ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/libev/ev.pod

Comparing libev/ev.pod (file contents):
Revision 1.30 by root, Fri Nov 23 04:36:03 2007 UTC vs.
Revision 1.115 by root, Mon Dec 31 01:32:59 2007 UTC

…		…
4		4
5	=head1 SYNOPSIS	5	=head1 SYNOPSIS
6		6
7	#include <ev.h>	7	#include <ev.h>
8		8
		9	=head2 EXAMPLE PROGRAM
		10
		11	#include <ev.h>
		12
		13	ev_io stdin_watcher;
		14	ev_timer timeout_watcher;
		15
		16	/* called when data readable on stdin */
		17	static void
		18	stdin_cb (EV_P_ struct ev_io *w, int revents)
		19	{
		20	/* puts ("stdin ready"); */
		21	ev_io_stop (EV_A_ w); /* just a syntax example */
		22	ev_unloop (EV_A_ EVUNLOOP_ALL); /* leave all loop calls */
		23	}
		24
		25	static void
		26	timeout_cb (EV_P_ struct ev_timer *w, int revents)
		27	{
		28	/* puts ("timeout"); */
		29	ev_unloop (EV_A_ EVUNLOOP_ONE); /* leave one loop call */
		30	}
		31
		32	int
		33	main (void)
		34	{
		35	struct ev_loop *loop = ev_default_loop (0);
		36
		37	/* initialise an io watcher, then start it */
		38	ev_io_init (&stdin_watcher, stdin_cb, /*STDIN_FILENO*/ 0, EV_READ);
		39	ev_io_start (loop, &stdin_watcher);
		40
		41	/* simple non-repeating 5.5 second timeout */
		42	ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);
		43	ev_timer_start (loop, &timeout_watcher);
		44
		45	/* loop till timeout or data ready */
		46	ev_loop (loop, 0);
		47
		48	return 0;
		49	}
		50
9	=head1 DESCRIPTION	51	=head1 DESCRIPTION
10		52
		53	The newest version of this document is also available as a html-formatted
		54	web page you might find easier to navigate when reading it for the first
		55	time: L<http://cvs.schmorp.de/libev/ev.html>.
		56
11	Libev is an event loop: you register interest in certain events (such as a	57	Libev is an event loop: you register interest in certain events (such as a
12	file descriptor being readable or a timeout occuring), and it will manage	58	file descriptor being readable or a timeout occurring), and it will manage
13	these event sources and provide your program with events.	59	these event sources and provide your program with events.
14		60
15	To do this, it must take more or less complete control over your process	61	To do this, it must take more or less complete control over your process
16	(or thread) by executing the I<event loop> handler, and will then	62	(or thread) by executing the I<event loop> handler, and will then
17	communicate events via a callback mechanism.	63	communicate events via a callback mechanism.
…		…
19	You register interest in certain events by registering so-called I<event	65	You register interest in certain events by registering so-called I<event
20	watchers>, which are relatively small C structures you initialise with the	66	watchers>, which are relatively small C structures you initialise with the
21	details of the event, and then hand it over to libev by I<starting> the	67	details of the event, and then hand it over to libev by I<starting> the
22	watcher.	68	watcher.
23		69
24	=head1 FEATURES	70	=head2 FEATURES
25		71
26	Libev supports select, poll, the linux-specific epoll and the bsd-specific	72	Libev supports C<select>, C<poll>, the Linux-specific C<epoll>, the
27	kqueue mechanisms for file descriptor events, relative timers, absolute	73	BSD-specific C<kqueue> and the Solaris-specific event port mechanisms
28	timers with customised rescheduling, signal events, process status change	74	for file descriptor events (C<ev_io>), the Linux C<inotify> interface
29	events (related to SIGCHLD), and event watchers dealing with the event	75	(for C<ev_stat>), relative timers (C<ev_timer>), absolute timers
30	loop mechanism itself (idle, prepare and check watchers). It also is quite	76	with customised rescheduling (C<ev_periodic>), synchronous signals
		77	(C<ev_signal>), process status change events (C<ev_child>), and event
		78	watchers dealing with the event loop mechanism itself (C<ev_idle>,
		79	C<ev_embed>, C<ev_prepare> and C<ev_check> watchers) as well as
		80	file watchers (C<ev_stat>) and even limited support for fork events
		81	(C<ev_fork>).
		82
		83	It also is quite fast (see this
31	fast (see this L<benchmark|http://libev.schmorp.de/bench.html> comparing	84	L<benchmark|http://libev.schmorp.de/bench.html> comparing it to libevent
32	it to libevent for example).	85	for example).
33		86
34	=head1 CONVENTIONS	87	=head2 CONVENTIONS
35		88
36	Libev is very configurable. In this manual the default configuration	89	Libev is very configurable. In this manual the default configuration will
37	will be described, which supports multiple event loops. For more info	90	be described, which supports multiple event loops. For more info about
38	about various configuration options please have a look at the file	91	various configuration options please have a look at B<EMBED> section in
39	F<README.embed> in the libev distribution. If libev was configured without	92	this manual. If libev was configured without support for multiple event
40	support for multiple event loops, then all functions taking an initial	93	loops, then all functions taking an initial argument of name C<loop>
41	argument of name C<loop> (which is always of type C<struct ev_loop *>)	94	(which is always of type C<struct ev_loop *>) will not have this argument.
42	will not have this argument.
43		95
44	=head1 TIME REPRESENTATION	96	=head2 TIME REPRESENTATION
45		97
46	Libev represents time as a single floating point number, representing the	98	Libev represents time as a single floating point number, representing the
47	(fractional) number of seconds since the (POSIX) epoch (somewhere near	99	(fractional) number of seconds since the (POSIX) epoch (somewhere near
48	the beginning of 1970, details are complicated, don't ask). This type is	100	the beginning of 1970, details are complicated, don't ask). This type is
49	called C<ev_tstamp>, which is what you should use too. It usually aliases	101	called C<ev_tstamp>, which is what you should use too. It usually aliases
50	to the double type in C.	102	to the C<double> type in C, and when you need to do any calculations on
		103	it, you should treat it as some floatingpoint value. Unlike the name
		104	component C<stamp> might indicate, it is also used for time differences
		105	throughout libev.
51		106
52	=head1 GLOBAL FUNCTIONS	107	=head1 GLOBAL FUNCTIONS
53		108
54	These functions can be called anytime, even before initialising the	109	These functions can be called anytime, even before initialising the
55	library in any way.	110	library in any way.
…		…
60		115
61	Returns the current time as libev would use it. Please note that the	116	Returns the current time as libev would use it. Please note that the
62	C<ev_now> function is usually faster and also often returns the timestamp	117	C<ev_now> function is usually faster and also often returns the timestamp
63	you actually want to know.	118	you actually want to know.
64		119
		120	=item ev_sleep (ev_tstamp interval)
		121
		122	Sleep for the given interval: The current thread will be blocked until
		123	either it is interrupted or the given time interval has passed. Basically
		124	this is a subsecond-resolution C<sleep ()>.
		125
65	=item int ev_version_major ()	126	=item int ev_version_major ()
66		127
67	=item int ev_version_minor ()	128	=item int ev_version_minor ()
68		129
69	You can find out the major and minor version numbers of the library	130	You can find out the major and minor ABI version numbers of the library
70	you linked against by calling the functions C<ev_version_major> and	131	you linked against by calling the functions C<ev_version_major> and
71	C<ev_version_minor>. If you want, you can compare against the global	132	C<ev_version_minor>. If you want, you can compare against the global
72	symbols C<EV_VERSION_MAJOR> and C<EV_VERSION_MINOR>, which specify the	133	symbols C<EV_VERSION_MAJOR> and C<EV_VERSION_MINOR>, which specify the
73	version of the library your program was compiled against.	134	version of the library your program was compiled against.
74		135
		136	These version numbers refer to the ABI version of the library, not the
		137	release version.
		138
75	Usually, it's a good idea to terminate if the major versions mismatch,	139	Usually, it's a good idea to terminate if the major versions mismatch,
76	as this indicates an incompatible change. Minor versions are usually	140	as this indicates an incompatible change. Minor versions are usually
77	compatible to older versions, so a larger minor version alone is usually	141	compatible to older versions, so a larger minor version alone is usually
78	not a problem.	142	not a problem.
79		143
		144	Example: Make sure we haven't accidentally been linked against the wrong
		145	version.
		146
		147	assert (("libev version mismatch",
		148	ev_version_major () == EV_VERSION_MAJOR
		149	&& ev_version_minor () >= EV_VERSION_MINOR));
		150
		151	=item unsigned int ev_supported_backends ()
		152
		153	Return the set of all backends (i.e. their corresponding C<EV_BACKEND_*>
		154	value) compiled into this binary of libev (independent of their
		155	availability on the system you are running on). See C<ev_default_loop> for
		156	a description of the set values.
		157
		158	Example: make sure we have the epoll method, because yeah this is cool and
		159	a must have and can we have a torrent of it please!!!11
		160
		161	assert (("sorry, no epoll, no sex",
		162	ev_supported_backends () & EVBACKEND_EPOLL));
		163
		164	=item unsigned int ev_recommended_backends ()
		165
		166	Return the set of all backends compiled into this binary of libev and also
		167	recommended for this platform. This set is often smaller than the one
		168	returned by C<ev_supported_backends>, as for example kqueue is broken on
		169	most BSDs and will not be autodetected unless you explicitly request it
		170	(assuming you know what you are doing). This is the set of backends that
		171	libev will probe for if you specify no backends explicitly.
		172
		173	=item unsigned int ev_embeddable_backends ()
		174
		175	Returns the set of backends that are embeddable in other event loops. This
		176	is the theoretical, all-platform, value. To find which backends
		177	might be supported on the current system, you would need to look at
		178	C<ev_embeddable_backends () & ev_supported_backends ()>, likewise for
		179	recommended ones.
		180
		181	See the description of C<ev_embed> watchers for more info.
		182
80	=item ev_set_allocator (void *(*cb)(void *ptr, long size))	183	=item ev_set_allocator (void *(*cb)(void *ptr, long size))
81		184
82	Sets the allocation function to use (the prototype is similar to the	185	Sets the allocation function to use (the prototype is similar - the
83	realloc C function, the semantics are identical). It is used to allocate	186	semantics is identical - to the realloc C function). It is used to
84	and free memory (no surprises here). If it returns zero when memory	187	allocate and free memory (no surprises here). If it returns zero when
85	needs to be allocated, the library might abort or take some potentially	188	memory needs to be allocated, the library might abort or take some
86	destructive action. The default is your system realloc function.	189	potentially destructive action. The default is your system realloc
		190	function.
87		191
88	You could override this function in high-availability programs to, say,	192	You could override this function in high-availability programs to, say,
89	free some memory if it cannot allocate memory, to use a special allocator,	193	free some memory if it cannot allocate memory, to use a special allocator,
90	or even to sleep a while and retry until some memory is available.	194	or even to sleep a while and retry until some memory is available.
		195
		196	Example: Replace the libev allocator with one that waits a bit and then
		197	retries).
		198
		199	static void *
		200	persistent_realloc (void *ptr, size_t size)
		201	{
		202	for (;;)
		203	{
		204	void *newptr = realloc (ptr, size);
		205
		206	if (newptr)
		207	return newptr;
		208
		209	sleep (60);
		210	}
		211	}
		212
		213	...
		214	ev_set_allocator (persistent_realloc);
91		215
92	=item ev_set_syserr_cb (void (*cb)(const char *msg));	216	=item ev_set_syserr_cb (void (*cb)(const char *msg));
93		217
94	Set the callback function to call on a retryable syscall error (such	218	Set the callback function to call on a retryable syscall error (such
95	as failed select, poll, epoll_wait). The message is a printable string	219	as failed select, poll, epoll_wait). The message is a printable string
…		…
97	callback is set, then libev will expect it to remedy the sitution, no	221	callback is set, then libev will expect it to remedy the sitution, no
98	matter what, when it returns. That is, libev will generally retry the	222	matter what, when it returns. That is, libev will generally retry the
99	requested operation, or, if the condition doesn't go away, do bad stuff	223	requested operation, or, if the condition doesn't go away, do bad stuff
100	(such as abort).	224	(such as abort).
101		225
		226	Example: This is basically the same thing that libev does internally, too.
		227
		228	static void
		229	fatal_error (const char *msg)
		230	{
		231	perror (msg);
		232	abort ();
		233	}
		234
		235	...
		236	ev_set_syserr_cb (fatal_error);
		237
102	=back	238	=back
103		239
104	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP	240	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP
105		241
106	An event loop is described by a C<struct ev_loop *>. The library knows two	242	An event loop is described by a C<struct ev_loop *>. The library knows two
…		…
119	=item struct ev_loop *ev_default_loop (unsigned int flags)	255	=item struct ev_loop *ev_default_loop (unsigned int flags)
120		256
121	This will initialise the default event loop if it hasn't been initialised	257	This will initialise the default event loop if it hasn't been initialised
122	yet and return it. If the default loop could not be initialised, returns	258	yet and return it. If the default loop could not be initialised, returns
123	false. If it already was initialised it simply returns it (and ignores the	259	false. If it already was initialised it simply returns it (and ignores the
124	flags).	260	flags. If that is troubling you, check C<ev_backend ()> afterwards).
125		261
126	If you don't know what event loop to use, use the one returned from this	262	If you don't know what event loop to use, use the one returned from this
127	function.	263	function.
128		264
129	The flags argument can be used to specify special behaviour or specific	265	The flags argument can be used to specify special behaviour or specific
130	backends to use, and is usually specified as 0 (or EVFLAG_AUTO).	266	backends to use, and is usually specified as C<0> (or C<EVFLAG_AUTO>).
131		267
132	It supports the following flags:	268	The following flags are supported:
133		269
134	=over 4	270	=over 4
135		271
136	=item C<EVFLAG_AUTO>	272	=item C<EVFLAG_AUTO>
137		273
…		…
145	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will	281	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will
146	override the flags completely if it is found in the environment. This is	282	override the flags completely if it is found in the environment. This is
147	useful to try out specific backends to test their performance, or to work	283	useful to try out specific backends to test their performance, or to work
148	around bugs.	284	around bugs.
149		285
		286	=item C<EVFLAG_FORKCHECK>
		287
		288	Instead of calling C<ev_default_fork> or C<ev_loop_fork> manually after
		289	a fork, you can also make libev check for a fork in each iteration by
		290	enabling this flag.
		291
		292	This works by calling C<getpid ()> on every iteration of the loop,
		293	and thus this might slow down your event loop if you do a lot of loop
		294	iterations and little real work, but is usually not noticeable (on my
		295	Linux system for example, C<getpid> is actually a simple 5-insn sequence
		296	without a syscall and thus I<very> fast, but my Linux system also has
		297	C<pthread_atfork> which is even faster).
		298
		299	The big advantage of this flag is that you can forget about fork (and
		300	forget about forgetting to tell libev about forking) when you use this
		301	flag.
		302
		303	This flag setting cannot be overriden or specified in the C<LIBEV_FLAGS>
		304	environment variable.
		305
150	=item C<EVMETHOD_SELECT> (value 1, portable select backend)	306	=item C<EVBACKEND_SELECT> (value 1, portable select backend)
151		307
152	This is your standard select(2) backend. Not I<completely> standard, as	308	This is your standard select(2) backend. Not I<completely> standard, as
153	libev tries to roll its own fd_set with no limits on the number of fds,	309	libev tries to roll its own fd_set with no limits on the number of fds,
154	but if that fails, expect a fairly low limit on the number of fds when	310	but if that fails, expect a fairly low limit on the number of fds when
155	using this backend. It doesn't scale too well (O(highest_fd)), but its usually	311	using this backend. It doesn't scale too well (O(highest_fd)), but its
156	the fastest backend for a low number of fds.	312	usually the fastest backend for a low number of (low-numbered :) fds.
157		313
		314	To get good performance out of this backend you need a high amount of
		315	parallelity (most of the file descriptors should be busy). If you are
		316	writing a server, you should C<accept ()> in a loop to accept as many
		317	connections as possible during one iteration. You might also want to have
		318	a look at C<ev_set_io_collect_interval ()> to increase the amount of
		319	readyness notifications you get per iteration.
