ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/libev/ev.pod

Comparing libev/ev.pod (file contents):
Revision 1.108 by root, Mon Dec 24 10:39:21 2007 UTC vs.
Revision 1.273 by root, Tue Nov 24 14:46:59 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	=head2 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 occurring), 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
…		…
70	=head2 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	=head2 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	=head2 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 some floatingpoint value. Unlike the name	133	on it, you should treat it as some floating point value. Unlike the name
104	component C<stamp> might indicate, it is also used for time differences	134	component C<stamp> might indicate, it is also used for time differences
105	throughout libev.	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
106		157
107	=head1 GLOBAL FUNCTIONS	158	=head1 GLOBAL FUNCTIONS
108		159
109	These functions can be called anytime, even before initialising the	160	These functions can be called anytime, even before initialising the
110	library in any way.	161	library in any way.
…		…
119		170
120	=item ev_sleep (ev_tstamp interval)	171	=item ev_sleep (ev_tstamp interval)
121		172
122	Sleep for the given interval: The current thread will be blocked until	173	Sleep for the given interval: The current thread will be blocked until
123	either it is interrupted or the given time interval has passed. Basically	174	either it is interrupted or the given time interval has passed. Basically
124	this is a subsecond-resolution C<sleep ()>.	175	this is a sub-second-resolution C<sleep ()>.
125		176
126	=item int ev_version_major ()	177	=item int ev_version_major ()
127		178
128	=item int ev_version_minor ()	179	=item int ev_version_minor ()
129		180
…		…
142	not a problem.	193	not a problem.
143		194
144	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
145	version.	196	version.
146		197
147	assert (("libev version mismatch",	198	assert (("libev version mismatch",
148	ev_version_major () == EV_VERSION_MAJOR	199	ev_version_major () == EV_VERSION_MAJOR
149	&& ev_version_minor () >= EV_VERSION_MINOR));	200	&& ev_version_minor () >= EV_VERSION_MINOR));
150		201
151	=item unsigned int ev_supported_backends ()	202	=item unsigned int ev_supported_backends ()
152		203
153	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_*>
154	value) compiled into this binary of libev (independent of their	205	value) compiled into this binary of libev (independent of their
…		…
156	a description of the set values.	207	a description of the set values.
157		208
158	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
159	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
160		211
161	assert (("sorry, no epoll, no sex",	212	assert (("sorry, no epoll, no sex",
162	ev_supported_backends () & EVBACKEND_EPOLL));	213	ev_supported_backends () & EVBACKEND_EPOLL));
163		214
164	=item unsigned int ev_recommended_backends ()	215	=item unsigned int ev_recommended_backends ()
165		216
166	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
167	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
168	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
169	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
170	(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
171	libev will probe for if you specify no backends explicitly.	222	libev will probe for if you specify no backends explicitly.
172		223
173	=item unsigned int ev_embeddable_backends ()	224	=item unsigned int ev_embeddable_backends ()
174		225
…		…
178	C<ev_embeddable_backends () & ev_supported_backends ()>, likewise for	229	C<ev_embeddable_backends () & ev_supported_backends ()>, likewise for
179	recommended ones.	230	recommended ones.
180		231
181	See the description of C<ev_embed> watchers for more info.	232	See the description of C<ev_embed> watchers for more info.
182		233
183	=item ev_set_allocator (void *(*cb)(void *ptr, long size))	234	=item ev_set_allocator (void *(*cb)(void *ptr, long size)) [NOT REENTRANT]
184		235
185	Sets the allocation function to use (the prototype is similar - the	236	Sets the allocation function to use (the prototype is similar - the
186	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
187	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
188	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
189	potentially destructive action. The default is your system realloc	240	or take some potentially destructive action.
190	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.
191		245
192	You could override this function in high-availability programs to, say,	246	You could override this function in high-availability programs to, say,
193	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,
194	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.
195		249
196	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
197	retries).	251	retries (example requires a standards-compliant C<realloc>).
198		252
199	static void *	253	static void *
200	persistent_realloc (void *ptr, size_t size)	254	persistent_realloc (void *ptr, size_t size)
201	{	255	{
202	for (;;)	256	for (;;)
…		…
211	}	265	}
212		266
213	...	267	...
214	ev_set_allocator (persistent_realloc);	268	ev_set_allocator (persistent_realloc);
215		269
216	=item ev_set_syserr_cb (void (*cb)(const char *msg));	270	=item ev_set_syserr_cb (void (*cb)(const char *msg)); [NOT REENTRANT]
217		271
218	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
219	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
220	indicating the system call or subsystem causing the problem. If this	274	indicating the system call or subsystem causing the problem. If this
221	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
222	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
223	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
224	(such as abort).	278	(such as abort).
225		279
226	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.
…		…
237		291
238	=back	292	=back
239		293
240	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP	294	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP
241		295
242	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>
243	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>
244	events, and dynamically created loops which do not.	298	I<function>).
245		299
246	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
247	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
248	create, you also create another event loop. Libev itself does no locking	302	not.
249	whatsoever, so if you mix calls to the same event loop in different
250	threads, make sure you lock (this is usually a bad idea, though, even if
251	done correctly, because it's hideous and inefficient).
252		303
253	=over 4	304	=over 4
254		305
255	=item struct ev_loop *ev_default_loop (unsigned int flags)	306	=item struct ev_loop *ev_default_loop (unsigned int flags)
256		307
…		…
260	flags. If that is troubling you, check C<ev_backend ()> afterwards).	311	flags. If that is troubling you, check C<ev_backend ()> afterwards).
261		312
262	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
263	function.	314	function.
264		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
265	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
266	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>).
267		329
268	The following flags are supported:	330	The following flags are supported:
269		331
…		…
274	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
275	thing, believe me).	337	thing, believe me).
276		338
277	=item C<EVFLAG_NOENV>	339	=item C<EVFLAG_NOENV>
278		340
279	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
280	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
281	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will	343	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will
282	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
283	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
284	around bugs.	346	around bugs.
…		…
290	enabling this flag.	352	enabling this flag.
291		353
292	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,
293	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
294	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
295	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
296	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
297	C<pthread_atfork> which is even faster).	359	C<pthread_atfork> which is even faster).
298		360
299	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
300	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
301	flag.	363	flag.
302		364
303	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>
304	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_NOSIGFD>
		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.
305		382
306	=item C<EVBACKEND_SELECT> (value 1, portable select backend)	383	=item C<EVBACKEND_SELECT> (value 1, portable select backend)
307		384
308	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
309	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,
310	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
311	using this backend. It doesn't scale too well (O(highest_fd)), but its	388	using this backend. It doesn't scale too well (O(highest_fd)), but its
312	usually the fastest backend for a low number of (low-numbered :) fds.	389	usually the fastest backend for a low number of (low-numbered :) fds.
313		390
314	To get good performance out of this backend you need a high amount of	391	To get good performance out of this backend you need a high amount of
315	parallelity (most of the file descriptors should be busy). If you are	392	parallelism (most of the file descriptors should be busy). If you are
316	writing a server, you should C<accept ()> in a loop to accept as many	393	writing a server, you should C<accept ()> in a loop to accept as many
317	connections as possible during one iteration. You might also want to have	394	connections as possible during one iteration. You might also want to have
318	a look at C<ev_set_io_collect_interval ()> to increase the amount of	395	a look at C<ev_set_io_collect_interval ()> to increase the amount of
319	readyness notifications you get per iteration.	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).
320		401
321	=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)
322		403
323	And this is your standard poll(2) backend. It's more complicated	404	And this is your standard poll(2) backend. It's more complicated
324	than select, but handles sparse fds better and has no artificial	405	than select, but handles sparse fds better and has no artificial
325	limit on the number of fds you can use (except it will slow down	406	limit on the number of fds you can use (except it will slow down
326	considerably with a lot of inactive fds). It scales similarly to select,	407	considerably with a lot of inactive fds). It scales similarly to select,
327	i.e. O(total_fds). See the entry for C<EVBACKEND_SELECT>, above, for	408	i.e. O(total_fds). See the entry for C<EVBACKEND_SELECT>, above, for
328	performance tips.	409	performance tips.
329		410
		411	This backend maps C<EV_READ> to C<POLLIN | POLLERR | POLLHUP>, and
		412	C<EV_WRITE> to C<POLLOUT | POLLERR | POLLHUP>.
		413
330	=item C<EVBACKEND_EPOLL> (value 4, Linux)	414	=item C<EVBACKEND_EPOLL> (value 4, Linux)
		415
		416	Use the linux-specific epoll(7) interface (for both pre- and post-2.6.9
		417	kernels).
331		418
332	For few fds, this backend is a bit little slower than poll and select,	419	For few fds, this backend is a bit little slower than poll and select,
333	but it scales phenomenally better. While poll and select usually scale	420	but it scales phenomenally better. While poll and select usually scale
334	like O(total_fds) where n is the total number of fds (or the highest fd),	421	like O(total_fds) where n is the total number of fds (or the highest fd),
335	epoll scales either O(1) or O(active_fds). The epoll design has a number	422	epoll scales either O(1) or O(active_fds).
336	of shortcomings, such as silently dropping events in some hard-to-detect	423
337	cases and rewiring a syscall per fd change, no fork support and bad	424	The epoll mechanism deserves honorable mention as the most misdesigned
338	support for dup.	425	of the more advanced event mechanisms: mere annoyances include silently
		426	dropping file descriptors, requiring a system call per change per file
		427	descriptor (and unnecessary guessing of parameters), problems with dup and
		428	so on. The biggest issue is fork races, however - if a program forks then
		429	I<both> parent and child process have to recreate the epoll set, which can
		430	take considerable time (one syscall per file descriptor) and is of course
		431	hard to detect.
		432
		433	Epoll is also notoriously buggy - embedding epoll fds I<should> work, but
		434	of course I<doesn't>, and epoll just loves to report events for totally
		435	I<different> file descriptors (even already closed ones, so one cannot
		436	even remove them from the set) than registered in the set (especially
		437	on SMP systems). Libev tries to counter these spurious notifications by
		438	employing an additional generation counter and comparing that against the
		439	events to filter out spurious ones, recreating the set when required.
339		440
340	While stopping, setting and starting an I/O watcher in the same iteration	441	While stopping, setting and starting an I/O watcher in the same iteration
341	will result in some caching, there is still a syscall per such incident	442	will result in some caching, there is still a system call per such
342	(because the fd could point to a different file description now), so its	443	incident (because the same I<file descriptor> could point to a different
343	best to avoid that. Also, C<dup ()>'ed file descriptors might not work	444	I<file description> now), so its best to avoid that. Also, C<dup ()>'ed
344	very well if you register events for both fds.	445	file descriptors might not work very well if you register events for both
345		446	file descriptors.
346	Please note that epoll sometimes generates spurious notifications, so you
347	need to use non-blocking I/O or other means to avoid blocking when no data
348	(or space) is available.
349		447
350	Best performance from this backend is achieved by not unregistering all	448	Best performance from this backend is achieved by not unregistering all
351	watchers for a file descriptor until it has been closed, if possible, i.e.	449	watchers for a file descriptor until it has been closed, if possible,
352	keep at least one watcher active per fd at all times.	450	i.e. keep at least one watcher active per fd at all times. Stopping and
		451	starting a watcher (without re-setting it) also usually doesn't cause
		452	extra overhead. A fork can both result in spurious notifications as well
		453	as in libev having to destroy and recreate the epoll object, which can
		454	take considerable time and thus should be avoided.
353		455
		456	All this means that, in practice, C<EVBACKEND_SELECT> can be as fast or
		457	faster than epoll for maybe up to a hundred file descriptors, depending on
		458	the usage. So sad.
		459
354	While nominally embeddeble in other event loops, this feature is broken in	460	While nominally embeddable in other event loops, this feature is broken in
355	all kernel versions tested so far.	461	all kernel versions tested so far.
		462
		463	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
		464	C<EVBACKEND_POLL>.
356		465
357	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)	466	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)
358		467
359	Kqueue deserves special mention, as at the time of this writing, it	468	Kqueue deserves special mention, as at the time of this writing, it
360	was broken on all BSDs except NetBSD (usually it doesn't work reliably	469	was broken on all BSDs except NetBSD (usually it doesn't work reliably
361	with anything but sockets and pipes, except on Darwin, where of course	470	with anything but sockets and pipes, except on Darwin, where of course
362	it's completely useless). For this reason it's not being "autodetected"	471	it's completely useless). Unlike epoll, however, whose brokenness
		472	is by design, these kqueue bugs can (and eventually will) be fixed
		473	without API changes to existing programs. For this reason it's not being
363	unless you explicitly specify it explicitly in the flags (i.e. using	474	"auto-detected" unless you explicitly specify it in the flags (i.e. using
364	C<EVBACKEND_KQUEUE>) or libev was compiled on a known-to-be-good (-enough)	475	C<EVBACKEND_KQUEUE>) or libev was compiled on a known-to-be-good (-enough)
365	system like NetBSD.	476	system like NetBSD.
366		477
367	You still can embed kqueue into a normal poll or select backend and use it	478	You still can embed kqueue into a normal poll or select backend and use it
368	only for sockets (after having made sure that sockets work with kqueue on	479	only for sockets (after having made sure that sockets work with kqueue on
369	the target platform). See C<ev_embed> watchers for more info.	480	the target platform). See C<ev_embed> watchers for more info.
370		481
371	It scales in the same way as the epoll backend, but the interface to the	482	It scales in the same way as the epoll backend, but the interface to the
372	kernel is more efficient (which says nothing about its actual speed, of	483	kernel is more efficient (which says nothing about its actual speed, of
373	course). While stopping, setting and starting an I/O watcher does never	484	course). While stopping, setting and starting an I/O watcher does never
374	cause an extra syscall as with C<EVBACKEND_EPOLL>, it still adds up to	485	cause an extra system call as with C<EVBACKEND_EPOLL>, it still adds up to
375	two event changes per incident, support for C<fork ()> is very bad and it	486	two event changes per incident. Support for C<fork ()> is very bad (but
376	drops fds silently in similarly hard-to-detect cases.	487	sane, unlike epoll) and it drops fds silently in similarly hard-to-detect
		488	cases
377		489
378	This backend usually performs well under most conditions.	490	This backend usually performs well under most conditions.
379		491
380	While nominally embeddable in other event loops, this doesn't work	492	While nominally embeddable in other event loops, this doesn't work
381	everywhere, so you might need to test for this. And since it is broken	493	everywhere, so you might need to test for this. And since it is broken
382	almost everywhere, you should only use it when you have a lot of sockets	494	almost everywhere, you should only use it when you have a lot of sockets
383	(for which it usually works), by embedding it into another event loop	495	(for which it usually works), by embedding it into another event loop
384	(e.g. C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>) and using it only for	496	(e.g. C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> (but C<poll> is of course
385	sockets.	497	also broken on OS X)) and, did I mention it, using it only for sockets.
		498
		499	This backend maps C<EV_READ> into an C<EVFILT_READ> kevent with
		500	C<NOTE_EOF>, and C<EV_WRITE> into an C<EVFILT_WRITE> kevent with
		501	C<NOTE_EOF>.
386		502
387	=item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8)	503	=item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8)
388		504
389	This is not implemented yet (and might never be, unless you send me an	505	This is not implemented yet (and might never be, unless you send me an
390	implementation). According to reports, C</dev/poll> only supports sockets	506	implementation). According to reports, C</dev/poll> only supports sockets
…		…
394	=item C<EVBACKEND_PORT> (value 32, Solaris 10)	510	=item C<EVBACKEND_PORT> (value 32, Solaris 10)
395		511
396	This uses the Solaris 10 event port mechanism. As with everything on Solaris,	512	This uses the Solaris 10 event port mechanism. As with everything on Solaris,
397	it's really slow, but it still scales very well (O(active_fds)).	513	it's really slow, but it still scales very well (O(active_fds)).
398		514
399	Please note that solaris event ports can deliver a lot of spurious	515	Please note that Solaris event ports can deliver a lot of spurious
400	notifications, so you need to use non-blocking I/O or other means to avoid	516	notifications, so you need to use non-blocking I/O or other means to avoid
401	blocking when no data (or space) is available.	517	blocking when no data (or space) is available.
402		518
403	While this backend scales well, it requires one system call per active	519	While this backend scales well, it requires one system call per active
404	file descriptor per loop iteration. For small and medium numbers of file	520	file descriptor per loop iteration. For small and medium numbers of file
405	descriptors a "slow" C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> backend	521	descriptors a "slow" C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> backend
406	might perform better.	522	might perform better.
407		523
		524	On the positive side, with the exception of the spurious readiness
		525	notifications, this backend actually performed fully to specification
		526	in all tests and is fully embeddable, which is a rare feat among the
		527	OS-specific backends (I vastly prefer correctness over speed hacks).
		528
		529	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
		530	C<EVBACKEND_POLL>.
		531
408	=item C<EVBACKEND_ALL>	532	=item C<EVBACKEND_ALL>
409		533
410	Try all backends (even potentially broken ones that wouldn't be tried	534	Try all backends (even potentially broken ones that wouldn't be tried
411	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as	535	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as
412	C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>.	536	C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>.
413		537
414	It is definitely not recommended to use this flag.	538	It is definitely not recommended to use this flag.
415		539
416	=back	540	=back
417		541
418	If one or more of these are ored into the flags value, then only these	542	If one or more of the backend flags are or'ed into the flags value,
419	backends will be tried (in the reverse order as given here). If none are	543	then only these backends will be tried (in the reverse order as listed
420	specified, most compiled-in backend will be tried, usually in reverse	544	here). If none are specified, all backends in C<ev_recommended_backends
421	order of their flag values :)	545	()> will be tried.
422		546
423	The most typical usage is like this:	547	Example: This is the most typical usage.
424		548
425	if (!ev_default_loop (0))	549	if (!ev_default_loop (0))
426	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");	550	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
427		551
428	Restrict libev to the select and poll backends, and do not allow	552	Example: Restrict libev to the select and poll backends, and do not allow
429	environment settings to be taken into account:	553	environment settings to be taken into account:
430		554
431	ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);	555	ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);
432		556
433	Use whatever libev has to offer, but make sure that kqueue is used if	557	Example: Use whatever libev has to offer, but make sure that kqueue is
434	available (warning, breaks stuff, best use only with your own private	558	used if available (warning, breaks stuff, best use only with your own
435	event loop and only if you know the OS supports your types of fds):	559	private event loop and only if you know the OS supports your types of
		560	fds):
436		561
437	ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);	562	ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);
438		563
439	=item struct ev_loop *ev_loop_new (unsigned int flags)	564	=item struct ev_loop *ev_loop_new (unsigned int flags)
440		565
441	Similar to C<ev_default_loop>, but always creates a new event loop that is	566	Similar to C<ev_default_loop>, but always creates a new event loop that is
442	always distinct from the default loop. Unlike the default loop, it cannot	567	always distinct from the default loop. Unlike the default loop, it cannot
443	handle signal and child watchers, and attempts to do so will be greeted by	568	handle signal and child watchers, and attempts to do so will be greeted by
444	undefined behaviour (or a failed assertion if assertions are enabled).	569	undefined behaviour (or a failed assertion if assertions are enabled).
445		570
		571	Note that this function I<is> thread-safe, and the recommended way to use
		572	libev with threads is indeed to create one loop per thread, and using the
		573	default loop in the "main" or "initial" thread.
		574
446	Example: Try to create a event loop that uses epoll and nothing else.	575	Example: Try to create a event loop that uses epoll and nothing else.
447		576
448	struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);	577	struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);
449	if (!epoller)	578	if (!epoller)
450	fatal ("no epoll found here, maybe it hides under your chair");	579	fatal ("no epoll found here, maybe it hides under your chair");
451		580
452	=item ev_default_destroy ()	581	=item ev_default_destroy ()
453		582
454	Destroys the default loop again (frees all memory and kernel state	583	Destroys the default loop again (frees all memory and kernel state
455	etc.). None of the active event watchers will be stopped in the normal	584	etc.). None of the active event watchers will be stopped in the normal
456	sense, so e.g. C<ev_is_active> might still return true. It is your	585	sense, so e.g. C<ev_is_active> might still return true. It is your
457	responsibility to either stop all watchers cleanly yoursef I<before>	586	responsibility to either stop all watchers cleanly yourself I<before>
458	calling this function, or cope with the fact afterwards (which is usually	587	calling this function, or cope with the fact afterwards (which is usually
459	the easiest thing, you can just ignore the watchers and/or C<free ()> them	588	the easiest thing, you can just ignore the watchers and/or C<free ()> them
460	for example).	589	for example).
461		590
462	Note that certain global state, such as signal state, will not be freed by	591	Note that certain global state, such as signal state (and installed signal
463	this function, and related watchers (such as signal and child watchers)	592	handlers), will not be freed by this function, and related watchers (such
464	would need to be stopped manually.	593	as signal and child watchers) would need to be stopped manually.
465		594
466	In general it is not advisable to call this function except in the	595	In general it is not advisable to call this function except in the
467	rare occasion where you really need to free e.g. the signal handling	596	rare occasion where you really need to free e.g. the signal handling
468	pipe fds. If you need dynamically allocated loops it is better to use	597	pipe fds. If you need dynamically allocated loops it is better to use
469	C<ev_loop_new> and C<ev_loop_destroy>).	598	C<ev_loop_new> and C<ev_loop_destroy>.
470		599
471	=item ev_loop_destroy (loop)	600	=item ev_loop_destroy (loop)
472		601
473	Like C<ev_default_destroy>, but destroys an event loop created by an	602	Like C<ev_default_destroy>, but destroys an event loop created by an
474	earlier call to C<ev_loop_new>.	603	earlier call to C<ev_loop_new>.
475		604
476	=item ev_default_fork ()	605	=item ev_default_fork ()
477		606
		607	This function sets a flag that causes subsequent C<ev_loop> iterations
478	This function reinitialises the kernel state for backends that have	608	to reinitialise the kernel state for backends that have one. Despite the
479	one. Despite the name, you can call it anytime, but it makes most sense	609	name, you can call it anytime, but it makes most sense after forking, in
480	after forking, in either the parent or child process (or both, but that	610	the child process (or both child and parent, but that again makes little
481	again makes little sense).	611	sense). You I<must> call it in the child before using any of the libev
		612	functions, and it will only take effect at the next C<ev_loop> iteration.
482		613
483	You I<must> call this function in the child process after forking if and	614	On the other hand, you only need to call this function in the child
484	only if you want to use the event library in both processes. If you just	615	process if and only if you want to use the event library in the child. If
485	fork+exec, you don't have to call it.	616	you just fork+exec, you don't have to call it at all.
486		617
487	The function itself is quite fast and it's usually not a problem to call	618	The function itself is quite fast and it's usually not a problem to call
488	it just in case after a fork. To make this easy, the function will fit in	619	it just in case after a fork. To make this easy, the function will fit in
489	quite nicely into a call to C<pthread_atfork>:	620	quite nicely into a call to C<pthread_atfork>:
490		621
491	pthread_atfork (0, 0, ev_default_fork);	622	pthread_atfork (0, 0, ev_default_fork);
492		623
493	At the moment, C<EVBACKEND_SELECT> and C<EVBACKEND_POLL> are safe to use
494	without calling this function, so if you force one of those backends you
495	do not need to care.
496
497	=item ev_loop_fork (loop)	624	=item ev_loop_fork (loop)
498		625
499	Like C<ev_default_fork>, but acts on an event loop created by	626	Like C<ev_default_fork>, but acts on an event loop created by
500	C<ev_loop_new>. Yes, you have to call this on every allocated event loop	627	C<ev_loop_new>. Yes, you have to call this on every allocated event loop
501	after fork, and how you do this is entirely your own problem.	628	after fork that you want to re-use in the child, and how you do this is
		629	entirely your own problem.
		630
		631	=item int ev_is_default_loop (loop)
		632
		633	Returns true when the given loop is, in fact, the default loop, and false
		634	otherwise.
502		635
503	=item unsigned int ev_loop_count (loop)	636	=item unsigned int ev_loop_count (loop)
504		637
505	Returns the count of loop iterations for the loop, which is identical to	638	Returns the count of loop iterations for the loop, which is identical to
506	the number of times libev did poll for new events. It starts at C<0> and	639	the number of times libev did poll for new events. It starts at C<0> and
507	happily wraps around with enough iterations.	640	happily wraps around with enough iterations.
508		641
509	This value can sometimes be useful as a generation counter of sorts (it	642	This value can sometimes be useful as a generation counter of sorts (it
510	"ticks" the number of loop iterations), as it roughly corresponds with	643	"ticks" the number of loop iterations), as it roughly corresponds with
511	C<ev_prepare> and C<ev_check> calls.	644	C<ev_prepare> and C<ev_check> calls.
		645
		646	=item unsigned int ev_loop_depth (loop)
		647
		648	Returns the number of times C<ev_loop> was entered minus the number of
		649	times C<ev_loop> was exited, in other words, the recursion depth.
		650
		651	Outside C<ev_loop>, this number is zero. In a callback, this number is
		652	C<1>, unless C<ev_loop> was invoked recursively (or from another thread),
		653	in which case it is higher.
		654
		655	Leaving C<ev_loop> abnormally (setjmp/longjmp, cancelling the thread
		656	etc.), doesn't count as exit.
512		657
513	=item unsigned int ev_backend (loop)	658	=item unsigned int ev_backend (loop)
514		659
515	Returns one of the C<EVBACKEND_*> flags indicating the event backend in	660	Returns one of the C<EVBACKEND_*> flags indicating the event backend in
516	use.	661	use.
…		…
521	received events and started processing them. This timestamp does not	666	received events and started processing them. This timestamp does not
522	change as long as callbacks are being processed, and this is also the base	667	change as long as callbacks are being processed, and this is also the base
523	time used for relative timers. You can treat it as the timestamp of the	668	time used for relative timers. You can treat it as the timestamp of the
524	event occurring (or more correctly, libev finding out about it).	669	event occurring (or more correctly, libev finding out about it).
525		670
		671	=item ev_now_update (loop)
		672
		673	Establishes the current time by querying the kernel, updating the time
		674	returned by C<ev_now ()> in the progress. This is a costly operation and
		675	is usually done automatically within C<ev_loop ()>.
		676
		677	This function is rarely useful, but when some event callback runs for a
		678	very long time without entering the event loop, updating libev's idea of
		679	the current time is a good idea.
		680
		681	See also L<The special problem of time updates> in the C<ev_timer> section.
		682
		683	=item ev_suspend (loop)
		684
		685	=item ev_resume (loop)
		686
		687	These two functions suspend and resume a loop, for use when the loop is
		688	not used for a while and timeouts should not be processed.
		689
		690	A typical use case would be an interactive program such as a game: When
		691	the user presses C<^Z> to suspend the game and resumes it an hour later it
		692	would be best to handle timeouts as if no time had actually passed while
		693	the program was suspended. This can be achieved by calling C<ev_suspend>
		694	in your C<SIGTSTP> handler, sending yourself a C<SIGSTOP> and calling
		695	C<ev_resume> directly afterwards to resume timer processing.
		696
		697	Effectively, all C<ev_timer> watchers will be delayed by the time spend
		698	between C<ev_suspend> and C<ev_resume>, and all C<ev_periodic> watchers
		699	will be rescheduled (that is, they will lose any events that would have
		700	occured while suspended).
		701
		702	After calling C<ev_suspend> you B<must not> call I<any> function on the
		703	given loop other than C<ev_resume>, and you B<must not> call C<ev_resume>
		704	without a previous call to C<ev_suspend>.
		705
		706	Calling C<ev_suspend>/C<ev_resume> has the side effect of updating the
		707	event loop time (see C<ev_now_update>).
		708
526	=item ev_loop (loop, int flags)	709	=item ev_loop (loop, int flags)
527		710
528	Finally, this is it, the event handler. This function usually is called	711	Finally, this is it, the event handler. This function usually is called
529	after you initialised all your watchers and you want to start handling	712	after you have initialised all your watchers and you want to start
530	events.	713	handling events.
531		714
532	If the flags argument is specified as C<0>, it will not return until	715	If the flags argument is specified as C<0>, it will not return until
533	either no event watchers are active anymore or C<ev_unloop> was called.	716	either no event watchers are active anymore or C<ev_unloop> was called.
534		717
535	Please note that an explicit C<ev_unloop> is usually better than	718	Please note that an explicit C<ev_unloop> is usually better than
536	relying on all watchers to be stopped when deciding when a program has	719	relying on all watchers to be stopped when deciding when a program has
537	finished (especially in interactive programs), but having a program that	720	finished (especially in interactive programs), but having a program
538	automatically loops as long as it has to and no longer by virtue of	721	that automatically loops as long as it has to and no longer by virtue
539	relying on its watchers stopping correctly is a thing of beauty.	722	of relying on its watchers stopping correctly, that is truly a thing of
		723	beauty.
540		724
541	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle	725	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle
542	those events and any outstanding ones, but will not block your process in	726	those events and any already outstanding ones, but will not block your
543	case there are no events and will return after one iteration of the loop.	727	process in case there are no events and will return after one iteration of
		728	the loop.
544		729
545	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if	730	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if
546	neccessary) and will handle those and any outstanding ones. It will block	731	necessary) and will handle those and any already outstanding ones. It
547	your process until at least one new event arrives, and will return after	732	will block your process until at least one new event arrives (which could
548	one iteration of the loop. This is useful if you are waiting for some	733	be an event internal to libev itself, so there is no guarantee that a
549	external event in conjunction with something not expressible using other	734	user-registered callback will be called), and will return after one
		735	iteration of the loop.
		736
		737	This is useful if you are waiting for some external event in conjunction
		738	with something not expressible using other libev watchers (i.e. "roll your
550	libev watchers. However, a pair of C<ev_prepare>/C<ev_check> watchers is	739	own C<ev_loop>"). However, a pair of C<ev_prepare>/C<ev_check> watchers is
551	usually a better approach for this kind of thing.	740	usually a better approach for this kind of thing.
552		741
553	Here are the gory details of what C<ev_loop> does:	742	Here are the gory details of what C<ev_loop> does:
554		743
555	- Before the first iteration, call any pending watchers.	744	- Before the first iteration, call any pending watchers.
556	* If there are no active watchers (reference count is zero), return.	745	* If EVFLAG_FORKCHECK was used, check for a fork.
557	- Queue all prepare watchers and then call all outstanding watchers.	746	- If a fork was detected (by any means), queue and call all fork watchers.
		747	- Queue and call all prepare watchers.
