ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/libev/ev.pod

Comparing libev/ev.pod (file contents):
Revision 1.81 by root, Wed Dec 12 04:53:58 2007 UTC vs.
Revision 1.266 by root, Wed Aug 26 17:15:05 2009 UTC

…		…
2		2
3	libev - a high performance full-featured event loop written in C	3	libev - a high performance full-featured event loop written in C
4		4
5	=head1 SYNOPSIS	5	=head1 SYNOPSIS
6		6
7	#include <ev.h>	7	#include <ev.h>
8		8
9	=head1 EXAMPLE PROGRAM	9	=head2 EXAMPLE PROGRAM
10		10
		11	// a single header file is required
11	#include <ev.h>	12	#include <ev.h>
12		13
		14	#include <stdio.h> // for puts
		15
		16	// every watcher type has its own typedef'd struct
		17	// with the name ev_TYPE
13	ev_io stdin_watcher;	18	ev_io stdin_watcher;
14	ev_timer timeout_watcher;	19	ev_timer timeout_watcher;
15		20
16	/* called when data readable on stdin */	21	// all watcher callbacks have a similar signature
		22	// this callback is called when data is readable on stdin
17	static void	23	static void
18	stdin_cb (EV_P_ struct ev_io *w, int revents)	24	stdin_cb (EV_P_ ev_io *w, int revents)
19	{	25	{
20	/* puts ("stdin ready"); */	26	puts ("stdin ready");
21	ev_io_stop (EV_A_ w); /* just a syntax example */	27	// for one-shot events, one must manually stop the watcher
22	ev_unloop (EV_A_ EVUNLOOP_ALL); /* leave all loop calls */	28	// with its corresponding stop function.
		29	ev_io_stop (EV_A_ w);
		30
		31	// this causes all nested ev_loop's to stop iterating
		32	ev_unloop (EV_A_ EVUNLOOP_ALL);
23	}	33	}
24		34
		35	// another callback, this time for a time-out
25	static void	36	static void
26	timeout_cb (EV_P_ struct ev_timer *w, int revents)	37	timeout_cb (EV_P_ ev_timer *w, int revents)
27	{	38	{
28	/* puts ("timeout"); */	39	puts ("timeout");
29	ev_unloop (EV_A_ EVUNLOOP_ONE); /* leave one loop call */	40	// this causes the innermost ev_loop to stop iterating
		41	ev_unloop (EV_A_ EVUNLOOP_ONE);
30	}	42	}
31		43
32	int	44	int
33	main (void)	45	main (void)
34	{	46	{
		47	// use the default event loop unless you have special needs
35	struct ev_loop *loop = ev_default_loop (0);	48	struct ev_loop *loop = ev_default_loop (0);
36		49
37	/* initialise an io watcher, then start it */	50	// initialise an io watcher, then start it
		51	// this one will watch for stdin to become readable
38	ev_io_init (&stdin_watcher, stdin_cb, /*STDIN_FILENO*/ 0, EV_READ);	52	ev_io_init (&stdin_watcher, stdin_cb, /*STDIN_FILENO*/ 0, EV_READ);
39	ev_io_start (loop, &stdin_watcher);	53	ev_io_start (loop, &stdin_watcher);
40		54
		55	// initialise a timer watcher, then start it
41	/* simple non-repeating 5.5 second timeout */	56	// simple non-repeating 5.5 second timeout
42	ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);	57	ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);
43	ev_timer_start (loop, &timeout_watcher);	58	ev_timer_start (loop, &timeout_watcher);
44		59
45	/* loop till timeout or data ready */	60	// now wait for events to arrive
46	ev_loop (loop, 0);	61	ev_loop (loop, 0);
47		62
		63	// unloop was called, so exit
48	return 0;	64	return 0;
49	}	65	}
50		66
51	=head1 DESCRIPTION	67	=head1 ABOUT THIS DOCUMENT
52		68
		69	This document documents the libev software package.
		70
53	The newest version of this document is also available as a html-formatted	71	The newest version of this document is also available as an html-formatted
54	web page you might find easier to navigate when reading it for the first	72	web page you might find easier to navigate when reading it for the first
55	time: L<http://cvs.schmorp.de/libev/ev.html>.	73	time: L<http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod>.
		74
		75	While this document tries to be as complete as possible in documenting
		76	libev, its usage and the rationale behind its design, it is not a tutorial
		77	on event-based programming, nor will it introduce event-based programming
		78	with libev.
		79
		80	Familarity with event based programming techniques in general is assumed
		81	throughout this document.
		82
		83	=head1 ABOUT LIBEV
56		84
57	Libev is an event loop: you register interest in certain events (such as a	85	Libev is an event loop: you register interest in certain events (such as a
58	file descriptor being readable or a timeout occuring), and it will manage	86	file descriptor being readable or a timeout occurring), and it will manage
59	these event sources and provide your program with events.	87	these event sources and provide your program with events.
60		88
61	To do this, it must take more or less complete control over your process	89	To do this, it must take more or less complete control over your process
62	(or thread) by executing the I<event loop> handler, and will then	90	(or thread) by executing the I<event loop> handler, and will then
63	communicate events via a callback mechanism.	91	communicate events via a callback mechanism.
…		…
65	You register interest in certain events by registering so-called I<event	93	You register interest in certain events by registering so-called I<event
66	watchers>, which are relatively small C structures you initialise with the	94	watchers>, which are relatively small C structures you initialise with the
67	details of the event, and then hand it over to libev by I<starting> the	95	details of the event, and then hand it over to libev by I<starting> the
68	watcher.	96	watcher.
69		97
70	=head1 FEATURES	98	=head2 FEATURES
71		99
72	Libev supports C<select>, C<poll>, the Linux-specific C<epoll>, the	100	Libev supports C<select>, C<poll>, the Linux-specific C<epoll>, the
73	BSD-specific C<kqueue> and the Solaris-specific event port mechanisms	101	BSD-specific C<kqueue> and the Solaris-specific event port mechanisms
74	for file descriptor events (C<ev_io>), the Linux C<inotify> interface	102	for file descriptor events (C<ev_io>), the Linux C<inotify> interface
75	(for C<ev_stat>), relative timers (C<ev_timer>), absolute timers	103	(for C<ev_stat>), Linux eventfd/signalfd (for faster and cleaner
76	with customised rescheduling (C<ev_periodic>), synchronous signals	104	inter-thread wakeup (C<ev_async>)/signal handling (C<ev_signal>)) relative
77	(C<ev_signal>), process status change events (C<ev_child>), and event	105	timers (C<ev_timer>), absolute timers with customised rescheduling
78	watchers dealing with the event loop mechanism itself (C<ev_idle>,	106	(C<ev_periodic>), synchronous signals (C<ev_signal>), process status
79	C<ev_embed>, C<ev_prepare> and C<ev_check> watchers) as well as	107	change events (C<ev_child>), and event watchers dealing with the event
80	file watchers (C<ev_stat>) and even limited support for fork events	108	loop mechanism itself (C<ev_idle>, C<ev_embed>, C<ev_prepare> and
81	(C<ev_fork>).	109	C<ev_check> watchers) as well as file watchers (C<ev_stat>) and even
		110	limited support for fork events (C<ev_fork>).
82		111
83	It also is quite fast (see this	112	It also is quite fast (see this
84	L<benchmark|http://libev.schmorp.de/bench.html> comparing it to libevent	113	L<benchmark|http://libev.schmorp.de/bench.html> comparing it to libevent
85	for example).	114	for example).
86		115
87	=head1 CONVENTIONS	116	=head2 CONVENTIONS
88		117
89	Libev is very configurable. In this manual the default configuration will	118	Libev is very configurable. In this manual the default (and most common)
90	be described, which supports multiple event loops. For more info about	119	configuration will be described, which supports multiple event loops. For
91	various configuration options please have a look at B<EMBED> section in	120	more info about various configuration options please have a look at
92	this manual. If libev was configured without support for multiple event	121	B<EMBED> section in this manual. If libev was configured without support
93	loops, then all functions taking an initial argument of name C<loop>	122	for multiple event loops, then all functions taking an initial argument of
94	(which is always of type C<struct ev_loop *>) will not have this argument.	123	name C<loop> (which is always of type C<ev_loop *>) will not have
		124	this argument.
95		125
96	=head1 TIME REPRESENTATION	126	=head2 TIME REPRESENTATION
97		127
98	Libev represents time as a single floating point number, representing the	128	Libev represents time as a single floating point number, representing
99	(fractional) number of seconds since the (POSIX) epoch (somewhere near	129	the (fractional) number of seconds since the (POSIX) epoch (somewhere
100	the beginning of 1970, details are complicated, don't ask). This type is	130	near the beginning of 1970, details are complicated, don't ask). This
101	called C<ev_tstamp>, which is what you should use too. It usually aliases	131	type is called C<ev_tstamp>, which is what you should use too. It usually
102	to the C<double> type in C, and when you need to do any calculations on	132	aliases to the C<double> type in C. When you need to do any calculations
103	it, you should treat it as such.	133	on it, you should treat it as some floating point value. Unlike the name
		134	component C<stamp> might indicate, it is also used for time differences
		135	throughout libev.
		136
		137	=head1 ERROR HANDLING
		138
		139	Libev knows three classes of errors: operating system errors, usage errors
		140	and internal errors (bugs).
		141
		142	When libev catches an operating system error it cannot handle (for example
		143	a system call indicating a condition libev cannot fix), it calls the callback
		144	set via C<ev_set_syserr_cb>, which is supposed to fix the problem or
		145	abort. The default is to print a diagnostic message and to call C<abort
		146	()>.
		147
		148	When libev detects a usage error such as a negative timer interval, then
		149	it will print a diagnostic message and abort (via the C<assert> mechanism,
		150	so C<NDEBUG> will disable this checking): these are programming errors in
		151	the libev caller and need to be fixed there.
		152
		153	Libev also has a few internal error-checking C<assert>ions, and also has
		154	extensive consistency checking code. These do not trigger under normal
		155	circumstances, as they indicate either a bug in libev or worse.
		156
104		157
105	=head1 GLOBAL FUNCTIONS	158	=head1 GLOBAL FUNCTIONS
106		159
107	These functions can be called anytime, even before initialising the	160	These functions can be called anytime, even before initialising the
108	library in any way.	161	library in any way.
…		…
112	=item ev_tstamp ev_time ()	165	=item ev_tstamp ev_time ()
113		166
114	Returns the current time as libev would use it. Please note that the	167	Returns the current time as libev would use it. Please note that the
115	C<ev_now> function is usually faster and also often returns the timestamp	168	C<ev_now> function is usually faster and also often returns the timestamp
116	you actually want to know.	169	you actually want to know.
		170
		171	=item ev_sleep (ev_tstamp interval)
		172
		173	Sleep for the given interval: The current thread will be blocked until
		174	either it is interrupted or the given time interval has passed. Basically
		175	this is a sub-second-resolution C<sleep ()>.
117		176
118	=item int ev_version_major ()	177	=item int ev_version_major ()
119		178
120	=item int ev_version_minor ()	179	=item int ev_version_minor ()
121		180
…		…
134	not a problem.	193	not a problem.
135		194
136	Example: Make sure we haven't accidentally been linked against the wrong	195	Example: Make sure we haven't accidentally been linked against the wrong
137	version.	196	version.
138		197
139	assert (("libev version mismatch",	198	assert (("libev version mismatch",
140	ev_version_major () == EV_VERSION_MAJOR	199	ev_version_major () == EV_VERSION_MAJOR
141	&& ev_version_minor () >= EV_VERSION_MINOR));	200	&& ev_version_minor () >= EV_VERSION_MINOR));
142		201
143	=item unsigned int ev_supported_backends ()	202	=item unsigned int ev_supported_backends ()
144		203
145	Return the set of all backends (i.e. their corresponding C<EV_BACKEND_*>	204	Return the set of all backends (i.e. their corresponding C<EV_BACKEND_*>
146	value) compiled into this binary of libev (independent of their	205	value) compiled into this binary of libev (independent of their
…		…
148	a description of the set values.	207	a description of the set values.
149		208
150	Example: make sure we have the epoll method, because yeah this is cool and	209	Example: make sure we have the epoll method, because yeah this is cool and
151	a must have and can we have a torrent of it please!!!11	210	a must have and can we have a torrent of it please!!!11
152		211
153	assert (("sorry, no epoll, no sex",	212	assert (("sorry, no epoll, no sex",
154	ev_supported_backends () & EVBACKEND_EPOLL));	213	ev_supported_backends () & EVBACKEND_EPOLL));
155		214
156	=item unsigned int ev_recommended_backends ()	215	=item unsigned int ev_recommended_backends ()
157		216
158	Return the set of all backends compiled into this binary of libev and also	217	Return the set of all backends compiled into this binary of libev and also
159	recommended for this platform. This set is often smaller than the one	218	recommended for this platform. This set is often smaller than the one
160	returned by C<ev_supported_backends>, as for example kqueue is broken on	219	returned by C<ev_supported_backends>, as for example kqueue is broken on
161	most BSDs and will not be autodetected unless you explicitly request it	220	most BSDs and will not be auto-detected unless you explicitly request it
162	(assuming you know what you are doing). This is the set of backends that	221	(assuming you know what you are doing). This is the set of backends that
163	libev will probe for if you specify no backends explicitly.	222	libev will probe for if you specify no backends explicitly.
164		223
165	=item unsigned int ev_embeddable_backends ()	224	=item unsigned int ev_embeddable_backends ()
166		225
…		…
170	C<ev_embeddable_backends () & ev_supported_backends ()>, likewise for	229	C<ev_embeddable_backends () & ev_supported_backends ()>, likewise for
171	recommended ones.	230	recommended ones.
172		231
173	See the description of C<ev_embed> watchers for more info.	232	See the description of C<ev_embed> watchers for more info.
174		233
175	=item ev_set_allocator (void *(*cb)(void *ptr, long size))	234	=item ev_set_allocator (void *(*cb)(void *ptr, long size)) [NOT REENTRANT]
176		235
177	Sets the allocation function to use (the prototype is similar - the	236	Sets the allocation function to use (the prototype is similar - the
178	semantics is identical - to the realloc C function). It is used to	237	semantics are identical to the C<realloc> C89/SuS/POSIX function). It is
179	allocate and free memory (no surprises here). If it returns zero when	238	used to allocate and free memory (no surprises here). If it returns zero
180	memory needs to be allocated, the library might abort or take some	239	when memory needs to be allocated (C<size != 0>), the library might abort
181	potentially destructive action. The default is your system realloc	240	or take some potentially destructive action.
182	function.	241
		242	Since some systems (at least OpenBSD and Darwin) fail to implement
		243	correct C<realloc> semantics, libev will use a wrapper around the system
		244	C<realloc> and C<free> functions by default.
183		245
184	You could override this function in high-availability programs to, say,	246	You could override this function in high-availability programs to, say,
185	free some memory if it cannot allocate memory, to use a special allocator,	247	free some memory if it cannot allocate memory, to use a special allocator,
186	or even to sleep a while and retry until some memory is available.	248	or even to sleep a while and retry until some memory is available.
187		249
188	Example: Replace the libev allocator with one that waits a bit and then	250	Example: Replace the libev allocator with one that waits a bit and then
189	retries).	251	retries (example requires a standards-compliant C<realloc>).
190		252
191	static void *	253	static void *
192	persistent_realloc (void *ptr, size_t size)	254	persistent_realloc (void *ptr, size_t size)
193	{	255	{
194	for (;;)	256	for (;;)
…		…
203	}	265	}
204		266
205	...	267	...
206	ev_set_allocator (persistent_realloc);	268	ev_set_allocator (persistent_realloc);
207		269
208	=item ev_set_syserr_cb (void (*cb)(const char *msg));	270	=item ev_set_syserr_cb (void (*cb)(const char *msg)); [NOT REENTRANT]
209		271
210	Set the callback function to call on a retryable syscall error (such	272	Set the callback function to call on a retryable system call error (such
211	as failed select, poll, epoll_wait). The message is a printable string	273	as failed select, poll, epoll_wait). The message is a printable string
212	indicating the system call or subsystem causing the problem. If this	274	indicating the system call or subsystem causing the problem. If this
213	callback is set, then libev will expect it to remedy the sitution, no	275	callback is set, then libev will expect it to remedy the situation, no
214	matter what, when it returns. That is, libev will generally retry the	276	matter what, when it returns. That is, libev will generally retry the
215	requested operation, or, if the condition doesn't go away, do bad stuff	277	requested operation, or, if the condition doesn't go away, do bad stuff
216	(such as abort).	278	(such as abort).
217		279
218	Example: This is basically the same thing that libev does internally, too.	280	Example: This is basically the same thing that libev does internally, too.
…		…
229		291
230	=back	292	=back
231		293
232	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP	294	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP
233		295
234	An event loop is described by a C<struct ev_loop *>. The library knows two	296	An event loop is described by a C<struct ev_loop *> (the C<struct>
235	types of such loops, the I<default> loop, which supports signals and child	297	is I<not> optional in this case, as there is also an C<ev_loop>
236	events, and dynamically created loops which do not.	298	I<function>).
237		299
238	If you use threads, a common model is to run the default event loop	300	The library knows two types of such loops, the I<default> loop, which
239	in your main thread (or in a separate thread) and for each thread you	301	supports signals and child events, and dynamically created loops which do
240	create, you also create another event loop. Libev itself does no locking	302	not.
241	whatsoever, so if you mix calls to the same event loop in different
242	threads, make sure you lock (this is usually a bad idea, though, even if
243	done correctly, because it's hideous and inefficient).
244		303
245	=over 4	304	=over 4
246		305
247	=item struct ev_loop *ev_default_loop (unsigned int flags)	306	=item struct ev_loop *ev_default_loop (unsigned int flags)
248		307
…		…
252	flags. If that is troubling you, check C<ev_backend ()> afterwards).	311	flags. If that is troubling you, check C<ev_backend ()> afterwards).
253		312
254	If you don't know what event loop to use, use the one returned from this	313	If you don't know what event loop to use, use the one returned from this
255	function.	314	function.
256		315
		316	Note that this function is I<not> thread-safe, so if you want to use it
		317	from multiple threads, you have to lock (note also that this is unlikely,
		318	as loops cannot be shared easily between threads anyway).
		319
		320	The default loop is the only loop that can handle C<ev_signal> and
		321	C<ev_child> watchers, and to do this, it always registers a handler
		322	for C<SIGCHLD>. If this is a problem for your application you can either
		323	create a dynamic loop with C<ev_loop_new> that doesn't do that, or you
		324	can simply overwrite the C<SIGCHLD> signal handler I<after> calling
		325	C<ev_default_init>.
		326
257	The flags argument can be used to specify special behaviour or specific	327	The flags argument can be used to specify special behaviour or specific
258	backends to use, and is usually specified as C<0> (or C<EVFLAG_AUTO>).	328	backends to use, and is usually specified as C<0> (or C<EVFLAG_AUTO>).
259		329
260	The following flags are supported:	330	The following flags are supported:
261		331
…		…
266	The default flags value. Use this if you have no clue (it's the right	336	The default flags value. Use this if you have no clue (it's the right
267	thing, believe me).	337	thing, believe me).
268		338
269	=item C<EVFLAG_NOENV>	339	=item C<EVFLAG_NOENV>
270		340
271	If this flag bit is ored into the flag value (or the program runs setuid	341	If this flag bit is or'ed into the flag value (or the program runs setuid
272	or setgid) then libev will I<not> look at the environment variable	342	or setgid) then libev will I<not> look at the environment variable
273	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will	343	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will
274	override the flags completely if it is found in the environment. This is	344	override the flags completely if it is found in the environment. This is
275	useful to try out specific backends to test their performance, or to work	345	useful to try out specific backends to test their performance, or to work
276	around bugs.	346	around bugs.
…		…
282	enabling this flag.	352	enabling this flag.
283		353
284	This works by calling C<getpid ()> on every iteration of the loop,	354	This works by calling C<getpid ()> on every iteration of the loop,
285	and thus this might slow down your event loop if you do a lot of loop	355	and thus this might slow down your event loop if you do a lot of loop
286	iterations and little real work, but is usually not noticeable (on my	356	iterations and little real work, but is usually not noticeable (on my
287	Linux system for example, C<getpid> is actually a simple 5-insn sequence	357	GNU/Linux system for example, C<getpid> is actually a simple 5-insn sequence
288	without a syscall and thus I<very> fast, but my Linux system also has	358	without a system call and thus I<very> fast, but my GNU/Linux system also has
289	C<pthread_atfork> which is even faster).	359	C<pthread_atfork> which is even faster).
290		360
291	The big advantage of this flag is that you can forget about fork (and	361	The big advantage of this flag is that you can forget about fork (and
292	forget about forgetting to tell libev about forking) when you use this	362	forget about forgetting to tell libev about forking) when you use this
293	flag.	363	flag.
294		364
295	This flag setting cannot be overriden or specified in the C<LIBEV_FLAGS>	365	This flag setting cannot be overridden or specified in the C<LIBEV_FLAGS>
296	environment variable.	366	environment variable.
		367
		368	=item C<EVFLAG_NOINOTIFY>
		369
		370	When this flag is specified, then libev will not attempt to use the
		371	I<inotify> API for it's C<ev_stat> watchers. Apart from debugging and
		372	testing, this flag can be useful to conserve inotify file descriptors, as
		373	otherwise each loop using C<ev_stat> watchers consumes one inotify handle.
		374
		375	=item C<EVFLAG_NOSIGNALFD>
		376
		377	When this flag is specified, then libev will not attempt to use the
		378	I<signalfd> API for it's C<ev_signal> (and C<ev_child>) watchers. This is
		379	probably only useful to work around any bugs in libev. Consequently, this
		380	flag might go away once the signalfd functionality is considered stable,
		381	so it's useful mostly in environment variables and not in program code.
297		382
298	=item C<EVBACKEND_SELECT> (value 1, portable select backend)	383	=item C<EVBACKEND_SELECT> (value 1, portable select backend)
299		384
300	This is your standard select(2) backend. Not I<completely> standard, as	385	This is your standard select(2) backend. Not I<completely> standard, as
301	libev tries to roll its own fd_set with no limits on the number of fds,	386	libev tries to roll its own fd_set with no limits on the number of fds,
302	but if that fails, expect a fairly low limit on the number of fds when	387	but if that fails, expect a fairly low limit on the number of fds when
303	using this backend. It doesn't scale too well (O(highest_fd)), but its usually	388	using this backend. It doesn't scale too well (O(highest_fd)), but its
304	the fastest backend for a low number of fds.	389	usually the fastest backend for a low number of (low-numbered :) fds.
		390
		391	To get good performance out of this backend you need a high amount of
		392	parallelism (most of the file descriptors should be busy). If you are
		393	writing a server, you should C<accept ()> in a loop to accept as many
		394	connections as possible during one iteration. You might also want to have
		395	a look at C<ev_set_io_collect_interval ()> to increase the amount of
		396	readiness notifications you get per iteration.
		397
		398	This backend maps C<EV_READ> to the C<readfds> set and C<EV_WRITE> to the
		399	C<writefds> set (and to work around Microsoft Windows bugs, also onto the
		400	C<exceptfds> set on that platform).
305		401
306	=item C<EVBACKEND_POLL> (value 2, poll backend, available everywhere except on windows)	402	=item C<EVBACKEND_POLL> (value 2, poll backend, available everywhere except on windows)
307		403
308	And this is your standard poll(2) backend. It's more complicated than	404	And this is your standard poll(2) backend. It's more complicated
309	select, but handles sparse fds better and has no artificial limit on the	405	than select, but handles sparse fds better and has no artificial
310	number of fds you can use (except it will slow down considerably with a	406	limit on the number of fds you can use (except it will slow down
311	lot of inactive fds). It scales similarly to select, i.e. O(total_fds).	407	considerably with a lot of inactive fds). It scales similarly to select,
		408	i.e. O(total_fds). See the entry for C<EVBACKEND_SELECT>, above, for
		409	performance tips.
		410
		411	This backend maps C<EV_READ> to C<POLLIN | POLLERR | POLLHUP>, and
		412	C<EV_WRITE> to C<POLLOUT | POLLERR | POLLHUP>.
312		413
313	=item C<EVBACKEND_EPOLL> (value 4, Linux)	414	=item C<EVBACKEND_EPOLL> (value 4, Linux)
314		415
315	For few fds, this backend is a bit little slower than poll and select,	416	For few fds, this backend is a bit little slower than poll and select,
316	but it scales phenomenally better. While poll and select usually scale like	417	but it scales phenomenally better. While poll and select usually scale
317	O(total_fds) where n is the total number of fds (or the highest fd), epoll scales	418	like O(total_fds) where n is the total number of fds (or the highest fd),
318	either O(1) or O(active_fds).	419	epoll scales either O(1) or O(active_fds).
319		420
		421	The epoll mechanism deserves honorable mention as the most misdesigned
		422	of the more advanced event mechanisms: mere annoyances include silently
		423	dropping file descriptors, requiring a system call per change per file
		424	descriptor (and unnecessary guessing of parameters), problems with dup and
		425	so on. The biggest issue is fork races, however - if a program forks then
		426	I<both> parent and child process have to recreate the epoll set, which can
		427	take considerable time (one syscall per file descriptor) and is of course
		428	hard to detect.
		429
		430	Epoll is also notoriously buggy - embedding epoll fds I<should> work, but
		431	of course I<doesn't>, and epoll just loves to report events for totally
		432	I<different> file descriptors (even already closed ones, so one cannot
		433	even remove them from the set) than registered in the set (especially
		434	on SMP systems). Libev tries to counter these spurious notifications by
		435	employing an additional generation counter and comparing that against the
		436	events to filter out spurious ones, recreating the set when required.
		437
320	While stopping and starting an I/O watcher in the same iteration will	438	While stopping, setting and starting an I/O watcher in the same iteration
321	result in some caching, there is still a syscall per such incident	439	will result in some caching, there is still a system call per such
322	(because the fd could point to a different file description now), so its	440	incident (because the same I<file descriptor> could point to a different
323	best to avoid that. Also, dup()ed file descriptors might not work very	441	I<file description> now), so its best to avoid that. Also, C<dup ()>'ed
324	well if you register events for both fds.	442	file descriptors might not work very well if you register events for both
		443	file descriptors.
325		444
326	Please note that epoll sometimes generates spurious notifications, so you	445	Best performance from this backend is achieved by not unregistering all
327	need to use non-blocking I/O or other means to avoid blocking when no data	446	watchers for a file descriptor until it has been closed, if possible,
328	(or space) is available.	447	i.e. keep at least one watcher active per fd at all times. Stopping and
		448	starting a watcher (without re-setting it) also usually doesn't cause
		449	extra overhead. A fork can both result in spurious notifications as well
		450	as in libev having to destroy and recreate the epoll object, which can
		451	take considerable time and thus should be avoided.
		452
		453	All this means that, in practice, C<EVBACKEND_SELECT> can be as fast or
		454	faster than epoll for maybe up to a hundred file descriptors, depending on
		455	the usage. So sad.
		456
		457	While nominally embeddable in other event loops, this feature is broken in
		458	all kernel versions tested so far.
		459
		460	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
		461	C<EVBACKEND_POLL>.
329		462
330	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)	463	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)
331		464
332	Kqueue deserves special mention, as at the time of this writing, it	465	Kqueue deserves special mention, as at the time of this writing, it
333	was broken on all BSDs except NetBSD (usually it doesn't work with	466	was broken on all BSDs except NetBSD (usually it doesn't work reliably
334	anything but sockets and pipes, except on Darwin, where of course its	467	with anything but sockets and pipes, except on Darwin, where of course
335	completely useless). For this reason its not being "autodetected"	468	it's completely useless). Unlike epoll, however, whose brokenness
		469	is by design, these kqueue bugs can (and eventually will) be fixed
		470	without API changes to existing programs. For this reason it's not being
336	unless you explicitly specify it explicitly in the flags (i.e. using	471	"auto-detected" unless you explicitly specify it in the flags (i.e. using
337	C<EVBACKEND_KQUEUE>).	472	C<EVBACKEND_KQUEUE>) or libev was compiled on a known-to-be-good (-enough)
		473	system like NetBSD.
		474
		475	You still can embed kqueue into a normal poll or select backend and use it
		476	only for sockets (after having made sure that sockets work with kqueue on
		477	the target platform). See C<ev_embed> watchers for more info.
338		478
339	It scales in the same way as the epoll backend, but the interface to the	479	It scales in the same way as the epoll backend, but the interface to the
340	kernel is more efficient (which says nothing about its actual speed, of	480	kernel is more efficient (which says nothing about its actual speed, of
341	course). While starting and stopping an I/O watcher does not cause an	481	course). While stopping, setting and starting an I/O watcher does never
342	extra syscall as with epoll, it still adds up to four event changes per	482	cause an extra system call as with C<EVBACKEND_EPOLL>, it still adds up to
343	incident, so its best to avoid that.	483	two event changes per incident. Support for C<fork ()> is very bad (but
		484	sane, unlike epoll) and it drops fds silently in similarly hard-to-detect
		485	cases
		486
		487	This backend usually performs well under most conditions.
		488
		489	While nominally embeddable in other event loops, this doesn't work
		490	everywhere, so you might need to test for this. And since it is broken
		491	almost everywhere, you should only use it when you have a lot of sockets
		492	(for which it usually works), by embedding it into another event loop
		493	(e.g. C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> (but C<poll> is of course
		494	also broken on OS X)) and, did I mention it, using it only for sockets.
		495
		496	This backend maps C<EV_READ> into an C<EVFILT_READ> kevent with
		497	C<NOTE_EOF>, and C<EV_WRITE> into an C<EVFILT_WRITE> kevent with
		498	C<NOTE_EOF>.
344		499
345	=item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8)	500	=item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8)
346		501
347	This is not implemented yet (and might never be).	502	This is not implemented yet (and might never be, unless you send me an
		503	implementation). According to reports, C</dev/poll> only supports sockets
		504	and is not embeddable, which would limit the usefulness of this backend
		505	immensely.
348		506
349	=item C<EVBACKEND_PORT> (value 32, Solaris 10)	507	=item C<EVBACKEND_PORT> (value 32, Solaris 10)
350		508
351	This uses the Solaris 10 port mechanism. As with everything on Solaris,	509	This uses the Solaris 10 event port mechanism. As with everything on Solaris,
352	it's really slow, but it still scales very well (O(active_fds)).	510	it's really slow, but it still scales very well (O(active_fds)).
353		511
354	Please note that solaris ports can result in a lot of spurious	512	Please note that Solaris event ports can deliver a lot of spurious
355	notifications, so you need to use non-blocking I/O or other means to avoid	513	notifications, so you need to use non-blocking I/O or other means to avoid
356	blocking when no data (or space) is available.	514	blocking when no data (or space) is available.
		515
		516	While this backend scales well, it requires one system call per active
		517	file descriptor per loop iteration. For small and medium numbers of file
		518	descriptors a "slow" C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> backend
		519	might perform better.
		520
		521	On the positive side, with the exception of the spurious readiness
		522	notifications, this backend actually performed fully to specification
		523	in all tests and is fully embeddable, which is a rare feat among the
		524	OS-specific backends (I vastly prefer correctness over speed hacks).
		525
		526	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
		527	C<EVBACKEND_POLL>.
357		528
358	=item C<EVBACKEND_ALL>	529	=item C<EVBACKEND_ALL>
359		530
360	Try all backends (even potentially broken ones that wouldn't be tried	531	Try all backends (even potentially broken ones that wouldn't be tried
361	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as	532	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as
362	C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>.	533	C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>.
363		534
		535	It is definitely not recommended to use this flag.
		536
364	=back	537	=back
365		538
366	If one or more of these are ored into the flags value, then only these	539	If one or more of the backend flags are or'ed into the flags value,
367	backends will be tried (in the reverse order as given here). If none are	540	then only these backends will be tried (in the reverse order as listed
368	specified, most compiled-in backend will be tried, usually in reverse	541	here). If none are specified, all backends in C<ev_recommended_backends
369	order of their flag values :)	542	()> will be tried.