		320
158	=item C<EVMETHOD_POLL> (value 2, poll backend, available everywhere except on windows)	321	=item C<EVBACKEND_POLL> (value 2, poll backend, available everywhere except on windows)
159		322
160	And this is your standard poll(2) backend. It's more complicated than	323	And this is your standard poll(2) backend. It's more complicated
161	select, but handles sparse fds better and has no artificial limit on the	324	than select, but handles sparse fds better and has no artificial
162	number of fds you can use (except it will slow down considerably with a	325	limit on the number of fds you can use (except it will slow down
163	lot of inactive fds). It scales similarly to select, i.e. O(total_fds).	326	considerably with a lot of inactive fds). It scales similarly to select,
		327	i.e. O(total_fds). See the entry for C<EVBACKEND_SELECT>, above, for
		328	performance tips.
164		329
165	=item C<EVMETHOD_EPOLL> (value 4, Linux)	330	=item C<EVBACKEND_EPOLL> (value 4, Linux)
166		331
167	For few fds, this backend is a bit little slower than poll and select,	332	For few fds, this backend is a bit little slower than poll and select,
168	but it scales phenomenally better. While poll and select usually scale like	333	but it scales phenomenally better. While poll and select usually scale
169	O(total_fds) where n is the total number of fds (or the highest fd), epoll scales	334	like O(total_fds) where n is the total number of fds (or the highest fd),
170	either O(1) or O(active_fds).	335	epoll scales either O(1) or O(active_fds). The epoll design has a number
		336	of shortcomings, such as silently dropping events in some hard-to-detect
		337	cases and rewiring a syscall per fd change, no fork support and bad
		338	support for dup.
171		339
172	While stopping and starting an I/O watcher in the same iteration will	340	While stopping, setting and starting an I/O watcher in the same iteration
173	result in some caching, there is still a syscall per such incident	341	will result in some caching, there is still a syscall per such incident
174	(because the fd could point to a different file description now), so its	342	(because the fd could point to a different file description now), so its
175	best to avoid that. Also, dup()ed file descriptors might not work very	343	best to avoid that. Also, C<dup ()>'ed file descriptors might not work
176	well if you register events for both fds.	344	very well if you register events for both fds.
177		345
		346	Please note that epoll sometimes generates spurious notifications, so you
		347	need to use non-blocking I/O or other means to avoid blocking when no data
		348	(or space) is available.
		349
		350	Best performance from this backend is achieved by not unregistering all
		351	watchers for a file descriptor until it has been closed, if possible, i.e.
		352	keep at least one watcher active per fd at all times.
		353
		354	While nominally embeddeble in other event loops, this feature is broken in
		355	all kernel versions tested so far.
		356
178	=item C<EVMETHOD_KQUEUE> (value 8, most BSD clones)	357	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)
179		358
180	Kqueue deserves special mention, as at the time of this writing, it	359	Kqueue deserves special mention, as at the time of this writing, it
181	was broken on all BSDs except NetBSD (usually it doesn't work with	360	was broken on all BSDs except NetBSD (usually it doesn't work reliably
182	anything but sockets and pipes, except on Darwin, where of course its	361	with anything but sockets and pipes, except on Darwin, where of course
183	completely useless). For this reason its not being "autodetected" unless	362	it's completely useless). For this reason it's not being "autodetected"
184	you explicitly specify the flags (i.e. you don't use EVFLAG_AUTO).	363	unless you explicitly specify it explicitly in the flags (i.e. using
		364	C<EVBACKEND_KQUEUE>) or libev was compiled on a known-to-be-good (-enough)
		365	system like NetBSD.
		366
		367	You still can embed kqueue into a normal poll or select backend and use it
		368	only for sockets (after having made sure that sockets work with kqueue on
		369	the target platform). See C<ev_embed> watchers for more info.
185		370
186	It scales in the same way as the epoll backend, but the interface to the	371	It scales in the same way as the epoll backend, but the interface to the
187	kernel is more efficient (which says nothing about its actual speed, of	372	kernel is more efficient (which says nothing about its actual speed, of
188	course). While starting and stopping an I/O watcher does not cause an	373	course). While stopping, setting and starting an I/O watcher does never
189	extra syscall as with epoll, it still adds up to four event changes per	374	cause an extra syscall as with C<EVBACKEND_EPOLL>, it still adds up to
190	incident, so its best to avoid that.	375	two event changes per incident, support for C<fork ()> is very bad and it
		376	drops fds silently in similarly hard-to-detect cases.
191		377
		378	This backend usually performs well under most conditions.
		379
		380	While nominally embeddable in other event loops, this doesn't work
		381	everywhere, so you might need to test for this. And since it is broken
		382	almost everywhere, you should only use it when you have a lot of sockets
		383	(for which it usually works), by embedding it into another event loop
		384	(e.g. C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>) and using it only for
		385	sockets.
		386
192	=item C<EVMETHOD_DEVPOLL> (value 16, Solaris 8)	387	=item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8)
193		388
194	This is not implemented yet (and might never be).	389	This is not implemented yet (and might never be, unless you send me an
		390	implementation). According to reports, C</dev/poll> only supports sockets
		391	and is not embeddable, which would limit the usefulness of this backend
		392	immensely.
195		393
196	=item C<EVMETHOD_PORT> (value 32, Solaris 10)	394	=item C<EVBACKEND_PORT> (value 32, Solaris 10)
197		395
198	This uses the Solaris 10 port mechanism. As with everything on Solaris,	396	This uses the Solaris 10 event port mechanism. As with everything on Solaris,
199	it's really slow, but it still scales very well (O(active_fds)).	397	it's really slow, but it still scales very well (O(active_fds)).
200		398
		399	Please note that solaris event ports can deliver a lot of spurious
		400	notifications, so you need to use non-blocking I/O or other means to avoid
		401	blocking when no data (or space) is available.
		402
		403	While this backend scales well, it requires one system call per active
		404	file descriptor per loop iteration. For small and medium numbers of file
		405	descriptors a "slow" C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> backend
		406	might perform better.
		407
201	=item C<EVMETHOD_ALL>	408	=item C<EVBACKEND_ALL>
202		409
203	Try all backends (even potentially broken ones that wouldn't be tried	410	Try all backends (even potentially broken ones that wouldn't be tried
204	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as	411	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as
205	C<EVMETHOD_ALL & ~EVMETHOD_KQUEUE>.	412	C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>.
		413
		414	It is definitely not recommended to use this flag.
206		415
207	=back	416	=back
208		417
209	If one or more of these are ored into the flags value, then only these	418	If one or more of these are ored into the flags value, then only these
210	backends will be tried (in the reverse order as given here). If none are	419	backends will be tried (in the reverse order as given here). If none are
211	specified, most compiled-in backend will be tried, usually in reverse	420	specified, most compiled-in backend will be tried, usually in reverse
212	order of their flag values :)	421	order of their flag values :)
213		422
		423	The most typical usage is like this:
		424
		425	if (!ev_default_loop (0))
		426	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
		427
		428	Restrict libev to the select and poll backends, and do not allow
		429	environment settings to be taken into account:
		430
		431	ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);
		432
		433	Use whatever libev has to offer, but make sure that kqueue is used if
		434	available (warning, breaks stuff, best use only with your own private
		435	event loop and only if you know the OS supports your types of fds):
		436
		437	ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);
		438
214	=item struct ev_loop *ev_loop_new (unsigned int flags)	439	=item struct ev_loop *ev_loop_new (unsigned int flags)
215		440
216	Similar to C<ev_default_loop>, but always creates a new event loop that is	441	Similar to C<ev_default_loop>, but always creates a new event loop that is
217	always distinct from the default loop. Unlike the default loop, it cannot	442	always distinct from the default loop. Unlike the default loop, it cannot
218	handle signal and child watchers, and attempts to do so will be greeted by	443	handle signal and child watchers, and attempts to do so will be greeted by
219	undefined behaviour (or a failed assertion if assertions are enabled).	444	undefined behaviour (or a failed assertion if assertions are enabled).
220		445
		446	Example: Try to create a event loop that uses epoll and nothing else.
		447
		448	struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);
		449	if (!epoller)
		450	fatal ("no epoll found here, maybe it hides under your chair");
		451
221	=item ev_default_destroy ()	452	=item ev_default_destroy ()
222		453
223	Destroys the default loop again (frees all memory and kernel state	454	Destroys the default loop again (frees all memory and kernel state
224	etc.). This stops all registered event watchers (by not touching them in	455	etc.). None of the active event watchers will be stopped in the normal
225	any way whatsoever, although you cannot rely on this :).	456	sense, so e.g. C<ev_is_active> might still return true. It is your
		457	responsibility to either stop all watchers cleanly yoursef I<before>
		458	calling this function, or cope with the fact afterwards (which is usually
		459	the easiest thing, you can just ignore the watchers and/or C<free ()> them
		460	for example).
		461
		462	Note that certain global state, such as signal state, will not be freed by
		463	this function, and related watchers (such as signal and child watchers)
		464	would need to be stopped manually.
		465
		466	In general it is not advisable to call this function except in the
		467	rare occasion where you really need to free e.g. the signal handling
		468	pipe fds. If you need dynamically allocated loops it is better to use
		469	C<ev_loop_new> and C<ev_loop_destroy>).
226		470
227	=item ev_loop_destroy (loop)	471	=item ev_loop_destroy (loop)
228		472
229	Like C<ev_default_destroy>, but destroys an event loop created by an	473	Like C<ev_default_destroy>, but destroys an event loop created by an
230	earlier call to C<ev_loop_new>.	474	earlier call to C<ev_loop_new>.
…		…
244	it just in case after a fork. To make this easy, the function will fit in	488	it just in case after a fork. To make this easy, the function will fit in
245	quite nicely into a call to C<pthread_atfork>:	489	quite nicely into a call to C<pthread_atfork>:
246		490
247	pthread_atfork (0, 0, ev_default_fork);	491	pthread_atfork (0, 0, ev_default_fork);
248		492
		493	At the moment, C<EVBACKEND_SELECT> and C<EVBACKEND_POLL> are safe to use
		494	without calling this function, so if you force one of those backends you
		495	do not need to care.
		496
249	=item ev_loop_fork (loop)	497	=item ev_loop_fork (loop)
250		498
251	Like C<ev_default_fork>, but acts on an event loop created by	499	Like C<ev_default_fork>, but acts on an event loop created by
252	C<ev_loop_new>. Yes, you have to call this on every allocated event loop	500	C<ev_loop_new>. Yes, you have to call this on every allocated event loop
253	after fork, and how you do this is entirely your own problem.	501	after fork, and how you do this is entirely your own problem.
254		502
		503	=item unsigned int ev_loop_count (loop)
		504
		505	Returns the count of loop iterations for the loop, which is identical to
		506	the number of times libev did poll for new events. It starts at C<0> and
		507	happily wraps around with enough iterations.
		508
		509	This value can sometimes be useful as a generation counter of sorts (it
		510	"ticks" the number of loop iterations), as it roughly corresponds with
		511	C<ev_prepare> and C<ev_check> calls.
		512
255	=item unsigned int ev_method (loop)	513	=item unsigned int ev_backend (loop)
256		514
257	Returns one of the C<EVMETHOD_*> flags indicating the event backend in	515	Returns one of the C<EVBACKEND_*> flags indicating the event backend in
258	use.	516	use.
259		517
260	=item ev_tstamp ev_now (loop)	518	=item ev_tstamp ev_now (loop)
261		519
262	Returns the current "event loop time", which is the time the event loop	520	Returns the current "event loop time", which is the time the event loop
263	got events and started processing them. This timestamp does not change	521	received events and started processing them. This timestamp does not
264	as long as callbacks are being processed, and this is also the base time	522	change as long as callbacks are being processed, and this is also the base
265	used for relative timers. You can treat it as the timestamp of the event	523	time used for relative timers. You can treat it as the timestamp of the
266	occuring (or more correctly, the mainloop finding out about it).	524	event occurring (or more correctly, libev finding out about it).
267		525
268	=item ev_loop (loop, int flags)	526	=item ev_loop (loop, int flags)
269		527
270	Finally, this is it, the event handler. This function usually is called	528	Finally, this is it, the event handler. This function usually is called
271	after you initialised all your watchers and you want to start handling	529	after you initialised all your watchers and you want to start handling
272	events.	530	events.
273		531
274	If the flags argument is specified as 0, it will not return until either	532	If the flags argument is specified as C<0>, it will not return until
275	no event watchers are active anymore or C<ev_unloop> was called.	533	either no event watchers are active anymore or C<ev_unloop> was called.
		534
		535	Please note that an explicit C<ev_unloop> is usually better than
		536	relying on all watchers to be stopped when deciding when a program has
		537	finished (especially in interactive programs), but having a program that
		538	automatically loops as long as it has to and no longer by virtue of
		539	relying on its watchers stopping correctly is a thing of beauty.
276		540
277	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle	541	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle
278	those events and any outstanding ones, but will not block your process in	542	those events and any outstanding ones, but will not block your process in
279	case there are no events and will return after one iteration of the loop.	543	case there are no events and will return after one iteration of the loop.
280		544
281	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if	545	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if
282	neccessary) and will handle those and any outstanding ones. It will block	546	neccessary) and will handle those and any outstanding ones. It will block
283	your process until at least one new event arrives, and will return after	547	your process until at least one new event arrives, and will return after
284	one iteration of the loop.	548	one iteration of the loop. This is useful if you are waiting for some
		549	external event in conjunction with something not expressible using other
		550	libev watchers. However, a pair of C<ev_prepare>/C<ev_check> watchers is
		551	usually a better approach for this kind of thing.
285		552
286	This flags value could be used to implement alternative looping
287	constructs, but the C<prepare> and C<check> watchers provide a better and
288	more generic mechanism.
289
290	Here are the gory details of what ev_loop does:	553	Here are the gory details of what C<ev_loop> does:
291		554
292	1. If there are no active watchers (reference count is zero), return.	555	- Before the first iteration, call any pending watchers.
		556	* If EVFLAG_FORKCHECK was used, check for a fork.
		557	- If a fork was detected, queue and call all fork watchers.
293	2. Queue and immediately call all prepare watchers.	558	- Queue and call all prepare watchers.
294	3. If we have been forked, recreate the kernel state.	559	- If we have been forked, recreate the kernel state.
295	4. Update the kernel state with all outstanding changes.	560	- Update the kernel state with all outstanding changes.
296	5. Update the "event loop time".	561	- Update the "event loop time".
297	6. Calculate for how long to block.	562	- Calculate for how long to sleep or block, if at all
		563	(active idle watchers, EVLOOP_NONBLOCK or not having
		564	any active watchers at all will result in not sleeping).
		565	- Sleep if the I/O and timer collect interval say so.
298	7. Block the process, waiting for events.	566	- Block the process, waiting for any events.
		567	- Queue all outstanding I/O (fd) events.
299	8. Update the "event loop time" and do time jump handling.	568	- Update the "event loop time" and do time jump handling.
300	9. Queue all outstanding timers.	569	- Queue all outstanding timers.
301	10. Queue all outstanding periodics.	570	- Queue all outstanding periodics.
302	11. If no events are pending now, queue all idle watchers.	571	- If no events are pending now, queue all idle watchers.
303	12. Queue all check watchers.	572	- Queue all check watchers.
304	13. Call all queued watchers in reverse order (i.e. check watchers first).	573	- Call all queued watchers in reverse order (i.e. check watchers first).
		574	Signals and child watchers are implemented as I/O watchers, and will
		575	be handled here by queueing them when their watcher gets executed.
305	14. If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK	576	- If ev_unloop has been called, or EVLOOP_ONESHOT or EVLOOP_NONBLOCK
306	was used, return, otherwise continue with step #1.	577	were used, or there are no active watchers, return, otherwise
		578	continue with step *.
		579
		580	Example: Queue some jobs and then loop until no events are outstanding
		581	anymore.
		582
		583	... queue jobs here, make sure they register event watchers as long
		584	... as they still have work to do (even an idle watcher will do..)
		585	ev_loop (my_loop, 0);
		586	... jobs done. yeah!