558	- If we have been forked, recreate the kernel state.	748	- If we have been forked, detach and recreate the kernel state
		749	as to not disturb the other process.
559	- Update the kernel state with all outstanding changes.	750	- Update the kernel state with all outstanding changes.
560	- Update the "event loop time".	751	- Update the "event loop time" (ev_now ()).
561	- Calculate for how long to block.	752	- Calculate for how long to sleep or block, if at all
		753	(active idle watchers, EVLOOP_NONBLOCK or not having
		754	any active watchers at all will result in not sleeping).
		755	- Sleep if the I/O and timer collect interval say so.
562	- Block the process, waiting for any events.	756	- Block the process, waiting for any events.
563	- Queue all outstanding I/O (fd) events.	757	- Queue all outstanding I/O (fd) events.
564	- Update the "event loop time" and do time jump handling.	758	- Update the "event loop time" (ev_now ()), and do time jump adjustments.
565	- Queue all outstanding timers.	759	- Queue all expired timers.
566	- Queue all outstanding periodics.	760	- Queue all expired periodics.
567	- If no events are pending now, queue all idle watchers.	761	- Unless any events are pending now, queue all idle watchers.
568	- Queue all check watchers.	762	- Queue all check watchers.
569	- Call all queued watchers in reverse order (i.e. check watchers first).	763	- Call all queued watchers in reverse order (i.e. check watchers first).
570	Signals and child watchers are implemented as I/O watchers, and will	764	Signals and child watchers are implemented as I/O watchers, and will
571	be handled here by queueing them when their watcher gets executed.	765	be handled here by queueing them when their watcher gets executed.
572	- If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK	766	- If ev_unloop has been called, or EVLOOP_ONESHOT or EVLOOP_NONBLOCK
573	were used, return, otherwise continue with step *.	767	were used, or there are no active watchers, return, otherwise
		768	continue with step *.
574		769
575	Example: Queue some jobs and then loop until no events are outsanding	770	Example: Queue some jobs and then loop until no events are outstanding
576	anymore.	771	anymore.
577		772
578	... queue jobs here, make sure they register event watchers as long	773	... queue jobs here, make sure they register event watchers as long
579	... as they still have work to do (even an idle watcher will do..)	774	... as they still have work to do (even an idle watcher will do..)
580	ev_loop (my_loop, 0);	775	ev_loop (my_loop, 0);
581	... jobs done. yeah!	776	... jobs done or somebody called unloop. yeah!
582		777
583	=item ev_unloop (loop, how)	778	=item ev_unloop (loop, how)
584		779
585	Can be used to make a call to C<ev_loop> return early (but only after it	780	Can be used to make a call to C<ev_loop> return early (but only after it
586	has processed all outstanding events). The C<how> argument must be either	781	has processed all outstanding events). The C<how> argument must be either
587	C<EVUNLOOP_ONE>, which will make the innermost C<ev_loop> call return, or	782	C<EVUNLOOP_ONE>, which will make the innermost C<ev_loop> call return, or
588	C<EVUNLOOP_ALL>, which will make all nested C<ev_loop> calls return.	783	C<EVUNLOOP_ALL>, which will make all nested C<ev_loop> calls return.
589		784
		785	This "unloop state" will be cleared when entering C<ev_loop> again.
		786
		787	It is safe to call C<ev_unloop> from otuside any C<ev_loop> calls.
		788
590	=item ev_ref (loop)	789	=item ev_ref (loop)
591		790
592	=item ev_unref (loop)	791	=item ev_unref (loop)
593		792
594	Ref/unref can be used to add or remove a reference count on the event	793	Ref/unref can be used to add or remove a reference count on the event
595	loop: Every watcher keeps one reference, and as long as the reference	794	loop: Every watcher keeps one reference, and as long as the reference
596	count is nonzero, C<ev_loop> will not return on its own. If you have	795	count is nonzero, C<ev_loop> will not return on its own.
		796
597	a watcher you never unregister that should not keep C<ev_loop> from	797	If you have a watcher you never unregister that should not keep C<ev_loop>
598	returning, ev_unref() after starting, and ev_ref() before stopping it. For	798	from returning, call ev_unref() after starting, and ev_ref() before
		799	stopping it.
		800
599	example, libev itself uses this for its internal signal pipe: It is not	801	As an example, libev itself uses this for its internal signal pipe: It
600	visible to the libev user and should not keep C<ev_loop> from exiting if	802	is not visible to the libev user and should not keep C<ev_loop> from
601	no event watchers registered by it are active. It is also an excellent	803	exiting if no event watchers registered by it are active. It is also an
602	way to do this for generic recurring timers or from within third-party	804	excellent way to do this for generic recurring timers or from within
603	libraries. Just remember to I<unref after start> and I<ref before stop>.	805	third-party libraries. Just remember to I<unref after start> and I<ref
		806	before stop> (but only if the watcher wasn't active before, or was active
		807	before, respectively. Note also that libev might stop watchers itself
		808	(e.g. non-repeating timers) in which case you have to C<ev_ref>
		809	in the callback).
604		810
605	Example: Create a signal watcher, but keep it from keeping C<ev_loop>	811	Example: Create a signal watcher, but keep it from keeping C<ev_loop>
606	running when nothing else is active.	812	running when nothing else is active.
607		813
608	struct ev_signal exitsig;	814	ev_signal exitsig;
609	ev_signal_init (&exitsig, sig_cb, SIGINT);	815	ev_signal_init (&exitsig, sig_cb, SIGINT);
610	ev_signal_start (loop, &exitsig);	816	ev_signal_start (loop, &exitsig);
611	evf_unref (loop);	817	evf_unref (loop);
612		818
613	Example: For some weird reason, unregister the above signal handler again.	819	Example: For some weird reason, unregister the above signal handler again.
614		820
615	ev_ref (loop);	821	ev_ref (loop);
616	ev_signal_stop (loop, &exitsig);	822	ev_signal_stop (loop, &exitsig);
617		823
618	=item ev_set_io_collect_interval (loop, ev_tstamp interval)	824	=item ev_set_io_collect_interval (loop, ev_tstamp interval)
619		825
620	=item ev_set_timeout_collect_interval (loop, ev_tstamp interval)	826	=item ev_set_timeout_collect_interval (loop, ev_tstamp interval)
621		827
622	These advanced functions influence the time that libev will spend waiting	828	These advanced functions influence the time that libev will spend waiting
623	for events. Both are by default C<0>, meaning that libev will try to	829	for events. Both time intervals are by default C<0>, meaning that libev
624	invoke timer/periodic callbacks and I/O callbacks with minimum latency.	830	will try to invoke timer/periodic callbacks and I/O callbacks with minimum
		831	latency.
625		832
626	Setting these to a higher value (the C<interval> I<must> be >= C<0>)	833	Setting these to a higher value (the C<interval> I<must> be >= C<0>)
627	allows libev to delay invocation of I/O and timer/periodic callbacks to	834	allows libev to delay invocation of I/O and timer/periodic callbacks
628	increase efficiency of loop iterations.	835	to increase efficiency of loop iterations (or to increase power-saving
		836	opportunities).
629		837
630	The background is that sometimes your program runs just fast enough to	838	The idea is that sometimes your program runs just fast enough to handle
631	handle one (or very few) event(s) per loop iteration. While this makes	839	one (or very few) event(s) per loop iteration. While this makes the
632	the program responsive, it also wastes a lot of CPU time to poll for new	840	program responsive, it also wastes a lot of CPU time to poll for new
633	events, especially with backends like C<select ()> which have a high	841	events, especially with backends like C<select ()> which have a high
634	overhead for the actual polling but can deliver many events at once.	842	overhead for the actual polling but can deliver many events at once.
635		843
636	By setting a higher I<io collect interval> you allow libev to spend more	844	By setting a higher I<io collect interval> you allow libev to spend more
637	time collecting I/O events, so you can handle more events per iteration,	845	time collecting I/O events, so you can handle more events per iteration,
638	at the cost of increasing latency. Timeouts (both C<ev_periodic> and	846	at the cost of increasing latency. Timeouts (both C<ev_periodic> and
639	C<ev_timer>) will be not affected. Setting this to a non-null value will	847	C<ev_timer>) will be not affected. Setting this to a non-null value will
640	introduce an additional C<ev_sleep ()> call into most loop iterations.	848	introduce an additional C<ev_sleep ()> call into most loop iterations. The
		849	sleep time ensures that libev will not poll for I/O events more often then
		850	once per this interval, on average.
641		851
642	Likewise, by setting a higher I<timeout collect interval> you allow libev	852	Likewise, by setting a higher I<timeout collect interval> you allow libev
643	to spend more time collecting timeouts, at the expense of increased	853	to spend more time collecting timeouts, at the expense of increased
644	latency (the watcher callback will be called later). C<ev_io> watchers	854	latency/jitter/inexactness (the watcher callback will be called
645	will not be affected. Setting this to a non-null value will not introduce	855	later). C<ev_io> watchers will not be affected. Setting this to a non-null
646	any overhead in libev.	856	value will not introduce any overhead in libev.
647		857
648	Many (busy) programs can usually benefit by setting the io collect	858	Many (busy) programs can usually benefit by setting the I/O collect
649	interval to a value near C<0.1> or so, which is often enough for	859	interval to a value near C<0.1> or so, which is often enough for
650	interactive servers (of course not for games), likewise for timeouts. It	860	interactive servers (of course not for games), likewise for timeouts. It
651	usually doesn't make much sense to set it to a lower value than C<0.01>,	861	usually doesn't make much sense to set it to a lower value than C<0.01>,
652	as this approsaches the timing granularity of most systems.	862	as this approaches the timing granularity of most systems. Note that if
		863	you do transactions with the outside world and you can't increase the
		864	parallelity, then this setting will limit your transaction rate (if you
		865	need to poll once per transaction and the I/O collect interval is 0.01,
		866	then you can't do more than 100 transations per second).
		867
		868	Setting the I<timeout collect interval> can improve the opportunity for
		869	saving power, as the program will "bundle" timer callback invocations that
		870	are "near" in time together, by delaying some, thus reducing the number of
		871	times the process sleeps and wakes up again. Another useful technique to
		872	reduce iterations/wake-ups is to use C<ev_periodic> watchers and make sure
		873	they fire on, say, one-second boundaries only.
		874
		875	Example: we only need 0.1s timeout granularity, and we wish not to poll
		876	more often than 100 times per second:
		877
		878	ev_set_timeout_collect_interval (EV_DEFAULT_UC_ 0.1);
		879	ev_set_io_collect_interval (EV_DEFAULT_UC_ 0.01);
		880
		881	=item ev_invoke_pending (loop)
		882
		883	This call will simply invoke all pending watchers while resetting their
		884	pending state. Normally, C<ev_loop> does this automatically when required,
		885	but when overriding the invoke callback this call comes handy.
		886
		887	=item int ev_pending_count (loop)
		888
		889	Returns the number of pending watchers - zero indicates that no watchers
		890	are pending.
		891
		892	=item ev_set_invoke_pending_cb (loop, void (*invoke_pending_cb)(EV_P))
		893
		894	This overrides the invoke pending functionality of the loop: Instead of
		895	invoking all pending watchers when there are any, C<ev_loop> will call
		896	this callback instead. This is useful, for example, when you want to
		897	invoke the actual watchers inside another context (another thread etc.).
		898
		899	If you want to reset the callback, use C<ev_invoke_pending> as new
		900	callback.
		901
		902	=item ev_set_loop_release_cb (loop, void (*release)(EV_P), void (*acquire)(EV_P))
		903
		904	Sometimes you want to share the same loop between multiple threads. This
		905	can be done relatively simply by putting mutex_lock/unlock calls around
		906	each call to a libev function.
		907
		908	However, C<ev_loop> can run an indefinite time, so it is not feasible to
		909	wait for it to return. One way around this is to wake up the loop via
		910	C<ev_unloop> and C<av_async_send>, another way is to set these I<release>
		911	and I<acquire> callbacks on the loop.
		912
		913	When set, then C<release> will be called just before the thread is
		914	suspended waiting for new events, and C<acquire> is called just
		915	afterwards.
		916
		917	Ideally, C<release> will just call your mutex_unlock function, and
		918	C<acquire> will just call the mutex_lock function again.
		919
		920	While event loop modifications are allowed between invocations of
		921	C<release> and C<acquire> (that's their only purpose after all), no
		922	modifications done will affect the event loop, i.e. adding watchers will
		923	have no effect on the set of file descriptors being watched, or the time
		924	waited. USe an C<ev_async> watcher to wake up C<ev_loop> when you want it
		925	to take note of any changes you made.
		926
		927	In theory, threads executing C<ev_loop> will be async-cancel safe between
		928	invocations of C<release> and C<acquire>.
		929
		930	See also the locking example in the C<THREADS> section later in this
		931	document.
		932
		933	=item ev_set_userdata (loop, void *data)
		934
		935	=item ev_userdata (loop)
		936
		937	Set and retrieve a single C<void *> associated with a loop. When
		938	C<ev_set_userdata> has never been called, then C<ev_userdata> returns
		939	C<0.>
		940
		941	These two functions can be used to associate arbitrary data with a loop,
		942	and are intended solely for the C<invoke_pending_cb>, C<release> and
		943	C<acquire> callbacks described above, but of course can be (ab-)used for
		944	any other purpose as well.
		945
		946	=item ev_loop_verify (loop)
		947
		948	This function only does something when C<EV_VERIFY> support has been
		949	compiled in, which is the default for non-minimal builds. It tries to go
		950	through all internal structures and checks them for validity. If anything
		951	is found to be inconsistent, it will print an error message to standard
		952	error and call C<abort ()>.
		953
		954	This can be used to catch bugs inside libev itself: under normal
		955	circumstances, this function will never abort as of course libev keeps its
		956	data structures consistent.
653		957
654	=back	958	=back
655		959
656		960
657	=head1 ANATOMY OF A WATCHER	961	=head1 ANATOMY OF A WATCHER
		962
		963	In the following description, uppercase C<TYPE> in names stands for the
		964	watcher type, e.g. C<ev_TYPE_start> can mean C<ev_timer_start> for timer
		965	watchers and C<ev_io_start> for I/O watchers.
658		966
659	A watcher is a structure that you create and register to record your	967	A watcher is a structure that you create and register to record your
660	interest in some event. For instance, if you want to wait for STDIN to	968	interest in some event. For instance, if you want to wait for STDIN to
661	become readable, you would create an C<ev_io> watcher for that:	969	become readable, you would create an C<ev_io> watcher for that:
662		970
663	static void my_cb (struct ev_loop *loop, struct ev_io *w, int revents)	971	static void my_cb (struct ev_loop *loop, ev_io *w, int revents)
664	{	972	{
665	ev_io_stop (w);	973	ev_io_stop (w);
666	ev_unloop (loop, EVUNLOOP_ALL);	974	ev_unloop (loop, EVUNLOOP_ALL);
667	}	975	}
668		976
669	struct ev_loop *loop = ev_default_loop (0);	977	struct ev_loop *loop = ev_default_loop (0);
		978
670	struct ev_io stdin_watcher;	979	ev_io stdin_watcher;
		980
671	ev_init (&stdin_watcher, my_cb);	981	ev_init (&stdin_watcher, my_cb);
672	ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);	982	ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);
673	ev_io_start (loop, &stdin_watcher);	983	ev_io_start (loop, &stdin_watcher);
		984
674	ev_loop (loop, 0);	985	ev_loop (loop, 0);
675		986
676	As you can see, you are responsible for allocating the memory for your	987	As you can see, you are responsible for allocating the memory for your
677	watcher structures (and it is usually a bad idea to do this on the stack,	988	watcher structures (and it is I<usually> a bad idea to do this on the
678	although this can sometimes be quite valid).	989	stack).
		990
		991	Each watcher has an associated watcher structure (called C<struct ev_TYPE>
		992	or simply C<ev_TYPE>, as typedefs are provided for all watcher structs).
679		993
680	Each watcher structure must be initialised by a call to C<ev_init	994	Each watcher structure must be initialised by a call to C<ev_init
681	(watcher *, callback)>, which expects a callback to be provided. This	995	(watcher *, callback)>, which expects a callback to be provided. This
682	callback gets invoked each time the event occurs (or, in the case of io	996	callback gets invoked each time the event occurs (or, in the case of I/O
683	watchers, each time the event loop detects that the file descriptor given	997	watchers, each time the event loop detects that the file descriptor given
684	is readable and/or writable).	998	is readable and/or writable).
685		999
686	Each watcher type has its own C<< ev_<type>_set (watcher *, ...) >> macro	1000	Each watcher type further has its own C<< ev_TYPE_set (watcher *, ...) >>
687	with arguments specific to this watcher type. There is also a macro	1001	macro to configure it, with arguments specific to the watcher type. There
688	to combine initialisation and setting in one call: C<< ev_<type>_init	1002	is also a macro to combine initialisation and setting in one call: C<<
689	(watcher *, callback, ...) >>.	1003	ev_TYPE_init (watcher *, callback, ...) >>.
690		1004
691	To make the watcher actually watch out for events, you have to start it	1005	To make the watcher actually watch out for events, you have to start it
692	with a watcher-specific start function (C<< ev_<type>_start (loop, watcher	1006	with a watcher-specific start function (C<< ev_TYPE_start (loop, watcher
693	*) >>), and you can stop watching for events at any time by calling the	1007	*) >>), and you can stop watching for events at any time by calling the
694	corresponding stop function (C<< ev_<type>_stop (loop, watcher *) >>.	1008	corresponding stop function (C<< ev_TYPE_stop (loop, watcher *) >>.
695		1009
696	As long as your watcher is active (has been started but not stopped) you	1010	As long as your watcher is active (has been started but not stopped) you
697	must not touch the values stored in it. Most specifically you must never	1011	must not touch the values stored in it. Most specifically you must never
698	reinitialise it or call its C<set> macro.	1012	reinitialise it or call its C<ev_TYPE_set> macro.
699		1013
700	Each and every callback receives the event loop pointer as first, the	1014	Each and every callback receives the event loop pointer as first, the
701	registered watcher structure as second, and a bitset of received events as	1015	registered watcher structure as second, and a bitset of received events as
702	third argument.	1016	third argument.
703		1017
…		…
757	=item C<EV_FORK>	1071	=item C<EV_FORK>
758		1072
759	The event loop has been resumed in the child process after fork (see	1073	The event loop has been resumed in the child process after fork (see
760	C<ev_fork>).	1074	C<ev_fork>).
761		1075
		1076	=item C<EV_ASYNC>
		1077
		1078	The given async watcher has been asynchronously notified (see C<ev_async>).
		1079
		1080	=item C<EV_CUSTOM>
		1081
		1082	Not ever sent (or otherwise used) by libev itself, but can be freely used
		1083	by libev users to signal watchers (e.g. via C<ev_feed_event>).
		1084
762	=item C<EV_ERROR>	1085	=item C<EV_ERROR>
763		1086
764	An unspecified error has occured, the watcher has been stopped. This might	1087	An unspecified error has occurred, the watcher has been stopped. This might
765	happen because the watcher could not be properly started because libev	1088	happen because the watcher could not be properly started because libev
766	ran out of memory, a file descriptor was found to be closed or any other	1089	ran out of memory, a file descriptor was found to be closed or any other
		1090	problem. Libev considers these application bugs.
		1091
767	problem. You best act on it by reporting the problem and somehow coping	1092	You best act on it by reporting the problem and somehow coping with the
768	with the watcher being stopped.	1093	watcher being stopped. Note that well-written programs should not receive
		1094	an error ever, so when your watcher receives it, this usually indicates a
		1095	bug in your program.
769		1096
770	Libev will usually signal a few "dummy" events together with an error,	1097	Libev will usually signal a few "dummy" events together with an error, for
771	for example it might indicate that a fd is readable or writable, and if	1098	example it might indicate that a fd is readable or writable, and if your
772	your callbacks is well-written it can just attempt the operation and cope	1099	callbacks is well-written it can just attempt the operation and cope with
773	with the error from read() or write(). This will not work in multithreaded	1100	the error from read() or write(). This will not work in multi-threaded
774	programs, though, so beware.	1101	programs, though, as the fd could already be closed and reused for another
		1102	thing, so beware.
775		1103
776	=back	1104	=back
777		1105
778	=head2 GENERIC WATCHER FUNCTIONS	1106	=head2 GENERIC WATCHER FUNCTIONS
779
780	In the following description, C<TYPE> stands for the watcher type,
781	e.g. C<timer> for C<ev_timer> watchers and C<io> for C<ev_io> watchers.
782		1107
783	=over 4	1108	=over 4
784		1109
785	=item C<ev_init> (ev_TYPE *watcher, callback)	1110	=item C<ev_init> (ev_TYPE *watcher, callback)
786		1111
…		…
792	which rolls both calls into one.	1117	which rolls both calls into one.
793		1118
794	You can reinitialise a watcher at any time as long as it has been stopped	1119	You can reinitialise a watcher at any time as long as it has been stopped
795	(or never started) and there are no pending events outstanding.	1120	(or never started) and there are no pending events outstanding.
796		1121
797	The callback is always of type C<void (*)(ev_loop *loop, ev_TYPE *watcher,	1122	The callback is always of type C<void (*)(struct ev_loop *loop, ev_TYPE *watcher,
798	int revents)>.	1123	int revents)>.
		1124
		1125	Example: Initialise an C<ev_io> watcher in two steps.
		1126
		1127	ev_io w;
		1128	ev_init (&w, my_cb);
		1129	ev_io_set (&w, STDIN_FILENO, EV_READ);
799		1130
800	=item C<ev_TYPE_set> (ev_TYPE *, [args])	1131	=item C<ev_TYPE_set> (ev_TYPE *, [args])
801		1132
802	This macro initialises the type-specific parts of a watcher. You need to	1133	This macro initialises the type-specific parts of a watcher. You need to
803	call C<ev_init> at least once before you call this macro, but you can	1134	call C<ev_init> at least once before you call this macro, but you can
…		…
806	difference to the C<ev_init> macro).	1137	difference to the C<ev_init> macro).
807		1138
808	Although some watcher types do not have type-specific arguments	1139	Although some watcher types do not have type-specific arguments
809	(e.g. C<ev_prepare>) you still need to call its C<set> macro.	1140	(e.g. C<ev_prepare>) you still need to call its C<set> macro.
810		1141
		1142	See C<ev_init>, above, for an example.
		1143
811	=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])	1144	=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])
812		1145
813	This convinience macro rolls both C<ev_init> and C<ev_TYPE_set> macro	1146	This convenience macro rolls both C<ev_init> and C<ev_TYPE_set> macro
814	calls into a single call. This is the most convinient method to initialise	1147	calls into a single call. This is the most convenient method to initialise
815	a watcher. The same limitations apply, of course.	1148	a watcher. The same limitations apply, of course.
		1149
		1150	Example: Initialise and set an C<ev_io> watcher in one step.
		1151
		1152	ev_io_init (&w, my_cb, STDIN_FILENO, EV_READ);
816		1153
817	=item C<ev_TYPE_start> (loop *, ev_TYPE *watcher)	1154	=item C<ev_TYPE_start> (loop *, ev_TYPE *watcher)
818		1155
819	Starts (activates) the given watcher. Only active watchers will receive	1156	Starts (activates) the given watcher. Only active watchers will receive
820	events. If the watcher is already active nothing will happen.	1157	events. If the watcher is already active nothing will happen.
821		1158
		1159	Example: Start the C<ev_io> watcher that is being abused as example in this
		1160	whole section.
		1161
		1162	ev_io_start (EV_DEFAULT_UC, &w);
		1163
822	=item C<ev_TYPE_stop> (loop *, ev_TYPE *watcher)	1164	=item C<ev_TYPE_stop> (loop *, ev_TYPE *watcher)
823		1165
824	Stops the given watcher again (if active) and clears the pending	1166	Stops the given watcher if active, and clears the pending status (whether
		1167	the watcher was active or not).
		1168
825	status. It is possible that stopped watchers are pending (for example,	1169	It is possible that stopped watchers are pending - for example,
826	non-repeating timers are being stopped when they become pending), but	1170	non-repeating timers are being stopped when they become pending - but
827	C<ev_TYPE_stop> ensures that the watcher is neither active nor pending. If	1171	calling C<ev_TYPE_stop> ensures that the watcher is neither active nor
828	you want to free or reuse the memory used by the watcher it is therefore a	1172	pending. If you want to free or reuse the memory used by the watcher it is
829	good idea to always call its C<ev_TYPE_stop> function.	1173	therefore a good idea to always call its C<ev_TYPE_stop> function.
830		1174
831	=item bool ev_is_active (ev_TYPE *watcher)	1175	=item bool ev_is_active (ev_TYPE *watcher)
832		1176
833	Returns a true value iff the watcher is active (i.e. it has been started	1177	Returns a true value iff the watcher is active (i.e. it has been started
834	and not yet been stopped). As long as a watcher is active you must not modify	1178	and not yet been stopped). As long as a watcher is active you must not modify
…		…
860	integer between C<EV_MAXPRI> (default: C<2>) and C<EV_MINPRI>	1204	integer between C<EV_MAXPRI> (default: C<2>) and C<EV_MINPRI>
861	(default: C<-2>). Pending watchers with higher priority will be invoked	1205	(default: C<-2>). Pending watchers with higher priority will be invoked
862	before watchers with lower priority, but priority will not keep watchers	1206	before watchers with lower priority, but priority will not keep watchers
863	from being executed (except for C<ev_idle> watchers).	1207	from being executed (except for C<ev_idle> watchers).
864		1208
865	This means that priorities are I<only> used for ordering callback
866	invocation after new events have been received. This is useful, for
867	example, to reduce latency after idling, or more often, to bind two
868	watchers on the same event and make sure one is called first.
869
870	If you need to suppress invocation when higher priority events are pending	1209	If you need to suppress invocation when higher priority events are pending
871	you need to look at C<ev_idle> watchers, which provide this functionality.	1210	you need to look at C<ev_idle> watchers, which provide this functionality.
872		1211
873	You I<must not> change the priority of a watcher as long as it is active or	1212	You I<must not> change the priority of a watcher as long as it is active or
874	pending.	1213	pending.
875		1214
		1215	Setting a priority outside the range of C<EV_MINPRI> to C<EV_MAXPRI> is
		1216	fine, as long as you do not mind that the priority value you query might
		1217	or might not have been clamped to the valid range.
		1218
876	The default priority used by watchers when no priority has been set is	1219	The default priority used by watchers when no priority has been set is
877	always C<0>, which is supposed to not be too high and not be too low :).	1220	always C<0>, which is supposed to not be too high and not be too low :).
878		1221
879	Setting a priority outside the range of C<EV_MINPRI> to C<EV_MAXPRI> is	1222	See L<WATCHER PRIORITY MODELS>, below, for a more thorough treatment of
880	fine, as long as you do not mind that the priority value you query might	1223	priorities.
881	or might not have been adjusted to be within valid range.
882		1224
883	=item ev_invoke (loop, ev_TYPE *watcher, int revents)	1225	=item ev_invoke (loop, ev_TYPE *watcher, int revents)
884		1226
885	Invoke the C<watcher> with the given C<loop> and C<revents>. Neither	1227	Invoke the C<watcher> with the given C<loop> and C<revents>. Neither
886	C<loop> nor C<revents> need to be valid as long as the watcher callback	1228	C<loop> nor C<revents> need to be valid as long as the watcher callback
887	can deal with that fact.	1229	can deal with that fact, as both are simply passed through to the
		1230	callback.
888		1231
889	=item int ev_clear_pending (loop, ev_TYPE *watcher)	1232	=item int ev_clear_pending (loop, ev_TYPE *watcher)
890		1233
891	If the watcher is pending, this function returns clears its pending status	1234	If the watcher is pending, this function clears its pending status and
892	and returns its C<revents> bitset (as if its callback was invoked). If the	1235	returns its C<revents> bitset (as if its callback was invoked). If the
893	watcher isn't pending it does nothing and returns C<0>.	1236	watcher isn't pending it does nothing and returns C<0>.
894		1237
		1238	Sometimes it can be useful to "poll" a watcher instead of waiting for its
		1239	callback to be invoked, which can be accomplished with this function.
		1240
		1241	=item ev_feed_event (struct ev_loop *, watcher *, int revents)
		1242
		1243	Feeds the given event set into the event loop, as if the specified event
		1244	had happened for the specified watcher (which must be a pointer to an
		1245	initialised but not necessarily started event watcher). Obviously you must
		1246	not free the watcher as long as it has pending events.
		1247
		1248	Stopping the watcher, letting libev invoke it, or calling
		1249	C<ev_clear_pending> will clear the pending event, even if the watcher was
		1250	not started in the first place.
		1251
		1252	See also C<ev_feed_fd_event> and C<ev_feed_signal_event> for related
		1253	functions that do not need a watcher.
		1254
895	=back	1255	=back
896		1256
897		1257
898	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER	1258	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER
899		1259
900	Each watcher has, by default, a member C<void *data> that you can change	1260	Each watcher has, by default, a member C<void *data> that you can change
901	and read at any time, libev will completely ignore it. This can be used	1261	and read at any time: libev will completely ignore it. This can be used
902	to associate arbitrary data with your watcher. If you need more data and	1262	to associate arbitrary data with your watcher. If you need more data and
903	don't want to allocate memory and store a pointer to it in that data	1263	don't want to allocate memory and store a pointer to it in that data
904	member, you can also "subclass" the watcher type and provide your own	1264	member, you can also "subclass" the watcher type and provide your own
905	data:	1265	data:
906		1266
907	struct my_io	1267	struct my_io
908	{	1268	{
909	struct ev_io io;	1269	ev_io io;
910	int otherfd;	1270	int otherfd;
911	void *somedata;	1271	void *somedata;
912	struct whatever *mostinteresting;	1272	struct whatever *mostinteresting;
913	}	1273	};
		1274
		1275	...
		1276	struct my_io w;
		1277	ev_io_init (&w.io, my_cb, fd, EV_READ);
914		1278
915	And since your callback will be called with a pointer to the watcher, you	1279	And since your callback will be called with a pointer to the watcher, you
916	can cast it back to your own type:	1280	can cast it back to your own type:
917		1281
918	static void my_cb (struct ev_loop *loop, struct ev_io *w_, int revents)	1282	static void my_cb (struct ev_loop *loop, ev_io *w_, int revents)
919	{	1283	{
920	struct my_io *w = (struct my_io *)w_;	1284	struct my_io *w = (struct my_io *)w_;
921	...	1285	...