370		543
371	The most typical usage is like this:	544	Example: This is the most typical usage.
372		545
373	if (!ev_default_loop (0))	546	if (!ev_default_loop (0))
374	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");	547	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
375		548
376	Restrict libev to the select and poll backends, and do not allow	549	Example: Restrict libev to the select and poll backends, and do not allow
377	environment settings to be taken into account:	550	environment settings to be taken into account:
378		551
379	ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);	552	ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);
380		553
381	Use whatever libev has to offer, but make sure that kqueue is used if	554	Example: Use whatever libev has to offer, but make sure that kqueue is
382	available (warning, breaks stuff, best use only with your own private	555	used if available (warning, breaks stuff, best use only with your own
383	event loop and only if you know the OS supports your types of fds):	556	private event loop and only if you know the OS supports your types of
		557	fds):
384		558
385	ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);	559	ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);
386		560
387	=item struct ev_loop *ev_loop_new (unsigned int flags)	561	=item struct ev_loop *ev_loop_new (unsigned int flags)
388		562
389	Similar to C<ev_default_loop>, but always creates a new event loop that is	563	Similar to C<ev_default_loop>, but always creates a new event loop that is
390	always distinct from the default loop. Unlike the default loop, it cannot	564	always distinct from the default loop. Unlike the default loop, it cannot
391	handle signal and child watchers, and attempts to do so will be greeted by	565	handle signal and child watchers, and attempts to do so will be greeted by
392	undefined behaviour (or a failed assertion if assertions are enabled).	566	undefined behaviour (or a failed assertion if assertions are enabled).
393		567
		568	Note that this function I<is> thread-safe, and the recommended way to use
		569	libev with threads is indeed to create one loop per thread, and using the
		570	default loop in the "main" or "initial" thread.
		571
394	Example: Try to create a event loop that uses epoll and nothing else.	572	Example: Try to create a event loop that uses epoll and nothing else.
395		573
396	struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);	574	struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);
397	if (!epoller)	575	if (!epoller)
398	fatal ("no epoll found here, maybe it hides under your chair");	576	fatal ("no epoll found here, maybe it hides under your chair");
399		577
400	=item ev_default_destroy ()	578	=item ev_default_destroy ()
401		579
402	Destroys the default loop again (frees all memory and kernel state	580	Destroys the default loop again (frees all memory and kernel state
403	etc.). None of the active event watchers will be stopped in the normal	581	etc.). None of the active event watchers will be stopped in the normal
404	sense, so e.g. C<ev_is_active> might still return true. It is your	582	sense, so e.g. C<ev_is_active> might still return true. It is your
405	responsibility to either stop all watchers cleanly yoursef I<before>	583	responsibility to either stop all watchers cleanly yourself I<before>
406	calling this function, or cope with the fact afterwards (which is usually	584	calling this function, or cope with the fact afterwards (which is usually
407	the easiest thing, youc na just ignore the watchers and/or C<free ()> them	585	the easiest thing, you can just ignore the watchers and/or C<free ()> them
408	for example).	586	for example).
		587
		588	Note that certain global state, such as signal state (and installed signal
		589	handlers), will not be freed by this function, and related watchers (such
		590	as signal and child watchers) would need to be stopped manually.
		591
		592	In general it is not advisable to call this function except in the
		593	rare occasion where you really need to free e.g. the signal handling
		594	pipe fds. If you need dynamically allocated loops it is better to use
		595	C<ev_loop_new> and C<ev_loop_destroy>).
409		596
410	=item ev_loop_destroy (loop)	597	=item ev_loop_destroy (loop)
411		598
412	Like C<ev_default_destroy>, but destroys an event loop created by an	599	Like C<ev_default_destroy>, but destroys an event loop created by an
413	earlier call to C<ev_loop_new>.	600	earlier call to C<ev_loop_new>.
414		601
415	=item ev_default_fork ()	602	=item ev_default_fork ()
416		603
		604	This function sets a flag that causes subsequent C<ev_loop> iterations
417	This function reinitialises the kernel state for backends that have	605	to reinitialise the kernel state for backends that have one. Despite the
418	one. Despite the name, you can call it anytime, but it makes most sense	606	name, you can call it anytime, but it makes most sense after forking, in
419	after forking, in either the parent or child process (or both, but that	607	the child process (or both child and parent, but that again makes little
420	again makes little sense).	608	sense). You I<must> call it in the child before using any of the libev
		609	functions, and it will only take effect at the next C<ev_loop> iteration.
421		610
422	You I<must> call this function in the child process after forking if and	611	On the other hand, you only need to call this function in the child
423	only if you want to use the event library in both processes. If you just	612	process if and only if you want to use the event library in the child. If
424	fork+exec, you don't have to call it.	613	you just fork+exec, you don't have to call it at all.
425		614
426	The function itself is quite fast and it's usually not a problem to call	615	The function itself is quite fast and it's usually not a problem to call
427	it just in case after a fork. To make this easy, the function will fit in	616	it just in case after a fork. To make this easy, the function will fit in
428	quite nicely into a call to C<pthread_atfork>:	617	quite nicely into a call to C<pthread_atfork>:
429		618
430	pthread_atfork (0, 0, ev_default_fork);	619	pthread_atfork (0, 0, ev_default_fork);
431		620
432	At the moment, C<EVBACKEND_SELECT> and C<EVBACKEND_POLL> are safe to use
433	without calling this function, so if you force one of those backends you
434	do not need to care.
435
436	=item ev_loop_fork (loop)	621	=item ev_loop_fork (loop)
437		622
438	Like C<ev_default_fork>, but acts on an event loop created by	623	Like C<ev_default_fork>, but acts on an event loop created by
439	C<ev_loop_new>. Yes, you have to call this on every allocated event loop	624	C<ev_loop_new>. Yes, you have to call this on every allocated event loop
440	after fork, and how you do this is entirely your own problem.	625	after fork that you want to re-use in the child, and how you do this is
		626	entirely your own problem.
		627
		628	=item int ev_is_default_loop (loop)
		629
		630	Returns true when the given loop is, in fact, the default loop, and false
		631	otherwise.
441		632
442	=item unsigned int ev_loop_count (loop)	633	=item unsigned int ev_loop_count (loop)
443		634
444	Returns the count of loop iterations for the loop, which is identical to	635	Returns the count of loop iterations for the loop, which is identical to
445	the number of times libev did poll for new events. It starts at C<0> and	636	the number of times libev did poll for new events. It starts at C<0> and
446	happily wraps around with enough iterations.	637	happily wraps around with enough iterations.
447		638
448	This value can sometimes be useful as a generation counter of sorts (it	639	This value can sometimes be useful as a generation counter of sorts (it
449	"ticks" the number of loop iterations), as it roughly corresponds with	640	"ticks" the number of loop iterations), as it roughly corresponds with
450	C<ev_prepare> and C<ev_check> calls.	641	C<ev_prepare> and C<ev_check> calls.
		642
		643	=item unsigned int ev_loop_depth (loop)
		644
		645	Returns the number of times C<ev_loop> was entered minus the number of
		646	times C<ev_loop> was exited, in other words, the recursion depth.
		647
		648	Outside C<ev_loop>, this number is zero. In a callback, this number is
		649	C<1>, unless C<ev_loop> was invoked recursively (or from another thread),
		650	in which case it is higher.
		651
		652	Leaving C<ev_loop> abnormally (setjmp/longjmp, cancelling the thread
		653	etc.), doesn't count as exit.
451		654
452	=item unsigned int ev_backend (loop)	655	=item unsigned int ev_backend (loop)
453		656
454	Returns one of the C<EVBACKEND_*> flags indicating the event backend in	657	Returns one of the C<EVBACKEND_*> flags indicating the event backend in
455	use.	658	use.
…		…
458		661
459	Returns the current "event loop time", which is the time the event loop	662	Returns the current "event loop time", which is the time the event loop
460	received events and started processing them. This timestamp does not	663	received events and started processing them. This timestamp does not
461	change as long as callbacks are being processed, and this is also the base	664	change as long as callbacks are being processed, and this is also the base
462	time used for relative timers. You can treat it as the timestamp of the	665	time used for relative timers. You can treat it as the timestamp of the
463	event occuring (or more correctly, libev finding out about it).	666	event occurring (or more correctly, libev finding out about it).
		667
		668	=item ev_now_update (loop)
		669
		670	Establishes the current time by querying the kernel, updating the time
		671	returned by C<ev_now ()> in the progress. This is a costly operation and
		672	is usually done automatically within C<ev_loop ()>.
		673
		674	This function is rarely useful, but when some event callback runs for a
		675	very long time without entering the event loop, updating libev's idea of
		676	the current time is a good idea.
		677
		678	See also L<The special problem of time updates> in the C<ev_timer> section.
		679
		680	=item ev_suspend (loop)
		681
		682	=item ev_resume (loop)
		683
		684	These two functions suspend and resume a loop, for use when the loop is
		685	not used for a while and timeouts should not be processed.
		686
		687	A typical use case would be an interactive program such as a game: When
		688	the user presses C<^Z> to suspend the game and resumes it an hour later it
		689	would be best to handle timeouts as if no time had actually passed while
		690	the program was suspended. This can be achieved by calling C<ev_suspend>
		691	in your C<SIGTSTP> handler, sending yourself a C<SIGSTOP> and calling
		692	C<ev_resume> directly afterwards to resume timer processing.
		693
		694	Effectively, all C<ev_timer> watchers will be delayed by the time spend
		695	between C<ev_suspend> and C<ev_resume>, and all C<ev_periodic> watchers
		696	will be rescheduled (that is, they will lose any events that would have
		697	occured while suspended).
		698
		699	After calling C<ev_suspend> you B<must not> call I<any> function on the
		700	given loop other than C<ev_resume>, and you B<must not> call C<ev_resume>
		701	without a previous call to C<ev_suspend>.
		702
		703	Calling C<ev_suspend>/C<ev_resume> has the side effect of updating the
		704	event loop time (see C<ev_now_update>).
464		705
465	=item ev_loop (loop, int flags)	706	=item ev_loop (loop, int flags)
466		707
467	Finally, this is it, the event handler. This function usually is called	708	Finally, this is it, the event handler. This function usually is called
468	after you initialised all your watchers and you want to start handling	709	after you initialised all your watchers and you want to start handling
…		…
471	If the flags argument is specified as C<0>, it will not return until	712	If the flags argument is specified as C<0>, it will not return until
472	either no event watchers are active anymore or C<ev_unloop> was called.	713	either no event watchers are active anymore or C<ev_unloop> was called.
473		714
474	Please note that an explicit C<ev_unloop> is usually better than	715	Please note that an explicit C<ev_unloop> is usually better than
475	relying on all watchers to be stopped when deciding when a program has	716	relying on all watchers to be stopped when deciding when a program has
476	finished (especially in interactive programs), but having a program that	717	finished (especially in interactive programs), but having a program
477	automatically loops as long as it has to and no longer by virtue of	718	that automatically loops as long as it has to and no longer by virtue
478	relying on its watchers stopping correctly is a thing of beauty.	719	of relying on its watchers stopping correctly, that is truly a thing of
		720	beauty.
479		721
480	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle	722	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle
481	those events and any outstanding ones, but will not block your process in	723	those events and any already outstanding ones, but will not block your
482	case there are no events and will return after one iteration of the loop.	724	process in case there are no events and will return after one iteration of
		725	the loop.
483		726
484	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if	727	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if
485	neccessary) and will handle those and any outstanding ones. It will block	728	necessary) and will handle those and any already outstanding ones. It
486	your process until at least one new event arrives, and will return after	729	will block your process until at least one new event arrives (which could
487	one iteration of the loop. This is useful if you are waiting for some	730	be an event internal to libev itself, so there is no guarantee that a
488	external event in conjunction with something not expressible using other	731	user-registered callback will be called), and will return after one
		732	iteration of the loop.
		733
		734	This is useful if you are waiting for some external event in conjunction
		735	with something not expressible using other libev watchers (i.e. "roll your
489	libev watchers. However, a pair of C<ev_prepare>/C<ev_check> watchers is	736	own C<ev_loop>"). However, a pair of C<ev_prepare>/C<ev_check> watchers is
490	usually a better approach for this kind of thing.	737	usually a better approach for this kind of thing.
491		738
492	Here are the gory details of what C<ev_loop> does:	739	Here are the gory details of what C<ev_loop> does:
493		740
494	- Before the first iteration, call any pending watchers.	741	- Before the first iteration, call any pending watchers.
495	* If there are no active watchers (reference count is zero), return.	742	* If EVFLAG_FORKCHECK was used, check for a fork.
496	- Queue all prepare watchers and then call all outstanding watchers.	743	- If a fork was detected (by any means), queue and call all fork watchers.
		744	- Queue and call all prepare watchers.
497	- If we have been forked, recreate the kernel state.	745	- If we have been forked, detach and recreate the kernel state
		746	as to not disturb the other process.
498	- Update the kernel state with all outstanding changes.	747	- Update the kernel state with all outstanding changes.
499	- Update the "event loop time".	748	- Update the "event loop time" (ev_now ()).
500	- Calculate for how long to block.	749	- Calculate for how long to sleep or block, if at all
		750	(active idle watchers, EVLOOP_NONBLOCK or not having
		751	any active watchers at all will result in not sleeping).
		752	- Sleep if the I/O and timer collect interval say so.
501	- Block the process, waiting for any events.	753	- Block the process, waiting for any events.
502	- Queue all outstanding I/O (fd) events.	754	- Queue all outstanding I/O (fd) events.
503	- Update the "event loop time" and do time jump handling.	755	- Update the "event loop time" (ev_now ()), and do time jump adjustments.
504	- Queue all outstanding timers.	756	- Queue all expired timers.
505	- Queue all outstanding periodics.	757	- Queue all expired periodics.
506	- If no events are pending now, queue all idle watchers.	758	- Unless any events are pending now, queue all idle watchers.
507	- Queue all check watchers.	759	- Queue all check watchers.
508	- Call all queued watchers in reverse order (i.e. check watchers first).	760	- Call all queued watchers in reverse order (i.e. check watchers first).
509	Signals and child watchers are implemented as I/O watchers, and will	761	Signals and child watchers are implemented as I/O watchers, and will
510	be handled here by queueing them when their watcher gets executed.	762	be handled here by queueing them when their watcher gets executed.
511	- If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK	763	- If ev_unloop has been called, or EVLOOP_ONESHOT or EVLOOP_NONBLOCK
512	were used, return, otherwise continue with step *.	764	were used, or there are no active watchers, return, otherwise
		765	continue with step *.
513		766
514	Example: Queue some jobs and then loop until no events are outsanding	767	Example: Queue some jobs and then loop until no events are outstanding
515	anymore.	768	anymore.
516		769
517	... queue jobs here, make sure they register event watchers as long	770	... queue jobs here, make sure they register event watchers as long
518	... as they still have work to do (even an idle watcher will do..)	771	... as they still have work to do (even an idle watcher will do..)
519	ev_loop (my_loop, 0);	772	ev_loop (my_loop, 0);
520	... jobs done. yeah!	773	... jobs done or somebody called unloop. yeah!
521		774
522	=item ev_unloop (loop, how)	775	=item ev_unloop (loop, how)
523		776
524	Can be used to make a call to C<ev_loop> return early (but only after it	777	Can be used to make a call to C<ev_loop> return early (but only after it
525	has processed all outstanding events). The C<how> argument must be either	778	has processed all outstanding events). The C<how> argument must be either
526	C<EVUNLOOP_ONE>, which will make the innermost C<ev_loop> call return, or	779	C<EVUNLOOP_ONE>, which will make the innermost C<ev_loop> call return, or
527	C<EVUNLOOP_ALL>, which will make all nested C<ev_loop> calls return.	780	C<EVUNLOOP_ALL>, which will make all nested C<ev_loop> calls return.
528		781
		782	This "unloop state" will be cleared when entering C<ev_loop> again.
		783
		784	It is safe to call C<ev_unloop> from otuside any C<ev_loop> calls.
		785
529	=item ev_ref (loop)	786	=item ev_ref (loop)
530		787
531	=item ev_unref (loop)	788	=item ev_unref (loop)
532		789
533	Ref/unref can be used to add or remove a reference count on the event	790	Ref/unref can be used to add or remove a reference count on the event
534	loop: Every watcher keeps one reference, and as long as the reference	791	loop: Every watcher keeps one reference, and as long as the reference
535	count is nonzero, C<ev_loop> will not return on its own. If you have	792	count is nonzero, C<ev_loop> will not return on its own.
		793
536	a watcher you never unregister that should not keep C<ev_loop> from	794	If you have a watcher you never unregister that should not keep C<ev_loop>
537	returning, ev_unref() after starting, and ev_ref() before stopping it. For	795	from returning, call ev_unref() after starting, and ev_ref() before
		796	stopping it.
		797
538	example, libev itself uses this for its internal signal pipe: It is not	798	As an example, libev itself uses this for its internal signal pipe: It
539	visible to the libev user and should not keep C<ev_loop> from exiting if	799	is not visible to the libev user and should not keep C<ev_loop> from
540	no event watchers registered by it are active. It is also an excellent	800	exiting if no event watchers registered by it are active. It is also an
541	way to do this for generic recurring timers or from within third-party	801	excellent way to do this for generic recurring timers or from within
542	libraries. Just remember to I<unref after start> and I<ref before stop>.	802	third-party libraries. Just remember to I<unref after start> and I<ref
		803	before stop> (but only if the watcher wasn't active before, or was active
		804	before, respectively. Note also that libev might stop watchers itself
		805	(e.g. non-repeating timers) in which case you have to C<ev_ref>
		806	in the callback).
543		807
544	Example: Create a signal watcher, but keep it from keeping C<ev_loop>	808	Example: Create a signal watcher, but keep it from keeping C<ev_loop>
545	running when nothing else is active.	809	running when nothing else is active.
546		810
547	struct ev_signal exitsig;	811	ev_signal exitsig;
548	ev_signal_init (&exitsig, sig_cb, SIGINT);	812	ev_signal_init (&exitsig, sig_cb, SIGINT);
549	ev_signal_start (loop, &exitsig);	813	ev_signal_start (loop, &exitsig);
550	evf_unref (loop);	814	evf_unref (loop);
551		815
552	Example: For some weird reason, unregister the above signal handler again.	816	Example: For some weird reason, unregister the above signal handler again.
553		817
554	ev_ref (loop);	818	ev_ref (loop);
555	ev_signal_stop (loop, &exitsig);	819	ev_signal_stop (loop, &exitsig);
		820
		821	=item ev_set_io_collect_interval (loop, ev_tstamp interval)
		822
		823	=item ev_set_timeout_collect_interval (loop, ev_tstamp interval)
		824
		825	These advanced functions influence the time that libev will spend waiting
		826	for events. Both time intervals are by default C<0>, meaning that libev
		827	will try to invoke timer/periodic callbacks and I/O callbacks with minimum
		828	latency.
		829
		830	Setting these to a higher value (the C<interval> I<must> be >= C<0>)
		831	allows libev to delay invocation of I/O and timer/periodic callbacks
		832	to increase efficiency of loop iterations (or to increase power-saving
		833	opportunities).
		834
		835	The idea is that sometimes your program runs just fast enough to handle
		836	one (or very few) event(s) per loop iteration. While this makes the
		837	program responsive, it also wastes a lot of CPU time to poll for new
		838	events, especially with backends like C<select ()> which have a high
		839	overhead for the actual polling but can deliver many events at once.
		840
		841	By setting a higher I<io collect interval> you allow libev to spend more
		842	time collecting I/O events, so you can handle more events per iteration,
		843	at the cost of increasing latency. Timeouts (both C<ev_periodic> and
		844	C<ev_timer>) will be not affected. Setting this to a non-null value will
		845	introduce an additional C<ev_sleep ()> call into most loop iterations. The
		846	sleep time ensures that libev will not poll for I/O events more often then
		847	once per this interval, on average.
		848
		849	Likewise, by setting a higher I<timeout collect interval> you allow libev
		850	to spend more time collecting timeouts, at the expense of increased
		851	latency/jitter/inexactness (the watcher callback will be called
		852	later). C<ev_io> watchers will not be affected. Setting this to a non-null
		853	value will not introduce any overhead in libev.
		854
		855	Many (busy) programs can usually benefit by setting the I/O collect
		856	interval to a value near C<0.1> or so, which is often enough for
		857	interactive servers (of course not for games), likewise for timeouts. It
		858	usually doesn't make much sense to set it to a lower value than C<0.01>,
		859	as this approaches the timing granularity of most systems. Note that if
		860	you do transactions with the outside world and you can't increase the
		861	parallelity, then this setting will limit your transaction rate (if you
		862	need to poll once per transaction and the I/O collect interval is 0.01,
		863	then you can't do more than 100 transations per second).
		864
		865	Setting the I<timeout collect interval> can improve the opportunity for
		866	saving power, as the program will "bundle" timer callback invocations that
		867	are "near" in time together, by delaying some, thus reducing the number of
		868	times the process sleeps and wakes up again. Another useful technique to
		869	reduce iterations/wake-ups is to use C<ev_periodic> watchers and make sure
		870	they fire on, say, one-second boundaries only.
		871
		872	Example: we only need 0.1s timeout granularity, and we wish not to poll
		873	more often than 100 times per second:
		874
		875	ev_set_timeout_collect_interval (EV_DEFAULT_UC_ 0.1);
		876	ev_set_io_collect_interval (EV_DEFAULT_UC_ 0.01);
		877
		878	=item ev_invoke_pending (loop)
		879
		880	This call will simply invoke all pending watchers while resetting their
		881	pending state. Normally, C<ev_loop> does this automatically when required,
		882	but when overriding the invoke callback this call comes handy.
		883
		884	=item int ev_pending_count (loop)
		885
		886	Returns the number of pending watchers - zero indicates that no watchers
		887	are pending.
		888
		889	=item ev_set_invoke_pending_cb (loop, void (*invoke_pending_cb)(EV_P))
		890
		891	This overrides the invoke pending functionality of the loop: Instead of
		892	invoking all pending watchers when there are any, C<ev_loop> will call
		893	this callback instead. This is useful, for example, when you want to
		894	invoke the actual watchers inside another context (another thread etc.).
		895
		896	If you want to reset the callback, use C<ev_invoke_pending> as new
		897	callback.
		898
		899	=item ev_set_loop_release_cb (loop, void (*release)(EV_P), void (*acquire)(EV_P))
		900
		901	Sometimes you want to share the same loop between multiple threads. This
		902	can be done relatively simply by putting mutex_lock/unlock calls around
		903	each call to a libev function.
		904
		905	However, C<ev_loop> can run an indefinite time, so it is not feasible to
		906	wait for it to return. One way around this is to wake up the loop via
		907	C<ev_unloop> and C<av_async_send>, another way is to set these I<release>
		908	and I<acquire> callbacks on the loop.
		909
		910	When set, then C<release> will be called just before the thread is
		911	suspended waiting for new events, and C<acquire> is called just
		912	afterwards.
		913
		914	Ideally, C<release> will just call your mutex_unlock function, and
		915	C<acquire> will just call the mutex_lock function again.
		916
		917	While event loop modifications are allowed between invocations of
		918	C<release> and C<acquire> (that's their only purpose after all), no
		919	modifications done will affect the event loop, i.e. adding watchers will
		920	have no effect on the set of file descriptors being watched, or the time
		921	waited. USe an C<ev_async> watcher to wake up C<ev_loop> when you want it
		922	to take note of any changes you made.
		923
		924	In theory, threads executing C<ev_loop> will be async-cancel safe between
		925	invocations of C<release> and C<acquire>.
		926
		927	See also the locking example in the C<THREADS> section later in this
		928	document.
		929
		930	=item ev_set_userdata (loop, void *data)
		931
		932	=item ev_userdata (loop)
		933
		934	Set and retrieve a single C<void *> associated with a loop. When
		935	C<ev_set_userdata> has never been called, then C<ev_userdata> returns
		936	C<0.>
		937
		938	These two functions can be used to associate arbitrary data with a loop,
		939	and are intended solely for the C<invoke_pending_cb>, C<release> and
		940	C<acquire> callbacks described above, but of course can be (ab-)used for
		941	any other purpose as well.
		942
		943	=item ev_loop_verify (loop)
		944
		945	This function only does something when C<EV_VERIFY> support has been
		946	compiled in, which is the default for non-minimal builds. It tries to go
		947	through all internal structures and checks them for validity. If anything
		948	is found to be inconsistent, it will print an error message to standard
		949	error and call C<abort ()>.
		950
		951	This can be used to catch bugs inside libev itself: under normal
		952	circumstances, this function will never abort as of course libev keeps its
		953	data structures consistent.
556		954
557	=back	955	=back
558		956
559		957
560	=head1 ANATOMY OF A WATCHER	958	=head1 ANATOMY OF A WATCHER
		959
		960	In the following description, uppercase C<TYPE> in names stands for the
		961	watcher type, e.g. C<ev_TYPE_start> can mean C<ev_timer_start> for timer
		962	watchers and C<ev_io_start> for I/O watchers.
561		963
562	A watcher is a structure that you create and register to record your	964	A watcher is a structure that you create and register to record your
563	interest in some event. For instance, if you want to wait for STDIN to	965	interest in some event. For instance, if you want to wait for STDIN to
564	become readable, you would create an C<ev_io> watcher for that:	966	become readable, you would create an C<ev_io> watcher for that:
565		967
566	static void my_cb (struct ev_loop *loop, struct ev_io *w, int revents)	968	static void my_cb (struct ev_loop *loop, ev_io *w, int revents)
567	{	969	{
568	ev_io_stop (w);	970	ev_io_stop (w);
569	ev_unloop (loop, EVUNLOOP_ALL);	971	ev_unloop (loop, EVUNLOOP_ALL);
570	}	972	}
571		973
572	struct ev_loop *loop = ev_default_loop (0);	974	struct ev_loop *loop = ev_default_loop (0);
		975
573	struct ev_io stdin_watcher;	976	ev_io stdin_watcher;
		977
574	ev_init (&stdin_watcher, my_cb);	978	ev_init (&stdin_watcher, my_cb);
575	ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);	979	ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);
576	ev_io_start (loop, &stdin_watcher);	980	ev_io_start (loop, &stdin_watcher);
		981
577	ev_loop (loop, 0);	982	ev_loop (loop, 0);
578		983
579	As you can see, you are responsible for allocating the memory for your	984	As you can see, you are responsible for allocating the memory for your
580	watcher structures (and it is usually a bad idea to do this on the stack,	985	watcher structures (and it is I<usually> a bad idea to do this on the
581	although this can sometimes be quite valid).	986	stack).
		987
		988	Each watcher has an associated watcher structure (called C<struct ev_TYPE>
		989	or simply C<ev_TYPE>, as typedefs are provided for all watcher structs).
582		990
583	Each watcher structure must be initialised by a call to C<ev_init	991	Each watcher structure must be initialised by a call to C<ev_init
584	(watcher *, callback)>, which expects a callback to be provided. This	992	(watcher *, callback)>, which expects a callback to be provided. This
585	callback gets invoked each time the event occurs (or, in the case of io	993	callback gets invoked each time the event occurs (or, in the case of I/O
586	watchers, each time the event loop detects that the file descriptor given	994	watchers, each time the event loop detects that the file descriptor given
587	is readable and/or writable).	995	is readable and/or writable).
588		996
589	Each watcher type has its own C<< ev_<type>_set (watcher *, ...) >> macro	997	Each watcher type further has its own C<< ev_TYPE_set (watcher *, ...) >>
590	with arguments specific to this watcher type. There is also a macro	998	macro to configure it, with arguments specific to the watcher type. There
591	to combine initialisation and setting in one call: C<< ev_<type>_init	999	is also a macro to combine initialisation and setting in one call: C<<
592	(watcher *, callback, ...) >>.	1000	ev_TYPE_init (watcher *, callback, ...) >>.
593		1001
594	To make the watcher actually watch out for events, you have to start it	1002	To make the watcher actually watch out for events, you have to start it
595	with a watcher-specific start function (C<< ev_<type>_start (loop, watcher	1003	with a watcher-specific start function (C<< ev_TYPE_start (loop, watcher
596	*) >>), and you can stop watching for events at any time by calling the	1004	*) >>), and you can stop watching for events at any time by calling the
597	corresponding stop function (C<< ev_<type>_stop (loop, watcher *) >>.	1005	corresponding stop function (C<< ev_TYPE_stop (loop, watcher *) >>.
598		1006
599	As long as your watcher is active (has been started but not stopped) you	1007	As long as your watcher is active (has been started but not stopped) you
600	must not touch the values stored in it. Most specifically you must never	1008	must not touch the values stored in it. Most specifically you must never
601	reinitialise it or call its C<set> macro.	1009	reinitialise it or call its C<ev_TYPE_set> macro.
602		1010
603	Each and every callback receives the event loop pointer as first, the	1011	Each and every callback receives the event loop pointer as first, the
604	registered watcher structure as second, and a bitset of received events as	1012	registered watcher structure as second, and a bitset of received events as
605	third argument.	1013	third argument.
606		1014
…		…
660	=item C<EV_FORK>	1068	=item C<EV_FORK>
661		1069
662	The event loop has been resumed in the child process after fork (see	1070	The event loop has been resumed in the child process after fork (see
663	C<ev_fork>).	1071	C<ev_fork>).
664		1072
		1073	=item C<EV_ASYNC>
		1074
		1075	The given async watcher has been asynchronously notified (see C<ev_async>).
		1076
		1077	=item C<EV_CUSTOM>
		1078
		1079	Not ever sent (or otherwise used) by libev itself, but can be freely used
		1080	by libev users to signal watchers (e.g. via C<ev_feed_event>).
		1081
665	=item C<EV_ERROR>	1082	=item C<EV_ERROR>
666		1083
667	An unspecified error has occured, the watcher has been stopped. This might	1084	An unspecified error has occurred, the watcher has been stopped. This might
668	happen because the watcher could not be properly started because libev	1085	happen because the watcher could not be properly started because libev
669	ran out of memory, a file descriptor was found to be closed or any other	1086	ran out of memory, a file descriptor was found to be closed or any other
		1087	problem. Libev considers these application bugs.
		1088
670	problem. You best act on it by reporting the problem and somehow coping	1089	You best act on it by reporting the problem and somehow coping with the
671	with the watcher being stopped.	1090	watcher being stopped. Note that well-written programs should not receive
		1091	an error ever, so when your watcher receives it, this usually indicates a
		1092	bug in your program.
672		1093
673	Libev will usually signal a few "dummy" events together with an error,	1094	Libev will usually signal a few "dummy" events together with an error, for
674	for example it might indicate that a fd is readable or writable, and if	1095	example it might indicate that a fd is readable or writable, and if your
675	your callbacks is well-written it can just attempt the operation and cope	1096	callbacks is well-written it can just attempt the operation and cope with
676	with the error from read() or write(). This will not work in multithreaded	1097	the error from read() or write(). This will not work in multi-threaded
677	programs, though, so beware.	1098	programs, though, as the fd could already be closed and reused for another
		1099	thing, so beware.
678		1100
679	=back	1101	=back
680		1102
681	=head2 GENERIC WATCHER FUNCTIONS	1103	=head2 GENERIC WATCHER FUNCTIONS
682
683	In the following description, C<TYPE> stands for the watcher type,
684	e.g. C<timer> for C<ev_timer> watchers and C<io> for C<ev_io> watchers.
685		1104
686	=over 4	1105	=over 4
687		1106
688	=item C<ev_init> (ev_TYPE *watcher, callback)	1107	=item C<ev_init> (ev_TYPE *watcher, callback)
689		1108
…		…
695	which rolls both calls into one.	1114	which rolls both calls into one.
