ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/libev/ev.pod

Comparing libev/ev.pod (file contents):
Revision 1.122 by root, Thu Jan 31 13:10:56 2008 UTC vs.
Revision 1.198 by root, Thu Oct 23 06:30:48 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	=head2 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_ 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_ 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	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
…		…
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	=head2 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<ev_loop *>) will not have
		109	this argument.
95		110
96	=head2 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
…		…
178	C<ev_embeddable_backends () & ev_supported_backends ()>, likewise for	214	C<ev_embeddable_backends () & ev_supported_backends ()>, likewise for
179	recommended ones.	215	recommended ones.
180		216
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)) [NOT REENTRANT]
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 (;;)
…		…
211	}	250	}
212		251
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)); [NOT REENTRANT]
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.
…		…
237		276
238	=back	277	=back
239		278
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<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
265	The default loop is the only loop that can handle C<ev_signal> and	301	The default loop is the only loop that can handle C<ev_signal> and
266	C<ev_child> watchers, and to do this, it always registers a handler	302	C<ev_child> watchers, and to do this, it always registers a handler
267	for C<SIGCHLD>. If this is a problem for your app you can either	303	for C<SIGCHLD>. If this is a problem for your application you can either
268	create a dynamic loop with C<ev_loop_new> that doesn't do that, or you	304	create a dynamic loop with C<ev_loop_new> that doesn't do that, or you
269	can simply overwrite the C<SIGCHLD> signal handler I<after> calling	305	can simply overwrite the C<SIGCHLD> signal handler I<after> calling
270	C<ev_default_init>.	306	C<ev_default_init>.
271		307
272	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
…		…
281	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
282	thing, believe me).	318	thing, believe me).
283		319
284	=item C<EVFLAG_NOENV>	320	=item C<EVFLAG_NOENV>
285		321
286	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
287	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
288	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will	324	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will
289	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
290	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
291	around bugs.	327	around bugs.
…		…
297	enabling this flag.	333	enabling this flag.
298		334
299	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,
300	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
301	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
302	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
303	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
304	C<pthread_atfork> which is even faster).	340	C<pthread_atfork> which is even faster).
305		341
306	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
307	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
308	flag.	344	flag.
309		345
310	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>
311	environment variable.	347	environment variable.
312		348
313	=item C<EVBACKEND_SELECT> (value 1, portable select backend)	349	=item C<EVBACKEND_SELECT> (value 1, portable select backend)
314		350
315	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
…		…
317	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
318	using this backend. It doesn't scale too well (O(highest_fd)), but its	354	using this backend. It doesn't scale too well (O(highest_fd)), but its
319	usually the fastest backend for a low number of (low-numbered :) fds.	355	usually the fastest backend for a low number of (low-numbered :) fds.
320		356
321	To get good performance out of this backend you need a high amount of	357	To get good performance out of this backend you need a high amount of
322	parallelity (most of the file descriptors should be busy). If you are	358	parallelism (most of the file descriptors should be busy). If you are
323	writing a server, you should C<accept ()> in a loop to accept as many	359	writing a server, you should C<accept ()> in a loop to accept as many
324	connections as possible during one iteration. You might also want to have	360	connections as possible during one iteration. You might also want to have
325	a look at C<ev_set_io_collect_interval ()> to increase the amount of	361	a look at C<ev_set_io_collect_interval ()> to increase the amount of
326	readyness notifications you get per iteration.	362	readiness notifications you get per iteration.
		363
		364	This backend maps C<EV_READ> to the C<readfds> set and C<EV_WRITE> to the
		365	C<writefds> set (and to work around Microsoft Windows bugs, also onto the
		366	C<exceptfds> set on that platform).
327		367
328	=item C<EVBACKEND_POLL> (value 2, poll backend, available everywhere except on windows)	368	=item C<EVBACKEND_POLL> (value 2, poll backend, available everywhere except on windows)
329		369
330	And this is your standard poll(2) backend. It's more complicated	370	And this is your standard poll(2) backend. It's more complicated
331	than select, but handles sparse fds better and has no artificial	371	than select, but handles sparse fds better and has no artificial
332	limit on the number of fds you can use (except it will slow down	372	limit on the number of fds you can use (except it will slow down
333	considerably with a lot of inactive fds). It scales similarly to select,	373	considerably with a lot of inactive fds). It scales similarly to select,
334	i.e. O(total_fds). See the entry for C<EVBACKEND_SELECT>, above, for	374	i.e. O(total_fds). See the entry for C<EVBACKEND_SELECT>, above, for
335	performance tips.	375	performance tips.
336		376
		377	This backend maps C<EV_READ> to C<POLLIN | POLLERR | POLLHUP>, and
		378	C<EV_WRITE> to C<POLLOUT | POLLERR | POLLHUP>.
		379
337	=item C<EVBACKEND_EPOLL> (value 4, Linux)	380	=item C<EVBACKEND_EPOLL> (value 4, Linux)
338		381
339	For few fds, this backend is a bit little slower than poll and select,	382	For few fds, this backend is a bit little slower than poll and select,
340	but it scales phenomenally better. While poll and select usually scale	383	but it scales phenomenally better. While poll and select usually scale
341	like O(total_fds) where n is the total number of fds (or the highest fd),	384	like O(total_fds) where n is the total number of fds (or the highest fd),
342	epoll scales either O(1) or O(active_fds). The epoll design has a number	385	epoll scales either O(1) or O(active_fds). The epoll design has a number
343	of shortcomings, such as silently dropping events in some hard-to-detect	386	of shortcomings, such as silently dropping events in some hard-to-detect
344	cases and rewiring a syscall per fd change, no fork support and bad	387	cases and requiring a system call per fd change, no fork support and bad
345	support for dup.	388	support for dup.
346		389
347	While stopping, setting and starting an I/O watcher in the same iteration	390	While stopping, setting and starting an I/O watcher in the same iteration
348	will result in some caching, there is still a syscall per such incident	391	will result in some caching, there is still a system call per such incident
349	(because the fd could point to a different file description now), so its	392	(because the fd could point to a different file description now), so its
350	best to avoid that. Also, C<dup ()>'ed file descriptors might not work	393	best to avoid that. Also, C<dup ()>'ed file descriptors might not work
351	very well if you register events for both fds.	394	very well if you register events for both fds.
352		395
353	Please note that epoll sometimes generates spurious notifications, so you	396	Please note that epoll sometimes generates spurious notifications, so you
354	need to use non-blocking I/O or other means to avoid blocking when no data	397	need to use non-blocking I/O or other means to avoid blocking when no data
355	(or space) is available.	398	(or space) is available.
356		399
357	Best performance from this backend is achieved by not unregistering all	400	Best performance from this backend is achieved by not unregistering all
358	watchers for a file descriptor until it has been closed, if possible, i.e.	401	watchers for a file descriptor until it has been closed, if possible,
359	keep at least one watcher active per fd at all times.	402	i.e. keep at least one watcher active per fd at all times. Stopping and
		403	starting a watcher (without re-setting it) also usually doesn't cause
		404	extra overhead.
360		405
361	While nominally embeddeble in other event loops, this feature is broken in	406	While nominally embeddable in other event loops, this feature is broken in
362	all kernel versions tested so far.	407	all kernel versions tested so far.
