ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/libev/ev.pod

Comparing libev/ev.pod (file contents):
Revision 1.35 by root, Fri Nov 23 19:35:09 2007 UTC vs.
Revision 1.345 by root, Wed Nov 10 14:36:42 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	=head1 DESCRIPTION	9	=head2 EXAMPLE PROGRAM
		10
		11	// a single header file is required
		12	#include <ev.h>
		13
		14	#include <stdio.h> // for puts
		15
		16	// every watcher type has its own typedef'd struct
		17	// with the name ev_TYPE
		18	ev_io stdin_watcher;
		19	ev_timer timeout_watcher;
		20
		21	// all watcher callbacks have a similar signature
		22	// this callback is called when data is readable on stdin
		23	static void
		24	stdin_cb (EV_P_ ev_io *w, int revents)
		25	{
		26	puts ("stdin ready");
		27	// for one-shot events, one must manually stop the watcher
		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);
		33	}
		34
		35	// another callback, this time for a time-out
		36	static void
		37	timeout_cb (EV_P_ ev_timer *w, int revents)
		38	{
		39	puts ("timeout");
		40	// this causes the innermost ev_run to stop iterating
		41	ev_break (EV_A_ EVBREAK_ONE);
		42	}
		43
		44	int
		45	main (void)
		46	{
		47	// use the default event loop unless you have special needs
		48	struct ev_loop *loop = EV_DEFAULT;
		49
		50	// initialise an io watcher, then start it
		51	// this one will watch for stdin to become readable
		52	ev_io_init (&stdin_watcher, stdin_cb, /*STDIN_FILENO*/ 0, EV_READ);
		53	ev_io_start (loop, &stdin_watcher);
		54
		55	// initialise a timer watcher, then start it
		56	// simple non-repeating 5.5 second timeout
		57	ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);
		58	ev_timer_start (loop, &timeout_watcher);
		59
		60	// now wait for events to arrive
		61	ev_run (loop, 0);
		62
		63	// unloop was called, so exit
		64	return 0;
		65	}
		66
		67	=head1 ABOUT THIS DOCUMENT
		68
		69	This document documents the libev software package.
		70
		71	The newest version of this document is also available as an html-formatted
		72	web page you might find easier to navigate when reading it for the first
		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 WHAT TO READ WHEN IN A HURRY
		84
		85	This manual tries to be very detailed, but unfortunately, this also makes
		86	it very long. If you just want to know the basics of libev, I suggest
		87	reading L<ANATOMY OF A WATCHER>, then the L<EXAMPLE PROGRAM> above and
		88	look up the missing functions in L<GLOBAL FUNCTIONS> and the C<ev_io> and
		89	C<ev_timer> sections in L<WATCHER TYPES>.
		90
		91	=head1 ABOUT LIBEV
10		92
11	Libev is an event loop: you register interest in certain events (such as a	93	Libev is an event loop: you register interest in certain events (such as a
12	file descriptor being readable or a timeout occuring), and it will manage	94	file descriptor being readable or a timeout occurring), and it will manage
13	these event sources and provide your program with events.	95	these event sources and provide your program with events.
14		96
15	To do this, it must take more or less complete control over your process	97	To do this, it must take more or less complete control over your process
16	(or thread) by executing the I<event loop> handler, and will then	98	(or thread) by executing the I<event loop> handler, and will then
17	communicate events via a callback mechanism.	99	communicate events via a callback mechanism.
…		…
19	You register interest in certain events by registering so-called I<event	101	You register interest in certain events by registering so-called I<event
20	watchers>, which are relatively small C structures you initialise with the	102	watchers>, which are relatively small C structures you initialise with the
21	details of the event, and then hand it over to libev by I<starting> the	103	details of the event, and then hand it over to libev by I<starting> the
22	watcher.	104	watcher.
23		105
24	=head1 FEATURES	106	=head2 FEATURES
25		107
26	Libev supports select, poll, the linux-specific epoll and the bsd-specific	108	Libev supports C<select>, C<poll>, the Linux-specific C<epoll>, the
27	kqueue mechanisms for file descriptor events, relative timers, absolute	109	BSD-specific C<kqueue> and the Solaris-specific event port mechanisms
28	timers with customised rescheduling, signal events, process status change	110	for file descriptor events (C<ev_io>), the Linux C<inotify> interface
		111	(for C<ev_stat>), Linux eventfd/signalfd (for faster and cleaner
		112	inter-thread wakeup (C<ev_async>)/signal handling (C<ev_signal>)) relative
		113	timers (C<ev_timer>), absolute timers with customised rescheduling
		114	(C<ev_periodic>), synchronous signals (C<ev_signal>), process status
29	events (related to SIGCHLD), and event watchers dealing with the event	115	change events (C<ev_child>), and event watchers dealing with the event
30	loop mechanism itself (idle, prepare and check watchers). It also is quite	116	loop mechanism itself (C<ev_idle>, C<ev_embed>, C<ev_prepare> and
		117	C<ev_check> watchers) as well as file watchers (C<ev_stat>) and even
		118	limited support for fork events (C<ev_fork>).
		119
		120	It also is quite fast (see this
31	fast (see this L<benchmark|http://libev.schmorp.de/bench.html> comparing	121	L<benchmark|http://libev.schmorp.de/bench.html> comparing it to libevent
32	it to libevent for example).	122	for example).
33		123
34	=head1 CONVENTIONS	124	=head2 CONVENTIONS
35		125
36	Libev is very configurable. In this manual the default configuration	126	Libev is very configurable. In this manual the default (and most common)
37	will be described, which supports multiple event loops. For more info	127	configuration will be described, which supports multiple event loops. For
38	about various configuration options please have a look at the file	128	more info about various configuration options please have a look at
39	F<README.embed> in the libev distribution. If libev was configured without	129	B<EMBED> section in this manual. If libev was configured without support
40	support for multiple event loops, then all functions taking an initial	130	for multiple event loops, then all functions taking an initial argument of
41	argument of name C<loop> (which is always of type C<struct ev_loop *>)	131	name C<loop> (which is always of type C<struct ev_loop *>) will not have
42	will not have this argument.	132	this argument.
43		133
44	=head1 TIME REPRESENTATION	134	=head2 TIME REPRESENTATION
45		135
46	Libev represents time as a single floating point number, representing the	136	Libev represents time as a single floating point number, representing
47	(fractional) number of seconds since the (POSIX) epoch (somewhere near	137	the (fractional) number of seconds since the (POSIX) epoch (in practice
48	the beginning of 1970, details are complicated, don't ask). This type is	138	somewhere near the beginning of 1970, details are complicated, don't
49	called C<ev_tstamp>, which is what you should use too. It usually aliases	139	ask). This type is called C<ev_tstamp>, which is what you should use
50	to the C<double> type in C, and when you need to do any calculations on	140	too. It usually aliases to the C<double> type in C. When you need to do
51	it, you should treat it as such.	141	any calculations on it, you should treat it as some floating point value.
		142
		143	Unlike the name component C<stamp> might indicate, it is also used for
		144	time differences (e.g. delays) throughout libev.
		145
		146	=head1 ERROR HANDLING
		147
		148	Libev knows three classes of errors: operating system errors, usage errors
		149	and internal errors (bugs).
		150
		151	When libev catches an operating system error it cannot handle (for example
		152	a system call indicating a condition libev cannot fix), it calls the callback
		153	set via C<ev_set_syserr_cb>, which is supposed to fix the problem or
		154	abort. The default is to print a diagnostic message and to call C<abort
		155	()>.
		156
		157	When libev detects a usage error such as a negative timer interval, then
		158	it will print a diagnostic message and abort (via the C<assert> mechanism,
		159	so C<NDEBUG> will disable this checking): these are programming errors in
		160	the libev caller and need to be fixed there.
		161
		162	Libev also has a few internal error-checking C<assert>ions, and also has
		163	extensive consistency checking code. These do not trigger under normal
		164	circumstances, as they indicate either a bug in libev or worse.
52		165
53		166
54	=head1 GLOBAL FUNCTIONS	167	=head1 GLOBAL FUNCTIONS
55		168
56	These functions can be called anytime, even before initialising the	169	These functions can be called anytime, even before initialising the
…		…
60		173
61	=item ev_tstamp ev_time ()	174	=item ev_tstamp ev_time ()
62		175
63	Returns the current time as libev would use it. Please note that the	176	Returns the current time as libev would use it. Please note that the
64	C<ev_now> function is usually faster and also often returns the timestamp	177	C<ev_now> function is usually faster and also often returns the timestamp
65	you actually want to know.	178	you actually want to know. Also interesting is the combination of
		179	C<ev_update_now> and C<ev_now>.
		180
		181	=item ev_sleep (ev_tstamp interval)
		182
		183	Sleep for the given interval: The current thread will be blocked until
		184	either it is interrupted or the given time interval has passed. Basically
		185	this is a sub-second-resolution C<sleep ()>.
66		186
67	=item int ev_version_major ()	187	=item int ev_version_major ()
68		188
69	=item int ev_version_minor ()	189	=item int ev_version_minor ()
70		190
71	You can find out the major and minor version numbers of the library	191	You can find out the major and minor ABI version numbers of the library
72	you linked against by calling the functions C<ev_version_major> and	192	you linked against by calling the functions C<ev_version_major> and
73	C<ev_version_minor>. If you want, you can compare against the global	193	C<ev_version_minor>. If you want, you can compare against the global
74	symbols C<EV_VERSION_MAJOR> and C<EV_VERSION_MINOR>, which specify the	194	symbols C<EV_VERSION_MAJOR> and C<EV_VERSION_MINOR>, which specify the
75	version of the library your program was compiled against.	195	version of the library your program was compiled against.
76		196
		197	These version numbers refer to the ABI version of the library, not the
		198	release version.
		199
77	Usually, it's a good idea to terminate if the major versions mismatch,	200	Usually, it's a good idea to terminate if the major versions mismatch,
78	as this indicates an incompatible change. Minor versions are usually	201	as this indicates an incompatible change. Minor versions are usually
79	compatible to older versions, so a larger minor version alone is usually	202	compatible to older versions, so a larger minor version alone is usually
80	not a problem.	203	not a problem.
81		204
82	Example: make sure we haven't accidentally been linked against the wrong	205	Example: Make sure we haven't accidentally been linked against the wrong
83	version:	206	version (note, however, that this will not detect other ABI mismatches,
		207	such as LFS or reentrancy).
84		208
85	assert (("libev version mismatch",	209	assert (("libev version mismatch",
86	ev_version_major () == EV_VERSION_MAJOR	210	ev_version_major () == EV_VERSION_MAJOR
87	&& ev_version_minor () >= EV_VERSION_MINOR));	211	&& ev_version_minor () >= EV_VERSION_MINOR));
88		212
89	=item unsigned int ev_supported_backends ()	213	=item unsigned int ev_supported_backends ()
90		214
91	Return the set of all backends (i.e. their corresponding C<EV_BACKEND_*>	215	Return the set of all backends (i.e. their corresponding C<EV_BACKEND_*>
92	value) compiled into this binary of libev (independent of their	216	value) compiled into this binary of libev (independent of their
…		…
94	a description of the set values.	218	a description of the set values.
95		219
96	Example: make sure we have the epoll method, because yeah this is cool and	220	Example: make sure we have the epoll method, because yeah this is cool and
97	a must have and can we have a torrent of it please!!!11	221	a must have and can we have a torrent of it please!!!11
98		222
99	assert (("sorry, no epoll, no sex",	223	assert (("sorry, no epoll, no sex",
100	ev_supported_backends () & EVBACKEND_EPOLL));	224	ev_supported_backends () & EVBACKEND_EPOLL));
101		225
102	=item unsigned int ev_recommended_backends ()	226	=item unsigned int ev_recommended_backends ()
103		227
104	Return the set of all backends compiled into this binary of libev and also	228	Return the set of all backends compiled into this binary of libev and
105	recommended for this platform. This set is often smaller than the one	229	also recommended for this platform, meaning it will work for most file
		230	descriptor types. This set is often smaller than the one returned by
106	returned by C<ev_supported_backends>, as for example kqueue is broken on	231	C<ev_supported_backends>, as for example kqueue is broken on most BSDs
107	most BSDs and will not be autodetected unless you explicitly request it	232	and will not be auto-detected unless you explicitly request it (assuming
108	(assuming you know what you are doing). This is the set of backends that	233	you know what you are doing). This is the set of backends that libev will
109	libev will probe for if you specify no backends explicitly.	234	probe for if you specify no backends explicitly.
110		235
111	=item unsigned int ev_embeddable_backends ()	236	=item unsigned int ev_embeddable_backends ()
112		237
113	Returns the set of backends that are embeddable in other event loops. This	238	Returns the set of backends that are embeddable in other event loops. This
114	is the theoretical, all-platform, value. To find which backends	239	value is platform-specific but can include backends not available on the
115	might be supported on the current system, you would need to look at	240	current system. To find which embeddable backends might be supported on
116	C<ev_embeddable_backends () & ev_supported_backends ()>, likewise for	241	the current system, you would need to look at C<ev_embeddable_backends ()
117	recommended ones.	242	& ev_supported_backends ()>, likewise for recommended ones.
118		243
119	See the description of C<ev_embed> watchers for more info.	244	See the description of C<ev_embed> watchers for more info.
120		245
121	=item ev_set_allocator (void *(*cb)(void *ptr, long size))	246	=item ev_set_allocator (void *(*cb)(void *ptr, long size))
122		247
123	Sets the allocation function to use (the prototype is similar to the	248	Sets the allocation function to use (the prototype is similar - the
124	realloc C function, the semantics are identical). It is used to allocate	249	semantics are identical to the C<realloc> C89/SuS/POSIX function). It is
125	and free memory (no surprises here). If it returns zero when memory	250	used to allocate and free memory (no surprises here). If it returns zero
126	needs to be allocated, the library might abort or take some potentially	251	when memory needs to be allocated (C<size != 0>), the library might abort
127	destructive action. The default is your system realloc function.	252	or take some potentially destructive action.
		253
		254	Since some systems (at least OpenBSD and Darwin) fail to implement
		255	correct C<realloc> semantics, libev will use a wrapper around the system
		256	C<realloc> and C<free> functions by default.
128		257
129	You could override this function in high-availability programs to, say,	258	You could override this function in high-availability programs to, say,
130	free some memory if it cannot allocate memory, to use a special allocator,	259	free some memory if it cannot allocate memory, to use a special allocator,
131	or even to sleep a while and retry until some memory is available.	260	or even to sleep a while and retry until some memory is available.
132		261
133	Example: replace the libev allocator with one that waits a bit and then	262	Example: Replace the libev allocator with one that waits a bit and then
134	retries: better than mine).	263	retries (example requires a standards-compliant C<realloc>).
135		264
136	static void *	265	static void *
137	persistent_realloc (void *ptr, long size)	266	persistent_realloc (void *ptr, size_t size)
138	{	267	{
139	for (;;)	268	for (;;)
140	{	269	{
141	void *newptr = realloc (ptr, size);	270	void *newptr = realloc (ptr, size);
142		271
…		…
148	}	277	}
149		278
150	...	279	...
151	ev_set_allocator (persistent_realloc);	280	ev_set_allocator (persistent_realloc);
152		281
153	=item ev_set_syserr_cb (void (*cb)(const char *msg));	282	=item ev_set_syserr_cb (void (*cb)(const char *msg))
154		283
155	Set the callback function to call on a retryable syscall error (such	284	Set the callback function to call on a retryable system call error (such
156	as failed select, poll, epoll_wait). The message is a printable string	285	as failed select, poll, epoll_wait). The message is a printable string
157	indicating the system call or subsystem causing the problem. If this	286	indicating the system call or subsystem causing the problem. If this
158	callback is set, then libev will expect it to remedy the sitution, no	287	callback is set, then libev will expect it to remedy the situation, no
159	matter what, when it returns. That is, libev will generally retry the	288	matter what, when it returns. That is, libev will generally retry the
160	requested operation, or, if the condition doesn't go away, do bad stuff	289	requested operation, or, if the condition doesn't go away, do bad stuff
161	(such as abort).	290	(such as abort).
162		291
163	Example: do the same thing as libev does internally:	292	Example: This is basically the same thing that libev does internally, too.
164		293
165	static void	294	static void
166	fatal_error (const char *msg)	295	fatal_error (const char *msg)
167	{	296	{
168	perror (msg);	297	perror (msg);
…		…
172	...	301	...
173	ev_set_syserr_cb (fatal_error);	302	ev_set_syserr_cb (fatal_error);
174		303
175	=back	304	=back
176		305
177	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP	306	=head1 FUNCTIONS CONTROLLING EVENT LOOPS
178		307
179	An event loop is described by a C<struct ev_loop *>. The library knows two	308	An event loop is described by a C<struct ev_loop *> (the C<struct> is
180	types of such loops, the I<default> loop, which supports signals and child	309	I<not> optional in this case unless libev 3 compatibility is disabled, as
181	events, and dynamically created loops which do not.	310	libev 3 had an C<ev_loop> function colliding with the struct name).
182		311
183	If you use threads, a common model is to run the default event loop	312	The library knows two types of such loops, the I<default> loop, which
184	in your main thread (or in a separate thread) and for each thread you	313	supports child process events, and dynamically created event loops which
185	create, you also create another event loop. Libev itself does no locking	314	do not.
186	whatsoever, so if you mix calls to the same event loop in different
187	threads, make sure you lock (this is usually a bad idea, though, even if
188	done correctly, because it's hideous and inefficient).
189		315
190	=over 4	316	=over 4
191		317
192	=item struct ev_loop *ev_default_loop (unsigned int flags)	318	=item struct ev_loop *ev_default_loop (unsigned int flags)
193		319
194	This will initialise the default event loop if it hasn't been initialised	320	This returns the "default" event loop object, which is what you should
195	yet and return it. If the default loop could not be initialised, returns	321	normally use when you just need "the event loop". Event loop objects and
196	false. If it already was initialised it simply returns it (and ignores the	322	the C<flags> parameter are described in more detail in the entry for
197	flags. If that is troubling you, check C<ev_backend ()> afterwards).	323	C<ev_loop_new>.
		324
		325	If the default loop is already initialised then this function simply
		326	returns it (and ignores the flags. If that is troubling you, check
		327	C<ev_backend ()> afterwards). Otherwise it will create it with the given
		328	flags, which should almost always be C<0>, unless the caller is also the
		329	one calling C<ev_run> or otherwise qualifies as "the main program".
198		330
199	If you don't know what event loop to use, use the one returned from this	331	If you don't know what event loop to use, use the one returned from this
200	function.	332	function (or via the C<EV_DEFAULT> macro).
		333
		334	Note that this function is I<not> thread-safe, so if you want to use it
		335	from multiple threads, you have to employ some kind of mutex (note also
		336	that this case is unlikely, as loops cannot be shared easily between
		337	threads anyway).
		338
		339	The default loop is the only loop that can handle C<ev_child> watchers,
		340	and to do this, it always registers a handler for C<SIGCHLD>. If this is
		341	a problem for your application you can either create a dynamic loop with
		342	C<ev_loop_new> which doesn't do that, or you can simply overwrite the
		343	C<SIGCHLD> signal handler I<after> calling C<ev_default_init>.
		344
		345	Example: This is the most typical usage.
		346
		347	if (!ev_default_loop (0))
		348	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
		349
		350	Example: Restrict libev to the select and poll backends, and do not allow
		351	environment settings to be taken into account:
		352
		353	ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);
		354
		355	=item struct ev_loop *ev_loop_new (unsigned int flags)
		356
		357	This will create and initialise a new event loop object. If the loop
		358	could not be initialised, returns false.
		359
		360	This function is thread-safe, and one common way to use libev with
		361	threads is indeed to create one loop per thread, and using the default
		362	loop in the "main" or "initial" thread.
201		363
202	The flags argument can be used to specify special behaviour or specific	364	The flags argument can be used to specify special behaviour or specific
203	backends to use, and is usually specified as C<0> (or C<EVFLAG_AUTO>).	365	backends to use, and is usually specified as C<0> (or C<EVFLAG_AUTO>).
204		366
205	The following flags are supported:	367	The following flags are supported:
…		…
211	The default flags value. Use this if you have no clue (it's the right	373	The default flags value. Use this if you have no clue (it's the right
212	thing, believe me).	374	thing, believe me).
213		375
214	=item C<EVFLAG_NOENV>	376	=item C<EVFLAG_NOENV>
215		377
216	If this flag bit is ored into the flag value (or the program runs setuid	378	If this flag bit is or'ed into the flag value (or the program runs setuid
217	or setgid) then libev will I<not> look at the environment variable	379	or setgid) then libev will I<not> look at the environment variable
218	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will	380	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will
219	override the flags completely if it is found in the environment. This is	381	override the flags completely if it is found in the environment. This is
220	useful to try out specific backends to test their performance, or to work	382	useful to try out specific backends to test their performance, or to work
221	around bugs.	383	around bugs.
222		384
		385	=item C<EVFLAG_FORKCHECK>
		386
		387	Instead of calling C<ev_loop_fork> manually after a fork, you can also
		388	make libev check for a fork in each iteration by enabling this flag.
		389
		390	This works by calling C<getpid ()> on every iteration of the loop,
		391	and thus this might slow down your event loop if you do a lot of loop
		392	iterations and little real work, but is usually not noticeable (on my
		393	GNU/Linux system for example, C<getpid> is actually a simple 5-insn sequence
		394	without a system call and thus I<very> fast, but my GNU/Linux system also has
		395	C<pthread_atfork> which is even faster).
		396
		397	The big advantage of this flag is that you can forget about fork (and
		398	forget about forgetting to tell libev about forking) when you use this
		399	flag.
		400
		401	This flag setting cannot be overridden or specified in the C<LIBEV_FLAGS>
		402	environment variable.
		403
		404	=item C<EVFLAG_NOINOTIFY>
		405
		406	When this flag is specified, then libev will not attempt to use the
		407	I<inotify> API for its C<ev_stat> watchers. Apart from debugging and
		408	testing, this flag can be useful to conserve inotify file descriptors, as
		409	otherwise each loop using C<ev_stat> watchers consumes one inotify handle.
		410
		411	=item C<EVFLAG_SIGNALFD>
		412
		413	When this flag is specified, then libev will attempt to use the
		414	I<signalfd> API for its C<ev_signal> (and C<ev_child>) watchers. This API
		415	delivers signals synchronously, which makes it both faster and might make
		416	it possible to get the queued signal data. It can also simplify signal
		417	handling with threads, as long as you properly block signals in your
		418	threads that are not interested in handling them.
		419
		420	Signalfd will not be used by default as this changes your signal mask, and
		421	there are a lot of shoddy libraries and programs (glib's threadpool for
		422	example) that can't properly initialise their signal masks.
		423
223	=item C<EVBACKEND_SELECT> (value 1, portable select backend)	424	=item C<EVBACKEND_SELECT> (value 1, portable select backend)
224		425
225	This is your standard select(2) backend. Not I<completely> standard, as	426	This is your standard select(2) backend. Not I<completely> standard, as
226	libev tries to roll its own fd_set with no limits on the number of fds,	427	libev tries to roll its own fd_set with no limits on the number of fds,
227	but if that fails, expect a fairly low limit on the number of fds when	428	but if that fails, expect a fairly low limit on the number of fds when
228	using this backend. It doesn't scale too well (O(highest_fd)), but its usually	429	using this backend. It doesn't scale too well (O(highest_fd)), but its
229	the fastest backend for a low number of fds.	430	usually the fastest backend for a low number of (low-numbered :) fds.
		431
		432	To get good performance out of this backend you need a high amount of
		433	parallelism (most of the file descriptors should be busy). If you are
		434	writing a server, you should C<accept ()> in a loop to accept as many
		435	connections as possible during one iteration. You might also want to have
		436	a look at C<ev_set_io_collect_interval ()> to increase the amount of
		437	readiness notifications you get per iteration.
		438
		439	This backend maps C<EV_READ> to the C<readfds> set and C<EV_WRITE> to the
		440	C<writefds> set (and to work around Microsoft Windows bugs, also onto the
		441	C<exceptfds> set on that platform).
230		442
231	=item C<EVBACKEND_POLL> (value 2, poll backend, available everywhere except on windows)	443	=item C<EVBACKEND_POLL> (value 2, poll backend, available everywhere except on windows)
232		444
233	And this is your standard poll(2) backend. It's more complicated than	445	And this is your standard poll(2) backend. It's more complicated
234	select, but handles sparse fds better and has no artificial limit on the	446	than select, but handles sparse fds better and has no artificial
235	number of fds you can use (except it will slow down considerably with a	447	limit on the number of fds you can use (except it will slow down
236	lot of inactive fds). It scales similarly to select, i.e. O(total_fds).	448	considerably with a lot of inactive fds). It scales similarly to select,
		449	i.e. O(total_fds). See the entry for C<EVBACKEND_SELECT>, above, for
		450	performance tips.
		451
		452	This backend maps C<EV_READ> to C<POLLIN | POLLERR | POLLHUP>, and
		453	C<EV_WRITE> to C<POLLOUT | POLLERR | POLLHUP>.
237		454
238	=item C<EVBACKEND_EPOLL> (value 4, Linux)	455	=item C<EVBACKEND_EPOLL> (value 4, Linux)
239		456
		457	Use the linux-specific epoll(7) interface (for both pre- and post-2.6.9
		458	kernels).
		459
240	For few fds, this backend is a bit little slower than poll and select,	460	For few fds, this backend is a bit little slower than poll and select,
241	but it scales phenomenally better. While poll and select usually scale like	461	but it scales phenomenally better. While poll and select usually scale
242	O(total_fds) where n is the total number of fds (or the highest fd), epoll scales	462	like O(total_fds) where n is the total number of fds (or the highest fd),
243	either O(1) or O(active_fds).	463	epoll scales either O(1) or O(active_fds).
244		464
		465	The epoll mechanism deserves honorable mention as the most misdesigned
		466	of the more advanced event mechanisms: mere annoyances include silently
		467	dropping file descriptors, requiring a system call per change per file
		468	descriptor (and unnecessary guessing of parameters), problems with dup,
		469	returning before the timeout value, resulting in additional iterations
		470	(and only giving 5ms accuracy while select on the same platform gives
		471	0.1ms) and so on. The biggest issue is fork races, however - if a program
		472	forks then I<both> parent and child process have to recreate the epoll
		473	set, which can take considerable time (one syscall per file descriptor)
		474	and is of course hard to detect.
		475
		476	Epoll is also notoriously buggy - embedding epoll fds I<should> work, but
		477	of course I<doesn't>, and epoll just loves to report events for totally
		478	I<different> file descriptors (even already closed ones, so one cannot
		479	even remove them from the set) than registered in the set (especially
		480	on SMP systems). Libev tries to counter these spurious notifications by
		481	employing an additional generation counter and comparing that against the
		482	events to filter out spurious ones, recreating the set when required. Last
		483	not least, it also refuses to work with some file descriptors which work
		484	perfectly fine with C<select> (files, many character devices...).
		485
		486	Epoll is truly the train wreck analog among event poll mechanisms.
		487
245	While stopping and starting an I/O watcher in the same iteration will	488	While stopping, setting and starting an I/O watcher in the same iteration
246	result in some caching, there is still a syscall per such incident	489	will result in some caching, there is still a system call per such
247	(because the fd could point to a different file description now), so its	490	incident (because the same I<file descriptor> could point to a different
248	best to avoid that. Also, dup()ed file descriptors might not work very	491	I<file description> now), so its best to avoid that. Also, C<dup ()>'ed
249	well if you register events for both fds.	492	file descriptors might not work very well if you register events for both
		493	file descriptors.
250		494
251	Please note that epoll sometimes generates spurious notifications, so you	495	Best performance from this backend is achieved by not unregistering all
252	need to use non-blocking I/O or other means to avoid blocking when no data	496	watchers for a file descriptor until it has been closed, if possible,
253	(or space) is available.	497	i.e. keep at least one watcher active per fd at all times. Stopping and
		498	starting a watcher (without re-setting it) also usually doesn't cause
		499	extra overhead. A fork can both result in spurious notifications as well
		500	as in libev having to destroy and recreate the epoll object, which can
		501	take considerable time and thus should be avoided.
		502
		503	All this means that, in practice, C<EVBACKEND_SELECT> can be as fast or
		504	faster than epoll for maybe up to a hundred file descriptors, depending on
		505	the usage. So sad.
		506
		507	While nominally embeddable in other event loops, this feature is broken in
		508	all kernel versions tested so far.
		509
		510	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
		511	C<EVBACKEND_POLL>.
254		512
255	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)	513	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)
256		514
257	Kqueue deserves special mention, as at the time of this writing, it	515	Kqueue deserves special mention, as at the time of this writing, it
258	was broken on all BSDs except NetBSD (usually it doesn't work with	516	was broken on all BSDs except NetBSD (usually it doesn't work reliably
259	anything but sockets and pipes, except on Darwin, where of course its	517	with anything but sockets and pipes, except on Darwin, where of course
260	completely useless). For this reason its not being "autodetected"	518	it's completely useless). Unlike epoll, however, whose brokenness
		519	is by design, these kqueue bugs can (and eventually will) be fixed
		520	without API changes to existing programs. For this reason it's not being
261	unless you explicitly specify it explicitly in the flags (i.e. using	521	"auto-detected" unless you explicitly specify it in the flags (i.e. using
262	C<EVBACKEND_KQUEUE>).	522	C<EVBACKEND_KQUEUE>) or libev was compiled on a known-to-be-good (-enough)
		523	system like NetBSD.
		524
		525	You still can embed kqueue into a normal poll or select backend and use it
		526	only for sockets (after having made sure that sockets work with kqueue on
		527	the target platform). See C<ev_embed> watchers for more info.
263		528
264	It scales in the same way as the epoll backend, but the interface to the	529	It scales in the same way as the epoll backend, but the interface to the
265	kernel is more efficient (which says nothing about its actual speed, of	530	kernel is more efficient (which says nothing about its actual speed, of
266	course). While starting and stopping an I/O watcher does not cause an	531	course). While stopping, setting and starting an I/O watcher does never
267	extra syscall as with epoll, it still adds up to four event changes per	532	cause an extra system call as with C<EVBACKEND_EPOLL>, it still adds up to
268	incident, so its best to avoid that.	533	two event changes per incident. Support for C<fork ()> is very bad (but
		534	sane, unlike epoll) and it drops fds silently in similarly hard-to-detect
		535	cases
		536
		537	This backend usually performs well under most conditions.
		538
		539	While nominally embeddable in other event loops, this doesn't work
		540	everywhere, so you might need to test for this. And since it is broken
		541	almost everywhere, you should only use it when you have a lot of sockets
		542	(for which it usually works), by embedding it into another event loop
		543	(e.g. C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> (but C<poll> is of course
		544	also broken on OS X)) and, did I mention it, using it only for sockets.
		545
		546	This backend maps C<EV_READ> into an C<EVFILT_READ> kevent with
		547	C<NOTE_EOF>, and C<EV_WRITE> into an C<EVFILT_WRITE> kevent with
		548	C<NOTE_EOF>.
269		549
270	=item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8)	550	=item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8)
271		551
272	This is not implemented yet (and might never be).	552	This is not implemented yet (and might never be, unless you send me an
		553	implementation). According to reports, C</dev/poll> only supports sockets
		554	and is not embeddable, which would limit the usefulness of this backend
		555	immensely.
273		556
274	=item C<EVBACKEND_PORT> (value 32, Solaris 10)	557	=item C<EVBACKEND_PORT> (value 32, Solaris 10)
275		558
276	This uses the Solaris 10 port mechanism. As with everything on Solaris,	559	This uses the Solaris 10 event port mechanism. As with everything on Solaris,
277	it's really slow, but it still scales very well (O(active_fds)).	560	it's really slow, but it still scales very well (O(active_fds)).