696		1115
697	You can reinitialise a watcher at any time as long as it has been stopped	1116	You can reinitialise a watcher at any time as long as it has been stopped
698	(or never started) and there are no pending events outstanding.	1117	(or never started) and there are no pending events outstanding.
699		1118
700	The callback is always of type C<void (*)(ev_loop *loop, ev_TYPE *watcher,	1119	The callback is always of type C<void (*)(struct ev_loop *loop, ev_TYPE *watcher,
701	int revents)>.	1120	int revents)>.
		1121
		1122	Example: Initialise an C<ev_io> watcher in two steps.
		1123
		1124	ev_io w;
		1125	ev_init (&w, my_cb);
		1126	ev_io_set (&w, STDIN_FILENO, EV_READ);
702		1127
703	=item C<ev_TYPE_set> (ev_TYPE *, [args])	1128	=item C<ev_TYPE_set> (ev_TYPE *, [args])
704		1129
705	This macro initialises the type-specific parts of a watcher. You need to	1130	This macro initialises the type-specific parts of a watcher. You need to
706	call C<ev_init> at least once before you call this macro, but you can	1131	call C<ev_init> at least once before you call this macro, but you can
…		…
709	difference to the C<ev_init> macro).	1134	difference to the C<ev_init> macro).
710		1135
711	Although some watcher types do not have type-specific arguments	1136	Although some watcher types do not have type-specific arguments
712	(e.g. C<ev_prepare>) you still need to call its C<set> macro.	1137	(e.g. C<ev_prepare>) you still need to call its C<set> macro.
713		1138
		1139	See C<ev_init>, above, for an example.
		1140
714	=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])	1141	=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])
715		1142
716	This convinience macro rolls both C<ev_init> and C<ev_TYPE_set> macro	1143	This convenience macro rolls both C<ev_init> and C<ev_TYPE_set> macro
717	calls into a single call. This is the most convinient method to initialise	1144	calls into a single call. This is the most convenient method to initialise
718	a watcher. The same limitations apply, of course.	1145	a watcher. The same limitations apply, of course.
		1146
		1147	Example: Initialise and set an C<ev_io> watcher in one step.
		1148
		1149	ev_io_init (&w, my_cb, STDIN_FILENO, EV_READ);
719		1150
720	=item C<ev_TYPE_start> (loop *, ev_TYPE *watcher)	1151	=item C<ev_TYPE_start> (loop *, ev_TYPE *watcher)
721		1152
722	Starts (activates) the given watcher. Only active watchers will receive	1153	Starts (activates) the given watcher. Only active watchers will receive
723	events. If the watcher is already active nothing will happen.	1154	events. If the watcher is already active nothing will happen.
724		1155
		1156	Example: Start the C<ev_io> watcher that is being abused as example in this
		1157	whole section.
		1158
		1159	ev_io_start (EV_DEFAULT_UC, &w);
		1160
725	=item C<ev_TYPE_stop> (loop *, ev_TYPE *watcher)	1161	=item C<ev_TYPE_stop> (loop *, ev_TYPE *watcher)
726		1162
727	Stops the given watcher again (if active) and clears the pending	1163	Stops the given watcher if active, and clears the pending status (whether
		1164	the watcher was active or not).
		1165
728	status. It is possible that stopped watchers are pending (for example,	1166	It is possible that stopped watchers are pending - for example,
729	non-repeating timers are being stopped when they become pending), but	1167	non-repeating timers are being stopped when they become pending - but
730	C<ev_TYPE_stop> ensures that the watcher is neither active nor pending. If	1168	calling C<ev_TYPE_stop> ensures that the watcher is neither active nor
731	you want to free or reuse the memory used by the watcher it is therefore a	1169	pending. If you want to free or reuse the memory used by the watcher it is
732	good idea to always call its C<ev_TYPE_stop> function.	1170	therefore a good idea to always call its C<ev_TYPE_stop> function.
733		1171
734	=item bool ev_is_active (ev_TYPE *watcher)	1172	=item bool ev_is_active (ev_TYPE *watcher)
735		1173
736	Returns a true value iff the watcher is active (i.e. it has been started	1174	Returns a true value iff the watcher is active (i.e. it has been started
737	and not yet been stopped). As long as a watcher is active you must not modify	1175	and not yet been stopped). As long as a watcher is active you must not modify
…		…
763	integer between C<EV_MAXPRI> (default: C<2>) and C<EV_MINPRI>	1201	integer between C<EV_MAXPRI> (default: C<2>) and C<EV_MINPRI>
764	(default: C<-2>). Pending watchers with higher priority will be invoked	1202	(default: C<-2>). Pending watchers with higher priority will be invoked
765	before watchers with lower priority, but priority will not keep watchers	1203	before watchers with lower priority, but priority will not keep watchers
766	from being executed (except for C<ev_idle> watchers).	1204	from being executed (except for C<ev_idle> watchers).
767		1205
768	This means that priorities are I<only> used for ordering callback
769	invocation after new events have been received. This is useful, for
770	example, to reduce latency after idling, or more often, to bind two
771	watchers on the same event and make sure one is called first.
772
773	If you need to suppress invocation when higher priority events are pending	1206	If you need to suppress invocation when higher priority events are pending
774	you need to look at C<ev_idle> watchers, which provide this functionality.	1207	you need to look at C<ev_idle> watchers, which provide this functionality.
775		1208
776	You I<must not> change the priority of a watcher as long as it is active or	1209	You I<must not> change the priority of a watcher as long as it is active or
777	pending.	1210	pending.
778		1211
		1212	Setting a priority outside the range of C<EV_MINPRI> to C<EV_MAXPRI> is
		1213	fine, as long as you do not mind that the priority value you query might
		1214	or might not have been clamped to the valid range.
		1215
779	The default priority used by watchers when no priority has been set is	1216	The default priority used by watchers when no priority has been set is
780	always C<0>, which is supposed to not be too high and not be too low :).	1217	always C<0>, which is supposed to not be too high and not be too low :).
781		1218
782	Setting a priority outside the range of C<EV_MINPRI> to C<EV_MAXPRI> is	1219	See L<WATCHER PRIORITY MODELS>, below, for a more thorough treatment of
783	fine, as long as you do not mind that the priority value you query might	1220	priorities.
784	or might not have been adjusted to be within valid range.
785		1221
786	=item ev_invoke (loop, ev_TYPE *watcher, int revents)	1222	=item ev_invoke (loop, ev_TYPE *watcher, int revents)
787		1223
788	Invoke the C<watcher> with the given C<loop> and C<revents>. Neither	1224	Invoke the C<watcher> with the given C<loop> and C<revents>. Neither
789	C<loop> nor C<revents> need to be valid as long as the watcher callback	1225	C<loop> nor C<revents> need to be valid as long as the watcher callback
790	can deal with that fact.	1226	can deal with that fact, as both are simply passed through to the
		1227	callback.
791		1228
792	=item int ev_clear_pending (loop, ev_TYPE *watcher)	1229	=item int ev_clear_pending (loop, ev_TYPE *watcher)
793		1230
794	If the watcher is pending, this function returns clears its pending status	1231	If the watcher is pending, this function clears its pending status and
795	and returns its C<revents> bitset (as if its callback was invoked). If the	1232	returns its C<revents> bitset (as if its callback was invoked). If the
796	watcher isn't pending it does nothing and returns C<0>.	1233	watcher isn't pending it does nothing and returns C<0>.
797		1234
		1235	Sometimes it can be useful to "poll" a watcher instead of waiting for its
		1236	callback to be invoked, which can be accomplished with this function.
		1237
798	=back	1238	=back
799		1239
800		1240
801	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER	1241	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER
802		1242
803	Each watcher has, by default, a member C<void *data> that you can change	1243	Each watcher has, by default, a member C<void *data> that you can change
804	and read at any time, libev will completely ignore it. This can be used	1244	and read at any time: libev will completely ignore it. This can be used
805	to associate arbitrary data with your watcher. If you need more data and	1245	to associate arbitrary data with your watcher. If you need more data and
806	don't want to allocate memory and store a pointer to it in that data	1246	don't want to allocate memory and store a pointer to it in that data
807	member, you can also "subclass" the watcher type and provide your own	1247	member, you can also "subclass" the watcher type and provide your own
808	data:	1248	data:
809		1249
810	struct my_io	1250	struct my_io
811	{	1251	{
812	struct ev_io io;	1252	ev_io io;
813	int otherfd;	1253	int otherfd;
814	void *somedata;	1254	void *somedata;
815	struct whatever *mostinteresting;	1255	struct whatever *mostinteresting;
816	}	1256	};
		1257
		1258	...
		1259	struct my_io w;
		1260	ev_io_init (&w.io, my_cb, fd, EV_READ);
817		1261
818	And since your callback will be called with a pointer to the watcher, you	1262	And since your callback will be called with a pointer to the watcher, you
819	can cast it back to your own type:	1263	can cast it back to your own type:
820		1264
821	static void my_cb (struct ev_loop *loop, struct ev_io *w_, int revents)	1265	static void my_cb (struct ev_loop *loop, ev_io *w_, int revents)
822	{	1266	{
823	struct my_io *w = (struct my_io *)w_;	1267	struct my_io *w = (struct my_io *)w_;
824	...	1268	...
825	}	1269	}
826		1270
827	More interesting and less C-conformant ways of casting your callback type	1271	More interesting and less C-conformant ways of casting your callback type
828	instead have been omitted.	1272	instead have been omitted.
829		1273
830	Another common scenario is having some data structure with multiple	1274	Another common scenario is to use some data structure with multiple
831	watchers:	1275	embedded watchers:
832		1276
833	struct my_biggy	1277	struct my_biggy
834	{	1278	{
835	int some_data;	1279	int some_data;
836	ev_timer t1;	1280	ev_timer t1;
837	ev_timer t2;	1281	ev_timer t2;
838	}	1282	}
839		1283
840	In this case getting the pointer to C<my_biggy> is a bit more complicated,	1284	In this case getting the pointer to C<my_biggy> is a bit more
841	you need to use C<offsetof>:	1285	complicated: Either you store the address of your C<my_biggy> struct
		1286	in the C<data> member of the watcher (for woozies), or you need to use
		1287	some pointer arithmetic using C<offsetof> inside your watchers (for real
		1288	programmers):
842		1289
843	#include <stddef.h>	1290	#include <stddef.h>
844		1291
845	static void	1292	static void
846	t1_cb (EV_P_ struct ev_timer *w, int revents)	1293	t1_cb (EV_P_ ev_timer *w, int revents)
847	{	1294	{
848	struct my_biggy big = (struct my_biggy *	1295	struct my_biggy big = (struct my_biggy *)
849	(((char *)w) - offsetof (struct my_biggy, t1));	1296	(((char *)w) - offsetof (struct my_biggy, t1));
850	}	1297	}
851		1298
852	static void	1299	static void
853	t2_cb (EV_P_ struct ev_timer *w, int revents)	1300	t2_cb (EV_P_ ev_timer *w, int revents)
854	{	1301	{
855	struct my_biggy big = (struct my_biggy *	1302	struct my_biggy big = (struct my_biggy *)
856	(((char *)w) - offsetof (struct my_biggy, t2));	1303	(((char *)w) - offsetof (struct my_biggy, t2));
857	}	1304	}
		1305
		1306	=head2 WATCHER PRIORITY MODELS
		1307
		1308	Many event loops support I<watcher priorities>, which are usually small
		1309	integers that influence the ordering of event callback invocation
		1310	between watchers in some way, all else being equal.
		1311
		1312	In libev, Watcher priorities can be set using C<ev_set_priority>. See its
		1313	description for the more technical details such as the actual priority
		1314	range.
		1315
		1316	There are two common ways how these these priorities are being interpreted
		1317	by event loops:
		1318
		1319	In the more common lock-out model, higher priorities "lock out" invocation
		1320	of lower priority watchers, which means as long as higher priority
		1321	watchers receive events, lower priority watchers are not being invoked.
		1322
		1323	The less common only-for-ordering model uses priorities solely to order
		1324	callback invocation within a single event loop iteration: Higher priority
		1325	watchers are invoked before lower priority ones, but they all get invoked
		1326	before polling for new events.
		1327
		1328	Libev uses the second (only-for-ordering) model for all its watchers
		1329	except for idle watchers (which use the lock-out model).
		1330
		1331	The rationale behind this is that implementing the lock-out model for
		1332	watchers is not well supported by most kernel interfaces, and most event
		1333	libraries will just poll for the same events again and again as long as
		1334	their callbacks have not been executed, which is very inefficient in the
		1335	common case of one high-priority watcher locking out a mass of lower
		1336	priority ones.
		1337
		1338	Static (ordering) priorities are most useful when you have two or more
		1339	watchers handling the same resource: a typical usage example is having an
		1340	C<ev_io> watcher to receive data, and an associated C<ev_timer> to handle
		1341	timeouts. Under load, data might be received while the program handles
		1342	other jobs, but since timers normally get invoked first, the timeout
		1343	handler will be executed before checking for data. In that case, giving
		1344	the timer a lower priority than the I/O watcher ensures that I/O will be
		1345	handled first even under adverse conditions (which is usually, but not
		1346	always, what you want).
		1347
		1348	Since idle watchers use the "lock-out" model, meaning that idle watchers
		1349	will only be executed when no same or higher priority watchers have
		1350	received events, they can be used to implement the "lock-out" model when
		1351	required.
		1352
		1353	For example, to emulate how many other event libraries handle priorities,
		1354	you can associate an C<ev_idle> watcher to each such watcher, and in
		1355	the normal watcher callback, you just start the idle watcher. The real
		1356	processing is done in the idle watcher callback. This causes libev to
		1357	continously poll and process kernel event data for the watcher, but when
		1358	the lock-out case is known to be rare (which in turn is rare :), this is
		1359	workable.
		1360
		1361	Usually, however, the lock-out model implemented that way will perform
		1362	miserably under the type of load it was designed to handle. In that case,
		1363	it might be preferable to stop the real watcher before starting the
		1364	idle watcher, so the kernel will not have to process the event in case
		1365	the actual processing will be delayed for considerable time.
		1366
		1367	Here is an example of an I/O watcher that should run at a strictly lower
		1368	priority than the default, and which should only process data when no
		1369	other events are pending:
		1370
		1371	ev_idle idle; // actual processing watcher
		1372	ev_io io; // actual event watcher
		1373
		1374	static void
		1375	io_cb (EV_P_ ev_io *w, int revents)
		1376	{
		1377	// stop the I/O watcher, we received the event, but
		1378	// are not yet ready to handle it.
		1379	ev_io_stop (EV_A_ w);
		1380
		1381	// start the idle watcher to ahndle the actual event.
		1382	// it will not be executed as long as other watchers
		1383	// with the default priority are receiving events.
		1384	ev_idle_start (EV_A_ &idle);
		1385	}
		1386
		1387	static void
		1388	idle_cb (EV_P_ ev_idle *w, int revents)
		1389	{
		1390	// actual processing
		1391	read (STDIN_FILENO, ...);
		1392
		1393	// have to start the I/O watcher again, as
		1394	// we have handled the event
		1395	ev_io_start (EV_P_ &io);
		1396	}
		1397
		1398	// initialisation
		1399	ev_idle_init (&idle, idle_cb);
		1400	ev_io_init (&io, io_cb, STDIN_FILENO, EV_READ);
		1401	ev_io_start (EV_DEFAULT_ &io);
		1402
		1403	In the "real" world, it might also be beneficial to start a timer, so that
		1404	low-priority connections can not be locked out forever under load. This
		1405	enables your program to keep a lower latency for important connections
		1406	during short periods of high load, while not completely locking out less
		1407	important ones.
858		1408
859		1409
860	=head1 WATCHER TYPES	1410	=head1 WATCHER TYPES
861		1411
862	This section describes each watcher in detail, but will not repeat	1412	This section describes each watcher in detail, but will not repeat
…		…
886	In general you can register as many read and/or write event watchers per	1436	In general you can register as many read and/or write event watchers per
887	fd as you want (as long as you don't confuse yourself). Setting all file	1437	fd as you want (as long as you don't confuse yourself). Setting all file
888	descriptors to non-blocking mode is also usually a good idea (but not	1438	descriptors to non-blocking mode is also usually a good idea (but not
889	required if you know what you are doing).	1439	required if you know what you are doing).
890		1440
891	You have to be careful with dup'ed file descriptors, though. Some backends	1441	If you cannot use non-blocking mode, then force the use of a
892	(the linux epoll backend is a notable example) cannot handle dup'ed file	1442	known-to-be-good backend (at the time of this writing, this includes only
893	descriptors correctly if you register interest in two or more fds pointing	1443	C<EVBACKEND_SELECT> and C<EVBACKEND_POLL>). The same applies to file
894	to the same underlying file/socket/etc. description (that is, they share	1444	descriptors for which non-blocking operation makes no sense (such as
895	the same underlying "file open").	1445	files) - libev doesn't guarentee any specific behaviour in that case.
896
897	If you must do this, then force the use of a known-to-be-good backend
898	(at the time of this writing, this includes only C<EVBACKEND_SELECT> and
899	C<EVBACKEND_POLL>).
900		1446
901	Another thing you have to watch out for is that it is quite easy to	1447	Another thing you have to watch out for is that it is quite easy to
902	receive "spurious" readyness notifications, that is your callback might	1448	receive "spurious" readiness notifications, that is your callback might
903	be called with C<EV_READ> but a subsequent C<read>(2) will actually block	1449	be called with C<EV_READ> but a subsequent C<read>(2) will actually block
904	because there is no data. Not only are some backends known to create a	1450	because there is no data. Not only are some backends known to create a
905	lot of those (for example solaris ports), it is very easy to get into	1451	lot of those (for example Solaris ports), it is very easy to get into
906	this situation even with a relatively standard program structure. Thus	1452	this situation even with a relatively standard program structure. Thus
907	it is best to always use non-blocking I/O: An extra C<read>(2) returning	1453	it is best to always use non-blocking I/O: An extra C<read>(2) returning
908	C<EAGAIN> is far preferable to a program hanging until some data arrives.	1454	C<EAGAIN> is far preferable to a program hanging until some data arrives.
909		1455
910	If you cannot run the fd in non-blocking mode (for example you should not	1456	If you cannot run the fd in non-blocking mode (for example you should
911	play around with an Xlib connection), then you have to seperately re-test	1457	not play around with an Xlib connection), then you have to separately
912	whether a file descriptor is really ready with a known-to-be good interface	1458	re-test whether a file descriptor is really ready with a known-to-be good
913	such as poll (fortunately in our Xlib example, Xlib already does this on	1459	interface such as poll (fortunately in our Xlib example, Xlib already
914	its own, so its quite safe to use).	1460	does this on its own, so its quite safe to use). Some people additionally
		1461	use C<SIGALRM> and an interval timer, just to be sure you won't block
		1462	indefinitely.
		1463
		1464	But really, best use non-blocking mode.
915		1465
916	=head3 The special problem of disappearing file descriptors	1466	=head3 The special problem of disappearing file descriptors
917		1467
918	Some backends (e.g kqueue, epoll) need to be told about closing a file	1468	Some backends (e.g. kqueue, epoll) need to be told about closing a file
919	descriptor (either by calling C<close> explicitly or by any other means,	1469	descriptor (either due to calling C<close> explicitly or any other means,
920	such as C<dup>). The reason is that you register interest in some file	1470	such as C<dup2>). The reason is that you register interest in some file
921	descriptor, but when it goes away, the operating system will silently drop	1471	descriptor, but when it goes away, the operating system will silently drop
922	this interest. If another file descriptor with the same number then is	1472	this interest. If another file descriptor with the same number then is
923	registered with libev, there is no efficient way to see that this is, in	1473	registered with libev, there is no efficient way to see that this is, in
924	fact, a different file descriptor.	1474	fact, a different file descriptor.
925		1475
…		…
932		1482
933	This is how one would do it normally anyway, the important point is that	1483	This is how one would do it normally anyway, the important point is that
934	the libev application should not optimise around libev but should leave	1484	the libev application should not optimise around libev but should leave
935	optimisations to libev.	1485	optimisations to libev.
936		1486
		1487	=head3 The special problem of dup'ed file descriptors
		1488
		1489	Some backends (e.g. epoll), cannot register events for file descriptors,
		1490	but only events for the underlying file descriptions. That means when you
		1491	have C<dup ()>'ed file descriptors or weirder constellations, and register
		1492	events for them, only one file descriptor might actually receive events.
		1493
		1494	There is no workaround possible except not registering events
		1495	for potentially C<dup ()>'ed file descriptors, or to resort to
		1496	C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>.
		1497
		1498	=head3 The special problem of fork
		1499
		1500	Some backends (epoll, kqueue) do not support C<fork ()> at all or exhibit
		1501	useless behaviour. Libev fully supports fork, but needs to be told about
		1502	it in the child.
		1503
		1504	To support fork in your programs, you either have to call
		1505	C<ev_default_fork ()> or C<ev_loop_fork ()> after a fork in the child,
		1506	enable C<EVFLAG_FORKCHECK>, or resort to C<EVBACKEND_SELECT> or
		1507	C<EVBACKEND_POLL>.
		1508
		1509	=head3 The special problem of SIGPIPE
		1510
		1511	While not really specific to libev, it is easy to forget about C<SIGPIPE>:
		1512	when writing to a pipe whose other end has been closed, your program gets
		1513	sent a SIGPIPE, which, by default, aborts your program. For most programs
		1514	this is sensible behaviour, for daemons, this is usually undesirable.
		1515
		1516	So when you encounter spurious, unexplained daemon exits, make sure you
		1517	ignore SIGPIPE (and maybe make sure you log the exit status of your daemon
		1518	somewhere, as that would have given you a big clue).
		1519
		1520
		1521	=head3 Watcher-Specific Functions
937		1522
938	=over 4	1523	=over 4
939		1524
940	=item ev_io_init (ev_io *, callback, int fd, int events)	1525	=item ev_io_init (ev_io *, callback, int fd, int events)
941		1526
942	=item ev_io_set (ev_io *, int fd, int events)	1527	=item ev_io_set (ev_io *, int fd, int events)
943		1528
944	Configures an C<ev_io> watcher. The C<fd> is the file descriptor to	1529	Configures an C<ev_io> watcher. The C<fd> is the file descriptor to
945	rceeive events for and events is either C<EV_READ>, C<EV_WRITE> or	1530	receive events for and C<events> is either C<EV_READ>, C<EV_WRITE> or
946	C<EV_READ | EV_WRITE> to receive the given events.	1531	C<EV_READ | EV_WRITE>, to express the desire to receive the given events.
947		1532
948	=item int fd [read-only]	1533	=item int fd [read-only]
949		1534
950	The file descriptor being watched.	1535	The file descriptor being watched.
951		1536
952	=item int events [read-only]	1537	=item int events [read-only]
953		1538
954	The events being watched.	1539	The events being watched.
955		1540
956	=back	1541	=back
		1542
		1543	=head3 Examples
957		1544
958	Example: Call C<stdin_readable_cb> when STDIN_FILENO has become, well	1545	Example: Call C<stdin_readable_cb> when STDIN_FILENO has become, well
959	readable, but only once. Since it is likely line-buffered, you could	1546	readable, but only once. Since it is likely line-buffered, you could
960	attempt to read a whole line in the callback.	1547	attempt to read a whole line in the callback.
961		1548
962	static void	1549	static void
963	stdin_readable_cb (struct ev_loop *loop, struct ev_io *w, int revents)	1550	stdin_readable_cb (struct ev_loop *loop, ev_io *w, int revents)
964	{	1551	{
965	ev_io_stop (loop, w);	1552	ev_io_stop (loop, w);
966	.. read from stdin here (or from w->fd) and haqndle any I/O errors	1553	.. read from stdin here (or from w->fd) and handle any I/O errors
967	}	1554	}
968		1555
969	...	1556	...
970	struct ev_loop *loop = ev_default_init (0);	1557	struct ev_loop *loop = ev_default_init (0);
971	struct ev_io stdin_readable;	1558	ev_io stdin_readable;
972	ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);	1559	ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
973	ev_io_start (loop, &stdin_readable);	1560	ev_io_start (loop, &stdin_readable);
974	ev_loop (loop, 0);	1561	ev_loop (loop, 0);
975		1562
976		1563
977	=head2 C<ev_timer> - relative and optionally repeating timeouts	1564	=head2 C<ev_timer> - relative and optionally repeating timeouts
978		1565
979	Timer watchers are simple relative timers that generate an event after a	1566	Timer watchers are simple relative timers that generate an event after a
980	given time, and optionally repeating in regular intervals after that.	1567	given time, and optionally repeating in regular intervals after that.
981		1568
982	The timers are based on real time, that is, if you register an event that	1569	The timers are based on real time, that is, if you register an event that
983	times out after an hour and you reset your system clock to last years	1570	times out after an hour and you reset your system clock to January last
984	time, it will still time out after (roughly) and hour. "Roughly" because	1571	year, it will still time out after (roughly) one hour. "Roughly" because
985	detecting time jumps is hard, and some inaccuracies are unavoidable (the	1572	detecting time jumps is hard, and some inaccuracies are unavoidable (the
986	monotonic clock option helps a lot here).	1573	monotonic clock option helps a lot here).
		1574
		1575	The callback is guaranteed to be invoked only I<after> its timeout has
		1576	passed (not I<at>, so on systems with very low-resolution clocks this
		1577	might introduce a small delay). If multiple timers become ready during the
		1578	same loop iteration then the ones with earlier time-out values are invoked
		1579	before ones of the same priority with later time-out values (but this is
		1580	no longer true when a callback calls C<ev_loop> recursively).
		1581
		1582	=head3 Be smart about timeouts
		1583
		1584	Many real-world problems involve some kind of timeout, usually for error
		1585	recovery. A typical example is an HTTP request - if the other side hangs,
		1586	you want to raise some error after a while.
		1587
		1588	What follows are some ways to handle this problem, from obvious and
		1589	inefficient to smart and efficient.
		1590
		1591	In the following, a 60 second activity timeout is assumed - a timeout that
		1592	gets reset to 60 seconds each time there is activity (e.g. each time some
		1593	data or other life sign was received).
		1594
		1595	=over 4
		1596
		1597	=item 1. Use a timer and stop, reinitialise and start it on activity.
		1598
		1599	This is the most obvious, but not the most simple way: In the beginning,
		1600	start the watcher:
		1601
		1602	ev_timer_init (timer, callback, 60., 0.);
		1603	ev_timer_start (loop, timer);
		1604
		1605	Then, each time there is some activity, C<ev_timer_stop> it, initialise it
		1606	and start it again:
		1607
		1608	ev_timer_stop (loop, timer);
		1609	ev_timer_set (timer, 60., 0.);
		1610	ev_timer_start (loop, timer);
		1611
		1612	This is relatively simple to implement, but means that each time there is
		1613	some activity, libev will first have to remove the timer from its internal
		1614	data structure and then add it again. Libev tries to be fast, but it's
		1615	still not a constant-time operation.
		1616
		1617	=item 2. Use a timer and re-start it with C<ev_timer_again> inactivity.
		1618
		1619	This is the easiest way, and involves using C<ev_timer_again> instead of
		1620	C<ev_timer_start>.
		1621
		1622	To implement this, configure an C<ev_timer> with a C<repeat> value
		1623	of C<60> and then call C<ev_timer_again> at start and each time you
		1624	successfully read or write some data. If you go into an idle state where
		1625	you do not expect data to travel on the socket, you can C<ev_timer_stop>
		1626	the timer, and C<ev_timer_again> will automatically restart it if need be.
		1627
		1628	That means you can ignore both the C<ev_timer_start> function and the
		1629	C<after> argument to C<ev_timer_set>, and only ever use the C<repeat>
		1630	member and C<ev_timer_again>.
		1631
		1632	At start:
		1633
		1634	ev_init (timer, callback);
		1635	timer->repeat = 60.;
		1636	ev_timer_again (loop, timer);
		1637
		1638	Each time there is some activity:
		1639
		1640	ev_timer_again (loop, timer);
		1641
		1642	It is even possible to change the time-out on the fly, regardless of
		1643	whether the watcher is active or not:
		1644
		1645	timer->repeat = 30.;
		1646	ev_timer_again (loop, timer);
		1647
		1648	This is slightly more efficient then stopping/starting the timer each time
		1649	you want to modify its timeout value, as libev does not have to completely
		1650	remove and re-insert the timer from/into its internal data structure.
		1651
		1652	It is, however, even simpler than the "obvious" way to do it.
		1653
		1654	=item 3. Let the timer time out, but then re-arm it as required.
		1655
		1656	This method is more tricky, but usually most efficient: Most timeouts are
		1657	relatively long compared to the intervals between other activity - in
		1658	our example, within 60 seconds, there are usually many I/O events with
		1659	associated activity resets.
		1660
		1661	In this case, it would be more efficient to leave the C<ev_timer> alone,
		1662	but remember the time of last activity, and check for a real timeout only
		1663	within the callback:
		1664
		1665	ev_tstamp last_activity; // time of last activity
		1666
		1667	static void
		1668	callback (EV_P_ ev_timer *w, int revents)
		1669	{
		1670	ev_tstamp now = ev_now (EV_A);
		1671	ev_tstamp timeout = last_activity + 60.;
		1672
		1673	// if last_activity + 60. is older than now, we did time out
		1674	if (timeout < now)
		1675	{
		1676	// timeout occured, take action
		1677	}
		1678	else
		1679	{
		1680	// callback was invoked, but there was some activity, re-arm
		1681	// the watcher to fire in last_activity + 60, which is
		1682	// guaranteed to be in the future, so "again" is positive:
		1683	w->repeat = timeout - now;
		1684	ev_timer_again (EV_A_ w);
		1685	}
		1686	}
		1687
		1688	To summarise the callback: first calculate the real timeout (defined
		1689	as "60 seconds after the last activity"), then check if that time has
		1690	been reached, which means something I<did>, in fact, time out. Otherwise
		1691	the callback was invoked too early (C<timeout> is in the future), so
		1692	re-schedule the timer to fire at that future time, to see if maybe we have
		1693	a timeout then.
		1694
		1695	Note how C<ev_timer_again> is used, taking advantage of the
		1696	C<ev_timer_again> optimisation when the timer is already running.
		1697
		1698	This scheme causes more callback invocations (about one every 60 seconds
		1699	minus half the average time between activity), but virtually no calls to
		1700	libev to change the timeout.
		1701
		1702	To start the timer, simply initialise the watcher and set C<last_activity>
		1703	to the current time (meaning we just have some activity :), then call the
		1704	callback, which will "do the right thing" and start the timer:
		1705
		1706	ev_init (timer, callback);
		1707	last_activity = ev_now (loop);
		1708	callback (loop, timer, EV_TIMEOUT);
		1709
		1710	And when there is some activity, simply store the current time in
		1711	C<last_activity>, no libev calls at all:
		1712
		1713	last_actiivty = ev_now (loop);
		1714
		1715	This technique is slightly more complex, but in most cases where the
		1716	time-out is unlikely to be triggered, much more efficient.
		1717
		1718	Changing the timeout is trivial as well (if it isn't hard-coded in the
		1719	callback :) - just change the timeout and invoke the callback, which will
		1720	fix things for you.
		1721
		1722	=item 4. Wee, just use a double-linked list for your timeouts.
		1723
		1724	If there is not one request, but many thousands (millions...), all
		1725	employing some kind of timeout with the same timeout value, then one can
		1726	do even better:
		1727
		1728	When starting the timeout, calculate the timeout value and put the timeout
		1729	at the I<end> of the list.
		1730
		1731	Then use an C<ev_timer> to fire when the timeout at the I<beginning> of
		1732	the list is expected to fire (for example, using the technique #3).
		1733
		1734	When there is some activity, remove the timer from the list, recalculate
		1735	the timeout, append it to the end of the list again, and make sure to
		1736	update the C<ev_timer> if it was taken from the beginning of the list.
