ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/libev/ev.pod

Comparing libev/ev.pod (file contents):
Revision 1.198 by root, Thu Oct 23 06:30:48 2008 UTC vs.
Revision 1.320 by root, Fri Oct 22 10:48:54 2010 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
…		…
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_ 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	ev_loop *loop = ev_default_loop (0);	48	struct ev_loop *loop = ev_default_loop (0);
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	// unloop 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 ABOUT LIBEV
70		84
71	Libev is an event loop: you register interest in certain events (such as a	85	Libev is an event loop: you register interest in certain events (such as a
72	file descriptor being readable or a timeout occurring), and it will manage	86	file descriptor being readable or a timeout occurring), and it will manage
73	these event sources and provide your program with events.	87	these event sources and provide your program with events.
74		88
…		…
84	=head2 FEATURES	98	=head2 FEATURES
85		99
86	Libev supports C<select>, C<poll>, the Linux-specific C<epoll>, the	100	Libev supports C<select>, C<poll>, the Linux-specific C<epoll>, the
87	BSD-specific C<kqueue> and the Solaris-specific event port mechanisms	101	BSD-specific C<kqueue> and the Solaris-specific event port mechanisms
88	for file descriptor events (C<ev_io>), the Linux C<inotify> interface	102	for file descriptor events (C<ev_io>), the Linux C<inotify> interface
89	(for C<ev_stat>), relative timers (C<ev_timer>), absolute timers	103	(for C<ev_stat>), Linux eventfd/signalfd (for faster and cleaner
90	with customised rescheduling (C<ev_periodic>), synchronous signals	104	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	105	timers (C<ev_timer>), absolute timers with customised rescheduling
92	watchers dealing with the event loop mechanism itself (C<ev_idle>,	106	(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	107	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	108	loop mechanism itself (C<ev_idle>, C<ev_embed>, C<ev_prepare> and
95	(C<ev_fork>).	109	C<ev_check> watchers) as well as file watchers (C<ev_stat>) and even
		110	limited support for fork events (C<ev_fork>).
96		111
97	It also is quite fast (see this	112	It also is quite fast (see this
98	L<benchmark|http://libev.schmorp.de/bench.html> comparing it to libevent	113	L<benchmark|http://libev.schmorp.de/bench.html> comparing it to libevent
99	for example).	114	for example).
100		115
…		…
103	Libev is very configurable. In this manual the default (and most common)	118	Libev is very configurable. In this manual the default (and most common)
104	configuration will be described, which supports multiple event loops. For	119	configuration will be described, which supports multiple event loops. For
105	more info about various configuration options please have a look at	120	more info about various configuration options please have a look at
106	B<EMBED> section in this manual. If libev was configured without support	121	B<EMBED> section in this manual. If libev was configured without support
107	for multiple event loops, then all functions taking an initial argument of	122	for multiple event loops, then all functions taking an initial argument of
108	name C<loop> (which is always of type C<ev_loop *>) will not have	123	name C<loop> (which is always of type C<struct ev_loop *>) will not have
109	this argument.	124	this argument.
110		125
111	=head2 TIME REPRESENTATION	126	=head2 TIME REPRESENTATION
112		127
113	Libev represents time as a single floating point number, representing the	128	Libev represents time as a single floating point number, representing
114	(fractional) number of seconds since the (POSIX) epoch (somewhere near	129	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	130	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	131	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	132	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	133	any calculations on it, you should treat it as some floating point value.
		134
119	component C<stamp> might indicate, it is also used for time differences	135	Unlike the name component C<stamp> might indicate, it is also used for
120	throughout libev.	136	time differences (e.g. delays) throughout libev.
121		137
122	=head1 ERROR HANDLING	138	=head1 ERROR HANDLING
123		139
124	Libev knows three classes of errors: operating system errors, usage errors	140	Libev knows three classes of errors: operating system errors, usage errors
125	and internal errors (bugs).	141	and internal errors (bugs).
…		…
149		165
150	=item ev_tstamp ev_time ()	166	=item ev_tstamp ev_time ()
151		167
152	Returns the current time as libev would use it. Please note that the	168	Returns the current time as libev would use it. Please note that the
153	C<ev_now> function is usually faster and also often returns the timestamp	169	C<ev_now> function is usually faster and also often returns the timestamp
154	you actually want to know.	170	you actually want to know. Also interetsing is the combination of
		171	C<ev_update_now> and C<ev_now>.
155		172
156	=item ev_sleep (ev_tstamp interval)	173	=item ev_sleep (ev_tstamp interval)
157		174
158	Sleep for the given interval: The current thread will be blocked until	175	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	176	either it is interrupted or the given time interval has passed. Basically
…		…
176	as this indicates an incompatible change. Minor versions are usually	193	as this indicates an incompatible change. Minor versions are usually
177	compatible to older versions, so a larger minor version alone is usually	194	compatible to older versions, so a larger minor version alone is usually
178	not a problem.	195	not a problem.
179		196
180	Example: Make sure we haven't accidentally been linked against the wrong	197	Example: Make sure we haven't accidentally been linked against the wrong
181	version.	198	version (note, however, that this will not detect other ABI mismatches,
		199	such as LFS or reentrancy).
182		200
183	assert (("libev version mismatch",	201	assert (("libev version mismatch",
184	ev_version_major () == EV_VERSION_MAJOR	202	ev_version_major () == EV_VERSION_MAJOR
185	&& ev_version_minor () >= EV_VERSION_MINOR));	203	&& ev_version_minor () >= EV_VERSION_MINOR));
186		204
…		…
197	assert (("sorry, no epoll, no sex",	215	assert (("sorry, no epoll, no sex",
198	ev_supported_backends () & EVBACKEND_EPOLL));	216	ev_supported_backends () & EVBACKEND_EPOLL));
199		217
200	=item unsigned int ev_recommended_backends ()	218	=item unsigned int ev_recommended_backends ()
201		219
202	Return the set of all backends compiled into this binary of libev and also	220	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	221	also recommended for this platform, meaning it will work for most file
		222	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	223	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	224	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	225	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.	226	probe for if you specify no backends explicitly.
208		227
209	=item unsigned int ev_embeddable_backends ()	228	=item unsigned int ev_embeddable_backends ()
210		229
211	Returns the set of backends that are embeddable in other event loops. This	230	Returns the set of backends that are embeddable in other event loops. This
212	is the theoretical, all-platform, value. To find which backends	231	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	232	current system. To find which embeddable backends might be supported on
214	C<ev_embeddable_backends () & ev_supported_backends ()>, likewise for	233	the current system, you would need to look at C<ev_embeddable_backends ()
215	recommended ones.	234	& ev_supported_backends ()>, likewise for recommended ones.
216		235
217	See the description of C<ev_embed> watchers for more info.	236	See the description of C<ev_embed> watchers for more info.
218		237
219	=item ev_set_allocator (void *(*cb)(void *ptr, long size)) [NOT REENTRANT]	238	=item ev_set_allocator (void *(*cb)(void *ptr, long size)) [NOT REENTRANT]
220		239
…		…
276		295
277	=back	296	=back
278		297
279	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP	298	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP
280		299
281	An event loop is described by a C<ev_loop *>. The library knows two	300	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	301	I<not> optional in this case unless libev 3 compatibility is disabled, as
283	events, and dynamically created loops which do not.	302	libev 3 had an C<ev_loop> function colliding with the struct name).
		303
		304	The library knows two types of such loops, the I<default> loop, which
		305	supports signals and child events, and dynamically created event loops
		306	which do not.
284		307
285	=over 4	308	=over 4
286		309
287	=item struct ev_loop *ev_default_loop (unsigned int flags)	310	=item struct ev_loop *ev_default_loop (unsigned int flags)
288		311
…		…
294	If you don't know what event loop to use, use the one returned from this	317	If you don't know what event loop to use, use the one returned from this
295	function.	318	function.
296		319
297	Note that this function is I<not> thread-safe, so if you want to use it	320	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,	321	from multiple threads, you have to lock (note also that this is unlikely,
299	as loops cannot bes hared easily between threads anyway).	322	as loops cannot be shared easily between threads anyway).
300		323
301	The default loop is the only loop that can handle C<ev_signal> and	324	The default loop is the only loop that can handle C<ev_signal> and
302	C<ev_child> watchers, and to do this, it always registers a handler	325	C<ev_child> watchers, and to do this, it always registers a handler
303	for C<SIGCHLD>. If this is a problem for your application you can either	326	for C<SIGCHLD>. If this is a problem for your application you can either
304	create a dynamic loop with C<ev_loop_new> that doesn't do that, or you	327	create a dynamic loop with C<ev_loop_new> that doesn't do that, or you
…		…
326	useful to try out specific backends to test their performance, or to work	349	useful to try out specific backends to test their performance, or to work
327	around bugs.	350	around bugs.
328		351
329	=item C<EVFLAG_FORKCHECK>	352	=item C<EVFLAG_FORKCHECK>
330		353
331	Instead of calling C<ev_default_fork> or C<ev_loop_fork> manually after	354	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	355	make libev check for a fork in each iteration by enabling this flag.
333	enabling this flag.
334		356
335	This works by calling C<getpid ()> on every iteration of the loop,	357	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	358	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	359	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	360	GNU/Linux system for example, C<getpid> is actually a simple 5-insn sequence
…		…
344	flag.	366	flag.
345		367
346	This flag setting cannot be overridden or specified in the C<LIBEV_FLAGS>	368	This flag setting cannot be overridden or specified in the C<LIBEV_FLAGS>
347	environment variable.	369	environment variable.
348		370
		371	=item C<EVFLAG_NOINOTIFY>
		372
		373	When this flag is specified, then libev will not attempt to use the
		374	I<inotify> API for it's C<ev_stat> watchers. Apart from debugging and
		375	testing, this flag can be useful to conserve inotify file descriptors, as
		376	otherwise each loop using C<ev_stat> watchers consumes one inotify handle.
		377
		378	=item C<EVFLAG_SIGNALFD>
		379
		380	When this flag is specified, then libev will attempt to use the
		381	I<signalfd> API for it's C<ev_signal> (and C<ev_child>) watchers. This API
		382	delivers signals synchronously, which makes it both faster and might make
		383	it possible to get the queued signal data. It can also simplify signal
		384	handling with threads, as long as you properly block signals in your
		385	threads that are not interested in handling them.
		386
		387	Signalfd will not be used by default as this changes your signal mask, and
		388	there are a lot of shoddy libraries and programs (glib's threadpool for
		389	example) that can't properly initialise their signal masks.
		390
349	=item C<EVBACKEND_SELECT> (value 1, portable select backend)	391	=item C<EVBACKEND_SELECT> (value 1, portable select backend)
350		392
351	This is your standard select(2) backend. Not I<completely> standard, as	393	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,	394	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	395	but if that fails, expect a fairly low limit on the number of fds when
…		…
377	This backend maps C<EV_READ> to C<POLLIN | POLLERR | POLLHUP>, and	419	This backend maps C<EV_READ> to C<POLLIN | POLLERR | POLLHUP>, and
378	C<EV_WRITE> to C<POLLOUT | POLLERR | POLLHUP>.	420	C<EV_WRITE> to C<POLLOUT | POLLERR | POLLHUP>.
379		421
380	=item C<EVBACKEND_EPOLL> (value 4, Linux)	422	=item C<EVBACKEND_EPOLL> (value 4, Linux)
381		423
		424	Use the linux-specific epoll(7) interface (for both pre- and post-2.6.9
		425	kernels).
		426
382	For few fds, this backend is a bit little slower than poll and select,	427	For few fds, this backend is a bit little slower than poll and select,
383	but it scales phenomenally better. While poll and select usually scale	428	but it scales phenomenally better. While poll and select usually scale
384	like O(total_fds) where n is the total number of fds (or the highest fd),	429	like O(total_fds) where n is the total number of fds (or the highest fd),
385	epoll scales either O(1) or O(active_fds). The epoll design has a number	430	epoll scales either O(1) or O(active_fds).
386	of shortcomings, such as silently dropping events in some hard-to-detect	431
387	cases and requiring a system call per fd change, no fork support and bad	432	The epoll mechanism deserves honorable mention as the most misdesigned
388	support for dup.	433	of the more advanced event mechanisms: mere annoyances include silently
		434	dropping file descriptors, requiring a system call per change per file
		435	descriptor (and unnecessary guessing of parameters), problems with dup and
		436	so on. The biggest issue is fork races, however - if a program forks then
		437	I<both> parent and child process have to recreate the epoll set, which can
		438	take considerable time (one syscall per file descriptor) and is of course
		439	hard to detect.
		440
		441	Epoll is also notoriously buggy - embedding epoll fds I<should> work, but
		442	of course I<doesn't>, and epoll just loves to report events for totally
		443	I<different> file descriptors (even already closed ones, so one cannot
		444	even remove them from the set) than registered in the set (especially
		445	on SMP systems). Libev tries to counter these spurious notifications by
		446	employing an additional generation counter and comparing that against the
		447	events to filter out spurious ones, recreating the set when required. Last
		448	not least, it also refuses to work with some file descriptors which work
		449	perfectly fine with C<select> (files, many character devices...).
389		450
390	While stopping, setting and starting an I/O watcher in the same iteration	451	While stopping, setting and starting an I/O watcher in the same iteration
391	will result in some caching, there is still a system call per such incident	452	will result in some caching, there is still a system call per such
392	(because the fd could point to a different file description now), so its	453	incident (because the same I<file descriptor> could point to a different
393	best to avoid that. Also, C<dup ()>'ed file descriptors might not work	454	I<file description> now), so its best to avoid that. Also, C<dup ()>'ed
394	very well if you register events for both fds.	455	file descriptors might not work very well if you register events for both
395		456	file descriptors.
396	Please note that epoll sometimes generates spurious notifications, so you
397	need to use non-blocking I/O or other means to avoid blocking when no data
398	(or space) is available.
399		457
400	Best performance from this backend is achieved by not unregistering all	458	Best performance from this backend is achieved by not unregistering all
401	watchers for a file descriptor until it has been closed, if possible,	459	watchers for a file descriptor until it has been closed, if possible,
402	i.e. keep at least one watcher active per fd at all times. Stopping and	460	i.e. keep at least one watcher active per fd at all times. Stopping and
403	starting a watcher (without re-setting it) also usually doesn't cause	461	starting a watcher (without re-setting it) also usually doesn't cause
404	extra overhead.	462	extra overhead. A fork can both result in spurious notifications as well
		463	as in libev having to destroy and recreate the epoll object, which can
		464	take considerable time and thus should be avoided.
		465
		466	All this means that, in practice, C<EVBACKEND_SELECT> can be as fast or
		467	faster than epoll for maybe up to a hundred file descriptors, depending on
		468	the usage. So sad.
405		469
406	While nominally embeddable in other event loops, this feature is broken in	470	While nominally embeddable in other event loops, this feature is broken in
407	all kernel versions tested so far.	471	all kernel versions tested so far.
408		472
409	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as	473	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
410	C<EVBACKEND_POLL>.	474	C<EVBACKEND_POLL>.
411		475
412	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)	476	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)
413		477
414	Kqueue deserves special mention, as at the time of this writing, it was	478	Kqueue deserves special mention, as at the time of this writing, it
415	broken on all BSDs except NetBSD (usually it doesn't work reliably with	479	was broken on all BSDs except NetBSD (usually it doesn't work reliably
416	anything but sockets and pipes, except on Darwin, where of course it's	480	with anything but sockets and pipes, except on Darwin, where of course
417	completely useless). For this reason it's not being "auto-detected" unless	481	it's completely useless). Unlike epoll, however, whose brokenness
418	you explicitly specify it in the flags (i.e. using C<EVBACKEND_KQUEUE>) or	482	is by design, these kqueue bugs can (and eventually will) be fixed
419	libev was compiled on a known-to-be-good (-enough) system like NetBSD.	483	without API changes to existing programs. For this reason it's not being
		484	"auto-detected" unless you explicitly specify it in the flags (i.e. using
		485	C<EVBACKEND_KQUEUE>) or libev was compiled on a known-to-be-good (-enough)
		486	system like NetBSD.
420		487
421	You still can embed kqueue into a normal poll or select backend and use it	488	You still can embed kqueue into a normal poll or select backend and use it
422	only for sockets (after having made sure that sockets work with kqueue on	489	only for sockets (after having made sure that sockets work with kqueue on
423	the target platform). See C<ev_embed> watchers for more info.	490	the target platform). See C<ev_embed> watchers for more info.
424		491
425	It scales in the same way as the epoll backend, but the interface to the	492	It scales in the same way as the epoll backend, but the interface to the
426	kernel is more efficient (which says nothing about its actual speed, of	493	kernel is more efficient (which says nothing about its actual speed, of
427	course). While stopping, setting and starting an I/O watcher does never	494	course). While stopping, setting and starting an I/O watcher does never
428	cause an extra system call as with C<EVBACKEND_EPOLL>, it still adds up to	495	cause an extra system call as with C<EVBACKEND_EPOLL>, it still adds up to
429	two event changes per incident. Support for C<fork ()> is very bad and it	496	two event changes per incident. Support for C<fork ()> is very bad (but
430	drops fds silently in similarly hard-to-detect cases.	497	sane, unlike epoll) and it drops fds silently in similarly hard-to-detect
		498	cases
431		499
432	This backend usually performs well under most conditions.	500	This backend usually performs well under most conditions.
433		501
434	While nominally embeddable in other event loops, this doesn't work	502	While nominally embeddable in other event loops, this doesn't work
435	everywhere, so you might need to test for this. And since it is broken	503	everywhere, so you might need to test for this. And since it is broken
436	almost everywhere, you should only use it when you have a lot of sockets	504	almost everywhere, you should only use it when you have a lot of sockets
437	(for which it usually works), by embedding it into another event loop	505	(for which it usually works), by embedding it into another event loop
438	(e.g. C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>) and, did I mention it,	506	(e.g. C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> (but C<poll> is of course
439	using it only for sockets.	507	also broken on OS X)) and, did I mention it, using it only for sockets.
440		508
441	This backend maps C<EV_READ> into an C<EVFILT_READ> kevent with	509	This backend maps C<EV_READ> into an C<EVFILT_READ> kevent with
442	C<NOTE_EOF>, and C<EV_WRITE> into an C<EVFILT_WRITE> kevent with	510	C<NOTE_EOF>, and C<EV_WRITE> into an C<EVFILT_WRITE> kevent with
443	C<NOTE_EOF>.	511	C<NOTE_EOF>.
444		512
…		…
464	might perform better.	532	might perform better.
465		533
466	On the positive side, with the exception of the spurious readiness	534	On the positive side, with the exception of the spurious readiness
467	notifications, this backend actually performed fully to specification	535	notifications, this backend actually performed fully to specification
468	in all tests and is fully embeddable, which is a rare feat among the	536	in all tests and is fully embeddable, which is a rare feat among the
469	OS-specific backends.	537	OS-specific backends (I vastly prefer correctness over speed hacks).
470		538
471	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as	539	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
472	C<EVBACKEND_POLL>.	540	C<EVBACKEND_POLL>.
473		541
474	=item C<EVBACKEND_ALL>	542	=item C<EVBACKEND_ALL>
…		…
479		547
480	It is definitely not recommended to use this flag.	548	It is definitely not recommended to use this flag.
481		549
482	=back	550	=back
483		551
484	If one or more of these are or'ed into the flags value, then only these	552	If one or more of the backend flags are or'ed into the flags value,
485	backends will be tried (in the reverse order as listed here). If none are	553	then only these backends will be tried (in the reverse order as listed
486	specified, all backends in C<ev_recommended_backends ()> will be tried.	554	here). If none are specified, all backends in C<ev_recommended_backends
		555	()> will be tried.
487		556
488	Example: This is the most typical usage.	557	Example: This is the most typical usage.
489		558
490	if (!ev_default_loop (0))	559	if (!ev_default_loop (0))
491	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");	560	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
…		…
503	ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);	572	ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);
504		573
505	=item struct ev_loop *ev_loop_new (unsigned int flags)	574	=item struct ev_loop *ev_loop_new (unsigned int flags)
506		575
507	Similar to C<ev_default_loop>, but always creates a new event loop that is	576	Similar to C<ev_default_loop>, but always creates a new event loop that is
508	always distinct from the default loop. Unlike the default loop, it cannot	577	always distinct from the default loop.
509	handle signal and child watchers, and attempts to do so will be greeted by
510	undefined behaviour (or a failed assertion if assertions are enabled).
511		578
512	Note that this function I<is> thread-safe, and the recommended way to use	579	Note that this function I<is> thread-safe, and one common way to use
513	libev with threads is indeed to create one loop per thread, and using the	580	libev with threads is indeed to create one loop per thread, and using the
514	default loop in the "main" or "initial" thread.	581	default loop in the "main" or "initial" thread.
515		582
516	Example: Try to create a event loop that uses epoll and nothing else.	583	Example: Try to create a event loop that uses epoll and nothing else.
517		584
…		…
519	if (!epoller)	586	if (!epoller)
520	fatal ("no epoll found here, maybe it hides under your chair");	587	fatal ("no epoll found here, maybe it hides under your chair");
521		588
522	=item ev_default_destroy ()	589	=item ev_default_destroy ()
523		590
524	Destroys the default loop again (frees all memory and kernel state	591	Destroys the default loop (frees all memory and kernel state etc.). None
525	etc.). None of the active event watchers will be stopped in the normal	592	of the active event watchers will be stopped in the normal sense, so
526	sense, so e.g. C<ev_is_active> might still return true. It is your	593	e.g. C<ev_is_active> might still return true. It is your responsibility to
527	responsibility to either stop all watchers cleanly yourself I<before>	594	either stop all watchers cleanly yourself I<before> calling this function,
528	calling this function, or cope with the fact afterwards (which is usually	595	or cope with the fact afterwards (which is usually the easiest thing, you
529	the easiest thing, you can just ignore the watchers and/or C<free ()> them	596	can just ignore the watchers and/or C<free ()> them for example).
530	for example).
531		597
532	Note that certain global state, such as signal state, will not be freed by	598	Note that certain global state, such as signal state (and installed signal
533	this function, and related watchers (such as signal and child watchers)	599	handlers), will not be freed by this function, and related watchers (such
534	would need to be stopped manually.	600	as signal and child watchers) would need to be stopped manually.
535		601
536	In general it is not advisable to call this function except in the	602	In general it is not advisable to call this function except in the
537	rare occasion where you really need to free e.g. the signal handling	603	rare occasion where you really need to free e.g. the signal handling
538	pipe fds. If you need dynamically allocated loops it is better to use	604	pipe fds. If you need dynamically allocated loops it is better to use
539	C<ev_loop_new> and C<ev_loop_destroy>).	605	C<ev_loop_new> and C<ev_loop_destroy>.
540		606
541	=item ev_loop_destroy (loop)	607	=item ev_loop_destroy (loop)
542		608
543	Like C<ev_default_destroy>, but destroys an event loop created by an	609	Like C<ev_default_destroy>, but destroys an event loop created by an
544	earlier call to C<ev_loop_new>.	610	earlier call to C<ev_loop_new>.
545		611
546	=item ev_default_fork ()	612	=item ev_default_fork ()
547		613
548	This function sets a flag that causes subsequent C<ev_loop> iterations	614	This function sets a flag that causes subsequent C<ev_run> iterations
549	to reinitialise the kernel state for backends that have one. Despite the	615	to reinitialise the kernel state for backends that have one. Despite the
550	name, you can call it anytime, but it makes most sense after forking, in	616	name, you can call it anytime, but it makes most sense after forking, in
551	the child process (or both child and parent, but that again makes little	617	the child process (or both child and parent, but that again makes little
552	sense). You I<must> call it in the child before using any of the libev	618	sense). You I<must> call it in the child before using any of the libev
553	functions, and it will only take effect at the next C<ev_loop> iteration.	619	functions, and it will only take effect at the next C<ev_run> iteration.
		620
		621	Again, you I<have> to call it on I<any> loop that you want to re-use after
		622	a fork, I<even if you do not plan to use the loop in the parent>. This is
		623	because some kernel interfaces *cough* I<kqueue> *cough* do funny things
		624	during fork.
554		625
555	On the other hand, you only need to call this function in the child	626	On the other hand, you only need to call this function in the child
556	process if and only if you want to use the event library in the child. If	627	process if and only if you want to use the event loop in the child. If
557	you just fork+exec, you don't have to call it at all.	628	you just fork+exec or create a new loop in the child, you don't have to
		629	call it at all (in fact, C<epoll> is so badly broken that it makes a
		630	difference, but libev will usually detect this case on its own and do a
		631	costly reset of the backend).
558		632
559	The function itself is quite fast and it's usually not a problem to call	633	The function itself is quite fast and it's usually not a problem to call
560	it just in case after a fork. To make this easy, the function will fit in	634	it just in case after a fork. To make this easy, the function will fit in
561	quite nicely into a call to C<pthread_atfork>:	635	quite nicely into a call to C<pthread_atfork>:
562		636
…		…
564		638
565	=item ev_loop_fork (loop)	639	=item ev_loop_fork (loop)
566		640
567	Like C<ev_default_fork>, but acts on an event loop created by	641	Like C<ev_default_fork>, but acts on an event loop created by
568	C<ev_loop_new>. Yes, you have to call this on every allocated event loop	642	C<ev_loop_new>. Yes, you have to call this on every allocated event loop
569	after fork that you want to re-use in the child, and how you do this is	643	after fork that you want to re-use in the child, and how you keep track of
570	entirely your own problem.	644	them is entirely your own problem.
