ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/libev/ev.pod

Comparing libev/ev.pod (file contents):
Revision 1.8 by root, Mon Nov 12 08:20:02 2007 UTC vs.
Revision 1.114 by root, Mon Dec 31 01:31:30 2007 UTC

…		…
4		4
5	=head1 SYNOPSIS	5	=head1 SYNOPSIS
6		6
7	#include <ev.h>	7	#include <ev.h>
8		8
		9	=head2 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
10		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>.
		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 occurring), 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
15	To do this, it must take more or less complete control over your process	61	To do this, it must take more or less complete control over your process
16	(or thread) by executing the I<event loop> handler, and will then	62	(or thread) by executing the I<event loop> handler, and will then
17	communicate events via a callback mechanism.	63	communicate events via a callback mechanism.
…		…
19	You register interest in certain events by registering so-called I<event	65	You register interest in certain events by registering so-called I<event
20	watchers>, which are relatively small C structures you initialise with the	66	watchers>, which are relatively small C structures you initialise with the
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	=head2 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	=head2 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 AND OTHER GLOBAL FUNCTIONS	96	=head2 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 some floatingpoint value. Unlike the name
		104	component C<stamp> might indicate, it is also used for time differences
		105	throughout libev.
		106
		107	=head1 GLOBAL FUNCTIONS
		108
		109	These functions can be called anytime, even before initialising the
		110	library in any way.
51		111
52	=over 4	112	=over 4
53		113
54	=item ev_tstamp ev_time ()	114	=item ev_tstamp ev_time ()
55		115
56	Returns the current time as libev would use it.	116	Returns the current time as libev would use it. Please note that the
		117	C<ev_now> function is usually faster and also often returns the timestamp
		118	you actually want to know.
		119
		120	=item ev_sleep (ev_tstamp interval)
		121
		122	Sleep for the given interval: The current thread will be blocked until
		123	either it is interrupted or the given time interval has passed. Basically
		124	this is a subsecond-resolution C<sleep ()>.
57		125
58	=item int ev_version_major ()	126	=item int ev_version_major ()
59		127
60	=item int ev_version_minor ()	128	=item int ev_version_minor ()
61		129
62	You can find out the major and minor version numbers of the library	130	You can find out the major and minor ABI version numbers of the library
63	you linked against by calling the functions C<ev_version_major> and	131	you linked against by calling the functions C<ev_version_major> and
64	C<ev_version_minor>. If you want, you can compare against the global	132	C<ev_version_minor>. If you want, you can compare against the global
65	symbols C<EV_VERSION_MAJOR> and C<EV_VERSION_MINOR>, which specify the	133	symbols C<EV_VERSION_MAJOR> and C<EV_VERSION_MINOR>, which specify the
66	version of the library your program was compiled against.	134	version of the library your program was compiled against.
67		135
		136	These version numbers refer to the ABI version of the library, not the
		137	release version.
		138
68	Usually, its a good idea to terminate if the major versions mismatch,	139	Usually, it's a good idea to terminate if the major versions mismatch,
69	as this indicates an incompatible change. Minor versions are usually	140	as this indicates an incompatible change. Minor versions are usually
70	compatible to older versions, so a larger minor version alone is usually	141	compatible to older versions, so a larger minor version alone is usually
71	not a problem.	142	not a problem.
72		143
		144	Example: Make sure we haven't accidentally been linked against the wrong
		145	version.
		146
		147	assert (("libev version mismatch",
		148	ev_version_major () == EV_VERSION_MAJOR
		149	&& ev_version_minor () >= EV_VERSION_MINOR));
		150
		151	=item unsigned int ev_supported_backends ()
		152
		153	Return the set of all backends (i.e. their corresponding C<EV_BACKEND_*>
		154	value) compiled into this binary of libev (independent of their
		155	availability on the system you are running on). See C<ev_default_loop> for
		156	a description of the set values.
		157
		158	Example: make sure we have the epoll method, because yeah this is cool and
		159	a must have and can we have a torrent of it please!!!11
		160
		161	assert (("sorry, no epoll, no sex",
		162	ev_supported_backends () & EVBACKEND_EPOLL));
		163
		164	=item unsigned int ev_recommended_backends ()
		165
		166	Return the set of all backends compiled into this binary of libev and also
		167	recommended for this platform. This set is often smaller than the one
		168	returned by C<ev_supported_backends>, as for example kqueue is broken on
		169	most BSDs and will not be autodetected unless you explicitly request it
		170	(assuming you know what you are doing). This is the set of backends that
		171	libev will probe for if you specify no backends explicitly.
		172
		173	=item unsigned int ev_embeddable_backends ()
		174
		175	Returns the set of backends that are embeddable in other event loops. This
		176	is the theoretical, all-platform, value. To find which backends
		177	might be supported on the current system, you would need to look at
		178	C<ev_embeddable_backends () & ev_supported_backends ()>, likewise for
		179	recommended ones.
		180
		181	See the description of C<ev_embed> watchers for more info.
		182
73	=item ev_set_allocator (void *(*cb)(void *ptr, long size))	183	=item ev_set_allocator (void *(*cb)(void *ptr, long size))
74		184
75	Sets the allocation function to use (the prototype is similar to the	185	Sets the allocation function to use (the prototype is similar - the
76	realloc C function, the semantics are identical). It is used to allocate	186	semantics is identical - to the realloc C function). It is used to
77	and free memory (no surprises here). If it returns zero when memory	187	allocate and free memory (no surprises here). If it returns zero when
78	needs to be allocated, the library might abort or take some potentially	188	memory needs to be allocated, the library might abort or take some
79	destructive action. The default is your system realloc function.	189	potentially destructive action. The default is your system realloc
		190	function.
80		191
81	You could override this function in high-availability programs to, say,	192	You could override this function in high-availability programs to, say,
82	free some memory if it cannot allocate memory, to use a special allocator,	193	free some memory if it cannot allocate memory, to use a special allocator,
83	or even to sleep a while and retry until some memory is available.	194	or even to sleep a while and retry until some memory is available.
		195
		196	Example: Replace the libev allocator with one that waits a bit and then
		197	retries).
		198
		199	static void *
		200	persistent_realloc (void *ptr, size_t size)
		201	{
		202	for (;;)
		203	{
		204	void *newptr = realloc (ptr, size);
		205
		206	if (newptr)
		207	return newptr;
		208
		209	sleep (60);
		210	}
		211	}
		212
		213	...
		214	ev_set_allocator (persistent_realloc);
84		215
85	=item ev_set_syserr_cb (void (*cb)(const char *msg));	216	=item ev_set_syserr_cb (void (*cb)(const char *msg));
86		217
87	Set the callback function to call on a retryable syscall error (such	218	Set the callback function to call on a retryable syscall error (such
88	as failed select, poll, epoll_wait). The message is a printable string	219	as failed select, poll, epoll_wait). The message is a printable string
…		…
90	callback is set, then libev will expect it to remedy the sitution, no	221	callback is set, then libev will expect it to remedy the sitution, no
91	matter what, when it returns. That is, libev will generally retry the	222	matter what, when it returns. That is, libev will generally retry the
92	requested operation, or, if the condition doesn't go away, do bad stuff	223	requested operation, or, if the condition doesn't go away, do bad stuff
93	(such as abort).	224	(such as abort).
94		225
		226	Example: This is basically the same thing that libev does internally, too.
		227
		228	static void
		229	fatal_error (const char *msg)
		230	{
		231	perror (msg);
		232	abort ();
		233	}
		234
		235	...
		236	ev_set_syserr_cb (fatal_error);
		237
95	=back	238	=back
96		239
97	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP	240	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP
98		241
99	An event loop is described by a C<struct ev_loop *>. The library knows two	242	An event loop is described by a C<struct ev_loop *>. The library knows two
100	types of such loops, the I<default> loop, which supports signals and child	243	types of such loops, the I<default> loop, which supports signals and child
101	events, and dynamically created loops which do not.	244	events, and dynamically created loops which do not.
102		245
103	If you use threads, a common model is to run the default event loop	246	If you use threads, a common model is to run the default event loop
104	in your main thread (or in a separate thrad) and for each thread you	247	in your main thread (or in a separate thread) and for each thread you
105	create, you also create another event loop. Libev itself does no locking	248	create, you also create another event loop. Libev itself does no locking
106	whatsoever, so if you mix calls to the same event loop in different	249	whatsoever, so if you mix calls to the same event loop in different
107	threads, make sure you lock (this is usually a bad idea, though, even if	250	threads, make sure you lock (this is usually a bad idea, though, even if
108	done correctly, because its hideous and inefficient).	251	done correctly, because it's hideous and inefficient).
109		252
110	=over 4	253	=over 4
111		254
112	=item struct ev_loop *ev_default_loop (unsigned int flags)	255	=item struct ev_loop *ev_default_loop (unsigned int flags)
113		256
114	This will initialise the default event loop if it hasn't been initialised	257	This will initialise the default event loop if it hasn't been initialised
115	yet and return it. If the default loop could not be initialised, returns	258	yet and return it. If the default loop could not be initialised, returns
116	false. If it already was initialised it simply returns it (and ignores the	259	false. If it already was initialised it simply returns it (and ignores the
117	flags).	260	flags. If that is troubling you, check C<ev_backend ()> afterwards).
118		261
119	If you don't know what event loop to use, use the one returned from this	262	If you don't know what event loop to use, use the one returned from this
120	function.	263	function.
121		264
122	The flags argument can be used to specify special behaviour or specific	265	The flags argument can be used to specify special behaviour or specific
123	backends to use, and is usually specified as 0 (or EVFLAG_AUTO).	266	backends to use, and is usually specified as C<0> (or C<EVFLAG_AUTO>).
124		267
125	It supports the following flags:	268	The following flags are supported:
126		269
127	=over 4	270	=over 4
128		271
129	=item EVFLAG_AUTO	272	=item C<EVFLAG_AUTO>
130		273
131	The default flags value. Use this if you have no clue (its the right	274	The default flags value. Use this if you have no clue (it's the right
132	thing, believe me).	275	thing, believe me).
133		276
134	=item EVFLAG_NOENV	277	=item C<EVFLAG_NOENV>
135		278
136	If this flag bit is ored into the flag value (or the program runs setuid	279	If this flag bit is ored into the flag value (or the program runs setuid
137	or setgid) then libev will I<not> look at the environment variable	280	or setgid) then libev will I<not> look at the environment variable
138	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will	281	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will
139	override the flags completely if it is found in the environment. This is	282	override the flags completely if it is found in the environment. This is
140	useful to try out specific backends to test their performance, or to work	283	useful to try out specific backends to test their performance, or to work
141	around bugs.	284	around bugs.
142		285
		286	=item C<EVFLAG_FORKCHECK>
		287
		288	Instead of calling C<ev_default_fork> or C<ev_loop_fork> manually after
		289	a fork, you can also make libev check for a fork in each iteration by
		290	enabling this flag.
		291
		292	This works by calling C<getpid ()> on every iteration of the loop,
		293	and thus this might slow down your event loop if you do a lot of loop
		294	iterations and little real work, but is usually not noticeable (on my
		295	Linux system for example, C<getpid> is actually a simple 5-insn sequence
		296	without a syscall and thus I<very> fast, but my Linux system also has
		297	C<pthread_atfork> which is even faster).
		298
		299	The big advantage of this flag is that you can forget about fork (and
		300	forget about forgetting to tell libev about forking) when you use this
		301	flag.
		302
		303	This flag setting cannot be overriden or specified in the C<LIBEV_FLAGS>
		304	environment variable.
		305
143	=item EVMETHOD_SELECT portable select backend	306	=item C<EVBACKEND_SELECT> (value 1, portable select backend)
144		307
145	=item EVMETHOD_POLL poll backend (everywhere except windows)	308	This is your standard select(2) backend. Not I<completely> standard, as
		309	libev tries to roll its own fd_set with no limits on the number of fds,
		310	but if that fails, expect a fairly low limit on the number of fds when
		311	using this backend. It doesn't scale too well (O(highest_fd)), but its
		312	usually the fastest backend for a low number of (low-numbered :) fds.
146		313
147	=item EVMETHOD_EPOLL linux only	314	To get good performance out of this backend you need a high amount of
		315	parallelity (most of the file descriptors should be busy). If you are
		316	writing a server, you should C<accept ()> in a loop to accept as many
		317	connections as possible during one iteration. You might also want to have
		318	a look at C<ev_set_io_collect_interval ()> to increase the amount of
		319	readyness notifications you get per iteration.
148		320
149	=item EVMETHOD_KQUEUE some bsds only	321	=item C<EVBACKEND_POLL> (value 2, poll backend, available everywhere except on windows)
150		322
151	=item EVMETHOD_DEVPOLL solaris 8 only	323	And this is your standard poll(2) backend. It's more complicated
		324	than select, but handles sparse fds better and has no artificial
		325	limit on the number of fds you can use (except it will slow down
		326	considerably with a lot of inactive fds). It scales similarly to select,
		327	i.e. O(total_fds). See the entry for C<EVBACKEND_SELECT>, above, for
		328	performance tips.
152		329
153	=item EVMETHOD_PORT solaris 10 only	330	=item C<EVBACKEND_EPOLL> (value 4, Linux)
		331
		332	For few fds, this backend is a bit little slower than poll and select,
		333	but it scales phenomenally better. While poll and select usually scale
		334	like O(total_fds) where n is the total number of fds (or the highest fd),
		335	epoll scales either O(1) or O(active_fds). The epoll design has a number
		336	of shortcomings, such as silently dropping events in some hard-to-detect
		337	cases and rewiring a syscall per fd change, no fork support and bad
		338	support for dup.
		339
		340	While stopping, setting and starting an I/O watcher in the same iteration
		341	will result in some caching, there is still a syscall per such incident
		342	(because the fd could point to a different file description now), so its
		343	best to avoid that. Also, C<dup ()>'ed file descriptors might not work
		344	very well if you register events for both fds.
		345
		346	Please note that epoll sometimes generates spurious notifications, so you
		347	need to use non-blocking I/O or other means to avoid blocking when no data
		348	(or space) is available.
		349
		350	Best performance from this backend is achieved by not unregistering all
		351	watchers for a file descriptor until it has been closed, if possible, i.e.
		352	keep at least one watcher active per fd at all times.
		353
		354	While nominally embeddeble in other event loops, this feature is broken in
		355	all kernel versions tested so far.
		356
		357	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)
		358
		359	Kqueue deserves special mention, as at the time of this writing, it
		360	was broken on all BSDs except NetBSD (usually it doesn't work reliably
		361	with anything but sockets and pipes, except on Darwin, where of course
		362	it's completely useless). For this reason it's not being "autodetected"
		363	unless you explicitly specify it explicitly in the flags (i.e. using
		364	C<EVBACKEND_KQUEUE>) or libev was compiled on a known-to-be-good (-enough)
		365	system like NetBSD.
		366
		367	You still can embed kqueue into a normal poll or select backend and use it
		368	only for sockets (after having made sure that sockets work with kqueue on
		369	the target platform). See C<ev_embed> watchers for more info.
		370
		371	It scales in the same way as the epoll backend, but the interface to the
		372	kernel is more efficient (which says nothing about its actual speed, of
		373	course). While stopping, setting and starting an I/O watcher does never
		374	cause an extra syscall as with C<EVBACKEND_EPOLL>, it still adds up to
		375	two event changes per incident, support for C<fork ()> is very bad and it
		376	drops fds silently in similarly hard-to-detect cases.
		377
		378	This backend usually performs well under most conditions.
		379
		380	While nominally embeddable in other event loops, this doesn't work
		381	everywhere, so you might need to test for this. And since it is broken
		382	almost everywhere, you should only use it when you have a lot of sockets
		383	(for which it usually works), by embedding it into another event loop
		384	(e.g. C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>) and using it only for
		385	sockets.
		386
		387	=item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8)
		388
		389	This is not implemented yet (and might never be, unless you send me an
		390	implementation). According to reports, C</dev/poll> only supports sockets
		391	and is not embeddable, which would limit the usefulness of this backend
		392	immensely.
		393
		394	=item C<EVBACKEND_PORT> (value 32, Solaris 10)
		395
		396	This uses the Solaris 10 event port mechanism. As with everything on Solaris,
		397	it's really slow, but it still scales very well (O(active_fds)).
		398
		399	Please note that solaris event ports can deliver a lot of spurious
		400	notifications, so you need to use non-blocking I/O or other means to avoid
		401	blocking when no data (or space) is available.
		402
		403	While this backend scales well, it requires one system call per active
		404	file descriptor per loop iteration. For small and medium numbers of file
		405	descriptors a "slow" C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> backend
		406	might perform better.