307		587
308	=item ev_unloop (loop, how)	588	=item ev_unloop (loop, how)
309		589
310	Can be used to make a call to C<ev_loop> return early (but only after it	590	Can be used to make a call to C<ev_loop> return early (but only after it
311	has processed all outstanding events). The C<how> argument must be either	591	has processed all outstanding events). The C<how> argument must be either
312	C<EVUNLOOP_ONE>, which will make the innermost C<ev_loop> call return, or	592	C<EVUNLOOP_ONE>, which will make the innermost C<ev_loop> call return, or
313	C<EVUNLOOP_ALL>, which will make all nested C<ev_loop> calls return.	593	C<EVUNLOOP_ALL>, which will make all nested C<ev_loop> calls return.
		594
		595	This "unloop state" will be cleared when entering C<ev_loop> again.
314		596
315	=item ev_ref (loop)	597	=item ev_ref (loop)
316		598
317	=item ev_unref (loop)	599	=item ev_unref (loop)
318		600
…		…
325	visible to the libev user and should not keep C<ev_loop> from exiting if	607	visible to the libev user and should not keep C<ev_loop> from exiting if
326	no event watchers registered by it are active. It is also an excellent	608	no event watchers registered by it are active. It is also an excellent
327	way to do this for generic recurring timers or from within third-party	609	way to do this for generic recurring timers or from within third-party
328	libraries. Just remember to I<unref after start> and I<ref before stop>.	610	libraries. Just remember to I<unref after start> and I<ref before stop>.
329		611
		612	Example: Create a signal watcher, but keep it from keeping C<ev_loop>
		613	running when nothing else is active.
		614
		615	struct ev_signal exitsig;
		616	ev_signal_init (&exitsig, sig_cb, SIGINT);
		617	ev_signal_start (loop, &exitsig);
		618	evf_unref (loop);
		619
		620	Example: For some weird reason, unregister the above signal handler again.
		621
		622	ev_ref (loop);
		623	ev_signal_stop (loop, &exitsig);
		624
		625	=item ev_set_io_collect_interval (loop, ev_tstamp interval)
		626
		627	=item ev_set_timeout_collect_interval (loop, ev_tstamp interval)
		628
		629	These advanced functions influence the time that libev will spend waiting
		630	for events. Both are by default C<0>, meaning that libev will try to
		631	invoke timer/periodic callbacks and I/O callbacks with minimum latency.
		632
		633	Setting these to a higher value (the C<interval> I<must> be >= C<0>)
		634	allows libev to delay invocation of I/O and timer/periodic callbacks to
		635	increase efficiency of loop iterations.
		636
		637	The background is that sometimes your program runs just fast enough to
		638	handle one (or very few) event(s) per loop iteration. While this makes
		639	the program responsive, it also wastes a lot of CPU time to poll for new
		640	events, especially with backends like C<select ()> which have a high
		641	overhead for the actual polling but can deliver many events at once.
		642
		643	By setting a higher I<io collect interval> you allow libev to spend more
		644	time collecting I/O events, so you can handle more events per iteration,
		645	at the cost of increasing latency. Timeouts (both C<ev_periodic> and
		646	C<ev_timer>) will be not affected. Setting this to a non-null value will
		647	introduce an additional C<ev_sleep ()> call into most loop iterations.
		648
		649	Likewise, by setting a higher I<timeout collect interval> you allow libev
		650	to spend more time collecting timeouts, at the expense of increased
		651	latency (the watcher callback will be called later). C<ev_io> watchers
		652	will not be affected. Setting this to a non-null value will not introduce
		653	any overhead in libev.
		654
		655	Many (busy) programs can usually benefit by setting the io collect
		656	interval to a value near C<0.1> or so, which is often enough for
		657	interactive servers (of course not for games), likewise for timeouts. It
		658	usually doesn't make much sense to set it to a lower value than C<0.01>,
		659	as this approsaches the timing granularity of most systems.
		660
330	=back	661	=back
		662
331		663
332	=head1 ANATOMY OF A WATCHER	664	=head1 ANATOMY OF A WATCHER
333		665
334	A watcher is a structure that you create and register to record your	666	A watcher is a structure that you create and register to record your
335	interest in some event. For instance, if you want to wait for STDIN to	667	interest in some event. For instance, if you want to wait for STDIN to
…		…
368	*) >>), and you can stop watching for events at any time by calling the	700	*) >>), and you can stop watching for events at any time by calling the
369	corresponding stop function (C<< ev_<type>_stop (loop, watcher *) >>.	701	corresponding stop function (C<< ev_<type>_stop (loop, watcher *) >>.
370		702
371	As long as your watcher is active (has been started but not stopped) you	703	As long as your watcher is active (has been started but not stopped) you
372	must not touch the values stored in it. Most specifically you must never	704	must not touch the values stored in it. Most specifically you must never
373	reinitialise it or call its set method.	705	reinitialise it or call its C<set> macro.
374
375	You can check whether an event is active by calling the C<ev_is_active
376	(watcher *)> macro. To see whether an event is outstanding (but the
377	callback for it has not been called yet) you can use the C<ev_is_pending
378	(watcher *)> macro.
379		706
380	Each and every callback receives the event loop pointer as first, the	707	Each and every callback receives the event loop pointer as first, the
381	registered watcher structure as second, and a bitset of received events as	708	registered watcher structure as second, and a bitset of received events as
382	third argument.	709	third argument.
383		710
…		…
407	The signal specified in the C<ev_signal> watcher has been received by a thread.	734	The signal specified in the C<ev_signal> watcher has been received by a thread.
408		735
409	=item C<EV_CHILD>	736	=item C<EV_CHILD>
410		737
411	The pid specified in the C<ev_child> watcher has received a status change.	738	The pid specified in the C<ev_child> watcher has received a status change.
		739
		740	=item C<EV_STAT>
		741
		742	The path specified in the C<ev_stat> watcher changed its attributes somehow.
412		743
413	=item C<EV_IDLE>	744	=item C<EV_IDLE>
414		745
415	The C<ev_idle> watcher has determined that you have nothing better to do.	746	The C<ev_idle> watcher has determined that you have nothing better to do.
416		747
…		…
424	received events. Callbacks of both watcher types can start and stop as	755	received events. Callbacks of both watcher types can start and stop as
425	many watchers as they want, and all of them will be taken into account	756	many watchers as they want, and all of them will be taken into account
426	(for example, a C<ev_prepare> watcher might start an idle watcher to keep	757	(for example, a C<ev_prepare> watcher might start an idle watcher to keep
427	C<ev_loop> from blocking).	758	C<ev_loop> from blocking).
428		759
		760	=item C<EV_EMBED>
		761
		762	The embedded event loop specified in the C<ev_embed> watcher needs attention.
		763
		764	=item C<EV_FORK>
		765
		766	The event loop has been resumed in the child process after fork (see
		767	C<ev_fork>).
		768
429	=item C<EV_ERROR>	769	=item C<EV_ERROR>
430		770
431	An unspecified error has occured, the watcher has been stopped. This might	771	An unspecified error has occured, the watcher has been stopped. This might
432	happen because the watcher could not be properly started because libev	772	happen because the watcher could not be properly started because libev
433	ran out of memory, a file descriptor was found to be closed or any other	773	ran out of memory, a file descriptor was found to be closed or any other
…		…
439	your callbacks is well-written it can just attempt the operation and cope	779	your callbacks is well-written it can just attempt the operation and cope
440	with the error from read() or write(). This will not work in multithreaded	780	with the error from read() or write(). This will not work in multithreaded
441	programs, though, so beware.	781	programs, though, so beware.
442		782
443	=back	783	=back
		784
		785	=head2 GENERIC WATCHER FUNCTIONS
		786
		787	In the following description, C<TYPE> stands for the watcher type,
		788	e.g. C<timer> for C<ev_timer> watchers and C<io> for C<ev_io> watchers.
		789
		790	=over 4
		791
		792	=item C<ev_init> (ev_TYPE *watcher, callback)
		793
		794	This macro initialises the generic portion of a watcher. The contents
		795	of the watcher object can be arbitrary (so C<malloc> will do). Only
		796	the generic parts of the watcher are initialised, you I<need> to call
		797	the type-specific C<ev_TYPE_set> macro afterwards to initialise the
		798	type-specific parts. For each type there is also a C<ev_TYPE_init> macro
		799	which rolls both calls into one.
		800
		801	You can reinitialise a watcher at any time as long as it has been stopped
		802	(or never started) and there are no pending events outstanding.
		803
		804	The callback is always of type C<void (*)(ev_loop *loop, ev_TYPE *watcher,
		805	int revents)>.
		806
		807	=item C<ev_TYPE_set> (ev_TYPE *, [args])
		808
		809	This macro initialises the type-specific parts of a watcher. You need to
		810	call C<ev_init> at least once before you call this macro, but you can
		811	call C<ev_TYPE_set> any number of times. You must not, however, call this
		812	macro on a watcher that is active (it can be pending, however, which is a
		813	difference to the C<ev_init> macro).
		814
		815	Although some watcher types do not have type-specific arguments
		816	(e.g. C<ev_prepare>) you still need to call its C<set> macro.
		817
		818	=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])
		819
		820	This convinience macro rolls both C<ev_init> and C<ev_TYPE_set> macro
		821	calls into a single call. This is the most convinient method to initialise
		822	a watcher. The same limitations apply, of course.
		823
		824	=item C<ev_TYPE_start> (loop *, ev_TYPE *watcher)
		825
		826	Starts (activates) the given watcher. Only active watchers will receive
		827	events. If the watcher is already active nothing will happen.
		828
		829	=item C<ev_TYPE_stop> (loop *, ev_TYPE *watcher)
		830
		831	Stops the given watcher again (if active) and clears the pending
		832	status. It is possible that stopped watchers are pending (for example,
		833	non-repeating timers are being stopped when they become pending), but
		834	C<ev_TYPE_stop> ensures that the watcher is neither active nor pending. If
		835	you want to free or reuse the memory used by the watcher it is therefore a
		836	good idea to always call its C<ev_TYPE_stop> function.
		837
		838	=item bool ev_is_active (ev_TYPE *watcher)
		839
		840	Returns a true value iff the watcher is active (i.e. it has been started
		841	and not yet been stopped). As long as a watcher is active you must not modify
		842	it.
		843
		844	=item bool ev_is_pending (ev_TYPE *watcher)
		845
		846	Returns a true value iff the watcher is pending, (i.e. it has outstanding
		847	events but its callback has not yet been invoked). As long as a watcher
		848	is pending (but not active) you must not call an init function on it (but
		849	C<ev_TYPE_set> is safe), you must not change its priority, and you must
		850	make sure the watcher is available to libev (e.g. you cannot C<free ()>
		851	it).
		852
		853	=item callback ev_cb (ev_TYPE *watcher)
		854
		855	Returns the callback currently set on the watcher.
		856
		857	=item ev_cb_set (ev_TYPE *watcher, callback)
		858
		859	Change the callback. You can change the callback at virtually any time
		860	(modulo threads).
		861
		862	=item ev_set_priority (ev_TYPE *watcher, priority)
		863
		864	=item int ev_priority (ev_TYPE *watcher)
		865
		866	Set and query the priority of the watcher. The priority is a small
		867	integer between C<EV_MAXPRI> (default: C<2>) and C<EV_MINPRI>
		868	(default: C<-2>). Pending watchers with higher priority will be invoked
		869	before watchers with lower priority, but priority will not keep watchers
		870	from being executed (except for C<ev_idle> watchers).
		871
		872	This means that priorities are I<only> used for ordering callback
		873	invocation after new events have been received. This is useful, for
		874	example, to reduce latency after idling, or more often, to bind two
		875	watchers on the same event and make sure one is called first.
		876
		877	If you need to suppress invocation when higher priority events are pending
		878	you need to look at C<ev_idle> watchers, which provide this functionality.
		879
		880	You I<must not> change the priority of a watcher as long as it is active or
		881	pending.
		882
		883	The default priority used by watchers when no priority has been set is
		884	always C<0>, which is supposed to not be too high and not be too low :).
		885
		886	Setting a priority outside the range of C<EV_MINPRI> to C<EV_MAXPRI> is
		887	fine, as long as you do not mind that the priority value you query might
		888	or might not have been adjusted to be within valid range.
		889
		890	=item ev_invoke (loop, ev_TYPE *watcher, int revents)
		891
		892	Invoke the C<watcher> with the given C<loop> and C<revents>. Neither
		893	C<loop> nor C<revents> need to be valid as long as the watcher callback
		894	can deal with that fact.
		895
		896	=item int ev_clear_pending (loop, ev_TYPE *watcher)
		897
		898	If the watcher is pending, this function returns clears its pending status
		899	and returns its C<revents> bitset (as if its callback was invoked). If the
		900	watcher isn't pending it does nothing and returns C<0>.
		901
		902	=back
		903
444		904
445	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER	905	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER
446		906
447	Each watcher has, by default, a member C<void *data> that you can change	907	Each watcher has, by default, a member C<void *data> that you can change
448	and read at any time, libev will completely ignore it. This can be used	908	and read at any time, libev will completely ignore it. This can be used
…		…
466	{	926	{
467	struct my_io *w = (struct my_io *)w_;	927	struct my_io *w = (struct my_io *)w_;
468	...	928	...
469	}	929	}
470		930
471	More interesting and less C-conformant ways of catsing your callback type	931	More interesting and less C-conformant ways of casting your callback type
472	have been omitted....	932	instead have been omitted.
		933
		934	Another common scenario is having some data structure with multiple
		935	watchers:
		936
		937	struct my_biggy
		938	{
		939	int some_data;
		940	ev_timer t1;
		941	ev_timer t2;
		942	}
		943
		944	In this case getting the pointer to C<my_biggy> is a bit more complicated,
		945	you need to use C<offsetof>:
		946
		947	#include <stddef.h>
		948
		949	static void
		950	t1_cb (EV_P_ struct ev_timer *w, int revents)
		951	{
		952	struct my_biggy big = (struct my_biggy *
		953	(((char *)w) - offsetof (struct my_biggy, t1));
		954	}
		955
		956	static void
		957	t2_cb (EV_P_ struct ev_timer *w, int revents)
		958	{
		959	struct my_biggy big = (struct my_biggy *
		960	(((char *)w) - offsetof (struct my_biggy, t2));
		961	}
473		962
474		963
475	=head1 WATCHER TYPES	964	=head1 WATCHER TYPES
476		965
477	This section describes each watcher in detail, but will not repeat	966	This section describes each watcher in detail, but will not repeat
478	information given in the last section.	967	information given in the last section. Any initialisation/set macros,
		968	functions and members specific to the watcher type are explained.
479		969
		970	Members are additionally marked with either I<[read-only]>, meaning that,
		971	while the watcher is active, you can look at the member and expect some
		972	sensible content, but you must not modify it (you can modify it while the
		973	watcher is stopped to your hearts content), or I<[read-write]>, which
		974	means you can expect it to have some sensible content while the watcher
		975	is active, but you can also modify it. Modifying it may not do something
		976	sensible or take immediate effect (or do anything at all), but libev will
		977	not crash or malfunction in any way.
		978
		979
480	=head2 C<ev_io> - is this file descriptor readable or writable	980	=head2 C<ev_io> - is this file descriptor readable or writable?
481		981
482	I/O watchers check whether a file descriptor is readable or writable	982	I/O watchers check whether a file descriptor is readable or writable
483	in each iteration of the event loop (This behaviour is called	983	in each iteration of the event loop, or, more precisely, when reading
484	level-triggering because you keep receiving events as long as the	984	would not block the process and writing would at least be able to write
485	condition persists. Remember you can stop the watcher if you don't want to	985	some data. This behaviour is called level-triggering because you keep
486	act on the event and neither want to receive future events).	986	receiving events as long as the condition persists. Remember you can stop
		987	the watcher if you don't want to act on the event and neither want to
		988	receive future events.
487		989
488	In general you can register as many read and/or write event watchers per	990	In general you can register as many read and/or write event watchers per
489	fd as you want (as long as you don't confuse yourself). Setting all file	991	fd as you want (as long as you don't confuse yourself). Setting all file
490	descriptors to non-blocking mode is also usually a good idea (but not	992	descriptors to non-blocking mode is also usually a good idea (but not
491	required if you know what you are doing).	993	required if you know what you are doing).
