[ViewVC] Diff of: cvs/libev/ev.pod

Comparing libev/ev.pod (file contents):
Revision 1.172 by root, Wed Aug 6 07:01:25 2008 UTC vs.
Revision 1.366 by sf-exg, Thu Feb 3 16:21:08 2011 UTC

…		…
9	=head2 EXAMPLE PROGRAM	9	=head2 EXAMPLE PROGRAM
10		10
11	// a single header file is required	11	// a single header file is required
12	#include <ev.h>	12	#include <ev.h>
13		13
		14	#include <stdio.h> // for puts
		15
14	// every watcher type has its own typedef'd struct	16	// every watcher type has its own typedef'd struct
15	// with the name ev_<type>	17	// with the name ev_TYPE
16	ev_io stdin_watcher;	18	ev_io stdin_watcher;
17	ev_timer timeout_watcher;	19	ev_timer timeout_watcher;
18		20
19	// all watcher callbacks have a similar signature	21	// all watcher callbacks have a similar signature
20	// this callback is called when data is readable on stdin	22	// this callback is called when data is readable on stdin
21	static void	23	static void
22	stdin_cb (EV_P_ struct ev_io *w, int revents)	24	stdin_cb (EV_P_ ev_io *w, int revents)
23	{	25	{
24	puts ("stdin ready");	26	puts ("stdin ready");
25	// for one-shot events, one must manually stop the watcher	27	// for one-shot events, one must manually stop the watcher
26	// with its corresponding stop function.	28	// with its corresponding stop function.
27	ev_io_stop (EV_A_ w);	29	ev_io_stop (EV_A_ w);
28		30
29	// this causes all nested ev_loop's to stop iterating	31	// this causes all nested ev_run's to stop iterating
30	ev_unloop (EV_A_ EVUNLOOP_ALL);	32	ev_break (EV_A_ EVBREAK_ALL);
31	}	33	}
32		34
33	// another callback, this time for a time-out	35	// another callback, this time for a time-out
34	static void	36	static void
35	timeout_cb (EV_P_ struct ev_timer *w, int revents)	37	timeout_cb (EV_P_ ev_timer *w, int revents)
36	{	38	{
37	puts ("timeout");	39	puts ("timeout");
38	// this causes the innermost ev_loop to stop iterating	40	// this causes the innermost ev_run to stop iterating
39	ev_unloop (EV_A_ EVUNLOOP_ONE);	41	ev_break (EV_A_ EVBREAK_ONE);
40	}	42	}
41		43
42	int	44	int
43	main (void)	45	main (void)
44	{	46	{
45	// use the default event loop unless you have special needs	47	// use the default event loop unless you have special needs
46	struct ev_loop *loop = ev_default_loop (0);	48	struct ev_loop *loop = EV_DEFAULT;
47		49
48	// initialise an io watcher, then start it	50	// initialise an io watcher, then start it
49	// this one will watch for stdin to become readable	51	// this one will watch for stdin to become readable
50	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);
51	ev_io_start (loop, &stdin_watcher);	53	ev_io_start (loop, &stdin_watcher);
…		…
54	// simple non-repeating 5.5 second timeout	56	// simple non-repeating 5.5 second timeout
55	ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);	57	ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);
56	ev_timer_start (loop, &timeout_watcher);	58	ev_timer_start (loop, &timeout_watcher);
57		59
58	// now wait for events to arrive	60	// now wait for events to arrive
59	ev_loop (loop, 0);	61	ev_run (loop, 0);
60		62
61	// unloop was called, so exit	63	// break was called, so exit
62	return 0;	64	return 0;
63	}	65	}
64		66
65	=head1 DESCRIPTION	67	=head1 ABOUT THIS DOCUMENT
		68
		69	This document documents the libev software package.
66		70
67	The newest version of this document is also available as an html-formatted	71	The newest version of this document is also available as an html-formatted
68	web page you might find easier to navigate when reading it for the first	72	web page you might find easier to navigate when reading it for the first
69	time: L<http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod>.	73	time: L<http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod>.
		74
		75	While this document tries to be as complete as possible in documenting
		76	libev, its usage and the rationale behind its design, it is not a tutorial
		77	on event-based programming, nor will it introduce event-based programming
		78	with libev.
		79
		80	Familiarity with event based programming techniques in general is assumed
		81	throughout this document.
		82
		83	=head1 WHAT TO READ WHEN IN A HURRY
		84
		85	This manual tries to be very detailed, but unfortunately, this also makes
		86	it very long. If you just want to know the basics of libev, I suggest
		87	reading L<ANATOMY OF A WATCHER>, then the L<EXAMPLE PROGRAM> above and
		88	look up the missing functions in L<GLOBAL FUNCTIONS> and the C<ev_io> and
		89	C<ev_timer> sections in L<WATCHER TYPES>.
		90
		91	=head1 ABOUT LIBEV
70		92
71	Libev is an event loop: you register interest in certain events (such as a	93	Libev is an event loop: you register interest in certain events (such as a
72	file descriptor being readable or a timeout occurring), and it will manage	94	file descriptor being readable or a timeout occurring), and it will manage
73	these event sources and provide your program with events.	95	these event sources and provide your program with events.
74		96
…		…
84	=head2 FEATURES	106	=head2 FEATURES
85		107
86	Libev supports C<select>, C<poll>, the Linux-specific C<epoll>, the	108	Libev supports C<select>, C<poll>, the Linux-specific C<epoll>, the
87	BSD-specific C<kqueue> and the Solaris-specific event port mechanisms	109	BSD-specific C<kqueue> and the Solaris-specific event port mechanisms
88	for file descriptor events (C<ev_io>), the Linux C<inotify> interface	110	for file descriptor events (C<ev_io>), the Linux C<inotify> interface
89	(for C<ev_stat>), relative timers (C<ev_timer>), absolute timers	111	(for C<ev_stat>), Linux eventfd/signalfd (for faster and cleaner
90	with customised rescheduling (C<ev_periodic>), synchronous signals	112	inter-thread wakeup (C<ev_async>)/signal handling (C<ev_signal>)) relative
91	(C<ev_signal>), process status change events (C<ev_child>), and event	113	timers (C<ev_timer>), absolute timers with customised rescheduling
92	watchers dealing with the event loop mechanism itself (C<ev_idle>,	114	(C<ev_periodic>), synchronous signals (C<ev_signal>), process status
93	C<ev_embed>, C<ev_prepare> and C<ev_check> watchers) as well as	115	change events (C<ev_child>), and event watchers dealing with the event
94	file watchers (C<ev_stat>) and even limited support for fork events	116	loop mechanism itself (C<ev_idle>, C<ev_embed>, C<ev_prepare> and
95	(C<ev_fork>).	117	C<ev_check> watchers) as well as file watchers (C<ev_stat>) and even
		118	limited support for fork events (C<ev_fork>).
96		119
97	It also is quite fast (see this	120	It also is quite fast (see this
98	L<benchmark\|http://libev.schmorp.de/bench.html> comparing it to libevent	121	L<benchmark\|http://libev.schmorp.de/bench.html> comparing it to libevent
99	for example).	122	for example).
100		123
…		…
108	name C<loop> (which is always of type C<struct ev_loop *>) will not have	131	name C<loop> (which is always of type C<struct ev_loop *>) will not have
109	this argument.	132	this argument.
110		133
111	=head2 TIME REPRESENTATION	134	=head2 TIME REPRESENTATION
112		135
113	Libev represents time as a single floating point number, representing the	136	Libev represents time as a single floating point number, representing
114	(fractional) number of seconds since the (POSIX) epoch (somewhere near	137	the (fractional) number of seconds since the (POSIX) epoch (in practice
115	the beginning of 1970, details are complicated, don't ask). This type is	138	somewhere near the beginning of 1970, details are complicated, don't
116	called C<ev_tstamp>, which is what you should use too. It usually aliases	139	ask). This type is called C<ev_tstamp>, which is what you should use
117	to the C<double> type in C, and when you need to do any calculations on	140	too. It usually aliases to the C<double> type in C. When you need to do
118	it, you should treat it as some floating point value. Unlike the name	141	any calculations on it, you should treat it as some floating point value.
		142
119	component C<stamp> might indicate, it is also used for time differences	143	Unlike the name component C<stamp> might indicate, it is also used for
120	throughout libev.	144	time differences (e.g. delays) throughout libev.
121		145
122	=head1 ERROR HANDLING	146	=head1 ERROR HANDLING
123		147
124	Libev knows three classes of errors: operating system errors, usage errors	148	Libev knows three classes of errors: operating system errors, usage errors
125	and internal errors (bugs).	149	and internal errors (bugs).
…		…
149		173
150	=item ev_tstamp ev_time ()	174	=item ev_tstamp ev_time ()
151		175
152	Returns the current time as libev would use it. Please note that the	176	Returns the current time as libev would use it. Please note that the
153	C<ev_now> function is usually faster and also often returns the timestamp	177	C<ev_now> function is usually faster and also often returns the timestamp
154	you actually want to know.	178	you actually want to know. Also interesting is the combination of
		179	C<ev_update_now> and C<ev_now>.
155		180
156	=item ev_sleep (ev_tstamp interval)	181	=item ev_sleep (ev_tstamp interval)
157		182
158	Sleep for the given interval: The current thread will be blocked until	183	Sleep for the given interval: The current thread will be blocked until
159	either it is interrupted or the given time interval has passed. Basically	184	either it is interrupted or the given time interval has passed. Basically
…		…
176	as this indicates an incompatible change. Minor versions are usually	201	as this indicates an incompatible change. Minor versions are usually
177	compatible to older versions, so a larger minor version alone is usually	202	compatible to older versions, so a larger minor version alone is usually
178	not a problem.	203	not a problem.
179		204
180	Example: Make sure we haven't accidentally been linked against the wrong	205	Example: Make sure we haven't accidentally been linked against the wrong
181	version.	206	version (note, however, that this will not detect other ABI mismatches,
		207	such as LFS or reentrancy).
182		208
183	assert (("libev version mismatch",	209	assert (("libev version mismatch",
184	ev_version_major () == EV_VERSION_MAJOR	210	ev_version_major () == EV_VERSION_MAJOR
185	&& ev_version_minor () >= EV_VERSION_MINOR));	211	&& ev_version_minor () >= EV_VERSION_MINOR));
186		212
…		…
197	assert (("sorry, no epoll, no sex",	223	assert (("sorry, no epoll, no sex",
198	ev_supported_backends () & EVBACKEND_EPOLL));	224	ev_supported_backends () & EVBACKEND_EPOLL));
199		225
200	=item unsigned int ev_recommended_backends ()	226	=item unsigned int ev_recommended_backends ()
201		227
202	Return the set of all backends compiled into this binary of libev and also	228	Return the set of all backends compiled into this binary of libev and
203	recommended for this platform. This set is often smaller than the one	229	also recommended for this platform, meaning it will work for most file
		230	descriptor types. This set is often smaller than the one returned by
204	returned by C<ev_supported_backends>, as for example kqueue is broken on	231	C<ev_supported_backends>, as for example kqueue is broken on most BSDs
205	most BSDs and will not be auto-detected unless you explicitly request it	232	and will not be auto-detected unless you explicitly request it (assuming
206	(assuming you know what you are doing). This is the set of backends that	233	you know what you are doing). This is the set of backends that libev will
207	libev will probe for if you specify no backends explicitly.	234	probe for if you specify no backends explicitly.
208		235
209	=item unsigned int ev_embeddable_backends ()	236	=item unsigned int ev_embeddable_backends ()
210		237
211	Returns the set of backends that are embeddable in other event loops. This	238	Returns the set of backends that are embeddable in other event loops. This
212	is the theoretical, all-platform, value. To find which backends	239	value is platform-specific but can include backends not available on the
213	might be supported on the current system, you would need to look at	240	current system. To find which embeddable backends might be supported on
214	C<ev_embeddable_backends () & ev_supported_backends ()>, likewise for	241	the current system, you would need to look at C<ev_embeddable_backends ()
215	recommended ones.	242	& ev_supported_backends ()>, likewise for recommended ones.
216		243
217	See the description of C<ev_embed> watchers for more info.	244	See the description of C<ev_embed> watchers for more info.
218		245
219	=item ev_set_allocator (void (cb)(void *ptr, long size))	246	=item ev_set_allocator (void (cb)(void *ptr, long size))
220		247
…		…
250	}	277	}
251		278
252	...	279	...
253	ev_set_allocator (persistent_realloc);	280	ev_set_allocator (persistent_realloc);
254		281
255	=item ev_set_syserr_cb (void (cb)(const char msg));	282	=item ev_set_syserr_cb (void (cb)(const char msg))
256		283
257	Set the callback function to call on a retryable system call error (such	284	Set the callback function to call on a retryable system call error (such
258	as failed select, poll, epoll_wait). The message is a printable string	285	as failed select, poll, epoll_wait). The message is a printable string
259	indicating the system call or subsystem causing the problem. If this	286	indicating the system call or subsystem causing the problem. If this
260	callback is set, then libev will expect it to remedy the situation, no	287	callback is set, then libev will expect it to remedy the situation, no
…		…
272	}	299	}
273		300
274	...	301	...
275	ev_set_syserr_cb (fatal_error);	302	ev_set_syserr_cb (fatal_error);
276		303
		304	=item ev_feed_signal (int signum)
		305
		306	This function can be used to "simulate" a signal receive. It is completely
		307	safe to call this function at any time, from any context, including signal
		308	handlers or random threads.
		309
		310	Its main use is to customise signal handling in your process, especially
		311	in the presence of threads. For example, you could block signals
		312	by default in all threads (and specifying C<EVFLAG_NOSIGMASK> when
		313	creating any loops), and in one thread, use C<sigwait> or any other
		314	mechanism to wait for signals, then "deliver" them to libev by calling
		315	C<ev_feed_signal>.
		316
277	=back	317	=back
278		318
279	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP	319	=head1 FUNCTIONS CONTROLLING EVENT LOOPS
280		320
281	An event loop is described by a C<struct ev_loop *>. The library knows two	321	An event loop is described by a C<struct ev_loop *> (the C<struct> is
282	types of such loops, the I<default> loop, which supports signals and child	322	I<not> optional in this case unless libev 3 compatibility is disabled, as
283	events, and dynamically created loops which do not.	323	libev 3 had an C<ev_loop> function colliding with the struct name).
		324
		325	The library knows two types of such loops, the I<default> loop, which
		326	supports child process events, and dynamically created event loops which
		327	do not.
284		328
285	=over 4	329	=over 4
286		330
287	=item struct ev_loop *ev_default_loop (unsigned int flags)	331	=item struct ev_loop *ev_default_loop (unsigned int flags)
288		332
289	This will initialise the default event loop if it hasn't been initialised	333	This returns the "default" event loop object, which is what you should
290	yet and return it. If the default loop could not be initialised, returns	334	normally use when you just need "the event loop". Event loop objects and
291	false. If it already was initialised it simply returns it (and ignores the	335	the C<flags> parameter are described in more detail in the entry for
292	flags. If that is troubling you, check C<ev_backend ()> afterwards).	336	C<ev_loop_new>.
		337
		338	If the default loop is already initialised then this function simply
		339	returns it (and ignores the flags. If that is troubling you, check
		340	C<ev_backend ()> afterwards). Otherwise it will create it with the given
		341	flags, which should almost always be C<0>, unless the caller is also the
		342	one calling C<ev_run> or otherwise qualifies as "the main program".
293		343
294	If you don't know what event loop to use, use the one returned from this	344	If you don't know what event loop to use, use the one returned from this
295	function.	345	function (or via the C<EV_DEFAULT> macro).
296		346
297	Note that this function is I<not> thread-safe, so if you want to use it	347	Note that this function is I<not> thread-safe, so if you want to use it
298	from multiple threads, you have to lock (note also that this is unlikely,	348	from multiple threads, you have to employ some kind of mutex (note also
299	as loops cannot bes hared easily between threads anyway).	349	that this case is unlikely, as loops cannot be shared easily between
		350	threads anyway).
300		351
301	The default loop is the only loop that can handle C<ev_signal> and	352	The default loop is the only loop that can handle C<ev_child> watchers,
302	C<ev_child> watchers, and to do this, it always registers a handler	353	and to do this, it always registers a handler for C<SIGCHLD>. If this is
303	for C<SIGCHLD>. If this is a problem for your application you can either	354	a problem for your application you can either create a dynamic loop with
304	create a dynamic loop with C<ev_loop_new> that doesn't do that, or you	355	C<ev_loop_new> which doesn't do that, or you can simply overwrite the
305	can simply overwrite the C<SIGCHLD> signal handler I<after> calling	356	C<SIGCHLD> signal handler I<after> calling C<ev_default_init>.
306	C<ev_default_init>.	357
		358	Example: This is the most typical usage.
		359
		360	if (!ev_default_loop (0))
		361	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
		362
		363	Example: Restrict libev to the select and poll backends, and do not allow
		364	environment settings to be taken into account:
		365
		366	ev_default_loop (EVBACKEND_POLL \| EVBACKEND_SELECT \| EVFLAG_NOENV);
		367
		368	=item struct ev_loop *ev_loop_new (unsigned int flags)
		369
		370	This will create and initialise a new event loop object. If the loop
		371	could not be initialised, returns false.
		372
		373	This function is thread-safe, and one common way to use libev with
		374	threads is indeed to create one loop per thread, and using the default
		375	loop in the "main" or "initial" thread.
307		376
308	The flags argument can be used to specify special behaviour or specific	377	The flags argument can be used to specify special behaviour or specific
309	backends to use, and is usually specified as C<0> (or C<EVFLAG_AUTO>).	378	backends to use, and is usually specified as C<0> (or C<EVFLAG_AUTO>).
310		379
311	The following flags are supported:	380	The following flags are supported:
…		…
326	useful to try out specific backends to test their performance, or to work	395	useful to try out specific backends to test their performance, or to work
327	around bugs.	396	around bugs.
328		397
329	=item C<EVFLAG_FORKCHECK>	398	=item C<EVFLAG_FORKCHECK>
330		399
331	Instead of calling C<ev_default_fork> or C<ev_loop_fork> manually after	400	Instead of calling C<ev_loop_fork> manually after a fork, you can also
332	a fork, you can also make libev check for a fork in each iteration by	401	make libev check for a fork in each iteration by enabling this flag.
333	enabling this flag.
334		402
335	This works by calling C<getpid ()> on every iteration of the loop,	403	This works by calling C<getpid ()> on every iteration of the loop,
336	and thus this might slow down your event loop if you do a lot of loop	404	and thus this might slow down your event loop if you do a lot of loop
337	iterations and little real work, but is usually not noticeable (on my	405	iterations and little real work, but is usually not noticeable (on my
338	GNU/Linux system for example, C<getpid> is actually a simple 5-insn sequence	406	GNU/Linux system for example, C<getpid> is actually a simple 5-insn sequence
…		…
344	flag.	412	flag.
345		413
346	This flag setting cannot be overridden or specified in the C<LIBEV_FLAGS>	414	This flag setting cannot be overridden or specified in the C<LIBEV_FLAGS>
347	environment variable.	415	environment variable.
348		416
		417	=item C<EVFLAG_NOINOTIFY>
		418
		419	When this flag is specified, then libev will not attempt to use the
		420	I<inotify> API for its C<ev_stat> watchers. Apart from debugging and
		421	testing, this flag can be useful to conserve inotify file descriptors, as
		422	otherwise each loop using C<ev_stat> watchers consumes one inotify handle.
		423
		424	=item C<EVFLAG_SIGNALFD>
		425
		426	When this flag is specified, then libev will attempt to use the
		427	I<signalfd> API for its C<ev_signal> (and C<ev_child>) watchers. This API
		428	delivers signals synchronously, which makes it both faster and might make
		429	it possible to get the queued signal data. It can also simplify signal
		430	handling with threads, as long as you properly block signals in your
		431	threads that are not interested in handling them.
		432
		433	Signalfd will not be used by default as this changes your signal mask, and
		434	there are a lot of shoddy libraries and programs (glib's threadpool for
		435	example) that can't properly initialise their signal masks.
		436
		437	=item C<EVFLAG_NOSIGMASK>
		438
		439	When this flag is specified, then libev will avoid to modify the signal
		440	mask. Specifically, this means you ahve to make sure signals are unblocked
		441	when you want to receive them.
		442
		443	This behaviour is useful when you want to do your own signal handling, or
		444	want to handle signals only in specific threads and want to avoid libev
		445	unblocking the signals.
		446
		447	It's also required by POSIX in a threaded program, as libev calls
		448	C<sigprocmask>, whose behaviour is officially unspecified.
		449
		450	This flag's behaviour will become the default in future versions of libev.
		451
349	=item C<EVBACKEND_SELECT> (value 1, portable select backend)	452	=item C<EVBACKEND_SELECT> (value 1, portable select backend)
350		453
351	This is your standard select(2) backend. Not I<completely> standard, as	454	This is your standard select(2) backend. Not I<completely> standard, as
352	libev tries to roll its own fd_set with no limits on the number of fds,	455	libev tries to roll its own fd_set with no limits on the number of fds,
353	but if that fails, expect a fairly low limit on the number of fds when	456	but if that fails, expect a fairly low limit on the number of fds when
…		…
359	writing a server, you should C<accept ()> in a loop to accept as many	462	writing a server, you should C<accept ()> in a loop to accept as many
360	connections as possible during one iteration. You might also want to have	463	connections as possible during one iteration. You might also want to have
361	a look at C<ev_set_io_collect_interval ()> to increase the amount of	464	a look at C<ev_set_io_collect_interval ()> to increase the amount of
362	readiness notifications you get per iteration.	465	readiness notifications you get per iteration.
363		466
		467	This backend maps C<EV_READ> to the C<readfds> set and C<EV_WRITE> to the
		468	C<writefds> set (and to work around Microsoft Windows bugs, also onto the
		469	C<exceptfds> set on that platform).
		470
364	=item C<EVBACKEND_POLL> (value 2, poll backend, available everywhere except on windows)	471	=item C<EVBACKEND_POLL> (value 2, poll backend, available everywhere except on windows)
365		472
366	And this is your standard poll(2) backend. It's more complicated	473	And this is your standard poll(2) backend. It's more complicated
367	than select, but handles sparse fds better and has no artificial	474	than select, but handles sparse fds better and has no artificial
368	limit on the number of fds you can use (except it will slow down	475	limit on the number of fds you can use (except it will slow down
369	considerably with a lot of inactive fds). It scales similarly to select,	476	considerably with a lot of inactive fds). It scales similarly to select,
370	i.e. O(total_fds). See the entry for C<EVBACKEND_SELECT>, above, for	477	i.e. O(total_fds). See the entry for C<EVBACKEND_SELECT>, above, for
371	performance tips.	478	performance tips.
372		479
		480	This backend maps C<EV_READ> to C<POLLIN \| POLLERR \| POLLHUP>, and
		481	C<EV_WRITE> to C<POLLOUT \| POLLERR \| POLLHUP>.
		482
373	=item C<EVBACKEND_EPOLL> (value 4, Linux)	483	=item C<EVBACKEND_EPOLL> (value 4, Linux)
		484
		485	Use the linux-specific epoll(7) interface (for both pre- and post-2.6.9
		486	kernels).
374		487
375	For few fds, this backend is a bit little slower than poll and select,	488	For few fds, this backend is a bit little slower than poll and select,
376	but it scales phenomenally better. While poll and select usually scale	489	but it scales phenomenally better. While poll and select usually scale
377	like O(total_fds) where n is the total number of fds (or the highest fd),	490	like O(total_fds) where n is the total number of fds (or the highest fd),
378	epoll scales either O(1) or O(active_fds). The epoll design has a number	491	epoll scales either O(1) or O(active_fds).
379	of shortcomings, such as silently dropping events in some hard-to-detect	492
380	cases and requiring a system call per fd change, no fork support and bad	493	The epoll mechanism deserves honorable mention as the most misdesigned
381	support for dup.	494	of the more advanced event mechanisms: mere annoyances include silently
		495	dropping file descriptors, requiring a system call per change per file
		496	descriptor (and unnecessary guessing of parameters), problems with dup,
		497	returning before the timeout value, resulting in additional iterations
		498	(and only giving 5ms accuracy while select on the same platform gives
		499	0.1ms) and so on. The biggest issue is fork races, however - if a program
		500	forks then I<both> parent and child process have to recreate the epoll
		501	set, which can take considerable time (one syscall per file descriptor)
		502	and is of course hard to detect.
		503
		504	Epoll is also notoriously buggy - embedding epoll fds I<should> work, but
		505	of course I<doesn't>, and epoll just loves to report events for totally
		506	I<different> file descriptors (even already closed ones, so one cannot
		507	even remove them from the set) than registered in the set (especially
		508	on SMP systems). Libev tries to counter these spurious notifications by
		509	employing an additional generation counter and comparing that against the
		510	events to filter out spurious ones, recreating the set when required. Last
		511	not least, it also refuses to work with some file descriptors which work
		512	perfectly fine with C<select> (files, many character devices...).
		513
		514	Epoll is truly the train wreck analog among event poll mechanisms,
		515	a frankenpoll, cobbled together in a hurry, no thought to design or
		516	interaction with others.
382		517
383	While stopping, setting and starting an I/O watcher in the same iteration	518	While stopping, setting and starting an I/O watcher in the same iteration
384	will result in some caching, there is still a system call per such incident	519	will result in some caching, there is still a system call per such
385	(because the fd could point to a different file description now), so its	520	incident (because the same I<file descriptor> could point to a different
386	best to avoid that. Also, C<dup ()>'ed file descriptors might not work	521	I<file description> now), so its best to avoid that. Also, C<dup ()>'ed
387	very well if you register events for both fds.	522	file descriptors might not work very well if you register events for both
388		523	file descriptors.
389	Please note that epoll sometimes generates spurious notifications, so you
390	need to use non-blocking I/O or other means to avoid blocking when no data
391	(or space) is available.
392		524
393	Best performance from this backend is achieved by not unregistering all	525	Best performance from this backend is achieved by not unregistering all
394	watchers for a file descriptor until it has been closed, if possible, i.e.	526	watchers for a file descriptor until it has been closed, if possible,
395	keep at least one watcher active per fd at all times.	527	i.e. keep at least one watcher active per fd at all times. Stopping and
		528	starting a watcher (without re-setting it) also usually doesn't cause
		529	extra overhead. A fork can both result in spurious notifications as well
		530	as in libev having to destroy and recreate the epoll object, which can
		531	take considerable time and thus should be avoided.
		532
		533	All this means that, in practice, C<EVBACKEND_SELECT> can be as fast or
		534	faster than epoll for maybe up to a hundred file descriptors, depending on
		535	the usage. So sad.
396		536
397	While nominally embeddable in other event loops, this feature is broken in	537	While nominally embeddable in other event loops, this feature is broken in
398	all kernel versions tested so far.	538	all kernel versions tested so far.
		539
		540	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
		541	C<EVBACKEND_POLL>.
399		542
400	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)	543	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)
401		544
402	Kqueue deserves special mention, as at the time of this writing, it	545	Kqueue deserves special mention, as at the time of this writing, it
403	was broken on all BSDs except NetBSD (usually it doesn't work reliably	546	was broken on all BSDs except NetBSD (usually it doesn't work reliably
404	with anything but sockets and pipes, except on Darwin, where of course	547	with anything but sockets and pipes, except on Darwin, where of course
405	it's completely useless). For this reason it's not being "auto-detected"	548	it's completely useless). Unlike epoll, however, whose brokenness
		549	is by design, these kqueue bugs can (and eventually will) be fixed
		550	without API changes to existing programs. For this reason it's not being
406	unless you explicitly specify it explicitly in the flags (i.e. using	551	"auto-detected" unless you explicitly specify it in the flags (i.e. using
407	C<EVBACKEND_KQUEUE>) or libev was compiled on a known-to-be-good (-enough)	552	C<EVBACKEND_KQUEUE>) or libev was compiled on a known-to-be-good (-enough)
408	system like NetBSD.	553	system like NetBSD.
409		554
410	You still can embed kqueue into a normal poll or select backend and use it	555	You still can embed kqueue into a normal poll or select backend and use it
411	only for sockets (after having made sure that sockets work with kqueue on	556	only for sockets (after having made sure that sockets work with kqueue on
…		…
413		558
414	It scales in the same way as the epoll backend, but the interface to the	559	It scales in the same way as the epoll backend, but the interface to the
415	kernel is more efficient (which says nothing about its actual speed, of	560	kernel is more efficient (which says nothing about its actual speed, of
416	course). While stopping, setting and starting an I/O watcher does never	561	course). While stopping, setting and starting an I/O watcher does never
417	cause an extra system call as with C<EVBACKEND_EPOLL>, it still adds up to	562	cause an extra system call as with C<EVBACKEND_EPOLL>, it still adds up to
418	two event changes per incident, support for C<fork ()> is very bad and it	563	two event changes per incident. Support for C<fork ()> is very bad (but
419	drops fds silently in similarly hard-to-detect cases.	564	sane, unlike epoll) and it drops fds silently in similarly hard-to-detect
		565	cases
420		566
421	This backend usually performs well under most conditions.	567	This backend usually performs well under most conditions.
422		568
423	While nominally embeddable in other event loops, this doesn't work	569	While nominally embeddable in other event loops, this doesn't work
424	everywhere, so you might need to test for this. And since it is broken	570	everywhere, so you might need to test for this. And since it is broken
425	almost everywhere, you should only use it when you have a lot of sockets	571	almost everywhere, you should only use it when you have a lot of sockets
426	(for which it usually works), by embedding it into another event loop	572	(for which it usually works), by embedding it into another event loop
427	(e.g. C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>) and using it only for	573	(e.g. C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> (but C<poll> is of course
428	sockets.	574	also broken on OS X)) and, did I mention it, using it only for sockets.
		575
		576	This backend maps C<EV_READ> into an C<EVFILT_READ> kevent with
		577	C<NOTE_EOF>, and C<EV_WRITE> into an C<EVFILT_WRITE> kevent with
		578	C<NOTE_EOF>.
429		579
430	=item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8)	580	=item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8)
431		581
432	This is not implemented yet (and might never be, unless you send me an	582	This is not implemented yet (and might never be, unless you send me an
433	implementation). According to reports, C</dev/poll> only supports sockets	583	implementation). According to reports, C</dev/poll> only supports sockets
…		…
437	=item C<EVBACKEND_PORT> (value 32, Solaris 10)	587	=item C<EVBACKEND_PORT> (value 32, Solaris 10)
438		588
439	This uses the Solaris 10 event port mechanism. As with everything on Solaris,	589	This uses the Solaris 10 event port mechanism. As with everything on Solaris,
440	it's really slow, but it still scales very well (O(active_fds)).	590	it's really slow, but it still scales very well (O(active_fds)).
441		591
442	Please note that Solaris event ports can deliver a lot of spurious
443	notifications, so you need to use non-blocking I/O or other means to avoid
444	blocking when no data (or space) is available.
445
446	While this backend scales well, it requires one system call per active	592	While this backend scales well, it requires one system call per active
447	file descriptor per loop iteration. For small and medium numbers of file	593	file descriptor per loop iteration. For small and medium numbers of file
448	descriptors a "slow" C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> backend	594	descriptors a "slow" C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> backend
449	might perform better.	595	might perform better.
450		596
451	On the positive side, ignoring the spurious readiness notifications, this	597	On the positive side, this backend actually performed fully to
452	backend actually performed to specification in all tests and is fully	598	specification in all tests and is fully embeddable, which is a rare feat
453	embeddable, which is a rare feat among the OS-specific backends.	599	among the OS-specific backends (I vastly prefer correctness over speed
		600	hacks).
		601
		602	On the negative side, the interface is I<bizarre> - so bizarre that
		603	even sun itself gets it wrong in their code examples: The event polling
		604	function sometimes returning events to the caller even though an error
		605	occurred, but with no indication whether it has done so or not (yes, it's
		606	even documented that way) - deadly for edge-triggered interfaces where
		607	you absolutely have to know whether an event occurred or not because you
		608	have to re-arm the watcher.
		609
		610	Fortunately libev seems to be able to work around these idiocies.
		611
		612	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
		613	C<EVBACKEND_POLL>.
454		614
455	=item C<EVBACKEND_ALL>	615	=item C<EVBACKEND_ALL>
456		616
457	Try all backends (even potentially broken ones that wouldn't be tried	617	Try all backends (even potentially broken ones that wouldn't be tried
458	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as	618	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as
459	C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>.	619	C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>.
460		620
461	It is definitely not recommended to use this flag.	621	It is definitely not recommended to use this flag, use whatever
		622	C<ev_recommended_backends ()> returns, or simply do not specify a backend
		623	at all.
		624
		625	=item C<EVBACKEND_MASK>
		626
		627	Not a backend at all, but a mask to select all backend bits from a
		628	C<flags> value, in case you want to mask out any backends from a flags
		629	value (e.g. when modifying the C<LIBEV_FLAGS> environment variable).
462		630
463	=back	631	=back
464		632
465	If one or more of these are or'ed into the flags value, then only these	633	If one or more of the backend flags are or'ed into the flags value,
466	backends will be tried (in the reverse order as listed here). If none are	634	then only these backends will be tried (in the reverse order as listed
467	specified, all backends in C<ev_recommended_backends ()> will be tried.	635	here). If none are specified, all backends in C<ev_recommended_backends
468		636	()> will be tried.
469	The most typical usage is like this:
470
471	if (!ev_default_loop (0))
472	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
473
474	Restrict libev to the select and poll backends, and do not allow
475	environment settings to be taken into account:
476
477	ev_default_loop (EVBACKEND_POLL \| EVBACKEND_SELECT \| EVFLAG_NOENV);
478
479	Use whatever libev has to offer, but make sure that kqueue is used if
480	available (warning, breaks stuff, best use only with your own private
481	event loop and only if you know the OS supports your types of fds):
482
483	ev_default_loop (ev_recommended_backends () \| EVBACKEND_KQUEUE);
484
485	=item struct ev_loop *ev_loop_new (unsigned int flags)
486
487	Similar to C<ev_default_loop>, but always creates a new event loop that is
488	always distinct from the default loop. Unlike the default loop, it cannot
489	handle signal and child watchers, and attempts to do so will be greeted by
490	undefined behaviour (or a failed assertion if assertions are enabled).
491
492	Note that this function I<is> thread-safe, and the recommended way to use
493	libev with threads is indeed to create one loop per thread, and using the
494	default loop in the "main" or "initial" thread.
495		637
496	Example: Try to create a event loop that uses epoll and nothing else.	638	Example: Try to create a event loop that uses epoll and nothing else.
