ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/libev/ev.html
(Generate patch)

Comparing libev/ev.html (file contents):
Revision 1.13 by root, Mon Nov 12 08:35:36 2007 UTC vs.
Revision 1.49 by root, Tue Nov 27 08:20:42 2007 UTC

4<head> 4<head>
5 <title>libev</title> 5 <title>libev</title>
6 <meta name="description" content="Pod documentation for libev" /> 6 <meta name="description" content="Pod documentation for libev" />
7 <meta name="inputfile" content="&lt;standard input&gt;" /> 7 <meta name="inputfile" content="&lt;standard input&gt;" />
8 <meta name="outputfile" content="&lt;standard output&gt;" /> 8 <meta name="outputfile" content="&lt;standard output&gt;" />
9 <meta name="created" content="Mon Nov 12 09:35:36 2007" /> 9 <meta name="created" content="Tue Nov 27 09:20:40 2007" />
10 <meta name="generator" content="Pod::Xhtml 1.57" /> 10 <meta name="generator" content="Pod::Xhtml 1.57" />
11<link rel="stylesheet" href="http://res.tst.eu/pod.css"/></head> 11<link rel="stylesheet" href="http://res.tst.eu/pod.css"/></head>
12<body> 12<body>
13<div class="pod"> 13<div class="pod">
14<!-- INDEX START --> 14<!-- INDEX START -->
17<ul><li><a href="#NAME">NAME</a></li> 17<ul><li><a href="#NAME">NAME</a></li>
18<li><a href="#SYNOPSIS">SYNOPSIS</a></li> 18<li><a href="#SYNOPSIS">SYNOPSIS</a></li>
19<li><a href="#DESCRIPTION">DESCRIPTION</a></li> 19<li><a href="#DESCRIPTION">DESCRIPTION</a></li>
20<li><a href="#FEATURES">FEATURES</a></li> 20<li><a href="#FEATURES">FEATURES</a></li>
21<li><a href="#CONVENTIONS">CONVENTIONS</a></li> 21<li><a href="#CONVENTIONS">CONVENTIONS</a></li>
22<li><a href="#TIME_REPRESENTATION">TIME REPRESENTATION</a></li>
22<li><a href="#TIME_AND_OTHER_GLOBAL_FUNCTIONS">TIME AND OTHER GLOBAL FUNCTIONS</a></li> 23<li><a href="#GLOBAL_FUNCTIONS">GLOBAL FUNCTIONS</a></li>
23<li><a href="#FUNCTIONS_CONTROLLING_THE_EVENT_LOOP">FUNCTIONS CONTROLLING THE EVENT LOOP</a></li> 24<li><a href="#FUNCTIONS_CONTROLLING_THE_EVENT_LOOP">FUNCTIONS CONTROLLING THE EVENT LOOP</a></li>
24<li><a href="#ANATOMY_OF_A_WATCHER">ANATOMY OF A WATCHER</a> 25<li><a href="#ANATOMY_OF_A_WATCHER">ANATOMY OF A WATCHER</a>
26<ul><li><a href="#GENERIC_WATCHER_FUNCTIONS">GENERIC WATCHER FUNCTIONS</a></li>
25<ul><li><a href="#ASSOCIATING_CUSTOM_DATA_WITH_A_WATCH">ASSOCIATING CUSTOM DATA WITH A WATCHER</a></li> 27<li><a href="#ASSOCIATING_CUSTOM_DATA_WITH_A_WATCH">ASSOCIATING CUSTOM DATA WITH A WATCHER</a></li>
26</ul> 28</ul>
27</li> 29</li>
28<li><a href="#WATCHER_TYPES">WATCHER TYPES</a> 30<li><a href="#WATCHER_TYPES">WATCHER TYPES</a>
29<ul><li><a href="#code_ev_io_code_is_this_file_descrip"><code>ev_io</code> - is this file descriptor readable or writable</a></li> 31<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>
30<li><a href="#code_ev_timer_code_relative_and_opti"><code>ev_timer</code> - relative and optionally recurring timeouts</a></li> 32<li><a href="#code_ev_timer_code_relative_and_opti"><code>ev_timer</code> - relative and optionally repeating timeouts</a></li>
31<li><a href="#code_ev_periodic_code_to_cron_or_not"><code>ev_periodic</code> - to cron or not to cron it</a></li> 33<li><a href="#code_ev_periodic_code_to_cron_or_not"><code>ev_periodic</code> - to cron or not to cron?</a></li>
32<li><a href="#code_ev_signal_code_signal_me_when_a"><code>ev_signal</code> - signal me when a signal gets signalled</a></li> 34<li><a href="#code_ev_signal_code_signal_me_when_a"><code>ev_signal</code> - signal me when a signal gets signalled!</a></li>
33<li><a href="#code_ev_child_code_wait_for_pid_stat"><code>ev_child</code> - wait for pid status changes</a></li> 35<li><a href="#code_ev_child_code_watch_out_for_pro"><code>ev_child</code> - watch out for process status changes</a></li>
36<li><a href="#code_ev_stat_code_did_the_file_attri"><code>ev_stat</code> - did the file attributes just change?</a></li>
34<li><a href="#code_ev_idle_code_when_you_ve_got_no"><code>ev_idle</code> - when you've got nothing better to do</a></li> 37<li><a href="#code_ev_idle_code_when_you_ve_got_no"><code>ev_idle</code> - when you've got nothing better to do...</a></li>
35<li><a href="#prepare_and_check_your_hooks_into_th">prepare and check - your hooks into the event loop</a></li> 38<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>
39<li><a href="#code_ev_embed_code_when_one_backend_"><code>ev_embed</code> - when one backend isn't enough...</a></li>
36</ul> 40</ul>
37</li> 41</li>
38<li><a href="#OTHER_FUNCTIONS">OTHER FUNCTIONS</a></li> 42<li><a href="#OTHER_FUNCTIONS">OTHER FUNCTIONS</a></li>
43<li><a href="#LIBEVENT_EMULATION">LIBEVENT EMULATION</a></li>
44<li><a href="#C_SUPPORT">C++ SUPPORT</a></li>
45<li><a href="#EMBEDDING">EMBEDDING</a>
46<ul><li><a href="#FILESETS">FILESETS</a>
47<ul><li><a href="#CORE_EVENT_LOOP">CORE EVENT LOOP</a></li>
48<li><a href="#LIBEVENT_COMPATIBILITY_API">LIBEVENT COMPATIBILITY API</a></li>
49<li><a href="#AUTOCONF_SUPPORT">AUTOCONF SUPPORT</a></li>
50</ul>
51</li>
52<li><a href="#PREPROCESSOR_SYMBOLS_MACROS">PREPROCESSOR SYMBOLS/MACROS</a></li>
53<li><a href="#EXAMPLES">EXAMPLES</a></li>
54</ul>
55</li>
56<li><a href="#COMPLEXITIES">COMPLEXITIES</a></li>
39<li><a href="#AUTHOR">AUTHOR</a> 57<li><a href="#AUTHOR">AUTHOR</a>
40</li> 58</li>
41</ul><hr /> 59</ul><hr />
42<!-- INDEX END --> 60<!-- INDEX END -->
43 61
87support for multiple event loops, then all functions taking an initial 105support for multiple event loops, then all functions taking an initial
88argument of name <code>loop</code> (which is always of type <code>struct ev_loop *</code>) 106argument of name <code>loop</code> (which is always of type <code>struct ev_loop *</code>)
89will not have this argument.</p> 107will not have this argument.</p>
90 108
91</div> 109</div>
92<h1 id="TIME_AND_OTHER_GLOBAL_FUNCTIONS">TIME AND OTHER GLOBAL FUNCTIONS</h1><p><a href="#TOP" class="toplink">Top</a></p> 110<h1 id="TIME_REPRESENTATION">TIME REPRESENTATION</h1><p><a href="#TOP" class="toplink">Top</a></p>
93<div id="TIME_AND_OTHER_GLOBAL_FUNCTIONS_CONT"> 111<div id="TIME_REPRESENTATION_CONTENT">
94<p>Libev represents time as a single floating point number, representing the 112<p>Libev represents time as a single floating point number, representing the
95(fractional) number of seconds since the (POSIX) epoch (somewhere near 113(fractional) number of seconds since the (POSIX) epoch (somewhere near
96the beginning of 1970, details are complicated, don't ask). This type is 114the beginning of 1970, details are complicated, don't ask). This type is
97called <code>ev_tstamp</code>, which is what you should use too. It usually aliases 115called <code>ev_tstamp</code>, which is what you should use too. It usually aliases
98to the double type in C.</p> 116to the <code>double</code> type in C, and when you need to do any calculations on
117it, you should treat it as such.</p>
118
119
120
121
122
123</div>
124<h1 id="GLOBAL_FUNCTIONS">GLOBAL FUNCTIONS</h1><p><a href="#TOP" class="toplink">Top</a></p>
125<div id="GLOBAL_FUNCTIONS_CONTENT">
126<p>These functions can be called anytime, even before initialising the
127library in any way.</p>
99<dl> 128<dl>
100 <dt>ev_tstamp ev_time ()</dt> 129 <dt>ev_tstamp ev_time ()</dt>
101 <dd> 130 <dd>
102 <p>Returns the current time as libev would use it.</p> 131 <p>Returns the current time as libev would use it. Please note that the
132<code>ev_now</code> function is usually faster and also often returns the timestamp
133you actually want to know.</p>
103 </dd> 134 </dd>
104 <dt>int ev_version_major ()</dt> 135 <dt>int ev_version_major ()</dt>
105 <dt>int ev_version_minor ()</dt> 136 <dt>int ev_version_minor ()</dt>
106 <dd> 137 <dd>
107 <p>You can find out the major and minor version numbers of the library 138 <p>You can find out the major and minor version numbers of the library
111version of the library your program was compiled against.</p> 142version of the library your program was compiled against.</p>
112 <p>Usually, it's a good idea to terminate if the major versions mismatch, 143 <p>Usually, it's a good idea to terminate if the major versions mismatch,
113as this indicates an incompatible change. Minor versions are usually 144as this indicates an incompatible change. Minor versions are usually
114compatible to older versions, so a larger minor version alone is usually 145compatible to older versions, so a larger minor version alone is usually
115not a problem.</p> 146not a problem.</p>
147 <p>Example: make sure we haven't accidentally been linked against the wrong
148version:</p>
149<pre> assert ((&quot;libev version mismatch&quot;,
150 ev_version_major () == EV_VERSION_MAJOR
151 &amp;&amp; ev_version_minor () &gt;= EV_VERSION_MINOR));
152
153</pre>
154 </dd>
155 <dt>unsigned int ev_supported_backends ()</dt>
156 <dd>
157 <p>Return the set of all backends (i.e. their corresponding <code>EV_BACKEND_*</code>
158value) compiled into this binary of libev (independent of their
159availability on the system you are running on). See <code>ev_default_loop</code> for
160a description of the set values.</p>
161 <p>Example: make sure we have the epoll method, because yeah this is cool and
162a must have and can we have a torrent of it please!!!11</p>
163<pre> assert ((&quot;sorry, no epoll, no sex&quot;,
164 ev_supported_backends () &amp; EVBACKEND_EPOLL));
165
166</pre>
167 </dd>
168 <dt>unsigned int ev_recommended_backends ()</dt>
169 <dd>
170 <p>Return the set of all backends compiled into this binary of libev and also
171recommended for this platform. This set is often smaller than the one
172returned by <code>ev_supported_backends</code>, as for example kqueue is broken on
173most BSDs and will not be autodetected unless you explicitly request it
174(assuming you know what you are doing). This is the set of backends that
175libev will probe for if you specify no backends explicitly.</p>
176 </dd>
177 <dt>unsigned int ev_embeddable_backends ()</dt>
178 <dd>
179 <p>Returns the set of backends that are embeddable in other event loops. This
180is the theoretical, all-platform, value. To find which backends
181might be supported on the current system, you would need to look at
182<code>ev_embeddable_backends () &amp; ev_supported_backends ()</code>, likewise for
183recommended ones.</p>
184 <p>See the description of <code>ev_embed</code> watchers for more info.</p>
116 </dd> 185 </dd>
117 <dt>ev_set_allocator (void *(*cb)(void *ptr, long size))</dt> 186 <dt>ev_set_allocator (void *(*cb)(void *ptr, long size))</dt>
118 <dd> 187 <dd>
119 <p>Sets the allocation function to use (the prototype is similar to the 188 <p>Sets the allocation function to use (the prototype is similar to the
120realloc C function, the semantics are identical). It is used to allocate 189realloc C function, the semantics are identical). It is used to allocate
122needs to be allocated, the library might abort or take some potentially 191needs to be allocated, the library might abort or take some potentially
123destructive action. The default is your system realloc function.</p> 192destructive action. The default is your system realloc function.</p>
124 <p>You could override this function in high-availability programs to, say, 193 <p>You could override this function in high-availability programs to, say,
125free some memory if it cannot allocate memory, to use a special allocator, 194free some memory if it cannot allocate memory, to use a special allocator,
126or even to sleep a while and retry until some memory is available.</p> 195or even to sleep a while and retry until some memory is available.</p>
196 <p>Example: replace the libev allocator with one that waits a bit and then
197retries: better than mine).</p>
198<pre> static void *
199 persistent_realloc (void *ptr, long size)
200 {
201 for (;;)
202 {
203 void *newptr = realloc (ptr, size);
204
205 if (newptr)
206 return newptr;
207
208 sleep (60);
209 }
210 }
211
212 ...
213 ev_set_allocator (persistent_realloc);
214
215</pre>
127 </dd> 216 </dd>
128 <dt>ev_set_syserr_cb (void (*cb)(const char *msg));</dt> 217 <dt>ev_set_syserr_cb (void (*cb)(const char *msg));</dt>
129 <dd> 218 <dd>
130 <p>Set the callback function to call on a retryable syscall error (such 219 <p>Set the callback function to call on a retryable syscall error (such
131as failed select, poll, epoll_wait). The message is a printable string 220as failed select, poll, epoll_wait). The message is a printable string
132indicating the system call or subsystem causing the problem. If this 221indicating the system call or subsystem causing the problem. If this
133callback is set, then libev will expect it to remedy the sitution, no 222callback is set, then libev will expect it to remedy the sitution, no
134matter what, when it returns. That is, libev will generally retry the 223matter what, when it returns. That is, libev will generally retry the
135requested operation, or, if the condition doesn't go away, do bad stuff 224requested operation, or, if the condition doesn't go away, do bad stuff
136(such as abort).</p> 225(such as abort).</p>
226 <p>Example: do the same thing as libev does internally:</p>
227<pre> static void
228 fatal_error (const char *msg)
229 {
230 perror (msg);
231 abort ();
232 }
233
234 ...