278		561
279	Please note that solaris ports can result in a lot of spurious	562	Please note that Solaris event ports can deliver a lot of spurious
280	notifications, so you need to use non-blocking I/O or other means to avoid	563	notifications, so you need to use non-blocking I/O or other means to avoid
281	blocking when no data (or space) is available.	564	blocking when no data (or space) is available.
		565
		566	While this backend scales well, it requires one system call per active
		567	file descriptor per loop iteration. For small and medium numbers of file
		568	descriptors a "slow" C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> backend
		569	might perform better.
		570
		571	On the positive side, with the exception of the spurious readiness
		572	notifications, this backend actually performed fully to specification
		573	in all tests and is fully embeddable, which is a rare feat among the
		574	OS-specific backends (I vastly prefer correctness over speed hacks).
		575
		576	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
		577	C<EVBACKEND_POLL>.
282		578
283	=item C<EVBACKEND_ALL>	579	=item C<EVBACKEND_ALL>
284		580
285	Try all backends (even potentially broken ones that wouldn't be tried	581	Try all backends (even potentially broken ones that wouldn't be tried
286	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as	582	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as
287	C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>.	583	C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>.
288		584
		585	It is definitely not recommended to use this flag.
		586
289	=back	587	=back
290		588
291	If one or more of these are ored into the flags value, then only these	589	If one or more of the backend flags are or'ed into the flags value,
292	backends will be tried (in the reverse order as given here). If none are	590	then only these backends will be tried (in the reverse order as listed
293	specified, most compiled-in backend will be tried, usually in reverse	591	here). If none are specified, all backends in C<ev_recommended_backends
294	order of their flag values :)	592	()> will be tried.
295		593
296	The most typical usage is like this:
297
298	if (!ev_default_loop (0))
299	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
300
301	Restrict libev to the select and poll backends, and do not allow
302	environment settings to be taken into account:
303
304	ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);
305
306	Use whatever libev has to offer, but make sure that kqueue is used if
307	available (warning, breaks stuff, best use only with your own private
308	event loop and only if you know the OS supports your types of fds):
309
310	ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);
311
312	=item struct ev_loop *ev_loop_new (unsigned int flags)
313
314	Similar to C<ev_default_loop>, but always creates a new event loop that is
315	always distinct from the default loop. Unlike the default loop, it cannot
316	handle signal and child watchers, and attempts to do so will be greeted by
317	undefined behaviour (or a failed assertion if assertions are enabled).
318
319	Example: try to create a event loop that uses epoll and nothing else.	594	Example: Try to create a event loop that uses epoll and nothing else.
320		595
321	struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);	596	struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);
322	if (!epoller)	597	if (!epoller)
323	fatal ("no epoll found here, maybe it hides under your chair");	598	fatal ("no epoll found here, maybe it hides under your chair");
324		599
325	=item ev_default_destroy ()	600	Example: Use whatever libev has to offer, but make sure that kqueue is
		601	used if available.
326		602
327	Destroys the default loop again (frees all memory and kernel state	603	struct ev_loop *loop = ev_loop_new (ev_recommended_backends () | EVBACKEND_KQUEUE);
328	etc.). This stops all registered event watchers (by not touching them in
329	any way whatsoever, although you cannot rely on this :).
330		604
331	=item ev_loop_destroy (loop)	605	=item ev_loop_destroy (loop)
332		606
333	Like C<ev_default_destroy>, but destroys an event loop created by an	607	Destroys an event loop object (frees all memory and kernel state
334	earlier call to C<ev_loop_new>.	608	etc.). None of the active event watchers will be stopped in the normal
		609	sense, so e.g. C<ev_is_active> might still return true. It is your
		610	responsibility to either stop all watchers cleanly yourself I<before>
		611	calling this function, or cope with the fact afterwards (which is usually
		612	the easiest thing, you can just ignore the watchers and/or C<free ()> them
		613	for example).
335		614
336	=item ev_default_fork ()	615	Note that certain global state, such as signal state (and installed signal
		616	handlers), will not be freed by this function, and related watchers (such
		617	as signal and child watchers) would need to be stopped manually.
337		618
		619	This function is normally used on loop objects allocated by
		620	C<ev_loop_new>, but it can also be used on the default loop returned by
		621	C<ev_default_loop>, in which case it is not thread-safe.
		622
		623	Note that it is not advisable to call this function on the default loop
		624	except in the rare occasion where you really need to free its resources.
		625	If you need dynamically allocated loops it is better to use C<ev_loop_new>
		626	and C<ev_loop_destroy>.
		627
		628	=item ev_loop_fork (loop)
		629
		630	This function sets a flag that causes subsequent C<ev_run> iterations to
338	This function reinitialises the kernel state for backends that have	631	reinitialise the kernel state for backends that have one. Despite the
339	one. Despite the name, you can call it anytime, but it makes most sense	632	name, you can call it anytime, but it makes most sense after forking, in
340	after forking, in either the parent or child process (or both, but that	633	the child process. You I<must> call it (or use C<EVFLAG_FORKCHECK>) in the
341	again makes little sense).	634	child before resuming or calling C<ev_run>.
342		635
343	You I<must> call this function in the child process after forking if and	636	Again, you I<have> to call it on I<any> loop that you want to re-use after
344	only if you want to use the event library in both processes. If you just	637	a fork, I<even if you do not plan to use the loop in the parent>. This is
345	fork+exec, you don't have to call it.	638	because some kernel interfaces *cough* I<kqueue> *cough* do funny things
		639	during fork.
		640
		641	On the other hand, you only need to call this function in the child
		642	process if and only if you want to use the event loop in the child. If
		643	you just fork+exec or create a new loop in the child, you don't have to
		644	call it at all (in fact, C<epoll> is so badly broken that it makes a
		645	difference, but libev will usually detect this case on its own and do a
		646	costly reset of the backend).
346		647
347	The function itself is quite fast and it's usually not a problem to call	648	The function itself is quite fast and it's usually not a problem to call
348	it just in case after a fork. To make this easy, the function will fit in	649	it just in case after a fork.
349	quite nicely into a call to C<pthread_atfork>:
350		650
		651	Example: Automate calling C<ev_loop_fork> on the default loop when
		652	using pthreads.
		653
		654	static void
		655	post_fork_child (void)
		656	{
		657	ev_loop_fork (EV_DEFAULT);
		658	}
		659
		660	...
351	pthread_atfork (0, 0, ev_default_fork);	661	pthread_atfork (0, 0, post_fork_child);
352		662
353	At the moment, C<EVBACKEND_SELECT> and C<EVBACKEND_POLL> are safe to use	663	=item int ev_is_default_loop (loop)
354	without calling this function, so if you force one of those backends you
355	do not need to care.
356		664
357	=item ev_loop_fork (loop)	665	Returns true when the given loop is, in fact, the default loop, and false
		666	otherwise.
358		667
359	Like C<ev_default_fork>, but acts on an event loop created by	668	=item unsigned int ev_iteration (loop)
360	C<ev_loop_new>. Yes, you have to call this on every allocated event loop	669
361	after fork, and how you do this is entirely your own problem.	670	Returns the current iteration count for the event loop, which is identical
		671	to the number of times libev did poll for new events. It starts at C<0>
		672	and happily wraps around with enough iterations.
		673
		674	This value can sometimes be useful as a generation counter of sorts (it
		675	"ticks" the number of loop iterations), as it roughly corresponds with
		676	C<ev_prepare> and C<ev_check> calls - and is incremented between the
		677	prepare and check phases.
		678
		679	=item unsigned int ev_depth (loop)
		680
		681	Returns the number of times C<ev_run> was entered minus the number of
		682	times C<ev_run> was exited normally, in other words, the recursion depth.
		683
		684	Outside C<ev_run>, this number is zero. In a callback, this number is
		685	C<1>, unless C<ev_run> was invoked recursively (or from another thread),
		686	in which case it is higher.
		687
		688	Leaving C<ev_run> abnormally (setjmp/longjmp, cancelling the thread,
		689	throwing an exception etc.), doesn't count as "exit" - consider this
		690	as a hint to avoid such ungentleman-like behaviour unless it's really
		691	convenient, in which case it is fully supported.
362		692
363	=item unsigned int ev_backend (loop)	693	=item unsigned int ev_backend (loop)
364		694
365	Returns one of the C<EVBACKEND_*> flags indicating the event backend in	695	Returns one of the C<EVBACKEND_*> flags indicating the event backend in
366	use.	696	use.
…		…
369		699
370	Returns the current "event loop time", which is the time the event loop	700	Returns the current "event loop time", which is the time the event loop
371	received events and started processing them. This timestamp does not	701	received events and started processing them. This timestamp does not
372	change as long as callbacks are being processed, and this is also the base	702	change as long as callbacks are being processed, and this is also the base
373	time used for relative timers. You can treat it as the timestamp of the	703	time used for relative timers. You can treat it as the timestamp of the
374	event occuring (or more correctly, libev finding out about it).	704	event occurring (or more correctly, libev finding out about it).
375		705
		706	=item ev_now_update (loop)
		707
		708	Establishes the current time by querying the kernel, updating the time
		709	returned by C<ev_now ()> in the progress. This is a costly operation and
		710	is usually done automatically within C<ev_run ()>.
		711
		712	This function is rarely useful, but when some event callback runs for a
		713	very long time without entering the event loop, updating libev's idea of
		714	the current time is a good idea.
		715
		716	See also L<The special problem of time updates> in the C<ev_timer> section.
		717
		718	=item ev_suspend (loop)
		719
		720	=item ev_resume (loop)
		721
		722	These two functions suspend and resume an event loop, for use when the
		723	loop is not used for a while and timeouts should not be processed.
		724
		725	A typical use case would be an interactive program such as a game: When
		726	the user presses C<^Z> to suspend the game and resumes it an hour later it
		727	would be best to handle timeouts as if no time had actually passed while
		728	the program was suspended. This can be achieved by calling C<ev_suspend>
		729	in your C<SIGTSTP> handler, sending yourself a C<SIGSTOP> and calling
		730	C<ev_resume> directly afterwards to resume timer processing.
		731
		732	Effectively, all C<ev_timer> watchers will be delayed by the time spend
		733	between C<ev_suspend> and C<ev_resume>, and all C<ev_periodic> watchers
		734	will be rescheduled (that is, they will lose any events that would have
		735	occurred while suspended).
		736
		737	After calling C<ev_suspend> you B<must not> call I<any> function on the
		738	given loop other than C<ev_resume>, and you B<must not> call C<ev_resume>
		739	without a previous call to C<ev_suspend>.
		740
		741	Calling C<ev_suspend>/C<ev_resume> has the side effect of updating the
		742	event loop time (see C<ev_now_update>).
		743
376	=item ev_loop (loop, int flags)	744	=item ev_run (loop, int flags)
377		745
378	Finally, this is it, the event handler. This function usually is called	746	Finally, this is it, the event handler. This function usually is called
379	after you initialised all your watchers and you want to start handling	747	after you have initialised all your watchers and you want to start
380	events.	748	handling events. It will ask the operating system for any new events, call
		749	the watcher callbacks, an then repeat the whole process indefinitely: This
		750	is why event loops are called I<loops>.
381		751
382	If the flags argument is specified as C<0>, it will not return until	752	If the flags argument is specified as C<0>, it will keep handling events
383	either no event watchers are active anymore or C<ev_unloop> was called.	753	until either no event watchers are active anymore or C<ev_break> was
		754	called.
384		755
385	Please note that an explicit C<ev_unloop> is usually better than	756	Please note that an explicit C<ev_break> is usually better than
386	relying on all watchers to be stopped when deciding when a program has	757	relying on all watchers to be stopped when deciding when a program has
387	finished (especially in interactive programs), but having a program that	758	finished (especially in interactive programs), but having a program
388	automatically loops as long as it has to and no longer by virtue of	759	that automatically loops as long as it has to and no longer by virtue
389	relying on its watchers stopping correctly is a thing of beauty.	760	of relying on its watchers stopping correctly, that is truly a thing of
		761	beauty.
390		762
		763	This function is also I<mostly> exception-safe - you can break out of
		764	a C<ev_run> call by calling C<longjmp> in a callback, throwing a C++
		765	exception and so on. This does not decrement the C<ev_depth> value, nor
		766	will it clear any outstanding C<EVBREAK_ONE> breaks.
		767
391	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle	768	A flags value of C<EVRUN_NOWAIT> will look for new events, will handle
392	those events and any outstanding ones, but will not block your process in	769	those events and any already outstanding ones, but will not wait and
393	case there are no events and will return after one iteration of the loop.	770	block your process in case there are no events and will return after one
		771	iteration of the loop. This is sometimes useful to poll and handle new
		772	events while doing lengthy calculations, to keep the program responsive.
394		773
395	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if	774	A flags value of C<EVRUN_ONCE> will look for new events (waiting if
396	neccessary) and will handle those and any outstanding ones. It will block	775	necessary) and will handle those and any already outstanding ones. It
397	your process until at least one new event arrives, and will return after	776	will block your process until at least one new event arrives (which could
398	one iteration of the loop. This is useful if you are waiting for some	777	be an event internal to libev itself, so there is no guarantee that a
399	external event in conjunction with something not expressible using other	778	user-registered callback will be called), and will return after one
		779	iteration of the loop.
		780
		781	This is useful if you are waiting for some external event in conjunction
		782	with something not expressible using other libev watchers (i.e. "roll your
400	libev watchers. However, a pair of C<ev_prepare>/C<ev_check> watchers is	783	own C<ev_run>"). However, a pair of C<ev_prepare>/C<ev_check> watchers is
401	usually a better approach for this kind of thing.	784	usually a better approach for this kind of thing.
402		785
403	Here are the gory details of what C<ev_loop> does:	786	Here are the gory details of what C<ev_run> does:
404		787
405	* If there are no active watchers (reference count is zero), return.	788	- Increment loop depth.
406	- Queue prepare watchers and then call all outstanding watchers.	789	- Reset the ev_break status.
		790	- Before the first iteration, call any pending watchers.
		791	LOOP:
		792	- If EVFLAG_FORKCHECK was used, check for a fork.
		793	- If a fork was detected (by any means), queue and call all fork watchers.
		794	- Queue and call all prepare watchers.
		795	- If ev_break was called, goto FINISH.
407	- If we have been forked, recreate the kernel state.	796	- If we have been forked, detach and recreate the kernel state
		797	as to not disturb the other process.
408	- Update the kernel state with all outstanding changes.	798	- Update the kernel state with all outstanding changes.
409	- Update the "event loop time".	799	- Update the "event loop time" (ev_now ()).
410	- Calculate for how long to block.	800	- Calculate for how long to sleep or block, if at all
		801	(active idle watchers, EVRUN_NOWAIT or not having
		802	any active watchers at all will result in not sleeping).
		803	- Sleep if the I/O and timer collect interval say so.
		804	- Increment loop iteration counter.
411	- Block the process, waiting for any events.	805	- Block the process, waiting for any events.
412	- Queue all outstanding I/O (fd) events.	806	- Queue all outstanding I/O (fd) events.
413	- Update the "event loop time" and do time jump handling.	807	- Update the "event loop time" (ev_now ()), and do time jump adjustments.
414	- Queue all outstanding timers.	808	- Queue all expired timers.
415	- Queue all outstanding periodics.	809	- Queue all expired periodics.
416	- If no events are pending now, queue all idle watchers.	810	- Queue all idle watchers with priority higher than that of pending events.
417	- Queue all check watchers.	811	- Queue all check watchers.
418	- Call all queued watchers in reverse order (i.e. check watchers first).	812	- Call all queued watchers in reverse order (i.e. check watchers first).
419	Signals and child watchers are implemented as I/O watchers, and will	813	Signals and child watchers are implemented as I/O watchers, and will
420	be handled here by queueing them when their watcher gets executed.	814	be handled here by queueing them when their watcher gets executed.
421	- If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK	815	- If ev_break has been called, or EVRUN_ONCE or EVRUN_NOWAIT
422	were used, return, otherwise continue with step *.	816	were used, or there are no active watchers, goto FINISH, otherwise
		817	continue with step LOOP.
		818	FINISH:
		819	- Reset the ev_break status iff it was EVBREAK_ONE.
		820	- Decrement the loop depth.
		821	- Return.
423		822
424	Example: queue some jobs and then loop until no events are outsanding	823	Example: Queue some jobs and then loop until no events are outstanding
425	anymore.	824	anymore.
426		825
427	... queue jobs here, make sure they register event watchers as long	826	... queue jobs here, make sure they register event watchers as long
428	... as they still have work to do (even an idle watcher will do..)	827	... as they still have work to do (even an idle watcher will do..)
429	ev_loop (my_loop, 0);	828	ev_run (my_loop, 0);
430	... jobs done. yeah!	829	... jobs done or somebody called unloop. yeah!
431		830
432	=item ev_unloop (loop, how)	831	=item ev_break (loop, how)
433		832
434	Can be used to make a call to C<ev_loop> return early (but only after it	833	Can be used to make a call to C<ev_run> return early (but only after it
435	has processed all outstanding events). The C<how> argument must be either	834	has processed all outstanding events). The C<how> argument must be either
436	C<EVUNLOOP_ONE>, which will make the innermost C<ev_loop> call return, or	835	C<EVBREAK_ONE>, which will make the innermost C<ev_run> call return, or
437	C<EVUNLOOP_ALL>, which will make all nested C<ev_loop> calls return.	836	C<EVBREAK_ALL>, which will make all nested C<ev_run> calls return.
		837
		838	This "break state" will be cleared on the next call to C<ev_run>.
		839
		840	It is safe to call C<ev_break> from outside any C<ev_run> calls, too, in
		841	which case it will have no effect.
438		842
439	=item ev_ref (loop)	843	=item ev_ref (loop)
440		844
441	=item ev_unref (loop)	845	=item ev_unref (loop)
442		846
443	Ref/unref can be used to add or remove a reference count on the event	847	Ref/unref can be used to add or remove a reference count on the event
444	loop: Every watcher keeps one reference, and as long as the reference	848	loop: Every watcher keeps one reference, and as long as the reference
445	count is nonzero, C<ev_loop> will not return on its own. If you have	849	count is nonzero, C<ev_run> will not return on its own.
446	a watcher you never unregister that should not keep C<ev_loop> from	850
447	returning, ev_unref() after starting, and ev_ref() before stopping it. For	851	This is useful when you have a watcher that you never intend to
		852	unregister, but that nevertheless should not keep C<ev_run> from
		853	returning. In such a case, call C<ev_unref> after starting, and C<ev_ref>
		854	before stopping it.
		855
448	example, libev itself uses this for its internal signal pipe: It is not	856	As an example, libev itself uses this for its internal signal pipe: It
449	visible to the libev user and should not keep C<ev_loop> from exiting if	857	is not visible to the libev user and should not keep C<ev_run> from
450	no event watchers registered by it are active. It is also an excellent	858	exiting if no event watchers registered by it are active. It is also an
451	way to do this for generic recurring timers or from within third-party	859	excellent way to do this for generic recurring timers or from within
452	libraries. Just remember to I<unref after start> and I<ref before stop>.	860	third-party libraries. Just remember to I<unref after start> and I<ref
		861	before stop> (but only if the watcher wasn't active before, or was active
		862	before, respectively. Note also that libev might stop watchers itself
		863	(e.g. non-repeating timers) in which case you have to C<ev_ref>
		864	in the callback).
453		865
454	Example: create a signal watcher, but keep it from keeping C<ev_loop>	866	Example: Create a signal watcher, but keep it from keeping C<ev_run>
455	running when nothing else is active.	867	running when nothing else is active.
456		868
457	struct dv_signal exitsig;	869	ev_signal exitsig;
458	ev_signal_init (&exitsig, sig_cb, SIGINT);	870	ev_signal_init (&exitsig, sig_cb, SIGINT);
459	ev_signal_start (myloop, &exitsig);	871	ev_signal_start (loop, &exitsig);
460	evf_unref (myloop);	872	evf_unref (loop);
461		873
462	Example: for some weird reason, unregister the above signal handler again.	874	Example: For some weird reason, unregister the above signal handler again.
463		875
464	ev_ref (myloop);	876	ev_ref (loop);
465	ev_signal_stop (myloop, &exitsig);	877	ev_signal_stop (loop, &exitsig);
		878
		879	=item ev_set_io_collect_interval (loop, ev_tstamp interval)
		880
		881	=item ev_set_timeout_collect_interval (loop, ev_tstamp interval)
		882
		883	These advanced functions influence the time that libev will spend waiting
		884	for events. Both time intervals are by default C<0>, meaning that libev
		885	will try to invoke timer/periodic callbacks and I/O callbacks with minimum
		886	latency.
		887
		888	Setting these to a higher value (the C<interval> I<must> be >= C<0>)
		889	allows libev to delay invocation of I/O and timer/periodic callbacks
		890	to increase efficiency of loop iterations (or to increase power-saving
		891	opportunities).
		892
		893	The idea is that sometimes your program runs just fast enough to handle
		894	one (or very few) event(s) per loop iteration. While this makes the
		895	program responsive, it also wastes a lot of CPU time to poll for new
		896	events, especially with backends like C<select ()> which have a high
		897	overhead for the actual polling but can deliver many events at once.
		898
		899	By setting a higher I<io collect interval> you allow libev to spend more
		900	time collecting I/O events, so you can handle more events per iteration,
		901	at the cost of increasing latency. Timeouts (both C<ev_periodic> and
		902	C<ev_timer>) will be not affected. Setting this to a non-null value will
		903	introduce an additional C<ev_sleep ()> call into most loop iterations. The
		904	sleep time ensures that libev will not poll for I/O events more often then
		905	once per this interval, on average.
		906
		907	Likewise, by setting a higher I<timeout collect interval> you allow libev
		908	to spend more time collecting timeouts, at the expense of increased
		909	latency/jitter/inexactness (the watcher callback will be called
		910	later). C<ev_io> watchers will not be affected. Setting this to a non-null
		911	value will not introduce any overhead in libev.
		912
		913	Many (busy) programs can usually benefit by setting the I/O collect
		914	interval to a value near C<0.1> or so, which is often enough for
		915	interactive servers (of course not for games), likewise for timeouts. It
		916	usually doesn't make much sense to set it to a lower value than C<0.01>,
		917	as this approaches the timing granularity of most systems. Note that if
		918	you do transactions with the outside world and you can't increase the
		919	parallelity, then this setting will limit your transaction rate (if you
		920	need to poll once per transaction and the I/O collect interval is 0.01,
		921	then you can't do more than 100 transactions per second).
		922
		923	Setting the I<timeout collect interval> can improve the opportunity for
		924	saving power, as the program will "bundle" timer callback invocations that
		925	are "near" in time together, by delaying some, thus reducing the number of
		926	times the process sleeps and wakes up again. Another useful technique to
		927	reduce iterations/wake-ups is to use C<ev_periodic> watchers and make sure
		928	they fire on, say, one-second boundaries only.
		929
		930	Example: we only need 0.1s timeout granularity, and we wish not to poll
		931	more often than 100 times per second:
		932
		933	ev_set_timeout_collect_interval (EV_DEFAULT_UC_ 0.1);
		934	ev_set_io_collect_interval (EV_DEFAULT_UC_ 0.01);
		935
		936	=item ev_invoke_pending (loop)
		937
		938	This call will simply invoke all pending watchers while resetting their
		939	pending state. Normally, C<ev_run> does this automatically when required,
		940	but when overriding the invoke callback this call comes handy. This
		941	function can be invoked from a watcher - this can be useful for example
		942	when you want to do some lengthy calculation and want to pass further
		943	event handling to another thread (you still have to make sure only one
		944	thread executes within C<ev_invoke_pending> or C<ev_run> of course).
		945
		946	=item int ev_pending_count (loop)
		947
		948	Returns the number of pending watchers - zero indicates that no watchers
		949	are pending.
		950
		951	=item ev_set_invoke_pending_cb (loop, void (*invoke_pending_cb)(EV_P))
		952
		953	This overrides the invoke pending functionality of the loop: Instead of
		954	invoking all pending watchers when there are any, C<ev_run> will call
		955	this callback instead. This is useful, for example, when you want to
		956	invoke the actual watchers inside another context (another thread etc.).
		957
		958	If you want to reset the callback, use C<ev_invoke_pending> as new
		959	callback.
		960
		961	=item ev_set_loop_release_cb (loop, void (*release)(EV_P), void (*acquire)(EV_P))
		962
		963	Sometimes you want to share the same loop between multiple threads. This
		964	can be done relatively simply by putting mutex_lock/unlock calls around
		965	each call to a libev function.
		966
		967	However, C<ev_run> can run an indefinite time, so it is not feasible
		968	to wait for it to return. One way around this is to wake up the event
		969	loop via C<ev_break> and C<av_async_send>, another way is to set these
		970	I<release> and I<acquire> callbacks on the loop.
		971
		972	When set, then C<release> will be called just before the thread is
		973	suspended waiting for new events, and C<acquire> is called just
		974	afterwards.
		975
		976	Ideally, C<release> will just call your mutex_unlock function, and
		977	C<acquire> will just call the mutex_lock function again.
		978
		979	While event loop modifications are allowed between invocations of
		980	C<release> and C<acquire> (that's their only purpose after all), no
		981	modifications done will affect the event loop, i.e. adding watchers will
		982	have no effect on the set of file descriptors being watched, or the time
		983	waited. Use an C<ev_async> watcher to wake up C<ev_run> when you want it
		984	to take note of any changes you made.
		985
		986	In theory, threads executing C<ev_run> will be async-cancel safe between
		987	invocations of C<release> and C<acquire>.
		988
		989	See also the locking example in the C<THREADS> section later in this
		990	document.
		991
		992	=item ev_set_userdata (loop, void *data)
		993
		994	=item void *ev_userdata (loop)
		995
		996	Set and retrieve a single C<void *> associated with a loop. When
		997	C<ev_set_userdata> has never been called, then C<ev_userdata> returns
		998	C<0>.
		999
		1000	These two functions can be used to associate arbitrary data with a loop,
		1001	and are intended solely for the C<invoke_pending_cb>, C<release> and
		1002	C<acquire> callbacks described above, but of course can be (ab-)used for
		1003	any other purpose as well.
		1004
		1005	=item ev_verify (loop)
		1006
		1007	This function only does something when C<EV_VERIFY> support has been
		1008	compiled in, which is the default for non-minimal builds. It tries to go
		1009	through all internal structures and checks them for validity. If anything
		1010	is found to be inconsistent, it will print an error message to standard
		1011	error and call C<abort ()>.
		1012
		1013	This can be used to catch bugs inside libev itself: under normal
		1014	circumstances, this function will never abort as of course libev keeps its
		1015	data structures consistent.
466		1016
467	=back	1017	=back
468		1018
		1019
469	=head1 ANATOMY OF A WATCHER	1020	=head1 ANATOMY OF A WATCHER
470		1021
		1022	In the following description, uppercase C<TYPE> in names stands for the
		1023	watcher type, e.g. C<ev_TYPE_start> can mean C<ev_timer_start> for timer
		1024	watchers and C<ev_io_start> for I/O watchers.
		1025
471	A watcher is a structure that you create and register to record your	1026	A watcher is an opaque structure that you allocate and register to record
472	interest in some event. For instance, if you want to wait for STDIN to	1027	your interest in some event. To make a concrete example, imagine you want
473	become readable, you would create an C<ev_io> watcher for that:	1028	to wait for STDIN to become readable, you would create an C<ev_io> watcher
		1029	for that:
474		1030
475	static void my_cb (struct ev_loop *loop, struct ev_io *w, int revents)	1031	static void my_cb (struct ev_loop *loop, ev_io *w, int revents)
476	{	1032	{
477	ev_io_stop (w);	1033	ev_io_stop (w);
478	ev_unloop (loop, EVUNLOOP_ALL);	1034	ev_break (loop, EVBREAK_ALL);
479	}	1035	}
480		1036
481	struct ev_loop *loop = ev_default_loop (0);	1037	struct ev_loop *loop = ev_default_loop (0);
		1038
482	struct ev_io stdin_watcher;	1039	ev_io stdin_watcher;
		1040
483	ev_init (&stdin_watcher, my_cb);	1041	ev_init (&stdin_watcher, my_cb);
484	ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);	1042	ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);
485	ev_io_start (loop, &stdin_watcher);	1043	ev_io_start (loop, &stdin_watcher);
		1044
486	ev_loop (loop, 0);	1045	ev_run (loop, 0);
487		1046
488	As you can see, you are responsible for allocating the memory for your	1047	As you can see, you are responsible for allocating the memory for your
489	watcher structures (and it is usually a bad idea to do this on the stack,	1048	watcher structures (and it is I<usually> a bad idea to do this on the
490	although this can sometimes be quite valid).	1049	stack).
491		1050
		1051	Each watcher has an associated watcher structure (called C<struct ev_TYPE>
		1052	or simply C<ev_TYPE>, as typedefs are provided for all watcher structs).
		1053
492	Each watcher structure must be initialised by a call to C<ev_init	1054	Each watcher structure must be initialised by a call to C<ev_init (watcher
493	(watcher *, callback)>, which expects a callback to be provided. This	1055	*, callback)>, which expects a callback to be provided. This callback is
494	callback gets invoked each time the event occurs (or, in the case of io	1056	invoked each time the event occurs (or, in the case of I/O watchers, each
495	watchers, each time the event loop detects that the file descriptor given	1057	time the event loop detects that the file descriptor given is readable
496	is readable and/or writable).	1058	and/or writable).
497		1059
498	Each watcher type has its own C<< ev_<type>_set (watcher *, ...) >> macro	1060	Each watcher type further has its own C<< ev_TYPE_set (watcher *, ...) >>
499	with arguments specific to this watcher type. There is also a macro	1061	macro to configure it, with arguments specific to the watcher type. There
500	to combine initialisation and setting in one call: C<< ev_<type>_init	1062	is also a macro to combine initialisation and setting in one call: C<<
501	(watcher *, callback, ...) >>.	1063	ev_TYPE_init (watcher *, callback, ...) >>.
502		1064
503	To make the watcher actually watch out for events, you have to start it	1065	To make the watcher actually watch out for events, you have to start it
504	with a watcher-specific start function (C<< ev_<type>_start (loop, watcher	1066	with a watcher-specific start function (C<< ev_TYPE_start (loop, watcher
505	*) >>), and you can stop watching for events at any time by calling the	1067	*) >>), and you can stop watching for events at any time by calling the
506	corresponding stop function (C<< ev_<type>_stop (loop, watcher *) >>.	1068	corresponding stop function (C<< ev_TYPE_stop (loop, watcher *) >>.
507		1069
508	As long as your watcher is active (has been started but not stopped) you	1070	As long as your watcher is active (has been started but not stopped) you
509	must not touch the values stored in it. Most specifically you must never	1071	must not touch the values stored in it. Most specifically you must never
510	reinitialise it or call its set macro.	1072	reinitialise it or call its C<ev_TYPE_set> macro.
511
512	You can check whether an event is active by calling the C<ev_is_active
513	(watcher *)> macro. To see whether an event is outstanding (but the
514	callback for it has not been called yet) you can use the C<ev_is_pending
515	(watcher *)> macro.
516		1073
517	Each and every callback receives the event loop pointer as first, the	1074	Each and every callback receives the event loop pointer as first, the
518	registered watcher structure as second, and a bitset of received events as	1075	registered watcher structure as second, and a bitset of received events as
519	third argument.	1076	third argument.
520		1077
…		…
529	=item C<EV_WRITE>	1086	=item C<EV_WRITE>
530		1087
531	The file descriptor in the C<ev_io> watcher has become readable and/or	1088	The file descriptor in the C<ev_io> watcher has become readable and/or
532	writable.	1089	writable.