497		639
498	struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL \| EVFLAG_NOENV);	640	struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL \| EVFLAG_NOENV);
499	if (!epoller)	641	if (!epoller)
500	fatal ("no epoll found here, maybe it hides under your chair");	642	fatal ("no epoll found here, maybe it hides under your chair");
501		643
		644	Example: Use whatever libev has to offer, but make sure that kqueue is
		645	used if available.
		646
		647	struct ev_loop *loop = ev_loop_new (ev_recommended_backends () \| EVBACKEND_KQUEUE);
		648
502	=item ev_default_destroy ()	649	=item ev_loop_destroy (loop)
503		650
504	Destroys the default loop again (frees all memory and kernel state	651	Destroys an event loop object (frees all memory and kernel state
505	etc.). None of the active event watchers will be stopped in the normal	652	etc.). None of the active event watchers will be stopped in the normal
506	sense, so e.g. C<ev_is_active> might still return true. It is your	653	sense, so e.g. C<ev_is_active> might still return true. It is your
507	responsibility to either stop all watchers cleanly yourself I<before>	654	responsibility to either stop all watchers cleanly yourself I<before>
508	calling this function, or cope with the fact afterwards (which is usually	655	calling this function, or cope with the fact afterwards (which is usually
509	the easiest thing, you can just ignore the watchers and/or C<free ()> them	656	the easiest thing, you can just ignore the watchers and/or C<free ()> them
510	for example).	657	for example).
511		658
512	Note that certain global state, such as signal state, will not be freed by	659	Note that certain global state, such as signal state (and installed signal
513	this function, and related watchers (such as signal and child watchers)	660	handlers), will not be freed by this function, and related watchers (such
514	would need to be stopped manually.	661	as signal and child watchers) would need to be stopped manually.
515		662
516	In general it is not advisable to call this function except in the	663	This function is normally used on loop objects allocated by
517	rare occasion where you really need to free e.g. the signal handling	664	C<ev_loop_new>, but it can also be used on the default loop returned by
		665	C<ev_default_loop>, in which case it is not thread-safe.
		666
		667	Note that it is not advisable to call this function on the default loop
		668	except in the rare occasion where you really need to free its resources.
518	pipe fds. If you need dynamically allocated loops it is better to use	669	If you need dynamically allocated loops it is better to use C<ev_loop_new>
519	C<ev_loop_new> and C<ev_loop_destroy>).	670	and C<ev_loop_destroy>.
520		671
521	=item ev_loop_destroy (loop)	672	=item ev_loop_fork (loop)
522		673
523	Like C<ev_default_destroy>, but destroys an event loop created by an
524	earlier call to C<ev_loop_new>.
525
526	=item ev_default_fork ()
527
528	This function sets a flag that causes subsequent C<ev_loop> iterations	674	This function sets a flag that causes subsequent C<ev_run> iterations to
529	to reinitialise the kernel state for backends that have one. Despite the	675	reinitialise the kernel state for backends that have one. Despite the
530	name, you can call it anytime, but it makes most sense after forking, in	676	name, you can call it anytime, but it makes most sense after forking, in
531	the child process (or both child and parent, but that again makes little	677	the child process. You I<must> call it (or use C<EVFLAG_FORKCHECK>) in the
532	sense). You I<must> call it in the child before using any of the libev	678	child before resuming or calling C<ev_run>.
533	functions, and it will only take effect at the next C<ev_loop> iteration.	679
		680	Again, you I<have> to call it on I<any> loop that you want to re-use after
		681	a fork, I<even if you do not plan to use the loop in the parent>. This is
		682	because some kernel interfaces cough I<kqueue> cough do funny things
		683	during fork.
534		684
535	On the other hand, you only need to call this function in the child	685	On the other hand, you only need to call this function in the child
536	process if and only if you want to use the event library in the child. If	686	process if and only if you want to use the event loop in the child. If
537	you just fork+exec, you don't have to call it at all.	687	you just fork+exec or create a new loop in the child, you don't have to
		688	call it at all (in fact, C<epoll> is so badly broken that it makes a
		689	difference, but libev will usually detect this case on its own and do a
		690	costly reset of the backend).
538		691
539	The function itself is quite fast and it's usually not a problem to call	692	The function itself is quite fast and it's usually not a problem to call
540	it just in case after a fork. To make this easy, the function will fit in	693	it just in case after a fork.
541	quite nicely into a call to C<pthread_atfork>:
542		694
		695	Example: Automate calling C<ev_loop_fork> on the default loop when
		696	using pthreads.
		697
		698	static void
		699	post_fork_child (void)
		700	{
		701	ev_loop_fork (EV_DEFAULT);
		702	}
		703
		704	...
543	pthread_atfork (0, 0, ev_default_fork);	705	pthread_atfork (0, 0, post_fork_child);
544
545	=item ev_loop_fork (loop)
546
547	Like C<ev_default_fork>, but acts on an event loop created by
548	C<ev_loop_new>. Yes, you have to call this on every allocated event loop
549	after fork, and how you do this is entirely your own problem.
550		706
551	=item int ev_is_default_loop (loop)	707	=item int ev_is_default_loop (loop)
552		708
553	Returns true when the given loop actually is the default loop, false otherwise.	709	Returns true when the given loop is, in fact, the default loop, and false
		710	otherwise.
554		711
555	=item unsigned int ev_loop_count (loop)	712	=item unsigned int ev_iteration (loop)
556		713
557	Returns the count of loop iterations for the loop, which is identical to	714	Returns the current iteration count for the event loop, which is identical
558	the number of times libev did poll for new events. It starts at C<0> and	715	to the number of times libev did poll for new events. It starts at C<0>
559	happily wraps around with enough iterations.	716	and happily wraps around with enough iterations.
560		717
561	This value can sometimes be useful as a generation counter of sorts (it	718	This value can sometimes be useful as a generation counter of sorts (it
562	"ticks" the number of loop iterations), as it roughly corresponds with	719	"ticks" the number of loop iterations), as it roughly corresponds with
563	C<ev_prepare> and C<ev_check> calls.	720	C<ev_prepare> and C<ev_check> calls - and is incremented between the
		721	prepare and check phases.
		722
		723	=item unsigned int ev_depth (loop)
		724
		725	Returns the number of times C<ev_run> was entered minus the number of
		726	times C<ev_run> was exited normally, in other words, the recursion depth.
		727
		728	Outside C<ev_run>, this number is zero. In a callback, this number is
		729	C<1>, unless C<ev_run> was invoked recursively (or from another thread),
		730	in which case it is higher.
		731
		732	Leaving C<ev_run> abnormally (setjmp/longjmp, cancelling the thread,
		733	throwing an exception etc.), doesn't count as "exit" - consider this
		734	as a hint to avoid such ungentleman-like behaviour unless it's really
		735	convenient, in which case it is fully supported.
564		736
565	=item unsigned int ev_backend (loop)	737	=item unsigned int ev_backend (loop)
566		738
567	Returns one of the C<EVBACKEND_*> flags indicating the event backend in	739	Returns one of the C<EVBACKEND_*> flags indicating the event backend in
568	use.	740	use.
…		…
573	received events and started processing them. This timestamp does not	745	received events and started processing them. This timestamp does not
574	change as long as callbacks are being processed, and this is also the base	746	change as long as callbacks are being processed, and this is also the base
575	time used for relative timers. You can treat it as the timestamp of the	747	time used for relative timers. You can treat it as the timestamp of the
576	event occurring (or more correctly, libev finding out about it).	748	event occurring (or more correctly, libev finding out about it).
577		749
		750	=item ev_now_update (loop)
		751
		752	Establishes the current time by querying the kernel, updating the time
		753	returned by C<ev_now ()> in the progress. This is a costly operation and
		754	is usually done automatically within C<ev_run ()>.
		755
		756	This function is rarely useful, but when some event callback runs for a
		757	very long time without entering the event loop, updating libev's idea of
		758	the current time is a good idea.
		759
		760	See also L<The special problem of time updates> in the C<ev_timer> section.
		761
		762	=item ev_suspend (loop)
		763
		764	=item ev_resume (loop)
		765
		766	These two functions suspend and resume an event loop, for use when the
		767	loop is not used for a while and timeouts should not be processed.
		768
		769	A typical use case would be an interactive program such as a game: When
		770	the user presses C<^Z> to suspend the game and resumes it an hour later it
		771	would be best to handle timeouts as if no time had actually passed while
		772	the program was suspended. This can be achieved by calling C<ev_suspend>
		773	in your C<SIGTSTP> handler, sending yourself a C<SIGSTOP> and calling
		774	C<ev_resume> directly afterwards to resume timer processing.
		775
		776	Effectively, all C<ev_timer> watchers will be delayed by the time spend
		777	between C<ev_suspend> and C<ev_resume>, and all C<ev_periodic> watchers
		778	will be rescheduled (that is, they will lose any events that would have
		779	occurred while suspended).
		780
		781	After calling C<ev_suspend> you B<must not> call I<any> function on the
		782	given loop other than C<ev_resume>, and you B<must not> call C<ev_resume>
		783	without a previous call to C<ev_suspend>.
		784
		785	Calling C<ev_suspend>/C<ev_resume> has the side effect of updating the
		786	event loop time (see C<ev_now_update>).
		787
578	=item ev_loop (loop, int flags)	788	=item ev_run (loop, int flags)
579		789
580	Finally, this is it, the event handler. This function usually is called	790	Finally, this is it, the event handler. This function usually is called
581	after you initialised all your watchers and you want to start handling	791	after you have initialised all your watchers and you want to start
582	events.	792	handling events. It will ask the operating system for any new events, call
		793	the watcher callbacks, an then repeat the whole process indefinitely: This
		794	is why event loops are called I<loops>.
583		795
584	If the flags argument is specified as C<0>, it will not return until	796	If the flags argument is specified as C<0>, it will keep handling events
585	either no event watchers are active anymore or C<ev_unloop> was called.	797	until either no event watchers are active anymore or C<ev_break> was
		798	called.
586		799
587	Please note that an explicit C<ev_unloop> is usually better than	800	Please note that an explicit C<ev_break> is usually better than
588	relying on all watchers to be stopped when deciding when a program has	801	relying on all watchers to be stopped when deciding when a program has
589	finished (especially in interactive programs), but having a program that	802	finished (especially in interactive programs), but having a program
590	automatically loops as long as it has to and no longer by virtue of	803	that automatically loops as long as it has to and no longer by virtue
591	relying on its watchers stopping correctly is a thing of beauty.	804	of relying on its watchers stopping correctly, that is truly a thing of
		805	beauty.
592		806
		807	This function is also I<mostly> exception-safe - you can break out of
		808	a C<ev_run> call by calling C<longjmp> in a callback, throwing a C++
		809	exception and so on. This does not decrement the C<ev_depth> value, nor
		810	will it clear any outstanding C<EVBREAK_ONE> breaks.
		811
593	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle	812	A flags value of C<EVRUN_NOWAIT> will look for new events, will handle
594	those events and any outstanding ones, but will not block your process in	813	those events and any already outstanding ones, but will not wait and
595	case there are no events and will return after one iteration of the loop.	814	block your process in case there are no events and will return after one
		815	iteration of the loop. This is sometimes useful to poll and handle new
		816	events while doing lengthy calculations, to keep the program responsive.
596		817
597	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if	818	A flags value of C<EVRUN_ONCE> will look for new events (waiting if
598	necessary) and will handle those and any outstanding ones. It will block	819	necessary) and will handle those and any already outstanding ones. It
599	your process until at least one new event arrives, and will return after	820	will block your process until at least one new event arrives (which could
600	one iteration of the loop. This is useful if you are waiting for some	821	be an event internal to libev itself, so there is no guarantee that a
601	external event in conjunction with something not expressible using other	822	user-registered callback will be called), and will return after one
		823	iteration of the loop.
		824
		825	This is useful if you are waiting for some external event in conjunction
		826	with something not expressible using other libev watchers (i.e. "roll your
602	libev watchers. However, a pair of C<ev_prepare>/C<ev_check> watchers is	827	own C<ev_run>"). However, a pair of C<ev_prepare>/C<ev_check> watchers is
603	usually a better approach for this kind of thing.	828	usually a better approach for this kind of thing.
604		829
605	Here are the gory details of what C<ev_loop> does:	830	Here are the gory details of what C<ev_run> does:
606		831
		832	- Increment loop depth.
		833	- Reset the ev_break status.
607	- Before the first iteration, call any pending watchers.	834	- Before the first iteration, call any pending watchers.
		835	LOOP:
608	* If EVFLAG_FORKCHECK was used, check for a fork.	836	- If EVFLAG_FORKCHECK was used, check for a fork.
609	- If a fork was detected (by any means), queue and call all fork watchers.	837	- If a fork was detected (by any means), queue and call all fork watchers.
610	- Queue and call all prepare watchers.	838	- Queue and call all prepare watchers.
		839	- If ev_break was called, goto FINISH.
611	- If we have been forked, detach and recreate the kernel state	840	- If we have been forked, detach and recreate the kernel state
612	as to not disturb the other process.	841	as to not disturb the other process.
613	- Update the kernel state with all outstanding changes.	842	- Update the kernel state with all outstanding changes.
614	- Update the "event loop time" (ev_now ()).	843	- Update the "event loop time" (ev_now ()).
615	- Calculate for how long to sleep or block, if at all	844	- Calculate for how long to sleep or block, if at all
616	(active idle watchers, EVLOOP_NONBLOCK or not having	845	(active idle watchers, EVRUN_NOWAIT or not having
617	any active watchers at all will result in not sleeping).	846	any active watchers at all will result in not sleeping).
618	- Sleep if the I/O and timer collect interval say so.	847	- Sleep if the I/O and timer collect interval say so.
		848	- Increment loop iteration counter.
619	- Block the process, waiting for any events.	849	- Block the process, waiting for any events.
620	- Queue all outstanding I/O (fd) events.	850	- Queue all outstanding I/O (fd) events.
621	- Update the "event loop time" (ev_now ()), and do time jump adjustments.	851	- Update the "event loop time" (ev_now ()), and do time jump adjustments.
622	- Queue all outstanding timers.	852	- Queue all expired timers.
623	- Queue all outstanding periodics.	853	- Queue all expired periodics.
624	- Unless any events are pending now, queue all idle watchers.	854	- Queue all idle watchers with priority higher than that of pending events.
625	- Queue all check watchers.	855	- Queue all check watchers.
626	- Call all queued watchers in reverse order (i.e. check watchers first).	856	- Call all queued watchers in reverse order (i.e. check watchers first).
627	Signals and child watchers are implemented as I/O watchers, and will	857	Signals and child watchers are implemented as I/O watchers, and will
628	be handled here by queueing them when their watcher gets executed.	858	be handled here by queueing them when their watcher gets executed.
629	- If ev_unloop has been called, or EVLOOP_ONESHOT or EVLOOP_NONBLOCK	859	- If ev_break has been called, or EVRUN_ONCE or EVRUN_NOWAIT
630	were used, or there are no active watchers, return, otherwise	860	were used, or there are no active watchers, goto FINISH, otherwise
631	continue with step *.	861	continue with step LOOP.
		862	FINISH:
		863	- Reset the ev_break status iff it was EVBREAK_ONE.
		864	- Decrement the loop depth.
		865	- Return.
632		866
633	Example: Queue some jobs and then loop until no events are outstanding	867	Example: Queue some jobs and then loop until no events are outstanding
634	anymore.	868	anymore.
635		869
636	... queue jobs here, make sure they register event watchers as long	870	... queue jobs here, make sure they register event watchers as long
637	... as they still have work to do (even an idle watcher will do..)	871	... as they still have work to do (even an idle watcher will do..)
638	ev_loop (my_loop, 0);	872	ev_run (my_loop, 0);
639	... jobs done or somebody called unloop. yeah!	873	... jobs done or somebody called break. yeah!
640		874
641	=item ev_unloop (loop, how)	875	=item ev_break (loop, how)
642		876
643	Can be used to make a call to C<ev_loop> return early (but only after it	877	Can be used to make a call to C<ev_run> return early (but only after it
644	has processed all outstanding events). The C<how> argument must be either	878	has processed all outstanding events). The C<how> argument must be either
645	C<EVUNLOOP_ONE>, which will make the innermost C<ev_loop> call return, or	879	C<EVBREAK_ONE>, which will make the innermost C<ev_run> call return, or
646	C<EVUNLOOP_ALL>, which will make all nested C<ev_loop> calls return.	880	C<EVBREAK_ALL>, which will make all nested C<ev_run> calls return.
647		881
648	This "unloop state" will be cleared when entering C<ev_loop> again.	882	This "break state" will be cleared on the next call to C<ev_run>.
		883
		884	It is safe to call C<ev_break> from outside any C<ev_run> calls, too, in
		885	which case it will have no effect.
649		886
650	=item ev_ref (loop)	887	=item ev_ref (loop)
651		888
652	=item ev_unref (loop)	889	=item ev_unref (loop)
653		890
654	Ref/unref can be used to add or remove a reference count on the event	891	Ref/unref can be used to add or remove a reference count on the event
655	loop: Every watcher keeps one reference, and as long as the reference	892	loop: Every watcher keeps one reference, and as long as the reference
656	count is nonzero, C<ev_loop> will not return on its own. If you have	893	count is nonzero, C<ev_run> will not return on its own.
657	a watcher you never unregister that should not keep C<ev_loop> from	894
658	returning, ev_unref() after starting, and ev_ref() before stopping it. For	895	This is useful when you have a watcher that you never intend to
		896	unregister, but that nevertheless should not keep C<ev_run> from
		897	returning. In such a case, call C<ev_unref> after starting, and C<ev_ref>
		898	before stopping it.
		899
659	example, libev itself uses this for its internal signal pipe: It is not	900	As an example, libev itself uses this for its internal signal pipe: It
660	visible to the libev user and should not keep C<ev_loop> from exiting if	901	is not visible to the libev user and should not keep C<ev_run> from
661	no event watchers registered by it are active. It is also an excellent	902	exiting if no event watchers registered by it are active. It is also an
662	way to do this for generic recurring timers or from within third-party	903	excellent way to do this for generic recurring timers or from within
663	libraries. Just remember to I<unref after start> and I<ref before stop>	904	third-party libraries. Just remember to I<unref after start> and I<ref
664	(but only if the watcher wasn't active before, or was active before,	905	before stop> (but only if the watcher wasn't active before, or was active
665	respectively).	906	before, respectively. Note also that libev might stop watchers itself
		907	(e.g. non-repeating timers) in which case you have to C<ev_ref>
		908	in the callback).
666		909
667	Example: Create a signal watcher, but keep it from keeping C<ev_loop>	910	Example: Create a signal watcher, but keep it from keeping C<ev_run>
668	running when nothing else is active.	911	running when nothing else is active.
669		912
670	struct ev_signal exitsig;	913	ev_signal exitsig;
671	ev_signal_init (&exitsig, sig_cb, SIGINT);	914	ev_signal_init (&exitsig, sig_cb, SIGINT);
672	ev_signal_start (loop, &exitsig);	915	ev_signal_start (loop, &exitsig);
673	evf_unref (loop);	916	ev_unref (loop);
674		917
675	Example: For some weird reason, unregister the above signal handler again.	918	Example: For some weird reason, unregister the above signal handler again.
676		919
677	ev_ref (loop);	920	ev_ref (loop);
678	ev_signal_stop (loop, &exitsig);	921	ev_signal_stop (loop, &exitsig);
…		…
689	Setting these to a higher value (the C<interval> I<must> be >= C<0>)	932	Setting these to a higher value (the C<interval> I<must> be >= C<0>)
690	allows libev to delay invocation of I/O and timer/periodic callbacks	933	allows libev to delay invocation of I/O and timer/periodic callbacks
691	to increase efficiency of loop iterations (or to increase power-saving	934	to increase efficiency of loop iterations (or to increase power-saving
692	opportunities).	935	opportunities).
693		936
694	The background is that sometimes your program runs just fast enough to	937	The idea is that sometimes your program runs just fast enough to handle
695	handle one (or very few) event(s) per loop iteration. While this makes	938	one (or very few) event(s) per loop iteration. While this makes the
696	the program responsive, it also wastes a lot of CPU time to poll for new	939	program responsive, it also wastes a lot of CPU time to poll for new
697	events, especially with backends like C<select ()> which have a high	940	events, especially with backends like C<select ()> which have a high
698	overhead for the actual polling but can deliver many events at once.	941	overhead for the actual polling but can deliver many events at once.
699		942
700	By setting a higher I<io collect interval> you allow libev to spend more	943	By setting a higher I<io collect interval> you allow libev to spend more
701	time collecting I/O events, so you can handle more events per iteration,	944	time collecting I/O events, so you can handle more events per iteration,
702	at the cost of increasing latency. Timeouts (both C<ev_periodic> and	945	at the cost of increasing latency. Timeouts (both C<ev_periodic> and
703	C<ev_timer>) will be not affected. Setting this to a non-null value will	946	C<ev_timer>) will be not affected. Setting this to a non-null value will
704	introduce an additional C<ev_sleep ()> call into most loop iterations.	947	introduce an additional C<ev_sleep ()> call into most loop iterations. The
		948	sleep time ensures that libev will not poll for I/O events more often then
		949	once per this interval, on average.
705		950
706	Likewise, by setting a higher I<timeout collect interval> you allow libev	951	Likewise, by setting a higher I<timeout collect interval> you allow libev
707	to spend more time collecting timeouts, at the expense of increased	952	to spend more time collecting timeouts, at the expense of increased
708	latency (the watcher callback will be called later). C<ev_io> watchers	953	latency/jitter/inexactness (the watcher callback will be called
709	will not be affected. Setting this to a non-null value will not introduce	954	later). C<ev_io> watchers will not be affected. Setting this to a non-null
710	any overhead in libev.	955	value will not introduce any overhead in libev.
711		956
712	Many (busy) programs can usually benefit by setting the I/O collect	957	Many (busy) programs can usually benefit by setting the I/O collect
713	interval to a value near C<0.1> or so, which is often enough for	958	interval to a value near C<0.1> or so, which is often enough for
714	interactive servers (of course not for games), likewise for timeouts. It	959	interactive servers (of course not for games), likewise for timeouts. It
715	usually doesn't make much sense to set it to a lower value than C<0.01>,	960	usually doesn't make much sense to set it to a lower value than C<0.01>,
716	as this approaches the timing granularity of most systems.	961	as this approaches the timing granularity of most systems. Note that if
		962	you do transactions with the outside world and you can't increase the
		963	parallelity, then this setting will limit your transaction rate (if you
		964	need to poll once per transaction and the I/O collect interval is 0.01,
		965	then you can't do more than 100 transactions per second).
717		966
718	Setting the I<timeout collect interval> can improve the opportunity for	967	Setting the I<timeout collect interval> can improve the opportunity for
719	saving power, as the program will "bundle" timer callback invocations that	968	saving power, as the program will "bundle" timer callback invocations that
720	are "near" in time together, by delaying some, thus reducing the number of	969	are "near" in time together, by delaying some, thus reducing the number of
721	times the process sleeps and wakes up again. Another useful technique to	970	times the process sleeps and wakes up again. Another useful technique to
722	reduce iterations/wake-ups is to use C<ev_periodic> watchers and make sure	971	reduce iterations/wake-ups is to use C<ev_periodic> watchers and make sure
723	they fire on, say, one-second boundaries only.	972	they fire on, say, one-second boundaries only.
724		973
		974	Example: we only need 0.1s timeout granularity, and we wish not to poll
		975	more often than 100 times per second:
		976
		977	ev_set_timeout_collect_interval (EV_DEFAULT_UC_ 0.1);
		978	ev_set_io_collect_interval (EV_DEFAULT_UC_ 0.01);
		979
		980	=item ev_invoke_pending (loop)
		981
		982	This call will simply invoke all pending watchers while resetting their
		983	pending state. Normally, C<ev_run> does this automatically when required,
		984	but when overriding the invoke callback this call comes handy. This
		985	function can be invoked from a watcher - this can be useful for example
		986	when you want to do some lengthy calculation and want to pass further
		987	event handling to another thread (you still have to make sure only one
		988	thread executes within C<ev_invoke_pending> or C<ev_run> of course).
		989
		990	=item int ev_pending_count (loop)
		991
		992	Returns the number of pending watchers - zero indicates that no watchers
		993	are pending.
		994
		995	=item ev_set_invoke_pending_cb (loop, void (*invoke_pending_cb)(EV_P))
		996
		997	This overrides the invoke pending functionality of the loop: Instead of
		998	invoking all pending watchers when there are any, C<ev_run> will call
		999	this callback instead. This is useful, for example, when you want to
		1000	invoke the actual watchers inside another context (another thread etc.).
		1001
		1002	If you want to reset the callback, use C<ev_invoke_pending> as new
		1003	callback.
		1004
		1005	=item ev_set_loop_release_cb (loop, void (release)(EV_P), void (acquire)(EV_P))
		1006
		1007	Sometimes you want to share the same loop between multiple threads. This
		1008	can be done relatively simply by putting mutex_lock/unlock calls around
		1009	each call to a libev function.
		1010
		1011	However, C<ev_run> can run an indefinite time, so it is not feasible
		1012	to wait for it to return. One way around this is to wake up the event
		1013	loop via C<ev_break> and C<av_async_send>, another way is to set these
		1014	I<release> and I<acquire> callbacks on the loop.
		1015
		1016	When set, then C<release> will be called just before the thread is
		1017	suspended waiting for new events, and C<acquire> is called just
		1018	afterwards.
		1019
		1020	Ideally, C<release> will just call your mutex_unlock function, and
		1021	C<acquire> will just call the mutex_lock function again.
		1022
		1023	While event loop modifications are allowed between invocations of
		1024	C<release> and C<acquire> (that's their only purpose after all), no
		1025	modifications done will affect the event loop, i.e. adding watchers will
		1026	have no effect on the set of file descriptors being watched, or the time
		1027	waited. Use an C<ev_async> watcher to wake up C<ev_run> when you want it
		1028	to take note of any changes you made.
		1029
		1030	In theory, threads executing C<ev_run> will be async-cancel safe between
		1031	invocations of C<release> and C<acquire>.
		1032
		1033	See also the locking example in the C<THREADS> section later in this
		1034	document.
		1035
		1036	=item ev_set_userdata (loop, void *data)
		1037
		1038	=item void *ev_userdata (loop)
		1039
		1040	Set and retrieve a single C<void *> associated with a loop. When
		1041	C<ev_set_userdata> has never been called, then C<ev_userdata> returns
		1042	C<0>.
		1043
		1044	These two functions can be used to associate arbitrary data with a loop,
		1045	and are intended solely for the C<invoke_pending_cb>, C<release> and
		1046	C<acquire> callbacks described above, but of course can be (ab-)used for
		1047	any other purpose as well.
		1048
725	=item ev_loop_verify (loop)	1049	=item ev_verify (loop)
726		1050
727	This function only does something when C<EV_VERIFY> support has been	1051	This function only does something when C<EV_VERIFY> support has been
728	compiled in. It tries to go through all internal structures and checks	1052	compiled in, which is the default for non-minimal builds. It tries to go
729	them for validity. If anything is found to be inconsistent, it will print	1053	through all internal structures and checks them for validity. If anything
730	an error message to standard error and call C<abort ()>.	1054	is found to be inconsistent, it will print an error message to standard
		1055	error and call C<abort ()>.
731		1056
732	This can be used to catch bugs inside libev itself: under normal	1057	This can be used to catch bugs inside libev itself: under normal
733	circumstances, this function will never abort as of course libev keeps its	1058	circumstances, this function will never abort as of course libev keeps its
734	data structures consistent.	1059	data structures consistent.
735		1060
736	=back	1061	=back
737		1062
738		1063
739	=head1 ANATOMY OF A WATCHER	1064	=head1 ANATOMY OF A WATCHER
740		1065
		1066	In the following description, uppercase C<TYPE> in names stands for the
		1067	watcher type, e.g. C<ev_TYPE_start> can mean C<ev_timer_start> for timer
		1068	watchers and C<ev_io_start> for I/O watchers.
		1069
741	A watcher is a structure that you create and register to record your	1070	A watcher is an opaque structure that you allocate and register to record
742	interest in some event. For instance, if you want to wait for STDIN to	1071	your interest in some event. To make a concrete example, imagine you want
743	become readable, you would create an C<ev_io> watcher for that:	1072	to wait for STDIN to become readable, you would create an C<ev_io> watcher
		1073	for that:
744		1074
745	static void my_cb (struct ev_loop loop, struct ev_io w, int revents)	1075	static void my_cb (struct ev_loop loop, ev_io w, int revents)
746	{	1076	{
747	ev_io_stop (w);	1077	ev_io_stop (w);
748	ev_unloop (loop, EVUNLOOP_ALL);	1078	ev_break (loop, EVBREAK_ALL);
749	}	1079	}
750		1080
751	struct ev_loop *loop = ev_default_loop (0);	1081	struct ev_loop *loop = ev_default_loop (0);
		1082
752	struct ev_io stdin_watcher;	1083	ev_io stdin_watcher;
		1084
753	ev_init (&stdin_watcher, my_cb);	1085	ev_init (&stdin_watcher, my_cb);
754	ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);	1086	ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);
755	ev_io_start (loop, &stdin_watcher);	1087	ev_io_start (loop, &stdin_watcher);
		1088
756	ev_loop (loop, 0);	1089	ev_run (loop, 0);
757		1090
758	As you can see, you are responsible for allocating the memory for your	1091	As you can see, you are responsible for allocating the memory for your
759	watcher structures (and it is usually a bad idea to do this on the stack,	1092	watcher structures (and it is I<usually> a bad idea to do this on the
760	although this can sometimes be quite valid).	1093	stack).
761		1094
		1095	Each watcher has an associated watcher structure (called C<struct ev_TYPE>
		1096	or simply C<ev_TYPE>, as typedefs are provided for all watcher structs).
		1097
762	Each watcher structure must be initialised by a call to C<ev_init	1098	Each watcher structure must be initialised by a call to C<ev_init (watcher
763	(watcher *, callback)>, which expects a callback to be provided. This	1099	*, callback)>, which expects a callback to be provided. This callback is
764	callback gets invoked each time the event occurs (or, in the case of I/O	1100	invoked each time the event occurs (or, in the case of I/O watchers, each
765	watchers, each time the event loop detects that the file descriptor given	1101	time the event loop detects that the file descriptor given is readable
766	is readable and/or writable).	1102	and/or writable).
767		1103
768	Each watcher type has its own C<< ev_<type>_set (watcher *, ...) >> macro	1104	Each watcher type further has its own C<< ev_TYPE_set (watcher *, ...) >>
769	with arguments specific to this watcher type. There is also a macro	1105	macro to configure it, with arguments specific to the watcher type. There
770	to combine initialisation and setting in one call: C<< ev_<type>_init	1106	is also a macro to combine initialisation and setting in one call: C<<
771	(watcher *, callback, ...) >>.	1107	ev_TYPE_init (watcher *, callback, ...) >>.
772		1108
773	To make the watcher actually watch out for events, you have to start it	1109	To make the watcher actually watch out for events, you have to start it
774	with a watcher-specific start function (C<< ev_<type>_start (loop, watcher	1110	with a watcher-specific start function (C<< ev_TYPE_start (loop, watcher
775	*) >>), and you can stop watching for events at any time by calling the	1111	*) >>), and you can stop watching for events at any time by calling the
776	corresponding stop function (C<< ev_<type>_stop (loop, watcher *) >>.	1112	corresponding stop function (C<< ev_TYPE_stop (loop, watcher *) >>.
777		1113
778	As long as your watcher is active (has been started but not stopped) you	1114	As long as your watcher is active (has been started but not stopped) you
779	must not touch the values stored in it. Most specifically you must never	1115	must not touch the values stored in it. Most specifically you must never
780	reinitialise it or call its C<set> macro.	1116	reinitialise it or call its C<ev_TYPE_set> macro.
781		1117
782	Each and every callback receives the event loop pointer as first, the	1118	Each and every callback receives the event loop pointer as first, the
783	registered watcher structure as second, and a bitset of received events as	1119	registered watcher structure as second, and a bitset of received events as
784	third argument.	1120	third argument.
785		1121
…		…
794	=item C<EV_WRITE>	1130	=item C<EV_WRITE>
795		1131
796	The file descriptor in the C<ev_io> watcher has become readable and/or	1132	The file descriptor in the C<ev_io> watcher has become readable and/or
797	writable.	1133	writable.
798		1134
799	=item C<EV_TIMEOUT>	1135	=item C<EV_TIMER>
800		1136
801	The C<ev_timer> watcher has timed out.	1137	The C<ev_timer> watcher has timed out.
802		1138
803	=item C<EV_PERIODIC>	1139	=item C<EV_PERIODIC>
804		1140
…		…
822		1158
823	=item C<EV_PREPARE>	1159	=item C<EV_PREPARE>
824		1160
825	=item C<EV_CHECK>	1161	=item C<EV_CHECK>
826		1162
827	All C<ev_prepare> watchers are invoked just I<before> C<ev_loop> starts	1163	All C<ev_prepare> watchers are invoked just I<before> C<ev_run> starts
828	to gather new events, and all C<ev_check> watchers are invoked just after	1164	to gather new events, and all C<ev_check> watchers are invoked just after
829	C<ev_loop> has gathered them, but before it invokes any callbacks for any	1165	C<ev_run> has gathered them, but before it invokes any callbacks for any
830	received events. Callbacks of both watcher types can start and stop as	1166	received events. Callbacks of both watcher types can start and stop as
831	many watchers as they want, and all of them will be taken into account	1167	many watchers as they want, and all of them will be taken into account
832	(for example, a C<ev_prepare> watcher might start an idle watcher to keep	1168	(for example, a C<ev_prepare> watcher might start an idle watcher to keep
833	C<ev_loop> from blocking).	1169	C<ev_run> from blocking).
834		1170
835	=item C<EV_EMBED>	1171	=item C<EV_EMBED>
836		1172
837	The embedded event loop specified in the C<ev_embed> watcher needs attention.	1173	The embedded event loop specified in the C<ev_embed> watcher needs attention.
838		1174
839	=item C<EV_FORK>	1175	=item C<EV_FORK>
840		1176
841	The event loop has been resumed in the child process after fork (see	1177	The event loop has been resumed in the child process after fork (see
842	C<ev_fork>).	1178	C<ev_fork>).
843		1179
		1180	=item C<EV_CLEANUP>
		1181
		1182	The event loop is about to be destroyed (see C<ev_cleanup>).
		1183
844	=item C<EV_ASYNC>	1184	=item C<EV_ASYNC>
845		1185
846	The given async watcher has been asynchronously notified (see C<ev_async>).	1186	The given async watcher has been asynchronously notified (see C<ev_async>).
