ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/libev/ev.pod

Comparing libev/ev.pod (file contents):
Revision 1.11 by root, Mon Nov 12 08:32:52 2007 UTC vs.
Revision 1.253 by root, Tue Jul 14 18:33:48 2009 UTC

…		…
2		2
3	libev - a high performance full-featured event loop written in C	3	libev - a high performance full-featured event loop written in C
4		4
5	=head1 SYNOPSIS	5	=head1 SYNOPSIS
6		6
7	#include <ev.h>	7	#include <ev.h>
8		8
9	=head1 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_loop's to stop iterating
		32	ev_unloop (EV_A_ EVUNLOOP_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_loop to stop iterating
		41	ev_unloop (EV_A_ EVUNLOOP_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_loop (0);
		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_loop (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	Familarity with event based programming techniques in general is assumed
		81	throughout this document.
		82
		83	=head1 ABOUT LIBEV
10		84
11	Libev is an event loop: you register interest in certain events (such as a	85	Libev is an event loop: you register interest in certain events (such as a
12	file descriptor being readable or a timeout occuring), and it will manage	86	file descriptor being readable or a timeout occurring), and it will manage
13	these event sources and provide your program with events.	87	these event sources and provide your program with events.
14		88
15	To do this, it must take more or less complete control over your process	89	To do this, it must take more or less complete control over your process
16	(or thread) by executing the I<event loop> handler, and will then	90	(or thread) by executing the I<event loop> handler, and will then
17	communicate events via a callback mechanism.	91	communicate events via a callback mechanism.
…		…
19	You register interest in certain events by registering so-called I<event	93	You register interest in certain events by registering so-called I<event
20	watchers>, which are relatively small C structures you initialise with the	94	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	95	details of the event, and then hand it over to libev by I<starting> the
22	watcher.	96	watcher.
23		97
24	=head1 FEATURES	98	=head2 FEATURES
25		99
26	Libev supports select, poll, the linux-specific epoll and the bsd-specific	100	Libev supports C<select>, C<poll>, the Linux-specific C<epoll>, the
27	kqueue mechanisms for file descriptor events, relative timers, absolute	101	BSD-specific C<kqueue> and the Solaris-specific event port mechanisms
28	timers with customised rescheduling, signal events, process status change	102	for file descriptor events (C<ev_io>), the Linux C<inotify> interface
29	events (related to SIGCHLD), and event watchers dealing with the event	103	(for C<ev_stat>), relative timers (C<ev_timer>), absolute timers
30	loop mechanism itself (idle, prepare and check watchers). It also is quite	104	with customised rescheduling (C<ev_periodic>), synchronous signals
		105	(C<ev_signal>), process status change events (C<ev_child>), and event
		106	watchers dealing with the event loop mechanism itself (C<ev_idle>,
		107	C<ev_embed>, C<ev_prepare> and C<ev_check> watchers) as well as
		108	file watchers (C<ev_stat>) and even limited support for fork events
		109	(C<ev_fork>).
		110
		111	It also is quite fast (see this
31	fast (see this L<benchmark|http://libev.schmorp.de/bench.html> comparing	112	L<benchmark|http://libev.schmorp.de/bench.html> comparing it to libevent
32	it to libevent for example).	113	for example).
33		114
34	=head1 CONVENTIONS	115	=head2 CONVENTIONS
35		116
36	Libev is very configurable. In this manual the default configuration	117	Libev is very configurable. In this manual the default (and most common)
37	will be described, which supports multiple event loops. For more info	118	configuration will be described, which supports multiple event loops. For
38	about various configuration options please have a look at the file	119	more info about various configuration options please have a look at
39	F<README.embed> in the libev distribution. If libev was configured without	120	B<EMBED> section in this manual. If libev was configured without support
40	support for multiple event loops, then all functions taking an initial	121	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 *>)	122	name C<loop> (which is always of type C<ev_loop *>) will not have
42	will not have this argument.	123	this argument.
43		124
44	=head1 TIME AND OTHER GLOBAL FUNCTIONS	125	=head2 TIME REPRESENTATION
45		126
46	Libev represents time as a single floating point number, representing the	127	Libev represents time as a single floating point number, representing
47	(fractional) number of seconds since the (POSIX) epoch (somewhere near	128	the (fractional) number of seconds since the (POSIX) epoch (somewhere
48	the beginning of 1970, details are complicated, don't ask). This type is	129	near the beginning of 1970, details are complicated, don't ask). This
49	called C<ev_tstamp>, which is what you should use too. It usually aliases	130	type is called C<ev_tstamp>, which is what you should use too. It usually
50	to the double type in C.	131	aliases to the C<double> type in C. When you need to do any calculations
		132	on it, you should treat it as some floating point value. Unlike the name
		133	component C<stamp> might indicate, it is also used for time differences
		134	throughout libev.
		135
		136	=head1 ERROR HANDLING
		137
		138	Libev knows three classes of errors: operating system errors, usage errors
		139	and internal errors (bugs).
		140
		141	When libev catches an operating system error it cannot handle (for example
		142	a system call indicating a condition libev cannot fix), it calls the callback
		143	set via C<ev_set_syserr_cb>, which is supposed to fix the problem or
		144	abort. The default is to print a diagnostic message and to call C<abort
		145	()>.
		146
		147	When libev detects a usage error such as a negative timer interval, then
		148	it will print a diagnostic message and abort (via the C<assert> mechanism,
		149	so C<NDEBUG> will disable this checking): these are programming errors in
		150	the libev caller and need to be fixed there.
		151
		152	Libev also has a few internal error-checking C<assert>ions, and also has
		153	extensive consistency checking code. These do not trigger under normal
		154	circumstances, as they indicate either a bug in libev or worse.
		155
		156
		157	=head1 GLOBAL FUNCTIONS
		158
		159	These functions can be called anytime, even before initialising the
		160	library in any way.
51		161
52	=over 4	162	=over 4
53		163
54	=item ev_tstamp ev_time ()	164	=item ev_tstamp ev_time ()
55		165
56	Returns the current time as libev would use it.	166	Returns the current time as libev would use it. Please note that the
		167	C<ev_now> function is usually faster and also often returns the timestamp
		168	you actually want to know.
		169
		170	=item ev_sleep (ev_tstamp interval)
		171
		172	Sleep for the given interval: The current thread will be blocked until
		173	either it is interrupted or the given time interval has passed. Basically
		174	this is a sub-second-resolution C<sleep ()>.
57		175
58	=item int ev_version_major ()	176	=item int ev_version_major ()
59		177
60	=item int ev_version_minor ()	178	=item int ev_version_minor ()
61		179
62	You can find out the major and minor version numbers of the library	180	You can find out the major and minor ABI version numbers of the library
63	you linked against by calling the functions C<ev_version_major> and	181	you linked against by calling the functions C<ev_version_major> and
64	C<ev_version_minor>. If you want, you can compare against the global	182	C<ev_version_minor>. If you want, you can compare against the global
65	symbols C<EV_VERSION_MAJOR> and C<EV_VERSION_MINOR>, which specify the	183	symbols C<EV_VERSION_MAJOR> and C<EV_VERSION_MINOR>, which specify the
66	version of the library your program was compiled against.	184	version of the library your program was compiled against.
67		185
		186	These version numbers refer to the ABI version of the library, not the
		187	release version.
		188
68	Usually, it's a good idea to terminate if the major versions mismatch,	189	Usually, it's a good idea to terminate if the major versions mismatch,
69	as this indicates an incompatible change. Minor versions are usually	190	as this indicates an incompatible change. Minor versions are usually
70	compatible to older versions, so a larger minor version alone is usually	191	compatible to older versions, so a larger minor version alone is usually
71	not a problem.	192	not a problem.
72		193
		194	Example: Make sure we haven't accidentally been linked against the wrong
		195	version.
		196
		197	assert (("libev version mismatch",
		198	ev_version_major () == EV_VERSION_MAJOR
		199	&& ev_version_minor () >= EV_VERSION_MINOR));
		200
		201	=item unsigned int ev_supported_backends ()
		202
		203	Return the set of all backends (i.e. their corresponding C<EV_BACKEND_*>
		204	value) compiled into this binary of libev (independent of their
		205	availability on the system you are running on). See C<ev_default_loop> for
		206	a description of the set values.
		207
		208	Example: make sure we have the epoll method, because yeah this is cool and
		209	a must have and can we have a torrent of it please!!!11
		210
		211	assert (("sorry, no epoll, no sex",
		212	ev_supported_backends () & EVBACKEND_EPOLL));
		213
		214	=item unsigned int ev_recommended_backends ()
		215
		216	Return the set of all backends compiled into this binary of libev and also
		217	recommended for this platform. This set is often smaller than the one
		218	returned by C<ev_supported_backends>, as for example kqueue is broken on
		219	most BSDs and will not be auto-detected unless you explicitly request it
		220	(assuming you know what you are doing). This is the set of backends that
		221	libev will probe for if you specify no backends explicitly.
		222
		223	=item unsigned int ev_embeddable_backends ()
		224
		225	Returns the set of backends that are embeddable in other event loops. This
		226	is the theoretical, all-platform, value. To find which backends
		227	might be supported on the current system, you would need to look at
		228	C<ev_embeddable_backends () & ev_supported_backends ()>, likewise for
		229	recommended ones.
		230
		231	See the description of C<ev_embed> watchers for more info.
		232
73	=item ev_set_allocator (void *(*cb)(void *ptr, long size))	233	=item ev_set_allocator (void *(*cb)(void *ptr, long size)) [NOT REENTRANT]
74		234
75	Sets the allocation function to use (the prototype is similar to the	235	Sets the allocation function to use (the prototype is similar - the
76	realloc C function, the semantics are identical). It is used to allocate	236	semantics are identical to the C<realloc> C89/SuS/POSIX function). It is
77	and free memory (no surprises here). If it returns zero when memory	237	used to allocate and free memory (no surprises here). If it returns zero
78	needs to be allocated, the library might abort or take some potentially	238	when memory needs to be allocated (C<size != 0>), the library might abort
79	destructive action. The default is your system realloc function.	239	or take some potentially destructive action.
		240
		241	Since some systems (at least OpenBSD and Darwin) fail to implement
		242	correct C<realloc> semantics, libev will use a wrapper around the system
		243	C<realloc> and C<free> functions by default.
80		244
81	You could override this function in high-availability programs to, say,	245	You could override this function in high-availability programs to, say,
82	free some memory if it cannot allocate memory, to use a special allocator,	246	free some memory if it cannot allocate memory, to use a special allocator,
83	or even to sleep a while and retry until some memory is available.	247	or even to sleep a while and retry until some memory is available.
84		248
		249	Example: Replace the libev allocator with one that waits a bit and then
		250	retries (example requires a standards-compliant C<realloc>).
		251
		252	static void *
		253	persistent_realloc (void *ptr, size_t size)
		254	{
		255	for (;;)
		256	{
		257	void *newptr = realloc (ptr, size);
		258
		259	if (newptr)
		260	return newptr;
		261
		262	sleep (60);
		263	}
		264	}
		265
		266	...
		267	ev_set_allocator (persistent_realloc);
		268
85	=item ev_set_syserr_cb (void (*cb)(const char *msg));	269	=item ev_set_syserr_cb (void (*cb)(const char *msg)); [NOT REENTRANT]
86		270
87	Set the callback function to call on a retryable syscall error (such	271	Set the callback function to call on a retryable system call error (such
88	as failed select, poll, epoll_wait). The message is a printable string	272	as failed select, poll, epoll_wait). The message is a printable string
89	indicating the system call or subsystem causing the problem. If this	273	indicating the system call or subsystem causing the problem. If this
90	callback is set, then libev will expect it to remedy the sitution, no	274	callback is set, then libev will expect it to remedy the situation, no
91	matter what, when it returns. That is, libev will generally retry the	275	matter what, when it returns. That is, libev will generally retry the
92	requested operation, or, if the condition doesn't go away, do bad stuff	276	requested operation, or, if the condition doesn't go away, do bad stuff
93	(such as abort).	277	(such as abort).
94		278
		279	Example: This is basically the same thing that libev does internally, too.
		280
		281	static void
		282	fatal_error (const char *msg)
		283	{
		284	perror (msg);
		285	abort ();
		286	}
		287
		288	...
		289	ev_set_syserr_cb (fatal_error);
		290
95	=back	291	=back
96		292
97	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP	293	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP
98		294
99	An event loop is described by a C<struct ev_loop *>. The library knows two	295	An event loop is described by a C<struct ev_loop *> (the C<struct>
100	types of such loops, the I<default> loop, which supports signals and child	296	is I<not> optional in this case, as there is also an C<ev_loop>
101	events, and dynamically created loops which do not.	297	I<function>).
102		298
103	If you use threads, a common model is to run the default event loop	299	The library knows two types of such loops, the I<default> loop, which
104	in your main thread (or in a separate thrad) and for each thread you	300	supports signals and child events, and dynamically created loops which do
105	create, you also create another event loop. Libev itself does no locking	301	not.
106	whatsoever, so if you mix calls to the same event loop in different
107	threads, make sure you lock (this is usually a bad idea, though, even if
108	done correctly, because it's hideous and inefficient).
109		302
110	=over 4	303	=over 4
111		304
112	=item struct ev_loop *ev_default_loop (unsigned int flags)	305	=item struct ev_loop *ev_default_loop (unsigned int flags)
113		306
114	This will initialise the default event loop if it hasn't been initialised	307	This will initialise the default event loop if it hasn't been initialised
115	yet and return it. If the default loop could not be initialised, returns	308	yet and return it. If the default loop could not be initialised, returns
116	false. If it already was initialised it simply returns it (and ignores the	309	false. If it already was initialised it simply returns it (and ignores the
117	flags).	310	flags. If that is troubling you, check C<ev_backend ()> afterwards).
118		311
119	If you don't know what event loop to use, use the one returned from this	312	If you don't know what event loop to use, use the one returned from this
120	function.	313	function.
121		314
		315	Note that this function is I<not> thread-safe, so if you want to use it
		316	from multiple threads, you have to lock (note also that this is unlikely,
		317	as loops cannot be shared easily between threads anyway).
		318
		319	The default loop is the only loop that can handle C<ev_signal> and
		320	C<ev_child> watchers, and to do this, it always registers a handler
		321	for C<SIGCHLD>. If this is a problem for your application you can either
		322	create a dynamic loop with C<ev_loop_new> that doesn't do that, or you
		323	can simply overwrite the C<SIGCHLD> signal handler I<after> calling
		324	C<ev_default_init>.
		325
122	The flags argument can be used to specify special behaviour or specific	326	The flags argument can be used to specify special behaviour or specific
123	backends to use, and is usually specified as 0 (or EVFLAG_AUTO).	327	backends to use, and is usually specified as C<0> (or C<EVFLAG_AUTO>).
124		328
125	It supports the following flags:	329	The following flags are supported:
126		330
127	=over 4	331	=over 4
128		332
129	=item C<EVFLAG_AUTO>	333	=item C<EVFLAG_AUTO>
130		334
131	The default flags value. Use this if you have no clue (it's the right	335	The default flags value. Use this if you have no clue (it's the right
132	thing, believe me).	336	thing, believe me).
133		337
134	=item C<EVFLAG_NOENV>	338	=item C<EVFLAG_NOENV>
135		339
136	If this flag bit is ored into the flag value (or the program runs setuid	340	If this flag bit is or'ed into the flag value (or the program runs setuid
137	or setgid) then libev will I<not> look at the environment variable	341	or setgid) then libev will I<not> look at the environment variable
138	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will	342	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will
139	override the flags completely if it is found in the environment. This is	343	override the flags completely if it is found in the environment. This is
140	useful to try out specific backends to test their performance, or to work	344	useful to try out specific backends to test their performance, or to work
141	around bugs.	345	around bugs.
142		346
		347	=item C<EVFLAG_FORKCHECK>
		348
		349	Instead of calling C<ev_default_fork> or C<ev_loop_fork> manually after
		350	a fork, you can also make libev check for a fork in each iteration by
		351	enabling this flag.
		352
		353	This works by calling C<getpid ()> on every iteration of the loop,
		354	and thus this might slow down your event loop if you do a lot of loop
		355	iterations and little real work, but is usually not noticeable (on my
		356	GNU/Linux system for example, C<getpid> is actually a simple 5-insn sequence
		357	without a system call and thus I<very> fast, but my GNU/Linux system also has
		358	C<pthread_atfork> which is even faster).
		359
		360	The big advantage of this flag is that you can forget about fork (and
		361	forget about forgetting to tell libev about forking) when you use this
		362	flag.
		363
		364	This flag setting cannot be overridden or specified in the C<LIBEV_FLAGS>
		365	environment variable.
		366
143	=item C<EVMETHOD_SELECT> (portable select backend)	367	=item C<EVBACKEND_SELECT> (value 1, portable select backend)
144		368
		369	This is your standard select(2) backend. Not I<completely> standard, as
		370	libev tries to roll its own fd_set with no limits on the number of fds,
		371	but if that fails, expect a fairly low limit on the number of fds when
		372	using this backend. It doesn't scale too well (O(highest_fd)), but its
		373	usually the fastest backend for a low number of (low-numbered :) fds.
		374
		375	To get good performance out of this backend you need a high amount of
		376	parallelism (most of the file descriptors should be busy). If you are
		377	writing a server, you should C<accept ()> in a loop to accept as many
		378	connections as possible during one iteration. You might also want to have
		379	a look at C<ev_set_io_collect_interval ()> to increase the amount of
		380	readiness notifications you get per iteration.
		381
		382	This backend maps C<EV_READ> to the C<readfds> set and C<EV_WRITE> to the
		383	C<writefds> set (and to work around Microsoft Windows bugs, also onto the
		384	C<exceptfds> set on that platform).
		385
145	=item C<EVMETHOD_POLL> (poll backend, available everywhere except on windows)	386	=item C<EVBACKEND_POLL> (value 2, poll backend, available everywhere except on windows)
146		387
147	=item C<EVMETHOD_EPOLL> (linux only)	388	And this is your standard poll(2) backend. It's more complicated
		389	than select, but handles sparse fds better and has no artificial
		390	limit on the number of fds you can use (except it will slow down
		391	considerably with a lot of inactive fds). It scales similarly to select,
		392	i.e. O(total_fds). See the entry for C<EVBACKEND_SELECT>, above, for
		393	performance tips.
148		394
149	=item C<EVMETHOD_KQUEUE> (some bsds only)	395	This backend maps C<EV_READ> to C<POLLIN | POLLERR | POLLHUP>, and
		396	C<EV_WRITE> to C<POLLOUT | POLLERR | POLLHUP>.
150		397
151	=item C<EVMETHOD_DEVPOLL> (solaris 8 only)	398	=item C<EVBACKEND_EPOLL> (value 4, Linux)
152		399
153	=item C<EVMETHOD_PORT> (solaris 10 only)	400	For few fds, this backend is a bit little slower than poll and select,
		401	but it scales phenomenally better. While poll and select usually scale
		402	like O(total_fds) where n is the total number of fds (or the highest fd),
		403	epoll scales either O(1) or O(active_fds).
154		404
155	If one or more of these are ored into the flags value, then only these	405	The epoll mechanism deserves honorable mention as the most misdesigned
156	backends will be tried (in the reverse order as given here). If one are	406	of the more advanced event mechanisms: mere annoyances include silently
157	specified, any backend will do.	407	dropping file descriptors, requiring a system call per change per file
		408	descriptor (and unnecessary guessing of parameters), problems with dup and
		409	so on. The biggest issue is fork races, however - if a program forks then
		410	I<both> parent and child process have to recreate the epoll set, which can
		411	take considerable time (one syscall per file descriptor) and is of course
		412	hard to detect.
		413
		414	Epoll is also notoriously buggy - embedding epoll fds I<should> work, but
		415	of course I<doesn't>, and epoll just loves to report events for totally
		416	I<different> file descriptors (even already closed ones, so one cannot
		417	even remove them from the set) than registered in the set (especially
		418	on SMP systems). Libev tries to counter these spurious notifications by
		419	employing an additional generation counter and comparing that against the
		420	events to filter out spurious ones, recreating the set when required.
		421
		422	While stopping, setting and starting an I/O watcher in the same iteration
		423	will result in some caching, there is still a system call per such
		424	incident (because the same I<file descriptor> could point to a different
		425	I<file description> now), so its best to avoid that. Also, C<dup ()>'ed
		426	file descriptors might not work very well if you register events for both
		427	file descriptors.
		428
		429	Best performance from this backend is achieved by not unregistering all
		430	watchers for a file descriptor until it has been closed, if possible,
		431	i.e. keep at least one watcher active per fd at all times. Stopping and
		432	starting a watcher (without re-setting it) also usually doesn't cause
		433	extra overhead. A fork can both result in spurious notifications as well
		434	as in libev having to destroy and recreate the epoll object, which can
		435	take considerable time and thus should be avoided.
		436
		437	All this means that, in practice, C<EVBACKEND_SELECT> can be as fast or
		438	faster than epoll for maybe up to a hundred file descriptors, depending on
		439	the usage. So sad.
		440
		441	While nominally embeddable in other event loops, this feature is broken in
		442	all kernel versions tested so far.
		443
		444	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
		445	C<EVBACKEND_POLL>.
		446
		447	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)
		448
		449	Kqueue deserves special mention, as at the time of this writing, it
		450	was broken on all BSDs except NetBSD (usually it doesn't work reliably
		451	with anything but sockets and pipes, except on Darwin, where of course
		452	it's completely useless). Unlike epoll, however, whose brokenness
		453	is by design, these kqueue bugs can (and eventually will) be fixed
		454	without API changes to existing programs. For this reason it's not being
		455	"auto-detected" unless you explicitly specify it in the flags (i.e. using
		456	C<EVBACKEND_KQUEUE>) or libev was compiled on a known-to-be-good (-enough)
		457	system like NetBSD.
		458
		459	You still can embed kqueue into a normal poll or select backend and use it
		460	only for sockets (after having made sure that sockets work with kqueue on
		461	the target platform). See C<ev_embed> watchers for more info.
		462
		463	It scales in the same way as the epoll backend, but the interface to the
		464	kernel is more efficient (which says nothing about its actual speed, of
		465	course). While stopping, setting and starting an I/O watcher does never
		466	cause an extra system call as with C<EVBACKEND_EPOLL>, it still adds up to
		467	two event changes per incident. Support for C<fork ()> is very bad (but
		468	sane, unlike epoll) and it drops fds silently in similarly hard-to-detect
		469	cases
		470
		471	This backend usually performs well under most conditions.
		472
		473	While nominally embeddable in other event loops, this doesn't work
		474	everywhere, so you might need to test for this. And since it is broken
		475	almost everywhere, you should only use it when you have a lot of sockets
		476	(for which it usually works), by embedding it into another event loop
		477	(e.g. C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> (but C<poll> is of course
		478	also broken on OS X)) and, did I mention it, using it only for sockets.
		479
		480	This backend maps C<EV_READ> into an C<EVFILT_READ> kevent with
		481	C<NOTE_EOF>, and C<EV_WRITE> into an C<EVFILT_WRITE> kevent with
		482	C<NOTE_EOF>.
		483
		484	=item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8)
		485
		486	This is not implemented yet (and might never be, unless you send me an
		487	implementation). According to reports, C</dev/poll> only supports sockets
		488	and is not embeddable, which would limit the usefulness of this backend
		489	immensely.
		490
		491	=item C<EVBACKEND_PORT> (value 32, Solaris 10)
		492
		493	This uses the Solaris 10 event port mechanism. As with everything on Solaris,
		494	it's really slow, but it still scales very well (O(active_fds)).
		495
		496	Please note that Solaris event ports can deliver a lot of spurious
		497	notifications, so you need to use non-blocking I/O or other means to avoid
		498	blocking when no data (or space) is available.
		499
		500	While this backend scales well, it requires one system call per active
		501	file descriptor per loop iteration. For small and medium numbers of file
		502	descriptors a "slow" C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> backend
		503	might perform better.
		504
		505	On the positive side, with the exception of the spurious readiness
		506	notifications, this backend actually performed fully to specification
		507	in all tests and is fully embeddable, which is a rare feat among the
		508	OS-specific backends (I vastly prefer correctness over speed hacks).
		509
		510	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
		511	C<EVBACKEND_POLL>.
		512
		513	=item C<EVBACKEND_ALL>
		514
		515	Try all backends (even potentially broken ones that wouldn't be tried
		516	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as
		517	C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>.
		518
		519	It is definitely not recommended to use this flag.
158		520
159	=back	521	=back
		522
		523	If one or more of these are or'ed into the flags value, then only these
		524	backends will be tried (in the reverse order as listed here). If none are
		525	specified, all backends in C<ev_recommended_backends ()> will be tried.
		526
		527	Example: This is the most typical usage.
		528
		529	if (!ev_default_loop (0))
		530	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
		531
		532	Example: Restrict libev to the select and poll backends, and do not allow
		533	environment settings to be taken into account:
		534
		535	ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);
		536
		537	Example: Use whatever libev has to offer, but make sure that kqueue is
		538	used if available (warning, breaks stuff, best use only with your own
		539	private event loop and only if you know the OS supports your types of
		540	fds):
		541
		542	ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);
160		543
161	=item struct ev_loop *ev_loop_new (unsigned int flags)	544	=item struct ev_loop *ev_loop_new (unsigned int flags)
162		545
163	Similar to C<ev_default_loop>, but always creates a new event loop that is	546	Similar to C<ev_default_loop>, but always creates a new event loop that is
164	always distinct from the default loop. Unlike the default loop, it cannot	547	always distinct from the default loop. Unlike the default loop, it cannot
165	handle signal and child watchers, and attempts to do so will be greeted by	548	handle signal and child watchers, and attempts to do so will be greeted by
166	undefined behaviour (or a failed assertion if assertions are enabled).	549	undefined behaviour (or a failed assertion if assertions are enabled).
