev_signal *, int signum)</dt>
1258 <dd>
1259 <p>Configures the watcher to trigger on the given signal number (usually one
1260 of the <code>SIGxxx</code> constants).</p>
1261 </dd>
1262 <dt>int signum [read-only]</dt>
1263 <dd>
1264 <p>The signal the watcher watches out for.</p>
1265 </dd>
1266 </dl>
1267
1268
1269
1270
1271
1272 </div>
1273 <h2 id="code_ev_child_code_watch_out_for_pro"><code>ev_child</code> - watch out for process status changes</h2>
1274 <div id="code_ev_child_code_watch_out_for_pro-2">
1275 <p>Child watchers trigger when your process receives a SIGCHLD in response to
1276 some child status changes (most typically when a child of yours dies).</p>
1277 <dl>
1278 <dt>ev_child_init (ev_child *, callback, int pid)</dt>
1279 <dt>ev_child_set (ev_child *, int pid)</dt>
1280 <dd>
1281 <p>Configures the watcher to wait for status changes of process <code>pid</code> (or
1282 <i>any</i> process if <code>pid</code> is specified as <code>0</code>). The callback can look
1283 at the <code>rstatus</code> member of the <code>ev_child</code> watcher structure to see
1284 the status word (use the macros from <code>sys/wait.h</code> and see your systems
1285 <code>waitpid</code> documentation). The <code>rpid</code> member contains the pid of the
1286 process causing the status change.</p>
1287 </dd>
1288 <dt>int pid [read-only]</dt>
1289 <dd>
1290 <p>The process id this watcher watches out for, or <code>0</code>, meaning any process id.</p>
1291 </dd>
1292 <dt>int rpid [read-write]</dt>
1293 <dd>
1294 <p>The process id that detected a status change.</p>
1295 </dd>
1296 <dt>int rstatus [read-write]</dt>
1297 <dd>
1298 <p>The process exit/trace status caused by <code>rpid</code> (see your systems
1299 <code>waitpid</code> and <code>sys/wait.h</code> documentation for details).</p>
1300 </dd>
1301 </dl>
1302 <p>Example: Try to exit cleanly on SIGINT and SIGTERM.</p>
1303 <pre> static void
1304 sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)
1305 {
1306 ev_unloop (loop, EVUNLOOP_ALL);
1307 }
1308
1309 struct ev_signal signal_watcher;
1310 ev_signal_init (&amp;signal_watcher, sigint_cb, SIGINT);
1311 ev_signal_start (loop, &amp;sigint_cb);
1312
1313
1314
1315
1316 </pre>
1317
1318 </div>
1319 <h2 id="code_ev_stat_code_did_the_file_attri"><code>ev_stat</code> - did the file attributes just change?</h2>
1320 <div id="code_ev_stat_code_did_the_file_attri-2">
1321 <p>This watches a filesystem path for attribute changes. That is, it calls
1322 <code>stat</code> regularly (or when the OS says it changed) and sees if it changed
1323 compared to the last time, invoking the callback if it did.</p>
1324 <p>The path does not need to exist: changing from &quot;path exists&quot; to &quot;path does
1325 not exist&quot; is a status change like any other. The condition &quot;path does
1326 not exist&quot; is signified by the <code>st_nlink</code> field being zero (which is
1327 otherwise always forced to be at least one) and all the other fields of
1328 the stat buffer having unspecified contents.</p>
1329 <p>The path <i>should</i> be absolute and <i>must not</i> end in a slash. If it is
1330 relative and your working directory changes, the behaviour is undefined.</p>
1331 <p>Since there is no standard to do this, the portable implementation simply
1332 calls <code>stat (2)</code> regularly on the path to see if it changed somehow. You
1333 can specify a recommended polling interval for this case. If you specify
1334 a polling interval of <code>0</code> (highly recommended!) then a <i>suitable,
1335 unspecified default</i> value will be used (which you can expect to be around
1336 five seconds, although this might change dynamically). Libev will also
1337 impose a minimum interval which is currently around <code>0.1</code>, but thats
1338 usually overkill.</p>
1339 <p>This watcher type is not meant for massive numbers of stat watchers,
1340 as even with OS-supported change notifications, this can be
1341 resource-intensive.</p>
1342 <p>At the time of this writing, only the Linux inotify interface is
1343 implemented (implementing kqueue support is left as an exercise for the
1344 reader). Inotify will be used to give hints only and should not change the
1345 semantics of <code>ev_stat</code> watchers, which means that libev sometimes needs
1346 to fall back to regular polling again even with inotify, but changes are
1347 usually detected immediately, and if the file exists there will be no
1348 polling.</p>
1349 <dl>
1350 <dt>ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)</dt>
1351 <dt>ev_stat_set (ev_stat *, const char *path, ev_tstamp interval)</dt>
1352 <dd>
1353 <p>Configures the watcher to wait for status changes of the given
1354 <code>path</code>. The <code>interval</code> is a hint on how quickly a change is expected to
1355 be detected and should normally be specified as <code>0</code> to let libev choose
1356 a suitable value. The memory pointed to by <code>path</code> must point to the same
1357 path for as long as the watcher is active.</p>
1358 <p>The callback will be receive <code>EV_STAT</code> when a change was detected,
1359 relative to the attributes at the time the watcher was started (or the
1360 last change was detected).</p>
1361 </dd>
1362 <dt>ev_stat_stat (ev_stat *)</dt>
1363 <dd>
1364 <p>Updates the stat buffer immediately with new values. If you change the
1365 watched path in your callback, you could call this fucntion to avoid
1366 detecting this change (while introducing a race condition). Can also be
1367 useful simply to find out the new values.</p>
1368 </dd>
1369 <dt>ev_statdata attr [read-only]</dt>
1370 <dd>
1371 <p>The most-recently detected attributes of the file. Although the type is of
1372 <code>ev_statdata</code>, this is usually the (or one of the) <code>struct stat</code> types
1373 suitable for your system. If the <code>st_nlink</code> member is <code>0</code>, then there
1374 was some error while <code>stat</code>ing the file.</p>
1375 </dd>
1376 <dt>ev_statdata prev [read-only]</dt>
1377 <dd>
1378 <p>The previous attributes of the file. The callback gets invoked whenever
1379 <code>prev</code> != <code>attr</code>.</p>
1380 </dd>
1381 <dt>ev_tstamp interval [read-only]</dt>
1382 <dd>
1383 <p>The specified interval.</p>
1384 </dd>
1385 <dt>const char *path [read-only]</dt>
1386 <dd>
1387 <p>The filesystem path that is being watched.</p>
1388 </dd>
1389 </dl>
1390 <p>Example: Watch <code>/etc/passwd</code> for attribute changes.</p>
1391 <pre> static void
1392 passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)
1393 {
1394 /* /etc/passwd changed in some way */
1395 if (w-&gt;attr.st_nlink)
1396 {
1397 printf (&quot;passwd current size %ld\n&quot;, (long)w-&gt;attr.st_size);
1398 printf (&quot;passwd current atime %ld\n&quot;, (long)w-&gt;attr.st_mtime);
1399 printf (&quot;passwd current mtime %ld\n&quot;, (long)w-&gt;attr.st_mtime);
1400 }
1401 else
1402 /* you shalt not abuse printf for puts */
1403 puts (&quot;wow, /etc/passwd is not there, expect problems. &quot;
1404 &quot;if this is windows, they already arrived\n&quot;);
1405 }
1406
1407 ...