		407
		408	=item C<EVBACKEND_ALL>
		409
		410	Try all backends (even potentially broken ones that wouldn't be tried
		411	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as
		412	C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>.
		413
		414	It is definitely not recommended to use this flag.
		415
		416	=back
154		417
155	If one or more of these are ored into the flags value, then only these	418	If one or more of these are ored into the flags value, then only these
156	backends will be tried (in the reverse order as given here). If one are	419	backends will be tried (in the reverse order as given here). If none are
157	specified, any backend will do.	420	specified, most compiled-in backend will be tried, usually in reverse
		421	order of their flag values :)
158		422
159	=back	423	The most typical usage is like this:
		424
		425	if (!ev_default_loop (0))
		426	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
		427
		428	Restrict libev to the select and poll backends, and do not allow
		429	environment settings to be taken into account:
		430
		431	ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);
		432
		433	Use whatever libev has to offer, but make sure that kqueue is used if
		434	available (warning, breaks stuff, best use only with your own private
		435	event loop and only if you know the OS supports your types of fds):
		436
		437	ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);
160		438
161	=item struct ev_loop *ev_loop_new (unsigned int flags)	439	=item struct ev_loop *ev_loop_new (unsigned int flags)
162		440
163	Similar to C<ev_default_loop>, but always creates a new event loop that is	441	Similar to C<ev_default_loop>, but always creates a new event loop that is
164	always distinct from the default loop. Unlike the default loop, it cannot	442	always distinct from the default loop. Unlike the default loop, it cannot
165	handle signal and child watchers, and attempts to do so will be greeted by	443	handle signal and child watchers, and attempts to do so will be greeted by
166	undefined behaviour (or a failed assertion if assertions are enabled).	444	undefined behaviour (or a failed assertion if assertions are enabled).
167		445
		446	Example: Try to create a event loop that uses epoll and nothing else.
		447
		448	struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);
		449	if (!epoller)
		450	fatal ("no epoll found here, maybe it hides under your chair");
		451
168	=item ev_default_destroy ()	452	=item ev_default_destroy ()
169		453
170	Destroys the default loop again (frees all memory and kernel state	454	Destroys the default loop again (frees all memory and kernel state
171	etc.). This stops all registered event watchers (by not touching them in	455	etc.). None of the active event watchers will be stopped in the normal
172	any way whatsoever, although you cnanot rely on this :).	456	sense, so e.g. C<ev_is_active> might still return true. It is your
		457	responsibility to either stop all watchers cleanly yoursef I<before>
		458	calling this function, or cope with the fact afterwards (which is usually
		459	the easiest thing, you can just ignore the watchers and/or C<free ()> them
		460	for example).
		461
		462	Note that certain global state, such as signal state, will not be freed by
		463	this function, and related watchers (such as signal and child watchers)
		464	would need to be stopped manually.
		465
		466	In general it is not advisable to call this function except in the
		467	rare occasion where you really need to free e.g. the signal handling
		468	pipe fds. If you need dynamically allocated loops it is better to use
		469	C<ev_loop_new> and C<ev_loop_destroy>).
173		470
174	=item ev_loop_destroy (loop)	471	=item ev_loop_destroy (loop)
175		472
176	Like C<ev_default_destroy>, but destroys an event loop created by an	473	Like C<ev_default_destroy>, but destroys an event loop created by an
177	earlier call to C<ev_loop_new>.	474	earlier call to C<ev_loop_new>.
…		…
181	This function reinitialises the kernel state for backends that have	478	This function reinitialises the kernel state for backends that have
182	one. Despite the name, you can call it anytime, but it makes most sense	479	one. Despite the name, you can call it anytime, but it makes most sense
183	after forking, in either the parent or child process (or both, but that	480	after forking, in either the parent or child process (or both, but that
184	again makes little sense).	481	again makes little sense).
185		482
186	You I<must> call this function after forking if and only if you want to	483	You I<must> call this function in the child process after forking if and
187	use the event library in both processes. If you just fork+exec, you don't	484	only if you want to use the event library in both processes. If you just
188	have to call it.	485	fork+exec, you don't have to call it.
189		486
190	The function itself is quite fast and its usually not a problem to call	487	The function itself is quite fast and it's usually not a problem to call
191	it just in case after a fork. To make this easy, the function will fit in	488	it just in case after a fork. To make this easy, the function will fit in
192	quite nicely into a call to C<pthread_atfork>:	489	quite nicely into a call to C<pthread_atfork>:
193		490
194	pthread_atfork (0, 0, ev_default_fork);	491	pthread_atfork (0, 0, ev_default_fork);
		492
		493	At the moment, C<EVBACKEND_SELECT> and C<EVBACKEND_POLL> are safe to use
		494	without calling this function, so if you force one of those backends you
		495	do not need to care.
195		496
196	=item ev_loop_fork (loop)	497	=item ev_loop_fork (loop)
197		498
198	Like C<ev_default_fork>, but acts on an event loop created by	499	Like C<ev_default_fork>, but acts on an event loop created by
199	C<ev_loop_new>. Yes, you have to call this on every allocated event loop	500	C<ev_loop_new>. Yes, you have to call this on every allocated event loop
200	after fork, and how you do this is entirely your own problem.	501	after fork, and how you do this is entirely your own problem.
201		502
		503	=item unsigned int ev_loop_count (loop)
		504
		505	Returns the count of loop iterations for the loop, which is identical to
		506	the number of times libev did poll for new events. It starts at C<0> and
		507	happily wraps around with enough iterations.
		508
		509	This value can sometimes be useful as a generation counter of sorts (it
		510	"ticks" the number of loop iterations), as it roughly corresponds with
		511	C<ev_prepare> and C<ev_check> calls.
		512
202	=item unsigned int ev_method (loop)	513	=item unsigned int ev_backend (loop)
203		514
204	Returns one of the C<EVMETHOD_*> flags indicating the event backend in	515	Returns one of the C<EVBACKEND_*> flags indicating the event backend in
205	use.	516	use.
206		517
207	=item ev_tstamp = ev_now (loop)	518	=item ev_tstamp ev_now (loop)
208		519
209	Returns the current "event loop time", which is the time the event loop	520	Returns the current "event loop time", which is the time the event loop
210	got events and started processing them. This timestamp does not change	521	received events and started processing them. This timestamp does not
211	as long as callbacks are being processed, and this is also the base time	522	change as long as callbacks are being processed, and this is also the base
212	used for relative timers. You can treat it as the timestamp of the event	523	time used for relative timers. You can treat it as the timestamp of the
213	occuring (or more correctly, the mainloop finding out about it).	524	event occurring (or more correctly, libev finding out about it).
214		525
215	=item ev_loop (loop, int flags)	526	=item ev_loop (loop, int flags)
216		527
217	Finally, this is it, the event handler. This function usually is called	528	Finally, this is it, the event handler. This function usually is called
218	after you initialised all your watchers and you want to start handling	529	after you initialised all your watchers and you want to start handling
219	events.	530	events.
220		531
221	If the flags argument is specified as 0, it will not return until either	532	If the flags argument is specified as C<0>, it will not return until
222	no event watchers are active anymore or C<ev_unloop> was called.	533	either no event watchers are active anymore or C<ev_unloop> was called.
		534
		535	Please note that an explicit C<ev_unloop> is usually better than
		536	relying on all watchers to be stopped when deciding when a program has
		537	finished (especially in interactive programs), but having a program that
		538	automatically loops as long as it has to and no longer by virtue of
		539	relying on its watchers stopping correctly is a thing of beauty.
223		540
224	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle	541	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle
225	those events and any outstanding ones, but will not block your process in	542	those events and any outstanding ones, but will not block your process in
226	case there are no events.	543	case there are no events and will return after one iteration of the loop.
227		544
228	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if	545	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if
229	neccessary) and will handle those and any outstanding ones. It will block	546	neccessary) and will handle those and any outstanding ones. It will block
230	your process until at least one new event arrives.	547	your process until at least one new event arrives, and will return after
		548	one iteration of the loop. This is useful if you are waiting for some
		549	external event in conjunction with something not expressible using other
		550	libev watchers. However, a pair of C<ev_prepare>/C<ev_check> watchers is
		551	usually a better approach for this kind of thing.
231		552
232	This flags value could be used to implement alternative looping	553	Here are the gory details of what C<ev_loop> does:
233	constructs, but the C<prepare> and C<check> watchers provide a better and	554
234	more generic mechanism.	555	- Before the first iteration, call any pending watchers.
		556	* If EVFLAG_FORKCHECK was used, check for a fork.
		557	- If a fork was detected, queue and call all fork watchers.
		558	- Queue and call all prepare watchers.
		559	- If we have been forked, recreate the kernel state.
		560	- Update the kernel state with all outstanding changes.
		561	- Update the "event loop time".
		562	- Calculate for how long to sleep or block, if at all
		563	(active idle watchers, EVLOOP_NONBLOCK or not having
		564	any active watchers at all will result in not sleeping).
		565	- Sleep if the I/O and timer collect interval say so.
		566	- Block the process, waiting for any events.
		567	- Queue all outstanding I/O (fd) events.
		568	- Update the "event loop time" and do time jump handling.
		569	- Queue all outstanding timers.
		570	- Queue all outstanding periodics.
		571	- If no events are pending now, queue all idle watchers.
		572	- Queue all check watchers.
		573	- Call all queued watchers in reverse order (i.e. check watchers first).
		574	Signals and child watchers are implemented as I/O watchers, and will
		575	be handled here by queueing them when their watcher gets executed.
		576	- If ev_unloop has been called, or EVLOOP_ONESHOT or EVLOOP_NONBLOCK
		577	were used, or there are no active watchers, return, otherwise
		578	continue with step *.
		579
		580	Example: Queue some jobs and then loop until no events are outstanding
		581	anymore.
		582
		583	... queue jobs here, make sure they register event watchers as long
		584	... as they still have work to do (even an idle watcher will do..)
		585	ev_loop (my_loop, 0);
		586	... jobs done. yeah!
235		587
236	=item ev_unloop (loop, how)	588	=item ev_unloop (loop, how)
237		589
238	Can be used to make a call to C<ev_loop> return early. The C<how> argument	590	Can be used to make a call to C<ev_loop> return early (but only after it
		591	has processed all outstanding events). The C<how> argument must be either
239	must be either C<EVUNLOOP_ONCE>, which will make the innermost C<ev_loop>	592	C<EVUNLOOP_ONE>, which will make the innermost C<ev_loop> call return, or
240	call return, or C<EVUNLOOP_ALL>, which will make all nested C<ev_loop>	593	C<EVUNLOOP_ALL>, which will make all nested C<ev_loop> calls return.
241	calls return.
242		594
243	=item ev_ref (loop)	595	=item ev_ref (loop)
244		596
245	=item ev_unref (loop)	597	=item ev_unref (loop)
246		598
247	Ref/unref can be used to add or remove a refcount on the event loop: Every	599	Ref/unref can be used to add or remove a reference count on the event
248	watcher keeps one reference. If you have a long-runing watcher you never	600	loop: Every watcher keeps one reference, and as long as the reference
249	unregister that should not keep ev_loop from running, ev_unref() after	601	count is nonzero, C<ev_loop> will not return on its own. If you have
250	starting, and ev_ref() before stopping it. Libev itself uses this for	602	a watcher you never unregister that should not keep C<ev_loop> from
251	example for its internal signal pipe: It is not visible to you as a user	603	returning, ev_unref() after starting, and ev_ref() before stopping it. For
252	and should not keep C<ev_loop> from exiting if the work is done. It is	604	example, libev itself uses this for its internal signal pipe: It is not
253	also an excellent way to do this for generic recurring timers or from	605	visible to the libev user and should not keep C<ev_loop> from exiting if
254	within third-party libraries. Just remember to unref after start and ref	606	no event watchers registered by it are active. It is also an excellent
255	before stop.	607	way to do this for generic recurring timers or from within third-party
		608	libraries. Just remember to I<unref after start> and I<ref before stop>.
		609
		610	Example: Create a signal watcher, but keep it from keeping C<ev_loop>
		611	running when nothing else is active.
		612
		613	struct ev_signal exitsig;
		614	ev_signal_init (&exitsig, sig_cb, SIGINT);
		615	ev_signal_start (loop, &exitsig);
		616	evf_unref (loop);
		617
		618	Example: For some weird reason, unregister the above signal handler again.
		619
		620	ev_ref (loop);
		621	ev_signal_stop (loop, &exitsig);
		622
		623	=item ev_set_io_collect_interval (loop, ev_tstamp interval)
		624
		625	=item ev_set_timeout_collect_interval (loop, ev_tstamp interval)
		626
		627	These advanced functions influence the time that libev will spend waiting
		628	for events. Both are by default C<0>, meaning that libev will try to
		629	invoke timer/periodic callbacks and I/O callbacks with minimum latency.
		630
		631	Setting these to a higher value (the C<interval> I<must> be >= C<0>)
		632	allows libev to delay invocation of I/O and timer/periodic callbacks to
		633	increase efficiency of loop iterations.
		634
		635	The background is that sometimes your program runs just fast enough to
		636	handle one (or very few) event(s) per loop iteration. While this makes
		637	the program responsive, it also wastes a lot of CPU time to poll for new
		638	events, especially with backends like C<select ()> which have a high
		639	overhead for the actual polling but can deliver many events at once.
		640
		641	By setting a higher I<io collect interval> you allow libev to spend more
		642	time collecting I/O events, so you can handle more events per iteration,
		643	at the cost of increasing latency. Timeouts (both C<ev_periodic> and
		644	C<ev_timer>) will be not affected. Setting this to a non-null value will
		645	introduce an additional C<ev_sleep ()> call into most loop iterations.
		646
		647	Likewise, by setting a higher I<timeout collect interval> you allow libev
		648	to spend more time collecting timeouts, at the expense of increased
		649	latency (the watcher callback will be called later). C<ev_io> watchers
		650	will not be affected. Setting this to a non-null value will not introduce
		651	any overhead in libev.
		652
		653	Many (busy) programs can usually benefit by setting the io collect
		654	interval to a value near C<0.1> or so, which is often enough for
		655	interactive servers (of course not for games), likewise for timeouts. It
		656	usually doesn't make much sense to set it to a lower value than C<0.01>,
		657	as this approsaches the timing granularity of most systems.
256		658
257	=back	659	=back
		660
258		661
259	=head1 ANATOMY OF A WATCHER	662	=head1 ANATOMY OF A WATCHER
260		663
261	A watcher is a structure that you create and register to record your	664	A watcher is a structure that you create and register to record your
262	interest in some event. For instance, if you want to wait for STDIN to	665	interest in some event. For instance, if you want to wait for STDIN to
263	become readable, you would create an ev_io watcher for that:	666	become readable, you would create an C<ev_io> watcher for that:
264		667
265	static void my_cb (struct ev_loop *loop, struct ev_io *w, int revents)	668	static void my_cb (struct ev_loop *loop, struct ev_io *w, int revents)
266	{	669	{
267	ev_io_stop (w);	670	ev_io_stop (w);
268	ev_unloop (loop, EVUNLOOP_ALL);	671	ev_unloop (loop, EVUNLOOP_ALL);
…		…
295	*) >>), and you can stop watching for events at any time by calling the	698	*) >>), and you can stop watching for events at any time by calling the
296	corresponding stop function (C<< ev_<type>_stop (loop, watcher *) >>.	699	corresponding stop function (C<< ev_<type>_stop (loop, watcher *) >>.
297		700
298	As long as your watcher is active (has been started but not stopped) you	701	As long as your watcher is active (has been started but not stopped) you
299	must not touch the values stored in it. Most specifically you must never	702	must not touch the values stored in it. Most specifically you must never
300	reinitialise it or call its set method.	703	reinitialise it or call its C<set> macro.
301
302	You cna check whether an event is active by calling the C<ev_is_active
303	(watcher *)> macro. To see whether an event is outstanding (but the
304	callback for it has not been called yet) you cna use the C<ev_is_pending
305	(watcher *)> macro.
306		704
307	Each and every callback receives the event loop pointer as first, the	705	Each and every callback receives the event loop pointer as first, the
308	registered watcher structure as second, and a bitset of received events as	706	registered watcher structure as second, and a bitset of received events as
309	third argument.	707	third argument.
310		708
311	The rceeived events usually include a single bit per event type received	709	The received events usually include a single bit per event type received
312	(you can receive multiple events at the same time). The possible bit masks	710	(you can receive multiple events at the same time). The possible bit masks
313	are:	711	are:
314		712
315	=over 4	713	=over 4
316		714
317	=item EV_READ	715	=item C<EV_READ>
318		716
319	=item EV_WRITE	717	=item C<EV_WRITE>
320		718
321	The file descriptor in the ev_io watcher has become readable and/or	719	The file descriptor in the C<ev_io> watcher has become readable and/or
322	writable.	720	writable.
