ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/libev/ev.pod

Comparing libev/ev.pod (file contents):
Revision 1.97 by root, Sat Dec 22 05:48:02 2007 UTC vs.
Revision 1.171 by root, Tue Jul 8 09:49:15 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		8
9	=head1 EXAMPLE PROGRAM	9	=head2 EXAMPLE PROGRAM
10		10
		11	// a single header file is required
11	#include <ev.h>	12	#include <ev.h>
12		13
		14	// every watcher type has its own typedef'd struct
		15	// with the name ev_<type>
13	ev_io stdin_watcher;	16	ev_io stdin_watcher;
14	ev_timer timeout_watcher;	17	ev_timer timeout_watcher;
15		18
16	/* called when data readable on stdin */	19	// all watcher callbacks have a similar signature
		20	// this callback is called when data is readable on stdin
17	static void	21	static void
18	stdin_cb (EV_P_ struct ev_io *w, int revents)	22	stdin_cb (EV_P_ struct ev_io *w, int revents)
19	{	23	{
20	/* puts ("stdin ready"); */	24	puts ("stdin ready");
21	ev_io_stop (EV_A_ w); /* just a syntax example */	25	// for one-shot events, one must manually stop the watcher
22	ev_unloop (EV_A_ EVUNLOOP_ALL); /* leave all loop calls */	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);
23	}	31	}
24		32
		33	// another callback, this time for a time-out
25	static void	34	static void
26	timeout_cb (EV_P_ struct ev_timer *w, int revents)	35	timeout_cb (EV_P_ struct ev_timer *w, int revents)
27	{	36	{
28	/* puts ("timeout"); */	37	puts ("timeout");
29	ev_unloop (EV_A_ EVUNLOOP_ONE); /* leave one loop call */	38	// this causes the innermost ev_loop to stop iterating
		39	ev_unloop (EV_A_ EVUNLOOP_ONE);
30	}	40	}
31		41
32	int	42	int
33	main (void)	43	main (void)
34	{	44	{
		45	// use the default event loop unless you have special needs
35	struct ev_loop *loop = ev_default_loop (0);	46	struct ev_loop *loop = ev_default_loop (0);
36		47
37	/* initialise an io watcher, then start it */	48	// initialise an io watcher, then start it
		49	// this one will watch for stdin to become readable
38	ev_io_init (&stdin_watcher, stdin_cb, /*STDIN_FILENO*/ 0, EV_READ);	50	ev_io_init (&stdin_watcher, stdin_cb, /*STDIN_FILENO*/ 0, EV_READ);
39	ev_io_start (loop, &stdin_watcher);	51	ev_io_start (loop, &stdin_watcher);
40		52
		53	// initialise a timer watcher, then start it
41	/* simple non-repeating 5.5 second timeout */	54	// simple non-repeating 5.5 second timeout
42	ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);	55	ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);
43	ev_timer_start (loop, &timeout_watcher);	56	ev_timer_start (loop, &timeout_watcher);
44		57
45	/* loop till timeout or data ready */	58	// now wait for events to arrive
46	ev_loop (loop, 0);	59	ev_loop (loop, 0);
47		60
		61	// unloop was called, so exit
48	return 0;	62	return 0;
49	}	63	}
50		64
51	=head1 DESCRIPTION	65	=head1 DESCRIPTION
52		66
53	The newest version of this document is also available as a html-formatted	67	The newest version of this document is also available as an html-formatted
54	web page you might find easier to navigate when reading it for the first	68	web page you might find easier to navigate when reading it for the first
55	time: L<http://cvs.schmorp.de/libev/ev.html>.	69	time: L<http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod>.
56		70
57	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
58	file descriptor being readable or a timeout occurring), and it will manage	72	file descriptor being readable or a timeout occurring), and it will manage
59	these event sources and provide your program with events.	73	these event sources and provide your program with events.
60		74
…		…
65	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
66	watchers>, which are relatively small C structures you initialise with the	80	watchers>, which are relatively small C structures you initialise with the
67	details of the event, and then hand it over to libev by I<starting> the	81	details of the event, and then hand it over to libev by I<starting> the
68	watcher.	82	watcher.
69		83
70	=head1 FEATURES	84	=head2 FEATURES
71		85
72	Libev supports C<select>, C<poll>, the Linux-specific C<epoll>, the	86	Libev supports C<select>, C<poll>, the Linux-specific C<epoll>, the
73	BSD-specific C<kqueue> and the Solaris-specific event port mechanisms	87	BSD-specific C<kqueue> and the Solaris-specific event port mechanisms
74	for file descriptor events (C<ev_io>), the Linux C<inotify> interface	88	for file descriptor events (C<ev_io>), the Linux C<inotify> interface
75	(for C<ev_stat>), relative timers (C<ev_timer>), absolute timers	89	(for C<ev_stat>), relative timers (C<ev_timer>), absolute timers
…		…
82		96
83	It also is quite fast (see this	97	It also is quite fast (see this
84	L<benchmark|http://libev.schmorp.de/bench.html> comparing it to libevent	98	L<benchmark|http://libev.schmorp.de/bench.html> comparing it to libevent
85	for example).	99	for example).
86		100
87	=head1 CONVENTIONS	101	=head2 CONVENTIONS
88		102
89	Libev is very configurable. In this manual the default configuration will	103	Libev is very configurable. In this manual the default (and most common)
90	be described, which supports multiple event loops. For more info about	104	configuration will be described, which supports multiple event loops. For
91	various configuration options please have a look at B<EMBED> section in	105	more info about various configuration options please have a look at
92	this manual. If libev was configured without support for multiple event	106	B<EMBED> section in this manual. If libev was configured without support
93	loops, then all functions taking an initial argument of name C<loop>	107	for multiple event loops, then all functions taking an initial argument of
94	(which is always of type C<struct ev_loop *>) will not have this argument.	108	name C<loop> (which is always of type C<struct ev_loop *>) will not have
		109	this argument.
95		110
96	=head1 TIME REPRESENTATION	111	=head2 TIME REPRESENTATION
97		112
98	Libev represents time as a single floating point number, representing the	113	Libev represents time as a single floating point number, representing the
99	(fractional) number of seconds since the (POSIX) epoch (somewhere near	114	(fractional) number of seconds since the (POSIX) epoch (somewhere near
100	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
101	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
102	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
103	it, you should treat it as some floatingpoint value. Unlike the name	118	it, you should treat it as some floating point value. Unlike the name
104	component C<stamp> might indicate, it is also used for time differences	119	component C<stamp> might indicate, it is also used for time differences
105	throughout libev.	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.
		141
106		142
107	=head1 GLOBAL FUNCTIONS	143	=head1 GLOBAL FUNCTIONS
108		144
109	These functions can be called anytime, even before initialising the	145	These functions can be called anytime, even before initialising the
110	library in any way.	146	library in any way.
…		…
119		155
120	=item ev_sleep (ev_tstamp interval)	156	=item ev_sleep (ev_tstamp interval)
121		157
122	Sleep for the given interval: The current thread will be blocked until	158	Sleep for the given interval: The current thread will be blocked until
123	either it is interrupted or the given time interval has passed. Basically	159	either it is interrupted or the given time interval has passed. Basically
124	this is a subsecond-resolution C<sleep ()>.	160	this is a sub-second-resolution C<sleep ()>.
125		161
126	=item int ev_version_major ()	162	=item int ev_version_major ()
127		163
128	=item int ev_version_minor ()	164	=item int ev_version_minor ()
129		165
…		…
142	not a problem.	178	not a problem.
143		179
144	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
145	version.	181	version.
146		182
147	assert (("libev version mismatch",	183	assert (("libev version mismatch",
148	ev_version_major () == EV_VERSION_MAJOR	184	ev_version_major () == EV_VERSION_MAJOR
149	&& ev_version_minor () >= EV_VERSION_MINOR));	185	&& ev_version_minor () >= EV_VERSION_MINOR));
150		186
151	=item unsigned int ev_supported_backends ()	187	=item unsigned int ev_supported_backends ()
152		188
153	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_*>
154	value) compiled into this binary of libev (independent of their	190	value) compiled into this binary of libev (independent of their
…		…
156	a description of the set values.	192	a description of the set values.
157		193
158	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
159	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
160		196
161	assert (("sorry, no epoll, no sex",	197	assert (("sorry, no epoll, no sex",
162	ev_supported_backends () & EVBACKEND_EPOLL));	198	ev_supported_backends () & EVBACKEND_EPOLL));
163		199
164	=item unsigned int ev_recommended_backends ()	200	=item unsigned int ev_recommended_backends ()
165		201
166	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
167	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
168	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
169	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
170	(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
171	libev will probe for if you specify no backends explicitly.	207	libev will probe for if you specify no backends explicitly.
172		208
173	=item unsigned int ev_embeddable_backends ()	209	=item unsigned int ev_embeddable_backends ()
174		210
…		…
181	See the description of C<ev_embed> watchers for more info.	217	See the description of C<ev_embed> watchers for more info.
182		218
183	=item ev_set_allocator (void *(*cb)(void *ptr, long size))	219	=item ev_set_allocator (void *(*cb)(void *ptr, long size))
184		220
185	Sets the allocation function to use (the prototype is similar - the	221	Sets the allocation function to use (the prototype is similar - the
186	semantics is identical - to the realloc C function). It is used to	222	semantics are identical to the C<realloc> C89/SuS/POSIX function). It is
187	allocate and free memory (no surprises here). If it returns zero when	223	used to allocate and free memory (no surprises here). If it returns zero
188	memory needs to be allocated, the library might abort or take some	224	when memory needs to be allocated (C<size != 0>), the library might abort
189	potentially destructive action. The default is your system realloc	225	or take some potentially destructive action.
190	function.	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.
191		230
192	You could override this function in high-availability programs to, say,	231	You could override this function in high-availability programs to, say,
193	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,
194	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.
195		234
196	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
197	retries).	236	retries (example requires a standards-compliant C<realloc>).
198		237
199	static void *	238	static void *
200	persistent_realloc (void *ptr, size_t size)	239	persistent_realloc (void *ptr, size_t size)
201	{	240	{
202	for (;;)	241	for (;;)
…		…
213	...	252	...
214	ev_set_allocator (persistent_realloc);	253	ev_set_allocator (persistent_realloc);
215		254
216	=item ev_set_syserr_cb (void (*cb)(const char *msg));	255	=item ev_set_syserr_cb (void (*cb)(const char *msg));
217		256
218	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
219	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
220	indicating the system call or subsystem causing the problem. If this	259	indicating the system call or subsystem causing the problem. If this
221	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
222	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
223	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
224	(such as abort).	263	(such as abort).
225		264
226	Example: This is basically the same thing that libev does internally, too.	265	Example: This is basically the same thing that libev does internally, too.
…		…
240	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP	279	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP
241		280
242	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 library knows two
243	types of such loops, the I<default> loop, which supports signals and child	282	types of such loops, the I<default> loop, which supports signals and child
244	events, and dynamically created loops which do not.	283	events, and dynamically created loops which do not.
245
246	If you use threads, a common model is to run the default event loop
247	in your main thread (or in a separate thread) and for each thread you
248	create, you also create another event loop. Libev itself does no locking
249	whatsoever, so if you mix calls to the same event loop in different
250	threads, make sure you lock (this is usually a bad idea, though, even if
251	done correctly, because it's hideous and inefficient).
252		284
253	=over 4	285	=over 4
254		286
255	=item struct ev_loop *ev_default_loop (unsigned int flags)	287	=item struct ev_loop *ev_default_loop (unsigned int flags)
256		288
…		…
260	flags. If that is troubling you, check C<ev_backend ()> afterwards).	292	flags. If that is troubling you, check C<ev_backend ()> afterwards).
261		293
262	If you don't know what event loop to use, use the one returned from this	294	If you don't know what event loop to use, use the one returned from this
263	function.	295	function.
264		296
		297	Note that this function is I<not> thread-safe, so if you want to use it
		298	from multiple threads, you have to lock (note also that this is unlikely,
		299	as loops cannot bes hared easily between threads anyway).
		300
		301	The default loop is the only loop that can handle C<ev_signal> and
		302	C<ev_child> watchers, and to do this, it always registers a handler
		303	for C<SIGCHLD>. If this is a problem for your application you can either
		304	create a dynamic loop with C<ev_loop_new> that doesn't do that, or you
		305	can simply overwrite the C<SIGCHLD> signal handler I<after> calling
		306	C<ev_default_init>.
		307
265	The flags argument can be used to specify special behaviour or specific	308	The flags argument can be used to specify special behaviour or specific
266	backends to use, and is usually specified as C<0> (or C<EVFLAG_AUTO>).	309	backends to use, and is usually specified as C<0> (or C<EVFLAG_AUTO>).
267		310
268	The following flags are supported:	311	The following flags are supported:
269		312
…		…
274	The default flags value. Use this if you have no clue (it's the right	317	The default flags value. Use this if you have no clue (it's the right
275	thing, believe me).	318	thing, believe me).
276		319
277	=item C<EVFLAG_NOENV>	320	=item C<EVFLAG_NOENV>
278		321
279	If this flag bit is ored into the flag value (or the program runs setuid	322	If this flag bit is or'ed into the flag value (or the program runs setuid
280	or setgid) then libev will I<not> look at the environment variable	323	or setgid) then libev will I<not> look at the environment variable
281	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will	324	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will
282	override the flags completely if it is found in the environment. This is	325	override the flags completely if it is found in the environment. This is
283	useful to try out specific backends to test their performance, or to work	326	useful to try out specific backends to test their performance, or to work
284	around bugs.	327	around bugs.
…		…
290	enabling this flag.	333	enabling this flag.
291		334
292	This works by calling C<getpid ()> on every iteration of the loop,	335	This works by calling C<getpid ()> on every iteration of the loop,
293	and thus this might slow down your event loop if you do a lot of loop	336	and thus this might slow down your event loop if you do a lot of loop
294	iterations and little real work, but is usually not noticeable (on my	337	iterations and little real work, but is usually not noticeable (on my
295	Linux system for example, C<getpid> is actually a simple 5-insn sequence	338	GNU/Linux system for example, C<getpid> is actually a simple 5-insn sequence
296	without a syscall and thus I<very> fast, but my Linux system also has	339	without a system call and thus I<very> fast, but my GNU/Linux system also has
297	C<pthread_atfork> which is even faster).	340	C<pthread_atfork> which is even faster).
298		341
299	The big advantage of this flag is that you can forget about fork (and	342	The big advantage of this flag is that you can forget about fork (and
300	forget about forgetting to tell libev about forking) when you use this	343	forget about forgetting to tell libev about forking) when you use this
301	flag.	344	flag.
302		345
303	This flag setting cannot be overriden or specified in the C<LIBEV_FLAGS>	346	This flag setting cannot be overridden or specified in the C<LIBEV_FLAGS>
304	environment variable.	347	environment variable.
305		348
306	=item C<EVBACKEND_SELECT> (value 1, portable select backend)	349	=item C<EVBACKEND_SELECT> (value 1, portable select backend)
307		350
308	This is your standard select(2) backend. Not I<completely> standard, as	351	This is your standard select(2) backend. Not I<completely> standard, as
309	libev tries to roll its own fd_set with no limits on the number of fds,	352	libev tries to roll its own fd_set with no limits on the number of fds,
310	but if that fails, expect a fairly low limit on the number of fds when	353	but if that fails, expect a fairly low limit on the number of fds when
311	using this backend. It doesn't scale too well (O(highest_fd)), but its usually	354	using this backend. It doesn't scale too well (O(highest_fd)), but its
312	the fastest backend for a low number of fds.	355	usually the fastest backend for a low number of (low-numbered :) fds.
		356
		357	To get good performance out of this backend you need a high amount of
		358	parallelism (most of the file descriptors should be busy). If you are
		359	writing a server, you should C<accept ()> in a loop to accept as many
		360	connections as possible during one iteration. You might also want to have
		361	a look at C<ev_set_io_collect_interval ()> to increase the amount of
		362	readiness notifications you get per iteration.
313		363
314	=item C<EVBACKEND_POLL> (value 2, poll backend, available everywhere except on windows)	364	=item C<EVBACKEND_POLL> (value 2, poll backend, available everywhere except on windows)
315		365
316	And this is your standard poll(2) backend. It's more complicated than	366	And this is your standard poll(2) backend. It's more complicated
317	select, but handles sparse fds better and has no artificial limit on the	367	than select, but handles sparse fds better and has no artificial
318	number of fds you can use (except it will slow down considerably with a	368	limit on the number of fds you can use (except it will slow down
319	lot of inactive fds). It scales similarly to select, i.e. O(total_fds).	369	considerably with a lot of inactive fds). It scales similarly to select,
		370	i.e. O(total_fds). See the entry for C<EVBACKEND_SELECT>, above, for
		371	performance tips.
320		372
321	=item C<EVBACKEND_EPOLL> (value 4, Linux)	373	=item C<EVBACKEND_EPOLL> (value 4, Linux)
322		374
323	For few fds, this backend is a bit little slower than poll and select,	375	For few fds, this backend is a bit little slower than poll and select,
324	but it scales phenomenally better. While poll and select usually scale	376	but it scales phenomenally better. While poll and select usually scale
325	like O(total_fds) where n is the total number of fds (or the highest fd),	377	like O(total_fds) where n is the total number of fds (or the highest fd),
326	epoll scales either O(1) or O(active_fds). The epoll design has a number	378	epoll scales either O(1) or O(active_fds). The epoll design has a number
327	of shortcomings, such as silently dropping events in some hard-to-detect	379	of shortcomings, such as silently dropping events in some hard-to-detect
328	cases and rewiring a syscall per fd change, no fork support and bad	380	cases and requiring a system call per fd change, no fork support and bad
329	support for dup:	381	support for dup.
330		382
331	While stopping, setting and starting an I/O watcher in the same iteration	383	While stopping, setting and starting an I/O watcher in the same iteration
332	will result in some caching, there is still a syscall per such incident	384	will result in some caching, there is still a system call per such incident
333	(because the fd could point to a different file description now), so its	385	(because the fd could point to a different file description now), so its
334	best to avoid that. Also, C<dup ()>'ed file descriptors might not work	386	best to avoid that. Also, C<dup ()>'ed file descriptors might not work
335	very well if you register events for both fds.	387	very well if you register events for both fds.
336		388
337	Please note that epoll sometimes generates spurious notifications, so you	389	Please note that epoll sometimes generates spurious notifications, so you
338	need to use non-blocking I/O or other means to avoid blocking when no data	390	need to use non-blocking I/O or other means to avoid blocking when no data
339	(or space) is available.	391	(or space) is available.
340		392
		393	Best performance from this backend is achieved by not unregistering all
		394	watchers for a file descriptor until it has been closed, if possible, i.e.
		395	keep at least one watcher active per fd at all times.
		396
		397	While nominally embeddable in other event loops, this feature is broken in
		398	all kernel versions tested so far.
		399
341	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)	400	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)
342		401
343	Kqueue deserves special mention, as at the time of this writing, it	402	Kqueue deserves special mention, as at the time of this writing, it
344	was broken on I<all> BSDs (usually it doesn't work with anything but	403	was broken on all BSDs except NetBSD (usually it doesn't work reliably
345	sockets and pipes, except on Darwin, where of course it's completely	404	with anything but sockets and pipes, except on Darwin, where of course
346	useless. On NetBSD, it seems to work for all the FD types I tested, so it
347	is used by default there). For this reason it's not being "autodetected"	405	it's completely useless). For this reason it's not being "auto-detected"
348	unless you explicitly specify it explicitly in the flags (i.e. using	406	unless you explicitly specify it explicitly in the flags (i.e. using
349	C<EVBACKEND_KQUEUE>) or libev was compiled on a known-to-be-good (-enough)	407	C<EVBACKEND_KQUEUE>) or libev was compiled on a known-to-be-good (-enough)
350	system like NetBSD.	408	system like NetBSD.
351		409
		410	You still can embed kqueue into a normal poll or select backend and use it
		411	only for sockets (after having made sure that sockets work with kqueue on
		412	the target platform). See C<ev_embed> watchers for more info.
		413
352	It scales in the same way as the epoll backend, but the interface to the	414	It scales in the same way as the epoll backend, but the interface to the
353	kernel is more efficient (which says nothing about its actual speed,	415	kernel is more efficient (which says nothing about its actual speed, of
354	of course). While stopping, setting and starting an I/O watcher does	416	course). While stopping, setting and starting an I/O watcher does never
355	never cause an extra syscall as with epoll, it still adds up to two event	417	cause an extra system call as with C<EVBACKEND_EPOLL>, it still adds up to
356	changes per incident, support for C<fork ()> is very bad and it drops fds	418	two event changes per incident, support for C<fork ()> is very bad and it
357	silently in similarly hard-to-detetc cases.	419	drops fds silently in similarly hard-to-detect cases.
		420
		421	This backend usually performs well under most conditions.
		422
		423	While nominally embeddable in other event loops, this doesn't work
		424	everywhere, so you might need to test for this. And since it is broken
		425	almost everywhere, you should only use it when you have a lot of sockets
		426	(for which it usually works), by embedding it into another event loop
		427	(e.g. C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>) and using it only for
		428	sockets.
358		429
359	=item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8)	430	=item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8)
360		431
361	This is not implemented yet (and might never be).	432	This is not implemented yet (and might never be, unless you send me an
		433	implementation). According to reports, C</dev/poll> only supports sockets
		434	and is not embeddable, which would limit the usefulness of this backend
		435	immensely.
362		436
363	=item C<EVBACKEND_PORT> (value 32, Solaris 10)	437	=item C<EVBACKEND_PORT> (value 32, Solaris 10)
364		438
365	This uses the Solaris 10 event port mechanism. As with everything on Solaris,	439	This uses the Solaris 10 event port mechanism. As with everything on Solaris,
366	it's really slow, but it still scales very well (O(active_fds)).	440	it's really slow, but it still scales very well (O(active_fds)).
367		441
368	Please note that solaris event ports can deliver a lot of spurious	442	Please note that Solaris event ports can deliver a lot of spurious
369	notifications, so you need to use non-blocking I/O or other means to avoid	443	notifications, so you need to use non-blocking I/O or other means to avoid
370	blocking when no data (or space) is available.	444	blocking when no data (or space) is available.