571		645
572	=item int ev_is_default_loop (loop)	646	=item int ev_is_default_loop (loop)
573		647
574	Returns true when the given loop is, in fact, the default loop, and false	648	Returns true when the given loop is, in fact, the default loop, and false
575	otherwise.	649	otherwise.
576		650
577	=item unsigned int ev_loop_count (loop)	651	=item unsigned int ev_iteration (loop)
578		652
579	Returns the count of loop iterations for the loop, which is identical to	653	Returns the current iteration count for the event loop, which is identical
580	the number of times libev did poll for new events. It starts at C<0> and	654	to the number of times libev did poll for new events. It starts at C<0>
581	happily wraps around with enough iterations.	655	and happily wraps around with enough iterations.
582		656
583	This value can sometimes be useful as a generation counter of sorts (it	657	This value can sometimes be useful as a generation counter of sorts (it
584	"ticks" the number of loop iterations), as it roughly corresponds with	658	"ticks" the number of loop iterations), as it roughly corresponds with
585	C<ev_prepare> and C<ev_check> calls.	659	C<ev_prepare> and C<ev_check> calls - and is incremented between the
		660	prepare and check phases.
		661
		662	=item unsigned int ev_depth (loop)
		663
		664	Returns the number of times C<ev_run> was entered minus the number of
		665	times C<ev_run> was exited, in other words, the recursion depth.
		666
		667	Outside C<ev_run>, this number is zero. In a callback, this number is
		668	C<1>, unless C<ev_run> was invoked recursively (or from another thread),
		669	in which case it is higher.
		670
		671	Leaving C<ev_run> abnormally (setjmp/longjmp, cancelling the thread
		672	etc.), doesn't count as "exit" - consider this as a hint to avoid such
		673	ungentleman-like behaviour unless it's really convenient.
586		674
587	=item unsigned int ev_backend (loop)	675	=item unsigned int ev_backend (loop)
588		676
589	Returns one of the C<EVBACKEND_*> flags indicating the event backend in	677	Returns one of the C<EVBACKEND_*> flags indicating the event backend in
590	use.	678	use.
…		…
599		687
600	=item ev_now_update (loop)	688	=item ev_now_update (loop)
601		689
602	Establishes the current time by querying the kernel, updating the time	690	Establishes the current time by querying the kernel, updating the time
603	returned by C<ev_now ()> in the progress. This is a costly operation and	691	returned by C<ev_now ()> in the progress. This is a costly operation and
604	is usually done automatically within C<ev_loop ()>.	692	is usually done automatically within C<ev_run ()>.
605		693
606	This function is rarely useful, but when some event callback runs for a	694	This function is rarely useful, but when some event callback runs for a
607	very long time without entering the event loop, updating libev's idea of	695	very long time without entering the event loop, updating libev's idea of
608	the current time is a good idea.	696	the current time is a good idea.
609		697
610	See also "The special problem of time updates" in the C<ev_timer> section.	698	See also L<The special problem of time updates> in the C<ev_timer> section.
611		699
		700	=item ev_suspend (loop)
		701
		702	=item ev_resume (loop)
		703
		704	These two functions suspend and resume an event loop, for use when the
		705	loop is not used for a while and timeouts should not be processed.
		706
		707	A typical use case would be an interactive program such as a game: When
		708	the user presses C<^Z> to suspend the game and resumes it an hour later it
		709	would be best to handle timeouts as if no time had actually passed while
		710	the program was suspended. This can be achieved by calling C<ev_suspend>
		711	in your C<SIGTSTP> handler, sending yourself a C<SIGSTOP> and calling
		712	C<ev_resume> directly afterwards to resume timer processing.
		713
		714	Effectively, all C<ev_timer> watchers will be delayed by the time spend
		715	between C<ev_suspend> and C<ev_resume>, and all C<ev_periodic> watchers
		716	will be rescheduled (that is, they will lose any events that would have
		717	occurred while suspended).
		718
		719	After calling C<ev_suspend> you B<must not> call I<any> function on the
		720	given loop other than C<ev_resume>, and you B<must not> call C<ev_resume>
		721	without a previous call to C<ev_suspend>.
		722
		723	Calling C<ev_suspend>/C<ev_resume> has the side effect of updating the
		724	event loop time (see C<ev_now_update>).
		725
612	=item ev_loop (loop, int flags)	726	=item ev_run (loop, int flags)
613		727
614	Finally, this is it, the event handler. This function usually is called	728	Finally, this is it, the event handler. This function usually is called
615	after you initialised all your watchers and you want to start handling	729	after you have initialised all your watchers and you want to start
616	events.	730	handling events. It will ask the operating system for any new events, call
		731	the watcher callbacks, an then repeat the whole process indefinitely: This
		732	is why event loops are called I<loops>.
617		733
618	If the flags argument is specified as C<0>, it will not return until	734	If the flags argument is specified as C<0>, it will keep handling events
619	either no event watchers are active anymore or C<ev_unloop> was called.	735	until either no event watchers are active anymore or C<ev_break> was
		736	called.
620		737
621	Please note that an explicit C<ev_unloop> is usually better than	738	Please note that an explicit C<ev_break> is usually better than
622	relying on all watchers to be stopped when deciding when a program has	739	relying on all watchers to be stopped when deciding when a program has
623	finished (especially in interactive programs), but having a program	740	finished (especially in interactive programs), but having a program
624	that automatically loops as long as it has to and no longer by virtue	741	that automatically loops as long as it has to and no longer by virtue
625	of relying on its watchers stopping correctly, that is truly a thing of	742	of relying on its watchers stopping correctly, that is truly a thing of
626	beauty.	743	beauty.
627		744
628	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle	745	A flags value of C<EVRUN_NOWAIT> will look for new events, will handle
629	those events and any already outstanding ones, but will not block your	746	those events and any already outstanding ones, but will not wait and
630	process in case there are no events and will return after one iteration of	747	block your process in case there are no events and will return after one
631	the loop.	748	iteration of the loop. This is sometimes useful to poll and handle new
		749	events while doing lengthy calculations, to keep the program responsive.
632		750
633	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if	751	A flags value of C<EVRUN_ONCE> will look for new events (waiting if
634	necessary) and will handle those and any already outstanding ones. It	752	necessary) and will handle those and any already outstanding ones. It
635	will block your process until at least one new event arrives (which could	753	will block your process until at least one new event arrives (which could
636	be an event internal to libev itself, so there is no guarentee that a	754	be an event internal to libev itself, so there is no guarantee that a
637	user-registered callback will be called), and will return after one	755	user-registered callback will be called), and will return after one
638	iteration of the loop.	756	iteration of the loop.
639		757
640	This is useful if you are waiting for some external event in conjunction	758	This is useful if you are waiting for some external event in conjunction
641	with something not expressible using other libev watchers (i.e. "roll your	759	with something not expressible using other libev watchers (i.e. "roll your
642	own C<ev_loop>"). However, a pair of C<ev_prepare>/C<ev_check> watchers is	760	own C<ev_run>"). However, a pair of C<ev_prepare>/C<ev_check> watchers is
643	usually a better approach for this kind of thing.	761	usually a better approach for this kind of thing.
644		762
645	Here are the gory details of what C<ev_loop> does:	763	Here are the gory details of what C<ev_run> does:
646		764
		765	- Increment loop depth.
		766	- Reset the ev_break status.
647	- Before the first iteration, call any pending watchers.	767	- Before the first iteration, call any pending watchers.
		768	LOOP:
648	* If EVFLAG_FORKCHECK was used, check for a fork.	769	- If EVFLAG_FORKCHECK was used, check for a fork.
649	- If a fork was detected (by any means), queue and call all fork watchers.	770	- If a fork was detected (by any means), queue and call all fork watchers.
650	- Queue and call all prepare watchers.	771	- Queue and call all prepare watchers.
		772	- If ev_break was called, goto FINISH.
651	- If we have been forked, detach and recreate the kernel state	773	- If we have been forked, detach and recreate the kernel state
652	as to not disturb the other process.	774	as to not disturb the other process.
653	- Update the kernel state with all outstanding changes.	775	- Update the kernel state with all outstanding changes.
654	- Update the "event loop time" (ev_now ()).	776	- Update the "event loop time" (ev_now ()).
655	- Calculate for how long to sleep or block, if at all	777	- Calculate for how long to sleep or block, if at all
656	(active idle watchers, EVLOOP_NONBLOCK or not having	778	(active idle watchers, EVRUN_NOWAIT or not having
657	any active watchers at all will result in not sleeping).	779	any active watchers at all will result in not sleeping).
658	- Sleep if the I/O and timer collect interval say so.	780	- Sleep if the I/O and timer collect interval say so.
		781	- Increment loop iteration counter.
659	- Block the process, waiting for any events.	782	- Block the process, waiting for any events.
660	- Queue all outstanding I/O (fd) events.	783	- Queue all outstanding I/O (fd) events.
661	- Update the "event loop time" (ev_now ()), and do time jump adjustments.	784	- Update the "event loop time" (ev_now ()), and do time jump adjustments.
662	- Queue all expired timers.	785	- Queue all expired timers.
663	- Queue all expired periodics.	786	- Queue all expired periodics.
664	- Unless any events are pending now, queue all idle watchers.	787	- Queue all idle watchers with priority higher than that of pending events.
665	- Queue all check watchers.	788	- Queue all check watchers.
666	- Call all queued watchers in reverse order (i.e. check watchers first).	789	- Call all queued watchers in reverse order (i.e. check watchers first).
667	Signals and child watchers are implemented as I/O watchers, and will	790	Signals and child watchers are implemented as I/O watchers, and will
668	be handled here by queueing them when their watcher gets executed.	791	be handled here by queueing them when their watcher gets executed.
669	- If ev_unloop has been called, or EVLOOP_ONESHOT or EVLOOP_NONBLOCK	792	- If ev_break has been called, or EVRUN_ONCE or EVRUN_NOWAIT
670	were used, or there are no active watchers, return, otherwise	793	were used, or there are no active watchers, goto FINISH, otherwise
671	continue with step *.	794	continue with step LOOP.
		795	FINISH:
		796	- Reset the ev_break status iff it was EVBREAK_ONE.
		797	- Decrement the loop depth.
		798	- Return.
672		799
673	Example: Queue some jobs and then loop until no events are outstanding	800	Example: Queue some jobs and then loop until no events are outstanding
674	anymore.	801	anymore.
675		802
676	... queue jobs here, make sure they register event watchers as long	803	... queue jobs here, make sure they register event watchers as long
677	... as they still have work to do (even an idle watcher will do..)	804	... as they still have work to do (even an idle watcher will do..)
678	ev_loop (my_loop, 0);	805	ev_run (my_loop, 0);
679	... jobs done or somebody called unloop. yeah!	806	... jobs done or somebody called unloop. yeah!
680		807
681	=item ev_unloop (loop, how)	808	=item ev_break (loop, how)
682		809
683	Can be used to make a call to C<ev_loop> return early (but only after it	810	Can be used to make a call to C<ev_run> return early (but only after it
684	has processed all outstanding events). The C<how> argument must be either	811	has processed all outstanding events). The C<how> argument must be either
685	C<EVUNLOOP_ONE>, which will make the innermost C<ev_loop> call return, or	812	C<EVBREAK_ONE>, which will make the innermost C<ev_run> call return, or
686	C<EVUNLOOP_ALL>, which will make all nested C<ev_loop> calls return.	813	C<EVBREAK_ALL>, which will make all nested C<ev_run> calls return.
687		814
688	This "unloop state" will be cleared when entering C<ev_loop> again.	815	This "unloop state" will be cleared when entering C<ev_run> again.
689		816
690	It is safe to call C<ev_unloop> from otuside any C<ev_loop> calls.	817	It is safe to call C<ev_break> from outside any C<ev_run> calls. ##TODO##
691		818
692	=item ev_ref (loop)	819	=item ev_ref (loop)
693		820
694	=item ev_unref (loop)	821	=item ev_unref (loop)
695		822
696	Ref/unref can be used to add or remove a reference count on the event	823	Ref/unref can be used to add or remove a reference count on the event
697	loop: Every watcher keeps one reference, and as long as the reference	824	loop: Every watcher keeps one reference, and as long as the reference
698	count is nonzero, C<ev_loop> will not return on its own.	825	count is nonzero, C<ev_run> will not return on its own.
699		826
700	If you have a watcher you never unregister that should not keep C<ev_loop>	827	This is useful when you have a watcher that you never intend to
701	from returning, call ev_unref() after starting, and ev_ref() before	828	unregister, but that nevertheless should not keep C<ev_run> from
		829	returning. In such a case, call C<ev_unref> after starting, and C<ev_ref>
702	stopping it.	830	before stopping it.
703		831
704	As an example, libev itself uses this for its internal signal pipe: It is	832	As an example, libev itself uses this for its internal signal pipe: It
705	not visible to the libev user and should not keep C<ev_loop> from exiting	833	is not visible to the libev user and should not keep C<ev_run> from
706	if no event watchers registered by it are active. It is also an excellent	834	exiting if no event watchers registered by it are active. It is also an
707	way to do this for generic recurring timers or from within third-party	835	excellent way to do this for generic recurring timers or from within
708	libraries. Just remember to I<unref after start> and I<ref before stop>	836	third-party libraries. Just remember to I<unref after start> and I<ref
709	(but only if the watcher wasn't active before, or was active before,	837	before stop> (but only if the watcher wasn't active before, or was active
710	respectively).	838	before, respectively. Note also that libev might stop watchers itself
		839	(e.g. non-repeating timers) in which case you have to C<ev_ref>
		840	in the callback).
711		841
712	Example: Create a signal watcher, but keep it from keeping C<ev_loop>	842	Example: Create a signal watcher, but keep it from keeping C<ev_run>
713	running when nothing else is active.	843	running when nothing else is active.
714		844
715	ev_signal exitsig;	845	ev_signal exitsig;
716	ev_signal_init (&exitsig, sig_cb, SIGINT);	846	ev_signal_init (&exitsig, sig_cb, SIGINT);
717	ev_signal_start (loop, &exitsig);	847	ev_signal_start (loop, &exitsig);
…		…
744		874
745	By setting a higher I<io collect interval> you allow libev to spend more	875	By setting a higher I<io collect interval> you allow libev to spend more
746	time collecting I/O events, so you can handle more events per iteration,	876	time collecting I/O events, so you can handle more events per iteration,
747	at the cost of increasing latency. Timeouts (both C<ev_periodic> and	877	at the cost of increasing latency. Timeouts (both C<ev_periodic> and
748	C<ev_timer>) will be not affected. Setting this to a non-null value will	878	C<ev_timer>) will be not affected. Setting this to a non-null value will
749	introduce an additional C<ev_sleep ()> call into most loop iterations.	879	introduce an additional C<ev_sleep ()> call into most loop iterations. The
		880	sleep time ensures that libev will not poll for I/O events more often then
		881	once per this interval, on average.
750		882
751	Likewise, by setting a higher I<timeout collect interval> you allow libev	883	Likewise, by setting a higher I<timeout collect interval> you allow libev
752	to spend more time collecting timeouts, at the expense of increased	884	to spend more time collecting timeouts, at the expense of increased
753	latency/jitter/inexactness (the watcher callback will be called	885	latency/jitter/inexactness (the watcher callback will be called
754	later). C<ev_io> watchers will not be affected. Setting this to a non-null	886	later). C<ev_io> watchers will not be affected. Setting this to a non-null
…		…
756		888
757	Many (busy) programs can usually benefit by setting the I/O collect	889	Many (busy) programs can usually benefit by setting the I/O collect
758	interval to a value near C<0.1> or so, which is often enough for	890	interval to a value near C<0.1> or so, which is often enough for
759	interactive servers (of course not for games), likewise for timeouts. It	891	interactive servers (of course not for games), likewise for timeouts. It
760	usually doesn't make much sense to set it to a lower value than C<0.01>,	892	usually doesn't make much sense to set it to a lower value than C<0.01>,
761	as this approaches the timing granularity of most systems.	893	as this approaches the timing granularity of most systems. Note that if
		894	you do transactions with the outside world and you can't increase the
		895	parallelity, then this setting will limit your transaction rate (if you
		896	need to poll once per transaction and the I/O collect interval is 0.01,
		897	then you can't do more than 100 transactions per second).
762		898
763	Setting the I<timeout collect interval> can improve the opportunity for	899	Setting the I<timeout collect interval> can improve the opportunity for
764	saving power, as the program will "bundle" timer callback invocations that	900	saving power, as the program will "bundle" timer callback invocations that
765	are "near" in time together, by delaying some, thus reducing the number of	901	are "near" in time together, by delaying some, thus reducing the number of
766	times the process sleeps and wakes up again. Another useful technique to	902	times the process sleeps and wakes up again. Another useful technique to
767	reduce iterations/wake-ups is to use C<ev_periodic> watchers and make sure	903	reduce iterations/wake-ups is to use C<ev_periodic> watchers and make sure
768	they fire on, say, one-second boundaries only.	904	they fire on, say, one-second boundaries only.
769		905
		906	Example: we only need 0.1s timeout granularity, and we wish not to poll
		907	more often than 100 times per second:
		908
		909	ev_set_timeout_collect_interval (EV_DEFAULT_UC_ 0.1);
		910	ev_set_io_collect_interval (EV_DEFAULT_UC_ 0.01);
		911
		912	=item ev_invoke_pending (loop)
		913
		914	This call will simply invoke all pending watchers while resetting their
		915	pending state. Normally, C<ev_run> does this automatically when required,
		916	but when overriding the invoke callback this call comes handy. This
		917	function can be invoked from a watcher - this can be useful for example
		918	when you want to do some lengthy calculation and want to pass further
		919	event handling to another thread (you still have to make sure only one
		920	thread executes within C<ev_invoke_pending> or C<ev_run> of course).
		921
		922	=item int ev_pending_count (loop)
		923
		924	Returns the number of pending watchers - zero indicates that no watchers
		925	are pending.
		926
		927	=item ev_set_invoke_pending_cb (loop, void (*invoke_pending_cb)(EV_P))
		928
		929	This overrides the invoke pending functionality of the loop: Instead of
		930	invoking all pending watchers when there are any, C<ev_run> will call
		931	this callback instead. This is useful, for example, when you want to
		932	invoke the actual watchers inside another context (another thread etc.).
		933
		934	If you want to reset the callback, use C<ev_invoke_pending> as new
		935	callback.
		936
		937	=item ev_set_loop_release_cb (loop, void (*release)(EV_P), void (*acquire)(EV_P))
		938
		939	Sometimes you want to share the same loop between multiple threads. This
		940	can be done relatively simply by putting mutex_lock/unlock calls around
		941	each call to a libev function.
		942
		943	However, C<ev_run> can run an indefinite time, so it is not feasible
		944	to wait for it to return. One way around this is to wake up the event
		945	loop via C<ev_break> and C<av_async_send>, another way is to set these
		946	I<release> and I<acquire> callbacks on the loop.
		947
		948	When set, then C<release> will be called just before the thread is
		949	suspended waiting for new events, and C<acquire> is called just
		950	afterwards.
		951
		952	Ideally, C<release> will just call your mutex_unlock function, and
		953	C<acquire> will just call the mutex_lock function again.
		954
		955	While event loop modifications are allowed between invocations of
		956	C<release> and C<acquire> (that's their only purpose after all), no
		957	modifications done will affect the event loop, i.e. adding watchers will
		958	have no effect on the set of file descriptors being watched, or the time
		959	waited. Use an C<ev_async> watcher to wake up C<ev_run> when you want it
		960	to take note of any changes you made.
		961
		962	In theory, threads executing C<ev_run> will be async-cancel safe between
		963	invocations of C<release> and C<acquire>.
		964
		965	See also the locking example in the C<THREADS> section later in this
		966	document.
		967
		968	=item ev_set_userdata (loop, void *data)
		969
		970	=item ev_userdata (loop)
		971
		972	Set and retrieve a single C<void *> associated with a loop. When
		973	C<ev_set_userdata> has never been called, then C<ev_userdata> returns
		974	C<0.>
		975
		976	These two functions can be used to associate arbitrary data with a loop,
		977	and are intended solely for the C<invoke_pending_cb>, C<release> and
		978	C<acquire> callbacks described above, but of course can be (ab-)used for
		979	any other purpose as well.
		980
770	=item ev_loop_verify (loop)	981	=item ev_verify (loop)
771		982
772	This function only does something when C<EV_VERIFY> support has been	983	This function only does something when C<EV_VERIFY> support has been
773	compiled in. which is the default for non-minimal builds. It tries to go	984	compiled in, which is the default for non-minimal builds. It tries to go
774	through all internal structures and checks them for validity. If anything	985	through all internal structures and checks them for validity. If anything
775	is found to be inconsistent, it will print an error message to standard	986	is found to be inconsistent, it will print an error message to standard
776	error and call C<abort ()>.	987	error and call C<abort ()>.
777		988
778	This can be used to catch bugs inside libev itself: under normal	989	This can be used to catch bugs inside libev itself: under normal
…		…
782	=back	993	=back
783		994
784		995
785	=head1 ANATOMY OF A WATCHER	996	=head1 ANATOMY OF A WATCHER
786		997
		998	In the following description, uppercase C<TYPE> in names stands for the
		999	watcher type, e.g. C<ev_TYPE_start> can mean C<ev_timer_start> for timer
		1000	watchers and C<ev_io_start> for I/O watchers.
		1001
787	A watcher is a structure that you create and register to record your	1002	A watcher is an opaque structure that you allocate and register to record
788	interest in some event. For instance, if you want to wait for STDIN to	1003	your interest in some event. To make a concrete example, imagine you want
789	become readable, you would create an C<ev_io> watcher for that:	1004	to wait for STDIN to become readable, you would create an C<ev_io> watcher
		1005	for that:
790		1006
791	static void my_cb (struct ev_loop *loop, ev_io *w, int revents)	1007	static void my_cb (struct ev_loop *loop, ev_io *w, int revents)
792	{	1008	{
793	ev_io_stop (w);	1009	ev_io_stop (w);
794	ev_unloop (loop, EVUNLOOP_ALL);	1010	ev_break (loop, EVBREAK_ALL);
795	}	1011	}
796		1012
797	struct ev_loop *loop = ev_default_loop (0);	1013	struct ev_loop *loop = ev_default_loop (0);
		1014
798	ev_io stdin_watcher;	1015	ev_io stdin_watcher;
		1016
799	ev_init (&stdin_watcher, my_cb);	1017	ev_init (&stdin_watcher, my_cb);
800	ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);	1018	ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);
801	ev_io_start (loop, &stdin_watcher);	1019	ev_io_start (loop, &stdin_watcher);
		1020
802	ev_loop (loop, 0);	1021	ev_run (loop, 0);
803		1022
804	As you can see, you are responsible for allocating the memory for your	1023	As you can see, you are responsible for allocating the memory for your
805	watcher structures (and it is usually a bad idea to do this on the stack,	1024	watcher structures (and it is I<usually> a bad idea to do this on the
806	although this can sometimes be quite valid).	1025	stack).
807		1026
		1027	Each watcher has an associated watcher structure (called C<struct ev_TYPE>
		1028	or simply C<ev_TYPE>, as typedefs are provided for all watcher structs).
		1029
808	Each watcher structure must be initialised by a call to C<ev_init	1030	Each watcher structure must be initialised by a call to C<ev_init (watcher
809	(watcher *, callback)>, which expects a callback to be provided. This	1031	*, callback)>, which expects a callback to be provided. This callback is
810	callback gets invoked each time the event occurs (or, in the case of I/O	1032	invoked each time the event occurs (or, in the case of I/O watchers, each
811	watchers, each time the event loop detects that the file descriptor given	1033	time the event loop detects that the file descriptor given is readable
812	is readable and/or writable).	1034	and/or writable).
813		1035
814	Each watcher type has its own C<< ev_<type>_set (watcher *, ...) >> macro	1036	Each watcher type further has its own C<< ev_TYPE_set (watcher *, ...) >>
815	with arguments specific to this watcher type. There is also a macro	1037	macro to configure it, with arguments specific to the watcher type. There
816	to combine initialisation and setting in one call: C<< ev_<type>_init	1038	is also a macro to combine initialisation and setting in one call: C<<
817	(watcher *, callback, ...) >>.	1039	ev_TYPE_init (watcher *, callback, ...) >>.
818		1040
819	To make the watcher actually watch out for events, you have to start it	1041	To make the watcher actually watch out for events, you have to start it
820	with a watcher-specific start function (C<< ev_<type>_start (loop, watcher	1042	with a watcher-specific start function (C<< ev_TYPE_start (loop, watcher
821	*) >>), and you can stop watching for events at any time by calling the	1043	*) >>), and you can stop watching for events at any time by calling the
822	corresponding stop function (C<< ev_<type>_stop (loop, watcher *) >>.	1044	corresponding stop function (C<< ev_TYPE_stop (loop, watcher *) >>.