363		408
		409	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
		410	C<EVBACKEND_POLL>.
		411
364	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)	412	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)
365		413
366	Kqueue deserves special mention, as at the time of this writing, it	414	Kqueue deserves special mention, as at the time of this writing, it was
367	was broken on all BSDs except NetBSD (usually it doesn't work reliably	415	broken on all BSDs except NetBSD (usually it doesn't work reliably with
368	with anything but sockets and pipes, except on Darwin, where of course	416	anything but sockets and pipes, except on Darwin, where of course it's
369	it's completely useless). For this reason it's not being "autodetected"	417	completely useless). For this reason it's not being "auto-detected" unless
370	unless you explicitly specify it explicitly in the flags (i.e. using	418	you explicitly specify it in the flags (i.e. using C<EVBACKEND_KQUEUE>) or
371	C<EVBACKEND_KQUEUE>) or libev was compiled on a known-to-be-good (-enough)	419	libev was compiled on a known-to-be-good (-enough) system like NetBSD.
372	system like NetBSD.
373		420
374	You still can embed kqueue into a normal poll or select backend and use it	421	You still can embed kqueue into a normal poll or select backend and use it
375	only for sockets (after having made sure that sockets work with kqueue on	422	only for sockets (after having made sure that sockets work with kqueue on
376	the target platform). See C<ev_embed> watchers for more info.	423	the target platform). See C<ev_embed> watchers for more info.
377		424
378	It scales in the same way as the epoll backend, but the interface to the	425	It scales in the same way as the epoll backend, but the interface to the
379	kernel is more efficient (which says nothing about its actual speed, of	426	kernel is more efficient (which says nothing about its actual speed, of
380	course). While stopping, setting and starting an I/O watcher does never	427	course). While stopping, setting and starting an I/O watcher does never
381	cause an extra syscall as with C<EVBACKEND_EPOLL>, it still adds up to	428	cause an extra system call as with C<EVBACKEND_EPOLL>, it still adds up to
382	two event changes per incident, support for C<fork ()> is very bad and it	429	two event changes per incident. Support for C<fork ()> is very bad and it
383	drops fds silently in similarly hard-to-detect cases.	430	drops fds silently in similarly hard-to-detect cases.
384		431
385	This backend usually performs well under most conditions.	432	This backend usually performs well under most conditions.
386		433
387	While nominally embeddable in other event loops, this doesn't work	434	While nominally embeddable in other event loops, this doesn't work
388	everywhere, so you might need to test for this. And since it is broken	435	everywhere, so you might need to test for this. And since it is broken
389	almost everywhere, you should only use it when you have a lot of sockets	436	almost everywhere, you should only use it when you have a lot of sockets
390	(for which it usually works), by embedding it into another event loop	437	(for which it usually works), by embedding it into another event loop
391	(e.g. C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>) and using it only for	438	(e.g. C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>) and, did I mention it,
392	sockets.	439	using it only for sockets.
		440
		441	This backend maps C<EV_READ> into an C<EVFILT_READ> kevent with
		442	C<NOTE_EOF>, and C<EV_WRITE> into an C<EVFILT_WRITE> kevent with
		443	C<NOTE_EOF>.
393		444
394	=item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8)	445	=item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8)
395		446
396	This is not implemented yet (and might never be, unless you send me an	447	This is not implemented yet (and might never be, unless you send me an
397	implementation). According to reports, C</dev/poll> only supports sockets	448	implementation). According to reports, C</dev/poll> only supports sockets
…		…
401	=item C<EVBACKEND_PORT> (value 32, Solaris 10)	452	=item C<EVBACKEND_PORT> (value 32, Solaris 10)
402		453
403	This uses the Solaris 10 event port mechanism. As with everything on Solaris,	454	This uses the Solaris 10 event port mechanism. As with everything on Solaris,
404	it's really slow, but it still scales very well (O(active_fds)).	455	it's really slow, but it still scales very well (O(active_fds)).
405		456
406	Please note that solaris event ports can deliver a lot of spurious	457	Please note that Solaris event ports can deliver a lot of spurious
407	notifications, so you need to use non-blocking I/O or other means to avoid	458	notifications, so you need to use non-blocking I/O or other means to avoid
408	blocking when no data (or space) is available.	459	blocking when no data (or space) is available.
409		460
410	While this backend scales well, it requires one system call per active	461	While this backend scales well, it requires one system call per active
411	file descriptor per loop iteration. For small and medium numbers of file	462	file descriptor per loop iteration. For small and medium numbers of file
412	descriptors a "slow" C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> backend	463	descriptors a "slow" C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> backend
413	might perform better.	464	might perform better.
414		465
415	On the positive side, ignoring the spurious readyness notifications, this	466	On the positive side, with the exception of the spurious readiness
416	backend actually performed to specification in all tests and is fully	467	notifications, this backend actually performed fully to specification
417	embeddable, which is a rare feat among the OS-specific backends.	468	in all tests and is fully embeddable, which is a rare feat among the
		469	OS-specific backends.
		470
		471	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
		472	C<EVBACKEND_POLL>.
418		473
419	=item C<EVBACKEND_ALL>	474	=item C<EVBACKEND_ALL>
420		475
421	Try all backends (even potentially broken ones that wouldn't be tried	476	Try all backends (even potentially broken ones that wouldn't be tried
422	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as	477	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as
…		…
424		479
425	It is definitely not recommended to use this flag.	480	It is definitely not recommended to use this flag.
426		481
427	=back	482	=back
428		483
429	If one or more of these are ored into the flags value, then only these	484	If one or more of these are or'ed into the flags value, then only these
430	backends will be tried (in the reverse order as listed here). If none are	485	backends will be tried (in the reverse order as listed here). If none are
431	specified, all backends in C<ev_recommended_backends ()> will be tried.	486	specified, all backends in C<ev_recommended_backends ()> will be tried.
432		487
433	The most typical usage is like this:	488	Example: This is the most typical usage.
434		489
435	if (!ev_default_loop (0))	490	if (!ev_default_loop (0))
436	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");	491	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
437		492
438	Restrict libev to the select and poll backends, and do not allow	493	Example: Restrict libev to the select and poll backends, and do not allow
439	environment settings to be taken into account:	494	environment settings to be taken into account:
440		495
441	ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);	496	ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);
442		497
443	Use whatever libev has to offer, but make sure that kqueue is used if	498	Example: Use whatever libev has to offer, but make sure that kqueue is
444	available (warning, breaks stuff, best use only with your own private	499	used if available (warning, breaks stuff, best use only with your own
445	event loop and only if you know the OS supports your types of fds):	500	private event loop and only if you know the OS supports your types of
		501	fds):
446		502
447	ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);	503	ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);
448		504
449	=item struct ev_loop *ev_loop_new (unsigned int flags)	505	=item struct ev_loop *ev_loop_new (unsigned int flags)
450		506
451	Similar to C<ev_default_loop>, but always creates a new event loop that is	507	Similar to C<ev_default_loop>, but always creates a new event loop that is
452	always distinct from the default loop. Unlike the default loop, it cannot	508	always distinct from the default loop. Unlike the default loop, it cannot
453	handle signal and child watchers, and attempts to do so will be greeted by	509	handle signal and child watchers, and attempts to do so will be greeted by
454	undefined behaviour (or a failed assertion if assertions are enabled).	510	undefined behaviour (or a failed assertion if assertions are enabled).
455		511
		512	Note that this function I<is> thread-safe, and the recommended way to use
		513	libev with threads is indeed to create one loop per thread, and using the
		514	default loop in the "main" or "initial" thread.
		515
456	Example: Try to create a event loop that uses epoll and nothing else.	516	Example: Try to create a event loop that uses epoll and nothing else.