		1187
		1188	=item C<EV_CUSTOM>
		1189
		1190	Not ever sent (or otherwise used) by libev itself, but can be freely used
		1191	by libev users to signal watchers (e.g. via C<ev_feed_event>).
847		1192
848	=item C<EV_ERROR>	1193	=item C<EV_ERROR>
849		1194
850	An unspecified error has occurred, the watcher has been stopped. This might	1195	An unspecified error has occurred, the watcher has been stopped. This might
851	happen because the watcher could not be properly started because libev	1196	happen because the watcher could not be properly started because libev
852	ran out of memory, a file descriptor was found to be closed or any other	1197	ran out of memory, a file descriptor was found to be closed or any other
		1198	problem. Libev considers these application bugs.
		1199
853	problem. You best act on it by reporting the problem and somehow coping	1200	You best act on it by reporting the problem and somehow coping with the
854	with the watcher being stopped.	1201	watcher being stopped. Note that well-written programs should not receive
		1202	an error ever, so when your watcher receives it, this usually indicates a
		1203	bug in your program.
855		1204
856	Libev will usually signal a few "dummy" events together with an error,	1205	Libev will usually signal a few "dummy" events together with an error, for
857	for example it might indicate that a fd is readable or writable, and if	1206	example it might indicate that a fd is readable or writable, and if your
858	your callbacks is well-written it can just attempt the operation and cope	1207	callbacks is well-written it can just attempt the operation and cope with
859	with the error from read() or write(). This will not work in multi-threaded	1208	the error from read() or write(). This will not work in multi-threaded
860	programs, though, so beware.	1209	programs, though, as the fd could already be closed and reused for another
		1210	thing, so beware.
861		1211
862	=back	1212	=back
863		1213
864	=head2 GENERIC WATCHER FUNCTIONS	1214	=head2 GENERIC WATCHER FUNCTIONS
865
866	In the following description, C<TYPE> stands for the watcher type,
867	e.g. C<timer> for C<ev_timer> watchers and C<io> for C<ev_io> watchers.
868		1215
869	=over 4	1216	=over 4
870		1217
871	=item C<ev_init> (ev_TYPE *watcher, callback)	1218	=item C<ev_init> (ev_TYPE *watcher, callback)
872		1219
…		…
878	which rolls both calls into one.	1225	which rolls both calls into one.
879		1226
880	You can reinitialise a watcher at any time as long as it has been stopped	1227	You can reinitialise a watcher at any time as long as it has been stopped
881	(or never started) and there are no pending events outstanding.	1228	(or never started) and there are no pending events outstanding.
882		1229
883	The callback is always of type C<void ()(ev_loop loop, ev_TYPE *watcher,	1230	The callback is always of type C<void ()(struct ev_loop loop, ev_TYPE *watcher,
884	int revents)>.	1231	int revents)>.
885		1232
		1233	Example: Initialise an C<ev_io> watcher in two steps.
		1234
		1235	ev_io w;
		1236	ev_init (&w, my_cb);
		1237	ev_io_set (&w, STDIN_FILENO, EV_READ);
		1238
886	=item C<ev_TYPE_set> (ev_TYPE *, [args])	1239	=item C<ev_TYPE_set> (ev_TYPE *watcher, [args])
887		1240
888	This macro initialises the type-specific parts of a watcher. You need to	1241	This macro initialises the type-specific parts of a watcher. You need to
889	call C<ev_init> at least once before you call this macro, but you can	1242	call C<ev_init> at least once before you call this macro, but you can
890	call C<ev_TYPE_set> any number of times. You must not, however, call this	1243	call C<ev_TYPE_set> any number of times. You must not, however, call this
891	macro on a watcher that is active (it can be pending, however, which is a	1244	macro on a watcher that is active (it can be pending, however, which is a
892	difference to the C<ev_init> macro).	1245	difference to the C<ev_init> macro).
893		1246
894	Although some watcher types do not have type-specific arguments	1247	Although some watcher types do not have type-specific arguments
895	(e.g. C<ev_prepare>) you still need to call its C<set> macro.	1248	(e.g. C<ev_prepare>) you still need to call its C<set> macro.
896		1249
		1250	See C<ev_init>, above, for an example.
		1251
897	=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])	1252	=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])
898		1253
899	This convenience macro rolls both C<ev_init> and C<ev_TYPE_set> macro	1254	This convenience macro rolls both C<ev_init> and C<ev_TYPE_set> macro
900	calls into a single call. This is the most convenient method to initialise	1255	calls into a single call. This is the most convenient method to initialise
901	a watcher. The same limitations apply, of course.	1256	a watcher. The same limitations apply, of course.
902		1257
		1258	Example: Initialise and set an C<ev_io> watcher in one step.
		1259
		1260	ev_io_init (&w, my_cb, STDIN_FILENO, EV_READ);
		1261
903	=item C<ev_TYPE_start> (loop , ev_TYPE watcher)	1262	=item C<ev_TYPE_start> (loop, ev_TYPE *watcher)
904		1263
905	Starts (activates) the given watcher. Only active watchers will receive	1264	Starts (activates) the given watcher. Only active watchers will receive
906	events. If the watcher is already active nothing will happen.	1265	events. If the watcher is already active nothing will happen.
907		1266
		1267	Example: Start the C<ev_io> watcher that is being abused as example in this
		1268	whole section.
		1269
		1270	ev_io_start (EV_DEFAULT_UC, &w);
		1271
908	=item C<ev_TYPE_stop> (loop , ev_TYPE watcher)	1272	=item C<ev_TYPE_stop> (loop, ev_TYPE *watcher)
909		1273
910	Stops the given watcher again (if active) and clears the pending	1274	Stops the given watcher if active, and clears the pending status (whether
		1275	the watcher was active or not).
		1276
911	status. It is possible that stopped watchers are pending (for example,	1277	It is possible that stopped watchers are pending - for example,
912	non-repeating timers are being stopped when they become pending), but	1278	non-repeating timers are being stopped when they become pending - but
913	C<ev_TYPE_stop> ensures that the watcher is neither active nor pending. If	1279	calling C<ev_TYPE_stop> ensures that the watcher is neither active nor
914	you want to free or reuse the memory used by the watcher it is therefore a	1280	pending. If you want to free or reuse the memory used by the watcher it is
915	good idea to always call its C<ev_TYPE_stop> function.	1281	therefore a good idea to always call its C<ev_TYPE_stop> function.
916		1282
917	=item bool ev_is_active (ev_TYPE *watcher)	1283	=item bool ev_is_active (ev_TYPE *watcher)
918		1284
919	Returns a true value iff the watcher is active (i.e. it has been started	1285	Returns a true value iff the watcher is active (i.e. it has been started
920	and not yet been stopped). As long as a watcher is active you must not modify	1286	and not yet been stopped). As long as a watcher is active you must not modify
…		…
936	=item ev_cb_set (ev_TYPE *watcher, callback)	1302	=item ev_cb_set (ev_TYPE *watcher, callback)
937		1303
938	Change the callback. You can change the callback at virtually any time	1304	Change the callback. You can change the callback at virtually any time
939	(modulo threads).	1305	(modulo threads).
940		1306
941	=item ev_set_priority (ev_TYPE *watcher, priority)	1307	=item ev_set_priority (ev_TYPE *watcher, int priority)
942		1308
943	=item int ev_priority (ev_TYPE *watcher)	1309	=item int ev_priority (ev_TYPE *watcher)
944		1310
945	Set and query the priority of the watcher. The priority is a small	1311	Set and query the priority of the watcher. The priority is a small
946	integer between C<EV_MAXPRI> (default: C<2>) and C<EV_MINPRI>	1312	integer between C<EV_MAXPRI> (default: C<2>) and C<EV_MINPRI>
947	(default: C<-2>). Pending watchers with higher priority will be invoked	1313	(default: C<-2>). Pending watchers with higher priority will be invoked
948	before watchers with lower priority, but priority will not keep watchers	1314	before watchers with lower priority, but priority will not keep watchers
949	from being executed (except for C<ev_idle> watchers).	1315	from being executed (except for C<ev_idle> watchers).
950		1316
951	This means that priorities are I<only> used for ordering callback
952	invocation after new events have been received. This is useful, for
953	example, to reduce latency after idling, or more often, to bind two
954	watchers on the same event and make sure one is called first.
955
956	If you need to suppress invocation when higher priority events are pending	1317	If you need to suppress invocation when higher priority events are pending
957	you need to look at C<ev_idle> watchers, which provide this functionality.	1318	you need to look at C<ev_idle> watchers, which provide this functionality.
958		1319
959	You I<must not> change the priority of a watcher as long as it is active or	1320	You I<must not> change the priority of a watcher as long as it is active or
960	pending.	1321	pending.
961		1322
		1323	Setting a priority outside the range of C<EV_MINPRI> to C<EV_MAXPRI> is
		1324	fine, as long as you do not mind that the priority value you query might
		1325	or might not have been clamped to the valid range.
		1326
962	The default priority used by watchers when no priority has been set is	1327	The default priority used by watchers when no priority has been set is
963	always C<0>, which is supposed to not be too high and not be too low :).	1328	always C<0>, which is supposed to not be too high and not be too low :).
964		1329
965	Setting a priority outside the range of C<EV_MINPRI> to C<EV_MAXPRI> is	1330	See L<WATCHER PRIORITY MODELS>, below, for a more thorough treatment of
966	fine, as long as you do not mind that the priority value you query might	1331	priorities.
967	or might not have been adjusted to be within valid range.
968		1332
969	=item ev_invoke (loop, ev_TYPE *watcher, int revents)	1333	=item ev_invoke (loop, ev_TYPE *watcher, int revents)
970		1334
971	Invoke the C<watcher> with the given C<loop> and C<revents>. Neither	1335	Invoke the C<watcher> with the given C<loop> and C<revents>. Neither
972	C<loop> nor C<revents> need to be valid as long as the watcher callback	1336	C<loop> nor C<revents> need to be valid as long as the watcher callback
973	can deal with that fact.	1337	can deal with that fact, as both are simply passed through to the
		1338	callback.
974		1339
975	=item int ev_clear_pending (loop, ev_TYPE *watcher)	1340	=item int ev_clear_pending (loop, ev_TYPE *watcher)
976		1341
977	If the watcher is pending, this function returns clears its pending status	1342	If the watcher is pending, this function clears its pending status and
978	and returns its C<revents> bitset (as if its callback was invoked). If the	1343	returns its C<revents> bitset (as if its callback was invoked). If the
979	watcher isn't pending it does nothing and returns C<0>.	1344	watcher isn't pending it does nothing and returns C<0>.
980		1345
		1346	Sometimes it can be useful to "poll" a watcher instead of waiting for its
		1347	callback to be invoked, which can be accomplished with this function.
		1348
		1349	=item ev_feed_event (loop, ev_TYPE *watcher, int revents)
		1350
		1351	Feeds the given event set into the event loop, as if the specified event
		1352	had happened for the specified watcher (which must be a pointer to an
		1353	initialised but not necessarily started event watcher). Obviously you must
		1354	not free the watcher as long as it has pending events.
		1355
		1356	Stopping the watcher, letting libev invoke it, or calling
		1357	C<ev_clear_pending> will clear the pending event, even if the watcher was
		1358	not started in the first place.
		1359
		1360	See also C<ev_feed_fd_event> and C<ev_feed_signal_event> for related
		1361	functions that do not need a watcher.
		1362
981	=back	1363	=back
982		1364
		1365	See also the L<ASSOCIATING CUSTOM DATA WITH A WATCHER> and L<BUILDING YOUR
		1366	OWN COMPOSITE WATCHERS> idioms.
983		1367
984	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER	1368	=head2 WATCHER STATES
985		1369
986	Each watcher has, by default, a member C<void *data> that you can change	1370	There are various watcher states mentioned throughout this manual -
987	and read at any time, libev will completely ignore it. This can be used	1371	active, pending and so on. In this section these states and the rules to
988	to associate arbitrary data with your watcher. If you need more data and	1372	transition between them will be described in more detail - and while these
989	don't want to allocate memory and store a pointer to it in that data	1373	rules might look complicated, they usually do "the right thing".
990	member, you can also "subclass" the watcher type and provide your own
991	data:
992		1374
993	struct my_io	1375	=over 4
		1376
		1377	=item initialiased
		1378
		1379	Before a watcher can be registered with the event looop it has to be
		1380	initialised. This can be done with a call to C<ev_TYPE_init>, or calls to
		1381	C<ev_init> followed by the watcher-specific C<ev_TYPE_set> function.
		1382
		1383	In this state it is simply some block of memory that is suitable for
		1384	use in an event loop. It can be moved around, freed, reused etc. at
		1385	will - as long as you either keep the memory contents intact, or call
		1386	C<ev_TYPE_init> again.
		1387
		1388	=item started/running/active
		1389
		1390	Once a watcher has been started with a call to C<ev_TYPE_start> it becomes
		1391	property of the event loop, and is actively waiting for events. While in
		1392	this state it cannot be accessed (except in a few documented ways), moved,
		1393	freed or anything else - the only legal thing is to keep a pointer to it,
		1394	and call libev functions on it that are documented to work on active watchers.
		1395
		1396	=item pending
		1397
		1398	If a watcher is active and libev determines that an event it is interested
		1399	in has occurred (such as a timer expiring), it will become pending. It will
		1400	stay in this pending state until either it is stopped or its callback is
		1401	about to be invoked, so it is not normally pending inside the watcher
		1402	callback.
		1403
		1404	The watcher might or might not be active while it is pending (for example,
		1405	an expired non-repeating timer can be pending but no longer active). If it
		1406	is stopped, it can be freely accessed (e.g. by calling C<ev_TYPE_set>),
		1407	but it is still property of the event loop at this time, so cannot be
		1408	moved, freed or reused. And if it is active the rules described in the
		1409	previous item still apply.
		1410
		1411	It is also possible to feed an event on a watcher that is not active (e.g.
		1412	via C<ev_feed_event>), in which case it becomes pending without being
		1413	active.
		1414
		1415	=item stopped
		1416
		1417	A watcher can be stopped implicitly by libev (in which case it might still
		1418	be pending), or explicitly by calling its C<ev_TYPE_stop> function. The
		1419	latter will clear any pending state the watcher might be in, regardless
		1420	of whether it was active or not, so stopping a watcher explicitly before
		1421	freeing it is often a good idea.
		1422
		1423	While stopped (and not pending) the watcher is essentially in the
		1424	initialised state, that is, it can be reused, moved, modified in any way
		1425	you wish (but when you trash the memory block, you need to C<ev_TYPE_init>
		1426	it again).
		1427
		1428	=back
		1429
		1430	=head2 WATCHER PRIORITY MODELS
		1431
		1432	Many event loops support I<watcher priorities>, which are usually small
		1433	integers that influence the ordering of event callback invocation
		1434	between watchers in some way, all else being equal.
		1435
		1436	In libev, Watcher priorities can be set using C<ev_set_priority>. See its
		1437	description for the more technical details such as the actual priority
		1438	range.
		1439
		1440	There are two common ways how these these priorities are being interpreted
		1441	by event loops:
		1442
		1443	In the more common lock-out model, higher priorities "lock out" invocation
		1444	of lower priority watchers, which means as long as higher priority
		1445	watchers receive events, lower priority watchers are not being invoked.
		1446
		1447	The less common only-for-ordering model uses priorities solely to order
		1448	callback invocation within a single event loop iteration: Higher priority
		1449	watchers are invoked before lower priority ones, but they all get invoked
		1450	before polling for new events.
		1451
		1452	Libev uses the second (only-for-ordering) model for all its watchers
		1453	except for idle watchers (which use the lock-out model).
		1454
		1455	The rationale behind this is that implementing the lock-out model for
		1456	watchers is not well supported by most kernel interfaces, and most event
		1457	libraries will just poll for the same events again and again as long as
		1458	their callbacks have not been executed, which is very inefficient in the
		1459	common case of one high-priority watcher locking out a mass of lower
		1460	priority ones.
		1461
		1462	Static (ordering) priorities are most useful when you have two or more
		1463	watchers handling the same resource: a typical usage example is having an
		1464	C<ev_io> watcher to receive data, and an associated C<ev_timer> to handle
		1465	timeouts. Under load, data might be received while the program handles
		1466	other jobs, but since timers normally get invoked first, the timeout
		1467	handler will be executed before checking for data. In that case, giving
		1468	the timer a lower priority than the I/O watcher ensures that I/O will be
		1469	handled first even under adverse conditions (which is usually, but not
		1470	always, what you want).
		1471
		1472	Since idle watchers use the "lock-out" model, meaning that idle watchers
		1473	will only be executed when no same or higher priority watchers have
		1474	received events, they can be used to implement the "lock-out" model when
		1475	required.
		1476
		1477	For example, to emulate how many other event libraries handle priorities,
		1478	you can associate an C<ev_idle> watcher to each such watcher, and in
		1479	the normal watcher callback, you just start the idle watcher. The real
		1480	processing is done in the idle watcher callback. This causes libev to
		1481	continuously poll and process kernel event data for the watcher, but when
		1482	the lock-out case is known to be rare (which in turn is rare :), this is
		1483	workable.
		1484
		1485	Usually, however, the lock-out model implemented that way will perform
		1486	miserably under the type of load it was designed to handle. In that case,
		1487	it might be preferable to stop the real watcher before starting the
		1488	idle watcher, so the kernel will not have to process the event in case
		1489	the actual processing will be delayed for considerable time.
		1490
		1491	Here is an example of an I/O watcher that should run at a strictly lower
		1492	priority than the default, and which should only process data when no
		1493	other events are pending:
		1494
		1495	ev_idle idle; // actual processing watcher
		1496	ev_io io; // actual event watcher
		1497
		1498	static void
		1499	io_cb (EV_P_ ev_io *w, int revents)
994	{	1500	{
995	struct ev_io io;	1501	// stop the I/O watcher, we received the event, but
996	int otherfd;	1502	// are not yet ready to handle it.
997	void *somedata;	1503	ev_io_stop (EV_A_ w);
998	struct whatever *mostinteresting;	1504
		1505	// start the idle watcher to handle the actual event.
		1506	// it will not be executed as long as other watchers
		1507	// with the default priority are receiving events.
		1508	ev_idle_start (EV_A_ &idle);
999	}	1509	}
1000		1510
1001	And since your callback will be called with a pointer to the watcher, you	1511	static void
1002	can cast it back to your own type:	1512	idle_cb (EV_P_ ev_idle *w, int revents)
1003
1004	static void my_cb (struct ev_loop loop, struct ev_io w_, int revents)
1005	{	1513	{
1006	struct my_io w = (struct my_io )w_;	1514	// actual processing
1007	...	1515	read (STDIN_FILENO, ...);
		1516
		1517	// have to start the I/O watcher again, as
		1518	// we have handled the event
		1519	ev_io_start (EV_P_ &io);
1008	}	1520	}
1009		1521
1010	More interesting and less C-conformant ways of casting your callback type	1522	// initialisation
1011	instead have been omitted.	1523	ev_idle_init (&idle, idle_cb);
		1524	ev_io_init (&io, io_cb, STDIN_FILENO, EV_READ);
		1525	ev_io_start (EV_DEFAULT_ &io);
1012		1526
1013	Another common scenario is having some data structure with multiple	1527	In the "real" world, it might also be beneficial to start a timer, so that
1014	watchers:	1528	low-priority connections can not be locked out forever under load. This
1015		1529	enables your program to keep a lower latency for important connections
1016	struct my_biggy	1530	during short periods of high load, while not completely locking out less
1017	{	1531	important ones.
1018	int some_data;
1019	ev_timer t1;
1020	ev_timer t2;
1021	}
1022
1023	In this case getting the pointer to C<my_biggy> is a bit more complicated,
1024	you need to use C<offsetof>:
1025
1026	#include <stddef.h>
1027
1028	static void
1029	t1_cb (EV_P_ struct ev_timer *w, int revents)
1030	{
1031	struct my_biggy big = (struct my_biggy *
1032	(((char *)w) - offsetof (struct my_biggy, t1));
1033	}
1034
1035	static void
1036	t2_cb (EV_P_ struct ev_timer *w, int revents)
1037	{
1038	struct my_biggy big = (struct my_biggy *
1039	(((char *)w) - offsetof (struct my_biggy, t2));
1040	}
1041		1532
1042		1533
1043	=head1 WATCHER TYPES	1534	=head1 WATCHER TYPES
1044		1535
1045	This section describes each watcher in detail, but will not repeat	1536	This section describes each watcher in detail, but will not repeat
…		…
1069	In general you can register as many read and/or write event watchers per	1560	In general you can register as many read and/or write event watchers per
1070	fd as you want (as long as you don't confuse yourself). Setting all file	1561	fd as you want (as long as you don't confuse yourself). Setting all file
1071	descriptors to non-blocking mode is also usually a good idea (but not	1562	descriptors to non-blocking mode is also usually a good idea (but not
1072	required if you know what you are doing).	1563	required if you know what you are doing).
1073		1564
1074	If you must do this, then force the use of a known-to-be-good backend
1075	(at the time of this writing, this includes only C<EVBACKEND_SELECT> and
1076	C<EVBACKEND_POLL>).
1077
1078	Another thing you have to watch out for is that it is quite easy to	1565	Another thing you have to watch out for is that it is quite easy to
1079	receive "spurious" readiness notifications, that is your callback might	1566	receive "spurious" readiness notifications, that is, your callback might
1080	be called with C<EV_READ> but a subsequent C<read>(2) will actually block	1567	be called with C<EV_READ> but a subsequent C<read>(2) will actually block
1081	because there is no data. Not only are some backends known to create a	1568	because there is no data. It is very easy to get into this situation even
1082	lot of those (for example Solaris ports), it is very easy to get into	1569	with a relatively standard program structure. Thus it is best to always
1083	this situation even with a relatively standard program structure. Thus	1570	use non-blocking I/O: An extra C<read>(2) returning C<EAGAIN> is far
1084	it is best to always use non-blocking I/O: An extra C<read>(2) returning
1085	C<EAGAIN> is far preferable to a program hanging until some data arrives.	1571	preferable to a program hanging until some data arrives.
1086		1572
1087	If you cannot run the fd in non-blocking mode (for example you should not	1573	If you cannot run the fd in non-blocking mode (for example you should
1088	play around with an Xlib connection), then you have to separately re-test	1574	not play around with an Xlib connection), then you have to separately
1089	whether a file descriptor is really ready with a known-to-be good interface	1575	re-test whether a file descriptor is really ready with a known-to-be good
1090	such as poll (fortunately in our Xlib example, Xlib already does this on	1576	interface such as poll (fortunately in the case of Xlib, it already does
1091	its own, so its quite safe to use).	1577	this on its own, so its quite safe to use). Some people additionally
		1578	use C<SIGALRM> and an interval timer, just to be sure you won't block
		1579	indefinitely.
		1580
		1581	But really, best use non-blocking mode.
1092		1582
1093	=head3 The special problem of disappearing file descriptors	1583	=head3 The special problem of disappearing file descriptors
1094		1584
1095	Some backends (e.g. kqueue, epoll) need to be told about closing a file	1585	Some backends (e.g. kqueue, epoll) need to be told about closing a file
1096	descriptor (either by calling C<close> explicitly or by any other means,	1586	descriptor (either due to calling C<close> explicitly or any other means,
1097	such as C<dup>). The reason is that you register interest in some file	1587	such as C<dup2>). The reason is that you register interest in some file
1098	descriptor, but when it goes away, the operating system will silently drop	1588	descriptor, but when it goes away, the operating system will silently drop
1099	this interest. If another file descriptor with the same number then is	1589	this interest. If another file descriptor with the same number then is
1100	registered with libev, there is no efficient way to see that this is, in	1590	registered with libev, there is no efficient way to see that this is, in
1101	fact, a different file descriptor.	1591	fact, a different file descriptor.
1102		1592
…		…
1120		1610
1121	There is no workaround possible except not registering events	1611	There is no workaround possible except not registering events
1122	for potentially C<dup ()>'ed file descriptors, or to resort to	1612	for potentially C<dup ()>'ed file descriptors, or to resort to
1123	C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>.	1613	C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>.
1124		1614
		1615	=head3 The special problem of files
		1616
		1617	Many people try to use C<select> (or libev) on file descriptors
		1618	representing files, and expect it to become ready when their program
		1619	doesn't block on disk accesses (which can take a long time on their own).
		1620
		1621	However, this cannot ever work in the "expected" way - you get a readiness
		1622	notification as soon as the kernel knows whether and how much data is
		1623	there, and in the case of open files, that's always the case, so you
		1624	always get a readiness notification instantly, and your read (or possibly
		1625	write) will still block on the disk I/O.
		1626
		1627	Another way to view it is that in the case of sockets, pipes, character
		1628	devices and so on, there is another party (the sender) that delivers data
		1629	on its own, but in the case of files, there is no such thing: the disk
		1630	will not send data on its own, simply because it doesn't know what you
		1631	wish to read - you would first have to request some data.
		1632
		1633	Since files are typically not-so-well supported by advanced notification
		1634	mechanism, libev tries hard to emulate POSIX behaviour with respect
		1635	to files, even though you should not use it. The reason for this is
		1636	convenience: sometimes you want to watch STDIN or STDOUT, which is
		1637	usually a tty, often a pipe, but also sometimes files or special devices
		1638	(for example, C<epoll> on Linux works with F</dev/random> but not with
		1639	F</dev/urandom>), and even though the file might better be served with
		1640	asynchronous I/O instead of with non-blocking I/O, it is still useful when
		1641	it "just works" instead of freezing.
		1642
		1643	So avoid file descriptors pointing to files when you know it (e.g. use
		1644	libeio), but use them when it is convenient, e.g. for STDIN/STDOUT, or
		1645	when you rarely read from a file instead of from a socket, and want to
		1646	reuse the same code path.
		1647
1125	=head3 The special problem of fork	1648	=head3 The special problem of fork
1126		1649
1127	Some backends (epoll, kqueue) do not support C<fork ()> at all or exhibit	1650	Some backends (epoll, kqueue) do not support C<fork ()> at all or exhibit
1128	useless behaviour. Libev fully supports fork, but needs to be told about	1651	useless behaviour. Libev fully supports fork, but needs to be told about
1129	it in the child.	1652	it in the child if you want to continue to use it in the child.
1130		1653
1131	To support fork in your programs, you either have to call	1654	To support fork in your child processes, you have to call C<ev_loop_fork
1132	C<ev_default_fork ()> or C<ev_loop_fork ()> after a fork in the child,	1655	()> after a fork in the child, enable C<EVFLAG_FORKCHECK>, or resort to
1133	enable C<EVFLAG_FORKCHECK>, or resort to C<EVBACKEND_SELECT> or	1656	C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>.
1134	C<EVBACKEND_POLL>.
1135		1657
1136	=head3 The special problem of SIGPIPE	1658	=head3 The special problem of SIGPIPE
1137		1659
1138	While not really specific to libev, it is easy to forget about SIGPIPE:	1660	While not really specific to libev, it is easy to forget about C<SIGPIPE>:
1139	when reading from a pipe whose other end has been closed, your program	1661	when writing to a pipe whose other end has been closed, your program gets
1140	gets send a SIGPIPE, which, by default, aborts your program. For most	1662	sent a SIGPIPE, which, by default, aborts your program. For most programs
1141	programs this is sensible behaviour, for daemons, this is usually	1663	this is sensible behaviour, for daemons, this is usually undesirable.
1142	undesirable.
1143		1664
1144	So when you encounter spurious, unexplained daemon exits, make sure you	1665	So when you encounter spurious, unexplained daemon exits, make sure you
1145	ignore SIGPIPE (and maybe make sure you log the exit status of your daemon	1666	ignore SIGPIPE (and maybe make sure you log the exit status of your daemon
1146	somewhere, as that would have given you a big clue).	1667	somewhere, as that would have given you a big clue).
1147		1668
		1669	=head3 The special problem of accept()ing when you can't
		1670
		1671	Many implementations of the POSIX C<accept> function (for example,
		1672	found in post-2004 Linux) have the peculiar behaviour of not removing a
		1673	connection from the pending queue in all error cases.
		1674
		1675	For example, larger servers often run out of file descriptors (because
		1676	of resource limits), causing C<accept> to fail with C<ENFILE> but not
		1677	rejecting the connection, leading to libev signalling readiness on
		1678	the next iteration again (the connection still exists after all), and
		1679	typically causing the program to loop at 100% CPU usage.
		1680
		1681	Unfortunately, the set of errors that cause this issue differs between
		1682	operating systems, there is usually little the app can do to remedy the
		1683	situation, and no known thread-safe method of removing the connection to
		1684	cope with overload is known (to me).
		1685
		1686	One of the easiest ways to handle this situation is to just ignore it
		1687	- when the program encounters an overload, it will just loop until the
		1688	situation is over. While this is a form of busy waiting, no OS offers an
		1689	event-based way to handle this situation, so it's the best one can do.
		1690
		1691	A better way to handle the situation is to log any errors other than
		1692	C<EAGAIN> and C<EWOULDBLOCK>, making sure not to flood the log with such
		1693	messages, and continue as usual, which at least gives the user an idea of
		1694	what could be wrong ("raise the ulimit!"). For extra points one could stop
		1695	the C<ev_io> watcher on the listening fd "for a while", which reduces CPU
		1696	usage.
		1697
		1698	If your program is single-threaded, then you could also keep a dummy file
		1699	descriptor for overload situations (e.g. by opening F</dev/null>), and
		1700	when you run into C<ENFILE> or C<EMFILE>, close it, run C<accept>,
		1701	close that fd, and create a new dummy fd. This will gracefully refuse
		1702	clients under typical overload conditions.
		1703
		1704	The last way to handle it is to simply log the error and C<exit>, as
		1705	is often done with C<malloc> failures, but this results in an easy
		1706	opportunity for a DoS attack.
1148		1707
1149	=head3 Watcher-Specific Functions	1708	=head3 Watcher-Specific Functions
1150		1709
1151	=over 4	1710	=over 4
1152		1711
1153	=item ev_io_init (ev_io *, callback, int fd, int events)	1712	=item ev_io_init (ev_io *, callback, int fd, int events)
1154		1713
1155	=item ev_io_set (ev_io *, int fd, int events)	1714	=item ev_io_set (ev_io *, int fd, int events)
1156		1715
1157	Configures an C<ev_io> watcher. The C<fd> is the file descriptor to	1716	Configures an C<ev_io> watcher. The C<fd> is the file descriptor to
1158	receive events for and events is either C<EV_READ>, C<EV_WRITE> or	1717	receive events for and C<events> is either C<EV_READ>, C<EV_WRITE> or
1159	C<EV_READ \| EV_WRITE> to receive the given events.	1718	C<EV_READ \| EV_WRITE>, to express the desire to receive the given events.
1160		1719
1161	=item int fd [read-only]	1720	=item int fd [read-only]
1162		1721
1163	The file descriptor being watched.	1722	The file descriptor being watched.
1164		1723
…		…
1173	Example: Call C<stdin_readable_cb> when STDIN_FILENO has become, well	1732	Example: Call C<stdin_readable_cb> when STDIN_FILENO has become, well
1174	readable, but only once. Since it is likely line-buffered, you could	1733	readable, but only once. Since it is likely line-buffered, you could
1175	attempt to read a whole line in the callback.	1734	attempt to read a whole line in the callback.
1176		1735
1177	static void	1736	static void
1178	stdin_readable_cb (struct ev_loop loop, struct ev_io w, int revents)	1737	stdin_readable_cb (struct ev_loop loop, ev_io w, int revents)
1179	{	1738	{
1180	ev_io_stop (loop, w);	1739	ev_io_stop (loop, w);
1181	.. read from stdin here (or from w->fd) and haqndle any I/O errors	1740	.. read from stdin here (or from w->fd) and handle any I/O errors
1182	}	1741	}
1183		1742
1184	...	1743	...
1185	struct ev_loop *loop = ev_default_init (0);	1744	struct ev_loop *loop = ev_default_init (0);
1186	struct ev_io stdin_readable;	1745	ev_io stdin_readable;
1187	ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);	1746	ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
1188	ev_io_start (loop, &stdin_readable);	1747	ev_io_start (loop, &stdin_readable);
1189	ev_loop (loop, 0);	1748	ev_run (loop, 0);
1190		1749
1191		1750
1192	=head2 C<ev_timer> - relative and optionally repeating timeouts	1751	=head2 C<ev_timer> - relative and optionally repeating timeouts
1193		1752
1194	Timer watchers are simple relative timers that generate an event after a	1753	Timer watchers are simple relative timers that generate an event after a
1195	given time, and optionally repeating in regular intervals after that.	1754	given time, and optionally repeating in regular intervals after that.
1196		1755
1197	The timers are based on real time, that is, if you register an event that	1756	The timers are based on real time, that is, if you register an event that
1198	times out after an hour and you reset your system clock to January last	1757	times out after an hour and you reset your system clock to January last
1199	year, it will still time out after (roughly) and hour. "Roughly" because	1758	year, it will still time out after (roughly) one hour. "Roughly" because
1200	detecting time jumps is hard, and some inaccuracies are unavoidable (the	1759	detecting time jumps is hard, and some inaccuracies are unavoidable (the
1201	monotonic clock option helps a lot here).	1760	monotonic clock option helps a lot here).
		1761
		1762	The callback is guaranteed to be invoked only I<after> its timeout has
		1763	passed (not I<at>, so on systems with very low-resolution clocks this
		1764	might introduce a small delay). If multiple timers become ready during the
		1765	same loop iteration then the ones with earlier time-out values are invoked
		1766	before ones of the same priority with later time-out values (but this is
		1767	no longer true when a callback calls C<ev_run> recursively).
		1768
		1769	=head3 Be smart about timeouts
		1770
		1771	Many real-world problems involve some kind of timeout, usually for error
		1772	recovery. A typical example is an HTTP request - if the other side hangs,
		1773	you want to raise some error after a while.
		1774
		1775	What follows are some ways to handle this problem, from obvious and
		1776	inefficient to smart and efficient.
		1777
		1778	In the following, a 60 second activity timeout is assumed - a timeout that
		1779	gets reset to 60 seconds each time there is activity (e.g. each time some
		1780	data or other life sign was received).
		1781
		1782	=over 4
		1783
		1784	=item 1. Use a timer and stop, reinitialise and start it on activity.
		1785
		1786	This is the most obvious, but not the most simple way: In the beginning,
		1787	start the watcher:
		1788
		1789	ev_timer_init (timer, callback, 60., 0.);
		1790	ev_timer_start (loop, timer);
		1791
		1792	Then, each time there is some activity, C<ev_timer_stop> it, initialise it
		1793	and start it again:
		1794
		1795	ev_timer_stop (loop, timer);
		1796	ev_timer_set (timer, 60., 0.);
		1797	ev_timer_start (loop, timer);
		1798
		1799	This is relatively simple to implement, but means that each time there is
		1800	some activity, libev will first have to remove the timer from its internal
		1801	data structure and then add it again. Libev tries to be fast, but it's
		1802	still not a constant-time operation.