922	}	1286	}
923		1287
924	More interesting and less C-conformant ways of casting your callback type	1288	More interesting and less C-conformant ways of casting your callback type
925	instead have been omitted.	1289	instead have been omitted.
926		1290
927	Another common scenario is having some data structure with multiple	1291	Another common scenario is to use some data structure with multiple
928	watchers:	1292	embedded watchers:
929		1293
930	struct my_biggy	1294	struct my_biggy
931	{	1295	{
932	int some_data;	1296	int some_data;
933	ev_timer t1;	1297	ev_timer t1;
934	ev_timer t2;	1298	ev_timer t2;
935	}	1299	}
936		1300
937	In this case getting the pointer to C<my_biggy> is a bit more complicated,	1301	In this case getting the pointer to C<my_biggy> is a bit more
938	you need to use C<offsetof>:	1302	complicated: Either you store the address of your C<my_biggy> struct
		1303	in the C<data> member of the watcher (for woozies), or you need to use
		1304	some pointer arithmetic using C<offsetof> inside your watchers (for real
		1305	programmers):
939		1306
940	#include <stddef.h>	1307	#include <stddef.h>
941		1308
942	static void	1309	static void
943	t1_cb (EV_P_ struct ev_timer *w, int revents)	1310	t1_cb (EV_P_ ev_timer *w, int revents)
944	{	1311	{
945	struct my_biggy big = (struct my_biggy *	1312	struct my_biggy big = (struct my_biggy *)
946	(((char *)w) - offsetof (struct my_biggy, t1));	1313	(((char *)w) - offsetof (struct my_biggy, t1));
947	}	1314	}
948		1315
949	static void	1316	static void
950	t2_cb (EV_P_ struct ev_timer *w, int revents)	1317	t2_cb (EV_P_ ev_timer *w, int revents)
951	{	1318	{
952	struct my_biggy big = (struct my_biggy *	1319	struct my_biggy big = (struct my_biggy *)
953	(((char *)w) - offsetof (struct my_biggy, t2));	1320	(((char *)w) - offsetof (struct my_biggy, t2));
954	}	1321	}
		1322
		1323	=head2 WATCHER PRIORITY MODELS
		1324
		1325	Many event loops support I<watcher priorities>, which are usually small
		1326	integers that influence the ordering of event callback invocation
		1327	between watchers in some way, all else being equal.
		1328
		1329	In libev, Watcher priorities can be set using C<ev_set_priority>. See its
		1330	description for the more technical details such as the actual priority
		1331	range.
		1332
		1333	There are two common ways how these these priorities are being interpreted
		1334	by event loops:
		1335
		1336	In the more common lock-out model, higher priorities "lock out" invocation
		1337	of lower priority watchers, which means as long as higher priority
		1338	watchers receive events, lower priority watchers are not being invoked.
		1339
		1340	The less common only-for-ordering model uses priorities solely to order
		1341	callback invocation within a single event loop iteration: Higher priority
		1342	watchers are invoked before lower priority ones, but they all get invoked
		1343	before polling for new events.
		1344
		1345	Libev uses the second (only-for-ordering) model for all its watchers
		1346	except for idle watchers (which use the lock-out model).
		1347
		1348	The rationale behind this is that implementing the lock-out model for
		1349	watchers is not well supported by most kernel interfaces, and most event
		1350	libraries will just poll for the same events again and again as long as
		1351	their callbacks have not been executed, which is very inefficient in the
		1352	common case of one high-priority watcher locking out a mass of lower
		1353	priority ones.
		1354
		1355	Static (ordering) priorities are most useful when you have two or more
		1356	watchers handling the same resource: a typical usage example is having an
		1357	C<ev_io> watcher to receive data, and an associated C<ev_timer> to handle
		1358	timeouts. Under load, data might be received while the program handles
		1359	other jobs, but since timers normally get invoked first, the timeout
		1360	handler will be executed before checking for data. In that case, giving
		1361	the timer a lower priority than the I/O watcher ensures that I/O will be
		1362	handled first even under adverse conditions (which is usually, but not
		1363	always, what you want).
		1364
		1365	Since idle watchers use the "lock-out" model, meaning that idle watchers
		1366	will only be executed when no same or higher priority watchers have
		1367	received events, they can be used to implement the "lock-out" model when
		1368	required.
		1369
		1370	For example, to emulate how many other event libraries handle priorities,
		1371	you can associate an C<ev_idle> watcher to each such watcher, and in
		1372	the normal watcher callback, you just start the idle watcher. The real
		1373	processing is done in the idle watcher callback. This causes libev to
		1374	continously poll and process kernel event data for the watcher, but when
		1375	the lock-out case is known to be rare (which in turn is rare :), this is
		1376	workable.
		1377
		1378	Usually, however, the lock-out model implemented that way will perform
		1379	miserably under the type of load it was designed to handle. In that case,
		1380	it might be preferable to stop the real watcher before starting the
		1381	idle watcher, so the kernel will not have to process the event in case
		1382	the actual processing will be delayed for considerable time.
		1383
		1384	Here is an example of an I/O watcher that should run at a strictly lower
		1385	priority than the default, and which should only process data when no
		1386	other events are pending:
		1387
		1388	ev_idle idle; // actual processing watcher
		1389	ev_io io; // actual event watcher
		1390
		1391	static void
		1392	io_cb (EV_P_ ev_io *w, int revents)
		1393	{
		1394	// stop the I/O watcher, we received the event, but
		1395	// are not yet ready to handle it.
		1396	ev_io_stop (EV_A_ w);
		1397
		1398	// start the idle watcher to ahndle the actual event.
		1399	// it will not be executed as long as other watchers
		1400	// with the default priority are receiving events.
		1401	ev_idle_start (EV_A_ &idle);
		1402	}
		1403
		1404	static void
		1405	idle_cb (EV_P_ ev_idle *w, int revents)
		1406	{
		1407	// actual processing
		1408	read (STDIN_FILENO, ...);
		1409
		1410	// have to start the I/O watcher again, as
		1411	// we have handled the event
		1412	ev_io_start (EV_P_ &io);
		1413	}
		1414
		1415	// initialisation
		1416	ev_idle_init (&idle, idle_cb);
		1417	ev_io_init (&io, io_cb, STDIN_FILENO, EV_READ);
		1418	ev_io_start (EV_DEFAULT_ &io);
		1419
		1420	In the "real" world, it might also be beneficial to start a timer, so that
		1421	low-priority connections can not be locked out forever under load. This
		1422	enables your program to keep a lower latency for important connections
		1423	during short periods of high load, while not completely locking out less
		1424	important ones.
955		1425
956		1426
957	=head1 WATCHER TYPES	1427	=head1 WATCHER TYPES
958		1428
959	This section describes each watcher in detail, but will not repeat	1429	This section describes each watcher in detail, but will not repeat
…		…
983	In general you can register as many read and/or write event watchers per	1453	In general you can register as many read and/or write event watchers per
984	fd as you want (as long as you don't confuse yourself). Setting all file	1454	fd as you want (as long as you don't confuse yourself). Setting all file
985	descriptors to non-blocking mode is also usually a good idea (but not	1455	descriptors to non-blocking mode is also usually a good idea (but not
986	required if you know what you are doing).	1456	required if you know what you are doing).
987		1457
988	You have to be careful with dup'ed file descriptors, though. Some backends	1458	If you cannot use non-blocking mode, then force the use of a
989	(the linux epoll backend is a notable example) cannot handle dup'ed file	1459	known-to-be-good backend (at the time of this writing, this includes only
990	descriptors correctly if you register interest in two or more fds pointing	1460	C<EVBACKEND_SELECT> and C<EVBACKEND_POLL>). The same applies to file
991	to the same underlying file/socket/etc. description (that is, they share	1461	descriptors for which non-blocking operation makes no sense (such as
992	the same underlying "file open").	1462	files) - libev doesn't guarentee any specific behaviour in that case.
993
994	If you must do this, then force the use of a known-to-be-good backend
995	(at the time of this writing, this includes only C<EVBACKEND_SELECT> and
996	C<EVBACKEND_POLL>).
997		1463
998	Another thing you have to watch out for is that it is quite easy to	1464	Another thing you have to watch out for is that it is quite easy to
999	receive "spurious" readyness notifications, that is your callback might	1465	receive "spurious" readiness notifications, that is your callback might
1000	be called with C<EV_READ> but a subsequent C<read>(2) will actually block	1466	be called with C<EV_READ> but a subsequent C<read>(2) will actually block
1001	because there is no data. Not only are some backends known to create a	1467	because there is no data. Not only are some backends known to create a
1002	lot of those (for example solaris ports), it is very easy to get into	1468	lot of those (for example Solaris ports), it is very easy to get into
1003	this situation even with a relatively standard program structure. Thus	1469	this situation even with a relatively standard program structure. Thus
1004	it is best to always use non-blocking I/O: An extra C<read>(2) returning	1470	it is best to always use non-blocking I/O: An extra C<read>(2) returning
1005	C<EAGAIN> is far preferable to a program hanging until some data arrives.	1471	C<EAGAIN> is far preferable to a program hanging until some data arrives.
1006		1472
1007	If you cannot run the fd in non-blocking mode (for example you should not	1473	If you cannot run the fd in non-blocking mode (for example you should
1008	play around with an Xlib connection), then you have to seperately re-test	1474	not play around with an Xlib connection), then you have to separately
1009	whether a file descriptor is really ready with a known-to-be good interface	1475	re-test whether a file descriptor is really ready with a known-to-be good
1010	such as poll (fortunately in our Xlib example, Xlib already does this on	1476	interface such as poll (fortunately in our Xlib example, Xlib already
1011	its own, so its quite safe to use).	1477	does this on its own, so its quite safe to use). Some people additionally
		1478	use C<SIGALRM> and an interval timer, just to be sure you won't block
		1479	indefinitely.
		1480
		1481	But really, best use non-blocking mode.
1012		1482
1013	=head3 The special problem of disappearing file descriptors	1483	=head3 The special problem of disappearing file descriptors
1014		1484
1015	Some backends (e.g. kqueue, epoll) need to be told about closing a file	1485	Some backends (e.g. kqueue, epoll) need to be told about closing a file
1016	descriptor (either by calling C<close> explicitly or by any other means,	1486	descriptor (either due to calling C<close> explicitly or any other means,
1017	such as C<dup>). The reason is that you register interest in some file	1487	such as C<dup2>). The reason is that you register interest in some file
1018	descriptor, but when it goes away, the operating system will silently drop	1488	descriptor, but when it goes away, the operating system will silently drop
1019	this interest. If another file descriptor with the same number then is	1489	this interest. If another file descriptor with the same number then is
1020	registered with libev, there is no efficient way to see that this is, in	1490	registered with libev, there is no efficient way to see that this is, in
1021	fact, a different file descriptor.	1491	fact, a different file descriptor.
1022		1492
…		…
1033		1503
1034	=head3 The special problem of dup'ed file descriptors	1504	=head3 The special problem of dup'ed file descriptors
1035		1505
1036	Some backends (e.g. epoll), cannot register events for file descriptors,	1506	Some backends (e.g. epoll), cannot register events for file descriptors,
1037	but only events for the underlying file descriptions. That means when you	1507	but only events for the underlying file descriptions. That means when you
1038	have C<dup ()>'ed file descriptors and register events for them, only one	1508	have C<dup ()>'ed file descriptors or weirder constellations, and register
1039	file descriptor might actually receive events.	1509	events for them, only one file descriptor might actually receive events.
1040		1510
1041	There is no workaround possible except not registering events	1511	There is no workaround possible except not registering events
1042	for potentially C<dup ()>'ed file descriptors, or to resort to	1512	for potentially C<dup ()>'ed file descriptors, or to resort to
1043	C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>.	1513	C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>.
1044		1514
…		…
1051	To support fork in your programs, you either have to call	1521	To support fork in your programs, you either have to call
1052	C<ev_default_fork ()> or C<ev_loop_fork ()> after a fork in the child,	1522	C<ev_default_fork ()> or C<ev_loop_fork ()> after a fork in the child,
1053	enable C<EVFLAG_FORKCHECK>, or resort to C<EVBACKEND_SELECT> or	1523	enable C<EVFLAG_FORKCHECK>, or resort to C<EVBACKEND_SELECT> or
1054	C<EVBACKEND_POLL>.	1524	C<EVBACKEND_POLL>.
1055		1525
		1526	=head3 The special problem of SIGPIPE
		1527
		1528	While not really specific to libev, it is easy to forget about C<SIGPIPE>:
		1529	when writing to a pipe whose other end has been closed, your program gets
		1530	sent a SIGPIPE, which, by default, aborts your program. For most programs
		1531	this is sensible behaviour, for daemons, this is usually undesirable.
		1532
		1533	So when you encounter spurious, unexplained daemon exits, make sure you
		1534	ignore SIGPIPE (and maybe make sure you log the exit status of your daemon
		1535	somewhere, as that would have given you a big clue).
		1536
1056		1537
1057	=head3 Watcher-Specific Functions	1538	=head3 Watcher-Specific Functions
1058		1539
1059	=over 4	1540	=over 4
1060		1541
1061	=item ev_io_init (ev_io *, callback, int fd, int events)	1542	=item ev_io_init (ev_io *, callback, int fd, int events)
1062		1543
1063	=item ev_io_set (ev_io *, int fd, int events)	1544	=item ev_io_set (ev_io *, int fd, int events)
1064		1545
1065	Configures an C<ev_io> watcher. The C<fd> is the file descriptor to	1546	Configures an C<ev_io> watcher. The C<fd> is the file descriptor to
1066	rceeive events for and events is either C<EV_READ>, C<EV_WRITE> or	1547	receive events for and C<events> is either C<EV_READ>, C<EV_WRITE> or
1067	C<EV_READ | EV_WRITE> to receive the given events.	1548	C<EV_READ | EV_WRITE>, to express the desire to receive the given events.
1068		1549
1069	=item int fd [read-only]	1550	=item int fd [read-only]
1070		1551
1071	The file descriptor being watched.	1552	The file descriptor being watched.
1072		1553
1073	=item int events [read-only]	1554	=item int events [read-only]
1074		1555
1075	The events being watched.	1556	The events being watched.
1076		1557
1077	=back	1558	=back
		1559
		1560	=head3 Examples
1078		1561
1079	Example: Call C<stdin_readable_cb> when STDIN_FILENO has become, well	1562	Example: Call C<stdin_readable_cb> when STDIN_FILENO has become, well
1080	readable, but only once. Since it is likely line-buffered, you could	1563	readable, but only once. Since it is likely line-buffered, you could
1081	attempt to read a whole line in the callback.	1564	attempt to read a whole line in the callback.
1082		1565
1083	static void	1566	static void
1084	stdin_readable_cb (struct ev_loop *loop, struct ev_io *w, int revents)	1567	stdin_readable_cb (struct ev_loop *loop, ev_io *w, int revents)
1085	{	1568	{
1086	ev_io_stop (loop, w);	1569	ev_io_stop (loop, w);
1087	.. read from stdin here (or from w->fd) and haqndle any I/O errors	1570	.. read from stdin here (or from w->fd) and handle any I/O errors
1088	}	1571	}
1089		1572
1090	...	1573	...
1091	struct ev_loop *loop = ev_default_init (0);	1574	struct ev_loop *loop = ev_default_init (0);
1092	struct ev_io stdin_readable;	1575	ev_io stdin_readable;
1093	ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);	1576	ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
1094	ev_io_start (loop, &stdin_readable);	1577	ev_io_start (loop, &stdin_readable);
1095	ev_loop (loop, 0);	1578	ev_loop (loop, 0);
1096		1579
1097		1580
1098	=head2 C<ev_timer> - relative and optionally repeating timeouts	1581	=head2 C<ev_timer> - relative and optionally repeating timeouts
1099		1582
1100	Timer watchers are simple relative timers that generate an event after a	1583	Timer watchers are simple relative timers that generate an event after a
1101	given time, and optionally repeating in regular intervals after that.	1584	given time, and optionally repeating in regular intervals after that.
1102		1585
1103	The timers are based on real time, that is, if you register an event that	1586	The timers are based on real time, that is, if you register an event that
1104	times out after an hour and you reset your system clock to last years	1587	times out after an hour and you reset your system clock to January last
1105	time, it will still time out after (roughly) and hour. "Roughly" because	1588	year, it will still time out after (roughly) one hour. "Roughly" because
1106	detecting time jumps is hard, and some inaccuracies are unavoidable (the	1589	detecting time jumps is hard, and some inaccuracies are unavoidable (the
1107	monotonic clock option helps a lot here).	1590	monotonic clock option helps a lot here).
		1591
		1592	The callback is guaranteed to be invoked only I<after> its timeout has
		1593	passed (not I<at>, so on systems with very low-resolution clocks this
		1594	might introduce a small delay). If multiple timers become ready during the
		1595	same loop iteration then the ones with earlier time-out values are invoked
		1596	before ones of the same priority with later time-out values (but this is
		1597	no longer true when a callback calls C<ev_loop> recursively).
		1598
		1599	=head3 Be smart about timeouts
		1600
		1601	Many real-world problems involve some kind of timeout, usually for error
		1602	recovery. A typical example is an HTTP request - if the other side hangs,
		1603	you want to raise some error after a while.
		1604
		1605	What follows are some ways to handle this problem, from obvious and
		1606	inefficient to smart and efficient.
		1607
		1608	In the following, a 60 second activity timeout is assumed - a timeout that
		1609	gets reset to 60 seconds each time there is activity (e.g. each time some
		1610	data or other life sign was received).
		1611
		1612	=over 4
		1613
		1614	=item 1. Use a timer and stop, reinitialise and start it on activity.
		1615
		1616	This is the most obvious, but not the most simple way: In the beginning,
		1617	start the watcher:
		1618
		1619	ev_timer_init (timer, callback, 60., 0.);
		1620	ev_timer_start (loop, timer);
		1621
		1622	Then, each time there is some activity, C<ev_timer_stop> it, initialise it
		1623	and start it again:
		1624
		1625	ev_timer_stop (loop, timer);
		1626	ev_timer_set (timer, 60., 0.);
		1627	ev_timer_start (loop, timer);
		1628
		1629	This is relatively simple to implement, but means that each time there is
		1630	some activity, libev will first have to remove the timer from its internal
		1631	data structure and then add it again. Libev tries to be fast, but it's
		1632	still not a constant-time operation.
		1633
		1634	=item 2. Use a timer and re-start it with C<ev_timer_again> inactivity.
		1635
		1636	This is the easiest way, and involves using C<ev_timer_again> instead of
		1637	C<ev_timer_start>.
		1638
		1639	To implement this, configure an C<ev_timer> with a C<repeat> value
		1640	of C<60> and then call C<ev_timer_again> at start and each time you
		1641	successfully read or write some data. If you go into an idle state where
		1642	you do not expect data to travel on the socket, you can C<ev_timer_stop>
		1643	the timer, and C<ev_timer_again> will automatically restart it if need be.
		1644
		1645	That means you can ignore both the C<ev_timer_start> function and the
		1646	C<after> argument to C<ev_timer_set>, and only ever use the C<repeat>
		1647	member and C<ev_timer_again>.
		1648
		1649	At start:
		1650
		1651	ev_init (timer, callback);
		1652	timer->repeat = 60.;
		1653	ev_timer_again (loop, timer);
		1654
		1655	Each time there is some activity:
		1656
		1657	ev_timer_again (loop, timer);
		1658
		1659	It is even possible to change the time-out on the fly, regardless of
		1660	whether the watcher is active or not:
		1661
		1662	timer->repeat = 30.;
		1663	ev_timer_again (loop, timer);
		1664
		1665	This is slightly more efficient then stopping/starting the timer each time
		1666	you want to modify its timeout value, as libev does not have to completely
		1667	remove and re-insert the timer from/into its internal data structure.
		1668
		1669	It is, however, even simpler than the "obvious" way to do it.
		1670
		1671	=item 3. Let the timer time out, but then re-arm it as required.
		1672
		1673	This method is more tricky, but usually most efficient: Most timeouts are
		1674	relatively long compared to the intervals between other activity - in
		1675	our example, within 60 seconds, there are usually many I/O events with
		1676	associated activity resets.
		1677
		1678	In this case, it would be more efficient to leave the C<ev_timer> alone,
		1679	but remember the time of last activity, and check for a real timeout only
		1680	within the callback:
		1681
		1682	ev_tstamp last_activity; // time of last activity
		1683
		1684	static void
		1685	callback (EV_P_ ev_timer *w, int revents)
		1686	{
		1687	ev_tstamp now = ev_now (EV_A);
		1688	ev_tstamp timeout = last_activity + 60.;
		1689
		1690	// if last_activity + 60. is older than now, we did time out
		1691	if (timeout < now)
		1692	{
		1693	// timeout occured, take action
		1694	}
		1695	else
		1696	{
		1697	// callback was invoked, but there was some activity, re-arm
		1698	// the watcher to fire in last_activity + 60, which is
		1699	// guaranteed to be in the future, so "again" is positive:
		1700	w->repeat = timeout - now;
		1701	ev_timer_again (EV_A_ w);
		1702	}
		1703	}
		1704
		1705	To summarise the callback: first calculate the real timeout (defined
		1706	as "60 seconds after the last activity"), then check if that time has
		1707	been reached, which means something I<did>, in fact, time out. Otherwise
		1708	the callback was invoked too early (C<timeout> is in the future), so
		1709	re-schedule the timer to fire at that future time, to see if maybe we have
		1710	a timeout then.
		1711
		1712	Note how C<ev_timer_again> is used, taking advantage of the
		1713	C<ev_timer_again> optimisation when the timer is already running.
		1714
		1715	This scheme causes more callback invocations (about one every 60 seconds
		1716	minus half the average time between activity), but virtually no calls to
		1717	libev to change the timeout.
		1718
		1719	To start the timer, simply initialise the watcher and set C<last_activity>
		1720	to the current time (meaning we just have some activity :), then call the
		1721	callback, which will "do the right thing" and start the timer:
		1722
		1723	ev_init (timer, callback);
		1724	last_activity = ev_now (loop);
		1725	callback (loop, timer, EV_TIMEOUT);
		1726
		1727	And when there is some activity, simply store the current time in
		1728	C<last_activity>, no libev calls at all:
		1729
		1730	last_actiivty = ev_now (loop);
		1731
		1732	This technique is slightly more complex, but in most cases where the
		1733	time-out is unlikely to be triggered, much more efficient.
		1734
		1735	Changing the timeout is trivial as well (if it isn't hard-coded in the
		1736	callback :) - just change the timeout and invoke the callback, which will
		1737	fix things for you.
		1738
		1739	=item 4. Wee, just use a double-linked list for your timeouts.
		1740
		1741	If there is not one request, but many thousands (millions...), all
		1742	employing some kind of timeout with the same timeout value, then one can
		1743	do even better:
		1744
		1745	When starting the timeout, calculate the timeout value and put the timeout
		1746	at the I<end> of the list.
		1747
		1748	Then use an C<ev_timer> to fire when the timeout at the I<beginning> of
		1749	the list is expected to fire (for example, using the technique #3).
		1750
		1751	When there is some activity, remove the timer from the list, recalculate
		1752	the timeout, append it to the end of the list again, and make sure to
		1753	update the C<ev_timer> if it was taken from the beginning of the list.
		1754
		1755	This way, one can manage an unlimited number of timeouts in O(1) time for
		1756	starting, stopping and updating the timers, at the expense of a major
		1757	complication, and having to use a constant timeout. The constant timeout
		1758	ensures that the list stays sorted.
		1759
		1760	=back
		1761
		1762	So which method the best?
		1763
		1764	Method #2 is a simple no-brain-required solution that is adequate in most
		1765	situations. Method #3 requires a bit more thinking, but handles many cases
		1766	better, and isn't very complicated either. In most case, choosing either
		1767	one is fine, with #3 being better in typical situations.
		1768
		1769	Method #1 is almost always a bad idea, and buys you nothing. Method #4 is
		1770	rather complicated, but extremely efficient, something that really pays
		1771	off after the first million or so of active timers, i.e. it's usually
		1772	overkill :)
		1773
		1774	=head3 The special problem of time updates
		1775
		1776	Establishing the current time is a costly operation (it usually takes at
		1777	least two system calls): EV therefore updates its idea of the current
		1778	time only before and after C<ev_loop> collects new events, which causes a
		1779	growing difference between C<ev_now ()> and C<ev_time ()> when handling
		1780	lots of events in one iteration.
1108		1781
1109	The relative timeouts are calculated relative to the C<ev_now ()>	1782	The relative timeouts are calculated relative to the C<ev_now ()>
1110	time. This is usually the right thing as this timestamp refers to the time	1783	time. This is usually the right thing as this timestamp refers to the time
1111	of the event triggering whatever timeout you are modifying/starting. If	1784	of the event triggering whatever timeout you are modifying/starting. If
1112	you suspect event processing to be delayed and you I<need> to base the timeout	1785	you suspect event processing to be delayed and you I<need> to base the
1113	on the current time, use something like this to adjust for this:	1786	timeout on the current time, use something like this to adjust for this:
1114		1787
1115	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);	1788	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);
1116		1789
1117	The callback is guarenteed to be invoked only when its timeout has passed,	1790	If the event loop is suspended for a long time, you can also force an
1118	but if multiple timers become ready during the same loop iteration then	1791	update of the time returned by C<ev_now ()> by calling C<ev_now_update
1119	order of execution is undefined.	1792	()>.
		1793
		1794	=head3 The special problems of suspended animation
		1795
		1796	When you leave the server world it is quite customary to hit machines that
		1797	can suspend/hibernate - what happens to the clocks during such a suspend?
		1798
		1799	Some quick tests made with a Linux 2.6.28 indicate that a suspend freezes
		1800	all processes, while the clocks (C<times>, C<CLOCK_MONOTONIC>) continue
		1801	to run until the system is suspended, but they will not advance while the
		1802	system is suspended. That means, on resume, it will be as if the program
		1803	was frozen for a few seconds, but the suspend time will not be counted
		1804	towards C<ev_timer> when a monotonic clock source is used. The real time
		1805	clock advanced as expected, but if it is used as sole clocksource, then a
		1806	long suspend would be detected as a time jump by libev, and timers would
		1807	be adjusted accordingly.
		1808
		1809	I would not be surprised to see different behaviour in different between
		1810	operating systems, OS versions or even different hardware.
		1811
		1812	The other form of suspend (job control, or sending a SIGSTOP) will see a
		1813	time jump in the monotonic clocks and the realtime clock. If the program
		1814	is suspended for a very long time, and monotonic clock sources are in use,
		1815	then you can expect C<ev_timer>s to expire as the full suspension time
		1816	will be counted towards the timers. When no monotonic clock source is in
		1817	use, then libev will again assume a timejump and adjust accordingly.
		1818
		1819	It might be beneficial for this latter case to call C<ev_suspend>
		1820	and C<ev_resume> in code that handles C<SIGTSTP>, to at least get
		1821	deterministic behaviour in this case (you can do nothing against
		1822	C<SIGSTOP>).
1120		1823
1121	=head3 Watcher-Specific Functions and Data Members	1824	=head3 Watcher-Specific Functions and Data Members
1122		1825
1123	=over 4	1826	=over 4
1124		1827
1125	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)	1828	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)
1126		1829
1127	=item ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)	1830	=item ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)
1128		1831
1129	Configure the timer to trigger after C<after> seconds. If C<repeat> is	1832	Configure the timer to trigger after C<after> seconds. If C<repeat>
1130	C<0.>, then it will automatically be stopped. If it is positive, then the	1833	is C<0.>, then it will automatically be stopped once the timeout is
1131	timer will automatically be configured to trigger again C<repeat> seconds	1834	reached. If it is positive, then the timer will automatically be
1132	later, again, and again, until stopped manually.	1835	configured to trigger again C<repeat> seconds later, again, and again,
		1836	until stopped manually.
1133		1837
1134	The timer itself will do a best-effort at avoiding drift, that is, if you	1838	The timer itself will do a best-effort at avoiding drift, that is, if
1135	configure a timer to trigger every 10 seconds, then it will trigger at	1839	you configure a timer to trigger every 10 seconds, then it will normally
1136	exactly 10 second intervals. If, however, your program cannot keep up with	1840	trigger at exactly 10 second intervals. If, however, your program cannot
1137	the timer (because it takes longer than those 10 seconds to do stuff) the	1841	keep up with the timer (because it takes longer than those 10 seconds to
1138	timer will not fire more than once per event loop iteration.	1842	do stuff) the timer will not fire more than once per event loop iteration.
1139		1843
1140	=item ev_timer_again (loop)	1844	=item ev_timer_again (loop, ev_timer *)
1141		1845
1142	This will act as if the timer timed out and restart it again if it is	1846	This will act as if the timer timed out and restart it again if it is
1143	repeating. The exact semantics are:	1847	repeating. The exact semantics are:
1144		1848
1145	If the timer is pending, its pending status is cleared.	1849	If the timer is pending, its pending status is cleared.
1146		1850
1147	If the timer is started but nonrepeating, stop it (as if it timed out).	1851	If the timer is started but non-repeating, stop it (as if it timed out).
1148		1852
1149	If the timer is repeating, either start it if necessary (with the	1853	If the timer is repeating, either start it if necessary (with the
1150	C<repeat> value), or reset the running timer to the C<repeat> value.	1854	C<repeat> value), or reset the running timer to the C<repeat> value.
1151		1855
1152	This sounds a bit complicated, but here is a useful and typical	1856	This sounds a bit complicated, see L<Be smart about timeouts>, above, for a
1153	example: Imagine you have a tcp connection and you want a so-called idle	1857	usage example.
1154	timeout, that is, you want to be called when there have been, say, 60
1155	seconds of inactivity on the socket. The easiest way to do this is to
1156	configure an C<ev_timer> with a C<repeat> value of C<60> and then call
1157	C<ev_timer_again> each time you successfully read or write some data. If
1158	you go into an idle state where you do not expect data to travel on the
1159	socket, you can C<ev_timer_stop> the timer, and C<ev_timer_again> will
1160	automatically restart it if need be.
1161		1858
1162	That means you can ignore the C<after> value and C<ev_timer_start>	1859	=item ev_timer_remaining (loop, ev_timer *)
1163	altogether and only ever use the C<repeat> value and C<ev_timer_again>:
1164		1860
1165	ev_timer_init (timer, callback, 0., 5.);	1861	Returns the remaining time until a timer fires. If the timer is active,
1166	ev_timer_again (loop, timer);	1862	then this time is relative to the current event loop time, otherwise it's
1167	...	1863	the timeout value currently configured.
