ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/libev/ev.pod

Comparing libev/ev.pod (file contents):
Revision 1.121 by root, Mon Jan 28 12:13:54 2008 UTC vs.
Revision 1.323 by root, Sun Oct 24 18:01:26 2010 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_run's to stop iterating
		32	ev_break (EV_A_ EVBREAK_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_run to stop iterating
		41	ev_break (EV_A_ EVBREAK_ONE);
30	}	42	}
31		43
32	int	44	int
33	main (void)	45	main (void)
34	{	46	{
35	struct ev_loop *loop = ev_default_loop (0);	47	// use the default event loop unless you have special needs
		48	struct ev_loop *loop = EV_DEFAULT;
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_run (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	Familiarity 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<struct 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 (in practice
100	the beginning of 1970, details are complicated, don't ask). This type is	130	somewhere near the beginning of 1970, details are complicated, don't
101	called C<ev_tstamp>, which is what you should use too. It usually aliases	131	ask). This type is called C<ev_tstamp>, which is what you should use
102	to the C<double> type in C, and when you need to do any calculations on	132	too. It usually aliases to the C<double> type in C. When you need to do
103	it, you should treat it as some floatingpoint value. Unlike the name	133	any calculations on it, you should treat it as some floating point value.
		134
104	component C<stamp> might indicate, it is also used for time differences	135	Unlike the name component C<stamp> might indicate, it is also used for
105	throughout libev.	136	time differences (e.g. delays) throughout libev.
		137
		138	=head1 ERROR HANDLING
		139
		140	Libev knows three classes of errors: operating system errors, usage errors
		141	and internal errors (bugs).
		142
		143	When libev catches an operating system error it cannot handle (for example
		144	a system call indicating a condition libev cannot fix), it calls the callback
		145	set via C<ev_set_syserr_cb>, which is supposed to fix the problem or
		146	abort. The default is to print a diagnostic message and to call C<abort
		147	()>.
		148
		149	When libev detects a usage error such as a negative timer interval, then
		150	it will print a diagnostic message and abort (via the C<assert> mechanism,
		151	so C<NDEBUG> will disable this checking): these are programming errors in
		152	the libev caller and need to be fixed there.
		153
		154	Libev also has a few internal error-checking C<assert>ions, and also has
		155	extensive consistency checking code. These do not trigger under normal
		156	circumstances, as they indicate either a bug in libev or worse.
		157
106		158
107	=head1 GLOBAL FUNCTIONS	159	=head1 GLOBAL FUNCTIONS
108		160
109	These functions can be called anytime, even before initialising the	161	These functions can be called anytime, even before initialising the
110	library in any way.	162	library in any way.
…		…
113		165
114	=item ev_tstamp ev_time ()	166	=item ev_tstamp ev_time ()
115		167
116	Returns the current time as libev would use it. Please note that the	168	Returns the current time as libev would use it. Please note that the
117	C<ev_now> function is usually faster and also often returns the timestamp	169	C<ev_now> function is usually faster and also often returns the timestamp
118	you actually want to know.	170	you actually want to know. Also interesting is the combination of
		171	C<ev_update_now> and C<ev_now>.
119		172
120	=item ev_sleep (ev_tstamp interval)	173	=item ev_sleep (ev_tstamp interval)
121		174
122	Sleep for the given interval: The current thread will be blocked until	175	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	176	either it is interrupted or the given time interval has passed. Basically
124	this is a subsecond-resolution C<sleep ()>.	177	this is a sub-second-resolution C<sleep ()>.
125		178
126	=item int ev_version_major ()	179	=item int ev_version_major ()
127		180
128	=item int ev_version_minor ()	181	=item int ev_version_minor ()
129		182
…		…
140	as this indicates an incompatible change. Minor versions are usually	193	as this indicates an incompatible change. Minor versions are usually
141	compatible to older versions, so a larger minor version alone is usually	194	compatible to older versions, so a larger minor version alone is usually
142	not a problem.	195	not a problem.
143		196
144	Example: Make sure we haven't accidentally been linked against the wrong	197	Example: Make sure we haven't accidentally been linked against the wrong
145	version.	198	version (note, however, that this will not detect other ABI mismatches,
		199	such as LFS or reentrancy).
146		200
147	assert (("libev version mismatch",	201	assert (("libev version mismatch",
148	ev_version_major () == EV_VERSION_MAJOR	202	ev_version_major () == EV_VERSION_MAJOR
149	&& ev_version_minor () >= EV_VERSION_MINOR));	203	&& ev_version_minor () >= EV_VERSION_MINOR));
150		204
151	=item unsigned int ev_supported_backends ()	205	=item unsigned int ev_supported_backends ()
152		206
153	Return the set of all backends (i.e. their corresponding C<EV_BACKEND_*>	207	Return the set of all backends (i.e. their corresponding C<EV_BACKEND_*>
154	value) compiled into this binary of libev (independent of their	208	value) compiled into this binary of libev (independent of their
…		…
156	a description of the set values.	210	a description of the set values.
157		211
158	Example: make sure we have the epoll method, because yeah this is cool and	212	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	213	a must have and can we have a torrent of it please!!!11
160		214
161	assert (("sorry, no epoll, no sex",	215	assert (("sorry, no epoll, no sex",
162	ev_supported_backends () & EVBACKEND_EPOLL));	216	ev_supported_backends () & EVBACKEND_EPOLL));
163		217
164	=item unsigned int ev_recommended_backends ()	218	=item unsigned int ev_recommended_backends ()
165		219
166	Return the set of all backends compiled into this binary of libev and also	220	Return the set of all backends compiled into this binary of libev and
167	recommended for this platform. This set is often smaller than the one	221	also recommended for this platform, meaning it will work for most file
		222	descriptor types. This set is often smaller than the one returned by
168	returned by C<ev_supported_backends>, as for example kqueue is broken on	223	C<ev_supported_backends>, as for example kqueue is broken on most BSDs
169	most BSDs and will not be autodetected unless you explicitly request it	224	and will not be auto-detected unless you explicitly request it (assuming
170	(assuming you know what you are doing). This is the set of backends that	225	you know what you are doing). This is the set of backends that libev will
171	libev will probe for if you specify no backends explicitly.	226	probe for if you specify no backends explicitly.
172		227
173	=item unsigned int ev_embeddable_backends ()	228	=item unsigned int ev_embeddable_backends ()
174		229
175	Returns the set of backends that are embeddable in other event loops. This	230	Returns the set of backends that are embeddable in other event loops. This
176	is the theoretical, all-platform, value. To find which backends	231	value is platform-specific but can include backends not available on the
177	might be supported on the current system, you would need to look at	232	current system. To find which embeddable backends might be supported on
178	C<ev_embeddable_backends () & ev_supported_backends ()>, likewise for	233	the current system, you would need to look at C<ev_embeddable_backends ()
179	recommended ones.	234	& ev_supported_backends ()>, likewise for recommended ones.
180		235
181	See the description of C<ev_embed> watchers for more info.	236	See the description of C<ev_embed> watchers for more info.
182		237
183	=item ev_set_allocator (void *(*cb)(void *ptr, long size))	238	=item ev_set_allocator (void *(*cb)(void *ptr, long size)) [NOT REENTRANT]
184		239
185	Sets the allocation function to use (the prototype is similar - the	240	Sets the allocation function to use (the prototype is similar - the
186	semantics is identical - to the realloc C function). It is used to	241	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	242	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	243	when memory needs to be allocated (C<size != 0>), the library might abort
189	potentially destructive action. The default is your system realloc	244	or take some potentially destructive action.
190	function.	245
		246	Since some systems (at least OpenBSD and Darwin) fail to implement
		247	correct C<realloc> semantics, libev will use a wrapper around the system
		248	C<realloc> and C<free> functions by default.
191		249
192	You could override this function in high-availability programs to, say,	250	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,	251	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.	252	or even to sleep a while and retry until some memory is available.
195		253
196	Example: Replace the libev allocator with one that waits a bit and then	254	Example: Replace the libev allocator with one that waits a bit and then
197	retries).	255	retries (example requires a standards-compliant C<realloc>).
198		256
199	static void *	257	static void *
200	persistent_realloc (void *ptr, size_t size)	258	persistent_realloc (void *ptr, size_t size)
201	{	259	{
202	for (;;)	260	for (;;)
…		…
211	}	269	}
212		270
213	...	271	...
214	ev_set_allocator (persistent_realloc);	272	ev_set_allocator (persistent_realloc);
215		273
216	=item ev_set_syserr_cb (void (*cb)(const char *msg));	274	=item ev_set_syserr_cb (void (*cb)(const char *msg)); [NOT REENTRANT]
217		275
218	Set the callback function to call on a retryable syscall error (such	276	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	277	as failed select, poll, epoll_wait). The message is a printable string
220	indicating the system call or subsystem causing the problem. If this	278	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	279	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	280	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	281	requested operation, or, if the condition doesn't go away, do bad stuff
224	(such as abort).	282	(such as abort).
225		283
226	Example: This is basically the same thing that libev does internally, too.	284	Example: This is basically the same thing that libev does internally, too.
…		…
235	...	293	...
236	ev_set_syserr_cb (fatal_error);	294	ev_set_syserr_cb (fatal_error);
237		295
238	=back	296	=back
239		297
240	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP	298	=head1 FUNCTIONS CONTROLLING EVENT LOOPS
241		299
242	An event loop is described by a C<struct ev_loop *>. The library knows two	300	An event loop is described by a C<struct ev_loop *> (the C<struct> is
243	types of such loops, the I<default> loop, which supports signals and child	301	I<not> optional in this case unless libev 3 compatibility is disabled, as
244	events, and dynamically created loops which do not.	302	libev 3 had an C<ev_loop> function colliding with the struct name).
245		303
246	If you use threads, a common model is to run the default event loop	304	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	305	supports signals and child events, and dynamically created event loops
248	create, you also create another event loop. Libev itself does no locking	306	which do 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		307
253	=over 4	308	=over 4
254		309
255	=item struct ev_loop *ev_default_loop (unsigned int flags)	310	=item struct ev_loop *ev_default_loop (unsigned int flags)
256		311
257	This will initialise the default event loop if it hasn't been initialised	312	This returns the "default" event loop object, which is what you should
258	yet and return it. If the default loop could not be initialised, returns	313	normally use when you just need "the event loop". Event loop objects and
259	false. If it already was initialised it simply returns it (and ignores the	314	the C<flags> parameter are described in more detail in the entry for
260	flags. If that is troubling you, check C<ev_backend ()> afterwards).	315	C<ev_loop_new>.
		316
		317	If the default loop is already initialised then this function simply
		318	returns it (and ignores the flags. If that is troubling you, check
		319	C<ev_backend ()> afterwards). Otherwise it will create it with the given
		320	flags, which should almost always be C<0>, unless the caller is also the
		321	one calling C<ev_run> or otherwise qualifies as "the main program".
261		322
262	If you don't know what event loop to use, use the one returned from this	323	If you don't know what event loop to use, use the one returned from this
263	function.	324	function (or via the C<EV_DEFAULT> macro).
264		325
		326	Note that this function is I<not> thread-safe, so if you want to use it
		327	from multiple threads, you have to employ some kind of mutex (note also
		328	that this case is unlikely, as loops cannot be shared easily between
		329	threads anyway).
		330
265	The default loop is the only loop that can handle C<ev_signal> and	331	The default loop is the only loop that can handle C<ev_child> watchers,
266	C<ev_child> watchers, and to do this, it always registers a handler	332	and to do this, it always registers a handler for C<SIGCHLD>. If this is
267	for C<SIGCHLD>. If this is a problem for your app you can either	333	a problem for your application you can either create a dynamic loop with
268	create a dynamic loop with C<ev_loop_new> that doesn't do that, or you	334	C<ev_loop_new> which doesn't do that, or you can simply overwrite the
269	can simply overwrite the C<SIGCHLD> signal handler I<after> calling	335	C<SIGCHLD> signal handler I<after> calling C<ev_default_init>.
270	C<ev_default_init>.	336
		337	Example: This is the most typical usage.
		338
		339	if (!ev_default_loop (0))
		340	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
		341
		342	Example: Restrict libev to the select and poll backends, and do not allow
		343	environment settings to be taken into account:
		344
		345	ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);
		346
		347	=item struct ev_loop *ev_loop_new (unsigned int flags)
		348
		349	This will create and initialise a new event loop object. If the loop
		350	could not be initialised, returns false.
		351
		352	Note that this function I<is> thread-safe, and one common way to use
		353	libev with threads is indeed to create one loop per thread, and using the
		354	default loop in the "main" or "initial" thread.
271		355
272	The flags argument can be used to specify special behaviour or specific	356	The flags argument can be used to specify special behaviour or specific
273	backends to use, and is usually specified as C<0> (or C<EVFLAG_AUTO>).	357	backends to use, and is usually specified as C<0> (or C<EVFLAG_AUTO>).
274		358
275	The following flags are supported:	359	The following flags are supported:
…		…
281	The default flags value. Use this if you have no clue (it's the right	365	The default flags value. Use this if you have no clue (it's the right
282	thing, believe me).	366	thing, believe me).
283		367
284	=item C<EVFLAG_NOENV>	368	=item C<EVFLAG_NOENV>
285		369
286	If this flag bit is ored into the flag value (or the program runs setuid	370	If this flag bit is or'ed into the flag value (or the program runs setuid
287	or setgid) then libev will I<not> look at the environment variable	371	or setgid) then libev will I<not> look at the environment variable
288	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will	372	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will
289	override the flags completely if it is found in the environment. This is	373	override the flags completely if it is found in the environment. This is
290	useful to try out specific backends to test their performance, or to work	374	useful to try out specific backends to test their performance, or to work
291	around bugs.	375	around bugs.
292		376
293	=item C<EVFLAG_FORKCHECK>	377	=item C<EVFLAG_FORKCHECK>
294		378
295	Instead of calling C<ev_default_fork> or C<ev_loop_fork> manually after	379	Instead of calling C<ev_loop_fork> manually after a fork, you can also
296	a fork, you can also make libev check for a fork in each iteration by	380	make libev check for a fork in each iteration by enabling this flag.
297	enabling this flag.
298		381
299	This works by calling C<getpid ()> on every iteration of the loop,	382	This works by calling C<getpid ()> on every iteration of the loop,
300	and thus this might slow down your event loop if you do a lot of loop	383	and thus this might slow down your event loop if you do a lot of loop
301	iterations and little real work, but is usually not noticeable (on my	384	iterations and little real work, but is usually not noticeable (on my
302	Linux system for example, C<getpid> is actually a simple 5-insn sequence	385	GNU/Linux system for example, C<getpid> is actually a simple 5-insn sequence
303	without a syscall and thus I<very> fast, but my Linux system also has	386	without a system call and thus I<very> fast, but my GNU/Linux system also has
304	C<pthread_atfork> which is even faster).	387	C<pthread_atfork> which is even faster).
305		388
306	The big advantage of this flag is that you can forget about fork (and	389	The big advantage of this flag is that you can forget about fork (and
307	forget about forgetting to tell libev about forking) when you use this	390	forget about forgetting to tell libev about forking) when you use this
308	flag.	391	flag.
309		392
310	This flag setting cannot be overriden or specified in the C<LIBEV_FLAGS>	393	This flag setting cannot be overridden or specified in the C<LIBEV_FLAGS>
311	environment variable.	394	environment variable.
		395
		396	=item C<EVFLAG_NOINOTIFY>
		397
		398	When this flag is specified, then libev will not attempt to use the
		399	I<inotify> API for it's C<ev_stat> watchers. Apart from debugging and
		400	testing, this flag can be useful to conserve inotify file descriptors, as
		401	otherwise each loop using C<ev_stat> watchers consumes one inotify handle.
		402
		403	=item C<EVFLAG_SIGNALFD>
		404
		405	When this flag is specified, then libev will attempt to use the
		406	I<signalfd> API for it's C<ev_signal> (and C<ev_child>) watchers. This API
		407	delivers signals synchronously, which makes it both faster and might make
		408	it possible to get the queued signal data. It can also simplify signal
		409	handling with threads, as long as you properly block signals in your
		410	threads that are not interested in handling them.
		411
		412	Signalfd will not be used by default as this changes your signal mask, and
		413	there are a lot of shoddy libraries and programs (glib's threadpool for
		414	example) that can't properly initialise their signal masks.
312		415
313	=item C<EVBACKEND_SELECT> (value 1, portable select backend)	416	=item C<EVBACKEND_SELECT> (value 1, portable select backend)
314		417
315	This is your standard select(2) backend. Not I<completely> standard, as	418	This is your standard select(2) backend. Not I<completely> standard, as
316	libev tries to roll its own fd_set with no limits on the number of fds,	419	libev tries to roll its own fd_set with no limits on the number of fds,
317	but if that fails, expect a fairly low limit on the number of fds when	420	but if that fails, expect a fairly low limit on the number of fds when
318	using this backend. It doesn't scale too well (O(highest_fd)), but its	421	using this backend. It doesn't scale too well (O(highest_fd)), but its
319	usually the fastest backend for a low number of (low-numbered :) fds.	422	usually the fastest backend for a low number of (low-numbered :) fds.
320		423
321	To get good performance out of this backend you need a high amount of	424	To get good performance out of this backend you need a high amount of
322	parallelity (most of the file descriptors should be busy). If you are	425	parallelism (most of the file descriptors should be busy). If you are
323	writing a server, you should C<accept ()> in a loop to accept as many	426	writing a server, you should C<accept ()> in a loop to accept as many
324	connections as possible during one iteration. You might also want to have	427	connections as possible during one iteration. You might also want to have
325	a look at C<ev_set_io_collect_interval ()> to increase the amount of	428	a look at C<ev_set_io_collect_interval ()> to increase the amount of
326	readyness notifications you get per iteration.	429	readiness notifications you get per iteration.
		430
		431	This backend maps C<EV_READ> to the C<readfds> set and C<EV_WRITE> to the
		432	C<writefds> set (and to work around Microsoft Windows bugs, also onto the
		433	C<exceptfds> set on that platform).
327		434
328	=item C<EVBACKEND_POLL> (value 2, poll backend, available everywhere except on windows)	435	=item C<EVBACKEND_POLL> (value 2, poll backend, available everywhere except on windows)
329		436
330	And this is your standard poll(2) backend. It's more complicated	437	And this is your standard poll(2) backend. It's more complicated
331	than select, but handles sparse fds better and has no artificial	438	than select, but handles sparse fds better and has no artificial
332	limit on the number of fds you can use (except it will slow down	439	limit on the number of fds you can use (except it will slow down
333	considerably with a lot of inactive fds). It scales similarly to select,	440	considerably with a lot of inactive fds). It scales similarly to select,
334	i.e. O(total_fds). See the entry for C<EVBACKEND_SELECT>, above, for	441	i.e. O(total_fds). See the entry for C<EVBACKEND_SELECT>, above, for
335	performance tips.	442	performance tips.
336		443
		444	This backend maps C<EV_READ> to C<POLLIN | POLLERR | POLLHUP>, and
		445	C<EV_WRITE> to C<POLLOUT | POLLERR | POLLHUP>.
		446
337	=item C<EVBACKEND_EPOLL> (value 4, Linux)	447	=item C<EVBACKEND_EPOLL> (value 4, Linux)
		448
		449	Use the linux-specific epoll(7) interface (for both pre- and post-2.6.9
		450	kernels).
338		451
339	For few fds, this backend is a bit little slower than poll and select,	452	For few fds, this backend is a bit little slower than poll and select,
340	but it scales phenomenally better. While poll and select usually scale	453	but it scales phenomenally better. While poll and select usually scale
341	like O(total_fds) where n is the total number of fds (or the highest fd),	454	like O(total_fds) where n is the total number of fds (or the highest fd),
342	epoll scales either O(1) or O(active_fds). The epoll design has a number	455	epoll scales either O(1) or O(active_fds).
343	of shortcomings, such as silently dropping events in some hard-to-detect	456
344	cases and rewiring a syscall per fd change, no fork support and bad	457	The epoll mechanism deserves honorable mention as the most misdesigned
345	support for dup.	458	of the more advanced event mechanisms: mere annoyances include silently
		459	dropping file descriptors, requiring a system call per change per file
		460	descriptor (and unnecessary guessing of parameters), problems with dup and
		461	so on. The biggest issue is fork races, however - if a program forks then
		462	I<both> parent and child process have to recreate the epoll set, which can
		463	take considerable time (one syscall per file descriptor) and is of course
		464	hard to detect.
		465
		466	Epoll is also notoriously buggy - embedding epoll fds I<should> work, but
		467	of course I<doesn't>, and epoll just loves to report events for totally
		468	I<different> file descriptors (even already closed ones, so one cannot
		469	even remove them from the set) than registered in the set (especially
		470	on SMP systems). Libev tries to counter these spurious notifications by
		471	employing an additional generation counter and comparing that against the
		472	events to filter out spurious ones, recreating the set when required. Last
		473	not least, it also refuses to work with some file descriptors which work
		474	perfectly fine with C<select> (files, many character devices...).
346		475
347	While stopping, setting and starting an I/O watcher in the same iteration	476	While stopping, setting and starting an I/O watcher in the same iteration
348	will result in some caching, there is still a syscall per such incident	477	will result in some caching, there is still a system call per such
349	(because the fd could point to a different file description now), so its	478	incident (because the same I<file descriptor> could point to a different
350	best to avoid that. Also, C<dup ()>'ed file descriptors might not work	479	I<file description> now), so its best to avoid that. Also, C<dup ()>'ed
351	very well if you register events for both fds.	480	file descriptors might not work very well if you register events for both
352		481	file descriptors.
353	Please note that epoll sometimes generates spurious notifications, so you
354	need to use non-blocking I/O or other means to avoid blocking when no data
355	(or space) is available.
356		482
357	Best performance from this backend is achieved by not unregistering all	483	Best performance from this backend is achieved by not unregistering all
358	watchers for a file descriptor until it has been closed, if possible, i.e.	484	watchers for a file descriptor until it has been closed, if possible,
359	keep at least one watcher active per fd at all times.	485	i.e. keep at least one watcher active per fd at all times. Stopping and
		486	starting a watcher (without re-setting it) also usually doesn't cause
		487	extra overhead. A fork can both result in spurious notifications as well
		488	as in libev having to destroy and recreate the epoll object, which can
		489	take considerable time and thus should be avoided.
360		490
		491	All this means that, in practice, C<EVBACKEND_SELECT> can be as fast or
		492	faster than epoll for maybe up to a hundred file descriptors, depending on
		493	the usage. So sad.
		494
361	While nominally embeddeble in other event loops, this feature is broken in	495	While nominally embeddable in other event loops, this feature is broken in
362	all kernel versions tested so far.	496	all kernel versions tested so far.
		497
		498	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
		499	C<EVBACKEND_POLL>.
363		500
364	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)	501	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)
365		502
366	Kqueue deserves special mention, as at the time of this writing, it	503	Kqueue deserves special mention, as at the time of this writing, it
367	was broken on all BSDs except NetBSD (usually it doesn't work reliably	504	was broken on all BSDs except NetBSD (usually it doesn't work reliably
368	with anything but sockets and pipes, except on Darwin, where of course	505	with anything but sockets and pipes, except on Darwin, where of course
369	it's completely useless). For this reason it's not being "autodetected"	506	it's completely useless). Unlike epoll, however, whose brokenness
		507	is by design, these kqueue bugs can (and eventually will) be fixed
		508	without API changes to existing programs. For this reason it's not being
370	unless you explicitly specify it explicitly in the flags (i.e. using	509	"auto-detected" unless you explicitly specify it in the flags (i.e. using
371	C<EVBACKEND_KQUEUE>) or libev was compiled on a known-to-be-good (-enough)	510	C<EVBACKEND_KQUEUE>) or libev was compiled on a known-to-be-good (-enough)
372	system like NetBSD.	511	system like NetBSD.
373		512
374	You still can embed kqueue into a normal poll or select backend and use it	513	You still can embed kqueue into a normal poll or select backend and use it
375	only for sockets (after having made sure that sockets work with kqueue on	514	only for sockets (after having made sure that sockets work with kqueue on
376	the target platform). See C<ev_embed> watchers for more info.	515	the target platform). See C<ev_embed> watchers for more info.
377		516
378	It scales in the same way as the epoll backend, but the interface to the	517	It scales in the same way as the epoll backend, but the interface to the
379	kernel is more efficient (which says nothing about its actual speed, of	518	kernel is more efficient (which says nothing about its actual speed, of
380	course). While stopping, setting and starting an I/O watcher does never	519	course). While stopping, setting and starting an I/O watcher does never
381	cause an extra syscall as with C<EVBACKEND_EPOLL>, it still adds up to	520	cause an extra system call as with C<EVBACKEND_EPOLL>, it still adds up to
382	two event changes per incident, support for C<fork ()> is very bad and it	521	two event changes per incident. Support for C<fork ()> is very bad (but
383	drops fds silently in similarly hard-to-detect cases.	522	sane, unlike epoll) and it drops fds silently in similarly hard-to-detect
		523	cases
384		524
385	This backend usually performs well under most conditions.	525	This backend usually performs well under most conditions.
386		526
387	While nominally embeddable in other event loops, this doesn't work	527	While nominally embeddable in other event loops, this doesn't work
388	everywhere, so you might need to test for this. And since it is broken	528	everywhere, so you might need to test for this. And since it is broken
389	almost everywhere, you should only use it when you have a lot of sockets	529	almost everywhere, you should only use it when you have a lot of sockets
390	(for which it usually works), by embedding it into another event loop	530	(for which it usually works), by embedding it into another event loop
391	(e.g. C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>) and using it only for	531	(e.g. C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> (but C<poll> is of course
392	sockets.	532	also broken on OS X)) and, did I mention it, using it only for sockets.
		533
		534	This backend maps C<EV_READ> into an C<EVFILT_READ> kevent with
		535	C<NOTE_EOF>, and C<EV_WRITE> into an C<EVFILT_WRITE> kevent with
		536	C<NOTE_EOF>.
393		537
394	=item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8)	538	=item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8)
395		539
396	This is not implemented yet (and might never be, unless you send me an	540	This is not implemented yet (and might never be, unless you send me an
397	implementation). According to reports, C</dev/poll> only supports sockets	541	implementation). According to reports, C</dev/poll> only supports sockets
…		…
401	=item C<EVBACKEND_PORT> (value 32, Solaris 10)	545	=item C<EVBACKEND_PORT> (value 32, Solaris 10)
402		546
403	This uses the Solaris 10 event port mechanism. As with everything on Solaris,	547	This uses the Solaris 10 event port mechanism. As with everything on Solaris,
404	it's really slow, but it still scales very well (O(active_fds)).	548	it's really slow, but it still scales very well (O(active_fds)).
405		549
406	Please note that solaris event ports can deliver a lot of spurious	550	Please note that Solaris event ports can deliver a lot of spurious
407	notifications, so you need to use non-blocking I/O or other means to avoid	551	notifications, so you need to use non-blocking I/O or other means to avoid
408	blocking when no data (or space) is available.	552	blocking when no data (or space) is available.
409		553
410	While this backend scales well, it requires one system call per active	554	While this backend scales well, it requires one system call per active
411	file descriptor per loop iteration. For small and medium numbers of file	555	file descriptor per loop iteration. For small and medium numbers of file
412	descriptors a "slow" C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> backend	556	descriptors a "slow" C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> backend
413	might perform better.	557	might perform better.
414		558
415	On the positive side, ignoring the spurious readyness notifications, this	559	On the positive side, with the exception of the spurious readiness
416	backend actually performed to specification in all tests and is fully	560	notifications, this backend actually performed fully to specification
417	embeddable, which is a rare feat among the OS-specific backends.	561	in all tests and is fully embeddable, which is a rare feat among the
		562	OS-specific backends (I vastly prefer correctness over speed hacks).
		563
		564	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
		565	C<EVBACKEND_POLL>.
418		566
419	=item C<EVBACKEND_ALL>	567	=item C<EVBACKEND_ALL>
420		568
421	Try all backends (even potentially broken ones that wouldn't be tried	569	Try all backends (even potentially broken ones that wouldn't be tried
422	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as	570	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as
…		…
424		572
425	It is definitely not recommended to use this flag.	573	It is definitely not recommended to use this flag.
426		574
427	=back	575	=back
428		576
429	If one or more of these are ored into the flags value, then only these	577	If one or more of the backend flags are or'ed into the flags value,
430	backends will be tried (in the reverse order as listed here). If none are	578	then only these backends will be tried (in the reverse order as listed
431	specified, all backends in C<ev_recommended_backends ()> will be tried.	579	here). If none are specified, all backends in C<ev_recommended_backends
432		580	()> will be tried.
433	The most typical usage is like this:
434
435	if (!ev_default_loop (0))
436	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
437
438	Restrict libev to the select and poll backends, and do not allow
439	environment settings to be taken into account:
440
441	ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);
442
443	Use whatever libev has to offer, but make sure that kqueue is used if
444	available (warning, breaks stuff, best use only with your own private
445	event loop and only if you know the OS supports your types of fds):
446
447	ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);
448
449	=item struct ev_loop *ev_loop_new (unsigned int flags)
450
451	Similar to C<ev_default_loop>, but always creates a new event loop that is
452	always distinct from the default loop. Unlike the default loop, it cannot
453	handle signal and child watchers, and attempts to do so will be greeted by
454	undefined behaviour (or a failed assertion if assertions are enabled).
455		581
456	Example: Try to create a event loop that uses epoll and nothing else.	582	Example: Try to create a event loop that uses epoll and nothing else.
457		583
458	struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);	584	struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);
459	if (!epoller)	585	if (!epoller)
460	fatal ("no epoll found here, maybe it hides under your chair");	586	fatal ("no epoll found here, maybe it hides under your chair");
461		587
		588	Example: Use whatever libev has to offer, but make sure that kqueue is
		589	used if available.
		590
		591	struct ev_loop *loop = ev_loop_new (ev_recommended_backends () | EVBACKEND_KQUEUE);
		592
462	=item ev_default_destroy ()	593	=item ev_loop_destroy (loop)
463		594
464	Destroys the default loop again (frees all memory and kernel state	595	Destroys an event loop object (frees all memory and kernel state
465	etc.). None of the active event watchers will be stopped in the normal	596	etc.). None of the active event watchers will be stopped in the normal
466	sense, so e.g. C<ev_is_active> might still return true. It is your	597	sense, so e.g. C<ev_is_active> might still return true. It is your
467	responsibility to either stop all watchers cleanly yoursef I<before>	598	responsibility to either stop all watchers cleanly yourself I<before>
468	calling this function, or cope with the fact afterwards (which is usually	599	calling this function, or cope with the fact afterwards (which is usually
469	the easiest thing, you can just ignore the watchers and/or C<free ()> them	600	the easiest thing, you can just ignore the watchers and/or C<free ()> them
470	for example).	601	for example).
471		602
472	Note that certain global state, such as signal state, will not be freed by	603	Note that certain global state, such as signal state (and installed signal
473	this function, and related watchers (such as signal and child watchers)	604	handlers), will not be freed by this function, and related watchers (such
474	would need to be stopped manually.	605	as signal and child watchers) would need to be stopped manually.
475		606
476	In general it is not advisable to call this function except in the	607	This function is normally used on loop objects allocated by
477	rare occasion where you really need to free e.g. the signal handling	608	C<ev_loop_new>, but it can also be used on the default loop returned by
		609	C<ev_default_loop>, in which case it is not thread-safe.
		610
		611	Note that it is not advisable to call this function on the default loop
		612	except in the rare occasion where you really need to free it's resources.
478	pipe fds. If you need dynamically allocated loops it is better to use	613	If you need dynamically allocated loops it is better to use C<ev_loop_new>
479	C<ev_loop_new> and C<ev_loop_destroy>).	614	and C<ev_loop_destroy>.
480		615
481	=item ev_loop_destroy (loop)	616	=item ev_loop_fork (loop)
482		617
483	Like C<ev_default_destroy>, but destroys an event loop created by an
484	earlier call to C<ev_loop_new>.
485
486	=item ev_default_fork ()
487
488	This function sets a flag that causes subsequent C<ev_loop> iterations	618	This function sets a flag that causes subsequent C<ev_run> iterations to
489	to reinitialise the kernel state for backends that have one. Despite the	619	reinitialise the kernel state for backends that have one. Despite the
490	name, you can call it anytime, but it makes most sense after forking, in	620	name, you can call it anytime, but it makes most sense after forking, in
491	the child process (or both child and parent, but that again makes little	621	the child process. You I<must> call it (or use C<EVFLAG_FORKCHECK>) in the
492	sense). You I<must> call it in the child before using any of the libev	622	child before resuming or calling C<ev_run>.
493	functions, and it will only take effect at the next C<ev_loop> iteration.	623
		624	Again, you I<have> to call it on I<any> loop that you want to re-use after
		625	a fork, I<even if you do not plan to use the loop in the parent>. This is
		626	because some kernel interfaces *cough* I<kqueue> *cough* do funny things
		627	during fork.
494		628
495	On the other hand, you only need to call this function in the child	629	On the other hand, you only need to call this function in the child
496	process if and only if you want to use the event library in the child. If	630	process if and only if you want to use the event loop in the child. If
497	you just fork+exec, you don't have to call it at all.	631	you just fork+exec or create a new loop in the child, you don't have to
		632	call it at all (in fact, C<epoll> is so badly broken that it makes a
		633	difference, but libev will usually detect this case on its own and do a
		634	costly reset of the backend).
498		635
499	The function itself is quite fast and it's usually not a problem to call	636	The function itself is quite fast and it's usually not a problem to call
500	it just in case after a fork. To make this easy, the function will fit in	637	it just in case after a fork.
501	quite nicely into a call to C<pthread_atfork>:
502		638
		639	Example: Automate calling C<ev_loop_fork> on the default loop when
		640	using pthreads.
		641
		642	static void
		643	post_fork_child (void)
		644	{
		645	ev_loop_fork (EV_DEFAULT);
		646	}
		647
		648	...
503	pthread_atfork (0, 0, ev_default_fork);	649	pthread_atfork (0, 0, post_fork_child);
504		650
505	=item ev_loop_fork (loop)	651	=item int ev_is_default_loop (loop)
506		652
507	Like C<ev_default_fork>, but acts on an event loop created by	653	Returns true when the given loop is, in fact, the default loop, and false
508	C<ev_loop_new>. Yes, you have to call this on every allocated event loop	654	otherwise.
509	after fork, and how you do this is entirely your own problem.
510		655
511	=item unsigned int ev_loop_count (loop)	656	=item unsigned int ev_iteration (loop)
512		657
513	Returns the count of loop iterations for the loop, which is identical to	658	Returns the current iteration count for the event loop, which is identical
514	the number of times libev did poll for new events. It starts at C<0> and	659	to the number of times libev did poll for new events. It starts at C<0>
515	happily wraps around with enough iterations.	660	and happily wraps around with enough iterations.