		1803
		1804	=item 2. Use a timer and re-start it with C<ev_timer_again> inactivity.
		1805
		1806	This is the easiest way, and involves using C<ev_timer_again> instead of
		1807	C<ev_timer_start>.
		1808
		1809	To implement this, configure an C<ev_timer> with a C<repeat> value
		1810	of C<60> and then call C<ev_timer_again> at start and each time you
		1811	successfully read or write some data. If you go into an idle state where
		1812	you do not expect data to travel on the socket, you can C<ev_timer_stop>
		1813	the timer, and C<ev_timer_again> will automatically restart it if need be.
		1814
		1815	That means you can ignore both the C<ev_timer_start> function and the
		1816	C<after> argument to C<ev_timer_set>, and only ever use the C<repeat>
		1817	member and C<ev_timer_again>.
		1818
		1819	At start:
		1820
		1821	ev_init (timer, callback);
		1822	timer->repeat = 60.;
		1823	ev_timer_again (loop, timer);
		1824
		1825	Each time there is some activity:
		1826
		1827	ev_timer_again (loop, timer);
		1828
		1829	It is even possible to change the time-out on the fly, regardless of
		1830	whether the watcher is active or not:
		1831
		1832	timer->repeat = 30.;
		1833	ev_timer_again (loop, timer);
		1834
		1835	This is slightly more efficient then stopping/starting the timer each time
		1836	you want to modify its timeout value, as libev does not have to completely
		1837	remove and re-insert the timer from/into its internal data structure.
		1838
		1839	It is, however, even simpler than the "obvious" way to do it.
		1840
		1841	=item 3. Let the timer time out, but then re-arm it as required.
		1842
		1843	This method is more tricky, but usually most efficient: Most timeouts are
		1844	relatively long compared to the intervals between other activity - in
		1845	our example, within 60 seconds, there are usually many I/O events with
		1846	associated activity resets.
		1847
		1848	In this case, it would be more efficient to leave the C<ev_timer> alone,
		1849	but remember the time of last activity, and check for a real timeout only
		1850	within the callback:
		1851
		1852	ev_tstamp last_activity; // time of last activity
		1853
		1854	static void
		1855	callback (EV_P_ ev_timer *w, int revents)
		1856	{
		1857	ev_tstamp now = ev_now (EV_A);
		1858	ev_tstamp timeout = last_activity + 60.;
		1859
		1860	// if last_activity + 60. is older than now, we did time out
		1861	if (timeout < now)
		1862	{
		1863	// timeout occurred, take action
		1864	}
		1865	else
		1866	{
		1867	// callback was invoked, but there was some activity, re-arm
		1868	// the watcher to fire in last_activity + 60, which is
		1869	// guaranteed to be in the future, so "again" is positive:
		1870	w->repeat = timeout - now;
		1871	ev_timer_again (EV_A_ w);
		1872	}
		1873	}
		1874
		1875	To summarise the callback: first calculate the real timeout (defined
		1876	as "60 seconds after the last activity"), then check if that time has
		1877	been reached, which means something I<did>, in fact, time out. Otherwise
		1878	the callback was invoked too early (C<timeout> is in the future), so
		1879	re-schedule the timer to fire at that future time, to see if maybe we have
		1880	a timeout then.
		1881
		1882	Note how C<ev_timer_again> is used, taking advantage of the
		1883	C<ev_timer_again> optimisation when the timer is already running.
		1884
		1885	This scheme causes more callback invocations (about one every 60 seconds
		1886	minus half the average time between activity), but virtually no calls to
		1887	libev to change the timeout.
		1888
		1889	To start the timer, simply initialise the watcher and set C<last_activity>
		1890	to the current time (meaning we just have some activity :), then call the
		1891	callback, which will "do the right thing" and start the timer:
		1892
		1893	ev_init (timer, callback);
		1894	last_activity = ev_now (loop);
		1895	callback (loop, timer, EV_TIMER);
		1896
		1897	And when there is some activity, simply store the current time in
		1898	C<last_activity>, no libev calls at all:
		1899
		1900	last_activity = ev_now (loop);
		1901
		1902	This technique is slightly more complex, but in most cases where the
		1903	time-out is unlikely to be triggered, much more efficient.
		1904
		1905	Changing the timeout is trivial as well (if it isn't hard-coded in the
		1906	callback :) - just change the timeout and invoke the callback, which will
		1907	fix things for you.
		1908
		1909	=item 4. Wee, just use a double-linked list for your timeouts.
		1910
		1911	If there is not one request, but many thousands (millions...), all
		1912	employing some kind of timeout with the same timeout value, then one can
		1913	do even better:
		1914
		1915	When starting the timeout, calculate the timeout value and put the timeout
		1916	at the I<end> of the list.
		1917
		1918	Then use an C<ev_timer> to fire when the timeout at the I<beginning> of
		1919	the list is expected to fire (for example, using the technique #3).
		1920
		1921	When there is some activity, remove the timer from the list, recalculate
		1922	the timeout, append it to the end of the list again, and make sure to
		1923	update the C<ev_timer> if it was taken from the beginning of the list.
		1924
		1925	This way, one can manage an unlimited number of timeouts in O(1) time for
		1926	starting, stopping and updating the timers, at the expense of a major
		1927	complication, and having to use a constant timeout. The constant timeout
		1928	ensures that the list stays sorted.
		1929
		1930	=back
		1931
		1932	So which method the best?
		1933
		1934	Method #2 is a simple no-brain-required solution that is adequate in most
		1935	situations. Method #3 requires a bit more thinking, but handles many cases
		1936	better, and isn't very complicated either. In most case, choosing either
		1937	one is fine, with #3 being better in typical situations.
		1938
		1939	Method #1 is almost always a bad idea, and buys you nothing. Method #4 is
		1940	rather complicated, but extremely efficient, something that really pays
		1941	off after the first million or so of active timers, i.e. it's usually
		1942	overkill :)
		1943
		1944	=head3 The special problem of time updates
		1945
		1946	Establishing the current time is a costly operation (it usually takes at
		1947	least two system calls): EV therefore updates its idea of the current
		1948	time only before and after C<ev_run> collects new events, which causes a
		1949	growing difference between C<ev_now ()> and C<ev_time ()> when handling
		1950	lots of events in one iteration.
1202		1951
1203	The relative timeouts are calculated relative to the C<ev_now ()>	1952	The relative timeouts are calculated relative to the C<ev_now ()>
1204	time. This is usually the right thing as this timestamp refers to the time	1953	time. This is usually the right thing as this timestamp refers to the time
1205	of the event triggering whatever timeout you are modifying/starting. If	1954	of the event triggering whatever timeout you are modifying/starting. If
1206	you suspect event processing to be delayed and you I<need> to base the timeout	1955	you suspect event processing to be delayed and you I<need> to base the
1207	on the current time, use something like this to adjust for this:	1956	timeout on the current time, use something like this to adjust for this:
1208		1957
1209	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);	1958	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);
1210		1959
1211	The callback is guaranteed to be invoked only after its timeout has passed,	1960	If the event loop is suspended for a long time, you can also force an
1212	but if multiple timers become ready during the same loop iteration then	1961	update of the time returned by C<ev_now ()> by calling C<ev_now_update
1213	order of execution is undefined.	1962	()>.
		1963
		1964	=head3 The special problems of suspended animation
		1965
		1966	When you leave the server world it is quite customary to hit machines that
		1967	can suspend/hibernate - what happens to the clocks during such a suspend?
		1968
		1969	Some quick tests made with a Linux 2.6.28 indicate that a suspend freezes
		1970	all processes, while the clocks (C<times>, C<CLOCK_MONOTONIC>) continue
		1971	to run until the system is suspended, but they will not advance while the
		1972	system is suspended. That means, on resume, it will be as if the program
		1973	was frozen for a few seconds, but the suspend time will not be counted
		1974	towards C<ev_timer> when a monotonic clock source is used. The real time
		1975	clock advanced as expected, but if it is used as sole clocksource, then a
		1976	long suspend would be detected as a time jump by libev, and timers would
		1977	be adjusted accordingly.
		1978
		1979	I would not be surprised to see different behaviour in different between
		1980	operating systems, OS versions or even different hardware.
		1981
		1982	The other form of suspend (job control, or sending a SIGSTOP) will see a
		1983	time jump in the monotonic clocks and the realtime clock. If the program
		1984	is suspended for a very long time, and monotonic clock sources are in use,
		1985	then you can expect C<ev_timer>s to expire as the full suspension time
		1986	will be counted towards the timers. When no monotonic clock source is in
		1987	use, then libev will again assume a timejump and adjust accordingly.
		1988
		1989	It might be beneficial for this latter case to call C<ev_suspend>
		1990	and C<ev_resume> in code that handles C<SIGTSTP>, to at least get
		1991	deterministic behaviour in this case (you can do nothing against
		1992	C<SIGSTOP>).
1214		1993
1215	=head3 Watcher-Specific Functions and Data Members	1994	=head3 Watcher-Specific Functions and Data Members
1216		1995
1217	=over 4	1996	=over 4
1218		1997
…		…
1242	If the timer is started but non-repeating, stop it (as if it timed out).	2021	If the timer is started but non-repeating, stop it (as if it timed out).
1243		2022
1244	If the timer is repeating, either start it if necessary (with the	2023	If the timer is repeating, either start it if necessary (with the
1245	C<repeat> value), or reset the running timer to the C<repeat> value.	2024	C<repeat> value), or reset the running timer to the C<repeat> value.
1246		2025
1247	This sounds a bit complicated, but here is a useful and typical	2026	This sounds a bit complicated, see L<Be smart about timeouts>, above, for a
1248	example: Imagine you have a TCP connection and you want a so-called idle	2027	usage example.
1249	timeout, that is, you want to be called when there have been, say, 60
1250	seconds of inactivity on the socket. The easiest way to do this is to
1251	configure an C<ev_timer> with a C<repeat> value of C<60> and then call
1252	C<ev_timer_again> each time you successfully read or write some data. If
1253	you go into an idle state where you do not expect data to travel on the
1254	socket, you can C<ev_timer_stop> the timer, and C<ev_timer_again> will
1255	automatically restart it if need be.
1256		2028
1257	That means you can ignore the C<after> value and C<ev_timer_start>	2029	=item ev_tstamp ev_timer_remaining (loop, ev_timer *)
1258	altogether and only ever use the C<repeat> value and C<ev_timer_again>:
1259		2030
1260	ev_timer_init (timer, callback, 0., 5.);	2031	Returns the remaining time until a timer fires. If the timer is active,
1261	ev_timer_again (loop, timer);	2032	then this time is relative to the current event loop time, otherwise it's
1262	...	2033	the timeout value currently configured.
1263	timer->again = 17.;
1264	ev_timer_again (loop, timer);
1265	...
1266	timer->again = 10.;
1267	ev_timer_again (loop, timer);
1268		2034
1269	This is more slightly efficient then stopping/starting the timer each time	2035	That is, after an C<ev_timer_set (w, 5, 7)>, C<ev_timer_remaining> returns
1270	you want to modify its timeout value.	2036	C<5>. When the timer is started and one second passes, C<ev_timer_remaining>
		2037	will return C<4>. When the timer expires and is restarted, it will return
		2038	roughly C<7> (likely slightly less as callback invocation takes some time,
		2039	too), and so on.
1271		2040
1272	=item ev_tstamp repeat [read-write]	2041	=item ev_tstamp repeat [read-write]
1273		2042
1274	The current C<repeat> value. Will be used each time the watcher times out	2043	The current C<repeat> value. Will be used each time the watcher times out
1275	or C<ev_timer_again> is called and determines the next timeout (if any),	2044	or C<ev_timer_again> is called, and determines the next timeout (if any),
1276	which is also when any modifications are taken into account.	2045	which is also when any modifications are taken into account.
1277		2046
1278	=back	2047	=back
1279		2048
1280	=head3 Examples	2049	=head3 Examples
1281		2050
1282	Example: Create a timer that fires after 60 seconds.	2051	Example: Create a timer that fires after 60 seconds.
1283		2052
1284	static void	2053	static void
1285	one_minute_cb (struct ev_loop loop, struct ev_timer w, int revents)	2054	one_minute_cb (struct ev_loop loop, ev_timer w, int revents)
1286	{	2055	{
1287	.. one minute over, w is actually stopped right here	2056	.. one minute over, w is actually stopped right here
1288	}	2057	}
1289		2058
1290	struct ev_timer mytimer;	2059	ev_timer mytimer;
1291	ev_timer_init (&mytimer, one_minute_cb, 60., 0.);	2060	ev_timer_init (&mytimer, one_minute_cb, 60., 0.);
1292	ev_timer_start (loop, &mytimer);	2061	ev_timer_start (loop, &mytimer);
1293		2062
1294	Example: Create a timeout timer that times out after 10 seconds of	2063	Example: Create a timeout timer that times out after 10 seconds of
1295	inactivity.	2064	inactivity.
1296		2065
1297	static void	2066	static void
1298	timeout_cb (struct ev_loop loop, struct ev_timer w, int revents)	2067	timeout_cb (struct ev_loop loop, ev_timer w, int revents)
1299	{	2068	{
1300	.. ten seconds without any activity	2069	.. ten seconds without any activity
1301	}	2070	}
1302		2071
1303	struct ev_timer mytimer;	2072	ev_timer mytimer;
1304	ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */	2073	ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */
1305	ev_timer_again (&mytimer); /* start timer */	2074	ev_timer_again (&mytimer); /* start timer */
1306	ev_loop (loop, 0);	2075	ev_run (loop, 0);
1307		2076
1308	// and in some piece of code that gets executed on any "activity":	2077	// and in some piece of code that gets executed on any "activity":
1309	// reset the timeout to start ticking again at 10 seconds	2078	// reset the timeout to start ticking again at 10 seconds
1310	ev_timer_again (&mytimer);	2079	ev_timer_again (&mytimer);
1311		2080
…		…
1313	=head2 C<ev_periodic> - to cron or not to cron?	2082	=head2 C<ev_periodic> - to cron or not to cron?
1314		2083
1315	Periodic watchers are also timers of a kind, but they are very versatile	2084	Periodic watchers are also timers of a kind, but they are very versatile
1316	(and unfortunately a bit complex).	2085	(and unfortunately a bit complex).
1317		2086
1318	Unlike C<ev_timer>'s, they are not based on real time (or relative time)	2087	Unlike C<ev_timer>, periodic watchers are not based on real time (or
1319	but on wall clock time (absolute time). You can tell a periodic watcher	2088	relative time, the physical time that passes) but on wall clock time
1320	to trigger after some specific point in time. For example, if you tell a	2089	(absolute time, the thing you can read on your calender or clock). The
1321	periodic watcher to trigger in 10 seconds (by specifying e.g. C<ev_now ()	2090	difference is that wall clock time can run faster or slower than real
1322	+ 10.>, that is, an absolute time not a delay) and then reset your system	2091	time, and time jumps are not uncommon (e.g. when you adjust your
1323	clock to January of the previous year, then it will take more than year	2092	wrist-watch).
1324	to trigger the event (unlike an C<ev_timer>, which would still trigger
1325	roughly 10 seconds later as it uses a relative timeout).
1326		2093
		2094	You can tell a periodic watcher to trigger after some specific point
		2095	in time: for example, if you tell a periodic watcher to trigger "in 10
		2096	seconds" (by specifying e.g. C<ev_now () + 10.>, that is, an absolute time
		2097	not a delay) and then reset your system clock to January of the previous
		2098	year, then it will take a year or more to trigger the event (unlike an
		2099	C<ev_timer>, which would still trigger roughly 10 seconds after starting
		2100	it, as it uses a relative timeout).
		2101
1327	C<ev_periodic>s can also be used to implement vastly more complex timers,	2102	C<ev_periodic> watchers can also be used to implement vastly more complex
1328	such as triggering an event on each "midnight, local time", or other	2103	timers, such as triggering an event on each "midnight, local time", or
1329	complicated, rules.	2104	other complicated rules. This cannot be done with C<ev_timer> watchers, as
		2105	those cannot react to time jumps.
1330		2106
1331	As with timers, the callback is guaranteed to be invoked only when the	2107	As with timers, the callback is guaranteed to be invoked only when the
1332	time (C<at>) has passed, but if multiple periodic timers become ready	2108	point in time where it is supposed to trigger has passed. If multiple
1333	during the same loop iteration then order of execution is undefined.	2109	timers become ready during the same loop iteration then the ones with
		2110	earlier time-out values are invoked before ones with later time-out values
		2111	(but this is no longer true when a callback calls C<ev_run> recursively).
1334		2112
1335	=head3 Watcher-Specific Functions and Data Members	2113	=head3 Watcher-Specific Functions and Data Members
1336		2114
1337	=over 4	2115	=over 4
1338		2116
1339	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)	2117	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp offset, ev_tstamp interval, reschedule_cb)
1340		2118
1341	=item ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)	2119	=item ev_periodic_set (ev_periodic *, ev_tstamp offset, ev_tstamp interval, reschedule_cb)
1342		2120
1343	Lots of arguments, lets sort it out... There are basically three modes of	2121	Lots of arguments, let's sort it out... There are basically three modes of
1344	operation, and we will explain them from simplest to complex:	2122	operation, and we will explain them from simplest to most complex:
1345		2123
1346	=over 4	2124	=over 4
1347		2125
1348	=item * absolute timer (at = time, interval = reschedule_cb = 0)	2126	=item * absolute timer (offset = absolute time, interval = 0, reschedule_cb = 0)
1349		2127
1350	In this configuration the watcher triggers an event after the wall clock	2128	In this configuration the watcher triggers an event after the wall clock
1351	time C<at> has passed and doesn't repeat. It will not adjust when a time	2129	time C<offset> has passed. It will not repeat and will not adjust when a
1352	jump occurs, that is, if it is to be run at January 1st 2011 then it will	2130	time jump occurs, that is, if it is to be run at January 1st 2011 then it
1353	run when the system time reaches or surpasses this time.	2131	will be stopped and invoked when the system clock reaches or surpasses
		2132	this point in time.
1354		2133
1355	=item * repeating interval timer (at = offset, interval > 0, reschedule_cb = 0)	2134	=item * repeating interval timer (offset = offset within interval, interval > 0, reschedule_cb = 0)
1356		2135
1357	In this mode the watcher will always be scheduled to time out at the next	2136	In this mode the watcher will always be scheduled to time out at the next
1358	C<at + N * interval> time (for some integer N, which can also be negative)	2137	C<offset + N * interval> time (for some integer N, which can also be
1359	and then repeat, regardless of any time jumps.	2138	negative) and then repeat, regardless of any time jumps. The C<offset>
		2139	argument is merely an offset into the C<interval> periods.
1360		2140
1361	This can be used to create timers that do not drift with respect to system	2141	This can be used to create timers that do not drift with respect to the
1362	time, for example, here is a C<ev_periodic> that triggers each hour, on	2142	system clock, for example, here is an C<ev_periodic> that triggers each
1363	the hour:	2143	hour, on the hour (with respect to UTC):
1364		2144
1365	ev_periodic_set (&periodic, 0., 3600., 0);	2145	ev_periodic_set (&periodic, 0., 3600., 0);
1366		2146
1367	This doesn't mean there will always be 3600 seconds in between triggers,	2147	This doesn't mean there will always be 3600 seconds in between triggers,
1368	but only that the callback will be called when the system time shows a	2148	but only that the callback will be called when the system time shows a
1369	full hour (UTC), or more correctly, when the system time is evenly divisible	2149	full hour (UTC), or more correctly, when the system time is evenly divisible
1370	by 3600.	2150	by 3600.
1371		2151
1372	Another way to think about it (for the mathematically inclined) is that	2152	Another way to think about it (for the mathematically inclined) is that
1373	C<ev_periodic> will try to run the callback in this mode at the next possible	2153	C<ev_periodic> will try to run the callback in this mode at the next possible
1374	time where C<time = at (mod interval)>, regardless of any time jumps.	2154	time where C<time = offset (mod interval)>, regardless of any time jumps.
1375		2155
1376	For numerical stability it is preferable that the C<at> value is near	2156	For numerical stability it is preferable that the C<offset> value is near
1377	C<ev_now ()> (the current time), but there is no range requirement for	2157	C<ev_now ()> (the current time), but there is no range requirement for
1378	this value, and in fact is often specified as zero.	2158	this value, and in fact is often specified as zero.
1379		2159
1380	Note also that there is an upper limit to how often a timer can fire (CPU	2160	Note also that there is an upper limit to how often a timer can fire (CPU
1381	speed for example), so if C<interval> is very small then timing stability	2161	speed for example), so if C<interval> is very small then timing stability
1382	will of course deteriorate. Libev itself tries to be exact to be about one	2162	will of course deteriorate. Libev itself tries to be exact to be about one
1383	millisecond (if the OS supports it and the machine is fast enough).	2163	millisecond (if the OS supports it and the machine is fast enough).
1384		2164
1385	=item * manual reschedule mode (at and interval ignored, reschedule_cb = callback)	2165	=item * manual reschedule mode (offset ignored, interval ignored, reschedule_cb = callback)
1386		2166
1387	In this mode the values for C<interval> and C<at> are both being	2167	In this mode the values for C<interval> and C<offset> are both being
1388	ignored. Instead, each time the periodic watcher gets scheduled, the	2168	ignored. Instead, each time the periodic watcher gets scheduled, the
1389	reschedule callback will be called with the watcher as first, and the	2169	reschedule callback will be called with the watcher as first, and the
1390	current time as second argument.	2170	current time as second argument.
1391		2171
1392	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher,	2172	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher, ever,
1393	ever, or make ANY event loop modifications whatsoever>.	2173	or make ANY other event loop modifications whatsoever, unless explicitly
		2174	allowed by documentation here>.
1394		2175
1395	If you need to stop it, return C<now + 1e30> (or so, fudge fudge) and stop	2176	If you need to stop it, return C<now + 1e30> (or so, fudge fudge) and stop
1396	it afterwards (e.g. by starting an C<ev_prepare> watcher, which is the	2177	it afterwards (e.g. by starting an C<ev_prepare> watcher, which is the
1397	only event loop modification you are allowed to do).	2178	only event loop modification you are allowed to do).
1398		2179
1399	The callback prototype is C<ev_tstamp (*reschedule_cb)(struct ev_periodic	2180	The callback prototype is C<ev_tstamp (*reschedule_cb)(ev_periodic
1400	*w, ev_tstamp now)>, e.g.:	2181	*w, ev_tstamp now)>, e.g.:
1401		2182
		2183	static ev_tstamp
1402	static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)	2184	my_rescheduler (ev_periodic *w, ev_tstamp now)
1403	{	2185	{
1404	return now + 60.;	2186	return now + 60.;
1405	}	2187	}
1406		2188
1407	It must return the next time to trigger, based on the passed time value	2189	It must return the next time to trigger, based on the passed time value
…		…
1427	a different time than the last time it was called (e.g. in a crond like	2209	a different time than the last time it was called (e.g. in a crond like
1428	program when the crontabs have changed).	2210	program when the crontabs have changed).
1429		2211
1430	=item ev_tstamp ev_periodic_at (ev_periodic *)	2212	=item ev_tstamp ev_periodic_at (ev_periodic *)
1431		2213
1432	When active, returns the absolute time that the watcher is supposed to	2214	When active, returns the absolute time that the watcher is supposed
1433	trigger next.	2215	to trigger next. This is not the same as the C<offset> argument to
		2216	C<ev_periodic_set>, but indeed works even in interval and manual
		2217	rescheduling modes.
1434		2218
1435	=item ev_tstamp offset [read-write]	2219	=item ev_tstamp offset [read-write]
1436		2220
1437	When repeating, this contains the offset value, otherwise this is the	2221	When repeating, this contains the offset value, otherwise this is the
1438	absolute point in time (the C<at> value passed to C<ev_periodic_set>).	2222	absolute point in time (the C<offset> value passed to C<ev_periodic_set>,
		2223	although libev might modify this value for better numerical stability).
1439		2224
1440	Can be modified any time, but changes only take effect when the periodic	2225	Can be modified any time, but changes only take effect when the periodic
1441	timer fires or C<ev_periodic_again> is being called.	2226	timer fires or C<ev_periodic_again> is being called.
1442		2227
1443	=item ev_tstamp interval [read-write]	2228	=item ev_tstamp interval [read-write]
1444		2229
1445	The current interval value. Can be modified any time, but changes only	2230	The current interval value. Can be modified any time, but changes only
1446	take effect when the periodic timer fires or C<ev_periodic_again> is being	2231	take effect when the periodic timer fires or C<ev_periodic_again> is being
1447	called.	2232	called.
1448		2233
1449	=item ev_tstamp (reschedule_cb)(struct ev_periodic w, ev_tstamp now) [read-write]	2234	=item ev_tstamp (reschedule_cb)(ev_periodic w, ev_tstamp now) [read-write]
1450		2235
1451	The current reschedule callback, or C<0>, if this functionality is	2236	The current reschedule callback, or C<0>, if this functionality is
1452	switched off. Can be changed any time, but changes only take effect when	2237	switched off. Can be changed any time, but changes only take effect when
1453	the periodic timer fires or C<ev_periodic_again> is being called.	2238	the periodic timer fires or C<ev_periodic_again> is being called.
1454		2239
1455	=back	2240	=back
1456		2241
1457	=head3 Examples	2242	=head3 Examples
1458		2243
1459	Example: Call a callback every hour, or, more precisely, whenever the	2244	Example: Call a callback every hour, or, more precisely, whenever the
1460	system clock is divisible by 3600. The callback invocation times have	2245	system time is divisible by 3600. The callback invocation times have
1461	potentially a lot of jitter, but good long-term stability.	2246	potentially a lot of jitter, but good long-term stability.
1462		2247
1463	static void	2248	static void
1464	clock_cb (struct ev_loop loop, struct ev_io w, int revents)	2249	clock_cb (struct ev_loop loop, ev_periodic w, int revents)
1465	{	2250	{
1466	... its now a full hour (UTC, or TAI or whatever your clock follows)	2251	... its now a full hour (UTC, or TAI or whatever your clock follows)
1467	}	2252	}
1468		2253
1469	struct ev_periodic hourly_tick;	2254	ev_periodic hourly_tick;
1470	ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);	2255	ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);
1471	ev_periodic_start (loop, &hourly_tick);	2256	ev_periodic_start (loop, &hourly_tick);
1472		2257
1473	Example: The same as above, but use a reschedule callback to do it:	2258	Example: The same as above, but use a reschedule callback to do it:
1474		2259
1475	#include <math.h>	2260	#include <math.h>
1476		2261
1477	static ev_tstamp	2262	static ev_tstamp
1478	my_scheduler_cb (struct ev_periodic *w, ev_tstamp now)	2263	my_scheduler_cb (ev_periodic *w, ev_tstamp now)
1479	{	2264	{
1480	return fmod (now, 3600.) + 3600.;	2265	return now + (3600. - fmod (now, 3600.));
1481	}	2266	}
1482		2267
1483	ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);	2268	ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
1484		2269
1485	Example: Call a callback every hour, starting now:	2270	Example: Call a callback every hour, starting now:
1486		2271
1487	struct ev_periodic hourly_tick;	2272	ev_periodic hourly_tick;
1488	ev_periodic_init (&hourly_tick, clock_cb,	2273	ev_periodic_init (&hourly_tick, clock_cb,
1489	fmod (ev_now (loop), 3600.), 3600., 0);	2274	fmod (ev_now (loop), 3600.), 3600., 0);
1490	ev_periodic_start (loop, &hourly_tick);	2275	ev_periodic_start (loop, &hourly_tick);
1491		2276
1492		2277
1493	=head2 C<ev_signal> - signal me when a signal gets signalled!	2278	=head2 C<ev_signal> - signal me when a signal gets signalled!
1494		2279
1495	Signal watchers will trigger an event when the process receives a specific	2280	Signal watchers will trigger an event when the process receives a specific
1496	signal one or more times. Even though signals are very asynchronous, libev	2281	signal one or more times. Even though signals are very asynchronous, libev
1497	will try it's best to deliver signals synchronously, i.e. as part of the	2282	will try its best to deliver signals synchronously, i.e. as part of the
1498	normal event processing, like any other event.	2283	normal event processing, like any other event.
1499		2284
		2285	If you want signals to be delivered truly asynchronously, just use
		2286	C<sigaction> as you would do without libev and forget about sharing
		2287	the signal. You can even use C<ev_async> from a signal handler to
		2288	synchronously wake up an event loop.
		2289
1500	You can configure as many watchers as you like per signal. Only when the	2290	You can configure as many watchers as you like for the same signal, but
		2291	only within the same loop, i.e. you can watch for C<SIGINT> in your
		2292	default loop and for C<SIGIO> in another loop, but you cannot watch for
		2293	C<SIGINT> in both the default loop and another loop at the same time. At
		2294	the moment, C<SIGCHLD> is permanently tied to the default loop.
		2295
1501	first watcher gets started will libev actually register a signal watcher	2296	When the first watcher gets started will libev actually register something
1502	with the kernel (thus it coexists with your own signal handlers as long	2297	with the kernel (thus it coexists with your own signal handlers as long as
1503	as you don't register any with libev). Similarly, when the last signal	2298	you don't register any with libev for the same signal).
1504	watcher for a signal is stopped libev will reset the signal handler to
1505	SIG_DFL (regardless of what it was set to before).
1506		2299
1507	If possible and supported, libev will install its handlers with	2300	If possible and supported, libev will install its handlers with
1508	C<SA_RESTART> behaviour enabled, so system calls should not be unduly	2301	C<SA_RESTART> (or equivalent) behaviour enabled, so system calls should
1509	interrupted. If you have a problem with system calls getting interrupted by	2302	not be unduly interrupted. If you have a problem with system calls getting
1510	signals you can block all signals in an C<ev_check> watcher and unblock	2303	interrupted by signals you can block all signals in an C<ev_check> watcher
1511	them in an C<ev_prepare> watcher.	2304	and unblock them in an C<ev_prepare> watcher.
		2305
		2306	=head3 The special problem of inheritance over fork/execve/pthread_create
		2307
		2308	Both the signal mask (C<sigprocmask>) and the signal disposition
		2309	(C<sigaction>) are unspecified after starting a signal watcher (and after
		2310	stopping it again), that is, libev might or might not block the signal,
		2311	and might or might not set or restore the installed signal handler (but
		2312	see C<EVFLAG_NOSIGMASK>).
		2313
		2314	While this does not matter for the signal disposition (libev never
		2315	sets signals to C<SIG_IGN>, so handlers will be reset to C<SIG_DFL> on
		2316	C<execve>), this matters for the signal mask: many programs do not expect
		2317	certain signals to be blocked.
		2318
		2319	This means that before calling C<exec> (from the child) you should reset
		2320	the signal mask to whatever "default" you expect (all clear is a good
		2321	choice usually).
		2322
		2323	The simplest way to ensure that the signal mask is reset in the child is
		2324	to install a fork handler with C<pthread_atfork> that resets it. That will
		2325	catch fork calls done by libraries (such as the libc) as well.
		2326
		2327	In current versions of libev, the signal will not be blocked indefinitely
		2328	unless you use the C<signalfd> API (C<EV_SIGNALFD>). While this reduces
		2329	the window of opportunity for problems, it will not go away, as libev
		2330	I<has> to modify the signal mask, at least temporarily.
		2331
		2332	So I can't stress this enough: I<If you do not reset your signal mask when
		2333	you expect it to be empty, you have a race condition in your code>. This
		2334	is not a libev-specific thing, this is true for most event libraries.
		2335
		2336	=head3 The special problem of threads signal handling
		2337
		2338	POSIX threads has problematic signal handling semantics, specifically,
		2339	a lot of functionality (sigfd, sigwait etc.) only really works if all
		2340	threads in a process block signals, which is hard to achieve.
		2341
		2342	When you want to use sigwait (or mix libev signal handling with your own
		2343	for the same signals), you can tackle this problem by globally blocking
		2344	all signals before creating any threads (or creating them with a fully set
		2345	sigprocmask) and also specifying the C<EVFLAG_NOSIGMASK> when creating
		2346	loops. Then designate one thread as "signal receiver thread" which handles
		2347	these signals. You can pass on any signals that libev might be interested
		2348	in by calling C<ev_feed_signal>.
1512		2349
1513	=head3 Watcher-Specific Functions and Data Members	2350	=head3 Watcher-Specific Functions and Data Members
1514		2351
1515	=over 4	2352	=over 4
1516		2353
…		…
1527		2364
1528	=back	2365	=back
1529		2366
1530	=head3 Examples	2367	=head3 Examples
1531		2368
1532	Example: Try to exit cleanly on SIGINT and SIGTERM.	2369	Example: Try to exit cleanly on SIGINT.
1533		2370
1534	static void	2371	static void
1535	sigint_cb (struct ev_loop loop, struct ev_signal w, int revents)	2372	sigint_cb (struct ev_loop loop, ev_signal w, int revents)
1536	{	2373	{
1537	ev_unloop (loop, EVUNLOOP_ALL);	2374	ev_break (loop, EVBREAK_ALL);
1538	}	2375	}
1539		2376
1540	struct ev_signal signal_watcher;	2377	ev_signal signal_watcher;
1541	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);	2378	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
1542	ev_signal_start (loop, &sigint_cb);	2379	ev_signal_start (loop, &signal_watcher);
1543		2380
1544		2381
1545	=head2 C<ev_child> - watch out for process status changes	2382	=head2 C<ev_child> - watch out for process status changes
1546		2383
1547	Child watchers trigger when your process receives a SIGCHLD in response to	2384	Child watchers trigger when your process receives a SIGCHLD in response to
1548	some child status changes (most typically when a child of yours dies). It	2385	some child status changes (most typically when a child of yours dies or
1549	is permissible to install a child watcher I<after> the child has been	2386	exits). It is permissible to install a child watcher I<after> the child
1550	forked (which implies it might have already exited), as long as the event	2387	has been forked (which implies it might have already exited), as long
1551	loop isn't entered (or is continued from a watcher).	2388	as the event loop isn't entered (or is continued from a watcher), i.e.,
		2389	forking and then immediately registering a watcher for the child is fine,
		2390	but forking and registering a watcher a few event loop iterations later or
		2391	in the next callback invocation is not.
1552		2392
1553	Only the default event loop is capable of handling signals, and therefore	2393	Only the default event loop is capable of handling signals, and therefore
1554	you can only register child watchers in the default event loop.	2394	you can only register child watchers in the default event loop.