533		1090
534	=item C<EV_TIMEOUT>	1091	=item C<EV_TIMER>
535		1092
536	The C<ev_timer> watcher has timed out.	1093	The C<ev_timer> watcher has timed out.
537		1094
538	=item C<EV_PERIODIC>	1095	=item C<EV_PERIODIC>
539		1096
…		…
545		1102
546	=item C<EV_CHILD>	1103	=item C<EV_CHILD>
547		1104
548	The pid specified in the C<ev_child> watcher has received a status change.	1105	The pid specified in the C<ev_child> watcher has received a status change.
549		1106
		1107	=item C<EV_STAT>
		1108
		1109	The path specified in the C<ev_stat> watcher changed its attributes somehow.
		1110
550	=item C<EV_IDLE>	1111	=item C<EV_IDLE>
551		1112
552	The C<ev_idle> watcher has determined that you have nothing better to do.	1113	The C<ev_idle> watcher has determined that you have nothing better to do.
553		1114
554	=item C<EV_PREPARE>	1115	=item C<EV_PREPARE>
555		1116
556	=item C<EV_CHECK>	1117	=item C<EV_CHECK>
557		1118
558	All C<ev_prepare> watchers are invoked just I<before> C<ev_loop> starts	1119	All C<ev_prepare> watchers are invoked just I<before> C<ev_run> starts
559	to gather new events, and all C<ev_check> watchers are invoked just after	1120	to gather new events, and all C<ev_check> watchers are invoked just after
560	C<ev_loop> has gathered them, but before it invokes any callbacks for any	1121	C<ev_run> has gathered them, but before it invokes any callbacks for any
561	received events. Callbacks of both watcher types can start and stop as	1122	received events. Callbacks of both watcher types can start and stop as
562	many watchers as they want, and all of them will be taken into account	1123	many watchers as they want, and all of them will be taken into account
563	(for example, a C<ev_prepare> watcher might start an idle watcher to keep	1124	(for example, a C<ev_prepare> watcher might start an idle watcher to keep
564	C<ev_loop> from blocking).	1125	C<ev_run> from blocking).
		1126
		1127	=item C<EV_EMBED>
		1128
		1129	The embedded event loop specified in the C<ev_embed> watcher needs attention.
		1130
		1131	=item C<EV_FORK>
		1132
		1133	The event loop has been resumed in the child process after fork (see
		1134	C<ev_fork>).
		1135
		1136	=item C<EV_CLEANUP>
		1137
		1138	The event loop is about to be destroyed (see C<ev_cleanup>).
		1139
		1140	=item C<EV_ASYNC>
		1141
		1142	The given async watcher has been asynchronously notified (see C<ev_async>).
		1143
		1144	=item C<EV_CUSTOM>
		1145
		1146	Not ever sent (or otherwise used) by libev itself, but can be freely used
		1147	by libev users to signal watchers (e.g. via C<ev_feed_event>).
565		1148
566	=item C<EV_ERROR>	1149	=item C<EV_ERROR>
567		1150
568	An unspecified error has occured, the watcher has been stopped. This might	1151	An unspecified error has occurred, the watcher has been stopped. This might
569	happen because the watcher could not be properly started because libev	1152	happen because the watcher could not be properly started because libev
570	ran out of memory, a file descriptor was found to be closed or any other	1153	ran out of memory, a file descriptor was found to be closed or any other
		1154	problem. Libev considers these application bugs.
		1155
571	problem. You best act on it by reporting the problem and somehow coping	1156	You best act on it by reporting the problem and somehow coping with the
572	with the watcher being stopped.	1157	watcher being stopped. Note that well-written programs should not receive
		1158	an error ever, so when your watcher receives it, this usually indicates a
		1159	bug in your program.
573		1160
574	Libev will usually signal a few "dummy" events together with an error,	1161	Libev will usually signal a few "dummy" events together with an error, for
575	for example it might indicate that a fd is readable or writable, and if	1162	example it might indicate that a fd is readable or writable, and if your
576	your callbacks is well-written it can just attempt the operation and cope	1163	callbacks is well-written it can just attempt the operation and cope with
577	with the error from read() or write(). This will not work in multithreaded	1164	the error from read() or write(). This will not work in multi-threaded
578	programs, though, so beware.	1165	programs, though, as the fd could already be closed and reused for another
		1166	thing, so beware.
579		1167
580	=back	1168	=back
581		1169
		1170	=head2 GENERIC WATCHER FUNCTIONS
		1171
		1172	=over 4
		1173
		1174	=item C<ev_init> (ev_TYPE *watcher, callback)
		1175
		1176	This macro initialises the generic portion of a watcher. The contents
		1177	of the watcher object can be arbitrary (so C<malloc> will do). Only
		1178	the generic parts of the watcher are initialised, you I<need> to call
		1179	the type-specific C<ev_TYPE_set> macro afterwards to initialise the
		1180	type-specific parts. For each type there is also a C<ev_TYPE_init> macro
		1181	which rolls both calls into one.
		1182
		1183	You can reinitialise a watcher at any time as long as it has been stopped
		1184	(or never started) and there are no pending events outstanding.
		1185
		1186	The callback is always of type C<void (*)(struct ev_loop *loop, ev_TYPE *watcher,
		1187	int revents)>.
		1188
		1189	Example: Initialise an C<ev_io> watcher in two steps.
		1190
		1191	ev_io w;
		1192	ev_init (&w, my_cb);
		1193	ev_io_set (&w, STDIN_FILENO, EV_READ);
		1194
		1195	=item C<ev_TYPE_set> (ev_TYPE *watcher, [args])
		1196
		1197	This macro initialises the type-specific parts of a watcher. You need to
		1198	call C<ev_init> at least once before you call this macro, but you can
		1199	call C<ev_TYPE_set> any number of times. You must not, however, call this
		1200	macro on a watcher that is active (it can be pending, however, which is a
		1201	difference to the C<ev_init> macro).
		1202
		1203	Although some watcher types do not have type-specific arguments
		1204	(e.g. C<ev_prepare>) you still need to call its C<set> macro.
		1205
		1206	See C<ev_init>, above, for an example.
		1207
		1208	=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])
		1209
		1210	This convenience macro rolls both C<ev_init> and C<ev_TYPE_set> macro
		1211	calls into a single call. This is the most convenient method to initialise
		1212	a watcher. The same limitations apply, of course.
		1213
		1214	Example: Initialise and set an C<ev_io> watcher in one step.
		1215
		1216	ev_io_init (&w, my_cb, STDIN_FILENO, EV_READ);
		1217
		1218	=item C<ev_TYPE_start> (loop, ev_TYPE *watcher)
		1219
		1220	Starts (activates) the given watcher. Only active watchers will receive
		1221	events. If the watcher is already active nothing will happen.
		1222
		1223	Example: Start the C<ev_io> watcher that is being abused as example in this
		1224	whole section.
		1225
		1226	ev_io_start (EV_DEFAULT_UC, &w);
		1227
		1228	=item C<ev_TYPE_stop> (loop, ev_TYPE *watcher)
		1229
		1230	Stops the given watcher if active, and clears the pending status (whether
		1231	the watcher was active or not).
		1232
		1233	It is possible that stopped watchers are pending - for example,
		1234	non-repeating timers are being stopped when they become pending - but
		1235	calling C<ev_TYPE_stop> ensures that the watcher is neither active nor
		1236	pending. If you want to free or reuse the memory used by the watcher it is
		1237	therefore a good idea to always call its C<ev_TYPE_stop> function.
		1238
		1239	=item bool ev_is_active (ev_TYPE *watcher)
		1240
		1241	Returns a true value iff the watcher is active (i.e. it has been started
		1242	and not yet been stopped). As long as a watcher is active you must not modify
		1243	it.
		1244
		1245	=item bool ev_is_pending (ev_TYPE *watcher)
		1246
		1247	Returns a true value iff the watcher is pending, (i.e. it has outstanding
		1248	events but its callback has not yet been invoked). As long as a watcher
		1249	is pending (but not active) you must not call an init function on it (but
		1250	C<ev_TYPE_set> is safe), you must not change its priority, and you must
		1251	make sure the watcher is available to libev (e.g. you cannot C<free ()>
		1252	it).
		1253
		1254	=item callback ev_cb (ev_TYPE *watcher)
		1255
		1256	Returns the callback currently set on the watcher.
		1257
		1258	=item ev_cb_set (ev_TYPE *watcher, callback)
		1259
		1260	Change the callback. You can change the callback at virtually any time
		1261	(modulo threads).
		1262
		1263	=item ev_set_priority (ev_TYPE *watcher, int priority)
		1264
		1265	=item int ev_priority (ev_TYPE *watcher)
		1266
		1267	Set and query the priority of the watcher. The priority is a small
		1268	integer between C<EV_MAXPRI> (default: C<2>) and C<EV_MINPRI>
		1269	(default: C<-2>). Pending watchers with higher priority will be invoked
		1270	before watchers with lower priority, but priority will not keep watchers
		1271	from being executed (except for C<ev_idle> watchers).
		1272
		1273	If you need to suppress invocation when higher priority events are pending
		1274	you need to look at C<ev_idle> watchers, which provide this functionality.
		1275
		1276	You I<must not> change the priority of a watcher as long as it is active or
		1277	pending.
		1278
		1279	Setting a priority outside the range of C<EV_MINPRI> to C<EV_MAXPRI> is
		1280	fine, as long as you do not mind that the priority value you query might
		1281	or might not have been clamped to the valid range.
		1282
		1283	The default priority used by watchers when no priority has been set is
		1284	always C<0>, which is supposed to not be too high and not be too low :).
		1285
		1286	See L<WATCHER PRIORITY MODELS>, below, for a more thorough treatment of
		1287	priorities.
		1288
		1289	=item ev_invoke (loop, ev_TYPE *watcher, int revents)
		1290
		1291	Invoke the C<watcher> with the given C<loop> and C<revents>. Neither
		1292	C<loop> nor C<revents> need to be valid as long as the watcher callback
		1293	can deal with that fact, as both are simply passed through to the
		1294	callback.
		1295
		1296	=item int ev_clear_pending (loop, ev_TYPE *watcher)
		1297
		1298	If the watcher is pending, this function clears its pending status and
		1299	returns its C<revents> bitset (as if its callback was invoked). If the
		1300	watcher isn't pending it does nothing and returns C<0>.
		1301
		1302	Sometimes it can be useful to "poll" a watcher instead of waiting for its
		1303	callback to be invoked, which can be accomplished with this function.
		1304
		1305	=item ev_feed_event (loop, ev_TYPE *watcher, int revents)
		1306
		1307	Feeds the given event set into the event loop, as if the specified event
		1308	had happened for the specified watcher (which must be a pointer to an
		1309	initialised but not necessarily started event watcher). Obviously you must
		1310	not free the watcher as long as it has pending events.
		1311
		1312	Stopping the watcher, letting libev invoke it, or calling
		1313	C<ev_clear_pending> will clear the pending event, even if the watcher was
		1314	not started in the first place.
		1315
		1316	See also C<ev_feed_fd_event> and C<ev_feed_signal_event> for related
		1317	functions that do not need a watcher.
		1318
		1319	=back
		1320
582	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER	1321	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER
583		1322
584	Each watcher has, by default, a member C<void *data> that you can change	1323	Each watcher has, by default, a member C<void *data> that you can change
585	and read at any time, libev will completely ignore it. This can be used	1324	and read at any time: libev will completely ignore it. This can be used
586	to associate arbitrary data with your watcher. If you need more data and	1325	to associate arbitrary data with your watcher. If you need more data and
587	don't want to allocate memory and store a pointer to it in that data	1326	don't want to allocate memory and store a pointer to it in that data
588	member, you can also "subclass" the watcher type and provide your own	1327	member, you can also "subclass" the watcher type and provide your own
589	data:	1328	data:
590		1329
591	struct my_io	1330	struct my_io
592	{	1331	{
593	struct ev_io io;	1332	ev_io io;
594	int otherfd;	1333	int otherfd;
595	void *somedata;	1334	void *somedata;
596	struct whatever *mostinteresting;	1335	struct whatever *mostinteresting;
597	}	1336	};
		1337
		1338	...
		1339	struct my_io w;
		1340	ev_io_init (&w.io, my_cb, fd, EV_READ);
598		1341
599	And since your callback will be called with a pointer to the watcher, you	1342	And since your callback will be called with a pointer to the watcher, you
600	can cast it back to your own type:	1343	can cast it back to your own type:
601		1344
602	static void my_cb (struct ev_loop *loop, struct ev_io *w_, int revents)	1345	static void my_cb (struct ev_loop *loop, ev_io *w_, int revents)
603	{	1346	{
604	struct my_io *w = (struct my_io *)w_;	1347	struct my_io *w = (struct my_io *)w_;
605	...	1348	...
606	}	1349	}
607		1350
608	More interesting and less C-conformant ways of catsing your callback type	1351	More interesting and less C-conformant ways of casting your callback type
609	have been omitted....	1352	instead have been omitted.
		1353
		1354	Another common scenario is to use some data structure with multiple
		1355	embedded watchers:
		1356
		1357	struct my_biggy
		1358	{
		1359	int some_data;
		1360	ev_timer t1;
		1361	ev_timer t2;
		1362	}
		1363
		1364	In this case getting the pointer to C<my_biggy> is a bit more
		1365	complicated: Either you store the address of your C<my_biggy> struct
		1366	in the C<data> member of the watcher (for woozies), or you need to use
		1367	some pointer arithmetic using C<offsetof> inside your watchers (for real
		1368	programmers):
		1369
		1370	#include <stddef.h>
		1371
		1372	static void
		1373	t1_cb (EV_P_ ev_timer *w, int revents)
		1374	{
		1375	struct my_biggy big = (struct my_biggy *)
		1376	(((char *)w) - offsetof (struct my_biggy, t1));
		1377	}
		1378
		1379	static void
		1380	t2_cb (EV_P_ ev_timer *w, int revents)
		1381	{
		1382	struct my_biggy big = (struct my_biggy *)
		1383	(((char *)w) - offsetof (struct my_biggy, t2));
		1384	}
		1385
		1386	=head2 WATCHER STATES
		1387
		1388	There are various watcher states mentioned throughout this manual -
		1389	active, pending and so on. In this section these states and the rules to
		1390	transition between them will be described in more detail - and while these
		1391	rules might look complicated, they usually do "the right thing".
		1392
		1393	=over 4
		1394
		1395	=item initialiased
		1396
		1397	Before a watcher can be registered with the event looop it has to be
		1398	initialised. This can be done with a call to C<ev_TYPE_init>, or calls to
		1399	C<ev_init> followed by the watcher-specific C<ev_TYPE_set> function.
		1400
		1401	In this state it is simply some block of memory that is suitable for use
		1402	in an event loop. It can be moved around, freed, reused etc. at will.
		1403
		1404	=item started/running/active
		1405
		1406	Once a watcher has been started with a call to C<ev_TYPE_start> it becomes
		1407	property of the event loop, and is actively waiting for events. While in
		1408	this state it cannot be accessed (except in a few documented ways), moved,
		1409	freed or anything else - the only legal thing is to keep a pointer to it,
		1410	and call libev functions on it that are documented to work on active watchers.
		1411
		1412	=item pending
		1413
		1414	If a watcher is active and libev determines that an event it is interested
		1415	in has occurred (such as a timer expiring), it will become pending. It will
		1416	stay in this pending state until either it is stopped or its callback is
		1417	about to be invoked, so it is not normally pending inside the watcher
		1418	callback.
		1419
		1420	The watcher might or might not be active while it is pending (for example,
		1421	an expired non-repeating timer can be pending but no longer active). If it
		1422	is stopped, it can be freely accessed (e.g. by calling C<ev_TYPE_set>),
		1423	but it is still property of the event loop at this time, so cannot be
		1424	moved, freed or reused. And if it is active the rules described in the
		1425	previous item still apply.
		1426
		1427	It is also possible to feed an event on a watcher that is not active (e.g.
		1428	via C<ev_feed_event>), in which case it becomes pending without being
		1429	active.
		1430
		1431	=item stopped
		1432
		1433	A watcher can be stopped implicitly by libev (in which case it might still
		1434	be pending), or explicitly by calling its C<ev_TYPE_stop> function. The
		1435	latter will clear any pending state the watcher might be in, regardless
		1436	of whether it was active or not, so stopping a watcher explicitly before
		1437	freeing it is often a good idea.
		1438
		1439	While stopped (and not pending) the watcher is essentially in the
		1440	initialised state, that is it can be reused, moved, modified in any way
		1441	you wish.
		1442
		1443	=back
		1444
		1445	=head2 WATCHER PRIORITY MODELS
		1446
		1447	Many event loops support I<watcher priorities>, which are usually small
		1448	integers that influence the ordering of event callback invocation
		1449	between watchers in some way, all else being equal.
		1450
		1451	In libev, Watcher priorities can be set using C<ev_set_priority>. See its
		1452	description for the more technical details such as the actual priority
		1453	range.
		1454
		1455	There are two common ways how these these priorities are being interpreted
		1456	by event loops:
		1457
		1458	In the more common lock-out model, higher priorities "lock out" invocation
		1459	of lower priority watchers, which means as long as higher priority
		1460	watchers receive events, lower priority watchers are not being invoked.
		1461
		1462	The less common only-for-ordering model uses priorities solely to order
		1463	callback invocation within a single event loop iteration: Higher priority
		1464	watchers are invoked before lower priority ones, but they all get invoked
		1465	before polling for new events.
		1466
		1467	Libev uses the second (only-for-ordering) model for all its watchers
		1468	except for idle watchers (which use the lock-out model).
		1469
		1470	The rationale behind this is that implementing the lock-out model for
		1471	watchers is not well supported by most kernel interfaces, and most event
		1472	libraries will just poll for the same events again and again as long as
		1473	their callbacks have not been executed, which is very inefficient in the
		1474	common case of one high-priority watcher locking out a mass of lower
		1475	priority ones.
		1476
		1477	Static (ordering) priorities are most useful when you have two or more
		1478	watchers handling the same resource: a typical usage example is having an
		1479	C<ev_io> watcher to receive data, and an associated C<ev_timer> to handle
		1480	timeouts. Under load, data might be received while the program handles
		1481	other jobs, but since timers normally get invoked first, the timeout
		1482	handler will be executed before checking for data. In that case, giving
		1483	the timer a lower priority than the I/O watcher ensures that I/O will be
		1484	handled first even under adverse conditions (which is usually, but not
		1485	always, what you want).
		1486
		1487	Since idle watchers use the "lock-out" model, meaning that idle watchers
		1488	will only be executed when no same or higher priority watchers have
		1489	received events, they can be used to implement the "lock-out" model when
		1490	required.
		1491
		1492	For example, to emulate how many other event libraries handle priorities,
		1493	you can associate an C<ev_idle> watcher to each such watcher, and in
		1494	the normal watcher callback, you just start the idle watcher. The real
		1495	processing is done in the idle watcher callback. This causes libev to
		1496	continuously poll and process kernel event data for the watcher, but when
		1497	the lock-out case is known to be rare (which in turn is rare :), this is
		1498	workable.
		1499
		1500	Usually, however, the lock-out model implemented that way will perform
		1501	miserably under the type of load it was designed to handle. In that case,
		1502	it might be preferable to stop the real watcher before starting the
		1503	idle watcher, so the kernel will not have to process the event in case
		1504	the actual processing will be delayed for considerable time.
		1505
		1506	Here is an example of an I/O watcher that should run at a strictly lower
		1507	priority than the default, and which should only process data when no
		1508	other events are pending:
		1509
		1510	ev_idle idle; // actual processing watcher
		1511	ev_io io; // actual event watcher
		1512
		1513	static void
		1514	io_cb (EV_P_ ev_io *w, int revents)
		1515	{
		1516	// stop the I/O watcher, we received the event, but
		1517	// are not yet ready to handle it.
		1518	ev_io_stop (EV_A_ w);
		1519
		1520	// start the idle watcher to handle the actual event.
		1521	// it will not be executed as long as other watchers
		1522	// with the default priority are receiving events.
		1523	ev_idle_start (EV_A_ &idle);
		1524	}
		1525
		1526	static void
		1527	idle_cb (EV_P_ ev_idle *w, int revents)
		1528	{
		1529	// actual processing
		1530	read (STDIN_FILENO, ...);
		1531
		1532	// have to start the I/O watcher again, as
		1533	// we have handled the event
		1534	ev_io_start (EV_P_ &io);
		1535	}
		1536
		1537	// initialisation
		1538	ev_idle_init (&idle, idle_cb);
		1539	ev_io_init (&io, io_cb, STDIN_FILENO, EV_READ);
		1540	ev_io_start (EV_DEFAULT_ &io);
		1541
		1542	In the "real" world, it might also be beneficial to start a timer, so that
		1543	low-priority connections can not be locked out forever under load. This
		1544	enables your program to keep a lower latency for important connections
		1545	during short periods of high load, while not completely locking out less
		1546	important ones.
610		1547
611		1548
612	=head1 WATCHER TYPES	1549	=head1 WATCHER TYPES
613		1550
614	This section describes each watcher in detail, but will not repeat	1551	This section describes each watcher in detail, but will not repeat
615	information given in the last section.	1552	information given in the last section. Any initialisation/set macros,
		1553	functions and members specific to the watcher type are explained.
616		1554
		1555	Members are additionally marked with either I<[read-only]>, meaning that,
		1556	while the watcher is active, you can look at the member and expect some
		1557	sensible content, but you must not modify it (you can modify it while the
		1558	watcher is stopped to your hearts content), or I<[read-write]>, which
		1559	means you can expect it to have some sensible content while the watcher
		1560	is active, but you can also modify it. Modifying it may not do something
		1561	sensible or take immediate effect (or do anything at all), but libev will
		1562	not crash or malfunction in any way.
617		1563
		1564
618	=head2 C<ev_io> - is this file descriptor readable or writable	1565	=head2 C<ev_io> - is this file descriptor readable or writable?
619		1566
620	I/O watchers check whether a file descriptor is readable or writable	1567	I/O watchers check whether a file descriptor is readable or writable
621	in each iteration of the event loop (This behaviour is called	1568	in each iteration of the event loop, or, more precisely, when reading
622	level-triggering because you keep receiving events as long as the	1569	would not block the process and writing would at least be able to write
623	condition persists. Remember you can stop the watcher if you don't want to	1570	some data. This behaviour is called level-triggering because you keep
624	act on the event and neither want to receive future events).	1571	receiving events as long as the condition persists. Remember you can stop
		1572	the watcher if you don't want to act on the event and neither want to
		1573	receive future events.
625		1574
626	In general you can register as many read and/or write event watchers per	1575	In general you can register as many read and/or write event watchers per
627	fd as you want (as long as you don't confuse yourself). Setting all file	1576	fd as you want (as long as you don't confuse yourself). Setting all file
628	descriptors to non-blocking mode is also usually a good idea (but not	1577	descriptors to non-blocking mode is also usually a good idea (but not
629	required if you know what you are doing).	1578	required if you know what you are doing).
630		1579
631	You have to be careful with dup'ed file descriptors, though. Some backends	1580	If you cannot use non-blocking mode, then force the use of a
632	(the linux epoll backend is a notable example) cannot handle dup'ed file	1581	known-to-be-good backend (at the time of this writing, this includes only
633	descriptors correctly if you register interest in two or more fds pointing	1582	C<EVBACKEND_SELECT> and C<EVBACKEND_POLL>). The same applies to file
634	to the same underlying file/socket etc. description (that is, they share	1583	descriptors for which non-blocking operation makes no sense (such as
635	the same underlying "file open").	1584	files) - libev doesn't guarantee any specific behaviour in that case.
636		1585
637	If you must do this, then force the use of a known-to-be-good backend	1586	Another thing you have to watch out for is that it is quite easy to
638	(at the time of this writing, this includes only C<EVBACKEND_SELECT> and	1587	receive "spurious" readiness notifications, that is your callback might
		1588	be called with C<EV_READ> but a subsequent C<read>(2) will actually block
		1589	because there is no data. Not only are some backends known to create a
		1590	lot of those (for example Solaris ports), it is very easy to get into
		1591	this situation even with a relatively standard program structure. Thus
		1592	it is best to always use non-blocking I/O: An extra C<read>(2) returning
		1593	C<EAGAIN> is far preferable to a program hanging until some data arrives.
		1594
		1595	If you cannot run the fd in non-blocking mode (for example you should
		1596	not play around with an Xlib connection), then you have to separately
		1597	re-test whether a file descriptor is really ready with a known-to-be good
		1598	interface such as poll (fortunately in our Xlib example, Xlib already
		1599	does this on its own, so its quite safe to use). Some people additionally
		1600	use C<SIGALRM> and an interval timer, just to be sure you won't block
		1601	indefinitely.
		1602
		1603	But really, best use non-blocking mode.
		1604
		1605	=head3 The special problem of disappearing file descriptors
		1606
		1607	Some backends (e.g. kqueue, epoll) need to be told about closing a file
		1608	descriptor (either due to calling C<close> explicitly or any other means,
		1609	such as C<dup2>). The reason is that you register interest in some file
		1610	descriptor, but when it goes away, the operating system will silently drop
		1611	this interest. If another file descriptor with the same number then is
		1612	registered with libev, there is no efficient way to see that this is, in
		1613	fact, a different file descriptor.
		1614
		1615	To avoid having to explicitly tell libev about such cases, libev follows
		1616	the following policy: Each time C<ev_io_set> is being called, libev
		1617	will assume that this is potentially a new file descriptor, otherwise
		1618	it is assumed that the file descriptor stays the same. That means that
		1619	you I<have> to call C<ev_io_set> (or C<ev_io_init>) when you change the
		1620	descriptor even if the file descriptor number itself did not change.
		1621
		1622	This is how one would do it normally anyway, the important point is that
		1623	the libev application should not optimise around libev but should leave
		1624	optimisations to libev.
		1625
		1626	=head3 The special problem of dup'ed file descriptors
		1627
		1628	Some backends (e.g. epoll), cannot register events for file descriptors,
		1629	but only events for the underlying file descriptions. That means when you
		1630	have C<dup ()>'ed file descriptors or weirder constellations, and register
		1631	events for them, only one file descriptor might actually receive events.
		1632
		1633	There is no workaround possible except not registering events
		1634	for potentially C<dup ()>'ed file descriptors, or to resort to
		1635	C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>.
		1636
		1637	=head3 The special problem of fork
		1638
		1639	Some backends (epoll, kqueue) do not support C<fork ()> at all or exhibit
		1640	useless behaviour. Libev fully supports fork, but needs to be told about
		1641	it in the child.
		1642
		1643	To support fork in your programs, you either have to call
		1644	C<ev_default_fork ()> or C<ev_loop_fork ()> after a fork in the child,
		1645	enable C<EVFLAG_FORKCHECK>, or resort to C<EVBACKEND_SELECT> or
639	C<EVBACKEND_POLL>).	1646	C<EVBACKEND_POLL>.
		1647
		1648	=head3 The special problem of SIGPIPE
		1649
		1650	While not really specific to libev, it is easy to forget about C<SIGPIPE>:
		1651	when writing to a pipe whose other end has been closed, your program gets
		1652	sent a SIGPIPE, which, by default, aborts your program. For most programs
		1653	this is sensible behaviour, for daemons, this is usually undesirable.
		1654
		1655	So when you encounter spurious, unexplained daemon exits, make sure you
		1656	ignore SIGPIPE (and maybe make sure you log the exit status of your daemon
		1657	somewhere, as that would have given you a big clue).
		1658
		1659	=head3 The special problem of accept()ing when you can't
		1660
		1661	Many implementations of the POSIX C<accept> function (for example,
		1662	found in post-2004 Linux) have the peculiar behaviour of not removing a
		1663	connection from the pending queue in all error cases.
		1664
		1665	For example, larger servers often run out of file descriptors (because
		1666	of resource limits), causing C<accept> to fail with C<ENFILE> but not
		1667	rejecting the connection, leading to libev signalling readiness on
		1668	the next iteration again (the connection still exists after all), and
		1669	typically causing the program to loop at 100% CPU usage.
		1670
		1671	Unfortunately, the set of errors that cause this issue differs between
		1672	operating systems, there is usually little the app can do to remedy the
		1673	situation, and no known thread-safe method of removing the connection to
		1674	cope with overload is known (to me).
		1675
		1676	One of the easiest ways to handle this situation is to just ignore it
		1677	- when the program encounters an overload, it will just loop until the
		1678	situation is over. While this is a form of busy waiting, no OS offers an
		1679	event-based way to handle this situation, so it's the best one can do.
		1680
		1681	A better way to handle the situation is to log any errors other than
		1682	C<EAGAIN> and C<EWOULDBLOCK>, making sure not to flood the log with such
		1683	messages, and continue as usual, which at least gives the user an idea of
		1684	what could be wrong ("raise the ulimit!"). For extra points one could stop
		1685	the C<ev_io> watcher on the listening fd "for a while", which reduces CPU
		1686	usage.
		1687
		1688	If your program is single-threaded, then you could also keep a dummy file
		1689	descriptor for overload situations (e.g. by opening F</dev/null>), and
		1690	when you run into C<ENFILE> or C<EMFILE>, close it, run C<accept>,
		1691	close that fd, and create a new dummy fd. This will gracefully refuse
		1692	clients under typical overload conditions.
		1693
		1694	The last way to handle it is to simply log the error and C<exit>, as
		1695	is often done with C<malloc> failures, but this results in an easy
		1696	opportunity for a DoS attack.
		1697
		1698	=head3 Watcher-Specific Functions
640		1699
641	=over 4	1700	=over 4
642		1701
643	=item ev_io_init (ev_io *, callback, int fd, int events)	1702	=item ev_io_init (ev_io *, callback, int fd, int events)
644		1703
645	=item ev_io_set (ev_io *, int fd, int events)	1704	=item ev_io_set (ev_io *, int fd, int events)
646		1705
647	Configures an C<ev_io> watcher. The fd is the file descriptor to rceeive	1706	Configures an C<ev_io> watcher. The C<fd> is the file descriptor to
648	events for and events is either C<EV_READ>, C<EV_WRITE> or C<EV_READ |	1707	receive events for and C<events> is either C<EV_READ>, C<EV_WRITE> or
649	EV_WRITE> to receive the given events.	1708	C<EV_READ | EV_WRITE>, to express the desire to receive the given events.
650		1709
651	Please note that most of the more scalable backend mechanisms (for example	1710	=item int fd [read-only]
652	epoll and solaris ports) can result in spurious readyness notifications	1711
653	for file descriptors, so you practically need to use non-blocking I/O (and	1712	The file descriptor being watched.
654	treat callback invocation as hint only), or retest separately with a safe	1713
655	interface before doing I/O (XLib can do this), or force the use of either	1714	=item int events [read-only]
656	C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>, which don't suffer from this	1715
657	problem. Also note that it is quite easy to have your callback invoked	1716	The events being watched.
658	when the readyness condition is no longer valid even when employing
659	typical ways of handling events, so its a good idea to use non-blocking
660	I/O unconditionally.
661		1717
662	=back	1718	=back
663		1719
		1720	=head3 Examples
		1721
664	Example: call C<stdin_readable_cb> when STDIN_FILENO has become, well	1722	Example: Call C<stdin_readable_cb> when STDIN_FILENO has become, well
665	readable, but only once. Since it is likely line-buffered, you could	1723	readable, but only once. Since it is likely line-buffered, you could
666	attempt to read a whole line in the callback:	1724	attempt to read a whole line in the callback.
667		1725
668	static void	1726	static void
669	stdin_readable_cb (struct ev_loop *loop, struct ev_io *w, int revents)	1727	stdin_readable_cb (struct ev_loop *loop, ev_io *w, int revents)
670	{	1728	{
671	ev_io_stop (loop, w);	1729	ev_io_stop (loop, w);
672	.. read from stdin here (or from w->fd) and haqndle any I/O errors	1730	.. read from stdin here (or from w->fd) and handle any I/O errors
673	}	1731	}
674		1732
675	...	1733	...
