ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/libev/ev.pod

Comparing libev/ev.pod (file contents):
Revision 1.30 by root, Fri Nov 23 04:36:03 2007 UTC vs.
Revision 1.58 by root, Wed Nov 28 11:31:34 2007 UTC

…		…
3	libev - a high performance full-featured event loop written in C	3	libev - a high performance full-featured event loop written in C
4		4
5	=head1 SYNOPSIS	5	=head1 SYNOPSIS
6		6
7	#include <ev.h>	7	#include <ev.h>
		8
		9	=head1 EXAMPLE PROGRAM
		10
		11	#include <ev.h>
		12
		13	ev_io stdin_watcher;
		14	ev_timer timeout_watcher;
		15
		16	/* called when data readable on stdin */
		17	static void
		18	stdin_cb (EV_P_ struct ev_io *w, int revents)
		19	{
		20	/* puts ("stdin ready"); */
		21	ev_io_stop (EV_A_ w); /* just a syntax example */
		22	ev_unloop (EV_A_ EVUNLOOP_ALL); /* leave all loop calls */
		23	}
		24
		25	static void
		26	timeout_cb (EV_P_ struct ev_timer *w, int revents)
		27	{
		28	/* puts ("timeout"); */
		29	ev_unloop (EV_A_ EVUNLOOP_ONE); /* leave one loop call */
		30	}
		31
		32	int
		33	main (void)
		34	{
		35	struct ev_loop *loop = ev_default_loop (0);
		36
		37	/* initialise an io watcher, then start it */
		38	ev_io_init (&stdin_watcher, stdin_cb, /*STDIN_FILENO*/ 0, EV_READ);
		39	ev_io_start (loop, &stdin_watcher);
		40
		41	/* simple non-repeating 5.5 second timeout */
		42	ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);
		43	ev_timer_start (loop, &timeout_watcher);
		44
		45	/* loop till timeout or data ready */
		46	ev_loop (loop, 0);
		47
		48	return 0;
		49	}
8		50
9	=head1 DESCRIPTION	51	=head1 DESCRIPTION
10		52
11	Libev is an event loop: you register interest in certain events (such as a	53	Libev is an event loop: you register interest in certain events (such as a
12	file descriptor being readable or a timeout occuring), and it will manage	54	file descriptor being readable or a timeout occuring), and it will manage
…		…
21	details of the event, and then hand it over to libev by I<starting> the	63	details of the event, and then hand it over to libev by I<starting> the
22	watcher.	64	watcher.
23		65
24	=head1 FEATURES	66	=head1 FEATURES
25		67
26	Libev supports select, poll, the linux-specific epoll and the bsd-specific	68	Libev supports C<select>, C<poll>, the Linux-specific C<epoll>, the
27	kqueue mechanisms for file descriptor events, relative timers, absolute	69	BSD-specific C<kqueue> and the Solaris-specific event port mechanisms
28	timers with customised rescheduling, signal events, process status change	70	for file descriptor events (C<ev_io>), the Linux C<inotify> interface
29	events (related to SIGCHLD), and event watchers dealing with the event	71	(for C<ev_stat>), relative timers (C<ev_timer>), absolute timers
30	loop mechanism itself (idle, prepare and check watchers). It also is quite	72	with customised rescheduling (C<ev_periodic>), synchronous signals
		73	(C<ev_signal>), process status change events (C<ev_child>), and event
		74	watchers dealing with the event loop mechanism itself (C<ev_idle>,
		75	C<ev_embed>, C<ev_prepare> and C<ev_check> watchers) as well as
		76	file watchers (C<ev_stat>) and even limited support for fork events
		77	(C<ev_fork>).
		78
		79	It also is quite fast (see this
31	fast (see this L<benchmark|http://libev.schmorp.de/bench.html> comparing	80	L<benchmark|http://libev.schmorp.de/bench.html> comparing it to libevent
32	it to libevent for example).	81	for example).
33		82
34	=head1 CONVENTIONS	83	=head1 CONVENTIONS
35		84
36	Libev is very configurable. In this manual the default configuration	85	Libev is very configurable. In this manual the default configuration will
37	will be described, which supports multiple event loops. For more info	86	be described, which supports multiple event loops. For more info about
38	about various configuration options please have a look at the file	87	various configuration options please have a look at B<EMBED> section in
39	F<README.embed> in the libev distribution. If libev was configured without	88	this manual. If libev was configured without support for multiple event
40	support for multiple event loops, then all functions taking an initial	89	loops, then all functions taking an initial argument of name C<loop>
41	argument of name C<loop> (which is always of type C<struct ev_loop *>)	90	(which is always of type C<struct ev_loop *>) will not have this argument.
42	will not have this argument.
43		91
44	=head1 TIME REPRESENTATION	92	=head1 TIME REPRESENTATION
45		93
46	Libev represents time as a single floating point number, representing the	94	Libev represents time as a single floating point number, representing the
47	(fractional) number of seconds since the (POSIX) epoch (somewhere near	95	(fractional) number of seconds since the (POSIX) epoch (somewhere near
48	the beginning of 1970, details are complicated, don't ask). This type is	96	the beginning of 1970, details are complicated, don't ask). This type is
49	called C<ev_tstamp>, which is what you should use too. It usually aliases	97	called C<ev_tstamp>, which is what you should use too. It usually aliases
50	to the double type in C.	98	to the C<double> type in C, and when you need to do any calculations on
		99	it, you should treat it as such.
51		100
52	=head1 GLOBAL FUNCTIONS	101	=head1 GLOBAL FUNCTIONS
53		102
54	These functions can be called anytime, even before initialising the	103	These functions can be called anytime, even before initialising the
55	library in any way.	104	library in any way.
…		…
75	Usually, it's a good idea to terminate if the major versions mismatch,	124	Usually, it's a good idea to terminate if the major versions mismatch,
76	as this indicates an incompatible change. Minor versions are usually	125	as this indicates an incompatible change. Minor versions are usually
77	compatible to older versions, so a larger minor version alone is usually	126	compatible to older versions, so a larger minor version alone is usually
78	not a problem.	127	not a problem.
79		128
		129	Example: Make sure we haven't accidentally been linked against the wrong
		130	version.
		131
		132	assert (("libev version mismatch",
		133	ev_version_major () == EV_VERSION_MAJOR
		134	&& ev_version_minor () >= EV_VERSION_MINOR));
		135
		136	=item unsigned int ev_supported_backends ()
		137
		138	Return the set of all backends (i.e. their corresponding C<EV_BACKEND_*>
		139	value) compiled into this binary of libev (independent of their
		140	availability on the system you are running on). See C<ev_default_loop> for
		141	a description of the set values.
		142
		143	Example: make sure we have the epoll method, because yeah this is cool and
		144	a must have and can we have a torrent of it please!!!11
		145
		146	assert (("sorry, no epoll, no sex",
		147	ev_supported_backends () & EVBACKEND_EPOLL));
		148
		149	=item unsigned int ev_recommended_backends ()
		150
		151	Return the set of all backends compiled into this binary of libev and also
		152	recommended for this platform. This set is often smaller than the one
		153	returned by C<ev_supported_backends>, as for example kqueue is broken on
		154	most BSDs and will not be autodetected unless you explicitly request it
		155	(assuming you know what you are doing). This is the set of backends that
		156	libev will probe for if you specify no backends explicitly.
		157
		158	=item unsigned int ev_embeddable_backends ()
		159
		160	Returns the set of backends that are embeddable in other event loops. This
		161	is the theoretical, all-platform, value. To find which backends
		162	might be supported on the current system, you would need to look at
		163	C<ev_embeddable_backends () & ev_supported_backends ()>, likewise for
		164	recommended ones.
		165
		166	See the description of C<ev_embed> watchers for more info.
		167
80	=item ev_set_allocator (void *(*cb)(void *ptr, long size))	168	=item ev_set_allocator (void *(*cb)(void *ptr, size_t size))
81		169
82	Sets the allocation function to use (the prototype is similar to the	170	Sets the allocation function to use (the prototype and semantics are
83	realloc C function, the semantics are identical). It is used to allocate	171	identical to the realloc C function). It is used to allocate and free
84	and free memory (no surprises here). If it returns zero when memory	172	memory (no surprises here). If it returns zero when memory needs to be
85	needs to be allocated, the library might abort or take some potentially	173	allocated, the library might abort or take some potentially destructive
86	destructive action. The default is your system realloc function.	174	action. The default is your system realloc function.
87		175
88	You could override this function in high-availability programs to, say,	176	You could override this function in high-availability programs to, say,
89	free some memory if it cannot allocate memory, to use a special allocator,	177	free some memory if it cannot allocate memory, to use a special allocator,
90	or even to sleep a while and retry until some memory is available.	178	or even to sleep a while and retry until some memory is available.
		179
		180	Example: Replace the libev allocator with one that waits a bit and then
		181	retries).
		182
		183	static void *
		184	persistent_realloc (void *ptr, size_t size)
		185	{
		186	for (;;)
		187	{
		188	void *newptr = realloc (ptr, size);
		189
		190	if (newptr)
		191	return newptr;
		192
		193	sleep (60);
		194	}
		195	}
		196
		197	...
		198	ev_set_allocator (persistent_realloc);
91		199
92	=item ev_set_syserr_cb (void (*cb)(const char *msg));	200	=item ev_set_syserr_cb (void (*cb)(const char *msg));
93		201
94	Set the callback function to call on a retryable syscall error (such	202	Set the callback function to call on a retryable syscall error (such
95	as failed select, poll, epoll_wait). The message is a printable string	203	as failed select, poll, epoll_wait). The message is a printable string
…		…
97	callback is set, then libev will expect it to remedy the sitution, no	205	callback is set, then libev will expect it to remedy the sitution, no
98	matter what, when it returns. That is, libev will generally retry the	206	matter what, when it returns. That is, libev will generally retry the
99	requested operation, or, if the condition doesn't go away, do bad stuff	207	requested operation, or, if the condition doesn't go away, do bad stuff
100	(such as abort).	208	(such as abort).
101		209
		210	Example: This is basically the same thing that libev does internally, too.
		211
		212	static void
		213	fatal_error (const char *msg)
		214	{
		215	perror (msg);
		216	abort ();
		217	}
		218
		219	...
		220	ev_set_syserr_cb (fatal_error);
		221
102	=back	222	=back
103		223
104	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP	224	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP
105		225
106	An event loop is described by a C<struct ev_loop *>. The library knows two	226	An event loop is described by a C<struct ev_loop *>. The library knows two
…		…
119	=item struct ev_loop *ev_default_loop (unsigned int flags)	239	=item struct ev_loop *ev_default_loop (unsigned int flags)
120		240
121	This will initialise the default event loop if it hasn't been initialised	241	This will initialise the default event loop if it hasn't been initialised
122	yet and return it. If the default loop could not be initialised, returns	242	yet and return it. If the default loop could not be initialised, returns
123	false. If it already was initialised it simply returns it (and ignores the	243	false. If it already was initialised it simply returns it (and ignores the
124	flags).	244	flags. If that is troubling you, check C<ev_backend ()> afterwards).
125		245
126	If you don't know what event loop to use, use the one returned from this	246	If you don't know what event loop to use, use the one returned from this
127	function.	247	function.
