ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/libev/ev.pod

Comparing libev/ev.pod (file contents):
Revision 1.38 by root, Sat Nov 24 09:48:38 2007 UTC vs.
Revision 1.205 by root, Mon Oct 27 12:20:32 2008 UTC

…		…
2		2
3	libev - a high performance full-featured event loop written in C	3	libev - a high performance full-featured event loop written in C
4		4
5	=head1 SYNOPSIS	5	=head1 SYNOPSIS
6		6
7	#include <ev.h>	7	#include <ev.h>
		8
		9	=head2 EXAMPLE PROGRAM
		10
		11	// a single header file is required
		12	#include <ev.h>
		13
		14	// every watcher type has its own typedef'd struct
		15	// with the name ev_TYPE
		16	ev_io stdin_watcher;
		17	ev_timer timeout_watcher;
		18
		19	// all watcher callbacks have a similar signature
		20	// this callback is called when data is readable on stdin
		21	static void
		22	stdin_cb (EV_P_ ev_io *w, int revents)
		23	{
		24	puts ("stdin ready");
		25	// for one-shot events, one must manually stop the watcher
		26	// with its corresponding stop function.
		27	ev_io_stop (EV_A_ w);
		28
		29	// this causes all nested ev_loop's to stop iterating
		30	ev_unloop (EV_A_ EVUNLOOP_ALL);
		31	}
		32
		33	// another callback, this time for a time-out
		34	static void
		35	timeout_cb (EV_P_ ev_timer *w, int revents)
		36	{
		37	puts ("timeout");
		38	// this causes the innermost ev_loop to stop iterating
		39	ev_unloop (EV_A_ EVUNLOOP_ONE);
		40	}
		41
		42	int
		43	main (void)
		44	{
		45	// use the default event loop unless you have special needs
		46	ev_loop *loop = ev_default_loop (0);
		47
		48	// initialise an io watcher, then start it
		49	// this one will watch for stdin to become readable
		50	ev_io_init (&stdin_watcher, stdin_cb, /*STDIN_FILENO*/ 0, EV_READ);
		51	ev_io_start (loop, &stdin_watcher);
		52
		53	// initialise a timer watcher, then start it
		54	// simple non-repeating 5.5 second timeout
		55	ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);
		56	ev_timer_start (loop, &timeout_watcher);
		57
		58	// now wait for events to arrive
		59	ev_loop (loop, 0);
		60
		61	// unloop was called, so exit
		62	return 0;
		63	}
8		64
9	=head1 DESCRIPTION	65	=head1 DESCRIPTION
10		66
		67	The newest version of this document is also available as an html-formatted
		68	web page you might find easier to navigate when reading it for the first
		69	time: L<http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod>.
		70
11	Libev is an event loop: you register interest in certain events (such as a	71	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	72	file descriptor being readable or a timeout occurring), and it will manage
13	these event sources and provide your program with events.	73	these event sources and provide your program with events.
14		74
15	To do this, it must take more or less complete control over your process	75	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	76	(or thread) by executing the I<event loop> handler, and will then
17	communicate events via a callback mechanism.	77	communicate events via a callback mechanism.
…		…
19	You register interest in certain events by registering so-called I<event	79	You register interest in certain events by registering so-called I<event
20	watchers>, which are relatively small C structures you initialise with the	80	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	81	details of the event, and then hand it over to libev by I<starting> the
22	watcher.	82	watcher.
23		83
24	=head1 FEATURES	84	=head2 FEATURES
25		85
26	Libev supports select, poll, the linux-specific epoll and the bsd-specific	86	Libev supports C<select>, C<poll>, the Linux-specific C<epoll>, the
27	kqueue mechanisms for file descriptor events, relative timers, absolute	87	BSD-specific C<kqueue> and the Solaris-specific event port mechanisms
28	timers with customised rescheduling, signal events, process status change	88	for file descriptor events (C<ev_io>), the Linux C<inotify> interface
29	events (related to SIGCHLD), and event watchers dealing with the event	89	(for C<ev_stat>), relative timers (C<ev_timer>), absolute timers
30	loop mechanism itself (idle, prepare and check watchers). It also is quite	90	with customised rescheduling (C<ev_periodic>), synchronous signals
		91	(C<ev_signal>), process status change events (C<ev_child>), and event
		92	watchers dealing with the event loop mechanism itself (C<ev_idle>,
		93	C<ev_embed>, C<ev_prepare> and C<ev_check> watchers) as well as
		94	file watchers (C<ev_stat>) and even limited support for fork events
		95	(C<ev_fork>).
		96
		97	It also is quite fast (see this
31	fast (see this L<benchmark|http://libev.schmorp.de/bench.html> comparing	98	L<benchmark|http://libev.schmorp.de/bench.html> comparing it to libevent
32	it to libevent for example).	99	for example).
33		100
34	=head1 CONVENTIONS	101	=head2 CONVENTIONS
35		102
36	Libev is very configurable. In this manual the default configuration	103	Libev is very configurable. In this manual the default (and most common)
37	will be described, which supports multiple event loops. For more info	104	configuration will be described, which supports multiple event loops. For
38	about various configuration options please have a look at the file	105	more info about various configuration options please have a look at
39	F<README.embed> in the libev distribution. If libev was configured without	106	B<EMBED> section in this manual. If libev was configured without support
40	support for multiple event loops, then all functions taking an initial	107	for multiple event loops, then all functions taking an initial argument of
41	argument of name C<loop> (which is always of type C<struct ev_loop *>)	108	name C<loop> (which is always of type C<ev_loop *>) will not have
42	will not have this argument.	109	this argument.
43		110
44	=head1 TIME REPRESENTATION	111	=head2 TIME REPRESENTATION
45		112
46	Libev represents time as a single floating point number, representing the	113	Libev represents time as a single floating point number, representing the
47	(fractional) number of seconds since the (POSIX) epoch (somewhere near	114	(fractional) number of seconds since the (POSIX) epoch (somewhere near
48	the beginning of 1970, details are complicated, don't ask). This type is	115	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	116	called C<ev_tstamp>, which is what you should use too. It usually aliases
50	to the C<double> type in C, and when you need to do any calculations on	117	to the C<double> type in C, and when you need to do any calculations on
51	it, you should treat it as such.	118	it, you should treat it as some floating point value. Unlike the name
		119	component C<stamp> might indicate, it is also used for time differences
		120	throughout libev.
		121
		122	=head1 ERROR HANDLING
		123
		124	Libev knows three classes of errors: operating system errors, usage errors
		125	and internal errors (bugs).
		126
		127	When libev catches an operating system error it cannot handle (for example
		128	a system call indicating a condition libev cannot fix), it calls the callback
		129	set via C<ev_set_syserr_cb>, which is supposed to fix the problem or
		130	abort. The default is to print a diagnostic message and to call C<abort
		131	()>.
		132
		133	When libev detects a usage error such as a negative timer interval, then
		134	it will print a diagnostic message and abort (via the C<assert> mechanism,
		135	so C<NDEBUG> will disable this checking): these are programming errors in
		136	the libev caller and need to be fixed there.
		137
		138	Libev also has a few internal error-checking C<assert>ions, and also has
		139	extensive consistency checking code. These do not trigger under normal
		140	circumstances, as they indicate either a bug in libev or worse.
52		141
53		142
54	=head1 GLOBAL FUNCTIONS	143	=head1 GLOBAL FUNCTIONS
55		144
56	These functions can be called anytime, even before initialising the	145	These functions can be called anytime, even before initialising the
…		…
62		151
63	Returns the current time as libev would use it. Please note that the	152	Returns the current time as libev would use it. Please note that the
64	C<ev_now> function is usually faster and also often returns the timestamp	153	C<ev_now> function is usually faster and also often returns the timestamp
65	you actually want to know.	154	you actually want to know.
66		155
		156	=item ev_sleep (ev_tstamp interval)
		157
		158	Sleep for the given interval: The current thread will be blocked until
		159	either it is interrupted or the given time interval has passed. Basically
		160	this is a sub-second-resolution C<sleep ()>.
		161
67	=item int ev_version_major ()	162	=item int ev_version_major ()
68		163
69	=item int ev_version_minor ()	164	=item int ev_version_minor ()
70		165
71	You can find out the major and minor version numbers of the library	166	You can find out the major and minor ABI version numbers of the library
72	you linked against by calling the functions C<ev_version_major> and	167	you linked against by calling the functions C<ev_version_major> and
73	C<ev_version_minor>. If you want, you can compare against the global	168	C<ev_version_minor>. If you want, you can compare against the global
74	symbols C<EV_VERSION_MAJOR> and C<EV_VERSION_MINOR>, which specify the	169	symbols C<EV_VERSION_MAJOR> and C<EV_VERSION_MINOR>, which specify the
75	version of the library your program was compiled against.	170	version of the library your program was compiled against.
76		171
		172	These version numbers refer to the ABI version of the library, not the
		173	release version.
		174
77	Usually, it's a good idea to terminate if the major versions mismatch,	175	Usually, it's a good idea to terminate if the major versions mismatch,
78	as this indicates an incompatible change. Minor versions are usually	176	as this indicates an incompatible change. Minor versions are usually
79	compatible to older versions, so a larger minor version alone is usually	177	compatible to older versions, so a larger minor version alone is usually
80	not a problem.	178	not a problem.
81		179
82	Example: make sure we haven't accidentally been linked against the wrong	180	Example: Make sure we haven't accidentally been linked against the wrong
83	version:	181	version.
84		182
85	assert (("libev version mismatch",	183	assert (("libev version mismatch",
86	ev_version_major () == EV_VERSION_MAJOR	184	ev_version_major () == EV_VERSION_MAJOR
87	&& ev_version_minor () >= EV_VERSION_MINOR));	185	&& ev_version_minor () >= EV_VERSION_MINOR));
88		186
89	=item unsigned int ev_supported_backends ()	187	=item unsigned int ev_supported_backends ()
90		188
91	Return the set of all backends (i.e. their corresponding C<EV_BACKEND_*>	189	Return the set of all backends (i.e. their corresponding C<EV_BACKEND_*>
92	value) compiled into this binary of libev (independent of their	190	value) compiled into this binary of libev (independent of their
…		…
94	a description of the set values.	192	a description of the set values.
95		193
96	Example: make sure we have the epoll method, because yeah this is cool and	194	Example: make sure we have the epoll method, because yeah this is cool and
97	a must have and can we have a torrent of it please!!!11	195	a must have and can we have a torrent of it please!!!11
98		196
99	assert (("sorry, no epoll, no sex",	197	assert (("sorry, no epoll, no sex",
100	ev_supported_backends () & EVBACKEND_EPOLL));	198	ev_supported_backends () & EVBACKEND_EPOLL));
101		199
102	=item unsigned int ev_recommended_backends ()	200	=item unsigned int ev_recommended_backends ()
103		201
104	Return the set of all backends compiled into this binary of libev and also	202	Return the set of all backends compiled into this binary of libev and also
105	recommended for this platform. This set is often smaller than the one	203	recommended for this platform. This set is often smaller than the one
106	returned by C<ev_supported_backends>, as for example kqueue is broken on	204	returned by C<ev_supported_backends>, as for example kqueue is broken on
107	most BSDs and will not be autodetected unless you explicitly request it	205	most BSDs and will not be auto-detected unless you explicitly request it
108	(assuming you know what you are doing). This is the set of backends that	206	(assuming you know what you are doing). This is the set of backends that
109	libev will probe for if you specify no backends explicitly.	207	libev will probe for if you specify no backends explicitly.
110		208
111	=item unsigned int ev_embeddable_backends ()	209	=item unsigned int ev_embeddable_backends ()
112		210
…		…
116	C<ev_embeddable_backends () & ev_supported_backends ()>, likewise for	214	C<ev_embeddable_backends () & ev_supported_backends ()>, likewise for
117	recommended ones.	215	recommended ones.
118		216
119	See the description of C<ev_embed> watchers for more info.	217	See the description of C<ev_embed> watchers for more info.
120		218
121	=item ev_set_allocator (void *(*cb)(void *ptr, long size))	219	=item ev_set_allocator (void *(*cb)(void *ptr, long size)) [NOT REENTRANT]
122		220
123	Sets the allocation function to use (the prototype is similar to the	221	Sets the allocation function to use (the prototype is similar - the
124	realloc C function, the semantics are identical). It is used to allocate	222	semantics are identical to the C<realloc> C89/SuS/POSIX function). It is
125	and free memory (no surprises here). If it returns zero when memory	223	used to allocate and free memory (no surprises here). If it returns zero
126	needs to be allocated, the library might abort or take some potentially	224	when memory needs to be allocated (C<size != 0>), the library might abort
127	destructive action. The default is your system realloc function.	225	or take some potentially destructive action.
		226
		227	Since some systems (at least OpenBSD and Darwin) fail to implement
		228	correct C<realloc> semantics, libev will use a wrapper around the system
		229	C<realloc> and C<free> functions by default.
128		230
129	You could override this function in high-availability programs to, say,	231	You could override this function in high-availability programs to, say,
130	free some memory if it cannot allocate memory, to use a special allocator,	232	free some memory if it cannot allocate memory, to use a special allocator,
131	or even to sleep a while and retry until some memory is available.	233	or even to sleep a while and retry until some memory is available.
132		234
133	Example: replace the libev allocator with one that waits a bit and then	235	Example: Replace the libev allocator with one that waits a bit and then
134	retries: better than mine).	236	retries (example requires a standards-compliant C<realloc>).
135		237
136	static void *	238	static void *
137	persistent_realloc (void *ptr, long size)	239	persistent_realloc (void *ptr, size_t size)
138	{	240	{
139	for (;;)	241	for (;;)
140	{	242	{
141	void *newptr = realloc (ptr, size);	243	void *newptr = realloc (ptr, size);
142		244
…		…
148	}	250	}
149		251
150	...	252	...
151	ev_set_allocator (persistent_realloc);	253	ev_set_allocator (persistent_realloc);
152		254
153	=item ev_set_syserr_cb (void (*cb)(const char *msg));	255	=item ev_set_syserr_cb (void (*cb)(const char *msg)); [NOT REENTRANT]
154		256
155	Set the callback function to call on a retryable syscall error (such	257	Set the callback function to call on a retryable system call error (such
156	as failed select, poll, epoll_wait). The message is a printable string	258	as failed select, poll, epoll_wait). The message is a printable string
157	indicating the system call or subsystem causing the problem. If this	259	indicating the system call or subsystem causing the problem. If this
158	callback is set, then libev will expect it to remedy the sitution, no	260	callback is set, then libev will expect it to remedy the situation, no
159	matter what, when it returns. That is, libev will generally retry the	261	matter what, when it returns. That is, libev will generally retry the
160	requested operation, or, if the condition doesn't go away, do bad stuff	262	requested operation, or, if the condition doesn't go away, do bad stuff
161	(such as abort).	263	(such as abort).
162		264
163	Example: do the same thing as libev does internally:	265	Example: This is basically the same thing that libev does internally, too.
164		266
165	static void	267	static void
166	fatal_error (const char *msg)	268	fatal_error (const char *msg)
167	{	269	{
168	perror (msg);	270	perror (msg);
…		…
174		276
175	=back	277	=back
176		278
177	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP	279	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP
178		280
179	An event loop is described by a C<struct ev_loop *>. The library knows two	281	An event loop is described by a C<struct ev_loop *> (the C<struct>
180	types of such loops, the I<default> loop, which supports signals and child	282	is I<not> optional in this case, as there is also an C<ev_loop>
181	events, and dynamically created loops which do not.	283	I<function>).
182		284
183	If you use threads, a common model is to run the default event loop	285	The library knows two types of such loops, the I<default> loop, which
184	in your main thread (or in a separate thread) and for each thread you	286	supports signals and child events, and dynamically created loops which do
185	create, you also create another event loop. Libev itself does no locking	287	not.
186	whatsoever, so if you mix calls to the same event loop in different
187	threads, make sure you lock (this is usually a bad idea, though, even if
188	done correctly, because it's hideous and inefficient).
189		288
190	=over 4	289	=over 4
191		290
192	=item struct ev_loop *ev_default_loop (unsigned int flags)	291	=item struct ev_loop *ev_default_loop (unsigned int flags)
193		292
…		…
197	flags. If that is troubling you, check C<ev_backend ()> afterwards).	296	flags. If that is troubling you, check C<ev_backend ()> afterwards).
198		297
199	If you don't know what event loop to use, use the one returned from this	298	If you don't know what event loop to use, use the one returned from this
200	function.	299	function.
201		300
		301	Note that this function is I<not> thread-safe, so if you want to use it
		302	from multiple threads, you have to lock (note also that this is unlikely,
		303	as loops cannot bes hared easily between threads anyway).
		304
		305	The default loop is the only loop that can handle C<ev_signal> and
		306	C<ev_child> watchers, and to do this, it always registers a handler
		307	for C<SIGCHLD>. If this is a problem for your application you can either
		308	create a dynamic loop with C<ev_loop_new> that doesn't do that, or you
		309	can simply overwrite the C<SIGCHLD> signal handler I<after> calling
		310	C<ev_default_init>.
		311
202	The flags argument can be used to specify special behaviour or specific	312	The flags argument can be used to specify special behaviour or specific
203	backends to use, and is usually specified as C<0> (or C<EVFLAG_AUTO>).	313	backends to use, and is usually specified as C<0> (or C<EVFLAG_AUTO>).
204		314
205	The following flags are supported:	315	The following flags are supported:
206		316
…		…
211	The default flags value. Use this if you have no clue (it's the right	321	The default flags value. Use this if you have no clue (it's the right
212	thing, believe me).	322	thing, believe me).
213		323
214	=item C<EVFLAG_NOENV>	324	=item C<EVFLAG_NOENV>
215		325
216	If this flag bit is ored into the flag value (or the program runs setuid	326	If this flag bit is or'ed into the flag value (or the program runs setuid
217	or setgid) then libev will I<not> look at the environment variable	327	or setgid) then libev will I<not> look at the environment variable
218	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will	328	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will
219	override the flags completely if it is found in the environment. This is	329	override the flags completely if it is found in the environment. This is
220	useful to try out specific backends to test their performance, or to work	330	useful to try out specific backends to test their performance, or to work
221	around bugs.	331	around bugs.
222		332
		333	=item C<EVFLAG_FORKCHECK>
		334
		335	Instead of calling C<ev_default_fork> or C<ev_loop_fork> manually after
		336	a fork, you can also make libev check for a fork in each iteration by
		337	enabling this flag.
		338
		339	This works by calling C<getpid ()> on every iteration of the loop,
		340	and thus this might slow down your event loop if you do a lot of loop
		341	iterations and little real work, but is usually not noticeable (on my
		342	GNU/Linux system for example, C<getpid> is actually a simple 5-insn sequence
		343	without a system call and thus I<very> fast, but my GNU/Linux system also has
		344	C<pthread_atfork> which is even faster).
		345
		346	The big advantage of this flag is that you can forget about fork (and
		347	forget about forgetting to tell libev about forking) when you use this
		348	flag.
		349
		350	This flag setting cannot be overridden or specified in the C<LIBEV_FLAGS>
		351	environment variable.
		352
223	=item C<EVBACKEND_SELECT> (value 1, portable select backend)	353	=item C<EVBACKEND_SELECT> (value 1, portable select backend)
224		354
225	This is your standard select(2) backend. Not I<completely> standard, as	355	This is your standard select(2) backend. Not I<completely> standard, as
226	libev tries to roll its own fd_set with no limits on the number of fds,	356	libev tries to roll its own fd_set with no limits on the number of fds,
227	but if that fails, expect a fairly low limit on the number of fds when	357	but if that fails, expect a fairly low limit on the number of fds when
228	using this backend. It doesn't scale too well (O(highest_fd)), but its usually	358	using this backend. It doesn't scale too well (O(highest_fd)), but its
229	the fastest backend for a low number of fds.	359	usually the fastest backend for a low number of (low-numbered :) fds.
		360
		361	To get good performance out of this backend you need a high amount of
		362	parallelism (most of the file descriptors should be busy). If you are
		363	writing a server, you should C<accept ()> in a loop to accept as many
		364	connections as possible during one iteration. You might also want to have
		365	a look at C<ev_set_io_collect_interval ()> to increase the amount of
		366	readiness notifications you get per iteration.
		367
		368	This backend maps C<EV_READ> to the C<readfds> set and C<EV_WRITE> to the
		369	C<writefds> set (and to work around Microsoft Windows bugs, also onto the
		370	C<exceptfds> set on that platform).
230		371
231	=item C<EVBACKEND_POLL> (value 2, poll backend, available everywhere except on windows)	372	=item C<EVBACKEND_POLL> (value 2, poll backend, available everywhere except on windows)
232		373
233	And this is your standard poll(2) backend. It's more complicated than	374	And this is your standard poll(2) backend. It's more complicated
234	select, but handles sparse fds better and has no artificial limit on the	375	than select, but handles sparse fds better and has no artificial
235	number of fds you can use (except it will slow down considerably with a	376	limit on the number of fds you can use (except it will slow down
236	lot of inactive fds). It scales similarly to select, i.e. O(total_fds).	377	considerably with a lot of inactive fds). It scales similarly to select,
		378	i.e. O(total_fds). See the entry for C<EVBACKEND_SELECT>, above, for
		379	performance tips.
		380
		381	This backend maps C<EV_READ> to C<POLLIN | POLLERR | POLLHUP>, and
		382	C<EV_WRITE> to C<POLLOUT | POLLERR | POLLHUP>.
237		383
238	=item C<EVBACKEND_EPOLL> (value 4, Linux)	384	=item C<EVBACKEND_EPOLL> (value 4, Linux)
239		385
240	For few fds, this backend is a bit little slower than poll and select,	386	For few fds, this backend is a bit little slower than poll and select,
241	but it scales phenomenally better. While poll and select usually scale like	387	but it scales phenomenally better. While poll and select usually scale
242	O(total_fds) where n is the total number of fds (or the highest fd), epoll scales	388	like O(total_fds) where n is the total number of fds (or the highest fd),
243	either O(1) or O(active_fds).	389	epoll scales either O(1) or O(active_fds).
244		390
		391	The epoll syscalls are the most misdesigned of the more advanced
		392	event mechanisms: probelsm include silently dropping events in some
		393	hard-to-detect cases, requiring a system call per fd change, no fork
		394	support, problems with dup and so on.
		395
		396	Epoll is also notoriously buggy - embedding epoll fds should work, but
		397	of course doesn't, and epoll just loves to report events for totally
		398	I<different> file descriptors (even already closed ones, so one cannot
		399	even remove them from the set) than registered in the set (especially
		400	on SMP systems). Libev tries to counter these spurious notifications by
		401	employing an additional generation counter and comparing that against the
		402	events to filter out spurious ones.
		403
245	While stopping and starting an I/O watcher in the same iteration will	404	While stopping, setting and starting an I/O watcher in the same iteration
246	result in some caching, there is still a syscall per such incident	405	will result in some caching, there is still a system call per such incident
247	(because the fd could point to a different file description now), so its	406	(because the fd could point to a different file description now), so its
248	best to avoid that. Also, dup()ed file descriptors might not work very	407	best to avoid that. Also, C<dup ()>'ed file descriptors might not work
249	well if you register events for both fds.	408	very well if you register events for both fds.
250		409
251	Please note that epoll sometimes generates spurious notifications, so you	410	Best performance from this backend is achieved by not unregistering all
252	need to use non-blocking I/O or other means to avoid blocking when no data	411	watchers for a file descriptor until it has been closed, if possible,
253	(or space) is available.	412	i.e. keep at least one watcher active per fd at all times. Stopping and
		413	starting a watcher (without re-setting it) also usually doesn't cause
		414	extra overhead.
		415
		416	While nominally embeddable in other event loops, this feature is broken in
		417	all kernel versions tested so far.
		418
		419	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
		420	C<EVBACKEND_POLL>.
254		421
255	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)	422	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)
256		423
257	Kqueue deserves special mention, as at the time of this writing, it	424	Kqueue deserves special mention, as at the time of this writing, it was
258	was broken on all BSDs except NetBSD (usually it doesn't work with	425	broken on all BSDs except NetBSD (usually it doesn't work reliably with
259	anything but sockets and pipes, except on Darwin, where of course its	426	anything but sockets and pipes, except on Darwin, where of course it's
260	completely useless). For this reason its not being "autodetected"	427	completely useless). For this reason it's not being "auto-detected" unless
261	unless you explicitly specify it explicitly in the flags (i.e. using	428	you explicitly specify it in the flags (i.e. using C<EVBACKEND_KQUEUE>) or
262	C<EVBACKEND_KQUEUE>).	429	libev was compiled on a known-to-be-good (-enough) system like NetBSD.
		430
		431	You still can embed kqueue into a normal poll or select backend and use it
		432	only for sockets (after having made sure that sockets work with kqueue on
		433	the target platform). See C<ev_embed> watchers for more info.
263		434
264	It scales in the same way as the epoll backend, but the interface to the	435	It scales in the same way as the epoll backend, but the interface to the
265	kernel is more efficient (which says nothing about its actual speed, of	436	kernel is more efficient (which says nothing about its actual speed, of
266	course). While starting and stopping an I/O watcher does not cause an	437	course). While stopping, setting and starting an I/O watcher does never
267	extra syscall as with epoll, it still adds up to four event changes per	438	cause an extra system call as with C<EVBACKEND_EPOLL>, it still adds up to
268	incident, so its best to avoid that.	439	two event changes per incident. Support for C<fork ()> is very bad and it
		440	drops fds silently in similarly hard-to-detect cases.