323		721
324	=item EV_TIMEOUT	722	=item C<EV_TIMEOUT>
325		723
326	The ev_timer watcher has timed out.	724	The C<ev_timer> watcher has timed out.
327		725
328	=item EV_PERIODIC	726	=item C<EV_PERIODIC>
329		727
330	The ev_periodic watcher has timed out.	728	The C<ev_periodic> watcher has timed out.
331		729
332	=item EV_SIGNAL	730	=item C<EV_SIGNAL>
333		731
334	The signal specified in the ev_signal watcher has been received by a thread.	732	The signal specified in the C<ev_signal> watcher has been received by a thread.
335		733
336	=item EV_CHILD	734	=item C<EV_CHILD>
337		735
338	The pid specified in the ev_child watcher has received a status change.	736	The pid specified in the C<ev_child> watcher has received a status change.
339		737
		738	=item C<EV_STAT>
		739
		740	The path specified in the C<ev_stat> watcher changed its attributes somehow.
		741
340	=item EV_IDLE	742	=item C<EV_IDLE>
341		743
342	The ev_idle watcher has determined that you have nothing better to do.	744	The C<ev_idle> watcher has determined that you have nothing better to do.
343		745
344	=item EV_PREPARE	746	=item C<EV_PREPARE>
345		747
346	=item EV_CHECK	748	=item C<EV_CHECK>
347		749
348	All ev_prepare watchers are invoked just I<before> C<ev_loop> starts	750	All C<ev_prepare> watchers are invoked just I<before> C<ev_loop> starts
349	to gather new events, and all ev_check watchers are invoked just after	751	to gather new events, and all C<ev_check> watchers are invoked just after
350	C<ev_loop> has gathered them, but before it invokes any callbacks for any	752	C<ev_loop> has gathered them, but before it invokes any callbacks for any
351	received events. Callbacks of both watcher types can start and stop as	753	received events. Callbacks of both watcher types can start and stop as
352	many watchers as they want, and all of them will be taken into account	754	many watchers as they want, and all of them will be taken into account
353	(for example, a ev_prepare watcher might start an idle watcher to keep	755	(for example, a C<ev_prepare> watcher might start an idle watcher to keep
354	C<ev_loop> from blocking).	756	C<ev_loop> from blocking).
355		757
		758	=item C<EV_EMBED>
		759
		760	The embedded event loop specified in the C<ev_embed> watcher needs attention.
		761
		762	=item C<EV_FORK>
		763
		764	The event loop has been resumed in the child process after fork (see
		765	C<ev_fork>).
		766
356	=item EV_ERROR	767	=item C<EV_ERROR>
357		768
358	An unspecified error has occured, the watcher has been stopped. This might	769	An unspecified error has occured, the watcher has been stopped. This might
359	happen because the watcher could not be properly started because libev	770	happen because the watcher could not be properly started because libev
360	ran out of memory, a file descriptor was found to be closed or any other	771	ran out of memory, a file descriptor was found to be closed or any other
361	problem. You best act on it by reporting the problem and somehow coping	772	problem. You best act on it by reporting the problem and somehow coping
…		…
367	with the error from read() or write(). This will not work in multithreaded	778	with the error from read() or write(). This will not work in multithreaded
368	programs, though, so beware.	779	programs, though, so beware.
369		780
370	=back	781	=back
371		782
		783	=head2 GENERIC WATCHER FUNCTIONS
		784
		785	In the following description, C<TYPE> stands for the watcher type,
		786	e.g. C<timer> for C<ev_timer> watchers and C<io> for C<ev_io> watchers.
		787
		788	=over 4
		789
		790	=item C<ev_init> (ev_TYPE *watcher, callback)
		791
		792	This macro initialises the generic portion of a watcher. The contents
		793	of the watcher object can be arbitrary (so C<malloc> will do). Only
		794	the generic parts of the watcher are initialised, you I<need> to call
		795	the type-specific C<ev_TYPE_set> macro afterwards to initialise the
		796	type-specific parts. For each type there is also a C<ev_TYPE_init> macro
		797	which rolls both calls into one.
		798
		799	You can reinitialise a watcher at any time as long as it has been stopped
		800	(or never started) and there are no pending events outstanding.
		801
		802	The callback is always of type C<void (*)(ev_loop *loop, ev_TYPE *watcher,
		803	int revents)>.
		804
		805	=item C<ev_TYPE_set> (ev_TYPE *, [args])
		806
		807	This macro initialises the type-specific parts of a watcher. You need to
		808	call C<ev_init> at least once before you call this macro, but you can
		809	call C<ev_TYPE_set> any number of times. You must not, however, call this
		810	macro on a watcher that is active (it can be pending, however, which is a
		811	difference to the C<ev_init> macro).
		812
		813	Although some watcher types do not have type-specific arguments
		814	(e.g. C<ev_prepare>) you still need to call its C<set> macro.
		815
		816	=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])
		817
		818	This convinience macro rolls both C<ev_init> and C<ev_TYPE_set> macro
		819	calls into a single call. This is the most convinient method to initialise
		820	a watcher. The same limitations apply, of course.
		821
		822	=item C<ev_TYPE_start> (loop *, ev_TYPE *watcher)
		823
		824	Starts (activates) the given watcher. Only active watchers will receive
		825	events. If the watcher is already active nothing will happen.
		826
		827	=item C<ev_TYPE_stop> (loop *, ev_TYPE *watcher)
		828
		829	Stops the given watcher again (if active) and clears the pending
		830	status. It is possible that stopped watchers are pending (for example,
		831	non-repeating timers are being stopped when they become pending), but
		832	C<ev_TYPE_stop> ensures that the watcher is neither active nor pending. If
		833	you want to free or reuse the memory used by the watcher it is therefore a
		834	good idea to always call its C<ev_TYPE_stop> function.
		835
		836	=item bool ev_is_active (ev_TYPE *watcher)
		837
		838	Returns a true value iff the watcher is active (i.e. it has been started
		839	and not yet been stopped). As long as a watcher is active you must not modify
		840	it.
		841
		842	=item bool ev_is_pending (ev_TYPE *watcher)
		843
		844	Returns a true value iff the watcher is pending, (i.e. it has outstanding
		845	events but its callback has not yet been invoked). As long as a watcher
		846	is pending (but not active) you must not call an init function on it (but
		847	C<ev_TYPE_set> is safe), you must not change its priority, and you must
		848	make sure the watcher is available to libev (e.g. you cannot C<free ()>
		849	it).
		850
		851	=item callback ev_cb (ev_TYPE *watcher)
		852
		853	Returns the callback currently set on the watcher.
		854
		855	=item ev_cb_set (ev_TYPE *watcher, callback)
		856
		857	Change the callback. You can change the callback at virtually any time
		858	(modulo threads).
		859
		860	=item ev_set_priority (ev_TYPE *watcher, priority)
		861
		862	=item int ev_priority (ev_TYPE *watcher)
		863
		864	Set and query the priority of the watcher. The priority is a small
		865	integer between C<EV_MAXPRI> (default: C<2>) and C<EV_MINPRI>
		866	(default: C<-2>). Pending watchers with higher priority will be invoked
		867	before watchers with lower priority, but priority will not keep watchers
		868	from being executed (except for C<ev_idle> watchers).
		869
		870	This means that priorities are I<only> used for ordering callback
		871	invocation after new events have been received. This is useful, for
		872	example, to reduce latency after idling, or more often, to bind two
		873	watchers on the same event and make sure one is called first.
		874
		875	If you need to suppress invocation when higher priority events are pending
		876	you need to look at C<ev_idle> watchers, which provide this functionality.
		877
		878	You I<must not> change the priority of a watcher as long as it is active or
		879	pending.
		880
		881	The default priority used by watchers when no priority has been set is
		882	always C<0>, which is supposed to not be too high and not be too low :).
		883
		884	Setting a priority outside the range of C<EV_MINPRI> to C<EV_MAXPRI> is
		885	fine, as long as you do not mind that the priority value you query might
		886	or might not have been adjusted to be within valid range.
		887
		888	=item ev_invoke (loop, ev_TYPE *watcher, int revents)
		889
		890	Invoke the C<watcher> with the given C<loop> and C<revents>. Neither
		891	C<loop> nor C<revents> need to be valid as long as the watcher callback
		892	can deal with that fact.
		893
		894	=item int ev_clear_pending (loop, ev_TYPE *watcher)
		895
		896	If the watcher is pending, this function returns clears its pending status
		897	and returns its C<revents> bitset (as if its callback was invoked). If the
		898	watcher isn't pending it does nothing and returns C<0>.
		899
		900	=back
		901
		902
372	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER	903	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER
373		904
374	Each watcher has, by default, a member C<void *data> that you can change	905	Each watcher has, by default, a member C<void *data> that you can change
375	and read at any time, libev will completely ignore it. This cna be used	906	and read at any time, libev will completely ignore it. This can be used
376	to associate arbitrary data with your watcher. If you need more data and	907	to associate arbitrary data with your watcher. If you need more data and
377	don't want to allocate memory and store a pointer to it in that data	908	don't want to allocate memory and store a pointer to it in that data
378	member, you can also "subclass" the watcher type and provide your own	909	member, you can also "subclass" the watcher type and provide your own
379	data:	910	data:
380		911
…		…
393	{	924	{
394	struct my_io *w = (struct my_io *)w_;	925	struct my_io *w = (struct my_io *)w_;
395	...	926	...
396	}	927	}
397		928
398	More interesting and less C-conformant ways of catsing your callback type	929	More interesting and less C-conformant ways of casting your callback type
399	have been omitted....	930	instead have been omitted.
		931
		932	Another common scenario is having some data structure with multiple
		933	watchers:
		934
		935	struct my_biggy
		936	{
		937	int some_data;
		938	ev_timer t1;
		939	ev_timer t2;
		940	}
		941
		942	In this case getting the pointer to C<my_biggy> is a bit more complicated,
		943	you need to use C<offsetof>:
		944
		945	#include <stddef.h>
		946
		947	static void
		948	t1_cb (EV_P_ struct ev_timer *w, int revents)
		949	{
		950	struct my_biggy big = (struct my_biggy *
		951	(((char *)w) - offsetof (struct my_biggy, t1));
		952	}
		953
		954	static void
		955	t2_cb (EV_P_ struct ev_timer *w, int revents)
		956	{
		957	struct my_biggy big = (struct my_biggy *
		958	(((char *)w) - offsetof (struct my_biggy, t2));
		959	}
400		960
401		961
402	=head1 WATCHER TYPES	962	=head1 WATCHER TYPES
403		963
404	This section describes each watcher in detail, but will not repeat	964	This section describes each watcher in detail, but will not repeat
405	information given in the last section.	965	information given in the last section. Any initialisation/set macros,
		966	functions and members specific to the watcher type are explained.
406		967
		968	Members are additionally marked with either I<[read-only]>, meaning that,
		969	while the watcher is active, you can look at the member and expect some
		970	sensible content, but you must not modify it (you can modify it while the
		971	watcher is stopped to your hearts content), or I<[read-write]>, which
		972	means you can expect it to have some sensible content while the watcher
		973	is active, but you can also modify it. Modifying it may not do something
		974	sensible or take immediate effect (or do anything at all), but libev will
		975	not crash or malfunction in any way.
		976
		977
407	=head2 struct ev_io - is my file descriptor readable or writable	978	=head2 C<ev_io> - is this file descriptor readable or writable?
408		979
409	I/O watchers check whether a file descriptor is readable or writable	980	I/O watchers check whether a file descriptor is readable or writable
410	in each iteration of the event loop (This behaviour is called	981	in each iteration of the event loop, or, more precisely, when reading
411	level-triggering because you keep receiving events as long as the	982	would not block the process and writing would at least be able to write
412	condition persists. Remember you cna stop the watcher if you don't want to	983	some data. This behaviour is called level-triggering because you keep
413	act on the event and neither want to receive future events).	984	receiving events as long as the condition persists. Remember you can stop
		985	the watcher if you don't want to act on the event and neither want to
		986	receive future events.
414		987
415	In general you can register as many read and/or write event watchers oer	988	In general you can register as many read and/or write event watchers per
416	fd as you want (as long as you don't confuse yourself). Setting all file	989	fd as you want (as long as you don't confuse yourself). Setting all file
417	descriptors to non-blocking mode is also usually a good idea (but not	990	descriptors to non-blocking mode is also usually a good idea (but not
418	required if you know what you are doing).	991	required if you know what you are doing).
419		992
420	You have to be careful with dup'ed file descriptors, though. Some backends
421	(the linux epoll backend is a notable example) cannot handle dup'ed file
422	descriptors correctly if you register interest in two or more fds pointing
423	to the same file/socket etc. description.
424
425	If you must do this, then force the use of a known-to-be-good backend	993	If you must do this, then force the use of a known-to-be-good backend
426	(at the time of this writing, this includes only EVMETHOD_SELECT and	994	(at the time of this writing, this includes only C<EVBACKEND_SELECT> and
427	EVMETHOD_POLL).	995	C<EVBACKEND_POLL>).
		996
		997	Another thing you have to watch out for is that it is quite easy to
		998	receive "spurious" readyness notifications, that is your callback might
		999	be called with C<EV_READ> but a subsequent C<read>(2) will actually block
		1000	because there is no data. Not only are some backends known to create a
		1001	lot of those (for example solaris ports), it is very easy to get into
		1002	this situation even with a relatively standard program structure. Thus
		1003	it is best to always use non-blocking I/O: An extra C<read>(2) returning
		1004	C<EAGAIN> is far preferable to a program hanging until some data arrives.
		1005
		1006	If you cannot run the fd in non-blocking mode (for example you should not
		1007	play around with an Xlib connection), then you have to seperately re-test
		1008	whether a file descriptor is really ready with a known-to-be good interface
		1009	such as poll (fortunately in our Xlib example, Xlib already does this on
		1010	its own, so its quite safe to use).
		1011
		1012	=head3 The special problem of disappearing file descriptors
		1013
		1014	Some backends (e.g. kqueue, epoll) need to be told about closing a file
		1015	descriptor (either by calling C<close> explicitly or by any other means,
		1016	such as C<dup>). The reason is that you register interest in some file
		1017	descriptor, but when it goes away, the operating system will silently drop
		1018	this interest. If another file descriptor with the same number then is
		1019	registered with libev, there is no efficient way to see that this is, in
		1020	fact, a different file descriptor.
		1021
		1022	To avoid having to explicitly tell libev about such cases, libev follows
		1023	the following policy: Each time C<ev_io_set> is being called, libev
		1024	will assume that this is potentially a new file descriptor, otherwise
		1025	it is assumed that the file descriptor stays the same. That means that
		1026	you I<have> to call C<ev_io_set> (or C<ev_io_init>) when you change the
		1027	descriptor even if the file descriptor number itself did not change.
		1028
		1029	This is how one would do it normally anyway, the important point is that
		1030	the libev application should not optimise around libev but should leave
		1031	optimisations to libev.
		1032
		1033	=head3 The special problem of dup'ed file descriptors
		1034
		1035	Some backends (e.g. epoll), cannot register events for file descriptors,
		1036	but only events for the underlying file descriptions. That means when you
		1037	have C<dup ()>'ed file descriptors or weirder constellations, and register
		1038	events for them, only one file descriptor might actually receive events.
		1039
		1040	There is no workaround possible except not registering events
		1041	for potentially C<dup ()>'ed file descriptors, or to resort to
		1042	C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>.
		1043
		1044	=head3 The special problem of fork
		1045
		1046	Some backends (epoll, kqueue) do not support C<fork ()> at all or exhibit
		1047	useless behaviour. Libev fully supports fork, but needs to be told about
		1048	it in the child.
		1049
		1050	To support fork in your programs, you either have to call
		1051	C<ev_default_fork ()> or C<ev_loop_fork ()> after a fork in the child,
		1052	enable C<EVFLAG_FORKCHECK>, or resort to C<EVBACKEND_SELECT> or
		1053	C<EVBACKEND_POLL>.
		1054
		1055
		1056	=head3 Watcher-Specific Functions
428		1057
429	=over 4	1058	=over 4
430		1059
431	=item ev_io_init (ev_io *, callback, int fd, int events)	1060	=item ev_io_init (ev_io *, callback, int fd, int events)
432		1061
433	=item ev_io_set (ev_io *, int fd, int events)	1062	=item ev_io_set (ev_io *, int fd, int events)
434		1063
435	Configures an ev_io watcher. The fd is the file descriptor to rceeive	1064	Configures an C<ev_io> watcher. The C<fd> is the file descriptor to
436	events for and events is either C<EV_READ>, C<EV_WRITE> or C<EV_READ |	1065	rceeive events for and events is either C<EV_READ>, C<EV_WRITE> or
437	EV_WRITE> to receive the given events.	1066	C<EV_READ | EV_WRITE> to receive the given events.
		1067
		1068	=item int fd [read-only]
		1069
		1070	The file descriptor being watched.
		1071
		1072	=item int events [read-only]
		1073
		1074	The events being watched.
438		1075
439	=back	1076	=back
440		1077
		1078	=head3 Examples
		1079
		1080	Example: Call C<stdin_readable_cb> when STDIN_FILENO has become, well
		1081	readable, but only once. Since it is likely line-buffered, you could
		1082	attempt to read a whole line in the callback.