		1737
		1738	This way, one can manage an unlimited number of timeouts in O(1) time for
		1739	starting, stopping and updating the timers, at the expense of a major
		1740	complication, and having to use a constant timeout. The constant timeout
		1741	ensures that the list stays sorted.
		1742
		1743	=back
		1744
		1745	So which method the best?
		1746
		1747	Method #2 is a simple no-brain-required solution that is adequate in most
		1748	situations. Method #3 requires a bit more thinking, but handles many cases
		1749	better, and isn't very complicated either. In most case, choosing either
		1750	one is fine, with #3 being better in typical situations.
		1751
		1752	Method #1 is almost always a bad idea, and buys you nothing. Method #4 is
		1753	rather complicated, but extremely efficient, something that really pays
		1754	off after the first million or so of active timers, i.e. it's usually
		1755	overkill :)
		1756
		1757	=head3 The special problem of time updates
		1758
		1759	Establishing the current time is a costly operation (it usually takes at
		1760	least two system calls): EV therefore updates its idea of the current
		1761	time only before and after C<ev_loop> collects new events, which causes a
		1762	growing difference between C<ev_now ()> and C<ev_time ()> when handling
		1763	lots of events in one iteration.
987		1764
988	The relative timeouts are calculated relative to the C<ev_now ()>	1765	The relative timeouts are calculated relative to the C<ev_now ()>
989	time. This is usually the right thing as this timestamp refers to the time	1766	time. This is usually the right thing as this timestamp refers to the time
990	of the event triggering whatever timeout you are modifying/starting. If	1767	of the event triggering whatever timeout you are modifying/starting. If
991	you suspect event processing to be delayed and you I<need> to base the timeout	1768	you suspect event processing to be delayed and you I<need> to base the
992	on the current time, use something like this to adjust for this:	1769	timeout on the current time, use something like this to adjust for this:
993		1770
994	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);	1771	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);
995		1772
996	The callback is guarenteed to be invoked only when its timeout has passed,	1773	If the event loop is suspended for a long time, you can also force an
997	but if multiple timers become ready during the same loop iteration then	1774	update of the time returned by C<ev_now ()> by calling C<ev_now_update
998	order of execution is undefined.	1775	()>.
		1776
		1777	=head3 The special problems of suspended animation
		1778
		1779	When you leave the server world it is quite customary to hit machines that
		1780	can suspend/hibernate - what happens to the clocks during such a suspend?
		1781
		1782	Some quick tests made with a Linux 2.6.28 indicate that a suspend freezes
		1783	all processes, while the clocks (C<times>, C<CLOCK_MONOTONIC>) continue
		1784	to run until the system is suspended, but they will not advance while the
		1785	system is suspended. That means, on resume, it will be as if the program
		1786	was frozen for a few seconds, but the suspend time will not be counted
		1787	towards C<ev_timer> when a monotonic clock source is used. The real time
		1788	clock advanced as expected, but if it is used as sole clocksource, then a
		1789	long suspend would be detected as a time jump by libev, and timers would
		1790	be adjusted accordingly.
		1791
		1792	I would not be surprised to see different behaviour in different between
		1793	operating systems, OS versions or even different hardware.
		1794
		1795	The other form of suspend (job control, or sending a SIGSTOP) will see a
		1796	time jump in the monotonic clocks and the realtime clock. If the program
		1797	is suspended for a very long time, and monotonic clock sources are in use,
		1798	then you can expect C<ev_timer>s to expire as the full suspension time
		1799	will be counted towards the timers. When no monotonic clock source is in
		1800	use, then libev will again assume a timejump and adjust accordingly.
		1801
		1802	It might be beneficial for this latter case to call C<ev_suspend>
		1803	and C<ev_resume> in code that handles C<SIGTSTP>, to at least get
		1804	deterministic behaviour in this case (you can do nothing against
		1805	C<SIGSTOP>).
		1806
		1807	=head3 Watcher-Specific Functions and Data Members
999		1808
1000	=over 4	1809	=over 4
1001		1810
1002	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)	1811	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)
1003		1812
1004	=item ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)	1813	=item ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)
1005		1814
1006	Configure the timer to trigger after C<after> seconds. If C<repeat> is	1815	Configure the timer to trigger after C<after> seconds. If C<repeat>
1007	C<0.>, then it will automatically be stopped. If it is positive, then the	1816	is C<0.>, then it will automatically be stopped once the timeout is
1008	timer will automatically be configured to trigger again C<repeat> seconds	1817	reached. If it is positive, then the timer will automatically be
1009	later, again, and again, until stopped manually.	1818	configured to trigger again C<repeat> seconds later, again, and again,
		1819	until stopped manually.
1010		1820
1011	The timer itself will do a best-effort at avoiding drift, that is, if you	1821	The timer itself will do a best-effort at avoiding drift, that is, if
1012	configure a timer to trigger every 10 seconds, then it will trigger at	1822	you configure a timer to trigger every 10 seconds, then it will normally
1013	exactly 10 second intervals. If, however, your program cannot keep up with	1823	trigger at exactly 10 second intervals. If, however, your program cannot
1014	the timer (because it takes longer than those 10 seconds to do stuff) the	1824	keep up with the timer (because it takes longer than those 10 seconds to
1015	timer will not fire more than once per event loop iteration.	1825	do stuff) the timer will not fire more than once per event loop iteration.
1016		1826
1017	=item ev_timer_again (loop)	1827	=item ev_timer_again (loop, ev_timer *)
1018		1828
1019	This will act as if the timer timed out and restart it again if it is	1829	This will act as if the timer timed out and restart it again if it is
1020	repeating. The exact semantics are:	1830	repeating. The exact semantics are:
1021		1831
1022	If the timer is pending, its pending status is cleared.	1832	If the timer is pending, its pending status is cleared.
1023		1833
1024	If the timer is started but nonrepeating, stop it (as if it timed out).	1834	If the timer is started but non-repeating, stop it (as if it timed out).
1025		1835
1026	If the timer is repeating, either start it if necessary (with the	1836	If the timer is repeating, either start it if necessary (with the
1027	C<repeat> value), or reset the running timer to the C<repeat> value.	1837	C<repeat> value), or reset the running timer to the C<repeat> value.
1028		1838
1029	This sounds a bit complicated, but here is a useful and typical	1839	This sounds a bit complicated, see L<Be smart about timeouts>, above, for a
1030	example: Imagine you have a tcp connection and you want a so-called idle	1840	usage example.
1031	timeout, that is, you want to be called when there have been, say, 60
1032	seconds of inactivity on the socket. The easiest way to do this is to
1033	configure an C<ev_timer> with a C<repeat> value of C<60> and then call
1034	C<ev_timer_again> each time you successfully read or write some data. If
1035	you go into an idle state where you do not expect data to travel on the
1036	socket, you can C<ev_timer_stop> the timer, and C<ev_timer_again> will
1037	automatically restart it if need be.
1038		1841
1039	That means you can ignore the C<after> value and C<ev_timer_start>	1842	=item ev_timer_remaining (loop, ev_timer *)
1040	altogether and only ever use the C<repeat> value and C<ev_timer_again>:
1041		1843
1042	ev_timer_init (timer, callback, 0., 5.);	1844	Returns the remaining time until a timer fires. If the timer is active,
1043	ev_timer_again (loop, timer);	1845	then this time is relative to the current event loop time, otherwise it's
1044	...	1846	the timeout value currently configured.
1045	timer->again = 17.;
1046	ev_timer_again (loop, timer);
1047	...
1048	timer->again = 10.;
1049	ev_timer_again (loop, timer);
1050		1847
1051	This is more slightly efficient then stopping/starting the timer each time	1848	That is, after an C<ev_timer_set (w, 5, 7)>, C<ev_timer_remaining> returns
1052	you want to modify its timeout value.	1849	C<5>. When the timer is started and one second passes, C<ev_timer_remain>
		1850	will return C<4>. When the timer expires and is restarted, it will return
		1851	roughly C<7> (likely slightly less as callback invocation takes some time,
		1852	too), and so on.
1053		1853
1054	=item ev_tstamp repeat [read-write]	1854	=item ev_tstamp repeat [read-write]
1055		1855
1056	The current C<repeat> value. Will be used each time the watcher times out	1856	The current C<repeat> value. Will be used each time the watcher times out
1057	or C<ev_timer_again> is called and determines the next timeout (if any),	1857	or C<ev_timer_again> is called, and determines the next timeout (if any),
1058	which is also when any modifications are taken into account.	1858	which is also when any modifications are taken into account.
1059		1859
1060	=back	1860	=back
1061		1861
		1862	=head3 Examples
		1863
1062	Example: Create a timer that fires after 60 seconds.	1864	Example: Create a timer that fires after 60 seconds.
1063		1865
1064	static void	1866	static void
1065	one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)	1867	one_minute_cb (struct ev_loop *loop, ev_timer *w, int revents)
1066	{	1868	{
1067	.. one minute over, w is actually stopped right here	1869	.. one minute over, w is actually stopped right here
1068	}	1870	}
1069		1871
1070	struct ev_timer mytimer;	1872	ev_timer mytimer;
1071	ev_timer_init (&mytimer, one_minute_cb, 60., 0.);	1873	ev_timer_init (&mytimer, one_minute_cb, 60., 0.);
1072	ev_timer_start (loop, &mytimer);	1874	ev_timer_start (loop, &mytimer);
1073		1875
1074	Example: Create a timeout timer that times out after 10 seconds of	1876	Example: Create a timeout timer that times out after 10 seconds of
1075	inactivity.	1877	inactivity.
1076		1878
1077	static void	1879	static void
1078	timeout_cb (struct ev_loop *loop, struct ev_timer *w, int revents)	1880	timeout_cb (struct ev_loop *loop, ev_timer *w, int revents)
1079	{	1881	{
1080	.. ten seconds without any activity	1882	.. ten seconds without any activity
1081	}	1883	}
1082		1884
1083	struct ev_timer mytimer;	1885	ev_timer mytimer;
1084	ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */	1886	ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */
1085	ev_timer_again (&mytimer); /* start timer */	1887	ev_timer_again (&mytimer); /* start timer */
1086	ev_loop (loop, 0);	1888	ev_loop (loop, 0);
1087		1889
1088	// and in some piece of code that gets executed on any "activity":	1890	// and in some piece of code that gets executed on any "activity":
1089	// reset the timeout to start ticking again at 10 seconds	1891	// reset the timeout to start ticking again at 10 seconds
1090	ev_timer_again (&mytimer);	1892	ev_timer_again (&mytimer);
1091		1893
1092		1894
1093	=head2 C<ev_periodic> - to cron or not to cron?	1895	=head2 C<ev_periodic> - to cron or not to cron?
1094		1896
1095	Periodic watchers are also timers of a kind, but they are very versatile	1897	Periodic watchers are also timers of a kind, but they are very versatile
1096	(and unfortunately a bit complex).	1898	(and unfortunately a bit complex).
1097		1899
1098	Unlike C<ev_timer>'s, they are not based on real time (or relative time)	1900	Unlike C<ev_timer>, periodic watchers are not based on real time (or
1099	but on wallclock time (absolute time). You can tell a periodic watcher	1901	relative time, the physical time that passes) but on wall clock time
1100	to trigger "at" some specific point in time. For example, if you tell a	1902	(absolute time, the thing you can read on your calender or clock). The
1101	periodic watcher to trigger in 10 seconds (by specifiying e.g. C<ev_now ()	1903	difference is that wall clock time can run faster or slower than real
1102	+ 10.>) and then reset your system clock to the last year, then it will	1904	time, and time jumps are not uncommon (e.g. when you adjust your
1103	take a year to trigger the event (unlike an C<ev_timer>, which would trigger	1905	wrist-watch).
1104	roughly 10 seconds later).
1105		1906
1106	They can also be used to implement vastly more complex timers, such as	1907	You can tell a periodic watcher to trigger after some specific point
1107	triggering an event on each midnight, local time or other, complicated,	1908	in time: for example, if you tell a periodic watcher to trigger "in 10
1108	rules.	1909	seconds" (by specifying e.g. C<ev_now () + 10.>, that is, an absolute time
		1910	not a delay) and then reset your system clock to January of the previous
		1911	year, then it will take a year or more to trigger the event (unlike an
		1912	C<ev_timer>, which would still trigger roughly 10 seconds after starting
		1913	it, as it uses a relative timeout).
1109		1914
		1915	C<ev_periodic> watchers can also be used to implement vastly more complex
		1916	timers, such as triggering an event on each "midnight, local time", or
		1917	other complicated rules. This cannot be done with C<ev_timer> watchers, as
		1918	those cannot react to time jumps.
		1919
1110	As with timers, the callback is guarenteed to be invoked only when the	1920	As with timers, the callback is guaranteed to be invoked only when the
1111	time (C<at>) has been passed, but if multiple periodic timers become ready	1921	point in time where it is supposed to trigger has passed. If multiple
1112	during the same loop iteration then order of execution is undefined.	1922	timers become ready during the same loop iteration then the ones with
		1923	earlier time-out values are invoked before ones with later time-out values
		1924	(but this is no longer true when a callback calls C<ev_loop> recursively).
		1925
		1926	=head3 Watcher-Specific Functions and Data Members
1113		1927
1114	=over 4	1928	=over 4
1115		1929
1116	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)	1930	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp offset, ev_tstamp interval, reschedule_cb)
1117		1931
1118	=item ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)	1932	=item ev_periodic_set (ev_periodic *, ev_tstamp offset, ev_tstamp interval, reschedule_cb)
1119		1933
1120	Lots of arguments, lets sort it out... There are basically three modes of	1934	Lots of arguments, let's sort it out... There are basically three modes of
1121	operation, and we will explain them from simplest to complex:	1935	operation, and we will explain them from simplest to most complex:
1122		1936
1123	=over 4	1937	=over 4
1124		1938
1125	=item * absolute timer (at = time, interval = reschedule_cb = 0)	1939	=item * absolute timer (offset = absolute time, interval = 0, reschedule_cb = 0)
1126		1940
1127	In this configuration the watcher triggers an event at the wallclock time	1941	In this configuration the watcher triggers an event after the wall clock
1128	C<at> and doesn't repeat. It will not adjust when a time jump occurs,	1942	time C<offset> has passed. It will not repeat and will not adjust when a
1129	that is, if it is to be run at January 1st 2011 then it will run when the	1943	time jump occurs, that is, if it is to be run at January 1st 2011 then it
1130	system time reaches or surpasses this time.	1944	will be stopped and invoked when the system clock reaches or surpasses
		1945	this point in time.
1131		1946
1132	=item * non-repeating interval timer (at = offset, interval > 0, reschedule_cb = 0)	1947	=item * repeating interval timer (offset = offset within interval, interval > 0, reschedule_cb = 0)
1133		1948
1134	In this mode the watcher will always be scheduled to time out at the next	1949	In this mode the watcher will always be scheduled to time out at the next
1135	C<at + N * interval> time (for some integer N, which can also be negative)	1950	C<offset + N * interval> time (for some integer N, which can also be
1136	and then repeat, regardless of any time jumps.	1951	negative) and then repeat, regardless of any time jumps. The C<offset>
		1952	argument is merely an offset into the C<interval> periods.
1137		1953
1138	This can be used to create timers that do not drift with respect to system	1954	This can be used to create timers that do not drift with respect to the
1139	time:	1955	system clock, for example, here is an C<ev_periodic> that triggers each
		1956	hour, on the hour (with respect to UTC):
1140		1957
1141	ev_periodic_set (&periodic, 0., 3600., 0);	1958	ev_periodic_set (&periodic, 0., 3600., 0);
1142		1959
1143	This doesn't mean there will always be 3600 seconds in between triggers,	1960	This doesn't mean there will always be 3600 seconds in between triggers,
1144	but only that the the callback will be called when the system time shows a	1961	but only that the callback will be called when the system time shows a
1145	full hour (UTC), or more correctly, when the system time is evenly divisible	1962	full hour (UTC), or more correctly, when the system time is evenly divisible
1146	by 3600.	1963	by 3600.
1147		1964
1148	Another way to think about it (for the mathematically inclined) is that	1965	Another way to think about it (for the mathematically inclined) is that
1149	C<ev_periodic> will try to run the callback in this mode at the next possible	1966	C<ev_periodic> will try to run the callback in this mode at the next possible
1150	time where C<time = at (mod interval)>, regardless of any time jumps.	1967	time where C<time = offset (mod interval)>, regardless of any time jumps.
1151		1968
1152	For numerical stability it is preferable that the C<at> value is near	1969	For numerical stability it is preferable that the C<offset> value is near
1153	C<ev_now ()> (the current time), but there is no range requirement for	1970	C<ev_now ()> (the current time), but there is no range requirement for
1154	this value.	1971	this value, and in fact is often specified as zero.
1155		1972
		1973	Note also that there is an upper limit to how often a timer can fire (CPU
		1974	speed for example), so if C<interval> is very small then timing stability
		1975	will of course deteriorate. Libev itself tries to be exact to be about one
		1976	millisecond (if the OS supports it and the machine is fast enough).
		1977
1156	=item * manual reschedule mode (at and interval ignored, reschedule_cb = callback)	1978	=item * manual reschedule mode (offset ignored, interval ignored, reschedule_cb = callback)
1157		1979
1158	In this mode the values for C<interval> and C<at> are both being	1980	In this mode the values for C<interval> and C<offset> are both being
1159	ignored. Instead, each time the periodic watcher gets scheduled, the	1981	ignored. Instead, each time the periodic watcher gets scheduled, the
1160	reschedule callback will be called with the watcher as first, and the	1982	reschedule callback will be called with the watcher as first, and the
1161	current time as second argument.	1983	current time as second argument.
1162		1984
1163	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher,	1985	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher, ever,
1164	ever, or make any event loop modifications>. If you need to stop it,	1986	or make ANY other event loop modifications whatsoever, unless explicitly
1165	return C<now + 1e30> (or so, fudge fudge) and stop it afterwards (e.g. by	1987	allowed by documentation here>.
1166	starting an C<ev_prepare> watcher, which is legal).
1167		1988
		1989	If you need to stop it, return C<now + 1e30> (or so, fudge fudge) and stop
		1990	it afterwards (e.g. by starting an C<ev_prepare> watcher, which is the
		1991	only event loop modification you are allowed to do).
		1992
1168	Its prototype is C<ev_tstamp (*reschedule_cb)(struct ev_periodic *w,	1993	The callback prototype is C<ev_tstamp (*reschedule_cb)(ev_periodic
1169	ev_tstamp now)>, e.g.:	1994	*w, ev_tstamp now)>, e.g.:
1170		1995
		1996	static ev_tstamp
1171	static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)	1997	my_rescheduler (ev_periodic *w, ev_tstamp now)
1172	{	1998	{
1173	return now + 60.;	1999	return now + 60.;
1174	}	2000	}
1175		2001
1176	It must return the next time to trigger, based on the passed time value	2002	It must return the next time to trigger, based on the passed time value
1177	(that is, the lowest time value larger than to the second argument). It	2003	(that is, the lowest time value larger than to the second argument). It
1178	will usually be called just before the callback will be triggered, but	2004	will usually be called just before the callback will be triggered, but
1179	might be called at other times, too.	2005	might be called at other times, too.
1180		2006
1181	NOTE: I<< This callback must always return a time that is later than the	2007	NOTE: I<< This callback must always return a time that is higher than or
1182	passed C<now> value >>. Not even C<now> itself will do, it I<must> be larger.	2008	equal to the passed C<now> value >>.
1183		2009
1184	This can be used to create very complex timers, such as a timer that	2010	This can be used to create very complex timers, such as a timer that
1185	triggers on each midnight, local time. To do this, you would calculate the	2011	triggers on "next midnight, local time". To do this, you would calculate the
1186	next midnight after C<now> and return the timestamp value for this. How	2012	next midnight after C<now> and return the timestamp value for this. How
1187	you do this is, again, up to you (but it is not trivial, which is the main	2013	you do this is, again, up to you (but it is not trivial, which is the main
1188	reason I omitted it as an example).	2014	reason I omitted it as an example).
1189		2015
1190	=back	2016	=back
…		…
1194	Simply stops and restarts the periodic watcher again. This is only useful	2020	Simply stops and restarts the periodic watcher again. This is only useful
1195	when you changed some parameters or the reschedule callback would return	2021	when you changed some parameters or the reschedule callback would return
1196	a different time than the last time it was called (e.g. in a crond like	2022	a different time than the last time it was called (e.g. in a crond like
1197	program when the crontabs have changed).	2023	program when the crontabs have changed).
1198		2024
		2025	=item ev_tstamp ev_periodic_at (ev_periodic *)
		2026
		2027	When active, returns the absolute time that the watcher is supposed
		2028	to trigger next. This is not the same as the C<offset> argument to
		2029	C<ev_periodic_set>, but indeed works even in interval and manual
		2030	rescheduling modes.
		2031
1199	=item ev_tstamp offset [read-write]	2032	=item ev_tstamp offset [read-write]
1200		2033
1201	When repeating, this contains the offset value, otherwise this is the	2034	When repeating, this contains the offset value, otherwise this is the
1202	absolute point in time (the C<at> value passed to C<ev_periodic_set>).	2035	absolute point in time (the C<offset> value passed to C<ev_periodic_set>,
		2036	although libev might modify this value for better numerical stability).
1203		2037
1204	Can be modified any time, but changes only take effect when the periodic	2038	Can be modified any time, but changes only take effect when the periodic
1205	timer fires or C<ev_periodic_again> is being called.	2039	timer fires or C<ev_periodic_again> is being called.
1206		2040
1207	=item ev_tstamp interval [read-write]	2041	=item ev_tstamp interval [read-write]
1208		2042
1209	The current interval value. Can be modified any time, but changes only	2043	The current interval value. Can be modified any time, but changes only
1210	take effect when the periodic timer fires or C<ev_periodic_again> is being	2044	take effect when the periodic timer fires or C<ev_periodic_again> is being
1211	called.	2045	called.
1212		2046
1213	=item ev_tstamp (*reschedule_cb)(struct ev_periodic *w, ev_tstamp now) [read-write]	2047	=item ev_tstamp (*reschedule_cb)(ev_periodic *w, ev_tstamp now) [read-write]
1214		2048
1215	The current reschedule callback, or C<0>, if this functionality is	2049	The current reschedule callback, or C<0>, if this functionality is
1216	switched off. Can be changed any time, but changes only take effect when	2050	switched off. Can be changed any time, but changes only take effect when
1217	the periodic timer fires or C<ev_periodic_again> is being called.	2051	the periodic timer fires or C<ev_periodic_again> is being called.
1218		2052
1219	=back	2053	=back
1220		2054
		2055	=head3 Examples
		2056
1221	Example: Call a callback every hour, or, more precisely, whenever the	2057	Example: Call a callback every hour, or, more precisely, whenever the
1222	system clock is divisible by 3600. The callback invocation times have	2058	system time is divisible by 3600. The callback invocation times have
1223	potentially a lot of jittering, but good long-term stability.	2059	potentially a lot of jitter, but good long-term stability.
1224		2060
1225	static void	2061	static void
1226	clock_cb (struct ev_loop *loop, struct ev_io *w, int revents)	2062	clock_cb (struct ev_loop *loop, ev_io *w, int revents)
1227	{	2063	{
1228	... its now a full hour (UTC, or TAI or whatever your clock follows)	2064	... its now a full hour (UTC, or TAI or whatever your clock follows)
1229	}	2065	}
1230		2066
1231	struct ev_periodic hourly_tick;	2067	ev_periodic hourly_tick;
1232	ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);	2068	ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);
1233	ev_periodic_start (loop, &hourly_tick);	2069	ev_periodic_start (loop, &hourly_tick);
1234		2070
1235	Example: The same as above, but use a reschedule callback to do it:	2071	Example: The same as above, but use a reschedule callback to do it:
1236		2072
1237	#include <math.h>	2073	#include <math.h>
1238		2074
1239	static ev_tstamp	2075	static ev_tstamp
1240	my_scheduler_cb (struct ev_periodic *w, ev_tstamp now)	2076	my_scheduler_cb (ev_periodic *w, ev_tstamp now)
1241	{	2077	{
1242	return fmod (now, 3600.) + 3600.;	2078	return now + (3600. - fmod (now, 3600.));
1243	}	2079	}
1244		2080
1245	ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);	2081	ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
1246		2082
1247	Example: Call a callback every hour, starting now:	2083	Example: Call a callback every hour, starting now:
1248		2084
1249	struct ev_periodic hourly_tick;	2085	ev_periodic hourly_tick;
1250	ev_periodic_init (&hourly_tick, clock_cb,	2086	ev_periodic_init (&hourly_tick, clock_cb,
1251	fmod (ev_now (loop), 3600.), 3600., 0);	2087	fmod (ev_now (loop), 3600.), 3600., 0);
1252	ev_periodic_start (loop, &hourly_tick);	2088	ev_periodic_start (loop, &hourly_tick);
1253		2089
1254		2090
1255	=head2 C<ev_signal> - signal me when a signal gets signalled!	2091	=head2 C<ev_signal> - signal me when a signal gets signalled!
1256		2092
1257	Signal watchers will trigger an event when the process receives a specific	2093	Signal watchers will trigger an event when the process receives a specific
1258	signal one or more times. Even though signals are very asynchronous, libev	2094	signal one or more times. Even though signals are very asynchronous, libev
1259	will try it's best to deliver signals synchronously, i.e. as part of the	2095	will try it's best to deliver signals synchronously, i.e. as part of the
1260	normal event processing, like any other event.	2096	normal event processing, like any other event.
1261		2097
		2098	If you want signals to be delivered truly asynchronously, just use
		2099	C<sigaction> as you would do without libev and forget about sharing
		2100	the signal. You can even use C<ev_async> from a signal handler to
		2101	synchronously wake up an event loop.
		2102
1262	You can configure as many watchers as you like per signal. Only when the	2103	You can configure as many watchers as you like for the same signal, but
		2104	only within the same loop, i.e. you can watch for C<SIGINT> in your
		2105	default loop and for C<SIGIO> in another loop, but you cannot watch for
		2106	C<SIGINT> in both the default loop and another loop at the same time. At
		2107	the moment, C<SIGCHLD> is permanently tied to the default loop.
		2108
1263	first watcher gets started will libev actually register a signal watcher	2109	When the first watcher gets started will libev actually register something
1264	with the kernel (thus it coexists with your own signal handlers as long	2110	with the kernel (thus it coexists with your own signal handlers as long as
1265	as you don't register any with libev). Similarly, when the last signal	2111	you don't register any with libev for the same signal).
1266	watcher for a signal is stopped libev will reset the signal handler to	2112
1267	SIG_DFL (regardless of what it was set to before).	2113	If possible and supported, libev will install its handlers with
		2114	C<SA_RESTART> (or equivalent) behaviour enabled, so system calls should
		2115	not be unduly interrupted. If you have a problem with system calls getting
		2116	interrupted by signals you can block all signals in an C<ev_check> watcher
		2117	and unblock them in an C<ev_prepare> watcher.
		2118
		2119	=head3 The special problem of inheritance over execve
		2120
		2121	Both the signal mask (C<sigprocmask>) and the signal disposition
		2122	(C<sigaction>) are unspecified after starting a signal watcher (and after
		2123	stopping it again), that is, libev might or might not block the signal,
		2124	and might or might not set or restore the installed signal handler.
		2125
		2126	While this does not matter for the signal disposition (libev never
		2127	sets signals to C<SIG_IGN>, so handlers will be reset to C<SIG_DFL> on
		2128	C<execve>), this matters for the signal mask: many programs do not expect
		2129	certain signals to be blocked.
		2130
		2131	This means that before calling C<exec> (from the child) you should reset
		2132	the signal mask to whatever "default" you expect (all clear is a good
		2133	choice usually).
		2134
		2135	In current versions of libev, you can ensure that the signal mask is not
		2136	blocking any signals (except temporarily, so thread users watch out) by
		2137	specifying the C<EVFLAG_NOSIGNALFD> when creating the event loop. This is
		2138	not guaranteed for future versions, however.
		2139
		2140	=head3 Watcher-Specific Functions and Data Members
1268		2141
1269	=over 4	2142	=over 4
1270		2143
1271	=item ev_signal_init (ev_signal *, callback, int signum)	2144	=item ev_signal_init (ev_signal *, callback, int signum)
1272		2145
…		…
1279		2152
1280	The signal the watcher watches out for.	2153	The signal the watcher watches out for.
1281		2154
1282	=back	2155	=back
1283		2156
		2157	=head3 Examples
		2158
		2159	Example: Try to exit cleanly on SIGINT.
		2160
		2161	static void
		2162	sigint_cb (struct ev_loop *loop, ev_signal *w, int revents)
		2163	{
		2164	ev_unloop (loop, EVUNLOOP_ALL);
		2165	}
		2166
		2167	ev_signal signal_watcher;
		2168	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
		2169	ev_signal_start (loop, &signal_watcher);
		2170
1284		2171
1285	=head2 C<ev_child> - watch out for process status changes	2172	=head2 C<ev_child> - watch out for process status changes
1286		2173
1287	Child watchers trigger when your process receives a SIGCHLD in response to	2174	Child watchers trigger when your process receives a SIGCHLD in response to
1288	some child status changes (most typically when a child of yours dies).	2175	some child status changes (most typically when a child of yours dies or
		2176	exits). It is permissible to install a child watcher I<after> the child
		2177	has been forked (which implies it might have already exited), as long
		2178	as the event loop isn't entered (or is continued from a watcher), i.e.,
		2179	forking and then immediately registering a watcher for the child is fine,
		2180	but forking and registering a watcher a few event loop iterations later or
		2181	in the next callback invocation is not.
		2182
		2183	Only the default event loop is capable of handling signals, and therefore
		2184	you can only register child watchers in the default event loop.
		2185
		2186	Due to some design glitches inside libev, child watchers will always be
		2187	handled at maximum priority (their priority is set to C<EV_MAXPRI> by
		2188	libev)
		2189
		2190	=head3 Process Interaction
		2191
		2192	Libev grabs C<SIGCHLD> as soon as the default event loop is
		2193	initialised. This is necessary to guarantee proper behaviour even if the
		2194	first child watcher is started after the child exits. The occurrence
		2195	of C<SIGCHLD> is recorded asynchronously, but child reaping is done
		2196	synchronously as part of the event loop processing. Libev always reaps all
		2197	children, even ones not watched.
		2198
		2199	=head3 Overriding the Built-In Processing
		2200
		2201	Libev offers no special support for overriding the built-in child
		2202	processing, but if your application collides with libev's default child
		2203	handler, you can override it easily by installing your own handler for
		2204	C<SIGCHLD> after initialising the default loop, and making sure the
		2205	default loop never gets destroyed. You are encouraged, however, to use an
		2206	event-based approach to child reaping and thus use libev's support for
		2207	that, so other libev users can use C<ev_child> watchers freely.
		2208
		2209	=head3 Stopping the Child Watcher
		2210
		2211	Currently, the child watcher never gets stopped, even when the
		2212	child terminates, so normally one needs to stop the watcher in the
		2213	callback. Future versions of libev might stop the watcher automatically
		2214	when a child exit is detected (calling C<ev_child_stop> twice is not a
		2215	problem).
		2216
		2217	=head3 Watcher-Specific Functions and Data Members
1289		2218
1290	=over 4	2219	=over 4
1291		2220
1292	=item ev_child_init (ev_child *, callback, int pid)	2221	=item ev_child_init (ev_child *, callback, int pid, int trace)
1293		2222
1294	=item ev_child_set (ev_child *, int pid)	2223	=item ev_child_set (ev_child *, int pid, int trace)
1295		2224
1296	Configures the watcher to wait for status changes of process C<pid> (or	2225	Configures the watcher to wait for status changes of process C<pid> (or
1297	I<any> process if C<pid> is specified as C<0>). The callback can look	2226	I<any> process if C<pid> is specified as C<0>). The callback can look
1298	at the C<rstatus> member of the C<ev_child> watcher structure to see	2227	at the C<rstatus> member of the C<ev_child> watcher structure to see
1299	the status word (use the macros from C<sys/wait.h> and see your systems	2228	the status word (use the macros from C<sys/wait.h> and see your systems
1300	C<waitpid> documentation). The C<rpid> member contains the pid of the	2229	C<waitpid> documentation). The C<rpid> member contains the pid of the
1301	process causing the status change.	2230	process causing the status change. C<trace> must be either C<0> (only
		2231	activate the watcher when the process terminates) or C<1> (additionally
		2232	activate the watcher when the process is stopped or continued).