235 ev_set_syserr_cb (fatal_error);
236
237</pre>
137 </dd> 238 </dd>
138</dl> 239</dl>
139 240
140</div> 241</div>
141<h1 id="FUNCTIONS_CONTROLLING_THE_EVENT_LOOP">FUNCTIONS CONTROLLING THE EVENT LOOP</h1><p><a href="#TOP" class="toplink">Top</a></p> 242<h1 id="FUNCTIONS_CONTROLLING_THE_EVENT_LOOP">FUNCTIONS CONTROLLING THE EVENT LOOP</h1><p><a href="#TOP" class="toplink">Top</a></p>
142<div id="FUNCTIONS_CONTROLLING_THE_EVENT_LOOP-2"> 243<div id="FUNCTIONS_CONTROLLING_THE_EVENT_LOOP-2">
143<p>An event loop is described by a <code>struct ev_loop *</code>. The library knows two 244<p>An event loop is described by a <code>struct ev_loop *</code>. The library knows two
144types of such loops, the <i>default</i> loop, which supports signals and child 245types of such loops, the <i>default</i> loop, which supports signals and child
145events, and dynamically created loops which do not.</p> 246events, and dynamically created loops which do not.</p>
146<p>If you use threads, a common model is to run the default event loop 247<p>If you use threads, a common model is to run the default event loop
147in your main thread (or in a separate thrad) and for each thread you 248in your main thread (or in a separate thread) and for each thread you
148create, you also create another event loop. Libev itself does no locking 249create, you also create another event loop. Libev itself does no locking
149whatsoever, so if you mix calls to the same event loop in different 250whatsoever, so if you mix calls to the same event loop in different
150threads, make sure you lock (this is usually a bad idea, though, even if 251threads, make sure you lock (this is usually a bad idea, though, even if
151done correctly, because it's hideous and inefficient).</p> 252done correctly, because it's hideous and inefficient).</p>
152<dl> 253<dl>
153 <dt>struct ev_loop *ev_default_loop (unsigned int flags)</dt> 254 <dt>struct ev_loop *ev_default_loop (unsigned int flags)</dt>
154 <dd> 255 <dd>
155 <p>This will initialise the default event loop if it hasn't been initialised 256 <p>This will initialise the default event loop if it hasn't been initialised
156yet and return it. If the default loop could not be initialised, returns 257yet and return it. If the default loop could not be initialised, returns
157false. If it already was initialised it simply returns it (and ignores the 258false. If it already was initialised it simply returns it (and ignores the
158flags).</p> 259flags. If that is troubling you, check <code>ev_backend ()</code> afterwards).</p>
159 <p>If you don't know what event loop to use, use the one returned from this 260 <p>If you don't know what event loop to use, use the one returned from this
160function.</p> 261function.</p>
161 <p>The flags argument can be used to specify special behaviour or specific 262 <p>The flags argument can be used to specify special behaviour or specific
162backends to use, and is usually specified as 0 (or EVFLAG_AUTO).</p> 263backends to use, and is usually specified as <code>0</code> (or <code>EVFLAG_AUTO</code>).</p>
163 <p>It supports the following flags:</p> 264 <p>The following flags are supported:</p>
164 <p> 265 <p>
165 <dl> 266 <dl>
166 <dt><code>EVFLAG_AUTO</code></dt> 267 <dt><code>EVFLAG_AUTO</code></dt>
167 <dd> 268 <dd>
168 <p>The default flags value. Use this if you have no clue (it's the right 269 <p>The default flags value. Use this if you have no clue (it's the right
175<code>LIBEV_FLAGS</code>. Otherwise (the default), this environment variable will 276<code>LIBEV_FLAGS</code>. Otherwise (the default), this environment variable will
176override the flags completely if it is found in the environment. This is 277override the flags completely if it is found in the environment. This is
177useful to try out specific backends to test their performance, or to work 278useful to try out specific backends to test their performance, or to work
178around bugs.</p> 279around bugs.</p>
179 </dd> 280 </dd>
180 <dt><code>EVMETHOD_SELECT</code> (portable select backend)</dt> 281 <dt><code>EVBACKEND_SELECT</code> (value 1, portable select backend)</dt>
181 <dt><code>EVMETHOD_POLL</code> (poll backend, available everywhere except on windows)</dt>
182 <dt><code>EVMETHOD_EPOLL</code> (linux only)</dt>
183 <dt><code>EVMETHOD_KQUEUE</code> (some bsds only)</dt>
184 <dt><code>EVMETHOD_DEVPOLL</code> (solaris 8 only)</dt>
185 <dt><code>EVMETHOD_PORT</code> (solaris 10 only)</dt>
186 <dd> 282 <dd>
187 <p>If one or more of these are ored into the flags value, then only these 283 <p>This is your standard select(2) backend. Not <i>completely</i> standard, as
188backends will be tried (in the reverse order as given here). If one are 284libev tries to roll its own fd_set with no limits on the number of fds,
189specified, any backend will do.</p> 285but if that fails, expect a fairly low limit on the number of fds when
286using this backend. It doesn't scale too well (O(highest_fd)), but its usually
287the fastest backend for a low number of fds.</p>
288 </dd>
289 <dt><code>EVBACKEND_POLL</code> (value 2, poll backend, available everywhere except on windows)</dt>
290 <dd>
291 <p>And this is your standard poll(2) backend. It's more complicated than
292select, but handles sparse fds better and has no artificial limit on the
293number of fds you can use (except it will slow down considerably with a
294lot of inactive fds). It scales similarly to select, i.e. O(total_fds).</p>
295 </dd>
296 <dt><code>EVBACKEND_EPOLL</code> (value 4, Linux)</dt>
297 <dd>
298 <p>For few fds, this backend is a bit little slower than poll and select,
299but it scales phenomenally better. While poll and select usually scale like
300O(total_fds) where n is the total number of fds (or the highest fd), epoll scales
301either O(1) or O(active_fds).</p>
302 <p>While stopping and starting an I/O watcher in the same iteration will
303result in some caching, there is still a syscall per such incident
304(because the fd could point to a different file description now), so its
305best to avoid that. Also, dup()ed file descriptors might not work very
306well if you register events for both fds.</p>
307 <p>Please note that epoll sometimes generates spurious notifications, so you
308need to use non-blocking I/O or other means to avoid blocking when no data
309(or space) is available.</p>
310 </dd>
311 <dt><code>EVBACKEND_KQUEUE</code> (value 8, most BSD clones)</dt>
312 <dd>
313 <p>Kqueue deserves special mention, as at the time of this writing, it
314was broken on all BSDs except NetBSD (usually it doesn't work with
315anything but sockets and pipes, except on Darwin, where of course its
316completely useless). For this reason its not being &quot;autodetected&quot;
317unless you explicitly specify it explicitly in the flags (i.e. using
318<code>EVBACKEND_KQUEUE</code>).</p>
319 <p>It scales in the same way as the epoll backend, but the interface to the
320kernel is more efficient (which says nothing about its actual speed, of
321course). While starting and stopping an I/O watcher does not cause an
322extra syscall as with epoll, it still adds up to four event changes per
323incident, so its best to avoid that.</p>
324 </dd>
325 <dt><code>EVBACKEND_DEVPOLL</code> (value 16, Solaris 8)</dt>
326 <dd>
327 <p>This is not implemented yet (and might never be).</p>
328 </dd>
329 <dt><code>EVBACKEND_PORT</code> (value 32, Solaris 10)</dt>
330 <dd>
331 <p>This uses the Solaris 10 port mechanism. As with everything on Solaris,
332it's really slow, but it still scales very well (O(active_fds)).</p>
333 <p>Please note that solaris ports can result in a lot of spurious
334notifications, so you need to use non-blocking I/O or other means to avoid
335blocking when no data (or space) is available.</p>
336 </dd>
337 <dt><code>EVBACKEND_ALL</code></dt>
338 <dd>
339 <p>Try all backends (even potentially broken ones that wouldn't be tried
340with <code>EVFLAG_AUTO</code>). Since this is a mask, you can do stuff such as
341<code>EVBACKEND_ALL &amp; ~EVBACKEND_KQUEUE</code>.</p>
190 </dd> 342 </dd>
191 </dl> 343 </dl>
192 </p> 344 </p>
345 <p>If one or more of these are ored into the flags value, then only these
346backends will be tried (in the reverse order as given here). If none are
347specified, most compiled-in backend will be tried, usually in reverse
348order of their flag values :)</p>
349 <p>The most typical usage is like this:</p>
350<pre> if (!ev_default_loop (0))
351 fatal (&quot;could not initialise libev, bad $LIBEV_FLAGS in environment?&quot;);
352
353</pre>
354 <p>Restrict libev to the select and poll backends, and do not allow
355environment settings to be taken into account:</p>
356<pre> ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);
357
358</pre>
359 <p>Use whatever libev has to offer, but make sure that kqueue is used if
360available (warning, breaks stuff, best use only with your own private
361event loop and only if you know the OS supports your types of fds):</p>
362<pre> ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);
363
364</pre>
193 </dd> 365 </dd>
194 <dt>struct ev_loop *ev_loop_new (unsigned int flags)</dt> 366 <dt>struct ev_loop *ev_loop_new (unsigned int flags)</dt>
195 <dd> 367 <dd>
196 <p>Similar to <code>ev_default_loop</code>, but always creates a new event loop that is 368 <p>Similar to <code>ev_default_loop</code>, but always creates a new event loop that is
197always distinct from the default loop. Unlike the default loop, it cannot 369always distinct from the default loop. Unlike the default loop, it cannot
198handle signal and child watchers, and attempts to do so will be greeted by 370handle signal and child watchers, and attempts to do so will be greeted by
199undefined behaviour (or a failed assertion if assertions are enabled).</p> 371undefined behaviour (or a failed assertion if assertions are enabled).</p>
372 <p>Example: try to create a event loop that uses epoll and nothing else.</p>
373<pre> struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);
374 if (!epoller)
375 fatal (&quot;no epoll found here, maybe it hides under your chair&quot;);
376
377</pre>
200 </dd> 378 </dd>
201 <dt>ev_default_destroy ()</dt> 379 <dt>ev_default_destroy ()</dt>
202 <dd> 380 <dd>
203 <p>Destroys the default loop again (frees all memory and kernel state 381 <p>Destroys the default loop again (frees all memory and kernel state
204etc.). This stops all registered event watchers (by not touching them in 382etc.). None of the active event watchers will be stopped in the normal
205any way whatsoever, although you cannot rely on this :).</p> 383sense, so e.g. <code>ev_is_active</code> might still return true. It is your
384responsibility to either stop all watchers cleanly yoursef <i>before</i>
385calling this function, or cope with the fact afterwards (which is usually
386the easiest thing, youc na just ignore the watchers and/or <code>free ()</code> them
387for example).</p>
206 </dd> 388 </dd>
207 <dt>ev_loop_destroy (loop)</dt> 389 <dt>ev_loop_destroy (loop)</dt>
208 <dd> 390 <dd>
209 <p>Like <code>ev_default_destroy</code>, but destroys an event loop created by an 391 <p>Like <code>ev_default_destroy</code>, but destroys an event loop created by an
210earlier call to <code>ev_loop_new</code>.</p> 392earlier call to <code>ev_loop_new</code>.</p>
213 <dd> 395 <dd>
214 <p>This function reinitialises the kernel state for backends that have 396 <p>This function reinitialises the kernel state for backends that have
215one. Despite the name, you can call it anytime, but it makes most sense 397one. Despite the name, you can call it anytime, but it makes most sense
216after forking, in either the parent or child process (or both, but that 398after forking, in either the parent or child process (or both, but that
217again makes little sense).</p> 399again makes little sense).</p>
218 <p>You <i>must</i> call this function after forking if and only if you want to 400 <p>You <i>must</i> call this function in the child process after forking if and
219use the event library in both processes. If you just fork+exec, you don't 401only if you want to use the event library in both processes. If you just
220have to call it.</p> 402fork+exec, you don't have to call it.</p>
221 <p>The function itself is quite fast and it's usually not a problem to call 403 <p>The function itself is quite fast and it's usually not a problem to call
222it just in case after a fork. To make this easy, the function will fit in 404it just in case after a fork. To make this easy, the function will fit in
223quite nicely into a call to <code>pthread_atfork</code>:</p> 405quite nicely into a call to <code>pthread_atfork</code>:</p>
224<pre> pthread_atfork (0, 0, ev_default_fork); 406<pre> pthread_atfork (0, 0, ev_default_fork);
225 407
226</pre> 408</pre>
409 <p>At the moment, <code>EVBACKEND_SELECT</code> and <code>EVBACKEND_POLL</code> are safe to use
410without calling this function, so if you force one of those backends you
411do not need to care.</p>
227 </dd> 412 </dd>
228 <dt>ev_loop_fork (loop)</dt> 413 <dt>ev_loop_fork (loop)</dt>
229 <dd> 414 <dd>
230 <p>Like <code>ev_default_fork</code>, but acts on an event loop created by 415 <p>Like <code>ev_default_fork</code>, but acts on an event loop created by
231<code>ev_loop_new</code>. Yes, you have to call this on every allocated event loop 416<code>ev_loop_new</code>. Yes, you have to call this on every allocated event loop
232after fork, and how you do this is entirely your own problem.</p> 417after fork, and how you do this is entirely your own problem.</p>
233 </dd> 418 </dd>
234 <dt>unsigned int ev_method (loop)</dt> 419 <dt>unsigned int ev_backend (loop)</dt>
235 <dd> 420 <dd>
236 <p>Returns one of the <code>EVMETHOD_*</code> flags indicating the event backend in 421 <p>Returns one of the <code>EVBACKEND_*</code> flags indicating the event backend in
237use.</p> 422use.</p>
238 </dd> 423 </dd>
239 <dt>ev_tstamp ev_now (loop)</dt> 424 <dt>ev_tstamp ev_now (loop)</dt>
240 <dd> 425 <dd>
241 <p>Returns the current &quot;event loop time&quot;, which is the time the event loop 426 <p>Returns the current &quot;event loop time&quot;, which is the time the event loop
242got events and started processing them. This timestamp does not change 427received events and started processing them. This timestamp does not
243as long as callbacks are being processed, and this is also the base time 428change as long as callbacks are being processed, and this is also the base
244used for relative timers. You can treat it as the timestamp of the event 429time used for relative timers. You can treat it as the timestamp of the
245occuring (or more correctly, the mainloop finding out about it).</p> 430event occuring (or more correctly, libev finding out about it).</p>
246 </dd> 431 </dd>
247 <dt>ev_loop (loop, int flags)</dt> 432 <dt>ev_loop (loop, int flags)</dt>
248 <dd> 433 <dd>
249 <p>Finally, this is it, the event handler. This function usually is called 434 <p>Finally, this is it, the event handler. This function usually is called
250after you initialised all your watchers and you want to start handling 435after you initialised all your watchers and you want to start handling
251events.</p> 436events.</p>
252 <p>If the flags argument is specified as 0, it will not return until either 437 <p>If the flags argument is specified as <code>0</code>, it will not return until
253no event watchers are active anymore or <code>ev_unloop</code> was called.</p> 438either no event watchers are active anymore or <code>ev_unloop</code> was called.</p>
439 <p>Please note that an explicit <code>ev_unloop</code> is usually better than
440relying on all watchers to be stopped when deciding when a program has
441finished (especially in interactive programs), but having a program that
442automatically loops as long as it has to and no longer by virtue of
443relying on its watchers stopping correctly is a thing of beauty.</p>
254 <p>A flags value of <code>EVLOOP_NONBLOCK</code> will look for new events, will handle 444 <p>A flags value of <code>EVLOOP_NONBLOCK</code> will look for new events, will handle
255those events and any outstanding ones, but will not block your process in 445those events and any outstanding ones, but will not block your process in
256case there are no events and will return after one iteration of the loop.</p> 446case there are no events and will return after one iteration of the loop.</p>
257 <p>A flags value of <code>EVLOOP_ONESHOT</code> will look for new events (waiting if 447 <p>A flags value of <code>EVLOOP_ONESHOT</code> will look for new events (waiting if
258neccessary) and will handle those and any outstanding ones. It will block 448neccessary) and will handle those and any outstanding ones. It will block
259your process until at least one new event arrives, and will return after 449your process until at least one new event arrives, and will return after
260one iteration of the loop.</p> 450one iteration of the loop. This is useful if you are waiting for some
261 <p>This flags value could be used to implement alternative looping 451external event in conjunction with something not expressible using other
262constructs, but the <code>prepare</code> and <code>check</code> watchers provide a better and 452libev watchers. However, a pair of <code>ev_prepare</code>/<code>ev_check</code> watchers is
263more generic mechanism.</p> 453usually a better approach for this kind of thing.</p>
454 <p>Here are the gory details of what <code>ev_loop</code> does:</p>
455<pre> * If there are no active watchers (reference count is zero), return.