516		661
517	This value can sometimes be useful as a generation counter of sorts (it	662	This value can sometimes be useful as a generation counter of sorts (it
518	"ticks" the number of loop iterations), as it roughly corresponds with	663	"ticks" the number of loop iterations), as it roughly corresponds with
519	C<ev_prepare> and C<ev_check> calls.	664	C<ev_prepare> and C<ev_check> calls - and is incremented between the
		665	prepare and check phases.
		666
		667	=item unsigned int ev_depth (loop)
		668
		669	Returns the number of times C<ev_run> was entered minus the number of
		670	times C<ev_run> was exited, in other words, the recursion depth.
		671
		672	Outside C<ev_run>, this number is zero. In a callback, this number is
		673	C<1>, unless C<ev_run> was invoked recursively (or from another thread),
		674	in which case it is higher.
		675
		676	Leaving C<ev_run> abnormally (setjmp/longjmp, cancelling the thread
		677	etc.), doesn't count as "exit" - consider this as a hint to avoid such
		678	ungentleman-like behaviour unless it's really convenient.
520		679
521	=item unsigned int ev_backend (loop)	680	=item unsigned int ev_backend (loop)
522		681
523	Returns one of the C<EVBACKEND_*> flags indicating the event backend in	682	Returns one of the C<EVBACKEND_*> flags indicating the event backend in
524	use.	683	use.
…		…
529	received events and started processing them. This timestamp does not	688	received events and started processing them. This timestamp does not
530	change as long as callbacks are being processed, and this is also the base	689	change as long as callbacks are being processed, and this is also the base
531	time used for relative timers. You can treat it as the timestamp of the	690	time used for relative timers. You can treat it as the timestamp of the
532	event occurring (or more correctly, libev finding out about it).	691	event occurring (or more correctly, libev finding out about it).
533		692
		693	=item ev_now_update (loop)
		694
		695	Establishes the current time by querying the kernel, updating the time
		696	returned by C<ev_now ()> in the progress. This is a costly operation and
		697	is usually done automatically within C<ev_run ()>.
		698
		699	This function is rarely useful, but when some event callback runs for a
		700	very long time without entering the event loop, updating libev's idea of
		701	the current time is a good idea.
		702
		703	See also L<The special problem of time updates> in the C<ev_timer> section.
		704
		705	=item ev_suspend (loop)
		706
		707	=item ev_resume (loop)
		708
		709	These two functions suspend and resume an event loop, for use when the
		710	loop is not used for a while and timeouts should not be processed.
		711
		712	A typical use case would be an interactive program such as a game: When
		713	the user presses C<^Z> to suspend the game and resumes it an hour later it
		714	would be best to handle timeouts as if no time had actually passed while
		715	the program was suspended. This can be achieved by calling C<ev_suspend>
		716	in your C<SIGTSTP> handler, sending yourself a C<SIGSTOP> and calling
		717	C<ev_resume> directly afterwards to resume timer processing.
		718
		719	Effectively, all C<ev_timer> watchers will be delayed by the time spend
		720	between C<ev_suspend> and C<ev_resume>, and all C<ev_periodic> watchers
		721	will be rescheduled (that is, they will lose any events that would have
		722	occurred while suspended).
		723
		724	After calling C<ev_suspend> you B<must not> call I<any> function on the
		725	given loop other than C<ev_resume>, and you B<must not> call C<ev_resume>
		726	without a previous call to C<ev_suspend>.
		727
		728	Calling C<ev_suspend>/C<ev_resume> has the side effect of updating the
		729	event loop time (see C<ev_now_update>).
		730
534	=item ev_loop (loop, int flags)	731	=item ev_run (loop, int flags)
535		732
536	Finally, this is it, the event handler. This function usually is called	733	Finally, this is it, the event handler. This function usually is called
537	after you initialised all your watchers and you want to start handling	734	after you have initialised all your watchers and you want to start
538	events.	735	handling events. It will ask the operating system for any new events, call
		736	the watcher callbacks, an then repeat the whole process indefinitely: This
		737	is why event loops are called I<loops>.
539		738
540	If the flags argument is specified as C<0>, it will not return until	739	If the flags argument is specified as C<0>, it will keep handling events
541	either no event watchers are active anymore or C<ev_unloop> was called.	740	until either no event watchers are active anymore or C<ev_break> was
		741	called.
542		742
543	Please note that an explicit C<ev_unloop> is usually better than	743	Please note that an explicit C<ev_break> is usually better than
544	relying on all watchers to be stopped when deciding when a program has	744	relying on all watchers to be stopped when deciding when a program has
545	finished (especially in interactive programs), but having a program that	745	finished (especially in interactive programs), but having a program
546	automatically loops as long as it has to and no longer by virtue of	746	that automatically loops as long as it has to and no longer by virtue
547	relying on its watchers stopping correctly is a thing of beauty.	747	of relying on its watchers stopping correctly, that is truly a thing of
		748	beauty.
548		749
549	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle	750	A flags value of C<EVRUN_NOWAIT> will look for new events, will handle
550	those events and any outstanding ones, but will not block your process in	751	those events and any already outstanding ones, but will not wait and
551	case there are no events and will return after one iteration of the loop.	752	block your process in case there are no events and will return after one
		753	iteration of the loop. This is sometimes useful to poll and handle new
		754	events while doing lengthy calculations, to keep the program responsive.
552		755
553	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if	756	A flags value of C<EVRUN_ONCE> will look for new events (waiting if
554	neccessary) and will handle those and any outstanding ones. It will block	757	necessary) and will handle those and any already outstanding ones. It
555	your process until at least one new event arrives, and will return after	758	will block your process until at least one new event arrives (which could
556	one iteration of the loop. This is useful if you are waiting for some	759	be an event internal to libev itself, so there is no guarantee that a
557	external event in conjunction with something not expressible using other	760	user-registered callback will be called), and will return after one
		761	iteration of the loop.
		762
		763	This is useful if you are waiting for some external event in conjunction
		764	with something not expressible using other libev watchers (i.e. "roll your
558	libev watchers. However, a pair of C<ev_prepare>/C<ev_check> watchers is	765	own C<ev_run>"). However, a pair of C<ev_prepare>/C<ev_check> watchers is
559	usually a better approach for this kind of thing.	766	usually a better approach for this kind of thing.
560		767
561	Here are the gory details of what C<ev_loop> does:	768	Here are the gory details of what C<ev_run> does:
562		769
		770	- Increment loop depth.
		771	- Reset the ev_break status.
563	- Before the first iteration, call any pending watchers.	772	- Before the first iteration, call any pending watchers.
		773	LOOP:
564	* If EVFLAG_FORKCHECK was used, check for a fork.	774	- If EVFLAG_FORKCHECK was used, check for a fork.
565	- If a fork was detected, queue and call all fork watchers.	775	- If a fork was detected (by any means), queue and call all fork watchers.
566	- Queue and call all prepare watchers.	776	- Queue and call all prepare watchers.
		777	- If ev_break was called, goto FINISH.
567	- If we have been forked, recreate the kernel state.	778	- If we have been forked, detach and recreate the kernel state
		779	as to not disturb the other process.
568	- Update the kernel state with all outstanding changes.	780	- Update the kernel state with all outstanding changes.
569	- Update the "event loop time".	781	- Update the "event loop time" (ev_now ()).
570	- Calculate for how long to sleep or block, if at all	782	- Calculate for how long to sleep or block, if at all
571	(active idle watchers, EVLOOP_NONBLOCK or not having	783	(active idle watchers, EVRUN_NOWAIT or not having
572	any active watchers at all will result in not sleeping).	784	any active watchers at all will result in not sleeping).
573	- Sleep if the I/O and timer collect interval say so.	785	- Sleep if the I/O and timer collect interval say so.
		786	- Increment loop iteration counter.
574	- Block the process, waiting for any events.	787	- Block the process, waiting for any events.
575	- Queue all outstanding I/O (fd) events.	788	- Queue all outstanding I/O (fd) events.
576	- Update the "event loop time" and do time jump handling.	789	- Update the "event loop time" (ev_now ()), and do time jump adjustments.
577	- Queue all outstanding timers.	790	- Queue all expired timers.
578	- Queue all outstanding periodics.	791	- Queue all expired periodics.
579	- If no events are pending now, queue all idle watchers.	792	- Queue all idle watchers with priority higher than that of pending events.
580	- Queue all check watchers.	793	- Queue all check watchers.
581	- Call all queued watchers in reverse order (i.e. check watchers first).	794	- Call all queued watchers in reverse order (i.e. check watchers first).
582	Signals and child watchers are implemented as I/O watchers, and will	795	Signals and child watchers are implemented as I/O watchers, and will
583	be handled here by queueing them when their watcher gets executed.	796	be handled here by queueing them when their watcher gets executed.
584	- If ev_unloop has been called, or EVLOOP_ONESHOT or EVLOOP_NONBLOCK	797	- If ev_break has been called, or EVRUN_ONCE or EVRUN_NOWAIT
585	were used, or there are no active watchers, return, otherwise	798	were used, or there are no active watchers, goto FINISH, otherwise
586	continue with step *.	799	continue with step LOOP.
		800	FINISH:
		801	- Reset the ev_break status iff it was EVBREAK_ONE.
		802	- Decrement the loop depth.
		803	- Return.
587		804
588	Example: Queue some jobs and then loop until no events are outstanding	805	Example: Queue some jobs and then loop until no events are outstanding
589	anymore.	806	anymore.
590		807
591	... queue jobs here, make sure they register event watchers as long	808	... queue jobs here, make sure they register event watchers as long
592	... as they still have work to do (even an idle watcher will do..)	809	... as they still have work to do (even an idle watcher will do..)
593	ev_loop (my_loop, 0);	810	ev_run (my_loop, 0);
594	... jobs done. yeah!	811	... jobs done or somebody called unloop. yeah!
595		812
596	=item ev_unloop (loop, how)	813	=item ev_break (loop, how)
597		814
598	Can be used to make a call to C<ev_loop> return early (but only after it	815	Can be used to make a call to C<ev_run> return early (but only after it
599	has processed all outstanding events). The C<how> argument must be either	816	has processed all outstanding events). The C<how> argument must be either
600	C<EVUNLOOP_ONE>, which will make the innermost C<ev_loop> call return, or	817	C<EVBREAK_ONE>, which will make the innermost C<ev_run> call return, or
601	C<EVUNLOOP_ALL>, which will make all nested C<ev_loop> calls return.	818	C<EVBREAK_ALL>, which will make all nested C<ev_run> calls return.
602		819
603	This "unloop state" will be cleared when entering C<ev_loop> again.	820	This "unloop state" will be cleared when entering C<ev_run> again.
		821
		822	It is safe to call C<ev_break> from outside any C<ev_run> calls. ##TODO##
604		823
605	=item ev_ref (loop)	824	=item ev_ref (loop)
606		825
607	=item ev_unref (loop)	826	=item ev_unref (loop)
608		827
609	Ref/unref can be used to add or remove a reference count on the event	828	Ref/unref can be used to add or remove a reference count on the event
610	loop: Every watcher keeps one reference, and as long as the reference	829	loop: Every watcher keeps one reference, and as long as the reference
611	count is nonzero, C<ev_loop> will not return on its own. If you have	830	count is nonzero, C<ev_run> will not return on its own.
612	a watcher you never unregister that should not keep C<ev_loop> from	831
613	returning, ev_unref() after starting, and ev_ref() before stopping it. For	832	This is useful when you have a watcher that you never intend to
		833	unregister, but that nevertheless should not keep C<ev_run> from
		834	returning. In such a case, call C<ev_unref> after starting, and C<ev_ref>
		835	before stopping it.
		836
614	example, libev itself uses this for its internal signal pipe: It is not	837	As an example, libev itself uses this for its internal signal pipe: It
615	visible to the libev user and should not keep C<ev_loop> from exiting if	838	is not visible to the libev user and should not keep C<ev_run> from
616	no event watchers registered by it are active. It is also an excellent	839	exiting if no event watchers registered by it are active. It is also an
617	way to do this for generic recurring timers or from within third-party	840	excellent way to do this for generic recurring timers or from within
618	libraries. Just remember to I<unref after start> and I<ref before stop>	841	third-party libraries. Just remember to I<unref after start> and I<ref
619	(but only if the watcher wasn't active before, or was active before,	842	before stop> (but only if the watcher wasn't active before, or was active
620	respectively).	843	before, respectively. Note also that libev might stop watchers itself
		844	(e.g. non-repeating timers) in which case you have to C<ev_ref>
		845	in the callback).
621		846
622	Example: Create a signal watcher, but keep it from keeping C<ev_loop>	847	Example: Create a signal watcher, but keep it from keeping C<ev_run>
623	running when nothing else is active.	848	running when nothing else is active.
624		849
625	struct ev_signal exitsig;	850	ev_signal exitsig;
626	ev_signal_init (&exitsig, sig_cb, SIGINT);	851	ev_signal_init (&exitsig, sig_cb, SIGINT);
627	ev_signal_start (loop, &exitsig);	852	ev_signal_start (loop, &exitsig);
628	evf_unref (loop);	853	evf_unref (loop);
629		854
630	Example: For some weird reason, unregister the above signal handler again.	855	Example: For some weird reason, unregister the above signal handler again.
631		856
632	ev_ref (loop);	857	ev_ref (loop);
633	ev_signal_stop (loop, &exitsig);	858	ev_signal_stop (loop, &exitsig);
634		859
635	=item ev_set_io_collect_interval (loop, ev_tstamp interval)	860	=item ev_set_io_collect_interval (loop, ev_tstamp interval)
636		861
637	=item ev_set_timeout_collect_interval (loop, ev_tstamp interval)	862	=item ev_set_timeout_collect_interval (loop, ev_tstamp interval)
638		863
639	These advanced functions influence the time that libev will spend waiting	864	These advanced functions influence the time that libev will spend waiting
640	for events. Both are by default C<0>, meaning that libev will try to	865	for events. Both time intervals are by default C<0>, meaning that libev
641	invoke timer/periodic callbacks and I/O callbacks with minimum latency.	866	will try to invoke timer/periodic callbacks and I/O callbacks with minimum
		867	latency.
642		868
643	Setting these to a higher value (the C<interval> I<must> be >= C<0>)	869	Setting these to a higher value (the C<interval> I<must> be >= C<0>)
644	allows libev to delay invocation of I/O and timer/periodic callbacks to	870	allows libev to delay invocation of I/O and timer/periodic callbacks
645	increase efficiency of loop iterations.	871	to increase efficiency of loop iterations (or to increase power-saving
		872	opportunities).
646		873
647	The background is that sometimes your program runs just fast enough to	874	The idea is that sometimes your program runs just fast enough to handle
648	handle one (or very few) event(s) per loop iteration. While this makes	875	one (or very few) event(s) per loop iteration. While this makes the
649	the program responsive, it also wastes a lot of CPU time to poll for new	876	program responsive, it also wastes a lot of CPU time to poll for new
650	events, especially with backends like C<select ()> which have a high	877	events, especially with backends like C<select ()> which have a high
651	overhead for the actual polling but can deliver many events at once.	878	overhead for the actual polling but can deliver many events at once.
652		879
653	By setting a higher I<io collect interval> you allow libev to spend more	880	By setting a higher I<io collect interval> you allow libev to spend more
654	time collecting I/O events, so you can handle more events per iteration,	881	time collecting I/O events, so you can handle more events per iteration,
655	at the cost of increasing latency. Timeouts (both C<ev_periodic> and	882	at the cost of increasing latency. Timeouts (both C<ev_periodic> and
656	C<ev_timer>) will be not affected. Setting this to a non-null value will	883	C<ev_timer>) will be not affected. Setting this to a non-null value will
657	introduce an additional C<ev_sleep ()> call into most loop iterations.	884	introduce an additional C<ev_sleep ()> call into most loop iterations. The
		885	sleep time ensures that libev will not poll for I/O events more often then
		886	once per this interval, on average.
658		887
659	Likewise, by setting a higher I<timeout collect interval> you allow libev	888	Likewise, by setting a higher I<timeout collect interval> you allow libev
660	to spend more time collecting timeouts, at the expense of increased	889	to spend more time collecting timeouts, at the expense of increased
661	latency (the watcher callback will be called later). C<ev_io> watchers	890	latency/jitter/inexactness (the watcher callback will be called
662	will not be affected. Setting this to a non-null value will not introduce	891	later). C<ev_io> watchers will not be affected. Setting this to a non-null
663	any overhead in libev.	892	value will not introduce any overhead in libev.
664		893
665	Many (busy) programs can usually benefit by setting the io collect	894	Many (busy) programs can usually benefit by setting the I/O collect
666	interval to a value near C<0.1> or so, which is often enough for	895	interval to a value near C<0.1> or so, which is often enough for
667	interactive servers (of course not for games), likewise for timeouts. It	896	interactive servers (of course not for games), likewise for timeouts. It
668	usually doesn't make much sense to set it to a lower value than C<0.01>,	897	usually doesn't make much sense to set it to a lower value than C<0.01>,
669	as this approsaches the timing granularity of most systems.	898	as this approaches the timing granularity of most systems. Note that if
		899	you do transactions with the outside world and you can't increase the
		900	parallelity, then this setting will limit your transaction rate (if you
		901	need to poll once per transaction and the I/O collect interval is 0.01,
		902	then you can't do more than 100 transactions per second).
		903
		904	Setting the I<timeout collect interval> can improve the opportunity for
		905	saving power, as the program will "bundle" timer callback invocations that
		906	are "near" in time together, by delaying some, thus reducing the number of
		907	times the process sleeps and wakes up again. Another useful technique to
		908	reduce iterations/wake-ups is to use C<ev_periodic> watchers and make sure
		909	they fire on, say, one-second boundaries only.
		910
		911	Example: we only need 0.1s timeout granularity, and we wish not to poll
		912	more often than 100 times per second:
		913
		914	ev_set_timeout_collect_interval (EV_DEFAULT_UC_ 0.1);
		915	ev_set_io_collect_interval (EV_DEFAULT_UC_ 0.01);
		916
		917	=item ev_invoke_pending (loop)
		918
		919	This call will simply invoke all pending watchers while resetting their
		920	pending state. Normally, C<ev_run> does this automatically when required,
		921	but when overriding the invoke callback this call comes handy. This
		922	function can be invoked from a watcher - this can be useful for example
		923	when you want to do some lengthy calculation and want to pass further
		924	event handling to another thread (you still have to make sure only one
		925	thread executes within C<ev_invoke_pending> or C<ev_run> of course).
		926
		927	=item int ev_pending_count (loop)
		928
		929	Returns the number of pending watchers - zero indicates that no watchers
		930	are pending.
		931
		932	=item ev_set_invoke_pending_cb (loop, void (*invoke_pending_cb)(EV_P))
		933
		934	This overrides the invoke pending functionality of the loop: Instead of
		935	invoking all pending watchers when there are any, C<ev_run> will call
		936	this callback instead. This is useful, for example, when you want to
		937	invoke the actual watchers inside another context (another thread etc.).
		938
		939	If you want to reset the callback, use C<ev_invoke_pending> as new
		940	callback.
		941
		942	=item ev_set_loop_release_cb (loop, void (*release)(EV_P), void (*acquire)(EV_P))
		943
		944	Sometimes you want to share the same loop between multiple threads. This
		945	can be done relatively simply by putting mutex_lock/unlock calls around
		946	each call to a libev function.
		947
		948	However, C<ev_run> can run an indefinite time, so it is not feasible
		949	to wait for it to return. One way around this is to wake up the event
		950	loop via C<ev_break> and C<av_async_send>, another way is to set these
		951	I<release> and I<acquire> callbacks on the loop.
		952
		953	When set, then C<release> will be called just before the thread is
		954	suspended waiting for new events, and C<acquire> is called just
		955	afterwards.
		956
		957	Ideally, C<release> will just call your mutex_unlock function, and
		958	C<acquire> will just call the mutex_lock function again.
		959
		960	While event loop modifications are allowed between invocations of
		961	C<release> and C<acquire> (that's their only purpose after all), no
		962	modifications done will affect the event loop, i.e. adding watchers will
		963	have no effect on the set of file descriptors being watched, or the time
		964	waited. Use an C<ev_async> watcher to wake up C<ev_run> when you want it
		965	to take note of any changes you made.
		966
		967	In theory, threads executing C<ev_run> will be async-cancel safe between
		968	invocations of C<release> and C<acquire>.
		969
		970	See also the locking example in the C<THREADS> section later in this
		971	document.
		972
		973	=item ev_set_userdata (loop, void *data)
		974
		975	=item ev_userdata (loop)
		976
		977	Set and retrieve a single C<void *> associated with a loop. When
		978	C<ev_set_userdata> has never been called, then C<ev_userdata> returns
		979	C<0.>
		980
		981	These two functions can be used to associate arbitrary data with a loop,
		982	and are intended solely for the C<invoke_pending_cb>, C<release> and
		983	C<acquire> callbacks described above, but of course can be (ab-)used for
		984	any other purpose as well.
		985
		986	=item ev_verify (loop)
		987
		988	This function only does something when C<EV_VERIFY> support has been
		989	compiled in, which is the default for non-minimal builds. It tries to go
		990	through all internal structures and checks them for validity. If anything
		991	is found to be inconsistent, it will print an error message to standard
		992	error and call C<abort ()>.
		993
		994	This can be used to catch bugs inside libev itself: under normal
		995	circumstances, this function will never abort as of course libev keeps its
		996	data structures consistent.
670		997
671	=back	998	=back
672		999
673		1000
674	=head1 ANATOMY OF A WATCHER	1001	=head1 ANATOMY OF A WATCHER
675		1002
		1003	In the following description, uppercase C<TYPE> in names stands for the
		1004	watcher type, e.g. C<ev_TYPE_start> can mean C<ev_timer_start> for timer
		1005	watchers and C<ev_io_start> for I/O watchers.
		1006
676	A watcher is a structure that you create and register to record your	1007	A watcher is an opaque structure that you allocate and register to record
677	interest in some event. For instance, if you want to wait for STDIN to	1008	your interest in some event. To make a concrete example, imagine you want
678	become readable, you would create an C<ev_io> watcher for that:	1009	to wait for STDIN to become readable, you would create an C<ev_io> watcher
		1010	for that:
679		1011
680	static void my_cb (struct ev_loop *loop, struct ev_io *w, int revents)	1012	static void my_cb (struct ev_loop *loop, ev_io *w, int revents)
681	{	1013	{
682	ev_io_stop (w);	1014	ev_io_stop (w);
683	ev_unloop (loop, EVUNLOOP_ALL);	1015	ev_break (loop, EVBREAK_ALL);
684	}	1016	}
685		1017
686	struct ev_loop *loop = ev_default_loop (0);	1018	struct ev_loop *loop = ev_default_loop (0);
		1019
687	struct ev_io stdin_watcher;	1020	ev_io stdin_watcher;
		1021
688	ev_init (&stdin_watcher, my_cb);	1022	ev_init (&stdin_watcher, my_cb);
689	ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);	1023	ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);
690	ev_io_start (loop, &stdin_watcher);	1024	ev_io_start (loop, &stdin_watcher);
		1025
691	ev_loop (loop, 0);	1026	ev_run (loop, 0);
692		1027
693	As you can see, you are responsible for allocating the memory for your	1028	As you can see, you are responsible for allocating the memory for your
694	watcher structures (and it is usually a bad idea to do this on the stack,	1029	watcher structures (and it is I<usually> a bad idea to do this on the
695	although this can sometimes be quite valid).	1030	stack).
696		1031
		1032	Each watcher has an associated watcher structure (called C<struct ev_TYPE>
		1033	or simply C<ev_TYPE>, as typedefs are provided for all watcher structs).
		1034
697	Each watcher structure must be initialised by a call to C<ev_init	1035	Each watcher structure must be initialised by a call to C<ev_init (watcher
698	(watcher *, callback)>, which expects a callback to be provided. This	1036	*, callback)>, which expects a callback to be provided. This callback is
699	callback gets invoked each time the event occurs (or, in the case of io	1037	invoked each time the event occurs (or, in the case of I/O watchers, each
700	watchers, each time the event loop detects that the file descriptor given	1038	time the event loop detects that the file descriptor given is readable
701	is readable and/or writable).	1039	and/or writable).
702		1040
703	Each watcher type has its own C<< ev_<type>_set (watcher *, ...) >> macro	1041	Each watcher type further has its own C<< ev_TYPE_set (watcher *, ...) >>
704	with arguments specific to this watcher type. There is also a macro	1042	macro to configure it, with arguments specific to the watcher type. There
705	to combine initialisation and setting in one call: C<< ev_<type>_init	1043	is also a macro to combine initialisation and setting in one call: C<<
706	(watcher *, callback, ...) >>.	1044	ev_TYPE_init (watcher *, callback, ...) >>.
707		1045
708	To make the watcher actually watch out for events, you have to start it	1046	To make the watcher actually watch out for events, you have to start it
709	with a watcher-specific start function (C<< ev_<type>_start (loop, watcher	1047	with a watcher-specific start function (C<< ev_TYPE_start (loop, watcher
710	*) >>), and you can stop watching for events at any time by calling the	1048	*) >>), and you can stop watching for events at any time by calling the
711	corresponding stop function (C<< ev_<type>_stop (loop, watcher *) >>.	1049	corresponding stop function (C<< ev_TYPE_stop (loop, watcher *) >>.
712		1050
713	As long as your watcher is active (has been started but not stopped) you	1051	As long as your watcher is active (has been started but not stopped) you
714	must not touch the values stored in it. Most specifically you must never	1052	must not touch the values stored in it. Most specifically you must never
715	reinitialise it or call its C<set> macro.	1053	reinitialise it or call its C<ev_TYPE_set> macro.
716		1054
717	Each and every callback receives the event loop pointer as first, the	1055	Each and every callback receives the event loop pointer as first, the
718	registered watcher structure as second, and a bitset of received events as	1056	registered watcher structure as second, and a bitset of received events as
719	third argument.	1057	third argument.
720		1058
…		…
729	=item C<EV_WRITE>	1067	=item C<EV_WRITE>
730		1068
731	The file descriptor in the C<ev_io> watcher has become readable and/or	1069	The file descriptor in the C<ev_io> watcher has become readable and/or
732	writable.	1070	writable.
733		1071
734	=item C<EV_TIMEOUT>	1072	=item C<EV_TIMER>
735		1073
736	The C<ev_timer> watcher has timed out.	1074	The C<ev_timer> watcher has timed out.
737		1075
738	=item C<EV_PERIODIC>	1076	=item C<EV_PERIODIC>
739		1077
…		…
757		1095
758	=item C<EV_PREPARE>	1096	=item C<EV_PREPARE>
759		1097
760	=item C<EV_CHECK>	1098	=item C<EV_CHECK>
761		1099
762	All C<ev_prepare> watchers are invoked just I<before> C<ev_loop> starts	1100	All C<ev_prepare> watchers are invoked just I<before> C<ev_run> starts
763	to gather new events, and all C<ev_check> watchers are invoked just after	1101	to gather new events, and all C<ev_check> watchers are invoked just after
764	C<ev_loop> has gathered them, but before it invokes any callbacks for any	1102	C<ev_run> has gathered them, but before it invokes any callbacks for any
765	received events. Callbacks of both watcher types can start and stop as	1103	received events. Callbacks of both watcher types can start and stop as
766	many watchers as they want, and all of them will be taken into account	1104	many watchers as they want, and all of them will be taken into account
767	(for example, a C<ev_prepare> watcher might start an idle watcher to keep	1105	(for example, a C<ev_prepare> watcher might start an idle watcher to keep
768	C<ev_loop> from blocking).	1106	C<ev_run> from blocking).
769		1107
770	=item C<EV_EMBED>	1108	=item C<EV_EMBED>
771		1109
772	The embedded event loop specified in the C<ev_embed> watcher needs attention.	1110	The embedded event loop specified in the C<ev_embed> watcher needs attention.
773		1111
774	=item C<EV_FORK>	1112	=item C<EV_FORK>
775		1113
776	The event loop has been resumed in the child process after fork (see	1114	The event loop has been resumed in the child process after fork (see
777	C<ev_fork>).	1115	C<ev_fork>).
778		1116
		1117	=item C<EV_ASYNC>
		1118
		1119	The given async watcher has been asynchronously notified (see C<ev_async>).
		1120
		1121	=item C<EV_CUSTOM>
		1122
		1123	Not ever sent (or otherwise used) by libev itself, but can be freely used
		1124	by libev users to signal watchers (e.g. via C<ev_feed_event>).
		1125
779	=item C<EV_ERROR>	1126	=item C<EV_ERROR>
780		1127
781	An unspecified error has occured, the watcher has been stopped. This might	1128	An unspecified error has occurred, the watcher has been stopped. This might
782	happen because the watcher could not be properly started because libev	1129	happen because the watcher could not be properly started because libev
783	ran out of memory, a file descriptor was found to be closed or any other	1130	ran out of memory, a file descriptor was found to be closed or any other
		1131	problem. Libev considers these application bugs.
		1132
784	problem. You best act on it by reporting the problem and somehow coping	1133	You best act on it by reporting the problem and somehow coping with the
785	with the watcher being stopped.	1134	watcher being stopped. Note that well-written programs should not receive
		1135	an error ever, so when your watcher receives it, this usually indicates a
		1136	bug in your program.
786		1137
787	Libev will usually signal a few "dummy" events together with an error,	1138	Libev will usually signal a few "dummy" events together with an error, for
788	for example it might indicate that a fd is readable or writable, and if	1139	example it might indicate that a fd is readable or writable, and if your
789	your callbacks is well-written it can just attempt the operation and cope	1140	callbacks is well-written it can just attempt the operation and cope with
790	with the error from read() or write(). This will not work in multithreaded	1141	the error from read() or write(). This will not work in multi-threaded
791	programs, though, so beware.	1142	programs, though, as the fd could already be closed and reused for another
		1143	thing, so beware.
792		1144
793	=back	1145	=back
794		1146
		1147	=head2 WATCHER STATES
		1148
		1149	There are various watcher states mentioned throughout this manual -
		1150	active, pending and so on. In this section these states and the rules to
		1151	transition between them will be described in more detail - and while these
		1152	rules might look complicated, they usually do "the right thing".
		1153
		1154	=over 4
		1155
		1156	=item initialiased
		1157
		1158	Before a watcher can be registered with the event looop it has to be
		1159	initialised. This can be done with a call to C<ev_TYPE_init>, or calls to
		1160	C<ev_init> followed by the watcher-specific C<ev_TYPE_set> function.
		1161
		1162	In this state it is simply some block of memory that is suitable for use
		1163	in an event loop. It can be moved around, freed, reused etc. at will.
		1164
		1165	=item started/running/active
		1166
		1167	Once a watcher has been started with a call to C<ev_TYPE_start> it becomes
		1168	property of the event loop, and is actively waiting for events. While in
		1169	this state it cannot be accessed (except in a few documented ways), moved,
		1170	freed or anything else - the only legal thing is to keep a pointer to it,
		1171	and call libev functions on it that are documented to work on active watchers.
		1172
		1173	=item pending
		1174
		1175	If a watcher is active and libev determines that an event it is interested
		1176	in has occurred (such as a timer expiring), it will become pending. It will
		1177	stay in this pending state until either it is stopped or its callback is
		1178	about to be invoked, so it is not normally pending inside the watcher
		1179	callback.
		1180
		1181	The watcher might or might not be active while it is pending (for example,
		1182	an expired non-repeating timer can be pending but no longer active). If it
		1183	is stopped, it can be freely accessed (e.g. by calling C<ev_TYPE_set>),
		1184	but it is still property of the event loop at this time, so cannot be
		1185	moved, freed or reused. And if it is active the rules described in the
		1186	previous item still apply.
		1187
		1188	It is also possible to feed an event on a watcher that is not active (e.g.
		1189	via C<ev_feed_event>), in which case it becomes pending without being
		1190	active.
		1191
		1192	=item stopped
		1193
		1194	A watcher can be stopped implicitly by libev (in which case it might still
		1195	be pending), or explicitly by calling its C<ev_TYPE_stop> function. The
		1196	latter will clear any pending state the watcher might be in, regardless
		1197	of whether it was active or not, so stopping a watcher explicitly before
		1198	freeing it is often a good idea.
		1199
		1200	While stopped (and not pending) the watcher is essentially in the
		1201	initialised state, that is it can be reused, moved, modified in any way
		1202	you wish.
		1203
		1204	=back
		1205
795	=head2 GENERIC WATCHER FUNCTIONS	1206	=head2 GENERIC WATCHER FUNCTIONS
796
797	In the following description, C<TYPE> stands for the watcher type,
798	e.g. C<timer> for C<ev_timer> watchers and C<io> for C<ev_io> watchers.
799		1207
800	=over 4	1208	=over 4
801		1209
802	=item C<ev_init> (ev_TYPE *watcher, callback)	1210	=item C<ev_init> (ev_TYPE *watcher, callback)
803		1211
…		…
809	which rolls both calls into one.	1217	which rolls both calls into one.
810		1218
811	You can reinitialise a watcher at any time as long as it has been stopped	1219	You can reinitialise a watcher at any time as long as it has been stopped
812	(or never started) and there are no pending events outstanding.	1220	(or never started) and there are no pending events outstanding.
813		1221
814	The callback is always of type C<void (*)(ev_loop *loop, ev_TYPE *watcher,	1222	The callback is always of type C<void (*)(struct ev_loop *loop, ev_TYPE *watcher,
815	int revents)>.	1223	int revents)>.
816		1224
		1225	Example: Initialise an C<ev_io> watcher in two steps.
		1226
		1227	ev_io w;
		1228	ev_init (&w, my_cb);
		1229	ev_io_set (&w, STDIN_FILENO, EV_READ);
		1230
817	=item C<ev_TYPE_set> (ev_TYPE *, [args])	1231	=item C<ev_TYPE_set> (ev_TYPE *watcher, [args])
818		1232
819	This macro initialises the type-specific parts of a watcher. You need to	1233	This macro initialises the type-specific parts of a watcher. You need to
820	call C<ev_init> at least once before you call this macro, but you can	1234	call C<ev_init> at least once before you call this macro, but you can
821	call C<ev_TYPE_set> any number of times. You must not, however, call this	1235	call C<ev_TYPE_set> any number of times. You must not, however, call this
822	macro on a watcher that is active (it can be pending, however, which is a	1236	macro on a watcher that is active (it can be pending, however, which is a
823	difference to the C<ev_init> macro).	1237	difference to the C<ev_init> macro).
824		1238
825	Although some watcher types do not have type-specific arguments	1239	Although some watcher types do not have type-specific arguments
826	(e.g. C<ev_prepare>) you still need to call its C<set> macro.	1240	(e.g. C<ev_prepare>) you still need to call its C<set> macro.