		441
		442	This backend usually performs well under most conditions.
		443
		444	While nominally embeddable in other event loops, this doesn't work
		445	everywhere, so you might need to test for this. And since it is broken
		446	almost everywhere, you should only use it when you have a lot of sockets
		447	(for which it usually works), by embedding it into another event loop
		448	(e.g. C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>) and, did I mention it,
		449	using it only for sockets.
		450
		451	This backend maps C<EV_READ> into an C<EVFILT_READ> kevent with
		452	C<NOTE_EOF>, and C<EV_WRITE> into an C<EVFILT_WRITE> kevent with
		453	C<NOTE_EOF>.
269		454
270	=item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8)	455	=item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8)
271		456
272	This is not implemented yet (and might never be).	457	This is not implemented yet (and might never be, unless you send me an
		458	implementation). According to reports, C</dev/poll> only supports sockets
		459	and is not embeddable, which would limit the usefulness of this backend
		460	immensely.
273		461
274	=item C<EVBACKEND_PORT> (value 32, Solaris 10)	462	=item C<EVBACKEND_PORT> (value 32, Solaris 10)
275		463
276	This uses the Solaris 10 port mechanism. As with everything on Solaris,	464	This uses the Solaris 10 event port mechanism. As with everything on Solaris,
277	it's really slow, but it still scales very well (O(active_fds)).	465	it's really slow, but it still scales very well (O(active_fds)).
278		466
279	Please note that solaris ports can result in a lot of spurious	467	Please note that Solaris event ports can deliver a lot of spurious
280	notifications, so you need to use non-blocking I/O or other means to avoid	468	notifications, so you need to use non-blocking I/O or other means to avoid
281	blocking when no data (or space) is available.	469	blocking when no data (or space) is available.
		470
		471	While this backend scales well, it requires one system call per active
		472	file descriptor per loop iteration. For small and medium numbers of file
		473	descriptors a "slow" C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> backend
		474	might perform better.
		475
		476	On the positive side, with the exception of the spurious readiness
		477	notifications, this backend actually performed fully to specification
		478	in all tests and is fully embeddable, which is a rare feat among the
		479	OS-specific backends.
		480
		481	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
		482	C<EVBACKEND_POLL>.
282		483
283	=item C<EVBACKEND_ALL>	484	=item C<EVBACKEND_ALL>
284		485
285	Try all backends (even potentially broken ones that wouldn't be tried	486	Try all backends (even potentially broken ones that wouldn't be tried
286	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as	487	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as
287	C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>.	488	C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>.
288		489
		490	It is definitely not recommended to use this flag.
		491
289	=back	492	=back
290		493
291	If one or more of these are ored into the flags value, then only these	494	If one or more of these are or'ed into the flags value, then only these
292	backends will be tried (in the reverse order as given here). If none are	495	backends will be tried (in the reverse order as listed here). If none are
293	specified, most compiled-in backend will be tried, usually in reverse	496	specified, all backends in C<ev_recommended_backends ()> will be tried.
294	order of their flag values :)
295		497
296	The most typical usage is like this:	498	Example: This is the most typical usage.
297		499
298	if (!ev_default_loop (0))	500	if (!ev_default_loop (0))
299	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");	501	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
300		502
301	Restrict libev to the select and poll backends, and do not allow	503	Example: Restrict libev to the select and poll backends, and do not allow
302	environment settings to be taken into account:	504	environment settings to be taken into account:
303		505
304	ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);	506	ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);
305		507
306	Use whatever libev has to offer, but make sure that kqueue is used if	508	Example: Use whatever libev has to offer, but make sure that kqueue is
307	available (warning, breaks stuff, best use only with your own private	509	used if available (warning, breaks stuff, best use only with your own
308	event loop and only if you know the OS supports your types of fds):	510	private event loop and only if you know the OS supports your types of
		511	fds):
309		512
310	ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);	513	ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);
311		514
312	=item struct ev_loop *ev_loop_new (unsigned int flags)	515	=item struct ev_loop *ev_loop_new (unsigned int flags)
313		516
314	Similar to C<ev_default_loop>, but always creates a new event loop that is	517	Similar to C<ev_default_loop>, but always creates a new event loop that is
315	always distinct from the default loop. Unlike the default loop, it cannot	518	always distinct from the default loop. Unlike the default loop, it cannot
316	handle signal and child watchers, and attempts to do so will be greeted by	519	handle signal and child watchers, and attempts to do so will be greeted by
317	undefined behaviour (or a failed assertion if assertions are enabled).	520	undefined behaviour (or a failed assertion if assertions are enabled).
318		521
		522	Note that this function I<is> thread-safe, and the recommended way to use
		523	libev with threads is indeed to create one loop per thread, and using the
		524	default loop in the "main" or "initial" thread.
		525
319	Example: try to create a event loop that uses epoll and nothing else.	526	Example: Try to create a event loop that uses epoll and nothing else.
320		527
321	struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);	528	struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);
322	if (!epoller)	529	if (!epoller)
323	fatal ("no epoll found here, maybe it hides under your chair");	530	fatal ("no epoll found here, maybe it hides under your chair");
324		531
325	=item ev_default_destroy ()	532	=item ev_default_destroy ()
326		533
327	Destroys the default loop again (frees all memory and kernel state	534	Destroys the default loop again (frees all memory and kernel state
328	etc.). None of the active event watchers will be stopped in the normal	535	etc.). None of the active event watchers will be stopped in the normal
329	sense, so e.g. C<ev_is_active> might still return true. It is your	536	sense, so e.g. C<ev_is_active> might still return true. It is your
330	responsibility to either stop all watchers cleanly yoursef I<before>	537	responsibility to either stop all watchers cleanly yourself I<before>
331	calling this function, or cope with the fact afterwards (which is usually	538	calling this function, or cope with the fact afterwards (which is usually
332	the easiest thing, youc na just ignore the watchers and/or C<free ()> them	539	the easiest thing, you can just ignore the watchers and/or C<free ()> them
333	for example).	540	for example).
		541
		542	Note that certain global state, such as signal state (and installed signal
		543	handlers), will not be freed by this function, and related watchers (such
		544	as signal and child watchers) would need to be stopped manually.
		545
		546	In general it is not advisable to call this function except in the
		547	rare occasion where you really need to free e.g. the signal handling
		548	pipe fds. If you need dynamically allocated loops it is better to use
		549	C<ev_loop_new> and C<ev_loop_destroy>).
334		550
335	=item ev_loop_destroy (loop)	551	=item ev_loop_destroy (loop)
336		552
337	Like C<ev_default_destroy>, but destroys an event loop created by an	553	Like C<ev_default_destroy>, but destroys an event loop created by an
338	earlier call to C<ev_loop_new>.	554	earlier call to C<ev_loop_new>.
339		555
340	=item ev_default_fork ()	556	=item ev_default_fork ()
341		557
		558	This function sets a flag that causes subsequent C<ev_loop> iterations
342	This function reinitialises the kernel state for backends that have	559	to reinitialise the kernel state for backends that have one. Despite the
343	one. Despite the name, you can call it anytime, but it makes most sense	560	name, you can call it anytime, but it makes most sense after forking, in
344	after forking, in either the parent or child process (or both, but that	561	the child process (or both child and parent, but that again makes little
345	again makes little sense).	562	sense). You I<must> call it in the child before using any of the libev
		563	functions, and it will only take effect at the next C<ev_loop> iteration.
346		564
347	You I<must> call this function in the child process after forking if and	565	On the other hand, you only need to call this function in the child
348	only if you want to use the event library in both processes. If you just	566	process if and only if you want to use the event library in the child. If
349	fork+exec, you don't have to call it.	567	you just fork+exec, you don't have to call it at all.
350		568
351	The function itself is quite fast and it's usually not a problem to call	569	The function itself is quite fast and it's usually not a problem to call
352	it just in case after a fork. To make this easy, the function will fit in	570	it just in case after a fork. To make this easy, the function will fit in
353	quite nicely into a call to C<pthread_atfork>:	571	quite nicely into a call to C<pthread_atfork>:
354		572
355	pthread_atfork (0, 0, ev_default_fork);	573	pthread_atfork (0, 0, ev_default_fork);
356		574
357	At the moment, C<EVBACKEND_SELECT> and C<EVBACKEND_POLL> are safe to use
358	without calling this function, so if you force one of those backends you
359	do not need to care.
360
361	=item ev_loop_fork (loop)	575	=item ev_loop_fork (loop)
362		576
363	Like C<ev_default_fork>, but acts on an event loop created by	577	Like C<ev_default_fork>, but acts on an event loop created by
364	C<ev_loop_new>. Yes, you have to call this on every allocated event loop	578	C<ev_loop_new>. Yes, you have to call this on every allocated event loop
365	after fork, and how you do this is entirely your own problem.	579	after fork that you want to re-use in the child, and how you do this is
		580	entirely your own problem.
		581
		582	=item int ev_is_default_loop (loop)
		583
		584	Returns true when the given loop is, in fact, the default loop, and false
		585	otherwise.
		586
		587	=item unsigned int ev_loop_count (loop)
		588
		589	Returns the count of loop iterations for the loop, which is identical to
		590	the number of times libev did poll for new events. It starts at C<0> and
		591	happily wraps around with enough iterations.
		592
		593	This value can sometimes be useful as a generation counter of sorts (it
		594	"ticks" the number of loop iterations), as it roughly corresponds with
		595	C<ev_prepare> and C<ev_check> calls.
366		596
367	=item unsigned int ev_backend (loop)	597	=item unsigned int ev_backend (loop)
368		598
369	Returns one of the C<EVBACKEND_*> flags indicating the event backend in	599	Returns one of the C<EVBACKEND_*> flags indicating the event backend in
370	use.	600	use.
…		…
373		603
374	Returns the current "event loop time", which is the time the event loop	604	Returns the current "event loop time", which is the time the event loop
375	received events and started processing them. This timestamp does not	605	received events and started processing them. This timestamp does not
376	change as long as callbacks are being processed, and this is also the base	606	change as long as callbacks are being processed, and this is also the base
377	time used for relative timers. You can treat it as the timestamp of the	607	time used for relative timers. You can treat it as the timestamp of the
378	event occuring (or more correctly, libev finding out about it).	608	event occurring (or more correctly, libev finding out about it).
		609
		610	=item ev_now_update (loop)
		611
		612	Establishes the current time by querying the kernel, updating the time
		613	returned by C<ev_now ()> in the progress. This is a costly operation and
		614	is usually done automatically within C<ev_loop ()>.
		615
		616	This function is rarely useful, but when some event callback runs for a
		617	very long time without entering the event loop, updating libev's idea of
		618	the current time is a good idea.
		619
		620	See also "The special problem of time updates" in the C<ev_timer> section.
379		621
380	=item ev_loop (loop, int flags)	622	=item ev_loop (loop, int flags)
381		623
382	Finally, this is it, the event handler. This function usually is called	624	Finally, this is it, the event handler. This function usually is called
383	after you initialised all your watchers and you want to start handling	625	after you initialised all your watchers and you want to start handling
…		…
386	If the flags argument is specified as C<0>, it will not return until	628	If the flags argument is specified as C<0>, it will not return until
387	either no event watchers are active anymore or C<ev_unloop> was called.	629	either no event watchers are active anymore or C<ev_unloop> was called.
388		630
389	Please note that an explicit C<ev_unloop> is usually better than	631	Please note that an explicit C<ev_unloop> is usually better than
390	relying on all watchers to be stopped when deciding when a program has	632	relying on all watchers to be stopped when deciding when a program has
391	finished (especially in interactive programs), but having a program that	633	finished (especially in interactive programs), but having a program
392	automatically loops as long as it has to and no longer by virtue of	634	that automatically loops as long as it has to and no longer by virtue
393	relying on its watchers stopping correctly is a thing of beauty.	635	of relying on its watchers stopping correctly, that is truly a thing of
		636	beauty.
394		637
395	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle	638	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle
396	those events and any outstanding ones, but will not block your process in	639	those events and any already outstanding ones, but will not block your
397	case there are no events and will return after one iteration of the loop.	640	process in case there are no events and will return after one iteration of
		641	the loop.
398		642
399	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if	643	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if
400	neccessary) and will handle those and any outstanding ones. It will block	644	necessary) and will handle those and any already outstanding ones. It
401	your process until at least one new event arrives, and will return after	645	will block your process until at least one new event arrives (which could
402	one iteration of the loop. This is useful if you are waiting for some	646	be an event internal to libev itself, so there is no guarentee that a
403	external event in conjunction with something not expressible using other	647	user-registered callback will be called), and will return after one
		648	iteration of the loop.
		649
		650	This is useful if you are waiting for some external event in conjunction
		651	with something not expressible using other libev watchers (i.e. "roll your
404	libev watchers. However, a pair of C<ev_prepare>/C<ev_check> watchers is	652	own C<ev_loop>"). However, a pair of C<ev_prepare>/C<ev_check> watchers is
405	usually a better approach for this kind of thing.	653	usually a better approach for this kind of thing.
406		654
407	Here are the gory details of what C<ev_loop> does:	655	Here are the gory details of what C<ev_loop> does:
408		656
409	* If there are no active watchers (reference count is zero), return.	657	- Before the first iteration, call any pending watchers.
410	- Queue prepare watchers and then call all outstanding watchers.	658	* If EVFLAG_FORKCHECK was used, check for a fork.
		659	- If a fork was detected (by any means), queue and call all fork watchers.
		660	- Queue and call all prepare watchers.
411	- If we have been forked, recreate the kernel state.	661	- If we have been forked, detach and recreate the kernel state
		662	as to not disturb the other process.
412	- Update the kernel state with all outstanding changes.	663	- Update the kernel state with all outstanding changes.
413	- Update the "event loop time".	664	- Update the "event loop time" (ev_now ()).
414	- Calculate for how long to block.	665	- Calculate for how long to sleep or block, if at all
		666	(active idle watchers, EVLOOP_NONBLOCK or not having
		667	any active watchers at all will result in not sleeping).
		668	- Sleep if the I/O and timer collect interval say so.
415	- Block the process, waiting for any events.	669	- Block the process, waiting for any events.
416	- Queue all outstanding I/O (fd) events.	670	- Queue all outstanding I/O (fd) events.
417	- Update the "event loop time" and do time jump handling.	671	- Update the "event loop time" (ev_now ()), and do time jump adjustments.
418	- Queue all outstanding timers.	672	- Queue all expired timers.
419	- Queue all outstanding periodics.	673	- Queue all expired periodics.
420	- If no events are pending now, queue all idle watchers.	674	- Unless any events are pending now, queue all idle watchers.
421	- Queue all check watchers.	675	- Queue all check watchers.
422	- Call all queued watchers in reverse order (i.e. check watchers first).	676	- Call all queued watchers in reverse order (i.e. check watchers first).
423	Signals and child watchers are implemented as I/O watchers, and will	677	Signals and child watchers are implemented as I/O watchers, and will
424	be handled here by queueing them when their watcher gets executed.	678	be handled here by queueing them when their watcher gets executed.
425	- If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK	679	- If ev_unloop has been called, or EVLOOP_ONESHOT or EVLOOP_NONBLOCK
426	were used, return, otherwise continue with step *.	680	were used, or there are no active watchers, return, otherwise
		681	continue with step *.
427		682
428	Example: queue some jobs and then loop until no events are outsanding	683	Example: Queue some jobs and then loop until no events are outstanding
429	anymore.	684	anymore.
430		685
431	... queue jobs here, make sure they register event watchers as long	686	... queue jobs here, make sure they register event watchers as long
432	... as they still have work to do (even an idle watcher will do..)	687	... as they still have work to do (even an idle watcher will do..)
433	ev_loop (my_loop, 0);	688	ev_loop (my_loop, 0);
434	... jobs done. yeah!	689	... jobs done or somebody called unloop. yeah!
435		690
436	=item ev_unloop (loop, how)	691	=item ev_unloop (loop, how)
437		692
438	Can be used to make a call to C<ev_loop> return early (but only after it	693	Can be used to make a call to C<ev_loop> return early (but only after it
439	has processed all outstanding events). The C<how> argument must be either	694	has processed all outstanding events). The C<how> argument must be either
440	C<EVUNLOOP_ONE>, which will make the innermost C<ev_loop> call return, or	695	C<EVUNLOOP_ONE>, which will make the innermost C<ev_loop> call return, or
441	C<EVUNLOOP_ALL>, which will make all nested C<ev_loop> calls return.	696	C<EVUNLOOP_ALL>, which will make all nested C<ev_loop> calls return.
442		697
		698	This "unloop state" will be cleared when entering C<ev_loop> again.
		699
		700	It is safe to call C<ev_unloop> from otuside any C<ev_loop> calls.
		701
443	=item ev_ref (loop)	702	=item ev_ref (loop)
444		703
445	=item ev_unref (loop)	704	=item ev_unref (loop)
446		705
447	Ref/unref can be used to add or remove a reference count on the event	706	Ref/unref can be used to add or remove a reference count on the event
448	loop: Every watcher keeps one reference, and as long as the reference	707	loop: Every watcher keeps one reference, and as long as the reference
449	count is nonzero, C<ev_loop> will not return on its own. If you have	708	count is nonzero, C<ev_loop> will not return on its own.
		709
450	a watcher you never unregister that should not keep C<ev_loop> from	710	If you have a watcher you never unregister that should not keep C<ev_loop>
451	returning, ev_unref() after starting, and ev_ref() before stopping it. For	711	from returning, call ev_unref() after starting, and ev_ref() before
		712	stopping it.
		713
452	example, libev itself uses this for its internal signal pipe: It is not	714	As an example, libev itself uses this for its internal signal pipe: It is
453	visible to the libev user and should not keep C<ev_loop> from exiting if	715	not visible to the libev user and should not keep C<ev_loop> from exiting
454	no event watchers registered by it are active. It is also an excellent	716	if no event watchers registered by it are active. It is also an excellent
455	way to do this for generic recurring timers or from within third-party	717	way to do this for generic recurring timers or from within third-party
456	libraries. Just remember to I<unref after start> and I<ref before stop>.	718	libraries. Just remember to I<unref after start> and I<ref before stop>
		719	(but only if the watcher wasn't active before, or was active before,
		720	respectively).
457		721
458	Example: create a signal watcher, but keep it from keeping C<ev_loop>	722	Example: Create a signal watcher, but keep it from keeping C<ev_loop>
459	running when nothing else is active.	723	running when nothing else is active.
460		724
461	struct dv_signal exitsig;	725	ev_signal exitsig;
462	ev_signal_init (&exitsig, sig_cb, SIGINT);	726	ev_signal_init (&exitsig, sig_cb, SIGINT);
463	ev_signal_start (myloop, &exitsig);	727	ev_signal_start (loop, &exitsig);
464	evf_unref (myloop);	728	evf_unref (loop);
465		729
466	Example: for some weird reason, unregister the above signal handler again.	730	Example: For some weird reason, unregister the above signal handler again.
467		731
468	ev_ref (myloop);	732	ev_ref (loop);
469	ev_signal_stop (myloop, &exitsig);	733	ev_signal_stop (loop, &exitsig);
		734
		735	=item ev_set_io_collect_interval (loop, ev_tstamp interval)
		736
		737	=item ev_set_timeout_collect_interval (loop, ev_tstamp interval)
		738
		739	These advanced functions influence the time that libev will spend waiting
		740	for events. Both time intervals are by default C<0>, meaning that libev
		741	will try to invoke timer/periodic callbacks and I/O callbacks with minimum
		742	latency.
		743
		744	Setting these to a higher value (the C<interval> I<must> be >= C<0>)
		745	allows libev to delay invocation of I/O and timer/periodic callbacks
		746	to increase efficiency of loop iterations (or to increase power-saving
		747	opportunities).
		748
		749	The idea is that sometimes your program runs just fast enough to handle
		750	one (or very few) event(s) per loop iteration. While this makes the
		751	program responsive, it also wastes a lot of CPU time to poll for new
		752	events, especially with backends like C<select ()> which have a high
		753	overhead for the actual polling but can deliver many events at once.
		754
		755	By setting a higher I<io collect interval> you allow libev to spend more
		756	time collecting I/O events, so you can handle more events per iteration,
		757	at the cost of increasing latency. Timeouts (both C<ev_periodic> and
		758	C<ev_timer>) will be not affected. Setting this to a non-null value will
		759	introduce an additional C<ev_sleep ()> call into most loop iterations.
		760
		761	Likewise, by setting a higher I<timeout collect interval> you allow libev
		762	to spend more time collecting timeouts, at the expense of increased
		763	latency/jitter/inexactness (the watcher callback will be called
		764	later). C<ev_io> watchers will not be affected. Setting this to a non-null
		765	value will not introduce any overhead in libev.
		766
		767	Many (busy) programs can usually benefit by setting the I/O collect
		768	interval to a value near C<0.1> or so, which is often enough for
		769	interactive servers (of course not for games), likewise for timeouts. It
		770	usually doesn't make much sense to set it to a lower value than C<0.01>,
		771	as this approaches the timing granularity of most systems.
		772
		773	Setting the I<timeout collect interval> can improve the opportunity for
		774	saving power, as the program will "bundle" timer callback invocations that
		775	are "near" in time together, by delaying some, thus reducing the number of
		776	times the process sleeps and wakes up again. Another useful technique to
		777	reduce iterations/wake-ups is to use C<ev_periodic> watchers and make sure
		778	they fire on, say, one-second boundaries only.
		779
		780	=item ev_loop_verify (loop)
		781
		782	This function only does something when C<EV_VERIFY> support has been
		783	compiled in, which is the default for non-minimal builds. It tries to go
		784	through all internal structures and checks them for validity. If anything
		785	is found to be inconsistent, it will print an error message to standard
		786	error and call C<abort ()>.
		787
		788	This can be used to catch bugs inside libev itself: under normal
		789	circumstances, this function will never abort as of course libev keeps its
		790	data structures consistent.
470		791
471	=back	792	=back
472		793
		794
473	=head1 ANATOMY OF A WATCHER	795	=head1 ANATOMY OF A WATCHER
		796
		797	In the following description, uppercase C<TYPE> in names stands for the
		798	watcher type, e.g. C<ev_TYPE_start> can mean C<ev_timer_start> for timer
		799	watchers and C<ev_io_start> for I/O watchers.
474		800
475	A watcher is a structure that you create and register to record your	801	A watcher is a structure that you create and register to record your
476	interest in some event. For instance, if you want to wait for STDIN to	802	interest in some event. For instance, if you want to wait for STDIN to
477	become readable, you would create an C<ev_io> watcher for that:	803	become readable, you would create an C<ev_io> watcher for that:
478		804
479	static void my_cb (struct ev_loop *loop, struct ev_io *w, int revents)	805	static void my_cb (struct ev_loop *loop, ev_io *w, int revents)
480	{	806	{
481	ev_io_stop (w);	807	ev_io_stop (w);
482	ev_unloop (loop, EVUNLOOP_ALL);	808	ev_unloop (loop, EVUNLOOP_ALL);
483	}	809	}
484		810
485	struct ev_loop *loop = ev_default_loop (0);	811	struct ev_loop *loop = ev_default_loop (0);
		812
486	struct ev_io stdin_watcher;	813	ev_io stdin_watcher;
		814
487	ev_init (&stdin_watcher, my_cb);	815	ev_init (&stdin_watcher, my_cb);
488	ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);	816	ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);
489	ev_io_start (loop, &stdin_watcher);	817	ev_io_start (loop, &stdin_watcher);
		818
490	ev_loop (loop, 0);	819	ev_loop (loop, 0);
491		820
492	As you can see, you are responsible for allocating the memory for your	821	As you can see, you are responsible for allocating the memory for your
493	watcher structures (and it is usually a bad idea to do this on the stack,	822	watcher structures (and it is I<usually> a bad idea to do this on the
494	although this can sometimes be quite valid).	823	stack).
		824
		825	Each watcher has an associated watcher structure (called C<struct ev_TYPE>
		826	or simply C<ev_TYPE>, as typedefs are provided for all watcher structs).
495		827
496	Each watcher structure must be initialised by a call to C<ev_init	828	Each watcher structure must be initialised by a call to C<ev_init
497	(watcher *, callback)>, which expects a callback to be provided. This	829	(watcher *, callback)>, which expects a callback to be provided. This
498	callback gets invoked each time the event occurs (or, in the case of io	830	callback gets invoked each time the event occurs (or, in the case of I/O
499	watchers, each time the event loop detects that the file descriptor given	831	watchers, each time the event loop detects that the file descriptor given
500	is readable and/or writable).	832	is readable and/or writable).
501		833
502	Each watcher type has its own C<< ev_<type>_set (watcher *, ...) >> macro	834	Each watcher type further has its own C<< ev_TYPE_set (watcher *, ...) >>
503	with arguments specific to this watcher type. There is also a macro	835	macro to configure it, with arguments specific to the watcher type. There
504	to combine initialisation and setting in one call: C<< ev_<type>_init	836	is also a macro to combine initialisation and setting in one call: C<<
505	(watcher *, callback, ...) >>.	837	ev_TYPE_init (watcher *, callback, ...) >>.