456 - Queue prepare watchers and then call all outstanding watchers.
457 - If we have been forked, recreate the kernel state.
458 - Update the kernel state with all outstanding changes.
459 - Update the &quot;event loop time&quot;.
460 - Calculate for how long to block.
461 - Block the process, waiting for any events.
462 - Queue all outstanding I/O (fd) events.
463 - Update the &quot;event loop time&quot; and do time jump handling.
464 - Queue all outstanding timers.
465 - Queue all outstanding periodics.
466 - If no events are pending now, queue all idle watchers.
467 - Queue all check watchers.
468 - Call all queued watchers in reverse order (i.e. check watchers first).
469 Signals and child watchers are implemented as I/O watchers, and will
470 be handled here by queueing them when their watcher gets executed.
471 - If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK
472 were used, return, otherwise continue with step *.
473
474</pre>
475 <p>Example: queue some jobs and then loop until no events are outsanding
476anymore.</p>
477<pre> ... queue jobs here, make sure they register event watchers as long
478 ... as they still have work to do (even an idle watcher will do..)
479 ev_loop (my_loop, 0);
480 ... jobs done. yeah!
481
482</pre>
264 </dd> 483 </dd>
265 <dt>ev_unloop (loop, how)</dt> 484 <dt>ev_unloop (loop, how)</dt>
266 <dd> 485 <dd>
267 <p>Can be used to make a call to <code>ev_loop</code> return early (but only after it 486 <p>Can be used to make a call to <code>ev_loop</code> return early (but only after it
268has processed all outstanding events). The <code>how</code> argument must be either 487has processed all outstanding events). The <code>how</code> argument must be either
269<code>EVUNLOOP_ONCE</code>, which will make the innermost <code>ev_loop</code> call return, or 488<code>EVUNLOOP_ONE</code>, which will make the innermost <code>ev_loop</code> call return, or
270<code>EVUNLOOP_ALL</code>, which will make all nested <code>ev_loop</code> calls return.</p> 489<code>EVUNLOOP_ALL</code>, which will make all nested <code>ev_loop</code> calls return.</p>
271 </dd> 490 </dd>
272 <dt>ev_ref (loop)</dt> 491 <dt>ev_ref (loop)</dt>
273 <dt>ev_unref (loop)</dt> 492 <dt>ev_unref (loop)</dt>
274 <dd> 493 <dd>
280example, libev itself uses this for its internal signal pipe: It is not 499example, libev itself uses this for its internal signal pipe: It is not
281visible to the libev user and should not keep <code>ev_loop</code> from exiting if 500visible to the libev user and should not keep <code>ev_loop</code> from exiting if
282no event watchers registered by it are active. It is also an excellent 501no event watchers registered by it are active. It is also an excellent
283way to do this for generic recurring timers or from within third-party 502way to do this for generic recurring timers or from within third-party
284libraries. Just remember to <i>unref after start</i> and <i>ref before stop</i>.</p> 503libraries. Just remember to <i>unref after start</i> and <i>ref before stop</i>.</p>
504 <p>Example: create a signal watcher, but keep it from keeping <code>ev_loop</code>
505running when nothing else is active.</p>
506<pre> struct dv_signal exitsig;
507 ev_signal_init (&amp;exitsig, sig_cb, SIGINT);
508 ev_signal_start (myloop, &amp;exitsig);
509 evf_unref (myloop);
510
511</pre>
512 <p>Example: for some weird reason, unregister the above signal handler again.</p>
513<pre> ev_ref (myloop);
514 ev_signal_stop (myloop, &amp;exitsig);
515
516</pre>
285 </dd> 517 </dd>
286</dl> 518</dl>
519
520
521
522
287 523
288</div> 524</div>
289<h1 id="ANATOMY_OF_A_WATCHER">ANATOMY OF A WATCHER</h1><p><a href="#TOP" class="toplink">Top</a></p> 525<h1 id="ANATOMY_OF_A_WATCHER">ANATOMY OF A WATCHER</h1><p><a href="#TOP" class="toplink">Top</a></p>
290<div id="ANATOMY_OF_A_WATCHER_CONTENT"> 526<div id="ANATOMY_OF_A_WATCHER_CONTENT">
291<p>A watcher is a structure that you create and register to record your 527<p>A watcher is a structure that you create and register to record your
321with a watcher-specific start function (<code>ev_&lt;type&gt;_start (loop, watcher 557with a watcher-specific start function (<code>ev_&lt;type&gt;_start (loop, watcher
322*)</code>), and you can stop watching for events at any time by calling the 558*)</code>), and you can stop watching for events at any time by calling the
323corresponding stop function (<code>ev_&lt;type&gt;_stop (loop, watcher *)</code>.</p> 559corresponding stop function (<code>ev_&lt;type&gt;_stop (loop, watcher *)</code>.</p>
324<p>As long as your watcher is active (has been started but not stopped) you 560<p>As long as your watcher is active (has been started but not stopped) you
325must not touch the values stored in it. Most specifically you must never 561must not touch the values stored in it. Most specifically you must never
326reinitialise it or call its set method.</p> 562reinitialise it or call its <code>set</code> macro.</p>
327<p>You cna check whether an event is active by calling the <code>ev_is_active
328(watcher *)</code> macro. To see whether an event is outstanding (but the
329callback for it has not been called yet) you cna use the <code>ev_is_pending
330(watcher *)</code> macro.</p>
331<p>Each and every callback receives the event loop pointer as first, the 563<p>Each and every callback receives the event loop pointer as first, the
332registered watcher structure as second, and a bitset of received events as 564registered watcher structure as second, and a bitset of received events as
333third argument.</p> 565third argument.</p>
334<p>The rceeived events usually include a single bit per event type received 566<p>The received events usually include a single bit per event type received
335(you can receive multiple events at the same time). The possible bit masks 567(you can receive multiple events at the same time). The possible bit masks
336are:</p> 568are:</p>
337<dl> 569<dl>
338 <dt><code>EV_READ</code></dt> 570 <dt><code>EV_READ</code></dt>
339 <dt><code>EV_WRITE</code></dt> 571 <dt><code>EV_WRITE</code></dt>
354 <p>The signal specified in the <code>ev_signal</code> watcher has been received by a thread.</p> 586 <p>The signal specified in the <code>ev_signal</code> watcher has been received by a thread.</p>
355 </dd> 587 </dd>
356 <dt><code>EV_CHILD</code></dt> 588 <dt><code>EV_CHILD</code></dt>
357 <dd> 589 <dd>
358 <p>The pid specified in the <code>ev_child</code> watcher has received a status change.</p> 590 <p>The pid specified in the <code>ev_child</code> watcher has received a status change.</p>
591 </dd>
592 <dt><code>EV_STAT</code></dt>
593 <dd>
594 <p>The path specified in the <code>ev_stat</code> watcher changed its attributes somehow.</p>
359 </dd> 595 </dd>
360 <dt><code>EV_IDLE</code></dt> 596 <dt><code>EV_IDLE</code></dt>
361 <dd> 597 <dd>
362 <p>The <code>ev_idle</code> watcher has determined that you have nothing better to do.</p> 598 <p>The <code>ev_idle</code> watcher has determined that you have nothing better to do.</p>
363 </dd> 599 </dd>
386programs, though, so beware.</p> 622programs, though, so beware.</p>
387 </dd> 623 </dd>
388</dl> 624</dl>
389 625
390</div> 626</div>
627<h2 id="GENERIC_WATCHER_FUNCTIONS">GENERIC WATCHER FUNCTIONS</h2>
628<div id="GENERIC_WATCHER_FUNCTIONS_CONTENT">
629<p>In the following description, <code>TYPE</code> stands for the watcher type,
630e.g. <code>timer</code> for <code>ev_timer</code> watchers and <code>io</code> for <code>ev_io</code> watchers.</p>
631<dl>
632 <dt><code>ev_init</code> (ev_TYPE *watcher, callback)</dt>
633 <dd>
634 <p>This macro initialises the generic portion of a watcher. The contents
635of the watcher object can be arbitrary (so <code>malloc</code> will do). Only
636the generic parts of the watcher are initialised, you <i>need</i> to call
637the type-specific <code>ev_TYPE_set</code> macro afterwards to initialise the
638type-specific parts. For each type there is also a <code>ev_TYPE_init</code> macro
639which rolls both calls into one.</p>
640 <p>You can reinitialise a watcher at any time as long as it has been stopped
641(or never started) and there are no pending events outstanding.</p>
642 <p>The callback is always of type <code>void (*)(ev_loop *loop, ev_TYPE *watcher,
643int revents)</code>.</p>
644 </dd>
645 <dt><code>ev_TYPE_set</code> (ev_TYPE *, [args])</dt>
646 <dd>
647 <p>This macro initialises the type-specific parts of a watcher. You need to
648call <code>ev_init</code> at least once before you call this macro, but you can
649call <code>ev_TYPE_set</code> any number of times. You must not, however, call this
650macro on a watcher that is active (it can be pending, however, which is a
651difference to the <code>ev_init</code> macro).</p>
652 <p>Although some watcher types do not have type-specific arguments
653(e.g. <code>ev_prepare</code>) you still need to call its <code>set</code> macro.</p>
654 </dd>
655 <dt><code>ev_TYPE_init</code> (ev_TYPE *watcher, callback, [args])</dt>
656 <dd>
657 <p>This convinience macro rolls both <code>ev_init</code> and <code>ev_TYPE_set</code> macro
658calls into a single call. This is the most convinient method to initialise
659a watcher. The same limitations apply, of course.</p>
660 </dd>
661 <dt><code>ev_TYPE_start</code> (loop *, ev_TYPE *watcher)</dt>
662 <dd>
663 <p>Starts (activates) the given watcher. Only active watchers will receive
664events. If the watcher is already active nothing will happen.</p>
665 </dd>
666 <dt><code>ev_TYPE_stop</code> (loop *, ev_TYPE *watcher)</dt>
667 <dd>
668 <p>Stops the given watcher again (if active) and clears the pending
669status. It is possible that stopped watchers are pending (for example,
670non-repeating timers are being stopped when they become pending), but
671<code>ev_TYPE_stop</code> ensures that the watcher is neither active nor pending. If
672you want to free or reuse the memory used by the watcher it is therefore a
673good idea to always call its <code>ev_TYPE_stop</code> function.</p>
674 </dd>
675 <dt>bool ev_is_active (ev_TYPE *watcher)</dt>
676 <dd>
677 <p>Returns a true value iff the watcher is active (i.e. it has been started
678and not yet been stopped). As long as a watcher is active you must not modify
679it.</p>
680 </dd>
681 <dt>bool ev_is_pending (ev_TYPE *watcher)</dt>
682 <dd>
683 <p>Returns a true value iff the watcher is pending, (i.e. it has outstanding
684events but its callback has not yet been invoked). As long as a watcher
685is pending (but not active) you must not call an init function on it (but
686<code>ev_TYPE_set</code> is safe) and you must make sure the watcher is available to
687libev (e.g. you cnanot <code>free ()</code> it).</p>
688 </dd>
689 <dt>callback = ev_cb (ev_TYPE *watcher)</dt>
690 <dd>
691 <p>Returns the callback currently set on the watcher.</p>
692 </dd>
693 <dt>ev_cb_set (ev_TYPE *watcher, callback)</dt>
694 <dd>
695 <p>Change the callback. You can change the callback at virtually any time
696(modulo threads).</p>
697 </dd>
698</dl>
699
700
701
702
703
704</div>
391<h2 id="ASSOCIATING_CUSTOM_DATA_WITH_A_WATCH">ASSOCIATING CUSTOM DATA WITH A WATCHER</h2> 705<h2 id="ASSOCIATING_CUSTOM_DATA_WITH_A_WATCH">ASSOCIATING CUSTOM DATA WITH A WATCHER</h2>
392<div id="ASSOCIATING_CUSTOM_DATA_WITH_A_WATCH-2"> 706<div id="ASSOCIATING_CUSTOM_DATA_WITH_A_WATCH-2">
393<p>Each watcher has, by default, a member <code>void *data</code> that you can change 707<p>Each watcher has, by default, a member <code>void *data</code> that you can change
394and read at any time, libev will completely ignore it. This cna be used 708and read at any time, libev will completely ignore it. This can be used
395to associate arbitrary data with your watcher. If you need more data and 709to associate arbitrary data with your watcher. If you need more data and
396don't want to allocate memory and store a pointer to it in that data 710don't want to allocate memory and store a pointer to it in that data
397member, you can also &quot;subclass&quot; the watcher type and provide your own 711member, you can also &quot;subclass&quot; the watcher type and provide your own
398data:</p> 712data:</p>
399<pre> struct my_io 713<pre> struct my_io
423 737
424</div> 738</div>
425<h1 id="WATCHER_TYPES">WATCHER TYPES</h1><p><a href="#TOP" class="toplink">Top</a></p> 739<h1 id="WATCHER_TYPES">WATCHER TYPES</h1><p><a href="#TOP" class="toplink">Top</a></p>
426<div id="WATCHER_TYPES_CONTENT"> 740<div id="WATCHER_TYPES_CONTENT">
427<p>This section describes each watcher in detail, but will not repeat 741<p>This section describes each watcher in detail, but will not repeat