827		1241
		1242	See C<ev_init>, above, for an example.
		1243
828	=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])	1244	=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])
829		1245
830	This convinience macro rolls both C<ev_init> and C<ev_TYPE_set> macro	1246	This convenience macro rolls both C<ev_init> and C<ev_TYPE_set> macro
831	calls into a single call. This is the most convinient method to initialise	1247	calls into a single call. This is the most convenient method to initialise
832	a watcher. The same limitations apply, of course.	1248	a watcher. The same limitations apply, of course.
833		1249
		1250	Example: Initialise and set an C<ev_io> watcher in one step.
		1251
		1252	ev_io_init (&w, my_cb, STDIN_FILENO, EV_READ);
		1253
834	=item C<ev_TYPE_start> (loop *, ev_TYPE *watcher)	1254	=item C<ev_TYPE_start> (loop, ev_TYPE *watcher)
835		1255
836	Starts (activates) the given watcher. Only active watchers will receive	1256	Starts (activates) the given watcher. Only active watchers will receive
837	events. If the watcher is already active nothing will happen.	1257	events. If the watcher is already active nothing will happen.
838		1258
		1259	Example: Start the C<ev_io> watcher that is being abused as example in this
		1260	whole section.
		1261
		1262	ev_io_start (EV_DEFAULT_UC, &w);
		1263
839	=item C<ev_TYPE_stop> (loop *, ev_TYPE *watcher)	1264	=item C<ev_TYPE_stop> (loop, ev_TYPE *watcher)
840		1265
841	Stops the given watcher again (if active) and clears the pending	1266	Stops the given watcher if active, and clears the pending status (whether
		1267	the watcher was active or not).
		1268
842	status. It is possible that stopped watchers are pending (for example,	1269	It is possible that stopped watchers are pending - for example,
843	non-repeating timers are being stopped when they become pending), but	1270	non-repeating timers are being stopped when they become pending - but
844	C<ev_TYPE_stop> ensures that the watcher is neither active nor pending. If	1271	calling C<ev_TYPE_stop> ensures that the watcher is neither active nor
845	you want to free or reuse the memory used by the watcher it is therefore a	1272	pending. If you want to free or reuse the memory used by the watcher it is
846	good idea to always call its C<ev_TYPE_stop> function.	1273	therefore a good idea to always call its C<ev_TYPE_stop> function.
847		1274
848	=item bool ev_is_active (ev_TYPE *watcher)	1275	=item bool ev_is_active (ev_TYPE *watcher)
849		1276
850	Returns a true value iff the watcher is active (i.e. it has been started	1277	Returns a true value iff the watcher is active (i.e. it has been started
851	and not yet been stopped). As long as a watcher is active you must not modify	1278	and not yet been stopped). As long as a watcher is active you must not modify
…		…
867	=item ev_cb_set (ev_TYPE *watcher, callback)	1294	=item ev_cb_set (ev_TYPE *watcher, callback)
868		1295
869	Change the callback. You can change the callback at virtually any time	1296	Change the callback. You can change the callback at virtually any time
870	(modulo threads).	1297	(modulo threads).
871		1298
872	=item ev_set_priority (ev_TYPE *watcher, priority)	1299	=item ev_set_priority (ev_TYPE *watcher, int priority)
873		1300
874	=item int ev_priority (ev_TYPE *watcher)	1301	=item int ev_priority (ev_TYPE *watcher)
875		1302
876	Set and query the priority of the watcher. The priority is a small	1303	Set and query the priority of the watcher. The priority is a small
877	integer between C<EV_MAXPRI> (default: C<2>) and C<EV_MINPRI>	1304	integer between C<EV_MAXPRI> (default: C<2>) and C<EV_MINPRI>
878	(default: C<-2>). Pending watchers with higher priority will be invoked	1305	(default: C<-2>). Pending watchers with higher priority will be invoked
879	before watchers with lower priority, but priority will not keep watchers	1306	before watchers with lower priority, but priority will not keep watchers
880	from being executed (except for C<ev_idle> watchers).	1307	from being executed (except for C<ev_idle> watchers).
881		1308
882	This means that priorities are I<only> used for ordering callback
883	invocation after new events have been received. This is useful, for
884	example, to reduce latency after idling, or more often, to bind two
885	watchers on the same event and make sure one is called first.
886
887	If you need to suppress invocation when higher priority events are pending	1309	If you need to suppress invocation when higher priority events are pending
888	you need to look at C<ev_idle> watchers, which provide this functionality.	1310	you need to look at C<ev_idle> watchers, which provide this functionality.
889		1311
890	You I<must not> change the priority of a watcher as long as it is active or	1312	You I<must not> change the priority of a watcher as long as it is active or
891	pending.	1313	pending.
892		1314
		1315	Setting a priority outside the range of C<EV_MINPRI> to C<EV_MAXPRI> is
		1316	fine, as long as you do not mind that the priority value you query might
		1317	or might not have been clamped to the valid range.
		1318
893	The default priority used by watchers when no priority has been set is	1319	The default priority used by watchers when no priority has been set is
894	always C<0>, which is supposed to not be too high and not be too low :).	1320	always C<0>, which is supposed to not be too high and not be too low :).
895		1321
896	Setting a priority outside the range of C<EV_MINPRI> to C<EV_MAXPRI> is	1322	See L<WATCHER PRIORITY MODELS>, below, for a more thorough treatment of
897	fine, as long as you do not mind that the priority value you query might	1323	priorities.
898	or might not have been adjusted to be within valid range.
899		1324
900	=item ev_invoke (loop, ev_TYPE *watcher, int revents)	1325	=item ev_invoke (loop, ev_TYPE *watcher, int revents)
901		1326
902	Invoke the C<watcher> with the given C<loop> and C<revents>. Neither	1327	Invoke the C<watcher> with the given C<loop> and C<revents>. Neither
903	C<loop> nor C<revents> need to be valid as long as the watcher callback	1328	C<loop> nor C<revents> need to be valid as long as the watcher callback
904	can deal with that fact.	1329	can deal with that fact, as both are simply passed through to the
		1330	callback.
905		1331
906	=item int ev_clear_pending (loop, ev_TYPE *watcher)	1332	=item int ev_clear_pending (loop, ev_TYPE *watcher)
907		1333
908	If the watcher is pending, this function returns clears its pending status	1334	If the watcher is pending, this function clears its pending status and
909	and returns its C<revents> bitset (as if its callback was invoked). If the	1335	returns its C<revents> bitset (as if its callback was invoked). If the
910	watcher isn't pending it does nothing and returns C<0>.	1336	watcher isn't pending it does nothing and returns C<0>.
911		1337
		1338	Sometimes it can be useful to "poll" a watcher instead of waiting for its
		1339	callback to be invoked, which can be accomplished with this function.
		1340
		1341	=item ev_feed_event (loop, ev_TYPE *watcher, int revents)
		1342
		1343	Feeds the given event set into the event loop, as if the specified event
		1344	had happened for the specified watcher (which must be a pointer to an
		1345	initialised but not necessarily started event watcher). Obviously you must
		1346	not free the watcher as long as it has pending events.
		1347
		1348	Stopping the watcher, letting libev invoke it, or calling
		1349	C<ev_clear_pending> will clear the pending event, even if the watcher was
		1350	not started in the first place.
		1351
		1352	See also C<ev_feed_fd_event> and C<ev_feed_signal_event> for related
		1353	functions that do not need a watcher.
		1354
912	=back	1355	=back
913		1356
914		1357
915	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER	1358	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER
916		1359
917	Each watcher has, by default, a member C<void *data> that you can change	1360	Each watcher has, by default, a member C<void *data> that you can change
918	and read at any time, libev will completely ignore it. This can be used	1361	and read at any time: libev will completely ignore it. This can be used
919	to associate arbitrary data with your watcher. If you need more data and	1362	to associate arbitrary data with your watcher. If you need more data and
920	don't want to allocate memory and store a pointer to it in that data	1363	don't want to allocate memory and store a pointer to it in that data
921	member, you can also "subclass" the watcher type and provide your own	1364	member, you can also "subclass" the watcher type and provide your own
922	data:	1365	data:
923		1366
924	struct my_io	1367	struct my_io
925	{	1368	{
926	struct ev_io io;	1369	ev_io io;
927	int otherfd;	1370	int otherfd;
928	void *somedata;	1371	void *somedata;
929	struct whatever *mostinteresting;	1372	struct whatever *mostinteresting;
930	}	1373	};
		1374
		1375	...
		1376	struct my_io w;
		1377	ev_io_init (&w.io, my_cb, fd, EV_READ);
931		1378
932	And since your callback will be called with a pointer to the watcher, you	1379	And since your callback will be called with a pointer to the watcher, you
933	can cast it back to your own type:	1380	can cast it back to your own type:
934		1381
935	static void my_cb (struct ev_loop *loop, struct ev_io *w_, int revents)	1382	static void my_cb (struct ev_loop *loop, ev_io *w_, int revents)
936	{	1383	{
937	struct my_io *w = (struct my_io *)w_;	1384	struct my_io *w = (struct my_io *)w_;
938	...	1385	...
939	}	1386	}
940		1387
941	More interesting and less C-conformant ways of casting your callback type	1388	More interesting and less C-conformant ways of casting your callback type
942	instead have been omitted.	1389	instead have been omitted.
943		1390
944	Another common scenario is having some data structure with multiple	1391	Another common scenario is to use some data structure with multiple
945	watchers:	1392	embedded watchers:
946		1393
947	struct my_biggy	1394	struct my_biggy
948	{	1395	{
949	int some_data;	1396	int some_data;
950	ev_timer t1;	1397	ev_timer t1;
951	ev_timer t2;	1398	ev_timer t2;
952	}	1399	}
953		1400
954	In this case getting the pointer to C<my_biggy> is a bit more complicated,	1401	In this case getting the pointer to C<my_biggy> is a bit more
955	you need to use C<offsetof>:	1402	complicated: Either you store the address of your C<my_biggy> struct
		1403	in the C<data> member of the watcher (for woozies), or you need to use
		1404	some pointer arithmetic using C<offsetof> inside your watchers (for real
		1405	programmers):
956		1406
957	#include <stddef.h>	1407	#include <stddef.h>
958		1408
959	static void	1409	static void
960	t1_cb (EV_P_ struct ev_timer *w, int revents)	1410	t1_cb (EV_P_ ev_timer *w, int revents)
961	{	1411	{
962	struct my_biggy big = (struct my_biggy *	1412	struct my_biggy big = (struct my_biggy *)
963	(((char *)w) - offsetof (struct my_biggy, t1));	1413	(((char *)w) - offsetof (struct my_biggy, t1));
964	}	1414	}
965		1415
966	static void	1416	static void
967	t2_cb (EV_P_ struct ev_timer *w, int revents)	1417	t2_cb (EV_P_ ev_timer *w, int revents)
968	{	1418	{
969	struct my_biggy big = (struct my_biggy *	1419	struct my_biggy big = (struct my_biggy *)
970	(((char *)w) - offsetof (struct my_biggy, t2));	1420	(((char *)w) - offsetof (struct my_biggy, t2));
971	}	1421	}
		1422
		1423	=head2 WATCHER PRIORITY MODELS
		1424
		1425	Many event loops support I<watcher priorities>, which are usually small
		1426	integers that influence the ordering of event callback invocation
		1427	between watchers in some way, all else being equal.
		1428
		1429	In libev, Watcher priorities can be set using C<ev_set_priority>. See its
		1430	description for the more technical details such as the actual priority
		1431	range.
		1432
		1433	There are two common ways how these these priorities are being interpreted
		1434	by event loops:
		1435
		1436	In the more common lock-out model, higher priorities "lock out" invocation
		1437	of lower priority watchers, which means as long as higher priority
		1438	watchers receive events, lower priority watchers are not being invoked.
		1439
		1440	The less common only-for-ordering model uses priorities solely to order
		1441	callback invocation within a single event loop iteration: Higher priority
		1442	watchers are invoked before lower priority ones, but they all get invoked
		1443	before polling for new events.
		1444
		1445	Libev uses the second (only-for-ordering) model for all its watchers
		1446	except for idle watchers (which use the lock-out model).
		1447
		1448	The rationale behind this is that implementing the lock-out model for
		1449	watchers is not well supported by most kernel interfaces, and most event
		1450	libraries will just poll for the same events again and again as long as
		1451	their callbacks have not been executed, which is very inefficient in the
		1452	common case of one high-priority watcher locking out a mass of lower
		1453	priority ones.
		1454
		1455	Static (ordering) priorities are most useful when you have two or more
		1456	watchers handling the same resource: a typical usage example is having an
		1457	C<ev_io> watcher to receive data, and an associated C<ev_timer> to handle
		1458	timeouts. Under load, data might be received while the program handles
		1459	other jobs, but since timers normally get invoked first, the timeout
		1460	handler will be executed before checking for data. In that case, giving
		1461	the timer a lower priority than the I/O watcher ensures that I/O will be
		1462	handled first even under adverse conditions (which is usually, but not
		1463	always, what you want).
		1464
		1465	Since idle watchers use the "lock-out" model, meaning that idle watchers
		1466	will only be executed when no same or higher priority watchers have
		1467	received events, they can be used to implement the "lock-out" model when
		1468	required.
		1469
		1470	For example, to emulate how many other event libraries handle priorities,
		1471	you can associate an C<ev_idle> watcher to each such watcher, and in
		1472	the normal watcher callback, you just start the idle watcher. The real
		1473	processing is done in the idle watcher callback. This causes libev to
		1474	continuously poll and process kernel event data for the watcher, but when
		1475	the lock-out case is known to be rare (which in turn is rare :), this is
		1476	workable.
		1477
		1478	Usually, however, the lock-out model implemented that way will perform
		1479	miserably under the type of load it was designed to handle. In that case,
		1480	it might be preferable to stop the real watcher before starting the
		1481	idle watcher, so the kernel will not have to process the event in case
		1482	the actual processing will be delayed for considerable time.
		1483
		1484	Here is an example of an I/O watcher that should run at a strictly lower
		1485	priority than the default, and which should only process data when no
		1486	other events are pending:
		1487
		1488	ev_idle idle; // actual processing watcher
		1489	ev_io io; // actual event watcher
		1490
		1491	static void
		1492	io_cb (EV_P_ ev_io *w, int revents)
		1493	{
		1494	// stop the I/O watcher, we received the event, but
		1495	// are not yet ready to handle it.
		1496	ev_io_stop (EV_A_ w);
		1497
		1498	// start the idle watcher to handle the actual event.
		1499	// it will not be executed as long as other watchers
		1500	// with the default priority are receiving events.
		1501	ev_idle_start (EV_A_ &idle);
		1502	}
		1503
		1504	static void
		1505	idle_cb (EV_P_ ev_idle *w, int revents)
		1506	{
		1507	// actual processing
		1508	read (STDIN_FILENO, ...);
		1509
		1510	// have to start the I/O watcher again, as
		1511	// we have handled the event
		1512	ev_io_start (EV_P_ &io);
		1513	}
		1514
		1515	// initialisation
		1516	ev_idle_init (&idle, idle_cb);
		1517	ev_io_init (&io, io_cb, STDIN_FILENO, EV_READ);
		1518	ev_io_start (EV_DEFAULT_ &io);
		1519
		1520	In the "real" world, it might also be beneficial to start a timer, so that
		1521	low-priority connections can not be locked out forever under load. This
		1522	enables your program to keep a lower latency for important connections
		1523	during short periods of high load, while not completely locking out less
		1524	important ones.
972		1525
973		1526
974	=head1 WATCHER TYPES	1527	=head1 WATCHER TYPES
975		1528
976	This section describes each watcher in detail, but will not repeat	1529	This section describes each watcher in detail, but will not repeat
…		…
1000	In general you can register as many read and/or write event watchers per	1553	In general you can register as many read and/or write event watchers per
1001	fd as you want (as long as you don't confuse yourself). Setting all file	1554	fd as you want (as long as you don't confuse yourself). Setting all file
1002	descriptors to non-blocking mode is also usually a good idea (but not	1555	descriptors to non-blocking mode is also usually a good idea (but not
1003	required if you know what you are doing).	1556	required if you know what you are doing).
1004		1557
1005	If you must do this, then force the use of a known-to-be-good backend	1558	If you cannot use non-blocking mode, then force the use of a
1006	(at the time of this writing, this includes only C<EVBACKEND_SELECT> and	1559	known-to-be-good backend (at the time of this writing, this includes only
1007	C<EVBACKEND_POLL>).	1560	C<EVBACKEND_SELECT> and C<EVBACKEND_POLL>). The same applies to file
		1561	descriptors for which non-blocking operation makes no sense (such as
		1562	files) - libev doesn't guarantee any specific behaviour in that case.
1008		1563
1009	Another thing you have to watch out for is that it is quite easy to	1564	Another thing you have to watch out for is that it is quite easy to
1010	receive "spurious" readyness notifications, that is your callback might	1565	receive "spurious" readiness notifications, that is your callback might
1011	be called with C<EV_READ> but a subsequent C<read>(2) will actually block	1566	be called with C<EV_READ> but a subsequent C<read>(2) will actually block
1012	because there is no data. Not only are some backends known to create a	1567	because there is no data. Not only are some backends known to create a
1013	lot of those (for example solaris ports), it is very easy to get into	1568	lot of those (for example Solaris ports), it is very easy to get into
1014	this situation even with a relatively standard program structure. Thus	1569	this situation even with a relatively standard program structure. Thus
1015	it is best to always use non-blocking I/O: An extra C<read>(2) returning	1570	it is best to always use non-blocking I/O: An extra C<read>(2) returning
1016	C<EAGAIN> is far preferable to a program hanging until some data arrives.	1571	C<EAGAIN> is far preferable to a program hanging until some data arrives.
1017		1572
1018	If you cannot run the fd in non-blocking mode (for example you should not	1573	If you cannot run the fd in non-blocking mode (for example you should
1019	play around with an Xlib connection), then you have to seperately re-test	1574	not play around with an Xlib connection), then you have to separately
1020	whether a file descriptor is really ready with a known-to-be good interface	1575	re-test whether a file descriptor is really ready with a known-to-be good
1021	such as poll (fortunately in our Xlib example, Xlib already does this on	1576	interface such as poll (fortunately in our Xlib example, Xlib already
1022	its own, so its quite safe to use).	1577	does this on its own, so its quite safe to use). Some people additionally
		1578	use C<SIGALRM> and an interval timer, just to be sure you won't block
		1579	indefinitely.
		1580
		1581	But really, best use non-blocking mode.
1023		1582
1024	=head3 The special problem of disappearing file descriptors	1583	=head3 The special problem of disappearing file descriptors
1025		1584
1026	Some backends (e.g. kqueue, epoll) need to be told about closing a file	1585	Some backends (e.g. kqueue, epoll) need to be told about closing a file
1027	descriptor (either by calling C<close> explicitly or by any other means,	1586	descriptor (either due to calling C<close> explicitly or any other means,
1028	such as C<dup>). The reason is that you register interest in some file	1587	such as C<dup2>). The reason is that you register interest in some file
1029	descriptor, but when it goes away, the operating system will silently drop	1588	descriptor, but when it goes away, the operating system will silently drop
1030	this interest. If another file descriptor with the same number then is	1589	this interest. If another file descriptor with the same number then is
1031	registered with libev, there is no efficient way to see that this is, in	1590	registered with libev, there is no efficient way to see that this is, in
1032	fact, a different file descriptor.	1591	fact, a different file descriptor.
1033		1592
…		…
1062	To support fork in your programs, you either have to call	1621	To support fork in your programs, you either have to call
1063	C<ev_default_fork ()> or C<ev_loop_fork ()> after a fork in the child,	1622	C<ev_default_fork ()> or C<ev_loop_fork ()> after a fork in the child,
1064	enable C<EVFLAG_FORKCHECK>, or resort to C<EVBACKEND_SELECT> or	1623	enable C<EVFLAG_FORKCHECK>, or resort to C<EVBACKEND_SELECT> or
1065	C<EVBACKEND_POLL>.	1624	C<EVBACKEND_POLL>.
1066		1625
		1626	=head3 The special problem of SIGPIPE
		1627
		1628	While not really specific to libev, it is easy to forget about C<SIGPIPE>:
		1629	when writing to a pipe whose other end has been closed, your program gets
		1630	sent a SIGPIPE, which, by default, aborts your program. For most programs
		1631	this is sensible behaviour, for daemons, this is usually undesirable.
		1632
		1633	So when you encounter spurious, unexplained daemon exits, make sure you
		1634	ignore SIGPIPE (and maybe make sure you log the exit status of your daemon
		1635	somewhere, as that would have given you a big clue).
		1636
		1637	=head3 The special problem of accept()ing when you can't
		1638
		1639	Many implementations of the POSIX C<accept> function (for example,
		1640	found in post-2004 Linux) have the peculiar behaviour of not removing a
		1641	connection from the pending queue in all error cases.
		1642
		1643	For example, larger servers often run out of file descriptors (because
		1644	of resource limits), causing C<accept> to fail with C<ENFILE> but not
		1645	rejecting the connection, leading to libev signalling readiness on
		1646	the next iteration again (the connection still exists after all), and
		1647	typically causing the program to loop at 100% CPU usage.
		1648
		1649	Unfortunately, the set of errors that cause this issue differs between
		1650	operating systems, there is usually little the app can do to remedy the
		1651	situation, and no known thread-safe method of removing the connection to
		1652	cope with overload is known (to me).
		1653
		1654	One of the easiest ways to handle this situation is to just ignore it
		1655	- when the program encounters an overload, it will just loop until the
		1656	situation is over. While this is a form of busy waiting, no OS offers an
		1657	event-based way to handle this situation, so it's the best one can do.
		1658
		1659	A better way to handle the situation is to log any errors other than
		1660	C<EAGAIN> and C<EWOULDBLOCK>, making sure not to flood the log with such
		1661	messages, and continue as usual, which at least gives the user an idea of
		1662	what could be wrong ("raise the ulimit!"). For extra points one could stop
		1663	the C<ev_io> watcher on the listening fd "for a while", which reduces CPU
		1664	usage.
		1665
		1666	If your program is single-threaded, then you could also keep a dummy file
		1667	descriptor for overload situations (e.g. by opening F</dev/null>), and
		1668	when you run into C<ENFILE> or C<EMFILE>, close it, run C<accept>,
		1669	close that fd, and create a new dummy fd. This will gracefully refuse
		1670	clients under typical overload conditions.
		1671
		1672	The last way to handle it is to simply log the error and C<exit>, as
		1673	is often done with C<malloc> failures, but this results in an easy
		1674	opportunity for a DoS attack.
1067		1675
1068	=head3 Watcher-Specific Functions	1676	=head3 Watcher-Specific Functions
1069		1677
1070	=over 4	1678	=over 4
1071		1679
1072	=item ev_io_init (ev_io *, callback, int fd, int events)	1680	=item ev_io_init (ev_io *, callback, int fd, int events)
1073		1681
1074	=item ev_io_set (ev_io *, int fd, int events)	1682	=item ev_io_set (ev_io *, int fd, int events)
1075		1683
1076	Configures an C<ev_io> watcher. The C<fd> is the file descriptor to	1684	Configures an C<ev_io> watcher. The C<fd> is the file descriptor to
1077	rceeive events for and events is either C<EV_READ>, C<EV_WRITE> or	1685	receive events for and C<events> is either C<EV_READ>, C<EV_WRITE> or
1078	C<EV_READ | EV_WRITE> to receive the given events.	1686	C<EV_READ | EV_WRITE>, to express the desire to receive the given events.
1079		1687
1080	=item int fd [read-only]	1688	=item int fd [read-only]
1081		1689
1082	The file descriptor being watched.	1690	The file descriptor being watched.
1083		1691
…		…
1091		1699
1092	Example: Call C<stdin_readable_cb> when STDIN_FILENO has become, well	1700	Example: Call C<stdin_readable_cb> when STDIN_FILENO has become, well
1093	readable, but only once. Since it is likely line-buffered, you could	1701	readable, but only once. Since it is likely line-buffered, you could
1094	attempt to read a whole line in the callback.	1702	attempt to read a whole line in the callback.
1095		1703
1096	static void	1704	static void
1097	stdin_readable_cb (struct ev_loop *loop, struct ev_io *w, int revents)	1705	stdin_readable_cb (struct ev_loop *loop, ev_io *w, int revents)
1098	{	1706	{
1099	ev_io_stop (loop, w);	1707	ev_io_stop (loop, w);
1100	.. read from stdin here (or from w->fd) and haqndle any I/O errors	1708	.. read from stdin here (or from w->fd) and handle any I/O errors
1101	}	1709	}
1102		1710
1103	...	1711	...
1104	struct ev_loop *loop = ev_default_init (0);	1712	struct ev_loop *loop = ev_default_init (0);
1105	struct ev_io stdin_readable;	1713	ev_io stdin_readable;
1106	ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);	1714	ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
1107	ev_io_start (loop, &stdin_readable);	1715	ev_io_start (loop, &stdin_readable);
1108	ev_loop (loop, 0);	1716	ev_run (loop, 0);
1109		1717
1110		1718
1111	=head2 C<ev_timer> - relative and optionally repeating timeouts	1719	=head2 C<ev_timer> - relative and optionally repeating timeouts
1112		1720
1113	Timer watchers are simple relative timers that generate an event after a	1721	Timer watchers are simple relative timers that generate an event after a
1114	given time, and optionally repeating in regular intervals after that.	1722	given time, and optionally repeating in regular intervals after that.
1115		1723
1116	The timers are based on real time, that is, if you register an event that	1724	The timers are based on real time, that is, if you register an event that
1117	times out after an hour and you reset your system clock to last years	1725	times out after an hour and you reset your system clock to January last
1118	time, it will still time out after (roughly) and hour. "Roughly" because	1726	year, it will still time out after (roughly) one hour. "Roughly" because
1119	detecting time jumps is hard, and some inaccuracies are unavoidable (the	1727	detecting time jumps is hard, and some inaccuracies are unavoidable (the
1120	monotonic clock option helps a lot here).	1728	monotonic clock option helps a lot here).
		1729
		1730	The callback is guaranteed to be invoked only I<after> its timeout has
		1731	passed (not I<at>, so on systems with very low-resolution clocks this
		1732	might introduce a small delay). If multiple timers become ready during the
		1733	same loop iteration then the ones with earlier time-out values are invoked
		1734	before ones of the same priority with later time-out values (but this is
		1735	no longer true when a callback calls C<ev_run> recursively).
		1736
		1737	=head3 Be smart about timeouts
		1738
		1739	Many real-world problems involve some kind of timeout, usually for error
		1740	recovery. A typical example is an HTTP request - if the other side hangs,
		1741	you want to raise some error after a while.
		1742
		1743	What follows are some ways to handle this problem, from obvious and
		1744	inefficient to smart and efficient.
		1745
		1746	In the following, a 60 second activity timeout is assumed - a timeout that
		1747	gets reset to 60 seconds each time there is activity (e.g. each time some
		1748	data or other life sign was received).
		1749
		1750	=over 4
		1751
		1752	=item 1. Use a timer and stop, reinitialise and start it on activity.
		1753
		1754	This is the most obvious, but not the most simple way: In the beginning,
		1755	start the watcher:
		1756
		1757	ev_timer_init (timer, callback, 60., 0.);
		1758	ev_timer_start (loop, timer);
		1759
		1760	Then, each time there is some activity, C<ev_timer_stop> it, initialise it
		1761	and start it again:
		1762
		1763	ev_timer_stop (loop, timer);
		1764	ev_timer_set (timer, 60., 0.);
		1765	ev_timer_start (loop, timer);
		1766
		1767	This is relatively simple to implement, but means that each time there is
		1768	some activity, libev will first have to remove the timer from its internal
		1769	data structure and then add it again. Libev tries to be fast, but it's
		1770	still not a constant-time operation.
		1771
		1772	=item 2. Use a timer and re-start it with C<ev_timer_again> inactivity.
		1773
		1774	This is the easiest way, and involves using C<ev_timer_again> instead of
		1775	C<ev_timer_start>.
		1776
		1777	To implement this, configure an C<ev_timer> with a C<repeat> value
		1778	of C<60> and then call C<ev_timer_again> at start and each time you
		1779	successfully read or write some data. If you go into an idle state where
		1780	you do not expect data to travel on the socket, you can C<ev_timer_stop>
		1781	the timer, and C<ev_timer_again> will automatically restart it if need be.
		1782
		1783	That means you can ignore both the C<ev_timer_start> function and the
		1784	C<after> argument to C<ev_timer_set>, and only ever use the C<repeat>
		1785	member and C<ev_timer_again>.
		1786
		1787	At start:
		1788
		1789	ev_init (timer, callback);
		1790	timer->repeat = 60.;
		1791	ev_timer_again (loop, timer);
		1792
		1793	Each time there is some activity:
		1794
		1795	ev_timer_again (loop, timer);
		1796
		1797	It is even possible to change the time-out on the fly, regardless of
		1798	whether the watcher is active or not:
		1799
		1800	timer->repeat = 30.;
		1801	ev_timer_again (loop, timer);
		1802
		1803	This is slightly more efficient then stopping/starting the timer each time
		1804	you want to modify its timeout value, as libev does not have to completely
		1805	remove and re-insert the timer from/into its internal data structure.
		1806
		1807	It is, however, even simpler than the "obvious" way to do it.
		1808
		1809	=item 3. Let the timer time out, but then re-arm it as required.
		1810
		1811	This method is more tricky, but usually most efficient: Most timeouts are
		1812	relatively long compared to the intervals between other activity - in
		1813	our example, within 60 seconds, there are usually many I/O events with
		1814	associated activity resets.
		1815
		1816	In this case, it would be more efficient to leave the C<ev_timer> alone,
		1817	but remember the time of last activity, and check for a real timeout only
		1818	within the callback:
		1819
		1820	ev_tstamp last_activity; // time of last activity
		1821
		1822	static void
		1823	callback (EV_P_ ev_timer *w, int revents)
		1824	{
		1825	ev_tstamp now = ev_now (EV_A);
		1826	ev_tstamp timeout = last_activity + 60.;
		1827
		1828	// if last_activity + 60. is older than now, we did time out
		1829	if (timeout < now)
		1830	{
		1831	// timeout occurred, take action
		1832	}
		1833	else
		1834	{
		1835	// callback was invoked, but there was some activity, re-arm
		1836	// the watcher to fire in last_activity + 60, which is
		1837	// guaranteed to be in the future, so "again" is positive:
		1838	w->repeat = timeout - now;
		1839	ev_timer_again (EV_A_ w);
		1840	}
		1841	}
		1842
		1843	To summarise the callback: first calculate the real timeout (defined
		1844	as "60 seconds after the last activity"), then check if that time has
		1845	been reached, which means something I<did>, in fact, time out. Otherwise
		1846	the callback was invoked too early (C<timeout> is in the future), so
		1847	re-schedule the timer to fire at that future time, to see if maybe we have
		1848	a timeout then.
		1849
		1850	Note how C<ev_timer_again> is used, taking advantage of the
		1851	C<ev_timer_again> optimisation when the timer is already running.
		1852
		1853	This scheme causes more callback invocations (about one every 60 seconds
		1854	minus half the average time between activity), but virtually no calls to
		1855	libev to change the timeout.
		1856
		1857	To start the timer, simply initialise the watcher and set C<last_activity>
		1858	to the current time (meaning we just have some activity :), then call the
		1859	callback, which will "do the right thing" and start the timer:
		1860
		1861	ev_init (timer, callback);
		1862	last_activity = ev_now (loop);
		1863	callback (loop, timer, EV_TIMER);
		1864
		1865	And when there is some activity, simply store the current time in
		1866	C<last_activity>, no libev calls at all:
		1867
		1868	last_activity = ev_now (loop);
		1869
		1870	This technique is slightly more complex, but in most cases where the
		1871	time-out is unlikely to be triggered, much more efficient.
		1872
		1873	Changing the timeout is trivial as well (if it isn't hard-coded in the
		1874	callback :) - just change the timeout and invoke the callback, which will
		1875	fix things for you.
		1876
		1877	=item 4. Wee, just use a double-linked list for your timeouts.
		1878
		1879	If there is not one request, but many thousands (millions...), all
		1880	employing some kind of timeout with the same timeout value, then one can
		1881	do even better:
		1882
		1883	When starting the timeout, calculate the timeout value and put the timeout
		1884	at the I<end> of the list.
		1885
		1886	Then use an C<ev_timer> to fire when the timeout at the I<beginning> of
		1887	the list is expected to fire (for example, using the technique #3).
		1888
		1889	When there is some activity, remove the timer from the list, recalculate
		1890	the timeout, append it to the end of the list again, and make sure to
		1891	update the C<ev_timer> if it was taken from the beginning of the list.
		1892
		1893	This way, one can manage an unlimited number of timeouts in O(1) time for
		1894	starting, stopping and updating the timers, at the expense of a major
		1895	complication, and having to use a constant timeout. The constant timeout
		1896	ensures that the list stays sorted.
		1897
		1898	=back
		1899
		1900	So which method the best?
		1901
		1902	Method #2 is a simple no-brain-required solution that is adequate in most
		1903	situations. Method #3 requires a bit more thinking, but handles many cases
		1904	better, and isn't very complicated either. In most case, choosing either
		1905	one is fine, with #3 being better in typical situations.
		1906
		1907	Method #1 is almost always a bad idea, and buys you nothing. Method #4 is
		1908	rather complicated, but extremely efficient, something that really pays
		1909	off after the first million or so of active timers, i.e. it's usually
		1910	overkill :)
		1911
		1912	=head3 The special problem of time updates
		1913
		1914	Establishing the current time is a costly operation (it usually takes at
		1915	least two system calls): EV therefore updates its idea of the current
		1916	time only before and after C<ev_run> collects new events, which causes a
		1917	growing difference between C<ev_now ()> and C<ev_time ()> when handling
		1918	lots of events in one iteration.
1121		1919
1122	The relative timeouts are calculated relative to the C<ev_now ()>	1920	The relative timeouts are calculated relative to the C<ev_now ()>
1123	time. This is usually the right thing as this timestamp refers to the time	1921	time. This is usually the right thing as this timestamp refers to the time
1124	of the event triggering whatever timeout you are modifying/starting. If	1922	of the event triggering whatever timeout you are modifying/starting. If
1125	you suspect event processing to be delayed and you I<need> to base the timeout	1923	you suspect event processing to be delayed and you I<need> to base the
1126	on the current time, use something like this to adjust for this:	1924	timeout on the current time, use something like this to adjust for this:
1127		1925
1128	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);	1926	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);
1129		1927
1130	The callback is guarenteed to be invoked only when its timeout has passed,	1928	If the event loop is suspended for a long time, you can also force an
1131	but if multiple timers become ready during the same loop iteration then	1929	update of the time returned by C<ev_now ()> by calling C<ev_now_update
1132	order of execution is undefined.	1930	()>.