		445
		446	While this backend scales well, it requires one system call per active
		447	file descriptor per loop iteration. For small and medium numbers of file
		448	descriptors a "slow" C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> backend
		449	might perform better.
		450
		451	On the positive side, ignoring the spurious readiness notifications, this
		452	backend actually performed to specification in all tests and is fully
		453	embeddable, which is a rare feat among the OS-specific backends.
371		454
372	=item C<EVBACKEND_ALL>	455	=item C<EVBACKEND_ALL>
373		456
374	Try all backends (even potentially broken ones that wouldn't be tried	457	Try all backends (even potentially broken ones that wouldn't be tried
375	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as	458	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as
376	C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>.	459	C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>.
377		460
		461	It is definitely not recommended to use this flag.
		462
378	=back	463	=back
379		464
380	If one or more of these are ored into the flags value, then only these	465	If one or more of these are or'ed into the flags value, then only these
381	backends will be tried (in the reverse order as given here). If none are	466	backends will be tried (in the reverse order as listed here). If none are
382	specified, most compiled-in backend will be tried, usually in reverse	467	specified, all backends in C<ev_recommended_backends ()> will be tried.
383	order of their flag values :)
384		468
385	The most typical usage is like this:	469	The most typical usage is like this:
386		470
387	if (!ev_default_loop (0))	471	if (!ev_default_loop (0))
388	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");	472	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
389		473
390	Restrict libev to the select and poll backends, and do not allow	474	Restrict libev to the select and poll backends, and do not allow
391	environment settings to be taken into account:	475	environment settings to be taken into account:
392		476
393	ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);	477	ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);
394		478
395	Use whatever libev has to offer, but make sure that kqueue is used if	479	Use whatever libev has to offer, but make sure that kqueue is used if
396	available (warning, breaks stuff, best use only with your own private	480	available (warning, breaks stuff, best use only with your own private
397	event loop and only if you know the OS supports your types of fds):	481	event loop and only if you know the OS supports your types of fds):
398		482
399	ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);	483	ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);
400		484
401	=item struct ev_loop *ev_loop_new (unsigned int flags)	485	=item struct ev_loop *ev_loop_new (unsigned int flags)
402		486
403	Similar to C<ev_default_loop>, but always creates a new event loop that is	487	Similar to C<ev_default_loop>, but always creates a new event loop that is
404	always distinct from the default loop. Unlike the default loop, it cannot	488	always distinct from the default loop. Unlike the default loop, it cannot
405	handle signal and child watchers, and attempts to do so will be greeted by	489	handle signal and child watchers, and attempts to do so will be greeted by
406	undefined behaviour (or a failed assertion if assertions are enabled).	490	undefined behaviour (or a failed assertion if assertions are enabled).
407		491
		492	Note that this function I<is> thread-safe, and the recommended way to use
		493	libev with threads is indeed to create one loop per thread, and using the
		494	default loop in the "main" or "initial" thread.
		495
408	Example: Try to create a event loop that uses epoll and nothing else.	496	Example: Try to create a event loop that uses epoll and nothing else.
409		497
410	struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);	498	struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);
411	if (!epoller)	499	if (!epoller)
412	fatal ("no epoll found here, maybe it hides under your chair");	500	fatal ("no epoll found here, maybe it hides under your chair");
413		501
414	=item ev_default_destroy ()	502	=item ev_default_destroy ()
415		503
416	Destroys the default loop again (frees all memory and kernel state	504	Destroys the default loop again (frees all memory and kernel state
417	etc.). None of the active event watchers will be stopped in the normal	505	etc.). None of the active event watchers will be stopped in the normal
418	sense, so e.g. C<ev_is_active> might still return true. It is your	506	sense, so e.g. C<ev_is_active> might still return true. It is your
419	responsibility to either stop all watchers cleanly yoursef I<before>	507	responsibility to either stop all watchers cleanly yourself I<before>
420	calling this function, or cope with the fact afterwards (which is usually	508	calling this function, or cope with the fact afterwards (which is usually
421	the easiest thing, you can just ignore the watchers and/or C<free ()> them	509	the easiest thing, you can just ignore the watchers and/or C<free ()> them
422	for example).	510	for example).
423		511
424	Note that certain global state, such as signal state, will not be freed by	512	Note that certain global state, such as signal state, will not be freed by
…		…
435	Like C<ev_default_destroy>, but destroys an event loop created by an	523	Like C<ev_default_destroy>, but destroys an event loop created by an
436	earlier call to C<ev_loop_new>.	524	earlier call to C<ev_loop_new>.
437		525
438	=item ev_default_fork ()	526	=item ev_default_fork ()
439		527
		528	This function sets a flag that causes subsequent C<ev_loop> iterations
440	This function reinitialises the kernel state for backends that have	529	to reinitialise the kernel state for backends that have one. Despite the
441	one. Despite the name, you can call it anytime, but it makes most sense	530	name, you can call it anytime, but it makes most sense after forking, in
442	after forking, in either the parent or child process (or both, but that	531	the child process (or both child and parent, but that again makes little
443	again makes little sense).	532	sense). You I<must> call it in the child before using any of the libev
		533	functions, and it will only take effect at the next C<ev_loop> iteration.
444		534
445	You I<must> call this function in the child process after forking if and	535	On the other hand, you only need to call this function in the child
446	only if you want to use the event library in both processes. If you just	536	process if and only if you want to use the event library in the child. If
447	fork+exec, you don't have to call it.	537	you just fork+exec, you don't have to call it at all.
448		538
449	The function itself is quite fast and it's usually not a problem to call	539	The function itself is quite fast and it's usually not a problem to call
450	it just in case after a fork. To make this easy, the function will fit in	540	it just in case after a fork. To make this easy, the function will fit in
451	quite nicely into a call to C<pthread_atfork>:	541	quite nicely into a call to C<pthread_atfork>:
452		542
453	pthread_atfork (0, 0, ev_default_fork);	543	pthread_atfork (0, 0, ev_default_fork);
454		544
455	At the moment, C<EVBACKEND_SELECT> and C<EVBACKEND_POLL> are safe to use
456	without calling this function, so if you force one of those backends you
457	do not need to care.
458
459	=item ev_loop_fork (loop)	545	=item ev_loop_fork (loop)
460		546
461	Like C<ev_default_fork>, but acts on an event loop created by	547	Like C<ev_default_fork>, but acts on an event loop created by
462	C<ev_loop_new>. Yes, you have to call this on every allocated event loop	548	C<ev_loop_new>. Yes, you have to call this on every allocated event loop
463	after fork, and how you do this is entirely your own problem.	549	after fork, and how you do this is entirely your own problem.
		550
		551	=item int ev_is_default_loop (loop)
		552
		553	Returns true when the given loop actually is the default loop, false otherwise.
464		554
465	=item unsigned int ev_loop_count (loop)	555	=item unsigned int ev_loop_count (loop)
466		556
467	Returns the count of loop iterations for the loop, which is identical to	557	Returns the count of loop iterations for the loop, which is identical to
468	the number of times libev did poll for new events. It starts at C<0> and	558	the number of times libev did poll for new events. It starts at C<0> and
…		…
503	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle	593	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle
504	those events and any outstanding ones, but will not block your process in	594	those events and any outstanding ones, but will not block your process in
505	case there are no events and will return after one iteration of the loop.	595	case there are no events and will return after one iteration of the loop.
506		596
507	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if	597	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if
508	neccessary) and will handle those and any outstanding ones. It will block	598	necessary) and will handle those and any outstanding ones. It will block
509	your process until at least one new event arrives, and will return after	599	your process until at least one new event arrives, and will return after
510	one iteration of the loop. This is useful if you are waiting for some	600	one iteration of the loop. This is useful if you are waiting for some
511	external event in conjunction with something not expressible using other	601	external event in conjunction with something not expressible using other
512	libev watchers. However, a pair of C<ev_prepare>/C<ev_check> watchers is	602	libev watchers. However, a pair of C<ev_prepare>/C<ev_check> watchers is
513	usually a better approach for this kind of thing.	603	usually a better approach for this kind of thing.
514		604
515	Here are the gory details of what C<ev_loop> does:	605	Here are the gory details of what C<ev_loop> does:
516		606
517	- Before the first iteration, call any pending watchers.	607	- Before the first iteration, call any pending watchers.
518	* If there are no active watchers (reference count is zero), return.	608	* If EVFLAG_FORKCHECK was used, check for a fork.
519	- Queue all prepare watchers and then call all outstanding watchers.	609	- If a fork was detected (by any means), queue and call all fork watchers.
		610	- Queue and call all prepare watchers.
520	- If we have been forked, recreate the kernel state.	611	- If we have been forked, detach and recreate the kernel state
		612	as to not disturb the other process.
521	- Update the kernel state with all outstanding changes.	613	- Update the kernel state with all outstanding changes.
522	- Update the "event loop time".	614	- Update the "event loop time" (ev_now ()).
523	- Calculate for how long to block.	615	- Calculate for how long to sleep or block, if at all
		616	(active idle watchers, EVLOOP_NONBLOCK or not having
		617	any active watchers at all will result in not sleeping).
		618	- Sleep if the I/O and timer collect interval say so.
524	- Block the process, waiting for any events.	619	- Block the process, waiting for any events.
525	- Queue all outstanding I/O (fd) events.	620	- Queue all outstanding I/O (fd) events.
526	- Update the "event loop time" and do time jump handling.	621	- Update the "event loop time" (ev_now ()), and do time jump adjustments.
527	- Queue all outstanding timers.	622	- Queue all outstanding timers.
528	- Queue all outstanding periodics.	623	- Queue all outstanding periodics.
529	- If no events are pending now, queue all idle watchers.	624	- Unless any events are pending now, queue all idle watchers.
530	- Queue all check watchers.	625	- Queue all check watchers.
531	- Call all queued watchers in reverse order (i.e. check watchers first).	626	- Call all queued watchers in reverse order (i.e. check watchers first).
532	Signals and child watchers are implemented as I/O watchers, and will	627	Signals and child watchers are implemented as I/O watchers, and will
533	be handled here by queueing them when their watcher gets executed.	628	be handled here by queueing them when their watcher gets executed.
534	- If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK	629	- If ev_unloop has been called, or EVLOOP_ONESHOT or EVLOOP_NONBLOCK
535	were used, return, otherwise continue with step *.	630	were used, or there are no active watchers, return, otherwise
		631	continue with step *.
536		632
537	Example: Queue some jobs and then loop until no events are outsanding	633	Example: Queue some jobs and then loop until no events are outstanding
538	anymore.	634	anymore.
539		635
540	... queue jobs here, make sure they register event watchers as long	636	... queue jobs here, make sure they register event watchers as long
541	... as they still have work to do (even an idle watcher will do..)	637	... as they still have work to do (even an idle watcher will do..)
542	ev_loop (my_loop, 0);	638	ev_loop (my_loop, 0);
543	... jobs done. yeah!	639	... jobs done or somebody called unloop. yeah!
544		640
545	=item ev_unloop (loop, how)	641	=item ev_unloop (loop, how)
546		642
547	Can be used to make a call to C<ev_loop> return early (but only after it	643	Can be used to make a call to C<ev_loop> return early (but only after it
548	has processed all outstanding events). The C<how> argument must be either	644	has processed all outstanding events). The C<how> argument must be either
549	C<EVUNLOOP_ONE>, which will make the innermost C<ev_loop> call return, or	645	C<EVUNLOOP_ONE>, which will make the innermost C<ev_loop> call return, or
550	C<EVUNLOOP_ALL>, which will make all nested C<ev_loop> calls return.	646	C<EVUNLOOP_ALL>, which will make all nested C<ev_loop> calls return.
		647
		648	This "unloop state" will be cleared when entering C<ev_loop> again.
551		649
552	=item ev_ref (loop)	650	=item ev_ref (loop)
553		651
554	=item ev_unref (loop)	652	=item ev_unref (loop)
555		653
…		…
560	returning, ev_unref() after starting, and ev_ref() before stopping it. For	658	returning, ev_unref() after starting, and ev_ref() before stopping it. For
561	example, libev itself uses this for its internal signal pipe: It is not	659	example, libev itself uses this for its internal signal pipe: It is not
562	visible to the libev user and should not keep C<ev_loop> from exiting if	660	visible to the libev user and should not keep C<ev_loop> from exiting if
563	no event watchers registered by it are active. It is also an excellent	661	no event watchers registered by it are active. It is also an excellent
564	way to do this for generic recurring timers or from within third-party	662	way to do this for generic recurring timers or from within third-party
565	libraries. Just remember to I<unref after start> and I<ref before stop>.	663	libraries. Just remember to I<unref after start> and I<ref before stop>
		664	(but only if the watcher wasn't active before, or was active before,
		665	respectively).
566		666
567	Example: Create a signal watcher, but keep it from keeping C<ev_loop>	667	Example: Create a signal watcher, but keep it from keeping C<ev_loop>
568	running when nothing else is active.	668	running when nothing else is active.
569		669
570	struct ev_signal exitsig;	670	struct ev_signal exitsig;
571	ev_signal_init (&exitsig, sig_cb, SIGINT);	671	ev_signal_init (&exitsig, sig_cb, SIGINT);
572	ev_signal_start (loop, &exitsig);	672	ev_signal_start (loop, &exitsig);
573	evf_unref (loop);	673	evf_unref (loop);
574		674
575	Example: For some weird reason, unregister the above signal handler again.	675	Example: For some weird reason, unregister the above signal handler again.
576		676
577	ev_ref (loop);	677	ev_ref (loop);
578	ev_signal_stop (loop, &exitsig);	678	ev_signal_stop (loop, &exitsig);
579		679
580	=item ev_set_io_collect_interval (loop, ev_tstamp interval)	680	=item ev_set_io_collect_interval (loop, ev_tstamp interval)
581		681
582	=item ev_set_timeout_collect_interval (loop, ev_tstamp interval)	682	=item ev_set_timeout_collect_interval (loop, ev_tstamp interval)
583		683
584	These advanced functions influence the time that libev will spend waiting	684	These advanced functions influence the time that libev will spend waiting
585	for events. Both are by default C<0>, meaning that libev will try to	685	for events. Both time intervals are by default C<0>, meaning that libev
586	invoke timer/periodic callbacks and I/O callbacks with minimum latency.	686	will try to invoke timer/periodic callbacks and I/O callbacks with minimum
		687	latency.
587		688
588	Setting these to a higher value (the C<interval> I<must> be >= C<0>)	689	Setting these to a higher value (the C<interval> I<must> be >= C<0>)
589	allows libev to delay invocation of I/O and timer/periodic callbacks to	690	allows libev to delay invocation of I/O and timer/periodic callbacks
590	increase efficiency of loop iterations.	691	to increase efficiency of loop iterations (or to increase power-saving
		692	opportunities).
591		693
592	The background is that sometimes your program runs just fast enough to	694	The background is that sometimes your program runs just fast enough to
593	handle one (or very few) event(s) per loop iteration. While this makes	695	handle one (or very few) event(s) per loop iteration. While this makes
594	the program responsive, it also wastes a lot of CPU time to poll for new	696	the program responsive, it also wastes a lot of CPU time to poll for new
595	events, especially with backends like C<select ()> which have a high	697	events, especially with backends like C<select ()> which have a high
596	overhead for the actual polling but can deliver many events at once.	698	overhead for the actual polling but can deliver many events at once.
597		699
598	By setting a higher I<io collect interval> you allow libev to spend more	700	By setting a higher I<io collect interval> you allow libev to spend more
599	time collecting I/O events, so you can handle more events per iteration,	701	time collecting I/O events, so you can handle more events per iteration,
600	at the cost of increasing latency. Timeouts (both C<ev_periodic> and	702	at the cost of increasing latency. Timeouts (both C<ev_periodic> and
601	C<ev_timer>) will be not affected.	703	C<ev_timer>) will be not affected. Setting this to a non-null value will
		704	introduce an additional C<ev_sleep ()> call into most loop iterations.
602		705
603	Likewise, by setting a higher I<timeout collect interval> you allow libev	706	Likewise, by setting a higher I<timeout collect interval> you allow libev
604	to spend more time collecting timeouts, at the expense of increased	707	to spend more time collecting timeouts, at the expense of increased
605	latency (the watcher callback will be called later). C<ev_io> watchers	708	latency (the watcher callback will be called later). C<ev_io> watchers
606	will not be affected.	709	will not be affected. Setting this to a non-null value will not introduce
		710	any overhead in libev.
607		711
608	Many programs can usually benefit by setting the io collect interval to	712	Many (busy) programs can usually benefit by setting the I/O collect
609	a value near C<0.1> or so, which is often enough for interactive servers	713	interval to a value near C<0.1> or so, which is often enough for
610	(of course not for games), likewise for timeouts. It usually doesn't make	714	interactive servers (of course not for games), likewise for timeouts. It
611	much sense to set it to a lower value than C<0.01>, as this approsaches	715	usually doesn't make much sense to set it to a lower value than C<0.01>,
612	the timing granularity of most systems.	716	as this approaches the timing granularity of most systems.
		717
		718	Setting the I<timeout collect interval> can improve the opportunity for
		719	saving power, as the program will "bundle" timer callback invocations that
		720	are "near" in time together, by delaying some, thus reducing the number of
		721	times the process sleeps and wakes up again. Another useful technique to
		722	reduce iterations/wake-ups is to use C<ev_periodic> watchers and make sure
		723	they fire on, say, one-second boundaries only.
		724
		725	=item ev_loop_verify (loop)
		726
		727	This function only does something when C<EV_VERIFY> support has been
		728	compiled in. It tries to go through all internal structures and checks
		729	them for validity. If anything is found to be inconsistent, it will print
		730	an error message to standard error and call C<abort ()>.
		731
		732	This can be used to catch bugs inside libev itself: under normal
		733	circumstances, this function will never abort as of course libev keeps its
		734	data structures consistent.
613		735
614	=back	736	=back
615		737
616		738
617	=head1 ANATOMY OF A WATCHER	739	=head1 ANATOMY OF A WATCHER
618		740
619	A watcher is a structure that you create and register to record your	741	A watcher is a structure that you create and register to record your
620	interest in some event. For instance, if you want to wait for STDIN to	742	interest in some event. For instance, if you want to wait for STDIN to
621	become readable, you would create an C<ev_io> watcher for that:	743	become readable, you would create an C<ev_io> watcher for that:
622		744
623	static void my_cb (struct ev_loop *loop, struct ev_io *w, int revents)	745	static void my_cb (struct ev_loop *loop, struct ev_io *w, int revents)
624	{	746	{
625	ev_io_stop (w);	747	ev_io_stop (w);
626	ev_unloop (loop, EVUNLOOP_ALL);	748	ev_unloop (loop, EVUNLOOP_ALL);
627	}	749	}
628		750
629	struct ev_loop *loop = ev_default_loop (0);	751	struct ev_loop *loop = ev_default_loop (0);
630	struct ev_io stdin_watcher;	752	struct ev_io stdin_watcher;
631	ev_init (&stdin_watcher, my_cb);	753	ev_init (&stdin_watcher, my_cb);
632	ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);	754	ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);
633	ev_io_start (loop, &stdin_watcher);	755	ev_io_start (loop, &stdin_watcher);
634	ev_loop (loop, 0);	756	ev_loop (loop, 0);
635		757
636	As you can see, you are responsible for allocating the memory for your	758	As you can see, you are responsible for allocating the memory for your
637	watcher structures (and it is usually a bad idea to do this on the stack,	759	watcher structures (and it is usually a bad idea to do this on the stack,
638	although this can sometimes be quite valid).	760	although this can sometimes be quite valid).
639		761
640	Each watcher structure must be initialised by a call to C<ev_init	762	Each watcher structure must be initialised by a call to C<ev_init
641	(watcher *, callback)>, which expects a callback to be provided. This	763	(watcher *, callback)>, which expects a callback to be provided. This
642	callback gets invoked each time the event occurs (or, in the case of io	764	callback gets invoked each time the event occurs (or, in the case of I/O
643	watchers, each time the event loop detects that the file descriptor given	765	watchers, each time the event loop detects that the file descriptor given
644	is readable and/or writable).	766	is readable and/or writable).
645		767
646	Each watcher type has its own C<< ev_<type>_set (watcher *, ...) >> macro	768	Each watcher type has its own C<< ev_<type>_set (watcher *, ...) >> macro
647	with arguments specific to this watcher type. There is also a macro	769	with arguments specific to this watcher type. There is also a macro
…		…
717	=item C<EV_FORK>	839	=item C<EV_FORK>
718		840
719	The event loop has been resumed in the child process after fork (see	841	The event loop has been resumed in the child process after fork (see
720	C<ev_fork>).	842	C<ev_fork>).
721		843
		844	=item C<EV_ASYNC>
		845
		846	The given async watcher has been asynchronously notified (see C<ev_async>).
		847
722	=item C<EV_ERROR>	848	=item C<EV_ERROR>
723		849
724	An unspecified error has occured, the watcher has been stopped. This might	850	An unspecified error has occurred, the watcher has been stopped. This might
725	happen because the watcher could not be properly started because libev	851	happen because the watcher could not be properly started because libev
726	ran out of memory, a file descriptor was found to be closed or any other	852	ran out of memory, a file descriptor was found to be closed or any other
727	problem. You best act on it by reporting the problem and somehow coping	853	problem. You best act on it by reporting the problem and somehow coping
728	with the watcher being stopped.	854	with the watcher being stopped.
729		855
730	Libev will usually signal a few "dummy" events together with an error,	856	Libev will usually signal a few "dummy" events together with an error,
731	for example it might indicate that a fd is readable or writable, and if	857	for example it might indicate that a fd is readable or writable, and if
732	your callbacks is well-written it can just attempt the operation and cope	858	your callbacks is well-written it can just attempt the operation and cope
733	with the error from read() or write(). This will not work in multithreaded	859	with the error from read() or write(). This will not work in multi-threaded
734	programs, though, so beware.	860	programs, though, so beware.