1408 ev_stat passwd;
1409
1410 ev_stat_init (&amp;passwd, passwd_cb, &quot;/etc/passwd&quot;);
1411 ev_stat_start (loop, &amp;passwd);
1412
1413
1414
1415
1416 </pre>
1417
1418 </div>
1419 <h2 id="code_ev_idle_code_when_you_ve_got_no"><code>ev_idle</code> - when you've got nothing better to do...</h2>
1420 <div id="code_ev_idle_code_when_you_ve_got_no-2">
1421 <p>Idle watchers trigger events when no other events of the same or higher
1422 priority are pending (prepare, check and other idle watchers do not
1423 count).</p>
1424 <p>That is, as long as your process is busy handling sockets or timeouts
1425 (or even signals, imagine) of the same or higher priority it will not be
1426 triggered. But when your process is idle (or only lower-priority watchers
1427 are pending), the idle watchers are being called once per event loop
1428 iteration - until stopped, that is, or your process receives more events
1429 and becomes busy again with higher priority stuff.</p>
1430 <p>The most noteworthy effect is that as long as any idle watchers are
1431 active, the process will not block when waiting for new events.</p>
1432 <p>Apart from keeping your process non-blocking (which is a useful
1433 effect on its own sometimes), idle watchers are a good place to do
1434 &quot;pseudo-background processing&quot;, or delay processing stuff to after the
1435 event loop has handled all outstanding events.</p>
1436 <dl>
1437 <dt>ev_idle_init (ev_signal *, callback)</dt>
1438 <dd>
1439 <p>Initialises and configures the idle watcher - it has no parameters of any
1440 kind. There is a <code>ev_idle_set</code> macro, but using it is utterly pointless,
1441 believe me.</p>
1442 </dd>
1443 </dl>
1444 <p>Example: Dynamically allocate an <code>ev_idle</code> watcher, start it, and in the
1445 callback, free it. Also, use no error checking, as usual.</p>
1446 <pre> static void
1447 idle_cb (struct ev_loop *loop, struct ev_idle *w, int revents)
1448 {
1449 free (w);
1450 // now do something you wanted to do when the program has
1451 // no longer asnything immediate to do.
1452 }
1453
1454 struct ev_idle *idle_watcher = malloc (sizeof (struct ev_idle));
1455 ev_idle_init (idle_watcher, idle_cb);
1456 ev_idle_start (loop, idle_cb);
1457
1458
1459
1460
1461 </pre>
1462
1463 </div>
1464 <h2 id="code_ev_prepare_code_and_code_ev_che"><code>ev_prepare</code> and <code>ev_check</code> - customise your event loop!</h2>
1465 <div id="code_ev_prepare_code_and_code_ev_che-2">
1466 <p>Prepare and check watchers are usually (but not always) used in tandem:
1467 prepare watchers get invoked before the process blocks and check watchers
1468 afterwards.</p>
1469 <p>You <i>must not</i> call <code>ev_loop</code> or similar functions that enter
1470 the current event loop from either <code>ev_prepare</code> or <code>ev_check</code>
1471 watchers. Other loops than the current one are fine, however. The
1472 rationale behind this is that you do not need to check for recursion in
1473 those watchers, i.e. the sequence will always be <code>ev_prepare</code>, blocking,
1474 <code>ev_check</code> so if you have one watcher of each kind they will always be
1475 called in pairs bracketing the blocking call.</p>
1476 <p>Their main purpose is to integrate other event mechanisms into libev and
1477 their use is somewhat advanced. This could be used, for example, to track
1478 variable changes, implement your own watchers, integrate net-snmp or a
1479 coroutine library and lots more. They are also occasionally useful if
1480 you cache some data and want to flush it before blocking (for example,
1481 in X programs you might want to do an <code>XFlush ()</code> in an <code>ev_prepare</code>
1482 watcher).</p>
1483 <p>This is done by examining in each prepare call which file descriptors need
1484 to be watched by the other library, registering <code>ev_io</code> watchers for
1485 them and starting an <code>ev_timer</code> watcher for any timeouts (many libraries
1486 provide just this functionality). Then, in the check watcher you check for
1487 any events that occured (by checking the pending status of all watchers
1488 and stopping them) and call back into the library. The I/O and timer
1489 callbacks will never actually be called (but must be valid nevertheless,
1490 because you never know, you know?).</p>
1491 <p>As another example, the Perl Coro module uses these hooks to integrate
1492 coroutines into libev programs, by yielding to other active coroutines
1493 during each prepare and only letting the process block if no coroutines
1494 are ready to run (it's actually more complicated: it only runs coroutines
1495 with priority higher than or equal to the event loop and one coroutine
1496 of lower priority, but only once, using idle watchers to keep the event
1497 loop from blocking if lower-priority coroutines are active, thus mapping
1498 low-priority coroutines to idle/background tasks).</p>
1499 <p>It is recommended to give <code>ev_check</code> watchers highest (<code>EV_MAXPRI</code>)
1500 priority, to ensure that they are being run before any other watchers
1501 after the poll. Also, <code>ev_check</code> watchers (and <code>ev_prepare</code> watchers,
1502 too) should not activate (&quot;feed&quot;) events into libev. While libev fully
1503 supports this, they will be called before other <code>ev_check</code> watchers did
1504 their job. As <code>ev_check</code> watchers are often used to embed other event
1505 loops those other event loops might be in an unusable state until their
1506 <code>ev_check</code> watcher ran (always remind yourself to coexist peacefully with
1507 others).</p>
1508 <dl>
1509 <dt>ev_prepare_init (ev_prepare *, callback)</dt>
1510 <dt>ev_check_init (ev_check *, callback)</dt>
1511 <dd>
1512 <p>Initialises and configures the prepare or check watcher - they have no
1513 parameters of any kind. There are <code>ev_prepare_set</code> and <code>ev_check_set</code>
1514 macros, but using them is utterly, utterly and completely pointless.</p>
1515 </dd>
1516 </dl>
1517 <p>There are a number of principal ways to embed other event loops or modules
1518 into libev. Here are some ideas on how to include libadns into libev
1519 (there is a Perl module named <code>EV::ADNS</code> that does this, which you could
1520 use for an actually working example. Another Perl module named <code>EV::Glib</code>
1521 embeds a Glib main context into libev, and finally, <code>Glib::EV</code> embeds EV
1522 into the Glib event loop).</p>
1523 <p>Method 1: Add IO watchers and a timeout watcher in a prepare handler,
1524 and in a check watcher, destroy them and call into libadns. What follows
1525 is pseudo-code only of course. This requires you to either use a low
1526 priority for the check watcher or use <code>ev_clear_pending</code> explicitly, as
1527 the callbacks for the IO/timeout watchers might not have been called yet.</p>
1528 <pre> static ev_io iow [nfd];
1529 static ev_timer tw;
1530
1531 static void
1532 io_cb (ev_loop *loop, ev_io *w, int revents)
1533 {
1534 }
1535
1536 // create io watchers for each fd and a timer before blocking
1537 static void
1538 adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents)
1539 {
1540 int timeout = 3600000;
1541 struct pollfd fds [nfd];
1542 // actual code will need to loop here and realloc etc.