1168	timer->again = 17.;
1169	ev_timer_again (loop, timer);
1170	...
1171	timer->again = 10.;
1172	ev_timer_again (loop, timer);
1173		1864
1174	This is more slightly efficient then stopping/starting the timer each time	1865	That is, after an C<ev_timer_set (w, 5, 7)>, C<ev_timer_remaining> returns
1175	you want to modify its timeout value.	1866	C<5>. When the timer is started and one second passes, C<ev_timer_remain>
		1867	will return C<4>. When the timer expires and is restarted, it will return
		1868	roughly C<7> (likely slightly less as callback invocation takes some time,
		1869	too), and so on.
1176		1870
1177	=item ev_tstamp repeat [read-write]	1871	=item ev_tstamp repeat [read-write]
1178		1872
1179	The current C<repeat> value. Will be used each time the watcher times out	1873	The current C<repeat> value. Will be used each time the watcher times out
1180	or C<ev_timer_again> is called and determines the next timeout (if any),	1874	or C<ev_timer_again> is called, and determines the next timeout (if any),
1181	which is also when any modifications are taken into account.	1875	which is also when any modifications are taken into account.
1182		1876
1183	=back	1877	=back
1184		1878
		1879	=head3 Examples
		1880
1185	Example: Create a timer that fires after 60 seconds.	1881	Example: Create a timer that fires after 60 seconds.
1186		1882
1187	static void	1883	static void
1188	one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)	1884	one_minute_cb (struct ev_loop *loop, ev_timer *w, int revents)
1189	{	1885	{
1190	.. one minute over, w is actually stopped right here	1886	.. one minute over, w is actually stopped right here
1191	}	1887	}
1192		1888
1193	struct ev_timer mytimer;	1889	ev_timer mytimer;
1194	ev_timer_init (&mytimer, one_minute_cb, 60., 0.);	1890	ev_timer_init (&mytimer, one_minute_cb, 60., 0.);
1195	ev_timer_start (loop, &mytimer);	1891	ev_timer_start (loop, &mytimer);
1196		1892
1197	Example: Create a timeout timer that times out after 10 seconds of	1893	Example: Create a timeout timer that times out after 10 seconds of
1198	inactivity.	1894	inactivity.
1199		1895
1200	static void	1896	static void
1201	timeout_cb (struct ev_loop *loop, struct ev_timer *w, int revents)	1897	timeout_cb (struct ev_loop *loop, ev_timer *w, int revents)
1202	{	1898	{
1203	.. ten seconds without any activity	1899	.. ten seconds without any activity
1204	}	1900	}
1205		1901
1206	struct ev_timer mytimer;	1902	ev_timer mytimer;
1207	ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */	1903	ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */
1208	ev_timer_again (&mytimer); /* start timer */	1904	ev_timer_again (&mytimer); /* start timer */
1209	ev_loop (loop, 0);	1905	ev_loop (loop, 0);
1210		1906
1211	// and in some piece of code that gets executed on any "activity":	1907	// and in some piece of code that gets executed on any "activity":
1212	// reset the timeout to start ticking again at 10 seconds	1908	// reset the timeout to start ticking again at 10 seconds
1213	ev_timer_again (&mytimer);	1909	ev_timer_again (&mytimer);
1214		1910
1215		1911
1216	=head2 C<ev_periodic> - to cron or not to cron?	1912	=head2 C<ev_periodic> - to cron or not to cron?
1217		1913
1218	Periodic watchers are also timers of a kind, but they are very versatile	1914	Periodic watchers are also timers of a kind, but they are very versatile
1219	(and unfortunately a bit complex).	1915	(and unfortunately a bit complex).
1220		1916
1221	Unlike C<ev_timer>'s, they are not based on real time (or relative time)	1917	Unlike C<ev_timer>, periodic watchers are not based on real time (or
1222	but on wallclock time (absolute time). You can tell a periodic watcher	1918	relative time, the physical time that passes) but on wall clock time
1223	to trigger "at" some specific point in time. For example, if you tell a	1919	(absolute time, the thing you can read on your calender or clock). The
1224	periodic watcher to trigger in 10 seconds (by specifiying e.g. C<ev_now ()	1920	difference is that wall clock time can run faster or slower than real
1225	+ 10.>) and then reset your system clock to the last year, then it will	1921	time, and time jumps are not uncommon (e.g. when you adjust your
1226	take a year to trigger the event (unlike an C<ev_timer>, which would trigger	1922	wrist-watch).
1227	roughly 10 seconds later).
1228		1923
1229	They can also be used to implement vastly more complex timers, such as	1924	You can tell a periodic watcher to trigger after some specific point
1230	triggering an event on each midnight, local time or other, complicated,	1925	in time: for example, if you tell a periodic watcher to trigger "in 10
1231	rules.	1926	seconds" (by specifying e.g. C<ev_now () + 10.>, that is, an absolute time
		1927	not a delay) and then reset your system clock to January of the previous
		1928	year, then it will take a year or more to trigger the event (unlike an
		1929	C<ev_timer>, which would still trigger roughly 10 seconds after starting
		1930	it, as it uses a relative timeout).
1232		1931
		1932	C<ev_periodic> watchers can also be used to implement vastly more complex
		1933	timers, such as triggering an event on each "midnight, local time", or
		1934	other complicated rules. This cannot be done with C<ev_timer> watchers, as
		1935	those cannot react to time jumps.
		1936
1233	As with timers, the callback is guarenteed to be invoked only when the	1937	As with timers, the callback is guaranteed to be invoked only when the
1234	time (C<at>) has been passed, but if multiple periodic timers become ready	1938	point in time where it is supposed to trigger has passed. If multiple
1235	during the same loop iteration then order of execution is undefined.	1939	timers become ready during the same loop iteration then the ones with
		1940	earlier time-out values are invoked before ones with later time-out values
		1941	(but this is no longer true when a callback calls C<ev_loop> recursively).
1236		1942
1237	=head3 Watcher-Specific Functions and Data Members	1943	=head3 Watcher-Specific Functions and Data Members
1238		1944
1239	=over 4	1945	=over 4
1240		1946
1241	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)	1947	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp offset, ev_tstamp interval, reschedule_cb)
1242		1948
1243	=item ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)	1949	=item ev_periodic_set (ev_periodic *, ev_tstamp offset, ev_tstamp interval, reschedule_cb)
1244		1950
1245	Lots of arguments, lets sort it out... There are basically three modes of	1951	Lots of arguments, let's sort it out... There are basically three modes of
1246	operation, and we will explain them from simplest to complex:	1952	operation, and we will explain them from simplest to most complex:
1247		1953
1248	=over 4	1954	=over 4
1249		1955
1250	=item * absolute timer (at = time, interval = reschedule_cb = 0)	1956	=item * absolute timer (offset = absolute time, interval = 0, reschedule_cb = 0)
1251		1957
1252	In this configuration the watcher triggers an event at the wallclock time	1958	In this configuration the watcher triggers an event after the wall clock
1253	C<at> and doesn't repeat. It will not adjust when a time jump occurs,	1959	time C<offset> has passed. It will not repeat and will not adjust when a
1254	that is, if it is to be run at January 1st 2011 then it will run when the	1960	time jump occurs, that is, if it is to be run at January 1st 2011 then it
1255	system time reaches or surpasses this time.	1961	will be stopped and invoked when the system clock reaches or surpasses
		1962	this point in time.
1256		1963
1257	=item * non-repeating interval timer (at = offset, interval > 0, reschedule_cb = 0)	1964	=item * repeating interval timer (offset = offset within interval, interval > 0, reschedule_cb = 0)
1258		1965
1259	In this mode the watcher will always be scheduled to time out at the next	1966	In this mode the watcher will always be scheduled to time out at the next
1260	C<at + N * interval> time (for some integer N, which can also be negative)	1967	C<offset + N * interval> time (for some integer N, which can also be
1261	and then repeat, regardless of any time jumps.	1968	negative) and then repeat, regardless of any time jumps. The C<offset>
		1969	argument is merely an offset into the C<interval> periods.
1262		1970
1263	This can be used to create timers that do not drift with respect to system	1971	This can be used to create timers that do not drift with respect to the
1264	time:	1972	system clock, for example, here is an C<ev_periodic> that triggers each
		1973	hour, on the hour (with respect to UTC):
1265		1974
1266	ev_periodic_set (&periodic, 0., 3600., 0);	1975	ev_periodic_set (&periodic, 0., 3600., 0);
1267		1976
1268	This doesn't mean there will always be 3600 seconds in between triggers,	1977	This doesn't mean there will always be 3600 seconds in between triggers,
1269	but only that the the callback will be called when the system time shows a	1978	but only that the callback will be called when the system time shows a
1270	full hour (UTC), or more correctly, when the system time is evenly divisible	1979	full hour (UTC), or more correctly, when the system time is evenly divisible
1271	by 3600.	1980	by 3600.
1272		1981
1273	Another way to think about it (for the mathematically inclined) is that	1982	Another way to think about it (for the mathematically inclined) is that
1274	C<ev_periodic> will try to run the callback in this mode at the next possible	1983	C<ev_periodic> will try to run the callback in this mode at the next possible
1275	time where C<time = at (mod interval)>, regardless of any time jumps.	1984	time where C<time = offset (mod interval)>, regardless of any time jumps.
1276		1985
1277	For numerical stability it is preferable that the C<at> value is near	1986	For numerical stability it is preferable that the C<offset> value is near
1278	C<ev_now ()> (the current time), but there is no range requirement for	1987	C<ev_now ()> (the current time), but there is no range requirement for
1279	this value.	1988	this value, and in fact is often specified as zero.
1280		1989
		1990	Note also that there is an upper limit to how often a timer can fire (CPU
		1991	speed for example), so if C<interval> is very small then timing stability
		1992	will of course deteriorate. Libev itself tries to be exact to be about one
		1993	millisecond (if the OS supports it and the machine is fast enough).
		1994
1281	=item * manual reschedule mode (at and interval ignored, reschedule_cb = callback)	1995	=item * manual reschedule mode (offset ignored, interval ignored, reschedule_cb = callback)
1282		1996
1283	In this mode the values for C<interval> and C<at> are both being	1997	In this mode the values for C<interval> and C<offset> are both being
1284	ignored. Instead, each time the periodic watcher gets scheduled, the	1998	ignored. Instead, each time the periodic watcher gets scheduled, the
1285	reschedule callback will be called with the watcher as first, and the	1999	reschedule callback will be called with the watcher as first, and the
1286	current time as second argument.	2000	current time as second argument.
1287		2001
1288	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher,	2002	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher, ever,
1289	ever, or make any event loop modifications>. If you need to stop it,	2003	or make ANY other event loop modifications whatsoever, unless explicitly
1290	return C<now + 1e30> (or so, fudge fudge) and stop it afterwards (e.g. by	2004	allowed by documentation here>.
1291	starting an C<ev_prepare> watcher, which is legal).
1292		2005
		2006	If you need to stop it, return C<now + 1e30> (or so, fudge fudge) and stop
		2007	it afterwards (e.g. by starting an C<ev_prepare> watcher, which is the
		2008	only event loop modification you are allowed to do).
		2009
1293	Its prototype is C<ev_tstamp (*reschedule_cb)(struct ev_periodic *w,	2010	The callback prototype is C<ev_tstamp (*reschedule_cb)(ev_periodic
1294	ev_tstamp now)>, e.g.:	2011	*w, ev_tstamp now)>, e.g.:
1295		2012
		2013	static ev_tstamp
1296	static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)	2014	my_rescheduler (ev_periodic *w, ev_tstamp now)
1297	{	2015	{
1298	return now + 60.;	2016	return now + 60.;
1299	}	2017	}
1300		2018
1301	It must return the next time to trigger, based on the passed time value	2019	It must return the next time to trigger, based on the passed time value
1302	(that is, the lowest time value larger than to the second argument). It	2020	(that is, the lowest time value larger than to the second argument). It
1303	will usually be called just before the callback will be triggered, but	2021	will usually be called just before the callback will be triggered, but
1304	might be called at other times, too.	2022	might be called at other times, too.
1305		2023
1306	NOTE: I<< This callback must always return a time that is later than the	2024	NOTE: I<< This callback must always return a time that is higher than or
1307	passed C<now> value >>. Not even C<now> itself will do, it I<must> be larger.	2025	equal to the passed C<now> value >>.
1308		2026
1309	This can be used to create very complex timers, such as a timer that	2027	This can be used to create very complex timers, such as a timer that
1310	triggers on each midnight, local time. To do this, you would calculate the	2028	triggers on "next midnight, local time". To do this, you would calculate the
1311	next midnight after C<now> and return the timestamp value for this. How	2029	next midnight after C<now> and return the timestamp value for this. How
1312	you do this is, again, up to you (but it is not trivial, which is the main	2030	you do this is, again, up to you (but it is not trivial, which is the main
1313	reason I omitted it as an example).	2031	reason I omitted it as an example).
1314		2032
1315	=back	2033	=back
…		…
1319	Simply stops and restarts the periodic watcher again. This is only useful	2037	Simply stops and restarts the periodic watcher again. This is only useful
1320	when you changed some parameters or the reschedule callback would return	2038	when you changed some parameters or the reschedule callback would return
1321	a different time than the last time it was called (e.g. in a crond like	2039	a different time than the last time it was called (e.g. in a crond like
1322	program when the crontabs have changed).	2040	program when the crontabs have changed).
1323		2041
		2042	=item ev_tstamp ev_periodic_at (ev_periodic *)
		2043
		2044	When active, returns the absolute time that the watcher is supposed
		2045	to trigger next. This is not the same as the C<offset> argument to
		2046	C<ev_periodic_set>, but indeed works even in interval and manual
		2047	rescheduling modes.
		2048
1324	=item ev_tstamp offset [read-write]	2049	=item ev_tstamp offset [read-write]
1325		2050
1326	When repeating, this contains the offset value, otherwise this is the	2051	When repeating, this contains the offset value, otherwise this is the
1327	absolute point in time (the C<at> value passed to C<ev_periodic_set>).	2052	absolute point in time (the C<offset> value passed to C<ev_periodic_set>,
		2053	although libev might modify this value for better numerical stability).
1328		2054
1329	Can be modified any time, but changes only take effect when the periodic	2055	Can be modified any time, but changes only take effect when the periodic
1330	timer fires or C<ev_periodic_again> is being called.	2056	timer fires or C<ev_periodic_again> is being called.
1331		2057
1332	=item ev_tstamp interval [read-write]	2058	=item ev_tstamp interval [read-write]
1333		2059
1334	The current interval value. Can be modified any time, but changes only	2060	The current interval value. Can be modified any time, but changes only
1335	take effect when the periodic timer fires or C<ev_periodic_again> is being	2061	take effect when the periodic timer fires or C<ev_periodic_again> is being
1336	called.	2062	called.
1337		2063
1338	=item ev_tstamp (*reschedule_cb)(struct ev_periodic *w, ev_tstamp now) [read-write]	2064	=item ev_tstamp (*reschedule_cb)(ev_periodic *w, ev_tstamp now) [read-write]
1339		2065
1340	The current reschedule callback, or C<0>, if this functionality is	2066	The current reschedule callback, or C<0>, if this functionality is
1341	switched off. Can be changed any time, but changes only take effect when	2067	switched off. Can be changed any time, but changes only take effect when
1342	the periodic timer fires or C<ev_periodic_again> is being called.	2068	the periodic timer fires or C<ev_periodic_again> is being called.
1343		2069
1344	=item ev_tstamp at [read-only]
1345
1346	When active, contains the absolute time that the watcher is supposed to
1347	trigger next.
1348
1349	=back	2070	=back
1350		2071
		2072	=head3 Examples
		2073
1351	Example: Call a callback every hour, or, more precisely, whenever the	2074	Example: Call a callback every hour, or, more precisely, whenever the
1352	system clock is divisible by 3600. The callback invocation times have	2075	system time is divisible by 3600. The callback invocation times have
1353	potentially a lot of jittering, but good long-term stability.	2076	potentially a lot of jitter, but good long-term stability.
1354		2077
1355	static void	2078	static void
1356	clock_cb (struct ev_loop *loop, struct ev_io *w, int revents)	2079	clock_cb (struct ev_loop *loop, ev_io *w, int revents)
1357	{	2080	{
1358	... its now a full hour (UTC, or TAI or whatever your clock follows)	2081	... its now a full hour (UTC, or TAI or whatever your clock follows)
1359	}	2082	}
1360		2083
1361	struct ev_periodic hourly_tick;	2084	ev_periodic hourly_tick;
1362	ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);	2085	ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);
1363	ev_periodic_start (loop, &hourly_tick);	2086	ev_periodic_start (loop, &hourly_tick);
1364		2087
1365	Example: The same as above, but use a reschedule callback to do it:	2088	Example: The same as above, but use a reschedule callback to do it:
1366		2089
1367	#include <math.h>	2090	#include <math.h>
1368		2091
1369	static ev_tstamp	2092	static ev_tstamp
1370	my_scheduler_cb (struct ev_periodic *w, ev_tstamp now)	2093	my_scheduler_cb (ev_periodic *w, ev_tstamp now)
1371	{	2094	{
1372	return fmod (now, 3600.) + 3600.;	2095	return now + (3600. - fmod (now, 3600.));
1373	}	2096	}
1374		2097
1375	ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);	2098	ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
1376		2099
1377	Example: Call a callback every hour, starting now:	2100	Example: Call a callback every hour, starting now:
1378		2101
1379	struct ev_periodic hourly_tick;	2102	ev_periodic hourly_tick;
1380	ev_periodic_init (&hourly_tick, clock_cb,	2103	ev_periodic_init (&hourly_tick, clock_cb,
1381	fmod (ev_now (loop), 3600.), 3600., 0);	2104	fmod (ev_now (loop), 3600.), 3600., 0);
1382	ev_periodic_start (loop, &hourly_tick);	2105	ev_periodic_start (loop, &hourly_tick);
1383		2106
1384		2107
1385	=head2 C<ev_signal> - signal me when a signal gets signalled!	2108	=head2 C<ev_signal> - signal me when a signal gets signalled!
1386		2109
1387	Signal watchers will trigger an event when the process receives a specific	2110	Signal watchers will trigger an event when the process receives a specific
1388	signal one or more times. Even though signals are very asynchronous, libev	2111	signal one or more times. Even though signals are very asynchronous, libev
1389	will try it's best to deliver signals synchronously, i.e. as part of the	2112	will try it's best to deliver signals synchronously, i.e. as part of the
1390	normal event processing, like any other event.	2113	normal event processing, like any other event.
1391		2114
		2115	If you want signals to be delivered truly asynchronously, just use
		2116	C<sigaction> as you would do without libev and forget about sharing
		2117	the signal. You can even use C<ev_async> from a signal handler to
		2118	synchronously wake up an event loop.
		2119
1392	You can configure as many watchers as you like per signal. Only when the	2120	You can configure as many watchers as you like for the same signal, but
		2121	only within the same loop, i.e. you can watch for C<SIGINT> in your
		2122	default loop and for C<SIGIO> in another loop, but you cannot watch for
		2123	C<SIGINT> in both the default loop and another loop at the same time. At
		2124	the moment, C<SIGCHLD> is permanently tied to the default loop.
		2125
1393	first watcher gets started will libev actually register a signal watcher	2126	When the first watcher gets started will libev actually register something
1394	with the kernel (thus it coexists with your own signal handlers as long	2127	with the kernel (thus it coexists with your own signal handlers as long as
1395	as you don't register any with libev). Similarly, when the last signal	2128	you don't register any with libev for the same signal).
1396	watcher for a signal is stopped libev will reset the signal handler to	2129
1397	SIG_DFL (regardless of what it was set to before).	2130	If possible and supported, libev will install its handlers with
		2131	C<SA_RESTART> (or equivalent) behaviour enabled, so system calls should
		2132	not be unduly interrupted. If you have a problem with system calls getting
		2133	interrupted by signals you can block all signals in an C<ev_check> watcher
		2134	and unblock them in an C<ev_prepare> watcher.
		2135
		2136	=head3 The special problem of inheritance over execve
		2137
		2138	Both the signal mask (C<sigprocmask>) and the signal disposition
		2139	(C<sigaction>) are unspecified after starting a signal watcher (and after
		2140	stopping it again), that is, libev might or might not block the signal,
		2141	and might or might not set or restore the installed signal handler.
		2142
		2143	While this does not matter for the signal disposition (libev never
		2144	sets signals to C<SIG_IGN>, so handlers will be reset to C<SIG_DFL> on
		2145	C<execve>), this matters for the signal mask: many programs do not expect
		2146	certain signals to be blocked.
		2147
		2148	This means that before calling C<exec> (from the child) you should reset
		2149	the signal mask to whatever "default" you expect (all clear is a good
		2150	choice usually).
		2151
		2152	The simplest way to ensure that the signal mask is reset in the child is
		2153	to install a fork handler with C<pthread_atfork> that resets it. That will
		2154	catch fork calls done by libraries (such as the libc) as well.
		2155
		2156	In current versions of libev, you can also ensure that the signal mask is
		2157	not blocking any signals (except temporarily, so thread users watch out)
		2158	by specifying the C<EVFLAG_NOSIGFD> when creating the event loop. This
		2159	is not guaranteed for future versions, however.
1398		2160
1399	=head3 Watcher-Specific Functions and Data Members	2161	=head3 Watcher-Specific Functions and Data Members
1400		2162
1401	=over 4	2163	=over 4
1402		2164
…		…
1411		2173
1412	The signal the watcher watches out for.	2174	The signal the watcher watches out for.
1413		2175
1414	=back	2176	=back
1415		2177
		2178	=head3 Examples
		2179
		2180	Example: Try to exit cleanly on SIGINT.
		2181
		2182	static void
		2183	sigint_cb (struct ev_loop *loop, ev_signal *w, int revents)
		2184	{
		2185	ev_unloop (loop, EVUNLOOP_ALL);
		2186	}
		2187
		2188	ev_signal signal_watcher;
		2189	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
		2190	ev_signal_start (loop, &signal_watcher);
		2191
1416		2192
1417	=head2 C<ev_child> - watch out for process status changes	2193	=head2 C<ev_child> - watch out for process status changes
1418		2194
1419	Child watchers trigger when your process receives a SIGCHLD in response to	2195	Child watchers trigger when your process receives a SIGCHLD in response to
1420	some child status changes (most typically when a child of yours dies).	2196	some child status changes (most typically when a child of yours dies or
		2197	exits). It is permissible to install a child watcher I<after> the child
		2198	has been forked (which implies it might have already exited), as long
		2199	as the event loop isn't entered (or is continued from a watcher), i.e.,
		2200	forking and then immediately registering a watcher for the child is fine,
		2201	but forking and registering a watcher a few event loop iterations later or
		2202	in the next callback invocation is not.
		2203
		2204	Only the default event loop is capable of handling signals, and therefore
		2205	you can only register child watchers in the default event loop.
		2206
		2207	Due to some design glitches inside libev, child watchers will always be
		2208	handled at maximum priority (their priority is set to C<EV_MAXPRI> by
		2209	libev)
		2210
		2211	=head3 Process Interaction
		2212
		2213	Libev grabs C<SIGCHLD> as soon as the default event loop is
		2214	initialised. This is necessary to guarantee proper behaviour even if the
		2215	first child watcher is started after the child exits. The occurrence
		2216	of C<SIGCHLD> is recorded asynchronously, but child reaping is done
		2217	synchronously as part of the event loop processing. Libev always reaps all
		2218	children, even ones not watched.
		2219
		2220	=head3 Overriding the Built-In Processing
		2221
		2222	Libev offers no special support for overriding the built-in child
		2223	processing, but if your application collides with libev's default child
		2224	handler, you can override it easily by installing your own handler for
		2225	C<SIGCHLD> after initialising the default loop, and making sure the
		2226	default loop never gets destroyed. You are encouraged, however, to use an
		2227	event-based approach to child reaping and thus use libev's support for
		2228	that, so other libev users can use C<ev_child> watchers freely.
		2229
		2230	=head3 Stopping the Child Watcher
		2231
		2232	Currently, the child watcher never gets stopped, even when the
		2233	child terminates, so normally one needs to stop the watcher in the
		2234	callback. Future versions of libev might stop the watcher automatically
		2235	when a child exit is detected (calling C<ev_child_stop> twice is not a
		2236	problem).
1421		2237
1422	=head3 Watcher-Specific Functions and Data Members	2238	=head3 Watcher-Specific Functions and Data Members
1423		2239
1424	=over 4	2240	=over 4
1425		2241
1426	=item ev_child_init (ev_child *, callback, int pid)	2242	=item ev_child_init (ev_child *, callback, int pid, int trace)
1427		2243
1428	=item ev_child_set (ev_child *, int pid)	2244	=item ev_child_set (ev_child *, int pid, int trace)
1429		2245
1430	Configures the watcher to wait for status changes of process C<pid> (or	2246	Configures the watcher to wait for status changes of process C<pid> (or
1431	I<any> process if C<pid> is specified as C<0>). The callback can look	2247	I<any> process if C<pid> is specified as C<0>). The callback can look
1432	at the C<rstatus> member of the C<ev_child> watcher structure to see	2248	at the C<rstatus> member of the C<ev_child> watcher structure to see
1433	the status word (use the macros from C<sys/wait.h> and see your systems	2249	the status word (use the macros from C<sys/wait.h> and see your systems
1434	C<waitpid> documentation). The C<rpid> member contains the pid of the	2250	C<waitpid> documentation). The C<rpid> member contains the pid of the
1435	process causing the status change.	2251	process causing the status change. C<trace> must be either C<0> (only
		2252	activate the watcher when the process terminates) or C<1> (additionally
		2253	activate the watcher when the process is stopped or continued).
1436		2254
1437	=item int pid [read-only]	2255	=item int pid [read-only]
1438		2256
1439	The process id this watcher watches out for, or C<0>, meaning any process id.	2257	The process id this watcher watches out for, or C<0>, meaning any process id.
1440		2258
…		…
1447	The process exit/trace status caused by C<rpid> (see your systems	2265	The process exit/trace status caused by C<rpid> (see your systems
1448	C<waitpid> and C<sys/wait.h> documentation for details).	2266	C<waitpid> and C<sys/wait.h> documentation for details).
1449		2267
1450	=back	2268	=back
1451		2269
1452	Example: Try to exit cleanly on SIGINT and SIGTERM.	2270	=head3 Examples
1453		2271
		2272	Example: C<fork()> a new process and install a child handler to wait for
		2273	its completion.
		2274
		2275	ev_child cw;
		2276
1454	static void	2277	static void
1455	sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)	2278	child_cb (EV_P_ ev_child *w, int revents)
1456	{	2279	{
1457	ev_unloop (loop, EVUNLOOP_ALL);	2280	ev_child_stop (EV_A_ w);
		2281	printf ("process %d exited with status %x\n", w->rpid, w->rstatus);
1458	}	2282	}
1459		2283
1460	struct ev_signal signal_watcher;	2284	pid_t pid = fork ();
1461	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);	2285
1462	ev_signal_start (loop, &sigint_cb);	2286	if (pid < 0)
		2287	// error
		2288	else if (pid == 0)
		2289	{
		2290	// the forked child executes here
		2291	exit (1);
		2292	}
		2293	else
		2294	{
		2295	ev_child_init (&cw, child_cb, pid, 0);
		2296	ev_child_start (EV_DEFAULT_ &cw);
		2297	}
1463		2298
1464		2299
1465	=head2 C<ev_stat> - did the file attributes just change?	2300	=head2 C<ev_stat> - did the file attributes just change?
1466		2301
1467	This watches a filesystem path for attribute changes. That is, it calls	2302	This watches a file system path for attribute changes. That is, it calls
1468	C<stat> regularly (or when the OS says it changed) and sees if it changed	2303	C<stat> on that path in regular intervals (or when the OS says it changed)
1469	compared to the last time, invoking the callback if it did.	2304	and sees if it changed compared to the last time, invoking the callback if
		2305	it did.
1470		2306
1471	The path does not need to exist: changing from "path exists" to "path does	2307	The path does not need to exist: changing from "path exists" to "path does
1472	not exist" is a status change like any other. The condition "path does	2308	not exist" is a status change like any other. The condition "path does not
1473	not exist" is signified by the C<st_nlink> field being zero (which is	2309	exist" (or more correctly "path cannot be stat'ed") is signified by the
1474	otherwise always forced to be at least one) and all the other fields of	2310	C<st_nlink> field being zero (which is otherwise always forced to be at
1475	the stat buffer having unspecified contents.	2311	least one) and all the other fields of the stat buffer having unspecified
		2312	contents.
1476		2313
1477	The path I<should> be absolute and I<must not> end in a slash. If it is	2314	The path I<must not> end in a slash or contain special components such as
		2315	C<.> or C<..>. The path I<should> be absolute: If it is relative and
1478	relative and your working directory changes, the behaviour is undefined.	2316	your working directory changes, then the behaviour is undefined.
1479		2317
1480	Since there is no standard to do this, the portable implementation simply	2318	Since there is no portable change notification interface available, the
1481	calls C<stat (2)> regularly on the path to see if it changed somehow. You	2319	portable implementation simply calls C<stat(2)> regularly on the path
1482	can specify a recommended polling interval for this case. If you specify	2320	to see if it changed somehow. You can specify a recommended polling
1483	a polling interval of C<0> (highly recommended!) then a I<suitable,	2321	interval for this case. If you specify a polling interval of C<0> (highly
1484	unspecified default> value will be used (which you can expect to be around	2322	recommended!) then a I<suitable, unspecified default> value will be used
1485	five seconds, although this might change dynamically). Libev will also	2323	(which you can expect to be around five seconds, although this might
1486	impose a minimum interval which is currently around C<0.1>, but thats	2324	change dynamically). Libev will also impose a minimum interval which is
1487	usually overkill.	2325	currently around C<0.1>, but that's usually overkill.
1488		2326
1489	This watcher type is not meant for massive numbers of stat watchers,	2327	This watcher type is not meant for massive numbers of stat watchers,
1490	as even with OS-supported change notifications, this can be	2328	as even with OS-supported change notifications, this can be
1491	resource-intensive.	2329	resource-intensive.
1492		2330
1493	At the time of this writing, only the Linux inotify interface is	2331	At the time of this writing, the only OS-specific interface implemented
1494	implemented (implementing kqueue support is left as an exercise for the	2332	is the Linux inotify interface (implementing kqueue support is left as an
1495	reader). Inotify will be used to give hints only and should not change the	2333	exercise for the reader. Note, however, that the author sees no way of
1496	semantics of C<ev_stat> watchers, which means that libev sometimes needs	2334	implementing C<ev_stat> semantics with kqueue, except as a hint).