492		994
493	You have to be careful with dup'ed file descriptors, though. Some backends
494	(the linux epoll backend is a notable example) cannot handle dup'ed file
495	descriptors correctly if you register interest in two or more fds pointing
496	to the same underlying file/socket etc. description (that is, they share
497	the same underlying "file open").
498
499	If you must do this, then force the use of a known-to-be-good backend	995	If you must do this, then force the use of a known-to-be-good backend
500	(at the time of this writing, this includes only EVMETHOD_SELECT and	996	(at the time of this writing, this includes only C<EVBACKEND_SELECT> and
501	EVMETHOD_POLL).	997	C<EVBACKEND_POLL>).
		998
		999	Another thing you have to watch out for is that it is quite easy to
		1000	receive "spurious" readyness notifications, that is your callback might
		1001	be called with C<EV_READ> but a subsequent C<read>(2) will actually block
		1002	because there is no data. Not only are some backends known to create a
		1003	lot of those (for example solaris ports), it is very easy to get into
		1004	this situation even with a relatively standard program structure. Thus
		1005	it is best to always use non-blocking I/O: An extra C<read>(2) returning
		1006	C<EAGAIN> is far preferable to a program hanging until some data arrives.
		1007
		1008	If you cannot run the fd in non-blocking mode (for example you should not
		1009	play around with an Xlib connection), then you have to seperately re-test
		1010	whether a file descriptor is really ready with a known-to-be good interface
		1011	such as poll (fortunately in our Xlib example, Xlib already does this on
		1012	its own, so its quite safe to use).
		1013
		1014	=head3 The special problem of disappearing file descriptors
		1015
		1016	Some backends (e.g. kqueue, epoll) need to be told about closing a file
		1017	descriptor (either by calling C<close> explicitly or by any other means,
		1018	such as C<dup>). The reason is that you register interest in some file
		1019	descriptor, but when it goes away, the operating system will silently drop
		1020	this interest. If another file descriptor with the same number then is
		1021	registered with libev, there is no efficient way to see that this is, in
		1022	fact, a different file descriptor.
		1023
		1024	To avoid having to explicitly tell libev about such cases, libev follows
		1025	the following policy: Each time C<ev_io_set> is being called, libev
		1026	will assume that this is potentially a new file descriptor, otherwise
		1027	it is assumed that the file descriptor stays the same. That means that
		1028	you I<have> to call C<ev_io_set> (or C<ev_io_init>) when you change the
		1029	descriptor even if the file descriptor number itself did not change.
		1030
		1031	This is how one would do it normally anyway, the important point is that
		1032	the libev application should not optimise around libev but should leave
		1033	optimisations to libev.
		1034
		1035	=head3 The special problem of dup'ed file descriptors
		1036
		1037	Some backends (e.g. epoll), cannot register events for file descriptors,
		1038	but only events for the underlying file descriptions. That means when you
		1039	have C<dup ()>'ed file descriptors or weirder constellations, and register
		1040	events for them, only one file descriptor might actually receive events.
		1041
		1042	There is no workaround possible except not registering events
		1043	for potentially C<dup ()>'ed file descriptors, or to resort to
		1044	C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>.
		1045
		1046	=head3 The special problem of fork
		1047
		1048	Some backends (epoll, kqueue) do not support C<fork ()> at all or exhibit
		1049	useless behaviour. Libev fully supports fork, but needs to be told about
		1050	it in the child.
		1051
		1052	To support fork in your programs, you either have to call
		1053	C<ev_default_fork ()> or C<ev_loop_fork ()> after a fork in the child,
		1054	enable C<EVFLAG_FORKCHECK>, or resort to C<EVBACKEND_SELECT> or
		1055	C<EVBACKEND_POLL>.
		1056
		1057
		1058	=head3 Watcher-Specific Functions
502		1059
503	=over 4	1060	=over 4
504		1061
505	=item ev_io_init (ev_io *, callback, int fd, int events)	1062	=item ev_io_init (ev_io *, callback, int fd, int events)
506		1063
507	=item ev_io_set (ev_io *, int fd, int events)	1064	=item ev_io_set (ev_io *, int fd, int events)
508		1065
509	Configures an C<ev_io> watcher. The fd is the file descriptor to rceeive	1066	Configures an C<ev_io> watcher. The C<fd> is the file descriptor to
510	events for and events is either C<EV_READ>, C<EV_WRITE> or C<EV_READ |	1067	rceeive events for and events is either C<EV_READ>, C<EV_WRITE> or
511	EV_WRITE> to receive the given events.	1068	C<EV_READ | EV_WRITE> to receive the given events.
		1069
		1070	=item int fd [read-only]
		1071
		1072	The file descriptor being watched.
		1073
		1074	=item int events [read-only]
		1075
		1076	The events being watched.
512		1077
513	=back	1078	=back
514		1079
		1080	=head3 Examples
		1081
		1082	Example: Call C<stdin_readable_cb> when STDIN_FILENO has become, well
		1083	readable, but only once. Since it is likely line-buffered, you could
		1084	attempt to read a whole line in the callback.
		1085
		1086	static void
		1087	stdin_readable_cb (struct ev_loop *loop, struct ev_io *w, int revents)
		1088	{
		1089	ev_io_stop (loop, w);
		1090	.. read from stdin here (or from w->fd) and haqndle any I/O errors
		1091	}
		1092
		1093	...
		1094	struct ev_loop *loop = ev_default_init (0);
		1095	struct ev_io stdin_readable;
		1096	ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
		1097	ev_io_start (loop, &stdin_readable);
		1098	ev_loop (loop, 0);
		1099
		1100
515	=head2 C<ev_timer> - relative and optionally recurring timeouts	1101	=head2 C<ev_timer> - relative and optionally repeating timeouts
516		1102
517	Timer watchers are simple relative timers that generate an event after a	1103	Timer watchers are simple relative timers that generate an event after a
518	given time, and optionally repeating in regular intervals after that.	1104	given time, and optionally repeating in regular intervals after that.
519		1105
520	The timers are based on real time, that is, if you register an event that	1106	The timers are based on real time, that is, if you register an event that
…		…
533		1119
534	The callback is guarenteed to be invoked only when its timeout has passed,	1120	The callback is guarenteed to be invoked only when its timeout has passed,
535	but if multiple timers become ready during the same loop iteration then	1121	but if multiple timers become ready during the same loop iteration then
536	order of execution is undefined.	1122	order of execution is undefined.
537		1123
		1124	=head3 Watcher-Specific Functions and Data Members
		1125
538	=over 4	1126	=over 4
539		1127
540	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)	1128	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)
541		1129
542	=item ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)	1130	=item ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)
…		…
555	=item ev_timer_again (loop)	1143	=item ev_timer_again (loop)
556		1144
557	This will act as if the timer timed out and restart it again if it is	1145	This will act as if the timer timed out and restart it again if it is
558	repeating. The exact semantics are:	1146	repeating. The exact semantics are:
559		1147
		1148	If the timer is pending, its pending status is cleared.
		1149
560	If the timer is started but nonrepeating, stop it.	1150	If the timer is started but nonrepeating, stop it (as if it timed out).
561		1151
562	If the timer is repeating, either start it if necessary (with the repeat	1152	If the timer is repeating, either start it if necessary (with the
563	value), or reset the running timer to the repeat value.	1153	C<repeat> value), or reset the running timer to the C<repeat> value.
564		1154
565	This sounds a bit complicated, but here is a useful and typical	1155	This sounds a bit complicated, but here is a useful and typical
566	example: Imagine you have a tcp connection and you want a so-called idle	1156	example: Imagine you have a tcp connection and you want a so-called idle
567	timeout, that is, you want to be called when there have been, say, 60	1157	timeout, that is, you want to be called when there have been, say, 60
568	seconds of inactivity on the socket. The easiest way to do this is to	1158	seconds of inactivity on the socket. The easiest way to do this is to
569	configure an C<ev_timer> with after=repeat=60 and calling ev_timer_again each	1159	configure an C<ev_timer> with a C<repeat> value of C<60> and then call
570	time you successfully read or write some data. If you go into an idle	1160	C<ev_timer_again> each time you successfully read or write some data. If
571	state where you do not expect data to travel on the socket, you can stop	1161	you go into an idle state where you do not expect data to travel on the
572	the timer, and again will automatically restart it if need be.	1162	socket, you can C<ev_timer_stop> the timer, and C<ev_timer_again> will
		1163	automatically restart it if need be.
		1164
		1165	That means you can ignore the C<after> value and C<ev_timer_start>
		1166	altogether and only ever use the C<repeat> value and C<ev_timer_again>:
		1167
		1168	ev_timer_init (timer, callback, 0., 5.);
		1169	ev_timer_again (loop, timer);
		1170	...
		1171	timer->again = 17.;
		1172	ev_timer_again (loop, timer);
		1173	...
		1174	timer->again = 10.;
		1175	ev_timer_again (loop, timer);
		1176
		1177	This is more slightly efficient then stopping/starting the timer each time
		1178	you want to modify its timeout value.
		1179
		1180	=item ev_tstamp repeat [read-write]
		1181
		1182	The current C<repeat> value. Will be used each time the watcher times out
		1183	or C<ev_timer_again> is called and determines the next timeout (if any),
		1184	which is also when any modifications are taken into account.
573		1185
574	=back	1186	=back
575		1187
		1188	=head3 Examples
		1189
		1190	Example: Create a timer that fires after 60 seconds.
		1191
		1192	static void
		1193	one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
		1194	{
		1195	.. one minute over, w is actually stopped right here
		1196	}
		1197
		1198	struct ev_timer mytimer;
		1199	ev_timer_init (&mytimer, one_minute_cb, 60., 0.);
		1200	ev_timer_start (loop, &mytimer);
		1201
		1202	Example: Create a timeout timer that times out after 10 seconds of
		1203	inactivity.
		1204
		1205	static void
		1206	timeout_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
		1207	{
		1208	.. ten seconds without any activity
		1209	}
		1210
		1211	struct ev_timer mytimer;
		1212	ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */
		1213	ev_timer_again (&mytimer); /* start timer */
		1214	ev_loop (loop, 0);
		1215
		1216	// and in some piece of code that gets executed on any "activity":
		1217	// reset the timeout to start ticking again at 10 seconds
		1218	ev_timer_again (&mytimer);
		1219
		1220
576	=head2 C<ev_periodic> - to cron or not to cron	1221	=head2 C<ev_periodic> - to cron or not to cron?
577		1222
578	Periodic watchers are also timers of a kind, but they are very versatile	1223	Periodic watchers are also timers of a kind, but they are very versatile
579	(and unfortunately a bit complex).	1224	(and unfortunately a bit complex).
580		1225
581	Unlike C<ev_timer>'s, they are not based on real time (or relative time)	1226	Unlike C<ev_timer>'s, they are not based on real time (or relative time)
582	but on wallclock time (absolute time). You can tell a periodic watcher	1227	but on wallclock time (absolute time). You can tell a periodic watcher
583	to trigger "at" some specific point in time. For example, if you tell a	1228	to trigger "at" some specific point in time. For example, if you tell a
584	periodic watcher to trigger in 10 seconds (by specifiying e.g. c<ev_now ()	1229	periodic watcher to trigger in 10 seconds (by specifiying e.g. C<ev_now ()
585	+ 10.>) and then reset your system clock to the last year, then it will	1230	+ 10.>) and then reset your system clock to the last year, then it will
586	take a year to trigger the event (unlike an C<ev_timer>, which would trigger	1231	take a year to trigger the event (unlike an C<ev_timer>, which would trigger
587	roughly 10 seconds later and of course not if you reset your system time	1232	roughly 10 seconds later).
588	again).
589		1233
590	They can also be used to implement vastly more complex timers, such as	1234	They can also be used to implement vastly more complex timers, such as
591	triggering an event on eahc midnight, local time.	1235	triggering an event on each midnight, local time or other, complicated,
		1236	rules.
592		1237
593	As with timers, the callback is guarenteed to be invoked only when the	1238	As with timers, the callback is guarenteed to be invoked only when the
594	time (C<at>) has been passed, but if multiple periodic timers become ready	1239	time (C<at>) has been passed, but if multiple periodic timers become ready
595	during the same loop iteration then order of execution is undefined.	1240	during the same loop iteration then order of execution is undefined.
596		1241
		1242	=head3 Watcher-Specific Functions and Data Members
		1243
597	=over 4	1244	=over 4
598		1245
599	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)	1246	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)
600		1247
601	=item ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)	1248	=item ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)
…		…
603	Lots of arguments, lets sort it out... There are basically three modes of	1250	Lots of arguments, lets sort it out... There are basically three modes of
604	operation, and we will explain them from simplest to complex:	1251	operation, and we will explain them from simplest to complex:
605		1252
606	=over 4	1253	=over 4
607		1254
608	=item * absolute timer (interval = reschedule_cb = 0)	1255	=item * absolute timer (at = time, interval = reschedule_cb = 0)
609		1256
610	In this configuration the watcher triggers an event at the wallclock time	1257	In this configuration the watcher triggers an event at the wallclock time
611	C<at> and doesn't repeat. It will not adjust when a time jump occurs,	1258	C<at> and doesn't repeat. It will not adjust when a time jump occurs,
612	that is, if it is to be run at January 1st 2011 then it will run when the	1259	that is, if it is to be run at January 1st 2011 then it will run when the
613	system time reaches or surpasses this time.	1260	system time reaches or surpasses this time.
614		1261
615	=item * non-repeating interval timer (interval > 0, reschedule_cb = 0)	1262	=item * non-repeating interval timer (at = offset, interval > 0, reschedule_cb = 0)
616		1263
617	In this mode the watcher will always be scheduled to time out at the next	1264	In this mode the watcher will always be scheduled to time out at the next
618	C<at + N * interval> time (for some integer N) and then repeat, regardless	1265	C<at + N * interval> time (for some integer N, which can also be negative)
619	of any time jumps.	1266	and then repeat, regardless of any time jumps.
620		1267
621	This can be used to create timers that do not drift with respect to system	1268	This can be used to create timers that do not drift with respect to system
622	time:	1269	time:
623		1270
624	ev_periodic_set (&periodic, 0., 3600., 0);	1271	ev_periodic_set (&periodic, 0., 3600., 0);
…		…
630		1277
631	Another way to think about it (for the mathematically inclined) is that	1278	Another way to think about it (for the mathematically inclined) is that
632	C<ev_periodic> will try to run the callback in this mode at the next possible	1279	C<ev_periodic> will try to run the callback in this mode at the next possible
633	time where C<time = at (mod interval)>, regardless of any time jumps.	1280	time where C<time = at (mod interval)>, regardless of any time jumps.
634		1281
		1282	For numerical stability it is preferable that the C<at> value is near
		1283	C<ev_now ()> (the current time), but there is no range requirement for
		1284	this value.
		1285
635	=item * manual reschedule mode (reschedule_cb = callback)	1286	=item * manual reschedule mode (at and interval ignored, reschedule_cb = callback)
636		1287
637	In this mode the values for C<interval> and C<at> are both being	1288	In this mode the values for C<interval> and C<at> are both being
638	ignored. Instead, each time the periodic watcher gets scheduled, the	1289	ignored. Instead, each time the periodic watcher gets scheduled, the
639	reschedule callback will be called with the watcher as first, and the	1290	reschedule callback will be called with the watcher as first, and the
640	current time as second argument.	1291	current time as second argument.
641		1292
642	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher,	1293	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher,
643	ever, or make any event loop modifications>. If you need to stop it,	1294	ever, or make any event loop modifications>. If you need to stop it,
644	return C<now + 1e30> (or so, fudge fudge) and stop it afterwards (e.g. by	1295	return C<now + 1e30> (or so, fudge fudge) and stop it afterwards (e.g. by
645	starting a prepare watcher).	1296	starting an C<ev_prepare> watcher, which is legal).
646		1297
647	Its prototype is C<ev_tstamp (*reschedule_cb)(struct ev_periodic *w,	1298	Its prototype is C<ev_tstamp (*reschedule_cb)(struct ev_periodic *w,
648	ev_tstamp now)>, e.g.:	1299	ev_tstamp now)>, e.g.:
649		1300
650	static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)	1301	static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)
…		…
673	Simply stops and restarts the periodic watcher again. This is only useful	1324	Simply stops and restarts the periodic watcher again. This is only useful
674	when you changed some parameters or the reschedule callback would return	1325	when you changed some parameters or the reschedule callback would return
675	a different time than the last time it was called (e.g. in a crond like	1326	a different time than the last time it was called (e.g. in a crond like
676	program when the crontabs have changed).	1327	program when the crontabs have changed).
