ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/libev/ev.pod

Comparing libev/ev.pod (file contents):
Revision 1.95 by root, Fri Dec 21 05:10:01 2007 UTC vs.
Revision 1.227 by root, Wed Mar 4 14:33:10 2009 UTC

…		…
2		2
3	libev - a high performance full-featured event loop written in C	3	libev - a high performance full-featured event loop written in C
4		4
5	=head1 SYNOPSIS	5	=head1 SYNOPSIS
6		6
7	#include <ev.h>	7	#include <ev.h>
8		8
9	=head1 EXAMPLE PROGRAM	9	=head2 EXAMPLE PROGRAM
10		10
		11	// a single header file is required
11	#include <ev.h>	12	#include <ev.h>
12		13
		14	#include <stdio.h> // for puts
		15
		16	// every watcher type has its own typedef'd struct
		17	// with the name ev_TYPE
13	ev_io stdin_watcher;	18	ev_io stdin_watcher;
14	ev_timer timeout_watcher;	19	ev_timer timeout_watcher;
15		20
16	/* called when data readable on stdin */	21	// all watcher callbacks have a similar signature
		22	// this callback is called when data is readable on stdin
17	static void	23	static void
18	stdin_cb (EV_P_ struct ev_io *w, int revents)	24	stdin_cb (EV_P_ ev_io *w, int revents)
19	{	25	{
20	/* puts ("stdin ready"); */	26	puts ("stdin ready");
21	ev_io_stop (EV_A_ w); /* just a syntax example */	27	// for one-shot events, one must manually stop the watcher
22	ev_unloop (EV_A_ EVUNLOOP_ALL); /* leave all loop calls */	28	// with its corresponding stop function.
		29	ev_io_stop (EV_A_ w);
		30
		31	// this causes all nested ev_loop's to stop iterating
		32	ev_unloop (EV_A_ EVUNLOOP_ALL);
23	}	33	}
24		34
		35	// another callback, this time for a time-out
25	static void	36	static void
26	timeout_cb (EV_P_ struct ev_timer *w, int revents)	37	timeout_cb (EV_P_ ev_timer *w, int revents)
27	{	38	{
28	/* puts ("timeout"); */	39	puts ("timeout");
29	ev_unloop (EV_A_ EVUNLOOP_ONE); /* leave one loop call */	40	// this causes the innermost ev_loop to stop iterating
		41	ev_unloop (EV_A_ EVUNLOOP_ONE);
30	}	42	}
31		43
32	int	44	int
33	main (void)	45	main (void)
34	{	46	{
		47	// use the default event loop unless you have special needs
35	struct ev_loop *loop = ev_default_loop (0);	48	struct ev_loop *loop = ev_default_loop (0);
36		49
37	/* initialise an io watcher, then start it */	50	// initialise an io watcher, then start it
		51	// this one will watch for stdin to become readable
38	ev_io_init (&stdin_watcher, stdin_cb, /*STDIN_FILENO*/ 0, EV_READ);	52	ev_io_init (&stdin_watcher, stdin_cb, /*STDIN_FILENO*/ 0, EV_READ);
39	ev_io_start (loop, &stdin_watcher);	53	ev_io_start (loop, &stdin_watcher);
40		54
		55	// initialise a timer watcher, then start it
41	/* simple non-repeating 5.5 second timeout */	56	// simple non-repeating 5.5 second timeout
42	ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);	57	ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);
43	ev_timer_start (loop, &timeout_watcher);	58	ev_timer_start (loop, &timeout_watcher);
44		59
45	/* loop till timeout or data ready */	60	// now wait for events to arrive
46	ev_loop (loop, 0);	61	ev_loop (loop, 0);
47		62
		63	// unloop was called, so exit
48	return 0;	64	return 0;
49	}	65	}
50		66
51	=head1 DESCRIPTION	67	=head1 DESCRIPTION
52		68
53	The newest version of this document is also available as a html-formatted	69	The newest version of this document is also available as an html-formatted
54	web page you might find easier to navigate when reading it for the first	70	web page you might find easier to navigate when reading it for the first
55	time: L<http://cvs.schmorp.de/libev/ev.html>.	71	time: L<http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod>.
56		72
57	Libev is an event loop: you register interest in certain events (such as a	73	Libev is an event loop: you register interest in certain events (such as a
58	file descriptor being readable or a timeout occurring), and it will manage	74	file descriptor being readable or a timeout occurring), and it will manage
59	these event sources and provide your program with events.	75	these event sources and provide your program with events.
60		76
…		…
65	You register interest in certain events by registering so-called I<event	81	You register interest in certain events by registering so-called I<event
66	watchers>, which are relatively small C structures you initialise with the	82	watchers>, which are relatively small C structures you initialise with the
67	details of the event, and then hand it over to libev by I<starting> the	83	details of the event, and then hand it over to libev by I<starting> the
68	watcher.	84	watcher.
69		85
70	=head1 FEATURES	86	=head2 FEATURES
71		87
72	Libev supports C<select>, C<poll>, the Linux-specific C<epoll>, the	88	Libev supports C<select>, C<poll>, the Linux-specific C<epoll>, the
73	BSD-specific C<kqueue> and the Solaris-specific event port mechanisms	89	BSD-specific C<kqueue> and the Solaris-specific event port mechanisms
74	for file descriptor events (C<ev_io>), the Linux C<inotify> interface	90	for file descriptor events (C<ev_io>), the Linux C<inotify> interface
75	(for C<ev_stat>), relative timers (C<ev_timer>), absolute timers	91	(for C<ev_stat>), relative timers (C<ev_timer>), absolute timers
…		…
82		98
83	It also is quite fast (see this	99	It also is quite fast (see this
84	L<benchmark|http://libev.schmorp.de/bench.html> comparing it to libevent	100	L<benchmark|http://libev.schmorp.de/bench.html> comparing it to libevent
85	for example).	101	for example).
86		102
87	=head1 CONVENTIONS	103	=head2 CONVENTIONS
88		104
89	Libev is very configurable. In this manual the default configuration will	105	Libev is very configurable. In this manual the default (and most common)
90	be described, which supports multiple event loops. For more info about	106	configuration will be described, which supports multiple event loops. For
91	various configuration options please have a look at B<EMBED> section in	107	more info about various configuration options please have a look at
92	this manual. If libev was configured without support for multiple event	108	B<EMBED> section in this manual. If libev was configured without support
93	loops, then all functions taking an initial argument of name C<loop>	109	for multiple event loops, then all functions taking an initial argument of
94	(which is always of type C<struct ev_loop *>) will not have this argument.	110	name C<loop> (which is always of type C<ev_loop *>) will not have
		111	this argument.
95		112
96	=head1 TIME REPRESENTATION	113	=head2 TIME REPRESENTATION
97		114
98	Libev represents time as a single floating point number, representing the	115	Libev represents time as a single floating point number, representing the
99	(fractional) number of seconds since the (POSIX) epoch (somewhere near	116	(fractional) number of seconds since the (POSIX) epoch (somewhere near
100	the beginning of 1970, details are complicated, don't ask). This type is	117	the beginning of 1970, details are complicated, don't ask). This type is
101	called C<ev_tstamp>, which is what you should use too. It usually aliases	118	called C<ev_tstamp>, which is what you should use too. It usually aliases
102	to the C<double> type in C, and when you need to do any calculations on	119	to the C<double> type in C, and when you need to do any calculations on
103	it, you should treat it as some floatingpoint value. Unlike the name	120	it, you should treat it as some floating point value. Unlike the name
104	component C<stamp> might indicate, it is also used for time differences	121	component C<stamp> might indicate, it is also used for time differences
105	throughout libev.	122	throughout libev.
		123
		124	=head1 ERROR HANDLING
		125
		126	Libev knows three classes of errors: operating system errors, usage errors
		127	and internal errors (bugs).
		128
		129	When libev catches an operating system error it cannot handle (for example
		130	a system call indicating a condition libev cannot fix), it calls the callback
		131	set via C<ev_set_syserr_cb>, which is supposed to fix the problem or
		132	abort. The default is to print a diagnostic message and to call C<abort
		133	()>.
		134
		135	When libev detects a usage error such as a negative timer interval, then
		136	it will print a diagnostic message and abort (via the C<assert> mechanism,
		137	so C<NDEBUG> will disable this checking): these are programming errors in
		138	the libev caller and need to be fixed there.
		139
		140	Libev also has a few internal error-checking C<assert>ions, and also has
		141	extensive consistency checking code. These do not trigger under normal
		142	circumstances, as they indicate either a bug in libev or worse.
		143
106		144
107	=head1 GLOBAL FUNCTIONS	145	=head1 GLOBAL FUNCTIONS
108		146
109	These functions can be called anytime, even before initialising the	147	These functions can be called anytime, even before initialising the
110	library in any way.	148	library in any way.
…		…
114	=item ev_tstamp ev_time ()	152	=item ev_tstamp ev_time ()
115		153
116	Returns the current time as libev would use it. Please note that the	154	Returns the current time as libev would use it. Please note that the
117	C<ev_now> function is usually faster and also often returns the timestamp	155	C<ev_now> function is usually faster and also often returns the timestamp
118	you actually want to know.	156	you actually want to know.
		157
		158	=item ev_sleep (ev_tstamp interval)
		159
		160	Sleep for the given interval: The current thread will be blocked until
		161	either it is interrupted or the given time interval has passed. Basically
		162	this is a sub-second-resolution C<sleep ()>.
119		163
120	=item int ev_version_major ()	164	=item int ev_version_major ()
121		165
122	=item int ev_version_minor ()	166	=item int ev_version_minor ()
123		167
…		…
136	not a problem.	180	not a problem.
137		181
138	Example: Make sure we haven't accidentally been linked against the wrong	182	Example: Make sure we haven't accidentally been linked against the wrong
139	version.	183	version.
140		184
141	assert (("libev version mismatch",	185	assert (("libev version mismatch",
142	ev_version_major () == EV_VERSION_MAJOR	186	ev_version_major () == EV_VERSION_MAJOR
143	&& ev_version_minor () >= EV_VERSION_MINOR));	187	&& ev_version_minor () >= EV_VERSION_MINOR));
144		188
145	=item unsigned int ev_supported_backends ()	189	=item unsigned int ev_supported_backends ()
146		190
147	Return the set of all backends (i.e. their corresponding C<EV_BACKEND_*>	191	Return the set of all backends (i.e. their corresponding C<EV_BACKEND_*>
148	value) compiled into this binary of libev (independent of their	192	value) compiled into this binary of libev (independent of their
…		…
150	a description of the set values.	194	a description of the set values.
151		195
152	Example: make sure we have the epoll method, because yeah this is cool and	196	Example: make sure we have the epoll method, because yeah this is cool and
153	a must have and can we have a torrent of it please!!!11	197	a must have and can we have a torrent of it please!!!11
154		198
155	assert (("sorry, no epoll, no sex",	199	assert (("sorry, no epoll, no sex",
156	ev_supported_backends () & EVBACKEND_EPOLL));	200	ev_supported_backends () & EVBACKEND_EPOLL));
157		201
158	=item unsigned int ev_recommended_backends ()	202	=item unsigned int ev_recommended_backends ()
159		203
160	Return the set of all backends compiled into this binary of libev and also	204	Return the set of all backends compiled into this binary of libev and also
161	recommended for this platform. This set is often smaller than the one	205	recommended for this platform. This set is often smaller than the one
162	returned by C<ev_supported_backends>, as for example kqueue is broken on	206	returned by C<ev_supported_backends>, as for example kqueue is broken on
163	most BSDs and will not be autodetected unless you explicitly request it	207	most BSDs and will not be auto-detected unless you explicitly request it
164	(assuming you know what you are doing). This is the set of backends that	208	(assuming you know what you are doing). This is the set of backends that
165	libev will probe for if you specify no backends explicitly.	209	libev will probe for if you specify no backends explicitly.
166		210
167	=item unsigned int ev_embeddable_backends ()	211	=item unsigned int ev_embeddable_backends ()
168		212
…		…
172	C<ev_embeddable_backends () & ev_supported_backends ()>, likewise for	216	C<ev_embeddable_backends () & ev_supported_backends ()>, likewise for
173	recommended ones.	217	recommended ones.
174		218
175	See the description of C<ev_embed> watchers for more info.	219	See the description of C<ev_embed> watchers for more info.
176		220
177	=item ev_set_allocator (void *(*cb)(void *ptr, long size))	221	=item ev_set_allocator (void *(*cb)(void *ptr, long size)) [NOT REENTRANT]
178		222
179	Sets the allocation function to use (the prototype is similar - the	223	Sets the allocation function to use (the prototype is similar - the
180	semantics is identical - to the realloc C function). It is used to	224	semantics are identical to the C<realloc> C89/SuS/POSIX function). It is
181	allocate and free memory (no surprises here). If it returns zero when	225	used to allocate and free memory (no surprises here). If it returns zero
182	memory needs to be allocated, the library might abort or take some	226	when memory needs to be allocated (C<size != 0>), the library might abort
183	potentially destructive action. The default is your system realloc	227	or take some potentially destructive action.
184	function.	228
		229	Since some systems (at least OpenBSD and Darwin) fail to implement
		230	correct C<realloc> semantics, libev will use a wrapper around the system
		231	C<realloc> and C<free> functions by default.
185		232
186	You could override this function in high-availability programs to, say,	233	You could override this function in high-availability programs to, say,
187	free some memory if it cannot allocate memory, to use a special allocator,	234	free some memory if it cannot allocate memory, to use a special allocator,
188	or even to sleep a while and retry until some memory is available.	235	or even to sleep a while and retry until some memory is available.
189		236
190	Example: Replace the libev allocator with one that waits a bit and then	237	Example: Replace the libev allocator with one that waits a bit and then
191	retries).	238	retries (example requires a standards-compliant C<realloc>).
192		239
193	static void *	240	static void *
194	persistent_realloc (void *ptr, size_t size)	241	persistent_realloc (void *ptr, size_t size)
195	{	242	{
196	for (;;)	243	for (;;)
…		…
205	}	252	}
206		253
207	...	254	...
208	ev_set_allocator (persistent_realloc);	255	ev_set_allocator (persistent_realloc);
209		256
210	=item ev_set_syserr_cb (void (*cb)(const char *msg));	257	=item ev_set_syserr_cb (void (*cb)(const char *msg)); [NOT REENTRANT]
211		258
212	Set the callback function to call on a retryable syscall error (such	259	Set the callback function to call on a retryable system call error (such
213	as failed select, poll, epoll_wait). The message is a printable string	260	as failed select, poll, epoll_wait). The message is a printable string
214	indicating the system call or subsystem causing the problem. If this	261	indicating the system call or subsystem causing the problem. If this
215	callback is set, then libev will expect it to remedy the sitution, no	262	callback is set, then libev will expect it to remedy the situation, no
216	matter what, when it returns. That is, libev will generally retry the	263	matter what, when it returns. That is, libev will generally retry the
217	requested operation, or, if the condition doesn't go away, do bad stuff	264	requested operation, or, if the condition doesn't go away, do bad stuff
218	(such as abort).	265	(such as abort).
219		266
220	Example: This is basically the same thing that libev does internally, too.	267	Example: This is basically the same thing that libev does internally, too.
…		…
231		278
232	=back	279	=back
233		280
234	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP	281	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP
235		282
236	An event loop is described by a C<struct ev_loop *>. The library knows two	283	An event loop is described by a C<struct ev_loop *> (the C<struct>
237	types of such loops, the I<default> loop, which supports signals and child	284	is I<not> optional in this case, as there is also an C<ev_loop>
238	events, and dynamically created loops which do not.	285	I<function>).
239		286
240	If you use threads, a common model is to run the default event loop	287	The library knows two types of such loops, the I<default> loop, which
241	in your main thread (or in a separate thread) and for each thread you	288	supports signals and child events, and dynamically created loops which do
242	create, you also create another event loop. Libev itself does no locking	289	not.
243	whatsoever, so if you mix calls to the same event loop in different
244	threads, make sure you lock (this is usually a bad idea, though, even if
245	done correctly, because it's hideous and inefficient).
246		290
247	=over 4	291	=over 4
248		292
249	=item struct ev_loop *ev_default_loop (unsigned int flags)	293	=item struct ev_loop *ev_default_loop (unsigned int flags)
250		294
…		…
254	flags. If that is troubling you, check C<ev_backend ()> afterwards).	298	flags. If that is troubling you, check C<ev_backend ()> afterwards).
255		299
256	If you don't know what event loop to use, use the one returned from this	300	If you don't know what event loop to use, use the one returned from this
257	function.	301	function.
258		302
		303	Note that this function is I<not> thread-safe, so if you want to use it
		304	from multiple threads, you have to lock (note also that this is unlikely,
		305	as loops cannot be shared easily between threads anyway).
		306
		307	The default loop is the only loop that can handle C<ev_signal> and
		308	C<ev_child> watchers, and to do this, it always registers a handler
		309	for C<SIGCHLD>. If this is a problem for your application you can either
		310	create a dynamic loop with C<ev_loop_new> that doesn't do that, or you
		311	can simply overwrite the C<SIGCHLD> signal handler I<after> calling
		312	C<ev_default_init>.
		313
259	The flags argument can be used to specify special behaviour or specific	314	The flags argument can be used to specify special behaviour or specific
260	backends to use, and is usually specified as C<0> (or C<EVFLAG_AUTO>).	315	backends to use, and is usually specified as C<0> (or C<EVFLAG_AUTO>).
261		316
262	The following flags are supported:	317	The following flags are supported:
263		318
…		…
268	The default flags value. Use this if you have no clue (it's the right	323	The default flags value. Use this if you have no clue (it's the right
269	thing, believe me).	324	thing, believe me).
270		325
271	=item C<EVFLAG_NOENV>	326	=item C<EVFLAG_NOENV>
272		327
273	If this flag bit is ored into the flag value (or the program runs setuid	328	If this flag bit is or'ed into the flag value (or the program runs setuid
274	or setgid) then libev will I<not> look at the environment variable	329	or setgid) then libev will I<not> look at the environment variable
275	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will	330	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will
276	override the flags completely if it is found in the environment. This is	331	override the flags completely if it is found in the environment. This is
277	useful to try out specific backends to test their performance, or to work	332	useful to try out specific backends to test their performance, or to work
278	around bugs.	333	around bugs.
…		…
284	enabling this flag.	339	enabling this flag.
285		340
286	This works by calling C<getpid ()> on every iteration of the loop,	341	This works by calling C<getpid ()> on every iteration of the loop,
287	and thus this might slow down your event loop if you do a lot of loop	342	and thus this might slow down your event loop if you do a lot of loop
288	iterations and little real work, but is usually not noticeable (on my	343	iterations and little real work, but is usually not noticeable (on my
289	Linux system for example, C<getpid> is actually a simple 5-insn sequence	344	GNU/Linux system for example, C<getpid> is actually a simple 5-insn sequence
290	without a syscall and thus I<very> fast, but my Linux system also has	345	without a system call and thus I<very> fast, but my GNU/Linux system also has
291	C<pthread_atfork> which is even faster).	346	C<pthread_atfork> which is even faster).
292		347
293	The big advantage of this flag is that you can forget about fork (and	348	The big advantage of this flag is that you can forget about fork (and
294	forget about forgetting to tell libev about forking) when you use this	349	forget about forgetting to tell libev about forking) when you use this
295	flag.	350	flag.
296		351
297	This flag setting cannot be overriden or specified in the C<LIBEV_FLAGS>	352	This flag setting cannot be overridden or specified in the C<LIBEV_FLAGS>
298	environment variable.	353	environment variable.
299		354
300	=item C<EVBACKEND_SELECT> (value 1, portable select backend)	355	=item C<EVBACKEND_SELECT> (value 1, portable select backend)
301		356
302	This is your standard select(2) backend. Not I<completely> standard, as	357	This is your standard select(2) backend. Not I<completely> standard, as
303	libev tries to roll its own fd_set with no limits on the number of fds,	358	libev tries to roll its own fd_set with no limits on the number of fds,
304	but if that fails, expect a fairly low limit on the number of fds when	359	but if that fails, expect a fairly low limit on the number of fds when
305	using this backend. It doesn't scale too well (O(highest_fd)), but its usually	360	using this backend. It doesn't scale too well (O(highest_fd)), but its
306	the fastest backend for a low number of fds.	361	usually the fastest backend for a low number of (low-numbered :) fds.
		362
		363	To get good performance out of this backend you need a high amount of
		364	parallelism (most of the file descriptors should be busy). If you are
		365	writing a server, you should C<accept ()> in a loop to accept as many
		366	connections as possible during one iteration. You might also want to have
		367	a look at C<ev_set_io_collect_interval ()> to increase the amount of
		368	readiness notifications you get per iteration.
		369
		370	This backend maps C<EV_READ> to the C<readfds> set and C<EV_WRITE> to the
		371	C<writefds> set (and to work around Microsoft Windows bugs, also onto the
		372	C<exceptfds> set on that platform).
307		373
308	=item C<EVBACKEND_POLL> (value 2, poll backend, available everywhere except on windows)	374	=item C<EVBACKEND_POLL> (value 2, poll backend, available everywhere except on windows)
309		375
310	And this is your standard poll(2) backend. It's more complicated than	376	And this is your standard poll(2) backend. It's more complicated
311	select, but handles sparse fds better and has no artificial limit on the	377	than select, but handles sparse fds better and has no artificial
312	number of fds you can use (except it will slow down considerably with a	378	limit on the number of fds you can use (except it will slow down
313	lot of inactive fds). It scales similarly to select, i.e. O(total_fds).	379	considerably with a lot of inactive fds). It scales similarly to select,
		380	i.e. O(total_fds). See the entry for C<EVBACKEND_SELECT>, above, for
		381	performance tips.
		382
		383	This backend maps C<EV_READ> to C<POLLIN | POLLERR | POLLHUP>, and
		384	C<EV_WRITE> to C<POLLOUT | POLLERR | POLLHUP>.
314		385
315	=item C<EVBACKEND_EPOLL> (value 4, Linux)	386	=item C<EVBACKEND_EPOLL> (value 4, Linux)
316		387
317	For few fds, this backend is a bit little slower than poll and select,	388	For few fds, this backend is a bit little slower than poll and select,
318	but it scales phenomenally better. While poll and select usually scale	389	but it scales phenomenally better. While poll and select usually scale
319	like O(total_fds) where n is the total number of fds (or the highest fd),	390	like O(total_fds) where n is the total number of fds (or the highest fd),
320	epoll scales either O(1) or O(active_fds). The epoll design has a number	391	epoll scales either O(1) or O(active_fds).
321	of shortcomings, such as silently dropping events in some hard-to-detect	392
322	cases and rewuiring a syscall per fd change, no fork support and bad	393	The epoll mechanism deserves honorable mention as the most misdesigned
323	support for dup:	394	of the more advanced event mechanisms: mere annoyances include silently
		395	dropping file descriptors, requiring a system call per change per file
		396	descriptor (and unnecessary guessing of parameters), problems with dup and
		397	so on. The biggest issue is fork races, however - if a program forks then
		398	I<both> parent and child process have to recreate the epoll set, which can
		399	take considerable time (one syscall per file descriptor) and is of course
		400	hard to detect.
		401
		402	Epoll is also notoriously buggy - embedding epoll fds I<should> work, but
		403	of course I<doesn't>, and epoll just loves to report events for totally
		404	I<different> file descriptors (even already closed ones, so one cannot
		405	even remove them from the set) than registered in the set (especially
		406	on SMP systems). Libev tries to counter these spurious notifications by
		407	employing an additional generation counter and comparing that against the
		408	events to filter out spurious ones, recreating the set when required.
324		409
325	While stopping, setting and starting an I/O watcher in the same iteration	410	While stopping, setting and starting an I/O watcher in the same iteration
326	will result in some caching, there is still a syscall per such incident	411	will result in some caching, there is still a system call per such
327	(because the fd could point to a different file description now), so its	412	incident (because the same I<file descriptor> could point to a different
328	best to avoid that. Also, C<dup ()>'ed file descriptors might not work	413	I<file description> now), so its best to avoid that. Also, C<dup ()>'ed
329	very well if you register events for both fds.	414	file descriptors might not work very well if you register events for both
		415	file descriptors.
330		416
331	Please note that epoll sometimes generates spurious notifications, so you	417	Best performance from this backend is achieved by not unregistering all
332	need to use non-blocking I/O or other means to avoid blocking when no data	418	watchers for a file descriptor until it has been closed, if possible,
333	(or space) is available.	419	i.e. keep at least one watcher active per fd at all times. Stopping and
		420	starting a watcher (without re-setting it) also usually doesn't cause
		421	extra overhead. A fork can both result in spurious notifications as well
		422	as in libev having to destroy and recreate the epoll object, which can
		423	take considerable time and thus should be avoided.
		424
		425	All this means that, in practice, C<EVBACKEND_SELECT> can be as fast or
		426	faster than epoll for maybe up to a hundred file descriptors, depending on
		427	the usage. So sad.
		428
		429	While nominally embeddable in other event loops, this feature is broken in
		430	all kernel versions tested so far.
		431
		432	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
		433	C<EVBACKEND_POLL>.
334		434
335	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)	435	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)
336		436
337	Kqueue deserves special mention, as at the time of this writing, it	437	Kqueue deserves special mention, as at the time of this writing, it
338	was broken on I<all> BSDs (usually it doesn't work with anything but	438	was broken on all BSDs except NetBSD (usually it doesn't work reliably
339	sockets and pipes, except on Darwin, where of course it's completely	439	with anything but sockets and pipes, except on Darwin, where of course
340	useless. On NetBSD, it seems to work for all the FD types I tested, so it	440	it's completely useless). Unlike epoll, however, whose brokenness
341	is used by default there). For this reason it's not being "autodetected"	441	is by design, these kqueue bugs can (and eventually will) be fixed
		442	without API changes to existing programs. For this reason it's not being
342	unless you explicitly specify it explicitly in the flags (i.e. using	443	"auto-detected" unless you explicitly specify it in the flags (i.e. using
343	C<EVBACKEND_KQUEUE>) or libev was compiled on a known-to-be-good (-enough)	444	C<EVBACKEND_KQUEUE>) or libev was compiled on a known-to-be-good (-enough)
344	system like NetBSD.	445	system like NetBSD.
345		446
		447	You still can embed kqueue into a normal poll or select backend and use it
		448	only for sockets (after having made sure that sockets work with kqueue on
		449	the target platform). See C<ev_embed> watchers for more info.
		450
346	It scales in the same way as the epoll backend, but the interface to the	451	It scales in the same way as the epoll backend, but the interface to the
347	kernel is more efficient (which says nothing about its actual speed,	452	kernel is more efficient (which says nothing about its actual speed, of
348	of course). While stopping, setting and starting an I/O watcher does	453	course). While stopping, setting and starting an I/O watcher does never
349	never cause an extra syscall as with epoll, it still adds up to two event	454	cause an extra system call as with C<EVBACKEND_EPOLL>, it still adds up to
350	changes per incident, support for C<fork ()> is very bad and it drops fds	455	two event changes per incident. Support for C<fork ()> is very bad (but
351	silently in similarly hard-to-detetc cases.	456	sane, unlike epoll) and it drops fds silently in similarly hard-to-detect
		457	cases
		458
		459	This backend usually performs well under most conditions.
		460
		461	While nominally embeddable in other event loops, this doesn't work
		462	everywhere, so you might need to test for this. And since it is broken
		463	almost everywhere, you should only use it when you have a lot of sockets
		464	(for which it usually works), by embedding it into another event loop
		465	(e.g. C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> (but C<poll> is of course
		466	also broken on OS X)) and, did I mention it, using it only for sockets.
		467
		468	This backend maps C<EV_READ> into an C<EVFILT_READ> kevent with
		469	C<NOTE_EOF>, and C<EV_WRITE> into an C<EVFILT_WRITE> kevent with
		470	C<NOTE_EOF>.
352		471
353	=item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8)	472	=item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8)
354		473
355	This is not implemented yet (and might never be).	474	This is not implemented yet (and might never be, unless you send me an
		475	implementation). According to reports, C</dev/poll> only supports sockets
		476	and is not embeddable, which would limit the usefulness of this backend
		477	immensely.