1302		2233
1303	=item int pid [read-only]	2234	=item int pid [read-only]
1304		2235
1305	The process id this watcher watches out for, or C<0>, meaning any process id.	2236	The process id this watcher watches out for, or C<0>, meaning any process id.
1306		2237
…		…
1313	The process exit/trace status caused by C<rpid> (see your systems	2244	The process exit/trace status caused by C<rpid> (see your systems
1314	C<waitpid> and C<sys/wait.h> documentation for details).	2245	C<waitpid> and C<sys/wait.h> documentation for details).
1315		2246
1316	=back	2247	=back
1317		2248
1318	Example: Try to exit cleanly on SIGINT and SIGTERM.	2249	=head3 Examples
1319		2250
		2251	Example: C<fork()> a new process and install a child handler to wait for
		2252	its completion.
		2253
		2254	ev_child cw;
		2255
1320	static void	2256	static void
1321	sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)	2257	child_cb (EV_P_ ev_child *w, int revents)
1322	{	2258	{
1323	ev_unloop (loop, EVUNLOOP_ALL);	2259	ev_child_stop (EV_A_ w);
		2260	printf ("process %d exited with status %x\n", w->rpid, w->rstatus);
1324	}	2261	}
1325		2262
1326	struct ev_signal signal_watcher;	2263	pid_t pid = fork ();
1327	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);	2264
1328	ev_signal_start (loop, &sigint_cb);	2265	if (pid < 0)
		2266	// error
		2267	else if (pid == 0)
		2268	{
		2269	// the forked child executes here
		2270	exit (1);
		2271	}
		2272	else
		2273	{
		2274	ev_child_init (&cw, child_cb, pid, 0);
		2275	ev_child_start (EV_DEFAULT_ &cw);
		2276	}
1329		2277
1330		2278
1331	=head2 C<ev_stat> - did the file attributes just change?	2279	=head2 C<ev_stat> - did the file attributes just change?
1332		2280
1333	This watches a filesystem path for attribute changes. That is, it calls	2281	This watches a file system path for attribute changes. That is, it calls
1334	C<stat> regularly (or when the OS says it changed) and sees if it changed	2282	C<stat> on that path in regular intervals (or when the OS says it changed)
1335	compared to the last time, invoking the callback if it did.	2283	and sees if it changed compared to the last time, invoking the callback if
		2284	it did.
1336		2285
1337	The path does not need to exist: changing from "path exists" to "path does	2286	The path does not need to exist: changing from "path exists" to "path does
1338	not exist" is a status change like any other. The condition "path does	2287	not exist" is a status change like any other. The condition "path does not
1339	not exist" is signified by the C<st_nlink> field being zero (which is	2288	exist" (or more correctly "path cannot be stat'ed") is signified by the
1340	otherwise always forced to be at least one) and all the other fields of	2289	C<st_nlink> field being zero (which is otherwise always forced to be at
1341	the stat buffer having unspecified contents.	2290	least one) and all the other fields of the stat buffer having unspecified
		2291	contents.
1342		2292
1343	The path I<should> be absolute and I<must not> end in a slash. If it is	2293	The path I<must not> end in a slash or contain special components such as
		2294	C<.> or C<..>. The path I<should> be absolute: If it is relative and
1344	relative and your working directory changes, the behaviour is undefined.	2295	your working directory changes, then the behaviour is undefined.
1345		2296
1346	Since there is no standard to do this, the portable implementation simply	2297	Since there is no portable change notification interface available, the
1347	calls C<stat (2)> regularly on the path to see if it changed somehow. You	2298	portable implementation simply calls C<stat(2)> regularly on the path
1348	can specify a recommended polling interval for this case. If you specify	2299	to see if it changed somehow. You can specify a recommended polling
1349	a polling interval of C<0> (highly recommended!) then a I<suitable,	2300	interval for this case. If you specify a polling interval of C<0> (highly
1350	unspecified default> value will be used (which you can expect to be around	2301	recommended!) then a I<suitable, unspecified default> value will be used
1351	five seconds, although this might change dynamically). Libev will also	2302	(which you can expect to be around five seconds, although this might
1352	impose a minimum interval which is currently around C<0.1>, but thats	2303	change dynamically). Libev will also impose a minimum interval which is
1353	usually overkill.	2304	currently around C<0.1>, but that's usually overkill.
1354		2305
1355	This watcher type is not meant for massive numbers of stat watchers,	2306	This watcher type is not meant for massive numbers of stat watchers,
1356	as even with OS-supported change notifications, this can be	2307	as even with OS-supported change notifications, this can be
1357	resource-intensive.	2308	resource-intensive.
1358		2309
1359	At the time of this writing, only the Linux inotify interface is	2310	At the time of this writing, the only OS-specific interface implemented
1360	implemented (implementing kqueue support is left as an exercise for the	2311	is the Linux inotify interface (implementing kqueue support is left as an
1361	reader). Inotify will be used to give hints only and should not change the	2312	exercise for the reader. Note, however, that the author sees no way of
1362	semantics of C<ev_stat> watchers, which means that libev sometimes needs	2313	implementing C<ev_stat> semantics with kqueue, except as a hint).
1363	to fall back to regular polling again even with inotify, but changes are	2314
1364	usually detected immediately, and if the file exists there will be no	2315	=head3 ABI Issues (Largefile Support)
1365	polling.	2316
		2317	Libev by default (unless the user overrides this) uses the default
		2318	compilation environment, which means that on systems with large file
		2319	support disabled by default, you get the 32 bit version of the stat
		2320	structure. When using the library from programs that change the ABI to
		2321	use 64 bit file offsets the programs will fail. In that case you have to
		2322	compile libev with the same flags to get binary compatibility. This is
		2323	obviously the case with any flags that change the ABI, but the problem is
		2324	most noticeably displayed with ev_stat and large file support.
		2325
		2326	The solution for this is to lobby your distribution maker to make large
		2327	file interfaces available by default (as e.g. FreeBSD does) and not
		2328	optional. Libev cannot simply switch on large file support because it has
		2329	to exchange stat structures with application programs compiled using the
		2330	default compilation environment.
		2331
		2332	=head3 Inotify and Kqueue
		2333
		2334	When C<inotify (7)> support has been compiled into libev and present at
		2335	runtime, it will be used to speed up change detection where possible. The
		2336	inotify descriptor will be created lazily when the first C<ev_stat>
		2337	watcher is being started.
		2338
		2339	Inotify presence does not change the semantics of C<ev_stat> watchers
		2340	except that changes might be detected earlier, and in some cases, to avoid
		2341	making regular C<stat> calls. Even in the presence of inotify support
		2342	there are many cases where libev has to resort to regular C<stat> polling,
		2343	but as long as kernel 2.6.25 or newer is used (2.6.24 and older have too
		2344	many bugs), the path exists (i.e. stat succeeds), and the path resides on
		2345	a local filesystem (libev currently assumes only ext2/3, jfs, reiserfs and
		2346	xfs are fully working) libev usually gets away without polling.
		2347
		2348	There is no support for kqueue, as apparently it cannot be used to
		2349	implement this functionality, due to the requirement of having a file
		2350	descriptor open on the object at all times, and detecting renames, unlinks
		2351	etc. is difficult.
		2352
		2353	=head3 C<stat ()> is a synchronous operation
		2354
		2355	Libev doesn't normally do any kind of I/O itself, and so is not blocking
		2356	the process. The exception are C<ev_stat> watchers - those call C<stat
		2357	()>, which is a synchronous operation.
		2358
		2359	For local paths, this usually doesn't matter: unless the system is very
		2360	busy or the intervals between stat's are large, a stat call will be fast,
		2361	as the path data is usually in memory already (except when starting the
		2362	watcher).
		2363
		2364	For networked file systems, calling C<stat ()> can block an indefinite
		2365	time due to network issues, and even under good conditions, a stat call
		2366	often takes multiple milliseconds.
		2367
		2368	Therefore, it is best to avoid using C<ev_stat> watchers on networked
		2369	paths, although this is fully supported by libev.
		2370
		2371	=head3 The special problem of stat time resolution
		2372
		2373	The C<stat ()> system call only supports full-second resolution portably,
		2374	and even on systems where the resolution is higher, most file systems
		2375	still only support whole seconds.
		2376
		2377	That means that, if the time is the only thing that changes, you can
		2378	easily miss updates: on the first update, C<ev_stat> detects a change and
		2379	calls your callback, which does something. When there is another update
		2380	within the same second, C<ev_stat> will be unable to detect unless the
		2381	stat data does change in other ways (e.g. file size).
		2382
		2383	The solution to this is to delay acting on a change for slightly more
		2384	than a second (or till slightly after the next full second boundary), using
		2385	a roughly one-second-delay C<ev_timer> (e.g. C<ev_timer_set (w, 0., 1.02);
		2386	ev_timer_again (loop, w)>).
		2387
		2388	The C<.02> offset is added to work around small timing inconsistencies
		2389	of some operating systems (where the second counter of the current time
		2390	might be be delayed. One such system is the Linux kernel, where a call to
		2391	C<gettimeofday> might return a timestamp with a full second later than
		2392	a subsequent C<time> call - if the equivalent of C<time ()> is used to
		2393	update file times then there will be a small window where the kernel uses
		2394	the previous second to update file times but libev might already execute
		2395	the timer callback).
		2396
		2397	=head3 Watcher-Specific Functions and Data Members
1366		2398
1367	=over 4	2399	=over 4
1368		2400
1369	=item ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)	2401	=item ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)
1370		2402
…		…
1374	C<path>. The C<interval> is a hint on how quickly a change is expected to	2406	C<path>. The C<interval> is a hint on how quickly a change is expected to
1375	be detected and should normally be specified as C<0> to let libev choose	2407	be detected and should normally be specified as C<0> to let libev choose
1376	a suitable value. The memory pointed to by C<path> must point to the same	2408	a suitable value. The memory pointed to by C<path> must point to the same
1377	path for as long as the watcher is active.	2409	path for as long as the watcher is active.
1378		2410
1379	The callback will be receive C<EV_STAT> when a change was detected,	2411	The callback will receive an C<EV_STAT> event when a change was detected,
1380	relative to the attributes at the time the watcher was started (or the	2412	relative to the attributes at the time the watcher was started (or the
1381	last change was detected).	2413	last change was detected).
1382		2414
1383	=item ev_stat_stat (ev_stat *)	2415	=item ev_stat_stat (loop, ev_stat *)
1384		2416
1385	Updates the stat buffer immediately with new values. If you change the	2417	Updates the stat buffer immediately with new values. If you change the
1386	watched path in your callback, you could call this fucntion to avoid	2418	watched path in your callback, you could call this function to avoid
1387	detecting this change (while introducing a race condition). Can also be	2419	detecting this change (while introducing a race condition if you are not
1388	useful simply to find out the new values.	2420	the only one changing the path). Can also be useful simply to find out the
		2421	new values.
1389		2422
1390	=item ev_statdata attr [read-only]	2423	=item ev_statdata attr [read-only]
1391		2424
1392	The most-recently detected attributes of the file. Although the type is of	2425	The most-recently detected attributes of the file. Although the type is
1393	C<ev_statdata>, this is usually the (or one of the) C<struct stat> types	2426	C<ev_statdata>, this is usually the (or one of the) C<struct stat> types
1394	suitable for your system. If the C<st_nlink> member is C<0>, then there	2427	suitable for your system, but you can only rely on the POSIX-standardised
		2428	members to be present. If the C<st_nlink> member is C<0>, then there was
1395	was some error while C<stat>ing the file.	2429	some error while C<stat>ing the file.
1396		2430
1397	=item ev_statdata prev [read-only]	2431	=item ev_statdata prev [read-only]
1398		2432
1399	The previous attributes of the file. The callback gets invoked whenever	2433	The previous attributes of the file. The callback gets invoked whenever
1400	C<prev> != C<attr>.	2434	C<prev> != C<attr>, or, more precisely, one or more of these members
		2435	differ: C<st_dev>, C<st_ino>, C<st_mode>, C<st_nlink>, C<st_uid>,
		2436	C<st_gid>, C<st_rdev>, C<st_size>, C<st_atime>, C<st_mtime>, C<st_ctime>.
1401		2437
1402	=item ev_tstamp interval [read-only]	2438	=item ev_tstamp interval [read-only]
1403		2439
1404	The specified interval.	2440	The specified interval.
1405		2441
1406	=item const char *path [read-only]	2442	=item const char *path [read-only]
1407		2443
1408	The filesystem path that is being watched.	2444	The file system path that is being watched.
1409		2445
1410	=back	2446	=back
1411		2447
		2448	=head3 Examples
		2449
1412	Example: Watch C</etc/passwd> for attribute changes.	2450	Example: Watch C</etc/passwd> for attribute changes.
1413		2451
1414	static void	2452	static void
1415	passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)	2453	passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)
1416	{	2454	{
1417	/* /etc/passwd changed in some way */	2455	/* /etc/passwd changed in some way */
1418	if (w->attr.st_nlink)	2456	if (w->attr.st_nlink)
1419	{	2457	{
1420	printf ("passwd current size %ld\n", (long)w->attr.st_size);	2458	printf ("passwd current size %ld\n", (long)w->attr.st_size);
1421	printf ("passwd current atime %ld\n", (long)w->attr.st_mtime);	2459	printf ("passwd current atime %ld\n", (long)w->attr.st_mtime);
1422	printf ("passwd current mtime %ld\n", (long)w->attr.st_mtime);	2460	printf ("passwd current mtime %ld\n", (long)w->attr.st_mtime);
1423	}	2461	}
1424	else	2462	else
1425	/* you shalt not abuse printf for puts */	2463	/* you shalt not abuse printf for puts */
1426	puts ("wow, /etc/passwd is not there, expect problems. "	2464	puts ("wow, /etc/passwd is not there, expect problems. "
1427	"if this is windows, they already arrived\n");	2465	"if this is windows, they already arrived\n");
1428	}	2466	}
1429		2467
1430	...	2468	...
1431	ev_stat passwd;	2469	ev_stat passwd;
1432		2470
1433	ev_stat_init (&passwd, passwd_cb, "/etc/passwd");	2471	ev_stat_init (&passwd, passwd_cb, "/etc/passwd", 0.);
1434	ev_stat_start (loop, &passwd);	2472	ev_stat_start (loop, &passwd);
		2473
		2474	Example: Like above, but additionally use a one-second delay so we do not
		2475	miss updates (however, frequent updates will delay processing, too, so
		2476	one might do the work both on C<ev_stat> callback invocation I<and> on
		2477	C<ev_timer> callback invocation).
		2478
		2479	static ev_stat passwd;
		2480	static ev_timer timer;
		2481
		2482	static void
		2483	timer_cb (EV_P_ ev_timer *w, int revents)
		2484	{
		2485	ev_timer_stop (EV_A_ w);
		2486
		2487	/* now it's one second after the most recent passwd change */
		2488	}
		2489
		2490	static void
		2491	stat_cb (EV_P_ ev_stat *w, int revents)
		2492	{
		2493	/* reset the one-second timer */
		2494	ev_timer_again (EV_A_ &timer);
		2495	}
		2496
		2497	...
		2498	ev_stat_init (&passwd, stat_cb, "/etc/passwd", 0.);
		2499	ev_stat_start (loop, &passwd);
		2500	ev_timer_init (&timer, timer_cb, 0., 1.02);
1435		2501
1436		2502
1437	=head2 C<ev_idle> - when you've got nothing better to do...	2503	=head2 C<ev_idle> - when you've got nothing better to do...
1438		2504
1439	Idle watchers trigger events when no other events of the same or higher	2505	Idle watchers trigger events when no other events of the same or higher
1440	priority are pending (prepare, check and other idle watchers do not	2506	priority are pending (prepare, check and other idle watchers do not count
1441	count).	2507	as receiving "events").
1442		2508
1443	That is, as long as your process is busy handling sockets or timeouts	2509	That is, as long as your process is busy handling sockets or timeouts
1444	(or even signals, imagine) of the same or higher priority it will not be	2510	(or even signals, imagine) of the same or higher priority it will not be
1445	triggered. But when your process is idle (or only lower-priority watchers	2511	triggered. But when your process is idle (or only lower-priority watchers
1446	are pending), the idle watchers are being called once per event loop	2512	are pending), the idle watchers are being called once per event loop
…		…
1453	Apart from keeping your process non-blocking (which is a useful	2519	Apart from keeping your process non-blocking (which is a useful
1454	effect on its own sometimes), idle watchers are a good place to do	2520	effect on its own sometimes), idle watchers are a good place to do
1455	"pseudo-background processing", or delay processing stuff to after the	2521	"pseudo-background processing", or delay processing stuff to after the
1456	event loop has handled all outstanding events.	2522	event loop has handled all outstanding events.
1457		2523
		2524	=head3 Watcher-Specific Functions and Data Members
		2525
1458	=over 4	2526	=over 4
1459		2527
1460	=item ev_idle_init (ev_signal *, callback)	2528	=item ev_idle_init (ev_idle *, callback)
1461		2529
1462	Initialises and configures the idle watcher - it has no parameters of any	2530	Initialises and configures the idle watcher - it has no parameters of any
1463	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,	2531	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,
1464	believe me.	2532	believe me.
1465		2533
1466	=back	2534	=back
1467		2535
		2536	=head3 Examples
		2537
1468	Example: Dynamically allocate an C<ev_idle> watcher, start it, and in the	2538	Example: Dynamically allocate an C<ev_idle> watcher, start it, and in the
1469	callback, free it. Also, use no error checking, as usual.	2539	callback, free it. Also, use no error checking, as usual.
1470		2540
1471	static void	2541	static void
1472	idle_cb (struct ev_loop *loop, struct ev_idle *w, int revents)	2542	idle_cb (struct ev_loop *loop, ev_idle *w, int revents)
1473	{	2543	{
1474	free (w);	2544	free (w);
1475	// now do something you wanted to do when the program has	2545	// now do something you wanted to do when the program has
1476	// no longer asnything immediate to do.	2546	// no longer anything immediate to do.
1477	}	2547	}
1478		2548
1479	struct ev_idle *idle_watcher = malloc (sizeof (struct ev_idle));	2549	ev_idle *idle_watcher = malloc (sizeof (ev_idle));
1480	ev_idle_init (idle_watcher, idle_cb);	2550	ev_idle_init (idle_watcher, idle_cb);
1481	ev_idle_start (loop, idle_cb);	2551	ev_idle_start (loop, idle_watcher);
1482		2552
1483		2553
1484	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!	2554	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!
1485		2555
1486	Prepare and check watchers are usually (but not always) used in tandem:	2556	Prepare and check watchers are usually (but not always) used in pairs:
1487	prepare watchers get invoked before the process blocks and check watchers	2557	prepare watchers get invoked before the process blocks and check watchers
1488	afterwards.	2558	afterwards.
1489		2559
1490	You I<must not> call C<ev_loop> or similar functions that enter	2560	You I<must not> call C<ev_loop> or similar functions that enter
1491	the current event loop from either C<ev_prepare> or C<ev_check>	2561	the current event loop from either C<ev_prepare> or C<ev_check>
…		…
1494	those watchers, i.e. the sequence will always be C<ev_prepare>, blocking,	2564	those watchers, i.e. the sequence will always be C<ev_prepare>, blocking,
1495	C<ev_check> so if you have one watcher of each kind they will always be	2565	C<ev_check> so if you have one watcher of each kind they will always be
1496	called in pairs bracketing the blocking call.	2566	called in pairs bracketing the blocking call.
1497		2567
1498	Their main purpose is to integrate other event mechanisms into libev and	2568	Their main purpose is to integrate other event mechanisms into libev and
1499	their use is somewhat advanced. This could be used, for example, to track	2569	their use is somewhat advanced. They could be used, for example, to track
1500	variable changes, implement your own watchers, integrate net-snmp or a	2570	variable changes, implement your own watchers, integrate net-snmp or a
1501	coroutine library and lots more. They are also occasionally useful if	2571	coroutine library and lots more. They are also occasionally useful if
1502	you cache some data and want to flush it before blocking (for example,	2572	you cache some data and want to flush it before blocking (for example,
1503	in X programs you might want to do an C<XFlush ()> in an C<ev_prepare>	2573	in X programs you might want to do an C<XFlush ()> in an C<ev_prepare>
1504	watcher).	2574	watcher).
1505		2575
1506	This is done by examining in each prepare call which file descriptors need	2576	This is done by examining in each prepare call which file descriptors
1507	to be watched by the other library, registering C<ev_io> watchers for	2577	need to be watched by the other library, registering C<ev_io> watchers
1508	them and starting an C<ev_timer> watcher for any timeouts (many libraries	2578	for them and starting an C<ev_timer> watcher for any timeouts (many
1509	provide just this functionality). Then, in the check watcher you check for	2579	libraries provide exactly this functionality). Then, in the check watcher,
1510	any events that occured (by checking the pending status of all watchers	2580	you check for any events that occurred (by checking the pending status
1511	and stopping them) and call back into the library. The I/O and timer	2581	of all watchers and stopping them) and call back into the library. The
1512	callbacks will never actually be called (but must be valid nevertheless,	2582	I/O and timer callbacks will never actually be called (but must be valid
1513	because you never know, you know?).	2583	nevertheless, because you never know, you know?).
1514		2584
1515	As another example, the Perl Coro module uses these hooks to integrate	2585	As another example, the Perl Coro module uses these hooks to integrate
1516	coroutines into libev programs, by yielding to other active coroutines	2586	coroutines into libev programs, by yielding to other active coroutines
1517	during each prepare and only letting the process block if no coroutines	2587	during each prepare and only letting the process block if no coroutines
1518	are ready to run (it's actually more complicated: it only runs coroutines	2588	are ready to run (it's actually more complicated: it only runs coroutines
…		…
1521	loop from blocking if lower-priority coroutines are active, thus mapping	2591	loop from blocking if lower-priority coroutines are active, thus mapping
1522	low-priority coroutines to idle/background tasks).	2592	low-priority coroutines to idle/background tasks).
1523		2593
1524	It is recommended to give C<ev_check> watchers highest (C<EV_MAXPRI>)	2594	It is recommended to give C<ev_check> watchers highest (C<EV_MAXPRI>)
1525	priority, to ensure that they are being run before any other watchers	2595	priority, to ensure that they are being run before any other watchers
		2596	after the poll (this doesn't matter for C<ev_prepare> watchers).
		2597
1526	after the poll. Also, C<ev_check> watchers (and C<ev_prepare> watchers,	2598	Also, C<ev_check> watchers (and C<ev_prepare> watchers, too) should not
1527	too) should not activate ("feed") events into libev. While libev fully	2599	activate ("feed") events into libev. While libev fully supports this, they
1528	supports this, they will be called before other C<ev_check> watchers did	2600	might get executed before other C<ev_check> watchers did their job. As
1529	their job. As C<ev_check> watchers are often used to embed other event	2601	C<ev_check> watchers are often used to embed other (non-libev) event
1530	loops those other event loops might be in an unusable state until their	2602	loops those other event loops might be in an unusable state until their
1531	C<ev_check> watcher ran (always remind yourself to coexist peacefully with	2603	C<ev_check> watcher ran (always remind yourself to coexist peacefully with
1532	others).	2604	others).
1533		2605
		2606	=head3 Watcher-Specific Functions and Data Members
		2607
1534	=over 4	2608	=over 4
1535		2609
1536	=item ev_prepare_init (ev_prepare *, callback)	2610	=item ev_prepare_init (ev_prepare *, callback)
1537		2611
1538	=item ev_check_init (ev_check *, callback)	2612	=item ev_check_init (ev_check *, callback)
1539		2613
1540	Initialises and configures the prepare or check watcher - they have no	2614	Initialises and configures the prepare or check watcher - they have no
1541	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>	2615	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>
1542	macros, but using them is utterly, utterly and completely pointless.	2616	macros, but using them is utterly, utterly, utterly and completely
		2617	pointless.
1543		2618
1544	=back	2619	=back
		2620
		2621	=head3 Examples
1545		2622
1546	There are a number of principal ways to embed other event loops or modules	2623	There are a number of principal ways to embed other event loops or modules
1547	into libev. Here are some ideas on how to include libadns into libev	2624	into libev. Here are some ideas on how to include libadns into libev
1548	(there is a Perl module named C<EV::ADNS> that does this, which you could	2625	(there is a Perl module named C<EV::ADNS> that does this, which you could
1549	use for an actually working example. Another Perl module named C<EV::Glib>	2626	use as a working example. Another Perl module named C<EV::Glib> embeds a
1550	embeds a Glib main context into libev, and finally, C<Glib::EV> embeds EV	2627	Glib main context into libev, and finally, C<Glib::EV> embeds EV into the
1551	into the Glib event loop).	2628	Glib event loop).
1552		2629
1553	Method 1: Add IO watchers and a timeout watcher in a prepare handler,	2630	Method 1: Add IO watchers and a timeout watcher in a prepare handler,
1554	and in a check watcher, destroy them and call into libadns. What follows	2631	and in a check watcher, destroy them and call into libadns. What follows
1555	is pseudo-code only of course. This requires you to either use a low	2632	is pseudo-code only of course. This requires you to either use a low
1556	priority for the check watcher or use C<ev_clear_pending> explicitly, as	2633	priority for the check watcher or use C<ev_clear_pending> explicitly, as
1557	the callbacks for the IO/timeout watchers might not have been called yet.	2634	the callbacks for the IO/timeout watchers might not have been called yet.
1558		2635
1559	static ev_io iow [nfd];	2636	static ev_io iow [nfd];
1560	static ev_timer tw;	2637	static ev_timer tw;
1561		2638
1562	static void	2639	static void
1563	io_cb (ev_loop *loop, ev_io *w, int revents)	2640	io_cb (struct ev_loop *loop, ev_io *w, int revents)
1564	{	2641	{
1565	}	2642	}
1566		2643
1567	// create io watchers for each fd and a timer before blocking	2644	// create io watchers for each fd and a timer before blocking
1568	static void	2645	static void
1569	adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents)	2646	adns_prepare_cb (struct ev_loop *loop, ev_prepare *w, int revents)
1570	{	2647	{
1571	int timeout = 3600000;	2648	int timeout = 3600000;
1572	struct pollfd fds [nfd];	2649	struct pollfd fds [nfd];
1573	// actual code will need to loop here and realloc etc.	2650	// actual code will need to loop here and realloc etc.
1574	adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));	2651	adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));
1575		2652
1576	/* the callback is illegal, but won't be called as we stop during check */	2653	/* the callback is illegal, but won't be called as we stop during check */
1577	ev_timer_init (&tw, 0, timeout * 1e-3);	2654	ev_timer_init (&tw, 0, timeout * 1e-3, 0.);
1578	ev_timer_start (loop, &tw);	2655	ev_timer_start (loop, &tw);
1579		2656
1580	// create one ev_io per pollfd	2657	// create one ev_io per pollfd
1581	for (int i = 0; i < nfd; ++i)	2658	for (int i = 0; i < nfd; ++i)
1582	{	2659	{
1583	ev_io_init (iow + i, io_cb, fds [i].fd,	2660	ev_io_init (iow + i, io_cb, fds [i].fd,
1584	((fds [i].events & POLLIN ? EV_READ : 0)	2661	((fds [i].events & POLLIN ? EV_READ : 0)
1585	| (fds [i].events & POLLOUT ? EV_WRITE : 0)));	2662	| (fds [i].events & POLLOUT ? EV_WRITE : 0)));
1586		2663
1587	fds [i].revents = 0;	2664	fds [i].revents = 0;
1588	ev_io_start (loop, iow + i);	2665	ev_io_start (loop, iow + i);
1589	}	2666	}
1590	}	2667	}
1591		2668
1592	// stop all watchers after blocking	2669	// stop all watchers after blocking
1593	static void	2670	static void
1594	adns_check_cb (ev_loop *loop, ev_check *w, int revents)	2671	adns_check_cb (struct ev_loop *loop, ev_check *w, int revents)
1595	{	2672	{
1596	ev_timer_stop (loop, &tw);	2673	ev_timer_stop (loop, &tw);
1597		2674
1598	for (int i = 0; i < nfd; ++i)	2675	for (int i = 0; i < nfd; ++i)
1599	{	2676	{
1600	// set the relevant poll flags	2677	// set the relevant poll flags
1601	// could also call adns_processreadable etc. here	2678	// could also call adns_processreadable etc. here
1602	struct pollfd *fd = fds + i;	2679	struct pollfd *fd = fds + i;
1603	int revents = ev_clear_pending (iow + i);	2680	int revents = ev_clear_pending (iow + i);
1604	if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;	2681	if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;
1605	if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;	2682	if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;
1606		2683
1607	// now stop the watcher	2684	// now stop the watcher
1608	ev_io_stop (loop, iow + i);	2685	ev_io_stop (loop, iow + i);
1609	}	2686	}
1610		2687
1611	adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));	2688	adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));
1612	}	2689	}
1613		2690
1614	Method 2: This would be just like method 1, but you run C<adns_afterpoll>	2691	Method 2: This would be just like method 1, but you run C<adns_afterpoll>
1615	in the prepare watcher and would dispose of the check watcher.	2692	in the prepare watcher and would dispose of the check watcher.
1616		2693
1617	Method 3: If the module to be embedded supports explicit event	2694	Method 3: If the module to be embedded supports explicit event
1618	notification (adns does), you can also make use of the actual watcher	2695	notification (libadns does), you can also make use of the actual watcher
1619	callbacks, and only destroy/create the watchers in the prepare watcher.	2696	callbacks, and only destroy/create the watchers in the prepare watcher.
1620		2697
1621	static void	2698	static void
1622	timer_cb (EV_P_ ev_timer *w, int revents)	2699	timer_cb (EV_P_ ev_timer *w, int revents)
1623	{	2700	{
1624	adns_state ads = (adns_state)w->data;	2701	adns_state ads = (adns_state)w->data;
1625	update_now (EV_A);	2702	update_now (EV_A);
1626		2703
1627	adns_processtimeouts (ads, &tv_now);	2704	adns_processtimeouts (ads, &tv_now);
1628	}	2705	}
1629		2706
1630	static void	2707	static void
1631	io_cb (EV_P_ ev_io *w, int revents)	2708	io_cb (EV_P_ ev_io *w, int revents)
1632	{	2709	{
1633	adns_state ads = (adns_state)w->data;	2710	adns_state ads = (adns_state)w->data;
1634	update_now (EV_A);	2711	update_now (EV_A);
1635		2712
1636	if (revents & EV_READ ) adns_processreadable (ads, w->fd, &tv_now);	2713	if (revents & EV_READ ) adns_processreadable (ads, w->fd, &tv_now);
1637	if (revents & EV_WRITE) adns_processwriteable (ads, w->fd, &tv_now);	2714	if (revents & EV_WRITE) adns_processwriteable (ads, w->fd, &tv_now);
1638	}	2715	}
1639		2716
1640	// do not ever call adns_afterpoll	2717	// do not ever call adns_afterpoll
1641		2718
1642	Method 4: Do not use a prepare or check watcher because the module you	2719	Method 4: Do not use a prepare or check watcher because the module you
1643	want to embed is too inflexible to support it. Instead, youc na override	2720	want to embed is not flexible enough to support it. Instead, you can
1644	their poll function. The drawback with this solution is that the main	2721	override their poll function. The drawback with this solution is that the
1645	loop is now no longer controllable by EV. The C<Glib::EV> module does	2722	main loop is now no longer controllable by EV. The C<Glib::EV> module uses
1646	this.	2723	this approach, effectively embedding EV as a client into the horrible
		2724	libglib event loop.