823		1045
824	As long as your watcher is active (has been started but not stopped) you	1046	As long as your watcher is active (has been started but not stopped) you
825	must not touch the values stored in it. Most specifically you must never	1047	must not touch the values stored in it. Most specifically you must never
826	reinitialise it or call its C<set> macro.	1048	reinitialise it or call its C<ev_TYPE_set> macro.
827		1049
828	Each and every callback receives the event loop pointer as first, the	1050	Each and every callback receives the event loop pointer as first, the
829	registered watcher structure as second, and a bitset of received events as	1051	registered watcher structure as second, and a bitset of received events as
830	third argument.	1052	third argument.
831		1053
…		…
840	=item C<EV_WRITE>	1062	=item C<EV_WRITE>
841		1063
842	The file descriptor in the C<ev_io> watcher has become readable and/or	1064	The file descriptor in the C<ev_io> watcher has become readable and/or
843	writable.	1065	writable.
844		1066
845	=item C<EV_TIMEOUT>	1067	=item C<EV_TIMER>
846		1068
847	The C<ev_timer> watcher has timed out.	1069	The C<ev_timer> watcher has timed out.
848		1070
849	=item C<EV_PERIODIC>	1071	=item C<EV_PERIODIC>
850		1072
…		…
868		1090
869	=item C<EV_PREPARE>	1091	=item C<EV_PREPARE>
870		1092
871	=item C<EV_CHECK>	1093	=item C<EV_CHECK>
872		1094
873	All C<ev_prepare> watchers are invoked just I<before> C<ev_loop> starts	1095	All C<ev_prepare> watchers are invoked just I<before> C<ev_run> starts
874	to gather new events, and all C<ev_check> watchers are invoked just after	1096	to gather new events, and all C<ev_check> watchers are invoked just after
875	C<ev_loop> has gathered them, but before it invokes any callbacks for any	1097	C<ev_run> has gathered them, but before it invokes any callbacks for any
876	received events. Callbacks of both watcher types can start and stop as	1098	received events. Callbacks of both watcher types can start and stop as
877	many watchers as they want, and all of them will be taken into account	1099	many watchers as they want, and all of them will be taken into account
878	(for example, a C<ev_prepare> watcher might start an idle watcher to keep	1100	(for example, a C<ev_prepare> watcher might start an idle watcher to keep
879	C<ev_loop> from blocking).	1101	C<ev_run> from blocking).
880		1102
881	=item C<EV_EMBED>	1103	=item C<EV_EMBED>
882		1104
883	The embedded event loop specified in the C<ev_embed> watcher needs attention.	1105	The embedded event loop specified in the C<ev_embed> watcher needs attention.
884		1106
…		…
888	C<ev_fork>).	1110	C<ev_fork>).
889		1111
890	=item C<EV_ASYNC>	1112	=item C<EV_ASYNC>
891		1113
892	The given async watcher has been asynchronously notified (see C<ev_async>).	1114	The given async watcher has been asynchronously notified (see C<ev_async>).
		1115
		1116	=item C<EV_CUSTOM>
		1117
		1118	Not ever sent (or otherwise used) by libev itself, but can be freely used
		1119	by libev users to signal watchers (e.g. via C<ev_feed_event>).
893		1120
894	=item C<EV_ERROR>	1121	=item C<EV_ERROR>
895		1122
896	An unspecified error has occurred, the watcher has been stopped. This might	1123	An unspecified error has occurred, the watcher has been stopped. This might
897	happen because the watcher could not be properly started because libev	1124	happen because the watcher could not be properly started because libev
…		…
910	programs, though, as the fd could already be closed and reused for another	1137	programs, though, as the fd could already be closed and reused for another
911	thing, so beware.	1138	thing, so beware.
912		1139
913	=back	1140	=back
914		1141
		1142	=head2 WATCHER STATES
		1143
		1144	There are various watcher states mentioned throughout this manual -
		1145	active, pending and so on. In this section these states and the rules to
		1146	transition between them will be described in more detail - and while these
		1147	rules might look complicated, they usually do "the right thing".
		1148
		1149	=over 4
		1150
		1151	=item initialiased
		1152
		1153	Before a watcher can be registered with the event looop it has to be
		1154	initialised. This can be done with a call to C<ev_TYPE_init>, or calls to
		1155	C<ev_init> followed by the watcher-specific C<ev_TYPE_set> function.
		1156
		1157	In this state it is simply some block of memory that is suitable for use
		1158	in an event loop. It can be moved around, freed, reused etc. at will.
		1159
		1160	=item started/running/active
		1161
		1162	Once a watcher has been started with a call to C<ev_TYPE_start> it becomes
		1163	property of the event loop, and is actively waiting for events. While in
		1164	this state it cannot be accessed (except in a few documented ways), moved,
		1165	freed or anything else - the only legal thing is to keep a pointer to it,
		1166	and call libev functions on it that are documented to work on active watchers.
		1167
		1168	=item pending
		1169
		1170	If a watcher is active and libev determines that an event it is interested
		1171	in has occurred (such as a timer expiring), it will become pending. It will
		1172	stay in this pending state until either it is stopped or its callback is
		1173	about to be invoked, so it is not normally pending inside the watcher
		1174	callback.
		1175
		1176	The watcher might or might not be active while it is pending (for example,
		1177	an expired non-repeating timer can be pending but no longer active). If it
		1178	is stopped, it can be freely accessed (e.g. by calling C<ev_TYPE_set>),
		1179	but it is still property of the event loop at this time, so cannot be
		1180	moved, freed or reused. And if it is active the rules described in the
		1181	previous item still apply.
		1182
		1183	It is also possible to feed an event on a watcher that is not active (e.g.
		1184	via C<ev_feed_event>), in which case it becomes pending without being
		1185	active.
		1186
		1187	=item stopped
		1188
		1189	A watcher can be stopped implicitly by libev (in which case it might still
		1190	be pending), or explicitly by calling its C<ev_TYPE_stop> function. The
		1191	latter will clear any pending state the watcher might be in, regardless
		1192	of whether it was active or not, so stopping a watcher explicitly before
		1193	freeing it is often a good idea.
		1194
		1195	While stopped (and not pending) the watcher is essentially in the
		1196	initialised state, that is it can be reused, moved, modified in any way
		1197	you wish.
		1198
		1199	=back
		1200
915	=head2 GENERIC WATCHER FUNCTIONS	1201	=head2 GENERIC WATCHER FUNCTIONS
916
917	In the following description, C<TYPE> stands for the watcher type,
918	e.g. C<timer> for C<ev_timer> watchers and C<io> for C<ev_io> watchers.
919		1202
920	=over 4	1203	=over 4
921		1204
922	=item C<ev_init> (ev_TYPE *watcher, callback)	1205	=item C<ev_init> (ev_TYPE *watcher, callback)
923		1206
…		…
938		1221
939	ev_io w;	1222	ev_io w;
940	ev_init (&w, my_cb);	1223	ev_init (&w, my_cb);
941	ev_io_set (&w, STDIN_FILENO, EV_READ);	1224	ev_io_set (&w, STDIN_FILENO, EV_READ);
942		1225
943	=item C<ev_TYPE_set> (ev_TYPE *, [args])	1226	=item C<ev_TYPE_set> (ev_TYPE *watcher, [args])
944		1227
945	This macro initialises the type-specific parts of a watcher. You need to	1228	This macro initialises the type-specific parts of a watcher. You need to
946	call C<ev_init> at least once before you call this macro, but you can	1229	call C<ev_init> at least once before you call this macro, but you can
947	call C<ev_TYPE_set> any number of times. You must not, however, call this	1230	call C<ev_TYPE_set> any number of times. You must not, however, call this
948	macro on a watcher that is active (it can be pending, however, which is a	1231	macro on a watcher that is active (it can be pending, however, which is a
…		…
961		1244
962	Example: Initialise and set an C<ev_io> watcher in one step.	1245	Example: Initialise and set an C<ev_io> watcher in one step.
963		1246
964	ev_io_init (&w, my_cb, STDIN_FILENO, EV_READ);	1247	ev_io_init (&w, my_cb, STDIN_FILENO, EV_READ);
965		1248
966	=item C<ev_TYPE_start> (loop *, ev_TYPE *watcher)	1249	=item C<ev_TYPE_start> (loop, ev_TYPE *watcher)
967		1250
968	Starts (activates) the given watcher. Only active watchers will receive	1251	Starts (activates) the given watcher. Only active watchers will receive
969	events. If the watcher is already active nothing will happen.	1252	events. If the watcher is already active nothing will happen.
970		1253
971	Example: Start the C<ev_io> watcher that is being abused as example in this	1254	Example: Start the C<ev_io> watcher that is being abused as example in this
972	whole section.	1255	whole section.
973		1256
974	ev_io_start (EV_DEFAULT_UC, &w);	1257	ev_io_start (EV_DEFAULT_UC, &w);
975		1258
976	=item C<ev_TYPE_stop> (loop *, ev_TYPE *watcher)	1259	=item C<ev_TYPE_stop> (loop, ev_TYPE *watcher)
977		1260
978	Stops the given watcher if active, and clears the pending status (whether	1261	Stops the given watcher if active, and clears the pending status (whether
979	the watcher was active or not).	1262	the watcher was active or not).
980		1263
981	It is possible that stopped watchers are pending - for example,	1264	It is possible that stopped watchers are pending - for example,
…		…
1006	=item ev_cb_set (ev_TYPE *watcher, callback)	1289	=item ev_cb_set (ev_TYPE *watcher, callback)
1007		1290
1008	Change the callback. You can change the callback at virtually any time	1291	Change the callback. You can change the callback at virtually any time
1009	(modulo threads).	1292	(modulo threads).
1010		1293
1011	=item ev_set_priority (ev_TYPE *watcher, priority)	1294	=item ev_set_priority (ev_TYPE *watcher, int priority)
1012		1295
1013	=item int ev_priority (ev_TYPE *watcher)	1296	=item int ev_priority (ev_TYPE *watcher)
1014		1297
1015	Set and query the priority of the watcher. The priority is a small	1298	Set and query the priority of the watcher. The priority is a small
1016	integer between C<EV_MAXPRI> (default: C<2>) and C<EV_MINPRI>	1299	integer between C<EV_MAXPRI> (default: C<2>) and C<EV_MINPRI>
1017	(default: C<-2>). Pending watchers with higher priority will be invoked	1300	(default: C<-2>). Pending watchers with higher priority will be invoked
1018	before watchers with lower priority, but priority will not keep watchers	1301	before watchers with lower priority, but priority will not keep watchers
1019	from being executed (except for C<ev_idle> watchers).	1302	from being executed (except for C<ev_idle> watchers).
1020		1303
1021	This means that priorities are I<only> used for ordering callback
1022	invocation after new events have been received. This is useful, for
1023	example, to reduce latency after idling, or more often, to bind two
1024	watchers on the same event and make sure one is called first.
1025
1026	If you need to suppress invocation when higher priority events are pending	1304	If you need to suppress invocation when higher priority events are pending
1027	you need to look at C<ev_idle> watchers, which provide this functionality.	1305	you need to look at C<ev_idle> watchers, which provide this functionality.
1028		1306
1029	You I<must not> change the priority of a watcher as long as it is active or	1307	You I<must not> change the priority of a watcher as long as it is active or
1030	pending.	1308	pending.
1031		1309
		1310	Setting a priority outside the range of C<EV_MINPRI> to C<EV_MAXPRI> is
		1311	fine, as long as you do not mind that the priority value you query might
		1312	or might not have been clamped to the valid range.
		1313
1032	The default priority used by watchers when no priority has been set is	1314	The default priority used by watchers when no priority has been set is
1033	always C<0>, which is supposed to not be too high and not be too low :).	1315	always C<0>, which is supposed to not be too high and not be too low :).
1034		1316
1035	Setting a priority outside the range of C<EV_MINPRI> to C<EV_MAXPRI> is	1317	See L<WATCHER PRIORITY MODELS>, below, for a more thorough treatment of
1036	fine, as long as you do not mind that the priority value you query might	1318	priorities.
1037	or might not have been adjusted to be within valid range.
1038		1319
1039	=item ev_invoke (loop, ev_TYPE *watcher, int revents)	1320	=item ev_invoke (loop, ev_TYPE *watcher, int revents)
1040		1321
1041	Invoke the C<watcher> with the given C<loop> and C<revents>. Neither	1322	Invoke the C<watcher> with the given C<loop> and C<revents>. Neither
1042	C<loop> nor C<revents> need to be valid as long as the watcher callback	1323	C<loop> nor C<revents> need to be valid as long as the watcher callback
…		…
1049	returns its C<revents> bitset (as if its callback was invoked). If the	1330	returns its C<revents> bitset (as if its callback was invoked). If the
1050	watcher isn't pending it does nothing and returns C<0>.	1331	watcher isn't pending it does nothing and returns C<0>.
1051		1332
1052	Sometimes it can be useful to "poll" a watcher instead of waiting for its	1333	Sometimes it can be useful to "poll" a watcher instead of waiting for its
1053	callback to be invoked, which can be accomplished with this function.	1334	callback to be invoked, which can be accomplished with this function.
		1335
		1336	=item ev_feed_event (loop, ev_TYPE *watcher, int revents)
		1337
		1338	Feeds the given event set into the event loop, as if the specified event
		1339	had happened for the specified watcher (which must be a pointer to an
		1340	initialised but not necessarily started event watcher). Obviously you must
		1341	not free the watcher as long as it has pending events.
		1342
		1343	Stopping the watcher, letting libev invoke it, or calling
		1344	C<ev_clear_pending> will clear the pending event, even if the watcher was
		1345	not started in the first place.
		1346
		1347	See also C<ev_feed_fd_event> and C<ev_feed_signal_event> for related
		1348	functions that do not need a watcher.
1054		1349
1055	=back	1350	=back
1056		1351
1057		1352
1058	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER	1353	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER
…		…
1107	#include <stddef.h>	1402	#include <stddef.h>
1108		1403
1109	static void	1404	static void
1110	t1_cb (EV_P_ ev_timer *w, int revents)	1405	t1_cb (EV_P_ ev_timer *w, int revents)
1111	{	1406	{
1112	struct my_biggy big = (struct my_biggy *	1407	struct my_biggy big = (struct my_biggy *)
1113	(((char *)w) - offsetof (struct my_biggy, t1));	1408	(((char *)w) - offsetof (struct my_biggy, t1));
1114	}	1409	}
1115		1410
1116	static void	1411	static void
1117	t2_cb (EV_P_ ev_timer *w, int revents)	1412	t2_cb (EV_P_ ev_timer *w, int revents)
1118	{	1413	{
1119	struct my_biggy big = (struct my_biggy *	1414	struct my_biggy big = (struct my_biggy *)
1120	(((char *)w) - offsetof (struct my_biggy, t2));	1415	(((char *)w) - offsetof (struct my_biggy, t2));
1121	}	1416	}
		1417
		1418	=head2 WATCHER PRIORITY MODELS
		1419
		1420	Many event loops support I<watcher priorities>, which are usually small
		1421	integers that influence the ordering of event callback invocation
		1422	between watchers in some way, all else being equal.
		1423
		1424	In libev, Watcher priorities can be set using C<ev_set_priority>. See its
		1425	description for the more technical details such as the actual priority
		1426	range.
		1427
		1428	There are two common ways how these these priorities are being interpreted
		1429	by event loops:
		1430
		1431	In the more common lock-out model, higher priorities "lock out" invocation
		1432	of lower priority watchers, which means as long as higher priority
		1433	watchers receive events, lower priority watchers are not being invoked.
		1434
		1435	The less common only-for-ordering model uses priorities solely to order
		1436	callback invocation within a single event loop iteration: Higher priority
		1437	watchers are invoked before lower priority ones, but they all get invoked
		1438	before polling for new events.
		1439
		1440	Libev uses the second (only-for-ordering) model for all its watchers
		1441	except for idle watchers (which use the lock-out model).
		1442
		1443	The rationale behind this is that implementing the lock-out model for
		1444	watchers is not well supported by most kernel interfaces, and most event
		1445	libraries will just poll for the same events again and again as long as
		1446	their callbacks have not been executed, which is very inefficient in the
		1447	common case of one high-priority watcher locking out a mass of lower
		1448	priority ones.
		1449
		1450	Static (ordering) priorities are most useful when you have two or more
		1451	watchers handling the same resource: a typical usage example is having an
		1452	C<ev_io> watcher to receive data, and an associated C<ev_timer> to handle
		1453	timeouts. Under load, data might be received while the program handles
		1454	other jobs, but since timers normally get invoked first, the timeout
		1455	handler will be executed before checking for data. In that case, giving
		1456	the timer a lower priority than the I/O watcher ensures that I/O will be
		1457	handled first even under adverse conditions (which is usually, but not
		1458	always, what you want).
		1459
		1460	Since idle watchers use the "lock-out" model, meaning that idle watchers
		1461	will only be executed when no same or higher priority watchers have
		1462	received events, they can be used to implement the "lock-out" model when
		1463	required.
		1464
		1465	For example, to emulate how many other event libraries handle priorities,
		1466	you can associate an C<ev_idle> watcher to each such watcher, and in
		1467	the normal watcher callback, you just start the idle watcher. The real
		1468	processing is done in the idle watcher callback. This causes libev to
		1469	continuously poll and process kernel event data for the watcher, but when
		1470	the lock-out case is known to be rare (which in turn is rare :), this is
		1471	workable.
		1472
		1473	Usually, however, the lock-out model implemented that way will perform
		1474	miserably under the type of load it was designed to handle. In that case,
		1475	it might be preferable to stop the real watcher before starting the
		1476	idle watcher, so the kernel will not have to process the event in case
		1477	the actual processing will be delayed for considerable time.
		1478
		1479	Here is an example of an I/O watcher that should run at a strictly lower
		1480	priority than the default, and which should only process data when no
		1481	other events are pending:
		1482
		1483	ev_idle idle; // actual processing watcher
		1484	ev_io io; // actual event watcher
		1485
		1486	static void
		1487	io_cb (EV_P_ ev_io *w, int revents)
		1488	{
		1489	// stop the I/O watcher, we received the event, but
		1490	// are not yet ready to handle it.
		1491	ev_io_stop (EV_A_ w);
		1492
		1493	// start the idle watcher to handle the actual event.
		1494	// it will not be executed as long as other watchers
		1495	// with the default priority are receiving events.
		1496	ev_idle_start (EV_A_ &idle);
		1497	}
		1498
		1499	static void
		1500	idle_cb (EV_P_ ev_idle *w, int revents)
		1501	{
		1502	// actual processing
		1503	read (STDIN_FILENO, ...);
		1504
		1505	// have to start the I/O watcher again, as
		1506	// we have handled the event
		1507	ev_io_start (EV_P_ &io);
		1508	}
		1509
		1510	// initialisation
		1511	ev_idle_init (&idle, idle_cb);
		1512	ev_io_init (&io, io_cb, STDIN_FILENO, EV_READ);
		1513	ev_io_start (EV_DEFAULT_ &io);
		1514
		1515	In the "real" world, it might also be beneficial to start a timer, so that
		1516	low-priority connections can not be locked out forever under load. This
		1517	enables your program to keep a lower latency for important connections
		1518	during short periods of high load, while not completely locking out less
		1519	important ones.
1122		1520
1123		1521
1124	=head1 WATCHER TYPES	1522	=head1 WATCHER TYPES
1125		1523
1126	This section describes each watcher in detail, but will not repeat	1524	This section describes each watcher in detail, but will not repeat
…		…
1152	descriptors to non-blocking mode is also usually a good idea (but not	1550	descriptors to non-blocking mode is also usually a good idea (but not
1153	required if you know what you are doing).	1551	required if you know what you are doing).
1154		1552
1155	If you cannot use non-blocking mode, then force the use of a	1553	If you cannot use non-blocking mode, then force the use of a
1156	known-to-be-good backend (at the time of this writing, this includes only	1554	known-to-be-good backend (at the time of this writing, this includes only
1157	C<EVBACKEND_SELECT> and C<EVBACKEND_POLL>).	1555	C<EVBACKEND_SELECT> and C<EVBACKEND_POLL>). The same applies to file
		1556	descriptors for which non-blocking operation makes no sense (such as
		1557	files) - libev doesn't guarantee any specific behaviour in that case.
1158		1558
1159	Another thing you have to watch out for is that it is quite easy to	1559	Another thing you have to watch out for is that it is quite easy to
1160	receive "spurious" readiness notifications, that is your callback might	1560	receive "spurious" readiness notifications, that is your callback might
1161	be called with C<EV_READ> but a subsequent C<read>(2) will actually block	1561	be called with C<EV_READ> but a subsequent C<read>(2) will actually block
1162	because there is no data. Not only are some backends known to create a	1562	because there is no data. Not only are some backends known to create a
…		…
1227		1627
1228	So when you encounter spurious, unexplained daemon exits, make sure you	1628	So when you encounter spurious, unexplained daemon exits, make sure you
1229	ignore SIGPIPE (and maybe make sure you log the exit status of your daemon	1629	ignore SIGPIPE (and maybe make sure you log the exit status of your daemon
1230	somewhere, as that would have given you a big clue).	1630	somewhere, as that would have given you a big clue).
1231		1631
		1632	=head3 The special problem of accept()ing when you can't
		1633
		1634	Many implementations of the POSIX C<accept> function (for example,
		1635	found in post-2004 Linux) have the peculiar behaviour of not removing a
		1636	connection from the pending queue in all error cases.
		1637
		1638	For example, larger servers often run out of file descriptors (because
		1639	of resource limits), causing C<accept> to fail with C<ENFILE> but not
		1640	rejecting the connection, leading to libev signalling readiness on
		1641	the next iteration again (the connection still exists after all), and
		1642	typically causing the program to loop at 100% CPU usage.
		1643
		1644	Unfortunately, the set of errors that cause this issue differs between
		1645	operating systems, there is usually little the app can do to remedy the
		1646	situation, and no known thread-safe method of removing the connection to
		1647	cope with overload is known (to me).
		1648
		1649	One of the easiest ways to handle this situation is to just ignore it
		1650	- when the program encounters an overload, it will just loop until the
		1651	situation is over. While this is a form of busy waiting, no OS offers an
		1652	event-based way to handle this situation, so it's the best one can do.
		1653
		1654	A better way to handle the situation is to log any errors other than
		1655	C<EAGAIN> and C<EWOULDBLOCK>, making sure not to flood the log with such
		1656	messages, and continue as usual, which at least gives the user an idea of
		1657	what could be wrong ("raise the ulimit!"). For extra points one could stop
		1658	the C<ev_io> watcher on the listening fd "for a while", which reduces CPU
		1659	usage.
		1660
		1661	If your program is single-threaded, then you could also keep a dummy file
		1662	descriptor for overload situations (e.g. by opening F</dev/null>), and
		1663	when you run into C<ENFILE> or C<EMFILE>, close it, run C<accept>,
		1664	close that fd, and create a new dummy fd. This will gracefully refuse
		1665	clients under typical overload conditions.
		1666
		1667	The last way to handle it is to simply log the error and C<exit>, as
		1668	is often done with C<malloc> failures, but this results in an easy
		1669	opportunity for a DoS attack.
1232		1670
1233	=head3 Watcher-Specific Functions	1671	=head3 Watcher-Specific Functions
1234		1672
1235	=over 4	1673	=over 4
1236		1674
…		…
1268	...	1706	...
1269	struct ev_loop *loop = ev_default_init (0);	1707	struct ev_loop *loop = ev_default_init (0);
1270	ev_io stdin_readable;	1708	ev_io stdin_readable;
1271	ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);	1709	ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
1272	ev_io_start (loop, &stdin_readable);	1710	ev_io_start (loop, &stdin_readable);
1273	ev_loop (loop, 0);	1711	ev_run (loop, 0);
1274		1712
1275		1713
1276	=head2 C<ev_timer> - relative and optionally repeating timeouts	1714	=head2 C<ev_timer> - relative and optionally repeating timeouts
1277		1715
1278	Timer watchers are simple relative timers that generate an event after a	1716	Timer watchers are simple relative timers that generate an event after a
…		…
1283	year, it will still time out after (roughly) one hour. "Roughly" because	1721	year, it will still time out after (roughly) one hour. "Roughly" because
1284	detecting time jumps is hard, and some inaccuracies are unavoidable (the	1722	detecting time jumps is hard, and some inaccuracies are unavoidable (the
1285	monotonic clock option helps a lot here).	1723	monotonic clock option helps a lot here).
1286		1724
1287	The callback is guaranteed to be invoked only I<after> its timeout has	1725	The callback is guaranteed to be invoked only I<after> its timeout has
1288	passed, but if multiple timers become ready during the same loop iteration	1726	passed (not I<at>, so on systems with very low-resolution clocks this
1289	then order of execution is undefined.	1727	might introduce a small delay). If multiple timers become ready during the
		1728	same loop iteration then the ones with earlier time-out values are invoked
		1729	before ones of the same priority with later time-out values (but this is
		1730	no longer true when a callback calls C<ev_run> recursively).