356		478
357	=item C<EVBACKEND_PORT> (value 32, Solaris 10)	479	=item C<EVBACKEND_PORT> (value 32, Solaris 10)
358		480
359	This uses the Solaris 10 event port mechanism. As with everything on Solaris,	481	This uses the Solaris 10 event port mechanism. As with everything on Solaris,
360	it's really slow, but it still scales very well (O(active_fds)).	482	it's really slow, but it still scales very well (O(active_fds)).
361		483
362	Please note that solaris event ports can deliver a lot of spurious	484	Please note that Solaris event ports can deliver a lot of spurious
363	notifications, so you need to use non-blocking I/O or other means to avoid	485	notifications, so you need to use non-blocking I/O or other means to avoid
364	blocking when no data (or space) is available.	486	blocking when no data (or space) is available.
		487
		488	While this backend scales well, it requires one system call per active
		489	file descriptor per loop iteration. For small and medium numbers of file
		490	descriptors a "slow" C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> backend
		491	might perform better.
		492
		493	On the positive side, with the exception of the spurious readiness
		494	notifications, this backend actually performed fully to specification
		495	in all tests and is fully embeddable, which is a rare feat among the
		496	OS-specific backends (I vastly prefer correctness over speed hacks).
		497
		498	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
		499	C<EVBACKEND_POLL>.
365		500
366	=item C<EVBACKEND_ALL>	501	=item C<EVBACKEND_ALL>
367		502
368	Try all backends (even potentially broken ones that wouldn't be tried	503	Try all backends (even potentially broken ones that wouldn't be tried
369	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as	504	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as
370	C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>.	505	C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>.
371		506
		507	It is definitely not recommended to use this flag.
		508
372	=back	509	=back
373		510
374	If one or more of these are ored into the flags value, then only these	511	If one or more of these are or'ed into the flags value, then only these
375	backends will be tried (in the reverse order as given here). If none are	512	backends will be tried (in the reverse order as listed here). If none are
376	specified, most compiled-in backend will be tried, usually in reverse	513	specified, all backends in C<ev_recommended_backends ()> will be tried.
377	order of their flag values :)
378		514
379	The most typical usage is like this:	515	Example: This is the most typical usage.
380		516
381	if (!ev_default_loop (0))	517	if (!ev_default_loop (0))
382	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");	518	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
383		519
384	Restrict libev to the select and poll backends, and do not allow	520	Example: Restrict libev to the select and poll backends, and do not allow
385	environment settings to be taken into account:	521	environment settings to be taken into account:
386		522
387	ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);	523	ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);
388		524
389	Use whatever libev has to offer, but make sure that kqueue is used if	525	Example: Use whatever libev has to offer, but make sure that kqueue is
390	available (warning, breaks stuff, best use only with your own private	526	used if available (warning, breaks stuff, best use only with your own
391	event loop and only if you know the OS supports your types of fds):	527	private event loop and only if you know the OS supports your types of
		528	fds):
392		529
393	ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);	530	ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);
394		531
395	=item struct ev_loop *ev_loop_new (unsigned int flags)	532	=item struct ev_loop *ev_loop_new (unsigned int flags)
396		533
397	Similar to C<ev_default_loop>, but always creates a new event loop that is	534	Similar to C<ev_default_loop>, but always creates a new event loop that is
398	always distinct from the default loop. Unlike the default loop, it cannot	535	always distinct from the default loop. Unlike the default loop, it cannot
399	handle signal and child watchers, and attempts to do so will be greeted by	536	handle signal and child watchers, and attempts to do so will be greeted by
400	undefined behaviour (or a failed assertion if assertions are enabled).	537	undefined behaviour (or a failed assertion if assertions are enabled).
401		538
		539	Note that this function I<is> thread-safe, and the recommended way to use
		540	libev with threads is indeed to create one loop per thread, and using the
		541	default loop in the "main" or "initial" thread.
		542
402	Example: Try to create a event loop that uses epoll and nothing else.	543	Example: Try to create a event loop that uses epoll and nothing else.
403		544
404	struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);	545	struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);
405	if (!epoller)	546	if (!epoller)
406	fatal ("no epoll found here, maybe it hides under your chair");	547	fatal ("no epoll found here, maybe it hides under your chair");
407		548
408	=item ev_default_destroy ()	549	=item ev_default_destroy ()
409		550
410	Destroys the default loop again (frees all memory and kernel state	551	Destroys the default loop again (frees all memory and kernel state
411	etc.). None of the active event watchers will be stopped in the normal	552	etc.). None of the active event watchers will be stopped in the normal
412	sense, so e.g. C<ev_is_active> might still return true. It is your	553	sense, so e.g. C<ev_is_active> might still return true. It is your
413	responsibility to either stop all watchers cleanly yoursef I<before>	554	responsibility to either stop all watchers cleanly yourself I<before>
414	calling this function, or cope with the fact afterwards (which is usually	555	calling this function, or cope with the fact afterwards (which is usually
415	the easiest thing, you can just ignore the watchers and/or C<free ()> them	556	the easiest thing, you can just ignore the watchers and/or C<free ()> them
416	for example).	557	for example).
417		558
418	Note that certain global state, such as signal state, will not be freed by	559	Note that certain global state, such as signal state (and installed signal
419	this function, and related watchers (such as signal and child watchers)	560	handlers), will not be freed by this function, and related watchers (such
420	would need to be stopped manually.	561	as signal and child watchers) would need to be stopped manually.
421		562
422	In general it is not advisable to call this function except in the	563	In general it is not advisable to call this function except in the
423	rare occasion where you really need to free e.g. the signal handling	564	rare occasion where you really need to free e.g. the signal handling
424	pipe fds. If you need dynamically allocated loops it is better to use	565	pipe fds. If you need dynamically allocated loops it is better to use
425	C<ev_loop_new> and C<ev_loop_destroy>).	566	C<ev_loop_new> and C<ev_loop_destroy>).
…		…
429	Like C<ev_default_destroy>, but destroys an event loop created by an	570	Like C<ev_default_destroy>, but destroys an event loop created by an
430	earlier call to C<ev_loop_new>.	571	earlier call to C<ev_loop_new>.
431		572
432	=item ev_default_fork ()	573	=item ev_default_fork ()
433		574
		575	This function sets a flag that causes subsequent C<ev_loop> iterations
434	This function reinitialises the kernel state for backends that have	576	to reinitialise the kernel state for backends that have one. Despite the
435	one. Despite the name, you can call it anytime, but it makes most sense	577	name, you can call it anytime, but it makes most sense after forking, in
436	after forking, in either the parent or child process (or both, but that	578	the child process (or both child and parent, but that again makes little
437	again makes little sense).	579	sense). You I<must> call it in the child before using any of the libev
		580	functions, and it will only take effect at the next C<ev_loop> iteration.
438		581
439	You I<must> call this function in the child process after forking if and	582	On the other hand, you only need to call this function in the child
440	only if you want to use the event library in both processes. If you just	583	process if and only if you want to use the event library in the child. If
441	fork+exec, you don't have to call it.	584	you just fork+exec, you don't have to call it at all.
442		585
443	The function itself is quite fast and it's usually not a problem to call	586	The function itself is quite fast and it's usually not a problem to call
444	it just in case after a fork. To make this easy, the function will fit in	587	it just in case after a fork. To make this easy, the function will fit in
445	quite nicely into a call to C<pthread_atfork>:	588	quite nicely into a call to C<pthread_atfork>:
446		589
447	pthread_atfork (0, 0, ev_default_fork);	590	pthread_atfork (0, 0, ev_default_fork);
448		591
449	At the moment, C<EVBACKEND_SELECT> and C<EVBACKEND_POLL> are safe to use
450	without calling this function, so if you force one of those backends you
451	do not need to care.
452
453	=item ev_loop_fork (loop)	592	=item ev_loop_fork (loop)
454		593
455	Like C<ev_default_fork>, but acts on an event loop created by	594	Like C<ev_default_fork>, but acts on an event loop created by
456	C<ev_loop_new>. Yes, you have to call this on every allocated event loop	595	C<ev_loop_new>. Yes, you have to call this on every allocated event loop
457	after fork, and how you do this is entirely your own problem.	596	after fork that you want to re-use in the child, and how you do this is
		597	entirely your own problem.
		598
		599	=item int ev_is_default_loop (loop)
		600
		601	Returns true when the given loop is, in fact, the default loop, and false
		602	otherwise.
458		603
459	=item unsigned int ev_loop_count (loop)	604	=item unsigned int ev_loop_count (loop)
460		605
461	Returns the count of loop iterations for the loop, which is identical to	606	Returns the count of loop iterations for the loop, which is identical to
462	the number of times libev did poll for new events. It starts at C<0> and	607	the number of times libev did poll for new events. It starts at C<0> and
…		…
477	received events and started processing them. This timestamp does not	622	received events and started processing them. This timestamp does not
478	change as long as callbacks are being processed, and this is also the base	623	change as long as callbacks are being processed, and this is also the base
479	time used for relative timers. You can treat it as the timestamp of the	624	time used for relative timers. You can treat it as the timestamp of the
480	event occurring (or more correctly, libev finding out about it).	625	event occurring (or more correctly, libev finding out about it).
481		626
		627	=item ev_now_update (loop)
		628
		629	Establishes the current time by querying the kernel, updating the time
		630	returned by C<ev_now ()> in the progress. This is a costly operation and
		631	is usually done automatically within C<ev_loop ()>.
		632
		633	This function is rarely useful, but when some event callback runs for a
		634	very long time without entering the event loop, updating libev's idea of
		635	the current time is a good idea.
		636
		637	See also "The special problem of time updates" in the C<ev_timer> section.
		638
482	=item ev_loop (loop, int flags)	639	=item ev_loop (loop, int flags)
483		640
484	Finally, this is it, the event handler. This function usually is called	641	Finally, this is it, the event handler. This function usually is called
485	after you initialised all your watchers and you want to start handling	642	after you initialised all your watchers and you want to start handling
486	events.	643	events.
…		…
488	If the flags argument is specified as C<0>, it will not return until	645	If the flags argument is specified as C<0>, it will not return until
489	either no event watchers are active anymore or C<ev_unloop> was called.	646	either no event watchers are active anymore or C<ev_unloop> was called.
490		647
491	Please note that an explicit C<ev_unloop> is usually better than	648	Please note that an explicit C<ev_unloop> is usually better than
492	relying on all watchers to be stopped when deciding when a program has	649	relying on all watchers to be stopped when deciding when a program has
493	finished (especially in interactive programs), but having a program that	650	finished (especially in interactive programs), but having a program
494	automatically loops as long as it has to and no longer by virtue of	651	that automatically loops as long as it has to and no longer by virtue
495	relying on its watchers stopping correctly is a thing of beauty.	652	of relying on its watchers stopping correctly, that is truly a thing of
		653	beauty.
496		654
497	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle	655	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle
498	those events and any outstanding ones, but will not block your process in	656	those events and any already outstanding ones, but will not block your
499	case there are no events and will return after one iteration of the loop.	657	process in case there are no events and will return after one iteration of
		658	the loop.
500		659
501	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if	660	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if
502	neccessary) and will handle those and any outstanding ones. It will block	661	necessary) and will handle those and any already outstanding ones. It
503	your process until at least one new event arrives, and will return after	662	will block your process until at least one new event arrives (which could
504	one iteration of the loop. This is useful if you are waiting for some	663	be an event internal to libev itself, so there is no guarantee that a
505	external event in conjunction with something not expressible using other	664	user-registered callback will be called), and will return after one
		665	iteration of the loop.
		666
		667	This is useful if you are waiting for some external event in conjunction
		668	with something not expressible using other libev watchers (i.e. "roll your
506	libev watchers. However, a pair of C<ev_prepare>/C<ev_check> watchers is	669	own C<ev_loop>"). However, a pair of C<ev_prepare>/C<ev_check> watchers is
507	usually a better approach for this kind of thing.	670	usually a better approach for this kind of thing.
508		671
509	Here are the gory details of what C<ev_loop> does:	672	Here are the gory details of what C<ev_loop> does:
510		673
511	- Before the first iteration, call any pending watchers.	674	- Before the first iteration, call any pending watchers.
512	* If there are no active watchers (reference count is zero), return.	675	* If EVFLAG_FORKCHECK was used, check for a fork.
513	- Queue all prepare watchers and then call all outstanding watchers.	676	- If a fork was detected (by any means), queue and call all fork watchers.
		677	- Queue and call all prepare watchers.
514	- If we have been forked, recreate the kernel state.	678	- If we have been forked, detach and recreate the kernel state
		679	as to not disturb the other process.
515	- Update the kernel state with all outstanding changes.	680	- Update the kernel state with all outstanding changes.
516	- Update the "event loop time".	681	- Update the "event loop time" (ev_now ()).
517	- Calculate for how long to block.	682	- Calculate for how long to sleep or block, if at all
		683	(active idle watchers, EVLOOP_NONBLOCK or not having
		684	any active watchers at all will result in not sleeping).
		685	- Sleep if the I/O and timer collect interval say so.
518	- Block the process, waiting for any events.	686	- Block the process, waiting for any events.
519	- Queue all outstanding I/O (fd) events.	687	- Queue all outstanding I/O (fd) events.
520	- Update the "event loop time" and do time jump handling.	688	- Update the "event loop time" (ev_now ()), and do time jump adjustments.
521	- Queue all outstanding timers.	689	- Queue all expired timers.
522	- Queue all outstanding periodics.	690	- Queue all expired periodics.
523	- If no events are pending now, queue all idle watchers.	691	- Unless any events are pending now, queue all idle watchers.
524	- Queue all check watchers.	692	- Queue all check watchers.
525	- Call all queued watchers in reverse order (i.e. check watchers first).	693	- Call all queued watchers in reverse order (i.e. check watchers first).
526	Signals and child watchers are implemented as I/O watchers, and will	694	Signals and child watchers are implemented as I/O watchers, and will
527	be handled here by queueing them when their watcher gets executed.	695	be handled here by queueing them when their watcher gets executed.
528	- If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK	696	- If ev_unloop has been called, or EVLOOP_ONESHOT or EVLOOP_NONBLOCK
529	were used, return, otherwise continue with step *.	697	were used, or there are no active watchers, return, otherwise
		698	continue with step *.
530		699
531	Example: Queue some jobs and then loop until no events are outsanding	700	Example: Queue some jobs and then loop until no events are outstanding
532	anymore.	701	anymore.
533		702
534	... queue jobs here, make sure they register event watchers as long	703	... queue jobs here, make sure they register event watchers as long
535	... as they still have work to do (even an idle watcher will do..)	704	... as they still have work to do (even an idle watcher will do..)
536	ev_loop (my_loop, 0);	705	ev_loop (my_loop, 0);
537	... jobs done. yeah!	706	... jobs done or somebody called unloop. yeah!
538		707
539	=item ev_unloop (loop, how)	708	=item ev_unloop (loop, how)
540		709
541	Can be used to make a call to C<ev_loop> return early (but only after it	710	Can be used to make a call to C<ev_loop> return early (but only after it
542	has processed all outstanding events). The C<how> argument must be either	711	has processed all outstanding events). The C<how> argument must be either
543	C<EVUNLOOP_ONE>, which will make the innermost C<ev_loop> call return, or	712	C<EVUNLOOP_ONE>, which will make the innermost C<ev_loop> call return, or
544	C<EVUNLOOP_ALL>, which will make all nested C<ev_loop> calls return.	713	C<EVUNLOOP_ALL>, which will make all nested C<ev_loop> calls return.
545		714
		715	This "unloop state" will be cleared when entering C<ev_loop> again.
		716
		717	It is safe to call C<ev_unloop> from otuside any C<ev_loop> calls.
		718
546	=item ev_ref (loop)	719	=item ev_ref (loop)
547		720
548	=item ev_unref (loop)	721	=item ev_unref (loop)
549		722
550	Ref/unref can be used to add or remove a reference count on the event	723	Ref/unref can be used to add or remove a reference count on the event
551	loop: Every watcher keeps one reference, and as long as the reference	724	loop: Every watcher keeps one reference, and as long as the reference
552	count is nonzero, C<ev_loop> will not return on its own. If you have	725	count is nonzero, C<ev_loop> will not return on its own.
		726
553	a watcher you never unregister that should not keep C<ev_loop> from	727	If you have a watcher you never unregister that should not keep C<ev_loop>
554	returning, ev_unref() after starting, and ev_ref() before stopping it. For	728	from returning, call ev_unref() after starting, and ev_ref() before
		729	stopping it.
		730
555	example, libev itself uses this for its internal signal pipe: It is not	731	As an example, libev itself uses this for its internal signal pipe: It is
556	visible to the libev user and should not keep C<ev_loop> from exiting if	732	not visible to the libev user and should not keep C<ev_loop> from exiting
557	no event watchers registered by it are active. It is also an excellent	733	if no event watchers registered by it are active. It is also an excellent
558	way to do this for generic recurring timers or from within third-party	734	way to do this for generic recurring timers or from within third-party
559	libraries. Just remember to I<unref after start> and I<ref before stop>.	735	libraries. Just remember to I<unref after start> and I<ref before stop>
		736	(but only if the watcher wasn't active before, or was active before,
		737	respectively).
560		738
561	Example: Create a signal watcher, but keep it from keeping C<ev_loop>	739	Example: Create a signal watcher, but keep it from keeping C<ev_loop>
562	running when nothing else is active.	740	running when nothing else is active.
563		741
564	struct ev_signal exitsig;	742	ev_signal exitsig;
565	ev_signal_init (&exitsig, sig_cb, SIGINT);	743	ev_signal_init (&exitsig, sig_cb, SIGINT);
566	ev_signal_start (loop, &exitsig);	744	ev_signal_start (loop, &exitsig);
567	evf_unref (loop);	745	evf_unref (loop);
568		746
569	Example: For some weird reason, unregister the above signal handler again.	747	Example: For some weird reason, unregister the above signal handler again.
570		748
571	ev_ref (loop);	749	ev_ref (loop);
572	ev_signal_stop (loop, &exitsig);	750	ev_signal_stop (loop, &exitsig);
		751
		752	=item ev_set_io_collect_interval (loop, ev_tstamp interval)
		753
		754	=item ev_set_timeout_collect_interval (loop, ev_tstamp interval)
		755
		756	These advanced functions influence the time that libev will spend waiting
		757	for events. Both time intervals are by default C<0>, meaning that libev
		758	will try to invoke timer/periodic callbacks and I/O callbacks with minimum
		759	latency.
		760
		761	Setting these to a higher value (the C<interval> I<must> be >= C<0>)
		762	allows libev to delay invocation of I/O and timer/periodic callbacks
		763	to increase efficiency of loop iterations (or to increase power-saving
		764	opportunities).
		765
		766	The idea is that sometimes your program runs just fast enough to handle
		767	one (or very few) event(s) per loop iteration. While this makes the
		768	program responsive, it also wastes a lot of CPU time to poll for new
		769	events, especially with backends like C<select ()> which have a high
		770	overhead for the actual polling but can deliver many events at once.
		771
		772	By setting a higher I<io collect interval> you allow libev to spend more
		773	time collecting I/O events, so you can handle more events per iteration,
		774	at the cost of increasing latency. Timeouts (both C<ev_periodic> and
		775	C<ev_timer>) will be not affected. Setting this to a non-null value will
		776	introduce an additional C<ev_sleep ()> call into most loop iterations.
		777
		778	Likewise, by setting a higher I<timeout collect interval> you allow libev
		779	to spend more time collecting timeouts, at the expense of increased
		780	latency/jitter/inexactness (the watcher callback will be called
		781	later). C<ev_io> watchers will not be affected. Setting this to a non-null
		782	value will not introduce any overhead in libev.
		783
		784	Many (busy) programs can usually benefit by setting the I/O collect
		785	interval to a value near C<0.1> or so, which is often enough for
		786	interactive servers (of course not for games), likewise for timeouts. It
		787	usually doesn't make much sense to set it to a lower value than C<0.01>,
		788	as this approaches the timing granularity of most systems.
		789
		790	Setting the I<timeout collect interval> can improve the opportunity for
		791	saving power, as the program will "bundle" timer callback invocations that
		792	are "near" in time together, by delaying some, thus reducing the number of
		793	times the process sleeps and wakes up again. Another useful technique to
		794	reduce iterations/wake-ups is to use C<ev_periodic> watchers and make sure
		795	they fire on, say, one-second boundaries only.
		796
		797	=item ev_loop_verify (loop)
		798
		799	This function only does something when C<EV_VERIFY> support has been
		800	compiled in, which is the default for non-minimal builds. It tries to go
		801	through all internal structures and checks them for validity. If anything
		802	is found to be inconsistent, it will print an error message to standard
		803	error and call C<abort ()>.
		804
		805	This can be used to catch bugs inside libev itself: under normal
		806	circumstances, this function will never abort as of course libev keeps its
		807	data structures consistent.
573		808
574	=back	809	=back
575		810
576		811
577	=head1 ANATOMY OF A WATCHER	812	=head1 ANATOMY OF A WATCHER
		813
		814	In the following description, uppercase C<TYPE> in names stands for the
		815	watcher type, e.g. C<ev_TYPE_start> can mean C<ev_timer_start> for timer
		816	watchers and C<ev_io_start> for I/O watchers.
578		817
579	A watcher is a structure that you create and register to record your	818	A watcher is a structure that you create and register to record your
580	interest in some event. For instance, if you want to wait for STDIN to	819	interest in some event. For instance, if you want to wait for STDIN to
581	become readable, you would create an C<ev_io> watcher for that:	820	become readable, you would create an C<ev_io> watcher for that:
582		821
583	static void my_cb (struct ev_loop *loop, struct ev_io *w, int revents)	822	static void my_cb (struct ev_loop *loop, ev_io *w, int revents)
584	{	823	{
585	ev_io_stop (w);	824	ev_io_stop (w);
586	ev_unloop (loop, EVUNLOOP_ALL);	825	ev_unloop (loop, EVUNLOOP_ALL);
587	}	826	}
588		827
589	struct ev_loop *loop = ev_default_loop (0);	828	struct ev_loop *loop = ev_default_loop (0);
		829
590	struct ev_io stdin_watcher;	830	ev_io stdin_watcher;
		831
591	ev_init (&stdin_watcher, my_cb);	832	ev_init (&stdin_watcher, my_cb);
592	ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);	833	ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);
593	ev_io_start (loop, &stdin_watcher);	834	ev_io_start (loop, &stdin_watcher);
		835
594	ev_loop (loop, 0);	836	ev_loop (loop, 0);
595		837
596	As you can see, you are responsible for allocating the memory for your	838	As you can see, you are responsible for allocating the memory for your
597	watcher structures (and it is usually a bad idea to do this on the stack,	839	watcher structures (and it is I<usually> a bad idea to do this on the
598	although this can sometimes be quite valid).	840	stack).
		841
		842	Each watcher has an associated watcher structure (called C<struct ev_TYPE>
		843	or simply C<ev_TYPE>, as typedefs are provided for all watcher structs).
599		844
600	Each watcher structure must be initialised by a call to C<ev_init	845	Each watcher structure must be initialised by a call to C<ev_init
601	(watcher *, callback)>, which expects a callback to be provided. This	846	(watcher *, callback)>, which expects a callback to be provided. This
602	callback gets invoked each time the event occurs (or, in the case of io	847	callback gets invoked each time the event occurs (or, in the case of I/O
603	watchers, each time the event loop detects that the file descriptor given	848	watchers, each time the event loop detects that the file descriptor given
604	is readable and/or writable).	849	is readable and/or writable).
605		850
606	Each watcher type has its own C<< ev_<type>_set (watcher *, ...) >> macro	851	Each watcher type further has its own C<< ev_TYPE_set (watcher *, ...) >>
607	with arguments specific to this watcher type. There is also a macro	852	macro to configure it, with arguments specific to the watcher type. There
608	to combine initialisation and setting in one call: C<< ev_<type>_init	853	is also a macro to combine initialisation and setting in one call: C<<
609	(watcher *, callback, ...) >>.	854	ev_TYPE_init (watcher *, callback, ...) >>.
610		855
611	To make the watcher actually watch out for events, you have to start it	856	To make the watcher actually watch out for events, you have to start it
612	with a watcher-specific start function (C<< ev_<type>_start (loop, watcher	857	with a watcher-specific start function (C<< ev_TYPE_start (loop, watcher
613	*) >>), and you can stop watching for events at any time by calling the	858	*) >>), and you can stop watching for events at any time by calling the
614	corresponding stop function (C<< ev_<type>_stop (loop, watcher *) >>.	859	corresponding stop function (C<< ev_TYPE_stop (loop, watcher *) >>.
615		860
616	As long as your watcher is active (has been started but not stopped) you	861	As long as your watcher is active (has been started but not stopped) you
617	must not touch the values stored in it. Most specifically you must never	862	must not touch the values stored in it. Most specifically you must never
618	reinitialise it or call its C<set> macro.	863	reinitialise it or call its C<ev_TYPE_set> macro.
619		864
620	Each and every callback receives the event loop pointer as first, the	865	Each and every callback receives the event loop pointer as first, the
621	registered watcher structure as second, and a bitset of received events as	866	registered watcher structure as second, and a bitset of received events as
622	third argument.	867	third argument.
623		868
…		…
677	=item C<EV_FORK>	922	=item C<EV_FORK>
678		923
679	The event loop has been resumed in the child process after fork (see	924	The event loop has been resumed in the child process after fork (see
680	C<ev_fork>).	925	C<ev_fork>).
681		926
		927	=item C<EV_ASYNC>
		928
		929	The given async watcher has been asynchronously notified (see C<ev_async>).
		930
682	=item C<EV_ERROR>	931	=item C<EV_ERROR>
683		932
684	An unspecified error has occured, the watcher has been stopped. This might	933	An unspecified error has occurred, the watcher has been stopped. This might
685	happen because the watcher could not be properly started because libev	934	happen because the watcher could not be properly started because libev
686	ran out of memory, a file descriptor was found to be closed or any other	935	ran out of memory, a file descriptor was found to be closed or any other
		936	problem. Libev considers these application bugs.
		937
687	problem. You best act on it by reporting the problem and somehow coping	938	You best act on it by reporting the problem and somehow coping with the
688	with the watcher being stopped.	939	watcher being stopped. Note that well-written programs should not receive
		940	an error ever, so when your watcher receives it, this usually indicates a
		941	bug in your program.
689		942
690	Libev will usually signal a few "dummy" events together with an error,	943	Libev will usually signal a few "dummy" events together with an error, for
691	for example it might indicate that a fd is readable or writable, and if	944	example it might indicate that a fd is readable or writable, and if your
692	your callbacks is well-written it can just attempt the operation and cope	945	callbacks is well-written it can just attempt the operation and cope with
693	with the error from read() or write(). This will not work in multithreaded	946	the error from read() or write(). This will not work in multi-threaded
694	programs, though, so beware.	947	programs, though, as the fd could already be closed and reused for another
		948	thing, so beware.
695		949
696	=back	950	=back
697		951
698	=head2 GENERIC WATCHER FUNCTIONS	952	=head2 GENERIC WATCHER FUNCTIONS
699
700	In the following description, C<TYPE> stands for the watcher type,
701	e.g. C<timer> for C<ev_timer> watchers and C<io> for C<ev_io> watchers.
702		953
703	=over 4	954	=over 4
704		955
705	=item C<ev_init> (ev_TYPE *watcher, callback)	956	=item C<ev_init> (ev_TYPE *watcher, callback)
706		957
…		…
712	which rolls both calls into one.	963	which rolls both calls into one.
713		964
714	You can reinitialise a watcher at any time as long as it has been stopped	965	You can reinitialise a watcher at any time as long as it has been stopped
715	(or never started) and there are no pending events outstanding.	966	(or never started) and there are no pending events outstanding.
716		967
717	The callback is always of type C<void (*)(ev_loop *loop, ev_TYPE *watcher,	968	The callback is always of type C<void (*)(struct ev_loop *loop, ev_TYPE *watcher,
718	int revents)>.	969	int revents)>.