		1083
		1084	static void
		1085	stdin_readable_cb (struct ev_loop *loop, struct ev_io *w, int revents)
		1086	{
		1087	ev_io_stop (loop, w);
		1088	.. read from stdin here (or from w->fd) and haqndle any I/O errors
		1089	}
		1090
		1091	...
		1092	struct ev_loop *loop = ev_default_init (0);
		1093	struct ev_io stdin_readable;
		1094	ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
		1095	ev_io_start (loop, &stdin_readable);
		1096	ev_loop (loop, 0);
		1097
		1098
441	=head2 struct ev_timer - relative and optionally recurring timeouts	1099	=head2 C<ev_timer> - relative and optionally repeating timeouts
442		1100
443	Timer watchers are simple relative timers that generate an event after a	1101	Timer watchers are simple relative timers that generate an event after a
444	given time, and optionally repeating in regular intervals after that.	1102	given time, and optionally repeating in regular intervals after that.
445		1103
446	The timers are based on real time, that is, if you register an event that	1104	The timers are based on real time, that is, if you register an event that
447	times out after an hour and youreset your system clock to last years	1105	times out after an hour and you reset your system clock to last years
448	time, it will still time out after (roughly) and hour. "Roughly" because	1106	time, it will still time out after (roughly) and hour. "Roughly" because
449	detecting time jumps is hard, and soem inaccuracies are unavoidable (the	1107	detecting time jumps is hard, and some inaccuracies are unavoidable (the
450	monotonic clock option helps a lot here).	1108	monotonic clock option helps a lot here).
		1109
		1110	The relative timeouts are calculated relative to the C<ev_now ()>
		1111	time. This is usually the right thing as this timestamp refers to the time
		1112	of the event triggering whatever timeout you are modifying/starting. If
		1113	you suspect event processing to be delayed and you I<need> to base the timeout
		1114	on the current time, use something like this to adjust for this:
		1115
		1116	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);
		1117
		1118	The callback is guarenteed to be invoked only when its timeout has passed,
		1119	but if multiple timers become ready during the same loop iteration then
		1120	order of execution is undefined.
		1121
		1122	=head3 Watcher-Specific Functions and Data Members
451		1123
452	=over 4	1124	=over 4
453		1125
454	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)	1126	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)
455		1127
…		…
461	later, again, and again, until stopped manually.	1133	later, again, and again, until stopped manually.
462		1134
463	The timer itself will do a best-effort at avoiding drift, that is, if you	1135	The timer itself will do a best-effort at avoiding drift, that is, if you
464	configure a timer to trigger every 10 seconds, then it will trigger at	1136	configure a timer to trigger every 10 seconds, then it will trigger at
465	exactly 10 second intervals. If, however, your program cannot keep up with	1137	exactly 10 second intervals. If, however, your program cannot keep up with
466	the timer (ecause it takes longer than those 10 seconds to do stuff) the	1138	the timer (because it takes longer than those 10 seconds to do stuff) the
467	timer will not fire more than once per event loop iteration.	1139	timer will not fire more than once per event loop iteration.
468		1140
469	=item ev_timer_again (loop)	1141	=item ev_timer_again (loop)
470		1142
471	This will act as if the timer timed out and restart it again if it is	1143	This will act as if the timer timed out and restart it again if it is
472	repeating. The exact semantics are:	1144	repeating. The exact semantics are:
473		1145
		1146	If the timer is pending, its pending status is cleared.
		1147
474	If the timer is started but nonrepeating, stop it.	1148	If the timer is started but nonrepeating, stop it (as if it timed out).
475		1149
476	If the timer is repeating, either start it if necessary (with the repeat	1150	If the timer is repeating, either start it if necessary (with the
477	value), or reset the running timer to the repeat value.	1151	C<repeat> value), or reset the running timer to the C<repeat> value.
478		1152
479	This sounds a bit complicated, but here is a useful and typical	1153	This sounds a bit complicated, but here is a useful and typical
480	example: Imagine you have a tcp connection and you want a so-called idle	1154	example: Imagine you have a tcp connection and you want a so-called idle
481	timeout, that is, you want to be called when there have been, say, 60	1155	timeout, that is, you want to be called when there have been, say, 60
482	seconds of inactivity on the socket. The easiest way to do this is to	1156	seconds of inactivity on the socket. The easiest way to do this is to
483	configure an ev_timer with after=repeat=60 and calling ev_timer_again each	1157	configure an C<ev_timer> with a C<repeat> value of C<60> and then call
484	time you successfully read or write some data. If you go into an idle	1158	C<ev_timer_again> each time you successfully read or write some data. If
485	state where you do not expect data to travel on the socket, you can stop	1159	you go into an idle state where you do not expect data to travel on the
486	the timer, and again will automatically restart it if need be.	1160	socket, you can C<ev_timer_stop> the timer, and C<ev_timer_again> will
		1161	automatically restart it if need be.
		1162
		1163	That means you can ignore the C<after> value and C<ev_timer_start>
		1164	altogether and only ever use the C<repeat> value and C<ev_timer_again>:
		1165
		1166	ev_timer_init (timer, callback, 0., 5.);
		1167	ev_timer_again (loop, timer);
		1168	...
		1169	timer->again = 17.;
		1170	ev_timer_again (loop, timer);
		1171	...
		1172	timer->again = 10.;
		1173	ev_timer_again (loop, timer);
		1174
		1175	This is more slightly efficient then stopping/starting the timer each time
		1176	you want to modify its timeout value.
		1177
		1178	=item ev_tstamp repeat [read-write]
		1179
		1180	The current C<repeat> value. Will be used each time the watcher times out
		1181	or C<ev_timer_again> is called and determines the next timeout (if any),
		1182	which is also when any modifications are taken into account.
487		1183
488	=back	1184	=back
489		1185
		1186	=head3 Examples
		1187
		1188	Example: Create a timer that fires after 60 seconds.
		1189
		1190	static void
		1191	one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
		1192	{
		1193	.. one minute over, w is actually stopped right here
		1194	}
		1195
		1196	struct ev_timer mytimer;
		1197	ev_timer_init (&mytimer, one_minute_cb, 60., 0.);
		1198	ev_timer_start (loop, &mytimer);
		1199
		1200	Example: Create a timeout timer that times out after 10 seconds of
		1201	inactivity.
		1202
		1203	static void
		1204	timeout_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
		1205	{
		1206	.. ten seconds without any activity
		1207	}
		1208
		1209	struct ev_timer mytimer;
		1210	ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */
		1211	ev_timer_again (&mytimer); /* start timer */
		1212	ev_loop (loop, 0);
		1213
		1214	// and in some piece of code that gets executed on any "activity":
		1215	// reset the timeout to start ticking again at 10 seconds
		1216	ev_timer_again (&mytimer);
		1217
		1218
490	=head2 ev_periodic - to cron or not to cron it	1219	=head2 C<ev_periodic> - to cron or not to cron?
491		1220
492	Periodic watchers are also timers of a kind, but they are very versatile	1221	Periodic watchers are also timers of a kind, but they are very versatile
493	(and unfortunately a bit complex).	1222	(and unfortunately a bit complex).
494		1223
495	Unlike ev_timer's, they are not based on real time (or relative time)	1224	Unlike C<ev_timer>'s, they are not based on real time (or relative time)
496	but on wallclock time (absolute time). You can tell a periodic watcher	1225	but on wallclock time (absolute time). You can tell a periodic watcher
497	to trigger "at" some specific point in time. For example, if you tell a	1226	to trigger "at" some specific point in time. For example, if you tell a
498	periodic watcher to trigger in 10 seconds (by specifiying e.g. c<ev_now ()	1227	periodic watcher to trigger in 10 seconds (by specifiying e.g. C<ev_now ()
499	+ 10.>) and then reset your system clock to the last year, then it will	1228	+ 10.>) and then reset your system clock to the last year, then it will
500	take a year to trigger the event (unlike an ev_timer, which would trigger	1229	take a year to trigger the event (unlike an C<ev_timer>, which would trigger
501	roughly 10 seconds later and of course not if you reset your system time	1230	roughly 10 seconds later).
502	again).
503		1231
504	They can also be used to implement vastly more complex timers, such as	1232	They can also be used to implement vastly more complex timers, such as
505	triggering an event on eahc midnight, local time.	1233	triggering an event on each midnight, local time or other, complicated,
		1234	rules.
		1235
		1236	As with timers, the callback is guarenteed to be invoked only when the
		1237	time (C<at>) has been passed, but if multiple periodic timers become ready
		1238	during the same loop iteration then order of execution is undefined.
		1239
		1240	=head3 Watcher-Specific Functions and Data Members
506		1241
507	=over 4	1242	=over 4
508		1243
509	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)	1244	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)
510		1245
511	=item ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)	1246	=item ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)
512		1247
513	Lots of arguments, lets sort it out... There are basically three modes of	1248	Lots of arguments, lets sort it out... There are basically three modes of
514	operation, and we will explain them from simplest to complex:	1249	operation, and we will explain them from simplest to complex:
515		1250
516
517	=over 4	1251	=over 4
518		1252
519	=item * absolute timer (interval = reschedule_cb = 0)	1253	=item * absolute timer (at = time, interval = reschedule_cb = 0)
520		1254
521	In this configuration the watcher triggers an event at the wallclock time	1255	In this configuration the watcher triggers an event at the wallclock time
522	C<at> and doesn't repeat. It will not adjust when a time jump occurs,	1256	C<at> and doesn't repeat. It will not adjust when a time jump occurs,
523	that is, if it is to be run at January 1st 2011 then it will run when the	1257	that is, if it is to be run at January 1st 2011 then it will run when the
524	system time reaches or surpasses this time.	1258	system time reaches or surpasses this time.
525		1259
526	=item * non-repeating interval timer (interval > 0, reschedule_cb = 0)	1260	=item * non-repeating interval timer (at = offset, interval > 0, reschedule_cb = 0)
527		1261
528	In this mode the watcher will always be scheduled to time out at the next	1262	In this mode the watcher will always be scheduled to time out at the next
529	C<at + N * interval> time (for some integer N) and then repeat, regardless	1263	C<at + N * interval> time (for some integer N, which can also be negative)
530	of any time jumps.	1264	and then repeat, regardless of any time jumps.
531		1265
532	This can be used to create timers that do not drift with respect to system	1266	This can be used to create timers that do not drift with respect to system
533	time:	1267	time:
534		1268
535	ev_periodic_set (&periodic, 0., 3600., 0);	1269	ev_periodic_set (&periodic, 0., 3600., 0);
536		1270
537	This doesn't mean there will always be 3600 seconds in between triggers,	1271	This doesn't mean there will always be 3600 seconds in between triggers,
538	but only that the the callback will be called when the system time shows a	1272	but only that the the callback will be called when the system time shows a
539	full hour (UTC), or more correct, when the system time is evenly divisible	1273	full hour (UTC), or more correctly, when the system time is evenly divisible
540	by 3600.	1274	by 3600.
541		1275
542	Another way to think about it (for the mathematically inclined) is that	1276	Another way to think about it (for the mathematically inclined) is that
543	ev_periodic will try to run the callback in this mode at the next possible	1277	C<ev_periodic> will try to run the callback in this mode at the next possible
544	time where C<time = at (mod interval)>, regardless of any time jumps.	1278	time where C<time = at (mod interval)>, regardless of any time jumps.
545		1279
		1280	For numerical stability it is preferable that the C<at> value is near
		1281	C<ev_now ()> (the current time), but there is no range requirement for
		1282	this value.
		1283
546	=item * manual reschedule mode (reschedule_cb = callback)	1284	=item * manual reschedule mode (at and interval ignored, reschedule_cb = callback)
547		1285
548	In this mode the values for C<interval> and C<at> are both being	1286	In this mode the values for C<interval> and C<at> are both being
549	ignored. Instead, each time the periodic watcher gets scheduled, the	1287	ignored. Instead, each time the periodic watcher gets scheduled, the
550	reschedule callback will be called with the watcher as first, and the	1288	reschedule callback will be called with the watcher as first, and the
551	current time as second argument.	1289	current time as second argument.
552		1290
553	NOTE: I<This callback MUST NOT stop or destroy the periodic or any other	1291	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher,
554	periodic watcher, ever, or make any event loop modificstions>. If you need	1292	ever, or make any event loop modifications>. If you need to stop it,
555	to stop it, return 1e30 (or so, fudge fudge) and stop it afterwards.	1293	return C<now + 1e30> (or so, fudge fudge) and stop it afterwards (e.g. by
		1294	starting an C<ev_prepare> watcher, which is legal).
556		1295
557	Its prototype is c<ev_tstamp (*reschedule_cb)(struct ev_periodic *w,	1296	Its prototype is C<ev_tstamp (*reschedule_cb)(struct ev_periodic *w,
558	ev_tstamp now)>, e.g.:	1297	ev_tstamp now)>, e.g.:
559		1298
560	static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)	1299	static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)
561	{	1300	{
562	return now + 60.;	1301	return now + 60.;
…		…
565	It must return the next time to trigger, based on the passed time value	1304	It must return the next time to trigger, based on the passed time value
566	(that is, the lowest time value larger than to the second argument). It	1305	(that is, the lowest time value larger than to the second argument). It
567	will usually be called just before the callback will be triggered, but	1306	will usually be called just before the callback will be triggered, but
568	might be called at other times, too.	1307	might be called at other times, too.
569		1308
		1309	NOTE: I<< This callback must always return a time that is later than the
		1310	passed C<now> value >>. Not even C<now> itself will do, it I<must> be larger.
		1311
570	This can be used to create very complex timers, such as a timer that	1312	This can be used to create very complex timers, such as a timer that
571	triggers on each midnight, local time. To do this, you would calculate the	1313	triggers on each midnight, local time. To do this, you would calculate the
572	next midnight after C<now> and return the timestamp value for this. How you do this	1314	next midnight after C<now> and return the timestamp value for this. How
573	is, again, up to you (but it is not trivial).	1315	you do this is, again, up to you (but it is not trivial, which is the main
		1316	reason I omitted it as an example).
574		1317
575	=back	1318	=back
576		1319
577	=item ev_periodic_again (loop, ev_periodic *)	1320	=item ev_periodic_again (loop, ev_periodic *)
578		1321
579	Simply stops and restarts the periodic watcher again. This is only useful	1322	Simply stops and restarts the periodic watcher again. This is only useful
580	when you changed some parameters or the reschedule callback would return	1323	when you changed some parameters or the reschedule callback would return
581	a different time than the last time it was called (e.g. in a crond like	1324	a different time than the last time it was called (e.g. in a crond like
582	program when the crontabs have changed).	1325	program when the crontabs have changed).
583		1326
		1327	=item ev_tstamp offset [read-write]
		1328
		1329	When repeating, this contains the offset value, otherwise this is the
		1330	absolute point in time (the C<at> value passed to C<ev_periodic_set>).
		1331
		1332	Can be modified any time, but changes only take effect when the periodic
		1333	timer fires or C<ev_periodic_again> is being called.
		1334
		1335	=item ev_tstamp interval [read-write]
		1336
		1337	The current interval value. Can be modified any time, but changes only
		1338	take effect when the periodic timer fires or C<ev_periodic_again> is being
		1339	called.
		1340
		1341	=item ev_tstamp (*reschedule_cb)(struct ev_periodic *w, ev_tstamp now) [read-write]
		1342
		1343	The current reschedule callback, or C<0>, if this functionality is
		1344	switched off. Can be changed any time, but changes only take effect when
		1345	the periodic timer fires or C<ev_periodic_again> is being called.
		1346
		1347	=item ev_tstamp at [read-only]
		1348
		1349	When active, contains the absolute time that the watcher is supposed to
		1350	trigger next.
		1351
584	=back	1352	=back
585		1353
		1354	=head3 Examples
		1355
		1356	Example: Call a callback every hour, or, more precisely, whenever the
		1357	system clock is divisible by 3600. The callback invocation times have
		1358	potentially a lot of jittering, but good long-term stability.