1290		1731
1291	=head3 Be smart about timeouts	1732	=head3 Be smart about timeouts
1292		1733
1293	Many real-world problems invole some kind of time-out, usually for error	1734	Many real-world problems involve some kind of timeout, usually for error
1294	recovery. A typical example is an HTTP request - if the other side hangs,	1735	recovery. A typical example is an HTTP request - if the other side hangs,
1295	you want to raise some error after a while.	1736	you want to raise some error after a while.
1296		1737
1297	Here are some ways on how to handle this problem, from simple and	1738	What follows are some ways to handle this problem, from obvious and
1298	inefficient to very efficient.	1739	inefficient to smart and efficient.
1299		1740
1300	In the following examples a 60 second activity timeout is assumed - a	1741	In the following, a 60 second activity timeout is assumed - a timeout that
1301	timeout that gets reset to 60 seconds each time some data ("a lifesign")	1742	gets reset to 60 seconds each time there is activity (e.g. each time some
1302	was received.	1743	data or other life sign was received).
1303		1744
1304	=over 4	1745	=over 4
1305		1746
1306	=item 1. Use a timer and stop, reinitialise, start it on activity.	1747	=item 1. Use a timer and stop, reinitialise and start it on activity.
1307		1748
1308	This is the most obvious, but not the most simple way: In the beginning,	1749	This is the most obvious, but not the most simple way: In the beginning,
1309	start the watcher:	1750	start the watcher:
1310		1751
1311	ev_timer_init (timer, callback, 60., 0.);	1752	ev_timer_init (timer, callback, 60., 0.);
1312	ev_timer_start (loop, timer);	1753	ev_timer_start (loop, timer);
1313		1754
1314	Then, each time there is some activity, C<ev_timer_stop> the timer,	1755	Then, each time there is some activity, C<ev_timer_stop> it, initialise it
1315	initialise it again, and start it:	1756	and start it again:
1316		1757
1317	ev_timer_stop (loop, timer);	1758	ev_timer_stop (loop, timer);
1318	ev_timer_set (timer, 60., 0.);	1759	ev_timer_set (timer, 60., 0.);
1319	ev_timer_start (loop, timer);	1760	ev_timer_start (loop, timer);
1320		1761
1321	This is relatively simple to implement, but means that each time there	1762	This is relatively simple to implement, but means that each time there is
1322	is some activity, libev will first have to remove the timer from it's	1763	some activity, libev will first have to remove the timer from its internal
1323	internal data strcuture and then add it again.	1764	data structure and then add it again. Libev tries to be fast, but it's
		1765	still not a constant-time operation.
1324		1766
1325	=item 2. Use a timer and re-start it with C<ev_timer_again> inactivity.	1767	=item 2. Use a timer and re-start it with C<ev_timer_again> inactivity.
1326		1768
1327	This is the easiest way, and involves using C<ev_timer_again> instead of	1769	This is the easiest way, and involves using C<ev_timer_again> instead of
1328	C<ev_timer_start>.	1770	C<ev_timer_start>.
1329		1771
1330	For this, configure an C<ev_timer> with a C<repeat> value of C<60> and	1772	To implement this, configure an C<ev_timer> with a C<repeat> value
1331	then call C<ev_timer_again> at start and each time you successfully read	1773	of C<60> and then call C<ev_timer_again> at start and each time you
1332	or write some data. If you go into an idle state where you do not expect	1774	successfully read or write some data. If you go into an idle state where
1333	data to travel on the socket, you can C<ev_timer_stop> the timer, and	1775	you do not expect data to travel on the socket, you can C<ev_timer_stop>
1334	C<ev_timer_again> will automatically restart it if need be.	1776	the timer, and C<ev_timer_again> will automatically restart it if need be.
1335		1777
1336	That means you can ignore the C<after> value and C<ev_timer_start>	1778	That means you can ignore both the C<ev_timer_start> function and the
1337	altogether and only ever use the C<repeat> value and C<ev_timer_again>.	1779	C<after> argument to C<ev_timer_set>, and only ever use the C<repeat>
		1780	member and C<ev_timer_again>.
1338		1781
1339	At start:	1782	At start:
1340		1783
1341	ev_timer_init (timer, callback, 0., 60.);	1784	ev_init (timer, callback);
		1785	timer->repeat = 60.;
1342	ev_timer_again (loop, timer);	1786	ev_timer_again (loop, timer);
1343		1787
1344	Each time you receive some data:	1788	Each time there is some activity:
1345		1789
1346	ev_timer_again (loop, timer);	1790	ev_timer_again (loop, timer);
1347		1791
1348	It is even possible to change the time-out on the fly:	1792	It is even possible to change the time-out on the fly, regardless of
		1793	whether the watcher is active or not:
1349		1794
1350	timer->repeat = 30.;	1795	timer->repeat = 30.;
1351	ev_timer_again (loop, timer);	1796	ev_timer_again (loop, timer);
1352		1797
1353	This is slightly more efficient then stopping/starting the timer each time	1798	This is slightly more efficient then stopping/starting the timer each time
1354	you want to modify its timeout value, as libev does not have to completely	1799	you want to modify its timeout value, as libev does not have to completely
1355	remove and re-insert the timer from/into it's internal data structure.	1800	remove and re-insert the timer from/into its internal data structure.
		1801
		1802	It is, however, even simpler than the "obvious" way to do it.
1356		1803
1357	=item 3. Let the timer time out, but then re-arm it as required.	1804	=item 3. Let the timer time out, but then re-arm it as required.
1358		1805
1359	This method is more tricky, but usually most efficient: Most timeouts are	1806	This method is more tricky, but usually most efficient: Most timeouts are
1360	relatively long compared to the loop iteration time - in our example,	1807	relatively long compared to the intervals between other activity - in
1361	within 60 seconds, there are usually many I/O events with associated	1808	our example, within 60 seconds, there are usually many I/O events with
1362	activity resets.	1809	associated activity resets.
1363		1810
1364	In this case, it would be more efficient to leave the C<ev_timer> alone,	1811	In this case, it would be more efficient to leave the C<ev_timer> alone,
1365	but remember the time of last activity, and check for a real timeout only	1812	but remember the time of last activity, and check for a real timeout only
1366	within the callback:	1813	within the callback:
1367		1814
1368	ev_tstamp last_activity; // time of last activity	1815	ev_tstamp last_activity; // time of last activity
1369		1816
1370	static void	1817	static void
1371	callback (EV_P_ ev_timer *w, int revents)	1818	callback (EV_P_ ev_timer *w, int revents)
1372	{	1819	{
1373	ev_tstamp now = ev_now (EV_A);	1820	ev_tstamp now = ev_now (EV_A);
1374	ev_tstamp timeout = last_activity + 60.;	1821	ev_tstamp timeout = last_activity + 60.;
1375		1822
1376	// if last_activity is older than now - timeout, we did time out	1823	// if last_activity + 60. is older than now, we did time out
1377	if (timeout < now)	1824	if (timeout < now)
1378	{	1825	{
1379	// timeout occured, take action	1826	// timeout occurred, take action
1380	}	1827	}
1381	else	1828	else
1382	{	1829	{
1383	// callback was invoked, but there was some activity, re-arm	1830	// callback was invoked, but there was some activity, re-arm
1384	// to fire in last_activity + 60.	1831	// the watcher to fire in last_activity + 60, which is
		1832	// guaranteed to be in the future, so "again" is positive:
1385	w->again = timeout - now;	1833	w->repeat = timeout - now;
1386	ev_timer_again (EV_A_ w);	1834	ev_timer_again (EV_A_ w);
1387	}	1835	}
1388	}	1836	}
1389		1837
1390	To summarise the callback: first calculate the real time-out (defined as	1838	To summarise the callback: first calculate the real timeout (defined
1391	"60 seconds after the last activity"), then check if that time has been	1839	as "60 seconds after the last activity"), then check if that time has
1392	reached, which means there was a real timeout. Otherwise the callback was	1840	been reached, which means something I<did>, in fact, time out. Otherwise
1393	invoked too early (timeout is in the future), so re-schedule the timer to	1841	the callback was invoked too early (C<timeout> is in the future), so
1394	fire at that future time.	1842	re-schedule the timer to fire at that future time, to see if maybe we have
		1843	a timeout then.
1395		1844
1396	Note how C<ev_timer_again> is used, taking advantage of the	1845	Note how C<ev_timer_again> is used, taking advantage of the
1397	C<ev_timer_again> optimisation when the timer is already running.	1846	C<ev_timer_again> optimisation when the timer is already running.
1398		1847
1399	This scheme causes more callback invocations (about one every 60 seconds),	1848	This scheme causes more callback invocations (about one every 60 seconds
1400	but virtually no calls to libev to change the timeout.	1849	minus half the average time between activity), but virtually no calls to
		1850	libev to change the timeout.
1401		1851
1402	To start the timer, simply intiialise the watcher and C<last_activity>,	1852	To start the timer, simply initialise the watcher and set C<last_activity>
1403	then call the callback:	1853	to the current time (meaning we just have some activity :), then call the
		1854	callback, which will "do the right thing" and start the timer:
1404		1855
1405	ev_timer_init (timer, callback);	1856	ev_init (timer, callback);
1406	last_activity = ev_now (loop);	1857	last_activity = ev_now (loop);
1407	callback (loop, timer, EV_TIMEOUT);	1858	callback (loop, timer, EV_TIMER);
1408		1859
1409	And when there is some activity, simply remember the time in	1860	And when there is some activity, simply store the current time in
1410	C<last_activity>:	1861	C<last_activity>, no libev calls at all:
1411		1862
1412	last_actiivty = ev_now (loop);	1863	last_activity = ev_now (loop);
1413		1864
1414	This technique is slightly more complex, but in most cases where the	1865	This technique is slightly more complex, but in most cases where the
1415	time-out is unlikely to be triggered, much more efficient.	1866	time-out is unlikely to be triggered, much more efficient.
1416		1867
		1868	Changing the timeout is trivial as well (if it isn't hard-coded in the
		1869	callback :) - just change the timeout and invoke the callback, which will
		1870	fix things for you.
		1871
		1872	=item 4. Wee, just use a double-linked list for your timeouts.
		1873
		1874	If there is not one request, but many thousands (millions...), all
		1875	employing some kind of timeout with the same timeout value, then one can
		1876	do even better:
		1877
		1878	When starting the timeout, calculate the timeout value and put the timeout
		1879	at the I<end> of the list.
		1880
		1881	Then use an C<ev_timer> to fire when the timeout at the I<beginning> of
		1882	the list is expected to fire (for example, using the technique #3).
		1883
		1884	When there is some activity, remove the timer from the list, recalculate
		1885	the timeout, append it to the end of the list again, and make sure to
		1886	update the C<ev_timer> if it was taken from the beginning of the list.
		1887
		1888	This way, one can manage an unlimited number of timeouts in O(1) time for
		1889	starting, stopping and updating the timers, at the expense of a major
		1890	complication, and having to use a constant timeout. The constant timeout
		1891	ensures that the list stays sorted.
		1892
1417	=back	1893	=back
		1894
		1895	So which method the best?
		1896
		1897	Method #2 is a simple no-brain-required solution that is adequate in most
		1898	situations. Method #3 requires a bit more thinking, but handles many cases
		1899	better, and isn't very complicated either. In most case, choosing either
		1900	one is fine, with #3 being better in typical situations.
		1901
		1902	Method #1 is almost always a bad idea, and buys you nothing. Method #4 is
		1903	rather complicated, but extremely efficient, something that really pays
		1904	off after the first million or so of active timers, i.e. it's usually
		1905	overkill :)
1418		1906
1419	=head3 The special problem of time updates	1907	=head3 The special problem of time updates
1420		1908
1421	Establishing the current time is a costly operation (it usually takes at	1909	Establishing the current time is a costly operation (it usually takes at
1422	least two system calls): EV therefore updates its idea of the current	1910	least two system calls): EV therefore updates its idea of the current
1423	time only before and after C<ev_loop> collects new events, which causes a	1911	time only before and after C<ev_run> collects new events, which causes a
1424	growing difference between C<ev_now ()> and C<ev_time ()> when handling	1912	growing difference between C<ev_now ()> and C<ev_time ()> when handling
1425	lots of events in one iteration.	1913	lots of events in one iteration.
1426		1914
1427	The relative timeouts are calculated relative to the C<ev_now ()>	1915	The relative timeouts are calculated relative to the C<ev_now ()>
1428	time. This is usually the right thing as this timestamp refers to the time	1916	time. This is usually the right thing as this timestamp refers to the time
…		…
1434		1922
1435	If the event loop is suspended for a long time, you can also force an	1923	If the event loop is suspended for a long time, you can also force an
1436	update of the time returned by C<ev_now ()> by calling C<ev_now_update	1924	update of the time returned by C<ev_now ()> by calling C<ev_now_update
1437	()>.	1925	()>.
1438		1926
		1927	=head3 The special problems of suspended animation
		1928
		1929	When you leave the server world it is quite customary to hit machines that
		1930	can suspend/hibernate - what happens to the clocks during such a suspend?
		1931
		1932	Some quick tests made with a Linux 2.6.28 indicate that a suspend freezes
		1933	all processes, while the clocks (C<times>, C<CLOCK_MONOTONIC>) continue
		1934	to run until the system is suspended, but they will not advance while the
		1935	system is suspended. That means, on resume, it will be as if the program
		1936	was frozen for a few seconds, but the suspend time will not be counted
		1937	towards C<ev_timer> when a monotonic clock source is used. The real time
		1938	clock advanced as expected, but if it is used as sole clocksource, then a
		1939	long suspend would be detected as a time jump by libev, and timers would
		1940	be adjusted accordingly.
		1941
		1942	I would not be surprised to see different behaviour in different between
		1943	operating systems, OS versions or even different hardware.
		1944
		1945	The other form of suspend (job control, or sending a SIGSTOP) will see a
		1946	time jump in the monotonic clocks and the realtime clock. If the program
		1947	is suspended for a very long time, and monotonic clock sources are in use,
		1948	then you can expect C<ev_timer>s to expire as the full suspension time
		1949	will be counted towards the timers. When no monotonic clock source is in
		1950	use, then libev will again assume a timejump and adjust accordingly.
		1951
		1952	It might be beneficial for this latter case to call C<ev_suspend>
		1953	and C<ev_resume> in code that handles C<SIGTSTP>, to at least get
		1954	deterministic behaviour in this case (you can do nothing against
		1955	C<SIGSTOP>).
		1956
1439	=head3 Watcher-Specific Functions and Data Members	1957	=head3 Watcher-Specific Functions and Data Members
1440		1958
1441	=over 4	1959	=over 4
1442		1960
1443	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)	1961	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)
…		…
1466	If the timer is started but non-repeating, stop it (as if it timed out).	1984	If the timer is started but non-repeating, stop it (as if it timed out).
1467		1985
1468	If the timer is repeating, either start it if necessary (with the	1986	If the timer is repeating, either start it if necessary (with the
1469	C<repeat> value), or reset the running timer to the C<repeat> value.	1987	C<repeat> value), or reset the running timer to the C<repeat> value.
1470		1988
1471	This sounds a bit complicated, see "Be smart about timeouts", above, for a	1989	This sounds a bit complicated, see L<Be smart about timeouts>, above, for a
1472	usage example.	1990	usage example.
		1991
		1992	=item ev_tstamp ev_timer_remaining (loop, ev_timer *)
		1993
		1994	Returns the remaining time until a timer fires. If the timer is active,
		1995	then this time is relative to the current event loop time, otherwise it's
		1996	the timeout value currently configured.
		1997
		1998	That is, after an C<ev_timer_set (w, 5, 7)>, C<ev_timer_remaining> returns
		1999	C<5>. When the timer is started and one second passes, C<ev_timer_remaining>
		2000	will return C<4>. When the timer expires and is restarted, it will return
		2001	roughly C<7> (likely slightly less as callback invocation takes some time,
		2002	too), and so on.
1473		2003
1474	=item ev_tstamp repeat [read-write]	2004	=item ev_tstamp repeat [read-write]
1475		2005
1476	The current C<repeat> value. Will be used each time the watcher times out	2006	The current C<repeat> value. Will be used each time the watcher times out
1477	or C<ev_timer_again> is called, and determines the next timeout (if any),	2007	or C<ev_timer_again> is called, and determines the next timeout (if any),
…		…
1503	}	2033	}
1504		2034
1505	ev_timer mytimer;	2035	ev_timer mytimer;
1506	ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */	2036	ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */
1507	ev_timer_again (&mytimer); /* start timer */	2037	ev_timer_again (&mytimer); /* start timer */
1508	ev_loop (loop, 0);	2038	ev_run (loop, 0);
1509		2039
1510	// and in some piece of code that gets executed on any "activity":	2040	// and in some piece of code that gets executed on any "activity":
1511	// reset the timeout to start ticking again at 10 seconds	2041	// reset the timeout to start ticking again at 10 seconds
1512	ev_timer_again (&mytimer);	2042	ev_timer_again (&mytimer);
1513		2043
…		…
1515	=head2 C<ev_periodic> - to cron or not to cron?	2045	=head2 C<ev_periodic> - to cron or not to cron?
1516		2046
1517	Periodic watchers are also timers of a kind, but they are very versatile	2047	Periodic watchers are also timers of a kind, but they are very versatile
1518	(and unfortunately a bit complex).	2048	(and unfortunately a bit complex).
1519		2049
1520	Unlike C<ev_timer>'s, they are not based on real time (or relative time)	2050	Unlike C<ev_timer>, periodic watchers are not based on real time (or
1521	but on wall clock time (absolute time). You can tell a periodic watcher	2051	relative time, the physical time that passes) but on wall clock time
1522	to trigger after some specific point in time. For example, if you tell a	2052	(absolute time, the thing you can read on your calender or clock). The
1523	periodic watcher to trigger in 10 seconds (by specifying e.g. C<ev_now ()	2053	difference is that wall clock time can run faster or slower than real
1524	+ 10.>, that is, an absolute time not a delay) and then reset your system	2054	time, and time jumps are not uncommon (e.g. when you adjust your
1525	clock to January of the previous year, then it will take more than year	2055	wrist-watch).
1526	to trigger the event (unlike an C<ev_timer>, which would still trigger
1527	roughly 10 seconds later as it uses a relative timeout).
1528		2056
		2057	You can tell a periodic watcher to trigger after some specific point
		2058	in time: for example, if you tell a periodic watcher to trigger "in 10
		2059	seconds" (by specifying e.g. C<ev_now () + 10.>, that is, an absolute time
		2060	not a delay) and then reset your system clock to January of the previous
		2061	year, then it will take a year or more to trigger the event (unlike an
		2062	C<ev_timer>, which would still trigger roughly 10 seconds after starting
		2063	it, as it uses a relative timeout).
		2064
1529	C<ev_periodic>s can also be used to implement vastly more complex timers,	2065	C<ev_periodic> watchers can also be used to implement vastly more complex
1530	such as triggering an event on each "midnight, local time", or other	2066	timers, such as triggering an event on each "midnight, local time", or
1531	complicated rules.	2067	other complicated rules. This cannot be done with C<ev_timer> watchers, as
		2068	those cannot react to time jumps.
1532		2069
1533	As with timers, the callback is guaranteed to be invoked only when the	2070	As with timers, the callback is guaranteed to be invoked only when the
1534	time (C<at>) has passed, but if multiple periodic timers become ready	2071	point in time where it is supposed to trigger has passed. If multiple
1535	during the same loop iteration, then order of execution is undefined.	2072	timers become ready during the same loop iteration then the ones with
		2073	earlier time-out values are invoked before ones with later time-out values
		2074	(but this is no longer true when a callback calls C<ev_run> recursively).
1536		2075
1537	=head3 Watcher-Specific Functions and Data Members	2076	=head3 Watcher-Specific Functions and Data Members
1538		2077
1539	=over 4	2078	=over 4
1540		2079
1541	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)	2080	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp offset, ev_tstamp interval, reschedule_cb)
1542		2081
1543	=item ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)	2082	=item ev_periodic_set (ev_periodic *, ev_tstamp offset, ev_tstamp interval, reschedule_cb)
1544		2083
1545	Lots of arguments, lets sort it out... There are basically three modes of	2084	Lots of arguments, let's sort it out... There are basically three modes of
1546	operation, and we will explain them from simplest to most complex:	2085	operation, and we will explain them from simplest to most complex:
1547		2086
1548	=over 4	2087	=over 4
1549		2088
1550	=item * absolute timer (at = time, interval = reschedule_cb = 0)	2089	=item * absolute timer (offset = absolute time, interval = 0, reschedule_cb = 0)
1551		2090
1552	In this configuration the watcher triggers an event after the wall clock	2091	In this configuration the watcher triggers an event after the wall clock
1553	time C<at> has passed. It will not repeat and will not adjust when a time	2092	time C<offset> has passed. It will not repeat and will not adjust when a
1554	jump occurs, that is, if it is to be run at January 1st 2011 then it will	2093	time jump occurs, that is, if it is to be run at January 1st 2011 then it
1555	only run when the system clock reaches or surpasses this time.	2094	will be stopped and invoked when the system clock reaches or surpasses
		2095	this point in time.
1556		2096
1557	=item * repeating interval timer (at = offset, interval > 0, reschedule_cb = 0)	2097	=item * repeating interval timer (offset = offset within interval, interval > 0, reschedule_cb = 0)
1558		2098
1559	In this mode the watcher will always be scheduled to time out at the next	2099	In this mode the watcher will always be scheduled to time out at the next
1560	C<at + N * interval> time (for some integer N, which can also be negative)	2100	C<offset + N * interval> time (for some integer N, which can also be
1561	and then repeat, regardless of any time jumps.	2101	negative) and then repeat, regardless of any time jumps. The C<offset>
		2102	argument is merely an offset into the C<interval> periods.
1562		2103
1563	This can be used to create timers that do not drift with respect to the	2104	This can be used to create timers that do not drift with respect to the
1564	system clock, for example, here is a C<ev_periodic> that triggers each	2105	system clock, for example, here is an C<ev_periodic> that triggers each
1565	hour, on the hour:	2106	hour, on the hour (with respect to UTC):
1566		2107
1567	ev_periodic_set (&periodic, 0., 3600., 0);	2108	ev_periodic_set (&periodic, 0., 3600., 0);
1568		2109
1569	This doesn't mean there will always be 3600 seconds in between triggers,	2110	This doesn't mean there will always be 3600 seconds in between triggers,
1570	but only that the callback will be called when the system time shows a	2111	but only that the callback will be called when the system time shows a
1571	full hour (UTC), or more correctly, when the system time is evenly divisible	2112	full hour (UTC), or more correctly, when the system time is evenly divisible
1572	by 3600.	2113	by 3600.
1573		2114
1574	Another way to think about it (for the mathematically inclined) is that	2115	Another way to think about it (for the mathematically inclined) is that
1575	C<ev_periodic> will try to run the callback in this mode at the next possible	2116	C<ev_periodic> will try to run the callback in this mode at the next possible
1576	time where C<time = at (mod interval)>, regardless of any time jumps.	2117	time where C<time = offset (mod interval)>, regardless of any time jumps.
1577		2118
1578	For numerical stability it is preferable that the C<at> value is near	2119	For numerical stability it is preferable that the C<offset> value is near
1579	C<ev_now ()> (the current time), but there is no range requirement for	2120	C<ev_now ()> (the current time), but there is no range requirement for
1580	this value, and in fact is often specified as zero.	2121	this value, and in fact is often specified as zero.
1581		2122
1582	Note also that there is an upper limit to how often a timer can fire (CPU	2123	Note also that there is an upper limit to how often a timer can fire (CPU
1583	speed for example), so if C<interval> is very small then timing stability	2124	speed for example), so if C<interval> is very small then timing stability
1584	will of course deteriorate. Libev itself tries to be exact to be about one	2125	will of course deteriorate. Libev itself tries to be exact to be about one
1585	millisecond (if the OS supports it and the machine is fast enough).	2126	millisecond (if the OS supports it and the machine is fast enough).
1586		2127
1587	=item * manual reschedule mode (at and interval ignored, reschedule_cb = callback)	2128	=item * manual reschedule mode (offset ignored, interval ignored, reschedule_cb = callback)
1588		2129
1589	In this mode the values for C<interval> and C<at> are both being	2130	In this mode the values for C<interval> and C<offset> are both being
1590	ignored. Instead, each time the periodic watcher gets scheduled, the	2131	ignored. Instead, each time the periodic watcher gets scheduled, the
1591	reschedule callback will be called with the watcher as first, and the	2132	reschedule callback will be called with the watcher as first, and the
1592	current time as second argument.	2133	current time as second argument.
1593		2134
1594	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher,	2135	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher, ever,
1595	ever, or make ANY event loop modifications whatsoever>.	2136	or make ANY other event loop modifications whatsoever, unless explicitly
		2137	allowed by documentation here>.