		970
		971	Example: Initialise an C<ev_io> watcher in two steps.
		972
		973	ev_io w;
		974	ev_init (&w, my_cb);
		975	ev_io_set (&w, STDIN_FILENO, EV_READ);
719		976
720	=item C<ev_TYPE_set> (ev_TYPE *, [args])	977	=item C<ev_TYPE_set> (ev_TYPE *, [args])
721		978
722	This macro initialises the type-specific parts of a watcher. You need to	979	This macro initialises the type-specific parts of a watcher. You need to
723	call C<ev_init> at least once before you call this macro, but you can	980	call C<ev_init> at least once before you call this macro, but you can
…		…
726	difference to the C<ev_init> macro).	983	difference to the C<ev_init> macro).
727		984
728	Although some watcher types do not have type-specific arguments	985	Although some watcher types do not have type-specific arguments
729	(e.g. C<ev_prepare>) you still need to call its C<set> macro.	986	(e.g. C<ev_prepare>) you still need to call its C<set> macro.
730		987
		988	See C<ev_init>, above, for an example.
		989
731	=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])	990	=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])
732		991
733	This convinience macro rolls both C<ev_init> and C<ev_TYPE_set> macro	992	This convenience macro rolls both C<ev_init> and C<ev_TYPE_set> macro
734	calls into a single call. This is the most convinient method to initialise	993	calls into a single call. This is the most convenient method to initialise
735	a watcher. The same limitations apply, of course.	994	a watcher. The same limitations apply, of course.
		995
		996	Example: Initialise and set an C<ev_io> watcher in one step.
		997
		998	ev_io_init (&w, my_cb, STDIN_FILENO, EV_READ);
736		999
737	=item C<ev_TYPE_start> (loop *, ev_TYPE *watcher)	1000	=item C<ev_TYPE_start> (loop *, ev_TYPE *watcher)
738		1001
739	Starts (activates) the given watcher. Only active watchers will receive	1002	Starts (activates) the given watcher. Only active watchers will receive
740	events. If the watcher is already active nothing will happen.	1003	events. If the watcher is already active nothing will happen.
741		1004
		1005	Example: Start the C<ev_io> watcher that is being abused as example in this
		1006	whole section.
		1007
		1008	ev_io_start (EV_DEFAULT_UC, &w);
		1009
742	=item C<ev_TYPE_stop> (loop *, ev_TYPE *watcher)	1010	=item C<ev_TYPE_stop> (loop *, ev_TYPE *watcher)
743		1011
744	Stops the given watcher again (if active) and clears the pending	1012	Stops the given watcher if active, and clears the pending status (whether
		1013	the watcher was active or not).
		1014
745	status. It is possible that stopped watchers are pending (for example,	1015	It is possible that stopped watchers are pending - for example,
746	non-repeating timers are being stopped when they become pending), but	1016	non-repeating timers are being stopped when they become pending - but
747	C<ev_TYPE_stop> ensures that the watcher is neither active nor pending. If	1017	calling C<ev_TYPE_stop> ensures that the watcher is neither active nor
748	you want to free or reuse the memory used by the watcher it is therefore a	1018	pending. If you want to free or reuse the memory used by the watcher it is
749	good idea to always call its C<ev_TYPE_stop> function.	1019	therefore a good idea to always call its C<ev_TYPE_stop> function.
750		1020
751	=item bool ev_is_active (ev_TYPE *watcher)	1021	=item bool ev_is_active (ev_TYPE *watcher)
752		1022
753	Returns a true value iff the watcher is active (i.e. it has been started	1023	Returns a true value iff the watcher is active (i.e. it has been started
754	and not yet been stopped). As long as a watcher is active you must not modify	1024	and not yet been stopped). As long as a watcher is active you must not modify
…		…
796	The default priority used by watchers when no priority has been set is	1066	The default priority used by watchers when no priority has been set is
797	always C<0>, which is supposed to not be too high and not be too low :).	1067	always C<0>, which is supposed to not be too high and not be too low :).
798		1068
799	Setting a priority outside the range of C<EV_MINPRI> to C<EV_MAXPRI> is	1069	Setting a priority outside the range of C<EV_MINPRI> to C<EV_MAXPRI> is
800	fine, as long as you do not mind that the priority value you query might	1070	fine, as long as you do not mind that the priority value you query might
801	or might not have been adjusted to be within valid range.	1071	or might not have been clamped to the valid range.
802		1072
803	=item ev_invoke (loop, ev_TYPE *watcher, int revents)	1073	=item ev_invoke (loop, ev_TYPE *watcher, int revents)
804		1074
805	Invoke the C<watcher> with the given C<loop> and C<revents>. Neither	1075	Invoke the C<watcher> with the given C<loop> and C<revents>. Neither
806	C<loop> nor C<revents> need to be valid as long as the watcher callback	1076	C<loop> nor C<revents> need to be valid as long as the watcher callback
807	can deal with that fact.	1077	can deal with that fact, as both are simply passed through to the
		1078	callback.
808		1079
809	=item int ev_clear_pending (loop, ev_TYPE *watcher)	1080	=item int ev_clear_pending (loop, ev_TYPE *watcher)
810		1081
811	If the watcher is pending, this function returns clears its pending status	1082	If the watcher is pending, this function clears its pending status and
812	and returns its C<revents> bitset (as if its callback was invoked). If the	1083	returns its C<revents> bitset (as if its callback was invoked). If the
813	watcher isn't pending it does nothing and returns C<0>.	1084	watcher isn't pending it does nothing and returns C<0>.
814		1085
		1086	Sometimes it can be useful to "poll" a watcher instead of waiting for its
		1087	callback to be invoked, which can be accomplished with this function.
		1088
815	=back	1089	=back
816		1090
817		1091
818	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER	1092	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER
819		1093
820	Each watcher has, by default, a member C<void *data> that you can change	1094	Each watcher has, by default, a member C<void *data> that you can change
821	and read at any time, libev will completely ignore it. This can be used	1095	and read at any time: libev will completely ignore it. This can be used
822	to associate arbitrary data with your watcher. If you need more data and	1096	to associate arbitrary data with your watcher. If you need more data and
823	don't want to allocate memory and store a pointer to it in that data	1097	don't want to allocate memory and store a pointer to it in that data
824	member, you can also "subclass" the watcher type and provide your own	1098	member, you can also "subclass" the watcher type and provide your own
825	data:	1099	data:
826		1100
827	struct my_io	1101	struct my_io
828	{	1102	{
829	struct ev_io io;	1103	ev_io io;
830	int otherfd;	1104	int otherfd;
831	void *somedata;	1105	void *somedata;
832	struct whatever *mostinteresting;	1106	struct whatever *mostinteresting;
833	}	1107	};
		1108
		1109	...
		1110	struct my_io w;
		1111	ev_io_init (&w.io, my_cb, fd, EV_READ);
834		1112
835	And since your callback will be called with a pointer to the watcher, you	1113	And since your callback will be called with a pointer to the watcher, you
836	can cast it back to your own type:	1114	can cast it back to your own type:
837		1115
838	static void my_cb (struct ev_loop *loop, struct ev_io *w_, int revents)	1116	static void my_cb (struct ev_loop *loop, ev_io *w_, int revents)
839	{	1117	{
840	struct my_io *w = (struct my_io *)w_;	1118	struct my_io *w = (struct my_io *)w_;
841	...	1119	...
842	}	1120	}
843		1121
844	More interesting and less C-conformant ways of casting your callback type	1122	More interesting and less C-conformant ways of casting your callback type
845	instead have been omitted.	1123	instead have been omitted.
846		1124
847	Another common scenario is having some data structure with multiple	1125	Another common scenario is to use some data structure with multiple
848	watchers:	1126	embedded watchers:
849		1127
850	struct my_biggy	1128	struct my_biggy
851	{	1129	{
852	int some_data;	1130	int some_data;
853	ev_timer t1;	1131	ev_timer t1;
854	ev_timer t2;	1132	ev_timer t2;
855	}	1133	}
856		1134
857	In this case getting the pointer to C<my_biggy> is a bit more complicated,	1135	In this case getting the pointer to C<my_biggy> is a bit more
858	you need to use C<offsetof>:	1136	complicated: Either you store the address of your C<my_biggy> struct
		1137	in the C<data> member of the watcher (for woozies), or you need to use
		1138	some pointer arithmetic using C<offsetof> inside your watchers (for real
		1139	programmers):
859		1140
860	#include <stddef.h>	1141	#include <stddef.h>
861		1142
862	static void	1143	static void
863	t1_cb (EV_P_ struct ev_timer *w, int revents)	1144	t1_cb (EV_P_ ev_timer *w, int revents)
864	{	1145	{
865	struct my_biggy big = (struct my_biggy *	1146	struct my_biggy big = (struct my_biggy *
866	(((char *)w) - offsetof (struct my_biggy, t1));	1147	(((char *)w) - offsetof (struct my_biggy, t1));
867	}	1148	}
868		1149
869	static void	1150	static void
870	t2_cb (EV_P_ struct ev_timer *w, int revents)	1151	t2_cb (EV_P_ ev_timer *w, int revents)
871	{	1152	{
872	struct my_biggy big = (struct my_biggy *	1153	struct my_biggy big = (struct my_biggy *
873	(((char *)w) - offsetof (struct my_biggy, t2));	1154	(((char *)w) - offsetof (struct my_biggy, t2));
874	}	1155	}
875		1156
876		1157
877	=head1 WATCHER TYPES	1158	=head1 WATCHER TYPES
878		1159
879	This section describes each watcher in detail, but will not repeat	1160	This section describes each watcher in detail, but will not repeat
…		…
903	In general you can register as many read and/or write event watchers per	1184	In general you can register as many read and/or write event watchers per
904	fd as you want (as long as you don't confuse yourself). Setting all file	1185	fd as you want (as long as you don't confuse yourself). Setting all file
905	descriptors to non-blocking mode is also usually a good idea (but not	1186	descriptors to non-blocking mode is also usually a good idea (but not
906	required if you know what you are doing).	1187	required if you know what you are doing).
907		1188
908	You have to be careful with dup'ed file descriptors, though. Some backends	1189	If you cannot use non-blocking mode, then force the use of a
909	(the linux epoll backend is a notable example) cannot handle dup'ed file	1190	known-to-be-good backend (at the time of this writing, this includes only
910	descriptors correctly if you register interest in two or more fds pointing	1191	C<EVBACKEND_SELECT> and C<EVBACKEND_POLL>).
911	to the same underlying file/socket/etc. description (that is, they share
912	the same underlying "file open").
913
914	If you must do this, then force the use of a known-to-be-good backend
915	(at the time of this writing, this includes only C<EVBACKEND_SELECT> and
916	C<EVBACKEND_POLL>).
917		1192
918	Another thing you have to watch out for is that it is quite easy to	1193	Another thing you have to watch out for is that it is quite easy to
919	receive "spurious" readyness notifications, that is your callback might	1194	receive "spurious" readiness notifications, that is your callback might
920	be called with C<EV_READ> but a subsequent C<read>(2) will actually block	1195	be called with C<EV_READ> but a subsequent C<read>(2) will actually block
921	because there is no data. Not only are some backends known to create a	1196	because there is no data. Not only are some backends known to create a
922	lot of those (for example solaris ports), it is very easy to get into	1197	lot of those (for example Solaris ports), it is very easy to get into
923	this situation even with a relatively standard program structure. Thus	1198	this situation even with a relatively standard program structure. Thus
924	it is best to always use non-blocking I/O: An extra C<read>(2) returning	1199	it is best to always use non-blocking I/O: An extra C<read>(2) returning
925	C<EAGAIN> is far preferable to a program hanging until some data arrives.	1200	C<EAGAIN> is far preferable to a program hanging until some data arrives.
926		1201
927	If you cannot run the fd in non-blocking mode (for example you should not	1202	If you cannot run the fd in non-blocking mode (for example you should
928	play around with an Xlib connection), then you have to seperately re-test	1203	not play around with an Xlib connection), then you have to separately
929	whether a file descriptor is really ready with a known-to-be good interface	1204	re-test whether a file descriptor is really ready with a known-to-be good
930	such as poll (fortunately in our Xlib example, Xlib already does this on	1205	interface such as poll (fortunately in our Xlib example, Xlib already
931	its own, so its quite safe to use).	1206	does this on its own, so its quite safe to use). Some people additionally
		1207	use C<SIGALRM> and an interval timer, just to be sure you won't block
		1208	indefinitely.
		1209
		1210	But really, best use non-blocking mode.
932		1211
933	=head3 The special problem of disappearing file descriptors	1212	=head3 The special problem of disappearing file descriptors
934		1213
935	Some backends (e.g. kqueue, epoll) need to be told about closing a file	1214	Some backends (e.g. kqueue, epoll) need to be told about closing a file
936	descriptor (either by calling C<close> explicitly or by any other means,	1215	descriptor (either due to calling C<close> explicitly or any other means,
937	such as C<dup>). The reason is that you register interest in some file	1216	such as C<dup2>). The reason is that you register interest in some file
938	descriptor, but when it goes away, the operating system will silently drop	1217	descriptor, but when it goes away, the operating system will silently drop
939	this interest. If another file descriptor with the same number then is	1218	this interest. If another file descriptor with the same number then is
940	registered with libev, there is no efficient way to see that this is, in	1219	registered with libev, there is no efficient way to see that this is, in
941	fact, a different file descriptor.	1220	fact, a different file descriptor.
942		1221
…		…
952	optimisations to libev.	1231	optimisations to libev.
953		1232
954	=head3 The special problem of dup'ed file descriptors	1233	=head3 The special problem of dup'ed file descriptors
955		1234
956	Some backends (e.g. epoll), cannot register events for file descriptors,	1235	Some backends (e.g. epoll), cannot register events for file descriptors,
957	but only events for the underlying file descriptions. That menas when you	1236	but only events for the underlying file descriptions. That means when you
958	have C<dup ()>'ed file descriptors and register events for them, only one	1237	have C<dup ()>'ed file descriptors or weirder constellations, and register
959	file descriptor might actually receive events.	1238	events for them, only one file descriptor might actually receive events.
960		1239
961	There is no workaorund possible except not registering events	1240	There is no workaround possible except not registering events
962	for potentially C<dup ()>'ed file descriptors or to resort to	1241	for potentially C<dup ()>'ed file descriptors, or to resort to
963	C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>.	1242	C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>.
964		1243
965	=head3 The special problem of fork	1244	=head3 The special problem of fork
966		1245
967	Some backends (epoll, kqueue) do not support C<fork ()> at all or exhibit	1246	Some backends (epoll, kqueue) do not support C<fork ()> at all or exhibit
…		…
971	To support fork in your programs, you either have to call	1250	To support fork in your programs, you either have to call
972	C<ev_default_fork ()> or C<ev_loop_fork ()> after a fork in the child,	1251	C<ev_default_fork ()> or C<ev_loop_fork ()> after a fork in the child,
973	enable C<EVFLAG_FORKCHECK>, or resort to C<EVBACKEND_SELECT> or	1252	enable C<EVFLAG_FORKCHECK>, or resort to C<EVBACKEND_SELECT> or
974	C<EVBACKEND_POLL>.	1253	C<EVBACKEND_POLL>.
975		1254
		1255	=head3 The special problem of SIGPIPE
		1256
		1257	While not really specific to libev, it is easy to forget about C<SIGPIPE>:
		1258	when writing to a pipe whose other end has been closed, your program gets
		1259	sent a SIGPIPE, which, by default, aborts your program. For most programs
		1260	this is sensible behaviour, for daemons, this is usually undesirable.
		1261
		1262	So when you encounter spurious, unexplained daemon exits, make sure you
		1263	ignore SIGPIPE (and maybe make sure you log the exit status of your daemon
		1264	somewhere, as that would have given you a big clue).
		1265
976		1266
977	=head3 Watcher-Specific Functions	1267	=head3 Watcher-Specific Functions
978		1268
979	=over 4	1269	=over 4
980		1270
981	=item ev_io_init (ev_io *, callback, int fd, int events)	1271	=item ev_io_init (ev_io *, callback, int fd, int events)
982		1272
983	=item ev_io_set (ev_io *, int fd, int events)	1273	=item ev_io_set (ev_io *, int fd, int events)
984		1274
985	Configures an C<ev_io> watcher. The C<fd> is the file descriptor to	1275	Configures an C<ev_io> watcher. The C<fd> is the file descriptor to
986	rceeive events for and events is either C<EV_READ>, C<EV_WRITE> or	1276	receive events for and C<events> is either C<EV_READ>, C<EV_WRITE> or
987	C<EV_READ | EV_WRITE> to receive the given events.	1277	C<EV_READ | EV_WRITE>, to express the desire to receive the given events.
988		1278
989	=item int fd [read-only]	1279	=item int fd [read-only]
990		1280
991	The file descriptor being watched.	1281	The file descriptor being watched.
992		1282
993	=item int events [read-only]	1283	=item int events [read-only]
994		1284
995	The events being watched.	1285	The events being watched.
996		1286
997	=back	1287	=back
		1288
		1289	=head3 Examples
998		1290
999	Example: Call C<stdin_readable_cb> when STDIN_FILENO has become, well	1291	Example: Call C<stdin_readable_cb> when STDIN_FILENO has become, well
1000	readable, but only once. Since it is likely line-buffered, you could	1292	readable, but only once. Since it is likely line-buffered, you could
1001	attempt to read a whole line in the callback.	1293	attempt to read a whole line in the callback.
1002		1294
1003	static void	1295	static void
1004	stdin_readable_cb (struct ev_loop *loop, struct ev_io *w, int revents)	1296	stdin_readable_cb (struct ev_loop *loop, ev_io *w, int revents)
1005	{	1297	{
1006	ev_io_stop (loop, w);	1298	ev_io_stop (loop, w);
1007	.. read from stdin here (or from w->fd) and haqndle any I/O errors	1299	.. read from stdin here (or from w->fd) and handle any I/O errors
1008	}	1300	}
1009		1301
1010	...	1302	...
1011	struct ev_loop *loop = ev_default_init (0);	1303	struct ev_loop *loop = ev_default_init (0);
1012	struct ev_io stdin_readable;	1304	ev_io stdin_readable;
1013	ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);	1305	ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
1014	ev_io_start (loop, &stdin_readable);	1306	ev_io_start (loop, &stdin_readable);
1015	ev_loop (loop, 0);	1307	ev_loop (loop, 0);
1016		1308
1017		1309
1018	=head2 C<ev_timer> - relative and optionally repeating timeouts	1310	=head2 C<ev_timer> - relative and optionally repeating timeouts
1019		1311
1020	Timer watchers are simple relative timers that generate an event after a	1312	Timer watchers are simple relative timers that generate an event after a
1021	given time, and optionally repeating in regular intervals after that.	1313	given time, and optionally repeating in regular intervals after that.
1022		1314
1023	The timers are based on real time, that is, if you register an event that	1315	The timers are based on real time, that is, if you register an event that
1024	times out after an hour and you reset your system clock to last years	1316	times out after an hour and you reset your system clock to January last
1025	time, it will still time out after (roughly) and hour. "Roughly" because	1317	year, it will still time out after (roughly) one hour. "Roughly" because
1026	detecting time jumps is hard, and some inaccuracies are unavoidable (the	1318	detecting time jumps is hard, and some inaccuracies are unavoidable (the
1027	monotonic clock option helps a lot here).	1319	monotonic clock option helps a lot here).
		1320
		1321	The callback is guaranteed to be invoked only I<after> its timeout has
		1322	passed, but if multiple timers become ready during the same loop iteration
		1323	then order of execution is undefined.
		1324
		1325	=head3 Be smart about timeouts
		1326
		1327	Many real-world problems involve some kind of timeout, usually for error
		1328	recovery. A typical example is an HTTP request - if the other side hangs,
		1329	you want to raise some error after a while.
		1330
		1331	What follows are some ways to handle this problem, from obvious and
		1332	inefficient to smart and efficient.
		1333
		1334	In the following, a 60 second activity timeout is assumed - a timeout that
		1335	gets reset to 60 seconds each time there is activity (e.g. each time some
		1336	data or other life sign was received).
		1337
		1338	=over 4
		1339
		1340	=item 1. Use a timer and stop, reinitialise and start it on activity.
		1341
		1342	This is the most obvious, but not the most simple way: In the beginning,
		1343	start the watcher:
		1344
		1345	ev_timer_init (timer, callback, 60., 0.);
		1346	ev_timer_start (loop, timer);
		1347
		1348	Then, each time there is some activity, C<ev_timer_stop> it, initialise it
		1349	and start it again:
		1350
		1351	ev_timer_stop (loop, timer);
		1352	ev_timer_set (timer, 60., 0.);
		1353	ev_timer_start (loop, timer);
		1354
		1355	This is relatively simple to implement, but means that each time there is
		1356	some activity, libev will first have to remove the timer from its internal
		1357	data structure and then add it again. Libev tries to be fast, but it's
		1358	still not a constant-time operation.
		1359
		1360	=item 2. Use a timer and re-start it with C<ev_timer_again> inactivity.
		1361
		1362	This is the easiest way, and involves using C<ev_timer_again> instead of
		1363	C<ev_timer_start>.
		1364
		1365	To implement this, configure an C<ev_timer> with a C<repeat> value
		1366	of C<60> and then call C<ev_timer_again> at start and each time you
		1367	successfully read or write some data. If you go into an idle state where
		1368	you do not expect data to travel on the socket, you can C<ev_timer_stop>
		1369	the timer, and C<ev_timer_again> will automatically restart it if need be.
		1370
		1371	That means you can ignore both the C<ev_timer_start> function and the
		1372	C<after> argument to C<ev_timer_set>, and only ever use the C<repeat>
		1373	member and C<ev_timer_again>.
		1374
		1375	At start:
		1376
		1377	ev_timer_init (timer, callback);
		1378	timer->repeat = 60.;
		1379	ev_timer_again (loop, timer);
		1380
		1381	Each time there is some activity:
		1382
		1383	ev_timer_again (loop, timer);
		1384
		1385	It is even possible to change the time-out on the fly, regardless of
		1386	whether the watcher is active or not:
		1387
		1388	timer->repeat = 30.;
		1389	ev_timer_again (loop, timer);
		1390
		1391	This is slightly more efficient then stopping/starting the timer each time
		1392	you want to modify its timeout value, as libev does not have to completely
		1393	remove and re-insert the timer from/into its internal data structure.
		1394
		1395	It is, however, even simpler than the "obvious" way to do it.
		1396
		1397	=item 3. Let the timer time out, but then re-arm it as required.
		1398
		1399	This method is more tricky, but usually most efficient: Most timeouts are
		1400	relatively long compared to the intervals between other activity - in
		1401	our example, within 60 seconds, there are usually many I/O events with
		1402	associated activity resets.
		1403
		1404	In this case, it would be more efficient to leave the C<ev_timer> alone,
		1405	but remember the time of last activity, and check for a real timeout only
		1406	within the callback:
		1407
		1408	ev_tstamp last_activity; // time of last activity
		1409
		1410	static void
		1411	callback (EV_P_ ev_timer *w, int revents)
		1412	{
		1413	ev_tstamp now = ev_now (EV_A);
		1414	ev_tstamp timeout = last_activity + 60.;
		1415
		1416	// if last_activity + 60. is older than now, we did time out
		1417	if (timeout < now)
		1418	{
		1419	// timeout occured, take action
		1420	}
		1421	else
		1422	{
		1423	// callback was invoked, but there was some activity, re-arm
		1424	// the watcher to fire in last_activity + 60, which is
		1425	// guaranteed to be in the future, so "again" is positive:
		1426	w->repeat = timeout - now;
		1427	ev_timer_again (EV_A_ w);
		1428	}
		1429	}
		1430
		1431	To summarise the callback: first calculate the real timeout (defined
		1432	as "60 seconds after the last activity"), then check if that time has
		1433	been reached, which means something I<did>, in fact, time out. Otherwise
		1434	the callback was invoked too early (C<timeout> is in the future), so
		1435	re-schedule the timer to fire at that future time, to see if maybe we have
		1436	a timeout then.
		1437
		1438	Note how C<ev_timer_again> is used, taking advantage of the
		1439	C<ev_timer_again> optimisation when the timer is already running.
		1440
		1441	This scheme causes more callback invocations (about one every 60 seconds
		1442	minus half the average time between activity), but virtually no calls to
		1443	libev to change the timeout.
		1444
		1445	To start the timer, simply initialise the watcher and set C<last_activity>
		1446	to the current time (meaning we just have some activity :), then call the
		1447	callback, which will "do the right thing" and start the timer:
		1448
		1449	ev_timer_init (timer, callback);
		1450	last_activity = ev_now (loop);
		1451	callback (loop, timer, EV_TIMEOUT);
		1452
		1453	And when there is some activity, simply store the current time in
		1454	C<last_activity>, no libev calls at all:
		1455
		1456	last_actiivty = ev_now (loop);
		1457
		1458	This technique is slightly more complex, but in most cases where the
		1459	time-out is unlikely to be triggered, much more efficient.
		1460
		1461	Changing the timeout is trivial as well (if it isn't hard-coded in the
		1462	callback :) - just change the timeout and invoke the callback, which will
		1463	fix things for you.
		1464
		1465	=item 4. Wee, just use a double-linked list for your timeouts.
		1466
		1467	If there is not one request, but many thousands (millions...), all
		1468	employing some kind of timeout with the same timeout value, then one can
		1469	do even better:
		1470
		1471	When starting the timeout, calculate the timeout value and put the timeout
		1472	at the I<end> of the list.
		1473
		1474	Then use an C<ev_timer> to fire when the timeout at the I<beginning> of
		1475	the list is expected to fire (for example, using the technique #3).
		1476
		1477	When there is some activity, remove the timer from the list, recalculate
		1478	the timeout, append it to the end of the list again, and make sure to
		1479	update the C<ev_timer> if it was taken from the beginning of the list.
		1480
		1481	This way, one can manage an unlimited number of timeouts in O(1) time for
		1482	starting, stopping and updating the timers, at the expense of a major
		1483	complication, and having to use a constant timeout. The constant timeout
		1484	ensures that the list stays sorted.
		1485
		1486	=back
		1487
		1488	So which method the best?
		1489
		1490	Method #2 is a simple no-brain-required solution that is adequate in most
		1491	situations. Method #3 requires a bit more thinking, but handles many cases
		1492	better, and isn't very complicated either. In most case, choosing either
		1493	one is fine, with #3 being better in typical situations.
		1494
		1495	Method #1 is almost always a bad idea, and buys you nothing. Method #4 is
		1496	rather complicated, but extremely efficient, something that really pays
		1497	off after the first million or so of active timers, i.e. it's usually
		1498	overkill :)
		1499
		1500	=head3 The special problem of time updates
		1501
		1502	Establishing the current time is a costly operation (it usually takes at
		1503	least two system calls): EV therefore updates its idea of the current
		1504	time only before and after C<ev_loop> collects new events, which causes a
		1505	growing difference between C<ev_now ()> and C<ev_time ()> when handling
		1506	lots of events in one iteration.
1028		1507
1029	The relative timeouts are calculated relative to the C<ev_now ()>	1508	The relative timeouts are calculated relative to the C<ev_now ()>
1030	time. This is usually the right thing as this timestamp refers to the time	1509	time. This is usually the right thing as this timestamp refers to the time
1031	of the event triggering whatever timeout you are modifying/starting. If	1510	of the event triggering whatever timeout you are modifying/starting. If
1032	you suspect event processing to be delayed and you I<need> to base the timeout	1511	you suspect event processing to be delayed and you I<need> to base the
1033	on the current time, use something like this to adjust for this:	1512	timeout on the current time, use something like this to adjust for this:
1034		1513
1035	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);	1514	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);
1036		1515
1037	The callback is guarenteed to be invoked only when its timeout has passed,	1516	If the event loop is suspended for a long time, you can also force an
1038	but if multiple timers become ready during the same loop iteration then	1517	update of the time returned by C<ev_now ()> by calling C<ev_now_update
1039	order of execution is undefined.	1518	()>.