676	struct ev_loop *loop = ev_default_init (0);	1734	struct ev_loop *loop = ev_default_init (0);
677	struct ev_io stdin_readable;	1735	ev_io stdin_readable;
678	ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);	1736	ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
679	ev_io_start (loop, &stdin_readable);	1737	ev_io_start (loop, &stdin_readable);
680	ev_loop (loop, 0);	1738	ev_run (loop, 0);
681		1739
682		1740
683	=head2 C<ev_timer> - relative and optionally recurring timeouts	1741	=head2 C<ev_timer> - relative and optionally repeating timeouts
684		1742
685	Timer watchers are simple relative timers that generate an event after a	1743	Timer watchers are simple relative timers that generate an event after a
686	given time, and optionally repeating in regular intervals after that.	1744	given time, and optionally repeating in regular intervals after that.
687		1745
688	The timers are based on real time, that is, if you register an event that	1746	The timers are based on real time, that is, if you register an event that
689	times out after an hour and you reset your system clock to last years	1747	times out after an hour and you reset your system clock to January last
690	time, it will still time out after (roughly) and hour. "Roughly" because	1748	year, it will still time out after (roughly) one hour. "Roughly" because
691	detecting time jumps is hard, and some inaccuracies are unavoidable (the	1749	detecting time jumps is hard, and some inaccuracies are unavoidable (the
692	monotonic clock option helps a lot here).	1750	monotonic clock option helps a lot here).
		1751
		1752	The callback is guaranteed to be invoked only I<after> its timeout has
		1753	passed (not I<at>, so on systems with very low-resolution clocks this
		1754	might introduce a small delay). If multiple timers become ready during the
		1755	same loop iteration then the ones with earlier time-out values are invoked
		1756	before ones of the same priority with later time-out values (but this is
		1757	no longer true when a callback calls C<ev_run> recursively).
		1758
		1759	=head3 Be smart about timeouts
		1760
		1761	Many real-world problems involve some kind of timeout, usually for error
		1762	recovery. A typical example is an HTTP request - if the other side hangs,
		1763	you want to raise some error after a while.
		1764
		1765	What follows are some ways to handle this problem, from obvious and
		1766	inefficient to smart and efficient.
		1767
		1768	In the following, a 60 second activity timeout is assumed - a timeout that
		1769	gets reset to 60 seconds each time there is activity (e.g. each time some
		1770	data or other life sign was received).
		1771
		1772	=over 4
		1773
		1774	=item 1. Use a timer and stop, reinitialise and start it on activity.
		1775
		1776	This is the most obvious, but not the most simple way: In the beginning,
		1777	start the watcher:
		1778
		1779	ev_timer_init (timer, callback, 60., 0.);
		1780	ev_timer_start (loop, timer);
		1781
		1782	Then, each time there is some activity, C<ev_timer_stop> it, initialise it
		1783	and start it again:
		1784
		1785	ev_timer_stop (loop, timer);
		1786	ev_timer_set (timer, 60., 0.);
		1787	ev_timer_start (loop, timer);
		1788
		1789	This is relatively simple to implement, but means that each time there is
		1790	some activity, libev will first have to remove the timer from its internal
		1791	data structure and then add it again. Libev tries to be fast, but it's
		1792	still not a constant-time operation.
		1793
		1794	=item 2. Use a timer and re-start it with C<ev_timer_again> inactivity.
		1795
		1796	This is the easiest way, and involves using C<ev_timer_again> instead of
		1797	C<ev_timer_start>.
		1798
		1799	To implement this, configure an C<ev_timer> with a C<repeat> value
		1800	of C<60> and then call C<ev_timer_again> at start and each time you
		1801	successfully read or write some data. If you go into an idle state where
		1802	you do not expect data to travel on the socket, you can C<ev_timer_stop>
		1803	the timer, and C<ev_timer_again> will automatically restart it if need be.
		1804
		1805	That means you can ignore both the C<ev_timer_start> function and the
		1806	C<after> argument to C<ev_timer_set>, and only ever use the C<repeat>
		1807	member and C<ev_timer_again>.
		1808
		1809	At start:
		1810
		1811	ev_init (timer, callback);
		1812	timer->repeat = 60.;
		1813	ev_timer_again (loop, timer);
		1814
		1815	Each time there is some activity:
		1816
		1817	ev_timer_again (loop, timer);
		1818
		1819	It is even possible to change the time-out on the fly, regardless of
		1820	whether the watcher is active or not:
		1821
		1822	timer->repeat = 30.;
		1823	ev_timer_again (loop, timer);
		1824
		1825	This is slightly more efficient then stopping/starting the timer each time
		1826	you want to modify its timeout value, as libev does not have to completely
		1827	remove and re-insert the timer from/into its internal data structure.
		1828
		1829	It is, however, even simpler than the "obvious" way to do it.
		1830
		1831	=item 3. Let the timer time out, but then re-arm it as required.
		1832
		1833	This method is more tricky, but usually most efficient: Most timeouts are
		1834	relatively long compared to the intervals between other activity - in
		1835	our example, within 60 seconds, there are usually many I/O events with
		1836	associated activity resets.
		1837
		1838	In this case, it would be more efficient to leave the C<ev_timer> alone,
		1839	but remember the time of last activity, and check for a real timeout only
		1840	within the callback:
		1841
		1842	ev_tstamp last_activity; // time of last activity
		1843
		1844	static void
		1845	callback (EV_P_ ev_timer *w, int revents)
		1846	{
		1847	ev_tstamp now = ev_now (EV_A);
		1848	ev_tstamp timeout = last_activity + 60.;
		1849
		1850	// if last_activity + 60. is older than now, we did time out
		1851	if (timeout < now)
		1852	{
		1853	// timeout occurred, take action
		1854	}
		1855	else
		1856	{
		1857	// callback was invoked, but there was some activity, re-arm
		1858	// the watcher to fire in last_activity + 60, which is
		1859	// guaranteed to be in the future, so "again" is positive:
		1860	w->repeat = timeout - now;
		1861	ev_timer_again (EV_A_ w);
		1862	}
		1863	}
		1864
		1865	To summarise the callback: first calculate the real timeout (defined
		1866	as "60 seconds after the last activity"), then check if that time has
		1867	been reached, which means something I<did>, in fact, time out. Otherwise
		1868	the callback was invoked too early (C<timeout> is in the future), so
		1869	re-schedule the timer to fire at that future time, to see if maybe we have
		1870	a timeout then.
		1871
		1872	Note how C<ev_timer_again> is used, taking advantage of the
		1873	C<ev_timer_again> optimisation when the timer is already running.
		1874
		1875	This scheme causes more callback invocations (about one every 60 seconds
		1876	minus half the average time between activity), but virtually no calls to
		1877	libev to change the timeout.
		1878
		1879	To start the timer, simply initialise the watcher and set C<last_activity>
		1880	to the current time (meaning we just have some activity :), then call the
		1881	callback, which will "do the right thing" and start the timer:
		1882
		1883	ev_init (timer, callback);
		1884	last_activity = ev_now (loop);
		1885	callback (loop, timer, EV_TIMER);
		1886
		1887	And when there is some activity, simply store the current time in
		1888	C<last_activity>, no libev calls at all:
		1889
		1890	last_activity = ev_now (loop);
		1891
		1892	This technique is slightly more complex, but in most cases where the
		1893	time-out is unlikely to be triggered, much more efficient.
		1894
		1895	Changing the timeout is trivial as well (if it isn't hard-coded in the
		1896	callback :) - just change the timeout and invoke the callback, which will
		1897	fix things for you.
		1898
		1899	=item 4. Wee, just use a double-linked list for your timeouts.
		1900
		1901	If there is not one request, but many thousands (millions...), all
		1902	employing some kind of timeout with the same timeout value, then one can
		1903	do even better:
		1904
		1905	When starting the timeout, calculate the timeout value and put the timeout
		1906	at the I<end> of the list.
		1907
		1908	Then use an C<ev_timer> to fire when the timeout at the I<beginning> of
		1909	the list is expected to fire (for example, using the technique #3).
		1910
		1911	When there is some activity, remove the timer from the list, recalculate
		1912	the timeout, append it to the end of the list again, and make sure to
		1913	update the C<ev_timer> if it was taken from the beginning of the list.
		1914
		1915	This way, one can manage an unlimited number of timeouts in O(1) time for
		1916	starting, stopping and updating the timers, at the expense of a major
		1917	complication, and having to use a constant timeout. The constant timeout
		1918	ensures that the list stays sorted.
		1919
		1920	=back
		1921
		1922	So which method the best?
		1923
		1924	Method #2 is a simple no-brain-required solution that is adequate in most
		1925	situations. Method #3 requires a bit more thinking, but handles many cases
		1926	better, and isn't very complicated either. In most case, choosing either
		1927	one is fine, with #3 being better in typical situations.
		1928
		1929	Method #1 is almost always a bad idea, and buys you nothing. Method #4 is
		1930	rather complicated, but extremely efficient, something that really pays
		1931	off after the first million or so of active timers, i.e. it's usually
		1932	overkill :)
		1933
		1934	=head3 The special problem of time updates
		1935
		1936	Establishing the current time is a costly operation (it usually takes at
		1937	least two system calls): EV therefore updates its idea of the current
		1938	time only before and after C<ev_run> collects new events, which causes a
		1939	growing difference between C<ev_now ()> and C<ev_time ()> when handling
		1940	lots of events in one iteration.
693		1941
694	The relative timeouts are calculated relative to the C<ev_now ()>	1942	The relative timeouts are calculated relative to the C<ev_now ()>
695	time. This is usually the right thing as this timestamp refers to the time	1943	time. This is usually the right thing as this timestamp refers to the time
696	of the event triggering whatever timeout you are modifying/starting. If	1944	of the event triggering whatever timeout you are modifying/starting. If
697	you suspect event processing to be delayed and you I<need> to base the timeout	1945	you suspect event processing to be delayed and you I<need> to base the
698	on the current time, use something like this to adjust for this:	1946	timeout on the current time, use something like this to adjust for this:
699		1947
700	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);	1948	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);
701		1949
702	The callback is guarenteed to be invoked only when its timeout has passed,	1950	If the event loop is suspended for a long time, you can also force an
703	but if multiple timers become ready during the same loop iteration then	1951	update of the time returned by C<ev_now ()> by calling C<ev_now_update
704	order of execution is undefined.	1952	()>.
		1953
		1954	=head3 The special problems of suspended animation
		1955
		1956	When you leave the server world it is quite customary to hit machines that
		1957	can suspend/hibernate - what happens to the clocks during such a suspend?
		1958
		1959	Some quick tests made with a Linux 2.6.28 indicate that a suspend freezes
		1960	all processes, while the clocks (C<times>, C<CLOCK_MONOTONIC>) continue
		1961	to run until the system is suspended, but they will not advance while the
		1962	system is suspended. That means, on resume, it will be as if the program
		1963	was frozen for a few seconds, but the suspend time will not be counted
		1964	towards C<ev_timer> when a monotonic clock source is used. The real time
		1965	clock advanced as expected, but if it is used as sole clocksource, then a
		1966	long suspend would be detected as a time jump by libev, and timers would
		1967	be adjusted accordingly.
		1968
		1969	I would not be surprised to see different behaviour in different between
		1970	operating systems, OS versions or even different hardware.
		1971
		1972	The other form of suspend (job control, or sending a SIGSTOP) will see a
		1973	time jump in the monotonic clocks and the realtime clock. If the program
		1974	is suspended for a very long time, and monotonic clock sources are in use,
		1975	then you can expect C<ev_timer>s to expire as the full suspension time
		1976	will be counted towards the timers. When no monotonic clock source is in
		1977	use, then libev will again assume a timejump and adjust accordingly.
		1978
		1979	It might be beneficial for this latter case to call C<ev_suspend>
		1980	and C<ev_resume> in code that handles C<SIGTSTP>, to at least get
		1981	deterministic behaviour in this case (you can do nothing against
		1982	C<SIGSTOP>).
		1983
		1984	=head3 Watcher-Specific Functions and Data Members
705		1985
706	=over 4	1986	=over 4
707		1987
708	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)	1988	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)
709		1989
710	=item ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)	1990	=item ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)
711		1991
712	Configure the timer to trigger after C<after> seconds. If C<repeat> is	1992	Configure the timer to trigger after C<after> seconds. If C<repeat>
713	C<0.>, then it will automatically be stopped. If it is positive, then the	1993	is C<0.>, then it will automatically be stopped once the timeout is
714	timer will automatically be configured to trigger again C<repeat> seconds	1994	reached. If it is positive, then the timer will automatically be
715	later, again, and again, until stopped manually.	1995	configured to trigger again C<repeat> seconds later, again, and again,
		1996	until stopped manually.
716		1997
717	The timer itself will do a best-effort at avoiding drift, that is, if you	1998	The timer itself will do a best-effort at avoiding drift, that is, if
718	configure a timer to trigger every 10 seconds, then it will trigger at	1999	you configure a timer to trigger every 10 seconds, then it will normally
719	exactly 10 second intervals. If, however, your program cannot keep up with	2000	trigger at exactly 10 second intervals. If, however, your program cannot
720	the timer (because it takes longer than those 10 seconds to do stuff) the	2001	keep up with the timer (because it takes longer than those 10 seconds to
721	timer will not fire more than once per event loop iteration.	2002	do stuff) the timer will not fire more than once per event loop iteration.
722		2003
723	=item ev_timer_again (loop)	2004	=item ev_timer_again (loop, ev_timer *)
724		2005
725	This will act as if the timer timed out and restart it again if it is	2006	This will act as if the timer timed out and restart it again if it is
726	repeating. The exact semantics are:	2007	repeating. The exact semantics are:
727		2008
		2009	If the timer is pending, its pending status is cleared.
		2010
728	If the timer is started but nonrepeating, stop it.	2011	If the timer is started but non-repeating, stop it (as if it timed out).
729		2012
730	If the timer is repeating, either start it if necessary (with the repeat	2013	If the timer is repeating, either start it if necessary (with the
731	value), or reset the running timer to the repeat value.	2014	C<repeat> value), or reset the running timer to the C<repeat> value.
732		2015
733	This sounds a bit complicated, but here is a useful and typical	2016	This sounds a bit complicated, see L<Be smart about timeouts>, above, for a
734	example: Imagine you have a tcp connection and you want a so-called idle	2017	usage example.
735	timeout, that is, you want to be called when there have been, say, 60	2018
736	seconds of inactivity on the socket. The easiest way to do this is to	2019	=item ev_tstamp ev_timer_remaining (loop, ev_timer *)
737	configure an C<ev_timer> with after=repeat=60 and calling ev_timer_again each	2020
738	time you successfully read or write some data. If you go into an idle	2021	Returns the remaining time until a timer fires. If the timer is active,
739	state where you do not expect data to travel on the socket, you can stop	2022	then this time is relative to the current event loop time, otherwise it's
740	the timer, and again will automatically restart it if need be.	2023	the timeout value currently configured.
		2024
		2025	That is, after an C<ev_timer_set (w, 5, 7)>, C<ev_timer_remaining> returns
		2026	C<5>. When the timer is started and one second passes, C<ev_timer_remaining>
		2027	will return C<4>. When the timer expires and is restarted, it will return
		2028	roughly C<7> (likely slightly less as callback invocation takes some time,
		2029	too), and so on.
		2030
		2031	=item ev_tstamp repeat [read-write]
		2032
		2033	The current C<repeat> value. Will be used each time the watcher times out
		2034	or C<ev_timer_again> is called, and determines the next timeout (if any),
		2035	which is also when any modifications are taken into account.
741		2036
742	=back	2037	=back
743		2038
		2039	=head3 Examples
		2040
744	Example: create a timer that fires after 60 seconds.	2041	Example: Create a timer that fires after 60 seconds.
745		2042
746	static void	2043	static void
747	one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)	2044	one_minute_cb (struct ev_loop *loop, ev_timer *w, int revents)
748	{	2045	{
749	.. one minute over, w is actually stopped right here	2046	.. one minute over, w is actually stopped right here
750	}	2047	}
751		2048
752	struct ev_timer mytimer;	2049	ev_timer mytimer;
753	ev_timer_init (&mytimer, one_minute_cb, 60., 0.);	2050	ev_timer_init (&mytimer, one_minute_cb, 60., 0.);
754	ev_timer_start (loop, &mytimer);	2051	ev_timer_start (loop, &mytimer);
755		2052
756	Example: create a timeout timer that times out after 10 seconds of	2053	Example: Create a timeout timer that times out after 10 seconds of
757	inactivity.	2054	inactivity.
758		2055
759	static void	2056	static void
760	timeout_cb (struct ev_loop *loop, struct ev_timer *w, int revents)	2057	timeout_cb (struct ev_loop *loop, ev_timer *w, int revents)
761	{	2058	{
762	.. ten seconds without any activity	2059	.. ten seconds without any activity
763	}	2060	}
764		2061
765	struct ev_timer mytimer;	2062	ev_timer mytimer;
766	ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */	2063	ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */
767	ev_timer_again (&mytimer); /* start timer */	2064	ev_timer_again (&mytimer); /* start timer */
768	ev_loop (loop, 0);	2065	ev_run (loop, 0);
769		2066
770	// and in some piece of code that gets executed on any "activity":	2067	// and in some piece of code that gets executed on any "activity":
771	// reset the timeout to start ticking again at 10 seconds	2068	// reset the timeout to start ticking again at 10 seconds
772	ev_timer_again (&mytimer);	2069	ev_timer_again (&mytimer);
773		2070
774		2071
775	=head2 C<ev_periodic> - to cron or not to cron	2072	=head2 C<ev_periodic> - to cron or not to cron?
776		2073
777	Periodic watchers are also timers of a kind, but they are very versatile	2074	Periodic watchers are also timers of a kind, but they are very versatile
778	(and unfortunately a bit complex).	2075	(and unfortunately a bit complex).
779		2076
780	Unlike C<ev_timer>'s, they are not based on real time (or relative time)	2077	Unlike C<ev_timer>, periodic watchers are not based on real time (or
781	but on wallclock time (absolute time). You can tell a periodic watcher	2078	relative time, the physical time that passes) but on wall clock time
782	to trigger "at" some specific point in time. For example, if you tell a	2079	(absolute time, the thing you can read on your calender or clock). The
783	periodic watcher to trigger in 10 seconds (by specifiying e.g. c<ev_now ()	2080	difference is that wall clock time can run faster or slower than real
784	+ 10.>) and then reset your system clock to the last year, then it will	2081	time, and time jumps are not uncommon (e.g. when you adjust your
785	take a year to trigger the event (unlike an C<ev_timer>, which would trigger	2082	wrist-watch).
786	roughly 10 seconds later and of course not if you reset your system time
787	again).
788		2083
789	They can also be used to implement vastly more complex timers, such as	2084	You can tell a periodic watcher to trigger after some specific point
		2085	in time: for example, if you tell a periodic watcher to trigger "in 10
		2086	seconds" (by specifying e.g. C<ev_now () + 10.>, that is, an absolute time
		2087	not a delay) and then reset your system clock to January of the previous
		2088	year, then it will take a year or more to trigger the event (unlike an
		2089	C<ev_timer>, which would still trigger roughly 10 seconds after starting
		2090	it, as it uses a relative timeout).
		2091
		2092	C<ev_periodic> watchers can also be used to implement vastly more complex
790	triggering an event on eahc midnight, local time.	2093	timers, such as triggering an event on each "midnight, local time", or
		2094	other complicated rules. This cannot be done with C<ev_timer> watchers, as
		2095	those cannot react to time jumps.
791		2096
792	As with timers, the callback is guarenteed to be invoked only when the	2097	As with timers, the callback is guaranteed to be invoked only when the
793	time (C<at>) has been passed, but if multiple periodic timers become ready	2098	point in time where it is supposed to trigger has passed. If multiple
794	during the same loop iteration then order of execution is undefined.	2099	timers become ready during the same loop iteration then the ones with
		2100	earlier time-out values are invoked before ones with later time-out values
		2101	(but this is no longer true when a callback calls C<ev_run> recursively).
		2102
		2103	=head3 Watcher-Specific Functions and Data Members
795		2104
796	=over 4	2105	=over 4
797		2106
798	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)	2107	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp offset, ev_tstamp interval, reschedule_cb)
799		2108
800	=item ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)	2109	=item ev_periodic_set (ev_periodic *, ev_tstamp offset, ev_tstamp interval, reschedule_cb)
801		2110
802	Lots of arguments, lets sort it out... There are basically three modes of	2111	Lots of arguments, let's sort it out... There are basically three modes of
803	operation, and we will explain them from simplest to complex:	2112	operation, and we will explain them from simplest to most complex:
804		2113
805	=over 4	2114	=over 4
806		2115
807	=item * absolute timer (interval = reschedule_cb = 0)	2116	=item * absolute timer (offset = absolute time, interval = 0, reschedule_cb = 0)
808		2117
809	In this configuration the watcher triggers an event at the wallclock time	2118	In this configuration the watcher triggers an event after the wall clock
810	C<at> and doesn't repeat. It will not adjust when a time jump occurs,	2119	time C<offset> has passed. It will not repeat and will not adjust when a
811	that is, if it is to be run at January 1st 2011 then it will run when the	2120	time jump occurs, that is, if it is to be run at January 1st 2011 then it
812	system time reaches or surpasses this time.	2121	will be stopped and invoked when the system clock reaches or surpasses
		2122	this point in time.
813		2123
814	=item * non-repeating interval timer (interval > 0, reschedule_cb = 0)	2124	=item * repeating interval timer (offset = offset within interval, interval > 0, reschedule_cb = 0)
815		2125
816	In this mode the watcher will always be scheduled to time out at the next	2126	In this mode the watcher will always be scheduled to time out at the next
817	C<at + N * interval> time (for some integer N) and then repeat, regardless	2127	C<offset + N * interval> time (for some integer N, which can also be
818	of any time jumps.	2128	negative) and then repeat, regardless of any time jumps. The C<offset>
		2129	argument is merely an offset into the C<interval> periods.
819		2130
820	This can be used to create timers that do not drift with respect to system	2131	This can be used to create timers that do not drift with respect to the
821	time:	2132	system clock, for example, here is an C<ev_periodic> that triggers each
		2133	hour, on the hour (with respect to UTC):
822		2134
823	ev_periodic_set (&periodic, 0., 3600., 0);	2135	ev_periodic_set (&periodic, 0., 3600., 0);
824		2136
825	This doesn't mean there will always be 3600 seconds in between triggers,	2137	This doesn't mean there will always be 3600 seconds in between triggers,
826	but only that the the callback will be called when the system time shows a	2138	but only that the callback will be called when the system time shows a
827	full hour (UTC), or more correctly, when the system time is evenly divisible	2139	full hour (UTC), or more correctly, when the system time is evenly divisible
828	by 3600.	2140	by 3600.
829		2141
830	Another way to think about it (for the mathematically inclined) is that	2142	Another way to think about it (for the mathematically inclined) is that
831	C<ev_periodic> will try to run the callback in this mode at the next possible	2143	C<ev_periodic> will try to run the callback in this mode at the next possible
832	time where C<time = at (mod interval)>, regardless of any time jumps.	2144	time where C<time = offset (mod interval)>, regardless of any time jumps.
833		2145
		2146	For numerical stability it is preferable that the C<offset> value is near
		2147	C<ev_now ()> (the current time), but there is no range requirement for
		2148	this value, and in fact is often specified as zero.
		2149
		2150	Note also that there is an upper limit to how often a timer can fire (CPU
		2151	speed for example), so if C<interval> is very small then timing stability
		2152	will of course deteriorate. Libev itself tries to be exact to be about one
		2153	millisecond (if the OS supports it and the machine is fast enough).
		2154
834	=item * manual reschedule mode (reschedule_cb = callback)	2155	=item * manual reschedule mode (offset ignored, interval ignored, reschedule_cb = callback)
835		2156
836	In this mode the values for C<interval> and C<at> are both being	2157	In this mode the values for C<interval> and C<offset> are both being
837	ignored. Instead, each time the periodic watcher gets scheduled, the	2158	ignored. Instead, each time the periodic watcher gets scheduled, the
838	reschedule callback will be called with the watcher as first, and the	2159	reschedule callback will be called with the watcher as first, and the
839	current time as second argument.	2160	current time as second argument.
840		2161
841	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher,	2162	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher, ever,
842	ever, or make any event loop modifications>. If you need to stop it,	2163	or make ANY other event loop modifications whatsoever, unless explicitly
843	return C<now + 1e30> (or so, fudge fudge) and stop it afterwards (e.g. by	2164	allowed by documentation here>.
844	starting a prepare watcher).
845		2165
		2166	If you need to stop it, return C<now + 1e30> (or so, fudge fudge) and stop
		2167	it afterwards (e.g. by starting an C<ev_prepare> watcher, which is the
		2168	only event loop modification you are allowed to do).
		2169
846	Its prototype is C<ev_tstamp (*reschedule_cb)(struct ev_periodic *w,	2170	The callback prototype is C<ev_tstamp (*reschedule_cb)(ev_periodic
847	ev_tstamp now)>, e.g.:	2171	*w, ev_tstamp now)>, e.g.:
848		2172
		2173	static ev_tstamp
849	static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)	2174	my_rescheduler (ev_periodic *w, ev_tstamp now)
850	{	2175	{
851	return now + 60.;	2176	return now + 60.;
852	}	2177	}
853		2178
854	It must return the next time to trigger, based on the passed time value	2179	It must return the next time to trigger, based on the passed time value
855	(that is, the lowest time value larger than to the second argument). It	2180	(that is, the lowest time value larger than to the second argument). It
856	will usually be called just before the callback will be triggered, but	2181	will usually be called just before the callback will be triggered, but
857	might be called at other times, too.	2182	might be called at other times, too.
858		2183
859	NOTE: I<< This callback must always return a time that is later than the	2184	NOTE: I<< This callback must always return a time that is higher than or
860	passed C<now> value >>. Not even C<now> itself will do, it I<must> be larger.	2185	equal to the passed C<now> value >>.
861		2186
862	This can be used to create very complex timers, such as a timer that	2187	This can be used to create very complex timers, such as a timer that
863	triggers on each midnight, local time. To do this, you would calculate the	2188	triggers on "next midnight, local time". To do this, you would calculate the
864	next midnight after C<now> and return the timestamp value for this. How	2189	next midnight after C<now> and return the timestamp value for this. How
865	you do this is, again, up to you (but it is not trivial, which is the main	2190	you do this is, again, up to you (but it is not trivial, which is the main
866	reason I omitted it as an example).	2191	reason I omitted it as an example).
867		2192
868	=back	2193	=back
…		…
872	Simply stops and restarts the periodic watcher again. This is only useful	2197	Simply stops and restarts the periodic watcher again. This is only useful
873	when you changed some parameters or the reschedule callback would return	2198	when you changed some parameters or the reschedule callback would return
874	a different time than the last time it was called (e.g. in a crond like	2199	a different time than the last time it was called (e.g. in a crond like
875	program when the crontabs have changed).	2200	program when the crontabs have changed).
876		2201
		2202	=item ev_tstamp ev_periodic_at (ev_periodic *)
		2203
		2204	When active, returns the absolute time that the watcher is supposed
		2205	to trigger next. This is not the same as the C<offset> argument to
		2206	C<ev_periodic_set>, but indeed works even in interval and manual
		2207	rescheduling modes.
		2208
		2209	=item ev_tstamp offset [read-write]
		2210
		2211	When repeating, this contains the offset value, otherwise this is the
		2212	absolute point in time (the C<offset> value passed to C<ev_periodic_set>,
		2213	although libev might modify this value for better numerical stability).
		2214
		2215	Can be modified any time, but changes only take effect when the periodic
		2216	timer fires or C<ev_periodic_again> is being called.
		2217
		2218	=item ev_tstamp interval [read-write]
		2219
		2220	The current interval value. Can be modified any time, but changes only
		2221	take effect when the periodic timer fires or C<ev_periodic_again> is being
		2222	called.
		2223
		2224	=item ev_tstamp (*reschedule_cb)(ev_periodic *w, ev_tstamp now) [read-write]
		2225
		2226	The current reschedule callback, or C<0>, if this functionality is
		2227	switched off. Can be changed any time, but changes only take effect when
		2228	the periodic timer fires or C<ev_periodic_again> is being called.
		2229
877	=back	2230	=back
878		2231
		2232	=head3 Examples
		2233
879	Example: call a callback every hour, or, more precisely, whenever the	2234	Example: Call a callback every hour, or, more precisely, whenever the
880	system clock is divisible by 3600. The callback invocation times have	2235	system time is divisible by 3600. The callback invocation times have
881	potentially a lot of jittering, but good long-term stability.	2236	potentially a lot of jitter, but good long-term stability.
882		2237
883	static void	2238	static void
884	clock_cb (struct ev_loop *loop, struct ev_io *w, int revents)	2239	clock_cb (struct ev_loop *loop, ev_periodic *w, int revents)
885	{	2240	{
886	... its now a full hour (UTC, or TAI or whatever your clock follows)	2241	... its now a full hour (UTC, or TAI or whatever your clock follows)
887	}	2242	}
888		2243
889	struct ev_periodic hourly_tick;	2244	ev_periodic hourly_tick;
890	ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);	2245	ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);
891	ev_periodic_start (loop, &hourly_tick);	2246	ev_periodic_start (loop, &hourly_tick);
892		2247
893	Example: the same as above, but use a reschedule callback to do it:	2248	Example: The same as above, but use a reschedule callback to do it:
894		2249
895	#include <math.h>	2250	#include <math.h>
896		2251
897	static ev_tstamp	2252	static ev_tstamp
898	my_scheduler_cb (struct ev_periodic *w, ev_tstamp now)	2253	my_scheduler_cb (ev_periodic *w, ev_tstamp now)
899	{	2254	{
900	return fmod (now, 3600.) + 3600.;	2255	return now + (3600. - fmod (now, 3600.));
901	}	2256	}
902		2257
903	ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);	2258	ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
904		2259
905	Example: call a callback every hour, starting now:	2260	Example: Call a callback every hour, starting now:
906		2261
907	struct ev_periodic hourly_tick;	2262	ev_periodic hourly_tick;
908	ev_periodic_init (&hourly_tick, clock_cb,	2263	ev_periodic_init (&hourly_tick, clock_cb,
909	fmod (ev_now (loop), 3600.), 3600., 0);	2264	fmod (ev_now (loop), 3600.), 3600., 0);
910	ev_periodic_start (loop, &hourly_tick);	2265	ev_periodic_start (loop, &hourly_tick);
911		2266
912		2267
913	=head2 C<ev_signal> - signal me when a signal gets signalled	2268	=head2 C<ev_signal> - signal me when a signal gets signalled!
914		2269
915	Signal watchers will trigger an event when the process receives a specific	2270	Signal watchers will trigger an event when the process receives a specific
916	signal one or more times. Even though signals are very asynchronous, libev	2271	signal one or more times. Even though signals are very asynchronous, libev
917	will try it's best to deliver signals synchronously, i.e. as part of the	2272	will try its best to deliver signals synchronously, i.e. as part of the
918	normal event processing, like any other event.	2273	normal event processing, like any other event.
919		2274
		2275	If you want signals to be delivered truly asynchronously, just use
		2276	C<sigaction> as you would do without libev and forget about sharing
		2277	the signal. You can even use C<ev_async> from a signal handler to
		2278	synchronously wake up an event loop.
		2279
920	You can configure as many watchers as you like per signal. Only when the	2280	You can configure as many watchers as you like for the same signal, but
		2281	only within the same loop, i.e. you can watch for C<SIGINT> in your
		2282	default loop and for C<SIGIO> in another loop, but you cannot watch for
		2283	C<SIGINT> in both the default loop and another loop at the same time. At
		2284	the moment, C<SIGCHLD> is permanently tied to the default loop.