1596		2138
1597	If you need to stop it, return C<now + 1e30> (or so, fudge fudge) and stop	2139	If you need to stop it, return C<now + 1e30> (or so, fudge fudge) and stop
1598	it afterwards (e.g. by starting an C<ev_prepare> watcher, which is the	2140	it afterwards (e.g. by starting an C<ev_prepare> watcher, which is the
1599	only event loop modification you are allowed to do).	2141	only event loop modification you are allowed to do).
1600		2142
…		…
1630	a different time than the last time it was called (e.g. in a crond like	2172	a different time than the last time it was called (e.g. in a crond like
1631	program when the crontabs have changed).	2173	program when the crontabs have changed).
1632		2174
1633	=item ev_tstamp ev_periodic_at (ev_periodic *)	2175	=item ev_tstamp ev_periodic_at (ev_periodic *)
1634		2176
1635	When active, returns the absolute time that the watcher is supposed to	2177	When active, returns the absolute time that the watcher is supposed
1636	trigger next.	2178	to trigger next. This is not the same as the C<offset> argument to
		2179	C<ev_periodic_set>, but indeed works even in interval and manual
		2180	rescheduling modes.
1637		2181
1638	=item ev_tstamp offset [read-write]	2182	=item ev_tstamp offset [read-write]
1639		2183
1640	When repeating, this contains the offset value, otherwise this is the	2184	When repeating, this contains the offset value, otherwise this is the
1641	absolute point in time (the C<at> value passed to C<ev_periodic_set>).	2185	absolute point in time (the C<offset> value passed to C<ev_periodic_set>,
		2186	although libev might modify this value for better numerical stability).
1642		2187
1643	Can be modified any time, but changes only take effect when the periodic	2188	Can be modified any time, but changes only take effect when the periodic
1644	timer fires or C<ev_periodic_again> is being called.	2189	timer fires or C<ev_periodic_again> is being called.
1645		2190
1646	=item ev_tstamp interval [read-write]	2191	=item ev_tstamp interval [read-write]
…		…
1662	Example: Call a callback every hour, or, more precisely, whenever the	2207	Example: Call a callback every hour, or, more precisely, whenever the
1663	system time is divisible by 3600. The callback invocation times have	2208	system time is divisible by 3600. The callback invocation times have
1664	potentially a lot of jitter, but good long-term stability.	2209	potentially a lot of jitter, but good long-term stability.
1665		2210
1666	static void	2211	static void
1667	clock_cb (struct ev_loop *loop, ev_io *w, int revents)	2212	clock_cb (struct ev_loop *loop, ev_periodic *w, int revents)
1668	{	2213	{
1669	... its now a full hour (UTC, or TAI or whatever your clock follows)	2214	... its now a full hour (UTC, or TAI or whatever your clock follows)
1670	}	2215	}
1671		2216
1672	ev_periodic hourly_tick;	2217	ev_periodic hourly_tick;
…		…
1698	Signal watchers will trigger an event when the process receives a specific	2243	Signal watchers will trigger an event when the process receives a specific
1699	signal one or more times. Even though signals are very asynchronous, libev	2244	signal one or more times. Even though signals are very asynchronous, libev
1700	will try it's best to deliver signals synchronously, i.e. as part of the	2245	will try it's best to deliver signals synchronously, i.e. as part of the
1701	normal event processing, like any other event.	2246	normal event processing, like any other event.
1702		2247
1703	If you want signals asynchronously, just use C<sigaction> as you would	2248	If you want signals to be delivered truly asynchronously, just use
1704	do without libev and forget about sharing the signal. You can even use	2249	C<sigaction> as you would do without libev and forget about sharing
1705	C<ev_async> from a signal handler to synchronously wake up an event loop.	2250	the signal. You can even use C<ev_async> from a signal handler to
		2251	synchronously wake up an event loop.
1706		2252
1707	You can configure as many watchers as you like per signal. Only when the	2253	You can configure as many watchers as you like for the same signal, but
		2254	only within the same loop, i.e. you can watch for C<SIGINT> in your
		2255	default loop and for C<SIGIO> in another loop, but you cannot watch for
		2256	C<SIGINT> in both the default loop and another loop at the same time. At
		2257	the moment, C<SIGCHLD> is permanently tied to the default loop.
		2258
1708	first watcher gets started will libev actually register a signal handler	2259	When the first watcher gets started will libev actually register something
1709	with the kernel (thus it coexists with your own signal handlers as long as	2260	with the kernel (thus it coexists with your own signal handlers as long as
1710	you don't register any with libev for the same signal). Similarly, when	2261	you don't register any with libev for the same signal).
1711	the last signal watcher for a signal is stopped, libev will reset the
1712	signal handler to SIG_DFL (regardless of what it was set to before).
1713		2262
1714	If possible and supported, libev will install its handlers with	2263	If possible and supported, libev will install its handlers with
1715	C<SA_RESTART> behaviour enabled, so system calls should not be unduly	2264	C<SA_RESTART> (or equivalent) behaviour enabled, so system calls should
1716	interrupted. If you have a problem with system calls getting interrupted by	2265	not be unduly interrupted. If you have a problem with system calls getting
1717	signals you can block all signals in an C<ev_check> watcher and unblock	2266	interrupted by signals you can block all signals in an C<ev_check> watcher
1718	them in an C<ev_prepare> watcher.	2267	and unblock them in an C<ev_prepare> watcher.
		2268
		2269	=head3 The special problem of inheritance over fork/execve/pthread_create
		2270
		2271	Both the signal mask (C<sigprocmask>) and the signal disposition
		2272	(C<sigaction>) are unspecified after starting a signal watcher (and after
		2273	stopping it again), that is, libev might or might not block the signal,
		2274	and might or might not set or restore the installed signal handler.
		2275
		2276	While this does not matter for the signal disposition (libev never
		2277	sets signals to C<SIG_IGN>, so handlers will be reset to C<SIG_DFL> on
		2278	C<execve>), this matters for the signal mask: many programs do not expect
		2279	certain signals to be blocked.
		2280
		2281	This means that before calling C<exec> (from the child) you should reset
		2282	the signal mask to whatever "default" you expect (all clear is a good
		2283	choice usually).
		2284
		2285	The simplest way to ensure that the signal mask is reset in the child is
		2286	to install a fork handler with C<pthread_atfork> that resets it. That will
		2287	catch fork calls done by libraries (such as the libc) as well.
		2288
		2289	In current versions of libev, the signal will not be blocked indefinitely
		2290	unless you use the C<signalfd> API (C<EV_SIGNALFD>). While this reduces
		2291	the window of opportunity for problems, it will not go away, as libev
		2292	I<has> to modify the signal mask, at least temporarily.
		2293
		2294	So I can't stress this enough: I<If you do not reset your signal mask when
		2295	you expect it to be empty, you have a race condition in your code>. This
		2296	is not a libev-specific thing, this is true for most event libraries.
1719		2297
1720	=head3 Watcher-Specific Functions and Data Members	2298	=head3 Watcher-Specific Functions and Data Members
1721		2299
1722	=over 4	2300	=over 4
1723		2301
…		…
1739	Example: Try to exit cleanly on SIGINT.	2317	Example: Try to exit cleanly on SIGINT.
1740		2318
1741	static void	2319	static void
1742	sigint_cb (struct ev_loop *loop, ev_signal *w, int revents)	2320	sigint_cb (struct ev_loop *loop, ev_signal *w, int revents)
1743	{	2321	{
1744	ev_unloop (loop, EVUNLOOP_ALL);	2322	ev_break (loop, EVBREAK_ALL);
1745	}	2323	}
1746		2324
1747	ev_signal signal_watcher;	2325	ev_signal signal_watcher;
1748	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);	2326	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
1749	ev_signal_start (loop, &signal_watcher);	2327	ev_signal_start (loop, &signal_watcher);
…		…
1755	some child status changes (most typically when a child of yours dies or	2333	some child status changes (most typically when a child of yours dies or
1756	exits). It is permissible to install a child watcher I<after> the child	2334	exits). It is permissible to install a child watcher I<after> the child
1757	has been forked (which implies it might have already exited), as long	2335	has been forked (which implies it might have already exited), as long
1758	as the event loop isn't entered (or is continued from a watcher), i.e.,	2336	as the event loop isn't entered (or is continued from a watcher), i.e.,
1759	forking and then immediately registering a watcher for the child is fine,	2337	forking and then immediately registering a watcher for the child is fine,
1760	but forking and registering a watcher a few event loop iterations later is	2338	but forking and registering a watcher a few event loop iterations later or
1761	not.	2339	in the next callback invocation is not.
1762		2340
1763	Only the default event loop is capable of handling signals, and therefore	2341	Only the default event loop is capable of handling signals, and therefore
1764	you can only register child watchers in the default event loop.	2342	you can only register child watchers in the default event loop.
1765		2343
		2344	Due to some design glitches inside libev, child watchers will always be
		2345	handled at maximum priority (their priority is set to C<EV_MAXPRI> by
		2346	libev)
		2347
1766	=head3 Process Interaction	2348	=head3 Process Interaction
1767		2349
1768	Libev grabs C<SIGCHLD> as soon as the default event loop is	2350	Libev grabs C<SIGCHLD> as soon as the default event loop is
1769	initialised. This is necessary to guarantee proper behaviour even if	2351	initialised. This is necessary to guarantee proper behaviour even if the
1770	the first child watcher is started after the child exits. The occurrence	2352	first child watcher is started after the child exits. The occurrence
1771	of C<SIGCHLD> is recorded asynchronously, but child reaping is done	2353	of C<SIGCHLD> is recorded asynchronously, but child reaping is done
1772	synchronously as part of the event loop processing. Libev always reaps all	2354	synchronously as part of the event loop processing. Libev always reaps all
1773	children, even ones not watched.	2355	children, even ones not watched.
1774		2356
1775	=head3 Overriding the Built-In Processing	2357	=head3 Overriding the Built-In Processing
…		…
1785	=head3 Stopping the Child Watcher	2367	=head3 Stopping the Child Watcher
1786		2368
1787	Currently, the child watcher never gets stopped, even when the	2369	Currently, the child watcher never gets stopped, even when the
1788	child terminates, so normally one needs to stop the watcher in the	2370	child terminates, so normally one needs to stop the watcher in the
1789	callback. Future versions of libev might stop the watcher automatically	2371	callback. Future versions of libev might stop the watcher automatically
1790	when a child exit is detected.	2372	when a child exit is detected (calling C<ev_child_stop> twice is not a
		2373	problem).
1791		2374
1792	=head3 Watcher-Specific Functions and Data Members	2375	=head3 Watcher-Specific Functions and Data Members
1793		2376
1794	=over 4	2377	=over 4
1795		2378
…		…
1852		2435
1853		2436
1854	=head2 C<ev_stat> - did the file attributes just change?	2437	=head2 C<ev_stat> - did the file attributes just change?
1855		2438
1856	This watches a file system path for attribute changes. That is, it calls	2439	This watches a file system path for attribute changes. That is, it calls
1857	C<stat> regularly (or when the OS says it changed) and sees if it changed	2440	C<stat> on that path in regular intervals (or when the OS says it changed)
1858	compared to the last time, invoking the callback if it did.	2441	and sees if it changed compared to the last time, invoking the callback if
		2442	it did.
1859		2443
1860	The path does not need to exist: changing from "path exists" to "path does	2444	The path does not need to exist: changing from "path exists" to "path does
1861	not exist" is a status change like any other. The condition "path does	2445	not exist" is a status change like any other. The condition "path does not
1862	not exist" is signified by the C<st_nlink> field being zero (which is	2446	exist" (or more correctly "path cannot be stat'ed") is signified by the
1863	otherwise always forced to be at least one) and all the other fields of	2447	C<st_nlink> field being zero (which is otherwise always forced to be at
1864	the stat buffer having unspecified contents.	2448	least one) and all the other fields of the stat buffer having unspecified
		2449	contents.
1865		2450
1866	The path I<should> be absolute and I<must not> end in a slash. If it is	2451	The path I<must not> end in a slash or contain special components such as
		2452	C<.> or C<..>. The path I<should> be absolute: If it is relative and
1867	relative and your working directory changes, the behaviour is undefined.	2453	your working directory changes, then the behaviour is undefined.
1868		2454
1869	Since there is no standard kernel interface to do this, the portable	2455	Since there is no portable change notification interface available, the
1870	implementation simply calls C<stat (2)> regularly on the path to see if	2456	portable implementation simply calls C<stat(2)> regularly on the path
1871	it changed somehow. You can specify a recommended polling interval for	2457	to see if it changed somehow. You can specify a recommended polling
1872	this case. If you specify a polling interval of C<0> (highly recommended!)	2458	interval for this case. If you specify a polling interval of C<0> (highly
1873	then a I<suitable, unspecified default> value will be used (which	2459	recommended!) then a I<suitable, unspecified default> value will be used
1874	you can expect to be around five seconds, although this might change	2460	(which you can expect to be around five seconds, although this might
1875	dynamically). Libev will also impose a minimum interval which is currently	2461	change dynamically). Libev will also impose a minimum interval which is
1876	around C<0.1>, but thats usually overkill.	2462	currently around C<0.1>, but that's usually overkill.
1877		2463
1878	This watcher type is not meant for massive numbers of stat watchers,	2464	This watcher type is not meant for massive numbers of stat watchers,
1879	as even with OS-supported change notifications, this can be	2465	as even with OS-supported change notifications, this can be
1880	resource-intensive.	2466	resource-intensive.
1881		2467
1882	At the time of this writing, the only OS-specific interface implemented	2468	At the time of this writing, the only OS-specific interface implemented
1883	is the Linux inotify interface (implementing kqueue support is left as	2469	is the Linux inotify interface (implementing kqueue support is left as an
1884	an exercise for the reader. Note, however, that the author sees no way	2470	exercise for the reader. Note, however, that the author sees no way of
1885	of implementing C<ev_stat> semantics with kqueue).	2471	implementing C<ev_stat> semantics with kqueue, except as a hint).
1886		2472
1887	=head3 ABI Issues (Largefile Support)	2473	=head3 ABI Issues (Largefile Support)
1888		2474
1889	Libev by default (unless the user overrides this) uses the default	2475	Libev by default (unless the user overrides this) uses the default
1890	compilation environment, which means that on systems with large file	2476	compilation environment, which means that on systems with large file
1891	support disabled by default, you get the 32 bit version of the stat	2477	support disabled by default, you get the 32 bit version of the stat
1892	structure. When using the library from programs that change the ABI to	2478	structure. When using the library from programs that change the ABI to
1893	use 64 bit file offsets the programs will fail. In that case you have to	2479	use 64 bit file offsets the programs will fail. In that case you have to
1894	compile libev with the same flags to get binary compatibility. This is	2480	compile libev with the same flags to get binary compatibility. This is
1895	obviously the case with any flags that change the ABI, but the problem is	2481	obviously the case with any flags that change the ABI, but the problem is
1896	most noticeably disabled with ev_stat and large file support.	2482	most noticeably displayed with ev_stat and large file support.
1897		2483
1898	The solution for this is to lobby your distribution maker to make large	2484	The solution for this is to lobby your distribution maker to make large
1899	file interfaces available by default (as e.g. FreeBSD does) and not	2485	file interfaces available by default (as e.g. FreeBSD does) and not
1900	optional. Libev cannot simply switch on large file support because it has	2486	optional. Libev cannot simply switch on large file support because it has
1901	to exchange stat structures with application programs compiled using the	2487	to exchange stat structures with application programs compiled using the
1902	default compilation environment.	2488	default compilation environment.
1903		2489
1904	=head3 Inotify and Kqueue	2490	=head3 Inotify and Kqueue
1905		2491
1906	When C<inotify (7)> support has been compiled into libev (generally	2492	When C<inotify (7)> support has been compiled into libev and present at
1907	only available with Linux 2.6.25 or above due to bugs in earlier	2493	runtime, it will be used to speed up change detection where possible. The
1908	implementations) and present at runtime, it will be used to speed up	2494	inotify descriptor will be created lazily when the first C<ev_stat>
1909	change detection where possible. The inotify descriptor will be created	2495	watcher is being started.
1910	lazily when the first C<ev_stat> watcher is being started.
1911		2496
1912	Inotify presence does not change the semantics of C<ev_stat> watchers	2497	Inotify presence does not change the semantics of C<ev_stat> watchers
1913	except that changes might be detected earlier, and in some cases, to avoid	2498	except that changes might be detected earlier, and in some cases, to avoid
1914	making regular C<stat> calls. Even in the presence of inotify support	2499	making regular C<stat> calls. Even in the presence of inotify support
1915	there are many cases where libev has to resort to regular C<stat> polling,	2500	there are many cases where libev has to resort to regular C<stat> polling,
1916	but as long as the path exists, libev usually gets away without polling.	2501	but as long as kernel 2.6.25 or newer is used (2.6.24 and older have too
		2502	many bugs), the path exists (i.e. stat succeeds), and the path resides on
		2503	a local filesystem (libev currently assumes only ext2/3, jfs, reiserfs and
		2504	xfs are fully working) libev usually gets away without polling.
1917		2505
1918	There is no support for kqueue, as apparently it cannot be used to	2506	There is no support for kqueue, as apparently it cannot be used to
1919	implement this functionality, due to the requirement of having a file	2507	implement this functionality, due to the requirement of having a file
1920	descriptor open on the object at all times, and detecting renames, unlinks	2508	descriptor open on the object at all times, and detecting renames, unlinks
1921	etc. is difficult.	2509	etc. is difficult.
1922		2510
		2511	=head3 C<stat ()> is a synchronous operation
		2512
		2513	Libev doesn't normally do any kind of I/O itself, and so is not blocking
		2514	the process. The exception are C<ev_stat> watchers - those call C<stat
		2515	()>, which is a synchronous operation.
		2516
		2517	For local paths, this usually doesn't matter: unless the system is very
		2518	busy or the intervals between stat's are large, a stat call will be fast,
		2519	as the path data is usually in memory already (except when starting the
		2520	watcher).
		2521
		2522	For networked file systems, calling C<stat ()> can block an indefinite
		2523	time due to network issues, and even under good conditions, a stat call
		2524	often takes multiple milliseconds.
		2525
		2526	Therefore, it is best to avoid using C<ev_stat> watchers on networked
		2527	paths, although this is fully supported by libev.
		2528
1923	=head3 The special problem of stat time resolution	2529	=head3 The special problem of stat time resolution
1924		2530
1925	The C<stat ()> system call only supports full-second resolution portably, and	2531	The C<stat ()> system call only supports full-second resolution portably,
1926	even on systems where the resolution is higher, most file systems still	2532	and even on systems where the resolution is higher, most file systems
1927	only support whole seconds.	2533	still only support whole seconds.
1928		2534
1929	That means that, if the time is the only thing that changes, you can	2535	That means that, if the time is the only thing that changes, you can
1930	easily miss updates: on the first update, C<ev_stat> detects a change and	2536	easily miss updates: on the first update, C<ev_stat> detects a change and
1931	calls your callback, which does something. When there is another update	2537	calls your callback, which does something. When there is another update
1932	within the same second, C<ev_stat> will be unable to detect unless the	2538	within the same second, C<ev_stat> will be unable to detect unless the
…		…
2075		2681
2076	=head3 Watcher-Specific Functions and Data Members	2682	=head3 Watcher-Specific Functions and Data Members
2077		2683
2078	=over 4	2684	=over 4
2079		2685
2080	=item ev_idle_init (ev_signal *, callback)	2686	=item ev_idle_init (ev_idle *, callback)
2081		2687
2082	Initialises and configures the idle watcher - it has no parameters of any	2688	Initialises and configures the idle watcher - it has no parameters of any
2083	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,	2689	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,
2084	believe me.	2690	believe me.
2085		2691
…		…
2098	// no longer anything immediate to do.	2704	// no longer anything immediate to do.
2099	}	2705	}
2100		2706
2101	ev_idle *idle_watcher = malloc (sizeof (ev_idle));	2707	ev_idle *idle_watcher = malloc (sizeof (ev_idle));
2102	ev_idle_init (idle_watcher, idle_cb);	2708	ev_idle_init (idle_watcher, idle_cb);
2103	ev_idle_start (loop, idle_cb);	2709	ev_idle_start (loop, idle_watcher);
2104		2710
2105		2711
2106	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!	2712	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!
2107		2713
2108	Prepare and check watchers are usually (but not always) used in pairs:	2714	Prepare and check watchers are usually (but not always) used in pairs:
2109	prepare watchers get invoked before the process blocks and check watchers	2715	prepare watchers get invoked before the process blocks and check watchers
2110	afterwards.	2716	afterwards.
2111		2717
2112	You I<must not> call C<ev_loop> or similar functions that enter	2718	You I<must not> call C<ev_run> or similar functions that enter
2113	the current event loop from either C<ev_prepare> or C<ev_check>	2719	the current event loop from either C<ev_prepare> or C<ev_check>
2114	watchers. Other loops than the current one are fine, however. The	2720	watchers. Other loops than the current one are fine, however. The
2115	rationale behind this is that you do not need to check for recursion in	2721	rationale behind this is that you do not need to check for recursion in
2116	those watchers, i.e. the sequence will always be C<ev_prepare>, blocking,	2722	those watchers, i.e. the sequence will always be C<ev_prepare>, blocking,
2117	C<ev_check> so if you have one watcher of each kind they will always be	2723	C<ev_check> so if you have one watcher of each kind they will always be
…		…
2201	struct pollfd fds [nfd];	2807	struct pollfd fds [nfd];
2202	// actual code will need to loop here and realloc etc.	2808	// actual code will need to loop here and realloc etc.
2203	adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));	2809	adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));
2204		2810
2205	/* the callback is illegal, but won't be called as we stop during check */	2811	/* the callback is illegal, but won't be called as we stop during check */
2206	ev_timer_init (&tw, 0, timeout * 1e-3);	2812	ev_timer_init (&tw, 0, timeout * 1e-3, 0.);
2207	ev_timer_start (loop, &tw);	2813	ev_timer_start (loop, &tw);
2208		2814
2209	// create one ev_io per pollfd	2815	// create one ev_io per pollfd
2210	for (int i = 0; i < nfd; ++i)	2816	for (int i = 0; i < nfd; ++i)
2211	{	2817	{
…		…
2285		2891
2286	if (timeout >= 0)	2892	if (timeout >= 0)
2287	// create/start timer	2893	// create/start timer
2288		2894
2289	// poll	2895	// poll
2290	ev_loop (EV_A_ 0);	2896	ev_run (EV_A_ 0);
2291		2897
2292	// stop timer again	2898	// stop timer again
2293	if (timeout >= 0)	2899	if (timeout >= 0)
2294	ev_timer_stop (EV_A_ &to);	2900	ev_timer_stop (EV_A_ &to);
2295		2901
…		…
2324	some fds have to be watched and handled very quickly (with low latency),	2930	some fds have to be watched and handled very quickly (with low latency),
2325	and even priorities and idle watchers might have too much overhead. In	2931	and even priorities and idle watchers might have too much overhead. In
2326	this case you would put all the high priority stuff in one loop and all	2932	this case you would put all the high priority stuff in one loop and all
2327	the rest in a second one, and embed the second one in the first.	2933	the rest in a second one, and embed the second one in the first.
2328		2934
2329	As long as the watcher is active, the callback will be invoked every time	2935	As long as the watcher is active, the callback will be invoked every
2330	there might be events pending in the embedded loop. The callback must then	2936	time there might be events pending in the embedded loop. The callback
2331	call C<ev_embed_sweep (mainloop, watcher)> to make a single sweep and invoke	2937	must then call C<ev_embed_sweep (mainloop, watcher)> to make a single
2332	their callbacks (you could also start an idle watcher to give the embedded	2938	sweep and invoke their callbacks (the callback doesn't need to invoke the
2333	loop strictly lower priority for example). You can also set the callback	2939	C<ev_embed_sweep> function directly, it could also start an idle watcher
2334	to C<0>, in which case the embed watcher will automatically execute the	2940	to give the embedded loop strictly lower priority for example).
2335	embedded loop sweep.
2336		2941
2337	As long as the watcher is started it will automatically handle events. The	2942	You can also set the callback to C<0>, in which case the embed watcher
2338	callback will be invoked whenever some events have been handled. You can	2943	will automatically execute the embedded loop sweep whenever necessary.
2339	set the callback to C<0> to avoid having to specify one if you are not
2340	interested in that.
2341		2944
2342	Also, there have not currently been made special provisions for forking:	2945	Fork detection will be handled transparently while the C<ev_embed> watcher
2343	when you fork, you not only have to call C<ev_loop_fork> on both loops,	2946	is active, i.e., the embedded loop will automatically be forked when the
2344	but you will also have to stop and restart any C<ev_embed> watchers	2947	embedding loop forks. In other cases, the user is responsible for calling
2345	yourself - but you can use a fork watcher to handle this automatically,	2948	C<ev_loop_fork> on the embedded loop.
2346	and future versions of libev might do just that.
2347		2949
2348	Unfortunately, not all backends are embeddable: only the ones returned by	2950	Unfortunately, not all backends are embeddable: only the ones returned by
2349	C<ev_embeddable_backends> are, which, unfortunately, does not include any	2951	C<ev_embeddable_backends> are, which, unfortunately, does not include any
2350	portable one.	2952	portable one.