1040		1519
1041	=head3 Watcher-Specific Functions and Data Members	1520	=head3 Watcher-Specific Functions and Data Members
1042		1521
1043	=over 4	1522	=over 4
1044		1523
1045	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)	1524	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)
1046		1525
1047	=item ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)	1526	=item ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)
1048		1527
1049	Configure the timer to trigger after C<after> seconds. If C<repeat> is	1528	Configure the timer to trigger after C<after> seconds. If C<repeat>
1050	C<0.>, then it will automatically be stopped. If it is positive, then the	1529	is C<0.>, then it will automatically be stopped once the timeout is
1051	timer will automatically be configured to trigger again C<repeat> seconds	1530	reached. If it is positive, then the timer will automatically be
1052	later, again, and again, until stopped manually.	1531	configured to trigger again C<repeat> seconds later, again, and again,
		1532	until stopped manually.
1053		1533
1054	The timer itself will do a best-effort at avoiding drift, that is, if you	1534	The timer itself will do a best-effort at avoiding drift, that is, if
1055	configure a timer to trigger every 10 seconds, then it will trigger at	1535	you configure a timer to trigger every 10 seconds, then it will normally
1056	exactly 10 second intervals. If, however, your program cannot keep up with	1536	trigger at exactly 10 second intervals. If, however, your program cannot
1057	the timer (because it takes longer than those 10 seconds to do stuff) the	1537	keep up with the timer (because it takes longer than those 10 seconds to
1058	timer will not fire more than once per event loop iteration.	1538	do stuff) the timer will not fire more than once per event loop iteration.
1059		1539
1060	=item ev_timer_again (loop)	1540	=item ev_timer_again (loop, ev_timer *)
1061		1541
1062	This will act as if the timer timed out and restart it again if it is	1542	This will act as if the timer timed out and restart it again if it is
1063	repeating. The exact semantics are:	1543	repeating. The exact semantics are:
1064		1544
1065	If the timer is pending, its pending status is cleared.	1545	If the timer is pending, its pending status is cleared.
1066		1546
1067	If the timer is started but nonrepeating, stop it (as if it timed out).	1547	If the timer is started but non-repeating, stop it (as if it timed out).
1068		1548
1069	If the timer is repeating, either start it if necessary (with the	1549	If the timer is repeating, either start it if necessary (with the
1070	C<repeat> value), or reset the running timer to the C<repeat> value.	1550	C<repeat> value), or reset the running timer to the C<repeat> value.
1071		1551
1072	This sounds a bit complicated, but here is a useful and typical	1552	This sounds a bit complicated, see "Be smart about timeouts", above, for a
1073	example: Imagine you have a tcp connection and you want a so-called idle	1553	usage example.
1074	timeout, that is, you want to be called when there have been, say, 60
1075	seconds of inactivity on the socket. The easiest way to do this is to
1076	configure an C<ev_timer> with a C<repeat> value of C<60> and then call
1077	C<ev_timer_again> each time you successfully read or write some data. If
1078	you go into an idle state where you do not expect data to travel on the
1079	socket, you can C<ev_timer_stop> the timer, and C<ev_timer_again> will
1080	automatically restart it if need be.
1081
1082	That means you can ignore the C<after> value and C<ev_timer_start>
1083	altogether and only ever use the C<repeat> value and C<ev_timer_again>:
1084
1085	ev_timer_init (timer, callback, 0., 5.);
1086	ev_timer_again (loop, timer);
1087	...
1088	timer->again = 17.;
1089	ev_timer_again (loop, timer);
1090	...
1091	timer->again = 10.;
1092	ev_timer_again (loop, timer);
1093
1094	This is more slightly efficient then stopping/starting the timer each time
1095	you want to modify its timeout value.
1096		1554
1097	=item ev_tstamp repeat [read-write]	1555	=item ev_tstamp repeat [read-write]
1098		1556
1099	The current C<repeat> value. Will be used each time the watcher times out	1557	The current C<repeat> value. Will be used each time the watcher times out
1100	or C<ev_timer_again> is called and determines the next timeout (if any),	1558	or C<ev_timer_again> is called, and determines the next timeout (if any),
1101	which is also when any modifications are taken into account.	1559	which is also when any modifications are taken into account.
1102		1560
1103	=back	1561	=back
1104		1562
		1563	=head3 Examples
		1564
1105	Example: Create a timer that fires after 60 seconds.	1565	Example: Create a timer that fires after 60 seconds.
1106		1566
1107	static void	1567	static void
1108	one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)	1568	one_minute_cb (struct ev_loop *loop, ev_timer *w, int revents)
1109	{	1569	{
1110	.. one minute over, w is actually stopped right here	1570	.. one minute over, w is actually stopped right here
1111	}	1571	}
1112		1572
1113	struct ev_timer mytimer;	1573	ev_timer mytimer;
1114	ev_timer_init (&mytimer, one_minute_cb, 60., 0.);	1574	ev_timer_init (&mytimer, one_minute_cb, 60., 0.);
1115	ev_timer_start (loop, &mytimer);	1575	ev_timer_start (loop, &mytimer);
1116		1576
1117	Example: Create a timeout timer that times out after 10 seconds of	1577	Example: Create a timeout timer that times out after 10 seconds of
1118	inactivity.	1578	inactivity.
1119		1579
1120	static void	1580	static void
1121	timeout_cb (struct ev_loop *loop, struct ev_timer *w, int revents)	1581	timeout_cb (struct ev_loop *loop, ev_timer *w, int revents)
1122	{	1582	{
1123	.. ten seconds without any activity	1583	.. ten seconds without any activity
1124	}	1584	}
1125		1585
1126	struct ev_timer mytimer;	1586	ev_timer mytimer;
1127	ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */	1587	ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */
1128	ev_timer_again (&mytimer); /* start timer */	1588	ev_timer_again (&mytimer); /* start timer */
1129	ev_loop (loop, 0);	1589	ev_loop (loop, 0);
1130		1590
1131	// and in some piece of code that gets executed on any "activity":	1591	// and in some piece of code that gets executed on any "activity":
1132	// reset the timeout to start ticking again at 10 seconds	1592	// reset the timeout to start ticking again at 10 seconds
1133	ev_timer_again (&mytimer);	1593	ev_timer_again (&mytimer);
1134		1594
1135		1595
1136	=head2 C<ev_periodic> - to cron or not to cron?	1596	=head2 C<ev_periodic> - to cron or not to cron?
1137		1597
1138	Periodic watchers are also timers of a kind, but they are very versatile	1598	Periodic watchers are also timers of a kind, but they are very versatile
1139	(and unfortunately a bit complex).	1599	(and unfortunately a bit complex).
1140		1600
1141	Unlike C<ev_timer>'s, they are not based on real time (or relative time)	1601	Unlike C<ev_timer>, periodic watchers are not based on real time (or
1142	but on wallclock time (absolute time). You can tell a periodic watcher	1602	relative time, the physical time that passes) but on wall clock time
1143	to trigger "at" some specific point in time. For example, if you tell a	1603	(absolute time, the thing you can read on your calender or clock). The
1144	periodic watcher to trigger in 10 seconds (by specifiying e.g. C<ev_now ()	1604	difference is that wall clock time can run faster or slower than real
1145	+ 10.>) and then reset your system clock to the last year, then it will	1605	time, and time jumps are not uncommon (e.g. when you adjust your
1146	take a year to trigger the event (unlike an C<ev_timer>, which would trigger	1606	wrist-watch).
1147	roughly 10 seconds later).
1148		1607
1149	They can also be used to implement vastly more complex timers, such as	1608	You can tell a periodic watcher to trigger after some specific point
1150	triggering an event on each midnight, local time or other, complicated,	1609	in time: for example, if you tell a periodic watcher to trigger "in 10
1151	rules.	1610	seconds" (by specifying e.g. C<ev_now () + 10.>, that is, an absolute time
		1611	not a delay) and then reset your system clock to January of the previous
		1612	year, then it will take a year or more to trigger the event (unlike an
		1613	C<ev_timer>, which would still trigger roughly 10 seconds after starting
		1614	it, as it uses a relative timeout).
1152		1615
		1616	C<ev_periodic> watchers can also be used to implement vastly more complex
		1617	timers, such as triggering an event on each "midnight, local time", or
		1618	other complicated rules. This cannot be done with C<ev_timer> watchers, as
		1619	those cannot react to time jumps.
		1620
1153	As with timers, the callback is guarenteed to be invoked only when the	1621	As with timers, the callback is guaranteed to be invoked only when the
1154	time (C<at>) has been passed, but if multiple periodic timers become ready	1622	point in time where it is supposed to trigger has passed, but if multiple
1155	during the same loop iteration then order of execution is undefined.	1623	periodic timers become ready during the same loop iteration, then order of
		1624	execution is undefined.
1156		1625
1157	=head3 Watcher-Specific Functions and Data Members	1626	=head3 Watcher-Specific Functions and Data Members
1158		1627
1159	=over 4	1628	=over 4
1160		1629
1161	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)	1630	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp offset, ev_tstamp interval, reschedule_cb)
1162		1631
1163	=item ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)	1632	=item ev_periodic_set (ev_periodic *, ev_tstamp offset, ev_tstamp interval, reschedule_cb)
1164		1633
1165	Lots of arguments, lets sort it out... There are basically three modes of	1634	Lots of arguments, let's sort it out... There are basically three modes of
1166	operation, and we will explain them from simplest to complex:	1635	operation, and we will explain them from simplest to most complex:
1167		1636
1168	=over 4	1637	=over 4
1169		1638
1170	=item * absolute timer (at = time, interval = reschedule_cb = 0)	1639	=item * absolute timer (offset = absolute time, interval = 0, reschedule_cb = 0)
1171		1640
1172	In this configuration the watcher triggers an event at the wallclock time	1641	In this configuration the watcher triggers an event after the wall clock
1173	C<at> and doesn't repeat. It will not adjust when a time jump occurs,	1642	time C<offset> has passed. It will not repeat and will not adjust when a
1174	that is, if it is to be run at January 1st 2011 then it will run when the	1643	time jump occurs, that is, if it is to be run at January 1st 2011 then it
1175	system time reaches or surpasses this time.	1644	will be stopped and invoked when the system clock reaches or surpasses
		1645	this point in time.
1176		1646
1177	=item * non-repeating interval timer (at = offset, interval > 0, reschedule_cb = 0)	1647	=item * repeating interval timer (offset = offset within interval, interval > 0, reschedule_cb = 0)
1178		1648
1179	In this mode the watcher will always be scheduled to time out at the next	1649	In this mode the watcher will always be scheduled to time out at the next
1180	C<at + N * interval> time (for some integer N, which can also be negative)	1650	C<offset + N * interval> time (for some integer N, which can also be
1181	and then repeat, regardless of any time jumps.	1651	negative) and then repeat, regardless of any time jumps. The C<offset>
		1652	argument is merely an offset into the C<interval> periods.
1182		1653
1183	This can be used to create timers that do not drift with respect to system	1654	This can be used to create timers that do not drift with respect to the
1184	time:	1655	system clock, for example, here is an C<ev_periodic> that triggers each
		1656	hour, on the hour (with respect to UTC):
1185		1657
1186	ev_periodic_set (&periodic, 0., 3600., 0);	1658	ev_periodic_set (&periodic, 0., 3600., 0);
1187		1659
1188	This doesn't mean there will always be 3600 seconds in between triggers,	1660	This doesn't mean there will always be 3600 seconds in between triggers,
1189	but only that the the callback will be called when the system time shows a	1661	but only that the callback will be called when the system time shows a
1190	full hour (UTC), or more correctly, when the system time is evenly divisible	1662	full hour (UTC), or more correctly, when the system time is evenly divisible
1191	by 3600.	1663	by 3600.
1192		1664
1193	Another way to think about it (for the mathematically inclined) is that	1665	Another way to think about it (for the mathematically inclined) is that
1194	C<ev_periodic> will try to run the callback in this mode at the next possible	1666	C<ev_periodic> will try to run the callback in this mode at the next possible
1195	time where C<time = at (mod interval)>, regardless of any time jumps.	1667	time where C<time = offset (mod interval)>, regardless of any time jumps.
1196		1668
1197	For numerical stability it is preferable that the C<at> value is near	1669	For numerical stability it is preferable that the C<offset> value is near
1198	C<ev_now ()> (the current time), but there is no range requirement for	1670	C<ev_now ()> (the current time), but there is no range requirement for
1199	this value.	1671	this value, and in fact is often specified as zero.
1200		1672
		1673	Note also that there is an upper limit to how often a timer can fire (CPU
		1674	speed for example), so if C<interval> is very small then timing stability
		1675	will of course deteriorate. Libev itself tries to be exact to be about one
		1676	millisecond (if the OS supports it and the machine is fast enough).
		1677
1201	=item * manual reschedule mode (at and interval ignored, reschedule_cb = callback)	1678	=item * manual reschedule mode (offset ignored, interval ignored, reschedule_cb = callback)
1202		1679
1203	In this mode the values for C<interval> and C<at> are both being	1680	In this mode the values for C<interval> and C<offset> are both being
1204	ignored. Instead, each time the periodic watcher gets scheduled, the	1681	ignored. Instead, each time the periodic watcher gets scheduled, the
1205	reschedule callback will be called with the watcher as first, and the	1682	reschedule callback will be called with the watcher as first, and the
1206	current time as second argument.	1683	current time as second argument.
1207		1684
1208	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher,	1685	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher, ever,
1209	ever, or make any event loop modifications>. If you need to stop it,	1686	or make ANY other event loop modifications whatsoever, unless explicitly
1210	return C<now + 1e30> (or so, fudge fudge) and stop it afterwards (e.g. by	1687	allowed by documentation here>.
1211	starting an C<ev_prepare> watcher, which is legal).
1212		1688
		1689	If you need to stop it, return C<now + 1e30> (or so, fudge fudge) and stop
		1690	it afterwards (e.g. by starting an C<ev_prepare> watcher, which is the
		1691	only event loop modification you are allowed to do).
		1692
1213	Its prototype is C<ev_tstamp (*reschedule_cb)(struct ev_periodic *w,	1693	The callback prototype is C<ev_tstamp (*reschedule_cb)(ev_periodic
1214	ev_tstamp now)>, e.g.:	1694	*w, ev_tstamp now)>, e.g.:
1215		1695
		1696	static ev_tstamp
1216	static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)	1697	my_rescheduler (ev_periodic *w, ev_tstamp now)
1217	{	1698	{
1218	return now + 60.;	1699	return now + 60.;
1219	}	1700	}
1220		1701
1221	It must return the next time to trigger, based on the passed time value	1702	It must return the next time to trigger, based on the passed time value
1222	(that is, the lowest time value larger than to the second argument). It	1703	(that is, the lowest time value larger than to the second argument). It
1223	will usually be called just before the callback will be triggered, but	1704	will usually be called just before the callback will be triggered, but
1224	might be called at other times, too.	1705	might be called at other times, too.
1225		1706
1226	NOTE: I<< This callback must always return a time that is later than the	1707	NOTE: I<< This callback must always return a time that is higher than or
1227	passed C<now> value >>. Not even C<now> itself will do, it I<must> be larger.	1708	equal to the passed C<now> value >>.
1228		1709
1229	This can be used to create very complex timers, such as a timer that	1710	This can be used to create very complex timers, such as a timer that
1230	triggers on each midnight, local time. To do this, you would calculate the	1711	triggers on "next midnight, local time". To do this, you would calculate the
1231	next midnight after C<now> and return the timestamp value for this. How	1712	next midnight after C<now> and return the timestamp value for this. How
1232	you do this is, again, up to you (but it is not trivial, which is the main	1713	you do this is, again, up to you (but it is not trivial, which is the main
1233	reason I omitted it as an example).	1714	reason I omitted it as an example).
1234		1715
1235	=back	1716	=back
…		…
1239	Simply stops and restarts the periodic watcher again. This is only useful	1720	Simply stops and restarts the periodic watcher again. This is only useful
1240	when you changed some parameters or the reschedule callback would return	1721	when you changed some parameters or the reschedule callback would return
1241	a different time than the last time it was called (e.g. in a crond like	1722	a different time than the last time it was called (e.g. in a crond like
1242	program when the crontabs have changed).	1723	program when the crontabs have changed).
1243		1724
		1725	=item ev_tstamp ev_periodic_at (ev_periodic *)
		1726
		1727	When active, returns the absolute time that the watcher is supposed
		1728	to trigger next. This is not the same as the C<offset> argument to
		1729	C<ev_periodic_set>, but indeed works even in interval and manual
		1730	rescheduling modes.
		1731
1244	=item ev_tstamp offset [read-write]	1732	=item ev_tstamp offset [read-write]
1245		1733
1246	When repeating, this contains the offset value, otherwise this is the	1734	When repeating, this contains the offset value, otherwise this is the
1247	absolute point in time (the C<at> value passed to C<ev_periodic_set>).	1735	absolute point in time (the C<offset> value passed to C<ev_periodic_set>,
		1736	although libev might modify this value for better numerical stability).
1248		1737
1249	Can be modified any time, but changes only take effect when the periodic	1738	Can be modified any time, but changes only take effect when the periodic
1250	timer fires or C<ev_periodic_again> is being called.	1739	timer fires or C<ev_periodic_again> is being called.
1251		1740
1252	=item ev_tstamp interval [read-write]	1741	=item ev_tstamp interval [read-write]
1253		1742
1254	The current interval value. Can be modified any time, but changes only	1743	The current interval value. Can be modified any time, but changes only
1255	take effect when the periodic timer fires or C<ev_periodic_again> is being	1744	take effect when the periodic timer fires or C<ev_periodic_again> is being
1256	called.	1745	called.
1257		1746
1258	=item ev_tstamp (*reschedule_cb)(struct ev_periodic *w, ev_tstamp now) [read-write]	1747	=item ev_tstamp (*reschedule_cb)(ev_periodic *w, ev_tstamp now) [read-write]
1259		1748
1260	The current reschedule callback, or C<0>, if this functionality is	1749	The current reschedule callback, or C<0>, if this functionality is
1261	switched off. Can be changed any time, but changes only take effect when	1750	switched off. Can be changed any time, but changes only take effect when
1262	the periodic timer fires or C<ev_periodic_again> is being called.	1751	the periodic timer fires or C<ev_periodic_again> is being called.
1263		1752
1264	=item ev_tstamp at [read-only]
1265
1266	When active, contains the absolute time that the watcher is supposed to
1267	trigger next.
1268
1269	=back	1753	=back
1270		1754
		1755	=head3 Examples
		1756
1271	Example: Call a callback every hour, or, more precisely, whenever the	1757	Example: Call a callback every hour, or, more precisely, whenever the
1272	system clock is divisible by 3600. The callback invocation times have	1758	system time is divisible by 3600. The callback invocation times have
1273	potentially a lot of jittering, but good long-term stability.	1759	potentially a lot of jitter, but good long-term stability.
1274		1760
1275	static void	1761	static void
1276	clock_cb (struct ev_loop *loop, struct ev_io *w, int revents)	1762	clock_cb (struct ev_loop *loop, ev_io *w, int revents)
1277	{	1763	{
1278	... its now a full hour (UTC, or TAI or whatever your clock follows)	1764	... its now a full hour (UTC, or TAI or whatever your clock follows)
1279	}	1765	}
1280		1766
1281	struct ev_periodic hourly_tick;	1767	ev_periodic hourly_tick;
1282	ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);	1768	ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);
1283	ev_periodic_start (loop, &hourly_tick);	1769	ev_periodic_start (loop, &hourly_tick);
1284		1770
1285	Example: The same as above, but use a reschedule callback to do it:	1771	Example: The same as above, but use a reschedule callback to do it:
1286		1772
1287	#include <math.h>	1773	#include <math.h>
1288		1774
1289	static ev_tstamp	1775	static ev_tstamp
1290	my_scheduler_cb (struct ev_periodic *w, ev_tstamp now)	1776	my_scheduler_cb (ev_periodic *w, ev_tstamp now)
1291	{	1777	{
1292	return fmod (now, 3600.) + 3600.;	1778	return now + (3600. - fmod (now, 3600.));
1293	}	1779	}
1294		1780
1295	ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);	1781	ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
1296		1782
1297	Example: Call a callback every hour, starting now:	1783	Example: Call a callback every hour, starting now:
1298		1784
1299	struct ev_periodic hourly_tick;	1785	ev_periodic hourly_tick;
1300	ev_periodic_init (&hourly_tick, clock_cb,	1786	ev_periodic_init (&hourly_tick, clock_cb,
1301	fmod (ev_now (loop), 3600.), 3600., 0);	1787	fmod (ev_now (loop), 3600.), 3600., 0);
1302	ev_periodic_start (loop, &hourly_tick);	1788	ev_periodic_start (loop, &hourly_tick);
1303		1789
1304		1790
1305	=head2 C<ev_signal> - signal me when a signal gets signalled!	1791	=head2 C<ev_signal> - signal me when a signal gets signalled!
1306		1792
1307	Signal watchers will trigger an event when the process receives a specific	1793	Signal watchers will trigger an event when the process receives a specific
1308	signal one or more times. Even though signals are very asynchronous, libev	1794	signal one or more times. Even though signals are very asynchronous, libev
1309	will try it's best to deliver signals synchronously, i.e. as part of the	1795	will try it's best to deliver signals synchronously, i.e. as part of the
1310	normal event processing, like any other event.	1796	normal event processing, like any other event.
1311		1797
		1798	If you want signals asynchronously, just use C<sigaction> as you would
		1799	do without libev and forget about sharing the signal. You can even use
		1800	C<ev_async> from a signal handler to synchronously wake up an event loop.
		1801
1312	You can configure as many watchers as you like per signal. Only when the	1802	You can configure as many watchers as you like per signal. Only when the
1313	first watcher gets started will libev actually register a signal watcher	1803	first watcher gets started will libev actually register a signal handler
1314	with the kernel (thus it coexists with your own signal handlers as long	1804	with the kernel (thus it coexists with your own signal handlers as long as
1315	as you don't register any with libev). Similarly, when the last signal	1805	you don't register any with libev for the same signal). Similarly, when
1316	watcher for a signal is stopped libev will reset the signal handler to	1806	the last signal watcher for a signal is stopped, libev will reset the
1317	SIG_DFL (regardless of what it was set to before).	1807	signal handler to SIG_DFL (regardless of what it was set to before).
		1808
		1809	If possible and supported, libev will install its handlers with
		1810	C<SA_RESTART> behaviour enabled, so system calls should not be unduly
		1811	interrupted. If you have a problem with system calls getting interrupted by
		1812	signals you can block all signals in an C<ev_check> watcher and unblock
		1813	them in an C<ev_prepare> watcher.
1318		1814
1319	=head3 Watcher-Specific Functions and Data Members	1815	=head3 Watcher-Specific Functions and Data Members
1320		1816
1321	=over 4	1817	=over 4
1322		1818
…		…
1331		1827
1332	The signal the watcher watches out for.	1828	The signal the watcher watches out for.
1333		1829
1334	=back	1830	=back
1335		1831
		1832	=head3 Examples
		1833
		1834	Example: Try to exit cleanly on SIGINT.
		1835
		1836	static void
		1837	sigint_cb (struct ev_loop *loop, ev_signal *w, int revents)
		1838	{
		1839	ev_unloop (loop, EVUNLOOP_ALL);
		1840	}
		1841
		1842	ev_signal signal_watcher;
		1843	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
		1844	ev_signal_start (loop, &signal_watcher);
		1845
1336		1846
1337	=head2 C<ev_child> - watch out for process status changes	1847	=head2 C<ev_child> - watch out for process status changes
1338		1848
1339	Child watchers trigger when your process receives a SIGCHLD in response to	1849	Child watchers trigger when your process receives a SIGCHLD in response to
1340	some child status changes (most typically when a child of yours dies).	1850	some child status changes (most typically when a child of yours dies or
		1851	exits). It is permissible to install a child watcher I<after> the child
		1852	has been forked (which implies it might have already exited), as long
		1853	as the event loop isn't entered (or is continued from a watcher), i.e.,
		1854	forking and then immediately registering a watcher for the child is fine,
		1855	but forking and registering a watcher a few event loop iterations later is
		1856	not.
		1857
		1858	Only the default event loop is capable of handling signals, and therefore
		1859	you can only register child watchers in the default event loop.
		1860
		1861	=head3 Process Interaction
		1862
		1863	Libev grabs C<SIGCHLD> as soon as the default event loop is
		1864	initialised. This is necessary to guarantee proper behaviour even if
		1865	the first child watcher is started after the child exits. The occurrence
		1866	of C<SIGCHLD> is recorded asynchronously, but child reaping is done
		1867	synchronously as part of the event loop processing. Libev always reaps all
		1868	children, even ones not watched.
		1869
		1870	=head3 Overriding the Built-In Processing
		1871
		1872	Libev offers no special support for overriding the built-in child
		1873	processing, but if your application collides with libev's default child
		1874	handler, you can override it easily by installing your own handler for
		1875	C<SIGCHLD> after initialising the default loop, and making sure the
		1876	default loop never gets destroyed. You are encouraged, however, to use an
		1877	event-based approach to child reaping and thus use libev's support for
		1878	that, so other libev users can use C<ev_child> watchers freely.
		1879
		1880	=head3 Stopping the Child Watcher
		1881
		1882	Currently, the child watcher never gets stopped, even when the
		1883	child terminates, so normally one needs to stop the watcher in the
		1884	callback. Future versions of libev might stop the watcher automatically
		1885	when a child exit is detected.
1341		1886
1342	=head3 Watcher-Specific Functions and Data Members	1887	=head3 Watcher-Specific Functions and Data Members
1343		1888
1344	=over 4	1889	=over 4
1345		1890
1346	=item ev_child_init (ev_child *, callback, int pid)	1891	=item ev_child_init (ev_child *, callback, int pid, int trace)
1347		1892
1348	=item ev_child_set (ev_child *, int pid)	1893	=item ev_child_set (ev_child *, int pid, int trace)
1349		1894
1350	Configures the watcher to wait for status changes of process C<pid> (or	1895	Configures the watcher to wait for status changes of process C<pid> (or
1351	I<any> process if C<pid> is specified as C<0>). The callback can look	1896	I<any> process if C<pid> is specified as C<0>). The callback can look
1352	at the C<rstatus> member of the C<ev_child> watcher structure to see	1897	at the C<rstatus> member of the C<ev_child> watcher structure to see
1353	the status word (use the macros from C<sys/wait.h> and see your systems	1898	the status word (use the macros from C<sys/wait.h> and see your systems
1354	C<waitpid> documentation). The C<rpid> member contains the pid of the	1899	C<waitpid> documentation). The C<rpid> member contains the pid of the
1355	process causing the status change.	1900	process causing the status change. C<trace> must be either C<0> (only
		1901	activate the watcher when the process terminates) or C<1> (additionally
		1902	activate the watcher when the process is stopped or continued).
1356		1903
1357	=item int pid [read-only]	1904	=item int pid [read-only]
1358		1905
1359	The process id this watcher watches out for, or C<0>, meaning any process id.	1906	The process id this watcher watches out for, or C<0>, meaning any process id.
1360		1907
…		…
1367	The process exit/trace status caused by C<rpid> (see your systems	1914	The process exit/trace status caused by C<rpid> (see your systems
1368	C<waitpid> and C<sys/wait.h> documentation for details).	1915	C<waitpid> and C<sys/wait.h> documentation for details).
1369		1916
1370	=back	1917	=back
1371		1918
1372	Example: Try to exit cleanly on SIGINT and SIGTERM.	1919	=head3 Examples
1373		1920
		1921	Example: C<fork()> a new process and install a child handler to wait for
		1922	its completion.
		1923
		1924	ev_child cw;
		1925
1374	static void	1926	static void
1375	sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)	1927	child_cb (EV_P_ ev_child *w, int revents)
1376	{	1928	{
1377	ev_unloop (loop, EVUNLOOP_ALL);	1929	ev_child_stop (EV_A_ w);
		1930	printf ("process %d exited with status %x\n", w->rpid, w->rstatus);
1378	}	1931	}
1379		1932
1380	struct ev_signal signal_watcher;	1933	pid_t pid = fork ();
1381	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);	1934
1382	ev_signal_start (loop, &sigint_cb);	1935	if (pid < 0)
		1936	// error
		1937	else if (pid == 0)
		1938	{
		1939	// the forked child executes here
		1940	exit (1);
		1941	}
		1942	else
		1943	{
		1944	ev_child_init (&cw, child_cb, pid, 0);
		1945	ev_child_start (EV_DEFAULT_ &cw);
		1946	}
1383		1947
1384		1948
1385	=head2 C<ev_stat> - did the file attributes just change?	1949	=head2 C<ev_stat> - did the file attributes just change?