		1359
		1360	static void
		1361	clock_cb (struct ev_loop *loop, struct ev_io *w, int revents)
		1362	{
		1363	... its now a full hour (UTC, or TAI or whatever your clock follows)
		1364	}
		1365
		1366	struct ev_periodic hourly_tick;
		1367	ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);
		1368	ev_periodic_start (loop, &hourly_tick);
		1369
		1370	Example: The same as above, but use a reschedule callback to do it:
		1371
		1372	#include <math.h>
		1373
		1374	static ev_tstamp
		1375	my_scheduler_cb (struct ev_periodic *w, ev_tstamp now)
		1376	{
		1377	return fmod (now, 3600.) + 3600.;
		1378	}
		1379
		1380	ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
		1381
		1382	Example: Call a callback every hour, starting now:
		1383
		1384	struct ev_periodic hourly_tick;
		1385	ev_periodic_init (&hourly_tick, clock_cb,
		1386	fmod (ev_now (loop), 3600.), 3600., 0);
		1387	ev_periodic_start (loop, &hourly_tick);
		1388
		1389
586	=head2 ev_signal - signal me when a signal gets signalled	1390	=head2 C<ev_signal> - signal me when a signal gets signalled!
587		1391
588	Signal watchers will trigger an event when the process receives a specific	1392	Signal watchers will trigger an event when the process receives a specific
589	signal one or more times. Even though signals are very asynchronous, libev	1393	signal one or more times. Even though signals are very asynchronous, libev
590	will try its best to deliver signals synchronously, i.e. as part of the	1394	will try it's best to deliver signals synchronously, i.e. as part of the
591	normal event processing, like any other event.	1395	normal event processing, like any other event.
592		1396
593	You cna configure as many watchers as you like per signal. Only when the	1397	You can configure as many watchers as you like per signal. Only when the
594	first watcher gets started will libev actually register a signal watcher	1398	first watcher gets started will libev actually register a signal watcher
595	with the kernel (thus it coexists with your own signal handlers as long	1399	with the kernel (thus it coexists with your own signal handlers as long
596	as you don't register any with libev). Similarly, when the last signal	1400	as you don't register any with libev). Similarly, when the last signal
597	watcher for a signal is stopped libev will reset the signal handler to	1401	watcher for a signal is stopped libev will reset the signal handler to
598	SIG_DFL (regardless of what it was set to before).	1402	SIG_DFL (regardless of what it was set to before).
599		1403
		1404	=head3 Watcher-Specific Functions and Data Members
		1405
600	=over 4	1406	=over 4
601		1407
602	=item ev_signal_init (ev_signal *, callback, int signum)	1408	=item ev_signal_init (ev_signal *, callback, int signum)
603		1409
604	=item ev_signal_set (ev_signal *, int signum)	1410	=item ev_signal_set (ev_signal *, int signum)
605		1411
606	Configures the watcher to trigger on the given signal number (usually one	1412	Configures the watcher to trigger on the given signal number (usually one
607	of the C<SIGxxx> constants).	1413	of the C<SIGxxx> constants).
608		1414
		1415	=item int signum [read-only]
		1416
		1417	The signal the watcher watches out for.
		1418
609	=back	1419	=back
610		1420
		1421
611	=head2 ev_child - wait for pid status changes	1422	=head2 C<ev_child> - watch out for process status changes
612		1423
613	Child watchers trigger when your process receives a SIGCHLD in response to	1424	Child watchers trigger when your process receives a SIGCHLD in response to
614	some child status changes (most typically when a child of yours dies).	1425	some child status changes (most typically when a child of yours dies).
		1426
		1427	=head3 Watcher-Specific Functions and Data Members
615		1428
616	=over 4	1429	=over 4
617		1430
618	=item ev_child_init (ev_child *, callback, int pid)	1431	=item ev_child_init (ev_child *, callback, int pid)
619		1432
620	=item ev_child_set (ev_child *, int pid)	1433	=item ev_child_set (ev_child *, int pid)
621		1434
622	Configures the watcher to wait for status changes of process C<pid> (or	1435	Configures the watcher to wait for status changes of process C<pid> (or
623	I<any> process if C<pid> is specified as C<0>). The callback can look	1436	I<any> process if C<pid> is specified as C<0>). The callback can look
624	at the C<rstatus> member of the C<ev_child> watcher structure to see	1437	at the C<rstatus> member of the C<ev_child> watcher structure to see
625	the status word (use the macros from C<sys/wait.h>). The C<rpid> member	1438	the status word (use the macros from C<sys/wait.h> and see your systems
626	contains the pid of the process causing the status change.	1439	C<waitpid> documentation). The C<rpid> member contains the pid of the
		1440	process causing the status change.
		1441
		1442	=item int pid [read-only]
		1443
		1444	The process id this watcher watches out for, or C<0>, meaning any process id.
		1445
		1446	=item int rpid [read-write]
		1447
		1448	The process id that detected a status change.
		1449
		1450	=item int rstatus [read-write]
		1451
		1452	The process exit/trace status caused by C<rpid> (see your systems
		1453	C<waitpid> and C<sys/wait.h> documentation for details).
627		1454
628	=back	1455	=back
629		1456
		1457	=head3 Examples
		1458
		1459	Example: Try to exit cleanly on SIGINT and SIGTERM.
		1460
		1461	static void
		1462	sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)
		1463	{
		1464	ev_unloop (loop, EVUNLOOP_ALL);
		1465	}
		1466
		1467	struct ev_signal signal_watcher;
		1468	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
		1469	ev_signal_start (loop, &sigint_cb);
		1470
		1471
		1472	=head2 C<ev_stat> - did the file attributes just change?
		1473
		1474	This watches a filesystem path for attribute changes. That is, it calls
		1475	C<stat> regularly (or when the OS says it changed) and sees if it changed
		1476	compared to the last time, invoking the callback if it did.
		1477
		1478	The path does not need to exist: changing from "path exists" to "path does
		1479	not exist" is a status change like any other. The condition "path does
		1480	not exist" is signified by the C<st_nlink> field being zero (which is
		1481	otherwise always forced to be at least one) and all the other fields of
		1482	the stat buffer having unspecified contents.
		1483
		1484	The path I<should> be absolute and I<must not> end in a slash. If it is
		1485	relative and your working directory changes, the behaviour is undefined.
		1486
		1487	Since there is no standard to do this, the portable implementation simply
		1488	calls C<stat (2)> regularly on the path to see if it changed somehow. You
		1489	can specify a recommended polling interval for this case. If you specify
		1490	a polling interval of C<0> (highly recommended!) then a I<suitable,
		1491	unspecified default> value will be used (which you can expect to be around
		1492	five seconds, although this might change dynamically). Libev will also
		1493	impose a minimum interval which is currently around C<0.1>, but thats
		1494	usually overkill.
		1495
		1496	This watcher type is not meant for massive numbers of stat watchers,
		1497	as even with OS-supported change notifications, this can be
		1498	resource-intensive.
		1499
		1500	At the time of this writing, only the Linux inotify interface is
		1501	implemented (implementing kqueue support is left as an exercise for the
		1502	reader). Inotify will be used to give hints only and should not change the
		1503	semantics of C<ev_stat> watchers, which means that libev sometimes needs
		1504	to fall back to regular polling again even with inotify, but changes are
		1505	usually detected immediately, and if the file exists there will be no
		1506	polling.
		1507
		1508	=head3 Inotify
		1509
		1510	When C<inotify (7)> support has been compiled into libev (generally only
		1511	available on Linux) and present at runtime, it will be used to speed up
		1512	change detection where possible. The inotify descriptor will be created lazily
		1513	when the first C<ev_stat> watcher is being started.
		1514
		1515	Inotify presense does not change the semantics of C<ev_stat> watchers
		1516	except that changes might be detected earlier, and in some cases, to avoid
		1517	making regular C<stat> calls. Even in the presense of inotify support
		1518	there are many cases where libev has to resort to regular C<stat> polling.
		1519
		1520	(There is no support for kqueue, as apparently it cannot be used to
		1521	implement this functionality, due to the requirement of having a file
		1522	descriptor open on the object at all times).
		1523
		1524	=head3 The special problem of stat time resolution
		1525
		1526	The C<stat ()> syscall only supports full-second resolution portably, and
		1527	even on systems where the resolution is higher, many filesystems still
		1528	only support whole seconds.
		1529
		1530	That means that, if the time is the only thing that changes, you might
		1531	miss updates: on the first update, C<ev_stat> detects a change and calls
		1532	your callback, which does something. When there is another update within
		1533	the same second, C<ev_stat> will be unable to detect it.
		1534
		1535	The solution to this is to delay acting on a change for a second (or till
		1536	the next second boundary), using a roughly one-second delay C<ev_timer>
		1537	(C<ev_timer_set (w, 0., 1.01); ev_timer_again (loop, w)>). The C<.01>
		1538	is added to work around small timing inconsistencies of some operating
		1539	systems.
		1540
		1541	=head3 Watcher-Specific Functions and Data Members
		1542
		1543	=over 4
		1544
		1545	=item ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)
		1546
		1547	=item ev_stat_set (ev_stat *, const char *path, ev_tstamp interval)
		1548
		1549	Configures the watcher to wait for status changes of the given
		1550	C<path>. The C<interval> is a hint on how quickly a change is expected to
		1551	be detected and should normally be specified as C<0> to let libev choose
		1552	a suitable value. The memory pointed to by C<path> must point to the same
		1553	path for as long as the watcher is active.
		1554
		1555	The callback will be receive C<EV_STAT> when a change was detected,
		1556	relative to the attributes at the time the watcher was started (or the
		1557	last change was detected).
		1558
		1559	=item ev_stat_stat (ev_stat *)
		1560
		1561	Updates the stat buffer immediately with new values. If you change the
		1562	watched path in your callback, you could call this fucntion to avoid
		1563	detecting this change (while introducing a race condition). Can also be
		1564	useful simply to find out the new values.
		1565
		1566	=item ev_statdata attr [read-only]
		1567
		1568	The most-recently detected attributes of the file. Although the type is of
		1569	C<ev_statdata>, this is usually the (or one of the) C<struct stat> types
		1570	suitable for your system. If the C<st_nlink> member is C<0>, then there
		1571	was some error while C<stat>ing the file.
		1572
		1573	=item ev_statdata prev [read-only]
		1574
		1575	The previous attributes of the file. The callback gets invoked whenever
		1576	C<prev> != C<attr>.
		1577
		1578	=item ev_tstamp interval [read-only]
		1579
		1580	The specified interval.
		1581
		1582	=item const char *path [read-only]
		1583
		1584	The filesystem path that is being watched.
		1585
		1586	=back
		1587
		1588	=head3 Examples
		1589
		1590	Example: Watch C</etc/passwd> for attribute changes.
		1591
		1592	static void
		1593	passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)
		1594	{
		1595	/* /etc/passwd changed in some way */
		1596	if (w->attr.st_nlink)
		1597	{
		1598	printf ("passwd current size %ld\n", (long)w->attr.st_size);
		1599	printf ("passwd current atime %ld\n", (long)w->attr.st_mtime);
		1600	printf ("passwd current mtime %ld\n", (long)w->attr.st_mtime);
		1601	}
		1602	else
		1603	/* you shalt not abuse printf for puts */
		1604	puts ("wow, /etc/passwd is not there, expect problems. "
		1605	"if this is windows, they already arrived\n");
		1606	}
		1607
		1608	...
		1609	ev_stat passwd;
		1610
		1611	ev_stat_init (&passwd, passwd_cb, "/etc/passwd", 0.);
		1612	ev_stat_start (loop, &passwd);
		1613
		1614	Example: Like above, but additionally use a one-second delay so we do not
		1615	miss updates (however, frequent updates will delay processing, too, so
		1616	one might do the work both on C<ev_stat> callback invocation I<and> on
		1617	C<ev_timer> callback invocation).
		1618
		1619	static ev_stat passwd;
		1620	static ev_timer timer;
		1621
		1622	static void
		1623	timer_cb (EV_P_ ev_timer *w, int revents)
		1624	{
		1625	ev_timer_stop (EV_A_ w);
		1626
		1627	/* now it's one second after the most recent passwd change */
		1628	}
		1629
		1630	static void
		1631	stat_cb (EV_P_ ev_stat *w, int revents)
		1632	{
		1633	/* reset the one-second timer */
		1634	ev_timer_again (EV_A_ &timer);
		1635	}
		1636
		1637	...
		1638	ev_stat_init (&passwd, stat_cb, "/etc/passwd", 0.);
		1639	ev_stat_start (loop, &passwd);
		1640	ev_timer_init (&timer, timer_cb, 0., 1.01);
		1641
		1642
630	=head2 ev_idle - when you've got nothing better to do	1643	=head2 C<ev_idle> - when you've got nothing better to do...
631		1644
632	Idle watchers trigger events when there are no other I/O or timer (or	1645	Idle watchers trigger events when no other events of the same or higher
633	periodic) events pending. That is, as long as your process is busy	1646	priority are pending (prepare, check and other idle watchers do not
634	handling sockets or timeouts it will not be called. But when your process	1647	count).
635	is idle all idle watchers are being called again and again - until	1648
		1649	That is, as long as your process is busy handling sockets or timeouts
		1650	(or even signals, imagine) of the same or higher priority it will not be
		1651	triggered. But when your process is idle (or only lower-priority watchers
		1652	are pending), the idle watchers are being called once per event loop
636	stopped, that is, or your process receives more events.	1653	iteration - until stopped, that is, or your process receives more events
		1654	and becomes busy again with higher priority stuff.
637		1655
638	The most noteworthy effect is that as long as any idle watchers are	1656	The most noteworthy effect is that as long as any idle watchers are
639	active, the process will not block when waiting for new events.	1657	active, the process will not block when waiting for new events.
640		1658
641	Apart from keeping your process non-blocking (which is a useful	1659	Apart from keeping your process non-blocking (which is a useful
642	effect on its own sometimes), idle watchers are a good place to do	1660	effect on its own sometimes), idle watchers are a good place to do
643	"pseudo-background processing", or delay processing stuff to after the	1661	"pseudo-background processing", or delay processing stuff to after the
644	event loop has handled all outstanding events.	1662	event loop has handled all outstanding events.
645		1663
		1664	=head3 Watcher-Specific Functions and Data Members
		1665
646	=over 4	1666	=over 4
647		1667
648	=item ev_idle_init (ev_signal *, callback)	1668	=item ev_idle_init (ev_signal *, callback)
649		1669
650	Initialises and configures the idle watcher - it has no parameters of any	1670	Initialises and configures the idle watcher - it has no parameters of any
651	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,	1671	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,
652	believe me.	1672	believe me.
653		1673
654	=back	1674	=back
655		1675
656	=head2 prepare and check - your hooks into the event loop	1676	=head3 Examples
657		1677
		1678	Example: Dynamically allocate an C<ev_idle> watcher, start it, and in the
		1679	callback, free it. Also, use no error checking, as usual.
		1680
		1681	static void
		1682	idle_cb (struct ev_loop *loop, struct ev_idle *w, int revents)
		1683	{
		1684	free (w);
		1685	// now do something you wanted to do when the program has
		1686	// no longer asnything immediate to do.
		1687	}
		1688
		1689	struct ev_idle *idle_watcher = malloc (sizeof (struct ev_idle));
		1690	ev_idle_init (idle_watcher, idle_cb);
		1691	ev_idle_start (loop, idle_cb);
		1692
		1693
		1694	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!
		1695
658	Prepare and check watchers usually (but not always) are used in	1696	Prepare and check watchers are usually (but not always) used in tandem:
659	tandom. Prepare watchers get invoked before the process blocks and check	1697	prepare watchers get invoked before the process blocks and check watchers
660	watchers afterwards.	1698	afterwards.
661		1699
		1700	You I<must not> call C<ev_loop> or similar functions that enter
		1701	the current event loop from either C<ev_prepare> or C<ev_check>
		1702	watchers. Other loops than the current one are fine, however. The
		1703	rationale behind this is that you do not need to check for recursion in
		1704	those watchers, i.e. the sequence will always be C<ev_prepare>, blocking,
		1705	C<ev_check> so if you have one watcher of each kind they will always be
		1706	called in pairs bracketing the blocking call.
		1707
662	Their main purpose is to integrate other event mechanisms into libev. This	1708	Their main purpose is to integrate other event mechanisms into libev and
663	could be used, for example, to track variable changes, implement your own	1709	their use is somewhat advanced. This could be used, for example, to track
664	watchers, integrate net-snmp or a coroutine library and lots more.	1710	variable changes, implement your own watchers, integrate net-snmp or a
		1711	coroutine library and lots more. They are also occasionally useful if
		1712	you cache some data and want to flush it before blocking (for example,
		1713	in X programs you might want to do an C<XFlush ()> in an C<ev_prepare>
		1714	watcher).