		1931
		1932	=head3 The special problems of suspended animation
		1933
		1934	When you leave the server world it is quite customary to hit machines that
		1935	can suspend/hibernate - what happens to the clocks during such a suspend?
		1936
		1937	Some quick tests made with a Linux 2.6.28 indicate that a suspend freezes
		1938	all processes, while the clocks (C<times>, C<CLOCK_MONOTONIC>) continue
		1939	to run until the system is suspended, but they will not advance while the
		1940	system is suspended. That means, on resume, it will be as if the program
		1941	was frozen for a few seconds, but the suspend time will not be counted
		1942	towards C<ev_timer> when a monotonic clock source is used. The real time
		1943	clock advanced as expected, but if it is used as sole clocksource, then a
		1944	long suspend would be detected as a time jump by libev, and timers would
		1945	be adjusted accordingly.
		1946
		1947	I would not be surprised to see different behaviour in different between
		1948	operating systems, OS versions or even different hardware.
		1949
		1950	The other form of suspend (job control, or sending a SIGSTOP) will see a
		1951	time jump in the monotonic clocks and the realtime clock. If the program
		1952	is suspended for a very long time, and monotonic clock sources are in use,
		1953	then you can expect C<ev_timer>s to expire as the full suspension time
		1954	will be counted towards the timers. When no monotonic clock source is in
		1955	use, then libev will again assume a timejump and adjust accordingly.
		1956
		1957	It might be beneficial for this latter case to call C<ev_suspend>
		1958	and C<ev_resume> in code that handles C<SIGTSTP>, to at least get
		1959	deterministic behaviour in this case (you can do nothing against
		1960	C<SIGSTOP>).
1133		1961
1134	=head3 Watcher-Specific Functions and Data Members	1962	=head3 Watcher-Specific Functions and Data Members
1135		1963
1136	=over 4	1964	=over 4
1137		1965
1138	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)	1966	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)
1139		1967
1140	=item ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)	1968	=item ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)
1141		1969
1142	Configure the timer to trigger after C<after> seconds. If C<repeat> is	1970	Configure the timer to trigger after C<after> seconds. If C<repeat>
1143	C<0.>, then it will automatically be stopped. If it is positive, then the	1971	is C<0.>, then it will automatically be stopped once the timeout is
1144	timer will automatically be configured to trigger again C<repeat> seconds	1972	reached. If it is positive, then the timer will automatically be
1145	later, again, and again, until stopped manually.	1973	configured to trigger again C<repeat> seconds later, again, and again,
		1974	until stopped manually.
1146		1975
1147	The timer itself will do a best-effort at avoiding drift, that is, if you	1976	The timer itself will do a best-effort at avoiding drift, that is, if
1148	configure a timer to trigger every 10 seconds, then it will trigger at	1977	you configure a timer to trigger every 10 seconds, then it will normally
1149	exactly 10 second intervals. If, however, your program cannot keep up with	1978	trigger at exactly 10 second intervals. If, however, your program cannot
1150	the timer (because it takes longer than those 10 seconds to do stuff) the	1979	keep up with the timer (because it takes longer than those 10 seconds to
1151	timer will not fire more than once per event loop iteration.	1980	do stuff) the timer will not fire more than once per event loop iteration.
1152		1981
1153	=item ev_timer_again (loop)	1982	=item ev_timer_again (loop, ev_timer *)
1154		1983
1155	This will act as if the timer timed out and restart it again if it is	1984	This will act as if the timer timed out and restart it again if it is
1156	repeating. The exact semantics are:	1985	repeating. The exact semantics are:
1157		1986
1158	If the timer is pending, its pending status is cleared.	1987	If the timer is pending, its pending status is cleared.
1159		1988
1160	If the timer is started but nonrepeating, stop it (as if it timed out).	1989	If the timer is started but non-repeating, stop it (as if it timed out).
1161		1990
1162	If the timer is repeating, either start it if necessary (with the	1991	If the timer is repeating, either start it if necessary (with the
1163	C<repeat> value), or reset the running timer to the C<repeat> value.	1992	C<repeat> value), or reset the running timer to the C<repeat> value.
1164		1993
1165	This sounds a bit complicated, but here is a useful and typical	1994	This sounds a bit complicated, see L<Be smart about timeouts>, above, for a
1166	example: Imagine you have a tcp connection and you want a so-called idle	1995	usage example.
1167	timeout, that is, you want to be called when there have been, say, 60
1168	seconds of inactivity on the socket. The easiest way to do this is to
1169	configure an C<ev_timer> with a C<repeat> value of C<60> and then call
1170	C<ev_timer_again> each time you successfully read or write some data. If
1171	you go into an idle state where you do not expect data to travel on the
1172	socket, you can C<ev_timer_stop> the timer, and C<ev_timer_again> will
1173	automatically restart it if need be.
1174		1996
1175	That means you can ignore the C<after> value and C<ev_timer_start>	1997	=item ev_tstamp ev_timer_remaining (loop, ev_timer *)
1176	altogether and only ever use the C<repeat> value and C<ev_timer_again>:
1177		1998
1178	ev_timer_init (timer, callback, 0., 5.);	1999	Returns the remaining time until a timer fires. If the timer is active,
1179	ev_timer_again (loop, timer);	2000	then this time is relative to the current event loop time, otherwise it's
1180	...	2001	the timeout value currently configured.
1181	timer->again = 17.;
1182	ev_timer_again (loop, timer);
1183	...
1184	timer->again = 10.;
1185	ev_timer_again (loop, timer);
1186		2002
1187	This is more slightly efficient then stopping/starting the timer each time	2003	That is, after an C<ev_timer_set (w, 5, 7)>, C<ev_timer_remaining> returns
1188	you want to modify its timeout value.	2004	C<5>. When the timer is started and one second passes, C<ev_timer_remaining>
		2005	will return C<4>. When the timer expires and is restarted, it will return
		2006	roughly C<7> (likely slightly less as callback invocation takes some time,
		2007	too), and so on.
1189		2008
1190	=item ev_tstamp repeat [read-write]	2009	=item ev_tstamp repeat [read-write]
1191		2010
1192	The current C<repeat> value. Will be used each time the watcher times out	2011	The current C<repeat> value. Will be used each time the watcher times out
1193	or C<ev_timer_again> is called and determines the next timeout (if any),	2012	or C<ev_timer_again> is called, and determines the next timeout (if any),
1194	which is also when any modifications are taken into account.	2013	which is also when any modifications are taken into account.
1195		2014
1196	=back	2015	=back
1197		2016
1198	=head3 Examples	2017	=head3 Examples
1199		2018
1200	Example: Create a timer that fires after 60 seconds.	2019	Example: Create a timer that fires after 60 seconds.
1201		2020
1202	static void	2021	static void
1203	one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)	2022	one_minute_cb (struct ev_loop *loop, ev_timer *w, int revents)
1204	{	2023	{
1205	.. one minute over, w is actually stopped right here	2024	.. one minute over, w is actually stopped right here
1206	}	2025	}
1207		2026
1208	struct ev_timer mytimer;	2027	ev_timer mytimer;
1209	ev_timer_init (&mytimer, one_minute_cb, 60., 0.);	2028	ev_timer_init (&mytimer, one_minute_cb, 60., 0.);
1210	ev_timer_start (loop, &mytimer);	2029	ev_timer_start (loop, &mytimer);
1211		2030
1212	Example: Create a timeout timer that times out after 10 seconds of	2031	Example: Create a timeout timer that times out after 10 seconds of
1213	inactivity.	2032	inactivity.
1214		2033
1215	static void	2034	static void
1216	timeout_cb (struct ev_loop *loop, struct ev_timer *w, int revents)	2035	timeout_cb (struct ev_loop *loop, ev_timer *w, int revents)
1217	{	2036	{
1218	.. ten seconds without any activity	2037	.. ten seconds without any activity
1219	}	2038	}
1220		2039
1221	struct ev_timer mytimer;	2040	ev_timer mytimer;
1222	ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */	2041	ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */
1223	ev_timer_again (&mytimer); /* start timer */	2042	ev_timer_again (&mytimer); /* start timer */
1224	ev_loop (loop, 0);	2043	ev_run (loop, 0);
1225		2044
1226	// and in some piece of code that gets executed on any "activity":	2045	// and in some piece of code that gets executed on any "activity":
1227	// reset the timeout to start ticking again at 10 seconds	2046	// reset the timeout to start ticking again at 10 seconds
1228	ev_timer_again (&mytimer);	2047	ev_timer_again (&mytimer);
1229		2048
1230		2049
1231	=head2 C<ev_periodic> - to cron or not to cron?	2050	=head2 C<ev_periodic> - to cron or not to cron?
1232		2051
1233	Periodic watchers are also timers of a kind, but they are very versatile	2052	Periodic watchers are also timers of a kind, but they are very versatile
1234	(and unfortunately a bit complex).	2053	(and unfortunately a bit complex).
1235		2054
1236	Unlike C<ev_timer>'s, they are not based on real time (or relative time)	2055	Unlike C<ev_timer>, periodic watchers are not based on real time (or
1237	but on wallclock time (absolute time). You can tell a periodic watcher	2056	relative time, the physical time that passes) but on wall clock time
1238	to trigger "at" some specific point in time. For example, if you tell a	2057	(absolute time, the thing you can read on your calender or clock). The
1239	periodic watcher to trigger in 10 seconds (by specifiying e.g. C<ev_now ()	2058	difference is that wall clock time can run faster or slower than real
1240	+ 10.>) and then reset your system clock to the last year, then it will	2059	time, and time jumps are not uncommon (e.g. when you adjust your
1241	take a year to trigger the event (unlike an C<ev_timer>, which would trigger	2060	wrist-watch).
1242	roughly 10 seconds later).
1243		2061
1244	They can also be used to implement vastly more complex timers, such as	2062	You can tell a periodic watcher to trigger after some specific point
1245	triggering an event on each midnight, local time or other, complicated,	2063	in time: for example, if you tell a periodic watcher to trigger "in 10
1246	rules.	2064	seconds" (by specifying e.g. C<ev_now () + 10.>, that is, an absolute time
		2065	not a delay) and then reset your system clock to January of the previous
		2066	year, then it will take a year or more to trigger the event (unlike an
		2067	C<ev_timer>, which would still trigger roughly 10 seconds after starting
		2068	it, as it uses a relative timeout).
1247		2069
		2070	C<ev_periodic> watchers can also be used to implement vastly more complex
		2071	timers, such as triggering an event on each "midnight, local time", or
		2072	other complicated rules. This cannot be done with C<ev_timer> watchers, as
		2073	those cannot react to time jumps.
		2074
1248	As with timers, the callback is guarenteed to be invoked only when the	2075	As with timers, the callback is guaranteed to be invoked only when the
1249	time (C<at>) has been passed, but if multiple periodic timers become ready	2076	point in time where it is supposed to trigger has passed. If multiple
1250	during the same loop iteration then order of execution is undefined.	2077	timers become ready during the same loop iteration then the ones with
		2078	earlier time-out values are invoked before ones with later time-out values
		2079	(but this is no longer true when a callback calls C<ev_run> recursively).
1251		2080
1252	=head3 Watcher-Specific Functions and Data Members	2081	=head3 Watcher-Specific Functions and Data Members
1253		2082
1254	=over 4	2083	=over 4
1255		2084
1256	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)	2085	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp offset, ev_tstamp interval, reschedule_cb)
1257		2086
1258	=item ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)	2087	=item ev_periodic_set (ev_periodic *, ev_tstamp offset, ev_tstamp interval, reschedule_cb)
1259		2088
1260	Lots of arguments, lets sort it out... There are basically three modes of	2089	Lots of arguments, let's sort it out... There are basically three modes of
1261	operation, and we will explain them from simplest to complex:	2090	operation, and we will explain them from simplest to most complex:
1262		2091
1263	=over 4	2092	=over 4
1264		2093
1265	=item * absolute timer (at = time, interval = reschedule_cb = 0)	2094	=item * absolute timer (offset = absolute time, interval = 0, reschedule_cb = 0)
1266		2095
1267	In this configuration the watcher triggers an event at the wallclock time	2096	In this configuration the watcher triggers an event after the wall clock
1268	C<at> and doesn't repeat. It will not adjust when a time jump occurs,	2097	time C<offset> has passed. It will not repeat and will not adjust when a
1269	that is, if it is to be run at January 1st 2011 then it will run when the	2098	time jump occurs, that is, if it is to be run at January 1st 2011 then it
1270	system time reaches or surpasses this time.	2099	will be stopped and invoked when the system clock reaches or surpasses
		2100	this point in time.
1271		2101
1272	=item * non-repeating interval timer (at = offset, interval > 0, reschedule_cb = 0)	2102	=item * repeating interval timer (offset = offset within interval, interval > 0, reschedule_cb = 0)
1273		2103
1274	In this mode the watcher will always be scheduled to time out at the next	2104	In this mode the watcher will always be scheduled to time out at the next
1275	C<at + N * interval> time (for some integer N, which can also be negative)	2105	C<offset + N * interval> time (for some integer N, which can also be
1276	and then repeat, regardless of any time jumps.	2106	negative) and then repeat, regardless of any time jumps. The C<offset>
		2107	argument is merely an offset into the C<interval> periods.
1277		2108
1278	This can be used to create timers that do not drift with respect to system	2109	This can be used to create timers that do not drift with respect to the
1279	time:	2110	system clock, for example, here is an C<ev_periodic> that triggers each
		2111	hour, on the hour (with respect to UTC):
1280		2112
1281	ev_periodic_set (&periodic, 0., 3600., 0);	2113	ev_periodic_set (&periodic, 0., 3600., 0);
1282		2114
1283	This doesn't mean there will always be 3600 seconds in between triggers,	2115	This doesn't mean there will always be 3600 seconds in between triggers,
1284	but only that the the callback will be called when the system time shows a	2116	but only that the callback will be called when the system time shows a
1285	full hour (UTC), or more correctly, when the system time is evenly divisible	2117	full hour (UTC), or more correctly, when the system time is evenly divisible
1286	by 3600.	2118	by 3600.
1287		2119
1288	Another way to think about it (for the mathematically inclined) is that	2120	Another way to think about it (for the mathematically inclined) is that
1289	C<ev_periodic> will try to run the callback in this mode at the next possible	2121	C<ev_periodic> will try to run the callback in this mode at the next possible
1290	time where C<time = at (mod interval)>, regardless of any time jumps.	2122	time where C<time = offset (mod interval)>, regardless of any time jumps.
1291		2123
1292	For numerical stability it is preferable that the C<at> value is near	2124	For numerical stability it is preferable that the C<offset> value is near
1293	C<ev_now ()> (the current time), but there is no range requirement for	2125	C<ev_now ()> (the current time), but there is no range requirement for
1294	this value.	2126	this value, and in fact is often specified as zero.
1295		2127
		2128	Note also that there is an upper limit to how often a timer can fire (CPU
		2129	speed for example), so if C<interval> is very small then timing stability
		2130	will of course deteriorate. Libev itself tries to be exact to be about one
		2131	millisecond (if the OS supports it and the machine is fast enough).
		2132
1296	=item * manual reschedule mode (at and interval ignored, reschedule_cb = callback)	2133	=item * manual reschedule mode (offset ignored, interval ignored, reschedule_cb = callback)
1297		2134
1298	In this mode the values for C<interval> and C<at> are both being	2135	In this mode the values for C<interval> and C<offset> are both being
1299	ignored. Instead, each time the periodic watcher gets scheduled, the	2136	ignored. Instead, each time the periodic watcher gets scheduled, the
1300	reschedule callback will be called with the watcher as first, and the	2137	reschedule callback will be called with the watcher as first, and the
1301	current time as second argument.	2138	current time as second argument.
1302		2139
1303	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher,	2140	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher, ever,
1304	ever, or make any event loop modifications>. If you need to stop it,	2141	or make ANY other event loop modifications whatsoever, unless explicitly
1305	return C<now + 1e30> (or so, fudge fudge) and stop it afterwards (e.g. by	2142	allowed by documentation here>.
1306	starting an C<ev_prepare> watcher, which is legal).
1307		2143
		2144	If you need to stop it, return C<now + 1e30> (or so, fudge fudge) and stop
		2145	it afterwards (e.g. by starting an C<ev_prepare> watcher, which is the
		2146	only event loop modification you are allowed to do).
		2147
1308	Its prototype is C<ev_tstamp (*reschedule_cb)(struct ev_periodic *w,	2148	The callback prototype is C<ev_tstamp (*reschedule_cb)(ev_periodic
1309	ev_tstamp now)>, e.g.:	2149	*w, ev_tstamp now)>, e.g.:
1310		2150
		2151	static ev_tstamp
1311	static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)	2152	my_rescheduler (ev_periodic *w, ev_tstamp now)
1312	{	2153	{
1313	return now + 60.;	2154	return now + 60.;
1314	}	2155	}
1315		2156
1316	It must return the next time to trigger, based on the passed time value	2157	It must return the next time to trigger, based on the passed time value
1317	(that is, the lowest time value larger than to the second argument). It	2158	(that is, the lowest time value larger than to the second argument). It
1318	will usually be called just before the callback will be triggered, but	2159	will usually be called just before the callback will be triggered, but
1319	might be called at other times, too.	2160	might be called at other times, too.
1320		2161
1321	NOTE: I<< This callback must always return a time that is later than the	2162	NOTE: I<< This callback must always return a time that is higher than or
1322	passed C<now> value >>. Not even C<now> itself will do, it I<must> be larger.	2163	equal to the passed C<now> value >>.
1323		2164
1324	This can be used to create very complex timers, such as a timer that	2165	This can be used to create very complex timers, such as a timer that
1325	triggers on each midnight, local time. To do this, you would calculate the	2166	triggers on "next midnight, local time". To do this, you would calculate the
1326	next midnight after C<now> and return the timestamp value for this. How	2167	next midnight after C<now> and return the timestamp value for this. How
1327	you do this is, again, up to you (but it is not trivial, which is the main	2168	you do this is, again, up to you (but it is not trivial, which is the main
1328	reason I omitted it as an example).	2169	reason I omitted it as an example).
1329		2170
1330	=back	2171	=back
…		…
1334	Simply stops and restarts the periodic watcher again. This is only useful	2175	Simply stops and restarts the periodic watcher again. This is only useful
1335	when you changed some parameters or the reschedule callback would return	2176	when you changed some parameters or the reschedule callback would return
1336	a different time than the last time it was called (e.g. in a crond like	2177	a different time than the last time it was called (e.g. in a crond like
1337	program when the crontabs have changed).	2178	program when the crontabs have changed).
1338		2179
		2180	=item ev_tstamp ev_periodic_at (ev_periodic *)
		2181
		2182	When active, returns the absolute time that the watcher is supposed
		2183	to trigger next. This is not the same as the C<offset> argument to
		2184	C<ev_periodic_set>, but indeed works even in interval and manual
		2185	rescheduling modes.
		2186
1339	=item ev_tstamp offset [read-write]	2187	=item ev_tstamp offset [read-write]
1340		2188
1341	When repeating, this contains the offset value, otherwise this is the	2189	When repeating, this contains the offset value, otherwise this is the
1342	absolute point in time (the C<at> value passed to C<ev_periodic_set>).	2190	absolute point in time (the C<offset> value passed to C<ev_periodic_set>,
		2191	although libev might modify this value for better numerical stability).
1343		2192
1344	Can be modified any time, but changes only take effect when the periodic	2193	Can be modified any time, but changes only take effect when the periodic
1345	timer fires or C<ev_periodic_again> is being called.	2194	timer fires or C<ev_periodic_again> is being called.
1346		2195
1347	=item ev_tstamp interval [read-write]	2196	=item ev_tstamp interval [read-write]
1348		2197
1349	The current interval value. Can be modified any time, but changes only	2198	The current interval value. Can be modified any time, but changes only
1350	take effect when the periodic timer fires or C<ev_periodic_again> is being	2199	take effect when the periodic timer fires or C<ev_periodic_again> is being
1351	called.	2200	called.
1352		2201
1353	=item ev_tstamp (*reschedule_cb)(struct ev_periodic *w, ev_tstamp now) [read-write]	2202	=item ev_tstamp (*reschedule_cb)(ev_periodic *w, ev_tstamp now) [read-write]
1354		2203
1355	The current reschedule callback, or C<0>, if this functionality is	2204	The current reschedule callback, or C<0>, if this functionality is
1356	switched off. Can be changed any time, but changes only take effect when	2205	switched off. Can be changed any time, but changes only take effect when
1357	the periodic timer fires or C<ev_periodic_again> is being called.	2206	the periodic timer fires or C<ev_periodic_again> is being called.
1358		2207
1359	=item ev_tstamp at [read-only]
1360
1361	When active, contains the absolute time that the watcher is supposed to
1362	trigger next.
1363
1364	=back	2208	=back
1365		2209
1366	=head3 Examples	2210	=head3 Examples
1367		2211
1368	Example: Call a callback every hour, or, more precisely, whenever the	2212	Example: Call a callback every hour, or, more precisely, whenever the
1369	system clock is divisible by 3600. The callback invocation times have	2213	system time is divisible by 3600. The callback invocation times have
1370	potentially a lot of jittering, but good long-term stability.	2214	potentially a lot of jitter, but good long-term stability.
1371		2215
1372	static void	2216	static void
1373	clock_cb (struct ev_loop *loop, struct ev_io *w, int revents)	2217	clock_cb (struct ev_loop *loop, ev_periodic *w, int revents)
1374	{	2218	{
1375	... its now a full hour (UTC, or TAI or whatever your clock follows)	2219	... its now a full hour (UTC, or TAI or whatever your clock follows)
1376	}	2220	}
1377		2221
1378	struct ev_periodic hourly_tick;	2222	ev_periodic hourly_tick;
1379	ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);	2223	ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);
1380	ev_periodic_start (loop, &hourly_tick);	2224	ev_periodic_start (loop, &hourly_tick);
1381		2225
1382	Example: The same as above, but use a reschedule callback to do it:	2226	Example: The same as above, but use a reschedule callback to do it:
1383		2227
1384	#include <math.h>	2228	#include <math.h>
1385		2229
1386	static ev_tstamp	2230	static ev_tstamp
1387	my_scheduler_cb (struct ev_periodic *w, ev_tstamp now)	2231	my_scheduler_cb (ev_periodic *w, ev_tstamp now)
1388	{	2232	{
1389	return fmod (now, 3600.) + 3600.;	2233	return now + (3600. - fmod (now, 3600.));
1390	}	2234	}
1391		2235
1392	ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);	2236	ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
1393		2237
1394	Example: Call a callback every hour, starting now:	2238	Example: Call a callback every hour, starting now:
1395		2239
1396	struct ev_periodic hourly_tick;	2240	ev_periodic hourly_tick;
1397	ev_periodic_init (&hourly_tick, clock_cb,	2241	ev_periodic_init (&hourly_tick, clock_cb,
1398	fmod (ev_now (loop), 3600.), 3600., 0);	2242	fmod (ev_now (loop), 3600.), 3600., 0);
1399	ev_periodic_start (loop, &hourly_tick);	2243	ev_periodic_start (loop, &hourly_tick);
1400		2244
1401		2245
1402	=head2 C<ev_signal> - signal me when a signal gets signalled!	2246	=head2 C<ev_signal> - signal me when a signal gets signalled!
1403		2247
1404	Signal watchers will trigger an event when the process receives a specific	2248	Signal watchers will trigger an event when the process receives a specific
1405	signal one or more times. Even though signals are very asynchronous, libev	2249	signal one or more times. Even though signals are very asynchronous, libev
1406	will try it's best to deliver signals synchronously, i.e. as part of the	2250	will try it's best to deliver signals synchronously, i.e. as part of the
1407	normal event processing, like any other event.	2251	normal event processing, like any other event.
1408		2252
		2253	If you want signals to be delivered truly asynchronously, just use
		2254	C<sigaction> as you would do without libev and forget about sharing
		2255	the signal. You can even use C<ev_async> from a signal handler to
		2256	synchronously wake up an event loop.
		2257
1409	You can configure as many watchers as you like per signal. Only when the	2258	You can configure as many watchers as you like for the same signal, but
		2259	only within the same loop, i.e. you can watch for C<SIGINT> in your
		2260	default loop and for C<SIGIO> in another loop, but you cannot watch for
		2261	C<SIGINT> in both the default loop and another loop at the same time. At
		2262	the moment, C<SIGCHLD> is permanently tied to the default loop.
		2263
1410	first watcher gets started will libev actually register a signal watcher	2264	When the first watcher gets started will libev actually register something
1411	with the kernel (thus it coexists with your own signal handlers as long	2265	with the kernel (thus it coexists with your own signal handlers as long as
1412	as you don't register any with libev). Similarly, when the last signal	2266	you don't register any with libev for the same signal).
1413	watcher for a signal is stopped libev will reset the signal handler to	2267
1414	SIG_DFL (regardless of what it was set to before).	2268	If possible and supported, libev will install its handlers with
		2269	C<SA_RESTART> (or equivalent) behaviour enabled, so system calls should
		2270	not be unduly interrupted. If you have a problem with system calls getting
		2271	interrupted by signals you can block all signals in an C<ev_check> watcher
		2272	and unblock them in an C<ev_prepare> watcher.
		2273
		2274	=head3 The special problem of inheritance over fork/execve/pthread_create
		2275
		2276	Both the signal mask (C<sigprocmask>) and the signal disposition
		2277	(C<sigaction>) are unspecified after starting a signal watcher (and after
		2278	stopping it again), that is, libev might or might not block the signal,
		2279	and might or might not set or restore the installed signal handler.
		2280
		2281	While this does not matter for the signal disposition (libev never
		2282	sets signals to C<SIG_IGN>, so handlers will be reset to C<SIG_DFL> on
		2283	C<execve>), this matters for the signal mask: many programs do not expect
		2284	certain signals to be blocked.
		2285
		2286	This means that before calling C<exec> (from the child) you should reset
		2287	the signal mask to whatever "default" you expect (all clear is a good
		2288	choice usually).
		2289
		2290	The simplest way to ensure that the signal mask is reset in the child is
		2291	to install a fork handler with C<pthread_atfork> that resets it. That will
		2292	catch fork calls done by libraries (such as the libc) as well.
		2293
		2294	In current versions of libev, the signal will not be blocked indefinitely
		2295	unless you use the C<signalfd> API (C<EV_SIGNALFD>). While this reduces
		2296	the window of opportunity for problems, it will not go away, as libev
		2297	I<has> to modify the signal mask, at least temporarily.
		2298
		2299	So I can't stress this enough: I<If you do not reset your signal mask when
		2300	you expect it to be empty, you have a race condition in your code>. This
		2301	is not a libev-specific thing, this is true for most event libraries.
1415		2302
1416	=head3 Watcher-Specific Functions and Data Members	2303	=head3 Watcher-Specific Functions and Data Members
1417		2304
1418	=over 4	2305	=over 4
1419		2306
…		…
1428		2315
1429	The signal the watcher watches out for.	2316	The signal the watcher watches out for.
1430		2317
1431	=back	2318	=back
1432		2319
		2320	=head3 Examples
		2321
		2322	Example: Try to exit cleanly on SIGINT.
		2323
		2324	static void
		2325	sigint_cb (struct ev_loop *loop, ev_signal *w, int revents)
		2326	{
		2327	ev_break (loop, EVBREAK_ALL);
		2328	}
		2329
		2330	ev_signal signal_watcher;
		2331	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
		2332	ev_signal_start (loop, &signal_watcher);
		2333
1433		2334
1434	=head2 C<ev_child> - watch out for process status changes	2335	=head2 C<ev_child> - watch out for process status changes
1435		2336
1436	Child watchers trigger when your process receives a SIGCHLD in response to	2337	Child watchers trigger when your process receives a SIGCHLD in response to
1437	some child status changes (most typically when a child of yours dies).	2338	some child status changes (most typically when a child of yours dies or
		2339	exits). It is permissible to install a child watcher I<after> the child
		2340	has been forked (which implies it might have already exited), as long
		2341	as the event loop isn't entered (or is continued from a watcher), i.e.,
		2342	forking and then immediately registering a watcher for the child is fine,
		2343	but forking and registering a watcher a few event loop iterations later or
		2344	in the next callback invocation is not.
		2345
		2346	Only the default event loop is capable of handling signals, and therefore
		2347	you can only register child watchers in the default event loop.
		2348
		2349	Due to some design glitches inside libev, child watchers will always be
		2350	handled at maximum priority (their priority is set to C<EV_MAXPRI> by
		2351	libev)
		2352
		2353	=head3 Process Interaction
		2354
		2355	Libev grabs C<SIGCHLD> as soon as the default event loop is
		2356	initialised. This is necessary to guarantee proper behaviour even if the
		2357	first child watcher is started after the child exits. The occurrence
		2358	of C<SIGCHLD> is recorded asynchronously, but child reaping is done
		2359	synchronously as part of the event loop processing. Libev always reaps all
		2360	children, even ones not watched.
		2361
		2362	=head3 Overriding the Built-In Processing
		2363
		2364	Libev offers no special support for overriding the built-in child
		2365	processing, but if your application collides with libev's default child
		2366	handler, you can override it easily by installing your own handler for
		2367	C<SIGCHLD> after initialising the default loop, and making sure the
		2368	default loop never gets destroyed. You are encouraged, however, to use an
		2369	event-based approach to child reaping and thus use libev's support for
		2370	that, so other libev users can use C<ev_child> watchers freely.
		2371
		2372	=head3 Stopping the Child Watcher
		2373
		2374	Currently, the child watcher never gets stopped, even when the
		2375	child terminates, so normally one needs to stop the watcher in the
		2376	callback. Future versions of libev might stop the watcher automatically
		2377	when a child exit is detected (calling C<ev_child_stop> twice is not a
		2378	problem).
1438		2379
1439	=head3 Watcher-Specific Functions and Data Members	2380	=head3 Watcher-Specific Functions and Data Members
1440		2381
1441	=over 4	2382	=over 4
1442		2383
…		…
1468		2409
1469	=back	2410	=back
1470		2411
1471	=head3 Examples	2412	=head3 Examples
1472		2413
1473	Example: Try to exit cleanly on SIGINT and SIGTERM.	2414	Example: C<fork()> a new process and install a child handler to wait for
		2415	its completion.
1474		2416
		2417	ev_child cw;
		2418
1475	static void	2419	static void
1476	sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)	2420	child_cb (EV_P_ ev_child *w, int revents)
1477	{	2421	{
1478	ev_unloop (loop, EVUNLOOP_ALL);	2422	ev_child_stop (EV_A_ w);
		2423	printf ("process %d exited with status %x\n", w->rpid, w->rstatus);
1479	}	2424	}
1480		2425
1481	struct ev_signal signal_watcher;	2426	pid_t pid = fork ();
1482	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);	2427
1483	ev_signal_start (loop, &sigint_cb);	2428	if (pid < 0)
		2429	// error
		2430	else if (pid == 0)
		2431	{
		2432	// the forked child executes here
		2433	exit (1);
		2434	}
		2435	else
		2436	{
		2437	ev_child_init (&cw, child_cb, pid, 0);
		2438	ev_child_start (EV_DEFAULT_ &cw);
		2439	}
1484		2440
1485		2441
1486	=head2 C<ev_stat> - did the file attributes just change?	2442	=head2 C<ev_stat> - did the file attributes just change?
1487		2443
1488	This watches a filesystem path for attribute changes. That is, it calls	2444	This watches a file system path for attribute changes. That is, it calls
1489	C<stat> regularly (or when the OS says it changed) and sees if it changed	2445	C<stat> on that path in regular intervals (or when the OS says it changed)
1490	compared to the last time, invoking the callback if it did.	2446	and sees if it changed compared to the last time, invoking the callback if
		2447	it did.
1491		2448
1492	The path does not need to exist: changing from "path exists" to "path does	2449	The path does not need to exist: changing from "path exists" to "path does
1493	not exist" is a status change like any other. The condition "path does	2450	not exist" is a status change like any other. The condition "path does not
1494	not exist" is signified by the C<st_nlink> field being zero (which is	2451	exist" (or more correctly "path cannot be stat'ed") is signified by the
1495	otherwise always forced to be at least one) and all the other fields of	2452	C<st_nlink> field being zero (which is otherwise always forced to be at
1496	the stat buffer having unspecified contents.	2453	least one) and all the other fields of the stat buffer having unspecified
		2454	contents.
1497		2455
1498	The path I<should> be absolute and I<must not> end in a slash. If it is	2456	The path I<must not> end in a slash or contain special components such as
		2457	C<.> or C<..>. The path I<should> be absolute: If it is relative and
1499	relative and your working directory changes, the behaviour is undefined.	2458	your working directory changes, then the behaviour is undefined.
1500		2459
1501	Since there is no standard to do this, the portable implementation simply	2460	Since there is no portable change notification interface available, the
1502	calls C<stat (2)> regularly on the path to see if it changed somehow. You	2461	portable implementation simply calls C<stat(2)> regularly on the path
1503	can specify a recommended polling interval for this case. If you specify	2462	to see if it changed somehow. You can specify a recommended polling
1504	a polling interval of C<0> (highly recommended!) then a I<suitable,	2463	interval for this case. If you specify a polling interval of C<0> (highly
1505	unspecified default> value will be used (which you can expect to be around	2464	recommended!) then a I<suitable, unspecified default> value will be used
1506	five seconds, although this might change dynamically). Libev will also	2465	(which you can expect to be around five seconds, although this might
1507	impose a minimum interval which is currently around C<0.1>, but thats	2466	change dynamically). Libev will also impose a minimum interval which is
1508	usually overkill.	2467	currently around C<0.1>, but that's usually overkill.
1509		2468
1510	This watcher type is not meant for massive numbers of stat watchers,	2469	This watcher type is not meant for massive numbers of stat watchers,
1511	as even with OS-supported change notifications, this can be	2470	as even with OS-supported change notifications, this can be
1512	resource-intensive.	2471	resource-intensive.
1513		2472
1514	At the time of this writing, only the Linux inotify interface is	2473	At the time of this writing, the only OS-specific interface implemented
1515	implemented (implementing kqueue support is left as an exercise for the	2474	is the Linux inotify interface (implementing kqueue support is left as an
1516	reader). Inotify will be used to give hints only and should not change the	2475	exercise for the reader. Note, however, that the author sees no way of
1517	semantics of C<ev_stat> watchers, which means that libev sometimes needs	2476	implementing C<ev_stat> semantics with kqueue, except as a hint).
1518	to fall back to regular polling again even with inotify, but changes are
1519	usually detected immediately, and if the file exists there will be no
1520	polling.