1386		1950
1387	This watches a filesystem path for attribute changes. That is, it calls	1951	This watches a file system path for attribute changes. That is, it calls
1388	C<stat> regularly (or when the OS says it changed) and sees if it changed	1952	C<stat> on that path in regular intervals (or when the OS says it changed)
1389	compared to the last time, invoking the callback if it did.	1953	and sees if it changed compared to the last time, invoking the callback if
		1954	it did.
1390		1955
1391	The path does not need to exist: changing from "path exists" to "path does	1956	The path does not need to exist: changing from "path exists" to "path does
1392	not exist" is a status change like any other. The condition "path does	1957	not exist" is a status change like any other. The condition "path does not
1393	not exist" is signified by the C<st_nlink> field being zero (which is	1958	exist" (or more correctly "path cannot be stat'ed") is signified by the
1394	otherwise always forced to be at least one) and all the other fields of	1959	C<st_nlink> field being zero (which is otherwise always forced to be at
1395	the stat buffer having unspecified contents.	1960	least one) and all the other fields of the stat buffer having unspecified
		1961	contents.
1396		1962
1397	The path I<should> be absolute and I<must not> end in a slash. If it is	1963	The path I<must not> end in a slash or contain special components such as
		1964	C<.> or C<..>. The path I<should> be absolute: If it is relative and
1398	relative and your working directory changes, the behaviour is undefined.	1965	your working directory changes, then the behaviour is undefined.
1399		1966
1400	Since there is no standard to do this, the portable implementation simply	1967	Since there is no portable change notification interface available, the
1401	calls C<stat (2)> regularly on the path to see if it changed somehow. You	1968	portable implementation simply calls C<stat(2)> regularly on the path
1402	can specify a recommended polling interval for this case. If you specify	1969	to see if it changed somehow. You can specify a recommended polling
1403	a polling interval of C<0> (highly recommended!) then a I<suitable,	1970	interval for this case. If you specify a polling interval of C<0> (highly
1404	unspecified default> value will be used (which you can expect to be around	1971	recommended!) then a I<suitable, unspecified default> value will be used
1405	five seconds, although this might change dynamically). Libev will also	1972	(which you can expect to be around five seconds, although this might
1406	impose a minimum interval which is currently around C<0.1>, but thats	1973	change dynamically). Libev will also impose a minimum interval which is
1407	usually overkill.	1974	currently around C<0.1>, but that's usually overkill.
1408		1975
1409	This watcher type is not meant for massive numbers of stat watchers,	1976	This watcher type is not meant for massive numbers of stat watchers,
1410	as even with OS-supported change notifications, this can be	1977	as even with OS-supported change notifications, this can be
1411	resource-intensive.	1978	resource-intensive.
1412		1979
1413	At the time of this writing, only the Linux inotify interface is	1980	At the time of this writing, the only OS-specific interface implemented
1414	implemented (implementing kqueue support is left as an exercise for the	1981	is the Linux inotify interface (implementing kqueue support is left as an
1415	reader). Inotify will be used to give hints only and should not change the	1982	exercise for the reader. Note, however, that the author sees no way of
1416	semantics of C<ev_stat> watchers, which means that libev sometimes needs	1983	implementing C<ev_stat> semantics with kqueue, except as a hint).
1417	to fall back to regular polling again even with inotify, but changes are	1984
1418	usually detected immediately, and if the file exists there will be no	1985	=head3 ABI Issues (Largefile Support)
1419	polling.	1986
		1987	Libev by default (unless the user overrides this) uses the default
		1988	compilation environment, which means that on systems with large file
		1989	support disabled by default, you get the 32 bit version of the stat
		1990	structure. When using the library from programs that change the ABI to
		1991	use 64 bit file offsets the programs will fail. In that case you have to
		1992	compile libev with the same flags to get binary compatibility. This is
		1993	obviously the case with any flags that change the ABI, but the problem is
		1994	most noticeably displayed with ev_stat and large file support.
		1995
		1996	The solution for this is to lobby your distribution maker to make large
		1997	file interfaces available by default (as e.g. FreeBSD does) and not
		1998	optional. Libev cannot simply switch on large file support because it has
		1999	to exchange stat structures with application programs compiled using the
		2000	default compilation environment.
		2001
		2002	=head3 Inotify and Kqueue
		2003
		2004	When C<inotify (7)> support has been compiled into libev and present at
		2005	runtime, it will be used to speed up change detection where possible. The
		2006	inotify descriptor will be created lazily when the first C<ev_stat>
		2007	watcher is being started.
		2008
		2009	Inotify presence does not change the semantics of C<ev_stat> watchers
		2010	except that changes might be detected earlier, and in some cases, to avoid
		2011	making regular C<stat> calls. Even in the presence of inotify support
		2012	there are many cases where libev has to resort to regular C<stat> polling,
		2013	but as long as kernel 2.6.25 or newer is used (2.6.24 and older have too
		2014	many bugs), the path exists (i.e. stat succeeds), and the path resides on
		2015	a local filesystem (libev currently assumes only ext2/3, jfs, reiserfs and
		2016	xfs are fully working) libev usually gets away without polling.
		2017
		2018	There is no support for kqueue, as apparently it cannot be used to
		2019	implement this functionality, due to the requirement of having a file
		2020	descriptor open on the object at all times, and detecting renames, unlinks
		2021	etc. is difficult.
		2022
		2023	=head3 C<stat ()> is a synchronous operation
		2024
		2025	Libev doesn't normally do any kind of I/O itself, and so is not blocking
		2026	the process. The exception are C<ev_stat> watchers - those call C<stat
		2027	()>, which is a synchronous operation.
		2028
		2029	For local paths, this usually doesn't matter: unless the system is very
		2030	busy or the intervals between stat's are large, a stat call will be fast,
		2031	as the path data is usually in memory already (except when starting the
		2032	watcher).
		2033
		2034	For networked file systems, calling C<stat ()> can block an indefinite
		2035	time due to network issues, and even under good conditions, a stat call
		2036	often takes multiple milliseconds.
		2037
		2038	Therefore, it is best to avoid using C<ev_stat> watchers on networked
		2039	paths, although this is fully supported by libev.
		2040
		2041	=head3 The special problem of stat time resolution
		2042
		2043	The C<stat ()> system call only supports full-second resolution portably,
		2044	and even on systems where the resolution is higher, most file systems
		2045	still only support whole seconds.
		2046
		2047	That means that, if the time is the only thing that changes, you can
		2048	easily miss updates: on the first update, C<ev_stat> detects a change and
		2049	calls your callback, which does something. When there is another update
		2050	within the same second, C<ev_stat> will be unable to detect unless the
		2051	stat data does change in other ways (e.g. file size).
		2052
		2053	The solution to this is to delay acting on a change for slightly more
		2054	than a second (or till slightly after the next full second boundary), using
		2055	a roughly one-second-delay C<ev_timer> (e.g. C<ev_timer_set (w, 0., 1.02);
		2056	ev_timer_again (loop, w)>).
		2057
		2058	The C<.02> offset is added to work around small timing inconsistencies
		2059	of some operating systems (where the second counter of the current time
		2060	might be be delayed. One such system is the Linux kernel, where a call to
		2061	C<gettimeofday> might return a timestamp with a full second later than
		2062	a subsequent C<time> call - if the equivalent of C<time ()> is used to
		2063	update file times then there will be a small window where the kernel uses
		2064	the previous second to update file times but libev might already execute
		2065	the timer callback).
1420		2066
1421	=head3 Watcher-Specific Functions and Data Members	2067	=head3 Watcher-Specific Functions and Data Members
1422		2068
1423	=over 4	2069	=over 4
1424		2070
…		…
1430	C<path>. The C<interval> is a hint on how quickly a change is expected to	2076	C<path>. The C<interval> is a hint on how quickly a change is expected to
1431	be detected and should normally be specified as C<0> to let libev choose	2077	be detected and should normally be specified as C<0> to let libev choose
1432	a suitable value. The memory pointed to by C<path> must point to the same	2078	a suitable value. The memory pointed to by C<path> must point to the same
1433	path for as long as the watcher is active.	2079	path for as long as the watcher is active.
1434		2080
1435	The callback will be receive C<EV_STAT> when a change was detected,	2081	The callback will receive an C<EV_STAT> event when a change was detected,
1436	relative to the attributes at the time the watcher was started (or the	2082	relative to the attributes at the time the watcher was started (or the
1437	last change was detected).	2083	last change was detected).
1438		2084
1439	=item ev_stat_stat (ev_stat *)	2085	=item ev_stat_stat (loop, ev_stat *)
1440		2086
1441	Updates the stat buffer immediately with new values. If you change the	2087	Updates the stat buffer immediately with new values. If you change the
1442	watched path in your callback, you could call this fucntion to avoid	2088	watched path in your callback, you could call this function to avoid
1443	detecting this change (while introducing a race condition). Can also be	2089	detecting this change (while introducing a race condition if you are not
1444	useful simply to find out the new values.	2090	the only one changing the path). Can also be useful simply to find out the
		2091	new values.
1445		2092
1446	=item ev_statdata attr [read-only]	2093	=item ev_statdata attr [read-only]
1447		2094
1448	The most-recently detected attributes of the file. Although the type is of	2095	The most-recently detected attributes of the file. Although the type is
1449	C<ev_statdata>, this is usually the (or one of the) C<struct stat> types	2096	C<ev_statdata>, this is usually the (or one of the) C<struct stat> types
1450	suitable for your system. If the C<st_nlink> member is C<0>, then there	2097	suitable for your system, but you can only rely on the POSIX-standardised
		2098	members to be present. If the C<st_nlink> member is C<0>, then there was
1451	was some error while C<stat>ing the file.	2099	some error while C<stat>ing the file.
1452		2100
1453	=item ev_statdata prev [read-only]	2101	=item ev_statdata prev [read-only]
1454		2102
1455	The previous attributes of the file. The callback gets invoked whenever	2103	The previous attributes of the file. The callback gets invoked whenever
1456	C<prev> != C<attr>.	2104	C<prev> != C<attr>, or, more precisely, one or more of these members
		2105	differ: C<st_dev>, C<st_ino>, C<st_mode>, C<st_nlink>, C<st_uid>,
		2106	C<st_gid>, C<st_rdev>, C<st_size>, C<st_atime>, C<st_mtime>, C<st_ctime>.
1457		2107
1458	=item ev_tstamp interval [read-only]	2108	=item ev_tstamp interval [read-only]
1459		2109
1460	The specified interval.	2110	The specified interval.
1461		2111
1462	=item const char *path [read-only]	2112	=item const char *path [read-only]
1463		2113
1464	The filesystem path that is being watched.	2114	The file system path that is being watched.
1465		2115
1466	=back	2116	=back
1467		2117
		2118	=head3 Examples
		2119
1468	Example: Watch C</etc/passwd> for attribute changes.	2120	Example: Watch C</etc/passwd> for attribute changes.
1469		2121
1470	static void	2122	static void
1471	passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)	2123	passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)
1472	{	2124	{
1473	/* /etc/passwd changed in some way */	2125	/* /etc/passwd changed in some way */
1474	if (w->attr.st_nlink)	2126	if (w->attr.st_nlink)
1475	{	2127	{
1476	printf ("passwd current size %ld\n", (long)w->attr.st_size);	2128	printf ("passwd current size %ld\n", (long)w->attr.st_size);
1477	printf ("passwd current atime %ld\n", (long)w->attr.st_mtime);	2129	printf ("passwd current atime %ld\n", (long)w->attr.st_mtime);
1478	printf ("passwd current mtime %ld\n", (long)w->attr.st_mtime);	2130	printf ("passwd current mtime %ld\n", (long)w->attr.st_mtime);
1479	}	2131	}
1480	else	2132	else
1481	/* you shalt not abuse printf for puts */	2133	/* you shalt not abuse printf for puts */
1482	puts ("wow, /etc/passwd is not there, expect problems. "	2134	puts ("wow, /etc/passwd is not there, expect problems. "
1483	"if this is windows, they already arrived\n");	2135	"if this is windows, they already arrived\n");
1484	}	2136	}
1485		2137
1486	...	2138	...
1487	ev_stat passwd;	2139	ev_stat passwd;
1488		2140
1489	ev_stat_init (&passwd, passwd_cb, "/etc/passwd");	2141	ev_stat_init (&passwd, passwd_cb, "/etc/passwd", 0.);
1490	ev_stat_start (loop, &passwd);	2142	ev_stat_start (loop, &passwd);
		2143
		2144	Example: Like above, but additionally use a one-second delay so we do not
		2145	miss updates (however, frequent updates will delay processing, too, so
		2146	one might do the work both on C<ev_stat> callback invocation I<and> on
		2147	C<ev_timer> callback invocation).
		2148
		2149	static ev_stat passwd;
		2150	static ev_timer timer;
		2151
		2152	static void
		2153	timer_cb (EV_P_ ev_timer *w, int revents)
		2154	{
		2155	ev_timer_stop (EV_A_ w);
		2156
		2157	/* now it's one second after the most recent passwd change */
		2158	}
		2159
		2160	static void
		2161	stat_cb (EV_P_ ev_stat *w, int revents)
		2162	{
		2163	/* reset the one-second timer */
		2164	ev_timer_again (EV_A_ &timer);
		2165	}
		2166
		2167	...
		2168	ev_stat_init (&passwd, stat_cb, "/etc/passwd", 0.);
		2169	ev_stat_start (loop, &passwd);
		2170	ev_timer_init (&timer, timer_cb, 0., 1.02);
1491		2171
1492		2172
1493	=head2 C<ev_idle> - when you've got nothing better to do...	2173	=head2 C<ev_idle> - when you've got nothing better to do...
1494		2174
1495	Idle watchers trigger events when no other events of the same or higher	2175	Idle watchers trigger events when no other events of the same or higher
1496	priority are pending (prepare, check and other idle watchers do not	2176	priority are pending (prepare, check and other idle watchers do not count
1497	count).	2177	as receiving "events").
1498		2178
1499	That is, as long as your process is busy handling sockets or timeouts	2179	That is, as long as your process is busy handling sockets or timeouts
1500	(or even signals, imagine) of the same or higher priority it will not be	2180	(or even signals, imagine) of the same or higher priority it will not be
1501	triggered. But when your process is idle (or only lower-priority watchers	2181	triggered. But when your process is idle (or only lower-priority watchers
1502	are pending), the idle watchers are being called once per event loop	2182	are pending), the idle watchers are being called once per event loop
…		…
1513		2193
1514	=head3 Watcher-Specific Functions and Data Members	2194	=head3 Watcher-Specific Functions and Data Members
1515		2195
1516	=over 4	2196	=over 4
1517		2197
1518	=item ev_idle_init (ev_signal *, callback)	2198	=item ev_idle_init (ev_idle *, callback)
1519		2199
1520	Initialises and configures the idle watcher - it has no parameters of any	2200	Initialises and configures the idle watcher - it has no parameters of any
1521	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,	2201	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,
1522	believe me.	2202	believe me.
1523		2203
1524	=back	2204	=back
1525		2205
		2206	=head3 Examples
		2207
1526	Example: Dynamically allocate an C<ev_idle> watcher, start it, and in the	2208	Example: Dynamically allocate an C<ev_idle> watcher, start it, and in the
1527	callback, free it. Also, use no error checking, as usual.	2209	callback, free it. Also, use no error checking, as usual.
1528		2210
1529	static void	2211	static void
1530	idle_cb (struct ev_loop *loop, struct ev_idle *w, int revents)	2212	idle_cb (struct ev_loop *loop, ev_idle *w, int revents)
1531	{	2213	{
1532	free (w);	2214	free (w);
1533	// now do something you wanted to do when the program has	2215	// now do something you wanted to do when the program has
1534	// no longer asnything immediate to do.	2216	// no longer anything immediate to do.
1535	}	2217	}
1536		2218
1537	struct ev_idle *idle_watcher = malloc (sizeof (struct ev_idle));	2219	ev_idle *idle_watcher = malloc (sizeof (ev_idle));
1538	ev_idle_init (idle_watcher, idle_cb);	2220	ev_idle_init (idle_watcher, idle_cb);
1539	ev_idle_start (loop, idle_cb);	2221	ev_idle_start (loop, idle_cb);
1540		2222
1541		2223
1542	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!	2224	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!
1543		2225
1544	Prepare and check watchers are usually (but not always) used in tandem:	2226	Prepare and check watchers are usually (but not always) used in pairs:
1545	prepare watchers get invoked before the process blocks and check watchers	2227	prepare watchers get invoked before the process blocks and check watchers
1546	afterwards.	2228	afterwards.
1547		2229
1548	You I<must not> call C<ev_loop> or similar functions that enter	2230	You I<must not> call C<ev_loop> or similar functions that enter
1549	the current event loop from either C<ev_prepare> or C<ev_check>	2231	the current event loop from either C<ev_prepare> or C<ev_check>
…		…
1552	those watchers, i.e. the sequence will always be C<ev_prepare>, blocking,	2234	those watchers, i.e. the sequence will always be C<ev_prepare>, blocking,
1553	C<ev_check> so if you have one watcher of each kind they will always be	2235	C<ev_check> so if you have one watcher of each kind they will always be
1554	called in pairs bracketing the blocking call.	2236	called in pairs bracketing the blocking call.
1555		2237
1556	Their main purpose is to integrate other event mechanisms into libev and	2238	Their main purpose is to integrate other event mechanisms into libev and
1557	their use is somewhat advanced. This could be used, for example, to track	2239	their use is somewhat advanced. They could be used, for example, to track
1558	variable changes, implement your own watchers, integrate net-snmp or a	2240	variable changes, implement your own watchers, integrate net-snmp or a
1559	coroutine library and lots more. They are also occasionally useful if	2241	coroutine library and lots more. They are also occasionally useful if
1560	you cache some data and want to flush it before blocking (for example,	2242	you cache some data and want to flush it before blocking (for example,
1561	in X programs you might want to do an C<XFlush ()> in an C<ev_prepare>	2243	in X programs you might want to do an C<XFlush ()> in an C<ev_prepare>
1562	watcher).	2244	watcher).
1563		2245
1564	This is done by examining in each prepare call which file descriptors need	2246	This is done by examining in each prepare call which file descriptors
1565	to be watched by the other library, registering C<ev_io> watchers for	2247	need to be watched by the other library, registering C<ev_io> watchers
1566	them and starting an C<ev_timer> watcher for any timeouts (many libraries	2248	for them and starting an C<ev_timer> watcher for any timeouts (many
1567	provide just this functionality). Then, in the check watcher you check for	2249	libraries provide exactly this functionality). Then, in the check watcher,
1568	any events that occured (by checking the pending status of all watchers	2250	you check for any events that occurred (by checking the pending status
1569	and stopping them) and call back into the library. The I/O and timer	2251	of all watchers and stopping them) and call back into the library. The
1570	callbacks will never actually be called (but must be valid nevertheless,	2252	I/O and timer callbacks will never actually be called (but must be valid
1571	because you never know, you know?).	2253	nevertheless, because you never know, you know?).
1572		2254
1573	As another example, the Perl Coro module uses these hooks to integrate	2255	As another example, the Perl Coro module uses these hooks to integrate
1574	coroutines into libev programs, by yielding to other active coroutines	2256	coroutines into libev programs, by yielding to other active coroutines
1575	during each prepare and only letting the process block if no coroutines	2257	during each prepare and only letting the process block if no coroutines
1576	are ready to run (it's actually more complicated: it only runs coroutines	2258	are ready to run (it's actually more complicated: it only runs coroutines
…		…
1579	loop from blocking if lower-priority coroutines are active, thus mapping	2261	loop from blocking if lower-priority coroutines are active, thus mapping
1580	low-priority coroutines to idle/background tasks).	2262	low-priority coroutines to idle/background tasks).
1581		2263
1582	It is recommended to give C<ev_check> watchers highest (C<EV_MAXPRI>)	2264	It is recommended to give C<ev_check> watchers highest (C<EV_MAXPRI>)
1583	priority, to ensure that they are being run before any other watchers	2265	priority, to ensure that they are being run before any other watchers
		2266	after the poll (this doesn't matter for C<ev_prepare> watchers).
		2267
1584	after the poll. Also, C<ev_check> watchers (and C<ev_prepare> watchers,	2268	Also, C<ev_check> watchers (and C<ev_prepare> watchers, too) should not
1585	too) should not activate ("feed") events into libev. While libev fully	2269	activate ("feed") events into libev. While libev fully supports this, they
1586	supports this, they will be called before other C<ev_check> watchers did	2270	might get executed before other C<ev_check> watchers did their job. As
1587	their job. As C<ev_check> watchers are often used to embed other event	2271	C<ev_check> watchers are often used to embed other (non-libev) event
1588	loops those other event loops might be in an unusable state until their	2272	loops those other event loops might be in an unusable state until their
1589	C<ev_check> watcher ran (always remind yourself to coexist peacefully with	2273	C<ev_check> watcher ran (always remind yourself to coexist peacefully with
1590	others).	2274	others).
1591		2275
1592	=head3 Watcher-Specific Functions and Data Members	2276	=head3 Watcher-Specific Functions and Data Members
…		…
1597		2281
1598	=item ev_check_init (ev_check *, callback)	2282	=item ev_check_init (ev_check *, callback)
1599		2283
1600	Initialises and configures the prepare or check watcher - they have no	2284	Initialises and configures the prepare or check watcher - they have no
1601	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>	2285	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>
1602	macros, but using them is utterly, utterly and completely pointless.	2286	macros, but using them is utterly, utterly, utterly and completely
		2287	pointless.
1603		2288
1604	=back	2289	=back
		2290
		2291	=head3 Examples
1605		2292
1606	There are a number of principal ways to embed other event loops or modules	2293	There are a number of principal ways to embed other event loops or modules
1607	into libev. Here are some ideas on how to include libadns into libev	2294	into libev. Here are some ideas on how to include libadns into libev
1608	(there is a Perl module named C<EV::ADNS> that does this, which you could	2295	(there is a Perl module named C<EV::ADNS> that does this, which you could
1609	use for an actually working example. Another Perl module named C<EV::Glib>	2296	use as a working example. Another Perl module named C<EV::Glib> embeds a
1610	embeds a Glib main context into libev, and finally, C<Glib::EV> embeds EV	2297	Glib main context into libev, and finally, C<Glib::EV> embeds EV into the
1611	into the Glib event loop).	2298	Glib event loop).
1612		2299
1613	Method 1: Add IO watchers and a timeout watcher in a prepare handler,	2300	Method 1: Add IO watchers and a timeout watcher in a prepare handler,
1614	and in a check watcher, destroy them and call into libadns. What follows	2301	and in a check watcher, destroy them and call into libadns. What follows
1615	is pseudo-code only of course. This requires you to either use a low	2302	is pseudo-code only of course. This requires you to either use a low
1616	priority for the check watcher or use C<ev_clear_pending> explicitly, as	2303	priority for the check watcher or use C<ev_clear_pending> explicitly, as
1617	the callbacks for the IO/timeout watchers might not have been called yet.	2304	the callbacks for the IO/timeout watchers might not have been called yet.
1618		2305
1619	static ev_io iow [nfd];	2306	static ev_io iow [nfd];
1620	static ev_timer tw;	2307	static ev_timer tw;
1621		2308
1622	static void	2309	static void
1623	io_cb (ev_loop *loop, ev_io *w, int revents)	2310	io_cb (struct ev_loop *loop, ev_io *w, int revents)
1624	{	2311	{
1625	}	2312	}
1626		2313
1627	// create io watchers for each fd and a timer before blocking	2314	// create io watchers for each fd and a timer before blocking
1628	static void	2315	static void
1629	adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents)	2316	adns_prepare_cb (struct ev_loop *loop, ev_prepare *w, int revents)
1630	{	2317	{
1631	int timeout = 3600000;	2318	int timeout = 3600000;
1632	struct pollfd fds [nfd];	2319	struct pollfd fds [nfd];
1633	// actual code will need to loop here and realloc etc.	2320	// actual code will need to loop here and realloc etc.
1634	adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));	2321	adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));
1635		2322
1636	/* the callback is illegal, but won't be called as we stop during check */	2323	/* the callback is illegal, but won't be called as we stop during check */
1637	ev_timer_init (&tw, 0, timeout * 1e-3);	2324	ev_timer_init (&tw, 0, timeout * 1e-3);
1638	ev_timer_start (loop, &tw);	2325	ev_timer_start (loop, &tw);
1639		2326
1640	// create one ev_io per pollfd	2327	// create one ev_io per pollfd
1641	for (int i = 0; i < nfd; ++i)	2328	for (int i = 0; i < nfd; ++i)
1642	{	2329	{
1643	ev_io_init (iow + i, io_cb, fds [i].fd,	2330	ev_io_init (iow + i, io_cb, fds [i].fd,
1644	((fds [i].events & POLLIN ? EV_READ : 0)	2331	((fds [i].events & POLLIN ? EV_READ : 0)
1645	| (fds [i].events & POLLOUT ? EV_WRITE : 0)));	2332	| (fds [i].events & POLLOUT ? EV_WRITE : 0)));
1646		2333
1647	fds [i].revents = 0;	2334	fds [i].revents = 0;
1648	ev_io_start (loop, iow + i);	2335	ev_io_start (loop, iow + i);
1649	}	2336	}
1650	}	2337	}
1651		2338
1652	// stop all watchers after blocking	2339	// stop all watchers after blocking
1653	static void	2340	static void
1654	adns_check_cb (ev_loop *loop, ev_check *w, int revents)	2341	adns_check_cb (struct ev_loop *loop, ev_check *w, int revents)
1655	{	2342	{
1656	ev_timer_stop (loop, &tw);	2343	ev_timer_stop (loop, &tw);
1657		2344
1658	for (int i = 0; i < nfd; ++i)	2345	for (int i = 0; i < nfd; ++i)
1659	{	2346	{
1660	// set the relevant poll flags	2347	// set the relevant poll flags
1661	// could also call adns_processreadable etc. here	2348	// could also call adns_processreadable etc. here
1662	struct pollfd *fd = fds + i;	2349	struct pollfd *fd = fds + i;
1663	int revents = ev_clear_pending (iow + i);	2350	int revents = ev_clear_pending (iow + i);
1664	if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;	2351	if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;
1665	if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;	2352	if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;
1666		2353
1667	// now stop the watcher	2354	// now stop the watcher
1668	ev_io_stop (loop, iow + i);	2355	ev_io_stop (loop, iow + i);
1669	}	2356	}
1670		2357
1671	adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));	2358	adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));
1672	}	2359	}
1673		2360
1674	Method 2: This would be just like method 1, but you run C<adns_afterpoll>	2361	Method 2: This would be just like method 1, but you run C<adns_afterpoll>
1675	in the prepare watcher and would dispose of the check watcher.	2362	in the prepare watcher and would dispose of the check watcher.
1676		2363
1677	Method 3: If the module to be embedded supports explicit event	2364	Method 3: If the module to be embedded supports explicit event
1678	notification (adns does), you can also make use of the actual watcher	2365	notification (libadns does), you can also make use of the actual watcher
1679	callbacks, and only destroy/create the watchers in the prepare watcher.	2366	callbacks, and only destroy/create the watchers in the prepare watcher.
