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

Comparing libev/ev.html (file contents):
Revision 1.18 by root, Mon Nov 12 08:57:03 2007 UTC vs.
Revision 1.75 by root, Sun Dec 9 19:46:56 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 09:57:02 2007" /> 9 <meta name="created" content="Sun Dec 9 20:45:16 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 -->
15<h3 id="TOP">Index</h3> 15<h3 id="TOP">Index</h3>
16 16
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="#EXAMPLE_PROGRAM">EXAMPLE PROGRAM</a></li>
19<li><a href="#DESCRIPTION">DESCRIPTION</a></li> 20<li><a href="#DESCRIPTION">DESCRIPTION</a></li>
20<li><a href="#FEATURES">FEATURES</a></li> 21<li><a href="#FEATURES">FEATURES</a></li>
21<li><a href="#CONVENTIONS">CONVENTIONS</a></li> 22<li><a href="#CONVENTIONS">CONVENTIONS</a></li>
22<li><a href="#TIME_REPRESENTATION">TIME REPRESENTATION</a></li> 23<li><a href="#TIME_REPRESENTATION">TIME REPRESENTATION</a></li>
23<li><a href="#GLOBAL_FUNCTIONS">GLOBAL FUNCTIONS</a></li> 24<li><a href="#GLOBAL_FUNCTIONS">GLOBAL FUNCTIONS</a></li>
24<li><a href="#FUNCTIONS_CONTROLLING_THE_EVENT_LOOP">FUNCTIONS CONTROLLING THE EVENT LOOP</a></li> 25<li><a href="#FUNCTIONS_CONTROLLING_THE_EVENT_LOOP">FUNCTIONS CONTROLLING THE EVENT LOOP</a></li>
25<li><a href="#ANATOMY_OF_A_WATCHER">ANATOMY OF A WATCHER</a> 26<li><a href="#ANATOMY_OF_A_WATCHER">ANATOMY OF A WATCHER</a>
27<ul><li><a href="#GENERIC_WATCHER_FUNCTIONS">GENERIC WATCHER FUNCTIONS</a></li>
26<ul><li><a href="#ASSOCIATING_CUSTOM_DATA_WITH_A_WATCH">ASSOCIATING CUSTOM DATA WITH A WATCHER</a></li> 28<li><a href="#ASSOCIATING_CUSTOM_DATA_WITH_A_WATCH">ASSOCIATING CUSTOM DATA WITH A WATCHER</a></li>
27</ul> 29</ul>
28</li> 30</li>
29<li><a href="#WATCHER_TYPES">WATCHER TYPES</a> 31<li><a href="#WATCHER_TYPES">WATCHER TYPES</a>
30<ul><li><a href="#code_ev_io_code_is_this_file_descrip"><code>ev_io</code> - is this file descriptor readable or writable</a></li> 32<ul><li><a href="#code_ev_io_code_is_this_file_descrip"><code>ev_io</code> - is this file descriptor readable or writable?</a></li>
31<li><a href="#code_ev_timer_code_relative_and_opti"><code>ev_timer</code> - relative and optionally recurring timeouts</a></li> 33<li><a href="#code_ev_timer_code_relative_and_opti"><code>ev_timer</code> - relative and optionally repeating timeouts</a></li>
32<li><a href="#code_ev_periodic_code_to_cron_or_not"><code>ev_periodic</code> - to cron or not to cron</a></li> 34<li><a href="#code_ev_periodic_code_to_cron_or_not"><code>ev_periodic</code> - to cron or not to cron?</a></li>
33<li><a href="#code_ev_signal_code_signal_me_when_a"><code>ev_signal</code> - signal me when a signal gets signalled</a></li> 35<li><a href="#code_ev_signal_code_signal_me_when_a"><code>ev_signal</code> - signal me when a signal gets signalled!</a></li>
34<li><a href="#code_ev_child_code_wait_for_pid_stat"><code>ev_child</code> - wait for pid status changes</a></li> 36<li><a href="#code_ev_child_code_watch_out_for_pro"><code>ev_child</code> - watch out for process status changes</a></li>
37<li><a href="#code_ev_stat_code_did_the_file_attri"><code>ev_stat</code> - did the file attributes just change?</a></li>
35<li><a href="#code_ev_idle_code_when_you_ve_got_no"><code>ev_idle</code> - when you've got nothing better to do</a></li> 38<li><a href="#code_ev_idle_code_when_you_ve_got_no"><code>ev_idle</code> - when you've got nothing better to do...</a></li>
36<li><a href="#code_ev_prepare_code_and_code_ev_che"><code>ev_prepare</code> and <code>ev_check</code> - customise your event loop</a></li> 39<li><a href="#code_ev_prepare_code_and_code_ev_che"><code>ev_prepare</code> and <code>ev_check</code> - customise your event loop!</a></li>
40<li><a href="#code_ev_embed_code_when_one_backend_"><code>ev_embed</code> - when one backend isn't enough...</a></li>
41<li><a href="#code_ev_fork_code_the_audacity_to_re"><code>ev_fork</code> - the audacity to resume the event loop after a fork</a></li>
37</ul> 42</ul>
38</li> 43</li>
39<li><a href="#OTHER_FUNCTIONS">OTHER FUNCTIONS</a></li> 44<li><a href="#OTHER_FUNCTIONS">OTHER FUNCTIONS</a></li>
45<li><a href="#LIBEVENT_EMULATION">LIBEVENT EMULATION</a></li>
46<li><a href="#C_SUPPORT">C++ SUPPORT</a></li>
47<li><a href="#MACRO_MAGIC">MACRO MAGIC</a></li>
48<li><a href="#EMBEDDING">EMBEDDING</a>
49<ul><li><a href="#FILESETS">FILESETS</a>
50<ul><li><a href="#CORE_EVENT_LOOP">CORE EVENT LOOP</a></li>
51<li><a href="#LIBEVENT_COMPATIBILITY_API">LIBEVENT COMPATIBILITY API</a></li>
52<li><a href="#AUTOCONF_SUPPORT">AUTOCONF SUPPORT</a></li>
53</ul>
54</li>
55<li><a href="#PREPROCESSOR_SYMBOLS_MACROS">PREPROCESSOR SYMBOLS/MACROS</a></li>
56<li><a href="#EXAMPLES">EXAMPLES</a></li>
57</ul>
58</li>
59<li><a href="#COMPLEXITIES">COMPLEXITIES</a></li>
40<li><a href="#AUTHOR">AUTHOR</a> 60<li><a href="#AUTHOR">AUTHOR</a>
41</li> 61</li>
42</ul><hr /> 62</ul><hr />
43<!-- INDEX END --> 63<!-- INDEX END -->
44 64
45<h1 id="NAME">NAME</h1><p><a href="#TOP" class="toplink">Top</a></p> 65<h1 id="NAME">NAME</h1>
46<div id="NAME_CONTENT"> 66<div id="NAME_CONTENT">
47<p>libev - a high performance full-featured event loop written in C</p> 67<p>libev - a high performance full-featured event loop written in C</p>
48 68
49</div> 69</div>
50<h1 id="SYNOPSIS">SYNOPSIS</h1><p><a href="#TOP" class="toplink">Top</a></p> 70<h1 id="SYNOPSIS">SYNOPSIS</h1>
51<div id="SYNOPSIS_CONTENT"> 71<div id="SYNOPSIS_CONTENT">
52<pre> #include &lt;ev.h&gt; 72<pre> #include &lt;ev.h&gt;
53 73
54</pre> 74</pre>
55 75
56</div> 76</div>
57<h1 id="DESCRIPTION">DESCRIPTION</h1><p><a href="#TOP" class="toplink">Top</a></p> 77<h1 id="EXAMPLE_PROGRAM">EXAMPLE PROGRAM</h1>
78<div id="EXAMPLE_PROGRAM_CONTENT">
79<pre> #include &lt;ev.h&gt;
80
81 ev_io stdin_watcher;
82 ev_timer timeout_watcher;
83
84 /* called when data readable on stdin */
85 static void
86 stdin_cb (EV_P_ struct ev_io *w, int revents)
87 {
88 /* puts (&quot;stdin ready&quot;); */
89 ev_io_stop (EV_A_ w); /* just a syntax example */
90 ev_unloop (EV_A_ EVUNLOOP_ALL); /* leave all loop calls */
91 }
92
93 static void
94 timeout_cb (EV_P_ struct ev_timer *w, int revents)
95 {
96 /* puts (&quot;timeout&quot;); */
97 ev_unloop (EV_A_ EVUNLOOP_ONE); /* leave one loop call */
98 }
99
100 int
101 main (void)
102 {
103 struct ev_loop *loop = ev_default_loop (0);
104
105 /* initialise an io watcher, then start it */
106 ev_io_init (&amp;stdin_watcher, stdin_cb, /*STDIN_FILENO*/ 0, EV_READ);
107 ev_io_start (loop, &amp;stdin_watcher);
108
109 /* simple non-repeating 5.5 second timeout */
110 ev_timer_init (&amp;timeout_watcher, timeout_cb, 5.5, 0.);
111 ev_timer_start (loop, &amp;timeout_watcher);
112
113 /* loop till timeout or data ready */
114 ev_loop (loop, 0);
115
116 return 0;
117 }
118
119</pre>
120
121</div>
122<h1 id="DESCRIPTION">DESCRIPTION</h1>
58<div id="DESCRIPTION_CONTENT"> 123<div id="DESCRIPTION_CONTENT">
124<p>The newest version of this document is also available as a html-formatted
125web page you might find easier to navigate when reading it for the first
126time: <a href="http://cvs.schmorp.de/libev/ev.html">http://cvs.schmorp.de/libev/ev.html</a>.</p>
59<p>Libev is an event loop: you register interest in certain events (such as a 127<p>Libev is an event loop: you register interest in certain events (such as a
60file descriptor being readable or a timeout occuring), and it will manage 128file descriptor being readable or a timeout occuring), and it will manage
61these event sources and provide your program with events.</p> 129these event sources and provide your program with events.</p>
62<p>To do this, it must take more or less complete control over your process 130<p>To do this, it must take more or less complete control over your process
63(or thread) by executing the <i>event loop</i> handler, and will then 131(or thread) by executing the <i>event loop</i> handler, and will then
66watchers</i>, which are relatively small C structures you initialise with the 134watchers</i>, which are relatively small C structures you initialise with the
67details of the event, and then hand it over to libev by <i>starting</i> the 135details of the event, and then hand it over to libev by <i>starting</i> the
68watcher.</p> 136watcher.</p>
69 137
70</div> 138</div>
71<h1 id="FEATURES">FEATURES</h1><p><a href="#TOP" class="toplink">Top</a></p> 139<h1 id="FEATURES">FEATURES</h1>
72<div id="FEATURES_CONTENT"> 140<div id="FEATURES_CONTENT">
73<p>Libev supports select, poll, the linux-specific epoll and the bsd-specific 141<p>Libev supports <code>select</code>, <code>poll</code>, the Linux-specific <code>epoll</code>, the
74kqueue mechanisms for file descriptor events, relative timers, absolute 142BSD-specific <code>kqueue</code> and the Solaris-specific event port mechanisms
75timers with customised rescheduling, signal events, process status change 143for file descriptor events (<code>ev_io</code>), the Linux <code>inotify</code> interface
76events (related to SIGCHLD), and event watchers dealing with the event 144(for <code>ev_stat</code>), relative timers (<code>ev_timer</code>), absolute timers
77loop mechanism itself (idle, prepare and check watchers). It also is quite 145with customised rescheduling (<code>ev_periodic</code>), synchronous signals
146(<code>ev_signal</code>), process status change events (<code>ev_child</code>), and event
147watchers dealing with the event loop mechanism itself (<code>ev_idle</code>,
148<code>ev_embed</code>, <code>ev_prepare</code> and <code>ev_check</code> watchers) as well as
149file watchers (<code>ev_stat</code>) and even limited support for fork events
150(<code>ev_fork</code>).</p>
151<p>It also is quite fast (see this
78fast (see this <a href="http://libev.schmorp.de/bench.html">benchmark</a> comparing 152<a href="http://libev.schmorp.de/bench.html">benchmark</a> comparing it to libevent
79it to libevent for example).</p> 153for example).</p>
80 154
81</div> 155</div>
82<h1 id="CONVENTIONS">CONVENTIONS</h1><p><a href="#TOP" class="toplink">Top</a></p> 156<h1 id="CONVENTIONS">CONVENTIONS</h1>
83<div id="CONVENTIONS_CONTENT"> 157<div id="CONVENTIONS_CONTENT">
84<p>Libev is very configurable. In this manual the default configuration 158<p>Libev is very configurable. In this manual the default configuration will
85will be described, which supports multiple event loops. For more info 159be described, which supports multiple event loops. For more info about
86about various configuration options please have a look at the file 160various configuration options please have a look at <strong>EMBED</strong> section in
87<cite>README.embed</cite> in the libev distribution. If libev was configured without 161this manual. If libev was configured without support for multiple event
88support for multiple event loops, then all functions taking an initial 162loops, then all functions taking an initial argument of name <code>loop</code>
89argument of name <code>loop</code> (which is always of type <code>struct ev_loop *</code>) 163(which is always of type <code>struct ev_loop *</code>) will not have this argument.</p>
90will not have this argument.</p>
91 164
92</div> 165</div>
93<h1 id="TIME_REPRESENTATION">TIME REPRESENTATION</h1><p><a href="#TOP" class="toplink">Top</a></p> 166<h1 id="TIME_REPRESENTATION">TIME REPRESENTATION</h1>
94<div id="TIME_REPRESENTATION_CONTENT"> 167<div id="TIME_REPRESENTATION_CONTENT">
95<p>Libev represents time as a single floating point number, representing the 168<p>Libev represents time as a single floating point number, representing the
96(fractional) number of seconds since the (POSIX) epoch (somewhere near 169(fractional) number of seconds since the (POSIX) epoch (somewhere near
97the beginning of 1970, details are complicated, don't ask). This type is 170the beginning of 1970, details are complicated, don't ask). This type is
98called <code>ev_tstamp</code>, which is what you should use too. It usually aliases 171called <code>ev_tstamp</code>, which is what you should use too. It usually aliases
99to the double type in C.</p> 172to the <code>double</code> type in C, and when you need to do any calculations on
173it, you should treat it as such.</p>
100 174
101</div> 175</div>
102<h1 id="GLOBAL_FUNCTIONS">GLOBAL FUNCTIONS</h1><p><a href="#TOP" class="toplink">Top</a></p> 176<h1 id="GLOBAL_FUNCTIONS">GLOBAL FUNCTIONS</h1>
103<div id="GLOBAL_FUNCTIONS_CONTENT"> 177<div id="GLOBAL_FUNCTIONS_CONTENT">
178<p>These functions can be called anytime, even before initialising the
179library in any way.</p>
104<dl> 180<dl>
105 <dt>ev_tstamp ev_time ()</dt> 181 <dt>ev_tstamp ev_time ()</dt>
106 <dd> 182 <dd>
107 <p>Returns the current time as libev would use it.</p> 183 <p>Returns the current time as libev would use it. Please note that the
184<code>ev_now</code> function is usually faster and also often returns the timestamp
185you actually want to know.</p>
108 </dd> 186 </dd>
109 <dt>int ev_version_major ()</dt> 187 <dt>int ev_version_major ()</dt>
110 <dt>int ev_version_minor ()</dt> 188 <dt>int ev_version_minor ()</dt>
111 <dd> 189 <dd>
112 <p>You can find out the major and minor version numbers of the library 190 <p>You can find out the major and minor API/ABI version numbers of the library
113you linked against by calling the functions <code>ev_version_major</code> and 191you linked against by calling the functions <code>ev_version_major</code> and
114<code>ev_version_minor</code>. If you want, you can compare against the global 192<code>ev_version_minor</code>. If you want, you can compare against the global
115symbols <code>EV_VERSION_MAJOR</code> and <code>EV_VERSION_MINOR</code>, which specify the 193symbols <code>EV_VERSION_MAJOR</code> and <code>EV_VERSION_MINOR</code>, which specify the
116version of the library your program was compiled against.</p> 194version of the library your program was compiled against.</p>
195 <p>These version numbers refer to the API and ABI version of the library, not
196the release version.</p>
117 <p>Usually, it's a good idea to terminate if the major versions mismatch, 197 <p>Usually, it's a good idea to terminate if the major versions mismatch,
118as this indicates an incompatible change. Minor versions are usually 198as this indicates an incompatible change. Minor versions are usually
119compatible to older versions, so a larger minor version alone is usually 199compatible to older versions, so a larger minor version alone is usually
120not a problem.</p> 200not a problem.</p>
201 <p>Example: Make sure we haven't accidentally been linked against the wrong
202version.</p>
203<pre> assert ((&quot;libev version mismatch&quot;,
204 ev_version_major () == EV_VERSION_MAJOR
205 &amp;&amp; ev_version_minor () &gt;= EV_VERSION_MINOR));
206
207</pre>
208 </dd>
209 <dt>unsigned int ev_supported_backends ()</dt>
210 <dd>
211 <p>Return the set of all backends (i.e. their corresponding <code>EV_BACKEND_*</code>
212value) compiled into this binary of libev (independent of their
213availability on the system you are running on). See <code>ev_default_loop</code> for
214a description of the set values.</p>
215 <p>Example: make sure we have the epoll method, because yeah this is cool and
216a must have and can we have a torrent of it please!!!11</p>
217<pre> assert ((&quot;sorry, no epoll, no sex&quot;,
218 ev_supported_backends () &amp; EVBACKEND_EPOLL));
219
220</pre>
221 </dd>
222 <dt>unsigned int ev_recommended_backends ()</dt>
223 <dd>
224 <p>Return the set of all backends compiled into this binary of libev and also
225recommended for this platform. This set is often smaller than the one
226returned by <code>ev_supported_backends</code>, as for example kqueue is broken on
227most BSDs and will not be autodetected unless you explicitly request it
228(assuming you know what you are doing). This is the set of backends that
229libev will probe for if you specify no backends explicitly.</p>
230 </dd>
231 <dt>unsigned int ev_embeddable_backends ()</dt>
232 <dd>
233 <p>Returns the set of backends that are embeddable in other event loops. This
234is the theoretical, all-platform, value. To find which backends
235might be supported on the current system, you would need to look at
236<code>ev_embeddable_backends () &amp; ev_supported_backends ()</code>, likewise for
237recommended ones.</p>
238 <p>See the description of <code>ev_embed</code> watchers for more info.</p>
121 </dd> 239 </dd>
122 <dt>ev_set_allocator (void *(*cb)(void *ptr, long size))</dt> 240 <dt>ev_set_allocator (void *(*cb)(void *ptr, long size))</dt>
123 <dd> 241 <dd>
124 <p>Sets the allocation function to use (the prototype is similar to the 242 <p>Sets the allocation function to use (the prototype is similar - the
125realloc C function, the semantics are identical). It is used to allocate 243semantics is identical - to the realloc C function). It is used to
126and free memory (no surprises here). If it returns zero when memory 244allocate and free memory (no surprises here). If it returns zero when
127needs to be allocated, the library might abort or take some potentially 245memory needs to be allocated, the library might abort or take some
128destructive action. The default is your system realloc function.</p> 246potentially destructive action. The default is your system realloc
247function.</p>
129 <p>You could override this function in high-availability programs to, say, 248 <p>You could override this function in high-availability programs to, say,
130free some memory if it cannot allocate memory, to use a special allocator, 249free some memory if it cannot allocate memory, to use a special allocator,
131or even to sleep a while and retry until some memory is available.</p> 250or even to sleep a while and retry until some memory is available.</p>
251 <p>Example: Replace the libev allocator with one that waits a bit and then
252retries).</p>
253<pre> static void *
254 persistent_realloc (void *ptr, size_t size)
255 {
256 for (;;)
257 {
258 void *newptr = realloc (ptr, size);
259
260 if (newptr)
261 return newptr;
262
263 sleep (60);
264 }
265 }
266
267 ...