1521		2477
1522	=head3 Inotify	2478	=head3 ABI Issues (Largefile Support)
1523		2479
		2480	Libev by default (unless the user overrides this) uses the default
		2481	compilation environment, which means that on systems with large file
		2482	support disabled by default, you get the 32 bit version of the stat
		2483	structure. When using the library from programs that change the ABI to
		2484	use 64 bit file offsets the programs will fail. In that case you have to
		2485	compile libev with the same flags to get binary compatibility. This is
		2486	obviously the case with any flags that change the ABI, but the problem is
		2487	most noticeably displayed with ev_stat and large file support.
		2488
		2489	The solution for this is to lobby your distribution maker to make large
		2490	file interfaces available by default (as e.g. FreeBSD does) and not
		2491	optional. Libev cannot simply switch on large file support because it has
		2492	to exchange stat structures with application programs compiled using the
		2493	default compilation environment.
		2494
		2495	=head3 Inotify and Kqueue
		2496
1524	When C<inotify (7)> support has been compiled into libev (generally only	2497	When C<inotify (7)> support has been compiled into libev and present at
1525	available on Linux) and present at runtime, it will be used to speed up	2498	runtime, it will be used to speed up change detection where possible. The
1526	change detection where possible. The inotify descriptor will be created lazily	2499	inotify descriptor will be created lazily when the first C<ev_stat>
1527	when the first C<ev_stat> watcher is being started.	2500	watcher is being started.
1528		2501
1529	Inotify presense does not change the semantics of C<ev_stat> watchers	2502	Inotify presence does not change the semantics of C<ev_stat> watchers
1530	except that changes might be detected earlier, and in some cases, to avoid	2503	except that changes might be detected earlier, and in some cases, to avoid
1531	making regular C<stat> calls. Even in the presense of inotify support	2504	making regular C<stat> calls. Even in the presence of inotify support
1532	there are many cases where libev has to resort to regular C<stat> polling.	2505	there are many cases where libev has to resort to regular C<stat> polling,
		2506	but as long as kernel 2.6.25 or newer is used (2.6.24 and older have too
		2507	many bugs), the path exists (i.e. stat succeeds), and the path resides on
		2508	a local filesystem (libev currently assumes only ext2/3, jfs, reiserfs and
		2509	xfs are fully working) libev usually gets away without polling.
1533		2510
1534	(There is no support for kqueue, as apparently it cannot be used to	2511	There is no support for kqueue, as apparently it cannot be used to
1535	implement this functionality, due to the requirement of having a file	2512	implement this functionality, due to the requirement of having a file
1536	descriptor open on the object at all times).	2513	descriptor open on the object at all times, and detecting renames, unlinks
		2514	etc. is difficult.
		2515
		2516	=head3 C<stat ()> is a synchronous operation
		2517
		2518	Libev doesn't normally do any kind of I/O itself, and so is not blocking
		2519	the process. The exception are C<ev_stat> watchers - those call C<stat
		2520	()>, which is a synchronous operation.
		2521
		2522	For local paths, this usually doesn't matter: unless the system is very
		2523	busy or the intervals between stat's are large, a stat call will be fast,
		2524	as the path data is usually in memory already (except when starting the
		2525	watcher).
		2526
		2527	For networked file systems, calling C<stat ()> can block an indefinite
		2528	time due to network issues, and even under good conditions, a stat call
		2529	often takes multiple milliseconds.
		2530
		2531	Therefore, it is best to avoid using C<ev_stat> watchers on networked
		2532	paths, although this is fully supported by libev.
1537		2533
1538	=head3 The special problem of stat time resolution	2534	=head3 The special problem of stat time resolution
1539		2535
1540	The C<stat ()> syscall only supports full-second resolution portably, and	2536	The C<stat ()> system call only supports full-second resolution portably,
1541	even on systems where the resolution is higher, many filesystems still	2537	and even on systems where the resolution is higher, most file systems
1542	only support whole seconds.	2538	still only support whole seconds.
1543		2539
1544	That means that, if the time is the only thing that changes, you might	2540	That means that, if the time is the only thing that changes, you can
1545	miss updates: on the first update, C<ev_stat> detects a change and calls	2541	easily miss updates: on the first update, C<ev_stat> detects a change and
1546	your callback, which does something. When there is another update within	2542	calls your callback, which does something. When there is another update
1547	the same second, C<ev_stat> will be unable to detect it.	2543	within the same second, C<ev_stat> will be unable to detect unless the
		2544	stat data does change in other ways (e.g. file size).
1548		2545
1549	The solution to this is to delay acting on a change for a second (or till	2546	The solution to this is to delay acting on a change for slightly more
1550	the next second boundary), using a roughly one-second delay C<ev_timer>	2547	than a second (or till slightly after the next full second boundary), using
1551	(C<ev_timer_set (w, 0., 1.01); ev_timer_again (loop, w)>). The C<.01>	2548	a roughly one-second-delay C<ev_timer> (e.g. C<ev_timer_set (w, 0., 1.02);
1552	is added to work around small timing inconsistencies of some operating	2549	ev_timer_again (loop, w)>).
1553	systems.	2550
		2551	The C<.02> offset is added to work around small timing inconsistencies
		2552	of some operating systems (where the second counter of the current time
		2553	might be be delayed. One such system is the Linux kernel, where a call to
		2554	C<gettimeofday> might return a timestamp with a full second later than
		2555	a subsequent C<time> call - if the equivalent of C<time ()> is used to
		2556	update file times then there will be a small window where the kernel uses
		2557	the previous second to update file times but libev might already execute
		2558	the timer callback).
1554		2559
1555	=head3 Watcher-Specific Functions and Data Members	2560	=head3 Watcher-Specific Functions and Data Members
1556		2561
1557	=over 4	2562	=over 4
1558		2563
…		…
1564	C<path>. The C<interval> is a hint on how quickly a change is expected to	2569	C<path>. The C<interval> is a hint on how quickly a change is expected to
1565	be detected and should normally be specified as C<0> to let libev choose	2570	be detected and should normally be specified as C<0> to let libev choose
1566	a suitable value. The memory pointed to by C<path> must point to the same	2571	a suitable value. The memory pointed to by C<path> must point to the same
1567	path for as long as the watcher is active.	2572	path for as long as the watcher is active.
1568		2573
1569	The callback will be receive C<EV_STAT> when a change was detected,	2574	The callback will receive an C<EV_STAT> event when a change was detected,
1570	relative to the attributes at the time the watcher was started (or the	2575	relative to the attributes at the time the watcher was started (or the
1571	last change was detected).	2576	last change was detected).
1572		2577
1573	=item ev_stat_stat (ev_stat *)	2578	=item ev_stat_stat (loop, ev_stat *)
1574		2579
1575	Updates the stat buffer immediately with new values. If you change the	2580	Updates the stat buffer immediately with new values. If you change the
1576	watched path in your callback, you could call this fucntion to avoid	2581	watched path in your callback, you could call this function to avoid
1577	detecting this change (while introducing a race condition). Can also be	2582	detecting this change (while introducing a race condition if you are not
1578	useful simply to find out the new values.	2583	the only one changing the path). Can also be useful simply to find out the
		2584	new values.
1579		2585
1580	=item ev_statdata attr [read-only]	2586	=item ev_statdata attr [read-only]
1581		2587
1582	The most-recently detected attributes of the file. Although the type is of	2588	The most-recently detected attributes of the file. Although the type is
1583	C<ev_statdata>, this is usually the (or one of the) C<struct stat> types	2589	C<ev_statdata>, this is usually the (or one of the) C<struct stat> types
1584	suitable for your system. If the C<st_nlink> member is C<0>, then there	2590	suitable for your system, but you can only rely on the POSIX-standardised
		2591	members to be present. If the C<st_nlink> member is C<0>, then there was
1585	was some error while C<stat>ing the file.	2592	some error while C<stat>ing the file.
1586		2593
1587	=item ev_statdata prev [read-only]	2594	=item ev_statdata prev [read-only]
1588		2595
1589	The previous attributes of the file. The callback gets invoked whenever	2596	The previous attributes of the file. The callback gets invoked whenever
1590	C<prev> != C<attr>.	2597	C<prev> != C<attr>, or, more precisely, one or more of these members
		2598	differ: C<st_dev>, C<st_ino>, C<st_mode>, C<st_nlink>, C<st_uid>,
		2599	C<st_gid>, C<st_rdev>, C<st_size>, C<st_atime>, C<st_mtime>, C<st_ctime>.
1591		2600
1592	=item ev_tstamp interval [read-only]	2601	=item ev_tstamp interval [read-only]
1593		2602
1594	The specified interval.	2603	The specified interval.
1595		2604
1596	=item const char *path [read-only]	2605	=item const char *path [read-only]
1597		2606
1598	The filesystem path that is being watched.	2607	The file system path that is being watched.
1599		2608
1600	=back	2609	=back
1601		2610
1602	=head3 Examples	2611	=head3 Examples
1603		2612
1604	Example: Watch C</etc/passwd> for attribute changes.	2613	Example: Watch C</etc/passwd> for attribute changes.
1605		2614
1606	static void	2615	static void
1607	passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)	2616	passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)
1608	{	2617	{
1609	/* /etc/passwd changed in some way */	2618	/* /etc/passwd changed in some way */
1610	if (w->attr.st_nlink)	2619	if (w->attr.st_nlink)
1611	{	2620	{
1612	printf ("passwd current size %ld\n", (long)w->attr.st_size);	2621	printf ("passwd current size %ld\n", (long)w->attr.st_size);
1613	printf ("passwd current atime %ld\n", (long)w->attr.st_mtime);	2622	printf ("passwd current atime %ld\n", (long)w->attr.st_mtime);
1614	printf ("passwd current mtime %ld\n", (long)w->attr.st_mtime);	2623	printf ("passwd current mtime %ld\n", (long)w->attr.st_mtime);
1615	}	2624	}
1616	else	2625	else
1617	/* you shalt not abuse printf for puts */	2626	/* you shalt not abuse printf for puts */
1618	puts ("wow, /etc/passwd is not there, expect problems. "	2627	puts ("wow, /etc/passwd is not there, expect problems. "
1619	"if this is windows, they already arrived\n");	2628	"if this is windows, they already arrived\n");
1620	}	2629	}
1621		2630
1622	...	2631	...
1623	ev_stat passwd;	2632	ev_stat passwd;
1624		2633
1625	ev_stat_init (&passwd, passwd_cb, "/etc/passwd", 0.);	2634	ev_stat_init (&passwd, passwd_cb, "/etc/passwd", 0.);
1626	ev_stat_start (loop, &passwd);	2635	ev_stat_start (loop, &passwd);
1627		2636
1628	Example: Like above, but additionally use a one-second delay so we do not	2637	Example: Like above, but additionally use a one-second delay so we do not
1629	miss updates (however, frequent updates will delay processing, too, so	2638	miss updates (however, frequent updates will delay processing, too, so
1630	one might do the work both on C<ev_stat> callback invocation I<and> on	2639	one might do the work both on C<ev_stat> callback invocation I<and> on
1631	C<ev_timer> callback invocation).	2640	C<ev_timer> callback invocation).
1632		2641
1633	static ev_stat passwd;	2642	static ev_stat passwd;
1634	static ev_timer timer;	2643	static ev_timer timer;
1635		2644
1636	static void	2645	static void
1637	timer_cb (EV_P_ ev_timer *w, int revents)	2646	timer_cb (EV_P_ ev_timer *w, int revents)
1638	{	2647	{
1639	ev_timer_stop (EV_A_ w);	2648	ev_timer_stop (EV_A_ w);
1640		2649
1641	/* now it's one second after the most recent passwd change */	2650	/* now it's one second after the most recent passwd change */
1642	}	2651	}
1643		2652
1644	static void	2653	static void
1645	stat_cb (EV_P_ ev_stat *w, int revents)	2654	stat_cb (EV_P_ ev_stat *w, int revents)
1646	{	2655	{
1647	/* reset the one-second timer */	2656	/* reset the one-second timer */
1648	ev_timer_again (EV_A_ &timer);	2657	ev_timer_again (EV_A_ &timer);
1649	}	2658	}
1650		2659
1651	...	2660	...
1652	ev_stat_init (&passwd, stat_cb, "/etc/passwd", 0.);	2661	ev_stat_init (&passwd, stat_cb, "/etc/passwd", 0.);
1653	ev_stat_start (loop, &passwd);	2662	ev_stat_start (loop, &passwd);
1654	ev_timer_init (&timer, timer_cb, 0., 1.01);	2663	ev_timer_init (&timer, timer_cb, 0., 1.02);
1655		2664
1656		2665
1657	=head2 C<ev_idle> - when you've got nothing better to do...	2666	=head2 C<ev_idle> - when you've got nothing better to do...
1658		2667
1659	Idle watchers trigger events when no other events of the same or higher	2668	Idle watchers trigger events when no other events of the same or higher
1660	priority are pending (prepare, check and other idle watchers do not	2669	priority are pending (prepare, check and other idle watchers do not count
1661	count).	2670	as receiving "events").
1662		2671
1663	That is, as long as your process is busy handling sockets or timeouts	2672	That is, as long as your process is busy handling sockets or timeouts
1664	(or even signals, imagine) of the same or higher priority it will not be	2673	(or even signals, imagine) of the same or higher priority it will not be
1665	triggered. But when your process is idle (or only lower-priority watchers	2674	triggered. But when your process is idle (or only lower-priority watchers
1666	are pending), the idle watchers are being called once per event loop	2675	are pending), the idle watchers are being called once per event loop
…		…
1677		2686
1678	=head3 Watcher-Specific Functions and Data Members	2687	=head3 Watcher-Specific Functions and Data Members
1679		2688
1680	=over 4	2689	=over 4
1681		2690
1682	=item ev_idle_init (ev_signal *, callback)	2691	=item ev_idle_init (ev_idle *, callback)
1683		2692
1684	Initialises and configures the idle watcher - it has no parameters of any	2693	Initialises and configures the idle watcher - it has no parameters of any
1685	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,	2694	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,
1686	believe me.	2695	believe me.
1687		2696
…		…
1690	=head3 Examples	2699	=head3 Examples
1691		2700
1692	Example: Dynamically allocate an C<ev_idle> watcher, start it, and in the	2701	Example: Dynamically allocate an C<ev_idle> watcher, start it, and in the
1693	callback, free it. Also, use no error checking, as usual.	2702	callback, free it. Also, use no error checking, as usual.
1694		2703
1695	static void	2704	static void
1696	idle_cb (struct ev_loop *loop, struct ev_idle *w, int revents)	2705	idle_cb (struct ev_loop *loop, ev_idle *w, int revents)
1697	{	2706	{
1698	free (w);	2707	free (w);
1699	// now do something you wanted to do when the program has	2708	// now do something you wanted to do when the program has
1700	// no longer anything immediate to do.	2709	// no longer anything immediate to do.
1701	}	2710	}
1702		2711
1703	struct ev_idle *idle_watcher = malloc (sizeof (struct ev_idle));	2712	ev_idle *idle_watcher = malloc (sizeof (ev_idle));
1704	ev_idle_init (idle_watcher, idle_cb);	2713	ev_idle_init (idle_watcher, idle_cb);
1705	ev_idle_start (loop, idle_cb);	2714	ev_idle_start (loop, idle_watcher);
1706		2715
1707		2716
1708	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!	2717	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!
1709		2718
1710	Prepare and check watchers are usually (but not always) used in tandem:	2719	Prepare and check watchers are usually (but not always) used in pairs:
1711	prepare watchers get invoked before the process blocks and check watchers	2720	prepare watchers get invoked before the process blocks and check watchers
1712	afterwards.	2721	afterwards.
1713		2722
1714	You I<must not> call C<ev_loop> or similar functions that enter	2723	You I<must not> call C<ev_run> or similar functions that enter
1715	the current event loop from either C<ev_prepare> or C<ev_check>	2724	the current event loop from either C<ev_prepare> or C<ev_check>
1716	watchers. Other loops than the current one are fine, however. The	2725	watchers. Other loops than the current one are fine, however. The
1717	rationale behind this is that you do not need to check for recursion in	2726	rationale behind this is that you do not need to check for recursion in
1718	those watchers, i.e. the sequence will always be C<ev_prepare>, blocking,	2727	those watchers, i.e. the sequence will always be C<ev_prepare>, blocking,
1719	C<ev_check> so if you have one watcher of each kind they will always be	2728	C<ev_check> so if you have one watcher of each kind they will always be
1720	called in pairs bracketing the blocking call.	2729	called in pairs bracketing the blocking call.
1721		2730
1722	Their main purpose is to integrate other event mechanisms into libev and	2731	Their main purpose is to integrate other event mechanisms into libev and
1723	their use is somewhat advanced. This could be used, for example, to track	2732	their use is somewhat advanced. They could be used, for example, to track
1724	variable changes, implement your own watchers, integrate net-snmp or a	2733	variable changes, implement your own watchers, integrate net-snmp or a
1725	coroutine library and lots more. They are also occasionally useful if	2734	coroutine library and lots more. They are also occasionally useful if
1726	you cache some data and want to flush it before blocking (for example,	2735	you cache some data and want to flush it before blocking (for example,
1727	in X programs you might want to do an C<XFlush ()> in an C<ev_prepare>	2736	in X programs you might want to do an C<XFlush ()> in an C<ev_prepare>
1728	watcher).	2737	watcher).
1729		2738
1730	This is done by examining in each prepare call which file descriptors need	2739	This is done by examining in each prepare call which file descriptors
1731	to be watched by the other library, registering C<ev_io> watchers for	2740	need to be watched by the other library, registering C<ev_io> watchers
1732	them and starting an C<ev_timer> watcher for any timeouts (many libraries	2741	for them and starting an C<ev_timer> watcher for any timeouts (many
1733	provide just this functionality). Then, in the check watcher you check for	2742	libraries provide exactly this functionality). Then, in the check watcher,
1734	any events that occured (by checking the pending status of all watchers	2743	you check for any events that occurred (by checking the pending status
1735	and stopping them) and call back into the library. The I/O and timer	2744	of all watchers and stopping them) and call back into the library. The
1736	callbacks will never actually be called (but must be valid nevertheless,	2745	I/O and timer callbacks will never actually be called (but must be valid
1737	because you never know, you know?).	2746	nevertheless, because you never know, you know?).
1738		2747
1739	As another example, the Perl Coro module uses these hooks to integrate	2748	As another example, the Perl Coro module uses these hooks to integrate
1740	coroutines into libev programs, by yielding to other active coroutines	2749	coroutines into libev programs, by yielding to other active coroutines
1741	during each prepare and only letting the process block if no coroutines	2750	during each prepare and only letting the process block if no coroutines
1742	are ready to run (it's actually more complicated: it only runs coroutines	2751	are ready to run (it's actually more complicated: it only runs coroutines
…		…
1745	loop from blocking if lower-priority coroutines are active, thus mapping	2754	loop from blocking if lower-priority coroutines are active, thus mapping
1746	low-priority coroutines to idle/background tasks).	2755	low-priority coroutines to idle/background tasks).
1747		2756
1748	It is recommended to give C<ev_check> watchers highest (C<EV_MAXPRI>)	2757	It is recommended to give C<ev_check> watchers highest (C<EV_MAXPRI>)
1749	priority, to ensure that they are being run before any other watchers	2758	priority, to ensure that they are being run before any other watchers
		2759	after the poll (this doesn't matter for C<ev_prepare> watchers).
		2760
1750	after the poll. Also, C<ev_check> watchers (and C<ev_prepare> watchers,	2761	Also, C<ev_check> watchers (and C<ev_prepare> watchers, too) should not
1751	too) should not activate ("feed") events into libev. While libev fully	2762	activate ("feed") events into libev. While libev fully supports this, they
1752	supports this, they will be called before other C<ev_check> watchers	2763	might get executed before other C<ev_check> watchers did their job. As
1753	did their job. As C<ev_check> watchers are often used to embed other	2764	C<ev_check> watchers are often used to embed other (non-libev) event
1754	(non-libev) event loops those other event loops might be in an unusable	2765	loops those other event loops might be in an unusable state until their
1755	state until their C<ev_check> watcher ran (always remind yourself to	2766	C<ev_check> watcher ran (always remind yourself to coexist peacefully with
1756	coexist peacefully with others).	2767	others).
1757		2768
1758	=head3 Watcher-Specific Functions and Data Members	2769	=head3 Watcher-Specific Functions and Data Members
1759		2770
1760	=over 4	2771	=over 4
1761		2772
…		…
1763		2774
1764	=item ev_check_init (ev_check *, callback)	2775	=item ev_check_init (ev_check *, callback)
1765		2776
1766	Initialises and configures the prepare or check watcher - they have no	2777	Initialises and configures the prepare or check watcher - they have no
1767	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>	2778	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>
1768	macros, but using them is utterly, utterly and completely pointless.	2779	macros, but using them is utterly, utterly, utterly and completely
		2780	pointless.
1769		2781
1770	=back	2782	=back
1771		2783
1772	=head3 Examples	2784	=head3 Examples
1773		2785
1774	There are a number of principal ways to embed other event loops or modules	2786	There are a number of principal ways to embed other event loops or modules
1775	into libev. Here are some ideas on how to include libadns into libev	2787	into libev. Here are some ideas on how to include libadns into libev
1776	(there is a Perl module named C<EV::ADNS> that does this, which you could	2788	(there is a Perl module named C<EV::ADNS> that does this, which you could
1777	use for an actually working example. Another Perl module named C<EV::Glib>	2789	use as a working example. Another Perl module named C<EV::Glib> embeds a
1778	embeds a Glib main context into libev, and finally, C<Glib::EV> embeds EV	2790	Glib main context into libev, and finally, C<Glib::EV> embeds EV into the
1779	into the Glib event loop).	2791	Glib event loop).
1780		2792
1781	Method 1: Add IO watchers and a timeout watcher in a prepare handler,	2793	Method 1: Add IO watchers and a timeout watcher in a prepare handler,
1782	and in a check watcher, destroy them and call into libadns. What follows	2794	and in a check watcher, destroy them and call into libadns. What follows
1783	is pseudo-code only of course. This requires you to either use a low	2795	is pseudo-code only of course. This requires you to either use a low
1784	priority for the check watcher or use C<ev_clear_pending> explicitly, as	2796	priority for the check watcher or use C<ev_clear_pending> explicitly, as
1785	the callbacks for the IO/timeout watchers might not have been called yet.	2797	the callbacks for the IO/timeout watchers might not have been called yet.
1786		2798
1787	static ev_io iow [nfd];	2799	static ev_io iow [nfd];
1788	static ev_timer tw;	2800	static ev_timer tw;
1789		2801
1790	static void	2802	static void
1791	io_cb (ev_loop *loop, ev_io *w, int revents)	2803	io_cb (struct ev_loop *loop, ev_io *w, int revents)
1792	{	2804	{
1793	}	2805	}
1794		2806
1795	// create io watchers for each fd and a timer before blocking	2807	// create io watchers for each fd and a timer before blocking
1796	static void	2808	static void
1797	adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents)	2809	adns_prepare_cb (struct ev_loop *loop, ev_prepare *w, int revents)
1798	{	2810	{
1799	int timeout = 3600000;	2811	int timeout = 3600000;
1800	struct pollfd fds [nfd];	2812	struct pollfd fds [nfd];
1801	// actual code will need to loop here and realloc etc.	2813	// actual code will need to loop here and realloc etc.
1802	adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));	2814	adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));
1803		2815
1804	/* the callback is illegal, but won't be called as we stop during check */	2816	/* the callback is illegal, but won't be called as we stop during check */
1805	ev_timer_init (&tw, 0, timeout * 1e-3);	2817	ev_timer_init (&tw, 0, timeout * 1e-3, 0.);
1806	ev_timer_start (loop, &tw);	2818	ev_timer_start (loop, &tw);
1807		2819
1808	// create one ev_io per pollfd	2820	// create one ev_io per pollfd
1809	for (int i = 0; i < nfd; ++i)	2821	for (int i = 0; i < nfd; ++i)
1810	{	2822	{
1811	ev_io_init (iow + i, io_cb, fds [i].fd,	2823	ev_io_init (iow + i, io_cb, fds [i].fd,
1812	((fds [i].events & POLLIN ? EV_READ : 0)	2824	((fds [i].events & POLLIN ? EV_READ : 0)
1813	| (fds [i].events & POLLOUT ? EV_WRITE : 0)));	2825	| (fds [i].events & POLLOUT ? EV_WRITE : 0)));
1814		2826
1815	fds [i].revents = 0;	2827	fds [i].revents = 0;
1816	ev_io_start (loop, iow + i);	2828	ev_io_start (loop, iow + i);
1817	}	2829	}
1818	}	2830	}
1819		2831
1820	// stop all watchers after blocking	2832	// stop all watchers after blocking
1821	static void	2833	static void
1822	adns_check_cb (ev_loop *loop, ev_check *w, int revents)	2834	adns_check_cb (struct ev_loop *loop, ev_check *w, int revents)
1823	{	2835	{
1824	ev_timer_stop (loop, &tw);	2836	ev_timer_stop (loop, &tw);
1825		2837
1826	for (int i = 0; i < nfd; ++i)	2838	for (int i = 0; i < nfd; ++i)
1827	{	2839	{
1828	// set the relevant poll flags	2840	// set the relevant poll flags
1829	// could also call adns_processreadable etc. here	2841	// could also call adns_processreadable etc. here
1830	struct pollfd *fd = fds + i;	2842	struct pollfd *fd = fds + i;
1831	int revents = ev_clear_pending (iow + i);	2843	int revents = ev_clear_pending (iow + i);
1832	if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;	2844	if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;
1833	if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;	2845	if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;
1834		2846
1835	// now stop the watcher	2847	// now stop the watcher
1836	ev_io_stop (loop, iow + i);	2848	ev_io_stop (loop, iow + i);
1837	}	2849	}
1838		2850
1839	adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));	2851	adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));
1840	}	2852	}
1841		2853
1842	Method 2: This would be just like method 1, but you run C<adns_afterpoll>	2854	Method 2: This would be just like method 1, but you run C<adns_afterpoll>
1843	in the prepare watcher and would dispose of the check watcher.	2855	in the prepare watcher and would dispose of the check watcher.
1844		2856
1845	Method 3: If the module to be embedded supports explicit event	2857	Method 3: If the module to be embedded supports explicit event
1846	notification (adns does), you can also make use of the actual watcher	2858	notification (libadns does), you can also make use of the actual watcher
1847	callbacks, and only destroy/create the watchers in the prepare watcher.	2859	callbacks, and only destroy/create the watchers in the prepare watcher.
1848		2860
1849	static void	2861	static void
1850	timer_cb (EV_P_ ev_timer *w, int revents)	2862	timer_cb (EV_P_ ev_timer *w, int revents)
1851	{	2863	{
1852	adns_state ads = (adns_state)w->data;	2864	adns_state ads = (adns_state)w->data;
1853	update_now (EV_A);	2865	update_now (EV_A);
1854		2866
1855	adns_processtimeouts (ads, &tv_now);	2867	adns_processtimeouts (ads, &tv_now);
1856	}	2868	}
1857		2869
1858	static void	2870	static void
1859	io_cb (EV_P_ ev_io *w, int revents)	2871	io_cb (EV_P_ ev_io *w, int revents)
1860	{	2872	{
1861	adns_state ads = (adns_state)w->data;	2873	adns_state ads = (adns_state)w->data;
1862	update_now (EV_A);	2874	update_now (EV_A);
1863		2875
1864	if (revents & EV_READ ) adns_processreadable (ads, w->fd, &tv_now);	2876	if (revents & EV_READ ) adns_processreadable (ads, w->fd, &tv_now);
1865	if (revents & EV_WRITE) adns_processwriteable (ads, w->fd, &tv_now);	2877	if (revents & EV_WRITE) adns_processwriteable (ads, w->fd, &tv_now);
1866	}	2878	}
1867		2879
1868	// do not ever call adns_afterpoll	2880	// do not ever call adns_afterpoll
1869		2881
1870	Method 4: Do not use a prepare or check watcher because the module you	2882	Method 4: Do not use a prepare or check watcher because the module you
1871	want to embed is too inflexible to support it. Instead, youc na override	2883	want to embed is not flexible enough to support it. Instead, you can
1872	their poll function. The drawback with this solution is that the main	2884	override their poll function. The drawback with this solution is that the
1873	loop is now no longer controllable by EV. The C<Glib::EV> module does	2885	main loop is now no longer controllable by EV. The C<Glib::EV> module uses
1874	this.	2886	this approach, effectively embedding EV as a client into the horrible
		2887	libglib event loop.
1875		2888
1876	static gint	2889	static gint
1877	event_poll_func (GPollFD *fds, guint nfds, gint timeout)	2890	event_poll_func (GPollFD *fds, guint nfds, gint timeout)
1878	{	2891	{
1879	int got_events = 0;	2892	int got_events = 0;
1880		2893
1881	for (n = 0; n < nfds; ++n)	2894	for (n = 0; n < nfds; ++n)
1882	// create/start io watcher that sets the relevant bits in fds[n] and increment got_events	2895	// create/start io watcher that sets the relevant bits in fds[n] and increment got_events
1883		2896
1884	if (timeout >= 0)	2897	if (timeout >= 0)
1885	// create/start timer	2898	// create/start timer
1886		2899
1887	// poll	2900	// poll
1888	ev_loop (EV_A_ 0);	2901	ev_run (EV_A_ 0);
1889		2902
1890	// stop timer again	2903	// stop timer again
1891	if (timeout >= 0)	2904	if (timeout >= 0)
1892	ev_timer_stop (EV_A_ &to);	2905	ev_timer_stop (EV_A_ &to);
1893		2906
1894	// stop io watchers again - their callbacks should have set	2907	// stop io watchers again - their callbacks should have set
1895	for (n = 0; n < nfds; ++n)	2908	for (n = 0; n < nfds; ++n)
1896	ev_io_stop (EV_A_ iow [n]);	2909	ev_io_stop (EV_A_ iow [n]);
1897		2910
1898	return got_events;	2911	return got_events;
1899	}	2912	}
1900		2913
1901		2914
1902	=head2 C<ev_embed> - when one backend isn't enough...	2915	=head2 C<ev_embed> - when one backend isn't enough...
1903		2916
1904	This is a rather advanced watcher type that lets you embed one event loop	2917	This is a rather advanced watcher type that lets you embed one event loop
…		…
1910	prioritise I/O.	2923	prioritise I/O.
1911		2924
1912	As an example for a bug workaround, the kqueue backend might only support	2925	As an example for a bug workaround, the kqueue backend might only support
1913	sockets on some platform, so it is unusable as generic backend, but you	2926	sockets on some platform, so it is unusable as generic backend, but you
1914	still want to make use of it because you have many sockets and it scales	2927	still want to make use of it because you have many sockets and it scales
1915	so nicely. In this case, you would create a kqueue-based loop and embed it	2928	so nicely. In this case, you would create a kqueue-based loop and embed
1916	into your default loop (which might use e.g. poll). Overall operation will	2929	it into your default loop (which might use e.g. poll). Overall operation
1917	be a bit slower because first libev has to poll and then call kevent, but	2930	will be a bit slower because first libev has to call C<poll> and then
1918	at least you can use both at what they are best.	2931	C<kevent>, but at least you can use both mechanisms for what they are
		2932	best: C<kqueue> for scalable sockets and C<poll> if you want it to work :)
1919		2933
1920	As for prioritising I/O: rarely you have the case where some fds have	2934	As for prioritising I/O: under rare circumstances you have the case where
1921	to be watched and handled very quickly (with low latency), and even	2935	some fds have to be watched and handled very quickly (with low latency),
1922	priorities and idle watchers might have too much overhead. In this case	2936	and even priorities and idle watchers might have too much overhead. In
1923	you would put all the high priority stuff in one loop and all the rest in	2937	this case you would put all the high priority stuff in one loop and all
1924	a second one, and embed the second one in the first.	2938	the rest in a second one, and embed the second one in the first.
1925		2939
1926	As long as the watcher is active, the callback will be invoked every time	2940	As long as the watcher is active, the callback will be invoked every
1927	there might be events pending in the embedded loop. The callback must then	2941	time there might be events pending in the embedded loop. The callback
1928	call C<ev_embed_sweep (mainloop, watcher)> to make a single sweep and invoke	2942	must then call C<ev_embed_sweep (mainloop, watcher)> to make a single
1929	their callbacks (you could also start an idle watcher to give the embedded	2943	sweep and invoke their callbacks (the callback doesn't need to invoke the
1930	loop strictly lower priority for example). You can also set the callback	2944	C<ev_embed_sweep> function directly, it could also start an idle watcher
1931	to C<0>, in which case the embed watcher will automatically execute the	2945	to give the embedded loop strictly lower priority for example).
1932	embedded loop sweep.
1933		2946
1934	As long as the watcher is started it will automatically handle events. The	2947	You can also set the callback to C<0>, in which case the embed watcher
1935	callback will be invoked whenever some events have been handled. You can	2948	will automatically execute the embedded loop sweep whenever necessary.
1936	set the callback to C<0> to avoid having to specify one if you are not
1937	interested in that.
1938		2949
1939	Also, there have not currently been made special provisions for forking:	2950	Fork detection will be handled transparently while the C<ev_embed> watcher
1940	when you fork, you not only have to call C<ev_loop_fork> on both loops,	2951	is active, i.e., the embedded loop will automatically be forked when the
1941	but you will also have to stop and restart any C<ev_embed> watchers	2952	embedding loop forks. In other cases, the user is responsible for calling
1942	yourself.	2953	C<ev_loop_fork> on the embedded loop.
1943		2954
1944	Unfortunately, not all backends are embeddable, only the ones returned by	2955	Unfortunately, not all backends are embeddable: only the ones returned by
1945	C<ev_embeddable_backends> are, which, unfortunately, does not include any	2956	C<ev_embeddable_backends> are, which, unfortunately, does not include any
1946	portable one.	2957	portable one.
1947		2958
1948	So when you want to use this feature you will always have to be prepared	2959	So when you want to use this feature you will always have to be prepared
1949	that you cannot get an embeddable loop. The recommended way to get around	2960	that you cannot get an embeddable loop. The recommended way to get around
1950	this is to have a separate variables for your embeddable loop, try to	2961	this is to have a separate variables for your embeddable loop, try to
1951	create it, and if that fails, use the normal loop for everything.	2962	create it, and if that fails, use the normal loop for everything.
1952		2963
		2964	=head3 C<ev_embed> and fork
		2965
		2966	While the C<ev_embed> watcher is running, forks in the embedding loop will
		2967	automatically be applied to the embedded loop as well, so no special
		2968	fork handling is required in that case. When the watcher is not running,
		2969	however, it is still the task of the libev user to call C<ev_loop_fork ()>
		2970	as applicable.