1680		2367
1681	static void	2368	static void
1682	timer_cb (EV_P_ ev_timer *w, int revents)	2369	timer_cb (EV_P_ ev_timer *w, int revents)
1683	{	2370	{
1684	adns_state ads = (adns_state)w->data;	2371	adns_state ads = (adns_state)w->data;
1685	update_now (EV_A);	2372	update_now (EV_A);
1686		2373
1687	adns_processtimeouts (ads, &tv_now);	2374	adns_processtimeouts (ads, &tv_now);
1688	}	2375	}
1689		2376
1690	static void	2377	static void
1691	io_cb (EV_P_ ev_io *w, int revents)	2378	io_cb (EV_P_ ev_io *w, int revents)
1692	{	2379	{
1693	adns_state ads = (adns_state)w->data;	2380	adns_state ads = (adns_state)w->data;
1694	update_now (EV_A);	2381	update_now (EV_A);
1695		2382
1696	if (revents & EV_READ ) adns_processreadable (ads, w->fd, &tv_now);	2383	if (revents & EV_READ ) adns_processreadable (ads, w->fd, &tv_now);
1697	if (revents & EV_WRITE) adns_processwriteable (ads, w->fd, &tv_now);	2384	if (revents & EV_WRITE) adns_processwriteable (ads, w->fd, &tv_now);
1698	}	2385	}
1699		2386
1700	// do not ever call adns_afterpoll	2387	// do not ever call adns_afterpoll
1701		2388
1702	Method 4: Do not use a prepare or check watcher because the module you	2389	Method 4: Do not use a prepare or check watcher because the module you
1703	want to embed is too inflexible to support it. Instead, youc na override	2390	want to embed is not flexible enough to support it. Instead, you can
1704	their poll function. The drawback with this solution is that the main	2391	override their poll function. The drawback with this solution is that the
1705	loop is now no longer controllable by EV. The C<Glib::EV> module does	2392	main loop is now no longer controllable by EV. The C<Glib::EV> module uses
1706	this.	2393	this approach, effectively embedding EV as a client into the horrible
		2394	libglib event loop.
1707		2395
1708	static gint	2396	static gint
1709	event_poll_func (GPollFD *fds, guint nfds, gint timeout)	2397	event_poll_func (GPollFD *fds, guint nfds, gint timeout)
1710	{	2398	{
1711	int got_events = 0;	2399	int got_events = 0;
1712		2400
1713	for (n = 0; n < nfds; ++n)	2401	for (n = 0; n < nfds; ++n)
1714	// create/start io watcher that sets the relevant bits in fds[n] and increment got_events	2402	// create/start io watcher that sets the relevant bits in fds[n] and increment got_events
1715		2403
1716	if (timeout >= 0)	2404	if (timeout >= 0)
1717	// create/start timer	2405	// create/start timer
1718		2406
1719	// poll	2407	// poll
1720	ev_loop (EV_A_ 0);	2408	ev_loop (EV_A_ 0);
1721		2409
1722	// stop timer again	2410	// stop timer again
1723	if (timeout >= 0)	2411	if (timeout >= 0)
1724	ev_timer_stop (EV_A_ &to);	2412	ev_timer_stop (EV_A_ &to);
1725		2413
1726	// stop io watchers again - their callbacks should have set	2414	// stop io watchers again - their callbacks should have set
1727	for (n = 0; n < nfds; ++n)	2415	for (n = 0; n < nfds; ++n)
1728	ev_io_stop (EV_A_ iow [n]);	2416	ev_io_stop (EV_A_ iow [n]);
1729		2417
1730	return got_events;	2418	return got_events;
1731	}	2419	}
1732		2420
1733		2421
1734	=head2 C<ev_embed> - when one backend isn't enough...	2422	=head2 C<ev_embed> - when one backend isn't enough...
1735		2423
1736	This is a rather advanced watcher type that lets you embed one event loop	2424	This is a rather advanced watcher type that lets you embed one event loop
1737	into another (currently only C<ev_io> events are supported in the embedded	2425	into another (currently only C<ev_io> events are supported in the embedded
1738	loop, other types of watchers might be handled in a delayed or incorrect	2426	loop, other types of watchers might be handled in a delayed or incorrect
1739	fashion and must not be used). (See portability notes, below).	2427	fashion and must not be used).
1740		2428
1741	There are primarily two reasons you would want that: work around bugs and	2429	There are primarily two reasons you would want that: work around bugs and
1742	prioritise I/O.	2430	prioritise I/O.
1743		2431
1744	As an example for a bug workaround, the kqueue backend might only support	2432	As an example for a bug workaround, the kqueue backend might only support
1745	sockets on some platform, so it is unusable as generic backend, but you	2433	sockets on some platform, so it is unusable as generic backend, but you
1746	still want to make use of it because you have many sockets and it scales	2434	still want to make use of it because you have many sockets and it scales
1747	so nicely. In this case, you would create a kqueue-based loop and embed it	2435	so nicely. In this case, you would create a kqueue-based loop and embed
1748	into your default loop (which might use e.g. poll). Overall operation will	2436	it into your default loop (which might use e.g. poll). Overall operation
1749	be a bit slower because first libev has to poll and then call kevent, but	2437	will be a bit slower because first libev has to call C<poll> and then
1750	at least you can use both at what they are best.	2438	C<kevent>, but at least you can use both mechanisms for what they are
		2439	best: C<kqueue> for scalable sockets and C<poll> if you want it to work :)
1751		2440
1752	As for prioritising I/O: rarely you have the case where some fds have	2441	As for prioritising I/O: under rare circumstances you have the case where
1753	to be watched and handled very quickly (with low latency), and even	2442	some fds have to be watched and handled very quickly (with low latency),
1754	priorities and idle watchers might have too much overhead. In this case	2443	and even priorities and idle watchers might have too much overhead. In
1755	you would put all the high priority stuff in one loop and all the rest in	2444	this case you would put all the high priority stuff in one loop and all
1756	a second one, and embed the second one in the first.	2445	the rest in a second one, and embed the second one in the first.
1757		2446
1758	As long as the watcher is active, the callback will be invoked every time	2447	As long as the watcher is active, the callback will be invoked every
1759	there might be events pending in the embedded loop. The callback must then	2448	time there might be events pending in the embedded loop. The callback
1760	call C<ev_embed_sweep (mainloop, watcher)> to make a single sweep and invoke	2449	must then call C<ev_embed_sweep (mainloop, watcher)> to make a single
1761	their callbacks (you could also start an idle watcher to give the embedded	2450	sweep and invoke their callbacks (the callback doesn't need to invoke the
1762	loop strictly lower priority for example). You can also set the callback	2451	C<ev_embed_sweep> function directly, it could also start an idle watcher
1763	to C<0>, in which case the embed watcher will automatically execute the	2452	to give the embedded loop strictly lower priority for example).
1764	embedded loop sweep.
1765		2453
1766	As long as the watcher is started it will automatically handle events. The	2454	You can also set the callback to C<0>, in which case the embed watcher
1767	callback will be invoked whenever some events have been handled. You can	2455	will automatically execute the embedded loop sweep whenever necessary.
1768	set the callback to C<0> to avoid having to specify one if you are not
1769	interested in that.
1770		2456
1771	Also, there have not currently been made special provisions for forking:	2457	Fork detection will be handled transparently while the C<ev_embed> watcher
1772	when you fork, you not only have to call C<ev_loop_fork> on both loops,	2458	is active, i.e., the embedded loop will automatically be forked when the
1773	but you will also have to stop and restart any C<ev_embed> watchers	2459	embedding loop forks. In other cases, the user is responsible for calling
1774	yourself.	2460	C<ev_loop_fork> on the embedded loop.
1775		2461
1776	Unfortunately, not all backends are embeddable, only the ones returned by	2462	Unfortunately, not all backends are embeddable: only the ones returned by
1777	C<ev_embeddable_backends> are, which, unfortunately, does not include any	2463	C<ev_embeddable_backends> are, which, unfortunately, does not include any
1778	portable one.	2464	portable one.
1779		2465
1780	So when you want to use this feature you will always have to be prepared	2466	So when you want to use this feature you will always have to be prepared
1781	that you cannot get an embeddable loop. The recommended way to get around	2467	that you cannot get an embeddable loop. The recommended way to get around
1782	this is to have a separate variables for your embeddable loop, try to	2468	this is to have a separate variables for your embeddable loop, try to
1783	create it, and if that fails, use the normal loop for everything:	2469	create it, and if that fails, use the normal loop for everything.
1784		2470
1785	struct ev_loop *loop_hi = ev_default_init (0);	2471	=head3 C<ev_embed> and fork
1786	struct ev_loop *loop_lo = 0;
1787	struct ev_embed embed;
1788
1789	// see if there is a chance of getting one that works
1790	// (remember that a flags value of 0 means autodetection)
1791	loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
1792	? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
1793	: 0;
1794		2472
1795	// if we got one, then embed it, otherwise default to loop_hi	2473	While the C<ev_embed> watcher is running, forks in the embedding loop will
1796	if (loop_lo)	2474	automatically be applied to the embedded loop as well, so no special
1797	{	2475	fork handling is required in that case. When the watcher is not running,
1798	ev_embed_init (&embed, 0, loop_lo);	2476	however, it is still the task of the libev user to call C<ev_loop_fork ()>
1799	ev_embed_start (loop_hi, &embed);	2477	as applicable.
1800	}
1801	else
1802	loop_lo = loop_hi;
1803
1804	=head2 Portability notes
1805
1806	Kqueue is nominally embeddable, but this is broken on all BSDs that I
1807	tried, in various ways. Usually the embedded event loop will simply never
1808	receive events, sometimes it will only trigger a few times, sometimes in a
1809	loop. Epoll is also nominally embeddable, but many Linux kernel versions
1810	will always eport the epoll fd as ready, even when no events are pending.
1811
1812	While libev allows embedding these backends (they are contained in
1813	C<ev_embeddable_backends ()>), take extreme care that it will actually
1814	work.
1815
1816	When in doubt, create a dynamic event loop forced to use sockets (this
1817	usually works) and possibly another thread and a pipe or so to report to
1818	your main event loop.
1819		2478
1820	=head3 Watcher-Specific Functions and Data Members	2479	=head3 Watcher-Specific Functions and Data Members
1821		2480
1822	=over 4	2481	=over 4
1823		2482
…		…
1827		2486
1828	Configures the watcher to embed the given loop, which must be	2487	Configures the watcher to embed the given loop, which must be
1829	embeddable. If the callback is C<0>, then C<ev_embed_sweep> will be	2488	embeddable. If the callback is C<0>, then C<ev_embed_sweep> will be
1830	invoked automatically, otherwise it is the responsibility of the callback	2489	invoked automatically, otherwise it is the responsibility of the callback
1831	to invoke it (it will continue to be called until the sweep has been done,	2490	to invoke it (it will continue to be called until the sweep has been done,
1832	if you do not want thta, you need to temporarily stop the embed watcher).	2491	if you do not want that, you need to temporarily stop the embed watcher).
1833		2492
1834	=item ev_embed_sweep (loop, ev_embed *)	2493	=item ev_embed_sweep (loop, ev_embed *)
1835		2494
1836	Make a single, non-blocking sweep over the embedded loop. This works	2495	Make a single, non-blocking sweep over the embedded loop. This works
1837	similarly to C<ev_loop (embedded_loop, EVLOOP_NONBLOCK)>, but in the most	2496	similarly to C<ev_loop (embedded_loop, EVLOOP_NONBLOCK)>, but in the most
1838	apropriate way for embedded loops.	2497	appropriate way for embedded loops.
1839		2498
1840	=item struct ev_loop *other [read-only]	2499	=item struct ev_loop *other [read-only]
1841		2500
1842	The embedded event loop.	2501	The embedded event loop.
1843		2502
1844	=back	2503	=back
		2504
		2505	=head3 Examples
		2506
		2507	Example: Try to get an embeddable event loop and embed it into the default
		2508	event loop. If that is not possible, use the default loop. The default
		2509	loop is stored in C<loop_hi>, while the embeddable loop is stored in
		2510	C<loop_lo> (which is C<loop_hi> in the case no embeddable loop can be
		2511	used).
		2512
		2513	struct ev_loop *loop_hi = ev_default_init (0);
		2514	struct ev_loop *loop_lo = 0;
		2515	ev_embed embed;
		2516
		2517	// see if there is a chance of getting one that works
		2518	// (remember that a flags value of 0 means autodetection)
		2519	loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
		2520	? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
		2521	: 0;
		2522
		2523	// if we got one, then embed it, otherwise default to loop_hi
		2524	if (loop_lo)
		2525	{
		2526	ev_embed_init (&embed, 0, loop_lo);
		2527	ev_embed_start (loop_hi, &embed);
		2528	}
		2529	else
		2530	loop_lo = loop_hi;
		2531
		2532	Example: Check if kqueue is available but not recommended and create
		2533	a kqueue backend for use with sockets (which usually work with any
		2534	kqueue implementation). Store the kqueue/socket-only event loop in
		2535	C<loop_socket>. (One might optionally use C<EVFLAG_NOENV>, too).
		2536
		2537	struct ev_loop *loop = ev_default_init (0);
		2538	struct ev_loop *loop_socket = 0;
		2539	ev_embed embed;
		2540
		2541	if (ev_supported_backends () & ~ev_recommended_backends () & EVBACKEND_KQUEUE)
		2542	if ((loop_socket = ev_loop_new (EVBACKEND_KQUEUE))
		2543	{
		2544	ev_embed_init (&embed, 0, loop_socket);
		2545	ev_embed_start (loop, &embed);
		2546	}
		2547
		2548	if (!loop_socket)
		2549	loop_socket = loop;
		2550
		2551	// now use loop_socket for all sockets, and loop for everything else
1845		2552
1846		2553
1847	=head2 C<ev_fork> - the audacity to resume the event loop after a fork	2554	=head2 C<ev_fork> - the audacity to resume the event loop after a fork
1848		2555
1849	Fork watchers are called when a C<fork ()> was detected (usually because	2556	Fork watchers are called when a C<fork ()> was detected (usually because
…		…
1865	believe me.	2572	believe me.
1866		2573
1867	=back	2574	=back
1868		2575
1869		2576
		2577	=head2 C<ev_async> - how to wake up another event loop
		2578
		2579	In general, you cannot use an C<ev_loop> from multiple threads or other
		2580	asynchronous sources such as signal handlers (as opposed to multiple event
		2581	loops - those are of course safe to use in different threads).
		2582
		2583	Sometimes, however, you need to wake up another event loop you do not
		2584	control, for example because it belongs to another thread. This is what
		2585	C<ev_async> watchers do: as long as the C<ev_async> watcher is active, you
		2586	can signal it by calling C<ev_async_send>, which is thread- and signal
		2587	safe.
		2588
		2589	This functionality is very similar to C<ev_signal> watchers, as signals,
		2590	too, are asynchronous in nature, and signals, too, will be compressed
		2591	(i.e. the number of callback invocations may be less than the number of
		2592	C<ev_async_sent> calls).
		2593
		2594	Unlike C<ev_signal> watchers, C<ev_async> works with any event loop, not
		2595	just the default loop.
		2596
		2597	=head3 Queueing
		2598
		2599	C<ev_async> does not support queueing of data in any way. The reason
		2600	is that the author does not know of a simple (or any) algorithm for a
		2601	multiple-writer-single-reader queue that works in all cases and doesn't
		2602	need elaborate support such as pthreads.
		2603
		2604	That means that if you want to queue data, you have to provide your own
		2605	queue. But at least I can tell you how to implement locking around your
		2606	queue:
		2607
		2608	=over 4
		2609
		2610	=item queueing from a signal handler context
		2611
		2612	To implement race-free queueing, you simply add to the queue in the signal
		2613	handler but you block the signal handler in the watcher callback. Here is
		2614	an example that does that for some fictitious SIGUSR1 handler:
		2615
		2616	static ev_async mysig;
		2617
		2618	static void
		2619	sigusr1_handler (void)
		2620	{
		2621	sometype data;
		2622
		2623	// no locking etc.
		2624	queue_put (data);
		2625	ev_async_send (EV_DEFAULT_ &mysig);
		2626	}
		2627
		2628	static void
		2629	mysig_cb (EV_P_ ev_async *w, int revents)
		2630	{
		2631	sometype data;
		2632	sigset_t block, prev;
		2633
		2634	sigemptyset (&block);
		2635	sigaddset (&block, SIGUSR1);
		2636	sigprocmask (SIG_BLOCK, &block, &prev);
		2637
		2638	while (queue_get (&data))
		2639	process (data);
		2640
		2641	if (sigismember (&prev, SIGUSR1)
		2642	sigprocmask (SIG_UNBLOCK, &block, 0);
		2643	}
		2644
		2645	(Note: pthreads in theory requires you to use C<pthread_setmask>
		2646	instead of C<sigprocmask> when you use threads, but libev doesn't do it
		2647	either...).
		2648
		2649	=item queueing from a thread context
		2650
		2651	The strategy for threads is different, as you cannot (easily) block
		2652	threads but you can easily preempt them, so to queue safely you need to
		2653	employ a traditional mutex lock, such as in this pthread example:
		2654
		2655	static ev_async mysig;
		2656	static pthread_mutex_t mymutex = PTHREAD_MUTEX_INITIALIZER;
		2657
		2658	static void
		2659	otherthread (void)
		2660	{
		2661	// only need to lock the actual queueing operation
		2662	pthread_mutex_lock (&mymutex);
		2663	queue_put (data);
		2664	pthread_mutex_unlock (&mymutex);
		2665
		2666	ev_async_send (EV_DEFAULT_ &mysig);
		2667	}
		2668
		2669	static void
		2670	mysig_cb (EV_P_ ev_async *w, int revents)
		2671	{
		2672	pthread_mutex_lock (&mymutex);
		2673
		2674	while (queue_get (&data))
		2675	process (data);
		2676
		2677	pthread_mutex_unlock (&mymutex);
		2678	}
		2679
		2680	=back
		2681
		2682
		2683	=head3 Watcher-Specific Functions and Data Members
		2684
		2685	=over 4
		2686
		2687	=item ev_async_init (ev_async *, callback)
		2688
		2689	Initialises and configures the async watcher - it has no parameters of any
		2690	kind. There is a C<ev_async_set> macro, but using it is utterly pointless,
		2691	trust me.
		2692
		2693	=item ev_async_send (loop, ev_async *)
		2694
		2695	Sends/signals/activates the given C<ev_async> watcher, that is, feeds
		2696	an C<EV_ASYNC> event on the watcher into the event loop. Unlike
		2697	C<ev_feed_event>, this call is safe to do from other threads, signal or
		2698	similar contexts (see the discussion of C<EV_ATOMIC_T> in the embedding
		2699	section below on what exactly this means).
		2700
		2701	Note that, as with other watchers in libev, multiple events might get
		2702	compressed into a single callback invocation (another way to look at this
		2703	is that C<ev_async> watchers are level-triggered, set on C<ev_async_send>,
		2704	reset when the event loop detects that).
		2705
		2706	This call incurs the overhead of a system call only once per event loop
		2707	iteration, so while the overhead might be noticeable, it doesn't apply to
		2708	repeated calls to C<ev_async_send> for the same event loop.
		2709
		2710	=item bool = ev_async_pending (ev_async *)
		2711
		2712	Returns a non-zero value when C<ev_async_send> has been called on the
		2713	watcher but the event has not yet been processed (or even noted) by the
		2714	event loop.
		2715
		2716	C<ev_async_send> sets a flag in the watcher and wakes up the loop. When
		2717	the loop iterates next and checks for the watcher to have become active,
		2718	it will reset the flag again. C<ev_async_pending> can be used to very
		2719	quickly check whether invoking the loop might be a good idea.
		2720
		2721	Not that this does I<not> check whether the watcher itself is pending,
		2722	only whether it has been requested to make this watcher pending: there
		2723	is a time window between the event loop checking and resetting the async
		2724	notification, and the callback being invoked.
		2725
		2726	=back
		2727
		2728
1870	=head1 OTHER FUNCTIONS	2729	=head1 OTHER FUNCTIONS
1871		2730
1872	There are some other functions of possible interest. Described. Here. Now.	2731	There are some other functions of possible interest. Described. Here. Now.
1873		2732
1874	=over 4	2733	=over 4
1875		2734
1876	=item ev_once (loop, int fd, int events, ev_tstamp timeout, callback)	2735	=item ev_once (loop, int fd, int events, ev_tstamp timeout, callback)
1877		2736
1878	This function combines a simple timer and an I/O watcher, calls your	2737	This function combines a simple timer and an I/O watcher, calls your
1879	callback on whichever event happens first and automatically stop both	2738	callback on whichever event happens first and automatically stops both
1880	watchers. This is useful if you want to wait for a single event on an fd	2739	watchers. This is useful if you want to wait for a single event on an fd
1881	or timeout without having to allocate/configure/start/stop/free one or	2740	or timeout without having to allocate/configure/start/stop/free one or
1882	more watchers yourself.	2741	more watchers yourself.
1883		2742
1884	If C<fd> is less than 0, then no I/O watcher will be started and events	2743	If C<fd> is less than 0, then no I/O watcher will be started and the
1885	is being ignored. Otherwise, an C<ev_io> watcher for the given C<fd> and	2744	C<events> argument is being ignored. Otherwise, an C<ev_io> watcher for
1886	C<events> set will be craeted and started.	2745	the given C<fd> and C<events> set will be created and started.
1887		2746
1888	If C<timeout> is less than 0, then no timeout watcher will be	2747	If C<timeout> is less than 0, then no timeout watcher will be
1889	started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and	2748	started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and
1890	repeat = 0) will be started. While C<0> is a valid timeout, it is of	2749	repeat = 0) will be started. C<0> is a valid timeout.
1891	dubious value.
1892		2750
1893	The callback has the type C<void (*cb)(int revents, void *arg)> and gets	2751	The callback has the type C<void (*cb)(int revents, void *arg)> and gets
1894	passed an C<revents> set like normal event callbacks (a combination of	2752	passed an C<revents> set like normal event callbacks (a combination of
1895	C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMEOUT>) and the C<arg>	2753	C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMEOUT>) and the C<arg>
1896	value passed to C<ev_once>:	2754	value passed to C<ev_once>. Note that it is possible to receive I<both>
		2755	a timeout and an io event at the same time - you probably should give io
		2756	events precedence.
1897		2757
		2758	Example: wait up to ten seconds for data to appear on STDIN_FILENO.
		2759
1898	static void stdin_ready (int revents, void *arg)	2760	static void stdin_ready (int revents, void *arg)
1899	{	2761	{
1900	if (revents & EV_TIMEOUT)
1901	/* doh, nothing entered */;
1902	else if (revents & EV_READ)	2762	if (revents & EV_READ)
1903	/* stdin might have data for us, joy! */;	2763	/* stdin might have data for us, joy! */;
		2764	else if (revents & EV_TIMEOUT)
		2765	/* doh, nothing entered */;
1904	}	2766	}
1905		2767
1906	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);	2768	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
1907		2769
1908	=item ev_feed_event (ev_loop *, watcher *, int revents)	2770	=item ev_feed_event (struct ev_loop *, watcher *, int revents)
1909		2771
1910	Feeds the given event set into the event loop, as if the specified event	2772	Feeds the given event set into the event loop, as if the specified event
1911	had happened for the specified watcher (which must be a pointer to an	2773	had happened for the specified watcher (which must be a pointer to an
1912	initialised but not necessarily started event watcher).	2774	initialised but not necessarily started event watcher).
1913		2775
1914	=item ev_feed_fd_event (ev_loop *, int fd, int revents)	2776	=item ev_feed_fd_event (struct ev_loop *, int fd, int revents)
1915		2777
1916	Feed an event on the given fd, as if a file descriptor backend detected	2778	Feed an event on the given fd, as if a file descriptor backend detected
1917	the given events it.	2779	the given events it.
1918		2780
1919	=item ev_feed_signal_event (ev_loop *loop, int signum)	2781	=item ev_feed_signal_event (struct ev_loop *loop, int signum)
1920		2782
1921	Feed an event as if the given signal occured (C<loop> must be the default	2783	Feed an event as if the given signal occurred (C<loop> must be the default
1922	loop!).	2784	loop!).
1923		2785
1924	=back	2786	=back
1925		2787
1926		2788
…		…
1942		2804
1943	=item * Priorities are not currently supported. Initialising priorities	2805	=item * Priorities are not currently supported. Initialising priorities
1944	will fail and all watchers will have the same priority, even though there	2806	will fail and all watchers will have the same priority, even though there
1945	is an ev_pri field.	2807	is an ev_pri field.
1946		2808
		2809	=item * In libevent, the last base created gets the signals, in libev, the
		2810	first base created (== the default loop) gets the signals.
		2811
1947	=item * Other members are not supported.	2812	=item * Other members are not supported.
1948		2813
1949	=item * The libev emulation is I<not> ABI compatible to libevent, you need	2814	=item * The libev emulation is I<not> ABI compatible to libevent, you need
1950	to use the libev header file and library.	2815	to use the libev header file and library.
1951		2816
1952	=back	2817	=back
1953		2818
1954	=head1 C++ SUPPORT	2819	=head1 C++ SUPPORT
1955		2820
1956	Libev comes with some simplistic wrapper classes for C++ that mainly allow	2821	Libev comes with some simplistic wrapper classes for C++ that mainly allow
1957	you to use some convinience methods to start/stop watchers and also change	2822	you to use some convenience methods to start/stop watchers and also change
1958	the callback model to a model using method callbacks on objects.	2823	the callback model to a model using method callbacks on objects.
1959		2824
1960	To use it,	2825	To use it,
1961		2826
1962	#include <ev++.h>	2827	#include <ev++.h>
1963		2828
1964	This automatically includes F<ev.h> and puts all of its definitions (many	2829	This automatically includes F<ev.h> and puts all of its definitions (many
1965	of them macros) into the global namespace. All C++ specific things are	2830	of them macros) into the global namespace. All C++ specific things are
1966	put into the C<ev> namespace. It should support all the same embedding	2831	put into the C<ev> namespace. It should support all the same embedding
1967	options as F<ev.h>, most notably C<EV_MULTIPLICITY>.	2832	options as F<ev.h>, most notably C<EV_MULTIPLICITY>.
…		…
2034	your compiler is good :), then the method will be fully inlined into the	2899	your compiler is good :), then the method will be fully inlined into the
2035	thunking function, making it as fast as a direct C callback.	2900	thunking function, making it as fast as a direct C callback.
2036		2901
2037	Example: simple class declaration and watcher initialisation	2902	Example: simple class declaration and watcher initialisation
2038		2903
2039	struct myclass	2904	struct myclass
2040	{	2905	{
2041	void io_cb (ev::io &w, int revents) { }	2906	void io_cb (ev::io &w, int revents) { }
2042	}	2907	}
2043		2908
2044	myclass obj;	2909	myclass obj;
2045	ev::io iow;	2910	ev::io iow;
2046	iow.set <myclass, &myclass::io_cb> (&obj);	2911	iow.set <myclass, &myclass::io_cb> (&obj);
		2912
		2913	=item w->set (object *)
		2914
		2915	This is an B<experimental> feature that might go away in a future version.
		2916
		2917	This is a variation of a method callback - leaving out the method to call
		2918	will default the method to C<operator ()>, which makes it possible to use
		2919	functor objects without having to manually specify the C<operator ()> all
		2920	the time. Incidentally, you can then also leave out the template argument
		2921	list.
		2922
		2923	The C<operator ()> method prototype must be C<void operator ()(watcher &w,
		2924	int revents)>.
		2925
		2926	See the method-C<set> above for more details.
		2927
		2928	Example: use a functor object as callback.
		2929
		2930	struct myfunctor
		2931	{
		2932	void operator() (ev::io &w, int revents)
		2933	{
		2934	...
		2935	}
		2936	}
		2937
		2938	myfunctor f;
		2939
		2940	ev::io w;
		2941	w.set (&f);
2047		2942
2048	=item w->set<function> (void *data = 0)	2943	=item w->set<function> (void *data = 0)
2049		2944
2050	Also sets a callback, but uses a static method or plain function as	2945	Also sets a callback, but uses a static method or plain function as
2051	callback. The optional C<data> argument will be stored in the watcher's	2946	callback. The optional C<data> argument will be stored in the watcher's
…		…
2053		2948
2054	The prototype of the C<function> must be C<void (*)(ev::TYPE &w, int)>.	2949	The prototype of the C<function> must be C<void (*)(ev::TYPE &w, int)>.