677		1328
		1329	=item ev_tstamp offset [read-write]
		1330
		1331	When repeating, this contains the offset value, otherwise this is the
		1332	absolute point in time (the C<at> value passed to C<ev_periodic_set>).
		1333
		1334	Can be modified any time, but changes only take effect when the periodic
		1335	timer fires or C<ev_periodic_again> is being called.
		1336
		1337	=item ev_tstamp interval [read-write]
		1338
		1339	The current interval value. Can be modified any time, but changes only
		1340	take effect when the periodic timer fires or C<ev_periodic_again> is being
		1341	called.
		1342
		1343	=item ev_tstamp (*reschedule_cb)(struct ev_periodic *w, ev_tstamp now) [read-write]
		1344
		1345	The current reschedule callback, or C<0>, if this functionality is
		1346	switched off. Can be changed any time, but changes only take effect when
		1347	the periodic timer fires or C<ev_periodic_again> is being called.
		1348
		1349	=item ev_tstamp at [read-only]
		1350
		1351	When active, contains the absolute time that the watcher is supposed to
		1352	trigger next.
		1353
678	=back	1354	=back
679		1355
		1356	=head3 Examples
		1357
		1358	Example: Call a callback every hour, or, more precisely, whenever the
		1359	system clock is divisible by 3600. The callback invocation times have
		1360	potentially a lot of jittering, but good long-term stability.
		1361
		1362	static void
		1363	clock_cb (struct ev_loop *loop, struct ev_io *w, int revents)
		1364	{
		1365	... its now a full hour (UTC, or TAI or whatever your clock follows)
		1366	}
		1367
		1368	struct ev_periodic hourly_tick;
		1369	ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);
		1370	ev_periodic_start (loop, &hourly_tick);
		1371
		1372	Example: The same as above, but use a reschedule callback to do it:
		1373
		1374	#include <math.h>
		1375
		1376	static ev_tstamp
		1377	my_scheduler_cb (struct ev_periodic *w, ev_tstamp now)
		1378	{
		1379	return fmod (now, 3600.) + 3600.;
		1380	}
		1381
		1382	ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
		1383
		1384	Example: Call a callback every hour, starting now:
		1385
		1386	struct ev_periodic hourly_tick;
		1387	ev_periodic_init (&hourly_tick, clock_cb,
		1388	fmod (ev_now (loop), 3600.), 3600., 0);
		1389	ev_periodic_start (loop, &hourly_tick);
		1390
		1391
680	=head2 C<ev_signal> - signal me when a signal gets signalled	1392	=head2 C<ev_signal> - signal me when a signal gets signalled!
681		1393
682	Signal watchers will trigger an event when the process receives a specific	1394	Signal watchers will trigger an event when the process receives a specific
683	signal one or more times. Even though signals are very asynchronous, libev	1395	signal one or more times. Even though signals are very asynchronous, libev
684	will try it's best to deliver signals synchronously, i.e. as part of the	1396	will try it's best to deliver signals synchronously, i.e. as part of the
685	normal event processing, like any other event.	1397	normal event processing, like any other event.
…		…
689	with the kernel (thus it coexists with your own signal handlers as long	1401	with the kernel (thus it coexists with your own signal handlers as long
690	as you don't register any with libev). Similarly, when the last signal	1402	as you don't register any with libev). Similarly, when the last signal
691	watcher for a signal is stopped libev will reset the signal handler to	1403	watcher for a signal is stopped libev will reset the signal handler to
692	SIG_DFL (regardless of what it was set to before).	1404	SIG_DFL (regardless of what it was set to before).
693		1405
		1406	=head3 Watcher-Specific Functions and Data Members
		1407
694	=over 4	1408	=over 4
695		1409
696	=item ev_signal_init (ev_signal *, callback, int signum)	1410	=item ev_signal_init (ev_signal *, callback, int signum)
697		1411
698	=item ev_signal_set (ev_signal *, int signum)	1412	=item ev_signal_set (ev_signal *, int signum)
699		1413
700	Configures the watcher to trigger on the given signal number (usually one	1414	Configures the watcher to trigger on the given signal number (usually one
701	of the C<SIGxxx> constants).	1415	of the C<SIGxxx> constants).
702		1416
		1417	=item int signum [read-only]
		1418
		1419	The signal the watcher watches out for.
		1420
703	=back	1421	=back
704		1422
		1423
705	=head2 C<ev_child> - wait for pid status changes	1424	=head2 C<ev_child> - watch out for process status changes
706		1425
707	Child watchers trigger when your process receives a SIGCHLD in response to	1426	Child watchers trigger when your process receives a SIGCHLD in response to
708	some child status changes (most typically when a child of yours dies).	1427	some child status changes (most typically when a child of yours dies).
		1428
		1429	=head3 Watcher-Specific Functions and Data Members
709		1430
710	=over 4	1431	=over 4
711		1432
712	=item ev_child_init (ev_child *, callback, int pid)	1433	=item ev_child_init (ev_child *, callback, int pid)
713		1434
…		…
718	at the C<rstatus> member of the C<ev_child> watcher structure to see	1439	at the C<rstatus> member of the C<ev_child> watcher structure to see
719	the status word (use the macros from C<sys/wait.h> and see your systems	1440	the status word (use the macros from C<sys/wait.h> and see your systems
720	C<waitpid> documentation). The C<rpid> member contains the pid of the	1441	C<waitpid> documentation). The C<rpid> member contains the pid of the
721	process causing the status change.	1442	process causing the status change.
722		1443
		1444	=item int pid [read-only]
		1445
		1446	The process id this watcher watches out for, or C<0>, meaning any process id.
		1447
		1448	=item int rpid [read-write]
		1449
		1450	The process id that detected a status change.
		1451
		1452	=item int rstatus [read-write]
		1453
		1454	The process exit/trace status caused by C<rpid> (see your systems
		1455	C<waitpid> and C<sys/wait.h> documentation for details).
		1456
723	=back	1457	=back
724		1458
		1459	=head3 Examples
		1460
		1461	Example: Try to exit cleanly on SIGINT and SIGTERM.
		1462
		1463	static void
		1464	sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)
		1465	{
		1466	ev_unloop (loop, EVUNLOOP_ALL);
		1467	}
		1468
		1469	struct ev_signal signal_watcher;
		1470	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
		1471	ev_signal_start (loop, &sigint_cb);
		1472
		1473
		1474	=head2 C<ev_stat> - did the file attributes just change?
		1475
		1476	This watches a filesystem path for attribute changes. That is, it calls
		1477	C<stat> regularly (or when the OS says it changed) and sees if it changed
		1478	compared to the last time, invoking the callback if it did.
		1479
		1480	The path does not need to exist: changing from "path exists" to "path does
		1481	not exist" is a status change like any other. The condition "path does
		1482	not exist" is signified by the C<st_nlink> field being zero (which is
		1483	otherwise always forced to be at least one) and all the other fields of
		1484	the stat buffer having unspecified contents.
		1485
		1486	The path I<should> be absolute and I<must not> end in a slash. If it is
		1487	relative and your working directory changes, the behaviour is undefined.
		1488
		1489	Since there is no standard to do this, the portable implementation simply
		1490	calls C<stat (2)> regularly on the path to see if it changed somehow. You
		1491	can specify a recommended polling interval for this case. If you specify
		1492	a polling interval of C<0> (highly recommended!) then a I<suitable,
		1493	unspecified default> value will be used (which you can expect to be around
		1494	five seconds, although this might change dynamically). Libev will also
		1495	impose a minimum interval which is currently around C<0.1>, but thats
		1496	usually overkill.
		1497
		1498	This watcher type is not meant for massive numbers of stat watchers,
		1499	as even with OS-supported change notifications, this can be
		1500	resource-intensive.
		1501
		1502	At the time of this writing, only the Linux inotify interface is
		1503	implemented (implementing kqueue support is left as an exercise for the
		1504	reader). Inotify will be used to give hints only and should not change the
		1505	semantics of C<ev_stat> watchers, which means that libev sometimes needs
		1506	to fall back to regular polling again even with inotify, but changes are
		1507	usually detected immediately, and if the file exists there will be no
		1508	polling.
		1509
		1510	=head3 Inotify
		1511
		1512	When C<inotify (7)> support has been compiled into libev (generally only
		1513	available on Linux) and present at runtime, it will be used to speed up
		1514	change detection where possible. The inotify descriptor will be created lazily
		1515	when the first C<ev_stat> watcher is being started.
		1516
		1517	Inotify presense does not change the semantics of C<ev_stat> watchers
		1518	except that changes might be detected earlier, and in some cases, to avoid
		1519	making regular C<stat> calls. Even in the presense of inotify support
		1520	there are many cases where libev has to resort to regular C<stat> polling.
		1521
		1522	(There is no support for kqueue, as apparently it cannot be used to
		1523	implement this functionality, due to the requirement of having a file
		1524	descriptor open on the object at all times).
		1525
		1526	=head3 The special problem of stat time resolution
		1527
		1528	The C<stat ()> syscall only supports full-second resolution portably, and
		1529	even on systems where the resolution is higher, many filesystems still
		1530	only support whole seconds.
		1531
		1532	That means that, if the time is the only thing that changes, you might
		1533	miss updates: on the first update, C<ev_stat> detects a change and calls
		1534	your callback, which does something. When there is another update within
		1535	the same second, C<ev_stat> will be unable to detect it.
		1536
		1537	The solution to this is to delay acting on a change for a second (or till
		1538	the next second boundary), using a roughly one-second delay C<ev_timer>
		1539	(C<ev_timer_set (w, 0., 1.01); ev_timer_again (loop, w)>). The C<.01>
		1540	is added to work around small timing inconsistencies of some operating
		1541	systems.
		1542
		1543	=head3 Watcher-Specific Functions and Data Members
		1544
		1545	=over 4
		1546
		1547	=item ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)
		1548
		1549	=item ev_stat_set (ev_stat *, const char *path, ev_tstamp interval)
		1550
		1551	Configures the watcher to wait for status changes of the given
		1552	C<path>. The C<interval> is a hint on how quickly a change is expected to
		1553	be detected and should normally be specified as C<0> to let libev choose
		1554	a suitable value. The memory pointed to by C<path> must point to the same
		1555	path for as long as the watcher is active.
		1556
		1557	The callback will be receive C<EV_STAT> when a change was detected,
		1558	relative to the attributes at the time the watcher was started (or the
		1559	last change was detected).
		1560
		1561	=item ev_stat_stat (ev_stat *)
		1562
		1563	Updates the stat buffer immediately with new values. If you change the
		1564	watched path in your callback, you could call this fucntion to avoid
		1565	detecting this change (while introducing a race condition). Can also be
		1566	useful simply to find out the new values.
		1567
		1568	=item ev_statdata attr [read-only]
		1569
		1570	The most-recently detected attributes of the file. Although the type is of
		1571	C<ev_statdata>, this is usually the (or one of the) C<struct stat> types
		1572	suitable for your system. If the C<st_nlink> member is C<0>, then there
		1573	was some error while C<stat>ing the file.
		1574
		1575	=item ev_statdata prev [read-only]
		1576
		1577	The previous attributes of the file. The callback gets invoked whenever
		1578	C<prev> != C<attr>.
		1579
		1580	=item ev_tstamp interval [read-only]
		1581
		1582	The specified interval.
		1583
		1584	=item const char *path [read-only]
		1585
		1586	The filesystem path that is being watched.
		1587
		1588	=back
		1589
		1590	=head3 Examples
		1591
		1592	Example: Watch C</etc/passwd> for attribute changes.
		1593
		1594	static void
		1595	passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)
		1596	{
		1597	/* /etc/passwd changed in some way */
		1598	if (w->attr.st_nlink)
		1599	{
		1600	printf ("passwd current size %ld\n", (long)w->attr.st_size);
		1601	printf ("passwd current atime %ld\n", (long)w->attr.st_mtime);
		1602	printf ("passwd current mtime %ld\n", (long)w->attr.st_mtime);
		1603	}
		1604	else
		1605	/* you shalt not abuse printf for puts */
		1606	puts ("wow, /etc/passwd is not there, expect problems. "
		1607	"if this is windows, they already arrived\n");
		1608	}
		1609
		1610	...
		1611	ev_stat passwd;
		1612
		1613	ev_stat_init (&passwd, passwd_cb, "/etc/passwd", 0.);
		1614	ev_stat_start (loop, &passwd);
		1615
		1616	Example: Like above, but additionally use a one-second delay so we do not
		1617	miss updates (however, frequent updates will delay processing, too, so
		1618	one might do the work both on C<ev_stat> callback invocation I<and> on
		1619	C<ev_timer> callback invocation).
		1620
		1621	static ev_stat passwd;
		1622	static ev_timer timer;
		1623
		1624	static void
		1625	timer_cb (EV_P_ ev_timer *w, int revents)
		1626	{
		1627	ev_timer_stop (EV_A_ w);
		1628
		1629	/* now it's one second after the most recent passwd change */
		1630	}
		1631
		1632	static void
		1633	stat_cb (EV_P_ ev_stat *w, int revents)
		1634	{
		1635	/* reset the one-second timer */
		1636	ev_timer_again (EV_A_ &timer);
		1637	}
		1638
		1639	...
		1640	ev_stat_init (&passwd, stat_cb, "/etc/passwd", 0.);
		1641	ev_stat_start (loop, &passwd);
		1642	ev_timer_init (&timer, timer_cb, 0., 1.01);
		1643
		1644
725	=head2 C<ev_idle> - when you've got nothing better to do	1645	=head2 C<ev_idle> - when you've got nothing better to do...
726		1646
727	Idle watchers trigger events when there are no other events are pending	1647	Idle watchers trigger events when no other events of the same or higher
728	(prepare, check and other idle watchers do not count). That is, as long	1648	priority are pending (prepare, check and other idle watchers do not
729	as your process is busy handling sockets or timeouts (or even signals,	1649	count).
730	imagine) it will not be triggered. But when your process is idle all idle	1650
731	watchers are being called again and again, once per event loop iteration -	1651	That is, as long as your process is busy handling sockets or timeouts
		1652	(or even signals, imagine) of the same or higher priority it will not be
		1653	triggered. But when your process is idle (or only lower-priority watchers
		1654	are pending), the idle watchers are being called once per event loop
732	until stopped, that is, or your process receives more events and becomes	1655	iteration - until stopped, that is, or your process receives more events
733	busy.	1656	and becomes busy again with higher priority stuff.
734		1657
735	The most noteworthy effect is that as long as any idle watchers are	1658	The most noteworthy effect is that as long as any idle watchers are
736	active, the process will not block when waiting for new events.	1659	active, the process will not block when waiting for new events.
737		1660
738	Apart from keeping your process non-blocking (which is a useful	1661	Apart from keeping your process non-blocking (which is a useful
739	effect on its own sometimes), idle watchers are a good place to do	1662	effect on its own sometimes), idle watchers are a good place to do
740	"pseudo-background processing", or delay processing stuff to after the	1663	"pseudo-background processing", or delay processing stuff to after the
741	event loop has handled all outstanding events.	1664	event loop has handled all outstanding events.
742		1665
		1666	=head3 Watcher-Specific Functions and Data Members
		1667
743	=over 4	1668	=over 4
744		1669
745	=item ev_idle_init (ev_signal *, callback)	1670	=item ev_idle_init (ev_signal *, callback)
746		1671
747	Initialises and configures the idle watcher - it has no parameters of any	1672	Initialises and configures the idle watcher - it has no parameters of any
748	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,	1673	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,
749	believe me.	1674	believe me.
750		1675
751	=back	1676	=back
752		1677
		1678	=head3 Examples
		1679
		1680	Example: Dynamically allocate an C<ev_idle> watcher, start it, and in the
		1681	callback, free it. Also, use no error checking, as usual.