		2971
1953	=head3 Watcher-Specific Functions and Data Members	2972	=head3 Watcher-Specific Functions and Data Members
1954		2973
1955	=over 4	2974	=over 4
1956		2975
1957	=item ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)	2976	=item ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)
…		…
1960		2979
1961	Configures the watcher to embed the given loop, which must be	2980	Configures the watcher to embed the given loop, which must be
1962	embeddable. If the callback is C<0>, then C<ev_embed_sweep> will be	2981	embeddable. If the callback is C<0>, then C<ev_embed_sweep> will be
1963	invoked automatically, otherwise it is the responsibility of the callback	2982	invoked automatically, otherwise it is the responsibility of the callback
1964	to invoke it (it will continue to be called until the sweep has been done,	2983	to invoke it (it will continue to be called until the sweep has been done,
1965	if you do not want thta, you need to temporarily stop the embed watcher).	2984	if you do not want that, you need to temporarily stop the embed watcher).
1966		2985
1967	=item ev_embed_sweep (loop, ev_embed *)	2986	=item ev_embed_sweep (loop, ev_embed *)
1968		2987
1969	Make a single, non-blocking sweep over the embedded loop. This works	2988	Make a single, non-blocking sweep over the embedded loop. This works
1970	similarly to C<ev_loop (embedded_loop, EVLOOP_NONBLOCK)>, but in the most	2989	similarly to C<ev_run (embedded_loop, EVRUN_NOWAIT)>, but in the most
1971	apropriate way for embedded loops.	2990	appropriate way for embedded loops.
1972		2991
1973	=item struct ev_loop *other [read-only]	2992	=item struct ev_loop *other [read-only]
1974		2993
1975	The embedded event loop.	2994	The embedded event loop.
1976		2995
…		…
1978		2997
1979	=head3 Examples	2998	=head3 Examples
1980		2999
1981	Example: Try to get an embeddable event loop and embed it into the default	3000	Example: Try to get an embeddable event loop and embed it into the default
1982	event loop. If that is not possible, use the default loop. The default	3001	event loop. If that is not possible, use the default loop. The default
1983	loop is stored in C<loop_hi>, while the mebeddable loop is stored in	3002	loop is stored in C<loop_hi>, while the embeddable loop is stored in
1984	C<loop_lo> (which is C<loop_hi> in the acse no embeddable loop can be	3003	C<loop_lo> (which is C<loop_hi> in the case no embeddable loop can be
1985	used).	3004	used).
1986		3005
1987	struct ev_loop *loop_hi = ev_default_init (0);	3006	struct ev_loop *loop_hi = ev_default_init (0);
1988	struct ev_loop *loop_lo = 0;	3007	struct ev_loop *loop_lo = 0;
1989	struct ev_embed embed;	3008	ev_embed embed;
1990		3009
1991	// see if there is a chance of getting one that works	3010	// see if there is a chance of getting one that works
1992	// (remember that a flags value of 0 means autodetection)	3011	// (remember that a flags value of 0 means autodetection)
1993	loop_lo = ev_embeddable_backends () & ev_recommended_backends ()	3012	loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
1994	? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())	3013	? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
1995	: 0;	3014	: 0;
1996		3015
1997	// if we got one, then embed it, otherwise default to loop_hi	3016	// if we got one, then embed it, otherwise default to loop_hi
1998	if (loop_lo)	3017	if (loop_lo)
1999	{	3018	{
2000	ev_embed_init (&embed, 0, loop_lo);	3019	ev_embed_init (&embed, 0, loop_lo);
2001	ev_embed_start (loop_hi, &embed);	3020	ev_embed_start (loop_hi, &embed);
2002	}	3021	}
2003	else	3022	else
2004	loop_lo = loop_hi;	3023	loop_lo = loop_hi;
2005		3024
2006	Example: Check if kqueue is available but not recommended and create	3025	Example: Check if kqueue is available but not recommended and create
2007	a kqueue backend for use with sockets (which usually work with any	3026	a kqueue backend for use with sockets (which usually work with any
2008	kqueue implementation). Store the kqueue/socket-only event loop in	3027	kqueue implementation). Store the kqueue/socket-only event loop in
2009	C<loop_socket>. (One might optionally use C<EVFLAG_NOENV>, too).	3028	C<loop_socket>. (One might optionally use C<EVFLAG_NOENV>, too).
2010		3029
2011	struct ev_loop *loop = ev_default_init (0);	3030	struct ev_loop *loop = ev_default_init (0);
2012	struct ev_loop *loop_socket = 0;	3031	struct ev_loop *loop_socket = 0;
2013	struct ev_embed embed;	3032	ev_embed embed;
2014		3033
2015	if (ev_supported_backends () & ~ev_recommended_backends () & EVBACKEND_KQUEUE)	3034	if (ev_supported_backends () & ~ev_recommended_backends () & EVBACKEND_KQUEUE)
2016	if ((loop_socket = ev_loop_new (EVBACKEND_KQUEUE))	3035	if ((loop_socket = ev_loop_new (EVBACKEND_KQUEUE))
2017	{	3036	{
2018	ev_embed_init (&embed, 0, loop_socket);	3037	ev_embed_init (&embed, 0, loop_socket);
2019	ev_embed_start (loop, &embed);	3038	ev_embed_start (loop, &embed);
2020	}	3039	}
2021		3040
2022	if (!loop_socket)	3041	if (!loop_socket)
2023	loop_socket = loop;	3042	loop_socket = loop;
2024		3043
2025	// now use loop_socket for all sockets, and loop for everything else	3044	// now use loop_socket for all sockets, and loop for everything else
2026		3045
2027		3046
2028	=head2 C<ev_fork> - the audacity to resume the event loop after a fork	3047	=head2 C<ev_fork> - the audacity to resume the event loop after a fork
2029		3048
2030	Fork watchers are called when a C<fork ()> was detected (usually because	3049	Fork watchers are called when a C<fork ()> was detected (usually because
…		…
2033	event loop blocks next and before C<ev_check> watchers are being called,	3052	event loop blocks next and before C<ev_check> watchers are being called,
2034	and only in the child after the fork. If whoever good citizen calling	3053	and only in the child after the fork. If whoever good citizen calling
2035	C<ev_default_fork> cheats and calls it in the wrong process, the fork	3054	C<ev_default_fork> cheats and calls it in the wrong process, the fork
2036	handlers will be invoked, too, of course.	3055	handlers will be invoked, too, of course.
2037		3056
		3057	=head3 The special problem of life after fork - how is it possible?
		3058
		3059	Most uses of C<fork()> consist of forking, then some simple calls to set
		3060	up/change the process environment, followed by a call to C<exec()>. This
		3061	sequence should be handled by libev without any problems.
		3062
		3063	This changes when the application actually wants to do event handling
		3064	in the child, or both parent in child, in effect "continuing" after the
		3065	fork.
		3066
		3067	The default mode of operation (for libev, with application help to detect
		3068	forks) is to duplicate all the state in the child, as would be expected
		3069	when I<either> the parent I<or> the child process continues.
		3070
		3071	When both processes want to continue using libev, then this is usually the
		3072	wrong result. In that case, usually one process (typically the parent) is
		3073	supposed to continue with all watchers in place as before, while the other
		3074	process typically wants to start fresh, i.e. without any active watchers.
		3075
		3076	The cleanest and most efficient way to achieve that with libev is to
		3077	simply create a new event loop, which of course will be "empty", and
		3078	use that for new watchers. This has the advantage of not touching more
		3079	memory than necessary, and thus avoiding the copy-on-write, and the
		3080	disadvantage of having to use multiple event loops (which do not support
		3081	signal watchers).
		3082
		3083	When this is not possible, or you want to use the default loop for
		3084	other reasons, then in the process that wants to start "fresh", call
		3085	C<ev_loop_destroy (EV_DEFAULT)> followed by C<ev_default_loop (...)>.
		3086	Destroying the default loop will "orphan" (not stop) all registered
		3087	watchers, so you have to be careful not to execute code that modifies
		3088	those watchers. Note also that in that case, you have to re-register any
		3089	signal watchers.
		3090
2038	=head3 Watcher-Specific Functions and Data Members	3091	=head3 Watcher-Specific Functions and Data Members
2039		3092
2040	=over 4	3093	=over 4
2041		3094
2042	=item ev_fork_init (ev_signal *, callback)	3095	=item ev_fork_init (ev_signal *, callback)
…		…
2046	believe me.	3099	believe me.
2047		3100
2048	=back	3101	=back
2049		3102
2050		3103
		3104	=head2 C<ev_async> - how to wake up an event loop
		3105
		3106	In general, you cannot use an C<ev_run> from multiple threads or other
		3107	asynchronous sources such as signal handlers (as opposed to multiple event
		3108	loops - those are of course safe to use in different threads).
		3109
		3110	Sometimes, however, you need to wake up an event loop you do not control,
		3111	for example because it belongs to another thread. This is what C<ev_async>
		3112	watchers do: as long as the C<ev_async> watcher is active, you can signal
		3113	it by calling C<ev_async_send>, which is thread- and signal safe.
		3114
		3115	This functionality is very similar to C<ev_signal> watchers, as signals,
		3116	too, are asynchronous in nature, and signals, too, will be compressed
		3117	(i.e. the number of callback invocations may be less than the number of
		3118	C<ev_async_sent> calls).
		3119
		3120	Unlike C<ev_signal> watchers, C<ev_async> works with any event loop, not
		3121	just the default loop.
		3122
		3123	=head3 Queueing
		3124
		3125	C<ev_async> does not support queueing of data in any way. The reason
		3126	is that the author does not know of a simple (or any) algorithm for a
		3127	multiple-writer-single-reader queue that works in all cases and doesn't
		3128	need elaborate support such as pthreads or unportable memory access
		3129	semantics.
		3130
		3131	That means that if you want to queue data, you have to provide your own
		3132	queue. But at least I can tell you how to implement locking around your
		3133	queue:
		3134
		3135	=over 4
		3136
		3137	=item queueing from a signal handler context
		3138
		3139	To implement race-free queueing, you simply add to the queue in the signal
		3140	handler but you block the signal handler in the watcher callback. Here is
		3141	an example that does that for some fictitious SIGUSR1 handler:
		3142
		3143	static ev_async mysig;
		3144
		3145	static void
		3146	sigusr1_handler (void)
		3147	{
		3148	sometype data;
		3149
		3150	// no locking etc.
		3151	queue_put (data);
		3152	ev_async_send (EV_DEFAULT_ &mysig);
		3153	}
		3154
		3155	static void
		3156	mysig_cb (EV_P_ ev_async *w, int revents)
		3157	{
		3158	sometype data;
		3159	sigset_t block, prev;
		3160
		3161	sigemptyset (&block);
		3162	sigaddset (&block, SIGUSR1);
		3163	sigprocmask (SIG_BLOCK, &block, &prev);
		3164
		3165	while (queue_get (&data))
		3166	process (data);
		3167
		3168	if (sigismember (&prev, SIGUSR1)
		3169	sigprocmask (SIG_UNBLOCK, &block, 0);
		3170	}
		3171
		3172	(Note: pthreads in theory requires you to use C<pthread_setmask>
		3173	instead of C<sigprocmask> when you use threads, but libev doesn't do it
		3174	either...).
		3175
		3176	=item queueing from a thread context
		3177
		3178	The strategy for threads is different, as you cannot (easily) block
		3179	threads but you can easily preempt them, so to queue safely you need to
		3180	employ a traditional mutex lock, such as in this pthread example:
		3181
		3182	static ev_async mysig;
		3183	static pthread_mutex_t mymutex = PTHREAD_MUTEX_INITIALIZER;
		3184
		3185	static void
		3186	otherthread (void)
		3187	{
		3188	// only need to lock the actual queueing operation
		3189	pthread_mutex_lock (&mymutex);
		3190	queue_put (data);
		3191	pthread_mutex_unlock (&mymutex);
		3192
		3193	ev_async_send (EV_DEFAULT_ &mysig);
		3194	}
		3195
		3196	static void
		3197	mysig_cb (EV_P_ ev_async *w, int revents)
		3198	{
		3199	pthread_mutex_lock (&mymutex);
		3200
		3201	while (queue_get (&data))
		3202	process (data);
		3203
		3204	pthread_mutex_unlock (&mymutex);
		3205	}
		3206
		3207	=back
		3208
		3209
		3210	=head3 Watcher-Specific Functions and Data Members
		3211
		3212	=over 4
		3213
		3214	=item ev_async_init (ev_async *, callback)
		3215
		3216	Initialises and configures the async watcher - it has no parameters of any
		3217	kind. There is a C<ev_async_set> macro, but using it is utterly pointless,
		3218	trust me.
		3219
		3220	=item ev_async_send (loop, ev_async *)
		3221
		3222	Sends/signals/activates the given C<ev_async> watcher, that is, feeds
		3223	an C<EV_ASYNC> event on the watcher into the event loop. Unlike
		3224	C<ev_feed_event>, this call is safe to do from other threads, signal or
		3225	similar contexts (see the discussion of C<EV_ATOMIC_T> in the embedding
		3226	section below on what exactly this means).
		3227
		3228	Note that, as with other watchers in libev, multiple events might get
		3229	compressed into a single callback invocation (another way to look at this
		3230	is that C<ev_async> watchers are level-triggered, set on C<ev_async_send>,
		3231	reset when the event loop detects that).
		3232
		3233	This call incurs the overhead of a system call only once per event loop
		3234	iteration, so while the overhead might be noticeable, it doesn't apply to
		3235	repeated calls to C<ev_async_send> for the same event loop.
		3236
		3237	=item bool = ev_async_pending (ev_async *)
		3238
		3239	Returns a non-zero value when C<ev_async_send> has been called on the
		3240	watcher but the event has not yet been processed (or even noted) by the
		3241	event loop.
		3242
		3243	C<ev_async_send> sets a flag in the watcher and wakes up the loop. When
		3244	the loop iterates next and checks for the watcher to have become active,
		3245	it will reset the flag again. C<ev_async_pending> can be used to very
		3246	quickly check whether invoking the loop might be a good idea.
		3247
		3248	Not that this does I<not> check whether the watcher itself is pending,
		3249	only whether it has been requested to make this watcher pending: there
		3250	is a time window between the event loop checking and resetting the async
		3251	notification, and the callback being invoked.
		3252
		3253	=back
		3254
		3255
2051	=head1 OTHER FUNCTIONS	3256	=head1 OTHER FUNCTIONS
2052		3257
2053	There are some other functions of possible interest. Described. Here. Now.	3258	There are some other functions of possible interest. Described. Here. Now.
2054		3259
2055	=over 4	3260	=over 4
2056		3261
2057	=item ev_once (loop, int fd, int events, ev_tstamp timeout, callback)	3262	=item ev_once (loop, int fd, int events, ev_tstamp timeout, callback)
2058		3263
2059	This function combines a simple timer and an I/O watcher, calls your	3264	This function combines a simple timer and an I/O watcher, calls your
2060	callback on whichever event happens first and automatically stop both	3265	callback on whichever event happens first and automatically stops both
2061	watchers. This is useful if you want to wait for a single event on an fd	3266	watchers. This is useful if you want to wait for a single event on an fd
2062	or timeout without having to allocate/configure/start/stop/free one or	3267	or timeout without having to allocate/configure/start/stop/free one or
2063	more watchers yourself.	3268	more watchers yourself.
2064		3269
2065	If C<fd> is less than 0, then no I/O watcher will be started and events	3270	If C<fd> is less than 0, then no I/O watcher will be started and the
2066	is being ignored. Otherwise, an C<ev_io> watcher for the given C<fd> and	3271	C<events> argument is being ignored. Otherwise, an C<ev_io> watcher for
2067	C<events> set will be craeted and started.	3272	the given C<fd> and C<events> set will be created and started.
2068		3273
2069	If C<timeout> is less than 0, then no timeout watcher will be	3274	If C<timeout> is less than 0, then no timeout watcher will be
2070	started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and	3275	started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and
2071	repeat = 0) will be started. While C<0> is a valid timeout, it is of	3276	repeat = 0) will be started. C<0> is a valid timeout.
2072	dubious value.
2073		3277
2074	The callback has the type C<void (*cb)(int revents, void *arg)> and gets	3278	The callback has the type C<void (*cb)(int revents, void *arg)> and is
2075	passed an C<revents> set like normal event callbacks (a combination of	3279	passed an C<revents> set like normal event callbacks (a combination of
2076	C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMEOUT>) and the C<arg>	3280	C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMER>) and the C<arg>
2077	value passed to C<ev_once>:	3281	value passed to C<ev_once>. Note that it is possible to receive I<both>
		3282	a timeout and an io event at the same time - you probably should give io
		3283	events precedence.
2078		3284
		3285	Example: wait up to ten seconds for data to appear on STDIN_FILENO.
		3286
2079	static void stdin_ready (int revents, void *arg)	3287	static void stdin_ready (int revents, void *arg)
2080	{	3288	{
2081	if (revents & EV_TIMEOUT)
2082	/* doh, nothing entered */;
2083	else if (revents & EV_READ)	3289	if (revents & EV_READ)
2084	/* stdin might have data for us, joy! */;	3290	/* stdin might have data for us, joy! */;
		3291	else if (revents & EV_TIMER)
		3292	/* doh, nothing entered */;
2085	}	3293	}
2086		3294
2087	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);	3295	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
2088		3296
2089	=item ev_feed_event (ev_loop *, watcher *, int revents)
2090
2091	Feeds the given event set into the event loop, as if the specified event
2092	had happened for the specified watcher (which must be a pointer to an
2093	initialised but not necessarily started event watcher).
2094
2095	=item ev_feed_fd_event (ev_loop *, int fd, int revents)	3297	=item ev_feed_fd_event (loop, int fd, int revents)
2096		3298
2097	Feed an event on the given fd, as if a file descriptor backend detected	3299	Feed an event on the given fd, as if a file descriptor backend detected
2098	the given events it.	3300	the given events it.
2099		3301
2100	=item ev_feed_signal_event (ev_loop *loop, int signum)	3302	=item ev_feed_signal_event (loop, int signum)
2101		3303
2102	Feed an event as if the given signal occured (C<loop> must be the default	3304	Feed an event as if the given signal occurred (C<loop> must be the default
2103	loop!).	3305	loop!).
2104		3306
2105	=back	3307	=back
2106		3308
2107		3309
…		…
2123		3325
2124	=item * Priorities are not currently supported. Initialising priorities	3326	=item * Priorities are not currently supported. Initialising priorities
2125	will fail and all watchers will have the same priority, even though there	3327	will fail and all watchers will have the same priority, even though there
2126	is an ev_pri field.	3328	is an ev_pri field.
2127		3329
		3330	=item * In libevent, the last base created gets the signals, in libev, the
		3331	first base created (== the default loop) gets the signals.
		3332
2128	=item * Other members are not supported.	3333	=item * Other members are not supported.
2129		3334
2130	=item * The libev emulation is I<not> ABI compatible to libevent, you need	3335	=item * The libev emulation is I<not> ABI compatible to libevent, you need
2131	to use the libev header file and library.	3336	to use the libev header file and library.
2132		3337
2133	=back	3338	=back
2134		3339
2135	=head1 C++ SUPPORT	3340	=head1 C++ SUPPORT
2136		3341
2137	Libev comes with some simplistic wrapper classes for C++ that mainly allow	3342	Libev comes with some simplistic wrapper classes for C++ that mainly allow
2138	you to use some convinience methods to start/stop watchers and also change	3343	you to use some convenience methods to start/stop watchers and also change
2139	the callback model to a model using method callbacks on objects.	3344	the callback model to a model using method callbacks on objects.
2140		3345
2141	To use it,	3346	To use it,
2142		3347
2143	#include <ev++.h>	3348	#include <ev++.h>
2144		3349
2145	This automatically includes F<ev.h> and puts all of its definitions (many	3350	This automatically includes F<ev.h> and puts all of its definitions (many
2146	of them macros) into the global namespace. All C++ specific things are	3351	of them macros) into the global namespace. All C++ specific things are
2147	put into the C<ev> namespace. It should support all the same embedding	3352	put into the C<ev> namespace. It should support all the same embedding
2148	options as F<ev.h>, most notably C<EV_MULTIPLICITY>.	3353	options as F<ev.h>, most notably C<EV_MULTIPLICITY>.
…		…
2182		3387
2183	=over 4	3388	=over 4
2184		3389
2185	=item ev::TYPE::TYPE ()	3390	=item ev::TYPE::TYPE ()
2186		3391
2187	=item ev::TYPE::TYPE (struct ev_loop *)	3392	=item ev::TYPE::TYPE (loop)
2188		3393
2189	=item ev::TYPE::~TYPE	3394	=item ev::TYPE::~TYPE
2190		3395
2191	The constructor (optionally) takes an event loop to associate the watcher	3396	The constructor (optionally) takes an event loop to associate the watcher
2192	with. If it is omitted, it will use C<EV_DEFAULT>.	3397	with. If it is omitted, it will use C<EV_DEFAULT>.
…		…
2215	your compiler is good :), then the method will be fully inlined into the	3420	your compiler is good :), then the method will be fully inlined into the
2216	thunking function, making it as fast as a direct C callback.	3421	thunking function, making it as fast as a direct C callback.
2217		3422
2218	Example: simple class declaration and watcher initialisation	3423	Example: simple class declaration and watcher initialisation
2219		3424
2220	struct myclass	3425	struct myclass
2221	{	3426	{
2222	void io_cb (ev::io &w, int revents) { }	3427	void io_cb (ev::io &w, int revents) { }
2223	}	3428	}
2224		3429
2225	myclass obj;	3430	myclass obj;
2226	ev::io iow;	3431	ev::io iow;
2227	iow.set <myclass, &myclass::io_cb> (&obj);	3432	iow.set <myclass, &myclass::io_cb> (&obj);
		3433
		3434	=item w->set (object *)
		3435
		3436	This is a variation of a method callback - leaving out the method to call
		3437	will default the method to C<operator ()>, which makes it possible to use
		3438	functor objects without having to manually specify the C<operator ()> all
		3439	the time. Incidentally, you can then also leave out the template argument
		3440	list.
		3441
		3442	The C<operator ()> method prototype must be C<void operator ()(watcher &w,
		3443	int revents)>.
		3444
		3445	See the method-C<set> above for more details.
		3446
		3447	Example: use a functor object as callback.
		3448
		3449	struct myfunctor
		3450	{
		3451	void operator() (ev::io &w, int revents)
		3452	{
		3453	...
		3454	}
		3455	}
		3456
		3457	myfunctor f;
		3458
		3459	ev::io w;
		3460	w.set (&f);
2228		3461
2229	=item w->set<function> (void *data = 0)	3462	=item w->set<function> (void *data = 0)
2230		3463
2231	Also sets a callback, but uses a static method or plain function as	3464	Also sets a callback, but uses a static method or plain function as
2232	callback. The optional C<data> argument will be stored in the watcher's	3465	callback. The optional C<data> argument will be stored in the watcher's
…		…
2234		3467
2235	The prototype of the C<function> must be C<void (*)(ev::TYPE &w, int)>.	3468	The prototype of the C<function> must be C<void (*)(ev::TYPE &w, int)>.
2236		3469
2237	See the method-C<set> above for more details.	3470	See the method-C<set> above for more details.
2238		3471
2239	Example:	3472	Example: Use a plain function as callback.
2240		3473
2241	static void io_cb (ev::io &w, int revents) { }	3474	static void io_cb (ev::io &w, int revents) { }
2242	iow.set <io_cb> ();	3475	iow.set <io_cb> ();
2243		3476
2244	=item w->set (struct ev_loop *)	3477	=item w->set (loop)
2245		3478
2246	Associates a different C<struct ev_loop> with this watcher. You can only	3479	Associates a different C<struct ev_loop> with this watcher. You can only
2247	do this when the watcher is inactive (and not pending either).	3480	do this when the watcher is inactive (and not pending either).
2248		3481
2249	=item w->set ([args])	3482	=item w->set ([arguments])
2250		3483
2251	Basically the same as C<ev_TYPE_set>, with the same args. Must be	3484	Basically the same as C<ev_TYPE_set>, with the same arguments. Either this
2252	called at least once. Unlike the C counterpart, an active watcher gets	3485	method or a suitable start method must be called at least once. Unlike the
2253	automatically stopped and restarted when reconfiguring it with this	3486	C counterpart, an active watcher gets automatically stopped and restarted
2254	method.	3487	when reconfiguring it with this method.
2255		3488
2256	=item w->start ()	3489	=item w->start ()
2257		3490
2258	Starts the watcher. Note that there is no C<loop> argument, as the	3491	Starts the watcher. Note that there is no C<loop> argument, as the
2259	constructor already stores the event loop.	3492	constructor already stores the event loop.
2260		3493
		3494	=item w->start ([arguments])
		3495
		3496	Instead of calling C<set> and C<start> methods separately, it is often
		3497	convenient to wrap them in one call. Uses the same type of arguments as
		3498	the configure C<set> method of the watcher.
		3499
2261	=item w->stop ()	3500	=item w->stop ()
2262		3501
2263	Stops the watcher if it is active. Again, no C<loop> argument.	3502	Stops the watcher if it is active. Again, no C<loop> argument.
2264		3503
2265	=item w->again () (C<ev::timer>, C<ev::periodic> only)	3504	=item w->again () (C<ev::timer>, C<ev::periodic> only)
…		…
2277		3516
2278	=back	3517	=back
2279		3518
2280	=back	3519	=back
2281		3520
2282	Example: Define a class with an IO and idle watcher, start one of them in	3521	Example: Define a class with two I/O and idle watchers, start the I/O
2283	the constructor.	3522	watchers in the constructor.
2284		3523
2285	class myclass	3524	class myclass
2286	{	3525	{
2287	ev::io io; void io_cb (ev::io &w, int revents);	3526	ev::io io ; void io_cb (ev::io &w, int revents);
		3527	ev::io2 io2 ; void io2_cb (ev::io &w, int revents);
2288	ev:idle idle void idle_cb (ev::idle &w, int revents);	3528	ev::idle idle; void idle_cb (ev::idle &w, int revents);
2289		3529
2290	myclass (int fd)	3530	myclass (int fd)
2291	{	3531	{
2292	io .set <myclass, &myclass::io_cb > (this);	3532	io .set <myclass, &myclass::io_cb > (this);
		3533	io2 .set <myclass, &myclass::io2_cb > (this);
2293	idle.set <myclass, &myclass::idle_cb> (this);	3534	idle.set <myclass, &myclass::idle_cb> (this);
2294		3535
2295	io.start (fd, ev::READ);	3536	io.set (fd, ev::WRITE); // configure the watcher
		3537	io.start (); // start it whenever convenient
		3538
		3539	io2.start (fd, ev::READ); // set + start in one call
2296	}	3540	}
2297	};	3541	};
		3542
		3543
		3544	=head1 OTHER LANGUAGE BINDINGS
		3545
		3546	Libev does not offer other language bindings itself, but bindings for a
		3547	number of languages exist in the form of third-party packages. If you know
		3548	any interesting language binding in addition to the ones listed here, drop
		3549	me a note.
		3550
		3551	=over 4
		3552
		3553	=item Perl
		3554
		3555	The EV module implements the full libev API and is actually used to test
		3556	libev. EV is developed together with libev. Apart from the EV core module,
		3557	there are additional modules that implement libev-compatible interfaces
		3558	to C<libadns> (C<EV::ADNS>, but C<AnyEvent::DNS> is preferred nowadays),
		3559	C<Net::SNMP> (C<Net::SNMP::EV>) and the C<libglib> event core (C<Glib::EV>
		3560	and C<EV::Glib>).
		3561
		3562	It can be found and installed via CPAN, its homepage is at
		3563	L<http://software.schmorp.de/pkg/EV>.
		3564
		3565	=item Python
		3566
		3567	Python bindings can be found at L<http://code.google.com/p/pyev/>. It
		3568	seems to be quite complete and well-documented.
		3569
		3570	=item Ruby
		3571
		3572	Tony Arcieri has written a ruby extension that offers access to a subset
		3573	of the libev API and adds file handle abstractions, asynchronous DNS and
		3574	more on top of it. It can be found via gem servers. Its homepage is at
		3575	L<http://rev.rubyforge.org/>.
		3576
		3577	Roger Pack reports that using the link order C<-lws2_32 -lmsvcrt-ruby-190>
		3578	makes rev work even on mingw.
		3579
		3580	=item Haskell
		3581
		3582	A haskell binding to libev is available at
		3583	L<http://hackage.haskell.org/cgi-bin/hackage-scripts/package/hlibev>.
		3584
		3585	=item D
		3586
		3587	Leandro Lucarella has written a D language binding (F<ev.d>) for libev, to
		3588	be found at L<http://proj.llucax.com.ar/wiki/evd>.
		3589
		3590	=item Ocaml
		3591
		3592	Erkki Seppala has written Ocaml bindings for libev, to be found at
		3593	L<http://modeemi.cs.tut.fi/~flux/software/ocaml-ev/>.
		3594
		3595	=item Lua
		3596
		3597	Brian Maher has written a partial interface to libev for lua (at the
		3598	time of this writing, only C<ev_io> and C<ev_timer>), to be found at
		3599	L<http://github.com/brimworks/lua-ev>.
		3600
		3601	=back
2298		3602
2299		3603
2300	=head1 MACRO MAGIC	3604	=head1 MACRO MAGIC
2301		3605
2302	Libev can be compiled with a variety of options, the most fundamantal	3606	Libev can be compiled with a variety of options, the most fundamental
2303	of which is C<EV_MULTIPLICITY>. This option determines whether (most)	3607	of which is C<EV_MULTIPLICITY>. This option determines whether (most)
2304	functions and callbacks have an initial C<struct ev_loop *> argument.	3608	functions and callbacks have an initial C<struct ev_loop *> argument.
2305		3609
2306	To make it easier to write programs that cope with either variant, the	3610	To make it easier to write programs that cope with either variant, the
2307	following macros are defined:	3611	following macros are defined:
…		…
2312		3616
2313	This provides the loop I<argument> for functions, if one is required ("ev	3617	This provides the loop I<argument> for functions, if one is required ("ev
2314	loop argument"). The C<EV_A> form is used when this is the sole argument,	3618	loop argument"). The C<EV_A> form is used when this is the sole argument,
2315	C<EV_A_> is used when other arguments are following. Example:	3619	C<EV_A_> is used when other arguments are following. Example:
2316		3620
2317	ev_unref (EV_A);	3621	ev_unref (EV_A);
2318	ev_timer_add (EV_A_ watcher);	3622	ev_timer_add (EV_A_ watcher);
2319	ev_loop (EV_A_ 0);	3623	ev_run (EV_A_ 0);
2320		3624
2321	It assumes the variable C<loop> of type C<struct ev_loop *> is in scope,	3625	It assumes the variable C<loop> of type C<struct ev_loop *> is in scope,
2322	which is often provided by the following macro.	3626	which is often provided by the following macro.
2323		3627
2324	=item C<EV_P>, C<EV_P_>	3628	=item C<EV_P>, C<EV_P_>
2325		3629
2326	This provides the loop I<parameter> for functions, if one is required ("ev	3630	This provides the loop I<parameter> for functions, if one is required ("ev
2327	loop parameter"). The C<EV_P> form is used when this is the sole parameter,	3631	loop parameter"). The C<EV_P> form is used when this is the sole parameter,
2328	C<EV_P_> is used when other parameters are following. Example:	3632	C<EV_P_> is used when other parameters are following. Example:
2329		3633
2330	// this is how ev_unref is being declared	3634	// this is how ev_unref is being declared
2331	static void ev_unref (EV_P);	3635	static void ev_unref (EV_P);
2332		3636
2333	// this is how you can declare your typical callback	3637	// this is how you can declare your typical callback
2334	static void cb (EV_P_ ev_timer *w, int revents)	3638	static void cb (EV_P_ ev_timer *w, int revents)
2335		3639
2336	It declares a parameter C<loop> of type C<struct ev_loop *>, quite	3640	It declares a parameter C<loop> of type C<struct ev_loop *>, quite
2337	suitable for use with C<EV_A>.	3641	suitable for use with C<EV_A>.
2338		3642
2339	=item C<EV_DEFAULT>, C<EV_DEFAULT_>	3643	=item C<EV_DEFAULT>, C<EV_DEFAULT_>
2340		3644
2341	Similar to the other two macros, this gives you the value of the default	3645	Similar to the other two macros, this gives you the value of the default
2342	loop, if multiple loops are supported ("ev loop default").	3646	loop, if multiple loops are supported ("ev loop default").
		3647
		3648	=item C<EV_DEFAULT_UC>, C<EV_DEFAULT_UC_>
		3649
		3650	Usage identical to C<EV_DEFAULT> and C<EV_DEFAULT_>, but requires that the
		3651	default loop has been initialised (C<UC> == unchecked). Their behaviour
		3652	is undefined when the default loop has not been initialised by a previous
		3653	execution of C<EV_DEFAULT>, C<EV_DEFAULT_> or C<ev_default_init (...)>.
		3654
		3655	It is often prudent to use C<EV_DEFAULT> when initialising the first
		3656	watcher in a function but use C<EV_DEFAULT_UC> afterwards.
2343		3657
2344	=back	3658	=back
2345		3659
2346	Example: Declare and initialise a check watcher, utilising the above	3660	Example: Declare and initialise a check watcher, utilising the above
2347	macros so it will work regardless of whether multiple loops are supported	3661	macros so it will work regardless of whether multiple loops are supported
2348	or not.	3662	or not.
2349		3663
2350	static void	3664	static void
2351	check_cb (EV_P_ ev_timer *w, int revents)	3665	check_cb (EV_P_ ev_timer *w, int revents)
2352	{	3666	{
2353	ev_check_stop (EV_A_ w);	3667	ev_check_stop (EV_A_ w);
2354	}	3668	}
2355		3669
2356	ev_check check;	3670	ev_check check;
2357	ev_check_init (&check, check_cb);	3671	ev_check_init (&check, check_cb);
2358	ev_check_start (EV_DEFAULT_ &check);	3672	ev_check_start (EV_DEFAULT_ &check);
2359	ev_loop (EV_DEFAULT_ 0);	3673	ev_run (EV_DEFAULT_ 0);
2360		3674
2361	=head1 EMBEDDING	3675	=head1 EMBEDDING
2362		3676
2363	Libev can (and often is) directly embedded into host	3677	Libev can (and often is) directly embedded into host
2364	applications. Examples of applications that embed it include the Deliantra	3678	applications. Examples of applications that embed it include the Deliantra
…		…
2371	libev somewhere in your source tree).	3685	libev somewhere in your source tree).
2372		3686
2373	=head2 FILESETS	3687	=head2 FILESETS
2374		3688
2375	Depending on what features you need you need to include one or more sets of files	3689	Depending on what features you need you need to include one or more sets of files
2376	in your app.	3690	in your application.