665		1715
666	This is done by examining in each prepare call which file descriptors need	1716	This is done by examining in each prepare call which file descriptors need
667	to be watched by the other library, registering ev_io watchers for them	1717	to be watched by the other library, registering C<ev_io> watchers for
668	and starting an ev_timer watcher for any timeouts (many libraries provide	1718	them and starting an C<ev_timer> watcher for any timeouts (many libraries
669	just this functionality). Then, in the check watcher you check for any	1719	provide just this functionality). Then, in the check watcher you check for
670	events that occured (by making your callbacks set soem flags for example)	1720	any events that occured (by checking the pending status of all watchers
671	and call back into the library.	1721	and stopping them) and call back into the library. The I/O and timer
		1722	callbacks will never actually be called (but must be valid nevertheless,
		1723	because you never know, you know?).
672		1724
673	As another example, the perl Coro module uses these hooks to integrate	1725	As another example, the Perl Coro module uses these hooks to integrate
674	coroutines into libev programs, by yielding to other active coroutines	1726	coroutines into libev programs, by yielding to other active coroutines
675	during each prepare and only letting the process block if no coroutines	1727	during each prepare and only letting the process block if no coroutines
676	are ready to run.	1728	are ready to run (it's actually more complicated: it only runs coroutines
		1729	with priority higher than or equal to the event loop and one coroutine
		1730	of lower priority, but only once, using idle watchers to keep the event
		1731	loop from blocking if lower-priority coroutines are active, thus mapping
		1732	low-priority coroutines to idle/background tasks).
		1733
		1734	It is recommended to give C<ev_check> watchers highest (C<EV_MAXPRI>)
		1735	priority, to ensure that they are being run before any other watchers
		1736	after the poll. Also, C<ev_check> watchers (and C<ev_prepare> watchers,
		1737	too) should not activate ("feed") events into libev. While libev fully
		1738	supports this, they will be called before other C<ev_check> watchers
		1739	did their job. As C<ev_check> watchers are often used to embed other
		1740	(non-libev) event loops those other event loops might be in an unusable
		1741	state until their C<ev_check> watcher ran (always remind yourself to
		1742	coexist peacefully with others).
		1743
		1744	=head3 Watcher-Specific Functions and Data Members
677		1745
678	=over 4	1746	=over 4
679		1747
680	=item ev_prepare_init (ev_prepare *, callback)	1748	=item ev_prepare_init (ev_prepare *, callback)
681		1749
682	=item ev_check_init (ev_check *, callback)	1750	=item ev_check_init (ev_check *, callback)
683		1751
684	Initialises and configures the prepare or check watcher - they have no	1752	Initialises and configures the prepare or check watcher - they have no
685	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>	1753	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>
686	macros, but using them is utterly, utterly pointless.	1754	macros, but using them is utterly, utterly and completely pointless.
687		1755
688	=back	1756	=back
689		1757
		1758	=head3 Examples
		1759
		1760	There are a number of principal ways to embed other event loops or modules
		1761	into libev. Here are some ideas on how to include libadns into libev
		1762	(there is a Perl module named C<EV::ADNS> that does this, which you could
		1763	use for an actually working example. Another Perl module named C<EV::Glib>
		1764	embeds a Glib main context into libev, and finally, C<Glib::EV> embeds EV
		1765	into the Glib event loop).
		1766
		1767	Method 1: Add IO watchers and a timeout watcher in a prepare handler,
		1768	and in a check watcher, destroy them and call into libadns. What follows
		1769	is pseudo-code only of course. This requires you to either use a low
		1770	priority for the check watcher or use C<ev_clear_pending> explicitly, as
		1771	the callbacks for the IO/timeout watchers might not have been called yet.
		1772
		1773	static ev_io iow [nfd];
		1774	static ev_timer tw;
		1775
		1776	static void
		1777	io_cb (ev_loop *loop, ev_io *w, int revents)
		1778	{
		1779	}
		1780
		1781	// create io watchers for each fd and a timer before blocking
		1782	static void
		1783	adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents)
		1784	{
		1785	int timeout = 3600000;
		1786	struct pollfd fds [nfd];
		1787	// actual code will need to loop here and realloc etc.
		1788	adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));
		1789
		1790	/* the callback is illegal, but won't be called as we stop during check */
		1791	ev_timer_init (&tw, 0, timeout * 1e-3);
		1792	ev_timer_start (loop, &tw);
		1793
		1794	// create one ev_io per pollfd
		1795	for (int i = 0; i < nfd; ++i)
		1796	{
		1797	ev_io_init (iow + i, io_cb, fds [i].fd,
		1798	((fds [i].events & POLLIN ? EV_READ : 0)
		1799	| (fds [i].events & POLLOUT ? EV_WRITE : 0)));
		1800
		1801	fds [i].revents = 0;
		1802	ev_io_start (loop, iow + i);
		1803	}
		1804	}
		1805
		1806	// stop all watchers after blocking
		1807	static void
		1808	adns_check_cb (ev_loop *loop, ev_check *w, int revents)
		1809	{
		1810	ev_timer_stop (loop, &tw);
		1811
		1812	for (int i = 0; i < nfd; ++i)
		1813	{
		1814	// set the relevant poll flags
		1815	// could also call adns_processreadable etc. here
		1816	struct pollfd *fd = fds + i;
		1817	int revents = ev_clear_pending (iow + i);
		1818	if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;
		1819	if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;
		1820
		1821	// now stop the watcher
		1822	ev_io_stop (loop, iow + i);
		1823	}
		1824
		1825	adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));
		1826	}
		1827
		1828	Method 2: This would be just like method 1, but you run C<adns_afterpoll>
		1829	in the prepare watcher and would dispose of the check watcher.
		1830
		1831	Method 3: If the module to be embedded supports explicit event
		1832	notification (adns does), you can also make use of the actual watcher
		1833	callbacks, and only destroy/create the watchers in the prepare watcher.
		1834
		1835	static void
		1836	timer_cb (EV_P_ ev_timer *w, int revents)
		1837	{
		1838	adns_state ads = (adns_state)w->data;
		1839	update_now (EV_A);
		1840
		1841	adns_processtimeouts (ads, &tv_now);
		1842	}
		1843
		1844	static void
		1845	io_cb (EV_P_ ev_io *w, int revents)
		1846	{
		1847	adns_state ads = (adns_state)w->data;
		1848	update_now (EV_A);
		1849
		1850	if (revents & EV_READ ) adns_processreadable (ads, w->fd, &tv_now);
		1851	if (revents & EV_WRITE) adns_processwriteable (ads, w->fd, &tv_now);
		1852	}
		1853
		1854	// do not ever call adns_afterpoll
		1855
		1856	Method 4: Do not use a prepare or check watcher because the module you
		1857	want to embed is too inflexible to support it. Instead, youc na override
		1858	their poll function. The drawback with this solution is that the main
		1859	loop is now no longer controllable by EV. The C<Glib::EV> module does
		1860	this.
		1861
		1862	static gint
		1863	event_poll_func (GPollFD *fds, guint nfds, gint timeout)
		1864	{
		1865	int got_events = 0;
		1866
		1867	for (n = 0; n < nfds; ++n)
		1868	// create/start io watcher that sets the relevant bits in fds[n] and increment got_events
		1869
		1870	if (timeout >= 0)
		1871	// create/start timer
		1872
		1873	// poll
		1874	ev_loop (EV_A_ 0);
		1875
		1876	// stop timer again
		1877	if (timeout >= 0)
		1878	ev_timer_stop (EV_A_ &to);
		1879
		1880	// stop io watchers again - their callbacks should have set
		1881	for (n = 0; n < nfds; ++n)
		1882	ev_io_stop (EV_A_ iow [n]);
		1883
		1884	return got_events;
		1885	}
		1886
		1887
		1888	=head2 C<ev_embed> - when one backend isn't enough...
		1889
		1890	This is a rather advanced watcher type that lets you embed one event loop
		1891	into another (currently only C<ev_io> events are supported in the embedded
		1892	loop, other types of watchers might be handled in a delayed or incorrect
		1893	fashion and must not be used).
		1894
		1895	There are primarily two reasons you would want that: work around bugs and
		1896	prioritise I/O.
		1897
		1898	As an example for a bug workaround, the kqueue backend might only support
		1899	sockets on some platform, so it is unusable as generic backend, but you
		1900	still want to make use of it because you have many sockets and it scales
		1901	so nicely. In this case, you would create a kqueue-based loop and embed it
		1902	into your default loop (which might use e.g. poll). Overall operation will
		1903	be a bit slower because first libev has to poll and then call kevent, but
		1904	at least you can use both at what they are best.
		1905
		1906	As for prioritising I/O: rarely you have the case where some fds have
		1907	to be watched and handled very quickly (with low latency), and even
		1908	priorities and idle watchers might have too much overhead. In this case
		1909	you would put all the high priority stuff in one loop and all the rest in
		1910	a second one, and embed the second one in the first.
		1911
		1912	As long as the watcher is active, the callback will be invoked every time
		1913	there might be events pending in the embedded loop. The callback must then
		1914	call C<ev_embed_sweep (mainloop, watcher)> to make a single sweep and invoke
		1915	their callbacks (you could also start an idle watcher to give the embedded
		1916	loop strictly lower priority for example). You can also set the callback
		1917	to C<0>, in which case the embed watcher will automatically execute the
		1918	embedded loop sweep.
		1919
		1920	As long as the watcher is started it will automatically handle events. The
		1921	callback will be invoked whenever some events have been handled. You can
		1922	set the callback to C<0> to avoid having to specify one if you are not
		1923	interested in that.
		1924
		1925	Also, there have not currently been made special provisions for forking:
		1926	when you fork, you not only have to call C<ev_loop_fork> on both loops,
		1927	but you will also have to stop and restart any C<ev_embed> watchers
		1928	yourself.
		1929
		1930	Unfortunately, not all backends are embeddable, only the ones returned by
		1931	C<ev_embeddable_backends> are, which, unfortunately, does not include any
		1932	portable one.
		1933
		1934	So when you want to use this feature you will always have to be prepared
		1935	that you cannot get an embeddable loop. The recommended way to get around
		1936	this is to have a separate variables for your embeddable loop, try to
		1937	create it, and if that fails, use the normal loop for everything.
		1938
		1939	=head3 Watcher-Specific Functions and Data Members
		1940
		1941	=over 4
		1942
		1943	=item ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)
		1944
		1945	=item ev_embed_set (ev_embed *, callback, struct ev_loop *embedded_loop)
		1946
		1947	Configures the watcher to embed the given loop, which must be
		1948	embeddable. If the callback is C<0>, then C<ev_embed_sweep> will be
		1949	invoked automatically, otherwise it is the responsibility of the callback
		1950	to invoke it (it will continue to be called until the sweep has been done,
		1951	if you do not want thta, you need to temporarily stop the embed watcher).
		1952
		1953	=item ev_embed_sweep (loop, ev_embed *)
		1954
		1955	Make a single, non-blocking sweep over the embedded loop. This works
		1956	similarly to C<ev_loop (embedded_loop, EVLOOP_NONBLOCK)>, but in the most
		1957	apropriate way for embedded loops.
		1958
		1959	=item struct ev_loop *other [read-only]
		1960
		1961	The embedded event loop.
		1962
		1963	=back
		1964
		1965	=head3 Examples
		1966
		1967	Example: Try to get an embeddable event loop and embed it into the default
		1968	event loop. If that is not possible, use the default loop. The default
		1969	loop is stored in C<loop_hi>, while the mebeddable loop is stored in
		1970	C<loop_lo> (which is C<loop_hi> in the acse no embeddable loop can be
		1971	used).
		1972
		1973	struct ev_loop *loop_hi = ev_default_init (0);
		1974	struct ev_loop *loop_lo = 0;
		1975	struct ev_embed embed;
		1976
		1977	// see if there is a chance of getting one that works
		1978	// (remember that a flags value of 0 means autodetection)
		1979	loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
		1980	? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
		1981	: 0;
		1982
		1983	// if we got one, then embed it, otherwise default to loop_hi
		1984	if (loop_lo)
		1985	{
		1986	ev_embed_init (&embed, 0, loop_lo);
		1987	ev_embed_start (loop_hi, &embed);
		1988	}
		1989	else
		1990	loop_lo = loop_hi;
		1991
		1992	Example: Check if kqueue is available but not recommended and create
		1993	a kqueue backend for use with sockets (which usually work with any
		1994	kqueue implementation). Store the kqueue/socket-only event loop in
		1995	C<loop_socket>. (One might optionally use C<EVFLAG_NOENV>, too).
		1996
		1997	struct ev_loop *loop = ev_default_init (0);
		1998	struct ev_loop *loop_socket = 0;
		1999	struct ev_embed embed;
		2000
		2001	if (ev_supported_backends () & ~ev_recommended_backends () & EVBACKEND_KQUEUE)
		2002	if ((loop_socket = ev_loop_new (EVBACKEND_KQUEUE))
		2003	{
		2004	ev_embed_init (&embed, 0, loop_socket);
		2005	ev_embed_start (loop, &embed);
		2006	}
		2007
		2008	if (!loop_socket)
		2009	loop_socket = loop;
		2010
		2011	// now use loop_socket for all sockets, and loop for everything else
		2012
		2013
		2014	=head2 C<ev_fork> - the audacity to resume the event loop after a fork
		2015
		2016	Fork watchers are called when a C<fork ()> was detected (usually because
		2017	whoever is a good citizen cared to tell libev about it by calling
		2018	C<ev_default_fork> or C<ev_loop_fork>). The invocation is done before the
		2019	event loop blocks next and before C<ev_check> watchers are being called,
		2020	and only in the child after the fork. If whoever good citizen calling
		2021	C<ev_default_fork> cheats and calls it in the wrong process, the fork
		2022	handlers will be invoked, too, of course.
		2023
		2024	=head3 Watcher-Specific Functions and Data Members
		2025
		2026	=over 4
		2027
		2028	=item ev_fork_init (ev_signal *, callback)
		2029
		2030	Initialises and configures the fork watcher - it has no parameters of any
		2031	kind. There is a C<ev_fork_set> macro, but using it is utterly pointless,
		2032	believe me.
		2033
		2034	=back
		2035
		2036
690	=head1 OTHER FUNCTIONS	2037	=head1 OTHER FUNCTIONS
691		2038
692	There are some other fucntions of possible interest. Described. Here. Now.	2039	There are some other functions of possible interest. Described. Here. Now.
693		2040
694	=over 4	2041	=over 4
695		2042
696	=item ev_once (loop, int fd, int events, ev_tstamp timeout, callback)	2043	=item ev_once (loop, int fd, int events, ev_tstamp timeout, callback)
697		2044
698	This function combines a simple timer and an I/O watcher, calls your	2045	This function combines a simple timer and an I/O watcher, calls your
699	callback on whichever event happens first and automatically stop both	2046	callback on whichever event happens first and automatically stop both
700	watchers. This is useful if you want to wait for a single event on an fd	2047	watchers. This is useful if you want to wait for a single event on an fd
701	or timeout without havign to allocate/configure/start/stop/free one or	2048	or timeout without having to allocate/configure/start/stop/free one or
702	more watchers yourself.	2049	more watchers yourself.
703		2050
704	If C<fd> is less than 0, then no I/O watcher will be started and events is	2051	If C<fd> is less than 0, then no I/O watcher will be started and events
705	ignored. Otherwise, an ev_io watcher for the given C<fd> and C<events> set	2052	is being ignored. Otherwise, an C<ev_io> watcher for the given C<fd> and
706	will be craeted and started.	2053	C<events> set will be craeted and started.
707		2054
708	If C<timeout> is less than 0, then no timeout watcher will be	2055	If C<timeout> is less than 0, then no timeout watcher will be
709	started. Otherwise an ev_timer watcher with after = C<timeout> (and repeat	2056	started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and
710	= 0) will be started.	2057	repeat = 0) will be started. While C<0> is a valid timeout, it is of
		2058	dubious value.