428information given in the last section.</p> 742information given in the last section. Any initialisation/set macros,
743functions and members specific to the watcher type are explained.</p>
744<p>Members are additionally marked with either <i>[read-only]</i>, meaning that,
745while the watcher is active, you can look at the member and expect some
746sensible content, but you must not modify it (you can modify it while the
747watcher is stopped to your hearts content), or <i>[read-write]</i>, which
748means you can expect it to have some sensible content while the watcher
749is active, but you can also modify it. Modifying it may not do something
750sensible or take immediate effect (or do anything at all), but libev will
751not crash or malfunction in any way.</p>
429 752
753
754
755
756
430</div> 757</div>
431<h2 id="code_ev_io_code_is_this_file_descrip"><code>ev_io</code> - is this file descriptor readable or writable</h2> 758<h2 id="code_ev_io_code_is_this_file_descrip"><code>ev_io</code> - is this file descriptor readable or writable?</h2>
432<div id="code_ev_io_code_is_this_file_descrip-2"> 759<div id="code_ev_io_code_is_this_file_descrip-2">
433<p>I/O watchers check whether a file descriptor is readable or writable 760<p>I/O watchers check whether a file descriptor is readable or writable
434in each iteration of the event loop (This behaviour is called 761in each iteration of the event loop, or, more precisely, when reading
435level-triggering because you keep receiving events as long as the 762would not block the process and writing would at least be able to write
436condition persists. Remember you cna stop the watcher if you don't want to 763some data. This behaviour is called level-triggering because you keep
437act on the event and neither want to receive future events).</p> 764receiving events as long as the condition persists. Remember you can stop
765the watcher if you don't want to act on the event and neither want to
766receive future events.</p>
438<p>In general you can register as many read and/or write event watchers oer 767<p>In general you can register as many read and/or write event watchers per
439fd as you want (as long as you don't confuse yourself). Setting all file 768fd as you want (as long as you don't confuse yourself). Setting all file
440descriptors to non-blocking mode is also usually a good idea (but not 769descriptors to non-blocking mode is also usually a good idea (but not
441required if you know what you are doing).</p> 770required if you know what you are doing).</p>
442<p>You have to be careful with dup'ed file descriptors, though. Some backends 771<p>You have to be careful with dup'ed file descriptors, though. Some backends
443(the linux epoll backend is a notable example) cannot handle dup'ed file 772(the linux epoll backend is a notable example) cannot handle dup'ed file
444descriptors correctly if you register interest in two or more fds pointing 773descriptors correctly if you register interest in two or more fds pointing
445to the same file/socket etc. description.</p> 774to the same underlying file/socket/etc. description (that is, they share
775the same underlying &quot;file open&quot;).</p>
446<p>If you must do this, then force the use of a known-to-be-good backend 776<p>If you must do this, then force the use of a known-to-be-good backend
447(at the time of this writing, this includes only EVMETHOD_SELECT and 777(at the time of this writing, this includes only <code>EVBACKEND_SELECT</code> and
448EVMETHOD_POLL).</p> 778<code>EVBACKEND_POLL</code>).</p>
779<p>Another thing you have to watch out for is that it is quite easy to
780receive &quot;spurious&quot; readyness notifications, that is your callback might
781be called with <code>EV_READ</code> but a subsequent <code>read</code>(2) will actually block
782because there is no data. Not only are some backends known to create a
783lot of those (for example solaris ports), it is very easy to get into
784this situation even with a relatively standard program structure. Thus
785it is best to always use non-blocking I/O: An extra <code>read</code>(2) returning
786<code>EAGAIN</code> is far preferable to a program hanging until some data arrives.</p>
787<p>If you cannot run the fd in non-blocking mode (for example you should not
788play around with an Xlib connection), then you have to seperately re-test
789wether a file descriptor is really ready with a known-to-be good interface
790such as poll (fortunately in our Xlib example, Xlib already does this on
791its own, so its quite safe to use).</p>
449<dl> 792<dl>
450 <dt>ev_io_init (ev_io *, callback, int fd, int events)</dt> 793 <dt>ev_io_init (ev_io *, callback, int fd, int events)</dt>
451 <dt>ev_io_set (ev_io *, int fd, int events)</dt> 794 <dt>ev_io_set (ev_io *, int fd, int events)</dt>
452 <dd> 795 <dd>
453 <p>Configures an <code>ev_io</code> watcher. The fd is the file descriptor to rceeive 796 <p>Configures an <code>ev_io</code> watcher. The <code>fd</code> is the file descriptor to
454events for and events is either <code>EV_READ</code>, <code>EV_WRITE</code> or <code>EV_READ | 797rceeive events for and events is either <code>EV_READ</code>, <code>EV_WRITE</code> or
455EV_WRITE</code> to receive the given events.</p> 798<code>EV_READ | EV_WRITE</code> to receive the given events.</p>
799 </dd>
800 <dt>int fd [read-only]</dt>
801 <dd>
802 <p>The file descriptor being watched.</p>
803 </dd>
804 <dt>int events [read-only]</dt>
805 <dd>
806 <p>The events being watched.</p>
456 </dd> 807 </dd>
457</dl> 808</dl>
809<p>Example: call <code>stdin_readable_cb</code> when STDIN_FILENO has become, well
810readable, but only once. Since it is likely line-buffered, you could
811attempt to read a whole line in the callback:</p>
812<pre> static void
813 stdin_readable_cb (struct ev_loop *loop, struct ev_io *w, int revents)
814 {
815 ev_io_stop (loop, w);
816 .. read from stdin here (or from w-&gt;fd) and haqndle any I/O errors
817 }
458 818
819 ...
820 struct ev_loop *loop = ev_default_init (0);
821 struct ev_io stdin_readable;
822 ev_io_init (&amp;stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
823 ev_io_start (loop, &amp;stdin_readable);
824 ev_loop (loop, 0);
825
826
827
828
829</pre>
830
459</div> 831</div>
460<h2 id="code_ev_timer_code_relative_and_opti"><code>ev_timer</code> - relative and optionally recurring timeouts</h2> 832<h2 id="code_ev_timer_code_relative_and_opti"><code>ev_timer</code> - relative and optionally repeating timeouts</h2>
461<div id="code_ev_timer_code_relative_and_opti-2"> 833<div id="code_ev_timer_code_relative_and_opti-2">
462<p>Timer watchers are simple relative timers that generate an event after a 834<p>Timer watchers are simple relative timers that generate an event after a
463given time, and optionally repeating in regular intervals after that.</p> 835given time, and optionally repeating in regular intervals after that.</p>
464<p>The timers are based on real time, that is, if you register an event that 836<p>The timers are based on real time, that is, if you register an event that
465times out after an hour and youreset your system clock to last years 837times out after an hour and you reset your system clock to last years
466time, it will still time out after (roughly) and hour. &quot;Roughly&quot; because 838time, it will still time out after (roughly) and hour. &quot;Roughly&quot; because
467detecting time jumps is hard, and soem inaccuracies are unavoidable (the 839detecting time jumps is hard, and some inaccuracies are unavoidable (the
468monotonic clock option helps a lot here).</p> 840monotonic clock option helps a lot here).</p>
469<p>The relative timeouts are calculated relative to the <code>ev_now ()</code> 841<p>The relative timeouts are calculated relative to the <code>ev_now ()</code>
470time. This is usually the right thing as this timestamp refers to the time 842time. This is usually the right thing as this timestamp refers to the time
471of the event triggering whatever timeout you are modifying/starting. If 843of the event triggering whatever timeout you are modifying/starting. If
472you suspect event processing to be delayed and you *need* to base the timeout 844you suspect event processing to be delayed and you <i>need</i> to base the timeout
473ion the current time, use something like this to adjust for this:</p> 845on the current time, use something like this to adjust for this:</p>
474<pre> ev_timer_set (&amp;timer, after + ev_now () - ev_time (), 0.); 846<pre> ev_timer_set (&amp;timer, after + ev_now () - ev_time (), 0.);
475 847
476</pre> 848</pre>
849<p>The callback is guarenteed to be invoked only when its timeout has passed,
850but if multiple timers become ready during the same loop iteration then
851order of execution is undefined.</p>
477<dl> 852<dl>
478 <dt>ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)</dt> 853 <dt>ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)</dt>
479 <dt>ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)</dt> 854 <dt>ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)</dt>
480 <dd> 855 <dd>
481 <p>Configure the timer to trigger after <code>after</code> seconds. If <code>repeat</code> is 856 <p>Configure the timer to trigger after <code>after</code> seconds. If <code>repeat</code> is
483timer will automatically be configured to trigger again <code>repeat</code> seconds 858timer will automatically be configured to trigger again <code>repeat</code> seconds
484later, again, and again, until stopped manually.</p> 859later, again, and again, until stopped manually.</p>
485 <p>The timer itself will do a best-effort at avoiding drift, that is, if you 860 <p>The timer itself will do a best-effort at avoiding drift, that is, if you
486configure a timer to trigger every 10 seconds, then it will trigger at 861configure a timer to trigger every 10 seconds, then it will trigger at
487exactly 10 second intervals. If, however, your program cannot keep up with 862exactly 10 second intervals. If, however, your program cannot keep up with
488the timer (ecause it takes longer than those 10 seconds to do stuff) the 863the timer (because it takes longer than those 10 seconds to do stuff) the
489timer will not fire more than once per event loop iteration.</p> 864timer will not fire more than once per event loop iteration.</p>
490 </dd> 865 </dd>
491 <dt>ev_timer_again (loop)</dt> 866 <dt>ev_timer_again (loop)</dt>
492 <dd> 867 <dd>
493 <p>This will act as if the timer timed out and restart it again if it is 868 <p>This will act as if the timer timed out and restart it again if it is
494repeating. The exact semantics are:</p> 869repeating. The exact semantics are:</p>
495 <p>If the timer is started but nonrepeating, stop it.</p> 870 <p>If the timer is started but nonrepeating, stop it.</p>
496 <p>If the timer is repeating, either start it if necessary (with the repeat 871 <p>If the timer is repeating, either start it if necessary (with the repeat
497value), or reset the running timer to the repeat value.</p> 872value), or reset the running timer to the repeat value.</p>
498 <p>This sounds a bit complicated, but here is a useful and typical 873 <p>This sounds a bit complicated, but here is a useful and typical
499example: Imagine you have a tcp connection and you want a so-called idle 874example: Imagine you have a tcp connection and you want a so-called
500timeout, that is, you want to be called when there have been, say, 60 875idle timeout, that is, you want to be called when there have been,
501seconds of inactivity on the socket. The easiest way to do this is to 876say, 60 seconds of inactivity on the socket. The easiest way to do
502configure an <code>ev_timer</code> with after=repeat=60 and calling ev_timer_again each 877this is to configure an <code>ev_timer</code> with <code>after</code>=<code>repeat</code>=<code>60</code> and calling
503time you successfully read or write some data. If you go into an idle 878<code>ev_timer_again</code> each time you successfully read or write some data. If
504state where you do not expect data to travel on the socket, you can stop 879you go into an idle state where you do not expect data to travel on the
505the timer, and again will automatically restart it if need be.</p> 880socket, you can stop the timer, and again will automatically restart it if
881need be.</p>
882 <p>You can also ignore the <code>after</code> value and <code>ev_timer_start</code> altogether
883and only ever use the <code>repeat</code> value:</p>
884<pre> ev_timer_init (timer, callback, 0., 5.);
885 ev_timer_again (loop, timer);
886 ...
887 timer-&gt;again = 17.;
888 ev_timer_again (loop, timer);
889 ...