		2285
921	first watcher gets started will libev actually register a signal watcher	2286	When the first watcher gets started will libev actually register something
922	with the kernel (thus it coexists with your own signal handlers as long	2287	with the kernel (thus it coexists with your own signal handlers as long as
923	as you don't register any with libev). Similarly, when the last signal	2288	you don't register any with libev for the same signal).
924	watcher for a signal is stopped libev will reset the signal handler to	2289
925	SIG_DFL (regardless of what it was set to before).	2290	If possible and supported, libev will install its handlers with
		2291	C<SA_RESTART> (or equivalent) behaviour enabled, so system calls should
		2292	not be unduly interrupted. If you have a problem with system calls getting
		2293	interrupted by signals you can block all signals in an C<ev_check> watcher
		2294	and unblock them in an C<ev_prepare> watcher.
		2295
		2296	=head3 The special problem of inheritance over fork/execve/pthread_create
		2297
		2298	Both the signal mask (C<sigprocmask>) and the signal disposition
		2299	(C<sigaction>) are unspecified after starting a signal watcher (and after
		2300	stopping it again), that is, libev might or might not block the signal,
		2301	and might or might not set or restore the installed signal handler.
		2302
		2303	While this does not matter for the signal disposition (libev never
		2304	sets signals to C<SIG_IGN>, so handlers will be reset to C<SIG_DFL> on
		2305	C<execve>), this matters for the signal mask: many programs do not expect
		2306	certain signals to be blocked.
		2307
		2308	This means that before calling C<exec> (from the child) you should reset
		2309	the signal mask to whatever "default" you expect (all clear is a good
		2310	choice usually).
		2311
		2312	The simplest way to ensure that the signal mask is reset in the child is
		2313	to install a fork handler with C<pthread_atfork> that resets it. That will
		2314	catch fork calls done by libraries (such as the libc) as well.
		2315
		2316	In current versions of libev, the signal will not be blocked indefinitely
		2317	unless you use the C<signalfd> API (C<EV_SIGNALFD>). While this reduces
		2318	the window of opportunity for problems, it will not go away, as libev
		2319	I<has> to modify the signal mask, at least temporarily.
		2320
		2321	So I can't stress this enough: I<If you do not reset your signal mask when
		2322	you expect it to be empty, you have a race condition in your code>. This
		2323	is not a libev-specific thing, this is true for most event libraries.
		2324
		2325	=head3 Watcher-Specific Functions and Data Members
926		2326
927	=over 4	2327	=over 4
928		2328
929	=item ev_signal_init (ev_signal *, callback, int signum)	2329	=item ev_signal_init (ev_signal *, callback, int signum)
930		2330
931	=item ev_signal_set (ev_signal *, int signum)	2331	=item ev_signal_set (ev_signal *, int signum)
932		2332
933	Configures the watcher to trigger on the given signal number (usually one	2333	Configures the watcher to trigger on the given signal number (usually one
934	of the C<SIGxxx> constants).	2334	of the C<SIGxxx> constants).
935		2335
		2336	=item int signum [read-only]
		2337
		2338	The signal the watcher watches out for.
		2339
936	=back	2340	=back
937		2341
		2342	=head3 Examples
938		2343
		2344	Example: Try to exit cleanly on SIGINT.
		2345
		2346	static void
		2347	sigint_cb (struct ev_loop *loop, ev_signal *w, int revents)
		2348	{
		2349	ev_break (loop, EVBREAK_ALL);
		2350	}
		2351
		2352	ev_signal signal_watcher;
		2353	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
		2354	ev_signal_start (loop, &signal_watcher);
		2355
		2356
939	=head2 C<ev_child> - wait for pid status changes	2357	=head2 C<ev_child> - watch out for process status changes
940		2358
941	Child watchers trigger when your process receives a SIGCHLD in response to	2359	Child watchers trigger when your process receives a SIGCHLD in response to
942	some child status changes (most typically when a child of yours dies).	2360	some child status changes (most typically when a child of yours dies or
		2361	exits). It is permissible to install a child watcher I<after> the child
		2362	has been forked (which implies it might have already exited), as long
		2363	as the event loop isn't entered (or is continued from a watcher), i.e.,
		2364	forking and then immediately registering a watcher for the child is fine,
		2365	but forking and registering a watcher a few event loop iterations later or
		2366	in the next callback invocation is not.
		2367
		2368	Only the default event loop is capable of handling signals, and therefore
		2369	you can only register child watchers in the default event loop.
		2370
		2371	Due to some design glitches inside libev, child watchers will always be
		2372	handled at maximum priority (their priority is set to C<EV_MAXPRI> by
		2373	libev)
		2374
		2375	=head3 Process Interaction
		2376
		2377	Libev grabs C<SIGCHLD> as soon as the default event loop is
		2378	initialised. This is necessary to guarantee proper behaviour even if the
		2379	first child watcher is started after the child exits. The occurrence
		2380	of C<SIGCHLD> is recorded asynchronously, but child reaping is done
		2381	synchronously as part of the event loop processing. Libev always reaps all
		2382	children, even ones not watched.
		2383
		2384	=head3 Overriding the Built-In Processing
		2385
		2386	Libev offers no special support for overriding the built-in child
		2387	processing, but if your application collides with libev's default child
		2388	handler, you can override it easily by installing your own handler for
		2389	C<SIGCHLD> after initialising the default loop, and making sure the
		2390	default loop never gets destroyed. You are encouraged, however, to use an
		2391	event-based approach to child reaping and thus use libev's support for
		2392	that, so other libev users can use C<ev_child> watchers freely.
		2393
		2394	=head3 Stopping the Child Watcher
		2395
		2396	Currently, the child watcher never gets stopped, even when the
		2397	child terminates, so normally one needs to stop the watcher in the
		2398	callback. Future versions of libev might stop the watcher automatically
		2399	when a child exit is detected (calling C<ev_child_stop> twice is not a
		2400	problem).
		2401
		2402	=head3 Watcher-Specific Functions and Data Members
943		2403
944	=over 4	2404	=over 4
945		2405
946	=item ev_child_init (ev_child *, callback, int pid)	2406	=item ev_child_init (ev_child *, callback, int pid, int trace)
947		2407
948	=item ev_child_set (ev_child *, int pid)	2408	=item ev_child_set (ev_child *, int pid, int trace)
949		2409
950	Configures the watcher to wait for status changes of process C<pid> (or	2410	Configures the watcher to wait for status changes of process C<pid> (or
951	I<any> process if C<pid> is specified as C<0>). The callback can look	2411	I<any> process if C<pid> is specified as C<0>). The callback can look
952	at the C<rstatus> member of the C<ev_child> watcher structure to see	2412	at the C<rstatus> member of the C<ev_child> watcher structure to see
953	the status word (use the macros from C<sys/wait.h> and see your systems	2413	the status word (use the macros from C<sys/wait.h> and see your systems
954	C<waitpid> documentation). The C<rpid> member contains the pid of the	2414	C<waitpid> documentation). The C<rpid> member contains the pid of the
955	process causing the status change.	2415	process causing the status change. C<trace> must be either C<0> (only
		2416	activate the watcher when the process terminates) or C<1> (additionally
		2417	activate the watcher when the process is stopped or continued).
		2418
		2419	=item int pid [read-only]
		2420
		2421	The process id this watcher watches out for, or C<0>, meaning any process id.
		2422
		2423	=item int rpid [read-write]
		2424
		2425	The process id that detected a status change.
		2426
		2427	=item int rstatus [read-write]
		2428
		2429	The process exit/trace status caused by C<rpid> (see your systems
		2430	C<waitpid> and C<sys/wait.h> documentation for details).
956		2431
957	=back	2432	=back
958		2433
959	Example: try to exit cleanly on SIGINT and SIGTERM.	2434	=head3 Examples
960		2435
		2436	Example: C<fork()> a new process and install a child handler to wait for
		2437	its completion.
		2438
		2439	ev_child cw;
		2440
961	static void	2441	static void
962	sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)	2442	child_cb (EV_P_ ev_child *w, int revents)
963	{	2443	{
964	ev_unloop (loop, EVUNLOOP_ALL);	2444	ev_child_stop (EV_A_ w);
		2445	printf ("process %d exited with status %x\n", w->rpid, w->rstatus);
965	}	2446	}
966		2447
967	struct ev_signal signal_watcher;	2448	pid_t pid = fork ();
968	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
969	ev_signal_start (loop, &sigint_cb);
970		2449
		2450	if (pid < 0)
		2451	// error
		2452	else if (pid == 0)
		2453	{
		2454	// the forked child executes here
		2455	exit (1);
		2456	}
		2457	else
		2458	{
		2459	ev_child_init (&cw, child_cb, pid, 0);
		2460	ev_child_start (EV_DEFAULT_ &cw);
		2461	}
971		2462
		2463
		2464	=head2 C<ev_stat> - did the file attributes just change?
		2465
		2466	This watches a file system path for attribute changes. That is, it calls
		2467	C<stat> on that path in regular intervals (or when the OS says it changed)
		2468	and sees if it changed compared to the last time, invoking the callback if
		2469	it did.
		2470
		2471	The path does not need to exist: changing from "path exists" to "path does
		2472	not exist" is a status change like any other. The condition "path does not
		2473	exist" (or more correctly "path cannot be stat'ed") is signified by the
		2474	C<st_nlink> field being zero (which is otherwise always forced to be at
		2475	least one) and all the other fields of the stat buffer having unspecified
		2476	contents.
		2477
		2478	The path I<must not> end in a slash or contain special components such as
		2479	C<.> or C<..>. The path I<should> be absolute: If it is relative and
		2480	your working directory changes, then the behaviour is undefined.
		2481
		2482	Since there is no portable change notification interface available, the
		2483	portable implementation simply calls C<stat(2)> regularly on the path
		2484	to see if it changed somehow. You can specify a recommended polling
		2485	interval for this case. If you specify a polling interval of C<0> (highly
		2486	recommended!) then a I<suitable, unspecified default> value will be used
		2487	(which you can expect to be around five seconds, although this might
		2488	change dynamically). Libev will also impose a minimum interval which is
		2489	currently around C<0.1>, but that's usually overkill.
		2490
		2491	This watcher type is not meant for massive numbers of stat watchers,
		2492	as even with OS-supported change notifications, this can be
		2493	resource-intensive.
		2494
		2495	At the time of this writing, the only OS-specific interface implemented
		2496	is the Linux inotify interface (implementing kqueue support is left as an
		2497	exercise for the reader. Note, however, that the author sees no way of
		2498	implementing C<ev_stat> semantics with kqueue, except as a hint).
		2499
		2500	=head3 ABI Issues (Largefile Support)
		2501
		2502	Libev by default (unless the user overrides this) uses the default
		2503	compilation environment, which means that on systems with large file
		2504	support disabled by default, you get the 32 bit version of the stat
		2505	structure. When using the library from programs that change the ABI to
		2506	use 64 bit file offsets the programs will fail. In that case you have to
		2507	compile libev with the same flags to get binary compatibility. This is
		2508	obviously the case with any flags that change the ABI, but the problem is
		2509	most noticeably displayed with ev_stat and large file support.
		2510
		2511	The solution for this is to lobby your distribution maker to make large
		2512	file interfaces available by default (as e.g. FreeBSD does) and not
		2513	optional. Libev cannot simply switch on large file support because it has
		2514	to exchange stat structures with application programs compiled using the
		2515	default compilation environment.
		2516
		2517	=head3 Inotify and Kqueue
		2518
		2519	When C<inotify (7)> support has been compiled into libev and present at
		2520	runtime, it will be used to speed up change detection where possible. The
		2521	inotify descriptor will be created lazily when the first C<ev_stat>
		2522	watcher is being started.
		2523
		2524	Inotify presence does not change the semantics of C<ev_stat> watchers
		2525	except that changes might be detected earlier, and in some cases, to avoid
		2526	making regular C<stat> calls. Even in the presence of inotify support
		2527	there are many cases where libev has to resort to regular C<stat> polling,
		2528	but as long as kernel 2.6.25 or newer is used (2.6.24 and older have too
		2529	many bugs), the path exists (i.e. stat succeeds), and the path resides on
		2530	a local filesystem (libev currently assumes only ext2/3, jfs, reiserfs and
		2531	xfs are fully working) libev usually gets away without polling.
		2532
		2533	There is no support for kqueue, as apparently it cannot be used to
		2534	implement this functionality, due to the requirement of having a file
		2535	descriptor open on the object at all times, and detecting renames, unlinks
		2536	etc. is difficult.
		2537
		2538	=head3 C<stat ()> is a synchronous operation
		2539
		2540	Libev doesn't normally do any kind of I/O itself, and so is not blocking
		2541	the process. The exception are C<ev_stat> watchers - those call C<stat
		2542	()>, which is a synchronous operation.
		2543
		2544	For local paths, this usually doesn't matter: unless the system is very
		2545	busy or the intervals between stat's are large, a stat call will be fast,
		2546	as the path data is usually in memory already (except when starting the
		2547	watcher).
		2548
		2549	For networked file systems, calling C<stat ()> can block an indefinite
		2550	time due to network issues, and even under good conditions, a stat call
		2551	often takes multiple milliseconds.
		2552
		2553	Therefore, it is best to avoid using C<ev_stat> watchers on networked
		2554	paths, although this is fully supported by libev.
		2555
		2556	=head3 The special problem of stat time resolution
		2557
		2558	The C<stat ()> system call only supports full-second resolution portably,
		2559	and even on systems where the resolution is higher, most file systems
		2560	still only support whole seconds.
		2561
		2562	That means that, if the time is the only thing that changes, you can
		2563	easily miss updates: on the first update, C<ev_stat> detects a change and
		2564	calls your callback, which does something. When there is another update
		2565	within the same second, C<ev_stat> will be unable to detect unless the
		2566	stat data does change in other ways (e.g. file size).
		2567
		2568	The solution to this is to delay acting on a change for slightly more
		2569	than a second (or till slightly after the next full second boundary), using
		2570	a roughly one-second-delay C<ev_timer> (e.g. C<ev_timer_set (w, 0., 1.02);
		2571	ev_timer_again (loop, w)>).
		2572
		2573	The C<.02> offset is added to work around small timing inconsistencies
		2574	of some operating systems (where the second counter of the current time
		2575	might be be delayed. One such system is the Linux kernel, where a call to
		2576	C<gettimeofday> might return a timestamp with a full second later than
		2577	a subsequent C<time> call - if the equivalent of C<time ()> is used to
		2578	update file times then there will be a small window where the kernel uses
		2579	the previous second to update file times but libev might already execute
		2580	the timer callback).
		2581
		2582	=head3 Watcher-Specific Functions and Data Members
		2583
		2584	=over 4
		2585
		2586	=item ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)
		2587
		2588	=item ev_stat_set (ev_stat *, const char *path, ev_tstamp interval)
		2589
		2590	Configures the watcher to wait for status changes of the given
		2591	C<path>. The C<interval> is a hint on how quickly a change is expected to
		2592	be detected and should normally be specified as C<0> to let libev choose
		2593	a suitable value. The memory pointed to by C<path> must point to the same
		2594	path for as long as the watcher is active.
		2595
		2596	The callback will receive an C<EV_STAT> event when a change was detected,
		2597	relative to the attributes at the time the watcher was started (or the
		2598	last change was detected).
		2599
		2600	=item ev_stat_stat (loop, ev_stat *)
		2601
		2602	Updates the stat buffer immediately with new values. If you change the
		2603	watched path in your callback, you could call this function to avoid
		2604	detecting this change (while introducing a race condition if you are not
		2605	the only one changing the path). Can also be useful simply to find out the
		2606	new values.
		2607
		2608	=item ev_statdata attr [read-only]
		2609
		2610	The most-recently detected attributes of the file. Although the type is
		2611	C<ev_statdata>, this is usually the (or one of the) C<struct stat> types
		2612	suitable for your system, but you can only rely on the POSIX-standardised
		2613	members to be present. If the C<st_nlink> member is C<0>, then there was
		2614	some error while C<stat>ing the file.
		2615
		2616	=item ev_statdata prev [read-only]
		2617
		2618	The previous attributes of the file. The callback gets invoked whenever
		2619	C<prev> != C<attr>, or, more precisely, one or more of these members
		2620	differ: C<st_dev>, C<st_ino>, C<st_mode>, C<st_nlink>, C<st_uid>,
		2621	C<st_gid>, C<st_rdev>, C<st_size>, C<st_atime>, C<st_mtime>, C<st_ctime>.
		2622
		2623	=item ev_tstamp interval [read-only]
		2624
		2625	The specified interval.
		2626
		2627	=item const char *path [read-only]
		2628
		2629	The file system path that is being watched.
		2630
		2631	=back
		2632
		2633	=head3 Examples
		2634
		2635	Example: Watch C</etc/passwd> for attribute changes.
		2636
		2637	static void
		2638	passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)
		2639	{
		2640	/* /etc/passwd changed in some way */
		2641	if (w->attr.st_nlink)
		2642	{
		2643	printf ("passwd current size %ld\n", (long)w->attr.st_size);
		2644	printf ("passwd current atime %ld\n", (long)w->attr.st_mtime);
		2645	printf ("passwd current mtime %ld\n", (long)w->attr.st_mtime);
		2646	}
		2647	else
		2648	/* you shalt not abuse printf for puts */
		2649	puts ("wow, /etc/passwd is not there, expect problems. "
		2650	"if this is windows, they already arrived\n");
		2651	}
		2652
		2653	...
		2654	ev_stat passwd;
		2655
		2656	ev_stat_init (&passwd, passwd_cb, "/etc/passwd", 0.);
		2657	ev_stat_start (loop, &passwd);
		2658
		2659	Example: Like above, but additionally use a one-second delay so we do not
		2660	miss updates (however, frequent updates will delay processing, too, so
		2661	one might do the work both on C<ev_stat> callback invocation I<and> on
		2662	C<ev_timer> callback invocation).
		2663
		2664	static ev_stat passwd;
		2665	static ev_timer timer;
		2666
		2667	static void
		2668	timer_cb (EV_P_ ev_timer *w, int revents)
		2669	{
		2670	ev_timer_stop (EV_A_ w);
		2671
		2672	/* now it's one second after the most recent passwd change */
		2673	}
		2674
		2675	static void
		2676	stat_cb (EV_P_ ev_stat *w, int revents)
		2677	{
		2678	/* reset the one-second timer */
		2679	ev_timer_again (EV_A_ &timer);
		2680	}
		2681
		2682	...
		2683	ev_stat_init (&passwd, stat_cb, "/etc/passwd", 0.);
		2684	ev_stat_start (loop, &passwd);
		2685	ev_timer_init (&timer, timer_cb, 0., 1.02);
		2686
		2687
972	=head2 C<ev_idle> - when you've got nothing better to do	2688	=head2 C<ev_idle> - when you've got nothing better to do...
973		2689
974	Idle watchers trigger events when there are no other events are pending	2690	Idle watchers trigger events when no other events of the same or higher
975	(prepare, check and other idle watchers do not count). That is, as long	2691	priority are pending (prepare, check and other idle watchers do not count
976	as your process is busy handling sockets or timeouts (or even signals,	2692	as receiving "events").
977	imagine) it will not be triggered. But when your process is idle all idle	2693
978	watchers are being called again and again, once per event loop iteration -	2694	That is, as long as your process is busy handling sockets or timeouts
		2695	(or even signals, imagine) of the same or higher priority it will not be
		2696	triggered. But when your process is idle (or only lower-priority watchers
		2697	are pending), the idle watchers are being called once per event loop
979	until stopped, that is, or your process receives more events and becomes	2698	iteration - until stopped, that is, or your process receives more events
980	busy.	2699	and becomes busy again with higher priority stuff.
981		2700
982	The most noteworthy effect is that as long as any idle watchers are	2701	The most noteworthy effect is that as long as any idle watchers are
983	active, the process will not block when waiting for new events.	2702	active, the process will not block when waiting for new events.
984		2703
985	Apart from keeping your process non-blocking (which is a useful	2704	Apart from keeping your process non-blocking (which is a useful
986	effect on its own sometimes), idle watchers are a good place to do	2705	effect on its own sometimes), idle watchers are a good place to do
987	"pseudo-background processing", or delay processing stuff to after the	2706	"pseudo-background processing", or delay processing stuff to after the
988	event loop has handled all outstanding events.	2707	event loop has handled all outstanding events.
989		2708
		2709	=head3 Watcher-Specific Functions and Data Members
		2710
990	=over 4	2711	=over 4
991		2712
992	=item ev_idle_init (ev_signal *, callback)	2713	=item ev_idle_init (ev_idle *, callback)
993		2714
994	Initialises and configures the idle watcher - it has no parameters of any	2715	Initialises and configures the idle watcher - it has no parameters of any
995	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,	2716	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,
996	believe me.	2717	believe me.
997		2718
998	=back	2719	=back
999		2720
		2721	=head3 Examples
		2722
1000	Example: dynamically allocate an C<ev_idle>, start it, and in the	2723	Example: Dynamically allocate an C<ev_idle> watcher, start it, and in the
1001	callback, free it. Alos, use no error checking, as usual.	2724	callback, free it. Also, use no error checking, as usual.
1002		2725
1003	static void	2726	static void
1004	idle_cb (struct ev_loop *loop, struct ev_idle *w, int revents)	2727	idle_cb (struct ev_loop *loop, ev_idle *w, int revents)
1005	{	2728	{
1006	free (w);	2729	free (w);
1007	// now do something you wanted to do when the program has	2730	// now do something you wanted to do when the program has
1008	// no longer asnything immediate to do.	2731	// no longer anything immediate to do.
1009	}	2732	}
1010		2733
1011	struct ev_idle *idle_watcher = malloc (sizeof (struct ev_idle));	2734	ev_idle *idle_watcher = malloc (sizeof (ev_idle));
1012	ev_idle_init (idle_watcher, idle_cb);	2735	ev_idle_init (idle_watcher, idle_cb);
1013	ev_idle_start (loop, idle_cb);	2736	ev_idle_start (loop, idle_watcher);
1014		2737
1015		2738
1016	=head2 C<ev_prepare> and C<ev_check> - customise your event loop	2739	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!
1017		2740
1018	Prepare and check watchers are usually (but not always) used in tandem:	2741	Prepare and check watchers are usually (but not always) used in pairs:
1019	prepare watchers get invoked before the process blocks and check watchers	2742	prepare watchers get invoked before the process blocks and check watchers
1020	afterwards.	2743	afterwards.
1021		2744
		2745	You I<must not> call C<ev_run> or similar functions that enter
		2746	the current event loop from either C<ev_prepare> or C<ev_check>
		2747	watchers. Other loops than the current one are fine, however. The
		2748	rationale behind this is that you do not need to check for recursion in
		2749	those watchers, i.e. the sequence will always be C<ev_prepare>, blocking,
		2750	C<ev_check> so if you have one watcher of each kind they will always be
		2751	called in pairs bracketing the blocking call.
		2752
1022	Their main purpose is to integrate other event mechanisms into libev and	2753	Their main purpose is to integrate other event mechanisms into libev and
1023	their use is somewhat advanced. This could be used, for example, to track	2754	their use is somewhat advanced. They could be used, for example, to track
1024	variable changes, implement your own watchers, integrate net-snmp or a	2755	variable changes, implement your own watchers, integrate net-snmp or a
1025	coroutine library and lots more.	2756	coroutine library and lots more. They are also occasionally useful if
		2757	you cache some data and want to flush it before blocking (for example,
		2758	in X programs you might want to do an C<XFlush ()> in an C<ev_prepare>
		2759	watcher).
1026		2760
1027	This is done by examining in each prepare call which file descriptors need	2761	This is done by examining in each prepare call which file descriptors
1028	to be watched by the other library, registering C<ev_io> watchers for	2762	need to be watched by the other library, registering C<ev_io> watchers
1029	them and starting an C<ev_timer> watcher for any timeouts (many libraries	2763	for them and starting an C<ev_timer> watcher for any timeouts (many
1030	provide just this functionality). Then, in the check watcher you check for	2764	libraries provide exactly this functionality). Then, in the check watcher,
1031	any events that occured (by checking the pending status of all watchers	2765	you check for any events that occurred (by checking the pending status
1032	and stopping them) and call back into the library. The I/O and timer	2766	of all watchers and stopping them) and call back into the library. The
1033	callbacks will never actually be called (but must be valid nevertheless,	2767	I/O and timer callbacks will never actually be called (but must be valid
1034	because you never know, you know?).	2768	nevertheless, because you never know, you know?).
1035		2769
1036	As another example, the Perl Coro module uses these hooks to integrate	2770	As another example, the Perl Coro module uses these hooks to integrate
1037	coroutines into libev programs, by yielding to other active coroutines	2771	coroutines into libev programs, by yielding to other active coroutines
1038	during each prepare and only letting the process block if no coroutines	2772	during each prepare and only letting the process block if no coroutines
1039	are ready to run (it's actually more complicated: it only runs coroutines	2773	are ready to run (it's actually more complicated: it only runs coroutines
1040	with priority higher than or equal to the event loop and one coroutine	2774	with priority higher than or equal to the event loop and one coroutine
1041	of lower priority, but only once, using idle watchers to keep the event	2775	of lower priority, but only once, using idle watchers to keep the event
1042	loop from blocking if lower-priority coroutines are active, thus mapping	2776	loop from blocking if lower-priority coroutines are active, thus mapping
1043	low-priority coroutines to idle/background tasks).	2777	low-priority coroutines to idle/background tasks).
1044		2778
		2779	It is recommended to give C<ev_check> watchers highest (C<EV_MAXPRI>)
		2780	priority, to ensure that they are being run before any other watchers
		2781	after the poll (this doesn't matter for C<ev_prepare> watchers).
		2782
		2783	Also, C<ev_check> watchers (and C<ev_prepare> watchers, too) should not
		2784	activate ("feed") events into libev. While libev fully supports this, they
		2785	might get executed before other C<ev_check> watchers did their job. As
		2786	C<ev_check> watchers are often used to embed other (non-libev) event
		2787	loops those other event loops might be in an unusable state until their
		2788	C<ev_check> watcher ran (always remind yourself to coexist peacefully with
		2789	others).
		2790
		2791	=head3 Watcher-Specific Functions and Data Members
		2792
1045	=over 4	2793	=over 4
1046		2794
1047	=item ev_prepare_init (ev_prepare *, callback)	2795	=item ev_prepare_init (ev_prepare *, callback)
1048		2796
1049	=item ev_check_init (ev_check *, callback)	2797	=item ev_check_init (ev_check *, callback)
1050		2798
1051	Initialises and configures the prepare or check watcher - they have no	2799	Initialises and configures the prepare or check watcher - they have no
1052	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>	2800	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>
1053	macros, but using them is utterly, utterly and completely pointless.	2801	macros, but using them is utterly, utterly, utterly and completely
		2802	pointless.
1054		2803
1055	=back	2804	=back
1056		2805
1057	Example: *TODO*.	2806	=head3 Examples
1058		2807
		2808	There are a number of principal ways to embed other event loops or modules
		2809	into libev. Here are some ideas on how to include libadns into libev
		2810	(there is a Perl module named C<EV::ADNS> that does this, which you could
		2811	use as a working example. Another Perl module named C<EV::Glib> embeds a
		2812	Glib main context into libev, and finally, C<Glib::EV> embeds EV into the
		2813	Glib event loop).
1059		2814
		2815	Method 1: Add IO watchers and a timeout watcher in a prepare handler,
		2816	and in a check watcher, destroy them and call into libadns. What follows
		2817	is pseudo-code only of course. This requires you to either use a low
		2818	priority for the check watcher or use C<ev_clear_pending> explicitly, as
		2819	the callbacks for the IO/timeout watchers might not have been called yet.
		2820
		2821	static ev_io iow [nfd];
		2822	static ev_timer tw;
		2823
		2824	static void
		2825	io_cb (struct ev_loop *loop, ev_io *w, int revents)
		2826	{
		2827	}
		2828
		2829	// create io watchers for each fd and a timer before blocking
		2830	static void
		2831	adns_prepare_cb (struct ev_loop *loop, ev_prepare *w, int revents)
		2832	{
		2833	int timeout = 3600000;
		2834	struct pollfd fds [nfd];
		2835	// actual code will need to loop here and realloc etc.
		2836	adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));
		2837
		2838	/* the callback is illegal, but won't be called as we stop during check */
		2839	ev_timer_init (&tw, 0, timeout * 1e-3, 0.);
		2840	ev_timer_start (loop, &tw);
		2841
		2842	// create one ev_io per pollfd
		2843	for (int i = 0; i < nfd; ++i)
		2844	{
		2845	ev_io_init (iow + i, io_cb, fds [i].fd,
		2846	((fds [i].events & POLLIN ? EV_READ : 0)
		2847	| (fds [i].events & POLLOUT ? EV_WRITE : 0)));
		2848
		2849	fds [i].revents = 0;
		2850	ev_io_start (loop, iow + i);
		2851	}
		2852	}
		2853
		2854	// stop all watchers after blocking
		2855	static void
		2856	adns_check_cb (struct ev_loop *loop, ev_check *w, int revents)
		2857	{
		2858	ev_timer_stop (loop, &tw);
		2859
		2860	for (int i = 0; i < nfd; ++i)
		2861	{
		2862	// set the relevant poll flags
		2863	// could also call adns_processreadable etc. here
		2864	struct pollfd *fd = fds + i;
		2865	int revents = ev_clear_pending (iow + i);
		2866	if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;
		2867	if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;
		2868
		2869	// now stop the watcher
		2870	ev_io_stop (loop, iow + i);
		2871	}
		2872
		2873	adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));
		2874	}
		2875
		2876	Method 2: This would be just like method 1, but you run C<adns_afterpoll>
		2877	in the prepare watcher and would dispose of the check watcher.
		2878
		2879	Method 3: If the module to be embedded supports explicit event
		2880	notification (libadns does), you can also make use of the actual watcher
		2881	callbacks, and only destroy/create the watchers in the prepare watcher.
		2882
		2883	static void
		2884	timer_cb (EV_P_ ev_timer *w, int revents)
		2885	{
		2886	adns_state ads = (adns_state)w->data;
		2887	update_now (EV_A);
		2888
		2889	adns_processtimeouts (ads, &tv_now);
		2890	}
		2891
		2892	static void
		2893	io_cb (EV_P_ ev_io *w, int revents)
		2894	{
		2895	adns_state ads = (adns_state)w->data;
		2896	update_now (EV_A);
		2897
		2898	if (revents & EV_READ ) adns_processreadable (ads, w->fd, &tv_now);
		2899	if (revents & EV_WRITE) adns_processwriteable (ads, w->fd, &tv_now);
		2900	}
		2901
		2902	// do not ever call adns_afterpoll
		2903
		2904	Method 4: Do not use a prepare or check watcher because the module you
		2905	want to embed is not flexible enough to support it. Instead, you can
		2906	override their poll function. The drawback with this solution is that the
		2907	main loop is now no longer controllable by EV. The C<Glib::EV> module uses
		2908	this approach, effectively embedding EV as a client into the horrible
		2909	libglib event loop.
		2910
		2911	static gint
		2912	event_poll_func (GPollFD *fds, guint nfds, gint timeout)
		2913	{
		2914	int got_events = 0;
		2915
		2916	for (n = 0; n < nfds; ++n)
		2917	// create/start io watcher that sets the relevant bits in fds[n] and increment got_events
		2918
		2919	if (timeout >= 0)
		2920	// create/start timer
		2921
		2922	// poll
		2923	ev_run (EV_A_ 0);
		2924
		2925	// stop timer again
		2926	if (timeout >= 0)
		2927	ev_timer_stop (EV_A_ &to);
		2928
		2929	// stop io watchers again - their callbacks should have set
		2930	for (n = 0; n < nfds; ++n)
		2931	ev_io_stop (EV_A_ iow [n]);
		2932
		2933	return got_events;
		2934	}
		2935
		2936
1060	=head2 C<ev_embed> - when one backend isn't enough	2937	=head2 C<ev_embed> - when one backend isn't enough...