1497	to fall back to regular polling again even with inotify, but changes are
1498	usually detected immediately, and if the file exists there will be no
1499	polling.
1500		2335
1501	=head3 Inotify	2336	=head3 ABI Issues (Largefile Support)
1502		2337
		2338	Libev by default (unless the user overrides this) uses the default
		2339	compilation environment, which means that on systems with large file
		2340	support disabled by default, you get the 32 bit version of the stat
		2341	structure. When using the library from programs that change the ABI to
		2342	use 64 bit file offsets the programs will fail. In that case you have to
		2343	compile libev with the same flags to get binary compatibility. This is
		2344	obviously the case with any flags that change the ABI, but the problem is
		2345	most noticeably displayed with ev_stat and large file support.
		2346
		2347	The solution for this is to lobby your distribution maker to make large
		2348	file interfaces available by default (as e.g. FreeBSD does) and not
		2349	optional. Libev cannot simply switch on large file support because it has
		2350	to exchange stat structures with application programs compiled using the
		2351	default compilation environment.
		2352
		2353	=head3 Inotify and Kqueue
		2354
1503	When C<inotify (7)> support has been compiled into libev (generally only	2355	When C<inotify (7)> support has been compiled into libev and present at
1504	available on Linux) and present at runtime, it will be used to speed up	2356	runtime, it will be used to speed up change detection where possible. The
1505	change detection where possible. The inotify descriptor will be created lazily	2357	inotify descriptor will be created lazily when the first C<ev_stat>
1506	when the first C<ev_stat> watcher is being started.	2358	watcher is being started.
1507		2359
1508	Inotify presense does not change the semantics of C<ev_stat> watchers	2360	Inotify presence does not change the semantics of C<ev_stat> watchers
1509	except that changes might be detected earlier, and in some cases, to avoid	2361	except that changes might be detected earlier, and in some cases, to avoid
1510	making regular C<stat> calls. Even in the presense of inotify support	2362	making regular C<stat> calls. Even in the presence of inotify support
1511	there are many cases where libev has to resort to regular C<stat> polling.	2363	there are many cases where libev has to resort to regular C<stat> polling,
		2364	but as long as kernel 2.6.25 or newer is used (2.6.24 and older have too
		2365	many bugs), the path exists (i.e. stat succeeds), and the path resides on
		2366	a local filesystem (libev currently assumes only ext2/3, jfs, reiserfs and
		2367	xfs are fully working) libev usually gets away without polling.
1512		2368
1513	(There is no support for kqueue, as apparently it cannot be used to	2369	There is no support for kqueue, as apparently it cannot be used to
1514	implement this functionality, due to the requirement of having a file	2370	implement this functionality, due to the requirement of having a file
1515	descriptor open on the object at all times).	2371	descriptor open on the object at all times, and detecting renames, unlinks
		2372	etc. is difficult.
		2373
		2374	=head3 C<stat ()> is a synchronous operation
		2375
		2376	Libev doesn't normally do any kind of I/O itself, and so is not blocking
		2377	the process. The exception are C<ev_stat> watchers - those call C<stat
		2378	()>, which is a synchronous operation.
		2379
		2380	For local paths, this usually doesn't matter: unless the system is very
		2381	busy or the intervals between stat's are large, a stat call will be fast,
		2382	as the path data is usually in memory already (except when starting the
		2383	watcher).
		2384
		2385	For networked file systems, calling C<stat ()> can block an indefinite
		2386	time due to network issues, and even under good conditions, a stat call
		2387	often takes multiple milliseconds.
		2388
		2389	Therefore, it is best to avoid using C<ev_stat> watchers on networked
		2390	paths, although this is fully supported by libev.
1516		2391
1517	=head3 The special problem of stat time resolution	2392	=head3 The special problem of stat time resolution
1518		2393
1519	The C<stat ()> syscall only supports full-second resolution portably, and	2394	The C<stat ()> system call only supports full-second resolution portably,
1520	even on systems where the resolution is higher, many filesystems still	2395	and even on systems where the resolution is higher, most file systems
1521	only support whole seconds.	2396	still only support whole seconds.
1522		2397
1523	That means that, if the time is the only thing that changes, you might	2398	That means that, if the time is the only thing that changes, you can
1524	miss updates: on the first update, C<ev_stat> detects a change and calls	2399	easily miss updates: on the first update, C<ev_stat> detects a change and
1525	your callback, which does something. When there is another update within	2400	calls your callback, which does something. When there is another update
1526	the same second, C<ev_stat> will be unable to detect it.	2401	within the same second, C<ev_stat> will be unable to detect unless the
		2402	stat data does change in other ways (e.g. file size).
1527		2403
1528	The solution to this is to delay acting on a change for a second (or till	2404	The solution to this is to delay acting on a change for slightly more
1529	the next second boundary), using a roughly one-second delay C<ev_timer>	2405	than a second (or till slightly after the next full second boundary), using
1530	(C<ev_timer_set (w, 0., 1.01); ev_timer_again (loop, w)>). The C<.01>	2406	a roughly one-second-delay C<ev_timer> (e.g. C<ev_timer_set (w, 0., 1.02);
1531	is added to work around small timing inconsistencies of some operating	2407	ev_timer_again (loop, w)>).
1532	systems.	2408
		2409	The C<.02> offset is added to work around small timing inconsistencies
		2410	of some operating systems (where the second counter of the current time
		2411	might be be delayed. One such system is the Linux kernel, where a call to
		2412	C<gettimeofday> might return a timestamp with a full second later than
		2413	a subsequent C<time> call - if the equivalent of C<time ()> is used to
		2414	update file times then there will be a small window where the kernel uses
		2415	the previous second to update file times but libev might already execute
		2416	the timer callback).
1533		2417
1534	=head3 Watcher-Specific Functions and Data Members	2418	=head3 Watcher-Specific Functions and Data Members
1535		2419
1536	=over 4	2420	=over 4
1537		2421
…		…
1543	C<path>. The C<interval> is a hint on how quickly a change is expected to	2427	C<path>. The C<interval> is a hint on how quickly a change is expected to
1544	be detected and should normally be specified as C<0> to let libev choose	2428	be detected and should normally be specified as C<0> to let libev choose
1545	a suitable value. The memory pointed to by C<path> must point to the same	2429	a suitable value. The memory pointed to by C<path> must point to the same
1546	path for as long as the watcher is active.	2430	path for as long as the watcher is active.
1547		2431
1548	The callback will be receive C<EV_STAT> when a change was detected,	2432	The callback will receive an C<EV_STAT> event when a change was detected,
1549	relative to the attributes at the time the watcher was started (or the	2433	relative to the attributes at the time the watcher was started (or the
1550	last change was detected).	2434	last change was detected).
1551		2435
1552	=item ev_stat_stat (ev_stat *)	2436	=item ev_stat_stat (loop, ev_stat *)
1553		2437
1554	Updates the stat buffer immediately with new values. If you change the	2438	Updates the stat buffer immediately with new values. If you change the
1555	watched path in your callback, you could call this fucntion to avoid	2439	watched path in your callback, you could call this function to avoid
1556	detecting this change (while introducing a race condition). Can also be	2440	detecting this change (while introducing a race condition if you are not
1557	useful simply to find out the new values.	2441	the only one changing the path). Can also be useful simply to find out the
		2442	new values.
1558		2443
1559	=item ev_statdata attr [read-only]	2444	=item ev_statdata attr [read-only]
1560		2445
1561	The most-recently detected attributes of the file. Although the type is of	2446	The most-recently detected attributes of the file. Although the type is
1562	C<ev_statdata>, this is usually the (or one of the) C<struct stat> types	2447	C<ev_statdata>, this is usually the (or one of the) C<struct stat> types
1563	suitable for your system. If the C<st_nlink> member is C<0>, then there	2448	suitable for your system, but you can only rely on the POSIX-standardised
		2449	members to be present. If the C<st_nlink> member is C<0>, then there was
1564	was some error while C<stat>ing the file.	2450	some error while C<stat>ing the file.
1565		2451
1566	=item ev_statdata prev [read-only]	2452	=item ev_statdata prev [read-only]
1567		2453
1568	The previous attributes of the file. The callback gets invoked whenever	2454	The previous attributes of the file. The callback gets invoked whenever
1569	C<prev> != C<attr>.	2455	C<prev> != C<attr>, or, more precisely, one or more of these members
		2456	differ: C<st_dev>, C<st_ino>, C<st_mode>, C<st_nlink>, C<st_uid>,
		2457	C<st_gid>, C<st_rdev>, C<st_size>, C<st_atime>, C<st_mtime>, C<st_ctime>.
1570		2458
1571	=item ev_tstamp interval [read-only]	2459	=item ev_tstamp interval [read-only]
1572		2460
1573	The specified interval.	2461	The specified interval.
1574		2462
1575	=item const char *path [read-only]	2463	=item const char *path [read-only]
1576		2464
1577	The filesystem path that is being watched.	2465	The file system path that is being watched.
1578		2466
1579	=back	2467	=back
1580		2468
1581	=head3 Examples	2469	=head3 Examples
1582		2470
1583	Example: Watch C</etc/passwd> for attribute changes.	2471	Example: Watch C</etc/passwd> for attribute changes.
1584		2472
1585	static void	2473	static void
1586	passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)	2474	passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)
1587	{	2475	{
1588	/* /etc/passwd changed in some way */	2476	/* /etc/passwd changed in some way */
1589	if (w->attr.st_nlink)	2477	if (w->attr.st_nlink)
1590	{	2478	{
1591	printf ("passwd current size %ld\n", (long)w->attr.st_size);	2479	printf ("passwd current size %ld\n", (long)w->attr.st_size);
1592	printf ("passwd current atime %ld\n", (long)w->attr.st_mtime);	2480	printf ("passwd current atime %ld\n", (long)w->attr.st_mtime);
1593	printf ("passwd current mtime %ld\n", (long)w->attr.st_mtime);	2481	printf ("passwd current mtime %ld\n", (long)w->attr.st_mtime);
1594	}	2482	}
1595	else	2483	else
1596	/* you shalt not abuse printf for puts */	2484	/* you shalt not abuse printf for puts */
1597	puts ("wow, /etc/passwd is not there, expect problems. "	2485	puts ("wow, /etc/passwd is not there, expect problems. "
1598	"if this is windows, they already arrived\n");	2486	"if this is windows, they already arrived\n");
1599	}	2487	}
1600		2488
1601	...	2489	...
1602	ev_stat passwd;	2490	ev_stat passwd;
1603		2491
1604	ev_stat_init (&passwd, passwd_cb, "/etc/passwd", 0.);	2492	ev_stat_init (&passwd, passwd_cb, "/etc/passwd", 0.);
1605	ev_stat_start (loop, &passwd);	2493	ev_stat_start (loop, &passwd);
1606		2494
1607	Example: Like above, but additionally use a one-second delay so we do not	2495	Example: Like above, but additionally use a one-second delay so we do not
1608	miss updates (however, frequent updates will delay processing, too, so	2496	miss updates (however, frequent updates will delay processing, too, so
1609	one might do the work both on C<ev_stat> callback invocation I<and> on	2497	one might do the work both on C<ev_stat> callback invocation I<and> on
1610	C<ev_timer> callback invocation).	2498	C<ev_timer> callback invocation).
1611		2499
1612	static ev_stat passwd;	2500	static ev_stat passwd;
1613	static ev_timer timer;	2501	static ev_timer timer;
1614		2502
1615	static void	2503	static void
1616	timer_cb (EV_P_ ev_timer *w, int revents)	2504	timer_cb (EV_P_ ev_timer *w, int revents)
1617	{	2505	{
1618	ev_timer_stop (EV_A_ w);	2506	ev_timer_stop (EV_A_ w);
1619		2507
1620	/* now it's one second after the most recent passwd change */	2508	/* now it's one second after the most recent passwd change */
1621	}	2509	}
1622		2510
1623	static void	2511	static void
1624	stat_cb (EV_P_ ev_stat *w, int revents)	2512	stat_cb (EV_P_ ev_stat *w, int revents)
1625	{	2513	{
1626	/* reset the one-second timer */	2514	/* reset the one-second timer */
1627	ev_timer_again (EV_A_ &timer);	2515	ev_timer_again (EV_A_ &timer);
1628	}	2516	}
1629		2517
1630	...	2518	...
1631	ev_stat_init (&passwd, stat_cb, "/etc/passwd", 0.);	2519	ev_stat_init (&passwd, stat_cb, "/etc/passwd", 0.);
1632	ev_stat_start (loop, &passwd);	2520	ev_stat_start (loop, &passwd);
1633	ev_timer_init (&timer, timer_cb, 0., 1.01);	2521	ev_timer_init (&timer, timer_cb, 0., 1.02);
1634		2522
1635		2523
1636	=head2 C<ev_idle> - when you've got nothing better to do...	2524	=head2 C<ev_idle> - when you've got nothing better to do...
1637		2525
1638	Idle watchers trigger events when no other events of the same or higher	2526	Idle watchers trigger events when no other events of the same or higher
1639	priority are pending (prepare, check and other idle watchers do not	2527	priority are pending (prepare, check and other idle watchers do not count
1640	count).	2528	as receiving "events").
1641		2529
1642	That is, as long as your process is busy handling sockets or timeouts	2530	That is, as long as your process is busy handling sockets or timeouts
1643	(or even signals, imagine) of the same or higher priority it will not be	2531	(or even signals, imagine) of the same or higher priority it will not be
1644	triggered. But when your process is idle (or only lower-priority watchers	2532	triggered. But when your process is idle (or only lower-priority watchers
1645	are pending), the idle watchers are being called once per event loop	2533	are pending), the idle watchers are being called once per event loop
…		…
1656		2544
1657	=head3 Watcher-Specific Functions and Data Members	2545	=head3 Watcher-Specific Functions and Data Members
1658		2546
1659	=over 4	2547	=over 4
1660		2548
1661	=item ev_idle_init (ev_signal *, callback)	2549	=item ev_idle_init (ev_idle *, callback)
1662		2550
1663	Initialises and configures the idle watcher - it has no parameters of any	2551	Initialises and configures the idle watcher - it has no parameters of any
1664	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,	2552	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,
1665	believe me.	2553	believe me.
1666		2554
1667	=back	2555	=back
1668		2556
		2557	=head3 Examples
		2558
1669	Example: Dynamically allocate an C<ev_idle> watcher, start it, and in the	2559	Example: Dynamically allocate an C<ev_idle> watcher, start it, and in the
1670	callback, free it. Also, use no error checking, as usual.	2560	callback, free it. Also, use no error checking, as usual.
1671		2561
1672	static void	2562	static void
1673	idle_cb (struct ev_loop *loop, struct ev_idle *w, int revents)	2563	idle_cb (struct ev_loop *loop, ev_idle *w, int revents)
1674	{	2564	{
1675	free (w);	2565	free (w);
1676	// now do something you wanted to do when the program has	2566	// now do something you wanted to do when the program has
1677	// no longer asnything immediate to do.	2567	// no longer anything immediate to do.
1678	}	2568	}
1679		2569
1680	struct ev_idle *idle_watcher = malloc (sizeof (struct ev_idle));	2570	ev_idle *idle_watcher = malloc (sizeof (ev_idle));
1681	ev_idle_init (idle_watcher, idle_cb);	2571	ev_idle_init (idle_watcher, idle_cb);
1682	ev_idle_start (loop, idle_cb);	2572	ev_idle_start (loop, idle_watcher);
1683		2573
1684		2574
1685	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!	2575	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!
1686		2576
1687	Prepare and check watchers are usually (but not always) used in tandem:	2577	Prepare and check watchers are usually (but not always) used in pairs:
1688	prepare watchers get invoked before the process blocks and check watchers	2578	prepare watchers get invoked before the process blocks and check watchers
1689	afterwards.	2579	afterwards.
1690		2580
1691	You I<must not> call C<ev_loop> or similar functions that enter	2581	You I<must not> call C<ev_loop> or similar functions that enter
1692	the current event loop from either C<ev_prepare> or C<ev_check>	2582	the current event loop from either C<ev_prepare> or C<ev_check>
…		…
1695	those watchers, i.e. the sequence will always be C<ev_prepare>, blocking,	2585	those watchers, i.e. the sequence will always be C<ev_prepare>, blocking,
1696	C<ev_check> so if you have one watcher of each kind they will always be	2586	C<ev_check> so if you have one watcher of each kind they will always be
1697	called in pairs bracketing the blocking call.	2587	called in pairs bracketing the blocking call.
1698		2588
1699	Their main purpose is to integrate other event mechanisms into libev and	2589	Their main purpose is to integrate other event mechanisms into libev and
1700	their use is somewhat advanced. This could be used, for example, to track	2590	their use is somewhat advanced. They could be used, for example, to track
1701	variable changes, implement your own watchers, integrate net-snmp or a	2591	variable changes, implement your own watchers, integrate net-snmp or a
1702	coroutine library and lots more. They are also occasionally useful if	2592	coroutine library and lots more. They are also occasionally useful if
1703	you cache some data and want to flush it before blocking (for example,	2593	you cache some data and want to flush it before blocking (for example,
1704	in X programs you might want to do an C<XFlush ()> in an C<ev_prepare>	2594	in X programs you might want to do an C<XFlush ()> in an C<ev_prepare>
1705	watcher).	2595	watcher).
1706		2596
1707	This is done by examining in each prepare call which file descriptors need	2597	This is done by examining in each prepare call which file descriptors
1708	to be watched by the other library, registering C<ev_io> watchers for	2598	need to be watched by the other library, registering C<ev_io> watchers
1709	them and starting an C<ev_timer> watcher for any timeouts (many libraries	2599	for them and starting an C<ev_timer> watcher for any timeouts (many
1710	provide just this functionality). Then, in the check watcher you check for	2600	libraries provide exactly this functionality). Then, in the check watcher,
1711	any events that occured (by checking the pending status of all watchers	2601	you check for any events that occurred (by checking the pending status
1712	and stopping them) and call back into the library. The I/O and timer	2602	of all watchers and stopping them) and call back into the library. The
1713	callbacks will never actually be called (but must be valid nevertheless,	2603	I/O and timer callbacks will never actually be called (but must be valid
1714	because you never know, you know?).	2604	nevertheless, because you never know, you know?).
1715		2605
1716	As another example, the Perl Coro module uses these hooks to integrate	2606	As another example, the Perl Coro module uses these hooks to integrate
1717	coroutines into libev programs, by yielding to other active coroutines	2607	coroutines into libev programs, by yielding to other active coroutines
1718	during each prepare and only letting the process block if no coroutines	2608	during each prepare and only letting the process block if no coroutines
1719	are ready to run (it's actually more complicated: it only runs coroutines	2609	are ready to run (it's actually more complicated: it only runs coroutines
…		…
1722	loop from blocking if lower-priority coroutines are active, thus mapping	2612	loop from blocking if lower-priority coroutines are active, thus mapping
1723	low-priority coroutines to idle/background tasks).	2613	low-priority coroutines to idle/background tasks).
1724		2614
1725	It is recommended to give C<ev_check> watchers highest (C<EV_MAXPRI>)	2615	It is recommended to give C<ev_check> watchers highest (C<EV_MAXPRI>)
1726	priority, to ensure that they are being run before any other watchers	2616	priority, to ensure that they are being run before any other watchers
		2617	after the poll (this doesn't matter for C<ev_prepare> watchers).
		2618
1727	after the poll. Also, C<ev_check> watchers (and C<ev_prepare> watchers,	2619	Also, C<ev_check> watchers (and C<ev_prepare> watchers, too) should not
1728	too) should not activate ("feed") events into libev. While libev fully	2620	activate ("feed") events into libev. While libev fully supports this, they
1729	supports this, they will be called before other C<ev_check> watchers	2621	might get executed before other C<ev_check> watchers did their job. As
1730	did their job. As C<ev_check> watchers are often used to embed other	2622	C<ev_check> watchers are often used to embed other (non-libev) event
1731	(non-libev) event loops those other event loops might be in an unusable	2623	loops those other event loops might be in an unusable state until their
1732	state until their C<ev_check> watcher ran (always remind yourself to	2624	C<ev_check> watcher ran (always remind yourself to coexist peacefully with
1733	coexist peacefully with others).	2625	others).
1734		2626
1735	=head3 Watcher-Specific Functions and Data Members	2627	=head3 Watcher-Specific Functions and Data Members
1736		2628
1737	=over 4	2629	=over 4
1738		2630
…		…
1740		2632
1741	=item ev_check_init (ev_check *, callback)	2633	=item ev_check_init (ev_check *, callback)
1742		2634
1743	Initialises and configures the prepare or check watcher - they have no	2635	Initialises and configures the prepare or check watcher - they have no
1744	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>	2636	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>
1745	macros, but using them is utterly, utterly and completely pointless.	2637	macros, but using them is utterly, utterly, utterly and completely
		2638	pointless.
1746		2639
1747	=back	2640	=back
		2641
		2642	=head3 Examples
1748		2643
1749	There are a number of principal ways to embed other event loops or modules	2644	There are a number of principal ways to embed other event loops or modules
1750	into libev. Here are some ideas on how to include libadns into libev	2645	into libev. Here are some ideas on how to include libadns into libev
1751	(there is a Perl module named C<EV::ADNS> that does this, which you could	2646	(there is a Perl module named C<EV::ADNS> that does this, which you could
1752	use for an actually working example. Another Perl module named C<EV::Glib>	2647	use as a working example. Another Perl module named C<EV::Glib> embeds a
1753	embeds a Glib main context into libev, and finally, C<Glib::EV> embeds EV	2648	Glib main context into libev, and finally, C<Glib::EV> embeds EV into the
1754	into the Glib event loop).	2649	Glib event loop).
1755		2650
1756	Method 1: Add IO watchers and a timeout watcher in a prepare handler,	2651	Method 1: Add IO watchers and a timeout watcher in a prepare handler,
1757	and in a check watcher, destroy them and call into libadns. What follows	2652	and in a check watcher, destroy them and call into libadns. What follows
1758	is pseudo-code only of course. This requires you to either use a low	2653	is pseudo-code only of course. This requires you to either use a low
1759	priority for the check watcher or use C<ev_clear_pending> explicitly, as	2654	priority for the check watcher or use C<ev_clear_pending> explicitly, as
1760	the callbacks for the IO/timeout watchers might not have been called yet.	2655	the callbacks for the IO/timeout watchers might not have been called yet.
1761		2656
1762	static ev_io iow [nfd];	2657	static ev_io iow [nfd];
1763	static ev_timer tw;	2658	static ev_timer tw;
1764		2659
1765	static void	2660	static void
1766	io_cb (ev_loop *loop, ev_io *w, int revents)	2661	io_cb (struct ev_loop *loop, ev_io *w, int revents)
1767	{	2662	{
1768	}	2663	}
1769		2664
1770	// create io watchers for each fd and a timer before blocking	2665	// create io watchers for each fd and a timer before blocking
1771	static void	2666	static void
1772	adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents)	2667	adns_prepare_cb (struct ev_loop *loop, ev_prepare *w, int revents)
1773	{	2668	{
1774	int timeout = 3600000;	2669	int timeout = 3600000;
1775	struct pollfd fds [nfd];	2670	struct pollfd fds [nfd];
1776	// actual code will need to loop here and realloc etc.	2671	// actual code will need to loop here and realloc etc.
1777	adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));	2672	adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));
1778		2673
1779	/* the callback is illegal, but won't be called as we stop during check */	2674	/* the callback is illegal, but won't be called as we stop during check */
1780	ev_timer_init (&tw, 0, timeout * 1e-3);	2675	ev_timer_init (&tw, 0, timeout * 1e-3, 0.);
1781	ev_timer_start (loop, &tw);	2676	ev_timer_start (loop, &tw);
1782		2677
1783	// create one ev_io per pollfd	2678	// create one ev_io per pollfd
1784	for (int i = 0; i < nfd; ++i)	2679	for (int i = 0; i < nfd; ++i)
1785	{	2680	{
1786	ev_io_init (iow + i, io_cb, fds [i].fd,	2681	ev_io_init (iow + i, io_cb, fds [i].fd,
1787	((fds [i].events & POLLIN ? EV_READ : 0)	2682	((fds [i].events & POLLIN ? EV_READ : 0)
1788	| (fds [i].events & POLLOUT ? EV_WRITE : 0)));	2683	| (fds [i].events & POLLOUT ? EV_WRITE : 0)));
1789		2684
1790	fds [i].revents = 0;	2685	fds [i].revents = 0;
1791	ev_io_start (loop, iow + i);	2686	ev_io_start (loop, iow + i);
1792	}	2687	}
1793	}	2688	}
1794		2689
1795	// stop all watchers after blocking	2690	// stop all watchers after blocking
1796	static void	2691	static void
1797	adns_check_cb (ev_loop *loop, ev_check *w, int revents)	2692	adns_check_cb (struct ev_loop *loop, ev_check *w, int revents)
1798	{	2693	{
1799	ev_timer_stop (loop, &tw);	2694	ev_timer_stop (loop, &tw);
1800		2695
1801	for (int i = 0; i < nfd; ++i)	2696	for (int i = 0; i < nfd; ++i)
1802	{	2697	{
1803	// set the relevant poll flags	2698	// set the relevant poll flags
1804	// could also call adns_processreadable etc. here	2699	// could also call adns_processreadable etc. here
1805	struct pollfd *fd = fds + i;	2700	struct pollfd *fd = fds + i;
1806	int revents = ev_clear_pending (iow + i);	2701	int revents = ev_clear_pending (iow + i);
1807	if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;	2702	if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;
1808	if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;	2703	if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;
1809		2704
1810	// now stop the watcher	2705	// now stop the watcher
1811	ev_io_stop (loop, iow + i);	2706	ev_io_stop (loop, iow + i);
1812	}	2707	}
1813		2708
1814	adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));	2709	adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));
1815	}	2710	}
1816		2711
1817	Method 2: This would be just like method 1, but you run C<adns_afterpoll>	2712	Method 2: This would be just like method 1, but you run C<adns_afterpoll>
1818	in the prepare watcher and would dispose of the check watcher.	2713	in the prepare watcher and would dispose of the check watcher.
1819		2714
1820	Method 3: If the module to be embedded supports explicit event	2715	Method 3: If the module to be embedded supports explicit event
1821	notification (adns does), you can also make use of the actual watcher	2716	notification (libadns does), you can also make use of the actual watcher
1822	callbacks, and only destroy/create the watchers in the prepare watcher.	2717	callbacks, and only destroy/create the watchers in the prepare watcher.
1823		2718
1824	static void	2719	static void
1825	timer_cb (EV_P_ ev_timer *w, int revents)	2720	timer_cb (EV_P_ ev_timer *w, int revents)
1826	{	2721	{
1827	adns_state ads = (adns_state)w->data;	2722	adns_state ads = (adns_state)w->data;
1828	update_now (EV_A);	2723	update_now (EV_A);
1829		2724
1830	adns_processtimeouts (ads, &tv_now);	2725	adns_processtimeouts (ads, &tv_now);
1831	}	2726	}
1832		2727
1833	static void	2728	static void
1834	io_cb (EV_P_ ev_io *w, int revents)	2729	io_cb (EV_P_ ev_io *w, int revents)
1835	{	2730	{
1836	adns_state ads = (adns_state)w->data;	2731	adns_state ads = (adns_state)w->data;
1837	update_now (EV_A);	2732	update_now (EV_A);
1838		2733
1839	if (revents & EV_READ ) adns_processreadable (ads, w->fd, &tv_now);	2734	if (revents & EV_READ ) adns_processreadable (ads, w->fd, &tv_now);
1840	if (revents & EV_WRITE) adns_processwriteable (ads, w->fd, &tv_now);	2735	if (revents & EV_WRITE) adns_processwriteable (ads, w->fd, &tv_now);
1841	}	2736	}
1842		2737
1843	// do not ever call adns_afterpoll	2738	// do not ever call adns_afterpoll
1844		2739
1845	Method 4: Do not use a prepare or check watcher because the module you	2740	Method 4: Do not use a prepare or check watcher because the module you
1846	want to embed is too inflexible to support it. Instead, youc na override	2741	want to embed is not flexible enough to support it. Instead, you can
1847	their poll function. The drawback with this solution is that the main	2742	override their poll function. The drawback with this solution is that the
1848	loop is now no longer controllable by EV. The C<Glib::EV> module does	2743	main loop is now no longer controllable by EV. The C<Glib::EV> module uses
1849	this.	2744	this approach, effectively embedding EV as a client into the horrible
		2745	libglib event loop.
1850		2746
1851	static gint	2747	static gint
1852	event_poll_func (GPollFD *fds, guint nfds, gint timeout)	2748	event_poll_func (GPollFD *fds, guint nfds, gint timeout)
1853	{	2749	{
1854	int got_events = 0;	2750	int got_events = 0;
1855		2751
1856	for (n = 0; n < nfds; ++n)	2752	for (n = 0; n < nfds; ++n)
1857	// create/start io watcher that sets the relevant bits in fds[n] and increment got_events	2753	// create/start io watcher that sets the relevant bits in fds[n] and increment got_events
1858		2754
1859	if (timeout >= 0)	2755	if (timeout >= 0)
1860	// create/start timer	2756	// create/start timer
1861		2757
1862	// poll	2758	// poll
1863	ev_loop (EV_A_ 0);	2759	ev_loop (EV_A_ 0);
1864		2760
1865	// stop timer again	2761	// stop timer again
1866	if (timeout >= 0)	2762	if (timeout >= 0)
1867	ev_timer_stop (EV_A_ &to);	2763	ev_timer_stop (EV_A_ &to);
1868		2764
1869	// stop io watchers again - their callbacks should have set	2765	// stop io watchers again - their callbacks should have set
1870	for (n = 0; n < nfds; ++n)	2766	for (n = 0; n < nfds; ++n)
1871	ev_io_stop (EV_A_ iow [n]);	2767	ev_io_stop (EV_A_ iow [n]);
1872		2768
1873	return got_events;	2769	return got_events;
1874	}	2770	}
1875		2771
1876		2772
1877	=head2 C<ev_embed> - when one backend isn't enough...	2773	=head2 C<ev_embed> - when one backend isn't enough...
1878		2774
1879	This is a rather advanced watcher type that lets you embed one event loop	2775	This is a rather advanced watcher type that lets you embed one event loop
…		…
1885	prioritise I/O.	2781	prioritise I/O.