1647		2725
1648	static gint	2726	static gint
1649	event_poll_func (GPollFD *fds, guint nfds, gint timeout)	2727	event_poll_func (GPollFD *fds, guint nfds, gint timeout)
1650	{	2728	{
1651	int got_events = 0;	2729	int got_events = 0;
1652		2730
1653	for (n = 0; n < nfds; ++n)	2731	for (n = 0; n < nfds; ++n)
1654	// create/start io watcher that sets the relevant bits in fds[n] and increment got_events	2732	// create/start io watcher that sets the relevant bits in fds[n] and increment got_events
1655		2733
1656	if (timeout >= 0)	2734	if (timeout >= 0)
1657	// create/start timer	2735	// create/start timer
1658		2736
1659	// poll	2737	// poll
1660	ev_loop (EV_A_ 0);	2738	ev_loop (EV_A_ 0);
1661		2739
1662	// stop timer again	2740	// stop timer again
1663	if (timeout >= 0)	2741	if (timeout >= 0)
1664	ev_timer_stop (EV_A_ &to);	2742	ev_timer_stop (EV_A_ &to);
1665		2743
1666	// stop io watchers again - their callbacks should have set	2744	// stop io watchers again - their callbacks should have set
1667	for (n = 0; n < nfds; ++n)	2745	for (n = 0; n < nfds; ++n)
1668	ev_io_stop (EV_A_ iow [n]);	2746	ev_io_stop (EV_A_ iow [n]);
1669		2747
1670	return got_events;	2748	return got_events;
1671	}	2749	}
1672		2750
1673		2751
1674	=head2 C<ev_embed> - when one backend isn't enough...	2752	=head2 C<ev_embed> - when one backend isn't enough...
1675		2753
1676	This is a rather advanced watcher type that lets you embed one event loop	2754	This is a rather advanced watcher type that lets you embed one event loop
…		…
1682	prioritise I/O.	2760	prioritise I/O.
1683		2761
1684	As an example for a bug workaround, the kqueue backend might only support	2762	As an example for a bug workaround, the kqueue backend might only support
1685	sockets on some platform, so it is unusable as generic backend, but you	2763	sockets on some platform, so it is unusable as generic backend, but you
1686	still want to make use of it because you have many sockets and it scales	2764	still want to make use of it because you have many sockets and it scales
1687	so nicely. In this case, you would create a kqueue-based loop and embed it	2765	so nicely. In this case, you would create a kqueue-based loop and embed
1688	into your default loop (which might use e.g. poll). Overall operation will	2766	it into your default loop (which might use e.g. poll). Overall operation
1689	be a bit slower because first libev has to poll and then call kevent, but	2767	will be a bit slower because first libev has to call C<poll> and then
1690	at least you can use both at what they are best.	2768	C<kevent>, but at least you can use both mechanisms for what they are
		2769	best: C<kqueue> for scalable sockets and C<poll> if you want it to work :)
1691		2770
1692	As for prioritising I/O: rarely you have the case where some fds have	2771	As for prioritising I/O: under rare circumstances you have the case where
1693	to be watched and handled very quickly (with low latency), and even	2772	some fds have to be watched and handled very quickly (with low latency),
1694	priorities and idle watchers might have too much overhead. In this case	2773	and even priorities and idle watchers might have too much overhead. In
1695	you would put all the high priority stuff in one loop and all the rest in	2774	this case you would put all the high priority stuff in one loop and all
1696	a second one, and embed the second one in the first.	2775	the rest in a second one, and embed the second one in the first.
1697		2776
1698	As long as the watcher is active, the callback will be invoked every time	2777	As long as the watcher is active, the callback will be invoked every
1699	there might be events pending in the embedded loop. The callback must then	2778	time there might be events pending in the embedded loop. The callback
1700	call C<ev_embed_sweep (mainloop, watcher)> to make a single sweep and invoke	2779	must then call C<ev_embed_sweep (mainloop, watcher)> to make a single
1701	their callbacks (you could also start an idle watcher to give the embedded	2780	sweep and invoke their callbacks (the callback doesn't need to invoke the
1702	loop strictly lower priority for example). You can also set the callback	2781	C<ev_embed_sweep> function directly, it could also start an idle watcher
1703	to C<0>, in which case the embed watcher will automatically execute the	2782	to give the embedded loop strictly lower priority for example).
1704	embedded loop sweep.
1705		2783
1706	As long as the watcher is started it will automatically handle events. The	2784	You can also set the callback to C<0>, in which case the embed watcher
1707	callback will be invoked whenever some events have been handled. You can	2785	will automatically execute the embedded loop sweep whenever necessary.
1708	set the callback to C<0> to avoid having to specify one if you are not
1709	interested in that.
1710		2786
1711	Also, there have not currently been made special provisions for forking:	2787	Fork detection will be handled transparently while the C<ev_embed> watcher
1712	when you fork, you not only have to call C<ev_loop_fork> on both loops,	2788	is active, i.e., the embedded loop will automatically be forked when the
1713	but you will also have to stop and restart any C<ev_embed> watchers	2789	embedding loop forks. In other cases, the user is responsible for calling
1714	yourself.	2790	C<ev_loop_fork> on the embedded loop.
1715		2791
1716	Unfortunately, not all backends are embeddable, only the ones returned by	2792	Unfortunately, not all backends are embeddable: only the ones returned by
1717	C<ev_embeddable_backends> are, which, unfortunately, does not include any	2793	C<ev_embeddable_backends> are, which, unfortunately, does not include any
1718	portable one.	2794	portable one.
1719		2795
1720	So when you want to use this feature you will always have to be prepared	2796	So when you want to use this feature you will always have to be prepared
1721	that you cannot get an embeddable loop. The recommended way to get around	2797	that you cannot get an embeddable loop. The recommended way to get around
1722	this is to have a separate variables for your embeddable loop, try to	2798	this is to have a separate variables for your embeddable loop, try to
1723	create it, and if that fails, use the normal loop for everything:	2799	create it, and if that fails, use the normal loop for everything.
1724		2800
1725	struct ev_loop *loop_hi = ev_default_init (0);	2801	=head3 C<ev_embed> and fork
1726	struct ev_loop *loop_lo = 0;
1727	struct ev_embed embed;
1728
1729	// see if there is a chance of getting one that works
1730	// (remember that a flags value of 0 means autodetection)
1731	loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
1732	? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
1733	: 0;
1734		2802
1735	// if we got one, then embed it, otherwise default to loop_hi	2803	While the C<ev_embed> watcher is running, forks in the embedding loop will
1736	if (loop_lo)	2804	automatically be applied to the embedded loop as well, so no special
1737	{	2805	fork handling is required in that case. When the watcher is not running,
1738	ev_embed_init (&embed, 0, loop_lo);	2806	however, it is still the task of the libev user to call C<ev_loop_fork ()>
1739	ev_embed_start (loop_hi, &embed);	2807	as applicable.
1740	}	2808
1741	else	2809	=head3 Watcher-Specific Functions and Data Members
1742	loop_lo = loop_hi;
1743		2810
1744	=over 4	2811	=over 4
1745		2812
1746	=item ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)	2813	=item ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)
1747		2814
…		…
1749		2816
1750	Configures the watcher to embed the given loop, which must be	2817	Configures the watcher to embed the given loop, which must be
1751	embeddable. If the callback is C<0>, then C<ev_embed_sweep> will be	2818	embeddable. If the callback is C<0>, then C<ev_embed_sweep> will be
1752	invoked automatically, otherwise it is the responsibility of the callback	2819	invoked automatically, otherwise it is the responsibility of the callback
1753	to invoke it (it will continue to be called until the sweep has been done,	2820	to invoke it (it will continue to be called until the sweep has been done,
1754	if you do not want thta, you need to temporarily stop the embed watcher).	2821	if you do not want that, you need to temporarily stop the embed watcher).
1755		2822
1756	=item ev_embed_sweep (loop, ev_embed *)	2823	=item ev_embed_sweep (loop, ev_embed *)
1757		2824
1758	Make a single, non-blocking sweep over the embedded loop. This works	2825	Make a single, non-blocking sweep over the embedded loop. This works
1759	similarly to C<ev_loop (embedded_loop, EVLOOP_NONBLOCK)>, but in the most	2826	similarly to C<ev_loop (embedded_loop, EVLOOP_NONBLOCK)>, but in the most
1760	apropriate way for embedded loops.	2827	appropriate way for embedded loops.
1761		2828
1762	=item struct ev_loop *loop [read-only]	2829	=item struct ev_loop *other [read-only]
1763		2830
1764	The embedded event loop.	2831	The embedded event loop.
1765		2832
1766	=back	2833	=back
		2834
		2835	=head3 Examples
		2836
		2837	Example: Try to get an embeddable event loop and embed it into the default
		2838	event loop. If that is not possible, use the default loop. The default
		2839	loop is stored in C<loop_hi>, while the embeddable loop is stored in
		2840	C<loop_lo> (which is C<loop_hi> in the case no embeddable loop can be
		2841	used).
		2842
		2843	struct ev_loop *loop_hi = ev_default_init (0);
		2844	struct ev_loop *loop_lo = 0;
		2845	ev_embed embed;
		2846
		2847	// see if there is a chance of getting one that works
		2848	// (remember that a flags value of 0 means autodetection)
		2849	loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
		2850	? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
		2851	: 0;
		2852
		2853	// if we got one, then embed it, otherwise default to loop_hi
		2854	if (loop_lo)
		2855	{
		2856	ev_embed_init (&embed, 0, loop_lo);
		2857	ev_embed_start (loop_hi, &embed);
		2858	}
		2859	else
		2860	loop_lo = loop_hi;
		2861
		2862	Example: Check if kqueue is available but not recommended and create
		2863	a kqueue backend for use with sockets (which usually work with any
		2864	kqueue implementation). Store the kqueue/socket-only event loop in
		2865	C<loop_socket>. (One might optionally use C<EVFLAG_NOENV>, too).
		2866
		2867	struct ev_loop *loop = ev_default_init (0);
		2868	struct ev_loop *loop_socket = 0;
		2869	ev_embed embed;
		2870
		2871	if (ev_supported_backends () & ~ev_recommended_backends () & EVBACKEND_KQUEUE)
		2872	if ((loop_socket = ev_loop_new (EVBACKEND_KQUEUE))
		2873	{
		2874	ev_embed_init (&embed, 0, loop_socket);
		2875	ev_embed_start (loop, &embed);
		2876	}
		2877
		2878	if (!loop_socket)
		2879	loop_socket = loop;
		2880
		2881	// now use loop_socket for all sockets, and loop for everything else
1767		2882
1768		2883
1769	=head2 C<ev_fork> - the audacity to resume the event loop after a fork	2884	=head2 C<ev_fork> - the audacity to resume the event loop after a fork
1770		2885
1771	Fork watchers are called when a C<fork ()> was detected (usually because	2886	Fork watchers are called when a C<fork ()> was detected (usually because
…		…
1774	event loop blocks next and before C<ev_check> watchers are being called,	2889	event loop blocks next and before C<ev_check> watchers are being called,
1775	and only in the child after the fork. If whoever good citizen calling	2890	and only in the child after the fork. If whoever good citizen calling
1776	C<ev_default_fork> cheats and calls it in the wrong process, the fork	2891	C<ev_default_fork> cheats and calls it in the wrong process, the fork
1777	handlers will be invoked, too, of course.	2892	handlers will be invoked, too, of course.
1778		2893
		2894	=head3 The special problem of life after fork - how is it possible?
		2895
		2896	Most uses of C<fork()> consist of forking, then some simple calls to ste
		2897	up/change the process environment, followed by a call to C<exec()>. This
		2898	sequence should be handled by libev without any problems.
		2899
		2900	This changes when the application actually wants to do event handling
		2901	in the child, or both parent in child, in effect "continuing" after the
		2902	fork.
		2903
		2904	The default mode of operation (for libev, with application help to detect
		2905	forks) is to duplicate all the state in the child, as would be expected
		2906	when I<either> the parent I<or> the child process continues.
		2907
		2908	When both processes want to continue using libev, then this is usually the
		2909	wrong result. In that case, usually one process (typically the parent) is
		2910	supposed to continue with all watchers in place as before, while the other
		2911	process typically wants to start fresh, i.e. without any active watchers.
		2912
		2913	The cleanest and most efficient way to achieve that with libev is to
		2914	simply create a new event loop, which of course will be "empty", and
		2915	use that for new watchers. This has the advantage of not touching more
		2916	memory than necessary, and thus avoiding the copy-on-write, and the
		2917	disadvantage of having to use multiple event loops (which do not support
		2918	signal watchers).
		2919
		2920	When this is not possible, or you want to use the default loop for
		2921	other reasons, then in the process that wants to start "fresh", call
		2922	C<ev_default_destroy ()> followed by C<ev_default_loop (...)>. Destroying
		2923	the default loop will "orphan" (not stop) all registered watchers, so you
		2924	have to be careful not to execute code that modifies those watchers. Note
		2925	also that in that case, you have to re-register any signal watchers.
		2926
		2927	=head3 Watcher-Specific Functions and Data Members
		2928
1779	=over 4	2929	=over 4
1780		2930
1781	=item ev_fork_init (ev_signal *, callback)	2931	=item ev_fork_init (ev_signal *, callback)
1782		2932
1783	Initialises and configures the fork watcher - it has no parameters of any	2933	Initialises and configures the fork watcher - it has no parameters of any
…		…
1785	believe me.	2935	believe me.
1786		2936
1787	=back	2937	=back
1788		2938
1789		2939
		2940	=head2 C<ev_async> - how to wake up another event loop
		2941
		2942	In general, you cannot use an C<ev_loop> from multiple threads or other
		2943	asynchronous sources such as signal handlers (as opposed to multiple event
		2944	loops - those are of course safe to use in different threads).
		2945
		2946	Sometimes, however, you need to wake up another event loop you do not
		2947	control, for example because it belongs to another thread. This is what
		2948	C<ev_async> watchers do: as long as the C<ev_async> watcher is active, you
		2949	can signal it by calling C<ev_async_send>, which is thread- and signal
		2950	safe.
		2951
		2952	This functionality is very similar to C<ev_signal> watchers, as signals,
		2953	too, are asynchronous in nature, and signals, too, will be compressed
		2954	(i.e. the number of callback invocations may be less than the number of
		2955	C<ev_async_sent> calls).
		2956
		2957	Unlike C<ev_signal> watchers, C<ev_async> works with any event loop, not
		2958	just the default loop.
		2959
		2960	=head3 Queueing
		2961
		2962	C<ev_async> does not support queueing of data in any way. The reason
		2963	is that the author does not know of a simple (or any) algorithm for a
		2964	multiple-writer-single-reader queue that works in all cases and doesn't
		2965	need elaborate support such as pthreads.
		2966
		2967	That means that if you want to queue data, you have to provide your own
		2968	queue. But at least I can tell you how to implement locking around your
		2969	queue:
		2970
		2971	=over 4
		2972
		2973	=item queueing from a signal handler context
		2974
		2975	To implement race-free queueing, you simply add to the queue in the signal
		2976	handler but you block the signal handler in the watcher callback. Here is
		2977	an example that does that for some fictitious SIGUSR1 handler:
		2978
		2979	static ev_async mysig;
		2980
		2981	static void
		2982	sigusr1_handler (void)
		2983	{
		2984	sometype data;
		2985
		2986	// no locking etc.
		2987	queue_put (data);
		2988	ev_async_send (EV_DEFAULT_ &mysig);
		2989	}
		2990
		2991	static void
		2992	mysig_cb (EV_P_ ev_async *w, int revents)
		2993	{
		2994	sometype data;
		2995	sigset_t block, prev;
		2996
		2997	sigemptyset (&block);
		2998	sigaddset (&block, SIGUSR1);
		2999	sigprocmask (SIG_BLOCK, &block, &prev);
		3000
		3001	while (queue_get (&data))
		3002	process (data);
		3003
		3004	if (sigismember (&prev, SIGUSR1)
		3005	sigprocmask (SIG_UNBLOCK, &block, 0);
		3006	}
		3007
		3008	(Note: pthreads in theory requires you to use C<pthread_setmask>
		3009	instead of C<sigprocmask> when you use threads, but libev doesn't do it
		3010	either...).
		3011
		3012	=item queueing from a thread context
		3013
		3014	The strategy for threads is different, as you cannot (easily) block
		3015	threads but you can easily preempt them, so to queue safely you need to
		3016	employ a traditional mutex lock, such as in this pthread example:
		3017
		3018	static ev_async mysig;
		3019	static pthread_mutex_t mymutex = PTHREAD_MUTEX_INITIALIZER;
		3020
		3021	static void
		3022	otherthread (void)
		3023	{
		3024	// only need to lock the actual queueing operation
		3025	pthread_mutex_lock (&mymutex);
		3026	queue_put (data);
		3027	pthread_mutex_unlock (&mymutex);
		3028
		3029	ev_async_send (EV_DEFAULT_ &mysig);
		3030	}
		3031
		3032	static void
		3033	mysig_cb (EV_P_ ev_async *w, int revents)
		3034	{
		3035	pthread_mutex_lock (&mymutex);
		3036
		3037	while (queue_get (&data))
		3038	process (data);
		3039
		3040	pthread_mutex_unlock (&mymutex);
		3041	}
		3042
		3043	=back
		3044
		3045
		3046	=head3 Watcher-Specific Functions and Data Members
		3047
		3048	=over 4
		3049
		3050	=item ev_async_init (ev_async *, callback)
		3051
		3052	Initialises and configures the async watcher - it has no parameters of any
		3053	kind. There is a C<ev_async_set> macro, but using it is utterly pointless,
		3054	trust me.
		3055
		3056	=item ev_async_send (loop, ev_async *)
		3057
		3058	Sends/signals/activates the given C<ev_async> watcher, that is, feeds
		3059	an C<EV_ASYNC> event on the watcher into the event loop. Unlike
		3060	C<ev_feed_event>, this call is safe to do from other threads, signal or
		3061	similar contexts (see the discussion of C<EV_ATOMIC_T> in the embedding
		3062	section below on what exactly this means).
		3063
		3064	Note that, as with other watchers in libev, multiple events might get
		3065	compressed into a single callback invocation (another way to look at this
		3066	is that C<ev_async> watchers are level-triggered, set on C<ev_async_send>,
		3067	reset when the event loop detects that).
		3068
		3069	This call incurs the overhead of a system call only once per event loop
		3070	iteration, so while the overhead might be noticeable, it doesn't apply to
		3071	repeated calls to C<ev_async_send> for the same event loop.
		3072
		3073	=item bool = ev_async_pending (ev_async *)
		3074
		3075	Returns a non-zero value when C<ev_async_send> has been called on the
		3076	watcher but the event has not yet been processed (or even noted) by the
		3077	event loop.
		3078
		3079	C<ev_async_send> sets a flag in the watcher and wakes up the loop. When
		3080	the loop iterates next and checks for the watcher to have become active,
		3081	it will reset the flag again. C<ev_async_pending> can be used to very
		3082	quickly check whether invoking the loop might be a good idea.
		3083
		3084	Not that this does I<not> check whether the watcher itself is pending,
		3085	only whether it has been requested to make this watcher pending: there
		3086	is a time window between the event loop checking and resetting the async
		3087	notification, and the callback being invoked.
		3088
		3089	=back
		3090
		3091
1790	=head1 OTHER FUNCTIONS	3092	=head1 OTHER FUNCTIONS
1791		3093
1792	There are some other functions of possible interest. Described. Here. Now.	3094	There are some other functions of possible interest. Described. Here. Now.
1793		3095
1794	=over 4	3096	=over 4
1795		3097
1796	=item ev_once (loop, int fd, int events, ev_tstamp timeout, callback)	3098	=item ev_once (loop, int fd, int events, ev_tstamp timeout, callback)
1797		3099
1798	This function combines a simple timer and an I/O watcher, calls your	3100	This function combines a simple timer and an I/O watcher, calls your
1799	callback on whichever event happens first and automatically stop both	3101	callback on whichever event happens first and automatically stops both
1800	watchers. This is useful if you want to wait for a single event on an fd	3102	watchers. This is useful if you want to wait for a single event on an fd
1801	or timeout without having to allocate/configure/start/stop/free one or	3103	or timeout without having to allocate/configure/start/stop/free one or
1802	more watchers yourself.	3104	more watchers yourself.
1803		3105
1804	If C<fd> is less than 0, then no I/O watcher will be started and events	3106	If C<fd> is less than 0, then no I/O watcher will be started and the
1805	is being ignored. Otherwise, an C<ev_io> watcher for the given C<fd> and	3107	C<events> argument is being ignored. Otherwise, an C<ev_io> watcher for
1806	C<events> set will be craeted and started.	3108	the given C<fd> and C<events> set will be created and started.
1807		3109
1808	If C<timeout> is less than 0, then no timeout watcher will be	3110	If C<timeout> is less than 0, then no timeout watcher will be
1809	started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and	3111	started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and
1810	repeat = 0) will be started. While C<0> is a valid timeout, it is of	3112	repeat = 0) will be started. C<0> is a valid timeout.
1811	dubious value.
1812		3113
1813	The callback has the type C<void (*cb)(int revents, void *arg)> and gets	3114	The callback has the type C<void (*cb)(int revents, void *arg)> and gets
1814	passed an C<revents> set like normal event callbacks (a combination of	3115	passed an C<revents> set like normal event callbacks (a combination of
1815	C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMEOUT>) and the C<arg>	3116	C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMEOUT>) and the C<arg>
1816	value passed to C<ev_once>:	3117	value passed to C<ev_once>. Note that it is possible to receive I<both>
		3118	a timeout and an io event at the same time - you probably should give io
		3119	events precedence.
1817		3120
		3121	Example: wait up to ten seconds for data to appear on STDIN_FILENO.
		3122
1818	static void stdin_ready (int revents, void *arg)	3123	static void stdin_ready (int revents, void *arg)
1819	{	3124	{
1820	if (revents & EV_TIMEOUT)
1821	/* doh, nothing entered */;
1822	else if (revents & EV_READ)	3125	if (revents & EV_READ)
1823	/* stdin might have data for us, joy! */;	3126	/* stdin might have data for us, joy! */;
		3127	else if (revents & EV_TIMEOUT)
		3128	/* doh, nothing entered */;
1824	}	3129	}
1825		3130
1826	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);	3131	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
1827		3132
1828	=item ev_feed_event (ev_loop *, watcher *, int revents)	3133	=item ev_feed_event (struct ev_loop *, watcher *, int revents)
1829		3134
1830	Feeds the given event set into the event loop, as if the specified event	3135	Feeds the given event set into the event loop, as if the specified event
1831	had happened for the specified watcher (which must be a pointer to an	3136	had happened for the specified watcher (which must be a pointer to an
1832	initialised but not necessarily started event watcher).	3137	initialised but not necessarily started event watcher).
1833		3138
1834	=item ev_feed_fd_event (ev_loop *, int fd, int revents)	3139	=item ev_feed_fd_event (struct ev_loop *, int fd, int revents)
1835		3140
1836	Feed an event on the given fd, as if a file descriptor backend detected	3141	Feed an event on the given fd, as if a file descriptor backend detected
1837	the given events it.	3142	the given events it.
1838		3143
1839	=item ev_feed_signal_event (ev_loop *loop, int signum)	3144	=item ev_feed_signal_event (struct ev_loop *loop, int signum)
1840		3145
1841	Feed an event as if the given signal occured (C<loop> must be the default	3146	Feed an event as if the given signal occurred (C<loop> must be the default
1842	loop!).	3147	loop!).
1843		3148
1844	=back	3149	=back
1845		3150
1846		3151
…		…
1862		3167
1863	=item * Priorities are not currently supported. Initialising priorities	3168	=item * Priorities are not currently supported. Initialising priorities
1864	will fail and all watchers will have the same priority, even though there	3169	will fail and all watchers will have the same priority, even though there
1865	is an ev_pri field.	3170	is an ev_pri field.
1866		3171
		3172	=item * In libevent, the last base created gets the signals, in libev, the
		3173	first base created (== the default loop) gets the signals.
		3174
1867	=item * Other members are not supported.	3175	=item * Other members are not supported.
1868		3176
1869	=item * The libev emulation is I<not> ABI compatible to libevent, you need	3177	=item * The libev emulation is I<not> ABI compatible to libevent, you need
1870	to use the libev header file and library.	3178	to use the libev header file and library.
1871		3179
1872	=back	3180	=back
1873		3181
1874	=head1 C++ SUPPORT	3182	=head1 C++ SUPPORT
1875		3183
1876	Libev comes with some simplistic wrapper classes for C++ that mainly allow	3184	Libev comes with some simplistic wrapper classes for C++ that mainly allow
1877	you to use some convinience methods to start/stop watchers and also change	3185	you to use some convenience methods to start/stop watchers and also change
1878	the callback model to a model using method callbacks on objects.	3186	the callback model to a model using method callbacks on objects.
1879		3187
1880	To use it,	3188	To use it,
1881		3189
1882	#include <ev++.h>	3190	#include <ev++.h>
1883		3191
1884	This automatically includes F<ev.h> and puts all of its definitions (many	3192	This automatically includes F<ev.h> and puts all of its definitions (many
1885	of them macros) into the global namespace. All C++ specific things are	3193	of them macros) into the global namespace. All C++ specific things are
1886	put into the C<ev> namespace. It should support all the same embedding	3194	put into the C<ev> namespace. It should support all the same embedding
1887	options as F<ev.h>, most notably C<EV_MULTIPLICITY>.	3195	options as F<ev.h>, most notably C<EV_MULTIPLICITY>.
…		…
1954	your compiler is good :), then the method will be fully inlined into the	3262	your compiler is good :), then the method will be fully inlined into the
1955	thunking function, making it as fast as a direct C callback.	3263	thunking function, making it as fast as a direct C callback.
1956		3264
1957	Example: simple class declaration and watcher initialisation	3265	Example: simple class declaration and watcher initialisation
1958		3266
1959	struct myclass	3267	struct myclass
1960	{	3268	{
1961	void io_cb (ev::io &w, int revents) { }	3269	void io_cb (ev::io &w, int revents) { }
1962	}	3270	}
1963		3271
1964	myclass obj;	3272	myclass obj;
1965	ev::io iow;	3273	ev::io iow;
1966	iow.set <myclass, &myclass::io_cb> (&obj);	3274	iow.set <myclass, &myclass::io_cb> (&obj);
		3275
		3276	=item w->set (object *)
		3277
		3278	This is an B<experimental> feature that might go away in a future version.
		3279
		3280	This is a variation of a method callback - leaving out the method to call
		3281	will default the method to C<operator ()>, which makes it possible to use
		3282	functor objects without having to manually specify the C<operator ()> all
		3283	the time. Incidentally, you can then also leave out the template argument
		3284	list.
		3285
		3286	The C<operator ()> method prototype must be C<void operator ()(watcher &w,
		3287	int revents)>.
		3288
		3289	See the method-C<set> above for more details.
		3290
		3291	Example: use a functor object as callback.
		3292
		3293	struct myfunctor
		3294	{
		3295	void operator() (ev::io &w, int revents)
		3296	{
		3297	...
		3298	}
		3299	}
		3300
		3301	myfunctor f;
		3302
		3303	ev::io w;
		3304	w.set (&f);
1967		3305
1968	=item w->set<function> (void *data = 0)	3306	=item w->set<function> (void *data = 0)
1969		3307
1970	Also sets a callback, but uses a static method or plain function as	3308	Also sets a callback, but uses a static method or plain function as
1971	callback. The optional C<data> argument will be stored in the watcher's	3309	callback. The optional C<data> argument will be stored in the watcher's
…		…
1973		3311
1974	The prototype of the C<function> must be C<void (*)(ev::TYPE &w, int)>.	3312	The prototype of the C<function> must be C<void (*)(ev::TYPE &w, int)>.
1975		3313
1976	See the method-C<set> above for more details.	3314	See the method-C<set> above for more details.
1977		3315
1978	Example:	3316	Example: Use a plain function as callback.
1979		3317
1980	static void io_cb (ev::io &w, int revents) { }	3318	static void io_cb (ev::io &w, int revents) { }
1981	iow.set <io_cb> ();	3319	iow.set <io_cb> ();
1982		3320
1983	=item w->set (struct ev_loop *)	3321	=item w->set (struct ev_loop *)
1984		3322
1985	Associates a different C<struct ev_loop> with this watcher. You can only	3323	Associates a different C<struct ev_loop> with this watcher. You can only
1986	do this when the watcher is inactive (and not pending either).	3324	do this when the watcher is inactive (and not pending either).
1987		3325
1988	=item w->set ([args])	3326	=item w->set ([arguments])
1989		3327
1990	Basically the same as C<ev_TYPE_set>, with the same args. Must be	3328	Basically the same as C<ev_TYPE_set>, with the same arguments. Must be
1991	called at least once. Unlike the C counterpart, an active watcher gets	3329	called at least once. Unlike the C counterpart, an active watcher gets
1992	automatically stopped and restarted when reconfiguring it with this	3330	automatically stopped and restarted when reconfiguring it with this
1993	method.	3331	method.
1994		3332
1995	=item w->start ()	3333	=item w->start ()
…		…
1999		3337
2000	=item w->stop ()	3338	=item w->stop ()
2001		3339
2002	Stops the watcher if it is active. Again, no C<loop> argument.	3340	Stops the watcher if it is active. Again, no C<loop> argument.
2003		3341
2004	=item w->again () C<ev::timer>, C<ev::periodic> only	3342	=item w->again () (C<ev::timer>, C<ev::periodic> only)
2005		3343
2006	For C<ev::timer> and C<ev::periodic>, this invokes the corresponding	3344	For C<ev::timer> and C<ev::periodic>, this invokes the corresponding
2007	C<ev_TYPE_again> function.	3345	C<ev_TYPE_again> function.
2008		3346
2009	=item w->sweep () C<ev::embed> only	3347	=item w->sweep () (C<ev::embed> only)
2010		3348
2011	Invokes C<ev_embed_sweep>.	3349	Invokes C<ev_embed_sweep>.
2012		3350
2013	=item w->update () C<ev::stat> only	3351	=item w->update () (C<ev::stat> only)
2014		3352
2015	Invokes C<ev_stat_stat>.	3353	Invokes C<ev_stat_stat>.
2016		3354
2017	=back	3355	=back
2018		3356
2019	=back	3357	=back
2020		3358
2021	Example: Define a class with an IO and idle watcher, start one of them in	3359	Example: Define a class with an IO and idle watcher, start one of them in
2022	the constructor.	3360	the constructor.
2023		3361
2024	class myclass	3362	class myclass
2025	{	3363	{
2026	ev_io io; void io_cb (ev::io &w, int revents);	3364	ev::io io ; void io_cb (ev::io &w, int revents);
2027	ev_idle idle void idle_cb (ev::idle &w, int revents);	3365	ev::idle idle; void idle_cb (ev::idle &w, int revents);
2028		3366
2029	myclass ();	3367	myclass (int fd)
2030	}	3368	{
2031
2032	myclass::myclass (int fd)
2033	{
2034	io .set <myclass, &myclass::io_cb > (this);	3369	io .set <myclass, &myclass::io_cb > (this);
2035	idle.set <myclass, &myclass::idle_cb> (this);	3370	idle.set <myclass, &myclass::idle_cb> (this);
2036		3371
2037	io.start (fd, ev::READ);	3372	io.start (fd, ev::READ);
		3373	}
2038	}	3374	};
		3375
		3376
		3377	=head1 OTHER LANGUAGE BINDINGS
		3378
		3379	Libev does not offer other language bindings itself, but bindings for a
		3380	number of languages exist in the form of third-party packages. If you know
		3381	any interesting language binding in addition to the ones listed here, drop
		3382	me a note.