1061		2938
1062	This is a rather advanced watcher type that lets you embed one event loop	2939	This is a rather advanced watcher type that lets you embed one event loop
1063	into another.	2940	into another (currently only C<ev_io> events are supported in the embedded
		2941	loop, other types of watchers might be handled in a delayed or incorrect
		2942	fashion and must not be used).
1064		2943
1065	There are primarily two reasons you would want that: work around bugs and	2944	There are primarily two reasons you would want that: work around bugs and
1066	prioritise I/O.	2945	prioritise I/O.
1067		2946
1068	As an example for a bug workaround, the kqueue backend might only support	2947	As an example for a bug workaround, the kqueue backend might only support
1069	sockets on some platform, so it is unusable as generic backend, but you	2948	sockets on some platform, so it is unusable as generic backend, but you
1070	still want to make use of it because you have many sockets and it scales	2949	still want to make use of it because you have many sockets and it scales
1071	so nicely. In this case, you would create a kqueue-based loop and embed it	2950	so nicely. In this case, you would create a kqueue-based loop and embed
1072	into your default loop (which might use e.g. poll). Overall operation will	2951	it into your default loop (which might use e.g. poll). Overall operation
1073	be a bit slower because first libev has to poll and then call kevent, but	2952	will be a bit slower because first libev has to call C<poll> and then
1074	at least you can use both at what they are best.	2953	C<kevent>, but at least you can use both mechanisms for what they are
		2954	best: C<kqueue> for scalable sockets and C<poll> if you want it to work :)
1075		2955
1076	As for prioritising I/O: rarely you have the case where some fds have	2956	As for prioritising I/O: under rare circumstances you have the case where
1077	to be watched and handled very quickly (with low latency), and even	2957	some fds have to be watched and handled very quickly (with low latency),
1078	priorities and idle watchers might have too much overhead. In this case	2958	and even priorities and idle watchers might have too much overhead. In
1079	you would put all the high priority stuff in one loop and all the rest in	2959	this case you would put all the high priority stuff in one loop and all
1080	a second one, and embed the second one in the first.	2960	the rest in a second one, and embed the second one in the first.
1081		2961
1082	As long as the watcher is started it will automatically handle events. The	2962	As long as the watcher is active, the callback will be invoked every
1083	callback will be invoked whenever some events have been handled. You can	2963	time there might be events pending in the embedded loop. The callback
1084	set the callback to C<0> to avoid having to specify one if you are not	2964	must then call C<ev_embed_sweep (mainloop, watcher)> to make a single
1085	interested in that.	2965	sweep and invoke their callbacks (the callback doesn't need to invoke the
		2966	C<ev_embed_sweep> function directly, it could also start an idle watcher
		2967	to give the embedded loop strictly lower priority for example).
1086		2968
1087	Also, there have not currently been made special provisions for forking:	2969	You can also set the callback to C<0>, in which case the embed watcher
1088	when you fork, you not only have to call C<ev_loop_fork> on both loops,	2970	will automatically execute the embedded loop sweep whenever necessary.
1089	but you will also have to stop and restart any C<ev_embed> watchers
1090	yourself.
1091		2971
		2972	Fork detection will be handled transparently while the C<ev_embed> watcher
		2973	is active, i.e., the embedded loop will automatically be forked when the
		2974	embedding loop forks. In other cases, the user is responsible for calling
		2975	C<ev_loop_fork> on the embedded loop.
		2976
1092	Unfortunately, not all backends are embeddable, only the ones returned by	2977	Unfortunately, not all backends are embeddable: only the ones returned by
1093	C<ev_embeddable_backends> are, which, unfortunately, does not include any	2978	C<ev_embeddable_backends> are, which, unfortunately, does not include any
1094	portable one.	2979	portable one.
1095		2980
1096	So when you want to use this feature you will always have to be prepared	2981	So when you want to use this feature you will always have to be prepared
1097	that you cannot get an embeddable loop. The recommended way to get around	2982	that you cannot get an embeddable loop. The recommended way to get around
1098	this is to have a separate variables for your embeddable loop, try to	2983	this is to have a separate variables for your embeddable loop, try to
1099	create it, and if that fails, use the normal loop for everything:	2984	create it, and if that fails, use the normal loop for everything.
1100		2985
1101	struct ev_loop *loop_hi = ev_default_init (0);	2986	=head3 C<ev_embed> and fork
1102	struct ev_loop *loop_lo = 0;
1103	struct ev_embed embed;
1104
1105	// see if there is a chance of getting one that works
1106	// (remember that a flags value of 0 means autodetection)
1107	loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
1108	? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
1109	: 0;
1110		2987
1111	// if we got one, then embed it, otherwise default to loop_hi	2988	While the C<ev_embed> watcher is running, forks in the embedding loop will
1112	if (loop_lo)	2989	automatically be applied to the embedded loop as well, so no special
1113	{	2990	fork handling is required in that case. When the watcher is not running,
1114	ev_embed_init (&embed, 0, loop_lo);	2991	however, it is still the task of the libev user to call C<ev_loop_fork ()>
1115	ev_embed_start (loop_hi, &embed);	2992	as applicable.
1116	}	2993
1117	else	2994	=head3 Watcher-Specific Functions and Data Members
1118	loop_lo = loop_hi;
1119		2995
1120	=over 4	2996	=over 4
1121		2997
1122	=item ev_embed_init (ev_embed *, callback, struct ev_loop *loop)	2998	=item ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)
1123		2999
1124	=item ev_embed_set (ev_embed *, callback, struct ev_loop *loop)	3000	=item ev_embed_set (ev_embed *, callback, struct ev_loop *embedded_loop)
1125		3001
1126	Configures the watcher to embed the given loop, which must be embeddable.	3002	Configures the watcher to embed the given loop, which must be
		3003	embeddable. If the callback is C<0>, then C<ev_embed_sweep> will be
		3004	invoked automatically, otherwise it is the responsibility of the callback
		3005	to invoke it (it will continue to be called until the sweep has been done,
		3006	if you do not want that, you need to temporarily stop the embed watcher).
		3007
		3008	=item ev_embed_sweep (loop, ev_embed *)
		3009
		3010	Make a single, non-blocking sweep over the embedded loop. This works
		3011	similarly to C<ev_run (embedded_loop, EVRUN_NOWAIT)>, but in the most
		3012	appropriate way for embedded loops.
		3013
		3014	=item struct ev_loop *other [read-only]
		3015
		3016	The embedded event loop.
1127		3017
1128	=back	3018	=back
1129		3019
		3020	=head3 Examples
		3021
		3022	Example: Try to get an embeddable event loop and embed it into the default
		3023	event loop. If that is not possible, use the default loop. The default
		3024	loop is stored in C<loop_hi>, while the embeddable loop is stored in
		3025	C<loop_lo> (which is C<loop_hi> in the case no embeddable loop can be
		3026	used).
		3027
		3028	struct ev_loop *loop_hi = ev_default_init (0);
		3029	struct ev_loop *loop_lo = 0;
		3030	ev_embed embed;
		3031
		3032	// see if there is a chance of getting one that works
		3033	// (remember that a flags value of 0 means autodetection)
		3034	loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
		3035	? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
		3036	: 0;
		3037
		3038	// if we got one, then embed it, otherwise default to loop_hi
		3039	if (loop_lo)
		3040	{
		3041	ev_embed_init (&embed, 0, loop_lo);
		3042	ev_embed_start (loop_hi, &embed);
		3043	}
		3044	else
		3045	loop_lo = loop_hi;
		3046
		3047	Example: Check if kqueue is available but not recommended and create
		3048	a kqueue backend for use with sockets (which usually work with any
		3049	kqueue implementation). Store the kqueue/socket-only event loop in
		3050	C<loop_socket>. (One might optionally use C<EVFLAG_NOENV>, too).
		3051
		3052	struct ev_loop *loop = ev_default_init (0);
		3053	struct ev_loop *loop_socket = 0;
		3054	ev_embed embed;
		3055
		3056	if (ev_supported_backends () & ~ev_recommended_backends () & EVBACKEND_KQUEUE)
		3057	if ((loop_socket = ev_loop_new (EVBACKEND_KQUEUE))
		3058	{
		3059	ev_embed_init (&embed, 0, loop_socket);
		3060	ev_embed_start (loop, &embed);
		3061	}
		3062
		3063	if (!loop_socket)
		3064	loop_socket = loop;
		3065
		3066	// now use loop_socket for all sockets, and loop for everything else
		3067
		3068
		3069	=head2 C<ev_fork> - the audacity to resume the event loop after a fork
		3070
		3071	Fork watchers are called when a C<fork ()> was detected (usually because
		3072	whoever is a good citizen cared to tell libev about it by calling
		3073	C<ev_default_fork> or C<ev_loop_fork>). The invocation is done before the
		3074	event loop blocks next and before C<ev_check> watchers are being called,
		3075	and only in the child after the fork. If whoever good citizen calling
		3076	C<ev_default_fork> cheats and calls it in the wrong process, the fork
		3077	handlers will be invoked, too, of course.
		3078
		3079	=head3 The special problem of life after fork - how is it possible?
		3080
		3081	Most uses of C<fork()> consist of forking, then some simple calls to set
		3082	up/change the process environment, followed by a call to C<exec()>. This
		3083	sequence should be handled by libev without any problems.
		3084
		3085	This changes when the application actually wants to do event handling
		3086	in the child, or both parent in child, in effect "continuing" after the
		3087	fork.
		3088
		3089	The default mode of operation (for libev, with application help to detect
		3090	forks) is to duplicate all the state in the child, as would be expected
		3091	when I<either> the parent I<or> the child process continues.
		3092
		3093	When both processes want to continue using libev, then this is usually the
		3094	wrong result. In that case, usually one process (typically the parent) is
		3095	supposed to continue with all watchers in place as before, while the other
		3096	process typically wants to start fresh, i.e. without any active watchers.
		3097
		3098	The cleanest and most efficient way to achieve that with libev is to
		3099	simply create a new event loop, which of course will be "empty", and
		3100	use that for new watchers. This has the advantage of not touching more
		3101	memory than necessary, and thus avoiding the copy-on-write, and the
		3102	disadvantage of having to use multiple event loops (which do not support
		3103	signal watchers).
		3104
		3105	When this is not possible, or you want to use the default loop for
		3106	other reasons, then in the process that wants to start "fresh", call
		3107	C<ev_loop_destroy (EV_DEFAULT)> followed by C<ev_default_loop (...)>.
		3108	Destroying the default loop will "orphan" (not stop) all registered
		3109	watchers, so you have to be careful not to execute code that modifies
		3110	those watchers. Note also that in that case, you have to re-register any
		3111	signal watchers.
		3112
		3113	=head3 Watcher-Specific Functions and Data Members
		3114
		3115	=over 4
		3116
		3117	=item ev_fork_init (ev_fork *, callback)
		3118
		3119	Initialises and configures the fork watcher - it has no parameters of any
		3120	kind. There is a C<ev_fork_set> macro, but using it is utterly pointless,
		3121	really.
		3122
		3123	=back
		3124
		3125
		3126	=head2 C<ev_cleanup> - even the best things end
		3127
		3128	Cleanup watchers are called just before the event loop is being destroyed
		3129	by a call to C<ev_loop_destroy>.
		3130
		3131	While there is no guarantee that the event loop gets destroyed, cleanup
		3132	watchers provide a convenient method to install cleanup hooks for your
		3133	program, worker threads and so on - you just to make sure to destroy the
		3134	loop when you want them to be invoked.
		3135
		3136	Cleanup watchers are invoked in the same way as any other watcher. Unlike
		3137	all other watchers, they do not keep a reference to the event loop (which
		3138	makes a lot of sense if you think about it). Like all other watchers, you
		3139	can call libev functions in the callback, except C<ev_cleanup_start>.
		3140
		3141	=head3 Watcher-Specific Functions and Data Members
		3142
		3143	=over 4
		3144
		3145	=item ev_cleanup_init (ev_cleanup *, callback)
		3146
		3147	Initialises and configures the cleanup watcher - it has no parameters of
		3148	any kind. There is a C<ev_cleanup_set> macro, but using it is utterly
		3149	pointless, I assure you.
		3150
		3151	=back
		3152
		3153	Example: Register an atexit handler to destroy the default loop, so any
		3154	cleanup functions are called.
		3155
		3156	static void
		3157	program_exits (void)
		3158	{
		3159	ev_loop_destroy (EV_DEFAULT_UC);
		3160	}
		3161
		3162	...
		3163	atexit (program_exits);
		3164
		3165
		3166	=head2 C<ev_async> - how to wake up an event loop
		3167
		3168	In general, you cannot use an C<ev_run> from multiple threads or other
		3169	asynchronous sources such as signal handlers (as opposed to multiple event
		3170	loops - those are of course safe to use in different threads).
		3171
		3172	Sometimes, however, you need to wake up an event loop you do not control,
		3173	for example because it belongs to another thread. This is what C<ev_async>
		3174	watchers do: as long as the C<ev_async> watcher is active, you can signal
		3175	it by calling C<ev_async_send>, which is thread- and signal safe.
		3176
		3177	This functionality is very similar to C<ev_signal> watchers, as signals,
		3178	too, are asynchronous in nature, and signals, too, will be compressed
		3179	(i.e. the number of callback invocations may be less than the number of
		3180	C<ev_async_sent> calls).
		3181
		3182	Unlike C<ev_signal> watchers, C<ev_async> works with any event loop, not
		3183	just the default loop.
		3184
		3185	=head3 Queueing
		3186
		3187	C<ev_async> does not support queueing of data in any way. The reason
		3188	is that the author does not know of a simple (or any) algorithm for a
		3189	multiple-writer-single-reader queue that works in all cases and doesn't
		3190	need elaborate support such as pthreads or unportable memory access
		3191	semantics.
		3192
		3193	That means that if you want to queue data, you have to provide your own
		3194	queue. But at least I can tell you how to implement locking around your
		3195	queue:
		3196
		3197	=over 4
		3198
		3199	=item queueing from a signal handler context
		3200
		3201	To implement race-free queueing, you simply add to the queue in the signal
		3202	handler but you block the signal handler in the watcher callback. Here is
		3203	an example that does that for some fictitious SIGUSR1 handler:
		3204
		3205	static ev_async mysig;
		3206
		3207	static void
		3208	sigusr1_handler (void)
		3209	{
		3210	sometype data;
		3211
		3212	// no locking etc.
		3213	queue_put (data);
		3214	ev_async_send (EV_DEFAULT_ &mysig);
		3215	}
		3216
		3217	static void
		3218	mysig_cb (EV_P_ ev_async *w, int revents)
		3219	{
		3220	sometype data;
		3221	sigset_t block, prev;
		3222
		3223	sigemptyset (&block);
		3224	sigaddset (&block, SIGUSR1);
		3225	sigprocmask (SIG_BLOCK, &block, &prev);
		3226
		3227	while (queue_get (&data))
		3228	process (data);
		3229
		3230	if (sigismember (&prev, SIGUSR1)
		3231	sigprocmask (SIG_UNBLOCK, &block, 0);
		3232	}
		3233
		3234	(Note: pthreads in theory requires you to use C<pthread_setmask>
		3235	instead of C<sigprocmask> when you use threads, but libev doesn't do it
		3236	either...).
		3237
		3238	=item queueing from a thread context
		3239
		3240	The strategy for threads is different, as you cannot (easily) block
		3241	threads but you can easily preempt them, so to queue safely you need to
		3242	employ a traditional mutex lock, such as in this pthread example:
		3243
		3244	static ev_async mysig;
		3245	static pthread_mutex_t mymutex = PTHREAD_MUTEX_INITIALIZER;
		3246
		3247	static void
		3248	otherthread (void)
		3249	{
		3250	// only need to lock the actual queueing operation
		3251	pthread_mutex_lock (&mymutex);
		3252	queue_put (data);
		3253	pthread_mutex_unlock (&mymutex);
		3254
		3255	ev_async_send (EV_DEFAULT_ &mysig);
		3256	}
		3257
		3258	static void
		3259	mysig_cb (EV_P_ ev_async *w, int revents)
		3260	{
		3261	pthread_mutex_lock (&mymutex);
		3262
		3263	while (queue_get (&data))
		3264	process (data);
		3265
		3266	pthread_mutex_unlock (&mymutex);
		3267	}
		3268
		3269	=back
		3270
		3271
		3272	=head3 Watcher-Specific Functions and Data Members
		3273
		3274	=over 4
		3275
		3276	=item ev_async_init (ev_async *, callback)
		3277
		3278	Initialises and configures the async watcher - it has no parameters of any
		3279	kind. There is a C<ev_async_set> macro, but using it is utterly pointless,
		3280	trust me.
		3281
		3282	=item ev_async_send (loop, ev_async *)
		3283
		3284	Sends/signals/activates the given C<ev_async> watcher, that is, feeds
		3285	an C<EV_ASYNC> event on the watcher into the event loop. Unlike
		3286	C<ev_feed_event>, this call is safe to do from other threads, signal or
		3287	similar contexts (see the discussion of C<EV_ATOMIC_T> in the embedding
		3288	section below on what exactly this means).
		3289
		3290	Note that, as with other watchers in libev, multiple events might get
		3291	compressed into a single callback invocation (another way to look at this
		3292	is that C<ev_async> watchers are level-triggered, set on C<ev_async_send>,
		3293	reset when the event loop detects that).
		3294
		3295	This call incurs the overhead of a system call only once per event loop
		3296	iteration, so while the overhead might be noticeable, it doesn't apply to
		3297	repeated calls to C<ev_async_send> for the same event loop.
		3298
		3299	=item bool = ev_async_pending (ev_async *)
		3300
		3301	Returns a non-zero value when C<ev_async_send> has been called on the
		3302	watcher but the event has not yet been processed (or even noted) by the
		3303	event loop.
		3304
		3305	C<ev_async_send> sets a flag in the watcher and wakes up the loop. When
		3306	the loop iterates next and checks for the watcher to have become active,
		3307	it will reset the flag again. C<ev_async_pending> can be used to very
		3308	quickly check whether invoking the loop might be a good idea.
		3309
		3310	Not that this does I<not> check whether the watcher itself is pending,
		3311	only whether it has been requested to make this watcher pending: there
		3312	is a time window between the event loop checking and resetting the async
		3313	notification, and the callback being invoked.
		3314
		3315	=back
		3316
1130		3317
1131	=head1 OTHER FUNCTIONS	3318	=head1 OTHER FUNCTIONS
1132		3319
1133	There are some other functions of possible interest. Described. Here. Now.	3320	There are some other functions of possible interest. Described. Here. Now.
1134		3321
1135	=over 4	3322	=over 4
1136		3323
1137	=item ev_once (loop, int fd, int events, ev_tstamp timeout, callback)	3324	=item ev_once (loop, int fd, int events, ev_tstamp timeout, callback)
1138		3325
1139	This function combines a simple timer and an I/O watcher, calls your	3326	This function combines a simple timer and an I/O watcher, calls your
1140	callback on whichever event happens first and automatically stop both	3327	callback on whichever event happens first and automatically stops both
1141	watchers. This is useful if you want to wait for a single event on an fd	3328	watchers. This is useful if you want to wait for a single event on an fd
1142	or timeout without having to allocate/configure/start/stop/free one or	3329	or timeout without having to allocate/configure/start/stop/free one or
1143	more watchers yourself.	3330	more watchers yourself.
1144		3331
1145	If C<fd> is less than 0, then no I/O watcher will be started and events	3332	If C<fd> is less than 0, then no I/O watcher will be started and the
1146	is being ignored. Otherwise, an C<ev_io> watcher for the given C<fd> and	3333	C<events> argument is being ignored. Otherwise, an C<ev_io> watcher for
1147	C<events> set will be craeted and started.	3334	the given C<fd> and C<events> set will be created and started.
1148		3335
1149	If C<timeout> is less than 0, then no timeout watcher will be	3336	If C<timeout> is less than 0, then no timeout watcher will be
1150	started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and	3337	started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and
1151	repeat = 0) will be started. While C<0> is a valid timeout, it is of	3338	repeat = 0) will be started. C<0> is a valid timeout.
1152	dubious value.
1153		3339
1154	The callback has the type C<void (*cb)(int revents, void *arg)> and gets	3340	The callback has the type C<void (*cb)(int revents, void *arg)> and is
1155	passed an C<revents> set like normal event callbacks (a combination of	3341	passed an C<revents> set like normal event callbacks (a combination of
1156	C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMEOUT>) and the C<arg>	3342	C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMER>) and the C<arg>
1157	value passed to C<ev_once>:	3343	value passed to C<ev_once>. Note that it is possible to receive I<both>
		3344	a timeout and an io event at the same time - you probably should give io
		3345	events precedence.
1158		3346
		3347	Example: wait up to ten seconds for data to appear on STDIN_FILENO.
		3348
1159	static void stdin_ready (int revents, void *arg)	3349	static void stdin_ready (int revents, void *arg)
1160	{	3350	{
1161	if (revents & EV_TIMEOUT)
1162	/* doh, nothing entered */;
1163	else if (revents & EV_READ)	3351	if (revents & EV_READ)
1164	/* stdin might have data for us, joy! */;	3352	/* stdin might have data for us, joy! */;
		3353	else if (revents & EV_TIMER)
		3354	/* doh, nothing entered */;
1165	}	3355	}
1166		3356
1167	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);	3357	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
1168
1169	=item ev_feed_event (loop, watcher, int events)
1170
1171	Feeds the given event set into the event loop, as if the specified event
1172	had happened for the specified watcher (which must be a pointer to an
1173	initialised but not necessarily started event watcher).
1174		3358
1175	=item ev_feed_fd_event (loop, int fd, int revents)	3359	=item ev_feed_fd_event (loop, int fd, int revents)
1176		3360
1177	Feed an event on the given fd, as if a file descriptor backend detected	3361	Feed an event on the given fd, as if a file descriptor backend detected
1178	the given events it.	3362	the given events it.
1179		3363
1180	=item ev_feed_signal_event (loop, int signum)	3364	=item ev_feed_signal_event (loop, int signum)
1181		3365
1182	Feed an event as if the given signal occured (loop must be the default loop!).	3366	Feed an event as if the given signal occurred (C<loop> must be the default
		3367	loop!).
		3368
		3369	=back
		3370
		3371
		3372	=head1 COMMON OR USEFUL IDIOMS (OR BOTH)
		3373
		3374	This section explains some common idioms that are not immediately
		3375	obvious. Note that examples are sprinkled over the whole manual, and this
		3376	section only contains stuff that wouldn't fit anywhere else.
		3377
		3378	=over 4
		3379
		3380	=item Model/nested event loop invocations and exit conditions.
		3381
		3382	Often (especially in GUI toolkits) there are places where you have
		3383	I<modal> interaction, which is most easily implemented by recursively
		3384	invoking C<ev_run>.
		3385
		3386	This brings the problem of exiting - a callback might want to finish the
		3387	main C<ev_run> call, but not the nested one (e.g. user clicked "Quit", but
		3388	a modal "Are you sure?" dialog is still waiting), or just the nested one
		3389	and not the main one (e.g. user clocked "Ok" in a modal dialog), or some
		3390	other combination: In these cases, C<ev_break> will not work alone.
		3391
		3392	The solution is to maintain "break this loop" variable for each C<ev_run>
		3393	invocation, and use a loop around C<ev_run> until the condition is
		3394	triggered, using C<EVRUN_ONCE>:
		3395
		3396	// main loop
		3397	int exit_main_loop = 0;
		3398
		3399	while (!exit_main_loop)
		3400	ev_run (EV_DEFAULT_ EVRUN_ONCE);
		3401
		3402	// in a model watcher
		3403	int exit_nested_loop = 0;
		3404
		3405	while (!exit_nested_loop)
		3406	ev_run (EV_A_ EVRUN_ONCE);
		3407
		3408	To exit from any of these loops, just set the corresponding exit variable:
		3409
		3410	// exit modal loop
		3411	exit_nested_loop = 1;
		3412
		3413	// exit main program, after modal loop is finished
		3414	exit_main_loop = 1;
		3415
		3416	// exit both
		3417	exit_main_loop = exit_nested_loop = 1;
1183		3418
1184	=back	3419	=back
1185		3420
1186		3421
1187	=head1 LIBEVENT EMULATION	3422	=head1 LIBEVENT EMULATION
1188		3423
1189	Libev offers a compatibility emulation layer for libevent. It cannot	3424	Libev offers a compatibility emulation layer for libevent. It cannot
1190	emulate the internals of libevent, so here are some usage hints:	3425	emulate the internals of libevent, so here are some usage hints:
1191		3426
1192	=over 4	3427	=over 4
		3428
		3429	=item * Only the libevent-1.4.1-beta API is being emulated.
		3430
		3431	This was the newest libevent version available when libev was implemented,
		3432	and is still mostly uncanged in 2010.
1193		3433
1194	=item * Use it by including <event.h>, as usual.	3434	=item * Use it by including <event.h>, as usual.
1195		3435
1196	=item * The following members are fully supported: ev_base, ev_callback,	3436	=item * The following members are fully supported: ev_base, ev_callback,
1197	ev_arg, ev_fd, ev_res, ev_events.	3437	ev_arg, ev_fd, ev_res, ev_events.
…		…
1202		3442
1203	=item * Priorities are not currently supported. Initialising priorities	3443	=item * Priorities are not currently supported. Initialising priorities
1204	will fail and all watchers will have the same priority, even though there	3444	will fail and all watchers will have the same priority, even though there
1205	is an ev_pri field.	3445	is an ev_pri field.
1206		3446
		3447	=item * In libevent, the last base created gets the signals, in libev, the
		3448	base that registered the signal gets the signals.
		3449
1207	=item * Other members are not supported.	3450	=item * Other members are not supported.
1208		3451
1209	=item * The libev emulation is I<not> ABI compatible to libevent, you need	3452	=item * The libev emulation is I<not> ABI compatible to libevent, you need
1210	to use the libev header file and library.	3453	to use the libev header file and library.
1211		3454
1212	=back	3455	=back
1213		3456
1214	=head1 C++ SUPPORT	3457	=head1 C++ SUPPORT
1215		3458
1216	TBD.	3459	Libev comes with some simplistic wrapper classes for C++ that mainly allow
		3460	you to use some convenience methods to start/stop watchers and also change
		3461	the callback model to a model using method callbacks on objects.
		3462
		3463	To use it,
		3464
		3465	#include <ev++.h>
		3466
		3467	This automatically includes F<ev.h> and puts all of its definitions (many
		3468	of them macros) into the global namespace. All C++ specific things are
		3469	put into the C<ev> namespace. It should support all the same embedding
		3470	options as F<ev.h>, most notably C<EV_MULTIPLICITY>.
		3471
		3472	Care has been taken to keep the overhead low. The only data member the C++
		3473	classes add (compared to plain C-style watchers) is the event loop pointer
		3474	that the watcher is associated with (or no additional members at all if
		3475	you disable C<EV_MULTIPLICITY> when embedding libev).
		3476
		3477	Currently, functions, and static and non-static member functions can be
		3478	used as callbacks. Other types should be easy to add as long as they only
		3479	need one additional pointer for context. If you need support for other
		3480	types of functors please contact the author (preferably after implementing
		3481	it).
		3482
		3483	Here is a list of things available in the C<ev> namespace:
		3484
		3485	=over 4
		3486
		3487	=item C<ev::READ>, C<ev::WRITE> etc.
		3488
		3489	These are just enum values with the same values as the C<EV_READ> etc.
		3490	macros from F<ev.h>.
		3491
		3492	=item C<ev::tstamp>, C<ev::now>
		3493
		3494	Aliases to the same types/functions as with the C<ev_> prefix.
		3495
		3496	=item C<ev::io>, C<ev::timer>, C<ev::periodic>, C<ev::idle>, C<ev::sig> etc.
		3497
		3498	For each C<ev_TYPE> watcher in F<ev.h> there is a corresponding class of
		3499	the same name in the C<ev> namespace, with the exception of C<ev_signal>
		3500	which is called C<ev::sig> to avoid clashes with the C<signal> macro
		3501	defines by many implementations.
		3502
		3503	All of those classes have these methods:
		3504
		3505	=over 4
		3506
		3507	=item ev::TYPE::TYPE ()
		3508
		3509	=item ev::TYPE::TYPE (loop)
		3510
		3511	=item ev::TYPE::~TYPE
		3512
		3513	The constructor (optionally) takes an event loop to associate the watcher
		3514	with. If it is omitted, it will use C<EV_DEFAULT>.
		3515
		3516	The constructor calls C<ev_init> for you, which means you have to call the
		3517	C<set> method before starting it.
		3518
		3519	It will not set a callback, however: You have to call the templated C<set>
		3520	method to set a callback before you can start the watcher.
		3521
		3522	(The reason why you have to use a method is a limitation in C++ which does
		3523	not allow explicit template arguments for constructors).
		3524
		3525	The destructor automatically stops the watcher if it is active.
		3526
		3527	=item w->set<class, &class::method> (object *)
		3528
		3529	This method sets the callback method to call. The method has to have a
		3530	signature of C<void (*)(ev_TYPE &, int)>, it receives the watcher as
		3531	first argument and the C<revents> as second. The object must be given as
		3532	parameter and is stored in the C<data> member of the watcher.
		3533
		3534	This method synthesizes efficient thunking code to call your method from
		3535	the C callback that libev requires. If your compiler can inline your
		3536	callback (i.e. it is visible to it at the place of the C<set> call and
		3537	your compiler is good :), then the method will be fully inlined into the
		3538	thunking function, making it as fast as a direct C callback.
		3539
		3540	Example: simple class declaration and watcher initialisation
		3541
		3542	struct myclass
		3543	{
		3544	void io_cb (ev::io &w, int revents) { }
		3545	}
		3546
		3547	myclass obj;
		3548	ev::io iow;
		3549	iow.set <myclass, &myclass::io_cb> (&obj);
		3550
		3551	=item w->set (object *)
		3552
		3553	This is a variation of a method callback - leaving out the method to call
		3554	will default the method to C<operator ()>, which makes it possible to use
		3555	functor objects without having to manually specify the C<operator ()> all
		3556	the time. Incidentally, you can then also leave out the template argument
		3557	list.
		3558
		3559	The C<operator ()> method prototype must be C<void operator ()(watcher &w,
		3560	int revents)>.
		3561
		3562	See the method-C<set> above for more details.
		3563
		3564	Example: use a functor object as callback.
		3565
		3566	struct myfunctor
		3567	{
		3568	void operator() (ev::io &w, int revents)
		3569	{
		3570	...
		3571	}
		3572	}
		3573
		3574	myfunctor f;
		3575
		3576	ev::io w;
		3577	w.set (&f);
		3578
		3579	=item w->set<function> (void *data = 0)
		3580
		3581	Also sets a callback, but uses a static method or plain function as
		3582	callback. The optional C<data> argument will be stored in the watcher's
		3583	C<data> member and is free for you to use.
		3584
		3585	The prototype of the C<function> must be C<void (*)(ev::TYPE &w, int)>.
		3586
		3587	See the method-C<set> above for more details.
		3588
		3589	Example: Use a plain function as callback.
		3590
		3591	static void io_cb (ev::io &w, int revents) { }
		3592	iow.set <io_cb> ();
		3593
		3594	=item w->set (loop)
		3595
		3596	Associates a different C<struct ev_loop> with this watcher. You can only
		3597	do this when the watcher is inactive (and not pending either).
		3598
		3599	=item w->set ([arguments])
		3600
		3601	Basically the same as C<ev_TYPE_set>, with the same arguments. Either this
		3602	method or a suitable start method must be called at least once. Unlike the
		3603	C counterpart, an active watcher gets automatically stopped and restarted
		3604	when reconfiguring it with this method.
		3605
		3606	=item w->start ()
		3607
		3608	Starts the watcher. Note that there is no C<loop> argument, as the
		3609	constructor already stores the event loop.
		3610
		3611	=item w->start ([arguments])
		3612
		3613	Instead of calling C<set> and C<start> methods separately, it is often
		3614	convenient to wrap them in one call. Uses the same type of arguments as
		3615	the configure C<set> method of the watcher.
		3616
		3617	=item w->stop ()
		3618
		3619	Stops the watcher if it is active. Again, no C<loop> argument.
		3620
		3621	=item w->again () (C<ev::timer>, C<ev::periodic> only)
		3622
		3623	For C<ev::timer> and C<ev::periodic>, this invokes the corresponding
		3624	C<ev_TYPE_again> function.
		3625
		3626	=item w->sweep () (C<ev::embed> only)
		3627
		3628	Invokes C<ev_embed_sweep>.
		3629
		3630	=item w->update () (C<ev::stat> only)
		3631
		3632	Invokes C<ev_stat_stat>.
		3633
		3634	=back
		3635
		3636	=back
		3637
		3638	Example: Define a class with two I/O and idle watchers, start the I/O
		3639	watchers in the constructor.
		3640
		3641	class myclass
		3642	{
		3643	ev::io io ; void io_cb (ev::io &w, int revents);
		3644	ev::io2 io2 ; void io2_cb (ev::io &w, int revents);
		3645	ev::idle idle; void idle_cb (ev::idle &w, int revents);
		3646
		3647	myclass (int fd)
		3648	{
		3649	io .set <myclass, &myclass::io_cb > (this);
		3650	io2 .set <myclass, &myclass::io2_cb > (this);
		3651	idle.set <myclass, &myclass::idle_cb> (this);
		3652
		3653	io.set (fd, ev::WRITE); // configure the watcher
		3654	io.start (); // start it whenever convenient
		3655
		3656	io2.start (fd, ev::READ); // set + start in one call
		3657	}
		3658	};
		3659
		3660
		3661	=head1 OTHER LANGUAGE BINDINGS
		3662
		3663	Libev does not offer other language bindings itself, but bindings for a
		3664	number of languages exist in the form of third-party packages. If you know
		3665	any interesting language binding in addition to the ones listed here, drop
		3666	me a note.
		3667
		3668	=over 4
		3669
		3670	=item Perl
		3671
		3672	The EV module implements the full libev API and is actually used to test
		3673	libev. EV is developed together with libev. Apart from the EV core module,
		3674	there are additional modules that implement libev-compatible interfaces
		3675	to C<libadns> (C<EV::ADNS>, but C<AnyEvent::DNS> is preferred nowadays),
		3676	C<Net::SNMP> (C<Net::SNMP::EV>) and the C<libglib> event core (C<Glib::EV>
		3677	and C<EV::Glib>).
		3678
		3679	It can be found and installed via CPAN, its homepage is at
		3680	L<http://software.schmorp.de/pkg/EV>.
		3681
		3682	=item Python
		3683
		3684	Python bindings can be found at L<http://code.google.com/p/pyev/>. It
		3685	seems to be quite complete and well-documented.
		3686
		3687	=item Ruby
		3688
		3689	Tony Arcieri has written a ruby extension that offers access to a subset
		3690	of the libev API and adds file handle abstractions, asynchronous DNS and
		3691	more on top of it. It can be found via gem servers. Its homepage is at
		3692	L<http://rev.rubyforge.org/>.
		3693
		3694	Roger Pack reports that using the link order C<-lws2_32 -lmsvcrt-ruby-190>
		3695	makes rev work even on mingw.
		3696
		3697	=item Haskell
		3698
		3699	A haskell binding to libev is available at
		3700	L<http://hackage.haskell.org/cgi-bin/hackage-scripts/package/hlibev>.
		3701
		3702	=item D
		3703
		3704	Leandro Lucarella has written a D language binding (F<ev.d>) for libev, to
		3705	be found at L<http://proj.llucax.com.ar/wiki/evd>.
		3706
		3707	=item Ocaml
		3708
		3709	Erkki Seppala has written Ocaml bindings for libev, to be found at
		3710	L<http://modeemi.cs.tut.fi/~flux/software/ocaml-ev/>.
		3711
		3712	=item Lua
		3713
		3714	Brian Maher has written a partial interface to libev for lua (at the
		3715	time of this writing, only C<ev_io> and C<ev_timer>), to be found at
		3716	L<http://github.com/brimworks/lua-ev>.
		3717
		3718	=back
		3719
		3720
		3721	=head1 MACRO MAGIC
		3722
		3723	Libev can be compiled with a variety of options, the most fundamental
		3724	of which is C<EV_MULTIPLICITY>. This option determines whether (most)
		3725	functions and callbacks have an initial C<struct ev_loop *> argument.
		3726
		3727	To make it easier to write programs that cope with either variant, the
		3728	following macros are defined:
		3729
		3730	=over 4
		3731
		3732	=item C<EV_A>, C<EV_A_>
		3733
		3734	This provides the loop I<argument> for functions, if one is required ("ev
		3735	loop argument"). The C<EV_A> form is used when this is the sole argument,
		3736	C<EV_A_> is used when other arguments are following. Example:
		3737
		3738	ev_unref (EV_A);
		3739	ev_timer_add (EV_A_ watcher);
		3740	ev_run (EV_A_ 0);
		3741
		3742	It assumes the variable C<loop> of type C<struct ev_loop *> is in scope,
		3743	which is often provided by the following macro.
		3744
		3745	=item C<EV_P>, C<EV_P_>
		3746
		3747	This provides the loop I<parameter> for functions, if one is required ("ev
		3748	loop parameter"). The C<EV_P> form is used when this is the sole parameter,
		3749	C<EV_P_> is used when other parameters are following. Example:
		3750
		3751	// this is how ev_unref is being declared
		3752	static void ev_unref (EV_P);
		3753
		3754	// this is how you can declare your typical callback
		3755	static void cb (EV_P_ ev_timer *w, int revents)
		3756
		3757	It declares a parameter C<loop> of type C<struct ev_loop *>, quite
		3758	suitable for use with C<EV_A>.
		3759
		3760	=item C<EV_DEFAULT>, C<EV_DEFAULT_>
		3761
		3762	Similar to the other two macros, this gives you the value of the default
		3763	loop, if multiple loops are supported ("ev loop default").
		3764
		3765	=item C<EV_DEFAULT_UC>, C<EV_DEFAULT_UC_>
		3766
		3767	Usage identical to C<EV_DEFAULT> and C<EV_DEFAULT_>, but requires that the
		3768	default loop has been initialised (C<UC> == unchecked). Their behaviour
		3769	is undefined when the default loop has not been initialised by a previous
		3770	execution of C<EV_DEFAULT>, C<EV_DEFAULT_> or C<ev_default_init (...)>.
		3771
		3772	It is often prudent to use C<EV_DEFAULT> when initialising the first
		3773	watcher in a function but use C<EV_DEFAULT_UC> afterwards.
		3774
		3775	=back
		3776
		3777	Example: Declare and initialise a check watcher, utilising the above
		3778	macros so it will work regardless of whether multiple loops are supported
		3779	or not.
		3780
		3781	static void
		3782	check_cb (EV_P_ ev_timer *w, int revents)
		3783	{
		3784	ev_check_stop (EV_A_ w);
		3785	}
		3786
		3787	ev_check check;
		3788	ev_check_init (&check, check_cb);
		3789	ev_check_start (EV_DEFAULT_ &check);
		3790	ev_run (EV_DEFAULT_ 0);
		3791
		3792	=head1 EMBEDDING
		3793
		3794	Libev can (and often is) directly embedded into host
		3795	applications. Examples of applications that embed it include the Deliantra
		3796	Game Server, the EV perl module, the GNU Virtual Private Ethernet (gvpe)
		3797	and rxvt-unicode.
		3798
		3799	The goal is to enable you to just copy the necessary files into your
		3800	source directory without having to change even a single line in them, so
		3801	you can easily upgrade by simply copying (or having a checked-out copy of
		3802	libev somewhere in your source tree).
		3803
		3804	=head2 FILESETS
		3805
		3806	Depending on what features you need you need to include one or more sets of files
		3807	in your application.
		3808
		3809	=head3 CORE EVENT LOOP
		3810
		3811	To include only the libev core (all the C<ev_*> functions), with manual
		3812	configuration (no autoconf):
		3813
		3814	#define EV_STANDALONE 1
		3815	#include "ev.c"
		3816
		3817	This will automatically include F<ev.h>, too, and should be done in a
		3818	single C source file only to provide the function implementations. To use
		3819	it, do the same for F<ev.h> in all files wishing to use this API (best
		3820	done by writing a wrapper around F<ev.h> that you can include instead and
		3821	where you can put other configuration options):
		3822
		3823	#define EV_STANDALONE 1
		3824	#include "ev.h"
		3825
		3826	Both header files and implementation files can be compiled with a C++
		3827	compiler (at least, that's a stated goal, and breakage will be treated
		3828	as a bug).
		3829
		3830	You need the following files in your source tree, or in a directory
		3831	in your include path (e.g. in libev/ when using -Ilibev):
		3832
		3833	ev.h
		3834	ev.c
		3835	ev_vars.h
		3836	ev_wrap.h
		3837
		3838	ev_win32.c required on win32 platforms only
		3839
		3840	ev_select.c only when select backend is enabled (which is enabled by default)
		3841	ev_poll.c only when poll backend is enabled (disabled by default)
		3842	ev_epoll.c only when the epoll backend is enabled (disabled by default)
		3843	ev_kqueue.c only when the kqueue backend is enabled (disabled by default)
		3844	ev_port.c only when the solaris port backend is enabled (disabled by default)
		3845
		3846	F<ev.c> includes the backend files directly when enabled, so you only need
		3847	to compile this single file.
		3848
		3849	=head3 LIBEVENT COMPATIBILITY API
		3850
		3851	To include the libevent compatibility API, also include:
		3852
		3853	#include "event.c"
		3854
		3855	in the file including F<ev.c>, and:
		3856
		3857	#include "event.h"
		3858
		3859	in the files that want to use the libevent API. This also includes F<ev.h>.
		3860
		3861	You need the following additional files for this:
		3862
		3863	event.h
		3864	event.c
		3865
		3866	=head3 AUTOCONF SUPPORT
		3867
		3868	Instead of using C<EV_STANDALONE=1> and providing your configuration in
		3869	whatever way you want, you can also C<m4_include([libev.m4])> in your
		3870	F<configure.ac> and leave C<EV_STANDALONE> undefined. F<ev.c> will then
		3871	include F<config.h> and configure itself accordingly.
		3872
		3873	For this of course you need the m4 file:
		3874
		3875	libev.m4
		3876
		3877	=head2 PREPROCESSOR SYMBOLS/MACROS
		3878
		3879	Libev can be configured via a variety of preprocessor symbols you have to
		3880	define before including (or compiling) any of its files. The default in
		3881	the absence of autoconf is documented for every option.
		3882
		3883	Symbols marked with "(h)" do not change the ABI, and can have different
		3884	values when compiling libev vs. including F<ev.h>, so it is permissible
		3885	to redefine them before including F<ev.h> without breaking compatibility
		3886	to a compiled library. All other symbols change the ABI, which means all
		3887	users of libev and the libev code itself must be compiled with compatible
		3888	settings.
		3889
		3890	=over 4
		3891
		3892	=item EV_COMPAT3 (h)
		3893
		3894	Backwards compatibility is a major concern for libev. This is why this
		3895	release of libev comes with wrappers for the functions and symbols that
		3896	have been renamed between libev version 3 and 4.
		3897
		3898	You can disable these wrappers (to test compatibility with future
		3899	versions) by defining C<EV_COMPAT3> to C<0> when compiling your
		3900	sources. This has the additional advantage that you can drop the C<struct>
		3901	from C<struct ev_loop> declarations, as libev will provide an C<ev_loop>
		3902	typedef in that case.
		3903
		3904	In some future version, the default for C<EV_COMPAT3> will become C<0>,
		3905	and in some even more future version the compatibility code will be
		3906	removed completely.
		3907
		3908	=item EV_STANDALONE (h)
		3909
		3910	Must always be C<1> if you do not use autoconf configuration, which
		3911	keeps libev from including F<config.h>, and it also defines dummy
		3912	implementations for some libevent functions (such as logging, which is not
		3913	supported). It will also not define any of the structs usually found in
		3914	F<event.h> that are not directly supported by the libev core alone.
		3915
		3916	In standalone mode, libev will still try to automatically deduce the
		3917	configuration, but has to be more conservative.
		3918
		3919	=item EV_USE_MONOTONIC
		3920
		3921	If defined to be C<1>, libev will try to detect the availability of the
		3922	monotonic clock option at both compile time and runtime. Otherwise no
		3923	use of the monotonic clock option will be attempted. If you enable this,
		3924	you usually have to link against librt or something similar. Enabling it
		3925	when the functionality isn't available is safe, though, although you have
		3926	to make sure you link against any libraries where the C<clock_gettime>
		3927	function is hiding in (often F<-lrt>). See also C<EV_USE_CLOCK_SYSCALL>.
		3928
		3929	=item EV_USE_REALTIME
		3930
		3931	If defined to be C<1>, libev will try to detect the availability of the
		3932	real-time clock option at compile time (and assume its availability
		3933	at runtime if successful). Otherwise no use of the real-time clock
		3934	option will be attempted. This effectively replaces C<gettimeofday>
		3935	by C<clock_get (CLOCK_REALTIME, ...)> and will not normally affect
		3936	correctness. See the note about libraries in the description of
		3937	C<EV_USE_MONOTONIC>, though. Defaults to the opposite value of
		3938	C<EV_USE_CLOCK_SYSCALL>.
		3939
		3940	=item EV_USE_CLOCK_SYSCALL
		3941
		3942	If defined to be C<1>, libev will try to use a direct syscall instead
		3943	of calling the system-provided C<clock_gettime> function. This option
		3944	exists because on GNU/Linux, C<clock_gettime> is in C<librt>, but C<librt>
		3945	unconditionally pulls in C<libpthread>, slowing down single-threaded
		3946	programs needlessly. Using a direct syscall is slightly slower (in
		3947	theory), because no optimised vdso implementation can be used, but avoids
		3948	the pthread dependency. Defaults to C<1> on GNU/Linux with glibc 2.x or
		3949	higher, as it simplifies linking (no need for C<-lrt>).
		3950
		3951	=item EV_USE_NANOSLEEP
		3952
		3953	If defined to be C<1>, libev will assume that C<nanosleep ()> is available
		3954	and will use it for delays. Otherwise it will use C<select ()>.
		3955
		3956	=item EV_USE_EVENTFD
		3957
		3958	If defined to be C<1>, then libev will assume that C<eventfd ()> is
		3959	available and will probe for kernel support at runtime. This will improve
		3960	C<ev_signal> and C<ev_async> performance and reduce resource consumption.
		3961	If undefined, it will be enabled if the headers indicate GNU/Linux + Glibc
		3962	2.7 or newer, otherwise disabled.
		3963
		3964	=item EV_USE_SELECT
		3965
		3966	If undefined or defined to be C<1>, libev will compile in support for the
		3967	C<select>(2) backend. No attempt at auto-detection will be done: if no
		3968	other method takes over, select will be it. Otherwise the select backend
		3969	will not be compiled in.
		3970
		3971	=item EV_SELECT_USE_FD_SET
		3972
		3973	If defined to C<1>, then the select backend will use the system C<fd_set>
		3974	structure. This is useful if libev doesn't compile due to a missing
		3975	C<NFDBITS> or C<fd_mask> definition or it mis-guesses the bitset layout
		3976	on exotic systems. This usually limits the range of file descriptors to
		3977	some low limit such as 1024 or might have other limitations (winsocket
		3978	only allows 64 sockets). The C<FD_SETSIZE> macro, set before compilation,
		3979	configures the maximum size of the C<fd_set>.
		3980
		3981	=item EV_SELECT_IS_WINSOCKET
		3982
		3983	When defined to C<1>, the select backend will assume that
		3984	select/socket/connect etc. don't understand file descriptors but
		3985	wants osf handles on win32 (this is the case when the select to
		3986	be used is the winsock select). This means that it will call
		3987	C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,
		3988	it is assumed that all these functions actually work on fds, even
		3989	on win32. Should not be defined on non-win32 platforms.
		3990
		3991	=item EV_FD_TO_WIN32_HANDLE(fd)
		3992
		3993	If C<EV_SELECT_IS_WINSOCKET> is enabled, then libev needs a way to map
		3994	file descriptors to socket handles. When not defining this symbol (the
		3995	default), then libev will call C<_get_osfhandle>, which is usually
		3996	correct. In some cases, programs use their own file descriptor management,
		3997	in which case they can provide this function to map fds to socket handles.
		3998
		3999	=item EV_WIN32_HANDLE_TO_FD(handle)
		4000
		4001	If C<EV_SELECT_IS_WINSOCKET> then libev maps handles to file descriptors
		4002	using the standard C<_open_osfhandle> function. For programs implementing
		4003	their own fd to handle mapping, overwriting this function makes it easier
		4004	to do so. This can be done by defining this macro to an appropriate value.
		4005
		4006	=item EV_WIN32_CLOSE_FD(fd)
		4007
		4008	If programs implement their own fd to handle mapping on win32, then this
		4009	macro can be used to override the C<close> function, useful to unregister
		4010	file descriptors again. Note that the replacement function has to close
		4011	the underlying OS handle.
		4012
		4013	=item EV_USE_POLL
		4014
		4015	If defined to be C<1>, libev will compile in support for the C<poll>(2)
		4016	backend. Otherwise it will be enabled on non-win32 platforms. It
		4017	takes precedence over select.
		4018
		4019	=item EV_USE_EPOLL
		4020
		4021	If defined to be C<1>, libev will compile in support for the Linux
		4022	C<epoll>(7) backend. Its availability will be detected at runtime,
		4023	otherwise another method will be used as fallback. This is the preferred
		4024	backend for GNU/Linux systems. If undefined, it will be enabled if the
		4025	headers indicate GNU/Linux + Glibc 2.4 or newer, otherwise disabled.
		4026
		4027	=item EV_USE_KQUEUE
		4028
		4029	If defined to be C<1>, libev will compile in support for the BSD style
		4030	C<kqueue>(2) backend. Its actual availability will be detected at runtime,
		4031	otherwise another method will be used as fallback. This is the preferred
		4032	backend for BSD and BSD-like systems, although on most BSDs kqueue only
		4033	supports some types of fds correctly (the only platform we found that
		4034	supports ptys for example was NetBSD), so kqueue might be compiled in, but
		4035	not be used unless explicitly requested. The best way to use it is to find
		4036	out whether kqueue supports your type of fd properly and use an embedded
		4037	kqueue loop.
		4038
		4039	=item EV_USE_PORT
		4040
		4041	If defined to be C<1>, libev will compile in support for the Solaris
		4042	10 port style backend. Its availability will be detected at runtime,
		4043	otherwise another method will be used as fallback. This is the preferred
		4044	backend for Solaris 10 systems.
		4045
		4046	=item EV_USE_DEVPOLL
		4047
		4048	Reserved for future expansion, works like the USE symbols above.
		4049
		4050	=item EV_USE_INOTIFY
		4051
		4052	If defined to be C<1>, libev will compile in support for the Linux inotify
		4053	interface to speed up C<ev_stat> watchers. Its actual availability will
		4054	be detected at runtime. If undefined, it will be enabled if the headers
		4055	indicate GNU/Linux + Glibc 2.4 or newer, otherwise disabled.
		4056
		4057	=item EV_ATOMIC_T
		4058
		4059	Libev requires an integer type (suitable for storing C<0> or C<1>) whose
		4060	access is atomic with respect to other threads or signal contexts. No such
		4061	type is easily found in the C language, so you can provide your own type
		4062	that you know is safe for your purposes. It is used both for signal handler "locking"
		4063	as well as for signal and thread safety in C<ev_async> watchers.
		4064
		4065	In the absence of this define, libev will use C<sig_atomic_t volatile>
		4066	(from F<signal.h>), which is usually good enough on most platforms.
		4067
		4068	=item EV_H (h)
		4069
		4070	The name of the F<ev.h> header file used to include it. The default if
		4071	undefined is C<"ev.h"> in F<event.h>, F<ev.c> and F<ev++.h>. This can be
		4072	used to virtually rename the F<ev.h> header file in case of conflicts.
		4073
		4074	=item EV_CONFIG_H (h)
		4075
		4076	If C<EV_STANDALONE> isn't C<1>, this variable can be used to override
		4077	F<ev.c>'s idea of where to find the F<config.h> file, similarly to
		4078	C<EV_H>, above.
		4079
		4080	=item EV_EVENT_H (h)
		4081
		4082	Similarly to C<EV_H>, this macro can be used to override F<event.c>'s idea
		4083	of how the F<event.h> header can be found, the default is C<"event.h">.
		4084
		4085	=item EV_PROTOTYPES (h)
		4086
		4087	If defined to be C<0>, then F<ev.h> will not define any function
		4088	prototypes, but still define all the structs and other symbols. This is
		4089	occasionally useful if you want to provide your own wrapper functions
		4090	around libev functions.
		4091
		4092	=item EV_MULTIPLICITY
		4093
		4094	If undefined or defined to C<1>, then all event-loop-specific functions
		4095	will have the C<struct ev_loop *> as first argument, and you can create
		4096	additional independent event loops. Otherwise there will be no support
		4097	for multiple event loops and there is no first event loop pointer
		4098	argument. Instead, all functions act on the single default loop.
		4099
		4100	=item EV_MINPRI
		4101
		4102	=item EV_MAXPRI
		4103
		4104	The range of allowed priorities. C<EV_MINPRI> must be smaller or equal to
		4105	C<EV_MAXPRI>, but otherwise there are no non-obvious limitations. You can
		4106	provide for more priorities by overriding those symbols (usually defined
		4107	to be C<-2> and C<2>, respectively).
		4108
		4109	When doing priority-based operations, libev usually has to linearly search
		4110	all the priorities, so having many of them (hundreds) uses a lot of space
		4111	and time, so using the defaults of five priorities (-2 .. +2) is usually
		4112	fine.
		4113
		4114	If your embedding application does not need any priorities, defining these
		4115	both to C<0> will save some memory and CPU.
		4116
		4117	=item EV_PERIODIC_ENABLE, EV_IDLE_ENABLE, EV_EMBED_ENABLE, EV_STAT_ENABLE,
		4118	EV_PREPARE_ENABLE, EV_CHECK_ENABLE, EV_FORK_ENABLE, EV_SIGNAL_ENABLE,
		4119	EV_ASYNC_ENABLE, EV_CHILD_ENABLE.
		4120
		4121	If undefined or defined to be C<1> (and the platform supports it), then
		4122	the respective watcher type is supported. If defined to be C<0>, then it
		4123	is not. Disabling watcher types mainly saves code size.
		4124
		4125	=item EV_FEATURES
		4126
		4127	If you need to shave off some kilobytes of code at the expense of some
		4128	speed (but with the full API), you can define this symbol to request
		4129	certain subsets of functionality. The default is to enable all features
		4130	that can be enabled on the platform.
		4131
		4132	A typical way to use this symbol is to define it to C<0> (or to a bitset
		4133	with some broad features you want) and then selectively re-enable
		4134	additional parts you want, for example if you want everything minimal,
		4135	but multiple event loop support, async and child watchers and the poll
		4136	backend, use this:
		4137
		4138	#define EV_FEATURES 0
		4139	#define EV_MULTIPLICITY 1
		4140	#define EV_USE_POLL 1
		4141	#define EV_CHILD_ENABLE 1
		4142	#define EV_ASYNC_ENABLE 1
		4143
		4144	The actual value is a bitset, it can be a combination of the following
		4145	values:
		4146
		4147	=over 4
		4148
		4149	=item C<1> - faster/larger code
		4150
		4151	Use larger code to speed up some operations.
		4152
		4153	Currently this is used to override some inlining decisions (enlarging the
		4154	code size by roughly 30% on amd64).
		4155
		4156	When optimising for size, use of compiler flags such as C<-Os> with
		4157	gcc is recommended, as well as C<-DNDEBUG>, as libev contains a number of
		4158	assertions.
		4159
		4160	=item C<2> - faster/larger data structures
		4161
		4162	Replaces the small 2-heap for timer management by a faster 4-heap, larger
		4163	hash table sizes and so on. This will usually further increase code size
		4164	and can additionally have an effect on the size of data structures at
		4165	runtime.
		4166
		4167	=item C<4> - full API configuration
		4168
		4169	This enables priorities (sets C<EV_MAXPRI>=2 and C<EV_MINPRI>=-2), and
		4170	enables multiplicity (C<EV_MULTIPLICITY>=1).
		4171
		4172	=item C<8> - full API
		4173
		4174	This enables a lot of the "lesser used" API functions. See C<ev.h> for
		4175	details on which parts of the API are still available without this
		4176	feature, and do not complain if this subset changes over time.
		4177
		4178	=item C<16> - enable all optional watcher types
		4179
		4180	Enables all optional watcher types. If you want to selectively enable
		4181	only some watcher types other than I/O and timers (e.g. prepare,
		4182	embed, async, child...) you can enable them manually by defining
		4183	C<EV_watchertype_ENABLE> to C<1> instead.
		4184
		4185	=item C<32> - enable all backends
		4186
		4187	This enables all backends - without this feature, you need to enable at
		4188	least one backend manually (C<EV_USE_SELECT> is a good choice).
		4189
		4190	=item C<64> - enable OS-specific "helper" APIs
		4191
		4192	Enable inotify, eventfd, signalfd and similar OS-specific helper APIs by
		4193	default.
		4194
		4195	=back
		4196
		4197	Compiling with C<gcc -Os -DEV_STANDALONE -DEV_USE_EPOLL=1 -DEV_FEATURES=0>
		4198	reduces the compiled size of libev from 24.7Kb code/2.8Kb data to 6.5Kb
		4199	code/0.3Kb data on my GNU/Linux amd64 system, while still giving you I/O
		4200	watchers, timers and monotonic clock support.
		4201
		4202	With an intelligent-enough linker (gcc+binutils are intelligent enough
		4203	when you use C<-Wl,--gc-sections -ffunction-sections>) functions unused by
		4204	your program might be left out as well - a binary starting a timer and an
		4205	I/O watcher then might come out at only 5Kb.
		4206
		4207	=item EV_AVOID_STDIO
		4208
		4209	If this is set to C<1> at compiletime, then libev will avoid using stdio
		4210	functions (printf, scanf, perror etc.). This will increase the code size
		4211	somewhat, but if your program doesn't otherwise depend on stdio and your
		4212	libc allows it, this avoids linking in the stdio library which is quite
		4213	big.
		4214
		4215	Note that error messages might become less precise when this option is
		4216	enabled.
		4217
		4218	=item EV_NSIG
		4219
		4220	The highest supported signal number, +1 (or, the number of
		4221	signals): Normally, libev tries to deduce the maximum number of signals
		4222	automatically, but sometimes this fails, in which case it can be
		4223	specified. Also, using a lower number than detected (C<32> should be
		4224	good for about any system in existence) can save some memory, as libev
		4225	statically allocates some 12-24 bytes per signal number.
		4226
		4227	=item EV_PID_HASHSIZE
		4228
		4229	C<ev_child> watchers use a small hash table to distribute workload by
		4230	pid. The default size is C<16> (or C<1> with C<EV_FEATURES> disabled),
		4231	usually more than enough. If you need to manage thousands of children you
		4232	might want to increase this value (I<must> be a power of two).
		4233
		4234	=item EV_INOTIFY_HASHSIZE
		4235
		4236	C<ev_stat> watchers use a small hash table to distribute workload by
		4237	inotify watch id. The default size is C<16> (or C<1> with C<EV_FEATURES>
		4238	disabled), usually more than enough. If you need to manage thousands of
		4239	C<ev_stat> watchers you might want to increase this value (I<must> be a
		4240	power of two).
		4241
		4242	=item EV_USE_4HEAP
		4243
		4244	Heaps are not very cache-efficient. To improve the cache-efficiency of the
		4245	timer and periodics heaps, libev uses a 4-heap when this symbol is defined
		4246	to C<1>. The 4-heap uses more complicated (longer) code but has noticeably
		4247	faster performance with many (thousands) of watchers.
		4248
		4249	The default is C<1>, unless C<EV_FEATURES> overrides it, in which case it
		4250	will be C<0>.
		4251
		4252	=item EV_HEAP_CACHE_AT
		4253
		4254	Heaps are not very cache-efficient. To improve the cache-efficiency of the
		4255	timer and periodics heaps, libev can cache the timestamp (I<at>) within
		4256	the heap structure (selected by defining C<EV_HEAP_CACHE_AT> to C<1>),
		4257	which uses 8-12 bytes more per watcher and a few hundred bytes more code,
		4258	but avoids random read accesses on heap changes. This improves performance
		4259	noticeably with many (hundreds) of watchers.
		4260
		4261	The default is C<1>, unless C<EV_FEATURES> overrides it, in which case it
		4262	will be C<0>.
		4263
		4264	=item EV_VERIFY
		4265
		4266	Controls how much internal verification (see C<ev_verify ()>) will
		4267	be done: If set to C<0>, no internal verification code will be compiled
		4268	in. If set to C<1>, then verification code will be compiled in, but not
		4269	called. If set to C<2>, then the internal verification code will be
		4270	called once per loop, which can slow down libev. If set to C<3>, then the
		4271	verification code will be called very frequently, which will slow down
		4272	libev considerably.
		4273
		4274	The default is C<1>, unless C<EV_FEATURES> overrides it, in which case it
		4275	will be C<0>.
		4276
		4277	=item EV_COMMON
		4278
		4279	By default, all watchers have a C<void *data> member. By redefining
		4280	this macro to something else you can include more and other types of
		4281	members. You have to define it each time you include one of the files,
		4282	though, and it must be identical each time.
		4283
		4284	For example, the perl EV module uses something like this:
		4285
		4286	#define EV_COMMON \
		4287	SV *self; /* contains this struct */ \
		4288	SV *cb_sv, *fh /* note no trailing ";" */
		4289
		4290	=item EV_CB_DECLARE (type)
		4291
		4292	=item EV_CB_INVOKE (watcher, revents)
		4293
		4294	=item ev_set_cb (ev, cb)
		4295
		4296	Can be used to change the callback member declaration in each watcher,
		4297	and the way callbacks are invoked and set. Must expand to a struct member
		4298	definition and a statement, respectively. See the F<ev.h> header file for
		4299	their default definitions. One possible use for overriding these is to
		4300	avoid the C<struct ev_loop *> as first argument in all cases, or to use
		4301	method calls instead of plain function calls in C++.
		4302
		4303	=back
		4304
		4305	=head2 EXPORTED API SYMBOLS
		4306
		4307	If you need to re-export the API (e.g. via a DLL) and you need a list of
		4308	exported symbols, you can use the provided F<Symbol.*> files which list
		4309	all public symbols, one per line:
		4310
		4311	Symbols.ev for libev proper
		4312	Symbols.event for the libevent emulation
		4313
		4314	This can also be used to rename all public symbols to avoid clashes with
		4315	multiple versions of libev linked together (which is obviously bad in
		4316	itself, but sometimes it is inconvenient to avoid this).
		4317
		4318	A sed command like this will create wrapper C<#define>'s that you need to
		4319	include before including F<ev.h>:
		4320
		4321	<Symbols.ev sed -e "s/.*/#define & myprefix_&/" >wrap.h
		4322
		4323	This would create a file F<wrap.h> which essentially looks like this:
		4324
		4325	#define ev_backend myprefix_ev_backend
		4326	#define ev_check_start myprefix_ev_check_start
		4327	#define ev_check_stop myprefix_ev_check_stop
		4328	...