735		861
736	=back	862	=back
737		863
738	=head2 GENERIC WATCHER FUNCTIONS	864	=head2 GENERIC WATCHER FUNCTIONS
…		…
768	Although some watcher types do not have type-specific arguments	894	Although some watcher types do not have type-specific arguments
769	(e.g. C<ev_prepare>) you still need to call its C<set> macro.	895	(e.g. C<ev_prepare>) you still need to call its C<set> macro.
770		896
771	=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])	897	=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])
772		898
773	This convinience macro rolls both C<ev_init> and C<ev_TYPE_set> macro	899	This convenience macro rolls both C<ev_init> and C<ev_TYPE_set> macro
774	calls into a single call. This is the most convinient method to initialise	900	calls into a single call. This is the most convenient method to initialise
775	a watcher. The same limitations apply, of course.	901	a watcher. The same limitations apply, of course.
776		902
777	=item C<ev_TYPE_start> (loop *, ev_TYPE *watcher)	903	=item C<ev_TYPE_start> (loop *, ev_TYPE *watcher)
778		904
779	Starts (activates) the given watcher. Only active watchers will receive	905	Starts (activates) the given watcher. Only active watchers will receive
…		…
862	to associate arbitrary data with your watcher. If you need more data and	988	to associate arbitrary data with your watcher. If you need more data and
863	don't want to allocate memory and store a pointer to it in that data	989	don't want to allocate memory and store a pointer to it in that data
864	member, you can also "subclass" the watcher type and provide your own	990	member, you can also "subclass" the watcher type and provide your own
865	data:	991	data:
866		992
867	struct my_io	993	struct my_io
868	{	994	{
869	struct ev_io io;	995	struct ev_io io;
870	int otherfd;	996	int otherfd;
871	void *somedata;	997	void *somedata;
872	struct whatever *mostinteresting;	998	struct whatever *mostinteresting;
873	}	999	}
874		1000
875	And since your callback will be called with a pointer to the watcher, you	1001	And since your callback will be called with a pointer to the watcher, you
876	can cast it back to your own type:	1002	can cast it back to your own type:
877		1003
878	static void my_cb (struct ev_loop *loop, struct ev_io *w_, int revents)	1004	static void my_cb (struct ev_loop *loop, struct ev_io *w_, int revents)
879	{	1005	{
880	struct my_io *w = (struct my_io *)w_;	1006	struct my_io *w = (struct my_io *)w_;
881	...	1007	...
882	}	1008	}
883		1009
884	More interesting and less C-conformant ways of casting your callback type	1010	More interesting and less C-conformant ways of casting your callback type
885	instead have been omitted.	1011	instead have been omitted.
886		1012
887	Another common scenario is having some data structure with multiple	1013	Another common scenario is having some data structure with multiple
888	watchers:	1014	watchers:
889		1015
890	struct my_biggy	1016	struct my_biggy
891	{	1017	{
892	int some_data;	1018	int some_data;
893	ev_timer t1;	1019	ev_timer t1;
894	ev_timer t2;	1020	ev_timer t2;
895	}	1021	}
896		1022
897	In this case getting the pointer to C<my_biggy> is a bit more complicated,	1023	In this case getting the pointer to C<my_biggy> is a bit more complicated,
898	you need to use C<offsetof>:	1024	you need to use C<offsetof>:
899		1025
900	#include <stddef.h>	1026	#include <stddef.h>
901		1027
902	static void	1028	static void
903	t1_cb (EV_P_ struct ev_timer *w, int revents)	1029	t1_cb (EV_P_ struct ev_timer *w, int revents)
904	{	1030	{
905	struct my_biggy big = (struct my_biggy *	1031	struct my_biggy big = (struct my_biggy *
906	(((char *)w) - offsetof (struct my_biggy, t1));	1032	(((char *)w) - offsetof (struct my_biggy, t1));
907	}	1033	}
908		1034
909	static void	1035	static void
910	t2_cb (EV_P_ struct ev_timer *w, int revents)	1036	t2_cb (EV_P_ struct ev_timer *w, int revents)
911	{	1037	{
912	struct my_biggy big = (struct my_biggy *	1038	struct my_biggy big = (struct my_biggy *
913	(((char *)w) - offsetof (struct my_biggy, t2));	1039	(((char *)w) - offsetof (struct my_biggy, t2));
914	}	1040	}
915		1041
916		1042
917	=head1 WATCHER TYPES	1043	=head1 WATCHER TYPES
918		1044
919	This section describes each watcher in detail, but will not repeat	1045	This section describes each watcher in detail, but will not repeat
…		…
943	In general you can register as many read and/or write event watchers per	1069	In general you can register as many read and/or write event watchers per
944	fd as you want (as long as you don't confuse yourself). Setting all file	1070	fd as you want (as long as you don't confuse yourself). Setting all file
945	descriptors to non-blocking mode is also usually a good idea (but not	1071	descriptors to non-blocking mode is also usually a good idea (but not
946	required if you know what you are doing).	1072	required if you know what you are doing).
947		1073
948	You have to be careful with dup'ed file descriptors, though. Some backends
949	(the linux epoll backend is a notable example) cannot handle dup'ed file
950	descriptors correctly if you register interest in two or more fds pointing
951	to the same underlying file/socket/etc. description (that is, they share
952	the same underlying "file open").
953
954	If you must do this, then force the use of a known-to-be-good backend	1074	If you must do this, then force the use of a known-to-be-good backend
955	(at the time of this writing, this includes only C<EVBACKEND_SELECT> and	1075	(at the time of this writing, this includes only C<EVBACKEND_SELECT> and
956	C<EVBACKEND_POLL>).	1076	C<EVBACKEND_POLL>).
957		1077
958	Another thing you have to watch out for is that it is quite easy to	1078	Another thing you have to watch out for is that it is quite easy to
959	receive "spurious" readyness notifications, that is your callback might	1079	receive "spurious" readiness notifications, that is your callback might
960	be called with C<EV_READ> but a subsequent C<read>(2) will actually block	1080	be called with C<EV_READ> but a subsequent C<read>(2) will actually block
961	because there is no data. Not only are some backends known to create a	1081	because there is no data. Not only are some backends known to create a
962	lot of those (for example solaris ports), it is very easy to get into	1082	lot of those (for example Solaris ports), it is very easy to get into
963	this situation even with a relatively standard program structure. Thus	1083	this situation even with a relatively standard program structure. Thus
964	it is best to always use non-blocking I/O: An extra C<read>(2) returning	1084	it is best to always use non-blocking I/O: An extra C<read>(2) returning
965	C<EAGAIN> is far preferable to a program hanging until some data arrives.	1085	C<EAGAIN> is far preferable to a program hanging until some data arrives.
966		1086
967	If you cannot run the fd in non-blocking mode (for example you should not	1087	If you cannot run the fd in non-blocking mode (for example you should not
968	play around with an Xlib connection), then you have to seperately re-test	1088	play around with an Xlib connection), then you have to separately re-test
969	whether a file descriptor is really ready with a known-to-be good interface	1089	whether a file descriptor is really ready with a known-to-be good interface
970	such as poll (fortunately in our Xlib example, Xlib already does this on	1090	such as poll (fortunately in our Xlib example, Xlib already does this on
971	its own, so its quite safe to use).	1091	its own, so its quite safe to use).
972		1092
973	=head3 The special problem of disappearing file descriptors	1093	=head3 The special problem of disappearing file descriptors
…		…
992	optimisations to libev.	1112	optimisations to libev.
993		1113
994	=head3 The special problem of dup'ed file descriptors	1114	=head3 The special problem of dup'ed file descriptors
995		1115
996	Some backends (e.g. epoll), cannot register events for file descriptors,	1116	Some backends (e.g. epoll), cannot register events for file descriptors,
997	but only events for the underlying file descriptions. That menas when you	1117	but only events for the underlying file descriptions. That means when you
998	have C<dup ()>'ed file descriptors and register events for them, only one	1118	have C<dup ()>'ed file descriptors or weirder constellations, and register
999	file descriptor might actually receive events.	1119	events for them, only one file descriptor might actually receive events.
1000		1120
1001	There is no workaorund possible except not registering events	1121	There is no workaround possible except not registering events
1002	for potentially C<dup ()>'ed file descriptors or to resort to	1122	for potentially C<dup ()>'ed file descriptors, or to resort to
1003	C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>.	1123	C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>.
1004		1124
1005	=head3 The special problem of fork	1125	=head3 The special problem of fork
1006		1126
1007	Some backends (epoll, kqueue) do not support C<fork ()> at all or exhibit	1127	Some backends (epoll, kqueue) do not support C<fork ()> at all or exhibit
…		…
1011	To support fork in your programs, you either have to call	1131	To support fork in your programs, you either have to call
1012	C<ev_default_fork ()> or C<ev_loop_fork ()> after a fork in the child,	1132	C<ev_default_fork ()> or C<ev_loop_fork ()> after a fork in the child,
1013	enable C<EVFLAG_FORKCHECK>, or resort to C<EVBACKEND_SELECT> or	1133	enable C<EVFLAG_FORKCHECK>, or resort to C<EVBACKEND_SELECT> or
1014	C<EVBACKEND_POLL>.	1134	C<EVBACKEND_POLL>.
1015		1135
		1136	=head3 The special problem of SIGPIPE
		1137
		1138	While not really specific to libev, it is easy to forget about SIGPIPE:
		1139	when reading from a pipe whose other end has been closed, your program
		1140	gets send a SIGPIPE, which, by default, aborts your program. For most
		1141	programs this is sensible behaviour, for daemons, this is usually
		1142	undesirable.
		1143
		1144	So when you encounter spurious, unexplained daemon exits, make sure you
		1145	ignore SIGPIPE (and maybe make sure you log the exit status of your daemon
		1146	somewhere, as that would have given you a big clue).
		1147
1016		1148
1017	=head3 Watcher-Specific Functions	1149	=head3 Watcher-Specific Functions
1018		1150
1019	=over 4	1151	=over 4
1020		1152
1021	=item ev_io_init (ev_io *, callback, int fd, int events)	1153	=item ev_io_init (ev_io *, callback, int fd, int events)
1022		1154
1023	=item ev_io_set (ev_io *, int fd, int events)	1155	=item ev_io_set (ev_io *, int fd, int events)
1024		1156
1025	Configures an C<ev_io> watcher. The C<fd> is the file descriptor to	1157	Configures an C<ev_io> watcher. The C<fd> is the file descriptor to
1026	rceeive events for and events is either C<EV_READ>, C<EV_WRITE> or	1158	receive events for and events is either C<EV_READ>, C<EV_WRITE> or
1027	C<EV_READ | EV_WRITE> to receive the given events.	1159	C<EV_READ | EV_WRITE> to receive the given events.
1028		1160
1029	=item int fd [read-only]	1161	=item int fd [read-only]
1030		1162
1031	The file descriptor being watched.	1163	The file descriptor being watched.
…		…
1033	=item int events [read-only]	1165	=item int events [read-only]
1034		1166
1035	The events being watched.	1167	The events being watched.
1036		1168
1037	=back	1169	=back
		1170
		1171	=head3 Examples
1038		1172
1039	Example: Call C<stdin_readable_cb> when STDIN_FILENO has become, well	1173	Example: Call C<stdin_readable_cb> when STDIN_FILENO has become, well
1040	readable, but only once. Since it is likely line-buffered, you could	1174	readable, but only once. Since it is likely line-buffered, you could
1041	attempt to read a whole line in the callback.	1175	attempt to read a whole line in the callback.
1042		1176
1043	static void	1177	static void
1044	stdin_readable_cb (struct ev_loop *loop, struct ev_io *w, int revents)	1178	stdin_readable_cb (struct ev_loop *loop, struct ev_io *w, int revents)
1045	{	1179	{
1046	ev_io_stop (loop, w);	1180	ev_io_stop (loop, w);
1047	.. read from stdin here (or from w->fd) and haqndle any I/O errors	1181	.. read from stdin here (or from w->fd) and haqndle any I/O errors
1048	}	1182	}
1049		1183
1050	...	1184	...
1051	struct ev_loop *loop = ev_default_init (0);	1185	struct ev_loop *loop = ev_default_init (0);
1052	struct ev_io stdin_readable;	1186	struct ev_io stdin_readable;
1053	ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);	1187	ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
1054	ev_io_start (loop, &stdin_readable);	1188	ev_io_start (loop, &stdin_readable);
1055	ev_loop (loop, 0);	1189	ev_loop (loop, 0);
1056		1190
1057		1191
1058	=head2 C<ev_timer> - relative and optionally repeating timeouts	1192	=head2 C<ev_timer> - relative and optionally repeating timeouts
1059		1193
1060	Timer watchers are simple relative timers that generate an event after a	1194	Timer watchers are simple relative timers that generate an event after a
1061	given time, and optionally repeating in regular intervals after that.	1195	given time, and optionally repeating in regular intervals after that.
1062		1196
1063	The timers are based on real time, that is, if you register an event that	1197	The timers are based on real time, that is, if you register an event that
1064	times out after an hour and you reset your system clock to last years	1198	times out after an hour and you reset your system clock to January last
1065	time, it will still time out after (roughly) and hour. "Roughly" because	1199	year, it will still time out after (roughly) and hour. "Roughly" because
1066	detecting time jumps is hard, and some inaccuracies are unavoidable (the	1200	detecting time jumps is hard, and some inaccuracies are unavoidable (the
1067	monotonic clock option helps a lot here).	1201	monotonic clock option helps a lot here).
1068		1202
1069	The relative timeouts are calculated relative to the C<ev_now ()>	1203	The relative timeouts are calculated relative to the C<ev_now ()>
1070	time. This is usually the right thing as this timestamp refers to the time	1204	time. This is usually the right thing as this timestamp refers to the time
…		…
1072	you suspect event processing to be delayed and you I<need> to base the timeout	1206	you suspect event processing to be delayed and you I<need> to base the timeout
1073	on the current time, use something like this to adjust for this:	1207	on the current time, use something like this to adjust for this:
1074		1208
1075	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);	1209	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);
1076		1210
1077	The callback is guarenteed to be invoked only when its timeout has passed,	1211	The callback is guaranteed to be invoked only after its timeout has passed,
1078	but if multiple timers become ready during the same loop iteration then	1212	but if multiple timers become ready during the same loop iteration then
1079	order of execution is undefined.	1213	order of execution is undefined.
1080		1214
1081	=head3 Watcher-Specific Functions and Data Members	1215	=head3 Watcher-Specific Functions and Data Members
1082		1216
…		…
1084		1218
1085	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)	1219	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)
1086		1220
1087	=item ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)	1221	=item ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)
1088		1222
1089	Configure the timer to trigger after C<after> seconds. If C<repeat> is	1223	Configure the timer to trigger after C<after> seconds. If C<repeat>
1090	C<0.>, then it will automatically be stopped. If it is positive, then the	1224	is C<0.>, then it will automatically be stopped once the timeout is
1091	timer will automatically be configured to trigger again C<repeat> seconds	1225	reached. If it is positive, then the timer will automatically be
1092	later, again, and again, until stopped manually.	1226	configured to trigger again C<repeat> seconds later, again, and again,
		1227	until stopped manually.
1093		1228
1094	The timer itself will do a best-effort at avoiding drift, that is, if you	1229	The timer itself will do a best-effort at avoiding drift, that is, if
1095	configure a timer to trigger every 10 seconds, then it will trigger at	1230	you configure a timer to trigger every 10 seconds, then it will normally
1096	exactly 10 second intervals. If, however, your program cannot keep up with	1231	trigger at exactly 10 second intervals. If, however, your program cannot
1097	the timer (because it takes longer than those 10 seconds to do stuff) the	1232	keep up with the timer (because it takes longer than those 10 seconds to
1098	timer will not fire more than once per event loop iteration.	1233	do stuff) the timer will not fire more than once per event loop iteration.
1099		1234
1100	=item ev_timer_again (loop)	1235	=item ev_timer_again (loop, ev_timer *)
1101		1236
1102	This will act as if the timer timed out and restart it again if it is	1237	This will act as if the timer timed out and restart it again if it is
1103	repeating. The exact semantics are:	1238	repeating. The exact semantics are:
1104		1239
1105	If the timer is pending, its pending status is cleared.	1240	If the timer is pending, its pending status is cleared.
1106		1241
1107	If the timer is started but nonrepeating, stop it (as if it timed out).	1242	If the timer is started but non-repeating, stop it (as if it timed out).
1108		1243
1109	If the timer is repeating, either start it if necessary (with the	1244	If the timer is repeating, either start it if necessary (with the
1110	C<repeat> value), or reset the running timer to the C<repeat> value.	1245	C<repeat> value), or reset the running timer to the C<repeat> value.
1111		1246
1112	This sounds a bit complicated, but here is a useful and typical	1247	This sounds a bit complicated, but here is a useful and typical
1113	example: Imagine you have a tcp connection and you want a so-called idle	1248	example: Imagine you have a TCP connection and you want a so-called idle
1114	timeout, that is, you want to be called when there have been, say, 60	1249	timeout, that is, you want to be called when there have been, say, 60
1115	seconds of inactivity on the socket. The easiest way to do this is to	1250	seconds of inactivity on the socket. The easiest way to do this is to
1116	configure an C<ev_timer> with a C<repeat> value of C<60> and then call	1251	configure an C<ev_timer> with a C<repeat> value of C<60> and then call
1117	C<ev_timer_again> each time you successfully read or write some data. If	1252	C<ev_timer_again> each time you successfully read or write some data. If
1118	you go into an idle state where you do not expect data to travel on the	1253	you go into an idle state where you do not expect data to travel on the
…		…
1140	or C<ev_timer_again> is called and determines the next timeout (if any),	1275	or C<ev_timer_again> is called and determines the next timeout (if any),
1141	which is also when any modifications are taken into account.	1276	which is also when any modifications are taken into account.
1142		1277
1143	=back	1278	=back
1144		1279
		1280	=head3 Examples
		1281
1145	Example: Create a timer that fires after 60 seconds.	1282	Example: Create a timer that fires after 60 seconds.
1146		1283
1147	static void	1284	static void
1148	one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)	1285	one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
1149	{	1286	{
1150	.. one minute over, w is actually stopped right here	1287	.. one minute over, w is actually stopped right here
1151	}	1288	}
1152		1289
1153	struct ev_timer mytimer;	1290	struct ev_timer mytimer;
1154	ev_timer_init (&mytimer, one_minute_cb, 60., 0.);	1291	ev_timer_init (&mytimer, one_minute_cb, 60., 0.);
1155	ev_timer_start (loop, &mytimer);	1292	ev_timer_start (loop, &mytimer);
1156		1293
1157	Example: Create a timeout timer that times out after 10 seconds of	1294	Example: Create a timeout timer that times out after 10 seconds of
1158	inactivity.	1295	inactivity.
1159		1296
1160	static void	1297	static void
1161	timeout_cb (struct ev_loop *loop, struct ev_timer *w, int revents)	1298	timeout_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
1162	{	1299	{
1163	.. ten seconds without any activity	1300	.. ten seconds without any activity
1164	}	1301	}
1165		1302
1166	struct ev_timer mytimer;	1303	struct ev_timer mytimer;
1167	ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */	1304	ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */
1168	ev_timer_again (&mytimer); /* start timer */	1305	ev_timer_again (&mytimer); /* start timer */
1169	ev_loop (loop, 0);	1306	ev_loop (loop, 0);
1170		1307
1171	// and in some piece of code that gets executed on any "activity":	1308	// and in some piece of code that gets executed on any "activity":
1172	// reset the timeout to start ticking again at 10 seconds	1309	// reset the timeout to start ticking again at 10 seconds
1173	ev_timer_again (&mytimer);	1310	ev_timer_again (&mytimer);
1174		1311
1175		1312
1176	=head2 C<ev_periodic> - to cron or not to cron?	1313	=head2 C<ev_periodic> - to cron or not to cron?
1177		1314
1178	Periodic watchers are also timers of a kind, but they are very versatile	1315	Periodic watchers are also timers of a kind, but they are very versatile
1179	(and unfortunately a bit complex).	1316	(and unfortunately a bit complex).
1180		1317
1181	Unlike C<ev_timer>'s, they are not based on real time (or relative time)	1318	Unlike C<ev_timer>'s, they are not based on real time (or relative time)
1182	but on wallclock time (absolute time). You can tell a periodic watcher	1319	but on wall clock time (absolute time). You can tell a periodic watcher
1183	to trigger "at" some specific point in time. For example, if you tell a	1320	to trigger after some specific point in time. For example, if you tell a
1184	periodic watcher to trigger in 10 seconds (by specifiying e.g. C<ev_now ()	1321	periodic watcher to trigger in 10 seconds (by specifying e.g. C<ev_now ()
1185	+ 10.>) and then reset your system clock to the last year, then it will	1322	+ 10.>, that is, an absolute time not a delay) and then reset your system
		1323	clock to January of the previous year, then it will take more than year
1186	take a year to trigger the event (unlike an C<ev_timer>, which would trigger	1324	to trigger the event (unlike an C<ev_timer>, which would still trigger
1187	roughly 10 seconds later).	1325	roughly 10 seconds later as it uses a relative timeout).
1188		1326
1189	They can also be used to implement vastly more complex timers, such as	1327	C<ev_periodic>s can also be used to implement vastly more complex timers,
1190	triggering an event on each midnight, local time or other, complicated,	1328	such as triggering an event on each "midnight, local time", or other
1191	rules.	1329	complicated, rules.
1192		1330
1193	As with timers, the callback is guarenteed to be invoked only when the	1331	As with timers, the callback is guaranteed to be invoked only when the
1194	time (C<at>) has been passed, but if multiple periodic timers become ready	1332	time (C<at>) has passed, but if multiple periodic timers become ready
1195	during the same loop iteration then order of execution is undefined.	1333	during the same loop iteration then order of execution is undefined.
1196		1334
1197	=head3 Watcher-Specific Functions and Data Members	1335	=head3 Watcher-Specific Functions and Data Members
1198		1336
1199	=over 4	1337	=over 4
…		…
1207		1345
1208	=over 4	1346	=over 4
1209		1347
1210	=item * absolute timer (at = time, interval = reschedule_cb = 0)	1348	=item * absolute timer (at = time, interval = reschedule_cb = 0)
1211		1349
1212	In this configuration the watcher triggers an event at the wallclock time	1350	In this configuration the watcher triggers an event after the wall clock
1213	C<at> and doesn't repeat. It will not adjust when a time jump occurs,	1351	time C<at> has passed and doesn't repeat. It will not adjust when a time
1214	that is, if it is to be run at January 1st 2011 then it will run when the	1352	jump occurs, that is, if it is to be run at January 1st 2011 then it will
1215	system time reaches or surpasses this time.	1353	run when the system time reaches or surpasses this time.