268 ev_set_allocator (persistent_realloc);
269
270</pre>
132 </dd> 271 </dd>
133 <dt>ev_set_syserr_cb (void (*cb)(const char *msg));</dt> 272 <dt>ev_set_syserr_cb (void (*cb)(const char *msg));</dt>
134 <dd> 273 <dd>
135 <p>Set the callback function to call on a retryable syscall error (such 274 <p>Set the callback function to call on a retryable syscall error (such
136as failed select, poll, epoll_wait). The message is a printable string 275as failed select, poll, epoll_wait). The message is a printable string
137indicating the system call or subsystem causing the problem. If this 276indicating the system call or subsystem causing the problem. If this
138callback is set, then libev will expect it to remedy the sitution, no 277callback is set, then libev will expect it to remedy the sitution, no
139matter what, when it returns. That is, libev will generally retry the 278matter what, when it returns. That is, libev will generally retry the
140requested operation, or, if the condition doesn't go away, do bad stuff 279requested operation, or, if the condition doesn't go away, do bad stuff
141(such as abort).</p> 280(such as abort).</p>
281 <p>Example: This is basically the same thing that libev does internally, too.</p>
282<pre> static void
283 fatal_error (const char *msg)
284 {
285 perror (msg);
286 abort ();
287 }
288
289 ...
290 ev_set_syserr_cb (fatal_error);
291
292</pre>
142 </dd> 293 </dd>
143</dl> 294</dl>
144 295
145</div> 296</div>
146<h1 id="FUNCTIONS_CONTROLLING_THE_EVENT_LOOP">FUNCTIONS CONTROLLING THE EVENT LOOP</h1><p><a href="#TOP" class="toplink">Top</a></p> 297<h1 id="FUNCTIONS_CONTROLLING_THE_EVENT_LOOP">FUNCTIONS CONTROLLING THE EVENT LOOP</h1>
147<div id="FUNCTIONS_CONTROLLING_THE_EVENT_LOOP-2"> 298<div id="FUNCTIONS_CONTROLLING_THE_EVENT_LOOP-2">
148<p>An event loop is described by a <code>struct ev_loop *</code>. The library knows two 299<p>An event loop is described by a <code>struct ev_loop *</code>. The library knows two
149types of such loops, the <i>default</i> loop, which supports signals and child 300types of such loops, the <i>default</i> loop, which supports signals and child
150events, and dynamically created loops which do not.</p> 301events, and dynamically created loops which do not.</p>
151<p>If you use threads, a common model is to run the default event loop 302<p>If you use threads, a common model is to run the default event loop
158 <dt>struct ev_loop *ev_default_loop (unsigned int flags)</dt> 309 <dt>struct ev_loop *ev_default_loop (unsigned int flags)</dt>
159 <dd> 310 <dd>
160 <p>This will initialise the default event loop if it hasn't been initialised 311 <p>This will initialise the default event loop if it hasn't been initialised
161yet and return it. If the default loop could not be initialised, returns 312yet and return it. If the default loop could not be initialised, returns
162false. If it already was initialised it simply returns it (and ignores the 313false. If it already was initialised it simply returns it (and ignores the
163flags).</p> 314flags. If that is troubling you, check <code>ev_backend ()</code> afterwards).</p>
164 <p>If you don't know what event loop to use, use the one returned from this 315 <p>If you don't know what event loop to use, use the one returned from this
165function.</p> 316function.</p>
166 <p>The flags argument can be used to specify special behaviour or specific 317 <p>The flags argument can be used to specify special behaviour or specific
167backends to use, and is usually specified as 0 (or EVFLAG_AUTO).</p> 318backends to use, and is usually specified as <code>0</code> (or <code>EVFLAG_AUTO</code>).</p>
168 <p>It supports the following flags:</p> 319 <p>The following flags are supported:</p>
169 <p> 320 <p>
170 <dl> 321 <dl>
171 <dt><code>EVFLAG_AUTO</code></dt> 322 <dt><code>EVFLAG_AUTO</code></dt>
172 <dd> 323 <dd>
173 <p>The default flags value. Use this if you have no clue (it's the right 324 <p>The default flags value. Use this if you have no clue (it's the right
180<code>LIBEV_FLAGS</code>. Otherwise (the default), this environment variable will 331<code>LIBEV_FLAGS</code>. Otherwise (the default), this environment variable will
181override the flags completely if it is found in the environment. This is 332override the flags completely if it is found in the environment. This is
182useful to try out specific backends to test their performance, or to work 333useful to try out specific backends to test their performance, or to work
183around bugs.</p> 334around bugs.</p>
184 </dd> 335 </dd>
185 <dt><code>EVMETHOD_SELECT</code> (portable select backend)</dt> 336 <dt><code>EVFLAG_FORKCHECK</code></dt>
186 <dt><code>EVMETHOD_POLL</code> (poll backend, available everywhere except on windows)</dt> 337 <dd>
187 <dt><code>EVMETHOD_EPOLL</code> (linux only)</dt> 338 <p>Instead of calling <code>ev_default_fork</code> or <code>ev_loop_fork</code> manually after
188 <dt><code>EVMETHOD_KQUEUE</code> (some bsds only)</dt> 339a fork, you can also make libev check for a fork in each iteration by
189 <dt><code>EVMETHOD_DEVPOLL</code> (solaris 8 only)</dt> 340enabling this flag.</p>
190 <dt><code>EVMETHOD_PORT</code> (solaris 10 only)</dt> 341 <p>This works by calling <code>getpid ()</code> on every iteration of the loop,
342and thus this might slow down your event loop if you do a lot of loop
343iterations and little real work, but is usually not noticeable (on my
344Linux system for example, <code>getpid</code> is actually a simple 5-insn sequence
345without a syscall and thus <i>very</i> fast, but my Linux system also has
346<code>pthread_atfork</code> which is even faster).</p>
347 <p>The big advantage of this flag is that you can forget about fork (and
348forget about forgetting to tell libev about forking) when you use this
349flag.</p>
350 <p>This flag setting cannot be overriden or specified in the <code>LIBEV_FLAGS</code>
351environment variable.</p>
191 <dd> 352 </dd>
192 <p>If one or more of these are ored into the flags value, then only these 353 <dt><code>EVBACKEND_SELECT</code> (value 1, portable select backend)</dt>
193backends will be tried (in the reverse order as given here). If one are 354 <dd>
194specified, any backend will do.</p> 355 <p>This is your standard select(2) backend. Not <i>completely</i> standard, as
356libev tries to roll its own fd_set with no limits on the number of fds,
357but if that fails, expect a fairly low limit on the number of fds when
358using this backend. It doesn't scale too well (O(highest_fd)), but its usually
359the fastest backend for a low number of fds.</p>
360 </dd>
361 <dt><code>EVBACKEND_POLL</code> (value 2, poll backend, available everywhere except on windows)</dt>
362 <dd>
363 <p>And this is your standard poll(2) backend. It's more complicated than
364select, but handles sparse fds better and has no artificial limit on the
365number of fds you can use (except it will slow down considerably with a
366lot of inactive fds). It scales similarly to select, i.e. O(total_fds).</p>
367 </dd>
368 <dt><code>EVBACKEND_EPOLL</code> (value 4, Linux)</dt>
369 <dd>
370 <p>For few fds, this backend is a bit little slower than poll and select,
371but it scales phenomenally better. While poll and select usually scale like
372O(total_fds) where n is the total number of fds (or the highest fd), epoll scales
373either O(1) or O(active_fds).</p>
374 <p>While stopping and starting an I/O watcher in the same iteration will
375result in some caching, there is still a syscall per such incident
376(because the fd could point to a different file description now), so its
377best to avoid that. Also, dup()ed file descriptors might not work very
378well if you register events for both fds.</p>
379 <p>Please note that epoll sometimes generates spurious notifications, so you
380need to use non-blocking I/O or other means to avoid blocking when no data
381(or space) is available.</p>
382 </dd>
383 <dt><code>EVBACKEND_KQUEUE</code> (value 8, most BSD clones)</dt>
384 <dd>
385 <p>Kqueue deserves special mention, as at the time of this writing, it
386was broken on all BSDs except NetBSD (usually it doesn't work with
387anything but sockets and pipes, except on Darwin, where of course its
388completely useless). For this reason its not being &quot;autodetected&quot;
389unless you explicitly specify it explicitly in the flags (i.e. using
390<code>EVBACKEND_KQUEUE</code>).</p>
391 <p>It scales in the same way as the epoll backend, but the interface to the
392kernel is more efficient (which says nothing about its actual speed, of
393course). While starting and stopping an I/O watcher does not cause an
394extra syscall as with epoll, it still adds up to four event changes per
395incident, so its best to avoid that.</p>
396 </dd>
397 <dt><code>EVBACKEND_DEVPOLL</code> (value 16, Solaris 8)</dt>
398 <dd>
399 <p>This is not implemented yet (and might never be).</p>
400 </dd>
401 <dt><code>EVBACKEND_PORT</code> (value 32, Solaris 10)</dt>
402 <dd>
403 <p>This uses the Solaris 10 port mechanism. As with everything on Solaris,
404it's really slow, but it still scales very well (O(active_fds)).</p>
405 <p>Please note that solaris ports can result in a lot of spurious
406notifications, so you need to use non-blocking I/O or other means to avoid
407blocking when no data (or space) is available.</p>
408 </dd>
409 <dt><code>EVBACKEND_ALL</code></dt>
410 <dd>
411 <p>Try all backends (even potentially broken ones that wouldn't be tried
412with <code>EVFLAG_AUTO</code>). Since this is a mask, you can do stuff such as
413<code>EVBACKEND_ALL &amp; ~EVBACKEND_KQUEUE</code>.</p>
195 </dd> 414 </dd>
196 </dl> 415 </dl>
197 </p> 416 </p>
417 <p>If one or more of these are ored into the flags value, then only these
418backends will be tried (in the reverse order as given here). If none are
419specified, most compiled-in backend will be tried, usually in reverse
420order of their flag values :)</p>
421 <p>The most typical usage is like this:</p>
422<pre> if (!ev_default_loop (0))
423 fatal (&quot;could not initialise libev, bad $LIBEV_FLAGS in environment?&quot;);
424
425</pre>
426 <p>Restrict libev to the select and poll backends, and do not allow
427environment settings to be taken into account:</p>
428<pre> ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);
429
430</pre>
431 <p>Use whatever libev has to offer, but make sure that kqueue is used if
432available (warning, breaks stuff, best use only with your own private
433event loop and only if you know the OS supports your types of fds):</p>
434<pre> ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);
435
436</pre>
198 </dd> 437 </dd>
199 <dt>struct ev_loop *ev_loop_new (unsigned int flags)</dt> 438 <dt>struct ev_loop *ev_loop_new (unsigned int flags)</dt>
200 <dd> 439 <dd>
201 <p>Similar to <code>ev_default_loop</code>, but always creates a new event loop that is 440 <p>Similar to <code>ev_default_loop</code>, but always creates a new event loop that is
202always distinct from the default loop. Unlike the default loop, it cannot 441always distinct from the default loop. Unlike the default loop, it cannot
203handle signal and child watchers, and attempts to do so will be greeted by 442handle signal and child watchers, and attempts to do so will be greeted by
204undefined behaviour (or a failed assertion if assertions are enabled).</p> 443undefined behaviour (or a failed assertion if assertions are enabled).</p>
444 <p>Example: Try to create a event loop that uses epoll and nothing else.</p>
445<pre> struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);
446 if (!epoller)
447 fatal (&quot;no epoll found here, maybe it hides under your chair&quot;);
448
449</pre>
205 </dd> 450 </dd>
206 <dt>ev_default_destroy ()</dt> 451 <dt>ev_default_destroy ()</dt>
207 <dd> 452 <dd>
208 <p>Destroys the default loop again (frees all memory and kernel state 453 <p>Destroys the default loop again (frees all memory and kernel state
209etc.). This stops all registered event watchers (by not touching them in 454etc.). None of the active event watchers will be stopped in the normal
210any way whatsoever, although you cannot rely on this :).</p> 455sense, so e.g. <code>ev_is_active</code> might still return true. It is your
456responsibility to either stop all watchers cleanly yoursef <i>before</i>
457calling this function, or cope with the fact afterwards (which is usually
458the easiest thing, youc na just ignore the watchers and/or <code>free ()</code> them
459for example).</p>
211 </dd> 460 </dd>
212 <dt>ev_loop_destroy (loop)</dt> 461 <dt>ev_loop_destroy (loop)</dt>
213 <dd> 462 <dd>
214 <p>Like <code>ev_default_destroy</code>, but destroys an event loop created by an 463 <p>Like <code>ev_default_destroy</code>, but destroys an event loop created by an
215earlier call to <code>ev_loop_new</code>.</p> 464earlier call to <code>ev_loop_new</code>.</p>
218 <dd> 467 <dd>
219 <p>This function reinitialises the kernel state for backends that have 468 <p>This function reinitialises the kernel state for backends that have
220one. Despite the name, you can call it anytime, but it makes most sense 469one. Despite the name, you can call it anytime, but it makes most sense
221after forking, in either the parent or child process (or both, but that 470after forking, in either the parent or child process (or both, but that
222again makes little sense).</p> 471again makes little sense).</p>
223 <p>You <i>must</i> call this function after forking if and only if you want to 472 <p>You <i>must</i> call this function in the child process after forking if and
224use the event library in both processes. If you just fork+exec, you don't 473only if you want to use the event library in both processes. If you just
225have to call it.</p> 474fork+exec, you don't have to call it.</p>
226 <p>The function itself is quite fast and it's usually not a problem to call 475 <p>The function itself is quite fast and it's usually not a problem to call
227it just in case after a fork. To make this easy, the function will fit in 476it just in case after a fork. To make this easy, the function will fit in
228quite nicely into a call to <code>pthread_atfork</code>:</p> 477quite nicely into a call to <code>pthread_atfork</code>:</p>
229<pre> pthread_atfork (0, 0, ev_default_fork); 478<pre> pthread_atfork (0, 0, ev_default_fork);
230 479
231</pre> 480</pre>
481 <p>At the moment, <code>EVBACKEND_SELECT</code> and <code>EVBACKEND_POLL</code> are safe to use
482without calling this function, so if you force one of those backends you
483do not need to care.</p>
232 </dd> 484 </dd>
233 <dt>ev_loop_fork (loop)</dt> 485 <dt>ev_loop_fork (loop)</dt>
234 <dd> 486 <dd>
235 <p>Like <code>ev_default_fork</code>, but acts on an event loop created by 487 <p>Like <code>ev_default_fork</code>, but acts on an event loop created by
236<code>ev_loop_new</code>. Yes, you have to call this on every allocated event loop 488<code>ev_loop_new</code>. Yes, you have to call this on every allocated event loop
237after fork, and how you do this is entirely your own problem.</p> 489after fork, and how you do this is entirely your own problem.</p>
238 </dd> 490 </dd>
239 <dt>unsigned int ev_method (loop)</dt> 491 <dt>unsigned int ev_loop_count (loop)</dt>
492 <dd>
493 <p>Returns the count of loop iterations for the loop, which is identical to
494the number of times libev did poll for new events. It starts at <code>0</code> and
495happily wraps around with enough iterations.</p>
496 <p>This value can sometimes be useful as a generation counter of sorts (it
497&quot;ticks&quot; the number of loop iterations), as it roughly corresponds with
498<code>ev_prepare</code> and <code>ev_check</code> calls.</p>
240 <dd> 499 </dd>
500 <dt>unsigned int ev_backend (loop)</dt>
501 <dd>
241 <p>Returns one of the <code>EVMETHOD_*</code> flags indicating the event backend in 502 <p>Returns one of the <code>EVBACKEND_*</code> flags indicating the event backend in
242use.</p> 503use.</p>
243 </dd> 504 </dd>
244 <dt>ev_tstamp ev_now (loop)</dt> 505 <dt>ev_tstamp ev_now (loop)</dt>
245 <dd> 506 <dd>
246 <p>Returns the current &quot;event loop time&quot;, which is the time the event loop 507 <p>Returns the current &quot;event loop time&quot;, which is the time the event loop
247got events and started processing them. This timestamp does not change 508received events and started processing them. This timestamp does not
248as long as callbacks are being processed, and this is also the base time 509change as long as callbacks are being processed, and this is also the base
249used for relative timers. You can treat it as the timestamp of the event 510time used for relative timers. You can treat it as the timestamp of the
250occuring (or more correctly, the mainloop finding out about it).</p> 511event occuring (or more correctly, libev finding out about it).</p>
251 </dd> 512 </dd>
252 <dt>ev_loop (loop, int flags)</dt> 513 <dt>ev_loop (loop, int flags)</dt>
253 <dd> 514 <dd>
254 <p>Finally, this is it, the event handler. This function usually is called 515 <p>Finally, this is it, the event handler. This function usually is called
255after you initialised all your watchers and you want to start handling 516after you initialised all your watchers and you want to start handling
256events.</p> 517events.</p>
257 <p>If the flags argument is specified as 0, it will not return until either 518 <p>If the flags argument is specified as <code>0</code>, it will not return until
258no event watchers are active anymore or <code>ev_unloop</code> was called.</p> 519either no event watchers are active anymore or <code>ev_unloop</code> was called.</p>
520 <p>Please note that an explicit <code>ev_unloop</code> is usually better than
521relying on all watchers to be stopped when deciding when a program has
522finished (especially in interactive programs), but having a program that
523automatically loops as long as it has to and no longer by virtue of
524relying on its watchers stopping correctly is a thing of beauty.</p>
259 <p>A flags value of <code>EVLOOP_NONBLOCK</code> will look for new events, will handle 525 <p>A flags value of <code>EVLOOP_NONBLOCK</code> will look for new events, will handle
260those events and any outstanding ones, but will not block your process in 526those events and any outstanding ones, but will not block your process in
261case there are no events and will return after one iteration of the loop.</p> 527case there are no events and will return after one iteration of the loop.</p>
262 <p>A flags value of <code>EVLOOP_ONESHOT</code> will look for new events (waiting if 528 <p>A flags value of <code>EVLOOP_ONESHOT</code> will look for new events (waiting if
263neccessary) and will handle those and any outstanding ones. It will block 529neccessary) and will handle those and any outstanding ones. It will block
264your process until at least one new event arrives, and will return after 530your process until at least one new event arrives, and will return after
265one iteration of the loop.</p> 531one iteration of the loop. This is useful if you are waiting for some
266 <p>This flags value could be used to implement alternative looping 532external event in conjunction with something not expressible using other
267constructs, but the <code>prepare</code> and <code>check</code> watchers provide a better and 533libev watchers. However, a pair of <code>ev_prepare</code>/<code>ev_check</code> watchers is
268more generic mechanism.</p> 534usually a better approach for this kind of thing.</p>
535 <p>Here are the gory details of what <code>ev_loop</code> does:</p>
536<pre> - Before the first iteration, call any pending watchers.