2377		3691
2378	=head3 CORE EVENT LOOP	3692	=head3 CORE EVENT LOOP
2379		3693
2380	To include only the libev core (all the C<ev_*> functions), with manual	3694	To include only the libev core (all the C<ev_*> functions), with manual
2381	configuration (no autoconf):	3695	configuration (no autoconf):
2382		3696
2383	#define EV_STANDALONE 1	3697	#define EV_STANDALONE 1
2384	#include "ev.c"	3698	#include "ev.c"
2385		3699
2386	This will automatically include F<ev.h>, too, and should be done in a	3700	This will automatically include F<ev.h>, too, and should be done in a
2387	single C source file only to provide the function implementations. To use	3701	single C source file only to provide the function implementations. To use
2388	it, do the same for F<ev.h> in all files wishing to use this API (best	3702	it, do the same for F<ev.h> in all files wishing to use this API (best
2389	done by writing a wrapper around F<ev.h> that you can include instead and	3703	done by writing a wrapper around F<ev.h> that you can include instead and
2390	where you can put other configuration options):	3704	where you can put other configuration options):
2391		3705
2392	#define EV_STANDALONE 1	3706	#define EV_STANDALONE 1
2393	#include "ev.h"	3707	#include "ev.h"
2394		3708
2395	Both header files and implementation files can be compiled with a C++	3709	Both header files and implementation files can be compiled with a C++
2396	compiler (at least, thats a stated goal, and breakage will be treated	3710	compiler (at least, that's a stated goal, and breakage will be treated
2397	as a bug).	3711	as a bug).
2398		3712
2399	You need the following files in your source tree, or in a directory	3713	You need the following files in your source tree, or in a directory
2400	in your include path (e.g. in libev/ when using -Ilibev):	3714	in your include path (e.g. in libev/ when using -Ilibev):
2401		3715
2402	ev.h	3716	ev.h
2403	ev.c	3717	ev.c
2404	ev_vars.h	3718	ev_vars.h
2405	ev_wrap.h	3719	ev_wrap.h
2406		3720
2407	ev_win32.c required on win32 platforms only	3721	ev_win32.c required on win32 platforms only
2408		3722
2409	ev_select.c only when select backend is enabled (which is enabled by default)	3723	ev_select.c only when select backend is enabled (which is enabled by default)
2410	ev_poll.c only when poll backend is enabled (disabled by default)	3724	ev_poll.c only when poll backend is enabled (disabled by default)
2411	ev_epoll.c only when the epoll backend is enabled (disabled by default)	3725	ev_epoll.c only when the epoll backend is enabled (disabled by default)
2412	ev_kqueue.c only when the kqueue backend is enabled (disabled by default)	3726	ev_kqueue.c only when the kqueue backend is enabled (disabled by default)
2413	ev_port.c only when the solaris port backend is enabled (disabled by default)	3727	ev_port.c only when the solaris port backend is enabled (disabled by default)
2414		3728
2415	F<ev.c> includes the backend files directly when enabled, so you only need	3729	F<ev.c> includes the backend files directly when enabled, so you only need
2416	to compile this single file.	3730	to compile this single file.
2417		3731
2418	=head3 LIBEVENT COMPATIBILITY API	3732	=head3 LIBEVENT COMPATIBILITY API
2419		3733
2420	To include the libevent compatibility API, also include:	3734	To include the libevent compatibility API, also include:
2421		3735
2422	#include "event.c"	3736	#include "event.c"
2423		3737
2424	in the file including F<ev.c>, and:	3738	in the file including F<ev.c>, and:
2425		3739
2426	#include "event.h"	3740	#include "event.h"
2427		3741
2428	in the files that want to use the libevent API. This also includes F<ev.h>.	3742	in the files that want to use the libevent API. This also includes F<ev.h>.
2429		3743
2430	You need the following additional files for this:	3744	You need the following additional files for this:
2431		3745
2432	event.h	3746	event.h
2433	event.c	3747	event.c
2434		3748
2435	=head3 AUTOCONF SUPPORT	3749	=head3 AUTOCONF SUPPORT
2436		3750
2437	Instead of using C<EV_STANDALONE=1> and providing your config in	3751	Instead of using C<EV_STANDALONE=1> and providing your configuration in
2438	whatever way you want, you can also C<m4_include([libev.m4])> in your	3752	whatever way you want, you can also C<m4_include([libev.m4])> in your
2439	F<configure.ac> and leave C<EV_STANDALONE> undefined. F<ev.c> will then	3753	F<configure.ac> and leave C<EV_STANDALONE> undefined. F<ev.c> will then
2440	include F<config.h> and configure itself accordingly.	3754	include F<config.h> and configure itself accordingly.
2441		3755
2442	For this of course you need the m4 file:	3756	For this of course you need the m4 file:
2443		3757
2444	libev.m4	3758	libev.m4
2445		3759
2446	=head2 PREPROCESSOR SYMBOLS/MACROS	3760	=head2 PREPROCESSOR SYMBOLS/MACROS
2447		3761
2448	Libev can be configured via a variety of preprocessor symbols you have to define	3762	Libev can be configured via a variety of preprocessor symbols you have to
2449	before including any of its files. The default is not to build for multiplicity	3763	define before including (or compiling) any of its files. The default in
2450	and only include the select backend.	3764	the absence of autoconf is documented for every option.
		3765
		3766	Symbols marked with "(h)" do not change the ABI, and can have different
		3767	values when compiling libev vs. including F<ev.h>, so it is permissible
		3768	to redefine them before including F<ev.h> without breaking compatibility
		3769	to a compiled library. All other symbols change the ABI, which means all
		3770	users of libev and the libev code itself must be compiled with compatible
		3771	settings.
2451		3772
2452	=over 4	3773	=over 4
2453		3774
		3775	=item EV_COMPAT3 (h)
		3776
		3777	Backwards compatibility is a major concern for libev. This is why this
		3778	release of libev comes with wrappers for the functions and symbols that
		3779	have been renamed between libev version 3 and 4.
		3780
		3781	You can disable these wrappers (to test compatibility with future
		3782	versions) by defining C<EV_COMPAT3> to C<0> when compiling your
		3783	sources. This has the additional advantage that you can drop the C<struct>
		3784	from C<struct ev_loop> declarations, as libev will provide an C<ev_loop>
		3785	typedef in that case.
		3786
		3787	In some future version, the default for C<EV_COMPAT3> will become C<0>,
		3788	and in some even more future version the compatibility code will be
		3789	removed completely.
		3790
2454	=item EV_STANDALONE	3791	=item EV_STANDALONE (h)
2455		3792
2456	Must always be C<1> if you do not use autoconf configuration, which	3793	Must always be C<1> if you do not use autoconf configuration, which
2457	keeps libev from including F<config.h>, and it also defines dummy	3794	keeps libev from including F<config.h>, and it also defines dummy
2458	implementations for some libevent functions (such as logging, which is not	3795	implementations for some libevent functions (such as logging, which is not
2459	supported). It will also not define any of the structs usually found in	3796	supported). It will also not define any of the structs usually found in
2460	F<event.h> that are not directly supported by the libev core alone.	3797	F<event.h> that are not directly supported by the libev core alone.
2461		3798
		3799	In standalone mode, libev will still try to automatically deduce the
		3800	configuration, but has to be more conservative.
		3801
2462	=item EV_USE_MONOTONIC	3802	=item EV_USE_MONOTONIC
2463		3803
2464	If defined to be C<1>, libev will try to detect the availability of the	3804	If defined to be C<1>, libev will try to detect the availability of the
2465	monotonic clock option at both compiletime and runtime. Otherwise no use	3805	monotonic clock option at both compile time and runtime. Otherwise no
2466	of the monotonic clock option will be attempted. If you enable this, you	3806	use of the monotonic clock option will be attempted. If you enable this,
2467	usually have to link against librt or something similar. Enabling it when	3807	you usually have to link against librt or something similar. Enabling it
2468	the functionality isn't available is safe, though, although you have	3808	when the functionality isn't available is safe, though, although you have
2469	to make sure you link against any libraries where the C<clock_gettime>	3809	to make sure you link against any libraries where the C<clock_gettime>
2470	function is hiding in (often F<-lrt>).	3810	function is hiding in (often F<-lrt>). See also C<EV_USE_CLOCK_SYSCALL>.
2471		3811
2472	=item EV_USE_REALTIME	3812	=item EV_USE_REALTIME
2473		3813
2474	If defined to be C<1>, libev will try to detect the availability of the	3814	If defined to be C<1>, libev will try to detect the availability of the
2475	realtime clock option at compiletime (and assume its availability at	3815	real-time clock option at compile time (and assume its availability
2476	runtime if successful). Otherwise no use of the realtime clock option will	3816	at runtime if successful). Otherwise no use of the real-time clock
2477	be attempted. This effectively replaces C<gettimeofday> by C<clock_get	3817	option will be attempted. This effectively replaces C<gettimeofday>
2478	(CLOCK_REALTIME, ...)> and will not normally affect correctness. See the	3818	by C<clock_get (CLOCK_REALTIME, ...)> and will not normally affect
2479	note about libraries in the description of C<EV_USE_MONOTONIC>, though.	3819	correctness. See the note about libraries in the description of
		3820	C<EV_USE_MONOTONIC>, though. Defaults to the opposite value of
		3821	C<EV_USE_CLOCK_SYSCALL>.
		3822
		3823	=item EV_USE_CLOCK_SYSCALL
		3824
		3825	If defined to be C<1>, libev will try to use a direct syscall instead
		3826	of calling the system-provided C<clock_gettime> function. This option
		3827	exists because on GNU/Linux, C<clock_gettime> is in C<librt>, but C<librt>
		3828	unconditionally pulls in C<libpthread>, slowing down single-threaded
		3829	programs needlessly. Using a direct syscall is slightly slower (in
		3830	theory), because no optimised vdso implementation can be used, but avoids
		3831	the pthread dependency. Defaults to C<1> on GNU/Linux with glibc 2.x or
		3832	higher, as it simplifies linking (no need for C<-lrt>).
2480		3833
2481	=item EV_USE_NANOSLEEP	3834	=item EV_USE_NANOSLEEP
2482		3835
2483	If defined to be C<1>, libev will assume that C<nanosleep ()> is available	3836	If defined to be C<1>, libev will assume that C<nanosleep ()> is available
2484	and will use it for delays. Otherwise it will use C<select ()>.	3837	and will use it for delays. Otherwise it will use C<select ()>.
2485		3838
		3839	=item EV_USE_EVENTFD
		3840
		3841	If defined to be C<1>, then libev will assume that C<eventfd ()> is
		3842	available and will probe for kernel support at runtime. This will improve
		3843	C<ev_signal> and C<ev_async> performance and reduce resource consumption.
		3844	If undefined, it will be enabled if the headers indicate GNU/Linux + Glibc
		3845	2.7 or newer, otherwise disabled.
		3846
2486	=item EV_USE_SELECT	3847	=item EV_USE_SELECT
2487		3848
2488	If undefined or defined to be C<1>, libev will compile in support for the	3849	If undefined or defined to be C<1>, libev will compile in support for the
2489	C<select>(2) backend. No attempt at autodetection will be done: if no	3850	C<select>(2) backend. No attempt at auto-detection will be done: if no
2490	other method takes over, select will be it. Otherwise the select backend	3851	other method takes over, select will be it. Otherwise the select backend
2491	will not be compiled in.	3852	will not be compiled in.
2492		3853
2493	=item EV_SELECT_USE_FD_SET	3854	=item EV_SELECT_USE_FD_SET
2494		3855
2495	If defined to C<1>, then the select backend will use the system C<fd_set>	3856	If defined to C<1>, then the select backend will use the system C<fd_set>
2496	structure. This is useful if libev doesn't compile due to a missing	3857	structure. This is useful if libev doesn't compile due to a missing
2497	C<NFDBITS> or C<fd_mask> definition or it misguesses the bitset layout on	3858	C<NFDBITS> or C<fd_mask> definition or it mis-guesses the bitset layout
2498	exotic systems. This usually limits the range of file descriptors to some	3859	on exotic systems. This usually limits the range of file descriptors to
2499	low limit such as 1024 or might have other limitations (winsocket only	3860	some low limit such as 1024 or might have other limitations (winsocket
2500	allows 64 sockets). The C<FD_SETSIZE> macro, set before compilation, might	3861	only allows 64 sockets). The C<FD_SETSIZE> macro, set before compilation,
2501	influence the size of the C<fd_set> used.	3862	configures the maximum size of the C<fd_set>.
2502		3863
2503	=item EV_SELECT_IS_WINSOCKET	3864	=item EV_SELECT_IS_WINSOCKET
2504		3865
2505	When defined to C<1>, the select backend will assume that	3866	When defined to C<1>, the select backend will assume that
2506	select/socket/connect etc. don't understand file descriptors but	3867	select/socket/connect etc. don't understand file descriptors but
…		…
2508	be used is the winsock select). This means that it will call	3869	be used is the winsock select). This means that it will call
2509	C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,	3870	C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,
2510	it is assumed that all these functions actually work on fds, even	3871	it is assumed that all these functions actually work on fds, even
2511	on win32. Should not be defined on non-win32 platforms.	3872	on win32. Should not be defined on non-win32 platforms.
2512		3873
2513	=item EV_FD_TO_WIN32_HANDLE	3874	=item EV_FD_TO_WIN32_HANDLE(fd)
2514		3875
2515	If C<EV_SELECT_IS_WINSOCKET> is enabled, then libev needs a way to map	3876	If C<EV_SELECT_IS_WINSOCKET> is enabled, then libev needs a way to map
2516	file descriptors to socket handles. When not defining this symbol (the	3877	file descriptors to socket handles. When not defining this symbol (the
2517	default), then libev will call C<_get_osfhandle>, which is usually	3878	default), then libev will call C<_get_osfhandle>, which is usually
2518	correct. In some cases, programs use their own file descriptor management,	3879	correct. In some cases, programs use their own file descriptor management,
2519	in which case they can provide this function to map fds to socket handles.	3880	in which case they can provide this function to map fds to socket handles.
2520		3881
		3882	=item EV_WIN32_HANDLE_TO_FD(handle)
		3883
		3884	If C<EV_SELECT_IS_WINSOCKET> then libev maps handles to file descriptors
		3885	using the standard C<_open_osfhandle> function. For programs implementing
		3886	their own fd to handle mapping, overwriting this function makes it easier
		3887	to do so. This can be done by defining this macro to an appropriate value.
		3888
		3889	=item EV_WIN32_CLOSE_FD(fd)
		3890
		3891	If programs implement their own fd to handle mapping on win32, then this
		3892	macro can be used to override the C<close> function, useful to unregister
		3893	file descriptors again. Note that the replacement function has to close
		3894	the underlying OS handle.
		3895
2521	=item EV_USE_POLL	3896	=item EV_USE_POLL
2522		3897
2523	If defined to be C<1>, libev will compile in support for the C<poll>(2)	3898	If defined to be C<1>, libev will compile in support for the C<poll>(2)
2524	backend. Otherwise it will be enabled on non-win32 platforms. It	3899	backend. Otherwise it will be enabled on non-win32 platforms. It
2525	takes precedence over select.	3900	takes precedence over select.
2526		3901
2527	=item EV_USE_EPOLL	3902	=item EV_USE_EPOLL
2528		3903
2529	If defined to be C<1>, libev will compile in support for the Linux	3904	If defined to be C<1>, libev will compile in support for the Linux
2530	C<epoll>(7) backend. Its availability will be detected at runtime,	3905	C<epoll>(7) backend. Its availability will be detected at runtime,
2531	otherwise another method will be used as fallback. This is the	3906	otherwise another method will be used as fallback. This is the preferred
2532	preferred backend for GNU/Linux systems.	3907	backend for GNU/Linux systems. If undefined, it will be enabled if the
		3908	headers indicate GNU/Linux + Glibc 2.4 or newer, otherwise disabled.
2533		3909
2534	=item EV_USE_KQUEUE	3910	=item EV_USE_KQUEUE
2535		3911
2536	If defined to be C<1>, libev will compile in support for the BSD style	3912	If defined to be C<1>, libev will compile in support for the BSD style
2537	C<kqueue>(2) backend. Its actual availability will be detected at runtime,	3913	C<kqueue>(2) backend. Its actual availability will be detected at runtime,
…		…
2550	otherwise another method will be used as fallback. This is the preferred	3926	otherwise another method will be used as fallback. This is the preferred
2551	backend for Solaris 10 systems.	3927	backend for Solaris 10 systems.
2552		3928
2553	=item EV_USE_DEVPOLL	3929	=item EV_USE_DEVPOLL
2554		3930
2555	reserved for future expansion, works like the USE symbols above.	3931	Reserved for future expansion, works like the USE symbols above.
2556		3932
2557	=item EV_USE_INOTIFY	3933	=item EV_USE_INOTIFY
2558		3934
2559	If defined to be C<1>, libev will compile in support for the Linux inotify	3935	If defined to be C<1>, libev will compile in support for the Linux inotify
2560	interface to speed up C<ev_stat> watchers. Its actual availability will	3936	interface to speed up C<ev_stat> watchers. Its actual availability will
2561	be detected at runtime.	3937	be detected at runtime. If undefined, it will be enabled if the headers
		3938	indicate GNU/Linux + Glibc 2.4 or newer, otherwise disabled.
2562		3939
		3940	=item EV_ATOMIC_T
		3941
		3942	Libev requires an integer type (suitable for storing C<0> or C<1>) whose
		3943	access is atomic with respect to other threads or signal contexts. No such
		3944	type is easily found in the C language, so you can provide your own type
		3945	that you know is safe for your purposes. It is used both for signal handler "locking"
		3946	as well as for signal and thread safety in C<ev_async> watchers.
		3947
		3948	In the absence of this define, libev will use C<sig_atomic_t volatile>
		3949	(from F<signal.h>), which is usually good enough on most platforms.
		3950
2563	=item EV_H	3951	=item EV_H (h)
2564		3952
2565	The name of the F<ev.h> header file used to include it. The default if	3953	The name of the F<ev.h> header file used to include it. The default if
2566	undefined is C<"ev.h"> in F<event.h>, F<ev.c> and F<ev++.h>. This can be	3954	undefined is C<"ev.h"> in F<event.h>, F<ev.c> and F<ev++.h>. This can be
2567	used to virtually rename the F<ev.h> header file in case of conflicts.	3955	used to virtually rename the F<ev.h> header file in case of conflicts.
2568		3956
2569	=item EV_CONFIG_H	3957	=item EV_CONFIG_H (h)
2570		3958
2571	If C<EV_STANDALONE> isn't C<1>, this variable can be used to override	3959	If C<EV_STANDALONE> isn't C<1>, this variable can be used to override
2572	F<ev.c>'s idea of where to find the F<config.h> file, similarly to	3960	F<ev.c>'s idea of where to find the F<config.h> file, similarly to
2573	C<EV_H>, above.	3961	C<EV_H>, above.
2574		3962
2575	=item EV_EVENT_H	3963	=item EV_EVENT_H (h)
2576		3964
2577	Similarly to C<EV_H>, this macro can be used to override F<event.c>'s idea	3965	Similarly to C<EV_H>, this macro can be used to override F<event.c>'s idea
2578	of how the F<event.h> header can be found, the default is C<"event.h">.	3966	of how the F<event.h> header can be found, the default is C<"event.h">.
2579		3967
2580	=item EV_PROTOTYPES	3968	=item EV_PROTOTYPES (h)
2581		3969
2582	If defined to be C<0>, then F<ev.h> will not define any function	3970	If defined to be C<0>, then F<ev.h> will not define any function
2583	prototypes, but still define all the structs and other symbols. This is	3971	prototypes, but still define all the structs and other symbols. This is
2584	occasionally useful if you want to provide your own wrapper functions	3972	occasionally useful if you want to provide your own wrapper functions
2585	around libev functions.	3973	around libev functions.
…		…
2604	When doing priority-based operations, libev usually has to linearly search	3992	When doing priority-based operations, libev usually has to linearly search
2605	all the priorities, so having many of them (hundreds) uses a lot of space	3993	all the priorities, so having many of them (hundreds) uses a lot of space
2606	and time, so using the defaults of five priorities (-2 .. +2) is usually	3994	and time, so using the defaults of five priorities (-2 .. +2) is usually
2607	fine.	3995	fine.
2608		3996
2609	If your embedding app does not need any priorities, defining these both to	3997	If your embedding application does not need any priorities, defining these
2610	C<0> will save some memory and cpu.	3998	both to C<0> will save some memory and CPU.
2611		3999
2612	=item EV_PERIODIC_ENABLE	4000	=item EV_PERIODIC_ENABLE, EV_IDLE_ENABLE, EV_EMBED_ENABLE, EV_STAT_ENABLE,
		4001	EV_PREPARE_ENABLE, EV_CHECK_ENABLE, EV_FORK_ENABLE, EV_SIGNAL_ENABLE,
		4002	EV_ASYNC_ENABLE, EV_CHILD_ENABLE.
2613		4003
2614	If undefined or defined to be C<1>, then periodic timers are supported. If	4004	If undefined or defined to be C<1> (and the platform supports it), then
2615	defined to be C<0>, then they are not. Disabling them saves a few kB of	4005	the respective watcher type is supported. If defined to be C<0>, then it
2616	code.	4006	is not. Disabling watcher types mainly saves code size.
2617		4007
2618	=item EV_IDLE_ENABLE	4008	=item EV_FEATURES
2619
2620	If undefined or defined to be C<1>, then idle watchers are supported. If
2621	defined to be C<0>, then they are not. Disabling them saves a few kB of
2622	code.
2623
2624	=item EV_EMBED_ENABLE
2625
2626	If undefined or defined to be C<1>, then embed watchers are supported. If
2627	defined to be C<0>, then they are not.
2628
2629	=item EV_STAT_ENABLE
2630
2631	If undefined or defined to be C<1>, then stat watchers are supported. If
2632	defined to be C<0>, then they are not.
2633
2634	=item EV_FORK_ENABLE
2635
2636	If undefined or defined to be C<1>, then fork watchers are supported. If
2637	defined to be C<0>, then they are not.
2638
2639	=item EV_MINIMAL
2640		4009
2641	If you need to shave off some kilobytes of code at the expense of some	4010	If you need to shave off some kilobytes of code at the expense of some
2642	speed, define this symbol to C<1>. Currently only used for gcc to override	4011	speed (but with the full API), you can define this symbol to request
2643	some inlining decisions, saves roughly 30% codesize of amd64.	4012	certain subsets of functionality. The default is to enable all features
		4013	that can be enabled on the platform.
		4014
		4015	A typical way to use this symbol is to define it to C<0> (or to a bitset
		4016	with some broad features you want) and then selectively re-enable
		4017	additional parts you want, for example if you want everything minimal,
		4018	but multiple event loop support, async and child watchers and the poll
		4019	backend, use this:
		4020
		4021	#define EV_FEATURES 0
		4022	#define EV_MULTIPLICITY 1
		4023	#define EV_USE_POLL 1
		4024	#define EV_CHILD_ENABLE 1
		4025	#define EV_ASYNC_ENABLE 1
		4026
		4027	The actual value is a bitset, it can be a combination of the following
		4028	values:
		4029
		4030	=over 4
		4031
		4032	=item C<1> - faster/larger code
		4033
		4034	Use larger code to speed up some operations.
		4035
		4036	Currently this is used to override some inlining decisions (enlarging the
		4037	code size by roughly 30% on amd64).
		4038
		4039	When optimising for size, use of compiler flags such as C<-Os> with
		4040	gcc is recommended, as well as C<-DNDEBUG>, as libev contains a number of
		4041	assertions.
		4042
		4043	=item C<2> - faster/larger data structures
		4044
		4045	Replaces the small 2-heap for timer management by a faster 4-heap, larger
		4046	hash table sizes and so on. This will usually further increase code size
		4047	and can additionally have an effect on the size of data structures at
		4048	runtime.
		4049
		4050	=item C<4> - full API configuration
		4051
		4052	This enables priorities (sets C<EV_MAXPRI>=2 and C<EV_MINPRI>=-2), and
		4053	enables multiplicity (C<EV_MULTIPLICITY>=1).
		4054
		4055	=item C<8> - full API
		4056
		4057	This enables a lot of the "lesser used" API functions. See C<ev.h> for
		4058	details on which parts of the API are still available without this
		4059	feature, and do not complain if this subset changes over time.
		4060
		4061	=item C<16> - enable all optional watcher types
		4062
		4063	Enables all optional watcher types. If you want to selectively enable
		4064	only some watcher types other than I/O and timers (e.g. prepare,
		4065	embed, async, child...) you can enable them manually by defining
		4066	C<EV_watchertype_ENABLE> to C<1> instead.
		4067
		4068	=item C<32> - enable all backends
		4069
		4070	This enables all backends - without this feature, you need to enable at
		4071	least one backend manually (C<EV_USE_SELECT> is a good choice).
		4072
		4073	=item C<64> - enable OS-specific "helper" APIs
		4074
		4075	Enable inotify, eventfd, signalfd and similar OS-specific helper APIs by
		4076	default.
		4077
		4078	=back
		4079
		4080	Compiling with C<gcc -Os -DEV_STANDALONE -DEV_USE_EPOLL=1 -DEV_FEATURES=0>
		4081	reduces the compiled size of libev from 24.7Kb code/2.8Kb data to 6.5Kb
		4082	code/0.3Kb data on my GNU/Linux amd64 system, while still giving you I/O
		4083	watchers, timers and monotonic clock support.
		4084
		4085	With an intelligent-enough linker (gcc+binutils are intelligent enough
		4086	when you use C<-Wl,--gc-sections -ffunction-sections>) functions unused by
		4087	your program might be left out as well - a binary starting a timer and an
		4088	I/O watcher then might come out at only 5Kb.
		4089
		4090	=item EV_AVOID_STDIO
		4091
		4092	If this is set to C<1> at compiletime, then libev will avoid using stdio
		4093	functions (printf, scanf, perror etc.). This will increase the code size
		4094	somewhat, but if your program doesn't otherwise depend on stdio and your
		4095	libc allows it, this avoids linking in the stdio library which is quite
		4096	big.
		4097
		4098	Note that error messages might become less precise when this option is
		4099	enabled.
		4100
		4101	=item EV_NSIG
		4102
		4103	The highest supported signal number, +1 (or, the number of
		4104	signals): Normally, libev tries to deduce the maximum number of signals
		4105	automatically, but sometimes this fails, in which case it can be
		4106	specified. Also, using a lower number than detected (C<32> should be
		4107	good for about any system in existence) can save some memory, as libev
		4108	statically allocates some 12-24 bytes per signal number.
2644		4109
2645	=item EV_PID_HASHSIZE	4110	=item EV_PID_HASHSIZE
2646		4111
2647	C<ev_child> watchers use a small hash table to distribute workload by	4112	C<ev_child> watchers use a small hash table to distribute workload by
2648	pid. The default size is C<16> (or C<1> with C<EV_MINIMAL>), usually more	4113	pid. The default size is C<16> (or C<1> with C<EV_FEATURES> disabled),
2649	than enough. If you need to manage thousands of children you might want to	4114	usually more than enough. If you need to manage thousands of children you
2650	increase this value (I<must> be a power of two).	4115	might want to increase this value (I<must> be a power of two).
2651		4116
2652	=item EV_INOTIFY_HASHSIZE	4117	=item EV_INOTIFY_HASHSIZE
2653		4118
2654	C<ev_stat> watchers use a small hash table to distribute workload by	4119	C<ev_stat> watchers use a small hash table to distribute workload by
2655	inotify watch id. The default size is C<16> (or C<1> with C<EV_MINIMAL>),	4120	inotify watch id. The default size is C<16> (or C<1> with C<EV_FEATURES>
2656	usually more than enough. If you need to manage thousands of C<ev_stat>	4121	disabled), usually more than enough. If you need to manage thousands of
2657	watchers you might want to increase this value (I<must> be a power of	4122	C<ev_stat> watchers you might want to increase this value (I<must> be a
2658	two).	4123	power of two).
		4124
		4125	=item EV_USE_4HEAP
		4126
		4127	Heaps are not very cache-efficient. To improve the cache-efficiency of the
		4128	timer and periodics heaps, libev uses a 4-heap when this symbol is defined
		4129	to C<1>. The 4-heap uses more complicated (longer) code but has noticeably
		4130	faster performance with many (thousands) of watchers.
		4131
		4132	The default is C<1>, unless C<EV_FEATURES> overrides it, in which case it
		4133	will be C<0>.
		4134
		4135	=item EV_HEAP_CACHE_AT
		4136
		4137	Heaps are not very cache-efficient. To improve the cache-efficiency of the
		4138	timer and periodics heaps, libev can cache the timestamp (I<at>) within
		4139	the heap structure (selected by defining C<EV_HEAP_CACHE_AT> to C<1>),
		4140	which uses 8-12 bytes more per watcher and a few hundred bytes more code,
		4141	but avoids random read accesses on heap changes. This improves performance
		4142	noticeably with many (hundreds) of watchers.
		4143
		4144	The default is C<1>, unless C<EV_FEATURES> overrides it, in which case it
		4145	will be C<0>.
		4146
		4147	=item EV_VERIFY
		4148
		4149	Controls how much internal verification (see C<ev_verify ()>) will
		4150	be done: If set to C<0>, no internal verification code will be compiled
		4151	in. If set to C<1>, then verification code will be compiled in, but not
		4152	called. If set to C<2>, then the internal verification code will be
		4153	called once per loop, which can slow down libev. If set to C<3>, then the
		4154	verification code will be called very frequently, which will slow down
		4155	libev considerably.
		4156
		4157	The default is C<1>, unless C<EV_FEATURES> overrides it, in which case it
		4158	will be C<0>.
2659		4159
2660	=item EV_COMMON	4160	=item EV_COMMON
2661		4161
2662	By default, all watchers have a C<void *data> member. By redefining	4162	By default, all watchers have a C<void *data> member. By redefining
2663	this macro to a something else you can include more and other types of	4163	this macro to something else you can include more and other types of
2664	members. You have to define it each time you include one of the files,	4164	members. You have to define it each time you include one of the files,
2665	though, and it must be identical each time.	4165	though, and it must be identical each time.
2666		4166
2667	For example, the perl EV module uses something like this:	4167	For example, the perl EV module uses something like this:
2668		4168
2669	#define EV_COMMON \	4169	#define EV_COMMON \
2670	SV *self; /* contains this struct */ \	4170	SV *self; /* contains this struct */ \
2671	SV *cb_sv, *fh /* note no trailing ";" */	4171	SV *cb_sv, *fh /* note no trailing ";" */
2672		4172
2673	=item EV_CB_DECLARE (type)	4173	=item EV_CB_DECLARE (type)
2674		4174
2675	=item EV_CB_INVOKE (watcher, revents)	4175	=item EV_CB_INVOKE (watcher, revents)
2676		4176
…		…
2681	definition and a statement, respectively. See the F<ev.h> header file for	4181	definition and a statement, respectively. See the F<ev.h> header file for
2682	their default definitions. One possible use for overriding these is to	4182	their default definitions. One possible use for overriding these is to
2683	avoid the C<struct ev_loop *> as first argument in all cases, or to use	4183	avoid the C<struct ev_loop *> as first argument in all cases, or to use
2684	method calls instead of plain function calls in C++.	4184	method calls instead of plain function calls in C++.
2685		4185
		4186	=back
		4187
2686	=head2 EXPORTED API SYMBOLS	4188	=head2 EXPORTED API SYMBOLS
2687		4189
2688	If you need to re-export the API (e.g. via a dll) and you need a list of	4190	If you need to re-export the API (e.g. via a DLL) and you need a list of
2689	exported symbols, you can use the provided F<Symbol.*> files which list	4191	exported symbols, you can use the provided F<Symbol.*> files which list
2690	all public symbols, one per line:	4192	all public symbols, one per line:
2691		4193
2692	Symbols.ev for libev proper	4194	Symbols.ev for libev proper
2693	Symbols.event for the libevent emulation	4195	Symbols.event for the libevent emulation
2694		4196
2695	This can also be used to rename all public symbols to avoid clashes with	4197	This can also be used to rename all public symbols to avoid clashes with
2696	multiple versions of libev linked together (which is obviously bad in	4198	multiple versions of libev linked together (which is obviously bad in
2697	itself, but sometimes it is inconvinient to avoid this).	4199	itself, but sometimes it is inconvenient to avoid this).
2698		4200
2699	A sed command like this will create wrapper C<#define>'s that you need to	4201	A sed command like this will create wrapper C<#define>'s that you need to
2700	include before including F<ev.h>:	4202	include before including F<ev.h>:
2701		4203
2702	<Symbols.ev sed -e "s/.*/#define & myprefix_&/" >wrap.h	4204	<Symbols.ev sed -e "s/.*/#define & myprefix_&/" >wrap.h
…		…
2719	file.	4221	file.
2720		4222
2721	The usage in rxvt-unicode is simpler. It has a F<ev_cpp.h> header file	4223	The usage in rxvt-unicode is simpler. It has a F<ev_cpp.h> header file
2722	that everybody includes and which overrides some configure choices:	4224	that everybody includes and which overrides some configure choices:
2723		4225
2724	#define EV_MINIMAL 1	4226	#define EV_FEATURES 8
2725	#define EV_USE_POLL 0	4227	#define EV_USE_SELECT 1
2726	#define EV_MULTIPLICITY 0
2727	#define EV_PERIODIC_ENABLE 0	4228	#define EV_PREPARE_ENABLE 1
		4229	#define EV_IDLE_ENABLE 1
2728	#define EV_STAT_ENABLE 0	4230	#define EV_SIGNAL_ENABLE 1
2729	#define EV_FORK_ENABLE 0	4231	#define EV_CHILD_ENABLE 1
		4232	#define EV_USE_STDEXCEPT 0
2730	#define EV_CONFIG_H <config.h>	4233	#define EV_CONFIG_H <config.h>
2731	#define EV_MINPRI 0
2732	#define EV_MAXPRI 0
2733		4234
2734	#include "ev++.h"	4235	#include "ev++.h"
2735		4236
2736	And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:	4237	And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:
2737		4238
2738	#include "ev_cpp.h"	4239	#include "ev_cpp.h"
2739	#include "ev.c"	4240	#include "ev.c"
2740		4241
		4242	=head1 INTERACTION WITH OTHER PROGRAMS OR LIBRARIES
2741		4243
2742	=head1 COMPLEXITIES	4244	=head2 THREADS AND COROUTINES
2743		4245
2744	In this section the complexities of (many of) the algorithms used inside	4246	=head3 THREADS
2745	libev will be explained. For complexity discussions about backends see the
2746	documentation for C<ev_default_init>.