1555		2395
		2396	Due to some design glitches inside libev, child watchers will always be
		2397	handled at maximum priority (their priority is set to C<EV_MAXPRI> by
		2398	libev)
		2399
1556	=head3 Process Interaction	2400	=head3 Process Interaction
1557		2401
1558	Libev grabs C<SIGCHLD> as soon as the default event loop is	2402	Libev grabs C<SIGCHLD> as soon as the default event loop is
1559	initialised. This is necessary to guarantee proper behaviour even if	2403	initialised. This is necessary to guarantee proper behaviour even if the
1560	the first child watcher is started after the child exits. The occurrence	2404	first child watcher is started after the child exits. The occurrence
1561	of C<SIGCHLD> is recorded asynchronously, but child reaping is done	2405	of C<SIGCHLD> is recorded asynchronously, but child reaping is done
1562	synchronously as part of the event loop processing. Libev always reaps all	2406	synchronously as part of the event loop processing. Libev always reaps all
1563	children, even ones not watched.	2407	children, even ones not watched.
1564		2408
1565	=head3 Overriding the Built-In Processing	2409	=head3 Overriding the Built-In Processing
…		…
1569	handler, you can override it easily by installing your own handler for	2413	handler, you can override it easily by installing your own handler for
1570	C<SIGCHLD> after initialising the default loop, and making sure the	2414	C<SIGCHLD> after initialising the default loop, and making sure the
1571	default loop never gets destroyed. You are encouraged, however, to use an	2415	default loop never gets destroyed. You are encouraged, however, to use an
1572	event-based approach to child reaping and thus use libev's support for	2416	event-based approach to child reaping and thus use libev's support for
1573	that, so other libev users can use C<ev_child> watchers freely.	2417	that, so other libev users can use C<ev_child> watchers freely.
		2418
		2419	=head3 Stopping the Child Watcher
		2420
		2421	Currently, the child watcher never gets stopped, even when the
		2422	child terminates, so normally one needs to stop the watcher in the
		2423	callback. Future versions of libev might stop the watcher automatically
		2424	when a child exit is detected (calling C<ev_child_stop> twice is not a
		2425	problem).
1574		2426
1575	=head3 Watcher-Specific Functions and Data Members	2427	=head3 Watcher-Specific Functions and Data Members
1576		2428
1577	=over 4	2429	=over 4
1578		2430
…		…
1610	its completion.	2462	its completion.
1611		2463
1612	ev_child cw;	2464	ev_child cw;
1613		2465
1614	static void	2466	static void
1615	child_cb (EV_P_ struct ev_child *w, int revents)	2467	child_cb (EV_P_ ev_child *w, int revents)
1616	{	2468	{
1617	ev_child_stop (EV_A_ w);	2469	ev_child_stop (EV_A_ w);
1618	printf ("process %d exited with status %x\n", w->rpid, w->rstatus);	2470	printf ("process %d exited with status %x\n", w->rpid, w->rstatus);
1619	}	2471	}
1620		2472
…		…
1635		2487
1636		2488
1637	=head2 C<ev_stat> - did the file attributes just change?	2489	=head2 C<ev_stat> - did the file attributes just change?
1638		2490
1639	This watches a file system path for attribute changes. That is, it calls	2491	This watches a file system path for attribute changes. That is, it calls
1640	C<stat> regularly (or when the OS says it changed) and sees if it changed	2492	C<stat> on that path in regular intervals (or when the OS says it changed)
1641	compared to the last time, invoking the callback if it did.	2493	and sees if it changed compared to the last time, invoking the callback if
		2494	it did.
1642		2495
1643	The path does not need to exist: changing from "path exists" to "path does	2496	The path does not need to exist: changing from "path exists" to "path does
1644	not exist" is a status change like any other. The condition "path does	2497	not exist" is a status change like any other. The condition "path does not
1645	not exist" is signified by the C<st_nlink> field being zero (which is	2498	exist" (or more correctly "path cannot be stat'ed") is signified by the
1646	otherwise always forced to be at least one) and all the other fields of	2499	C<st_nlink> field being zero (which is otherwise always forced to be at
1647	the stat buffer having unspecified contents.	2500	least one) and all the other fields of the stat buffer having unspecified
		2501	contents.
1648		2502
1649	The path I<should> be absolute and I<must not> end in a slash. If it is	2503	The path I<must not> end in a slash or contain special components such as
		2504	C<.> or C<..>. The path I<should> be absolute: If it is relative and
1650	relative and your working directory changes, the behaviour is undefined.	2505	your working directory changes, then the behaviour is undefined.
1651		2506
1652	Since there is no standard to do this, the portable implementation simply	2507	Since there is no portable change notification interface available, the
1653	calls C<stat (2)> regularly on the path to see if it changed somehow. You	2508	portable implementation simply calls C<stat(2)> regularly on the path
1654	can specify a recommended polling interval for this case. If you specify	2509	to see if it changed somehow. You can specify a recommended polling
1655	a polling interval of C<0> (highly recommended!) then a I<suitable,	2510	interval for this case. If you specify a polling interval of C<0> (highly
1656	unspecified default> value will be used (which you can expect to be around	2511	recommended!) then a I<suitable, unspecified default> value will be used
1657	five seconds, although this might change dynamically). Libev will also	2512	(which you can expect to be around five seconds, although this might
1658	impose a minimum interval which is currently around C<0.1>, but thats	2513	change dynamically). Libev will also impose a minimum interval which is
1659	usually overkill.	2514	currently around C<0.1>, but that's usually overkill.
1660		2515
1661	This watcher type is not meant for massive numbers of stat watchers,	2516	This watcher type is not meant for massive numbers of stat watchers,
1662	as even with OS-supported change notifications, this can be	2517	as even with OS-supported change notifications, this can be
1663	resource-intensive.	2518	resource-intensive.
1664		2519
1665	At the time of this writing, only the Linux inotify interface is	2520	At the time of this writing, the only OS-specific interface implemented
1666	implemented (implementing kqueue support is left as an exercise for the	2521	is the Linux inotify interface (implementing kqueue support is left as an
1667	reader, note, however, that the author sees no way of implementing ev_stat	2522	exercise for the reader. Note, however, that the author sees no way of
1668	semantics with kqueue). Inotify will be used to give hints only and should	2523	implementing C<ev_stat> semantics with kqueue, except as a hint).
1669	not change the semantics of C<ev_stat> watchers, which means that libev
1670	sometimes needs to fall back to regular polling again even with inotify,
1671	but changes are usually detected immediately, and if the file exists there
1672	will be no polling.
1673		2524
1674	=head3 ABI Issues (Largefile Support)	2525	=head3 ABI Issues (Largefile Support)
1675		2526
1676	Libev by default (unless the user overrides this) uses the default	2527	Libev by default (unless the user overrides this) uses the default
1677	compilation environment, which means that on systems with large file	2528	compilation environment, which means that on systems with large file
1678	support disabled by default, you get the 32 bit version of the stat	2529	support disabled by default, you get the 32 bit version of the stat
1679	structure. When using the library from programs that change the ABI to	2530	structure. When using the library from programs that change the ABI to
1680	use 64 bit file offsets the programs will fail. In that case you have to	2531	use 64 bit file offsets the programs will fail. In that case you have to
1681	compile libev with the same flags to get binary compatibility. This is	2532	compile libev with the same flags to get binary compatibility. This is
1682	obviously the case with any flags that change the ABI, but the problem is	2533	obviously the case with any flags that change the ABI, but the problem is
1683	most noticeably disabled with ev_stat and large file support.	2534	most noticeably displayed with ev_stat and large file support.
1684		2535
1685	The solution for this is to lobby your distribution maker to make large	2536	The solution for this is to lobby your distribution maker to make large
1686	file interfaces available by default (as e.g. FreeBSD does) and not	2537	file interfaces available by default (as e.g. FreeBSD does) and not
1687	optional. Libev cannot simply switch on large file support because it has	2538	optional. Libev cannot simply switch on large file support because it has
1688	to exchange stat structures with application programs compiled using the	2539	to exchange stat structures with application programs compiled using the
1689	default compilation environment.	2540	default compilation environment.
1690		2541
1691	=head3 Inotify	2542	=head3 Inotify and Kqueue
1692		2543
1693	When C<inotify (7)> support has been compiled into libev (generally only	2544	When C<inotify (7)> support has been compiled into libev and present at
1694	available on Linux) and present at runtime, it will be used to speed up	2545	runtime, it will be used to speed up change detection where possible. The
1695	change detection where possible. The inotify descriptor will be created lazily	2546	inotify descriptor will be created lazily when the first C<ev_stat>
1696	when the first C<ev_stat> watcher is being started.	2547	watcher is being started.
1697		2548
1698	Inotify presence does not change the semantics of C<ev_stat> watchers	2549	Inotify presence does not change the semantics of C<ev_stat> watchers
1699	except that changes might be detected earlier, and in some cases, to avoid	2550	except that changes might be detected earlier, and in some cases, to avoid
1700	making regular C<stat> calls. Even in the presence of inotify support	2551	making regular C<stat> calls. Even in the presence of inotify support
1701	there are many cases where libev has to resort to regular C<stat> polling.	2552	there are many cases where libev has to resort to regular C<stat> polling,
		2553	but as long as kernel 2.6.25 or newer is used (2.6.24 and older have too
		2554	many bugs), the path exists (i.e. stat succeeds), and the path resides on
		2555	a local filesystem (libev currently assumes only ext2/3, jfs, reiserfs and
		2556	xfs are fully working) libev usually gets away without polling.
1702		2557
1703	(There is no support for kqueue, as apparently it cannot be used to	2558	There is no support for kqueue, as apparently it cannot be used to
1704	implement this functionality, due to the requirement of having a file	2559	implement this functionality, due to the requirement of having a file
1705	descriptor open on the object at all times).	2560	descriptor open on the object at all times, and detecting renames, unlinks
		2561	etc. is difficult.
		2562
		2563	=head3 C<stat ()> is a synchronous operation
		2564
		2565	Libev doesn't normally do any kind of I/O itself, and so is not blocking
		2566	the process. The exception are C<ev_stat> watchers - those call C<stat
		2567	()>, which is a synchronous operation.
		2568
		2569	For local paths, this usually doesn't matter: unless the system is very
		2570	busy or the intervals between stat's are large, a stat call will be fast,
		2571	as the path data is usually in memory already (except when starting the
		2572	watcher).
		2573
		2574	For networked file systems, calling C<stat ()> can block an indefinite
		2575	time due to network issues, and even under good conditions, a stat call
		2576	often takes multiple milliseconds.
		2577
		2578	Therefore, it is best to avoid using C<ev_stat> watchers on networked
		2579	paths, although this is fully supported by libev.
1706		2580
1707	=head3 The special problem of stat time resolution	2581	=head3 The special problem of stat time resolution
1708		2582
1709	The C<stat ()> system call only supports full-second resolution portably, and	2583	The C<stat ()> system call only supports full-second resolution portably,
1710	even on systems where the resolution is higher, many file systems still	2584	and even on systems where the resolution is higher, most file systems
1711	only support whole seconds.	2585	still only support whole seconds.
1712		2586
1713	That means that, if the time is the only thing that changes, you can	2587	That means that, if the time is the only thing that changes, you can
1714	easily miss updates: on the first update, C<ev_stat> detects a change and	2588	easily miss updates: on the first update, C<ev_stat> detects a change and
1715	calls your callback, which does something. When there is another update	2589	calls your callback, which does something. When there is another update
1716	within the same second, C<ev_stat> will be unable to detect it as the stat	2590	within the same second, C<ev_stat> will be unable to detect unless the
1717	data does not change.	2591	stat data does change in other ways (e.g. file size).
1718		2592
1719	The solution to this is to delay acting on a change for slightly more	2593	The solution to this is to delay acting on a change for slightly more
1720	than a second (or till slightly after the next full second boundary), using	2594	than a second (or till slightly after the next full second boundary), using
1721	a roughly one-second-delay C<ev_timer> (e.g. C<ev_timer_set (w, 0., 1.02);	2595	a roughly one-second-delay C<ev_timer> (e.g. C<ev_timer_set (w, 0., 1.02);
1722	ev_timer_again (loop, w)>).	2596	ev_timer_again (loop, w)>).
…		…
1742	C<path>. The C<interval> is a hint on how quickly a change is expected to	2616	C<path>. The C<interval> is a hint on how quickly a change is expected to
1743	be detected and should normally be specified as C<0> to let libev choose	2617	be detected and should normally be specified as C<0> to let libev choose
1744	a suitable value. The memory pointed to by C<path> must point to the same	2618	a suitable value. The memory pointed to by C<path> must point to the same
1745	path for as long as the watcher is active.	2619	path for as long as the watcher is active.
1746		2620
1747	The callback will receive C<EV_STAT> when a change was detected, relative	2621	The callback will receive an C<EV_STAT> event when a change was detected,
1748	to the attributes at the time the watcher was started (or the last change	2622	relative to the attributes at the time the watcher was started (or the
1749	was detected).	2623	last change was detected).
1750		2624
1751	=item ev_stat_stat (loop, ev_stat *)	2625	=item ev_stat_stat (loop, ev_stat *)
1752		2626
1753	Updates the stat buffer immediately with new values. If you change the	2627	Updates the stat buffer immediately with new values. If you change the
1754	watched path in your callback, you could call this function to avoid	2628	watched path in your callback, you could call this function to avoid
…		…
1837		2711
1838		2712
1839	=head2 C<ev_idle> - when you've got nothing better to do...	2713	=head2 C<ev_idle> - when you've got nothing better to do...
1840		2714
1841	Idle watchers trigger events when no other events of the same or higher	2715	Idle watchers trigger events when no other events of the same or higher
1842	priority are pending (prepare, check and other idle watchers do not	2716	priority are pending (prepare, check and other idle watchers do not count
1843	count).	2717	as receiving "events").
1844		2718
1845	That is, as long as your process is busy handling sockets or timeouts	2719	That is, as long as your process is busy handling sockets or timeouts
1846	(or even signals, imagine) of the same or higher priority it will not be	2720	(or even signals, imagine) of the same or higher priority it will not be
1847	triggered. But when your process is idle (or only lower-priority watchers	2721	triggered. But when your process is idle (or only lower-priority watchers
1848	are pending), the idle watchers are being called once per event loop	2722	are pending), the idle watchers are being called once per event loop
…		…
1859		2733
1860	=head3 Watcher-Specific Functions and Data Members	2734	=head3 Watcher-Specific Functions and Data Members
1861		2735
1862	=over 4	2736	=over 4
1863		2737
1864	=item ev_idle_init (ev_signal *, callback)	2738	=item ev_idle_init (ev_idle *, callback)
1865		2739
1866	Initialises and configures the idle watcher - it has no parameters of any	2740	Initialises and configures the idle watcher - it has no parameters of any
1867	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,	2741	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,
1868	believe me.	2742	believe me.
1869		2743
…		…
1873		2747
1874	Example: Dynamically allocate an C<ev_idle> watcher, start it, and in the	2748	Example: Dynamically allocate an C<ev_idle> watcher, start it, and in the
1875	callback, free it. Also, use no error checking, as usual.	2749	callback, free it. Also, use no error checking, as usual.
1876		2750
1877	static void	2751	static void
1878	idle_cb (struct ev_loop loop, struct ev_idle w, int revents)	2752	idle_cb (struct ev_loop loop, ev_idle w, int revents)
1879	{	2753	{
1880	free (w);	2754	free (w);
1881	// now do something you wanted to do when the program has	2755	// now do something you wanted to do when the program has
1882	// no longer anything immediate to do.	2756	// no longer anything immediate to do.
1883	}	2757	}
1884		2758
1885	struct ev_idle *idle_watcher = malloc (sizeof (struct ev_idle));	2759	ev_idle *idle_watcher = malloc (sizeof (ev_idle));
1886	ev_idle_init (idle_watcher, idle_cb);	2760	ev_idle_init (idle_watcher, idle_cb);
1887	ev_idle_start (loop, idle_cb);	2761	ev_idle_start (loop, idle_watcher);
1888		2762
1889		2763
1890	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!	2764	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!
1891		2765
1892	Prepare and check watchers are usually (but not always) used in tandem:	2766	Prepare and check watchers are usually (but not always) used in pairs:
1893	prepare watchers get invoked before the process blocks and check watchers	2767	prepare watchers get invoked before the process blocks and check watchers
1894	afterwards.	2768	afterwards.
1895		2769
1896	You I<must not> call C<ev_loop> or similar functions that enter	2770	You I<must not> call C<ev_run> or similar functions that enter
1897	the current event loop from either C<ev_prepare> or C<ev_check>	2771	the current event loop from either C<ev_prepare> or C<ev_check>
1898	watchers. Other loops than the current one are fine, however. The	2772	watchers. Other loops than the current one are fine, however. The
1899	rationale behind this is that you do not need to check for recursion in	2773	rationale behind this is that you do not need to check for recursion in
1900	those watchers, i.e. the sequence will always be C<ev_prepare>, blocking,	2774	those watchers, i.e. the sequence will always be C<ev_prepare>, blocking,
1901	C<ev_check> so if you have one watcher of each kind they will always be	2775	C<ev_check> so if you have one watcher of each kind they will always be
1902	called in pairs bracketing the blocking call.	2776	called in pairs bracketing the blocking call.
1903		2777
1904	Their main purpose is to integrate other event mechanisms into libev and	2778	Their main purpose is to integrate other event mechanisms into libev and
1905	their use is somewhat advanced. This could be used, for example, to track	2779	their use is somewhat advanced. They could be used, for example, to track
1906	variable changes, implement your own watchers, integrate net-snmp or a	2780	variable changes, implement your own watchers, integrate net-snmp or a
1907	coroutine library and lots more. They are also occasionally useful if	2781	coroutine library and lots more. They are also occasionally useful if
1908	you cache some data and want to flush it before blocking (for example,	2782	you cache some data and want to flush it before blocking (for example,
1909	in X programs you might want to do an C<XFlush ()> in an C<ev_prepare>	2783	in X programs you might want to do an C<XFlush ()> in an C<ev_prepare>
1910	watcher).	2784	watcher).
1911		2785
1912	This is done by examining in each prepare call which file descriptors need	2786	This is done by examining in each prepare call which file descriptors
1913	to be watched by the other library, registering C<ev_io> watchers for	2787	need to be watched by the other library, registering C<ev_io> watchers
1914	them and starting an C<ev_timer> watcher for any timeouts (many libraries	2788	for them and starting an C<ev_timer> watcher for any timeouts (many
1915	provide just this functionality). Then, in the check watcher you check for	2789	libraries provide exactly this functionality). Then, in the check watcher,
1916	any events that occurred (by checking the pending status of all watchers	2790	you check for any events that occurred (by checking the pending status
1917	and stopping them) and call back into the library. The I/O and timer	2791	of all watchers and stopping them) and call back into the library. The
1918	callbacks will never actually be called (but must be valid nevertheless,	2792	I/O and timer callbacks will never actually be called (but must be valid
1919	because you never know, you know?).	2793	nevertheless, because you never know, you know?).
1920		2794
1921	As another example, the Perl Coro module uses these hooks to integrate	2795	As another example, the Perl Coro module uses these hooks to integrate
1922	coroutines into libev programs, by yielding to other active coroutines	2796	coroutines into libev programs, by yielding to other active coroutines
1923	during each prepare and only letting the process block if no coroutines	2797	during each prepare and only letting the process block if no coroutines
1924	are ready to run (it's actually more complicated: it only runs coroutines	2798	are ready to run (it's actually more complicated: it only runs coroutines
…		…
1927	loop from blocking if lower-priority coroutines are active, thus mapping	2801	loop from blocking if lower-priority coroutines are active, thus mapping
1928	low-priority coroutines to idle/background tasks).	2802	low-priority coroutines to idle/background tasks).
1929		2803
1930	It is recommended to give C<ev_check> watchers highest (C<EV_MAXPRI>)	2804	It is recommended to give C<ev_check> watchers highest (C<EV_MAXPRI>)
1931	priority, to ensure that they are being run before any other watchers	2805	priority, to ensure that they are being run before any other watchers
		2806	after the poll (this doesn't matter for C<ev_prepare> watchers).
		2807
1932	after the poll. Also, C<ev_check> watchers (and C<ev_prepare> watchers,	2808	Also, C<ev_check> watchers (and C<ev_prepare> watchers, too) should not
1933	too) should not activate ("feed") events into libev. While libev fully	2809	activate ("feed") events into libev. While libev fully supports this, they
1934	supports this, they might get executed before other C<ev_check> watchers	2810	might get executed before other C<ev_check> watchers did their job. As
1935	did their job. As C<ev_check> watchers are often used to embed other	2811	C<ev_check> watchers are often used to embed other (non-libev) event
1936	(non-libev) event loops those other event loops might be in an unusable	2812	loops those other event loops might be in an unusable state until their
1937	state until their C<ev_check> watcher ran (always remind yourself to	2813	C<ev_check> watcher ran (always remind yourself to coexist peacefully with
1938	coexist peacefully with others).	2814	others).
1939		2815
1940	=head3 Watcher-Specific Functions and Data Members	2816	=head3 Watcher-Specific Functions and Data Members
1941		2817
1942	=over 4	2818	=over 4
1943		2819
…		…
1945		2821
1946	=item ev_check_init (ev_check *, callback)	2822	=item ev_check_init (ev_check *, callback)
1947		2823
1948	Initialises and configures the prepare or check watcher - they have no	2824	Initialises and configures the prepare or check watcher - they have no
1949	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>	2825	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>
1950	macros, but using them is utterly, utterly and completely pointless.	2826	macros, but using them is utterly, utterly, utterly and completely
		2827	pointless.
1951		2828
1952	=back	2829	=back
1953		2830
1954	=head3 Examples	2831	=head3 Examples
1955		2832
…		…
1968		2845
1969	static ev_io iow [nfd];	2846	static ev_io iow [nfd];
1970	static ev_timer tw;	2847	static ev_timer tw;
1971		2848
1972	static void	2849	static void
1973	io_cb (ev_loop loop, ev_io w, int revents)	2850	io_cb (struct ev_loop loop, ev_io w, int revents)
1974	{	2851	{
1975	}	2852	}
1976		2853
1977	// create io watchers for each fd and a timer before blocking	2854	// create io watchers for each fd and a timer before blocking
1978	static void	2855	static void
1979	adns_prepare_cb (ev_loop loop, ev_prepare w, int revents)	2856	adns_prepare_cb (struct ev_loop loop, ev_prepare w, int revents)
1980	{	2857	{
1981	int timeout = 3600000;	2858	int timeout = 3600000;
1982	struct pollfd fds [nfd];	2859	struct pollfd fds [nfd];
1983	// actual code will need to loop here and realloc etc.	2860	// actual code will need to loop here and realloc etc.
1984	adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));	2861	adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));
1985		2862
1986	/* the callback is illegal, but won't be called as we stop during check */	2863	/* the callback is illegal, but won't be called as we stop during check */
1987	ev_timer_init (&tw, 0, timeout * 1e-3);	2864	ev_timer_init (&tw, 0, timeout * 1e-3, 0.);
1988	ev_timer_start (loop, &tw);	2865	ev_timer_start (loop, &tw);
1989		2866
1990	// create one ev_io per pollfd	2867	// create one ev_io per pollfd
1991	for (int i = 0; i < nfd; ++i)	2868	for (int i = 0; i < nfd; ++i)
1992	{	2869	{
…		…
1999	}	2876	}
2000	}	2877	}
2001		2878
2002	// stop all watchers after blocking	2879	// stop all watchers after blocking
2003	static void	2880	static void
2004	adns_check_cb (ev_loop loop, ev_check w, int revents)	2881	adns_check_cb (struct ev_loop loop, ev_check w, int revents)
2005	{	2882	{
2006	ev_timer_stop (loop, &tw);	2883	ev_timer_stop (loop, &tw);
2007		2884
2008	for (int i = 0; i < nfd; ++i)	2885	for (int i = 0; i < nfd; ++i)
2009	{	2886	{
…		…
2048	}	2925	}
2049		2926
2050	// do not ever call adns_afterpoll	2927	// do not ever call adns_afterpoll
2051		2928
2052	Method 4: Do not use a prepare or check watcher because the module you	2929	Method 4: Do not use a prepare or check watcher because the module you
2053	want to embed is too inflexible to support it. Instead, you can override	2930	want to embed is not flexible enough to support it. Instead, you can
2054	their poll function. The drawback with this solution is that the main	2931	override their poll function. The drawback with this solution is that the
2055	loop is now no longer controllable by EV. The C<Glib::EV> module does	2932	main loop is now no longer controllable by EV. The C<Glib::EV> module uses
2056	this.	2933	this approach, effectively embedding EV as a client into the horrible
		2934	libglib event loop.
2057		2935
2058	static gint	2936	static gint
2059	event_poll_func (GPollFD *fds, guint nfds, gint timeout)	2937	event_poll_func (GPollFD *fds, guint nfds, gint timeout)
2060	{	2938	{
2061	int got_events = 0;	2939	int got_events = 0;
…		…
2065		2943
2066	if (timeout >= 0)	2944	if (timeout >= 0)
2067	// create/start timer	2945	// create/start timer
2068		2946
2069	// poll	2947	// poll
2070	ev_loop (EV_A_ 0);	2948	ev_run (EV_A_ 0);
2071		2949
2072	// stop timer again	2950	// stop timer again
2073	if (timeout >= 0)	2951	if (timeout >= 0)
2074	ev_timer_stop (EV_A_ &to);	2952	ev_timer_stop (EV_A_ &to);
2075		2953
…		…
2092	prioritise I/O.	2970	prioritise I/O.
2093		2971
2094	As an example for a bug workaround, the kqueue backend might only support	2972	As an example for a bug workaround, the kqueue backend might only support
2095	sockets on some platform, so it is unusable as generic backend, but you	2973	sockets on some platform, so it is unusable as generic backend, but you
2096	still want to make use of it because you have many sockets and it scales	2974	still want to make use of it because you have many sockets and it scales
2097	so nicely. In this case, you would create a kqueue-based loop and embed it	2975	so nicely. In this case, you would create a kqueue-based loop and embed
2098	into your default loop (which might use e.g. poll). Overall operation will	2976	it into your default loop (which might use e.g. poll). Overall operation
2099	be a bit slower because first libev has to poll and then call kevent, but	2977	will be a bit slower because first libev has to call C<poll> and then
2100	at least you can use both at what they are best.	2978	C<kevent>, but at least you can use both mechanisms for what they are
		2979	best: C<kqueue> for scalable sockets and C<poll> if you want it to work :)
2101		2980
2102	As for prioritising I/O: rarely you have the case where some fds have	2981	As for prioritising I/O: under rare circumstances you have the case where
2103	to be watched and handled very quickly (with low latency), and even	2982	some fds have to be watched and handled very quickly (with low latency),
2104	priorities and idle watchers might have too much overhead. In this case	2983	and even priorities and idle watchers might have too much overhead. In
2105	you would put all the high priority stuff in one loop and all the rest in	2984	this case you would put all the high priority stuff in one loop and all
2106	a second one, and embed the second one in the first.	2985	the rest in a second one, and embed the second one in the first.
2107		2986
2108	As long as the watcher is active, the callback will be invoked every time	2987	As long as the watcher is active, the callback will be invoked every
2109	there might be events pending in the embedded loop. The callback must then	2988	time there might be events pending in the embedded loop. The callback
2110	call C<ev_embed_sweep (mainloop, watcher)> to make a single sweep and invoke	2989	must then call C<ev_embed_sweep (mainloop, watcher)> to make a single
2111	their callbacks (you could also start an idle watcher to give the embedded	2990	sweep and invoke their callbacks (the callback doesn't need to invoke the
2112	loop strictly lower priority for example). You can also set the callback	2991	C<ev_embed_sweep> function directly, it could also start an idle watcher
2113	to C<0>, in which case the embed watcher will automatically execute the	2992	to give the embedded loop strictly lower priority for example).
2114	embedded loop sweep.
2115		2993
2116	As long as the watcher is started it will automatically handle events. The	2994	You can also set the callback to C<0>, in which case the embed watcher
2117	callback will be invoked whenever some events have been handled. You can	2995	will automatically execute the embedded loop sweep whenever necessary.
2118	set the callback to C<0> to avoid having to specify one if you are not
2119	interested in that.
2120		2996
2121	Also, there have not currently been made special provisions for forking:	2997	Fork detection will be handled transparently while the C<ev_embed> watcher
2122	when you fork, you not only have to call C<ev_loop_fork> on both loops,	2998	is active, i.e., the embedded loop will automatically be forked when the
2123	but you will also have to stop and restart any C<ev_embed> watchers	2999	embedding loop forks. In other cases, the user is responsible for calling
2124	yourself.	3000	C<ev_loop_fork> on the embedded loop.
2125		3001
2126	Unfortunately, not all backends are embeddable, only the ones returned by	3002	Unfortunately, not all backends are embeddable: only the ones returned by
2127	C<ev_embeddable_backends> are, which, unfortunately, does not include any	3003	C<ev_embeddable_backends> are, which, unfortunately, does not include any
2128	portable one.	3004	portable one.
2129		3005
2130	So when you want to use this feature you will always have to be prepared	3006	So when you want to use this feature you will always have to be prepared
2131	that you cannot get an embeddable loop. The recommended way to get around	3007	that you cannot get an embeddable loop. The recommended way to get around
2132	this is to have a separate variables for your embeddable loop, try to	3008	this is to have a separate variables for your embeddable loop, try to
2133	create it, and if that fails, use the normal loop for everything.	3009	create it, and if that fails, use the normal loop for everything.
		3010
		3011	=head3 C<ev_embed> and fork
		3012
		3013	While the C<ev_embed> watcher is running, forks in the embedding loop will
		3014	automatically be applied to the embedded loop as well, so no special
		3015	fork handling is required in that case. When the watcher is not running,
		3016	however, it is still the task of the libev user to call C<ev_loop_fork ()>
		3017	as applicable.
2134		3018
2135	=head3 Watcher-Specific Functions and Data Members	3019	=head3 Watcher-Specific Functions and Data Members
2136		3020
2137	=over 4	3021	=over 4
2138		3022
…		…
2147	if you do not want that, you need to temporarily stop the embed watcher).	3031	if you do not want that, you need to temporarily stop the embed watcher).
2148		3032
2149	=item ev_embed_sweep (loop, ev_embed *)	3033	=item ev_embed_sweep (loop, ev_embed *)
2150		3034
2151	Make a single, non-blocking sweep over the embedded loop. This works	3035	Make a single, non-blocking sweep over the embedded loop. This works
2152	similarly to C<ev_loop (embedded_loop, EVLOOP_NONBLOCK)>, but in the most	3036	similarly to C<ev_run (embedded_loop, EVRUN_NOWAIT)>, but in the most
2153	appropriate way for embedded loops.	3037	appropriate way for embedded loops.
2154		3038
2155	=item struct ev_loop *other [read-only]	3039	=item struct ev_loop *other [read-only]
2156		3040
2157	The embedded event loop.	3041	The embedded event loop.
…		…
2166	C<loop_lo> (which is C<loop_hi> in the case no embeddable loop can be	3050	C<loop_lo> (which is C<loop_hi> in the case no embeddable loop can be
2167	used).	3051	used).
2168		3052
2169	struct ev_loop *loop_hi = ev_default_init (0);	3053	struct ev_loop *loop_hi = ev_default_init (0);
2170	struct ev_loop *loop_lo = 0;	3054	struct ev_loop *loop_lo = 0;
2171	struct ev_embed embed;	3055	ev_embed embed;
2172		3056
2173	// see if there is a chance of getting one that works	3057	// see if there is a chance of getting one that works
2174	// (remember that a flags value of 0 means autodetection)	3058	// (remember that a flags value of 0 means autodetection)
2175	loop_lo = ev_embeddable_backends () & ev_recommended_backends ()	3059	loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
2176	? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())	3060	? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
…		…
2190	kqueue implementation). Store the kqueue/socket-only event loop in	3074	kqueue implementation). Store the kqueue/socket-only event loop in
2191	C<loop_socket>. (One might optionally use C<EVFLAG_NOENV>, too).	3075	C<loop_socket>. (One might optionally use C<EVFLAG_NOENV>, too).
2192		3076
2193	struct ev_loop *loop = ev_default_init (0);	3077	struct ev_loop *loop = ev_default_init (0);
2194	struct ev_loop *loop_socket = 0;	3078	struct ev_loop *loop_socket = 0;
2195	struct ev_embed embed;	3079	ev_embed embed;
2196		3080
2197	if (ev_supported_backends () & ~ev_recommended_backends () & EVBACKEND_KQUEUE)	3081	if (ev_supported_backends () & ~ev_recommended_backends () & EVBACKEND_KQUEUE)
2198	if ((loop_socket = ev_loop_new (EVBACKEND_KQUEUE))	3082	if ((loop_socket = ev_loop_new (EVBACKEND_KQUEUE))
2199	{	3083	{
2200	ev_embed_init (&embed, 0, loop_socket);	3084	ev_embed_init (&embed, 0, loop_socket);
…		…
2215	event loop blocks next and before C<ev_check> watchers are being called,	3099	event loop blocks next and before C<ev_check> watchers are being called,
2216	and only in the child after the fork. If whoever good citizen calling	3100	and only in the child after the fork. If whoever good citizen calling
2217	C<ev_default_fork> cheats and calls it in the wrong process, the fork	3101	C<ev_default_fork> cheats and calls it in the wrong process, the fork
2218	handlers will be invoked, too, of course.	3102	handlers will be invoked, too, of course.
2219		3103
		3104	=head3 The special problem of life after fork - how is it possible?
		3105
		3106	Most uses of C<fork()> consist of forking, then some simple calls to set
		3107	up/change the process environment, followed by a call to C<exec()>. This
		3108	sequence should be handled by libev without any problems.
		3109
		3110	This changes when the application actually wants to do event handling
		3111	in the child, or both parent in child, in effect "continuing" after the
		3112	fork.
		3113
		3114	The default mode of operation (for libev, with application help to detect
		3115	forks) is to duplicate all the state in the child, as would be expected
		3116	when I<either> the parent I<or> the child process continues.
		3117
		3118	When both processes want to continue using libev, then this is usually the
		3119	wrong result. In that case, usually one process (typically the parent) is
		3120	supposed to continue with all watchers in place as before, while the other
		3121	process typically wants to start fresh, i.e. without any active watchers.
		3122
		3123	The cleanest and most efficient way to achieve that with libev is to
		3124	simply create a new event loop, which of course will be "empty", and
		3125	use that for new watchers. This has the advantage of not touching more
		3126	memory than necessary, and thus avoiding the copy-on-write, and the
		3127	disadvantage of having to use multiple event loops (which do not support
		3128	signal watchers).