2055		2950
2056	See the method-C<set> above for more details.	2951	See the method-C<set> above for more details.
2057		2952
2058	Example:	2953	Example: Use a plain function as callback.
2059		2954
2060	static void io_cb (ev::io &w, int revents) { }	2955	static void io_cb (ev::io &w, int revents) { }
2061	iow.set <io_cb> ();	2956	iow.set <io_cb> ();
2062		2957
2063	=item w->set (struct ev_loop *)	2958	=item w->set (struct ev_loop *)
2064		2959
2065	Associates a different C<struct ev_loop> with this watcher. You can only	2960	Associates a different C<struct ev_loop> with this watcher. You can only
2066	do this when the watcher is inactive (and not pending either).	2961	do this when the watcher is inactive (and not pending either).
2067		2962
2068	=item w->set ([args])	2963	=item w->set ([arguments])
2069		2964
2070	Basically the same as C<ev_TYPE_set>, with the same args. Must be	2965	Basically the same as C<ev_TYPE_set>, with the same arguments. Must be
2071	called at least once. Unlike the C counterpart, an active watcher gets	2966	called at least once. Unlike the C counterpart, an active watcher gets
2072	automatically stopped and restarted when reconfiguring it with this	2967	automatically stopped and restarted when reconfiguring it with this
2073	method.	2968	method.
2074		2969
2075	=item w->start ()	2970	=item w->start ()
…		…
2099	=back	2994	=back
2100		2995
2101	Example: Define a class with an IO and idle watcher, start one of them in	2996	Example: Define a class with an IO and idle watcher, start one of them in
2102	the constructor.	2997	the constructor.
2103		2998
2104	class myclass	2999	class myclass
2105	{	3000	{
2106	ev_io io; void io_cb (ev::io &w, int revents);	3001	ev::io io ; void io_cb (ev::io &w, int revents);
2107	ev_idle idle void idle_cb (ev::idle &w, int revents);	3002	ev::idle idle; void idle_cb (ev::idle &w, int revents);
2108		3003
2109	myclass ();	3004	myclass (int fd)
2110	}	3005	{
2111
2112	myclass::myclass (int fd)
2113	{
2114	io .set <myclass, &myclass::io_cb > (this);	3006	io .set <myclass, &myclass::io_cb > (this);
2115	idle.set <myclass, &myclass::idle_cb> (this);	3007	idle.set <myclass, &myclass::idle_cb> (this);
2116		3008
2117	io.start (fd, ev::READ);	3009	io.start (fd, ev::READ);
		3010	}
2118	}	3011	};
		3012
		3013
		3014	=head1 OTHER LANGUAGE BINDINGS
		3015
		3016	Libev does not offer other language bindings itself, but bindings for a
		3017	number of languages exist in the form of third-party packages. If you know
		3018	any interesting language binding in addition to the ones listed here, drop
		3019	me a note.
		3020
		3021	=over 4
		3022
		3023	=item Perl
		3024
		3025	The EV module implements the full libev API and is actually used to test
		3026	libev. EV is developed together with libev. Apart from the EV core module,
		3027	there are additional modules that implement libev-compatible interfaces
		3028	to C<libadns> (C<EV::ADNS>, but C<AnyEvent::DNS> is preferred nowadays),
		3029	C<Net::SNMP> (C<Net::SNMP::EV>) and the C<libglib> event core (C<Glib::EV>
		3030	and C<EV::Glib>).
		3031
		3032	It can be found and installed via CPAN, its homepage is at
		3033	L<http://software.schmorp.de/pkg/EV>.
		3034
		3035	=item Python
		3036
		3037	Python bindings can be found at L<http://code.google.com/p/pyev/>. It
		3038	seems to be quite complete and well-documented. Note, however, that the
		3039	patch they require for libev is outright dangerous as it breaks the ABI
		3040	for everybody else, and therefore, should never be applied in an installed
		3041	libev (if python requires an incompatible ABI then it needs to embed
		3042	libev).
		3043
		3044	=item Ruby
		3045
		3046	Tony Arcieri has written a ruby extension that offers access to a subset
		3047	of the libev API and adds file handle abstractions, asynchronous DNS and
		3048	more on top of it. It can be found via gem servers. Its homepage is at
		3049	L<http://rev.rubyforge.org/>.
		3050
		3051	Roger Pack reports that using the link order C<-lws2_32 -lmsvcrt-ruby-190>
		3052	makes rev work even on mingw.
		3053
		3054	=item D
		3055
		3056	Leandro Lucarella has written a D language binding (F<ev.d>) for libev, to
		3057	be found at L<http://proj.llucax.com.ar/wiki/evd>.
		3058
		3059	=item Ocaml
		3060
		3061	Erkki Seppala has written Ocaml bindings for libev, to be found at
		3062	L<http://modeemi.cs.tut.fi/~flux/software/ocaml-ev/>.
		3063
		3064	=back
2119		3065
2120		3066
2121	=head1 MACRO MAGIC	3067	=head1 MACRO MAGIC
2122		3068
2123	Libev can be compiled with a variety of options, the most fundamantal	3069	Libev can be compiled with a variety of options, the most fundamental
2124	of which is C<EV_MULTIPLICITY>. This option determines whether (most)	3070	of which is C<EV_MULTIPLICITY>. This option determines whether (most)
2125	functions and callbacks have an initial C<struct ev_loop *> argument.	3071	functions and callbacks have an initial C<struct ev_loop *> argument.
2126		3072
2127	To make it easier to write programs that cope with either variant, the	3073	To make it easier to write programs that cope with either variant, the
2128	following macros are defined:	3074	following macros are defined:
…		…
2133		3079
2134	This provides the loop I<argument> for functions, if one is required ("ev	3080	This provides the loop I<argument> for functions, if one is required ("ev
2135	loop argument"). The C<EV_A> form is used when this is the sole argument,	3081	loop argument"). The C<EV_A> form is used when this is the sole argument,
2136	C<EV_A_> is used when other arguments are following. Example:	3082	C<EV_A_> is used when other arguments are following. Example:
2137		3083
2138	ev_unref (EV_A);	3084	ev_unref (EV_A);
2139	ev_timer_add (EV_A_ watcher);	3085	ev_timer_add (EV_A_ watcher);
2140	ev_loop (EV_A_ 0);	3086	ev_loop (EV_A_ 0);
2141		3087
2142	It assumes the variable C<loop> of type C<struct ev_loop *> is in scope,	3088	It assumes the variable C<loop> of type C<struct ev_loop *> is in scope,
2143	which is often provided by the following macro.	3089	which is often provided by the following macro.
2144		3090
2145	=item C<EV_P>, C<EV_P_>	3091	=item C<EV_P>, C<EV_P_>
2146		3092
2147	This provides the loop I<parameter> for functions, if one is required ("ev	3093	This provides the loop I<parameter> for functions, if one is required ("ev
2148	loop parameter"). The C<EV_P> form is used when this is the sole parameter,	3094	loop parameter"). The C<EV_P> form is used when this is the sole parameter,
2149	C<EV_P_> is used when other parameters are following. Example:	3095	C<EV_P_> is used when other parameters are following. Example:
2150		3096
2151	// this is how ev_unref is being declared	3097	// this is how ev_unref is being declared
2152	static void ev_unref (EV_P);	3098	static void ev_unref (EV_P);
2153		3099
2154	// this is how you can declare your typical callback	3100	// this is how you can declare your typical callback
2155	static void cb (EV_P_ ev_timer *w, int revents)	3101	static void cb (EV_P_ ev_timer *w, int revents)
2156		3102
2157	It declares a parameter C<loop> of type C<struct ev_loop *>, quite	3103	It declares a parameter C<loop> of type C<struct ev_loop *>, quite
2158	suitable for use with C<EV_A>.	3104	suitable for use with C<EV_A>.
2159		3105
2160	=item C<EV_DEFAULT>, C<EV_DEFAULT_>	3106	=item C<EV_DEFAULT>, C<EV_DEFAULT_>
2161		3107
2162	Similar to the other two macros, this gives you the value of the default	3108	Similar to the other two macros, this gives you the value of the default
2163	loop, if multiple loops are supported ("ev loop default").	3109	loop, if multiple loops are supported ("ev loop default").
		3110
		3111	=item C<EV_DEFAULT_UC>, C<EV_DEFAULT_UC_>
		3112
		3113	Usage identical to C<EV_DEFAULT> and C<EV_DEFAULT_>, but requires that the
		3114	default loop has been initialised (C<UC> == unchecked). Their behaviour
		3115	is undefined when the default loop has not been initialised by a previous
		3116	execution of C<EV_DEFAULT>, C<EV_DEFAULT_> or C<ev_default_init (...)>.
		3117
		3118	It is often prudent to use C<EV_DEFAULT> when initialising the first
		3119	watcher in a function but use C<EV_DEFAULT_UC> afterwards.
2164		3120
2165	=back	3121	=back
2166		3122
2167	Example: Declare and initialise a check watcher, utilising the above	3123	Example: Declare and initialise a check watcher, utilising the above
2168	macros so it will work regardless of whether multiple loops are supported	3124	macros so it will work regardless of whether multiple loops are supported
2169	or not.	3125	or not.
2170		3126
2171	static void	3127	static void
2172	check_cb (EV_P_ ev_timer *w, int revents)	3128	check_cb (EV_P_ ev_timer *w, int revents)
2173	{	3129	{
2174	ev_check_stop (EV_A_ w);	3130	ev_check_stop (EV_A_ w);
2175	}	3131	}
2176		3132
2177	ev_check check;	3133	ev_check check;
2178	ev_check_init (&check, check_cb);	3134	ev_check_init (&check, check_cb);
2179	ev_check_start (EV_DEFAULT_ &check);	3135	ev_check_start (EV_DEFAULT_ &check);
2180	ev_loop (EV_DEFAULT_ 0);	3136	ev_loop (EV_DEFAULT_ 0);
2181		3137
2182	=head1 EMBEDDING	3138	=head1 EMBEDDING
2183		3139
2184	Libev can (and often is) directly embedded into host	3140	Libev can (and often is) directly embedded into host
2185	applications. Examples of applications that embed it include the Deliantra	3141	applications. Examples of applications that embed it include the Deliantra
…		…
2192	libev somewhere in your source tree).	3148	libev somewhere in your source tree).
2193		3149
2194	=head2 FILESETS	3150	=head2 FILESETS
2195		3151
2196	Depending on what features you need you need to include one or more sets of files	3152	Depending on what features you need you need to include one or more sets of files
2197	in your app.	3153	in your application.
2198		3154
2199	=head3 CORE EVENT LOOP	3155	=head3 CORE EVENT LOOP
2200		3156
2201	To include only the libev core (all the C<ev_*> functions), with manual	3157	To include only the libev core (all the C<ev_*> functions), with manual
2202	configuration (no autoconf):	3158	configuration (no autoconf):
2203		3159
2204	#define EV_STANDALONE 1	3160	#define EV_STANDALONE 1
2205	#include "ev.c"	3161	#include "ev.c"
2206		3162
2207	This will automatically include F<ev.h>, too, and should be done in a	3163	This will automatically include F<ev.h>, too, and should be done in a
2208	single C source file only to provide the function implementations. To use	3164	single C source file only to provide the function implementations. To use
2209	it, do the same for F<ev.h> in all files wishing to use this API (best	3165	it, do the same for F<ev.h> in all files wishing to use this API (best
2210	done by writing a wrapper around F<ev.h> that you can include instead and	3166	done by writing a wrapper around F<ev.h> that you can include instead and
2211	where you can put other configuration options):	3167	where you can put other configuration options):
2212		3168
2213	#define EV_STANDALONE 1	3169	#define EV_STANDALONE 1
2214	#include "ev.h"	3170	#include "ev.h"
2215		3171
2216	Both header files and implementation files can be compiled with a C++	3172	Both header files and implementation files can be compiled with a C++
2217	compiler (at least, thats a stated goal, and breakage will be treated	3173	compiler (at least, that's a stated goal, and breakage will be treated
2218	as a bug).	3174	as a bug).
2219		3175
2220	You need the following files in your source tree, or in a directory	3176	You need the following files in your source tree, or in a directory
2221	in your include path (e.g. in libev/ when using -Ilibev):	3177	in your include path (e.g. in libev/ when using -Ilibev):
2222		3178
2223	ev.h	3179	ev.h
2224	ev.c	3180	ev.c
2225	ev_vars.h	3181	ev_vars.h
2226	ev_wrap.h	3182	ev_wrap.h
2227		3183
2228	ev_win32.c required on win32 platforms only	3184	ev_win32.c required on win32 platforms only
2229		3185
2230	ev_select.c only when select backend is enabled (which is enabled by default)	3186	ev_select.c only when select backend is enabled (which is enabled by default)
2231	ev_poll.c only when poll backend is enabled (disabled by default)	3187	ev_poll.c only when poll backend is enabled (disabled by default)
2232	ev_epoll.c only when the epoll backend is enabled (disabled by default)	3188	ev_epoll.c only when the epoll backend is enabled (disabled by default)
2233	ev_kqueue.c only when the kqueue backend is enabled (disabled by default)	3189	ev_kqueue.c only when the kqueue backend is enabled (disabled by default)
2234	ev_port.c only when the solaris port backend is enabled (disabled by default)	3190	ev_port.c only when the solaris port backend is enabled (disabled by default)
2235		3191
2236	F<ev.c> includes the backend files directly when enabled, so you only need	3192	F<ev.c> includes the backend files directly when enabled, so you only need
2237	to compile this single file.	3193	to compile this single file.
2238		3194
2239	=head3 LIBEVENT COMPATIBILITY API	3195	=head3 LIBEVENT COMPATIBILITY API
2240		3196
2241	To include the libevent compatibility API, also include:	3197	To include the libevent compatibility API, also include:
2242		3198
2243	#include "event.c"	3199	#include "event.c"
2244		3200
2245	in the file including F<ev.c>, and:	3201	in the file including F<ev.c>, and:
2246		3202
2247	#include "event.h"	3203	#include "event.h"
2248		3204
2249	in the files that want to use the libevent API. This also includes F<ev.h>.	3205	in the files that want to use the libevent API. This also includes F<ev.h>.
2250		3206
2251	You need the following additional files for this:	3207	You need the following additional files for this:
2252		3208
2253	event.h	3209	event.h
2254	event.c	3210	event.c
2255		3211
2256	=head3 AUTOCONF SUPPORT	3212	=head3 AUTOCONF SUPPORT
2257		3213
2258	Instead of using C<EV_STANDALONE=1> and providing your config in	3214	Instead of using C<EV_STANDALONE=1> and providing your configuration in
2259	whatever way you want, you can also C<m4_include([libev.m4])> in your	3215	whatever way you want, you can also C<m4_include([libev.m4])> in your
2260	F<configure.ac> and leave C<EV_STANDALONE> undefined. F<ev.c> will then	3216	F<configure.ac> and leave C<EV_STANDALONE> undefined. F<ev.c> will then
2261	include F<config.h> and configure itself accordingly.	3217	include F<config.h> and configure itself accordingly.
2262		3218
2263	For this of course you need the m4 file:	3219	For this of course you need the m4 file:
2264		3220
2265	libev.m4	3221	libev.m4
2266		3222
2267	=head2 PREPROCESSOR SYMBOLS/MACROS	3223	=head2 PREPROCESSOR SYMBOLS/MACROS
2268		3224
2269	Libev can be configured via a variety of preprocessor symbols you have to define	3225	Libev can be configured via a variety of preprocessor symbols you have to
2270	before including any of its files. The default is not to build for multiplicity	3226	define before including any of its files. The default in the absence of
2271	and only include the select backend.	3227	autoconf is documented for every option.
2272		3228
2273	=over 4	3229	=over 4
2274		3230
2275	=item EV_STANDALONE	3231	=item EV_STANDALONE
2276		3232
…		…
2278	keeps libev from including F<config.h>, and it also defines dummy	3234	keeps libev from including F<config.h>, and it also defines dummy
2279	implementations for some libevent functions (such as logging, which is not	3235	implementations for some libevent functions (such as logging, which is not
2280	supported). It will also not define any of the structs usually found in	3236	supported). It will also not define any of the structs usually found in
2281	F<event.h> that are not directly supported by the libev core alone.	3237	F<event.h> that are not directly supported by the libev core alone.
2282		3238
		3239	In stanbdalone mode, libev will still try to automatically deduce the
		3240	configuration, but has to be more conservative.
		3241
2283	=item EV_USE_MONOTONIC	3242	=item EV_USE_MONOTONIC
2284		3243
2285	If defined to be C<1>, libev will try to detect the availability of the	3244	If defined to be C<1>, libev will try to detect the availability of the
2286	monotonic clock option at both compiletime and runtime. Otherwise no use	3245	monotonic clock option at both compile time and runtime. Otherwise no
2287	of the monotonic clock option will be attempted. If you enable this, you	3246	use of the monotonic clock option will be attempted. If you enable this,
2288	usually have to link against librt or something similar. Enabling it when	3247	you usually have to link against librt or something similar. Enabling it
2289	the functionality isn't available is safe, though, although you have	3248	when the functionality isn't available is safe, though, although you have
2290	to make sure you link against any libraries where the C<clock_gettime>	3249	to make sure you link against any libraries where the C<clock_gettime>
2291	function is hiding in (often F<-lrt>).	3250	function is hiding in (often F<-lrt>). See also C<EV_USE_CLOCK_SYSCALL>.
2292		3251
2293	=item EV_USE_REALTIME	3252	=item EV_USE_REALTIME
2294		3253
2295	If defined to be C<1>, libev will try to detect the availability of the	3254	If defined to be C<1>, libev will try to detect the availability of the
2296	realtime clock option at compiletime (and assume its availability at	3255	real-time clock option at compile time (and assume its availability
2297	runtime if successful). Otherwise no use of the realtime clock option will	3256	at runtime if successful). Otherwise no use of the real-time clock
2298	be attempted. This effectively replaces C<gettimeofday> by C<clock_get	3257	option will be attempted. This effectively replaces C<gettimeofday>
2299	(CLOCK_REALTIME, ...)> and will not normally affect correctness. See the	3258	by C<clock_get (CLOCK_REALTIME, ...)> and will not normally affect
2300	note about libraries in the description of C<EV_USE_MONOTONIC>, though.	3259	correctness. See the note about libraries in the description of
		3260	C<EV_USE_MONOTONIC>, though. Defaults to the opposite value of
		3261	C<EV_USE_CLOCK_SYSCALL>.
		3262
		3263	=item EV_USE_CLOCK_SYSCALL
		3264
		3265	If defined to be C<1>, libev will try to use a direct syscall instead
		3266	of calling the system-provided C<clock_gettime> function. This option
		3267	exists because on GNU/Linux, C<clock_gettime> is in C<librt>, but C<librt>
		3268	unconditionally pulls in C<libpthread>, slowing down single-threaded
		3269	programs needlessly. Using a direct syscall is slightly slower (in
		3270	theory), because no optimised vdso implementation can be used, but avoids
		3271	the pthread dependency. Defaults to C<1> on GNU/Linux with glibc 2.x or
		3272	higher, as it simplifies linking (no need for C<-lrt>).
		3273
		3274	=item EV_USE_NANOSLEEP
		3275
		3276	If defined to be C<1>, libev will assume that C<nanosleep ()> is available
		3277	and will use it for delays. Otherwise it will use C<select ()>.
		3278
		3279	=item EV_USE_EVENTFD
		3280
		3281	If defined to be C<1>, then libev will assume that C<eventfd ()> is
		3282	available and will probe for kernel support at runtime. This will improve
		3283	C<ev_signal> and C<ev_async> performance and reduce resource consumption.
		3284	If undefined, it will be enabled if the headers indicate GNU/Linux + Glibc
		3285	2.7 or newer, otherwise disabled.
2301		3286
2302	=item EV_USE_SELECT	3287	=item EV_USE_SELECT
2303		3288
2304	If undefined or defined to be C<1>, libev will compile in support for the	3289	If undefined or defined to be C<1>, libev will compile in support for the
2305	C<select>(2) backend. No attempt at autodetection will be done: if no	3290	C<select>(2) backend. No attempt at auto-detection will be done: if no
2306	other method takes over, select will be it. Otherwise the select backend	3291	other method takes over, select will be it. Otherwise the select backend
2307	will not be compiled in.	3292	will not be compiled in.
2308		3293
2309	=item EV_SELECT_USE_FD_SET	3294	=item EV_SELECT_USE_FD_SET
2310		3295
2311	If defined to C<1>, then the select backend will use the system C<fd_set>	3296	If defined to C<1>, then the select backend will use the system C<fd_set>
2312	structure. This is useful if libev doesn't compile due to a missing	3297	structure. This is useful if libev doesn't compile due to a missing
2313	C<NFDBITS> or C<fd_mask> definition or it misguesses the bitset layout on	3298	C<NFDBITS> or C<fd_mask> definition or it mis-guesses the bitset layout
2314	exotic systems. This usually limits the range of file descriptors to some	3299	on exotic systems. This usually limits the range of file descriptors to
2315	low limit such as 1024 or might have other limitations (winsocket only	3300	some low limit such as 1024 or might have other limitations (winsocket
2316	allows 64 sockets). The C<FD_SETSIZE> macro, set before compilation, might	3301	only allows 64 sockets). The C<FD_SETSIZE> macro, set before compilation,
2317	influence the size of the C<fd_set> used.	3302	configures the maximum size of the C<fd_set>.
2318		3303
2319	=item EV_SELECT_IS_WINSOCKET	3304	=item EV_SELECT_IS_WINSOCKET
2320		3305
2321	When defined to C<1>, the select backend will assume that	3306	When defined to C<1>, the select backend will assume that
2322	select/socket/connect etc. don't understand file descriptors but	3307	select/socket/connect etc. don't understand file descriptors but
…		…
2324	be used is the winsock select). This means that it will call	3309	be used is the winsock select). This means that it will call
2325	C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,	3310	C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,
2326	it is assumed that all these functions actually work on fds, even	3311	it is assumed that all these functions actually work on fds, even
2327	on win32. Should not be defined on non-win32 platforms.	3312	on win32. Should not be defined on non-win32 platforms.
2328		3313
		3314	=item EV_FD_TO_WIN32_HANDLE
		3315
		3316	If C<EV_SELECT_IS_WINSOCKET> is enabled, then libev needs a way to map
		3317	file descriptors to socket handles. When not defining this symbol (the
		3318	default), then libev will call C<_get_osfhandle>, which is usually
		3319	correct. In some cases, programs use their own file descriptor management,
		3320	in which case they can provide this function to map fds to socket handles.
		3321
2329	=item EV_USE_POLL	3322	=item EV_USE_POLL
2330		3323
2331	If defined to be C<1>, libev will compile in support for the C<poll>(2)	3324	If defined to be C<1>, libev will compile in support for the C<poll>(2)
2332	backend. Otherwise it will be enabled on non-win32 platforms. It	3325	backend. Otherwise it will be enabled on non-win32 platforms. It
2333	takes precedence over select.	3326	takes precedence over select.
2334		3327
2335	=item EV_USE_EPOLL	3328	=item EV_USE_EPOLL
2336		3329
2337	If defined to be C<1>, libev will compile in support for the Linux	3330	If defined to be C<1>, libev will compile in support for the Linux
2338	C<epoll>(7) backend. Its availability will be detected at runtime,	3331	C<epoll>(7) backend. Its availability will be detected at runtime,
2339	otherwise another method will be used as fallback. This is the	3332	otherwise another method will be used as fallback. This is the preferred
2340	preferred backend for GNU/Linux systems.	3333	backend for GNU/Linux systems. If undefined, it will be enabled if the
		3334	headers indicate GNU/Linux + Glibc 2.4 or newer, otherwise disabled.
2341		3335
2342	=item EV_USE_KQUEUE	3336	=item EV_USE_KQUEUE
2343		3337
2344	If defined to be C<1>, libev will compile in support for the BSD style	3338	If defined to be C<1>, libev will compile in support for the BSD style
2345	C<kqueue>(2) backend. Its actual availability will be detected at runtime,	3339	C<kqueue>(2) backend. Its actual availability will be detected at runtime,
…		…
2358	otherwise another method will be used as fallback. This is the preferred	3352	otherwise another method will be used as fallback. This is the preferred
2359	backend for Solaris 10 systems.	3353	backend for Solaris 10 systems.
2360		3354
2361	=item EV_USE_DEVPOLL	3355	=item EV_USE_DEVPOLL
2362		3356
2363	reserved for future expansion, works like the USE symbols above.	3357	Reserved for future expansion, works like the USE symbols above.
2364		3358
2365	=item EV_USE_INOTIFY	3359	=item EV_USE_INOTIFY
2366		3360
2367	If defined to be C<1>, libev will compile in support for the Linux inotify	3361	If defined to be C<1>, libev will compile in support for the Linux inotify
2368	interface to speed up C<ev_stat> watchers. Its actual availability will	3362	interface to speed up C<ev_stat> watchers. Its actual availability will
2369	be detected at runtime.	3363	be detected at runtime. If undefined, it will be enabled if the headers
		3364	indicate GNU/Linux + Glibc 2.4 or newer, otherwise disabled.
		3365
		3366	=item EV_ATOMIC_T
		3367
		3368	Libev requires an integer type (suitable for storing C<0> or C<1>) whose
		3369	access is atomic with respect to other threads or signal contexts. No such
		3370	type is easily found in the C language, so you can provide your own type
		3371	that you know is safe for your purposes. It is used both for signal handler "locking"
		3372	as well as for signal and thread safety in C<ev_async> watchers.
		3373
		3374	In the absence of this define, libev will use C<sig_atomic_t volatile>
		3375	(from F<signal.h>), which is usually good enough on most platforms.
2370		3376
2371	=item EV_H	3377	=item EV_H
2372		3378
2373	The name of the F<ev.h> header file used to include it. The default if	3379	The name of the F<ev.h> header file used to include it. The default if
2374	undefined is C<< <ev.h> >> in F<event.h> and C<"ev.h"> in F<ev.c>. This	3380	undefined is C<"ev.h"> in F<event.h>, F<ev.c> and F<ev++.h>. This can be
2375	can be used to virtually rename the F<ev.h> header file in case of conflicts.	3381	used to virtually rename the F<ev.h> header file in case of conflicts.
2376		3382
2377	=item EV_CONFIG_H	3383	=item EV_CONFIG_H
2378		3384
2379	If C<EV_STANDALONE> isn't C<1>, this variable can be used to override	3385	If C<EV_STANDALONE> isn't C<1>, this variable can be used to override
2380	F<ev.c>'s idea of where to find the F<config.h> file, similarly to	3386	F<ev.c>'s idea of where to find the F<config.h> file, similarly to
2381	C<EV_H>, above.	3387	C<EV_H>, above.
2382		3388
2383	=item EV_EVENT_H	3389	=item EV_EVENT_H
2384		3390
2385	Similarly to C<EV_H>, this macro can be used to override F<event.c>'s idea	3391	Similarly to C<EV_H>, this macro can be used to override F<event.c>'s idea
2386	of how the F<event.h> header can be found.	3392	of how the F<event.h> header can be found, the default is C<"event.h">.
2387		3393
2388	=item EV_PROTOTYPES	3394	=item EV_PROTOTYPES
2389		3395
2390	If defined to be C<0>, then F<ev.h> will not define any function	3396	If defined to be C<0>, then F<ev.h> will not define any function
2391	prototypes, but still define all the structs and other symbols. This is	3397	prototypes, but still define all the structs and other symbols. This is
…		…
2412	When doing priority-based operations, libev usually has to linearly search	3418	When doing priority-based operations, libev usually has to linearly search
2413	all the priorities, so having many of them (hundreds) uses a lot of space	3419	all the priorities, so having many of them (hundreds) uses a lot of space
2414	and time, so using the defaults of five priorities (-2 .. +2) is usually	3420	and time, so using the defaults of five priorities (-2 .. +2) is usually
2415	fine.	3421	fine.