167		550
		551	Note that this function I<is> thread-safe, and the recommended way to use
		552	libev with threads is indeed to create one loop per thread, and using the
		553	default loop in the "main" or "initial" thread.
		554
		555	Example: Try to create a event loop that uses epoll and nothing else.
		556
		557	struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);
		558	if (!epoller)
		559	fatal ("no epoll found here, maybe it hides under your chair");
		560
168	=item ev_default_destroy ()	561	=item ev_default_destroy ()
169		562
170	Destroys the default loop again (frees all memory and kernel state	563	Destroys the default loop again (frees all memory and kernel state
171	etc.). This stops all registered event watchers (by not touching them in	564	etc.). None of the active event watchers will be stopped in the normal
172	any way whatsoever, although you cannot rely on this :).	565	sense, so e.g. C<ev_is_active> might still return true. It is your
		566	responsibility to either stop all watchers cleanly yourself I<before>
		567	calling this function, or cope with the fact afterwards (which is usually
		568	the easiest thing, you can just ignore the watchers and/or C<free ()> them
		569	for example).
		570
		571	Note that certain global state, such as signal state (and installed signal
		572	handlers), will not be freed by this function, and related watchers (such
		573	as signal and child watchers) would need to be stopped manually.
		574
		575	In general it is not advisable to call this function except in the
		576	rare occasion where you really need to free e.g. the signal handling
		577	pipe fds. If you need dynamically allocated loops it is better to use
		578	C<ev_loop_new> and C<ev_loop_destroy>).
173		579
174	=item ev_loop_destroy (loop)	580	=item ev_loop_destroy (loop)
175		581
176	Like C<ev_default_destroy>, but destroys an event loop created by an	582	Like C<ev_default_destroy>, but destroys an event loop created by an
177	earlier call to C<ev_loop_new>.	583	earlier call to C<ev_loop_new>.
178		584
179	=item ev_default_fork ()	585	=item ev_default_fork ()
180		586
		587	This function sets a flag that causes subsequent C<ev_loop> iterations
181	This function reinitialises the kernel state for backends that have	588	to reinitialise the kernel state for backends that have one. Despite the
182	one. Despite the name, you can call it anytime, but it makes most sense	589	name, you can call it anytime, but it makes most sense after forking, in
183	after forking, in either the parent or child process (or both, but that	590	the child process (or both child and parent, but that again makes little
184	again makes little sense).	591	sense). You I<must> call it in the child before using any of the libev
		592	functions, and it will only take effect at the next C<ev_loop> iteration.
185		593
186	You I<must> call this function after forking if and only if you want to	594	On the other hand, you only need to call this function in the child
187	use the event library in both processes. If you just fork+exec, you don't	595	process if and only if you want to use the event library in the child. If
188	have to call it.	596	you just fork+exec, you don't have to call it at all.
189		597
190	The function itself is quite fast and it's usually not a problem to call	598	The function itself is quite fast and it's usually not a problem to call
191	it just in case after a fork. To make this easy, the function will fit in	599	it just in case after a fork. To make this easy, the function will fit in
192	quite nicely into a call to C<pthread_atfork>:	600	quite nicely into a call to C<pthread_atfork>:
193		601
…		…
195		603
196	=item ev_loop_fork (loop)	604	=item ev_loop_fork (loop)
197		605
198	Like C<ev_default_fork>, but acts on an event loop created by	606	Like C<ev_default_fork>, but acts on an event loop created by
199	C<ev_loop_new>. Yes, you have to call this on every allocated event loop	607	C<ev_loop_new>. Yes, you have to call this on every allocated event loop
200	after fork, and how you do this is entirely your own problem.	608	after fork that you want to re-use in the child, and how you do this is
		609	entirely your own problem.
201		610
		611	=item int ev_is_default_loop (loop)
		612
		613	Returns true when the given loop is, in fact, the default loop, and false
		614	otherwise.
		615
		616	=item unsigned int ev_loop_count (loop)
		617
		618	Returns the count of loop iterations for the loop, which is identical to
		619	the number of times libev did poll for new events. It starts at C<0> and
		620	happily wraps around with enough iterations.
		621
		622	This value can sometimes be useful as a generation counter of sorts (it
		623	"ticks" the number of loop iterations), as it roughly corresponds with
		624	C<ev_prepare> and C<ev_check> calls.
		625
		626	=item unsigned int ev_loop_depth (loop)
		627
		628	Returns the number of times C<ev_loop> was entered minus the number of
		629	times C<ev_loop> was exited, in other words, the recursion depth.
		630
		631	Outside C<ev_loop>, this number is zero. In a callback, this number is
		632	C<1>, unless C<ev_loop> was invoked recursively (or from another thread),
		633	in which case it is higher.
		634
		635	Leaving C<ev_loop> abnormally (setjmp/longjmp, cancelling the thread
		636	etc.), doesn't count as exit.
		637
202	=item unsigned int ev_method (loop)	638	=item unsigned int ev_backend (loop)
203		639
204	Returns one of the C<EVMETHOD_*> flags indicating the event backend in	640	Returns one of the C<EVBACKEND_*> flags indicating the event backend in
205	use.	641	use.
206		642
207	=item ev_tstamp ev_now (loop)	643	=item ev_tstamp ev_now (loop)
208		644
209	Returns the current "event loop time", which is the time the event loop	645	Returns the current "event loop time", which is the time the event loop
210	got events and started processing them. This timestamp does not change	646	received events and started processing them. This timestamp does not
211	as long as callbacks are being processed, and this is also the base time	647	change as long as callbacks are being processed, and this is also the base
212	used for relative timers. You can treat it as the timestamp of the event	648	time used for relative timers. You can treat it as the timestamp of the
213	occuring (or more correctly, the mainloop finding out about it).	649	event occurring (or more correctly, libev finding out about it).
		650
		651	=item ev_now_update (loop)
		652
		653	Establishes the current time by querying the kernel, updating the time
		654	returned by C<ev_now ()> in the progress. This is a costly operation and
		655	is usually done automatically within C<ev_loop ()>.
		656
		657	This function is rarely useful, but when some event callback runs for a
		658	very long time without entering the event loop, updating libev's idea of
		659	the current time is a good idea.
		660
		661	See also L<The special problem of time updates> in the C<ev_timer> section.
		662
		663	=item ev_suspend (loop)
		664
		665	=item ev_resume (loop)
		666
		667	These two functions suspend and resume a loop, for use when the loop is
		668	not used for a while and timeouts should not be processed.
		669
		670	A typical use case would be an interactive program such as a game: When
		671	the user presses C<^Z> to suspend the game and resumes it an hour later it
		672	would be best to handle timeouts as if no time had actually passed while
		673	the program was suspended. This can be achieved by calling C<ev_suspend>
		674	in your C<SIGTSTP> handler, sending yourself a C<SIGSTOP> and calling
		675	C<ev_resume> directly afterwards to resume timer processing.
		676
		677	Effectively, all C<ev_timer> watchers will be delayed by the time spend
		678	between C<ev_suspend> and C<ev_resume>, and all C<ev_periodic> watchers
		679	will be rescheduled (that is, they will lose any events that would have
		680	occured while suspended).
		681
		682	After calling C<ev_suspend> you B<must not> call I<any> function on the
		683	given loop other than C<ev_resume>, and you B<must not> call C<ev_resume>
		684	without a previous call to C<ev_suspend>.
		685
		686	Calling C<ev_suspend>/C<ev_resume> has the side effect of updating the
		687	event loop time (see C<ev_now_update>).
214		688
215	=item ev_loop (loop, int flags)	689	=item ev_loop (loop, int flags)
216		690
217	Finally, this is it, the event handler. This function usually is called	691	Finally, this is it, the event handler. This function usually is called
218	after you initialised all your watchers and you want to start handling	692	after you initialised all your watchers and you want to start handling
219	events.	693	events.
220		694
221	If the flags argument is specified as 0, it will not return until either	695	If the flags argument is specified as C<0>, it will not return until
222	no event watchers are active anymore or C<ev_unloop> was called.	696	either no event watchers are active anymore or C<ev_unloop> was called.
		697
		698	Please note that an explicit C<ev_unloop> is usually better than
		699	relying on all watchers to be stopped when deciding when a program has
		700	finished (especially in interactive programs), but having a program
		701	that automatically loops as long as it has to and no longer by virtue
		702	of relying on its watchers stopping correctly, that is truly a thing of
		703	beauty.
223		704
224	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle	705	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle
225	those events and any outstanding ones, but will not block your process in	706	those events and any already outstanding ones, but will not block your
226	case there are no events and will return after one iteration of the loop.	707	process in case there are no events and will return after one iteration of
		708	the loop.
227		709
228	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if	710	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if
229	neccessary) and will handle those and any outstanding ones. It will block	711	necessary) and will handle those and any already outstanding ones. It
230	your process until at least one new event arrives, and will return after	712	will block your process until at least one new event arrives (which could
		713	be an event internal to libev itself, so there is no guarantee that a
		714	user-registered callback will be called), and will return after one
231	one iteration of the loop.	715	iteration of the loop.
232		716
233	This flags value could be used to implement alternative looping	717	This is useful if you are waiting for some external event in conjunction
234	constructs, but the C<prepare> and C<check> watchers provide a better and	718	with something not expressible using other libev watchers (i.e. "roll your
235	more generic mechanism.	719	own C<ev_loop>"). However, a pair of C<ev_prepare>/C<ev_check> watchers is
		720	usually a better approach for this kind of thing.
		721
		722	Here are the gory details of what C<ev_loop> does:
		723
		724	- Before the first iteration, call any pending watchers.
		725	* If EVFLAG_FORKCHECK was used, check for a fork.
		726	- If a fork was detected (by any means), queue and call all fork watchers.
		727	- Queue and call all prepare watchers.
		728	- If we have been forked, detach and recreate the kernel state
		729	as to not disturb the other process.
		730	- Update the kernel state with all outstanding changes.
		731	- Update the "event loop time" (ev_now ()).
		732	- Calculate for how long to sleep or block, if at all
		733	(active idle watchers, EVLOOP_NONBLOCK or not having
		734	any active watchers at all will result in not sleeping).
		735	- Sleep if the I/O and timer collect interval say so.
		736	- Block the process, waiting for any events.
		737	- Queue all outstanding I/O (fd) events.
		738	- Update the "event loop time" (ev_now ()), and do time jump adjustments.
		739	- Queue all expired timers.
		740	- Queue all expired periodics.
		741	- Unless any events are pending now, queue all idle watchers.
		742	- Queue all check watchers.
		743	- Call all queued watchers in reverse order (i.e. check watchers first).
		744	Signals and child watchers are implemented as I/O watchers, and will
		745	be handled here by queueing them when their watcher gets executed.
		746	- If ev_unloop has been called, or EVLOOP_ONESHOT or EVLOOP_NONBLOCK
		747	were used, or there are no active watchers, return, otherwise
		748	continue with step *.
		749
		750	Example: Queue some jobs and then loop until no events are outstanding
		751	anymore.
		752
		753	... queue jobs here, make sure they register event watchers as long
		754	... as they still have work to do (even an idle watcher will do..)
		755	ev_loop (my_loop, 0);
		756	... jobs done or somebody called unloop. yeah!
236		757
237	=item ev_unloop (loop, how)	758	=item ev_unloop (loop, how)
238		759
239	Can be used to make a call to C<ev_loop> return early (but only after it	760	Can be used to make a call to C<ev_loop> return early (but only after it
240	has processed all outstanding events). The C<how> argument must be either	761	has processed all outstanding events). The C<how> argument must be either
241	C<EVUNLOOP_ONCE>, which will make the innermost C<ev_loop> call return, or	762	C<EVUNLOOP_ONE>, which will make the innermost C<ev_loop> call return, or
242	C<EVUNLOOP_ALL>, which will make all nested C<ev_loop> calls return.	763	C<EVUNLOOP_ALL>, which will make all nested C<ev_loop> calls return.
		764
		765	This "unloop state" will be cleared when entering C<ev_loop> again.
		766
		767	It is safe to call C<ev_unloop> from otuside any C<ev_loop> calls.
243		768
244	=item ev_ref (loop)	769	=item ev_ref (loop)
245		770
246	=item ev_unref (loop)	771	=item ev_unref (loop)
247		772
248	Ref/unref can be used to add or remove a reference count on the event	773	Ref/unref can be used to add or remove a reference count on the event
249	loop: Every watcher keeps one reference, and as long as the reference	774	loop: Every watcher keeps one reference, and as long as the reference
250	count is nonzero, C<ev_loop> will not return on its own. If you have	775	count is nonzero, C<ev_loop> will not return on its own.
		776
251	a watcher you never unregister that should not keep C<ev_loop> from	777	If you have a watcher you never unregister that should not keep C<ev_loop>
252	returning, ev_unref() after starting, and ev_ref() before stopping it. For	778	from returning, call ev_unref() after starting, and ev_ref() before
		779	stopping it.
		780
253	example, libev itself uses this for its internal signal pipe: It is not	781	As an example, libev itself uses this for its internal signal pipe: It
254	visible to the libev user and should not keep C<ev_loop> from exiting if	782	is not visible to the libev user and should not keep C<ev_loop> from
255	no event watchers registered by it are active. It is also an excellent	783	exiting if no event watchers registered by it are active. It is also an
256	way to do this for generic recurring timers or from within third-party	784	excellent way to do this for generic recurring timers or from within
257	libraries. Just remember to I<unref after start> and I<ref before stop>.	785	third-party libraries. Just remember to I<unref after start> and I<ref
		786	before stop> (but only if the watcher wasn't active before, or was active
		787	before, respectively. Note also that libev might stop watchers itself
		788	(e.g. non-repeating timers) in which case you have to C<ev_ref>
		789	in the callback).
		790
		791	Example: Create a signal watcher, but keep it from keeping C<ev_loop>
		792	running when nothing else is active.
		793
		794	ev_signal exitsig;
		795	ev_signal_init (&exitsig, sig_cb, SIGINT);
		796	ev_signal_start (loop, &exitsig);
		797	evf_unref (loop);
		798
		799	Example: For some weird reason, unregister the above signal handler again.
		800
		801	ev_ref (loop);
		802	ev_signal_stop (loop, &exitsig);
		803
		804	=item ev_set_io_collect_interval (loop, ev_tstamp interval)
		805
		806	=item ev_set_timeout_collect_interval (loop, ev_tstamp interval)
		807
		808	These advanced functions influence the time that libev will spend waiting
		809	for events. Both time intervals are by default C<0>, meaning that libev
		810	will try to invoke timer/periodic callbacks and I/O callbacks with minimum
		811	latency.
		812
		813	Setting these to a higher value (the C<interval> I<must> be >= C<0>)
		814	allows libev to delay invocation of I/O and timer/periodic callbacks
		815	to increase efficiency of loop iterations (or to increase power-saving
		816	opportunities).
		817
		818	The idea is that sometimes your program runs just fast enough to handle
		819	one (or very few) event(s) per loop iteration. While this makes the
		820	program responsive, it also wastes a lot of CPU time to poll for new
		821	events, especially with backends like C<select ()> which have a high
		822	overhead for the actual polling but can deliver many events at once.
		823
		824	By setting a higher I<io collect interval> you allow libev to spend more
		825	time collecting I/O events, so you can handle more events per iteration,
		826	at the cost of increasing latency. Timeouts (both C<ev_periodic> and
		827	C<ev_timer>) will be not affected. Setting this to a non-null value will
		828	introduce an additional C<ev_sleep ()> call into most loop iterations. The
		829	sleep time ensures that libev will not poll for I/O events more often then
		830	once per this interval, on average.
		831
		832	Likewise, by setting a higher I<timeout collect interval> you allow libev
		833	to spend more time collecting timeouts, at the expense of increased
		834	latency/jitter/inexactness (the watcher callback will be called
		835	later). C<ev_io> watchers will not be affected. Setting this to a non-null
		836	value will not introduce any overhead in libev.
		837
		838	Many (busy) programs can usually benefit by setting the I/O collect
		839	interval to a value near C<0.1> or so, which is often enough for
		840	interactive servers (of course not for games), likewise for timeouts. It
		841	usually doesn't make much sense to set it to a lower value than C<0.01>,
		842	as this approaches the timing granularity of most systems. Note that if
		843	you do transactions with the outside world and you can't increase the
		844	parallelity, then this setting will limit your transaction rate (if you
		845	need to poll once per transaction and the I/O collect interval is 0.01,
		846	then you can't do more than 100 transations per second).
		847
		848	Setting the I<timeout collect interval> can improve the opportunity for
		849	saving power, as the program will "bundle" timer callback invocations that
		850	are "near" in time together, by delaying some, thus reducing the number of
		851	times the process sleeps and wakes up again. Another useful technique to
		852	reduce iterations/wake-ups is to use C<ev_periodic> watchers and make sure
		853	they fire on, say, one-second boundaries only.
		854
		855	Example: we only need 0.1s timeout granularity, and we wish not to poll
		856	more often than 100 times per second:
		857
		858	ev_set_timeout_collect_interval (EV_DEFAULT_UC_ 0.1);
		859	ev_set_io_collect_interval (EV_DEFAULT_UC_ 0.01);
		860
		861	=item ev_invoke_pending (loop)
		862
		863	This call will simply invoke all pending watchers while resetting their
		864	pending state. Normally, C<ev_loop> does this automatically when required,
		865	but when overriding the invoke callback this call comes handy.
		866
		867	=item ev_set_invoke_pending_cb (loop, void (*invoke_pending_cb)(EV_P))
		868
		869	This overrides the invoke pending functionality of the loop: Instead of
		870	invoking all pending watchers when there are any, C<ev_loop> will call
		871	this callback instead. This is useful, for example, when you want to
		872	invoke the actual watchers inside another context (another thread etc.).
		873
		874	If you want to reset the callback, use C<ev_invoke_pending> as new
		875	callback.
		876
		877	=item ev_set_loop_release_cb (loop, void (*release)(EV_P), void (*acquire)(EV_P))
		878
		879	Sometimes you want to share the same loop between multiple threads. This
		880	can be done relatively simply by putting mutex_lock/unlock calls around
		881	each call to a libev function.
		882
		883	However, C<ev_loop> can run an indefinite time, so it is not feasible to
		884	wait for it to return. One way around this is to wake up the loop via
		885	C<ev_unloop> and C<av_async_send>, another way is to set these I<release>
		886	and I<acquire> callbacks on the loop.
		887
		888	When set, then C<release> will be called just before the thread is
		889	suspended waiting for new events, and C<acquire> is called just
		890	afterwards.
		891
		892	Ideally, C<release> will just call your mutex_unlock function, and
		893	C<acquire> will just call the mutex_lock function again.
		894
		895	=item ev_set_userdata (loop, void *data)
		896
		897	=item ev_userdata (loop)
		898
		899	Set and retrieve a single C<void *> associated with a loop. When
		900	C<ev_set_userdata> has never been called, then C<ev_userdata> returns
		901	C<0.>
		902
		903	These two functions can be used to associate arbitrary data with a loop,
		904	and are intended solely for the C<invoke_pending_cb>, C<release> and
		905	C<acquire> callbacks described above, but of course can be (ab-)used for
		906	any other purpose as well.
		907
		908	=item ev_loop_verify (loop)
		909
		910	This function only does something when C<EV_VERIFY> support has been
		911	compiled in, which is the default for non-minimal builds. It tries to go
		912	through all internal structures and checks them for validity. If anything
		913	is found to be inconsistent, it will print an error message to standard
		914	error and call C<abort ()>.
		915
		916	This can be used to catch bugs inside libev itself: under normal
		917	circumstances, this function will never abort as of course libev keeps its
		918	data structures consistent.
258		919
259	=back	920	=back
260		921
		922
261	=head1 ANATOMY OF A WATCHER	923	=head1 ANATOMY OF A WATCHER
		924
		925	In the following description, uppercase C<TYPE> in names stands for the
		926	watcher type, e.g. C<ev_TYPE_start> can mean C<ev_timer_start> for timer
		927	watchers and C<ev_io_start> for I/O watchers.
262		928
263	A watcher is a structure that you create and register to record your	929	A watcher is a structure that you create and register to record your
264	interest in some event. For instance, if you want to wait for STDIN to	930	interest in some event. For instance, if you want to wait for STDIN to
265	become readable, you would create an C<ev_io> watcher for that:	931	become readable, you would create an C<ev_io> watcher for that:
266		932
267	static void my_cb (struct ev_loop *loop, struct ev_io *w, int revents)	933	static void my_cb (struct ev_loop *loop, ev_io *w, int revents)
268	{	934	{
269	ev_io_stop (w);	935	ev_io_stop (w);
270	ev_unloop (loop, EVUNLOOP_ALL);	936	ev_unloop (loop, EVUNLOOP_ALL);
271	}	937	}
272		938
273	struct ev_loop *loop = ev_default_loop (0);	939	struct ev_loop *loop = ev_default_loop (0);
		940
274	struct ev_io stdin_watcher;	941	ev_io stdin_watcher;
		942
275	ev_init (&stdin_watcher, my_cb);	943	ev_init (&stdin_watcher, my_cb);
276	ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);	944	ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);
277	ev_io_start (loop, &stdin_watcher);	945	ev_io_start (loop, &stdin_watcher);
		946
278	ev_loop (loop, 0);	947	ev_loop (loop, 0);
279		948
280	As you can see, you are responsible for allocating the memory for your	949	As you can see, you are responsible for allocating the memory for your
281	watcher structures (and it is usually a bad idea to do this on the stack,	950	watcher structures (and it is I<usually> a bad idea to do this on the
282	although this can sometimes be quite valid).	951	stack).
		952
		953	Each watcher has an associated watcher structure (called C<struct ev_TYPE>
		954	or simply C<ev_TYPE>, as typedefs are provided for all watcher structs).
283		955
284	Each watcher structure must be initialised by a call to C<ev_init	956	Each watcher structure must be initialised by a call to C<ev_init
285	(watcher *, callback)>, which expects a callback to be provided. This	957	(watcher *, callback)>, which expects a callback to be provided. This
286	callback gets invoked each time the event occurs (or, in the case of io	958	callback gets invoked each time the event occurs (or, in the case of I/O
287	watchers, each time the event loop detects that the file descriptor given	959	watchers, each time the event loop detects that the file descriptor given
288	is readable and/or writable).	960	is readable and/or writable).
289		961
290	Each watcher type has its own C<< ev_<type>_set (watcher *, ...) >> macro	962	Each watcher type further has its own C<< ev_TYPE_set (watcher *, ...) >>
291	with arguments specific to this watcher type. There is also a macro	963	macro to configure it, with arguments specific to the watcher type. There
292	to combine initialisation and setting in one call: C<< ev_<type>_init	964	is also a macro to combine initialisation and setting in one call: C<<
293	(watcher *, callback, ...) >>.	965	ev_TYPE_init (watcher *, callback, ...) >>.
294		966
295	To make the watcher actually watch out for events, you have to start it	967	To make the watcher actually watch out for events, you have to start it
296	with a watcher-specific start function (C<< ev_<type>_start (loop, watcher	968	with a watcher-specific start function (C<< ev_TYPE_start (loop, watcher
297	*) >>), and you can stop watching for events at any time by calling the	969	*) >>), and you can stop watching for events at any time by calling the
298	corresponding stop function (C<< ev_<type>_stop (loop, watcher *) >>.	970	corresponding stop function (C<< ev_TYPE_stop (loop, watcher *) >>.
299		971
300	As long as your watcher is active (has been started but not stopped) you	972	As long as your watcher is active (has been started but not stopped) you
301	must not touch the values stored in it. Most specifically you must never	973	must not touch the values stored in it. Most specifically you must never
302	reinitialise it or call its set method.	974	reinitialise it or call its C<ev_TYPE_set> macro.
303
304	You cna check whether an event is active by calling the C<ev_is_active
305	(watcher *)> macro. To see whether an event is outstanding (but the
306	callback for it has not been called yet) you cna use the C<ev_is_pending
307	(watcher *)> macro.
308		975
309	Each and every callback receives the event loop pointer as first, the	976	Each and every callback receives the event loop pointer as first, the
310	registered watcher structure as second, and a bitset of received events as	977	registered watcher structure as second, and a bitset of received events as
311	third argument.	978	third argument.
312		979
313	The rceeived events usually include a single bit per event type received	980	The received events usually include a single bit per event type received
314	(you can receive multiple events at the same time). The possible bit masks	981	(you can receive multiple events at the same time). The possible bit masks
315	are:	982	are:
316		983
317	=over 4	984	=over 4
318		985
…		…
336	The signal specified in the C<ev_signal> watcher has been received by a thread.	1003	The signal specified in the C<ev_signal> watcher has been received by a thread.
337		1004
338	=item C<EV_CHILD>	1005	=item C<EV_CHILD>
339		1006
340	The pid specified in the C<ev_child> watcher has received a status change.	1007	The pid specified in the C<ev_child> watcher has received a status change.