1886		2782
1887	As an example for a bug workaround, the kqueue backend might only support	2783	As an example for a bug workaround, the kqueue backend might only support
1888	sockets on some platform, so it is unusable as generic backend, but you	2784	sockets on some platform, so it is unusable as generic backend, but you
1889	still want to make use of it because you have many sockets and it scales	2785	still want to make use of it because you have many sockets and it scales
1890	so nicely. In this case, you would create a kqueue-based loop and embed it	2786	so nicely. In this case, you would create a kqueue-based loop and embed
1891	into your default loop (which might use e.g. poll). Overall operation will	2787	it into your default loop (which might use e.g. poll). Overall operation
1892	be a bit slower because first libev has to poll and then call kevent, but	2788	will be a bit slower because first libev has to call C<poll> and then
1893	at least you can use both at what they are best.	2789	C<kevent>, but at least you can use both mechanisms for what they are
		2790	best: C<kqueue> for scalable sockets and C<poll> if you want it to work :)
1894		2791
1895	As for prioritising I/O: rarely you have the case where some fds have	2792	As for prioritising I/O: under rare circumstances you have the case where
1896	to be watched and handled very quickly (with low latency), and even	2793	some fds have to be watched and handled very quickly (with low latency),
1897	priorities and idle watchers might have too much overhead. In this case	2794	and even priorities and idle watchers might have too much overhead. In
1898	you would put all the high priority stuff in one loop and all the rest in	2795	this case you would put all the high priority stuff in one loop and all
1899	a second one, and embed the second one in the first.	2796	the rest in a second one, and embed the second one in the first.
1900		2797
1901	As long as the watcher is active, the callback will be invoked every time	2798	As long as the watcher is active, the callback will be invoked every
1902	there might be events pending in the embedded loop. The callback must then	2799	time there might be events pending in the embedded loop. The callback
1903	call C<ev_embed_sweep (mainloop, watcher)> to make a single sweep and invoke	2800	must then call C<ev_embed_sweep (mainloop, watcher)> to make a single
1904	their callbacks (you could also start an idle watcher to give the embedded	2801	sweep and invoke their callbacks (the callback doesn't need to invoke the
1905	loop strictly lower priority for example). You can also set the callback	2802	C<ev_embed_sweep> function directly, it could also start an idle watcher
1906	to C<0>, in which case the embed watcher will automatically execute the	2803	to give the embedded loop strictly lower priority for example).
1907	embedded loop sweep.
1908		2804
1909	As long as the watcher is started it will automatically handle events. The	2805	You can also set the callback to C<0>, in which case the embed watcher
1910	callback will be invoked whenever some events have been handled. You can	2806	will automatically execute the embedded loop sweep whenever necessary.
1911	set the callback to C<0> to avoid having to specify one if you are not
1912	interested in that.
1913		2807
1914	Also, there have not currently been made special provisions for forking:	2808	Fork detection will be handled transparently while the C<ev_embed> watcher
1915	when you fork, you not only have to call C<ev_loop_fork> on both loops,	2809	is active, i.e., the embedded loop will automatically be forked when the
1916	but you will also have to stop and restart any C<ev_embed> watchers	2810	embedding loop forks. In other cases, the user is responsible for calling
1917	yourself.	2811	C<ev_loop_fork> on the embedded loop.
1918		2812
1919	Unfortunately, not all backends are embeddable, only the ones returned by	2813	Unfortunately, not all backends are embeddable: only the ones returned by
1920	C<ev_embeddable_backends> are, which, unfortunately, does not include any	2814	C<ev_embeddable_backends> are, which, unfortunately, does not include any
1921	portable one.	2815	portable one.
1922		2816
1923	So when you want to use this feature you will always have to be prepared	2817	So when you want to use this feature you will always have to be prepared
1924	that you cannot get an embeddable loop. The recommended way to get around	2818	that you cannot get an embeddable loop. The recommended way to get around
1925	this is to have a separate variables for your embeddable loop, try to	2819	this is to have a separate variables for your embeddable loop, try to
1926	create it, and if that fails, use the normal loop for everything:	2820	create it, and if that fails, use the normal loop for everything.
1927		2821
1928	struct ev_loop *loop_hi = ev_default_init (0);	2822	=head3 C<ev_embed> and fork
1929	struct ev_loop *loop_lo = 0;
1930	struct ev_embed embed;
1931
1932	// see if there is a chance of getting one that works
1933	// (remember that a flags value of 0 means autodetection)
1934	loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
1935	? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
1936	: 0;
1937		2823
1938	// if we got one, then embed it, otherwise default to loop_hi	2824	While the C<ev_embed> watcher is running, forks in the embedding loop will
1939	if (loop_lo)	2825	automatically be applied to the embedded loop as well, so no special
1940	{	2826	fork handling is required in that case. When the watcher is not running,
1941	ev_embed_init (&embed, 0, loop_lo);	2827	however, it is still the task of the libev user to call C<ev_loop_fork ()>
1942	ev_embed_start (loop_hi, &embed);	2828	as applicable.
1943	}
1944	else
1945	loop_lo = loop_hi;
1946		2829
1947	=head3 Watcher-Specific Functions and Data Members	2830	=head3 Watcher-Specific Functions and Data Members
1948		2831
1949	=over 4	2832	=over 4
1950		2833
…		…
1954		2837
1955	Configures the watcher to embed the given loop, which must be	2838	Configures the watcher to embed the given loop, which must be
1956	embeddable. If the callback is C<0>, then C<ev_embed_sweep> will be	2839	embeddable. If the callback is C<0>, then C<ev_embed_sweep> will be
1957	invoked automatically, otherwise it is the responsibility of the callback	2840	invoked automatically, otherwise it is the responsibility of the callback
1958	to invoke it (it will continue to be called until the sweep has been done,	2841	to invoke it (it will continue to be called until the sweep has been done,
1959	if you do not want thta, you need to temporarily stop the embed watcher).	2842	if you do not want that, you need to temporarily stop the embed watcher).
1960		2843
1961	=item ev_embed_sweep (loop, ev_embed *)	2844	=item ev_embed_sweep (loop, ev_embed *)
1962		2845
1963	Make a single, non-blocking sweep over the embedded loop. This works	2846	Make a single, non-blocking sweep over the embedded loop. This works
1964	similarly to C<ev_loop (embedded_loop, EVLOOP_NONBLOCK)>, but in the most	2847	similarly to C<ev_loop (embedded_loop, EVLOOP_NONBLOCK)>, but in the most
1965	apropriate way for embedded loops.	2848	appropriate way for embedded loops.
1966		2849
1967	=item struct ev_loop *other [read-only]	2850	=item struct ev_loop *other [read-only]
1968		2851
1969	The embedded event loop.	2852	The embedded event loop.
1970		2853
1971	=back	2854	=back
		2855
		2856	=head3 Examples
		2857
		2858	Example: Try to get an embeddable event loop and embed it into the default
		2859	event loop. If that is not possible, use the default loop. The default
		2860	loop is stored in C<loop_hi>, while the embeddable loop is stored in
		2861	C<loop_lo> (which is C<loop_hi> in the case no embeddable loop can be
		2862	used).
		2863
		2864	struct ev_loop *loop_hi = ev_default_init (0);
		2865	struct ev_loop *loop_lo = 0;
		2866	ev_embed embed;
		2867
		2868	// see if there is a chance of getting one that works
		2869	// (remember that a flags value of 0 means autodetection)
		2870	loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
		2871	? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
		2872	: 0;
		2873
		2874	// if we got one, then embed it, otherwise default to loop_hi
		2875	if (loop_lo)
		2876	{
		2877	ev_embed_init (&embed, 0, loop_lo);
		2878	ev_embed_start (loop_hi, &embed);
		2879	}
		2880	else
		2881	loop_lo = loop_hi;
		2882
		2883	Example: Check if kqueue is available but not recommended and create
		2884	a kqueue backend for use with sockets (which usually work with any
		2885	kqueue implementation). Store the kqueue/socket-only event loop in
		2886	C<loop_socket>. (One might optionally use C<EVFLAG_NOENV>, too).
		2887
		2888	struct ev_loop *loop = ev_default_init (0);
		2889	struct ev_loop *loop_socket = 0;
		2890	ev_embed embed;
		2891
		2892	if (ev_supported_backends () & ~ev_recommended_backends () & EVBACKEND_KQUEUE)
		2893	if ((loop_socket = ev_loop_new (EVBACKEND_KQUEUE))
		2894	{
		2895	ev_embed_init (&embed, 0, loop_socket);
		2896	ev_embed_start (loop, &embed);
		2897	}
		2898
		2899	if (!loop_socket)
		2900	loop_socket = loop;
		2901
		2902	// now use loop_socket for all sockets, and loop for everything else
1972		2903
1973		2904
1974	=head2 C<ev_fork> - the audacity to resume the event loop after a fork	2905	=head2 C<ev_fork> - the audacity to resume the event loop after a fork
1975		2906
1976	Fork watchers are called when a C<fork ()> was detected (usually because	2907	Fork watchers are called when a C<fork ()> was detected (usually because
…		…
1979	event loop blocks next and before C<ev_check> watchers are being called,	2910	event loop blocks next and before C<ev_check> watchers are being called,
1980	and only in the child after the fork. If whoever good citizen calling	2911	and only in the child after the fork. If whoever good citizen calling
1981	C<ev_default_fork> cheats and calls it in the wrong process, the fork	2912	C<ev_default_fork> cheats and calls it in the wrong process, the fork
1982	handlers will be invoked, too, of course.	2913	handlers will be invoked, too, of course.
1983		2914
		2915	=head3 The special problem of life after fork - how is it possible?
		2916
		2917	Most uses of C<fork()> consist of forking, then some simple calls to ste
		2918	up/change the process environment, followed by a call to C<exec()>. This
		2919	sequence should be handled by libev without any problems.
		2920
		2921	This changes when the application actually wants to do event handling
		2922	in the child, or both parent in child, in effect "continuing" after the
		2923	fork.
		2924
		2925	The default mode of operation (for libev, with application help to detect
		2926	forks) is to duplicate all the state in the child, as would be expected
		2927	when I<either> the parent I<or> the child process continues.
		2928
		2929	When both processes want to continue using libev, then this is usually the
		2930	wrong result. In that case, usually one process (typically the parent) is
		2931	supposed to continue with all watchers in place as before, while the other
		2932	process typically wants to start fresh, i.e. without any active watchers.
		2933
		2934	The cleanest and most efficient way to achieve that with libev is to
		2935	simply create a new event loop, which of course will be "empty", and
		2936	use that for new watchers. This has the advantage of not touching more
		2937	memory than necessary, and thus avoiding the copy-on-write, and the
		2938	disadvantage of having to use multiple event loops (which do not support
		2939	signal watchers).
		2940
		2941	When this is not possible, or you want to use the default loop for
		2942	other reasons, then in the process that wants to start "fresh", call
		2943	C<ev_default_destroy ()> followed by C<ev_default_loop (...)>. Destroying
		2944	the default loop will "orphan" (not stop) all registered watchers, so you
		2945	have to be careful not to execute code that modifies those watchers. Note
		2946	also that in that case, you have to re-register any signal watchers.
		2947
1984	=head3 Watcher-Specific Functions and Data Members	2948	=head3 Watcher-Specific Functions and Data Members
1985		2949
1986	=over 4	2950	=over 4
1987		2951
1988	=item ev_fork_init (ev_signal *, callback)	2952	=item ev_fork_init (ev_signal *, callback)
…		…
1992	believe me.	2956	believe me.
1993		2957
1994	=back	2958	=back
1995		2959
1996		2960
		2961	=head2 C<ev_async> - how to wake up another event loop
		2962
		2963	In general, you cannot use an C<ev_loop> from multiple threads or other
		2964	asynchronous sources such as signal handlers (as opposed to multiple event
		2965	loops - those are of course safe to use in different threads).
		2966
		2967	Sometimes, however, you need to wake up another event loop you do not
		2968	control, for example because it belongs to another thread. This is what
		2969	C<ev_async> watchers do: as long as the C<ev_async> watcher is active, you
		2970	can signal it by calling C<ev_async_send>, which is thread- and signal
		2971	safe.
		2972
		2973	This functionality is very similar to C<ev_signal> watchers, as signals,
		2974	too, are asynchronous in nature, and signals, too, will be compressed
		2975	(i.e. the number of callback invocations may be less than the number of
		2976	C<ev_async_sent> calls).
		2977
		2978	Unlike C<ev_signal> watchers, C<ev_async> works with any event loop, not
		2979	just the default loop.
		2980
		2981	=head3 Queueing
		2982
		2983	C<ev_async> does not support queueing of data in any way. The reason
		2984	is that the author does not know of a simple (or any) algorithm for a
		2985	multiple-writer-single-reader queue that works in all cases and doesn't
		2986	need elaborate support such as pthreads.
		2987
		2988	That means that if you want to queue data, you have to provide your own
		2989	queue. But at least I can tell you how to implement locking around your
		2990	queue:
		2991
		2992	=over 4
		2993
		2994	=item queueing from a signal handler context
		2995
		2996	To implement race-free queueing, you simply add to the queue in the signal
		2997	handler but you block the signal handler in the watcher callback. Here is
		2998	an example that does that for some fictitious SIGUSR1 handler:
		2999
		3000	static ev_async mysig;
		3001
		3002	static void
		3003	sigusr1_handler (void)
		3004	{
		3005	sometype data;
		3006
		3007	// no locking etc.
		3008	queue_put (data);
		3009	ev_async_send (EV_DEFAULT_ &mysig);
		3010	}
		3011
		3012	static void
		3013	mysig_cb (EV_P_ ev_async *w, int revents)
		3014	{
		3015	sometype data;
		3016	sigset_t block, prev;
		3017
		3018	sigemptyset (&block);
		3019	sigaddset (&block, SIGUSR1);
		3020	sigprocmask (SIG_BLOCK, &block, &prev);
		3021
		3022	while (queue_get (&data))
		3023	process (data);
		3024
		3025	if (sigismember (&prev, SIGUSR1)
		3026	sigprocmask (SIG_UNBLOCK, &block, 0);
		3027	}
		3028
		3029	(Note: pthreads in theory requires you to use C<pthread_setmask>
		3030	instead of C<sigprocmask> when you use threads, but libev doesn't do it
		3031	either...).
		3032
		3033	=item queueing from a thread context
		3034
		3035	The strategy for threads is different, as you cannot (easily) block
		3036	threads but you can easily preempt them, so to queue safely you need to
		3037	employ a traditional mutex lock, such as in this pthread example:
		3038
		3039	static ev_async mysig;
		3040	static pthread_mutex_t mymutex = PTHREAD_MUTEX_INITIALIZER;
		3041
		3042	static void
		3043	otherthread (void)
		3044	{
		3045	// only need to lock the actual queueing operation
		3046	pthread_mutex_lock (&mymutex);
		3047	queue_put (data);
		3048	pthread_mutex_unlock (&mymutex);
		3049
		3050	ev_async_send (EV_DEFAULT_ &mysig);
		3051	}
		3052
		3053	static void
		3054	mysig_cb (EV_P_ ev_async *w, int revents)
		3055	{
		3056	pthread_mutex_lock (&mymutex);
		3057
		3058	while (queue_get (&data))
		3059	process (data);
		3060
		3061	pthread_mutex_unlock (&mymutex);
		3062	}
		3063
		3064	=back
		3065
		3066
		3067	=head3 Watcher-Specific Functions and Data Members
		3068
		3069	=over 4
		3070
		3071	=item ev_async_init (ev_async *, callback)
		3072
		3073	Initialises and configures the async watcher - it has no parameters of any
		3074	kind. There is a C<ev_async_set> macro, but using it is utterly pointless,
		3075	trust me.
		3076
		3077	=item ev_async_send (loop, ev_async *)
		3078
		3079	Sends/signals/activates the given C<ev_async> watcher, that is, feeds
		3080	an C<EV_ASYNC> event on the watcher into the event loop. Unlike
		3081	C<ev_feed_event>, this call is safe to do from other threads, signal or
		3082	similar contexts (see the discussion of C<EV_ATOMIC_T> in the embedding
		3083	section below on what exactly this means).
		3084
		3085	Note that, as with other watchers in libev, multiple events might get
		3086	compressed into a single callback invocation (another way to look at this
		3087	is that C<ev_async> watchers are level-triggered, set on C<ev_async_send>,
		3088	reset when the event loop detects that).
		3089
		3090	This call incurs the overhead of a system call only once per event loop
		3091	iteration, so while the overhead might be noticeable, it doesn't apply to
		3092	repeated calls to C<ev_async_send> for the same event loop.
		3093
		3094	=item bool = ev_async_pending (ev_async *)
		3095
		3096	Returns a non-zero value when C<ev_async_send> has been called on the
		3097	watcher but the event has not yet been processed (or even noted) by the
		3098	event loop.
		3099
		3100	C<ev_async_send> sets a flag in the watcher and wakes up the loop. When
		3101	the loop iterates next and checks for the watcher to have become active,
		3102	it will reset the flag again. C<ev_async_pending> can be used to very
		3103	quickly check whether invoking the loop might be a good idea.
		3104
		3105	Not that this does I<not> check whether the watcher itself is pending,
		3106	only whether it has been requested to make this watcher pending: there
		3107	is a time window between the event loop checking and resetting the async
		3108	notification, and the callback being invoked.
		3109
		3110	=back
		3111
		3112
1997	=head1 OTHER FUNCTIONS	3113	=head1 OTHER FUNCTIONS
1998		3114
1999	There are some other functions of possible interest. Described. Here. Now.	3115	There are some other functions of possible interest. Described. Here. Now.
2000		3116
2001	=over 4	3117	=over 4
2002		3118
2003	=item ev_once (loop, int fd, int events, ev_tstamp timeout, callback)	3119	=item ev_once (loop, int fd, int events, ev_tstamp timeout, callback)
2004		3120
2005	This function combines a simple timer and an I/O watcher, calls your	3121	This function combines a simple timer and an I/O watcher, calls your
2006	callback on whichever event happens first and automatically stop both	3122	callback on whichever event happens first and automatically stops both
2007	watchers. This is useful if you want to wait for a single event on an fd	3123	watchers. This is useful if you want to wait for a single event on an fd
2008	or timeout without having to allocate/configure/start/stop/free one or	3124	or timeout without having to allocate/configure/start/stop/free one or
2009	more watchers yourself.	3125	more watchers yourself.
2010		3126
2011	If C<fd> is less than 0, then no I/O watcher will be started and events	3127	If C<fd> is less than 0, then no I/O watcher will be started and the
2012	is being ignored. Otherwise, an C<ev_io> watcher for the given C<fd> and	3128	C<events> argument is being ignored. Otherwise, an C<ev_io> watcher for
2013	C<events> set will be craeted and started.	3129	the given C<fd> and C<events> set will be created and started.
2014		3130
2015	If C<timeout> is less than 0, then no timeout watcher will be	3131	If C<timeout> is less than 0, then no timeout watcher will be
2016	started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and	3132	started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and
2017	repeat = 0) will be started. While C<0> is a valid timeout, it is of	3133	repeat = 0) will be started. C<0> is a valid timeout.
2018	dubious value.
2019		3134
2020	The callback has the type C<void (*cb)(int revents, void *arg)> and gets	3135	The callback has the type C<void (*cb)(int revents, void *arg)> and gets
2021	passed an C<revents> set like normal event callbacks (a combination of	3136	passed an C<revents> set like normal event callbacks (a combination of
2022	C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMEOUT>) and the C<arg>	3137	C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMEOUT>) and the C<arg>
2023	value passed to C<ev_once>:	3138	value passed to C<ev_once>. Note that it is possible to receive I<both>
		3139	a timeout and an io event at the same time - you probably should give io
		3140	events precedence.
2024		3141
		3142	Example: wait up to ten seconds for data to appear on STDIN_FILENO.
		3143
2025	static void stdin_ready (int revents, void *arg)	3144	static void stdin_ready (int revents, void *arg)
2026	{	3145	{
2027	if (revents & EV_TIMEOUT)
2028	/* doh, nothing entered */;
2029	else if (revents & EV_READ)	3146	if (revents & EV_READ)
2030	/* stdin might have data for us, joy! */;	3147	/* stdin might have data for us, joy! */;
		3148	else if (revents & EV_TIMEOUT)
		3149	/* doh, nothing entered */;
2031	}	3150	}
2032		3151
2033	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);	3152	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
2034		3153
2035	=item ev_feed_event (ev_loop *, watcher *, int revents)
2036
2037	Feeds the given event set into the event loop, as if the specified event
2038	had happened for the specified watcher (which must be a pointer to an
2039	initialised but not necessarily started event watcher).
2040
2041	=item ev_feed_fd_event (ev_loop *, int fd, int revents)	3154	=item ev_feed_fd_event (struct ev_loop *, int fd, int revents)
2042		3155
2043	Feed an event on the given fd, as if a file descriptor backend detected	3156	Feed an event on the given fd, as if a file descriptor backend detected
2044	the given events it.	3157	the given events it.
2045		3158
2046	=item ev_feed_signal_event (ev_loop *loop, int signum)	3159	=item ev_feed_signal_event (struct ev_loop *loop, int signum)
2047		3160
2048	Feed an event as if the given signal occured (C<loop> must be the default	3161	Feed an event as if the given signal occurred (C<loop> must be the default
2049	loop!).	3162	loop!).
2050		3163
2051	=back	3164	=back
2052		3165
2053		3166
…		…
2069		3182
2070	=item * Priorities are not currently supported. Initialising priorities	3183	=item * Priorities are not currently supported. Initialising priorities
2071	will fail and all watchers will have the same priority, even though there	3184	will fail and all watchers will have the same priority, even though there
2072	is an ev_pri field.	3185	is an ev_pri field.
2073		3186
		3187	=item * In libevent, the last base created gets the signals, in libev, the
		3188	first base created (== the default loop) gets the signals.
		3189
2074	=item * Other members are not supported.	3190	=item * Other members are not supported.
2075		3191
2076	=item * The libev emulation is I<not> ABI compatible to libevent, you need	3192	=item * The libev emulation is I<not> ABI compatible to libevent, you need
2077	to use the libev header file and library.	3193	to use the libev header file and library.
2078		3194
2079	=back	3195	=back
2080		3196
2081	=head1 C++ SUPPORT	3197	=head1 C++ SUPPORT
2082		3198
2083	Libev comes with some simplistic wrapper classes for C++ that mainly allow	3199	Libev comes with some simplistic wrapper classes for C++ that mainly allow
2084	you to use some convinience methods to start/stop watchers and also change	3200	you to use some convenience methods to start/stop watchers and also change
2085	the callback model to a model using method callbacks on objects.	3201	the callback model to a model using method callbacks on objects.
2086		3202
2087	To use it,	3203	To use it,
2088		3204
2089	#include <ev++.h>	3205	#include <ev++.h>
2090		3206
2091	This automatically includes F<ev.h> and puts all of its definitions (many	3207	This automatically includes F<ev.h> and puts all of its definitions (many
2092	of them macros) into the global namespace. All C++ specific things are	3208	of them macros) into the global namespace. All C++ specific things are
2093	put into the C<ev> namespace. It should support all the same embedding	3209	put into the C<ev> namespace. It should support all the same embedding
2094	options as F<ev.h>, most notably C<EV_MULTIPLICITY>.	3210	options as F<ev.h>, most notably C<EV_MULTIPLICITY>.
…		…
2161	your compiler is good :), then the method will be fully inlined into the	3277	your compiler is good :), then the method will be fully inlined into the
2162	thunking function, making it as fast as a direct C callback.	3278	thunking function, making it as fast as a direct C callback.
2163		3279
2164	Example: simple class declaration and watcher initialisation	3280	Example: simple class declaration and watcher initialisation
2165		3281
2166	struct myclass	3282	struct myclass
2167	{	3283	{
2168	void io_cb (ev::io &w, int revents) { }	3284	void io_cb (ev::io &w, int revents) { }
2169	}	3285	}
2170		3286
2171	myclass obj;	3287	myclass obj;
2172	ev::io iow;	3288	ev::io iow;
2173	iow.set <myclass, &myclass::io_cb> (&obj);	3289	iow.set <myclass, &myclass::io_cb> (&obj);
		3290
		3291	=item w->set (object *)
		3292
		3293	This is an B<experimental> feature that might go away in a future version.
		3294
		3295	This is a variation of a method callback - leaving out the method to call
		3296	will default the method to C<operator ()>, which makes it possible to use
		3297	functor objects without having to manually specify the C<operator ()> all
		3298	the time. Incidentally, you can then also leave out the template argument
		3299	list.
		3300
		3301	The C<operator ()> method prototype must be C<void operator ()(watcher &w,
		3302	int revents)>.
		3303
		3304	See the method-C<set> above for more details.
		3305
		3306	Example: use a functor object as callback.
		3307
		3308	struct myfunctor
		3309	{
		3310	void operator() (ev::io &w, int revents)
		3311	{
		3312	...
		3313	}
		3314	}
		3315
		3316	myfunctor f;
		3317
		3318	ev::io w;
		3319	w.set (&f);
2174		3320
2175	=item w->set<function> (void *data = 0)	3321	=item w->set<function> (void *data = 0)
2176		3322
2177	Also sets a callback, but uses a static method or plain function as	3323	Also sets a callback, but uses a static method or plain function as
2178	callback. The optional C<data> argument will be stored in the watcher's	3324	callback. The optional C<data> argument will be stored in the watcher's
…		…
2180		3326
2181	The prototype of the C<function> must be C<void (*)(ev::TYPE &w, int)>.	3327	The prototype of the C<function> must be C<void (*)(ev::TYPE &w, int)>.
2182		3328
2183	See the method-C<set> above for more details.	3329	See the method-C<set> above for more details.
2184		3330
2185	Example:	3331	Example: Use a plain function as callback.
2186		3332
2187	static void io_cb (ev::io &w, int revents) { }	3333	static void io_cb (ev::io &w, int revents) { }
2188	iow.set <io_cb> ();	3334	iow.set <io_cb> ();
2189		3335
2190	=item w->set (struct ev_loop *)	3336	=item w->set (struct ev_loop *)
2191		3337
2192	Associates a different C<struct ev_loop> with this watcher. You can only	3338	Associates a different C<struct ev_loop> with this watcher. You can only
2193	do this when the watcher is inactive (and not pending either).	3339	do this when the watcher is inactive (and not pending either).
2194		3340
2195	=item w->set ([args])	3341	=item w->set ([arguments])
2196		3342
2197	Basically the same as C<ev_TYPE_set>, with the same args. Must be	3343	Basically the same as C<ev_TYPE_set>, with the same arguments. Must be
2198	called at least once. Unlike the C counterpart, an active watcher gets	3344	called at least once. Unlike the C counterpart, an active watcher gets
2199	automatically stopped and restarted when reconfiguring it with this	3345	automatically stopped and restarted when reconfiguring it with this
2200	method.	3346	method.
2201		3347
2202	=item w->start ()	3348	=item w->start ()
…		…
2226	=back	3372	=back
2227		3373
2228	Example: Define a class with an IO and idle watcher, start one of them in	3374	Example: Define a class with an IO and idle watcher, start one of them in
2229	the constructor.	3375	the constructor.
2230		3376
2231	class myclass	3377	class myclass
2232	{	3378	{
2233	ev_io io; void io_cb (ev::io &w, int revents);	3379	ev::io io ; void io_cb (ev::io &w, int revents);
2234	ev_idle idle void idle_cb (ev::idle &w, int revents);	3380	ev::idle idle; void idle_cb (ev::idle &w, int revents);
2235		3381
2236	myclass ();	3382	myclass (int fd)
2237	}	3383	{
2238
2239	myclass::myclass (int fd)
2240	{
2241	io .set <myclass, &myclass::io_cb > (this);	3384	io .set <myclass, &myclass::io_cb > (this);
2242	idle.set <myclass, &myclass::idle_cb> (this);	3385	idle.set <myclass, &myclass::idle_cb> (this);
2243		3386
2244	io.start (fd, ev::READ);	3387	io.start (fd, ev::READ);
		3388	}
2245	}	3389	};
		3390
		3391
		3392	=head1 OTHER LANGUAGE BINDINGS
		3393
		3394	Libev does not offer other language bindings itself, but bindings for a
		3395	number of languages exist in the form of third-party packages. If you know
		3396	any interesting language binding in addition to the ones listed here, drop
		3397	me a note.
		3398
		3399	=over 4
		3400
		3401	=item Perl
		3402
		3403	The EV module implements the full libev API and is actually used to test
		3404	libev. EV is developed together with libev. Apart from the EV core module,
		3405	there are additional modules that implement libev-compatible interfaces
		3406	to C<libadns> (C<EV::ADNS>, but C<AnyEvent::DNS> is preferred nowadays),
		3407	C<Net::SNMP> (C<Net::SNMP::EV>) and the C<libglib> event core (C<Glib::EV>
		3408	and C<EV::Glib>).
		3409
		3410	It can be found and installed via CPAN, its homepage is at
		3411	L<http://software.schmorp.de/pkg/EV>.
		3412
		3413	=item Python
		3414
		3415	Python bindings can be found at L<http://code.google.com/p/pyev/>. It
		3416	seems to be quite complete and well-documented.
		3417
		3418	=item Ruby
		3419
		3420	Tony Arcieri has written a ruby extension that offers access to a subset
		3421	of the libev API and adds file handle abstractions, asynchronous DNS and
		3422	more on top of it. It can be found via gem servers. Its homepage is at
		3423	L<http://rev.rubyforge.org/>.
		3424
		3425	Roger Pack reports that using the link order C<-lws2_32 -lmsvcrt-ruby-190>
		3426	makes rev work even on mingw.
		3427
		3428	=item Haskell
		3429
		3430	A haskell binding to libev is available at
		3431	L<http://hackage.haskell.org/cgi-bin/hackage-scripts/package/hlibev>.
		3432
		3433	=item D
		3434
		3435	Leandro Lucarella has written a D language binding (F<ev.d>) for libev, to
		3436	be found at L<http://proj.llucax.com.ar/wiki/evd>.