2747		4247
2748	All of the following are about amortised time: If an array needs to be	4248	All libev functions are reentrant and thread-safe unless explicitly
2749	extended, libev needs to realloc and move the whole array, but this	4249	documented otherwise, but libev implements no locking itself. This means
2750	happens asymptotically never with higher number of elements, so O(1) might	4250	that you can use as many loops as you want in parallel, as long as there
2751	mean it might do a lengthy realloc operation in rare cases, but on average	4251	are no concurrent calls into any libev function with the same loop
2752	it is much faster and asymptotically approaches constant time.	4252	parameter (C<ev_default_*> calls have an implicit default loop parameter,
		4253	of course): libev guarantees that different event loops share no data
		4254	structures that need any locking.
		4255
		4256	Or to put it differently: calls with different loop parameters can be done
		4257	concurrently from multiple threads, calls with the same loop parameter
		4258	must be done serially (but can be done from different threads, as long as
		4259	only one thread ever is inside a call at any point in time, e.g. by using
		4260	a mutex per loop).
		4261
		4262	Specifically to support threads (and signal handlers), libev implements
		4263	so-called C<ev_async> watchers, which allow some limited form of
		4264	concurrency on the same event loop, namely waking it up "from the
		4265	outside".
		4266
		4267	If you want to know which design (one loop, locking, or multiple loops
		4268	without or something else still) is best for your problem, then I cannot
		4269	help you, but here is some generic advice:
2753		4270
2754	=over 4	4271	=over 4
2755		4272
2756	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)	4273	=item * most applications have a main thread: use the default libev loop
		4274	in that thread, or create a separate thread running only the default loop.
2757		4275
2758	This means that, when you have a watcher that triggers in one hour and	4276	This helps integrating other libraries or software modules that use libev
2759	there are 100 watchers that would trigger before that then inserting will	4277	themselves and don't care/know about threading.
2760	have to skip roughly seven (C<ld 100>) of these watchers.
2761		4278
2762	=item Changing timer/periodic watchers (by autorepeat or calling again): O(log skipped_other_timers)	4279	=item * one loop per thread is usually a good model.
2763		4280
2764	That means that changing a timer costs less than removing/adding them	4281	Doing this is almost never wrong, sometimes a better-performance model
2765	as only the relative motion in the event queue has to be paid for.	4282	exists, but it is always a good start.
2766		4283
2767	=item Starting io/check/prepare/idle/signal/child watchers: O(1)	4284	=item * other models exist, such as the leader/follower pattern, where one
		4285	loop is handed through multiple threads in a kind of round-robin fashion.
2768		4286
2769	These just add the watcher into an array or at the head of a list.	4287	Choosing a model is hard - look around, learn, know that usually you can do
		4288	better than you currently do :-)
2770		4289
2771	=item Stopping check/prepare/idle watchers: O(1)	4290	=item * often you need to talk to some other thread which blocks in the
		4291	event loop.
2772		4292
2773	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))	4293	C<ev_async> watchers can be used to wake them up from other threads safely
		4294	(or from signal contexts...).
2774		4295
2775	These watchers are stored in lists then need to be walked to find the	4296	An example use would be to communicate signals or other events that only
2776	correct watcher to remove. The lists are usually short (you don't usually	4297	work in the default loop by registering the signal watcher with the
2777	have many watchers waiting for the same fd or signal).	4298	default loop and triggering an C<ev_async> watcher from the default loop
2778		4299	watcher callback into the event loop interested in the signal.
2779	=item Finding the next timer in each loop iteration: O(1)
2780
2781	By virtue of using a binary heap, the next timer is always found at the
2782	beginning of the storage array.
2783
2784	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)
2785
2786	A change means an I/O watcher gets started or stopped, which requires
2787	libev to recalculate its status (and possibly tell the kernel, depending
2788	on backend and wether C<ev_io_set> was used).
2789
2790	=item Activating one watcher (putting it into the pending state): O(1)
2791
2792	=item Priority handling: O(number_of_priorities)
2793
2794	Priorities are implemented by allocating some space for each
2795	priority. When doing priority-based operations, libev usually has to
2796	linearly search all the priorities, but starting/stopping and activating
2797	watchers becomes O(1) w.r.t. prioritiy handling.
2798		4300
2799	=back	4301	=back
2800		4302
		4303	=head4 THREAD LOCKING EXAMPLE
2801		4304
2802	=head1 Win32 platform limitations and workarounds	4305	Here is a fictitious example of how to run an event loop in a different
		4306	thread than where callbacks are being invoked and watchers are
		4307	created/added/removed.
		4308
		4309	For a real-world example, see the C<EV::Loop::Async> perl module,
		4310	which uses exactly this technique (which is suited for many high-level
		4311	languages).
		4312
		4313	The example uses a pthread mutex to protect the loop data, a condition
		4314	variable to wait for callback invocations, an async watcher to notify the
		4315	event loop thread and an unspecified mechanism to wake up the main thread.
		4316
		4317	First, you need to associate some data with the event loop:
		4318
		4319	typedef struct {
		4320	mutex_t lock; /* global loop lock */
		4321	ev_async async_w;
		4322	thread_t tid;
		4323	cond_t invoke_cv;
		4324	} userdata;
		4325
		4326	void prepare_loop (EV_P)
		4327	{
		4328	// for simplicity, we use a static userdata struct.
		4329	static userdata u;
		4330
		4331	ev_async_init (&u->async_w, async_cb);
		4332	ev_async_start (EV_A_ &u->async_w);
		4333
		4334	pthread_mutex_init (&u->lock, 0);
		4335	pthread_cond_init (&u->invoke_cv, 0);
		4336
		4337	// now associate this with the loop
		4338	ev_set_userdata (EV_A_ u);
		4339	ev_set_invoke_pending_cb (EV_A_ l_invoke);
		4340	ev_set_loop_release_cb (EV_A_ l_release, l_acquire);
		4341
		4342	// then create the thread running ev_loop
		4343	pthread_create (&u->tid, 0, l_run, EV_A);
		4344	}
		4345
		4346	The callback for the C<ev_async> watcher does nothing: the watcher is used
		4347	solely to wake up the event loop so it takes notice of any new watchers
		4348	that might have been added:
		4349
		4350	static void
		4351	async_cb (EV_P_ ev_async *w, int revents)
		4352	{
		4353	// just used for the side effects
		4354	}
		4355
		4356	The C<l_release> and C<l_acquire> callbacks simply unlock/lock the mutex
		4357	protecting the loop data, respectively.
		4358
		4359	static void
		4360	l_release (EV_P)
		4361	{
		4362	userdata *u = ev_userdata (EV_A);
		4363	pthread_mutex_unlock (&u->lock);
		4364	}
		4365
		4366	static void
		4367	l_acquire (EV_P)
		4368	{
		4369	userdata *u = ev_userdata (EV_A);
		4370	pthread_mutex_lock (&u->lock);
		4371	}
		4372
		4373	The event loop thread first acquires the mutex, and then jumps straight
		4374	into C<ev_run>:
		4375
		4376	void *
		4377	l_run (void *thr_arg)
		4378	{
		4379	struct ev_loop *loop = (struct ev_loop *)thr_arg;
		4380
		4381	l_acquire (EV_A);
		4382	pthread_setcanceltype (PTHREAD_CANCEL_ASYNCHRONOUS, 0);
		4383	ev_run (EV_A_ 0);
		4384	l_release (EV_A);
		4385
		4386	return 0;
		4387	}
		4388
		4389	Instead of invoking all pending watchers, the C<l_invoke> callback will
		4390	signal the main thread via some unspecified mechanism (signals? pipe
		4391	writes? C<Async::Interrupt>?) and then waits until all pending watchers
		4392	have been called (in a while loop because a) spurious wakeups are possible
		4393	and b) skipping inter-thread-communication when there are no pending
		4394	watchers is very beneficial):
		4395
		4396	static void
		4397	l_invoke (EV_P)
		4398	{
		4399	userdata *u = ev_userdata (EV_A);
		4400
		4401	while (ev_pending_count (EV_A))
		4402	{
		4403	wake_up_other_thread_in_some_magic_or_not_so_magic_way ();
		4404	pthread_cond_wait (&u->invoke_cv, &u->lock);
		4405	}
		4406	}
		4407
		4408	Now, whenever the main thread gets told to invoke pending watchers, it
		4409	will grab the lock, call C<ev_invoke_pending> and then signal the loop
		4410	thread to continue:
		4411
		4412	static void
		4413	real_invoke_pending (EV_P)
		4414	{
		4415	userdata *u = ev_userdata (EV_A);
		4416
		4417	pthread_mutex_lock (&u->lock);
		4418	ev_invoke_pending (EV_A);
		4419	pthread_cond_signal (&u->invoke_cv);
		4420	pthread_mutex_unlock (&u->lock);
		4421	}
		4422
		4423	Whenever you want to start/stop a watcher or do other modifications to an
		4424	event loop, you will now have to lock:
		4425
		4426	ev_timer timeout_watcher;
		4427	userdata *u = ev_userdata (EV_A);
		4428
		4429	ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);
		4430
		4431	pthread_mutex_lock (&u->lock);
		4432	ev_timer_start (EV_A_ &timeout_watcher);
		4433	ev_async_send (EV_A_ &u->async_w);
		4434	pthread_mutex_unlock (&u->lock);
		4435
		4436	Note that sending the C<ev_async> watcher is required because otherwise
		4437	an event loop currently blocking in the kernel will have no knowledge
		4438	about the newly added timer. By waking up the loop it will pick up any new
		4439	watchers in the next event loop iteration.
		4440
		4441	=head3 COROUTINES
		4442
		4443	Libev is very accommodating to coroutines ("cooperative threads"):
		4444	libev fully supports nesting calls to its functions from different
		4445	coroutines (e.g. you can call C<ev_run> on the same loop from two
		4446	different coroutines, and switch freely between both coroutines running
		4447	the loop, as long as you don't confuse yourself). The only exception is
		4448	that you must not do this from C<ev_periodic> reschedule callbacks.
		4449
		4450	Care has been taken to ensure that libev does not keep local state inside
		4451	C<ev_run>, and other calls do not usually allow for coroutine switches as
		4452	they do not call any callbacks.
		4453
		4454	=head2 COMPILER WARNINGS
		4455
		4456	Depending on your compiler and compiler settings, you might get no or a
		4457	lot of warnings when compiling libev code. Some people are apparently
		4458	scared by this.
		4459
		4460	However, these are unavoidable for many reasons. For one, each compiler
		4461	has different warnings, and each user has different tastes regarding
		4462	warning options. "Warn-free" code therefore cannot be a goal except when
		4463	targeting a specific compiler and compiler-version.
		4464
		4465	Another reason is that some compiler warnings require elaborate
		4466	workarounds, or other changes to the code that make it less clear and less
		4467	maintainable.
		4468
		4469	And of course, some compiler warnings are just plain stupid, or simply
		4470	wrong (because they don't actually warn about the condition their message
		4471	seems to warn about). For example, certain older gcc versions had some
		4472	warnings that resulted in an extreme number of false positives. These have
		4473	been fixed, but some people still insist on making code warn-free with
		4474	such buggy versions.
		4475
		4476	While libev is written to generate as few warnings as possible,
		4477	"warn-free" code is not a goal, and it is recommended not to build libev
		4478	with any compiler warnings enabled unless you are prepared to cope with
		4479	them (e.g. by ignoring them). Remember that warnings are just that:
		4480	warnings, not errors, or proof of bugs.
		4481
		4482
		4483	=head2 VALGRIND
		4484
		4485	Valgrind has a special section here because it is a popular tool that is
		4486	highly useful. Unfortunately, valgrind reports are very hard to interpret.
		4487
		4488	If you think you found a bug (memory leak, uninitialised data access etc.)
		4489	in libev, then check twice: If valgrind reports something like:
		4490
		4491	==2274== definitely lost: 0 bytes in 0 blocks.
		4492	==2274== possibly lost: 0 bytes in 0 blocks.
		4493	==2274== still reachable: 256 bytes in 1 blocks.
		4494
		4495	Then there is no memory leak, just as memory accounted to global variables
		4496	is not a memleak - the memory is still being referenced, and didn't leak.
		4497
		4498	Similarly, under some circumstances, valgrind might report kernel bugs
		4499	as if it were a bug in libev (e.g. in realloc or in the poll backend,
		4500	although an acceptable workaround has been found here), or it might be
		4501	confused.
		4502
		4503	Keep in mind that valgrind is a very good tool, but only a tool. Don't
		4504	make it into some kind of religion.
		4505
		4506	If you are unsure about something, feel free to contact the mailing list
		4507	with the full valgrind report and an explanation on why you think this
		4508	is a bug in libev (best check the archives, too :). However, don't be
		4509	annoyed when you get a brisk "this is no bug" answer and take the chance
		4510	of learning how to interpret valgrind properly.
		4511
		4512	If you need, for some reason, empty reports from valgrind for your project
		4513	I suggest using suppression lists.
		4514
		4515
		4516	=head1 PORTABILITY NOTES
		4517
		4518	=head2 GNU/LINUX 32 BIT LIMITATIONS
		4519
		4520	GNU/Linux is the only common platform that supports 64 bit file/large file
		4521	interfaces but I<disables> them by default.
		4522
		4523	That means that libev compiled in the default environment doesn't support
		4524	files larger than 2GiB or so, which mainly affects C<ev_stat> watchers.
		4525
		4526	Unfortunately, many programs try to work around this GNU/Linux issue
		4527	by enabling the large file API, which makes them incompatible with the
		4528	standard libev compiled for their system.
		4529
		4530	Likewise, libev cannot enable the large file API itself as this would
		4531	suddenly make it incompatible to the default compile time environment,
		4532	i.e. all programs not using special compile switches.
		4533
		4534	=head2 OS/X AND DARWIN BUGS
		4535
		4536	The whole thing is a bug if you ask me - basically any system interface
		4537	you touch is broken, whether it is locales, poll, kqueue or even the
		4538	OpenGL drivers.
		4539
		4540	=head3 C<kqueue> is buggy
		4541
		4542	The kqueue syscall is broken in all known versions - most versions support
		4543	only sockets, many support pipes.
		4544
		4545	Libev tries to work around this by not using C<kqueue> by default on this
		4546	rotten platform, but of course you can still ask for it when creating a
		4547	loop - embedding a socket-only kqueue loop into a select-based one is
		4548	probably going to work well.
		4549
		4550	=head3 C<poll> is buggy
		4551
		4552	Instead of fixing C<kqueue>, Apple replaced their (working) C<poll>
		4553	implementation by something calling C<kqueue> internally around the 10.5.6
		4554	release, so now C<kqueue> I<and> C<poll> are broken.
		4555
		4556	Libev tries to work around this by not using C<poll> by default on
		4557	this rotten platform, but of course you can still ask for it when creating
		4558	a loop.
		4559
		4560	=head3 C<select> is buggy
		4561
		4562	All that's left is C<select>, and of course Apple found a way to fuck this
		4563	one up as well: On OS/X, C<select> actively limits the number of file
		4564	descriptors you can pass in to 1024 - your program suddenly crashes when
		4565	you use more.
		4566
		4567	There is an undocumented "workaround" for this - defining
		4568	C<_DARWIN_UNLIMITED_SELECT>, which libev tries to use, so select I<should>
		4569	work on OS/X.
		4570
		4571	=head2 SOLARIS PROBLEMS AND WORKAROUNDS
		4572
		4573	=head3 C<errno> reentrancy
		4574
		4575	The default compile environment on Solaris is unfortunately so
		4576	thread-unsafe that you can't even use components/libraries compiled
		4577	without C<-D_REENTRANT> in a threaded program, which, of course, isn't
		4578	defined by default. A valid, if stupid, implementation choice.
		4579
		4580	If you want to use libev in threaded environments you have to make sure
		4581	it's compiled with C<_REENTRANT> defined.
		4582
		4583	=head3 Event port backend
		4584
		4585	The scalable event interface for Solaris is called "event
		4586	ports". Unfortunately, this mechanism is very buggy in all major
		4587	releases. If you run into high CPU usage, your program freezes or you get
		4588	a large number of spurious wakeups, make sure you have all the relevant
		4589	and latest kernel patches applied. No, I don't know which ones, but there
		4590	are multiple ones to apply, and afterwards, event ports actually work
		4591	great.
		4592
		4593	If you can't get it to work, you can try running the program by setting
		4594	the environment variable C<LIBEV_FLAGS=3> to only allow C<poll> and
		4595	C<select> backends.
		4596
		4597	=head2 AIX POLL BUG
		4598
		4599	AIX unfortunately has a broken C<poll.h> header. Libev works around
		4600	this by trying to avoid the poll backend altogether (i.e. it's not even
		4601	compiled in), which normally isn't a big problem as C<select> works fine
		4602	with large bitsets on AIX, and AIX is dead anyway.
		4603
		4604	=head2 WIN32 PLATFORM LIMITATIONS AND WORKAROUNDS
		4605
		4606	=head3 General issues
2803		4607
2804	Win32 doesn't support any of the standards (e.g. POSIX) that libev	4608	Win32 doesn't support any of the standards (e.g. POSIX) that libev
2805	requires, and its I/O model is fundamentally incompatible with the POSIX	4609	requires, and its I/O model is fundamentally incompatible with the POSIX
2806	model. Libev still offers limited functionality on this platform in	4610	model. Libev still offers limited functionality on this platform in
2807	the form of the C<EVBACKEND_SELECT> backend, and only supports socket	4611	the form of the C<EVBACKEND_SELECT> backend, and only supports socket
2808	descriptors. This only applies when using Win32 natively, not when using	4612	descriptors. This only applies when using Win32 natively, not when using
2809	e.g. cygwin.	4613	e.g. cygwin. Actually, it only applies to the microsofts own compilers,
		4614	as every compielr comes with a slightly differently broken/incompatible
		4615	environment.
		4616
		4617	Lifting these limitations would basically require the full
		4618	re-implementation of the I/O system. If you are into this kind of thing,
		4619	then note that glib does exactly that for you in a very portable way (note
		4620	also that glib is the slowest event library known to man).
2810		4621
2811	There is no supported compilation method available on windows except	4622	There is no supported compilation method available on windows except
2812	embedding it into other applications.	4623	embedding it into other applications.
2813		4624
		4625	Sensible signal handling is officially unsupported by Microsoft - libev
		4626	tries its best, but under most conditions, signals will simply not work.
		4627
		4628	Not a libev limitation but worth mentioning: windows apparently doesn't
		4629	accept large writes: instead of resulting in a partial write, windows will
		4630	either accept everything or return C<ENOBUFS> if the buffer is too large,
		4631	so make sure you only write small amounts into your sockets (less than a
		4632	megabyte seems safe, but this apparently depends on the amount of memory
		4633	available).
		4634
2814	Due to the many, low, and arbitrary limits on the win32 platform and the	4635	Due to the many, low, and arbitrary limits on the win32 platform and
2815	abysmal performance of winsockets, using a large number of sockets is not	4636	the abysmal performance of winsockets, using a large number of sockets
2816	recommended (and not reasonable). If your program needs to use more than	4637	is not recommended (and not reasonable). If your program needs to use
2817	a hundred or so sockets, then likely it needs to use a totally different	4638	more than a hundred or so sockets, then likely it needs to use a totally
2818	implementation for windows, as libev offers the POSIX model, which cannot	4639	different implementation for windows, as libev offers the POSIX readiness
2819	be implemented efficiently on windows (microsoft monopoly games).	4640	notification model, which cannot be implemented efficiently on windows
		4641	(due to Microsoft monopoly games).
2820		4642
2821	=over 4	4643	A typical way to use libev under windows is to embed it (see the embedding
		4644	section for details) and use the following F<evwrap.h> header file instead
		4645	of F<ev.h>:
2822		4646
		4647	#define EV_STANDALONE /* keeps ev from requiring config.h */
		4648	#define EV_SELECT_IS_WINSOCKET 1 /* configure libev for windows select */
		4649
		4650	#include "ev.h"
		4651
		4652	And compile the following F<evwrap.c> file into your project (make sure
		4653	you do I<not> compile the F<ev.c> or any other embedded source files!):
		4654
		4655	#include "evwrap.h"
		4656	#include "ev.c"
		4657
2823	=item The winsocket select function	4658	=head3 The winsocket C<select> function
2824		4659
2825	The winsocket C<select> function doesn't follow POSIX in that it requires	4660	The winsocket C<select> function doesn't follow POSIX in that it
2826	socket I<handles> and not socket I<file descriptors>. This makes select	4661	requires socket I<handles> and not socket I<file descriptors> (it is
2827	very inefficient, and also requires a mapping from file descriptors	4662	also extremely buggy). This makes select very inefficient, and also
2828	to socket handles. See the discussion of the C<EV_SELECT_USE_FD_SET>,	4663	requires a mapping from file descriptors to socket handles (the Microsoft
2829	C<EV_SELECT_IS_WINSOCKET> and C<EV_FD_TO_WIN32_HANDLE> preprocessor	4664	C runtime provides the function C<_open_osfhandle> for this). See the
2830	symbols for more info.	4665	discussion of the C<EV_SELECT_USE_FD_SET>, C<EV_SELECT_IS_WINSOCKET> and
		4666	C<EV_FD_TO_WIN32_HANDLE> preprocessor symbols for more info.
2831		4667
2832	The configuration for a "naked" win32 using the microsoft runtime	4668	The configuration for a "naked" win32 using the Microsoft runtime
2833	libraries and raw winsocket select is:	4669	libraries and raw winsocket select is:
2834		4670
2835	#define EV_USE_SELECT 1	4671	#define EV_USE_SELECT 1
2836	#define EV_SELECT_IS_WINSOCKET 1 /* forces EV_SELECT_USE_FD_SET, too */	4672	#define EV_SELECT_IS_WINSOCKET 1 /* forces EV_SELECT_USE_FD_SET, too */
2837		4673
2838	Note that winsockets handling of fd sets is O(n), so you can easily get a	4674	Note that winsockets handling of fd sets is O(n), so you can easily get a
2839	complexity in the O(n²) range when using win32.	4675	complexity in the O(n²) range when using win32.
2840		4676
2841	=item Limited number of file descriptors	4677	=head3 Limited number of file descriptors
2842		4678
2843	Windows has numerous arbitrary (and low) limits on things. Early versions	4679	Windows has numerous arbitrary (and low) limits on things.
2844	of winsocket's select only supported waiting for a max. of C<64> handles	4680
		4681	Early versions of winsocket's select only supported waiting for a maximum
2845	(probably owning to the fact that all windows kernels can only wait for	4682	of C<64> handles (probably owning to the fact that all windows kernels
2846	C<64> things at the same time internally; microsoft recommends spawning a	4683	can only wait for C<64> things at the same time internally; Microsoft
2847	chain of threads and wait for 63 handles and the previous thread in each).	4684	recommends spawning a chain of threads and wait for 63 handles and the
		4685	previous thread in each. Sounds great!).
2848		4686
2849	Newer versions support more handles, but you need to define C<FD_SETSIZE>	4687	Newer versions support more handles, but you need to define C<FD_SETSIZE>
2850	to some high number (e.g. C<2048>) before compiling the winsocket select	4688	to some high number (e.g. C<2048>) before compiling the winsocket select
2851	call (which might be in libev or elsewhere, for example, perl does its own	4689	call (which might be in libev or elsewhere, for example, perl and many
2852	select emulation on windows).	4690	other interpreters do their own select emulation on windows).
2853		4691
2854	Another limit is the number of file descriptors in the microsoft runtime	4692	Another limit is the number of file descriptors in the Microsoft runtime
2855	libraries, which by default is C<64> (there must be a hidden I<64> fetish	4693	libraries, which by default is C<64> (there must be a hidden I<64>
2856	or something like this inside microsoft). You can increase this by calling	4694	fetish or something like this inside Microsoft). You can increase this
2857	C<_setmaxstdio>, which can increase this limit to C<2048> (another	4695	by calling C<_setmaxstdio>, which can increase this limit to C<2048>
2858	arbitrary limit), but is broken in many versions of the microsoft runtime	4696	(another arbitrary limit), but is broken in many versions of the Microsoft
2859	libraries.
2860
2861	This might get you to about C<512> or C<2048> sockets (depending on	4697	runtime libraries. This might get you to about C<512> or C<2048> sockets
2862	windows version and/or the phase of the moon). To get more, you need to	4698	(depending on windows version and/or the phase of the moon). To get more,
2863	wrap all I/O functions and provide your own fd management, but the cost of	4699	you need to wrap all I/O functions and provide your own fd management, but
2864	calling select (O(n²)) will likely make this unworkable.	4700	the cost of calling select (O(n²)) will likely make this unworkable.
		4701
		4702	=head2 PORTABILITY REQUIREMENTS
		4703
		4704	In addition to a working ISO-C implementation and of course the
		4705	backend-specific APIs, libev relies on a few additional extensions:
		4706
		4707	=over 4
		4708
		4709	=item C<void (*)(ev_watcher_type *, int revents)> must have compatible
		4710	calling conventions regardless of C<ev_watcher_type *>.
		4711
		4712	Libev assumes not only that all watcher pointers have the same internal
		4713	structure (guaranteed by POSIX but not by ISO C for example), but it also
		4714	assumes that the same (machine) code can be used to call any watcher
		4715	callback: The watcher callbacks have different type signatures, but libev
		4716	calls them using an C<ev_watcher *> internally.
		4717
		4718	=item C<sig_atomic_t volatile> must be thread-atomic as well
		4719
		4720	The type C<sig_atomic_t volatile> (or whatever is defined as
		4721	C<EV_ATOMIC_T>) must be atomic with respect to accesses from different
		4722	threads. This is not part of the specification for C<sig_atomic_t>, but is
		4723	believed to be sufficiently portable.
		4724
		4725	=item C<sigprocmask> must work in a threaded environment
		4726
		4727	Libev uses C<sigprocmask> to temporarily block signals. This is not
		4728	allowed in a threaded program (C<pthread_sigmask> has to be used). Typical
		4729	pthread implementations will either allow C<sigprocmask> in the "main
		4730	thread" or will block signals process-wide, both behaviours would
		4731	be compatible with libev. Interaction between C<sigprocmask> and
		4732	C<pthread_sigmask> could complicate things, however.
		4733
		4734	The most portable way to handle signals is to block signals in all threads
		4735	except the initial one, and run the default loop in the initial thread as
		4736	well.
		4737
		4738	=item C<long> must be large enough for common memory allocation sizes
		4739
		4740	To improve portability and simplify its API, libev uses C<long> internally
		4741	instead of C<size_t> when allocating its data structures. On non-POSIX
		4742	systems (Microsoft...) this might be unexpectedly low, but is still at
		4743	least 31 bits everywhere, which is enough for hundreds of millions of
		4744	watchers.
		4745
		4746	=item C<double> must hold a time value in seconds with enough accuracy
		4747
		4748	The type C<double> is used to represent timestamps. It is required to
		4749	have at least 51 bits of mantissa (and 9 bits of exponent), which is
		4750	good enough for at least into the year 4000 with millisecond accuracy
		4751	(the design goal for libev). This requirement is overfulfilled by
		4752	implementations using IEEE 754, which is basically all existing ones. With
		4753	IEEE 754 doubles, you get microsecond accuracy until at least 2200.
2865		4754
2866	=back	4755	=back
2867		4756
		4757	If you know of other additional requirements drop me a note.
		4758
		4759
		4760	=head1 ALGORITHMIC COMPLEXITIES
		4761
		4762	In this section the complexities of (many of) the algorithms used inside
		4763	libev will be documented. For complexity discussions about backends see
		4764	the documentation for C<ev_default_init>.
		4765
		4766	All of the following are about amortised time: If an array needs to be
		4767	extended, libev needs to realloc and move the whole array, but this
		4768	happens asymptotically rarer with higher number of elements, so O(1) might
		4769	mean that libev does a lengthy realloc operation in rare cases, but on
		4770	average it is much faster and asymptotically approaches constant time.
		4771
		4772	=over 4
		4773
		4774	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)
		4775
		4776	This means that, when you have a watcher that triggers in one hour and
		4777	there are 100 watchers that would trigger before that, then inserting will
		4778	have to skip roughly seven (C<ld 100>) of these watchers.
		4779
		4780	=item Changing timer/periodic watchers (by autorepeat or calling again): O(log skipped_other_timers)
		4781
		4782	That means that changing a timer costs less than removing/adding them,
		4783	as only the relative motion in the event queue has to be paid for.
		4784
		4785	=item Starting io/check/prepare/idle/signal/child/fork/async watchers: O(1)
		4786
		4787	These just add the watcher into an array or at the head of a list.
		4788
		4789	=item Stopping check/prepare/idle/fork/async watchers: O(1)
		4790
		4791	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))
		4792
		4793	These watchers are stored in lists, so they need to be walked to find the
		4794	correct watcher to remove. The lists are usually short (you don't usually
		4795	have many watchers waiting for the same fd or signal: one is typical, two
		4796	is rare).
		4797
		4798	=item Finding the next timer in each loop iteration: O(1)
		4799
		4800	By virtue of using a binary or 4-heap, the next timer is always found at a
		4801	fixed position in the storage array.
		4802
		4803	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)
		4804
		4805	A change means an I/O watcher gets started or stopped, which requires
		4806	libev to recalculate its status (and possibly tell the kernel, depending
		4807	on backend and whether C<ev_io_set> was used).
		4808
		4809	=item Activating one watcher (putting it into the pending state): O(1)
		4810
		4811	=item Priority handling: O(number_of_priorities)
		4812
		4813	Priorities are implemented by allocating some space for each
		4814	priority. When doing priority-based operations, libev usually has to
		4815	linearly search all the priorities, but starting/stopping and activating
		4816	watchers becomes O(1) with respect to priority handling.
		4817
		4818	=item Sending an ev_async: O(1)
		4819
		4820	=item Processing ev_async_send: O(number_of_async_watchers)
		4821
		4822	=item Processing signals: O(max_signal_number)
		4823
		4824	Sending involves a system call I<iff> there were no other C<ev_async_send>
		4825	calls in the current loop iteration. Checking for async and signal events
		4826	involves iterating over all running async watchers or all signal numbers.
		4827
		4828	=back
		4829
		4830
		4831	=head1 PORTING FROM LIBEV 3.X TO 4.X
		4832
		4833	The major version 4 introduced some minor incompatible changes to the API.
		4834
		4835	At the moment, the C<ev.h> header file tries to implement superficial
		4836	compatibility, so most programs should still compile. Those might be
		4837	removed in later versions of libev, so better update early than late.
		4838
		4839	=over 4
		4840
		4841	=item C<ev_default_destroy> and C<ev_default_fork> have been removed
		4842
		4843	These calls can be replaced easily by their C<ev_loop_xxx> counterparts:
		4844
		4845	ev_loop_destroy (EV_DEFAULT);
		4846	ev_loop_fork (EV_DEFAULT);
		4847
		4848	=item function/symbol renames
		4849
		4850	A number of functions and symbols have been renamed:
		4851
		4852	ev_loop => ev_run
		4853	EVLOOP_NONBLOCK => EVRUN_NOWAIT
		4854	EVLOOP_ONESHOT => EVRUN_ONCE
		4855
		4856	ev_unloop => ev_break
		4857	EVUNLOOP_CANCEL => EVBREAK_CANCEL
		4858	EVUNLOOP_ONE => EVBREAK_ONE
		4859	EVUNLOOP_ALL => EVBREAK_ALL
		4860
		4861	EV_TIMEOUT => EV_TIMER
		4862
		4863	ev_loop_count => ev_iteration
		4864	ev_loop_depth => ev_depth
		4865	ev_loop_verify => ev_verify
		4866
		4867	Most functions working on C<struct ev_loop> objects don't have an
		4868	C<ev_loop_> prefix, so it was removed; C<ev_loop>, C<ev_unloop> and
		4869	associated constants have been renamed to not collide with the C<struct
		4870	ev_loop> anymore and C<EV_TIMER> now follows the same naming scheme
		4871	as all other watcher types. Note that C<ev_loop_fork> is still called
		4872	C<ev_loop_fork> because it would otherwise clash with the C<ev_fork>
		4873	typedef.
		4874
		4875	=item C<EV_COMPAT3> backwards compatibility mechanism
		4876
		4877	The backward compatibility mechanism can be controlled by
		4878	C<EV_COMPAT3>. See L<PREPROCESSOR SYMBOLS/MACROS> in the L<EMBEDDING>
		4879	section.
		4880
		4881	=item C<EV_MINIMAL> mechanism replaced by C<EV_FEATURES>
		4882
		4883	The preprocessor symbol C<EV_MINIMAL> has been replaced by a different
		4884	mechanism, C<EV_FEATURES>. Programs using C<EV_MINIMAL> usually compile
		4885	and work, but the library code will of course be larger.
		4886
		4887	=back
		4888
		4889
		4890	=head1 GLOSSARY
		4891
		4892	=over 4
		4893
		4894	=item active
		4895
		4896	A watcher is active as long as it has been started and not yet stopped.
		4897	See L<WATCHER STATES> for details.
		4898
		4899	=item application
		4900
		4901	In this document, an application is whatever is using libev.
		4902
		4903	=item backend
		4904
		4905	The part of the code dealing with the operating system interfaces.
		4906
		4907	=item callback
		4908
		4909	The address of a function that is called when some event has been
		4910	detected. Callbacks are being passed the event loop, the watcher that
		4911	received the event, and the actual event bitset.
		4912
		4913	=item callback/watcher invocation
		4914
		4915	The act of calling the callback associated with a watcher.
		4916
		4917	=item event
		4918
		4919	A change of state of some external event, such as data now being available
		4920	for reading on a file descriptor, time having passed or simply not having
		4921	any other events happening anymore.
		4922
		4923	In libev, events are represented as single bits (such as C<EV_READ> or
		4924	C<EV_TIMER>).
		4925
		4926	=item event library
		4927
		4928	A software package implementing an event model and loop.
		4929
		4930	=item event loop
		4931
		4932	An entity that handles and processes external events and converts them
		4933	into callback invocations.
		4934
		4935	=item event model
		4936
		4937	The model used to describe how an event loop handles and processes
		4938	watchers and events.
		4939
		4940	=item pending
		4941
		4942	A watcher is pending as soon as the corresponding event has been
		4943	detected. See L<WATCHER STATES> for details.
		4944
		4945	=item real time
		4946
		4947	The physical time that is observed. It is apparently strictly monotonic :)
		4948
		4949	=item wall-clock time
		4950
		4951	The time and date as shown on clocks. Unlike real time, it can actually
		4952	be wrong and jump forwards and backwards, e.g. when the you adjust your
		4953	clock.
		4954
		4955	=item watcher
		4956
		4957	A data structure that describes interest in certain events. Watchers need
		4958	to be started (attached to an event loop) before they can receive events.
		4959
		4960	=back
2868		4961
2869	=head1 AUTHOR	4962	=head1 AUTHOR
2870		4963
2871	Marc Lehmann <libev@schmorp.de>.	4964	Marc Lehmann <libev@schmorp.de>, with repeated corrections by Mikael Magnusson.
2872		4965

Diff Legend

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