		3383
		3384	=over 4
		3385
		3386	=item Perl
		3387
		3388	The EV module implements the full libev API and is actually used to test
		3389	libev. EV is developed together with libev. Apart from the EV core module,
		3390	there are additional modules that implement libev-compatible interfaces
		3391	to C<libadns> (C<EV::ADNS>, but C<AnyEvent::DNS> is preferred nowadays),
		3392	C<Net::SNMP> (C<Net::SNMP::EV>) and the C<libglib> event core (C<Glib::EV>
		3393	and C<EV::Glib>).
		3394
		3395	It can be found and installed via CPAN, its homepage is at
		3396	L<http://software.schmorp.de/pkg/EV>.
		3397
		3398	=item Python
		3399
		3400	Python bindings can be found at L<http://code.google.com/p/pyev/>. It
		3401	seems to be quite complete and well-documented.
		3402
		3403	=item Ruby
		3404
		3405	Tony Arcieri has written a ruby extension that offers access to a subset
		3406	of the libev API and adds file handle abstractions, asynchronous DNS and
		3407	more on top of it. It can be found via gem servers. Its homepage is at
		3408	L<http://rev.rubyforge.org/>.
		3409
		3410	Roger Pack reports that using the link order C<-lws2_32 -lmsvcrt-ruby-190>
		3411	makes rev work even on mingw.
		3412
		3413	=item Haskell
		3414
		3415	A haskell binding to libev is available at
		3416	L<http://hackage.haskell.org/cgi-bin/hackage-scripts/package/hlibev>.
		3417
		3418	=item D
		3419
		3420	Leandro Lucarella has written a D language binding (F<ev.d>) for libev, to
		3421	be found at L<http://proj.llucax.com.ar/wiki/evd>.
		3422
		3423	=item Ocaml
		3424
		3425	Erkki Seppala has written Ocaml bindings for libev, to be found at
		3426	L<http://modeemi.cs.tut.fi/~flux/software/ocaml-ev/>.
		3427
		3428	=item Lua
		3429
		3430	Brian Maher has written a partial interface to libev
		3431	for lua (only C<ev_io> and C<ev_timer>), to be found at
		3432	L<http://github.com/brimworks/lua-ev>.
		3433
		3434	=back
2039		3435
2040		3436
2041	=head1 MACRO MAGIC	3437	=head1 MACRO MAGIC
2042		3438
2043	Libev can be compiled with a variety of options, the most fundemantal is	3439	Libev can be compiled with a variety of options, the most fundamental
2044	C<EV_MULTIPLICITY>. This option determines whether (most) functions and	3440	of which is C<EV_MULTIPLICITY>. This option determines whether (most)
2045	callbacks have an initial C<struct ev_loop *> argument.	3441	functions and callbacks have an initial C<struct ev_loop *> argument.
2046		3442
2047	To make it easier to write programs that cope with either variant, the	3443	To make it easier to write programs that cope with either variant, the
2048	following macros are defined:	3444	following macros are defined:
2049		3445
2050	=over 4	3446	=over 4
…		…
2053		3449
2054	This provides the loop I<argument> for functions, if one is required ("ev	3450	This provides the loop I<argument> for functions, if one is required ("ev
2055	loop argument"). The C<EV_A> form is used when this is the sole argument,	3451	loop argument"). The C<EV_A> form is used when this is the sole argument,
2056	C<EV_A_> is used when other arguments are following. Example:	3452	C<EV_A_> is used when other arguments are following. Example:
2057		3453
2058	ev_unref (EV_A);	3454	ev_unref (EV_A);
2059	ev_timer_add (EV_A_ watcher);	3455	ev_timer_add (EV_A_ watcher);
2060	ev_loop (EV_A_ 0);	3456	ev_loop (EV_A_ 0);
2061		3457
2062	It assumes the variable C<loop> of type C<struct ev_loop *> is in scope,	3458	It assumes the variable C<loop> of type C<struct ev_loop *> is in scope,
2063	which is often provided by the following macro.	3459	which is often provided by the following macro.
2064		3460
2065	=item C<EV_P>, C<EV_P_>	3461	=item C<EV_P>, C<EV_P_>
2066		3462
2067	This provides the loop I<parameter> for functions, if one is required ("ev	3463	This provides the loop I<parameter> for functions, if one is required ("ev
2068	loop parameter"). The C<EV_P> form is used when this is the sole parameter,	3464	loop parameter"). The C<EV_P> form is used when this is the sole parameter,
2069	C<EV_P_> is used when other parameters are following. Example:	3465	C<EV_P_> is used when other parameters are following. Example:
2070		3466
2071	// this is how ev_unref is being declared	3467	// this is how ev_unref is being declared
2072	static void ev_unref (EV_P);	3468	static void ev_unref (EV_P);
2073		3469
2074	// this is how you can declare your typical callback	3470	// this is how you can declare your typical callback
2075	static void cb (EV_P_ ev_timer *w, int revents)	3471	static void cb (EV_P_ ev_timer *w, int revents)
2076		3472
2077	It declares a parameter C<loop> of type C<struct ev_loop *>, quite	3473	It declares a parameter C<loop> of type C<struct ev_loop *>, quite
2078	suitable for use with C<EV_A>.	3474	suitable for use with C<EV_A>.
2079		3475
2080	=item C<EV_DEFAULT>, C<EV_DEFAULT_>	3476	=item C<EV_DEFAULT>, C<EV_DEFAULT_>
2081		3477
2082	Similar to the other two macros, this gives you the value of the default	3478	Similar to the other two macros, this gives you the value of the default
2083	loop, if multiple loops are supported ("ev loop default").	3479	loop, if multiple loops are supported ("ev loop default").
		3480
		3481	=item C<EV_DEFAULT_UC>, C<EV_DEFAULT_UC_>
		3482
		3483	Usage identical to C<EV_DEFAULT> and C<EV_DEFAULT_>, but requires that the
		3484	default loop has been initialised (C<UC> == unchecked). Their behaviour
		3485	is undefined when the default loop has not been initialised by a previous
		3486	execution of C<EV_DEFAULT>, C<EV_DEFAULT_> or C<ev_default_init (...)>.
		3487
		3488	It is often prudent to use C<EV_DEFAULT> when initialising the first
		3489	watcher in a function but use C<EV_DEFAULT_UC> afterwards.
2084		3490
2085	=back	3491	=back
2086		3492
2087	Example: Declare and initialise a check watcher, utilising the above	3493	Example: Declare and initialise a check watcher, utilising the above
2088	macros so it will work regardless of whether multiple loops are supported	3494	macros so it will work regardless of whether multiple loops are supported
2089	or not.	3495	or not.
2090		3496
2091	static void	3497	static void
2092	check_cb (EV_P_ ev_timer *w, int revents)	3498	check_cb (EV_P_ ev_timer *w, int revents)
2093	{	3499	{
2094	ev_check_stop (EV_A_ w);	3500	ev_check_stop (EV_A_ w);
2095	}	3501	}
2096		3502
2097	ev_check check;	3503	ev_check check;
2098	ev_check_init (&check, check_cb);	3504	ev_check_init (&check, check_cb);
2099	ev_check_start (EV_DEFAULT_ &check);	3505	ev_check_start (EV_DEFAULT_ &check);
2100	ev_loop (EV_DEFAULT_ 0);	3506	ev_loop (EV_DEFAULT_ 0);
2101		3507
2102	=head1 EMBEDDING	3508	=head1 EMBEDDING
2103		3509
2104	Libev can (and often is) directly embedded into host	3510	Libev can (and often is) directly embedded into host
2105	applications. Examples of applications that embed it include the Deliantra	3511	applications. Examples of applications that embed it include the Deliantra
2106	Game Server, the EV perl module, the GNU Virtual Private Ethernet (gvpe)	3512	Game Server, the EV perl module, the GNU Virtual Private Ethernet (gvpe)
2107	and rxvt-unicode.	3513	and rxvt-unicode.
2108		3514
2109	The goal is to enable you to just copy the neecssary files into your	3515	The goal is to enable you to just copy the necessary files into your
2110	source directory without having to change even a single line in them, so	3516	source directory without having to change even a single line in them, so
2111	you can easily upgrade by simply copying (or having a checked-out copy of	3517	you can easily upgrade by simply copying (or having a checked-out copy of
2112	libev somewhere in your source tree).	3518	libev somewhere in your source tree).
2113		3519
2114	=head2 FILESETS	3520	=head2 FILESETS
2115		3521
2116	Depending on what features you need you need to include one or more sets of files	3522	Depending on what features you need you need to include one or more sets of files
2117	in your app.	3523	in your application.
2118		3524
2119	=head3 CORE EVENT LOOP	3525	=head3 CORE EVENT LOOP
2120		3526
2121	To include only the libev core (all the C<ev_*> functions), with manual	3527	To include only the libev core (all the C<ev_*> functions), with manual
2122	configuration (no autoconf):	3528	configuration (no autoconf):
2123		3529
2124	#define EV_STANDALONE 1	3530	#define EV_STANDALONE 1
2125	#include "ev.c"	3531	#include "ev.c"
2126		3532
2127	This will automatically include F<ev.h>, too, and should be done in a	3533	This will automatically include F<ev.h>, too, and should be done in a
2128	single C source file only to provide the function implementations. To use	3534	single C source file only to provide the function implementations. To use
2129	it, do the same for F<ev.h> in all files wishing to use this API (best	3535	it, do the same for F<ev.h> in all files wishing to use this API (best
2130	done by writing a wrapper around F<ev.h> that you can include instead and	3536	done by writing a wrapper around F<ev.h> that you can include instead and
2131	where you can put other configuration options):	3537	where you can put other configuration options):
2132		3538
2133	#define EV_STANDALONE 1	3539	#define EV_STANDALONE 1
2134	#include "ev.h"	3540	#include "ev.h"
2135		3541
2136	Both header files and implementation files can be compiled with a C++	3542	Both header files and implementation files can be compiled with a C++
2137	compiler (at least, thats a stated goal, and breakage will be treated	3543	compiler (at least, that's a stated goal, and breakage will be treated
2138	as a bug).	3544	as a bug).
2139		3545
2140	You need the following files in your source tree, or in a directory	3546	You need the following files in your source tree, or in a directory
2141	in your include path (e.g. in libev/ when using -Ilibev):	3547	in your include path (e.g. in libev/ when using -Ilibev):
2142		3548
2143	ev.h	3549	ev.h
2144	ev.c	3550	ev.c
2145	ev_vars.h	3551	ev_vars.h
2146	ev_wrap.h	3552	ev_wrap.h
2147		3553
2148	ev_win32.c required on win32 platforms only	3554	ev_win32.c required on win32 platforms only
2149		3555
2150	ev_select.c only when select backend is enabled (which is enabled by default)	3556	ev_select.c only when select backend is enabled (which is enabled by default)
2151	ev_poll.c only when poll backend is enabled (disabled by default)	3557	ev_poll.c only when poll backend is enabled (disabled by default)
2152	ev_epoll.c only when the epoll backend is enabled (disabled by default)	3558	ev_epoll.c only when the epoll backend is enabled (disabled by default)
2153	ev_kqueue.c only when the kqueue backend is enabled (disabled by default)	3559	ev_kqueue.c only when the kqueue backend is enabled (disabled by default)
2154	ev_port.c only when the solaris port backend is enabled (disabled by default)	3560	ev_port.c only when the solaris port backend is enabled (disabled by default)
2155		3561
2156	F<ev.c> includes the backend files directly when enabled, so you only need	3562	F<ev.c> includes the backend files directly when enabled, so you only need
2157	to compile this single file.	3563	to compile this single file.
2158		3564
2159	=head3 LIBEVENT COMPATIBILITY API	3565	=head3 LIBEVENT COMPATIBILITY API
2160		3566
2161	To include the libevent compatibility API, also include:	3567	To include the libevent compatibility API, also include:
2162		3568
2163	#include "event.c"	3569	#include "event.c"
2164		3570
2165	in the file including F<ev.c>, and:	3571	in the file including F<ev.c>, and:
2166		3572
2167	#include "event.h"	3573	#include "event.h"
2168		3574
2169	in the files that want to use the libevent API. This also includes F<ev.h>.	3575	in the files that want to use the libevent API. This also includes F<ev.h>.
2170		3576
2171	You need the following additional files for this:	3577	You need the following additional files for this:
2172		3578
2173	event.h	3579	event.h
2174	event.c	3580	event.c
2175		3581
2176	=head3 AUTOCONF SUPPORT	3582	=head3 AUTOCONF SUPPORT
2177		3583
2178	Instead of using C<EV_STANDALONE=1> and providing your config in	3584	Instead of using C<EV_STANDALONE=1> and providing your configuration in
2179	whatever way you want, you can also C<m4_include([libev.m4])> in your	3585	whatever way you want, you can also C<m4_include([libev.m4])> in your
2180	F<configure.ac> and leave C<EV_STANDALONE> undefined. F<ev.c> will then	3586	F<configure.ac> and leave C<EV_STANDALONE> undefined. F<ev.c> will then
2181	include F<config.h> and configure itself accordingly.	3587	include F<config.h> and configure itself accordingly.
2182		3588
2183	For this of course you need the m4 file:	3589	For this of course you need the m4 file:
2184		3590
2185	libev.m4	3591	libev.m4
2186		3592
2187	=head2 PREPROCESSOR SYMBOLS/MACROS	3593	=head2 PREPROCESSOR SYMBOLS/MACROS
2188		3594
2189	Libev can be configured via a variety of preprocessor symbols you have to define	3595	Libev can be configured via a variety of preprocessor symbols you have to
2190	before including any of its files. The default is not to build for multiplicity	3596	define before including any of its files. The default in the absence of
2191	and only include the select backend.	3597	autoconf is documented for every option.
2192		3598
2193	=over 4	3599	=over 4
2194		3600
2195	=item EV_STANDALONE	3601	=item EV_STANDALONE
2196		3602
…		…
2198	keeps libev from including F<config.h>, and it also defines dummy	3604	keeps libev from including F<config.h>, and it also defines dummy
2199	implementations for some libevent functions (such as logging, which is not	3605	implementations for some libevent functions (such as logging, which is not
2200	supported). It will also not define any of the structs usually found in	3606	supported). It will also not define any of the structs usually found in
2201	F<event.h> that are not directly supported by the libev core alone.	3607	F<event.h> that are not directly supported by the libev core alone.
2202		3608
		3609	In standalone mode, libev will still try to automatically deduce the
		3610	configuration, but has to be more conservative.
		3611
2203	=item EV_USE_MONOTONIC	3612	=item EV_USE_MONOTONIC
2204		3613
2205	If defined to be C<1>, libev will try to detect the availability of the	3614	If defined to be C<1>, libev will try to detect the availability of the
2206	monotonic clock option at both compiletime and runtime. Otherwise no use	3615	monotonic clock option at both compile time and runtime. Otherwise no
2207	of the monotonic clock option will be attempted. If you enable this, you	3616	use of the monotonic clock option will be attempted. If you enable this,
2208	usually have to link against librt or something similar. Enabling it when	3617	you usually have to link against librt or something similar. Enabling it
2209	the functionality isn't available is safe, though, althoguh you have	3618	when the functionality isn't available is safe, though, although you have
2210	to make sure you link against any libraries where the C<clock_gettime>	3619	to make sure you link against any libraries where the C<clock_gettime>
2211	function is hiding in (often F<-lrt>).	3620	function is hiding in (often F<-lrt>). See also C<EV_USE_CLOCK_SYSCALL>.
2212		3621
2213	=item EV_USE_REALTIME	3622	=item EV_USE_REALTIME
2214		3623
2215	If defined to be C<1>, libev will try to detect the availability of the	3624	If defined to be C<1>, libev will try to detect the availability of the
2216	realtime clock option at compiletime (and assume its availability at	3625	real-time clock option at compile time (and assume its availability
2217	runtime if successful). Otherwise no use of the realtime clock option will	3626	at runtime if successful). Otherwise no use of the real-time clock
2218	be attempted. This effectively replaces C<gettimeofday> by C<clock_get	3627	option will be attempted. This effectively replaces C<gettimeofday>
2219	(CLOCK_REALTIME, ...)> and will not normally affect correctness. See tzhe note about libraries	3628	by C<clock_get (CLOCK_REALTIME, ...)> and will not normally affect
2220	in the description of C<EV_USE_MONOTONIC>, though.	3629	correctness. See the note about libraries in the description of
		3630	C<EV_USE_MONOTONIC>, though. Defaults to the opposite value of
		3631	C<EV_USE_CLOCK_SYSCALL>.
		3632
		3633	=item EV_USE_CLOCK_SYSCALL
		3634
		3635	If defined to be C<1>, libev will try to use a direct syscall instead
		3636	of calling the system-provided C<clock_gettime> function. This option
		3637	exists because on GNU/Linux, C<clock_gettime> is in C<librt>, but C<librt>
		3638	unconditionally pulls in C<libpthread>, slowing down single-threaded
		3639	programs needlessly. Using a direct syscall is slightly slower (in
		3640	theory), because no optimised vdso implementation can be used, but avoids
		3641	the pthread dependency. Defaults to C<1> on GNU/Linux with glibc 2.x or
		3642	higher, as it simplifies linking (no need for C<-lrt>).
		3643
		3644	=item EV_USE_NANOSLEEP
		3645
		3646	If defined to be C<1>, libev will assume that C<nanosleep ()> is available
		3647	and will use it for delays. Otherwise it will use C<select ()>.
		3648
		3649	=item EV_USE_EVENTFD
		3650
		3651	If defined to be C<1>, then libev will assume that C<eventfd ()> is
		3652	available and will probe for kernel support at runtime. This will improve
		3653	C<ev_signal> and C<ev_async> performance and reduce resource consumption.
		3654	If undefined, it will be enabled if the headers indicate GNU/Linux + Glibc
		3655	2.7 or newer, otherwise disabled.
2221		3656
2222	=item EV_USE_SELECT	3657	=item EV_USE_SELECT
2223		3658
2224	If undefined or defined to be C<1>, libev will compile in support for the	3659	If undefined or defined to be C<1>, libev will compile in support for the
2225	C<select>(2) backend. No attempt at autodetection will be done: if no	3660	C<select>(2) backend. No attempt at auto-detection will be done: if no
2226	other method takes over, select will be it. Otherwise the select backend	3661	other method takes over, select will be it. Otherwise the select backend
2227	will not be compiled in.	3662	will not be compiled in.
2228		3663
2229	=item EV_SELECT_USE_FD_SET	3664	=item EV_SELECT_USE_FD_SET
2230		3665
2231	If defined to C<1>, then the select backend will use the system C<fd_set>	3666	If defined to C<1>, then the select backend will use the system C<fd_set>
2232	structure. This is useful if libev doesn't compile due to a missing	3667	structure. This is useful if libev doesn't compile due to a missing
2233	C<NFDBITS> or C<fd_mask> definition or it misguesses the bitset layout on	3668	C<NFDBITS> or C<fd_mask> definition or it mis-guesses the bitset layout
2234	exotic systems. This usually limits the range of file descriptors to some	3669	on exotic systems. This usually limits the range of file descriptors to
2235	low limit such as 1024 or might have other limitations (winsocket only	3670	some low limit such as 1024 or might have other limitations (winsocket
2236	allows 64 sockets). The C<FD_SETSIZE> macro, set before compilation, might	3671	only allows 64 sockets). The C<FD_SETSIZE> macro, set before compilation,
2237	influence the size of the C<fd_set> used.	3672	configures the maximum size of the C<fd_set>.
2238		3673
2239	=item EV_SELECT_IS_WINSOCKET	3674	=item EV_SELECT_IS_WINSOCKET
2240		3675
2241	When defined to C<1>, the select backend will assume that	3676	When defined to C<1>, the select backend will assume that
2242	select/socket/connect etc. don't understand file descriptors but	3677	select/socket/connect etc. don't understand file descriptors but
…		…
2244	be used is the winsock select). This means that it will call	3679	be used is the winsock select). This means that it will call
2245	C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,	3680	C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,
2246	it is assumed that all these functions actually work on fds, even	3681	it is assumed that all these functions actually work on fds, even
2247	on win32. Should not be defined on non-win32 platforms.	3682	on win32. Should not be defined on non-win32 platforms.
2248		3683
		3684	=item EV_FD_TO_WIN32_HANDLE(fd)
		3685
		3686	If C<EV_SELECT_IS_WINSOCKET> is enabled, then libev needs a way to map
		3687	file descriptors to socket handles. When not defining this symbol (the
		3688	default), then libev will call C<_get_osfhandle>, which is usually
		3689	correct. In some cases, programs use their own file descriptor management,
		3690	in which case they can provide this function to map fds to socket handles.
		3691
		3692	=item EV_WIN32_HANDLE_TO_FD(handle)
		3693
		3694	If C<EV_SELECT_IS_WINSOCKET> then libev maps handles to file descriptors
		3695	using the standard C<_open_osfhandle> function. For programs implementing
		3696	their own fd to handle mapping, overwriting this function makes it easier
		3697	to do so. This can be done by defining this macro to an appropriate value.
		3698
		3699	=item EV_WIN32_CLOSE_FD(fd)
		3700
		3701	If programs implement their own fd to handle mapping on win32, then this
		3702	macro can be used to override the C<close> function, useful to unregister
		3703	file descriptors again. Note that the replacement function has to close
		3704	the underlying OS handle.
		3705
2249	=item EV_USE_POLL	3706	=item EV_USE_POLL
2250		3707
2251	If defined to be C<1>, libev will compile in support for the C<poll>(2)	3708	If defined to be C<1>, libev will compile in support for the C<poll>(2)
2252	backend. Otherwise it will be enabled on non-win32 platforms. It	3709	backend. Otherwise it will be enabled on non-win32 platforms. It
2253	takes precedence over select.	3710	takes precedence over select.
2254		3711
2255	=item EV_USE_EPOLL	3712	=item EV_USE_EPOLL
2256		3713
2257	If defined to be C<1>, libev will compile in support for the Linux	3714	If defined to be C<1>, libev will compile in support for the Linux
2258	C<epoll>(7) backend. Its availability will be detected at runtime,	3715	C<epoll>(7) backend. Its availability will be detected at runtime,
2259	otherwise another method will be used as fallback. This is the	3716	otherwise another method will be used as fallback. This is the preferred
2260	preferred backend for GNU/Linux systems.	3717	backend for GNU/Linux systems. If undefined, it will be enabled if the
		3718	headers indicate GNU/Linux + Glibc 2.4 or newer, otherwise disabled.
2261		3719
2262	=item EV_USE_KQUEUE	3720	=item EV_USE_KQUEUE
2263		3721
2264	If defined to be C<1>, libev will compile in support for the BSD style	3722	If defined to be C<1>, libev will compile in support for the BSD style
2265	C<kqueue>(2) backend. Its actual availability will be detected at runtime,	3723	C<kqueue>(2) backend. Its actual availability will be detected at runtime,
…		…
2278	otherwise another method will be used as fallback. This is the preferred	3736	otherwise another method will be used as fallback. This is the preferred
2279	backend for Solaris 10 systems.	3737	backend for Solaris 10 systems.
2280		3738
2281	=item EV_USE_DEVPOLL	3739	=item EV_USE_DEVPOLL
2282		3740
2283	reserved for future expansion, works like the USE symbols above.	3741	Reserved for future expansion, works like the USE symbols above.
2284		3742
2285	=item EV_USE_INOTIFY	3743	=item EV_USE_INOTIFY
2286		3744
2287	If defined to be C<1>, libev will compile in support for the Linux inotify	3745	If defined to be C<1>, libev will compile in support for the Linux inotify
2288	interface to speed up C<ev_stat> watchers. Its actual availability will	3746	interface to speed up C<ev_stat> watchers. Its actual availability will
2289	be detected at runtime.	3747	be detected at runtime. If undefined, it will be enabled if the headers
		3748	indicate GNU/Linux + Glibc 2.4 or newer, otherwise disabled.
		3749
		3750	=item EV_ATOMIC_T
		3751
		3752	Libev requires an integer type (suitable for storing C<0> or C<1>) whose
		3753	access is atomic with respect to other threads or signal contexts. No such
		3754	type is easily found in the C language, so you can provide your own type
		3755	that you know is safe for your purposes. It is used both for signal handler "locking"
		3756	as well as for signal and thread safety in C<ev_async> watchers.
		3757
		3758	In the absence of this define, libev will use C<sig_atomic_t volatile>
		3759	(from F<signal.h>), which is usually good enough on most platforms.
2290		3760
2291	=item EV_H	3761	=item EV_H
2292		3762
2293	The name of the F<ev.h> header file used to include it. The default if	3763	The name of the F<ev.h> header file used to include it. The default if
2294	undefined is C<< <ev.h> >> in F<event.h> and C<"ev.h"> in F<ev.c>. This	3764	undefined is C<"ev.h"> in F<event.h>, F<ev.c> and F<ev++.h>. This can be
2295	can be used to virtually rename the F<ev.h> header file in case of conflicts.	3765	used to virtually rename the F<ev.h> header file in case of conflicts.
2296		3766
2297	=item EV_CONFIG_H	3767	=item EV_CONFIG_H
2298		3768
2299	If C<EV_STANDALONE> isn't C<1>, this variable can be used to override	3769	If C<EV_STANDALONE> isn't C<1>, this variable can be used to override
2300	F<ev.c>'s idea of where to find the F<config.h> file, similarly to	3770	F<ev.c>'s idea of where to find the F<config.h> file, similarly to
2301	C<EV_H>, above.	3771	C<EV_H>, above.
2302		3772
2303	=item EV_EVENT_H	3773	=item EV_EVENT_H
2304		3774
2305	Similarly to C<EV_H>, this macro can be used to override F<event.c>'s idea	3775	Similarly to C<EV_H>, this macro can be used to override F<event.c>'s idea
2306	of how the F<event.h> header can be found.	3776	of how the F<event.h> header can be found, the default is C<"event.h">.
2307		3777
2308	=item EV_PROTOTYPES	3778	=item EV_PROTOTYPES
2309		3779
2310	If defined to be C<0>, then F<ev.h> will not define any function	3780	If defined to be C<0>, then F<ev.h> will not define any function
2311	prototypes, but still define all the structs and other symbols. This is	3781	prototypes, but still define all the structs and other symbols. This is
…		…
2332	When doing priority-based operations, libev usually has to linearly search	3802	When doing priority-based operations, libev usually has to linearly search
2333	all the priorities, so having many of them (hundreds) uses a lot of space	3803	all the priorities, so having many of them (hundreds) uses a lot of space
2334	and time, so using the defaults of five priorities (-2 .. +2) is usually	3804	and time, so using the defaults of five priorities (-2 .. +2) is usually
2335	fine.	3805	fine.
2336		3806
2337	If your embedding app does not need any priorities, defining these both to	3807	If your embedding application does not need any priorities, defining these
2338	C<0> will save some memory and cpu.	3808	both to C<0> will save some memory and CPU.
2339		3809
2340	=item EV_PERIODIC_ENABLE	3810	=item EV_PERIODIC_ENABLE
2341		3811
2342	If undefined or defined to be C<1>, then periodic timers are supported. If	3812	If undefined or defined to be C<1>, then periodic timers are supported. If
2343	defined to be C<0>, then they are not. Disabling them saves a few kB of	3813	defined to be C<0>, then they are not. Disabling them saves a few kB of
…		…
2350	code.	3820	code.
2351		3821
2352	=item EV_EMBED_ENABLE	3822	=item EV_EMBED_ENABLE
2353		3823
2354	If undefined or defined to be C<1>, then embed watchers are supported. If	3824	If undefined or defined to be C<1>, then embed watchers are supported. If
2355	defined to be C<0>, then they are not.	3825	defined to be C<0>, then they are not. Embed watchers rely on most other
		3826	watcher types, which therefore must not be disabled.
2356		3827
2357	=item EV_STAT_ENABLE	3828	=item EV_STAT_ENABLE
2358		3829
2359	If undefined or defined to be C<1>, then stat watchers are supported. If	3830	If undefined or defined to be C<1>, then stat watchers are supported. If
2360	defined to be C<0>, then they are not.	3831	defined to be C<0>, then they are not.
…		…
2362	=item EV_FORK_ENABLE	3833	=item EV_FORK_ENABLE
2363		3834
2364	If undefined or defined to be C<1>, then fork watchers are supported. If	3835	If undefined or defined to be C<1>, then fork watchers are supported. If
2365	defined to be C<0>, then they are not.	3836	defined to be C<0>, then they are not.
2366		3837
		3838	=item EV_ASYNC_ENABLE
		3839
		3840	If undefined or defined to be C<1>, then async watchers are supported. If
		3841	defined to be C<0>, then they are not.
		3842
2367	=item EV_MINIMAL	3843	=item EV_MINIMAL
2368		3844
2369	If you need to shave off some kilobytes of code at the expense of some	3845	If you need to shave off some kilobytes of code at the expense of some
2370	speed, define this symbol to C<1>. Currently only used for gcc to override	3846	speed (but with the full API), define this symbol to C<1>. Currently this
2371	some inlining decisions, saves roughly 30% codesize of amd64.	3847	is used to override some inlining decisions, saves roughly 30% code size
		3848	on amd64. It also selects a much smaller 2-heap for timer management over
		3849	the default 4-heap.
		3850
		3851	You can save even more by disabling watcher types you do not need
		3852	and setting C<EV_MAXPRI> == C<EV_MINPRI>. Also, disabling C<assert>
		3853	(C<-DNDEBUG>) will usually reduce code size a lot.
		3854
		3855	Defining C<EV_MINIMAL> to C<2> will additionally reduce the core API to
		3856	provide a bare-bones event library. See C<ev.h> for details on what parts
		3857	of the API are still available, and do not complain if this subset changes
		3858	over time.
		3859
		3860	=item EV_NSIG
		3861
		3862	The highest supported signal number, +1 (or, the number of
		3863	signals): Normally, libev tries to deduce the maximum number of signals
		3864	automatically, but sometimes this fails, in which case it can be
		3865	specified. Also, using a lower number than detected (C<32> should be
		3866	good for about any system in existance) can save some memory, as libev
		3867	statically allocates some 12-24 bytes per signal number.
2372		3868
2373	=item EV_PID_HASHSIZE	3869	=item EV_PID_HASHSIZE
2374		3870
2375	C<ev_child> watchers use a small hash table to distribute workload by	3871	C<ev_child> watchers use a small hash table to distribute workload by
2376	pid. The default size is C<16> (or C<1> with C<EV_MINIMAL>), usually more	3872	pid. The default size is C<16> (or C<1> with C<EV_MINIMAL>), usually more
2377	than enough. If you need to manage thousands of children you might want to	3873	than enough. If you need to manage thousands of children you might want to
2378	increase this value (I<must> be a power of two).	3874	increase this value (I<must> be a power of two).
2379		3875
2380	=item EV_INOTIFY_HASHSIZE	3876	=item EV_INOTIFY_HASHSIZE
2381		3877
2382	C<ev_staz> watchers use a small hash table to distribute workload by	3878	C<ev_stat> watchers use a small hash table to distribute workload by
2383	inotify watch id. The default size is C<16> (or C<1> with C<EV_MINIMAL>),	3879	inotify watch id. The default size is C<16> (or C<1> with C<EV_MINIMAL>),
2384	usually more than enough. If you need to manage thousands of C<ev_stat>	3880	usually more than enough. If you need to manage thousands of C<ev_stat>
2385	watchers you might want to increase this value (I<must> be a power of	3881	watchers you might want to increase this value (I<must> be a power of
2386	two).	3882	two).
2387		3883
		3884	=item EV_USE_4HEAP
		3885
		3886	Heaps are not very cache-efficient. To improve the cache-efficiency of the
		3887	timer and periodics heaps, libev uses a 4-heap when this symbol is defined
		3888	to C<1>. The 4-heap uses more complicated (longer) code but has noticeably
		3889	faster performance with many (thousands) of watchers.
		3890
		3891	The default is C<1> unless C<EV_MINIMAL> is set in which case it is C<0>
		3892	(disabled).
		3893
		3894	=item EV_HEAP_CACHE_AT
		3895
		3896	Heaps are not very cache-efficient. To improve the cache-efficiency of the
		3897	timer and periodics heaps, libev can cache the timestamp (I<at>) within
		3898	the heap structure (selected by defining C<EV_HEAP_CACHE_AT> to C<1>),
		3899	which uses 8-12 bytes more per watcher and a few hundred bytes more code,
		3900	but avoids random read accesses on heap changes. This improves performance
		3901	noticeably with many (hundreds) of watchers.
		3902
		3903	The default is C<1> unless C<EV_MINIMAL> is set in which case it is C<0>
		3904	(disabled).
		3905
		3906	=item EV_VERIFY
		3907
		3908	Controls how much internal verification (see C<ev_loop_verify ()>) will
		3909	be done: If set to C<0>, no internal verification code will be compiled
		3910	in. If set to C<1>, then verification code will be compiled in, but not
		3911	called. If set to C<2>, then the internal verification code will be
		3912	called once per loop, which can slow down libev. If set to C<3>, then the
		3913	verification code will be called very frequently, which will slow down
		3914	libev considerably.
		3915
		3916	The default is C<1>, unless C<EV_MINIMAL> is set, in which case it will be
		3917	C<0>.
		3918
2388	=item EV_COMMON	3919	=item EV_COMMON
2389		3920
2390	By default, all watchers have a C<void *data> member. By redefining	3921	By default, all watchers have a C<void *data> member. By redefining
2391	this macro to a something else you can include more and other types of	3922	this macro to a something else you can include more and other types of
2392	members. You have to define it each time you include one of the files,	3923	members. You have to define it each time you include one of the files,
2393	though, and it must be identical each time.	3924	though, and it must be identical each time.
2394		3925
2395	For example, the perl EV module uses something like this:	3926	For example, the perl EV module uses something like this:
2396		3927
2397	#define EV_COMMON \	3928	#define EV_COMMON \
2398	SV *self; /* contains this struct */ \	3929	SV *self; /* contains this struct */ \
2399	SV *cb_sv, *fh /* note no trailing ";" */	3930	SV *cb_sv, *fh /* note no trailing ";" */
2400		3931
2401	=item EV_CB_DECLARE (type)	3932	=item EV_CB_DECLARE (type)
2402		3933
2403	=item EV_CB_INVOKE (watcher, revents)	3934	=item EV_CB_INVOKE (watcher, revents)
2404		3935
2405	=item ev_set_cb (ev, cb)	3936	=item ev_set_cb (ev, cb)
2406		3937
2407	Can be used to change the callback member declaration in each watcher,	3938	Can be used to change the callback member declaration in each watcher,
2408	and the way callbacks are invoked and set. Must expand to a struct member	3939	and the way callbacks are invoked and set. Must expand to a struct member
2409	definition and a statement, respectively. See the F<ev.v> header file for	3940	definition and a statement, respectively. See the F<ev.h> header file for
2410	their default definitions. One possible use for overriding these is to	3941	their default definitions. One possible use for overriding these is to
2411	avoid the C<struct ev_loop *> as first argument in all cases, or to use	3942	avoid the C<struct ev_loop *> as first argument in all cases, or to use
2412	method calls instead of plain function calls in C++.	3943	method calls instead of plain function calls in C++.
		3944
		3945	=back
		3946
		3947	=head2 EXPORTED API SYMBOLS
		3948
		3949	If you need to re-export the API (e.g. via a DLL) and you need a list of
		3950	exported symbols, you can use the provided F<Symbol.*> files which list
		3951	all public symbols, one per line:
		3952
		3953	Symbols.ev for libev proper
		3954	Symbols.event for the libevent emulation
		3955
		3956	This can also be used to rename all public symbols to avoid clashes with
		3957	multiple versions of libev linked together (which is obviously bad in
		3958	itself, but sometimes it is inconvenient to avoid this).
		3959
		3960	A sed command like this will create wrapper C<#define>'s that you need to
		3961	include before including F<ev.h>:
		3962
		3963	<Symbols.ev sed -e "s/.*/#define & myprefix_&/" >wrap.h
		3964
		3965	This would create a file F<wrap.h> which essentially looks like this:
		3966
		3967	#define ev_backend myprefix_ev_backend
		3968	#define ev_check_start myprefix_ev_check_start
		3969	#define ev_check_stop myprefix_ev_check_stop
		3970	...
2413		3971
2414	=head2 EXAMPLES	3972	=head2 EXAMPLES
2415		3973
2416	For a real-world example of a program the includes libev	3974	For a real-world example of a program the includes libev
2417	verbatim, you can have a look at the EV perl module	3975	verbatim, you can have a look at the EV perl module
…		…
2422	file.	3980	file.
2423		3981
2424	The usage in rxvt-unicode is simpler. It has a F<ev_cpp.h> header file	3982	The usage in rxvt-unicode is simpler. It has a F<ev_cpp.h> header file
2425	that everybody includes and which overrides some configure choices:	3983	that everybody includes and which overrides some configure choices:
2426		3984
2427	#define EV_MINIMAL 1	3985	#define EV_MINIMAL 1
2428	#define EV_USE_POLL 0	3986	#define EV_USE_POLL 0
2429	#define EV_MULTIPLICITY 0	3987	#define EV_MULTIPLICITY 0
2430	#define EV_PERIODIC_ENABLE 0	3988	#define EV_PERIODIC_ENABLE 0
2431	#define EV_STAT_ENABLE 0	3989	#define EV_STAT_ENABLE 0
2432	#define EV_FORK_ENABLE 0	3990	#define EV_FORK_ENABLE 0
2433	#define EV_CONFIG_H <config.h>	3991	#define EV_CONFIG_H <config.h>
2434	#define EV_MINPRI 0	3992	#define EV_MINPRI 0
2435	#define EV_MAXPRI 0	3993	#define EV_MAXPRI 0
2436		3994
2437	#include "ev++.h"	3995	#include "ev++.h"
2438		3996
2439	And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:	3997	And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:
2440		3998
2441	#include "ev_cpp.h"	3999	#include "ev_cpp.h"
2442	#include "ev.c"	4000	#include "ev.c"
2443		4001
		4002	=head1 INTERACTION WITH OTHER PROGRAMS OR LIBRARIES
2444		4003
		4004	=head2 THREADS AND COROUTINES
		4005
		4006	=head3 THREADS
		4007
		4008	All libev functions are reentrant and thread-safe unless explicitly
		4009	documented otherwise, but libev implements no locking itself. This means
		4010	that you can use as many loops as you want in parallel, as long as there
		4011	are no concurrent calls into any libev function with the same loop
		4012	parameter (C<ev_default_*> calls have an implicit default loop parameter,
		4013	of course): libev guarantees that different event loops share no data
		4014	structures that need any locking.
		4015
		4016	Or to put it differently: calls with different loop parameters can be done
		4017	concurrently from multiple threads, calls with the same loop parameter
		4018	must be done serially (but can be done from different threads, as long as
		4019	only one thread ever is inside a call at any point in time, e.g. by using
		4020	a mutex per loop).
		4021
		4022	Specifically to support threads (and signal handlers), libev implements
		4023	so-called C<ev_async> watchers, which allow some limited form of
		4024	concurrency on the same event loop, namely waking it up "from the
		4025	outside".
		4026
		4027	If you want to know which design (one loop, locking, or multiple loops
		4028	without or something else still) is best for your problem, then I cannot
		4029	help you, but here is some generic advice:
		4030
		4031	=over 4
		4032
		4033	=item * most applications have a main thread: use the default libev loop
		4034	in that thread, or create a separate thread running only the default loop.
		4035
		4036	This helps integrating other libraries or software modules that use libev
		4037	themselves and don't care/know about threading.
		4038
		4039	=item * one loop per thread is usually a good model.
		4040
		4041	Doing this is almost never wrong, sometimes a better-performance model
		4042	exists, but it is always a good start.
		4043
		4044	=item * other models exist, such as the leader/follower pattern, where one
		4045	loop is handed through multiple threads in a kind of round-robin fashion.
		4046
		4047	Choosing a model is hard - look around, learn, know that usually you can do
		4048	better than you currently do :-)
		4049
		4050	=item * often you need to talk to some other thread which blocks in the
		4051	event loop.
		4052
		4053	C<ev_async> watchers can be used to wake them up from other threads safely
		4054	(or from signal contexts...).
		4055
		4056	An example use would be to communicate signals or other events that only
		4057	work in the default loop by registering the signal watcher with the
		4058	default loop and triggering an C<ev_async> watcher from the default loop
		4059	watcher callback into the event loop interested in the signal.
		4060
		4061	=back
		4062
		4063	=head4 THREAD LOCKING EXAMPLE
		4064
		4065	Here is a fictitious example of how to run an event loop in a different
		4066	thread than where callbacks are being invoked and watchers are
		4067	created/added/removed.
		4068
		4069	For a real-world example, see the C<EV::Loop::Async> perl module,
		4070	which uses exactly this technique (which is suited for many high-level
		4071	languages).
		4072
		4073	The example uses a pthread mutex to protect the loop data, a condition
		4074	variable to wait for callback invocations, an async watcher to notify the
		4075	event loop thread and an unspecified mechanism to wake up the main thread.
		4076
		4077	First, you need to associate some data with the event loop:
		4078
		4079	typedef struct {
		4080	mutex_t lock; /* global loop lock */
		4081	ev_async async_w;
		4082	thread_t tid;
		4083	cond_t invoke_cv;
		4084	} userdata;
		4085
		4086	void prepare_loop (EV_P)
		4087	{
		4088	// for simplicity, we use a static userdata struct.
		4089	static userdata u;
		4090
		4091	ev_async_init (&u->async_w, async_cb);
		4092	ev_async_start (EV_A_ &u->async_w);
		4093
		4094	pthread_mutex_init (&u->lock, 0);
		4095	pthread_cond_init (&u->invoke_cv, 0);
		4096
		4097	// now associate this with the loop
		4098	ev_set_userdata (EV_A_ u);
		4099	ev_set_invoke_pending_cb (EV_A_ l_invoke);
		4100	ev_set_loop_release_cb (EV_A_ l_release, l_acquire);
		4101
		4102	// then create the thread running ev_loop
		4103	pthread_create (&u->tid, 0, l_run, EV_A);
		4104	}
		4105
		4106	The callback for the C<ev_async> watcher does nothing: the watcher is used
		4107	solely to wake up the event loop so it takes notice of any new watchers
		4108	that might have been added:
		4109
		4110	static void
		4111	async_cb (EV_P_ ev_async *w, int revents)
		4112	{
		4113	// just used for the side effects
		4114	}
		4115
		4116	The C<l_release> and C<l_acquire> callbacks simply unlock/lock the mutex
		4117	protecting the loop data, respectively.
		4118
		4119	static void
		4120	l_release (EV_P)
		4121	{
		4122	userdata *u = ev_userdata (EV_A);
		4123	pthread_mutex_unlock (&u->lock);
		4124	}
		4125
		4126	static void
		4127	l_acquire (EV_P)
		4128	{
		4129	userdata *u = ev_userdata (EV_A);
		4130	pthread_mutex_lock (&u->lock);
		4131	}
		4132
		4133	The event loop thread first acquires the mutex, and then jumps straight
		4134	into C<ev_loop>:
		4135
		4136	void *
		4137	l_run (void *thr_arg)
		4138	{
		4139	struct ev_loop *loop = (struct ev_loop *)thr_arg;
		4140
		4141	l_acquire (EV_A);
		4142	pthread_setcanceltype (PTHREAD_CANCEL_ASYNCHRONOUS, 0);
		4143	ev_loop (EV_A_ 0);
		4144	l_release (EV_A);
		4145
		4146	return 0;
		4147	}
		4148
		4149	Instead of invoking all pending watchers, the C<l_invoke> callback will
		4150	signal the main thread via some unspecified mechanism (signals? pipe
		4151	writes? C<Async::Interrupt>?) and then waits until all pending watchers
		4152	have been called (in a while loop because a) spurious wakeups are possible
		4153	and b) skipping inter-thread-communication when there are no pending
		4154	watchers is very beneficial):
		4155
		4156	static void
		4157	l_invoke (EV_P)
		4158	{
		4159	userdata *u = ev_userdata (EV_A);
		4160
		4161	while (ev_pending_count (EV_A))
		4162	{
		4163	wake_up_other_thread_in_some_magic_or_not_so_magic_way ();
		4164	pthread_cond_wait (&u->invoke_cv, &u->lock);
		4165	}
		4166	}
		4167
		4168	Now, whenever the main thread gets told to invoke pending watchers, it
		4169	will grab the lock, call C<ev_invoke_pending> and then signal the loop
		4170	thread to continue:
		4171
		4172	static void
		4173	real_invoke_pending (EV_P)
		4174	{
		4175	userdata *u = ev_userdata (EV_A);
		4176
		4177	pthread_mutex_lock (&u->lock);
		4178	ev_invoke_pending (EV_A);
		4179	pthread_cond_signal (&u->invoke_cv);
		4180	pthread_mutex_unlock (&u->lock);
		4181	}
		4182
		4183	Whenever you want to start/stop a watcher or do other modifications to an
		4184	event loop, you will now have to lock:
		4185
		4186	ev_timer timeout_watcher;
		4187	userdata *u = ev_userdata (EV_A);
		4188
		4189	ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);
		4190
		4191	pthread_mutex_lock (&u->lock);
		4192	ev_timer_start (EV_A_ &timeout_watcher);
		4193	ev_async_send (EV_A_ &u->async_w);
		4194	pthread_mutex_unlock (&u->lock);
		4195
		4196	Note that sending the C<ev_async> watcher is required because otherwise
		4197	an event loop currently blocking in the kernel will have no knowledge
		4198	about the newly added timer. By waking up the loop it will pick up any new
		4199	watchers in the next event loop iteration.
		4200
		4201	=head3 COROUTINES
		4202
		4203	Libev is very accommodating to coroutines ("cooperative threads"):
		4204	libev fully supports nesting calls to its functions from different
		4205	coroutines (e.g. you can call C<ev_loop> on the same loop from two
		4206	different coroutines, and switch freely between both coroutines running
		4207	the loop, as long as you don't confuse yourself). The only exception is
		4208	that you must not do this from C<ev_periodic> reschedule callbacks.
		4209
		4210	Care has been taken to ensure that libev does not keep local state inside
		4211	C<ev_loop>, and other calls do not usually allow for coroutine switches as
		4212	they do not call any callbacks.
		4213
		4214	=head2 COMPILER WARNINGS
		4215
		4216	Depending on your compiler and compiler settings, you might get no or a
		4217	lot of warnings when compiling libev code. Some people are apparently
		4218	scared by this.
		4219
		4220	However, these are unavoidable for many reasons. For one, each compiler
		4221	has different warnings, and each user has different tastes regarding
		4222	warning options. "Warn-free" code therefore cannot be a goal except when
		4223	targeting a specific compiler and compiler-version.
		4224
		4225	Another reason is that some compiler warnings require elaborate
		4226	workarounds, or other changes to the code that make it less clear and less
		4227	maintainable.
		4228
		4229	And of course, some compiler warnings are just plain stupid, or simply
		4230	wrong (because they don't actually warn about the condition their message
		4231	seems to warn about). For example, certain older gcc versions had some
		4232	warnings that resulted an extreme number of false positives. These have
		4233	been fixed, but some people still insist on making code warn-free with
		4234	such buggy versions.
		4235
		4236	While libev is written to generate as few warnings as possible,
		4237	"warn-free" code is not a goal, and it is recommended not to build libev
		4238	with any compiler warnings enabled unless you are prepared to cope with
		4239	them (e.g. by ignoring them). Remember that warnings are just that:
		4240	warnings, not errors, or proof of bugs.
		4241
		4242
		4243	=head2 VALGRIND
		4244
		4245	Valgrind has a special section here because it is a popular tool that is
		4246	highly useful. Unfortunately, valgrind reports are very hard to interpret.
		4247
		4248	If you think you found a bug (memory leak, uninitialised data access etc.)
		4249	in libev, then check twice: If valgrind reports something like:
		4250
		4251	==2274== definitely lost: 0 bytes in 0 blocks.
		4252	==2274== possibly lost: 0 bytes in 0 blocks.
		4253	==2274== still reachable: 256 bytes in 1 blocks.
		4254
		4255	Then there is no memory leak, just as memory accounted to global variables
		4256	is not a memleak - the memory is still being referenced, and didn't leak.
		4257
		4258	Similarly, under some circumstances, valgrind might report kernel bugs
		4259	as if it were a bug in libev (e.g. in realloc or in the poll backend,
		4260	although an acceptable workaround has been found here), or it might be
		4261	confused.
		4262
		4263	Keep in mind that valgrind is a very good tool, but only a tool. Don't
		4264	make it into some kind of religion.
		4265
		4266	If you are unsure about something, feel free to contact the mailing list
		4267	with the full valgrind report and an explanation on why you think this
		4268	is a bug in libev (best check the archives, too :). However, don't be
		4269	annoyed when you get a brisk "this is no bug" answer and take the chance
		4270	of learning how to interpret valgrind properly.
		4271
		4272	If you need, for some reason, empty reports from valgrind for your project
		4273	I suggest using suppression lists.
		4274
		4275
		4276	=head1 PORTABILITY NOTES
		4277
		4278	=head2 WIN32 PLATFORM LIMITATIONS AND WORKAROUNDS
		4279
		4280	Win32 doesn't support any of the standards (e.g. POSIX) that libev
		4281	requires, and its I/O model is fundamentally incompatible with the POSIX
		4282	model. Libev still offers limited functionality on this platform in
		4283	the form of the C<EVBACKEND_SELECT> backend, and only supports socket
		4284	descriptors. This only applies when using Win32 natively, not when using
		4285	e.g. cygwin.
		4286
		4287	Lifting these limitations would basically require the full
		4288	re-implementation of the I/O system. If you are into these kinds of
		4289	things, then note that glib does exactly that for you in a very portable
		4290	way (note also that glib is the slowest event library known to man).
		4291
		4292	There is no supported compilation method available on windows except
		4293	embedding it into other applications.
		4294
		4295	Sensible signal handling is officially unsupported by Microsoft - libev
		4296	tries its best, but under most conditions, signals will simply not work.
		4297
		4298	Not a libev limitation but worth mentioning: windows apparently doesn't
		4299	accept large writes: instead of resulting in a partial write, windows will
		4300	either accept everything or return C<ENOBUFS> if the buffer is too large,
		4301	so make sure you only write small amounts into your sockets (less than a
		4302	megabyte seems safe, but this apparently depends on the amount of memory
		4303	available).
		4304
		4305	Due to the many, low, and arbitrary limits on the win32 platform and
		4306	the abysmal performance of winsockets, using a large number of sockets
		4307	is not recommended (and not reasonable). If your program needs to use
		4308	more than a hundred or so sockets, then likely it needs to use a totally
		4309	different implementation for windows, as libev offers the POSIX readiness
		4310	notification model, which cannot be implemented efficiently on windows
		4311	(due to Microsoft monopoly games).
		4312
		4313	A typical way to use libev under windows is to embed it (see the embedding
		4314	section for details) and use the following F<evwrap.h> header file instead
		4315	of F<ev.h>:
		4316
		4317	#define EV_STANDALONE /* keeps ev from requiring config.h */
		4318	#define EV_SELECT_IS_WINSOCKET 1 /* configure libev for windows select */
		4319
		4320	#include "ev.h"
		4321
		4322	And compile the following F<evwrap.c> file into your project (make sure
		4323	you do I<not> compile the F<ev.c> or any other embedded source files!):
		4324
		4325	#include "evwrap.h"
		4326	#include "ev.c"
		4327
		4328	=over 4
		4329
		4330	=item The winsocket select function
		4331
		4332	The winsocket C<select> function doesn't follow POSIX in that it
		4333	requires socket I<handles> and not socket I<file descriptors> (it is
		4334	also extremely buggy). This makes select very inefficient, and also
		4335	requires a mapping from file descriptors to socket handles (the Microsoft
		4336	C runtime provides the function C<_open_osfhandle> for this). See the
		4337	discussion of the C<EV_SELECT_USE_FD_SET>, C<EV_SELECT_IS_WINSOCKET> and
		4338	C<EV_FD_TO_WIN32_HANDLE> preprocessor symbols for more info.
		4339
		4340	The configuration for a "naked" win32 using the Microsoft runtime
		4341	libraries and raw winsocket select is:
		4342
		4343	#define EV_USE_SELECT 1
		4344	#define EV_SELECT_IS_WINSOCKET 1 /* forces EV_SELECT_USE_FD_SET, too */
		4345
		4346	Note that winsockets handling of fd sets is O(n), so you can easily get a
		4347	complexity in the O(n²) range when using win32.
		4348
		4349	=item Limited number of file descriptors
		4350
		4351	Windows has numerous arbitrary (and low) limits on things.
		4352
		4353	Early versions of winsocket's select only supported waiting for a maximum
		4354	of C<64> handles (probably owning to the fact that all windows kernels
		4355	can only wait for C<64> things at the same time internally; Microsoft
		4356	recommends spawning a chain of threads and wait for 63 handles and the
		4357	previous thread in each. Sounds great!).
		4358
		4359	Newer versions support more handles, but you need to define C<FD_SETSIZE>
		4360	to some high number (e.g. C<2048>) before compiling the winsocket select
		4361	call (which might be in libev or elsewhere, for example, perl and many
		4362	other interpreters do their own select emulation on windows).
		4363
		4364	Another limit is the number of file descriptors in the Microsoft runtime
		4365	libraries, which by default is C<64> (there must be a hidden I<64>
		4366	fetish or something like this inside Microsoft). You can increase this
		4367	by calling C<_setmaxstdio>, which can increase this limit to C<2048>
		4368	(another arbitrary limit), but is broken in many versions of the Microsoft
		4369	runtime libraries. This might get you to about C<512> or C<2048> sockets
		4370	(depending on windows version and/or the phase of the moon). To get more,
		4371	you need to wrap all I/O functions and provide your own fd management, but
		4372	the cost of calling select (O(n²)) will likely make this unworkable.
		4373
		4374	=back
		4375
		4376	=head2 PORTABILITY REQUIREMENTS
		4377
		4378	In addition to a working ISO-C implementation and of course the
		4379	backend-specific APIs, libev relies on a few additional extensions:
		4380
		4381	=over 4
		4382
		4383	=item C<void (*)(ev_watcher_type *, int revents)> must have compatible
		4384	calling conventions regardless of C<ev_watcher_type *>.
		4385
		4386	Libev assumes not only that all watcher pointers have the same internal
		4387	structure (guaranteed by POSIX but not by ISO C for example), but it also
		4388	assumes that the same (machine) code can be used to call any watcher
		4389	callback: The watcher callbacks have different type signatures, but libev
		4390	calls them using an C<ev_watcher *> internally.
		4391
		4392	=item C<sig_atomic_t volatile> must be thread-atomic as well
		4393
		4394	The type C<sig_atomic_t volatile> (or whatever is defined as
		4395	C<EV_ATOMIC_T>) must be atomic with respect to accesses from different
		4396	threads. This is not part of the specification for C<sig_atomic_t>, but is
		4397	believed to be sufficiently portable.
		4398
		4399	=item C<sigprocmask> must work in a threaded environment
		4400
		4401	Libev uses C<sigprocmask> to temporarily block signals. This is not
		4402	allowed in a threaded program (C<pthread_sigmask> has to be used). Typical
		4403	pthread implementations will either allow C<sigprocmask> in the "main
		4404	thread" or will block signals process-wide, both behaviours would
		4405	be compatible with libev. Interaction between C<sigprocmask> and
		4406	C<pthread_sigmask> could complicate things, however.
		4407
		4408	The most portable way to handle signals is to block signals in all threads
		4409	except the initial one, and run the default loop in the initial thread as
		4410	well.
		4411
		4412	=item C<long> must be large enough for common memory allocation sizes
		4413
		4414	To improve portability and simplify its API, libev uses C<long> internally
		4415	instead of C<size_t> when allocating its data structures. On non-POSIX
		4416	systems (Microsoft...) this might be unexpectedly low, but is still at
		4417	least 31 bits everywhere, which is enough for hundreds of millions of
		4418	watchers.
		4419
		4420	=item C<double> must hold a time value in seconds with enough accuracy
		4421
		4422	The type C<double> is used to represent timestamps. It is required to
		4423	have at least 51 bits of mantissa (and 9 bits of exponent), which is good
		4424	enough for at least into the year 4000. This requirement is fulfilled by
		4425	implementations implementing IEEE 754, which is basically all existing
		4426	ones. With IEEE 754 doubles, you get microsecond accuracy until at least
		4427	2200.
		4428
		4429	=back
		4430
		4431	If you know of other additional requirements drop me a note.
		4432
		4433
2445	=head1 COMPLEXITIES	4434	=head1 ALGORITHMIC COMPLEXITIES
2446		4435
2447	In this section the complexities of (many of) the algorithms used inside	4436	In this section the complexities of (many of) the algorithms used inside
2448	libev will be explained. For complexity discussions about backends see the	4437	libev will be documented. For complexity discussions about backends see
2449	documentation for C<ev_default_init>.	4438	the documentation for C<ev_default_init>.
2450		4439
2451	All of the following are about amortised time: If an array needs to be	4440	All of the following are about amortised time: If an array needs to be
2452	extended, libev needs to realloc and move the whole array, but this	4441	extended, libev needs to realloc and move the whole array, but this
2453	happens asymptotically never with higher number of elements, so O(1) might	4442	happens asymptotically rarer with higher number of elements, so O(1) might
2454	mean it might do a lengthy realloc operation in rare cases, but on average	4443	mean that libev does a lengthy realloc operation in rare cases, but on
2455	it is much faster and asymptotically approaches constant time.	4444	average it is much faster and asymptotically approaches constant time.
2456		4445
2457	=over 4	4446	=over 4
2458		4447
2459	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)	4448	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)
2460		4449
2461	This means that, when you have a watcher that triggers in one hour and	4450	This means that, when you have a watcher that triggers in one hour and
2462	there are 100 watchers that would trigger before that then inserting will	4451	there are 100 watchers that would trigger before that, then inserting will
2463	have to skip those 100 watchers.	4452	have to skip roughly seven (C<ld 100>) of these watchers.
2464		4453
2465	=item Changing timer/periodic watchers (by autorepeat, again): O(log skipped_other_timers)	4454	=item Changing timer/periodic watchers (by autorepeat or calling again): O(log skipped_other_timers)
2466		4455
2467	That means that for changing a timer costs less than removing/adding them	4456	That means that changing a timer costs less than removing/adding them,
2468	as only the relative motion in the event queue has to be paid for.	4457	as only the relative motion in the event queue has to be paid for.
2469		4458
2470	=item Starting io/check/prepare/idle/signal/child watchers: O(1)	4459	=item Starting io/check/prepare/idle/signal/child/fork/async watchers: O(1)
2471		4460
2472	These just add the watcher into an array or at the head of a list.	4461	These just add the watcher into an array or at the head of a list.
		4462
2473	=item Stopping check/prepare/idle watchers: O(1)	4463	=item Stopping check/prepare/idle/fork/async watchers: O(1)
2474		4464
2475	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))	4465	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))
2476		4466
2477	These watchers are stored in lists then need to be walked to find the	4467	These watchers are stored in lists, so they need to be walked to find the
2478	correct watcher to remove. The lists are usually short (you don't usually	4468	correct watcher to remove. The lists are usually short (you don't usually
2479	have many watchers waiting for the same fd or signal).	4469	have many watchers waiting for the same fd or signal: one is typical, two
		4470	is rare).
2480		4471
2481	=item Finding the next timer per loop iteration: O(1)	4472	=item Finding the next timer in each loop iteration: O(1)
		4473
		4474	By virtue of using a binary or 4-heap, the next timer is always found at a
		4475	fixed position in the storage array.
2482		4476
2483	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)	4477	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)
2484		4478
2485	A change means an I/O watcher gets started or stopped, which requires	4479	A change means an I/O watcher gets started or stopped, which requires
2486	libev to recalculate its status (and possibly tell the kernel).	4480	libev to recalculate its status (and possibly tell the kernel, depending
		4481	on backend and whether C<ev_io_set> was used).
2487		4482
2488	=item Activating one watcher: O(1)	4483	=item Activating one watcher (putting it into the pending state): O(1)
2489		4484
2490	=item Priority handling: O(number_of_priorities)	4485	=item Priority handling: O(number_of_priorities)
2491		4486
2492	Priorities are implemented by allocating some space for each	4487	Priorities are implemented by allocating some space for each
2493	priority. When doing priority-based operations, libev usually has to	4488	priority. When doing priority-based operations, libev usually has to
2494	linearly search all the priorities.	4489	linearly search all the priorities, but starting/stopping and activating
		4490	watchers becomes O(1) with respect to priority handling.
		4491
		4492	=item Sending an ev_async: O(1)
		4493
		4494	=item Processing ev_async_send: O(number_of_async_watchers)
		4495
		4496	=item Processing signals: O(max_signal_number)
		4497
		4498	Sending involves a system call I<iff> there were no other C<ev_async_send>
		4499	calls in the current loop iteration. Checking for async and signal events
		4500	involves iterating over all running async watchers or all signal numbers.
2495		4501
2496	=back	4502	=back
2497		4503
2498		4504
		4505	=head1 GLOSSARY
		4506
		4507	=over 4
		4508
		4509	=item active
		4510
		4511	A watcher is active as long as it has been started (has been attached to
		4512	an event loop) but not yet stopped (disassociated from the event loop).
		4513
		4514	=item application
		4515
		4516	In this document, an application is whatever is using libev.
		4517
		4518	=item callback
		4519
		4520	The address of a function that is called when some event has been
		4521	detected. Callbacks are being passed the event loop, the watcher that
		4522	received the event, and the actual event bitset.
		4523
		4524	=item callback invocation
		4525
		4526	The act of calling the callback associated with a watcher.
		4527
		4528	=item event
		4529
		4530	A change of state of some external event, such as data now being available
		4531	for reading on a file descriptor, time having passed or simply not having
		4532	any other events happening anymore.
		4533
		4534	In libev, events are represented as single bits (such as C<EV_READ> or
		4535	C<EV_TIMEOUT>).
		4536
		4537	=item event library
		4538
		4539	A software package implementing an event model and loop.
		4540
		4541	=item event loop
		4542
		4543	An entity that handles and processes external events and converts them
		4544	into callback invocations.
		4545
		4546	=item event model
		4547
		4548	The model used to describe how an event loop handles and processes
		4549	watchers and events.
		4550
		4551	=item pending
		4552
		4553	A watcher is pending as soon as the corresponding event has been detected,
		4554	and stops being pending as soon as the watcher will be invoked or its
		4555	pending status is explicitly cleared by the application.
		4556
		4557	A watcher can be pending, but not active. Stopping a watcher also clears
		4558	its pending status.
		4559
		4560	=item real time
		4561
		4562	The physical time that is observed. It is apparently strictly monotonic :)
		4563
		4564	=item wall-clock time
		4565
		4566	The time and date as shown on clocks. Unlike real time, it can actually
		4567	be wrong and jump forwards and backwards, e.g. when the you adjust your
		4568	clock.
		4569
		4570	=item watcher
		4571
		4572	A data structure that describes interest in certain events. Watchers need
		4573	to be started (attached to an event loop) before they can receive events.
		4574
		4575	=item watcher invocation
		4576
		4577	The act of calling the callback associated with a watcher.
		4578
		4579	=back
		4580
2499	=head1 AUTHOR	4581	=head1 AUTHOR
2500		4582
2501	Marc Lehmann <libev@schmorp.de>.	4583	Marc Lehmann <libev@schmorp.de>, with repeated corrections by Mikael Magnusson.
2502		4584

Diff Legend

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