		3437
		3438	=item Ocaml
		3439
		3440	Erkki Seppala has written Ocaml bindings for libev, to be found at
		3441	L<http://modeemi.cs.tut.fi/~flux/software/ocaml-ev/>.
		3442
		3443	=item Lua
		3444
		3445	Brian Maher has written a partial interface to libev
		3446	for lua (only C<ev_io> and C<ev_timer>), to be found at
		3447	L<http://github.com/brimworks/lua-ev>.
		3448
		3449	=back
2246		3450
2247		3451
2248	=head1 MACRO MAGIC	3452	=head1 MACRO MAGIC
2249		3453
2250	Libev can be compiled with a variety of options, the most fundamantal	3454	Libev can be compiled with a variety of options, the most fundamental
2251	of which is C<EV_MULTIPLICITY>. This option determines whether (most)	3455	of which is C<EV_MULTIPLICITY>. This option determines whether (most)
2252	functions and callbacks have an initial C<struct ev_loop *> argument.	3456	functions and callbacks have an initial C<struct ev_loop *> argument.
2253		3457
2254	To make it easier to write programs that cope with either variant, the	3458	To make it easier to write programs that cope with either variant, the
2255	following macros are defined:	3459	following macros are defined:
…		…
2260		3464
2261	This provides the loop I<argument> for functions, if one is required ("ev	3465	This provides the loop I<argument> for functions, if one is required ("ev
2262	loop argument"). The C<EV_A> form is used when this is the sole argument,	3466	loop argument"). The C<EV_A> form is used when this is the sole argument,
2263	C<EV_A_> is used when other arguments are following. Example:	3467	C<EV_A_> is used when other arguments are following. Example:
2264		3468
2265	ev_unref (EV_A);	3469	ev_unref (EV_A);
2266	ev_timer_add (EV_A_ watcher);	3470	ev_timer_add (EV_A_ watcher);
2267	ev_loop (EV_A_ 0);	3471	ev_loop (EV_A_ 0);
2268		3472
2269	It assumes the variable C<loop> of type C<struct ev_loop *> is in scope,	3473	It assumes the variable C<loop> of type C<struct ev_loop *> is in scope,
2270	which is often provided by the following macro.	3474	which is often provided by the following macro.
2271		3475
2272	=item C<EV_P>, C<EV_P_>	3476	=item C<EV_P>, C<EV_P_>
2273		3477
2274	This provides the loop I<parameter> for functions, if one is required ("ev	3478	This provides the loop I<parameter> for functions, if one is required ("ev
2275	loop parameter"). The C<EV_P> form is used when this is the sole parameter,	3479	loop parameter"). The C<EV_P> form is used when this is the sole parameter,
2276	C<EV_P_> is used when other parameters are following. Example:	3480	C<EV_P_> is used when other parameters are following. Example:
2277		3481
2278	// this is how ev_unref is being declared	3482	// this is how ev_unref is being declared
2279	static void ev_unref (EV_P);	3483	static void ev_unref (EV_P);
2280		3484
2281	// this is how you can declare your typical callback	3485	// this is how you can declare your typical callback
2282	static void cb (EV_P_ ev_timer *w, int revents)	3486	static void cb (EV_P_ ev_timer *w, int revents)
2283		3487
2284	It declares a parameter C<loop> of type C<struct ev_loop *>, quite	3488	It declares a parameter C<loop> of type C<struct ev_loop *>, quite
2285	suitable for use with C<EV_A>.	3489	suitable for use with C<EV_A>.
2286		3490
2287	=item C<EV_DEFAULT>, C<EV_DEFAULT_>	3491	=item C<EV_DEFAULT>, C<EV_DEFAULT_>
2288		3492
2289	Similar to the other two macros, this gives you the value of the default	3493	Similar to the other two macros, this gives you the value of the default
2290	loop, if multiple loops are supported ("ev loop default").	3494	loop, if multiple loops are supported ("ev loop default").
		3495
		3496	=item C<EV_DEFAULT_UC>, C<EV_DEFAULT_UC_>
		3497
		3498	Usage identical to C<EV_DEFAULT> and C<EV_DEFAULT_>, but requires that the
		3499	default loop has been initialised (C<UC> == unchecked). Their behaviour
		3500	is undefined when the default loop has not been initialised by a previous
		3501	execution of C<EV_DEFAULT>, C<EV_DEFAULT_> or C<ev_default_init (...)>.
		3502
		3503	It is often prudent to use C<EV_DEFAULT> when initialising the first
		3504	watcher in a function but use C<EV_DEFAULT_UC> afterwards.
2291		3505
2292	=back	3506	=back
2293		3507
2294	Example: Declare and initialise a check watcher, utilising the above	3508	Example: Declare and initialise a check watcher, utilising the above
2295	macros so it will work regardless of whether multiple loops are supported	3509	macros so it will work regardless of whether multiple loops are supported
2296	or not.	3510	or not.
2297		3511
2298	static void	3512	static void
2299	check_cb (EV_P_ ev_timer *w, int revents)	3513	check_cb (EV_P_ ev_timer *w, int revents)
2300	{	3514	{
2301	ev_check_stop (EV_A_ w);	3515	ev_check_stop (EV_A_ w);
2302	}	3516	}
2303		3517
2304	ev_check check;	3518	ev_check check;
2305	ev_check_init (&check, check_cb);	3519	ev_check_init (&check, check_cb);
2306	ev_check_start (EV_DEFAULT_ &check);	3520	ev_check_start (EV_DEFAULT_ &check);
2307	ev_loop (EV_DEFAULT_ 0);	3521	ev_loop (EV_DEFAULT_ 0);
2308		3522
2309	=head1 EMBEDDING	3523	=head1 EMBEDDING
2310		3524
2311	Libev can (and often is) directly embedded into host	3525	Libev can (and often is) directly embedded into host
2312	applications. Examples of applications that embed it include the Deliantra	3526	applications. Examples of applications that embed it include the Deliantra
…		…
2319	libev somewhere in your source tree).	3533	libev somewhere in your source tree).
2320		3534
2321	=head2 FILESETS	3535	=head2 FILESETS
2322		3536
2323	Depending on what features you need you need to include one or more sets of files	3537	Depending on what features you need you need to include one or more sets of files
2324	in your app.	3538	in your application.
2325		3539
2326	=head3 CORE EVENT LOOP	3540	=head3 CORE EVENT LOOP
2327		3541
2328	To include only the libev core (all the C<ev_*> functions), with manual	3542	To include only the libev core (all the C<ev_*> functions), with manual
2329	configuration (no autoconf):	3543	configuration (no autoconf):
2330		3544
2331	#define EV_STANDALONE 1	3545	#define EV_STANDALONE 1
2332	#include "ev.c"	3546	#include "ev.c"
2333		3547
2334	This will automatically include F<ev.h>, too, and should be done in a	3548	This will automatically include F<ev.h>, too, and should be done in a
2335	single C source file only to provide the function implementations. To use	3549	single C source file only to provide the function implementations. To use
2336	it, do the same for F<ev.h> in all files wishing to use this API (best	3550	it, do the same for F<ev.h> in all files wishing to use this API (best
2337	done by writing a wrapper around F<ev.h> that you can include instead and	3551	done by writing a wrapper around F<ev.h> that you can include instead and
2338	where you can put other configuration options):	3552	where you can put other configuration options):
2339		3553
2340	#define EV_STANDALONE 1	3554	#define EV_STANDALONE 1
2341	#include "ev.h"	3555	#include "ev.h"
2342		3556
2343	Both header files and implementation files can be compiled with a C++	3557	Both header files and implementation files can be compiled with a C++
2344	compiler (at least, thats a stated goal, and breakage will be treated	3558	compiler (at least, that's a stated goal, and breakage will be treated
2345	as a bug).	3559	as a bug).
2346		3560
2347	You need the following files in your source tree, or in a directory	3561	You need the following files in your source tree, or in a directory
2348	in your include path (e.g. in libev/ when using -Ilibev):	3562	in your include path (e.g. in libev/ when using -Ilibev):
2349		3563
2350	ev.h	3564	ev.h
2351	ev.c	3565	ev.c
2352	ev_vars.h	3566	ev_vars.h
2353	ev_wrap.h	3567	ev_wrap.h
2354		3568
2355	ev_win32.c required on win32 platforms only	3569	ev_win32.c required on win32 platforms only
2356		3570
2357	ev_select.c only when select backend is enabled (which is enabled by default)	3571	ev_select.c only when select backend is enabled (which is enabled by default)
2358	ev_poll.c only when poll backend is enabled (disabled by default)	3572	ev_poll.c only when poll backend is enabled (disabled by default)
2359	ev_epoll.c only when the epoll backend is enabled (disabled by default)	3573	ev_epoll.c only when the epoll backend is enabled (disabled by default)
2360	ev_kqueue.c only when the kqueue backend is enabled (disabled by default)	3574	ev_kqueue.c only when the kqueue backend is enabled (disabled by default)
2361	ev_port.c only when the solaris port backend is enabled (disabled by default)	3575	ev_port.c only when the solaris port backend is enabled (disabled by default)
2362		3576
2363	F<ev.c> includes the backend files directly when enabled, so you only need	3577	F<ev.c> includes the backend files directly when enabled, so you only need
2364	to compile this single file.	3578	to compile this single file.
2365		3579
2366	=head3 LIBEVENT COMPATIBILITY API	3580	=head3 LIBEVENT COMPATIBILITY API
2367		3581
2368	To include the libevent compatibility API, also include:	3582	To include the libevent compatibility API, also include:
2369		3583
2370	#include "event.c"	3584	#include "event.c"
2371		3585
2372	in the file including F<ev.c>, and:	3586	in the file including F<ev.c>, and:
2373		3587
2374	#include "event.h"	3588	#include "event.h"
2375		3589
2376	in the files that want to use the libevent API. This also includes F<ev.h>.	3590	in the files that want to use the libevent API. This also includes F<ev.h>.
2377		3591
2378	You need the following additional files for this:	3592	You need the following additional files for this:
2379		3593
2380	event.h	3594	event.h
2381	event.c	3595	event.c
2382		3596
2383	=head3 AUTOCONF SUPPORT	3597	=head3 AUTOCONF SUPPORT
2384		3598
2385	Instead of using C<EV_STANDALONE=1> and providing your config in	3599	Instead of using C<EV_STANDALONE=1> and providing your configuration in
2386	whatever way you want, you can also C<m4_include([libev.m4])> in your	3600	whatever way you want, you can also C<m4_include([libev.m4])> in your
2387	F<configure.ac> and leave C<EV_STANDALONE> undefined. F<ev.c> will then	3601	F<configure.ac> and leave C<EV_STANDALONE> undefined. F<ev.c> will then
2388	include F<config.h> and configure itself accordingly.	3602	include F<config.h> and configure itself accordingly.
2389		3603
2390	For this of course you need the m4 file:	3604	For this of course you need the m4 file:
2391		3605
2392	libev.m4	3606	libev.m4
2393		3607
2394	=head2 PREPROCESSOR SYMBOLS/MACROS	3608	=head2 PREPROCESSOR SYMBOLS/MACROS
2395		3609
2396	Libev can be configured via a variety of preprocessor symbols you have to define	3610	Libev can be configured via a variety of preprocessor symbols you have to
2397	before including any of its files. The default is not to build for multiplicity	3611	define before including any of its files. The default in the absence of
2398	and only include the select backend.	3612	autoconf is documented for every option.
2399		3613
2400	=over 4	3614	=over 4
2401		3615
2402	=item EV_STANDALONE	3616	=item EV_STANDALONE
2403		3617
…		…
2405	keeps libev from including F<config.h>, and it also defines dummy	3619	keeps libev from including F<config.h>, and it also defines dummy
2406	implementations for some libevent functions (such as logging, which is not	3620	implementations for some libevent functions (such as logging, which is not
2407	supported). It will also not define any of the structs usually found in	3621	supported). It will also not define any of the structs usually found in
2408	F<event.h> that are not directly supported by the libev core alone.	3622	F<event.h> that are not directly supported by the libev core alone.
2409		3623
		3624	In standalone mode, libev will still try to automatically deduce the
		3625	configuration, but has to be more conservative.
		3626
2410	=item EV_USE_MONOTONIC	3627	=item EV_USE_MONOTONIC
2411		3628
2412	If defined to be C<1>, libev will try to detect the availability of the	3629	If defined to be C<1>, libev will try to detect the availability of the
2413	monotonic clock option at both compiletime and runtime. Otherwise no use	3630	monotonic clock option at both compile time and runtime. Otherwise no
2414	of the monotonic clock option will be attempted. If you enable this, you	3631	use of the monotonic clock option will be attempted. If you enable this,
2415	usually have to link against librt or something similar. Enabling it when	3632	you usually have to link against librt or something similar. Enabling it
2416	the functionality isn't available is safe, though, although you have	3633	when the functionality isn't available is safe, though, although you have
2417	to make sure you link against any libraries where the C<clock_gettime>	3634	to make sure you link against any libraries where the C<clock_gettime>
2418	function is hiding in (often F<-lrt>).	3635	function is hiding in (often F<-lrt>). See also C<EV_USE_CLOCK_SYSCALL>.
2419		3636
2420	=item EV_USE_REALTIME	3637	=item EV_USE_REALTIME
2421		3638
2422	If defined to be C<1>, libev will try to detect the availability of the	3639	If defined to be C<1>, libev will try to detect the availability of the
2423	realtime clock option at compiletime (and assume its availability at	3640	real-time clock option at compile time (and assume its availability
2424	runtime if successful). Otherwise no use of the realtime clock option will	3641	at runtime if successful). Otherwise no use of the real-time clock
2425	be attempted. This effectively replaces C<gettimeofday> by C<clock_get	3642	option will be attempted. This effectively replaces C<gettimeofday>
2426	(CLOCK_REALTIME, ...)> and will not normally affect correctness. See the	3643	by C<clock_get (CLOCK_REALTIME, ...)> and will not normally affect
2427	note about libraries in the description of C<EV_USE_MONOTONIC>, though.	3644	correctness. See the note about libraries in the description of
		3645	C<EV_USE_MONOTONIC>, though. Defaults to the opposite value of
		3646	C<EV_USE_CLOCK_SYSCALL>.
		3647
		3648	=item EV_USE_CLOCK_SYSCALL
		3649
		3650	If defined to be C<1>, libev will try to use a direct syscall instead
		3651	of calling the system-provided C<clock_gettime> function. This option
		3652	exists because on GNU/Linux, C<clock_gettime> is in C<librt>, but C<librt>
		3653	unconditionally pulls in C<libpthread>, slowing down single-threaded
		3654	programs needlessly. Using a direct syscall is slightly slower (in
		3655	theory), because no optimised vdso implementation can be used, but avoids
		3656	the pthread dependency. Defaults to C<1> on GNU/Linux with glibc 2.x or
		3657	higher, as it simplifies linking (no need for C<-lrt>).
2428		3658
2429	=item EV_USE_NANOSLEEP	3659	=item EV_USE_NANOSLEEP
2430		3660
2431	If defined to be C<1>, libev will assume that C<nanosleep ()> is available	3661	If defined to be C<1>, libev will assume that C<nanosleep ()> is available
2432	and will use it for delays. Otherwise it will use C<select ()>.	3662	and will use it for delays. Otherwise it will use C<select ()>.
2433		3663
		3664	=item EV_USE_EVENTFD
		3665
		3666	If defined to be C<1>, then libev will assume that C<eventfd ()> is
		3667	available and will probe for kernel support at runtime. This will improve
		3668	C<ev_signal> and C<ev_async> performance and reduce resource consumption.
		3669	If undefined, it will be enabled if the headers indicate GNU/Linux + Glibc
		3670	2.7 or newer, otherwise disabled.
		3671
2434	=item EV_USE_SELECT	3672	=item EV_USE_SELECT
2435		3673
2436	If undefined or defined to be C<1>, libev will compile in support for the	3674	If undefined or defined to be C<1>, libev will compile in support for the
2437	C<select>(2) backend. No attempt at autodetection will be done: if no	3675	C<select>(2) backend. No attempt at auto-detection will be done: if no
2438	other method takes over, select will be it. Otherwise the select backend	3676	other method takes over, select will be it. Otherwise the select backend
2439	will not be compiled in.	3677	will not be compiled in.
2440		3678
2441	=item EV_SELECT_USE_FD_SET	3679	=item EV_SELECT_USE_FD_SET
2442		3680
2443	If defined to C<1>, then the select backend will use the system C<fd_set>	3681	If defined to C<1>, then the select backend will use the system C<fd_set>
2444	structure. This is useful if libev doesn't compile due to a missing	3682	structure. This is useful if libev doesn't compile due to a missing
2445	C<NFDBITS> or C<fd_mask> definition or it misguesses the bitset layout on	3683	C<NFDBITS> or C<fd_mask> definition or it mis-guesses the bitset layout
2446	exotic systems. This usually limits the range of file descriptors to some	3684	on exotic systems. This usually limits the range of file descriptors to
2447	low limit such as 1024 or might have other limitations (winsocket only	3685	some low limit such as 1024 or might have other limitations (winsocket
2448	allows 64 sockets). The C<FD_SETSIZE> macro, set before compilation, might	3686	only allows 64 sockets). The C<FD_SETSIZE> macro, set before compilation,
2449	influence the size of the C<fd_set> used.	3687	configures the maximum size of the C<fd_set>.
2450		3688
2451	=item EV_SELECT_IS_WINSOCKET	3689	=item EV_SELECT_IS_WINSOCKET
2452		3690
2453	When defined to C<1>, the select backend will assume that	3691	When defined to C<1>, the select backend will assume that
2454	select/socket/connect etc. don't understand file descriptors but	3692	select/socket/connect etc. don't understand file descriptors but
…		…
2456	be used is the winsock select). This means that it will call	3694	be used is the winsock select). This means that it will call
2457	C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,	3695	C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,
2458	it is assumed that all these functions actually work on fds, even	3696	it is assumed that all these functions actually work on fds, even
2459	on win32. Should not be defined on non-win32 platforms.	3697	on win32. Should not be defined on non-win32 platforms.
2460		3698
		3699	=item EV_FD_TO_WIN32_HANDLE(fd)
		3700
		3701	If C<EV_SELECT_IS_WINSOCKET> is enabled, then libev needs a way to map
		3702	file descriptors to socket handles. When not defining this symbol (the
		3703	default), then libev will call C<_get_osfhandle>, which is usually
		3704	correct. In some cases, programs use their own file descriptor management,
		3705	in which case they can provide this function to map fds to socket handles.
		3706
		3707	=item EV_WIN32_HANDLE_TO_FD(handle)
		3708
		3709	If C<EV_SELECT_IS_WINSOCKET> then libev maps handles to file descriptors
		3710	using the standard C<_open_osfhandle> function. For programs implementing
		3711	their own fd to handle mapping, overwriting this function makes it easier
		3712	to do so. This can be done by defining this macro to an appropriate value.
		3713
		3714	=item EV_WIN32_CLOSE_FD(fd)
		3715
		3716	If programs implement their own fd to handle mapping on win32, then this
		3717	macro can be used to override the C<close> function, useful to unregister
		3718	file descriptors again. Note that the replacement function has to close
		3719	the underlying OS handle.
		3720
2461	=item EV_USE_POLL	3721	=item EV_USE_POLL
2462		3722
2463	If defined to be C<1>, libev will compile in support for the C<poll>(2)	3723	If defined to be C<1>, libev will compile in support for the C<poll>(2)
2464	backend. Otherwise it will be enabled on non-win32 platforms. It	3724	backend. Otherwise it will be enabled on non-win32 platforms. It
2465	takes precedence over select.	3725	takes precedence over select.
2466		3726
2467	=item EV_USE_EPOLL	3727	=item EV_USE_EPOLL
2468		3728
2469	If defined to be C<1>, libev will compile in support for the Linux	3729	If defined to be C<1>, libev will compile in support for the Linux
2470	C<epoll>(7) backend. Its availability will be detected at runtime,	3730	C<epoll>(7) backend. Its availability will be detected at runtime,
2471	otherwise another method will be used as fallback. This is the	3731	otherwise another method will be used as fallback. This is the preferred
2472	preferred backend for GNU/Linux systems.	3732	backend for GNU/Linux systems. If undefined, it will be enabled if the
		3733	headers indicate GNU/Linux + Glibc 2.4 or newer, otherwise disabled.
2473		3734
2474	=item EV_USE_KQUEUE	3735	=item EV_USE_KQUEUE
2475		3736
2476	If defined to be C<1>, libev will compile in support for the BSD style	3737	If defined to be C<1>, libev will compile in support for the BSD style
2477	C<kqueue>(2) backend. Its actual availability will be detected at runtime,	3738	C<kqueue>(2) backend. Its actual availability will be detected at runtime,
…		…
2490	otherwise another method will be used as fallback. This is the preferred	3751	otherwise another method will be used as fallback. This is the preferred
2491	backend for Solaris 10 systems.	3752	backend for Solaris 10 systems.
2492		3753
2493	=item EV_USE_DEVPOLL	3754	=item EV_USE_DEVPOLL
2494		3755
2495	reserved for future expansion, works like the USE symbols above.	3756	Reserved for future expansion, works like the USE symbols above.
2496		3757
2497	=item EV_USE_INOTIFY	3758	=item EV_USE_INOTIFY
2498		3759
2499	If defined to be C<1>, libev will compile in support for the Linux inotify	3760	If defined to be C<1>, libev will compile in support for the Linux inotify
2500	interface to speed up C<ev_stat> watchers. Its actual availability will	3761	interface to speed up C<ev_stat> watchers. Its actual availability will
2501	be detected at runtime.	3762	be detected at runtime. If undefined, it will be enabled if the headers
		3763	indicate GNU/Linux + Glibc 2.4 or newer, otherwise disabled.
		3764
		3765	=item EV_ATOMIC_T
		3766
		3767	Libev requires an integer type (suitable for storing C<0> or C<1>) whose
		3768	access is atomic with respect to other threads or signal contexts. No such
		3769	type is easily found in the C language, so you can provide your own type
		3770	that you know is safe for your purposes. It is used both for signal handler "locking"
		3771	as well as for signal and thread safety in C<ev_async> watchers.
		3772
		3773	In the absence of this define, libev will use C<sig_atomic_t volatile>
		3774	(from F<signal.h>), which is usually good enough on most platforms.
2502		3775
2503	=item EV_H	3776	=item EV_H
2504		3777
2505	The name of the F<ev.h> header file used to include it. The default if	3778	The name of the F<ev.h> header file used to include it. The default if
2506	undefined is C<< <ev.h> >> in F<event.h> and C<"ev.h"> in F<ev.c>. This	3779	undefined is C<"ev.h"> in F<event.h>, F<ev.c> and F<ev++.h>. This can be
2507	can be used to virtually rename the F<ev.h> header file in case of conflicts.	3780	used to virtually rename the F<ev.h> header file in case of conflicts.
2508		3781
2509	=item EV_CONFIG_H	3782	=item EV_CONFIG_H
2510		3783
2511	If C<EV_STANDALONE> isn't C<1>, this variable can be used to override	3784	If C<EV_STANDALONE> isn't C<1>, this variable can be used to override
2512	F<ev.c>'s idea of where to find the F<config.h> file, similarly to	3785	F<ev.c>'s idea of where to find the F<config.h> file, similarly to
2513	C<EV_H>, above.	3786	C<EV_H>, above.
2514		3787
2515	=item EV_EVENT_H	3788	=item EV_EVENT_H
2516		3789
2517	Similarly to C<EV_H>, this macro can be used to override F<event.c>'s idea	3790	Similarly to C<EV_H>, this macro can be used to override F<event.c>'s idea
2518	of how the F<event.h> header can be found.	3791	of how the F<event.h> header can be found, the default is C<"event.h">.
2519		3792
2520	=item EV_PROTOTYPES	3793	=item EV_PROTOTYPES
2521		3794
2522	If defined to be C<0>, then F<ev.h> will not define any function	3795	If defined to be C<0>, then F<ev.h> will not define any function
2523	prototypes, but still define all the structs and other symbols. This is	3796	prototypes, but still define all the structs and other symbols. This is
…		…
2544	When doing priority-based operations, libev usually has to linearly search	3817	When doing priority-based operations, libev usually has to linearly search
2545	all the priorities, so having many of them (hundreds) uses a lot of space	3818	all the priorities, so having many of them (hundreds) uses a lot of space
2546	and time, so using the defaults of five priorities (-2 .. +2) is usually	3819	and time, so using the defaults of five priorities (-2 .. +2) is usually
2547	fine.	3820	fine.
2548		3821
2549	If your embedding app does not need any priorities, defining these both to	3822	If your embedding application does not need any priorities, defining these
2550	C<0> will save some memory and cpu.	3823	both to C<0> will save some memory and CPU.
2551		3824
2552	=item EV_PERIODIC_ENABLE	3825	=item EV_PERIODIC_ENABLE
2553		3826
2554	If undefined or defined to be C<1>, then periodic timers are supported. If	3827	If undefined or defined to be C<1>, then periodic timers are supported. If
2555	defined to be C<0>, then they are not. Disabling them saves a few kB of	3828	defined to be C<0>, then they are not. Disabling them saves a few kB of
…		…
2562	code.	3835	code.
2563		3836
2564	=item EV_EMBED_ENABLE	3837	=item EV_EMBED_ENABLE
2565		3838
2566	If undefined or defined to be C<1>, then embed watchers are supported. If	3839	If undefined or defined to be C<1>, then embed watchers are supported. If
2567	defined to be C<0>, then they are not.	3840	defined to be C<0>, then they are not. Embed watchers rely on most other
		3841	watcher types, which therefore must not be disabled.
2568		3842
2569	=item EV_STAT_ENABLE	3843	=item EV_STAT_ENABLE
2570		3844
2571	If undefined or defined to be C<1>, then stat watchers are supported. If	3845	If undefined or defined to be C<1>, then stat watchers are supported. If
2572	defined to be C<0>, then they are not.	3846	defined to be C<0>, then they are not.
…		…
2574	=item EV_FORK_ENABLE	3848	=item EV_FORK_ENABLE
2575		3849
2576	If undefined or defined to be C<1>, then fork watchers are supported. If	3850	If undefined or defined to be C<1>, then fork watchers are supported. If
2577	defined to be C<0>, then they are not.	3851	defined to be C<0>, then they are not.
2578		3852
		3853	=item EV_ASYNC_ENABLE
		3854
		3855	If undefined or defined to be C<1>, then async watchers are supported. If
		3856	defined to be C<0>, then they are not.
		3857
2579	=item EV_MINIMAL	3858	=item EV_MINIMAL
2580		3859
2581	If you need to shave off some kilobytes of code at the expense of some	3860	If you need to shave off some kilobytes of code at the expense of some
2582	speed, define this symbol to C<1>. Currently only used for gcc to override	3861	speed (but with the full API), define this symbol to C<1>. Currently this
2583	some inlining decisions, saves roughly 30% codesize of amd64.	3862	is used to override some inlining decisions, saves roughly 30% code size
		3863	on amd64. It also selects a much smaller 2-heap for timer management over
		3864	the default 4-heap.
		3865
		3866	You can save even more by disabling watcher types you do not need
		3867	and setting C<EV_MAXPRI> == C<EV_MINPRI>. Also, disabling C<assert>
		3868	(C<-DNDEBUG>) will usually reduce code size a lot.
		3869
		3870	Defining C<EV_MINIMAL> to C<2> will additionally reduce the core API to
		3871	provide a bare-bones event library. See C<ev.h> for details on what parts
		3872	of the API are still available, and do not complain if this subset changes
		3873	over time.
		3874
		3875	=item EV_NSIG
		3876
		3877	The highest supported signal number, +1 (or, the number of
		3878	signals): Normally, libev tries to deduce the maximum number of signals
		3879	automatically, but sometimes this fails, in which case it can be
		3880	specified. Also, using a lower number than detected (C<32> should be
		3881	good for about any system in existance) can save some memory, as libev
		3882	statically allocates some 12-24 bytes per signal number.
2584		3883
2585	=item EV_PID_HASHSIZE	3884	=item EV_PID_HASHSIZE
2586		3885
2587	C<ev_child> watchers use a small hash table to distribute workload by	3886	C<ev_child> watchers use a small hash table to distribute workload by
2588	pid. The default size is C<16> (or C<1> with C<EV_MINIMAL>), usually more	3887	pid. The default size is C<16> (or C<1> with C<EV_MINIMAL>), usually more
…		…
2595	inotify watch id. The default size is C<16> (or C<1> with C<EV_MINIMAL>),	3894	inotify watch id. The default size is C<16> (or C<1> with C<EV_MINIMAL>),
2596	usually more than enough. If you need to manage thousands of C<ev_stat>	3895	usually more than enough. If you need to manage thousands of C<ev_stat>
2597	watchers you might want to increase this value (I<must> be a power of	3896	watchers you might want to increase this value (I<must> be a power of
2598	two).	3897	two).
2599		3898
		3899	=item EV_USE_4HEAP
		3900
		3901	Heaps are not very cache-efficient. To improve the cache-efficiency of the
		3902	timer and periodics heaps, libev uses a 4-heap when this symbol is defined
		3903	to C<1>. The 4-heap uses more complicated (longer) code but has noticeably
		3904	faster performance with many (thousands) of watchers.
		3905
		3906	The default is C<1> unless C<EV_MINIMAL> is set in which case it is C<0>
		3907	(disabled).
		3908
		3909	=item EV_HEAP_CACHE_AT
		3910
		3911	Heaps are not very cache-efficient. To improve the cache-efficiency of the
		3912	timer and periodics heaps, libev can cache the timestamp (I<at>) within
		3913	the heap structure (selected by defining C<EV_HEAP_CACHE_AT> to C<1>),
		3914	which uses 8-12 bytes more per watcher and a few hundred bytes more code,
		3915	but avoids random read accesses on heap changes. This improves performance
		3916	noticeably with many (hundreds) of watchers.
		3917
		3918	The default is C<1> unless C<EV_MINIMAL> is set in which case it is C<0>
		3919	(disabled).
		3920
		3921	=item EV_VERIFY
		3922
		3923	Controls how much internal verification (see C<ev_loop_verify ()>) will
		3924	be done: If set to C<0>, no internal verification code will be compiled
		3925	in. If set to C<1>, then verification code will be compiled in, but not
		3926	called. If set to C<2>, then the internal verification code will be
		3927	called once per loop, which can slow down libev. If set to C<3>, then the
		3928	verification code will be called very frequently, which will slow down
		3929	libev considerably.