		3129
		3130	When this is not possible, or you want to use the default loop for
		3131	other reasons, then in the process that wants to start "fresh", call
		3132	C<ev_loop_destroy (EV_DEFAULT)> followed by C<ev_default_loop (...)>.
		3133	Destroying the default loop will "orphan" (not stop) all registered
		3134	watchers, so you have to be careful not to execute code that modifies
		3135	those watchers. Note also that in that case, you have to re-register any
		3136	signal watchers.
		3137
2220	=head3 Watcher-Specific Functions and Data Members	3138	=head3 Watcher-Specific Functions and Data Members
2221		3139
2222	=over 4	3140	=over 4
2223		3141
2224	=item ev_fork_init (ev_signal *, callback)	3142	=item ev_fork_init (ev_fork *, callback)
2225		3143
2226	Initialises and configures the fork watcher - it has no parameters of any	3144	Initialises and configures the fork watcher - it has no parameters of any
2227	kind. There is a C<ev_fork_set> macro, but using it is utterly pointless,	3145	kind. There is a C<ev_fork_set> macro, but using it is utterly pointless,
2228	believe me.	3146	really.
2229		3147
2230	=back	3148	=back
2231		3149
2232		3150
		3151	=head2 C<ev_cleanup> - even the best things end
		3152
		3153	Cleanup watchers are called just before the event loop is being destroyed
		3154	by a call to C<ev_loop_destroy>.
		3155
		3156	While there is no guarantee that the event loop gets destroyed, cleanup
		3157	watchers provide a convenient method to install cleanup hooks for your
		3158	program, worker threads and so on - you just to make sure to destroy the
		3159	loop when you want them to be invoked.
		3160
		3161	Cleanup watchers are invoked in the same way as any other watcher. Unlike
		3162	all other watchers, they do not keep a reference to the event loop (which
		3163	makes a lot of sense if you think about it). Like all other watchers, you
		3164	can call libev functions in the callback, except C<ev_cleanup_start>.
		3165
		3166	=head3 Watcher-Specific Functions and Data Members
		3167
		3168	=over 4
		3169
		3170	=item ev_cleanup_init (ev_cleanup *, callback)
		3171
		3172	Initialises and configures the cleanup watcher - it has no parameters of
		3173	any kind. There is a C<ev_cleanup_set> macro, but using it is utterly
		3174	pointless, I assure you.
		3175
		3176	=back
		3177
		3178	Example: Register an atexit handler to destroy the default loop, so any
		3179	cleanup functions are called.
		3180
		3181	static void
		3182	program_exits (void)
		3183	{
		3184	ev_loop_destroy (EV_DEFAULT_UC);
		3185	}
		3186
		3187	...
		3188	atexit (program_exits);
		3189
		3190
2233	=head2 C<ev_async> - how to wake up another event loop	3191	=head2 C<ev_async> - how to wake up an event loop
2234		3192
2235	In general, you cannot use an C<ev_loop> from multiple threads or other	3193	In general, you cannot use an C<ev_loop> from multiple threads or other
2236	asynchronous sources such as signal handlers (as opposed to multiple event	3194	asynchronous sources such as signal handlers (as opposed to multiple event
2237	loops - those are of course safe to use in different threads).	3195	loops - those are of course safe to use in different threads).
2238		3196
2239	Sometimes, however, you need to wake up another event loop you do not	3197	Sometimes, however, you need to wake up an event loop you do not control,
2240	control, for example because it belongs to another thread. This is what	3198	for example because it belongs to another thread. This is what C<ev_async>
2241	C<ev_async> watchers do: as long as the C<ev_async> watcher is active, you	3199	watchers do: as long as the C<ev_async> watcher is active, you can signal
2242	can signal it by calling C<ev_async_send>, which is thread- and signal	3200	it by calling C<ev_async_send>, which is thread- and signal safe.
2243	safe.
2244		3201
2245	This functionality is very similar to C<ev_signal> watchers, as signals,	3202	This functionality is very similar to C<ev_signal> watchers, as signals,
2246	too, are asynchronous in nature, and signals, too, will be compressed	3203	too, are asynchronous in nature, and signals, too, will be compressed
2247	(i.e. the number of callback invocations may be less than the number of	3204	(i.e. the number of callback invocations may be less than the number of
2248	C<ev_async_sent> calls).	3205	C<ev_async_sent> calls). In fact, you could use signal watchers as a kind
		3206	of "global async watchers" by using a watcher on an otherwise unused
		3207	signal, and C<ev_feed_signal> to signal this watcher from another thread,
		3208	even without knowing which loop owns the signal.
2249		3209
2250	Unlike C<ev_signal> watchers, C<ev_async> works with any event loop, not	3210	Unlike C<ev_signal> watchers, C<ev_async> works with any event loop, not
2251	just the default loop.	3211	just the default loop.
2252		3212
2253	=head3 Queueing	3213	=head3 Queueing
2254		3214
2255	C<ev_async> does not support queueing of data in any way. The reason	3215	C<ev_async> does not support queueing of data in any way. The reason
2256	is that the author does not know of a simple (or any) algorithm for a	3216	is that the author does not know of a simple (or any) algorithm for a
2257	multiple-writer-single-reader queue that works in all cases and doesn't	3217	multiple-writer-single-reader queue that works in all cases and doesn't
2258	need elaborate support such as pthreads.	3218	need elaborate support such as pthreads or unportable memory access
		3219	semantics.
2259		3220
2260	That means that if you want to queue data, you have to provide your own	3221	That means that if you want to queue data, you have to provide your own
2261	queue. But at least I can tell you would implement locking around your	3222	queue. But at least I can tell you how to implement locking around your
2262	queue:	3223	queue:
2263		3224
2264	=over 4	3225	=over 4
2265		3226
2266	=item queueing from a signal handler context	3227	=item queueing from a signal handler context
2267		3228
2268	To implement race-free queueing, you simply add to the queue in the signal	3229	To implement race-free queueing, you simply add to the queue in the signal
2269	handler but you block the signal handler in the watcher callback. Here is an example that does that for	3230	handler but you block the signal handler in the watcher callback. Here is
2270	some fictitious SIGUSR1 handler:	3231	an example that does that for some fictitious SIGUSR1 handler:
2271		3232
2272	static ev_async mysig;	3233	static ev_async mysig;
2273		3234
2274	static void	3235	static void
2275	sigusr1_handler (void)	3236	sigusr1_handler (void)
…		…
2341	=over 4	3302	=over 4
2342		3303
2343	=item ev_async_init (ev_async *, callback)	3304	=item ev_async_init (ev_async *, callback)
2344		3305
2345	Initialises and configures the async watcher - it has no parameters of any	3306	Initialises and configures the async watcher - it has no parameters of any
2346	kind. There is a C<ev_asynd_set> macro, but using it is utterly pointless,	3307	kind. There is a C<ev_async_set> macro, but using it is utterly pointless,
2347	believe me.	3308	trust me.
2348		3309
2349	=item ev_async_send (loop, ev_async *)	3310	=item ev_async_send (loop, ev_async *)
2350		3311
2351	Sends/signals/activates the given C<ev_async> watcher, that is, feeds	3312	Sends/signals/activates the given C<ev_async> watcher, that is, feeds
2352	an C<EV_ASYNC> event on the watcher into the event loop. Unlike	3313	an C<EV_ASYNC> event on the watcher into the event loop, and instantly
		3314	returns.
		3315
2353	C<ev_feed_event>, this call is safe to do in other threads, signal or	3316	Unlike C<ev_feed_event>, this call is safe to do from other threads,
2354	similar contexts (see the discussion of C<EV_ATOMIC_T> in the embedding	3317	signal or similar contexts (see the discussion of C<EV_ATOMIC_T> in the
2355	section below on what exactly this means).	3318	embedding section below on what exactly this means).
2356		3319
		3320	Note that, as with other watchers in libev, multiple events might get
		3321	compressed into a single callback invocation (another way to look at this
		3322	is that C<ev_async> watchers are level-triggered, set on C<ev_async_send>,
		3323	reset when the event loop detects that).
		3324
2357	This call incurs the overhead of a system call only once per loop iteration,	3325	This call incurs the overhead of a system call only once per event loop
2358	so while the overhead might be noticeable, it doesn't apply to repeated	3326	iteration, so while the overhead might be noticeable, it doesn't apply to
2359	calls to C<ev_async_send>.	3327	repeated calls to C<ev_async_send> for the same event loop.
2360		3328
2361	=item bool = ev_async_pending (ev_async *)	3329	=item bool = ev_async_pending (ev_async *)
2362		3330
2363	Returns a non-zero value when C<ev_async_send> has been called on the	3331	Returns a non-zero value when C<ev_async_send> has been called on the
2364	watcher but the event has not yet been processed (or even noted) by the	3332	watcher but the event has not yet been processed (or even noted) by the
…		…
2367	C<ev_async_send> sets a flag in the watcher and wakes up the loop. When	3335	C<ev_async_send> sets a flag in the watcher and wakes up the loop. When
2368	the loop iterates next and checks for the watcher to have become active,	3336	the loop iterates next and checks for the watcher to have become active,
2369	it will reset the flag again. C<ev_async_pending> can be used to very	3337	it will reset the flag again. C<ev_async_pending> can be used to very
2370	quickly check whether invoking the loop might be a good idea.	3338	quickly check whether invoking the loop might be a good idea.
2371		3339
2372	Not that this does I<not> check whether the watcher itself is pending, only	3340	Not that this does I<not> check whether the watcher itself is pending,
2373	whether it has been requested to make this watcher pending.	3341	only whether it has been requested to make this watcher pending: there
		3342	is a time window between the event loop checking and resetting the async
		3343	notification, and the callback being invoked.
2374		3344
2375	=back	3345	=back
2376		3346
2377		3347
2378	=head1 OTHER FUNCTIONS	3348	=head1 OTHER FUNCTIONS
…		…
2382	=over 4	3352	=over 4
2383		3353
2384	=item ev_once (loop, int fd, int events, ev_tstamp timeout, callback)	3354	=item ev_once (loop, int fd, int events, ev_tstamp timeout, callback)
2385		3355
2386	This function combines a simple timer and an I/O watcher, calls your	3356	This function combines a simple timer and an I/O watcher, calls your
2387	callback on whichever event happens first and automatically stop both	3357	callback on whichever event happens first and automatically stops both
2388	watchers. This is useful if you want to wait for a single event on an fd	3358	watchers. This is useful if you want to wait for a single event on an fd
2389	or timeout without having to allocate/configure/start/stop/free one or	3359	or timeout without having to allocate/configure/start/stop/free one or
2390	more watchers yourself.	3360	more watchers yourself.
2391		3361
2392	If C<fd> is less than 0, then no I/O watcher will be started and events	3362	If C<fd> is less than 0, then no I/O watcher will be started and the
2393	is being ignored. Otherwise, an C<ev_io> watcher for the given C<fd> and	3363	C<events> argument is being ignored. Otherwise, an C<ev_io> watcher for
2394	C<events> set will be created and started.	3364	the given C<fd> and C<events> set will be created and started.
2395		3365
2396	If C<timeout> is less than 0, then no timeout watcher will be	3366	If C<timeout> is less than 0, then no timeout watcher will be
2397	started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and	3367	started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and
2398	repeat = 0) will be started. While C<0> is a valid timeout, it is of	3368	repeat = 0) will be started. C<0> is a valid timeout.
2399	dubious value.
2400		3369
2401	The callback has the type C<void (cb)(int revents, void arg)> and gets	3370	The callback has the type C<void (cb)(int revents, void arg)> and is
2402	passed an C<revents> set like normal event callbacks (a combination of	3371	passed an C<revents> set like normal event callbacks (a combination of
2403	C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMEOUT>) and the C<arg>	3372	C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMER>) and the C<arg>
2404	value passed to C<ev_once>:	3373	value passed to C<ev_once>. Note that it is possible to receive I<both>
		3374	a timeout and an io event at the same time - you probably should give io
		3375	events precedence.
		3376
		3377	Example: wait up to ten seconds for data to appear on STDIN_FILENO.
2405		3378
2406	static void stdin_ready (int revents, void *arg)	3379	static void stdin_ready (int revents, void *arg)
2407	{	3380	{
		3381	if (revents & EV_READ)
		3382	/* stdin might have data for us, joy! */;
2408	if (revents & EV_TIMEOUT)	3383	else if (revents & EV_TIMER)
2409	/* doh, nothing entered */;	3384	/* doh, nothing entered */;
2410	else if (revents & EV_READ)
2411	/* stdin might have data for us, joy! */;
2412	}	3385	}
2413		3386
2414	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);	3387	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
2415		3388
2416	=item ev_feed_event (ev_loop , watcher , int revents)
2417
2418	Feeds the given event set into the event loop, as if the specified event
2419	had happened for the specified watcher (which must be a pointer to an
2420	initialised but not necessarily started event watcher).
2421
2422	=item ev_feed_fd_event (ev_loop *, int fd, int revents)	3389	=item ev_feed_fd_event (loop, int fd, int revents)
2423		3390
2424	Feed an event on the given fd, as if a file descriptor backend detected	3391	Feed an event on the given fd, as if a file descriptor backend detected
2425	the given events it.	3392	the given events it.
2426		3393
2427	=item ev_feed_signal_event (ev_loop *loop, int signum)	3394	=item ev_feed_signal_event (loop, int signum)
2428		3395
2429	Feed an event as if the given signal occurred (C<loop> must be the default	3396	Feed an event as if the given signal occurred. See also C<ev_feed_signal>,
2430	loop!).	3397	which is async-safe.
2431		3398
2432	=back	3399	=back
		3400
		3401
		3402	=head1 COMMON OR USEFUL IDIOMS (OR BOTH)
		3403
		3404	This section explains some common idioms that are not immediately
		3405	obvious. Note that examples are sprinkled over the whole manual, and this
		3406	section only contains stuff that wouldn't fit anywhere else.
		3407
		3408	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER
		3409
		3410	Each watcher has, by default, a C<void *data> member that you can read
		3411	or modify at any time: libev will completely ignore it. This can be used
		3412	to associate arbitrary data with your watcher. If you need more data and
		3413	don't want to allocate memory separately and store a pointer to it in that
		3414	data member, you can also "subclass" the watcher type and provide your own
		3415	data:
		3416
		3417	struct my_io
		3418	{
		3419	ev_io io;
		3420	int otherfd;
		3421	void *somedata;
		3422	struct whatever *mostinteresting;
		3423	};
		3424
		3425	...
		3426	struct my_io w;
		3427	ev_io_init (&w.io, my_cb, fd, EV_READ);
		3428
		3429	And since your callback will be called with a pointer to the watcher, you
		3430	can cast it back to your own type:
		3431
		3432	static void my_cb (struct ev_loop loop, ev_io w_, int revents)
		3433	{
		3434	struct my_io w = (struct my_io )w_;
		3435	...
		3436	}
		3437
		3438	More interesting and less C-conformant ways of casting your callback
		3439	function type instead have been omitted.
		3440
		3441	=head2 BUILDING YOUR OWN COMPOSITE WATCHERS
		3442
		3443	Another common scenario is to use some data structure with multiple
		3444	embedded watchers, in effect creating your own watcher that combines
		3445	multiple libev event sources into one "super-watcher":
		3446
		3447	struct my_biggy
		3448	{
		3449	int some_data;
		3450	ev_timer t1;
		3451	ev_timer t2;
		3452	}
		3453
		3454	In this case getting the pointer to C<my_biggy> is a bit more
		3455	complicated: Either you store the address of your C<my_biggy> struct in
		3456	the C<data> member of the watcher (for woozies or C++ coders), or you need
		3457	to use some pointer arithmetic using C<offsetof> inside your watchers (for
		3458	real programmers):
		3459
		3460	#include <stddef.h>
		3461
		3462	static void
		3463	t1_cb (EV_P_ ev_timer *w, int revents)
		3464	{
		3465	struct my_biggy big = (struct my_biggy *)
		3466	(((char *)w) - offsetof (struct my_biggy, t1));
		3467	}
		3468
		3469	static void
		3470	t2_cb (EV_P_ ev_timer *w, int revents)
		3471	{
		3472	struct my_biggy big = (struct my_biggy *)
		3473	(((char *)w) - offsetof (struct my_biggy, t2));
		3474	}
		3475
		3476	=head2 MODEL/NESTED EVENT LOOP INVOCATIONS AND EXIT CONDITIONS
		3477
		3478	Often (especially in GUI toolkits) there are places where you have
		3479	I<modal> interaction, which is most easily implemented by recursively
		3480	invoking C<ev_run>.
		3481
		3482	This brings the problem of exiting - a callback might want to finish the
		3483	main C<ev_run> call, but not the nested one (e.g. user clicked "Quit", but
		3484	a modal "Are you sure?" dialog is still waiting), or just the nested one
		3485	and not the main one (e.g. user clocked "Ok" in a modal dialog), or some
		3486	other combination: In these cases, C<ev_break> will not work alone.
		3487
		3488	The solution is to maintain "break this loop" variable for each C<ev_run>
		3489	invocation, and use a loop around C<ev_run> until the condition is
		3490	triggered, using C<EVRUN_ONCE>:
		3491
		3492	// main loop
		3493	int exit_main_loop = 0;
		3494
		3495	while (!exit_main_loop)
		3496	ev_run (EV_DEFAULT_ EVRUN_ONCE);
		3497
		3498	// in a model watcher
		3499	int exit_nested_loop = 0;
		3500
		3501	while (!exit_nested_loop)
		3502	ev_run (EV_A_ EVRUN_ONCE);
		3503
		3504	To exit from any of these loops, just set the corresponding exit variable:
		3505
		3506	// exit modal loop
		3507	exit_nested_loop = 1;
		3508
		3509	// exit main program, after modal loop is finished
		3510	exit_main_loop = 1;
		3511
		3512	// exit both
		3513	exit_main_loop = exit_nested_loop = 1;
		3514
		3515	=head2 THREAD LOCKING EXAMPLE
		3516
		3517	Here is a fictitious example of how to run an event loop in a different
		3518	thread from where callbacks are being invoked and watchers are
		3519	created/added/removed.
		3520
		3521	For a real-world example, see the C<EV::Loop::Async> perl module,
		3522	which uses exactly this technique (which is suited for many high-level
		3523	languages).
		3524
		3525	The example uses a pthread mutex to protect the loop data, a condition
		3526	variable to wait for callback invocations, an async watcher to notify the
		3527	event loop thread and an unspecified mechanism to wake up the main thread.
		3528
		3529	First, you need to associate some data with the event loop:
		3530
		3531	typedef struct {
		3532	mutex_t lock; /* global loop lock */
		3533	ev_async async_w;
		3534	thread_t tid;
		3535	cond_t invoke_cv;
		3536	} userdata;
		3537
		3538	void prepare_loop (EV_P)
		3539	{
		3540	// for simplicity, we use a static userdata struct.
		3541	static userdata u;
		3542
		3543	ev_async_init (&u->async_w, async_cb);
		3544	ev_async_start (EV_A_ &u->async_w);
		3545
		3546	pthread_mutex_init (&u->lock, 0);
		3547	pthread_cond_init (&u->invoke_cv, 0);
		3548
		3549	// now associate this with the loop
		3550	ev_set_userdata (EV_A_ u);
		3551	ev_set_invoke_pending_cb (EV_A_ l_invoke);
		3552	ev_set_loop_release_cb (EV_A_ l_release, l_acquire);
		3553
		3554	// then create the thread running ev_run
		3555	pthread_create (&u->tid, 0, l_run, EV_A);
		3556	}
		3557
		3558	The callback for the C<ev_async> watcher does nothing: the watcher is used
		3559	solely to wake up the event loop so it takes notice of any new watchers
		3560	that might have been added:
		3561
		3562	static void
		3563	async_cb (EV_P_ ev_async *w, int revents)
		3564	{
		3565	// just used for the side effects
		3566	}
		3567
		3568	The C<l_release> and C<l_acquire> callbacks simply unlock/lock the mutex
		3569	protecting the loop data, respectively.
		3570
		3571	static void
		3572	l_release (EV_P)
		3573	{
		3574	userdata *u = ev_userdata (EV_A);
		3575	pthread_mutex_unlock (&u->lock);
		3576	}
		3577
		3578	static void
		3579	l_acquire (EV_P)
		3580	{
		3581	userdata *u = ev_userdata (EV_A);
		3582	pthread_mutex_lock (&u->lock);
		3583	}
		3584
		3585	The event loop thread first acquires the mutex, and then jumps straight
		3586	into C<ev_run>:
		3587
		3588	void *
		3589	l_run (void *thr_arg)
		3590	{
		3591	struct ev_loop loop = (struct ev_loop )thr_arg;
		3592
		3593	l_acquire (EV_A);
		3594	pthread_setcanceltype (PTHREAD_CANCEL_ASYNCHRONOUS, 0);
		3595	ev_run (EV_A_ 0);
		3596	l_release (EV_A);
		3597
		3598	return 0;
		3599	}
		3600
		3601	Instead of invoking all pending watchers, the C<l_invoke> callback will
		3602	signal the main thread via some unspecified mechanism (signals? pipe
		3603	writes? C<Async::Interrupt>?) and then waits until all pending watchers
		3604	have been called (in a while loop because a) spurious wakeups are possible
		3605	and b) skipping inter-thread-communication when there are no pending
		3606	watchers is very beneficial):
		3607
		3608	static void
		3609	l_invoke (EV_P)
		3610	{
		3611	userdata *u = ev_userdata (EV_A);
		3612
		3613	while (ev_pending_count (EV_A))
		3614	{
		3615	wake_up_other_thread_in_some_magic_or_not_so_magic_way ();
		3616	pthread_cond_wait (&u->invoke_cv, &u->lock);
		3617	}
		3618	}
		3619
		3620	Now, whenever the main thread gets told to invoke pending watchers, it
		3621	will grab the lock, call C<ev_invoke_pending> and then signal the loop
		3622	thread to continue:
		3623
		3624	static void
		3625	real_invoke_pending (EV_P)
		3626	{
		3627	userdata *u = ev_userdata (EV_A);
		3628
		3629	pthread_mutex_lock (&u->lock);
		3630	ev_invoke_pending (EV_A);
		3631	pthread_cond_signal (&u->invoke_cv);
		3632	pthread_mutex_unlock (&u->lock);
		3633	}
		3634
		3635	Whenever you want to start/stop a watcher or do other modifications to an
		3636	event loop, you will now have to lock:
		3637
		3638	ev_timer timeout_watcher;
		3639	userdata *u = ev_userdata (EV_A);
		3640
		3641	ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);
		3642
		3643	pthread_mutex_lock (&u->lock);
		3644	ev_timer_start (EV_A_ &timeout_watcher);
		3645	ev_async_send (EV_A_ &u->async_w);
		3646	pthread_mutex_unlock (&u->lock);
		3647
		3648	Note that sending the C<ev_async> watcher is required because otherwise
		3649	an event loop currently blocking in the kernel will have no knowledge
		3650	about the newly added timer. By waking up the loop it will pick up any new
		3651	watchers in the next event loop iteration.
		3652
		3653	=head2 THREADS, COROUTINES, CONTINUATIONS, QUEUES... INSTEAD OF CALLBACKS
		3654
		3655	While the overhead of a callback that e.g. schedules a thread is small, it
		3656	is still an overhead. If you embed libev, and your main usage is with some
		3657	kind of threads or coroutines, you might want to customise libev so that
		3658	doesn't need callbacks anymore.
		3659
		3660	Imagine you have coroutines that you can switch to using a function
		3661	C<switch_to (coro)>, that libev runs in a coroutine called C<libev_coro>
		3662	and that due to some magic, the currently active coroutine is stored in a
		3663	global called C<current_coro>. Then you can build your own "wait for libev
		3664	event" primitive by changing C<EV_CB_DECLARE> and C<EV_CB_INVOKE> (note
		3665	the differing C<;> conventions):
		3666
		3667	#define EV_CB_DECLARE(type) struct my_coro *cb;
		3668	#define EV_CB_INVOKE(watcher) switch_to ((watcher)->cb)
		3669
		3670	That means instead of having a C callback function, you store the
		3671	coroutine to switch to in each watcher, and instead of having libev call
		3672	your callback, you instead have it switch to that coroutine.
		3673
		3674	A coroutine might now wait for an event with a function called
		3675	C<wait_for_event>. (the watcher needs to be started, as always, but it doesn't
		3676	matter when, or whether the watcher is active or not when this function is
		3677	called):
		3678
		3679	void
		3680	wait_for_event (ev_watcher *w)
		3681	{
		3682	ev_cb_set (w) = current_coro;
		3683	switch_to (libev_coro);
		3684	}
		3685
		3686	That basically suspends the coroutine inside C<wait_for_event> and
		3687	continues the libev coroutine, which, when appropriate, switches back to
		3688	this or any other coroutine. I am sure if you sue this your own :)
		3689
		3690	You can do similar tricks if you have, say, threads with an event queue -
		3691	instead of storing a coroutine, you store the queue object and instead of
		3692	switching to a coroutine, you push the watcher onto the queue and notify
		3693	any waiters.
		3694
		3695	To embed libev, see L<EMBEDDING>, but in short, it's easiest to create two
		3696	files, F<my_ev.h> and F<my_ev.c> that include the respective libev files:
		3697
		3698	// my_ev.h
		3699	#define EV_CB_DECLARE(type) struct my_coro *cb;
		3700	#define EV_CB_INVOKE(watcher) switch_to ((watcher)->cb);
		3701	#include "../libev/ev.h"
		3702
		3703	// my_ev.c
		3704	#define EV_H "my_ev.h"
		3705	#include "../libev/ev.c"
		3706
		3707	And then use F<my_ev.h> when you would normally use F<ev.h>, and compile
		3708	F<my_ev.c> into your project. When properly specifying include paths, you
		3709	can even use F<ev.h> as header file name directly.
2433		3710
2434		3711
2435	=head1 LIBEVENT EMULATION	3712	=head1 LIBEVENT EMULATION
2436		3713
2437	Libev offers a compatibility emulation layer for libevent. It cannot	3714	Libev offers a compatibility emulation layer for libevent. It cannot
2438	emulate the internals of libevent, so here are some usage hints:	3715	emulate the internals of libevent, so here are some usage hints:
2439		3716
2440	=over 4	3717	=over 4
		3718
		3719	=item * Only the libevent-1.4.1-beta API is being emulated.
		3720
		3721	This was the newest libevent version available when libev was implemented,
		3722	and is still mostly unchanged in 2010.
2441		3723
2442	=item * Use it by including <event.h>, as usual.	3724	=item * Use it by including <event.h>, as usual.
2443		3725
2444	=item * The following members are fully supported: ev_base, ev_callback,	3726	=item * The following members are fully supported: ev_base, ev_callback,
2445	ev_arg, ev_fd, ev_res, ev_events.	3727	ev_arg, ev_fd, ev_res, ev_events.
…		…
2451	=item * Priorities are not currently supported. Initialising priorities	3733	=item * Priorities are not currently supported. Initialising priorities
2452	will fail and all watchers will have the same priority, even though there	3734	will fail and all watchers will have the same priority, even though there
2453	is an ev_pri field.	3735	is an ev_pri field.
2454		3736
2455	=item * In libevent, the last base created gets the signals, in libev, the	3737	=item * In libevent, the last base created gets the signals, in libev, the
2456	first base created (== the default loop) gets the signals.	3738	base that registered the signal gets the signals.
2457		3739
2458	=item * Other members are not supported.	3740	=item * Other members are not supported.
2459		3741
2460	=item * The libev emulation is I<not> ABI compatible to libevent, you need	3742	=item * The libev emulation is I<not> ABI compatible to libevent, you need
2461	to use the libev header file and library.	3743	to use the libev header file and library.
…		…
2480	Care has been taken to keep the overhead low. The only data member the C++	3762	Care has been taken to keep the overhead low. The only data member the C++
2481	classes add (compared to plain C-style watchers) is the event loop pointer	3763	classes add (compared to plain C-style watchers) is the event loop pointer
2482	that the watcher is associated with (or no additional members at all if	3764	that the watcher is associated with (or no additional members at all if
2483	you disable C<EV_MULTIPLICITY> when embedding libev).	3765	you disable C<EV_MULTIPLICITY> when embedding libev).
2484		3766
2485	Currently, functions, and static and non-static member functions can be	3767	Currently, functions, static and non-static member functions and classes
2486	used as callbacks. Other types should be easy to add as long as they only	3768	with C<operator ()> can be used as callbacks. Other types should be easy
2487	need one additional pointer for context. If you need support for other	3769	to add as long as they only need one additional pointer for context. If
2488	types of functors please contact the author (preferably after implementing	3770	you need support for other types of functors please contact the author
2489	it).	3771	(preferably after implementing it).
2490		3772
2491	Here is a list of things available in the C<ev> namespace:	3773	Here is a list of things available in the C<ev> namespace:
2492		3774
2493	=over 4	3775	=over 4
2494		3776
…		…
2512		3794
2513	=over 4	3795	=over 4
2514		3796
2515	=item ev::TYPE::TYPE ()	3797	=item ev::TYPE::TYPE ()
2516		3798
2517	=item ev::TYPE::TYPE (struct ev_loop *)	3799	=item ev::TYPE::TYPE (loop)
2518		3800
2519	=item ev::TYPE::~TYPE	3801	=item ev::TYPE::~TYPE
2520		3802
2521	The constructor (optionally) takes an event loop to associate the watcher	3803	The constructor (optionally) takes an event loop to associate the watcher
2522	with. If it is omitted, it will use C<EV_DEFAULT>.	3804	with. If it is omitted, it will use C<EV_DEFAULT>.
…		…
2554		3836
2555	myclass obj;	3837	myclass obj;
2556	ev::io iow;	3838	ev::io iow;
2557	iow.set <myclass, &myclass::io_cb> (&obj);	3839	iow.set <myclass, &myclass::io_cb> (&obj);
2558		3840
		3841	=item w->set (object *)
		3842
		3843	This is a variation of a method callback - leaving out the method to call
		3844	will default the method to C<operator ()>, which makes it possible to use
		3845	functor objects without having to manually specify the C<operator ()> all
		3846	the time. Incidentally, you can then also leave out the template argument
		3847	list.
		3848
		3849	The C<operator ()> method prototype must be C<void operator ()(watcher &w,
		3850	int revents)>.
		3851
		3852	See the method-C<set> above for more details.
		3853
		3854	Example: use a functor object as callback.
		3855
		3856	struct myfunctor
		3857	{
		3858	void operator() (ev::io &w, int revents)
		3859	{
		3860	...
		3861	}
		3862	}
		3863
		3864	myfunctor f;
		3865
		3866	ev::io w;
		3867	w.set (&f);
		3868
2559	=item w->set<function> (void *data = 0)	3869	=item w->set<function> (void *data = 0)
2560		3870
2561	Also sets a callback, but uses a static method or plain function as	3871	Also sets a callback, but uses a static method or plain function as
2562	callback. The optional C<data> argument will be stored in the watcher's	3872	callback. The optional C<data> argument will be stored in the watcher's
2563	C<data> member and is free for you to use.	3873	C<data> member and is free for you to use.
2564		3874
2565	The prototype of the C<function> must be C<void (*)(ev::TYPE &w, int)>.	3875	The prototype of the C<function> must be C<void (*)(ev::TYPE &w, int)>.
2566		3876
2567	See the method-C<set> above for more details.	3877	See the method-C<set> above for more details.
2568		3878
2569	Example:	3879	Example: Use a plain function as callback.
2570		3880
2571	static void io_cb (ev::io &w, int revents) { }	3881	static void io_cb (ev::io &w, int revents) { }
2572	iow.set <io_cb> ();	3882	iow.set <io_cb> ();
2573		3883
2574	=item w->set (struct ev_loop *)	3884	=item w->set (loop)
2575		3885
2576	Associates a different C<struct ev_loop> with this watcher. You can only	3886	Associates a different C<struct ev_loop> with this watcher. You can only
2577	do this when the watcher is inactive (and not pending either).	3887	do this when the watcher is inactive (and not pending either).
2578		3888
2579	=item w->set ([arguments])	3889	=item w->set ([arguments])
2580		3890
2581	Basically the same as C<ev_TYPE_set>, with the same arguments. Must be	3891	Basically the same as C<ev_TYPE_set>, with the same arguments. Either this
2582	called at least once. Unlike the C counterpart, an active watcher gets	3892	method or a suitable start method must be called at least once. Unlike the
2583	automatically stopped and restarted when reconfiguring it with this	3893	C counterpart, an active watcher gets automatically stopped and restarted
2584	method.	3894	when reconfiguring it with this method.
2585		3895
2586	=item w->start ()	3896	=item w->start ()
2587		3897
2588	Starts the watcher. Note that there is no C<loop> argument, as the	3898	Starts the watcher. Note that there is no C<loop> argument, as the
2589	constructor already stores the event loop.	3899	constructor already stores the event loop.
2590		3900
		3901	=item w->start ([arguments])
		3902
		3903	Instead of calling C<set> and C<start> methods separately, it is often
		3904	convenient to wrap them in one call. Uses the same type of arguments as
		3905	the configure C<set> method of the watcher.
		3906
2591	=item w->stop ()	3907	=item w->stop ()
2592		3908
2593	Stops the watcher if it is active. Again, no C<loop> argument.	3909	Stops the watcher if it is active. Again, no C<loop> argument.
2594		3910
2595	=item w->again () (C<ev::timer>, C<ev::periodic> only)	3911	=item w->again () (C<ev::timer>, C<ev::periodic> only)
…		…
2607		3923
2608	=back	3924	=back
2609		3925
2610	=back	3926	=back
2611		3927
2612	Example: Define a class with an IO and idle watcher, start one of them in	3928	Example: Define a class with two I/O and idle watchers, start the I/O
2613	the constructor.	3929	watchers in the constructor.