		1008
		1009	=item C<EV_STAT>
		1010
		1011	The path specified in the C<ev_stat> watcher changed its attributes somehow.
341		1012
342	=item C<EV_IDLE>	1013	=item C<EV_IDLE>
343		1014
344	The C<ev_idle> watcher has determined that you have nothing better to do.	1015	The C<ev_idle> watcher has determined that you have nothing better to do.
345		1016
…		…
353	received events. Callbacks of both watcher types can start and stop as	1024	received events. Callbacks of both watcher types can start and stop as
354	many watchers as they want, and all of them will be taken into account	1025	many watchers as they want, and all of them will be taken into account
355	(for example, a C<ev_prepare> watcher might start an idle watcher to keep	1026	(for example, a C<ev_prepare> watcher might start an idle watcher to keep
356	C<ev_loop> from blocking).	1027	C<ev_loop> from blocking).
357		1028
		1029	=item C<EV_EMBED>
		1030
		1031	The embedded event loop specified in the C<ev_embed> watcher needs attention.
		1032
		1033	=item C<EV_FORK>
		1034
		1035	The event loop has been resumed in the child process after fork (see
		1036	C<ev_fork>).
		1037
		1038	=item C<EV_ASYNC>
		1039
		1040	The given async watcher has been asynchronously notified (see C<ev_async>).
		1041
		1042	=item C<EV_CUSTOM>
		1043
		1044	Not ever sent (or otherwise used) by libev itself, but can be freely used
		1045	by libev users to signal watchers (e.g. via C<ev_feed_event>).
		1046
358	=item C<EV_ERROR>	1047	=item C<EV_ERROR>
359		1048
360	An unspecified error has occured, the watcher has been stopped. This might	1049	An unspecified error has occurred, the watcher has been stopped. This might
361	happen because the watcher could not be properly started because libev	1050	happen because the watcher could not be properly started because libev
362	ran out of memory, a file descriptor was found to be closed or any other	1051	ran out of memory, a file descriptor was found to be closed or any other
		1052	problem. Libev considers these application bugs.
		1053
363	problem. You best act on it by reporting the problem and somehow coping	1054	You best act on it by reporting the problem and somehow coping with the
364	with the watcher being stopped.	1055	watcher being stopped. Note that well-written programs should not receive
		1056	an error ever, so when your watcher receives it, this usually indicates a
		1057	bug in your program.
365		1058
366	Libev will usually signal a few "dummy" events together with an error,	1059	Libev will usually signal a few "dummy" events together with an error, for
367	for example it might indicate that a fd is readable or writable, and if	1060	example it might indicate that a fd is readable or writable, and if your
368	your callbacks is well-written it can just attempt the operation and cope	1061	callbacks is well-written it can just attempt the operation and cope with
369	with the error from read() or write(). This will not work in multithreaded	1062	the error from read() or write(). This will not work in multi-threaded
370	programs, though, so beware.	1063	programs, though, as the fd could already be closed and reused for another
		1064	thing, so beware.
371		1065
372	=back	1066	=back
373		1067
		1068	=head2 GENERIC WATCHER FUNCTIONS
		1069
		1070	=over 4
		1071
		1072	=item C<ev_init> (ev_TYPE *watcher, callback)
		1073
		1074	This macro initialises the generic portion of a watcher. The contents
		1075	of the watcher object can be arbitrary (so C<malloc> will do). Only
		1076	the generic parts of the watcher are initialised, you I<need> to call
		1077	the type-specific C<ev_TYPE_set> macro afterwards to initialise the
		1078	type-specific parts. For each type there is also a C<ev_TYPE_init> macro
		1079	which rolls both calls into one.
		1080
		1081	You can reinitialise a watcher at any time as long as it has been stopped
		1082	(or never started) and there are no pending events outstanding.
		1083
		1084	The callback is always of type C<void (*)(struct ev_loop *loop, ev_TYPE *watcher,
		1085	int revents)>.
		1086
		1087	Example: Initialise an C<ev_io> watcher in two steps.
		1088
		1089	ev_io w;
		1090	ev_init (&w, my_cb);
		1091	ev_io_set (&w, STDIN_FILENO, EV_READ);
		1092
		1093	=item C<ev_TYPE_set> (ev_TYPE *, [args])
		1094
		1095	This macro initialises the type-specific parts of a watcher. You need to
		1096	call C<ev_init> at least once before you call this macro, but you can
		1097	call C<ev_TYPE_set> any number of times. You must not, however, call this
		1098	macro on a watcher that is active (it can be pending, however, which is a
		1099	difference to the C<ev_init> macro).
		1100
		1101	Although some watcher types do not have type-specific arguments
		1102	(e.g. C<ev_prepare>) you still need to call its C<set> macro.
		1103
		1104	See C<ev_init>, above, for an example.
		1105
		1106	=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])
		1107
		1108	This convenience macro rolls both C<ev_init> and C<ev_TYPE_set> macro
		1109	calls into a single call. This is the most convenient method to initialise
		1110	a watcher. The same limitations apply, of course.
		1111
		1112	Example: Initialise and set an C<ev_io> watcher in one step.
		1113
		1114	ev_io_init (&w, my_cb, STDIN_FILENO, EV_READ);
		1115
		1116	=item C<ev_TYPE_start> (loop *, ev_TYPE *watcher)
		1117
		1118	Starts (activates) the given watcher. Only active watchers will receive
		1119	events. If the watcher is already active nothing will happen.
		1120
		1121	Example: Start the C<ev_io> watcher that is being abused as example in this
		1122	whole section.
		1123
		1124	ev_io_start (EV_DEFAULT_UC, &w);
		1125
		1126	=item C<ev_TYPE_stop> (loop *, ev_TYPE *watcher)
		1127
		1128	Stops the given watcher if active, and clears the pending status (whether
		1129	the watcher was active or not).
		1130
		1131	It is possible that stopped watchers are pending - for example,
		1132	non-repeating timers are being stopped when they become pending - but
		1133	calling C<ev_TYPE_stop> ensures that the watcher is neither active nor
		1134	pending. If you want to free or reuse the memory used by the watcher it is
		1135	therefore a good idea to always call its C<ev_TYPE_stop> function.
		1136
		1137	=item bool ev_is_active (ev_TYPE *watcher)
		1138
		1139	Returns a true value iff the watcher is active (i.e. it has been started
		1140	and not yet been stopped). As long as a watcher is active you must not modify
		1141	it.
		1142
		1143	=item bool ev_is_pending (ev_TYPE *watcher)
		1144
		1145	Returns a true value iff the watcher is pending, (i.e. it has outstanding
		1146	events but its callback has not yet been invoked). As long as a watcher
		1147	is pending (but not active) you must not call an init function on it (but
		1148	C<ev_TYPE_set> is safe), you must not change its priority, and you must
		1149	make sure the watcher is available to libev (e.g. you cannot C<free ()>
		1150	it).
		1151
		1152	=item callback ev_cb (ev_TYPE *watcher)
		1153
		1154	Returns the callback currently set on the watcher.
		1155
		1156	=item ev_cb_set (ev_TYPE *watcher, callback)
		1157
		1158	Change the callback. You can change the callback at virtually any time
		1159	(modulo threads).
		1160
		1161	=item ev_set_priority (ev_TYPE *watcher, priority)
		1162
		1163	=item int ev_priority (ev_TYPE *watcher)
		1164
		1165	Set and query the priority of the watcher. The priority is a small
		1166	integer between C<EV_MAXPRI> (default: C<2>) and C<EV_MINPRI>
		1167	(default: C<-2>). Pending watchers with higher priority will be invoked
		1168	before watchers with lower priority, but priority will not keep watchers
		1169	from being executed (except for C<ev_idle> watchers).
		1170
		1171	If you need to suppress invocation when higher priority events are pending
		1172	you need to look at C<ev_idle> watchers, which provide this functionality.
		1173
		1174	You I<must not> change the priority of a watcher as long as it is active or
		1175	pending.
		1176
		1177	Setting a priority outside the range of C<EV_MINPRI> to C<EV_MAXPRI> is
		1178	fine, as long as you do not mind that the priority value you query might
		1179	or might not have been clamped to the valid range.
		1180
		1181	The default priority used by watchers when no priority has been set is
		1182	always C<0>, which is supposed to not be too high and not be too low :).
		1183
		1184	See L<WATCHER PRIORITY MODELS>, below, for a more thorough treatment of
		1185	priorities.
		1186
		1187	=item ev_invoke (loop, ev_TYPE *watcher, int revents)
		1188
		1189	Invoke the C<watcher> with the given C<loop> and C<revents>. Neither
		1190	C<loop> nor C<revents> need to be valid as long as the watcher callback
		1191	can deal with that fact, as both are simply passed through to the
		1192	callback.
		1193
		1194	=item int ev_clear_pending (loop, ev_TYPE *watcher)
		1195
		1196	If the watcher is pending, this function clears its pending status and
		1197	returns its C<revents> bitset (as if its callback was invoked). If the
		1198	watcher isn't pending it does nothing and returns C<0>.
		1199
		1200	Sometimes it can be useful to "poll" a watcher instead of waiting for its
		1201	callback to be invoked, which can be accomplished with this function.
		1202
		1203	=back
		1204
		1205
374	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER	1206	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER
375		1207
376	Each watcher has, by default, a member C<void *data> that you can change	1208	Each watcher has, by default, a member C<void *data> that you can change
377	and read at any time, libev will completely ignore it. This cna be used	1209	and read at any time: libev will completely ignore it. This can be used
378	to associate arbitrary data with your watcher. If you need more data and	1210	to associate arbitrary data with your watcher. If you need more data and
379	don't want to allocate memory and store a pointer to it in that data	1211	don't want to allocate memory and store a pointer to it in that data
380	member, you can also "subclass" the watcher type and provide your own	1212	member, you can also "subclass" the watcher type and provide your own
381	data:	1213	data:
382		1214
383	struct my_io	1215	struct my_io
384	{	1216	{
385	struct ev_io io;	1217	ev_io io;
386	int otherfd;	1218	int otherfd;
387	void *somedata;	1219	void *somedata;
388	struct whatever *mostinteresting;	1220	struct whatever *mostinteresting;
389	}	1221	};
		1222
		1223	...
		1224	struct my_io w;
		1225	ev_io_init (&w.io, my_cb, fd, EV_READ);
390		1226
391	And since your callback will be called with a pointer to the watcher, you	1227	And since your callback will be called with a pointer to the watcher, you
392	can cast it back to your own type:	1228	can cast it back to your own type:
393		1229
394	static void my_cb (struct ev_loop *loop, struct ev_io *w_, int revents)	1230	static void my_cb (struct ev_loop *loop, ev_io *w_, int revents)
395	{	1231	{
396	struct my_io *w = (struct my_io *)w_;	1232	struct my_io *w = (struct my_io *)w_;
397	...	1233	...
398	}	1234	}
399		1235
400	More interesting and less C-conformant ways of catsing your callback type	1236	More interesting and less C-conformant ways of casting your callback type
401	have been omitted....	1237	instead have been omitted.
		1238
		1239	Another common scenario is to use some data structure with multiple
		1240	embedded watchers:
		1241
		1242	struct my_biggy
		1243	{
		1244	int some_data;
		1245	ev_timer t1;
		1246	ev_timer t2;
		1247	}
		1248
		1249	In this case getting the pointer to C<my_biggy> is a bit more
		1250	complicated: Either you store the address of your C<my_biggy> struct
		1251	in the C<data> member of the watcher (for woozies), or you need to use
		1252	some pointer arithmetic using C<offsetof> inside your watchers (for real
		1253	programmers):
		1254
		1255	#include <stddef.h>
		1256
		1257	static void
		1258	t1_cb (EV_P_ ev_timer *w, int revents)
		1259	{
		1260	struct my_biggy big = (struct my_biggy *)
		1261	(((char *)w) - offsetof (struct my_biggy, t1));
		1262	}
		1263
		1264	static void
		1265	t2_cb (EV_P_ ev_timer *w, int revents)
		1266	{
		1267	struct my_biggy big = (struct my_biggy *)
		1268	(((char *)w) - offsetof (struct my_biggy, t2));
		1269	}
		1270
		1271	=head2 WATCHER PRIORITY MODELS
		1272
		1273	Many event loops support I<watcher priorities>, which are usually small
		1274	integers that influence the ordering of event callback invocation
		1275	between watchers in some way, all else being equal.
		1276
		1277	In libev, Watcher priorities can be set using C<ev_set_priority>. See its
		1278	description for the more technical details such as the actual priority
		1279	range.
		1280
		1281	There are two common ways how these these priorities are being interpreted
		1282	by event loops:
		1283
		1284	In the more common lock-out model, higher priorities "lock out" invocation
		1285	of lower priority watchers, which means as long as higher priority
		1286	watchers receive events, lower priority watchers are not being invoked.
		1287
		1288	The less common only-for-ordering model uses priorities solely to order
		1289	callback invocation within a single event loop iteration: Higher priority
		1290	watchers are invoked before lower priority ones, but they all get invoked
		1291	before polling for new events.
		1292
		1293	Libev uses the second (only-for-ordering) model for all its watchers
		1294	except for idle watchers (which use the lock-out model).
		1295
		1296	The rationale behind this is that implementing the lock-out model for
		1297	watchers is not well supported by most kernel interfaces, and most event
		1298	libraries will just poll for the same events again and again as long as
		1299	their callbacks have not been executed, which is very inefficient in the
		1300	common case of one high-priority watcher locking out a mass of lower
		1301	priority ones.
		1302
		1303	Static (ordering) priorities are most useful when you have two or more
		1304	watchers handling the same resource: a typical usage example is having an
		1305	C<ev_io> watcher to receive data, and an associated C<ev_timer> to handle
		1306	timeouts. Under load, data might be received while the program handles
		1307	other jobs, but since timers normally get invoked first, the timeout
		1308	handler will be executed before checking for data. In that case, giving
		1309	the timer a lower priority than the I/O watcher ensures that I/O will be
		1310	handled first even under adverse conditions (which is usually, but not
		1311	always, what you want).
		1312
		1313	Since idle watchers use the "lock-out" model, meaning that idle watchers
		1314	will only be executed when no same or higher priority watchers have
		1315	received events, they can be used to implement the "lock-out" model when
		1316	required.
		1317
		1318	For example, to emulate how many other event libraries handle priorities,
		1319	you can associate an C<ev_idle> watcher to each such watcher, and in
		1320	the normal watcher callback, you just start the idle watcher. The real
		1321	processing is done in the idle watcher callback. This causes libev to
		1322	continously poll and process kernel event data for the watcher, but when
		1323	the lock-out case is known to be rare (which in turn is rare :), this is
		1324	workable.
		1325
		1326	Usually, however, the lock-out model implemented that way will perform
		1327	miserably under the type of load it was designed to handle. In that case,
		1328	it might be preferable to stop the real watcher before starting the
		1329	idle watcher, so the kernel will not have to process the event in case
		1330	the actual processing will be delayed for considerable time.
		1331
		1332	Here is an example of an I/O watcher that should run at a strictly lower
		1333	priority than the default, and which should only process data when no
		1334	other events are pending:
		1335
		1336	ev_idle idle; // actual processing watcher
		1337	ev_io io; // actual event watcher
		1338
		1339	static void
		1340	io_cb (EV_P_ ev_io *w, int revents)
		1341	{
		1342	// stop the I/O watcher, we received the event, but
		1343	// are not yet ready to handle it.
		1344	ev_io_stop (EV_A_ w);
		1345
		1346	// start the idle watcher to ahndle the actual event.
		1347	// it will not be executed as long as other watchers
		1348	// with the default priority are receiving events.
		1349	ev_idle_start (EV_A_ &idle);
		1350	}
		1351
		1352	static void
		1353	idle_cb (EV_P_ ev_idle *w, int revents)
		1354	{
		1355	// actual processing
		1356	read (STDIN_FILENO, ...);
		1357
		1358	// have to start the I/O watcher again, as
		1359	// we have handled the event
		1360	ev_io_start (EV_P_ &io);
		1361	}
		1362
		1363	// initialisation
		1364	ev_idle_init (&idle, idle_cb);
		1365	ev_io_init (&io, io_cb, STDIN_FILENO, EV_READ);
		1366	ev_io_start (EV_DEFAULT_ &io);
		1367
		1368	In the "real" world, it might also be beneficial to start a timer, so that
		1369	low-priority connections can not be locked out forever under load. This
		1370	enables your program to keep a lower latency for important connections
		1371	during short periods of high load, while not completely locking out less
		1372	important ones.
402		1373
403		1374
404	=head1 WATCHER TYPES	1375	=head1 WATCHER TYPES
405		1376
406	This section describes each watcher in detail, but will not repeat	1377	This section describes each watcher in detail, but will not repeat
407	information given in the last section.	1378	information given in the last section. Any initialisation/set macros,
		1379	functions and members specific to the watcher type are explained.
408		1380
		1381	Members are additionally marked with either I<[read-only]>, meaning that,
		1382	while the watcher is active, you can look at the member and expect some
		1383	sensible content, but you must not modify it (you can modify it while the
		1384	watcher is stopped to your hearts content), or I<[read-write]>, which
		1385	means you can expect it to have some sensible content while the watcher
		1386	is active, but you can also modify it. Modifying it may not do something
		1387	sensible or take immediate effect (or do anything at all), but libev will
		1388	not crash or malfunction in any way.
		1389
		1390
409	=head2 C<ev_io> - is this file descriptor readable or writable	1391	=head2 C<ev_io> - is this file descriptor readable or writable?
410		1392
411	I/O watchers check whether a file descriptor is readable or writable	1393	I/O watchers check whether a file descriptor is readable or writable
412	in each iteration of the event loop (This behaviour is called	1394	in each iteration of the event loop, or, more precisely, when reading
413	level-triggering because you keep receiving events as long as the	1395	would not block the process and writing would at least be able to write
414	condition persists. Remember you cna stop the watcher if you don't want to	1396	some data. This behaviour is called level-triggering because you keep
415	act on the event and neither want to receive future events).	1397	receiving events as long as the condition persists. Remember you can stop
		1398	the watcher if you don't want to act on the event and neither want to
		1399	receive future events.
416		1400
417	In general you can register as many read and/or write event watchers oer	1401	In general you can register as many read and/or write event watchers per
418	fd as you want (as long as you don't confuse yourself). Setting all file	1402	fd as you want (as long as you don't confuse yourself). Setting all file
419	descriptors to non-blocking mode is also usually a good idea (but not	1403	descriptors to non-blocking mode is also usually a good idea (but not
420	required if you know what you are doing).	1404	required if you know what you are doing).
421		1405
422	You have to be careful with dup'ed file descriptors, though. Some backends	1406	If you cannot use non-blocking mode, then force the use of a
423	(the linux epoll backend is a notable example) cannot handle dup'ed file	1407	known-to-be-good backend (at the time of this writing, this includes only
424	descriptors correctly if you register interest in two or more fds pointing	1408	C<EVBACKEND_SELECT> and C<EVBACKEND_POLL>). The same applies to file
425	to the same file/socket etc. description.	1409	descriptors for which non-blocking operation makes no sense (such as
		1410	files) - libev doesn't guarentee any specific behaviour in that case.
426		1411
427	If you must do this, then force the use of a known-to-be-good backend	1412	Another thing you have to watch out for is that it is quite easy to
428	(at the time of this writing, this includes only EVMETHOD_SELECT and	1413	receive "spurious" readiness notifications, that is your callback might
429	EVMETHOD_POLL).	1414	be called with C<EV_READ> but a subsequent C<read>(2) will actually block
		1415	because there is no data. Not only are some backends known to create a
		1416	lot of those (for example Solaris ports), it is very easy to get into
		1417	this situation even with a relatively standard program structure. Thus
		1418	it is best to always use non-blocking I/O: An extra C<read>(2) returning
		1419	C<EAGAIN> is far preferable to a program hanging until some data arrives.
		1420
		1421	If you cannot run the fd in non-blocking mode (for example you should
		1422	not play around with an Xlib connection), then you have to separately
		1423	re-test whether a file descriptor is really ready with a known-to-be good
		1424	interface such as poll (fortunately in our Xlib example, Xlib already
		1425	does this on its own, so its quite safe to use). Some people additionally
		1426	use C<SIGALRM> and an interval timer, just to be sure you won't block
		1427	indefinitely.
		1428
		1429	But really, best use non-blocking mode.
		1430
		1431	=head3 The special problem of disappearing file descriptors
		1432
		1433	Some backends (e.g. kqueue, epoll) need to be told about closing a file
		1434	descriptor (either due to calling C<close> explicitly or any other means,
		1435	such as C<dup2>). The reason is that you register interest in some file
		1436	descriptor, but when it goes away, the operating system will silently drop
		1437	this interest. If another file descriptor with the same number then is
		1438	registered with libev, there is no efficient way to see that this is, in
		1439	fact, a different file descriptor.
		1440
		1441	To avoid having to explicitly tell libev about such cases, libev follows
		1442	the following policy: Each time C<ev_io_set> is being called, libev
		1443	will assume that this is potentially a new file descriptor, otherwise
		1444	it is assumed that the file descriptor stays the same. That means that
		1445	you I<have> to call C<ev_io_set> (or C<ev_io_init>) when you change the
		1446	descriptor even if the file descriptor number itself did not change.
		1447
		1448	This is how one would do it normally anyway, the important point is that
		1449	the libev application should not optimise around libev but should leave
		1450	optimisations to libev.
		1451
		1452	=head3 The special problem of dup'ed file descriptors
		1453
		1454	Some backends (e.g. epoll), cannot register events for file descriptors,
		1455	but only events for the underlying file descriptions. That means when you
		1456	have C<dup ()>'ed file descriptors or weirder constellations, and register
		1457	events for them, only one file descriptor might actually receive events.
		1458
		1459	There is no workaround possible except not registering events
		1460	for potentially C<dup ()>'ed file descriptors, or to resort to
		1461	C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>.
		1462
		1463	=head3 The special problem of fork
		1464
		1465	Some backends (epoll, kqueue) do not support C<fork ()> at all or exhibit
		1466	useless behaviour. Libev fully supports fork, but needs to be told about
		1467	it in the child.
		1468
		1469	To support fork in your programs, you either have to call
		1470	C<ev_default_fork ()> or C<ev_loop_fork ()> after a fork in the child,
		1471	enable C<EVFLAG_FORKCHECK>, or resort to C<EVBACKEND_SELECT> or
		1472	C<EVBACKEND_POLL>.
		1473
		1474	=head3 The special problem of SIGPIPE
		1475
		1476	While not really specific to libev, it is easy to forget about C<SIGPIPE>:
		1477	when writing to a pipe whose other end has been closed, your program gets
		1478	sent a SIGPIPE, which, by default, aborts your program. For most programs
		1479	this is sensible behaviour, for daemons, this is usually undesirable.
		1480
		1481	So when you encounter spurious, unexplained daemon exits, make sure you
		1482	ignore SIGPIPE (and maybe make sure you log the exit status of your daemon
		1483	somewhere, as that would have given you a big clue).
		1484
		1485
		1486	=head3 Watcher-Specific Functions
430		1487
431	=over 4	1488	=over 4
432		1489
433	=item ev_io_init (ev_io *, callback, int fd, int events)	1490	=item ev_io_init (ev_io *, callback, int fd, int events)
434		1491
435	=item ev_io_set (ev_io *, int fd, int events)	1492	=item ev_io_set (ev_io *, int fd, int events)
436		1493
437	Configures an C<ev_io> watcher. The fd is the file descriptor to rceeive	1494	Configures an C<ev_io> watcher. The C<fd> is the file descriptor to
438	events for and events is either C<EV_READ>, C<EV_WRITE> or C<EV_READ |	1495	receive events for and C<events> is either C<EV_READ>, C<EV_WRITE> or
439	EV_WRITE> to receive the given events.	1496	C<EV_READ | EV_WRITE>, to express the desire to receive the given events.
		1497
		1498	=item int fd [read-only]
		1499
		1500	The file descriptor being watched.
		1501
		1502	=item int events [read-only]
		1503
		1504	The events being watched.
440		1505
441	=back	1506	=back
442		1507
		1508	=head3 Examples
		1509
		1510	Example: Call C<stdin_readable_cb> when STDIN_FILENO has become, well
		1511	readable, but only once. Since it is likely line-buffered, you could
		1512	attempt to read a whole line in the callback.
		1513
		1514	static void
		1515	stdin_readable_cb (struct ev_loop *loop, ev_io *w, int revents)
		1516	{
		1517	ev_io_stop (loop, w);
		1518	.. read from stdin here (or from w->fd) and handle any I/O errors
		1519	}
		1520
		1521	...
		1522	struct ev_loop *loop = ev_default_init (0);
		1523	ev_io stdin_readable;
		1524	ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
		1525	ev_io_start (loop, &stdin_readable);
		1526	ev_loop (loop, 0);
		1527
		1528
443	=head2 C<ev_timer> - relative and optionally recurring timeouts	1529	=head2 C<ev_timer> - relative and optionally repeating timeouts
444		1530
445	Timer watchers are simple relative timers that generate an event after a	1531	Timer watchers are simple relative timers that generate an event after a
446	given time, and optionally repeating in regular intervals after that.	1532	given time, and optionally repeating in regular intervals after that.