		1682
		1683	static void
		1684	idle_cb (struct ev_loop *loop, struct ev_idle *w, int revents)
		1685	{
		1686	free (w);
		1687	// now do something you wanted to do when the program has
		1688	// no longer asnything immediate to do.
		1689	}
		1690
		1691	struct ev_idle *idle_watcher = malloc (sizeof (struct ev_idle));
		1692	ev_idle_init (idle_watcher, idle_cb);
		1693	ev_idle_start (loop, idle_cb);
		1694
		1695
753	=head2 C<ev_prepare> and C<ev_check> - customise your event loop	1696	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!
754		1697
755	Prepare and check watchers are usually (but not always) used in tandem:	1698	Prepare and check watchers are usually (but not always) used in tandem:
756	prepare watchers get invoked before the process blocks and check watchers	1699	prepare watchers get invoked before the process blocks and check watchers
757	afterwards.	1700	afterwards.
758		1701
		1702	You I<must not> call C<ev_loop> or similar functions that enter
		1703	the current event loop from either C<ev_prepare> or C<ev_check>
		1704	watchers. Other loops than the current one are fine, however. The
		1705	rationale behind this is that you do not need to check for recursion in
		1706	those watchers, i.e. the sequence will always be C<ev_prepare>, blocking,
		1707	C<ev_check> so if you have one watcher of each kind they will always be
		1708	called in pairs bracketing the blocking call.
		1709
759	Their main purpose is to integrate other event mechanisms into libev. This	1710	Their main purpose is to integrate other event mechanisms into libev and
760	could be used, for example, to track variable changes, implement your own	1711	their use is somewhat advanced. This could be used, for example, to track
761	watchers, integrate net-snmp or a coroutine library and lots more.	1712	variable changes, implement your own watchers, integrate net-snmp or a
		1713	coroutine library and lots more. They are also occasionally useful if
		1714	you cache some data and want to flush it before blocking (for example,
		1715	in X programs you might want to do an C<XFlush ()> in an C<ev_prepare>
		1716	watcher).
762		1717
763	This is done by examining in each prepare call which file descriptors need	1718	This is done by examining in each prepare call which file descriptors need
764	to be watched by the other library, registering C<ev_io> watchers for	1719	to be watched by the other library, registering C<ev_io> watchers for
765	them and starting an C<ev_timer> watcher for any timeouts (many libraries	1720	them and starting an C<ev_timer> watcher for any timeouts (many libraries
766	provide just this functionality). Then, in the check watcher you check for	1721	provide just this functionality). Then, in the check watcher you check for
…		…
776	with priority higher than or equal to the event loop and one coroutine	1731	with priority higher than or equal to the event loop and one coroutine
777	of lower priority, but only once, using idle watchers to keep the event	1732	of lower priority, but only once, using idle watchers to keep the event
778	loop from blocking if lower-priority coroutines are active, thus mapping	1733	loop from blocking if lower-priority coroutines are active, thus mapping
779	low-priority coroutines to idle/background tasks).	1734	low-priority coroutines to idle/background tasks).
780		1735
		1736	It is recommended to give C<ev_check> watchers highest (C<EV_MAXPRI>)
		1737	priority, to ensure that they are being run before any other watchers
		1738	after the poll. Also, C<ev_check> watchers (and C<ev_prepare> watchers,
		1739	too) should not activate ("feed") events into libev. While libev fully
		1740	supports this, they will be called before other C<ev_check> watchers
		1741	did their job. As C<ev_check> watchers are often used to embed other
		1742	(non-libev) event loops those other event loops might be in an unusable
		1743	state until their C<ev_check> watcher ran (always remind yourself to
		1744	coexist peacefully with others).
		1745
		1746	=head3 Watcher-Specific Functions and Data Members
		1747
781	=over 4	1748	=over 4
782		1749
783	=item ev_prepare_init (ev_prepare *, callback)	1750	=item ev_prepare_init (ev_prepare *, callback)
784		1751
785	=item ev_check_init (ev_check *, callback)	1752	=item ev_check_init (ev_check *, callback)
…		…
787	Initialises and configures the prepare or check watcher - they have no	1754	Initialises and configures the prepare or check watcher - they have no
788	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>	1755	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>
789	macros, but using them is utterly, utterly and completely pointless.	1756	macros, but using them is utterly, utterly and completely pointless.
790		1757
791	=back	1758	=back
		1759
		1760	=head3 Examples
		1761
		1762	There are a number of principal ways to embed other event loops or modules
		1763	into libev. Here are some ideas on how to include libadns into libev
		1764	(there is a Perl module named C<EV::ADNS> that does this, which you could
		1765	use for an actually working example. Another Perl module named C<EV::Glib>
		1766	embeds a Glib main context into libev, and finally, C<Glib::EV> embeds EV
		1767	into the Glib event loop).
		1768
		1769	Method 1: Add IO watchers and a timeout watcher in a prepare handler,
		1770	and in a check watcher, destroy them and call into libadns. What follows
		1771	is pseudo-code only of course. This requires you to either use a low
		1772	priority for the check watcher or use C<ev_clear_pending> explicitly, as
		1773	the callbacks for the IO/timeout watchers might not have been called yet.
		1774
		1775	static ev_io iow [nfd];
		1776	static ev_timer tw;
		1777
		1778	static void
		1779	io_cb (ev_loop *loop, ev_io *w, int revents)
		1780	{
		1781	}
		1782
		1783	// create io watchers for each fd and a timer before blocking
		1784	static void
		1785	adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents)
		1786	{
		1787	int timeout = 3600000;
		1788	struct pollfd fds [nfd];
		1789	// actual code will need to loop here and realloc etc.
		1790	adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));
		1791
		1792	/* the callback is illegal, but won't be called as we stop during check */
		1793	ev_timer_init (&tw, 0, timeout * 1e-3);
		1794	ev_timer_start (loop, &tw);
		1795
		1796	// create one ev_io per pollfd
		1797	for (int i = 0; i < nfd; ++i)
		1798	{
		1799	ev_io_init (iow + i, io_cb, fds [i].fd,
		1800	((fds [i].events & POLLIN ? EV_READ : 0)
		1801	| (fds [i].events & POLLOUT ? EV_WRITE : 0)));
		1802
		1803	fds [i].revents = 0;
		1804	ev_io_start (loop, iow + i);
		1805	}
		1806	}
		1807
		1808	// stop all watchers after blocking
		1809	static void
		1810	adns_check_cb (ev_loop *loop, ev_check *w, int revents)
		1811	{
		1812	ev_timer_stop (loop, &tw);
		1813
		1814	for (int i = 0; i < nfd; ++i)
		1815	{
		1816	// set the relevant poll flags
		1817	// could also call adns_processreadable etc. here
		1818	struct pollfd *fd = fds + i;
		1819	int revents = ev_clear_pending (iow + i);
		1820	if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;
		1821	if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;
		1822
		1823	// now stop the watcher
		1824	ev_io_stop (loop, iow + i);
		1825	}
		1826
		1827	adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));
		1828	}
		1829
		1830	Method 2: This would be just like method 1, but you run C<adns_afterpoll>
		1831	in the prepare watcher and would dispose of the check watcher.
		1832
		1833	Method 3: If the module to be embedded supports explicit event
		1834	notification (adns does), you can also make use of the actual watcher
		1835	callbacks, and only destroy/create the watchers in the prepare watcher.
		1836
		1837	static void
		1838	timer_cb (EV_P_ ev_timer *w, int revents)
		1839	{
		1840	adns_state ads = (adns_state)w->data;
		1841	update_now (EV_A);
		1842
		1843	adns_processtimeouts (ads, &tv_now);
		1844	}
		1845
		1846	static void
		1847	io_cb (EV_P_ ev_io *w, int revents)
		1848	{
		1849	adns_state ads = (adns_state)w->data;
		1850	update_now (EV_A);
		1851
		1852	if (revents & EV_READ ) adns_processreadable (ads, w->fd, &tv_now);
		1853	if (revents & EV_WRITE) adns_processwriteable (ads, w->fd, &tv_now);
		1854	}
		1855
		1856	// do not ever call adns_afterpoll
		1857
		1858	Method 4: Do not use a prepare or check watcher because the module you
		1859	want to embed is too inflexible to support it. Instead, youc na override
		1860	their poll function. The drawback with this solution is that the main
		1861	loop is now no longer controllable by EV. The C<Glib::EV> module does
		1862	this.
		1863
		1864	static gint
		1865	event_poll_func (GPollFD *fds, guint nfds, gint timeout)
		1866	{
		1867	int got_events = 0;
		1868
		1869	for (n = 0; n < nfds; ++n)
		1870	// create/start io watcher that sets the relevant bits in fds[n] and increment got_events
		1871
		1872	if (timeout >= 0)
		1873	// create/start timer
		1874
		1875	// poll
		1876	ev_loop (EV_A_ 0);
		1877
		1878	// stop timer again
		1879	if (timeout >= 0)
		1880	ev_timer_stop (EV_A_ &to);
		1881
		1882	// stop io watchers again - their callbacks should have set
		1883	for (n = 0; n < nfds; ++n)
		1884	ev_io_stop (EV_A_ iow [n]);
		1885
		1886	return got_events;
		1887	}
		1888
		1889
		1890	=head2 C<ev_embed> - when one backend isn't enough...
		1891
		1892	This is a rather advanced watcher type that lets you embed one event loop
		1893	into another (currently only C<ev_io> events are supported in the embedded
		1894	loop, other types of watchers might be handled in a delayed or incorrect
		1895	fashion and must not be used).
		1896
		1897	There are primarily two reasons you would want that: work around bugs and
		1898	prioritise I/O.
		1899
		1900	As an example for a bug workaround, the kqueue backend might only support
		1901	sockets on some platform, so it is unusable as generic backend, but you
		1902	still want to make use of it because you have many sockets and it scales
		1903	so nicely. In this case, you would create a kqueue-based loop and embed it
		1904	into your default loop (which might use e.g. poll). Overall operation will
		1905	be a bit slower because first libev has to poll and then call kevent, but
		1906	at least you can use both at what they are best.
		1907
		1908	As for prioritising I/O: rarely you have the case where some fds have
		1909	to be watched and handled very quickly (with low latency), and even
		1910	priorities and idle watchers might have too much overhead. In this case
		1911	you would put all the high priority stuff in one loop and all the rest in
		1912	a second one, and embed the second one in the first.
		1913
		1914	As long as the watcher is active, the callback will be invoked every time
		1915	there might be events pending in the embedded loop. The callback must then
		1916	call C<ev_embed_sweep (mainloop, watcher)> to make a single sweep and invoke
		1917	their callbacks (you could also start an idle watcher to give the embedded
		1918	loop strictly lower priority for example). You can also set the callback
		1919	to C<0>, in which case the embed watcher will automatically execute the
		1920	embedded loop sweep.
		1921
		1922	As long as the watcher is started it will automatically handle events. The
		1923	callback will be invoked whenever some events have been handled. You can
		1924	set the callback to C<0> to avoid having to specify one if you are not
		1925	interested in that.
		1926
		1927	Also, there have not currently been made special provisions for forking:
		1928	when you fork, you not only have to call C<ev_loop_fork> on both loops,
		1929	but you will also have to stop and restart any C<ev_embed> watchers
		1930	yourself.
		1931
		1932	Unfortunately, not all backends are embeddable, only the ones returned by
		1933	C<ev_embeddable_backends> are, which, unfortunately, does not include any
		1934	portable one.
		1935
		1936	So when you want to use this feature you will always have to be prepared
		1937	that you cannot get an embeddable loop. The recommended way to get around
		1938	this is to have a separate variables for your embeddable loop, try to
		1939	create it, and if that fails, use the normal loop for everything.
		1940
		1941	=head3 Watcher-Specific Functions and Data Members
		1942
		1943	=over 4
		1944
		1945	=item ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)
		1946
		1947	=item ev_embed_set (ev_embed *, callback, struct ev_loop *embedded_loop)
		1948
		1949	Configures the watcher to embed the given loop, which must be
		1950	embeddable. If the callback is C<0>, then C<ev_embed_sweep> will be
		1951	invoked automatically, otherwise it is the responsibility of the callback
		1952	to invoke it (it will continue to be called until the sweep has been done,
		1953	if you do not want thta, you need to temporarily stop the embed watcher).
		1954
		1955	=item ev_embed_sweep (loop, ev_embed *)
		1956
		1957	Make a single, non-blocking sweep over the embedded loop. This works
		1958	similarly to C<ev_loop (embedded_loop, EVLOOP_NONBLOCK)>, but in the most
		1959	apropriate way for embedded loops.
		1960
		1961	=item struct ev_loop *other [read-only]
		1962
		1963	The embedded event loop.
		1964
		1965	=back
		1966
		1967	=head3 Examples
		1968
		1969	Example: Try to get an embeddable event loop and embed it into the default
		1970	event loop. If that is not possible, use the default loop. The default
		1971	loop is stored in C<loop_hi>, while the mebeddable loop is stored in
		1972	C<loop_lo> (which is C<loop_hi> in the acse no embeddable loop can be
		1973	used).
		1974
		1975	struct ev_loop *loop_hi = ev_default_init (0);
		1976	struct ev_loop *loop_lo = 0;
		1977	struct ev_embed embed;
		1978
		1979	// see if there is a chance of getting one that works
		1980	// (remember that a flags value of 0 means autodetection)
		1981	loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
		1982	? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
		1983	: 0;
		1984
		1985	// if we got one, then embed it, otherwise default to loop_hi
		1986	if (loop_lo)
		1987	{
		1988	ev_embed_init (&embed, 0, loop_lo);
		1989	ev_embed_start (loop_hi, &embed);
		1990	}
		1991	else
		1992	loop_lo = loop_hi;
		1993
		1994	Example: Check if kqueue is available but not recommended and create
		1995	a kqueue backend for use with sockets (which usually work with any
		1996	kqueue implementation). Store the kqueue/socket-only event loop in
		1997	C<loop_socket>. (One might optionally use C<EVFLAG_NOENV>, too).
		1998
		1999	struct ev_loop *loop = ev_default_init (0);
		2000	struct ev_loop *loop_socket = 0;
		2001	struct ev_embed embed;
		2002
		2003	if (ev_supported_backends () & ~ev_recommended_backends () & EVBACKEND_KQUEUE)
		2004	if ((loop_socket = ev_loop_new (EVBACKEND_KQUEUE))
		2005	{
		2006	ev_embed_init (&embed, 0, loop_socket);
		2007	ev_embed_start (loop, &embed);
		2008	}
		2009
		2010	if (!loop_socket)
		2011	loop_socket = loop;
		2012
		2013	// now use loop_socket for all sockets, and loop for everything else
		2014
		2015
		2016	=head2 C<ev_fork> - the audacity to resume the event loop after a fork
		2017
		2018	Fork watchers are called when a C<fork ()> was detected (usually because
		2019	whoever is a good citizen cared to tell libev about it by calling
		2020	C<ev_default_fork> or C<ev_loop_fork>). The invocation is done before the
		2021	event loop blocks next and before C<ev_check> watchers are being called,
		2022	and only in the child after the fork. If whoever good citizen calling
		2023	C<ev_default_fork> cheats and calls it in the wrong process, the fork
		2024	handlers will be invoked, too, of course.
		2025
		2026	=head3 Watcher-Specific Functions and Data Members
		2027
		2028	=over 4
		2029
		2030	=item ev_fork_init (ev_signal *, callback)
		2031
		2032	Initialises and configures the fork watcher - it has no parameters of any
		2033	kind. There is a C<ev_fork_set> macro, but using it is utterly pointless,
		2034	believe me.
		2035
		2036	=back
		2037
792		2038
793	=head1 OTHER FUNCTIONS	2039	=head1 OTHER FUNCTIONS
794		2040
795	There are some other functions of possible interest. Described. Here. Now.	2041	There are some other functions of possible interest. Described. Here. Now.
796		2042
…		…
826	/* stdin might have data for us, joy! */;	2072	/* stdin might have data for us, joy! */;
827	}	2073	}
828		2074
829	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);	2075	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
830		2076
831	=item ev_feed_event (loop, watcher, int events)	2077	=item ev_feed_event (ev_loop *, watcher *, int revents)
832		2078
833	Feeds the given event set into the event loop, as if the specified event	2079	Feeds the given event set into the event loop, as if the specified event
834	had happened for the specified watcher (which must be a pointer to an	2080	had happened for the specified watcher (which must be a pointer to an
835	initialised but not necessarily started event watcher).	2081	initialised but not necessarily started event watcher).