711		2059
712	The callback has the type C<void (*cb)(int revents, void *arg)> and	2060	The callback has the type C<void (*cb)(int revents, void *arg)> and gets
713	gets passed an events set (normally a combination of EV_ERROR, EV_READ,	2061	passed an C<revents> set like normal event callbacks (a combination of
714	EV_WRITE or EV_TIMEOUT) and the C<arg> value passed to C<ev_once>:	2062	C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMEOUT>) and the C<arg>
		2063	value passed to C<ev_once>:
715		2064
716	static void stdin_ready (int revents, void *arg)	2065	static void stdin_ready (int revents, void *arg)
717	{	2066	{
718	if (revents & EV_TIMEOUT)	2067	if (revents & EV_TIMEOUT)
719	/* doh, nothing entered */	2068	/* doh, nothing entered */;
720	else if (revents & EV_READ)	2069	else if (revents & EV_READ)
721	/* stdin might have data for us, joy! */	2070	/* stdin might have data for us, joy! */;
722	}	2071	}
723		2072
724	ev_once (STDIN_FILENO, EV_READm 10., stdin_ready, 0);	2073	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
725		2074
726	=item ev_feed_event (loop, watcher, int events)	2075	=item ev_feed_event (ev_loop *, watcher *, int revents)
727		2076
728	Feeds the given event set into the event loop, as if the specified event	2077	Feeds the given event set into the event loop, as if the specified event
729	has happened for the specified watcher (which must be a pointer to an	2078	had happened for the specified watcher (which must be a pointer to an
730	initialised but not necessarily active event watcher).	2079	initialised but not necessarily started event watcher).
731		2080
732	=item ev_feed_fd_event (loop, int fd, int revents)	2081	=item ev_feed_fd_event (ev_loop *, int fd, int revents)
733		2082
734	Feed an event on the given fd, as if a file descriptor backend detected it.	2083	Feed an event on the given fd, as if a file descriptor backend detected
		2084	the given events it.
735		2085
736	=item ev_feed_signal_event (loop, int signum)	2086	=item ev_feed_signal_event (ev_loop *loop, int signum)
737		2087
738	Feed an event as if the given signal occured (loop must be the default loop!).	2088	Feed an event as if the given signal occured (C<loop> must be the default
		2089	loop!).
739		2090
740	=back	2091	=back
741		2092
		2093
		2094	=head1 LIBEVENT EMULATION
		2095
		2096	Libev offers a compatibility emulation layer for libevent. It cannot
		2097	emulate the internals of libevent, so here are some usage hints:
		2098
		2099	=over 4
		2100
		2101	=item * Use it by including <event.h>, as usual.
		2102
		2103	=item * The following members are fully supported: ev_base, ev_callback,
		2104	ev_arg, ev_fd, ev_res, ev_events.
		2105
		2106	=item * Avoid using ev_flags and the EVLIST_*-macros, while it is
		2107	maintained by libev, it does not work exactly the same way as in libevent (consider
		2108	it a private API).
		2109
		2110	=item * Priorities are not currently supported. Initialising priorities
		2111	will fail and all watchers will have the same priority, even though there
		2112	is an ev_pri field.
		2113
		2114	=item * Other members are not supported.
		2115
		2116	=item * The libev emulation is I<not> ABI compatible to libevent, you need
		2117	to use the libev header file and library.
		2118
		2119	=back
		2120
		2121	=head1 C++ SUPPORT
		2122
		2123	Libev comes with some simplistic wrapper classes for C++ that mainly allow
		2124	you to use some convinience methods to start/stop watchers and also change
		2125	the callback model to a model using method callbacks on objects.
		2126
		2127	To use it,
		2128
		2129	#include <ev++.h>
		2130
		2131	This automatically includes F<ev.h> and puts all of its definitions (many
		2132	of them macros) into the global namespace. All C++ specific things are
		2133	put into the C<ev> namespace. It should support all the same embedding
		2134	options as F<ev.h>, most notably C<EV_MULTIPLICITY>.
		2135
		2136	Care has been taken to keep the overhead low. The only data member the C++
		2137	classes add (compared to plain C-style watchers) is the event loop pointer
		2138	that the watcher is associated with (or no additional members at all if
		2139	you disable C<EV_MULTIPLICITY> when embedding libev).
		2140
		2141	Currently, functions, and static and non-static member functions can be
		2142	used as callbacks. Other types should be easy to add as long as they only
		2143	need one additional pointer for context. If you need support for other
		2144	types of functors please contact the author (preferably after implementing
		2145	it).
		2146
		2147	Here is a list of things available in the C<ev> namespace:
		2148
		2149	=over 4
		2150
		2151	=item C<ev::READ>, C<ev::WRITE> etc.
		2152
		2153	These are just enum values with the same values as the C<EV_READ> etc.
		2154	macros from F<ev.h>.
		2155
		2156	=item C<ev::tstamp>, C<ev::now>
		2157
		2158	Aliases to the same types/functions as with the C<ev_> prefix.
		2159
		2160	=item C<ev::io>, C<ev::timer>, C<ev::periodic>, C<ev::idle>, C<ev::sig> etc.
		2161
		2162	For each C<ev_TYPE> watcher in F<ev.h> there is a corresponding class of
		2163	the same name in the C<ev> namespace, with the exception of C<ev_signal>
		2164	which is called C<ev::sig> to avoid clashes with the C<signal> macro
		2165	defines by many implementations.
		2166
		2167	All of those classes have these methods:
		2168
		2169	=over 4
		2170
		2171	=item ev::TYPE::TYPE ()
		2172
		2173	=item ev::TYPE::TYPE (struct ev_loop *)
		2174
		2175	=item ev::TYPE::~TYPE
		2176
		2177	The constructor (optionally) takes an event loop to associate the watcher
		2178	with. If it is omitted, it will use C<EV_DEFAULT>.
		2179
		2180	The constructor calls C<ev_init> for you, which means you have to call the
		2181	C<set> method before starting it.
		2182
		2183	It will not set a callback, however: You have to call the templated C<set>
		2184	method to set a callback before you can start the watcher.
		2185
		2186	(The reason why you have to use a method is a limitation in C++ which does
		2187	not allow explicit template arguments for constructors).
		2188
		2189	The destructor automatically stops the watcher if it is active.
		2190
		2191	=item w->set<class, &class::method> (object *)
		2192
		2193	This method sets the callback method to call. The method has to have a
		2194	signature of C<void (*)(ev_TYPE &, int)>, it receives the watcher as
		2195	first argument and the C<revents> as second. The object must be given as
		2196	parameter and is stored in the C<data> member of the watcher.
		2197
		2198	This method synthesizes efficient thunking code to call your method from
		2199	the C callback that libev requires. If your compiler can inline your
		2200	callback (i.e. it is visible to it at the place of the C<set> call and
		2201	your compiler is good :), then the method will be fully inlined into the
		2202	thunking function, making it as fast as a direct C callback.
		2203
		2204	Example: simple class declaration and watcher initialisation
		2205
		2206	struct myclass
		2207	{
		2208	void io_cb (ev::io &w, int revents) { }
		2209	}
		2210
		2211	myclass obj;
		2212	ev::io iow;
		2213	iow.set <myclass, &myclass::io_cb> (&obj);
		2214
		2215	=item w->set<function> (void *data = 0)
		2216
		2217	Also sets a callback, but uses a static method or plain function as
		2218	callback. The optional C<data> argument will be stored in the watcher's
		2219	C<data> member and is free for you to use.
		2220
		2221	The prototype of the C<function> must be C<void (*)(ev::TYPE &w, int)>.
		2222
		2223	See the method-C<set> above for more details.
		2224
		2225	Example:
		2226
		2227	static void io_cb (ev::io &w, int revents) { }
		2228	iow.set <io_cb> ();
		2229
		2230	=item w->set (struct ev_loop *)
		2231
		2232	Associates a different C<struct ev_loop> with this watcher. You can only
		2233	do this when the watcher is inactive (and not pending either).
		2234
		2235	=item w->set ([args])
		2236
		2237	Basically the same as C<ev_TYPE_set>, with the same args. Must be
		2238	called at least once. Unlike the C counterpart, an active watcher gets
		2239	automatically stopped and restarted when reconfiguring it with this
		2240	method.
		2241
		2242	=item w->start ()
		2243
		2244	Starts the watcher. Note that there is no C<loop> argument, as the
		2245	constructor already stores the event loop.
		2246
		2247	=item w->stop ()
		2248
		2249	Stops the watcher if it is active. Again, no C<loop> argument.
		2250
		2251	=item w->again () (C<ev::timer>, C<ev::periodic> only)
		2252
		2253	For C<ev::timer> and C<ev::periodic>, this invokes the corresponding
		2254	C<ev_TYPE_again> function.
		2255
		2256	=item w->sweep () (C<ev::embed> only)
		2257
		2258	Invokes C<ev_embed_sweep>.
		2259
		2260	=item w->update () (C<ev::stat> only)
		2261
		2262	Invokes C<ev_stat_stat>.
		2263
		2264	=back
		2265
		2266	=back
		2267
		2268	Example: Define a class with an IO and idle watcher, start one of them in
		2269	the constructor.
		2270
		2271	class myclass
		2272	{
		2273	ev_io io; void io_cb (ev::io &w, int revents);
		2274	ev_idle idle void idle_cb (ev::idle &w, int revents);
		2275
		2276	myclass ();
		2277	}
		2278
		2279	myclass::myclass (int fd)
		2280	{
		2281	io .set <myclass, &myclass::io_cb > (this);
		2282	idle.set <myclass, &myclass::idle_cb> (this);
		2283
		2284	io.start (fd, ev::READ);
		2285	}
		2286
		2287
		2288	=head1 MACRO MAGIC
		2289
		2290	Libev can be compiled with a variety of options, the most fundamantal
		2291	of which is C<EV_MULTIPLICITY>. This option determines whether (most)
		2292	functions and callbacks have an initial C<struct ev_loop *> argument.
		2293
		2294	To make it easier to write programs that cope with either variant, the
		2295	following macros are defined:
		2296
		2297	=over 4
		2298
		2299	=item C<EV_A>, C<EV_A_>
		2300
		2301	This provides the loop I<argument> for functions, if one is required ("ev
		2302	loop argument"). The C<EV_A> form is used when this is the sole argument,
		2303	C<EV_A_> is used when other arguments are following. Example:
		2304
		2305	ev_unref (EV_A);
		2306	ev_timer_add (EV_A_ watcher);
		2307	ev_loop (EV_A_ 0);
		2308
		2309	It assumes the variable C<loop> of type C<struct ev_loop *> is in scope,
		2310	which is often provided by the following macro.
		2311
		2312	=item C<EV_P>, C<EV_P_>
		2313
		2314	This provides the loop I<parameter> for functions, if one is required ("ev
		2315	loop parameter"). The C<EV_P> form is used when this is the sole parameter,
		2316	C<EV_P_> is used when other parameters are following. Example:
		2317
		2318	// this is how ev_unref is being declared
		2319	static void ev_unref (EV_P);
		2320
		2321	// this is how you can declare your typical callback
		2322	static void cb (EV_P_ ev_timer *w, int revents)
		2323
		2324	It declares a parameter C<loop> of type C<struct ev_loop *>, quite
		2325	suitable for use with C<EV_A>.
		2326
		2327	=item C<EV_DEFAULT>, C<EV_DEFAULT_>
		2328
		2329	Similar to the other two macros, this gives you the value of the default
		2330	loop, if multiple loops are supported ("ev loop default").
		2331
		2332	=back
		2333
		2334	Example: Declare and initialise a check watcher, utilising the above
		2335	macros so it will work regardless of whether multiple loops are supported
		2336	or not.
		2337
		2338	static void
		2339	check_cb (EV_P_ ev_timer *w, int revents)
		2340	{
		2341	ev_check_stop (EV_A_ w);
		2342	}
		2343
		2344	ev_check check;
		2345	ev_check_init (&check, check_cb);
		2346	ev_check_start (EV_DEFAULT_ &check);
		2347	ev_loop (EV_DEFAULT_ 0);
		2348
		2349	=head1 EMBEDDING
		2350
		2351	Libev can (and often is) directly embedded into host
		2352	applications. Examples of applications that embed it include the Deliantra
		2353	Game Server, the EV perl module, the GNU Virtual Private Ethernet (gvpe)
		2354	and rxvt-unicode.
		2355
		2356	The goal is to enable you to just copy the necessary files into your
		2357	source directory without having to change even a single line in them, so
		2358	you can easily upgrade by simply copying (or having a checked-out copy of
		2359	libev somewhere in your source tree).
		2360
		2361	=head2 FILESETS
		2362
		2363	Depending on what features you need you need to include one or more sets of files
		2364	in your app.
		2365
		2366	=head3 CORE EVENT LOOP
		2367
		2368	To include only the libev core (all the C<ev_*> functions), with manual
		2369	configuration (no autoconf):
		2370
		2371	#define EV_STANDALONE 1
		2372	#include "ev.c"
		2373
		2374	This will automatically include F<ev.h>, too, and should be done in a
		2375	single C source file only to provide the function implementations. To use
		2376	it, do the same for F<ev.h> in all files wishing to use this API (best
		2377	done by writing a wrapper around F<ev.h> that you can include instead and
		2378	where you can put other configuration options):
		2379
		2380	#define EV_STANDALONE 1
		2381	#include "ev.h"
		2382
		2383	Both header files and implementation files can be compiled with a C++
		2384	compiler (at least, thats a stated goal, and breakage will be treated
		2385	as a bug).
		2386
		2387	You need the following files in your source tree, or in a directory
		2388	in your include path (e.g. in libev/ when using -Ilibev):
		2389
		2390	ev.h
		2391	ev.c
		2392	ev_vars.h
		2393	ev_wrap.h
		2394
		2395	ev_win32.c required on win32 platforms only
		2396
		2397	ev_select.c only when select backend is enabled (which is enabled by default)
		2398	ev_poll.c only when poll backend is enabled (disabled by default)
		2399	ev_epoll.c only when the epoll backend is enabled (disabled by default)
		2400	ev_kqueue.c only when the kqueue backend is enabled (disabled by default)
		2401	ev_port.c only when the solaris port backend is enabled (disabled by default)
		2402
		2403	F<ev.c> includes the backend files directly when enabled, so you only need
		2404	to compile this single file.
		2405
		2406	=head3 LIBEVENT COMPATIBILITY API
		2407
		2408	To include the libevent compatibility API, also include:
		2409
		2410	#include "event.c"
		2411
		2412	in the file including F<ev.c>, and:
		2413
		2414	#include "event.h"
		2415
		2416	in the files that want to use the libevent API. This also includes F<ev.h>.
		2417
		2418	You need the following additional files for this:
		2419
		2420	event.h
		2421	event.c
		2422
		2423	=head3 AUTOCONF SUPPORT
		2424
		2425	Instead of using C<EV_STANDALONE=1> and providing your config in
		2426	whatever way you want, you can also C<m4_include([libev.m4])> in your
		2427	F<configure.ac> and leave C<EV_STANDALONE> undefined. F<ev.c> will then
		2428	include F<config.h> and configure itself accordingly.
		2429
		2430	For this of course you need the m4 file:
		2431
		2432	libev.m4
		2433
		2434	=head2 PREPROCESSOR SYMBOLS/MACROS
		2435
		2436	Libev can be configured via a variety of preprocessor symbols you have to define
		2437	before including any of its files. The default is not to build for multiplicity
		2438	and only include the select backend.
		2439
		2440	=over 4
		2441
		2442	=item EV_STANDALONE
		2443
		2444	Must always be C<1> if you do not use autoconf configuration, which
		2445	keeps libev from including F<config.h>, and it also defines dummy
		2446	implementations for some libevent functions (such as logging, which is not
		2447	supported). It will also not define any of the structs usually found in
		2448	F<event.h> that are not directly supported by the libev core alone.
		2449
		2450	=item EV_USE_MONOTONIC
		2451
		2452	If defined to be C<1>, libev will try to detect the availability of the
		2453	monotonic clock option at both compiletime and runtime. Otherwise no use
		2454	of the monotonic clock option will be attempted. If you enable this, you
		2455	usually have to link against librt or something similar. Enabling it when
		2456	the functionality isn't available is safe, though, although you have
		2457	to make sure you link against any libraries where the C<clock_gettime>
		2458	function is hiding in (often F<-lrt>).
		2459
		2460	=item EV_USE_REALTIME
		2461
		2462	If defined to be C<1>, libev will try to detect the availability of the
		2463	realtime clock option at compiletime (and assume its availability at
		2464	runtime if successful). Otherwise no use of the realtime clock option will
		2465	be attempted. This effectively replaces C<gettimeofday> by C<clock_get
		2466	(CLOCK_REALTIME, ...)> and will not normally affect correctness. See the
		2467	note about libraries in the description of C<EV_USE_MONOTONIC>, though.
		2468
		2469	=item EV_USE_NANOSLEEP
		2470
		2471	If defined to be C<1>, libev will assume that C<nanosleep ()> is available
		2472	and will use it for delays. Otherwise it will use C<select ()>.
		2473
		2474	=item EV_USE_SELECT
		2475
		2476	If undefined or defined to be C<1>, libev will compile in support for the
		2477	C<select>(2) backend. No attempt at autodetection will be done: if no
		2478	other method takes over, select will be it. Otherwise the select backend
		2479	will not be compiled in.