128		248
129	The flags argument can be used to specify special behaviour or specific	249	The flags argument can be used to specify special behaviour or specific
130	backends to use, and is usually specified as 0 (or EVFLAG_AUTO).	250	backends to use, and is usually specified as C<0> (or C<EVFLAG_AUTO>).
131		251
132	It supports the following flags:	252	The following flags are supported:
133		253
134	=over 4	254	=over 4
135		255
136	=item C<EVFLAG_AUTO>	256	=item C<EVFLAG_AUTO>
137		257
…		…
145	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will	265	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will
146	override the flags completely if it is found in the environment. This is	266	override the flags completely if it is found in the environment. This is
147	useful to try out specific backends to test their performance, or to work	267	useful to try out specific backends to test their performance, or to work
148	around bugs.	268	around bugs.
149		269
150	=item C<EVMETHOD_SELECT> (value 1, portable select backend)	270	=item C<EVBACKEND_SELECT> (value 1, portable select backend)
151		271
152	This is your standard select(2) backend. Not I<completely> standard, as	272	This is your standard select(2) backend. Not I<completely> standard, as
153	libev tries to roll its own fd_set with no limits on the number of fds,	273	libev tries to roll its own fd_set with no limits on the number of fds,
154	but if that fails, expect a fairly low limit on the number of fds when	274	but if that fails, expect a fairly low limit on the number of fds when
155	using this backend. It doesn't scale too well (O(highest_fd)), but its usually	275	using this backend. It doesn't scale too well (O(highest_fd)), but its usually
156	the fastest backend for a low number of fds.	276	the fastest backend for a low number of fds.
157		277
158	=item C<EVMETHOD_POLL> (value 2, poll backend, available everywhere except on windows)	278	=item C<EVBACKEND_POLL> (value 2, poll backend, available everywhere except on windows)
159		279
160	And this is your standard poll(2) backend. It's more complicated than	280	And this is your standard poll(2) backend. It's more complicated than
161	select, but handles sparse fds better and has no artificial limit on the	281	select, but handles sparse fds better and has no artificial limit on the
162	number of fds you can use (except it will slow down considerably with a	282	number of fds you can use (except it will slow down considerably with a
163	lot of inactive fds). It scales similarly to select, i.e. O(total_fds).	283	lot of inactive fds). It scales similarly to select, i.e. O(total_fds).
164		284
165	=item C<EVMETHOD_EPOLL> (value 4, Linux)	285	=item C<EVBACKEND_EPOLL> (value 4, Linux)
166		286
167	For few fds, this backend is a bit little slower than poll and select,	287	For few fds, this backend is a bit little slower than poll and select,
168	but it scales phenomenally better. While poll and select usually scale like	288	but it scales phenomenally better. While poll and select usually scale like
169	O(total_fds) where n is the total number of fds (or the highest fd), epoll scales	289	O(total_fds) where n is the total number of fds (or the highest fd), epoll scales
170	either O(1) or O(active_fds).	290	either O(1) or O(active_fds).
…		…
173	result in some caching, there is still a syscall per such incident	293	result in some caching, there is still a syscall per such incident
174	(because the fd could point to a different file description now), so its	294	(because the fd could point to a different file description now), so its
175	best to avoid that. Also, dup()ed file descriptors might not work very	295	best to avoid that. Also, dup()ed file descriptors might not work very
176	well if you register events for both fds.	296	well if you register events for both fds.
177		297
		298	Please note that epoll sometimes generates spurious notifications, so you
		299	need to use non-blocking I/O or other means to avoid blocking when no data
		300	(or space) is available.
		301
178	=item C<EVMETHOD_KQUEUE> (value 8, most BSD clones)	302	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)
179		303
180	Kqueue deserves special mention, as at the time of this writing, it	304	Kqueue deserves special mention, as at the time of this writing, it
181	was broken on all BSDs except NetBSD (usually it doesn't work with	305	was broken on all BSDs except NetBSD (usually it doesn't work with
182	anything but sockets and pipes, except on Darwin, where of course its	306	anything but sockets and pipes, except on Darwin, where of course its
183	completely useless). For this reason its not being "autodetected" unless	307	completely useless). For this reason its not being "autodetected"
184	you explicitly specify the flags (i.e. you don't use EVFLAG_AUTO).	308	unless you explicitly specify it explicitly in the flags (i.e. using
		309	C<EVBACKEND_KQUEUE>).
185		310
186	It scales in the same way as the epoll backend, but the interface to the	311	It scales in the same way as the epoll backend, but the interface to the
187	kernel is more efficient (which says nothing about its actual speed, of	312	kernel is more efficient (which says nothing about its actual speed, of
188	course). While starting and stopping an I/O watcher does not cause an	313	course). While starting and stopping an I/O watcher does not cause an
189	extra syscall as with epoll, it still adds up to four event changes per	314	extra syscall as with epoll, it still adds up to four event changes per
190	incident, so its best to avoid that.	315	incident, so its best to avoid that.
191		316
192	=item C<EVMETHOD_DEVPOLL> (value 16, Solaris 8)	317	=item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8)
193		318
194	This is not implemented yet (and might never be).	319	This is not implemented yet (and might never be).
195		320
196	=item C<EVMETHOD_PORT> (value 32, Solaris 10)	321	=item C<EVBACKEND_PORT> (value 32, Solaris 10)
197		322
198	This uses the Solaris 10 port mechanism. As with everything on Solaris,	323	This uses the Solaris 10 port mechanism. As with everything on Solaris,
199	it's really slow, but it still scales very well (O(active_fds)).	324	it's really slow, but it still scales very well (O(active_fds)).
200		325
		326	Please note that solaris ports can result in a lot of spurious
		327	notifications, so you need to use non-blocking I/O or other means to avoid
		328	blocking when no data (or space) is available.
		329
201	=item C<EVMETHOD_ALL>	330	=item C<EVBACKEND_ALL>
202		331
203	Try all backends (even potentially broken ones that wouldn't be tried	332	Try all backends (even potentially broken ones that wouldn't be tried
204	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as	333	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as
205	C<EVMETHOD_ALL & ~EVMETHOD_KQUEUE>.	334	C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>.
206		335
207	=back	336	=back
208		337
209	If one or more of these are ored into the flags value, then only these	338	If one or more of these are ored into the flags value, then only these
210	backends will be tried (in the reverse order as given here). If none are	339	backends will be tried (in the reverse order as given here). If none are
211	specified, most compiled-in backend will be tried, usually in reverse	340	specified, most compiled-in backend will be tried, usually in reverse
212	order of their flag values :)	341	order of their flag values :)
213		342
		343	The most typical usage is like this:
		344
		345	if (!ev_default_loop (0))
		346	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
		347
		348	Restrict libev to the select and poll backends, and do not allow
		349	environment settings to be taken into account:
		350
		351	ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);
		352
		353	Use whatever libev has to offer, but make sure that kqueue is used if
		354	available (warning, breaks stuff, best use only with your own private
		355	event loop and only if you know the OS supports your types of fds):
		356
		357	ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);
		358
214	=item struct ev_loop *ev_loop_new (unsigned int flags)	359	=item struct ev_loop *ev_loop_new (unsigned int flags)
215		360
216	Similar to C<ev_default_loop>, but always creates a new event loop that is	361	Similar to C<ev_default_loop>, but always creates a new event loop that is
217	always distinct from the default loop. Unlike the default loop, it cannot	362	always distinct from the default loop. Unlike the default loop, it cannot
218	handle signal and child watchers, and attempts to do so will be greeted by	363	handle signal and child watchers, and attempts to do so will be greeted by
219	undefined behaviour (or a failed assertion if assertions are enabled).	364	undefined behaviour (or a failed assertion if assertions are enabled).
220		365
		366	Example: Try to create a event loop that uses epoll and nothing else.
		367
		368	struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);
		369	if (!epoller)
		370	fatal ("no epoll found here, maybe it hides under your chair");
		371
221	=item ev_default_destroy ()	372	=item ev_default_destroy ()
222		373
223	Destroys the default loop again (frees all memory and kernel state	374	Destroys the default loop again (frees all memory and kernel state
224	etc.). This stops all registered event watchers (by not touching them in	375	etc.). None of the active event watchers will be stopped in the normal
225	any way whatsoever, although you cannot rely on this :).	376	sense, so e.g. C<ev_is_active> might still return true. It is your
		377	responsibility to either stop all watchers cleanly yoursef I<before>
		378	calling this function, or cope with the fact afterwards (which is usually
		379	the easiest thing, youc na just ignore the watchers and/or C<free ()> them
		380	for example).
226		381
227	=item ev_loop_destroy (loop)	382	=item ev_loop_destroy (loop)
228		383
229	Like C<ev_default_destroy>, but destroys an event loop created by an	384	Like C<ev_default_destroy>, but destroys an event loop created by an
230	earlier call to C<ev_loop_new>.	385	earlier call to C<ev_loop_new>.
…		…
244	it just in case after a fork. To make this easy, the function will fit in	399	it just in case after a fork. To make this easy, the function will fit in
245	quite nicely into a call to C<pthread_atfork>:	400	quite nicely into a call to C<pthread_atfork>:
246		401
247	pthread_atfork (0, 0, ev_default_fork);	402	pthread_atfork (0, 0, ev_default_fork);
248		403
		404	At the moment, C<EVBACKEND_SELECT> and C<EVBACKEND_POLL> are safe to use
		405	without calling this function, so if you force one of those backends you
		406	do not need to care.
		407
249	=item ev_loop_fork (loop)	408	=item ev_loop_fork (loop)
250		409
251	Like C<ev_default_fork>, but acts on an event loop created by	410	Like C<ev_default_fork>, but acts on an event loop created by
252	C<ev_loop_new>. Yes, you have to call this on every allocated event loop	411	C<ev_loop_new>. Yes, you have to call this on every allocated event loop
253	after fork, and how you do this is entirely your own problem.	412	after fork, and how you do this is entirely your own problem.
254		413
255	=item unsigned int ev_method (loop)	414	=item unsigned int ev_backend (loop)
256		415
257	Returns one of the C<EVMETHOD_*> flags indicating the event backend in	416	Returns one of the C<EVBACKEND_*> flags indicating the event backend in
258	use.	417	use.
259		418
260	=item ev_tstamp ev_now (loop)	419	=item ev_tstamp ev_now (loop)
261		420
262	Returns the current "event loop time", which is the time the event loop	421	Returns the current "event loop time", which is the time the event loop
263	got events and started processing them. This timestamp does not change	422	received events and started processing them. This timestamp does not
264	as long as callbacks are being processed, and this is also the base time	423	change as long as callbacks are being processed, and this is also the base
265	used for relative timers. You can treat it as the timestamp of the event	424	time used for relative timers. You can treat it as the timestamp of the
266	occuring (or more correctly, the mainloop finding out about it).	425	event occuring (or more correctly, libev finding out about it).
267		426
268	=item ev_loop (loop, int flags)	427	=item ev_loop (loop, int flags)
269		428
270	Finally, this is it, the event handler. This function usually is called	429	Finally, this is it, the event handler. This function usually is called
271	after you initialised all your watchers and you want to start handling	430	after you initialised all your watchers and you want to start handling
272	events.	431	events.