836		2082
837	=item ev_feed_fd_event (loop, int fd, int revents)	2083	=item ev_feed_fd_event (ev_loop *, int fd, int revents)
838		2084
839	Feed an event on the given fd, as if a file descriptor backend detected	2085	Feed an event on the given fd, as if a file descriptor backend detected
840	the given events it.	2086	the given events it.
841		2087
842	=item ev_feed_signal_event (loop, int signum)	2088	=item ev_feed_signal_event (ev_loop *loop, int signum)
843		2089
844	Feed an event as if the given signal occured (loop must be the default loop!).	2090	Feed an event as if the given signal occured (C<loop> must be the default
		2091	loop!).
845		2092
846	=back	2093	=back
		2094
847		2095
848	=head1 LIBEVENT EMULATION	2096	=head1 LIBEVENT EMULATION
849		2097
850	Libev offers a compatibility emulation layer for libevent. It cannot	2098	Libev offers a compatibility emulation layer for libevent. It cannot
851	emulate the internals of libevent, so here are some usage hints:	2099	emulate the internals of libevent, so here are some usage hints:
…		…
872		2120
873	=back	2121	=back
874		2122
875	=head1 C++ SUPPORT	2123	=head1 C++ SUPPORT
876		2124
877	TBD.	2125	Libev comes with some simplistic wrapper classes for C++ that mainly allow
		2126	you to use some convinience methods to start/stop watchers and also change
		2127	the callback model to a model using method callbacks on objects.
		2128
		2129	To use it,
		2130
		2131	#include <ev++.h>
		2132
		2133	This automatically includes F<ev.h> and puts all of its definitions (many
		2134	of them macros) into the global namespace. All C++ specific things are
		2135	put into the C<ev> namespace. It should support all the same embedding
		2136	options as F<ev.h>, most notably C<EV_MULTIPLICITY>.
		2137
		2138	Care has been taken to keep the overhead low. The only data member the C++
		2139	classes add (compared to plain C-style watchers) is the event loop pointer
		2140	that the watcher is associated with (or no additional members at all if
		2141	you disable C<EV_MULTIPLICITY> when embedding libev).
		2142
		2143	Currently, functions, and static and non-static member functions can be
		2144	used as callbacks. Other types should be easy to add as long as they only
		2145	need one additional pointer for context. If you need support for other
		2146	types of functors please contact the author (preferably after implementing
		2147	it).
		2148
		2149	Here is a list of things available in the C<ev> namespace:
		2150
		2151	=over 4
		2152
		2153	=item C<ev::READ>, C<ev::WRITE> etc.
		2154
		2155	These are just enum values with the same values as the C<EV_READ> etc.
		2156	macros from F<ev.h>.
		2157
		2158	=item C<ev::tstamp>, C<ev::now>
		2159
		2160	Aliases to the same types/functions as with the C<ev_> prefix.
		2161
		2162	=item C<ev::io>, C<ev::timer>, C<ev::periodic>, C<ev::idle>, C<ev::sig> etc.
		2163
		2164	For each C<ev_TYPE> watcher in F<ev.h> there is a corresponding class of
		2165	the same name in the C<ev> namespace, with the exception of C<ev_signal>
		2166	which is called C<ev::sig> to avoid clashes with the C<signal> macro
		2167	defines by many implementations.
		2168
		2169	All of those classes have these methods:
		2170
		2171	=over 4
		2172
		2173	=item ev::TYPE::TYPE ()
		2174
		2175	=item ev::TYPE::TYPE (struct ev_loop *)
		2176
		2177	=item ev::TYPE::~TYPE
		2178
		2179	The constructor (optionally) takes an event loop to associate the watcher
		2180	with. If it is omitted, it will use C<EV_DEFAULT>.
		2181
		2182	The constructor calls C<ev_init> for you, which means you have to call the
		2183	C<set> method before starting it.
		2184
		2185	It will not set a callback, however: You have to call the templated C<set>
		2186	method to set a callback before you can start the watcher.
		2187
		2188	(The reason why you have to use a method is a limitation in C++ which does
		2189	not allow explicit template arguments for constructors).
		2190
		2191	The destructor automatically stops the watcher if it is active.
		2192
		2193	=item w->set<class, &class::method> (object *)
		2194
		2195	This method sets the callback method to call. The method has to have a
		2196	signature of C<void (*)(ev_TYPE &, int)>, it receives the watcher as
		2197	first argument and the C<revents> as second. The object must be given as
		2198	parameter and is stored in the C<data> member of the watcher.
		2199
		2200	This method synthesizes efficient thunking code to call your method from
		2201	the C callback that libev requires. If your compiler can inline your
		2202	callback (i.e. it is visible to it at the place of the C<set> call and
		2203	your compiler is good :), then the method will be fully inlined into the
		2204	thunking function, making it as fast as a direct C callback.
		2205
		2206	Example: simple class declaration and watcher initialisation
		2207
		2208	struct myclass
		2209	{
		2210	void io_cb (ev::io &w, int revents) { }
		2211	}
		2212
		2213	myclass obj;
		2214	ev::io iow;
		2215	iow.set <myclass, &myclass::io_cb> (&obj);
		2216
		2217	=item w->set<function> (void *data = 0)
		2218
		2219	Also sets a callback, but uses a static method or plain function as
		2220	callback. The optional C<data> argument will be stored in the watcher's
		2221	C<data> member and is free for you to use.
		2222
		2223	The prototype of the C<function> must be C<void (*)(ev::TYPE &w, int)>.
		2224
		2225	See the method-C<set> above for more details.
		2226
		2227	Example:
		2228
		2229	static void io_cb (ev::io &w, int revents) { }
		2230	iow.set <io_cb> ();
		2231
		2232	=item w->set (struct ev_loop *)
		2233
		2234	Associates a different C<struct ev_loop> with this watcher. You can only
		2235	do this when the watcher is inactive (and not pending either).
		2236
		2237	=item w->set ([args])
		2238
		2239	Basically the same as C<ev_TYPE_set>, with the same args. Must be
		2240	called at least once. Unlike the C counterpart, an active watcher gets
		2241	automatically stopped and restarted when reconfiguring it with this
		2242	method.
		2243
		2244	=item w->start ()
		2245
		2246	Starts the watcher. Note that there is no C<loop> argument, as the
		2247	constructor already stores the event loop.
		2248
		2249	=item w->stop ()
		2250
		2251	Stops the watcher if it is active. Again, no C<loop> argument.
		2252
		2253	=item w->again () (C<ev::timer>, C<ev::periodic> only)
		2254
		2255	For C<ev::timer> and C<ev::periodic>, this invokes the corresponding
		2256	C<ev_TYPE_again> function.
		2257
		2258	=item w->sweep () (C<ev::embed> only)
		2259
		2260	Invokes C<ev_embed_sweep>.
		2261
		2262	=item w->update () (C<ev::stat> only)
		2263
		2264	Invokes C<ev_stat_stat>.
		2265
		2266	=back
		2267
		2268	=back
		2269
		2270	Example: Define a class with an IO and idle watcher, start one of them in
		2271	the constructor.
		2272
		2273	class myclass
		2274	{
		2275	ev_io io; void io_cb (ev::io &w, int revents);
		2276	ev_idle idle void idle_cb (ev::idle &w, int revents);
		2277
		2278	myclass ();
		2279	}
		2280
		2281	myclass::myclass (int fd)
		2282	{
		2283	io .set <myclass, &myclass::io_cb > (this);
		2284	idle.set <myclass, &myclass::idle_cb> (this);
		2285
		2286	io.start (fd, ev::READ);
		2287	}
		2288
		2289
		2290	=head1 MACRO MAGIC
		2291
		2292	Libev can be compiled with a variety of options, the most fundamantal
		2293	of which is C<EV_MULTIPLICITY>. This option determines whether (most)
		2294	functions and callbacks have an initial C<struct ev_loop *> argument.
		2295
		2296	To make it easier to write programs that cope with either variant, the
		2297	following macros are defined:
		2298
		2299	=over 4
		2300
		2301	=item C<EV_A>, C<EV_A_>
		2302
		2303	This provides the loop I<argument> for functions, if one is required ("ev
		2304	loop argument"). The C<EV_A> form is used when this is the sole argument,
		2305	C<EV_A_> is used when other arguments are following. Example:
		2306
		2307	ev_unref (EV_A);
		2308	ev_timer_add (EV_A_ watcher);
		2309	ev_loop (EV_A_ 0);
		2310
		2311	It assumes the variable C<loop> of type C<struct ev_loop *> is in scope,
		2312	which is often provided by the following macro.
		2313
		2314	=item C<EV_P>, C<EV_P_>
		2315
		2316	This provides the loop I<parameter> for functions, if one is required ("ev
		2317	loop parameter"). The C<EV_P> form is used when this is the sole parameter,
		2318	C<EV_P_> is used when other parameters are following. Example:
		2319
		2320	// this is how ev_unref is being declared
		2321	static void ev_unref (EV_P);
		2322
		2323	// this is how you can declare your typical callback
		2324	static void cb (EV_P_ ev_timer *w, int revents)
		2325
		2326	It declares a parameter C<loop> of type C<struct ev_loop *>, quite
		2327	suitable for use with C<EV_A>.
		2328
		2329	=item C<EV_DEFAULT>, C<EV_DEFAULT_>
		2330
		2331	Similar to the other two macros, this gives you the value of the default
		2332	loop, if multiple loops are supported ("ev loop default").
		2333
		2334	=back
		2335
		2336	Example: Declare and initialise a check watcher, utilising the above
		2337	macros so it will work regardless of whether multiple loops are supported
		2338	or not.
		2339
		2340	static void
		2341	check_cb (EV_P_ ev_timer *w, int revents)
		2342	{
		2343	ev_check_stop (EV_A_ w);
		2344	}
		2345
		2346	ev_check check;
		2347	ev_check_init (&check, check_cb);
		2348	ev_check_start (EV_DEFAULT_ &check);
		2349	ev_loop (EV_DEFAULT_ 0);
		2350
		2351	=head1 EMBEDDING
		2352
		2353	Libev can (and often is) directly embedded into host
		2354	applications. Examples of applications that embed it include the Deliantra
		2355	Game Server, the EV perl module, the GNU Virtual Private Ethernet (gvpe)
		2356	and rxvt-unicode.
		2357
		2358	The goal is to enable you to just copy the necessary files into your
		2359	source directory without having to change even a single line in them, so
		2360	you can easily upgrade by simply copying (or having a checked-out copy of
		2361	libev somewhere in your source tree).
		2362
		2363	=head2 FILESETS
		2364
		2365	Depending on what features you need you need to include one or more sets of files
		2366	in your app.
		2367
		2368	=head3 CORE EVENT LOOP
		2369
		2370	To include only the libev core (all the C<ev_*> functions), with manual
		2371	configuration (no autoconf):
		2372
		2373	#define EV_STANDALONE 1
		2374	#include "ev.c"
		2375
		2376	This will automatically include F<ev.h>, too, and should be done in a
		2377	single C source file only to provide the function implementations. To use
		2378	it, do the same for F<ev.h> in all files wishing to use this API (best
		2379	done by writing a wrapper around F<ev.h> that you can include instead and
		2380	where you can put other configuration options):
		2381
		2382	#define EV_STANDALONE 1
		2383	#include "ev.h"
		2384
		2385	Both header files and implementation files can be compiled with a C++
		2386	compiler (at least, thats a stated goal, and breakage will be treated
		2387	as a bug).
		2388
		2389	You need the following files in your source tree, or in a directory
		2390	in your include path (e.g. in libev/ when using -Ilibev):
		2391
		2392	ev.h
		2393	ev.c
		2394	ev_vars.h
		2395	ev_wrap.h
		2396
		2397	ev_win32.c required on win32 platforms only
		2398
		2399	ev_select.c only when select backend is enabled (which is enabled by default)
		2400	ev_poll.c only when poll backend is enabled (disabled by default)
		2401	ev_epoll.c only when the epoll backend is enabled (disabled by default)
		2402	ev_kqueue.c only when the kqueue backend is enabled (disabled by default)
		2403	ev_port.c only when the solaris port backend is enabled (disabled by default)
		2404
		2405	F<ev.c> includes the backend files directly when enabled, so you only need
		2406	to compile this single file.
		2407
		2408	=head3 LIBEVENT COMPATIBILITY API
		2409
		2410	To include the libevent compatibility API, also include:
		2411
		2412	#include "event.c"
		2413
		2414	in the file including F<ev.c>, and:
		2415
		2416	#include "event.h"
		2417
		2418	in the files that want to use the libevent API. This also includes F<ev.h>.
		2419
		2420	You need the following additional files for this:
		2421
		2422	event.h
		2423	event.c
		2424
		2425	=head3 AUTOCONF SUPPORT
		2426
		2427	Instead of using C<EV_STANDALONE=1> and providing your config in
		2428	whatever way you want, you can also C<m4_include([libev.m4])> in your
		2429	F<configure.ac> and leave C<EV_STANDALONE> undefined. F<ev.c> will then
		2430	include F<config.h> and configure itself accordingly.
		2431
		2432	For this of course you need the m4 file:
		2433
		2434	libev.m4
		2435
		2436	=head2 PREPROCESSOR SYMBOLS/MACROS
		2437
		2438	Libev can be configured via a variety of preprocessor symbols you have to define
		2439	before including any of its files. The default is not to build for multiplicity
		2440	and only include the select backend.
		2441
		2442	=over 4
		2443
		2444	=item EV_STANDALONE
		2445
		2446	Must always be C<1> if you do not use autoconf configuration, which
		2447	keeps libev from including F<config.h>, and it also defines dummy
		2448	implementations for some libevent functions (such as logging, which is not
		2449	supported). It will also not define any of the structs usually found in
		2450	F<event.h> that are not directly supported by the libev core alone.
		2451
		2452	=item EV_USE_MONOTONIC
		2453
		2454	If defined to be C<1>, libev will try to detect the availability of the
		2455	monotonic clock option at both compiletime and runtime. Otherwise no use
		2456	of the monotonic clock option will be attempted. If you enable this, you
		2457	usually have to link against librt or something similar. Enabling it when
		2458	the functionality isn't available is safe, though, although you have
		2459	to make sure you link against any libraries where the C<clock_gettime>
		2460	function is hiding in (often F<-lrt>).
		2461
		2462	=item EV_USE_REALTIME
		2463
		2464	If defined to be C<1>, libev will try to detect the availability of the
		2465	realtime clock option at compiletime (and assume its availability at
		2466	runtime if successful). Otherwise no use of the realtime clock option will
		2467	be attempted. This effectively replaces C<gettimeofday> by C<clock_get
		2468	(CLOCK_REALTIME, ...)> and will not normally affect correctness. See the
		2469	note about libraries in the description of C<EV_USE_MONOTONIC>, though.
		2470
		2471	=item EV_USE_NANOSLEEP
		2472
		2473	If defined to be C<1>, libev will assume that C<nanosleep ()> is available
		2474	and will use it for delays. Otherwise it will use C<select ()>.
		2475
		2476	=item EV_USE_SELECT
		2477
		2478	If undefined or defined to be C<1>, libev will compile in support for the
		2479	C<select>(2) backend. No attempt at autodetection will be done: if no
		2480	other method takes over, select will be it. Otherwise the select backend
		2481	will not be compiled in.
		2482
		2483	=item EV_SELECT_USE_FD_SET
		2484
		2485	If defined to C<1>, then the select backend will use the system C<fd_set>
		2486	structure. This is useful if libev doesn't compile due to a missing
		2487	C<NFDBITS> or C<fd_mask> definition or it misguesses the bitset layout on
		2488	exotic systems. This usually limits the range of file descriptors to some
		2489	low limit such as 1024 or might have other limitations (winsocket only
		2490	allows 64 sockets). The C<FD_SETSIZE> macro, set before compilation, might
		2491	influence the size of the C<fd_set> used.
		2492
		2493	=item EV_SELECT_IS_WINSOCKET
		2494
		2495	When defined to C<1>, the select backend will assume that
		2496	select/socket/connect etc. don't understand file descriptors but
		2497	wants osf handles on win32 (this is the case when the select to
		2498	be used is the winsock select). This means that it will call
		2499	C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,
		2500	it is assumed that all these functions actually work on fds, even
		2501	on win32. Should not be defined on non-win32 platforms.
		2502
		2503	=item EV_FD_TO_WIN32_HANDLE
		2504
		2505	If C<EV_SELECT_IS_WINSOCKET> is enabled, then libev needs a way to map
		2506	file descriptors to socket handles. When not defining this symbol (the
		2507	default), then libev will call C<_get_osfhandle>, which is usually
		2508	correct. In some cases, programs use their own file descriptor management,
		2509	in which case they can provide this function to map fds to socket handles.
		2510
		2511	=item EV_USE_POLL
		2512
		2513	If defined to be C<1>, libev will compile in support for the C<poll>(2)
		2514	backend. Otherwise it will be enabled on non-win32 platforms. It
		2515	takes precedence over select.
		2516
		2517	=item EV_USE_EPOLL
		2518
		2519	If defined to be C<1>, libev will compile in support for the Linux
		2520	C<epoll>(7) backend. Its availability will be detected at runtime,
		2521	otherwise another method will be used as fallback. This is the
		2522	preferred backend for GNU/Linux systems.
		2523
		2524	=item EV_USE_KQUEUE
		2525
		2526	If defined to be C<1>, libev will compile in support for the BSD style
		2527	C<kqueue>(2) backend. Its actual availability will be detected at runtime,
		2528	otherwise another method will be used as fallback. This is the preferred
		2529	backend for BSD and BSD-like systems, although on most BSDs kqueue only
		2530	supports some types of fds correctly (the only platform we found that
		2531	supports ptys for example was NetBSD), so kqueue might be compiled in, but
		2532	not be used unless explicitly requested. The best way to use it is to find
		2533	out whether kqueue supports your type of fd properly and use an embedded
		2534	kqueue loop.
		2535
		2536	=item EV_USE_PORT
		2537
		2538	If defined to be C<1>, libev will compile in support for the Solaris
		2539	10 port style backend. Its availability will be detected at runtime,
		2540	otherwise another method will be used as fallback. This is the preferred
		2541	backend for Solaris 10 systems.
		2542
		2543	=item EV_USE_DEVPOLL
		2544
		2545	reserved for future expansion, works like the USE symbols above.
		2546
		2547	=item EV_USE_INOTIFY
		2548
		2549	If defined to be C<1>, libev will compile in support for the Linux inotify
		2550	interface to speed up C<ev_stat> watchers. Its actual availability will
		2551	be detected at runtime.
		2552
		2553	=item EV_H
		2554
		2555	The name of the F<ev.h> header file used to include it. The default if
		2556	undefined is C<"ev.h"> in F<event.h> and F<ev.c>. This can be used to
		2557	virtually rename the F<ev.h> header file in case of conflicts.
		2558
		2559	=item EV_CONFIG_H
		2560
		2561	If C<EV_STANDALONE> isn't C<1>, this variable can be used to override
		2562	F<ev.c>'s idea of where to find the F<config.h> file, similarly to
		2563	C<EV_H>, above.
		2564
		2565	=item EV_EVENT_H
		2566
		2567	Similarly to C<EV_H>, this macro can be used to override F<event.c>'s idea
		2568	of how the F<event.h> header can be found, the dfeault is C<"event.h">.
		2569
		2570	=item EV_PROTOTYPES
		2571
		2572	If defined to be C<0>, then F<ev.h> will not define any function
		2573	prototypes, but still define all the structs and other symbols. This is
		2574	occasionally useful if you want to provide your own wrapper functions
		2575	around libev functions.
		2576
		2577	=item EV_MULTIPLICITY
		2578
		2579	If undefined or defined to C<1>, then all event-loop-specific functions
		2580	will have the C<struct ev_loop *> as first argument, and you can create
		2581	additional independent event loops. Otherwise there will be no support
		2582	for multiple event loops and there is no first event loop pointer
		2583	argument. Instead, all functions act on the single default loop.
		2584
		2585	=item EV_MINPRI
		2586
		2587	=item EV_MAXPRI
		2588
		2589	The range of allowed priorities. C<EV_MINPRI> must be smaller or equal to
		2590	C<EV_MAXPRI>, but otherwise there are no non-obvious limitations. You can
		2591	provide for more priorities by overriding those symbols (usually defined
		2592	to be C<-2> and C<2>, respectively).
		2593
		2594	When doing priority-based operations, libev usually has to linearly search
		2595	all the priorities, so having many of them (hundreds) uses a lot of space
		2596	and time, so using the defaults of five priorities (-2 .. +2) is usually
		2597	fine.
		2598
		2599	If your embedding app does not need any priorities, defining these both to
		2600	C<0> will save some memory and cpu.
		2601
		2602	=item EV_PERIODIC_ENABLE
		2603
		2604	If undefined or defined to be C<1>, then periodic timers are supported. If
		2605	defined to be C<0>, then they are not. Disabling them saves a few kB of
		2606	code.
		2607
		2608	=item EV_IDLE_ENABLE
		2609
		2610	If undefined or defined to be C<1>, then idle watchers are supported. If
		2611	defined to be C<0>, then they are not. Disabling them saves a few kB of
		2612	code.
		2613
		2614	=item EV_EMBED_ENABLE
		2615
		2616	If undefined or defined to be C<1>, then embed watchers are supported. If
		2617	defined to be C<0>, then they are not.
		2618
		2619	=item EV_STAT_ENABLE
		2620
		2621	If undefined or defined to be C<1>, then stat watchers are supported. If
		2622	defined to be C<0>, then they are not.
		2623
		2624	=item EV_FORK_ENABLE
		2625
		2626	If undefined or defined to be C<1>, then fork watchers are supported. If
		2627	defined to be C<0>, then they are not.
		2628
		2629	=item EV_MINIMAL
		2630
		2631	If you need to shave off some kilobytes of code at the expense of some
		2632	speed, define this symbol to C<1>. Currently only used for gcc to override
		2633	some inlining decisions, saves roughly 30% codesize of amd64.
		2634
		2635	=item EV_PID_HASHSIZE
		2636
		2637	C<ev_child> watchers use a small hash table to distribute workload by
		2638	pid. The default size is C<16> (or C<1> with C<EV_MINIMAL>), usually more
		2639	than enough. If you need to manage thousands of children you might want to
		2640	increase this value (I<must> be a power of two).
		2641
		2642	=item EV_INOTIFY_HASHSIZE
		2643
		2644	C<ev_stat> watchers use a small hash table to distribute workload by
		2645	inotify watch id. The default size is C<16> (or C<1> with C<EV_MINIMAL>),
		2646	usually more than enough. If you need to manage thousands of C<ev_stat>
		2647	watchers you might want to increase this value (I<must> be a power of
		2648	two).
		2649
		2650	=item EV_COMMON
		2651
		2652	By default, all watchers have a C<void *data> member. By redefining
		2653	this macro to a something else you can include more and other types of
		2654	members. You have to define it each time you include one of the files,
		2655	though, and it must be identical each time.
		2656
		2657	For example, the perl EV module uses something like this:
		2658
		2659	#define EV_COMMON \
		2660	SV *self; /* contains this struct */ \
		2661	SV *cb_sv, *fh /* note no trailing ";" */
		2662
		2663	=item EV_CB_DECLARE (type)
		2664
		2665	=item EV_CB_INVOKE (watcher, revents)
		2666
		2667	=item ev_set_cb (ev, cb)
		2668
		2669	Can be used to change the callback member declaration in each watcher,
		2670	and the way callbacks are invoked and set. Must expand to a struct member
		2671	definition and a statement, respectively. See the F<ev.h> header file for
		2672	their default definitions. One possible use for overriding these is to
		2673	avoid the C<struct ev_loop *> as first argument in all cases, or to use
		2674	method calls instead of plain function calls in C++.
		2675
		2676	=head2 EXPORTED API SYMBOLS
		2677
		2678	If you need to re-export the API (e.g. via a dll) and you need a list of
		2679	exported symbols, you can use the provided F<Symbol.*> files which list
		2680	all public symbols, one per line:
		2681
		2682	Symbols.ev for libev proper
		2683	Symbols.event for the libevent emulation
		2684
		2685	This can also be used to rename all public symbols to avoid clashes with
		2686	multiple versions of libev linked together (which is obviously bad in
		2687	itself, but sometimes it is inconvinient to avoid this).
		2688
		2689	A sed command like this will create wrapper C<#define>'s that you need to
		2690	include before including F<ev.h>:
		2691
		2692	<Symbols.ev sed -e "s/.*/#define & myprefix_&/" >wrap.h
		2693
		2694	This would create a file F<wrap.h> which essentially looks like this:
		2695
		2696	#define ev_backend myprefix_ev_backend
		2697	#define ev_check_start myprefix_ev_check_start
		2698	#define ev_check_stop myprefix_ev_check_stop
		2699	...
		2700
		2701	=head2 EXAMPLES
		2702
		2703	For a real-world example of a program the includes libev
		2704	verbatim, you can have a look at the EV perl module
		2705	(L<http://software.schmorp.de/pkg/EV.html>). It has the libev files in
		2706	the F<libev/> subdirectory and includes them in the F<EV/EVAPI.h> (public
		2707	interface) and F<EV.xs> (implementation) files. Only the F<EV.xs> file
		2708	will be compiled. It is pretty complex because it provides its own header
		2709	file.
		2710
		2711	The usage in rxvt-unicode is simpler. It has a F<ev_cpp.h> header file
		2712	that everybody includes and which overrides some configure choices:
		2713
		2714	#define EV_MINIMAL 1
		2715	#define EV_USE_POLL 0
		2716	#define EV_MULTIPLICITY 0
		2717	#define EV_PERIODIC_ENABLE 0
		2718	#define EV_STAT_ENABLE 0
		2719	#define EV_FORK_ENABLE 0
		2720	#define EV_CONFIG_H <config.h>
		2721	#define EV_MINPRI 0
		2722	#define EV_MAXPRI 0
		2723
		2724	#include "ev++.h"
		2725
		2726	And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:
		2727
		2728	#include "ev_cpp.h"
		2729	#include "ev.c"
		2730
		2731
		2732	=head1 COMPLEXITIES
		2733
		2734	In this section the complexities of (many of) the algorithms used inside
		2735	libev will be explained. For complexity discussions about backends see the
		2736	documentation for C<ev_default_init>.
		2737
		2738	All of the following are about amortised time: If an array needs to be
		2739	extended, libev needs to realloc and move the whole array, but this
		2740	happens asymptotically never with higher number of elements, so O(1) might
		2741	mean it might do a lengthy realloc operation in rare cases, but on average
		2742	it is much faster and asymptotically approaches constant time.
		2743
		2744	=over 4
		2745
		2746	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)
		2747
		2748	This means that, when you have a watcher that triggers in one hour and
		2749	there are 100 watchers that would trigger before that then inserting will
		2750	have to skip roughly seven (C<ld 100>) of these watchers.
		2751
		2752	=item Changing timer/periodic watchers (by autorepeat or calling again): O(log skipped_other_timers)
		2753
		2754	That means that changing a timer costs less than removing/adding them
		2755	as only the relative motion in the event queue has to be paid for.
		2756
		2757	=item Starting io/check/prepare/idle/signal/child watchers: O(1)
		2758
		2759	These just add the watcher into an array or at the head of a list.
		2760
		2761	=item Stopping check/prepare/idle watchers: O(1)
		2762
		2763	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))
		2764
		2765	These watchers are stored in lists then need to be walked to find the
		2766	correct watcher to remove. The lists are usually short (you don't usually
		2767	have many watchers waiting for the same fd or signal).
		2768
		2769	=item Finding the next timer in each loop iteration: O(1)
		2770
		2771	By virtue of using a binary heap, the next timer is always found at the
		2772	beginning of the storage array.
		2773
		2774	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)
		2775
		2776	A change means an I/O watcher gets started or stopped, which requires
		2777	libev to recalculate its status (and possibly tell the kernel, depending
		2778	on backend and wether C<ev_io_set> was used).
		2779
		2780	=item Activating one watcher (putting it into the pending state): O(1)
		2781
		2782	=item Priority handling: O(number_of_priorities)
		2783
		2784	Priorities are implemented by allocating some space for each
		2785	priority. When doing priority-based operations, libev usually has to
		2786	linearly search all the priorities, but starting/stopping and activating
		2787	watchers becomes O(1) w.r.t. prioritiy handling.
		2788
		2789	=back
		2790
		2791
		2792	=head1 Win32 platform limitations and workarounds
		2793
		2794	Win32 doesn't support any of the standards (e.g. POSIX) that libev
		2795	requires, and its I/O model is fundamentally incompatible with the POSIX
		2796	model. Libev still offers limited functionality on this platform in
		2797	the form of the C<EVBACKEND_SELECT> backend, and only supports socket
		2798	descriptors. This only applies when using Win32 natively, not when using
		2799	e.g. cygwin.
		2800
		2801	There is no supported compilation method available on windows except
		2802	embedding it into other applications.
		2803
		2804	Due to the many, low, and arbitrary limits on the win32 platform and the
		2805	abysmal performance of winsockets, using a large number of sockets is not
		2806	recommended (and not reasonable). If your program needs to use more than
		2807	a hundred or so sockets, then likely it needs to use a totally different
		2808	implementation for windows, as libev offers the POSIX model, which cannot
		2809	be implemented efficiently on windows (microsoft monopoly games).
		2810
		2811	=over 4
		2812
		2813	=item The winsocket select function
		2814
		2815	The winsocket C<select> function doesn't follow POSIX in that it requires
		2816	socket I<handles> and not socket I<file descriptors>. This makes select
		2817	very inefficient, and also requires a mapping from file descriptors
		2818	to socket handles. See the discussion of the C<EV_SELECT_USE_FD_SET>,
		2819	C<EV_SELECT_IS_WINSOCKET> and C<EV_FD_TO_WIN32_HANDLE> preprocessor
		2820	symbols for more info.
		2821
		2822	The configuration for a "naked" win32 using the microsoft runtime
		2823	libraries and raw winsocket select is:
		2824
		2825	#define EV_USE_SELECT 1
		2826	#define EV_SELECT_IS_WINSOCKET 1 /* forces EV_SELECT_USE_FD_SET, too */
		2827
		2828	Note that winsockets handling of fd sets is O(n), so you can easily get a
		2829	complexity in the O(n²) range when using win32.
		2830
		2831	=item Limited number of file descriptors
		2832
		2833	Windows has numerous arbitrary (and low) limits on things. Early versions
		2834	of winsocket's select only supported waiting for a max. of C<64> handles
		2835	(probably owning to the fact that all windows kernels can only wait for
		2836	C<64> things at the same time internally; microsoft recommends spawning a
		2837	chain of threads and wait for 63 handles and the previous thread in each).
		2838
		2839	Newer versions support more handles, but you need to define C<FD_SETSIZE>
		2840	to some high number (e.g. C<2048>) before compiling the winsocket select
		2841	call (which might be in libev or elsewhere, for example, perl does its own
		2842	select emulation on windows).
		2843
		2844	Another limit is the number of file descriptors in the microsoft runtime
		2845	libraries, which by default is C<64> (there must be a hidden I<64> fetish
		2846	or something like this inside microsoft). You can increase this by calling
		2847	C<_setmaxstdio>, which can increase this limit to C<2048> (another
		2848	arbitrary limit), but is broken in many versions of the microsoft runtime
		2849	libraries.
		2850
		2851	This might get you to about C<512> or C<2048> sockets (depending on
		2852	windows version and/or the phase of the moon). To get more, you need to
		2853	wrap all I/O functions and provide your own fd management, but the cost of
		2854	calling select (O(n²)) will likely make this unworkable.
		2855
		2856	=back
		2857
878		2858
879	=head1 AUTHOR	2859	=head1 AUTHOR
880		2860
881	Marc Lehmann <libev@schmorp.de>.	2861	Marc Lehmann <libev@schmorp.de>.
882		2862

Diff Legend

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