447		1533
448	The timers are based on real time, that is, if you register an event that	1534	The timers are based on real time, that is, if you register an event that
449	times out after an hour and youreset your system clock to last years	1535	times out after an hour and you reset your system clock to January last
450	time, it will still time out after (roughly) and hour. "Roughly" because	1536	year, it will still time out after (roughly) one hour. "Roughly" because
451	detecting time jumps is hard, and soem inaccuracies are unavoidable (the	1537	detecting time jumps is hard, and some inaccuracies are unavoidable (the
452	monotonic clock option helps a lot here).	1538	monotonic clock option helps a lot here).
		1539
		1540	The callback is guaranteed to be invoked only I<after> its timeout has
		1541	passed (not I<at>, so on systems with very low-resolution clocks this
		1542	might introduce a small delay). If multiple timers become ready during the
		1543	same loop iteration then the ones with earlier time-out values are invoked
		1544	before ones of the same priority with later time-out values (but this is
		1545	no longer true when a callback calls C<ev_loop> recursively).
		1546
		1547	=head3 Be smart about timeouts
		1548
		1549	Many real-world problems involve some kind of timeout, usually for error
		1550	recovery. A typical example is an HTTP request - if the other side hangs,
		1551	you want to raise some error after a while.
		1552
		1553	What follows are some ways to handle this problem, from obvious and
		1554	inefficient to smart and efficient.
		1555
		1556	In the following, a 60 second activity timeout is assumed - a timeout that
		1557	gets reset to 60 seconds each time there is activity (e.g. each time some
		1558	data or other life sign was received).
		1559
		1560	=over 4
		1561
		1562	=item 1. Use a timer and stop, reinitialise and start it on activity.
		1563
		1564	This is the most obvious, but not the most simple way: In the beginning,
		1565	start the watcher:
		1566
		1567	ev_timer_init (timer, callback, 60., 0.);
		1568	ev_timer_start (loop, timer);
		1569
		1570	Then, each time there is some activity, C<ev_timer_stop> it, initialise it
		1571	and start it again:
		1572
		1573	ev_timer_stop (loop, timer);
		1574	ev_timer_set (timer, 60., 0.);
		1575	ev_timer_start (loop, timer);
		1576
		1577	This is relatively simple to implement, but means that each time there is
		1578	some activity, libev will first have to remove the timer from its internal
		1579	data structure and then add it again. Libev tries to be fast, but it's
		1580	still not a constant-time operation.
		1581
		1582	=item 2. Use a timer and re-start it with C<ev_timer_again> inactivity.
		1583
		1584	This is the easiest way, and involves using C<ev_timer_again> instead of
		1585	C<ev_timer_start>.
		1586
		1587	To implement this, configure an C<ev_timer> with a C<repeat> value
		1588	of C<60> and then call C<ev_timer_again> at start and each time you
		1589	successfully read or write some data. If you go into an idle state where
		1590	you do not expect data to travel on the socket, you can C<ev_timer_stop>
		1591	the timer, and C<ev_timer_again> will automatically restart it if need be.
		1592
		1593	That means you can ignore both the C<ev_timer_start> function and the
		1594	C<after> argument to C<ev_timer_set>, and only ever use the C<repeat>
		1595	member and C<ev_timer_again>.
		1596
		1597	At start:
		1598
		1599	ev_init (timer, callback);
		1600	timer->repeat = 60.;
		1601	ev_timer_again (loop, timer);
		1602
		1603	Each time there is some activity:
		1604
		1605	ev_timer_again (loop, timer);
		1606
		1607	It is even possible to change the time-out on the fly, regardless of
		1608	whether the watcher is active or not:
		1609
		1610	timer->repeat = 30.;
		1611	ev_timer_again (loop, timer);
		1612
		1613	This is slightly more efficient then stopping/starting the timer each time
		1614	you want to modify its timeout value, as libev does not have to completely
		1615	remove and re-insert the timer from/into its internal data structure.
		1616
		1617	It is, however, even simpler than the "obvious" way to do it.
		1618
		1619	=item 3. Let the timer time out, but then re-arm it as required.
		1620
		1621	This method is more tricky, but usually most efficient: Most timeouts are
		1622	relatively long compared to the intervals between other activity - in
		1623	our example, within 60 seconds, there are usually many I/O events with
		1624	associated activity resets.
		1625
		1626	In this case, it would be more efficient to leave the C<ev_timer> alone,
		1627	but remember the time of last activity, and check for a real timeout only
		1628	within the callback:
		1629
		1630	ev_tstamp last_activity; // time of last activity
		1631
		1632	static void
		1633	callback (EV_P_ ev_timer *w, int revents)
		1634	{
		1635	ev_tstamp now = ev_now (EV_A);
		1636	ev_tstamp timeout = last_activity + 60.;
		1637
		1638	// if last_activity + 60. is older than now, we did time out
		1639	if (timeout < now)
		1640	{
		1641	// timeout occured, take action
		1642	}
		1643	else
		1644	{
		1645	// callback was invoked, but there was some activity, re-arm
		1646	// the watcher to fire in last_activity + 60, which is
		1647	// guaranteed to be in the future, so "again" is positive:
		1648	w->repeat = timeout - now;
		1649	ev_timer_again (EV_A_ w);
		1650	}
		1651	}
		1652
		1653	To summarise the callback: first calculate the real timeout (defined
		1654	as "60 seconds after the last activity"), then check if that time has
		1655	been reached, which means something I<did>, in fact, time out. Otherwise
		1656	the callback was invoked too early (C<timeout> is in the future), so
		1657	re-schedule the timer to fire at that future time, to see if maybe we have
		1658	a timeout then.
		1659
		1660	Note how C<ev_timer_again> is used, taking advantage of the
		1661	C<ev_timer_again> optimisation when the timer is already running.
		1662
		1663	This scheme causes more callback invocations (about one every 60 seconds
		1664	minus half the average time between activity), but virtually no calls to
		1665	libev to change the timeout.
		1666
		1667	To start the timer, simply initialise the watcher and set C<last_activity>
		1668	to the current time (meaning we just have some activity :), then call the
		1669	callback, which will "do the right thing" and start the timer:
		1670
		1671	ev_init (timer, callback);
		1672	last_activity = ev_now (loop);
		1673	callback (loop, timer, EV_TIMEOUT);
		1674
		1675	And when there is some activity, simply store the current time in
		1676	C<last_activity>, no libev calls at all:
		1677
		1678	last_actiivty = ev_now (loop);
		1679
		1680	This technique is slightly more complex, but in most cases where the
		1681	time-out is unlikely to be triggered, much more efficient.
		1682
		1683	Changing the timeout is trivial as well (if it isn't hard-coded in the
		1684	callback :) - just change the timeout and invoke the callback, which will
		1685	fix things for you.
		1686
		1687	=item 4. Wee, just use a double-linked list for your timeouts.
		1688
		1689	If there is not one request, but many thousands (millions...), all
		1690	employing some kind of timeout with the same timeout value, then one can
		1691	do even better:
		1692
		1693	When starting the timeout, calculate the timeout value and put the timeout
		1694	at the I<end> of the list.
		1695
		1696	Then use an C<ev_timer> to fire when the timeout at the I<beginning> of
		1697	the list is expected to fire (for example, using the technique #3).
		1698
		1699	When there is some activity, remove the timer from the list, recalculate
		1700	the timeout, append it to the end of the list again, and make sure to
		1701	update the C<ev_timer> if it was taken from the beginning of the list.
		1702
		1703	This way, one can manage an unlimited number of timeouts in O(1) time for
		1704	starting, stopping and updating the timers, at the expense of a major
		1705	complication, and having to use a constant timeout. The constant timeout
		1706	ensures that the list stays sorted.
		1707
		1708	=back
		1709
		1710	So which method the best?
		1711
		1712	Method #2 is a simple no-brain-required solution that is adequate in most
		1713	situations. Method #3 requires a bit more thinking, but handles many cases
		1714	better, and isn't very complicated either. In most case, choosing either
		1715	one is fine, with #3 being better in typical situations.
		1716
		1717	Method #1 is almost always a bad idea, and buys you nothing. Method #4 is
		1718	rather complicated, but extremely efficient, something that really pays
		1719	off after the first million or so of active timers, i.e. it's usually
		1720	overkill :)
		1721
		1722	=head3 The special problem of time updates
		1723
		1724	Establishing the current time is a costly operation (it usually takes at
		1725	least two system calls): EV therefore updates its idea of the current
		1726	time only before and after C<ev_loop> collects new events, which causes a
		1727	growing difference between C<ev_now ()> and C<ev_time ()> when handling
		1728	lots of events in one iteration.
453		1729
454	The relative timeouts are calculated relative to the C<ev_now ()>	1730	The relative timeouts are calculated relative to the C<ev_now ()>
455	time. This is usually the right thing as this timestamp refers to the time	1731	time. This is usually the right thing as this timestamp refers to the time
456	of the event triggering whatever timeout you are modifying/starting. If	1732	of the event triggering whatever timeout you are modifying/starting. If
457	you suspect event processing to be delayed and you *need* to base the timeout	1733	you suspect event processing to be delayed and you I<need> to base the
458	ion the current time, use something like this to adjust for this:	1734	timeout on the current time, use something like this to adjust for this:
459		1735
460	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);	1736	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);
461		1737
		1738	If the event loop is suspended for a long time, you can also force an
		1739	update of the time returned by C<ev_now ()> by calling C<ev_now_update
		1740	()>.
		1741
		1742	=head3 Watcher-Specific Functions and Data Members
		1743
462	=over 4	1744	=over 4
463		1745
464	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)	1746	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)
465		1747
466	=item ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)	1748	=item ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)
467		1749
468	Configure the timer to trigger after C<after> seconds. If C<repeat> is	1750	Configure the timer to trigger after C<after> seconds. If C<repeat>
469	C<0.>, then it will automatically be stopped. If it is positive, then the	1751	is C<0.>, then it will automatically be stopped once the timeout is
470	timer will automatically be configured to trigger again C<repeat> seconds	1752	reached. If it is positive, then the timer will automatically be
471	later, again, and again, until stopped manually.	1753	configured to trigger again C<repeat> seconds later, again, and again,
		1754	until stopped manually.
472		1755
473	The timer itself will do a best-effort at avoiding drift, that is, if you	1756	The timer itself will do a best-effort at avoiding drift, that is, if
474	configure a timer to trigger every 10 seconds, then it will trigger at	1757	you configure a timer to trigger every 10 seconds, then it will normally
475	exactly 10 second intervals. If, however, your program cannot keep up with	1758	trigger at exactly 10 second intervals. If, however, your program cannot
476	the timer (ecause it takes longer than those 10 seconds to do stuff) the	1759	keep up with the timer (because it takes longer than those 10 seconds to
477	timer will not fire more than once per event loop iteration.	1760	do stuff) the timer will not fire more than once per event loop iteration.
478		1761
479	=item ev_timer_again (loop)	1762	=item ev_timer_again (loop, ev_timer *)
480		1763
481	This will act as if the timer timed out and restart it again if it is	1764	This will act as if the timer timed out and restart it again if it is
482	repeating. The exact semantics are:	1765	repeating. The exact semantics are:
483		1766
		1767	If the timer is pending, its pending status is cleared.
		1768
484	If the timer is started but nonrepeating, stop it.	1769	If the timer is started but non-repeating, stop it (as if it timed out).
485		1770
486	If the timer is repeating, either start it if necessary (with the repeat	1771	If the timer is repeating, either start it if necessary (with the
487	value), or reset the running timer to the repeat value.	1772	C<repeat> value), or reset the running timer to the C<repeat> value.
488		1773
489	This sounds a bit complicated, but here is a useful and typical	1774	This sounds a bit complicated, see L<Be smart about timeouts>, above, for a
490	example: Imagine you have a tcp connection and you want a so-called idle	1775	usage example.
491	timeout, that is, you want to be called when there have been, say, 60	1776
492	seconds of inactivity on the socket. The easiest way to do this is to	1777	=item ev_tstamp repeat [read-write]
493	configure an C<ev_timer> with after=repeat=60 and calling ev_timer_again each	1778
494	time you successfully read or write some data. If you go into an idle	1779	The current C<repeat> value. Will be used each time the watcher times out
495	state where you do not expect data to travel on the socket, you can stop	1780	or C<ev_timer_again> is called, and determines the next timeout (if any),
496	the timer, and again will automatically restart it if need be.	1781	which is also when any modifications are taken into account.
497		1782
498	=back	1783	=back
499		1784
		1785	=head3 Examples
		1786
		1787	Example: Create a timer that fires after 60 seconds.
		1788
		1789	static void
		1790	one_minute_cb (struct ev_loop *loop, ev_timer *w, int revents)
		1791	{
		1792	.. one minute over, w is actually stopped right here
		1793	}
		1794
		1795	ev_timer mytimer;
		1796	ev_timer_init (&mytimer, one_minute_cb, 60., 0.);
		1797	ev_timer_start (loop, &mytimer);
		1798
		1799	Example: Create a timeout timer that times out after 10 seconds of
		1800	inactivity.
		1801
		1802	static void
		1803	timeout_cb (struct ev_loop *loop, ev_timer *w, int revents)
		1804	{
		1805	.. ten seconds without any activity
		1806	}
		1807
		1808	ev_timer mytimer;
		1809	ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */
		1810	ev_timer_again (&mytimer); /* start timer */
		1811	ev_loop (loop, 0);
		1812
		1813	// and in some piece of code that gets executed on any "activity":
		1814	// reset the timeout to start ticking again at 10 seconds
		1815	ev_timer_again (&mytimer);
		1816
		1817
500	=head2 C<ev_periodic> - to cron or not to cron it	1818	=head2 C<ev_periodic> - to cron or not to cron?
501		1819
502	Periodic watchers are also timers of a kind, but they are very versatile	1820	Periodic watchers are also timers of a kind, but they are very versatile
503	(and unfortunately a bit complex).	1821	(and unfortunately a bit complex).
504		1822
505	Unlike C<ev_timer>'s, they are not based on real time (or relative time)	1823	Unlike C<ev_timer>, periodic watchers are not based on real time (or
506	but on wallclock time (absolute time). You can tell a periodic watcher	1824	relative time, the physical time that passes) but on wall clock time
507	to trigger "at" some specific point in time. For example, if you tell a	1825	(absolute time, the thing you can read on your calender or clock). The
508	periodic watcher to trigger in 10 seconds (by specifiying e.g. c<ev_now ()	1826	difference is that wall clock time can run faster or slower than real
509	+ 10.>) and then reset your system clock to the last year, then it will	1827	time, and time jumps are not uncommon (e.g. when you adjust your
510	take a year to trigger the event (unlike an C<ev_timer>, which would trigger	1828	wrist-watch).
511	roughly 10 seconds later and of course not if you reset your system time
512	again).
513		1829
514	They can also be used to implement vastly more complex timers, such as	1830	You can tell a periodic watcher to trigger after some specific point
		1831	in time: for example, if you tell a periodic watcher to trigger "in 10
		1832	seconds" (by specifying e.g. C<ev_now () + 10.>, that is, an absolute time
		1833	not a delay) and then reset your system clock to January of the previous
		1834	year, then it will take a year or more to trigger the event (unlike an
		1835	C<ev_timer>, which would still trigger roughly 10 seconds after starting
		1836	it, as it uses a relative timeout).
		1837
		1838	C<ev_periodic> watchers can also be used to implement vastly more complex
515	triggering an event on eahc midnight, local time.	1839	timers, such as triggering an event on each "midnight, local time", or
		1840	other complicated rules. This cannot be done with C<ev_timer> watchers, as
		1841	those cannot react to time jumps.
		1842
		1843	As with timers, the callback is guaranteed to be invoked only when the
		1844	point in time where it is supposed to trigger has passed. If multiple
		1845	timers become ready during the same loop iteration then the ones with
		1846	earlier time-out values are invoked before ones with later time-out values
		1847	(but this is no longer true when a callback calls C<ev_loop> recursively).
		1848
		1849	=head3 Watcher-Specific Functions and Data Members
516		1850
517	=over 4	1851	=over 4
518		1852
519	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)	1853	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp offset, ev_tstamp interval, reschedule_cb)
520		1854
521	=item ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)	1855	=item ev_periodic_set (ev_periodic *, ev_tstamp offset, ev_tstamp interval, reschedule_cb)
522		1856
523	Lots of arguments, lets sort it out... There are basically three modes of	1857	Lots of arguments, let's sort it out... There are basically three modes of
524	operation, and we will explain them from simplest to complex:	1858	operation, and we will explain them from simplest to most complex:
525
526		1859
527	=over 4	1860	=over 4
528		1861
529	=item * absolute timer (interval = reschedule_cb = 0)	1862	=item * absolute timer (offset = absolute time, interval = 0, reschedule_cb = 0)
530		1863
531	In this configuration the watcher triggers an event at the wallclock time	1864	In this configuration the watcher triggers an event after the wall clock
532	C<at> and doesn't repeat. It will not adjust when a time jump occurs,	1865	time C<offset> has passed. It will not repeat and will not adjust when a
533	that is, if it is to be run at January 1st 2011 then it will run when the	1866	time jump occurs, that is, if it is to be run at January 1st 2011 then it
534	system time reaches or surpasses this time.	1867	will be stopped and invoked when the system clock reaches or surpasses
		1868	this point in time.
535		1869
536	=item * non-repeating interval timer (interval > 0, reschedule_cb = 0)	1870	=item * repeating interval timer (offset = offset within interval, interval > 0, reschedule_cb = 0)
537		1871
538	In this mode the watcher will always be scheduled to time out at the next	1872	In this mode the watcher will always be scheduled to time out at the next
539	C<at + N * interval> time (for some integer N) and then repeat, regardless	1873	C<offset + N * interval> time (for some integer N, which can also be
540	of any time jumps.	1874	negative) and then repeat, regardless of any time jumps. The C<offset>
		1875	argument is merely an offset into the C<interval> periods.
541		1876
542	This can be used to create timers that do not drift with respect to system	1877	This can be used to create timers that do not drift with respect to the
543	time:	1878	system clock, for example, here is an C<ev_periodic> that triggers each
		1879	hour, on the hour (with respect to UTC):
544		1880
545	ev_periodic_set (&periodic, 0., 3600., 0);	1881	ev_periodic_set (&periodic, 0., 3600., 0);
546		1882
547	This doesn't mean there will always be 3600 seconds in between triggers,	1883	This doesn't mean there will always be 3600 seconds in between triggers,
548	but only that the the callback will be called when the system time shows a	1884	but only that the callback will be called when the system time shows a
549	full hour (UTC), or more correct, when the system time is evenly divisible	1885	full hour (UTC), or more correctly, when the system time is evenly divisible
550	by 3600.	1886	by 3600.
551		1887
552	Another way to think about it (for the mathematically inclined) is that	1888	Another way to think about it (for the mathematically inclined) is that
553	C<ev_periodic> will try to run the callback in this mode at the next possible	1889	C<ev_periodic> will try to run the callback in this mode at the next possible
554	time where C<time = at (mod interval)>, regardless of any time jumps.	1890	time where C<time = offset (mod interval)>, regardless of any time jumps.
555		1891
		1892	For numerical stability it is preferable that the C<offset> value is near
		1893	C<ev_now ()> (the current time), but there is no range requirement for
		1894	this value, and in fact is often specified as zero.
		1895
		1896	Note also that there is an upper limit to how often a timer can fire (CPU
		1897	speed for example), so if C<interval> is very small then timing stability
		1898	will of course deteriorate. Libev itself tries to be exact to be about one
		1899	millisecond (if the OS supports it and the machine is fast enough).
		1900
556	=item * manual reschedule mode (reschedule_cb = callback)	1901	=item * manual reschedule mode (offset ignored, interval ignored, reschedule_cb = callback)
557		1902
558	In this mode the values for C<interval> and C<at> are both being	1903	In this mode the values for C<interval> and C<offset> are both being
559	ignored. Instead, each time the periodic watcher gets scheduled, the	1904	ignored. Instead, each time the periodic watcher gets scheduled, the
560	reschedule callback will be called with the watcher as first, and the	1905	reschedule callback will be called with the watcher as first, and the
561	current time as second argument.	1906	current time as second argument.
562		1907
563	NOTE: I<This callback MUST NOT stop or destroy the periodic or any other	1908	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher, ever,
564	periodic watcher, ever, or make any event loop modificstions>. If you need	1909	or make ANY other event loop modifications whatsoever, unless explicitly
565	to stop it, return 1e30 (or so, fudge fudge) and stop it afterwards.	1910	allowed by documentation here>.
566		1911
		1912	If you need to stop it, return C<now + 1e30> (or so, fudge fudge) and stop
		1913	it afterwards (e.g. by starting an C<ev_prepare> watcher, which is the
		1914	only event loop modification you are allowed to do).
		1915
567	Its prototype is c<ev_tstamp (*reschedule_cb)(struct ev_periodic *w,	1916	The callback prototype is C<ev_tstamp (*reschedule_cb)(ev_periodic
568	ev_tstamp now)>, e.g.:	1917	*w, ev_tstamp now)>, e.g.:
569		1918
		1919	static ev_tstamp
570	static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)	1920	my_rescheduler (ev_periodic *w, ev_tstamp now)
571	{	1921	{
572	return now + 60.;	1922	return now + 60.;
573	}	1923	}
574		1924
575	It must return the next time to trigger, based on the passed time value	1925	It must return the next time to trigger, based on the passed time value
576	(that is, the lowest time value larger than to the second argument). It	1926	(that is, the lowest time value larger than to the second argument). It
577	will usually be called just before the callback will be triggered, but	1927	will usually be called just before the callback will be triggered, but
578	might be called at other times, too.	1928	might be called at other times, too.
579		1929
		1930	NOTE: I<< This callback must always return a time that is higher than or
		1931	equal to the passed C<now> value >>.
		1932
580	This can be used to create very complex timers, such as a timer that	1933	This can be used to create very complex timers, such as a timer that
581	triggers on each midnight, local time. To do this, you would calculate the	1934	triggers on "next midnight, local time". To do this, you would calculate the
582	next midnight after C<now> and return the timestamp value for this. How you do this	1935	next midnight after C<now> and return the timestamp value for this. How
583	is, again, up to you (but it is not trivial).	1936	you do this is, again, up to you (but it is not trivial, which is the main
		1937	reason I omitted it as an example).
584		1938
585	=back	1939	=back
586		1940
587	=item ev_periodic_again (loop, ev_periodic *)	1941	=item ev_periodic_again (loop, ev_periodic *)
588		1942
589	Simply stops and restarts the periodic watcher again. This is only useful	1943	Simply stops and restarts the periodic watcher again. This is only useful
590	when you changed some parameters or the reschedule callback would return	1944	when you changed some parameters or the reschedule callback would return
591	a different time than the last time it was called (e.g. in a crond like	1945	a different time than the last time it was called (e.g. in a crond like
592	program when the crontabs have changed).	1946	program when the crontabs have changed).
593		1947
		1948	=item ev_tstamp ev_periodic_at (ev_periodic *)
		1949
		1950	When active, returns the absolute time that the watcher is supposed
		1951	to trigger next. This is not the same as the C<offset> argument to
		1952	C<ev_periodic_set>, but indeed works even in interval and manual
		1953	rescheduling modes.
		1954
		1955	=item ev_tstamp offset [read-write]
		1956
		1957	When repeating, this contains the offset value, otherwise this is the
		1958	absolute point in time (the C<offset> value passed to C<ev_periodic_set>,
		1959	although libev might modify this value for better numerical stability).
		1960
		1961	Can be modified any time, but changes only take effect when the periodic
		1962	timer fires or C<ev_periodic_again> is being called.
		1963
		1964	=item ev_tstamp interval [read-write]
		1965
		1966	The current interval value. Can be modified any time, but changes only
		1967	take effect when the periodic timer fires or C<ev_periodic_again> is being
		1968	called.
		1969
		1970	=item ev_tstamp (*reschedule_cb)(ev_periodic *w, ev_tstamp now) [read-write]
		1971
		1972	The current reschedule callback, or C<0>, if this functionality is
		1973	switched off. Can be changed any time, but changes only take effect when
		1974	the periodic timer fires or C<ev_periodic_again> is being called.
		1975
594	=back	1976	=back
595		1977
		1978	=head3 Examples
		1979
		1980	Example: Call a callback every hour, or, more precisely, whenever the
		1981	system time is divisible by 3600. The callback invocation times have
		1982	potentially a lot of jitter, but good long-term stability.
		1983
		1984	static void
		1985	clock_cb (struct ev_loop *loop, ev_io *w, int revents)
		1986	{
		1987	... its now a full hour (UTC, or TAI or whatever your clock follows)
		1988	}
		1989
		1990	ev_periodic hourly_tick;
		1991	ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);
		1992	ev_periodic_start (loop, &hourly_tick);
		1993
		1994	Example: The same as above, but use a reschedule callback to do it:
		1995
		1996	#include <math.h>
		1997
		1998	static ev_tstamp
		1999	my_scheduler_cb (ev_periodic *w, ev_tstamp now)
		2000	{
		2001	return now + (3600. - fmod (now, 3600.));
		2002	}
		2003
		2004	ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
		2005
		2006	Example: Call a callback every hour, starting now:
		2007
		2008	ev_periodic hourly_tick;
		2009	ev_periodic_init (&hourly_tick, clock_cb,
		2010	fmod (ev_now (loop), 3600.), 3600., 0);
		2011	ev_periodic_start (loop, &hourly_tick);
		2012
		2013
596	=head2 C<ev_signal> - signal me when a signal gets signalled	2014	=head2 C<ev_signal> - signal me when a signal gets signalled!