537 * If there are no active watchers (reference count is zero), return.
538 - Queue all prepare watchers and then call all outstanding watchers.
539 - If we have been forked, recreate the kernel state.
540 - Update the kernel state with all outstanding changes.
541 - Update the &quot;event loop time&quot;.
542 - Calculate for how long to block.
543 - Block the process, waiting for any events.
544 - Queue all outstanding I/O (fd) events.
545 - Update the &quot;event loop time&quot; and do time jump handling.
546 - Queue all outstanding timers.
547 - Queue all outstanding periodics.
548 - If no events are pending now, queue all idle watchers.
549 - Queue all check watchers.
550 - Call all queued watchers in reverse order (i.e. check watchers first).
551 Signals and child watchers are implemented as I/O watchers, and will
552 be handled here by queueing them when their watcher gets executed.
553 - If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK
554 were used, return, otherwise continue with step *.
555
556</pre>
557 <p>Example: Queue some jobs and then loop until no events are outsanding
558anymore.</p>
559<pre> ... queue jobs here, make sure they register event watchers as long
560 ... as they still have work to do (even an idle watcher will do..)
561 ev_loop (my_loop, 0);
562 ... jobs done. yeah!
563
564</pre>
269 </dd> 565 </dd>
270 <dt>ev_unloop (loop, how)</dt> 566 <dt>ev_unloop (loop, how)</dt>
271 <dd> 567 <dd>
272 <p>Can be used to make a call to <code>ev_loop</code> return early (but only after it 568 <p>Can be used to make a call to <code>ev_loop</code> return early (but only after it
273has processed all outstanding events). The <code>how</code> argument must be either 569has processed all outstanding events). The <code>how</code> argument must be either
274<code>EVUNLOOP_ONCE</code>, which will make the innermost <code>ev_loop</code> call return, or 570<code>EVUNLOOP_ONE</code>, which will make the innermost <code>ev_loop</code> call return, or
275<code>EVUNLOOP_ALL</code>, which will make all nested <code>ev_loop</code> calls return.</p> 571<code>EVUNLOOP_ALL</code>, which will make all nested <code>ev_loop</code> calls return.</p>
276 </dd> 572 </dd>
277 <dt>ev_ref (loop)</dt> 573 <dt>ev_ref (loop)</dt>
278 <dt>ev_unref (loop)</dt> 574 <dt>ev_unref (loop)</dt>
279 <dd> 575 <dd>
285example, libev itself uses this for its internal signal pipe: It is not 581example, libev itself uses this for its internal signal pipe: It is not
286visible to the libev user and should not keep <code>ev_loop</code> from exiting if 582visible to the libev user and should not keep <code>ev_loop</code> from exiting if
287no event watchers registered by it are active. It is also an excellent 583no event watchers registered by it are active. It is also an excellent
288way to do this for generic recurring timers or from within third-party 584way to do this for generic recurring timers or from within third-party
289libraries. Just remember to <i>unref after start</i> and <i>ref before stop</i>.</p> 585libraries. Just remember to <i>unref after start</i> and <i>ref before stop</i>.</p>
586 <p>Example: Create a signal watcher, but keep it from keeping <code>ev_loop</code>
587running when nothing else is active.</p>
588<pre> struct ev_signal exitsig;
589 ev_signal_init (&amp;exitsig, sig_cb, SIGINT);
590 ev_signal_start (loop, &amp;exitsig);
591 evf_unref (loop);
592
593</pre>
594 <p>Example: For some weird reason, unregister the above signal handler again.</p>
595<pre> ev_ref (loop);
596 ev_signal_stop (loop, &amp;exitsig);
597
598</pre>
290 </dd> 599 </dd>
291</dl> 600</dl>
292 601
602
603
604
605
293</div> 606</div>
294<h1 id="ANATOMY_OF_A_WATCHER">ANATOMY OF A WATCHER</h1><p><a href="#TOP" class="toplink">Top</a></p> 607<h1 id="ANATOMY_OF_A_WATCHER">ANATOMY OF A WATCHER</h1>
295<div id="ANATOMY_OF_A_WATCHER_CONTENT"> 608<div id="ANATOMY_OF_A_WATCHER_CONTENT">
296<p>A watcher is a structure that you create and register to record your 609<p>A watcher is a structure that you create and register to record your
297interest in some event. For instance, if you want to wait for STDIN to 610interest in some event. For instance, if you want to wait for STDIN to
298become readable, you would create an <code>ev_io</code> watcher for that:</p> 611become readable, you would create an <code>ev_io</code> watcher for that:</p>
299<pre> static void my_cb (struct ev_loop *loop, struct ev_io *w, int revents) 612<pre> static void my_cb (struct ev_loop *loop, struct ev_io *w, int revents)
326with a watcher-specific start function (<code>ev_&lt;type&gt;_start (loop, watcher 639with a watcher-specific start function (<code>ev_&lt;type&gt;_start (loop, watcher
327*)</code>), and you can stop watching for events at any time by calling the 640*)</code>), and you can stop watching for events at any time by calling the
328corresponding stop function (<code>ev_&lt;type&gt;_stop (loop, watcher *)</code>.</p> 641corresponding stop function (<code>ev_&lt;type&gt;_stop (loop, watcher *)</code>.</p>
329<p>As long as your watcher is active (has been started but not stopped) you 642<p>As long as your watcher is active (has been started but not stopped) you
330must not touch the values stored in it. Most specifically you must never 643must not touch the values stored in it. Most specifically you must never
331reinitialise it or call its set method.</p> 644reinitialise it or call its <code>set</code> macro.</p>
332<p>You can check whether an event is active by calling the <code>ev_is_active
333(watcher *)</code> macro. To see whether an event is outstanding (but the
334callback for it has not been called yet) you can use the <code>ev_is_pending
335(watcher *)</code> macro.</p>
336<p>Each and every callback receives the event loop pointer as first, the 645<p>Each and every callback receives the event loop pointer as first, the
337registered watcher structure as second, and a bitset of received events as 646registered watcher structure as second, and a bitset of received events as
338third argument.</p> 647third argument.</p>
339<p>The received events usually include a single bit per event type received 648<p>The received events usually include a single bit per event type received
340(you can receive multiple events at the same time). The possible bit masks 649(you can receive multiple events at the same time). The possible bit masks
360 </dd> 669 </dd>
361 <dt><code>EV_CHILD</code></dt> 670 <dt><code>EV_CHILD</code></dt>
362 <dd> 671 <dd>
363 <p>The pid specified in the <code>ev_child</code> watcher has received a status change.</p> 672 <p>The pid specified in the <code>ev_child</code> watcher has received a status change.</p>
364 </dd> 673 </dd>
674 <dt><code>EV_STAT</code></dt>
675 <dd>
676 <p>The path specified in the <code>ev_stat</code> watcher changed its attributes somehow.</p>
677 </dd>
365 <dt><code>EV_IDLE</code></dt> 678 <dt><code>EV_IDLE</code></dt>
366 <dd> 679 <dd>
367 <p>The <code>ev_idle</code> watcher has determined that you have nothing better to do.</p> 680 <p>The <code>ev_idle</code> watcher has determined that you have nothing better to do.</p>
368 </dd> 681 </dd>
369 <dt><code>EV_PREPARE</code></dt> 682 <dt><code>EV_PREPARE</code></dt>
374<code>ev_loop</code> has gathered them, but before it invokes any callbacks for any 687<code>ev_loop</code> has gathered them, but before it invokes any callbacks for any
375received events. Callbacks of both watcher types can start and stop as 688received events. Callbacks of both watcher types can start and stop as
376many watchers as they want, and all of them will be taken into account 689many watchers as they want, and all of them will be taken into account
377(for example, a <code>ev_prepare</code> watcher might start an idle watcher to keep 690(for example, a <code>ev_prepare</code> watcher might start an idle watcher to keep
378<code>ev_loop</code> from blocking).</p> 691<code>ev_loop</code> from blocking).</p>
692 </dd>
693 <dt><code>EV_EMBED</code></dt>
694 <dd>
695 <p>The embedded event loop specified in the <code>ev_embed</code> watcher needs attention.</p>
696 </dd>
697 <dt><code>EV_FORK</code></dt>
698 <dd>
699 <p>The event loop has been resumed in the child process after fork (see
700<code>ev_fork</code>).</p>
379 </dd> 701 </dd>
380 <dt><code>EV_ERROR</code></dt> 702 <dt><code>EV_ERROR</code></dt>
381 <dd> 703 <dd>
382 <p>An unspecified error has occured, the watcher has been stopped. This might 704 <p>An unspecified error has occured, the watcher has been stopped. This might
383happen because the watcher could not be properly started because libev 705happen because the watcher could not be properly started because libev
391programs, though, so beware.</p> 713programs, though, so beware.</p>
392 </dd> 714 </dd>
393</dl> 715</dl>
394 716
395</div> 717</div>
718<h2 id="GENERIC_WATCHER_FUNCTIONS">GENERIC WATCHER FUNCTIONS</h2>
719<div id="GENERIC_WATCHER_FUNCTIONS_CONTENT">
720<p>In the following description, <code>TYPE</code> stands for the watcher type,
721e.g. <code>timer</code> for <code>ev_timer</code> watchers and <code>io</code> for <code>ev_io</code> watchers.</p>
722<dl>
723 <dt><code>ev_init</code> (ev_TYPE *watcher, callback)</dt>
724 <dd>
725 <p>This macro initialises the generic portion of a watcher. The contents
726of the watcher object can be arbitrary (so <code>malloc</code> will do). Only
727the generic parts of the watcher are initialised, you <i>need</i> to call
728the type-specific <code>ev_TYPE_set</code> macro afterwards to initialise the
729type-specific parts. For each type there is also a <code>ev_TYPE_init</code> macro
730which rolls both calls into one.</p>
731 <p>You can reinitialise a watcher at any time as long as it has been stopped
732(or never started) and there are no pending events outstanding.</p>
733 <p>The callback is always of type <code>void (*)(ev_loop *loop, ev_TYPE *watcher,
734int revents)</code>.</p>
735 </dd>
736 <dt><code>ev_TYPE_set</code> (ev_TYPE *, [args])</dt>
737 <dd>
738 <p>This macro initialises the type-specific parts of a watcher. You need to
739call <code>ev_init</code> at least once before you call this macro, but you can
740call <code>ev_TYPE_set</code> any number of times. You must not, however, call this
741macro on a watcher that is active (it can be pending, however, which is a
742difference to the <code>ev_init</code> macro).</p>
743 <p>Although some watcher types do not have type-specific arguments
744(e.g. <code>ev_prepare</code>) you still need to call its <code>set</code> macro.</p>
745 </dd>
746 <dt><code>ev_TYPE_init</code> (ev_TYPE *watcher, callback, [args])</dt>
747 <dd>
748 <p>This convinience macro rolls both <code>ev_init</code> and <code>ev_TYPE_set</code> macro
749calls into a single call. This is the most convinient method to initialise
750a watcher. The same limitations apply, of course.</p>
751 </dd>
752 <dt><code>ev_TYPE_start</code> (loop *, ev_TYPE *watcher)</dt>
753 <dd>
754 <p>Starts (activates) the given watcher. Only active watchers will receive
755events. If the watcher is already active nothing will happen.</p>
756 </dd>
757 <dt><code>ev_TYPE_stop</code> (loop *, ev_TYPE *watcher)</dt>
758 <dd>
759 <p>Stops the given watcher again (if active) and clears the pending
760status. It is possible that stopped watchers are pending (for example,
761non-repeating timers are being stopped when they become pending), but
762<code>ev_TYPE_stop</code> ensures that the watcher is neither active nor pending. If
763you want to free or reuse the memory used by the watcher it is therefore a
764good idea to always call its <code>ev_TYPE_stop</code> function.</p>
765 </dd>
766 <dt>bool ev_is_active (ev_TYPE *watcher)</dt>
767 <dd>
768 <p>Returns a true value iff the watcher is active (i.e. it has been started
769and not yet been stopped). As long as a watcher is active you must not modify
770it.</p>
771 </dd>
772 <dt>bool ev_is_pending (ev_TYPE *watcher)</dt>
773 <dd>
774 <p>Returns a true value iff the watcher is pending, (i.e. it has outstanding
775events but its callback has not yet been invoked). As long as a watcher
776is pending (but not active) you must not call an init function on it (but
777<code>ev_TYPE_set</code> is safe), you must not change its priority, and you must
778make sure the watcher is available to libev (e.g. you cannot <code>free ()</code>
779it).</p>
780 </dd>
781 <dt>callback ev_cb (ev_TYPE *watcher)</dt>
782 <dd>
783 <p>Returns the callback currently set on the watcher.</p>
784 </dd>
785 <dt>ev_cb_set (ev_TYPE *watcher, callback)</dt>
786 <dd>
787 <p>Change the callback. You can change the callback at virtually any time
788(modulo threads).</p>
789 </dd>
790 <dt>ev_set_priority (ev_TYPE *watcher, priority)</dt>
791 <dt>int ev_priority (ev_TYPE *watcher)</dt>
792 <dd>
793 <p>Set and query the priority of the watcher. The priority is a small
794integer between <code>EV_MAXPRI</code> (default: <code>2</code>) and <code>EV_MINPRI</code>
795(default: <code>-2</code>). Pending watchers with higher priority will be invoked
796before watchers with lower priority, but priority will not keep watchers
797from being executed (except for <code>ev_idle</code> watchers).</p>
798 <p>This means that priorities are <i>only</i> used for ordering callback
799invocation after new events have been received. This is useful, for
800example, to reduce latency after idling, or more often, to bind two
801watchers on the same event and make sure one is called first.</p>
802 <p>If you need to suppress invocation when higher priority events are pending
803you need to look at <code>ev_idle</code> watchers, which provide this functionality.</p>
804 <p>You <i>must not</i> change the priority of a watcher as long as it is active or
805pending.</p>
806 <p>The default priority used by watchers when no priority has been set is
807always <code>0</code>, which is supposed to not be too high and not be too low :).</p>
808 <p>Setting a priority outside the range of <code>EV_MINPRI</code> to <code>EV_MAXPRI</code> is
809fine, as long as you do not mind that the priority value you query might
810or might not have been adjusted to be within valid range.</p>
811 </dd>
812 <dt>ev_invoke (loop, ev_TYPE *watcher, int revents)</dt>
813 <dd>
814 <p>Invoke the <code>watcher</code> with the given <code>loop</code> and <code>revents</code>. Neither
815<code>loop</code> nor <code>revents</code> need to be valid as long as the watcher callback
816can deal with that fact.</p>
817 </dd>
818 <dt>int ev_clear_pending (loop, ev_TYPE *watcher)</dt>
819 <dd>
820 <p>If the watcher is pending, this function returns clears its pending status
821and returns its <code>revents</code> bitset (as if its callback was invoked). If the
822watcher isn't pending it does nothing and returns <code>0</code>.</p>
823 </dd>
824</dl>
825
826
827
828
829
830</div>
396<h2 id="ASSOCIATING_CUSTOM_DATA_WITH_A_WATCH">ASSOCIATING CUSTOM DATA WITH A WATCHER</h2> 831<h2 id="ASSOCIATING_CUSTOM_DATA_WITH_A_WATCH">ASSOCIATING CUSTOM DATA WITH A WATCHER</h2>
397<div id="ASSOCIATING_CUSTOM_DATA_WITH_A_WATCH-2"> 832<div id="ASSOCIATING_CUSTOM_DATA_WITH_A_WATCH-2">
398<p>Each watcher has, by default, a member <code>void *data</code> that you can change 833<p>Each watcher has, by default, a member <code>void *data</code> that you can change
399and read at any time, libev will completely ignore it. This can be used 834and read at any time, libev will completely ignore it. This can be used
400to associate arbitrary data with your watcher. If you need more data and 835to associate arbitrary data with your watcher. If you need more data and
417 struct my_io *w = (struct my_io *)w_; 852 struct my_io *w = (struct my_io *)w_;
418 ... 853 ...