2351		2953
…		…
2377	if you do not want that, you need to temporarily stop the embed watcher).	2979	if you do not want that, you need to temporarily stop the embed watcher).
2378		2980
2379	=item ev_embed_sweep (loop, ev_embed *)	2981	=item ev_embed_sweep (loop, ev_embed *)
2380		2982
2381	Make a single, non-blocking sweep over the embedded loop. This works	2983	Make a single, non-blocking sweep over the embedded loop. This works
2382	similarly to C<ev_loop (embedded_loop, EVLOOP_NONBLOCK)>, but in the most	2984	similarly to C<ev_run (embedded_loop, EVRUN_NOWAIT)>, but in the most
2383	appropriate way for embedded loops.	2985	appropriate way for embedded loops.
2384		2986
2385	=item struct ev_loop *other [read-only]	2987	=item struct ev_loop *other [read-only]
2386		2988
2387	The embedded event loop.	2989	The embedded event loop.
…		…
2445	event loop blocks next and before C<ev_check> watchers are being called,	3047	event loop blocks next and before C<ev_check> watchers are being called,
2446	and only in the child after the fork. If whoever good citizen calling	3048	and only in the child after the fork. If whoever good citizen calling
2447	C<ev_default_fork> cheats and calls it in the wrong process, the fork	3049	C<ev_default_fork> cheats and calls it in the wrong process, the fork
2448	handlers will be invoked, too, of course.	3050	handlers will be invoked, too, of course.
2449		3051
		3052	=head3 The special problem of life after fork - how is it possible?
		3053
		3054	Most uses of C<fork()> consist of forking, then some simple calls to set
		3055	up/change the process environment, followed by a call to C<exec()>. This
		3056	sequence should be handled by libev without any problems.
		3057
		3058	This changes when the application actually wants to do event handling
		3059	in the child, or both parent in child, in effect "continuing" after the
		3060	fork.
		3061
		3062	The default mode of operation (for libev, with application help to detect
		3063	forks) is to duplicate all the state in the child, as would be expected
		3064	when I<either> the parent I<or> the child process continues.
		3065
		3066	When both processes want to continue using libev, then this is usually the
		3067	wrong result. In that case, usually one process (typically the parent) is
		3068	supposed to continue with all watchers in place as before, while the other
		3069	process typically wants to start fresh, i.e. without any active watchers.
		3070
		3071	The cleanest and most efficient way to achieve that with libev is to
		3072	simply create a new event loop, which of course will be "empty", and
		3073	use that for new watchers. This has the advantage of not touching more
		3074	memory than necessary, and thus avoiding the copy-on-write, and the
		3075	disadvantage of having to use multiple event loops (which do not support
		3076	signal watchers).
		3077
		3078	When this is not possible, or you want to use the default loop for
		3079	other reasons, then in the process that wants to start "fresh", call
		3080	C<ev_default_destroy ()> followed by C<ev_default_loop (...)>. Destroying
		3081	the default loop will "orphan" (not stop) all registered watchers, so you
		3082	have to be careful not to execute code that modifies those watchers. Note
		3083	also that in that case, you have to re-register any signal watchers.
		3084
2450	=head3 Watcher-Specific Functions and Data Members	3085	=head3 Watcher-Specific Functions and Data Members
2451		3086
2452	=over 4	3087	=over 4
2453		3088
2454	=item ev_fork_init (ev_signal *, callback)	3089	=item ev_fork_init (ev_signal *, callback)
…		…
2458	believe me.	3093	believe me.
2459		3094
2460	=back	3095	=back
2461		3096
2462		3097
2463	=head2 C<ev_async> - how to wake up another event loop	3098	=head2 C<ev_async> - how to wake up an event loop
2464		3099
2465	In general, you cannot use an C<ev_loop> from multiple threads or other	3100	In general, you cannot use an C<ev_run> from multiple threads or other
2466	asynchronous sources such as signal handlers (as opposed to multiple event	3101	asynchronous sources such as signal handlers (as opposed to multiple event
2467	loops - those are of course safe to use in different threads).	3102	loops - those are of course safe to use in different threads).
2468		3103
2469	Sometimes, however, you need to wake up another event loop you do not	3104	Sometimes, however, you need to wake up an event loop you do not control,
2470	control, for example because it belongs to another thread. This is what	3105	for example because it belongs to another thread. This is what C<ev_async>
2471	C<ev_async> watchers do: as long as the C<ev_async> watcher is active, you	3106	watchers do: as long as the C<ev_async> watcher is active, you can signal
2472	can signal it by calling C<ev_async_send>, which is thread- and signal	3107	it by calling C<ev_async_send>, which is thread- and signal safe.
2473	safe.
2474		3108
2475	This functionality is very similar to C<ev_signal> watchers, as signals,	3109	This functionality is very similar to C<ev_signal> watchers, as signals,
2476	too, are asynchronous in nature, and signals, too, will be compressed	3110	too, are asynchronous in nature, and signals, too, will be compressed
2477	(i.e. the number of callback invocations may be less than the number of	3111	(i.e. the number of callback invocations may be less than the number of
2478	C<ev_async_sent> calls).	3112	C<ev_async_sent> calls).
…		…
2483	=head3 Queueing	3117	=head3 Queueing
2484		3118
2485	C<ev_async> does not support queueing of data in any way. The reason	3119	C<ev_async> does not support queueing of data in any way. The reason
2486	is that the author does not know of a simple (or any) algorithm for a	3120	is that the author does not know of a simple (or any) algorithm for a
2487	multiple-writer-single-reader queue that works in all cases and doesn't	3121	multiple-writer-single-reader queue that works in all cases and doesn't
2488	need elaborate support such as pthreads.	3122	need elaborate support such as pthreads or unportable memory access
		3123	semantics.
2489		3124
2490	That means that if you want to queue data, you have to provide your own	3125	That means that if you want to queue data, you have to provide your own
2491	queue. But at least I can tell you how to implement locking around your	3126	queue. But at least I can tell you how to implement locking around your
2492	queue:	3127	queue:
2493		3128
…		…
2571	=over 4	3206	=over 4
2572		3207
2573	=item ev_async_init (ev_async *, callback)	3208	=item ev_async_init (ev_async *, callback)
2574		3209
2575	Initialises and configures the async watcher - it has no parameters of any	3210	Initialises and configures the async watcher - it has no parameters of any
2576	kind. There is a C<ev_asynd_set> macro, but using it is utterly pointless,	3211	kind. There is a C<ev_async_set> macro, but using it is utterly pointless,
2577	trust me.	3212	trust me.
2578		3213
2579	=item ev_async_send (loop, ev_async *)	3214	=item ev_async_send (loop, ev_async *)
2580		3215
2581	Sends/signals/activates the given C<ev_async> watcher, that is, feeds	3216	Sends/signals/activates the given C<ev_async> watcher, that is, feeds
2582	an C<EV_ASYNC> event on the watcher into the event loop. Unlike	3217	an C<EV_ASYNC> event on the watcher into the event loop. Unlike
2583	C<ev_feed_event>, this call is safe to do from other threads, signal or	3218	C<ev_feed_event>, this call is safe to do from other threads, signal or
2584	similar contexts (see the discussion of C<EV_ATOMIC_T> in the embedding	3219	similar contexts (see the discussion of C<EV_ATOMIC_T> in the embedding
2585	section below on what exactly this means).	3220	section below on what exactly this means).
2586		3221
		3222	Note that, as with other watchers in libev, multiple events might get
		3223	compressed into a single callback invocation (another way to look at this
		3224	is that C<ev_async> watchers are level-triggered, set on C<ev_async_send>,
		3225	reset when the event loop detects that).
		3226
2587	This call incurs the overhead of a system call only once per loop iteration,	3227	This call incurs the overhead of a system call only once per event loop
2588	so while the overhead might be noticeable, it doesn't apply to repeated	3228	iteration, so while the overhead might be noticeable, it doesn't apply to
2589	calls to C<ev_async_send>.	3229	repeated calls to C<ev_async_send> for the same event loop.
2590		3230
2591	=item bool = ev_async_pending (ev_async *)	3231	=item bool = ev_async_pending (ev_async *)
2592		3232
2593	Returns a non-zero value when C<ev_async_send> has been called on the	3233	Returns a non-zero value when C<ev_async_send> has been called on the
2594	watcher but the event has not yet been processed (or even noted) by the	3234	watcher but the event has not yet been processed (or even noted) by the
…		…
2597	C<ev_async_send> sets a flag in the watcher and wakes up the loop. When	3237	C<ev_async_send> sets a flag in the watcher and wakes up the loop. When
2598	the loop iterates next and checks for the watcher to have become active,	3238	the loop iterates next and checks for the watcher to have become active,
2599	it will reset the flag again. C<ev_async_pending> can be used to very	3239	it will reset the flag again. C<ev_async_pending> can be used to very
2600	quickly check whether invoking the loop might be a good idea.	3240	quickly check whether invoking the loop might be a good idea.
2601		3241
2602	Not that this does I<not> check whether the watcher itself is pending, only	3242	Not that this does I<not> check whether the watcher itself is pending,
2603	whether it has been requested to make this watcher pending.	3243	only whether it has been requested to make this watcher pending: there
		3244	is a time window between the event loop checking and resetting the async
		3245	notification, and the callback being invoked.
2604		3246
2605	=back	3247	=back
2606		3248
2607		3249
2608	=head1 OTHER FUNCTIONS	3250	=head1 OTHER FUNCTIONS
…		…
2625		3267
2626	If C<timeout> is less than 0, then no timeout watcher will be	3268	If C<timeout> is less than 0, then no timeout watcher will be
2627	started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and	3269	started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and
2628	repeat = 0) will be started. C<0> is a valid timeout.	3270	repeat = 0) will be started. C<0> is a valid timeout.
2629		3271
2630	The callback has the type C<void (*cb)(int revents, void *arg)> and gets	3272	The callback has the type C<void (*cb)(int revents, void *arg)> and is
2631	passed an C<revents> set like normal event callbacks (a combination of	3273	passed an C<revents> set like normal event callbacks (a combination of
2632	C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMEOUT>) and the C<arg>	3274	C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMER>) and the C<arg>
2633	value passed to C<ev_once>. Note that it is possible to receive I<both>	3275	value passed to C<ev_once>. Note that it is possible to receive I<both>
2634	a timeout and an io event at the same time - you probably should give io	3276	a timeout and an io event at the same time - you probably should give io
2635	events precedence.	3277	events precedence.
2636		3278
2637	Example: wait up to ten seconds for data to appear on STDIN_FILENO.	3279	Example: wait up to ten seconds for data to appear on STDIN_FILENO.
2638		3280
2639	static void stdin_ready (int revents, void *arg)	3281	static void stdin_ready (int revents, void *arg)
2640	{	3282	{
2641	if (revents & EV_READ)	3283	if (revents & EV_READ)
2642	/* stdin might have data for us, joy! */;	3284	/* stdin might have data for us, joy! */;
2643	else if (revents & EV_TIMEOUT)	3285	else if (revents & EV_TIMER)
2644	/* doh, nothing entered */;	3286	/* doh, nothing entered */;
2645	}	3287	}
2646		3288
2647	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);	3289	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
2648		3290
2649	=item ev_feed_event (struct ev_loop *, watcher *, int revents)
2650
2651	Feeds the given event set into the event loop, as if the specified event
2652	had happened for the specified watcher (which must be a pointer to an
2653	initialised but not necessarily started event watcher).
2654
2655	=item ev_feed_fd_event (struct ev_loop *, int fd, int revents)	3291	=item ev_feed_fd_event (loop, int fd, int revents)
2656		3292
2657	Feed an event on the given fd, as if a file descriptor backend detected	3293	Feed an event on the given fd, as if a file descriptor backend detected
2658	the given events it.	3294	the given events it.
2659		3295
2660	=item ev_feed_signal_event (struct ev_loop *loop, int signum)	3296	=item ev_feed_signal_event (loop, int signum)
2661		3297
2662	Feed an event as if the given signal occurred (C<loop> must be the default	3298	Feed an event as if the given signal occurred (C<loop> must be the default
2663	loop!).	3299	loop!).
2664		3300
2665	=back	3301	=back
…		…
2745		3381
2746	=over 4	3382	=over 4
2747		3383
2748	=item ev::TYPE::TYPE ()	3384	=item ev::TYPE::TYPE ()
2749		3385
2750	=item ev::TYPE::TYPE (struct ev_loop *)	3386	=item ev::TYPE::TYPE (loop)
2751		3387
2752	=item ev::TYPE::~TYPE	3388	=item ev::TYPE::~TYPE
2753		3389
2754	The constructor (optionally) takes an event loop to associate the watcher	3390	The constructor (optionally) takes an event loop to associate the watcher
2755	with. If it is omitted, it will use C<EV_DEFAULT>.	3391	with. If it is omitted, it will use C<EV_DEFAULT>.
…		…
2787		3423
2788	myclass obj;	3424	myclass obj;
2789	ev::io iow;	3425	ev::io iow;
2790	iow.set <myclass, &myclass::io_cb> (&obj);	3426	iow.set <myclass, &myclass::io_cb> (&obj);
2791		3427
		3428	=item w->set (object *)
		3429
		3430	This is a variation of a method callback - leaving out the method to call
		3431	will default the method to C<operator ()>, which makes it possible to use
		3432	functor objects without having to manually specify the C<operator ()> all
		3433	the time. Incidentally, you can then also leave out the template argument
		3434	list.
		3435
		3436	The C<operator ()> method prototype must be C<void operator ()(watcher &w,
		3437	int revents)>.
		3438
		3439	See the method-C<set> above for more details.
		3440
		3441	Example: use a functor object as callback.
		3442
		3443	struct myfunctor
		3444	{
		3445	void operator() (ev::io &w, int revents)
		3446	{
		3447	...
		3448	}
		3449	}
		3450
		3451	myfunctor f;
		3452
		3453	ev::io w;
		3454	w.set (&f);
		3455
2792	=item w->set<function> (void *data = 0)	3456	=item w->set<function> (void *data = 0)
2793		3457
2794	Also sets a callback, but uses a static method or plain function as	3458	Also sets a callback, but uses a static method or plain function as
2795	callback. The optional C<data> argument will be stored in the watcher's	3459	callback. The optional C<data> argument will be stored in the watcher's
2796	C<data> member and is free for you to use.	3460	C<data> member and is free for you to use.
…		…
2802	Example: Use a plain function as callback.	3466	Example: Use a plain function as callback.
2803		3467
2804	static void io_cb (ev::io &w, int revents) { }	3468	static void io_cb (ev::io &w, int revents) { }
2805	iow.set <io_cb> ();	3469	iow.set <io_cb> ();
2806		3470
2807	=item w->set (struct ev_loop *)	3471	=item w->set (loop)
2808		3472
2809	Associates a different C<struct ev_loop> with this watcher. You can only	3473	Associates a different C<struct ev_loop> with this watcher. You can only
2810	do this when the watcher is inactive (and not pending either).	3474	do this when the watcher is inactive (and not pending either).
2811		3475
2812	=item w->set ([arguments])	3476	=item w->set ([arguments])
2813		3477
2814	Basically the same as C<ev_TYPE_set>, with the same arguments. Must be	3478	Basically the same as C<ev_TYPE_set>, with the same arguments. Either this
2815	called at least once. Unlike the C counterpart, an active watcher gets	3479	method or a suitable start method must be called at least once. Unlike the
2816	automatically stopped and restarted when reconfiguring it with this	3480	C counterpart, an active watcher gets automatically stopped and restarted
2817	method.	3481	when reconfiguring it with this method.
2818		3482
2819	=item w->start ()	3483	=item w->start ()
2820		3484
2821	Starts the watcher. Note that there is no C<loop> argument, as the	3485	Starts the watcher. Note that there is no C<loop> argument, as the
2822	constructor already stores the event loop.	3486	constructor already stores the event loop.
2823		3487
		3488	=item w->start ([arguments])
		3489
		3490	Instead of calling C<set> and C<start> methods separately, it is often
		3491	convenient to wrap them in one call. Uses the same type of arguments as
		3492	the configure C<set> method of the watcher.
		3493
2824	=item w->stop ()	3494	=item w->stop ()
2825		3495
2826	Stops the watcher if it is active. Again, no C<loop> argument.	3496	Stops the watcher if it is active. Again, no C<loop> argument.
2827		3497
2828	=item w->again () (C<ev::timer>, C<ev::periodic> only)	3498	=item w->again () (C<ev::timer>, C<ev::periodic> only)
…		…
2840		3510
2841	=back	3511	=back
2842		3512
2843	=back	3513	=back
2844		3514
2845	Example: Define a class with an IO and idle watcher, start one of them in	3515	Example: Define a class with two I/O and idle watchers, start the I/O
2846	the constructor.	3516	watchers in the constructor.
2847		3517
2848	class myclass	3518	class myclass
2849	{	3519	{
2850	ev::io io ; void io_cb (ev::io &w, int revents);	3520	ev::io io ; void io_cb (ev::io &w, int revents);
		3521	ev::io2 io2 ; void io2_cb (ev::io &w, int revents);
2851	ev::idle idle; void idle_cb (ev::idle &w, int revents);	3522	ev::idle idle; void idle_cb (ev::idle &w, int revents);
2852		3523
2853	myclass (int fd)	3524	myclass (int fd)
2854	{	3525	{
2855	io .set <myclass, &myclass::io_cb > (this);	3526	io .set <myclass, &myclass::io_cb > (this);
		3527	io2 .set <myclass, &myclass::io2_cb > (this);
2856	idle.set <myclass, &myclass::idle_cb> (this);	3528	idle.set <myclass, &myclass::idle_cb> (this);
2857		3529
2858	io.start (fd, ev::READ);	3530	io.set (fd, ev::WRITE); // configure the watcher
		3531	io.start (); // start it whenever convenient
		3532
		3533	io2.start (fd, ev::READ); // set + start in one call
2859	}	3534	}
2860	};	3535	};
2861		3536
2862		3537
2863	=head1 OTHER LANGUAGE BINDINGS	3538	=head1 OTHER LANGUAGE BINDINGS
…		…
2882	L<http://software.schmorp.de/pkg/EV>.	3557	L<http://software.schmorp.de/pkg/EV>.
2883		3558
2884	=item Python	3559	=item Python
2885		3560
2886	Python bindings can be found at L<http://code.google.com/p/pyev/>. It	3561	Python bindings can be found at L<http://code.google.com/p/pyev/>. It
2887	seems to be quite complete and well-documented. Note, however, that the	3562	seems to be quite complete and well-documented.
2888	patch they require for libev is outright dangerous as it breaks the ABI
2889	for everybody else, and therefore, should never be applied in an installed
2890	libev (if python requires an incompatible ABI then it needs to embed
2891	libev).
2892		3563
2893	=item Ruby	3564	=item Ruby
2894		3565
2895	Tony Arcieri has written a ruby extension that offers access to a subset	3566	Tony Arcieri has written a ruby extension that offers access to a subset
2896	of the libev API and adds file handle abstractions, asynchronous DNS and	3567	of the libev API and adds file handle abstractions, asynchronous DNS and
2897	more on top of it. It can be found via gem servers. Its homepage is at	3568	more on top of it. It can be found via gem servers. Its homepage is at
2898	L<http://rev.rubyforge.org/>.	3569	L<http://rev.rubyforge.org/>.
2899		3570
		3571	Roger Pack reports that using the link order C<-lws2_32 -lmsvcrt-ruby-190>
		3572	makes rev work even on mingw.
		3573
		3574	=item Haskell
		3575
		3576	A haskell binding to libev is available at
		3577	L<http://hackage.haskell.org/cgi-bin/hackage-scripts/package/hlibev>.
		3578
2900	=item D	3579	=item D
2901		3580
2902	Leandro Lucarella has written a D language binding (F<ev.d>) for libev, to	3581	Leandro Lucarella has written a D language binding (F<ev.d>) for libev, to
2903	be found at L<http://proj.llucax.com.ar/wiki/evd>.	3582	be found at L<http://proj.llucax.com.ar/wiki/evd>.
		3583
		3584	=item Ocaml
		3585
		3586	Erkki Seppala has written Ocaml bindings for libev, to be found at
		3587	L<http://modeemi.cs.tut.fi/~flux/software/ocaml-ev/>.
		3588
		3589	=item Lua
		3590
		3591	Brian Maher has written a partial interface to libev for lua (at the
		3592	time of this writing, only C<ev_io> and C<ev_timer>), to be found at
		3593	L<http://github.com/brimworks/lua-ev>.
2904		3594
2905	=back	3595	=back
2906		3596
2907		3597
2908	=head1 MACRO MAGIC	3598	=head1 MACRO MAGIC
…		…
2922	loop argument"). The C<EV_A> form is used when this is the sole argument,	3612	loop argument"). The C<EV_A> form is used when this is the sole argument,
2923	C<EV_A_> is used when other arguments are following. Example:	3613	C<EV_A_> is used when other arguments are following. Example:
2924		3614
2925	ev_unref (EV_A);	3615	ev_unref (EV_A);
2926	ev_timer_add (EV_A_ watcher);	3616	ev_timer_add (EV_A_ watcher);
2927	ev_loop (EV_A_ 0);	3617	ev_run (EV_A_ 0);
2928		3618
2929	It assumes the variable C<loop> of type C<struct ev_loop *> is in scope,	3619	It assumes the variable C<loop> of type C<struct ev_loop *> is in scope,
2930	which is often provided by the following macro.	3620	which is often provided by the following macro.
2931		3621
2932	=item C<EV_P>, C<EV_P_>	3622	=item C<EV_P>, C<EV_P_>
…		…
2972	}	3662	}
2973		3663
2974	ev_check check;	3664	ev_check check;
2975	ev_check_init (&check, check_cb);	3665	ev_check_init (&check, check_cb);
2976	ev_check_start (EV_DEFAULT_ &check);	3666	ev_check_start (EV_DEFAULT_ &check);
2977	ev_loop (EV_DEFAULT_ 0);	3667	ev_run (EV_DEFAULT_ 0);
2978		3668
2979	=head1 EMBEDDING	3669	=head1 EMBEDDING
2980		3670
2981	Libev can (and often is) directly embedded into host	3671	Libev can (and often is) directly embedded into host
2982	applications. Examples of applications that embed it include the Deliantra	3672	applications. Examples of applications that embed it include the Deliantra
…		…
3009		3699
3010	#define EV_STANDALONE 1	3700	#define EV_STANDALONE 1
3011	#include "ev.h"	3701	#include "ev.h"
3012		3702
3013	Both header files and implementation files can be compiled with a C++	3703	Both header files and implementation files can be compiled with a C++
3014	compiler (at least, thats a stated goal, and breakage will be treated	3704	compiler (at least, that's a stated goal, and breakage will be treated
3015	as a bug).	3705	as a bug).
3016		3706
3017	You need the following files in your source tree, or in a directory	3707	You need the following files in your source tree, or in a directory
3018	in your include path (e.g. in libev/ when using -Ilibev):	3708	in your include path (e.g. in libev/ when using -Ilibev):
3019		3709
…		…
3062	libev.m4	3752	libev.m4
3063		3753
3064	=head2 PREPROCESSOR SYMBOLS/MACROS	3754	=head2 PREPROCESSOR SYMBOLS/MACROS
3065		3755
3066	Libev can be configured via a variety of preprocessor symbols you have to	3756	Libev can be configured via a variety of preprocessor symbols you have to
3067	define before including any of its files. The default in the absence of	3757	define before including (or compiling) any of its files. The default in
3068	autoconf is documented for every option.	3758	the absence of autoconf is documented for every option.
		3759
		3760	Symbols marked with "(h)" do not change the ABI, and can have different
		3761	values when compiling libev vs. including F<ev.h>, so it is permissible
		3762	to redefine them before including F<ev.h> without breaking compatibility
		3763	to a compiled library. All other symbols change the ABI, which means all
		3764	users of libev and the libev code itself must be compiled with compatible
		3765	settings.
3069		3766
3070	=over 4	3767	=over 4
3071		3768
		3769	=item EV_COMPAT3 (h)
		3770
		3771	Backwards compatibility is a major concern for libev. This is why this
		3772	release of libev comes with wrappers for the functions and symbols that
		3773	have been renamed between libev version 3 and 4.
		3774
		3775	You can disable these wrappers (to test compatibility with future
		3776	versions) by defining C<EV_COMPAT3> to C<0> when compiling your
		3777	sources. This has the additional advantage that you can drop the C<struct>
		3778	from C<struct ev_loop> declarations, as libev will provide an C<ev_loop>
		3779	typedef in that case.
		3780
		3781	In some future version, the default for C<EV_COMPAT3> will become C<0>,
		3782	and in some even more future version the compatibility code will be
		3783	removed completely.