597		2015
598	Signal watchers will trigger an event when the process receives a specific	2016	Signal watchers will trigger an event when the process receives a specific
599	signal one or more times. Even though signals are very asynchronous, libev	2017	signal one or more times. Even though signals are very asynchronous, libev
600	will try it's best to deliver signals synchronously, i.e. as part of the	2018	will try it's best to deliver signals synchronously, i.e. as part of the
601	normal event processing, like any other event.	2019	normal event processing, like any other event.
602		2020
		2021	If you want signals asynchronously, just use C<sigaction> as you would
		2022	do without libev and forget about sharing the signal. You can even use
		2023	C<ev_async> from a signal handler to synchronously wake up an event loop.
		2024
603	You cna configure as many watchers as you like per signal. Only when the	2025	You can configure as many watchers as you like per signal. Only when the
604	first watcher gets started will libev actually register a signal watcher	2026	first watcher gets started will libev actually register a signal handler
605	with the kernel (thus it coexists with your own signal handlers as long	2027	with the kernel (thus it coexists with your own signal handlers as long as
606	as you don't register any with libev). Similarly, when the last signal	2028	you don't register any with libev for the same signal). Similarly, when
607	watcher for a signal is stopped libev will reset the signal handler to	2029	the last signal watcher for a signal is stopped, libev will reset the
608	SIG_DFL (regardless of what it was set to before).	2030	signal handler to SIG_DFL (regardless of what it was set to before).
		2031
		2032	If possible and supported, libev will install its handlers with
		2033	C<SA_RESTART> behaviour enabled, so system calls should not be unduly
		2034	interrupted. If you have a problem with system calls getting interrupted by
		2035	signals you can block all signals in an C<ev_check> watcher and unblock
		2036	them in an C<ev_prepare> watcher.
		2037
		2038	=head3 Watcher-Specific Functions and Data Members
609		2039
610	=over 4	2040	=over 4
611		2041
612	=item ev_signal_init (ev_signal *, callback, int signum)	2042	=item ev_signal_init (ev_signal *, callback, int signum)
613		2043
614	=item ev_signal_set (ev_signal *, int signum)	2044	=item ev_signal_set (ev_signal *, int signum)
615		2045
616	Configures the watcher to trigger on the given signal number (usually one	2046	Configures the watcher to trigger on the given signal number (usually one
617	of the C<SIGxxx> constants).	2047	of the C<SIGxxx> constants).
618		2048
		2049	=item int signum [read-only]
		2050
		2051	The signal the watcher watches out for.
		2052
619	=back	2053	=back
620		2054
		2055	=head3 Examples
		2056
		2057	Example: Try to exit cleanly on SIGINT.
		2058
		2059	static void
		2060	sigint_cb (struct ev_loop *loop, ev_signal *w, int revents)
		2061	{
		2062	ev_unloop (loop, EVUNLOOP_ALL);
		2063	}
		2064
		2065	ev_signal signal_watcher;
		2066	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
		2067	ev_signal_start (loop, &signal_watcher);
		2068
		2069
621	=head2 C<ev_child> - wait for pid status changes	2070	=head2 C<ev_child> - watch out for process status changes
622		2071
623	Child watchers trigger when your process receives a SIGCHLD in response to	2072	Child watchers trigger when your process receives a SIGCHLD in response to
624	some child status changes (most typically when a child of yours dies).	2073	some child status changes (most typically when a child of yours dies or
		2074	exits). It is permissible to install a child watcher I<after> the child
		2075	has been forked (which implies it might have already exited), as long
		2076	as the event loop isn't entered (or is continued from a watcher), i.e.,
		2077	forking and then immediately registering a watcher for the child is fine,
		2078	but forking and registering a watcher a few event loop iterations later or
		2079	in the next callback invocation is not.
		2080
		2081	Only the default event loop is capable of handling signals, and therefore
		2082	you can only register child watchers in the default event loop.
		2083
		2084	Due to some design glitches inside libev, child watchers will always be
		2085	handled at maximum priority (their priority is set to C<EV_MAXPRI> by
		2086	libev)
		2087
		2088	=head3 Process Interaction
		2089
		2090	Libev grabs C<SIGCHLD> as soon as the default event loop is
		2091	initialised. This is necessary to guarantee proper behaviour even if
		2092	the first child watcher is started after the child exits. The occurrence
		2093	of C<SIGCHLD> is recorded asynchronously, but child reaping is done
		2094	synchronously as part of the event loop processing. Libev always reaps all
		2095	children, even ones not watched.
		2096
		2097	=head3 Overriding the Built-In Processing
		2098
		2099	Libev offers no special support for overriding the built-in child
		2100	processing, but if your application collides with libev's default child
		2101	handler, you can override it easily by installing your own handler for
		2102	C<SIGCHLD> after initialising the default loop, and making sure the
		2103	default loop never gets destroyed. You are encouraged, however, to use an
		2104	event-based approach to child reaping and thus use libev's support for
		2105	that, so other libev users can use C<ev_child> watchers freely.
		2106
		2107	=head3 Stopping the Child Watcher
		2108
		2109	Currently, the child watcher never gets stopped, even when the
		2110	child terminates, so normally one needs to stop the watcher in the
		2111	callback. Future versions of libev might stop the watcher automatically
		2112	when a child exit is detected.
		2113
		2114	=head3 Watcher-Specific Functions and Data Members
625		2115
626	=over 4	2116	=over 4
627		2117
628	=item ev_child_init (ev_child *, callback, int pid)	2118	=item ev_child_init (ev_child *, callback, int pid, int trace)
629		2119
630	=item ev_child_set (ev_child *, int pid)	2120	=item ev_child_set (ev_child *, int pid, int trace)
631		2121
632	Configures the watcher to wait for status changes of process C<pid> (or	2122	Configures the watcher to wait for status changes of process C<pid> (or
633	I<any> process if C<pid> is specified as C<0>). The callback can look	2123	I<any> process if C<pid> is specified as C<0>). The callback can look
634	at the C<rstatus> member of the C<ev_child> watcher structure to see	2124	at the C<rstatus> member of the C<ev_child> watcher structure to see
635	the status word (use the macros from C<sys/wait.h>). The C<rpid> member	2125	the status word (use the macros from C<sys/wait.h> and see your systems
636	contains the pid of the process causing the status change.	2126	C<waitpid> documentation). The C<rpid> member contains the pid of the
		2127	process causing the status change. C<trace> must be either C<0> (only
		2128	activate the watcher when the process terminates) or C<1> (additionally
		2129	activate the watcher when the process is stopped or continued).
		2130
		2131	=item int pid [read-only]
		2132
		2133	The process id this watcher watches out for, or C<0>, meaning any process id.
		2134
		2135	=item int rpid [read-write]
		2136
		2137	The process id that detected a status change.
		2138
		2139	=item int rstatus [read-write]
		2140
		2141	The process exit/trace status caused by C<rpid> (see your systems
		2142	C<waitpid> and C<sys/wait.h> documentation for details).
637		2143
638	=back	2144	=back
639		2145
		2146	=head3 Examples
		2147
		2148	Example: C<fork()> a new process and install a child handler to wait for
		2149	its completion.
		2150
		2151	ev_child cw;
		2152
		2153	static void
		2154	child_cb (EV_P_ ev_child *w, int revents)
		2155	{
		2156	ev_child_stop (EV_A_ w);
		2157	printf ("process %d exited with status %x\n", w->rpid, w->rstatus);
		2158	}
		2159
		2160	pid_t pid = fork ();
		2161
		2162	if (pid < 0)
		2163	// error
		2164	else if (pid == 0)
		2165	{
		2166	// the forked child executes here
		2167	exit (1);
		2168	}
		2169	else
		2170	{
		2171	ev_child_init (&cw, child_cb, pid, 0);
		2172	ev_child_start (EV_DEFAULT_ &cw);
		2173	}
		2174
		2175
		2176	=head2 C<ev_stat> - did the file attributes just change?
		2177
		2178	This watches a file system path for attribute changes. That is, it calls
		2179	C<stat> on that path in regular intervals (or when the OS says it changed)
		2180	and sees if it changed compared to the last time, invoking the callback if
		2181	it did.
		2182
		2183	The path does not need to exist: changing from "path exists" to "path does
		2184	not exist" is a status change like any other. The condition "path does not
		2185	exist" (or more correctly "path cannot be stat'ed") is signified by the
		2186	C<st_nlink> field being zero (which is otherwise always forced to be at
		2187	least one) and all the other fields of the stat buffer having unspecified
		2188	contents.
		2189
		2190	The path I<must not> end in a slash or contain special components such as
		2191	C<.> or C<..>. The path I<should> be absolute: If it is relative and
		2192	your working directory changes, then the behaviour is undefined.
		2193
		2194	Since there is no portable change notification interface available, the
		2195	portable implementation simply calls C<stat(2)> regularly on the path
		2196	to see if it changed somehow. You can specify a recommended polling
		2197	interval for this case. If you specify a polling interval of C<0> (highly
		2198	recommended!) then a I<suitable, unspecified default> value will be used
		2199	(which you can expect to be around five seconds, although this might
		2200	change dynamically). Libev will also impose a minimum interval which is
		2201	currently around C<0.1>, but that's usually overkill.
		2202
		2203	This watcher type is not meant for massive numbers of stat watchers,
		2204	as even with OS-supported change notifications, this can be
		2205	resource-intensive.
		2206
		2207	At the time of this writing, the only OS-specific interface implemented
		2208	is the Linux inotify interface (implementing kqueue support is left as an
		2209	exercise for the reader. Note, however, that the author sees no way of
		2210	implementing C<ev_stat> semantics with kqueue, except as a hint).
		2211
		2212	=head3 ABI Issues (Largefile Support)
		2213
		2214	Libev by default (unless the user overrides this) uses the default
		2215	compilation environment, which means that on systems with large file
		2216	support disabled by default, you get the 32 bit version of the stat
		2217	structure. When using the library from programs that change the ABI to
		2218	use 64 bit file offsets the programs will fail. In that case you have to
		2219	compile libev with the same flags to get binary compatibility. This is
		2220	obviously the case with any flags that change the ABI, but the problem is
		2221	most noticeably displayed with ev_stat and large file support.
		2222
		2223	The solution for this is to lobby your distribution maker to make large
		2224	file interfaces available by default (as e.g. FreeBSD does) and not
		2225	optional. Libev cannot simply switch on large file support because it has
		2226	to exchange stat structures with application programs compiled using the
		2227	default compilation environment.
		2228
		2229	=head3 Inotify and Kqueue
		2230
		2231	When C<inotify (7)> support has been compiled into libev and present at
		2232	runtime, it will be used to speed up change detection where possible. The
		2233	inotify descriptor will be created lazily when the first C<ev_stat>
		2234	watcher is being started.
		2235
		2236	Inotify presence does not change the semantics of C<ev_stat> watchers
		2237	except that changes might be detected earlier, and in some cases, to avoid
		2238	making regular C<stat> calls. Even in the presence of inotify support
		2239	there are many cases where libev has to resort to regular C<stat> polling,
		2240	but as long as kernel 2.6.25 or newer is used (2.6.24 and older have too
		2241	many bugs), the path exists (i.e. stat succeeds), and the path resides on
		2242	a local filesystem (libev currently assumes only ext2/3, jfs, reiserfs and
		2243	xfs are fully working) libev usually gets away without polling.
		2244
		2245	There is no support for kqueue, as apparently it cannot be used to
		2246	implement this functionality, due to the requirement of having a file
		2247	descriptor open on the object at all times, and detecting renames, unlinks
		2248	etc. is difficult.
		2249
		2250	=head3 C<stat ()> is a synchronous operation
		2251
		2252	Libev doesn't normally do any kind of I/O itself, and so is not blocking
		2253	the process. The exception are C<ev_stat> watchers - those call C<stat
		2254	()>, which is a synchronous operation.
		2255
		2256	For local paths, this usually doesn't matter: unless the system is very
		2257	busy or the intervals between stat's are large, a stat call will be fast,
		2258	as the path data is usually in memory already (except when starting the
		2259	watcher).
		2260
		2261	For networked file systems, calling C<stat ()> can block an indefinite
		2262	time due to network issues, and even under good conditions, a stat call
		2263	often takes multiple milliseconds.
		2264
		2265	Therefore, it is best to avoid using C<ev_stat> watchers on networked
		2266	paths, although this is fully supported by libev.
		2267
		2268	=head3 The special problem of stat time resolution
		2269
		2270	The C<stat ()> system call only supports full-second resolution portably,
		2271	and even on systems where the resolution is higher, most file systems
		2272	still only support whole seconds.
		2273
		2274	That means that, if the time is the only thing that changes, you can
		2275	easily miss updates: on the first update, C<ev_stat> detects a change and
		2276	calls your callback, which does something. When there is another update
		2277	within the same second, C<ev_stat> will be unable to detect unless the
		2278	stat data does change in other ways (e.g. file size).
		2279
		2280	The solution to this is to delay acting on a change for slightly more
		2281	than a second (or till slightly after the next full second boundary), using
		2282	a roughly one-second-delay C<ev_timer> (e.g. C<ev_timer_set (w, 0., 1.02);
		2283	ev_timer_again (loop, w)>).
		2284
		2285	The C<.02> offset is added to work around small timing inconsistencies
		2286	of some operating systems (where the second counter of the current time
		2287	might be be delayed. One such system is the Linux kernel, where a call to
		2288	C<gettimeofday> might return a timestamp with a full second later than
		2289	a subsequent C<time> call - if the equivalent of C<time ()> is used to
		2290	update file times then there will be a small window where the kernel uses
		2291	the previous second to update file times but libev might already execute
		2292	the timer callback).
		2293
		2294	=head3 Watcher-Specific Functions and Data Members
		2295
		2296	=over 4
		2297
		2298	=item ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)
		2299
		2300	=item ev_stat_set (ev_stat *, const char *path, ev_tstamp interval)
		2301
		2302	Configures the watcher to wait for status changes of the given
		2303	C<path>. The C<interval> is a hint on how quickly a change is expected to
		2304	be detected and should normally be specified as C<0> to let libev choose
		2305	a suitable value. The memory pointed to by C<path> must point to the same
		2306	path for as long as the watcher is active.
		2307
		2308	The callback will receive an C<EV_STAT> event when a change was detected,
		2309	relative to the attributes at the time the watcher was started (or the
		2310	last change was detected).
		2311
		2312	=item ev_stat_stat (loop, ev_stat *)
		2313
		2314	Updates the stat buffer immediately with new values. If you change the
		2315	watched path in your callback, you could call this function to avoid
		2316	detecting this change (while introducing a race condition if you are not
		2317	the only one changing the path). Can also be useful simply to find out the
		2318	new values.
		2319
		2320	=item ev_statdata attr [read-only]
		2321
		2322	The most-recently detected attributes of the file. Although the type is
		2323	C<ev_statdata>, this is usually the (or one of the) C<struct stat> types
		2324	suitable for your system, but you can only rely on the POSIX-standardised
		2325	members to be present. If the C<st_nlink> member is C<0>, then there was
		2326	some error while C<stat>ing the file.
		2327
		2328	=item ev_statdata prev [read-only]
		2329
		2330	The previous attributes of the file. The callback gets invoked whenever
		2331	C<prev> != C<attr>, or, more precisely, one or more of these members
		2332	differ: C<st_dev>, C<st_ino>, C<st_mode>, C<st_nlink>, C<st_uid>,
		2333	C<st_gid>, C<st_rdev>, C<st_size>, C<st_atime>, C<st_mtime>, C<st_ctime>.
		2334
		2335	=item ev_tstamp interval [read-only]
		2336
		2337	The specified interval.
		2338
		2339	=item const char *path [read-only]
		2340
		2341	The file system path that is being watched.
		2342
		2343	=back
		2344
		2345	=head3 Examples
		2346
		2347	Example: Watch C</etc/passwd> for attribute changes.
		2348
		2349	static void
		2350	passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)
		2351	{
		2352	/* /etc/passwd changed in some way */
		2353	if (w->attr.st_nlink)
		2354	{
		2355	printf ("passwd current size %ld\n", (long)w->attr.st_size);
		2356	printf ("passwd current atime %ld\n", (long)w->attr.st_mtime);
		2357	printf ("passwd current mtime %ld\n", (long)w->attr.st_mtime);
		2358	}
		2359	else
		2360	/* you shalt not abuse printf for puts */
		2361	puts ("wow, /etc/passwd is not there, expect problems. "
		2362	"if this is windows, they already arrived\n");
		2363	}
		2364
		2365	...
		2366	ev_stat passwd;
		2367
		2368	ev_stat_init (&passwd, passwd_cb, "/etc/passwd", 0.);
		2369	ev_stat_start (loop, &passwd);
		2370
		2371	Example: Like above, but additionally use a one-second delay so we do not
		2372	miss updates (however, frequent updates will delay processing, too, so
		2373	one might do the work both on C<ev_stat> callback invocation I<and> on
		2374	C<ev_timer> callback invocation).
		2375
		2376	static ev_stat passwd;
		2377	static ev_timer timer;
		2378
		2379	static void
		2380	timer_cb (EV_P_ ev_timer *w, int revents)
		2381	{
		2382	ev_timer_stop (EV_A_ w);
		2383
		2384	/* now it's one second after the most recent passwd change */
		2385	}
		2386
		2387	static void
		2388	stat_cb (EV_P_ ev_stat *w, int revents)
		2389	{
		2390	/* reset the one-second timer */
		2391	ev_timer_again (EV_A_ &timer);
		2392	}
		2393
		2394	...
		2395	ev_stat_init (&passwd, stat_cb, "/etc/passwd", 0.);
		2396	ev_stat_start (loop, &passwd);
		2397	ev_timer_init (&timer, timer_cb, 0., 1.02);
		2398
		2399
640	=head2 C<ev_idle> - when you've got nothing better to do	2400	=head2 C<ev_idle> - when you've got nothing better to do...
641		2401
642	Idle watchers trigger events when there are no other I/O or timer (or	2402	Idle watchers trigger events when no other events of the same or higher
643	periodic) events pending. That is, as long as your process is busy	2403	priority are pending (prepare, check and other idle watchers do not count
644	handling sockets or timeouts it will not be called. But when your process	2404	as receiving "events").
645	is idle all idle watchers are being called again and again - until	2405
		2406	That is, as long as your process is busy handling sockets or timeouts
		2407	(or even signals, imagine) of the same or higher priority it will not be
		2408	triggered. But when your process is idle (or only lower-priority watchers
		2409	are pending), the idle watchers are being called once per event loop
646	stopped, that is, or your process receives more events.	2410	iteration - until stopped, that is, or your process receives more events
		2411	and becomes busy again with higher priority stuff.
647		2412
648	The most noteworthy effect is that as long as any idle watchers are	2413	The most noteworthy effect is that as long as any idle watchers are
649	active, the process will not block when waiting for new events.	2414	active, the process will not block when waiting for new events.
650		2415
651	Apart from keeping your process non-blocking (which is a useful	2416	Apart from keeping your process non-blocking (which is a useful
652	effect on its own sometimes), idle watchers are a good place to do	2417	effect on its own sometimes), idle watchers are a good place to do
653	"pseudo-background processing", or delay processing stuff to after the	2418	"pseudo-background processing", or delay processing stuff to after the
654	event loop has handled all outstanding events.	2419	event loop has handled all outstanding events.
655		2420
		2421	=head3 Watcher-Specific Functions and Data Members
		2422
656	=over 4	2423	=over 4
657		2424
658	=item ev_idle_init (ev_signal *, callback)	2425	=item ev_idle_init (ev_idle *, callback)
659		2426
660	Initialises and configures the idle watcher - it has no parameters of any	2427	Initialises and configures the idle watcher - it has no parameters of any
661	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,	2428	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,
662	believe me.	2429	believe me.
663		2430
664	=back	2431	=back
665		2432
666	=head2 prepare and check - your hooks into the event loop	2433	=head3 Examples
667		2434
		2435	Example: Dynamically allocate an C<ev_idle> watcher, start it, and in the
		2436	callback, free it. Also, use no error checking, as usual.
		2437
		2438	static void
		2439	idle_cb (struct ev_loop *loop, ev_idle *w, int revents)
		2440	{
		2441	free (w);
		2442	// now do something you wanted to do when the program has
		2443	// no longer anything immediate to do.
		2444	}
		2445
		2446	ev_idle *idle_watcher = malloc (sizeof (ev_idle));
		2447	ev_idle_init (idle_watcher, idle_cb);
		2448	ev_idle_start (loop, idle_watcher);
		2449
		2450
		2451	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!
		2452
668	Prepare and check watchers usually (but not always) are used in	2453	Prepare and check watchers are usually (but not always) used in pairs:
669	tandom. Prepare watchers get invoked before the process blocks and check	2454	prepare watchers get invoked before the process blocks and check watchers
670	watchers afterwards.	2455	afterwards.
671		2456
		2457	You I<must not> call C<ev_loop> or similar functions that enter
		2458	the current event loop from either C<ev_prepare> or C<ev_check>
		2459	watchers. Other loops than the current one are fine, however. The
		2460	rationale behind this is that you do not need to check for recursion in
		2461	those watchers, i.e. the sequence will always be C<ev_prepare>, blocking,
		2462	C<ev_check> so if you have one watcher of each kind they will always be
		2463	called in pairs bracketing the blocking call.
		2464
672	Their main purpose is to integrate other event mechanisms into libev. This	2465	Their main purpose is to integrate other event mechanisms into libev and
673	could be used, for example, to track variable changes, implement your own	2466	their use is somewhat advanced. They could be used, for example, to track
674	watchers, integrate net-snmp or a coroutine library and lots more.	2467	variable changes, implement your own watchers, integrate net-snmp or a
		2468	coroutine library and lots more. They are also occasionally useful if
		2469	you cache some data and want to flush it before blocking (for example,
		2470	in X programs you might want to do an C<XFlush ()> in an C<ev_prepare>
		2471	watcher).
675		2472
676	This is done by examining in each prepare call which file descriptors need	2473	This is done by examining in each prepare call which file descriptors
677	to be watched by the other library, registering C<ev_io> watchers for them	2474	need to be watched by the other library, registering C<ev_io> watchers
678	and starting an C<ev_timer> watcher for any timeouts (many libraries provide	2475	for them and starting an C<ev_timer> watcher for any timeouts (many
679	just this functionality). Then, in the check watcher you check for any	2476	libraries provide exactly this functionality). Then, in the check watcher,
680	events that occured (by making your callbacks set soem flags for example)	2477	you check for any events that occurred (by checking the pending status
681	and call back into the library.	2478	of all watchers and stopping them) and call back into the library. The
		2479	I/O and timer callbacks will never actually be called (but must be valid
		2480	nevertheless, because you never know, you know?).
682		2481
683	As another example, the perl Coro module uses these hooks to integrate	2482	As another example, the Perl Coro module uses these hooks to integrate
684	coroutines into libev programs, by yielding to other active coroutines	2483	coroutines into libev programs, by yielding to other active coroutines
685	during each prepare and only letting the process block if no coroutines	2484	during each prepare and only letting the process block if no coroutines
686	are ready to run.	2485	are ready to run (it's actually more complicated: it only runs coroutines
		2486	with priority higher than or equal to the event loop and one coroutine
		2487	of lower priority, but only once, using idle watchers to keep the event
		2488	loop from blocking if lower-priority coroutines are active, thus mapping
		2489	low-priority coroutines to idle/background tasks).
		2490
		2491	It is recommended to give C<ev_check> watchers highest (C<EV_MAXPRI>)
		2492	priority, to ensure that they are being run before any other watchers
		2493	after the poll (this doesn't matter for C<ev_prepare> watchers).
		2494
		2495	Also, C<ev_check> watchers (and C<ev_prepare> watchers, too) should not
		2496	activate ("feed") events into libev. While libev fully supports this, they
		2497	might get executed before other C<ev_check> watchers did their job. As
		2498	C<ev_check> watchers are often used to embed other (non-libev) event
		2499	loops those other event loops might be in an unusable state until their
		2500	C<ev_check> watcher ran (always remind yourself to coexist peacefully with
		2501	others).
		2502
		2503	=head3 Watcher-Specific Functions and Data Members
687		2504
688	=over 4	2505	=over 4
689		2506
690	=item ev_prepare_init (ev_prepare *, callback)	2507	=item ev_prepare_init (ev_prepare *, callback)
691		2508
692	=item ev_check_init (ev_check *, callback)	2509	=item ev_check_init (ev_check *, callback)
693		2510
694	Initialises and configures the prepare or check watcher - they have no	2511	Initialises and configures the prepare or check watcher - they have no
695	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>	2512	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>
696	macros, but using them is utterly, utterly pointless.	2513	macros, but using them is utterly, utterly, utterly and completely
		2514	pointless.
697		2515
698	=back	2516	=back
699		2517
		2518	=head3 Examples
		2519
		2520	There are a number of principal ways to embed other event loops or modules
		2521	into libev. Here are some ideas on how to include libadns into libev
		2522	(there is a Perl module named C<EV::ADNS> that does this, which you could
		2523	use as a working example. Another Perl module named C<EV::Glib> embeds a
		2524	Glib main context into libev, and finally, C<Glib::EV> embeds EV into the
		2525	Glib event loop).
		2526
		2527	Method 1: Add IO watchers and a timeout watcher in a prepare handler,
		2528	and in a check watcher, destroy them and call into libadns. What follows
		2529	is pseudo-code only of course. This requires you to either use a low
		2530	priority for the check watcher or use C<ev_clear_pending> explicitly, as
		2531	the callbacks for the IO/timeout watchers might not have been called yet.
		2532
		2533	static ev_io iow [nfd];
		2534	static ev_timer tw;
		2535
		2536	static void
		2537	io_cb (struct ev_loop *loop, ev_io *w, int revents)
		2538	{
		2539	}
		2540
		2541	// create io watchers for each fd and a timer before blocking
		2542	static void
		2543	adns_prepare_cb (struct ev_loop *loop, ev_prepare *w, int revents)
		2544	{
		2545	int timeout = 3600000;
		2546	struct pollfd fds [nfd];
		2547	// actual code will need to loop here and realloc etc.
		2548	adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));
		2549
		2550	/* the callback is illegal, but won't be called as we stop during check */
		2551	ev_timer_init (&tw, 0, timeout * 1e-3, 0.);
		2552	ev_timer_start (loop, &tw);
		2553
		2554	// create one ev_io per pollfd
		2555	for (int i = 0; i < nfd; ++i)
		2556	{
		2557	ev_io_init (iow + i, io_cb, fds [i].fd,
		2558	((fds [i].events & POLLIN ? EV_READ : 0)
		2559	| (fds [i].events & POLLOUT ? EV_WRITE : 0)));
		2560
		2561	fds [i].revents = 0;
		2562	ev_io_start (loop, iow + i);
		2563	}
		2564	}
		2565
		2566	// stop all watchers after blocking
		2567	static void
		2568	adns_check_cb (struct ev_loop *loop, ev_check *w, int revents)
		2569	{
		2570	ev_timer_stop (loop, &tw);
		2571
		2572	for (int i = 0; i < nfd; ++i)
		2573	{
		2574	// set the relevant poll flags
		2575	// could also call adns_processreadable etc. here
		2576	struct pollfd *fd = fds + i;
		2577	int revents = ev_clear_pending (iow + i);
		2578	if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;
		2579	if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;
		2580
		2581	// now stop the watcher
		2582	ev_io_stop (loop, iow + i);
		2583	}
		2584
		2585	adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));
		2586	}
		2587
		2588	Method 2: This would be just like method 1, but you run C<adns_afterpoll>
		2589	in the prepare watcher and would dispose of the check watcher.
		2590
		2591	Method 3: If the module to be embedded supports explicit event
		2592	notification (libadns does), you can also make use of the actual watcher
		2593	callbacks, and only destroy/create the watchers in the prepare watcher.
		2594
		2595	static void
		2596	timer_cb (EV_P_ ev_timer *w, int revents)
		2597	{
		2598	adns_state ads = (adns_state)w->data;
		2599	update_now (EV_A);
		2600
		2601	adns_processtimeouts (ads, &tv_now);
		2602	}
		2603
		2604	static void
		2605	io_cb (EV_P_ ev_io *w, int revents)
		2606	{
		2607	adns_state ads = (adns_state)w->data;
		2608	update_now (EV_A);
		2609
		2610	if (revents & EV_READ ) adns_processreadable (ads, w->fd, &tv_now);
		2611	if (revents & EV_WRITE) adns_processwriteable (ads, w->fd, &tv_now);
		2612	}
		2613
		2614	// do not ever call adns_afterpoll
		2615
		2616	Method 4: Do not use a prepare or check watcher because the module you
		2617	want to embed is not flexible enough to support it. Instead, you can
		2618	override their poll function. The drawback with this solution is that the
		2619	main loop is now no longer controllable by EV. The C<Glib::EV> module uses
		2620	this approach, effectively embedding EV as a client into the horrible
		2621	libglib event loop.
		2622
		2623	static gint
		2624	event_poll_func (GPollFD *fds, guint nfds, gint timeout)
		2625	{
		2626	int got_events = 0;
		2627
		2628	for (n = 0; n < nfds; ++n)
		2629	// create/start io watcher that sets the relevant bits in fds[n] and increment got_events
		2630
		2631	if (timeout >= 0)
		2632	// create/start timer
		2633
		2634	// poll
		2635	ev_loop (EV_A_ 0);
		2636
		2637	// stop timer again
		2638	if (timeout >= 0)
		2639	ev_timer_stop (EV_A_ &to);
		2640
		2641	// stop io watchers again - their callbacks should have set
		2642	for (n = 0; n < nfds; ++n)
		2643	ev_io_stop (EV_A_ iow [n]);
		2644
		2645	return got_events;
		2646	}
		2647
		2648
		2649	=head2 C<ev_embed> - when one backend isn't enough...
		2650
		2651	This is a rather advanced watcher type that lets you embed one event loop
		2652	into another (currently only C<ev_io> events are supported in the embedded
		2653	loop, other types of watchers might be handled in a delayed or incorrect
		2654	fashion and must not be used).
		2655
		2656	There are primarily two reasons you would want that: work around bugs and
		2657	prioritise I/O.
		2658
		2659	As an example for a bug workaround, the kqueue backend might only support
		2660	sockets on some platform, so it is unusable as generic backend, but you
		2661	still want to make use of it because you have many sockets and it scales
		2662	so nicely. In this case, you would create a kqueue-based loop and embed
		2663	it into your default loop (which might use e.g. poll). Overall operation
		2664	will be a bit slower because first libev has to call C<poll> and then
		2665	C<kevent>, but at least you can use both mechanisms for what they are
		2666	best: C<kqueue> for scalable sockets and C<poll> if you want it to work :)
		2667
		2668	As for prioritising I/O: under rare circumstances you have the case where
		2669	some fds have to be watched and handled very quickly (with low latency),
		2670	and even priorities and idle watchers might have too much overhead. In
		2671	this case you would put all the high priority stuff in one loop and all
		2672	the rest in a second one, and embed the second one in the first.
		2673
		2674	As long as the watcher is active, the callback will be invoked every
		2675	time there might be events pending in the embedded loop. The callback
		2676	must then call C<ev_embed_sweep (mainloop, watcher)> to make a single
		2677	sweep and invoke their callbacks (the callback doesn't need to invoke the
		2678	C<ev_embed_sweep> function directly, it could also start an idle watcher
		2679	to give the embedded loop strictly lower priority for example).
		2680
		2681	You can also set the callback to C<0>, in which case the embed watcher
		2682	will automatically execute the embedded loop sweep whenever necessary.
		2683
		2684	Fork detection will be handled transparently while the C<ev_embed> watcher
		2685	is active, i.e., the embedded loop will automatically be forked when the
		2686	embedding loop forks. In other cases, the user is responsible for calling
		2687	C<ev_loop_fork> on the embedded loop.
		2688
		2689	Unfortunately, not all backends are embeddable: only the ones returned by
		2690	C<ev_embeddable_backends> are, which, unfortunately, does not include any
		2691	portable one.
		2692
		2693	So when you want to use this feature you will always have to be prepared
		2694	that you cannot get an embeddable loop. The recommended way to get around
		2695	this is to have a separate variables for your embeddable loop, try to
		2696	create it, and if that fails, use the normal loop for everything.
		2697
		2698	=head3 C<ev_embed> and fork
		2699
		2700	While the C<ev_embed> watcher is running, forks in the embedding loop will
		2701	automatically be applied to the embedded loop as well, so no special
		2702	fork handling is required in that case. When the watcher is not running,
		2703	however, it is still the task of the libev user to call C<ev_loop_fork ()>
		2704	as applicable.
		2705
		2706	=head3 Watcher-Specific Functions and Data Members
		2707
		2708	=over 4
		2709
		2710	=item ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)
		2711
		2712	=item ev_embed_set (ev_embed *, callback, struct ev_loop *embedded_loop)
		2713
		2714	Configures the watcher to embed the given loop, which must be
		2715	embeddable. If the callback is C<0>, then C<ev_embed_sweep> will be
		2716	invoked automatically, otherwise it is the responsibility of the callback
		2717	to invoke it (it will continue to be called until the sweep has been done,
		2718	if you do not want that, you need to temporarily stop the embed watcher).
		2719
		2720	=item ev_embed_sweep (loop, ev_embed *)
		2721
		2722	Make a single, non-blocking sweep over the embedded loop. This works
		2723	similarly to C<ev_loop (embedded_loop, EVLOOP_NONBLOCK)>, but in the most
		2724	appropriate way for embedded loops.
		2725
		2726	=item struct ev_loop *other [read-only]
		2727
		2728	The embedded event loop.
		2729
		2730	=back
		2731
		2732	=head3 Examples
		2733
		2734	Example: Try to get an embeddable event loop and embed it into the default
		2735	event loop. If that is not possible, use the default loop. The default
		2736	loop is stored in C<loop_hi>, while the embeddable loop is stored in
		2737	C<loop_lo> (which is C<loop_hi> in the case no embeddable loop can be
		2738	used).
		2739
		2740	struct ev_loop *loop_hi = ev_default_init (0);
		2741	struct ev_loop *loop_lo = 0;
		2742	ev_embed embed;
		2743
		2744	// see if there is a chance of getting one that works
		2745	// (remember that a flags value of 0 means autodetection)
		2746	loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
		2747	? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
		2748	: 0;
		2749
		2750	// if we got one, then embed it, otherwise default to loop_hi
		2751	if (loop_lo)
		2752	{
		2753	ev_embed_init (&embed, 0, loop_lo);
		2754	ev_embed_start (loop_hi, &embed);
		2755	}
		2756	else
		2757	loop_lo = loop_hi;
		2758
		2759	Example: Check if kqueue is available but not recommended and create
		2760	a kqueue backend for use with sockets (which usually work with any
		2761	kqueue implementation). Store the kqueue/socket-only event loop in
		2762	C<loop_socket>. (One might optionally use C<EVFLAG_NOENV>, too).
		2763
		2764	struct ev_loop *loop = ev_default_init (0);
		2765	struct ev_loop *loop_socket = 0;
		2766	ev_embed embed;
		2767
		2768	if (ev_supported_backends () & ~ev_recommended_backends () & EVBACKEND_KQUEUE)
		2769	if ((loop_socket = ev_loop_new (EVBACKEND_KQUEUE))
		2770	{
		2771	ev_embed_init (&embed, 0, loop_socket);
		2772	ev_embed_start (loop, &embed);
		2773	}
		2774
		2775	if (!loop_socket)
		2776	loop_socket = loop;
		2777
		2778	// now use loop_socket for all sockets, and loop for everything else
		2779
		2780
		2781	=head2 C<ev_fork> - the audacity to resume the event loop after a fork
		2782
		2783	Fork watchers are called when a C<fork ()> was detected (usually because
		2784	whoever is a good citizen cared to tell libev about it by calling
		2785	C<ev_default_fork> or C<ev_loop_fork>). The invocation is done before the
		2786	event loop blocks next and before C<ev_check> watchers are being called,
		2787	and only in the child after the fork. If whoever good citizen calling
		2788	C<ev_default_fork> cheats and calls it in the wrong process, the fork
		2789	handlers will be invoked, too, of course.
		2790
		2791	=head3 The special problem of life after fork - how is it possible?
		2792
		2793	Most uses of C<fork()> consist of forking, then some simple calls to ste
		2794	up/change the process environment, followed by a call to C<exec()>. This
		2795	sequence should be handled by libev without any problems.
		2796
		2797	This changes when the application actually wants to do event handling
		2798	in the child, or both parent in child, in effect "continuing" after the
		2799	fork.
		2800
		2801	The default mode of operation (for libev, with application help to detect
		2802	forks) is to duplicate all the state in the child, as would be expected
		2803	when I<either> the parent I<or> the child process continues.
		2804
		2805	When both processes want to continue using libev, then this is usually the
		2806	wrong result. In that case, usually one process (typically the parent) is
		2807	supposed to continue with all watchers in place as before, while the other
		2808	process typically wants to start fresh, i.e. without any active watchers.
		2809
		2810	The cleanest and most efficient way to achieve that with libev is to
		2811	simply create a new event loop, which of course will be "empty", and
		2812	use that for new watchers. This has the advantage of not touching more
		2813	memory than necessary, and thus avoiding the copy-on-write, and the
		2814	disadvantage of having to use multiple event loops (which do not support
		2815	signal watchers).
		2816
		2817	When this is not possible, or you want to use the default loop for
		2818	other reasons, then in the process that wants to start "fresh", call
		2819	C<ev_default_destroy ()> followed by C<ev_default_loop (...)>. Destroying
		2820	the default loop will "orphan" (not stop) all registered watchers, so you
		2821	have to be careful not to execute code that modifies those watchers. Note
		2822	also that in that case, you have to re-register any signal watchers.
		2823
		2824	=head3 Watcher-Specific Functions and Data Members
		2825
		2826	=over 4
		2827
		2828	=item ev_fork_init (ev_signal *, callback)
		2829
		2830	Initialises and configures the fork watcher - it has no parameters of any
		2831	kind. There is a C<ev_fork_set> macro, but using it is utterly pointless,
		2832	believe me.
		2833
		2834	=back
		2835
		2836
		2837	=head2 C<ev_async> - how to wake up another event loop
		2838
		2839	In general, you cannot use an C<ev_loop> from multiple threads or other
		2840	asynchronous sources such as signal handlers (as opposed to multiple event
		2841	loops - those are of course safe to use in different threads).
		2842
		2843	Sometimes, however, you need to wake up another event loop you do not
		2844	control, for example because it belongs to another thread. This is what
		2845	C<ev_async> watchers do: as long as the C<ev_async> watcher is active, you
		2846	can signal it by calling C<ev_async_send>, which is thread- and signal
		2847	safe.
		2848
		2849	This functionality is very similar to C<ev_signal> watchers, as signals,
		2850	too, are asynchronous in nature, and signals, too, will be compressed
		2851	(i.e. the number of callback invocations may be less than the number of
		2852	C<ev_async_sent> calls).
		2853
		2854	Unlike C<ev_signal> watchers, C<ev_async> works with any event loop, not
		2855	just the default loop.
		2856
		2857	=head3 Queueing
		2858
		2859	C<ev_async> does not support queueing of data in any way. The reason
		2860	is that the author does not know of a simple (or any) algorithm for a
		2861	multiple-writer-single-reader queue that works in all cases and doesn't
		2862	need elaborate support such as pthreads.
		2863
		2864	That means that if you want to queue data, you have to provide your own
		2865	queue. But at least I can tell you how to implement locking around your
		2866	queue:
		2867
		2868	=over 4
		2869
		2870	=item queueing from a signal handler context
		2871
		2872	To implement race-free queueing, you simply add to the queue in the signal
		2873	handler but you block the signal handler in the watcher callback. Here is
		2874	an example that does that for some fictitious SIGUSR1 handler:
		2875
		2876	static ev_async mysig;
		2877
		2878	static void
		2879	sigusr1_handler (void)
		2880	{
		2881	sometype data;
		2882
		2883	// no locking etc.
		2884	queue_put (data);
		2885	ev_async_send (EV_DEFAULT_ &mysig);
		2886	}
		2887
		2888	static void
		2889	mysig_cb (EV_P_ ev_async *w, int revents)
		2890	{
		2891	sometype data;
		2892	sigset_t block, prev;
		2893
		2894	sigemptyset (&block);
		2895	sigaddset (&block, SIGUSR1);
		2896	sigprocmask (SIG_BLOCK, &block, &prev);
		2897
		2898	while (queue_get (&data))
		2899	process (data);
		2900
		2901	if (sigismember (&prev, SIGUSR1)
		2902	sigprocmask (SIG_UNBLOCK, &block, 0);
		2903	}
		2904
		2905	(Note: pthreads in theory requires you to use C<pthread_setmask>
		2906	instead of C<sigprocmask> when you use threads, but libev doesn't do it
		2907	either...).
		2908
		2909	=item queueing from a thread context
		2910
		2911	The strategy for threads is different, as you cannot (easily) block
		2912	threads but you can easily preempt them, so to queue safely you need to
		2913	employ a traditional mutex lock, such as in this pthread example:
		2914
		2915	static ev_async mysig;
		2916	static pthread_mutex_t mymutex = PTHREAD_MUTEX_INITIALIZER;
		2917
		2918	static void
		2919	otherthread (void)
		2920	{
		2921	// only need to lock the actual queueing operation
		2922	pthread_mutex_lock (&mymutex);
		2923	queue_put (data);
		2924	pthread_mutex_unlock (&mymutex);
		2925
		2926	ev_async_send (EV_DEFAULT_ &mysig);
		2927	}
		2928
		2929	static void
		2930	mysig_cb (EV_P_ ev_async *w, int revents)
		2931	{
		2932	pthread_mutex_lock (&mymutex);
		2933
		2934	while (queue_get (&data))
		2935	process (data);
		2936
		2937	pthread_mutex_unlock (&mymutex);
		2938	}
		2939
		2940	=back
		2941
		2942
		2943	=head3 Watcher-Specific Functions and Data Members
		2944
		2945	=over 4
		2946
		2947	=item ev_async_init (ev_async *, callback)
		2948
		2949	Initialises and configures the async watcher - it has no parameters of any
		2950	kind. There is a C<ev_async_set> macro, but using it is utterly pointless,
		2951	trust me.
		2952
		2953	=item ev_async_send (loop, ev_async *)
		2954
		2955	Sends/signals/activates the given C<ev_async> watcher, that is, feeds
		2956	an C<EV_ASYNC> event on the watcher into the event loop. Unlike
		2957	C<ev_feed_event>, this call is safe to do from other threads, signal or
		2958	similar contexts (see the discussion of C<EV_ATOMIC_T> in the embedding
		2959	section below on what exactly this means).
		2960
		2961	Note that, as with other watchers in libev, multiple events might get
		2962	compressed into a single callback invocation (another way to look at this
		2963	is that C<ev_async> watchers are level-triggered, set on C<ev_async_send>,
		2964	reset when the event loop detects that).
		2965
		2966	This call incurs the overhead of a system call only once per event loop
		2967	iteration, so while the overhead might be noticeable, it doesn't apply to
		2968	repeated calls to C<ev_async_send> for the same event loop.
		2969
		2970	=item bool = ev_async_pending (ev_async *)
		2971
		2972	Returns a non-zero value when C<ev_async_send> has been called on the
		2973	watcher but the event has not yet been processed (or even noted) by the
		2974	event loop.
		2975
		2976	C<ev_async_send> sets a flag in the watcher and wakes up the loop. When
		2977	the loop iterates next and checks for the watcher to have become active,
		2978	it will reset the flag again. C<ev_async_pending> can be used to very
		2979	quickly check whether invoking the loop might be a good idea.
		2980
		2981	Not that this does I<not> check whether the watcher itself is pending,
		2982	only whether it has been requested to make this watcher pending: there
		2983	is a time window between the event loop checking and resetting the async
		2984	notification, and the callback being invoked.
		2985
		2986	=back
		2987
		2988
700	=head1 OTHER FUNCTIONS	2989	=head1 OTHER FUNCTIONS
701		2990
702	There are some other fucntions of possible interest. Described. Here. Now.	2991	There are some other functions of possible interest. Described. Here. Now.
703		2992
704	=over 4	2993	=over 4
705		2994
706	=item ev_once (loop, int fd, int events, ev_tstamp timeout, callback)	2995	=item ev_once (loop, int fd, int events, ev_tstamp timeout, callback)
707		2996
708	This function combines a simple timer and an I/O watcher, calls your	2997	This function combines a simple timer and an I/O watcher, calls your
709	callback on whichever event happens first and automatically stop both	2998	callback on whichever event happens first and automatically stops both
710	watchers. This is useful if you want to wait for a single event on an fd	2999	watchers. This is useful if you want to wait for a single event on an fd
711	or timeout without havign to allocate/configure/start/stop/free one or	3000	or timeout without having to allocate/configure/start/stop/free one or
712	more watchers yourself.	3001	more watchers yourself.
713		3002
714	If C<fd> is less than 0, then no I/O watcher will be started and events is	3003	If C<fd> is less than 0, then no I/O watcher will be started and the
715	ignored. Otherwise, an C<ev_io> watcher for the given C<fd> and C<events> set	3004	C<events> argument is being ignored. Otherwise, an C<ev_io> watcher for
716	will be craeted and started.	3005	the given C<fd> and C<events> set will be created and started.
717		3006
718	If C<timeout> is less than 0, then no timeout watcher will be	3007	If C<timeout> is less than 0, then no timeout watcher will be
719	started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and repeat	3008	started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and
720	= 0) will be started.	3009	repeat = 0) will be started. C<0> is a valid timeout.
721		3010
722	The callback has the type C<void (*cb)(int revents, void *arg)> and	3011	The callback has the type C<void (*cb)(int revents, void *arg)> and gets
723	gets passed an events set (normally a combination of C<EV_ERROR>, C<EV_READ>,	3012	passed an C<revents> set like normal event callbacks (a combination of
724	C<EV_WRITE> or C<EV_TIMEOUT>) and the C<arg> value passed to C<ev_once>:	3013	C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMEOUT>) and the C<arg>
		3014	value passed to C<ev_once>. Note that it is possible to receive I<both>
		3015	a timeout and an io event at the same time - you probably should give io
		3016	events precedence.
725		3017
		3018	Example: wait up to ten seconds for data to appear on STDIN_FILENO.
		3019
726	static void stdin_ready (int revents, void *arg)	3020	static void stdin_ready (int revents, void *arg)
727	{	3021	{
728	if (revents & EV_TIMEOUT)
729	/* doh, nothing entered */
730	else if (revents & EV_READ)	3022	if (revents & EV_READ)
731	/* stdin might have data for us, joy! */	3023	/* stdin might have data for us, joy! */;
		3024	else if (revents & EV_TIMEOUT)
		3025	/* doh, nothing entered */;
732	}	3026	}
733		3027
734	ev_once (STDIN_FILENO, EV_READm 10., stdin_ready, 0);	3028	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
735		3029
736	=item ev_feed_event (loop, watcher, int events)	3030	=item ev_feed_event (struct ev_loop *, watcher *, int revents)
737		3031
738	Feeds the given event set into the event loop, as if the specified event	3032	Feeds the given event set into the event loop, as if the specified event
739	has happened for the specified watcher (which must be a pointer to an	3033	had happened for the specified watcher (which must be a pointer to an
740	initialised but not necessarily active event watcher).	3034	initialised but not necessarily started event watcher).
741		3035
742	=item ev_feed_fd_event (loop, int fd, int revents)	3036	=item ev_feed_fd_event (struct ev_loop *, int fd, int revents)
743		3037
744	Feed an event on the given fd, as if a file descriptor backend detected it.	3038	Feed an event on the given fd, as if a file descriptor backend detected
		3039	the given events it.
745		3040
746	=item ev_feed_signal_event (loop, int signum)	3041	=item ev_feed_signal_event (struct ev_loop *loop, int signum)
747		3042
748	Feed an event as if the given signal occured (loop must be the default loop!).	3043	Feed an event as if the given signal occurred (C<loop> must be the default
		3044	loop!).
749		3045
750	=back	3046	=back
751		3047
		3048
		3049	=head1 LIBEVENT EMULATION
		3050
		3051	Libev offers a compatibility emulation layer for libevent. It cannot
		3052	emulate the internals of libevent, so here are some usage hints:
		3053
		3054	=over 4
		3055
		3056	=item * Use it by including <event.h>, as usual.
		3057
		3058	=item * The following members are fully supported: ev_base, ev_callback,
		3059	ev_arg, ev_fd, ev_res, ev_events.
		3060
		3061	=item * Avoid using ev_flags and the EVLIST_*-macros, while it is
		3062	maintained by libev, it does not work exactly the same way as in libevent (consider
		3063	it a private API).
		3064
		3065	=item * Priorities are not currently supported. Initialising priorities
		3066	will fail and all watchers will have the same priority, even though there
		3067	is an ev_pri field.
		3068
		3069	=item * In libevent, the last base created gets the signals, in libev, the
		3070	first base created (== the default loop) gets the signals.
		3071
		3072	=item * Other members are not supported.
		3073
		3074	=item * The libev emulation is I<not> ABI compatible to libevent, you need
		3075	to use the libev header file and library.
		3076
		3077	=back
		3078
		3079	=head1 C++ SUPPORT
		3080
		3081	Libev comes with some simplistic wrapper classes for C++ that mainly allow
		3082	you to use some convenience methods to start/stop watchers and also change
		3083	the callback model to a model using method callbacks on objects.
		3084
		3085	To use it,
		3086
		3087	#include <ev++.h>
		3088
		3089	This automatically includes F<ev.h> and puts all of its definitions (many
		3090	of them macros) into the global namespace. All C++ specific things are
		3091	put into the C<ev> namespace. It should support all the same embedding
		3092	options as F<ev.h>, most notably C<EV_MULTIPLICITY>.
		3093
		3094	Care has been taken to keep the overhead low. The only data member the C++
		3095	classes add (compared to plain C-style watchers) is the event loop pointer
		3096	that the watcher is associated with (or no additional members at all if
		3097	you disable C<EV_MULTIPLICITY> when embedding libev).
		3098
		3099	Currently, functions, and static and non-static member functions can be
		3100	used as callbacks. Other types should be easy to add as long as they only
		3101	need one additional pointer for context. If you need support for other
		3102	types of functors please contact the author (preferably after implementing
		3103	it).
		3104
		3105	Here is a list of things available in the C<ev> namespace:
		3106
		3107	=over 4
		3108
		3109	=item C<ev::READ>, C<ev::WRITE> etc.
		3110
		3111	These are just enum values with the same values as the C<EV_READ> etc.
		3112	macros from F<ev.h>.
		3113
		3114	=item C<ev::tstamp>, C<ev::now>
		3115
		3116	Aliases to the same types/functions as with the C<ev_> prefix.
		3117
		3118	=item C<ev::io>, C<ev::timer>, C<ev::periodic>, C<ev::idle>, C<ev::sig> etc.
		3119
		3120	For each C<ev_TYPE> watcher in F<ev.h> there is a corresponding class of
		3121	the same name in the C<ev> namespace, with the exception of C<ev_signal>
		3122	which is called C<ev::sig> to avoid clashes with the C<signal> macro
		3123	defines by many implementations.
		3124
		3125	All of those classes have these methods:
		3126
		3127	=over 4
		3128
		3129	=item ev::TYPE::TYPE ()
		3130
		3131	=item ev::TYPE::TYPE (struct ev_loop *)
		3132
		3133	=item ev::TYPE::~TYPE
		3134
		3135	The constructor (optionally) takes an event loop to associate the watcher
		3136	with. If it is omitted, it will use C<EV_DEFAULT>.
		3137
		3138	The constructor calls C<ev_init> for you, which means you have to call the
		3139	C<set> method before starting it.
		3140
		3141	It will not set a callback, however: You have to call the templated C<set>
		3142	method to set a callback before you can start the watcher.
		3143
		3144	(The reason why you have to use a method is a limitation in C++ which does
		3145	not allow explicit template arguments for constructors).
		3146
		3147	The destructor automatically stops the watcher if it is active.
		3148
		3149	=item w->set<class, &class::method> (object *)
		3150
		3151	This method sets the callback method to call. The method has to have a
		3152	signature of C<void (*)(ev_TYPE &, int)>, it receives the watcher as
		3153	first argument and the C<revents> as second. The object must be given as
		3154	parameter and is stored in the C<data> member of the watcher.
		3155
		3156	This method synthesizes efficient thunking code to call your method from
		3157	the C callback that libev requires. If your compiler can inline your
		3158	callback (i.e. it is visible to it at the place of the C<set> call and
		3159	your compiler is good :), then the method will be fully inlined into the
		3160	thunking function, making it as fast as a direct C callback.
		3161
		3162	Example: simple class declaration and watcher initialisation
		3163
		3164	struct myclass
		3165	{
		3166	void io_cb (ev::io &w, int revents) { }
		3167	}
		3168
		3169	myclass obj;
		3170	ev::io iow;
		3171	iow.set <myclass, &myclass::io_cb> (&obj);
		3172
		3173	=item w->set (object *)
		3174
		3175	This is an B<experimental> feature that might go away in a future version.
		3176
		3177	This is a variation of a method callback - leaving out the method to call
		3178	will default the method to C<operator ()>, which makes it possible to use
		3179	functor objects without having to manually specify the C<operator ()> all
		3180	the time. Incidentally, you can then also leave out the template argument
		3181	list.
		3182
		3183	The C<operator ()> method prototype must be C<void operator ()(watcher &w,
		3184	int revents)>.
		3185
		3186	See the method-C<set> above for more details.
		3187
		3188	Example: use a functor object as callback.
		3189
		3190	struct myfunctor
		3191	{
		3192	void operator() (ev::io &w, int revents)
		3193	{
		3194	...
		3195	}
		3196	}
		3197
		3198	myfunctor f;
		3199
		3200	ev::io w;
		3201	w.set (&f);
		3202
		3203	=item w->set<function> (void *data = 0)
		3204
		3205	Also sets a callback, but uses a static method or plain function as
		3206	callback. The optional C<data> argument will be stored in the watcher's
		3207	C<data> member and is free for you to use.
		3208
		3209	The prototype of the C<function> must be C<void (*)(ev::TYPE &w, int)>.
		3210
		3211	See the method-C<set> above for more details.
		3212
		3213	Example: Use a plain function as callback.
		3214
		3215	static void io_cb (ev::io &w, int revents) { }
		3216	iow.set <io_cb> ();
		3217
		3218	=item w->set (struct ev_loop *)
		3219
		3220	Associates a different C<struct ev_loop> with this watcher. You can only
		3221	do this when the watcher is inactive (and not pending either).
		3222
		3223	=item w->set ([arguments])
		3224
		3225	Basically the same as C<ev_TYPE_set>, with the same arguments. Must be
		3226	called at least once. Unlike the C counterpart, an active watcher gets
		3227	automatically stopped and restarted when reconfiguring it with this
		3228	method.
		3229
		3230	=item w->start ()
		3231
		3232	Starts the watcher. Note that there is no C<loop> argument, as the
		3233	constructor already stores the event loop.
		3234
		3235	=item w->stop ()
		3236
		3237	Stops the watcher if it is active. Again, no C<loop> argument.
		3238
		3239	=item w->again () (C<ev::timer>, C<ev::periodic> only)
		3240
		3241	For C<ev::timer> and C<ev::periodic>, this invokes the corresponding
		3242	C<ev_TYPE_again> function.
		3243
		3244	=item w->sweep () (C<ev::embed> only)
		3245
		3246	Invokes C<ev_embed_sweep>.
		3247
		3248	=item w->update () (C<ev::stat> only)
		3249
		3250	Invokes C<ev_stat_stat>.
		3251
		3252	=back
		3253
		3254	=back
		3255
		3256	Example: Define a class with an IO and idle watcher, start one of them in
		3257	the constructor.
		3258
		3259	class myclass
		3260	{
		3261	ev::io io ; void io_cb (ev::io &w, int revents);
		3262	ev::idle idle; void idle_cb (ev::idle &w, int revents);
		3263
		3264	myclass (int fd)
		3265	{
		3266	io .set <myclass, &myclass::io_cb > (this);
		3267	idle.set <myclass, &myclass::idle_cb> (this);
		3268
		3269	io.start (fd, ev::READ);
		3270	}
		3271	};
		3272
		3273
		3274	=head1 OTHER LANGUAGE BINDINGS
		3275
		3276	Libev does not offer other language bindings itself, but bindings for a
		3277	number of languages exist in the form of third-party packages. If you know
		3278	any interesting language binding in addition to the ones listed here, drop
		3279	me a note.
		3280
		3281	=over 4
		3282
		3283	=item Perl
		3284
		3285	The EV module implements the full libev API and is actually used to test
		3286	libev. EV is developed together with libev. Apart from the EV core module,
		3287	there are additional modules that implement libev-compatible interfaces
		3288	to C<libadns> (C<EV::ADNS>, but C<AnyEvent::DNS> is preferred nowadays),
		3289	C<Net::SNMP> (C<Net::SNMP::EV>) and the C<libglib> event core (C<Glib::EV>
		3290	and C<EV::Glib>).
		3291
		3292	It can be found and installed via CPAN, its homepage is at
		3293	L<http://software.schmorp.de/pkg/EV>.
		3294
		3295	=item Python
		3296
		3297	Python bindings can be found at L<http://code.google.com/p/pyev/>. It
		3298	seems to be quite complete and well-documented.
		3299
		3300	=item Ruby
		3301
		3302	Tony Arcieri has written a ruby extension that offers access to a subset
		3303	of the libev API and adds file handle abstractions, asynchronous DNS and
		3304	more on top of it. It can be found via gem servers. Its homepage is at
		3305	L<http://rev.rubyforge.org/>.
		3306
		3307	Roger Pack reports that using the link order C<-lws2_32 -lmsvcrt-ruby-190>
		3308	makes rev work even on mingw.
		3309
		3310	=item Haskell
		3311
		3312	A haskell binding to libev is available at
		3313	L<http://hackage.haskell.org/cgi-bin/hackage-scripts/package/hlibev>.
		3314
		3315	=item D
		3316
		3317	Leandro Lucarella has written a D language binding (F<ev.d>) for libev, to
		3318	be found at L<http://proj.llucax.com.ar/wiki/evd>.
		3319
		3320	=item Ocaml
		3321
		3322	Erkki Seppala has written Ocaml bindings for libev, to be found at
		3323	L<http://modeemi.cs.tut.fi/~flux/software/ocaml-ev/>.
		3324
		3325	=back
		3326
		3327
		3328	=head1 MACRO MAGIC
		3329
		3330	Libev can be compiled with a variety of options, the most fundamental
		3331	of which is C<EV_MULTIPLICITY>. This option determines whether (most)
		3332	functions and callbacks have an initial C<struct ev_loop *> argument.
		3333
		3334	To make it easier to write programs that cope with either variant, the
		3335	following macros are defined:
		3336
		3337	=over 4
		3338
		3339	=item C<EV_A>, C<EV_A_>
		3340
		3341	This provides the loop I<argument> for functions, if one is required ("ev
		3342	loop argument"). The C<EV_A> form is used when this is the sole argument,
		3343	C<EV_A_> is used when other arguments are following. Example:
		3344
		3345	ev_unref (EV_A);
		3346	ev_timer_add (EV_A_ watcher);
		3347	ev_loop (EV_A_ 0);
		3348
		3349	It assumes the variable C<loop> of type C<struct ev_loop *> is in scope,
		3350	which is often provided by the following macro.
		3351
		3352	=item C<EV_P>, C<EV_P_>
		3353
		3354	This provides the loop I<parameter> for functions, if one is required ("ev
		3355	loop parameter"). The C<EV_P> form is used when this is the sole parameter,
		3356	C<EV_P_> is used when other parameters are following. Example:
		3357
		3358	// this is how ev_unref is being declared
		3359	static void ev_unref (EV_P);
		3360
		3361	// this is how you can declare your typical callback
		3362	static void cb (EV_P_ ev_timer *w, int revents)
		3363
		3364	It declares a parameter C<loop> of type C<struct ev_loop *>, quite
		3365	suitable for use with C<EV_A>.
		3366
		3367	=item C<EV_DEFAULT>, C<EV_DEFAULT_>
		3368
		3369	Similar to the other two macros, this gives you the value of the default
		3370	loop, if multiple loops are supported ("ev loop default").
		3371
		3372	=item C<EV_DEFAULT_UC>, C<EV_DEFAULT_UC_>
		3373
		3374	Usage identical to C<EV_DEFAULT> and C<EV_DEFAULT_>, but requires that the
		3375	default loop has been initialised (C<UC> == unchecked). Their behaviour
		3376	is undefined when the default loop has not been initialised by a previous
		3377	execution of C<EV_DEFAULT>, C<EV_DEFAULT_> or C<ev_default_init (...)>.
		3378
		3379	It is often prudent to use C<EV_DEFAULT> when initialising the first
		3380	watcher in a function but use C<EV_DEFAULT_UC> afterwards.
		3381
		3382	=back
		3383
		3384	Example: Declare and initialise a check watcher, utilising the above
		3385	macros so it will work regardless of whether multiple loops are supported
		3386	or not.
		3387
		3388	static void
		3389	check_cb (EV_P_ ev_timer *w, int revents)
		3390	{
		3391	ev_check_stop (EV_A_ w);
		3392	}
		3393
		3394	ev_check check;
		3395	ev_check_init (&check, check_cb);
		3396	ev_check_start (EV_DEFAULT_ &check);
		3397	ev_loop (EV_DEFAULT_ 0);
		3398
		3399	=head1 EMBEDDING
		3400
		3401	Libev can (and often is) directly embedded into host
		3402	applications. Examples of applications that embed it include the Deliantra
		3403	Game Server, the EV perl module, the GNU Virtual Private Ethernet (gvpe)
		3404	and rxvt-unicode.
		3405
		3406	The goal is to enable you to just copy the necessary files into your
		3407	source directory without having to change even a single line in them, so
		3408	you can easily upgrade by simply copying (or having a checked-out copy of
		3409	libev somewhere in your source tree).
		3410
		3411	=head2 FILESETS
		3412
		3413	Depending on what features you need you need to include one or more sets of files
		3414	in your application.
		3415
		3416	=head3 CORE EVENT LOOP
		3417
		3418	To include only the libev core (all the C<ev_*> functions), with manual
		3419	configuration (no autoconf):
		3420
		3421	#define EV_STANDALONE 1
		3422	#include "ev.c"
		3423
		3424	This will automatically include F<ev.h>, too, and should be done in a
		3425	single C source file only to provide the function implementations. To use
		3426	it, do the same for F<ev.h> in all files wishing to use this API (best
		3427	done by writing a wrapper around F<ev.h> that you can include instead and
		3428	where you can put other configuration options):
		3429
		3430	#define EV_STANDALONE 1
		3431	#include "ev.h"
		3432
		3433	Both header files and implementation files can be compiled with a C++
		3434	compiler (at least, that's a stated goal, and breakage will be treated
		3435	as a bug).
		3436
		3437	You need the following files in your source tree, or in a directory
		3438	in your include path (e.g. in libev/ when using -Ilibev):
		3439
		3440	ev.h
		3441	ev.c
		3442	ev_vars.h
		3443	ev_wrap.h
		3444
		3445	ev_win32.c required on win32 platforms only
		3446
		3447	ev_select.c only when select backend is enabled (which is enabled by default)
		3448	ev_poll.c only when poll backend is enabled (disabled by default)
		3449	ev_epoll.c only when the epoll backend is enabled (disabled by default)
		3450	ev_kqueue.c only when the kqueue backend is enabled (disabled by default)
		3451	ev_port.c only when the solaris port backend is enabled (disabled by default)
		3452
		3453	F<ev.c> includes the backend files directly when enabled, so you only need
		3454	to compile this single file.
		3455
		3456	=head3 LIBEVENT COMPATIBILITY API
		3457
		3458	To include the libevent compatibility API, also include:
		3459
		3460	#include "event.c"
		3461
		3462	in the file including F<ev.c>, and:
		3463
		3464	#include "event.h"
		3465
		3466	in the files that want to use the libevent API. This also includes F<ev.h>.
		3467
		3468	You need the following additional files for this:
		3469
		3470	event.h
		3471	event.c
		3472
		3473	=head3 AUTOCONF SUPPORT
		3474
		3475	Instead of using C<EV_STANDALONE=1> and providing your configuration in
		3476	whatever way you want, you can also C<m4_include([libev.m4])> in your
		3477	F<configure.ac> and leave C<EV_STANDALONE> undefined. F<ev.c> will then
		3478	include F<config.h> and configure itself accordingly.
		3479
		3480	For this of course you need the m4 file:
		3481
		3482	libev.m4
		3483
		3484	=head2 PREPROCESSOR SYMBOLS/MACROS
		3485
		3486	Libev can be configured via a variety of preprocessor symbols you have to
		3487	define before including any of its files. The default in the absence of
		3488	autoconf is documented for every option.
		3489
		3490	=over 4
		3491
		3492	=item EV_STANDALONE
		3493
		3494	Must always be C<1> if you do not use autoconf configuration, which
		3495	keeps libev from including F<config.h>, and it also defines dummy
		3496	implementations for some libevent functions (such as logging, which is not
		3497	supported). It will also not define any of the structs usually found in
		3498	F<event.h> that are not directly supported by the libev core alone.
		3499
		3500	In stanbdalone mode, libev will still try to automatically deduce the
		3501	configuration, but has to be more conservative.
		3502
		3503	=item EV_USE_MONOTONIC
		3504
		3505	If defined to be C<1>, libev will try to detect the availability of the
		3506	monotonic clock option at both compile time and runtime. Otherwise no
		3507	use of the monotonic clock option will be attempted. If you enable this,
		3508	you usually have to link against librt or something similar. Enabling it
		3509	when the functionality isn't available is safe, though, although you have
		3510	to make sure you link against any libraries where the C<clock_gettime>
		3511	function is hiding in (often F<-lrt>). See also C<EV_USE_CLOCK_SYSCALL>.
		3512
		3513	=item EV_USE_REALTIME
		3514
		3515	If defined to be C<1>, libev will try to detect the availability of the
		3516	real-time clock option at compile time (and assume its availability
		3517	at runtime if successful). Otherwise no use of the real-time clock
		3518	option will be attempted. This effectively replaces C<gettimeofday>
		3519	by C<clock_get (CLOCK_REALTIME, ...)> and will not normally affect
		3520	correctness. See the note about libraries in the description of
		3521	C<EV_USE_MONOTONIC>, though. Defaults to the opposite value of
		3522	C<EV_USE_CLOCK_SYSCALL>.
		3523
		3524	=item EV_USE_CLOCK_SYSCALL
		3525
		3526	If defined to be C<1>, libev will try to use a direct syscall instead
		3527	of calling the system-provided C<clock_gettime> function. This option
		3528	exists because on GNU/Linux, C<clock_gettime> is in C<librt>, but C<librt>
		3529	unconditionally pulls in C<libpthread>, slowing down single-threaded
		3530	programs needlessly. Using a direct syscall is slightly slower (in
		3531	theory), because no optimised vdso implementation can be used, but avoids
		3532	the pthread dependency. Defaults to C<1> on GNU/Linux with glibc 2.x or
		3533	higher, as it simplifies linking (no need for C<-lrt>).
		3534
		3535	=item EV_USE_NANOSLEEP
		3536
		3537	If defined to be C<1>, libev will assume that C<nanosleep ()> is available
		3538	and will use it for delays. Otherwise it will use C<select ()>.
		3539
		3540	=item EV_USE_EVENTFD
		3541
		3542	If defined to be C<1>, then libev will assume that C<eventfd ()> is
		3543	available and will probe for kernel support at runtime. This will improve
		3544	C<ev_signal> and C<ev_async> performance and reduce resource consumption.
		3545	If undefined, it will be enabled if the headers indicate GNU/Linux + Glibc
		3546	2.7 or newer, otherwise disabled.
		3547
		3548	=item EV_USE_SELECT
		3549
		3550	If undefined or defined to be C<1>, libev will compile in support for the
		3551	C<select>(2) backend. No attempt at auto-detection will be done: if no
		3552	other method takes over, select will be it. Otherwise the select backend
		3553	will not be compiled in.
		3554
		3555	=item EV_SELECT_USE_FD_SET
		3556
		3557	If defined to C<1>, then the select backend will use the system C<fd_set>
		3558	structure. This is useful if libev doesn't compile due to a missing
		3559	C<NFDBITS> or C<fd_mask> definition or it mis-guesses the bitset layout
		3560	on exotic systems. This usually limits the range of file descriptors to
		3561	some low limit such as 1024 or might have other limitations (winsocket
		3562	only allows 64 sockets). The C<FD_SETSIZE> macro, set before compilation,
		3563	configures the maximum size of the C<fd_set>.
		3564
		3565	=item EV_SELECT_IS_WINSOCKET
		3566
		3567	When defined to C<1>, the select backend will assume that
		3568	select/socket/connect etc. don't understand file descriptors but
		3569	wants osf handles on win32 (this is the case when the select to
		3570	be used is the winsock select). This means that it will call
		3571	C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,
		3572	it is assumed that all these functions actually work on fds, even
		3573	on win32. Should not be defined on non-win32 platforms.
		3574
		3575	=item EV_FD_TO_WIN32_HANDLE
		3576
		3577	If C<EV_SELECT_IS_WINSOCKET> is enabled, then libev needs a way to map
		3578	file descriptors to socket handles. When not defining this symbol (the
		3579	default), then libev will call C<_get_osfhandle>, which is usually
		3580	correct. In some cases, programs use their own file descriptor management,
		3581	in which case they can provide this function to map fds to socket handles.
		3582
		3583	=item EV_USE_POLL
		3584
		3585	If defined to be C<1>, libev will compile in support for the C<poll>(2)
		3586	backend. Otherwise it will be enabled on non-win32 platforms. It
		3587	takes precedence over select.
		3588
		3589	=item EV_USE_EPOLL
		3590
		3591	If defined to be C<1>, libev will compile in support for the Linux
		3592	C<epoll>(7) backend. Its availability will be detected at runtime,
		3593	otherwise another method will be used as fallback. This is the preferred
		3594	backend for GNU/Linux systems. If undefined, it will be enabled if the
		3595	headers indicate GNU/Linux + Glibc 2.4 or newer, otherwise disabled.
		3596
		3597	=item EV_USE_KQUEUE
		3598
		3599	If defined to be C<1>, libev will compile in support for the BSD style
		3600	C<kqueue>(2) backend. Its actual availability will be detected at runtime,
		3601	otherwise another method will be used as fallback. This is the preferred
		3602	backend for BSD and BSD-like systems, although on most BSDs kqueue only
		3603	supports some types of fds correctly (the only platform we found that
		3604	supports ptys for example was NetBSD), so kqueue might be compiled in, but
		3605	not be used unless explicitly requested. The best way to use it is to find
		3606	out whether kqueue supports your type of fd properly and use an embedded
		3607	kqueue loop.
		3608
		3609	=item EV_USE_PORT
		3610
		3611	If defined to be C<1>, libev will compile in support for the Solaris
		3612	10 port style backend. Its availability will be detected at runtime,
		3613	otherwise another method will be used as fallback. This is the preferred
		3614	backend for Solaris 10 systems.
		3615
		3616	=item EV_USE_DEVPOLL
		3617
		3618	Reserved for future expansion, works like the USE symbols above.
		3619
		3620	=item EV_USE_INOTIFY
		3621
		3622	If defined to be C<1>, libev will compile in support for the Linux inotify
		3623	interface to speed up C<ev_stat> watchers. Its actual availability will
		3624	be detected at runtime. If undefined, it will be enabled if the headers
		3625	indicate GNU/Linux + Glibc 2.4 or newer, otherwise disabled.
		3626
		3627	=item EV_ATOMIC_T
		3628
		3629	Libev requires an integer type (suitable for storing C<0> or C<1>) whose
		3630	access is atomic with respect to other threads or signal contexts. No such
		3631	type is easily found in the C language, so you can provide your own type
		3632	that you know is safe for your purposes. It is used both for signal handler "locking"
		3633	as well as for signal and thread safety in C<ev_async> watchers.
		3634
		3635	In the absence of this define, libev will use C<sig_atomic_t volatile>
		3636	(from F<signal.h>), which is usually good enough on most platforms.
		3637
		3638	=item EV_H
		3639
		3640	The name of the F<ev.h> header file used to include it. The default if
		3641	undefined is C<"ev.h"> in F<event.h>, F<ev.c> and F<ev++.h>. This can be
		3642	used to virtually rename the F<ev.h> header file in case of conflicts.
		3643
		3644	=item EV_CONFIG_H
		3645
		3646	If C<EV_STANDALONE> isn't C<1>, this variable can be used to override
		3647	F<ev.c>'s idea of where to find the F<config.h> file, similarly to
		3648	C<EV_H>, above.
		3649
		3650	=item EV_EVENT_H
		3651
		3652	Similarly to C<EV_H>, this macro can be used to override F<event.c>'s idea
		3653	of how the F<event.h> header can be found, the default is C<"event.h">.
		3654
		3655	=item EV_PROTOTYPES
		3656
		3657	If defined to be C<0>, then F<ev.h> will not define any function
		3658	prototypes, but still define all the structs and other symbols. This is
		3659	occasionally useful if you want to provide your own wrapper functions
		3660	around libev functions.
		3661
		3662	=item EV_MULTIPLICITY
		3663
		3664	If undefined or defined to C<1>, then all event-loop-specific functions
		3665	will have the C<struct ev_loop *> as first argument, and you can create
		3666	additional independent event loops. Otherwise there will be no support
		3667	for multiple event loops and there is no first event loop pointer
		3668	argument. Instead, all functions act on the single default loop.
		3669
		3670	=item EV_MINPRI
		3671
		3672	=item EV_MAXPRI
		3673
		3674	The range of allowed priorities. C<EV_MINPRI> must be smaller or equal to
		3675	C<EV_MAXPRI>, but otherwise there are no non-obvious limitations. You can
		3676	provide for more priorities by overriding those symbols (usually defined
		3677	to be C<-2> and C<2>, respectively).
		3678
		3679	When doing priority-based operations, libev usually has to linearly search
		3680	all the priorities, so having many of them (hundreds) uses a lot of space
		3681	and time, so using the defaults of five priorities (-2 .. +2) is usually
		3682	fine.
		3683
		3684	If your embedding application does not need any priorities, defining these
		3685	both to C<0> will save some memory and CPU.
		3686
		3687	=item EV_PERIODIC_ENABLE
		3688
		3689	If undefined or defined to be C<1>, then periodic timers are supported. If
		3690	defined to be C<0>, then they are not. Disabling them saves a few kB of
		3691	code.
		3692
		3693	=item EV_IDLE_ENABLE
		3694
		3695	If undefined or defined to be C<1>, then idle watchers are supported. If
		3696	defined to be C<0>, then they are not. Disabling them saves a few kB of
		3697	code.
		3698
		3699	=item EV_EMBED_ENABLE
		3700
		3701	If undefined or defined to be C<1>, then embed watchers are supported. If
		3702	defined to be C<0>, then they are not. Embed watchers rely on most other
		3703	watcher types, which therefore must not be disabled.
		3704
		3705	=item EV_STAT_ENABLE
		3706
		3707	If undefined or defined to be C<1>, then stat watchers are supported. If
		3708	defined to be C<0>, then they are not.
		3709
		3710	=item EV_FORK_ENABLE
		3711
		3712	If undefined or defined to be C<1>, then fork watchers are supported. If
		3713	defined to be C<0>, then they are not.
		3714
		3715	=item EV_ASYNC_ENABLE
		3716
		3717	If undefined or defined to be C<1>, then async watchers are supported. If
		3718	defined to be C<0>, then they are not.
		3719
		3720	=item EV_MINIMAL
		3721
		3722	If you need to shave off some kilobytes of code at the expense of some
		3723	speed (but with the full API), define this symbol to C<1>. Currently this
		3724	is used to override some inlining decisions, saves roughly 30% code size
		3725	on amd64. It also selects a much smaller 2-heap for timer management over
		3726	the default 4-heap.
		3727
		3728	You can save even more by disabling watcher types you do not need
		3729	and setting C<EV_MAXPRI> == C<EV_MINPRI>. Also, disabling C<assert>
		3730	(C<-DNDEBUG>) will usually reduce code size a lot.
		3731
		3732	Defining C<EV_MINIMAL> to C<2> will additionally reduce the core API to
		3733	provide a bare-bones event library. See C<ev.h> for details on what parts
		3734	of the API are still available, and do not complain if this subset changes
		3735	over time.
		3736
		3737	=item EV_PID_HASHSIZE
		3738
		3739	C<ev_child> watchers use a small hash table to distribute workload by
		3740	pid. The default size is C<16> (or C<1> with C<EV_MINIMAL>), usually more
		3741	than enough. If you need to manage thousands of children you might want to
		3742	increase this value (I<must> be a power of two).
		3743
		3744	=item EV_INOTIFY_HASHSIZE
		3745
		3746	C<ev_stat> watchers use a small hash table to distribute workload by
		3747	inotify watch id. The default size is C<16> (or C<1> with C<EV_MINIMAL>),
		3748	usually more than enough. If you need to manage thousands of C<ev_stat>
		3749	watchers you might want to increase this value (I<must> be a power of
		3750	two).
		3751
		3752	=item EV_USE_4HEAP
		3753
		3754	Heaps are not very cache-efficient. To improve the cache-efficiency of the
		3755	timer and periodics heaps, libev uses a 4-heap when this symbol is defined
		3756	to C<1>. The 4-heap uses more complicated (longer) code but has noticeably
		3757	faster performance with many (thousands) of watchers.
		3758
		3759	The default is C<1> unless C<EV_MINIMAL> is set in which case it is C<0>
		3760	(disabled).
		3761
		3762	=item EV_HEAP_CACHE_AT
		3763
		3764	Heaps are not very cache-efficient. To improve the cache-efficiency of the
		3765	timer and periodics heaps, libev can cache the timestamp (I<at>) within
		3766	the heap structure (selected by defining C<EV_HEAP_CACHE_AT> to C<1>),
		3767	which uses 8-12 bytes more per watcher and a few hundred bytes more code,
		3768	but avoids random read accesses on heap changes. This improves performance
		3769	noticeably with many (hundreds) of watchers.
		3770
		3771	The default is C<1> unless C<EV_MINIMAL> is set in which case it is C<0>
		3772	(disabled).
		3773
		3774	=item EV_VERIFY
		3775
		3776	Controls how much internal verification (see C<ev_loop_verify ()>) will
		3777	be done: If set to C<0>, no internal verification code will be compiled
		3778	in. If set to C<1>, then verification code will be compiled in, but not
		3779	called. If set to C<2>, then the internal verification code will be
		3780	called once per loop, which can slow down libev. If set to C<3>, then the
		3781	verification code will be called very frequently, which will slow down
		3782	libev considerably.
		3783
		3784	The default is C<1>, unless C<EV_MINIMAL> is set, in which case it will be
		3785	C<0>.
		3786
		3787	=item EV_COMMON
		3788
		3789	By default, all watchers have a C<void *data> member. By redefining
		3790	this macro to a something else you can include more and other types of
		3791	members. You have to define it each time you include one of the files,
		3792	though, and it must be identical each time.
		3793
		3794	For example, the perl EV module uses something like this:
		3795
		3796	#define EV_COMMON \
		3797	SV *self; /* contains this struct */ \
		3798	SV *cb_sv, *fh /* note no trailing ";" */
		3799
		3800	=item EV_CB_DECLARE (type)
		3801
		3802	=item EV_CB_INVOKE (watcher, revents)
		3803
		3804	=item ev_set_cb (ev, cb)
		3805
		3806	Can be used to change the callback member declaration in each watcher,
		3807	and the way callbacks are invoked and set. Must expand to a struct member
		3808	definition and a statement, respectively. See the F<ev.h> header file for
		3809	their default definitions. One possible use for overriding these is to
		3810	avoid the C<struct ev_loop *> as first argument in all cases, or to use
		3811	method calls instead of plain function calls in C++.
		3812
		3813	=back
		3814
		3815	=head2 EXPORTED API SYMBOLS
		3816
		3817	If you need to re-export the API (e.g. via a DLL) and you need a list of
		3818	exported symbols, you can use the provided F<Symbol.*> files which list
		3819	all public symbols, one per line:
		3820
		3821	Symbols.ev for libev proper
		3822	Symbols.event for the libevent emulation
		3823
		3824	This can also be used to rename all public symbols to avoid clashes with
		3825	multiple versions of libev linked together (which is obviously bad in
		3826	itself, but sometimes it is inconvenient to avoid this).
		3827
		3828	A sed command like this will create wrapper C<#define>'s that you need to
		3829	include before including F<ev.h>:
		3830
		3831	<Symbols.ev sed -e "s/.*/#define & myprefix_&/" >wrap.h
		3832
		3833	This would create a file F<wrap.h> which essentially looks like this:
		3834
		3835	#define ev_backend myprefix_ev_backend
		3836	#define ev_check_start myprefix_ev_check_start
		3837	#define ev_check_stop myprefix_ev_check_stop
		3838	...
		3839
		3840	=head2 EXAMPLES
		3841
		3842	For a real-world example of a program the includes libev
		3843	verbatim, you can have a look at the EV perl module
		3844	(L<http://software.schmorp.de/pkg/EV.html>). It has the libev files in
		3845	the F<libev/> subdirectory and includes them in the F<EV/EVAPI.h> (public
		3846	interface) and F<EV.xs> (implementation) files. Only the F<EV.xs> file
		3847	will be compiled. It is pretty complex because it provides its own header
		3848	file.
		3849
		3850	The usage in rxvt-unicode is simpler. It has a F<ev_cpp.h> header file
		3851	that everybody includes and which overrides some configure choices:
		3852
		3853	#define EV_MINIMAL 1
		3854	#define EV_USE_POLL 0
		3855	#define EV_MULTIPLICITY 0
		3856	#define EV_PERIODIC_ENABLE 0
		3857	#define EV_STAT_ENABLE 0
		3858	#define EV_FORK_ENABLE 0
		3859	#define EV_CONFIG_H <config.h>
		3860	#define EV_MINPRI 0
		3861	#define EV_MAXPRI 0
		3862
		3863	#include "ev++.h"
		3864
		3865	And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:
		3866
		3867	#include "ev_cpp.h"
		3868	#include "ev.c"
		3869
		3870	=head1 INTERACTION WITH OTHER PROGRAMS OR LIBRARIES
		3871
		3872	=head2 THREADS AND COROUTINES
		3873
		3874	=head3 THREADS
		3875
		3876	All libev functions are reentrant and thread-safe unless explicitly
		3877	documented otherwise, but libev implements no locking itself. This means
		3878	that you can use as many loops as you want in parallel, as long as there
		3879	are no concurrent calls into any libev function with the same loop
		3880	parameter (C<ev_default_*> calls have an implicit default loop parameter,
		3881	of course): libev guarantees that different event loops share no data
		3882	structures that need any locking.
		3883
		3884	Or to put it differently: calls with different loop parameters can be done
		3885	concurrently from multiple threads, calls with the same loop parameter
		3886	must be done serially (but can be done from different threads, as long as
		3887	only one thread ever is inside a call at any point in time, e.g. by using
		3888	a mutex per loop).
		3889
		3890	Specifically to support threads (and signal handlers), libev implements
		3891	so-called C<ev_async> watchers, which allow some limited form of
		3892	concurrency on the same event loop, namely waking it up "from the
		3893	outside".
		3894
		3895	If you want to know which design (one loop, locking, or multiple loops
		3896	without or something else still) is best for your problem, then I cannot
		3897	help you, but here is some generic advice:
		3898
		3899	=over 4
		3900
		3901	=item * most applications have a main thread: use the default libev loop
		3902	in that thread, or create a separate thread running only the default loop.
		3903
		3904	This helps integrating other libraries or software modules that use libev
		3905	themselves and don't care/know about threading.
		3906
		3907	=item * one loop per thread is usually a good model.
		3908
		3909	Doing this is almost never wrong, sometimes a better-performance model
		3910	exists, but it is always a good start.
		3911
		3912	=item * other models exist, such as the leader/follower pattern, where one
		3913	loop is handed through multiple threads in a kind of round-robin fashion.
		3914
		3915	Choosing a model is hard - look around, learn, know that usually you can do
		3916	better than you currently do :-)
		3917
		3918	=item * often you need to talk to some other thread which blocks in the
		3919	event loop.
		3920
		3921	C<ev_async> watchers can be used to wake them up from other threads safely
		3922	(or from signal contexts...).
		3923
		3924	An example use would be to communicate signals or other events that only
		3925	work in the default loop by registering the signal watcher with the
		3926	default loop and triggering an C<ev_async> watcher from the default loop
		3927	watcher callback into the event loop interested in the signal.
		3928
		3929	=back
		3930
		3931	=head4 THREAD LOCKING EXAMPLE
		3932
		3933	=head3 COROUTINES
		3934
		3935	Libev is very accommodating to coroutines ("cooperative threads"):
		3936	libev fully supports nesting calls to its functions from different
		3937	coroutines (e.g. you can call C<ev_loop> on the same loop from two
		3938	different coroutines, and switch freely between both coroutines running the
		3939	loop, as long as you don't confuse yourself). The only exception is that
		3940	you must not do this from C<ev_periodic> reschedule callbacks.
		3941
		3942	Care has been taken to ensure that libev does not keep local state inside
		3943	C<ev_loop>, and other calls do not usually allow for coroutine switches as
		3944	they do not call any callbacks.
		3945
		3946	=head2 COMPILER WARNINGS
		3947
		3948	Depending on your compiler and compiler settings, you might get no or a
		3949	lot of warnings when compiling libev code. Some people are apparently
		3950	scared by this.
		3951
		3952	However, these are unavoidable for many reasons. For one, each compiler
		3953	has different warnings, and each user has different tastes regarding
		3954	warning options. "Warn-free" code therefore cannot be a goal except when
		3955	targeting a specific compiler and compiler-version.
		3956
		3957	Another reason is that some compiler warnings require elaborate
		3958	workarounds, or other changes to the code that make it less clear and less
		3959	maintainable.
		3960
		3961	And of course, some compiler warnings are just plain stupid, or simply
		3962	wrong (because they don't actually warn about the condition their message
		3963	seems to warn about). For example, certain older gcc versions had some
		3964	warnings that resulted an extreme number of false positives. These have
		3965	been fixed, but some people still insist on making code warn-free with
		3966	such buggy versions.
		3967
		3968	While libev is written to generate as few warnings as possible,
		3969	"warn-free" code is not a goal, and it is recommended not to build libev
		3970	with any compiler warnings enabled unless you are prepared to cope with
		3971	them (e.g. by ignoring them). Remember that warnings are just that:
		3972	warnings, not errors, or proof of bugs.
		3973
		3974
		3975	=head2 VALGRIND
		3976
		3977	Valgrind has a special section here because it is a popular tool that is
		3978	highly useful. Unfortunately, valgrind reports are very hard to interpret.
		3979
		3980	If you think you found a bug (memory leak, uninitialised data access etc.)
		3981	in libev, then check twice: If valgrind reports something like:
		3982
		3983	==2274== definitely lost: 0 bytes in 0 blocks.
		3984	==2274== possibly lost: 0 bytes in 0 blocks.
		3985	==2274== still reachable: 256 bytes in 1 blocks.
		3986
		3987	Then there is no memory leak, just as memory accounted to global variables
		3988	is not a memleak - the memory is still being referenced, and didn't leak.
		3989
		3990	Similarly, under some circumstances, valgrind might report kernel bugs
		3991	as if it were a bug in libev (e.g. in realloc or in the poll backend,
		3992	although an acceptable workaround has been found here), or it might be
		3993	confused.
		3994
		3995	Keep in mind that valgrind is a very good tool, but only a tool. Don't
		3996	make it into some kind of religion.
		3997
		3998	If you are unsure about something, feel free to contact the mailing list
		3999	with the full valgrind report and an explanation on why you think this
		4000	is a bug in libev (best check the archives, too :). However, don't be
		4001	annoyed when you get a brisk "this is no bug" answer and take the chance
		4002	of learning how to interpret valgrind properly.
		4003
		4004	If you need, for some reason, empty reports from valgrind for your project
		4005	I suggest using suppression lists.
		4006
		4007
		4008	=head1 PORTABILITY NOTES
		4009
		4010	=head2 WIN32 PLATFORM LIMITATIONS AND WORKAROUNDS
		4011
		4012	Win32 doesn't support any of the standards (e.g. POSIX) that libev
		4013	requires, and its I/O model is fundamentally incompatible with the POSIX
		4014	model. Libev still offers limited functionality on this platform in
		4015	the form of the C<EVBACKEND_SELECT> backend, and only supports socket
		4016	descriptors. This only applies when using Win32 natively, not when using
		4017	e.g. cygwin.
		4018
		4019	Lifting these limitations would basically require the full
		4020	re-implementation of the I/O system. If you are into these kinds of
		4021	things, then note that glib does exactly that for you in a very portable
		4022	way (note also that glib is the slowest event library known to man).
		4023
		4024	There is no supported compilation method available on windows except
		4025	embedding it into other applications.
		4026
		4027	Sensible signal handling is officially unsupported by Microsoft - libev
		4028	tries its best, but under most conditions, signals will simply not work.
		4029
		4030	Not a libev limitation but worth mentioning: windows apparently doesn't
		4031	accept large writes: instead of resulting in a partial write, windows will
		4032	either accept everything or return C<ENOBUFS> if the buffer is too large,
		4033	so make sure you only write small amounts into your sockets (less than a
		4034	megabyte seems safe, but this apparently depends on the amount of memory
		4035	available).
		4036
		4037	Due to the many, low, and arbitrary limits on the win32 platform and
		4038	the abysmal performance of winsockets, using a large number of sockets
		4039	is not recommended (and not reasonable). If your program needs to use
		4040	more than a hundred or so sockets, then likely it needs to use a totally
		4041	different implementation for windows, as libev offers the POSIX readiness
		4042	notification model, which cannot be implemented efficiently on windows
		4043	(due to Microsoft monopoly games).
		4044
		4045	A typical way to use libev under windows is to embed it (see the embedding
		4046	section for details) and use the following F<evwrap.h> header file instead
		4047	of F<ev.h>:
		4048
		4049	#define EV_STANDALONE /* keeps ev from requiring config.h */
		4050	#define EV_SELECT_IS_WINSOCKET 1 /* configure libev for windows select */
		4051
		4052	#include "ev.h"
		4053
		4054	And compile the following F<evwrap.c> file into your project (make sure
		4055	you do I<not> compile the F<ev.c> or any other embedded source files!):
		4056
		4057	#include "evwrap.h"
		4058	#include "ev.c"
		4059
		4060	=over 4
		4061
		4062	=item The winsocket select function
		4063
		4064	The winsocket C<select> function doesn't follow POSIX in that it
		4065	requires socket I<handles> and not socket I<file descriptors> (it is
		4066	also extremely buggy). This makes select very inefficient, and also
		4067	requires a mapping from file descriptors to socket handles (the Microsoft
		4068	C runtime provides the function C<_open_osfhandle> for this). See the
		4069	discussion of the C<EV_SELECT_USE_FD_SET>, C<EV_SELECT_IS_WINSOCKET> and
		4070	C<EV_FD_TO_WIN32_HANDLE> preprocessor symbols for more info.
		4071
		4072	The configuration for a "naked" win32 using the Microsoft runtime
		4073	libraries and raw winsocket select is:
		4074
		4075	#define EV_USE_SELECT 1
		4076	#define EV_SELECT_IS_WINSOCKET 1 /* forces EV_SELECT_USE_FD_SET, too */
		4077
		4078	Note that winsockets handling of fd sets is O(n), so you can easily get a
		4079	complexity in the O(n²) range when using win32.
		4080
		4081	=item Limited number of file descriptors
		4082
		4083	Windows has numerous arbitrary (and low) limits on things.
		4084
		4085	Early versions of winsocket's select only supported waiting for a maximum
		4086	of C<64> handles (probably owning to the fact that all windows kernels
		4087	can only wait for C<64> things at the same time internally; Microsoft
		4088	recommends spawning a chain of threads and wait for 63 handles and the
		4089	previous thread in each. Sounds great!).
		4090
		4091	Newer versions support more handles, but you need to define C<FD_SETSIZE>
		4092	to some high number (e.g. C<2048>) before compiling the winsocket select
		4093	call (which might be in libev or elsewhere, for example, perl and many
		4094	other interpreters do their own select emulation on windows).
		4095
		4096	Another limit is the number of file descriptors in the Microsoft runtime
		4097	libraries, which by default is C<64> (there must be a hidden I<64>
		4098	fetish or something like this inside Microsoft). You can increase this
		4099	by calling C<_setmaxstdio>, which can increase this limit to C<2048>
		4100	(another arbitrary limit), but is broken in many versions of the Microsoft
		4101	runtime libraries. This might get you to about C<512> or C<2048> sockets
		4102	(depending on windows version and/or the phase of the moon). To get more,
		4103	you need to wrap all I/O functions and provide your own fd management, but
		4104	the cost of calling select (O(n²)) will likely make this unworkable.
		4105
		4106	=back
		4107
		4108	=head2 PORTABILITY REQUIREMENTS
		4109
		4110	In addition to a working ISO-C implementation and of course the
		4111	backend-specific APIs, libev relies on a few additional extensions:
		4112
		4113	=over 4
		4114
		4115	=item C<void (*)(ev_watcher_type *, int revents)> must have compatible
		4116	calling conventions regardless of C<ev_watcher_type *>.
		4117
		4118	Libev assumes not only that all watcher pointers have the same internal
		4119	structure (guaranteed by POSIX but not by ISO C for example), but it also
		4120	assumes that the same (machine) code can be used to call any watcher
		4121	callback: The watcher callbacks have different type signatures, but libev
		4122	calls them using an C<ev_watcher *> internally.
		4123
		4124	=item C<sig_atomic_t volatile> must be thread-atomic as well
		4125
		4126	The type C<sig_atomic_t volatile> (or whatever is defined as
		4127	C<EV_ATOMIC_T>) must be atomic with respect to accesses from different
		4128	threads. This is not part of the specification for C<sig_atomic_t>, but is
		4129	believed to be sufficiently portable.
		4130
		4131	=item C<sigprocmask> must work in a threaded environment
		4132
		4133	Libev uses C<sigprocmask> to temporarily block signals. This is not
		4134	allowed in a threaded program (C<pthread_sigmask> has to be used). Typical
		4135	pthread implementations will either allow C<sigprocmask> in the "main
		4136	thread" or will block signals process-wide, both behaviours would
		4137	be compatible with libev. Interaction between C<sigprocmask> and
		4138	C<pthread_sigmask> could complicate things, however.
		4139
		4140	The most portable way to handle signals is to block signals in all threads
		4141	except the initial one, and run the default loop in the initial thread as
		4142	well.
		4143
		4144	=item C<long> must be large enough for common memory allocation sizes
		4145
		4146	To improve portability and simplify its API, libev uses C<long> internally
		4147	instead of C<size_t> when allocating its data structures. On non-POSIX
		4148	systems (Microsoft...) this might be unexpectedly low, but is still at
		4149	least 31 bits everywhere, which is enough for hundreds of millions of
		4150	watchers.
		4151
		4152	=item C<double> must hold a time value in seconds with enough accuracy
		4153
		4154	The type C<double> is used to represent timestamps. It is required to
		4155	have at least 51 bits of mantissa (and 9 bits of exponent), which is good
		4156	enough for at least into the year 4000. This requirement is fulfilled by
		4157	implementations implementing IEEE 754, which is basically all existing
		4158	ones. With IEEE 754 doubles, you get microsecond accuracy until at least
		4159	2200.
		4160
		4161	=back
		4162
		4163	If you know of other additional requirements drop me a note.
		4164
		4165
		4166	=head1 ALGORITHMIC COMPLEXITIES
		4167
		4168	In this section the complexities of (many of) the algorithms used inside
		4169	libev will be documented. For complexity discussions about backends see
		4170	the documentation for C<ev_default_init>.
		4171
		4172	All of the following are about amortised time: If an array needs to be
		4173	extended, libev needs to realloc and move the whole array, but this
		4174	happens asymptotically rarer with higher number of elements, so O(1) might
		4175	mean that libev does a lengthy realloc operation in rare cases, but on
		4176	average it is much faster and asymptotically approaches constant time.
		4177
		4178	=over 4
		4179
		4180	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)
		4181
		4182	This means that, when you have a watcher that triggers in one hour and
		4183	there are 100 watchers that would trigger before that, then inserting will
		4184	have to skip roughly seven (C<ld 100>) of these watchers.
		4185
		4186	=item Changing timer/periodic watchers (by autorepeat or calling again): O(log skipped_other_timers)
		4187
		4188	That means that changing a timer costs less than removing/adding them,
		4189	as only the relative motion in the event queue has to be paid for.
		4190
		4191	=item Starting io/check/prepare/idle/signal/child/fork/async watchers: O(1)
		4192
		4193	These just add the watcher into an array or at the head of a list.
		4194
		4195	=item Stopping check/prepare/idle/fork/async watchers: O(1)
		4196
		4197	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))
		4198
		4199	These watchers are stored in lists, so they need to be walked to find the
		4200	correct watcher to remove. The lists are usually short (you don't usually
		4201	have many watchers waiting for the same fd or signal: one is typical, two
		4202	is rare).
		4203
		4204	=item Finding the next timer in each loop iteration: O(1)
		4205
		4206	By virtue of using a binary or 4-heap, the next timer is always found at a
		4207	fixed position in the storage array.
		4208
		4209	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)
		4210
		4211	A change means an I/O watcher gets started or stopped, which requires
		4212	libev to recalculate its status (and possibly tell the kernel, depending
		4213	on backend and whether C<ev_io_set> was used).
		4214
		4215	=item Activating one watcher (putting it into the pending state): O(1)
		4216
		4217	=item Priority handling: O(number_of_priorities)
		4218
		4219	Priorities are implemented by allocating some space for each
		4220	priority. When doing priority-based operations, libev usually has to
		4221	linearly search all the priorities, but starting/stopping and activating
		4222	watchers becomes O(1) with respect to priority handling.
		4223
		4224	=item Sending an ev_async: O(1)
		4225
		4226	=item Processing ev_async_send: O(number_of_async_watchers)
		4227
		4228	=item Processing signals: O(max_signal_number)
		4229
		4230	Sending involves a system call I<iff> there were no other C<ev_async_send>
		4231	calls in the current loop iteration. Checking for async and signal events
		4232	involves iterating over all running async watchers or all signal numbers.
		4233
		4234	=back
		4235
		4236
		4237	=head1 GLOSSARY
		4238
		4239	=over 4
		4240
		4241	=item active
		4242
		4243	A watcher is active as long as it has been started (has been attached to
		4244	an event loop) but not yet stopped (disassociated from the event loop).
		4245
		4246	=item application
		4247
		4248	In this document, an application is whatever is using libev.
		4249
		4250	=item callback
		4251
		4252	The address of a function that is called when some event has been
		4253	detected. Callbacks are being passed the event loop, the watcher that
		4254	received the event, and the actual event bitset.
		4255
		4256	=item callback invocation
		4257
		4258	The act of calling the callback associated with a watcher.
		4259
		4260	=item event
		4261
		4262	A change of state of some external event, such as data now being available
		4263	for reading on a file descriptor, time having passed or simply not having
		4264	any other events happening anymore.
		4265
		4266	In libev, events are represented as single bits (such as C<EV_READ> or
		4267	C<EV_TIMEOUT>).
		4268
		4269	=item event library
		4270
		4271	A software package implementing an event model and loop.
		4272
		4273	=item event loop
		4274
		4275	An entity that handles and processes external events and converts them
		4276	into callback invocations.
		4277
		4278	=item event model
		4279
		4280	The model used to describe how an event loop handles and processes
		4281	watchers and events.
		4282
		4283	=item pending
		4284
		4285	A watcher is pending as soon as the corresponding event has been detected,
		4286	and stops being pending as soon as the watcher will be invoked or its
		4287	pending status is explicitly cleared by the application.
		4288
		4289	A watcher can be pending, but not active. Stopping a watcher also clears
		4290	its pending status.
		4291
		4292	=item real time
		4293
		4294	The physical time that is observed. It is apparently strictly monotonic :)
		4295
		4296	=item wall-clock time
		4297
		4298	The time and date as shown on clocks. Unlike real time, it can actually
		4299	be wrong and jump forwards and backwards, e.g. when the you adjust your
		4300	clock.
		4301
		4302	=item watcher
		4303
		4304	A data structure that describes interest in certain events. Watchers need
		4305	to be started (attached to an event loop) before they can receive events.
		4306
		4307	=item watcher invocation
		4308
		4309	The act of calling the callback associated with a watcher.
		4310
		4311	=back
		4312
752	=head1 AUTHOR	4313	=head1 AUTHOR
753		4314
754	Marc Lehmann <libev@schmorp.de>.	4315	Marc Lehmann <libev@schmorp.de>, with repeated corrections by Mikael Magnusson.
755		4316

Diff Legend

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