506		838
507	To make the watcher actually watch out for events, you have to start it	839	To make the watcher actually watch out for events, you have to start it
508	with a watcher-specific start function (C<< ev_<type>_start (loop, watcher	840	with a watcher-specific start function (C<< ev_TYPE_start (loop, watcher
509	*) >>), and you can stop watching for events at any time by calling the	841	*) >>), and you can stop watching for events at any time by calling the
510	corresponding stop function (C<< ev_<type>_stop (loop, watcher *) >>.	842	corresponding stop function (C<< ev_TYPE_stop (loop, watcher *) >>.
511		843
512	As long as your watcher is active (has been started but not stopped) you	844	As long as your watcher is active (has been started but not stopped) you
513	must not touch the values stored in it. Most specifically you must never	845	must not touch the values stored in it. Most specifically you must never
514	reinitialise it or call its C<set> macro.	846	reinitialise it or call its C<ev_TYPE_set> macro.
515		847
516	Each and every callback receives the event loop pointer as first, the	848	Each and every callback receives the event loop pointer as first, the
517	registered watcher structure as second, and a bitset of received events as	849	registered watcher structure as second, and a bitset of received events as
518	third argument.	850	third argument.
519		851
…		…
543	The signal specified in the C<ev_signal> watcher has been received by a thread.	875	The signal specified in the C<ev_signal> watcher has been received by a thread.
544		876
545	=item C<EV_CHILD>	877	=item C<EV_CHILD>
546		878
547	The pid specified in the C<ev_child> watcher has received a status change.	879	The pid specified in the C<ev_child> watcher has received a status change.
		880
		881	=item C<EV_STAT>
		882
		883	The path specified in the C<ev_stat> watcher changed its attributes somehow.
548		884
549	=item C<EV_IDLE>	885	=item C<EV_IDLE>
550		886
551	The C<ev_idle> watcher has determined that you have nothing better to do.	887	The C<ev_idle> watcher has determined that you have nothing better to do.
552		888
…		…
560	received events. Callbacks of both watcher types can start and stop as	896	received events. Callbacks of both watcher types can start and stop as
561	many watchers as they want, and all of them will be taken into account	897	many watchers as they want, and all of them will be taken into account
562	(for example, a C<ev_prepare> watcher might start an idle watcher to keep	898	(for example, a C<ev_prepare> watcher might start an idle watcher to keep
563	C<ev_loop> from blocking).	899	C<ev_loop> from blocking).
564		900
		901	=item C<EV_EMBED>
		902
		903	The embedded event loop specified in the C<ev_embed> watcher needs attention.
		904
		905	=item C<EV_FORK>
		906
		907	The event loop has been resumed in the child process after fork (see
		908	C<ev_fork>).
		909
		910	=item C<EV_ASYNC>
		911
		912	The given async watcher has been asynchronously notified (see C<ev_async>).
		913
565	=item C<EV_ERROR>	914	=item C<EV_ERROR>
566		915
567	An unspecified error has occured, the watcher has been stopped. This might	916	An unspecified error has occurred, the watcher has been stopped. This might
568	happen because the watcher could not be properly started because libev	917	happen because the watcher could not be properly started because libev
569	ran out of memory, a file descriptor was found to be closed or any other	918	ran out of memory, a file descriptor was found to be closed or any other
		919	problem. Libev considers these application bugs.
		920
570	problem. You best act on it by reporting the problem and somehow coping	921	You best act on it by reporting the problem and somehow coping with the
571	with the watcher being stopped.	922	watcher being stopped. Note that well-written programs should not receive
		923	an error ever, so when your watcher receives it, this usually indicates a
		924	bug in your program.
572		925
573	Libev will usually signal a few "dummy" events together with an error,	926	Libev will usually signal a few "dummy" events together with an error, for
574	for example it might indicate that a fd is readable or writable, and if	927	example it might indicate that a fd is readable or writable, and if your
575	your callbacks is well-written it can just attempt the operation and cope	928	callbacks is well-written it can just attempt the operation and cope with
576	with the error from read() or write(). This will not work in multithreaded	929	the error from read() or write(). This will not work in multi-threaded
577	programs, though, so beware.	930	programs, though, as the fd could already be closed and reused for another
		931	thing, so beware.
578		932
579	=back	933	=back
580		934
581	=head2 SUMMARY OF GENERIC WATCHER FUNCTIONS	935	=head2 GENERIC WATCHER FUNCTIONS
582
583	In the following description, C<TYPE> stands for the watcher type,
584	e.g. C<timer> for C<ev_timer> watchers and C<io> for C<ev_io> watchers.
585		936
586	=over 4	937	=over 4
587		938
588	=item C<ev_init> (ev_TYPE *watcher, callback)	939	=item C<ev_init> (ev_TYPE *watcher, callback)
589		940
…		…
595	which rolls both calls into one.	946	which rolls both calls into one.
596		947
597	You can reinitialise a watcher at any time as long as it has been stopped	948	You can reinitialise a watcher at any time as long as it has been stopped
598	(or never started) and there are no pending events outstanding.	949	(or never started) and there are no pending events outstanding.
599		950
600	The callbakc is always of type C<void (*)(ev_loop *loop, ev_TYPE *watcher,	951	The callback is always of type C<void (*)(struct ev_loop *loop, ev_TYPE *watcher,
601	int revents)>.	952	int revents)>.
		953
		954	Example: Initialise an C<ev_io> watcher in two steps.
		955
		956	ev_io w;
		957	ev_init (&w, my_cb);
		958	ev_io_set (&w, STDIN_FILENO, EV_READ);
602		959
603	=item C<ev_TYPE_set> (ev_TYPE *, [args])	960	=item C<ev_TYPE_set> (ev_TYPE *, [args])
604		961
605	This macro initialises the type-specific parts of a watcher. You need to	962	This macro initialises the type-specific parts of a watcher. You need to
606	call C<ev_init> at least once before you call this macro, but you can	963	call C<ev_init> at least once before you call this macro, but you can
…		…
609	difference to the C<ev_init> macro).	966	difference to the C<ev_init> macro).
610		967
611	Although some watcher types do not have type-specific arguments	968	Although some watcher types do not have type-specific arguments
612	(e.g. C<ev_prepare>) you still need to call its C<set> macro.	969	(e.g. C<ev_prepare>) you still need to call its C<set> macro.
613		970
		971	See C<ev_init>, above, for an example.
		972
614	=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])	973	=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])
615		974
616	This convinience macro rolls both C<ev_init> and C<ev_TYPE_set> macro	975	This convenience macro rolls both C<ev_init> and C<ev_TYPE_set> macro
617	calls into a single call. This is the most convinient method to initialise	976	calls into a single call. This is the most convenient method to initialise
618	a watcher. The same limitations apply, of course.	977	a watcher. The same limitations apply, of course.
		978
		979	Example: Initialise and set an C<ev_io> watcher in one step.
		980
		981	ev_io_init (&w, my_cb, STDIN_FILENO, EV_READ);
619		982
620	=item C<ev_TYPE_start> (loop *, ev_TYPE *watcher)	983	=item C<ev_TYPE_start> (loop *, ev_TYPE *watcher)
621		984
622	Starts (activates) the given watcher. Only active watchers will receive	985	Starts (activates) the given watcher. Only active watchers will receive
623	events. If the watcher is already active nothing will happen.	986	events. If the watcher is already active nothing will happen.
624		987
		988	Example: Start the C<ev_io> watcher that is being abused as example in this
		989	whole section.
		990
		991	ev_io_start (EV_DEFAULT_UC, &w);
		992
625	=item C<ev_TYPE_stop> (loop *, ev_TYPE *watcher)	993	=item C<ev_TYPE_stop> (loop *, ev_TYPE *watcher)
626		994
627	Stops the given watcher again (if active) and clears the pending	995	Stops the given watcher if active, and clears the pending status (whether
		996	the watcher was active or not).
		997
628	status. It is possible that stopped watchers are pending (for example,	998	It is possible that stopped watchers are pending - for example,
629	non-repeating timers are being stopped when they become pending), but	999	non-repeating timers are being stopped when they become pending - but
630	C<ev_TYPE_stop> ensures that the watcher is neither active nor pending. If	1000	calling C<ev_TYPE_stop> ensures that the watcher is neither active nor
631	you want to free or reuse the memory used by the watcher it is therefore a	1001	pending. If you want to free or reuse the memory used by the watcher it is
632	good idea to always call its C<ev_TYPE_stop> function.	1002	therefore a good idea to always call its C<ev_TYPE_stop> function.
633		1003
634	=item bool ev_is_active (ev_TYPE *watcher)	1004	=item bool ev_is_active (ev_TYPE *watcher)
635		1005
636	Returns a true value iff the watcher is active (i.e. it has been started	1006	Returns a true value iff the watcher is active (i.e. it has been started
637	and not yet been stopped). As long as a watcher is active you must not modify	1007	and not yet been stopped). As long as a watcher is active you must not modify
…		…
640	=item bool ev_is_pending (ev_TYPE *watcher)	1010	=item bool ev_is_pending (ev_TYPE *watcher)
641		1011
642	Returns a true value iff the watcher is pending, (i.e. it has outstanding	1012	Returns a true value iff the watcher is pending, (i.e. it has outstanding
643	events but its callback has not yet been invoked). As long as a watcher	1013	events but its callback has not yet been invoked). As long as a watcher
644	is pending (but not active) you must not call an init function on it (but	1014	is pending (but not active) you must not call an init function on it (but
645	C<ev_TYPE_set> is safe) and you must make sure the watcher is available to	1015	C<ev_TYPE_set> is safe), you must not change its priority, and you must
646	libev (e.g. you cnanot C<free ()> it).	1016	make sure the watcher is available to libev (e.g. you cannot C<free ()>
		1017	it).
647		1018
648	=item callback = ev_cb (ev_TYPE *watcher)	1019	=item callback ev_cb (ev_TYPE *watcher)
649		1020
650	Returns the callback currently set on the watcher.	1021	Returns the callback currently set on the watcher.
651		1022
652	=item ev_cb_set (ev_TYPE *watcher, callback)	1023	=item ev_cb_set (ev_TYPE *watcher, callback)
653		1024
654	Change the callback. You can change the callback at virtually any time	1025	Change the callback. You can change the callback at virtually any time
655	(modulo threads).	1026	(modulo threads).
656		1027
		1028	=item ev_set_priority (ev_TYPE *watcher, priority)
		1029
		1030	=item int ev_priority (ev_TYPE *watcher)
		1031
		1032	Set and query the priority of the watcher. The priority is a small
		1033	integer between C<EV_MAXPRI> (default: C<2>) and C<EV_MINPRI>
		1034	(default: C<-2>). Pending watchers with higher priority will be invoked
		1035	before watchers with lower priority, but priority will not keep watchers
		1036	from being executed (except for C<ev_idle> watchers).
		1037
		1038	This means that priorities are I<only> used for ordering callback
		1039	invocation after new events have been received. This is useful, for
		1040	example, to reduce latency after idling, or more often, to bind two
		1041	watchers on the same event and make sure one is called first.
		1042
		1043	If you need to suppress invocation when higher priority events are pending
		1044	you need to look at C<ev_idle> watchers, which provide this functionality.
		1045
		1046	You I<must not> change the priority of a watcher as long as it is active or
		1047	pending.
		1048
		1049	The default priority used by watchers when no priority has been set is
		1050	always C<0>, which is supposed to not be too high and not be too low :).
		1051
		1052	Setting a priority outside the range of C<EV_MINPRI> to C<EV_MAXPRI> is
		1053	fine, as long as you do not mind that the priority value you query might
		1054	or might not have been clamped to the valid range.
		1055
		1056	=item ev_invoke (loop, ev_TYPE *watcher, int revents)
		1057
		1058	Invoke the C<watcher> with the given C<loop> and C<revents>. Neither
		1059	C<loop> nor C<revents> need to be valid as long as the watcher callback
		1060	can deal with that fact, as both are simply passed through to the
		1061	callback.
		1062
		1063	=item int ev_clear_pending (loop, ev_TYPE *watcher)
		1064
		1065	If the watcher is pending, this function clears its pending status and
		1066	returns its C<revents> bitset (as if its callback was invoked). If the
		1067	watcher isn't pending it does nothing and returns C<0>.
		1068
		1069	Sometimes it can be useful to "poll" a watcher instead of waiting for its
		1070	callback to be invoked, which can be accomplished with this function.
		1071
657	=back	1072	=back
658		1073
659		1074
660	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER	1075	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER
661		1076
662	Each watcher has, by default, a member C<void *data> that you can change	1077	Each watcher has, by default, a member C<void *data> that you can change
663	and read at any time, libev will completely ignore it. This can be used	1078	and read at any time: libev will completely ignore it. This can be used
664	to associate arbitrary data with your watcher. If you need more data and	1079	to associate arbitrary data with your watcher. If you need more data and
665	don't want to allocate memory and store a pointer to it in that data	1080	don't want to allocate memory and store a pointer to it in that data
666	member, you can also "subclass" the watcher type and provide your own	1081	member, you can also "subclass" the watcher type and provide your own
667	data:	1082	data:
668		1083
669	struct my_io	1084	struct my_io
670	{	1085	{
671	struct ev_io io;	1086	ev_io io;
672	int otherfd;	1087	int otherfd;
673	void *somedata;	1088	void *somedata;
674	struct whatever *mostinteresting;	1089	struct whatever *mostinteresting;
675	}	1090	};
		1091
		1092	...
		1093	struct my_io w;
		1094	ev_io_init (&w.io, my_cb, fd, EV_READ);
676		1095
677	And since your callback will be called with a pointer to the watcher, you	1096	And since your callback will be called with a pointer to the watcher, you
678	can cast it back to your own type:	1097	can cast it back to your own type:
679		1098
680	static void my_cb (struct ev_loop *loop, struct ev_io *w_, int revents)	1099	static void my_cb (struct ev_loop *loop, ev_io *w_, int revents)
681	{	1100	{
682	struct my_io *w = (struct my_io *)w_;	1101	struct my_io *w = (struct my_io *)w_;
683	...	1102	...
684	}	1103	}
685		1104
686	More interesting and less C-conformant ways of catsing your callback type	1105	More interesting and less C-conformant ways of casting your callback type
687	have been omitted....	1106	instead have been omitted.
		1107
		1108	Another common scenario is to use some data structure with multiple
		1109	embedded watchers:
		1110
		1111	struct my_biggy
		1112	{
		1113	int some_data;
		1114	ev_timer t1;
		1115	ev_timer t2;
		1116	}
		1117
		1118	In this case getting the pointer to C<my_biggy> is a bit more
		1119	complicated: Either you store the address of your C<my_biggy> struct
		1120	in the C<data> member of the watcher (for woozies), or you need to use
		1121	some pointer arithmetic using C<offsetof> inside your watchers (for real
		1122	programmers):
		1123
		1124	#include <stddef.h>
		1125
		1126	static void
		1127	t1_cb (EV_P_ ev_timer *w, int revents)
		1128	{
		1129	struct my_biggy big = (struct my_biggy *
		1130	(((char *)w) - offsetof (struct my_biggy, t1));
		1131	}
		1132
		1133	static void
		1134	t2_cb (EV_P_ ev_timer *w, int revents)
		1135	{
		1136	struct my_biggy big = (struct my_biggy *
		1137	(((char *)w) - offsetof (struct my_biggy, t2));
		1138	}
688		1139
689		1140
690	=head1 WATCHER TYPES	1141	=head1 WATCHER TYPES
691		1142
692	This section describes each watcher in detail, but will not repeat	1143	This section describes each watcher in detail, but will not repeat
693	information given in the last section.	1144	information given in the last section. Any initialisation/set macros,
		1145	functions and members specific to the watcher type are explained.
694		1146
		1147	Members are additionally marked with either I<[read-only]>, meaning that,
		1148	while the watcher is active, you can look at the member and expect some
		1149	sensible content, but you must not modify it (you can modify it while the
		1150	watcher is stopped to your hearts content), or I<[read-write]>, which
		1151	means you can expect it to have some sensible content while the watcher
		1152	is active, but you can also modify it. Modifying it may not do something
		1153	sensible or take immediate effect (or do anything at all), but libev will
		1154	not crash or malfunction in any way.
695		1155
		1156
696	=head2 C<ev_io> - is this file descriptor readable or writable	1157	=head2 C<ev_io> - is this file descriptor readable or writable?
697		1158
698	I/O watchers check whether a file descriptor is readable or writable	1159	I/O watchers check whether a file descriptor is readable or writable
699	in each iteration of the event loop (This behaviour is called	1160	in each iteration of the event loop, or, more precisely, when reading
700	level-triggering because you keep receiving events as long as the	1161	would not block the process and writing would at least be able to write
701	condition persists. Remember you can stop the watcher if you don't want to	1162	some data. This behaviour is called level-triggering because you keep
702	act on the event and neither want to receive future events).	1163	receiving events as long as the condition persists. Remember you can stop
		1164	the watcher if you don't want to act on the event and neither want to
		1165	receive future events.
703		1166
704	In general you can register as many read and/or write event watchers per	1167	In general you can register as many read and/or write event watchers per
705	fd as you want (as long as you don't confuse yourself). Setting all file	1168	fd as you want (as long as you don't confuse yourself). Setting all file
706	descriptors to non-blocking mode is also usually a good idea (but not	1169	descriptors to non-blocking mode is also usually a good idea (but not
707	required if you know what you are doing).	1170	required if you know what you are doing).
708		1171
709	You have to be careful with dup'ed file descriptors, though. Some backends	1172	If you cannot use non-blocking mode, then force the use of a
710	(the linux epoll backend is a notable example) cannot handle dup'ed file	1173	known-to-be-good backend (at the time of this writing, this includes only
711	descriptors correctly if you register interest in two or more fds pointing	1174	C<EVBACKEND_SELECT> and C<EVBACKEND_POLL>).
712	to the same underlying file/socket etc. description (that is, they share
713	the same underlying "file open").
714		1175
715	If you must do this, then force the use of a known-to-be-good backend	1176	Another thing you have to watch out for is that it is quite easy to
716	(at the time of this writing, this includes only C<EVBACKEND_SELECT> and	1177	receive "spurious" readiness notifications, that is your callback might
		1178	be called with C<EV_READ> but a subsequent C<read>(2) will actually block
		1179	because there is no data. Not only are some backends known to create a
		1180	lot of those (for example Solaris ports), it is very easy to get into
		1181	this situation even with a relatively standard program structure. Thus
		1182	it is best to always use non-blocking I/O: An extra C<read>(2) returning
		1183	C<EAGAIN> is far preferable to a program hanging until some data arrives.
		1184
		1185	If you cannot run the fd in non-blocking mode (for example you should
		1186	not play around with an Xlib connection), then you have to separately
		1187	re-test whether a file descriptor is really ready with a known-to-be good
		1188	interface such as poll (fortunately in our Xlib example, Xlib already
		1189	does this on its own, so its quite safe to use). Some people additionally
		1190	use C<SIGALRM> and an interval timer, just to be sure you won't block
		1191	indefinitely.
		1192
		1193	But really, best use non-blocking mode.
		1194
		1195	=head3 The special problem of disappearing file descriptors
		1196
		1197	Some backends (e.g. kqueue, epoll) need to be told about closing a file
		1198	descriptor (either due to calling C<close> explicitly or any other means,
		1199	such as C<dup2>). The reason is that you register interest in some file
		1200	descriptor, but when it goes away, the operating system will silently drop
		1201	this interest. If another file descriptor with the same number then is
		1202	registered with libev, there is no efficient way to see that this is, in
		1203	fact, a different file descriptor.
		1204
		1205	To avoid having to explicitly tell libev about such cases, libev follows
		1206	the following policy: Each time C<ev_io_set> is being called, libev
		1207	will assume that this is potentially a new file descriptor, otherwise
		1208	it is assumed that the file descriptor stays the same. That means that
		1209	you I<have> to call C<ev_io_set> (or C<ev_io_init>) when you change the
		1210	descriptor even if the file descriptor number itself did not change.
		1211
		1212	This is how one would do it normally anyway, the important point is that
		1213	the libev application should not optimise around libev but should leave
		1214	optimisations to libev.
		1215
		1216	=head3 The special problem of dup'ed file descriptors
		1217
		1218	Some backends (e.g. epoll), cannot register events for file descriptors,
		1219	but only events for the underlying file descriptions. That means when you
		1220	have C<dup ()>'ed file descriptors or weirder constellations, and register
		1221	events for them, only one file descriptor might actually receive events.
		1222
		1223	There is no workaround possible except not registering events
		1224	for potentially C<dup ()>'ed file descriptors, or to resort to
		1225	C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>.
		1226
		1227	=head3 The special problem of fork
		1228
		1229	Some backends (epoll, kqueue) do not support C<fork ()> at all or exhibit
		1230	useless behaviour. Libev fully supports fork, but needs to be told about
		1231	it in the child.
		1232
		1233	To support fork in your programs, you either have to call
		1234	C<ev_default_fork ()> or C<ev_loop_fork ()> after a fork in the child,
		1235	enable C<EVFLAG_FORKCHECK>, or resort to C<EVBACKEND_SELECT> or
717	C<EVBACKEND_POLL>).	1236	C<EVBACKEND_POLL>.
		1237
		1238	=head3 The special problem of SIGPIPE
		1239
		1240	While not really specific to libev, it is easy to forget about C<SIGPIPE>:
		1241	when writing to a pipe whose other end has been closed, your program gets
		1242	sent a SIGPIPE, which, by default, aborts your program. For most programs
		1243	this is sensible behaviour, for daemons, this is usually undesirable.
		1244
		1245	So when you encounter spurious, unexplained daemon exits, make sure you
		1246	ignore SIGPIPE (and maybe make sure you log the exit status of your daemon
		1247	somewhere, as that would have given you a big clue).
		1248
		1249
		1250	=head3 Watcher-Specific Functions
718		1251
719	=over 4	1252	=over 4
720		1253
721	=item ev_io_init (ev_io *, callback, int fd, int events)	1254	=item ev_io_init (ev_io *, callback, int fd, int events)
722		1255
723	=item ev_io_set (ev_io *, int fd, int events)	1256	=item ev_io_set (ev_io *, int fd, int events)
724		1257
725	Configures an C<ev_io> watcher. The fd is the file descriptor to rceeive	1258	Configures an C<ev_io> watcher. The C<fd> is the file descriptor to
726	events for and events is either C<EV_READ>, C<EV_WRITE> or C<EV_READ |	1259	receive events for and C<events> is either C<EV_READ>, C<EV_WRITE> or
727	EV_WRITE> to receive the given events.	1260	C<EV_READ | EV_WRITE>, to express the desire to receive the given events.
728		1261
729	Please note that most of the more scalable backend mechanisms (for example	1262	=item int fd [read-only]
730	epoll and solaris ports) can result in spurious readyness notifications	1263
731	for file descriptors, so you practically need to use non-blocking I/O (and	1264	The file descriptor being watched.
732	treat callback invocation as hint only), or retest separately with a safe	1265
733	interface before doing I/O (XLib can do this), or force the use of either	1266	=item int events [read-only]
734	C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>, which don't suffer from this	1267
735	problem. Also note that it is quite easy to have your callback invoked	1268	The events being watched.
736	when the readyness condition is no longer valid even when employing
737	typical ways of handling events, so its a good idea to use non-blocking
738	I/O unconditionally.
739		1269
740	=back	1270	=back
741		1271
		1272	=head3 Examples
		1273
742	Example: call C<stdin_readable_cb> when STDIN_FILENO has become, well	1274	Example: Call C<stdin_readable_cb> when STDIN_FILENO has become, well
743	readable, but only once. Since it is likely line-buffered, you could	1275	readable, but only once. Since it is likely line-buffered, you could
744	attempt to read a whole line in the callback:	1276	attempt to read a whole line in the callback.
745		1277
746	static void	1278	static void
747	stdin_readable_cb (struct ev_loop *loop, struct ev_io *w, int revents)	1279	stdin_readable_cb (struct ev_loop *loop, ev_io *w, int revents)
748	{	1280	{
749	ev_io_stop (loop, w);	1281	ev_io_stop (loop, w);
750	.. read from stdin here (or from w->fd) and haqndle any I/O errors	1282	.. read from stdin here (or from w->fd) and handle any I/O errors
751	}	1283	}
752		1284
753	...	1285	...
754	struct ev_loop *loop = ev_default_init (0);	1286	struct ev_loop *loop = ev_default_init (0);
755	struct ev_io stdin_readable;	1287	ev_io stdin_readable;
756	ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);	1288	ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
757	ev_io_start (loop, &stdin_readable);	1289	ev_io_start (loop, &stdin_readable);
758	ev_loop (loop, 0);	1290	ev_loop (loop, 0);
759		1291
760		1292
761	=head2 C<ev_timer> - relative and optionally recurring timeouts	1293	=head2 C<ev_timer> - relative and optionally repeating timeouts
762		1294
763	Timer watchers are simple relative timers that generate an event after a	1295	Timer watchers are simple relative timers that generate an event after a
764	given time, and optionally repeating in regular intervals after that.	1296	given time, and optionally repeating in regular intervals after that.