2614		3930
2615	class myclass	3931	class myclass
2616	{	3932	{
2617	ev::io io; void io_cb (ev::io &w, int revents);	3933	ev::io io ; void io_cb (ev::io &w, int revents);
		3934	ev::io2 io2 ; void io2_cb (ev::io &w, int revents);
2618	ev:idle idle void idle_cb (ev::idle &w, int revents);	3935	ev::idle idle; void idle_cb (ev::idle &w, int revents);
2619		3936
2620	myclass (int fd)	3937	myclass (int fd)
2621	{	3938	{
2622	io .set <myclass, &myclass::io_cb > (this);	3939	io .set <myclass, &myclass::io_cb > (this);
		3940	io2 .set <myclass, &myclass::io2_cb > (this);
2623	idle.set <myclass, &myclass::idle_cb> (this);	3941	idle.set <myclass, &myclass::idle_cb> (this);
2624		3942
2625	io.start (fd, ev::READ);	3943	io.set (fd, ev::WRITE); // configure the watcher
		3944	io.start (); // start it whenever convenient
		3945
		3946	io2.start (fd, ev::READ); // set + start in one call
2626	}	3947	}
2627	};	3948	};
2628		3949
2629		3950
2630	=head1 OTHER LANGUAGE BINDINGS	3951	=head1 OTHER LANGUAGE BINDINGS
…		…
2639	=item Perl	3960	=item Perl
2640		3961
2641	The EV module implements the full libev API and is actually used to test	3962	The EV module implements the full libev API and is actually used to test
2642	libev. EV is developed together with libev. Apart from the EV core module,	3963	libev. EV is developed together with libev. Apart from the EV core module,
2643	there are additional modules that implement libev-compatible interfaces	3964	there are additional modules that implement libev-compatible interfaces
2644	to C<libadns> (C<EV::ADNS>), C<Net::SNMP> (C<Net::SNMP::EV>) and the	3965	to C<libadns> (C<EV::ADNS>, but C<AnyEvent::DNS> is preferred nowadays),
2645	C<libglib> event core (C<Glib::EV> and C<EV::Glib>).	3966	C<Net::SNMP> (C<Net::SNMP::EV>) and the C<libglib> event core (C<Glib::EV>
		3967	and C<EV::Glib>).
2646		3968
2647	It can be found and installed via CPAN, its homepage is at	3969	It can be found and installed via CPAN, its homepage is at
2648	L<http://software.schmorp.de/pkg/EV>.	3970	L<http://software.schmorp.de/pkg/EV>.
2649		3971
2650	=item Python	3972	=item Python
2651		3973
2652	Python bindings can be found at L<http://code.google.com/p/pyev/>. It	3974	Python bindings can be found at L<http://code.google.com/p/pyev/>. It
2653	seems to be quite complete and well-documented. Note, however, that the	3975	seems to be quite complete and well-documented.
2654	patch they require for libev is outright dangerous as it breaks the ABI
2655	for everybody else, and therefore, should never be applied in an installed
2656	libev (if python requires an incompatible ABI then it needs to embed
2657	libev).
2658		3976
2659	=item Ruby	3977	=item Ruby
2660		3978
2661	Tony Arcieri has written a ruby extension that offers access to a subset	3979	Tony Arcieri has written a ruby extension that offers access to a subset
2662	of the libev API and adds file handle abstractions, asynchronous DNS and	3980	of the libev API and adds file handle abstractions, asynchronous DNS and
2663	more on top of it. It can be found via gem servers. Its homepage is at	3981	more on top of it. It can be found via gem servers. Its homepage is at
2664	L<http://rev.rubyforge.org/>.	3982	L<http://rev.rubyforge.org/>.
2665		3983
		3984	Roger Pack reports that using the link order C<-lws2_32 -lmsvcrt-ruby-190>
		3985	makes rev work even on mingw.
		3986
		3987	=item Haskell
		3988
		3989	A haskell binding to libev is available at
		3990	L<http://hackage.haskell.org/cgi-bin/hackage-scripts/package/hlibev>.
		3991
2666	=item D	3992	=item D
2667		3993
2668	Leandro Lucarella has written a D language binding (F<ev.d>) for libev, to	3994	Leandro Lucarella has written a D language binding (F<ev.d>) for libev, to
2669	be found at L<http://proj.llucax.com.ar/wiki/evd>.	3995	be found at L<http://proj.llucax.com.ar/wiki/evd>.
		3996
		3997	=item Ocaml
		3998
		3999	Erkki Seppala has written Ocaml bindings for libev, to be found at
		4000	L<http://modeemi.cs.tut.fi/~flux/software/ocaml-ev/>.
		4001
		4002	=item Lua
		4003
		4004	Brian Maher has written a partial interface to libev for lua (at the
		4005	time of this writing, only C<ev_io> and C<ev_timer>), to be found at
		4006	L<http://github.com/brimworks/lua-ev>.
2670		4007
2671	=back	4008	=back
2672		4009
2673		4010
2674	=head1 MACRO MAGIC	4011	=head1 MACRO MAGIC
…		…
2688	loop argument"). The C<EV_A> form is used when this is the sole argument,	4025	loop argument"). The C<EV_A> form is used when this is the sole argument,
2689	C<EV_A_> is used when other arguments are following. Example:	4026	C<EV_A_> is used when other arguments are following. Example:
2690		4027
2691	ev_unref (EV_A);	4028	ev_unref (EV_A);
2692	ev_timer_add (EV_A_ watcher);	4029	ev_timer_add (EV_A_ watcher);
2693	ev_loop (EV_A_ 0);	4030	ev_run (EV_A_ 0);
2694		4031
2695	It assumes the variable C<loop> of type C<struct ev_loop *> is in scope,	4032	It assumes the variable C<loop> of type C<struct ev_loop *> is in scope,
2696	which is often provided by the following macro.	4033	which is often provided by the following macro.
2697		4034
2698	=item C<EV_P>, C<EV_P_>	4035	=item C<EV_P>, C<EV_P_>
…		…
2738	}	4075	}
2739		4076
2740	ev_check check;	4077	ev_check check;
2741	ev_check_init (&check, check_cb);	4078	ev_check_init (&check, check_cb);
2742	ev_check_start (EV_DEFAULT_ &check);	4079	ev_check_start (EV_DEFAULT_ &check);
2743	ev_loop (EV_DEFAULT_ 0);	4080	ev_run (EV_DEFAULT_ 0);
2744		4081
2745	=head1 EMBEDDING	4082	=head1 EMBEDDING
2746		4083
2747	Libev can (and often is) directly embedded into host	4084	Libev can (and often is) directly embedded into host
2748	applications. Examples of applications that embed it include the Deliantra	4085	applications. Examples of applications that embed it include the Deliantra
…		…
2775		4112
2776	#define EV_STANDALONE 1	4113	#define EV_STANDALONE 1
2777	#include "ev.h"	4114	#include "ev.h"
2778		4115
2779	Both header files and implementation files can be compiled with a C++	4116	Both header files and implementation files can be compiled with a C++
2780	compiler (at least, thats a stated goal, and breakage will be treated	4117	compiler (at least, that's a stated goal, and breakage will be treated
2781	as a bug).	4118	as a bug).
2782		4119
2783	You need the following files in your source tree, or in a directory	4120	You need the following files in your source tree, or in a directory
2784	in your include path (e.g. in libev/ when using -Ilibev):	4121	in your include path (e.g. in libev/ when using -Ilibev):
2785		4122
…		…
2828	libev.m4	4165	libev.m4
2829		4166
2830	=head2 PREPROCESSOR SYMBOLS/MACROS	4167	=head2 PREPROCESSOR SYMBOLS/MACROS
2831		4168
2832	Libev can be configured via a variety of preprocessor symbols you have to	4169	Libev can be configured via a variety of preprocessor symbols you have to
2833	define before including any of its files. The default in the absence of	4170	define before including (or compiling) any of its files. The default in
2834	autoconf is noted for every option.	4171	the absence of autoconf is documented for every option.
		4172
		4173	Symbols marked with "(h)" do not change the ABI, and can have different
		4174	values when compiling libev vs. including F<ev.h>, so it is permissible
		4175	to redefine them before including F<ev.h> without breaking compatibility
		4176	to a compiled library. All other symbols change the ABI, which means all
		4177	users of libev and the libev code itself must be compiled with compatible
		4178	settings.
2835		4179
2836	=over 4	4180	=over 4
2837		4181
		4182	=item EV_COMPAT3 (h)
		4183
		4184	Backwards compatibility is a major concern for libev. This is why this
		4185	release of libev comes with wrappers for the functions and symbols that
		4186	have been renamed between libev version 3 and 4.
		4187
		4188	You can disable these wrappers (to test compatibility with future
		4189	versions) by defining C<EV_COMPAT3> to C<0> when compiling your
		4190	sources. This has the additional advantage that you can drop the C<struct>
		4191	from C<struct ev_loop> declarations, as libev will provide an C<ev_loop>
		4192	typedef in that case.
		4193
		4194	In some future version, the default for C<EV_COMPAT3> will become C<0>,
		4195	and in some even more future version the compatibility code will be
		4196	removed completely.
		4197
2838	=item EV_STANDALONE	4198	=item EV_STANDALONE (h)
2839		4199
2840	Must always be C<1> if you do not use autoconf configuration, which	4200	Must always be C<1> if you do not use autoconf configuration, which
2841	keeps libev from including F<config.h>, and it also defines dummy	4201	keeps libev from including F<config.h>, and it also defines dummy
2842	implementations for some libevent functions (such as logging, which is not	4202	implementations for some libevent functions (such as logging, which is not
2843	supported). It will also not define any of the structs usually found in	4203	supported). It will also not define any of the structs usually found in
2844	F<event.h> that are not directly supported by the libev core alone.	4204	F<event.h> that are not directly supported by the libev core alone.
2845		4205
		4206	In standalone mode, libev will still try to automatically deduce the
		4207	configuration, but has to be more conservative.
		4208
2846	=item EV_USE_MONOTONIC	4209	=item EV_USE_MONOTONIC
2847		4210
2848	If defined to be C<1>, libev will try to detect the availability of the	4211	If defined to be C<1>, libev will try to detect the availability of the
2849	monotonic clock option at both compile time and runtime. Otherwise no use	4212	monotonic clock option at both compile time and runtime. Otherwise no
2850	of the monotonic clock option will be attempted. If you enable this, you	4213	use of the monotonic clock option will be attempted. If you enable this,
2851	usually have to link against librt or something similar. Enabling it when	4214	you usually have to link against librt or something similar. Enabling it
2852	the functionality isn't available is safe, though, although you have	4215	when the functionality isn't available is safe, though, although you have
2853	to make sure you link against any libraries where the C<clock_gettime>	4216	to make sure you link against any libraries where the C<clock_gettime>
2854	function is hiding in (often F<-lrt>).	4217	function is hiding in (often F<-lrt>). See also C<EV_USE_CLOCK_SYSCALL>.
2855		4218
2856	=item EV_USE_REALTIME	4219	=item EV_USE_REALTIME
2857		4220
2858	If defined to be C<1>, libev will try to detect the availability of the	4221	If defined to be C<1>, libev will try to detect the availability of the
2859	real-time clock option at compile time (and assume its availability at	4222	real-time clock option at compile time (and assume its availability
2860	runtime if successful). Otherwise no use of the real-time clock option will	4223	at runtime if successful). Otherwise no use of the real-time clock
2861	be attempted. This effectively replaces C<gettimeofday> by C<clock_get	4224	option will be attempted. This effectively replaces C<gettimeofday>
2862	(CLOCK_REALTIME, ...)> and will not normally affect correctness. See the	4225	by C<clock_get (CLOCK_REALTIME, ...)> and will not normally affect
2863	note about libraries in the description of C<EV_USE_MONOTONIC>, though.	4226	correctness. See the note about libraries in the description of
		4227	C<EV_USE_MONOTONIC>, though. Defaults to the opposite value of
		4228	C<EV_USE_CLOCK_SYSCALL>.
		4229
		4230	=item EV_USE_CLOCK_SYSCALL
		4231
		4232	If defined to be C<1>, libev will try to use a direct syscall instead
		4233	of calling the system-provided C<clock_gettime> function. This option
		4234	exists because on GNU/Linux, C<clock_gettime> is in C<librt>, but C<librt>
		4235	unconditionally pulls in C<libpthread>, slowing down single-threaded
		4236	programs needlessly. Using a direct syscall is slightly slower (in
		4237	theory), because no optimised vdso implementation can be used, but avoids
		4238	the pthread dependency. Defaults to C<1> on GNU/Linux with glibc 2.x or
		4239	higher, as it simplifies linking (no need for C<-lrt>).
2864		4240
2865	=item EV_USE_NANOSLEEP	4241	=item EV_USE_NANOSLEEP
2866		4242
2867	If defined to be C<1>, libev will assume that C<nanosleep ()> is available	4243	If defined to be C<1>, libev will assume that C<nanosleep ()> is available
2868	and will use it for delays. Otherwise it will use C<select ()>.	4244	and will use it for delays. Otherwise it will use C<select ()>.
…		…
2884		4260
2885	=item EV_SELECT_USE_FD_SET	4261	=item EV_SELECT_USE_FD_SET
2886		4262
2887	If defined to C<1>, then the select backend will use the system C<fd_set>	4263	If defined to C<1>, then the select backend will use the system C<fd_set>
2888	structure. This is useful if libev doesn't compile due to a missing	4264	structure. This is useful if libev doesn't compile due to a missing
2889	C<NFDBITS> or C<fd_mask> definition or it mis-guesses the bitset layout on	4265	C<NFDBITS> or C<fd_mask> definition or it mis-guesses the bitset layout
2890	exotic systems. This usually limits the range of file descriptors to some	4266	on exotic systems. This usually limits the range of file descriptors to
2891	low limit such as 1024 or might have other limitations (winsocket only	4267	some low limit such as 1024 or might have other limitations (winsocket
2892	allows 64 sockets). The C<FD_SETSIZE> macro, set before compilation, might	4268	only allows 64 sockets). The C<FD_SETSIZE> macro, set before compilation,
2893	influence the size of the C<fd_set> used.	4269	configures the maximum size of the C<fd_set>.
2894		4270
2895	=item EV_SELECT_IS_WINSOCKET	4271	=item EV_SELECT_IS_WINSOCKET
2896		4272
2897	When defined to C<1>, the select backend will assume that	4273	When defined to C<1>, the select backend will assume that
2898	select/socket/connect etc. don't understand file descriptors but	4274	select/socket/connect etc. don't understand file descriptors but
…		…
2900	be used is the winsock select). This means that it will call	4276	be used is the winsock select). This means that it will call
2901	C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,	4277	C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,
2902	it is assumed that all these functions actually work on fds, even	4278	it is assumed that all these functions actually work on fds, even
2903	on win32. Should not be defined on non-win32 platforms.	4279	on win32. Should not be defined on non-win32 platforms.
2904		4280
2905	=item EV_FD_TO_WIN32_HANDLE	4281	=item EV_FD_TO_WIN32_HANDLE(fd)
2906		4282
2907	If C<EV_SELECT_IS_WINSOCKET> is enabled, then libev needs a way to map	4283	If C<EV_SELECT_IS_WINSOCKET> is enabled, then libev needs a way to map
2908	file descriptors to socket handles. When not defining this symbol (the	4284	file descriptors to socket handles. When not defining this symbol (the
2909	default), then libev will call C<_get_osfhandle>, which is usually	4285	default), then libev will call C<_get_osfhandle>, which is usually
2910	correct. In some cases, programs use their own file descriptor management,	4286	correct. In some cases, programs use their own file descriptor management,
2911	in which case they can provide this function to map fds to socket handles.	4287	in which case they can provide this function to map fds to socket handles.
		4288
		4289	=item EV_WIN32_HANDLE_TO_FD(handle)
		4290
		4291	If C<EV_SELECT_IS_WINSOCKET> then libev maps handles to file descriptors
		4292	using the standard C<_open_osfhandle> function. For programs implementing
		4293	their own fd to handle mapping, overwriting this function makes it easier
		4294	to do so. This can be done by defining this macro to an appropriate value.
		4295
		4296	=item EV_WIN32_CLOSE_FD(fd)
		4297
		4298	If programs implement their own fd to handle mapping on win32, then this
		4299	macro can be used to override the C<close> function, useful to unregister
		4300	file descriptors again. Note that the replacement function has to close
		4301	the underlying OS handle.
2912		4302
2913	=item EV_USE_POLL	4303	=item EV_USE_POLL
2914		4304
2915	If defined to be C<1>, libev will compile in support for the C<poll>(2)	4305	If defined to be C<1>, libev will compile in support for the C<poll>(2)
2916	backend. Otherwise it will be enabled on non-win32 platforms. It	4306	backend. Otherwise it will be enabled on non-win32 platforms. It
…		…
2963	as well as for signal and thread safety in C<ev_async> watchers.	4353	as well as for signal and thread safety in C<ev_async> watchers.
2964		4354
2965	In the absence of this define, libev will use C<sig_atomic_t volatile>	4355	In the absence of this define, libev will use C<sig_atomic_t volatile>
2966	(from F<signal.h>), which is usually good enough on most platforms.	4356	(from F<signal.h>), which is usually good enough on most platforms.
2967		4357
2968	=item EV_H	4358	=item EV_H (h)
2969		4359
2970	The name of the F<ev.h> header file used to include it. The default if	4360	The name of the F<ev.h> header file used to include it. The default if
2971	undefined is C<"ev.h"> in F<event.h>, F<ev.c> and F<ev++.h>. This can be	4361	undefined is C<"ev.h"> in F<event.h>, F<ev.c> and F<ev++.h>. This can be
2972	used to virtually rename the F<ev.h> header file in case of conflicts.	4362	used to virtually rename the F<ev.h> header file in case of conflicts.
2973		4363
2974	=item EV_CONFIG_H	4364	=item EV_CONFIG_H (h)
2975		4365
2976	If C<EV_STANDALONE> isn't C<1>, this variable can be used to override	4366	If C<EV_STANDALONE> isn't C<1>, this variable can be used to override
2977	F<ev.c>'s idea of where to find the F<config.h> file, similarly to	4367	F<ev.c>'s idea of where to find the F<config.h> file, similarly to
2978	C<EV_H>, above.	4368	C<EV_H>, above.
2979		4369
2980	=item EV_EVENT_H	4370	=item EV_EVENT_H (h)
2981		4371
2982	Similarly to C<EV_H>, this macro can be used to override F<event.c>'s idea	4372	Similarly to C<EV_H>, this macro can be used to override F<event.c>'s idea
2983	of how the F<event.h> header can be found, the default is C<"event.h">.	4373	of how the F<event.h> header can be found, the default is C<"event.h">.
2984		4374
2985	=item EV_PROTOTYPES	4375	=item EV_PROTOTYPES (h)
2986		4376
2987	If defined to be C<0>, then F<ev.h> will not define any function	4377	If defined to be C<0>, then F<ev.h> will not define any function
2988	prototypes, but still define all the structs and other symbols. This is	4378	prototypes, but still define all the structs and other symbols. This is
2989	occasionally useful if you want to provide your own wrapper functions	4379	occasionally useful if you want to provide your own wrapper functions
2990	around libev functions.	4380	around libev functions.
…		…
3009	When doing priority-based operations, libev usually has to linearly search	4399	When doing priority-based operations, libev usually has to linearly search
3010	all the priorities, so having many of them (hundreds) uses a lot of space	4400	all the priorities, so having many of them (hundreds) uses a lot of space
3011	and time, so using the defaults of five priorities (-2 .. +2) is usually	4401	and time, so using the defaults of five priorities (-2 .. +2) is usually
3012	fine.	4402	fine.
3013		4403
3014	If your embedding application does not need any priorities, defining these both to	4404	If your embedding application does not need any priorities, defining these
3015	C<0> will save some memory and CPU.	4405	both to C<0> will save some memory and CPU.
3016		4406
3017	=item EV_PERIODIC_ENABLE	4407	=item EV_PERIODIC_ENABLE, EV_IDLE_ENABLE, EV_EMBED_ENABLE, EV_STAT_ENABLE,
		4408	EV_PREPARE_ENABLE, EV_CHECK_ENABLE, EV_FORK_ENABLE, EV_SIGNAL_ENABLE,
		4409	EV_ASYNC_ENABLE, EV_CHILD_ENABLE.
3018		4410
3019	If undefined or defined to be C<1>, then periodic timers are supported. If	4411	If undefined or defined to be C<1> (and the platform supports it), then
3020	defined to be C<0>, then they are not. Disabling them saves a few kB of	4412	the respective watcher type is supported. If defined to be C<0>, then it
3021	code.	4413	is not. Disabling watcher types mainly saves code size.
3022		4414
3023	=item EV_IDLE_ENABLE	4415	=item EV_FEATURES
3024
3025	If undefined or defined to be C<1>, then idle watchers are supported. If
3026	defined to be C<0>, then they are not. Disabling them saves a few kB of
3027	code.
3028
3029	=item EV_EMBED_ENABLE
3030
3031	If undefined or defined to be C<1>, then embed watchers are supported. If
3032	defined to be C<0>, then they are not.
3033
3034	=item EV_STAT_ENABLE
3035
3036	If undefined or defined to be C<1>, then stat watchers are supported. If
3037	defined to be C<0>, then they are not.
3038
3039	=item EV_FORK_ENABLE
3040
3041	If undefined or defined to be C<1>, then fork watchers are supported. If
3042	defined to be C<0>, then they are not.
3043
3044	=item EV_ASYNC_ENABLE
3045
3046	If undefined or defined to be C<1>, then async watchers are supported. If
3047	defined to be C<0>, then they are not.
3048
3049	=item EV_MINIMAL
3050		4416
3051	If you need to shave off some kilobytes of code at the expense of some	4417	If you need to shave off some kilobytes of code at the expense of some
3052	speed, define this symbol to C<1>. Currently this is used to override some	4418	speed (but with the full API), you can define this symbol to request
3053	inlining decisions, saves roughly 30% code size on amd64. It also selects a	4419	certain subsets of functionality. The default is to enable all features
3054	much smaller 2-heap for timer management over the default 4-heap.	4420	that can be enabled on the platform.
		4421
		4422	A typical way to use this symbol is to define it to C<0> (or to a bitset
		4423	with some broad features you want) and then selectively re-enable
		4424	additional parts you want, for example if you want everything minimal,
		4425	but multiple event loop support, async and child watchers and the poll
		4426	backend, use this:
		4427
		4428	#define EV_FEATURES 0
		4429	#define EV_MULTIPLICITY 1
		4430	#define EV_USE_POLL 1
		4431	#define EV_CHILD_ENABLE 1
		4432	#define EV_ASYNC_ENABLE 1
		4433
		4434	The actual value is a bitset, it can be a combination of the following
		4435	values:
		4436
		4437	=over 4
		4438
		4439	=item C<1> - faster/larger code
		4440
		4441	Use larger code to speed up some operations.
		4442
		4443	Currently this is used to override some inlining decisions (enlarging the
		4444	code size by roughly 30% on amd64).
		4445
		4446	When optimising for size, use of compiler flags such as C<-Os> with
		4447	gcc is recommended, as well as C<-DNDEBUG>, as libev contains a number of
		4448	assertions.
		4449
		4450	=item C<2> - faster/larger data structures
		4451
		4452	Replaces the small 2-heap for timer management by a faster 4-heap, larger
		4453	hash table sizes and so on. This will usually further increase code size
		4454	and can additionally have an effect on the size of data structures at
		4455	runtime.
		4456
		4457	=item C<4> - full API configuration
		4458
		4459	This enables priorities (sets C<EV_MAXPRI>=2 and C<EV_MINPRI>=-2), and
		4460	enables multiplicity (C<EV_MULTIPLICITY>=1).
		4461
		4462	=item C<8> - full API
		4463
		4464	This enables a lot of the "lesser used" API functions. See C<ev.h> for
		4465	details on which parts of the API are still available without this
		4466	feature, and do not complain if this subset changes over time.
		4467
		4468	=item C<16> - enable all optional watcher types
		4469
		4470	Enables all optional watcher types. If you want to selectively enable
		4471	only some watcher types other than I/O and timers (e.g. prepare,
		4472	embed, async, child...) you can enable them manually by defining
		4473	C<EV_watchertype_ENABLE> to C<1> instead.
		4474
		4475	=item C<32> - enable all backends
		4476
		4477	This enables all backends - without this feature, you need to enable at
		4478	least one backend manually (C<EV_USE_SELECT> is a good choice).
		4479
		4480	=item C<64> - enable OS-specific "helper" APIs
		4481
		4482	Enable inotify, eventfd, signalfd and similar OS-specific helper APIs by
		4483	default.
		4484
		4485	=back
		4486
		4487	Compiling with C<gcc -Os -DEV_STANDALONE -DEV_USE_EPOLL=1 -DEV_FEATURES=0>
		4488	reduces the compiled size of libev from 24.7Kb code/2.8Kb data to 6.5Kb
		4489	code/0.3Kb data on my GNU/Linux amd64 system, while still giving you I/O
		4490	watchers, timers and monotonic clock support.
		4491
		4492	With an intelligent-enough linker (gcc+binutils are intelligent enough
		4493	when you use C<-Wl,--gc-sections -ffunction-sections>) functions unused by
		4494	your program might be left out as well - a binary starting a timer and an
		4495	I/O watcher then might come out at only 5Kb.
		4496
		4497	=item EV_AVOID_STDIO
		4498
		4499	If this is set to C<1> at compiletime, then libev will avoid using stdio
		4500	functions (printf, scanf, perror etc.). This will increase the code size
		4501	somewhat, but if your program doesn't otherwise depend on stdio and your
		4502	libc allows it, this avoids linking in the stdio library which is quite
		4503	big.
		4504
		4505	Note that error messages might become less precise when this option is
		4506	enabled.
		4507
		4508	=item EV_NSIG
		4509
		4510	The highest supported signal number, +1 (or, the number of
		4511	signals): Normally, libev tries to deduce the maximum number of signals
		4512	automatically, but sometimes this fails, in which case it can be
		4513	specified. Also, using a lower number than detected (C<32> should be
		4514	good for about any system in existence) can save some memory, as libev
		4515	statically allocates some 12-24 bytes per signal number.
3055		4516
3056	=item EV_PID_HASHSIZE	4517	=item EV_PID_HASHSIZE
3057		4518
3058	C<ev_child> watchers use a small hash table to distribute workload by	4519	C<ev_child> watchers use a small hash table to distribute workload by
3059	pid. The default size is C<16> (or C<1> with C<EV_MINIMAL>), usually more	4520	pid. The default size is C<16> (or C<1> with C<EV_FEATURES> disabled),
3060	than enough. If you need to manage thousands of children you might want to	4521	usually more than enough. If you need to manage thousands of children you
3061	increase this value (I<must> be a power of two).	4522	might want to increase this value (I<must> be a power of two).
3062		4523
3063	=item EV_INOTIFY_HASHSIZE	4524	=item EV_INOTIFY_HASHSIZE
3064		4525
3065	C<ev_stat> watchers use a small hash table to distribute workload by	4526	C<ev_stat> watchers use a small hash table to distribute workload by
3066	inotify watch id. The default size is C<16> (or C<1> with C<EV_MINIMAL>),	4527	inotify watch id. The default size is C<16> (or C<1> with C<EV_FEATURES>
3067	usually more than enough. If you need to manage thousands of C<ev_stat>	4528	disabled), usually more than enough. If you need to manage thousands of
3068	watchers you might want to increase this value (I<must> be a power of	4529	C<ev_stat> watchers you might want to increase this value (I<must> be a
3069	two).	4530	power of two).
3070		4531
3071	=item EV_USE_4HEAP	4532	=item EV_USE_4HEAP
3072		4533
3073	Heaps are not very cache-efficient. To improve the cache-efficiency of the	4534	Heaps are not very cache-efficient. To improve the cache-efficiency of the
3074	timer and periodics heap, libev uses a 4-heap when this symbol is defined	4535	timer and periodics heaps, libev uses a 4-heap when this symbol is defined
3075	to C<1>. The 4-heap uses more complicated (longer) code but has	4536	to C<1>. The 4-heap uses more complicated (longer) code but has noticeably
3076	noticeably faster performance with many (thousands) of watchers.	4537	faster performance with many (thousands) of watchers.
3077		4538
3078	The default is C<1> unless C<EV_MINIMAL> is set in which case it is C<0>	4539	The default is C<1>, unless C<EV_FEATURES> overrides it, in which case it
3079	(disabled).	4540	will be C<0>.
3080		4541
3081	=item EV_HEAP_CACHE_AT	4542	=item EV_HEAP_CACHE_AT
3082		4543
3083	Heaps are not very cache-efficient. To improve the cache-efficiency of the	4544	Heaps are not very cache-efficient. To improve the cache-efficiency of the
3084	timer and periodics heap, libev can cache the timestamp (I<at>) within	4545	timer and periodics heaps, libev can cache the timestamp (I<at>) within
3085	the heap structure (selected by defining C<EV_HEAP_CACHE_AT> to C<1>),	4546	the heap structure (selected by defining C<EV_HEAP_CACHE_AT> to C<1>),
3086	which uses 8-12 bytes more per watcher and a few hundred bytes more code,	4547	which uses 8-12 bytes more per watcher and a few hundred bytes more code,
3087	but avoids random read accesses on heap changes. This improves performance	4548	but avoids random read accesses on heap changes. This improves performance
3088	noticeably with with many (hundreds) of watchers.	4549	noticeably with many (hundreds) of watchers.
3089		4550
3090	The default is C<1> unless C<EV_MINIMAL> is set in which case it is C<0>	4551	The default is C<1>, unless C<EV_FEATURES> overrides it, in which case it
3091	(disabled).	4552	will be C<0>.
3092		4553
3093	=item EV_VERIFY	4554	=item EV_VERIFY
3094		4555
3095	Controls how much internal verification (see C<ev_loop_verify ()>) will	4556	Controls how much internal verification (see C<ev_verify ()>) will
3096	be done: If set to C<0>, no internal verification code will be compiled	4557	be done: If set to C<0>, no internal verification code will be compiled
3097	in. If set to C<1>, then verification code will be compiled in, but not	4558	in. If set to C<1>, then verification code will be compiled in, but not
3098	called. If set to C<2>, then the internal verification code will be	4559	called. If set to C<2>, then the internal verification code will be
3099	called once per loop, which can slow down libev. If set to C<3>, then the	4560	called once per loop, which can slow down libev. If set to C<3>, then the
3100	verification code will be called very frequently, which will slow down	4561	verification code will be called very frequently, which will slow down
3101	libev considerably.	4562	libev considerably.
3102		4563
3103	The default is C<1>, unless C<EV_MINIMAL> is set, in which case it will be	4564	The default is C<1>, unless C<EV_FEATURES> overrides it, in which case it
3104	C<0.>	4565	will be C<0>.
3105		4566
3106	=item EV_COMMON	4567	=item EV_COMMON
3107		4568
3108	By default, all watchers have a C<void *data> member. By redefining	4569	By default, all watchers have a C<void *data> member. By redefining
3109	this macro to a something else you can include more and other types of	4570	this macro to something else you can include more and other types of
3110	members. You have to define it each time you include one of the files,	4571	members. You have to define it each time you include one of the files,
3111	though, and it must be identical each time.	4572	though, and it must be identical each time.
3112		4573
3113	For example, the perl EV module uses something like this:	4574	For example, the perl EV module uses something like this:
3114		4575
…		…
3126	and the way callbacks are invoked and set. Must expand to a struct member	4587	and the way callbacks are invoked and set. Must expand to a struct member
3127	definition and a statement, respectively. See the F<ev.h> header file for	4588	definition and a statement, respectively. See the F<ev.h> header file for
3128	their default definitions. One possible use for overriding these is to	4589	their default definitions. One possible use for overriding these is to
3129	avoid the C<struct ev_loop *> as first argument in all cases, or to use	4590	avoid the C<struct ev_loop *> as first argument in all cases, or to use
3130	method calls instead of plain function calls in C++.	4591	method calls instead of plain function calls in C++.
		4592
		4593	=back
3131		4594
3132	=head2 EXPORTED API SYMBOLS	4595	=head2 EXPORTED API SYMBOLS
3133		4596
3134	If you need to re-export the API (e.g. via a DLL) and you need a list of	4597	If you need to re-export the API (e.g. via a DLL) and you need a list of
3135	exported symbols, you can use the provided F<Symbol.*> files which list	4598	exported symbols, you can use the provided F<Symbol.*> files which list
…		…
3165	file.	4628	file.
3166		4629
3167	The usage in rxvt-unicode is simpler. It has a F<ev_cpp.h> header file	4630	The usage in rxvt-unicode is simpler. It has a F<ev_cpp.h> header file
3168	that everybody includes and which overrides some configure choices:	4631	that everybody includes and which overrides some configure choices:
3169		4632
3170	#define EV_MINIMAL 1	4633	#define EV_FEATURES 8
3171	#define EV_USE_POLL 0	4634	#define EV_USE_SELECT 1
3172	#define EV_MULTIPLICITY 0
3173	#define EV_PERIODIC_ENABLE 0	4635	#define EV_PREPARE_ENABLE 1
		4636	#define EV_IDLE_ENABLE 1
3174	#define EV_STAT_ENABLE 0	4637	#define EV_SIGNAL_ENABLE 1
3175	#define EV_FORK_ENABLE 0	4638	#define EV_CHILD_ENABLE 1
		4639	#define EV_USE_STDEXCEPT 0
3176	#define EV_CONFIG_H <config.h>	4640	#define EV_CONFIG_H <config.h>
3177	#define EV_MINPRI 0
3178	#define EV_MAXPRI 0
3179		4641
3180	#include "ev++.h"	4642	#include "ev++.h"
3181		4643
3182	And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:	4644	And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:
3183		4645
3184	#include "ev_cpp.h"	4646	#include "ev_cpp.h"
3185	#include "ev.c"	4647	#include "ev.c"
3186		4648
		4649	=head1 INTERACTION WITH OTHER PROGRAMS, LIBRARIES OR THE ENVIRONMENT
3187		4650
3188	=head1 THREADS AND COROUTINES	4651	=head2 THREADS AND COROUTINES
3189		4652
3190	=head2 THREADS	4653	=head3 THREADS
3191		4654
3192	Libev itself is completely thread-safe, but it uses no locking. This	4655	All libev functions are reentrant and thread-safe unless explicitly
		4656	documented otherwise, but libev implements no locking itself. This means
3193	means that you can use as many loops as you want in parallel, as long as	4657	that you can use as many loops as you want in parallel, as long as there
3194	only one thread ever calls into one libev function with the same loop	4658	are no concurrent calls into any libev function with the same loop
3195	parameter.	4659	parameter (C<ev_default_*> calls have an implicit default loop parameter,
		4660	of course): libev guarantees that different event loops share no data
		4661	structures that need any locking.
3196		4662
3197	Or put differently: calls with different loop parameters can be done in	4663	Or to put it differently: calls with different loop parameters can be done
3198	parallel from multiple threads, calls with the same loop parameter must be	4664	concurrently from multiple threads, calls with the same loop parameter
3199	done serially (but can be done from different threads, as long as only one	4665	must be done serially (but can be done from different threads, as long as
3200	thread ever is inside a call at any point in time, e.g. by using a mutex	4666	only one thread ever is inside a call at any point in time, e.g. by using
3201	per loop).	4667	a mutex per loop).
		4668
		4669	Specifically to support threads (and signal handlers), libev implements
		4670	so-called C<ev_async> watchers, which allow some limited form of
		4671	concurrency on the same event loop, namely waking it up "from the
		4672	outside".