273		432
274	If the flags argument is specified as 0, it will not return until either	433	If the flags argument is specified as C<0>, it will not return until
275	no event watchers are active anymore or C<ev_unloop> was called.	434	either no event watchers are active anymore or C<ev_unloop> was called.
		435
		436	Please note that an explicit C<ev_unloop> is usually better than
		437	relying on all watchers to be stopped when deciding when a program has
		438	finished (especially in interactive programs), but having a program that
		439	automatically loops as long as it has to and no longer by virtue of
		440	relying on its watchers stopping correctly is a thing of beauty.
276		441
277	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle	442	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle
278	those events and any outstanding ones, but will not block your process in	443	those events and any outstanding ones, but will not block your process in
279	case there are no events and will return after one iteration of the loop.	444	case there are no events and will return after one iteration of the loop.
280		445
281	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if	446	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if
282	neccessary) and will handle those and any outstanding ones. It will block	447	neccessary) and will handle those and any outstanding ones. It will block
283	your process until at least one new event arrives, and will return after	448	your process until at least one new event arrives, and will return after
284	one iteration of the loop.	449	one iteration of the loop. This is useful if you are waiting for some
		450	external event in conjunction with something not expressible using other
		451	libev watchers. However, a pair of C<ev_prepare>/C<ev_check> watchers is
		452	usually a better approach for this kind of thing.
285		453
286	This flags value could be used to implement alternative looping
287	constructs, but the C<prepare> and C<check> watchers provide a better and
288	more generic mechanism.
289
290	Here are the gory details of what ev_loop does:	454	Here are the gory details of what C<ev_loop> does:
291		455
292	1. If there are no active watchers (reference count is zero), return.	456	* If there are no active watchers (reference count is zero), return.
293	2. Queue and immediately call all prepare watchers.	457	- Queue prepare watchers and then call all outstanding watchers.
294	3. If we have been forked, recreate the kernel state.	458	- If we have been forked, recreate the kernel state.
295	4. Update the kernel state with all outstanding changes.	459	- Update the kernel state with all outstanding changes.
296	5. Update the "event loop time".	460	- Update the "event loop time".
297	6. Calculate for how long to block.	461	- Calculate for how long to block.
298	7. Block the process, waiting for events.	462	- Block the process, waiting for any events.
		463	- Queue all outstanding I/O (fd) events.
299	8. Update the "event loop time" and do time jump handling.	464	- Update the "event loop time" and do time jump handling.
300	9. Queue all outstanding timers.	465	- Queue all outstanding timers.
301	10. Queue all outstanding periodics.	466	- Queue all outstanding periodics.
302	11. If no events are pending now, queue all idle watchers.	467	- If no events are pending now, queue all idle watchers.
303	12. Queue all check watchers.	468	- Queue all check watchers.
304	13. Call all queued watchers in reverse order (i.e. check watchers first).	469	- Call all queued watchers in reverse order (i.e. check watchers first).
		470	Signals and child watchers are implemented as I/O watchers, and will
		471	be handled here by queueing them when their watcher gets executed.
305	14. If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK	472	- If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK
306	was used, return, otherwise continue with step #1.	473	were used, return, otherwise continue with step *.
		474
		475	Example: Queue some jobs and then loop until no events are outsanding
		476	anymore.
		477
		478	... queue jobs here, make sure they register event watchers as long
		479	... as they still have work to do (even an idle watcher will do..)
		480	ev_loop (my_loop, 0);
		481	... jobs done. yeah!
307		482
308	=item ev_unloop (loop, how)	483	=item ev_unloop (loop, how)
309		484
310	Can be used to make a call to C<ev_loop> return early (but only after it	485	Can be used to make a call to C<ev_loop> return early (but only after it
311	has processed all outstanding events). The C<how> argument must be either	486	has processed all outstanding events). The C<how> argument must be either
…		…
325	visible to the libev user and should not keep C<ev_loop> from exiting if	500	visible to the libev user and should not keep C<ev_loop> from exiting if
326	no event watchers registered by it are active. It is also an excellent	501	no event watchers registered by it are active. It is also an excellent
327	way to do this for generic recurring timers or from within third-party	502	way to do this for generic recurring timers or from within third-party
328	libraries. Just remember to I<unref after start> and I<ref before stop>.	503	libraries. Just remember to I<unref after start> and I<ref before stop>.
329		504
		505	Example: Create a signal watcher, but keep it from keeping C<ev_loop>
		506	running when nothing else is active.
		507
		508	struct ev_signal exitsig;
		509	ev_signal_init (&exitsig, sig_cb, SIGINT);
		510	ev_signal_start (loop, &exitsig);
		511	evf_unref (loop);
		512
		513	Example: For some weird reason, unregister the above signal handler again.
		514
		515	ev_ref (loop);
		516	ev_signal_stop (loop, &exitsig);
		517
330	=back	518	=back
		519
331		520
332	=head1 ANATOMY OF A WATCHER	521	=head1 ANATOMY OF A WATCHER
333		522
334	A watcher is a structure that you create and register to record your	523	A watcher is a structure that you create and register to record your
335	interest in some event. For instance, if you want to wait for STDIN to	524	interest in some event. For instance, if you want to wait for STDIN to
…		…
368	*) >>), and you can stop watching for events at any time by calling the	557	*) >>), and you can stop watching for events at any time by calling the
369	corresponding stop function (C<< ev_<type>_stop (loop, watcher *) >>.	558	corresponding stop function (C<< ev_<type>_stop (loop, watcher *) >>.
370		559
371	As long as your watcher is active (has been started but not stopped) you	560	As long as your watcher is active (has been started but not stopped) you
372	must not touch the values stored in it. Most specifically you must never	561	must not touch the values stored in it. Most specifically you must never
373	reinitialise it or call its set method.	562	reinitialise it or call its C<set> macro.
374
375	You can check whether an event is active by calling the C<ev_is_active
376	(watcher *)> macro. To see whether an event is outstanding (but the
377	callback for it has not been called yet) you can use the C<ev_is_pending
378	(watcher *)> macro.
379		563
380	Each and every callback receives the event loop pointer as first, the	564	Each and every callback receives the event loop pointer as first, the
381	registered watcher structure as second, and a bitset of received events as	565	registered watcher structure as second, and a bitset of received events as
382	third argument.	566	third argument.
383		567
…		…
407	The signal specified in the C<ev_signal> watcher has been received by a thread.	591	The signal specified in the C<ev_signal> watcher has been received by a thread.
408		592
409	=item C<EV_CHILD>	593	=item C<EV_CHILD>
410		594
411	The pid specified in the C<ev_child> watcher has received a status change.	595	The pid specified in the C<ev_child> watcher has received a status change.
		596
		597	=item C<EV_STAT>
		598
		599	The path specified in the C<ev_stat> watcher changed its attributes somehow.
412		600
413	=item C<EV_IDLE>	601	=item C<EV_IDLE>
414		602
415	The C<ev_idle> watcher has determined that you have nothing better to do.	603	The C<ev_idle> watcher has determined that you have nothing better to do.
416		604
…		…
424	received events. Callbacks of both watcher types can start and stop as	612	received events. Callbacks of both watcher types can start and stop as
425	many watchers as they want, and all of them will be taken into account	613	many watchers as they want, and all of them will be taken into account
426	(for example, a C<ev_prepare> watcher might start an idle watcher to keep	614	(for example, a C<ev_prepare> watcher might start an idle watcher to keep
427	C<ev_loop> from blocking).	615	C<ev_loop> from blocking).
428		616
		617	=item C<EV_EMBED>
		618
		619	The embedded event loop specified in the C<ev_embed> watcher needs attention.
		620
		621	=item C<EV_FORK>
		622
		623	The event loop has been resumed in the child process after fork (see
		624	C<ev_fork>).
		625
429	=item C<EV_ERROR>	626	=item C<EV_ERROR>
430		627
431	An unspecified error has occured, the watcher has been stopped. This might	628	An unspecified error has occured, the watcher has been stopped. This might
432	happen because the watcher could not be properly started because libev	629	happen because the watcher could not be properly started because libev
433	ran out of memory, a file descriptor was found to be closed or any other	630	ran out of memory, a file descriptor was found to be closed or any other
…		…
439	your callbacks is well-written it can just attempt the operation and cope	636	your callbacks is well-written it can just attempt the operation and cope
440	with the error from read() or write(). This will not work in multithreaded	637	with the error from read() or write(). This will not work in multithreaded
441	programs, though, so beware.	638	programs, though, so beware.
442		639
443	=back	640	=back
		641
		642	=head2 GENERIC WATCHER FUNCTIONS
		643
		644	In the following description, C<TYPE> stands for the watcher type,
		645	e.g. C<timer> for C<ev_timer> watchers and C<io> for C<ev_io> watchers.
		646
		647	=over 4
		648
		649	=item C<ev_init> (ev_TYPE *watcher, callback)
		650
		651	This macro initialises the generic portion of a watcher. The contents
		652	of the watcher object can be arbitrary (so C<malloc> will do). Only
		653	the generic parts of the watcher are initialised, you I<need> to call
		654	the type-specific C<ev_TYPE_set> macro afterwards to initialise the
		655	type-specific parts. For each type there is also a C<ev_TYPE_init> macro
		656	which rolls both calls into one.
		657
		658	You can reinitialise a watcher at any time as long as it has been stopped
		659	(or never started) and there are no pending events outstanding.
		660
		661	The callback is always of type C<void (*)(ev_loop *loop, ev_TYPE *watcher,
		662	int revents)>.
		663
		664	=item C<ev_TYPE_set> (ev_TYPE *, [args])
		665
		666	This macro initialises the type-specific parts of a watcher. You need to
		667	call C<ev_init> at least once before you call this macro, but you can
		668	call C<ev_TYPE_set> any number of times. You must not, however, call this
		669	macro on a watcher that is active (it can be pending, however, which is a
		670	difference to the C<ev_init> macro).
		671
		672	Although some watcher types do not have type-specific arguments
		673	(e.g. C<ev_prepare>) you still need to call its C<set> macro.
		674
		675	=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])
		676
		677	This convinience macro rolls both C<ev_init> and C<ev_TYPE_set> macro
		678	calls into a single call. This is the most convinient method to initialise
		679	a watcher. The same limitations apply, of course.
		680
		681	=item C<ev_TYPE_start> (loop *, ev_TYPE *watcher)
		682
		683	Starts (activates) the given watcher. Only active watchers will receive
		684	events. If the watcher is already active nothing will happen.
		685
		686	=item C<ev_TYPE_stop> (loop *, ev_TYPE *watcher)
		687
		688	Stops the given watcher again (if active) and clears the pending
		689	status. It is possible that stopped watchers are pending (for example,
		690	non-repeating timers are being stopped when they become pending), but
		691	C<ev_TYPE_stop> ensures that the watcher is neither active nor pending. If
		692	you want to free or reuse the memory used by the watcher it is therefore a
		693	good idea to always call its C<ev_TYPE_stop> function.
		694
		695	=item bool ev_is_active (ev_TYPE *watcher)
		696
		697	Returns a true value iff the watcher is active (i.e. it has been started
		698	and not yet been stopped). As long as a watcher is active you must not modify
		699	it.