765		1297
766	The timers are based on real time, that is, if you register an event that	1298	The timers are based on real time, that is, if you register an event that
767	times out after an hour and you reset your system clock to last years	1299	times out after an hour and you reset your system clock to January last
768	time, it will still time out after (roughly) and hour. "Roughly" because	1300	year, it will still time out after (roughly) one hour. "Roughly" because
769	detecting time jumps is hard, and some inaccuracies are unavoidable (the	1301	detecting time jumps is hard, and some inaccuracies are unavoidable (the
770	monotonic clock option helps a lot here).	1302	monotonic clock option helps a lot here).
		1303
		1304	The callback is guaranteed to be invoked only I<after> its timeout has
		1305	passed, but if multiple timers become ready during the same loop iteration
		1306	then order of execution is undefined.
		1307
		1308	=head3 Be smart about timeouts
		1309
		1310	Many real-world problems involve some kind of timeout, usually for error
		1311	recovery. A typical example is an HTTP request - if the other side hangs,
		1312	you want to raise some error after a while.
		1313
		1314	What follows are some ways to handle this problem, from obvious and
		1315	inefficient to smart and efficient.
		1316
		1317	In the following, a 60 second activity timeout is assumed - a timeout that
		1318	gets reset to 60 seconds each time there is activity (e.g. each time some
		1319	data or other life sign was received).
		1320
		1321	=over 4
		1322
		1323	=item 1. Use a timer and stop, reinitialise and start it on activity.
		1324
		1325	This is the most obvious, but not the most simple way: In the beginning,
		1326	start the watcher:
		1327
		1328	ev_timer_init (timer, callback, 60., 0.);
		1329	ev_timer_start (loop, timer);
		1330
		1331	Then, each time there is some activity, C<ev_timer_stop> it, initialise it
		1332	and start it again:
		1333
		1334	ev_timer_stop (loop, timer);
		1335	ev_timer_set (timer, 60., 0.);
		1336	ev_timer_start (loop, timer);
		1337
		1338	This is relatively simple to implement, but means that each time there is
		1339	some activity, libev will first have to remove the timer from its internal
		1340	data structure and then add it again. Libev tries to be fast, but it's
		1341	still not a constant-time operation.
		1342
		1343	=item 2. Use a timer and re-start it with C<ev_timer_again> inactivity.
		1344
		1345	This is the easiest way, and involves using C<ev_timer_again> instead of
		1346	C<ev_timer_start>.
		1347
		1348	To implement this, configure an C<ev_timer> with a C<repeat> value
		1349	of C<60> and then call C<ev_timer_again> at start and each time you
		1350	successfully read or write some data. If you go into an idle state where
		1351	you do not expect data to travel on the socket, you can C<ev_timer_stop>
		1352	the timer, and C<ev_timer_again> will automatically restart it if need be.
		1353
		1354	That means you can ignore both the C<ev_timer_start> function and the
		1355	C<after> argument to C<ev_timer_set>, and only ever use the C<repeat>
		1356	member and C<ev_timer_again>.
		1357
		1358	At start:
		1359
		1360	ev_timer_init (timer, callback);
		1361	timer->repeat = 60.;
		1362	ev_timer_again (loop, timer);
		1363
		1364	Each time there is some activity:
		1365
		1366	ev_timer_again (loop, timer);
		1367
		1368	It is even possible to change the time-out on the fly, regardless of
		1369	whether the watcher is active or not:
		1370
		1371	timer->repeat = 30.;
		1372	ev_timer_again (loop, timer);
		1373
		1374	This is slightly more efficient then stopping/starting the timer each time
		1375	you want to modify its timeout value, as libev does not have to completely
		1376	remove and re-insert the timer from/into its internal data structure.
		1377
		1378	It is, however, even simpler than the "obvious" way to do it.
		1379
		1380	=item 3. Let the timer time out, but then re-arm it as required.
		1381
		1382	This method is more tricky, but usually most efficient: Most timeouts are
		1383	relatively long compared to the intervals between other activity - in
		1384	our example, within 60 seconds, there are usually many I/O events with
		1385	associated activity resets.
		1386
		1387	In this case, it would be more efficient to leave the C<ev_timer> alone,
		1388	but remember the time of last activity, and check for a real timeout only
		1389	within the callback:
		1390
		1391	ev_tstamp last_activity; // time of last activity
		1392
		1393	static void
		1394	callback (EV_P_ ev_timer *w, int revents)
		1395	{
		1396	ev_tstamp now = ev_now (EV_A);
		1397	ev_tstamp timeout = last_activity + 60.;
		1398
		1399	// if last_activity + 60. is older than now, we did time out
		1400	if (timeout < now)
		1401	{
		1402	// timeout occured, take action
		1403	}
		1404	else
		1405	{
		1406	// callback was invoked, but there was some activity, re-arm
		1407	// the watcher to fire in last_activity + 60, which is
		1408	// guaranteed to be in the future, so "again" is positive:
		1409	w->again = timeout - now;
		1410	ev_timer_again (EV_A_ w);
		1411	}
		1412	}
		1413
		1414	To summarise the callback: first calculate the real timeout (defined
		1415	as "60 seconds after the last activity"), then check if that time has
		1416	been reached, which means something I<did>, in fact, time out. Otherwise
		1417	the callback was invoked too early (C<timeout> is in the future), so
		1418	re-schedule the timer to fire at that future time, to see if maybe we have
		1419	a timeout then.
		1420
		1421	Note how C<ev_timer_again> is used, taking advantage of the
		1422	C<ev_timer_again> optimisation when the timer is already running.
		1423
		1424	This scheme causes more callback invocations (about one every 60 seconds
		1425	minus half the average time between activity), but virtually no calls to
		1426	libev to change the timeout.
		1427
		1428	To start the timer, simply initialise the watcher and set C<last_activity>
		1429	to the current time (meaning we just have some activity :), then call the
		1430	callback, which will "do the right thing" and start the timer:
		1431
		1432	ev_timer_init (timer, callback);
		1433	last_activity = ev_now (loop);
		1434	callback (loop, timer, EV_TIMEOUT);
		1435
		1436	And when there is some activity, simply store the current time in
		1437	C<last_activity>, no libev calls at all:
		1438
		1439	last_actiivty = ev_now (loop);
		1440
		1441	This technique is slightly more complex, but in most cases where the
		1442	time-out is unlikely to be triggered, much more efficient.
		1443
		1444	Changing the timeout is trivial as well (if it isn't hard-coded in the
		1445	callback :) - just change the timeout and invoke the callback, which will
		1446	fix things for you.
		1447
		1448	=item 4. Wee, just use a double-linked list for your timeouts.
		1449
		1450	If there is not one request, but many thousands (millions...), all
		1451	employing some kind of timeout with the same timeout value, then one can
		1452	do even better:
		1453
		1454	When starting the timeout, calculate the timeout value and put the timeout
		1455	at the I<end> of the list.
		1456
		1457	Then use an C<ev_timer> to fire when the timeout at the I<beginning> of
		1458	the list is expected to fire (for example, using the technique #3).
		1459
		1460	When there is some activity, remove the timer from the list, recalculate
		1461	the timeout, append it to the end of the list again, and make sure to
		1462	update the C<ev_timer> if it was taken from the beginning of the list.
		1463
		1464	This way, one can manage an unlimited number of timeouts in O(1) time for
		1465	starting, stopping and updating the timers, at the expense of a major
		1466	complication, and having to use a constant timeout. The constant timeout
		1467	ensures that the list stays sorted.
		1468
		1469	=back
		1470
		1471	So which method the best?
		1472
		1473	Method #2 is a simple no-brain-required solution that is adequate in most
		1474	situations. Method #3 requires a bit more thinking, but handles many cases
		1475	better, and isn't very complicated either. In most case, choosing either
		1476	one is fine, with #3 being better in typical situations.
		1477
		1478	Method #1 is almost always a bad idea, and buys you nothing. Method #4 is
		1479	rather complicated, but extremely efficient, something that really pays
		1480	off after the first million or so of active timers, i.e. it's usually
		1481	overkill :)
		1482
		1483	=head3 The special problem of time updates
		1484
		1485	Establishing the current time is a costly operation (it usually takes at
		1486	least two system calls): EV therefore updates its idea of the current
		1487	time only before and after C<ev_loop> collects new events, which causes a
		1488	growing difference between C<ev_now ()> and C<ev_time ()> when handling
		1489	lots of events in one iteration.
771		1490
772	The relative timeouts are calculated relative to the C<ev_now ()>	1491	The relative timeouts are calculated relative to the C<ev_now ()>
773	time. This is usually the right thing as this timestamp refers to the time	1492	time. This is usually the right thing as this timestamp refers to the time
774	of the event triggering whatever timeout you are modifying/starting. If	1493	of the event triggering whatever timeout you are modifying/starting. If
775	you suspect event processing to be delayed and you I<need> to base the timeout	1494	you suspect event processing to be delayed and you I<need> to base the
776	on the current time, use something like this to adjust for this:	1495	timeout on the current time, use something like this to adjust for this:
777		1496
778	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);	1497	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);
779		1498
780	The callback is guarenteed to be invoked only when its timeout has passed,	1499	If the event loop is suspended for a long time, you can also force an
781	but if multiple timers become ready during the same loop iteration then	1500	update of the time returned by C<ev_now ()> by calling C<ev_now_update
782	order of execution is undefined.	1501	()>.
		1502
		1503	=head3 Watcher-Specific Functions and Data Members
783		1504
784	=over 4	1505	=over 4
785		1506
786	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)	1507	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)
787		1508
788	=item ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)	1509	=item ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)
789		1510
790	Configure the timer to trigger after C<after> seconds. If C<repeat> is	1511	Configure the timer to trigger after C<after> seconds. If C<repeat>
791	C<0.>, then it will automatically be stopped. If it is positive, then the	1512	is C<0.>, then it will automatically be stopped once the timeout is
792	timer will automatically be configured to trigger again C<repeat> seconds	1513	reached. If it is positive, then the timer will automatically be
793	later, again, and again, until stopped manually.	1514	configured to trigger again C<repeat> seconds later, again, and again,
		1515	until stopped manually.
794		1516
795	The timer itself will do a best-effort at avoiding drift, that is, if you	1517	The timer itself will do a best-effort at avoiding drift, that is, if
796	configure a timer to trigger every 10 seconds, then it will trigger at	1518	you configure a timer to trigger every 10 seconds, then it will normally
797	exactly 10 second intervals. If, however, your program cannot keep up with	1519	trigger at exactly 10 second intervals. If, however, your program cannot
798	the timer (because it takes longer than those 10 seconds to do stuff) the	1520	keep up with the timer (because it takes longer than those 10 seconds to
799	timer will not fire more than once per event loop iteration.	1521	do stuff) the timer will not fire more than once per event loop iteration.
800		1522
801	=item ev_timer_again (loop)	1523	=item ev_timer_again (loop, ev_timer *)
802		1524
803	This will act as if the timer timed out and restart it again if it is	1525	This will act as if the timer timed out and restart it again if it is
804	repeating. The exact semantics are:	1526	repeating. The exact semantics are:
805		1527
		1528	If the timer is pending, its pending status is cleared.
		1529
806	If the timer is started but nonrepeating, stop it.	1530	If the timer is started but non-repeating, stop it (as if it timed out).
807		1531
808	If the timer is repeating, either start it if necessary (with the repeat	1532	If the timer is repeating, either start it if necessary (with the
809	value), or reset the running timer to the repeat value.	1533	C<repeat> value), or reset the running timer to the C<repeat> value.
810		1534
811	This sounds a bit complicated, but here is a useful and typical	1535	This sounds a bit complicated, see "Be smart about timeouts", above, for a
812	example: Imagine you have a tcp connection and you want a so-called idle	1536	usage example.
813	timeout, that is, you want to be called when there have been, say, 60	1537
814	seconds of inactivity on the socket. The easiest way to do this is to	1538	=item ev_tstamp repeat [read-write]
815	configure an C<ev_timer> with after=repeat=60 and calling ev_timer_again each	1539
816	time you successfully read or write some data. If you go into an idle	1540	The current C<repeat> value. Will be used each time the watcher times out
817	state where you do not expect data to travel on the socket, you can stop	1541	or C<ev_timer_again> is called, and determines the next timeout (if any),
818	the timer, and again will automatically restart it if need be.	1542	which is also when any modifications are taken into account.
819		1543
820	=back	1544	=back
821		1545
		1546	=head3 Examples
		1547
822	Example: create a timer that fires after 60 seconds.	1548	Example: Create a timer that fires after 60 seconds.
823		1549
824	static void	1550	static void
825	one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)	1551	one_minute_cb (struct ev_loop *loop, ev_timer *w, int revents)
826	{	1552	{
827	.. one minute over, w is actually stopped right here	1553	.. one minute over, w is actually stopped right here
828	}	1554	}
829		1555
830	struct ev_timer mytimer;	1556	ev_timer mytimer;
831	ev_timer_init (&mytimer, one_minute_cb, 60., 0.);	1557	ev_timer_init (&mytimer, one_minute_cb, 60., 0.);
832	ev_timer_start (loop, &mytimer);	1558	ev_timer_start (loop, &mytimer);
833		1559
834	Example: create a timeout timer that times out after 10 seconds of	1560	Example: Create a timeout timer that times out after 10 seconds of
835	inactivity.	1561	inactivity.
836		1562
837	static void	1563	static void
838	timeout_cb (struct ev_loop *loop, struct ev_timer *w, int revents)	1564	timeout_cb (struct ev_loop *loop, ev_timer *w, int revents)
839	{	1565	{
840	.. ten seconds without any activity	1566	.. ten seconds without any activity
841	}	1567	}
842		1568
843	struct ev_timer mytimer;	1569	ev_timer mytimer;
844	ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */	1570	ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */
845	ev_timer_again (&mytimer); /* start timer */	1571	ev_timer_again (&mytimer); /* start timer */
846	ev_loop (loop, 0);	1572	ev_loop (loop, 0);
847		1573
848	// and in some piece of code that gets executed on any "activity":	1574	// and in some piece of code that gets executed on any "activity":
849	// reset the timeout to start ticking again at 10 seconds	1575	// reset the timeout to start ticking again at 10 seconds
850	ev_timer_again (&mytimer);	1576	ev_timer_again (&mytimer);
851		1577
852		1578
853	=head2 C<ev_periodic> - to cron or not to cron	1579	=head2 C<ev_periodic> - to cron or not to cron?
854		1580
855	Periodic watchers are also timers of a kind, but they are very versatile	1581	Periodic watchers are also timers of a kind, but they are very versatile
856	(and unfortunately a bit complex).	1582	(and unfortunately a bit complex).
857		1583
858	Unlike C<ev_timer>'s, they are not based on real time (or relative time)	1584	Unlike C<ev_timer>'s, they are not based on real time (or relative time)
859	but on wallclock time (absolute time). You can tell a periodic watcher	1585	but on wall clock time (absolute time). You can tell a periodic watcher
860	to trigger "at" some specific point in time. For example, if you tell a	1586	to trigger after some specific point in time. For example, if you tell a
861	periodic watcher to trigger in 10 seconds (by specifiying e.g. C<ev_now ()	1587	periodic watcher to trigger in 10 seconds (by specifying e.g. C<ev_now ()
862	+ 10.>) and then reset your system clock to the last year, then it will	1588	+ 10.>, that is, an absolute time not a delay) and then reset your system
		1589	clock to January of the previous year, then it will take more than year
863	take a year to trigger the event (unlike an C<ev_timer>, which would trigger	1590	to trigger the event (unlike an C<ev_timer>, which would still trigger
864	roughly 10 seconds later and of course not if you reset your system time	1591	roughly 10 seconds later as it uses a relative timeout).
865	again).
866		1592
867	They can also be used to implement vastly more complex timers, such as	1593	C<ev_periodic>s can also be used to implement vastly more complex timers,
868	triggering an event on eahc midnight, local time.	1594	such as triggering an event on each "midnight, local time", or other
		1595	complicated rules.
869		1596
870	As with timers, the callback is guarenteed to be invoked only when the	1597	As with timers, the callback is guaranteed to be invoked only when the
871	time (C<at>) has been passed, but if multiple periodic timers become ready	1598	time (C<at>) has passed, but if multiple periodic timers become ready
872	during the same loop iteration then order of execution is undefined.	1599	during the same loop iteration, then order of execution is undefined.
		1600
		1601	=head3 Watcher-Specific Functions and Data Members
873		1602
874	=over 4	1603	=over 4
875		1604
876	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)	1605	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)
877		1606
878	=item ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)	1607	=item ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)
879		1608
880	Lots of arguments, lets sort it out... There are basically three modes of	1609	Lots of arguments, lets sort it out... There are basically three modes of
881	operation, and we will explain them from simplest to complex:	1610	operation, and we will explain them from simplest to most complex:
882		1611
883	=over 4	1612	=over 4
884		1613
885	=item * absolute timer (interval = reschedule_cb = 0)	1614	=item * absolute timer (at = time, interval = reschedule_cb = 0)
886		1615
887	In this configuration the watcher triggers an event at the wallclock time	1616	In this configuration the watcher triggers an event after the wall clock
888	C<at> and doesn't repeat. It will not adjust when a time jump occurs,	1617	time C<at> has passed. It will not repeat and will not adjust when a time
889	that is, if it is to be run at January 1st 2011 then it will run when the	1618	jump occurs, that is, if it is to be run at January 1st 2011 then it will
890	system time reaches or surpasses this time.	1619	only run when the system clock reaches or surpasses this time.
891		1620
892	=item * non-repeating interval timer (interval > 0, reschedule_cb = 0)	1621	=item * repeating interval timer (at = offset, interval > 0, reschedule_cb = 0)
893		1622
894	In this mode the watcher will always be scheduled to time out at the next	1623	In this mode the watcher will always be scheduled to time out at the next
895	C<at + N * interval> time (for some integer N) and then repeat, regardless	1624	C<at + N * interval> time (for some integer N, which can also be negative)
896	of any time jumps.	1625	and then repeat, regardless of any time jumps.
897		1626
898	This can be used to create timers that do not drift with respect to system	1627	This can be used to create timers that do not drift with respect to the
899	time:	1628	system clock, for example, here is a C<ev_periodic> that triggers each
		1629	hour, on the hour:
900		1630
901	ev_periodic_set (&periodic, 0., 3600., 0);	1631	ev_periodic_set (&periodic, 0., 3600., 0);
902		1632
903	This doesn't mean there will always be 3600 seconds in between triggers,	1633	This doesn't mean there will always be 3600 seconds in between triggers,
904	but only that the the callback will be called when the system time shows a	1634	but only that the callback will be called when the system time shows a
905	full hour (UTC), or more correctly, when the system time is evenly divisible	1635	full hour (UTC), or more correctly, when the system time is evenly divisible
906	by 3600.	1636	by 3600.
907		1637
908	Another way to think about it (for the mathematically inclined) is that	1638	Another way to think about it (for the mathematically inclined) is that
909	C<ev_periodic> will try to run the callback in this mode at the next possible	1639	C<ev_periodic> will try to run the callback in this mode at the next possible
910	time where C<time = at (mod interval)>, regardless of any time jumps.	1640	time where C<time = at (mod interval)>, regardless of any time jumps.
911		1641
		1642	For numerical stability it is preferable that the C<at> value is near
		1643	C<ev_now ()> (the current time), but there is no range requirement for
		1644	this value, and in fact is often specified as zero.
		1645
		1646	Note also that there is an upper limit to how often a timer can fire (CPU
		1647	speed for example), so if C<interval> is very small then timing stability
		1648	will of course deteriorate. Libev itself tries to be exact to be about one
		1649	millisecond (if the OS supports it and the machine is fast enough).
		1650
912	=item * manual reschedule mode (reschedule_cb = callback)	1651	=item * manual reschedule mode (at and interval ignored, reschedule_cb = callback)
913		1652
914	In this mode the values for C<interval> and C<at> are both being	1653	In this mode the values for C<interval> and C<at> are both being
915	ignored. Instead, each time the periodic watcher gets scheduled, the	1654	ignored. Instead, each time the periodic watcher gets scheduled, the
916	reschedule callback will be called with the watcher as first, and the	1655	reschedule callback will be called with the watcher as first, and the
917	current time as second argument.	1656	current time as second argument.
918		1657
919	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher,	1658	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher,
920	ever, or make any event loop modifications>. If you need to stop it,	1659	ever, or make ANY event loop modifications whatsoever>.
921	return C<now + 1e30> (or so, fudge fudge) and stop it afterwards (e.g. by
922	starting a prepare watcher).
923		1660
		1661	If you need to stop it, return C<now + 1e30> (or so, fudge fudge) and stop
		1662	it afterwards (e.g. by starting an C<ev_prepare> watcher, which is the
		1663	only event loop modification you are allowed to do).
		1664
924	Its prototype is C<ev_tstamp (*reschedule_cb)(struct ev_periodic *w,	1665	The callback prototype is C<ev_tstamp (*reschedule_cb)(ev_periodic
925	ev_tstamp now)>, e.g.:	1666	*w, ev_tstamp now)>, e.g.:
926		1667
		1668	static ev_tstamp
927	static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)	1669	my_rescheduler (ev_periodic *w, ev_tstamp now)
928	{	1670	{
929	return now + 60.;	1671	return now + 60.;
930	}	1672	}
931		1673
932	It must return the next time to trigger, based on the passed time value	1674	It must return the next time to trigger, based on the passed time value
933	(that is, the lowest time value larger than to the second argument). It	1675	(that is, the lowest time value larger than to the second argument). It
934	will usually be called just before the callback will be triggered, but	1676	will usually be called just before the callback will be triggered, but
935	might be called at other times, too.	1677	might be called at other times, too.
936		1678
937	NOTE: I<< This callback must always return a time that is later than the	1679	NOTE: I<< This callback must always return a time that is higher than or
938	passed C<now> value >>. Not even C<now> itself will do, it I<must> be larger.	1680	equal to the passed C<now> value >>.
939		1681
940	This can be used to create very complex timers, such as a timer that	1682	This can be used to create very complex timers, such as a timer that
941	triggers on each midnight, local time. To do this, you would calculate the	1683	triggers on "next midnight, local time". To do this, you would calculate the
942	next midnight after C<now> and return the timestamp value for this. How	1684	next midnight after C<now> and return the timestamp value for this. How
943	you do this is, again, up to you (but it is not trivial, which is the main	1685	you do this is, again, up to you (but it is not trivial, which is the main
944	reason I omitted it as an example).	1686	reason I omitted it as an example).
945		1687
946	=back	1688	=back
…		…
950	Simply stops and restarts the periodic watcher again. This is only useful	1692	Simply stops and restarts the periodic watcher again. This is only useful
951	when you changed some parameters or the reschedule callback would return	1693	when you changed some parameters or the reschedule callback would return
952	a different time than the last time it was called (e.g. in a crond like	1694	a different time than the last time it was called (e.g. in a crond like
953	program when the crontabs have changed).	1695	program when the crontabs have changed).
954		1696
		1697	=item ev_tstamp ev_periodic_at (ev_periodic *)
		1698
		1699	When active, returns the absolute time that the watcher is supposed to
		1700	trigger next.
		1701
		1702	=item ev_tstamp offset [read-write]
		1703
		1704	When repeating, this contains the offset value, otherwise this is the
		1705	absolute point in time (the C<at> value passed to C<ev_periodic_set>).
		1706
		1707	Can be modified any time, but changes only take effect when the periodic
		1708	timer fires or C<ev_periodic_again> is being called.
		1709
		1710	=item ev_tstamp interval [read-write]
		1711
		1712	The current interval value. Can be modified any time, but changes only
		1713	take effect when the periodic timer fires or C<ev_periodic_again> is being
		1714	called.
		1715
		1716	=item ev_tstamp (*reschedule_cb)(ev_periodic *w, ev_tstamp now) [read-write]
		1717
		1718	The current reschedule callback, or C<0>, if this functionality is
		1719	switched off. Can be changed any time, but changes only take effect when
		1720	the periodic timer fires or C<ev_periodic_again> is being called.
		1721
955	=back	1722	=back
956		1723
		1724	=head3 Examples
		1725
957	Example: call a callback every hour, or, more precisely, whenever the	1726	Example: Call a callback every hour, or, more precisely, whenever the
958	system clock is divisible by 3600. The callback invocation times have	1727	system time is divisible by 3600. The callback invocation times have
959	potentially a lot of jittering, but good long-term stability.	1728	potentially a lot of jitter, but good long-term stability.
960		1729
961	static void	1730	static void
962	clock_cb (struct ev_loop *loop, struct ev_io *w, int revents)	1731	clock_cb (struct ev_loop *loop, ev_io *w, int revents)
963	{	1732	{
964	... its now a full hour (UTC, or TAI or whatever your clock follows)	1733	... its now a full hour (UTC, or TAI or whatever your clock follows)
965	}	1734	}
966		1735
967	struct ev_periodic hourly_tick;	1736	ev_periodic hourly_tick;
968	ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);	1737	ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);
969	ev_periodic_start (loop, &hourly_tick);	1738	ev_periodic_start (loop, &hourly_tick);
970		1739
971	Example: the same as above, but use a reschedule callback to do it:	1740	Example: The same as above, but use a reschedule callback to do it:
972		1741
973	#include <math.h>	1742	#include <math.h>
974		1743
975	static ev_tstamp	1744	static ev_tstamp
976	my_scheduler_cb (struct ev_periodic *w, ev_tstamp now)	1745	my_scheduler_cb (ev_periodic *w, ev_tstamp now)
977	{	1746	{
978	return fmod (now, 3600.) + 3600.;	1747	return now + (3600. - fmod (now, 3600.));
979	}	1748	}
980		1749
981	ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);	1750	ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
982		1751
983	Example: call a callback every hour, starting now:	1752	Example: Call a callback every hour, starting now:
984		1753
985	struct ev_periodic hourly_tick;	1754	ev_periodic hourly_tick;
986	ev_periodic_init (&hourly_tick, clock_cb,	1755	ev_periodic_init (&hourly_tick, clock_cb,
987	fmod (ev_now (loop), 3600.), 3600., 0);	1756	fmod (ev_now (loop), 3600.), 3600., 0);
988	ev_periodic_start (loop, &hourly_tick);	1757	ev_periodic_start (loop, &hourly_tick);
989		1758
990		1759
991	=head2 C<ev_signal> - signal me when a signal gets signalled	1760	=head2 C<ev_signal> - signal me when a signal gets signalled!
992		1761
993	Signal watchers will trigger an event when the process receives a specific	1762	Signal watchers will trigger an event when the process receives a specific
994	signal one or more times. Even though signals are very asynchronous, libev	1763	signal one or more times. Even though signals are very asynchronous, libev
995	will try it's best to deliver signals synchronously, i.e. as part of the	1764	will try it's best to deliver signals synchronously, i.e. as part of the
996	normal event processing, like any other event.	1765	normal event processing, like any other event.
997		1766
		1767	If you want signals asynchronously, just use C<sigaction> as you would
		1768	do without libev and forget about sharing the signal. You can even use
		1769	C<ev_async> from a signal handler to synchronously wake up an event loop.
		1770
998	You can configure as many watchers as you like per signal. Only when the	1771	You can configure as many watchers as you like per signal. Only when the
999	first watcher gets started will libev actually register a signal watcher	1772	first watcher gets started will libev actually register a signal handler
1000	with the kernel (thus it coexists with your own signal handlers as long	1773	with the kernel (thus it coexists with your own signal handlers as long as
1001	as you don't register any with libev). Similarly, when the last signal	1774	you don't register any with libev for the same signal). Similarly, when
1002	watcher for a signal is stopped libev will reset the signal handler to	1775	the last signal watcher for a signal is stopped, libev will reset the
1003	SIG_DFL (regardless of what it was set to before).	1776	signal handler to SIG_DFL (regardless of what it was set to before).
		1777
		1778	If possible and supported, libev will install its handlers with
		1779	C<SA_RESTART> behaviour enabled, so system calls should not be unduly
		1780	interrupted. If you have a problem with system calls getting interrupted by
		1781	signals you can block all signals in an C<ev_check> watcher and unblock
		1782	them in an C<ev_prepare> watcher.
		1783
		1784	=head3 Watcher-Specific Functions and Data Members
1004		1785
1005	=over 4	1786	=over 4
1006		1787
1007	=item ev_signal_init (ev_signal *, callback, int signum)	1788	=item ev_signal_init (ev_signal *, callback, int signum)
1008		1789
1009	=item ev_signal_set (ev_signal *, int signum)	1790	=item ev_signal_set (ev_signal *, int signum)
1010		1791
1011	Configures the watcher to trigger on the given signal number (usually one	1792	Configures the watcher to trigger on the given signal number (usually one
1012	of the C<SIGxxx> constants).	1793	of the C<SIGxxx> constants).
1013		1794
		1795	=item int signum [read-only]
		1796
		1797	The signal the watcher watches out for.
		1798
1014	=back	1799	=back
1015		1800
		1801	=head3 Examples
1016		1802
		1803	Example: Try to exit cleanly on SIGINT.
		1804
		1805	static void
		1806	sigint_cb (struct ev_loop *loop, ev_signal *w, int revents)
		1807	{
		1808	ev_unloop (loop, EVUNLOOP_ALL);
		1809	}
		1810
		1811	ev_signal signal_watcher;
		1812	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
		1813	ev_signal_start (loop, &signal_watcher);
		1814
		1815
1017	=head2 C<ev_child> - wait for pid status changes	1816	=head2 C<ev_child> - watch out for process status changes
1018		1817
1019	Child watchers trigger when your process receives a SIGCHLD in response to	1818	Child watchers trigger when your process receives a SIGCHLD in response to
1020	some child status changes (most typically when a child of yours dies).	1819	some child status changes (most typically when a child of yours dies or
		1820	exits). It is permissible to install a child watcher I<after> the child
		1821	has been forked (which implies it might have already exited), as long
		1822	as the event loop isn't entered (or is continued from a watcher), i.e.,
		1823	forking and then immediately registering a watcher for the child is fine,
		1824	but forking and registering a watcher a few event loop iterations later is
		1825	not.
		1826
		1827	Only the default event loop is capable of handling signals, and therefore
		1828	you can only register child watchers in the default event loop.
		1829
		1830	=head3 Process Interaction
		1831
		1832	Libev grabs C<SIGCHLD> as soon as the default event loop is
		1833	initialised. This is necessary to guarantee proper behaviour even if
		1834	the first child watcher is started after the child exits. The occurrence
		1835	of C<SIGCHLD> is recorded asynchronously, but child reaping is done
		1836	synchronously as part of the event loop processing. Libev always reaps all
		1837	children, even ones not watched.
		1838
		1839	=head3 Overriding the Built-In Processing
		1840
		1841	Libev offers no special support for overriding the built-in child
		1842	processing, but if your application collides with libev's default child
		1843	handler, you can override it easily by installing your own handler for
		1844	C<SIGCHLD> after initialising the default loop, and making sure the
		1845	default loop never gets destroyed. You are encouraged, however, to use an
		1846	event-based approach to child reaping and thus use libev's support for
		1847	that, so other libev users can use C<ev_child> watchers freely.
		1848
		1849	=head3 Stopping the Child Watcher
		1850
		1851	Currently, the child watcher never gets stopped, even when the
		1852	child terminates, so normally one needs to stop the watcher in the
		1853	callback. Future versions of libev might stop the watcher automatically
		1854	when a child exit is detected.
		1855
		1856	=head3 Watcher-Specific Functions and Data Members
1021		1857
1022	=over 4	1858	=over 4
1023		1859
1024	=item ev_child_init (ev_child *, callback, int pid)	1860	=item ev_child_init (ev_child *, callback, int pid, int trace)
1025		1861
1026	=item ev_child_set (ev_child *, int pid)	1862	=item ev_child_set (ev_child *, int pid, int trace)
1027		1863
1028	Configures the watcher to wait for status changes of process C<pid> (or	1864	Configures the watcher to wait for status changes of process C<pid> (or
1029	I<any> process if C<pid> is specified as C<0>). The callback can look	1865	I<any> process if C<pid> is specified as C<0>). The callback can look
1030	at the C<rstatus> member of the C<ev_child> watcher structure to see	1866	at the C<rstatus> member of the C<ev_child> watcher structure to see
1031	the status word (use the macros from C<sys/wait.h> and see your systems	1867	the status word (use the macros from C<sys/wait.h> and see your systems
1032	C<waitpid> documentation). The C<rpid> member contains the pid of the	1868	C<waitpid> documentation). The C<rpid> member contains the pid of the
1033	process causing the status change.	1869	process causing the status change. C<trace> must be either C<0> (only
		1870	activate the watcher when the process terminates) or C<1> (additionally
		1871	activate the watcher when the process is stopped or continued).
		1872
		1873	=item int pid [read-only]
		1874
		1875	The process id this watcher watches out for, or C<0>, meaning any process id.
		1876
		1877	=item int rpid [read-write]
		1878
		1879	The process id that detected a status change.
		1880
		1881	=item int rstatus [read-write]
		1882
		1883	The process exit/trace status caused by C<rpid> (see your systems
		1884	C<waitpid> and C<sys/wait.h> documentation for details).
1034		1885
1035	=back	1886	=back
1036		1887
1037	Example: try to exit cleanly on SIGINT and SIGTERM.	1888	=head3 Examples
1038		1889
		1890	Example: C<fork()> a new process and install a child handler to wait for
		1891	its completion.
		1892
		1893	ev_child cw;
		1894
1039	static void	1895	static void
1040	sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)	1896	child_cb (EV_P_ ev_child *w, int revents)
1041	{	1897	{
1042	ev_unloop (loop, EVUNLOOP_ALL);	1898	ev_child_stop (EV_A_ w);
		1899	printf ("process %d exited with status %x\n", w->rpid, w->rstatus);
1043	}	1900	}
1044		1901
1045	struct ev_signal signal_watcher;	1902	pid_t pid = fork ();
1046	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
1047	ev_signal_start (loop, &sigint_cb);
1048		1903
		1904	if (pid < 0)
		1905	// error
		1906	else if (pid == 0)
		1907	{
		1908	// the forked child executes here
		1909	exit (1);
		1910	}
		1911	else
		1912	{
		1913	ev_child_init (&cw, child_cb, pid, 0);
		1914	ev_child_start (EV_DEFAULT_ &cw);
		1915	}
1049		1916
		1917
		1918	=head2 C<ev_stat> - did the file attributes just change?
		1919
		1920	This watches a file system path for attribute changes. That is, it calls
		1921	C<stat> regularly (or when the OS says it changed) and sees if it changed
		1922	compared to the last time, invoking the callback if it did.
		1923
		1924	The path does not need to exist: changing from "path exists" to "path does
		1925	not exist" is a status change like any other. The condition "path does
		1926	not exist" is signified by the C<st_nlink> field being zero (which is
		1927	otherwise always forced to be at least one) and all the other fields of
		1928	the stat buffer having unspecified contents.
		1929
		1930	The path I<should> be absolute and I<must not> end in a slash. If it is
		1931	relative and your working directory changes, the behaviour is undefined.
		1932
		1933	Since there is no standard kernel interface to do this, the portable
		1934	implementation simply calls C<stat (2)> regularly on the path to see if
		1935	it changed somehow. You can specify a recommended polling interval for
		1936	this case. If you specify a polling interval of C<0> (highly recommended!)
		1937	then a I<suitable, unspecified default> value will be used (which
		1938	you can expect to be around five seconds, although this might change
		1939	dynamically). Libev will also impose a minimum interval which is currently
		1940	around C<0.1>, but thats usually overkill.
		1941
		1942	This watcher type is not meant for massive numbers of stat watchers,
		1943	as even with OS-supported change notifications, this can be
		1944	resource-intensive.
		1945
		1946	At the time of this writing, the only OS-specific interface implemented
		1947	is the Linux inotify interface (implementing kqueue support is left as
		1948	an exercise for the reader. Note, however, that the author sees no way
		1949	of implementing C<ev_stat> semantics with kqueue).
		1950
		1951	=head3 ABI Issues (Largefile Support)
		1952
		1953	Libev by default (unless the user overrides this) uses the default
		1954	compilation environment, which means that on systems with large file
		1955	support disabled by default, you get the 32 bit version of the stat
		1956	structure. When using the library from programs that change the ABI to
		1957	use 64 bit file offsets the programs will fail. In that case you have to
		1958	compile libev with the same flags to get binary compatibility. This is
		1959	obviously the case with any flags that change the ABI, but the problem is
		1960	most noticeably disabled with ev_stat and large file support.
		1961
		1962	The solution for this is to lobby your distribution maker to make large
		1963	file interfaces available by default (as e.g. FreeBSD does) and not
		1964	optional. Libev cannot simply switch on large file support because it has
		1965	to exchange stat structures with application programs compiled using the
		1966	default compilation environment.
		1967
		1968	=head3 Inotify and Kqueue
		1969
		1970	When C<inotify (7)> support has been compiled into libev (generally
		1971	only available with Linux 2.6.25 or above due to bugs in earlier
		1972	implementations) and present at runtime, it will be used to speed up
		1973	change detection where possible. The inotify descriptor will be created
		1974	lazily when the first C<ev_stat> watcher is being started.
		1975
		1976	Inotify presence does not change the semantics of C<ev_stat> watchers
		1977	except that changes might be detected earlier, and in some cases, to avoid
		1978	making regular C<stat> calls. Even in the presence of inotify support
		1979	there are many cases where libev has to resort to regular C<stat> polling,
		1980	but as long as the path exists, libev usually gets away without polling.
		1981
		1982	There is no support for kqueue, as apparently it cannot be used to
		1983	implement this functionality, due to the requirement of having a file
		1984	descriptor open on the object at all times, and detecting renames, unlinks
		1985	etc. is difficult.
		1986
		1987	=head3 The special problem of stat time resolution
		1988
		1989	The C<stat ()> system call only supports full-second resolution portably, and
		1990	even on systems where the resolution is higher, most file systems still
		1991	only support whole seconds.
		1992
		1993	That means that, if the time is the only thing that changes, you can
		1994	easily miss updates: on the first update, C<ev_stat> detects a change and
		1995	calls your callback, which does something. When there is another update
		1996	within the same second, C<ev_stat> will be unable to detect unless the
		1997	stat data does change in other ways (e.g. file size).
		1998
		1999	The solution to this is to delay acting on a change for slightly more
		2000	than a second (or till slightly after the next full second boundary), using
		2001	a roughly one-second-delay C<ev_timer> (e.g. C<ev_timer_set (w, 0., 1.02);
		2002	ev_timer_again (loop, w)>).
		2003
		2004	The C<.02> offset is added to work around small timing inconsistencies
		2005	of some operating systems (where the second counter of the current time
		2006	might be be delayed. One such system is the Linux kernel, where a call to
		2007	C<gettimeofday> might return a timestamp with a full second later than
		2008	a subsequent C<time> call - if the equivalent of C<time ()> is used to
		2009	update file times then there will be a small window where the kernel uses
		2010	the previous second to update file times but libev might already execute
		2011	the timer callback).
		2012
		2013	=head3 Watcher-Specific Functions and Data Members
		2014
		2015	=over 4
		2016
		2017	=item ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)
		2018
		2019	=item ev_stat_set (ev_stat *, const char *path, ev_tstamp interval)
		2020
		2021	Configures the watcher to wait for status changes of the given
		2022	C<path>. The C<interval> is a hint on how quickly a change is expected to
		2023	be detected and should normally be specified as C<0> to let libev choose
		2024	a suitable value. The memory pointed to by C<path> must point to the same
		2025	path for as long as the watcher is active.
		2026
		2027	The callback will receive an C<EV_STAT> event when a change was detected,
		2028	relative to the attributes at the time the watcher was started (or the
		2029	last change was detected).
		2030
		2031	=item ev_stat_stat (loop, ev_stat *)
		2032
		2033	Updates the stat buffer immediately with new values. If you change the
		2034	watched path in your callback, you could call this function to avoid
		2035	detecting this change (while introducing a race condition if you are not
		2036	the only one changing the path). Can also be useful simply to find out the
		2037	new values.
		2038
		2039	=item ev_statdata attr [read-only]
		2040
		2041	The most-recently detected attributes of the file. Although the type is
		2042	C<ev_statdata>, this is usually the (or one of the) C<struct stat> types
		2043	suitable for your system, but you can only rely on the POSIX-standardised
		2044	members to be present. If the C<st_nlink> member is C<0>, then there was
		2045	some error while C<stat>ing the file.
		2046
		2047	=item ev_statdata prev [read-only]
		2048
		2049	The previous attributes of the file. The callback gets invoked whenever
		2050	C<prev> != C<attr>, or, more precisely, one or more of these members
		2051	differ: C<st_dev>, C<st_ino>, C<st_mode>, C<st_nlink>, C<st_uid>,
		2052	C<st_gid>, C<st_rdev>, C<st_size>, C<st_atime>, C<st_mtime>, C<st_ctime>.
		2053
		2054	=item ev_tstamp interval [read-only]
		2055
		2056	The specified interval.
		2057
		2058	=item const char *path [read-only]
		2059
		2060	The file system path that is being watched.
		2061
		2062	=back
		2063
		2064	=head3 Examples
		2065
		2066	Example: Watch C</etc/passwd> for attribute changes.
		2067
		2068	static void
		2069	passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)
		2070	{
		2071	/* /etc/passwd changed in some way */
		2072	if (w->attr.st_nlink)
		2073	{
		2074	printf ("passwd current size %ld\n", (long)w->attr.st_size);
		2075	printf ("passwd current atime %ld\n", (long)w->attr.st_mtime);
		2076	printf ("passwd current mtime %ld\n", (long)w->attr.st_mtime);
		2077	}
		2078	else
		2079	/* you shalt not abuse printf for puts */
		2080	puts ("wow, /etc/passwd is not there, expect problems. "
		2081	"if this is windows, they already arrived\n");
		2082	}
		2083
		2084	...
		2085	ev_stat passwd;
		2086
		2087	ev_stat_init (&passwd, passwd_cb, "/etc/passwd", 0.);
		2088	ev_stat_start (loop, &passwd);
		2089
		2090	Example: Like above, but additionally use a one-second delay so we do not
		2091	miss updates (however, frequent updates will delay processing, too, so
		2092	one might do the work both on C<ev_stat> callback invocation I<and> on
		2093	C<ev_timer> callback invocation).
		2094
		2095	static ev_stat passwd;
		2096	static ev_timer timer;
		2097
		2098	static void
		2099	timer_cb (EV_P_ ev_timer *w, int revents)
		2100	{
		2101	ev_timer_stop (EV_A_ w);
		2102
		2103	/* now it's one second after the most recent passwd change */
		2104	}
		2105
		2106	static void
		2107	stat_cb (EV_P_ ev_stat *w, int revents)
		2108	{
		2109	/* reset the one-second timer */
		2110	ev_timer_again (EV_A_ &timer);
		2111	}
		2112
		2113	...
		2114	ev_stat_init (&passwd, stat_cb, "/etc/passwd", 0.);
		2115	ev_stat_start (loop, &passwd);
		2116	ev_timer_init (&timer, timer_cb, 0., 1.02);
		2117
		2118
1050	=head2 C<ev_idle> - when you've got nothing better to do	2119	=head2 C<ev_idle> - when you've got nothing better to do...
1051		2120
1052	Idle watchers trigger events when there are no other events are pending	2121	Idle watchers trigger events when no other events of the same or higher
1053	(prepare, check and other idle watchers do not count). That is, as long	2122	priority are pending (prepare, check and other idle watchers do not count
1054	as your process is busy handling sockets or timeouts (or even signals,	2123	as receiving "events").
1055	imagine) it will not be triggered. But when your process is idle all idle	2124
1056	watchers are being called again and again, once per event loop iteration -	2125	That is, as long as your process is busy handling sockets or timeouts
		2126	(or even signals, imagine) of the same or higher priority it will not be
		2127	triggered. But when your process is idle (or only lower-priority watchers
		2128	are pending), the idle watchers are being called once per event loop
1057	until stopped, that is, or your process receives more events and becomes	2129	iteration - until stopped, that is, or your process receives more events
1058	busy.	2130	and becomes busy again with higher priority stuff.
1059		2131
1060	The most noteworthy effect is that as long as any idle watchers are	2132	The most noteworthy effect is that as long as any idle watchers are
1061	active, the process will not block when waiting for new events.	2133	active, the process will not block when waiting for new events.
1062		2134
1063	Apart from keeping your process non-blocking (which is a useful	2135	Apart from keeping your process non-blocking (which is a useful
1064	effect on its own sometimes), idle watchers are a good place to do	2136	effect on its own sometimes), idle watchers are a good place to do
1065	"pseudo-background processing", or delay processing stuff to after the	2137	"pseudo-background processing", or delay processing stuff to after the
1066	event loop has handled all outstanding events.	2138	event loop has handled all outstanding events.
1067		2139
		2140	=head3 Watcher-Specific Functions and Data Members
		2141
1068	=over 4	2142	=over 4
1069		2143
1070	=item ev_idle_init (ev_signal *, callback)	2144	=item ev_idle_init (ev_signal *, callback)
1071		2145
1072	Initialises and configures the idle watcher - it has no parameters of any	2146	Initialises and configures the idle watcher - it has no parameters of any
1073	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,	2147	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,
1074	believe me.	2148	believe me.
1075		2149
1076	=back	2150	=back
1077		2151
		2152	=head3 Examples
		2153
1078	Example: dynamically allocate an C<ev_idle>, start it, and in the	2154	Example: Dynamically allocate an C<ev_idle> watcher, start it, and in the
1079	callback, free it. Alos, use no error checking, as usual.	2155	callback, free it. Also, use no error checking, as usual.
1080		2156
1081	static void	2157	static void
1082	idle_cb (struct ev_loop *loop, struct ev_idle *w, int revents)	2158	idle_cb (struct ev_loop *loop, ev_idle *w, int revents)
1083	{	2159	{
1084	free (w);	2160	free (w);
1085	// now do something you wanted to do when the program has	2161	// now do something you wanted to do when the program has
1086	// no longer asnything immediate to do.	2162	// no longer anything immediate to do.
1087	}	2163	}
1088		2164
1089	struct ev_idle *idle_watcher = malloc (sizeof (struct ev_idle));	2165	ev_idle *idle_watcher = malloc (sizeof (ev_idle));
1090	ev_idle_init (idle_watcher, idle_cb);	2166	ev_idle_init (idle_watcher, idle_cb);
1091	ev_idle_start (loop, idle_cb);	2167	ev_idle_start (loop, idle_cb);
1092		2168
1093		2169
1094	=head2 C<ev_prepare> and C<ev_check> - customise your event loop	2170	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!
1095		2171
1096	Prepare and check watchers are usually (but not always) used in tandem:	2172	Prepare and check watchers are usually (but not always) used in pairs:
1097	prepare watchers get invoked before the process blocks and check watchers	2173	prepare watchers get invoked before the process blocks and check watchers
1098	afterwards.	2174	afterwards.
1099		2175
		2176	You I<must not> call C<ev_loop> or similar functions that enter
		2177	the current event loop from either C<ev_prepare> or C<ev_check>
		2178	watchers. Other loops than the current one are fine, however. The
		2179	rationale behind this is that you do not need to check for recursion in
		2180	those watchers, i.e. the sequence will always be C<ev_prepare>, blocking,
		2181	C<ev_check> so if you have one watcher of each kind they will always be
		2182	called in pairs bracketing the blocking call.
		2183
1100	Their main purpose is to integrate other event mechanisms into libev and	2184	Their main purpose is to integrate other event mechanisms into libev and
1101	their use is somewhat advanced. This could be used, for example, to track	2185	their use is somewhat advanced. They could be used, for example, to track
1102	variable changes, implement your own watchers, integrate net-snmp or a	2186	variable changes, implement your own watchers, integrate net-snmp or a
1103	coroutine library and lots more.	2187	coroutine library and lots more. They are also occasionally useful if
		2188	you cache some data and want to flush it before blocking (for example,
		2189	in X programs you might want to do an C<XFlush ()> in an C<ev_prepare>
		2190	watcher).
1104		2191
1105	This is done by examining in each prepare call which file descriptors need	2192	This is done by examining in each prepare call which file descriptors
1106	to be watched by the other library, registering C<ev_io> watchers for	2193	need to be watched by the other library, registering C<ev_io> watchers
1107	them and starting an C<ev_timer> watcher for any timeouts (many libraries	2194	for them and starting an C<ev_timer> watcher for any timeouts (many
1108	provide just this functionality). Then, in the check watcher you check for	2195	libraries provide exactly this functionality). Then, in the check watcher,
1109	any events that occured (by checking the pending status of all watchers	2196	you check for any events that occurred (by checking the pending status
1110	and stopping them) and call back into the library. The I/O and timer	2197	of all watchers and stopping them) and call back into the library. The
1111	callbacks will never actually be called (but must be valid nevertheless,	2198	I/O and timer callbacks will never actually be called (but must be valid
1112	because you never know, you know?).	2199	nevertheless, because you never know, you know?).
1113		2200
1114	As another example, the Perl Coro module uses these hooks to integrate	2201	As another example, the Perl Coro module uses these hooks to integrate
1115	coroutines into libev programs, by yielding to other active coroutines	2202	coroutines into libev programs, by yielding to other active coroutines
1116	during each prepare and only letting the process block if no coroutines	2203	during each prepare and only letting the process block if no coroutines
1117	are ready to run (it's actually more complicated: it only runs coroutines	2204	are ready to run (it's actually more complicated: it only runs coroutines
1118	with priority higher than or equal to the event loop and one coroutine	2205	with priority higher than or equal to the event loop and one coroutine
1119	of lower priority, but only once, using idle watchers to keep the event	2206	of lower priority, but only once, using idle watchers to keep the event
1120	loop from blocking if lower-priority coroutines are active, thus mapping	2207	loop from blocking if lower-priority coroutines are active, thus mapping
1121	low-priority coroutines to idle/background tasks).	2208	low-priority coroutines to idle/background tasks).
1122		2209
		2210	It is recommended to give C<ev_check> watchers highest (C<EV_MAXPRI>)
		2211	priority, to ensure that they are being run before any other watchers
		2212	after the poll (this doesn't matter for C<ev_prepare> watchers).
		2213
		2214	Also, C<ev_check> watchers (and C<ev_prepare> watchers, too) should not
		2215	activate ("feed") events into libev. While libev fully supports this, they
		2216	might get executed before other C<ev_check> watchers did their job. As
		2217	C<ev_check> watchers are often used to embed other (non-libev) event
		2218	loops those other event loops might be in an unusable state until their
		2219	C<ev_check> watcher ran (always remind yourself to coexist peacefully with
		2220	others).
		2221
		2222	=head3 Watcher-Specific Functions and Data Members
		2223
1123	=over 4	2224	=over 4
1124		2225
1125	=item ev_prepare_init (ev_prepare *, callback)	2226	=item ev_prepare_init (ev_prepare *, callback)
1126		2227
1127	=item ev_check_init (ev_check *, callback)	2228	=item ev_check_init (ev_check *, callback)
1128		2229
1129	Initialises and configures the prepare or check watcher - they have no	2230	Initialises and configures the prepare or check watcher - they have no
1130	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>	2231	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>
1131	macros, but using them is utterly, utterly and completely pointless.	2232	macros, but using them is utterly, utterly, utterly and completely
		2233	pointless.
1132		2234
1133	=back	2235	=back
1134		2236
1135	Example: *TODO*.	2237	=head3 Examples
1136		2238
		2239	There are a number of principal ways to embed other event loops or modules
		2240	into libev. Here are some ideas on how to include libadns into libev
		2241	(there is a Perl module named C<EV::ADNS> that does this, which you could
		2242	use as a working example. Another Perl module named C<EV::Glib> embeds a
		2243	Glib main context into libev, and finally, C<Glib::EV> embeds EV into the
		2244	Glib event loop).
1137		2245
		2246	Method 1: Add IO watchers and a timeout watcher in a prepare handler,
		2247	and in a check watcher, destroy them and call into libadns. What follows
		2248	is pseudo-code only of course. This requires you to either use a low
		2249	priority for the check watcher or use C<ev_clear_pending> explicitly, as
		2250	the callbacks for the IO/timeout watchers might not have been called yet.
		2251
		2252	static ev_io iow [nfd];
		2253	static ev_timer tw;
		2254
		2255	static void
		2256	io_cb (struct ev_loop *loop, ev_io *w, int revents)
		2257	{
		2258	}
		2259
		2260	// create io watchers for each fd and a timer before blocking
		2261	static void
		2262	adns_prepare_cb (struct ev_loop *loop, ev_prepare *w, int revents)
		2263	{
		2264	int timeout = 3600000;
		2265	struct pollfd fds [nfd];
		2266	// actual code will need to loop here and realloc etc.
		2267	adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));
		2268
		2269	/* the callback is illegal, but won't be called as we stop during check */
		2270	ev_timer_init (&tw, 0, timeout * 1e-3);
		2271	ev_timer_start (loop, &tw);
		2272
		2273	// create one ev_io per pollfd
		2274	for (int i = 0; i < nfd; ++i)
		2275	{
		2276	ev_io_init (iow + i, io_cb, fds [i].fd,
		2277	((fds [i].events & POLLIN ? EV_READ : 0)
		2278	| (fds [i].events & POLLOUT ? EV_WRITE : 0)));
		2279
		2280	fds [i].revents = 0;
		2281	ev_io_start (loop, iow + i);
		2282	}
		2283	}
		2284
		2285	// stop all watchers after blocking
		2286	static void
		2287	adns_check_cb (struct ev_loop *loop, ev_check *w, int revents)
		2288	{
		2289	ev_timer_stop (loop, &tw);
		2290
		2291	for (int i = 0; i < nfd; ++i)
		2292	{
		2293	// set the relevant poll flags
		2294	// could also call adns_processreadable etc. here
		2295	struct pollfd *fd = fds + i;
		2296	int revents = ev_clear_pending (iow + i);
		2297	if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;
		2298	if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;
		2299
		2300	// now stop the watcher
		2301	ev_io_stop (loop, iow + i);
		2302	}
		2303
		2304	adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));
		2305	}
		2306
		2307	Method 2: This would be just like method 1, but you run C<adns_afterpoll>
		2308	in the prepare watcher and would dispose of the check watcher.
		2309
		2310	Method 3: If the module to be embedded supports explicit event
		2311	notification (libadns does), you can also make use of the actual watcher
		2312	callbacks, and only destroy/create the watchers in the prepare watcher.
		2313
		2314	static void
		2315	timer_cb (EV_P_ ev_timer *w, int revents)
		2316	{
		2317	adns_state ads = (adns_state)w->data;
		2318	update_now (EV_A);
		2319
		2320	adns_processtimeouts (ads, &tv_now);
		2321	}
		2322
		2323	static void
		2324	io_cb (EV_P_ ev_io *w, int revents)
		2325	{
		2326	adns_state ads = (adns_state)w->data;
		2327	update_now (EV_A);
		2328
		2329	if (revents & EV_READ ) adns_processreadable (ads, w->fd, &tv_now);
		2330	if (revents & EV_WRITE) adns_processwriteable (ads, w->fd, &tv_now);
		2331	}
		2332
		2333	// do not ever call adns_afterpoll
		2334
		2335	Method 4: Do not use a prepare or check watcher because the module you
		2336	want to embed is not flexible enough to support it. Instead, you can
		2337	override their poll function. The drawback with this solution is that the
		2338	main loop is now no longer controllable by EV. The C<Glib::EV> module uses
		2339	this approach, effectively embedding EV as a client into the horrible
		2340	libglib event loop.
		2341
		2342	static gint
		2343	event_poll_func (GPollFD *fds, guint nfds, gint timeout)
		2344	{
		2345	int got_events = 0;
		2346
		2347	for (n = 0; n < nfds; ++n)
		2348	// create/start io watcher that sets the relevant bits in fds[n] and increment got_events
		2349
		2350	if (timeout >= 0)
		2351	// create/start timer
		2352
		2353	// poll
		2354	ev_loop (EV_A_ 0);
		2355
		2356	// stop timer again
		2357	if (timeout >= 0)
		2358	ev_timer_stop (EV_A_ &to);
		2359
		2360	// stop io watchers again - their callbacks should have set
		2361	for (n = 0; n < nfds; ++n)
		2362	ev_io_stop (EV_A_ iow [n]);
		2363
		2364	return got_events;
		2365	}
		2366
		2367
1138	=head2 C<ev_embed> - when one backend isn't enough	2368	=head2 C<ev_embed> - when one backend isn't enough...
1139		2369
1140	This is a rather advanced watcher type that lets you embed one event loop	2370	This is a rather advanced watcher type that lets you embed one event loop
1141	into another (currently only C<ev_io> events are supported in the embedded	2371	into another (currently only C<ev_io> events are supported in the embedded
1142	loop, other types of watchers might be handled in a delayed or incorrect	2372	loop, other types of watchers might be handled in a delayed or incorrect
1143	fashion and must not be used).	2373	fashion and must not be used).
…		…
1146	prioritise I/O.	2376	prioritise I/O.
1147		2377
1148	As an example for a bug workaround, the kqueue backend might only support	2378	As an example for a bug workaround, the kqueue backend might only support
1149	sockets on some platform, so it is unusable as generic backend, but you	2379	sockets on some platform, so it is unusable as generic backend, but you
1150	still want to make use of it because you have many sockets and it scales	2380	still want to make use of it because you have many sockets and it scales
1151	so nicely. In this case, you would create a kqueue-based loop and embed it	2381	so nicely. In this case, you would create a kqueue-based loop and embed
1152	into your default loop (which might use e.g. poll). Overall operation will	2382	it into your default loop (which might use e.g. poll). Overall operation
1153	be a bit slower because first libev has to poll and then call kevent, but	2383	will be a bit slower because first libev has to call C<poll> and then
1154	at least you can use both at what they are best.	2384	C<kevent>, but at least you can use both mechanisms for what they are
		2385	best: C<kqueue> for scalable sockets and C<poll> if you want it to work :)
1155		2386
1156	As for prioritising I/O: rarely you have the case where some fds have	2387	As for prioritising I/O: under rare circumstances you have the case where
1157	to be watched and handled very quickly (with low latency), and even	2388	some fds have to be watched and handled very quickly (with low latency),
1158	priorities and idle watchers might have too much overhead. In this case	2389	and even priorities and idle watchers might have too much overhead. In
1159	you would put all the high priority stuff in one loop and all the rest in	2390	this case you would put all the high priority stuff in one loop and all
1160	a second one, and embed the second one in the first.	2391	the rest in a second one, and embed the second one in the first.
1161		2392
1162	As long as the watcher is active, the callback will be invoked every time	2393	As long as the watcher is active, the callback will be invoked every time
1163	there might be events pending in the embedded loop. The callback must then	2394	there might be events pending in the embedded loop. The callback must then
1164	call C<ev_embed_sweep (mainloop, watcher)> to make a single sweep and invoke	2395	call C<ev_embed_sweep (mainloop, watcher)> to make a single sweep and invoke
1165	their callbacks (you could also start an idle watcher to give the embedded	2396	their callbacks (you could also start an idle watcher to give the embedded
…		…
1173	interested in that.	2404	interested in that.
1174		2405
1175	Also, there have not currently been made special provisions for forking:	2406	Also, there have not currently been made special provisions for forking:
1176	when you fork, you not only have to call C<ev_loop_fork> on both loops,	2407	when you fork, you not only have to call C<ev_loop_fork> on both loops,
1177	but you will also have to stop and restart any C<ev_embed> watchers	2408	but you will also have to stop and restart any C<ev_embed> watchers
1178	yourself.	2409	yourself - but you can use a fork watcher to handle this automatically,
		2410	and future versions of libev might do just that.
1179		2411
1180	Unfortunately, not all backends are embeddable, only the ones returned by	2412	Unfortunately, not all backends are embeddable: only the ones returned by
1181	C<ev_embeddable_backends> are, which, unfortunately, does not include any	2413	C<ev_embeddable_backends> are, which, unfortunately, does not include any
1182	portable one.	2414	portable one.
1183		2415
1184	So when you want to use this feature you will always have to be prepared	2416	So when you want to use this feature you will always have to be prepared
1185	that you cannot get an embeddable loop. The recommended way to get around	2417	that you cannot get an embeddable loop. The recommended way to get around
1186	this is to have a separate variables for your embeddable loop, try to	2418	this is to have a separate variables for your embeddable loop, try to
1187	create it, and if that fails, use the normal loop for everything:	2419	create it, and if that fails, use the normal loop for everything.
1188		2420
1189	struct ev_loop *loop_hi = ev_default_init (0);	2421	=head3 C<ev_embed> and fork
1190	struct ev_loop *loop_lo = 0;
1191	struct ev_embed embed;
1192
1193	// see if there is a chance of getting one that works
1194	// (remember that a flags value of 0 means autodetection)
1195	loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
1196	? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
1197	: 0;
1198		2422
1199	// if we got one, then embed it, otherwise default to loop_hi	2423	While the C<ev_embed> watcher is running, forks in the embedding loop will
1200	if (loop_lo)	2424	automatically be applied to the embedded loop as well, so no special
1201	{	2425	fork handling is required in that case. When the watcher is not running,
1202	ev_embed_init (&embed, 0, loop_lo);	2426	however, it is still the task of the libev user to call C<ev_loop_fork ()>
1203	ev_embed_start (loop_hi, &embed);	2427	as applicable.
1204	}	2428
1205	else	2429	=head3 Watcher-Specific Functions and Data Members
1206	loop_lo = loop_hi;
1207		2430
1208	=over 4	2431	=over 4
1209		2432
1210	=item ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)	2433	=item ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)
1211		2434
…		…
1213		2436
1214	Configures the watcher to embed the given loop, which must be	2437	Configures the watcher to embed the given loop, which must be
1215	embeddable. If the callback is C<0>, then C<ev_embed_sweep> will be	2438	embeddable. If the callback is C<0>, then C<ev_embed_sweep> will be
1216	invoked automatically, otherwise it is the responsibility of the callback	2439	invoked automatically, otherwise it is the responsibility of the callback
1217	to invoke it (it will continue to be called until the sweep has been done,	2440	to invoke it (it will continue to be called until the sweep has been done,
1218	if you do not want thta, you need to temporarily stop the embed watcher).	2441	if you do not want that, you need to temporarily stop the embed watcher).
1219		2442
1220	=item ev_embed_sweep (loop, ev_embed *)	2443	=item ev_embed_sweep (loop, ev_embed *)
1221		2444
1222	Make a single, non-blocking sweep over the embedded loop. This works	2445	Make a single, non-blocking sweep over the embedded loop. This works
1223	similarly to C<ev_loop (embedded_loop, EVLOOP_NONBLOCK)>, but in the most	2446	similarly to C<ev_loop (embedded_loop, EVLOOP_NONBLOCK)>, but in the most
1224	apropriate way for embedded loops.	2447	appropriate way for embedded loops.
		2448
		2449	=item struct ev_loop *other [read-only]
		2450
		2451	The embedded event loop.
1225		2452
1226	=back	2453	=back
1227		2454
		2455	=head3 Examples
		2456
		2457	Example: Try to get an embeddable event loop and embed it into the default
		2458	event loop. If that is not possible, use the default loop. The default
		2459	loop is stored in C<loop_hi>, while the embeddable loop is stored in
		2460	C<loop_lo> (which is C<loop_hi> in the case no embeddable loop can be
		2461	used).
		2462
		2463	struct ev_loop *loop_hi = ev_default_init (0);
		2464	struct ev_loop *loop_lo = 0;
		2465	ev_embed embed;
		2466
		2467	// see if there is a chance of getting one that works
		2468	// (remember that a flags value of 0 means autodetection)
		2469	loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
		2470	? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
		2471	: 0;
		2472
		2473	// if we got one, then embed it, otherwise default to loop_hi
		2474	if (loop_lo)
		2475	{
		2476	ev_embed_init (&embed, 0, loop_lo);
		2477	ev_embed_start (loop_hi, &embed);
		2478	}
		2479	else
		2480	loop_lo = loop_hi;
		2481
		2482	Example: Check if kqueue is available but not recommended and create
		2483	a kqueue backend for use with sockets (which usually work with any
		2484	kqueue implementation). Store the kqueue/socket-only event loop in
		2485	C<loop_socket>. (One might optionally use C<EVFLAG_NOENV>, too).
		2486
		2487	struct ev_loop *loop = ev_default_init (0);
		2488	struct ev_loop *loop_socket = 0;
		2489	ev_embed embed;
		2490
		2491	if (ev_supported_backends () & ~ev_recommended_backends () & EVBACKEND_KQUEUE)
		2492	if ((loop_socket = ev_loop_new (EVBACKEND_KQUEUE))
		2493	{
		2494	ev_embed_init (&embed, 0, loop_socket);
		2495	ev_embed_start (loop, &embed);
		2496	}
		2497
		2498	if (!loop_socket)
		2499	loop_socket = loop;
		2500
		2501	// now use loop_socket for all sockets, and loop for everything else
		2502
		2503
		2504	=head2 C<ev_fork> - the audacity to resume the event loop after a fork
		2505
		2506	Fork watchers are called when a C<fork ()> was detected (usually because
		2507	whoever is a good citizen cared to tell libev about it by calling
		2508	C<ev_default_fork> or C<ev_loop_fork>). The invocation is done before the
		2509	event loop blocks next and before C<ev_check> watchers are being called,
		2510	and only in the child after the fork. If whoever good citizen calling
		2511	C<ev_default_fork> cheats and calls it in the wrong process, the fork
		2512	handlers will be invoked, too, of course.
		2513
		2514	=head3 Watcher-Specific Functions and Data Members
		2515
		2516	=over 4
		2517
		2518	=item ev_fork_init (ev_signal *, callback)
		2519
		2520	Initialises and configures the fork watcher - it has no parameters of any
		2521	kind. There is a C<ev_fork_set> macro, but using it is utterly pointless,
		2522	believe me.
		2523
		2524	=back
		2525
		2526
		2527	=head2 C<ev_async> - how to wake up another event loop
		2528
		2529	In general, you cannot use an C<ev_loop> from multiple threads or other
		2530	asynchronous sources such as signal handlers (as opposed to multiple event
		2531	loops - those are of course safe to use in different threads).
		2532
		2533	Sometimes, however, you need to wake up another event loop you do not
		2534	control, for example because it belongs to another thread. This is what
		2535	C<ev_async> watchers do: as long as the C<ev_async> watcher is active, you
		2536	can signal it by calling C<ev_async_send>, which is thread- and signal
		2537	safe.
		2538
		2539	This functionality is very similar to C<ev_signal> watchers, as signals,
		2540	too, are asynchronous in nature, and signals, too, will be compressed
		2541	(i.e. the number of callback invocations may be less than the number of
		2542	C<ev_async_sent> calls).
		2543
		2544	Unlike C<ev_signal> watchers, C<ev_async> works with any event loop, not
		2545	just the default loop.
		2546
		2547	=head3 Queueing
		2548
		2549	C<ev_async> does not support queueing of data in any way. The reason
		2550	is that the author does not know of a simple (or any) algorithm for a
		2551	multiple-writer-single-reader queue that works in all cases and doesn't
		2552	need elaborate support such as pthreads.
		2553
		2554	That means that if you want to queue data, you have to provide your own
		2555	queue. But at least I can tell you how to implement locking around your
		2556	queue:
		2557
		2558	=over 4
		2559
		2560	=item queueing from a signal handler context
		2561
		2562	To implement race-free queueing, you simply add to the queue in the signal
		2563	handler but you block the signal handler in the watcher callback. Here is
		2564	an example that does that for some fictitious SIGUSR1 handler:
		2565
		2566	static ev_async mysig;
		2567
		2568	static void
		2569	sigusr1_handler (void)
		2570	{
		2571	sometype data;
		2572
		2573	// no locking etc.
		2574	queue_put (data);
		2575	ev_async_send (EV_DEFAULT_ &mysig);
		2576	}
		2577
		2578	static void
		2579	mysig_cb (EV_P_ ev_async *w, int revents)
		2580	{
		2581	sometype data;
		2582	sigset_t block, prev;
		2583
		2584	sigemptyset (&block);
		2585	sigaddset (&block, SIGUSR1);
		2586	sigprocmask (SIG_BLOCK, &block, &prev);
		2587
		2588	while (queue_get (&data))
		2589	process (data);
		2590
		2591	if (sigismember (&prev, SIGUSR1)
		2592	sigprocmask (SIG_UNBLOCK, &block, 0);
		2593	}
		2594
		2595	(Note: pthreads in theory requires you to use C<pthread_setmask>
		2596	instead of C<sigprocmask> when you use threads, but libev doesn't do it
		2597	either...).
		2598
		2599	=item queueing from a thread context
		2600
		2601	The strategy for threads is different, as you cannot (easily) block
		2602	threads but you can easily preempt them, so to queue safely you need to
		2603	employ a traditional mutex lock, such as in this pthread example:
		2604
		2605	static ev_async mysig;
		2606	static pthread_mutex_t mymutex = PTHREAD_MUTEX_INITIALIZER;
		2607
		2608	static void
		2609	otherthread (void)
		2610	{
		2611	// only need to lock the actual queueing operation
		2612	pthread_mutex_lock (&mymutex);
		2613	queue_put (data);
		2614	pthread_mutex_unlock (&mymutex);
		2615
		2616	ev_async_send (EV_DEFAULT_ &mysig);
		2617	}
		2618
		2619	static void
		2620	mysig_cb (EV_P_ ev_async *w, int revents)
		2621	{
		2622	pthread_mutex_lock (&mymutex);
		2623
		2624	while (queue_get (&data))
		2625	process (data);
		2626
		2627	pthread_mutex_unlock (&mymutex);
		2628	}
		2629
		2630	=back
		2631
		2632
		2633	=head3 Watcher-Specific Functions and Data Members
		2634
		2635	=over 4
		2636
		2637	=item ev_async_init (ev_async *, callback)
		2638
		2639	Initialises and configures the async watcher - it has no parameters of any
		2640	kind. There is a C<ev_asynd_set> macro, but using it is utterly pointless,
		2641	trust me.
		2642
		2643	=item ev_async_send (loop, ev_async *)
		2644
		2645	Sends/signals/activates the given C<ev_async> watcher, that is, feeds
		2646	an C<EV_ASYNC> event on the watcher into the event loop. Unlike
		2647	C<ev_feed_event>, this call is safe to do from other threads, signal or
		2648	similar contexts (see the discussion of C<EV_ATOMIC_T> in the embedding
		2649	section below on what exactly this means).
		2650
		2651	This call incurs the overhead of a system call only once per loop iteration,
		2652	so while the overhead might be noticeable, it doesn't apply to repeated
		2653	calls to C<ev_async_send>.
		2654
		2655	=item bool = ev_async_pending (ev_async *)
		2656
		2657	Returns a non-zero value when C<ev_async_send> has been called on the
		2658	watcher but the event has not yet been processed (or even noted) by the
		2659	event loop.
		2660
		2661	C<ev_async_send> sets a flag in the watcher and wakes up the loop. When
		2662	the loop iterates next and checks for the watcher to have become active,
		2663	it will reset the flag again. C<ev_async_pending> can be used to very
		2664	quickly check whether invoking the loop might be a good idea.
		2665
		2666	Not that this does I<not> check whether the watcher itself is pending, only
		2667	whether it has been requested to make this watcher pending.
		2668
		2669	=back
		2670
1228		2671
1229	=head1 OTHER FUNCTIONS	2672	=head1 OTHER FUNCTIONS
1230		2673
1231	There are some other functions of possible interest. Described. Here. Now.	2674	There are some other functions of possible interest. Described. Here. Now.
1232		2675
1233	=over 4	2676	=over 4
1234		2677
1235	=item ev_once (loop, int fd, int events, ev_tstamp timeout, callback)	2678	=item ev_once (loop, int fd, int events, ev_tstamp timeout, callback)
1236		2679
1237	This function combines a simple timer and an I/O watcher, calls your	2680	This function combines a simple timer and an I/O watcher, calls your
1238	callback on whichever event happens first and automatically stop both	2681	callback on whichever event happens first and automatically stops both
1239	watchers. This is useful if you want to wait for a single event on an fd	2682	watchers. This is useful if you want to wait for a single event on an fd
1240	or timeout without having to allocate/configure/start/stop/free one or	2683	or timeout without having to allocate/configure/start/stop/free one or
1241	more watchers yourself.	2684	more watchers yourself.
1242		2685
1243	If C<fd> is less than 0, then no I/O watcher will be started and events	2686	If C<fd> is less than 0, then no I/O watcher will be started and the
1244	is being ignored. Otherwise, an C<ev_io> watcher for the given C<fd> and	2687	C<events> argument is being ignored. Otherwise, an C<ev_io> watcher for
1245	C<events> set will be craeted and started.	2688	the given C<fd> and C<events> set will be created and started.
1246		2689
1247	If C<timeout> is less than 0, then no timeout watcher will be	2690	If C<timeout> is less than 0, then no timeout watcher will be
1248	started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and	2691	started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and
1249	repeat = 0) will be started. While C<0> is a valid timeout, it is of	2692	repeat = 0) will be started. C<0> is a valid timeout.
1250	dubious value.
1251		2693
1252	The callback has the type C<void (*cb)(int revents, void *arg)> and gets	2694	The callback has the type C<void (*cb)(int revents, void *arg)> and gets
1253	passed an C<revents> set like normal event callbacks (a combination of	2695	passed an C<revents> set like normal event callbacks (a combination of
1254	C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMEOUT>) and the C<arg>	2696	C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMEOUT>) and the C<arg>
1255	value passed to C<ev_once>:	2697	value passed to C<ev_once>. Note that it is possible to receive I<both>
		2698	a timeout and an io event at the same time - you probably should give io
		2699	events precedence.
1256		2700
		2701	Example: wait up to ten seconds for data to appear on STDIN_FILENO.
		2702
1257	static void stdin_ready (int revents, void *arg)	2703	static void stdin_ready (int revents, void *arg)
1258	{	2704	{
1259	if (revents & EV_TIMEOUT)
1260	/* doh, nothing entered */;
1261	else if (revents & EV_READ)	2705	if (revents & EV_READ)
1262	/* stdin might have data for us, joy! */;	2706	/* stdin might have data for us, joy! */;
		2707	else if (revents & EV_TIMEOUT)
		2708	/* doh, nothing entered */;
1263	}	2709	}
1264		2710
1265	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);	2711	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
1266		2712
1267	=item ev_feed_event (ev_loop *, watcher *, int revents)	2713	=item ev_feed_event (struct ev_loop *, watcher *, int revents)
1268		2714
1269	Feeds the given event set into the event loop, as if the specified event	2715	Feeds the given event set into the event loop, as if the specified event
1270	had happened for the specified watcher (which must be a pointer to an	2716	had happened for the specified watcher (which must be a pointer to an
1271	initialised but not necessarily started event watcher).	2717	initialised but not necessarily started event watcher).
1272		2718
1273	=item ev_feed_fd_event (ev_loop *, int fd, int revents)	2719	=item ev_feed_fd_event (struct ev_loop *, int fd, int revents)
1274		2720
1275	Feed an event on the given fd, as if a file descriptor backend detected	2721	Feed an event on the given fd, as if a file descriptor backend detected
1276	the given events it.	2722	the given events it.
1277		2723
1278	=item ev_feed_signal_event (ev_loop *loop, int signum)	2724	=item ev_feed_signal_event (struct ev_loop *loop, int signum)
1279		2725
1280	Feed an event as if the given signal occured (C<loop> must be the default	2726	Feed an event as if the given signal occurred (C<loop> must be the default
1281	loop!).	2727	loop!).
1282		2728
1283	=back	2729	=back
1284		2730
1285		2731
…		…
1301		2747
1302	=item * Priorities are not currently supported. Initialising priorities	2748	=item * Priorities are not currently supported. Initialising priorities
1303	will fail and all watchers will have the same priority, even though there	2749	will fail and all watchers will have the same priority, even though there
1304	is an ev_pri field.	2750	is an ev_pri field.
1305		2751
		2752	=item * In libevent, the last base created gets the signals, in libev, the
		2753	first base created (== the default loop) gets the signals.
		2754
1306	=item * Other members are not supported.	2755	=item * Other members are not supported.
1307		2756
1308	=item * The libev emulation is I<not> ABI compatible to libevent, you need	2757	=item * The libev emulation is I<not> ABI compatible to libevent, you need
1309	to use the libev header file and library.	2758	to use the libev header file and library.
1310		2759
1311	=back	2760	=back
1312		2761
1313	=head1 C++ SUPPORT	2762	=head1 C++ SUPPORT
1314		2763
1315	Libev comes with some simplistic wrapper classes for C++ that mainly allow	2764	Libev comes with some simplistic wrapper classes for C++ that mainly allow
1316	you to use some convinience methods to start/stop watchers and also change	2765	you to use some convenience methods to start/stop watchers and also change
1317	the callback model to a model using method callbacks on objects.	2766	the callback model to a model using method callbacks on objects.
1318		2767
1319	To use it,	2768	To use it,
1320		2769
1321	#include <ev++.h>	2770	#include <ev++.h>
1322		2771
1323	(it is not installed by default). This automatically includes F<ev.h>	2772	This automatically includes F<ev.h> and puts all of its definitions (many
1324	and puts all of its definitions (many of them macros) into the global	2773	of them macros) into the global namespace. All C++ specific things are
1325	namespace. All C++ specific things are put into the C<ev> namespace.	2774	put into the C<ev> namespace. It should support all the same embedding
		2775	options as F<ev.h>, most notably C<EV_MULTIPLICITY>.
1326		2776
1327	It should support all the same embedding options as F<ev.h>, most notably	2777	Care has been taken to keep the overhead low. The only data member the C++
1328	C<EV_MULTIPLICITY>.	2778	classes add (compared to plain C-style watchers) is the event loop pointer
		2779	that the watcher is associated with (or no additional members at all if
		2780	you disable C<EV_MULTIPLICITY> when embedding libev).
		2781
		2782	Currently, functions, and static and non-static member functions can be
		2783	used as callbacks. Other types should be easy to add as long as they only
		2784	need one additional pointer for context. If you need support for other
		2785	types of functors please contact the author (preferably after implementing
		2786	it).
1329		2787
1330	Here is a list of things available in the C<ev> namespace:	2788	Here is a list of things available in the C<ev> namespace:
1331		2789
1332	=over 4	2790	=over 4
1333		2791
…		…
1349		2807
1350	All of those classes have these methods:	2808	All of those classes have these methods:
1351		2809
1352	=over 4	2810	=over 4
1353		2811
1354	=item ev::TYPE::TYPE (object *, object::method *)	2812	=item ev::TYPE::TYPE ()
1355		2813
1356	=item ev::TYPE::TYPE (object *, object::method *, struct ev_loop *)	2814	=item ev::TYPE::TYPE (struct ev_loop *)
1357		2815
1358	=item ev::TYPE::~TYPE	2816	=item ev::TYPE::~TYPE
1359		2817
1360	The constructor takes a pointer to an object and a method pointer to	2818	The constructor (optionally) takes an event loop to associate the watcher
1361	the event handler callback to call in this class. The constructor calls	2819	with. If it is omitted, it will use C<EV_DEFAULT>.
1362	C<ev_init> for you, which means you have to call the C<set> method	2820
1363	before starting it. If you do not specify a loop then the constructor	2821	The constructor calls C<ev_init> for you, which means you have to call the
1364	automatically associates the default loop with this watcher.	2822	C<set> method before starting it.
		2823
		2824	It will not set a callback, however: You have to call the templated C<set>
		2825	method to set a callback before you can start the watcher.
		2826
		2827	(The reason why you have to use a method is a limitation in C++ which does
		2828	not allow explicit template arguments for constructors).
1365		2829
1366	The destructor automatically stops the watcher if it is active.	2830	The destructor automatically stops the watcher if it is active.
		2831
		2832	=item w->set<class, &class::method> (object *)
		2833
		2834	This method sets the callback method to call. The method has to have a
		2835	signature of C<void (*)(ev_TYPE &, int)>, it receives the watcher as
		2836	first argument and the C<revents> as second. The object must be given as
		2837	parameter and is stored in the C<data> member of the watcher.
		2838
		2839	This method synthesizes efficient thunking code to call your method from
		2840	the C callback that libev requires. If your compiler can inline your
		2841	callback (i.e. it is visible to it at the place of the C<set> call and
		2842	your compiler is good :), then the method will be fully inlined into the
		2843	thunking function, making it as fast as a direct C callback.
		2844
		2845	Example: simple class declaration and watcher initialisation
		2846
		2847	struct myclass
		2848	{
		2849	void io_cb (ev::io &w, int revents) { }
		2850	}
		2851
		2852	myclass obj;
		2853	ev::io iow;
		2854	iow.set <myclass, &myclass::io_cb> (&obj);
		2855
		2856	=item w->set<function> (void *data = 0)
		2857
		2858	Also sets a callback, but uses a static method or plain function as
		2859	callback. The optional C<data> argument will be stored in the watcher's
		2860	C<data> member and is free for you to use.
		2861
		2862	The prototype of the C<function> must be C<void (*)(ev::TYPE &w, int)>.
		2863
		2864	See the method-C<set> above for more details.
		2865
		2866	Example: Use a plain function as callback.
		2867
		2868	static void io_cb (ev::io &w, int revents) { }
		2869	iow.set <io_cb> ();
1367		2870
1368	=item w->set (struct ev_loop *)	2871	=item w->set (struct ev_loop *)
1369		2872
1370	Associates a different C<struct ev_loop> with this watcher. You can only	2873	Associates a different C<struct ev_loop> with this watcher. You can only
1371	do this when the watcher is inactive (and not pending either).	2874	do this when the watcher is inactive (and not pending either).
1372		2875
1373	=item w->set ([args])	2876	=item w->set ([arguments])
1374		2877
1375	Basically the same as C<ev_TYPE_set>, with the same args. Must be	2878	Basically the same as C<ev_TYPE_set>, with the same arguments. Must be
1376	called at least once. Unlike the C counterpart, an active watcher gets	2879	called at least once. Unlike the C counterpart, an active watcher gets
1377	automatically stopped and restarted.	2880	automatically stopped and restarted when reconfiguring it with this
		2881	method.
1378		2882
1379	=item w->start ()	2883	=item w->start ()
1380		2884
1381	Starts the watcher. Note that there is no C<loop> argument as the	2885	Starts the watcher. Note that there is no C<loop> argument, as the
1382	constructor already takes the loop.	2886	constructor already stores the event loop.
1383		2887
1384	=item w->stop ()	2888	=item w->stop ()
1385		2889
1386	Stops the watcher if it is active. Again, no C<loop> argument.	2890	Stops the watcher if it is active. Again, no C<loop> argument.
1387		2891
1388	=item w->again () C<ev::timer>, C<ev::periodic> only	2892	=item w->again () (C<ev::timer>, C<ev::periodic> only)
1389		2893
1390	For C<ev::timer> and C<ev::periodic>, this invokes the corresponding	2894	For C<ev::timer> and C<ev::periodic>, this invokes the corresponding
1391	C<ev_TYPE_again> function.	2895	C<ev_TYPE_again> function.
1392		2896
1393	=item w->sweep () C<ev::embed> only	2897	=item w->sweep () (C<ev::embed> only)
1394		2898
1395	Invokes C<ev_embed_sweep>.	2899	Invokes C<ev_embed_sweep>.
		2900
		2901	=item w->update () (C<ev::stat> only)
		2902
		2903	Invokes C<ev_stat_stat>.
1396		2904
1397	=back	2905	=back
1398		2906
1399	=back	2907	=back
1400		2908
1401	Example: Define a class with an IO and idle watcher, start one of them in	2909	Example: Define a class with an IO and idle watcher, start one of them in
1402	the constructor.	2910	the constructor.
1403		2911
1404	class myclass	2912	class myclass
1405	{	2913	{
1406	ev_io io; void io_cb (ev::io &w, int revents);	2914	ev::io io ; void io_cb (ev::io &w, int revents);
1407	ev_idle idle void idle_cb (ev::idle &w, int revents);	2915	ev::idle idle; void idle_cb (ev::idle &w, int revents);
1408		2916
1409	myclass ();	2917	myclass (int fd)
1410	}	2918	{
		2919	io .set <myclass, &myclass::io_cb > (this);
		2920	idle.set <myclass, &myclass::idle_cb> (this);
1411		2921
1412	myclass::myclass (int fd)
1413	: io (this, &myclass::io_cb),
1414	idle (this, &myclass::idle_cb)
1415	{
1416	io.start (fd, ev::READ);	2922	io.start (fd, ev::READ);
		2923	}
		2924	};
		2925
		2926
		2927	=head1 OTHER LANGUAGE BINDINGS
		2928
		2929	Libev does not offer other language bindings itself, but bindings for a
		2930	number of languages exist in the form of third-party packages. If you know
		2931	any interesting language binding in addition to the ones listed here, drop
		2932	me a note.
		2933
		2934	=over 4
		2935
		2936	=item Perl
		2937
		2938	The EV module implements the full libev API and is actually used to test
		2939	libev. EV is developed together with libev. Apart from the EV core module,
		2940	there are additional modules that implement libev-compatible interfaces
		2941	to C<libadns> (C<EV::ADNS>, but C<AnyEvent::DNS> is preferred nowadays),
		2942	C<Net::SNMP> (C<Net::SNMP::EV>) and the C<libglib> event core (C<Glib::EV>
		2943	and C<EV::Glib>).
		2944
		2945	It can be found and installed via CPAN, its homepage is at
		2946	L<http://software.schmorp.de/pkg/EV>.
		2947
		2948	=item Python
		2949
		2950	Python bindings can be found at L<http://code.google.com/p/pyev/>. It
		2951	seems to be quite complete and well-documented. Note, however, that the
		2952	patch they require for libev is outright dangerous as it breaks the ABI
		2953	for everybody else, and therefore, should never be applied in an installed
		2954	libev (if python requires an incompatible ABI then it needs to embed
		2955	libev).
		2956
		2957	=item Ruby
		2958
		2959	Tony Arcieri has written a ruby extension that offers access to a subset
		2960	of the libev API and adds file handle abstractions, asynchronous DNS and
		2961	more on top of it. It can be found via gem servers. Its homepage is at
		2962	L<http://rev.rubyforge.org/>.
		2963
		2964	=item D
		2965
		2966	Leandro Lucarella has written a D language binding (F<ev.d>) for libev, to
		2967	be found at L<http://proj.llucax.com.ar/wiki/evd>.
		2968
		2969	=item Ocaml
		2970
		2971	Erkki Seppala has written Ocaml bindings for libev, to be found at
		2972	L<http://modeemi.cs.tut.fi/~flux/software/ocaml-ev/>.
		2973
		2974	=back
		2975
		2976
		2977	=head1 MACRO MAGIC
		2978
		2979	Libev can be compiled with a variety of options, the most fundamental
		2980	of which is C<EV_MULTIPLICITY>. This option determines whether (most)
		2981	functions and callbacks have an initial C<struct ev_loop *> argument.
		2982
		2983	To make it easier to write programs that cope with either variant, the
		2984	following macros are defined:
		2985
		2986	=over 4
		2987
		2988	=item C<EV_A>, C<EV_A_>
		2989
		2990	This provides the loop I<argument> for functions, if one is required ("ev
		2991	loop argument"). The C<EV_A> form is used when this is the sole argument,
		2992	C<EV_A_> is used when other arguments are following. Example:
		2993
		2994	ev_unref (EV_A);
		2995	ev_timer_add (EV_A_ watcher);
		2996	ev_loop (EV_A_ 0);
		2997
		2998	It assumes the variable C<loop> of type C<struct ev_loop *> is in scope,
		2999	which is often provided by the following macro.
		3000
		3001	=item C<EV_P>, C<EV_P_>
		3002
		3003	This provides the loop I<parameter> for functions, if one is required ("ev
		3004	loop parameter"). The C<EV_P> form is used when this is the sole parameter,
		3005	C<EV_P_> is used when other parameters are following. Example:
		3006
		3007	// this is how ev_unref is being declared
		3008	static void ev_unref (EV_P);
		3009
		3010	// this is how you can declare your typical callback
		3011	static void cb (EV_P_ ev_timer *w, int revents)
		3012
		3013	It declares a parameter C<loop> of type C<struct ev_loop *>, quite
		3014	suitable for use with C<EV_A>.
		3015
		3016	=item C<EV_DEFAULT>, C<EV_DEFAULT_>
		3017
		3018	Similar to the other two macros, this gives you the value of the default
		3019	loop, if multiple loops are supported ("ev loop default").
		3020
		3021	=item C<EV_DEFAULT_UC>, C<EV_DEFAULT_UC_>
		3022
		3023	Usage identical to C<EV_DEFAULT> and C<EV_DEFAULT_>, but requires that the
		3024	default loop has been initialised (C<UC> == unchecked). Their behaviour
		3025	is undefined when the default loop has not been initialised by a previous
		3026	execution of C<EV_DEFAULT>, C<EV_DEFAULT_> or C<ev_default_init (...)>.
		3027
		3028	It is often prudent to use C<EV_DEFAULT> when initialising the first
		3029	watcher in a function but use C<EV_DEFAULT_UC> afterwards.
		3030
		3031	=back
		3032
		3033	Example: Declare and initialise a check watcher, utilising the above
		3034	macros so it will work regardless of whether multiple loops are supported
		3035	or not.
		3036
		3037	static void
		3038	check_cb (EV_P_ ev_timer *w, int revents)
		3039	{
		3040	ev_check_stop (EV_A_ w);
1417	}	3041	}
		3042
		3043	ev_check check;
		3044	ev_check_init (&check, check_cb);
		3045	ev_check_start (EV_DEFAULT_ &check);
		3046	ev_loop (EV_DEFAULT_ 0);
		3047
		3048	=head1 EMBEDDING
		3049
		3050	Libev can (and often is) directly embedded into host
		3051	applications. Examples of applications that embed it include the Deliantra
		3052	Game Server, the EV perl module, the GNU Virtual Private Ethernet (gvpe)
		3053	and rxvt-unicode.
		3054
		3055	The goal is to enable you to just copy the necessary files into your
		3056	source directory without having to change even a single line in them, so
		3057	you can easily upgrade by simply copying (or having a checked-out copy of
		3058	libev somewhere in your source tree).
		3059
		3060	=head2 FILESETS
		3061
		3062	Depending on what features you need you need to include one or more sets of files
		3063	in your application.
		3064
		3065	=head3 CORE EVENT LOOP
		3066
		3067	To include only the libev core (all the C<ev_*> functions), with manual
		3068	configuration (no autoconf):
		3069
		3070	#define EV_STANDALONE 1
		3071	#include "ev.c"
		3072
		3073	This will automatically include F<ev.h>, too, and should be done in a
		3074	single C source file only to provide the function implementations. To use
		3075	it, do the same for F<ev.h> in all files wishing to use this API (best
		3076	done by writing a wrapper around F<ev.h> that you can include instead and
		3077	where you can put other configuration options):
		3078
		3079	#define EV_STANDALONE 1
		3080	#include "ev.h"
		3081
		3082	Both header files and implementation files can be compiled with a C++
		3083	compiler (at least, thats a stated goal, and breakage will be treated
		3084	as a bug).
		3085
		3086	You need the following files in your source tree, or in a directory
		3087	in your include path (e.g. in libev/ when using -Ilibev):
		3088
		3089	ev.h
		3090	ev.c
		3091	ev_vars.h
		3092	ev_wrap.h
		3093
		3094	ev_win32.c required on win32 platforms only
		3095
		3096	ev_select.c only when select backend is enabled (which is enabled by default)
		3097	ev_poll.c only when poll backend is enabled (disabled by default)
		3098	ev_epoll.c only when the epoll backend is enabled (disabled by default)
		3099	ev_kqueue.c only when the kqueue backend is enabled (disabled by default)
		3100	ev_port.c only when the solaris port backend is enabled (disabled by default)
		3101
		3102	F<ev.c> includes the backend files directly when enabled, so you only need
		3103	to compile this single file.
		3104
		3105	=head3 LIBEVENT COMPATIBILITY API
		3106
		3107	To include the libevent compatibility API, also include:
		3108
		3109	#include "event.c"
		3110
		3111	in the file including F<ev.c>, and:
		3112
		3113	#include "event.h"
		3114
		3115	in the files that want to use the libevent API. This also includes F<ev.h>.
		3116
		3117	You need the following additional files for this:
		3118
		3119	event.h
		3120	event.c
		3121
		3122	=head3 AUTOCONF SUPPORT
		3123
		3124	Instead of using C<EV_STANDALONE=1> and providing your configuration in
		3125	whatever way you want, you can also C<m4_include([libev.m4])> in your
		3126	F<configure.ac> and leave C<EV_STANDALONE> undefined. F<ev.c> will then
		3127	include F<config.h> and configure itself accordingly.
		3128
		3129	For this of course you need the m4 file:
		3130
		3131	libev.m4
		3132
		3133	=head2 PREPROCESSOR SYMBOLS/MACROS
		3134
		3135	Libev can be configured via a variety of preprocessor symbols you have to
		3136	define before including any of its files. The default in the absence of
		3137	autoconf is documented for every option.
		3138
		3139	=over 4
		3140
		3141	=item EV_STANDALONE
		3142
		3143	Must always be C<1> if you do not use autoconf configuration, which
		3144	keeps libev from including F<config.h>, and it also defines dummy
		3145	implementations for some libevent functions (such as logging, which is not
		3146	supported). It will also not define any of the structs usually found in
		3147	F<event.h> that are not directly supported by the libev core alone.
		3148
		3149	=item EV_USE_MONOTONIC
		3150
		3151	If defined to be C<1>, libev will try to detect the availability of the
		3152	monotonic clock option at both compile time and runtime. Otherwise no use
		3153	of the monotonic clock option will be attempted. If you enable this, you
		3154	usually have to link against librt or something similar. Enabling it when
		3155	the functionality isn't available is safe, though, although you have
		3156	to make sure you link against any libraries where the C<clock_gettime>
		3157	function is hiding in (often F<-lrt>).
		3158
		3159	=item EV_USE_REALTIME
		3160
		3161	If defined to be C<1>, libev will try to detect the availability of the
		3162	real-time clock option at compile time (and assume its availability at
		3163	runtime if successful). Otherwise no use of the real-time clock option will
		3164	be attempted. This effectively replaces C<gettimeofday> by C<clock_get
		3165	(CLOCK_REALTIME, ...)> and will not normally affect correctness. See the
		3166	note about libraries in the description of C<EV_USE_MONOTONIC>, though.
		3167
		3168	=item EV_USE_NANOSLEEP
		3169
		3170	If defined to be C<1>, libev will assume that C<nanosleep ()> is available
		3171	and will use it for delays. Otherwise it will use C<select ()>.
		3172
		3173	=item EV_USE_EVENTFD
		3174
		3175	If defined to be C<1>, then libev will assume that C<eventfd ()> is
		3176	available and will probe for kernel support at runtime. This will improve
		3177	C<ev_signal> and C<ev_async> performance and reduce resource consumption.
		3178	If undefined, it will be enabled if the headers indicate GNU/Linux + Glibc
		3179	2.7 or newer, otherwise disabled.
		3180
		3181	=item EV_USE_SELECT
		3182
		3183	If undefined or defined to be C<1>, libev will compile in support for the
		3184	C<select>(2) backend. No attempt at auto-detection will be done: if no
		3185	other method takes over, select will be it. Otherwise the select backend
		3186	will not be compiled in.
		3187
		3188	=item EV_SELECT_USE_FD_SET
		3189
		3190	If defined to C<1>, then the select backend will use the system C<fd_set>
		3191	structure. This is useful if libev doesn't compile due to a missing
		3192	C<NFDBITS> or C<fd_mask> definition or it mis-guesses the bitset layout on
		3193	exotic systems. This usually limits the range of file descriptors to some
		3194	low limit such as 1024 or might have other limitations (winsocket only
		3195	allows 64 sockets). The C<FD_SETSIZE> macro, set before compilation, might
		3196	influence the size of the C<fd_set> used.
		3197
		3198	=item EV_SELECT_IS_WINSOCKET
		3199
		3200	When defined to C<1>, the select backend will assume that
		3201	select/socket/connect etc. don't understand file descriptors but
		3202	wants osf handles on win32 (this is the case when the select to
		3203	be used is the winsock select). This means that it will call
		3204	C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,
		3205	it is assumed that all these functions actually work on fds, even
		3206	on win32. Should not be defined on non-win32 platforms.
		3207
		3208	=item EV_FD_TO_WIN32_HANDLE
		3209
		3210	If C<EV_SELECT_IS_WINSOCKET> is enabled, then libev needs a way to map
		3211	file descriptors to socket handles. When not defining this symbol (the
		3212	default), then libev will call C<_get_osfhandle>, which is usually
		3213	correct. In some cases, programs use their own file descriptor management,
		3214	in which case they can provide this function to map fds to socket handles.
		3215
		3216	=item EV_USE_POLL
		3217
		3218	If defined to be C<1>, libev will compile in support for the C<poll>(2)
		3219	backend. Otherwise it will be enabled on non-win32 platforms. It
		3220	takes precedence over select.
		3221
		3222	=item EV_USE_EPOLL
		3223
		3224	If defined to be C<1>, libev will compile in support for the Linux
		3225	C<epoll>(7) backend. Its availability will be detected at runtime,
		3226	otherwise another method will be used as fallback. This is the preferred
		3227	backend for GNU/Linux systems. If undefined, it will be enabled if the
		3228	headers indicate GNU/Linux + Glibc 2.4 or newer, otherwise disabled.
		3229
		3230	=item EV_USE_KQUEUE
		3231
		3232	If defined to be C<1>, libev will compile in support for the BSD style
		3233	C<kqueue>(2) backend. Its actual availability will be detected at runtime,
		3234	otherwise another method will be used as fallback. This is the preferred
		3235	backend for BSD and BSD-like systems, although on most BSDs kqueue only
		3236	supports some types of fds correctly (the only platform we found that
		3237	supports ptys for example was NetBSD), so kqueue might be compiled in, but
		3238	not be used unless explicitly requested. The best way to use it is to find
		3239	out whether kqueue supports your type of fd properly and use an embedded
		3240	kqueue loop.
		3241
		3242	=item EV_USE_PORT
		3243
		3244	If defined to be C<1>, libev will compile in support for the Solaris
		3245	10 port style backend. Its availability will be detected at runtime,
		3246	otherwise another method will be used as fallback. This is the preferred
		3247	backend for Solaris 10 systems.
		3248
		3249	=item EV_USE_DEVPOLL
		3250
		3251	Reserved for future expansion, works like the USE symbols above.
		3252
		3253	=item EV_USE_INOTIFY
		3254
		3255	If defined to be C<1>, libev will compile in support for the Linux inotify
		3256	interface to speed up C<ev_stat> watchers. Its actual availability will
		3257	be detected at runtime. If undefined, it will be enabled if the headers
		3258	indicate GNU/Linux + Glibc 2.4 or newer, otherwise disabled.
		3259
		3260	=item EV_ATOMIC_T
		3261
		3262	Libev requires an integer type (suitable for storing C<0> or C<1>) whose
		3263	access is atomic with respect to other threads or signal contexts. No such
		3264	type is easily found in the C language, so you can provide your own type
		3265	that you know is safe for your purposes. It is used both for signal handler "locking"
		3266	as well as for signal and thread safety in C<ev_async> watchers.
		3267
		3268	In the absence of this define, libev will use C<sig_atomic_t volatile>
		3269	(from F<signal.h>), which is usually good enough on most platforms.
		3270
		3271	=item EV_H
		3272
		3273	The name of the F<ev.h> header file used to include it. The default if
		3274	undefined is C<"ev.h"> in F<event.h>, F<ev.c> and F<ev++.h>. This can be
		3275	used to virtually rename the F<ev.h> header file in case of conflicts.
		3276
		3277	=item EV_CONFIG_H
		3278
		3279	If C<EV_STANDALONE> isn't C<1>, this variable can be used to override
		3280	F<ev.c>'s idea of where to find the F<config.h> file, similarly to
		3281	C<EV_H>, above.
		3282
		3283	=item EV_EVENT_H
		3284
		3285	Similarly to C<EV_H>, this macro can be used to override F<event.c>'s idea
		3286	of how the F<event.h> header can be found, the default is C<"event.h">.
		3287
		3288	=item EV_PROTOTYPES
		3289
		3290	If defined to be C<0>, then F<ev.h> will not define any function
		3291	prototypes, but still define all the structs and other symbols. This is
		3292	occasionally useful if you want to provide your own wrapper functions
		3293	around libev functions.
		3294
		3295	=item EV_MULTIPLICITY
		3296
		3297	If undefined or defined to C<1>, then all event-loop-specific functions
		3298	will have the C<struct ev_loop *> as first argument, and you can create
		3299	additional independent event loops. Otherwise there will be no support
		3300	for multiple event loops and there is no first event loop pointer
		3301	argument. Instead, all functions act on the single default loop.
		3302
		3303	=item EV_MINPRI
		3304
		3305	=item EV_MAXPRI
		3306
		3307	The range of allowed priorities. C<EV_MINPRI> must be smaller or equal to
		3308	C<EV_MAXPRI>, but otherwise there are no non-obvious limitations. You can
		3309	provide for more priorities by overriding those symbols (usually defined
		3310	to be C<-2> and C<2>, respectively).
		3311
		3312	When doing priority-based operations, libev usually has to linearly search
		3313	all the priorities, so having many of them (hundreds) uses a lot of space
		3314	and time, so using the defaults of five priorities (-2 .. +2) is usually
		3315	fine.
		3316
		3317	If your embedding application does not need any priorities, defining these
		3318	both to C<0> will save some memory and CPU.
		3319
		3320	=item EV_PERIODIC_ENABLE
		3321
		3322	If undefined or defined to be C<1>, then periodic timers are supported. If
		3323	defined to be C<0>, then they are not. Disabling them saves a few kB of
		3324	code.
		3325
		3326	=item EV_IDLE_ENABLE
		3327
		3328	If undefined or defined to be C<1>, then idle watchers are supported. If
		3329	defined to be C<0>, then they are not. Disabling them saves a few kB of
		3330	code.
		3331
		3332	=item EV_EMBED_ENABLE
		3333
		3334	If undefined or defined to be C<1>, then embed watchers are supported. If
		3335	defined to be C<0>, then they are not. Embed watchers rely on most other
		3336	watcher types, which therefore must not be disabled.
		3337
		3338	=item EV_STAT_ENABLE
		3339
		3340	If undefined or defined to be C<1>, then stat watchers are supported. If
		3341	defined to be C<0>, then they are not.
		3342
		3343	=item EV_FORK_ENABLE
		3344
		3345	If undefined or defined to be C<1>, then fork watchers are supported. If
		3346	defined to be C<0>, then they are not.
		3347
		3348	=item EV_ASYNC_ENABLE
		3349
		3350	If undefined or defined to be C<1>, then async watchers are supported. If
		3351	defined to be C<0>, then they are not.
		3352
		3353	=item EV_MINIMAL
		3354
		3355	If you need to shave off some kilobytes of code at the expense of some
		3356	speed, define this symbol to C<1>. Currently this is used to override some
		3357	inlining decisions, saves roughly 30% code size on amd64. It also selects a
		3358	much smaller 2-heap for timer management over the default 4-heap.
		3359
		3360	=item EV_PID_HASHSIZE
		3361
		3362	C<ev_child> watchers use a small hash table to distribute workload by
		3363	pid. The default size is C<16> (or C<1> with C<EV_MINIMAL>), usually more
		3364	than enough. If you need to manage thousands of children you might want to
		3365	increase this value (I<must> be a power of two).
		3366
		3367	=item EV_INOTIFY_HASHSIZE
		3368
		3369	C<ev_stat> watchers use a small hash table to distribute workload by
		3370	inotify watch id. The default size is C<16> (or C<1> with C<EV_MINIMAL>),
		3371	usually more than enough. If you need to manage thousands of C<ev_stat>
		3372	watchers you might want to increase this value (I<must> be a power of
		3373	two).
		3374
		3375	=item EV_USE_4HEAP
		3376
		3377	Heaps are not very cache-efficient. To improve the cache-efficiency of the
		3378	timer and periodics heaps, libev uses a 4-heap when this symbol is defined
		3379	to C<1>. The 4-heap uses more complicated (longer) code but has noticeably
		3380	faster performance with many (thousands) of watchers.
		3381
		3382	The default is C<1> unless C<EV_MINIMAL> is set in which case it is C<0>
		3383	(disabled).
		3384
		3385	=item EV_HEAP_CACHE_AT
		3386
		3387	Heaps are not very cache-efficient. To improve the cache-efficiency of the
		3388	timer and periodics heaps, libev can cache the timestamp (I<at>) within
		3389	the heap structure (selected by defining C<EV_HEAP_CACHE_AT> to C<1>),
		3390	which uses 8-12 bytes more per watcher and a few hundred bytes more code,
		3391	but avoids random read accesses on heap changes. This improves performance
		3392	noticeably with many (hundreds) of watchers.
		3393
		3394	The default is C<1> unless C<EV_MINIMAL> is set in which case it is C<0>
		3395	(disabled).
		3396
		3397	=item EV_VERIFY
		3398
		3399	Controls how much internal verification (see C<ev_loop_verify ()>) will
		3400	be done: If set to C<0>, no internal verification code will be compiled
		3401	in. If set to C<1>, then verification code will be compiled in, but not
		3402	called. If set to C<2>, then the internal verification code will be
		3403	called once per loop, which can slow down libev. If set to C<3>, then the
		3404	verification code will be called very frequently, which will slow down
		3405	libev considerably.
		3406
		3407	The default is C<1>, unless C<EV_MINIMAL> is set, in which case it will be
		3408	C<0>.
		3409
		3410	=item EV_COMMON
		3411
		3412	By default, all watchers have a C<void *data> member. By redefining
		3413	this macro to a something else you can include more and other types of
		3414	members. You have to define it each time you include one of the files,
		3415	though, and it must be identical each time.
		3416
		3417	For example, the perl EV module uses something like this:
		3418
		3419	#define EV_COMMON \
		3420	SV *self; /* contains this struct */ \
		3421	SV *cb_sv, *fh /* note no trailing ";" */
		3422
		3423	=item EV_CB_DECLARE (type)
		3424
		3425	=item EV_CB_INVOKE (watcher, revents)
		3426
		3427	=item ev_set_cb (ev, cb)
		3428
		3429	Can be used to change the callback member declaration in each watcher,
		3430	and the way callbacks are invoked and set. Must expand to a struct member
		3431	definition and a statement, respectively. See the F<ev.h> header file for
		3432	their default definitions. One possible use for overriding these is to
		3433	avoid the C<struct ev_loop *> as first argument in all cases, or to use
		3434	method calls instead of plain function calls in C++.
		3435
		3436	=back
		3437
		3438	=head2 EXPORTED API SYMBOLS
		3439
		3440	If you need to re-export the API (e.g. via a DLL) and you need a list of
		3441	exported symbols, you can use the provided F<Symbol.*> files which list
		3442	all public symbols, one per line:
		3443
		3444	Symbols.ev for libev proper
		3445	Symbols.event for the libevent emulation
		3446
		3447	This can also be used to rename all public symbols to avoid clashes with
		3448	multiple versions of libev linked together (which is obviously bad in
		3449	itself, but sometimes it is inconvenient to avoid this).
		3450
		3451	A sed command like this will create wrapper C<#define>'s that you need to
		3452	include before including F<ev.h>:
		3453
		3454	<Symbols.ev sed -e "s/.*/#define & myprefix_&/" >wrap.h
		3455
		3456	This would create a file F<wrap.h> which essentially looks like this:
		3457
		3458	#define ev_backend myprefix_ev_backend
		3459	#define ev_check_start myprefix_ev_check_start
		3460	#define ev_check_stop myprefix_ev_check_stop
		3461	...
		3462
		3463	=head2 EXAMPLES
		3464
		3465	For a real-world example of a program the includes libev
		3466	verbatim, you can have a look at the EV perl module
		3467	(L<http://software.schmorp.de/pkg/EV.html>). It has the libev files in
		3468	the F<libev/> subdirectory and includes them in the F<EV/EVAPI.h> (public
		3469	interface) and F<EV.xs> (implementation) files. Only the F<EV.xs> file
		3470	will be compiled. It is pretty complex because it provides its own header
		3471	file.
		3472
		3473	The usage in rxvt-unicode is simpler. It has a F<ev_cpp.h> header file
		3474	that everybody includes and which overrides some configure choices:
		3475
		3476	#define EV_MINIMAL 1
		3477	#define EV_USE_POLL 0
		3478	#define EV_MULTIPLICITY 0
		3479	#define EV_PERIODIC_ENABLE 0
		3480	#define EV_STAT_ENABLE 0
		3481	#define EV_FORK_ENABLE 0
		3482	#define EV_CONFIG_H <config.h>
		3483	#define EV_MINPRI 0
		3484	#define EV_MAXPRI 0
		3485
		3486	#include "ev++.h"
		3487
		3488	And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:
		3489
		3490	#include "ev_cpp.h"
		3491	#include "ev.c"
		3492
		3493	=head1 INTERACTION WITH OTHER PROGRAMS OR LIBRARIES
		3494
		3495	=head2 THREADS AND COROUTINES
		3496
		3497	=head3 THREADS
		3498
		3499	All libev functions are reentrant and thread-safe unless explicitly
		3500	documented otherwise, but libev implements no locking itself. This means
		3501	that you can use as many loops as you want in parallel, as long as there
		3502	are no concurrent calls into any libev function with the same loop
		3503	parameter (C<ev_default_*> calls have an implicit default loop parameter,
		3504	of course): libev guarantees that different event loops share no data
		3505	structures that need any locking.
		3506
		3507	Or to put it differently: calls with different loop parameters can be done
		3508	concurrently from multiple threads, calls with the same loop parameter
		3509	must be done serially (but can be done from different threads, as long as
		3510	only one thread ever is inside a call at any point in time, e.g. by using
		3511	a mutex per loop).
		3512
		3513	Specifically to support threads (and signal handlers), libev implements
		3514	so-called C<ev_async> watchers, which allow some limited form of
		3515	concurrency on the same event loop, namely waking it up "from the
		3516	outside".
		3517
		3518	If you want to know which design (one loop, locking, or multiple loops
		3519	without or something else still) is best for your problem, then I cannot
		3520	help you, but here is some generic advice:
		3521
		3522	=over 4
		3523
		3524	=item * most applications have a main thread: use the default libev loop
		3525	in that thread, or create a separate thread running only the default loop.
		3526
		3527	This helps integrating other libraries or software modules that use libev
		3528	themselves and don't care/know about threading.
		3529
		3530	=item * one loop per thread is usually a good model.
		3531
		3532	Doing this is almost never wrong, sometimes a better-performance model
		3533	exists, but it is always a good start.
		3534
		3535	=item * other models exist, such as the leader/follower pattern, where one
		3536	loop is handed through multiple threads in a kind of round-robin fashion.
		3537
		3538	Choosing a model is hard - look around, learn, know that usually you can do
		3539	better than you currently do :-)
		3540
		3541	=item * often you need to talk to some other thread which blocks in the
		3542	event loop.
		3543
		3544	C<ev_async> watchers can be used to wake them up from other threads safely
		3545	(or from signal contexts...).
		3546
		3547	An example use would be to communicate signals or other events that only
		3548	work in the default loop by registering the signal watcher with the
		3549	default loop and triggering an C<ev_async> watcher from the default loop
		3550	watcher callback into the event loop interested in the signal.
		3551
		3552	=back
		3553
		3554	=head3 COROUTINES
		3555
		3556	Libev is very accommodating to coroutines ("cooperative threads"):
		3557	libev fully supports nesting calls to its functions from different
		3558	coroutines (e.g. you can call C<ev_loop> on the same loop from two
		3559	different coroutines, and switch freely between both coroutines running the
		3560	loop, as long as you don't confuse yourself). The only exception is that
		3561	you must not do this from C<ev_periodic> reschedule callbacks.
		3562
		3563	Care has been taken to ensure that libev does not keep local state inside
		3564	C<ev_loop>, and other calls do not usually allow for coroutine switches as
		3565	they do not clal any callbacks.
		3566
		3567	=head2 COMPILER WARNINGS
		3568
		3569	Depending on your compiler and compiler settings, you might get no or a
		3570	lot of warnings when compiling libev code. Some people are apparently
		3571	scared by this.
		3572
		3573	However, these are unavoidable for many reasons. For one, each compiler
		3574	has different warnings, and each user has different tastes regarding
		3575	warning options. "Warn-free" code therefore cannot be a goal except when
		3576	targeting a specific compiler and compiler-version.
		3577
		3578	Another reason is that some compiler warnings require elaborate
		3579	workarounds, or other changes to the code that make it less clear and less
		3580	maintainable.
		3581
		3582	And of course, some compiler warnings are just plain stupid, or simply
		3583	wrong (because they don't actually warn about the condition their message
		3584	seems to warn about). For example, certain older gcc versions had some
		3585	warnings that resulted an extreme number of false positives. These have
		3586	been fixed, but some people still insist on making code warn-free with
		3587	such buggy versions.
		3588
		3589	While libev is written to generate as few warnings as possible,
		3590	"warn-free" code is not a goal, and it is recommended not to build libev
		3591	with any compiler warnings enabled unless you are prepared to cope with
		3592	them (e.g. by ignoring them). Remember that warnings are just that:
		3593	warnings, not errors, or proof of bugs.
		3594
		3595
		3596	=head2 VALGRIND
		3597
		3598	Valgrind has a special section here because it is a popular tool that is
		3599	highly useful. Unfortunately, valgrind reports are very hard to interpret.
		3600
		3601	If you think you found a bug (memory leak, uninitialised data access etc.)
		3602	in libev, then check twice: If valgrind reports something like:
		3603
		3604	==2274== definitely lost: 0 bytes in 0 blocks.
		3605	==2274== possibly lost: 0 bytes in 0 blocks.
		3606	==2274== still reachable: 256 bytes in 1 blocks.
		3607
		3608	Then there is no memory leak, just as memory accounted to global variables
		3609	is not a memleak - the memory is still being refernced, and didn't leak.
		3610
		3611	Similarly, under some circumstances, valgrind might report kernel bugs
		3612	as if it were a bug in libev (e.g. in realloc or in the poll backend,
		3613	although an acceptable workaround has been found here), or it might be
		3614	confused.
		3615
		3616	Keep in mind that valgrind is a very good tool, but only a tool. Don't
		3617	make it into some kind of religion.
		3618
		3619	If you are unsure about something, feel free to contact the mailing list
		3620	with the full valgrind report and an explanation on why you think this
		3621	is a bug in libev (best check the archives, too :). However, don't be
		3622	annoyed when you get a brisk "this is no bug" answer and take the chance
		3623	of learning how to interpret valgrind properly.
		3624
		3625	If you need, for some reason, empty reports from valgrind for your project
		3626	I suggest using suppression lists.
		3627
		3628
		3629	=head1 PORTABILITY NOTES
		3630
		3631	=head2 WIN32 PLATFORM LIMITATIONS AND WORKAROUNDS
		3632
		3633	Win32 doesn't support any of the standards (e.g. POSIX) that libev
		3634	requires, and its I/O model is fundamentally incompatible with the POSIX
		3635	model. Libev still offers limited functionality on this platform in
		3636	the form of the C<EVBACKEND_SELECT> backend, and only supports socket
		3637	descriptors. This only applies when using Win32 natively, not when using
		3638	e.g. cygwin.
		3639
		3640	Lifting these limitations would basically require the full
		3641	re-implementation of the I/O system. If you are into these kinds of
		3642	things, then note that glib does exactly that for you in a very portable
		3643	way (note also that glib is the slowest event library known to man).
		3644
		3645	There is no supported compilation method available on windows except
		3646	embedding it into other applications.
		3647
		3648	Not a libev limitation but worth mentioning: windows apparently doesn't
		3649	accept large writes: instead of resulting in a partial write, windows will
		3650	either accept everything or return C<ENOBUFS> if the buffer is too large,
		3651	so make sure you only write small amounts into your sockets (less than a
		3652	megabyte seems safe, but this apparently depends on the amount of memory
		3653	available).
		3654
		3655	Due to the many, low, and arbitrary limits on the win32 platform and
		3656	the abysmal performance of winsockets, using a large number of sockets
		3657	is not recommended (and not reasonable). If your program needs to use
		3658	more than a hundred or so sockets, then likely it needs to use a totally
		3659	different implementation for windows, as libev offers the POSIX readiness
		3660	notification model, which cannot be implemented efficiently on windows
		3661	(Microsoft monopoly games).
		3662
		3663	A typical way to use libev under windows is to embed it (see the embedding
		3664	section for details) and use the following F<evwrap.h> header file instead
		3665	of F<ev.h>:
		3666
		3667	#define EV_STANDALONE /* keeps ev from requiring config.h */
		3668	#define EV_SELECT_IS_WINSOCKET 1 /* configure libev for windows select */
		3669
		3670	#include "ev.h"
		3671
		3672	And compile the following F<evwrap.c> file into your project (make sure
		3673	you do I<not> compile the F<ev.c> or any other embedded source files!):
		3674
		3675	#include "evwrap.h"
		3676	#include "ev.c"
		3677
		3678	=over 4
		3679
		3680	=item The winsocket select function
		3681
		3682	The winsocket C<select> function doesn't follow POSIX in that it
		3683	requires socket I<handles> and not socket I<file descriptors> (it is
		3684	also extremely buggy). This makes select very inefficient, and also
		3685	requires a mapping from file descriptors to socket handles (the Microsoft
		3686	C runtime provides the function C<_open_osfhandle> for this). See the
		3687	discussion of the C<EV_SELECT_USE_FD_SET>, C<EV_SELECT_IS_WINSOCKET> and
		3688	C<EV_FD_TO_WIN32_HANDLE> preprocessor symbols for more info.
		3689
		3690	The configuration for a "naked" win32 using the Microsoft runtime
		3691	libraries and raw winsocket select is:
		3692
		3693	#define EV_USE_SELECT 1
		3694	#define EV_SELECT_IS_WINSOCKET 1 /* forces EV_SELECT_USE_FD_SET, too */
		3695
		3696	Note that winsockets handling of fd sets is O(n), so you can easily get a
		3697	complexity in the O(n²) range when using win32.
		3698
		3699	=item Limited number of file descriptors
		3700
		3701	Windows has numerous arbitrary (and low) limits on things.
		3702
		3703	Early versions of winsocket's select only supported waiting for a maximum
		3704	of C<64> handles (probably owning to the fact that all windows kernels
		3705	can only wait for C<64> things at the same time internally; Microsoft
		3706	recommends spawning a chain of threads and wait for 63 handles and the
		3707	previous thread in each. Great).
		3708
		3709	Newer versions support more handles, but you need to define C<FD_SETSIZE>
		3710	to some high number (e.g. C<2048>) before compiling the winsocket select
		3711	call (which might be in libev or elsewhere, for example, perl does its own
		3712	select emulation on windows).
		3713
		3714	Another limit is the number of file descriptors in the Microsoft runtime
		3715	libraries, which by default is C<64> (there must be a hidden I<64> fetish
		3716	or something like this inside Microsoft). You can increase this by calling
		3717	C<_setmaxstdio>, which can increase this limit to C<2048> (another
		3718	arbitrary limit), but is broken in many versions of the Microsoft runtime
		3719	libraries.
		3720
		3721	This might get you to about C<512> or C<2048> sockets (depending on
		3722	windows version and/or the phase of the moon). To get more, you need to
		3723	wrap all I/O functions and provide your own fd management, but the cost of
		3724	calling select (O(n²)) will likely make this unworkable.
		3725
		3726	=back
		3727
		3728	=head2 PORTABILITY REQUIREMENTS
		3729
		3730	In addition to a working ISO-C implementation and of course the
		3731	backend-specific APIs, libev relies on a few additional extensions:
		3732
		3733	=over 4
		3734
		3735	=item C<void (*)(ev_watcher_type *, int revents)> must have compatible
		3736	calling conventions regardless of C<ev_watcher_type *>.
		3737
		3738	Libev assumes not only that all watcher pointers have the same internal
		3739	structure (guaranteed by POSIX but not by ISO C for example), but it also
		3740	assumes that the same (machine) code can be used to call any watcher
		3741	callback: The watcher callbacks have different type signatures, but libev
		3742	calls them using an C<ev_watcher *> internally.
		3743
		3744	=item C<sig_atomic_t volatile> must be thread-atomic as well
		3745
		3746	The type C<sig_atomic_t volatile> (or whatever is defined as
		3747	C<EV_ATOMIC_T>) must be atomic with respect to accesses from different
		3748	threads. This is not part of the specification for C<sig_atomic_t>, but is
		3749	believed to be sufficiently portable.
		3750
		3751	=item C<sigprocmask> must work in a threaded environment
		3752
		3753	Libev uses C<sigprocmask> to temporarily block signals. This is not
		3754	allowed in a threaded program (C<pthread_sigmask> has to be used). Typical
		3755	pthread implementations will either allow C<sigprocmask> in the "main
		3756	thread" or will block signals process-wide, both behaviours would
		3757	be compatible with libev. Interaction between C<sigprocmask> and
		3758	C<pthread_sigmask> could complicate things, however.
		3759
		3760	The most portable way to handle signals is to block signals in all threads
		3761	except the initial one, and run the default loop in the initial thread as
		3762	well.
		3763
		3764	=item C<long> must be large enough for common memory allocation sizes
		3765
		3766	To improve portability and simplify its API, libev uses C<long> internally
		3767	instead of C<size_t> when allocating its data structures. On non-POSIX
		3768	systems (Microsoft...) this might be unexpectedly low, but is still at
		3769	least 31 bits everywhere, which is enough for hundreds of millions of
		3770	watchers.
		3771
		3772	=item C<double> must hold a time value in seconds with enough accuracy
		3773
		3774	The type C<double> is used to represent timestamps. It is required to
		3775	have at least 51 bits of mantissa (and 9 bits of exponent), which is good
		3776	enough for at least into the year 4000. This requirement is fulfilled by
		3777	implementations implementing IEEE 754 (basically all existing ones).
		3778
		3779	=back
		3780
		3781	If you know of other additional requirements drop me a note.
		3782
		3783
		3784	=head1 ALGORITHMIC COMPLEXITIES
		3785
		3786	In this section the complexities of (many of) the algorithms used inside
		3787	libev will be documented. For complexity discussions about backends see
		3788	the documentation for C<ev_default_init>.
		3789
		3790	All of the following are about amortised time: If an array needs to be
		3791	extended, libev needs to realloc and move the whole array, but this
		3792	happens asymptotically rarer with higher number of elements, so O(1) might
		3793	mean that libev does a lengthy realloc operation in rare cases, but on
		3794	average it is much faster and asymptotically approaches constant time.
		3795
		3796	=over 4
		3797
		3798	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)
		3799
		3800	This means that, when you have a watcher that triggers in one hour and
		3801	there are 100 watchers that would trigger before that, then inserting will
		3802	have to skip roughly seven (C<ld 100>) of these watchers.
		3803
		3804	=item Changing timer/periodic watchers (by autorepeat or calling again): O(log skipped_other_timers)
		3805
		3806	That means that changing a timer costs less than removing/adding them,
		3807	as only the relative motion in the event queue has to be paid for.
		3808
		3809	=item Starting io/check/prepare/idle/signal/child/fork/async watchers: O(1)
		3810
		3811	These just add the watcher into an array or at the head of a list.
		3812
		3813	=item Stopping check/prepare/idle/fork/async watchers: O(1)
		3814
		3815	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))
		3816
		3817	These watchers are stored in lists, so they need to be walked to find the
		3818	correct watcher to remove. The lists are usually short (you don't usually
		3819	have many watchers waiting for the same fd or signal: one is typical, two
		3820	is rare).
		3821
		3822	=item Finding the next timer in each loop iteration: O(1)
		3823
		3824	By virtue of using a binary or 4-heap, the next timer is always found at a
		3825	fixed position in the storage array.
		3826
		3827	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)
		3828
		3829	A change means an I/O watcher gets started or stopped, which requires
		3830	libev to recalculate its status (and possibly tell the kernel, depending
		3831	on backend and whether C<ev_io_set> was used).
		3832
		3833	=item Activating one watcher (putting it into the pending state): O(1)
		3834
		3835	=item Priority handling: O(number_of_priorities)
		3836
		3837	Priorities are implemented by allocating some space for each
		3838	priority. When doing priority-based operations, libev usually has to
		3839	linearly search all the priorities, but starting/stopping and activating
		3840	watchers becomes O(1) with respect to priority handling.
		3841
		3842	=item Sending an ev_async: O(1)
		3843
		3844	=item Processing ev_async_send: O(number_of_async_watchers)
		3845
		3846	=item Processing signals: O(max_signal_number)
		3847
		3848	Sending involves a system call I<iff> there were no other C<ev_async_send>
		3849	calls in the current loop iteration. Checking for async and signal events
		3850	involves iterating over all running async watchers or all signal numbers.
		3851
		3852	=back
		3853
1418		3854
1419	=head1 AUTHOR	3855	=head1 AUTHOR
1420		3856
1421	Marc Lehmann <libev@schmorp.de>.	3857	Marc Lehmann <libev@schmorp.de>.
1422		3858

Diff Legend

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