2416		3422
2417	If your embedding app does not need any priorities, defining these both to	3423	If your embedding application does not need any priorities, defining these
2418	C<0> will save some memory and cpu.	3424	both to C<0> will save some memory and CPU.
2419		3425
2420	=item EV_PERIODIC_ENABLE	3426	=item EV_PERIODIC_ENABLE
2421		3427
2422	If undefined or defined to be C<1>, then periodic timers are supported. If	3428	If undefined or defined to be C<1>, then periodic timers are supported. If
2423	defined to be C<0>, then they are not. Disabling them saves a few kB of	3429	defined to be C<0>, then they are not. Disabling them saves a few kB of
…		…
2430	code.	3436	code.
2431		3437
2432	=item EV_EMBED_ENABLE	3438	=item EV_EMBED_ENABLE
2433		3439
2434	If undefined or defined to be C<1>, then embed watchers are supported. If	3440	If undefined or defined to be C<1>, then embed watchers are supported. If
2435	defined to be C<0>, then they are not.	3441	defined to be C<0>, then they are not. Embed watchers rely on most other
		3442	watcher types, which therefore must not be disabled.
2436		3443
2437	=item EV_STAT_ENABLE	3444	=item EV_STAT_ENABLE
2438		3445
2439	If undefined or defined to be C<1>, then stat watchers are supported. If	3446	If undefined or defined to be C<1>, then stat watchers are supported. If
2440	defined to be C<0>, then they are not.	3447	defined to be C<0>, then they are not.
…		…
2442	=item EV_FORK_ENABLE	3449	=item EV_FORK_ENABLE
2443		3450
2444	If undefined or defined to be C<1>, then fork watchers are supported. If	3451	If undefined or defined to be C<1>, then fork watchers are supported. If
2445	defined to be C<0>, then they are not.	3452	defined to be C<0>, then they are not.
2446		3453
		3454	=item EV_ASYNC_ENABLE
		3455
		3456	If undefined or defined to be C<1>, then async watchers are supported. If
		3457	defined to be C<0>, then they are not.
		3458
2447	=item EV_MINIMAL	3459	=item EV_MINIMAL
2448		3460
2449	If you need to shave off some kilobytes of code at the expense of some	3461	If you need to shave off some kilobytes of code at the expense of some
2450	speed, define this symbol to C<1>. Currently only used for gcc to override	3462	speed, define this symbol to C<1>. Currently this is used to override some
2451	some inlining decisions, saves roughly 30% codesize of amd64.	3463	inlining decisions, saves roughly 30% code size on amd64. It also selects a
		3464	much smaller 2-heap for timer management over the default 4-heap.
2452		3465
2453	=item EV_PID_HASHSIZE	3466	=item EV_PID_HASHSIZE
2454		3467
2455	C<ev_child> watchers use a small hash table to distribute workload by	3468	C<ev_child> watchers use a small hash table to distribute workload by
2456	pid. The default size is C<16> (or C<1> with C<EV_MINIMAL>), usually more	3469	pid. The default size is C<16> (or C<1> with C<EV_MINIMAL>), usually more
2457	than enough. If you need to manage thousands of children you might want to	3470	than enough. If you need to manage thousands of children you might want to
2458	increase this value (I<must> be a power of two).	3471	increase this value (I<must> be a power of two).
2459		3472
2460	=item EV_INOTIFY_HASHSIZE	3473	=item EV_INOTIFY_HASHSIZE
2461		3474
2462	C<ev_staz> watchers use a small hash table to distribute workload by	3475	C<ev_stat> watchers use a small hash table to distribute workload by
2463	inotify watch id. The default size is C<16> (or C<1> with C<EV_MINIMAL>),	3476	inotify watch id. The default size is C<16> (or C<1> with C<EV_MINIMAL>),
2464	usually more than enough. If you need to manage thousands of C<ev_stat>	3477	usually more than enough. If you need to manage thousands of C<ev_stat>
2465	watchers you might want to increase this value (I<must> be a power of	3478	watchers you might want to increase this value (I<must> be a power of
2466	two).	3479	two).
2467		3480
		3481	=item EV_USE_4HEAP
		3482
		3483	Heaps are not very cache-efficient. To improve the cache-efficiency of the
		3484	timer and periodics heaps, libev uses a 4-heap when this symbol is defined
		3485	to C<1>. The 4-heap uses more complicated (longer) code but has noticeably
		3486	faster performance with many (thousands) of watchers.
		3487
		3488	The default is C<1> unless C<EV_MINIMAL> is set in which case it is C<0>
		3489	(disabled).
		3490
		3491	=item EV_HEAP_CACHE_AT
		3492
		3493	Heaps are not very cache-efficient. To improve the cache-efficiency of the
		3494	timer and periodics heaps, libev can cache the timestamp (I<at>) within
		3495	the heap structure (selected by defining C<EV_HEAP_CACHE_AT> to C<1>),
		3496	which uses 8-12 bytes more per watcher and a few hundred bytes more code,
		3497	but avoids random read accesses on heap changes. This improves performance
		3498	noticeably with many (hundreds) of watchers.
		3499
		3500	The default is C<1> unless C<EV_MINIMAL> is set in which case it is C<0>
		3501	(disabled).
		3502
		3503	=item EV_VERIFY
		3504
		3505	Controls how much internal verification (see C<ev_loop_verify ()>) will
		3506	be done: If set to C<0>, no internal verification code will be compiled
		3507	in. If set to C<1>, then verification code will be compiled in, but not
		3508	called. If set to C<2>, then the internal verification code will be
		3509	called once per loop, which can slow down libev. If set to C<3>, then the
		3510	verification code will be called very frequently, which will slow down
		3511	libev considerably.
		3512
		3513	The default is C<1>, unless C<EV_MINIMAL> is set, in which case it will be
		3514	C<0>.
		3515
2468	=item EV_COMMON	3516	=item EV_COMMON
2469		3517
2470	By default, all watchers have a C<void *data> member. By redefining	3518	By default, all watchers have a C<void *data> member. By redefining
2471	this macro to a something else you can include more and other types of	3519	this macro to a something else you can include more and other types of
2472	members. You have to define it each time you include one of the files,	3520	members. You have to define it each time you include one of the files,
2473	though, and it must be identical each time.	3521	though, and it must be identical each time.
2474		3522
2475	For example, the perl EV module uses something like this:	3523	For example, the perl EV module uses something like this:
2476		3524
2477	#define EV_COMMON \	3525	#define EV_COMMON \
2478	SV *self; /* contains this struct */ \	3526	SV *self; /* contains this struct */ \
2479	SV *cb_sv, *fh /* note no trailing ";" */	3527	SV *cb_sv, *fh /* note no trailing ";" */
2480		3528
2481	=item EV_CB_DECLARE (type)	3529	=item EV_CB_DECLARE (type)
2482		3530
2483	=item EV_CB_INVOKE (watcher, revents)	3531	=item EV_CB_INVOKE (watcher, revents)
2484		3532
…		…
2489	definition and a statement, respectively. See the F<ev.h> header file for	3537	definition and a statement, respectively. See the F<ev.h> header file for
2490	their default definitions. One possible use for overriding these is to	3538	their default definitions. One possible use for overriding these is to
2491	avoid the C<struct ev_loop *> as first argument in all cases, or to use	3539	avoid the C<struct ev_loop *> as first argument in all cases, or to use
2492	method calls instead of plain function calls in C++.	3540	method calls instead of plain function calls in C++.
2493		3541
		3542	=back
		3543
2494	=head2 EXPORTED API SYMBOLS	3544	=head2 EXPORTED API SYMBOLS
2495		3545
2496	If you need to re-export the API (e.g. via a dll) and you need a list of	3546	If you need to re-export the API (e.g. via a DLL) and you need a list of
2497	exported symbols, you can use the provided F<Symbol.*> files which list	3547	exported symbols, you can use the provided F<Symbol.*> files which list
2498	all public symbols, one per line:	3548	all public symbols, one per line:
2499		3549
2500	Symbols.ev for libev proper	3550	Symbols.ev for libev proper
2501	Symbols.event for the libevent emulation	3551	Symbols.event for the libevent emulation
2502		3552
2503	This can also be used to rename all public symbols to avoid clashes with	3553	This can also be used to rename all public symbols to avoid clashes with
2504	multiple versions of libev linked together (which is obviously bad in	3554	multiple versions of libev linked together (which is obviously bad in
2505	itself, but sometimes it is inconvinient to avoid this).	3555	itself, but sometimes it is inconvenient to avoid this).
2506		3556
2507	A sed command like this will create wrapper C<#define>'s that you need to	3557	A sed command like this will create wrapper C<#define>'s that you need to
2508	include before including F<ev.h>:	3558	include before including F<ev.h>:
2509		3559
2510	<Symbols.ev sed -e "s/.*/#define & myprefix_&/" >wrap.h	3560	<Symbols.ev sed -e "s/.*/#define & myprefix_&/" >wrap.h
…		…
2527	file.	3577	file.
2528		3578
2529	The usage in rxvt-unicode is simpler. It has a F<ev_cpp.h> header file	3579	The usage in rxvt-unicode is simpler. It has a F<ev_cpp.h> header file
2530	that everybody includes and which overrides some configure choices:	3580	that everybody includes and which overrides some configure choices:
2531		3581
2532	#define EV_MINIMAL 1	3582	#define EV_MINIMAL 1
2533	#define EV_USE_POLL 0	3583	#define EV_USE_POLL 0
2534	#define EV_MULTIPLICITY 0	3584	#define EV_MULTIPLICITY 0
2535	#define EV_PERIODIC_ENABLE 0	3585	#define EV_PERIODIC_ENABLE 0
2536	#define EV_STAT_ENABLE 0	3586	#define EV_STAT_ENABLE 0
2537	#define EV_FORK_ENABLE 0	3587	#define EV_FORK_ENABLE 0
2538	#define EV_CONFIG_H <config.h>	3588	#define EV_CONFIG_H <config.h>
2539	#define EV_MINPRI 0	3589	#define EV_MINPRI 0
2540	#define EV_MAXPRI 0	3590	#define EV_MAXPRI 0
2541		3591
2542	#include "ev++.h"	3592	#include "ev++.h"
2543		3593
2544	And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:	3594	And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:
2545		3595
2546	#include "ev_cpp.h"	3596	#include "ev_cpp.h"
2547	#include "ev.c"	3597	#include "ev.c"
2548		3598
		3599	=head1 INTERACTION WITH OTHER PROGRAMS OR LIBRARIES
2549		3600
		3601	=head2 THREADS AND COROUTINES
		3602
		3603	=head3 THREADS
		3604
		3605	All libev functions are reentrant and thread-safe unless explicitly
		3606	documented otherwise, but libev implements no locking itself. This means
		3607	that you can use as many loops as you want in parallel, as long as there
		3608	are no concurrent calls into any libev function with the same loop
		3609	parameter (C<ev_default_*> calls have an implicit default loop parameter,
		3610	of course): libev guarantees that different event loops share no data
		3611	structures that need any locking.
		3612
		3613	Or to put it differently: calls with different loop parameters can be done
		3614	concurrently from multiple threads, calls with the same loop parameter
		3615	must be done serially (but can be done from different threads, as long as
		3616	only one thread ever is inside a call at any point in time, e.g. by using
		3617	a mutex per loop).
		3618
		3619	Specifically to support threads (and signal handlers), libev implements
		3620	so-called C<ev_async> watchers, which allow some limited form of
		3621	concurrency on the same event loop, namely waking it up "from the
		3622	outside".
		3623
		3624	If you want to know which design (one loop, locking, or multiple loops
		3625	without or something else still) is best for your problem, then I cannot
		3626	help you, but here is some generic advice:
		3627
		3628	=over 4
		3629
		3630	=item * most applications have a main thread: use the default libev loop
		3631	in that thread, or create a separate thread running only the default loop.
		3632
		3633	This helps integrating other libraries or software modules that use libev
		3634	themselves and don't care/know about threading.
		3635
		3636	=item * one loop per thread is usually a good model.
		3637
		3638	Doing this is almost never wrong, sometimes a better-performance model
		3639	exists, but it is always a good start.
		3640
		3641	=item * other models exist, such as the leader/follower pattern, where one
		3642	loop is handed through multiple threads in a kind of round-robin fashion.
		3643
		3644	Choosing a model is hard - look around, learn, know that usually you can do
		3645	better than you currently do :-)
		3646
		3647	=item * often you need to talk to some other thread which blocks in the
		3648	event loop.
		3649
		3650	C<ev_async> watchers can be used to wake them up from other threads safely
		3651	(or from signal contexts...).
		3652
		3653	An example use would be to communicate signals or other events that only
		3654	work in the default loop by registering the signal watcher with the
		3655	default loop and triggering an C<ev_async> watcher from the default loop
		3656	watcher callback into the event loop interested in the signal.
		3657
		3658	=back
		3659
		3660	=head3 COROUTINES
		3661
		3662	Libev is very accommodating to coroutines ("cooperative threads"):
		3663	libev fully supports nesting calls to its functions from different
		3664	coroutines (e.g. you can call C<ev_loop> on the same loop from two
		3665	different coroutines, and switch freely between both coroutines running the
		3666	loop, as long as you don't confuse yourself). The only exception is that
		3667	you must not do this from C<ev_periodic> reschedule callbacks.
		3668
		3669	Care has been taken to ensure that libev does not keep local state inside
		3670	C<ev_loop>, and other calls do not usually allow for coroutine switches as
		3671	they do not call any callbacks.
		3672
		3673	=head2 COMPILER WARNINGS
		3674
		3675	Depending on your compiler and compiler settings, you might get no or a
		3676	lot of warnings when compiling libev code. Some people are apparently
		3677	scared by this.
		3678
		3679	However, these are unavoidable for many reasons. For one, each compiler
		3680	has different warnings, and each user has different tastes regarding
		3681	warning options. "Warn-free" code therefore cannot be a goal except when
		3682	targeting a specific compiler and compiler-version.
		3683
		3684	Another reason is that some compiler warnings require elaborate
		3685	workarounds, or other changes to the code that make it less clear and less
		3686	maintainable.
		3687
		3688	And of course, some compiler warnings are just plain stupid, or simply
		3689	wrong (because they don't actually warn about the condition their message
		3690	seems to warn about). For example, certain older gcc versions had some
		3691	warnings that resulted an extreme number of false positives. These have
		3692	been fixed, but some people still insist on making code warn-free with
		3693	such buggy versions.
		3694
		3695	While libev is written to generate as few warnings as possible,
		3696	"warn-free" code is not a goal, and it is recommended not to build libev
		3697	with any compiler warnings enabled unless you are prepared to cope with
		3698	them (e.g. by ignoring them). Remember that warnings are just that:
		3699	warnings, not errors, or proof of bugs.
		3700
		3701
		3702	=head2 VALGRIND
		3703
		3704	Valgrind has a special section here because it is a popular tool that is
		3705	highly useful. Unfortunately, valgrind reports are very hard to interpret.
		3706
		3707	If you think you found a bug (memory leak, uninitialised data access etc.)
		3708	in libev, then check twice: If valgrind reports something like:
		3709
		3710	==2274== definitely lost: 0 bytes in 0 blocks.
		3711	==2274== possibly lost: 0 bytes in 0 blocks.
		3712	==2274== still reachable: 256 bytes in 1 blocks.
		3713
		3714	Then there is no memory leak, just as memory accounted to global variables
		3715	is not a memleak - the memory is still being referenced, and didn't leak.
		3716
		3717	Similarly, under some circumstances, valgrind might report kernel bugs
		3718	as if it were a bug in libev (e.g. in realloc or in the poll backend,
		3719	although an acceptable workaround has been found here), or it might be
		3720	confused.
		3721
		3722	Keep in mind that valgrind is a very good tool, but only a tool. Don't
		3723	make it into some kind of religion.
		3724
		3725	If you are unsure about something, feel free to contact the mailing list
		3726	with the full valgrind report and an explanation on why you think this
		3727	is a bug in libev (best check the archives, too :). However, don't be
		3728	annoyed when you get a brisk "this is no bug" answer and take the chance
		3729	of learning how to interpret valgrind properly.
		3730
		3731	If you need, for some reason, empty reports from valgrind for your project
		3732	I suggest using suppression lists.
		3733
		3734
		3735	=head1 PORTABILITY NOTES
		3736
		3737	=head2 WIN32 PLATFORM LIMITATIONS AND WORKAROUNDS
		3738
		3739	Win32 doesn't support any of the standards (e.g. POSIX) that libev
		3740	requires, and its I/O model is fundamentally incompatible with the POSIX
		3741	model. Libev still offers limited functionality on this platform in
		3742	the form of the C<EVBACKEND_SELECT> backend, and only supports socket
		3743	descriptors. This only applies when using Win32 natively, not when using
		3744	e.g. cygwin.
		3745
		3746	Lifting these limitations would basically require the full
		3747	re-implementation of the I/O system. If you are into these kinds of
		3748	things, then note that glib does exactly that for you in a very portable
		3749	way (note also that glib is the slowest event library known to man).
		3750
		3751	There is no supported compilation method available on windows except
		3752	embedding it into other applications.
		3753
		3754	Not a libev limitation but worth mentioning: windows apparently doesn't
		3755	accept large writes: instead of resulting in a partial write, windows will
		3756	either accept everything or return C<ENOBUFS> if the buffer is too large,
		3757	so make sure you only write small amounts into your sockets (less than a
		3758	megabyte seems safe, but this apparently depends on the amount of memory
		3759	available).
		3760
		3761	Due to the many, low, and arbitrary limits on the win32 platform and
		3762	the abysmal performance of winsockets, using a large number of sockets
		3763	is not recommended (and not reasonable). If your program needs to use
		3764	more than a hundred or so sockets, then likely it needs to use a totally
		3765	different implementation for windows, as libev offers the POSIX readiness
		3766	notification model, which cannot be implemented efficiently on windows
		3767	(Microsoft monopoly games).
		3768
		3769	A typical way to use libev under windows is to embed it (see the embedding
		3770	section for details) and use the following F<evwrap.h> header file instead
		3771	of F<ev.h>:
		3772
		3773	#define EV_STANDALONE /* keeps ev from requiring config.h */
		3774	#define EV_SELECT_IS_WINSOCKET 1 /* configure libev for windows select */
		3775
		3776	#include "ev.h"
		3777
		3778	And compile the following F<evwrap.c> file into your project (make sure
		3779	you do I<not> compile the F<ev.c> or any other embedded source files!):
		3780
		3781	#include "evwrap.h"
		3782	#include "ev.c"
		3783
		3784	=over 4
		3785
		3786	=item The winsocket select function
		3787
		3788	The winsocket C<select> function doesn't follow POSIX in that it
		3789	requires socket I<handles> and not socket I<file descriptors> (it is
		3790	also extremely buggy). This makes select very inefficient, and also
		3791	requires a mapping from file descriptors to socket handles (the Microsoft
		3792	C runtime provides the function C<_open_osfhandle> for this). See the
		3793	discussion of the C<EV_SELECT_USE_FD_SET>, C<EV_SELECT_IS_WINSOCKET> and
		3794	C<EV_FD_TO_WIN32_HANDLE> preprocessor symbols for more info.
		3795
		3796	The configuration for a "naked" win32 using the Microsoft runtime
		3797	libraries and raw winsocket select is:
		3798
		3799	#define EV_USE_SELECT 1
		3800	#define EV_SELECT_IS_WINSOCKET 1 /* forces EV_SELECT_USE_FD_SET, too */
		3801
		3802	Note that winsockets handling of fd sets is O(n), so you can easily get a
		3803	complexity in the O(n²) range when using win32.
		3804
		3805	=item Limited number of file descriptors
		3806
		3807	Windows has numerous arbitrary (and low) limits on things.
		3808
		3809	Early versions of winsocket's select only supported waiting for a maximum
		3810	of C<64> handles (probably owning to the fact that all windows kernels
		3811	can only wait for C<64> things at the same time internally; Microsoft
		3812	recommends spawning a chain of threads and wait for 63 handles and the
		3813	previous thread in each. Great).
		3814
		3815	Newer versions support more handles, but you need to define C<FD_SETSIZE>
		3816	to some high number (e.g. C<2048>) before compiling the winsocket select
		3817	call (which might be in libev or elsewhere, for example, perl does its own
		3818	select emulation on windows).
		3819
		3820	Another limit is the number of file descriptors in the Microsoft runtime
		3821	libraries, which by default is C<64> (there must be a hidden I<64> fetish
		3822	or something like this inside Microsoft). You can increase this by calling
		3823	C<_setmaxstdio>, which can increase this limit to C<2048> (another
		3824	arbitrary limit), but is broken in many versions of the Microsoft runtime
		3825	libraries.
		3826
		3827	This might get you to about C<512> or C<2048> sockets (depending on
		3828	windows version and/or the phase of the moon). To get more, you need to
		3829	wrap all I/O functions and provide your own fd management, but the cost of
		3830	calling select (O(n²)) will likely make this unworkable.
		3831
		3832	=back
		3833
		3834	=head2 PORTABILITY REQUIREMENTS
		3835
		3836	In addition to a working ISO-C implementation and of course the
		3837	backend-specific APIs, libev relies on a few additional extensions:
		3838
		3839	=over 4
		3840
		3841	=item C<void (*)(ev_watcher_type *, int revents)> must have compatible
		3842	calling conventions regardless of C<ev_watcher_type *>.
		3843
		3844	Libev assumes not only that all watcher pointers have the same internal
		3845	structure (guaranteed by POSIX but not by ISO C for example), but it also
		3846	assumes that the same (machine) code can be used to call any watcher
		3847	callback: The watcher callbacks have different type signatures, but libev
		3848	calls them using an C<ev_watcher *> internally.
		3849
		3850	=item C<sig_atomic_t volatile> must be thread-atomic as well
		3851
		3852	The type C<sig_atomic_t volatile> (or whatever is defined as
		3853	C<EV_ATOMIC_T>) must be atomic with respect to accesses from different
		3854	threads. This is not part of the specification for C<sig_atomic_t>, but is
		3855	believed to be sufficiently portable.
		3856
		3857	=item C<sigprocmask> must work in a threaded environment
		3858
		3859	Libev uses C<sigprocmask> to temporarily block signals. This is not
		3860	allowed in a threaded program (C<pthread_sigmask> has to be used). Typical
		3861	pthread implementations will either allow C<sigprocmask> in the "main
		3862	thread" or will block signals process-wide, both behaviours would
		3863	be compatible with libev. Interaction between C<sigprocmask> and
		3864	C<pthread_sigmask> could complicate things, however.
		3865
		3866	The most portable way to handle signals is to block signals in all threads
		3867	except the initial one, and run the default loop in the initial thread as
		3868	well.
		3869
		3870	=item C<long> must be large enough for common memory allocation sizes
		3871
		3872	To improve portability and simplify its API, libev uses C<long> internally
		3873	instead of C<size_t> when allocating its data structures. On non-POSIX
		3874	systems (Microsoft...) this might be unexpectedly low, but is still at
		3875	least 31 bits everywhere, which is enough for hundreds of millions of
		3876	watchers.
		3877
		3878	=item C<double> must hold a time value in seconds with enough accuracy
		3879
		3880	The type C<double> is used to represent timestamps. It is required to
		3881	have at least 51 bits of mantissa (and 9 bits of exponent), which is good
		3882	enough for at least into the year 4000. This requirement is fulfilled by
		3883	implementations implementing IEEE 754 (basically all existing ones).
		3884
		3885	=back
		3886
		3887	If you know of other additional requirements drop me a note.
		3888
		3889
2550	=head1 COMPLEXITIES	3890	=head1 ALGORITHMIC COMPLEXITIES
2551		3891
2552	In this section the complexities of (many of) the algorithms used inside	3892	In this section the complexities of (many of) the algorithms used inside
2553	libev will be explained. For complexity discussions about backends see the	3893	libev will be documented. For complexity discussions about backends see
2554	documentation for C<ev_default_init>.	3894	the documentation for C<ev_default_init>.
2555		3895
2556	All of the following are about amortised time: If an array needs to be	3896	All of the following are about amortised time: If an array needs to be
2557	extended, libev needs to realloc and move the whole array, but this	3897	extended, libev needs to realloc and move the whole array, but this
2558	happens asymptotically never with higher number of elements, so O(1) might	3898	happens asymptotically rarer with higher number of elements, so O(1) might
2559	mean it might do a lengthy realloc operation in rare cases, but on average	3899	mean that libev does a lengthy realloc operation in rare cases, but on
2560	it is much faster and asymptotically approaches constant time.	3900	average it is much faster and asymptotically approaches constant time.
2561		3901
2562	=over 4	3902	=over 4
2563		3903
2564	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)	3904	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)
2565		3905
2566	This means that, when you have a watcher that triggers in one hour and	3906	This means that, when you have a watcher that triggers in one hour and
2567	there are 100 watchers that would trigger before that then inserting will	3907	there are 100 watchers that would trigger before that, then inserting will
2568	have to skip those 100 watchers.	3908	have to skip roughly seven (C<ld 100>) of these watchers.
2569		3909
2570	=item Changing timer/periodic watchers (by autorepeat, again): O(log skipped_other_timers)	3910	=item Changing timer/periodic watchers (by autorepeat or calling again): O(log skipped_other_timers)
2571		3911
2572	That means that for changing a timer costs less than removing/adding them	3912	That means that changing a timer costs less than removing/adding them,
2573	as only the relative motion in the event queue has to be paid for.	3913	as only the relative motion in the event queue has to be paid for.
2574		3914
2575	=item Starting io/check/prepare/idle/signal/child watchers: O(1)	3915	=item Starting io/check/prepare/idle/signal/child/fork/async watchers: O(1)
2576		3916
2577	These just add the watcher into an array or at the head of a list.	3917	These just add the watcher into an array or at the head of a list.
		3918
2578	=item Stopping check/prepare/idle watchers: O(1)	3919	=item Stopping check/prepare/idle/fork/async watchers: O(1)
2579		3920
2580	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))	3921	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))
2581		3922
2582	These watchers are stored in lists then need to be walked to find the	3923	These watchers are stored in lists, so they need to be walked to find the
2583	correct watcher to remove. The lists are usually short (you don't usually	3924	correct watcher to remove. The lists are usually short (you don't usually
2584	have many watchers waiting for the same fd or signal).	3925	have many watchers waiting for the same fd or signal: one is typical, two
		3926	is rare).
2585		3927
2586	=item Finding the next timer per loop iteration: O(1)	3928	=item Finding the next timer in each loop iteration: O(1)
		3929
		3930	By virtue of using a binary or 4-heap, the next timer is always found at a
		3931	fixed position in the storage array.
2587		3932
2588	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)	3933	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)
2589		3934
2590	A change means an I/O watcher gets started or stopped, which requires	3935	A change means an I/O watcher gets started or stopped, which requires
2591	libev to recalculate its status (and possibly tell the kernel).	3936	libev to recalculate its status (and possibly tell the kernel, depending
		3937	on backend and whether C<ev_io_set> was used).
2592		3938
2593	=item Activating one watcher: O(1)	3939	=item Activating one watcher (putting it into the pending state): O(1)
2594		3940
2595	=item Priority handling: O(number_of_priorities)	3941	=item Priority handling: O(number_of_priorities)
2596		3942
2597	Priorities are implemented by allocating some space for each	3943	Priorities are implemented by allocating some space for each
2598	priority. When doing priority-based operations, libev usually has to	3944	priority. When doing priority-based operations, libev usually has to
2599	linearly search all the priorities.	3945	linearly search all the priorities, but starting/stopping and activating
		3946	watchers becomes O(1) with respect to priority handling.
		3947
		3948	=item Sending an ev_async: O(1)
		3949
		3950	=item Processing ev_async_send: O(number_of_async_watchers)
		3951
		3952	=item Processing signals: O(max_signal_number)
		3953
		3954	Sending involves a system call I<iff> there were no other C<ev_async_send>
		3955	calls in the current loop iteration. Checking for async and signal events
		3956	involves iterating over all running async watchers or all signal numbers.
2600		3957
2601	=back	3958	=back
2602		3959
2603		3960
2604	=head1 AUTHOR	3961	=head1 AUTHOR
2605		3962
2606	Marc Lehmann <libev@schmorp.de>.	3963	Marc Lehmann <libev@schmorp.de>, with repeated corrections by Mikael Magnusson.
2607		3964

Diff Legend

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