		3930
		3931	The default is C<1>, unless C<EV_MINIMAL> is set, in which case it will be
		3932	C<0>.
		3933
2600	=item EV_COMMON	3934	=item EV_COMMON
2601		3935
2602	By default, all watchers have a C<void *data> member. By redefining	3936	By default, all watchers have a C<void *data> member. By redefining
2603	this macro to a something else you can include more and other types of	3937	this macro to a something else you can include more and other types of
2604	members. You have to define it each time you include one of the files,	3938	members. You have to define it each time you include one of the files,
2605	though, and it must be identical each time.	3939	though, and it must be identical each time.
2606		3940
2607	For example, the perl EV module uses something like this:	3941	For example, the perl EV module uses something like this:
2608		3942
2609	#define EV_COMMON \	3943	#define EV_COMMON \
2610	SV *self; /* contains this struct */ \	3944	SV *self; /* contains this struct */ \
2611	SV *cb_sv, *fh /* note no trailing ";" */	3945	SV *cb_sv, *fh /* note no trailing ";" */
2612		3946
2613	=item EV_CB_DECLARE (type)	3947	=item EV_CB_DECLARE (type)
2614		3948
2615	=item EV_CB_INVOKE (watcher, revents)	3949	=item EV_CB_INVOKE (watcher, revents)
2616		3950
…		…
2621	definition and a statement, respectively. See the F<ev.h> header file for	3955	definition and a statement, respectively. See the F<ev.h> header file for
2622	their default definitions. One possible use for overriding these is to	3956	their default definitions. One possible use for overriding these is to
2623	avoid the C<struct ev_loop *> as first argument in all cases, or to use	3957	avoid the C<struct ev_loop *> as first argument in all cases, or to use
2624	method calls instead of plain function calls in C++.	3958	method calls instead of plain function calls in C++.
2625		3959
		3960	=back
		3961
2626	=head2 EXPORTED API SYMBOLS	3962	=head2 EXPORTED API SYMBOLS
2627		3963
2628	If you need to re-export the API (e.g. via a dll) and you need a list of	3964	If you need to re-export the API (e.g. via a DLL) and you need a list of
2629	exported symbols, you can use the provided F<Symbol.*> files which list	3965	exported symbols, you can use the provided F<Symbol.*> files which list
2630	all public symbols, one per line:	3966	all public symbols, one per line:
2631		3967
2632	Symbols.ev for libev proper	3968	Symbols.ev for libev proper
2633	Symbols.event for the libevent emulation	3969	Symbols.event for the libevent emulation
2634		3970
2635	This can also be used to rename all public symbols to avoid clashes with	3971	This can also be used to rename all public symbols to avoid clashes with
2636	multiple versions of libev linked together (which is obviously bad in	3972	multiple versions of libev linked together (which is obviously bad in
2637	itself, but sometimes it is inconvinient to avoid this).	3973	itself, but sometimes it is inconvenient to avoid this).
2638		3974
2639	A sed command like this will create wrapper C<#define>'s that you need to	3975	A sed command like this will create wrapper C<#define>'s that you need to
2640	include before including F<ev.h>:	3976	include before including F<ev.h>:
2641		3977
2642	<Symbols.ev sed -e "s/.*/#define & myprefix_&/" >wrap.h	3978	<Symbols.ev sed -e "s/.*/#define & myprefix_&/" >wrap.h
…		…
2659	file.	3995	file.
2660		3996
2661	The usage in rxvt-unicode is simpler. It has a F<ev_cpp.h> header file	3997	The usage in rxvt-unicode is simpler. It has a F<ev_cpp.h> header file
2662	that everybody includes and which overrides some configure choices:	3998	that everybody includes and which overrides some configure choices:
2663		3999
2664	#define EV_MINIMAL 1	4000	#define EV_MINIMAL 1
2665	#define EV_USE_POLL 0	4001	#define EV_USE_POLL 0
2666	#define EV_MULTIPLICITY 0	4002	#define EV_MULTIPLICITY 0
2667	#define EV_PERIODIC_ENABLE 0	4003	#define EV_PERIODIC_ENABLE 0
2668	#define EV_STAT_ENABLE 0	4004	#define EV_STAT_ENABLE 0
2669	#define EV_FORK_ENABLE 0	4005	#define EV_FORK_ENABLE 0
2670	#define EV_CONFIG_H <config.h>	4006	#define EV_CONFIG_H <config.h>
2671	#define EV_MINPRI 0	4007	#define EV_MINPRI 0
2672	#define EV_MAXPRI 0	4008	#define EV_MAXPRI 0
2673		4009
2674	#include "ev++.h"	4010	#include "ev++.h"
2675		4011
2676	And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:	4012	And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:
2677		4013
2678	#include "ev_cpp.h"	4014	#include "ev_cpp.h"
2679	#include "ev.c"	4015	#include "ev.c"
2680		4016
		4017	=head1 INTERACTION WITH OTHER PROGRAMS OR LIBRARIES
2681		4018
		4019	=head2 THREADS AND COROUTINES
		4020
		4021	=head3 THREADS
		4022
		4023	All libev functions are reentrant and thread-safe unless explicitly
		4024	documented otherwise, but libev implements no locking itself. This means
		4025	that you can use as many loops as you want in parallel, as long as there
		4026	are no concurrent calls into any libev function with the same loop
		4027	parameter (C<ev_default_*> calls have an implicit default loop parameter,
		4028	of course): libev guarantees that different event loops share no data
		4029	structures that need any locking.
		4030
		4031	Or to put it differently: calls with different loop parameters can be done
		4032	concurrently from multiple threads, calls with the same loop parameter
		4033	must be done serially (but can be done from different threads, as long as
		4034	only one thread ever is inside a call at any point in time, e.g. by using
		4035	a mutex per loop).
		4036
		4037	Specifically to support threads (and signal handlers), libev implements
		4038	so-called C<ev_async> watchers, which allow some limited form of
		4039	concurrency on the same event loop, namely waking it up "from the
		4040	outside".
		4041
		4042	If you want to know which design (one loop, locking, or multiple loops
		4043	without or something else still) is best for your problem, then I cannot
		4044	help you, but here is some generic advice:
		4045
		4046	=over 4
		4047
		4048	=item * most applications have a main thread: use the default libev loop
		4049	in that thread, or create a separate thread running only the default loop.
		4050
		4051	This helps integrating other libraries or software modules that use libev
		4052	themselves and don't care/know about threading.
		4053
		4054	=item * one loop per thread is usually a good model.
		4055
		4056	Doing this is almost never wrong, sometimes a better-performance model
		4057	exists, but it is always a good start.
		4058
		4059	=item * other models exist, such as the leader/follower pattern, where one
		4060	loop is handed through multiple threads in a kind of round-robin fashion.
		4061
		4062	Choosing a model is hard - look around, learn, know that usually you can do
		4063	better than you currently do :-)
		4064
		4065	=item * often you need to talk to some other thread which blocks in the
		4066	event loop.
		4067
		4068	C<ev_async> watchers can be used to wake them up from other threads safely
		4069	(or from signal contexts...).
		4070
		4071	An example use would be to communicate signals or other events that only
		4072	work in the default loop by registering the signal watcher with the
		4073	default loop and triggering an C<ev_async> watcher from the default loop
		4074	watcher callback into the event loop interested in the signal.
		4075
		4076	=back
		4077
		4078	=head4 THREAD LOCKING EXAMPLE
		4079
		4080	Here is a fictitious example of how to run an event loop in a different
		4081	thread than where callbacks are being invoked and watchers are
		4082	created/added/removed.
		4083
		4084	For a real-world example, see the C<EV::Loop::Async> perl module,
		4085	which uses exactly this technique (which is suited for many high-level
		4086	languages).
		4087
		4088	The example uses a pthread mutex to protect the loop data, a condition
		4089	variable to wait for callback invocations, an async watcher to notify the
		4090	event loop thread and an unspecified mechanism to wake up the main thread.
		4091
		4092	First, you need to associate some data with the event loop:
		4093
		4094	typedef struct {
		4095	mutex_t lock; /* global loop lock */
		4096	ev_async async_w;
		4097	thread_t tid;
		4098	cond_t invoke_cv;
		4099	} userdata;
		4100
		4101	void prepare_loop (EV_P)
		4102	{
		4103	// for simplicity, we use a static userdata struct.
		4104	static userdata u;
		4105
		4106	ev_async_init (&u->async_w, async_cb);
		4107	ev_async_start (EV_A_ &u->async_w);
		4108
		4109	pthread_mutex_init (&u->lock, 0);
		4110	pthread_cond_init (&u->invoke_cv, 0);
		4111
		4112	// now associate this with the loop
		4113	ev_set_userdata (EV_A_ u);
		4114	ev_set_invoke_pending_cb (EV_A_ l_invoke);
		4115	ev_set_loop_release_cb (EV_A_ l_release, l_acquire);
		4116
		4117	// then create the thread running ev_loop
		4118	pthread_create (&u->tid, 0, l_run, EV_A);
		4119	}
		4120
		4121	The callback for the C<ev_async> watcher does nothing: the watcher is used
		4122	solely to wake up the event loop so it takes notice of any new watchers
		4123	that might have been added:
		4124
		4125	static void
		4126	async_cb (EV_P_ ev_async *w, int revents)
		4127	{
		4128	// just used for the side effects
		4129	}
		4130
		4131	The C<l_release> and C<l_acquire> callbacks simply unlock/lock the mutex
		4132	protecting the loop data, respectively.
		4133
		4134	static void
		4135	l_release (EV_P)
		4136	{
		4137	userdata *u = ev_userdata (EV_A);
		4138	pthread_mutex_unlock (&u->lock);
		4139	}
		4140
		4141	static void
		4142	l_acquire (EV_P)
		4143	{
		4144	userdata *u = ev_userdata (EV_A);
		4145	pthread_mutex_lock (&u->lock);
		4146	}
		4147
		4148	The event loop thread first acquires the mutex, and then jumps straight
		4149	into C<ev_loop>:
		4150
		4151	void *
		4152	l_run (void *thr_arg)
		4153	{
		4154	struct ev_loop *loop = (struct ev_loop *)thr_arg;
		4155
		4156	l_acquire (EV_A);
		4157	pthread_setcanceltype (PTHREAD_CANCEL_ASYNCHRONOUS, 0);
		4158	ev_loop (EV_A_ 0);
		4159	l_release (EV_A);
		4160
		4161	return 0;
		4162	}
		4163
		4164	Instead of invoking all pending watchers, the C<l_invoke> callback will
		4165	signal the main thread via some unspecified mechanism (signals? pipe
		4166	writes? C<Async::Interrupt>?) and then waits until all pending watchers
		4167	have been called (in a while loop because a) spurious wakeups are possible
		4168	and b) skipping inter-thread-communication when there are no pending
		4169	watchers is very beneficial):
		4170
		4171	static void
		4172	l_invoke (EV_P)
		4173	{
		4174	userdata *u = ev_userdata (EV_A);
		4175
		4176	while (ev_pending_count (EV_A))
		4177	{
		4178	wake_up_other_thread_in_some_magic_or_not_so_magic_way ();
		4179	pthread_cond_wait (&u->invoke_cv, &u->lock);
		4180	}
		4181	}
		4182
		4183	Now, whenever the main thread gets told to invoke pending watchers, it
		4184	will grab the lock, call C<ev_invoke_pending> and then signal the loop
		4185	thread to continue:
		4186
		4187	static void
		4188	real_invoke_pending (EV_P)
		4189	{
		4190	userdata *u = ev_userdata (EV_A);
		4191
		4192	pthread_mutex_lock (&u->lock);
		4193	ev_invoke_pending (EV_A);
		4194	pthread_cond_signal (&u->invoke_cv);
		4195	pthread_mutex_unlock (&u->lock);
		4196	}
		4197
		4198	Whenever you want to start/stop a watcher or do other modifications to an
		4199	event loop, you will now have to lock:
		4200
		4201	ev_timer timeout_watcher;
		4202	userdata *u = ev_userdata (EV_A);
		4203
		4204	ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);
		4205
		4206	pthread_mutex_lock (&u->lock);
		4207	ev_timer_start (EV_A_ &timeout_watcher);
		4208	ev_async_send (EV_A_ &u->async_w);
		4209	pthread_mutex_unlock (&u->lock);
		4210
		4211	Note that sending the C<ev_async> watcher is required because otherwise
		4212	an event loop currently blocking in the kernel will have no knowledge
		4213	about the newly added timer. By waking up the loop it will pick up any new
		4214	watchers in the next event loop iteration.
		4215
		4216	=head3 COROUTINES
		4217
		4218	Libev is very accommodating to coroutines ("cooperative threads"):
		4219	libev fully supports nesting calls to its functions from different
		4220	coroutines (e.g. you can call C<ev_loop> on the same loop from two
		4221	different coroutines, and switch freely between both coroutines running
		4222	the loop, as long as you don't confuse yourself). The only exception is
		4223	that you must not do this from C<ev_periodic> reschedule callbacks.
		4224
		4225	Care has been taken to ensure that libev does not keep local state inside
		4226	C<ev_loop>, and other calls do not usually allow for coroutine switches as
		4227	they do not call any callbacks.
		4228
		4229	=head2 COMPILER WARNINGS
		4230
		4231	Depending on your compiler and compiler settings, you might get no or a
		4232	lot of warnings when compiling libev code. Some people are apparently
		4233	scared by this.
		4234
		4235	However, these are unavoidable for many reasons. For one, each compiler
		4236	has different warnings, and each user has different tastes regarding
		4237	warning options. "Warn-free" code therefore cannot be a goal except when
		4238	targeting a specific compiler and compiler-version.
		4239
		4240	Another reason is that some compiler warnings require elaborate
		4241	workarounds, or other changes to the code that make it less clear and less
		4242	maintainable.
		4243
		4244	And of course, some compiler warnings are just plain stupid, or simply
		4245	wrong (because they don't actually warn about the condition their message
		4246	seems to warn about). For example, certain older gcc versions had some
		4247	warnings that resulted an extreme number of false positives. These have
		4248	been fixed, but some people still insist on making code warn-free with
		4249	such buggy versions.
		4250
		4251	While libev is written to generate as few warnings as possible,
		4252	"warn-free" code is not a goal, and it is recommended not to build libev
		4253	with any compiler warnings enabled unless you are prepared to cope with
		4254	them (e.g. by ignoring them). Remember that warnings are just that:
		4255	warnings, not errors, or proof of bugs.
		4256
		4257
		4258	=head2 VALGRIND
		4259
		4260	Valgrind has a special section here because it is a popular tool that is
		4261	highly useful. Unfortunately, valgrind reports are very hard to interpret.
		4262
		4263	If you think you found a bug (memory leak, uninitialised data access etc.)
		4264	in libev, then check twice: If valgrind reports something like:
		4265
		4266	==2274== definitely lost: 0 bytes in 0 blocks.
		4267	==2274== possibly lost: 0 bytes in 0 blocks.
		4268	==2274== still reachable: 256 bytes in 1 blocks.
		4269
		4270	Then there is no memory leak, just as memory accounted to global variables
		4271	is not a memleak - the memory is still being referenced, and didn't leak.
		4272
		4273	Similarly, under some circumstances, valgrind might report kernel bugs
		4274	as if it were a bug in libev (e.g. in realloc or in the poll backend,
		4275	although an acceptable workaround has been found here), or it might be
		4276	confused.
		4277
		4278	Keep in mind that valgrind is a very good tool, but only a tool. Don't
		4279	make it into some kind of religion.
		4280
		4281	If you are unsure about something, feel free to contact the mailing list
		4282	with the full valgrind report and an explanation on why you think this
		4283	is a bug in libev (best check the archives, too :). However, don't be
		4284	annoyed when you get a brisk "this is no bug" answer and take the chance
		4285	of learning how to interpret valgrind properly.
		4286
		4287	If you need, for some reason, empty reports from valgrind for your project
		4288	I suggest using suppression lists.
		4289
		4290
		4291	=head1 PORTABILITY NOTES
		4292
		4293	=head2 WIN32 PLATFORM LIMITATIONS AND WORKAROUNDS
		4294
		4295	Win32 doesn't support any of the standards (e.g. POSIX) that libev
		4296	requires, and its I/O model is fundamentally incompatible with the POSIX
		4297	model. Libev still offers limited functionality on this platform in
		4298	the form of the C<EVBACKEND_SELECT> backend, and only supports socket
		4299	descriptors. This only applies when using Win32 natively, not when using
		4300	e.g. cygwin.
		4301
		4302	Lifting these limitations would basically require the full
		4303	re-implementation of the I/O system. If you are into these kinds of
		4304	things, then note that glib does exactly that for you in a very portable
		4305	way (note also that glib is the slowest event library known to man).
		4306
		4307	There is no supported compilation method available on windows except
		4308	embedding it into other applications.
		4309
		4310	Sensible signal handling is officially unsupported by Microsoft - libev
		4311	tries its best, but under most conditions, signals will simply not work.
		4312
		4313	Not a libev limitation but worth mentioning: windows apparently doesn't
		4314	accept large writes: instead of resulting in a partial write, windows will
		4315	either accept everything or return C<ENOBUFS> if the buffer is too large,
		4316	so make sure you only write small amounts into your sockets (less than a
		4317	megabyte seems safe, but this apparently depends on the amount of memory
		4318	available).
		4319
		4320	Due to the many, low, and arbitrary limits on the win32 platform and
		4321	the abysmal performance of winsockets, using a large number of sockets
		4322	is not recommended (and not reasonable). If your program needs to use
		4323	more than a hundred or so sockets, then likely it needs to use a totally
		4324	different implementation for windows, as libev offers the POSIX readiness
		4325	notification model, which cannot be implemented efficiently on windows
		4326	(due to Microsoft monopoly games).
		4327
		4328	A typical way to use libev under windows is to embed it (see the embedding
		4329	section for details) and use the following F<evwrap.h> header file instead
		4330	of F<ev.h>:
		4331
		4332	#define EV_STANDALONE /* keeps ev from requiring config.h */
		4333	#define EV_SELECT_IS_WINSOCKET 1 /* configure libev for windows select */
		4334
		4335	#include "ev.h"
		4336
		4337	And compile the following F<evwrap.c> file into your project (make sure
		4338	you do I<not> compile the F<ev.c> or any other embedded source files!):
		4339
		4340	#include "evwrap.h"
		4341	#include "ev.c"
		4342
		4343	=over 4
		4344
		4345	=item The winsocket select function
		4346
		4347	The winsocket C<select> function doesn't follow POSIX in that it
		4348	requires socket I<handles> and not socket I<file descriptors> (it is
		4349	also extremely buggy). This makes select very inefficient, and also
		4350	requires a mapping from file descriptors to socket handles (the Microsoft
		4351	C runtime provides the function C<_open_osfhandle> for this). See the
		4352	discussion of the C<EV_SELECT_USE_FD_SET>, C<EV_SELECT_IS_WINSOCKET> and
		4353	C<EV_FD_TO_WIN32_HANDLE> preprocessor symbols for more info.
		4354
		4355	The configuration for a "naked" win32 using the Microsoft runtime
		4356	libraries and raw winsocket select is:
		4357
		4358	#define EV_USE_SELECT 1
		4359	#define EV_SELECT_IS_WINSOCKET 1 /* forces EV_SELECT_USE_FD_SET, too */
		4360
		4361	Note that winsockets handling of fd sets is O(n), so you can easily get a
		4362	complexity in the O(n²) range when using win32.
		4363
		4364	=item Limited number of file descriptors
		4365
		4366	Windows has numerous arbitrary (and low) limits on things.
		4367
		4368	Early versions of winsocket's select only supported waiting for a maximum
		4369	of C<64> handles (probably owning to the fact that all windows kernels
		4370	can only wait for C<64> things at the same time internally; Microsoft
		4371	recommends spawning a chain of threads and wait for 63 handles and the
		4372	previous thread in each. Sounds great!).
		4373
		4374	Newer versions support more handles, but you need to define C<FD_SETSIZE>
		4375	to some high number (e.g. C<2048>) before compiling the winsocket select
		4376	call (which might be in libev or elsewhere, for example, perl and many
		4377	other interpreters do their own select emulation on windows).
		4378
		4379	Another limit is the number of file descriptors in the Microsoft runtime
		4380	libraries, which by default is C<64> (there must be a hidden I<64>
		4381	fetish or something like this inside Microsoft). You can increase this
		4382	by calling C<_setmaxstdio>, which can increase this limit to C<2048>
		4383	(another arbitrary limit), but is broken in many versions of the Microsoft
		4384	runtime libraries. This might get you to about C<512> or C<2048> sockets
		4385	(depending on windows version and/or the phase of the moon). To get more,
		4386	you need to wrap all I/O functions and provide your own fd management, but
		4387	the cost of calling select (O(n²)) will likely make this unworkable.
		4388
		4389	=back
		4390
		4391	=head2 PORTABILITY REQUIREMENTS
		4392
		4393	In addition to a working ISO-C implementation and of course the
		4394	backend-specific APIs, libev relies on a few additional extensions:
		4395
		4396	=over 4
		4397
		4398	=item C<void (*)(ev_watcher_type *, int revents)> must have compatible
		4399	calling conventions regardless of C<ev_watcher_type *>.
		4400
		4401	Libev assumes not only that all watcher pointers have the same internal
		4402	structure (guaranteed by POSIX but not by ISO C for example), but it also
		4403	assumes that the same (machine) code can be used to call any watcher
		4404	callback: The watcher callbacks have different type signatures, but libev
		4405	calls them using an C<ev_watcher *> internally.
		4406
		4407	=item C<sig_atomic_t volatile> must be thread-atomic as well
		4408
		4409	The type C<sig_atomic_t volatile> (or whatever is defined as
		4410	C<EV_ATOMIC_T>) must be atomic with respect to accesses from different
		4411	threads. This is not part of the specification for C<sig_atomic_t>, but is
		4412	believed to be sufficiently portable.
		4413
		4414	=item C<sigprocmask> must work in a threaded environment
		4415
		4416	Libev uses C<sigprocmask> to temporarily block signals. This is not
		4417	allowed in a threaded program (C<pthread_sigmask> has to be used). Typical
		4418	pthread implementations will either allow C<sigprocmask> in the "main
		4419	thread" or will block signals process-wide, both behaviours would
		4420	be compatible with libev. Interaction between C<sigprocmask> and
		4421	C<pthread_sigmask> could complicate things, however.
		4422
		4423	The most portable way to handle signals is to block signals in all threads
		4424	except the initial one, and run the default loop in the initial thread as
		4425	well.
		4426
		4427	=item C<long> must be large enough for common memory allocation sizes
		4428
		4429	To improve portability and simplify its API, libev uses C<long> internally
		4430	instead of C<size_t> when allocating its data structures. On non-POSIX
		4431	systems (Microsoft...) this might be unexpectedly low, but is still at
		4432	least 31 bits everywhere, which is enough for hundreds of millions of
		4433	watchers.
		4434
		4435	=item C<double> must hold a time value in seconds with enough accuracy
		4436
		4437	The type C<double> is used to represent timestamps. It is required to
		4438	have at least 51 bits of mantissa (and 9 bits of exponent), which is good
		4439	enough for at least into the year 4000. This requirement is fulfilled by
		4440	implementations implementing IEEE 754, which is basically all existing
		4441	ones. With IEEE 754 doubles, you get microsecond accuracy until at least
		4442	2200.
		4443
		4444	=back
		4445
		4446	If you know of other additional requirements drop me a note.
		4447
		4448
2682	=head1 COMPLEXITIES	4449	=head1 ALGORITHMIC COMPLEXITIES
2683		4450
2684	In this section the complexities of (many of) the algorithms used inside	4451	In this section the complexities of (many of) the algorithms used inside
2685	libev will be explained. For complexity discussions about backends see the	4452	libev will be documented. For complexity discussions about backends see
2686	documentation for C<ev_default_init>.	4453	the documentation for C<ev_default_init>.
2687		4454
2688	All of the following are about amortised time: If an array needs to be	4455	All of the following are about amortised time: If an array needs to be
2689	extended, libev needs to realloc and move the whole array, but this	4456	extended, libev needs to realloc and move the whole array, but this
2690	happens asymptotically never with higher number of elements, so O(1) might	4457	happens asymptotically rarer with higher number of elements, so O(1) might
2691	mean it might do a lengthy realloc operation in rare cases, but on average	4458	mean that libev does a lengthy realloc operation in rare cases, but on
2692	it is much faster and asymptotically approaches constant time.	4459	average it is much faster and asymptotically approaches constant time.
2693		4460
2694	=over 4	4461	=over 4
2695		4462
2696	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)	4463	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)
2697		4464
2698	This means that, when you have a watcher that triggers in one hour and	4465	This means that, when you have a watcher that triggers in one hour and
2699	there are 100 watchers that would trigger before that then inserting will	4466	there are 100 watchers that would trigger before that, then inserting will
2700	have to skip roughly seven (C<ld 100>) of these watchers.	4467	have to skip roughly seven (C<ld 100>) of these watchers.
2701		4468
2702	=item Changing timer/periodic watchers (by autorepeat or calling again): O(log skipped_other_timers)	4469	=item Changing timer/periodic watchers (by autorepeat or calling again): O(log skipped_other_timers)
2703		4470
2704	That means that changing a timer costs less than removing/adding them	4471	That means that changing a timer costs less than removing/adding them,
2705	as only the relative motion in the event queue has to be paid for.	4472	as only the relative motion in the event queue has to be paid for.
2706		4473
2707	=item Starting io/check/prepare/idle/signal/child watchers: O(1)	4474	=item Starting io/check/prepare/idle/signal/child/fork/async watchers: O(1)
2708		4475
2709	These just add the watcher into an array or at the head of a list.	4476	These just add the watcher into an array or at the head of a list.
2710		4477
2711	=item Stopping check/prepare/idle watchers: O(1)	4478	=item Stopping check/prepare/idle/fork/async watchers: O(1)
2712		4479
2713	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))	4480	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))
2714		4481
2715	These watchers are stored in lists then need to be walked to find the	4482	These watchers are stored in lists, so they need to be walked to find the
2716	correct watcher to remove. The lists are usually short (you don't usually	4483	correct watcher to remove. The lists are usually short (you don't usually
2717	have many watchers waiting for the same fd or signal).	4484	have many watchers waiting for the same fd or signal: one is typical, two
		4485	is rare).
2718		4486
2719	=item Finding the next timer in each loop iteration: O(1)	4487	=item Finding the next timer in each loop iteration: O(1)
2720		4488
2721	By virtue of using a binary heap, the next timer is always found at the	4489	By virtue of using a binary or 4-heap, the next timer is always found at a
2722	beginning of the storage array.	4490	fixed position in the storage array.
2723		4491
2724	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)	4492	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)
2725		4493
2726	A change means an I/O watcher gets started or stopped, which requires	4494	A change means an I/O watcher gets started or stopped, which requires
2727	libev to recalculate its status (and possibly tell the kernel, depending	4495	libev to recalculate its status (and possibly tell the kernel, depending
2728	on backend and wether C<ev_io_set> was used).	4496	on backend and whether C<ev_io_set> was used).
2729		4497
2730	=item Activating one watcher (putting it into the pending state): O(1)	4498	=item Activating one watcher (putting it into the pending state): O(1)
2731		4499
2732	=item Priority handling: O(number_of_priorities)	4500	=item Priority handling: O(number_of_priorities)
2733		4501
2734	Priorities are implemented by allocating some space for each	4502	Priorities are implemented by allocating some space for each
2735	priority. When doing priority-based operations, libev usually has to	4503	priority. When doing priority-based operations, libev usually has to
2736	linearly search all the priorities, but starting/stopping and activating	4504	linearly search all the priorities, but starting/stopping and activating
2737	watchers becomes O(1) w.r.t. prioritiy handling.	4505	watchers becomes O(1) with respect to priority handling.
		4506
		4507	=item Sending an ev_async: O(1)
		4508
		4509	=item Processing ev_async_send: O(number_of_async_watchers)
		4510
		4511	=item Processing signals: O(max_signal_number)
		4512
		4513	Sending involves a system call I<iff> there were no other C<ev_async_send>
		4514	calls in the current loop iteration. Checking for async and signal events
		4515	involves iterating over all running async watchers or all signal numbers.
2738		4516
2739	=back	4517	=back
2740		4518
2741		4519
		4520	=head1 GLOSSARY
		4521
		4522	=over 4
		4523
		4524	=item active
		4525
		4526	A watcher is active as long as it has been started (has been attached to
		4527	an event loop) but not yet stopped (disassociated from the event loop).
		4528
		4529	=item application
		4530
		4531	In this document, an application is whatever is using libev.
		4532
		4533	=item callback
		4534
		4535	The address of a function that is called when some event has been
		4536	detected. Callbacks are being passed the event loop, the watcher that
		4537	received the event, and the actual event bitset.
		4538
		4539	=item callback invocation
		4540
		4541	The act of calling the callback associated with a watcher.
		4542
		4543	=item event
		4544
		4545	A change of state of some external event, such as data now being available
		4546	for reading on a file descriptor, time having passed or simply not having
		4547	any other events happening anymore.
		4548
		4549	In libev, events are represented as single bits (such as C<EV_READ> or
		4550	C<EV_TIMEOUT>).
		4551
		4552	=item event library
		4553
		4554	A software package implementing an event model and loop.
		4555
		4556	=item event loop
		4557
		4558	An entity that handles and processes external events and converts them
		4559	into callback invocations.
		4560
		4561	=item event model
		4562
		4563	The model used to describe how an event loop handles and processes
		4564	watchers and events.
		4565
		4566	=item pending
		4567
		4568	A watcher is pending as soon as the corresponding event has been detected,
		4569	and stops being pending as soon as the watcher will be invoked or its
		4570	pending status is explicitly cleared by the application.
		4571
		4572	A watcher can be pending, but not active. Stopping a watcher also clears
		4573	its pending status.
		4574
		4575	=item real time
		4576
		4577	The physical time that is observed. It is apparently strictly monotonic :)
		4578
		4579	=item wall-clock time
		4580
		4581	The time and date as shown on clocks. Unlike real time, it can actually
		4582	be wrong and jump forwards and backwards, e.g. when the you adjust your
		4583	clock.
		4584
		4585	=item watcher
		4586
		4587	A data structure that describes interest in certain events. Watchers need
		4588	to be started (attached to an event loop) before they can receive events.
		4589
		4590	=item watcher invocation
		4591
		4592	The act of calling the callback associated with a watcher.
		4593
		4594	=back
		4595
2742	=head1 AUTHOR	4596	=head1 AUTHOR
2743		4597
2744	Marc Lehmann <libev@schmorp.de>.	4598	Marc Lehmann <libev@schmorp.de>, with repeated corrections by Mikael Magnusson.
2745		4599

Diff Legend

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