		4329
		4330	=head2 EXAMPLES
		4331
		4332	For a real-world example of a program the includes libev
		4333	verbatim, you can have a look at the EV perl module
		4334	(L<http://software.schmorp.de/pkg/EV.html>). It has the libev files in
		4335	the F<libev/> subdirectory and includes them in the F<EV/EVAPI.h> (public
		4336	interface) and F<EV.xs> (implementation) files. Only the F<EV.xs> file
		4337	will be compiled. It is pretty complex because it provides its own header
		4338	file.
		4339
		4340	The usage in rxvt-unicode is simpler. It has a F<ev_cpp.h> header file
		4341	that everybody includes and which overrides some configure choices:
		4342
		4343	#define EV_FEATURES 8
		4344	#define EV_USE_SELECT 1
		4345	#define EV_PREPARE_ENABLE 1
		4346	#define EV_IDLE_ENABLE 1
		4347	#define EV_SIGNAL_ENABLE 1
		4348	#define EV_CHILD_ENABLE 1
		4349	#define EV_USE_STDEXCEPT 0
		4350	#define EV_CONFIG_H <config.h>
		4351
		4352	#include "ev++.h"
		4353
		4354	And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:
		4355
		4356	#include "ev_cpp.h"
		4357	#include "ev.c"
		4358
		4359	=head1 INTERACTION WITH OTHER PROGRAMS OR LIBRARIES
		4360
		4361	=head2 THREADS AND COROUTINES
		4362
		4363	=head3 THREADS
		4364
		4365	All libev functions are reentrant and thread-safe unless explicitly
		4366	documented otherwise, but libev implements no locking itself. This means
		4367	that you can use as many loops as you want in parallel, as long as there
		4368	are no concurrent calls into any libev function with the same loop
		4369	parameter (C<ev_default_*> calls have an implicit default loop parameter,
		4370	of course): libev guarantees that different event loops share no data
		4371	structures that need any locking.
		4372
		4373	Or to put it differently: calls with different loop parameters can be done
		4374	concurrently from multiple threads, calls with the same loop parameter
		4375	must be done serially (but can be done from different threads, as long as
		4376	only one thread ever is inside a call at any point in time, e.g. by using
		4377	a mutex per loop).
		4378
		4379	Specifically to support threads (and signal handlers), libev implements
		4380	so-called C<ev_async> watchers, which allow some limited form of
		4381	concurrency on the same event loop, namely waking it up "from the
		4382	outside".
		4383
		4384	If you want to know which design (one loop, locking, or multiple loops
		4385	without or something else still) is best for your problem, then I cannot
		4386	help you, but here is some generic advice:
		4387
		4388	=over 4
		4389
		4390	=item * most applications have a main thread: use the default libev loop
		4391	in that thread, or create a separate thread running only the default loop.
		4392
		4393	This helps integrating other libraries or software modules that use libev
		4394	themselves and don't care/know about threading.
		4395
		4396	=item * one loop per thread is usually a good model.
		4397
		4398	Doing this is almost never wrong, sometimes a better-performance model
		4399	exists, but it is always a good start.
		4400
		4401	=item * other models exist, such as the leader/follower pattern, where one
		4402	loop is handed through multiple threads in a kind of round-robin fashion.
		4403
		4404	Choosing a model is hard - look around, learn, know that usually you can do
		4405	better than you currently do :-)
		4406
		4407	=item * often you need to talk to some other thread which blocks in the
		4408	event loop.
		4409
		4410	C<ev_async> watchers can be used to wake them up from other threads safely
		4411	(or from signal contexts...).
		4412
		4413	An example use would be to communicate signals or other events that only
		4414	work in the default loop by registering the signal watcher with the
		4415	default loop and triggering an C<ev_async> watcher from the default loop
		4416	watcher callback into the event loop interested in the signal.
		4417
		4418	=back
		4419
		4420	=head4 THREAD LOCKING EXAMPLE
		4421
		4422	Here is a fictitious example of how to run an event loop in a different
		4423	thread than where callbacks are being invoked and watchers are
		4424	created/added/removed.
		4425
		4426	For a real-world example, see the C<EV::Loop::Async> perl module,
		4427	which uses exactly this technique (which is suited for many high-level
		4428	languages).
		4429
		4430	The example uses a pthread mutex to protect the loop data, a condition
		4431	variable to wait for callback invocations, an async watcher to notify the
		4432	event loop thread and an unspecified mechanism to wake up the main thread.
		4433
		4434	First, you need to associate some data with the event loop:
		4435
		4436	typedef struct {
		4437	mutex_t lock; /* global loop lock */
		4438	ev_async async_w;
		4439	thread_t tid;
		4440	cond_t invoke_cv;
		4441	} userdata;
		4442
		4443	void prepare_loop (EV_P)
		4444	{
		4445	// for simplicity, we use a static userdata struct.
		4446	static userdata u;
		4447
		4448	ev_async_init (&u->async_w, async_cb);
		4449	ev_async_start (EV_A_ &u->async_w);
		4450
		4451	pthread_mutex_init (&u->lock, 0);
		4452	pthread_cond_init (&u->invoke_cv, 0);
		4453
		4454	// now associate this with the loop
		4455	ev_set_userdata (EV_A_ u);
		4456	ev_set_invoke_pending_cb (EV_A_ l_invoke);
		4457	ev_set_loop_release_cb (EV_A_ l_release, l_acquire);
		4458
		4459	// then create the thread running ev_loop
		4460	pthread_create (&u->tid, 0, l_run, EV_A);
		4461	}
		4462
		4463	The callback for the C<ev_async> watcher does nothing: the watcher is used
		4464	solely to wake up the event loop so it takes notice of any new watchers
		4465	that might have been added:
		4466
		4467	static void
		4468	async_cb (EV_P_ ev_async *w, int revents)
		4469	{
		4470	// just used for the side effects
		4471	}
		4472
		4473	The C<l_release> and C<l_acquire> callbacks simply unlock/lock the mutex
		4474	protecting the loop data, respectively.
		4475
		4476	static void
		4477	l_release (EV_P)
		4478	{
		4479	userdata *u = ev_userdata (EV_A);
		4480	pthread_mutex_unlock (&u->lock);
		4481	}
		4482
		4483	static void
		4484	l_acquire (EV_P)
		4485	{
		4486	userdata *u = ev_userdata (EV_A);
		4487	pthread_mutex_lock (&u->lock);
		4488	}
		4489
		4490	The event loop thread first acquires the mutex, and then jumps straight
		4491	into C<ev_run>:
		4492
		4493	void *
		4494	l_run (void *thr_arg)
		4495	{
		4496	struct ev_loop *loop = (struct ev_loop *)thr_arg;
		4497
		4498	l_acquire (EV_A);
		4499	pthread_setcanceltype (PTHREAD_CANCEL_ASYNCHRONOUS, 0);
		4500	ev_run (EV_A_ 0);
		4501	l_release (EV_A);
		4502
		4503	return 0;
		4504	}
		4505
		4506	Instead of invoking all pending watchers, the C<l_invoke> callback will
		4507	signal the main thread via some unspecified mechanism (signals? pipe
		4508	writes? C<Async::Interrupt>?) and then waits until all pending watchers
		4509	have been called (in a while loop because a) spurious wakeups are possible
		4510	and b) skipping inter-thread-communication when there are no pending
		4511	watchers is very beneficial):
		4512
		4513	static void
		4514	l_invoke (EV_P)
		4515	{
		4516	userdata *u = ev_userdata (EV_A);
		4517
		4518	while (ev_pending_count (EV_A))
		4519	{
		4520	wake_up_other_thread_in_some_magic_or_not_so_magic_way ();
		4521	pthread_cond_wait (&u->invoke_cv, &u->lock);
		4522	}
		4523	}
		4524
		4525	Now, whenever the main thread gets told to invoke pending watchers, it
		4526	will grab the lock, call C<ev_invoke_pending> and then signal the loop
		4527	thread to continue:
		4528
		4529	static void
		4530	real_invoke_pending (EV_P)
		4531	{
		4532	userdata *u = ev_userdata (EV_A);
		4533
		4534	pthread_mutex_lock (&u->lock);
		4535	ev_invoke_pending (EV_A);
		4536	pthread_cond_signal (&u->invoke_cv);
		4537	pthread_mutex_unlock (&u->lock);
		4538	}
		4539
		4540	Whenever you want to start/stop a watcher or do other modifications to an
		4541	event loop, you will now have to lock:
		4542
		4543	ev_timer timeout_watcher;
		4544	userdata *u = ev_userdata (EV_A);
		4545
		4546	ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);
		4547
		4548	pthread_mutex_lock (&u->lock);
		4549	ev_timer_start (EV_A_ &timeout_watcher);
		4550	ev_async_send (EV_A_ &u->async_w);
		4551	pthread_mutex_unlock (&u->lock);
		4552
		4553	Note that sending the C<ev_async> watcher is required because otherwise
		4554	an event loop currently blocking in the kernel will have no knowledge
		4555	about the newly added timer. By waking up the loop it will pick up any new
		4556	watchers in the next event loop iteration.
		4557
		4558	=head3 COROUTINES
		4559
		4560	Libev is very accommodating to coroutines ("cooperative threads"):
		4561	libev fully supports nesting calls to its functions from different
		4562	coroutines (e.g. you can call C<ev_run> on the same loop from two
		4563	different coroutines, and switch freely between both coroutines running
		4564	the loop, as long as you don't confuse yourself). The only exception is
		4565	that you must not do this from C<ev_periodic> reschedule callbacks.
		4566
		4567	Care has been taken to ensure that libev does not keep local state inside
		4568	C<ev_run>, and other calls do not usually allow for coroutine switches as
		4569	they do not call any callbacks.
		4570
		4571	=head2 COMPILER WARNINGS
		4572
		4573	Depending on your compiler and compiler settings, you might get no or a
		4574	lot of warnings when compiling libev code. Some people are apparently
		4575	scared by this.
		4576
		4577	However, these are unavoidable for many reasons. For one, each compiler
		4578	has different warnings, and each user has different tastes regarding
		4579	warning options. "Warn-free" code therefore cannot be a goal except when
		4580	targeting a specific compiler and compiler-version.
		4581
		4582	Another reason is that some compiler warnings require elaborate
		4583	workarounds, or other changes to the code that make it less clear and less
		4584	maintainable.
		4585
		4586	And of course, some compiler warnings are just plain stupid, or simply
		4587	wrong (because they don't actually warn about the condition their message
		4588	seems to warn about). For example, certain older gcc versions had some
		4589	warnings that resulted in an extreme number of false positives. These have
		4590	been fixed, but some people still insist on making code warn-free with
		4591	such buggy versions.
		4592
		4593	While libev is written to generate as few warnings as possible,
		4594	"warn-free" code is not a goal, and it is recommended not to build libev
		4595	with any compiler warnings enabled unless you are prepared to cope with
		4596	them (e.g. by ignoring them). Remember that warnings are just that:
		4597	warnings, not errors, or proof of bugs.
		4598
		4599
		4600	=head2 VALGRIND
		4601
		4602	Valgrind has a special section here because it is a popular tool that is
		4603	highly useful. Unfortunately, valgrind reports are very hard to interpret.
		4604
		4605	If you think you found a bug (memory leak, uninitialised data access etc.)
		4606	in libev, then check twice: If valgrind reports something like:
		4607
		4608	==2274== definitely lost: 0 bytes in 0 blocks.
		4609	==2274== possibly lost: 0 bytes in 0 blocks.
		4610	==2274== still reachable: 256 bytes in 1 blocks.
		4611
		4612	Then there is no memory leak, just as memory accounted to global variables
		4613	is not a memleak - the memory is still being referenced, and didn't leak.
		4614
		4615	Similarly, under some circumstances, valgrind might report kernel bugs
		4616	as if it were a bug in libev (e.g. in realloc or in the poll backend,
		4617	although an acceptable workaround has been found here), or it might be
		4618	confused.
		4619
		4620	Keep in mind that valgrind is a very good tool, but only a tool. Don't
		4621	make it into some kind of religion.
		4622
		4623	If you are unsure about something, feel free to contact the mailing list
		4624	with the full valgrind report and an explanation on why you think this
		4625	is a bug in libev (best check the archives, too :). However, don't be
		4626	annoyed when you get a brisk "this is no bug" answer and take the chance
		4627	of learning how to interpret valgrind properly.
		4628
		4629	If you need, for some reason, empty reports from valgrind for your project
		4630	I suggest using suppression lists.
		4631
		4632
		4633	=head1 PORTABILITY NOTES
		4634
		4635	=head2 GNU/LINUX 32 BIT LIMITATIONS
		4636
		4637	GNU/Linux is the only common platform that supports 64 bit file/large file
		4638	interfaces but I<disables> them by default.
		4639
		4640	That means that libev compiled in the default environment doesn't support
		4641	files larger than 2GiB or so, which mainly affects C<ev_stat> watchers.
		4642
		4643	Unfortunately, many programs try to work around this GNU/Linux issue
		4644	by enabling the large file API, which makes them incompatible with the
		4645	standard libev compiled for their system.
		4646
		4647	Likewise, libev cannot enable the large file API itself as this would
		4648	suddenly make it incompatible to the default compile time environment,
		4649	i.e. all programs not using special compile switches.
		4650
		4651	=head2 OS/X AND DARWIN BUGS
		4652
		4653	The whole thing is a bug if you ask me - basically any system interface
		4654	you touch is broken, whether it is locales, poll, kqueue or even the
		4655	OpenGL drivers.
		4656
		4657	=head3 C<kqueue> is buggy
		4658
		4659	The kqueue syscall is broken in all known versions - most versions support
		4660	only sockets, many support pipes.
		4661
		4662	Libev tries to work around this by not using C<kqueue> by default on this
		4663	rotten platform, but of course you can still ask for it when creating a
		4664	loop - embedding a socket-only kqueue loop into a select-based one is
		4665	probably going to work well.
		4666
		4667	=head3 C<poll> is buggy
		4668
		4669	Instead of fixing C<kqueue>, Apple replaced their (working) C<poll>
		4670	implementation by something calling C<kqueue> internally around the 10.5.6
		4671	release, so now C<kqueue> I<and> C<poll> are broken.
		4672
		4673	Libev tries to work around this by not using C<poll> by default on
		4674	this rotten platform, but of course you can still ask for it when creating
		4675	a loop.
		4676
		4677	=head3 C<select> is buggy
		4678
		4679	All that's left is C<select>, and of course Apple found a way to fuck this
		4680	one up as well: On OS/X, C<select> actively limits the number of file
		4681	descriptors you can pass in to 1024 - your program suddenly crashes when
		4682	you use more.
		4683
		4684	There is an undocumented "workaround" for this - defining
		4685	C<_DARWIN_UNLIMITED_SELECT>, which libev tries to use, so select I<should>
		4686	work on OS/X.
		4687
		4688	=head2 SOLARIS PROBLEMS AND WORKAROUNDS
		4689
		4690	=head3 C<errno> reentrancy
		4691
		4692	The default compile environment on Solaris is unfortunately so
		4693	thread-unsafe that you can't even use components/libraries compiled
		4694	without C<-D_REENTRANT> in a threaded program, which, of course, isn't
		4695	defined by default. A valid, if stupid, implementation choice.
		4696
		4697	If you want to use libev in threaded environments you have to make sure
		4698	it's compiled with C<_REENTRANT> defined.
		4699
		4700	=head3 Event port backend
		4701
		4702	The scalable event interface for Solaris is called "event
		4703	ports". Unfortunately, this mechanism is very buggy in all major
		4704	releases. If you run into high CPU usage, your program freezes or you get
		4705	a large number of spurious wakeups, make sure you have all the relevant
		4706	and latest kernel patches applied. No, I don't know which ones, but there
		4707	are multiple ones to apply, and afterwards, event ports actually work
		4708	great.
		4709
		4710	If you can't get it to work, you can try running the program by setting
		4711	the environment variable C<LIBEV_FLAGS=3> to only allow C<poll> and
		4712	C<select> backends.
		4713
		4714	=head2 AIX POLL BUG
		4715
		4716	AIX unfortunately has a broken C<poll.h> header. Libev works around
		4717	this by trying to avoid the poll backend altogether (i.e. it's not even
		4718	compiled in), which normally isn't a big problem as C<select> works fine
		4719	with large bitsets on AIX, and AIX is dead anyway.
		4720
		4721	=head2 WIN32 PLATFORM LIMITATIONS AND WORKAROUNDS
		4722
		4723	=head3 General issues
		4724
		4725	Win32 doesn't support any of the standards (e.g. POSIX) that libev
		4726	requires, and its I/O model is fundamentally incompatible with the POSIX
		4727	model. Libev still offers limited functionality on this platform in
		4728	the form of the C<EVBACKEND_SELECT> backend, and only supports socket
		4729	descriptors. This only applies when using Win32 natively, not when using
		4730	e.g. cygwin. Actually, it only applies to the microsofts own compilers,
		4731	as every compielr comes with a slightly differently broken/incompatible
		4732	environment.
		4733
		4734	Lifting these limitations would basically require the full
		4735	re-implementation of the I/O system. If you are into this kind of thing,
		4736	then note that glib does exactly that for you in a very portable way (note
		4737	also that glib is the slowest event library known to man).
		4738
		4739	There is no supported compilation method available on windows except
		4740	embedding it into other applications.
		4741
		4742	Sensible signal handling is officially unsupported by Microsoft - libev
		4743	tries its best, but under most conditions, signals will simply not work.
		4744
		4745	Not a libev limitation but worth mentioning: windows apparently doesn't
		4746	accept large writes: instead of resulting in a partial write, windows will
		4747	either accept everything or return C<ENOBUFS> if the buffer is too large,
		4748	so make sure you only write small amounts into your sockets (less than a
		4749	megabyte seems safe, but this apparently depends on the amount of memory
		4750	available).
		4751
		4752	Due to the many, low, and arbitrary limits on the win32 platform and
		4753	the abysmal performance of winsockets, using a large number of sockets
		4754	is not recommended (and not reasonable). If your program needs to use
		4755	more than a hundred or so sockets, then likely it needs to use a totally
		4756	different implementation for windows, as libev offers the POSIX readiness
		4757	notification model, which cannot be implemented efficiently on windows
		4758	(due to Microsoft monopoly games).
		4759
		4760	A typical way to use libev under windows is to embed it (see the embedding
		4761	section for details) and use the following F<evwrap.h> header file instead
		4762	of F<ev.h>:
		4763
		4764	#define EV_STANDALONE /* keeps ev from requiring config.h */
		4765	#define EV_SELECT_IS_WINSOCKET 1 /* configure libev for windows select */
		4766
		4767	#include "ev.h"
		4768
		4769	And compile the following F<evwrap.c> file into your project (make sure
		4770	you do I<not> compile the F<ev.c> or any other embedded source files!):
		4771
		4772	#include "evwrap.h"
		4773	#include "ev.c"
		4774
		4775	=head3 The winsocket C<select> function
		4776
		4777	The winsocket C<select> function doesn't follow POSIX in that it
		4778	requires socket I<handles> and not socket I<file descriptors> (it is
		4779	also extremely buggy). This makes select very inefficient, and also
		4780	requires a mapping from file descriptors to socket handles (the Microsoft
		4781	C runtime provides the function C<_open_osfhandle> for this). See the
		4782	discussion of the C<EV_SELECT_USE_FD_SET>, C<EV_SELECT_IS_WINSOCKET> and
		4783	C<EV_FD_TO_WIN32_HANDLE> preprocessor symbols for more info.
		4784
		4785	The configuration for a "naked" win32 using the Microsoft runtime
		4786	libraries and raw winsocket select is:
		4787
		4788	#define EV_USE_SELECT 1
		4789	#define EV_SELECT_IS_WINSOCKET 1 /* forces EV_SELECT_USE_FD_SET, too */
		4790
		4791	Note that winsockets handling of fd sets is O(n), so you can easily get a
		4792	complexity in the O(n²) range when using win32.
		4793
		4794	=head3 Limited number of file descriptors
		4795
		4796	Windows has numerous arbitrary (and low) limits on things.
		4797
		4798	Early versions of winsocket's select only supported waiting for a maximum
		4799	of C<64> handles (probably owning to the fact that all windows kernels
		4800	can only wait for C<64> things at the same time internally; Microsoft
		4801	recommends spawning a chain of threads and wait for 63 handles and the
		4802	previous thread in each. Sounds great!).
		4803
		4804	Newer versions support more handles, but you need to define C<FD_SETSIZE>
		4805	to some high number (e.g. C<2048>) before compiling the winsocket select
		4806	call (which might be in libev or elsewhere, for example, perl and many
		4807	other interpreters do their own select emulation on windows).
		4808
		4809	Another limit is the number of file descriptors in the Microsoft runtime
		4810	libraries, which by default is C<64> (there must be a hidden I<64>
		4811	fetish or something like this inside Microsoft). You can increase this
		4812	by calling C<_setmaxstdio>, which can increase this limit to C<2048>
		4813	(another arbitrary limit), but is broken in many versions of the Microsoft
		4814	runtime libraries. This might get you to about C<512> or C<2048> sockets
		4815	(depending on windows version and/or the phase of the moon). To get more,
		4816	you need to wrap all I/O functions and provide your own fd management, but
		4817	the cost of calling select (O(n²)) will likely make this unworkable.
		4818
		4819	=head2 PORTABILITY REQUIREMENTS
		4820
		4821	In addition to a working ISO-C implementation and of course the
		4822	backend-specific APIs, libev relies on a few additional extensions:
		4823
		4824	=over 4
		4825
		4826	=item C<void (*)(ev_watcher_type *, int revents)> must have compatible
		4827	calling conventions regardless of C<ev_watcher_type *>.
		4828
		4829	Libev assumes not only that all watcher pointers have the same internal
		4830	structure (guaranteed by POSIX but not by ISO C for example), but it also
		4831	assumes that the same (machine) code can be used to call any watcher
		4832	callback: The watcher callbacks have different type signatures, but libev
		4833	calls them using an C<ev_watcher *> internally.
		4834
		4835	=item pointer accesses must be thread-atomic
		4836
		4837	Accessing a pointer value must be atomic, it must both be readable and
		4838	writable in one piece - this is the case on all current architectures.
		4839
		4840	=item C<sig_atomic_t volatile> must be thread-atomic as well
		4841
		4842	The type C<sig_atomic_t volatile> (or whatever is defined as
		4843	C<EV_ATOMIC_T>) must be atomic with respect to accesses from different
		4844	threads. This is not part of the specification for C<sig_atomic_t>, but is
		4845	believed to be sufficiently portable.
		4846
		4847	=item C<sigprocmask> must work in a threaded environment
		4848
		4849	Libev uses C<sigprocmask> to temporarily block signals. This is not
		4850	allowed in a threaded program (C<pthread_sigmask> has to be used). Typical
		4851	pthread implementations will either allow C<sigprocmask> in the "main
		4852	thread" or will block signals process-wide, both behaviours would
		4853	be compatible with libev. Interaction between C<sigprocmask> and
		4854	C<pthread_sigmask> could complicate things, however.
		4855
		4856	The most portable way to handle signals is to block signals in all threads
		4857	except the initial one, and run the default loop in the initial thread as
		4858	well.
		4859
		4860	=item C<long> must be large enough for common memory allocation sizes
		4861
		4862	To improve portability and simplify its API, libev uses C<long> internally
		4863	instead of C<size_t> when allocating its data structures. On non-POSIX
		4864	systems (Microsoft...) this might be unexpectedly low, but is still at
		4865	least 31 bits everywhere, which is enough for hundreds of millions of
		4866	watchers.
		4867
		4868	=item C<double> must hold a time value in seconds with enough accuracy
		4869
		4870	The type C<double> is used to represent timestamps. It is required to
		4871	have at least 51 bits of mantissa (and 9 bits of exponent), which is
		4872	good enough for at least into the year 4000 with millisecond accuracy
		4873	(the design goal for libev). This requirement is overfulfilled by
		4874	implementations using IEEE 754, which is basically all existing ones. With
		4875	IEEE 754 doubles, you get microsecond accuracy until at least 2200.
		4876
		4877	=back
		4878
		4879	If you know of other additional requirements drop me a note.
		4880
		4881
		4882	=head1 ALGORITHMIC COMPLEXITIES
		4883
		4884	In this section the complexities of (many of) the algorithms used inside
		4885	libev will be documented. For complexity discussions about backends see
		4886	the documentation for C<ev_default_init>.
		4887
		4888	All of the following are about amortised time: If an array needs to be
		4889	extended, libev needs to realloc and move the whole array, but this
		4890	happens asymptotically rarer with higher number of elements, so O(1) might
		4891	mean that libev does a lengthy realloc operation in rare cases, but on
		4892	average it is much faster and asymptotically approaches constant time.
		4893
		4894	=over 4
		4895
		4896	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)
		4897
		4898	This means that, when you have a watcher that triggers in one hour and
		4899	there are 100 watchers that would trigger before that, then inserting will
		4900	have to skip roughly seven (C<ld 100>) of these watchers.
		4901
		4902	=item Changing timer/periodic watchers (by autorepeat or calling again): O(log skipped_other_timers)
		4903
		4904	That means that changing a timer costs less than removing/adding them,
		4905	as only the relative motion in the event queue has to be paid for.
		4906
		4907	=item Starting io/check/prepare/idle/signal/child/fork/async watchers: O(1)
		4908
		4909	These just add the watcher into an array or at the head of a list.
		4910
		4911	=item Stopping check/prepare/idle/fork/async watchers: O(1)
		4912
		4913	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))
		4914
		4915	These watchers are stored in lists, so they need to be walked to find the
		4916	correct watcher to remove. The lists are usually short (you don't usually
		4917	have many watchers waiting for the same fd or signal: one is typical, two
		4918	is rare).
		4919
		4920	=item Finding the next timer in each loop iteration: O(1)
		4921
		4922	By virtue of using a binary or 4-heap, the next timer is always found at a
		4923	fixed position in the storage array.
		4924
		4925	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)
		4926
		4927	A change means an I/O watcher gets started or stopped, which requires
		4928	libev to recalculate its status (and possibly tell the kernel, depending
		4929	on backend and whether C<ev_io_set> was used).
		4930
		4931	=item Activating one watcher (putting it into the pending state): O(1)
		4932
		4933	=item Priority handling: O(number_of_priorities)
		4934
		4935	Priorities are implemented by allocating some space for each
		4936	priority. When doing priority-based operations, libev usually has to
		4937	linearly search all the priorities, but starting/stopping and activating
		4938	watchers becomes O(1) with respect to priority handling.
		4939
		4940	=item Sending an ev_async: O(1)
		4941
		4942	=item Processing ev_async_send: O(number_of_async_watchers)
		4943
		4944	=item Processing signals: O(max_signal_number)
		4945
		4946	Sending involves a system call I<iff> there were no other C<ev_async_send>
		4947	calls in the current loop iteration. Checking for async and signal events
		4948	involves iterating over all running async watchers or all signal numbers.
		4949
		4950	=back
		4951
		4952
		4953	=head1 PORTING FROM LIBEV 3.X TO 4.X
		4954
		4955	The major version 4 introduced some incompatible changes to the API.
		4956
		4957	At the moment, the C<ev.h> header file provides compatibility definitions
		4958	for all changes, so most programs should still compile. The compatibility
		4959	layer might be removed in later versions of libev, so better update to the
		4960	new API early than late.
		4961
		4962	=over 4
		4963
		4964	=item C<EV_COMPAT3> backwards compatibility mechanism
		4965
		4966	The backward compatibility mechanism can be controlled by
		4967	C<EV_COMPAT3>. See L<PREPROCESSOR SYMBOLS/MACROS> in the L<EMBEDDING>
		4968	section.
		4969
		4970	=item C<ev_default_destroy> and C<ev_default_fork> have been removed
		4971
		4972	These calls can be replaced easily by their C<ev_loop_xxx> counterparts:
		4973
		4974	ev_loop_destroy (EV_DEFAULT_UC);
		4975	ev_loop_fork (EV_DEFAULT);
		4976
		4977	=item function/symbol renames
		4978
		4979	A number of functions and symbols have been renamed:
		4980
		4981	ev_loop => ev_run
		4982	EVLOOP_NONBLOCK => EVRUN_NOWAIT
		4983	EVLOOP_ONESHOT => EVRUN_ONCE
		4984
		4985	ev_unloop => ev_break
		4986	EVUNLOOP_CANCEL => EVBREAK_CANCEL
		4987	EVUNLOOP_ONE => EVBREAK_ONE
		4988	EVUNLOOP_ALL => EVBREAK_ALL
		4989
		4990	EV_TIMEOUT => EV_TIMER
		4991
		4992	ev_loop_count => ev_iteration
		4993	ev_loop_depth => ev_depth
		4994	ev_loop_verify => ev_verify
		4995
		4996	Most functions working on C<struct ev_loop> objects don't have an
		4997	C<ev_loop_> prefix, so it was removed; C<ev_loop>, C<ev_unloop> and
		4998	associated constants have been renamed to not collide with the C<struct
		4999	ev_loop> anymore and C<EV_TIMER> now follows the same naming scheme
		5000	as all other watcher types. Note that C<ev_loop_fork> is still called
		5001	C<ev_loop_fork> because it would otherwise clash with the C<ev_fork>
		5002	typedef.
		5003
		5004	=item C<EV_MINIMAL> mechanism replaced by C<EV_FEATURES>
		5005
		5006	The preprocessor symbol C<EV_MINIMAL> has been replaced by a different
		5007	mechanism, C<EV_FEATURES>. Programs using C<EV_MINIMAL> usually compile
		5008	and work, but the library code will of course be larger.
		5009
		5010	=back
		5011
		5012
		5013	=head1 GLOSSARY
		5014
		5015	=over 4
		5016
		5017	=item active
		5018
		5019	A watcher is active as long as it has been started and not yet stopped.
		5020	See L<WATCHER STATES> for details.
		5021
		5022	=item application
		5023
		5024	In this document, an application is whatever is using libev.
		5025
		5026	=item backend
		5027
		5028	The part of the code dealing with the operating system interfaces.
		5029
		5030	=item callback
		5031
		5032	The address of a function that is called when some event has been
		5033	detected. Callbacks are being passed the event loop, the watcher that
		5034	received the event, and the actual event bitset.
		5035
		5036	=item callback/watcher invocation
		5037
		5038	The act of calling the callback associated with a watcher.
		5039
		5040	=item event
		5041
		5042	A change of state of some external event, such as data now being available
		5043	for reading on a file descriptor, time having passed or simply not having
		5044	any other events happening anymore.
		5045
		5046	In libev, events are represented as single bits (such as C<EV_READ> or
		5047	C<EV_TIMER>).
		5048
		5049	=item event library
		5050
		5051	A software package implementing an event model and loop.
		5052
		5053	=item event loop
		5054
		5055	An entity that handles and processes external events and converts them
		5056	into callback invocations.
		5057
		5058	=item event model
		5059
		5060	The model used to describe how an event loop handles and processes
		5061	watchers and events.
		5062
		5063	=item pending
		5064
		5065	A watcher is pending as soon as the corresponding event has been
		5066	detected. See L<WATCHER STATES> for details.
		5067
		5068	=item real time
		5069
		5070	The physical time that is observed. It is apparently strictly monotonic :)
		5071
		5072	=item wall-clock time
		5073
		5074	The time and date as shown on clocks. Unlike real time, it can actually
		5075	be wrong and jump forwards and backwards, e.g. when the you adjust your
		5076	clock.
		5077
		5078	=item watcher
		5079
		5080	A data structure that describes interest in certain events. Watchers need
		5081	to be started (attached to an event loop) before they can receive events.
		5082
		5083	=back
1217		5084
1218	=head1 AUTHOR	5085	=head1 AUTHOR
1219		5086
1220	Marc Lehmann <libev@schmorp.de>.	5087	Marc Lehmann <libev@schmorp.de>, with repeated corrections by Mikael
		5088	Magnusson and Emanuele Giaquinta.
1221		5089

Diff Legend

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