		3784
3072	=item EV_STANDALONE	3785	=item EV_STANDALONE (h)
3073		3786
3074	Must always be C<1> if you do not use autoconf configuration, which	3787	Must always be C<1> if you do not use autoconf configuration, which
3075	keeps libev from including F<config.h>, and it also defines dummy	3788	keeps libev from including F<config.h>, and it also defines dummy
3076	implementations for some libevent functions (such as logging, which is not	3789	implementations for some libevent functions (such as logging, which is not
3077	supported). It will also not define any of the structs usually found in	3790	supported). It will also not define any of the structs usually found in
3078	F<event.h> that are not directly supported by the libev core alone.	3791	F<event.h> that are not directly supported by the libev core alone.
3079		3792
		3793	In standalone mode, libev will still try to automatically deduce the
		3794	configuration, but has to be more conservative.
		3795
3080	=item EV_USE_MONOTONIC	3796	=item EV_USE_MONOTONIC
3081		3797
3082	If defined to be C<1>, libev will try to detect the availability of the	3798	If defined to be C<1>, libev will try to detect the availability of the
3083	monotonic clock option at both compile time and runtime. Otherwise no use	3799	monotonic clock option at both compile time and runtime. Otherwise no
3084	of the monotonic clock option will be attempted. If you enable this, you	3800	use of the monotonic clock option will be attempted. If you enable this,
3085	usually have to link against librt or something similar. Enabling it when	3801	you usually have to link against librt or something similar. Enabling it
3086	the functionality isn't available is safe, though, although you have	3802	when the functionality isn't available is safe, though, although you have
3087	to make sure you link against any libraries where the C<clock_gettime>	3803	to make sure you link against any libraries where the C<clock_gettime>
3088	function is hiding in (often F<-lrt>).	3804	function is hiding in (often F<-lrt>). See also C<EV_USE_CLOCK_SYSCALL>.
3089		3805
3090	=item EV_USE_REALTIME	3806	=item EV_USE_REALTIME
3091		3807
3092	If defined to be C<1>, libev will try to detect the availability of the	3808	If defined to be C<1>, libev will try to detect the availability of the
3093	real-time clock option at compile time (and assume its availability at	3809	real-time clock option at compile time (and assume its availability
3094	runtime if successful). Otherwise no use of the real-time clock option will	3810	at runtime if successful). Otherwise no use of the real-time clock
3095	be attempted. This effectively replaces C<gettimeofday> by C<clock_get	3811	option will be attempted. This effectively replaces C<gettimeofday>
3096	(CLOCK_REALTIME, ...)> and will not normally affect correctness. See the	3812	by C<clock_get (CLOCK_REALTIME, ...)> and will not normally affect
3097	note about libraries in the description of C<EV_USE_MONOTONIC>, though.	3813	correctness. See the note about libraries in the description of
		3814	C<EV_USE_MONOTONIC>, though. Defaults to the opposite value of
		3815	C<EV_USE_CLOCK_SYSCALL>.
		3816
		3817	=item EV_USE_CLOCK_SYSCALL
		3818
		3819	If defined to be C<1>, libev will try to use a direct syscall instead
		3820	of calling the system-provided C<clock_gettime> function. This option
		3821	exists because on GNU/Linux, C<clock_gettime> is in C<librt>, but C<librt>
		3822	unconditionally pulls in C<libpthread>, slowing down single-threaded
		3823	programs needlessly. Using a direct syscall is slightly slower (in
		3824	theory), because no optimised vdso implementation can be used, but avoids
		3825	the pthread dependency. Defaults to C<1> on GNU/Linux with glibc 2.x or
		3826	higher, as it simplifies linking (no need for C<-lrt>).
3098		3827
3099	=item EV_USE_NANOSLEEP	3828	=item EV_USE_NANOSLEEP
3100		3829
3101	If defined to be C<1>, libev will assume that C<nanosleep ()> is available	3830	If defined to be C<1>, libev will assume that C<nanosleep ()> is available
3102	and will use it for delays. Otherwise it will use C<select ()>.	3831	and will use it for delays. Otherwise it will use C<select ()>.
…		…
3118		3847
3119	=item EV_SELECT_USE_FD_SET	3848	=item EV_SELECT_USE_FD_SET
3120		3849
3121	If defined to C<1>, then the select backend will use the system C<fd_set>	3850	If defined to C<1>, then the select backend will use the system C<fd_set>
3122	structure. This is useful if libev doesn't compile due to a missing	3851	structure. This is useful if libev doesn't compile due to a missing
3123	C<NFDBITS> or C<fd_mask> definition or it mis-guesses the bitset layout on	3852	C<NFDBITS> or C<fd_mask> definition or it mis-guesses the bitset layout
3124	exotic systems. This usually limits the range of file descriptors to some	3853	on exotic systems. This usually limits the range of file descriptors to
3125	low limit such as 1024 or might have other limitations (winsocket only	3854	some low limit such as 1024 or might have other limitations (winsocket
3126	allows 64 sockets). The C<FD_SETSIZE> macro, set before compilation, might	3855	only allows 64 sockets). The C<FD_SETSIZE> macro, set before compilation,
3127	influence the size of the C<fd_set> used.	3856	configures the maximum size of the C<fd_set>.
3128		3857
3129	=item EV_SELECT_IS_WINSOCKET	3858	=item EV_SELECT_IS_WINSOCKET
3130		3859
3131	When defined to C<1>, the select backend will assume that	3860	When defined to C<1>, the select backend will assume that
3132	select/socket/connect etc. don't understand file descriptors but	3861	select/socket/connect etc. don't understand file descriptors but
…		…
3134	be used is the winsock select). This means that it will call	3863	be used is the winsock select). This means that it will call
3135	C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,	3864	C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,
3136	it is assumed that all these functions actually work on fds, even	3865	it is assumed that all these functions actually work on fds, even
3137	on win32. Should not be defined on non-win32 platforms.	3866	on win32. Should not be defined on non-win32 platforms.
3138		3867
3139	=item EV_FD_TO_WIN32_HANDLE	3868	=item EV_FD_TO_WIN32_HANDLE(fd)
3140		3869
3141	If C<EV_SELECT_IS_WINSOCKET> is enabled, then libev needs a way to map	3870	If C<EV_SELECT_IS_WINSOCKET> is enabled, then libev needs a way to map
3142	file descriptors to socket handles. When not defining this symbol (the	3871	file descriptors to socket handles. When not defining this symbol (the
3143	default), then libev will call C<_get_osfhandle>, which is usually	3872	default), then libev will call C<_get_osfhandle>, which is usually
3144	correct. In some cases, programs use their own file descriptor management,	3873	correct. In some cases, programs use their own file descriptor management,
3145	in which case they can provide this function to map fds to socket handles.	3874	in which case they can provide this function to map fds to socket handles.
		3875
		3876	=item EV_WIN32_HANDLE_TO_FD(handle)
		3877
		3878	If C<EV_SELECT_IS_WINSOCKET> then libev maps handles to file descriptors
		3879	using the standard C<_open_osfhandle> function. For programs implementing
		3880	their own fd to handle mapping, overwriting this function makes it easier
		3881	to do so. This can be done by defining this macro to an appropriate value.
		3882
		3883	=item EV_WIN32_CLOSE_FD(fd)
		3884
		3885	If programs implement their own fd to handle mapping on win32, then this
		3886	macro can be used to override the C<close> function, useful to unregister
		3887	file descriptors again. Note that the replacement function has to close
		3888	the underlying OS handle.
3146		3889
3147	=item EV_USE_POLL	3890	=item EV_USE_POLL
3148		3891
3149	If defined to be C<1>, libev will compile in support for the C<poll>(2)	3892	If defined to be C<1>, libev will compile in support for the C<poll>(2)
3150	backend. Otherwise it will be enabled on non-win32 platforms. It	3893	backend. Otherwise it will be enabled on non-win32 platforms. It
…		…
3197	as well as for signal and thread safety in C<ev_async> watchers.	3940	as well as for signal and thread safety in C<ev_async> watchers.
3198		3941
3199	In the absence of this define, libev will use C<sig_atomic_t volatile>	3942	In the absence of this define, libev will use C<sig_atomic_t volatile>
3200	(from F<signal.h>), which is usually good enough on most platforms.	3943	(from F<signal.h>), which is usually good enough on most platforms.
3201		3944
3202	=item EV_H	3945	=item EV_H (h)
3203		3946
3204	The name of the F<ev.h> header file used to include it. The default if	3947	The name of the F<ev.h> header file used to include it. The default if
3205	undefined is C<"ev.h"> in F<event.h>, F<ev.c> and F<ev++.h>. This can be	3948	undefined is C<"ev.h"> in F<event.h>, F<ev.c> and F<ev++.h>. This can be
3206	used to virtually rename the F<ev.h> header file in case of conflicts.	3949	used to virtually rename the F<ev.h> header file in case of conflicts.
3207		3950
3208	=item EV_CONFIG_H	3951	=item EV_CONFIG_H (h)
3209		3952
3210	If C<EV_STANDALONE> isn't C<1>, this variable can be used to override	3953	If C<EV_STANDALONE> isn't C<1>, this variable can be used to override
3211	F<ev.c>'s idea of where to find the F<config.h> file, similarly to	3954	F<ev.c>'s idea of where to find the F<config.h> file, similarly to
3212	C<EV_H>, above.	3955	C<EV_H>, above.
3213		3956
3214	=item EV_EVENT_H	3957	=item EV_EVENT_H (h)
3215		3958
3216	Similarly to C<EV_H>, this macro can be used to override F<event.c>'s idea	3959	Similarly to C<EV_H>, this macro can be used to override F<event.c>'s idea
3217	of how the F<event.h> header can be found, the default is C<"event.h">.	3960	of how the F<event.h> header can be found, the default is C<"event.h">.
3218		3961
3219	=item EV_PROTOTYPES	3962	=item EV_PROTOTYPES (h)
3220		3963
3221	If defined to be C<0>, then F<ev.h> will not define any function	3964	If defined to be C<0>, then F<ev.h> will not define any function
3222	prototypes, but still define all the structs and other symbols. This is	3965	prototypes, but still define all the structs and other symbols. This is
3223	occasionally useful if you want to provide your own wrapper functions	3966	occasionally useful if you want to provide your own wrapper functions
3224	around libev functions.	3967	around libev functions.
…		…
3246	fine.	3989	fine.
3247		3990
3248	If your embedding application does not need any priorities, defining these	3991	If your embedding application does not need any priorities, defining these
3249	both to C<0> will save some memory and CPU.	3992	both to C<0> will save some memory and CPU.
3250		3993
3251	=item EV_PERIODIC_ENABLE	3994	=item EV_PERIODIC_ENABLE, EV_IDLE_ENABLE, EV_EMBED_ENABLE, EV_STAT_ENABLE,
		3995	EV_PREPARE_ENABLE, EV_CHECK_ENABLE, EV_FORK_ENABLE, EV_SIGNAL_ENABLE,
		3996	EV_ASYNC_ENABLE, EV_CHILD_ENABLE.
3252		3997
3253	If undefined or defined to be C<1>, then periodic timers are supported. If	3998	If undefined or defined to be C<1> (and the platform supports it), then
3254	defined to be C<0>, then they are not. Disabling them saves a few kB of	3999	the respective watcher type is supported. If defined to be C<0>, then it
3255	code.	4000	is not. Disabling watcher types mainly saves code size.
3256		4001
3257	=item EV_IDLE_ENABLE	4002	=item EV_FEATURES
3258
3259	If undefined or defined to be C<1>, then idle watchers are supported. If
3260	defined to be C<0>, then they are not. Disabling them saves a few kB of
3261	code.
3262
3263	=item EV_EMBED_ENABLE
3264
3265	If undefined or defined to be C<1>, then embed watchers are supported. If
3266	defined to be C<0>, then they are not. Embed watchers rely on most other
3267	watcher types, which therefore must not be disabled.
3268
3269	=item EV_STAT_ENABLE
3270
3271	If undefined or defined to be C<1>, then stat watchers are supported. If
3272	defined to be C<0>, then they are not.
3273
3274	=item EV_FORK_ENABLE
3275
3276	If undefined or defined to be C<1>, then fork watchers are supported. If
3277	defined to be C<0>, then they are not.
3278
3279	=item EV_ASYNC_ENABLE
3280
3281	If undefined or defined to be C<1>, then async watchers are supported. If
3282	defined to be C<0>, then they are not.
3283
3284	=item EV_MINIMAL
3285		4003
3286	If you need to shave off some kilobytes of code at the expense of some	4004	If you need to shave off some kilobytes of code at the expense of some
3287	speed, define this symbol to C<1>. Currently this is used to override some	4005	speed (but with the full API), you can define this symbol to request
3288	inlining decisions, saves roughly 30% code size on amd64. It also selects a	4006	certain subsets of functionality. The default is to enable all features
3289	much smaller 2-heap for timer management over the default 4-heap.	4007	that can be enabled on the platform.
		4008
		4009	A typical way to use this symbol is to define it to C<0> (or to a bitset
		4010	with some broad features you want) and then selectively re-enable
		4011	additional parts you want, for example if you want everything minimal,
		4012	but multiple event loop support, async and child watchers and the poll
		4013	backend, use this:
		4014
		4015	#define EV_FEATURES 0
		4016	#define EV_MULTIPLICITY 1
		4017	#define EV_USE_POLL 1
		4018	#define EV_CHILD_ENABLE 1
		4019	#define EV_ASYNC_ENABLE 1
		4020
		4021	The actual value is a bitset, it can be a combination of the following
		4022	values:
		4023
		4024	=over 4
		4025
		4026	=item C<1> - faster/larger code
		4027
		4028	Use larger code to speed up some operations.
		4029
		4030	Currently this is used to override some inlining decisions (enlarging the
		4031	code size by roughly 30% on amd64).
		4032
		4033	When optimising for size, use of compiler flags such as C<-Os> with
		4034	gcc is recommended, as well as C<-DNDEBUG>, as libev contains a number of
		4035	assertions.
		4036
		4037	=item C<2> - faster/larger data structures
		4038
		4039	Replaces the small 2-heap for timer management by a faster 4-heap, larger
		4040	hash table sizes and so on. This will usually further increase code size
		4041	and can additionally have an effect on the size of data structures at
		4042	runtime.
		4043
		4044	=item C<4> - full API configuration
		4045
		4046	This enables priorities (sets C<EV_MAXPRI>=2 and C<EV_MINPRI>=-2), and
		4047	enables multiplicity (C<EV_MULTIPLICITY>=1).
		4048
		4049	=item C<8> - full API
		4050
		4051	This enables a lot of the "lesser used" API functions. See C<ev.h> for
		4052	details on which parts of the API are still available without this
		4053	feature, and do not complain if this subset changes over time.
		4054
		4055	=item C<16> - enable all optional watcher types
		4056
		4057	Enables all optional watcher types. If you want to selectively enable
		4058	only some watcher types other than I/O and timers (e.g. prepare,
		4059	embed, async, child...) you can enable them manually by defining
		4060	C<EV_watchertype_ENABLE> to C<1> instead.
		4061
		4062	=item C<32> - enable all backends
		4063
		4064	This enables all backends - without this feature, you need to enable at
		4065	least one backend manually (C<EV_USE_SELECT> is a good choice).
		4066
		4067	=item C<64> - enable OS-specific "helper" APIs
		4068
		4069	Enable inotify, eventfd, signalfd and similar OS-specific helper APIs by
		4070	default.
		4071
		4072	=back
		4073
		4074	Compiling with C<gcc -Os -DEV_STANDALONE -DEV_USE_EPOLL=1 -DEV_FEATURES=0>
		4075	reduces the compiled size of libev from 24.7Kb code/2.8Kb data to 6.5Kb
		4076	code/0.3Kb data on my GNU/Linux amd64 system, while still giving you I/O
		4077	watchers, timers and monotonic clock support.
		4078
		4079	With an intelligent-enough linker (gcc+binutils are intelligent enough
		4080	when you use C<-Wl,--gc-sections -ffunction-sections>) functions unused by
		4081	your program might be left out as well - a binary starting a timer and an
		4082	I/O watcher then might come out at only 5Kb.
		4083
		4084	=item EV_AVOID_STDIO
		4085
		4086	If this is set to C<1> at compiletime, then libev will avoid using stdio
		4087	functions (printf, scanf, perror etc.). This will increase the code size
		4088	somewhat, but if your program doesn't otherwise depend on stdio and your
		4089	libc allows it, this avoids linking in the stdio library which is quite
		4090	big.
		4091
		4092	Note that error messages might become less precise when this option is
		4093	enabled.
		4094
		4095	=item EV_NSIG
		4096
		4097	The highest supported signal number, +1 (or, the number of
		4098	signals): Normally, libev tries to deduce the maximum number of signals
		4099	automatically, but sometimes this fails, in which case it can be
		4100	specified. Also, using a lower number than detected (C<32> should be
		4101	good for about any system in existence) can save some memory, as libev
		4102	statically allocates some 12-24 bytes per signal number.
3290		4103
3291	=item EV_PID_HASHSIZE	4104	=item EV_PID_HASHSIZE
3292		4105
3293	C<ev_child> watchers use a small hash table to distribute workload by	4106	C<ev_child> watchers use a small hash table to distribute workload by
3294	pid. The default size is C<16> (or C<1> with C<EV_MINIMAL>), usually more	4107	pid. The default size is C<16> (or C<1> with C<EV_FEATURES> disabled),
3295	than enough. If you need to manage thousands of children you might want to	4108	usually more than enough. If you need to manage thousands of children you
3296	increase this value (I<must> be a power of two).	4109	might want to increase this value (I<must> be a power of two).
3297		4110
3298	=item EV_INOTIFY_HASHSIZE	4111	=item EV_INOTIFY_HASHSIZE
3299		4112
3300	C<ev_stat> watchers use a small hash table to distribute workload by	4113	C<ev_stat> watchers use a small hash table to distribute workload by
3301	inotify watch id. The default size is C<16> (or C<1> with C<EV_MINIMAL>),	4114	inotify watch id. The default size is C<16> (or C<1> with C<EV_FEATURES>
3302	usually more than enough. If you need to manage thousands of C<ev_stat>	4115	disabled), usually more than enough. If you need to manage thousands of
3303	watchers you might want to increase this value (I<must> be a power of	4116	C<ev_stat> watchers you might want to increase this value (I<must> be a
3304	two).	4117	power of two).
3305		4118
3306	=item EV_USE_4HEAP	4119	=item EV_USE_4HEAP
3307		4120
3308	Heaps are not very cache-efficient. To improve the cache-efficiency of the	4121	Heaps are not very cache-efficient. To improve the cache-efficiency of the
3309	timer and periodics heaps, libev uses a 4-heap when this symbol is defined	4122	timer and periodics heaps, libev uses a 4-heap when this symbol is defined
3310	to C<1>. The 4-heap uses more complicated (longer) code but has noticeably	4123	to C<1>. The 4-heap uses more complicated (longer) code but has noticeably
3311	faster performance with many (thousands) of watchers.	4124	faster performance with many (thousands) of watchers.
3312		4125
3313	The default is C<1> unless C<EV_MINIMAL> is set in which case it is C<0>	4126	The default is C<1>, unless C<EV_FEATURES> overrides it, in which case it
3314	(disabled).	4127	will be C<0>.
3315		4128
3316	=item EV_HEAP_CACHE_AT	4129	=item EV_HEAP_CACHE_AT
3317		4130
3318	Heaps are not very cache-efficient. To improve the cache-efficiency of the	4131	Heaps are not very cache-efficient. To improve the cache-efficiency of the
3319	timer and periodics heaps, libev can cache the timestamp (I<at>) within	4132	timer and periodics heaps, libev can cache the timestamp (I<at>) within
3320	the heap structure (selected by defining C<EV_HEAP_CACHE_AT> to C<1>),	4133	the heap structure (selected by defining C<EV_HEAP_CACHE_AT> to C<1>),
3321	which uses 8-12 bytes more per watcher and a few hundred bytes more code,	4134	which uses 8-12 bytes more per watcher and a few hundred bytes more code,
3322	but avoids random read accesses on heap changes. This improves performance	4135	but avoids random read accesses on heap changes. This improves performance
3323	noticeably with many (hundreds) of watchers.	4136	noticeably with many (hundreds) of watchers.
3324		4137
3325	The default is C<1> unless C<EV_MINIMAL> is set in which case it is C<0>	4138	The default is C<1>, unless C<EV_FEATURES> overrides it, in which case it
3326	(disabled).	4139	will be C<0>.
3327		4140
3328	=item EV_VERIFY	4141	=item EV_VERIFY
3329		4142
3330	Controls how much internal verification (see C<ev_loop_verify ()>) will	4143	Controls how much internal verification (see C<ev_verify ()>) will
3331	be done: If set to C<0>, no internal verification code will be compiled	4144	be done: If set to C<0>, no internal verification code will be compiled
3332	in. If set to C<1>, then verification code will be compiled in, but not	4145	in. If set to C<1>, then verification code will be compiled in, but not
3333	called. If set to C<2>, then the internal verification code will be	4146	called. If set to C<2>, then the internal verification code will be
3334	called once per loop, which can slow down libev. If set to C<3>, then the	4147	called once per loop, which can slow down libev. If set to C<3>, then the
3335	verification code will be called very frequently, which will slow down	4148	verification code will be called very frequently, which will slow down
3336	libev considerably.	4149	libev considerably.
3337		4150
3338	The default is C<1>, unless C<EV_MINIMAL> is set, in which case it will be	4151	The default is C<1>, unless C<EV_FEATURES> overrides it, in which case it
3339	C<0>.	4152	will be C<0>.
3340		4153
3341	=item EV_COMMON	4154	=item EV_COMMON
3342		4155
3343	By default, all watchers have a C<void *data> member. By redefining	4156	By default, all watchers have a C<void *data> member. By redefining
3344	this macro to a something else you can include more and other types of	4157	this macro to something else you can include more and other types of
3345	members. You have to define it each time you include one of the files,	4158	members. You have to define it each time you include one of the files,
3346	though, and it must be identical each time.	4159	though, and it must be identical each time.
3347		4160
3348	For example, the perl EV module uses something like this:	4161	For example, the perl EV module uses something like this:
3349		4162
…		…
3402	file.	4215	file.
3403		4216
3404	The usage in rxvt-unicode is simpler. It has a F<ev_cpp.h> header file	4217	The usage in rxvt-unicode is simpler. It has a F<ev_cpp.h> header file
3405	that everybody includes and which overrides some configure choices:	4218	that everybody includes and which overrides some configure choices:
3406		4219
3407	#define EV_MINIMAL 1	4220	#define EV_FEATURES 8
3408	#define EV_USE_POLL 0	4221	#define EV_USE_SELECT 1
3409	#define EV_MULTIPLICITY 0
3410	#define EV_PERIODIC_ENABLE 0	4222	#define EV_PREPARE_ENABLE 1
		4223	#define EV_IDLE_ENABLE 1
3411	#define EV_STAT_ENABLE 0	4224	#define EV_SIGNAL_ENABLE 1
3412	#define EV_FORK_ENABLE 0	4225	#define EV_CHILD_ENABLE 1
		4226	#define EV_USE_STDEXCEPT 0
3413	#define EV_CONFIG_H <config.h>	4227	#define EV_CONFIG_H <config.h>
3414	#define EV_MINPRI 0
3415	#define EV_MAXPRI 0
3416		4228
3417	#include "ev++.h"	4229	#include "ev++.h"
3418		4230
3419	And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:	4231	And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:
3420		4232
…		…
3480	default loop and triggering an C<ev_async> watcher from the default loop	4292	default loop and triggering an C<ev_async> watcher from the default loop
3481	watcher callback into the event loop interested in the signal.	4293	watcher callback into the event loop interested in the signal.
3482		4294
3483	=back	4295	=back
3484		4296
		4297	=head4 THREAD LOCKING EXAMPLE
		4298
		4299	Here is a fictitious example of how to run an event loop in a different
		4300	thread than where callbacks are being invoked and watchers are
		4301	created/added/removed.
		4302
		4303	For a real-world example, see the C<EV::Loop::Async> perl module,
		4304	which uses exactly this technique (which is suited for many high-level
		4305	languages).
		4306
		4307	The example uses a pthread mutex to protect the loop data, a condition
		4308	variable to wait for callback invocations, an async watcher to notify the
		4309	event loop thread and an unspecified mechanism to wake up the main thread.
		4310
		4311	First, you need to associate some data with the event loop:
		4312
		4313	typedef struct {
		4314	mutex_t lock; /* global loop lock */
		4315	ev_async async_w;
		4316	thread_t tid;
		4317	cond_t invoke_cv;
		4318	} userdata;
		4319
		4320	void prepare_loop (EV_P)
		4321	{
		4322	// for simplicity, we use a static userdata struct.
		4323	static userdata u;
		4324
		4325	ev_async_init (&u->async_w, async_cb);
		4326	ev_async_start (EV_A_ &u->async_w);
		4327
		4328	pthread_mutex_init (&u->lock, 0);
		4329	pthread_cond_init (&u->invoke_cv, 0);
		4330
		4331	// now associate this with the loop
		4332	ev_set_userdata (EV_A_ u);
		4333	ev_set_invoke_pending_cb (EV_A_ l_invoke);
		4334	ev_set_loop_release_cb (EV_A_ l_release, l_acquire);
		4335
		4336	// then create the thread running ev_loop
		4337	pthread_create (&u->tid, 0, l_run, EV_A);
		4338	}
		4339
		4340	The callback for the C<ev_async> watcher does nothing: the watcher is used
		4341	solely to wake up the event loop so it takes notice of any new watchers
		4342	that might have been added:
		4343
		4344	static void
		4345	async_cb (EV_P_ ev_async *w, int revents)
		4346	{
		4347	// just used for the side effects
		4348	}
		4349
		4350	The C<l_release> and C<l_acquire> callbacks simply unlock/lock the mutex
		4351	protecting the loop data, respectively.