1543 adns_beforepoll (ads, fds, &amp;nfd, &amp;timeout, timeval_from (ev_time ()));
1544
1545 /* the callback is illegal, but won't be called as we stop during check */
1546 ev_timer_init (&amp;tw, 0, timeout * 1e-3);
1547 ev_timer_start (loop, &amp;tw);
1548
1549 // create one ev_io per pollfd
1550 for (int i = 0; i &lt; nfd; ++i)
1551 {
1552 ev_io_init (iow + i, io_cb, fds [i].fd,
1553 ((fds [i].events &amp; POLLIN ? EV_READ : 0)
1554 | (fds [i].events &amp; POLLOUT ? EV_WRITE : 0)));
1555
1556 fds [i].revents = 0;
1557 ev_io_start (loop, iow + i);
1558 }
1559 }
1560
1561 // stop all watchers after blocking
1562 static void
1563 adns_check_cb (ev_loop *loop, ev_check *w, int revents)
1564 {
1565 ev_timer_stop (loop, &amp;tw);
1566
1567 for (int i = 0; i &lt; nfd; ++i)
1568 {
1569 // set the relevant poll flags
1570 // could also call adns_processreadable etc. here
1571 struct pollfd *fd = fds + i;
1572 int revents = ev_clear_pending (iow + i);
1573 if (revents &amp; EV_READ ) fd-&gt;revents |= fd-&gt;events &amp; POLLIN;
1574 if (revents &amp; EV_WRITE) fd-&gt;revents |= fd-&gt;events &amp; POLLOUT;
1575
1576 // now stop the watcher
1577 ev_io_stop (loop, iow + i);
1578 }
1579
1580 adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));
1581 }
1582
1583 </pre>
1584 <p>Method 2: This would be just like method 1, but you run <code>adns_afterpoll</code>
1585 in the prepare watcher and would dispose of the check watcher.</p>
1586 <p>Method 3: If the module to be embedded supports explicit event
1587 notification (adns does), you can also make use of the actual watcher
1588 callbacks, and only destroy/create the watchers in the prepare watcher.</p>
1589 <pre> static void
1590 timer_cb (EV_P_ ev_timer *w, int revents)
1591 {
1592 adns_state ads = (adns_state)w-&gt;data;
1593 update_now (EV_A);
1594
1595 adns_processtimeouts (ads, &amp;tv_now);
1596 }
1597
1598 static void
1599 io_cb (EV_P_ ev_io *w, int revents)
1600 {
1601 adns_state ads = (adns_state)w-&gt;data;
1602 update_now (EV_A);
1603
1604 if (revents &amp; EV_READ ) adns_processreadable (ads, w-&gt;fd, &amp;tv_now);
1605 if (revents &amp; EV_WRITE) adns_processwriteable (ads, w-&gt;fd, &amp;tv_now);
1606 }
1607
1608 // do not ever call adns_afterpoll
1609
1610 </pre>
1611 <p>Method 4: Do not use a prepare or check watcher because the module you
1612 want to embed is too inflexible to support it. Instead, youc na override
1613 their poll function. The drawback with this solution is that the main
1614 loop is now no longer controllable by EV. The <code>Glib::EV</code> module does
1615 this.</p>
1616 <pre> static gint
1617 event_poll_func (GPollFD *fds, guint nfds, gint timeout)
1618 {
1619 int got_events = 0;
1620
1621 for (n = 0; n &lt; nfds; ++n)
1622 // create/start io watcher that sets the relevant bits in fds[n] and increment got_events
1623
1624 if (timeout &gt;= 0)
1625 // create/start timer
1626
1627 // poll
1628 ev_loop (EV_A_ 0);
1629
1630 // stop timer again
1631 if (timeout &gt;= 0)
1632 ev_timer_stop (EV_A_ &amp;to);
1633
1634 // stop io watchers again - their callbacks should have set
1635 for (n = 0; n &lt; nfds; ++n)
1636 ev_io_stop (EV_A_ iow [n]);
1637
1638 return got_events;
1639 }
1640
1641
1642
1643
1644 </pre>
1645
1646 </div>
1647 <h2 id="code_ev_embed_code_when_one_backend_"><code>ev_embed</code> - when one backend isn't enough...</h2>
1648 <div id="code_ev_embed_code_when_one_backend_-2">
1649 <p>This is a rather advanced watcher type that lets you embed one event loop
1650 into another (currently only <code>ev_io</code> events are supported in the embedded
1651 loop, other types of watchers might be handled in a delayed or incorrect
1652 fashion and must not be used).</p>
1653 <p>There are primarily two reasons you would want that: work around bugs and
1654 prioritise I/O.</p>
1655 <p>As an example for a bug workaround, the kqueue backend might only support
1656 sockets on some platform, so it is unusable as generic backend, but you
1657 still want to make use of it because you have many sockets and it scales
1658 so nicely. In this case, you would create a kqueue-based loop and embed it
1659 into your default loop (which might use e.g. poll). Overall operation will
1660 be a bit slower because first libev has to poll and then call kevent, but
1661 at least you can use both at what they are best.</p>
1662 <p>As for prioritising I/O: rarely you have the case where some fds have
1663 to be watched and handled very quickly (with low latency), and even
1664 priorities and idle watchers might have too much overhead. In this case
1665 you would put all the high priority stuff in one loop and all the rest in
1666 a second one, and embed the second one in the first.</p>
1667 <p>As long as the watcher is active, the callback will be invoked every time
1668 there might be events pending in the embedded loop. The callback must then
1669 call <code>ev_embed_sweep (mainloop, watcher)</code> to make a single sweep and invoke
1670 their callbacks (you could also start an idle watcher to give the embedded
1671 loop strictly lower priority for example). You can also set the callback
1672 to <code>0</code>, in which case the embed watcher will automatically execute the
1673 embedded loop sweep.</p>
1674 <p>As long as the watcher is started it will automatically handle events. The
1675 callback will be invoked whenever some events have been handled. You can
1676 set the callback to <code>0</code> to avoid having to specify one if you are not
1677 interested in that.</p>
1678 <p>Also, there have not currently been made special provisions for forking:
1679 when you fork, you not only have to call <code>ev_loop_fork</code> on both loops,
1680 but you will also have to stop and restart any <code>ev_embed</code> watchers
1681 yourself.</p>
1682 <p>Unfortunately, not all backends are embeddable, only the ones returned by
1683 <code>ev_embeddable_backends</code> are, which, unfortunately, does not include any
1684 portable one.</p>
1685 <p>So when you want to use this feature you will always have to be prepared
1686 that you cannot get an embeddable loop. The recommended way to get around
1687 this is to have a separate variables for your embeddable loop, try to
1688 create it, and if that fails, use the normal loop for everything:</p>
1689 <pre> struct ev_loop *loop_hi = ev_default_init (0);
1690 struct ev_loop *loop_lo = 0;
1691 struct ev_embed embed;
1692
1693 // see if there is a chance of getting one that works
1694 // (remember that a flags value of 0 means autodetection)
1695 loop_lo = ev_embeddable_backends () &amp; ev_recommended_backends ()
1696 ? ev_loop_new (ev_embeddable_backends () &amp; ev_recommended_backends ())
1697 : 0;
1698
1699 // if we got one, then embed it, otherwise default to loop_hi
1700 if (loop_lo)
1701 {
1702 ev_embed_init (&amp;embed, 0, loop_lo);
1703 ev_embed_start (loop_hi, &amp;embed);
1704 }
1705 else
1706 loop_lo = loop_hi;
1707
1708 </pre>
1709 <dl>
1710 <dt>ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)</dt>
1711 <dt>ev_embed_set (ev_embed *, callback, struct ev_loop *embedded_loop)</dt>
1712 <dd>
1713 <p>Configures the watcher to embed the given loop, which must be
1714 embeddable. If the callback is <code>0</code>, then <code>ev_embed_sweep</code> will be
1715 invoked automatically, otherwise it is the responsibility of the callback
1716 to invoke it (it will continue to be called until the sweep has been done,
1717 if you do not want thta, you need to temporarily stop the embed watcher).</p>
1718 </dd>
1719 <dt>ev_embed_sweep (loop, ev_embed *)</dt>
1720 <dd>
1721 <p>Make a single, non-blocking sweep over the embedded loop. This works
1722 similarly to <code>ev_loop (embedded_loop, EVLOOP_NONBLOCK)</code>, but in the most
1723 apropriate way for embedded loops.</p>
1724 </dd>
1725 <dt>struct ev_loop *loop [read-only]</dt>
1726 <dd>
1727 <p>The embedded event loop.</p>
1728 </dd>
1729 </dl>
1730
1731
1732
1733
1734
1735 </div>
1736 <h2 id="code_ev_fork_code_the_audacity_to_re"><code>ev_fork</code> - the audacity to resume the event loop after a fork</h2>
1737 <div id="code_ev_fork_code_the_audacity_to_re-2">
1738 <p>Fork watchers are called when a <code>fork ()</code> was detected (usually because
1739 whoever is a good citizen cared to tell libev about it by calling
1740 <code>ev_default_fork</code> or <code>ev_loop_fork</code>). The invocation is done before the
1741 event loop blocks next and before <code>ev_check</code> watchers are being called,
1742 and only in the child after the fork. If whoever good citizen calling
1743 <code>ev_default_fork</code> cheats and calls it in the wrong process, the fork
1744 handlers will be invoked, too, of course.</p>
1745 <dl>
1746 <dt>ev_fork_init (ev_signal *, callback)</dt>
1747 <dd>
1748 <p>Initialises and configures the fork watcher - it has no parameters of any
1749 kind. There is a <code>ev_fork_set</code> macro, but using it is utterly pointless,
1750 believe me.</p>
1751 </dd>
1752 </dl>
1753
1754
1755
1756
1757
1758 </div>
1759 <h1 id="OTHER_FUNCTIONS">OTHER FUNCTIONS</h1>
1760 <div id="OTHER_FUNCTIONS_CONTENT">
1761 <p>There are some other functions of possible interest. Described. Here. Now.</p>
1762 <dl>
1763 <dt>ev_once (loop, int fd, int events, ev_tstamp timeout, callback)</dt>
1764 <dd>
1765 <p>This function combines a simple timer and an I/O watcher, calls your
1766 callback on whichever event happens first and automatically stop both
1767 watchers. This is useful if you want to wait for a single event on an fd
1768 or timeout without having to allocate/configure/start/stop/free one or
1769 more watchers yourself.</p>
1770 <p>If <code>fd</code> is less than 0, then no I/O watcher will be started and events
1771 is being ignored. Otherwise, an <code>ev_io</code> watcher for the given <code>fd</code> and
1772 <code>events</code> set will be craeted and started.</p>
1773 <p>If <code>timeout</code> is less than 0, then no timeout watcher will be
1774 started. Otherwise an <code>ev_timer</code> watcher with after = <code>timeout</code> (and
1775 repeat = 0) will be started. While <code>0</code> is a valid timeout, it is of
1776 dubious value.</p>
1777 <p>The callback has the type <code>void (*cb)(int revents, void *arg)</code> and gets
1778 passed an <code>revents</code> set like normal event callbacks (a combination of
1779 <code>EV_ERROR</code>, <code>EV_READ</code>, <code>EV_WRITE</code> or <code>EV_TIMEOUT</code>) and the <code>arg</code>
1780 value passed to <code>ev_once</code>:</p>
1781 <pre> static void stdin_ready (int revents, void *arg)
1782 {
1783 if (revents &amp; EV_TIMEOUT)
1784 /* doh, nothing entered */;
1785 else if (revents &amp; EV_READ)
1786 /* stdin might have data for us, joy! */;
1787 }
1788
1789 ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
1790
1791 </pre>
1792 </dd>
1793 <dt>ev_feed_event (ev_loop *, watcher *, int revents)</dt>
1794 <dd>
1795 <p>Feeds the given event set into the event loop, as if the specified event
1796 had happened for the specified watcher (which must be a pointer to an
1797 initialised but not necessarily started event watcher).</p>
1798 </dd>
1799 <dt>ev_feed_fd_event (ev_loop *, int fd, int revents)</dt>
1800 <dd>
1801 <p>Feed an event on the given fd, as if a file descriptor backend detected
1802 the given events it.</p>
1803 </dd>
1804 <dt>ev_feed_signal_event (ev_loop *loop, int signum)</dt>
1805 <dd>
1806 <p>Feed an event as if the given signal occured (<code>loop</code> must be the default
1807 loop!).</p>
1808 </dd>
1809 </dl>
1810
1811
1812
1813
1814
1815 </div>
1816 <h1 id="LIBEVENT_EMULATION">LIBEVENT EMULATION</h1>
1817 <div id="LIBEVENT_EMULATION_CONTENT">
1818 <p>Libev offers a compatibility emulation layer for libevent. It cannot
1819 emulate the internals of libevent, so here are some usage hints:</p>
1820 <dl>
1821 <dt>* Use it by including &lt;event.h&gt;, as usual.</dt>
1822 <dt>* The following members are fully supported: ev_base, ev_callback,
1823 ev_arg, ev_fd, ev_res, ev_events.</dt>
1824 <dt>* Avoid using ev_flags and the EVLIST_*-macros, while it is
1825 maintained by libev, it does not work exactly the same way as in libevent (consider
1826 it a private API).</dt>
1827 <dt>* Priorities are not currently supported. Initialising priorities
1828 will fail and all watchers will have the same priority, even though there
1829 is an ev_pri field.</dt>
1830 <dt>* Other members are not supported.</dt>
1831 <dt>* The libev emulation is <i>not</i> ABI compatible to libevent, you need
1832 to use the libev header file and library.</dt>
1833 </dl>
1834
1835 </div>
1836 <h1 id="C_SUPPORT">C++ SUPPORT</h1>
1837 <div id="C_SUPPORT_CONTENT">
1838 <p>Libev comes with some simplistic wrapper classes for C++ that mainly allow
1839 you to use some convinience methods to start/stop watchers and also change
1840 the callback model to a model using method callbacks on objects.</p>
1841 <p>To use it,</p>
1842 <pre> #include &lt;ev++.h&gt;
1843
1844 </pre>
1845 <p>This automatically includes <cite>ev.h</cite> and puts all of its definitions (many
1846 of them macros) into the global namespace. All C++ specific things are
1847 put into the <code>ev</code> namespace. It should support all the same embedding
1848 options as <cite>ev.h</cite>, most notably <code>EV_MULTIPLICITY</code>.</p>
1849 <p>Care has been taken to keep the overhead low. The only data member the C++
1850 classes add (compared to plain C-style watchers) is the event loop pointer
1851 that the watcher is associated with (or no additional members at all if
1852 you disable <code>EV_MULTIPLICITY</code> when embedding libev).</p>
1853 <p>Currently, functions, and static and non-static member functions can be
1854 used as callbacks. Other types should be easy to add as long as they only
1855 need one additional pointer for context. If you need support for other
1856 types of functors please contact the author (preferably after implementing
1857 it).</p>
1858 <p>Here is a list of things available in the <code>ev</code> namespace:</p>
1859 <dl>
1860 <dt><code>ev::READ</code>, <code>ev::WRITE</code> etc.</dt>
1861 <dd>
1862 <p>These are just enum values with the same values as the <code>EV_READ</code> etc.
1863 macros from <cite>ev.h</cite>.</p>
1864 </dd>
1865 <dt><code>ev::tstamp</code>, <code>ev::now</code></dt>
1866 <dd>
1867 <p>Aliases to the same types/functions as with the <code>ev_</code> prefix.</p>
1868 </dd>
1869 <dt><code>ev::io</code>, <code>ev::timer</code>, <code>ev::periodic</code>, <code>ev::idle</code>, <code>ev::sig</code> etc.</dt>
1870 <dd>
1871 <p>For each <code>ev_TYPE</code> watcher in <cite>ev.h</cite> there is a corresponding class of
1872 the same name in the <code>ev</code> namespace, with the exception of <code>ev_signal</code>
1873 which is called <code>ev::sig</code> to avoid clashes with the <code>signal</code> macro
1874 defines by many implementations.</p>
1875 <p>All of those classes have these methods:</p>
1876 <p>
1877 <dl>
1878 <dt>ev::TYPE::TYPE ()</dt>
1879 <dt>ev::TYPE::TYPE (struct ev_loop *)</dt>
1880 <dt>ev::TYPE::~TYPE</dt>
1881 <dd>
1882 <p>The constructor (optionally) takes an event loop to associate the watcher
1883 with. If it is omitted, it will use <code>EV_DEFAULT</code>.</p>
1884 <p>The constructor calls <code>ev_init</code> for you, which means you have to call the
1885 <code>set</code> method before starting it.</p>
1886 <p>It will not set a callback, however: You have to call the templated <code>set</code>
1887 method to set a callback before you can start the watcher.</p>
1888 <p>(The reason why you have to use a method is a limitation in C++ which does
1889 not allow explicit template arguments for constructors).</p>
1890 <p>The destructor automatically stops the watcher if it is active.</p>
1891 </dd>
1892 <dt>w-&gt;set&lt;class, &amp;class::method&gt; (object *)</dt>
1893 <dd>
1894 <p>This method sets the callback method to call. The method has to have a
1895 signature of <code>void (*)(ev_TYPE &amp;, int)</code>, it receives the watcher as
1896 first argument and the <code>revents</code> as second. The object must be given as
1897 parameter and is stored in the <code>data</code> member of the watcher.</p>
1898 <p>This method synthesizes efficient thunking code to call your method from
1899 the C callback that libev requires. If your compiler can inline your
1900 callback (i.e. it is visible to it at the place of the <code>set</code> call and
1901 your compiler is good :), then the method will be fully inlined into the
1902 thunking function, making it as fast as a direct C callback.</p>
1903 <p>Example: simple class declaration and watcher initialisation</p>
1904 <pre> struct myclass
1905 {
1906 void io_cb (ev::io &amp;w, int revents) { }
1907 }
1908
1909 myclass obj;
1910 ev::io iow;
1911 iow.set &lt;myclass, &amp;myclass::io_cb&gt; (&amp;obj);
1912
1913 </pre>
1914 </dd>
1915 <dt>w-&gt;set&lt;function&gt; (void *data = 0)</dt>
1916 <dd>
1917 <p>Also sets a callback, but uses a static method or plain function as
1918 callback. The optional <code>data</code> argument will be stored in the watcher's
1919 <code>data</code> member and is free for you to use.</p>
1920 <p>The prototype of the <code>function</code> must be <code>void (*)(ev::TYPE &amp;w, int)</code>.</p>
1921 <p>See the method-<code>set</code> above for more details.</p>
1922 <p>Example:</p>
1923 <pre> static void io_cb (ev::io &amp;w, int revents) { }
1924 iow.set &lt;io_cb&gt; ();
1925
1926 </pre>
1927 </dd>
1928 <dt>w-&gt;set (struct ev_loop *)</dt>
1929 <dd>
1930 <p>Associates a different <code>struct ev_loop</code> with this watcher. You can only
1931 do this when the watcher is inactive (and not pending either).</p>
1932 </dd>
1933 <dt>w-&gt;set ([args])</dt>
1934 <dd>
1935 <p>Basically the same as <code>ev_TYPE_set</code>, with the same args. Must be
1936 called at least once. Unlike the C counterpart, an active watcher gets
1937 automatically stopped and restarted when reconfiguring it with this
1938 method.</p>
1939 </dd>
1940 <dt>w-&gt;start ()</dt>
1941 <dd>
1942 <p>Starts the watcher. Note that there is no <code>loop</code> argument, as the
1943 constructor already stores the event loop.</p>
1944 </dd>
1945 <dt>w-&gt;stop ()</dt>
1946 <dd>
1947 <p>Stops the watcher if it is active. Again, no <code>loop</code> argument.</p>
1948 </dd>
1949 <dt>w-&gt;again () <code>ev::timer</code>, <code>ev::periodic</code> only</dt>
1950 <dd>
1951 <p>For <code>ev::timer</code> and <code>ev::periodic</code>, this invokes the corresponding
1952 <code>ev_TYPE_again</code> function.</p>
1953 </dd>
1954 <dt>w-&gt;sweep () <code>ev::embed</code> only</dt>
1955 <dd>
1956 <p>Invokes <code>ev_embed_sweep</code>.</p>
1957 </dd>
1958 <dt>w-&gt;update () <code>ev::stat</code> only</dt>
1959 <dd>
1960 <p>Invokes <code>ev_stat_stat</code>.</p>
1961 </dd>
1962 </dl>
1963 </p>
1964 </dd>
1965 </dl>
1966 <p>Example: Define a class with an IO and idle watcher, start one of them in
1967 the constructor.</p>
1968 <pre> class myclass
1969 {
1970 ev_io io; void io_cb (ev::io &amp;w, int revents);
1971 ev_idle idle void idle_cb (ev::idle &amp;w, int revents);
1972
1973 myclass ();
1974 }
1975
1976 myclass::myclass (int fd)
1977 {
1978 io .set &lt;myclass, &amp;myclass::io_cb &gt; (this);
1979 idle.set &lt;myclass, &amp;myclass::idle_cb&gt; (this);
1980
1981 io.start (fd, ev::READ);
1982 }
1983
1984
1985
1986
1987 </pre>
1988
1989 </div>
1990 <h1 id="MACRO_MAGIC">MACRO MAGIC</h1>
1991 <div id="MACRO_MAGIC_CONTENT">
1992 <p>Libev can be compiled with a variety of options, the most fundemantal is
1993 <code>EV_MULTIPLICITY</code>. This option determines whether (most) functions and
1994 callbacks have an initial <code>struct ev_loop *</code> argument.</p>
1995 <p>To make it easier to write programs that cope with either variant, the
1996 following macros are defined:</p>
1997 <dl>
1998 <dt><code>EV_A</code>, <code>EV_A_</code></dt>
1999 <dd>
2000 <p>This provides the loop <i>argument</i> for functions, if one is required (&quot;ev
2001 loop argument&quot;). The <code>EV_A</code> form is used when this is the sole argument,
2002 <code>EV_A_</code> is used when other arguments are following. Example:</p>
2003 <pre> ev_unref (EV_A);
2004 ev_timer_add (EV_A_ watcher);
2005 ev_loop (EV_A_ 0);
2006
2007 </pre>
2008 <p>It assumes the variable <code>loop</code> of type <code>struct ev_loop *</code> is in scope,
2009 which is often provided by the following macro.</p>
2010 </dd>
2011 <dt><code>EV_P</code>, <code>EV_P_</code></dt>
2012 <dd>
2013 <p>This provides the loop <i>parameter</i> for functions, if one is required (&quot;ev
2014 loop parameter&quot;). The <code>EV_P</code> form is used when this is the sole parameter,
2015 <code>EV_P_</code> is used when other parameters are following. Example:</p>
2016 <pre> // this is how ev_unref is being declared
2017 static void ev_unref (EV_P);
2018
2019 // this is how you can declare your typical callback
2020 static void cb (EV_P_ ev_timer *w, int revents)
2021
2022 </pre>
2023 <p>It declares a parameter <code>loop</code> of type <code>struct ev_loop *</code>, quite
2024 suitable for use with <code>EV_A</code>.</p>
2025 </dd>
2026 <dt><code>EV_DEFAULT</code>, <code>EV_DEFAULT_</code></dt>
2027 <dd>
2028 <p>Similar to the other two macros, this gives you the value of the default
2029 loop, if multiple loops are supported (&quot;ev loop default&quot;).</p>
2030 </dd>
2031 </dl>
2032 <p>Example: Declare and initialise a check watcher, utilising the above
2033 macros so it will work regardless of whether multiple loops are supported
2034 or not.</p>
2035 <pre> static void
2036 check_cb (EV_P_ ev_timer *w, int revents)
2037 {
2038 ev_check_stop (EV_A_ w);
2039 }
2040
2041 ev_check check;
2042 ev_check_init (&amp;check, check_cb);
2043 ev_check_start (EV_DEFAULT_ &amp;check);
2044 ev_loop (EV_DEFAULT_ 0);
2045
2046 </pre>
2047
2048 </div>
2049 <h1 id="EMBEDDING">EMBEDDING</h1>
2050 <div id="EMBEDDING_CONTENT">
2051 <p>Libev can (and often is) directly embedded into host
2052 applications. Examples of applications that embed it include the Deliantra
2053 Game Server, the EV perl module, the GNU Virtual Private Ethernet (gvpe)
2054 and rxvt-unicode.</p>
2055 <p>The goal is to enable you to just copy the neecssary files into your
2056 source directory without having to change even a single line in them, so
2057 you can easily upgrade by simply copying (or having a checked-out copy of
2058 libev somewhere in your source tree).</p>
2059
2060 </div>
2061 <h2 id="FILESETS">FILESETS</h2>
2062 <div id="FILESETS_CONTENT">
2063 <p>Depending on what features you need you need to include one or more sets of files
2064 in your app.</p>
2065
2066 </div>
2067 <h3 id="CORE_EVENT_LOOP">CORE EVENT LOOP</h3>
2068 <div id="CORE_EVENT_LOOP_CONTENT">
2069 <p>To include only the libev core (all the <code>ev_*</code> functions), with manual
2070 configuration (no autoconf):</p>
2071 <pre> #define EV_STANDALONE 1
2072 #include &quot;ev.c&quot;
2073
2074 </pre>
2075 <p>This will automatically include <cite>ev.h</cite>, too, and should be done in a
2076 single C source file only to provide the function implementations. To use
2077 it, do the same for <cite>ev.h</cite> in all files wishing to use this API (best
2078 done by writing a wrapper around <cite>ev.h</cite> that you can include instead and
2079 where you can put other configuration options):</p>
2080 <pre> #define EV_STANDALONE 1
2081 #include &quot;ev.h&quot;
2082
2083 </pre>
2084 <p>Both header files and implementation files can be compiled with a C++
2085 compiler (at least, thats a stated goal, and breakage will be treated
2086 as a bug).</p>
2087 <p>You need the following files in your source tree, or in a directory
2088 in your include path (e.g. in libev/ when using -Ilibev):</p>
2089 <pre> ev.h
2090 ev.c
2091 ev_vars.h
2092 ev_wrap.h
2093
2094 ev_win32.c required on win32 platforms only
2095
2096 ev_select.c only when select backend is enabled (which is enabled by default)
2097 ev_poll.c only when poll backend is enabled (disabled by default)
2098 ev_epoll.c only when the epoll backend is enabled (disabled by default)
2099 ev_kqueue.c only when the kqueue backend is enabled (disabled by default)
2100 ev_port.c only when the solaris port backend is enabled (disabled by default)
2101
2102 </pre>
2103 <p><cite>ev.c</cite> includes the backend files directly when enabled, so you only need
2104 to compile this single file.</p>
2105
2106 </div>
2107 <h3 id="LIBEVENT_COMPATIBILITY_API">LIBEVENT COMPATIBILITY API</h3>
2108 <div id="LIBEVENT_COMPATIBILITY_API_CONTENT">
2109 <p>To include the libevent compatibility API, also include:</p>
2110 <pre> #include &quot;event.c&quot;
2111
2112 </pre>
2113 <p>in the file including <cite>ev.c</cite>, and:</p>
2114 <pre> #include &quot;event.h&quot;
2115
2116 </pre>
2117 <p>in the files that want to use the libevent API. This also includes <cite>ev.h</cite>.</p>
2118 <p>You need the following additional files for this:</p>
2119 <pre> event.h
2120 event.c
2121
2122 </pre>
2123
2124 </div>
2125 <h3 id="AUTOCONF_SUPPORT">AUTOCONF SUPPORT</h3>
2126 <div id="AUTOCONF_SUPPORT_CONTENT">
2127 <p>Instead of using <code>EV_STANDALONE=1</code> and providing your config in
2128 whatever way you want, you can also <code>m4_include([libev.m4])</code> in your
2129 <cite>configure.ac</cite> and leave <code>EV_STANDALONE</code> undefined. <cite>ev.c</cite> will then
2130 include <cite>config.h</cite> and configure itself accordingly.</p>
2131 <p>For this of course you need the m4 file:</p>
2132 <pre> libev.m4
2133
2134 </pre>
2135
2136 </div>
2137 <h2 id="PREPROCESSOR_SYMBOLS_MACROS">PREPROCESSOR SYMBOLS/MACROS</h2>
2138 <div id="PREPROCESSOR_SYMBOLS_MACROS_CONTENT">
2139 <p>Libev can be configured via a variety of preprocessor symbols you have to define
2140 before including any of its files. The default is not to build for multiplicity
2141 and only include the select backend.</p>
2142 <dl>
2143 <dt>EV_STANDALONE</dt>
2144 <dd>
2145 <p>Must always be <code>1</code> if you do not use autoconf configuration, which
2146 keeps libev from including <cite>config.h</cite>, and it also defines dummy
2147 implementations for some libevent functions (such as logging, which is not
2148 supported). It will also not define any of the structs usually found in
2149 <cite>event.h</cite> that are not directly supported by the libev core alone.</p>
2150 </dd>
2151 <dt>EV_USE_MONOTONIC</dt>
2152 <dd>
2153 <p>If defined to be <code>1</code>, libev will try to detect the availability of the
2154 monotonic clock option at both compiletime and runtime. Otherwise no use
2155 of the monotonic clock option will be attempted. If you enable this, you
2156 usually have to link against librt or something similar. Enabling it when
2157 the functionality isn't available is safe, though, althoguh you have
2158 to make sure you link against any libraries where the <code>clock_gettime</code>
2159 function is hiding in (often <cite>-lrt</cite>).</p>
2160 </dd>
2161 <dt>EV_USE_REALTIME</dt>
2162 <dd>
2163 <p>If defined to be <code>1</code>, libev will try to detect the availability of the
2164 realtime clock option at compiletime (and assume its availability at
2165 runtime if successful). Otherwise no use of the realtime clock option will
2166 be attempted. This effectively replaces <code>gettimeofday</code> by <code>clock_get
2167 (CLOCK_REALTIME, ...)</code> and will not normally affect correctness. See tzhe note about libraries
2168 in the description of <code>EV_USE_MONOTONIC</code>, though.</p>
2169 </dd>
2170 <dt>EV_USE_SELECT</dt>
2171 <dd>
2172 <p>If undefined or defined to be <code>1</code>, libev will compile in support for the
2173 <code>select</code>(2) backend. No attempt at autodetection will be done: if no
2174 other method takes over, select will be it. Otherwise the select backend
2175 will not be compiled in.</p>
2176 </dd>
2177 <dt>EV_SELECT_USE_FD_SET</dt>
2178 <dd>
2179 <p>If defined to <code>1</code>, then the select backend will use the system <code>fd_set</code>
2180 structure. This is useful if libev doesn't compile due to a missing
2181 <code>NFDBITS</code> or <code>fd_mask</code> definition or it misguesses the bitset layout on
2182 exotic systems. This usually limits the range of file descriptors to some
2183 low limit such as 1024 or might have other limitations (winsocket only
2184 allows 64 sockets). The <code>FD_SETSIZE</code> macro, set before compilation, might
2185 influence the size of the <code>fd_set</code> used.</p>
2186 </dd>
2187 <dt>EV_SELECT_IS_WINSOCKET</dt>
2188 <dd>
2189 <p>When defined to <code>1</code>, the select backend will assume that
2190 select/socket/connect etc. don't understand file descriptors but
2191 wants osf handles on win32 (this is the case when the select to
2192 be used is the winsock select). This means that it will call
2193 <code>_get_osfhandle</code> on the fd to convert it to an OS handle. Otherwise,
2194 it is assumed that all these functions actually work on fds, even
2195 on win32. Should not be defined on non-win32 platforms.</p>
2196 </dd>
2197 <dt>EV_USE_POLL</dt>
2198 <dd>
2199 <p>If defined to be <code>1</code>, libev will compile in support for the <code>poll</code>(2)
2200 backend. Otherwise it will be enabled on non-win32 platforms. It
2201 takes precedence over select.</p>
2202 </dd>
2203 <dt>EV_USE_EPOLL</dt>
2204 <dd>
2205 <p>If defined to be <code>1</code>, libev will compile in support for the Linux
2206 <code>epoll</code>(7) backend. Its availability will be detected at runtime,
2207 otherwise another method will be used as fallback. This is the
2208 preferred backend for GNU/Linux systems.</p>
2209 </dd>
2210 <dt>EV_USE_KQUEUE</dt>
2211 <dd>
2212 <p>If defined to be <code>1</code>, libev will compile in support for the BSD style
2213 <code>kqueue</code>(2) backend. Its actual availability will be detected at runtime,
2214 otherwise another method will be used as fallback. This is the preferred
2215 backend for BSD and BSD-like systems, although on most BSDs kqueue only
2216 supports some types of fds correctly (the only platform we found that
2217 supports ptys for example was NetBSD), so kqueue might be compiled in, but
2218 not be used unless explicitly requested. The best way to use it is to find
2219 out whether kqueue supports your type of fd properly and use an embedded
2220 kqueue loop.</p>
2221 </dd>
2222 <dt>EV_USE_PORT</dt>
2223 <dd>
2224 <p>If defined to be <code>1</code>, libev will compile in support for the Solaris
2225 10 port style backend. Its availability will be detected at runtime,
2226 otherwise another method will be used as fallback. This is the preferred
2227 backend for Solaris 10 systems.</p>
2228 </dd>
2229 <dt>EV_USE_DEVPOLL</dt>
2230 <dd>
2231 <p>reserved for future expansion, works like the USE symbols above.</p>
2232 </dd>
2233 <dt>EV_USE_INOTIFY</dt>
2234 <dd>
2235 <p>If defined to be <code>1</code>, libev will compile in support for the Linux inotify
2236 interface to speed up <code>ev_stat</code> watchers. Its actual availability will
2237 be detected at runtime.</p>
2238 </dd>
2239 <dt>EV_H</dt>
2240 <dd>
2241 <p>The name of the <cite>ev.h</cite> header file used to include it. The default if
2242 undefined is <code>&lt;ev.h&gt;</code> in <cite>event.h</cite> and <code>&quot;ev.h&quot;</code> in <cite>ev.c</cite>. This
2243 can be used to virtually rename the <cite>ev.h</cite> header file in case of conflicts.</p>
2244 </dd>
2245 <dt>EV_CONFIG_H</dt>
2246 <dd>
2247 <p>If <code>EV_STANDALONE</code> isn't <code>1</code>, this variable can be used to override
2248 <cite>ev.c</cite>'s idea of where to find the <cite>config.h</cite> file, similarly to
2249 <code>EV_H</code>, above.</p>
2250 </dd>
2251 <dt>EV_EVENT_H</dt>
2252 <dd>
2253 <p>Similarly to <code>EV_H</code>, this macro can be used to override <cite>event.c</cite>'s idea
2254 of how the <cite>event.h</cite> header can be found.</p>
2255 </dd>
2256 <dt>EV_PROTOTYPES</dt>
2257 <dd>
2258 <p>If defined to be <code>0</code>, then <cite>ev.h</cite> will not define any function
2259 prototypes, but still define all the structs and other symbols. This is
2260 occasionally useful if you want to provide your own wrapper functions
2261 around libev functions.</p>
2262 </dd>
2263 <dt>EV_MULTIPLICITY</dt>
2264 <dd>
2265 <p>If undefined or defined to <code>1</code>, then all event-loop-specific functions
2266 will have the <code>struct ev_loop *</code> as first argument, and you can create
2267 additional independent event loops. Otherwise there will be no support
2268 for multiple event loops and there is no first event loop pointer
2269 argument. Instead, all functions act on the single default loop.</p>
2270 </dd>
2271 <dt>EV_MINPRI</dt>
2272 <dt>EV_MAXPRI</dt>
2273 <dd>
2274 <p>The range of allowed priorities. <code>EV_MINPRI</code> must be smaller or equal to
2275 <code>EV_MAXPRI</code>, but otherwise there are no non-obvious limitations. You can
2276 provide for more priorities by overriding those symbols (usually defined
2277 to be <code>-2</code> and <code>2</code>, respectively).</p>
2278 <p>When doing priority-based operations, libev usually has to linearly search
2279 all the priorities, so having many of them (hundreds) uses a lot of space
2280 and time, so using the defaults of five priorities (-2 .. +2) is usually
2281 fine.</p>
2282 <p>If your embedding app does not need any priorities, defining these both to
2283 <code>0</code> will save some memory and cpu.</p>
2284 </dd>
2285 <dt>EV_PERIODIC_ENABLE</dt>
2286 <dd>
2287 <p>If undefined or defined to be <code>1</code>, then periodic timers are supported. If
2288 defined to be <code>0</code>, then they are not. Disabling them saves a few kB of
2289 code.</p>
2290 </dd>
2291 <dt>EV_IDLE_ENABLE</dt>
2292 <dd>
2293 <p>If undefined or defined to be <code>1</code>, then idle watchers are supported. If
2294 defined to be <code>0</code>, then they are not. Disabling them saves a few kB of
2295 code.</p>
2296 </dd>
2297 <dt>EV_EMBED_ENABLE</dt>
2298 <dd>
2299 <p>If undefined or defined to be <code>1</code>, then embed watchers are supported. If
2300 defined to be <code>0</code>, then they are not.</p>
2301 </dd>
2302 <dt>EV_STAT_ENABLE</dt>
2303 <dd>
2304 <p>If undefined or defined to be <code>1</code>, then stat watchers are supported. If
2305 defined to be <code>0</code>, then they are not.</p>
2306 </dd>
2307 <dt>EV_FORK_ENABLE</dt>
2308 <dd>
2309 <p>If undefined or defined to be <code>1</code>, then fork watchers are supported. If
2310 defined to be <code>0</code>, then they are not.</p>
2311 </dd>
2312 <dt>EV_MINIMAL</dt>
2313 <dd>
2314 <p>If you need to shave off some kilobytes of code at the expense of some
2315 speed, define this symbol to <code>1</code>. Currently only used for gcc to override
2316 some inlining decisions, saves roughly 30% codesize of amd64.</p>
2317 </dd>
2318 <dt>EV_PID_HASHSIZE</dt>
2319 <dd>
2320 <p><code>ev_child</code> watchers use a small hash table to distribute workload by
2321 pid. The default size is <code>16</code> (or <code>1</code> with <code>EV_MINIMAL</code>), usually more
2322 than enough. If you need to manage thousands of children you might want to
2323 increase this value (<i>must</i> be a power of two).</p>
2324 </dd>
2325 <dt>EV_INOTIFY_HASHSIZE</dt>
2326 <dd>
2327 <p><code>ev_staz</code> watchers use a small hash table to distribute workload by
2328 inotify watch id. The default size is <code>16</code> (or <code>1</code> with <code>EV_MINIMAL</code>),
2329 usually more than enough. If you need to manage thousands of <code>ev_stat</code>
2330 watchers you might want to increase this value (<i>must</i> be a power of
2331 two).</p>
2332 </dd>
2333 <dt>EV_COMMON</dt>
2334 <dd>
2335 <p>By default, all watchers have a <code>void *data</code> member. By redefining
2336 this macro to a something else you can include more and other types of
2337 members. You have to define it each time you include one of the files,
2338 though, and it must be identical each time.</p>
2339 <p>For example, the perl EV module uses something like this:</p>
2340 <pre> #define EV_COMMON \
2341 SV *self; /* contains this struct */ \
2342 SV *cb_sv, *fh /* note no trailing &quot;;&quot; */
2343
2344 </pre>
2345 </dd>
2346 <dt>EV_CB_DECLARE (type)</dt>
2347 <dt>EV_CB_INVOKE (watcher, revents)</dt>
2348 <dt>ev_set_cb (ev, cb)</dt>
2349 <dd>
2350 <p>Can be used to change the callback member declaration in each watcher,
2351 and the way callbacks are invoked and set. Must expand to a struct member
2352 definition and a statement, respectively. See the <cite>ev.v</cite> header file for
2353 their default definitions. One possible use for overriding these is to
2354 avoid the <code>struct ev_loop *</code> as first argument in all cases, or to use
2355 method calls instead of plain function calls in C++.</p>
2356
2357 </div>
2358 <h2 id="EXAMPLES">EXAMPLES</h2>
2359 <div id="EXAMPLES_CONTENT">
2360 <p>For a real-world example of a program the includes libev
2361 verbatim, you can have a look at the EV perl module
2362 (<a href="http://software.schmorp.de/pkg/EV.html">http://software.schmorp.de/pkg/EV.html</a>). It has the libev files in
2363 the <cite>libev/</cite> subdirectory and includes them in the <cite>EV/EVAPI.h</cite> (public
2364 interface) and <cite>EV.xs</cite> (implementation) files. Only the <cite>EV.xs</cite> file
2365 will be compiled. It is pretty complex because it provides its own header
2366 file.</p>
2367 <p>The usage in rxvt-unicode is simpler. It has a <cite>ev_cpp.h</cite> header file
2368 that everybody includes and which overrides some configure choices:</p>
2369 <pre> #define EV_MINIMAL 1
2370 #define EV_USE_POLL 0
2371 #define EV_MULTIPLICITY 0
2372 #define EV_PERIODIC_ENABLE 0
2373 #define EV_STAT_ENABLE 0
2374 #define EV_FORK_ENABLE 0
2375 #define EV_CONFIG_H &lt;config.h&gt;
2376 #define EV_MINPRI 0
2377 #define EV_MAXPRI 0
2378
2379 #include &quot;ev++.h&quot;
2380
2381 </pre>
2382 <p>And a <cite>ev_cpp.C</cite> implementation file that contains libev proper and is compiled:</p>
2383 <pre> #include &quot;ev_cpp.h&quot;
2384 #include &quot;ev.c&quot;
2385
2386
2387
2388
2389 </pre>
2390
2391 </div>
2392 <h1 id="COMPLEXITIES">COMPLEXITIES</h1>
2393 <div id="COMPLEXITIES_CONTENT">
2394 <p>In this section the complexities of (many of) the algorithms used inside
2395 libev will be explained. For complexity discussions about backends see the
2396 documentation for <code>ev_default_init</code>.</p>
2397 <p>All of the following are about amortised time: If an array needs to be
2398 extended, libev needs to realloc and move the whole array, but this
2399 happens asymptotically never with higher number of elements, so O(1) might
2400 mean it might do a lengthy realloc operation in rare cases, but on average
2401 it is much faster and asymptotically approaches constant time.</p>
2402 <p>
2403 <dl>
2404 <dt>Starting and stopping timer/periodic watchers: O(log skipped_other_timers)</dt>
2405 <dd>
2406 <p>This means that, when you have a watcher that triggers in one hour and
2407 there are 100 watchers that would trigger before that then inserting will
2408 have to skip those 100 watchers.</p>
2409 </dd>
2410 <dt>Changing timer/periodic watchers (by autorepeat, again): O(log skipped_other_timers)</dt>
2411 <dd>
2412 <p>That means that for changing a timer costs less than removing/adding them
2413 as only the relative motion in the event queue has to be paid for.</p>
2414 </dd>
2415 <dt>Starting io/check/prepare/idle/signal/child watchers: O(1)</dt>
2416 <dd>
2417 <p>These just add the watcher into an array or at the head of a list.
2418 =item Stopping check/prepare/idle watchers: O(1)</p>
2419 </dd>
2420 <dt>Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))</dt>
2421 <dd>
2422 <p>These watchers are stored in lists then need to be walked to find the
2423 correct watcher to remove. The lists are usually short (you don't usually
2424 have many watchers waiting for the same fd or signal).</p>
2425 </dd>
2426 <dt>Finding the next timer per loop iteration: O(1)</dt>
2427 <dt>Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)</dt>
2428 <dd>
2429 <p>A change means an I/O watcher gets started or stopped, which requires
2430 libev to recalculate its status (and possibly tell the kernel).</p>
2431 </dd>
2432 <dt>Activating one watcher: O(1)</dt>
2433 <dt>Priority handling: O(number_of_priorities)</dt>
2434 <dd>
2435 <p>Priorities are implemented by allocating some space for each
2436 priority. When doing priority-based operations, libev usually has to
2437 linearly search all the priorities.</p>
2438 </dd>
2439 </dl>
2440 </p>
2441
2442
2443
2444
2445
2446 </div>
2447 <h1 id="AUTHOR">AUTHOR</h1>
2448 <div id="AUTHOR_CONTENT">
2449 <p>Marc Lehmann &lt;libev@schmorp.de&gt;.</p>
2450
2451 </div>
2452 </div></body>
2453 </html>