		2480
		2481	=item EV_SELECT_USE_FD_SET
		2482
		2483	If defined to C<1>, then the select backend will use the system C<fd_set>
		2484	structure. This is useful if libev doesn't compile due to a missing
		2485	C<NFDBITS> or C<fd_mask> definition or it misguesses the bitset layout on
		2486	exotic systems. This usually limits the range of file descriptors to some
		2487	low limit such as 1024 or might have other limitations (winsocket only
		2488	allows 64 sockets). The C<FD_SETSIZE> macro, set before compilation, might
		2489	influence the size of the C<fd_set> used.
		2490
		2491	=item EV_SELECT_IS_WINSOCKET
		2492
		2493	When defined to C<1>, the select backend will assume that
		2494	select/socket/connect etc. don't understand file descriptors but
		2495	wants osf handles on win32 (this is the case when the select to
		2496	be used is the winsock select). This means that it will call
		2497	C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,
		2498	it is assumed that all these functions actually work on fds, even
		2499	on win32. Should not be defined on non-win32 platforms.
		2500
		2501	=item EV_FD_TO_WIN32_HANDLE
		2502
		2503	If C<EV_SELECT_IS_WINSOCKET> is enabled, then libev needs a way to map
		2504	file descriptors to socket handles. When not defining this symbol (the
		2505	default), then libev will call C<_get_osfhandle>, which is usually
		2506	correct. In some cases, programs use their own file descriptor management,
		2507	in which case they can provide this function to map fds to socket handles.
		2508
		2509	=item EV_USE_POLL
		2510
		2511	If defined to be C<1>, libev will compile in support for the C<poll>(2)
		2512	backend. Otherwise it will be enabled on non-win32 platforms. It
		2513	takes precedence over select.
		2514
		2515	=item EV_USE_EPOLL
		2516
		2517	If defined to be C<1>, libev will compile in support for the Linux
		2518	C<epoll>(7) backend. Its availability will be detected at runtime,
		2519	otherwise another method will be used as fallback. This is the
		2520	preferred backend for GNU/Linux systems.
		2521
		2522	=item EV_USE_KQUEUE
		2523
		2524	If defined to be C<1>, libev will compile in support for the BSD style
		2525	C<kqueue>(2) backend. Its actual availability will be detected at runtime,
		2526	otherwise another method will be used as fallback. This is the preferred
		2527	backend for BSD and BSD-like systems, although on most BSDs kqueue only
		2528	supports some types of fds correctly (the only platform we found that
		2529	supports ptys for example was NetBSD), so kqueue might be compiled in, but
		2530	not be used unless explicitly requested. The best way to use it is to find
		2531	out whether kqueue supports your type of fd properly and use an embedded
		2532	kqueue loop.
		2533
		2534	=item EV_USE_PORT
		2535
		2536	If defined to be C<1>, libev will compile in support for the Solaris
		2537	10 port style backend. Its availability will be detected at runtime,
		2538	otherwise another method will be used as fallback. This is the preferred
		2539	backend for Solaris 10 systems.
		2540
		2541	=item EV_USE_DEVPOLL
		2542
		2543	reserved for future expansion, works like the USE symbols above.
		2544
		2545	=item EV_USE_INOTIFY
		2546
		2547	If defined to be C<1>, libev will compile in support for the Linux inotify
		2548	interface to speed up C<ev_stat> watchers. Its actual availability will
		2549	be detected at runtime.
		2550
		2551	=item EV_H
		2552
		2553	The name of the F<ev.h> header file used to include it. The default if
		2554	undefined is C<"ev.h"> in F<event.h> and F<ev.c>. This can be used to
		2555	virtually rename the F<ev.h> header file in case of conflicts.
		2556
		2557	=item EV_CONFIG_H
		2558
		2559	If C<EV_STANDALONE> isn't C<1>, this variable can be used to override
		2560	F<ev.c>'s idea of where to find the F<config.h> file, similarly to
		2561	C<EV_H>, above.
		2562
		2563	=item EV_EVENT_H
		2564
		2565	Similarly to C<EV_H>, this macro can be used to override F<event.c>'s idea
		2566	of how the F<event.h> header can be found, the dfeault is C<"event.h">.
		2567
		2568	=item EV_PROTOTYPES
		2569
		2570	If defined to be C<0>, then F<ev.h> will not define any function
		2571	prototypes, but still define all the structs and other symbols. This is
		2572	occasionally useful if you want to provide your own wrapper functions
		2573	around libev functions.
		2574
		2575	=item EV_MULTIPLICITY
		2576
		2577	If undefined or defined to C<1>, then all event-loop-specific functions
		2578	will have the C<struct ev_loop *> as first argument, and you can create
		2579	additional independent event loops. Otherwise there will be no support
		2580	for multiple event loops and there is no first event loop pointer
		2581	argument. Instead, all functions act on the single default loop.
		2582
		2583	=item EV_MINPRI
		2584
		2585	=item EV_MAXPRI
		2586
		2587	The range of allowed priorities. C<EV_MINPRI> must be smaller or equal to
		2588	C<EV_MAXPRI>, but otherwise there are no non-obvious limitations. You can
		2589	provide for more priorities by overriding those symbols (usually defined
		2590	to be C<-2> and C<2>, respectively).
		2591
		2592	When doing priority-based operations, libev usually has to linearly search
		2593	all the priorities, so having many of them (hundreds) uses a lot of space
		2594	and time, so using the defaults of five priorities (-2 .. +2) is usually
		2595	fine.
		2596
		2597	If your embedding app does not need any priorities, defining these both to
		2598	C<0> will save some memory and cpu.
		2599
		2600	=item EV_PERIODIC_ENABLE
		2601
		2602	If undefined or defined to be C<1>, then periodic timers are supported. If
		2603	defined to be C<0>, then they are not. Disabling them saves a few kB of
		2604	code.
		2605
		2606	=item EV_IDLE_ENABLE
		2607
		2608	If undefined or defined to be C<1>, then idle watchers are supported. If
		2609	defined to be C<0>, then they are not. Disabling them saves a few kB of
		2610	code.
		2611
		2612	=item EV_EMBED_ENABLE
		2613
		2614	If undefined or defined to be C<1>, then embed watchers are supported. If
		2615	defined to be C<0>, then they are not.
		2616
		2617	=item EV_STAT_ENABLE
		2618
		2619	If undefined or defined to be C<1>, then stat watchers are supported. If
		2620	defined to be C<0>, then they are not.
		2621
		2622	=item EV_FORK_ENABLE
		2623
		2624	If undefined or defined to be C<1>, then fork watchers are supported. If
		2625	defined to be C<0>, then they are not.
		2626
		2627	=item EV_MINIMAL
		2628
		2629	If you need to shave off some kilobytes of code at the expense of some
		2630	speed, define this symbol to C<1>. Currently only used for gcc to override
		2631	some inlining decisions, saves roughly 30% codesize of amd64.
		2632
		2633	=item EV_PID_HASHSIZE
		2634
		2635	C<ev_child> watchers use a small hash table to distribute workload by
		2636	pid. The default size is C<16> (or C<1> with C<EV_MINIMAL>), usually more
		2637	than enough. If you need to manage thousands of children you might want to
		2638	increase this value (I<must> be a power of two).
		2639
		2640	=item EV_INOTIFY_HASHSIZE
		2641
		2642	C<ev_stat> watchers use a small hash table to distribute workload by
		2643	inotify watch id. The default size is C<16> (or C<1> with C<EV_MINIMAL>),
		2644	usually more than enough. If you need to manage thousands of C<ev_stat>
		2645	watchers you might want to increase this value (I<must> be a power of
		2646	two).
		2647
		2648	=item EV_COMMON
		2649
		2650	By default, all watchers have a C<void *data> member. By redefining
		2651	this macro to a something else you can include more and other types of
		2652	members. You have to define it each time you include one of the files,
		2653	though, and it must be identical each time.
		2654
		2655	For example, the perl EV module uses something like this:
		2656
		2657	#define EV_COMMON \
		2658	SV *self; /* contains this struct */ \
		2659	SV *cb_sv, *fh /* note no trailing ";" */
		2660
		2661	=item EV_CB_DECLARE (type)
		2662
		2663	=item EV_CB_INVOKE (watcher, revents)
		2664
		2665	=item ev_set_cb (ev, cb)
		2666
		2667	Can be used to change the callback member declaration in each watcher,
		2668	and the way callbacks are invoked and set. Must expand to a struct member
		2669	definition and a statement, respectively. See the F<ev.h> header file for
		2670	their default definitions. One possible use for overriding these is to
		2671	avoid the C<struct ev_loop *> as first argument in all cases, or to use
		2672	method calls instead of plain function calls in C++.
		2673
		2674	=head2 EXPORTED API SYMBOLS
		2675
		2676	If you need to re-export the API (e.g. via a dll) and you need a list of
		2677	exported symbols, you can use the provided F<Symbol.*> files which list
		2678	all public symbols, one per line:
		2679
		2680	Symbols.ev for libev proper
		2681	Symbols.event for the libevent emulation
		2682
		2683	This can also be used to rename all public symbols to avoid clashes with
		2684	multiple versions of libev linked together (which is obviously bad in
		2685	itself, but sometimes it is inconvinient to avoid this).
		2686
		2687	A sed command like this will create wrapper C<#define>'s that you need to
		2688	include before including F<ev.h>:
		2689
		2690	<Symbols.ev sed -e "s/.*/#define & myprefix_&/" >wrap.h
		2691
		2692	This would create a file F<wrap.h> which essentially looks like this:
		2693
		2694	#define ev_backend myprefix_ev_backend
		2695	#define ev_check_start myprefix_ev_check_start
		2696	#define ev_check_stop myprefix_ev_check_stop
		2697	...
		2698
		2699	=head2 EXAMPLES
		2700
		2701	For a real-world example of a program the includes libev
		2702	verbatim, you can have a look at the EV perl module
		2703	(L<http://software.schmorp.de/pkg/EV.html>). It has the libev files in
		2704	the F<libev/> subdirectory and includes them in the F<EV/EVAPI.h> (public
		2705	interface) and F<EV.xs> (implementation) files. Only the F<EV.xs> file
		2706	will be compiled. It is pretty complex because it provides its own header
		2707	file.
		2708
		2709	The usage in rxvt-unicode is simpler. It has a F<ev_cpp.h> header file
		2710	that everybody includes and which overrides some configure choices:
		2711
		2712	#define EV_MINIMAL 1
		2713	#define EV_USE_POLL 0
		2714	#define EV_MULTIPLICITY 0
		2715	#define EV_PERIODIC_ENABLE 0
		2716	#define EV_STAT_ENABLE 0
		2717	#define EV_FORK_ENABLE 0
		2718	#define EV_CONFIG_H <config.h>
		2719	#define EV_MINPRI 0
		2720	#define EV_MAXPRI 0
		2721
		2722	#include "ev++.h"
		2723
		2724	And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:
		2725
		2726	#include "ev_cpp.h"
		2727	#include "ev.c"
		2728
		2729
		2730	=head1 COMPLEXITIES
		2731
		2732	In this section the complexities of (many of) the algorithms used inside
		2733	libev will be explained. For complexity discussions about backends see the
		2734	documentation for C<ev_default_init>.
		2735
		2736	All of the following are about amortised time: If an array needs to be
		2737	extended, libev needs to realloc and move the whole array, but this
		2738	happens asymptotically never with higher number of elements, so O(1) might
		2739	mean it might do a lengthy realloc operation in rare cases, but on average
		2740	it is much faster and asymptotically approaches constant time.
		2741
		2742	=over 4
		2743
		2744	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)
		2745
		2746	This means that, when you have a watcher that triggers in one hour and
		2747	there are 100 watchers that would trigger before that then inserting will
		2748	have to skip roughly seven (C<ld 100>) of these watchers.
		2749
		2750	=item Changing timer/periodic watchers (by autorepeat or calling again): O(log skipped_other_timers)
		2751
		2752	That means that changing a timer costs less than removing/adding them
		2753	as only the relative motion in the event queue has to be paid for.
		2754
		2755	=item Starting io/check/prepare/idle/signal/child watchers: O(1)
		2756
		2757	These just add the watcher into an array or at the head of a list.
		2758
		2759	=item Stopping check/prepare/idle watchers: O(1)
		2760
		2761	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))
		2762
		2763	These watchers are stored in lists then need to be walked to find the
		2764	correct watcher to remove. The lists are usually short (you don't usually
		2765	have many watchers waiting for the same fd or signal).
		2766
		2767	=item Finding the next timer in each loop iteration: O(1)
		2768
		2769	By virtue of using a binary heap, the next timer is always found at the
		2770	beginning of the storage array.
		2771
		2772	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)
		2773
		2774	A change means an I/O watcher gets started or stopped, which requires
		2775	libev to recalculate its status (and possibly tell the kernel, depending
		2776	on backend and wether C<ev_io_set> was used).
		2777
		2778	=item Activating one watcher (putting it into the pending state): O(1)
		2779
		2780	=item Priority handling: O(number_of_priorities)
		2781
		2782	Priorities are implemented by allocating some space for each
		2783	priority. When doing priority-based operations, libev usually has to
		2784	linearly search all the priorities, but starting/stopping and activating
		2785	watchers becomes O(1) w.r.t. prioritiy handling.
		2786
		2787	=back
		2788
		2789
		2790	=head1 Win32 platform limitations and workarounds
		2791
		2792	Win32 doesn't support any of the standards (e.g. POSIX) that libev
		2793	requires, and its I/O model is fundamentally incompatible with the POSIX
		2794	model. Libev still offers limited functionality on this platform in
		2795	the form of the C<EVBACKEND_SELECT> backend, and only supports socket
		2796	descriptors. This only applies when using Win32 natively, not when using
		2797	e.g. cygwin.
		2798
		2799	There is no supported compilation method available on windows except
		2800	embedding it into other applications.
		2801
		2802	Due to the many, low, and arbitrary limits on the win32 platform and the
		2803	abysmal performance of winsockets, using a large number of sockets is not
		2804	recommended (and not reasonable). If your program needs to use more than
		2805	a hundred or so sockets, then likely it needs to use a totally different
		2806	implementation for windows, as libev offers the POSIX model, which cannot
		2807	be implemented efficiently on windows (microsoft monopoly games).
		2808
		2809	=over 4
		2810
		2811	=item The winsocket select function
		2812
		2813	The winsocket C<select> function doesn't follow POSIX in that it requires
		2814	socket I<handles> and not socket I<file descriptors>. This makes select
		2815	very inefficient, and also requires a mapping from file descriptors
		2816	to socket handles. See the discussion of the C<EV_SELECT_USE_FD_SET>,
		2817	C<EV_SELECT_IS_WINSOCKET> and C<EV_FD_TO_WIN32_HANDLE> preprocessor
		2818	symbols for more info.
		2819
		2820	The configuration for a "naked" win32 using the microsoft runtime
		2821	libraries and raw winsocket select is:
		2822
		2823	#define EV_USE_SELECT 1
		2824	#define EV_SELECT_IS_WINSOCKET 1 /* forces EV_SELECT_USE_FD_SET, too */
		2825
		2826	Note that winsockets handling of fd sets is O(n), so you can easily get a
		2827	complexity in the O(n²) range when using win32.
		2828
		2829	=item Limited number of file descriptors
		2830
		2831	Windows has numerous arbitrary (and low) limits on things. Early versions
		2832	of winsocket's select only supported waiting for a max. of C<64> handles
		2833	(probably owning to the fact that all windows kernels can only wait for
		2834	C<64> things at the same time internally; microsoft recommends spawning a
		2835	chain of threads and wait for 63 handles and the previous thread in each).
		2836
		2837	Newer versions support more handles, but you need to define C<FD_SETSIZE>
		2838	to some high number (e.g. C<2048>) before compiling the winsocket select
		2839	call (which might be in libev or elsewhere, for example, perl does its own
		2840	select emulation on windows).
		2841
		2842	Another limit is the number of file descriptors in the microsoft runtime
		2843	libraries, which by default is C<64> (there must be a hidden I<64> fetish
		2844	or something like this inside microsoft). You can increase this by calling
		2845	C<_setmaxstdio>, which can increase this limit to C<2048> (another
		2846	arbitrary limit), but is broken in many versions of the microsoft runtime
		2847	libraries.
		2848
		2849	This might get you to about C<512> or C<2048> sockets (depending on
		2850	windows version and/or the phase of the moon). To get more, you need to
		2851	wrap all I/O functions and provide your own fd management, but the cost of
		2852	calling select (O(n²)) will likely make this unworkable.
		2853
		2854	=back
		2855
		2856
742	=head1 AUTHOR	2857	=head1 AUTHOR
743		2858
744	Marc Lehmann <libev@schmorp.de>.	2859	Marc Lehmann <libev@schmorp.de>.
745		2860

Diff Legend

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