890 timer-&gt;again = 10.;
891 ev_timer_again (loop, timer);
892
893</pre>
894 <p>This is more efficient then stopping/starting the timer eahc time you want
895to modify its timeout value.</p>
896 </dd>
897 <dt>ev_tstamp repeat [read-write]</dt>
898 <dd>
899 <p>The current <code>repeat</code> value. Will be used each time the watcher times out
900or <code>ev_timer_again</code> is called and determines the next timeout (if any),
901which is also when any modifications are taken into account.</p>
506 </dd> 902 </dd>
507</dl> 903</dl>
904<p>Example: create a timer that fires after 60 seconds.</p>
905<pre> static void
906 one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
907 {
908 .. one minute over, w is actually stopped right here
909 }
508 910
911 struct ev_timer mytimer;
912 ev_timer_init (&amp;mytimer, one_minute_cb, 60., 0.);
913 ev_timer_start (loop, &amp;mytimer);
914
915</pre>
916<p>Example: create a timeout timer that times out after 10 seconds of
917inactivity.</p>
918<pre> static void
919 timeout_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
920 {
921 .. ten seconds without any activity
922 }
923
924 struct ev_timer mytimer;
925 ev_timer_init (&amp;mytimer, timeout_cb, 0., 10.); /* note, only repeat used */
926 ev_timer_again (&amp;mytimer); /* start timer */
927 ev_loop (loop, 0);
928
929 // and in some piece of code that gets executed on any &quot;activity&quot;:
930 // reset the timeout to start ticking again at 10 seconds
931 ev_timer_again (&amp;mytimer);
932
933
934
935
936</pre>
937
509</div> 938</div>
510<h2 id="code_ev_periodic_code_to_cron_or_not"><code>ev_periodic</code> - to cron or not to cron it</h2> 939<h2 id="code_ev_periodic_code_to_cron_or_not"><code>ev_periodic</code> - to cron or not to cron?</h2>
511<div id="code_ev_periodic_code_to_cron_or_not-2"> 940<div id="code_ev_periodic_code_to_cron_or_not-2">
512<p>Periodic watchers are also timers of a kind, but they are very versatile 941<p>Periodic watchers are also timers of a kind, but they are very versatile
513(and unfortunately a bit complex).</p> 942(and unfortunately a bit complex).</p>
514<p>Unlike <code>ev_timer</code>'s, they are not based on real time (or relative time) 943<p>Unlike <code>ev_timer</code>'s, they are not based on real time (or relative time)
515but on wallclock time (absolute time). You can tell a periodic watcher 944but on wallclock time (absolute time). You can tell a periodic watcher
516to trigger &quot;at&quot; some specific point in time. For example, if you tell a 945to trigger &quot;at&quot; some specific point in time. For example, if you tell a
517periodic watcher to trigger in 10 seconds (by specifiying e.g. c&lt;ev_now () 946periodic watcher to trigger in 10 seconds (by specifiying e.g. <code>ev_now ()
518+ 10.&gt;) and then reset your system clock to the last year, then it will 947+ 10.</code>) and then reset your system clock to the last year, then it will
519take a year to trigger the event (unlike an <code>ev_timer</code>, which would trigger 948take a year to trigger the event (unlike an <code>ev_timer</code>, which would trigger
520roughly 10 seconds later and of course not if you reset your system time 949roughly 10 seconds later and of course not if you reset your system time
521again).</p> 950again).</p>
522<p>They can also be used to implement vastly more complex timers, such as 951<p>They can also be used to implement vastly more complex timers, such as
523triggering an event on eahc midnight, local time.</p> 952triggering an event on eahc midnight, local time.</p>
953<p>As with timers, the callback is guarenteed to be invoked only when the
954time (<code>at</code>) has been passed, but if multiple periodic timers become ready
955during the same loop iteration then order of execution is undefined.</p>
524<dl> 956<dl>
525 <dt>ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)</dt> 957 <dt>ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)</dt>
526 <dt>ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)</dt> 958 <dt>ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)</dt>
527 <dd> 959 <dd>
528 <p>Lots of arguments, lets sort it out... There are basically three modes of 960 <p>Lots of arguments, lets sort it out... There are basically three modes of
529operation, and we will explain them from simplest to complex:</p> 961operation, and we will explain them from simplest to complex:</p>
530
531
532
533
534 <p> 962 <p>
535 <dl> 963 <dl>
536 <dt>* absolute timer (interval = reschedule_cb = 0)</dt> 964 <dt>* absolute timer (interval = reschedule_cb = 0)</dt>
537 <dd> 965 <dd>
538 <p>In this configuration the watcher triggers an event at the wallclock time 966 <p>In this configuration the watcher triggers an event at the wallclock time
562 <dd> 990 <dd>
563 <p>In this mode the values for <code>interval</code> and <code>at</code> are both being 991 <p>In this mode the values for <code>interval</code> and <code>at</code> are both being
564ignored. Instead, each time the periodic watcher gets scheduled, the 992ignored. Instead, each time the periodic watcher gets scheduled, the
565reschedule callback will be called with the watcher as first, and the 993reschedule callback will be called with the watcher as first, and the
566current time as second argument.</p> 994current time as second argument.</p>
567 <p>NOTE: <i>This callback MUST NOT stop or destroy the periodic or any other 995 <p>NOTE: <i>This callback MUST NOT stop or destroy any periodic watcher,
568periodic watcher, ever, or make any event loop modifications</i>. If you need 996ever, or make any event loop modifications</i>. If you need to stop it,
569to stop it, return <code>now + 1e30</code> (or so, fudge fudge) and stop it afterwards.</p> 997return <code>now + 1e30</code> (or so, fudge fudge) and stop it afterwards (e.g. by
570 <p>Also, <i>this callback must always return a time that is later than the 998starting a prepare watcher).</p>
571passed <code>now</code> value</i>. Not even <code>now</code> itself will be ok.</p>
572 <p>Its prototype is <code>ev_tstamp (*reschedule_cb)(struct ev_periodic *w, 999 <p>Its prototype is <code>ev_tstamp (*reschedule_cb)(struct ev_periodic *w,
573ev_tstamp now)</code>, e.g.:</p> 1000ev_tstamp now)</code>, e.g.:</p>
574<pre> static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now) 1001<pre> static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)
575 { 1002 {
576 return now + 60.; 1003 return now + 60.;
579</pre> 1006</pre>
580 <p>It must return the next time to trigger, based on the passed time value 1007 <p>It must return the next time to trigger, based on the passed time value
581(that is, the lowest time value larger than to the second argument). It 1008(that is, the lowest time value larger than to the second argument). It
582will usually be called just before the callback will be triggered, but 1009will usually be called just before the callback will be triggered, but
583might be called at other times, too.</p> 1010might be called at other times, too.</p>
1011 <p>NOTE: <i>This callback must always return a time that is later than the
1012passed <code>now</code> value</i>. Not even <code>now</code> itself will do, it <i>must</i> be larger.</p>
584 <p>This can be used to create very complex timers, such as a timer that 1013 <p>This can be used to create very complex timers, such as a timer that
585triggers on each midnight, local time. To do this, you would calculate the 1014triggers on each midnight, local time. To do this, you would calculate the
586next midnight after <code>now</code> and return the timestamp value for this. How you do this 1015next midnight after <code>now</code> and return the timestamp value for this. How
587is, again, up to you (but it is not trivial).</p> 1016you do this is, again, up to you (but it is not trivial, which is the main
1017reason I omitted it as an example).</p>
588 </dd> 1018 </dd>
589 </dl> 1019 </dl>
590 </p> 1020 </p>
591 </dd> 1021 </dd>
592 <dt>ev_periodic_again (loop, ev_periodic *)</dt> 1022 <dt>ev_periodic_again (loop, ev_periodic *)</dt>
594 <p>Simply stops and restarts the periodic watcher again. This is only useful 1024 <p>Simply stops and restarts the periodic watcher again. This is only useful
595when you changed some parameters or the reschedule callback would return 1025when you changed some parameters or the reschedule callback would return
596a different time than the last time it was called (e.g. in a crond like 1026a different time than the last time it was called (e.g. in a crond like
597program when the crontabs have changed).</p> 1027program when the crontabs have changed).</p>
598 </dd> 1028 </dd>
1029 <dt>ev_tstamp interval [read-write]</dt>
1030 <dd>
1031 <p>The current interval value. Can be modified any time, but changes only
1032take effect when the periodic timer fires or <code>ev_periodic_again</code> is being
1033called.</p>
1034 </dd>
1035 <dt>ev_tstamp (*reschedule_cb)(struct ev_periodic *w, ev_tstamp now) [read-write]</dt>
1036 <dd>
1037 <p>The current reschedule callback, or <code>0</code>, if this functionality is
1038switched off. Can be changed any time, but changes only take effect when
1039the periodic timer fires or <code>ev_periodic_again</code> is being called.</p>
1040 </dd>
599</dl> 1041</dl>
1042<p>Example: call a callback every hour, or, more precisely, whenever the
1043system clock is divisible by 3600. The callback invocation times have
1044potentially a lot of jittering, but good long-term stability.</p>
1045<pre> static void
1046 clock_cb (struct ev_loop *loop, struct ev_io *w, int revents)
1047 {
1048 ... its now a full hour (UTC, or TAI or whatever your clock follows)
1049 }
600 1050
1051 struct ev_periodic hourly_tick;
1052 ev_periodic_init (&amp;hourly_tick, clock_cb, 0., 3600., 0);
1053 ev_periodic_start (loop, &amp;hourly_tick);
1054
1055</pre>
1056<p>Example: the same as above, but use a reschedule callback to do it:</p>
1057<pre> #include &lt;math.h&gt;
1058
1059 static ev_tstamp
1060 my_scheduler_cb (struct ev_periodic *w, ev_tstamp now)
1061 {
1062 return fmod (now, 3600.) + 3600.;
1063 }
1064
1065 ev_periodic_init (&amp;hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
1066
1067</pre>
1068<p>Example: call a callback every hour, starting now:</p>
1069<pre> struct ev_periodic hourly_tick;
1070 ev_periodic_init (&amp;hourly_tick, clock_cb,
1071 fmod (ev_now (loop), 3600.), 3600., 0);
1072 ev_periodic_start (loop, &amp;hourly_tick);
1073
1074
1075
1076
1077</pre>
1078
601</div> 1079</div>
602<h2 id="code_ev_signal_code_signal_me_when_a"><code>ev_signal</code> - signal me when a signal gets signalled</h2> 1080<h2 id="code_ev_signal_code_signal_me_when_a"><code>ev_signal</code> - signal me when a signal gets signalled!</h2>
603<div id="code_ev_signal_code_signal_me_when_a-2"> 1081<div id="code_ev_signal_code_signal_me_when_a-2">
604<p>Signal watchers will trigger an event when the process receives a specific 1082<p>Signal watchers will trigger an event when the process receives a specific
605signal one or more times. Even though signals are very asynchronous, libev 1083signal one or more times. Even though signals are very asynchronous, libev
606will try it's best to deliver signals synchronously, i.e. as part of the 1084will try it's best to deliver signals synchronously, i.e. as part of the
607normal event processing, like any other event.</p> 1085normal event processing, like any other event.</p>
608<p>You cna configure as many watchers as you like per signal. Only when the 1086<p>You can configure as many watchers as you like per signal. Only when the
609first watcher gets started will libev actually register a signal watcher 1087first watcher gets started will libev actually register a signal watcher
610with the kernel (thus it coexists with your own signal handlers as long 1088with the kernel (thus it coexists with your own signal handlers as long
611as you don't register any with libev). Similarly, when the last signal 1089as you don't register any with libev). Similarly, when the last signal
612watcher for a signal is stopped libev will reset the signal handler to 1090watcher for a signal is stopped libev will reset the signal handler to
613SIG_DFL (regardless of what it was set to before).</p> 1091SIG_DFL (regardless of what it was set to before).</p>
616 <dt>ev_signal_set (ev_signal *, int signum)</dt> 1094 <dt>ev_signal_set (ev_signal *, int signum)</dt>
617 <dd> 1095 <dd>
618 <p>Configures the watcher to trigger on the given signal number (usually one 1096 <p>Configures the watcher to trigger on the given signal number (usually one
619of the <code>SIGxxx</code> constants).</p> 1097of the <code>SIGxxx</code> constants).</p>
620 </dd> 1098 </dd>
1099 <dt>int signum [read-only]</dt>
1100 <dd>
1101 <p>The signal the watcher watches out for.</p>
1102 </dd>
621</dl> 1103</dl>
622 1104
1105
1106
1107
1108
623</div> 1109</div>
624<h2 id="code_ev_child_code_wait_for_pid_stat"><code>ev_child</code> - wait for pid status changes</h2> 1110<h2 id="code_ev_child_code_watch_out_for_pro"><code>ev_child</code> - watch out for process status changes</h2>
625<div id="code_ev_child_code_wait_for_pid_stat-2"> 1111<div id="code_ev_child_code_watch_out_for_pro-2">
626<p>Child watchers trigger when your process receives a SIGCHLD in response to 1112<p>Child watchers trigger when your process receives a SIGCHLD in response to
627some child status changes (most typically when a child of yours dies).</p> 1113some child status changes (most typically when a child of yours dies).</p>
628<dl> 1114<dl>
629 <dt>ev_child_init (ev_child *, callback, int pid)</dt> 1115 <dt>ev_child_init (ev_child *, callback, int pid)</dt>
630 <dt>ev_child_set (ev_child *, int pid)</dt> 1116 <dt>ev_child_set (ev_child *, int pid)</dt>
631 <dd> 1117 <dd>
632 <p>Configures the watcher to wait for status changes of process <code>pid</code> (or 1118 <p>Configures the watcher to wait for status changes of process <code>pid</code> (or
633<i>any</i> process if <code>pid</code> is specified as <code>0</code>). The callback can look 1119<i>any</i> process if <code>pid</code> is specified as <code>0</code>). The callback can look
634at the <code>rstatus</code> member of the <code>ev_child</code> watcher structure to see 1120at the <code>rstatus</code> member of the <code>ev_child</code> watcher structure to see
635the status word (use the macros from <code>sys/wait.h</code>). The <code>rpid</code> member 1121the status word (use the macros from <code>sys/wait.h</code> and see your systems
1122<code>waitpid</code> documentation). The <code>rpid</code> member contains the pid of the
636contains the pid of the process causing the status change.</p> 1123process causing the status change.</p>
1124 </dd>
1125 <dt>int pid [read-only]</dt>
1126 <dd>
1127 <p>The process id this watcher watches out for, or <code>0</code>, meaning any process id.</p>
1128 </dd>
1129 <dt>int rpid [read-write]</dt>
1130 <dd>
1131 <p>The process id that detected a status change.</p>
1132 </dd>
1133 <dt>int rstatus [read-write]</dt>
1134 <dd>
1135 <p>The process exit/trace status caused by <code>rpid</code> (see your systems
1136<code>waitpid</code> and <code>sys/wait.h</code> documentation for details).</p>
637 </dd> 1137 </dd>
638</dl> 1138</dl>
1139<p>Example: try to exit cleanly on SIGINT and SIGTERM.</p>
1140<pre> static void
1141 sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)
1142 {
1143 ev_unloop (loop, EVUNLOOP_ALL);
1144 }
639 1145
1146 struct ev_signal signal_watcher;
1147 ev_signal_init (&amp;signal_watcher, sigint_cb, SIGINT);
1148 ev_signal_start (loop, &amp;sigint_cb);
1149
1150
1151
1152
1153</pre>
1154
640</div> 1155</div>
1156<h2 id="code_ev_stat_code_did_the_file_attri"><code>ev_stat</code> - did the file attributes just change?</h2>
1157<div id="code_ev_stat_code_did_the_file_attri-2">
1158<p>This watches a filesystem path for attribute changes. That is, it calls
1159<code>stat</code> regularly (or when the OS says it changed) and sees if it changed
1160compared to the last time, invoking the callback if it did.</p>
1161<p>The path does not need to exist: changing from &quot;path exists&quot; to &quot;path does
1162not exist&quot; is a status change like any other. The condition &quot;path does
1163not exist&quot; is signified by the <code>st_nlink</code> field being zero (which is
1164otherwise always forced to be at least one) and all the other fields of
1165the stat buffer having unspecified contents.</p>
1166<p>Since there is no standard to do this, the portable implementation simply
1167calls <code>stat (2)</code> regulalry on the path to see if it changed somehow. You
1168can specify a recommended polling interval for this case. If you specify
1169a polling interval of <code>0</code> (highly recommended!) then a <i>suitable,
1170unspecified default</i> value will be used (which you can expect to be around
1171five seconds, although this might change dynamically). Libev will also
1172impose a minimum interval which is currently around <code>0.1</code>, but thats
1173usually overkill.</p>
1174<p>This watcher type is not meant for massive numbers of stat watchers,
1175as even with OS-supported change notifications, this can be
1176resource-intensive.</p>
1177<p>At the time of this writing, no specific OS backends are implemented, but
1178if demand increases, at least a kqueue and inotify backend will be added.</p>
1179<dl>
1180 <dt>ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)</dt>
1181 <dt>ev_stat_set (ev_stat *, const char *path, ev_tstamp interval)</dt>
1182 <dd>
1183 <p>Configures the watcher to wait for status changes of the given
1184<code>path</code>. The <code>interval</code> is a hint on how quickly a change is expected to
1185be detected and should normally be specified as <code>0</code> to let libev choose
1186a suitable value. The memory pointed to by <code>path</code> must point to the same
1187path for as long as the watcher is active.</p>
1188 <p>The callback will be receive <code>EV_STAT</code> when a change was detected,
1189relative to the attributes at the time the watcher was started (or the
1190last change was detected).</p>
1191 </dd>
1192 <dt>ev_stat_stat (ev_stat *)</dt>
1193 <dd>
1194 <p>Updates the stat buffer immediately with new values. If you change the
1195watched path in your callback, you could call this fucntion to avoid
1196detecting this change (while introducing a race condition). Can also be
1197useful simply to find out the new values.</p>
1198 </dd>
1199 <dt>ev_statdata attr [read-only]</dt>
1200 <dd>
1201 <p>The most-recently detected attributes of the file. Although the type is of
1202<code>ev_statdata</code>, this is usually the (or one of the) <code>struct stat</code> types
1203suitable for your system. If the <code>st_nlink</code> member is <code>0</code>, then there
1204was some error while <code>stat</code>ing the file.</p>
1205 </dd>
1206 <dt>ev_statdata prev [read-only]</dt>
1207 <dd>
1208 <p>The previous attributes of the file. The callback gets invoked whenever
1209<code>prev</code> != <code>attr</code>.</p>
1210 </dd>
1211 <dt>ev_tstamp interval [read-only]</dt>
1212 <dd>
1213 <p>The specified interval.</p>
1214 </dd>
1215 <dt>const char *path [read-only]</dt>
1216 <dd>
1217 <p>The filesystem path that is being watched.</p>
1218 </dd>
1219</dl>
1220<p>Example: Watch <code>/etc/passwd</code> for attribute changes.</p>
1221<pre> static void
1222 passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)
1223 {
1224 /* /etc/passwd changed in some way */
1225 if (w-&gt;attr.st_nlink)
1226 {
1227 printf (&quot;passwd current size %ld\n&quot;, (long)w-&gt;attr.st_size);
1228 printf (&quot;passwd current atime %ld\n&quot;, (long)w-&gt;attr.st_mtime);
1229 printf (&quot;passwd current mtime %ld\n&quot;, (long)w-&gt;attr.st_mtime);
1230 }
1231 else
1232 /* you shalt not abuse printf for puts */
1233 puts (&quot;wow, /etc/passwd is not there, expect problems. &quot;
1234 &quot;if this is windows, they already arrived\n&quot;);
1235 }
1236
1237 ...