419 } 854 }
420 855
421</pre> 856</pre>
422<p>More interesting and less C-conformant ways of catsing your callback type 857<p>More interesting and less C-conformant ways of casting your callback type
423have been omitted....</p> 858instead have been omitted.</p>
859<p>Another common scenario is having some data structure with multiple
860watchers:</p>
861<pre> struct my_biggy
862 {
863 int some_data;
864 ev_timer t1;
865 ev_timer t2;
866 }
424 867
868</pre>
869<p>In this case getting the pointer to <code>my_biggy</code> is a bit more complicated,
870you need to use <code>offsetof</code>:</p>
871<pre> #include &lt;stddef.h&gt;
425 872
873 static void
874 t1_cb (EV_P_ struct ev_timer *w, int revents)
875 {
876 struct my_biggy big = (struct my_biggy *
877 (((char *)w) - offsetof (struct my_biggy, t1));
878 }
426 879
880 static void
881 t2_cb (EV_P_ struct ev_timer *w, int revents)
882 {
883 struct my_biggy big = (struct my_biggy *
884 (((char *)w) - offsetof (struct my_biggy, t2));
885 }
427 886
428 887
888
889
890</pre>
891
429</div> 892</div>
430<h1 id="WATCHER_TYPES">WATCHER TYPES</h1><p><a href="#TOP" class="toplink">Top</a></p> 893<h1 id="WATCHER_TYPES">WATCHER TYPES</h1>
431<div id="WATCHER_TYPES_CONTENT"> 894<div id="WATCHER_TYPES_CONTENT">
432<p>This section describes each watcher in detail, but will not repeat 895<p>This section describes each watcher in detail, but will not repeat
433information given in the last section.</p> 896information given in the last section. Any initialisation/set macros,
897functions and members specific to the watcher type are explained.</p>
898<p>Members are additionally marked with either <i>[read-only]</i>, meaning that,
899while the watcher is active, you can look at the member and expect some
900sensible content, but you must not modify it (you can modify it while the
901watcher is stopped to your hearts content), or <i>[read-write]</i>, which
902means you can expect it to have some sensible content while the watcher
903is active, but you can also modify it. Modifying it may not do something
904sensible or take immediate effect (or do anything at all), but libev will
905not crash or malfunction in any way.</p>
434 906
907
908
909
910
435</div> 911</div>
436<h2 id="code_ev_io_code_is_this_file_descrip"><code>ev_io</code> - is this file descriptor readable or writable</h2> 912<h2 id="code_ev_io_code_is_this_file_descrip"><code>ev_io</code> - is this file descriptor readable or writable?</h2>
437<div id="code_ev_io_code_is_this_file_descrip-2"> 913<div id="code_ev_io_code_is_this_file_descrip-2">
438<p>I/O watchers check whether a file descriptor is readable or writable 914<p>I/O watchers check whether a file descriptor is readable or writable
439in each iteration of the event loop (This behaviour is called 915in each iteration of the event loop, or, more precisely, when reading
440level-triggering because you keep receiving events as long as the 916would not block the process and writing would at least be able to write
441condition persists. Remember you can stop the watcher if you don't want to 917some data. This behaviour is called level-triggering because you keep
442act on the event and neither want to receive future events).</p> 918receiving events as long as the condition persists. Remember you can stop
919the watcher if you don't want to act on the event and neither want to
920receive future events.</p>
443<p>In general you can register as many read and/or write event watchers oer 921<p>In general you can register as many read and/or write event watchers per
444fd as you want (as long as you don't confuse yourself). Setting all file 922fd as you want (as long as you don't confuse yourself). Setting all file
445descriptors to non-blocking mode is also usually a good idea (but not 923descriptors to non-blocking mode is also usually a good idea (but not
446required if you know what you are doing).</p> 924required if you know what you are doing).</p>
447<p>You have to be careful with dup'ed file descriptors, though. Some backends 925<p>You have to be careful with dup'ed file descriptors, though. Some backends
448(the linux epoll backend is a notable example) cannot handle dup'ed file 926(the linux epoll backend is a notable example) cannot handle dup'ed file
449descriptors correctly if you register interest in two or more fds pointing 927descriptors correctly if you register interest in two or more fds pointing
450to the same file/socket etc. description.</p> 928to the same underlying file/socket/etc. description (that is, they share
929the same underlying &quot;file open&quot;).</p>
451<p>If you must do this, then force the use of a known-to-be-good backend 930<p>If you must do this, then force the use of a known-to-be-good backend
452(at the time of this writing, this includes only EVMETHOD_SELECT and 931(at the time of this writing, this includes only <code>EVBACKEND_SELECT</code> and
453EVMETHOD_POLL).</p> 932<code>EVBACKEND_POLL</code>).</p>
933<p>Another thing you have to watch out for is that it is quite easy to
934receive &quot;spurious&quot; readyness notifications, that is your callback might
935be called with <code>EV_READ</code> but a subsequent <code>read</code>(2) will actually block
936because there is no data. Not only are some backends known to create a
937lot of those (for example solaris ports), it is very easy to get into
938this situation even with a relatively standard program structure. Thus
939it is best to always use non-blocking I/O: An extra <code>read</code>(2) returning
940<code>EAGAIN</code> is far preferable to a program hanging until some data arrives.</p>
941<p>If you cannot run the fd in non-blocking mode (for example you should not
942play around with an Xlib connection), then you have to seperately re-test
943whether a file descriptor is really ready with a known-to-be good interface
944such as poll (fortunately in our Xlib example, Xlib already does this on
945its own, so its quite safe to use).</p>
454<dl> 946<dl>
455 <dt>ev_io_init (ev_io *, callback, int fd, int events)</dt> 947 <dt>ev_io_init (ev_io *, callback, int fd, int events)</dt>
456 <dt>ev_io_set (ev_io *, int fd, int events)</dt> 948 <dt>ev_io_set (ev_io *, int fd, int events)</dt>
457 <dd> 949 <dd>
458 <p>Configures an <code>ev_io</code> watcher. The fd is the file descriptor to rceeive 950 <p>Configures an <code>ev_io</code> watcher. The <code>fd</code> is the file descriptor to
459events for and events is either <code>EV_READ</code>, <code>EV_WRITE</code> or <code>EV_READ | 951rceeive events for and events is either <code>EV_READ</code>, <code>EV_WRITE</code> or
460EV_WRITE</code> to receive the given events.</p> 952<code>EV_READ | EV_WRITE</code> to receive the given events.</p>
953 </dd>
954 <dt>int fd [read-only]</dt>
955 <dd>
956 <p>The file descriptor being watched.</p>
957 </dd>
958 <dt>int events [read-only]</dt>
959 <dd>
960 <p>The events being watched.</p>
461 </dd> 961 </dd>
462</dl> 962</dl>
963<p>Example: Call <code>stdin_readable_cb</code> when STDIN_FILENO has become, well
964readable, but only once. Since it is likely line-buffered, you could
965attempt to read a whole line in the callback.</p>
966<pre> static void
967 stdin_readable_cb (struct ev_loop *loop, struct ev_io *w, int revents)
968 {
969 ev_io_stop (loop, w);
970 .. read from stdin here (or from w-&gt;fd) and haqndle any I/O errors
971 }
463 972
973 ...
974 struct ev_loop *loop = ev_default_init (0);
975 struct ev_io stdin_readable;
976 ev_io_init (&amp;stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
977 ev_io_start (loop, &amp;stdin_readable);
978 ev_loop (loop, 0);
979
980
981
982
983</pre>
984
464</div> 985</div>
465<h2 id="code_ev_timer_code_relative_and_opti"><code>ev_timer</code> - relative and optionally recurring timeouts</h2> 986<h2 id="code_ev_timer_code_relative_and_opti"><code>ev_timer</code> - relative and optionally repeating timeouts</h2>
466<div id="code_ev_timer_code_relative_and_opti-2"> 987<div id="code_ev_timer_code_relative_and_opti-2">
467<p>Timer watchers are simple relative timers that generate an event after a 988<p>Timer watchers are simple relative timers that generate an event after a
468given time, and optionally repeating in regular intervals after that.</p> 989given time, and optionally repeating in regular intervals after that.</p>
469<p>The timers are based on real time, that is, if you register an event that 990<p>The timers are based on real time, that is, if you register an event that
470times out after an hour and youreset your system clock to last years 991times out after an hour and you reset your system clock to last years
471time, it will still time out after (roughly) and hour. &quot;Roughly&quot; because 992time, it will still time out after (roughly) and hour. &quot;Roughly&quot; because
472detecting time jumps is hard, and soem inaccuracies are unavoidable (the 993detecting time jumps is hard, and some inaccuracies are unavoidable (the
473monotonic clock option helps a lot here).</p> 994monotonic clock option helps a lot here).</p>
474<p>The relative timeouts are calculated relative to the <code>ev_now ()</code> 995<p>The relative timeouts are calculated relative to the <code>ev_now ()</code>
475time. This is usually the right thing as this timestamp refers to the time 996time. This is usually the right thing as this timestamp refers to the time
476of the event triggering whatever timeout you are modifying/starting. If 997of the event triggering whatever timeout you are modifying/starting. If
477you suspect event processing to be delayed and you *need* to base the timeout 998you suspect event processing to be delayed and you <i>need</i> to base the timeout
478ion the current time, use something like this to adjust for this:</p> 999on the current time, use something like this to adjust for this:</p>
479<pre> ev_timer_set (&amp;timer, after + ev_now () - ev_time (), 0.); 1000<pre> ev_timer_set (&amp;timer, after + ev_now () - ev_time (), 0.);
480 1001
481</pre> 1002</pre>
1003<p>The callback is guarenteed to be invoked only when its timeout has passed,
1004but if multiple timers become ready during the same loop iteration then
1005order of execution is undefined.</p>
482<dl> 1006<dl>
483 <dt>ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)</dt> 1007 <dt>ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)</dt>
484 <dt>ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)</dt> 1008 <dt>ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)</dt>
485 <dd> 1009 <dd>
486 <p>Configure the timer to trigger after <code>after</code> seconds. If <code>repeat</code> is 1010 <p>Configure the timer to trigger after <code>after</code> seconds. If <code>repeat</code> is
488timer will automatically be configured to trigger again <code>repeat</code> seconds 1012timer will automatically be configured to trigger again <code>repeat</code> seconds
489later, again, and again, until stopped manually.</p> 1013later, again, and again, until stopped manually.</p>
490 <p>The timer itself will do a best-effort at avoiding drift, that is, if you 1014 <p>The timer itself will do a best-effort at avoiding drift, that is, if you
491configure a timer to trigger every 10 seconds, then it will trigger at 1015configure a timer to trigger every 10 seconds, then it will trigger at
492exactly 10 second intervals. If, however, your program cannot keep up with 1016exactly 10 second intervals. If, however, your program cannot keep up with
493the timer (ecause it takes longer than those 10 seconds to do stuff) the 1017the timer (because it takes longer than those 10 seconds to do stuff) the
494timer will not fire more than once per event loop iteration.</p> 1018timer will not fire more than once per event loop iteration.</p>
495 </dd> 1019 </dd>
496 <dt>ev_timer_again (loop)</dt> 1020 <dt>ev_timer_again (loop)</dt>
497 <dd> 1021 <dd>
498 <p>This will act as if the timer timed out and restart it again if it is 1022 <p>This will act as if the timer timed out and restart it again if it is
499repeating. The exact semantics are:</p> 1023repeating. The exact semantics are:</p>
1024 <p>If the timer is pending, its pending status is cleared.</p>
500 <p>If the timer is started but nonrepeating, stop it.</p> 1025 <p>If the timer is started but nonrepeating, stop it (as if it timed out).</p>
501 <p>If the timer is repeating, either start it if necessary (with the repeat 1026 <p>If the timer is repeating, either start it if necessary (with the
502value), or reset the running timer to the repeat value.</p> 1027<code>repeat</code> value), or reset the running timer to the <code>repeat</code> value.</p>
503 <p>This sounds a bit complicated, but here is a useful and typical 1028 <p>This sounds a bit complicated, but here is a useful and typical
504example: Imagine you have a tcp connection and you want a so-called idle 1029example: Imagine you have a tcp connection and you want a so-called idle
505timeout, that is, you want to be called when there have been, say, 60 1030timeout, that is, you want to be called when there have been, say, 60
506seconds of inactivity on the socket. The easiest way to do this is to 1031seconds of inactivity on the socket. The easiest way to do this is to
507configure an <code>ev_timer</code> with after=repeat=60 and calling ev_timer_again each 1032configure an <code>ev_timer</code> with a <code>repeat</code> value of <code>60</code> and then call
508time you successfully read or write some data. If you go into an idle 1033<code>ev_timer_again</code> each time you successfully read or write some data. If
509state where you do not expect data to travel on the socket, you can stop 1034you go into an idle state where you do not expect data to travel on the
1035socket, you can <code>ev_timer_stop</code> the timer, and <code>ev_timer_again</code> will
510the timer, and again will automatically restart it if need be.</p> 1036automatically restart it if need be.</p>
1037 <p>That means you can ignore the <code>after</code> value and <code>ev_timer_start</code>
1038altogether and only ever use the <code>repeat</code> value and <code>ev_timer_again</code>:</p>
1039<pre> ev_timer_init (timer, callback, 0., 5.);
1040 ev_timer_again (loop, timer);
1041 ...
1042 timer-&gt;again = 17.;
1043 ev_timer_again (loop, timer);
1044 ...
1045 timer-&gt;again = 10.;
1046 ev_timer_again (loop, timer);
1047
1048</pre>
1049 <p>This is more slightly efficient then stopping/starting the timer each time
1050you want to modify its timeout value.</p>
1051 </dd>
1052 <dt>ev_tstamp repeat [read-write]</dt>
1053 <dd>
1054 <p>The current <code>repeat</code> value. Will be used each time the watcher times out
1055or <code>ev_timer_again</code> is called and determines the next timeout (if any),
1056which is also when any modifications are taken into account.</p>
511 </dd> 1057 </dd>
512</dl> 1058</dl>
1059<p>Example: Create a timer that fires after 60 seconds.</p>
1060<pre> static void
1061 one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
1062 {
1063 .. one minute over, w is actually stopped right here
1064 }
513 1065
1066 struct ev_timer mytimer;
1067 ev_timer_init (&amp;mytimer, one_minute_cb, 60., 0.);
1068 ev_timer_start (loop, &amp;mytimer);
1069
1070</pre>
1071<p>Example: Create a timeout timer that times out after 10 seconds of
1072inactivity.</p>
1073<pre> static void
1074 timeout_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
1075 {
1076 .. ten seconds without any activity
1077 }
1078
1079 struct ev_timer mytimer;
1080 ev_timer_init (&amp;mytimer, timeout_cb, 0., 10.); /* note, only repeat used */
1081 ev_timer_again (&amp;mytimer); /* start timer */
1082 ev_loop (loop, 0);
1083
1084 // and in some piece of code that gets executed on any &quot;activity&quot;:
1085 // reset the timeout to start ticking again at 10 seconds
1086 ev_timer_again (&amp;mytimer);
1087
1088
1089
1090
1091</pre>
1092
514</div> 1093</div>
515<h2 id="code_ev_periodic_code_to_cron_or_not"><code>ev_periodic</code> - to cron or not to cron</h2> 1094<h2 id="code_ev_periodic_code_to_cron_or_not"><code>ev_periodic</code> - to cron or not to cron?</h2>
516<div id="code_ev_periodic_code_to_cron_or_not-2"> 1095<div id="code_ev_periodic_code_to_cron_or_not-2">
517<p>Periodic watchers are also timers of a kind, but they are very versatile 1096<p>Periodic watchers are also timers of a kind, but they are very versatile
518(and unfortunately a bit complex).</p> 1097(and unfortunately a bit complex).</p>
519<p>Unlike <code>ev_timer</code>'s, they are not based on real time (or relative time) 1098<p>Unlike <code>ev_timer</code>'s, they are not based on real time (or relative time)
520but on wallclock time (absolute time). You can tell a periodic watcher 1099but on wallclock time (absolute time). You can tell a periodic watcher
521to trigger &quot;at&quot; some specific point in time. For example, if you tell a 1100to trigger &quot;at&quot; some specific point in time. For example, if you tell a
522periodic watcher to trigger in 10 seconds (by specifiying e.g. c&lt;ev_now () 1101periodic watcher to trigger in 10 seconds (by specifiying e.g. <code>ev_now ()
523+ 10.&gt;) and then reset your system clock to the last year, then it will 1102+ 10.</code>) and then reset your system clock to the last year, then it will
524take a year to trigger the event (unlike an <code>ev_timer</code>, which would trigger 1103take a year to trigger the event (unlike an <code>ev_timer</code>, which would trigger
525roughly 10 seconds later and of course not if you reset your system time 1104roughly 10 seconds later).</p>
526again).</p>
527<p>They can also be used to implement vastly more complex timers, such as 1105<p>They can also be used to implement vastly more complex timers, such as
528triggering an event on eahc midnight, local time.</p> 1106triggering an event on each midnight, local time or other, complicated,
1107rules.</p>
1108<p>As with timers, the callback is guarenteed to be invoked only when the
1109time (<code>at</code>) has been passed, but if multiple periodic timers become ready
1110during the same loop iteration then order of execution is undefined.</p>
529<dl> 1111<dl>
530 <dt>ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)</dt> 1112 <dt>ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)</dt>
531 <dt>ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)</dt> 1113 <dt>ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)</dt>
532 <dd> 1114 <dd>
533 <p>Lots of arguments, lets sort it out... There are basically three modes of 1115 <p>Lots of arguments, lets sort it out... There are basically three modes of
534operation, and we will explain them from simplest to complex:</p> 1116operation, and we will explain them from simplest to complex:</p>
535
536
537
538
539 <p> 1117 <p>
540 <dl> 1118 <dl>
541 <dt>* absolute timer (interval = reschedule_cb = 0)</dt> 1119 <dt>* absolute timer (at = time, interval = reschedule_cb = 0)</dt>
542 <dd> 1120 <dd>
543 <p>In this configuration the watcher triggers an event at the wallclock time 1121 <p>In this configuration the watcher triggers an event at the wallclock time
544<code>at</code> and doesn't repeat. It will not adjust when a time jump occurs, 1122<code>at</code> and doesn't repeat. It will not adjust when a time jump occurs,
545that is, if it is to be run at January 1st 2011 then it will run when the 1123that is, if it is to be run at January 1st 2011 then it will run when the
546system time reaches or surpasses this time.</p> 1124system time reaches or surpasses this time.</p>
547 </dd> 1125 </dd>
548 <dt>* non-repeating interval timer (interval &gt; 0, reschedule_cb = 0)</dt> 1126 <dt>* non-repeating interval timer (at = offset, interval &gt; 0, reschedule_cb = 0)</dt>
549 <dd> 1127 <dd>
550 <p>In this mode the watcher will always be scheduled to time out at the next 1128 <p>In this mode the watcher will always be scheduled to time out at the next
551<code>at + N * interval</code> time (for some integer N) and then repeat, regardless 1129<code>at + N * interval</code> time (for some integer N, which can also be negative)
552of any time jumps.</p> 1130and then repeat, regardless of any time jumps.</p>
553 <p>This can be used to create timers that do not drift with respect to system 1131 <p>This can be used to create timers that do not drift with respect to system
554time:</p> 1132time:</p>
555<pre> ev_periodic_set (&amp;periodic, 0., 3600., 0); 1133<pre> ev_periodic_set (&amp;periodic, 0., 3600., 0);
556 1134
557</pre> 1135</pre>
560full hour (UTC), or more correctly, when the system time is evenly divisible 1138full hour (UTC), or more correctly, when the system time is evenly divisible
561by 3600.</p> 1139by 3600.</p>
562 <p>Another way to think about it (for the mathematically inclined) is that 1140 <p>Another way to think about it (for the mathematically inclined) is that
563<code>ev_periodic</code> will try to run the callback in this mode at the next possible 1141<code>ev_periodic</code> will try to run the callback in this mode at the next possible
564time where <code>time = at (mod interval)</code>, regardless of any time jumps.</p> 1142time where <code>time = at (mod interval)</code>, regardless of any time jumps.</p>
1143 <p>For numerical stability it is preferable that the <code>at</code> value is near
1144<code>ev_now ()</code> (the current time), but there is no range requirement for
1145this value.</p>
565 </dd> 1146 </dd>
566 <dt>* manual reschedule mode (reschedule_cb = callback)</dt> 1147 <dt>* manual reschedule mode (at and interval ignored, reschedule_cb = callback)</dt>
567 <dd> 1148 <dd>
568 <p>In this mode the values for <code>interval</code> and <code>at</code> are both being 1149 <p>In this mode the values for <code>interval</code> and <code>at</code> are both being
569ignored. Instead, each time the periodic watcher gets scheduled, the 1150ignored. Instead, each time the periodic watcher gets scheduled, the
570reschedule callback will be called with the watcher as first, and the 1151reschedule callback will be called with the watcher as first, and the
571current time as second argument.</p> 1152current time as second argument.</p>
572 <p>NOTE: <i>This callback MUST NOT stop or destroy the periodic or any other 1153 <p>NOTE: <i>This callback MUST NOT stop or destroy any periodic watcher,
573periodic watcher, ever, or make any event loop modifications</i>. If you need 1154ever, or make any event loop modifications</i>. If you need to stop it,
574to stop it, return <code>now + 1e30</code> (or so, fudge fudge) and stop it afterwards.</p> 1155return <code>now + 1e30</code> (or so, fudge fudge) and stop it afterwards (e.g. by
575 <p>Also, <i>this callback must always return a time that is later than the 1156starting an <code>ev_prepare</code> watcher, which is legal).</p>
576passed <code>now</code> value</i>. Not even <code>now</code> itself will be ok.</p>
577 <p>Its prototype is <code>ev_tstamp (*reschedule_cb)(struct ev_periodic *w, 1157 <p>Its prototype is <code>ev_tstamp (*reschedule_cb)(struct ev_periodic *w,
578ev_tstamp now)</code>, e.g.:</p> 1158ev_tstamp now)</code>, e.g.:</p>
579<pre> static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now) 1159<pre> static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)
580 { 1160 {
581 return now + 60.; 1161 return now + 60.;
584</pre> 1164</pre>
585 <p>It must return the next time to trigger, based on the passed time value 1165 <p>It must return the next time to trigger, based on the passed time value
586(that is, the lowest time value larger than to the second argument). It 1166(that is, the lowest time value larger than to the second argument). It
587will usually be called just before the callback will be triggered, but 1167will usually be called just before the callback will be triggered, but
588might be called at other times, too.</p> 1168might be called at other times, too.</p>
1169 <p>NOTE: <i>This callback must always return a time that is later than the
1170passed <code>now</code> value</i>. Not even <code>now</code> itself will do, it <i>must</i> be larger.</p>
589 <p>This can be used to create very complex timers, such as a timer that 1171 <p>This can be used to create very complex timers, such as a timer that
590triggers on each midnight, local time. To do this, you would calculate the 1172triggers on each midnight, local time. To do this, you would calculate the
591next midnight after <code>now</code> and return the timestamp value for this. How you do this 1173next midnight after <code>now</code> and return the timestamp value for this. How
592is, again, up to you (but it is not trivial).</p> 1174you do this is, again, up to you (but it is not trivial, which is the main
1175reason I omitted it as an example).</p>
593 </dd> 1176 </dd>
594 </dl> 1177 </dl>
595 </p> 1178 </p>
596 </dd> 1179 </dd>
597 <dt>ev_periodic_again (loop, ev_periodic *)</dt> 1180 <dt>ev_periodic_again (loop, ev_periodic *)</dt>
599 <p>Simply stops and restarts the periodic watcher again. This is only useful 1182 <p>Simply stops and restarts the periodic watcher again. This is only useful
600when you changed some parameters or the reschedule callback would return 1183when you changed some parameters or the reschedule callback would return
601a different time than the last time it was called (e.g. in a crond like 1184a different time than the last time it was called (e.g. in a crond like
602program when the crontabs have changed).</p> 1185program when the crontabs have changed).</p>
603 </dd> 1186 </dd>
1187 <dt>ev_tstamp offset [read-write]</dt>
1188 <dd>
1189 <p>When repeating, this contains the offset value, otherwise this is the
1190absolute point in time (the <code>at</code> value passed to <code>ev_periodic_set</code>).</p>
1191 <p>Can be modified any time, but changes only take effect when the periodic
1192timer fires or <code>ev_periodic_again</code> is being called.</p>
1193 </dd>
1194 <dt>ev_tstamp interval [read-write]</dt>
1195 <dd>
1196 <p>The current interval value. Can be modified any time, but changes only
1197take effect when the periodic timer fires or <code>ev_periodic_again</code> is being
1198called.</p>
1199 </dd>
1200 <dt>ev_tstamp (*reschedule_cb)(struct ev_periodic *w, ev_tstamp now) [read-write]</dt>
1201 <dd>
1202 <p>The current reschedule callback, or <code>0</code>, if this functionality is
1203switched off. Can be changed any time, but changes only take effect when
1204the periodic timer fires or <code>ev_periodic_again</code> is being called.</p>
1205 </dd>
604</dl> 1206</dl>
1207<p>Example: Call a callback every hour, or, more precisely, whenever the
1208system clock is divisible by 3600. The callback invocation times have
1209potentially a lot of jittering, but good long-term stability.</p>
1210<pre> static void
1211 clock_cb (struct ev_loop *loop, struct ev_io *w, int revents)
1212 {
1213 ... its now a full hour (UTC, or TAI or whatever your clock follows)
1214 }
605 1215
1216 struct ev_periodic hourly_tick;
1217 ev_periodic_init (&amp;hourly_tick, clock_cb, 0., 3600., 0);
1218 ev_periodic_start (loop, &amp;hourly_tick);
1219
1220</pre>
1221<p>Example: The same as above, but use a reschedule callback to do it:</p>
1222<pre> #include &lt;math.h&gt;
1223
1224 static ev_tstamp
1225 my_scheduler_cb (struct ev_periodic *w, ev_tstamp now)
1226 {
1227 return fmod (now, 3600.) + 3600.;
1228 }
1229
1230 ev_periodic_init (&amp;hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
1231
1232</pre>
1233<p>Example: Call a callback every hour, starting now:</p>
1234<pre> struct ev_periodic hourly_tick;
1235 ev_periodic_init (&amp;hourly_tick, clock_cb,
1236 fmod (ev_now (loop), 3600.), 3600., 0);
1237 ev_periodic_start (loop, &amp;hourly_tick);
1238
1239
1240
1241
1242</pre>
1243
606</div> 1244</div>
607<h2 id="code_ev_signal_code_signal_me_when_a"><code>ev_signal</code> - signal me when a signal gets signalled</h2> 1245<h2 id="code_ev_signal_code_signal_me_when_a"><code>ev_signal</code> - signal me when a signal gets signalled!</h2>
608<div id="code_ev_signal_code_signal_me_when_a-2"> 1246<div id="code_ev_signal_code_signal_me_when_a-2">
609<p>Signal watchers will trigger an event when the process receives a specific 1247<p>Signal watchers will trigger an event when the process receives a specific
610signal one or more times. Even though signals are very asynchronous, libev 1248signal one or more times. Even though signals are very asynchronous, libev
611will try it's best to deliver signals synchronously, i.e. as part of the 1249will try it's best to deliver signals synchronously, i.e. as part of the
612normal event processing, like any other event.</p> 1250normal event processing, like any other event.</p>
621 <dt>ev_signal_set (ev_signal *, int signum)</dt> 1259 <dt>ev_signal_set (ev_signal *, int signum)</dt>
622 <dd> 1260 <dd>
623 <p>Configures the watcher to trigger on the given signal number (usually one 1261 <p>Configures the watcher to trigger on the given signal number (usually one
624of the <code>SIGxxx</code> constants).</p> 1262of the <code>SIGxxx</code> constants).</p>
625 </dd> 1263 </dd>
1264 <dt>int signum [read-only]</dt>
1265 <dd>
1266 <p>The signal the watcher watches out for.</p>
1267 </dd>
626</dl> 1268</dl>
627 1269
1270
1271
1272
1273
628</div> 1274</div>
629<h2 id="code_ev_child_code_wait_for_pid_stat"><code>ev_child</code> - wait for pid status changes</h2> 1275<h2 id="code_ev_child_code_watch_out_for_pro"><code>ev_child</code> - watch out for process status changes</h2>
630<div id="code_ev_child_code_wait_for_pid_stat-2"> 1276<div id="code_ev_child_code_watch_out_for_pro-2">
631<p>Child watchers trigger when your process receives a SIGCHLD in response to 1277<p>Child watchers trigger when your process receives a SIGCHLD in response to
632some child status changes (most typically when a child of yours dies).</p> 1278some child status changes (most typically when a child of yours dies).</p>
633<dl> 1279<dl>
634 <dt>ev_child_init (ev_child *, callback, int pid)</dt> 1280 <dt>ev_child_init (ev_child *, callback, int pid)</dt>
635 <dt>ev_child_set (ev_child *, int pid)</dt> 1281 <dt>ev_child_set (ev_child *, int pid)</dt>
639at the <code>rstatus</code> member of the <code>ev_child</code> watcher structure to see 1285at the <code>rstatus</code> member of the <code>ev_child</code> watcher structure to see
640the status word (use the macros from <code>sys/wait.h</code> and see your systems 1286the status word (use the macros from <code>sys/wait.h</code> and see your systems
641<code>waitpid</code> documentation). The <code>rpid</code> member contains the pid of the 1287<code>waitpid</code> documentation). The <code>rpid</code> member contains the pid of the
642process causing the status change.</p> 1288process causing the status change.</p>
643 </dd> 1289 </dd>
1290 <dt>int pid [read-only]</dt>
1291 <dd>
1292 <p>The process id this watcher watches out for, or <code>0</code>, meaning any process id.</p>
1293 </dd>
1294 <dt>int rpid [read-write]</dt>
1295 <dd>
1296 <p>The process id that detected a status change.</p>
1297 </dd>
1298 <dt>int rstatus [read-write]</dt>
1299 <dd>
1300 <p>The process exit/trace status caused by <code>rpid</code> (see your systems
1301<code>waitpid</code> and <code>sys/wait.h</code> documentation for details).</p>
1302 </dd>
644</dl> 1303</dl>
1304<p>Example: Try to exit cleanly on SIGINT and SIGTERM.</p>
1305<pre> static void
1306 sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)
1307 {
1308 ev_unloop (loop, EVUNLOOP_ALL);
1309 }
645 1310
1311 struct ev_signal signal_watcher;
1312 ev_signal_init (&amp;signal_watcher, sigint_cb, SIGINT);
1313 ev_signal_start (loop, &amp;sigint_cb);
1314
1315
1316
1317
1318</pre>
1319
646</div> 1320</div>
1321<h2 id="code_ev_stat_code_did_the_file_attri"><code>ev_stat</code> - did the file attributes just change?</h2>
1322<div id="code_ev_stat_code_did_the_file_attri-2">
1323<p>This watches a filesystem path for attribute changes. That is, it calls
1324<code>stat</code> regularly (or when the OS says it changed) and sees if it changed
1325compared to the last time, invoking the callback if it did.</p>
1326<p>The path does not need to exist: changing from &quot;path exists&quot; to &quot;path does
1327not exist&quot; is a status change like any other. The condition &quot;path does
1328not exist&quot; is signified by the <code>st_nlink</code> field being zero (which is
1329otherwise always forced to be at least one) and all the other fields of
1330the stat buffer having unspecified contents.</p>
1331<p>The path <i>should</i> be absolute and <i>must not</i> end in a slash. If it is
1332relative and your working directory changes, the behaviour is undefined.</p>
1333<p>Since there is no standard to do this, the portable implementation simply
1334calls <code>stat (2)</code> regularly on the path to see if it changed somehow. You
1335can specify a recommended polling interval for this case. If you specify
1336a polling interval of <code>0</code> (highly recommended!) then a <i>suitable,
1337unspecified default</i> value will be used (which you can expect to be around
1338five seconds, although this might change dynamically). Libev will also
1339impose a minimum interval which is currently around <code>0.1</code>, but thats
1340usually overkill.</p>
1341<p>This watcher type is not meant for massive numbers of stat watchers,
1342as even with OS-supported change notifications, this can be
1343resource-intensive.</p>
1344<p>At the time of this writing, only the Linux inotify interface is
1345implemented (implementing kqueue support is left as an exercise for the
1346reader). Inotify will be used to give hints only and should not change the
1347semantics of <code>ev_stat</code> watchers, which means that libev sometimes needs
1348to fall back to regular polling again even with inotify, but changes are
1349usually detected immediately, and if the file exists there will be no
1350polling.</p>
1351<dl>
1352 <dt>ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)</dt>
1353 <dt>ev_stat_set (ev_stat *, const char *path, ev_tstamp interval)</dt>
1354 <dd>
1355 <p>Configures the watcher to wait for status changes of the given
1356<code>path</code>. The <code>interval</code> is a hint on how quickly a change is expected to
1357be detected and should normally be specified as <code>0</code> to let libev choose
1358a suitable value. The memory pointed to by <code>path</code> must point to the same
1359path for as long as the watcher is active.</p>
1360 <p>The callback will be receive <code>EV_STAT</code> when a change was detected,
1361relative to the attributes at the time the watcher was started (or the
1362last change was detected).</p>
1363 </dd>
1364 <dt>ev_stat_stat (ev_stat *)</dt>
1365 <dd>
1366 <p>Updates the stat buffer immediately with new values. If you change the
1367watched path in your callback, you could call this fucntion to avoid
1368detecting this change (while introducing a race condition). Can also be
1369useful simply to find out the new values.</p>
1370 </dd>
1371 <dt>ev_statdata attr [read-only]</dt>
1372 <dd>
1373 <p>The most-recently detected attributes of the file. Although the type is of
1374<code>ev_statdata</code>, this is usually the (or one of the) <code>struct stat</code> types
1375suitable for your system. If the <code>st_nlink</code> member is <code>0</code>, then there
1376was some error while <code>stat</code>ing the file.</p>
1377 </dd>
1378 <dt>ev_statdata prev [read-only]</dt>
1379 <dd>
1380 <p>The previous attributes of the file. The callback gets invoked whenever
1381<code>prev</code> != <code>attr</code>.</p>
1382 </dd>
1383 <dt>ev_tstamp interval [read-only]</dt>
1384 <dd>
1385 <p>The specified interval.</p>
1386 </dd>
1387 <dt>const char *path [read-only]</dt>
1388 <dd>
1389 <p>The filesystem path that is being watched.</p>
1390 </dd>
1391</dl>
1392<p>Example: Watch <code>/etc/passwd</code> for attribute changes.</p>
1393<pre> static void
1394 passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)
1395 {
1396 /* /etc/passwd changed in some way */
1397 if (w-&gt;attr.st_nlink)
1398 {
1399 printf (&quot;passwd current size %ld\n&quot;, (long)w-&gt;attr.st_size);
1400 printf (&quot;passwd current atime %ld\n&quot;, (long)w-&gt;attr.st_mtime);
1401 printf (&quot;passwd current mtime %ld\n&quot;, (long)w-&gt;attr.st_mtime);
1402 }
1403 else
1404 /* you shalt not abuse printf for puts */
1405 puts (&quot;wow, /etc/passwd is not there, expect problems. &quot;
1406 &quot;if this is windows, they already arrived\n&quot;);
1407 }
1408
1409 ...
1410 ev_stat passwd;
1411
1412 ev_stat_init (&amp;passwd, passwd_cb, &quot;/etc/passwd&quot;);
1413 ev_stat_start (loop, &amp;passwd);
1414
1415
1416
1417
1418</pre>
1419
1420</div>
647<h2 id="code_ev_idle_code_when_you_ve_got_no"><code>ev_idle</code> - when you've got nothing better to do</h2> 1421<h2 id="code_ev_idle_code_when_you_ve_got_no"><code>ev_idle</code> - when you've got nothing better to do...</h2>
648<div id="code_ev_idle_code_when_you_ve_got_no-2"> 1422<div id="code_ev_idle_code_when_you_ve_got_no-2">
649<p>Idle watchers trigger events when there are no other events are pending 1423<p>Idle watchers trigger events when no other events of the same or higher
650(prepare, check and other idle watchers do not count). That is, as long 1424priority are pending (prepare, check and other idle watchers do not
651as your process is busy handling sockets or timeouts (or even signals, 1425count).</p>
652imagine) it will not be triggered. But when your process is idle all idle 1426<p>That is, as long as your process is busy handling sockets or timeouts
653watchers are being called again and again, once per event loop iteration - 1427(or even signals, imagine) of the same or higher priority it will not be
1428triggered. But when your process is idle (or only lower-priority watchers
1429are pending), the idle watchers are being called once per event loop
654until stopped, that is, or your process receives more events and becomes 1430iteration - until stopped, that is, or your process receives more events
655busy.</p> 1431and becomes busy again with higher priority stuff.</p>
656<p>The most noteworthy effect is that as long as any idle watchers are 1432<p>The most noteworthy effect is that as long as any idle watchers are
657active, the process will not block when waiting for new events.</p> 1433active, the process will not block when waiting for new events.</p>
658<p>Apart from keeping your process non-blocking (which is a useful 1434<p>Apart from keeping your process non-blocking (which is a useful
659effect on its own sometimes), idle watchers are a good place to do 1435effect on its own sometimes), idle watchers are a good place to do
660&quot;pseudo-background processing&quot;, or delay processing stuff to after the 1436&quot;pseudo-background processing&quot;, or delay processing stuff to after the
665 <p>Initialises and configures the idle watcher - it has no parameters of any 1441 <p>Initialises and configures the idle watcher - it has no parameters of any
666kind. There is a <code>ev_idle_set</code> macro, but using it is utterly pointless, 1442kind. There is a <code>ev_idle_set</code> macro, but using it is utterly pointless,
667believe me.</p> 1443believe me.</p>
668 </dd> 1444 </dd>
669</dl> 1445</dl>
1446<p>Example: Dynamically allocate an <code>ev_idle</code> watcher, start it, and in the
1447callback, free it. Also, use no error checking, as usual.</p>
1448<pre> static void
1449 idle_cb (struct ev_loop *loop, struct ev_idle *w, int revents)
1450 {
1451 free (w);
1452 // now do something you wanted to do when the program has
1453 // no longer asnything immediate to do.
1454 }
670 1455
1456 struct ev_idle *idle_watcher = malloc (sizeof (struct ev_idle));
1457 ev_idle_init (idle_watcher, idle_cb);
1458 ev_idle_start (loop, idle_cb);
1459
1460
1461
1462
1463</pre>
1464
671</div> 1465</div>
672<h2 id="code_ev_prepare_code_and_code_ev_che"><code>ev_prepare</code> and <code>ev_check</code> - customise your event loop</h2> 1466<h2 id="code_ev_prepare_code_and_code_ev_che"><code>ev_prepare</code> and <code>ev_check</code> - customise your event loop!</h2>
673<div id="code_ev_prepare_code_and_code_ev_che-2"> 1467<div id="code_ev_prepare_code_and_code_ev_che-2">
674<p>Prepare and check watchers are usually (but not always) used in tandem: 1468<p>Prepare and check watchers are usually (but not always) used in tandem:
675Prepare watchers get invoked before the process blocks and check watchers 1469prepare watchers get invoked before the process blocks and check watchers
676afterwards.</p> 1470afterwards.</p>
1471<p>You <i>must not</i> call <code>ev_loop</code> or similar functions that enter
1472the current event loop from either <code>ev_prepare</code> or <code>ev_check</code>
1473watchers. Other loops than the current one are fine, however. The
1474rationale behind this is that you do not need to check for recursion in
1475those watchers, i.e. the sequence will always be <code>ev_prepare</code>, blocking,
1476<code>ev_check</code> so if you have one watcher of each kind they will always be
1477called in pairs bracketing the blocking call.</p>
677<p>Their main purpose is to integrate other event mechanisms into libev. This 1478<p>Their main purpose is to integrate other event mechanisms into libev and
678could be used, for example, to track variable changes, implement your own 1479their use is somewhat advanced. This could be used, for example, to track
679watchers, integrate net-snmp or a coroutine library and lots more.</p> 1480variable changes, implement your own watchers, integrate net-snmp or a
1481coroutine library and lots more. They are also occasionally useful if
1482you cache some data and want to flush it before blocking (for example,
1483in X programs you might want to do an <code>XFlush ()</code> in an <code>ev_prepare</code>
1484watcher).</p>
680<p>This is done by examining in each prepare call which file descriptors need 1485<p>This is done by examining in each prepare call which file descriptors need
681to be watched by the other library, registering <code>ev_io</code> watchers for 1486to be watched by the other library, registering <code>ev_io</code> watchers for
682them and starting an <code>ev_timer</code> watcher for any timeouts (many libraries 1487them and starting an <code>ev_timer</code> watcher for any timeouts (many libraries
683provide just this functionality). Then, in the check watcher you check for 1488provide just this functionality). Then, in the check watcher you check for
684any events that occured (by checking the pending status of all watchers 1489any events that occured (by checking the pending status of all watchers
685and stopping them) and call back into the library. The I/O and timer 1490and stopping them) and call back into the library. The I/O and timer
686callbacks will never actually be called (but must be valid neverthelles, 1491callbacks will never actually be called (but must be valid nevertheless,
687because you never know, you know?).</p> 1492because you never know, you know?).</p>
688<p>As another example, the Perl Coro module uses these hooks to integrate 1493<p>As another example, the Perl Coro module uses these hooks to integrate
689coroutines into libev programs, by yielding to other active coroutines 1494coroutines into libev programs, by yielding to other active coroutines
690during each prepare and only letting the process block if no coroutines 1495during each prepare and only letting the process block if no coroutines
691are ready to run (its actually more complicated, it only runs coroutines 1496are ready to run (it's actually more complicated: it only runs coroutines
692with priority higher than the event loop and one lower priority once, 1497with priority higher than or equal to the event loop and one coroutine
693using idle watchers to keep the event loop from blocking if lower-priority 1498of lower priority, but only once, using idle watchers to keep the event
694coroutines exist, thus mapping low-priority coroutines to idle/background 1499loop from blocking if lower-priority coroutines are active, thus mapping
695tasks).</p> 1500low-priority coroutines to idle/background tasks).</p>
1501<p>It is recommended to give <code>ev_check</code> watchers highest (<code>EV_MAXPRI</code>)
1502priority, to ensure that they are being run before any other watchers
1503after the poll. Also, <code>ev_check</code> watchers (and <code>ev_prepare</code> watchers,
1504too) should not activate (&quot;feed&quot;) events into libev. While libev fully
1505supports this, they will be called before other <code>ev_check</code> watchers did
1506their job. As <code>ev_check</code> watchers are often used to embed other event
1507loops those other event loops might be in an unusable state until their
1508<code>ev_check</code> watcher ran (always remind yourself to coexist peacefully with
1509others).</p>
696<dl> 1510<dl>
697 <dt>ev_prepare_init (ev_prepare *, callback)</dt> 1511 <dt>ev_prepare_init (ev_prepare *, callback)</dt>
698 <dt>ev_check_init (ev_check *, callback)</dt> 1512 <dt>ev_check_init (ev_check *, callback)</dt>
699 <dd> 1513 <dd>
700 <p>Initialises and configures the prepare or check watcher - they have no 1514 <p>Initialises and configures the prepare or check watcher - they have no
701parameters of any kind. There are <code>ev_prepare_set</code> and <code>ev_check_set</code> 1515parameters of any kind. There are <code>ev_prepare_set</code> and <code>ev_check_set</code>
702macros, but using them is utterly, utterly and completely pointless.</p> 1516macros, but using them is utterly, utterly and completely pointless.</p>
703 </dd> 1517 </dd>
704</dl> 1518</dl>
1519<p>There are a number of principal ways to embed other event loops or modules
1520into libev. Here are some ideas on how to include libadns into libev
1521(there is a Perl module named <code>EV::ADNS</code> that does this, which you could
1522use for an actually working example. Another Perl module named <code>EV::Glib</code>
1523embeds a Glib main context into libev, and finally, <code>Glib::EV</code> embeds EV
1524into the Glib event loop).</p>
1525<p>Method 1: Add IO watchers and a timeout watcher in a prepare handler,
1526and in a check watcher, destroy them and call into libadns. What follows
1527is pseudo-code only of course. This requires you to either use a low
1528priority for the check watcher or use <code>ev_clear_pending</code> explicitly, as
1529the callbacks for the IO/timeout watchers might not have been called yet.</p>
1530<pre> static ev_io iow [nfd];
1531 static ev_timer tw;
705 1532
1533 static void
1534 io_cb (ev_loop *loop, ev_io *w, int revents)
1535 {
1536 }
1537
1538 // create io watchers for each fd and a timer before blocking
1539 static void
1540 adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents)
1541 {
1542 int timeout = 3600000;
1543 struct pollfd fds [nfd];
1544 // actual code will need to loop here and realloc etc.
1545 adns_beforepoll (ads, fds, &amp;nfd, &amp;timeout, timeval_from (ev_time ()));
1546
1547 /* the callback is illegal, but won't be called as we stop during check */
1548 ev_timer_init (&amp;tw, 0, timeout * 1e-3);
1549 ev_timer_start (loop, &amp;tw);
1550
1551 // create one ev_io per pollfd
1552 for (int i = 0; i &lt; nfd; ++i)
1553 {
1554 ev_io_init (iow + i, io_cb, fds [i].fd,
1555 ((fds [i].events &amp; POLLIN ? EV_READ : 0)
1556 | (fds [i].events &amp; POLLOUT ? EV_WRITE : 0)));
1557
1558 fds [i].revents = 0;
1559 ev_io_start (loop, iow + i);
1560 }
1561 }
1562
1563 // stop all watchers after blocking
1564 static void
1565 adns_check_cb (ev_loop *loop, ev_check *w, int revents)
1566 {
1567 ev_timer_stop (loop, &amp;tw);
1568
1569 for (int i = 0; i &lt; nfd; ++i)
1570 {
1571 // set the relevant poll flags
1572 // could also call adns_processreadable etc. here
1573 struct pollfd *fd = fds + i;
1574 int revents = ev_clear_pending (iow + i);
1575 if (revents &amp; EV_READ ) fd-&gt;revents |= fd-&gt;events &amp; POLLIN;
1576 if (revents &amp; EV_WRITE) fd-&gt;revents |= fd-&gt;events &amp; POLLOUT;
1577
1578 // now stop the watcher
1579 ev_io_stop (loop, iow + i);
1580 }
1581
1582 adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));
1583 }
1584
1585</pre>
1586<p>Method 2: This would be just like method 1, but you run <code>adns_afterpoll</code>
1587in the prepare watcher and would dispose of the check watcher.</p>
1588<p>Method 3: If the module to be embedded supports explicit event
1589notification (adns does), you can also make use of the actual watcher
1590callbacks, and only destroy/create the watchers in the prepare watcher.</p>
1591<pre> static void
1592 timer_cb (EV_P_ ev_timer *w, int revents)
1593 {
1594 adns_state ads = (adns_state)w-&gt;data;
1595 update_now (EV_A);
1596
1597 adns_processtimeouts (ads, &amp;tv_now);
1598 }
1599
1600 static void
1601 io_cb (EV_P_ ev_io *w, int revents)
1602 {
1603 adns_state ads = (adns_state)w-&gt;data;
1604 update_now (EV_A);
1605
1606 if (revents &amp; EV_READ ) adns_processreadable (ads, w-&gt;fd, &amp;tv_now);
1607 if (revents &amp; EV_WRITE) adns_processwriteable (ads, w-&gt;fd, &amp;tv_now);
1608 }
1609
1610 // do not ever call adns_afterpoll
1611
1612</pre>
1613<p>Method 4: Do not use a prepare or check watcher because the module you
1614want to embed is too inflexible to support it. Instead, youc na override
1615their poll function. The drawback with this solution is that the main
1616loop is now no longer controllable by EV. The <code>Glib::EV</code> module does
1617this.</p>
1618<pre> static gint
1619 event_poll_func (GPollFD *fds, guint nfds, gint timeout)
1620 {
1621 int got_events = 0;
1622
1623 for (n = 0; n &lt; nfds; ++n)
1624 // create/start io watcher that sets the relevant bits in fds[n] and increment got_events
1625
1626 if (timeout &gt;= 0)
1627 // create/start timer
1628
1629 // poll
1630 ev_loop (EV_A_ 0);
1631
1632 // stop timer again
1633 if (timeout &gt;= 0)
1634 ev_timer_stop (EV_A_ &amp;to);
1635
1636 // stop io watchers again - their callbacks should have set
1637 for (n = 0; n &lt; nfds; ++n)
1638 ev_io_stop (EV_A_ iow [n]);
1639
1640 return got_events;
1641 }
1642
1643
1644
1645
1646</pre>
1647
706</div> 1648</div>
707<h1 id="OTHER_FUNCTIONS">OTHER FUNCTIONS</h1><p><a href="#TOP" class="toplink">Top</a></p> 1649<h2 id="code_ev_embed_code_when_one_backend_"><code>ev_embed</code> - when one backend isn't enough...</h2>
1650<div id="code_ev_embed_code_when_one_backend_-2">
1651<p>This is a rather advanced watcher type that lets you embed one event loop
1652into another (currently only <code>ev_io</code> events are supported in the embedded
1653loop, other types of watchers might be handled in a delayed or incorrect
1654fashion and must not be used).</p>
1655<p>There are primarily two reasons you would want that: work around bugs and
1656prioritise I/O.</p>
1657<p>As an example for a bug workaround, the kqueue backend might only support
1658sockets on some platform, so it is unusable as generic backend, but you
1659still want to make use of it because you have many sockets and it scales
1660so nicely. In this case, you would create a kqueue-based loop and embed it
1661into your default loop (which might use e.g. poll). Overall operation will
1662be a bit slower because first libev has to poll and then call kevent, but
1663at least you can use both at what they are best.</p>
1664<p>As for prioritising I/O: rarely you have the case where some fds have
1665to be watched and handled very quickly (with low latency), and even
1666priorities and idle watchers might have too much overhead. In this case
1667you would put all the high priority stuff in one loop and all the rest in
1668a second one, and embed the second one in the first.</p>
1669<p>As long as the watcher is active, the callback will be invoked every time
1670there might be events pending in the embedded loop. The callback must then
1671call <code>ev_embed_sweep (mainloop, watcher)</code> to make a single sweep and invoke
1672their callbacks (you could also start an idle watcher to give the embedded
1673loop strictly lower priority for example). You can also set the callback
1674to <code>0</code>, in which case the embed watcher will automatically execute the
1675embedded loop sweep.</p>
1676<p>As long as the watcher is started it will automatically handle events. The
1677callback will be invoked whenever some events have been handled. You can
1678set the callback to <code>0</code> to avoid having to specify one if you are not
1679interested in that.</p>
1680<p>Also, there have not currently been made special provisions for forking:
1681when you fork, you not only have to call <code>ev_loop_fork</code> on both loops,
1682but you will also have to stop and restart any <code>ev_embed</code> watchers
1683yourself.</p>
1684<p>Unfortunately, not all backends are embeddable, only the ones returned by
1685<code>ev_embeddable_backends</code> are, which, unfortunately, does not include any
1686portable one.</p>
1687<p>So when you want to use this feature you will always have to be prepared
1688that you cannot get an embeddable loop. The recommended way to get around
1689this is to have a separate variables for your embeddable loop, try to
1690create it, and if that fails, use the normal loop for everything:</p>
1691<pre> struct ev_loop *loop_hi = ev_default_init (0);
1692 struct ev_loop *loop_lo = 0;
1693 struct ev_embed embed;
1694
1695 // see if there is a chance of getting one that works
1696 // (remember that a flags value of 0 means autodetection)
1697 loop_lo = ev_embeddable_backends () &amp; ev_recommended_backends ()
1698 ? ev_loop_new (ev_embeddable_backends () &amp; ev_recommended_backends ())
1699 : 0;
1700
1701 // if we got one, then embed it, otherwise default to loop_hi
1702 if (loop_lo)
1703 {
1704 ev_embed_init (&amp;embed, 0, loop_lo);
1705 ev_embed_start (loop_hi, &amp;embed);
1706 }
1707 else
1708 loop_lo = loop_hi;
1709
1710</pre>
1711<dl>
1712 <dt>ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)</dt>
1713 <dt>ev_embed_set (ev_embed *, callback, struct ev_loop *embedded_loop)</dt>
1714 <dd>
1715 <p>Configures the watcher to embed the given loop, which must be
1716embeddable. If the callback is <code>0</code>, then <code>ev_embed_sweep</code> will be
1717invoked automatically, otherwise it is the responsibility of the callback
1718to invoke it (it will continue to be called until the sweep has been done,
1719if you do not want thta, you need to temporarily stop the embed watcher).</p>
1720 </dd>
1721 <dt>ev_embed_sweep (loop, ev_embed *)</dt>
1722 <dd>
1723 <p>Make a single, non-blocking sweep over the embedded loop. This works
1724similarly to <code>ev_loop (embedded_loop, EVLOOP_NONBLOCK)</code>, but in the most
1725apropriate way for embedded loops.</p>
1726 </dd>
1727 <dt>struct ev_loop *loop [read-only]</dt>
1728 <dd>
1729 <p>The embedded event loop.</p>
1730 </dd>
1731</dl>
1732
1733
1734
1735
1736
1737</div>
1738<h2 id="code_ev_fork_code_the_audacity_to_re"><code>ev_fork</code> - the audacity to resume the event loop after a fork</h2>
1739<div id="code_ev_fork_code_the_audacity_to_re-2">
1740<p>Fork watchers are called when a <code>fork ()</code> was detected (usually because
1741whoever is a good citizen cared to tell libev about it by calling
1742<code>ev_default_fork</code> or <code>ev_loop_fork</code>). The invocation is done before the
1743event loop blocks next and before <code>ev_check</code> watchers are being called,
1744and only in the child after the fork. If whoever good citizen calling
1745<code>ev_default_fork</code> cheats and calls it in the wrong process, the fork
1746handlers will be invoked, too, of course.</p>
1747<dl>
1748 <dt>ev_fork_init (ev_signal *, callback)</dt>
1749 <dd>
1750 <p>Initialises and configures the fork watcher - it has no parameters of any
1751kind. There is a <code>ev_fork_set</code> macro, but using it is utterly pointless,
1752believe me.</p>
1753 </dd>
1754</dl>
1755
1756
1757
1758
1759
1760</div>
1761<h1 id="OTHER_FUNCTIONS">OTHER FUNCTIONS</h1>
708<div id="OTHER_FUNCTIONS_CONTENT"> 1762<div id="OTHER_FUNCTIONS_CONTENT">
709<p>There are some other functions of possible interest. Described. Here. Now.</p> 1763<p>There are some other functions of possible interest. Described. Here. Now.</p>
710<dl> 1764<dl>
711 <dt>ev_once (loop, int fd, int events, ev_tstamp timeout, callback)</dt> 1765 <dt>ev_once (loop, int fd, int events, ev_tstamp timeout, callback)</dt>
712 <dd> 1766 <dd>
713 <p>This function combines a simple timer and an I/O watcher, calls your 1767 <p>This function combines a simple timer and an I/O watcher, calls your
714callback on whichever event happens first and automatically stop both 1768callback on whichever event happens first and automatically stop both
715watchers. This is useful if you want to wait for a single event on an fd 1769watchers. This is useful if you want to wait for a single event on an fd
716or timeout without havign to allocate/configure/start/stop/free one or 1770or timeout without having to allocate/configure/start/stop/free one or
717more watchers yourself.</p> 1771more watchers yourself.</p>
718 <p>If <code>fd</code> is less than 0, then no I/O watcher will be started and events 1772 <p>If <code>fd</code> is less than 0, then no I/O watcher will be started and events
719is being ignored. Otherwise, an <code>ev_io</code> watcher for the given <code>fd</code> and 1773is being ignored. Otherwise, an <code>ev_io</code> watcher for the given <code>fd</code> and
720<code>events</code> set will be craeted and started.</p> 1774<code>events</code> set will be craeted and started.</p>
721 <p>If <code>timeout</code> is less than 0, then no timeout watcher will be 1775 <p>If <code>timeout</code> is less than 0, then no timeout watcher will be
722started. Otherwise an <code>ev_timer</code> watcher with after = <code>timeout</code> (and 1776started. Otherwise an <code>ev_timer</code> watcher with after = <code>timeout</code> (and
723repeat = 0) will be started. While <code>0</code> is a valid timeout, it is of 1777repeat = 0) will be started. While <code>0</code> is a valid timeout, it is of
724dubious value.</p> 1778dubious value.</p>
725 <p>The callback has the type <code>void (*cb)(int revents, void *arg)</code> and gets 1779 <p>The callback has the type <code>void (*cb)(int revents, void *arg)</code> and gets
726passed an events set like normal event callbacks (with a combination of 1780passed an <code>revents</code> set like normal event callbacks (a combination of
727<code>EV_ERROR</code>, <code>EV_READ</code>, <code>EV_WRITE</code> or <code>EV_TIMEOUT</code>) and the <code>arg</code> 1781<code>EV_ERROR</code>, <code>EV_READ</code>, <code>EV_WRITE</code> or <code>EV_TIMEOUT</code>) and the <code>arg</code>
728value passed to <code>ev_once</code>:</p> 1782value passed to <code>ev_once</code>:</p>
729<pre> static void stdin_ready (int revents, void *arg) 1783<pre> static void stdin_ready (int revents, void *arg)
730 { 1784 {
731 if (revents &amp; EV_TIMEOUT) 1785 if (revents &amp; EV_TIMEOUT)
736 1790
737 ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0); 1791 ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
738 1792
739</pre> 1793</pre>
740 </dd> 1794 </dd>
741 <dt>ev_feed_event (loop, watcher, int events)</dt> 1795 <dt>ev_feed_event (ev_loop *, watcher *, int revents)</dt>
742 <dd> 1796 <dd>
743 <p>Feeds the given event set into the event loop, as if the specified event 1797 <p>Feeds the given event set into the event loop, as if the specified event
744had happened for the specified watcher (which must be a pointer to an 1798had happened for the specified watcher (which must be a pointer to an
745initialised but not necessarily started event watcher).</p> 1799initialised but not necessarily started event watcher).</p>
746 </dd> 1800 </dd>
747 <dt>ev_feed_fd_event (loop, int fd, int revents)</dt> 1801 <dt>ev_feed_fd_event (ev_loop *, int fd, int revents)</dt>
748 <dd> 1802 <dd>
749 <p>Feed an event on the given fd, as if a file descriptor backend detected 1803 <p>Feed an event on the given fd, as if a file descriptor backend detected
750the given events it.</p> 1804the given events it.</p>
751 </dd> 1805 </dd>
752 <dt>ev_feed_signal_event (loop, int signum)</dt> 1806 <dt>ev_feed_signal_event (ev_loop *loop, int signum)</dt>
753 <dd> 1807 <dd>
754 <p>Feed an event as if the given signal occured (loop must be the default loop!).</p> 1808 <p>Feed an event as if the given signal occured (<code>loop</code> must be the default
1809loop!).</p>
755 </dd> 1810 </dd>
756</dl> 1811</dl>
757 1812
1813
1814
1815
1816
758</div> 1817</div>
759<h1 id="AUTHOR">AUTHOR</h1><p><a href="#TOP" class="toplink">Top</a></p> 1818<h1 id="LIBEVENT_EMULATION">LIBEVENT EMULATION</h1>
1819<div id="LIBEVENT_EMULATION_CONTENT">
1820<p>Libev offers a compatibility emulation layer for libevent. It cannot
1821emulate the internals of libevent, so here are some usage hints:</p>
1822<dl>
1823 <dt>* Use it by including &lt;event.h&gt;, as usual.</dt>
1824 <dt>* The following members are fully supported: ev_base, ev_callback,
1825ev_arg, ev_fd, ev_res, ev_events.</dt>
1826 <dt>* Avoid using ev_flags and the EVLIST_*-macros, while it is
1827maintained by libev, it does not work exactly the same way as in libevent (consider
1828it a private API).</dt>
1829 <dt>* Priorities are not currently supported. Initialising priorities
1830will fail and all watchers will have the same priority, even though there
1831is an ev_pri field.</dt>
1832 <dt>* Other members are not supported.</dt>
1833 <dt>* The libev emulation is <i>not</i> ABI compatible to libevent, you need
1834to use the libev header file and library.</dt>
1835</dl>
1836
1837</div>
1838<h1 id="C_SUPPORT">C++ SUPPORT</h1>
1839<div id="C_SUPPORT_CONTENT">
1840<p>Libev comes with some simplistic wrapper classes for C++ that mainly allow
1841you to use some convinience methods to start/stop watchers and also change
1842the callback model to a model using method callbacks on objects.</p>
1843<p>To use it,</p>
1844<pre> #include &lt;ev++.h&gt;
1845
1846</pre>
1847<p>This automatically includes <cite>ev.h</cite> and puts all of its definitions (many
1848of them macros) into the global namespace. All C++ specific things are
1849put into the <code>ev</code> namespace. It should support all the same embedding
1850options as <cite>ev.h</cite>, most notably <code>EV_MULTIPLICITY</code>.</p>
1851<p>Care has been taken to keep the overhead low. The only data member the C++
1852classes add (compared to plain C-style watchers) is the event loop pointer
1853that the watcher is associated with (or no additional members at all if
1854you disable <code>EV_MULTIPLICITY</code> when embedding libev).</p>
1855<p>Currently, functions, and static and non-static member functions can be
1856used as callbacks. Other types should be easy to add as long as they only
1857need one additional pointer for context. If you need support for other
1858types of functors please contact the author (preferably after implementing
1859it).</p>
1860<p>Here is a list of things available in the <code>ev</code> namespace:</p>
1861<dl>
1862 <dt><code>ev::READ</code>, <code>ev::WRITE</code> etc.</dt>
1863 <dd>
1864 <p>These are just enum values with the same values as the <code>EV_READ</code> etc.
1865macros from <cite>ev.h</cite>.</p>
1866 </dd>
1867 <dt><code>ev::tstamp</code>, <code>ev::now</code></dt>
1868 <dd>
1869 <p>Aliases to the same types/functions as with the <code>ev_</code> prefix.</p>
1870 </dd>
1871 <dt><code>ev::io</code>, <code>ev::timer</code>, <code>ev::periodic</code>, <code>ev::idle</code>, <code>ev::sig</code> etc.</dt>
1872 <dd>
1873 <p>For each <code>ev_TYPE</code> watcher in <cite>ev.h</cite> there is a corresponding class of
1874the same name in the <code>ev</code> namespace, with the exception of <code>ev_signal</code>
1875which is called <code>ev::sig</code> to avoid clashes with the <code>signal</code> macro
1876defines by many implementations.</p>
1877 <p>All of those classes have these methods:</p>
1878 <p>
1879 <dl>
1880 <dt>ev::TYPE::TYPE ()</dt>
1881 <dt>ev::TYPE::TYPE (struct ev_loop *)</dt>
1882 <dt>ev::TYPE::~TYPE</dt>
1883 <dd>
1884 <p>The constructor (optionally) takes an event loop to associate the watcher
1885with. If it is omitted, it will use <code>EV_DEFAULT</code>.</p>
1886 <p>The constructor calls <code>ev_init</code> for you, which means you have to call the
1887<code>set</code> method before starting it.</p>
1888 <p>It will not set a callback, however: You have to call the templated <code>set</code>
1889method to set a callback before you can start the watcher.</p>
1890 <p>(The reason why you have to use a method is a limitation in C++ which does
1891not allow explicit template arguments for constructors).</p>
1892 <p>The destructor automatically stops the watcher if it is active.</p>
1893 </dd>
1894 <dt>w-&gt;set&lt;class, &amp;class::method&gt; (object *)</dt>
1895 <dd>
1896 <p>This method sets the callback method to call. The method has to have a
1897signature of <code>void (*)(ev_TYPE &amp;, int)</code>, it receives the watcher as
1898first argument and the <code>revents</code> as second. The object must be given as
1899parameter and is stored in the <code>data</code> member of the watcher.</p>
1900 <p>This method synthesizes efficient thunking code to call your method from
1901the C callback that libev requires. If your compiler can inline your
1902callback (i.e. it is visible to it at the place of the <code>set</code> call and
1903your compiler is good :), then the method will be fully inlined into the
1904thunking function, making it as fast as a direct C callback.</p>
1905 <p>Example: simple class declaration and watcher initialisation</p>
1906<pre> struct myclass
1907 {
1908 void io_cb (ev::io &amp;w, int revents) { }
1909 }
1910
1911 myclass obj;
1912 ev::io iow;
1913 iow.set &lt;myclass, &amp;myclass::io_cb&gt; (&amp;obj);
1914
1915</pre>
1916 </dd>
1917 <dt>w-&gt;set&lt;function&gt; (void *data = 0)</dt>
1918 <dd>
1919 <p>Also sets a callback, but uses a static method or plain function as
1920callback. The optional <code>data</code> argument will be stored in the watcher's
1921<code>data</code> member and is free for you to use.</p>
1922 <p>The prototype of the <code>function</code> must be <code>void (*)(ev::TYPE &amp;w, int)</code>.</p>
1923 <p>See the method-<code>set</code> above for more details.</p>
1924 <p>Example:</p>
1925<pre> static void io_cb (ev::io &amp;w, int revents) { }
1926 iow.set &lt;io_cb&gt; ();
1927
1928</pre>
1929 </dd>
1930 <dt>w-&gt;set (struct ev_loop *)</dt>
1931 <dd>
1932 <p>Associates a different <code>struct ev_loop</code> with this watcher. You can only
1933do this when the watcher is inactive (and not pending either).</p>
1934 </dd>
1935 <dt>w-&gt;set ([args])</dt>
1936 <dd>
1937 <p>Basically the same as <code>ev_TYPE_set</code>, with the same args. Must be
1938called at least once. Unlike the C counterpart, an active watcher gets
1939automatically stopped and restarted when reconfiguring it with this
1940method.</p>
1941 </dd>
1942 <dt>w-&gt;start ()</dt>
1943 <dd>
1944 <p>Starts the watcher. Note that there is no <code>loop</code> argument, as the
1945constructor already stores the event loop.</p>
1946 </dd>
1947 <dt>w-&gt;stop ()</dt>
1948 <dd>
1949 <p>Stops the watcher if it is active. Again, no <code>loop</code> argument.</p>
1950 </dd>
1951 <dt>w-&gt;again () <code>ev::timer</code>, <code>ev::periodic</code> only</dt>
1952 <dd>
1953 <p>For <code>ev::timer</code> and <code>ev::periodic</code>, this invokes the corresponding
1954<code>ev_TYPE_again</code> function.</p>
1955 </dd>
1956 <dt>w-&gt;sweep () <code>ev::embed</code> only</dt>
1957 <dd>
1958 <p>Invokes <code>ev_embed_sweep</code>.</p>
1959 </dd>
1960 <dt>w-&gt;update () <code>ev::stat</code> only</dt>
1961 <dd>
1962 <p>Invokes <code>ev_stat_stat</code>.</p>
1963 </dd>
1964 </dl>
1965 </p>
1966 </dd>
1967</dl>
1968<p>Example: Define a class with an IO and idle watcher, start one of them in
1969the constructor.</p>
1970<pre> class myclass
1971 {
1972 ev_io io; void io_cb (ev::io &amp;w, int revents);
1973 ev_idle idle void idle_cb (ev::idle &amp;w, int revents);
1974
1975 myclass ();
1976 }
1977
1978 myclass::myclass (int fd)
1979 {
1980 io .set &lt;myclass, &amp;myclass::io_cb &gt; (this);
1981 idle.set &lt;myclass, &amp;myclass::idle_cb&gt; (this);
1982
1983 io.start (fd, ev::READ);
1984 }
1985
1986
1987
1988
1989</pre>
1990
1991</div>
1992<h1 id="MACRO_MAGIC">MACRO MAGIC</h1>
1993<div id="MACRO_MAGIC_CONTENT">
1994<p>Libev can be compiled with a variety of options, the most fundemantal is
1995<code>EV_MULTIPLICITY</code>. This option determines whether (most) functions and
1996callbacks have an initial <code>struct ev_loop *</code> argument.</p>
1997<p>To make it easier to write programs that cope with either variant, the
1998following macros are defined:</p>
1999<dl>
2000 <dt><code>EV_A</code>, <code>EV_A_</code></dt>
2001 <dd>
2002 <p>This provides the loop <i>argument</i> for functions, if one is required (&quot;ev
2003loop argument&quot;). The <code>EV_A</code> form is used when this is the sole argument,
2004<code>EV_A_</code> is used when other arguments are following. Example:</p>
2005<pre> ev_unref (EV_A);
2006 ev_timer_add (EV_A_ watcher);
2007 ev_loop (EV_A_ 0);
2008
2009</pre>
2010 <p>It assumes the variable <code>loop</code> of type <code>struct ev_loop *</code> is in scope,
2011which is often provided by the following macro.</p>
2012 </dd>
2013 <dt><code>EV_P</code>, <code>EV_P_</code></dt>
2014 <dd>
2015 <p>This provides the loop <i>parameter</i> for functions, if one is required (&quot;ev
2016loop parameter&quot;). The <code>EV_P</code> form is used when this is the sole parameter,
2017<code>EV_P_</code> is used when other parameters are following. Example:</p>
2018<pre> // this is how ev_unref is being declared
2019 static void ev_unref (EV_P);
2020
2021 // this is how you can declare your typical callback
2022 static void cb (EV_P_ ev_timer *w, int revents)
2023
2024</pre>
2025 <p>It declares a parameter <code>loop</code> of type <code>struct ev_loop *</code>, quite
2026suitable for use with <code>EV_A</code>.</p>
2027 </dd>
2028 <dt><code>EV_DEFAULT</code>, <code>EV_DEFAULT_</code></dt>
2029 <dd>
2030 <p>Similar to the other two macros, this gives you the value of the default
2031loop, if multiple loops are supported (&quot;ev loop default&quot;).</p>
2032 </dd>
2033</dl>
2034<p>Example: Declare and initialise a check watcher, utilising the above
2035macros so it will work regardless of whether multiple loops are supported
2036or not.</p>
2037<pre> static void
2038 check_cb (EV_P_ ev_timer *w, int revents)
2039 {
2040 ev_check_stop (EV_A_ w);
2041 }
2042
2043 ev_check check;
2044 ev_check_init (&amp;check, check_cb);
2045 ev_check_start (EV_DEFAULT_ &amp;check);
2046 ev_loop (EV_DEFAULT_ 0);
2047
2048</pre>
2049
2050</div>
2051<h1 id="EMBEDDING">EMBEDDING</h1>
2052<div id="EMBEDDING_CONTENT">
2053<p>Libev can (and often is) directly embedded into host
2054applications. Examples of applications that embed it include the Deliantra
2055Game Server, the EV perl module, the GNU Virtual Private Ethernet (gvpe)
2056and rxvt-unicode.</p>
2057<p>The goal is to enable you to just copy the neecssary files into your
2058source directory without having to change even a single line in them, so
2059you can easily upgrade by simply copying (or having a checked-out copy of
2060libev somewhere in your source tree).</p>
2061
2062</div>
2063<h2 id="FILESETS">FILESETS</h2>
2064<div id="FILESETS_CONTENT">
2065<p>Depending on what features you need you need to include one or more sets of files
2066in your app.</p>
2067
2068</div>
2069<h3 id="CORE_EVENT_LOOP">CORE EVENT LOOP</h3>
2070<div id="CORE_EVENT_LOOP_CONTENT">
2071<p>To include only the libev core (all the <code>ev_*</code> functions), with manual
2072configuration (no autoconf):</p>
2073<pre> #define EV_STANDALONE 1
2074 #include &quot;ev.c&quot;
2075
2076</pre>
2077<p>This will automatically include <cite>ev.h</cite>, too, and should be done in a
2078single C source file only to provide the function implementations. To use
2079it, do the same for <cite>ev.h</cite> in all files wishing to use this API (best
2080done by writing a wrapper around <cite>ev.h</cite> that you can include instead and
2081where you can put other configuration options):</p>
2082<pre> #define EV_STANDALONE 1
2083 #include &quot;ev.h&quot;
2084
2085</pre>
2086<p>Both header files and implementation files can be compiled with a C++
2087compiler (at least, thats a stated goal, and breakage will be treated
2088as a bug).</p>
2089<p>You need the following files in your source tree, or in a directory
2090in your include path (e.g. in libev/ when using -Ilibev):</p>
2091<pre> ev.h
2092 ev.c
2093 ev_vars.h
2094 ev_wrap.h
2095
2096 ev_win32.c required on win32 platforms only
2097
2098 ev_select.c only when select backend is enabled (which is enabled by default)
2099 ev_poll.c only when poll backend is enabled (disabled by default)
2100 ev_epoll.c only when the epoll backend is enabled (disabled by default)
2101 ev_kqueue.c only when the kqueue backend is enabled (disabled by default)
2102 ev_port.c only when the solaris port backend is enabled (disabled by default)
2103
2104</pre>
2105<p><cite>ev.c</cite> includes the backend files directly when enabled, so you only need
2106to compile this single file.</p>
2107
2108</div>
2109<h3 id="LIBEVENT_COMPATIBILITY_API">LIBEVENT COMPATIBILITY API</h3>
2110<div id="LIBEVENT_COMPATIBILITY_API_CONTENT">
2111<p>To include the libevent compatibility API, also include:</p>
2112<pre> #include &quot;event.c&quot;
2113
2114</pre>
2115<p>in the file including <cite>ev.c</cite>, and:</p>
2116<pre> #include &quot;event.h&quot;
2117
2118</pre>
2119<p>in the files that want to use the libevent API. This also includes <cite>ev.h</cite>.</p>
2120<p>You need the following additional files for this:</p>
2121<pre> event.h
2122 event.c
2123
2124</pre>
2125
2126</div>
2127<h3 id="AUTOCONF_SUPPORT">AUTOCONF SUPPORT</h3>
2128<div id="AUTOCONF_SUPPORT_CONTENT">
2129<p>Instead of using <code>EV_STANDALONE=1</code> and providing your config in
2130whatever way you want, you can also <code>m4_include([libev.m4])</code> in your
2131<cite>configure.ac</cite> and leave <code>EV_STANDALONE</code> undefined. <cite>ev.c</cite> will then
2132include <cite>config.h</cite> and configure itself accordingly.</p>
2133<p>For this of course you need the m4 file:</p>
2134<pre> libev.m4
2135
2136</pre>
2137
2138</div>
2139<h2 id="PREPROCESSOR_SYMBOLS_MACROS">PREPROCESSOR SYMBOLS/MACROS</h2>
2140<div id="PREPROCESSOR_SYMBOLS_MACROS_CONTENT">
2141<p>Libev can be configured via a variety of preprocessor symbols you have to define
2142before including any of its files. The default is not to build for multiplicity
2143and only include the select backend.</p>
2144<dl>
2145 <dt>EV_STANDALONE</dt>
2146 <dd>
2147 <p>Must always be <code>1</code> if you do not use autoconf configuration, which
2148keeps libev from including <cite>config.h</cite>, and it also defines dummy
2149implementations for some libevent functions (such as logging, which is not
2150supported). It will also not define any of the structs usually found in
2151<cite>event.h</cite> that are not directly supported by the libev core alone.</p>
2152 </dd>
2153 <dt>EV_USE_MONOTONIC</dt>
2154 <dd>
2155 <p>If defined to be <code>1</code>, libev will try to detect the availability of the
2156monotonic clock option at both compiletime and runtime. Otherwise no use
2157of the monotonic clock option will be attempted. If you enable this, you
2158usually have to link against librt or something similar. Enabling it when
2159the functionality isn't available is safe, though, althoguh you have
2160to make sure you link against any libraries where the <code>clock_gettime</code>
2161function is hiding in (often <cite>-lrt</cite>).</p>
2162 </dd>
2163 <dt>EV_USE_REALTIME</dt>
2164 <dd>
2165 <p>If defined to be <code>1</code>, libev will try to detect the availability of the
2166realtime clock option at compiletime (and assume its availability at
2167runtime if successful). Otherwise no use of the realtime clock option will
2168be attempted. This effectively replaces <code>gettimeofday</code> by <code>clock_get
2169(CLOCK_REALTIME, ...)</code> and will not normally affect correctness. See tzhe note about libraries
2170in the description of <code>EV_USE_MONOTONIC</code>, though.</p>
2171 </dd>
2172 <dt>EV_USE_SELECT</dt>
2173 <dd>
2174 <p>If undefined or defined to be <code>1</code>, libev will compile in support for the
2175<code>select</code>(2) backend. No attempt at autodetection will be done: if no
2176other method takes over, select will be it. Otherwise the select backend
2177will not be compiled in.</p>
2178 </dd>
2179 <dt>EV_SELECT_USE_FD_SET</dt>
2180 <dd>
2181 <p>If defined to <code>1</code>, then the select backend will use the system <code>fd_set</code>
2182structure. This is useful if libev doesn't compile due to a missing
2183<code>NFDBITS</code> or <code>fd_mask</code> definition or it misguesses the bitset layout on
2184exotic systems. This usually limits the range of file descriptors to some
2185low limit such as 1024 or might have other limitations (winsocket only
2186allows 64 sockets). The <code>FD_SETSIZE</code> macro, set before compilation, might
2187influence the size of the <code>fd_set</code> used.</p>
2188 </dd>
2189 <dt>EV_SELECT_IS_WINSOCKET</dt>
2190 <dd>
2191 <p>When defined to <code>1</code>, the select backend will assume that
2192select/socket/connect etc. don't understand file descriptors but
2193wants osf handles on win32 (this is the case when the select to
2194be used is the winsock select). This means that it will call
2195<code>_get_osfhandle</code> on the fd to convert it to an OS handle. Otherwise,
2196it is assumed that all these functions actually work on fds, even
2197on win32. Should not be defined on non-win32 platforms.</p>
2198 </dd>
2199 <dt>EV_USE_POLL</dt>
2200 <dd>
2201 <p>If defined to be <code>1</code>, libev will compile in support for the <code>poll</code>(2)
2202backend. Otherwise it will be enabled on non-win32 platforms. It
2203takes precedence over select.</p>
2204 </dd>
2205 <dt>EV_USE_EPOLL</dt>
2206 <dd>
2207 <p>If defined to be <code>1</code>, libev will compile in support for the Linux
2208<code>epoll</code>(7) backend. Its availability will be detected at runtime,
2209otherwise another method will be used as fallback. This is the
2210preferred backend for GNU/Linux systems.</p>
2211 </dd>
2212 <dt>EV_USE_KQUEUE</dt>
2213 <dd>
2214 <p>If defined to be <code>1</code>, libev will compile in support for the BSD style
2215<code>kqueue</code>(2) backend. Its actual availability will be detected at runtime,
2216otherwise another method will be used as fallback. This is the preferred
2217backend for BSD and BSD-like systems, although on most BSDs kqueue only
2218supports some types of fds correctly (the only platform we found that
2219supports ptys for example was NetBSD), so kqueue might be compiled in, but
2220not be used unless explicitly requested. The best way to use it is to find
2221out whether kqueue supports your type of fd properly and use an embedded
2222kqueue loop.</p>
2223 </dd>
2224 <dt>EV_USE_PORT</dt>
2225 <dd>
2226 <p>If defined to be <code>1</code>, libev will compile in support for the Solaris
222710 port style backend. Its availability will be detected at runtime,
2228otherwise another method will be used as fallback. This is the preferred
2229backend for Solaris 10 systems.</p>
2230 </dd>
2231 <dt>EV_USE_DEVPOLL</dt>
2232 <dd>
2233 <p>reserved for future expansion, works like the USE symbols above.</p>
2234 </dd>
2235 <dt>EV_USE_INOTIFY</dt>
2236 <dd>
2237 <p>If defined to be <code>1</code>, libev will compile in support for the Linux inotify
2238interface to speed up <code>ev_stat</code> watchers. Its actual availability will
2239be detected at runtime.</p>
2240 </dd>
2241 <dt>EV_H</dt>
2242 <dd>
2243 <p>The name of the <cite>ev.h</cite> header file used to include it. The default if
2244undefined 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
2245can be used to virtually rename the <cite>ev.h</cite> header file in case of conflicts.</p>
2246 </dd>
2247 <dt>EV_CONFIG_H</dt>
2248 <dd>
2249 <p>If <code>EV_STANDALONE</code> isn't <code>1</code>, this variable can be used to override
2250<cite>ev.c</cite>'s idea of where to find the <cite>config.h</cite> file, similarly to
2251<code>EV_H</code>, above.</p>
2252 </dd>
2253 <dt>EV_EVENT_H</dt>
2254 <dd>
2255 <p>Similarly to <code>EV_H</code>, this macro can be used to override <cite>event.c</cite>'s idea
2256of how the <cite>event.h</cite> header can be found.</p>
2257 </dd>
2258 <dt>EV_PROTOTYPES</dt>
2259 <dd>
2260 <p>If defined to be <code>0</code>, then <cite>ev.h</cite> will not define any function
2261prototypes, but still define all the structs and other symbols. This is
2262occasionally useful if you want to provide your own wrapper functions
2263around libev functions.</p>
2264 </dd>
2265 <dt>EV_MULTIPLICITY</dt>
2266 <dd>
2267 <p>If undefined or defined to <code>1</code>, then all event-loop-specific functions
2268will have the <code>struct ev_loop *</code> as first argument, and you can create
2269additional independent event loops. Otherwise there will be no support
2270for multiple event loops and there is no first event loop pointer
2271argument. Instead, all functions act on the single default loop.</p>
2272 </dd>
2273 <dt>EV_MINPRI</dt>
2274 <dt>EV_MAXPRI</dt>
2275 <dd>
2276 <p>The range of allowed priorities. <code>EV_MINPRI</code> must be smaller or equal to
2277<code>EV_MAXPRI</code>, but otherwise there are no non-obvious limitations. You can
2278provide for more priorities by overriding those symbols (usually defined
2279to be <code>-2</code> and <code>2</code>, respectively).</p>
2280 <p>When doing priority-based operations, libev usually has to linearly search
2281all the priorities, so having many of them (hundreds) uses a lot of space
2282and time, so using the defaults of five priorities (-2 .. +2) is usually
2283fine.</p>
2284 <p>If your embedding app does not need any priorities, defining these both to
2285<code>0</code> will save some memory and cpu.</p>
2286 </dd>
2287 <dt>EV_PERIODIC_ENABLE</dt>
2288 <dd>
2289 <p>If undefined or defined to be <code>1</code>, then periodic timers are supported. If
2290defined to be <code>0</code>, then they are not. Disabling them saves a few kB of
2291code.</p>
2292 </dd>
2293 <dt>EV_IDLE_ENABLE</dt>
2294 <dd>
2295 <p>If undefined or defined to be <code>1</code>, then idle watchers are supported. If
2296defined to be <code>0</code>, then they are not. Disabling them saves a few kB of
2297code.</p>
2298 </dd>
2299 <dt>EV_EMBED_ENABLE</dt>
2300 <dd>
2301 <p>If undefined or defined to be <code>1</code>, then embed watchers are supported. If
2302defined to be <code>0</code>, then they are not.</p>
2303 </dd>
2304 <dt>EV_STAT_ENABLE</dt>
2305 <dd>
2306 <p>If undefined or defined to be <code>1</code>, then stat watchers are supported. If
2307defined to be <code>0</code>, then they are not.</p>
2308 </dd>
2309 <dt>EV_FORK_ENABLE</dt>
2310 <dd>
2311 <p>If undefined or defined to be <code>1</code>, then fork watchers are supported. If
2312defined to be <code>0</code>, then they are not.</p>
2313 </dd>
2314 <dt>EV_MINIMAL</dt>
2315 <dd>
2316 <p>If you need to shave off some kilobytes of code at the expense of some
2317speed, define this symbol to <code>1</code>. Currently only used for gcc to override
2318some inlining decisions, saves roughly 30% codesize of amd64.</p>
2319 </dd>
2320 <dt>EV_PID_HASHSIZE</dt>
2321 <dd>
2322 <p><code>ev_child</code> watchers use a small hash table to distribute workload by
2323pid. The default size is <code>16</code> (or <code>1</code> with <code>EV_MINIMAL</code>), usually more
2324than enough. If you need to manage thousands of children you might want to
2325increase this value (<i>must</i> be a power of two).</p>
2326 </dd>
2327 <dt>EV_INOTIFY_HASHSIZE</dt>
2328 <dd>
2329 <p><code>ev_staz</code> watchers use a small hash table to distribute workload by
2330inotify watch id. The default size is <code>16</code> (or <code>1</code> with <code>EV_MINIMAL</code>),
2331usually more than enough. If you need to manage thousands of <code>ev_stat</code>
2332watchers you might want to increase this value (<i>must</i> be a power of
2333two).</p>
2334 </dd>
2335 <dt>EV_COMMON</dt>
2336 <dd>
2337 <p>By default, all watchers have a <code>void *data</code> member. By redefining
2338this macro to a something else you can include more and other types of
2339members. You have to define it each time you include one of the files,
2340though, and it must be identical each time.</p>
2341 <p>For example, the perl EV module uses something like this:</p>
2342<pre> #define EV_COMMON \
2343 SV *self; /* contains this struct */ \
2344 SV *cb_sv, *fh /* note no trailing &quot;;&quot; */
2345
2346</pre>
2347 </dd>
2348 <dt>EV_CB_DECLARE (type)</dt>
2349 <dt>EV_CB_INVOKE (watcher, revents)</dt>
2350 <dt>ev_set_cb (ev, cb)</dt>
2351 <dd>
2352 <p>Can be used to change the callback member declaration in each watcher,
2353and the way callbacks are invoked and set. Must expand to a struct member
2354definition and a statement, respectively. See the <cite>ev.v</cite> header file for
2355their default definitions. One possible use for overriding these is to
2356avoid the <code>struct ev_loop *</code> as first argument in all cases, or to use
2357method calls instead of plain function calls in C++.</p>
2358
2359</div>
2360<h2 id="EXAMPLES">EXAMPLES</h2>
2361<div id="EXAMPLES_CONTENT">
2362 <p>For a real-world example of a program the includes libev
2363verbatim, you can have a look at the EV perl module
2364(<a href="http://software.schmorp.de/pkg/EV.html">http://software.schmorp.de/pkg/EV.html</a>). It has the libev files in
2365the <cite>libev/</cite> subdirectory and includes them in the <cite>EV/EVAPI.h</cite> (public
2366interface) and <cite>EV.xs</cite> (implementation) files. Only the <cite>EV.xs</cite> file
2367will be compiled. It is pretty complex because it provides its own header
2368file.</p>
2369 <p>The usage in rxvt-unicode is simpler. It has a <cite>ev_cpp.h</cite> header file
2370that everybody includes and which overrides some configure choices:</p>
2371<pre> #define EV_MINIMAL 1
2372 #define EV_USE_POLL 0
2373 #define EV_MULTIPLICITY 0
2374 #define EV_PERIODIC_ENABLE 0
2375 #define EV_STAT_ENABLE 0
2376 #define EV_FORK_ENABLE 0
2377 #define EV_CONFIG_H &lt;config.h&gt;
2378 #define EV_MINPRI 0
2379 #define EV_MAXPRI 0
2380
2381 #include &quot;ev++.h&quot;
2382
2383</pre>
2384 <p>And a <cite>ev_cpp.C</cite> implementation file that contains libev proper and is compiled:</p>
2385<pre> #include &quot;ev_cpp.h&quot;
2386 #include &quot;ev.c&quot;
2387
2388
2389
2390
2391</pre>
2392
2393</div>
2394<h1 id="COMPLEXITIES">COMPLEXITIES</h1>
2395<div id="COMPLEXITIES_CONTENT">
2396 <p>In this section the complexities of (many of) the algorithms used inside
2397libev will be explained. For complexity discussions about backends see the
2398documentation for <code>ev_default_init</code>.</p>
2399 <p>All of the following are about amortised time: If an array needs to be
2400extended, libev needs to realloc and move the whole array, but this
2401happens asymptotically never with higher number of elements, so O(1) might
2402mean it might do a lengthy realloc operation in rare cases, but on average
2403it is much faster and asymptotically approaches constant time.</p>
2404 <p>
2405 <dl>
2406 <dt>Starting and stopping timer/periodic watchers: O(log skipped_other_timers)</dt>
2407 <dd>
2408 <p>This means that, when you have a watcher that triggers in one hour and
2409there are 100 watchers that would trigger before that then inserting will
2410have to skip those 100 watchers.</p>
2411 </dd>
2412 <dt>Changing timer/periodic watchers (by autorepeat, again): O(log skipped_other_timers)</dt>
2413 <dd>
2414 <p>That means that for changing a timer costs less than removing/adding them
2415as only the relative motion in the event queue has to be paid for.</p>
2416 </dd>
2417 <dt>Starting io/check/prepare/idle/signal/child watchers: O(1)</dt>
2418 <dd>
2419 <p>These just add the watcher into an array or at the head of a list.
2420=item Stopping check/prepare/idle watchers: O(1)</p>
2421 </dd>
2422 <dt>Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))</dt>
2423 <dd>
2424 <p>These watchers are stored in lists then need to be walked to find the
2425correct watcher to remove. The lists are usually short (you don't usually
2426have many watchers waiting for the same fd or signal).</p>
2427 </dd>
2428 <dt>Finding the next timer per loop iteration: O(1)</dt>
2429 <dt>Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)</dt>
2430 <dd>
2431 <p>A change means an I/O watcher gets started or stopped, which requires
2432libev to recalculate its status (and possibly tell the kernel).</p>
2433 </dd>
2434 <dt>Activating one watcher: O(1)</dt>
2435 <dt>Priority handling: O(number_of_priorities)</dt>
2436 <dd>
2437 <p>Priorities are implemented by allocating some space for each
2438priority. When doing priority-based operations, libev usually has to
2439linearly search all the priorities.</p>
2440 </dd>
2441 </dl>
2442 </p>
2443
2444
2445
2446
2447
2448</div>
2449<h1 id="AUTHOR">AUTHOR</h1>
760<div id="AUTHOR_CONTENT"> 2450<div id="AUTHOR_CONTENT">
761<p>Marc Lehmann &lt;libev@schmorp.de&gt;.</p> 2451 <p>Marc Lehmann &lt;libev@schmorp.de&gt;.</p>
762 2452
763</div> 2453</div>
764</div></body> 2454</div></body>
765</html> 2455</html>

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines