ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/libev/ev.pod

Comparing libev/ev.pod (file contents):
Revision 1.46 by root, Mon Nov 26 10:20:43 2007 UTC vs.
Revision 1.310 by root, Thu Oct 21 12:32:47 2010 UTC

…		…
2		2
3	libev - a high performance full-featured event loop written in C	3	libev - a high performance full-featured event loop written in C
4		4
5	=head1 SYNOPSIS	5	=head1 SYNOPSIS
6		6
7	#include <ev.h>	7	#include <ev.h>
8		8
9	=head1 DESCRIPTION	9	=head2 EXAMPLE PROGRAM
		10
		11	// a single header file is required
		12	#include <ev.h>
		13
		14	#include <stdio.h> // for puts
		15
		16	// every watcher type has its own typedef'd struct
		17	// with the name ev_TYPE
		18	ev_io stdin_watcher;
		19	ev_timer timeout_watcher;
		20
		21	// all watcher callbacks have a similar signature
		22	// this callback is called when data is readable on stdin
		23	static void
		24	stdin_cb (EV_P_ ev_io *w, int revents)
		25	{
		26	puts ("stdin ready");
		27	// for one-shot events, one must manually stop the watcher
		28	// with its corresponding stop function.
		29	ev_io_stop (EV_A_ w);
		30
		31	// this causes all nested ev_run's to stop iterating
		32	ev_break (EV_A_ EVBREAK_ALL);
		33	}
		34
		35	// another callback, this time for a time-out
		36	static void
		37	timeout_cb (EV_P_ ev_timer *w, int revents)
		38	{
		39	puts ("timeout");
		40	// this causes the innermost ev_run to stop iterating
		41	ev_break (EV_A_ EVBREAK_ONE);
		42	}
		43
		44	int
		45	main (void)
		46	{
		47	// use the default event loop unless you have special needs
		48	struct ev_loop *loop = ev_default_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_run (loop, 0);
		62
		63	// unloop was called, so exit
		64	return 0;
		65	}
		66
		67	=head1 ABOUT THIS DOCUMENT
		68
		69	This document documents the libev software package.
		70
		71	The newest version of this document is also available as an html-formatted
		72	web page you might find easier to navigate when reading it for the first
		73	time: L<http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod>.
		74
		75	While this document tries to be as complete as possible in documenting
		76	libev, its usage and the rationale behind its design, it is not a tutorial
		77	on event-based programming, nor will it introduce event-based programming
		78	with libev.
		79
		80	Familiarity with event based programming techniques in general is assumed
		81	throughout this document.
		82
		83	=head1 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
		103	(for C<ev_stat>), Linux eventfd/signalfd (for faster and cleaner
		104	inter-thread wakeup (C<ev_async>)/signal handling (C<ev_signal>)) relative
		105	timers (C<ev_timer>), absolute timers with customised rescheduling
		106	(C<ev_periodic>), synchronous signals (C<ev_signal>), process status
29	events (related to SIGCHLD), and event watchers dealing with the event	107	change events (C<ev_child>), and event watchers dealing with the event
30	loop mechanism itself (idle, prepare and check watchers). It also is quite	108	loop mechanism itself (C<ev_idle>, C<ev_embed>, C<ev_prepare> and
		109	C<ev_check> watchers) as well as file watchers (C<ev_stat>) and even
		110	limited support for fork events (C<ev_fork>).
		111
		112	It also is quite fast (see this
31	fast (see this L<benchmark|http://libev.schmorp.de/bench.html> comparing	113	L<benchmark|http://libev.schmorp.de/bench.html> comparing it to libevent
32	it to libevent for example).	114	for example).
33		115
34	=head1 CONVENTIONS	116	=head2 CONVENTIONS
35		117
36	Libev is very configurable. In this manual the default configuration	118	Libev is very configurable. In this manual the default (and most common)
37	will be described, which supports multiple event loops. For more info	119	configuration will be described, which supports multiple event loops. For
38	about various configuration options please have a look at the file	120	more info about various configuration options please have a look at
39	F<README.embed> in the libev distribution. If libev was configured without	121	B<EMBED> section in this manual. If libev was configured without support
40	support for multiple event loops, then all functions taking an initial	122	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 *>)	123	name C<loop> (which is always of type C<struct ev_loop *>) will not have
42	will not have this argument.	124	this argument.
43		125
44	=head1 TIME REPRESENTATION	126	=head2 TIME REPRESENTATION
45		127
46	Libev represents time as a single floating point number, representing the	128	Libev represents time as a single floating point number, representing
47	(fractional) number of seconds since the (POSIX) epoch (somewhere near	129	the (fractional) number of seconds since the (POSIX) epoch (in practise
48	the beginning of 1970, details are complicated, don't ask). This type is	130	somewhere near the beginning of 1970, details are complicated, don't
49	called C<ev_tstamp>, which is what you should use too. It usually aliases	131	ask). This type is called C<ev_tstamp>, which is what you should use
50	to the C<double> type in C, and when you need to do any calculations on	132	too. It usually aliases to the C<double> type in C. When you need to do
51	it, you should treat it as such.	133	any calculations on it, you should treat it as some floating point value.
		134
		135	Unlike the name component C<stamp> might indicate, it is also used for
		136	time differences (e.g. delays) throughout libev.
		137
		138	=head1 ERROR HANDLING
		139
		140	Libev knows three classes of errors: operating system errors, usage errors
		141	and internal errors (bugs).
		142
		143	When libev catches an operating system error it cannot handle (for example
		144	a system call indicating a condition libev cannot fix), it calls the callback
		145	set via C<ev_set_syserr_cb>, which is supposed to fix the problem or
		146	abort. The default is to print a diagnostic message and to call C<abort
		147	()>.
		148
		149	When libev detects a usage error such as a negative timer interval, then
		150	it will print a diagnostic message and abort (via the C<assert> mechanism,
		151	so C<NDEBUG> will disable this checking): these are programming errors in
		152	the libev caller and need to be fixed there.
		153
		154	Libev also has a few internal error-checking C<assert>ions, and also has
		155	extensive consistency checking code. These do not trigger under normal
		156	circumstances, as they indicate either a bug in libev or worse.
52		157
53		158
54	=head1 GLOBAL FUNCTIONS	159	=head1 GLOBAL FUNCTIONS
55		160
56	These functions can be called anytime, even before initialising the	161	These functions can be called anytime, even before initialising the
…		…
62		167
63	Returns the current time as libev would use it. Please note that the	168	Returns the current time as libev would use it. Please note that the
64	C<ev_now> function is usually faster and also often returns the timestamp	169	C<ev_now> function is usually faster and also often returns the timestamp
65	you actually want to know.	170	you actually want to know.
66		171
		172	=item ev_sleep (ev_tstamp interval)
		173
		174	Sleep for the given interval: The current thread will be blocked until
		175	either it is interrupted or the given time interval has passed. Basically
		176	this is a sub-second-resolution C<sleep ()>.
		177
67	=item int ev_version_major ()	178	=item int ev_version_major ()
68		179
69	=item int ev_version_minor ()	180	=item int ev_version_minor ()
70		181
71	You can find out the major and minor version numbers of the library	182	You can find out the major and minor ABI version numbers of the library
72	you linked against by calling the functions C<ev_version_major> and	183	you linked against by calling the functions C<ev_version_major> and
73	C<ev_version_minor>. If you want, you can compare against the global	184	C<ev_version_minor>. If you want, you can compare against the global
74	symbols C<EV_VERSION_MAJOR> and C<EV_VERSION_MINOR>, which specify the	185	symbols C<EV_VERSION_MAJOR> and C<EV_VERSION_MINOR>, which specify the
75	version of the library your program was compiled against.	186	version of the library your program was compiled against.
76		187
		188	These version numbers refer to the ABI version of the library, not the
		189	release version.
		190
77	Usually, it's a good idea to terminate if the major versions mismatch,	191	Usually, it's a good idea to terminate if the major versions mismatch,
78	as this indicates an incompatible change. Minor versions are usually	192	as this indicates an incompatible change. Minor versions are usually
79	compatible to older versions, so a larger minor version alone is usually	193	compatible to older versions, so a larger minor version alone is usually
80	not a problem.	194	not a problem.
81		195
82	Example: make sure we haven't accidentally been linked against the wrong	196	Example: Make sure we haven't accidentally been linked against the wrong
83	version:	197	version (note, however, that this will not detect ABI mismatches :).
84		198
85	assert (("libev version mismatch",	199	assert (("libev version mismatch",
86	ev_version_major () == EV_VERSION_MAJOR	200	ev_version_major () == EV_VERSION_MAJOR
87	&& ev_version_minor () >= EV_VERSION_MINOR));	201	&& ev_version_minor () >= EV_VERSION_MINOR));
88		202
89	=item unsigned int ev_supported_backends ()	203	=item unsigned int ev_supported_backends ()
90		204
91	Return the set of all backends (i.e. their corresponding C<EV_BACKEND_*>	205	Return the set of all backends (i.e. their corresponding C<EV_BACKEND_*>
92	value) compiled into this binary of libev (independent of their	206	value) compiled into this binary of libev (independent of their
…		…
94	a description of the set values.	208	a description of the set values.
95		209
96	Example: make sure we have the epoll method, because yeah this is cool and	210	Example: make sure we have the epoll method, because yeah this is cool and
97	a must have and can we have a torrent of it please!!!11	211	a must have and can we have a torrent of it please!!!11
98		212
99	assert (("sorry, no epoll, no sex",	213	assert (("sorry, no epoll, no sex",
100	ev_supported_backends () & EVBACKEND_EPOLL));	214	ev_supported_backends () & EVBACKEND_EPOLL));
101		215
102	=item unsigned int ev_recommended_backends ()	216	=item unsigned int ev_recommended_backends ()
103		217
104	Return the set of all backends compiled into this binary of libev and also	218	Return the set of all backends compiled into this binary of libev and also
105	recommended for this platform. This set is often smaller than the one	219	recommended for this platform. This set is often smaller than the one
106	returned by C<ev_supported_backends>, as for example kqueue is broken on	220	returned by C<ev_supported_backends>, as for example kqueue is broken on
107	most BSDs and will not be autodetected unless you explicitly request it	221	most BSDs and will not be auto-detected unless you explicitly request it
108	(assuming you know what you are doing). This is the set of backends that	222	(assuming you know what you are doing). This is the set of backends that
109	libev will probe for if you specify no backends explicitly.	223	libev will probe for if you specify no backends explicitly.
110		224
111	=item unsigned int ev_embeddable_backends ()	225	=item unsigned int ev_embeddable_backends ()
112		226
…		…
116	C<ev_embeddable_backends () & ev_supported_backends ()>, likewise for	230	C<ev_embeddable_backends () & ev_supported_backends ()>, likewise for
117	recommended ones.	231	recommended ones.
118		232
119	See the description of C<ev_embed> watchers for more info.	233	See the description of C<ev_embed> watchers for more info.
120		234
121	=item ev_set_allocator (void *(*cb)(void *ptr, long size))	235	=item ev_set_allocator (void *(*cb)(void *ptr, long size)) [NOT REENTRANT]
122		236
123	Sets the allocation function to use (the prototype is similar to the	237	Sets the allocation function to use (the prototype is similar - the
124	realloc C function, the semantics are identical). It is used to allocate	238	semantics are identical to the C<realloc> C89/SuS/POSIX function). It is
125	and free memory (no surprises here). If it returns zero when memory	239	used to allocate and free memory (no surprises here). If it returns zero
126	needs to be allocated, the library might abort or take some potentially	240	when memory needs to be allocated (C<size != 0>), the library might abort
127	destructive action. The default is your system realloc function.	241	or take some potentially destructive action.
		242
		243	Since some systems (at least OpenBSD and Darwin) fail to implement
		244	correct C<realloc> semantics, libev will use a wrapper around the system
		245	C<realloc> and C<free> functions by default.
128		246
129	You could override this function in high-availability programs to, say,	247	You could override this function in high-availability programs to, say,
130	free some memory if it cannot allocate memory, to use a special allocator,	248	free some memory if it cannot allocate memory, to use a special allocator,
131	or even to sleep a while and retry until some memory is available.	249	or even to sleep a while and retry until some memory is available.
132		250
133	Example: replace the libev allocator with one that waits a bit and then	251	Example: Replace the libev allocator with one that waits a bit and then
134	retries: better than mine).	252	retries (example requires a standards-compliant C<realloc>).
135		253
136	static void *	254	static void *
137	persistent_realloc (void *ptr, long size)	255	persistent_realloc (void *ptr, size_t size)
138	{	256	{
139	for (;;)	257	for (;;)
140	{	258	{
141	void *newptr = realloc (ptr, size);	259	void *newptr = realloc (ptr, size);
142		260
…		…
148	}	266	}
149		267
150	...	268	...
151	ev_set_allocator (persistent_realloc);	269	ev_set_allocator (persistent_realloc);
152		270
153	=item ev_set_syserr_cb (void (*cb)(const char *msg));	271	=item ev_set_syserr_cb (void (*cb)(const char *msg)); [NOT REENTRANT]
154		272
155	Set the callback function to call on a retryable syscall error (such	273	Set the callback function to call on a retryable system call error (such
156	as failed select, poll, epoll_wait). The message is a printable string	274	as failed select, poll, epoll_wait). The message is a printable string
157	indicating the system call or subsystem causing the problem. If this	275	indicating the system call or subsystem causing the problem. If this
158	callback is set, then libev will expect it to remedy the sitution, no	276	callback is set, then libev will expect it to remedy the situation, no
159	matter what, when it returns. That is, libev will generally retry the	277	matter what, when it returns. That is, libev will generally retry the
160	requested operation, or, if the condition doesn't go away, do bad stuff	278	requested operation, or, if the condition doesn't go away, do bad stuff
161	(such as abort).	279	(such as abort).
162		280
163	Example: do the same thing as libev does internally:	281	Example: This is basically the same thing that libev does internally, too.
164		282
165	static void	283	static void
166	fatal_error (const char *msg)	284	fatal_error (const char *msg)
167	{	285	{
168	perror (msg);	286	perror (msg);
…		…
174		292
175	=back	293	=back
176		294
177	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP	295	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP
178		296
179	An event loop is described by a C<struct ev_loop *>. The library knows two	297	An event loop is described by a C<struct ev_loop *> (the C<struct> is
180	types of such loops, the I<default> loop, which supports signals and child	298	I<not> optional in case unless libev 3 compatibility is disabled, as libev
181	events, and dynamically created loops which do not.	299	3 had an C<ev_loop> function colliding with the struct name).
182		300
183	If you use threads, a common model is to run the default event loop	301	The library knows two types of such loops, the I<default> loop, which
184	in your main thread (or in a separate thread) and for each thread you	302	supports signals and child events, and dynamically created event loops
185	create, you also create another event loop. Libev itself does no locking	303	which do not.
186	whatsoever, so if you mix calls to the same event loop in different
187	threads, make sure you lock (this is usually a bad idea, though, even if
188	done correctly, because it's hideous and inefficient).
189		304
190	=over 4	305	=over 4
191		306
192	=item struct ev_loop *ev_default_loop (unsigned int flags)	307	=item struct ev_loop *ev_default_loop (unsigned int flags)
193		308
…		…
197	flags. If that is troubling you, check C<ev_backend ()> afterwards).	312	flags. If that is troubling you, check C<ev_backend ()> afterwards).
198		313
199	If you don't know what event loop to use, use the one returned from this	314	If you don't know what event loop to use, use the one returned from this
200	function.	315	function.
201		316
		317	Note that this function is I<not> thread-safe, so if you want to use it
		318	from multiple threads, you have to lock (note also that this is unlikely,
		319	as loops cannot be shared easily between threads anyway).
		320
		321	The default loop is the only loop that can handle C<ev_signal> and
		322	C<ev_child> watchers, and to do this, it always registers a handler
		323	for C<SIGCHLD>. If this is a problem for your application you can either
		324	create a dynamic loop with C<ev_loop_new> that doesn't do that, or you
		325	can simply overwrite the C<SIGCHLD> signal handler I<after> calling
		326	C<ev_default_init>.
		327
202	The flags argument can be used to specify special behaviour or specific	328	The flags argument can be used to specify special behaviour or specific
203	backends to use, and is usually specified as C<0> (or C<EVFLAG_AUTO>).	329	backends to use, and is usually specified as C<0> (or C<EVFLAG_AUTO>).
204		330
205	The following flags are supported:	331	The following flags are supported:
206		332
…		…
211	The default flags value. Use this if you have no clue (it's the right	337	The default flags value. Use this if you have no clue (it's the right
212	thing, believe me).	338	thing, believe me).
213		339
214	=item C<EVFLAG_NOENV>	340	=item C<EVFLAG_NOENV>
215		341
216	If this flag bit is ored into the flag value (or the program runs setuid	342	If this flag bit is or'ed into the flag value (or the program runs setuid
217	or setgid) then libev will I<not> look at the environment variable	343	or setgid) then libev will I<not> look at the environment variable
218	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will	344	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will
219	override the flags completely if it is found in the environment. This is	345	override the flags completely if it is found in the environment. This is
220	useful to try out specific backends to test their performance, or to work	346	useful to try out specific backends to test their performance, or to work
221	around bugs.	347	around bugs.
222		348
		349	=item C<EVFLAG_FORKCHECK>
		350
		351	Instead of calling C<ev_loop_fork> manually after a fork, you can also
		352	make libev check for a fork in each iteration by enabling this flag.
		353
		354	This works by calling C<getpid ()> on every iteration of the loop,
		355	and thus this might slow down your event loop if you do a lot of loop
		356	iterations and little real work, but is usually not noticeable (on my
		357	GNU/Linux system for example, C<getpid> is actually a simple 5-insn sequence
		358	without a system call and thus I<very> fast, but my GNU/Linux system also has
		359	C<pthread_atfork> which is even faster).
		360
		361	The big advantage of this flag is that you can forget about fork (and
		362	forget about forgetting to tell libev about forking) when you use this
		363	flag.
		364
		365	This flag setting cannot be overridden or specified in the C<LIBEV_FLAGS>
		366	environment variable.
		367
		368	=item C<EVFLAG_NOINOTIFY>
		369
		370	When this flag is specified, then libev will not attempt to use the
		371	I<inotify> API for it's C<ev_stat> watchers. Apart from debugging and
		372	testing, this flag can be useful to conserve inotify file descriptors, as
		373	otherwise each loop using C<ev_stat> watchers consumes one inotify handle.
		374
		375	=item C<EVFLAG_SIGNALFD>
		376
		377	When this flag is specified, then libev will attempt to use the
		378	I<signalfd> API for it's C<ev_signal> (and C<ev_child>) watchers. This API
		379	delivers signals synchronously, which makes it both faster and might make
		380	it possible to get the queued signal data. It can also simplify signal
		381	handling with threads, as long as you properly block signals in your
		382	threads that are not interested in handling them.
		383
		384	Signalfd will not be used by default as this changes your signal mask, and
		385	there are a lot of shoddy libraries and programs (glib's threadpool for
		386	example) that can't properly initialise their signal masks.
		387
223	=item C<EVBACKEND_SELECT> (value 1, portable select backend)	388	=item C<EVBACKEND_SELECT> (value 1, portable select backend)
224		389
225	This is your standard select(2) backend. Not I<completely> standard, as	390	This is your standard select(2) backend. Not I<completely> standard, as
226	libev tries to roll its own fd_set with no limits on the number of fds,	391	libev tries to roll its own fd_set with no limits on the number of fds,
227	but if that fails, expect a fairly low limit on the number of fds when	392	but if that fails, expect a fairly low limit on the number of fds when
228	using this backend. It doesn't scale too well (O(highest_fd)), but its usually	393	using this backend. It doesn't scale too well (O(highest_fd)), but its
229	the fastest backend for a low number of fds.	394	usually the fastest backend for a low number of (low-numbered :) fds.
		395
		396	To get good performance out of this backend you need a high amount of
		397	parallelism (most of the file descriptors should be busy). If you are
		398	writing a server, you should C<accept ()> in a loop to accept as many
		399	connections as possible during one iteration. You might also want to have
		400	a look at C<ev_set_io_collect_interval ()> to increase the amount of
		401	readiness notifications you get per iteration.
		402
		403	This backend maps C<EV_READ> to the C<readfds> set and C<EV_WRITE> to the
		404	C<writefds> set (and to work around Microsoft Windows bugs, also onto the
		405	C<exceptfds> set on that platform).
230		406
231	=item C<EVBACKEND_POLL> (value 2, poll backend, available everywhere except on windows)	407	=item C<EVBACKEND_POLL> (value 2, poll backend, available everywhere except on windows)
232		408
233	And this is your standard poll(2) backend. It's more complicated than	409	And this is your standard poll(2) backend. It's more complicated
234	select, but handles sparse fds better and has no artificial limit on the	410	than select, but handles sparse fds better and has no artificial
235	number of fds you can use (except it will slow down considerably with a	411	limit on the number of fds you can use (except it will slow down
236	lot of inactive fds). It scales similarly to select, i.e. O(total_fds).	412	considerably with a lot of inactive fds). It scales similarly to select,
		413	i.e. O(total_fds). See the entry for C<EVBACKEND_SELECT>, above, for
		414	performance tips.
		415
		416	This backend maps C<EV_READ> to C<POLLIN | POLLERR | POLLHUP>, and
		417	C<EV_WRITE> to C<POLLOUT | POLLERR | POLLHUP>.
237		418
238	=item C<EVBACKEND_EPOLL> (value 4, Linux)	419	=item C<EVBACKEND_EPOLL> (value 4, Linux)
239		420
		421	Use the linux-specific epoll(7) interface (for both pre- and post-2.6.9
		422	kernels).
		423
240	For few fds, this backend is a bit little slower than poll and select,	424	For few fds, this backend is a bit little slower than poll and select,
241	but it scales phenomenally better. While poll and select usually scale like	425	but it scales phenomenally better. While poll and select usually scale
242	O(total_fds) where n is the total number of fds (or the highest fd), epoll scales	426	like O(total_fds) where n is the total number of fds (or the highest fd),
243	either O(1) or O(active_fds).	427	epoll scales either O(1) or O(active_fds).
244		428
		429	The epoll mechanism deserves honorable mention as the most misdesigned
		430	of the more advanced event mechanisms: mere annoyances include silently
		431	dropping file descriptors, requiring a system call per change per file
		432	descriptor (and unnecessary guessing of parameters), problems with dup and
		433	so on. The biggest issue is fork races, however - if a program forks then
		434	I<both> parent and child process have to recreate the epoll set, which can
		435	take considerable time (one syscall per file descriptor) and is of course
		436	hard to detect.
		437
		438	Epoll is also notoriously buggy - embedding epoll fds I<should> work, but
		439	of course I<doesn't>, and epoll just loves to report events for totally
		440	I<different> file descriptors (even already closed ones, so one cannot
		441	even remove them from the set) than registered in the set (especially
		442	on SMP systems). Libev tries to counter these spurious notifications by
		443	employing an additional generation counter and comparing that against the
		444	events to filter out spurious ones, recreating the set when required. Last
		445	not least, it also refuses to work with some file descriptors which work
		446	perfectly fine with C<select> (files, many character devices...).
		447
245	While stopping and starting an I/O watcher in the same iteration will	448	While stopping, setting and starting an I/O watcher in the same iteration
246	result in some caching, there is still a syscall per such incident	449	will result in some caching, there is still a system call per such
247	(because the fd could point to a different file description now), so its	450	incident (because the same I<file descriptor> could point to a different
248	best to avoid that. Also, dup()ed file descriptors might not work very	451	I<file description> now), so its best to avoid that. Also, C<dup ()>'ed
249	well if you register events for both fds.	452	file descriptors might not work very well if you register events for both
		453	file descriptors.
250		454
251	Please note that epoll sometimes generates spurious notifications, so you	455	Best performance from this backend is achieved by not unregistering all
252	need to use non-blocking I/O or other means to avoid blocking when no data	456	watchers for a file descriptor until it has been closed, if possible,
253	(or space) is available.	457	i.e. keep at least one watcher active per fd at all times. Stopping and
		458	starting a watcher (without re-setting it) also usually doesn't cause
		459	extra overhead. A fork can both result in spurious notifications as well
		460	as in libev having to destroy and recreate the epoll object, which can
		461	take considerable time and thus should be avoided.
		462
		463	All this means that, in practice, C<EVBACKEND_SELECT> can be as fast or
		464	faster than epoll for maybe up to a hundred file descriptors, depending on
		465	the usage. So sad.
		466
		467	While nominally embeddable in other event loops, this feature is broken in
		468	all kernel versions tested so far.
		469
		470	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
		471	C<EVBACKEND_POLL>.
254		472
255	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)	473	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)
256		474
257	Kqueue deserves special mention, as at the time of this writing, it	475	Kqueue deserves special mention, as at the time of this writing, it
258	was broken on all BSDs except NetBSD (usually it doesn't work with	476	was broken on all BSDs except NetBSD (usually it doesn't work reliably
259	anything but sockets and pipes, except on Darwin, where of course its	477	with anything but sockets and pipes, except on Darwin, where of course
260	completely useless). For this reason its not being "autodetected"	478	it's completely useless). Unlike epoll, however, whose brokenness
		479	is by design, these kqueue bugs can (and eventually will) be fixed
		480	without API changes to existing programs. For this reason it's not being
261	unless you explicitly specify it explicitly in the flags (i.e. using	481	"auto-detected" unless you explicitly specify it in the flags (i.e. using
262	C<EVBACKEND_KQUEUE>).	482	C<EVBACKEND_KQUEUE>) or libev was compiled on a known-to-be-good (-enough)
		483	system like NetBSD.
		484
		485	You still can embed kqueue into a normal poll or select backend and use it
		486	only for sockets (after having made sure that sockets work with kqueue on
		487	the target platform). See C<ev_embed> watchers for more info.
263		488
264	It scales in the same way as the epoll backend, but the interface to the	489	It scales in the same way as the epoll backend, but the interface to the
265	kernel is more efficient (which says nothing about its actual speed, of	490	kernel is more efficient (which says nothing about its actual speed, of
266	course). While starting and stopping an I/O watcher does not cause an	491	course). While stopping, setting and starting an I/O watcher does never
267	extra syscall as with epoll, it still adds up to four event changes per	492	cause an extra system call as with C<EVBACKEND_EPOLL>, it still adds up to
268	incident, so its best to avoid that.	493	two event changes per incident. Support for C<fork ()> is very bad (but
		494	sane, unlike epoll) and it drops fds silently in similarly hard-to-detect
		495	cases
		496
		497	This backend usually performs well under most conditions.
		498
		499	While nominally embeddable in other event loops, this doesn't work
		500	everywhere, so you might need to test for this. And since it is broken
		501	almost everywhere, you should only use it when you have a lot of sockets
		502	(for which it usually works), by embedding it into another event loop
		503	(e.g. C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> (but C<poll> is of course
		504	also broken on OS X)) and, did I mention it, using it only for sockets.
		505
		506	This backend maps C<EV_READ> into an C<EVFILT_READ> kevent with
		507	C<NOTE_EOF>, and C<EV_WRITE> into an C<EVFILT_WRITE> kevent with
		508	C<NOTE_EOF>.
269		509
270	=item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8)	510	=item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8)
271		511
272	This is not implemented yet (and might never be).	512	This is not implemented yet (and might never be, unless you send me an
		513	implementation). According to reports, C</dev/poll> only supports sockets
		514	and is not embeddable, which would limit the usefulness of this backend
		515	immensely.
273		516
274	=item C<EVBACKEND_PORT> (value 32, Solaris 10)	517	=item C<EVBACKEND_PORT> (value 32, Solaris 10)
275		518
276	This uses the Solaris 10 port mechanism. As with everything on Solaris,	519	This uses the Solaris 10 event port mechanism. As with everything on Solaris,
277	it's really slow, but it still scales very well (O(active_fds)).	520	it's really slow, but it still scales very well (O(active_fds)).
278		521
279	Please note that solaris ports can result in a lot of spurious	522	Please note that Solaris event ports can deliver a lot of spurious
280	notifications, so you need to use non-blocking I/O or other means to avoid	523	notifications, so you need to use non-blocking I/O or other means to avoid
281	blocking when no data (or space) is available.	524	blocking when no data (or space) is available.
		525
		526	While this backend scales well, it requires one system call per active
		527	file descriptor per loop iteration. For small and medium numbers of file
		528	descriptors a "slow" C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> backend
		529	might perform better.
		530
		531	On the positive side, with the exception of the spurious readiness
		532	notifications, this backend actually performed fully to specification
		533	in all tests and is fully embeddable, which is a rare feat among the
		534	OS-specific backends (I vastly prefer correctness over speed hacks).
		535
		536	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
		537	C<EVBACKEND_POLL>.
282		538
283	=item C<EVBACKEND_ALL>	539	=item C<EVBACKEND_ALL>
284		540
285	Try all backends (even potentially broken ones that wouldn't be tried	541	Try all backends (even potentially broken ones that wouldn't be tried
286	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as	542	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as
287	C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>.	543	C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>.
288		544
		545	It is definitely not recommended to use this flag.
		546
289	=back	547	=back
290		548
291	If one or more of these are ored into the flags value, then only these	549	If one or more of the backend flags are or'ed into the flags value,
292	backends will be tried (in the reverse order as given here). If none are	550	then only these backends will be tried (in the reverse order as listed
293	specified, most compiled-in backend will be tried, usually in reverse	551	here). If none are specified, all backends in C<ev_recommended_backends
294	order of their flag values :)	552	()> will be tried.
295		553
296	The most typical usage is like this:	554	Example: This is the most typical usage.
297		555
298	if (!ev_default_loop (0))	556	if (!ev_default_loop (0))
299	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");	557	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
300		558
301	Restrict libev to the select and poll backends, and do not allow	559	Example: Restrict libev to the select and poll backends, and do not allow
302	environment settings to be taken into account:	560	environment settings to be taken into account:
303		561
304	ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);	562	ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);
305		563
306	Use whatever libev has to offer, but make sure that kqueue is used if	564	Example: Use whatever libev has to offer, but make sure that kqueue is
307	available (warning, breaks stuff, best use only with your own private	565	used if available (warning, breaks stuff, best use only with your own
308	event loop and only if you know the OS supports your types of fds):	566	private event loop and only if you know the OS supports your types of
		567	fds):
309		568
310	ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);	569	ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);
311		570
312	=item struct ev_loop *ev_loop_new (unsigned int flags)	571	=item struct ev_loop *ev_loop_new (unsigned int flags)
313		572
314	Similar to C<ev_default_loop>, but always creates a new event loop that is	573	Similar to C<ev_default_loop>, but always creates a new event loop that is
315	always distinct from the default loop. Unlike the default loop, it cannot	574	always distinct from the default loop.
316	handle signal and child watchers, and attempts to do so will be greeted by
317	undefined behaviour (or a failed assertion if assertions are enabled).
318		575
		576	Note that this function I<is> thread-safe, and one common way to use
		577	libev with threads is indeed to create one loop per thread, and using the
		578	default loop in the "main" or "initial" thread.
		579
319	Example: try to create a event loop that uses epoll and nothing else.	580	Example: Try to create a event loop that uses epoll and nothing else.
320		581
321	struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);	582	struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);
322	if (!epoller)	583	if (!epoller)
323	fatal ("no epoll found here, maybe it hides under your chair");	584	fatal ("no epoll found here, maybe it hides under your chair");
324		585
325	=item ev_default_destroy ()	586	=item ev_default_destroy ()
326		587
327	Destroys the default loop again (frees all memory and kernel state	588	Destroys the default loop (frees all memory and kernel state etc.). None
328	etc.). None of the active event watchers will be stopped in the normal	589	of the active event watchers will be stopped in the normal sense, so
329	sense, so e.g. C<ev_is_active> might still return true. It is your	590	e.g. C<ev_is_active> might still return true. It is your responsibility to
330	responsibility to either stop all watchers cleanly yoursef I<before>	591	either stop all watchers cleanly yourself I<before> calling this function,
331	calling this function, or cope with the fact afterwards (which is usually	592	or cope with the fact afterwards (which is usually the easiest thing, you
332	the easiest thing, youc na just ignore the watchers and/or C<free ()> them	593	can just ignore the watchers and/or C<free ()> them for example).
333	for example).	594
		595	Note that certain global state, such as signal state (and installed signal
		596	handlers), will not be freed by this function, and related watchers (such
		597	as signal and child watchers) would need to be stopped manually.
		598
		599	In general it is not advisable to call this function except in the
		600	rare occasion where you really need to free e.g. the signal handling
		601	pipe fds. If you need dynamically allocated loops it is better to use
		602	C<ev_loop_new> and C<ev_loop_destroy>.
334		603
335	=item ev_loop_destroy (loop)	604	=item ev_loop_destroy (loop)
336		605
337	Like C<ev_default_destroy>, but destroys an event loop created by an	606	Like C<ev_default_destroy>, but destroys an event loop created by an
338	earlier call to C<ev_loop_new>.	607	earlier call to C<ev_loop_new>.
339		608
340	=item ev_default_fork ()	609	=item ev_default_fork ()
341		610
		611	This function sets a flag that causes subsequent C<ev_run> iterations
342	This function reinitialises the kernel state for backends that have	612	to reinitialise the kernel state for backends that have one. Despite the
343	one. Despite the name, you can call it anytime, but it makes most sense	613	name, you can call it anytime, but it makes most sense after forking, in
344	after forking, in either the parent or child process (or both, but that	614	the child process (or both child and parent, but that again makes little
345	again makes little sense).	615	sense). You I<must> call it in the child before using any of the libev
		616	functions, and it will only take effect at the next C<ev_run> iteration.
346		617
347	You I<must> call this function in the child process after forking if and	618	Again, you I<have> to call it on I<any> loop that you want to re-use after
348	only if you want to use the event library in both processes. If you just	619	a fork, I<even if you do not plan to use the loop in the parent>. This is
349	fork+exec, you don't have to call it.	620	because some kernel interfaces *cough* I<kqueue> *cough* do funny things
		621	during fork.
		622
		623	On the other hand, you only need to call this function in the child
		624	process if and only if you want to use the event loop in the child. If
		625	you just fork+exec or create a new loop in the child, you don't have to
		626	call it at all (in fact, C<epoll> is so badly broken that it makes a
		627	difference, but libev will usually detect this case on its own and do a
		628	costly reset of the backend).
350		629
351	The function itself is quite fast and it's usually not a problem to call	630	The function itself is quite fast and it's usually not a problem to call
352	it just in case after a fork. To make this easy, the function will fit in	631	it just in case after a fork. To make this easy, the function will fit in
353	quite nicely into a call to C<pthread_atfork>:	632	quite nicely into a call to C<pthread_atfork>:
354		633
355	pthread_atfork (0, 0, ev_default_fork);	634	pthread_atfork (0, 0, ev_default_fork);
356		635
357	At the moment, C<EVBACKEND_SELECT> and C<EVBACKEND_POLL> are safe to use
358	without calling this function, so if you force one of those backends you
359	do not need to care.
360
361	=item ev_loop_fork (loop)	636	=item ev_loop_fork (loop)
362		637
363	Like C<ev_default_fork>, but acts on an event loop created by	638	Like C<ev_default_fork>, but acts on an event loop created by
364	C<ev_loop_new>. Yes, you have to call this on every allocated event loop	639	C<ev_loop_new>. Yes, you have to call this on every allocated event loop
365	after fork, and how you do this is entirely your own problem.	640	after fork that you want to re-use in the child, and how you keep track of
		641	them is entirely your own problem.
		642
		643	=item int ev_is_default_loop (loop)
		644
		645	Returns true when the given loop is, in fact, the default loop, and false
		646	otherwise.
		647
		648	=item unsigned int ev_iteration (loop)
		649
		650	Returns the current iteration count for the event loop, which is identical
		651	to the number of times libev did poll for new events. It starts at C<0>
		652	and happily wraps around with enough iterations.
		653
		654	This value can sometimes be useful as a generation counter of sorts (it
		655	"ticks" the number of loop iterations), as it roughly corresponds with
		656	C<ev_prepare> and C<ev_check> calls - and is incremented between the
		657	prepare and check phases.
		658
		659	=item unsigned int ev_depth (loop)
		660
		661	Returns the number of times C<ev_run> was entered minus the number of
		662	times C<ev_run> was exited, in other words, the recursion depth.
		663
		664	Outside C<ev_run>, this number is zero. In a callback, this number is
		665	C<1>, unless C<ev_run> was invoked recursively (or from another thread),
		666	in which case it is higher.
		667
		668	Leaving C<ev_run> abnormally (setjmp/longjmp, cancelling the thread
		669	etc.), doesn't count as "exit" - consider this as a hint to avoid such
		670	ungentleman-like behaviour unless it's really convenient.
366		671
367	=item unsigned int ev_backend (loop)	672	=item unsigned int ev_backend (loop)
368		673
369	Returns one of the C<EVBACKEND_*> flags indicating the event backend in	674	Returns one of the C<EVBACKEND_*> flags indicating the event backend in
370	use.	675	use.
…		…
373		678
374	Returns the current "event loop time", which is the time the event loop	679	Returns the current "event loop time", which is the time the event loop
375	received events and started processing them. This timestamp does not	680	received events and started processing them. This timestamp does not
376	change as long as callbacks are being processed, and this is also the base	681	change as long as callbacks are being processed, and this is also the base
377	time used for relative timers. You can treat it as the timestamp of the	682	time used for relative timers. You can treat it as the timestamp of the
378	event occuring (or more correctly, libev finding out about it).	683	event occurring (or more correctly, libev finding out about it).
379		684
		685	=item ev_now_update (loop)
		686
		687	Establishes the current time by querying the kernel, updating the time
		688	returned by C<ev_now ()> in the progress. This is a costly operation and
		689	is usually done automatically within C<ev_run ()>.
		690
		691	This function is rarely useful, but when some event callback runs for a
		692	very long time without entering the event loop, updating libev's idea of
		693	the current time is a good idea.
		694
		695	See also L<The special problem of time updates> in the C<ev_timer> section.
		696
		697	=item ev_suspend (loop)
		698
		699	=item ev_resume (loop)
		700
		701	These two functions suspend and resume an event loop, for use when the
		702	loop is not used for a while and timeouts should not be processed.
		703
		704	A typical use case would be an interactive program such as a game: When
		705	the user presses C<^Z> to suspend the game and resumes it an hour later it
		706	would be best to handle timeouts as if no time had actually passed while
		707	the program was suspended. This can be achieved by calling C<ev_suspend>
		708	in your C<SIGTSTP> handler, sending yourself a C<SIGSTOP> and calling
		709	C<ev_resume> directly afterwards to resume timer processing.
		710
		711	Effectively, all C<ev_timer> watchers will be delayed by the time spend
		712	between C<ev_suspend> and C<ev_resume>, and all C<ev_periodic> watchers
		713	will be rescheduled (that is, they will lose any events that would have
		714	occurred while suspended).
		715
		716	After calling C<ev_suspend> you B<must not> call I<any> function on the
		717	given loop other than C<ev_resume>, and you B<must not> call C<ev_resume>
		718	without a previous call to C<ev_suspend>.
		719
		720	Calling C<ev_suspend>/C<ev_resume> has the side effect of updating the
		721	event loop time (see C<ev_now_update>).
		722
380	=item ev_loop (loop, int flags)	723	=item ev_run (loop, int flags)
381		724
382	Finally, this is it, the event handler. This function usually is called	725	Finally, this is it, the event handler. This function usually is called
383	after you initialised all your watchers and you want to start handling	726	after you have initialised all your watchers and you want to start
384	events.	727	handling events. It will ask the operating system for any new events, call
		728	the watcher callbacks, an then repeat the whole process indefinitely: This
		729	is why event loops are called I<loops>.
385		730
386	If the flags argument is specified as C<0>, it will not return until	731	If the flags argument is specified as C<0>, it will keep handling events
387	either no event watchers are active anymore or C<ev_unloop> was called.	732	until either no event watchers are active anymore or C<ev_break> was
		733	called.
388		734
389	Please note that an explicit C<ev_unloop> is usually better than	735	Please note that an explicit C<ev_break> is usually better than
390	relying on all watchers to be stopped when deciding when a program has	736	relying on all watchers to be stopped when deciding when a program has
391	finished (especially in interactive programs), but having a program that	737	finished (especially in interactive programs), but having a program
392	automatically loops as long as it has to and no longer by virtue of	738	that automatically loops as long as it has to and no longer by virtue
393	relying on its watchers stopping correctly is a thing of beauty.	739	of relying on its watchers stopping correctly, that is truly a thing of
		740	beauty.
394		741
395	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle	742	A flags value of C<EVRUN_NOWAIT> will look for new events, will handle
396	those events and any outstanding ones, but will not block your process in	743	those events and any already outstanding ones, but will not wait and
397	case there are no events and will return after one iteration of the loop.	744	block your process in case there are no events and will return after one
		745	iteration of the loop. This is sometimes useful to poll and handle new
		746	events while doing lengthy calculations, to keep the program responsive.
398		747
399	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if	748	A flags value of C<EVRUN_ONCE> will look for new events (waiting if
400	neccessary) and will handle those and any outstanding ones. It will block	749	necessary) and will handle those and any already outstanding ones. It
401	your process until at least one new event arrives, and will return after	750	will block your process until at least one new event arrives (which could
402	one iteration of the loop. This is useful if you are waiting for some	751	be an event internal to libev itself, so there is no guarantee that a
403	external event in conjunction with something not expressible using other	752	user-registered callback will be called), and will return after one
		753	iteration of the loop.
		754
		755	This is useful if you are waiting for some external event in conjunction
		756	with something not expressible using other libev watchers (i.e. "roll your
404	libev watchers. However, a pair of C<ev_prepare>/C<ev_check> watchers is	757	own C<ev_run>"). However, a pair of C<ev_prepare>/C<ev_check> watchers is
405	usually a better approach for this kind of thing.	758	usually a better approach for this kind of thing.
406		759
407	Here are the gory details of what C<ev_loop> does:	760	Here are the gory details of what C<ev_run> does:
408		761
409	* If there are no active watchers (reference count is zero), return.	762	- Increment loop depth.
410	- Queue prepare watchers and then call all outstanding watchers.	763	- Reset the ev_break status.
		764	- Before the first iteration, call any pending watchers.
		765	LOOP:
		766	- If EVFLAG_FORKCHECK was used, check for a fork.
		767	- If a fork was detected (by any means), queue and call all fork watchers.
		768	- Queue and call all prepare watchers.
		769	- If ev_break was called, goto FINISH.
411	- If we have been forked, recreate the kernel state.	770	- If we have been forked, detach and recreate the kernel state
		771	as to not disturb the other process.
412	- Update the kernel state with all outstanding changes.	772	- Update the kernel state with all outstanding changes.
413	- Update the "event loop time".	773	- Update the "event loop time" (ev_now ()).
414	- Calculate for how long to block.	774	- Calculate for how long to sleep or block, if at all
		775	(active idle watchers, EVRUN_NOWAIT or not having
		776	any active watchers at all will result in not sleeping).
		777	- Sleep if the I/O and timer collect interval say so.
		778	- Increment loop iteration counter.
415	- Block the process, waiting for any events.	779	- Block the process, waiting for any events.
416	- Queue all outstanding I/O (fd) events.	780	- Queue all outstanding I/O (fd) events.
417	- Update the "event loop time" and do time jump handling.	781	- Update the "event loop time" (ev_now ()), and do time jump adjustments.
418	- Queue all outstanding timers.	782	- Queue all expired timers.
419	- Queue all outstanding periodics.	783	- Queue all expired periodics.
420	- If no events are pending now, queue all idle watchers.	784	- Queue all idle watchers with priority higher than that of pending events.
421	- Queue all check watchers.	785	- Queue all check watchers.
422	- Call all queued watchers in reverse order (i.e. check watchers first).	786	- Call all queued watchers in reverse order (i.e. check watchers first).
423	Signals and child watchers are implemented as I/O watchers, and will	787	Signals and child watchers are implemented as I/O watchers, and will
424	be handled here by queueing them when their watcher gets executed.	788	be handled here by queueing them when their watcher gets executed.
425	- If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK	789	- If ev_break has been called, or EVRUN_ONCE or EVRUN_NOWAIT
426	were used, return, otherwise continue with step *.	790	were used, or there are no active watchers, goto FINISH, otherwise
		791	continue with step LOOP.
		792	FINISH:
		793	- Reset the ev_break status iff it was EVBREAK_ONE.
		794	- Decrement the loop depth.
		795	- Return.
427		796
428	Example: queue some jobs and then loop until no events are outsanding	797	Example: Queue some jobs and then loop until no events are outstanding
429	anymore.	798	anymore.
430		799
431	... queue jobs here, make sure they register event watchers as long	800	... queue jobs here, make sure they register event watchers as long
432	... as they still have work to do (even an idle watcher will do..)	801	... as they still have work to do (even an idle watcher will do..)
433	ev_loop (my_loop, 0);	802	ev_run (my_loop, 0);
434	... jobs done. yeah!	803	... jobs done or somebody called unloop. yeah!
435		804
436	=item ev_unloop (loop, how)	805	=item ev_break (loop, how)
437		806
438	Can be used to make a call to C<ev_loop> return early (but only after it	807	Can be used to make a call to C<ev_run> return early (but only after it
439	has processed all outstanding events). The C<how> argument must be either	808	has processed all outstanding events). The C<how> argument must be either
440	C<EVUNLOOP_ONE>, which will make the innermost C<ev_loop> call return, or	809	C<EVBREAK_ONE>, which will make the innermost C<ev_run> call return, or
441	C<EVUNLOOP_ALL>, which will make all nested C<ev_loop> calls return.	810	C<EVBREAK_ALL>, which will make all nested C<ev_run> calls return.
		811
		812	This "unloop state" will be cleared when entering C<ev_run> again.
		813
		814	It is safe to call C<ev_break> from outside any C<ev_run> calls. ##TODO##
442		815
443	=item ev_ref (loop)	816	=item ev_ref (loop)
444		817
445	=item ev_unref (loop)	818	=item ev_unref (loop)
446		819
447	Ref/unref can be used to add or remove a reference count on the event	820	Ref/unref can be used to add or remove a reference count on the event
448	loop: Every watcher keeps one reference, and as long as the reference	821	loop: Every watcher keeps one reference, and as long as the reference
449	count is nonzero, C<ev_loop> will not return on its own. If you have	822	count is nonzero, C<ev_run> will not return on its own.
450	a watcher you never unregister that should not keep C<ev_loop> from	823
451	returning, ev_unref() after starting, and ev_ref() before stopping it. For	824	This is useful when you have a watcher that you never intend to
		825	unregister, but that nevertheless should not keep C<ev_run> from
		826	returning. In such a case, call C<ev_unref> after starting, and C<ev_ref>
		827	before stopping it.
		828
452	example, libev itself uses this for its internal signal pipe: It is not	829	As an example, libev itself uses this for its internal signal pipe: It
453	visible to the libev user and should not keep C<ev_loop> from exiting if	830	is not visible to the libev user and should not keep C<ev_run> from
454	no event watchers registered by it are active. It is also an excellent	831	exiting if no event watchers registered by it are active. It is also an
455	way to do this for generic recurring timers or from within third-party	832	excellent way to do this for generic recurring timers or from within
456	libraries. Just remember to I<unref after start> and I<ref before stop>.	833	third-party libraries. Just remember to I<unref after start> and I<ref
		834	before stop> (but only if the watcher wasn't active before, or was active
		835	before, respectively. Note also that libev might stop watchers itself
		836	(e.g. non-repeating timers) in which case you have to C<ev_ref>
		837	in the callback).
457		838
458	Example: create a signal watcher, but keep it from keeping C<ev_loop>	839	Example: Create a signal watcher, but keep it from keeping C<ev_run>
459	running when nothing else is active.	840	running when nothing else is active.
460		841
461	struct dv_signal exitsig;	842	ev_signal exitsig;
462	ev_signal_init (&exitsig, sig_cb, SIGINT);	843	ev_signal_init (&exitsig, sig_cb, SIGINT);
463	ev_signal_start (myloop, &exitsig);	844	ev_signal_start (loop, &exitsig);
464	evf_unref (myloop);	845	evf_unref (loop);
465		846
466	Example: for some weird reason, unregister the above signal handler again.	847	Example: For some weird reason, unregister the above signal handler again.
467		848
468	ev_ref (myloop);	849	ev_ref (loop);
469	ev_signal_stop (myloop, &exitsig);	850	ev_signal_stop (loop, &exitsig);
		851
		852	=item ev_set_io_collect_interval (loop, ev_tstamp interval)
		853
		854	=item ev_set_timeout_collect_interval (loop, ev_tstamp interval)
		855
		856	These advanced functions influence the time that libev will spend waiting
		857	for events. Both time intervals are by default C<0>, meaning that libev
		858	will try to invoke timer/periodic callbacks and I/O callbacks with minimum
		859	latency.
		860
		861	Setting these to a higher value (the C<interval> I<must> be >= C<0>)
		862	allows libev to delay invocation of I/O and timer/periodic callbacks
		863	to increase efficiency of loop iterations (or to increase power-saving
		864	opportunities).
		865
		866	The idea is that sometimes your program runs just fast enough to handle
		867	one (or very few) event(s) per loop iteration. While this makes the
		868	program responsive, it also wastes a lot of CPU time to poll for new
		869	events, especially with backends like C<select ()> which have a high
		870	overhead for the actual polling but can deliver many events at once.
		871
		872	By setting a higher I<io collect interval> you allow libev to spend more
		873	time collecting I/O events, so you can handle more events per iteration,
		874	at the cost of increasing latency. Timeouts (both C<ev_periodic> and
		875	C<ev_timer>) will be not affected. Setting this to a non-null value will
		876	introduce an additional C<ev_sleep ()> call into most loop iterations. The
		877	sleep time ensures that libev will not poll for I/O events more often then
		878	once per this interval, on average.
		879
		880	Likewise, by setting a higher I<timeout collect interval> you allow libev
		881	to spend more time collecting timeouts, at the expense of increased
		882	latency/jitter/inexactness (the watcher callback will be called
		883	later). C<ev_io> watchers will not be affected. Setting this to a non-null
		884	value will not introduce any overhead in libev.
		885
		886	Many (busy) programs can usually benefit by setting the I/O collect
		887	interval to a value near C<0.1> or so, which is often enough for
		888	interactive servers (of course not for games), likewise for timeouts. It
		889	usually doesn't make much sense to set it to a lower value than C<0.01>,
		890	as this approaches the timing granularity of most systems. Note that if
		891	you do transactions with the outside world and you can't increase the
		892	parallelity, then this setting will limit your transaction rate (if you
		893	need to poll once per transaction and the I/O collect interval is 0.01,
		894	then you can't do more than 100 transactions per second).
		895
		896	Setting the I<timeout collect interval> can improve the opportunity for
		897	saving power, as the program will "bundle" timer callback invocations that
		898	are "near" in time together, by delaying some, thus reducing the number of
		899	times the process sleeps and wakes up again. Another useful technique to
		900	reduce iterations/wake-ups is to use C<ev_periodic> watchers and make sure
		901	they fire on, say, one-second boundaries only.
		902
		903	Example: we only need 0.1s timeout granularity, and we wish not to poll
		904	more often than 100 times per second:
		905
		906	ev_set_timeout_collect_interval (EV_DEFAULT_UC_ 0.1);
		907	ev_set_io_collect_interval (EV_DEFAULT_UC_ 0.01);
		908
		909	=item ev_invoke_pending (loop)
		910
		911	This call will simply invoke all pending watchers while resetting their
		912	pending state. Normally, C<ev_run> does this automatically when required,
		913	but when overriding the invoke callback this call comes handy.
		914
		915	=item int ev_pending_count (loop)
		916
		917	Returns the number of pending watchers - zero indicates that no watchers
		918	are pending.
		919
		920	=item ev_set_invoke_pending_cb (loop, void (*invoke_pending_cb)(EV_P))
		921
		922	This overrides the invoke pending functionality of the loop: Instead of
		923	invoking all pending watchers when there are any, C<ev_run> will call
		924	this callback instead. This is useful, for example, when you want to
		925	invoke the actual watchers inside another context (another thread etc.).
		926
		927	If you want to reset the callback, use C<ev_invoke_pending> as new
		928	callback.
		929
		930	=item ev_set_loop_release_cb (loop, void (*release)(EV_P), void (*acquire)(EV_P))
		931
		932	Sometimes you want to share the same loop between multiple threads. This
		933	can be done relatively simply by putting mutex_lock/unlock calls around
		934	each call to a libev function.
		935
		936	However, C<ev_run> can run an indefinite time, so it is not feasible
		937	to wait for it to return. One way around this is to wake up the event
		938	loop via C<ev_break> and C<av_async_send>, another way is to set these
		939	I<release> and I<acquire> callbacks on the loop.
		940
		941	When set, then C<release> will be called just before the thread is
		942	suspended waiting for new events, and C<acquire> is called just
		943	afterwards.
		944
		945	Ideally, C<release> will just call your mutex_unlock function, and
		946	C<acquire> will just call the mutex_lock function again.
		947
		948	While event loop modifications are allowed between invocations of
		949	C<release> and C<acquire> (that's their only purpose after all), no
		950	modifications done will affect the event loop, i.e. adding watchers will
		951	have no effect on the set of file descriptors being watched, or the time
		952	waited. Use an C<ev_async> watcher to wake up C<ev_run> when you want it
		953	to take note of any changes you made.
		954
		955	In theory, threads executing C<ev_run> will be async-cancel safe between
		956	invocations of C<release> and C<acquire>.
		957
		958	See also the locking example in the C<THREADS> section later in this
		959	document.
		960
		961	=item ev_set_userdata (loop, void *data)
		962
		963	=item ev_userdata (loop)
		964
		965	Set and retrieve a single C<void *> associated with a loop. When
		966	C<ev_set_userdata> has never been called, then C<ev_userdata> returns
		967	C<0.>
		968
		969	These two functions can be used to associate arbitrary data with a loop,
		970	and are intended solely for the C<invoke_pending_cb>, C<release> and
		971	C<acquire> callbacks described above, but of course can be (ab-)used for
		972	any other purpose as well.
		973
		974	=item ev_verify (loop)
		975
		976	This function only does something when C<EV_VERIFY> support has been
		977	compiled in, which is the default for non-minimal builds. It tries to go
		978	through all internal structures and checks them for validity. If anything
		979	is found to be inconsistent, it will print an error message to standard
		980	error and call C<abort ()>.
		981
		982	This can be used to catch bugs inside libev itself: under normal
		983	circumstances, this function will never abort as of course libev keeps its
		984	data structures consistent.
470		985
471	=back	986	=back
472		987
473		988
474	=head1 ANATOMY OF A WATCHER	989	=head1 ANATOMY OF A WATCHER
		990
		991	In the following description, uppercase C<TYPE> in names stands for the
		992	watcher type, e.g. C<ev_TYPE_start> can mean C<ev_timer_start> for timer
		993	watchers and C<ev_io_start> for I/O watchers.
475		994
476	A watcher is a structure that you create and register to record your	995	A watcher is a structure that you create and register to record your
477	interest in some event. For instance, if you want to wait for STDIN to	996	interest in some event. For instance, if you want to wait for STDIN to
478	become readable, you would create an C<ev_io> watcher for that:	997	become readable, you would create an C<ev_io> watcher for that:
479		998
480	static void my_cb (struct ev_loop *loop, struct ev_io *w, int revents)	999	static void my_cb (struct ev_loop *loop, ev_io *w, int revents)
481	{	1000	{
482	ev_io_stop (w);	1001	ev_io_stop (w);
483	ev_unloop (loop, EVUNLOOP_ALL);	1002	ev_break (loop, EVBREAK_ALL);
484	}	1003	}
485		1004
486	struct ev_loop *loop = ev_default_loop (0);	1005	struct ev_loop *loop = ev_default_loop (0);
		1006
487	struct ev_io stdin_watcher;	1007	ev_io stdin_watcher;
		1008
488	ev_init (&stdin_watcher, my_cb);	1009	ev_init (&stdin_watcher, my_cb);
489	ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);	1010	ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);
490	ev_io_start (loop, &stdin_watcher);	1011	ev_io_start (loop, &stdin_watcher);
		1012
491	ev_loop (loop, 0);	1013	ev_run (loop, 0);
492		1014
493	As you can see, you are responsible for allocating the memory for your	1015	As you can see, you are responsible for allocating the memory for your
494	watcher structures (and it is usually a bad idea to do this on the stack,	1016	watcher structures (and it is I<usually> a bad idea to do this on the
495	although this can sometimes be quite valid).	1017	stack).
		1018
		1019	Each watcher has an associated watcher structure (called C<struct ev_TYPE>
		1020	or simply C<ev_TYPE>, as typedefs are provided for all watcher structs).
496		1021
497	Each watcher structure must be initialised by a call to C<ev_init	1022	Each watcher structure must be initialised by a call to C<ev_init
498	(watcher *, callback)>, which expects a callback to be provided. This	1023	(watcher *, callback)>, which expects a callback to be provided. This
499	callback gets invoked each time the event occurs (or, in the case of io	1024	callback gets invoked each time the event occurs (or, in the case of I/O
500	watchers, each time the event loop detects that the file descriptor given	1025	watchers, each time the event loop detects that the file descriptor given
501	is readable and/or writable).	1026	is readable and/or writable).
502		1027
503	Each watcher type has its own C<< ev_<type>_set (watcher *, ...) >> macro	1028	Each watcher type further has its own C<< ev_TYPE_set (watcher *, ...) >>
504	with arguments specific to this watcher type. There is also a macro	1029	macro to configure it, with arguments specific to the watcher type. There
505	to combine initialisation and setting in one call: C<< ev_<type>_init	1030	is also a macro to combine initialisation and setting in one call: C<<
506	(watcher *, callback, ...) >>.	1031	ev_TYPE_init (watcher *, callback, ...) >>.
507		1032
508	To make the watcher actually watch out for events, you have to start it	1033	To make the watcher actually watch out for events, you have to start it
509	with a watcher-specific start function (C<< ev_<type>_start (loop, watcher	1034	with a watcher-specific start function (C<< ev_TYPE_start (loop, watcher
510	*) >>), and you can stop watching for events at any time by calling the	1035	*) >>), and you can stop watching for events at any time by calling the
511	corresponding stop function (C<< ev_<type>_stop (loop, watcher *) >>.	1036	corresponding stop function (C<< ev_TYPE_stop (loop, watcher *) >>.
512		1037
513	As long as your watcher is active (has been started but not stopped) you	1038	As long as your watcher is active (has been started but not stopped) you
514	must not touch the values stored in it. Most specifically you must never	1039	must not touch the values stored in it. Most specifically you must never
515	reinitialise it or call its C<set> macro.	1040	reinitialise it or call its C<ev_TYPE_set> macro.
516		1041
517	Each and every callback receives the event loop pointer as first, the	1042	Each and every callback receives the event loop pointer as first, the
518	registered watcher structure as second, and a bitset of received events as	1043	registered watcher structure as second, and a bitset of received events as
519	third argument.	1044	third argument.
520		1045
…		…
529	=item C<EV_WRITE>	1054	=item C<EV_WRITE>
530		1055
531	The file descriptor in the C<ev_io> watcher has become readable and/or	1056	The file descriptor in the C<ev_io> watcher has become readable and/or
532	writable.	1057	writable.
533		1058
534	=item C<EV_TIMEOUT>	1059	=item C<EV_TIMER>
535		1060
536	The C<ev_timer> watcher has timed out.	1061	The C<ev_timer> watcher has timed out.
537		1062
538	=item C<EV_PERIODIC>	1063	=item C<EV_PERIODIC>
539		1064
…		…
545		1070
546	=item C<EV_CHILD>	1071	=item C<EV_CHILD>
547		1072
548	The pid specified in the C<ev_child> watcher has received a status change.	1073	The pid specified in the C<ev_child> watcher has received a status change.
549		1074
		1075	=item C<EV_STAT>
		1076
		1077	The path specified in the C<ev_stat> watcher changed its attributes somehow.
		1078
550	=item C<EV_IDLE>	1079	=item C<EV_IDLE>
551		1080
552	The C<ev_idle> watcher has determined that you have nothing better to do.	1081	The C<ev_idle> watcher has determined that you have nothing better to do.
553		1082
554	=item C<EV_PREPARE>	1083	=item C<EV_PREPARE>
555		1084
556	=item C<EV_CHECK>	1085	=item C<EV_CHECK>
557		1086
558	All C<ev_prepare> watchers are invoked just I<before> C<ev_loop> starts	1087	All C<ev_prepare> watchers are invoked just I<before> C<ev_run> starts
559	to gather new events, and all C<ev_check> watchers are invoked just after	1088	to gather new events, and all C<ev_check> watchers are invoked just after
560	C<ev_loop> has gathered them, but before it invokes any callbacks for any	1089	C<ev_run> has gathered them, but before it invokes any callbacks for any
561	received events. Callbacks of both watcher types can start and stop as	1090	received events. Callbacks of both watcher types can start and stop as
562	many watchers as they want, and all of them will be taken into account	1091	many watchers as they want, and all of them will be taken into account
563	(for example, a C<ev_prepare> watcher might start an idle watcher to keep	1092	(for example, a C<ev_prepare> watcher might start an idle watcher to keep
564	C<ev_loop> from blocking).	1093	C<ev_run> from blocking).
		1094
		1095	=item C<EV_EMBED>
		1096
		1097	The embedded event loop specified in the C<ev_embed> watcher needs attention.
		1098
		1099	=item C<EV_FORK>
		1100
		1101	The event loop has been resumed in the child process after fork (see
		1102	C<ev_fork>).
		1103
		1104	=item C<EV_ASYNC>
		1105
		1106	The given async watcher has been asynchronously notified (see C<ev_async>).
		1107
		1108	=item C<EV_CUSTOM>
		1109
		1110	Not ever sent (or otherwise used) by libev itself, but can be freely used
		1111	by libev users to signal watchers (e.g. via C<ev_feed_event>).
565		1112
566	=item C<EV_ERROR>	1113	=item C<EV_ERROR>
567		1114
568	An unspecified error has occured, the watcher has been stopped. This might	1115	An unspecified error has occurred, the watcher has been stopped. This might
569	happen because the watcher could not be properly started because libev	1116	happen because the watcher could not be properly started because libev
570	ran out of memory, a file descriptor was found to be closed or any other	1117	ran out of memory, a file descriptor was found to be closed or any other
		1118	problem. Libev considers these application bugs.
		1119
571	problem. You best act on it by reporting the problem and somehow coping	1120	You best act on it by reporting the problem and somehow coping with the
572	with the watcher being stopped.	1121	watcher being stopped. Note that well-written programs should not receive
		1122	an error ever, so when your watcher receives it, this usually indicates a
		1123	bug in your program.
573		1124
574	Libev will usually signal a few "dummy" events together with an error,	1125	Libev will usually signal a few "dummy" events together with an error, for
575	for example it might indicate that a fd is readable or writable, and if	1126	example it might indicate that a fd is readable or writable, and if your
576	your callbacks is well-written it can just attempt the operation and cope	1127	callbacks is well-written it can just attempt the operation and cope with
577	with the error from read() or write(). This will not work in multithreaded	1128	the error from read() or write(). This will not work in multi-threaded
578	programs, though, so beware.	1129	programs, though, as the fd could already be closed and reused for another
		1130	thing, so beware.
579		1131
580	=back	1132	=back
581		1133
582	=head2 GENERIC WATCHER FUNCTIONS	1134	=head2 GENERIC WATCHER FUNCTIONS
583
584	In the following description, C<TYPE> stands for the watcher type,
585	e.g. C<timer> for C<ev_timer> watchers and C<io> for C<ev_io> watchers.
586		1135
587	=over 4	1136	=over 4
588		1137
589	=item C<ev_init> (ev_TYPE *watcher, callback)	1138	=item C<ev_init> (ev_TYPE *watcher, callback)
590		1139
…		…
596	which rolls both calls into one.	1145	which rolls both calls into one.
597		1146
598	You can reinitialise a watcher at any time as long as it has been stopped	1147	You can reinitialise a watcher at any time as long as it has been stopped
599	(or never started) and there are no pending events outstanding.	1148	(or never started) and there are no pending events outstanding.
600		1149
601	The callback is always of type C<void (*)(ev_loop *loop, ev_TYPE *watcher,	1150	The callback is always of type C<void (*)(struct ev_loop *loop, ev_TYPE *watcher,
602	int revents)>.	1151	int revents)>.
603		1152
		1153	Example: Initialise an C<ev_io> watcher in two steps.
		1154
		1155	ev_io w;
		1156	ev_init (&w, my_cb);
		1157	ev_io_set (&w, STDIN_FILENO, EV_READ);
		1158
604	=item C<ev_TYPE_set> (ev_TYPE *, [args])	1159	=item C<ev_TYPE_set> (ev_TYPE *watcher, [args])
605		1160
606	This macro initialises the type-specific parts of a watcher. You need to	1161	This macro initialises the type-specific parts of a watcher. You need to
607	call C<ev_init> at least once before you call this macro, but you can	1162	call C<ev_init> at least once before you call this macro, but you can
608	call C<ev_TYPE_set> any number of times. You must not, however, call this	1163	call C<ev_TYPE_set> any number of times. You must not, however, call this
609	macro on a watcher that is active (it can be pending, however, which is a	1164	macro on a watcher that is active (it can be pending, however, which is a
610	difference to the C<ev_init> macro).	1165	difference to the C<ev_init> macro).
611		1166
612	Although some watcher types do not have type-specific arguments	1167	Although some watcher types do not have type-specific arguments
613	(e.g. C<ev_prepare>) you still need to call its C<set> macro.	1168	(e.g. C<ev_prepare>) you still need to call its C<set> macro.
614		1169
		1170	See C<ev_init>, above, for an example.
		1171
615	=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])	1172	=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])
616		1173
617	This convinience macro rolls both C<ev_init> and C<ev_TYPE_set> macro	1174	This convenience macro rolls both C<ev_init> and C<ev_TYPE_set> macro
618	calls into a single call. This is the most convinient method to initialise	1175	calls into a single call. This is the most convenient method to initialise
619	a watcher. The same limitations apply, of course.	1176	a watcher. The same limitations apply, of course.
620		1177
		1178	Example: Initialise and set an C<ev_io> watcher in one step.
		1179
		1180	ev_io_init (&w, my_cb, STDIN_FILENO, EV_READ);
		1181
621	=item C<ev_TYPE_start> (loop *, ev_TYPE *watcher)	1182	=item C<ev_TYPE_start> (loop, ev_TYPE *watcher)
622		1183
623	Starts (activates) the given watcher. Only active watchers will receive	1184	Starts (activates) the given watcher. Only active watchers will receive
624	events. If the watcher is already active nothing will happen.	1185	events. If the watcher is already active nothing will happen.
625		1186
		1187	Example: Start the C<ev_io> watcher that is being abused as example in this
		1188	whole section.
		1189
		1190	ev_io_start (EV_DEFAULT_UC, &w);
		1191
626	=item C<ev_TYPE_stop> (loop *, ev_TYPE *watcher)	1192	=item C<ev_TYPE_stop> (loop, ev_TYPE *watcher)
627		1193
628	Stops the given watcher again (if active) and clears the pending	1194	Stops the given watcher if active, and clears the pending status (whether
		1195	the watcher was active or not).
		1196
629	status. It is possible that stopped watchers are pending (for example,	1197	It is possible that stopped watchers are pending - for example,
630	non-repeating timers are being stopped when they become pending), but	1198	non-repeating timers are being stopped when they become pending - but
631	C<ev_TYPE_stop> ensures that the watcher is neither active nor pending. If	1199	calling C<ev_TYPE_stop> ensures that the watcher is neither active nor
632	you want to free or reuse the memory used by the watcher it is therefore a	1200	pending. If you want to free or reuse the memory used by the watcher it is
633	good idea to always call its C<ev_TYPE_stop> function.	1201	therefore a good idea to always call its C<ev_TYPE_stop> function.
634		1202
635	=item bool ev_is_active (ev_TYPE *watcher)	1203	=item bool ev_is_active (ev_TYPE *watcher)
636		1204
637	Returns a true value iff the watcher is active (i.e. it has been started	1205	Returns a true value iff the watcher is active (i.e. it has been started
638	and not yet been stopped). As long as a watcher is active you must not modify	1206	and not yet been stopped). As long as a watcher is active you must not modify
…		…
641	=item bool ev_is_pending (ev_TYPE *watcher)	1209	=item bool ev_is_pending (ev_TYPE *watcher)
642		1210
643	Returns a true value iff the watcher is pending, (i.e. it has outstanding	1211	Returns a true value iff the watcher is pending, (i.e. it has outstanding
644	events but its callback has not yet been invoked). As long as a watcher	1212	events but its callback has not yet been invoked). As long as a watcher
645	is pending (but not active) you must not call an init function on it (but	1213	is pending (but not active) you must not call an init function on it (but
646	C<ev_TYPE_set> is safe) and you must make sure the watcher is available to	1214	C<ev_TYPE_set> is safe), you must not change its priority, and you must
647	libev (e.g. you cnanot C<free ()> it).	1215	make sure the watcher is available to libev (e.g. you cannot C<free ()>
		1216	it).
648		1217
649	=item callback = ev_cb (ev_TYPE *watcher)	1218	=item callback ev_cb (ev_TYPE *watcher)
650		1219
651	Returns the callback currently set on the watcher.	1220	Returns the callback currently set on the watcher.
652		1221
653	=item ev_cb_set (ev_TYPE *watcher, callback)	1222	=item ev_cb_set (ev_TYPE *watcher, callback)
654		1223
655	Change the callback. You can change the callback at virtually any time	1224	Change the callback. You can change the callback at virtually any time
656	(modulo threads).	1225	(modulo threads).
657		1226
		1227	=item ev_set_priority (ev_TYPE *watcher, int priority)
		1228
		1229	=item int ev_priority (ev_TYPE *watcher)
		1230
		1231	Set and query the priority of the watcher. The priority is a small
		1232	integer between C<EV_MAXPRI> (default: C<2>) and C<EV_MINPRI>
		1233	(default: C<-2>). Pending watchers with higher priority will be invoked
		1234	before watchers with lower priority, but priority will not keep watchers
		1235	from being executed (except for C<ev_idle> watchers).
		1236
		1237	If you need to suppress invocation when higher priority events are pending
		1238	you need to look at C<ev_idle> watchers, which provide this functionality.
		1239
		1240	You I<must not> change the priority of a watcher as long as it is active or
		1241	pending.
		1242
		1243	Setting a priority outside the range of C<EV_MINPRI> to C<EV_MAXPRI> is
		1244	fine, as long as you do not mind that the priority value you query might
		1245	or might not have been clamped to the valid range.
		1246
		1247	The default priority used by watchers when no priority has been set is
		1248	always C<0>, which is supposed to not be too high and not be too low :).
		1249
		1250	See L<WATCHER PRIORITY MODELS>, below, for a more thorough treatment of
		1251	priorities.
		1252
		1253	=item ev_invoke (loop, ev_TYPE *watcher, int revents)
		1254
		1255	Invoke the C<watcher> with the given C<loop> and C<revents>. Neither
		1256	C<loop> nor C<revents> need to be valid as long as the watcher callback
		1257	can deal with that fact, as both are simply passed through to the
		1258	callback.
		1259
		1260	=item int ev_clear_pending (loop, ev_TYPE *watcher)
		1261
		1262	If the watcher is pending, this function clears its pending status and
		1263	returns its C<revents> bitset (as if its callback was invoked). If the
		1264	watcher isn't pending it does nothing and returns C<0>.
		1265
		1266	Sometimes it can be useful to "poll" a watcher instead of waiting for its
		1267	callback to be invoked, which can be accomplished with this function.
		1268
		1269	=item ev_feed_event (loop, ev_TYPE *watcher, int revents)
		1270
		1271	Feeds the given event set into the event loop, as if the specified event
		1272	had happened for the specified watcher (which must be a pointer to an
		1273	initialised but not necessarily started event watcher). Obviously you must
		1274	not free the watcher as long as it has pending events.
		1275
		1276	Stopping the watcher, letting libev invoke it, or calling
		1277	C<ev_clear_pending> will clear the pending event, even if the watcher was
		1278	not started in the first place.
		1279
		1280	See also C<ev_feed_fd_event> and C<ev_feed_signal_event> for related
		1281	functions that do not need a watcher.
		1282
658	=back	1283	=back
659		1284
660		1285
661	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER	1286	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER
662		1287
663	Each watcher has, by default, a member C<void *data> that you can change	1288	Each watcher has, by default, a member C<void *data> that you can change
664	and read at any time, libev will completely ignore it. This can be used	1289	and read at any time: libev will completely ignore it. This can be used
665	to associate arbitrary data with your watcher. If you need more data and	1290	to associate arbitrary data with your watcher. If you need more data and
666	don't want to allocate memory and store a pointer to it in that data	1291	don't want to allocate memory and store a pointer to it in that data
667	member, you can also "subclass" the watcher type and provide your own	1292	member, you can also "subclass" the watcher type and provide your own
668	data:	1293	data:
669		1294
670	struct my_io	1295	struct my_io
671	{	1296	{
672	struct ev_io io;	1297	ev_io io;
673	int otherfd;	1298	int otherfd;
674	void *somedata;	1299	void *somedata;
675	struct whatever *mostinteresting;	1300	struct whatever *mostinteresting;
676	}	1301	};
		1302
		1303	...
		1304	struct my_io w;
		1305	ev_io_init (&w.io, my_cb, fd, EV_READ);
677		1306
678	And since your callback will be called with a pointer to the watcher, you	1307	And since your callback will be called with a pointer to the watcher, you
679	can cast it back to your own type:	1308	can cast it back to your own type:
680		1309
681	static void my_cb (struct ev_loop *loop, struct ev_io *w_, int revents)	1310	static void my_cb (struct ev_loop *loop, ev_io *w_, int revents)
682	{	1311	{
683	struct my_io *w = (struct my_io *)w_;	1312	struct my_io *w = (struct my_io *)w_;
684	...	1313	...
685	}	1314	}
686		1315
687	More interesting and less C-conformant ways of catsing your callback type	1316	More interesting and less C-conformant ways of casting your callback type
688	have been omitted....	1317	instead have been omitted.
		1318
		1319	Another common scenario is to use some data structure with multiple
		1320	embedded watchers:
		1321
		1322	struct my_biggy
		1323	{
		1324	int some_data;
		1325	ev_timer t1;
		1326	ev_timer t2;
		1327	}
		1328
		1329	In this case getting the pointer to C<my_biggy> is a bit more
		1330	complicated: Either you store the address of your C<my_biggy> struct
		1331	in the C<data> member of the watcher (for woozies), or you need to use
		1332	some pointer arithmetic using C<offsetof> inside your watchers (for real
		1333	programmers):
		1334
		1335	#include <stddef.h>
		1336
		1337	static void
		1338	t1_cb (EV_P_ ev_timer *w, int revents)
		1339	{
		1340	struct my_biggy big = (struct my_biggy *)
		1341	(((char *)w) - offsetof (struct my_biggy, t1));
		1342	}
		1343
		1344	static void
		1345	t2_cb (EV_P_ ev_timer *w, int revents)
		1346	{
		1347	struct my_biggy big = (struct my_biggy *)
		1348	(((char *)w) - offsetof (struct my_biggy, t2));
		1349	}
		1350
		1351	=head2 WATCHER PRIORITY MODELS
		1352
		1353	Many event loops support I<watcher priorities>, which are usually small
		1354	integers that influence the ordering of event callback invocation
		1355	between watchers in some way, all else being equal.
		1356
		1357	In libev, Watcher priorities can be set using C<ev_set_priority>. See its
		1358	description for the more technical details such as the actual priority
		1359	range.
		1360
		1361	There are two common ways how these these priorities are being interpreted
		1362	by event loops:
		1363
		1364	In the more common lock-out model, higher priorities "lock out" invocation
		1365	of lower priority watchers, which means as long as higher priority
		1366	watchers receive events, lower priority watchers are not being invoked.
		1367
		1368	The less common only-for-ordering model uses priorities solely to order
		1369	callback invocation within a single event loop iteration: Higher priority
		1370	watchers are invoked before lower priority ones, but they all get invoked
		1371	before polling for new events.
		1372
		1373	Libev uses the second (only-for-ordering) model for all its watchers
		1374	except for idle watchers (which use the lock-out model).
		1375
		1376	The rationale behind this is that implementing the lock-out model for
		1377	watchers is not well supported by most kernel interfaces, and most event
		1378	libraries will just poll for the same events again and again as long as
		1379	their callbacks have not been executed, which is very inefficient in the
		1380	common case of one high-priority watcher locking out a mass of lower
		1381	priority ones.
		1382
		1383	Static (ordering) priorities are most useful when you have two or more
		1384	watchers handling the same resource: a typical usage example is having an
		1385	C<ev_io> watcher to receive data, and an associated C<ev_timer> to handle
		1386	timeouts. Under load, data might be received while the program handles
		1387	other jobs, but since timers normally get invoked first, the timeout
		1388	handler will be executed before checking for data. In that case, giving
		1389	the timer a lower priority than the I/O watcher ensures that I/O will be
		1390	handled first even under adverse conditions (which is usually, but not
		1391	always, what you want).
		1392
		1393	Since idle watchers use the "lock-out" model, meaning that idle watchers
		1394	will only be executed when no same or higher priority watchers have
		1395	received events, they can be used to implement the "lock-out" model when
		1396	required.
		1397
		1398	For example, to emulate how many other event libraries handle priorities,
		1399	you can associate an C<ev_idle> watcher to each such watcher, and in
		1400	the normal watcher callback, you just start the idle watcher. The real
		1401	processing is done in the idle watcher callback. This causes libev to
		1402	continuously poll and process kernel event data for the watcher, but when
		1403	the lock-out case is known to be rare (which in turn is rare :), this is
		1404	workable.
		1405
		1406	Usually, however, the lock-out model implemented that way will perform
		1407	miserably under the type of load it was designed to handle. In that case,
		1408	it might be preferable to stop the real watcher before starting the
		1409	idle watcher, so the kernel will not have to process the event in case
		1410	the actual processing will be delayed for considerable time.
		1411
		1412	Here is an example of an I/O watcher that should run at a strictly lower
		1413	priority than the default, and which should only process data when no
		1414	other events are pending:
		1415
		1416	ev_idle idle; // actual processing watcher
		1417	ev_io io; // actual event watcher
		1418
		1419	static void
		1420	io_cb (EV_P_ ev_io *w, int revents)
		1421	{
		1422	// stop the I/O watcher, we received the event, but
		1423	// are not yet ready to handle it.
		1424	ev_io_stop (EV_A_ w);
		1425
		1426	// start the idle watcher to handle the actual event.
		1427	// it will not be executed as long as other watchers
		1428	// with the default priority are receiving events.
		1429	ev_idle_start (EV_A_ &idle);
		1430	}
		1431
		1432	static void
		1433	idle_cb (EV_P_ ev_idle *w, int revents)
		1434	{
		1435	// actual processing
		1436	read (STDIN_FILENO, ...);
		1437
		1438	// have to start the I/O watcher again, as
		1439	// we have handled the event
		1440	ev_io_start (EV_P_ &io);
		1441	}
		1442
		1443	// initialisation
		1444	ev_idle_init (&idle, idle_cb);
		1445	ev_io_init (&io, io_cb, STDIN_FILENO, EV_READ);
		1446	ev_io_start (EV_DEFAULT_ &io);
		1447
		1448	In the "real" world, it might also be beneficial to start a timer, so that
		1449	low-priority connections can not be locked out forever under load. This
		1450	enables your program to keep a lower latency for important connections
		1451	during short periods of high load, while not completely locking out less
		1452	important ones.
689		1453
690		1454
691	=head1 WATCHER TYPES	1455	=head1 WATCHER TYPES
692		1456
693	This section describes each watcher in detail, but will not repeat	1457	This section describes each watcher in detail, but will not repeat
694	information given in the last section.	1458	information given in the last section. Any initialisation/set macros,
		1459	functions and members specific to the watcher type are explained.
		1460
		1461	Members are additionally marked with either I<[read-only]>, meaning that,
		1462	while the watcher is active, you can look at the member and expect some
		1463	sensible content, but you must not modify it (you can modify it while the
		1464	watcher is stopped to your hearts content), or I<[read-write]>, which
		1465	means you can expect it to have some sensible content while the watcher
		1466	is active, but you can also modify it. Modifying it may not do something
		1467	sensible or take immediate effect (or do anything at all), but libev will
		1468	not crash or malfunction in any way.
695		1469
696		1470
697	=head2 C<ev_io> - is this file descriptor readable or writable?	1471	=head2 C<ev_io> - is this file descriptor readable or writable?
698		1472
699	I/O watchers check whether a file descriptor is readable or writable	1473	I/O watchers check whether a file descriptor is readable or writable
…		…
707	In general you can register as many read and/or write event watchers per	1481	In general you can register as many read and/or write event watchers per
708	fd as you want (as long as you don't confuse yourself). Setting all file	1482	fd as you want (as long as you don't confuse yourself). Setting all file
709	descriptors to non-blocking mode is also usually a good idea (but not	1483	descriptors to non-blocking mode is also usually a good idea (but not
710	required if you know what you are doing).	1484	required if you know what you are doing).
711		1485
712	You have to be careful with dup'ed file descriptors, though. Some backends	1486	If you cannot use non-blocking mode, then force the use of a
713	(the linux epoll backend is a notable example) cannot handle dup'ed file	1487	known-to-be-good backend (at the time of this writing, this includes only
714	descriptors correctly if you register interest in two or more fds pointing	1488	C<EVBACKEND_SELECT> and C<EVBACKEND_POLL>). The same applies to file
715	to the same underlying file/socket/etc. description (that is, they share	1489	descriptors for which non-blocking operation makes no sense (such as
716	the same underlying "file open").	1490	files) - libev doesn't guarantee any specific behaviour in that case.
717
718	If you must do this, then force the use of a known-to-be-good backend
719	(at the time of this writing, this includes only C<EVBACKEND_SELECT> and
720	C<EVBACKEND_POLL>).
721		1491
722	Another thing you have to watch out for is that it is quite easy to	1492	Another thing you have to watch out for is that it is quite easy to
723	receive "spurious" readyness notifications, that is your callback might	1493	receive "spurious" readiness notifications, that is your callback might
724	be called with C<EV_READ> but a subsequent C<read>(2) will actually block	1494	be called with C<EV_READ> but a subsequent C<read>(2) will actually block
725	because there is no data. Not only are some backends known to create a	1495	because there is no data. Not only are some backends known to create a
726	lot of those (for example solaris ports), it is very easy to get into	1496	lot of those (for example Solaris ports), it is very easy to get into
727	this situation even with a relatively standard program structure. Thus	1497	this situation even with a relatively standard program structure. Thus
728	it is best to always use non-blocking I/O: An extra C<read>(2) returning	1498	it is best to always use non-blocking I/O: An extra C<read>(2) returning
729	C<EAGAIN> is far preferable to a program hanging until some data arrives.	1499	C<EAGAIN> is far preferable to a program hanging until some data arrives.
730		1500
731	If you cannot run the fd in non-blocking mode (for example you should not	1501	If you cannot run the fd in non-blocking mode (for example you should
732	play around with an Xlib connection), then you have to seperately re-test	1502	not play around with an Xlib connection), then you have to separately
733	wether a file descriptor is really ready with a known-to-be good interface	1503	re-test whether a file descriptor is really ready with a known-to-be good
734	such as poll (fortunately in our Xlib example, Xlib already does this on	1504	interface such as poll (fortunately in our Xlib example, Xlib already
735	its own, so its quite safe to use).	1505	does this on its own, so its quite safe to use). Some people additionally
		1506	use C<SIGALRM> and an interval timer, just to be sure you won't block
		1507	indefinitely.
		1508
		1509	But really, best use non-blocking mode.
		1510
		1511	=head3 The special problem of disappearing file descriptors
		1512
		1513	Some backends (e.g. kqueue, epoll) need to be told about closing a file
		1514	descriptor (either due to calling C<close> explicitly or any other means,
		1515	such as C<dup2>). The reason is that you register interest in some file
		1516	descriptor, but when it goes away, the operating system will silently drop
		1517	this interest. If another file descriptor with the same number then is
		1518	registered with libev, there is no efficient way to see that this is, in
		1519	fact, a different file descriptor.
		1520
		1521	To avoid having to explicitly tell libev about such cases, libev follows
		1522	the following policy: Each time C<ev_io_set> is being called, libev
		1523	will assume that this is potentially a new file descriptor, otherwise
		1524	it is assumed that the file descriptor stays the same. That means that
		1525	you I<have> to call C<ev_io_set> (or C<ev_io_init>) when you change the
		1526	descriptor even if the file descriptor number itself did not change.
		1527
		1528	This is how one would do it normally anyway, the important point is that
		1529	the libev application should not optimise around libev but should leave
		1530	optimisations to libev.
		1531
		1532	=head3 The special problem of dup'ed file descriptors
		1533
		1534	Some backends (e.g. epoll), cannot register events for file descriptors,
		1535	but only events for the underlying file descriptions. That means when you
		1536	have C<dup ()>'ed file descriptors or weirder constellations, and register
		1537	events for them, only one file descriptor might actually receive events.
		1538
		1539	There is no workaround possible except not registering events
		1540	for potentially C<dup ()>'ed file descriptors, or to resort to
		1541	C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>.
		1542
		1543	=head3 The special problem of fork
		1544
		1545	Some backends (epoll, kqueue) do not support C<fork ()> at all or exhibit
		1546	useless behaviour. Libev fully supports fork, but needs to be told about
		1547	it in the child.
		1548
		1549	To support fork in your programs, you either have to call
		1550	C<ev_default_fork ()> or C<ev_loop_fork ()> after a fork in the child,
		1551	enable C<EVFLAG_FORKCHECK>, or resort to C<EVBACKEND_SELECT> or
		1552	C<EVBACKEND_POLL>.
		1553
		1554	=head3 The special problem of SIGPIPE
		1555
		1556	While not really specific to libev, it is easy to forget about C<SIGPIPE>:
		1557	when writing to a pipe whose other end has been closed, your program gets
		1558	sent a SIGPIPE, which, by default, aborts your program. For most programs
		1559	this is sensible behaviour, for daemons, this is usually undesirable.
		1560
		1561	So when you encounter spurious, unexplained daemon exits, make sure you
		1562	ignore SIGPIPE (and maybe make sure you log the exit status of your daemon
		1563	somewhere, as that would have given you a big clue).
		1564
		1565	=head3 The special problem of accept()ing when you can't
		1566
		1567	Many implementations of the POSIX C<accept> function (for example,
		1568	found in post-2004 Linux) have the peculiar behaviour of not removing a
		1569	connection from the pending queue in all error cases.
		1570
		1571	For example, larger servers often run out of file descriptors (because
		1572	of resource limits), causing C<accept> to fail with C<ENFILE> but not
		1573	rejecting the connection, leading to libev signalling readiness on
		1574	the next iteration again (the connection still exists after all), and
		1575	typically causing the program to loop at 100% CPU usage.
		1576
		1577	Unfortunately, the set of errors that cause this issue differs between
		1578	operating systems, there is usually little the app can do to remedy the
		1579	situation, and no known thread-safe method of removing the connection to
		1580	cope with overload is known (to me).
		1581
		1582	One of the easiest ways to handle this situation is to just ignore it
		1583	- when the program encounters an overload, it will just loop until the
		1584	situation is over. While this is a form of busy waiting, no OS offers an
		1585	event-based way to handle this situation, so it's the best one can do.
		1586
		1587	A better way to handle the situation is to log any errors other than
		1588	C<EAGAIN> and C<EWOULDBLOCK>, making sure not to flood the log with such
		1589	messages, and continue as usual, which at least gives the user an idea of
		1590	what could be wrong ("raise the ulimit!"). For extra points one could stop
		1591	the C<ev_io> watcher on the listening fd "for a while", which reduces CPU
		1592	usage.
		1593
		1594	If your program is single-threaded, then you could also keep a dummy file
		1595	descriptor for overload situations (e.g. by opening F</dev/null>), and
		1596	when you run into C<ENFILE> or C<EMFILE>, close it, run C<accept>,
		1597	close that fd, and create a new dummy fd. This will gracefully refuse
		1598	clients under typical overload conditions.
		1599
		1600	The last way to handle it is to simply log the error and C<exit>, as
		1601	is often done with C<malloc> failures, but this results in an easy
		1602	opportunity for a DoS attack.
		1603
		1604	=head3 Watcher-Specific Functions
736		1605
737	=over 4	1606	=over 4
738		1607
739	=item ev_io_init (ev_io *, callback, int fd, int events)	1608	=item ev_io_init (ev_io *, callback, int fd, int events)
740		1609
741	=item ev_io_set (ev_io *, int fd, int events)	1610	=item ev_io_set (ev_io *, int fd, int events)
742		1611
743	Configures an C<ev_io> watcher. The C<fd> is the file descriptor to	1612	Configures an C<ev_io> watcher. The C<fd> is the file descriptor to
744	rceeive events for and events is either C<EV_READ>, C<EV_WRITE> or	1613	receive events for and C<events> is either C<EV_READ>, C<EV_WRITE> or
745	C<EV_READ | EV_WRITE> to receive the given events.	1614	C<EV_READ | EV_WRITE>, to express the desire to receive the given events.
		1615
		1616	=item int fd [read-only]
		1617
		1618	The file descriptor being watched.
		1619
		1620	=item int events [read-only]
		1621
		1622	The events being watched.
746		1623
747	=back	1624	=back
748		1625
		1626	=head3 Examples
		1627
749	Example: call C<stdin_readable_cb> when STDIN_FILENO has become, well	1628	Example: Call C<stdin_readable_cb> when STDIN_FILENO has become, well
750	readable, but only once. Since it is likely line-buffered, you could	1629	readable, but only once. Since it is likely line-buffered, you could
751	attempt to read a whole line in the callback:	1630	attempt to read a whole line in the callback.
752		1631
753	static void	1632	static void
754	stdin_readable_cb (struct ev_loop *loop, struct ev_io *w, int revents)	1633	stdin_readable_cb (struct ev_loop *loop, ev_io *w, int revents)
755	{	1634	{
756	ev_io_stop (loop, w);	1635	ev_io_stop (loop, w);
757	.. read from stdin here (or from w->fd) and haqndle any I/O errors	1636	.. read from stdin here (or from w->fd) and handle any I/O errors
758	}	1637	}
759		1638
760	...	1639	...
761	struct ev_loop *loop = ev_default_init (0);	1640	struct ev_loop *loop = ev_default_init (0);
762	struct ev_io stdin_readable;	1641	ev_io stdin_readable;
763	ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);	1642	ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
764	ev_io_start (loop, &stdin_readable);	1643	ev_io_start (loop, &stdin_readable);
765	ev_loop (loop, 0);	1644	ev_run (loop, 0);
766		1645
767		1646
768	=head2 C<ev_timer> - relative and optionally repeating timeouts	1647	=head2 C<ev_timer> - relative and optionally repeating timeouts
769		1648
770	Timer watchers are simple relative timers that generate an event after a	1649	Timer watchers are simple relative timers that generate an event after a
771	given time, and optionally repeating in regular intervals after that.	1650	given time, and optionally repeating in regular intervals after that.
772		1651
773	The timers are based on real time, that is, if you register an event that	1652	The timers are based on real time, that is, if you register an event that
774	times out after an hour and you reset your system clock to last years	1653	times out after an hour and you reset your system clock to January last
775	time, it will still time out after (roughly) and hour. "Roughly" because	1654	year, it will still time out after (roughly) one hour. "Roughly" because
776	detecting time jumps is hard, and some inaccuracies are unavoidable (the	1655	detecting time jumps is hard, and some inaccuracies are unavoidable (the
777	monotonic clock option helps a lot here).	1656	monotonic clock option helps a lot here).
		1657
		1658	The callback is guaranteed to be invoked only I<after> its timeout has
		1659	passed (not I<at>, so on systems with very low-resolution clocks this
		1660	might introduce a small delay). If multiple timers become ready during the
		1661	same loop iteration then the ones with earlier time-out values are invoked
		1662	before ones of the same priority with later time-out values (but this is
		1663	no longer true when a callback calls C<ev_run> recursively).
		1664
		1665	=head3 Be smart about timeouts
		1666
		1667	Many real-world problems involve some kind of timeout, usually for error
		1668	recovery. A typical example is an HTTP request - if the other side hangs,
		1669	you want to raise some error after a while.
		1670
		1671	What follows are some ways to handle this problem, from obvious and
		1672	inefficient to smart and efficient.
		1673
		1674	In the following, a 60 second activity timeout is assumed - a timeout that
		1675	gets reset to 60 seconds each time there is activity (e.g. each time some
		1676	data or other life sign was received).
		1677
		1678	=over 4
		1679
		1680	=item 1. Use a timer and stop, reinitialise and start it on activity.
		1681
		1682	This is the most obvious, but not the most simple way: In the beginning,
		1683	start the watcher:
		1684
		1685	ev_timer_init (timer, callback, 60., 0.);
		1686	ev_timer_start (loop, timer);
		1687
		1688	Then, each time there is some activity, C<ev_timer_stop> it, initialise it
		1689	and start it again:
		1690
		1691	ev_timer_stop (loop, timer);
		1692	ev_timer_set (timer, 60., 0.);
		1693	ev_timer_start (loop, timer);
		1694
		1695	This is relatively simple to implement, but means that each time there is
		1696	some activity, libev will first have to remove the timer from its internal
		1697	data structure and then add it again. Libev tries to be fast, but it's
		1698	still not a constant-time operation.
		1699
		1700	=item 2. Use a timer and re-start it with C<ev_timer_again> inactivity.
		1701
		1702	This is the easiest way, and involves using C<ev_timer_again> instead of
		1703	C<ev_timer_start>.
		1704
		1705	To implement this, configure an C<ev_timer> with a C<repeat> value
		1706	of C<60> and then call C<ev_timer_again> at start and each time you
		1707	successfully read or write some data. If you go into an idle state where
		1708	you do not expect data to travel on the socket, you can C<ev_timer_stop>
		1709	the timer, and C<ev_timer_again> will automatically restart it if need be.
		1710
		1711	That means you can ignore both the C<ev_timer_start> function and the
		1712	C<after> argument to C<ev_timer_set>, and only ever use the C<repeat>
		1713	member and C<ev_timer_again>.
		1714
		1715	At start:
		1716
		1717	ev_init (timer, callback);
		1718	timer->repeat = 60.;
		1719	ev_timer_again (loop, timer);
		1720
		1721	Each time there is some activity:
		1722
		1723	ev_timer_again (loop, timer);
		1724
		1725	It is even possible to change the time-out on the fly, regardless of
		1726	whether the watcher is active or not:
		1727
		1728	timer->repeat = 30.;
		1729	ev_timer_again (loop, timer);
		1730
		1731	This is slightly more efficient then stopping/starting the timer each time
		1732	you want to modify its timeout value, as libev does not have to completely
		1733	remove and re-insert the timer from/into its internal data structure.
		1734
		1735	It is, however, even simpler than the "obvious" way to do it.
		1736
		1737	=item 3. Let the timer time out, but then re-arm it as required.
		1738
		1739	This method is more tricky, but usually most efficient: Most timeouts are
		1740	relatively long compared to the intervals between other activity - in
		1741	our example, within 60 seconds, there are usually many I/O events with
		1742	associated activity resets.
		1743
		1744	In this case, it would be more efficient to leave the C<ev_timer> alone,
		1745	but remember the time of last activity, and check for a real timeout only
		1746	within the callback:
		1747
		1748	ev_tstamp last_activity; // time of last activity
		1749
		1750	static void
		1751	callback (EV_P_ ev_timer *w, int revents)
		1752	{
		1753	ev_tstamp now = ev_now (EV_A);
		1754	ev_tstamp timeout = last_activity + 60.;
		1755
		1756	// if last_activity + 60. is older than now, we did time out
		1757	if (timeout < now)
		1758	{
		1759	// timeout occurred, take action
		1760	}
		1761	else
		1762	{
		1763	// callback was invoked, but there was some activity, re-arm
		1764	// the watcher to fire in last_activity + 60, which is
		1765	// guaranteed to be in the future, so "again" is positive:
		1766	w->repeat = timeout - now;
		1767	ev_timer_again (EV_A_ w);
		1768	}
		1769	}
		1770
		1771	To summarise the callback: first calculate the real timeout (defined
		1772	as "60 seconds after the last activity"), then check if that time has
		1773	been reached, which means something I<did>, in fact, time out. Otherwise
		1774	the callback was invoked too early (C<timeout> is in the future), so
		1775	re-schedule the timer to fire at that future time, to see if maybe we have
		1776	a timeout then.
		1777
		1778	Note how C<ev_timer_again> is used, taking advantage of the
		1779	C<ev_timer_again> optimisation when the timer is already running.
		1780
		1781	This scheme causes more callback invocations (about one every 60 seconds
		1782	minus half the average time between activity), but virtually no calls to
		1783	libev to change the timeout.
		1784
		1785	To start the timer, simply initialise the watcher and set C<last_activity>
		1786	to the current time (meaning we just have some activity :), then call the
		1787	callback, which will "do the right thing" and start the timer:
		1788
		1789	ev_init (timer, callback);
		1790	last_activity = ev_now (loop);
		1791	callback (loop, timer, EV_TIMER);
		1792
		1793	And when there is some activity, simply store the current time in
		1794	C<last_activity>, no libev calls at all:
		1795
		1796	last_activity = ev_now (loop);
		1797
		1798	This technique is slightly more complex, but in most cases where the
		1799	time-out is unlikely to be triggered, much more efficient.
		1800
		1801	Changing the timeout is trivial as well (if it isn't hard-coded in the
		1802	callback :) - just change the timeout and invoke the callback, which will
		1803	fix things for you.
		1804
		1805	=item 4. Wee, just use a double-linked list for your timeouts.
		1806
		1807	If there is not one request, but many thousands (millions...), all
		1808	employing some kind of timeout with the same timeout value, then one can
		1809	do even better:
		1810
		1811	When starting the timeout, calculate the timeout value and put the timeout
		1812	at the I<end> of the list.
		1813
		1814	Then use an C<ev_timer> to fire when the timeout at the I<beginning> of
		1815	the list is expected to fire (for example, using the technique #3).
		1816
		1817	When there is some activity, remove the timer from the list, recalculate
		1818	the timeout, append it to the end of the list again, and make sure to
		1819	update the C<ev_timer> if it was taken from the beginning of the list.
		1820
		1821	This way, one can manage an unlimited number of timeouts in O(1) time for
		1822	starting, stopping and updating the timers, at the expense of a major
		1823	complication, and having to use a constant timeout. The constant timeout
		1824	ensures that the list stays sorted.
		1825
		1826	=back
		1827
		1828	So which method the best?
		1829
		1830	Method #2 is a simple no-brain-required solution that is adequate in most
		1831	situations. Method #3 requires a bit more thinking, but handles many cases
		1832	better, and isn't very complicated either. In most case, choosing either
		1833	one is fine, with #3 being better in typical situations.
		1834
		1835	Method #1 is almost always a bad idea, and buys you nothing. Method #4 is
		1836	rather complicated, but extremely efficient, something that really pays
		1837	off after the first million or so of active timers, i.e. it's usually
		1838	overkill :)
		1839
		1840	=head3 The special problem of time updates
		1841
		1842	Establishing the current time is a costly operation (it usually takes at
		1843	least two system calls): EV therefore updates its idea of the current
		1844	time only before and after C<ev_run> collects new events, which causes a
		1845	growing difference between C<ev_now ()> and C<ev_time ()> when handling
		1846	lots of events in one iteration.
778		1847
779	The relative timeouts are calculated relative to the C<ev_now ()>	1848	The relative timeouts are calculated relative to the C<ev_now ()>
780	time. This is usually the right thing as this timestamp refers to the time	1849	time. This is usually the right thing as this timestamp refers to the time
781	of the event triggering whatever timeout you are modifying/starting. If	1850	of the event triggering whatever timeout you are modifying/starting. If
782	you suspect event processing to be delayed and you I<need> to base the timeout	1851	you suspect event processing to be delayed and you I<need> to base the
783	on the current time, use something like this to adjust for this:	1852	timeout on the current time, use something like this to adjust for this:
784		1853
785	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);	1854	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);
786		1855
787	The callback is guarenteed to be invoked only when its timeout has passed,	1856	If the event loop is suspended for a long time, you can also force an
788	but if multiple timers become ready during the same loop iteration then	1857	update of the time returned by C<ev_now ()> by calling C<ev_now_update
789	order of execution is undefined.	1858	()>.
		1859
		1860	=head3 The special problems of suspended animation
		1861
		1862	When you leave the server world it is quite customary to hit machines that
		1863	can suspend/hibernate - what happens to the clocks during such a suspend?
		1864
		1865	Some quick tests made with a Linux 2.6.28 indicate that a suspend freezes
		1866	all processes, while the clocks (C<times>, C<CLOCK_MONOTONIC>) continue
		1867	to run until the system is suspended, but they will not advance while the
		1868	system is suspended. That means, on resume, it will be as if the program
		1869	was frozen for a few seconds, but the suspend time will not be counted
		1870	towards C<ev_timer> when a monotonic clock source is used. The real time
		1871	clock advanced as expected, but if it is used as sole clocksource, then a
		1872	long suspend would be detected as a time jump by libev, and timers would
		1873	be adjusted accordingly.
		1874
		1875	I would not be surprised to see different behaviour in different between
		1876	operating systems, OS versions or even different hardware.
		1877
		1878	The other form of suspend (job control, or sending a SIGSTOP) will see a
		1879	time jump in the monotonic clocks and the realtime clock. If the program
		1880	is suspended for a very long time, and monotonic clock sources are in use,
		1881	then you can expect C<ev_timer>s to expire as the full suspension time
		1882	will be counted towards the timers. When no monotonic clock source is in
		1883	use, then libev will again assume a timejump and adjust accordingly.
		1884
		1885	It might be beneficial for this latter case to call C<ev_suspend>
		1886	and C<ev_resume> in code that handles C<SIGTSTP>, to at least get
		1887	deterministic behaviour in this case (you can do nothing against
		1888	C<SIGSTOP>).
		1889
		1890	=head3 Watcher-Specific Functions and Data Members
790		1891
791	=over 4	1892	=over 4
792		1893
793	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)	1894	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)
794		1895
795	=item ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)	1896	=item ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)
796		1897
797	Configure the timer to trigger after C<after> seconds. If C<repeat> is	1898	Configure the timer to trigger after C<after> seconds. If C<repeat>
798	C<0.>, then it will automatically be stopped. If it is positive, then the	1899	is C<0.>, then it will automatically be stopped once the timeout is
799	timer will automatically be configured to trigger again C<repeat> seconds	1900	reached. If it is positive, then the timer will automatically be
800	later, again, and again, until stopped manually.	1901	configured to trigger again C<repeat> seconds later, again, and again,
		1902	until stopped manually.
801		1903
802	The timer itself will do a best-effort at avoiding drift, that is, if you	1904	The timer itself will do a best-effort at avoiding drift, that is, if
803	configure a timer to trigger every 10 seconds, then it will trigger at	1905	you configure a timer to trigger every 10 seconds, then it will normally
804	exactly 10 second intervals. If, however, your program cannot keep up with	1906	trigger at exactly 10 second intervals. If, however, your program cannot
805	the timer (because it takes longer than those 10 seconds to do stuff) the	1907	keep up with the timer (because it takes longer than those 10 seconds to
806	timer will not fire more than once per event loop iteration.	1908	do stuff) the timer will not fire more than once per event loop iteration.
807		1909
808	=item ev_timer_again (loop)	1910	=item ev_timer_again (loop, ev_timer *)
809		1911
810	This will act as if the timer timed out and restart it again if it is	1912	This will act as if the timer timed out and restart it again if it is
811	repeating. The exact semantics are:	1913	repeating. The exact semantics are:
812		1914
		1915	If the timer is pending, its pending status is cleared.
		1916
813	If the timer is started but nonrepeating, stop it.	1917	If the timer is started but non-repeating, stop it (as if it timed out).
814		1918
815	If the timer is repeating, either start it if necessary (with the repeat	1919	If the timer is repeating, either start it if necessary (with the
816	value), or reset the running timer to the repeat value.	1920	C<repeat> value), or reset the running timer to the C<repeat> value.
817		1921
818	This sounds a bit complicated, but here is a useful and typical	1922	This sounds a bit complicated, see L<Be smart about timeouts>, above, for a
819	example: Imagine you have a tcp connection and you want a so-called idle	1923	usage example.
820	timeout, that is, you want to be called when there have been, say, 60	1924
821	seconds of inactivity on the socket. The easiest way to do this is to	1925	=item ev_tstamp ev_timer_remaining (loop, ev_timer *)
822	configure an C<ev_timer> with after=repeat=60 and calling ev_timer_again each	1926
823	time you successfully read or write some data. If you go into an idle	1927	Returns the remaining time until a timer fires. If the timer is active,
824	state where you do not expect data to travel on the socket, you can stop	1928	then this time is relative to the current event loop time, otherwise it's
825	the timer, and again will automatically restart it if need be.	1929	the timeout value currently configured.
		1930
		1931	That is, after an C<ev_timer_set (w, 5, 7)>, C<ev_timer_remaining> returns
		1932	C<5>. When the timer is started and one second passes, C<ev_timer_remaining>
		1933	will return C<4>. When the timer expires and is restarted, it will return
		1934	roughly C<7> (likely slightly less as callback invocation takes some time,
		1935	too), and so on.
		1936
		1937	=item ev_tstamp repeat [read-write]
		1938
		1939	The current C<repeat> value. Will be used each time the watcher times out
		1940	or C<ev_timer_again> is called, and determines the next timeout (if any),
		1941	which is also when any modifications are taken into account.
826		1942
827	=back	1943	=back
828		1944
		1945	=head3 Examples
		1946
829	Example: create a timer that fires after 60 seconds.	1947	Example: Create a timer that fires after 60 seconds.
830		1948
831	static void	1949	static void
832	one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)	1950	one_minute_cb (struct ev_loop *loop, ev_timer *w, int revents)
833	{	1951	{
834	.. one minute over, w is actually stopped right here	1952	.. one minute over, w is actually stopped right here
835	}	1953	}
836		1954
837	struct ev_timer mytimer;	1955	ev_timer mytimer;
838	ev_timer_init (&mytimer, one_minute_cb, 60., 0.);	1956	ev_timer_init (&mytimer, one_minute_cb, 60., 0.);
839	ev_timer_start (loop, &mytimer);	1957	ev_timer_start (loop, &mytimer);
840		1958
841	Example: create a timeout timer that times out after 10 seconds of	1959	Example: Create a timeout timer that times out after 10 seconds of
842	inactivity.	1960	inactivity.
843		1961
844	static void	1962	static void
845	timeout_cb (struct ev_loop *loop, struct ev_timer *w, int revents)	1963	timeout_cb (struct ev_loop *loop, ev_timer *w, int revents)
846	{	1964	{
847	.. ten seconds without any activity	1965	.. ten seconds without any activity
848	}	1966	}
849		1967
850	struct ev_timer mytimer;	1968	ev_timer mytimer;
851	ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */	1969	ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */
852	ev_timer_again (&mytimer); /* start timer */	1970	ev_timer_again (&mytimer); /* start timer */
853	ev_loop (loop, 0);	1971	ev_run (loop, 0);
854		1972
855	// and in some piece of code that gets executed on any "activity":	1973	// and in some piece of code that gets executed on any "activity":
856	// reset the timeout to start ticking again at 10 seconds	1974	// reset the timeout to start ticking again at 10 seconds
857	ev_timer_again (&mytimer);	1975	ev_timer_again (&mytimer);
858		1976
859		1977
860	=head2 C<ev_periodic> - to cron or not to cron?	1978	=head2 C<ev_periodic> - to cron or not to cron?
861		1979
862	Periodic watchers are also timers of a kind, but they are very versatile	1980	Periodic watchers are also timers of a kind, but they are very versatile
863	(and unfortunately a bit complex).	1981	(and unfortunately a bit complex).
864		1982
865	Unlike C<ev_timer>'s, they are not based on real time (or relative time)	1983	Unlike C<ev_timer>, periodic watchers are not based on real time (or
866	but on wallclock time (absolute time). You can tell a periodic watcher	1984	relative time, the physical time that passes) but on wall clock time
867	to trigger "at" some specific point in time. For example, if you tell a	1985	(absolute time, the thing you can read on your calender or clock). The
868	periodic watcher to trigger in 10 seconds (by specifiying e.g. C<ev_now ()	1986	difference is that wall clock time can run faster or slower than real
869	+ 10.>) and then reset your system clock to the last year, then it will	1987	time, and time jumps are not uncommon (e.g. when you adjust your
870	take a year to trigger the event (unlike an C<ev_timer>, which would trigger	1988	wrist-watch).
871	roughly 10 seconds later and of course not if you reset your system time
872	again).
873		1989
874	They can also be used to implement vastly more complex timers, such as	1990	You can tell a periodic watcher to trigger after some specific point
		1991	in time: for example, if you tell a periodic watcher to trigger "in 10
		1992	seconds" (by specifying e.g. C<ev_now () + 10.>, that is, an absolute time
		1993	not a delay) and then reset your system clock to January of the previous
		1994	year, then it will take a year or more to trigger the event (unlike an
		1995	C<ev_timer>, which would still trigger roughly 10 seconds after starting
		1996	it, as it uses a relative timeout).
		1997
		1998	C<ev_periodic> watchers can also be used to implement vastly more complex
875	triggering an event on eahc midnight, local time.	1999	timers, such as triggering an event on each "midnight, local time", or
		2000	other complicated rules. This cannot be done with C<ev_timer> watchers, as
		2001	those cannot react to time jumps.
876		2002
877	As with timers, the callback is guarenteed to be invoked only when the	2003	As with timers, the callback is guaranteed to be invoked only when the
878	time (C<at>) has been passed, but if multiple periodic timers become ready	2004	point in time where it is supposed to trigger has passed. If multiple
879	during the same loop iteration then order of execution is undefined.	2005	timers become ready during the same loop iteration then the ones with
		2006	earlier time-out values are invoked before ones with later time-out values
		2007	(but this is no longer true when a callback calls C<ev_run> recursively).
		2008
		2009	=head3 Watcher-Specific Functions and Data Members
880		2010
881	=over 4	2011	=over 4
882		2012
883	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)	2013	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp offset, ev_tstamp interval, reschedule_cb)
884		2014
885	=item ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)	2015	=item ev_periodic_set (ev_periodic *, ev_tstamp offset, ev_tstamp interval, reschedule_cb)
886		2016
887	Lots of arguments, lets sort it out... There are basically three modes of	2017	Lots of arguments, let's sort it out... There are basically three modes of
888	operation, and we will explain them from simplest to complex:	2018	operation, and we will explain them from simplest to most complex:
889		2019
890	=over 4	2020	=over 4
891		2021
892	=item * absolute timer (interval = reschedule_cb = 0)	2022	=item * absolute timer (offset = absolute time, interval = 0, reschedule_cb = 0)
893		2023
894	In this configuration the watcher triggers an event at the wallclock time	2024	In this configuration the watcher triggers an event after the wall clock
895	C<at> and doesn't repeat. It will not adjust when a time jump occurs,	2025	time C<offset> has passed. It will not repeat and will not adjust when a
896	that is, if it is to be run at January 1st 2011 then it will run when the	2026	time jump occurs, that is, if it is to be run at January 1st 2011 then it
897	system time reaches or surpasses this time.	2027	will be stopped and invoked when the system clock reaches or surpasses
		2028	this point in time.
898		2029
899	=item * non-repeating interval timer (interval > 0, reschedule_cb = 0)	2030	=item * repeating interval timer (offset = offset within interval, interval > 0, reschedule_cb = 0)
900		2031
901	In this mode the watcher will always be scheduled to time out at the next	2032	In this mode the watcher will always be scheduled to time out at the next
902	C<at + N * interval> time (for some integer N) and then repeat, regardless	2033	C<offset + N * interval> time (for some integer N, which can also be
903	of any time jumps.	2034	negative) and then repeat, regardless of any time jumps. The C<offset>
		2035	argument is merely an offset into the C<interval> periods.
904		2036
905	This can be used to create timers that do not drift with respect to system	2037	This can be used to create timers that do not drift with respect to the
906	time:	2038	system clock, for example, here is an C<ev_periodic> that triggers each
		2039	hour, on the hour (with respect to UTC):
907		2040
908	ev_periodic_set (&periodic, 0., 3600., 0);	2041	ev_periodic_set (&periodic, 0., 3600., 0);
909		2042
910	This doesn't mean there will always be 3600 seconds in between triggers,	2043	This doesn't mean there will always be 3600 seconds in between triggers,
911	but only that the the callback will be called when the system time shows a	2044	but only that the callback will be called when the system time shows a
912	full hour (UTC), or more correctly, when the system time is evenly divisible	2045	full hour (UTC), or more correctly, when the system time is evenly divisible
913	by 3600.	2046	by 3600.
914		2047
915	Another way to think about it (for the mathematically inclined) is that	2048	Another way to think about it (for the mathematically inclined) is that
916	C<ev_periodic> will try to run the callback in this mode at the next possible	2049	C<ev_periodic> will try to run the callback in this mode at the next possible
917	time where C<time = at (mod interval)>, regardless of any time jumps.	2050	time where C<time = offset (mod interval)>, regardless of any time jumps.
918		2051
		2052	For numerical stability it is preferable that the C<offset> value is near
		2053	C<ev_now ()> (the current time), but there is no range requirement for
		2054	this value, and in fact is often specified as zero.
		2055
		2056	Note also that there is an upper limit to how often a timer can fire (CPU
		2057	speed for example), so if C<interval> is very small then timing stability
		2058	will of course deteriorate. Libev itself tries to be exact to be about one
		2059	millisecond (if the OS supports it and the machine is fast enough).
		2060
919	=item * manual reschedule mode (reschedule_cb = callback)	2061	=item * manual reschedule mode (offset ignored, interval ignored, reschedule_cb = callback)
920		2062
921	In this mode the values for C<interval> and C<at> are both being	2063	In this mode the values for C<interval> and C<offset> are both being
922	ignored. Instead, each time the periodic watcher gets scheduled, the	2064	ignored. Instead, each time the periodic watcher gets scheduled, the
923	reschedule callback will be called with the watcher as first, and the	2065	reschedule callback will be called with the watcher as first, and the
924	current time as second argument.	2066	current time as second argument.
925		2067
926	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher,	2068	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher, ever,
927	ever, or make any event loop modifications>. If you need to stop it,	2069	or make ANY other event loop modifications whatsoever, unless explicitly
928	return C<now + 1e30> (or so, fudge fudge) and stop it afterwards (e.g. by	2070	allowed by documentation here>.
929	starting a prepare watcher).
930		2071
		2072	If you need to stop it, return C<now + 1e30> (or so, fudge fudge) and stop
		2073	it afterwards (e.g. by starting an C<ev_prepare> watcher, which is the
		2074	only event loop modification you are allowed to do).
		2075
931	Its prototype is C<ev_tstamp (*reschedule_cb)(struct ev_periodic *w,	2076	The callback prototype is C<ev_tstamp (*reschedule_cb)(ev_periodic
932	ev_tstamp now)>, e.g.:	2077	*w, ev_tstamp now)>, e.g.:
933		2078
		2079	static ev_tstamp
934	static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)	2080	my_rescheduler (ev_periodic *w, ev_tstamp now)
935	{	2081	{
936	return now + 60.;	2082	return now + 60.;
937	}	2083	}
938		2084
939	It must return the next time to trigger, based on the passed time value	2085	It must return the next time to trigger, based on the passed time value
940	(that is, the lowest time value larger than to the second argument). It	2086	(that is, the lowest time value larger than to the second argument). It
941	will usually be called just before the callback will be triggered, but	2087	will usually be called just before the callback will be triggered, but
942	might be called at other times, too.	2088	might be called at other times, too.
943		2089
944	NOTE: I<< This callback must always return a time that is later than the	2090	NOTE: I<< This callback must always return a time that is higher than or
945	passed C<now> value >>. Not even C<now> itself will do, it I<must> be larger.	2091	equal to the passed C<now> value >>.
946		2092
947	This can be used to create very complex timers, such as a timer that	2093	This can be used to create very complex timers, such as a timer that
948	triggers on each midnight, local time. To do this, you would calculate the	2094	triggers on "next midnight, local time". To do this, you would calculate the
949	next midnight after C<now> and return the timestamp value for this. How	2095	next midnight after C<now> and return the timestamp value for this. How
950	you do this is, again, up to you (but it is not trivial, which is the main	2096	you do this is, again, up to you (but it is not trivial, which is the main
951	reason I omitted it as an example).	2097	reason I omitted it as an example).
952		2098
953	=back	2099	=back
…		…
957	Simply stops and restarts the periodic watcher again. This is only useful	2103	Simply stops and restarts the periodic watcher again. This is only useful
958	when you changed some parameters or the reschedule callback would return	2104	when you changed some parameters or the reschedule callback would return
959	a different time than the last time it was called (e.g. in a crond like	2105	a different time than the last time it was called (e.g. in a crond like
960	program when the crontabs have changed).	2106	program when the crontabs have changed).
961		2107
		2108	=item ev_tstamp ev_periodic_at (ev_periodic *)
		2109
		2110	When active, returns the absolute time that the watcher is supposed
		2111	to trigger next. This is not the same as the C<offset> argument to
		2112	C<ev_periodic_set>, but indeed works even in interval and manual
		2113	rescheduling modes.
		2114
		2115	=item ev_tstamp offset [read-write]
		2116
		2117	When repeating, this contains the offset value, otherwise this is the
		2118	absolute point in time (the C<offset> value passed to C<ev_periodic_set>,
		2119	although libev might modify this value for better numerical stability).
		2120
		2121	Can be modified any time, but changes only take effect when the periodic
		2122	timer fires or C<ev_periodic_again> is being called.
		2123
		2124	=item ev_tstamp interval [read-write]
		2125
		2126	The current interval value. Can be modified any time, but changes only
		2127	take effect when the periodic timer fires or C<ev_periodic_again> is being
		2128	called.
		2129
		2130	=item ev_tstamp (*reschedule_cb)(ev_periodic *w, ev_tstamp now) [read-write]
		2131
		2132	The current reschedule callback, or C<0>, if this functionality is
		2133	switched off. Can be changed any time, but changes only take effect when
		2134	the periodic timer fires or C<ev_periodic_again> is being called.
		2135
962	=back	2136	=back
963		2137
		2138	=head3 Examples
		2139
964	Example: call a callback every hour, or, more precisely, whenever the	2140	Example: Call a callback every hour, or, more precisely, whenever the
965	system clock is divisible by 3600. The callback invocation times have	2141	system time is divisible by 3600. The callback invocation times have
966	potentially a lot of jittering, but good long-term stability.	2142	potentially a lot of jitter, but good long-term stability.
967		2143
968	static void	2144	static void
969	clock_cb (struct ev_loop *loop, struct ev_io *w, int revents)	2145	clock_cb (struct ev_loop *loop, ev_periodic *w, int revents)
970	{	2146	{
971	... its now a full hour (UTC, or TAI or whatever your clock follows)	2147	... its now a full hour (UTC, or TAI or whatever your clock follows)
972	}	2148	}
973		2149
974	struct ev_periodic hourly_tick;	2150	ev_periodic hourly_tick;
975	ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);	2151	ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);
976	ev_periodic_start (loop, &hourly_tick);	2152	ev_periodic_start (loop, &hourly_tick);
977		2153
978	Example: the same as above, but use a reschedule callback to do it:	2154	Example: The same as above, but use a reschedule callback to do it:
979		2155
980	#include <math.h>	2156	#include <math.h>
981		2157
982	static ev_tstamp	2158	static ev_tstamp
983	my_scheduler_cb (struct ev_periodic *w, ev_tstamp now)	2159	my_scheduler_cb (ev_periodic *w, ev_tstamp now)
984	{	2160	{
985	return fmod (now, 3600.) + 3600.;	2161	return now + (3600. - fmod (now, 3600.));
986	}	2162	}
987		2163
988	ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);	2164	ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
989		2165
990	Example: call a callback every hour, starting now:	2166	Example: Call a callback every hour, starting now:
991		2167
992	struct ev_periodic hourly_tick;	2168	ev_periodic hourly_tick;
993	ev_periodic_init (&hourly_tick, clock_cb,	2169	ev_periodic_init (&hourly_tick, clock_cb,
994	fmod (ev_now (loop), 3600.), 3600., 0);	2170	fmod (ev_now (loop), 3600.), 3600., 0);
995	ev_periodic_start (loop, &hourly_tick);	2171	ev_periodic_start (loop, &hourly_tick);
996		2172
997		2173
998	=head2 C<ev_signal> - signal me when a signal gets signalled!	2174	=head2 C<ev_signal> - signal me when a signal gets signalled!
999		2175
1000	Signal watchers will trigger an event when the process receives a specific	2176	Signal watchers will trigger an event when the process receives a specific
1001	signal one or more times. Even though signals are very asynchronous, libev	2177	signal one or more times. Even though signals are very asynchronous, libev
1002	will try it's best to deliver signals synchronously, i.e. as part of the	2178	will try it's best to deliver signals synchronously, i.e. as part of the
1003	normal event processing, like any other event.	2179	normal event processing, like any other event.
1004		2180
		2181	If you want signals to be delivered truly asynchronously, just use
		2182	C<sigaction> as you would do without libev and forget about sharing
		2183	the signal. You can even use C<ev_async> from a signal handler to
		2184	synchronously wake up an event loop.
		2185
1005	You can configure as many watchers as you like per signal. Only when the	2186	You can configure as many watchers as you like for the same signal, but
		2187	only within the same loop, i.e. you can watch for C<SIGINT> in your
		2188	default loop and for C<SIGIO> in another loop, but you cannot watch for
		2189	C<SIGINT> in both the default loop and another loop at the same time. At
		2190	the moment, C<SIGCHLD> is permanently tied to the default loop.
		2191
1006	first watcher gets started will libev actually register a signal watcher	2192	When the first watcher gets started will libev actually register something
1007	with the kernel (thus it coexists with your own signal handlers as long	2193	with the kernel (thus it coexists with your own signal handlers as long as
1008	as you don't register any with libev). Similarly, when the last signal	2194	you don't register any with libev for the same signal).
1009	watcher for a signal is stopped libev will reset the signal handler to	2195
1010	SIG_DFL (regardless of what it was set to before).	2196	If possible and supported, libev will install its handlers with
		2197	C<SA_RESTART> (or equivalent) behaviour enabled, so system calls should
		2198	not be unduly interrupted. If you have a problem with system calls getting
		2199	interrupted by signals you can block all signals in an C<ev_check> watcher
		2200	and unblock them in an C<ev_prepare> watcher.
		2201
		2202	=head3 The special problem of inheritance over fork/execve/pthread_create
		2203
		2204	Both the signal mask (C<sigprocmask>) and the signal disposition
		2205	(C<sigaction>) are unspecified after starting a signal watcher (and after
		2206	stopping it again), that is, libev might or might not block the signal,
		2207	and might or might not set or restore the installed signal handler.
		2208
		2209	While this does not matter for the signal disposition (libev never
		2210	sets signals to C<SIG_IGN>, so handlers will be reset to C<SIG_DFL> on
		2211	C<execve>), this matters for the signal mask: many programs do not expect
		2212	certain signals to be blocked.
		2213
		2214	This means that before calling C<exec> (from the child) you should reset
		2215	the signal mask to whatever "default" you expect (all clear is a good
		2216	choice usually).
		2217
		2218	The simplest way to ensure that the signal mask is reset in the child is
		2219	to install a fork handler with C<pthread_atfork> that resets it. That will
		2220	catch fork calls done by libraries (such as the libc) as well.
		2221
		2222	In current versions of libev, the signal will not be blocked indefinitely
		2223	unless you use the C<signalfd> API (C<EV_SIGNALFD>). While this reduces
		2224	the window of opportunity for problems, it will not go away, as libev
		2225	I<has> to modify the signal mask, at least temporarily.
		2226
		2227	So I can't stress this enough: I<If you do not reset your signal mask when
		2228	you expect it to be empty, you have a race condition in your code>. This
		2229	is not a libev-specific thing, this is true for most event libraries.
		2230
		2231	=head3 Watcher-Specific Functions and Data Members
1011		2232
1012	=over 4	2233	=over 4
1013		2234
1014	=item ev_signal_init (ev_signal *, callback, int signum)	2235	=item ev_signal_init (ev_signal *, callback, int signum)
1015		2236
1016	=item ev_signal_set (ev_signal *, int signum)	2237	=item ev_signal_set (ev_signal *, int signum)
1017		2238
1018	Configures the watcher to trigger on the given signal number (usually one	2239	Configures the watcher to trigger on the given signal number (usually one
1019	of the C<SIGxxx> constants).	2240	of the C<SIGxxx> constants).
1020		2241
		2242	=item int signum [read-only]
		2243
		2244	The signal the watcher watches out for.
		2245
1021	=back	2246	=back
1022		2247
		2248	=head3 Examples
		2249
		2250	Example: Try to exit cleanly on SIGINT.
		2251
		2252	static void
		2253	sigint_cb (struct ev_loop *loop, ev_signal *w, int revents)
		2254	{
		2255	ev_break (loop, EVBREAK_ALL);
		2256	}
		2257
		2258	ev_signal signal_watcher;
		2259	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
		2260	ev_signal_start (loop, &signal_watcher);
		2261
1023		2262
1024	=head2 C<ev_child> - watch out for process status changes	2263	=head2 C<ev_child> - watch out for process status changes
1025		2264
1026	Child watchers trigger when your process receives a SIGCHLD in response to	2265	Child watchers trigger when your process receives a SIGCHLD in response to
1027	some child status changes (most typically when a child of yours dies).	2266	some child status changes (most typically when a child of yours dies or
		2267	exits). It is permissible to install a child watcher I<after> the child
		2268	has been forked (which implies it might have already exited), as long
		2269	as the event loop isn't entered (or is continued from a watcher), i.e.,
		2270	forking and then immediately registering a watcher for the child is fine,
		2271	but forking and registering a watcher a few event loop iterations later or
		2272	in the next callback invocation is not.
		2273
		2274	Only the default event loop is capable of handling signals, and therefore
		2275	you can only register child watchers in the default event loop.
		2276
		2277	Due to some design glitches inside libev, child watchers will always be
		2278	handled at maximum priority (their priority is set to C<EV_MAXPRI> by
		2279	libev)
		2280
		2281	=head3 Process Interaction
		2282
		2283	Libev grabs C<SIGCHLD> as soon as the default event loop is
		2284	initialised. This is necessary to guarantee proper behaviour even if the
		2285	first child watcher is started after the child exits. The occurrence
		2286	of C<SIGCHLD> is recorded asynchronously, but child reaping is done
		2287	synchronously as part of the event loop processing. Libev always reaps all
		2288	children, even ones not watched.
		2289
		2290	=head3 Overriding the Built-In Processing
		2291
		2292	Libev offers no special support for overriding the built-in child
		2293	processing, but if your application collides with libev's default child
		2294	handler, you can override it easily by installing your own handler for
		2295	C<SIGCHLD> after initialising the default loop, and making sure the
		2296	default loop never gets destroyed. You are encouraged, however, to use an
		2297	event-based approach to child reaping and thus use libev's support for
		2298	that, so other libev users can use C<ev_child> watchers freely.
		2299
		2300	=head3 Stopping the Child Watcher
		2301
		2302	Currently, the child watcher never gets stopped, even when the
		2303	child terminates, so normally one needs to stop the watcher in the
		2304	callback. Future versions of libev might stop the watcher automatically
		2305	when a child exit is detected (calling C<ev_child_stop> twice is not a
		2306	problem).
		2307
		2308	=head3 Watcher-Specific Functions and Data Members
1028		2309
1029	=over 4	2310	=over 4
1030		2311
1031	=item ev_child_init (ev_child *, callback, int pid)	2312	=item ev_child_init (ev_child *, callback, int pid, int trace)
1032		2313
1033	=item ev_child_set (ev_child *, int pid)	2314	=item ev_child_set (ev_child *, int pid, int trace)
1034		2315
1035	Configures the watcher to wait for status changes of process C<pid> (or	2316	Configures the watcher to wait for status changes of process C<pid> (or
1036	I<any> process if C<pid> is specified as C<0>). The callback can look	2317	I<any> process if C<pid> is specified as C<0>). The callback can look
1037	at the C<rstatus> member of the C<ev_child> watcher structure to see	2318	at the C<rstatus> member of the C<ev_child> watcher structure to see
1038	the status word (use the macros from C<sys/wait.h> and see your systems	2319	the status word (use the macros from C<sys/wait.h> and see your systems
1039	C<waitpid> documentation). The C<rpid> member contains the pid of the	2320	C<waitpid> documentation). The C<rpid> member contains the pid of the
1040	process causing the status change.	2321	process causing the status change. C<trace> must be either C<0> (only
		2322	activate the watcher when the process terminates) or C<1> (additionally
		2323	activate the watcher when the process is stopped or continued).
		2324
		2325	=item int pid [read-only]
		2326
		2327	The process id this watcher watches out for, or C<0>, meaning any process id.
		2328
		2329	=item int rpid [read-write]
		2330
		2331	The process id that detected a status change.
		2332
		2333	=item int rstatus [read-write]
		2334
		2335	The process exit/trace status caused by C<rpid> (see your systems
		2336	C<waitpid> and C<sys/wait.h> documentation for details).
1041		2337
1042	=back	2338	=back
1043		2339
1044	Example: try to exit cleanly on SIGINT and SIGTERM.	2340	=head3 Examples
1045		2341
		2342	Example: C<fork()> a new process and install a child handler to wait for
		2343	its completion.
		2344
		2345	ev_child cw;
		2346
1046	static void	2347	static void
1047	sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)	2348	child_cb (EV_P_ ev_child *w, int revents)
1048	{	2349	{
1049	ev_unloop (loop, EVUNLOOP_ALL);	2350	ev_child_stop (EV_A_ w);
		2351	printf ("process %d exited with status %x\n", w->rpid, w->rstatus);
1050	}	2352	}
1051		2353
1052	struct ev_signal signal_watcher;	2354	pid_t pid = fork ();
1053	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);	2355
1054	ev_signal_start (loop, &sigint_cb);	2356	if (pid < 0)
		2357	// error
		2358	else if (pid == 0)
		2359	{
		2360	// the forked child executes here
		2361	exit (1);
		2362	}
		2363	else
		2364	{
		2365	ev_child_init (&cw, child_cb, pid, 0);
		2366	ev_child_start (EV_DEFAULT_ &cw);
		2367	}
		2368
		2369
		2370	=head2 C<ev_stat> - did the file attributes just change?
		2371
		2372	This watches a file system path for attribute changes. That is, it calls
		2373	C<stat> on that path in regular intervals (or when the OS says it changed)
		2374	and sees if it changed compared to the last time, invoking the callback if
		2375	it did.
		2376
		2377	The path does not need to exist: changing from "path exists" to "path does
		2378	not exist" is a status change like any other. The condition "path does not
		2379	exist" (or more correctly "path cannot be stat'ed") is signified by the
		2380	C<st_nlink> field being zero (which is otherwise always forced to be at
		2381	least one) and all the other fields of the stat buffer having unspecified
		2382	contents.
		2383
		2384	The path I<must not> end in a slash or contain special components such as
		2385	C<.> or C<..>. The path I<should> be absolute: If it is relative and
		2386	your working directory changes, then the behaviour is undefined.
		2387
		2388	Since there is no portable change notification interface available, the
		2389	portable implementation simply calls C<stat(2)> regularly on the path
		2390	to see if it changed somehow. You can specify a recommended polling
		2391	interval for this case. If you specify a polling interval of C<0> (highly
		2392	recommended!) then a I<suitable, unspecified default> value will be used
		2393	(which you can expect to be around five seconds, although this might
		2394	change dynamically). Libev will also impose a minimum interval which is
		2395	currently around C<0.1>, but that's usually overkill.
		2396
		2397	This watcher type is not meant for massive numbers of stat watchers,
		2398	as even with OS-supported change notifications, this can be
		2399	resource-intensive.
		2400
		2401	At the time of this writing, the only OS-specific interface implemented
		2402	is the Linux inotify interface (implementing kqueue support is left as an
		2403	exercise for the reader. Note, however, that the author sees no way of
		2404	implementing C<ev_stat> semantics with kqueue, except as a hint).
		2405
		2406	=head3 ABI Issues (Largefile Support)
		2407
		2408	Libev by default (unless the user overrides this) uses the default
		2409	compilation environment, which means that on systems with large file
		2410	support disabled by default, you get the 32 bit version of the stat
		2411	structure. When using the library from programs that change the ABI to
		2412	use 64 bit file offsets the programs will fail. In that case you have to
		2413	compile libev with the same flags to get binary compatibility. This is
		2414	obviously the case with any flags that change the ABI, but the problem is
		2415	most noticeably displayed with ev_stat and large file support.
		2416
		2417	The solution for this is to lobby your distribution maker to make large
		2418	file interfaces available by default (as e.g. FreeBSD does) and not
		2419	optional. Libev cannot simply switch on large file support because it has
		2420	to exchange stat structures with application programs compiled using the
		2421	default compilation environment.
		2422
		2423	=head3 Inotify and Kqueue
		2424
		2425	When C<inotify (7)> support has been compiled into libev and present at
		2426	runtime, it will be used to speed up change detection where possible. The
		2427	inotify descriptor will be created lazily when the first C<ev_stat>
		2428	watcher is being started.
		2429
		2430	Inotify presence does not change the semantics of C<ev_stat> watchers
		2431	except that changes might be detected earlier, and in some cases, to avoid
		2432	making regular C<stat> calls. Even in the presence of inotify support
		2433	there are many cases where libev has to resort to regular C<stat> polling,
		2434	but as long as kernel 2.6.25 or newer is used (2.6.24 and older have too
		2435	many bugs), the path exists (i.e. stat succeeds), and the path resides on
		2436	a local filesystem (libev currently assumes only ext2/3, jfs, reiserfs and
		2437	xfs are fully working) libev usually gets away without polling.
		2438
		2439	There is no support for kqueue, as apparently it cannot be used to
		2440	implement this functionality, due to the requirement of having a file
		2441	descriptor open on the object at all times, and detecting renames, unlinks
		2442	etc. is difficult.
		2443
		2444	=head3 C<stat ()> is a synchronous operation
		2445
		2446	Libev doesn't normally do any kind of I/O itself, and so is not blocking
		2447	the process. The exception are C<ev_stat> watchers - those call C<stat
		2448	()>, which is a synchronous operation.
		2449
		2450	For local paths, this usually doesn't matter: unless the system is very
		2451	busy or the intervals between stat's are large, a stat call will be fast,
		2452	as the path data is usually in memory already (except when starting the
		2453	watcher).
		2454
		2455	For networked file systems, calling C<stat ()> can block an indefinite
		2456	time due to network issues, and even under good conditions, a stat call
		2457	often takes multiple milliseconds.
		2458
		2459	Therefore, it is best to avoid using C<ev_stat> watchers on networked
		2460	paths, although this is fully supported by libev.
		2461
		2462	=head3 The special problem of stat time resolution
		2463
		2464	The C<stat ()> system call only supports full-second resolution portably,
		2465	and even on systems where the resolution is higher, most file systems
		2466	still only support whole seconds.
		2467
		2468	That means that, if the time is the only thing that changes, you can
		2469	easily miss updates: on the first update, C<ev_stat> detects a change and
		2470	calls your callback, which does something. When there is another update
		2471	within the same second, C<ev_stat> will be unable to detect unless the
		2472	stat data does change in other ways (e.g. file size).
		2473
		2474	The solution to this is to delay acting on a change for slightly more
		2475	than a second (or till slightly after the next full second boundary), using
		2476	a roughly one-second-delay C<ev_timer> (e.g. C<ev_timer_set (w, 0., 1.02);
		2477	ev_timer_again (loop, w)>).
		2478
		2479	The C<.02> offset is added to work around small timing inconsistencies
		2480	of some operating systems (where the second counter of the current time
		2481	might be be delayed. One such system is the Linux kernel, where a call to
		2482	C<gettimeofday> might return a timestamp with a full second later than
		2483	a subsequent C<time> call - if the equivalent of C<time ()> is used to
		2484	update file times then there will be a small window where the kernel uses
		2485	the previous second to update file times but libev might already execute
		2486	the timer callback).
		2487
		2488	=head3 Watcher-Specific Functions and Data Members
		2489
		2490	=over 4
		2491
		2492	=item ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)
		2493
		2494	=item ev_stat_set (ev_stat *, const char *path, ev_tstamp interval)
		2495
		2496	Configures the watcher to wait for status changes of the given
		2497	C<path>. The C<interval> is a hint on how quickly a change is expected to
		2498	be detected and should normally be specified as C<0> to let libev choose
		2499	a suitable value. The memory pointed to by C<path> must point to the same
		2500	path for as long as the watcher is active.
		2501
		2502	The callback will receive an C<EV_STAT> event when a change was detected,
		2503	relative to the attributes at the time the watcher was started (or the
		2504	last change was detected).
		2505
		2506	=item ev_stat_stat (loop, ev_stat *)
		2507
		2508	Updates the stat buffer immediately with new values. If you change the
		2509	watched path in your callback, you could call this function to avoid
		2510	detecting this change (while introducing a race condition if you are not
		2511	the only one changing the path). Can also be useful simply to find out the
		2512	new values.
		2513
		2514	=item ev_statdata attr [read-only]
		2515
		2516	The most-recently detected attributes of the file. Although the type is
		2517	C<ev_statdata>, this is usually the (or one of the) C<struct stat> types
		2518	suitable for your system, but you can only rely on the POSIX-standardised
		2519	members to be present. If the C<st_nlink> member is C<0>, then there was
		2520	some error while C<stat>ing the file.
		2521
		2522	=item ev_statdata prev [read-only]
		2523
		2524	The previous attributes of the file. The callback gets invoked whenever
		2525	C<prev> != C<attr>, or, more precisely, one or more of these members
		2526	differ: C<st_dev>, C<st_ino>, C<st_mode>, C<st_nlink>, C<st_uid>,
		2527	C<st_gid>, C<st_rdev>, C<st_size>, C<st_atime>, C<st_mtime>, C<st_ctime>.
		2528
		2529	=item ev_tstamp interval [read-only]
		2530
		2531	The specified interval.
		2532
		2533	=item const char *path [read-only]
		2534
		2535	The file system path that is being watched.
		2536
		2537	=back
		2538
		2539	=head3 Examples
		2540
		2541	Example: Watch C</etc/passwd> for attribute changes.
		2542
		2543	static void
		2544	passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)
		2545	{
		2546	/* /etc/passwd changed in some way */
		2547	if (w->attr.st_nlink)
		2548	{
		2549	printf ("passwd current size %ld\n", (long)w->attr.st_size);
		2550	printf ("passwd current atime %ld\n", (long)w->attr.st_mtime);
		2551	printf ("passwd current mtime %ld\n", (long)w->attr.st_mtime);
		2552	}
		2553	else
		2554	/* you shalt not abuse printf for puts */
		2555	puts ("wow, /etc/passwd is not there, expect problems. "
		2556	"if this is windows, they already arrived\n");
		2557	}
		2558
		2559	...
		2560	ev_stat passwd;
		2561
		2562	ev_stat_init (&passwd, passwd_cb, "/etc/passwd", 0.);
		2563	ev_stat_start (loop, &passwd);
		2564
		2565	Example: Like above, but additionally use a one-second delay so we do not
		2566	miss updates (however, frequent updates will delay processing, too, so
		2567	one might do the work both on C<ev_stat> callback invocation I<and> on
		2568	C<ev_timer> callback invocation).
		2569
		2570	static ev_stat passwd;
		2571	static ev_timer timer;
		2572
		2573	static void
		2574	timer_cb (EV_P_ ev_timer *w, int revents)
		2575	{
		2576	ev_timer_stop (EV_A_ w);
		2577
		2578	/* now it's one second after the most recent passwd change */
		2579	}
		2580
		2581	static void
		2582	stat_cb (EV_P_ ev_stat *w, int revents)
		2583	{
		2584	/* reset the one-second timer */
		2585	ev_timer_again (EV_A_ &timer);
		2586	}
		2587
		2588	...
		2589	ev_stat_init (&passwd, stat_cb, "/etc/passwd", 0.);
		2590	ev_stat_start (loop, &passwd);
		2591	ev_timer_init (&timer, timer_cb, 0., 1.02);
1055		2592
1056		2593
1057	=head2 C<ev_idle> - when you've got nothing better to do...	2594	=head2 C<ev_idle> - when you've got nothing better to do...
1058		2595
1059	Idle watchers trigger events when there are no other events are pending	2596	Idle watchers trigger events when no other events of the same or higher
1060	(prepare, check and other idle watchers do not count). That is, as long	2597	priority are pending (prepare, check and other idle watchers do not count
1061	as your process is busy handling sockets or timeouts (or even signals,	2598	as receiving "events").
1062	imagine) it will not be triggered. But when your process is idle all idle	2599
1063	watchers are being called again and again, once per event loop iteration -	2600	That is, as long as your process is busy handling sockets or timeouts
		2601	(or even signals, imagine) of the same or higher priority it will not be
		2602	triggered. But when your process is idle (or only lower-priority watchers
		2603	are pending), the idle watchers are being called once per event loop
1064	until stopped, that is, or your process receives more events and becomes	2604	iteration - until stopped, that is, or your process receives more events
1065	busy.	2605	and becomes busy again with higher priority stuff.
1066		2606
1067	The most noteworthy effect is that as long as any idle watchers are	2607	The most noteworthy effect is that as long as any idle watchers are
1068	active, the process will not block when waiting for new events.	2608	active, the process will not block when waiting for new events.
1069		2609
1070	Apart from keeping your process non-blocking (which is a useful	2610	Apart from keeping your process non-blocking (which is a useful
1071	effect on its own sometimes), idle watchers are a good place to do	2611	effect on its own sometimes), idle watchers are a good place to do
1072	"pseudo-background processing", or delay processing stuff to after the	2612	"pseudo-background processing", or delay processing stuff to after the
1073	event loop has handled all outstanding events.	2613	event loop has handled all outstanding events.
1074		2614
		2615	=head3 Watcher-Specific Functions and Data Members
		2616
1075	=over 4	2617	=over 4
1076		2618
1077	=item ev_idle_init (ev_signal *, callback)	2619	=item ev_idle_init (ev_idle *, callback)
1078		2620
1079	Initialises and configures the idle watcher - it has no parameters of any	2621	Initialises and configures the idle watcher - it has no parameters of any
1080	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,	2622	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,
1081	believe me.	2623	believe me.
1082		2624
1083	=back	2625	=back
1084		2626
		2627	=head3 Examples
		2628
1085	Example: dynamically allocate an C<ev_idle>, start it, and in the	2629	Example: Dynamically allocate an C<ev_idle> watcher, start it, and in the
1086	callback, free it. Alos, use no error checking, as usual.	2630	callback, free it. Also, use no error checking, as usual.
1087		2631
1088	static void	2632	static void
1089	idle_cb (struct ev_loop *loop, struct ev_idle *w, int revents)	2633	idle_cb (struct ev_loop *loop, ev_idle *w, int revents)
1090	{	2634	{
1091	free (w);	2635	free (w);
1092	// now do something you wanted to do when the program has	2636	// now do something you wanted to do when the program has
1093	// no longer asnything immediate to do.	2637	// no longer anything immediate to do.
1094	}	2638	}
1095		2639
1096	struct ev_idle *idle_watcher = malloc (sizeof (struct ev_idle));	2640	ev_idle *idle_watcher = malloc (sizeof (ev_idle));
1097	ev_idle_init (idle_watcher, idle_cb);	2641	ev_idle_init (idle_watcher, idle_cb);
1098	ev_idle_start (loop, idle_cb);	2642	ev_idle_start (loop, idle_watcher);
1099		2643
1100		2644
1101	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!	2645	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!
1102		2646
1103	Prepare and check watchers are usually (but not always) used in tandem:	2647	Prepare and check watchers are usually (but not always) used in pairs:
1104	prepare watchers get invoked before the process blocks and check watchers	2648	prepare watchers get invoked before the process blocks and check watchers
1105	afterwards.	2649	afterwards.
1106		2650
1107	You I<must not> call C<ev_loop> or similar functions that enter	2651	You I<must not> call C<ev_run> or similar functions that enter
1108	the current event loop from either C<ev_prepare> or C<ev_check>	2652	the current event loop from either C<ev_prepare> or C<ev_check>
1109	watchers. Other loops than the current one are fine, however. The	2653	watchers. Other loops than the current one are fine, however. The
1110	rationale behind this is that you do not need to check for recursion in	2654	rationale behind this is that you do not need to check for recursion in
1111	those watchers, i.e. the sequence will always be C<ev_prepare>, blocking,	2655	those watchers, i.e. the sequence will always be C<ev_prepare>, blocking,
1112	C<ev_check> so if you have one watcher of each kind they will always be	2656	C<ev_check> so if you have one watcher of each kind they will always be
1113	called in pairs bracketing the blocking call.	2657	called in pairs bracketing the blocking call.
1114		2658
1115	Their main purpose is to integrate other event mechanisms into libev and	2659	Their main purpose is to integrate other event mechanisms into libev and
1116	their use is somewhat advanced. This could be used, for example, to track	2660	their use is somewhat advanced. They could be used, for example, to track
1117	variable changes, implement your own watchers, integrate net-snmp or a	2661	variable changes, implement your own watchers, integrate net-snmp or a
1118	coroutine library and lots more. They are also occasionally useful if	2662	coroutine library and lots more. They are also occasionally useful if
1119	you cache some data and want to flush it before blocking (for example,	2663	you cache some data and want to flush it before blocking (for example,
1120	in X programs you might want to do an C<XFlush ()> in an C<ev_prepare>	2664	in X programs you might want to do an C<XFlush ()> in an C<ev_prepare>
1121	watcher).	2665	watcher).
1122		2666
1123	This is done by examining in each prepare call which file descriptors need	2667	This is done by examining in each prepare call which file descriptors
1124	to be watched by the other library, registering C<ev_io> watchers for	2668	need to be watched by the other library, registering C<ev_io> watchers
1125	them and starting an C<ev_timer> watcher for any timeouts (many libraries	2669	for them and starting an C<ev_timer> watcher for any timeouts (many
1126	provide just this functionality). Then, in the check watcher you check for	2670	libraries provide exactly this functionality). Then, in the check watcher,
1127	any events that occured (by checking the pending status of all watchers	2671	you check for any events that occurred (by checking the pending status
1128	and stopping them) and call back into the library. The I/O and timer	2672	of all watchers and stopping them) and call back into the library. The
1129	callbacks will never actually be called (but must be valid nevertheless,	2673	I/O and timer callbacks will never actually be called (but must be valid
1130	because you never know, you know?).	2674	nevertheless, because you never know, you know?).
1131		2675
1132	As another example, the Perl Coro module uses these hooks to integrate	2676	As another example, the Perl Coro module uses these hooks to integrate
1133	coroutines into libev programs, by yielding to other active coroutines	2677	coroutines into libev programs, by yielding to other active coroutines
1134	during each prepare and only letting the process block if no coroutines	2678	during each prepare and only letting the process block if no coroutines
1135	are ready to run (it's actually more complicated: it only runs coroutines	2679	are ready to run (it's actually more complicated: it only runs coroutines
1136	with priority higher than or equal to the event loop and one coroutine	2680	with priority higher than or equal to the event loop and one coroutine
1137	of lower priority, but only once, using idle watchers to keep the event	2681	of lower priority, but only once, using idle watchers to keep the event
1138	loop from blocking if lower-priority coroutines are active, thus mapping	2682	loop from blocking if lower-priority coroutines are active, thus mapping
1139	low-priority coroutines to idle/background tasks).	2683	low-priority coroutines to idle/background tasks).
1140		2684
		2685	It is recommended to give C<ev_check> watchers highest (C<EV_MAXPRI>)
		2686	priority, to ensure that they are being run before any other watchers
		2687	after the poll (this doesn't matter for C<ev_prepare> watchers).
		2688
		2689	Also, C<ev_check> watchers (and C<ev_prepare> watchers, too) should not
		2690	activate ("feed") events into libev. While libev fully supports this, they
		2691	might get executed before other C<ev_check> watchers did their job. As
		2692	C<ev_check> watchers are often used to embed other (non-libev) event
		2693	loops those other event loops might be in an unusable state until their
		2694	C<ev_check> watcher ran (always remind yourself to coexist peacefully with
		2695	others).
		2696
		2697	=head3 Watcher-Specific Functions and Data Members
		2698
1141	=over 4	2699	=over 4
1142		2700
1143	=item ev_prepare_init (ev_prepare *, callback)	2701	=item ev_prepare_init (ev_prepare *, callback)
1144		2702
1145	=item ev_check_init (ev_check *, callback)	2703	=item ev_check_init (ev_check *, callback)
1146		2704
1147	Initialises and configures the prepare or check watcher - they have no	2705	Initialises and configures the prepare or check watcher - they have no
1148	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>	2706	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>
1149	macros, but using them is utterly, utterly and completely pointless.	2707	macros, but using them is utterly, utterly, utterly and completely
		2708	pointless.
1150		2709
1151	=back	2710	=back
1152		2711
1153	Example: To include a library such as adns, you would add IO watchers	2712	=head3 Examples
1154	and a timeout watcher in a prepare handler, as required by libadns, and	2713
		2714	There are a number of principal ways to embed other event loops or modules
		2715	into libev. Here are some ideas on how to include libadns into libev
		2716	(there is a Perl module named C<EV::ADNS> that does this, which you could
		2717	use as a working example. Another Perl module named C<EV::Glib> embeds a
		2718	Glib main context into libev, and finally, C<Glib::EV> embeds EV into the
		2719	Glib event loop).
		2720
		2721	Method 1: Add IO watchers and a timeout watcher in a prepare handler,
1155	in a check watcher, destroy them and call into libadns. What follows is	2722	and in a check watcher, destroy them and call into libadns. What follows
1156	pseudo-code only of course:	2723	is pseudo-code only of course. This requires you to either use a low
		2724	priority for the check watcher or use C<ev_clear_pending> explicitly, as
		2725	the callbacks for the IO/timeout watchers might not have been called yet.
1157		2726
1158	static ev_io iow [nfd];	2727	static ev_io iow [nfd];
1159	static ev_timer tw;	2728	static ev_timer tw;
1160		2729
1161	static void	2730	static void
1162	io_cb (ev_loop *loop, ev_io *w, int revents)	2731	io_cb (struct ev_loop *loop, ev_io *w, int revents)
1163	{	2732	{
1164	// set the relevant poll flags
1165	// could also call adns_processreadable etc. here
1166	struct pollfd *fd = (struct pollfd *)w->data;
1167	if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;
1168	if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;
1169	}	2733	}
1170		2734
1171	// create io watchers for each fd and a timer before blocking	2735	// create io watchers for each fd and a timer before blocking
1172	static void	2736	static void
1173	adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents)	2737	adns_prepare_cb (struct ev_loop *loop, ev_prepare *w, int revents)
1174	{	2738	{
1175	int timeout = 3600000;truct pollfd fds [nfd];	2739	int timeout = 3600000;
		2740	struct pollfd fds [nfd];
1176	// actual code will need to loop here and realloc etc.	2741	// actual code will need to loop here and realloc etc.
1177	adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));	2742	adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));
1178		2743
1179	/* the callback is illegal, but won't be called as we stop during check */	2744	/* the callback is illegal, but won't be called as we stop during check */
1180	ev_timer_init (&tw, 0, timeout * 1e-3);	2745	ev_timer_init (&tw, 0, timeout * 1e-3, 0.);
1181	ev_timer_start (loop, &tw);	2746	ev_timer_start (loop, &tw);
1182		2747
1183	// create on ev_io per pollfd	2748	// create one ev_io per pollfd
1184	for (int i = 0; i < nfd; ++i)	2749	for (int i = 0; i < nfd; ++i)
1185	{	2750	{
1186	ev_io_init (iow + i, io_cb, fds [i].fd,	2751	ev_io_init (iow + i, io_cb, fds [i].fd,
1187	((fds [i].events & POLLIN ? EV_READ : 0)	2752	((fds [i].events & POLLIN ? EV_READ : 0)
1188	| (fds [i].events & POLLOUT ? EV_WRITE : 0)));	2753	| (fds [i].events & POLLOUT ? EV_WRITE : 0)));
1189		2754
1190	fds [i].revents = 0;	2755	fds [i].revents = 0;
1191	iow [i].data = fds + i;
1192	ev_io_start (loop, iow + i);	2756	ev_io_start (loop, iow + i);
1193	}	2757	}
1194	}	2758	}
1195		2759
1196	// stop all watchers after blocking	2760	// stop all watchers after blocking
1197	static void	2761	static void
1198	adns_check_cb (ev_loop *loop, ev_check *w, int revents)	2762	adns_check_cb (struct ev_loop *loop, ev_check *w, int revents)
1199	{	2763	{
1200	ev_timer_stop (loop, &tw);	2764	ev_timer_stop (loop, &tw);
1201		2765
1202	for (int i = 0; i < nfd; ++i)	2766	for (int i = 0; i < nfd; ++i)
		2767	{
		2768	// set the relevant poll flags
		2769	// could also call adns_processreadable etc. here
		2770	struct pollfd *fd = fds + i;
		2771	int revents = ev_clear_pending (iow + i);
		2772	if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;
		2773	if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;
		2774
		2775	// now stop the watcher
1203	ev_io_stop (loop, iow + i);	2776	ev_io_stop (loop, iow + i);
		2777	}
1204		2778
1205	adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));	2779	adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));
1206	}	2780	}
		2781
		2782	Method 2: This would be just like method 1, but you run C<adns_afterpoll>
		2783	in the prepare watcher and would dispose of the check watcher.
		2784
		2785	Method 3: If the module to be embedded supports explicit event
		2786	notification (libadns does), you can also make use of the actual watcher
		2787	callbacks, and only destroy/create the watchers in the prepare watcher.
		2788
		2789	static void
		2790	timer_cb (EV_P_ ev_timer *w, int revents)
		2791	{
		2792	adns_state ads = (adns_state)w->data;
		2793	update_now (EV_A);
		2794
		2795	adns_processtimeouts (ads, &tv_now);
		2796	}
		2797
		2798	static void
		2799	io_cb (EV_P_ ev_io *w, int revents)
		2800	{
		2801	adns_state ads = (adns_state)w->data;
		2802	update_now (EV_A);
		2803
		2804	if (revents & EV_READ ) adns_processreadable (ads, w->fd, &tv_now);
		2805	if (revents & EV_WRITE) adns_processwriteable (ads, w->fd, &tv_now);
		2806	}
		2807
		2808	// do not ever call adns_afterpoll
		2809
		2810	Method 4: Do not use a prepare or check watcher because the module you
		2811	want to embed is not flexible enough to support it. Instead, you can
		2812	override their poll function. The drawback with this solution is that the
		2813	main loop is now no longer controllable by EV. The C<Glib::EV> module uses
		2814	this approach, effectively embedding EV as a client into the horrible
		2815	libglib event loop.
		2816
		2817	static gint
		2818	event_poll_func (GPollFD *fds, guint nfds, gint timeout)
		2819	{
		2820	int got_events = 0;
		2821
		2822	for (n = 0; n < nfds; ++n)
		2823	// create/start io watcher that sets the relevant bits in fds[n] and increment got_events
		2824
		2825	if (timeout >= 0)
		2826	// create/start timer
		2827
		2828	// poll
		2829	ev_run (EV_A_ 0);
		2830
		2831	// stop timer again
		2832	if (timeout >= 0)
		2833	ev_timer_stop (EV_A_ &to);
		2834
		2835	// stop io watchers again - their callbacks should have set
		2836	for (n = 0; n < nfds; ++n)
		2837	ev_io_stop (EV_A_ iow [n]);
		2838
		2839	return got_events;
		2840	}
1207		2841
1208		2842
1209	=head2 C<ev_embed> - when one backend isn't enough...	2843	=head2 C<ev_embed> - when one backend isn't enough...
1210		2844
1211	This is a rather advanced watcher type that lets you embed one event loop	2845	This is a rather advanced watcher type that lets you embed one event loop
…		…
1217	prioritise I/O.	2851	prioritise I/O.
1218		2852
1219	As an example for a bug workaround, the kqueue backend might only support	2853	As an example for a bug workaround, the kqueue backend might only support
1220	sockets on some platform, so it is unusable as generic backend, but you	2854	sockets on some platform, so it is unusable as generic backend, but you
1221	still want to make use of it because you have many sockets and it scales	2855	still want to make use of it because you have many sockets and it scales
1222	so nicely. In this case, you would create a kqueue-based loop and embed it	2856	so nicely. In this case, you would create a kqueue-based loop and embed
1223	into your default loop (which might use e.g. poll). Overall operation will	2857	it into your default loop (which might use e.g. poll). Overall operation
1224	be a bit slower because first libev has to poll and then call kevent, but	2858	will be a bit slower because first libev has to call C<poll> and then
1225	at least you can use both at what they are best.	2859	C<kevent>, but at least you can use both mechanisms for what they are
		2860	best: C<kqueue> for scalable sockets and C<poll> if you want it to work :)
1226		2861
1227	As for prioritising I/O: rarely you have the case where some fds have	2862	As for prioritising I/O: under rare circumstances you have the case where
1228	to be watched and handled very quickly (with low latency), and even	2863	some fds have to be watched and handled very quickly (with low latency),
1229	priorities and idle watchers might have too much overhead. In this case	2864	and even priorities and idle watchers might have too much overhead. In
1230	you would put all the high priority stuff in one loop and all the rest in	2865	this case you would put all the high priority stuff in one loop and all
1231	a second one, and embed the second one in the first.	2866	the rest in a second one, and embed the second one in the first.
1232		2867
1233	As long as the watcher is active, the callback will be invoked every time	2868	As long as the watcher is active, the callback will be invoked every
1234	there might be events pending in the embedded loop. The callback must then	2869	time there might be events pending in the embedded loop. The callback
1235	call C<ev_embed_sweep (mainloop, watcher)> to make a single sweep and invoke	2870	must then call C<ev_embed_sweep (mainloop, watcher)> to make a single
1236	their callbacks (you could also start an idle watcher to give the embedded	2871	sweep and invoke their callbacks (the callback doesn't need to invoke the
1237	loop strictly lower priority for example). You can also set the callback	2872	C<ev_embed_sweep> function directly, it could also start an idle watcher
1238	to C<0>, in which case the embed watcher will automatically execute the	2873	to give the embedded loop strictly lower priority for example).
1239	embedded loop sweep.
1240		2874
1241	As long as the watcher is started it will automatically handle events. The	2875	You can also set the callback to C<0>, in which case the embed watcher
1242	callback will be invoked whenever some events have been handled. You can	2876	will automatically execute the embedded loop sweep whenever necessary.
1243	set the callback to C<0> to avoid having to specify one if you are not
1244	interested in that.
1245		2877
1246	Also, there have not currently been made special provisions for forking:	2878	Fork detection will be handled transparently while the C<ev_embed> watcher
1247	when you fork, you not only have to call C<ev_loop_fork> on both loops,	2879	is active, i.e., the embedded loop will automatically be forked when the
1248	but you will also have to stop and restart any C<ev_embed> watchers	2880	embedding loop forks. In other cases, the user is responsible for calling
1249	yourself.	2881	C<ev_loop_fork> on the embedded loop.
1250		2882
1251	Unfortunately, not all backends are embeddable, only the ones returned by	2883	Unfortunately, not all backends are embeddable: only the ones returned by
1252	C<ev_embeddable_backends> are, which, unfortunately, does not include any	2884	C<ev_embeddable_backends> are, which, unfortunately, does not include any
1253	portable one.	2885	portable one.
1254		2886
1255	So when you want to use this feature you will always have to be prepared	2887	So when you want to use this feature you will always have to be prepared
1256	that you cannot get an embeddable loop. The recommended way to get around	2888	that you cannot get an embeddable loop. The recommended way to get around
1257	this is to have a separate variables for your embeddable loop, try to	2889	this is to have a separate variables for your embeddable loop, try to
1258	create it, and if that fails, use the normal loop for everything:	2890	create it, and if that fails, use the normal loop for everything.
1259		2891
1260	struct ev_loop *loop_hi = ev_default_init (0);	2892	=head3 C<ev_embed> and fork
1261	struct ev_loop *loop_lo = 0;
1262	struct ev_embed embed;
1263
1264	// see if there is a chance of getting one that works
1265	// (remember that a flags value of 0 means autodetection)
1266	loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
1267	? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
1268	: 0;
1269		2893
1270	// if we got one, then embed it, otherwise default to loop_hi	2894	While the C<ev_embed> watcher is running, forks in the embedding loop will
1271	if (loop_lo)	2895	automatically be applied to the embedded loop as well, so no special
1272	{	2896	fork handling is required in that case. When the watcher is not running,
1273	ev_embed_init (&embed, 0, loop_lo);	2897	however, it is still the task of the libev user to call C<ev_loop_fork ()>
1274	ev_embed_start (loop_hi, &embed);	2898	as applicable.
1275	}	2899
1276	else	2900	=head3 Watcher-Specific Functions and Data Members
1277	loop_lo = loop_hi;
1278		2901
1279	=over 4	2902	=over 4
1280		2903
1281	=item ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)	2904	=item ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)
1282		2905
…		…
1284		2907
1285	Configures the watcher to embed the given loop, which must be	2908	Configures the watcher to embed the given loop, which must be
1286	embeddable. If the callback is C<0>, then C<ev_embed_sweep> will be	2909	embeddable. If the callback is C<0>, then C<ev_embed_sweep> will be
1287	invoked automatically, otherwise it is the responsibility of the callback	2910	invoked automatically, otherwise it is the responsibility of the callback
1288	to invoke it (it will continue to be called until the sweep has been done,	2911	to invoke it (it will continue to be called until the sweep has been done,
1289	if you do not want thta, you need to temporarily stop the embed watcher).	2912	if you do not want that, you need to temporarily stop the embed watcher).
1290		2913
1291	=item ev_embed_sweep (loop, ev_embed *)	2914	=item ev_embed_sweep (loop, ev_embed *)
1292		2915
1293	Make a single, non-blocking sweep over the embedded loop. This works	2916	Make a single, non-blocking sweep over the embedded loop. This works
1294	similarly to C<ev_loop (embedded_loop, EVLOOP_NONBLOCK)>, but in the most	2917	similarly to C<ev_run (embedded_loop, EVRUN_NOWAIT)>, but in the most
1295	apropriate way for embedded loops.	2918	appropriate way for embedded loops.
		2919
		2920	=item struct ev_loop *other [read-only]
		2921
		2922	The embedded event loop.
1296		2923
1297	=back	2924	=back
1298		2925
		2926	=head3 Examples
		2927
		2928	Example: Try to get an embeddable event loop and embed it into the default
		2929	event loop. If that is not possible, use the default loop. The default
		2930	loop is stored in C<loop_hi>, while the embeddable loop is stored in
		2931	C<loop_lo> (which is C<loop_hi> in the case no embeddable loop can be
		2932	used).
		2933
		2934	struct ev_loop *loop_hi = ev_default_init (0);
		2935	struct ev_loop *loop_lo = 0;
		2936	ev_embed embed;
		2937
		2938	// see if there is a chance of getting one that works
		2939	// (remember that a flags value of 0 means autodetection)
		2940	loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
		2941	? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
		2942	: 0;
		2943
		2944	// if we got one, then embed it, otherwise default to loop_hi
		2945	if (loop_lo)
		2946	{
		2947	ev_embed_init (&embed, 0, loop_lo);
		2948	ev_embed_start (loop_hi, &embed);
		2949	}
		2950	else
		2951	loop_lo = loop_hi;
		2952
		2953	Example: Check if kqueue is available but not recommended and create
		2954	a kqueue backend for use with sockets (which usually work with any
		2955	kqueue implementation). Store the kqueue/socket-only event loop in
		2956	C<loop_socket>. (One might optionally use C<EVFLAG_NOENV>, too).
		2957
		2958	struct ev_loop *loop = ev_default_init (0);
		2959	struct ev_loop *loop_socket = 0;
		2960	ev_embed embed;
		2961
		2962	if (ev_supported_backends () & ~ev_recommended_backends () & EVBACKEND_KQUEUE)
		2963	if ((loop_socket = ev_loop_new (EVBACKEND_KQUEUE))
		2964	{
		2965	ev_embed_init (&embed, 0, loop_socket);
		2966	ev_embed_start (loop, &embed);
		2967	}
		2968
		2969	if (!loop_socket)
		2970	loop_socket = loop;
		2971
		2972	// now use loop_socket for all sockets, and loop for everything else
		2973
		2974
		2975	=head2 C<ev_fork> - the audacity to resume the event loop after a fork
		2976
		2977	Fork watchers are called when a C<fork ()> was detected (usually because
		2978	whoever is a good citizen cared to tell libev about it by calling
		2979	C<ev_default_fork> or C<ev_loop_fork>). The invocation is done before the
		2980	event loop blocks next and before C<ev_check> watchers are being called,
		2981	and only in the child after the fork. If whoever good citizen calling
		2982	C<ev_default_fork> cheats and calls it in the wrong process, the fork
		2983	handlers will be invoked, too, of course.
		2984
		2985	=head3 The special problem of life after fork - how is it possible?
		2986
		2987	Most uses of C<fork()> consist of forking, then some simple calls to set
		2988	up/change the process environment, followed by a call to C<exec()>. This
		2989	sequence should be handled by libev without any problems.
		2990
		2991	This changes when the application actually wants to do event handling
		2992	in the child, or both parent in child, in effect "continuing" after the
		2993	fork.
		2994
		2995	The default mode of operation (for libev, with application help to detect
		2996	forks) is to duplicate all the state in the child, as would be expected
		2997	when I<either> the parent I<or> the child process continues.
		2998
		2999	When both processes want to continue using libev, then this is usually the
		3000	wrong result. In that case, usually one process (typically the parent) is
		3001	supposed to continue with all watchers in place as before, while the other
		3002	process typically wants to start fresh, i.e. without any active watchers.
		3003
		3004	The cleanest and most efficient way to achieve that with libev is to
		3005	simply create a new event loop, which of course will be "empty", and
		3006	use that for new watchers. This has the advantage of not touching more
		3007	memory than necessary, and thus avoiding the copy-on-write, and the
		3008	disadvantage of having to use multiple event loops (which do not support
		3009	signal watchers).
		3010
		3011	When this is not possible, or you want to use the default loop for
		3012	other reasons, then in the process that wants to start "fresh", call
		3013	C<ev_default_destroy ()> followed by C<ev_default_loop (...)>. Destroying
		3014	the default loop will "orphan" (not stop) all registered watchers, so you
		3015	have to be careful not to execute code that modifies those watchers. Note
		3016	also that in that case, you have to re-register any signal watchers.
		3017
		3018	=head3 Watcher-Specific Functions and Data Members
		3019
		3020	=over 4
		3021
		3022	=item ev_fork_init (ev_signal *, callback)
		3023
		3024	Initialises and configures the fork watcher - it has no parameters of any
		3025	kind. There is a C<ev_fork_set> macro, but using it is utterly pointless,
		3026	believe me.
		3027
		3028	=back
		3029
		3030
		3031	=head2 C<ev_async> - how to wake up an event loop
		3032
		3033	In general, you cannot use an C<ev_run> from multiple threads or other
		3034	asynchronous sources such as signal handlers (as opposed to multiple event
		3035	loops - those are of course safe to use in different threads).
		3036
		3037	Sometimes, however, you need to wake up an event loop you do not control,
		3038	for example because it belongs to another thread. This is what C<ev_async>
		3039	watchers do: as long as the C<ev_async> watcher is active, you can signal
		3040	it by calling C<ev_async_send>, which is thread- and signal safe.
		3041
		3042	This functionality is very similar to C<ev_signal> watchers, as signals,
		3043	too, are asynchronous in nature, and signals, too, will be compressed
		3044	(i.e. the number of callback invocations may be less than the number of
		3045	C<ev_async_sent> calls).
		3046
		3047	Unlike C<ev_signal> watchers, C<ev_async> works with any event loop, not
		3048	just the default loop.
		3049
		3050	=head3 Queueing
		3051
		3052	C<ev_async> does not support queueing of data in any way. The reason
		3053	is that the author does not know of a simple (or any) algorithm for a
		3054	multiple-writer-single-reader queue that works in all cases and doesn't
		3055	need elaborate support such as pthreads or unportable memory access
		3056	semantics.
		3057
		3058	That means that if you want to queue data, you have to provide your own
		3059	queue. But at least I can tell you how to implement locking around your
		3060	queue:
		3061
		3062	=over 4
		3063
		3064	=item queueing from a signal handler context
		3065
		3066	To implement race-free queueing, you simply add to the queue in the signal
		3067	handler but you block the signal handler in the watcher callback. Here is
		3068	an example that does that for some fictitious SIGUSR1 handler:
		3069
		3070	static ev_async mysig;
		3071
		3072	static void
		3073	sigusr1_handler (void)
		3074	{
		3075	sometype data;
		3076
		3077	// no locking etc.
		3078	queue_put (data);
		3079	ev_async_send (EV_DEFAULT_ &mysig);
		3080	}
		3081
		3082	static void
		3083	mysig_cb (EV_P_ ev_async *w, int revents)
		3084	{
		3085	sometype data;
		3086	sigset_t block, prev;
		3087
		3088	sigemptyset (&block);
		3089	sigaddset (&block, SIGUSR1);
		3090	sigprocmask (SIG_BLOCK, &block, &prev);
		3091
		3092	while (queue_get (&data))
		3093	process (data);
		3094
		3095	if (sigismember (&prev, SIGUSR1)
		3096	sigprocmask (SIG_UNBLOCK, &block, 0);
		3097	}
		3098
		3099	(Note: pthreads in theory requires you to use C<pthread_setmask>
		3100	instead of C<sigprocmask> when you use threads, but libev doesn't do it
		3101	either...).
		3102
		3103	=item queueing from a thread context
		3104
		3105	The strategy for threads is different, as you cannot (easily) block
		3106	threads but you can easily preempt them, so to queue safely you need to
		3107	employ a traditional mutex lock, such as in this pthread example:
		3108
		3109	static ev_async mysig;
		3110	static pthread_mutex_t mymutex = PTHREAD_MUTEX_INITIALIZER;
		3111
		3112	static void
		3113	otherthread (void)
		3114	{
		3115	// only need to lock the actual queueing operation
		3116	pthread_mutex_lock (&mymutex);
		3117	queue_put (data);
		3118	pthread_mutex_unlock (&mymutex);
		3119
		3120	ev_async_send (EV_DEFAULT_ &mysig);
		3121	}
		3122
		3123	static void
		3124	mysig_cb (EV_P_ ev_async *w, int revents)
		3125	{
		3126	pthread_mutex_lock (&mymutex);
		3127
		3128	while (queue_get (&data))
		3129	process (data);
		3130
		3131	pthread_mutex_unlock (&mymutex);
		3132	}
		3133
		3134	=back
		3135
		3136
		3137	=head3 Watcher-Specific Functions and Data Members
		3138
		3139	=over 4
		3140
		3141	=item ev_async_init (ev_async *, callback)
		3142
		3143	Initialises and configures the async watcher - it has no parameters of any
		3144	kind. There is a C<ev_async_set> macro, but using it is utterly pointless,
		3145	trust me.
		3146
		3147	=item ev_async_send (loop, ev_async *)
		3148
		3149	Sends/signals/activates the given C<ev_async> watcher, that is, feeds
		3150	an C<EV_ASYNC> event on the watcher into the event loop. Unlike
		3151	C<ev_feed_event>, this call is safe to do from other threads, signal or
		3152	similar contexts (see the discussion of C<EV_ATOMIC_T> in the embedding
		3153	section below on what exactly this means).
		3154
		3155	Note that, as with other watchers in libev, multiple events might get
		3156	compressed into a single callback invocation (another way to look at this
		3157	is that C<ev_async> watchers are level-triggered, set on C<ev_async_send>,
		3158	reset when the event loop detects that).
		3159
		3160	This call incurs the overhead of a system call only once per event loop
		3161	iteration, so while the overhead might be noticeable, it doesn't apply to
		3162	repeated calls to C<ev_async_send> for the same event loop.
		3163
		3164	=item bool = ev_async_pending (ev_async *)
		3165
		3166	Returns a non-zero value when C<ev_async_send> has been called on the
		3167	watcher but the event has not yet been processed (or even noted) by the
		3168	event loop.
		3169
		3170	C<ev_async_send> sets a flag in the watcher and wakes up the loop. When
		3171	the loop iterates next and checks for the watcher to have become active,
		3172	it will reset the flag again. C<ev_async_pending> can be used to very
		3173	quickly check whether invoking the loop might be a good idea.
		3174
		3175	Not that this does I<not> check whether the watcher itself is pending,
		3176	only whether it has been requested to make this watcher pending: there
		3177	is a time window between the event loop checking and resetting the async
		3178	notification, and the callback being invoked.
		3179
		3180	=back
		3181
1299		3182
1300	=head1 OTHER FUNCTIONS	3183	=head1 OTHER FUNCTIONS
1301		3184
1302	There are some other functions of possible interest. Described. Here. Now.	3185	There are some other functions of possible interest. Described. Here. Now.
1303		3186
1304	=over 4	3187	=over 4
1305		3188
1306	=item ev_once (loop, int fd, int events, ev_tstamp timeout, callback)	3189	=item ev_once (loop, int fd, int events, ev_tstamp timeout, callback)
1307		3190
1308	This function combines a simple timer and an I/O watcher, calls your	3191	This function combines a simple timer and an I/O watcher, calls your
1309	callback on whichever event happens first and automatically stop both	3192	callback on whichever event happens first and automatically stops both
1310	watchers. This is useful if you want to wait for a single event on an fd	3193	watchers. This is useful if you want to wait for a single event on an fd
1311	or timeout without having to allocate/configure/start/stop/free one or	3194	or timeout without having to allocate/configure/start/stop/free one or
1312	more watchers yourself.	3195	more watchers yourself.
1313		3196
1314	If C<fd> is less than 0, then no I/O watcher will be started and events	3197	If C<fd> is less than 0, then no I/O watcher will be started and the
1315	is being ignored. Otherwise, an C<ev_io> watcher for the given C<fd> and	3198	C<events> argument is being ignored. Otherwise, an C<ev_io> watcher for
1316	C<events> set will be craeted and started.	3199	the given C<fd> and C<events> set will be created and started.
1317		3200
1318	If C<timeout> is less than 0, then no timeout watcher will be	3201	If C<timeout> is less than 0, then no timeout watcher will be
1319	started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and	3202	started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and
1320	repeat = 0) will be started. While C<0> is a valid timeout, it is of	3203	repeat = 0) will be started. C<0> is a valid timeout.
1321	dubious value.
1322		3204
1323	The callback has the type C<void (*cb)(int revents, void *arg)> and gets	3205	The callback has the type C<void (*cb)(int revents, void *arg)> and is
1324	passed an C<revents> set like normal event callbacks (a combination of	3206	passed an C<revents> set like normal event callbacks (a combination of
1325	C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMEOUT>) and the C<arg>	3207	C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMER>) and the C<arg>
1326	value passed to C<ev_once>:	3208	value passed to C<ev_once>. Note that it is possible to receive I<both>
		3209	a timeout and an io event at the same time - you probably should give io
		3210	events precedence.
1327		3211
		3212	Example: wait up to ten seconds for data to appear on STDIN_FILENO.
		3213
1328	static void stdin_ready (int revents, void *arg)	3214	static void stdin_ready (int revents, void *arg)
1329	{	3215	{
1330	if (revents & EV_TIMEOUT)
1331	/* doh, nothing entered */;
1332	else if (revents & EV_READ)	3216	if (revents & EV_READ)
1333	/* stdin might have data for us, joy! */;	3217	/* stdin might have data for us, joy! */;
		3218	else if (revents & EV_TIMER)
		3219	/* doh, nothing entered */;
1334	}	3220	}
1335		3221
1336	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);	3222	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
1337		3223
1338	=item ev_feed_event (ev_loop *, watcher *, int revents)
1339
1340	Feeds the given event set into the event loop, as if the specified event
1341	had happened for the specified watcher (which must be a pointer to an
1342	initialised but not necessarily started event watcher).
1343
1344	=item ev_feed_fd_event (ev_loop *, int fd, int revents)	3224	=item ev_feed_fd_event (loop, int fd, int revents)
1345		3225
1346	Feed an event on the given fd, as if a file descriptor backend detected	3226	Feed an event on the given fd, as if a file descriptor backend detected
1347	the given events it.	3227	the given events it.
1348		3228
1349	=item ev_feed_signal_event (ev_loop *loop, int signum)	3229	=item ev_feed_signal_event (loop, int signum)
1350		3230
1351	Feed an event as if the given signal occured (C<loop> must be the default	3231	Feed an event as if the given signal occurred (C<loop> must be the default
1352	loop!).	3232	loop!).
1353		3233
1354	=back	3234	=back
1355		3235
1356		3236
…		…
1372		3252
1373	=item * Priorities are not currently supported. Initialising priorities	3253	=item * Priorities are not currently supported. Initialising priorities
1374	will fail and all watchers will have the same priority, even though there	3254	will fail and all watchers will have the same priority, even though there
1375	is an ev_pri field.	3255	is an ev_pri field.
1376		3256
		3257	=item * In libevent, the last base created gets the signals, in libev, the
		3258	first base created (== the default loop) gets the signals.
		3259
1377	=item * Other members are not supported.	3260	=item * Other members are not supported.
1378		3261
1379	=item * The libev emulation is I<not> ABI compatible to libevent, you need	3262	=item * The libev emulation is I<not> ABI compatible to libevent, you need
1380	to use the libev header file and library.	3263	to use the libev header file and library.
1381		3264
1382	=back	3265	=back
1383		3266
1384	=head1 C++ SUPPORT	3267	=head1 C++ SUPPORT
1385		3268
1386	Libev comes with some simplistic wrapper classes for C++ that mainly allow	3269	Libev comes with some simplistic wrapper classes for C++ that mainly allow
1387	you to use some convinience methods to start/stop watchers and also change	3270	you to use some convenience methods to start/stop watchers and also change
1388	the callback model to a model using method callbacks on objects.	3271	the callback model to a model using method callbacks on objects.
1389		3272
1390	To use it,	3273	To use it,
1391		3274
1392	#include <ev++.h>	3275	#include <ev++.h>
1393		3276
1394	(it is not installed by default). This automatically includes F<ev.h>	3277	This automatically includes F<ev.h> and puts all of its definitions (many
1395	and puts all of its definitions (many of them macros) into the global	3278	of them macros) into the global namespace. All C++ specific things are
1396	namespace. All C++ specific things are put into the C<ev> namespace.	3279	put into the C<ev> namespace. It should support all the same embedding
		3280	options as F<ev.h>, most notably C<EV_MULTIPLICITY>.
1397		3281
1398	It should support all the same embedding options as F<ev.h>, most notably	3282	Care has been taken to keep the overhead low. The only data member the C++
1399	C<EV_MULTIPLICITY>.	3283	classes add (compared to plain C-style watchers) is the event loop pointer
		3284	that the watcher is associated with (or no additional members at all if
		3285	you disable C<EV_MULTIPLICITY> when embedding libev).
		3286
		3287	Currently, functions, and static and non-static member functions can be
		3288	used as callbacks. Other types should be easy to add as long as they only
		3289	need one additional pointer for context. If you need support for other
		3290	types of functors please contact the author (preferably after implementing
		3291	it).
1400		3292
1401	Here is a list of things available in the C<ev> namespace:	3293	Here is a list of things available in the C<ev> namespace:
1402		3294
1403	=over 4	3295	=over 4
1404		3296
…		…
1420		3312
1421	All of those classes have these methods:	3313	All of those classes have these methods:
1422		3314
1423	=over 4	3315	=over 4
1424		3316
1425	=item ev::TYPE::TYPE (object *, object::method *)	3317	=item ev::TYPE::TYPE ()
1426		3318
1427	=item ev::TYPE::TYPE (object *, object::method *, struct ev_loop *)	3319	=item ev::TYPE::TYPE (loop)
1428		3320
1429	=item ev::TYPE::~TYPE	3321	=item ev::TYPE::~TYPE
1430		3322
1431	The constructor takes a pointer to an object and a method pointer to	3323	The constructor (optionally) takes an event loop to associate the watcher
1432	the event handler callback to call in this class. The constructor calls	3324	with. If it is omitted, it will use C<EV_DEFAULT>.
1433	C<ev_init> for you, which means you have to call the C<set> method	3325
1434	before starting it. If you do not specify a loop then the constructor	3326	The constructor calls C<ev_init> for you, which means you have to call the
1435	automatically associates the default loop with this watcher.	3327	C<set> method before starting it.
		3328
		3329	It will not set a callback, however: You have to call the templated C<set>
		3330	method to set a callback before you can start the watcher.
		3331
		3332	(The reason why you have to use a method is a limitation in C++ which does
		3333	not allow explicit template arguments for constructors).
1436		3334
1437	The destructor automatically stops the watcher if it is active.	3335	The destructor automatically stops the watcher if it is active.
1438		3336
		3337	=item w->set<class, &class::method> (object *)
		3338
		3339	This method sets the callback method to call. The method has to have a
		3340	signature of C<void (*)(ev_TYPE &, int)>, it receives the watcher as
		3341	first argument and the C<revents> as second. The object must be given as
		3342	parameter and is stored in the C<data> member of the watcher.
		3343
		3344	This method synthesizes efficient thunking code to call your method from
		3345	the C callback that libev requires. If your compiler can inline your
		3346	callback (i.e. it is visible to it at the place of the C<set> call and
		3347	your compiler is good :), then the method will be fully inlined into the
		3348	thunking function, making it as fast as a direct C callback.
		3349
		3350	Example: simple class declaration and watcher initialisation
		3351
		3352	struct myclass
		3353	{
		3354	void io_cb (ev::io &w, int revents) { }
		3355	}
		3356
		3357	myclass obj;
		3358	ev::io iow;
		3359	iow.set <myclass, &myclass::io_cb> (&obj);
		3360
		3361	=item w->set (object *)
		3362
		3363	This is a variation of a method callback - leaving out the method to call
		3364	will default the method to C<operator ()>, which makes it possible to use
		3365	functor objects without having to manually specify the C<operator ()> all
		3366	the time. Incidentally, you can then also leave out the template argument
		3367	list.
		3368
		3369	The C<operator ()> method prototype must be C<void operator ()(watcher &w,
		3370	int revents)>.
		3371
		3372	See the method-C<set> above for more details.
		3373
		3374	Example: use a functor object as callback.
		3375
		3376	struct myfunctor
		3377	{
		3378	void operator() (ev::io &w, int revents)
		3379	{
		3380	...
		3381	}
		3382	}
		3383
		3384	myfunctor f;
		3385
		3386	ev::io w;
		3387	w.set (&f);
		3388
		3389	=item w->set<function> (void *data = 0)
		3390
		3391	Also sets a callback, but uses a static method or plain function as
		3392	callback. The optional C<data> argument will be stored in the watcher's
		3393	C<data> member and is free for you to use.
		3394
		3395	The prototype of the C<function> must be C<void (*)(ev::TYPE &w, int)>.
		3396
		3397	See the method-C<set> above for more details.
		3398
		3399	Example: Use a plain function as callback.
		3400
		3401	static void io_cb (ev::io &w, int revents) { }
		3402	iow.set <io_cb> ();
		3403
1439	=item w->set (struct ev_loop *)	3404	=item w->set (loop)
1440		3405
1441	Associates a different C<struct ev_loop> with this watcher. You can only	3406	Associates a different C<struct ev_loop> with this watcher. You can only
1442	do this when the watcher is inactive (and not pending either).	3407	do this when the watcher is inactive (and not pending either).
1443		3408
1444	=item w->set ([args])	3409	=item w->set ([arguments])
1445		3410
1446	Basically the same as C<ev_TYPE_set>, with the same args. Must be	3411	Basically the same as C<ev_TYPE_set>, with the same arguments. Either this
1447	called at least once. Unlike the C counterpart, an active watcher gets	3412	method or a suitable start method must be called at least once. Unlike the
1448	automatically stopped and restarted.	3413	C counterpart, an active watcher gets automatically stopped and restarted
		3414	when reconfiguring it with this method.
1449		3415
1450	=item w->start ()	3416	=item w->start ()
1451		3417
1452	Starts the watcher. Note that there is no C<loop> argument as the	3418	Starts the watcher. Note that there is no C<loop> argument, as the
1453	constructor already takes the loop.	3419	constructor already stores the event loop.
		3420
		3421	=item w->start ([arguments])
		3422
		3423	Instead of calling C<set> and C<start> methods separately, it is often
		3424	convenient to wrap them in one call. Uses the same type of arguments as
		3425	the configure C<set> method of the watcher.
1454		3426
1455	=item w->stop ()	3427	=item w->stop ()
1456		3428
1457	Stops the watcher if it is active. Again, no C<loop> argument.	3429	Stops the watcher if it is active. Again, no C<loop> argument.
1458		3430
1459	=item w->again () C<ev::timer>, C<ev::periodic> only	3431	=item w->again () (C<ev::timer>, C<ev::periodic> only)
1460		3432
1461	For C<ev::timer> and C<ev::periodic>, this invokes the corresponding	3433	For C<ev::timer> and C<ev::periodic>, this invokes the corresponding
1462	C<ev_TYPE_again> function.	3434	C<ev_TYPE_again> function.
1463		3435
1464	=item w->sweep () C<ev::embed> only	3436	=item w->sweep () (C<ev::embed> only)
1465		3437
1466	Invokes C<ev_embed_sweep>.	3438	Invokes C<ev_embed_sweep>.
1467		3439
		3440	=item w->update () (C<ev::stat> only)
		3441
		3442	Invokes C<ev_stat_stat>.
		3443
1468	=back	3444	=back
1469		3445
1470	=back	3446	=back
1471		3447
1472	Example: Define a class with an IO and idle watcher, start one of them in	3448	Example: Define a class with two I/O and idle watchers, start the I/O
1473	the constructor.	3449	watchers in the constructor.
1474		3450
1475	class myclass	3451	class myclass
1476	{	3452	{
1477	ev_io io; void io_cb (ev::io &w, int revents);	3453	ev::io io ; void io_cb (ev::io &w, int revents);
		3454	ev::io2 io2 ; void io2_cb (ev::io &w, int revents);
1478	ev_idle idle void idle_cb (ev::idle &w, int revents);	3455	ev::idle idle; void idle_cb (ev::idle &w, int revents);
1479		3456
1480	myclass ();	3457	myclass (int fd)
		3458	{
		3459	io .set <myclass, &myclass::io_cb > (this);
		3460	io2 .set <myclass, &myclass::io2_cb > (this);
		3461	idle.set <myclass, &myclass::idle_cb> (this);
		3462
		3463	io.set (fd, ev::WRITE); // configure the watcher
		3464	io.start (); // start it whenever convenient
		3465
		3466	io2.start (fd, ev::READ); // set + start in one call
		3467	}
1481	}	3468	};
1482		3469
1483	myclass::myclass (int fd)	3470
1484	: io (this, &myclass::io_cb),	3471	=head1 OTHER LANGUAGE BINDINGS
1485	idle (this, &myclass::idle_cb)	3472
		3473	Libev does not offer other language bindings itself, but bindings for a
		3474	number of languages exist in the form of third-party packages. If you know
		3475	any interesting language binding in addition to the ones listed here, drop
		3476	me a note.
		3477
		3478	=over 4
		3479
		3480	=item Perl
		3481
		3482	The EV module implements the full libev API and is actually used to test
		3483	libev. EV is developed together with libev. Apart from the EV core module,
		3484	there are additional modules that implement libev-compatible interfaces
		3485	to C<libadns> (C<EV::ADNS>, but C<AnyEvent::DNS> is preferred nowadays),
		3486	C<Net::SNMP> (C<Net::SNMP::EV>) and the C<libglib> event core (C<Glib::EV>
		3487	and C<EV::Glib>).
		3488
		3489	It can be found and installed via CPAN, its homepage is at
		3490	L<http://software.schmorp.de/pkg/EV>.
		3491
		3492	=item Python
		3493
		3494	Python bindings can be found at L<http://code.google.com/p/pyev/>. It
		3495	seems to be quite complete and well-documented.
		3496
		3497	=item Ruby
		3498
		3499	Tony Arcieri has written a ruby extension that offers access to a subset
		3500	of the libev API and adds file handle abstractions, asynchronous DNS and
		3501	more on top of it. It can be found via gem servers. Its homepage is at
		3502	L<http://rev.rubyforge.org/>.
		3503
		3504	Roger Pack reports that using the link order C<-lws2_32 -lmsvcrt-ruby-190>
		3505	makes rev work even on mingw.
		3506
		3507	=item Haskell
		3508
		3509	A haskell binding to libev is available at
		3510	L<http://hackage.haskell.org/cgi-bin/hackage-scripts/package/hlibev>.
		3511
		3512	=item D
		3513
		3514	Leandro Lucarella has written a D language binding (F<ev.d>) for libev, to
		3515	be found at L<http://proj.llucax.com.ar/wiki/evd>.
		3516
		3517	=item Ocaml
		3518
		3519	Erkki Seppala has written Ocaml bindings for libev, to be found at
		3520	L<http://modeemi.cs.tut.fi/~flux/software/ocaml-ev/>.
		3521
		3522	=item Lua
		3523
		3524	Brian Maher has written a partial interface to libev for lua (at the
		3525	time of this writing, only C<ev_io> and C<ev_timer>), to be found at
		3526	L<http://github.com/brimworks/lua-ev>.
		3527
		3528	=back
		3529
		3530
		3531	=head1 MACRO MAGIC
		3532
		3533	Libev can be compiled with a variety of options, the most fundamental
		3534	of which is C<EV_MULTIPLICITY>. This option determines whether (most)
		3535	functions and callbacks have an initial C<struct ev_loop *> argument.
		3536
		3537	To make it easier to write programs that cope with either variant, the
		3538	following macros are defined:
		3539
		3540	=over 4
		3541
		3542	=item C<EV_A>, C<EV_A_>
		3543
		3544	This provides the loop I<argument> for functions, if one is required ("ev
		3545	loop argument"). The C<EV_A> form is used when this is the sole argument,
		3546	C<EV_A_> is used when other arguments are following. Example:
		3547
		3548	ev_unref (EV_A);
		3549	ev_timer_add (EV_A_ watcher);
		3550	ev_run (EV_A_ 0);
		3551
		3552	It assumes the variable C<loop> of type C<struct ev_loop *> is in scope,
		3553	which is often provided by the following macro.
		3554
		3555	=item C<EV_P>, C<EV_P_>
		3556
		3557	This provides the loop I<parameter> for functions, if one is required ("ev
		3558	loop parameter"). The C<EV_P> form is used when this is the sole parameter,
		3559	C<EV_P_> is used when other parameters are following. Example:
		3560
		3561	// this is how ev_unref is being declared
		3562	static void ev_unref (EV_P);
		3563
		3564	// this is how you can declare your typical callback
		3565	static void cb (EV_P_ ev_timer *w, int revents)
		3566
		3567	It declares a parameter C<loop> of type C<struct ev_loop *>, quite
		3568	suitable for use with C<EV_A>.
		3569
		3570	=item C<EV_DEFAULT>, C<EV_DEFAULT_>
		3571
		3572	Similar to the other two macros, this gives you the value of the default
		3573	loop, if multiple loops are supported ("ev loop default").
		3574
		3575	=item C<EV_DEFAULT_UC>, C<EV_DEFAULT_UC_>
		3576
		3577	Usage identical to C<EV_DEFAULT> and C<EV_DEFAULT_>, but requires that the
		3578	default loop has been initialised (C<UC> == unchecked). Their behaviour
		3579	is undefined when the default loop has not been initialised by a previous
		3580	execution of C<EV_DEFAULT>, C<EV_DEFAULT_> or C<ev_default_init (...)>.
		3581
		3582	It is often prudent to use C<EV_DEFAULT> when initialising the first
		3583	watcher in a function but use C<EV_DEFAULT_UC> afterwards.
		3584
		3585	=back
		3586
		3587	Example: Declare and initialise a check watcher, utilising the above
		3588	macros so it will work regardless of whether multiple loops are supported
		3589	or not.
		3590
		3591	static void
		3592	check_cb (EV_P_ ev_timer *w, int revents)
1486	{	3593	{
1487	io.start (fd, ev::READ);	3594	ev_check_stop (EV_A_ w);
1488	}	3595	}
		3596
		3597	ev_check check;
		3598	ev_check_init (&check, check_cb);
		3599	ev_check_start (EV_DEFAULT_ &check);
		3600	ev_run (EV_DEFAULT_ 0);
1489		3601
1490	=head1 EMBEDDING	3602	=head1 EMBEDDING
1491		3603
1492	Libev can (and often is) directly embedded into host	3604	Libev can (and often is) directly embedded into host
1493	applications. Examples of applications that embed it include the Deliantra	3605	applications. Examples of applications that embed it include the Deliantra
1494	Game Server, the EV perl module, the GNU Virtual Private Ethernet (gvpe)	3606	Game Server, the EV perl module, the GNU Virtual Private Ethernet (gvpe)
1495	and rxvt-unicode.	3607	and rxvt-unicode.
1496		3608
1497	The goal is to enable you to just copy the neecssary files into your	3609	The goal is to enable you to just copy the necessary files into your
1498	source directory without having to change even a single line in them, so	3610	source directory without having to change even a single line in them, so
1499	you can easily upgrade by simply copying (or having a checked-out copy of	3611	you can easily upgrade by simply copying (or having a checked-out copy of
1500	libev somewhere in your source tree).	3612	libev somewhere in your source tree).
1501		3613
1502	=head2 FILESETS	3614	=head2 FILESETS
1503		3615
1504	Depending on what features you need you need to include one or more sets of files	3616	Depending on what features you need you need to include one or more sets of files
1505	in your app.	3617	in your application.
1506		3618
1507	=head3 CORE EVENT LOOP	3619	=head3 CORE EVENT LOOP
1508		3620
1509	To include only the libev core (all the C<ev_*> functions), with manual	3621	To include only the libev core (all the C<ev_*> functions), with manual
1510	configuration (no autoconf):	3622	configuration (no autoconf):
1511		3623
1512	#define EV_STANDALONE 1	3624	#define EV_STANDALONE 1
1513	#include "ev.c"	3625	#include "ev.c"
1514		3626
1515	This will automatically include F<ev.h>, too, and should be done in a	3627	This will automatically include F<ev.h>, too, and should be done in a
1516	single C source file only to provide the function implementations. To use	3628	single C source file only to provide the function implementations. To use
1517	it, do the same for F<ev.h> in all files wishing to use this API (best	3629	it, do the same for F<ev.h> in all files wishing to use this API (best
1518	done by writing a wrapper around F<ev.h> that you can include instead and	3630	done by writing a wrapper around F<ev.h> that you can include instead and
1519	where you can put other configuration options):	3631	where you can put other configuration options):
1520		3632
1521	#define EV_STANDALONE 1	3633	#define EV_STANDALONE 1
1522	#include "ev.h"	3634	#include "ev.h"
1523		3635
1524	Both header files and implementation files can be compiled with a C++	3636	Both header files and implementation files can be compiled with a C++
1525	compiler (at least, thats a stated goal, and breakage will be treated	3637	compiler (at least, that's a stated goal, and breakage will be treated
1526	as a bug).	3638	as a bug).
1527		3639
1528	You need the following files in your source tree, or in a directory	3640	You need the following files in your source tree, or in a directory
1529	in your include path (e.g. in libev/ when using -Ilibev):	3641	in your include path (e.g. in libev/ when using -Ilibev):
1530		3642
1531	ev.h	3643	ev.h
1532	ev.c	3644	ev.c
1533	ev_vars.h	3645	ev_vars.h
1534	ev_wrap.h	3646	ev_wrap.h
1535		3647
1536	ev_win32.c required on win32 platforms only	3648	ev_win32.c required on win32 platforms only
1537		3649
1538	ev_select.c only when select backend is enabled (which is by default)	3650	ev_select.c only when select backend is enabled (which is enabled by default)
1539	ev_poll.c only when poll backend is enabled (disabled by default)	3651	ev_poll.c only when poll backend is enabled (disabled by default)
1540	ev_epoll.c only when the epoll backend is enabled (disabled by default)	3652	ev_epoll.c only when the epoll backend is enabled (disabled by default)
1541	ev_kqueue.c only when the kqueue backend is enabled (disabled by default)	3653	ev_kqueue.c only when the kqueue backend is enabled (disabled by default)
1542	ev_port.c only when the solaris port backend is enabled (disabled by default)	3654	ev_port.c only when the solaris port backend is enabled (disabled by default)
1543		3655
1544	F<ev.c> includes the backend files directly when enabled, so you only need	3656	F<ev.c> includes the backend files directly when enabled, so you only need
1545	to compile this single file.	3657	to compile this single file.
1546		3658
1547	=head3 LIBEVENT COMPATIBILITY API	3659	=head3 LIBEVENT COMPATIBILITY API
1548		3660
1549	To include the libevent compatibility API, also include:	3661	To include the libevent compatibility API, also include:
1550		3662
1551	#include "event.c"	3663	#include "event.c"
1552		3664
1553	in the file including F<ev.c>, and:	3665	in the file including F<ev.c>, and:
1554		3666
1555	#include "event.h"	3667	#include "event.h"
1556		3668
1557	in the files that want to use the libevent API. This also includes F<ev.h>.	3669	in the files that want to use the libevent API. This also includes F<ev.h>.
1558		3670
1559	You need the following additional files for this:	3671	You need the following additional files for this:
1560		3672
1561	event.h	3673	event.h
1562	event.c	3674	event.c
1563		3675
1564	=head3 AUTOCONF SUPPORT	3676	=head3 AUTOCONF SUPPORT
1565		3677
1566	Instead of using C<EV_STANDALONE=1> and providing your config in	3678	Instead of using C<EV_STANDALONE=1> and providing your configuration in
1567	whatever way you want, you can also C<m4_include([libev.m4])> in your	3679	whatever way you want, you can also C<m4_include([libev.m4])> in your
1568	F<configure.ac> and leave C<EV_STANDALONE> undefined. F<ev.c> will then	3680	F<configure.ac> and leave C<EV_STANDALONE> undefined. F<ev.c> will then
1569	include F<config.h> and configure itself accordingly.	3681	include F<config.h> and configure itself accordingly.
1570		3682
1571	For this of course you need the m4 file:	3683	For this of course you need the m4 file:
1572		3684
1573	libev.m4	3685	libev.m4
1574		3686
1575	=head2 PREPROCESSOR SYMBOLS/MACROS	3687	=head2 PREPROCESSOR SYMBOLS/MACROS
1576		3688
1577	Libev can be configured via a variety of preprocessor symbols you have to define	3689	Libev can be configured via a variety of preprocessor symbols you have to
1578	before including any of its files. The default is not to build for multiplicity	3690	define before including (or compiling) any of its files. The default in
1579	and only include the select backend.	3691	the absence of autoconf is documented for every option.
		3692
		3693	Symbols marked with "(h)" do not change the ABI, and can have different
		3694	values when compiling libev vs. including F<ev.h>, so it is permissible
		3695	to redefine them before including F<ev.h> without breaking compatibility
		3696	to a compiled library. All other symbols change the ABI, which means all
		3697	users of libev and the libev code itself must be compiled with compatible
		3698	settings.
1580		3699
1581	=over 4	3700	=over 4
1582		3701
		3702	=item EV_COMPAT3 (h)
		3703
		3704	Backwards compatibility is a major concern for libev. This is why this
		3705	release of libev comes with wrappers for the functions and symbols that
		3706	have been renamed between libev version 3 and 4.
		3707
		3708	You can disable these wrappers (to test compatibility with future
		3709	versions) by defining C<EV_COMPAT3> to C<0> when compiling your
		3710	sources. This has the additional advantage that you can drop the C<struct>
		3711	from C<struct ev_loop> declarations, as libev will provide an C<ev_loop>
		3712	typedef in that case.
		3713
		3714	In some future version, the default for C<EV_COMPAT3> will become C<0>,
		3715	and in some even more future version the compatibility code will be
		3716	removed completely.
		3717
1583	=item EV_STANDALONE	3718	=item EV_STANDALONE (h)
1584		3719
1585	Must always be C<1> if you do not use autoconf configuration, which	3720	Must always be C<1> if you do not use autoconf configuration, which
1586	keeps libev from including F<config.h>, and it also defines dummy	3721	keeps libev from including F<config.h>, and it also defines dummy
1587	implementations for some libevent functions (such as logging, which is not	3722	implementations for some libevent functions (such as logging, which is not
1588	supported). It will also not define any of the structs usually found in	3723	supported). It will also not define any of the structs usually found in
1589	F<event.h> that are not directly supported by the libev core alone.	3724	F<event.h> that are not directly supported by the libev core alone.
1590		3725
		3726	In standalone mode, libev will still try to automatically deduce the
		3727	configuration, but has to be more conservative.
		3728
1591	=item EV_USE_MONOTONIC	3729	=item EV_USE_MONOTONIC
1592		3730
1593	If defined to be C<1>, libev will try to detect the availability of the	3731	If defined to be C<1>, libev will try to detect the availability of the
1594	monotonic clock option at both compiletime and runtime. Otherwise no use	3732	monotonic clock option at both compile time and runtime. Otherwise no
1595	of the monotonic clock option will be attempted. If you enable this, you	3733	use of the monotonic clock option will be attempted. If you enable this,
1596	usually have to link against librt or something similar. Enabling it when	3734	you usually have to link against librt or something similar. Enabling it
1597	the functionality isn't available is safe, though, althoguh you have	3735	when the functionality isn't available is safe, though, although you have
1598	to make sure you link against any libraries where the C<clock_gettime>	3736	to make sure you link against any libraries where the C<clock_gettime>
1599	function is hiding in (often F<-lrt>).	3737	function is hiding in (often F<-lrt>). See also C<EV_USE_CLOCK_SYSCALL>.
1600		3738
1601	=item EV_USE_REALTIME	3739	=item EV_USE_REALTIME
1602		3740
1603	If defined to be C<1>, libev will try to detect the availability of the	3741	If defined to be C<1>, libev will try to detect the availability of the
1604	realtime clock option at compiletime (and assume its availability at	3742	real-time clock option at compile time (and assume its availability
1605	runtime if successful). Otherwise no use of the realtime clock option will	3743	at runtime if successful). Otherwise no use of the real-time clock
1606	be attempted. This effectively replaces C<gettimeofday> by C<clock_get	3744	option will be attempted. This effectively replaces C<gettimeofday>
1607	(CLOCK_REALTIME, ...)> and will not normally affect correctness. See tzhe note about libraries	3745	by C<clock_get (CLOCK_REALTIME, ...)> and will not normally affect
1608	in the description of C<EV_USE_MONOTONIC>, though.	3746	correctness. See the note about libraries in the description of
		3747	C<EV_USE_MONOTONIC>, though. Defaults to the opposite value of
		3748	C<EV_USE_CLOCK_SYSCALL>.
		3749
		3750	=item EV_USE_CLOCK_SYSCALL
		3751
		3752	If defined to be C<1>, libev will try to use a direct syscall instead
		3753	of calling the system-provided C<clock_gettime> function. This option
		3754	exists because on GNU/Linux, C<clock_gettime> is in C<librt>, but C<librt>
		3755	unconditionally pulls in C<libpthread>, slowing down single-threaded
		3756	programs needlessly. Using a direct syscall is slightly slower (in
		3757	theory), because no optimised vdso implementation can be used, but avoids
		3758	the pthread dependency. Defaults to C<1> on GNU/Linux with glibc 2.x or
		3759	higher, as it simplifies linking (no need for C<-lrt>).
		3760
		3761	=item EV_USE_NANOSLEEP
		3762
		3763	If defined to be C<1>, libev will assume that C<nanosleep ()> is available
		3764	and will use it for delays. Otherwise it will use C<select ()>.
		3765
		3766	=item EV_USE_EVENTFD
		3767
		3768	If defined to be C<1>, then libev will assume that C<eventfd ()> is
		3769	available and will probe for kernel support at runtime. This will improve
		3770	C<ev_signal> and C<ev_async> performance and reduce resource consumption.
		3771	If undefined, it will be enabled if the headers indicate GNU/Linux + Glibc
		3772	2.7 or newer, otherwise disabled.
1609		3773
1610	=item EV_USE_SELECT	3774	=item EV_USE_SELECT
1611		3775
1612	If undefined or defined to be C<1>, libev will compile in support for the	3776	If undefined or defined to be C<1>, libev will compile in support for the
1613	C<select>(2) backend. No attempt at autodetection will be done: if no	3777	C<select>(2) backend. No attempt at auto-detection will be done: if no
1614	other method takes over, select will be it. Otherwise the select backend	3778	other method takes over, select will be it. Otherwise the select backend
1615	will not be compiled in.	3779	will not be compiled in.
1616		3780
1617	=item EV_SELECT_USE_FD_SET	3781	=item EV_SELECT_USE_FD_SET
1618		3782
1619	If defined to C<1>, then the select backend will use the system C<fd_set>	3783	If defined to C<1>, then the select backend will use the system C<fd_set>
1620	structure. This is useful if libev doesn't compile due to a missing	3784	structure. This is useful if libev doesn't compile due to a missing
1621	C<NFDBITS> or C<fd_mask> definition or it misguesses the bitset layout on	3785	C<NFDBITS> or C<fd_mask> definition or it mis-guesses the bitset layout
1622	exotic systems. This usually limits the range of file descriptors to some	3786	on exotic systems. This usually limits the range of file descriptors to
1623	low limit such as 1024 or might have other limitations (winsocket only	3787	some low limit such as 1024 or might have other limitations (winsocket
1624	allows 64 sockets). The C<FD_SETSIZE> macro, set before compilation, might	3788	only allows 64 sockets). The C<FD_SETSIZE> macro, set before compilation,
1625	influence the size of the C<fd_set> used.	3789	configures the maximum size of the C<fd_set>.
1626		3790
1627	=item EV_SELECT_IS_WINSOCKET	3791	=item EV_SELECT_IS_WINSOCKET
1628		3792
1629	When defined to C<1>, the select backend will assume that	3793	When defined to C<1>, the select backend will assume that
1630	select/socket/connect etc. don't understand file descriptors but	3794	select/socket/connect etc. don't understand file descriptors but
…		…
1632	be used is the winsock select). This means that it will call	3796	be used is the winsock select). This means that it will call
1633	C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,	3797	C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,
1634	it is assumed that all these functions actually work on fds, even	3798	it is assumed that all these functions actually work on fds, even
1635	on win32. Should not be defined on non-win32 platforms.	3799	on win32. Should not be defined on non-win32 platforms.
1636		3800
		3801	=item EV_FD_TO_WIN32_HANDLE(fd)
		3802
		3803	If C<EV_SELECT_IS_WINSOCKET> is enabled, then libev needs a way to map
		3804	file descriptors to socket handles. When not defining this symbol (the
		3805	default), then libev will call C<_get_osfhandle>, which is usually
		3806	correct. In some cases, programs use their own file descriptor management,
		3807	in which case they can provide this function to map fds to socket handles.
		3808
		3809	=item EV_WIN32_HANDLE_TO_FD(handle)
		3810
		3811	If C<EV_SELECT_IS_WINSOCKET> then libev maps handles to file descriptors
		3812	using the standard C<_open_osfhandle> function. For programs implementing
		3813	their own fd to handle mapping, overwriting this function makes it easier
		3814	to do so. This can be done by defining this macro to an appropriate value.
		3815
		3816	=item EV_WIN32_CLOSE_FD(fd)
		3817
		3818	If programs implement their own fd to handle mapping on win32, then this
		3819	macro can be used to override the C<close> function, useful to unregister
		3820	file descriptors again. Note that the replacement function has to close
		3821	the underlying OS handle.
		3822
1637	=item EV_USE_POLL	3823	=item EV_USE_POLL
1638		3824
1639	If defined to be C<1>, libev will compile in support for the C<poll>(2)	3825	If defined to be C<1>, libev will compile in support for the C<poll>(2)
1640	backend. Otherwise it will be enabled on non-win32 platforms. It	3826	backend. Otherwise it will be enabled on non-win32 platforms. It
1641	takes precedence over select.	3827	takes precedence over select.
1642		3828
1643	=item EV_USE_EPOLL	3829	=item EV_USE_EPOLL
1644		3830
1645	If defined to be C<1>, libev will compile in support for the Linux	3831	If defined to be C<1>, libev will compile in support for the Linux
1646	C<epoll>(7) backend. Its availability will be detected at runtime,	3832	C<epoll>(7) backend. Its availability will be detected at runtime,
1647	otherwise another method will be used as fallback. This is the	3833	otherwise another method will be used as fallback. This is the preferred
1648	preferred backend for GNU/Linux systems.	3834	backend for GNU/Linux systems. If undefined, it will be enabled if the
		3835	headers indicate GNU/Linux + Glibc 2.4 or newer, otherwise disabled.
1649		3836
1650	=item EV_USE_KQUEUE	3837	=item EV_USE_KQUEUE
1651		3838
1652	If defined to be C<1>, libev will compile in support for the BSD style	3839	If defined to be C<1>, libev will compile in support for the BSD style
1653	C<kqueue>(2) backend. Its actual availability will be detected at runtime,	3840	C<kqueue>(2) backend. Its actual availability will be detected at runtime,
…		…
1666	otherwise another method will be used as fallback. This is the preferred	3853	otherwise another method will be used as fallback. This is the preferred
1667	backend for Solaris 10 systems.	3854	backend for Solaris 10 systems.
1668		3855
1669	=item EV_USE_DEVPOLL	3856	=item EV_USE_DEVPOLL
1670		3857
1671	reserved for future expansion, works like the USE symbols above.	3858	Reserved for future expansion, works like the USE symbols above.
1672		3859
		3860	=item EV_USE_INOTIFY
		3861
		3862	If defined to be C<1>, libev will compile in support for the Linux inotify
		3863	interface to speed up C<ev_stat> watchers. Its actual availability will
		3864	be detected at runtime. If undefined, it will be enabled if the headers
		3865	indicate GNU/Linux + Glibc 2.4 or newer, otherwise disabled.
		3866
		3867	=item EV_ATOMIC_T
		3868
		3869	Libev requires an integer type (suitable for storing C<0> or C<1>) whose
		3870	access is atomic with respect to other threads or signal contexts. No such
		3871	type is easily found in the C language, so you can provide your own type
		3872	that you know is safe for your purposes. It is used both for signal handler "locking"
		3873	as well as for signal and thread safety in C<ev_async> watchers.
		3874
		3875	In the absence of this define, libev will use C<sig_atomic_t volatile>
		3876	(from F<signal.h>), which is usually good enough on most platforms.
		3877
1673	=item EV_H	3878	=item EV_H (h)
1674		3879
1675	The name of the F<ev.h> header file used to include it. The default if	3880	The name of the F<ev.h> header file used to include it. The default if
1676	undefined is C<< <ev.h> >> in F<event.h> and C<"ev.h"> in F<ev.c>. This	3881	undefined is C<"ev.h"> in F<event.h>, F<ev.c> and F<ev++.h>. This can be
1677	can be used to virtually rename the F<ev.h> header file in case of conflicts.	3882	used to virtually rename the F<ev.h> header file in case of conflicts.
1678		3883
1679	=item EV_CONFIG_H	3884	=item EV_CONFIG_H (h)
1680		3885
1681	If C<EV_STANDALONE> isn't C<1>, this variable can be used to override	3886	If C<EV_STANDALONE> isn't C<1>, this variable can be used to override
1682	F<ev.c>'s idea of where to find the F<config.h> file, similarly to	3887	F<ev.c>'s idea of where to find the F<config.h> file, similarly to
1683	C<EV_H>, above.	3888	C<EV_H>, above.
1684		3889
1685	=item EV_EVENT_H	3890	=item EV_EVENT_H (h)
1686		3891
1687	Similarly to C<EV_H>, this macro can be used to override F<event.c>'s idea	3892	Similarly to C<EV_H>, this macro can be used to override F<event.c>'s idea
1688	of how the F<event.h> header can be found.	3893	of how the F<event.h> header can be found, the default is C<"event.h">.
1689		3894
1690	=item EV_PROTOTYPES	3895	=item EV_PROTOTYPES (h)
1691		3896
1692	If defined to be C<0>, then F<ev.h> will not define any function	3897	If defined to be C<0>, then F<ev.h> will not define any function
1693	prototypes, but still define all the structs and other symbols. This is	3898	prototypes, but still define all the structs and other symbols. This is
1694	occasionally useful if you want to provide your own wrapper functions	3899	occasionally useful if you want to provide your own wrapper functions
1695	around libev functions.	3900	around libev functions.
…		…
1700	will have the C<struct ev_loop *> as first argument, and you can create	3905	will have the C<struct ev_loop *> as first argument, and you can create
1701	additional independent event loops. Otherwise there will be no support	3906	additional independent event loops. Otherwise there will be no support
1702	for multiple event loops and there is no first event loop pointer	3907	for multiple event loops and there is no first event loop pointer
1703	argument. Instead, all functions act on the single default loop.	3908	argument. Instead, all functions act on the single default loop.
1704		3909
1705	=item EV_PERIODICS	3910	=item EV_MINPRI
1706		3911
1707	If undefined or defined to be C<1>, then periodic timers are supported,	3912	=item EV_MAXPRI
1708	otherwise not. This saves a few kb of code.	3913
		3914	The range of allowed priorities. C<EV_MINPRI> must be smaller or equal to
		3915	C<EV_MAXPRI>, but otherwise there are no non-obvious limitations. You can
		3916	provide for more priorities by overriding those symbols (usually defined
		3917	to be C<-2> and C<2>, respectively).
		3918
		3919	When doing priority-based operations, libev usually has to linearly search
		3920	all the priorities, so having many of them (hundreds) uses a lot of space
		3921	and time, so using the defaults of five priorities (-2 .. +2) is usually
		3922	fine.
		3923
		3924	If your embedding application does not need any priorities, defining these
		3925	both to C<0> will save some memory and CPU.
		3926
		3927	=item EV_PERIODIC_ENABLE, EV_IDLE_ENABLE, EV_EMBED_ENABLE, EV_STAT_ENABLE,
		3928	EV_PREPARE_ENABLE, EV_CHECK_ENABLE, EV_FORK_ENABLE, EV_SIGNAL_ENABLE,
		3929	EV_ASYNC_ENABLE, EV_CHILD_ENABLE.
		3930
		3931	If undefined or defined to be C<1> (and the platform supports it), then
		3932	the respective watcher type is supported. If defined to be C<0>, then it
		3933	is not. Disabling watcher types mainly saves code size.
		3934
		3935	=item EV_FEATURES
		3936
		3937	If you need to shave off some kilobytes of code at the expense of some
		3938	speed (but with the full API), you can define this symbol to request
		3939	certain subsets of functionality. The default is to enable all features
		3940	that can be enabled on the platform.
		3941
		3942	A typical way to use this symbol is to define it to C<0> (or to a bitset
		3943	with some broad features you want) and then selectively re-enable
		3944	additional parts you want, for example if you want everything minimal,
		3945	but multiple event loop support, async and child watchers and the poll
		3946	backend, use this:
		3947
		3948	#define EV_FEATURES 0
		3949	#define EV_MULTIPLICITY 1
		3950	#define EV_USE_POLL 1
		3951	#define EV_CHILD_ENABLE 1
		3952	#define EV_ASYNC_ENABLE 1
		3953
		3954	The actual value is a bitset, it can be a combination of the following
		3955	values:
		3956
		3957	=over 4
		3958
		3959	=item C<1> - faster/larger code
		3960
		3961	Use larger code to speed up some operations.
		3962
		3963	Currently this is used to override some inlining decisions (enlarging the
		3964	code size by roughly 30% on amd64).
		3965
		3966	When optimising for size, use of compiler flags such as C<-Os> with
		3967	gcc is recommended, as well as C<-DNDEBUG>, as libev contains a number of
		3968	assertions.
		3969
		3970	=item C<2> - faster/larger data structures
		3971
		3972	Replaces the small 2-heap for timer management by a faster 4-heap, larger
		3973	hash table sizes and so on. This will usually further increase code size
		3974	and can additionally have an effect on the size of data structures at
		3975	runtime.
		3976
		3977	=item C<4> - full API configuration
		3978
		3979	This enables priorities (sets C<EV_MAXPRI>=2 and C<EV_MINPRI>=-2), and
		3980	enables multiplicity (C<EV_MULTIPLICITY>=1).
		3981
		3982	=item C<8> - full API
		3983
		3984	This enables a lot of the "lesser used" API functions. See C<ev.h> for
		3985	details on which parts of the API are still available without this
		3986	feature, and do not complain if this subset changes over time.
		3987
		3988	=item C<16> - enable all optional watcher types
		3989
		3990	Enables all optional watcher types. If you want to selectively enable
		3991	only some watcher types other than I/O and timers (e.g. prepare,
		3992	embed, async, child...) you can enable them manually by defining
		3993	C<EV_watchertype_ENABLE> to C<1> instead.
		3994
		3995	=item C<32> - enable all backends
		3996
		3997	This enables all backends - without this feature, you need to enable at
		3998	least one backend manually (C<EV_USE_SELECT> is a good choice).
		3999
		4000	=item C<64> - enable OS-specific "helper" APIs
		4001
		4002	Enable inotify, eventfd, signalfd and similar OS-specific helper APIs by
		4003	default.
		4004
		4005	=back
		4006
		4007	Compiling with C<gcc -Os -DEV_STANDALONE -DEV_USE_EPOLL=1 -DEV_FEATURES=0>
		4008	reduces the compiled size of libev from 24.7Kb code/2.8Kb data to 6.5Kb
		4009	code/0.3Kb data on my GNU/Linux amd64 system, while still giving you I/O
		4010	watchers, timers and monotonic clock support.
		4011
		4012	With an intelligent-enough linker (gcc+binutils are intelligent enough
		4013	when you use C<-Wl,--gc-sections -ffunction-sections>) functions unused by
		4014	your program might be left out as well - a binary starting a timer and an
		4015	I/O watcher then might come out at only 5Kb.
		4016
		4017	=item EV_AVOID_STDIO
		4018
		4019	If this is set to C<1> at compiletime, then libev will avoid using stdio
		4020	functions (printf, scanf, perror etc.). This will increase the code size
		4021	somewhat, but if your program doesn't otherwise depend on stdio and your
		4022	libc allows it, this avoids linking in the stdio library which is quite
		4023	big.
		4024
		4025	Note that error messages might become less precise when this option is
		4026	enabled.
		4027
		4028	=item EV_NSIG
		4029
		4030	The highest supported signal number, +1 (or, the number of
		4031	signals): Normally, libev tries to deduce the maximum number of signals
		4032	automatically, but sometimes this fails, in which case it can be
		4033	specified. Also, using a lower number than detected (C<32> should be
		4034	good for about any system in existence) can save some memory, as libev
		4035	statically allocates some 12-24 bytes per signal number.
		4036
		4037	=item EV_PID_HASHSIZE
		4038
		4039	C<ev_child> watchers use a small hash table to distribute workload by
		4040	pid. The default size is C<16> (or C<1> with C<EV_FEATURES> disabled),
		4041	usually more than enough. If you need to manage thousands of children you
		4042	might want to increase this value (I<must> be a power of two).
		4043
		4044	=item EV_INOTIFY_HASHSIZE
		4045
		4046	C<ev_stat> watchers use a small hash table to distribute workload by
		4047	inotify watch id. The default size is C<16> (or C<1> with C<EV_FEATURES>
		4048	disabled), usually more than enough. If you need to manage thousands of
		4049	C<ev_stat> watchers you might want to increase this value (I<must> be a
		4050	power of two).
		4051
		4052	=item EV_USE_4HEAP
		4053
		4054	Heaps are not very cache-efficient. To improve the cache-efficiency of the
		4055	timer and periodics heaps, libev uses a 4-heap when this symbol is defined
		4056	to C<1>. The 4-heap uses more complicated (longer) code but has noticeably
		4057	faster performance with many (thousands) of watchers.
		4058
		4059	The default is C<1>, unless C<EV_FEATURES> overrides it, in which case it
		4060	will be C<0>.
		4061
		4062	=item EV_HEAP_CACHE_AT
		4063
		4064	Heaps are not very cache-efficient. To improve the cache-efficiency of the
		4065	timer and periodics heaps, libev can cache the timestamp (I<at>) within
		4066	the heap structure (selected by defining C<EV_HEAP_CACHE_AT> to C<1>),
		4067	which uses 8-12 bytes more per watcher and a few hundred bytes more code,
		4068	but avoids random read accesses on heap changes. This improves performance
		4069	noticeably with many (hundreds) of watchers.
		4070
		4071	The default is C<1>, unless C<EV_FEATURES> overrides it, in which case it
		4072	will be C<0>.
		4073
		4074	=item EV_VERIFY
		4075
		4076	Controls how much internal verification (see C<ev_verify ()>) will
		4077	be done: If set to C<0>, no internal verification code will be compiled
		4078	in. If set to C<1>, then verification code will be compiled in, but not
		4079	called. If set to C<2>, then the internal verification code will be
		4080	called once per loop, which can slow down libev. If set to C<3>, then the
		4081	verification code will be called very frequently, which will slow down
		4082	libev considerably.
		4083
		4084	The default is C<1>, unless C<EV_FEATURES> overrides it, in which case it
		4085	will be C<0>.
1709		4086
1710	=item EV_COMMON	4087	=item EV_COMMON
1711		4088
1712	By default, all watchers have a C<void *data> member. By redefining	4089	By default, all watchers have a C<void *data> member. By redefining
1713	this macro to a something else you can include more and other types of	4090	this macro to something else you can include more and other types of
1714	members. You have to define it each time you include one of the files,	4091	members. You have to define it each time you include one of the files,
1715	though, and it must be identical each time.	4092	though, and it must be identical each time.
1716		4093
1717	For example, the perl EV module uses something like this:	4094	For example, the perl EV module uses something like this:
1718		4095
1719	#define EV_COMMON \	4096	#define EV_COMMON \
1720	SV *self; /* contains this struct */ \	4097	SV *self; /* contains this struct */ \
1721	SV *cb_sv, *fh /* note no trailing ";" */	4098	SV *cb_sv, *fh /* note no trailing ";" */
1722		4099
1723	=item EV_CB_DECLARE (type)	4100	=item EV_CB_DECLARE (type)
1724		4101
1725	=item EV_CB_INVOKE (watcher, revents)	4102	=item EV_CB_INVOKE (watcher, revents)
1726		4103
1727	=item ev_set_cb (ev, cb)	4104	=item ev_set_cb (ev, cb)
1728		4105
1729	Can be used to change the callback member declaration in each watcher,	4106	Can be used to change the callback member declaration in each watcher,
1730	and the way callbacks are invoked and set. Must expand to a struct member	4107	and the way callbacks are invoked and set. Must expand to a struct member
1731	definition and a statement, respectively. See the F<ev.v> header file for	4108	definition and a statement, respectively. See the F<ev.h> header file for
1732	their default definitions. One possible use for overriding these is to	4109	their default definitions. One possible use for overriding these is to
1733	avoid the C<struct ev_loop *> as first argument in all cases, or to use	4110	avoid the C<struct ev_loop *> as first argument in all cases, or to use
1734	method calls instead of plain function calls in C++.	4111	method calls instead of plain function calls in C++.
		4112
		4113	=back
		4114
		4115	=head2 EXPORTED API SYMBOLS
		4116
		4117	If you need to re-export the API (e.g. via a DLL) and you need a list of
		4118	exported symbols, you can use the provided F<Symbol.*> files which list
		4119	all public symbols, one per line:
		4120
		4121	Symbols.ev for libev proper
		4122	Symbols.event for the libevent emulation
		4123
		4124	This can also be used to rename all public symbols to avoid clashes with
		4125	multiple versions of libev linked together (which is obviously bad in
		4126	itself, but sometimes it is inconvenient to avoid this).
		4127
		4128	A sed command like this will create wrapper C<#define>'s that you need to
		4129	include before including F<ev.h>:
		4130
		4131	<Symbols.ev sed -e "s/.*/#define & myprefix_&/" >wrap.h
		4132
		4133	This would create a file F<wrap.h> which essentially looks like this:
		4134
		4135	#define ev_backend myprefix_ev_backend
		4136	#define ev_check_start myprefix_ev_check_start
		4137	#define ev_check_stop myprefix_ev_check_stop
		4138	...
1735		4139
1736	=head2 EXAMPLES	4140	=head2 EXAMPLES
1737		4141
1738	For a real-world example of a program the includes libev	4142	For a real-world example of a program the includes libev
1739	verbatim, you can have a look at the EV perl module	4143	verbatim, you can have a look at the EV perl module
…		…
1742	interface) and F<EV.xs> (implementation) files. Only the F<EV.xs> file	4146	interface) and F<EV.xs> (implementation) files. Only the F<EV.xs> file
1743	will be compiled. It is pretty complex because it provides its own header	4147	will be compiled. It is pretty complex because it provides its own header
1744	file.	4148	file.
1745		4149
1746	The usage in rxvt-unicode is simpler. It has a F<ev_cpp.h> header file	4150	The usage in rxvt-unicode is simpler. It has a F<ev_cpp.h> header file
1747	that everybody includes and which overrides some autoconf choices:	4151	that everybody includes and which overrides some configure choices:
1748		4152
		4153	#define EV_FEATURES 8
		4154	#define EV_USE_SELECT 1
		4155	#define EV_PREPARE_ENABLE 1
		4156	#define EV_IDLE_ENABLE 1
		4157	#define EV_SIGNAL_ENABLE 1
		4158	#define EV_CHILD_ENABLE 1
1749	#define EV_USE_POLL 0	4159	#define EV_USE_STDEXCEPT 0
1750	#define EV_MULTIPLICITY 0
1751	#define EV_PERIODICS 0
1752	#define EV_CONFIG_H <config.h>	4160	#define EV_CONFIG_H <config.h>
1753		4161
1754	#include "ev++.h"	4162	#include "ev++.h"
1755		4163
1756	And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:	4164	And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:
1757		4165
1758	#include "ev_cpp.h"	4166	#include "ev_cpp.h"
1759	#include "ev.c"	4167	#include "ev.c"
1760		4168
		4169	=head1 INTERACTION WITH OTHER PROGRAMS OR LIBRARIES
1761		4170
		4171	=head2 THREADS AND COROUTINES
		4172
		4173	=head3 THREADS
		4174
		4175	All libev functions are reentrant and thread-safe unless explicitly
		4176	documented otherwise, but libev implements no locking itself. This means
		4177	that you can use as many loops as you want in parallel, as long as there
		4178	are no concurrent calls into any libev function with the same loop
		4179	parameter (C<ev_default_*> calls have an implicit default loop parameter,
		4180	of course): libev guarantees that different event loops share no data
		4181	structures that need any locking.
		4182
		4183	Or to put it differently: calls with different loop parameters can be done
		4184	concurrently from multiple threads, calls with the same loop parameter
		4185	must be done serially (but can be done from different threads, as long as
		4186	only one thread ever is inside a call at any point in time, e.g. by using
		4187	a mutex per loop).
		4188
		4189	Specifically to support threads (and signal handlers), libev implements
		4190	so-called C<ev_async> watchers, which allow some limited form of
		4191	concurrency on the same event loop, namely waking it up "from the
		4192	outside".
		4193
		4194	If you want to know which design (one loop, locking, or multiple loops
		4195	without or something else still) is best for your problem, then I cannot
		4196	help you, but here is some generic advice:
		4197
		4198	=over 4
		4199
		4200	=item * most applications have a main thread: use the default libev loop
		4201	in that thread, or create a separate thread running only the default loop.
		4202
		4203	This helps integrating other libraries or software modules that use libev
		4204	themselves and don't care/know about threading.
		4205
		4206	=item * one loop per thread is usually a good model.
		4207
		4208	Doing this is almost never wrong, sometimes a better-performance model
		4209	exists, but it is always a good start.
		4210
		4211	=item * other models exist, such as the leader/follower pattern, where one
		4212	loop is handed through multiple threads in a kind of round-robin fashion.
		4213
		4214	Choosing a model is hard - look around, learn, know that usually you can do
		4215	better than you currently do :-)
		4216
		4217	=item * often you need to talk to some other thread which blocks in the
		4218	event loop.
		4219
		4220	C<ev_async> watchers can be used to wake them up from other threads safely
		4221	(or from signal contexts...).
		4222
		4223	An example use would be to communicate signals or other events that only
		4224	work in the default loop by registering the signal watcher with the
		4225	default loop and triggering an C<ev_async> watcher from the default loop
		4226	watcher callback into the event loop interested in the signal.
		4227
		4228	=back
		4229
		4230	=head4 THREAD LOCKING EXAMPLE
		4231
		4232	Here is a fictitious example of how to run an event loop in a different
		4233	thread than where callbacks are being invoked and watchers are
		4234	created/added/removed.
		4235
		4236	For a real-world example, see the C<EV::Loop::Async> perl module,
		4237	which uses exactly this technique (which is suited for many high-level
		4238	languages).
		4239
		4240	The example uses a pthread mutex to protect the loop data, a condition
		4241	variable to wait for callback invocations, an async watcher to notify the
		4242	event loop thread and an unspecified mechanism to wake up the main thread.
		4243
		4244	First, you need to associate some data with the event loop:
		4245
		4246	typedef struct {
		4247	mutex_t lock; /* global loop lock */
		4248	ev_async async_w;
		4249	thread_t tid;
		4250	cond_t invoke_cv;
		4251	} userdata;
		4252
		4253	void prepare_loop (EV_P)
		4254	{
		4255	// for simplicity, we use a static userdata struct.
		4256	static userdata u;
		4257
		4258	ev_async_init (&u->async_w, async_cb);
		4259	ev_async_start (EV_A_ &u->async_w);
		4260
		4261	pthread_mutex_init (&u->lock, 0);
		4262	pthread_cond_init (&u->invoke_cv, 0);
		4263
		4264	// now associate this with the loop
		4265	ev_set_userdata (EV_A_ u);
		4266	ev_set_invoke_pending_cb (EV_A_ l_invoke);
		4267	ev_set_loop_release_cb (EV_A_ l_release, l_acquire);
		4268
		4269	// then create the thread running ev_loop
		4270	pthread_create (&u->tid, 0, l_run, EV_A);
		4271	}
		4272
		4273	The callback for the C<ev_async> watcher does nothing: the watcher is used
		4274	solely to wake up the event loop so it takes notice of any new watchers
		4275	that might have been added:
		4276
		4277	static void
		4278	async_cb (EV_P_ ev_async *w, int revents)
		4279	{
		4280	// just used for the side effects
		4281	}
		4282
		4283	The C<l_release> and C<l_acquire> callbacks simply unlock/lock the mutex
		4284	protecting the loop data, respectively.
		4285
		4286	static void
		4287	l_release (EV_P)
		4288	{
		4289	userdata *u = ev_userdata (EV_A);
		4290	pthread_mutex_unlock (&u->lock);
		4291	}
		4292
		4293	static void
		4294	l_acquire (EV_P)
		4295	{
		4296	userdata *u = ev_userdata (EV_A);
		4297	pthread_mutex_lock (&u->lock);
		4298	}
		4299
		4300	The event loop thread first acquires the mutex, and then jumps straight
		4301	into C<ev_run>:
		4302
		4303	void *
		4304	l_run (void *thr_arg)
		4305	{
		4306	struct ev_loop *loop = (struct ev_loop *)thr_arg;
		4307
		4308	l_acquire (EV_A);
		4309	pthread_setcanceltype (PTHREAD_CANCEL_ASYNCHRONOUS, 0);
		4310	ev_run (EV_A_ 0);
		4311	l_release (EV_A);
		4312
		4313	return 0;
		4314	}
		4315
		4316	Instead of invoking all pending watchers, the C<l_invoke> callback will
		4317	signal the main thread via some unspecified mechanism (signals? pipe
		4318	writes? C<Async::Interrupt>?) and then waits until all pending watchers
		4319	have been called (in a while loop because a) spurious wakeups are possible
		4320	and b) skipping inter-thread-communication when there are no pending
		4321	watchers is very beneficial):
		4322
		4323	static void
		4324	l_invoke (EV_P)
		4325	{
		4326	userdata *u = ev_userdata (EV_A);
		4327
		4328	while (ev_pending_count (EV_A))
		4329	{
		4330	wake_up_other_thread_in_some_magic_or_not_so_magic_way ();
		4331	pthread_cond_wait (&u->invoke_cv, &u->lock);
		4332	}
		4333	}
		4334
		4335	Now, whenever the main thread gets told to invoke pending watchers, it
		4336	will grab the lock, call C<ev_invoke_pending> and then signal the loop
		4337	thread to continue:
		4338
		4339	static void
		4340	real_invoke_pending (EV_P)
		4341	{
		4342	userdata *u = ev_userdata (EV_A);
		4343
		4344	pthread_mutex_lock (&u->lock);
		4345	ev_invoke_pending (EV_A);
		4346	pthread_cond_signal (&u->invoke_cv);
		4347	pthread_mutex_unlock (&u->lock);
		4348	}
		4349
		4350	Whenever you want to start/stop a watcher or do other modifications to an
		4351	event loop, you will now have to lock:
		4352
		4353	ev_timer timeout_watcher;
		4354	userdata *u = ev_userdata (EV_A);
		4355
		4356	ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);
		4357
		4358	pthread_mutex_lock (&u->lock);
		4359	ev_timer_start (EV_A_ &timeout_watcher);
		4360	ev_async_send (EV_A_ &u->async_w);
		4361	pthread_mutex_unlock (&u->lock);
		4362
		4363	Note that sending the C<ev_async> watcher is required because otherwise
		4364	an event loop currently blocking in the kernel will have no knowledge
		4365	about the newly added timer. By waking up the loop it will pick up any new
		4366	watchers in the next event loop iteration.
		4367
		4368	=head3 COROUTINES
		4369
		4370	Libev is very accommodating to coroutines ("cooperative threads"):
		4371	libev fully supports nesting calls to its functions from different
		4372	coroutines (e.g. you can call C<ev_run> on the same loop from two
		4373	different coroutines, and switch freely between both coroutines running
		4374	the loop, as long as you don't confuse yourself). The only exception is
		4375	that you must not do this from C<ev_periodic> reschedule callbacks.
		4376
		4377	Care has been taken to ensure that libev does not keep local state inside
		4378	C<ev_run>, and other calls do not usually allow for coroutine switches as
		4379	they do not call any callbacks.
		4380
		4381	=head2 COMPILER WARNINGS
		4382
		4383	Depending on your compiler and compiler settings, you might get no or a
		4384	lot of warnings when compiling libev code. Some people are apparently
		4385	scared by this.
		4386
		4387	However, these are unavoidable for many reasons. For one, each compiler
		4388	has different warnings, and each user has different tastes regarding
		4389	warning options. "Warn-free" code therefore cannot be a goal except when
		4390	targeting a specific compiler and compiler-version.
		4391
		4392	Another reason is that some compiler warnings require elaborate
		4393	workarounds, or other changes to the code that make it less clear and less
		4394	maintainable.
		4395
		4396	And of course, some compiler warnings are just plain stupid, or simply
		4397	wrong (because they don't actually warn about the condition their message
		4398	seems to warn about). For example, certain older gcc versions had some
		4399	warnings that resulted in an extreme number of false positives. These have
		4400	been fixed, but some people still insist on making code warn-free with
		4401	such buggy versions.
		4402
		4403	While libev is written to generate as few warnings as possible,
		4404	"warn-free" code is not a goal, and it is recommended not to build libev
		4405	with any compiler warnings enabled unless you are prepared to cope with
		4406	them (e.g. by ignoring them). Remember that warnings are just that:
		4407	warnings, not errors, or proof of bugs.
		4408
		4409
		4410	=head2 VALGRIND
		4411
		4412	Valgrind has a special section here because it is a popular tool that is
		4413	highly useful. Unfortunately, valgrind reports are very hard to interpret.
		4414
		4415	If you think you found a bug (memory leak, uninitialised data access etc.)
		4416	in libev, then check twice: If valgrind reports something like:
		4417
		4418	==2274== definitely lost: 0 bytes in 0 blocks.
		4419	==2274== possibly lost: 0 bytes in 0 blocks.
		4420	==2274== still reachable: 256 bytes in 1 blocks.
		4421
		4422	Then there is no memory leak, just as memory accounted to global variables
		4423	is not a memleak - the memory is still being referenced, and didn't leak.
		4424
		4425	Similarly, under some circumstances, valgrind might report kernel bugs
		4426	as if it were a bug in libev (e.g. in realloc or in the poll backend,
		4427	although an acceptable workaround has been found here), or it might be
		4428	confused.
		4429
		4430	Keep in mind that valgrind is a very good tool, but only a tool. Don't
		4431	make it into some kind of religion.
		4432
		4433	If you are unsure about something, feel free to contact the mailing list
		4434	with the full valgrind report and an explanation on why you think this
		4435	is a bug in libev (best check the archives, too :). However, don't be
		4436	annoyed when you get a brisk "this is no bug" answer and take the chance
		4437	of learning how to interpret valgrind properly.
		4438
		4439	If you need, for some reason, empty reports from valgrind for your project
		4440	I suggest using suppression lists.
		4441
		4442
		4443	=head1 PORTABILITY NOTES
		4444
		4445	=head2 GNU/LINUX 32 BIT LIMITATIONS
		4446
		4447	GNU/Linux is the only common platform that supports 64 bit file/large file
		4448	interfaces but I<disables> them by default.
		4449
		4450	That means that libev compiled in the default environment doesn't support
		4451	files larger than 2GiB or so, which mainly affects C<ev_stat> watchers.
		4452
		4453	Unfortunately, many programs try to work around this GNU/Linux issue
		4454	by enabling the large file API, which makes them incompatible with the
		4455	standard libev compiled for their system.
		4456
		4457	Likewise, libev cannot enable the large file API itself as this would
		4458	suddenly make it incompatible to the default compile time environment,
		4459	i.e. all programs not using special compile switches.
		4460
		4461	=head2 OS/X AND DARWIN BUGS
		4462
		4463	The whole thing is a bug if you ask me - basically any system interface
		4464	you touch is broken, whether it is locales, poll, kqueue or even the
		4465	OpenGL drivers.
		4466
		4467	=head3 C<kqueue> is buggy
		4468
		4469	The kqueue syscall is broken in all known versions - most versions support
		4470	only sockets, many support pipes.
		4471
		4472	Libev tries to work around this by not using C<kqueue> by default on
		4473	this rotten platform, but of course you can still ask for it when creating
		4474	a loop.
		4475
		4476	=head3 C<poll> is buggy
		4477
		4478	Instead of fixing C<kqueue>, Apple replaced their (working) C<poll>
		4479	implementation by something calling C<kqueue> internally around the 10.5.6
		4480	release, so now C<kqueue> I<and> C<poll> are broken.
		4481
		4482	Libev tries to work around this by not using C<poll> by default on
		4483	this rotten platform, but of course you can still ask for it when creating
		4484	a loop.
		4485
		4486	=head3 C<select> is buggy
		4487
		4488	All that's left is C<select>, and of course Apple found a way to fuck this
		4489	one up as well: On OS/X, C<select> actively limits the number of file
		4490	descriptors you can pass in to 1024 - your program suddenly crashes when
		4491	you use more.
		4492
		4493	There is an undocumented "workaround" for this - defining
		4494	C<_DARWIN_UNLIMITED_SELECT>, which libev tries to use, so select I<should>
		4495	work on OS/X.
		4496
		4497	=head2 SOLARIS PROBLEMS AND WORKAROUNDS
		4498
		4499	=head3 C<errno> reentrancy
		4500
		4501	The default compile environment on Solaris is unfortunately so
		4502	thread-unsafe that you can't even use components/libraries compiled
		4503	without C<-D_REENTRANT> (as long as they use C<errno>), which, of course,
		4504	isn't defined by default.
		4505
		4506	If you want to use libev in threaded environments you have to make sure
		4507	it's compiled with C<_REENTRANT> defined.
		4508
		4509	=head3 Event port backend
		4510
		4511	The scalable event interface for Solaris is called "event ports". Unfortunately,
		4512	this mechanism is very buggy. If you run into high CPU usage, your program
		4513	freezes or you get a large number of spurious wakeups, make sure you have
		4514	all the relevant and latest kernel patches applied. No, I don't know which
		4515	ones, but there are multiple ones.
		4516
		4517	If you can't get it to work, you can try running the program by setting
		4518	the environment variable C<LIBEV_FLAGS=3> to only allow C<poll> and
		4519	C<select> backends.
		4520
		4521	=head2 AIX POLL BUG
		4522
		4523	AIX unfortunately has a broken C<poll.h> header. Libev works around
		4524	this by trying to avoid the poll backend altogether (i.e. it's not even
		4525	compiled in), which normally isn't a big problem as C<select> works fine
		4526	with large bitsets, and AIX is dead anyway.
		4527
		4528	=head2 WIN32 PLATFORM LIMITATIONS AND WORKAROUNDS
		4529
		4530	=head3 General issues
		4531
		4532	Win32 doesn't support any of the standards (e.g. POSIX) that libev
		4533	requires, and its I/O model is fundamentally incompatible with the POSIX
		4534	model. Libev still offers limited functionality on this platform in
		4535	the form of the C<EVBACKEND_SELECT> backend, and only supports socket
		4536	descriptors. This only applies when using Win32 natively, not when using
		4537	e.g. cygwin. Actually, it only applies to the microsofts own compilers,
		4538	as every compielr comes with a slightly differently broken/incompatible
		4539	environment.
		4540
		4541	Lifting these limitations would basically require the full
		4542	re-implementation of the I/O system. If you are into this kind of thing,
		4543	then note that glib does exactly that for you in a very portable way (note
		4544	also that glib is the slowest event library known to man).
		4545
		4546	There is no supported compilation method available on windows except
		4547	embedding it into other applications.
		4548
		4549	Sensible signal handling is officially unsupported by Microsoft - libev
		4550	tries its best, but under most conditions, signals will simply not work.
		4551
		4552	Not a libev limitation but worth mentioning: windows apparently doesn't
		4553	accept large writes: instead of resulting in a partial write, windows will
		4554	either accept everything or return C<ENOBUFS> if the buffer is too large,
		4555	so make sure you only write small amounts into your sockets (less than a
		4556	megabyte seems safe, but this apparently depends on the amount of memory
		4557	available).
		4558
		4559	Due to the many, low, and arbitrary limits on the win32 platform and
		4560	the abysmal performance of winsockets, using a large number of sockets
		4561	is not recommended (and not reasonable). If your program needs to use
		4562	more than a hundred or so sockets, then likely it needs to use a totally
		4563	different implementation for windows, as libev offers the POSIX readiness
		4564	notification model, which cannot be implemented efficiently on windows
		4565	(due to Microsoft monopoly games).
		4566
		4567	A typical way to use libev under windows is to embed it (see the embedding
		4568	section for details) and use the following F<evwrap.h> header file instead
		4569	of F<ev.h>:
		4570
		4571	#define EV_STANDALONE /* keeps ev from requiring config.h */
		4572	#define EV_SELECT_IS_WINSOCKET 1 /* configure libev for windows select */
		4573
		4574	#include "ev.h"
		4575
		4576	And compile the following F<evwrap.c> file into your project (make sure
		4577	you do I<not> compile the F<ev.c> or any other embedded source files!):
		4578
		4579	#include "evwrap.h"
		4580	#include "ev.c"
		4581
		4582	=head3 The winsocket C<select> function
		4583
		4584	The winsocket C<select> function doesn't follow POSIX in that it
		4585	requires socket I<handles> and not socket I<file descriptors> (it is
		4586	also extremely buggy). This makes select very inefficient, and also
		4587	requires a mapping from file descriptors to socket handles (the Microsoft
		4588	C runtime provides the function C<_open_osfhandle> for this). See the
		4589	discussion of the C<EV_SELECT_USE_FD_SET>, C<EV_SELECT_IS_WINSOCKET> and
		4590	C<EV_FD_TO_WIN32_HANDLE> preprocessor symbols for more info.
		4591
		4592	The configuration for a "naked" win32 using the Microsoft runtime
		4593	libraries and raw winsocket select is:
		4594
		4595	#define EV_USE_SELECT 1
		4596	#define EV_SELECT_IS_WINSOCKET 1 /* forces EV_SELECT_USE_FD_SET, too */
		4597
		4598	Note that winsockets handling of fd sets is O(n), so you can easily get a
		4599	complexity in the O(n²) range when using win32.
		4600
		4601	=head3 Limited number of file descriptors
		4602
		4603	Windows has numerous arbitrary (and low) limits on things.
		4604
		4605	Early versions of winsocket's select only supported waiting for a maximum
		4606	of C<64> handles (probably owning to the fact that all windows kernels
		4607	can only wait for C<64> things at the same time internally; Microsoft
		4608	recommends spawning a chain of threads and wait for 63 handles and the
		4609	previous thread in each. Sounds great!).
		4610
		4611	Newer versions support more handles, but you need to define C<FD_SETSIZE>
		4612	to some high number (e.g. C<2048>) before compiling the winsocket select
		4613	call (which might be in libev or elsewhere, for example, perl and many
		4614	other interpreters do their own select emulation on windows).
		4615
		4616	Another limit is the number of file descriptors in the Microsoft runtime
		4617	libraries, which by default is C<64> (there must be a hidden I<64>
		4618	fetish or something like this inside Microsoft). You can increase this
		4619	by calling C<_setmaxstdio>, which can increase this limit to C<2048>
		4620	(another arbitrary limit), but is broken in many versions of the Microsoft
		4621	runtime libraries. This might get you to about C<512> or C<2048> sockets
		4622	(depending on windows version and/or the phase of the moon). To get more,
		4623	you need to wrap all I/O functions and provide your own fd management, but
		4624	the cost of calling select (O(n²)) will likely make this unworkable.
		4625
		4626	=head2 PORTABILITY REQUIREMENTS
		4627
		4628	In addition to a working ISO-C implementation and of course the
		4629	backend-specific APIs, libev relies on a few additional extensions:
		4630
		4631	=over 4
		4632
		4633	=item C<void (*)(ev_watcher_type *, int revents)> must have compatible
		4634	calling conventions regardless of C<ev_watcher_type *>.
		4635
		4636	Libev assumes not only that all watcher pointers have the same internal
		4637	structure (guaranteed by POSIX but not by ISO C for example), but it also
		4638	assumes that the same (machine) code can be used to call any watcher
		4639	callback: The watcher callbacks have different type signatures, but libev
		4640	calls them using an C<ev_watcher *> internally.
		4641
		4642	=item C<sig_atomic_t volatile> must be thread-atomic as well
		4643
		4644	The type C<sig_atomic_t volatile> (or whatever is defined as
		4645	C<EV_ATOMIC_T>) must be atomic with respect to accesses from different
		4646	threads. This is not part of the specification for C<sig_atomic_t>, but is
		4647	believed to be sufficiently portable.
		4648
		4649	=item C<sigprocmask> must work in a threaded environment
		4650
		4651	Libev uses C<sigprocmask> to temporarily block signals. This is not
		4652	allowed in a threaded program (C<pthread_sigmask> has to be used). Typical
		4653	pthread implementations will either allow C<sigprocmask> in the "main
		4654	thread" or will block signals process-wide, both behaviours would
		4655	be compatible with libev. Interaction between C<sigprocmask> and
		4656	C<pthread_sigmask> could complicate things, however.
		4657
		4658	The most portable way to handle signals is to block signals in all threads
		4659	except the initial one, and run the default loop in the initial thread as
		4660	well.
		4661
		4662	=item C<long> must be large enough for common memory allocation sizes
		4663
		4664	To improve portability and simplify its API, libev uses C<long> internally
		4665	instead of C<size_t> when allocating its data structures. On non-POSIX
		4666	systems (Microsoft...) this might be unexpectedly low, but is still at
		4667	least 31 bits everywhere, which is enough for hundreds of millions of
		4668	watchers.
		4669
		4670	=item C<double> must hold a time value in seconds with enough accuracy
		4671
		4672	The type C<double> is used to represent timestamps. It is required to
		4673	have at least 51 bits of mantissa (and 9 bits of exponent), which is
		4674	good enough for at least into the year 4000 with millisecond accuracy
		4675	(the design goal for libev). This requirement is overfulfilled by
		4676	implementations using IEEE 754, which is basically all existing ones. With
		4677	IEEE 754 doubles, you get microsecond accuracy until at least 2200.
		4678
		4679	=back
		4680
		4681	If you know of other additional requirements drop me a note.
		4682
		4683
1762	=head1 COMPLEXITIES	4684	=head1 ALGORITHMIC COMPLEXITIES
1763		4685
1764	In this section the complexities of (many of) the algorithms used inside	4686	In this section the complexities of (many of) the algorithms used inside
1765	libev will be explained. For complexity discussions about backends see the	4687	libev will be documented. For complexity discussions about backends see
1766	documentation for C<ev_default_init>.	4688	the documentation for C<ev_default_init>.
		4689
		4690	All of the following are about amortised time: If an array needs to be
		4691	extended, libev needs to realloc and move the whole array, but this
		4692	happens asymptotically rarer with higher number of elements, so O(1) might
		4693	mean that libev does a lengthy realloc operation in rare cases, but on
		4694	average it is much faster and asymptotically approaches constant time.
1767		4695
1768	=over 4	4696	=over 4
1769		4697
1770	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)	4698	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)
1771		4699
		4700	This means that, when you have a watcher that triggers in one hour and
		4701	there are 100 watchers that would trigger before that, then inserting will
		4702	have to skip roughly seven (C<ld 100>) of these watchers.
		4703
1772	=item Changing timer/periodic watchers (by autorepeat, again): O(log skipped_other_timers)	4704	=item Changing timer/periodic watchers (by autorepeat or calling again): O(log skipped_other_timers)
1773		4705
		4706	That means that changing a timer costs less than removing/adding them,
		4707	as only the relative motion in the event queue has to be paid for.
		4708
1774	=item Starting io/check/prepare/idle/signal/child watchers: O(1)	4709	=item Starting io/check/prepare/idle/signal/child/fork/async watchers: O(1)
1775		4710
		4711	These just add the watcher into an array or at the head of a list.
		4712
1776	=item Stopping check/prepare/idle watchers: O(1)	4713	=item Stopping check/prepare/idle/fork/async watchers: O(1)
1777		4714
1778	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % 16))	4715	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))
1779		4716
		4717	These watchers are stored in lists, so they need to be walked to find the
		4718	correct watcher to remove. The lists are usually short (you don't usually
		4719	have many watchers waiting for the same fd or signal: one is typical, two
		4720	is rare).
		4721
1780	=item Finding the next timer per loop iteration: O(1)	4722	=item Finding the next timer in each loop iteration: O(1)
		4723
		4724	By virtue of using a binary or 4-heap, the next timer is always found at a
		4725	fixed position in the storage array.
1781		4726
1782	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)	4727	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)
1783		4728
1784	=item Activating one watcher: O(1)	4729	A change means an I/O watcher gets started or stopped, which requires
		4730	libev to recalculate its status (and possibly tell the kernel, depending
		4731	on backend and whether C<ev_io_set> was used).
		4732
		4733	=item Activating one watcher (putting it into the pending state): O(1)
		4734
		4735	=item Priority handling: O(number_of_priorities)
		4736
		4737	Priorities are implemented by allocating some space for each
		4738	priority. When doing priority-based operations, libev usually has to
		4739	linearly search all the priorities, but starting/stopping and activating
		4740	watchers becomes O(1) with respect to priority handling.
		4741
		4742	=item Sending an ev_async: O(1)
		4743
		4744	=item Processing ev_async_send: O(number_of_async_watchers)
		4745
		4746	=item Processing signals: O(max_signal_number)
		4747
		4748	Sending involves a system call I<iff> there were no other C<ev_async_send>
		4749	calls in the current loop iteration. Checking for async and signal events
		4750	involves iterating over all running async watchers or all signal numbers.
1785		4751
1786	=back	4752	=back
1787		4753
1788		4754
		4755	=head1 PORTING FROM LIBEV 3.X TO 4.X
		4756
		4757	The major version 4 introduced some minor incompatible changes to the API.
		4758
		4759	At the moment, the C<ev.h> header file tries to implement superficial
		4760	compatibility, so most programs should still compile. Those might be
		4761	removed in later versions of libev, so better update early than late.
		4762
		4763	=over 4
		4764
		4765	=item function/symbol renames
		4766
		4767	A number of functions and symbols have been renamed:
		4768
		4769	ev_loop => ev_run
		4770	EVLOOP_NONBLOCK => EVRUN_NOWAIT
		4771	EVLOOP_ONESHOT => EVRUN_ONCE
		4772
		4773	ev_unloop => ev_break
		4774	EVUNLOOP_CANCEL => EVBREAK_CANCEL
		4775	EVUNLOOP_ONE => EVBREAK_ONE
		4776	EVUNLOOP_ALL => EVBREAK_ALL
		4777
		4778	EV_TIMEOUT => EV_TIMER
		4779
		4780	ev_loop_count => ev_iteration
		4781	ev_loop_depth => ev_depth
		4782	ev_loop_verify => ev_verify
		4783
		4784	Most functions working on C<struct ev_loop> objects don't have an
		4785	C<ev_loop_> prefix, so it was removed; C<ev_loop>, C<ev_unloop> and
		4786	associated constants have been renamed to not collide with the C<struct
		4787	ev_loop> anymore and C<EV_TIMER> now follows the same naming scheme
		4788	as all other watcher types. Note that C<ev_loop_fork> is still called
		4789	C<ev_loop_fork> because it would otherwise clash with the C<ev_fork>
		4790	typedef.
		4791
		4792	=item C<EV_COMPAT3> backwards compatibility mechanism
		4793
		4794	The backward compatibility mechanism can be controlled by
		4795	C<EV_COMPAT3>. See L<PREPROCESSOR SYMBOLS/MACROS> in the L<EMBEDDING>
		4796	section.
		4797
		4798	=item C<EV_MINIMAL> mechanism replaced by C<EV_FEATURES>
		4799
		4800	The preprocessor symbol C<EV_MINIMAL> has been replaced by a different
		4801	mechanism, C<EV_FEATURES>. Programs using C<EV_MINIMAL> usually compile
		4802	and work, but the library code will of course be larger.
		4803
		4804	=back
		4805
		4806
		4807	=head1 GLOSSARY
		4808
		4809	=over 4
		4810
		4811	=item active
		4812
		4813	A watcher is active as long as it has been started (has been attached to
		4814	an event loop) but not yet stopped (disassociated from the event loop).
		4815
		4816	=item application
		4817
		4818	In this document, an application is whatever is using libev.
		4819
		4820	=item callback
		4821
		4822	The address of a function that is called when some event has been
		4823	detected. Callbacks are being passed the event loop, the watcher that
		4824	received the event, and the actual event bitset.
		4825
		4826	=item callback invocation
		4827
		4828	The act of calling the callback associated with a watcher.
		4829
		4830	=item event
		4831
		4832	A change of state of some external event, such as data now being available
		4833	for reading on a file descriptor, time having passed or simply not having
		4834	any other events happening anymore.
		4835
		4836	In libev, events are represented as single bits (such as C<EV_READ> or
		4837	C<EV_TIMER>).
		4838
		4839	=item event library
		4840
		4841	A software package implementing an event model and loop.
		4842
		4843	=item event loop
		4844
		4845	An entity that handles and processes external events and converts them
		4846	into callback invocations.
		4847
		4848	=item event model
		4849
		4850	The model used to describe how an event loop handles and processes
		4851	watchers and events.
		4852
		4853	=item pending
		4854
		4855	A watcher is pending as soon as the corresponding event has been detected,
		4856	and stops being pending as soon as the watcher will be invoked or its
		4857	pending status is explicitly cleared by the application.
		4858
		4859	A watcher can be pending, but not active. Stopping a watcher also clears
		4860	its pending status.
		4861
		4862	=item real time
		4863
		4864	The physical time that is observed. It is apparently strictly monotonic :)
		4865
		4866	=item wall-clock time
		4867
		4868	The time and date as shown on clocks. Unlike real time, it can actually
		4869	be wrong and jump forwards and backwards, e.g. when the you adjust your
		4870	clock.
		4871
		4872	=item watcher
		4873
		4874	A data structure that describes interest in certain events. Watchers need
		4875	to be started (attached to an event loop) before they can receive events.
		4876
		4877	=item watcher invocation
		4878
		4879	The act of calling the callback associated with a watcher.
		4880
		4881	=back
		4882
1789	=head1 AUTHOR	4883	=head1 AUTHOR
1790		4884
1791	Marc Lehmann <libev@schmorp.de>.	4885	Marc Lehmann <libev@schmorp.de>, with repeated corrections by Mikael Magnusson.
1792		4886

Diff Legend

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