3202		4673
3203	If you want to know which design (one loop, locking, or multiple loops	4674	If you want to know which design (one loop, locking, or multiple loops
3204	without or something else still) is best for your problem, then I cannot	4675	without or something else still) is best for your problem, then I cannot
3205	help you. I can give some generic advice however:	4676	help you, but here is some generic advice:
3206		4677
3207	=over 4	4678	=over 4
3208		4679
3209	=item * most applications have a main thread: use the default libev loop	4680	=item * most applications have a main thread: use the default libev loop
3210	in that thread, or create a separate thread running only the default loop.	4681	in that thread, or create a separate thread running only the default loop.
…		…
3222		4693
3223	Choosing a model is hard - look around, learn, know that usually you can do	4694	Choosing a model is hard - look around, learn, know that usually you can do
3224	better than you currently do :-)	4695	better than you currently do :-)
3225		4696
3226	=item * often you need to talk to some other thread which blocks in the	4697	=item * often you need to talk to some other thread which blocks in the
		4698	event loop.
		4699
3227	event loop - C<ev_async> watchers can be used to wake them up from other	4700	C<ev_async> watchers can be used to wake them up from other threads safely
3228	threads safely (or from signal contexts...).	4701	(or from signal contexts...).
		4702
		4703	An example use would be to communicate signals or other events that only
		4704	work in the default loop by registering the signal watcher with the
		4705	default loop and triggering an C<ev_async> watcher from the default loop
		4706	watcher callback into the event loop interested in the signal.
3229		4707
3230	=back	4708	=back
3231		4709
		4710	See also L<THREAD LOCKING EXAMPLE>.
		4711
3232	=head2 COROUTINES	4712	=head3 COROUTINES
3233		4713
3234	Libev is much more accommodating to coroutines ("cooperative threads"):	4714	Libev is very accommodating to coroutines ("cooperative threads"):
3235	libev fully supports nesting calls to it's functions from different	4715	libev fully supports nesting calls to its functions from different
3236	coroutines (e.g. you can call C<ev_loop> on the same loop from two	4716	coroutines (e.g. you can call C<ev_run> on the same loop from two
3237	different coroutines and switch freely between both coroutines running the	4717	different coroutines, and switch freely between both coroutines running
3238	loop, as long as you don't confuse yourself). The only exception is that	4718	the loop, as long as you don't confuse yourself). The only exception is
3239	you must not do this from C<ev_periodic> reschedule callbacks.	4719	that you must not do this from C<ev_periodic> reschedule callbacks.
3240		4720
3241	Care has been invested into making sure that libev does not keep local	4721	Care has been taken to ensure that libev does not keep local state inside
3242	state inside C<ev_loop>, and other calls do not usually allow coroutine	4722	C<ev_run>, and other calls do not usually allow for coroutine switches as
3243	switches.	4723	they do not call any callbacks.
3244		4724
		4725	=head2 COMPILER WARNINGS
3245		4726
3246	=head1 COMPLEXITIES	4727	Depending on your compiler and compiler settings, you might get no or a
		4728	lot of warnings when compiling libev code. Some people are apparently
		4729	scared by this.
3247		4730
3248	In this section the complexities of (many of) the algorithms used inside	4731	However, these are unavoidable for many reasons. For one, each compiler
3249	libev will be explained. For complexity discussions about backends see the	4732	has different warnings, and each user has different tastes regarding
3250	documentation for C<ev_default_init>.	4733	warning options. "Warn-free" code therefore cannot be a goal except when
		4734	targeting a specific compiler and compiler-version.
3251		4735
3252	All of the following are about amortised time: If an array needs to be	4736	Another reason is that some compiler warnings require elaborate
3253	extended, libev needs to realloc and move the whole array, but this	4737	workarounds, or other changes to the code that make it less clear and less
3254	happens asymptotically never with higher number of elements, so O(1) might	4738	maintainable.
3255	mean it might do a lengthy realloc operation in rare cases, but on average
3256	it is much faster and asymptotically approaches constant time.
3257		4739
3258	=over 4	4740	And of course, some compiler warnings are just plain stupid, or simply
		4741	wrong (because they don't actually warn about the condition their message
		4742	seems to warn about). For example, certain older gcc versions had some
		4743	warnings that resulted in an extreme number of false positives. These have
		4744	been fixed, but some people still insist on making code warn-free with
		4745	such buggy versions.
3259		4746
3260	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)	4747	While libev is written to generate as few warnings as possible,
		4748	"warn-free" code is not a goal, and it is recommended not to build libev
		4749	with any compiler warnings enabled unless you are prepared to cope with
		4750	them (e.g. by ignoring them). Remember that warnings are just that:
		4751	warnings, not errors, or proof of bugs.
3261		4752
3262	This means that, when you have a watcher that triggers in one hour and
3263	there are 100 watchers that would trigger before that then inserting will
3264	have to skip roughly seven (C<ld 100>) of these watchers.
3265		4753
3266	=item Changing timer/periodic watchers (by autorepeat or calling again): O(log skipped_other_timers)	4754	=head2 VALGRIND
3267		4755
3268	That means that changing a timer costs less than removing/adding them	4756	Valgrind has a special section here because it is a popular tool that is
3269	as only the relative motion in the event queue has to be paid for.	4757	highly useful. Unfortunately, valgrind reports are very hard to interpret.
3270		4758
3271	=item Starting io/check/prepare/idle/signal/child/fork/async watchers: O(1)	4759	If you think you found a bug (memory leak, uninitialised data access etc.)
		4760	in libev, then check twice: If valgrind reports something like:
3272		4761
3273	These just add the watcher into an array or at the head of a list.	4762	==2274== definitely lost: 0 bytes in 0 blocks.
		4763	==2274== possibly lost: 0 bytes in 0 blocks.
		4764	==2274== still reachable: 256 bytes in 1 blocks.
3274		4765
3275	=item Stopping check/prepare/idle/fork/async watchers: O(1)	4766	Then there is no memory leak, just as memory accounted to global variables
		4767	is not a memleak - the memory is still being referenced, and didn't leak.
3276		4768
3277	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))	4769	Similarly, under some circumstances, valgrind might report kernel bugs
		4770	as if it were a bug in libev (e.g. in realloc or in the poll backend,
		4771	although an acceptable workaround has been found here), or it might be
		4772	confused.
3278		4773
3279	These watchers are stored in lists then need to be walked to find the	4774	Keep in mind that valgrind is a very good tool, but only a tool. Don't
3280	correct watcher to remove. The lists are usually short (you don't usually	4775	make it into some kind of religion.
3281	have many watchers waiting for the same fd or signal).
3282		4776
3283	=item Finding the next timer in each loop iteration: O(1)	4777	If you are unsure about something, feel free to contact the mailing list
		4778	with the full valgrind report and an explanation on why you think this
		4779	is a bug in libev (best check the archives, too :). However, don't be
		4780	annoyed when you get a brisk "this is no bug" answer and take the chance
		4781	of learning how to interpret valgrind properly.
3284		4782
3285	By virtue of using a binary or 4-heap, the next timer is always found at a	4783	If you need, for some reason, empty reports from valgrind for your project
3286	fixed position in the storage array.	4784	I suggest using suppression lists.
3287		4785
3288	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)
3289		4786
3290	A change means an I/O watcher gets started or stopped, which requires	4787	=head1 PORTABILITY NOTES
3291	libev to recalculate its status (and possibly tell the kernel, depending
3292	on backend and whether C<ev_io_set> was used).
3293		4788
3294	=item Activating one watcher (putting it into the pending state): O(1)	4789	=head2 GNU/LINUX 32 BIT LIMITATIONS
3295		4790
3296	=item Priority handling: O(number_of_priorities)	4791	GNU/Linux is the only common platform that supports 64 bit file/large file
		4792	interfaces but I<disables> them by default.
3297		4793
3298	Priorities are implemented by allocating some space for each	4794	That means that libev compiled in the default environment doesn't support
3299	priority. When doing priority-based operations, libev usually has to	4795	files larger than 2GiB or so, which mainly affects C<ev_stat> watchers.
3300	linearly search all the priorities, but starting/stopping and activating
3301	watchers becomes O(1) w.r.t. priority handling.
3302		4796
3303	=item Sending an ev_async: O(1)	4797	Unfortunately, many programs try to work around this GNU/Linux issue
		4798	by enabling the large file API, which makes them incompatible with the
		4799	standard libev compiled for their system.
3304		4800
3305	=item Processing ev_async_send: O(number_of_async_watchers)	4801	Likewise, libev cannot enable the large file API itself as this would
		4802	suddenly make it incompatible to the default compile time environment,
		4803	i.e. all programs not using special compile switches.
3306		4804
3307	=item Processing signals: O(max_signal_number)	4805	=head2 OS/X AND DARWIN BUGS
3308		4806
3309	Sending involves a system call I<iff> there were no other C<ev_async_send>	4807	The whole thing is a bug if you ask me - basically any system interface
3310	calls in the current loop iteration. Checking for async and signal events	4808	you touch is broken, whether it is locales, poll, kqueue or even the
3311	involves iterating over all running async watchers or all signal numbers.	4809	OpenGL drivers.
3312		4810
3313	=back	4811	=head3 C<kqueue> is buggy
3314		4812
		4813	The kqueue syscall is broken in all known versions - most versions support
		4814	only sockets, many support pipes.
3315		4815
		4816	Libev tries to work around this by not using C<kqueue> by default on this
		4817	rotten platform, but of course you can still ask for it when creating a
		4818	loop - embedding a socket-only kqueue loop into a select-based one is
		4819	probably going to work well.
		4820
		4821	=head3 C<poll> is buggy
		4822
		4823	Instead of fixing C<kqueue>, Apple replaced their (working) C<poll>
		4824	implementation by something calling C<kqueue> internally around the 10.5.6
		4825	release, so now C<kqueue> I<and> C<poll> are broken.
		4826
		4827	Libev tries to work around this by not using C<poll> by default on
		4828	this rotten platform, but of course you can still ask for it when creating
		4829	a loop.
		4830
		4831	=head3 C<select> is buggy
		4832
		4833	All that's left is C<select>, and of course Apple found a way to fuck this
		4834	one up as well: On OS/X, C<select> actively limits the number of file
		4835	descriptors you can pass in to 1024 - your program suddenly crashes when
		4836	you use more.
		4837
		4838	There is an undocumented "workaround" for this - defining
		4839	C<_DARWIN_UNLIMITED_SELECT>, which libev tries to use, so select I<should>
		4840	work on OS/X.
		4841
		4842	=head2 SOLARIS PROBLEMS AND WORKAROUNDS
		4843
		4844	=head3 C<errno> reentrancy
		4845
		4846	The default compile environment on Solaris is unfortunately so
		4847	thread-unsafe that you can't even use components/libraries compiled
		4848	without C<-D_REENTRANT> in a threaded program, which, of course, isn't
		4849	defined by default. A valid, if stupid, implementation choice.
		4850
		4851	If you want to use libev in threaded environments you have to make sure
		4852	it's compiled with C<_REENTRANT> defined.
		4853
		4854	=head3 Event port backend
		4855
		4856	The scalable event interface for Solaris is called "event
		4857	ports". Unfortunately, this mechanism is very buggy in all major
		4858	releases. If you run into high CPU usage, your program freezes or you get
		4859	a large number of spurious wakeups, make sure you have all the relevant
		4860	and latest kernel patches applied. No, I don't know which ones, but there
		4861	are multiple ones to apply, and afterwards, event ports actually work
		4862	great.
		4863
		4864	If you can't get it to work, you can try running the program by setting
		4865	the environment variable C<LIBEV_FLAGS=3> to only allow C<poll> and
		4866	C<select> backends.
		4867
		4868	=head2 AIX POLL BUG
		4869
		4870	AIX unfortunately has a broken C<poll.h> header. Libev works around
		4871	this by trying to avoid the poll backend altogether (i.e. it's not even
		4872	compiled in), which normally isn't a big problem as C<select> works fine
		4873	with large bitsets on AIX, and AIX is dead anyway.
		4874
3316	=head1 WIN32 PLATFORM LIMITATIONS AND WORKAROUNDS	4875	=head2 WIN32 PLATFORM LIMITATIONS AND WORKAROUNDS
		4876
		4877	=head3 General issues
3317		4878
3318	Win32 doesn't support any of the standards (e.g. POSIX) that libev	4879	Win32 doesn't support any of the standards (e.g. POSIX) that libev
3319	requires, and its I/O model is fundamentally incompatible with the POSIX	4880	requires, and its I/O model is fundamentally incompatible with the POSIX
3320	model. Libev still offers limited functionality on this platform in	4881	model. Libev still offers limited functionality on this platform in
3321	the form of the C<EVBACKEND_SELECT> backend, and only supports socket	4882	the form of the C<EVBACKEND_SELECT> backend, and only supports socket
3322	descriptors. This only applies when using Win32 natively, not when using	4883	descriptors. This only applies when using Win32 natively, not when using
3323	e.g. cygwin.	4884	e.g. cygwin. Actually, it only applies to the microsofts own compilers,
		4885	as every compielr comes with a slightly differently broken/incompatible
		4886	environment.
3324		4887
3325	Lifting these limitations would basically require the full	4888	Lifting these limitations would basically require the full
3326	re-implementation of the I/O system. If you are into these kinds of	4889	re-implementation of the I/O system. If you are into this kind of thing,
3327	things, then note that glib does exactly that for you in a very portable	4890	then note that glib does exactly that for you in a very portable way (note
3328	way (note also that glib is the slowest event library known to man).	4891	also that glib is the slowest event library known to man).
3329		4892
3330	There is no supported compilation method available on windows except	4893	There is no supported compilation method available on windows except
3331	embedding it into other applications.	4894	embedding it into other applications.
		4895
		4896	Sensible signal handling is officially unsupported by Microsoft - libev
		4897	tries its best, but under most conditions, signals will simply not work.
3332		4898
3333	Not a libev limitation but worth mentioning: windows apparently doesn't	4899	Not a libev limitation but worth mentioning: windows apparently doesn't
3334	accept large writes: instead of resulting in a partial write, windows will	4900	accept large writes: instead of resulting in a partial write, windows will
3335	either accept everything or return C<ENOBUFS> if the buffer is too large,	4901	either accept everything or return C<ENOBUFS> if the buffer is too large,
3336	so make sure you only write small amounts into your sockets (less than a	4902	so make sure you only write small amounts into your sockets (less than a
3337	megabyte seems safe, but thsi apparently depends on the amount of memory	4903	megabyte seems safe, but this apparently depends on the amount of memory
3338	available).	4904	available).
3339		4905
3340	Due to the many, low, and arbitrary limits on the win32 platform and	4906	Due to the many, low, and arbitrary limits on the win32 platform and
3341	the abysmal performance of winsockets, using a large number of sockets	4907	the abysmal performance of winsockets, using a large number of sockets
3342	is not recommended (and not reasonable). If your program needs to use	4908	is not recommended (and not reasonable). If your program needs to use
3343	more than a hundred or so sockets, then likely it needs to use a totally	4909	more than a hundred or so sockets, then likely it needs to use a totally
3344	different implementation for windows, as libev offers the POSIX readiness	4910	different implementation for windows, as libev offers the POSIX readiness
3345	notification model, which cannot be implemented efficiently on windows	4911	notification model, which cannot be implemented efficiently on windows
3346	(Microsoft monopoly games).	4912	(due to Microsoft monopoly games).
3347		4913
3348	A typical way to use libev under windows is to embed it (see the embedding	4914	A typical way to use libev under windows is to embed it (see the embedding
3349	section for details) and use the following F<evwrap.h> header file instead	4915	section for details) and use the following F<evwrap.h> header file instead
3350	of F<ev.h>:	4916	of F<ev.h>:
3351		4917
…		…
3353	#define EV_SELECT_IS_WINSOCKET 1 /* configure libev for windows select */	4919	#define EV_SELECT_IS_WINSOCKET 1 /* configure libev for windows select */
3354		4920
3355	#include "ev.h"	4921	#include "ev.h"
3356		4922
3357	And compile the following F<evwrap.c> file into your project (make sure	4923	And compile the following F<evwrap.c> file into your project (make sure
3358	you do I<not> compile the F<ev.c> or any other embedded soruce files!):	4924	you do I<not> compile the F<ev.c> or any other embedded source files!):
3359		4925
3360	#include "evwrap.h"	4926	#include "evwrap.h"
3361	#include "ev.c"	4927	#include "ev.c"
3362		4928
3363	=over 4
3364
3365	=item The winsocket select function	4929	=head3 The winsocket C<select> function
3366		4930
3367	The winsocket C<select> function doesn't follow POSIX in that it	4931	The winsocket C<select> function doesn't follow POSIX in that it
3368	requires socket I<handles> and not socket I<file descriptors> (it is	4932	requires socket I<handles> and not socket I<file descriptors> (it is
3369	also extremely buggy). This makes select very inefficient, and also	4933	also extremely buggy). This makes select very inefficient, and also
3370	requires a mapping from file descriptors to socket handles (the Microsoft	4934	requires a mapping from file descriptors to socket handles (the Microsoft
…		…
3379	#define EV_SELECT_IS_WINSOCKET 1 /* forces EV_SELECT_USE_FD_SET, too */	4943	#define EV_SELECT_IS_WINSOCKET 1 /* forces EV_SELECT_USE_FD_SET, too */
3380		4944
3381	Note that winsockets handling of fd sets is O(n), so you can easily get a	4945	Note that winsockets handling of fd sets is O(n), so you can easily get a
3382	complexity in the O(n²) range when using win32.	4946	complexity in the O(n²) range when using win32.
3383		4947
3384	=item Limited number of file descriptors	4948	=head3 Limited number of file descriptors
3385		4949
3386	Windows has numerous arbitrary (and low) limits on things.	4950	Windows has numerous arbitrary (and low) limits on things.
3387		4951
3388	Early versions of winsocket's select only supported waiting for a maximum	4952	Early versions of winsocket's select only supported waiting for a maximum
3389	of C<64> handles (probably owning to the fact that all windows kernels	4953	of C<64> handles (probably owning to the fact that all windows kernels
3390	can only wait for C<64> things at the same time internally; Microsoft	4954	can only wait for C<64> things at the same time internally; Microsoft
3391	recommends spawning a chain of threads and wait for 63 handles and the	4955	recommends spawning a chain of threads and wait for 63 handles and the
3392	previous thread in each. Great).	4956	previous thread in each. Sounds great!).
3393		4957
3394	Newer versions support more handles, but you need to define C<FD_SETSIZE>	4958	Newer versions support more handles, but you need to define C<FD_SETSIZE>
3395	to some high number (e.g. C<2048>) before compiling the winsocket select	4959	to some high number (e.g. C<2048>) before compiling the winsocket select
3396	call (which might be in libev or elsewhere, for example, perl does its own	4960	call (which might be in libev or elsewhere, for example, perl and many
3397	select emulation on windows).	4961	other interpreters do their own select emulation on windows).
3398		4962
3399	Another limit is the number of file descriptors in the Microsoft runtime	4963	Another limit is the number of file descriptors in the Microsoft runtime
3400	libraries, which by default is C<64> (there must be a hidden I<64> fetish	4964	libraries, which by default is C<64> (there must be a hidden I<64>
3401	or something like this inside Microsoft). You can increase this by calling	4965	fetish or something like this inside Microsoft). You can increase this
3402	C<_setmaxstdio>, which can increase this limit to C<2048> (another	4966	by calling C<_setmaxstdio>, which can increase this limit to C<2048>
3403	arbitrary limit), but is broken in many versions of the Microsoft runtime	4967	(another arbitrary limit), but is broken in many versions of the Microsoft
3404	libraries.
3405
3406	This might get you to about C<512> or C<2048> sockets (depending on	4968	runtime libraries. This might get you to about C<512> or C<2048> sockets
3407	windows version and/or the phase of the moon). To get more, you need to	4969	(depending on windows version and/or the phase of the moon). To get more,
3408	wrap all I/O functions and provide your own fd management, but the cost of	4970	you need to wrap all I/O functions and provide your own fd management, but
3409	calling select (O(n²)) will likely make this unworkable.	4971	the cost of calling select (O(n²)) will likely make this unworkable.
3410		4972
3411	=back
3412
3413
3414	=head1 PORTABILITY REQUIREMENTS	4973	=head2 PORTABILITY REQUIREMENTS
3415		4974
3416	In addition to a working ISO-C implementation, libev relies on a few	4975	In addition to a working ISO-C implementation and of course the
3417	additional extensions:	4976	backend-specific APIs, libev relies on a few additional extensions:
3418		4977
3419	=over 4	4978	=over 4
3420		4979
3421	=item C<void ()(ev_watcher_type , int revents)> must have compatible	4980	=item C<void ()(ev_watcher_type , int revents)> must have compatible
3422	calling conventions regardless of C<ev_watcher_type *>.	4981	calling conventions regardless of C<ev_watcher_type *>.
…		…
3425	structure (guaranteed by POSIX but not by ISO C for example), but it also	4984	structure (guaranteed by POSIX but not by ISO C for example), but it also
3426	assumes that the same (machine) code can be used to call any watcher	4985	assumes that the same (machine) code can be used to call any watcher
3427	callback: The watcher callbacks have different type signatures, but libev	4986	callback: The watcher callbacks have different type signatures, but libev
3428	calls them using an C<ev_watcher *> internally.	4987	calls them using an C<ev_watcher *> internally.
3429		4988
		4989	=item pointer accesses must be thread-atomic
		4990
		4991	Accessing a pointer value must be atomic, it must both be readable and
		4992	writable in one piece - this is the case on all current architectures.
		4993
3430	=item C<sig_atomic_t volatile> must be thread-atomic as well	4994	=item C<sig_atomic_t volatile> must be thread-atomic as well
3431		4995
3432	The type C<sig_atomic_t volatile> (or whatever is defined as	4996	The type C<sig_atomic_t volatile> (or whatever is defined as
3433	C<EV_ATOMIC_T>) must be atomic w.r.t. accesses from different	4997	C<EV_ATOMIC_T>) must be atomic with respect to accesses from different
3434	threads. This is not part of the specification for C<sig_atomic_t>, but is	4998	threads. This is not part of the specification for C<sig_atomic_t>, but is
3435	believed to be sufficiently portable.	4999	believed to be sufficiently portable.
3436		5000
3437	=item C<sigprocmask> must work in a threaded environment	5001	=item C<sigprocmask> must work in a threaded environment
3438		5002
…		…
3447	except the initial one, and run the default loop in the initial thread as	5011	except the initial one, and run the default loop in the initial thread as
3448	well.	5012	well.
3449		5013
3450	=item C<long> must be large enough for common memory allocation sizes	5014	=item C<long> must be large enough for common memory allocation sizes
3451		5015
3452	To improve portability and simplify using libev, libev uses C<long>	5016	To improve portability and simplify its API, libev uses C<long> internally
3453	internally instead of C<size_t> when allocating its data structures. On	5017	instead of C<size_t> when allocating its data structures. On non-POSIX
3454	non-POSIX systems (Microsoft...) this might be unexpectedly low, but	5018	systems (Microsoft...) this might be unexpectedly low, but is still at
3455	is still at least 31 bits everywhere, which is enough for hundreds of	5019	least 31 bits everywhere, which is enough for hundreds of millions of
3456	millions of watchers.	5020	watchers.
3457		5021
3458	=item C<double> must hold a time value in seconds with enough accuracy	5022	=item C<double> must hold a time value in seconds with enough accuracy
3459		5023
3460	The type C<double> is used to represent timestamps. It is required to	5024	The type C<double> is used to represent timestamps. It is required to
3461	have at least 51 bits of mantissa (and 9 bits of exponent), which is good	5025	have at least 51 bits of mantissa (and 9 bits of exponent), which is
3462	enough for at least into the year 4000. This requirement is fulfilled by	5026	good enough for at least into the year 4000 with millisecond accuracy
		5027	(the design goal for libev). This requirement is overfulfilled by
3463	implementations implementing IEEE 754 (basically all existing ones).	5028	implementations using IEEE 754, which is basically all existing ones. With
		5029	IEEE 754 doubles, you get microsecond accuracy until at least 2200.
3464		5030
3465	=back	5031	=back
3466		5032
3467	If you know of other additional requirements drop me a note.	5033	If you know of other additional requirements drop me a note.
3468		5034
3469		5035
3470	=head1 COMPILER WARNINGS	5036	=head1 ALGORITHMIC COMPLEXITIES
3471		5037
3472	Depending on your compiler and compiler settings, you might get no or a	5038	In this section the complexities of (many of) the algorithms used inside
3473	lot of warnings when compiling libev code. Some people are apparently	5039	libev will be documented. For complexity discussions about backends see
3474	scared by this.	5040	the documentation for C<ev_default_init>.
3475		5041
3476	However, these are unavoidable for many reasons. For one, each compiler	5042	All of the following are about amortised time: If an array needs to be
3477	has different warnings, and each user has different tastes regarding	5043	extended, libev needs to realloc and move the whole array, but this
3478	warning options. "Warn-free" code therefore cannot be a goal except when	5044	happens asymptotically rarer with higher number of elements, so O(1) might
3479	targeting a specific compiler and compiler-version.	5045	mean that libev does a lengthy realloc operation in rare cases, but on
		5046	average it is much faster and asymptotically approaches constant time.
3480		5047
3481	Another reason is that some compiler warnings require elaborate	5048	=over 4
3482	workarounds, or other changes to the code that make it less clear and less
3483	maintainable.
3484		5049
3485	And of course, some compiler warnings are just plain stupid, or simply	5050	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)
3486	wrong (because they don't actually warn about the condition their message
3487	seems to warn about).
3488		5051
3489	While libev is written to generate as few warnings as possible,	5052	This means that, when you have a watcher that triggers in one hour and
3490	"warn-free" code is not a goal, and it is recommended not to build libev	5053	there are 100 watchers that would trigger before that, then inserting will
3491	with any compiler warnings enabled unless you are prepared to cope with	5054	have to skip roughly seven (C<ld 100>) of these watchers.
3492	them (e.g. by ignoring them). Remember that warnings are just that:
3493	warnings, not errors, or proof of bugs.
3494		5055
		5056	=item Changing timer/periodic watchers (by autorepeat or calling again): O(log skipped_other_timers)
3495		5057
3496	=head1 VALGRIND	5058	That means that changing a timer costs less than removing/adding them,
		5059	as only the relative motion in the event queue has to be paid for.
3497		5060
3498	Valgrind has a special section here because it is a popular tool that is	5061	=item Starting io/check/prepare/idle/signal/child/fork/async watchers: O(1)
3499	highly useful, but valgrind reports are very hard to interpret.
3500		5062
3501	If you think you found a bug (memory leak, uninitialised data access etc.)	5063	These just add the watcher into an array or at the head of a list.
3502	in libev, then check twice: If valgrind reports something like:
3503		5064
3504	==2274== definitely lost: 0 bytes in 0 blocks.	5065	=item Stopping check/prepare/idle/fork/async watchers: O(1)
3505	==2274== possibly lost: 0 bytes in 0 blocks.
3506	==2274== still reachable: 256 bytes in 1 blocks.
3507		5066
3508	Then there is no memory leak. Similarly, under some circumstances,	5067	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))
3509	valgrind might report kernel bugs as if it were a bug in libev, or it
3510	might be confused (it is a very good tool, but only a tool).
3511		5068
3512	If you are unsure about something, feel free to contact the mailing list	5069	These watchers are stored in lists, so they need to be walked to find the
3513	with the full valgrind report and an explanation on why you think this is	5070	correct watcher to remove. The lists are usually short (you don't usually
3514	a bug in libev. However, don't be annoyed when you get a brisk "this is	5071	have many watchers waiting for the same fd or signal: one is typical, two
3515	no bug" answer and take the chance of learning how to interpret valgrind	5072	is rare).
3516	properly.
3517		5073
3518	If you need, for some reason, empty reports from valgrind for your project	5074	=item Finding the next timer in each loop iteration: O(1)
3519	I suggest using suppression lists.
3520		5075
		5076	By virtue of using a binary or 4-heap, the next timer is always found at a
		5077	fixed position in the storage array.
		5078
		5079	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)
		5080
		5081	A change means an I/O watcher gets started or stopped, which requires
		5082	libev to recalculate its status (and possibly tell the kernel, depending
		5083	on backend and whether C<ev_io_set> was used).
		5084
		5085	=item Activating one watcher (putting it into the pending state): O(1)
		5086
		5087	=item Priority handling: O(number_of_priorities)
		5088
		5089	Priorities are implemented by allocating some space for each
		5090	priority. When doing priority-based operations, libev usually has to
		5091	linearly search all the priorities, but starting/stopping and activating
		5092	watchers becomes O(1) with respect to priority handling.
		5093
		5094	=item Sending an ev_async: O(1)
		5095
		5096	=item Processing ev_async_send: O(number_of_async_watchers)
		5097
		5098	=item Processing signals: O(max_signal_number)
		5099
		5100	Sending involves a system call I<iff> there were no other C<ev_async_send>
		5101	calls in the current loop iteration. Checking for async and signal events
		5102	involves iterating over all running async watchers or all signal numbers.
		5103
		5104	=back
		5105
		5106
		5107	=head1 PORTING FROM LIBEV 3.X TO 4.X
		5108
		5109	The major version 4 introduced some incompatible changes to the API.
		5110
		5111	At the moment, the C<ev.h> header file provides compatibility definitions
		5112	for all changes, so most programs should still compile. The compatibility
		5113	layer might be removed in later versions of libev, so better update to the
		5114	new API early than late.
		5115
		5116	=over 4
		5117
		5118	=item C<EV_COMPAT3> backwards compatibility mechanism
		5119
		5120	The backward compatibility mechanism can be controlled by
		5121	C<EV_COMPAT3>. See L<PREPROCESSOR SYMBOLS/MACROS> in the L<EMBEDDING>
		5122	section.
		5123
		5124	=item C<ev_default_destroy> and C<ev_default_fork> have been removed
		5125
		5126	These calls can be replaced easily by their C<ev_loop_xxx> counterparts:
		5127
		5128	ev_loop_destroy (EV_DEFAULT_UC);
		5129	ev_loop_fork (EV_DEFAULT);
		5130
		5131	=item function/symbol renames
		5132
		5133	A number of functions and symbols have been renamed:
		5134
		5135	ev_loop => ev_run
		5136	EVLOOP_NONBLOCK => EVRUN_NOWAIT
		5137	EVLOOP_ONESHOT => EVRUN_ONCE
		5138
		5139	ev_unloop => ev_break
		5140	EVUNLOOP_CANCEL => EVBREAK_CANCEL
		5141	EVUNLOOP_ONE => EVBREAK_ONE
		5142	EVUNLOOP_ALL => EVBREAK_ALL
		5143
		5144	EV_TIMEOUT => EV_TIMER
		5145
		5146	ev_loop_count => ev_iteration
		5147	ev_loop_depth => ev_depth
		5148	ev_loop_verify => ev_verify
		5149
		5150	Most functions working on C<struct ev_loop> objects don't have an
		5151	C<ev_loop_> prefix, so it was removed; C<ev_loop>, C<ev_unloop> and
		5152	associated constants have been renamed to not collide with the C<struct
		5153	ev_loop> anymore and C<EV_TIMER> now follows the same naming scheme
		5154	as all other watcher types. Note that C<ev_loop_fork> is still called
		5155	C<ev_loop_fork> because it would otherwise clash with the C<ev_fork>
		5156	typedef.
		5157
		5158	=item C<EV_MINIMAL> mechanism replaced by C<EV_FEATURES>
		5159
		5160	The preprocessor symbol C<EV_MINIMAL> has been replaced by a different
		5161	mechanism, C<EV_FEATURES>. Programs using C<EV_MINIMAL> usually compile
		5162	and work, but the library code will of course be larger.
		5163
		5164	=back
		5165
		5166
		5167	=head1 GLOSSARY
		5168
		5169	=over 4
		5170
		5171	=item active
		5172
		5173	A watcher is active as long as it has been started and not yet stopped.
		5174	See L<WATCHER STATES> for details.
		5175
		5176	=item application
		5177
		5178	In this document, an application is whatever is using libev.
		5179
		5180	=item backend
		5181
		5182	The part of the code dealing with the operating system interfaces.
		5183
		5184	=item callback
		5185
		5186	The address of a function that is called when some event has been
		5187	detected. Callbacks are being passed the event loop, the watcher that
		5188	received the event, and the actual event bitset.
		5189
		5190	=item callback/watcher invocation
		5191
		5192	The act of calling the callback associated with a watcher.
		5193
		5194	=item event
		5195
		5196	A change of state of some external event, such as data now being available
		5197	for reading on a file descriptor, time having passed or simply not having
		5198	any other events happening anymore.
		5199
		5200	In libev, events are represented as single bits (such as C<EV_READ> or
		5201	C<EV_TIMER>).
		5202
		5203	=item event library
		5204
		5205	A software package implementing an event model and loop.
		5206
		5207	=item event loop
		5208
		5209	An entity that handles and processes external events and converts them
		5210	into callback invocations.
		5211
		5212	=item event model
		5213
		5214	The model used to describe how an event loop handles and processes
		5215	watchers and events.
		5216
		5217	=item pending
		5218
		5219	A watcher is pending as soon as the corresponding event has been
		5220	detected. See L<WATCHER STATES> for details.
		5221
		5222	=item real time
		5223
		5224	The physical time that is observed. It is apparently strictly monotonic :)
		5225
		5226	=item wall-clock time
		5227
		5228	The time and date as shown on clocks. Unlike real time, it can actually
		5229	be wrong and jump forwards and backwards, e.g. when you adjust your
		5230	clock.
		5231
		5232	=item watcher
		5233
		5234	A data structure that describes interest in certain events. Watchers need
		5235	to be started (attached to an event loop) before they can receive events.
		5236
		5237	=back
3521		5238
3522	=head1 AUTHOR	5239	=head1 AUTHOR
3523		5240
3524	Marc Lehmann <libev@schmorp.de>.	5241	Marc Lehmann <libev@schmorp.de>, with repeated corrections by Mikael
		5242	Magnusson and Emanuele Giaquinta, and minor corrections by many others.
3525		5243

Diff Legend

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

Comparing libev/ev.pod (file contents): Revision 1.172 by root, Wed Aug 6 07:01:25 2008 UTC vs. Revision 1.366 by sf-exg, Thu Feb 3 16:21:08 2011 UTC

Diff Legend

Comparing libev/ev.pod (file contents):
Revision 1.172 by root, Wed Aug 6 07:01:25 2008 UTC vs.
Revision 1.366 by sf-exg, Thu Feb 3 16:21:08 2011 UTC