1216		1354
1217	=item * non-repeating interval timer (at = offset, interval > 0, reschedule_cb = 0)	1355	=item * repeating interval timer (at = offset, interval > 0, reschedule_cb = 0)
1218		1356
1219	In this mode the watcher will always be scheduled to time out at the next	1357	In this mode the watcher will always be scheduled to time out at the next
1220	C<at + N * interval> time (for some integer N, which can also be negative)	1358	C<at + N * interval> time (for some integer N, which can also be negative)
1221	and then repeat, regardless of any time jumps.	1359	and then repeat, regardless of any time jumps.
1222		1360
1223	This can be used to create timers that do not drift with respect to system	1361	This can be used to create timers that do not drift with respect to system
1224	time:	1362	time, for example, here is a C<ev_periodic> that triggers each hour, on
		1363	the hour:
1225		1364
1226	ev_periodic_set (&periodic, 0., 3600., 0);	1365	ev_periodic_set (&periodic, 0., 3600., 0);
1227		1366
1228	This doesn't mean there will always be 3600 seconds in between triggers,	1367	This doesn't mean there will always be 3600 seconds in between triggers,
1229	but only that the the callback will be called when the system time shows a	1368	but only that the callback will be called when the system time shows a
1230	full hour (UTC), or more correctly, when the system time is evenly divisible	1369	full hour (UTC), or more correctly, when the system time is evenly divisible
1231	by 3600.	1370	by 3600.
1232		1371
1233	Another way to think about it (for the mathematically inclined) is that	1372	Another way to think about it (for the mathematically inclined) is that
1234	C<ev_periodic> will try to run the callback in this mode at the next possible	1373	C<ev_periodic> will try to run the callback in this mode at the next possible
1235	time where C<time = at (mod interval)>, regardless of any time jumps.	1374	time where C<time = at (mod interval)>, regardless of any time jumps.
1236		1375
1237	For numerical stability it is preferable that the C<at> value is near	1376	For numerical stability it is preferable that the C<at> value is near
1238	C<ev_now ()> (the current time), but there is no range requirement for	1377	C<ev_now ()> (the current time), but there is no range requirement for
1239	this value.	1378	this value, and in fact is often specified as zero.
		1379
		1380	Note also that there is an upper limit to how often a timer can fire (CPU
		1381	speed for example), so if C<interval> is very small then timing stability
		1382	will of course deteriorate. Libev itself tries to be exact to be about one
		1383	millisecond (if the OS supports it and the machine is fast enough).
1240		1384
1241	=item * manual reschedule mode (at and interval ignored, reschedule_cb = callback)	1385	=item * manual reschedule mode (at and interval ignored, reschedule_cb = callback)
1242		1386
1243	In this mode the values for C<interval> and C<at> are both being	1387	In this mode the values for C<interval> and C<at> are both being
1244	ignored. Instead, each time the periodic watcher gets scheduled, the	1388	ignored. Instead, each time the periodic watcher gets scheduled, the
1245	reschedule callback will be called with the watcher as first, and the	1389	reschedule callback will be called with the watcher as first, and the
1246	current time as second argument.	1390	current time as second argument.
1247		1391
1248	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher,	1392	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher,
1249	ever, or make any event loop modifications>. If you need to stop it,	1393	ever, or make ANY event loop modifications whatsoever>.
1250	return C<now + 1e30> (or so, fudge fudge) and stop it afterwards (e.g. by
1251	starting an C<ev_prepare> watcher, which is legal).
1252		1394
		1395	If you need to stop it, return C<now + 1e30> (or so, fudge fudge) and stop
		1396	it afterwards (e.g. by starting an C<ev_prepare> watcher, which is the
		1397	only event loop modification you are allowed to do).
		1398
1253	Its prototype is C<ev_tstamp (*reschedule_cb)(struct ev_periodic *w,	1399	The callback prototype is C<ev_tstamp (*reschedule_cb)(struct ev_periodic
1254	ev_tstamp now)>, e.g.:	1400	*w, ev_tstamp now)>, e.g.:
1255		1401
1256	static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)	1402	static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)
1257	{	1403	{
1258	return now + 60.;	1404	return now + 60.;
1259	}	1405	}
…		…
1261	It must return the next time to trigger, based on the passed time value	1407	It must return the next time to trigger, based on the passed time value
1262	(that is, the lowest time value larger than to the second argument). It	1408	(that is, the lowest time value larger than to the second argument). It
1263	will usually be called just before the callback will be triggered, but	1409	will usually be called just before the callback will be triggered, but
1264	might be called at other times, too.	1410	might be called at other times, too.
1265		1411
1266	NOTE: I<< This callback must always return a time that is later than the	1412	NOTE: I<< This callback must always return a time that is higher than or
1267	passed C<now> value >>. Not even C<now> itself will do, it I<must> be larger.	1413	equal to the passed C<now> value >>.
1268		1414
1269	This can be used to create very complex timers, such as a timer that	1415	This can be used to create very complex timers, such as a timer that
1270	triggers on each midnight, local time. To do this, you would calculate the	1416	triggers on "next midnight, local time". To do this, you would calculate the
1271	next midnight after C<now> and return the timestamp value for this. How	1417	next midnight after C<now> and return the timestamp value for this. How
1272	you do this is, again, up to you (but it is not trivial, which is the main	1418	you do this is, again, up to you (but it is not trivial, which is the main
1273	reason I omitted it as an example).	1419	reason I omitted it as an example).
1274		1420
1275	=back	1421	=back
…		…
1279	Simply stops and restarts the periodic watcher again. This is only useful	1425	Simply stops and restarts the periodic watcher again. This is only useful
1280	when you changed some parameters or the reschedule callback would return	1426	when you changed some parameters or the reschedule callback would return
1281	a different time than the last time it was called (e.g. in a crond like	1427	a different time than the last time it was called (e.g. in a crond like
1282	program when the crontabs have changed).	1428	program when the crontabs have changed).
1283		1429
		1430	=item ev_tstamp ev_periodic_at (ev_periodic *)
		1431
		1432	When active, returns the absolute time that the watcher is supposed to
		1433	trigger next.
		1434
1284	=item ev_tstamp offset [read-write]	1435	=item ev_tstamp offset [read-write]
1285		1436
1286	When repeating, this contains the offset value, otherwise this is the	1437	When repeating, this contains the offset value, otherwise this is the
1287	absolute point in time (the C<at> value passed to C<ev_periodic_set>).	1438	absolute point in time (the C<at> value passed to C<ev_periodic_set>).
1288		1439
…		…
1299		1450
1300	The current reschedule callback, or C<0>, if this functionality is	1451	The current reschedule callback, or C<0>, if this functionality is
1301	switched off. Can be changed any time, but changes only take effect when	1452	switched off. Can be changed any time, but changes only take effect when
1302	the periodic timer fires or C<ev_periodic_again> is being called.	1453	the periodic timer fires or C<ev_periodic_again> is being called.
1303		1454
1304	=item ev_tstamp at [read-only]
1305
1306	When active, contains the absolute time that the watcher is supposed to
1307	trigger next.
1308
1309	=back	1455	=back
		1456
		1457	=head3 Examples
1310		1458
1311	Example: Call a callback every hour, or, more precisely, whenever the	1459	Example: Call a callback every hour, or, more precisely, whenever the
1312	system clock is divisible by 3600. The callback invocation times have	1460	system clock is divisible by 3600. The callback invocation times have
1313	potentially a lot of jittering, but good long-term stability.	1461	potentially a lot of jitter, but good long-term stability.
1314		1462
1315	static void	1463	static void
1316	clock_cb (struct ev_loop *loop, struct ev_io *w, int revents)	1464	clock_cb (struct ev_loop *loop, struct ev_io *w, int revents)
1317	{	1465	{
1318	... its now a full hour (UTC, or TAI or whatever your clock follows)	1466	... its now a full hour (UTC, or TAI or whatever your clock follows)
1319	}	1467	}
1320		1468
1321	struct ev_periodic hourly_tick;	1469	struct ev_periodic hourly_tick;
1322	ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);	1470	ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);
1323	ev_periodic_start (loop, &hourly_tick);	1471	ev_periodic_start (loop, &hourly_tick);
1324		1472
1325	Example: The same as above, but use a reschedule callback to do it:	1473	Example: The same as above, but use a reschedule callback to do it:
1326		1474
1327	#include <math.h>	1475	#include <math.h>
1328		1476
1329	static ev_tstamp	1477	static ev_tstamp
1330	my_scheduler_cb (struct ev_periodic *w, ev_tstamp now)	1478	my_scheduler_cb (struct ev_periodic *w, ev_tstamp now)
1331	{	1479	{
1332	return fmod (now, 3600.) + 3600.;	1480	return fmod (now, 3600.) + 3600.;
1333	}	1481	}
1334		1482
1335	ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);	1483	ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
1336		1484
1337	Example: Call a callback every hour, starting now:	1485	Example: Call a callback every hour, starting now:
1338		1486
1339	struct ev_periodic hourly_tick;	1487	struct ev_periodic hourly_tick;
1340	ev_periodic_init (&hourly_tick, clock_cb,	1488	ev_periodic_init (&hourly_tick, clock_cb,
1341	fmod (ev_now (loop), 3600.), 3600., 0);	1489	fmod (ev_now (loop), 3600.), 3600., 0);
1342	ev_periodic_start (loop, &hourly_tick);	1490	ev_periodic_start (loop, &hourly_tick);
1343		1491
1344		1492
1345	=head2 C<ev_signal> - signal me when a signal gets signalled!	1493	=head2 C<ev_signal> - signal me when a signal gets signalled!
1346		1494
1347	Signal watchers will trigger an event when the process receives a specific	1495	Signal watchers will trigger an event when the process receives a specific
…		…
1354	with the kernel (thus it coexists with your own signal handlers as long	1502	with the kernel (thus it coexists with your own signal handlers as long
1355	as you don't register any with libev). Similarly, when the last signal	1503	as you don't register any with libev). Similarly, when the last signal
1356	watcher for a signal is stopped libev will reset the signal handler to	1504	watcher for a signal is stopped libev will reset the signal handler to
1357	SIG_DFL (regardless of what it was set to before).	1505	SIG_DFL (regardless of what it was set to before).
1358		1506
		1507	If possible and supported, libev will install its handlers with
		1508	C<SA_RESTART> behaviour enabled, so system calls should not be unduly
		1509	interrupted. If you have a problem with system calls getting interrupted by
		1510	signals you can block all signals in an C<ev_check> watcher and unblock
		1511	them in an C<ev_prepare> watcher.
		1512
1359	=head3 Watcher-Specific Functions and Data Members	1513	=head3 Watcher-Specific Functions and Data Members
1360		1514
1361	=over 4	1515	=over 4
1362		1516
1363	=item ev_signal_init (ev_signal *, callback, int signum)	1517	=item ev_signal_init (ev_signal *, callback, int signum)
…		…
1371		1525
1372	The signal the watcher watches out for.	1526	The signal the watcher watches out for.
1373		1527
1374	=back	1528	=back
1375		1529
		1530	=head3 Examples
		1531
		1532	Example: Try to exit cleanly on SIGINT and SIGTERM.
		1533
		1534	static void
		1535	sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)
		1536	{
		1537	ev_unloop (loop, EVUNLOOP_ALL);
		1538	}
		1539
		1540	struct ev_signal signal_watcher;
		1541	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
		1542	ev_signal_start (loop, &sigint_cb);
		1543
1376		1544
1377	=head2 C<ev_child> - watch out for process status changes	1545	=head2 C<ev_child> - watch out for process status changes
1378		1546
1379	Child watchers trigger when your process receives a SIGCHLD in response to	1547	Child watchers trigger when your process receives a SIGCHLD in response to
1380	some child status changes (most typically when a child of yours dies).	1548	some child status changes (most typically when a child of yours dies). It
		1549	is permissible to install a child watcher I<after> the child has been
		1550	forked (which implies it might have already exited), as long as the event
		1551	loop isn't entered (or is continued from a watcher).
		1552
		1553	Only the default event loop is capable of handling signals, and therefore
		1554	you can only register child watchers in the default event loop.
		1555
		1556	=head3 Process Interaction
		1557
		1558	Libev grabs C<SIGCHLD> as soon as the default event loop is
		1559	initialised. This is necessary to guarantee proper behaviour even if
		1560	the first child watcher is started after the child exits. The occurrence
		1561	of C<SIGCHLD> is recorded asynchronously, but child reaping is done
		1562	synchronously as part of the event loop processing. Libev always reaps all
		1563	children, even ones not watched.
		1564
		1565	=head3 Overriding the Built-In Processing
		1566
		1567	Libev offers no special support for overriding the built-in child
		1568	processing, but if your application collides with libev's default child
		1569	handler, you can override it easily by installing your own handler for
		1570	C<SIGCHLD> after initialising the default loop, and making sure the
		1571	default loop never gets destroyed. You are encouraged, however, to use an
		1572	event-based approach to child reaping and thus use libev's support for
		1573	that, so other libev users can use C<ev_child> watchers freely.
1381		1574
1382	=head3 Watcher-Specific Functions and Data Members	1575	=head3 Watcher-Specific Functions and Data Members
1383		1576
1384	=over 4	1577	=over 4
1385		1578
1386	=item ev_child_init (ev_child *, callback, int pid)	1579	=item ev_child_init (ev_child *, callback, int pid, int trace)
1387		1580
1388	=item ev_child_set (ev_child *, int pid)	1581	=item ev_child_set (ev_child *, int pid, int trace)
1389		1582
1390	Configures the watcher to wait for status changes of process C<pid> (or	1583	Configures the watcher to wait for status changes of process C<pid> (or
1391	I<any> process if C<pid> is specified as C<0>). The callback can look	1584	I<any> process if C<pid> is specified as C<0>). The callback can look
1392	at the C<rstatus> member of the C<ev_child> watcher structure to see	1585	at the C<rstatus> member of the C<ev_child> watcher structure to see
1393	the status word (use the macros from C<sys/wait.h> and see your systems	1586	the status word (use the macros from C<sys/wait.h> and see your systems
1394	C<waitpid> documentation). The C<rpid> member contains the pid of the	1587	C<waitpid> documentation). The C<rpid> member contains the pid of the
1395	process causing the status change.	1588	process causing the status change. C<trace> must be either C<0> (only
		1589	activate the watcher when the process terminates) or C<1> (additionally
		1590	activate the watcher when the process is stopped or continued).
1396		1591
1397	=item int pid [read-only]	1592	=item int pid [read-only]
1398		1593
1399	The process id this watcher watches out for, or C<0>, meaning any process id.	1594	The process id this watcher watches out for, or C<0>, meaning any process id.
1400		1595
…		…
1407	The process exit/trace status caused by C<rpid> (see your systems	1602	The process exit/trace status caused by C<rpid> (see your systems
1408	C<waitpid> and C<sys/wait.h> documentation for details).	1603	C<waitpid> and C<sys/wait.h> documentation for details).
1409		1604
1410	=back	1605	=back
1411		1606
1412	Example: Try to exit cleanly on SIGINT and SIGTERM.	1607	=head3 Examples
1413		1608
		1609	Example: C<fork()> a new process and install a child handler to wait for
		1610	its completion.
		1611
		1612	ev_child cw;
		1613
1414	static void	1614	static void
1415	sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)	1615	child_cb (EV_P_ struct ev_child *w, int revents)
1416	{	1616	{
1417	ev_unloop (loop, EVUNLOOP_ALL);	1617	ev_child_stop (EV_A_ w);
		1618	printf ("process %d exited with status %x\n", w->rpid, w->rstatus);
1418	}	1619	}
1419		1620
1420	struct ev_signal signal_watcher;	1621	pid_t pid = fork ();
1421	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);	1622
1422	ev_signal_start (loop, &sigint_cb);	1623	if (pid < 0)
		1624	// error
		1625	else if (pid == 0)
		1626	{
		1627	// the forked child executes here
		1628	exit (1);
		1629	}
		1630	else
		1631	{
		1632	ev_child_init (&cw, child_cb, pid, 0);
		1633	ev_child_start (EV_DEFAULT_ &cw);
		1634	}
1423		1635
1424		1636
1425	=head2 C<ev_stat> - did the file attributes just change?	1637	=head2 C<ev_stat> - did the file attributes just change?
1426		1638
1427	This watches a filesystem path for attribute changes. That is, it calls	1639	This watches a file system path for attribute changes. That is, it calls
1428	C<stat> regularly (or when the OS says it changed) and sees if it changed	1640	C<stat> regularly (or when the OS says it changed) and sees if it changed
1429	compared to the last time, invoking the callback if it did.	1641	compared to the last time, invoking the callback if it did.
1430		1642
1431	The path does not need to exist: changing from "path exists" to "path does	1643	The path does not need to exist: changing from "path exists" to "path does
1432	not exist" is a status change like any other. The condition "path does	1644	not exist" is a status change like any other. The condition "path does
…		…
1450	as even with OS-supported change notifications, this can be	1662	as even with OS-supported change notifications, this can be
1451	resource-intensive.	1663	resource-intensive.
1452		1664
1453	At the time of this writing, only the Linux inotify interface is	1665	At the time of this writing, only the Linux inotify interface is
1454	implemented (implementing kqueue support is left as an exercise for the	1666	implemented (implementing kqueue support is left as an exercise for the
		1667	reader, note, however, that the author sees no way of implementing ev_stat
1455	reader). Inotify will be used to give hints only and should not change the	1668	semantics with kqueue). Inotify will be used to give hints only and should
1456	semantics of C<ev_stat> watchers, which means that libev sometimes needs	1669	not change the semantics of C<ev_stat> watchers, which means that libev
1457	to fall back to regular polling again even with inotify, but changes are	1670	sometimes needs to fall back to regular polling again even with inotify,
1458	usually detected immediately, and if the file exists there will be no	1671	but changes are usually detected immediately, and if the file exists there
1459	polling.	1672	will be no polling.
		1673
		1674	=head3 ABI Issues (Largefile Support)
		1675
		1676	Libev by default (unless the user overrides this) uses the default
		1677	compilation environment, which means that on systems with large file
		1678	support disabled by default, you get the 32 bit version of the stat
		1679	structure. When using the library from programs that change the ABI to
		1680	use 64 bit file offsets the programs will fail. In that case you have to
		1681	compile libev with the same flags to get binary compatibility. This is
		1682	obviously the case with any flags that change the ABI, but the problem is
		1683	most noticeably disabled with ev_stat and large file support.
		1684
		1685	The solution for this is to lobby your distribution maker to make large
		1686	file interfaces available by default (as e.g. FreeBSD does) and not
		1687	optional. Libev cannot simply switch on large file support because it has
		1688	to exchange stat structures with application programs compiled using the
		1689	default compilation environment.
		1690
		1691	=head3 Inotify
		1692
		1693	When C<inotify (7)> support has been compiled into libev (generally only
		1694	available on Linux) and present at runtime, it will be used to speed up
		1695	change detection where possible. The inotify descriptor will be created lazily
		1696	when the first C<ev_stat> watcher is being started.
		1697
		1698	Inotify presence does not change the semantics of C<ev_stat> watchers
		1699	except that changes might be detected earlier, and in some cases, to avoid
		1700	making regular C<stat> calls. Even in the presence of inotify support
		1701	there are many cases where libev has to resort to regular C<stat> polling.
		1702
		1703	(There is no support for kqueue, as apparently it cannot be used to
		1704	implement this functionality, due to the requirement of having a file
		1705	descriptor open on the object at all times).
		1706
		1707	=head3 The special problem of stat time resolution
		1708
		1709	The C<stat ()> system call only supports full-second resolution portably, and
		1710	even on systems where the resolution is higher, many file systems still
		1711	only support whole seconds.
		1712
		1713	That means that, if the time is the only thing that changes, you can
		1714	easily miss updates: on the first update, C<ev_stat> detects a change and
		1715	calls your callback, which does something. When there is another update
		1716	within the same second, C<ev_stat> will be unable to detect it as the stat
		1717	data does not change.
		1718
		1719	The solution to this is to delay acting on a change for slightly more
		1720	than a second (or till slightly after the next full second boundary), using
		1721	a roughly one-second-delay C<ev_timer> (e.g. C<ev_timer_set (w, 0., 1.02);
		1722	ev_timer_again (loop, w)>).
		1723
		1724	The C<.02> offset is added to work around small timing inconsistencies
		1725	of some operating systems (where the second counter of the current time
		1726	might be be delayed. One such system is the Linux kernel, where a call to
		1727	C<gettimeofday> might return a timestamp with a full second later than
		1728	a subsequent C<time> call - if the equivalent of C<time ()> is used to
		1729	update file times then there will be a small window where the kernel uses
		1730	the previous second to update file times but libev might already execute
		1731	the timer callback).
1460		1732
1461	=head3 Watcher-Specific Functions and Data Members	1733	=head3 Watcher-Specific Functions and Data Members
1462		1734
1463	=over 4	1735	=over 4
1464		1736
…		…
1470	C<path>. The C<interval> is a hint on how quickly a change is expected to	1742	C<path>. The C<interval> is a hint on how quickly a change is expected to
1471	be detected and should normally be specified as C<0> to let libev choose	1743	be detected and should normally be specified as C<0> to let libev choose
1472	a suitable value. The memory pointed to by C<path> must point to the same	1744	a suitable value. The memory pointed to by C<path> must point to the same
1473	path for as long as the watcher is active.	1745	path for as long as the watcher is active.
1474		1746
1475	The callback will be receive C<EV_STAT> when a change was detected,	1747	The callback will receive C<EV_STAT> when a change was detected, relative
1476	relative to the attributes at the time the watcher was started (or the	1748	to the attributes at the time the watcher was started (or the last change
1477	last change was detected).	1749	was detected).
1478		1750
1479	=item ev_stat_stat (ev_stat *)	1751	=item ev_stat_stat (loop, ev_stat *)
1480		1752
1481	Updates the stat buffer immediately with new values. If you change the	1753	Updates the stat buffer immediately with new values. If you change the
1482	watched path in your callback, you could call this fucntion to avoid	1754	watched path in your callback, you could call this function to avoid
1483	detecting this change (while introducing a race condition). Can also be	1755	detecting this change (while introducing a race condition if you are not
1484	useful simply to find out the new values.	1756	the only one changing the path). Can also be useful simply to find out the
		1757	new values.
1485		1758
1486	=item ev_statdata attr [read-only]	1759	=item ev_statdata attr [read-only]
1487		1760
1488	The most-recently detected attributes of the file. Although the type is of	1761	The most-recently detected attributes of the file. Although the type is
1489	C<ev_statdata>, this is usually the (or one of the) C<struct stat> types	1762	C<ev_statdata>, this is usually the (or one of the) C<struct stat> types
1490	suitable for your system. If the C<st_nlink> member is C<0>, then there	1763	suitable for your system, but you can only rely on the POSIX-standardised
		1764	members to be present. If the C<st_nlink> member is C<0>, then there was
1491	was some error while C<stat>ing the file.	1765	some error while C<stat>ing the file.
1492		1766
1493	=item ev_statdata prev [read-only]	1767	=item ev_statdata prev [read-only]
1494		1768
1495	The previous attributes of the file. The callback gets invoked whenever	1769	The previous attributes of the file. The callback gets invoked whenever
1496	C<prev> != C<attr>.	1770	C<prev> != C<attr>, or, more precisely, one or more of these members
		1771	differ: C<st_dev>, C<st_ino>, C<st_mode>, C<st_nlink>, C<st_uid>,
		1772	C<st_gid>, C<st_rdev>, C<st_size>, C<st_atime>, C<st_mtime>, C<st_ctime>.
1497		1773
1498	=item ev_tstamp interval [read-only]	1774	=item ev_tstamp interval [read-only]
1499		1775
1500	The specified interval.	1776	The specified interval.
1501		1777
1502	=item const char *path [read-only]	1778	=item const char *path [read-only]
1503		1779
1504	The filesystem path that is being watched.	1780	The file system path that is being watched.
1505		1781
1506	=back	1782	=back
1507		1783
		1784	=head3 Examples
		1785
1508	Example: Watch C</etc/passwd> for attribute changes.	1786	Example: Watch C</etc/passwd> for attribute changes.
1509		1787
1510	static void	1788	static void
1511	passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)	1789	passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)
1512	{	1790	{
1513	/* /etc/passwd changed in some way */	1791	/* /etc/passwd changed in some way */
1514	if (w->attr.st_nlink)	1792	if (w->attr.st_nlink)
1515	{	1793	{
1516	printf ("passwd current size %ld\n", (long)w->attr.st_size);	1794	printf ("passwd current size %ld\n", (long)w->attr.st_size);
1517	printf ("passwd current atime %ld\n", (long)w->attr.st_mtime);	1795	printf ("passwd current atime %ld\n", (long)w->attr.st_mtime);
1518	printf ("passwd current mtime %ld\n", (long)w->attr.st_mtime);	1796	printf ("passwd current mtime %ld\n", (long)w->attr.st_mtime);
1519	}	1797	}
1520	else	1798	else
1521	/* you shalt not abuse printf for puts */	1799	/* you shalt not abuse printf for puts */
1522	puts ("wow, /etc/passwd is not there, expect problems. "	1800	puts ("wow, /etc/passwd is not there, expect problems. "
1523	"if this is windows, they already arrived\n");	1801	"if this is windows, they already arrived\n");
1524	}	1802	}
1525		1803
1526	...	1804	...
1527	ev_stat passwd;	1805	ev_stat passwd;
1528		1806
1529	ev_stat_init (&passwd, passwd_cb, "/etc/passwd");	1807	ev_stat_init (&passwd, passwd_cb, "/etc/passwd", 0.);
1530	ev_stat_start (loop, &passwd);	1808	ev_stat_start (loop, &passwd);
		1809
		1810	Example: Like above, but additionally use a one-second delay so we do not
		1811	miss updates (however, frequent updates will delay processing, too, so
		1812	one might do the work both on C<ev_stat> callback invocation I<and> on
		1813	C<ev_timer> callback invocation).
		1814
		1815	static ev_stat passwd;
		1816	static ev_timer timer;
		1817
		1818	static void
		1819	timer_cb (EV_P_ ev_timer *w, int revents)
		1820	{
		1821	ev_timer_stop (EV_A_ w);
		1822
		1823	/* now it's one second after the most recent passwd change */
		1824	}
		1825
		1826	static void
		1827	stat_cb (EV_P_ ev_stat *w, int revents)
		1828	{
		1829	/* reset the one-second timer */
		1830	ev_timer_again (EV_A_ &timer);
		1831	}
		1832
		1833	...
		1834	ev_stat_init (&passwd, stat_cb, "/etc/passwd", 0.);
		1835	ev_stat_start (loop, &passwd);
		1836	ev_timer_init (&timer, timer_cb, 0., 1.02);
1531		1837
1532		1838
1533	=head2 C<ev_idle> - when you've got nothing better to do...	1839	=head2 C<ev_idle> - when you've got nothing better to do...
1534		1840
1535	Idle watchers trigger events when no other events of the same or higher	1841	Idle watchers trigger events when no other events of the same or higher
…		…
1561	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,	1867	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,
1562	believe me.	1868	believe me.
1563		1869
1564	=back	1870	=back
1565		1871
		1872	=head3 Examples
		1873
1566	Example: Dynamically allocate an C<ev_idle> watcher, start it, and in the	1874	Example: Dynamically allocate an C<ev_idle> watcher, start it, and in the
1567	callback, free it. Also, use no error checking, as usual.	1875	callback, free it. Also, use no error checking, as usual.
1568		1876
1569	static void	1877	static void
1570	idle_cb (struct ev_loop *loop, struct ev_idle *w, int revents)	1878	idle_cb (struct ev_loop *loop, struct ev_idle *w, int revents)
1571	{	1879	{
1572	free (w);	1880	free (w);
1573	// now do something you wanted to do when the program has	1881	// now do something you wanted to do when the program has
1574	// no longer asnything immediate to do.	1882	// no longer anything immediate to do.
1575	}	1883	}
1576		1884
1577	struct ev_idle *idle_watcher = malloc (sizeof (struct ev_idle));	1885	struct ev_idle *idle_watcher = malloc (sizeof (struct ev_idle));
1578	ev_idle_init (idle_watcher, idle_cb);	1886	ev_idle_init (idle_watcher, idle_cb);
1579	ev_idle_start (loop, idle_cb);	1887	ev_idle_start (loop, idle_cb);
1580		1888
1581		1889
1582	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!	1890	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!
1583		1891
1584	Prepare and check watchers are usually (but not always) used in tandem:	1892	Prepare and check watchers are usually (but not always) used in tandem:
…		…
1603		1911
1604	This is done by examining in each prepare call which file descriptors need	1912	This is done by examining in each prepare call which file descriptors need
1605	to be watched by the other library, registering C<ev_io> watchers for	1913	to be watched by the other library, registering C<ev_io> watchers for
1606	them and starting an C<ev_timer> watcher for any timeouts (many libraries	1914	them and starting an C<ev_timer> watcher for any timeouts (many libraries
1607	provide just this functionality). Then, in the check watcher you check for	1915	provide just this functionality). Then, in the check watcher you check for
1608	any events that occured (by checking the pending status of all watchers	1916	any events that occurred (by checking the pending status of all watchers
1609	and stopping them) and call back into the library. The I/O and timer	1917	and stopping them) and call back into the library. The I/O and timer
1610	callbacks will never actually be called (but must be valid nevertheless,	1918	callbacks will never actually be called (but must be valid nevertheless,
1611	because you never know, you know?).	1919	because you never know, you know?).
1612		1920
1613	As another example, the Perl Coro module uses these hooks to integrate	1921	As another example, the Perl Coro module uses these hooks to integrate
…		…
1621		1929
1622	It is recommended to give C<ev_check> watchers highest (C<EV_MAXPRI>)	1930	It is recommended to give C<ev_check> watchers highest (C<EV_MAXPRI>)
1623	priority, to ensure that they are being run before any other watchers	1931	priority, to ensure that they are being run before any other watchers
1624	after the poll. Also, C<ev_check> watchers (and C<ev_prepare> watchers,	1932	after the poll. Also, C<ev_check> watchers (and C<ev_prepare> watchers,
1625	too) should not activate ("feed") events into libev. While libev fully	1933	too) should not activate ("feed") events into libev. While libev fully
1626	supports this, they will be called before other C<ev_check> watchers did	1934	supports this, they might get executed before other C<ev_check> watchers
1627	their job. As C<ev_check> watchers are often used to embed other event	1935	did their job. As C<ev_check> watchers are often used to embed other
1628	loops those other event loops might be in an unusable state until their	1936	(non-libev) event loops those other event loops might be in an unusable
1629	C<ev_check> watcher ran (always remind yourself to coexist peacefully with	1937	state until their C<ev_check> watcher ran (always remind yourself to
1630	others).	1938	coexist peacefully with others).
1631		1939
1632	=head3 Watcher-Specific Functions and Data Members	1940	=head3 Watcher-Specific Functions and Data Members
1633		1941
1634	=over 4	1942	=over 4
1635		1943
…		…
1641	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>	1949	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>
1642	macros, but using them is utterly, utterly and completely pointless.	1950	macros, but using them is utterly, utterly and completely pointless.
1643		1951
1644	=back	1952	=back
1645		1953
		1954	=head3 Examples
		1955
1646	There are a number of principal ways to embed other event loops or modules	1956	There are a number of principal ways to embed other event loops or modules
1647	into libev. Here are some ideas on how to include libadns into libev	1957	into libev. Here are some ideas on how to include libadns into libev
1648	(there is a Perl module named C<EV::ADNS> that does this, which you could	1958	(there is a Perl module named C<EV::ADNS> that does this, which you could
1649	use for an actually working example. Another Perl module named C<EV::Glib>	1959	use as a working example. Another Perl module named C<EV::Glib> embeds a
1650	embeds a Glib main context into libev, and finally, C<Glib::EV> embeds EV	1960	Glib main context into libev, and finally, C<Glib::EV> embeds EV into the
1651	into the Glib event loop).	1961	Glib event loop).
1652		1962
1653	Method 1: Add IO watchers and a timeout watcher in a prepare handler,	1963	Method 1: Add IO watchers and a timeout watcher in a prepare handler,
1654	and in a check watcher, destroy them and call into libadns. What follows	1964	and in a check watcher, destroy them and call into libadns. What follows
1655	is pseudo-code only of course. This requires you to either use a low	1965	is pseudo-code only of course. This requires you to either use a low
1656	priority for the check watcher or use C<ev_clear_pending> explicitly, as	1966	priority for the check watcher or use C<ev_clear_pending> explicitly, as
1657	the callbacks for the IO/timeout watchers might not have been called yet.	1967	the callbacks for the IO/timeout watchers might not have been called yet.
1658		1968
1659	static ev_io iow [nfd];	1969	static ev_io iow [nfd];
1660	static ev_timer tw;	1970	static ev_timer tw;
1661		1971
1662	static void	1972	static void
1663	io_cb (ev_loop *loop, ev_io *w, int revents)	1973	io_cb (ev_loop *loop, ev_io *w, int revents)
1664	{	1974	{
1665	}	1975	}
1666		1976
1667	// create io watchers for each fd and a timer before blocking	1977	// create io watchers for each fd and a timer before blocking
1668	static void	1978	static void
1669	adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents)	1979	adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents)
1670	{	1980	{
1671	int timeout = 3600000;	1981	int timeout = 3600000;
1672	struct pollfd fds [nfd];	1982	struct pollfd fds [nfd];
1673	// actual code will need to loop here and realloc etc.	1983	// actual code will need to loop here and realloc etc.
1674	adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));	1984	adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));
1675		1985
1676	/* the callback is illegal, but won't be called as we stop during check */	1986	/* the callback is illegal, but won't be called as we stop during check */
1677	ev_timer_init (&tw, 0, timeout * 1e-3);	1987	ev_timer_init (&tw, 0, timeout * 1e-3);
1678	ev_timer_start (loop, &tw);	1988	ev_timer_start (loop, &tw);
1679		1989
1680	// create one ev_io per pollfd	1990	// create one ev_io per pollfd
1681	for (int i = 0; i < nfd; ++i)	1991	for (int i = 0; i < nfd; ++i)
1682	{	1992	{
1683	ev_io_init (iow + i, io_cb, fds [i].fd,	1993	ev_io_init (iow + i, io_cb, fds [i].fd,
1684	((fds [i].events & POLLIN ? EV_READ : 0)	1994	((fds [i].events & POLLIN ? EV_READ : 0)
1685	| (fds [i].events & POLLOUT ? EV_WRITE : 0)));	1995	| (fds [i].events & POLLOUT ? EV_WRITE : 0)));
1686		1996
1687	fds [i].revents = 0;	1997	fds [i].revents = 0;
1688	ev_io_start (loop, iow + i);	1998	ev_io_start (loop, iow + i);
1689	}	1999	}
1690	}	2000	}
1691		2001
1692	// stop all watchers after blocking	2002	// stop all watchers after blocking
1693	static void	2003	static void
1694	adns_check_cb (ev_loop *loop, ev_check *w, int revents)	2004	adns_check_cb (ev_loop *loop, ev_check *w, int revents)
1695	{	2005	{
1696	ev_timer_stop (loop, &tw);	2006	ev_timer_stop (loop, &tw);
1697		2007
1698	for (int i = 0; i < nfd; ++i)	2008	for (int i = 0; i < nfd; ++i)
1699	{	2009	{
1700	// set the relevant poll flags	2010	// set the relevant poll flags
1701	// could also call adns_processreadable etc. here	2011	// could also call adns_processreadable etc. here
1702	struct pollfd *fd = fds + i;	2012	struct pollfd *fd = fds + i;
1703	int revents = ev_clear_pending (iow + i);	2013	int revents = ev_clear_pending (iow + i);
1704	if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;	2014	if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;
1705	if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;	2015	if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;
1706		2016
1707	// now stop the watcher	2017	// now stop the watcher
1708	ev_io_stop (loop, iow + i);	2018	ev_io_stop (loop, iow + i);
1709	}	2019	}
1710		2020
1711	adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));	2021	adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));
1712	}	2022	}
1713		2023
1714	Method 2: This would be just like method 1, but you run C<adns_afterpoll>	2024	Method 2: This would be just like method 1, but you run C<adns_afterpoll>
1715	in the prepare watcher and would dispose of the check watcher.	2025	in the prepare watcher and would dispose of the check watcher.
1716		2026
1717	Method 3: If the module to be embedded supports explicit event	2027	Method 3: If the module to be embedded supports explicit event
1718	notification (adns does), you can also make use of the actual watcher	2028	notification (libadns does), you can also make use of the actual watcher
1719	callbacks, and only destroy/create the watchers in the prepare watcher.	2029	callbacks, and only destroy/create the watchers in the prepare watcher.
1720		2030
1721	static void	2031	static void
1722	timer_cb (EV_P_ ev_timer *w, int revents)	2032	timer_cb (EV_P_ ev_timer *w, int revents)
1723	{	2033	{
1724	adns_state ads = (adns_state)w->data;	2034	adns_state ads = (adns_state)w->data;
1725	update_now (EV_A);	2035	update_now (EV_A);
1726		2036
1727	adns_processtimeouts (ads, &tv_now);	2037	adns_processtimeouts (ads, &tv_now);
1728	}	2038	}
1729		2039
1730	static void	2040	static void
1731	io_cb (EV_P_ ev_io *w, int revents)	2041	io_cb (EV_P_ ev_io *w, int revents)
1732	{	2042	{
1733	adns_state ads = (adns_state)w->data;	2043	adns_state ads = (adns_state)w->data;
1734	update_now (EV_A);	2044	update_now (EV_A);
1735		2045
1736	if (revents & EV_READ ) adns_processreadable (ads, w->fd, &tv_now);	2046	if (revents & EV_READ ) adns_processreadable (ads, w->fd, &tv_now);
1737	if (revents & EV_WRITE) adns_processwriteable (ads, w->fd, &tv_now);	2047	if (revents & EV_WRITE) adns_processwriteable (ads, w->fd, &tv_now);
1738	}	2048	}
1739		2049
1740	// do not ever call adns_afterpoll	2050	// do not ever call adns_afterpoll
1741		2051
1742	Method 4: Do not use a prepare or check watcher because the module you	2052	Method 4: Do not use a prepare or check watcher because the module you
1743	want to embed is too inflexible to support it. Instead, youc na override	2053	want to embed is too inflexible to support it. Instead, you can override
1744	their poll function. The drawback with this solution is that the main	2054	their poll function. The drawback with this solution is that the main
1745	loop is now no longer controllable by EV. The C<Glib::EV> module does	2055	loop is now no longer controllable by EV. The C<Glib::EV> module does
1746	this.	2056	this.
1747		2057
1748	static gint	2058	static gint
1749	event_poll_func (GPollFD *fds, guint nfds, gint timeout)	2059	event_poll_func (GPollFD *fds, guint nfds, gint timeout)
1750	{	2060	{
1751	int got_events = 0;	2061	int got_events = 0;
1752		2062
1753	for (n = 0; n < nfds; ++n)	2063	for (n = 0; n < nfds; ++n)
1754	// create/start io watcher that sets the relevant bits in fds[n] and increment got_events	2064	// create/start io watcher that sets the relevant bits in fds[n] and increment got_events
1755		2065
1756	if (timeout >= 0)	2066	if (timeout >= 0)
1757	// create/start timer	2067	// create/start timer
1758		2068
1759	// poll	2069	// poll
1760	ev_loop (EV_A_ 0);	2070	ev_loop (EV_A_ 0);
1761		2071
1762	// stop timer again	2072	// stop timer again
1763	if (timeout >= 0)	2073	if (timeout >= 0)
1764	ev_timer_stop (EV_A_ &to);	2074	ev_timer_stop (EV_A_ &to);
1765		2075
1766	// stop io watchers again - their callbacks should have set	2076	// stop io watchers again - their callbacks should have set
1767	for (n = 0; n < nfds; ++n)	2077	for (n = 0; n < nfds; ++n)
1768	ev_io_stop (EV_A_ iow [n]);	2078	ev_io_stop (EV_A_ iow [n]);
1769		2079
1770	return got_events;	2080	return got_events;
1771	}	2081	}
1772		2082
1773		2083
1774	=head2 C<ev_embed> - when one backend isn't enough...	2084	=head2 C<ev_embed> - when one backend isn't enough...
1775		2085
1776	This is a rather advanced watcher type that lets you embed one event loop	2086	This is a rather advanced watcher type that lets you embed one event loop
1777	into another (currently only C<ev_io> events are supported in the embedded	2087	into another (currently only C<ev_io> events are supported in the embedded
1778	loop, other types of watchers might be handled in a delayed or incorrect	2088	loop, other types of watchers might be handled in a delayed or incorrect
1779	fashion and must not be used). (See portability notes, below).	2089	fashion and must not be used).
1780		2090
1781	There are primarily two reasons you would want that: work around bugs and	2091	There are primarily two reasons you would want that: work around bugs and
1782	prioritise I/O.	2092	prioritise I/O.
1783		2093
1784	As an example for a bug workaround, the kqueue backend might only support	2094	As an example for a bug workaround, the kqueue backend might only support
…		…
1818	portable one.	2128	portable one.
1819		2129
1820	So when you want to use this feature you will always have to be prepared	2130	So when you want to use this feature you will always have to be prepared
1821	that you cannot get an embeddable loop. The recommended way to get around	2131	that you cannot get an embeddable loop. The recommended way to get around
1822	this is to have a separate variables for your embeddable loop, try to	2132	this is to have a separate variables for your embeddable loop, try to
1823	create it, and if that fails, use the normal loop for everything:	2133	create it, and if that fails, use the normal loop for everything.
1824
1825	struct ev_loop *loop_hi = ev_default_init (0);
1826	struct ev_loop *loop_lo = 0;
1827	struct ev_embed embed;
1828
1829	// see if there is a chance of getting one that works
1830	// (remember that a flags value of 0 means autodetection)
1831	loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
1832	? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
1833	: 0;
1834
1835	// if we got one, then embed it, otherwise default to loop_hi
1836	if (loop_lo)
1837	{
1838	ev_embed_init (&embed, 0, loop_lo);
1839	ev_embed_start (loop_hi, &embed);
1840	}
1841	else
1842	loop_lo = loop_hi;
1843
1844	=head2 Portability notes
1845
1846	Kqueue is nominally embeddable, but this is broken on all BSDs that I
1847	tried, in various ways. Usually the embedded event loop will simply never
1848	receive events, sometimes it will only trigger a few times, sometimes in a
1849	loop. Epoll is also nominally embeddable, but many Linux kernel versions
1850	will always eport the epoll fd as ready, even when no events are pending.
1851
1852	While libev allows embedding these backends (they are contained in
1853	C<ev_embeddable_backends ()>), take extreme care that it will actually
1854	work.
1855
1856	When in doubt, create a dynamic event loop forced to use sockets (this
1857	usually works) and possibly another thread and a pipe or so to report to
1858	your main event loop.
1859		2134
1860	=head3 Watcher-Specific Functions and Data Members	2135	=head3 Watcher-Specific Functions and Data Members
1861		2136
1862	=over 4	2137	=over 4
1863		2138
…		…
1867		2142
1868	Configures the watcher to embed the given loop, which must be	2143	Configures the watcher to embed the given loop, which must be
1869	embeddable. If the callback is C<0>, then C<ev_embed_sweep> will be	2144	embeddable. If the callback is C<0>, then C<ev_embed_sweep> will be
1870	invoked automatically, otherwise it is the responsibility of the callback	2145	invoked automatically, otherwise it is the responsibility of the callback
1871	to invoke it (it will continue to be called until the sweep has been done,	2146	to invoke it (it will continue to be called until the sweep has been done,
1872	if you do not want thta, you need to temporarily stop the embed watcher).	2147	if you do not want that, you need to temporarily stop the embed watcher).
1873		2148
1874	=item ev_embed_sweep (loop, ev_embed *)	2149	=item ev_embed_sweep (loop, ev_embed *)
1875		2150
1876	Make a single, non-blocking sweep over the embedded loop. This works	2151	Make a single, non-blocking sweep over the embedded loop. This works
1877	similarly to C<ev_loop (embedded_loop, EVLOOP_NONBLOCK)>, but in the most	2152	similarly to C<ev_loop (embedded_loop, EVLOOP_NONBLOCK)>, but in the most
1878	apropriate way for embedded loops.	2153	appropriate way for embedded loops.
1879		2154
1880	=item struct ev_loop *other [read-only]	2155	=item struct ev_loop *other [read-only]
1881		2156
1882	The embedded event loop.	2157	The embedded event loop.
1883		2158
1884	=back	2159	=back
		2160
		2161	=head3 Examples
		2162
		2163	Example: Try to get an embeddable event loop and embed it into the default
		2164	event loop. If that is not possible, use the default loop. The default
		2165	loop is stored in C<loop_hi>, while the embeddable loop is stored in
		2166	C<loop_lo> (which is C<loop_hi> in the case no embeddable loop can be
		2167	used).
		2168
		2169	struct ev_loop *loop_hi = ev_default_init (0);
		2170	struct ev_loop *loop_lo = 0;
		2171	struct ev_embed embed;
		2172
		2173	// see if there is a chance of getting one that works
		2174	// (remember that a flags value of 0 means autodetection)
		2175	loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
		2176	? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
		2177	: 0;
		2178
		2179	// if we got one, then embed it, otherwise default to loop_hi
		2180	if (loop_lo)
		2181	{
		2182	ev_embed_init (&embed, 0, loop_lo);
		2183	ev_embed_start (loop_hi, &embed);
		2184	}
		2185	else
		2186	loop_lo = loop_hi;
		2187
		2188	Example: Check if kqueue is available but not recommended and create
		2189	a kqueue backend for use with sockets (which usually work with any
		2190	kqueue implementation). Store the kqueue/socket-only event loop in
		2191	C<loop_socket>. (One might optionally use C<EVFLAG_NOENV>, too).
		2192
		2193	struct ev_loop *loop = ev_default_init (0);
		2194	struct ev_loop *loop_socket = 0;
		2195	struct ev_embed embed;
		2196
		2197	if (ev_supported_backends () & ~ev_recommended_backends () & EVBACKEND_KQUEUE)
		2198	if ((loop_socket = ev_loop_new (EVBACKEND_KQUEUE))
		2199	{
		2200	ev_embed_init (&embed, 0, loop_socket);
		2201	ev_embed_start (loop, &embed);
		2202	}
		2203
		2204	if (!loop_socket)
		2205	loop_socket = loop;
		2206
		2207	// now use loop_socket for all sockets, and loop for everything else
1885		2208
1886		2209
1887	=head2 C<ev_fork> - the audacity to resume the event loop after a fork	2210	=head2 C<ev_fork> - the audacity to resume the event loop after a fork
1888		2211
1889	Fork watchers are called when a C<fork ()> was detected (usually because	2212	Fork watchers are called when a C<fork ()> was detected (usually because
…		…
1905	believe me.	2228	believe me.
1906		2229
1907	=back	2230	=back
1908		2231
1909		2232
		2233	=head2 C<ev_async> - how to wake up another event loop
		2234
		2235	In general, you cannot use an C<ev_loop> from multiple threads or other
		2236	asynchronous sources such as signal handlers (as opposed to multiple event
		2237	loops - those are of course safe to use in different threads).
		2238
		2239	Sometimes, however, you need to wake up another event loop you do not
		2240	control, for example because it belongs to another thread. This is what
		2241	C<ev_async> watchers do: as long as the C<ev_async> watcher is active, you
		2242	can signal it by calling C<ev_async_send>, which is thread- and signal
		2243	safe.
		2244
		2245	This functionality is very similar to C<ev_signal> watchers, as signals,
		2246	too, are asynchronous in nature, and signals, too, will be compressed
		2247	(i.e. the number of callback invocations may be less than the number of
		2248	C<ev_async_sent> calls).
		2249
		2250	Unlike C<ev_signal> watchers, C<ev_async> works with any event loop, not
		2251	just the default loop.
		2252
		2253	=head3 Queueing
		2254
		2255	C<ev_async> does not support queueing of data in any way. The reason
		2256	is that the author does not know of a simple (or any) algorithm for a
		2257	multiple-writer-single-reader queue that works in all cases and doesn't
		2258	need elaborate support such as pthreads.
		2259
		2260	That means that if you want to queue data, you have to provide your own
		2261	queue. But at least I can tell you would implement locking around your
		2262	queue:
		2263
		2264	=over 4
		2265
		2266	=item queueing from a signal handler context
		2267
		2268	To implement race-free queueing, you simply add to the queue in the signal
		2269	handler but you block the signal handler in the watcher callback. Here is an example that does that for
		2270	some fictitious SIGUSR1 handler:
		2271
		2272	static ev_async mysig;
		2273
		2274	static void
		2275	sigusr1_handler (void)
		2276	{
		2277	sometype data;
		2278
		2279	// no locking etc.
		2280	queue_put (data);
		2281	ev_async_send (EV_DEFAULT_ &mysig);
		2282	}
		2283
		2284	static void
		2285	mysig_cb (EV_P_ ev_async *w, int revents)
		2286	{
		2287	sometype data;
		2288	sigset_t block, prev;
		2289
		2290	sigemptyset (&block);
		2291	sigaddset (&block, SIGUSR1);
		2292	sigprocmask (SIG_BLOCK, &block, &prev);
		2293
		2294	while (queue_get (&data))
		2295	process (data);
		2296
		2297	if (sigismember (&prev, SIGUSR1)
		2298	sigprocmask (SIG_UNBLOCK, &block, 0);
		2299	}
		2300
		2301	(Note: pthreads in theory requires you to use C<pthread_setmask>
		2302	instead of C<sigprocmask> when you use threads, but libev doesn't do it
		2303	either...).
		2304
		2305	=item queueing from a thread context
		2306
		2307	The strategy for threads is different, as you cannot (easily) block
		2308	threads but you can easily preempt them, so to queue safely you need to
		2309	employ a traditional mutex lock, such as in this pthread example:
		2310
		2311	static ev_async mysig;
		2312	static pthread_mutex_t mymutex = PTHREAD_MUTEX_INITIALIZER;
		2313
		2314	static void
		2315	otherthread (void)
		2316	{
		2317	// only need to lock the actual queueing operation
		2318	pthread_mutex_lock (&mymutex);
		2319	queue_put (data);
		2320	pthread_mutex_unlock (&mymutex);
		2321
		2322	ev_async_send (EV_DEFAULT_ &mysig);
		2323	}
		2324
		2325	static void
		2326	mysig_cb (EV_P_ ev_async *w, int revents)
		2327	{
		2328	pthread_mutex_lock (&mymutex);
		2329
		2330	while (queue_get (&data))
		2331	process (data);
		2332
		2333	pthread_mutex_unlock (&mymutex);
		2334	}
		2335
		2336	=back
		2337
		2338
		2339	=head3 Watcher-Specific Functions and Data Members
		2340
		2341	=over 4
		2342
		2343	=item ev_async_init (ev_async *, callback)
		2344
		2345	Initialises and configures the async watcher - it has no parameters of any
		2346	kind. There is a C<ev_asynd_set> macro, but using it is utterly pointless,
		2347	believe me.
		2348
		2349	=item ev_async_send (loop, ev_async *)
		2350
		2351	Sends/signals/activates the given C<ev_async> watcher, that is, feeds
		2352	an C<EV_ASYNC> event on the watcher into the event loop. Unlike
		2353	C<ev_feed_event>, this call is safe to do in other threads, signal or
		2354	similar contexts (see the discussion of C<EV_ATOMIC_T> in the embedding
		2355	section below on what exactly this means).
		2356
		2357	This call incurs the overhead of a system call only once per loop iteration,
		2358	so while the overhead might be noticeable, it doesn't apply to repeated
		2359	calls to C<ev_async_send>.
		2360
		2361	=item bool = ev_async_pending (ev_async *)
		2362
		2363	Returns a non-zero value when C<ev_async_send> has been called on the
		2364	watcher but the event has not yet been processed (or even noted) by the
		2365	event loop.
		2366
		2367	C<ev_async_send> sets a flag in the watcher and wakes up the loop. When
		2368	the loop iterates next and checks for the watcher to have become active,
		2369	it will reset the flag again. C<ev_async_pending> can be used to very
		2370	quickly check whether invoking the loop might be a good idea.
		2371
		2372	Not that this does I<not> check whether the watcher itself is pending, only
		2373	whether it has been requested to make this watcher pending.
		2374
		2375	=back
		2376
		2377
1910	=head1 OTHER FUNCTIONS	2378	=head1 OTHER FUNCTIONS
1911		2379
1912	There are some other functions of possible interest. Described. Here. Now.	2380	There are some other functions of possible interest. Described. Here. Now.
1913		2381
1914	=over 4	2382	=over 4
…		…
1921	or timeout without having to allocate/configure/start/stop/free one or	2389	or timeout without having to allocate/configure/start/stop/free one or
1922	more watchers yourself.	2390	more watchers yourself.
1923		2391
1924	If C<fd> is less than 0, then no I/O watcher will be started and events	2392	If C<fd> is less than 0, then no I/O watcher will be started and events
1925	is being ignored. Otherwise, an C<ev_io> watcher for the given C<fd> and	2393	is being ignored. Otherwise, an C<ev_io> watcher for the given C<fd> and
1926	C<events> set will be craeted and started.	2394	C<events> set will be created and started.
1927		2395
1928	If C<timeout> is less than 0, then no timeout watcher will be	2396	If C<timeout> is less than 0, then no timeout watcher will be
1929	started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and	2397	started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and
1930	repeat = 0) will be started. While C<0> is a valid timeout, it is of	2398	repeat = 0) will be started. While C<0> is a valid timeout, it is of
1931	dubious value.	2399	dubious value.
…		…
1933	The callback has the type C<void (*cb)(int revents, void *arg)> and gets	2401	The callback has the type C<void (*cb)(int revents, void *arg)> and gets
1934	passed an C<revents> set like normal event callbacks (a combination of	2402	passed an C<revents> set like normal event callbacks (a combination of
1935	C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMEOUT>) and the C<arg>	2403	C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMEOUT>) and the C<arg>
1936	value passed to C<ev_once>:	2404	value passed to C<ev_once>:
1937		2405
1938	static void stdin_ready (int revents, void *arg)	2406	static void stdin_ready (int revents, void *arg)
1939	{	2407	{
1940	if (revents & EV_TIMEOUT)	2408	if (revents & EV_TIMEOUT)
1941	/* doh, nothing entered */;	2409	/* doh, nothing entered */;
1942	else if (revents & EV_READ)	2410	else if (revents & EV_READ)
1943	/* stdin might have data for us, joy! */;	2411	/* stdin might have data for us, joy! */;
1944	}	2412	}
1945		2413
1946	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);	2414	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
1947		2415
1948	=item ev_feed_event (ev_loop *, watcher *, int revents)	2416	=item ev_feed_event (ev_loop *, watcher *, int revents)
1949		2417
1950	Feeds the given event set into the event loop, as if the specified event	2418	Feeds the given event set into the event loop, as if the specified event
1951	had happened for the specified watcher (which must be a pointer to an	2419	had happened for the specified watcher (which must be a pointer to an
…		…
1956	Feed an event on the given fd, as if a file descriptor backend detected	2424	Feed an event on the given fd, as if a file descriptor backend detected
1957	the given events it.	2425	the given events it.
1958		2426
1959	=item ev_feed_signal_event (ev_loop *loop, int signum)	2427	=item ev_feed_signal_event (ev_loop *loop, int signum)
1960		2428
1961	Feed an event as if the given signal occured (C<loop> must be the default	2429	Feed an event as if the given signal occurred (C<loop> must be the default
1962	loop!).	2430	loop!).
1963		2431
1964	=back	2432	=back
1965		2433
1966		2434
…		…
1982		2450
1983	=item * Priorities are not currently supported. Initialising priorities	2451	=item * Priorities are not currently supported. Initialising priorities
1984	will fail and all watchers will have the same priority, even though there	2452	will fail and all watchers will have the same priority, even though there
1985	is an ev_pri field.	2453	is an ev_pri field.
1986		2454
		2455	=item * In libevent, the last base created gets the signals, in libev, the
		2456	first base created (== the default loop) gets the signals.
		2457
1987	=item * Other members are not supported.	2458	=item * Other members are not supported.
1988		2459
1989	=item * The libev emulation is I<not> ABI compatible to libevent, you need	2460	=item * The libev emulation is I<not> ABI compatible to libevent, you need
1990	to use the libev header file and library.	2461	to use the libev header file and library.
1991		2462
1992	=back	2463	=back
1993		2464
1994	=head1 C++ SUPPORT	2465	=head1 C++ SUPPORT
1995		2466
1996	Libev comes with some simplistic wrapper classes for C++ that mainly allow	2467	Libev comes with some simplistic wrapper classes for C++ that mainly allow
1997	you to use some convinience methods to start/stop watchers and also change	2468	you to use some convenience methods to start/stop watchers and also change
1998	the callback model to a model using method callbacks on objects.	2469	the callback model to a model using method callbacks on objects.
1999		2470
2000	To use it,	2471	To use it,
2001		2472
2002	#include <ev++.h>	2473	#include <ev++.h>
2003		2474
2004	This automatically includes F<ev.h> and puts all of its definitions (many	2475	This automatically includes F<ev.h> and puts all of its definitions (many
2005	of them macros) into the global namespace. All C++ specific things are	2476	of them macros) into the global namespace. All C++ specific things are
2006	put into the C<ev> namespace. It should support all the same embedding	2477	put into the C<ev> namespace. It should support all the same embedding
2007	options as F<ev.h>, most notably C<EV_MULTIPLICITY>.	2478	options as F<ev.h>, most notably C<EV_MULTIPLICITY>.
…		…
2074	your compiler is good :), then the method will be fully inlined into the	2545	your compiler is good :), then the method will be fully inlined into the
2075	thunking function, making it as fast as a direct C callback.	2546	thunking function, making it as fast as a direct C callback.
2076		2547
2077	Example: simple class declaration and watcher initialisation	2548	Example: simple class declaration and watcher initialisation
2078		2549
2079	struct myclass	2550	struct myclass
2080	{	2551	{
2081	void io_cb (ev::io &w, int revents) { }	2552	void io_cb (ev::io &w, int revents) { }
2082	}	2553	}
2083		2554
2084	myclass obj;	2555	myclass obj;
2085	ev::io iow;	2556	ev::io iow;
2086	iow.set <myclass, &myclass::io_cb> (&obj);	2557	iow.set <myclass, &myclass::io_cb> (&obj);
2087		2558
2088	=item w->set<function> (void *data = 0)	2559	=item w->set<function> (void *data = 0)
2089		2560
2090	Also sets a callback, but uses a static method or plain function as	2561	Also sets a callback, but uses a static method or plain function as
2091	callback. The optional C<data> argument will be stored in the watcher's	2562	callback. The optional C<data> argument will be stored in the watcher's
…		…
2095		2566
2096	See the method-C<set> above for more details.	2567	See the method-C<set> above for more details.
2097		2568
2098	Example:	2569	Example:
2099		2570
2100	static void io_cb (ev::io &w, int revents) { }	2571	static void io_cb (ev::io &w, int revents) { }
2101	iow.set <io_cb> ();	2572	iow.set <io_cb> ();
2102		2573
2103	=item w->set (struct ev_loop *)	2574	=item w->set (struct ev_loop *)
2104		2575
2105	Associates a different C<struct ev_loop> with this watcher. You can only	2576	Associates a different C<struct ev_loop> with this watcher. You can only
2106	do this when the watcher is inactive (and not pending either).	2577	do this when the watcher is inactive (and not pending either).
2107		2578
2108	=item w->set ([args])	2579	=item w->set ([arguments])
2109		2580
2110	Basically the same as C<ev_TYPE_set>, with the same args. Must be	2581	Basically the same as C<ev_TYPE_set>, with the same arguments. Must be
2111	called at least once. Unlike the C counterpart, an active watcher gets	2582	called at least once. Unlike the C counterpart, an active watcher gets
2112	automatically stopped and restarted when reconfiguring it with this	2583	automatically stopped and restarted when reconfiguring it with this
2113	method.	2584	method.
2114		2585
2115	=item w->start ()	2586	=item w->start ()
…		…
2139	=back	2610	=back
2140		2611
2141	Example: Define a class with an IO and idle watcher, start one of them in	2612	Example: Define a class with an IO and idle watcher, start one of them in
2142	the constructor.	2613	the constructor.
2143		2614
2144	class myclass	2615	class myclass
2145	{	2616	{
2146	ev_io io; void io_cb (ev::io &w, int revents);	2617	ev::io io; void io_cb (ev::io &w, int revents);
2147	ev_idle idle void idle_cb (ev::idle &w, int revents);	2618	ev:idle idle void idle_cb (ev::idle &w, int revents);
2148		2619
2149	myclass ();	2620	myclass (int fd)
2150	}	2621	{
2151
2152	myclass::myclass (int fd)
2153	{
2154	io .set <myclass, &myclass::io_cb > (this);	2622	io .set <myclass, &myclass::io_cb > (this);
2155	idle.set <myclass, &myclass::idle_cb> (this);	2623	idle.set <myclass, &myclass::idle_cb> (this);
2156		2624
2157	io.start (fd, ev::READ);	2625	io.start (fd, ev::READ);
		2626	}
2158	}	2627	};
		2628
		2629
		2630	=head1 OTHER LANGUAGE BINDINGS
		2631
		2632	Libev does not offer other language bindings itself, but bindings for a
		2633	number of languages exist in the form of third-party packages. If you know
		2634	any interesting language binding in addition to the ones listed here, drop
		2635	me a note.
		2636
		2637	=over 4
		2638
		2639	=item Perl
		2640
		2641	The EV module implements the full libev API and is actually used to test
		2642	libev. EV is developed together with libev. Apart from the EV core module,
		2643	there are additional modules that implement libev-compatible interfaces
		2644	to C<libadns> (C<EV::ADNS>), C<Net::SNMP> (C<Net::SNMP::EV>) and the
		2645	C<libglib> event core (C<Glib::EV> and C<EV::Glib>).
		2646
		2647	It can be found and installed via CPAN, its homepage is at
		2648	L<http://software.schmorp.de/pkg/EV>.
		2649
		2650	=item Python
		2651
		2652	Python bindings can be found at L<http://code.google.com/p/pyev/>. It
		2653	seems to be quite complete and well-documented. Note, however, that the
		2654	patch they require for libev is outright dangerous as it breaks the ABI
		2655	for everybody else, and therefore, should never be applied in an installed
		2656	libev (if python requires an incompatible ABI then it needs to embed
		2657	libev).
		2658
		2659	=item Ruby
		2660
		2661	Tony Arcieri has written a ruby extension that offers access to a subset
		2662	of the libev API and adds file handle abstractions, asynchronous DNS and
		2663	more on top of it. It can be found via gem servers. Its homepage is at
		2664	L<http://rev.rubyforge.org/>.
		2665
		2666	=item D
		2667
		2668	Leandro Lucarella has written a D language binding (F<ev.d>) for libev, to
		2669	be found at L<http://git.llucax.com.ar/?p=software/ev.d.git;a=summary>.
		2670
		2671	=back
2159		2672
2160		2673
2161	=head1 MACRO MAGIC	2674	=head1 MACRO MAGIC
2162		2675
2163	Libev can be compiled with a variety of options, the most fundamantal	2676	Libev can be compiled with a variety of options, the most fundamental
2164	of which is C<EV_MULTIPLICITY>. This option determines whether (most)	2677	of which is C<EV_MULTIPLICITY>. This option determines whether (most)
2165	functions and callbacks have an initial C<struct ev_loop *> argument.	2678	functions and callbacks have an initial C<struct ev_loop *> argument.
2166		2679
2167	To make it easier to write programs that cope with either variant, the	2680	To make it easier to write programs that cope with either variant, the
2168	following macros are defined:	2681	following macros are defined:
…		…
2173		2686
2174	This provides the loop I<argument> for functions, if one is required ("ev	2687	This provides the loop I<argument> for functions, if one is required ("ev
2175	loop argument"). The C<EV_A> form is used when this is the sole argument,	2688	loop argument"). The C<EV_A> form is used when this is the sole argument,
2176	C<EV_A_> is used when other arguments are following. Example:	2689	C<EV_A_> is used when other arguments are following. Example:
2177		2690
2178	ev_unref (EV_A);	2691	ev_unref (EV_A);
2179	ev_timer_add (EV_A_ watcher);	2692	ev_timer_add (EV_A_ watcher);
2180	ev_loop (EV_A_ 0);	2693	ev_loop (EV_A_ 0);
2181		2694
2182	It assumes the variable C<loop> of type C<struct ev_loop *> is in scope,	2695	It assumes the variable C<loop> of type C<struct ev_loop *> is in scope,
2183	which is often provided by the following macro.	2696	which is often provided by the following macro.
2184		2697
2185	=item C<EV_P>, C<EV_P_>	2698	=item C<EV_P>, C<EV_P_>
2186		2699
2187	This provides the loop I<parameter> for functions, if one is required ("ev	2700	This provides the loop I<parameter> for functions, if one is required ("ev
2188	loop parameter"). The C<EV_P> form is used when this is the sole parameter,	2701	loop parameter"). The C<EV_P> form is used when this is the sole parameter,
2189	C<EV_P_> is used when other parameters are following. Example:	2702	C<EV_P_> is used when other parameters are following. Example:
2190		2703
2191	// this is how ev_unref is being declared	2704	// this is how ev_unref is being declared
2192	static void ev_unref (EV_P);	2705	static void ev_unref (EV_P);
2193		2706
2194	// this is how you can declare your typical callback	2707	// this is how you can declare your typical callback
2195	static void cb (EV_P_ ev_timer *w, int revents)	2708	static void cb (EV_P_ ev_timer *w, int revents)
2196		2709
2197	It declares a parameter C<loop> of type C<struct ev_loop *>, quite	2710	It declares a parameter C<loop> of type C<struct ev_loop *>, quite
2198	suitable for use with C<EV_A>.	2711	suitable for use with C<EV_A>.
2199		2712
2200	=item C<EV_DEFAULT>, C<EV_DEFAULT_>	2713	=item C<EV_DEFAULT>, C<EV_DEFAULT_>
2201		2714
2202	Similar to the other two macros, this gives you the value of the default	2715	Similar to the other two macros, this gives you the value of the default
2203	loop, if multiple loops are supported ("ev loop default").	2716	loop, if multiple loops are supported ("ev loop default").
		2717
		2718	=item C<EV_DEFAULT_UC>, C<EV_DEFAULT_UC_>
		2719
		2720	Usage identical to C<EV_DEFAULT> and C<EV_DEFAULT_>, but requires that the
		2721	default loop has been initialised (C<UC> == unchecked). Their behaviour
		2722	is undefined when the default loop has not been initialised by a previous
		2723	execution of C<EV_DEFAULT>, C<EV_DEFAULT_> or C<ev_default_init (...)>.
		2724
		2725	It is often prudent to use C<EV_DEFAULT> when initialising the first
		2726	watcher in a function but use C<EV_DEFAULT_UC> afterwards.
2204		2727
2205	=back	2728	=back
2206		2729
2207	Example: Declare and initialise a check watcher, utilising the above	2730	Example: Declare and initialise a check watcher, utilising the above
2208	macros so it will work regardless of whether multiple loops are supported	2731	macros so it will work regardless of whether multiple loops are supported
2209	or not.	2732	or not.
2210		2733
2211	static void	2734	static void
2212	check_cb (EV_P_ ev_timer *w, int revents)	2735	check_cb (EV_P_ ev_timer *w, int revents)
2213	{	2736	{
2214	ev_check_stop (EV_A_ w);	2737	ev_check_stop (EV_A_ w);
2215	}	2738	}
2216		2739
2217	ev_check check;	2740	ev_check check;
2218	ev_check_init (&check, check_cb);	2741	ev_check_init (&check, check_cb);
2219	ev_check_start (EV_DEFAULT_ &check);	2742	ev_check_start (EV_DEFAULT_ &check);
2220	ev_loop (EV_DEFAULT_ 0);	2743	ev_loop (EV_DEFAULT_ 0);
2221		2744
2222	=head1 EMBEDDING	2745	=head1 EMBEDDING
2223		2746
2224	Libev can (and often is) directly embedded into host	2747	Libev can (and often is) directly embedded into host
2225	applications. Examples of applications that embed it include the Deliantra	2748	applications. Examples of applications that embed it include the Deliantra
…		…
2232	libev somewhere in your source tree).	2755	libev somewhere in your source tree).
2233		2756
2234	=head2 FILESETS	2757	=head2 FILESETS
2235		2758
2236	Depending on what features you need you need to include one or more sets of files	2759	Depending on what features you need you need to include one or more sets of files
2237	in your app.	2760	in your application.
2238		2761
2239	=head3 CORE EVENT LOOP	2762	=head3 CORE EVENT LOOP
2240		2763
2241	To include only the libev core (all the C<ev_*> functions), with manual	2764	To include only the libev core (all the C<ev_*> functions), with manual
2242	configuration (no autoconf):	2765	configuration (no autoconf):
2243		2766
2244	#define EV_STANDALONE 1	2767	#define EV_STANDALONE 1
2245	#include "ev.c"	2768	#include "ev.c"
2246		2769
2247	This will automatically include F<ev.h>, too, and should be done in a	2770	This will automatically include F<ev.h>, too, and should be done in a
2248	single C source file only to provide the function implementations. To use	2771	single C source file only to provide the function implementations. To use
2249	it, do the same for F<ev.h> in all files wishing to use this API (best	2772	it, do the same for F<ev.h> in all files wishing to use this API (best
2250	done by writing a wrapper around F<ev.h> that you can include instead and	2773	done by writing a wrapper around F<ev.h> that you can include instead and
2251	where you can put other configuration options):	2774	where you can put other configuration options):
2252		2775
2253	#define EV_STANDALONE 1	2776	#define EV_STANDALONE 1
2254	#include "ev.h"	2777	#include "ev.h"
2255		2778
2256	Both header files and implementation files can be compiled with a C++	2779	Both header files and implementation files can be compiled with a C++
2257	compiler (at least, thats a stated goal, and breakage will be treated	2780	compiler (at least, thats a stated goal, and breakage will be treated
2258	as a bug).	2781	as a bug).
2259		2782
2260	You need the following files in your source tree, or in a directory	2783	You need the following files in your source tree, or in a directory
2261	in your include path (e.g. in libev/ when using -Ilibev):	2784	in your include path (e.g. in libev/ when using -Ilibev):
2262		2785
2263	ev.h	2786	ev.h
2264	ev.c	2787	ev.c
2265	ev_vars.h	2788	ev_vars.h
2266	ev_wrap.h	2789	ev_wrap.h
2267		2790
2268	ev_win32.c required on win32 platforms only	2791	ev_win32.c required on win32 platforms only
2269		2792
2270	ev_select.c only when select backend is enabled (which is enabled by default)	2793	ev_select.c only when select backend is enabled (which is enabled by default)
2271	ev_poll.c only when poll backend is enabled (disabled by default)	2794	ev_poll.c only when poll backend is enabled (disabled by default)
2272	ev_epoll.c only when the epoll backend is enabled (disabled by default)	2795	ev_epoll.c only when the epoll backend is enabled (disabled by default)
2273	ev_kqueue.c only when the kqueue backend is enabled (disabled by default)	2796	ev_kqueue.c only when the kqueue backend is enabled (disabled by default)
2274	ev_port.c only when the solaris port backend is enabled (disabled by default)	2797	ev_port.c only when the solaris port backend is enabled (disabled by default)
2275		2798
2276	F<ev.c> includes the backend files directly when enabled, so you only need	2799	F<ev.c> includes the backend files directly when enabled, so you only need
2277	to compile this single file.	2800	to compile this single file.
2278		2801
2279	=head3 LIBEVENT COMPATIBILITY API	2802	=head3 LIBEVENT COMPATIBILITY API
2280		2803
2281	To include the libevent compatibility API, also include:	2804	To include the libevent compatibility API, also include:
2282		2805
2283	#include "event.c"	2806	#include "event.c"
2284		2807
2285	in the file including F<ev.c>, and:	2808	in the file including F<ev.c>, and:
2286		2809
2287	#include "event.h"	2810	#include "event.h"
2288		2811
2289	in the files that want to use the libevent API. This also includes F<ev.h>.	2812	in the files that want to use the libevent API. This also includes F<ev.h>.
2290		2813
2291	You need the following additional files for this:	2814	You need the following additional files for this:
2292		2815
2293	event.h	2816	event.h
2294	event.c	2817	event.c
2295		2818
2296	=head3 AUTOCONF SUPPORT	2819	=head3 AUTOCONF SUPPORT
2297		2820
2298	Instead of using C<EV_STANDALONE=1> and providing your config in	2821	Instead of using C<EV_STANDALONE=1> and providing your configuration in
2299	whatever way you want, you can also C<m4_include([libev.m4])> in your	2822	whatever way you want, you can also C<m4_include([libev.m4])> in your
2300	F<configure.ac> and leave C<EV_STANDALONE> undefined. F<ev.c> will then	2823	F<configure.ac> and leave C<EV_STANDALONE> undefined. F<ev.c> will then
2301	include F<config.h> and configure itself accordingly.	2824	include F<config.h> and configure itself accordingly.
2302		2825
2303	For this of course you need the m4 file:	2826	For this of course you need the m4 file:
2304		2827
2305	libev.m4	2828	libev.m4
2306		2829
2307	=head2 PREPROCESSOR SYMBOLS/MACROS	2830	=head2 PREPROCESSOR SYMBOLS/MACROS
2308		2831
2309	Libev can be configured via a variety of preprocessor symbols you have to define	2832	Libev can be configured via a variety of preprocessor symbols you have to
2310	before including any of its files. The default is not to build for multiplicity	2833	define before including any of its files. The default in the absence of
2311	and only include the select backend.	2834	autoconf is noted for every option.
2312		2835
2313	=over 4	2836	=over 4
2314		2837
2315	=item EV_STANDALONE	2838	=item EV_STANDALONE
2316		2839
…		…
2321	F<event.h> that are not directly supported by the libev core alone.	2844	F<event.h> that are not directly supported by the libev core alone.
2322		2845
2323	=item EV_USE_MONOTONIC	2846	=item EV_USE_MONOTONIC
2324		2847
2325	If defined to be C<1>, libev will try to detect the availability of the	2848	If defined to be C<1>, libev will try to detect the availability of the
2326	monotonic clock option at both compiletime and runtime. Otherwise no use	2849	monotonic clock option at both compile time and runtime. Otherwise no use
2327	of the monotonic clock option will be attempted. If you enable this, you	2850	of the monotonic clock option will be attempted. If you enable this, you
2328	usually have to link against librt or something similar. Enabling it when	2851	usually have to link against librt or something similar. Enabling it when
2329	the functionality isn't available is safe, though, although you have	2852	the functionality isn't available is safe, though, although you have
2330	to make sure you link against any libraries where the C<clock_gettime>	2853	to make sure you link against any libraries where the C<clock_gettime>
2331	function is hiding in (often F<-lrt>).	2854	function is hiding in (often F<-lrt>).
2332		2855
2333	=item EV_USE_REALTIME	2856	=item EV_USE_REALTIME
2334		2857
2335	If defined to be C<1>, libev will try to detect the availability of the	2858	If defined to be C<1>, libev will try to detect the availability of the
2336	realtime clock option at compiletime (and assume its availability at	2859	real-time clock option at compile time (and assume its availability at
2337	runtime if successful). Otherwise no use of the realtime clock option will	2860	runtime if successful). Otherwise no use of the real-time clock option will
2338	be attempted. This effectively replaces C<gettimeofday> by C<clock_get	2861	be attempted. This effectively replaces C<gettimeofday> by C<clock_get
2339	(CLOCK_REALTIME, ...)> and will not normally affect correctness. See the	2862	(CLOCK_REALTIME, ...)> and will not normally affect correctness. See the
2340	note about libraries in the description of C<EV_USE_MONOTONIC>, though.	2863	note about libraries in the description of C<EV_USE_MONOTONIC>, though.
2341		2864
2342	=item EV_USE_NANOSLEEP	2865	=item EV_USE_NANOSLEEP
2343		2866
2344	If defined to be C<1>, libev will assume that C<nanosleep ()> is available	2867	If defined to be C<1>, libev will assume that C<nanosleep ()> is available
2345	and will use it for delays. Otherwise it will use C<select ()>.	2868	and will use it for delays. Otherwise it will use C<select ()>.
2346		2869
		2870	=item EV_USE_EVENTFD
		2871
		2872	If defined to be C<1>, then libev will assume that C<eventfd ()> is
		2873	available and will probe for kernel support at runtime. This will improve
		2874	C<ev_signal> and C<ev_async> performance and reduce resource consumption.
		2875	If undefined, it will be enabled if the headers indicate GNU/Linux + Glibc
		2876	2.7 or newer, otherwise disabled.
		2877
2347	=item EV_USE_SELECT	2878	=item EV_USE_SELECT
2348		2879
2349	If undefined or defined to be C<1>, libev will compile in support for the	2880	If undefined or defined to be C<1>, libev will compile in support for the
2350	C<select>(2) backend. No attempt at autodetection will be done: if no	2881	C<select>(2) backend. No attempt at auto-detection will be done: if no
2351	other method takes over, select will be it. Otherwise the select backend	2882	other method takes over, select will be it. Otherwise the select backend
2352	will not be compiled in.	2883	will not be compiled in.
2353		2884
2354	=item EV_SELECT_USE_FD_SET	2885	=item EV_SELECT_USE_FD_SET
2355		2886
2356	If defined to C<1>, then the select backend will use the system C<fd_set>	2887	If defined to C<1>, then the select backend will use the system C<fd_set>
2357	structure. This is useful if libev doesn't compile due to a missing	2888	structure. This is useful if libev doesn't compile due to a missing
2358	C<NFDBITS> or C<fd_mask> definition or it misguesses the bitset layout on	2889	C<NFDBITS> or C<fd_mask> definition or it mis-guesses the bitset layout on
2359	exotic systems. This usually limits the range of file descriptors to some	2890	exotic systems. This usually limits the range of file descriptors to some
2360	low limit such as 1024 or might have other limitations (winsocket only	2891	low limit such as 1024 or might have other limitations (winsocket only
2361	allows 64 sockets). The C<FD_SETSIZE> macro, set before compilation, might	2892	allows 64 sockets). The C<FD_SETSIZE> macro, set before compilation, might
2362	influence the size of the C<fd_set> used.	2893	influence the size of the C<fd_set> used.
2363		2894
…		…
2369	be used is the winsock select). This means that it will call	2900	be used is the winsock select). This means that it will call
2370	C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,	2901	C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,
2371	it is assumed that all these functions actually work on fds, even	2902	it is assumed that all these functions actually work on fds, even
2372	on win32. Should not be defined on non-win32 platforms.	2903	on win32. Should not be defined on non-win32 platforms.
2373		2904
		2905	=item EV_FD_TO_WIN32_HANDLE
		2906
		2907	If C<EV_SELECT_IS_WINSOCKET> is enabled, then libev needs a way to map
		2908	file descriptors to socket handles. When not defining this symbol (the
		2909	default), then libev will call C<_get_osfhandle>, which is usually
		2910	correct. In some cases, programs use their own file descriptor management,
		2911	in which case they can provide this function to map fds to socket handles.
		2912
2374	=item EV_USE_POLL	2913	=item EV_USE_POLL
2375		2914
2376	If defined to be C<1>, libev will compile in support for the C<poll>(2)	2915	If defined to be C<1>, libev will compile in support for the C<poll>(2)
2377	backend. Otherwise it will be enabled on non-win32 platforms. It	2916	backend. Otherwise it will be enabled on non-win32 platforms. It
2378	takes precedence over select.	2917	takes precedence over select.
2379		2918
2380	=item EV_USE_EPOLL	2919	=item EV_USE_EPOLL
2381		2920
2382	If defined to be C<1>, libev will compile in support for the Linux	2921	If defined to be C<1>, libev will compile in support for the Linux
2383	C<epoll>(7) backend. Its availability will be detected at runtime,	2922	C<epoll>(7) backend. Its availability will be detected at runtime,
2384	otherwise another method will be used as fallback. This is the	2923	otherwise another method will be used as fallback. This is the preferred
2385	preferred backend for GNU/Linux systems.	2924	backend for GNU/Linux systems. If undefined, it will be enabled if the
		2925	headers indicate GNU/Linux + Glibc 2.4 or newer, otherwise disabled.
2386		2926
2387	=item EV_USE_KQUEUE	2927	=item EV_USE_KQUEUE
2388		2928
2389	If defined to be C<1>, libev will compile in support for the BSD style	2929	If defined to be C<1>, libev will compile in support for the BSD style
2390	C<kqueue>(2) backend. Its actual availability will be detected at runtime,	2930	C<kqueue>(2) backend. Its actual availability will be detected at runtime,
…		…
2403	otherwise another method will be used as fallback. This is the preferred	2943	otherwise another method will be used as fallback. This is the preferred
2404	backend for Solaris 10 systems.	2944	backend for Solaris 10 systems.
2405		2945
2406	=item EV_USE_DEVPOLL	2946	=item EV_USE_DEVPOLL
2407		2947
2408	reserved for future expansion, works like the USE symbols above.	2948	Reserved for future expansion, works like the USE symbols above.
2409		2949
2410	=item EV_USE_INOTIFY	2950	=item EV_USE_INOTIFY
2411		2951
2412	If defined to be C<1>, libev will compile in support for the Linux inotify	2952	If defined to be C<1>, libev will compile in support for the Linux inotify
2413	interface to speed up C<ev_stat> watchers. Its actual availability will	2953	interface to speed up C<ev_stat> watchers. Its actual availability will
2414	be detected at runtime.	2954	be detected at runtime. If undefined, it will be enabled if the headers
		2955	indicate GNU/Linux + Glibc 2.4 or newer, otherwise disabled.
		2956
		2957	=item EV_ATOMIC_T
		2958
		2959	Libev requires an integer type (suitable for storing C<0> or C<1>) whose
		2960	access is atomic with respect to other threads or signal contexts. No such
		2961	type is easily found in the C language, so you can provide your own type
		2962	that you know is safe for your purposes. It is used both for signal handler "locking"
		2963	as well as for signal and thread safety in C<ev_async> watchers.
		2964
		2965	In the absence of this define, libev will use C<sig_atomic_t volatile>
		2966	(from F<signal.h>), which is usually good enough on most platforms.
2415		2967
2416	=item EV_H	2968	=item EV_H
2417		2969
2418	The name of the F<ev.h> header file used to include it. The default if	2970	The name of the F<ev.h> header file used to include it. The default if
2419	undefined is C<< <ev.h> >> in F<event.h> and C<"ev.h"> in F<ev.c>. This	2971	undefined is C<"ev.h"> in F<event.h>, F<ev.c> and F<ev++.h>. This can be
2420	can be used to virtually rename the F<ev.h> header file in case of conflicts.	2972	used to virtually rename the F<ev.h> header file in case of conflicts.
2421		2973
2422	=item EV_CONFIG_H	2974	=item EV_CONFIG_H
2423		2975
2424	If C<EV_STANDALONE> isn't C<1>, this variable can be used to override	2976	If C<EV_STANDALONE> isn't C<1>, this variable can be used to override
2425	F<ev.c>'s idea of where to find the F<config.h> file, similarly to	2977	F<ev.c>'s idea of where to find the F<config.h> file, similarly to
2426	C<EV_H>, above.	2978	C<EV_H>, above.
2427		2979
2428	=item EV_EVENT_H	2980	=item EV_EVENT_H
2429		2981
2430	Similarly to C<EV_H>, this macro can be used to override F<event.c>'s idea	2982	Similarly to C<EV_H>, this macro can be used to override F<event.c>'s idea
2431	of how the F<event.h> header can be found.	2983	of how the F<event.h> header can be found, the default is C<"event.h">.
2432		2984
2433	=item EV_PROTOTYPES	2985	=item EV_PROTOTYPES
2434		2986
2435	If defined to be C<0>, then F<ev.h> will not define any function	2987	If defined to be C<0>, then F<ev.h> will not define any function
2436	prototypes, but still define all the structs and other symbols. This is	2988	prototypes, but still define all the structs and other symbols. This is
…		…
2457	When doing priority-based operations, libev usually has to linearly search	3009	When doing priority-based operations, libev usually has to linearly search
2458	all the priorities, so having many of them (hundreds) uses a lot of space	3010	all the priorities, so having many of them (hundreds) uses a lot of space
2459	and time, so using the defaults of five priorities (-2 .. +2) is usually	3011	and time, so using the defaults of five priorities (-2 .. +2) is usually
2460	fine.	3012	fine.
2461		3013
2462	If your embedding app does not need any priorities, defining these both to	3014	If your embedding application does not need any priorities, defining these both to
2463	C<0> will save some memory and cpu.	3015	C<0> will save some memory and CPU.
2464		3016
2465	=item EV_PERIODIC_ENABLE	3017	=item EV_PERIODIC_ENABLE
2466		3018
2467	If undefined or defined to be C<1>, then periodic timers are supported. If	3019	If undefined or defined to be C<1>, then periodic timers are supported. If
2468	defined to be C<0>, then they are not. Disabling them saves a few kB of	3020	defined to be C<0>, then they are not. Disabling them saves a few kB of
…		…
2487	=item EV_FORK_ENABLE	3039	=item EV_FORK_ENABLE
2488		3040
2489	If undefined or defined to be C<1>, then fork watchers are supported. If	3041	If undefined or defined to be C<1>, then fork watchers are supported. If
2490	defined to be C<0>, then they are not.	3042	defined to be C<0>, then they are not.
2491		3043
		3044	=item EV_ASYNC_ENABLE
		3045
		3046	If undefined or defined to be C<1>, then async watchers are supported. If
		3047	defined to be C<0>, then they are not.
		3048
2492	=item EV_MINIMAL	3049	=item EV_MINIMAL
2493		3050
2494	If you need to shave off some kilobytes of code at the expense of some	3051	If you need to shave off some kilobytes of code at the expense of some
2495	speed, define this symbol to C<1>. Currently only used for gcc to override	3052	speed, define this symbol to C<1>. Currently this is used to override some
2496	some inlining decisions, saves roughly 30% codesize of amd64.	3053	inlining decisions, saves roughly 30% code size on amd64. It also selects a
		3054	much smaller 2-heap for timer management over the default 4-heap.
2497		3055
2498	=item EV_PID_HASHSIZE	3056	=item EV_PID_HASHSIZE
2499		3057
2500	C<ev_child> watchers use a small hash table to distribute workload by	3058	C<ev_child> watchers use a small hash table to distribute workload by
2501	pid. The default size is C<16> (or C<1> with C<EV_MINIMAL>), usually more	3059	pid. The default size is C<16> (or C<1> with C<EV_MINIMAL>), usually more
2502	than enough. If you need to manage thousands of children you might want to	3060	than enough. If you need to manage thousands of children you might want to
2503	increase this value (I<must> be a power of two).	3061	increase this value (I<must> be a power of two).
2504		3062
2505	=item EV_INOTIFY_HASHSIZE	3063	=item EV_INOTIFY_HASHSIZE
2506		3064
2507	C<ev_staz> watchers use a small hash table to distribute workload by	3065	C<ev_stat> watchers use a small hash table to distribute workload by
2508	inotify watch id. The default size is C<16> (or C<1> with C<EV_MINIMAL>),	3066	inotify watch id. The default size is C<16> (or C<1> with C<EV_MINIMAL>),
2509	usually more than enough. If you need to manage thousands of C<ev_stat>	3067	usually more than enough. If you need to manage thousands of C<ev_stat>
2510	watchers you might want to increase this value (I<must> be a power of	3068	watchers you might want to increase this value (I<must> be a power of
2511	two).	3069	two).
2512		3070
		3071	=item EV_USE_4HEAP
		3072
		3073	Heaps are not very cache-efficient. To improve the cache-efficiency of the
		3074	timer and periodics heap, libev uses a 4-heap when this symbol is defined
		3075	to C<1>. The 4-heap uses more complicated (longer) code but has
		3076	noticeably faster performance with many (thousands) of watchers.
		3077
		3078	The default is C<1> unless C<EV_MINIMAL> is set in which case it is C<0>
		3079	(disabled).
		3080
		3081	=item EV_HEAP_CACHE_AT
		3082
		3083	Heaps are not very cache-efficient. To improve the cache-efficiency of the
		3084	timer and periodics heap, libev can cache the timestamp (I<at>) within
		3085	the heap structure (selected by defining C<EV_HEAP_CACHE_AT> to C<1>),
		3086	which uses 8-12 bytes more per watcher and a few hundred bytes more code,
		3087	but avoids random read accesses on heap changes. This improves performance
		3088	noticeably with with many (hundreds) of watchers.
		3089
		3090	The default is C<1> unless C<EV_MINIMAL> is set in which case it is C<0>
		3091	(disabled).
		3092
		3093	=item EV_VERIFY
		3094
		3095	Controls how much internal verification (see C<ev_loop_verify ()>) will
		3096	be done: If set to C<0>, no internal verification code will be compiled
		3097	in. If set to C<1>, then verification code will be compiled in, but not
		3098	called. If set to C<2>, then the internal verification code will be
		3099	called once per loop, which can slow down libev. If set to C<3>, then the
		3100	verification code will be called very frequently, which will slow down
		3101	libev considerably.
		3102
		3103	The default is C<1>, unless C<EV_MINIMAL> is set, in which case it will be
		3104	C<0.>
		3105
2513	=item EV_COMMON	3106	=item EV_COMMON
2514		3107
2515	By default, all watchers have a C<void *data> member. By redefining	3108	By default, all watchers have a C<void *data> member. By redefining
2516	this macro to a something else you can include more and other types of	3109	this macro to a something else you can include more and other types of
2517	members. You have to define it each time you include one of the files,	3110	members. You have to define it each time you include one of the files,
2518	though, and it must be identical each time.	3111	though, and it must be identical each time.
2519		3112
2520	For example, the perl EV module uses something like this:	3113	For example, the perl EV module uses something like this:
2521		3114
2522	#define EV_COMMON \	3115	#define EV_COMMON \
2523	SV *self; /* contains this struct */ \	3116	SV *self; /* contains this struct */ \
2524	SV *cb_sv, *fh /* note no trailing ";" */	3117	SV *cb_sv, *fh /* note no trailing ";" */
2525		3118
2526	=item EV_CB_DECLARE (type)	3119	=item EV_CB_DECLARE (type)
2527		3120
2528	=item EV_CB_INVOKE (watcher, revents)	3121	=item EV_CB_INVOKE (watcher, revents)
2529		3122
…		…
2536	avoid the C<struct ev_loop *> as first argument in all cases, or to use	3129	avoid the C<struct ev_loop *> as first argument in all cases, or to use
2537	method calls instead of plain function calls in C++.	3130	method calls instead of plain function calls in C++.
2538		3131
2539	=head2 EXPORTED API SYMBOLS	3132	=head2 EXPORTED API SYMBOLS
2540		3133
2541	If you need to re-export the API (e.g. via a dll) and you need a list of	3134	If you need to re-export the API (e.g. via a DLL) and you need a list of
2542	exported symbols, you can use the provided F<Symbol.*> files which list	3135	exported symbols, you can use the provided F<Symbol.*> files which list
2543	all public symbols, one per line:	3136	all public symbols, one per line:
2544		3137
2545	Symbols.ev for libev proper	3138	Symbols.ev for libev proper
2546	Symbols.event for the libevent emulation	3139	Symbols.event for the libevent emulation
2547		3140
2548	This can also be used to rename all public symbols to avoid clashes with	3141	This can also be used to rename all public symbols to avoid clashes with
2549	multiple versions of libev linked together (which is obviously bad in	3142	multiple versions of libev linked together (which is obviously bad in
2550	itself, but sometimes it is inconvinient to avoid this).	3143	itself, but sometimes it is inconvenient to avoid this).
2551		3144
2552	A sed command like this will create wrapper C<#define>'s that you need to	3145	A sed command like this will create wrapper C<#define>'s that you need to
2553	include before including F<ev.h>:	3146	include before including F<ev.h>:
2554		3147
2555	<Symbols.ev sed -e "s/.*/#define & myprefix_&/" >wrap.h	3148	<Symbols.ev sed -e "s/.*/#define & myprefix_&/" >wrap.h
…		…
2572	file.	3165	file.
2573		3166
2574	The usage in rxvt-unicode is simpler. It has a F<ev_cpp.h> header file	3167	The usage in rxvt-unicode is simpler. It has a F<ev_cpp.h> header file
2575	that everybody includes and which overrides some configure choices:	3168	that everybody includes and which overrides some configure choices:
2576		3169
2577	#define EV_MINIMAL 1	3170	#define EV_MINIMAL 1
2578	#define EV_USE_POLL 0	3171	#define EV_USE_POLL 0
2579	#define EV_MULTIPLICITY 0	3172	#define EV_MULTIPLICITY 0
2580	#define EV_PERIODIC_ENABLE 0	3173	#define EV_PERIODIC_ENABLE 0
2581	#define EV_STAT_ENABLE 0	3174	#define EV_STAT_ENABLE 0
2582	#define EV_FORK_ENABLE 0	3175	#define EV_FORK_ENABLE 0
2583	#define EV_CONFIG_H <config.h>	3176	#define EV_CONFIG_H <config.h>
2584	#define EV_MINPRI 0	3177	#define EV_MINPRI 0
2585	#define EV_MAXPRI 0	3178	#define EV_MAXPRI 0
2586		3179
2587	#include "ev++.h"	3180	#include "ev++.h"
2588		3181
2589	And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:	3182	And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:
2590		3183
2591	#include "ev_cpp.h"	3184	#include "ev_cpp.h"
2592	#include "ev.c"	3185	#include "ev.c"
		3186
		3187
		3188	=head1 THREADS AND COROUTINES
		3189
		3190	=head2 THREADS
		3191
		3192	Libev itself is completely thread-safe, but it uses no locking. This
		3193	means that you can use as many loops as you want in parallel, as long as
		3194	only one thread ever calls into one libev function with the same loop
		3195	parameter.
		3196
		3197	Or put differently: calls with different loop parameters can be done in
		3198	parallel from multiple threads, calls with the same loop parameter must be
		3199	done serially (but can be done from different threads, as long as only one
		3200	thread ever is inside a call at any point in time, e.g. by using a mutex
		3201	per loop).
		3202
		3203	If you want to know which design (one loop, locking, or multiple loops
		3204	without or something else still) is best for your problem, then I cannot
		3205	help you. I can give some generic advice however:
		3206
		3207	=over 4
		3208
		3209	=item * most applications have a main thread: use the default libev loop
		3210	in that thread, or create a separate thread running only the default loop.
		3211
		3212	This helps integrating other libraries or software modules that use libev
		3213	themselves and don't care/know about threading.
		3214
		3215	=item * one loop per thread is usually a good model.
		3216
		3217	Doing this is almost never wrong, sometimes a better-performance model
		3218	exists, but it is always a good start.
		3219
		3220	=item * other models exist, such as the leader/follower pattern, where one
		3221	loop is handed through multiple threads in a kind of round-robin fashion.
		3222
		3223	Choosing a model is hard - look around, learn, know that usually you can do
		3224	better than you currently do :-)
		3225
		3226	=item * often you need to talk to some other thread which blocks in the
		3227	event loop - C<ev_async> watchers can be used to wake them up from other
		3228	threads safely (or from signal contexts...).
		3229
		3230	=back
		3231
		3232	=head2 COROUTINES
		3233
		3234	Libev is much more accommodating to coroutines ("cooperative threads"):
		3235	libev fully supports nesting calls to it's functions from different
		3236	coroutines (e.g. you can call C<ev_loop> on the same loop from two
		3237	different coroutines and switch freely between both coroutines running the
		3238	loop, as long as you don't confuse yourself). The only exception is that
		3239	you must not do this from C<ev_periodic> reschedule callbacks.
		3240
		3241	Care has been invested into making sure that libev does not keep local
		3242	state inside C<ev_loop>, and other calls do not usually allow coroutine
		3243	switches.
2593		3244
2594		3245
2595	=head1 COMPLEXITIES	3246	=head1 COMPLEXITIES
2596		3247
2597	In this section the complexities of (many of) the algorithms used inside	3248	In this section the complexities of (many of) the algorithms used inside
…		…
2608		3259
2609	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)	3260	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)
2610		3261
2611	This means that, when you have a watcher that triggers in one hour and	3262	This means that, when you have a watcher that triggers in one hour and
2612	there are 100 watchers that would trigger before that then inserting will	3263	there are 100 watchers that would trigger before that then inserting will
2613	have to skip those 100 watchers.	3264	have to skip roughly seven (C<ld 100>) of these watchers.
2614		3265
2615	=item Changing timer/periodic watchers (by autorepeat, again): O(log skipped_other_timers)	3266	=item Changing timer/periodic watchers (by autorepeat or calling again): O(log skipped_other_timers)
2616		3267
2617	That means that for changing a timer costs less than removing/adding them	3268	That means that changing a timer costs less than removing/adding them
2618	as only the relative motion in the event queue has to be paid for.	3269	as only the relative motion in the event queue has to be paid for.
2619		3270
2620	=item Starting io/check/prepare/idle/signal/child watchers: O(1)	3271	=item Starting io/check/prepare/idle/signal/child/fork/async watchers: O(1)
2621		3272
2622	These just add the watcher into an array or at the head of a list.	3273	These just add the watcher into an array or at the head of a list.
		3274
2623	=item Stopping check/prepare/idle watchers: O(1)	3275	=item Stopping check/prepare/idle/fork/async watchers: O(1)
2624		3276
2625	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))	3277	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))
2626		3278
2627	These watchers are stored in lists then need to be walked to find the	3279	These watchers are stored in lists then need to be walked to find the
2628	correct watcher to remove. The lists are usually short (you don't usually	3280	correct watcher to remove. The lists are usually short (you don't usually
2629	have many watchers waiting for the same fd or signal).	3281	have many watchers waiting for the same fd or signal).
2630		3282
2631	=item Finding the next timer per loop iteration: O(1)	3283	=item Finding the next timer in each loop iteration: O(1)
		3284
		3285	By virtue of using a binary or 4-heap, the next timer is always found at a
		3286	fixed position in the storage array.
2632		3287
2633	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)	3288	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)
2634		3289
2635	A change means an I/O watcher gets started or stopped, which requires	3290	A change means an I/O watcher gets started or stopped, which requires
2636	libev to recalculate its status (and possibly tell the kernel).	3291	libev to recalculate its status (and possibly tell the kernel, depending
		3292	on backend and whether C<ev_io_set> was used).
2637		3293
2638	=item Activating one watcher: O(1)	3294	=item Activating one watcher (putting it into the pending state): O(1)
2639		3295
2640	=item Priority handling: O(number_of_priorities)	3296	=item Priority handling: O(number_of_priorities)
2641		3297
2642	Priorities are implemented by allocating some space for each	3298	Priorities are implemented by allocating some space for each
2643	priority. When doing priority-based operations, libev usually has to	3299	priority. When doing priority-based operations, libev usually has to
2644	linearly search all the priorities.	3300	linearly search all the priorities, but starting/stopping and activating
		3301	watchers becomes O(1) w.r.t. priority handling.
		3302
		3303	=item Sending an ev_async: O(1)
		3304
		3305	=item Processing ev_async_send: O(number_of_async_watchers)
		3306
		3307	=item Processing signals: O(max_signal_number)
		3308
		3309	Sending involves a system call I<iff> there were no other C<ev_async_send>
		3310	calls in the current loop iteration. Checking for async and signal events
		3311	involves iterating over all running async watchers or all signal numbers.
2645		3312
2646	=back	3313	=back
2647		3314
2648		3315
		3316	=head1 WIN32 PLATFORM LIMITATIONS AND WORKAROUNDS
		3317
		3318	Win32 doesn't support any of the standards (e.g. POSIX) that libev
		3319	requires, and its I/O model is fundamentally incompatible with the POSIX
		3320	model. Libev still offers limited functionality on this platform in
		3321	the form of the C<EVBACKEND_SELECT> backend, and only supports socket
		3322	descriptors. This only applies when using Win32 natively, not when using
		3323	e.g. cygwin.
		3324
		3325	Lifting these limitations would basically require the full
		3326	re-implementation of the I/O system. If you are into these kinds of
		3327	things, then note that glib does exactly that for you in a very portable
		3328	way (note also that glib is the slowest event library known to man).
		3329
		3330	There is no supported compilation method available on windows except
		3331	embedding it into other applications.
		3332
		3333	Not a libev limitation but worth mentioning: windows apparently doesn't
		3334	accept large writes: instead of resulting in a partial write, windows will
		3335	either accept everything or return C<ENOBUFS> if the buffer is too large,
		3336	so make sure you only write small amounts into your sockets (less than a
		3337	megabyte seems safe, but thsi apparently depends on the amount of memory
		3338	available).
		3339
		3340	Due to the many, low, and arbitrary limits on the win32 platform and
		3341	the abysmal performance of winsockets, using a large number of sockets
		3342	is not recommended (and not reasonable). If your program needs to use
		3343	more than a hundred or so sockets, then likely it needs to use a totally
		3344	different implementation for windows, as libev offers the POSIX readiness
		3345	notification model, which cannot be implemented efficiently on windows
		3346	(Microsoft monopoly games).
		3347
		3348	A typical way to use libev under windows is to embed it (see the embedding
		3349	section for details) and use the following F<evwrap.h> header file instead
		3350	of F<ev.h>:
		3351
		3352	#define EV_STANDALONE /* keeps ev from requiring config.h */
		3353	#define EV_SELECT_IS_WINSOCKET 1 /* configure libev for windows select */
		3354
		3355	#include "ev.h"
		3356
		3357	And compile the following F<evwrap.c> file into your project (make sure
		3358	you do I<not> compile the F<ev.c> or any other embedded soruce files!):
		3359
		3360	#include "evwrap.h"
		3361	#include "ev.c"
		3362
		3363	=over 4
		3364
		3365	=item The winsocket select function
		3366
		3367	The winsocket C<select> function doesn't follow POSIX in that it
		3368	requires socket I<handles> and not socket I<file descriptors> (it is
		3369	also extremely buggy). This makes select very inefficient, and also
		3370	requires a mapping from file descriptors to socket handles (the Microsoft
		3371	C runtime provides the function C<_open_osfhandle> for this). See the
		3372	discussion of the C<EV_SELECT_USE_FD_SET>, C<EV_SELECT_IS_WINSOCKET> and
		3373	C<EV_FD_TO_WIN32_HANDLE> preprocessor symbols for more info.
		3374
		3375	The configuration for a "naked" win32 using the Microsoft runtime
		3376	libraries and raw winsocket select is:
		3377
		3378	#define EV_USE_SELECT 1
		3379	#define EV_SELECT_IS_WINSOCKET 1 /* forces EV_SELECT_USE_FD_SET, too */
		3380
		3381	Note that winsockets handling of fd sets is O(n), so you can easily get a
		3382	complexity in the O(n²) range when using win32.
		3383
		3384	=item Limited number of file descriptors
		3385
		3386	Windows has numerous arbitrary (and low) limits on things.
		3387
		3388	Early versions of winsocket's select only supported waiting for a maximum
		3389	of C<64> handles (probably owning to the fact that all windows kernels
		3390	can only wait for C<64> things at the same time internally; Microsoft
		3391	recommends spawning a chain of threads and wait for 63 handles and the
		3392	previous thread in each. Great).
		3393
		3394	Newer versions support more handles, but you need to define C<FD_SETSIZE>
		3395	to some high number (e.g. C<2048>) before compiling the winsocket select
		3396	call (which might be in libev or elsewhere, for example, perl does its own
		3397	select emulation on windows).
		3398
		3399	Another limit is the number of file descriptors in the Microsoft runtime
		3400	libraries, which by default is C<64> (there must be a hidden I<64> fetish
		3401	or something like this inside Microsoft). You can increase this by calling
		3402	C<_setmaxstdio>, which can increase this limit to C<2048> (another
		3403	arbitrary limit), but is broken in many versions of the Microsoft runtime
		3404	libraries.
		3405
		3406	This might get you to about C<512> or C<2048> sockets (depending on
		3407	windows version and/or the phase of the moon). To get more, you need to
		3408	wrap all I/O functions and provide your own fd management, but the cost of
		3409	calling select (O(n²)) will likely make this unworkable.
		3410
		3411	=back
		3412
		3413
		3414	=head1 PORTABILITY REQUIREMENTS
		3415
		3416	In addition to a working ISO-C implementation, libev relies on a few
		3417	additional extensions:
		3418
		3419	=over 4
		3420
		3421	=item C<void (*)(ev_watcher_type *, int revents)> must have compatible
		3422	calling conventions regardless of C<ev_watcher_type *>.
		3423
		3424	Libev assumes not only that all watcher pointers have the same internal
		3425	structure (guaranteed by POSIX but not by ISO C for example), but it also
		3426	assumes that the same (machine) code can be used to call any watcher
		3427	callback: The watcher callbacks have different type signatures, but libev
		3428	calls them using an C<ev_watcher *> internally.
		3429
		3430	=item C<sig_atomic_t volatile> must be thread-atomic as well
		3431
		3432	The type C<sig_atomic_t volatile> (or whatever is defined as
		3433	C<EV_ATOMIC_T>) must be atomic w.r.t. accesses from different
		3434	threads. This is not part of the specification for C<sig_atomic_t>, but is
		3435	believed to be sufficiently portable.
		3436
		3437	=item C<sigprocmask> must work in a threaded environment
		3438
		3439	Libev uses C<sigprocmask> to temporarily block signals. This is not
		3440	allowed in a threaded program (C<pthread_sigmask> has to be used). Typical
		3441	pthread implementations will either allow C<sigprocmask> in the "main
		3442	thread" or will block signals process-wide, both behaviours would
		3443	be compatible with libev. Interaction between C<sigprocmask> and
		3444	C<pthread_sigmask> could complicate things, however.
		3445
		3446	The most portable way to handle signals is to block signals in all threads
		3447	except the initial one, and run the default loop in the initial thread as
		3448	well.
		3449
		3450	=item C<long> must be large enough for common memory allocation sizes
		3451
		3452	To improve portability and simplify using libev, libev uses C<long>
		3453	internally instead of C<size_t> when allocating its data structures. On
		3454	non-POSIX systems (Microsoft...) this might be unexpectedly low, but
		3455	is still at least 31 bits everywhere, which is enough for hundreds of
		3456	millions of watchers.
		3457
		3458	=item C<double> must hold a time value in seconds with enough accuracy
		3459
		3460	The type C<double> is used to represent timestamps. It is required to
		3461	have at least 51 bits of mantissa (and 9 bits of exponent), which is good
		3462	enough for at least into the year 4000. This requirement is fulfilled by
		3463	implementations implementing IEEE 754 (basically all existing ones).
		3464
		3465	=back
		3466
		3467	If you know of other additional requirements drop me a note.
		3468
		3469
		3470	=head1 COMPILER WARNINGS
		3471
		3472	Depending on your compiler and compiler settings, you might get no or a
		3473	lot of warnings when compiling libev code. Some people are apparently
		3474	scared by this.
		3475
		3476	However, these are unavoidable for many reasons. For one, each compiler
		3477	has different warnings, and each user has different tastes regarding
		3478	warning options. "Warn-free" code therefore cannot be a goal except when
		3479	targeting a specific compiler and compiler-version.
		3480
		3481	Another reason is that some compiler warnings require elaborate
		3482	workarounds, or other changes to the code that make it less clear and less
		3483	maintainable.
		3484
		3485	And of course, some compiler warnings are just plain stupid, or simply
		3486	wrong (because they don't actually warn about the condition their message
		3487	seems to warn about).
		3488
		3489	While libev is written to generate as few warnings as possible,
		3490	"warn-free" code is not a goal, and it is recommended not to build libev
		3491	with any compiler warnings enabled unless you are prepared to cope with
		3492	them (e.g. by ignoring them). Remember that warnings are just that:
		3493	warnings, not errors, or proof of bugs.
		3494
		3495
		3496	=head1 VALGRIND
		3497
		3498	Valgrind has a special section here because it is a popular tool that is
		3499	highly useful, but valgrind reports are very hard to interpret.
		3500
		3501	If you think you found a bug (memory leak, uninitialised data access etc.)
		3502	in libev, then check twice: If valgrind reports something like:
		3503
		3504	==2274== definitely lost: 0 bytes in 0 blocks.
		3505	==2274== possibly lost: 0 bytes in 0 blocks.
		3506	==2274== still reachable: 256 bytes in 1 blocks.
		3507
		3508	Then there is no memory leak. Similarly, under some circumstances,
		3509	valgrind might report kernel bugs as if it were a bug in libev, or it
		3510	might be confused (it is a very good tool, but only a tool).
		3511
		3512	If you are unsure about something, feel free to contact the mailing list
		3513	with the full valgrind report and an explanation on why you think this is
		3514	a bug in libev. However, don't be annoyed when you get a brisk "this is
		3515	no bug" answer and take the chance of learning how to interpret valgrind
		3516	properly.
		3517
		3518	If you need, for some reason, empty reports from valgrind for your project
		3519	I suggest using suppression lists.
		3520
		3521
2649	=head1 AUTHOR	3522	=head1 AUTHOR
2650		3523
2651	Marc Lehmann <libev@schmorp.de>.	3524	Marc Lehmann <libev@schmorp.de>.
2652		3525

Diff Legend

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