1238 ev_stat passwd;
1239
1240 ev_stat_init (&amp;passwd, passwd_cb, &quot;/etc/passwd&quot;);
1241 ev_stat_start (loop, &amp;passwd);
1242
1243
1244
1245
1246</pre>
1247
1248</div>
641<h2 id="code_ev_idle_code_when_you_ve_got_no"><code>ev_idle</code> - when you've got nothing better to do</h2> 1249<h2 id="code_ev_idle_code_when_you_ve_got_no"><code>ev_idle</code> - when you've got nothing better to do...</h2>
642<div id="code_ev_idle_code_when_you_ve_got_no-2"> 1250<div id="code_ev_idle_code_when_you_ve_got_no-2">
643<p>Idle watchers trigger events when there are no other I/O or timer (or 1251<p>Idle watchers trigger events when there are no other events are pending
644periodic) events pending. That is, as long as your process is busy 1252(prepare, check and other idle watchers do not count). That is, as long
645handling sockets or timeouts it will not be called. But when your process 1253as your process is busy handling sockets or timeouts (or even signals,
646is idle all idle watchers are being called again and again - until 1254imagine) it will not be triggered. But when your process is idle all idle
1255watchers are being called again and again, once per event loop iteration -
647stopped, that is, or your process receives more events.</p> 1256until stopped, that is, or your process receives more events and becomes
1257busy.</p>
648<p>The most noteworthy effect is that as long as any idle watchers are 1258<p>The most noteworthy effect is that as long as any idle watchers are
649active, the process will not block when waiting for new events.</p> 1259active, the process will not block when waiting for new events.</p>
650<p>Apart from keeping your process non-blocking (which is a useful 1260<p>Apart from keeping your process non-blocking (which is a useful
651effect on its own sometimes), idle watchers are a good place to do 1261effect on its own sometimes), idle watchers are a good place to do
652&quot;pseudo-background processing&quot;, or delay processing stuff to after the 1262&quot;pseudo-background processing&quot;, or delay processing stuff to after the
657 <p>Initialises and configures the idle watcher - it has no parameters of any 1267 <p>Initialises and configures the idle watcher - it has no parameters of any
658kind. There is a <code>ev_idle_set</code> macro, but using it is utterly pointless, 1268kind. There is a <code>ev_idle_set</code> macro, but using it is utterly pointless,
659believe me.</p> 1269believe me.</p>
660 </dd> 1270 </dd>
661</dl> 1271</dl>
1272<p>Example: dynamically allocate an <code>ev_idle</code>, start it, and in the
1273callback, free it. Alos, use no error checking, as usual.</p>
1274<pre> static void
1275 idle_cb (struct ev_loop *loop, struct ev_idle *w, int revents)
1276 {
1277 free (w);
1278 // now do something you wanted to do when the program has
1279 // no longer asnything immediate to do.
1280 }
662 1281
1282 struct ev_idle *idle_watcher = malloc (sizeof (struct ev_idle));
1283 ev_idle_init (idle_watcher, idle_cb);
1284 ev_idle_start (loop, idle_cb);
1285
1286
1287
1288
1289</pre>
1290
663</div> 1291</div>
664<h2 id="prepare_and_check_your_hooks_into_th">prepare and check - your hooks into the event loop</h2> 1292<h2 id="code_ev_prepare_code_and_code_ev_che"><code>ev_prepare</code> and <code>ev_check</code> - customise your event loop!</h2>
665<div id="prepare_and_check_your_hooks_into_th-2"> 1293<div id="code_ev_prepare_code_and_code_ev_che-2">
666<p>Prepare and check watchers usually (but not always) are used in 1294<p>Prepare and check watchers are usually (but not always) used in tandem:
667tandom. Prepare watchers get invoked before the process blocks and check 1295prepare watchers get invoked before the process blocks and check watchers
668watchers afterwards.</p> 1296afterwards.</p>
1297<p>You <i>must not</i> call <code>ev_loop</code> or similar functions that enter
1298the current event loop from either <code>ev_prepare</code> or <code>ev_check</code>
1299watchers. Other loops than the current one are fine, however. The
1300rationale behind this is that you do not need to check for recursion in
1301those watchers, i.e. the sequence will always be <code>ev_prepare</code>, blocking,
1302<code>ev_check</code> so if you have one watcher of each kind they will always be
1303called in pairs bracketing the blocking call.</p>
669<p>Their main purpose is to integrate other event mechanisms into libev. This 1304<p>Their main purpose is to integrate other event mechanisms into libev and
670could be used, for example, to track variable changes, implement your own 1305their use is somewhat advanced. This could be used, for example, to track
671watchers, integrate net-snmp or a coroutine library and lots more.</p> 1306variable changes, implement your own watchers, integrate net-snmp or a
1307coroutine library and lots more. They are also occasionally useful if
1308you cache some data and want to flush it before blocking (for example,
1309in X programs you might want to do an <code>XFlush ()</code> in an <code>ev_prepare</code>
1310watcher).</p>
672<p>This is done by examining in each prepare call which file descriptors need 1311<p>This is done by examining in each prepare call which file descriptors need
673to be watched by the other library, registering <code>ev_io</code> watchers for them 1312to be watched by the other library, registering <code>ev_io</code> watchers for
674and starting an <code>ev_timer</code> watcher for any timeouts (many libraries provide 1313them and starting an <code>ev_timer</code> watcher for any timeouts (many libraries
675just this functionality). Then, in the check watcher you check for any 1314provide just this functionality). Then, in the check watcher you check for
676events that occured (by making your callbacks set soem flags for example) 1315any events that occured (by checking the pending status of all watchers
677and call back into the library.</p> 1316and stopping them) and call back into the library. The I/O and timer
1317callbacks will never actually be called (but must be valid nevertheless,
1318because you never know, you know?).</p>
678<p>As another example, the perl Coro module uses these hooks to integrate 1319<p>As another example, the Perl Coro module uses these hooks to integrate
679coroutines into libev programs, by yielding to other active coroutines 1320coroutines into libev programs, by yielding to other active coroutines
680during each prepare and only letting the process block if no coroutines 1321during each prepare and only letting the process block if no coroutines
681are ready to run.</p> 1322are ready to run (it's actually more complicated: it only runs coroutines
1323with priority higher than or equal to the event loop and one coroutine
1324of lower priority, but only once, using idle watchers to keep the event
1325loop from blocking if lower-priority coroutines are active, thus mapping
1326low-priority coroutines to idle/background tasks).</p>
682<dl> 1327<dl>
683 <dt>ev_prepare_init (ev_prepare *, callback)</dt> 1328 <dt>ev_prepare_init (ev_prepare *, callback)</dt>
684 <dt>ev_check_init (ev_check *, callback)</dt> 1329 <dt>ev_check_init (ev_check *, callback)</dt>
685 <dd> 1330 <dd>
686 <p>Initialises and configures the prepare or check watcher - they have no 1331 <p>Initialises and configures the prepare or check watcher - they have no
687parameters of any kind. There are <code>ev_prepare_set</code> and <code>ev_check_set</code> 1332parameters of any kind. There are <code>ev_prepare_set</code> and <code>ev_check_set</code>
688macros, but using them is utterly, utterly pointless.</p> 1333macros, but using them is utterly, utterly and completely pointless.</p>
689 </dd> 1334 </dd>
690</dl> 1335</dl>
1336<p>Example: To include a library such as adns, you would add IO watchers
1337and a timeout watcher in a prepare handler, as required by libadns, and
1338in a check watcher, destroy them and call into libadns. What follows is
1339pseudo-code only of course:</p>
1340<pre> static ev_io iow [nfd];
1341 static ev_timer tw;
1342
1343 static void
1344 io_cb (ev_loop *loop, ev_io *w, int revents)
1345 {
1346 // set the relevant poll flags
1347 // could also call adns_processreadable etc. here
1348 struct pollfd *fd = (struct pollfd *)w-&gt;data;
1349 if (revents &amp; EV_READ ) fd-&gt;revents |= fd-&gt;events &amp; POLLIN;
1350 if (revents &amp; EV_WRITE) fd-&gt;revents |= fd-&gt;events &amp; POLLOUT;
1351 }
1352
1353 // create io watchers for each fd and a timer before blocking
1354 static void
1355 adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents)
1356 {
1357 int timeout = 3600000;truct pollfd fds [nfd];
1358 // actual code will need to loop here and realloc etc.
1359 adns_beforepoll (ads, fds, &amp;nfd, &amp;timeout, timeval_from (ev_time ()));
1360
1361 /* the callback is illegal, but won't be called as we stop during check */
1362 ev_timer_init (&amp;tw, 0, timeout * 1e-3);
1363 ev_timer_start (loop, &amp;tw);
1364
1365 // create on ev_io per pollfd
1366 for (int i = 0; i &lt; nfd; ++i)
1367 {
1368 ev_io_init (iow + i, io_cb, fds [i].fd,
1369 ((fds [i].events &amp; POLLIN ? EV_READ : 0)
1370 | (fds [i].events &amp; POLLOUT ? EV_WRITE : 0)));
1371
1372 fds [i].revents = 0;
1373 iow [i].data = fds + i;
1374 ev_io_start (loop, iow + i);
1375 }
1376 }
1377
1378 // stop all watchers after blocking
1379 static void
1380 adns_check_cb (ev_loop *loop, ev_check *w, int revents)
1381 {
1382 ev_timer_stop (loop, &amp;tw);
1383
1384 for (int i = 0; i &lt; nfd; ++i)
1385 ev_io_stop (loop, iow + i);
1386
1387 adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));
1388 }
1389
1390
1391
1392
1393</pre>
1394
1395</div>
1396<h2 id="code_ev_embed_code_when_one_backend_"><code>ev_embed</code> - when one backend isn't enough...</h2>
1397<div id="code_ev_embed_code_when_one_backend_-2">
1398<p>This is a rather advanced watcher type that lets you embed one event loop
1399into another (currently only <code>ev_io</code> events are supported in the embedded
1400loop, other types of watchers might be handled in a delayed or incorrect
1401fashion and must not be used).</p>
1402<p>There are primarily two reasons you would want that: work around bugs and
1403prioritise I/O.</p>
1404<p>As an example for a bug workaround, the kqueue backend might only support
1405sockets on some platform, so it is unusable as generic backend, but you
1406still want to make use of it because you have many sockets and it scales
1407so nicely. In this case, you would create a kqueue-based loop and embed it
1408into your default loop (which might use e.g. poll). Overall operation will
1409be a bit slower because first libev has to poll and then call kevent, but
1410at least you can use both at what they are best.</p>
1411<p>As for prioritising I/O: rarely you have the case where some fds have
1412to be watched and handled very quickly (with low latency), and even
1413priorities and idle watchers might have too much overhead. In this case
1414you would put all the high priority stuff in one loop and all the rest in
1415a second one, and embed the second one in the first.</p>
1416<p>As long as the watcher is active, the callback will be invoked every time
1417there might be events pending in the embedded loop. The callback must then
1418call <code>ev_embed_sweep (mainloop, watcher)</code> to make a single sweep and invoke
1419their callbacks (you could also start an idle watcher to give the embedded
1420loop strictly lower priority for example). You can also set the callback
1421to <code>0</code>, in which case the embed watcher will automatically execute the
1422embedded loop sweep.</p>
1423<p>As long as the watcher is started it will automatically handle events. The
1424callback will be invoked whenever some events have been handled. You can
1425set the callback to <code>0</code> to avoid having to specify one if you are not
1426interested in that.</p>
1427<p>Also, there have not currently been made special provisions for forking:
1428when you fork, you not only have to call <code>ev_loop_fork</code> on both loops,
1429but you will also have to stop and restart any <code>ev_embed</code> watchers
1430yourself.</p>
1431<p>Unfortunately, not all backends are embeddable, only the ones returned by
1432<code>ev_embeddable_backends</code> are, which, unfortunately, does not include any
1433portable one.</p>
1434<p>So when you want to use this feature you will always have to be prepared
1435that you cannot get an embeddable loop. The recommended way to get around
1436this is to have a separate variables for your embeddable loop, try to
1437create it, and if that fails, use the normal loop for everything:</p>
1438<pre> struct ev_loop *loop_hi = ev_default_init (0);
1439 struct ev_loop *loop_lo = 0;
1440 struct ev_embed embed;
1441
1442 // see if there is a chance of getting one that works
1443 // (remember that a flags value of 0 means autodetection)
1444 loop_lo = ev_embeddable_backends () &amp; ev_recommended_backends ()
1445 ? ev_loop_new (ev_embeddable_backends () &amp; ev_recommended_backends ())
1446 : 0;
1447
1448 // if we got one, then embed it, otherwise default to loop_hi
1449 if (loop_lo)
1450 {
1451 ev_embed_init (&amp;embed, 0, loop_lo);
1452 ev_embed_start (loop_hi, &amp;embed);
1453 }
1454 else
1455 loop_lo = loop_hi;
1456
1457</pre>
1458<dl>
1459 <dt>ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)</dt>
1460 <dt>ev_embed_set (ev_embed *, callback, struct ev_loop *embedded_loop)</dt>
1461 <dd>
1462 <p>Configures the watcher to embed the given loop, which must be
1463embeddable. If the callback is <code>0</code>, then <code>ev_embed_sweep</code> will be
1464invoked automatically, otherwise it is the responsibility of the callback
1465to invoke it (it will continue to be called until the sweep has been done,
1466if you do not want thta, you need to temporarily stop the embed watcher).</p>
1467 </dd>
1468 <dt>ev_embed_sweep (loop, ev_embed *)</dt>
1469 <dd>
1470 <p>Make a single, non-blocking sweep over the embedded loop. This works
1471similarly to <code>ev_loop (embedded_loop, EVLOOP_NONBLOCK)</code>, but in the most
1472apropriate way for embedded loops.</p>
1473 </dd>
1474 <dt>struct ev_loop *loop [read-only]</dt>
1475 <dd>
1476 <p>The embedded event loop.</p>
1477 </dd>
1478</dl>
1479
1480
1481
1482
691 1483
692</div> 1484</div>
693<h1 id="OTHER_FUNCTIONS">OTHER FUNCTIONS</h1><p><a href="#TOP" class="toplink">Top</a></p> 1485<h1 id="OTHER_FUNCTIONS">OTHER FUNCTIONS</h1><p><a href="#TOP" class="toplink">Top</a></p>
694<div id="OTHER_FUNCTIONS_CONTENT"> 1486<div id="OTHER_FUNCTIONS_CONTENT">
695<p>There are some other fucntions of possible interest. Described. Here. Now.</p> 1487<p>There are some other functions of possible interest. Described. Here. Now.</p>
696<dl> 1488<dl>
697 <dt>ev_once (loop, int fd, int events, ev_tstamp timeout, callback)</dt> 1489 <dt>ev_once (loop, int fd, int events, ev_tstamp timeout, callback)</dt>
698 <dd> 1490 <dd>
699 <p>This function combines a simple timer and an I/O watcher, calls your 1491 <p>This function combines a simple timer and an I/O watcher, calls your
700callback on whichever event happens first and automatically stop both 1492callback on whichever event happens first and automatically stop both
701watchers. This is useful if you want to wait for a single event on an fd 1493watchers. This is useful if you want to wait for a single event on an fd
702or timeout without havign to allocate/configure/start/stop/free one or 1494or timeout without having to allocate/configure/start/stop/free one or
703more watchers yourself.</p> 1495more watchers yourself.</p>
704 <p>If <code>fd</code> is less than 0, then no I/O watcher will be started and events is 1496 <p>If <code>fd</code> is less than 0, then no I/O watcher will be started and events
705ignored. Otherwise, an <code>ev_io</code> watcher for the given <code>fd</code> and <code>events</code> set 1497is being ignored. Otherwise, an <code>ev_io</code> watcher for the given <code>fd</code> and
706will be craeted and started.</p> 1498<code>events</code> set will be craeted and started.</p>
707 <p>If <code>timeout</code> is less than 0, then no timeout watcher will be 1499 <p>If <code>timeout</code> is less than 0, then no timeout watcher will be
708started. Otherwise an <code>ev_timer</code> watcher with after = <code>timeout</code> (and repeat 1500started. Otherwise an <code>ev_timer</code> watcher with after = <code>timeout</code> (and
709= 0) will be started.</p> 1501repeat = 0) will be started. While <code>0</code> is a valid timeout, it is of
1502dubious value.</p>
710 <p>The callback has the type <code>void (*cb)(int revents, void *arg)</code> and 1503 <p>The callback has the type <code>void (*cb)(int revents, void *arg)</code> and gets
711gets passed an events set (normally a combination of <code>EV_ERROR</code>, <code>EV_READ</code>, 1504passed an <code>revents</code> set like normal event callbacks (a combination of
712<code>EV_WRITE</code> or <code>EV_TIMEOUT</code>) and the <code>arg</code> value passed to <code>ev_once</code>:</p> 1505<code>EV_ERROR</code>, <code>EV_READ</code>, <code>EV_WRITE</code> or <code>EV_TIMEOUT</code>) and the <code>arg</code>
1506value passed to <code>ev_once</code>:</p>
713<pre> static void stdin_ready (int revents, void *arg) 1507<pre> static void stdin_ready (int revents, void *arg)
714 { 1508 {
715 if (revents &amp; EV_TIMEOUT) 1509 if (revents &amp; EV_TIMEOUT)
716 /* doh, nothing entered */ 1510 /* doh, nothing entered */;
717 else if (revents &amp; EV_READ) 1511 else if (revents &amp; EV_READ)
718 /* stdin might have data for us, joy! */ 1512 /* stdin might have data for us, joy! */;
719 } 1513 }
720 1514
721 ev_once (STDIN_FILENO, EV_READm 10., stdin_ready, 0); 1515 ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
722 1516
723</pre> 1517</pre>
724 </dd> 1518 </dd>
725 <dt>ev_feed_event (loop, watcher, int events)</dt> 1519 <dt>ev_feed_event (ev_loop *, watcher *, int revents)</dt>
726 <dd> 1520 <dd>
727 <p>Feeds the given event set into the event loop, as if the specified event 1521 <p>Feeds the given event set into the event loop, as if the specified event
728has happened for the specified watcher (which must be a pointer to an 1522had happened for the specified watcher (which must be a pointer to an
729initialised but not necessarily active event watcher).</p> 1523initialised but not necessarily started event watcher).</p>
730 </dd> 1524 </dd>
731 <dt>ev_feed_fd_event (loop, int fd, int revents)</dt> 1525 <dt>ev_feed_fd_event (ev_loop *, int fd, int revents)</dt>
732 <dd> 1526 <dd>
733 <p>Feed an event on the given fd, as if a file descriptor backend detected it.</p> 1527 <p>Feed an event on the given fd, as if a file descriptor backend detected
1528the given events it.</p>
734 </dd> 1529 </dd>
735 <dt>ev_feed_signal_event (loop, int signum)</dt> 1530 <dt>ev_feed_signal_event (ev_loop *loop, int signum)</dt>
736 <dd> 1531 <dd>
737 <p>Feed an event as if the given signal occured (loop must be the default loop!).</p> 1532 <p>Feed an event as if the given signal occured (<code>loop</code> must be the default
1533loop!).</p>
738 </dd> 1534 </dd>
739</dl> 1535</dl>
1536
1537
1538
1539
1540
1541</div>
1542<h1 id="LIBEVENT_EMULATION">LIBEVENT EMULATION</h1><p><a href="#TOP" class="toplink">Top</a></p>
1543<div id="LIBEVENT_EMULATION_CONTENT">
1544<p>Libev offers a compatibility emulation layer for libevent. It cannot
1545emulate the internals of libevent, so here are some usage hints:</p>
1546<dl>
1547 <dt>* Use it by including &lt;event.h&gt;, as usual.</dt>
1548 <dt>* The following members are fully supported: ev_base, ev_callback,
1549ev_arg, ev_fd, ev_res, ev_events.</dt>
1550 <dt>* Avoid using ev_flags and the EVLIST_*-macros, while it is
1551maintained by libev, it does not work exactly the same way as in libevent (consider
1552it a private API).</dt>
1553 <dt>* Priorities are not currently supported. Initialising priorities
1554will fail and all watchers will have the same priority, even though there
1555is an ev_pri field.</dt>
1556 <dt>* Other members are not supported.</dt>
1557 <dt>* The libev emulation is <i>not</i> ABI compatible to libevent, you need
1558to use the libev header file and library.</dt>
1559</dl>
1560
1561</div>
1562<h1 id="C_SUPPORT">C++ SUPPORT</h1><p><a href="#TOP" class="toplink">Top</a></p>
1563<div id="C_SUPPORT_CONTENT">
1564<p>Libev comes with some simplistic wrapper classes for C++ that mainly allow
1565you to use some convinience methods to start/stop watchers and also change
1566the callback model to a model using method callbacks on objects.</p>
1567<p>To use it,</p>
1568<pre> #include &lt;ev++.h&gt;
1569
1570</pre>
1571<p>(it is not installed by default). This automatically includes <cite>ev.h</cite>
1572and puts all of its definitions (many of them macros) into the global
1573namespace. All C++ specific things are put into the <code>ev</code> namespace.</p>
1574<p>It should support all the same embedding options as <cite>ev.h</cite>, most notably
1575<code>EV_MULTIPLICITY</code>.</p>
1576<p>Here is a list of things available in the <code>ev</code> namespace:</p>
1577<dl>
1578 <dt><code>ev::READ</code>, <code>ev::WRITE</code> etc.</dt>
1579 <dd>
1580 <p>These are just enum values with the same values as the <code>EV_READ</code> etc.
1581macros from <cite>ev.h</cite>.</p>
1582 </dd>
1583 <dt><code>ev::tstamp</code>, <code>ev::now</code></dt>
1584 <dd>
1585 <p>Aliases to the same types/functions as with the <code>ev_</code> prefix.</p>
1586 </dd>
1587 <dt><code>ev::io</code>, <code>ev::timer</code>, <code>ev::periodic</code>, <code>ev::idle</code>, <code>ev::sig</code> etc.</dt>
1588 <dd>
1589 <p>For each <code>ev_TYPE</code> watcher in <cite>ev.h</cite> there is a corresponding class of
1590the same name in the <code>ev</code> namespace, with the exception of <code>ev_signal</code>
1591which is called <code>ev::sig</code> to avoid clashes with the <code>signal</code> macro
1592defines by many implementations.</p>
1593 <p>All of those classes have these methods:</p>
1594 <p>
1595 <dl>
1596 <dt>ev::TYPE::TYPE (object *, object::method *)</dt>
1597 <dt>ev::TYPE::TYPE (object *, object::method *, struct ev_loop *)</dt>
1598 <dt>ev::TYPE::~TYPE</dt>
1599 <dd>
1600 <p>The constructor takes a pointer to an object and a method pointer to
1601the event handler callback to call in this class. The constructor calls
1602<code>ev_init</code> for you, which means you have to call the <code>set</code> method
1603before starting it. If you do not specify a loop then the constructor
1604automatically associates the default loop with this watcher.</p>
1605 <p>The destructor automatically stops the watcher if it is active.</p>
1606 </dd>
1607 <dt>w-&gt;set (struct ev_loop *)</dt>
1608 <dd>
1609 <p>Associates a different <code>struct ev_loop</code> with this watcher. You can only
1610do this when the watcher is inactive (and not pending either).</p>
1611 </dd>
1612 <dt>w-&gt;set ([args])</dt>
1613 <dd>
1614 <p>Basically the same as <code>ev_TYPE_set</code>, with the same args. Must be
1615called at least once. Unlike the C counterpart, an active watcher gets
1616automatically stopped and restarted.</p>
1617 </dd>
1618 <dt>w-&gt;start ()</dt>
1619 <dd>
1620 <p>Starts the watcher. Note that there is no <code>loop</code> argument as the
1621constructor already takes the loop.</p>
1622 </dd>
1623 <dt>w-&gt;stop ()</dt>
1624 <dd>
1625 <p>Stops the watcher if it is active. Again, no <code>loop</code> argument.</p>
1626 </dd>
1627 <dt>w-&gt;again () <code>ev::timer</code>, <code>ev::periodic</code> only</dt>
1628 <dd>
1629 <p>For <code>ev::timer</code> and <code>ev::periodic</code>, this invokes the corresponding
1630<code>ev_TYPE_again</code> function.</p>
1631 </dd>
1632 <dt>w-&gt;sweep () <code>ev::embed</code> only</dt>
1633 <dd>
1634 <p>Invokes <code>ev_embed_sweep</code>.</p>
1635 </dd>
1636 <dt>w-&gt;update () <code>ev::stat</code> only</dt>
1637 <dd>
1638 <p>Invokes <code>ev_stat_stat</code>.</p>
1639 </dd>
1640 </dl>
1641 </p>
1642 </dd>
1643</dl>
1644<p>Example: Define a class with an IO and idle watcher, start one of them in
1645the constructor.</p>
1646<pre> class myclass