		4352
		4353	static void
		4354	l_release (EV_P)
		4355	{
		4356	userdata *u = ev_userdata (EV_A);
		4357	pthread_mutex_unlock (&u->lock);
		4358	}
		4359
		4360	static void
		4361	l_acquire (EV_P)
		4362	{
		4363	userdata *u = ev_userdata (EV_A);
		4364	pthread_mutex_lock (&u->lock);
		4365	}
		4366
		4367	The event loop thread first acquires the mutex, and then jumps straight
		4368	into C<ev_run>:
		4369
		4370	void *
		4371	l_run (void *thr_arg)
		4372	{
		4373	struct ev_loop *loop = (struct ev_loop *)thr_arg;
		4374
		4375	l_acquire (EV_A);
		4376	pthread_setcanceltype (PTHREAD_CANCEL_ASYNCHRONOUS, 0);
		4377	ev_run (EV_A_ 0);
		4378	l_release (EV_A);
		4379
		4380	return 0;
		4381	}
		4382
		4383	Instead of invoking all pending watchers, the C<l_invoke> callback will
		4384	signal the main thread via some unspecified mechanism (signals? pipe
		4385	writes? C<Async::Interrupt>?) and then waits until all pending watchers
		4386	have been called (in a while loop because a) spurious wakeups are possible
		4387	and b) skipping inter-thread-communication when there are no pending
		4388	watchers is very beneficial):
		4389
		4390	static void
		4391	l_invoke (EV_P)
		4392	{
		4393	userdata *u = ev_userdata (EV_A);
		4394
		4395	while (ev_pending_count (EV_A))
		4396	{
		4397	wake_up_other_thread_in_some_magic_or_not_so_magic_way ();
		4398	pthread_cond_wait (&u->invoke_cv, &u->lock);
		4399	}
		4400	}
		4401
		4402	Now, whenever the main thread gets told to invoke pending watchers, it
		4403	will grab the lock, call C<ev_invoke_pending> and then signal the loop
		4404	thread to continue:
		4405
		4406	static void
		4407	real_invoke_pending (EV_P)
		4408	{
		4409	userdata *u = ev_userdata (EV_A);
		4410
		4411	pthread_mutex_lock (&u->lock);
		4412	ev_invoke_pending (EV_A);
		4413	pthread_cond_signal (&u->invoke_cv);
		4414	pthread_mutex_unlock (&u->lock);
		4415	}
		4416
		4417	Whenever you want to start/stop a watcher or do other modifications to an
		4418	event loop, you will now have to lock:
		4419
		4420	ev_timer timeout_watcher;
		4421	userdata *u = ev_userdata (EV_A);
		4422
		4423	ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);
		4424
		4425	pthread_mutex_lock (&u->lock);
		4426	ev_timer_start (EV_A_ &timeout_watcher);
		4427	ev_async_send (EV_A_ &u->async_w);
		4428	pthread_mutex_unlock (&u->lock);
		4429
		4430	Note that sending the C<ev_async> watcher is required because otherwise
		4431	an event loop currently blocking in the kernel will have no knowledge
		4432	about the newly added timer. By waking up the loop it will pick up any new
		4433	watchers in the next event loop iteration.
		4434
3485	=head3 COROUTINES	4435	=head3 COROUTINES
3486		4436
3487	Libev is very accommodating to coroutines ("cooperative threads"):	4437	Libev is very accommodating to coroutines ("cooperative threads"):
3488	libev fully supports nesting calls to its functions from different	4438	libev fully supports nesting calls to its functions from different
3489	coroutines (e.g. you can call C<ev_loop> on the same loop from two	4439	coroutines (e.g. you can call C<ev_run> on the same loop from two
3490	different coroutines, and switch freely between both coroutines running the	4440	different coroutines, and switch freely between both coroutines running
3491	loop, as long as you don't confuse yourself). The only exception is that	4441	the loop, as long as you don't confuse yourself). The only exception is
3492	you must not do this from C<ev_periodic> reschedule callbacks.	4442	that you must not do this from C<ev_periodic> reschedule callbacks.
3493		4443
3494	Care has been taken to ensure that libev does not keep local state inside	4444	Care has been taken to ensure that libev does not keep local state inside
3495	C<ev_loop>, and other calls do not usually allow for coroutine switches as	4445	C<ev_run>, and other calls do not usually allow for coroutine switches as
3496	they do not clal any callbacks.	4446	they do not call any callbacks.
3497		4447
3498	=head2 COMPILER WARNINGS	4448	=head2 COMPILER WARNINGS
3499		4449
3500	Depending on your compiler and compiler settings, you might get no or a	4450	Depending on your compiler and compiler settings, you might get no or a
3501	lot of warnings when compiling libev code. Some people are apparently	4451	lot of warnings when compiling libev code. Some people are apparently
…		…
3511	maintainable.	4461	maintainable.
3512		4462
3513	And of course, some compiler warnings are just plain stupid, or simply	4463	And of course, some compiler warnings are just plain stupid, or simply
3514	wrong (because they don't actually warn about the condition their message	4464	wrong (because they don't actually warn about the condition their message
3515	seems to warn about). For example, certain older gcc versions had some	4465	seems to warn about). For example, certain older gcc versions had some
3516	warnings that resulted an extreme number of false positives. These have	4466	warnings that resulted in an extreme number of false positives. These have
3517	been fixed, but some people still insist on making code warn-free with	4467	been fixed, but some people still insist on making code warn-free with
3518	such buggy versions.	4468	such buggy versions.
3519		4469
3520	While libev is written to generate as few warnings as possible,	4470	While libev is written to generate as few warnings as possible,
3521	"warn-free" code is not a goal, and it is recommended not to build libev	4471	"warn-free" code is not a goal, and it is recommended not to build libev
…		…
3535	==2274== definitely lost: 0 bytes in 0 blocks.	4485	==2274== definitely lost: 0 bytes in 0 blocks.
3536	==2274== possibly lost: 0 bytes in 0 blocks.	4486	==2274== possibly lost: 0 bytes in 0 blocks.
3537	==2274== still reachable: 256 bytes in 1 blocks.	4487	==2274== still reachable: 256 bytes in 1 blocks.
3538		4488
3539	Then there is no memory leak, just as memory accounted to global variables	4489	Then there is no memory leak, just as memory accounted to global variables
3540	is not a memleak - the memory is still being refernced, and didn't leak.	4490	is not a memleak - the memory is still being referenced, and didn't leak.
3541		4491
3542	Similarly, under some circumstances, valgrind might report kernel bugs	4492	Similarly, under some circumstances, valgrind might report kernel bugs
3543	as if it were a bug in libev (e.g. in realloc or in the poll backend,	4493	as if it were a bug in libev (e.g. in realloc or in the poll backend,
3544	although an acceptable workaround has been found here), or it might be	4494	although an acceptable workaround has been found here), or it might be
3545	confused.	4495	confused.
…		…
3557	I suggest using suppression lists.	4507	I suggest using suppression lists.
3558		4508
3559		4509
3560	=head1 PORTABILITY NOTES	4510	=head1 PORTABILITY NOTES
3561		4511
		4512	=head2 GNU/LINUX 32 BIT LIMITATIONS
		4513
		4514	GNU/Linux is the only common platform that supports 64 bit file/large file
		4515	interfaces but I<disables> them by default.
		4516
		4517	That means that libev compiled in the default environment doesn't support
		4518	files larger than 2GiB or so, which mainly affects C<ev_stat> watchers.
		4519
		4520	Unfortunately, many programs try to work around this GNU/Linux issue
		4521	by enabling the large file API, which makes them incompatible with the
		4522	standard libev compiled for their system.
		4523
		4524	Likewise, libev cannot enable the large file API itself as this would
		4525	suddenly make it incompatible to the default compile time environment,
		4526	i.e. all programs not using special compile switches.
		4527
		4528	=head2 OS/X AND DARWIN BUGS
		4529
		4530	The whole thing is a bug if you ask me - basically any system interface
		4531	you touch is broken, whether it is locales, poll, kqueue or even the
		4532	OpenGL drivers.
		4533
		4534	=head3 C<kqueue> is buggy
		4535
		4536	The kqueue syscall is broken in all known versions - most versions support
		4537	only sockets, many support pipes.
		4538
		4539	Libev tries to work around this by not using C<kqueue> by default on this
		4540	rotten platform, but of course you can still ask for it when creating a
		4541	loop - embedding a socket-only kqueue loop into a select-based one is
		4542	probably going to work well.
		4543
		4544	=head3 C<poll> is buggy
		4545
		4546	Instead of fixing C<kqueue>, Apple replaced their (working) C<poll>
		4547	implementation by something calling C<kqueue> internally around the 10.5.6
		4548	release, so now C<kqueue> I<and> C<poll> are broken.
		4549
		4550	Libev tries to work around this by not using C<poll> by default on
		4551	this rotten platform, but of course you can still ask for it when creating
		4552	a loop.
		4553
		4554	=head3 C<select> is buggy
		4555
		4556	All that's left is C<select>, and of course Apple found a way to fuck this
		4557	one up as well: On OS/X, C<select> actively limits the number of file
		4558	descriptors you can pass in to 1024 - your program suddenly crashes when
		4559	you use more.
		4560
		4561	There is an undocumented "workaround" for this - defining
		4562	C<_DARWIN_UNLIMITED_SELECT>, which libev tries to use, so select I<should>
		4563	work on OS/X.
		4564
		4565	=head2 SOLARIS PROBLEMS AND WORKAROUNDS
		4566
		4567	=head3 C<errno> reentrancy
		4568
		4569	The default compile environment on Solaris is unfortunately so
		4570	thread-unsafe that you can't even use components/libraries compiled
		4571	without C<-D_REENTRANT> in a threaded program, which, of course, isn't
		4572	defined by default. A valid, if stupid, implementation choice.
		4573
		4574	If you want to use libev in threaded environments you have to make sure
		4575	it's compiled with C<_REENTRANT> defined.
		4576
		4577	=head3 Event port backend
		4578
		4579	The scalable event interface for Solaris is called "event
		4580	ports". Unfortunately, this mechanism is very buggy in all major
		4581	releases. If you run into high CPU usage, your program freezes or you get
		4582	a large number of spurious wakeups, make sure you have all the relevant
		4583	and latest kernel patches applied. No, I don't know which ones, but there
		4584	are multiple ones to apply, and afterwards, event ports actually work
		4585	great.
		4586
		4587	If you can't get it to work, you can try running the program by setting
		4588	the environment variable C<LIBEV_FLAGS=3> to only allow C<poll> and
		4589	C<select> backends.
		4590
		4591	=head2 AIX POLL BUG
		4592
		4593	AIX unfortunately has a broken C<poll.h> header. Libev works around
		4594	this by trying to avoid the poll backend altogether (i.e. it's not even
		4595	compiled in), which normally isn't a big problem as C<select> works fine
		4596	with large bitsets on AIX, and AIX is dead anyway.
		4597
3562	=head2 WIN32 PLATFORM LIMITATIONS AND WORKAROUNDS	4598	=head2 WIN32 PLATFORM LIMITATIONS AND WORKAROUNDS
		4599
		4600	=head3 General issues
3563		4601
3564	Win32 doesn't support any of the standards (e.g. POSIX) that libev	4602	Win32 doesn't support any of the standards (e.g. POSIX) that libev
3565	requires, and its I/O model is fundamentally incompatible with the POSIX	4603	requires, and its I/O model is fundamentally incompatible with the POSIX
3566	model. Libev still offers limited functionality on this platform in	4604	model. Libev still offers limited functionality on this platform in
3567	the form of the C<EVBACKEND_SELECT> backend, and only supports socket	4605	the form of the C<EVBACKEND_SELECT> backend, and only supports socket
3568	descriptors. This only applies when using Win32 natively, not when using	4606	descriptors. This only applies when using Win32 natively, not when using
3569	e.g. cygwin.	4607	e.g. cygwin. Actually, it only applies to the microsofts own compilers,
		4608	as every compielr comes with a slightly differently broken/incompatible
		4609	environment.
3570		4610
3571	Lifting these limitations would basically require the full	4611	Lifting these limitations would basically require the full
3572	re-implementation of the I/O system. If you are into these kinds of	4612	re-implementation of the I/O system. If you are into this kind of thing,
3573	things, then note that glib does exactly that for you in a very portable	4613	then note that glib does exactly that for you in a very portable way (note
3574	way (note also that glib is the slowest event library known to man).	4614	also that glib is the slowest event library known to man).
3575		4615
3576	There is no supported compilation method available on windows except	4616	There is no supported compilation method available on windows except
3577	embedding it into other applications.	4617	embedding it into other applications.
		4618
		4619	Sensible signal handling is officially unsupported by Microsoft - libev
		4620	tries its best, but under most conditions, signals will simply not work.
3578		4621
3579	Not a libev limitation but worth mentioning: windows apparently doesn't	4622	Not a libev limitation but worth mentioning: windows apparently doesn't
3580	accept large writes: instead of resulting in a partial write, windows will	4623	accept large writes: instead of resulting in a partial write, windows will
3581	either accept everything or return C<ENOBUFS> if the buffer is too large,	4624	either accept everything or return C<ENOBUFS> if the buffer is too large,
3582	so make sure you only write small amounts into your sockets (less than a	4625	so make sure you only write small amounts into your sockets (less than a
…		…
3587	the abysmal performance of winsockets, using a large number of sockets	4630	the abysmal performance of winsockets, using a large number of sockets
3588	is not recommended (and not reasonable). If your program needs to use	4631	is not recommended (and not reasonable). If your program needs to use
3589	more than a hundred or so sockets, then likely it needs to use a totally	4632	more than a hundred or so sockets, then likely it needs to use a totally
3590	different implementation for windows, as libev offers the POSIX readiness	4633	different implementation for windows, as libev offers the POSIX readiness
3591	notification model, which cannot be implemented efficiently on windows	4634	notification model, which cannot be implemented efficiently on windows
3592	(Microsoft monopoly games).	4635	(due to Microsoft monopoly games).
3593		4636
3594	A typical way to use libev under windows is to embed it (see the embedding	4637	A typical way to use libev under windows is to embed it (see the embedding
3595	section for details) and use the following F<evwrap.h> header file instead	4638	section for details) and use the following F<evwrap.h> header file instead
3596	of F<ev.h>:	4639	of F<ev.h>:
3597		4640
…		…
3604	you do I<not> compile the F<ev.c> or any other embedded source files!):	4647	you do I<not> compile the F<ev.c> or any other embedded source files!):
3605		4648
3606	#include "evwrap.h"	4649	#include "evwrap.h"
3607	#include "ev.c"	4650	#include "ev.c"
3608		4651
3609	=over 4
3610
3611	=item The winsocket select function	4652	=head3 The winsocket C<select> function
3612		4653
3613	The winsocket C<select> function doesn't follow POSIX in that it	4654	The winsocket C<select> function doesn't follow POSIX in that it
3614	requires socket I<handles> and not socket I<file descriptors> (it is	4655	requires socket I<handles> and not socket I<file descriptors> (it is
3615	also extremely buggy). This makes select very inefficient, and also	4656	also extremely buggy). This makes select very inefficient, and also
3616	requires a mapping from file descriptors to socket handles (the Microsoft	4657	requires a mapping from file descriptors to socket handles (the Microsoft
…		…
3625	#define EV_SELECT_IS_WINSOCKET 1 /* forces EV_SELECT_USE_FD_SET, too */	4666	#define EV_SELECT_IS_WINSOCKET 1 /* forces EV_SELECT_USE_FD_SET, too */
3626		4667
3627	Note that winsockets handling of fd sets is O(n), so you can easily get a	4668	Note that winsockets handling of fd sets is O(n), so you can easily get a
3628	complexity in the O(n²) range when using win32.	4669	complexity in the O(n²) range when using win32.
3629		4670
3630	=item Limited number of file descriptors	4671	=head3 Limited number of file descriptors
3631		4672
3632	Windows has numerous arbitrary (and low) limits on things.	4673	Windows has numerous arbitrary (and low) limits on things.
3633		4674
3634	Early versions of winsocket's select only supported waiting for a maximum	4675	Early versions of winsocket's select only supported waiting for a maximum
3635	of C<64> handles (probably owning to the fact that all windows kernels	4676	of C<64> handles (probably owning to the fact that all windows kernels
3636	can only wait for C<64> things at the same time internally; Microsoft	4677	can only wait for C<64> things at the same time internally; Microsoft
3637	recommends spawning a chain of threads and wait for 63 handles and the	4678	recommends spawning a chain of threads and wait for 63 handles and the
3638	previous thread in each. Great).	4679	previous thread in each. Sounds great!).
3639		4680
3640	Newer versions support more handles, but you need to define C<FD_SETSIZE>	4681	Newer versions support more handles, but you need to define C<FD_SETSIZE>
3641	to some high number (e.g. C<2048>) before compiling the winsocket select	4682	to some high number (e.g. C<2048>) before compiling the winsocket select
3642	call (which might be in libev or elsewhere, for example, perl does its own	4683	call (which might be in libev or elsewhere, for example, perl and many
3643	select emulation on windows).	4684	other interpreters do their own select emulation on windows).
3644		4685
3645	Another limit is the number of file descriptors in the Microsoft runtime	4686	Another limit is the number of file descriptors in the Microsoft runtime
3646	libraries, which by default is C<64> (there must be a hidden I<64> fetish	4687	libraries, which by default is C<64> (there must be a hidden I<64>
3647	or something like this inside Microsoft). You can increase this by calling	4688	fetish or something like this inside Microsoft). You can increase this
3648	C<_setmaxstdio>, which can increase this limit to C<2048> (another	4689	by calling C<_setmaxstdio>, which can increase this limit to C<2048>
3649	arbitrary limit), but is broken in many versions of the Microsoft runtime	4690	(another arbitrary limit), but is broken in many versions of the Microsoft
3650	libraries.
3651
3652	This might get you to about C<512> or C<2048> sockets (depending on	4691	runtime libraries. This might get you to about C<512> or C<2048> sockets
3653	windows version and/or the phase of the moon). To get more, you need to	4692	(depending on windows version and/or the phase of the moon). To get more,
3654	wrap all I/O functions and provide your own fd management, but the cost of	4693	you need to wrap all I/O functions and provide your own fd management, but
3655	calling select (O(n²)) will likely make this unworkable.	4694	the cost of calling select (O(n²)) will likely make this unworkable.
3656
3657	=back
3658		4695
3659	=head2 PORTABILITY REQUIREMENTS	4696	=head2 PORTABILITY REQUIREMENTS
3660		4697
3661	In addition to a working ISO-C implementation and of course the	4698	In addition to a working ISO-C implementation and of course the
3662	backend-specific APIs, libev relies on a few additional extensions:	4699	backend-specific APIs, libev relies on a few additional extensions:
…		…
3701	watchers.	4738	watchers.
3702		4739
3703	=item C<double> must hold a time value in seconds with enough accuracy	4740	=item C<double> must hold a time value in seconds with enough accuracy
3704		4741
3705	The type C<double> is used to represent timestamps. It is required to	4742	The type C<double> is used to represent timestamps. It is required to
3706	have at least 51 bits of mantissa (and 9 bits of exponent), which is good	4743	have at least 51 bits of mantissa (and 9 bits of exponent), which is
3707	enough for at least into the year 4000. This requirement is fulfilled by	4744	good enough for at least into the year 4000 with millisecond accuracy
		4745	(the design goal for libev). This requirement is overfulfilled by
3708	implementations implementing IEEE 754 (basically all existing ones).	4746	implementations using IEEE 754, which is basically all existing ones. With
		4747	IEEE 754 doubles, you get microsecond accuracy until at least 2200.
3709		4748
3710	=back	4749	=back
3711		4750
3712	If you know of other additional requirements drop me a note.	4751	If you know of other additional requirements drop me a note.
3713		4752
…		…
3781	involves iterating over all running async watchers or all signal numbers.	4820	involves iterating over all running async watchers or all signal numbers.
3782		4821
3783	=back	4822	=back
3784		4823
3785		4824
		4825	=head1 PORTING FROM LIBEV 3.X TO 4.X
		4826
		4827	The major version 4 introduced some minor incompatible changes to the API.
		4828
		4829	At the moment, the C<ev.h> header file tries to implement superficial
		4830	compatibility, so most programs should still compile. Those might be
		4831	removed in later versions of libev, so better update early than late.
		4832
		4833	=over 4
		4834
		4835	=item function/symbol renames
		4836
		4837	A number of functions and symbols have been renamed:
		4838
		4839	ev_loop => ev_run
		4840	EVLOOP_NONBLOCK => EVRUN_NOWAIT
		4841	EVLOOP_ONESHOT => EVRUN_ONCE
		4842
		4843	ev_unloop => ev_break
		4844	EVUNLOOP_CANCEL => EVBREAK_CANCEL
		4845	EVUNLOOP_ONE => EVBREAK_ONE
		4846	EVUNLOOP_ALL => EVBREAK_ALL
		4847
		4848	EV_TIMEOUT => EV_TIMER
		4849
		4850	ev_loop_count => ev_iteration
		4851	ev_loop_depth => ev_depth
		4852	ev_loop_verify => ev_verify
		4853
		4854	Most functions working on C<struct ev_loop> objects don't have an
		4855	C<ev_loop_> prefix, so it was removed; C<ev_loop>, C<ev_unloop> and
		4856	associated constants have been renamed to not collide with the C<struct
		4857	ev_loop> anymore and C<EV_TIMER> now follows the same naming scheme
		4858	as all other watcher types. Note that C<ev_loop_fork> is still called
		4859	C<ev_loop_fork> because it would otherwise clash with the C<ev_fork>
		4860	typedef.
		4861
		4862	=item C<EV_COMPAT3> backwards compatibility mechanism
		4863
		4864	The backward compatibility mechanism can be controlled by
		4865	C<EV_COMPAT3>. See L<PREPROCESSOR SYMBOLS/MACROS> in the L<EMBEDDING>
		4866	section.
		4867
		4868	=item C<EV_MINIMAL> mechanism replaced by C<EV_FEATURES>
		4869
		4870	The preprocessor symbol C<EV_MINIMAL> has been replaced by a different
		4871	mechanism, C<EV_FEATURES>. Programs using C<EV_MINIMAL> usually compile
		4872	and work, but the library code will of course be larger.
		4873
		4874	=back
		4875
		4876
		4877	=head1 GLOSSARY
		4878
		4879	=over 4
		4880
		4881	=item active
		4882
		4883	A watcher is active as long as it has been started and not yet stopped.
		4884	See L<WATCHER STATES> for details.
		4885
		4886	=item application
		4887
		4888	In this document, an application is whatever is using libev.
		4889
		4890	=item backend
		4891
		4892	The part of the code dealing with the operating system interfaces.
		4893
		4894	=item callback
		4895
		4896	The address of a function that is called when some event has been
		4897	detected. Callbacks are being passed the event loop, the watcher that
		4898	received the event, and the actual event bitset.
		4899
		4900	=item callback/watcher invocation
		4901
		4902	The act of calling the callback associated with a watcher.
		4903
		4904	=item event
		4905
		4906	A change of state of some external event, such as data now being available
		4907	for reading on a file descriptor, time having passed or simply not having
		4908	any other events happening anymore.
		4909
		4910	In libev, events are represented as single bits (such as C<EV_READ> or
		4911	C<EV_TIMER>).
		4912
		4913	=item event library
		4914
		4915	A software package implementing an event model and loop.
		4916
		4917	=item event loop
		4918
		4919	An entity that handles and processes external events and converts them
		4920	into callback invocations.
		4921
		4922	=item event model
		4923
		4924	The model used to describe how an event loop handles and processes
		4925	watchers and events.
		4926
		4927	=item pending
		4928
		4929	A watcher is pending as soon as the corresponding event has been
		4930	detected. See L<WATCHER STATES> for details.
		4931
		4932	=item real time
		4933
		4934	The physical time that is observed. It is apparently strictly monotonic :)
		4935
		4936	=item wall-clock time
		4937
		4938	The time and date as shown on clocks. Unlike real time, it can actually
		4939	be wrong and jump forwards and backwards, e.g. when the you adjust your
		4940	clock.
		4941
		4942	=item watcher
		4943
		4944	A data structure that describes interest in certain events. Watchers need
		4945	to be started (attached to an event loop) before they can receive events.
		4946
		4947	=back
		4948
3786	=head1 AUTHOR	4949	=head1 AUTHOR
3787		4950
3788	Marc Lehmann <libev@schmorp.de>.	4951	Marc Lehmann <libev@schmorp.de>, with repeated corrections by Mikael Magnusson.
3789		4952

Diff Legend

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