		700
		701	=item bool ev_is_pending (ev_TYPE *watcher)
		702
		703	Returns a true value iff the watcher is pending, (i.e. it has outstanding
		704	events but its callback has not yet been invoked). As long as a watcher
		705	is pending (but not active) you must not call an init function on it (but
		706	C<ev_TYPE_set> is safe) and you must make sure the watcher is available to
		707	libev (e.g. you cnanot C<free ()> it).
		708
		709	=item callback ev_cb (ev_TYPE *watcher)
		710
		711	Returns the callback currently set on the watcher.
		712
		713	=item ev_cb_set (ev_TYPE *watcher, callback)
		714
		715	Change the callback. You can change the callback at virtually any time
		716	(modulo threads).
		717
		718	=back
		719
444		720
445	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER	721	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER
446		722
447	Each watcher has, by default, a member C<void *data> that you can change	723	Each watcher has, by default, a member C<void *data> that you can change
448	and read at any time, libev will completely ignore it. This can be used	724	and read at any time, libev will completely ignore it. This can be used
…		…
466	{	742	{
467	struct my_io *w = (struct my_io *)w_;	743	struct my_io *w = (struct my_io *)w_;
468	...	744	...
469	}	745	}
470		746
471	More interesting and less C-conformant ways of catsing your callback type	747	More interesting and less C-conformant ways of casting your callback type
472	have been omitted....	748	instead have been omitted.
		749
		750	Another common scenario is having some data structure with multiple
		751	watchers:
		752
		753	struct my_biggy
		754	{
		755	int some_data;
		756	ev_timer t1;
		757	ev_timer t2;
		758	}
		759
		760	In this case getting the pointer to C<my_biggy> is a bit more complicated,
		761	you need to use C<offsetof>:
		762
		763	#include <stddef.h>
		764
		765	static void
		766	t1_cb (EV_P_ struct ev_timer *w, int revents)
		767	{
		768	struct my_biggy big = (struct my_biggy *
		769	(((char *)w) - offsetof (struct my_biggy, t1));
		770	}
		771
		772	static void
		773	t2_cb (EV_P_ struct ev_timer *w, int revents)
		774	{
		775	struct my_biggy big = (struct my_biggy *
		776	(((char *)w) - offsetof (struct my_biggy, t2));
		777	}
473		778
474		779
475	=head1 WATCHER TYPES	780	=head1 WATCHER TYPES
476		781
477	This section describes each watcher in detail, but will not repeat	782	This section describes each watcher in detail, but will not repeat
478	information given in the last section.	783	information given in the last section. Any initialisation/set macros,
		784	functions and members specific to the watcher type are explained.
479		785
		786	Members are additionally marked with either I<[read-only]>, meaning that,
		787	while the watcher is active, you can look at the member and expect some
		788	sensible content, but you must not modify it (you can modify it while the
		789	watcher is stopped to your hearts content), or I<[read-write]>, which
		790	means you can expect it to have some sensible content while the watcher
		791	is active, but you can also modify it. Modifying it may not do something
		792	sensible or take immediate effect (or do anything at all), but libev will
		793	not crash or malfunction in any way.
		794
		795
480	=head2 C<ev_io> - is this file descriptor readable or writable	796	=head2 C<ev_io> - is this file descriptor readable or writable?
481		797
482	I/O watchers check whether a file descriptor is readable or writable	798	I/O watchers check whether a file descriptor is readable or writable
483	in each iteration of the event loop (This behaviour is called	799	in each iteration of the event loop, or, more precisely, when reading
484	level-triggering because you keep receiving events as long as the	800	would not block the process and writing would at least be able to write
485	condition persists. Remember you can stop the watcher if you don't want to	801	some data. This behaviour is called level-triggering because you keep
486	act on the event and neither want to receive future events).	802	receiving events as long as the condition persists. Remember you can stop
		803	the watcher if you don't want to act on the event and neither want to
		804	receive future events.
487		805
488	In general you can register as many read and/or write event watchers per	806	In general you can register as many read and/or write event watchers per
489	fd as you want (as long as you don't confuse yourself). Setting all file	807	fd as you want (as long as you don't confuse yourself). Setting all file
490	descriptors to non-blocking mode is also usually a good idea (but not	808	descriptors to non-blocking mode is also usually a good idea (but not
491	required if you know what you are doing).	809	required if you know what you are doing).
492		810
493	You have to be careful with dup'ed file descriptors, though. Some backends	811	You have to be careful with dup'ed file descriptors, though. Some backends
494	(the linux epoll backend is a notable example) cannot handle dup'ed file	812	(the linux epoll backend is a notable example) cannot handle dup'ed file
495	descriptors correctly if you register interest in two or more fds pointing	813	descriptors correctly if you register interest in two or more fds pointing
496	to the same underlying file/socket etc. description (that is, they share	814	to the same underlying file/socket/etc. description (that is, they share
497	the same underlying "file open").	815	the same underlying "file open").
498		816
499	If you must do this, then force the use of a known-to-be-good backend	817	If you must do this, then force the use of a known-to-be-good backend
500	(at the time of this writing, this includes only EVMETHOD_SELECT and	818	(at the time of this writing, this includes only C<EVBACKEND_SELECT> and
501	EVMETHOD_POLL).	819	C<EVBACKEND_POLL>).
		820
		821	Another thing you have to watch out for is that it is quite easy to
		822	receive "spurious" readyness notifications, that is your callback might
		823	be called with C<EV_READ> but a subsequent C<read>(2) will actually block
		824	because there is no data. Not only are some backends known to create a
		825	lot of those (for example solaris ports), it is very easy to get into
		826	this situation even with a relatively standard program structure. Thus
		827	it is best to always use non-blocking I/O: An extra C<read>(2) returning
		828	C<EAGAIN> is far preferable to a program hanging until some data arrives.
		829
		830	If you cannot run the fd in non-blocking mode (for example you should not
		831	play around with an Xlib connection), then you have to seperately re-test
		832	wether a file descriptor is really ready with a known-to-be good interface
		833	such as poll (fortunately in our Xlib example, Xlib already does this on
		834	its own, so its quite safe to use).
502		835
503	=over 4	836	=over 4
504		837
505	=item ev_io_init (ev_io *, callback, int fd, int events)	838	=item ev_io_init (ev_io *, callback, int fd, int events)
506		839
507	=item ev_io_set (ev_io *, int fd, int events)	840	=item ev_io_set (ev_io *, int fd, int events)
508		841
509	Configures an C<ev_io> watcher. The fd is the file descriptor to rceeive	842	Configures an C<ev_io> watcher. The C<fd> is the file descriptor to
510	events for and events is either C<EV_READ>, C<EV_WRITE> or C<EV_READ |	843	rceeive events for and events is either C<EV_READ>, C<EV_WRITE> or
511	EV_WRITE> to receive the given events.	844	C<EV_READ | EV_WRITE> to receive the given events.
		845
		846	=item int fd [read-only]
		847
		848	The file descriptor being watched.
		849
		850	=item int events [read-only]
		851
		852	The events being watched.
512		853
513	=back	854	=back
514		855
		856	Example: Call C<stdin_readable_cb> when STDIN_FILENO has become, well
		857	readable, but only once. Since it is likely line-buffered, you could
		858	attempt to read a whole line in the callback.
		859
		860	static void
		861	stdin_readable_cb (struct ev_loop *loop, struct ev_io *w, int revents)
		862	{
		863	ev_io_stop (loop, w);
		864	.. read from stdin here (or from w->fd) and haqndle any I/O errors
		865	}
		866
		867	...
		868	struct ev_loop *loop = ev_default_init (0);
		869	struct ev_io stdin_readable;
		870	ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
		871	ev_io_start (loop, &stdin_readable);
		872	ev_loop (loop, 0);
		873
		874
515	=head2 C<ev_timer> - relative and optionally recurring timeouts	875	=head2 C<ev_timer> - relative and optionally repeating timeouts
516		876
517	Timer watchers are simple relative timers that generate an event after a	877	Timer watchers are simple relative timers that generate an event after a
518	given time, and optionally repeating in regular intervals after that.	878	given time, and optionally repeating in regular intervals after that.
519		879
520	The timers are based on real time, that is, if you register an event that	880	The timers are based on real time, that is, if you register an event that
…		…
561		921
562	If the timer is repeating, either start it if necessary (with the repeat	922	If the timer is repeating, either start it if necessary (with the repeat
563	value), or reset the running timer to the repeat value.	923	value), or reset the running timer to the repeat value.
564		924
565	This sounds a bit complicated, but here is a useful and typical	925	This sounds a bit complicated, but here is a useful and typical
566	example: Imagine you have a tcp connection and you want a so-called idle	926	example: Imagine you have a tcp connection and you want a so-called
567	timeout, that is, you want to be called when there have been, say, 60	927	idle timeout, that is, you want to be called when there have been,
568	seconds of inactivity on the socket. The easiest way to do this is to	928	say, 60 seconds of inactivity on the socket. The easiest way to do
569	configure an C<ev_timer> with after=repeat=60 and calling ev_timer_again each	929	this is to configure an C<ev_timer> with C<after>=C<repeat>=C<60> and calling
570	time you successfully read or write some data. If you go into an idle	930	C<ev_timer_again> each time you successfully read or write some data. If
571	state where you do not expect data to travel on the socket, you can stop	931	you go into an idle state where you do not expect data to travel on the
572	the timer, and again will automatically restart it if need be.	932	socket, you can stop the timer, and again will automatically restart it if
		933	need be.
		934
		935	You can also ignore the C<after> value and C<ev_timer_start> altogether
		936	and only ever use the C<repeat> value:
		937
		938	ev_timer_init (timer, callback, 0., 5.);
		939	ev_timer_again (loop, timer);
		940	...
		941	timer->again = 17.;
		942	ev_timer_again (loop, timer);
		943	...
		944	timer->again = 10.;
		945	ev_timer_again (loop, timer);
		946
		947	This is more efficient then stopping/starting the timer eahc time you want
		948	to modify its timeout value.
		949
		950	=item ev_tstamp repeat [read-write]
		951
		952	The current C<repeat> value. Will be used each time the watcher times out
		953	or C<ev_timer_again> is called and determines the next timeout (if any),
		954	which is also when any modifications are taken into account.
573		955
574	=back	956	=back
575		957
		958	Example: Create a timer that fires after 60 seconds.
		959
		960	static void
		961	one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
		962	{
		963	.. one minute over, w is actually stopped right here
		964	}
		965
		966	struct ev_timer mytimer;
		967	ev_timer_init (&mytimer, one_minute_cb, 60., 0.);
		968	ev_timer_start (loop, &mytimer);
		969
		970	Example: Create a timeout timer that times out after 10 seconds of
		971	inactivity.
		972
		973	static void
		974	timeout_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
		975	{
		976	.. ten seconds without any activity
		977	}
		978
		979	struct ev_timer mytimer;
		980	ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */
		981	ev_timer_again (&mytimer); /* start timer */
		982	ev_loop (loop, 0);
		983
		984	// and in some piece of code that gets executed on any "activity":
		985	// reset the timeout to start ticking again at 10 seconds
		986	ev_timer_again (&mytimer);
		987
		988
576	=head2 C<ev_periodic> - to cron or not to cron	989	=head2 C<ev_periodic> - to cron or not to cron?
577		990
578	Periodic watchers are also timers of a kind, but they are very versatile	991	Periodic watchers are also timers of a kind, but they are very versatile
579	(and unfortunately a bit complex).	992	(and unfortunately a bit complex).
580		993
581	Unlike C<ev_timer>'s, they are not based on real time (or relative time)	994	Unlike C<ev_timer>'s, they are not based on real time (or relative time)
582	but on wallclock time (absolute time). You can tell a periodic watcher	995	but on wallclock time (absolute time). You can tell a periodic watcher
583	to trigger "at" some specific point in time. For example, if you tell a	996	to trigger "at" some specific point in time. For example, if you tell a
584	periodic watcher to trigger in 10 seconds (by specifiying e.g. c<ev_now ()	997	periodic watcher to trigger in 10 seconds (by specifiying e.g. C<ev_now ()
585	+ 10.>) and then reset your system clock to the last year, then it will	998	+ 10.>) and then reset your system clock to the last year, then it will
586	take a year to trigger the event (unlike an C<ev_timer>, which would trigger	999	take a year to trigger the event (unlike an C<ev_timer>, which would trigger
587	roughly 10 seconds later and of course not if you reset your system time	1000	roughly 10 seconds later and of course not if you reset your system time
588	again).	1001	again).
589		1002
…		…
673	Simply stops and restarts the periodic watcher again. This is only useful	1086	Simply stops and restarts the periodic watcher again. This is only useful
674	when you changed some parameters or the reschedule callback would return	1087	when you changed some parameters or the reschedule callback would return
675	a different time than the last time it was called (e.g. in a crond like	1088	a different time than the last time it was called (e.g. in a crond like
676	program when the crontabs have changed).	1089	program when the crontabs have changed).
677		1090
		1091	=item ev_tstamp interval [read-write]
		1092
		1093	The current interval value. Can be modified any time, but changes only
		1094	take effect when the periodic timer fires or C<ev_periodic_again> is being
		1095	called.
		1096
		1097	=item ev_tstamp (*reschedule_cb)(struct ev_periodic *w, ev_tstamp now) [read-write]
		1098
		1099	The current reschedule callback, or C<0>, if this functionality is
		1100	switched off. Can be changed any time, but changes only take effect when
		1101	the periodic timer fires or C<ev_periodic_again> is being called.
		1102
678	=back	1103	=back
679		1104
		1105	Example: Call a callback every hour, or, more precisely, whenever the
		1106	system clock is divisible by 3600. The callback invocation times have
		1107	potentially a lot of jittering, but good long-term stability.
		1108
		1109	static void
		1110	clock_cb (struct ev_loop *loop, struct ev_io *w, int revents)
		1111	{
		1112	... its now a full hour (UTC, or TAI or whatever your clock follows)
		1113	}
		1114
		1115	struct ev_periodic hourly_tick;
		1116	ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);
		1117	ev_periodic_start (loop, &hourly_tick);
		1118
		1119	Example: The same as above, but use a reschedule callback to do it:
		1120
		1121	#include <math.h>
		1122
		1123	static ev_tstamp
		1124	my_scheduler_cb (struct ev_periodic *w, ev_tstamp now)
		1125	{
		1126	return fmod (now, 3600.) + 3600.;
		1127	}
		1128
		1129	ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
		1130
		1131	Example: Call a callback every hour, starting now:
		1132
		1133	struct ev_periodic hourly_tick;
		1134	ev_periodic_init (&hourly_tick, clock_cb,
		1135	fmod (ev_now (loop), 3600.), 3600., 0);
		1136	ev_periodic_start (loop, &hourly_tick);
		1137
		1138
680	=head2 C<ev_signal> - signal me when a signal gets signalled	1139	=head2 C<ev_signal> - signal me when a signal gets signalled!
681		1140
682	Signal watchers will trigger an event when the process receives a specific	1141	Signal watchers will trigger an event when the process receives a specific
683	signal one or more times. Even though signals are very asynchronous, libev	1142	signal one or more times. Even though signals are very asynchronous, libev
684	will try it's best to deliver signals synchronously, i.e. as part of the	1143	will try it's best to deliver signals synchronously, i.e. as part of the
685	normal event processing, like any other event.	1144	normal event processing, like any other event.
…		…
698	=item ev_signal_set (ev_signal *, int signum)	1157	=item ev_signal_set (ev_signal *, int signum)
699		1158
700	Configures the watcher to trigger on the given signal number (usually one	1159	Configures the watcher to trigger on the given signal number (usually one
701	of the C<SIGxxx> constants).	1160	of the C<SIGxxx> constants).
702		1161
		1162	=item int signum [read-only]
		1163
		1164	The signal the watcher watches out for.
		1165
703	=back	1166	=back
704		1167
		1168
705	=head2 C<ev_child> - wait for pid status changes	1169	=head2 C<ev_child> - watch out for process status changes
706		1170
707	Child watchers trigger when your process receives a SIGCHLD in response to	1171	Child watchers trigger when your process receives a SIGCHLD in response to
708	some child status changes (most typically when a child of yours dies).	1172	some child status changes (most typically when a child of yours dies).
709		1173
710	=over 4	1174	=over 4
…		…
718	at the C<rstatus> member of the C<ev_child> watcher structure to see	1182	at the C<rstatus> member of the C<ev_child> watcher structure to see
719	the status word (use the macros from C<sys/wait.h> and see your systems	1183	the status word (use the macros from C<sys/wait.h> and see your systems
720	C<waitpid> documentation). The C<rpid> member contains the pid of the	1184	C<waitpid> documentation). The C<rpid> member contains the pid of the
721	process causing the status change.	1185	process causing the status change.
722		1186
		1187	=item int pid [read-only]
		1188
		1189	The process id this watcher watches out for, or C<0>, meaning any process id.
		1190
		1191	=item int rpid [read-write]
		1192
		1193	The process id that detected a status change.
		1194
		1195	=item int rstatus [read-write]
		1196
		1197	The process exit/trace status caused by C<rpid> (see your systems
		1198	C<waitpid> and C<sys/wait.h> documentation for details).
		1199
723	=back	1200	=back
724		1201
		1202	Example: Try to exit cleanly on SIGINT and SIGTERM.
		1203
		1204	static void
		1205	sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)
		1206	{
		1207	ev_unloop (loop, EVUNLOOP_ALL);
		1208	}
		1209
		1210	struct ev_signal signal_watcher;
		1211	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
		1212	ev_signal_start (loop, &sigint_cb);
		1213
		1214
		1215	=head2 C<ev_stat> - did the file attributes just change?
		1216
		1217	This watches a filesystem path for attribute changes. That is, it calls
		1218	C<stat> regularly (or when the OS says it changed) and sees if it changed
		1219	compared to the last time, invoking the callback if it did.
		1220
		1221	The path does not need to exist: changing from "path exists" to "path does
		1222	not exist" is a status change like any other. The condition "path does
		1223	not exist" is signified by the C<st_nlink> field being zero (which is
		1224	otherwise always forced to be at least one) and all the other fields of
		1225	the stat buffer having unspecified contents.
		1226
		1227	Since there is no standard to do this, the portable implementation simply
		1228	calls C<stat (2)> regularly on the path to see if it changed somehow. You
		1229	can specify a recommended polling interval for this case. If you specify
		1230	a polling interval of C<0> (highly recommended!) then a I<suitable,
		1231	unspecified default> value will be used (which you can expect to be around
		1232	five seconds, although this might change dynamically). Libev will also
		1233	impose a minimum interval which is currently around C<0.1>, but thats
		1234	usually overkill.
		1235
		1236	This watcher type is not meant for massive numbers of stat watchers,
		1237	as even with OS-supported change notifications, this can be
		1238	resource-intensive.
		1239
		1240	At the time of this writing, only the Linux inotify interface is
		1241	implemented (implementing kqueue support is left as an exercise for the
		1242	reader). Inotify will be used to give hints only and should not change the
		1243	semantics of C<ev_stat> watchers, which means that libev sometimes needs
		1244	to fall back to regular polling again even with inotify, but changes are
		1245	usually detected immediately, and if the file exists there will be no
		1246	polling.
		1247
		1248	=over 4
		1249
		1250	=item ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)
		1251
		1252	=item ev_stat_set (ev_stat *, const char *path, ev_tstamp interval)
		1253
		1254	Configures the watcher to wait for status changes of the given
		1255	C<path>. The C<interval> is a hint on how quickly a change is expected to
		1256	be detected and should normally be specified as C<0> to let libev choose
		1257	a suitable value. The memory pointed to by C<path> must point to the same
		1258	path for as long as the watcher is active.
		1259
		1260	The callback will be receive C<EV_STAT> when a change was detected,
		1261	relative to the attributes at the time the watcher was started (or the
		1262	last change was detected).
		1263
		1264	=item ev_stat_stat (ev_stat *)
		1265
		1266	Updates the stat buffer immediately with new values. If you change the
		1267	watched path in your callback, you could call this fucntion to avoid
		1268	detecting this change (while introducing a race condition). Can also be
		1269	useful simply to find out the new values.
		1270
		1271	=item ev_statdata attr [read-only]
		1272
		1273	The most-recently detected attributes of the file. Although the type is of
		1274	C<ev_statdata>, this is usually the (or one of the) C<struct stat> types
		1275	suitable for your system. If the C<st_nlink> member is C<0>, then there
		1276	was some error while C<stat>ing the file.
		1277
		1278	=item ev_statdata prev [read-only]
		1279
		1280	The previous attributes of the file. The callback gets invoked whenever
		1281	C<prev> != C<attr>.
		1282
		1283	=item ev_tstamp interval [read-only]
		1284
		1285	The specified interval.
		1286
		1287	=item const char *path [read-only]
		1288
		1289	The filesystem path that is being watched.
		1290
		1291	=back
		1292
		1293	Example: Watch C</etc/passwd> for attribute changes.
		1294
		1295	static void
		1296	passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)
		1297	{
		1298	/* /etc/passwd changed in some way */
		1299	if (w->attr.st_nlink)
		1300	{
		1301	printf ("passwd current size %ld\n", (long)w->attr.st_size);
		1302	printf ("passwd current atime %ld\n", (long)w->attr.st_mtime);
		1303	printf ("passwd current mtime %ld\n", (long)w->attr.st_mtime);
		1304	}
		1305	else
		1306	/* you shalt not abuse printf for puts */
		1307	puts ("wow, /etc/passwd is not there, expect problems. "
		1308	"if this is windows, they already arrived\n");
		1309	}
		1310
		1311	...
		1312	ev_stat passwd;
		1313
		1314	ev_stat_init (&passwd, passwd_cb, "/etc/passwd");
		1315	ev_stat_start (loop, &passwd);
		1316
		1317
725	=head2 C<ev_idle> - when you've got nothing better to do	1318	=head2 C<ev_idle> - when you've got nothing better to do...