1647 {
1648 ev_io io; void io_cb (ev::io &amp;w, int revents);
1649 ev_idle idle void idle_cb (ev::idle &amp;w, int revents);
1650
1651 myclass ();
1652 }
1653
1654 myclass::myclass (int fd)
1655 : io (this, &amp;myclass::io_cb),
1656 idle (this, &amp;myclass::idle_cb)
1657 {
1658 io.start (fd, ev::READ);
1659 }
1660
1661</pre>
1662
1663</div>
1664<h1 id="EMBEDDING">EMBEDDING</h1><p><a href="#TOP" class="toplink">Top</a></p>
1665<div id="EMBEDDING_CONTENT">
1666<p>Libev can (and often is) directly embedded into host
1667applications. Examples of applications that embed it include the Deliantra
1668Game Server, the EV perl module, the GNU Virtual Private Ethernet (gvpe)
1669and rxvt-unicode.</p>
1670<p>The goal is to enable you to just copy the neecssary files into your
1671source directory without having to change even a single line in them, so
1672you can easily upgrade by simply copying (or having a checked-out copy of
1673libev somewhere in your source tree).</p>
1674
1675</div>
1676<h2 id="FILESETS">FILESETS</h2>
1677<div id="FILESETS_CONTENT">
1678<p>Depending on what features you need you need to include one or more sets of files
1679in your app.</p>
1680
1681</div>
1682<h3 id="CORE_EVENT_LOOP">CORE EVENT LOOP</h3>
1683<div id="CORE_EVENT_LOOP_CONTENT">
1684<p>To include only the libev core (all the <code>ev_*</code> functions), with manual
1685configuration (no autoconf):</p>
1686<pre> #define EV_STANDALONE 1
1687 #include &quot;ev.c&quot;
1688
1689</pre>
1690<p>This will automatically include <cite>ev.h</cite>, too, and should be done in a
1691single C source file only to provide the function implementations. To use
1692it, do the same for <cite>ev.h</cite> in all files wishing to use this API (best
1693done by writing a wrapper around <cite>ev.h</cite> that you can include instead and
1694where you can put other configuration options):</p>
1695<pre> #define EV_STANDALONE 1
1696 #include &quot;ev.h&quot;
1697
1698</pre>
1699<p>Both header files and implementation files can be compiled with a C++
1700compiler (at least, thats a stated goal, and breakage will be treated
1701as a bug).</p>
1702<p>You need the following files in your source tree, or in a directory
1703in your include path (e.g. in libev/ when using -Ilibev):</p>
1704<pre> ev.h
1705 ev.c
1706 ev_vars.h
1707 ev_wrap.h
1708
1709 ev_win32.c required on win32 platforms only
1710
1711 ev_select.c only when select backend is enabled (which is by default)
1712 ev_poll.c only when poll backend is enabled (disabled by default)
1713 ev_epoll.c only when the epoll backend is enabled (disabled by default)
1714 ev_kqueue.c only when the kqueue backend is enabled (disabled by default)
1715 ev_port.c only when the solaris port backend is enabled (disabled by default)
1716
1717</pre>
1718<p><cite>ev.c</cite> includes the backend files directly when enabled, so you only need
1719to compile this single file.</p>
1720
1721</div>
1722<h3 id="LIBEVENT_COMPATIBILITY_API">LIBEVENT COMPATIBILITY API</h3>
1723<div id="LIBEVENT_COMPATIBILITY_API_CONTENT">
1724<p>To include the libevent compatibility API, also include:</p>
1725<pre> #include &quot;event.c&quot;
1726
1727</pre>
1728<p>in the file including <cite>ev.c</cite>, and:</p>
1729<pre> #include &quot;event.h&quot;
1730
1731</pre>
1732<p>in the files that want to use the libevent API. This also includes <cite>ev.h</cite>.</p>
1733<p>You need the following additional files for this:</p>
1734<pre> event.h
1735 event.c
1736
1737</pre>
1738
1739</div>
1740<h3 id="AUTOCONF_SUPPORT">AUTOCONF SUPPORT</h3>
1741<div id="AUTOCONF_SUPPORT_CONTENT">
1742<p>Instead of using <code>EV_STANDALONE=1</code> and providing your config in
1743whatever way you want, you can also <code>m4_include([libev.m4])</code> in your
1744<cite>configure.ac</cite> and leave <code>EV_STANDALONE</code> undefined. <cite>ev.c</cite> will then
1745include <cite>config.h</cite> and configure itself accordingly.</p>
1746<p>For this of course you need the m4 file:</p>
1747<pre> libev.m4
1748
1749</pre>
1750
1751</div>
1752<h2 id="PREPROCESSOR_SYMBOLS_MACROS">PREPROCESSOR SYMBOLS/MACROS</h2>
1753<div id="PREPROCESSOR_SYMBOLS_MACROS_CONTENT">
1754<p>Libev can be configured via a variety of preprocessor symbols you have to define
1755before including any of its files. The default is not to build for multiplicity
1756and only include the select backend.</p>
1757<dl>
1758 <dt>EV_STANDALONE</dt>
1759 <dd>
1760 <p>Must always be <code>1</code> if you do not use autoconf configuration, which
1761keeps libev from including <cite>config.h</cite>, and it also defines dummy
1762implementations for some libevent functions (such as logging, which is not
1763supported). It will also not define any of the structs usually found in
1764<cite>event.h</cite> that are not directly supported by the libev core alone.</p>
1765 </dd>
1766 <dt>EV_USE_MONOTONIC</dt>
1767 <dd>
1768 <p>If defined to be <code>1</code>, libev will try to detect the availability of the
1769monotonic clock option at both compiletime and runtime. Otherwise no use
1770of the monotonic clock option will be attempted. If you enable this, you
1771usually have to link against librt or something similar. Enabling it when
1772the functionality isn't available is safe, though, althoguh you have
1773to make sure you link against any libraries where the <code>clock_gettime</code>
1774function is hiding in (often <cite>-lrt</cite>).</p>
1775 </dd>
1776 <dt>EV_USE_REALTIME</dt>
1777 <dd>
1778 <p>If defined to be <code>1</code>, libev will try to detect the availability of the
1779realtime clock option at compiletime (and assume its availability at
1780runtime if successful). Otherwise no use of the realtime clock option will
1781be attempted. This effectively replaces <code>gettimeofday</code> by <code>clock_get
1782(CLOCK_REALTIME, ...)</code> and will not normally affect correctness. See tzhe note about libraries
1783in the description of <code>EV_USE_MONOTONIC</code>, though.</p>
1784 </dd>
1785 <dt>EV_USE_SELECT</dt>
1786 <dd>
1787 <p>If undefined or defined to be <code>1</code>, libev will compile in support for the
1788<code>select</code>(2) backend. No attempt at autodetection will be done: if no
1789other method takes over, select will be it. Otherwise the select backend
1790will not be compiled in.</p>
1791 </dd>
1792 <dt>EV_SELECT_USE_FD_SET</dt>
1793 <dd>
1794 <p>If defined to <code>1</code>, then the select backend will use the system <code>fd_set</code>
1795structure. This is useful if libev doesn't compile due to a missing
1796<code>NFDBITS</code> or <code>fd_mask</code> definition or it misguesses the bitset layout on
1797exotic systems. This usually limits the range of file descriptors to some
1798low limit such as 1024 or might have other limitations (winsocket only
1799allows 64 sockets). The <code>FD_SETSIZE</code> macro, set before compilation, might
1800influence the size of the <code>fd_set</code> used.</p>
1801 </dd>
1802 <dt>EV_SELECT_IS_WINSOCKET</dt>
1803 <dd>
1804 <p>When defined to <code>1</code>, the select backend will assume that
1805select/socket/connect etc. don't understand file descriptors but
1806wants osf handles on win32 (this is the case when the select to
1807be used is the winsock select). This means that it will call
1808<code>_get_osfhandle</code> on the fd to convert it to an OS handle. Otherwise,
1809it is assumed that all these functions actually work on fds, even
1810on win32. Should not be defined on non-win32 platforms.</p>
1811 </dd>
1812 <dt>EV_USE_POLL</dt>
1813 <dd>
1814 <p>If defined to be <code>1</code>, libev will compile in support for the <code>poll</code>(2)
1815backend. Otherwise it will be enabled on non-win32 platforms. It
1816takes precedence over select.</p>
1817 </dd>
1818 <dt>EV_USE_EPOLL</dt>
1819 <dd>
1820 <p>If defined to be <code>1</code>, libev will compile in support for the Linux
1821<code>epoll</code>(7) backend. Its availability will be detected at runtime,
1822otherwise another method will be used as fallback. This is the
1823preferred backend for GNU/Linux systems.</p>
1824 </dd>
1825 <dt>EV_USE_KQUEUE</dt>
1826 <dd>
1827 <p>If defined to be <code>1</code>, libev will compile in support for the BSD style
1828<code>kqueue</code>(2) backend. Its actual availability will be detected at runtime,
1829otherwise another method will be used as fallback. This is the preferred
1830backend for BSD and BSD-like systems, although on most BSDs kqueue only
1831supports some types of fds correctly (the only platform we found that
1832supports ptys for example was NetBSD), so kqueue might be compiled in, but
1833not be used unless explicitly requested. The best way to use it is to find
1834out whether kqueue supports your type of fd properly and use an embedded
1835kqueue loop.</p>
1836 </dd>
1837 <dt>EV_USE_PORT</dt>
1838 <dd>
1839 <p>If defined to be <code>1</code>, libev will compile in support for the Solaris
184010 port style backend. Its availability will be detected at runtime,
1841otherwise another method will be used as fallback. This is the preferred
1842backend for Solaris 10 systems.</p>
1843 </dd>
1844 <dt>EV_USE_DEVPOLL</dt>
1845 <dd>
1846 <p>reserved for future expansion, works like the USE symbols above.</p>
1847 </dd>
1848 <dt>EV_H</dt>
1849 <dd>
1850 <p>The name of the <cite>ev.h</cite> header file used to include it. The default if
1851undefined 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
1852can be used to virtually rename the <cite>ev.h</cite> header file in case of conflicts.</p>
1853 </dd>
1854 <dt>EV_CONFIG_H</dt>
1855 <dd>
1856 <p>If <code>EV_STANDALONE</code> isn't <code>1</code>, this variable can be used to override
1857<cite>ev.c</cite>'s idea of where to find the <cite>config.h</cite> file, similarly to
1858<code>EV_H</code>, above.</p>
1859 </dd>
1860 <dt>EV_EVENT_H</dt>
1861 <dd>
1862 <p>Similarly to <code>EV_H</code>, this macro can be used to override <cite>event.c</cite>'s idea
1863of how the <cite>event.h</cite> header can be found.</p>
1864 </dd>
1865 <dt>EV_PROTOTYPES</dt>
1866 <dd>
1867 <p>If defined to be <code>0</code>, then <cite>ev.h</cite> will not define any function
1868prototypes, but still define all the structs and other symbols. This is
1869occasionally useful if you want to provide your own wrapper functions
1870around libev functions.</p>
1871 </dd>
1872 <dt>EV_MULTIPLICITY</dt>
1873 <dd>
1874 <p>If undefined or defined to <code>1</code>, then all event-loop-specific functions
1875will have the <code>struct ev_loop *</code> as first argument, and you can create
1876additional independent event loops. Otherwise there will be no support
1877for multiple event loops and there is no first event loop pointer
1878argument. Instead, all functions act on the single default loop.</p>
1879 </dd>
1880 <dt>EV_PERIODIC_ENABLE</dt>
1881 <dd>
1882 <p>If undefined or defined to be <code>1</code>, then periodic timers are supported. If
1883defined to be <code>0</code>, then they are not. Disabling them saves a few kB of
1884code.</p>
1885 </dd>
1886 <dt>EV_EMBED_ENABLE</dt>
1887 <dd>
1888 <p>If undefined or defined to be <code>1</code>, then embed watchers are supported. If
1889defined to be <code>0</code>, then they are not.</p>
1890 </dd>
1891 <dt>EV_STAT_ENABLE</dt>
1892 <dd>
1893 <p>If undefined or defined to be <code>1</code>, then stat watchers are supported. If
1894defined to be <code>0</code>, then they are not.</p>
1895 </dd>
1896 <dt>EV_MINIMAL</dt>
1897 <dd>
1898 <p>If you need to shave off some kilobytes of code at the expense of some
1899speed, define this symbol to <code>1</code>. Currently only used for gcc to override
1900some inlining decisions, saves roughly 30% codesize of amd64.</p>
1901 </dd>
1902 <dt>EV_COMMON</dt>
1903 <dd>
1904 <p>By default, all watchers have a <code>void *data</code> member. By redefining
1905this macro to a something else you can include more and other types of
1906members. You have to define it each time you include one of the files,
1907though, and it must be identical each time.</p>
1908 <p>For example, the perl EV module uses something like this:</p>
1909<pre> #define EV_COMMON \
1910 SV *self; /* contains this struct */ \
1911 SV *cb_sv, *fh /* note no trailing &quot;;&quot; */
1912
1913</pre>
1914 </dd>
1915 <dt>EV_CB_DECLARE (type)</dt>
1916 <dt>EV_CB_INVOKE (watcher, revents)</dt>
1917 <dt>ev_set_cb (ev, cb)</dt>
1918 <dd>
1919 <p>Can be used to change the callback member declaration in each watcher,
1920and the way callbacks are invoked and set. Must expand to a struct member
1921definition and a statement, respectively. See the <cite>ev.v</cite> header file for
1922their default definitions. One possible use for overriding these is to
1923avoid the <code>struct ev_loop *</code> as first argument in all cases, or to use
1924method calls instead of plain function calls in C++.</p>
1925
1926</div>
1927<h2 id="EXAMPLES">EXAMPLES</h2>
1928<div id="EXAMPLES_CONTENT">
1929 <p>For a real-world example of a program the includes libev
1930verbatim, you can have a look at the EV perl module
1931(<a href="http://software.schmorp.de/pkg/EV.html">http://software.schmorp.de/pkg/EV.html</a>). It has the libev files in
1932the <cite>libev/</cite> subdirectory and includes them in the <cite>EV/EVAPI.h</cite> (public
1933interface) and <cite>EV.xs</cite> (implementation) files. Only the <cite>EV.xs</cite> file
1934will be compiled. It is pretty complex because it provides its own header
1935file.</p>
1936 <p>The usage in rxvt-unicode is simpler. It has a <cite>ev_cpp.h</cite> header file
1937that everybody includes and which overrides some autoconf choices:</p>
1938<pre> #define EV_USE_POLL 0
1939 #define EV_MULTIPLICITY 0
1940 #define EV_PERIODICS 0
1941 #define EV_CONFIG_H &lt;config.h&gt;
1942
1943 #include &quot;ev++.h&quot;
1944
1945</pre>
1946 <p>And a <cite>ev_cpp.C</cite> implementation file that contains libev proper and is compiled:</p>
1947<pre> #include &quot;ev_cpp.h&quot;
1948 #include &quot;ev.c&quot;
1949
1950
1951
1952
1953</pre>
1954
1955</div>
1956<h1 id="COMPLEXITIES">COMPLEXITIES</h1><p><a href="#TOP" class="toplink">Top</a></p>
1957<div id="COMPLEXITIES_CONTENT">
1958 <p>In this section the complexities of (many of) the algorithms used inside
1959libev will be explained. For complexity discussions about backends see the
1960documentation for <code>ev_default_init</code>.</p>
1961 <p>
1962 <dl>
1963 <dt>Starting and stopping timer/periodic watchers: O(log skipped_other_timers)</dt>
1964 <dt>Changing timer/periodic watchers (by autorepeat, again): O(log skipped_other_timers)</dt>
1965 <dt>Starting io/check/prepare/idle/signal/child watchers: O(1)</dt>
1966 <dt>Stopping check/prepare/idle watchers: O(1)</dt>
1967 <dt>Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % 16))</dt>
1968 <dt>Finding the next timer per loop iteration: O(1)</dt>
1969 <dt>Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)</dt>
1970 <dt>Activating one watcher: O(1)</dt>
1971 </dl>
1972 </p>
1973
1974
1975
1976
740 1977
741</div> 1978</div>
742<h1 id="AUTHOR">AUTHOR</h1><p><a href="#TOP" class="toplink">Top</a></p> 1979<h1 id="AUTHOR">AUTHOR</h1><p><a href="#TOP" class="toplink">Top</a></p>
743<div id="AUTHOR_CONTENT"> 1980<div id="AUTHOR_CONTENT">
744<p>Marc Lehmann &lt;libev@schmorp.de&gt;.</p> 1981 <p>Marc Lehmann &lt;libev@schmorp.de&gt;.</p>
745 1982
746</div> 1983</div>
747</div></body> 1984</div></body>
748</html> 1985</html>

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines