ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/libev/ev.pod

Comparing libev/ev.pod (file contents):
Revision 1.32 by root, Fri Nov 23 08:36:35 2007 UTC vs.
Revision 1.76 by root, Sat Dec 8 15:30:30 2007 UTC

…		…
4		4
5	=head1 SYNOPSIS	5	=head1 SYNOPSIS
6		6
7	#include <ev.h>	7	#include <ev.h>
8		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	}
		50
9	=head1 DESCRIPTION	51	=head1 DESCRIPTION
		52
		53	The newest version of this document is also available as a html-formatted
		54	web page you might find easier to navigate when reading it for the first
		55	time: L<http://cvs.schmorp.de/libev/ev.html>.
10		56
11	Libev is an event loop: you register interest in certain events (such as a	57	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	58	file descriptor being readable or a timeout occuring), and it will manage
13	these event sources and provide your program with events.	59	these event sources and provide your program with events.
14		60
…		…
21	details of the event, and then hand it over to libev by I<starting> the	67	details of the event, and then hand it over to libev by I<starting> the
22	watcher.	68	watcher.
23		69
24	=head1 FEATURES	70	=head1 FEATURES
25		71
26	Libev supports select, poll, the linux-specific epoll and the bsd-specific	72	Libev supports C<select>, C<poll>, the Linux-specific C<epoll>, the
27	kqueue mechanisms for file descriptor events, relative timers, absolute	73	BSD-specific C<kqueue> and the Solaris-specific event port mechanisms
28	timers with customised rescheduling, signal events, process status change	74	for file descriptor events (C<ev_io>), the Linux C<inotify> interface
29	events (related to SIGCHLD), and event watchers dealing with the event	75	(for C<ev_stat>), relative timers (C<ev_timer>), absolute timers
30	loop mechanism itself (idle, prepare and check watchers). It also is quite	76	with customised rescheduling (C<ev_periodic>), synchronous signals
		77	(C<ev_signal>), process status change events (C<ev_child>), and event
		78	watchers dealing with the event loop mechanism itself (C<ev_idle>,
		79	C<ev_embed>, C<ev_prepare> and C<ev_check> watchers) as well as
		80	file watchers (C<ev_stat>) and even limited support for fork events
		81	(C<ev_fork>).
		82
		83	It also is quite fast (see this
31	fast (see this L<benchmark|http://libev.schmorp.de/bench.html> comparing	84	L<benchmark|http://libev.schmorp.de/bench.html> comparing it to libevent
32	it to libevent for example).	85	for example).
33		86
34	=head1 CONVENTIONS	87	=head1 CONVENTIONS
35		88
36	Libev is very configurable. In this manual the default configuration	89	Libev is very configurable. In this manual the default configuration will
37	will be described, which supports multiple event loops. For more info	90	be described, which supports multiple event loops. For more info about
38	about various configuration options please have a look at the file	91	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	92	this manual. If libev was configured without support for multiple event
40	support for multiple event loops, then all functions taking an initial	93	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 *>)	94	(which is always of type C<struct ev_loop *>) will not have this argument.
42	will not have this argument.
43		95
44	=head1 TIME REPRESENTATION	96	=head1 TIME REPRESENTATION
45		97
46	Libev represents time as a single floating point number, representing the	98	Libev represents time as a single floating point number, representing the
47	(fractional) number of seconds since the (POSIX) epoch (somewhere near	99	(fractional) number of seconds since the (POSIX) epoch (somewhere near
48	the beginning of 1970, details are complicated, don't ask). This type is	100	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	101	called C<ev_tstamp>, which is what you should use too. It usually aliases
50	to the double type in C.	102	to the C<double> type in C, and when you need to do any calculations on
		103	it, you should treat it as such.
51		104
52	=head1 GLOBAL FUNCTIONS	105	=head1 GLOBAL FUNCTIONS
53		106
54	These functions can be called anytime, even before initialising the	107	These functions can be called anytime, even before initialising the
55	library in any way.	108	library in any way.
…		…
75	Usually, it's a good idea to terminate if the major versions mismatch,	128	Usually, it's a good idea to terminate if the major versions mismatch,
76	as this indicates an incompatible change. Minor versions are usually	129	as this indicates an incompatible change. Minor versions are usually
77	compatible to older versions, so a larger minor version alone is usually	130	compatible to older versions, so a larger minor version alone is usually
78	not a problem.	131	not a problem.
79		132
		133	Example: Make sure we haven't accidentally been linked against the wrong
		134	version.
		135
		136	assert (("libev version mismatch",
		137	ev_version_major () == EV_VERSION_MAJOR
		138	&& ev_version_minor () >= EV_VERSION_MINOR));
		139
80	=item unsigned int ev_supported_backends ()	140	=item unsigned int ev_supported_backends ()
81		141
82	Return the set of all backends (i.e. their corresponding C<EV_BACKEND_*>	142	Return the set of all backends (i.e. their corresponding C<EV_BACKEND_*>
83	value) compiled into this binary of libev (independent of their	143	value) compiled into this binary of libev (independent of their
84	availability on the system you are running on). See C<ev_default_loop> for	144	availability on the system you are running on). See C<ev_default_loop> for
85	a description of the set values.	145	a description of the set values.
		146
		147	Example: make sure we have the epoll method, because yeah this is cool and
		148	a must have and can we have a torrent of it please!!!11
		149
		150	assert (("sorry, no epoll, no sex",
		151	ev_supported_backends () & EVBACKEND_EPOLL));
86		152
87	=item unsigned int ev_recommended_backends ()	153	=item unsigned int ev_recommended_backends ()
88		154
89	Return the set of all backends compiled into this binary of libev and also	155	Return the set of all backends compiled into this binary of libev and also
90	recommended for this platform. This set is often smaller than the one	156	recommended for this platform. This set is often smaller than the one
91	returned by C<ev_supported_backends>, as for example kqueue is broken on	157	returned by C<ev_supported_backends>, as for example kqueue is broken on
92	most BSDs and will not be autodetected unless you explicitly request it	158	most BSDs and will not be autodetected unless you explicitly request it
93	(assuming you know what you are doing). This is the set of backends that	159	(assuming you know what you are doing). This is the set of backends that
94	C<EVFLAG_AUTO> will probe for.	160	libev will probe for if you specify no backends explicitly.
		161
		162	=item unsigned int ev_embeddable_backends ()
		163
		164	Returns the set of backends that are embeddable in other event loops. This
		165	is the theoretical, all-platform, value. To find which backends
		166	might be supported on the current system, you would need to look at
		167	C<ev_embeddable_backends () & ev_supported_backends ()>, likewise for
		168	recommended ones.
		169
		170	See the description of C<ev_embed> watchers for more info.
95		171
96	=item ev_set_allocator (void *(*cb)(void *ptr, long size))	172	=item ev_set_allocator (void *(*cb)(void *ptr, long size))
97		173
98	Sets the allocation function to use (the prototype is similar to the	174	Sets the allocation function to use (the prototype is similar - the
99	realloc C function, the semantics are identical). It is used to allocate	175	semantics is identical - to the realloc C function). It is used to
100	and free memory (no surprises here). If it returns zero when memory	176	allocate and free memory (no surprises here). If it returns zero when
101	needs to be allocated, the library might abort or take some potentially	177	memory needs to be allocated, the library might abort or take some
102	destructive action. The default is your system realloc function.	178	potentially destructive action. The default is your system realloc
		179	function.
103		180
104	You could override this function in high-availability programs to, say,	181	You could override this function in high-availability programs to, say,
105	free some memory if it cannot allocate memory, to use a special allocator,	182	free some memory if it cannot allocate memory, to use a special allocator,
106	or even to sleep a while and retry until some memory is available.	183	or even to sleep a while and retry until some memory is available.
		184
		185	Example: Replace the libev allocator with one that waits a bit and then
		186	retries).
		187
		188	static void *
		189	persistent_realloc (void *ptr, size_t size)
		190	{
		191	for (;;)
		192	{
		193	void *newptr = realloc (ptr, size);
		194
		195	if (newptr)
		196	return newptr;
		197
		198	sleep (60);
		199	}
		200	}
		201
		202	...
		203	ev_set_allocator (persistent_realloc);
107		204
108	=item ev_set_syserr_cb (void (*cb)(const char *msg));	205	=item ev_set_syserr_cb (void (*cb)(const char *msg));
109		206
110	Set the callback function to call on a retryable syscall error (such	207	Set the callback function to call on a retryable syscall error (such
111	as failed select, poll, epoll_wait). The message is a printable string	208	as failed select, poll, epoll_wait). The message is a printable string
…		…
113	callback is set, then libev will expect it to remedy the sitution, no	210	callback is set, then libev will expect it to remedy the sitution, no
114	matter what, when it returns. That is, libev will generally retry the	211	matter what, when it returns. That is, libev will generally retry the
115	requested operation, or, if the condition doesn't go away, do bad stuff	212	requested operation, or, if the condition doesn't go away, do bad stuff
116	(such as abort).	213	(such as abort).
117		214
		215	Example: This is basically the same thing that libev does internally, too.
		216
		217	static void
		218	fatal_error (const char *msg)
		219	{
		220	perror (msg);
		221	abort ();
		222	}
		223
		224	...
		225	ev_set_syserr_cb (fatal_error);
		226
118	=back	227	=back
119		228
120	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP	229	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP
121		230
122	An event loop is described by a C<struct ev_loop *>. The library knows two	231	An event loop is described by a C<struct ev_loop *>. The library knows two
…		…
141		250
142	If you don't know what event loop to use, use the one returned from this	251	If you don't know what event loop to use, use the one returned from this
143	function.	252	function.
144		253
145	The flags argument can be used to specify special behaviour or specific	254	The flags argument can be used to specify special behaviour or specific
146	backends to use, and is usually specified as C<0> (or EVFLAG_AUTO).	255	backends to use, and is usually specified as C<0> (or C<EVFLAG_AUTO>).
147		256
148	It supports the following flags:	257	The following flags are supported:
149		258
150	=over 4	259	=over 4
151		260
152	=item C<EVFLAG_AUTO>	261	=item C<EVFLAG_AUTO>
153		262
…		…
161	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will	270	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will
162	override the flags completely if it is found in the environment. This is	271	override the flags completely if it is found in the environment. This is
163	useful to try out specific backends to test their performance, or to work	272	useful to try out specific backends to test their performance, or to work
164	around bugs.	273	around bugs.
165		274
		275	=item C<EVFLAG_FORKCHECK>
		276
		277	Instead of calling C<ev_default_fork> or C<ev_loop_fork> manually after
		278	a fork, you can also make libev check for a fork in each iteration by
		279	enabling this flag.
		280
		281	This works by calling C<getpid ()> on every iteration of the loop,
		282	and thus this might slow down your event loop if you do a lot of loop
		283	iterations and little real work, but is usually not noticeable (on my
		284	Linux system for example, C<getpid> is actually a simple 5-insn sequence
		285	without a syscall and thus I<very> fast, but my Linux system also has
		286	C<pthread_atfork> which is even faster).
		287
		288	The big advantage of this flag is that you can forget about fork (and
		289	forget about forgetting to tell libev about forking) when you use this
		290	flag.
		291
		292	This flag setting cannot be overriden or specified in the C<LIBEV_FLAGS>
		293	environment variable.
		294
166	=item C<EVBACKEND_SELECT> (value 1, portable select backend)	295	=item C<EVBACKEND_SELECT> (value 1, portable select backend)
167		296
168	This is your standard select(2) backend. Not I<completely> standard, as	297	This is your standard select(2) backend. Not I<completely> standard, as
169	libev tries to roll its own fd_set with no limits on the number of fds,	298	libev tries to roll its own fd_set with no limits on the number of fds,
170	but if that fails, expect a fairly low limit on the number of fds when	299	but if that fails, expect a fairly low limit on the number of fds when
…		…
198	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)	327	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)
199		328
200	Kqueue deserves special mention, as at the time of this writing, it	329	Kqueue deserves special mention, as at the time of this writing, it
201	was broken on all BSDs except NetBSD (usually it doesn't work with	330	was broken on all BSDs except NetBSD (usually it doesn't work with
202	anything but sockets and pipes, except on Darwin, where of course its	331	anything but sockets and pipes, except on Darwin, where of course its
203	completely useless). For this reason its not being "autodetected" unless	332	completely useless). For this reason its not being "autodetected"
204	you explicitly specify the flags (i.e. you don't use EVFLAG_AUTO).	333	unless you explicitly specify it explicitly in the flags (i.e. using
		334	C<EVBACKEND_KQUEUE>).
205		335
206	It scales in the same way as the epoll backend, but the interface to the	336	It scales in the same way as the epoll backend, but the interface to the
207	kernel is more efficient (which says nothing about its actual speed, of	337	kernel is more efficient (which says nothing about its actual speed, of
208	course). While starting and stopping an I/O watcher does not cause an	338	course). While starting and stopping an I/O watcher does not cause an
209	extra syscall as with epoll, it still adds up to four event changes per	339	extra syscall as with epoll, it still adds up to four event changes per
…		…
233	If one or more of these are ored into the flags value, then only these	363	If one or more of these are ored into the flags value, then only these
234	backends will be tried (in the reverse order as given here). If none are	364	backends will be tried (in the reverse order as given here). If none are
235	specified, most compiled-in backend will be tried, usually in reverse	365	specified, most compiled-in backend will be tried, usually in reverse
236	order of their flag values :)	366	order of their flag values :)
237		367
		368	The most typical usage is like this:
		369
		370	if (!ev_default_loop (0))
		371	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
		372
		373	Restrict libev to the select and poll backends, and do not allow
		374	environment settings to be taken into account:
		375
		376	ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);
		377
		378	Use whatever libev has to offer, but make sure that kqueue is used if
		379	available (warning, breaks stuff, best use only with your own private
		380	event loop and only if you know the OS supports your types of fds):
		381
		382	ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);
		383
238	=item struct ev_loop *ev_loop_new (unsigned int flags)	384	=item struct ev_loop *ev_loop_new (unsigned int flags)
239		385
240	Similar to C<ev_default_loop>, but always creates a new event loop that is	386	Similar to C<ev_default_loop>, but always creates a new event loop that is
241	always distinct from the default loop. Unlike the default loop, it cannot	387	always distinct from the default loop. Unlike the default loop, it cannot
242	handle signal and child watchers, and attempts to do so will be greeted by	388	handle signal and child watchers, and attempts to do so will be greeted by
243	undefined behaviour (or a failed assertion if assertions are enabled).	389	undefined behaviour (or a failed assertion if assertions are enabled).
244		390
		391	Example: Try to create a event loop that uses epoll and nothing else.
		392
		393	struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);
		394	if (!epoller)
		395	fatal ("no epoll found here, maybe it hides under your chair");
		396
245	=item ev_default_destroy ()	397	=item ev_default_destroy ()
246		398
247	Destroys the default loop again (frees all memory and kernel state	399	Destroys the default loop again (frees all memory and kernel state
248	etc.). This stops all registered event watchers (by not touching them in	400	etc.). None of the active event watchers will be stopped in the normal
249	any way whatsoever, although you cannot rely on this :).	401	sense, so e.g. C<ev_is_active> might still return true. It is your
		402	responsibility to either stop all watchers cleanly yoursef I<before>
		403	calling this function, or cope with the fact afterwards (which is usually
		404	the easiest thing, youc na just ignore the watchers and/or C<free ()> them
		405	for example).
250		406
251	=item ev_loop_destroy (loop)	407	=item ev_loop_destroy (loop)
252		408
253	Like C<ev_default_destroy>, but destroys an event loop created by an	409	Like C<ev_default_destroy>, but destroys an event loop created by an
254	earlier call to C<ev_loop_new>.	410	earlier call to C<ev_loop_new>.
…		…
278		434
279	Like C<ev_default_fork>, but acts on an event loop created by	435	Like C<ev_default_fork>, but acts on an event loop created by
280	C<ev_loop_new>. Yes, you have to call this on every allocated event loop	436	C<ev_loop_new>. Yes, you have to call this on every allocated event loop
281	after fork, and how you do this is entirely your own problem.	437	after fork, and how you do this is entirely your own problem.
282		438
		439	=item unsigned int ev_loop_count (loop)
		440
		441	Returns the count of loop iterations for the loop, which is identical to
		442	the number of times libev did poll for new events. It starts at C<0> and
		443	happily wraps around with enough iterations.
		444
		445	This value can sometimes be useful as a generation counter of sorts (it
		446	"ticks" the number of loop iterations), as it roughly corresponds with
		447	C<ev_prepare> and C<ev_check> calls.
		448
283	=item unsigned int ev_backend (loop)	449	=item unsigned int ev_backend (loop)
284		450
285	Returns one of the C<EVBACKEND_*> flags indicating the event backend in	451	Returns one of the C<EVBACKEND_*> flags indicating the event backend in
286	use.	452	use.
287		453
288	=item ev_tstamp ev_now (loop)	454	=item ev_tstamp ev_now (loop)
289		455
290	Returns the current "event loop time", which is the time the event loop	456	Returns the current "event loop time", which is the time the event loop
291	got events and started processing them. This timestamp does not change	457	received events and started processing them. This timestamp does not
292	as long as callbacks are being processed, and this is also the base time	458	change as long as callbacks are being processed, and this is also the base
293	used for relative timers. You can treat it as the timestamp of the event	459	time used for relative timers. You can treat it as the timestamp of the
294	occuring (or more correctly, the mainloop finding out about it).	460	event occuring (or more correctly, libev finding out about it).
295		461
296	=item ev_loop (loop, int flags)	462	=item ev_loop (loop, int flags)
297		463
298	Finally, this is it, the event handler. This function usually is called	464	Finally, this is it, the event handler. This function usually is called
299	after you initialised all your watchers and you want to start handling	465	after you initialised all your watchers and you want to start handling
300	events.	466	events.
301		467
302	If the flags argument is specified as 0, it will not return until either	468	If the flags argument is specified as C<0>, it will not return until
303	no event watchers are active anymore or C<ev_unloop> was called.	469	either no event watchers are active anymore or C<ev_unloop> was called.
		470
		471	Please note that an explicit C<ev_unloop> is usually better than
		472	relying on all watchers to be stopped when deciding when a program has
		473	finished (especially in interactive programs), but having a program that
		474	automatically loops as long as it has to and no longer by virtue of
		475	relying on its watchers stopping correctly is a thing of beauty.
304		476
305	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle	477	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle
306	those events and any outstanding ones, but will not block your process in	478	those events and any outstanding ones, but will not block your process in
307	case there are no events and will return after one iteration of the loop.	479	case there are no events and will return after one iteration of the loop.
308		480
309	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if	481	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if
310	neccessary) and will handle those and any outstanding ones. It will block	482	neccessary) and will handle those and any outstanding ones. It will block
311	your process until at least one new event arrives, and will return after	483	your process until at least one new event arrives, and will return after
312	one iteration of the loop.	484	one iteration of the loop. This is useful if you are waiting for some
		485	external event in conjunction with something not expressible using other
		486	libev watchers. However, a pair of C<ev_prepare>/C<ev_check> watchers is
		487	usually a better approach for this kind of thing.
313		488
314	This flags value could be used to implement alternative looping
315	constructs, but the C<prepare> and C<check> watchers provide a better and
316	more generic mechanism.
317
318	Here are the gory details of what ev_loop does:	489	Here are the gory details of what C<ev_loop> does:
319		490
320	1. If there are no active watchers (reference count is zero), return.	491	* If there are no active watchers (reference count is zero), return.
321	2. Queue and immediately call all prepare watchers.	492	- Queue prepare watchers and then call all outstanding watchers.
322	3. If we have been forked, recreate the kernel state.	493	- If we have been forked, recreate the kernel state.
323	4. Update the kernel state with all outstanding changes.	494	- Update the kernel state with all outstanding changes.
324	5. Update the "event loop time".	495	- Update the "event loop time".
325	6. Calculate for how long to block.	496	- Calculate for how long to block.
326	7. Block the process, waiting for events.	497	- Block the process, waiting for any events.
		498	- Queue all outstanding I/O (fd) events.
327	8. Update the "event loop time" and do time jump handling.	499	- Update the "event loop time" and do time jump handling.
328	9. Queue all outstanding timers.	500	- Queue all outstanding timers.
329	10. Queue all outstanding periodics.	501	- Queue all outstanding periodics.
330	11. If no events are pending now, queue all idle watchers.	502	- If no events are pending now, queue all idle watchers.
331	12. Queue all check watchers.	503	- Queue all check watchers.
332	13. Call all queued watchers in reverse order (i.e. check watchers first).	504	- Call all queued watchers in reverse order (i.e. check watchers first).
		505	Signals and child watchers are implemented as I/O watchers, and will
		506	be handled here by queueing them when their watcher gets executed.
333	14. If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK	507	- If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK
334	was used, return, otherwise continue with step #1.	508	were used, return, otherwise continue with step *.
		509
		510	Example: Queue some jobs and then loop until no events are outsanding
		511	anymore.
		512
		513	... queue jobs here, make sure they register event watchers as long
		514	... as they still have work to do (even an idle watcher will do..)
		515	ev_loop (my_loop, 0);
		516	... jobs done. yeah!
335		517
336	=item ev_unloop (loop, how)	518	=item ev_unloop (loop, how)
337		519
338	Can be used to make a call to C<ev_loop> return early (but only after it	520	Can be used to make a call to C<ev_loop> return early (but only after it
339	has processed all outstanding events). The C<how> argument must be either	521	has processed all outstanding events). The C<how> argument must be either
…		…
353	visible to the libev user and should not keep C<ev_loop> from exiting if	535	visible to the libev user and should not keep C<ev_loop> from exiting if
354	no event watchers registered by it are active. It is also an excellent	536	no event watchers registered by it are active. It is also an excellent
355	way to do this for generic recurring timers or from within third-party	537	way to do this for generic recurring timers or from within third-party
356	libraries. Just remember to I<unref after start> and I<ref before stop>.	538	libraries. Just remember to I<unref after start> and I<ref before stop>.
357		539
		540	Example: Create a signal watcher, but keep it from keeping C<ev_loop>
		541	running when nothing else is active.
		542
		543	struct ev_signal exitsig;
		544	ev_signal_init (&exitsig, sig_cb, SIGINT);
		545	ev_signal_start (loop, &exitsig);
		546	evf_unref (loop);
		547
		548	Example: For some weird reason, unregister the above signal handler again.
		549
		550	ev_ref (loop);
		551	ev_signal_stop (loop, &exitsig);
		552
358	=back	553	=back
		554
359		555
360	=head1 ANATOMY OF A WATCHER	556	=head1 ANATOMY OF A WATCHER
361		557
362	A watcher is a structure that you create and register to record your	558	A watcher is a structure that you create and register to record your
363	interest in some event. For instance, if you want to wait for STDIN to	559	interest in some event. For instance, if you want to wait for STDIN to
…		…
396	*) >>), and you can stop watching for events at any time by calling the	592	*) >>), and you can stop watching for events at any time by calling the
397	corresponding stop function (C<< ev_<type>_stop (loop, watcher *) >>.	593	corresponding stop function (C<< ev_<type>_stop (loop, watcher *) >>.
398		594
399	As long as your watcher is active (has been started but not stopped) you	595	As long as your watcher is active (has been started but not stopped) you
400	must not touch the values stored in it. Most specifically you must never	596	must not touch the values stored in it. Most specifically you must never
401	reinitialise it or call its set macro.	597	reinitialise it or call its C<set> macro.
402
403	You can check whether an event is active by calling the C<ev_is_active
404	(watcher *)> macro. To see whether an event is outstanding (but the
405	callback for it has not been called yet) you can use the C<ev_is_pending
406	(watcher *)> macro.
407		598
408	Each and every callback receives the event loop pointer as first, the	599	Each and every callback receives the event loop pointer as first, the
409	registered watcher structure as second, and a bitset of received events as	600	registered watcher structure as second, and a bitset of received events as
410	third argument.	601	third argument.
411		602
…		…
435	The signal specified in the C<ev_signal> watcher has been received by a thread.	626	The signal specified in the C<ev_signal> watcher has been received by a thread.
436		627
437	=item C<EV_CHILD>	628	=item C<EV_CHILD>
438		629
439	The pid specified in the C<ev_child> watcher has received a status change.	630	The pid specified in the C<ev_child> watcher has received a status change.
		631
		632	=item C<EV_STAT>
		633
		634	The path specified in the C<ev_stat> watcher changed its attributes somehow.
440		635
441	=item C<EV_IDLE>	636	=item C<EV_IDLE>
442		637
443	The C<ev_idle> watcher has determined that you have nothing better to do.	638	The C<ev_idle> watcher has determined that you have nothing better to do.
444		639
…		…
452	received events. Callbacks of both watcher types can start and stop as	647	received events. Callbacks of both watcher types can start and stop as
453	many watchers as they want, and all of them will be taken into account	648	many watchers as they want, and all of them will be taken into account
454	(for example, a C<ev_prepare> watcher might start an idle watcher to keep	649	(for example, a C<ev_prepare> watcher might start an idle watcher to keep
455	C<ev_loop> from blocking).	650	C<ev_loop> from blocking).
456		651
		652	=item C<EV_EMBED>
		653
		654	The embedded event loop specified in the C<ev_embed> watcher needs attention.
		655
		656	=item C<EV_FORK>
		657
		658	The event loop has been resumed in the child process after fork (see
		659	C<ev_fork>).
		660
457	=item C<EV_ERROR>	661	=item C<EV_ERROR>
458		662
459	An unspecified error has occured, the watcher has been stopped. This might	663	An unspecified error has occured, the watcher has been stopped. This might
460	happen because the watcher could not be properly started because libev	664	happen because the watcher could not be properly started because libev
461	ran out of memory, a file descriptor was found to be closed or any other	665	ran out of memory, a file descriptor was found to be closed or any other
…		…
467	your callbacks is well-written it can just attempt the operation and cope	671	your callbacks is well-written it can just attempt the operation and cope
468	with the error from read() or write(). This will not work in multithreaded	672	with the error from read() or write(). This will not work in multithreaded
469	programs, though, so beware.	673	programs, though, so beware.
470		674
471	=back	675	=back
		676
		677	=head2 GENERIC WATCHER FUNCTIONS
		678
		679	In the following description, C<TYPE> stands for the watcher type,
		680	e.g. C<timer> for C<ev_timer> watchers and C<io> for C<ev_io> watchers.
		681
		682	=over 4
		683
		684	=item C<ev_init> (ev_TYPE *watcher, callback)
		685
		686	This macro initialises the generic portion of a watcher. The contents
		687	of the watcher object can be arbitrary (so C<malloc> will do). Only
		688	the generic parts of the watcher are initialised, you I<need> to call
		689	the type-specific C<ev_TYPE_set> macro afterwards to initialise the
		690	type-specific parts. For each type there is also a C<ev_TYPE_init> macro
		691	which rolls both calls into one.
		692
		693	You can reinitialise a watcher at any time as long as it has been stopped
		694	(or never started) and there are no pending events outstanding.
		695
		696	The callback is always of type C<void (*)(ev_loop *loop, ev_TYPE *watcher,
		697	int revents)>.
		698
		699	=item C<ev_TYPE_set> (ev_TYPE *, [args])
		700
		701	This macro initialises the type-specific parts of a watcher. You need to
		702	call C<ev_init> at least once before you call this macro, but you can
		703	call C<ev_TYPE_set> any number of times. You must not, however, call this
		704	macro on a watcher that is active (it can be pending, however, which is a
		705	difference to the C<ev_init> macro).
		706
		707	Although some watcher types do not have type-specific arguments
		708	(e.g. C<ev_prepare>) you still need to call its C<set> macro.
		709
		710	=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])
		711
		712	This convinience macro rolls both C<ev_init> and C<ev_TYPE_set> macro
		713	calls into a single call. This is the most convinient method to initialise
		714	a watcher. The same limitations apply, of course.
		715
		716	=item C<ev_TYPE_start> (loop *, ev_TYPE *watcher)
		717
		718	Starts (activates) the given watcher. Only active watchers will receive
		719	events. If the watcher is already active nothing will happen.
		720
		721	=item C<ev_TYPE_stop> (loop *, ev_TYPE *watcher)
		722
		723	Stops the given watcher again (if active) and clears the pending
		724	status. It is possible that stopped watchers are pending (for example,
		725	non-repeating timers are being stopped when they become pending), but
		726	C<ev_TYPE_stop> ensures that the watcher is neither active nor pending. If
		727	you want to free or reuse the memory used by the watcher it is therefore a
		728	good idea to always call its C<ev_TYPE_stop> function.
		729
		730	=item bool ev_is_active (ev_TYPE *watcher)
		731
		732	Returns a true value iff the watcher is active (i.e. it has been started
		733	and not yet been stopped). As long as a watcher is active you must not modify
		734	it.
		735
		736	=item bool ev_is_pending (ev_TYPE *watcher)
		737
		738	Returns a true value iff the watcher is pending, (i.e. it has outstanding
		739	events but its callback has not yet been invoked). As long as a watcher
		740	is pending (but not active) you must not call an init function on it (but
		741	C<ev_TYPE_set> is safe), you must not change its priority, and you must
		742	make sure the watcher is available to libev (e.g. you cannot C<free ()>
		743	it).
		744
		745	=item callback ev_cb (ev_TYPE *watcher)
		746
		747	Returns the callback currently set on the watcher.
		748
		749	=item ev_cb_set (ev_TYPE *watcher, callback)
		750
		751	Change the callback. You can change the callback at virtually any time
		752	(modulo threads).
		753
		754	=item ev_set_priority (ev_TYPE *watcher, priority)
		755
		756	=item int ev_priority (ev_TYPE *watcher)
		757
		758	Set and query the priority of the watcher. The priority is a small
		759	integer between C<EV_MAXPRI> (default: C<2>) and C<EV_MINPRI>
		760	(default: C<-2>). Pending watchers with higher priority will be invoked
		761	before watchers with lower priority, but priority will not keep watchers
		762	from being executed (except for C<ev_idle> watchers).
		763
		764	This means that priorities are I<only> used for ordering callback
		765	invocation after new events have been received. This is useful, for
		766	example, to reduce latency after idling, or more often, to bind two
		767	watchers on the same event and make sure one is called first.
		768
		769	If you need to suppress invocation when higher priority events are pending
		770	you need to look at C<ev_idle> watchers, which provide this functionality.
		771
		772	You I<must not> change the priority of a watcher as long as it is active or
		773	pending.
		774
		775	The default priority used by watchers when no priority has been set is
		776	always C<0>, which is supposed to not be too high and not be too low :).
		777
		778	Setting a priority outside the range of C<EV_MINPRI> to C<EV_MAXPRI> is
		779	fine, as long as you do not mind that the priority value you query might
		780	or might not have been adjusted to be within valid range.
		781
		782	=item ev_invoke (loop, ev_TYPE *watcher, int revents)
		783
		784	Invoke the C<watcher> with the given C<loop> and C<revents>. Neither
		785	C<loop> nor C<revents> need to be valid as long as the watcher callback
		786	can deal with that fact.
		787
		788	=item int ev_clear_pending (loop, ev_TYPE *watcher)
		789
		790	If the watcher is pending, this function returns clears its pending status
		791	and returns its C<revents> bitset (as if its callback was invoked). If the
		792	watcher isn't pending it does nothing and returns C<0>.
		793
		794	=back
		795
472		796
473	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER	797	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER
474		798
475	Each watcher has, by default, a member C<void *data> that you can change	799	Each watcher has, by default, a member C<void *data> that you can change
476	and read at any time, libev will completely ignore it. This can be used	800	and read at any time, libev will completely ignore it. This can be used
…		…
494	{	818	{
495	struct my_io *w = (struct my_io *)w_;	819	struct my_io *w = (struct my_io *)w_;
496	...	820	...
497	}	821	}
498		822
499	More interesting and less C-conformant ways of catsing your callback type	823	More interesting and less C-conformant ways of casting your callback type
500	have been omitted....	824	instead have been omitted.
		825
		826	Another common scenario is having some data structure with multiple
		827	watchers:
		828
		829	struct my_biggy
		830	{
		831	int some_data;
		832	ev_timer t1;
		833	ev_timer t2;
		834	}
		835
		836	In this case getting the pointer to C<my_biggy> is a bit more complicated,
		837	you need to use C<offsetof>:
		838
		839	#include <stddef.h>
		840
		841	static void
		842	t1_cb (EV_P_ struct ev_timer *w, int revents)
		843	{
		844	struct my_biggy big = (struct my_biggy *
		845	(((char *)w) - offsetof (struct my_biggy, t1));
		846	}
		847
		848	static void
		849	t2_cb (EV_P_ struct ev_timer *w, int revents)
		850	{
		851	struct my_biggy big = (struct my_biggy *
		852	(((char *)w) - offsetof (struct my_biggy, t2));
		853	}
501		854
502		855
503	=head1 WATCHER TYPES	856	=head1 WATCHER TYPES
504		857
505	This section describes each watcher in detail, but will not repeat	858	This section describes each watcher in detail, but will not repeat
506	information given in the last section.	859	information given in the last section. Any initialisation/set macros,
		860	functions and members specific to the watcher type are explained.
507		861
		862	Members are additionally marked with either I<[read-only]>, meaning that,
		863	while the watcher is active, you can look at the member and expect some
		864	sensible content, but you must not modify it (you can modify it while the
		865	watcher is stopped to your hearts content), or I<[read-write]>, which
		866	means you can expect it to have some sensible content while the watcher
		867	is active, but you can also modify it. Modifying it may not do something
		868	sensible or take immediate effect (or do anything at all), but libev will
		869	not crash or malfunction in any way.
		870
		871
508	=head2 C<ev_io> - is this file descriptor readable or writable	872	=head2 C<ev_io> - is this file descriptor readable or writable?
509		873
510	I/O watchers check whether a file descriptor is readable or writable	874	I/O watchers check whether a file descriptor is readable or writable
511	in each iteration of the event loop (This behaviour is called	875	in each iteration of the event loop, or, more precisely, when reading
512	level-triggering because you keep receiving events as long as the	876	would not block the process and writing would at least be able to write
513	condition persists. Remember you can stop the watcher if you don't want to	877	some data. This behaviour is called level-triggering because you keep
514	act on the event and neither want to receive future events).	878	receiving events as long as the condition persists. Remember you can stop
		879	the watcher if you don't want to act on the event and neither want to
		880	receive future events.
515		881
516	In general you can register as many read and/or write event watchers per	882	In general you can register as many read and/or write event watchers per
517	fd as you want (as long as you don't confuse yourself). Setting all file	883	fd as you want (as long as you don't confuse yourself). Setting all file
518	descriptors to non-blocking mode is also usually a good idea (but not	884	descriptors to non-blocking mode is also usually a good idea (but not
519	required if you know what you are doing).	885	required if you know what you are doing).
520		886
521	You have to be careful with dup'ed file descriptors, though. Some backends	887	You have to be careful with dup'ed file descriptors, though. Some backends
522	(the linux epoll backend is a notable example) cannot handle dup'ed file	888	(the linux epoll backend is a notable example) cannot handle dup'ed file
523	descriptors correctly if you register interest in two or more fds pointing	889	descriptors correctly if you register interest in two or more fds pointing
524	to the same underlying file/socket etc. description (that is, they share	890	to the same underlying file/socket/etc. description (that is, they share
525	the same underlying "file open").	891	the same underlying "file open").
526		892
527	If you must do this, then force the use of a known-to-be-good backend	893	If you must do this, then force the use of a known-to-be-good backend
528	(at the time of this writing, this includes only C<EVBACKEND_SELECT> and	894	(at the time of this writing, this includes only C<EVBACKEND_SELECT> and
529	C<EVBACKEND_POLL>).	895	C<EVBACKEND_POLL>).
530		896
		897	Another thing you have to watch out for is that it is quite easy to
		898	receive "spurious" readyness notifications, that is your callback might
		899	be called with C<EV_READ> but a subsequent C<read>(2) will actually block
		900	because there is no data. Not only are some backends known to create a
		901	lot of those (for example solaris ports), it is very easy to get into
		902	this situation even with a relatively standard program structure. Thus
		903	it is best to always use non-blocking I/O: An extra C<read>(2) returning
		904	C<EAGAIN> is far preferable to a program hanging until some data arrives.
		905
		906	If you cannot run the fd in non-blocking mode (for example you should not
		907	play around with an Xlib connection), then you have to seperately re-test
		908	whether a file descriptor is really ready with a known-to-be good interface
		909	such as poll (fortunately in our Xlib example, Xlib already does this on
		910	its own, so its quite safe to use).
		911
531	=over 4	912	=over 4
532		913
533	=item ev_io_init (ev_io *, callback, int fd, int events)	914	=item ev_io_init (ev_io *, callback, int fd, int events)
534		915
535	=item ev_io_set (ev_io *, int fd, int events)	916	=item ev_io_set (ev_io *, int fd, int events)
536		917
537	Configures an C<ev_io> watcher. The fd is the file descriptor to rceeive	918	Configures an C<ev_io> watcher. The C<fd> is the file descriptor to
538	events for and events is either C<EV_READ>, C<EV_WRITE> or C<EV_READ |	919	rceeive events for and events is either C<EV_READ>, C<EV_WRITE> or
539	EV_WRITE> to receive the given events.	920	C<EV_READ | EV_WRITE> to receive the given events.
540		921
541	Please note that most of the more scalable backend mechanisms (for example	922	=item int fd [read-only]
542	epoll and solaris ports) can result in spurious readyness notifications	923
543	for file descriptors, so you practically need to use non-blocking I/O (and	924	The file descriptor being watched.
544	treat callback invocation as hint only), or retest separately with a safe	925
545	interface before doing I/O (XLib can do this), or force the use of either	926	=item int events [read-only]
546	C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>, which don't suffer from this	927
547	problem. Also note that it is quite easy to have your callback invoked	928	The events being watched.
548	when the readyness condition is no longer valid even when employing
549	typical ways of handling events, so its a good idea to use non-blocking
550	I/O unconditionally.
551		929
552	=back	930	=back
553		931
		932	Example: Call C<stdin_readable_cb> when STDIN_FILENO has become, well
		933	readable, but only once. Since it is likely line-buffered, you could
		934	attempt to read a whole line in the callback.
		935
		936	static void
		937	stdin_readable_cb (struct ev_loop *loop, struct ev_io *w, int revents)
		938	{
		939	ev_io_stop (loop, w);
		940	.. read from stdin here (or from w->fd) and haqndle any I/O errors
		941	}
		942
		943	...
		944	struct ev_loop *loop = ev_default_init (0);
		945	struct ev_io stdin_readable;
		946	ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
		947	ev_io_start (loop, &stdin_readable);
		948	ev_loop (loop, 0);
		949
		950
554	=head2 C<ev_timer> - relative and optionally recurring timeouts	951	=head2 C<ev_timer> - relative and optionally repeating timeouts
555		952
556	Timer watchers are simple relative timers that generate an event after a	953	Timer watchers are simple relative timers that generate an event after a
557	given time, and optionally repeating in regular intervals after that.	954	given time, and optionally repeating in regular intervals after that.
558		955
559	The timers are based on real time, that is, if you register an event that	956	The timers are based on real time, that is, if you register an event that
…		…
594	=item ev_timer_again (loop)	991	=item ev_timer_again (loop)
595		992
596	This will act as if the timer timed out and restart it again if it is	993	This will act as if the timer timed out and restart it again if it is
597	repeating. The exact semantics are:	994	repeating. The exact semantics are:
598		995
		996	If the timer is pending, its pending status is cleared.
		997
599	If the timer is started but nonrepeating, stop it.	998	If the timer is started but nonrepeating, stop it (as if it timed out).
600		999
601	If the timer is repeating, either start it if necessary (with the repeat	1000	If the timer is repeating, either start it if necessary (with the
602	value), or reset the running timer to the repeat value.	1001	C<repeat> value), or reset the running timer to the C<repeat> value.
603		1002
604	This sounds a bit complicated, but here is a useful and typical	1003	This sounds a bit complicated, but here is a useful and typical
605	example: Imagine you have a tcp connection and you want a so-called idle	1004	example: Imagine you have a tcp connection and you want a so-called idle
606	timeout, that is, you want to be called when there have been, say, 60	1005	timeout, that is, you want to be called when there have been, say, 60
607	seconds of inactivity on the socket. The easiest way to do this is to	1006	seconds of inactivity on the socket. The easiest way to do this is to
608	configure an C<ev_timer> with after=repeat=60 and calling ev_timer_again each	1007	configure an C<ev_timer> with a C<repeat> value of C<60> and then call
609	time you successfully read or write some data. If you go into an idle	1008	C<ev_timer_again> each time you successfully read or write some data. If
610	state where you do not expect data to travel on the socket, you can stop	1009	you go into an idle state where you do not expect data to travel on the
611	the timer, and again will automatically restart it if need be.	1010	socket, you can C<ev_timer_stop> the timer, and C<ev_timer_again> will
		1011	automatically restart it if need be.
		1012
		1013	That means you can ignore the C<after> value and C<ev_timer_start>
		1014	altogether and only ever use the C<repeat> value and C<ev_timer_again>:
		1015
		1016	ev_timer_init (timer, callback, 0., 5.);
		1017	ev_timer_again (loop, timer);
		1018	...
		1019	timer->again = 17.;
		1020	ev_timer_again (loop, timer);
		1021	...
		1022	timer->again = 10.;
		1023	ev_timer_again (loop, timer);
		1024
		1025	This is more slightly efficient then stopping/starting the timer each time
		1026	you want to modify its timeout value.
		1027
		1028	=item ev_tstamp repeat [read-write]
		1029
		1030	The current C<repeat> value. Will be used each time the watcher times out
		1031	or C<ev_timer_again> is called and determines the next timeout (if any),
		1032	which is also when any modifications are taken into account.
612		1033
613	=back	1034	=back
614		1035
		1036	Example: Create a timer that fires after 60 seconds.
		1037
		1038	static void
		1039	one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
		1040	{
		1041	.. one minute over, w is actually stopped right here
		1042	}
		1043
		1044	struct ev_timer mytimer;
		1045	ev_timer_init (&mytimer, one_minute_cb, 60., 0.);
		1046	ev_timer_start (loop, &mytimer);
		1047
		1048	Example: Create a timeout timer that times out after 10 seconds of
		1049	inactivity.
		1050
		1051	static void
		1052	timeout_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
		1053	{
		1054	.. ten seconds without any activity
		1055	}
		1056
		1057	struct ev_timer mytimer;
		1058	ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */
		1059	ev_timer_again (&mytimer); /* start timer */
		1060	ev_loop (loop, 0);
		1061
		1062	// and in some piece of code that gets executed on any "activity":
		1063	// reset the timeout to start ticking again at 10 seconds
		1064	ev_timer_again (&mytimer);
		1065
		1066
615	=head2 C<ev_periodic> - to cron or not to cron	1067	=head2 C<ev_periodic> - to cron or not to cron?
616		1068
617	Periodic watchers are also timers of a kind, but they are very versatile	1069	Periodic watchers are also timers of a kind, but they are very versatile
618	(and unfortunately a bit complex).	1070	(and unfortunately a bit complex).
619		1071
620	Unlike C<ev_timer>'s, they are not based on real time (or relative time)	1072	Unlike C<ev_timer>'s, they are not based on real time (or relative time)
621	but on wallclock time (absolute time). You can tell a periodic watcher	1073	but on wallclock time (absolute time). You can tell a periodic watcher
622	to trigger "at" some specific point in time. For example, if you tell a	1074	to trigger "at" some specific point in time. For example, if you tell a
623	periodic watcher to trigger in 10 seconds (by specifiying e.g. c<ev_now ()	1075	periodic watcher to trigger in 10 seconds (by specifiying e.g. C<ev_now ()
624	+ 10.>) and then reset your system clock to the last year, then it will	1076	+ 10.>) and then reset your system clock to the last year, then it will
625	take a year to trigger the event (unlike an C<ev_timer>, which would trigger	1077	take a year to trigger the event (unlike an C<ev_timer>, which would trigger
626	roughly 10 seconds later and of course not if you reset your system time	1078	roughly 10 seconds later and of course not if you reset your system time
627	again).	1079	again).
628		1080
…		…
712	Simply stops and restarts the periodic watcher again. This is only useful	1164	Simply stops and restarts the periodic watcher again. This is only useful
713	when you changed some parameters or the reschedule callback would return	1165	when you changed some parameters or the reschedule callback would return
714	a different time than the last time it was called (e.g. in a crond like	1166	a different time than the last time it was called (e.g. in a crond like
715	program when the crontabs have changed).	1167	program when the crontabs have changed).
716		1168
		1169	=item ev_tstamp interval [read-write]
		1170
		1171	The current interval value. Can be modified any time, but changes only
		1172	take effect when the periodic timer fires or C<ev_periodic_again> is being
		1173	called.
		1174
		1175	=item ev_tstamp (*reschedule_cb)(struct ev_periodic *w, ev_tstamp now) [read-write]
		1176
		1177	The current reschedule callback, or C<0>, if this functionality is
		1178	switched off. Can be changed any time, but changes only take effect when
		1179	the periodic timer fires or C<ev_periodic_again> is being called.
		1180
717	=back	1181	=back
718		1182
		1183	Example: Call a callback every hour, or, more precisely, whenever the
		1184	system clock is divisible by 3600. The callback invocation times have
		1185	potentially a lot of jittering, but good long-term stability.
		1186
		1187	static void
		1188	clock_cb (struct ev_loop *loop, struct ev_io *w, int revents)
		1189	{
		1190	... its now a full hour (UTC, or TAI or whatever your clock follows)
		1191	}
		1192
		1193	struct ev_periodic hourly_tick;
		1194	ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);
		1195	ev_periodic_start (loop, &hourly_tick);
		1196
		1197	Example: The same as above, but use a reschedule callback to do it:
		1198
		1199	#include <math.h>
		1200
		1201	static ev_tstamp
		1202	my_scheduler_cb (struct ev_periodic *w, ev_tstamp now)
		1203	{
		1204	return fmod (now, 3600.) + 3600.;
		1205	}
		1206
		1207	ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
		1208
		1209	Example: Call a callback every hour, starting now:
		1210
		1211	struct ev_periodic hourly_tick;
		1212	ev_periodic_init (&hourly_tick, clock_cb,
		1213	fmod (ev_now (loop), 3600.), 3600., 0);
		1214	ev_periodic_start (loop, &hourly_tick);
		1215
		1216
719	=head2 C<ev_signal> - signal me when a signal gets signalled	1217	=head2 C<ev_signal> - signal me when a signal gets signalled!
720		1218
721	Signal watchers will trigger an event when the process receives a specific	1219	Signal watchers will trigger an event when the process receives a specific
722	signal one or more times. Even though signals are very asynchronous, libev	1220	signal one or more times. Even though signals are very asynchronous, libev
723	will try it's best to deliver signals synchronously, i.e. as part of the	1221	will try it's best to deliver signals synchronously, i.e. as part of the
724	normal event processing, like any other event.	1222	normal event processing, like any other event.
…		…
737	=item ev_signal_set (ev_signal *, int signum)	1235	=item ev_signal_set (ev_signal *, int signum)
738		1236
739	Configures the watcher to trigger on the given signal number (usually one	1237	Configures the watcher to trigger on the given signal number (usually one
740	of the C<SIGxxx> constants).	1238	of the C<SIGxxx> constants).
741		1239
		1240	=item int signum [read-only]
		1241
		1242	The signal the watcher watches out for.
		1243
742	=back	1244	=back
743		1245
		1246
744	=head2 C<ev_child> - wait for pid status changes	1247	=head2 C<ev_child> - watch out for process status changes
745		1248
746	Child watchers trigger when your process receives a SIGCHLD in response to	1249	Child watchers trigger when your process receives a SIGCHLD in response to
747	some child status changes (most typically when a child of yours dies).	1250	some child status changes (most typically when a child of yours dies).
748		1251
749	=over 4	1252	=over 4
…		…
757	at the C<rstatus> member of the C<ev_child> watcher structure to see	1260	at the C<rstatus> member of the C<ev_child> watcher structure to see
758	the status word (use the macros from C<sys/wait.h> and see your systems	1261	the status word (use the macros from C<sys/wait.h> and see your systems
759	C<waitpid> documentation). The C<rpid> member contains the pid of the	1262	C<waitpid> documentation). The C<rpid> member contains the pid of the
760	process causing the status change.	1263	process causing the status change.
761		1264
		1265	=item int pid [read-only]
		1266
		1267	The process id this watcher watches out for, or C<0>, meaning any process id.
		1268
		1269	=item int rpid [read-write]
		1270
		1271	The process id that detected a status change.
		1272
		1273	=item int rstatus [read-write]
		1274
		1275	The process exit/trace status caused by C<rpid> (see your systems
		1276	C<waitpid> and C<sys/wait.h> documentation for details).
		1277
762	=back	1278	=back
763		1279
		1280	Example: Try to exit cleanly on SIGINT and SIGTERM.
		1281
		1282	static void
		1283	sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)
		1284	{
		1285	ev_unloop (loop, EVUNLOOP_ALL);
		1286	}
		1287
		1288	struct ev_signal signal_watcher;
		1289	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
		1290	ev_signal_start (loop, &sigint_cb);
		1291
		1292
		1293	=head2 C<ev_stat> - did the file attributes just change?
		1294
		1295	This watches a filesystem path for attribute changes. That is, it calls
		1296	C<stat> regularly (or when the OS says it changed) and sees if it changed
		1297	compared to the last time, invoking the callback if it did.
		1298
		1299	The path does not need to exist: changing from "path exists" to "path does
		1300	not exist" is a status change like any other. The condition "path does
		1301	not exist" is signified by the C<st_nlink> field being zero (which is
		1302	otherwise always forced to be at least one) and all the other fields of
		1303	the stat buffer having unspecified contents.
		1304
		1305	The path I<should> be absolute and I<must not> end in a slash. If it is
		1306	relative and your working directory changes, the behaviour is undefined.
		1307
		1308	Since there is no standard to do this, the portable implementation simply
		1309	calls C<stat (2)> regularly on the path to see if it changed somehow. You
		1310	can specify a recommended polling interval for this case. If you specify
		1311	a polling interval of C<0> (highly recommended!) then a I<suitable,
		1312	unspecified default> value will be used (which you can expect to be around
		1313	five seconds, although this might change dynamically). Libev will also
		1314	impose a minimum interval which is currently around C<0.1>, but thats
		1315	usually overkill.
		1316
		1317	This watcher type is not meant for massive numbers of stat watchers,
		1318	as even with OS-supported change notifications, this can be
		1319	resource-intensive.
		1320
		1321	At the time of this writing, only the Linux inotify interface is
		1322	implemented (implementing kqueue support is left as an exercise for the
		1323	reader). Inotify will be used to give hints only and should not change the
		1324	semantics of C<ev_stat> watchers, which means that libev sometimes needs
		1325	to fall back to regular polling again even with inotify, but changes are
		1326	usually detected immediately, and if the file exists there will be no
		1327	polling.
		1328
		1329	=over 4
		1330
		1331	=item ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)
		1332
		1333	=item ev_stat_set (ev_stat *, const char *path, ev_tstamp interval)
		1334
		1335	Configures the watcher to wait for status changes of the given
		1336	C<path>. The C<interval> is a hint on how quickly a change is expected to
		1337	be detected and should normally be specified as C<0> to let libev choose
		1338	a suitable value. The memory pointed to by C<path> must point to the same
		1339	path for as long as the watcher is active.
		1340
		1341	The callback will be receive C<EV_STAT> when a change was detected,
		1342	relative to the attributes at the time the watcher was started (or the
		1343	last change was detected).
		1344
		1345	=item ev_stat_stat (ev_stat *)
		1346
		1347	Updates the stat buffer immediately with new values. If you change the
		1348	watched path in your callback, you could call this fucntion to avoid
		1349	detecting this change (while introducing a race condition). Can also be
		1350	useful simply to find out the new values.
		1351
		1352	=item ev_statdata attr [read-only]
		1353
		1354	The most-recently detected attributes of the file. Although the type is of
		1355	C<ev_statdata>, this is usually the (or one of the) C<struct stat> types
		1356	suitable for your system. If the C<st_nlink> member is C<0>, then there
		1357	was some error while C<stat>ing the file.
		1358
		1359	=item ev_statdata prev [read-only]
		1360
		1361	The previous attributes of the file. The callback gets invoked whenever
		1362	C<prev> != C<attr>.
		1363
		1364	=item ev_tstamp interval [read-only]
		1365
		1366	The specified interval.
		1367
		1368	=item const char *path [read-only]
		1369
		1370	The filesystem path that is being watched.
		1371
		1372	=back
		1373
		1374	Example: Watch C</etc/passwd> for attribute changes.
		1375
		1376	static void
		1377	passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)
		1378	{
		1379	/* /etc/passwd changed in some way */
		1380	if (w->attr.st_nlink)
		1381	{
		1382	printf ("passwd current size %ld\n", (long)w->attr.st_size);
		1383	printf ("passwd current atime %ld\n", (long)w->attr.st_mtime);
		1384	printf ("passwd current mtime %ld\n", (long)w->attr.st_mtime);
		1385	}
		1386	else
		1387	/* you shalt not abuse printf for puts */
		1388	puts ("wow, /etc/passwd is not there, expect problems. "
		1389	"if this is windows, they already arrived\n");
		1390	}
		1391
		1392	...
		1393	ev_stat passwd;
		1394
		1395	ev_stat_init (&passwd, passwd_cb, "/etc/passwd");
		1396	ev_stat_start (loop, &passwd);
		1397
		1398
764	=head2 C<ev_idle> - when you've got nothing better to do	1399	=head2 C<ev_idle> - when you've got nothing better to do...
765		1400
766	Idle watchers trigger events when there are no other events are pending	1401	Idle watchers trigger events when no other events of the same or higher
767	(prepare, check and other idle watchers do not count). That is, as long	1402	priority are pending (prepare, check and other idle watchers do not
768	as your process is busy handling sockets or timeouts (or even signals,	1403	count).
769	imagine) it will not be triggered. But when your process is idle all idle	1404
770	watchers are being called again and again, once per event loop iteration -	1405	That is, as long as your process is busy handling sockets or timeouts
		1406	(or even signals, imagine) of the same or higher priority it will not be
		1407	triggered. But when your process is idle (or only lower-priority watchers
		1408	are pending), the idle watchers are being called once per event loop
771	until stopped, that is, or your process receives more events and becomes	1409	iteration - until stopped, that is, or your process receives more events
772	busy.	1410	and becomes busy again with higher priority stuff.
773		1411
774	The most noteworthy effect is that as long as any idle watchers are	1412	The most noteworthy effect is that as long as any idle watchers are
775	active, the process will not block when waiting for new events.	1413	active, the process will not block when waiting for new events.
776		1414
777	Apart from keeping your process non-blocking (which is a useful	1415	Apart from keeping your process non-blocking (which is a useful
…		…
787	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,	1425	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,
788	believe me.	1426	believe me.
789		1427
790	=back	1428	=back
791		1429
		1430	Example: Dynamically allocate an C<ev_idle> watcher, start it, and in the
		1431	callback, free it. Also, use no error checking, as usual.
		1432
		1433	static void
		1434	idle_cb (struct ev_loop *loop, struct ev_idle *w, int revents)
		1435	{
		1436	free (w);
		1437	// now do something you wanted to do when the program has
		1438	// no longer asnything immediate to do.
		1439	}
		1440
		1441	struct ev_idle *idle_watcher = malloc (sizeof (struct ev_idle));
		1442	ev_idle_init (idle_watcher, idle_cb);
		1443	ev_idle_start (loop, idle_cb);
		1444
		1445
792	=head2 C<ev_prepare> and C<ev_check> - customise your event loop	1446	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!
793		1447
794	Prepare and check watchers are usually (but not always) used in tandem:	1448	Prepare and check watchers are usually (but not always) used in tandem:
795	prepare watchers get invoked before the process blocks and check watchers	1449	prepare watchers get invoked before the process blocks and check watchers
796	afterwards.	1450	afterwards.
797		1451
		1452	You I<must not> call C<ev_loop> or similar functions that enter
		1453	the current event loop from either C<ev_prepare> or C<ev_check>
		1454	watchers. Other loops than the current one are fine, however. The
		1455	rationale behind this is that you do not need to check for recursion in
		1456	those watchers, i.e. the sequence will always be C<ev_prepare>, blocking,
		1457	C<ev_check> so if you have one watcher of each kind they will always be
		1458	called in pairs bracketing the blocking call.
		1459
798	Their main purpose is to integrate other event mechanisms into libev. This	1460	Their main purpose is to integrate other event mechanisms into libev and
799	could be used, for example, to track variable changes, implement your own	1461	their use is somewhat advanced. This could be used, for example, to track
800	watchers, integrate net-snmp or a coroutine library and lots more.	1462	variable changes, implement your own watchers, integrate net-snmp or a
		1463	coroutine library and lots more. They are also occasionally useful if
		1464	you cache some data and want to flush it before blocking (for example,
		1465	in X programs you might want to do an C<XFlush ()> in an C<ev_prepare>
		1466	watcher).
801		1467
802	This is done by examining in each prepare call which file descriptors need	1468	This is done by examining in each prepare call which file descriptors need
803	to be watched by the other library, registering C<ev_io> watchers for	1469	to be watched by the other library, registering C<ev_io> watchers for
804	them and starting an C<ev_timer> watcher for any timeouts (many libraries	1470	them and starting an C<ev_timer> watcher for any timeouts (many libraries
805	provide just this functionality). Then, in the check watcher you check for	1471	provide just this functionality). Then, in the check watcher you check for
…		…
827	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>	1493	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>
828	macros, but using them is utterly, utterly and completely pointless.	1494	macros, but using them is utterly, utterly and completely pointless.
829		1495
830	=back	1496	=back
831		1497
		1498	There are a number of principal ways to embed other event loops or modules
		1499	into libev. Here are some ideas on how to include libadns into libev
		1500	(there is a Perl module named C<EV::ADNS> that does this, which you could
		1501	use for an actually working example. Another Perl module named C<EV::Glib>
		1502	embeds a Glib main context into libev, and finally, C<Glib::EV> embeds EV
		1503	into the Glib event loop).
		1504
		1505	Method 1: Add IO watchers and a timeout watcher in a prepare handler,
		1506	and in a check watcher, destroy them and call into libadns. What follows
		1507	is pseudo-code only of course. This requires you to either use a low
		1508	priority for the check watcher or use C<ev_clear_pending> explicitly, as
		1509	the callbacks for the IO/timeout watchers might not have been called yet.
		1510
		1511	static ev_io iow [nfd];
		1512	static ev_timer tw;
		1513
		1514	static void
		1515	io_cb (ev_loop *loop, ev_io *w, int revents)
		1516	{
		1517	}
		1518
		1519	// create io watchers for each fd and a timer before blocking
		1520	static void
		1521	adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents)
		1522	{
		1523	int timeout = 3600000;
		1524	struct pollfd fds [nfd];
		1525	// actual code will need to loop here and realloc etc.
		1526	adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));
		1527
		1528	/* the callback is illegal, but won't be called as we stop during check */
		1529	ev_timer_init (&tw, 0, timeout * 1e-3);
		1530	ev_timer_start (loop, &tw);
		1531
		1532	// create one ev_io per pollfd
		1533	for (int i = 0; i < nfd; ++i)
		1534	{
		1535	ev_io_init (iow + i, io_cb, fds [i].fd,
		1536	((fds [i].events & POLLIN ? EV_READ : 0)
		1537	| (fds [i].events & POLLOUT ? EV_WRITE : 0)));
		1538
		1539	fds [i].revents = 0;
		1540	ev_io_start (loop, iow + i);
		1541	}
		1542	}
		1543
		1544	// stop all watchers after blocking
		1545	static void
		1546	adns_check_cb (ev_loop *loop, ev_check *w, int revents)
		1547	{
		1548	ev_timer_stop (loop, &tw);
		1549
		1550	for (int i = 0; i < nfd; ++i)
		1551	{
		1552	// set the relevant poll flags
		1553	// could also call adns_processreadable etc. here
		1554	struct pollfd *fd = fds + i;
		1555	int revents = ev_clear_pending (iow + i);
		1556	if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;
		1557	if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;
		1558
		1559	// now stop the watcher
		1560	ev_io_stop (loop, iow + i);
		1561	}
		1562
		1563	adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));
		1564	}
		1565
		1566	Method 2: This would be just like method 1, but you run C<adns_afterpoll>
		1567	in the prepare watcher and would dispose of the check watcher.
		1568
		1569	Method 3: If the module to be embedded supports explicit event
		1570	notification (adns does), you can also make use of the actual watcher
		1571	callbacks, and only destroy/create the watchers in the prepare watcher.
		1572
		1573	static void
		1574	timer_cb (EV_P_ ev_timer *w, int revents)
		1575	{
		1576	adns_state ads = (adns_state)w->data;
		1577	update_now (EV_A);
		1578
		1579	adns_processtimeouts (ads, &tv_now);
		1580	}
		1581
		1582	static void
		1583	io_cb (EV_P_ ev_io *w, int revents)
		1584	{
		1585	adns_state ads = (adns_state)w->data;
		1586	update_now (EV_A);
		1587
		1588	if (revents & EV_READ ) adns_processreadable (ads, w->fd, &tv_now);
		1589	if (revents & EV_WRITE) adns_processwriteable (ads, w->fd, &tv_now);
		1590	}
		1591
		1592	// do not ever call adns_afterpoll
		1593
		1594	Method 4: Do not use a prepare or check watcher because the module you
		1595	want to embed is too inflexible to support it. Instead, youc na override
		1596	their poll function. The drawback with this solution is that the main
		1597	loop is now no longer controllable by EV. The C<Glib::EV> module does
		1598	this.
		1599
		1600	static gint
		1601	event_poll_func (GPollFD *fds, guint nfds, gint timeout)
		1602	{
		1603	int got_events = 0;
		1604
		1605	for (n = 0; n < nfds; ++n)
		1606	// create/start io watcher that sets the relevant bits in fds[n] and increment got_events
		1607
		1608	if (timeout >= 0)
		1609	// create/start timer
		1610
		1611	// poll
		1612	ev_loop (EV_A_ 0);
		1613
		1614	// stop timer again
		1615	if (timeout >= 0)
		1616	ev_timer_stop (EV_A_ &to);
		1617
		1618	// stop io watchers again - their callbacks should have set
		1619	for (n = 0; n < nfds; ++n)
		1620	ev_io_stop (EV_A_ iow [n]);
		1621
		1622	return got_events;
		1623	}
		1624
		1625
		1626	=head2 C<ev_embed> - when one backend isn't enough...
		1627
		1628	This is a rather advanced watcher type that lets you embed one event loop
		1629	into another (currently only C<ev_io> events are supported in the embedded
		1630	loop, other types of watchers might be handled in a delayed or incorrect
		1631	fashion and must not be used).
		1632
		1633	There are primarily two reasons you would want that: work around bugs and
		1634	prioritise I/O.
		1635
		1636	As an example for a bug workaround, the kqueue backend might only support
		1637	sockets on some platform, so it is unusable as generic backend, but you
		1638	still want to make use of it because you have many sockets and it scales
		1639	so nicely. In this case, you would create a kqueue-based loop and embed it
		1640	into your default loop (which might use e.g. poll). Overall operation will
		1641	be a bit slower because first libev has to poll and then call kevent, but
		1642	at least you can use both at what they are best.
		1643
		1644	As for prioritising I/O: rarely you have the case where some fds have
		1645	to be watched and handled very quickly (with low latency), and even
		1646	priorities and idle watchers might have too much overhead. In this case
		1647	you would put all the high priority stuff in one loop and all the rest in
		1648	a second one, and embed the second one in the first.
		1649
		1650	As long as the watcher is active, the callback will be invoked every time
		1651	there might be events pending in the embedded loop. The callback must then
		1652	call C<ev_embed_sweep (mainloop, watcher)> to make a single sweep and invoke
		1653	their callbacks (you could also start an idle watcher to give the embedded
		1654	loop strictly lower priority for example). You can also set the callback
		1655	to C<0>, in which case the embed watcher will automatically execute the
		1656	embedded loop sweep.
		1657
		1658	As long as the watcher is started it will automatically handle events. The
		1659	callback will be invoked whenever some events have been handled. You can
		1660	set the callback to C<0> to avoid having to specify one if you are not
		1661	interested in that.
		1662
		1663	Also, there have not currently been made special provisions for forking:
		1664	when you fork, you not only have to call C<ev_loop_fork> on both loops,
		1665	but you will also have to stop and restart any C<ev_embed> watchers
		1666	yourself.
		1667
		1668	Unfortunately, not all backends are embeddable, only the ones returned by
		1669	C<ev_embeddable_backends> are, which, unfortunately, does not include any
		1670	portable one.
		1671
		1672	So when you want to use this feature you will always have to be prepared
		1673	that you cannot get an embeddable loop. The recommended way to get around
		1674	this is to have a separate variables for your embeddable loop, try to
		1675	create it, and if that fails, use the normal loop for everything:
		1676
		1677	struct ev_loop *loop_hi = ev_default_init (0);
		1678	struct ev_loop *loop_lo = 0;
		1679	struct ev_embed embed;
		1680
		1681	// see if there is a chance of getting one that works
		1682	// (remember that a flags value of 0 means autodetection)
		1683	loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
		1684	? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
		1685	: 0;
		1686
		1687	// if we got one, then embed it, otherwise default to loop_hi
		1688	if (loop_lo)
		1689	{
		1690	ev_embed_init (&embed, 0, loop_lo);
		1691	ev_embed_start (loop_hi, &embed);
		1692	}
		1693	else
		1694	loop_lo = loop_hi;
		1695
		1696	=over 4
		1697
		1698	=item ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)
		1699
		1700	=item ev_embed_set (ev_embed *, callback, struct ev_loop *embedded_loop)
		1701
		1702	Configures the watcher to embed the given loop, which must be
		1703	embeddable. If the callback is C<0>, then C<ev_embed_sweep> will be
		1704	invoked automatically, otherwise it is the responsibility of the callback
		1705	to invoke it (it will continue to be called until the sweep has been done,
		1706	if you do not want thta, you need to temporarily stop the embed watcher).
		1707
		1708	=item ev_embed_sweep (loop, ev_embed *)
		1709
		1710	Make a single, non-blocking sweep over the embedded loop. This works
		1711	similarly to C<ev_loop (embedded_loop, EVLOOP_NONBLOCK)>, but in the most
		1712	apropriate way for embedded loops.
		1713
		1714	=item struct ev_loop *loop [read-only]
		1715
		1716	The embedded event loop.
		1717
		1718	=back
		1719
		1720
		1721	=head2 C<ev_fork> - the audacity to resume the event loop after a fork
		1722
		1723	Fork watchers are called when a C<fork ()> was detected (usually because
		1724	whoever is a good citizen cared to tell libev about it by calling
		1725	C<ev_default_fork> or C<ev_loop_fork>). The invocation is done before the
		1726	event loop blocks next and before C<ev_check> watchers are being called,
		1727	and only in the child after the fork. If whoever good citizen calling
		1728	C<ev_default_fork> cheats and calls it in the wrong process, the fork
		1729	handlers will be invoked, too, of course.
		1730
		1731	=over 4
		1732
		1733	=item ev_fork_init (ev_signal *, callback)
		1734
		1735	Initialises and configures the fork watcher - it has no parameters of any
		1736	kind. There is a C<ev_fork_set> macro, but using it is utterly pointless,
		1737	believe me.
		1738
		1739	=back
		1740
		1741
832	=head1 OTHER FUNCTIONS	1742	=head1 OTHER FUNCTIONS
833		1743
834	There are some other functions of possible interest. Described. Here. Now.	1744	There are some other functions of possible interest. Described. Here. Now.
835		1745
836	=over 4	1746	=over 4
…		…
865	/* stdin might have data for us, joy! */;	1775	/* stdin might have data for us, joy! */;
866	}	1776	}
867		1777
868	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);	1778	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
869		1779
870	=item ev_feed_event (loop, watcher, int events)	1780	=item ev_feed_event (ev_loop *, watcher *, int revents)
871		1781
872	Feeds the given event set into the event loop, as if the specified event	1782	Feeds the given event set into the event loop, as if the specified event
873	had happened for the specified watcher (which must be a pointer to an	1783	had happened for the specified watcher (which must be a pointer to an
874	initialised but not necessarily started event watcher).	1784	initialised but not necessarily started event watcher).
875		1785
876	=item ev_feed_fd_event (loop, int fd, int revents)	1786	=item ev_feed_fd_event (ev_loop *, int fd, int revents)
877		1787
878	Feed an event on the given fd, as if a file descriptor backend detected	1788	Feed an event on the given fd, as if a file descriptor backend detected
879	the given events it.	1789	the given events it.
880		1790
881	=item ev_feed_signal_event (loop, int signum)	1791	=item ev_feed_signal_event (ev_loop *loop, int signum)
882		1792
883	Feed an event as if the given signal occured (loop must be the default loop!).	1793	Feed an event as if the given signal occured (C<loop> must be the default
		1794	loop!).
884		1795
885	=back	1796	=back
		1797
886		1798
887	=head1 LIBEVENT EMULATION	1799	=head1 LIBEVENT EMULATION
888		1800
889	Libev offers a compatibility emulation layer for libevent. It cannot	1801	Libev offers a compatibility emulation layer for libevent. It cannot
890	emulate the internals of libevent, so here are some usage hints:	1802	emulate the internals of libevent, so here are some usage hints:
…		…
911		1823
912	=back	1824	=back
913		1825
914	=head1 C++ SUPPORT	1826	=head1 C++ SUPPORT
915		1827
916	TBD.	1828	Libev comes with some simplistic wrapper classes for C++ that mainly allow
		1829	you to use some convinience methods to start/stop watchers and also change
		1830	the callback model to a model using method callbacks on objects.
		1831
		1832	To use it,
		1833
		1834	#include <ev++.h>
		1835
		1836	This automatically includes F<ev.h> and puts all of its definitions (many
		1837	of them macros) into the global namespace. All C++ specific things are
		1838	put into the C<ev> namespace. It should support all the same embedding
		1839	options as F<ev.h>, most notably C<EV_MULTIPLICITY>.
		1840
		1841	Care has been taken to keep the overhead low. The only data member the C++
		1842	classes add (compared to plain C-style watchers) is the event loop pointer
		1843	that the watcher is associated with (or no additional members at all if
		1844	you disable C<EV_MULTIPLICITY> when embedding libev).
		1845
		1846	Currently, functions, and static and non-static member functions can be
		1847	used as callbacks. Other types should be easy to add as long as they only
		1848	need one additional pointer for context. If you need support for other
		1849	types of functors please contact the author (preferably after implementing
		1850	it).
		1851
		1852	Here is a list of things available in the C<ev> namespace:
		1853
		1854	=over 4
		1855
		1856	=item C<ev::READ>, C<ev::WRITE> etc.
		1857
		1858	These are just enum values with the same values as the C<EV_READ> etc.
		1859	macros from F<ev.h>.
		1860
		1861	=item C<ev::tstamp>, C<ev::now>
		1862
		1863	Aliases to the same types/functions as with the C<ev_> prefix.
		1864
		1865	=item C<ev::io>, C<ev::timer>, C<ev::periodic>, C<ev::idle>, C<ev::sig> etc.
		1866
		1867	For each C<ev_TYPE> watcher in F<ev.h> there is a corresponding class of
		1868	the same name in the C<ev> namespace, with the exception of C<ev_signal>
		1869	which is called C<ev::sig> to avoid clashes with the C<signal> macro
		1870	defines by many implementations.
		1871
		1872	All of those classes have these methods:
		1873
		1874	=over 4
		1875
		1876	=item ev::TYPE::TYPE ()
		1877
		1878	=item ev::TYPE::TYPE (struct ev_loop *)
		1879
		1880	=item ev::TYPE::~TYPE
		1881
		1882	The constructor (optionally) takes an event loop to associate the watcher
		1883	with. If it is omitted, it will use C<EV_DEFAULT>.
		1884
		1885	The constructor calls C<ev_init> for you, which means you have to call the
		1886	C<set> method before starting it.
		1887
		1888	It will not set a callback, however: You have to call the templated C<set>
		1889	method to set a callback before you can start the watcher.
		1890
		1891	(The reason why you have to use a method is a limitation in C++ which does
		1892	not allow explicit template arguments for constructors).
		1893
		1894	The destructor automatically stops the watcher if it is active.
		1895
		1896	=item w->set<class, &class::method> (object *)
		1897
		1898	This method sets the callback method to call. The method has to have a
		1899	signature of C<void (*)(ev_TYPE &, int)>, it receives the watcher as
		1900	first argument and the C<revents> as second. The object must be given as
		1901	parameter and is stored in the C<data> member of the watcher.
		1902
		1903	This method synthesizes efficient thunking code to call your method from
		1904	the C callback that libev requires. If your compiler can inline your
		1905	callback (i.e. it is visible to it at the place of the C<set> call and
		1906	your compiler is good :), then the method will be fully inlined into the
		1907	thunking function, making it as fast as a direct C callback.
		1908
		1909	Example: simple class declaration and watcher initialisation
		1910
		1911	struct myclass
		1912	{
		1913	void io_cb (ev::io &w, int revents) { }
		1914	}
		1915
		1916	myclass obj;
		1917	ev::io iow;
		1918	iow.set <myclass, &myclass::io_cb> (&obj);
		1919
		1920	=item w->set<function> (void *data = 0)
		1921
		1922	Also sets a callback, but uses a static method or plain function as
		1923	callback. The optional C<data> argument will be stored in the watcher's
		1924	C<data> member and is free for you to use.
		1925
		1926	The prototype of the C<function> must be C<void (*)(ev::TYPE &w, int)>.
		1927
		1928	See the method-C<set> above for more details.
		1929
		1930	Example:
		1931
		1932	static void io_cb (ev::io &w, int revents) { }
		1933	iow.set <io_cb> ();
		1934
		1935	=item w->set (struct ev_loop *)
		1936
		1937	Associates a different C<struct ev_loop> with this watcher. You can only
		1938	do this when the watcher is inactive (and not pending either).
		1939
		1940	=item w->set ([args])
		1941
		1942	Basically the same as C<ev_TYPE_set>, with the same args. Must be
		1943	called at least once. Unlike the C counterpart, an active watcher gets
		1944	automatically stopped and restarted when reconfiguring it with this
		1945	method.
		1946
		1947	=item w->start ()
		1948
		1949	Starts the watcher. Note that there is no C<loop> argument, as the
		1950	constructor already stores the event loop.
		1951
		1952	=item w->stop ()
		1953
		1954	Stops the watcher if it is active. Again, no C<loop> argument.
		1955
		1956	=item w->again () C<ev::timer>, C<ev::periodic> only
		1957
		1958	For C<ev::timer> and C<ev::periodic>, this invokes the corresponding
		1959	C<ev_TYPE_again> function.
		1960
		1961	=item w->sweep () C<ev::embed> only
		1962
		1963	Invokes C<ev_embed_sweep>.
		1964
		1965	=item w->update () C<ev::stat> only
		1966
		1967	Invokes C<ev_stat_stat>.
		1968
		1969	=back
		1970
		1971	=back
		1972
		1973	Example: Define a class with an IO and idle watcher, start one of them in
		1974	the constructor.
		1975
		1976	class myclass
		1977	{
		1978	ev_io io; void io_cb (ev::io &w, int revents);
		1979	ev_idle idle void idle_cb (ev::idle &w, int revents);
		1980
		1981	myclass ();
		1982	}
		1983
		1984	myclass::myclass (int fd)
		1985	{
		1986	io .set <myclass, &myclass::io_cb > (this);
		1987	idle.set <myclass, &myclass::idle_cb> (this);
		1988
		1989	io.start (fd, ev::READ);
		1990	}
		1991
		1992
		1993	=head1 MACRO MAGIC
		1994
		1995	Libev can be compiled with a variety of options, the most fundemantal is
		1996	C<EV_MULTIPLICITY>. This option determines whether (most) functions and
		1997	callbacks have an initial C<struct ev_loop *> argument.
		1998
		1999	To make it easier to write programs that cope with either variant, the
		2000	following macros are defined:
		2001
		2002	=over 4
		2003
		2004	=item C<EV_A>, C<EV_A_>
		2005
		2006	This provides the loop I<argument> for functions, if one is required ("ev
		2007	loop argument"). The C<EV_A> form is used when this is the sole argument,
		2008	C<EV_A_> is used when other arguments are following. Example:
		2009
		2010	ev_unref (EV_A);
		2011	ev_timer_add (EV_A_ watcher);
		2012	ev_loop (EV_A_ 0);
		2013
		2014	It assumes the variable C<loop> of type C<struct ev_loop *> is in scope,
		2015	which is often provided by the following macro.
		2016
		2017	=item C<EV_P>, C<EV_P_>
		2018
		2019	This provides the loop I<parameter> for functions, if one is required ("ev
		2020	loop parameter"). The C<EV_P> form is used when this is the sole parameter,
		2021	C<EV_P_> is used when other parameters are following. Example:
		2022
		2023	// this is how ev_unref is being declared
		2024	static void ev_unref (EV_P);
		2025
		2026	// this is how you can declare your typical callback
		2027	static void cb (EV_P_ ev_timer *w, int revents)
		2028
		2029	It declares a parameter C<loop> of type C<struct ev_loop *>, quite
		2030	suitable for use with C<EV_A>.
		2031
		2032	=item C<EV_DEFAULT>, C<EV_DEFAULT_>
		2033
		2034	Similar to the other two macros, this gives you the value of the default
		2035	loop, if multiple loops are supported ("ev loop default").
		2036
		2037	=back
		2038
		2039	Example: Declare and initialise a check watcher, utilising the above
		2040	macros so it will work regardless of whether multiple loops are supported
		2041	or not.
		2042
		2043	static void
		2044	check_cb (EV_P_ ev_timer *w, int revents)
		2045	{
		2046	ev_check_stop (EV_A_ w);
		2047	}
		2048
		2049	ev_check check;
		2050	ev_check_init (&check, check_cb);
		2051	ev_check_start (EV_DEFAULT_ &check);
		2052	ev_loop (EV_DEFAULT_ 0);
		2053
		2054	=head1 EMBEDDING
		2055
		2056	Libev can (and often is) directly embedded into host
		2057	applications. Examples of applications that embed it include the Deliantra
		2058	Game Server, the EV perl module, the GNU Virtual Private Ethernet (gvpe)
		2059	and rxvt-unicode.
		2060
		2061	The goal is to enable you to just copy the neecssary files into your
		2062	source directory without having to change even a single line in them, so
		2063	you can easily upgrade by simply copying (or having a checked-out copy of
		2064	libev somewhere in your source tree).
		2065
		2066	=head2 FILESETS
		2067
		2068	Depending on what features you need you need to include one or more sets of files
		2069	in your app.
		2070
		2071	=head3 CORE EVENT LOOP
		2072
		2073	To include only the libev core (all the C<ev_*> functions), with manual
		2074	configuration (no autoconf):
		2075
		2076	#define EV_STANDALONE 1
		2077	#include "ev.c"
		2078
		2079	This will automatically include F<ev.h>, too, and should be done in a
		2080	single C source file only to provide the function implementations. To use
		2081	it, do the same for F<ev.h> in all files wishing to use this API (best
		2082	done by writing a wrapper around F<ev.h> that you can include instead and
		2083	where you can put other configuration options):
		2084
		2085	#define EV_STANDALONE 1
		2086	#include "ev.h"
		2087
		2088	Both header files and implementation files can be compiled with a C++
		2089	compiler (at least, thats a stated goal, and breakage will be treated
		2090	as a bug).
		2091
		2092	You need the following files in your source tree, or in a directory
		2093	in your include path (e.g. in libev/ when using -Ilibev):
		2094
		2095	ev.h
		2096	ev.c
		2097	ev_vars.h
		2098	ev_wrap.h
		2099
		2100	ev_win32.c required on win32 platforms only
		2101
		2102	ev_select.c only when select backend is enabled (which is enabled by default)
		2103	ev_poll.c only when poll backend is enabled (disabled by default)
		2104	ev_epoll.c only when the epoll backend is enabled (disabled by default)
		2105	ev_kqueue.c only when the kqueue backend is enabled (disabled by default)
		2106	ev_port.c only when the solaris port backend is enabled (disabled by default)
		2107
		2108	F<ev.c> includes the backend files directly when enabled, so you only need
		2109	to compile this single file.
		2110
		2111	=head3 LIBEVENT COMPATIBILITY API
		2112
		2113	To include the libevent compatibility API, also include:
		2114
		2115	#include "event.c"
		2116
		2117	in the file including F<ev.c>, and:
		2118
		2119	#include "event.h"
		2120
		2121	in the files that want to use the libevent API. This also includes F<ev.h>.
		2122
		2123	You need the following additional files for this:
		2124
		2125	event.h
		2126	event.c
		2127
		2128	=head3 AUTOCONF SUPPORT
		2129
		2130	Instead of using C<EV_STANDALONE=1> and providing your config in
		2131	whatever way you want, you can also C<m4_include([libev.m4])> in your
		2132	F<configure.ac> and leave C<EV_STANDALONE> undefined. F<ev.c> will then
		2133	include F<config.h> and configure itself accordingly.
		2134
		2135	For this of course you need the m4 file:
		2136
		2137	libev.m4
		2138
		2139	=head2 PREPROCESSOR SYMBOLS/MACROS
		2140
		2141	Libev can be configured via a variety of preprocessor symbols you have to define
		2142	before including any of its files. The default is not to build for multiplicity
		2143	and only include the select backend.
		2144
		2145	=over 4
		2146
		2147	=item EV_STANDALONE
		2148
		2149	Must always be C<1> if you do not use autoconf configuration, which
		2150	keeps libev from including F<config.h>, and it also defines dummy
		2151	implementations for some libevent functions (such as logging, which is not
		2152	supported). It will also not define any of the structs usually found in
		2153	F<event.h> that are not directly supported by the libev core alone.
		2154
		2155	=item EV_USE_MONOTONIC
		2156
		2157	If defined to be C<1>, libev will try to detect the availability of the
		2158	monotonic clock option at both compiletime and runtime. Otherwise no use
		2159	of the monotonic clock option will be attempted. If you enable this, you
		2160	usually have to link against librt or something similar. Enabling it when
		2161	the functionality isn't available is safe, though, althoguh you have
		2162	to make sure you link against any libraries where the C<clock_gettime>
		2163	function is hiding in (often F<-lrt>).
		2164
		2165	=item EV_USE_REALTIME
		2166
		2167	If defined to be C<1>, libev will try to detect the availability of the
		2168	realtime clock option at compiletime (and assume its availability at
		2169	runtime if successful). Otherwise no use of the realtime clock option will
		2170	be attempted. This effectively replaces C<gettimeofday> by C<clock_get
		2171	(CLOCK_REALTIME, ...)> and will not normally affect correctness. See tzhe note about libraries
		2172	in the description of C<EV_USE_MONOTONIC>, though.
		2173
		2174	=item EV_USE_SELECT
		2175
		2176	If undefined or defined to be C<1>, libev will compile in support for the
		2177	C<select>(2) backend. No attempt at autodetection will be done: if no
		2178	other method takes over, select will be it. Otherwise the select backend
		2179	will not be compiled in.
		2180
		2181	=item EV_SELECT_USE_FD_SET
		2182
		2183	If defined to C<1>, then the select backend will use the system C<fd_set>
		2184	structure. This is useful if libev doesn't compile due to a missing
		2185	C<NFDBITS> or C<fd_mask> definition or it misguesses the bitset layout on
		2186	exotic systems. This usually limits the range of file descriptors to some
		2187	low limit such as 1024 or might have other limitations (winsocket only
		2188	allows 64 sockets). The C<FD_SETSIZE> macro, set before compilation, might
		2189	influence the size of the C<fd_set> used.
		2190
		2191	=item EV_SELECT_IS_WINSOCKET
		2192
		2193	When defined to C<1>, the select backend will assume that
		2194	select/socket/connect etc. don't understand file descriptors but
		2195	wants osf handles on win32 (this is the case when the select to
		2196	be used is the winsock select). This means that it will call
		2197	C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,
		2198	it is assumed that all these functions actually work on fds, even
		2199	on win32. Should not be defined on non-win32 platforms.
		2200
		2201	=item EV_USE_POLL
		2202
		2203	If defined to be C<1>, libev will compile in support for the C<poll>(2)
		2204	backend. Otherwise it will be enabled on non-win32 platforms. It
		2205	takes precedence over select.
		2206
		2207	=item EV_USE_EPOLL
		2208
		2209	If defined to be C<1>, libev will compile in support for the Linux
		2210	C<epoll>(7) backend. Its availability will be detected at runtime,
		2211	otherwise another method will be used as fallback. This is the
		2212	preferred backend for GNU/Linux systems.
		2213
		2214	=item EV_USE_KQUEUE
		2215
		2216	If defined to be C<1>, libev will compile in support for the BSD style
		2217	C<kqueue>(2) backend. Its actual availability will be detected at runtime,
		2218	otherwise another method will be used as fallback. This is the preferred
		2219	backend for BSD and BSD-like systems, although on most BSDs kqueue only
		2220	supports some types of fds correctly (the only platform we found that
		2221	supports ptys for example was NetBSD), so kqueue might be compiled in, but
		2222	not be used unless explicitly requested. The best way to use it is to find
		2223	out whether kqueue supports your type of fd properly and use an embedded
		2224	kqueue loop.
		2225
		2226	=item EV_USE_PORT
		2227
		2228	If defined to be C<1>, libev will compile in support for the Solaris
		2229	10 port style backend. Its availability will be detected at runtime,
		2230	otherwise another method will be used as fallback. This is the preferred
		2231	backend for Solaris 10 systems.
		2232
		2233	=item EV_USE_DEVPOLL
		2234
		2235	reserved for future expansion, works like the USE symbols above.
		2236
		2237	=item EV_USE_INOTIFY
		2238
		2239	If defined to be C<1>, libev will compile in support for the Linux inotify
		2240	interface to speed up C<ev_stat> watchers. Its actual availability will
		2241	be detected at runtime.
		2242
		2243	=item EV_H
		2244
		2245	The name of the F<ev.h> header file used to include it. The default if
		2246	undefined is C<< <ev.h> >> in F<event.h> and C<"ev.h"> in F<ev.c>. This
		2247	can be used to virtually rename the F<ev.h> header file in case of conflicts.
		2248
		2249	=item EV_CONFIG_H
		2250
		2251	If C<EV_STANDALONE> isn't C<1>, this variable can be used to override
		2252	F<ev.c>'s idea of where to find the F<config.h> file, similarly to
		2253	C<EV_H>, above.
		2254
		2255	=item EV_EVENT_H
		2256
		2257	Similarly to C<EV_H>, this macro can be used to override F<event.c>'s idea
		2258	of how the F<event.h> header can be found.
		2259
		2260	=item EV_PROTOTYPES
		2261
		2262	If defined to be C<0>, then F<ev.h> will not define any function
		2263	prototypes, but still define all the structs and other symbols. This is
		2264	occasionally useful if you want to provide your own wrapper functions
		2265	around libev functions.
		2266
		2267	=item EV_MULTIPLICITY
		2268
		2269	If undefined or defined to C<1>, then all event-loop-specific functions
		2270	will have the C<struct ev_loop *> as first argument, and you can create
		2271	additional independent event loops. Otherwise there will be no support
		2272	for multiple event loops and there is no first event loop pointer
		2273	argument. Instead, all functions act on the single default loop.
		2274
		2275	=item EV_MINPRI
		2276
		2277	=item EV_MAXPRI
		2278
		2279	The range of allowed priorities. C<EV_MINPRI> must be smaller or equal to
		2280	C<EV_MAXPRI>, but otherwise there are no non-obvious limitations. You can
		2281	provide for more priorities by overriding those symbols (usually defined
		2282	to be C<-2> and C<2>, respectively).
		2283
		2284	When doing priority-based operations, libev usually has to linearly search
		2285	all the priorities, so having many of them (hundreds) uses a lot of space
		2286	and time, so using the defaults of five priorities (-2 .. +2) is usually
		2287	fine.
		2288
		2289	If your embedding app does not need any priorities, defining these both to
		2290	C<0> will save some memory and cpu.
		2291
		2292	=item EV_PERIODIC_ENABLE
		2293
		2294	If undefined or defined to be C<1>, then periodic timers are supported. If
		2295	defined to be C<0>, then they are not. Disabling them saves a few kB of
		2296	code.
		2297
		2298	=item EV_IDLE_ENABLE
		2299
		2300	If undefined or defined to be C<1>, then idle watchers are supported. If
		2301	defined to be C<0>, then they are not. Disabling them saves a few kB of
		2302	code.
		2303
		2304	=item EV_EMBED_ENABLE
		2305
		2306	If undefined or defined to be C<1>, then embed watchers are supported. If
		2307	defined to be C<0>, then they are not.
		2308
		2309	=item EV_STAT_ENABLE
		2310
		2311	If undefined or defined to be C<1>, then stat watchers are supported. If
		2312	defined to be C<0>, then they are not.
		2313
		2314	=item EV_FORK_ENABLE
		2315
		2316	If undefined or defined to be C<1>, then fork watchers are supported. If
		2317	defined to be C<0>, then they are not.
		2318
		2319	=item EV_MINIMAL
		2320
		2321	If you need to shave off some kilobytes of code at the expense of some
		2322	speed, define this symbol to C<1>. Currently only used for gcc to override
		2323	some inlining decisions, saves roughly 30% codesize of amd64.
		2324
		2325	=item EV_PID_HASHSIZE
		2326
		2327	C<ev_child> watchers use a small hash table to distribute workload by
		2328	pid. The default size is C<16> (or C<1> with C<EV_MINIMAL>), usually more
		2329	than enough. If you need to manage thousands of children you might want to
		2330	increase this value (I<must> be a power of two).
		2331
		2332	=item EV_INOTIFY_HASHSIZE
		2333
		2334	C<ev_staz> watchers use a small hash table to distribute workload by
		2335	inotify watch id. The default size is C<16> (or C<1> with C<EV_MINIMAL>),
		2336	usually more than enough. If you need to manage thousands of C<ev_stat>
		2337	watchers you might want to increase this value (I<must> be a power of
		2338	two).
		2339
		2340	=item EV_COMMON
		2341
		2342	By default, all watchers have a C<void *data> member. By redefining
		2343	this macro to a something else you can include more and other types of
		2344	members. You have to define it each time you include one of the files,
		2345	though, and it must be identical each time.
		2346
		2347	For example, the perl EV module uses something like this:
		2348
		2349	#define EV_COMMON \
		2350	SV *self; /* contains this struct */ \
		2351	SV *cb_sv, *fh /* note no trailing ";" */
		2352
		2353	=item EV_CB_DECLARE (type)
		2354
		2355	=item EV_CB_INVOKE (watcher, revents)
		2356
		2357	=item ev_set_cb (ev, cb)
		2358
		2359	Can be used to change the callback member declaration in each watcher,
		2360	and the way callbacks are invoked and set. Must expand to a struct member
		2361	definition and a statement, respectively. See the F<ev.v> header file for
		2362	their default definitions. One possible use for overriding these is to
		2363	avoid the C<struct ev_loop *> as first argument in all cases, or to use
		2364	method calls instead of plain function calls in C++.
		2365
		2366	=head2 EXAMPLES
		2367
		2368	For a real-world example of a program the includes libev
		2369	verbatim, you can have a look at the EV perl module
		2370	(L<http://software.schmorp.de/pkg/EV.html>). It has the libev files in
		2371	the F<libev/> subdirectory and includes them in the F<EV/EVAPI.h> (public
		2372	interface) and F<EV.xs> (implementation) files. Only the F<EV.xs> file
		2373	will be compiled. It is pretty complex because it provides its own header
		2374	file.
		2375
		2376	The usage in rxvt-unicode is simpler. It has a F<ev_cpp.h> header file
		2377	that everybody includes and which overrides some configure choices:
		2378
		2379	#define EV_MINIMAL 1
		2380	#define EV_USE_POLL 0
		2381	#define EV_MULTIPLICITY 0
		2382	#define EV_PERIODIC_ENABLE 0
		2383	#define EV_STAT_ENABLE 0
		2384	#define EV_FORK_ENABLE 0
		2385	#define EV_CONFIG_H <config.h>
		2386	#define EV_MINPRI 0
		2387	#define EV_MAXPRI 0
		2388
		2389	#include "ev++.h"
		2390
		2391	And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:
		2392
		2393	#include "ev_cpp.h"
		2394	#include "ev.c"
		2395
		2396
		2397	=head1 COMPLEXITIES
		2398
		2399	In this section the complexities of (many of) the algorithms used inside
		2400	libev will be explained. For complexity discussions about backends see the
		2401	documentation for C<ev_default_init>.
		2402
		2403	All of the following are about amortised time: If an array needs to be
		2404	extended, libev needs to realloc and move the whole array, but this
		2405	happens asymptotically never with higher number of elements, so O(1) might
		2406	mean it might do a lengthy realloc operation in rare cases, but on average
		2407	it is much faster and asymptotically approaches constant time.
		2408
		2409	=over 4
		2410
		2411	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)
		2412
		2413	This means that, when you have a watcher that triggers in one hour and
		2414	there are 100 watchers that would trigger before that then inserting will
		2415	have to skip those 100 watchers.
		2416
		2417	=item Changing timer/periodic watchers (by autorepeat, again): O(log skipped_other_timers)
		2418
		2419	That means that for changing a timer costs less than removing/adding them
		2420	as only the relative motion in the event queue has to be paid for.
		2421
		2422	=item Starting io/check/prepare/idle/signal/child watchers: O(1)
		2423
		2424	These just add the watcher into an array or at the head of a list.
		2425	=item Stopping check/prepare/idle watchers: O(1)
		2426
		2427	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))
		2428
		2429	These watchers are stored in lists then need to be walked to find the
		2430	correct watcher to remove. The lists are usually short (you don't usually
		2431	have many watchers waiting for the same fd or signal).
		2432
		2433	=item Finding the next timer per loop iteration: O(1)
		2434
		2435	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)
		2436
		2437	A change means an I/O watcher gets started or stopped, which requires
		2438	libev to recalculate its status (and possibly tell the kernel).
		2439
		2440	=item Activating one watcher: O(1)
		2441
		2442	=item Priority handling: O(number_of_priorities)
		2443
		2444	Priorities are implemented by allocating some space for each
		2445	priority. When doing priority-based operations, libev usually has to
		2446	linearly search all the priorities.
		2447
		2448	=back
		2449
917		2450
918	=head1 AUTHOR	2451	=head1 AUTHOR
919		2452
920	Marc Lehmann <libev@schmorp.de>.	2453	Marc Lehmann <libev@schmorp.de>.
921		2454

Diff Legend

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