726		1319
727	Idle watchers trigger events when there are no other events are pending	1320	Idle watchers trigger events when there are no other events are pending
728	(prepare, check and other idle watchers do not count). That is, as long	1321	(prepare, check and other idle watchers do not count). That is, as long
729	as your process is busy handling sockets or timeouts (or even signals,	1322	as your process is busy handling sockets or timeouts (or even signals,
730	imagine) it will not be triggered. But when your process is idle all idle	1323	imagine) it will not be triggered. But when your process is idle all idle
…		…
748	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,	1341	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,
749	believe me.	1342	believe me.
750		1343
751	=back	1344	=back
752		1345
		1346	Example: Dynamically allocate an C<ev_idle> watcher, start it, and in the
		1347	callback, free it. Also, use no error checking, as usual.
		1348
		1349	static void
		1350	idle_cb (struct ev_loop *loop, struct ev_idle *w, int revents)
		1351	{
		1352	free (w);
		1353	// now do something you wanted to do when the program has
		1354	// no longer asnything immediate to do.
		1355	}
		1356
		1357	struct ev_idle *idle_watcher = malloc (sizeof (struct ev_idle));
		1358	ev_idle_init (idle_watcher, idle_cb);
		1359	ev_idle_start (loop, idle_cb);
		1360
		1361
753	=head2 C<ev_prepare> and C<ev_check> - customise your event loop	1362	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!
754		1363
755	Prepare and check watchers are usually (but not always) used in tandem:	1364	Prepare and check watchers are usually (but not always) used in tandem:
756	prepare watchers get invoked before the process blocks and check watchers	1365	prepare watchers get invoked before the process blocks and check watchers
757	afterwards.	1366	afterwards.
758		1367
		1368	You I<must not> call C<ev_loop> or similar functions that enter
		1369	the current event loop from either C<ev_prepare> or C<ev_check>
		1370	watchers. Other loops than the current one are fine, however. The
		1371	rationale behind this is that you do not need to check for recursion in
		1372	those watchers, i.e. the sequence will always be C<ev_prepare>, blocking,
		1373	C<ev_check> so if you have one watcher of each kind they will always be
		1374	called in pairs bracketing the blocking call.
		1375
759	Their main purpose is to integrate other event mechanisms into libev. This	1376	Their main purpose is to integrate other event mechanisms into libev and
760	could be used, for example, to track variable changes, implement your own	1377	their use is somewhat advanced. This could be used, for example, to track
761	watchers, integrate net-snmp or a coroutine library and lots more.	1378	variable changes, implement your own watchers, integrate net-snmp or a
		1379	coroutine library and lots more. They are also occasionally useful if
		1380	you cache some data and want to flush it before blocking (for example,
		1381	in X programs you might want to do an C<XFlush ()> in an C<ev_prepare>
		1382	watcher).
762		1383
763	This is done by examining in each prepare call which file descriptors need	1384	This is done by examining in each prepare call which file descriptors need
764	to be watched by the other library, registering C<ev_io> watchers for	1385	to be watched by the other library, registering C<ev_io> watchers for
765	them and starting an C<ev_timer> watcher for any timeouts (many libraries	1386	them and starting an C<ev_timer> watcher for any timeouts (many libraries
766	provide just this functionality). Then, in the check watcher you check for	1387	provide just this functionality). Then, in the check watcher you check for
…		…
788	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>	1409	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>
789	macros, but using them is utterly, utterly and completely pointless.	1410	macros, but using them is utterly, utterly and completely pointless.
790		1411
791	=back	1412	=back
792		1413
		1414	Example: To include a library such as adns, you would add IO watchers
		1415	and a timeout watcher in a prepare handler, as required by libadns, and
		1416	in a check watcher, destroy them and call into libadns. What follows is
		1417	pseudo-code only of course:
		1418
		1419	static ev_io iow [nfd];
		1420	static ev_timer tw;
		1421
		1422	static void
		1423	io_cb (ev_loop *loop, ev_io *w, int revents)
		1424	{
		1425	// set the relevant poll flags
		1426	// could also call adns_processreadable etc. here
		1427	struct pollfd *fd = (struct pollfd *)w->data;
		1428	if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;
		1429	if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;
		1430	}
		1431
		1432	// create io watchers for each fd and a timer before blocking
		1433	static void
		1434	adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents)
		1435	{
		1436	int timeout = 3600000;truct pollfd fds [nfd];
		1437	// actual code will need to loop here and realloc etc.
		1438	adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));
		1439
		1440	/* the callback is illegal, but won't be called as we stop during check */
		1441	ev_timer_init (&tw, 0, timeout * 1e-3);
		1442	ev_timer_start (loop, &tw);
		1443
		1444	// create on ev_io per pollfd
		1445	for (int i = 0; i < nfd; ++i)
		1446	{
		1447	ev_io_init (iow + i, io_cb, fds [i].fd,
		1448	((fds [i].events & POLLIN ? EV_READ : 0)
		1449	| (fds [i].events & POLLOUT ? EV_WRITE : 0)));
		1450
		1451	fds [i].revents = 0;
		1452	iow [i].data = fds + i;
		1453	ev_io_start (loop, iow + i);
		1454	}
		1455	}
		1456
		1457	// stop all watchers after blocking
		1458	static void
		1459	adns_check_cb (ev_loop *loop, ev_check *w, int revents)
		1460	{
		1461	ev_timer_stop (loop, &tw);
		1462
		1463	for (int i = 0; i < nfd; ++i)
		1464	ev_io_stop (loop, iow + i);
		1465
		1466	adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));
		1467	}
		1468
		1469
		1470	=head2 C<ev_embed> - when one backend isn't enough...
		1471
		1472	This is a rather advanced watcher type that lets you embed one event loop
		1473	into another (currently only C<ev_io> events are supported in the embedded
		1474	loop, other types of watchers might be handled in a delayed or incorrect
		1475	fashion and must not be used).
		1476
		1477	There are primarily two reasons you would want that: work around bugs and
		1478	prioritise I/O.
		1479
		1480	As an example for a bug workaround, the kqueue backend might only support
		1481	sockets on some platform, so it is unusable as generic backend, but you
		1482	still want to make use of it because you have many sockets and it scales
		1483	so nicely. In this case, you would create a kqueue-based loop and embed it
		1484	into your default loop (which might use e.g. poll). Overall operation will
		1485	be a bit slower because first libev has to poll and then call kevent, but
		1486	at least you can use both at what they are best.
		1487
		1488	As for prioritising I/O: rarely you have the case where some fds have
		1489	to be watched and handled very quickly (with low latency), and even
		1490	priorities and idle watchers might have too much overhead. In this case
		1491	you would put all the high priority stuff in one loop and all the rest in
		1492	a second one, and embed the second one in the first.
		1493
		1494	As long as the watcher is active, the callback will be invoked every time
		1495	there might be events pending in the embedded loop. The callback must then
		1496	call C<ev_embed_sweep (mainloop, watcher)> to make a single sweep and invoke
		1497	their callbacks (you could also start an idle watcher to give the embedded
		1498	loop strictly lower priority for example). You can also set the callback
		1499	to C<0>, in which case the embed watcher will automatically execute the
		1500	embedded loop sweep.
		1501
		1502	As long as the watcher is started it will automatically handle events. The
		1503	callback will be invoked whenever some events have been handled. You can
		1504	set the callback to C<0> to avoid having to specify one if you are not
		1505	interested in that.
		1506
		1507	Also, there have not currently been made special provisions for forking:
		1508	when you fork, you not only have to call C<ev_loop_fork> on both loops,
		1509	but you will also have to stop and restart any C<ev_embed> watchers
		1510	yourself.
		1511
		1512	Unfortunately, not all backends are embeddable, only the ones returned by
		1513	C<ev_embeddable_backends> are, which, unfortunately, does not include any
		1514	portable one.
		1515
		1516	So when you want to use this feature you will always have to be prepared
		1517	that you cannot get an embeddable loop. The recommended way to get around
		1518	this is to have a separate variables for your embeddable loop, try to
		1519	create it, and if that fails, use the normal loop for everything:
		1520
		1521	struct ev_loop *loop_hi = ev_default_init (0);
		1522	struct ev_loop *loop_lo = 0;
		1523	struct ev_embed embed;
		1524
		1525	// see if there is a chance of getting one that works
		1526	// (remember that a flags value of 0 means autodetection)
		1527	loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
		1528	? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
		1529	: 0;
		1530
		1531	// if we got one, then embed it, otherwise default to loop_hi
		1532	if (loop_lo)
		1533	{
		1534	ev_embed_init (&embed, 0, loop_lo);
		1535	ev_embed_start (loop_hi, &embed);
		1536	}
		1537	else
		1538	loop_lo = loop_hi;
		1539
		1540	=over 4
		1541
		1542	=item ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)
		1543
		1544	=item ev_embed_set (ev_embed *, callback, struct ev_loop *embedded_loop)
		1545
		1546	Configures the watcher to embed the given loop, which must be
		1547	embeddable. If the callback is C<0>, then C<ev_embed_sweep> will be
		1548	invoked automatically, otherwise it is the responsibility of the callback
		1549	to invoke it (it will continue to be called until the sweep has been done,
		1550	if you do not want thta, you need to temporarily stop the embed watcher).
		1551
		1552	=item ev_embed_sweep (loop, ev_embed *)
		1553
		1554	Make a single, non-blocking sweep over the embedded loop. This works
		1555	similarly to C<ev_loop (embedded_loop, EVLOOP_NONBLOCK)>, but in the most
		1556	apropriate way for embedded loops.
		1557
		1558	=item struct ev_loop *loop [read-only]
		1559
		1560	The embedded event loop.
		1561
		1562	=back
		1563
		1564
		1565	=head2 C<ev_fork> - the audacity to resume the event loop after a fork
		1566
		1567	Fork watchers are called when a C<fork ()> was detected (usually because
		1568	whoever is a good citizen cared to tell libev about it by calling
		1569	C<ev_default_fork> or C<ev_loop_fork>). The invocation is done before the
		1570	event loop blocks next and before C<ev_check> watchers are being called,
		1571	and only in the child after the fork. If whoever good citizen calling
		1572	C<ev_default_fork> cheats and calls it in the wrong process, the fork
		1573	handlers will be invoked, too, of course.
		1574
		1575	=over 4
		1576
		1577	=item ev_fork_init (ev_signal *, callback)
		1578
		1579	Initialises and configures the fork watcher - it has no parameters of any
		1580	kind. There is a C<ev_fork_set> macro, but using it is utterly pointless,
		1581	believe me.
		1582
		1583	=back
		1584
		1585
793	=head1 OTHER FUNCTIONS	1586	=head1 OTHER FUNCTIONS
794		1587
795	There are some other functions of possible interest. Described. Here. Now.	1588	There are some other functions of possible interest. Described. Here. Now.
796		1589
797	=over 4	1590	=over 4
…		…
826	/* stdin might have data for us, joy! */;	1619	/* stdin might have data for us, joy! */;
827	}	1620	}
828		1621
829	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);	1622	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
830		1623
831	=item ev_feed_event (loop, watcher, int events)	1624	=item ev_feed_event (ev_loop *, watcher *, int revents)
832		1625
833	Feeds the given event set into the event loop, as if the specified event	1626	Feeds the given event set into the event loop, as if the specified event
834	had happened for the specified watcher (which must be a pointer to an	1627	had happened for the specified watcher (which must be a pointer to an
835	initialised but not necessarily started event watcher).	1628	initialised but not necessarily started event watcher).
836		1629
837	=item ev_feed_fd_event (loop, int fd, int revents)	1630	=item ev_feed_fd_event (ev_loop *, int fd, int revents)
838		1631
839	Feed an event on the given fd, as if a file descriptor backend detected	1632	Feed an event on the given fd, as if a file descriptor backend detected
840	the given events it.	1633	the given events it.
841		1634
842	=item ev_feed_signal_event (loop, int signum)	1635	=item ev_feed_signal_event (ev_loop *loop, int signum)
843		1636
844	Feed an event as if the given signal occured (loop must be the default loop!).	1637	Feed an event as if the given signal occured (C<loop> must be the default
		1638	loop!).
845		1639
846	=back	1640	=back
		1641
847		1642
848	=head1 LIBEVENT EMULATION	1643	=head1 LIBEVENT EMULATION
849		1644
850	Libev offers a compatibility emulation layer for libevent. It cannot	1645	Libev offers a compatibility emulation layer for libevent. It cannot
851	emulate the internals of libevent, so here are some usage hints:	1646	emulate the internals of libevent, so here are some usage hints:
…		…
872		1667
873	=back	1668	=back
874		1669
875	=head1 C++ SUPPORT	1670	=head1 C++ SUPPORT
876		1671
877	TBD.	1672	Libev comes with some simplistic wrapper classes for C++ that mainly allow
		1673	you to use some convinience methods to start/stop watchers and also change
		1674	the callback model to a model using method callbacks on objects.
		1675
		1676	To use it,
		1677
		1678	#include <ev++.h>
		1679
		1680	(it is not installed by default). This automatically includes F<ev.h>
		1681	and puts all of its definitions (many of them macros) into the global
		1682	namespace. All C++ specific things are put into the C<ev> namespace.
		1683
		1684	It should support all the same embedding options as F<ev.h>, most notably
		1685	C<EV_MULTIPLICITY>.
		1686
		1687	Here is a list of things available in the C<ev> namespace:
		1688
		1689	=over 4
		1690
		1691	=item C<ev::READ>, C<ev::WRITE> etc.
		1692
		1693	These are just enum values with the same values as the C<EV_READ> etc.
		1694	macros from F<ev.h>.
		1695
		1696	=item C<ev::tstamp>, C<ev::now>
		1697
		1698	Aliases to the same types/functions as with the C<ev_> prefix.
		1699
		1700	=item C<ev::io>, C<ev::timer>, C<ev::periodic>, C<ev::idle>, C<ev::sig> etc.
		1701
		1702	For each C<ev_TYPE> watcher in F<ev.h> there is a corresponding class of
		1703	the same name in the C<ev> namespace, with the exception of C<ev_signal>
		1704	which is called C<ev::sig> to avoid clashes with the C<signal> macro
		1705	defines by many implementations.
		1706
		1707	All of those classes have these methods:
		1708
		1709	=over 4
		1710
		1711	=item ev::TYPE::TYPE (object *, object::method *)
		1712
		1713	=item ev::TYPE::TYPE (object *, object::method *, struct ev_loop *)
		1714
		1715	=item ev::TYPE::~TYPE
		1716
		1717	The constructor takes a pointer to an object and a method pointer to
		1718	the event handler callback to call in this class. The constructor calls
		1719	C<ev_init> for you, which means you have to call the C<set> method
		1720	before starting it. If you do not specify a loop then the constructor
		1721	automatically associates the default loop with this watcher.
		1722
		1723	The destructor automatically stops the watcher if it is active.
		1724
		1725	=item w->set (struct ev_loop *)
		1726
		1727	Associates a different C<struct ev_loop> with this watcher. You can only
		1728	do this when the watcher is inactive (and not pending either).
		1729
		1730	=item w->set ([args])
		1731
		1732	Basically the same as C<ev_TYPE_set>, with the same args. Must be
		1733	called at least once. Unlike the C counterpart, an active watcher gets
		1734	automatically stopped and restarted.
		1735
		1736	=item w->start ()
		1737
		1738	Starts the watcher. Note that there is no C<loop> argument as the
		1739	constructor already takes the loop.
		1740
		1741	=item w->stop ()
		1742
		1743	Stops the watcher if it is active. Again, no C<loop> argument.
		1744
		1745	=item w->again () C<ev::timer>, C<ev::periodic> only
		1746
		1747	For C<ev::timer> and C<ev::periodic>, this invokes the corresponding
		1748	C<ev_TYPE_again> function.
		1749
		1750	=item w->sweep () C<ev::embed> only
		1751
		1752	Invokes C<ev_embed_sweep>.
		1753
		1754	=item w->update () C<ev::stat> only
		1755
		1756	Invokes C<ev_stat_stat>.
		1757
		1758	=back
		1759
		1760	=back
		1761
		1762	Example: Define a class with an IO and idle watcher, start one of them in
		1763	the constructor.
		1764
		1765	class myclass
		1766	{
		1767	ev_io io; void io_cb (ev::io &w, int revents);
		1768	ev_idle idle void idle_cb (ev::idle &w, int revents);
		1769
		1770	myclass ();
		1771	}
		1772
		1773	myclass::myclass (int fd)
		1774	: io (this, &myclass::io_cb),
		1775	idle (this, &myclass::idle_cb)
		1776	{
		1777	io.start (fd, ev::READ);
		1778	}
		1779
		1780
		1781	=head1 MACRO MAGIC
		1782
		1783	Libev can be compiled with a variety of options, the most fundemantal is
		1784	C<EV_MULTIPLICITY>. This option determines wether (most) functions and
		1785	callbacks have an initial C<struct ev_loop *> argument.
		1786
		1787	To make it easier to write programs that cope with either variant, the
		1788	following macros are defined:
		1789
		1790	=over 4
		1791
		1792	=item C<EV_A>, C<EV_A_>
		1793
		1794	This provides the loop I<argument> for functions, if one is required ("ev
		1795	loop argument"). The C<EV_A> form is used when this is the sole argument,
		1796	C<EV_A_> is used when other arguments are following. Example:
		1797
		1798	ev_unref (EV_A);
		1799	ev_timer_add (EV_A_ watcher);
		1800	ev_loop (EV_A_ 0);
		1801
		1802	It assumes the variable C<loop> of type C<struct ev_loop *> is in scope,
		1803	which is often provided by the following macro.
		1804
		1805	=item C<EV_P>, C<EV_P_>
		1806
		1807	This provides the loop I<parameter> for functions, if one is required ("ev
		1808	loop parameter"). The C<EV_P> form is used when this is the sole parameter,
		1809	C<EV_P_> is used when other parameters are following. Example:
		1810
		1811	// this is how ev_unref is being declared
		1812	static void ev_unref (EV_P);
		1813
		1814	// this is how you can declare your typical callback
		1815	static void cb (EV_P_ ev_timer *w, int revents)
		1816
		1817	It declares a parameter C<loop> of type C<struct ev_loop *>, quite
		1818	suitable for use with C<EV_A>.
		1819
		1820	=item C<EV_DEFAULT>, C<EV_DEFAULT_>
		1821
		1822	Similar to the other two macros, this gives you the value of the default
		1823	loop, if multiple loops are supported ("ev loop default").
		1824
		1825	=back
		1826
		1827	Example: Declare and initialise a check watcher, working regardless of
		1828	wether multiple loops are supported or not.
		1829
		1830	static void
		1831	check_cb (EV_P_ ev_timer *w, int revents)
		1832	{
		1833	ev_check_stop (EV_A_ w);
		1834	}
		1835
		1836	ev_check check;
		1837	ev_check_init (&check, check_cb);
		1838	ev_check_start (EV_DEFAULT_ &check);
		1839	ev_loop (EV_DEFAULT_ 0);
		1840
		1841
		1842	=head1 EMBEDDING
		1843
		1844	Libev can (and often is) directly embedded into host
		1845	applications. Examples of applications that embed it include the Deliantra
		1846	Game Server, the EV perl module, the GNU Virtual Private Ethernet (gvpe)
		1847	and rxvt-unicode.
		1848
		1849	The goal is to enable you to just copy the neecssary files into your
		1850	source directory without having to change even a single line in them, so
		1851	you can easily upgrade by simply copying (or having a checked-out copy of
		1852	libev somewhere in your source tree).
		1853
		1854	=head2 FILESETS
		1855
		1856	Depending on what features you need you need to include one or more sets of files
		1857	in your app.
		1858
		1859	=head3 CORE EVENT LOOP
		1860
		1861	To include only the libev core (all the C<ev_*> functions), with manual
		1862	configuration (no autoconf):
		1863
		1864	#define EV_STANDALONE 1
		1865	#include "ev.c"
		1866
		1867	This will automatically include F<ev.h>, too, and should be done in a
		1868	single C source file only to provide the function implementations. To use
		1869	it, do the same for F<ev.h> in all files wishing to use this API (best
		1870	done by writing a wrapper around F<ev.h> that you can include instead and
		1871	where you can put other configuration options):
		1872
		1873	#define EV_STANDALONE 1
		1874	#include "ev.h"
		1875
		1876	Both header files and implementation files can be compiled with a C++
		1877	compiler (at least, thats a stated goal, and breakage will be treated
		1878	as a bug).
		1879
		1880	You need the following files in your source tree, or in a directory
		1881	in your include path (e.g. in libev/ when using -Ilibev):
		1882
		1883	ev.h
		1884	ev.c
		1885	ev_vars.h
		1886	ev_wrap.h
		1887
		1888	ev_win32.c required on win32 platforms only
		1889
		1890	ev_select.c only when select backend is enabled (which is by default)
		1891	ev_poll.c only when poll backend is enabled (disabled by default)
		1892	ev_epoll.c only when the epoll backend is enabled (disabled by default)
		1893	ev_kqueue.c only when the kqueue backend is enabled (disabled by default)
		1894	ev_port.c only when the solaris port backend is enabled (disabled by default)
		1895
		1896	F<ev.c> includes the backend files directly when enabled, so you only need
		1897	to compile this single file.
		1898
		1899	=head3 LIBEVENT COMPATIBILITY API
		1900
		1901	To include the libevent compatibility API, also include:
		1902
		1903	#include "event.c"
		1904
		1905	in the file including F<ev.c>, and:
		1906
		1907	#include "event.h"
		1908
		1909	in the files that want to use the libevent API. This also includes F<ev.h>.
		1910
		1911	You need the following additional files for this:
		1912
		1913	event.h
		1914	event.c
		1915
		1916	=head3 AUTOCONF SUPPORT
		1917
		1918	Instead of using C<EV_STANDALONE=1> and providing your config in
		1919	whatever way you want, you can also C<m4_include([libev.m4])> in your
		1920	F<configure.ac> and leave C<EV_STANDALONE> undefined. F<ev.c> will then
		1921	include F<config.h> and configure itself accordingly.
		1922
		1923	For this of course you need the m4 file:
		1924
		1925	libev.m4
		1926
		1927	=head2 PREPROCESSOR SYMBOLS/MACROS
		1928
		1929	Libev can be configured via a variety of preprocessor symbols you have to define
		1930	before including any of its files. The default is not to build for multiplicity
		1931	and only include the select backend.
		1932
		1933	=over 4
		1934
		1935	=item EV_STANDALONE
		1936
		1937	Must always be C<1> if you do not use autoconf configuration, which
		1938	keeps libev from including F<config.h>, and it also defines dummy
		1939	implementations for some libevent functions (such as logging, which is not
		1940	supported). It will also not define any of the structs usually found in
		1941	F<event.h> that are not directly supported by the libev core alone.
		1942
		1943	=item EV_USE_MONOTONIC
		1944
		1945	If defined to be C<1>, libev will try to detect the availability of the
		1946	monotonic clock option at both compiletime and runtime. Otherwise no use
		1947	of the monotonic clock option will be attempted. If you enable this, you
		1948	usually have to link against librt or something similar. Enabling it when
		1949	the functionality isn't available is safe, though, althoguh you have
		1950	to make sure you link against any libraries where the C<clock_gettime>
		1951	function is hiding in (often F<-lrt>).
		1952
		1953	=item EV_USE_REALTIME
		1954
		1955	If defined to be C<1>, libev will try to detect the availability of the
		1956	realtime clock option at compiletime (and assume its availability at
		1957	runtime if successful). Otherwise no use of the realtime clock option will
		1958	be attempted. This effectively replaces C<gettimeofday> by C<clock_get
		1959	(CLOCK_REALTIME, ...)> and will not normally affect correctness. See tzhe note about libraries
		1960	in the description of C<EV_USE_MONOTONIC>, though.
		1961
		1962	=item EV_USE_SELECT
		1963
		1964	If undefined or defined to be C<1>, libev will compile in support for the
		1965	C<select>(2) backend. No attempt at autodetection will be done: if no
		1966	other method takes over, select will be it. Otherwise the select backend
		1967	will not be compiled in.
		1968
		1969	=item EV_SELECT_USE_FD_SET
		1970
		1971	If defined to C<1>, then the select backend will use the system C<fd_set>
		1972	structure. This is useful if libev doesn't compile due to a missing
		1973	C<NFDBITS> or C<fd_mask> definition or it misguesses the bitset layout on
		1974	exotic systems. This usually limits the range of file descriptors to some
		1975	low limit such as 1024 or might have other limitations (winsocket only
		1976	allows 64 sockets). The C<FD_SETSIZE> macro, set before compilation, might
		1977	influence the size of the C<fd_set> used.
		1978
		1979	=item EV_SELECT_IS_WINSOCKET
		1980
		1981	When defined to C<1>, the select backend will assume that
		1982	select/socket/connect etc. don't understand file descriptors but
		1983	wants osf handles on win32 (this is the case when the select to
		1984	be used is the winsock select). This means that it will call
		1985	C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,
		1986	it is assumed that all these functions actually work on fds, even
		1987	on win32. Should not be defined on non-win32 platforms.
		1988
		1989	=item EV_USE_POLL
		1990
		1991	If defined to be C<1>, libev will compile in support for the C<poll>(2)
		1992	backend. Otherwise it will be enabled on non-win32 platforms. It
		1993	takes precedence over select.
		1994
		1995	=item EV_USE_EPOLL
		1996
		1997	If defined to be C<1>, libev will compile in support for the Linux
		1998	C<epoll>(7) backend. Its availability will be detected at runtime,
		1999	otherwise another method will be used as fallback. This is the
		2000	preferred backend for GNU/Linux systems.
		2001
		2002	=item EV_USE_KQUEUE
		2003
		2004	If defined to be C<1>, libev will compile in support for the BSD style
		2005	C<kqueue>(2) backend. Its actual availability will be detected at runtime,
		2006	otherwise another method will be used as fallback. This is the preferred
		2007	backend for BSD and BSD-like systems, although on most BSDs kqueue only
		2008	supports some types of fds correctly (the only platform we found that
		2009	supports ptys for example was NetBSD), so kqueue might be compiled in, but
		2010	not be used unless explicitly requested. The best way to use it is to find
		2011	out whether kqueue supports your type of fd properly and use an embedded
		2012	kqueue loop.
		2013
		2014	=item EV_USE_PORT
		2015
		2016	If defined to be C<1>, libev will compile in support for the Solaris
		2017	10 port style backend. Its availability will be detected at runtime,
		2018	otherwise another method will be used as fallback. This is the preferred
		2019	backend for Solaris 10 systems.
		2020
		2021	=item EV_USE_DEVPOLL
		2022
		2023	reserved for future expansion, works like the USE symbols above.
		2024
		2025	=item EV_USE_INOTIFY
		2026
		2027	If defined to be C<1>, libev will compile in support for the Linux inotify
		2028	interface to speed up C<ev_stat> watchers. Its actual availability will
		2029	be detected at runtime.
		2030
		2031	=item EV_H
		2032
		2033	The name of the F<ev.h> header file used to include it. The default if
		2034	undefined is C<< <ev.h> >> in F<event.h> and C<"ev.h"> in F<ev.c>. This
		2035	can be used to virtually rename the F<ev.h> header file in case of conflicts.
		2036
		2037	=item EV_CONFIG_H
		2038
		2039	If C<EV_STANDALONE> isn't C<1>, this variable can be used to override
		2040	F<ev.c>'s idea of where to find the F<config.h> file, similarly to
		2041	C<EV_H>, above.
		2042
		2043	=item EV_EVENT_H
		2044
		2045	Similarly to C<EV_H>, this macro can be used to override F<event.c>'s idea
		2046	of how the F<event.h> header can be found.
		2047
		2048	=item EV_PROTOTYPES
		2049
		2050	If defined to be C<0>, then F<ev.h> will not define any function
		2051	prototypes, but still define all the structs and other symbols. This is
		2052	occasionally useful if you want to provide your own wrapper functions
		2053	around libev functions.
		2054
		2055	=item EV_MULTIPLICITY
		2056
		2057	If undefined or defined to C<1>, then all event-loop-specific functions
		2058	will have the C<struct ev_loop *> as first argument, and you can create
		2059	additional independent event loops. Otherwise there will be no support
		2060	for multiple event loops and there is no first event loop pointer
		2061	argument. Instead, all functions act on the single default loop.
		2062
		2063	=item EV_PERIODIC_ENABLE
		2064
		2065	If undefined or defined to be C<1>, then periodic timers are supported. If
		2066	defined to be C<0>, then they are not. Disabling them saves a few kB of
		2067	code.
		2068
		2069	=item EV_EMBED_ENABLE
		2070
		2071	If undefined or defined to be C<1>, then embed watchers are supported. If
		2072	defined to be C<0>, then they are not.
		2073
		2074	=item EV_STAT_ENABLE
		2075
		2076	If undefined or defined to be C<1>, then stat watchers are supported. If
		2077	defined to be C<0>, then they are not.
		2078
		2079	=item EV_FORK_ENABLE
		2080
		2081	If undefined or defined to be C<1>, then fork watchers are supported. If
		2082	defined to be C<0>, then they are not.
		2083
		2084	=item EV_MINIMAL
		2085
		2086	If you need to shave off some kilobytes of code at the expense of some
		2087	speed, define this symbol to C<1>. Currently only used for gcc to override
		2088	some inlining decisions, saves roughly 30% codesize of amd64.
		2089
		2090	=item EV_PID_HASHSIZE
		2091
		2092	C<ev_child> watchers use a small hash table to distribute workload by
		2093	pid. The default size is C<16> (or C<1> with C<EV_MINIMAL>), usually more
		2094	than enough. If you need to manage thousands of children you might want to
		2095	increase this value (I<must> be a power of two).
		2096
		2097	=item EV_INOTIFY_HASHSIZE
		2098
		2099	C<ev_staz> watchers use a small hash table to distribute workload by
		2100	inotify watch id. The default size is C<16> (or C<1> with C<EV_MINIMAL>),
		2101	usually more than enough. If you need to manage thousands of C<ev_stat>
		2102	watchers you might want to increase this value (I<must> be a power of
		2103	two).
		2104
		2105	=item EV_COMMON
		2106
		2107	By default, all watchers have a C<void *data> member. By redefining
		2108	this macro to a something else you can include more and other types of
		2109	members. You have to define it each time you include one of the files,
		2110	though, and it must be identical each time.
		2111
		2112	For example, the perl EV module uses something like this:
		2113
		2114	#define EV_COMMON \
		2115	SV *self; /* contains this struct */ \
		2116	SV *cb_sv, *fh /* note no trailing ";" */
		2117
		2118	=item EV_CB_DECLARE (type)
		2119
		2120	=item EV_CB_INVOKE (watcher, revents)
		2121
		2122	=item ev_set_cb (ev, cb)
		2123
		2124	Can be used to change the callback member declaration in each watcher,
		2125	and the way callbacks are invoked and set. Must expand to a struct member
		2126	definition and a statement, respectively. See the F<ev.v> header file for
		2127	their default definitions. One possible use for overriding these is to
		2128	avoid the C<struct ev_loop *> as first argument in all cases, or to use
		2129	method calls instead of plain function calls in C++.
		2130
		2131	=head2 EXAMPLES
		2132
		2133	For a real-world example of a program the includes libev
		2134	verbatim, you can have a look at the EV perl module
		2135	(L<http://software.schmorp.de/pkg/EV.html>). It has the libev files in
		2136	the F<libev/> subdirectory and includes them in the F<EV/EVAPI.h> (public
		2137	interface) and F<EV.xs> (implementation) files. Only the F<EV.xs> file
		2138	will be compiled. It is pretty complex because it provides its own header
		2139	file.
		2140
		2141	The usage in rxvt-unicode is simpler. It has a F<ev_cpp.h> header file
		2142	that everybody includes and which overrides some autoconf choices:
		2143
		2144	#define EV_USE_POLL 0
		2145	#define EV_MULTIPLICITY 0
		2146	#define EV_PERIODICS 0
		2147	#define EV_CONFIG_H <config.h>
		2148
		2149	#include "ev++.h"
		2150
		2151	And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:
		2152
		2153	#include "ev_cpp.h"
		2154	#include "ev.c"
		2155
		2156
		2157	=head1 COMPLEXITIES
		2158
		2159	In this section the complexities of (many of) the algorithms used inside
		2160	libev will be explained. For complexity discussions about backends see the
		2161	documentation for C<ev_default_init>.
		2162
		2163	=over 4
		2164
		2165	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)
		2166
		2167	=item Changing timer/periodic watchers (by autorepeat, again): O(log skipped_other_timers)
		2168
		2169	=item Starting io/check/prepare/idle/signal/child watchers: O(1)
		2170
		2171	=item Stopping check/prepare/idle watchers: O(1)
		2172
		2173	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))
		2174
		2175	=item Finding the next timer per loop iteration: O(1)
		2176
		2177	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)
		2178
		2179	=item Activating one watcher: O(1)
		2180
		2181	=back
		2182
878		2183
879	=head1 AUTHOR	2184	=head1 AUTHOR
880		2185
881	Marc Lehmann <libev@schmorp.de>.	2186	Marc Lehmann <libev@schmorp.de>.
882		2187

Diff Legend

–	Removed lines
+	Added lines
<	Changed lines
>	Changed lines