457		517
458	struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);	518	struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);
459	if (!epoller)	519	if (!epoller)
460	fatal ("no epoll found here, maybe it hides under your chair");	520	fatal ("no epoll found here, maybe it hides under your chair");
461		521
462	=item ev_default_destroy ()	522	=item ev_default_destroy ()
463		523
464	Destroys the default loop again (frees all memory and kernel state	524	Destroys the default loop again (frees all memory and kernel state
465	etc.). None of the active event watchers will be stopped in the normal	525	etc.). None of the active event watchers will be stopped in the normal
466	sense, so e.g. C<ev_is_active> might still return true. It is your	526	sense, so e.g. C<ev_is_active> might still return true. It is your
467	responsibility to either stop all watchers cleanly yoursef I<before>	527	responsibility to either stop all watchers cleanly yourself I<before>
468	calling this function, or cope with the fact afterwards (which is usually	528	calling this function, or cope with the fact afterwards (which is usually
469	the easiest thing, you can just ignore the watchers and/or C<free ()> them	529	the easiest thing, you can just ignore the watchers and/or C<free ()> them
470	for example).	530	for example).
471		531
472	Note that certain global state, such as signal state, will not be freed by	532	Note that certain global state, such as signal state, will not be freed by
…		…
504		564
505	=item ev_loop_fork (loop)	565	=item ev_loop_fork (loop)
506		566
507	Like C<ev_default_fork>, but acts on an event loop created by	567	Like C<ev_default_fork>, but acts on an event loop created by
508	C<ev_loop_new>. Yes, you have to call this on every allocated event loop	568	C<ev_loop_new>. Yes, you have to call this on every allocated event loop
509	after fork, and how you do this is entirely your own problem.	569	after fork that you want to re-use in the child, and how you do this is
		570	entirely your own problem.
		571
		572	=item int ev_is_default_loop (loop)
		573
		574	Returns true when the given loop is, in fact, the default loop, and false
		575	otherwise.
510		576
511	=item unsigned int ev_loop_count (loop)	577	=item unsigned int ev_loop_count (loop)
512		578
513	Returns the count of loop iterations for the loop, which is identical to	579	Returns the count of loop iterations for the loop, which is identical to
514	the number of times libev did poll for new events. It starts at C<0> and	580	the number of times libev did poll for new events. It starts at C<0> and
…		…
529	received events and started processing them. This timestamp does not	595	received events and started processing them. This timestamp does not
530	change as long as callbacks are being processed, and this is also the base	596	change as long as callbacks are being processed, and this is also the base
531	time used for relative timers. You can treat it as the timestamp of the	597	time used for relative timers. You can treat it as the timestamp of the
532	event occurring (or more correctly, libev finding out about it).	598	event occurring (or more correctly, libev finding out about it).
533		599
		600	=item ev_now_update (loop)
		601
		602	Establishes the current time by querying the kernel, updating the time
		603	returned by C<ev_now ()> in the progress. This is a costly operation and
		604	is usually done automatically within C<ev_loop ()>.
		605
		606	This function is rarely useful, but when some event callback runs for a
		607	very long time without entering the event loop, updating libev's idea of
		608	the current time is a good idea.
		609
		610	See also "The special problem of time updates" in the C<ev_timer> section.
		611
534	=item ev_loop (loop, int flags)	612	=item ev_loop (loop, int flags)
535		613
536	Finally, this is it, the event handler. This function usually is called	614	Finally, this is it, the event handler. This function usually is called
537	after you initialised all your watchers and you want to start handling	615	after you initialised all your watchers and you want to start handling
538	events.	616	events.
…		…
540	If the flags argument is specified as C<0>, it will not return until	618	If the flags argument is specified as C<0>, it will not return until
541	either no event watchers are active anymore or C<ev_unloop> was called.	619	either no event watchers are active anymore or C<ev_unloop> was called.
542		620
543	Please note that an explicit C<ev_unloop> is usually better than	621	Please note that an explicit C<ev_unloop> is usually better than
544	relying on all watchers to be stopped when deciding when a program has	622	relying on all watchers to be stopped when deciding when a program has
545	finished (especially in interactive programs), but having a program that	623	finished (especially in interactive programs), but having a program
546	automatically loops as long as it has to and no longer by virtue of	624	that automatically loops as long as it has to and no longer by virtue
547	relying on its watchers stopping correctly is a thing of beauty.	625	of relying on its watchers stopping correctly, that is truly a thing of
		626	beauty.
548		627
549	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle	628	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle
550	those events and any outstanding ones, but will not block your process in	629	those events and any already outstanding ones, but will not block your
551	case there are no events and will return after one iteration of the loop.	630	process in case there are no events and will return after one iteration of
		631	the loop.
552		632
553	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if	633	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if
554	neccessary) and will handle those and any outstanding ones. It will block	634	necessary) and will handle those and any already outstanding ones. It
555	your process until at least one new event arrives, and will return after	635	will block your process until at least one new event arrives (which could
556	one iteration of the loop. This is useful if you are waiting for some	636	be an event internal to libev itself, so there is no guarentee that a
557	external event in conjunction with something not expressible using other	637	user-registered callback will be called), and will return after one
		638	iteration of the loop.
		639
		640	This is useful if you are waiting for some external event in conjunction
		641	with something not expressible using other libev watchers (i.e. "roll your
558	libev watchers. However, a pair of C<ev_prepare>/C<ev_check> watchers is	642	own C<ev_loop>"). However, a pair of C<ev_prepare>/C<ev_check> watchers is
559	usually a better approach for this kind of thing.	643	usually a better approach for this kind of thing.
560		644
561	Here are the gory details of what C<ev_loop> does:	645	Here are the gory details of what C<ev_loop> does:
562		646
563	- Before the first iteration, call any pending watchers.	647	- Before the first iteration, call any pending watchers.
564	* If EVFLAG_FORKCHECK was used, check for a fork.	648	* If EVFLAG_FORKCHECK was used, check for a fork.
565	- If a fork was detected, queue and call all fork watchers.	649	- If a fork was detected (by any means), queue and call all fork watchers.
566	- Queue and call all prepare watchers.	650	- Queue and call all prepare watchers.
567	- If we have been forked, recreate the kernel state.	651	- If we have been forked, detach and recreate the kernel state
		652	as to not disturb the other process.
568	- Update the kernel state with all outstanding changes.	653	- Update the kernel state with all outstanding changes.
569	- Update the "event loop time".	654	- Update the "event loop time" (ev_now ()).
570	- Calculate for how long to sleep or block, if at all	655	- Calculate for how long to sleep or block, if at all
571	(active idle watchers, EVLOOP_NONBLOCK or not having	656	(active idle watchers, EVLOOP_NONBLOCK or not having
572	any active watchers at all will result in not sleeping).	657	any active watchers at all will result in not sleeping).
573	- Sleep if the I/O and timer collect interval say so.	658	- Sleep if the I/O and timer collect interval say so.
574	- Block the process, waiting for any events.	659	- Block the process, waiting for any events.
575	- Queue all outstanding I/O (fd) events.	660	- Queue all outstanding I/O (fd) events.
576	- Update the "event loop time" and do time jump handling.	661	- Update the "event loop time" (ev_now ()), and do time jump adjustments.
577	- Queue all outstanding timers.	662	- Queue all expired timers.
578	- Queue all outstanding periodics.	663	- Queue all expired periodics.
579	- If no events are pending now, queue all idle watchers.	664	- Unless any events are pending now, queue all idle watchers.
580	- Queue all check watchers.	665	- Queue all check watchers.
581	- Call all queued watchers in reverse order (i.e. check watchers first).	666	- Call all queued watchers in reverse order (i.e. check watchers first).
582	Signals and child watchers are implemented as I/O watchers, and will	667	Signals and child watchers are implemented as I/O watchers, and will
583	be handled here by queueing them when their watcher gets executed.	668	be handled here by queueing them when their watcher gets executed.
584	- If ev_unloop has been called, or EVLOOP_ONESHOT or EVLOOP_NONBLOCK	669	- If ev_unloop has been called, or EVLOOP_ONESHOT or EVLOOP_NONBLOCK
…		…
589	anymore.	674	anymore.
590		675
591	... queue jobs here, make sure they register event watchers as long	676	... queue jobs here, make sure they register event watchers as long
592	... as they still have work to do (even an idle watcher will do..)	677	... as they still have work to do (even an idle watcher will do..)
593	ev_loop (my_loop, 0);	678	ev_loop (my_loop, 0);
594	... jobs done. yeah!	679	... jobs done or somebody called unloop. yeah!
595		680
596	=item ev_unloop (loop, how)	681	=item ev_unloop (loop, how)
597		682
598	Can be used to make a call to C<ev_loop> return early (but only after it	683	Can be used to make a call to C<ev_loop> return early (but only after it
599	has processed all outstanding events). The C<how> argument must be either	684	has processed all outstanding events). The C<how> argument must be either
600	C<EVUNLOOP_ONE>, which will make the innermost C<ev_loop> call return, or	685	C<EVUNLOOP_ONE>, which will make the innermost C<ev_loop> call return, or
601	C<EVUNLOOP_ALL>, which will make all nested C<ev_loop> calls return.	686	C<EVUNLOOP_ALL>, which will make all nested C<ev_loop> calls return.
602		687
603	This "unloop state" will be cleared when entering C<ev_loop> again.	688	This "unloop state" will be cleared when entering C<ev_loop> again.
604		689
		690	It is safe to call C<ev_unloop> from otuside any C<ev_loop> calls.
		691
605	=item ev_ref (loop)	692	=item ev_ref (loop)
606		693
607	=item ev_unref (loop)	694	=item ev_unref (loop)
608		695
609	Ref/unref can be used to add or remove a reference count on the event	696	Ref/unref can be used to add or remove a reference count on the event
610	loop: Every watcher keeps one reference, and as long as the reference	697	loop: Every watcher keeps one reference, and as long as the reference
611	count is nonzero, C<ev_loop> will not return on its own. If you have	698	count is nonzero, C<ev_loop> will not return on its own.
		699
612	a watcher you never unregister that should not keep C<ev_loop> from	700	If you have a watcher you never unregister that should not keep C<ev_loop>
613	returning, ev_unref() after starting, and ev_ref() before stopping it. For	701	from returning, call ev_unref() after starting, and ev_ref() before
		702	stopping it.
		703
614	example, libev itself uses this for its internal signal pipe: It is not	704	As an example, libev itself uses this for its internal signal pipe: It is
615	visible to the libev user and should not keep C<ev_loop> from exiting if	705	not visible to the libev user and should not keep C<ev_loop> from exiting
616	no event watchers registered by it are active. It is also an excellent	706	if no event watchers registered by it are active. It is also an excellent
617	way to do this for generic recurring timers or from within third-party	707	way to do this for generic recurring timers or from within third-party
618	libraries. Just remember to I<unref after start> and I<ref before stop>	708	libraries. Just remember to I<unref after start> and I<ref before stop>
619	(but only if the watcher wasn't active before, or was active before,	709	(but only if the watcher wasn't active before, or was active before,
620	respectively).	710	respectively).
621		711
622	Example: Create a signal watcher, but keep it from keeping C<ev_loop>	712	Example: Create a signal watcher, but keep it from keeping C<ev_loop>
623	running when nothing else is active.	713	running when nothing else is active.
624		714
625	struct ev_signal exitsig;	715	ev_signal exitsig;
626	ev_signal_init (&exitsig, sig_cb, SIGINT);	716	ev_signal_init (&exitsig, sig_cb, SIGINT);
627	ev_signal_start (loop, &exitsig);	717	ev_signal_start (loop, &exitsig);
628	evf_unref (loop);	718	evf_unref (loop);
629		719
630	Example: For some weird reason, unregister the above signal handler again.	720	Example: For some weird reason, unregister the above signal handler again.
631		721
632	ev_ref (loop);	722	ev_ref (loop);
633	ev_signal_stop (loop, &exitsig);	723	ev_signal_stop (loop, &exitsig);
634		724
635	=item ev_set_io_collect_interval (loop, ev_tstamp interval)	725	=item ev_set_io_collect_interval (loop, ev_tstamp interval)
636		726
637	=item ev_set_timeout_collect_interval (loop, ev_tstamp interval)	727	=item ev_set_timeout_collect_interval (loop, ev_tstamp interval)
638		728
639	These advanced functions influence the time that libev will spend waiting	729	These advanced functions influence the time that libev will spend waiting
640	for events. Both are by default C<0>, meaning that libev will try to	730	for events. Both time intervals are by default C<0>, meaning that libev
641	invoke timer/periodic callbacks and I/O callbacks with minimum latency.	731	will try to invoke timer/periodic callbacks and I/O callbacks with minimum
		732	latency.
642		733
643	Setting these to a higher value (the C<interval> I<must> be >= C<0>)	734	Setting these to a higher value (the C<interval> I<must> be >= C<0>)
644	allows libev to delay invocation of I/O and timer/periodic callbacks to	735	allows libev to delay invocation of I/O and timer/periodic callbacks
645	increase efficiency of loop iterations.	736	to increase efficiency of loop iterations (or to increase power-saving
		737	opportunities).
646		738
647	The background is that sometimes your program runs just fast enough to	739	The idea is that sometimes your program runs just fast enough to handle
648	handle one (or very few) event(s) per loop iteration. While this makes	740	one (or very few) event(s) per loop iteration. While this makes the
649	the program responsive, it also wastes a lot of CPU time to poll for new	741	program responsive, it also wastes a lot of CPU time to poll for new
650	events, especially with backends like C<select ()> which have a high	742	events, especially with backends like C<select ()> which have a high
651	overhead for the actual polling but can deliver many events at once.	743	overhead for the actual polling but can deliver many events at once.
652		744
653	By setting a higher I<io collect interval> you allow libev to spend more	745	By setting a higher I<io collect interval> you allow libev to spend more
654	time collecting I/O events, so you can handle more events per iteration,	746	time collecting I/O events, so you can handle more events per iteration,
…		…
656	C<ev_timer>) will be not affected. Setting this to a non-null value will	748	C<ev_timer>) will be not affected. Setting this to a non-null value will
657	introduce an additional C<ev_sleep ()> call into most loop iterations.	749	introduce an additional C<ev_sleep ()> call into most loop iterations.
658		750
659	Likewise, by setting a higher I<timeout collect interval> you allow libev	751	Likewise, by setting a higher I<timeout collect interval> you allow libev
660	to spend more time collecting timeouts, at the expense of increased	752	to spend more time collecting timeouts, at the expense of increased
661	latency (the watcher callback will be called later). C<ev_io> watchers	753	latency/jitter/inexactness (the watcher callback will be called
662	will not be affected. Setting this to a non-null value will not introduce	754	later). C<ev_io> watchers will not be affected. Setting this to a non-null
663	any overhead in libev.	755	value will not introduce any overhead in libev.
664		756
665	Many (busy) programs can usually benefit by setting the io collect	757	Many (busy) programs can usually benefit by setting the I/O collect
666	interval to a value near C<0.1> or so, which is often enough for	758	interval to a value near C<0.1> or so, which is often enough for
667	interactive servers (of course not for games), likewise for timeouts. It	759	interactive servers (of course not for games), likewise for timeouts. It
668	usually doesn't make much sense to set it to a lower value than C<0.01>,	760	usually doesn't make much sense to set it to a lower value than C<0.01>,
669	as this approsaches the timing granularity of most systems.	761	as this approaches the timing granularity of most systems.
		762
		763	Setting the I<timeout collect interval> can improve the opportunity for
		764	saving power, as the program will "bundle" timer callback invocations that
		765	are "near" in time together, by delaying some, thus reducing the number of
		766	times the process sleeps and wakes up again. Another useful technique to
		767	reduce iterations/wake-ups is to use C<ev_periodic> watchers and make sure
		768	they fire on, say, one-second boundaries only.
		769
		770	=item ev_loop_verify (loop)
		771
		772	This function only does something when C<EV_VERIFY> support has been
		773	compiled in. which is the default for non-minimal builds. It tries to go
		774	through all internal structures and checks them for validity. If anything
		775	is found to be inconsistent, it will print an error message to standard
		776	error and call C<abort ()>.
		777
		778	This can be used to catch bugs inside libev itself: under normal
		779	circumstances, this function will never abort as of course libev keeps its
		780	data structures consistent.
670		781
671	=back	782	=back
672		783
673		784
674	=head1 ANATOMY OF A WATCHER	785	=head1 ANATOMY OF A WATCHER
675		786
676	A watcher is a structure that you create and register to record your	787	A watcher is a structure that you create and register to record your
677	interest in some event. For instance, if you want to wait for STDIN to	788	interest in some event. For instance, if you want to wait for STDIN to
678	become readable, you would create an C<ev_io> watcher for that:	789	become readable, you would create an C<ev_io> watcher for that:
679		790
680	static void my_cb (struct ev_loop *loop, struct ev_io *w, int revents)	791	static void my_cb (struct ev_loop *loop, ev_io *w, int revents)
681	{	792	{
682	ev_io_stop (w);	793	ev_io_stop (w);
683	ev_unloop (loop, EVUNLOOP_ALL);	794	ev_unloop (loop, EVUNLOOP_ALL);
684	}	795	}
685		796
686	struct ev_loop *loop = ev_default_loop (0);	797	struct ev_loop *loop = ev_default_loop (0);
687	struct ev_io stdin_watcher;	798	ev_io stdin_watcher;
688	ev_init (&stdin_watcher, my_cb);	799	ev_init (&stdin_watcher, my_cb);
689	ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);	800	ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);
690	ev_io_start (loop, &stdin_watcher);	801	ev_io_start (loop, &stdin_watcher);
691	ev_loop (loop, 0);	802	ev_loop (loop, 0);
692		803
693	As you can see, you are responsible for allocating the memory for your	804	As you can see, you are responsible for allocating the memory for your
694	watcher structures (and it is usually a bad idea to do this on the stack,	805	watcher structures (and it is usually a bad idea to do this on the stack,
695	although this can sometimes be quite valid).	806	although this can sometimes be quite valid).
696		807
697	Each watcher structure must be initialised by a call to C<ev_init	808	Each watcher structure must be initialised by a call to C<ev_init
698	(watcher *, callback)>, which expects a callback to be provided. This	809	(watcher *, callback)>, which expects a callback to be provided. This
699	callback gets invoked each time the event occurs (or, in the case of io	810	callback gets invoked each time the event occurs (or, in the case of I/O
700	watchers, each time the event loop detects that the file descriptor given	811	watchers, each time the event loop detects that the file descriptor given
701	is readable and/or writable).	812	is readable and/or writable).
702		813
703	Each watcher type has its own C<< ev_<type>_set (watcher *, ...) >> macro	814	Each watcher type has its own C<< ev_<type>_set (watcher *, ...) >> macro
704	with arguments specific to this watcher type. There is also a macro	815	with arguments specific to this watcher type. There is also a macro
…		…
780		891
781	The given async watcher has been asynchronously notified (see C<ev_async>).	892	The given async watcher has been asynchronously notified (see C<ev_async>).
782		893
783	=item C<EV_ERROR>	894	=item C<EV_ERROR>
784		895
785	An unspecified error has occured, the watcher has been stopped. This might	896	An unspecified error has occurred, the watcher has been stopped. This might
786	happen because the watcher could not be properly started because libev	897	happen because the watcher could not be properly started because libev
787	ran out of memory, a file descriptor was found to be closed or any other	898	ran out of memory, a file descriptor was found to be closed or any other
		899	problem. Libev considers these application bugs.
		900
788	problem. You best act on it by reporting the problem and somehow coping	901	You best act on it by reporting the problem and somehow coping with the
789	with the watcher being stopped.	902	watcher being stopped. Note that well-written programs should not receive
		903	an error ever, so when your watcher receives it, this usually indicates a
		904	bug in your program.
790		905
791	Libev will usually signal a few "dummy" events together with an error,	906	Libev will usually signal a few "dummy" events together with an error, for
792	for example it might indicate that a fd is readable or writable, and if	907	example it might indicate that a fd is readable or writable, and if your
793	your callbacks is well-written it can just attempt the operation and cope	908	callbacks is well-written it can just attempt the operation and cope with
794	with the error from read() or write(). This will not work in multithreaded	909	the error from read() or write(). This will not work in multi-threaded
795	programs, though, so beware.	910	programs, though, as the fd could already be closed and reused for another
		911	thing, so beware.
796		912
797	=back	913	=back
798		914
799	=head2 GENERIC WATCHER FUNCTIONS	915	=head2 GENERIC WATCHER FUNCTIONS
800		916
…		…
813	which rolls both calls into one.	929	which rolls both calls into one.
814		930
815	You can reinitialise a watcher at any time as long as it has been stopped	931	You can reinitialise a watcher at any time as long as it has been stopped
816	(or never started) and there are no pending events outstanding.	932	(or never started) and there are no pending events outstanding.
817		933
818	The callback is always of type C<void (*)(ev_loop *loop, ev_TYPE *watcher,	934	The callback is always of type C<void (*)(struct ev_loop *loop, ev_TYPE *watcher,
819	int revents)>.	935	int revents)>.
		936
		937	Example: Initialise an C<ev_io> watcher in two steps.
		938
		939	ev_io w;
		940	ev_init (&w, my_cb);
		941	ev_io_set (&w, STDIN_FILENO, EV_READ);
820		942
821	=item C<ev_TYPE_set> (ev_TYPE *, [args])	943	=item C<ev_TYPE_set> (ev_TYPE *, [args])
822		944
823	This macro initialises the type-specific parts of a watcher. You need to	945	This macro initialises the type-specific parts of a watcher. You need to
824	call C<ev_init> at least once before you call this macro, but you can	946	call C<ev_init> at least once before you call this macro, but you can
…		…
827	difference to the C<ev_init> macro).	949	difference to the C<ev_init> macro).
828		950
829	Although some watcher types do not have type-specific arguments	951	Although some watcher types do not have type-specific arguments
830	(e.g. C<ev_prepare>) you still need to call its C<set> macro.	952	(e.g. C<ev_prepare>) you still need to call its C<set> macro.
831		953
		954	See C<ev_init>, above, for an example.
		955
832	=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])	956	=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])
833		957
834	This convinience macro rolls both C<ev_init> and C<ev_TYPE_set> macro	958	This convenience macro rolls both C<ev_init> and C<ev_TYPE_set> macro
835	calls into a single call. This is the most convinient method to initialise	959	calls into a single call. This is the most convenient method to initialise
836	a watcher. The same limitations apply, of course.	960	a watcher. The same limitations apply, of course.
		961
		962	Example: Initialise and set an C<ev_io> watcher in one step.
		963
		964	ev_io_init (&w, my_cb, STDIN_FILENO, EV_READ);
837		965
838	=item C<ev_TYPE_start> (loop *, ev_TYPE *watcher)	966	=item C<ev_TYPE_start> (loop *, ev_TYPE *watcher)
839		967
840	Starts (activates) the given watcher. Only active watchers will receive	968	Starts (activates) the given watcher. Only active watchers will receive
841	events. If the watcher is already active nothing will happen.	969	events. If the watcher is already active nothing will happen.
842		970
		971	Example: Start the C<ev_io> watcher that is being abused as example in this
		972	whole section.
		973
		974	ev_io_start (EV_DEFAULT_UC, &w);
		975
843	=item C<ev_TYPE_stop> (loop *, ev_TYPE *watcher)	976	=item C<ev_TYPE_stop> (loop *, ev_TYPE *watcher)
844		977
845	Stops the given watcher again (if active) and clears the pending	978	Stops the given watcher if active, and clears the pending status (whether
		979	the watcher was active or not).
		980
846	status. It is possible that stopped watchers are pending (for example,	981	It is possible that stopped watchers are pending - for example,
847	non-repeating timers are being stopped when they become pending), but	982	non-repeating timers are being stopped when they become pending - but
848	C<ev_TYPE_stop> ensures that the watcher is neither active nor pending. If	983	calling C<ev_TYPE_stop> ensures that the watcher is neither active nor
849	you want to free or reuse the memory used by the watcher it is therefore a	984	pending. If you want to free or reuse the memory used by the watcher it is
850	good idea to always call its C<ev_TYPE_stop> function.	985	therefore a good idea to always call its C<ev_TYPE_stop> function.
851		986
852	=item bool ev_is_active (ev_TYPE *watcher)	987	=item bool ev_is_active (ev_TYPE *watcher)
853		988
854	Returns a true value iff the watcher is active (i.e. it has been started	989	Returns a true value iff the watcher is active (i.e. it has been started
855	and not yet been stopped). As long as a watcher is active you must not modify	990	and not yet been stopped). As long as a watcher is active you must not modify
…		…
903		1038
904	=item ev_invoke (loop, ev_TYPE *watcher, int revents)	1039	=item ev_invoke (loop, ev_TYPE *watcher, int revents)
905		1040
906	Invoke the C<watcher> with the given C<loop> and C<revents>. Neither	1041	Invoke the C<watcher> with the given C<loop> and C<revents>. Neither
907	C<loop> nor C<revents> need to be valid as long as the watcher callback	1042	C<loop> nor C<revents> need to be valid as long as the watcher callback
908	can deal with that fact.	1043	can deal with that fact, as both are simply passed through to the
		1044	callback.
909		1045
910	=item int ev_clear_pending (loop, ev_TYPE *watcher)	1046	=item int ev_clear_pending (loop, ev_TYPE *watcher)
911		1047
912	If the watcher is pending, this function returns clears its pending status	1048	If the watcher is pending, this function clears its pending status and
913	and returns its C<revents> bitset (as if its callback was invoked). If the	1049	returns its C<revents> bitset (as if its callback was invoked). If the
914	watcher isn't pending it does nothing and returns C<0>.	1050	watcher isn't pending it does nothing and returns C<0>.
915		1051
		1052	Sometimes it can be useful to "poll" a watcher instead of waiting for its
		1053	callback to be invoked, which can be accomplished with this function.
		1054
916	=back	1055	=back
917		1056
918		1057
919	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER	1058	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER
920		1059
921	Each watcher has, by default, a member C<void *data> that you can change	1060	Each watcher has, by default, a member C<void *data> that you can change
922	and read at any time, libev will completely ignore it. This can be used	1061	and read at any time: libev will completely ignore it. This can be used
923	to associate arbitrary data with your watcher. If you need more data and	1062	to associate arbitrary data with your watcher. If you need more data and
924	don't want to allocate memory and store a pointer to it in that data	1063	don't want to allocate memory and store a pointer to it in that data
925	member, you can also "subclass" the watcher type and provide your own	1064	member, you can also "subclass" the watcher type and provide your own
926	data:	1065	data:
927		1066
928	struct my_io	1067	struct my_io
929	{	1068	{
930	struct ev_io io;	1069	ev_io io;
931	int otherfd;	1070	int otherfd;
932	void *somedata;	1071	void *somedata;
933	struct whatever *mostinteresting;	1072	struct whatever *mostinteresting;
934	}	1073	};
		1074
		1075	...
		1076	struct my_io w;
		1077	ev_io_init (&w.io, my_cb, fd, EV_READ);
935		1078
936	And since your callback will be called with a pointer to the watcher, you	1079	And since your callback will be called with a pointer to the watcher, you
937	can cast it back to your own type:	1080	can cast it back to your own type:
938		1081
939	static void my_cb (struct ev_loop *loop, struct ev_io *w_, int revents)	1082	static void my_cb (struct ev_loop *loop, ev_io *w_, int revents)
940	{	1083	{
941	struct my_io *w = (struct my_io *)w_;	1084	struct my_io *w = (struct my_io *)w_;
942	...	1085	...
943	}	1086	}
944		1087
945	More interesting and less C-conformant ways of casting your callback type	1088	More interesting and less C-conformant ways of casting your callback type
946	instead have been omitted.	1089	instead have been omitted.
947		1090
948	Another common scenario is having some data structure with multiple	1091	Another common scenario is to use some data structure with multiple
949	watchers:	1092	embedded watchers:
950		1093
951	struct my_biggy	1094	struct my_biggy
952	{	1095	{
953	int some_data;	1096	int some_data;
954	ev_timer t1;	1097	ev_timer t1;
955	ev_timer t2;	1098	ev_timer t2;
956	}	1099	}
957		1100
958	In this case getting the pointer to C<my_biggy> is a bit more complicated,	1101	In this case getting the pointer to C<my_biggy> is a bit more
959	you need to use C<offsetof>:	1102	complicated: Either you store the address of your C<my_biggy> struct
		1103	in the C<data> member of the watcher (for woozies), or you need to use
		1104	some pointer arithmetic using C<offsetof> inside your watchers (for real
		1105	programmers):
960		1106
961	#include <stddef.h>	1107	#include <stddef.h>
962		1108
963	static void	1109	static void
964	t1_cb (EV_P_ struct ev_timer *w, int revents)	1110	t1_cb (EV_P_ ev_timer *w, int revents)
965	{	1111	{
966	struct my_biggy big = (struct my_biggy *	1112	struct my_biggy big = (struct my_biggy *
967	(((char *)w) - offsetof (struct my_biggy, t1));	1113	(((char *)w) - offsetof (struct my_biggy, t1));
968	}	1114	}
969		1115
970	static void	1116	static void
971	t2_cb (EV_P_ struct ev_timer *w, int revents)	1117	t2_cb (EV_P_ ev_timer *w, int revents)
972	{	1118	{
973	struct my_biggy big = (struct my_biggy *	1119	struct my_biggy big = (struct my_biggy *
974	(((char *)w) - offsetof (struct my_biggy, t2));	1120	(((char *)w) - offsetof (struct my_biggy, t2));
975	}	1121	}
976		1122
977		1123
978	=head1 WATCHER TYPES	1124	=head1 WATCHER TYPES
979		1125
980	This section describes each watcher in detail, but will not repeat	1126	This section describes each watcher in detail, but will not repeat
…		…
1004	In general you can register as many read and/or write event watchers per	1150	In general you can register as many read and/or write event watchers per
1005	fd as you want (as long as you don't confuse yourself). Setting all file	1151	fd as you want (as long as you don't confuse yourself). Setting all file
1006	descriptors to non-blocking mode is also usually a good idea (but not	1152	descriptors to non-blocking mode is also usually a good idea (but not
1007	required if you know what you are doing).	1153	required if you know what you are doing).
1008		1154
1009	If you must do this, then force the use of a known-to-be-good backend	1155	If you cannot use non-blocking mode, then force the use of a
1010	(at the time of this writing, this includes only C<EVBACKEND_SELECT> and	1156	known-to-be-good backend (at the time of this writing, this includes only
1011	C<EVBACKEND_POLL>).	1157	C<EVBACKEND_SELECT> and C<EVBACKEND_POLL>).
1012		1158
1013	Another thing you have to watch out for is that it is quite easy to	1159	Another thing you have to watch out for is that it is quite easy to
1014	receive "spurious" readyness notifications, that is your callback might	1160	receive "spurious" readiness notifications, that is your callback might
1015	be called with C<EV_READ> but a subsequent C<read>(2) will actually block	1161	be called with C<EV_READ> but a subsequent C<read>(2) will actually block
1016	because there is no data. Not only are some backends known to create a	1162	because there is no data. Not only are some backends known to create a
1017	lot of those (for example solaris ports), it is very easy to get into	1163	lot of those (for example Solaris ports), it is very easy to get into
1018	this situation even with a relatively standard program structure. Thus	1164	this situation even with a relatively standard program structure. Thus
1019	it is best to always use non-blocking I/O: An extra C<read>(2) returning	1165	it is best to always use non-blocking I/O: An extra C<read>(2) returning
1020	C<EAGAIN> is far preferable to a program hanging until some data arrives.	1166	C<EAGAIN> is far preferable to a program hanging until some data arrives.
1021		1167
1022	If you cannot run the fd in non-blocking mode (for example you should not	1168	If you cannot run the fd in non-blocking mode (for example you should
1023	play around with an Xlib connection), then you have to seperately re-test	1169	not play around with an Xlib connection), then you have to separately
1024	whether a file descriptor is really ready with a known-to-be good interface	1170	re-test whether a file descriptor is really ready with a known-to-be good
1025	such as poll (fortunately in our Xlib example, Xlib already does this on	1171	interface such as poll (fortunately in our Xlib example, Xlib already
1026	its own, so its quite safe to use).	1172	does this on its own, so its quite safe to use). Some people additionally
		1173	use C<SIGALRM> and an interval timer, just to be sure you won't block
		1174	indefinitely.
		1175
		1176	But really, best use non-blocking mode.
1027		1177
1028	=head3 The special problem of disappearing file descriptors	1178	=head3 The special problem of disappearing file descriptors
1029		1179
1030	Some backends (e.g. kqueue, epoll) need to be told about closing a file	1180	Some backends (e.g. kqueue, epoll) need to be told about closing a file
1031	descriptor (either by calling C<close> explicitly or by any other means,	1181	descriptor (either due to calling C<close> explicitly or any other means,
1032	such as C<dup>). The reason is that you register interest in some file	1182	such as C<dup2>). The reason is that you register interest in some file
1033	descriptor, but when it goes away, the operating system will silently drop	1183	descriptor, but when it goes away, the operating system will silently drop
1034	this interest. If another file descriptor with the same number then is	1184	this interest. If another file descriptor with the same number then is
1035	registered with libev, there is no efficient way to see that this is, in	1185	registered with libev, there is no efficient way to see that this is, in
1036	fact, a different file descriptor.	1186	fact, a different file descriptor.
1037		1187
…		…
1066	To support fork in your programs, you either have to call	1216	To support fork in your programs, you either have to call
1067	C<ev_default_fork ()> or C<ev_loop_fork ()> after a fork in the child,	1217	C<ev_default_fork ()> or C<ev_loop_fork ()> after a fork in the child,
1068	enable C<EVFLAG_FORKCHECK>, or resort to C<EVBACKEND_SELECT> or	1218	enable C<EVFLAG_FORKCHECK>, or resort to C<EVBACKEND_SELECT> or
1069	C<EVBACKEND_POLL>.	1219	C<EVBACKEND_POLL>.
1070		1220
		1221	=head3 The special problem of SIGPIPE
		1222
		1223	While not really specific to libev, it is easy to forget about C<SIGPIPE>:
		1224	when writing to a pipe whose other end has been closed, your program gets
		1225	sent a SIGPIPE, which, by default, aborts your program. For most programs
		1226	this is sensible behaviour, for daemons, this is usually undesirable.
		1227
		1228	So when you encounter spurious, unexplained daemon exits, make sure you
		1229	ignore SIGPIPE (and maybe make sure you log the exit status of your daemon
		1230	somewhere, as that would have given you a big clue).
		1231
1071		1232
1072	=head3 Watcher-Specific Functions	1233	=head3 Watcher-Specific Functions
1073		1234
1074	=over 4	1235	=over 4
1075		1236
1076	=item ev_io_init (ev_io *, callback, int fd, int events)	1237	=item ev_io_init (ev_io *, callback, int fd, int events)
1077		1238
1078	=item ev_io_set (ev_io *, int fd, int events)	1239	=item ev_io_set (ev_io *, int fd, int events)
1079		1240
1080	Configures an C<ev_io> watcher. The C<fd> is the file descriptor to	1241	Configures an C<ev_io> watcher. The C<fd> is the file descriptor to
1081	rceeive events for and events is either C<EV_READ>, C<EV_WRITE> or	1242	receive events for and C<events> is either C<EV_READ>, C<EV_WRITE> or
1082	C<EV_READ | EV_WRITE> to receive the given events.	1243	C<EV_READ | EV_WRITE>, to express the desire to receive the given events.
1083		1244
1084	=item int fd [read-only]	1245	=item int fd [read-only]
1085		1246
1086	The file descriptor being watched.	1247	The file descriptor being watched.
1087		1248
…		…
1095		1256
1096	Example: Call C<stdin_readable_cb> when STDIN_FILENO has become, well	1257	Example: Call C<stdin_readable_cb> when STDIN_FILENO has become, well
1097	readable, but only once. Since it is likely line-buffered, you could	1258	readable, but only once. Since it is likely line-buffered, you could
1098	attempt to read a whole line in the callback.	1259	attempt to read a whole line in the callback.
1099		1260
1100	static void	1261	static void
1101	stdin_readable_cb (struct ev_loop *loop, struct ev_io *w, int revents)	1262	stdin_readable_cb (struct ev_loop *loop, ev_io *w, int revents)
1102	{	1263	{
1103	ev_io_stop (loop, w);	1264	ev_io_stop (loop, w);
1104	.. read from stdin here (or from w->fd) and haqndle any I/O errors	1265	.. read from stdin here (or from w->fd) and handle any I/O errors
1105	}	1266	}
1106		1267
1107	...	1268	...
1108	struct ev_loop *loop = ev_default_init (0);	1269	struct ev_loop *loop = ev_default_init (0);
1109	struct ev_io stdin_readable;	1270	ev_io stdin_readable;
1110	ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);	1271	ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
1111	ev_io_start (loop, &stdin_readable);	1272	ev_io_start (loop, &stdin_readable);
1112	ev_loop (loop, 0);	1273	ev_loop (loop, 0);
1113		1274
1114		1275
1115	=head2 C<ev_timer> - relative and optionally repeating timeouts	1276	=head2 C<ev_timer> - relative and optionally repeating timeouts
1116		1277
1117	Timer watchers are simple relative timers that generate an event after a	1278	Timer watchers are simple relative timers that generate an event after a
1118	given time, and optionally repeating in regular intervals after that.	1279	given time, and optionally repeating in regular intervals after that.
1119		1280
1120	The timers are based on real time, that is, if you register an event that	1281	The timers are based on real time, that is, if you register an event that
1121	times out after an hour and you reset your system clock to last years	1282	times out after an hour and you reset your system clock to January last
1122	time, it will still time out after (roughly) and hour. "Roughly" because	1283	year, it will still time out after (roughly) one hour. "Roughly" because
1123	detecting time jumps is hard, and some inaccuracies are unavoidable (the	1284	detecting time jumps is hard, and some inaccuracies are unavoidable (the
1124	monotonic clock option helps a lot here).	1285	monotonic clock option helps a lot here).
		1286
		1287	The callback is guaranteed to be invoked only I<after> its timeout has
		1288	passed, but if multiple timers become ready during the same loop iteration
		1289	then order of execution is undefined.
		1290
		1291	=head3 Be smart about timeouts
		1292
		1293	Many real-world problems invole some kind of time-out, usually for error
		1294	recovery. A typical example is an HTTP request - if the other side hangs,
		1295	you want to raise some error after a while.
		1296
		1297	Here are some ways on how to handle this problem, from simple and
		1298	inefficient to very efficient.
		1299
		1300	In the following examples a 60 second activity timeout is assumed - a
		1301	timeout that gets reset to 60 seconds each time some data ("a lifesign")
		1302	was received.
		1303
		1304	=over 4
		1305
		1306	=item 1. Use a timer and stop, reinitialise, start it on activity.
		1307
		1308	This is the most obvious, but not the most simple way: In the beginning,
		1309	start the watcher:
		1310
		1311	ev_timer_init (timer, callback, 60., 0.);
		1312	ev_timer_start (loop, timer);
		1313
		1314	Then, each time there is some activity, C<ev_timer_stop> the timer,
		1315	initialise it again, and start it:
		1316
		1317	ev_timer_stop (loop, timer);
		1318	ev_timer_set (timer, 60., 0.);
		1319	ev_timer_start (loop, timer);
		1320
		1321	This is relatively simple to implement, but means that each time there
		1322	is some activity, libev will first have to remove the timer from it's
		1323	internal data strcuture and then add it again.
		1324
		1325	=item 2. Use a timer and re-start it with C<ev_timer_again> inactivity.
		1326
		1327	This is the easiest way, and involves using C<ev_timer_again> instead of
		1328	C<ev_timer_start>.
		1329
		1330	For this, configure an C<ev_timer> with a C<repeat> value of C<60> and
		1331	then call C<ev_timer_again> at start and each time you successfully read
		1332	or write some data. If you go into an idle state where you do not expect
		1333	data to travel on the socket, you can C<ev_timer_stop> the timer, and
		1334	C<ev_timer_again> will automatically restart it if need be.
		1335
		1336	That means you can ignore the C<after> value and C<ev_timer_start>
		1337	altogether and only ever use the C<repeat> value and C<ev_timer_again>.
		1338
		1339	At start:
		1340
		1341	ev_timer_init (timer, callback, 0., 60.);
		1342	ev_timer_again (loop, timer);
		1343
		1344	Each time you receive some data:
		1345
		1346	ev_timer_again (loop, timer);
		1347
		1348	It is even possible to change the time-out on the fly:
		1349
		1350	timer->repeat = 30.;
		1351	ev_timer_again (loop, timer);
		1352
		1353	This is slightly more efficient then stopping/starting the timer each time
		1354	you want to modify its timeout value, as libev does not have to completely
		1355	remove and re-insert the timer from/into it's internal data structure.
		1356
		1357	=item 3. Let the timer time out, but then re-arm it as required.
		1358
		1359	This method is more tricky, but usually most efficient: Most timeouts are
		1360	relatively long compared to the loop iteration time - in our example,
		1361	within 60 seconds, there are usually many I/O events with associated
		1362	activity resets.
		1363
		1364	In this case, it would be more efficient to leave the C<ev_timer> alone,
		1365	but remember the time of last activity, and check for a real timeout only
		1366	within the callback:
		1367
		1368	ev_tstamp last_activity; // time of last activity
		1369
		1370	static void
		1371	callback (EV_P_ ev_timer *w, int revents)
		1372	{
		1373	ev_tstamp now = ev_now (EV_A);
		1374	ev_tstamp timeout = last_activity + 60.;
		1375
		1376	// if last_activity is older than now - timeout, we did time out
		1377	if (timeout < now)
		1378	{
		1379	// timeout occured, take action
		1380	}
		1381	else
		1382	{
		1383	// callback was invoked, but there was some activity, re-arm
		1384	// to fire in last_activity + 60.
		1385	w->again = timeout - now;
		1386	ev_timer_again (EV_A_ w);
		1387	}
		1388	}
		1389
		1390	To summarise the callback: first calculate the real time-out (defined as
		1391	"60 seconds after the last activity"), then check if that time has been
		1392	reached, which means there was a real timeout. Otherwise the callback was
		1393	invoked too early (timeout is in the future), so re-schedule the timer to
		1394	fire at that future time.
		1395
		1396	Note how C<ev_timer_again> is used, taking advantage of the
		1397	C<ev_timer_again> optimisation when the timer is already running.
		1398
		1399	This scheme causes more callback invocations (about one every 60 seconds),
		1400	but virtually no calls to libev to change the timeout.
		1401
		1402	To start the timer, simply intiialise the watcher and C<last_activity>,
		1403	then call the callback:
		1404
		1405	ev_timer_init (timer, callback);
		1406	last_activity = ev_now (loop);
		1407	callback (loop, timer, EV_TIMEOUT);
		1408
		1409	And when there is some activity, simply remember the time in
		1410	C<last_activity>:
		1411
		1412	last_actiivty = ev_now (loop);
		1413
		1414	This technique is slightly more complex, but in most cases where the
		1415	time-out is unlikely to be triggered, much more efficient.
		1416
		1417	=back
		1418
		1419	=head3 The special problem of time updates
		1420
		1421	Establishing the current time is a costly operation (it usually takes at
		1422	least two system calls): EV therefore updates its idea of the current
		1423	time only before and after C<ev_loop> collects new events, which causes a
		1424	growing difference between C<ev_now ()> and C<ev_time ()> when handling
		1425	lots of events in one iteration.
1125		1426
1126	The relative timeouts are calculated relative to the C<ev_now ()>	1427	The relative timeouts are calculated relative to the C<ev_now ()>
1127	time. This is usually the right thing as this timestamp refers to the time	1428	time. This is usually the right thing as this timestamp refers to the time
1128	of the event triggering whatever timeout you are modifying/starting. If	1429	of the event triggering whatever timeout you are modifying/starting. If
1129	you suspect event processing to be delayed and you I<need> to base the timeout	1430	you suspect event processing to be delayed and you I<need> to base the
1130	on the current time, use something like this to adjust for this:	1431	timeout on the current time, use something like this to adjust for this:
1131		1432
1132	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);	1433	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);
1133		1434
1134	The callback is guarenteed to be invoked only when its timeout has passed,	1435	If the event loop is suspended for a long time, you can also force an
1135	but if multiple timers become ready during the same loop iteration then	1436	update of the time returned by C<ev_now ()> by calling C<ev_now_update
1136	order of execution is undefined.	1437	()>.
1137		1438
1138	=head3 Watcher-Specific Functions and Data Members	1439	=head3 Watcher-Specific Functions and Data Members
1139		1440
1140	=over 4	1441	=over 4
1141		1442
1142	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)	1443	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)
1143		1444
1144	=item ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)	1445	=item ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)
1145		1446
1146	Configure the timer to trigger after C<after> seconds. If C<repeat> is	1447	Configure the timer to trigger after C<after> seconds. If C<repeat>
1147	C<0.>, then it will automatically be stopped. If it is positive, then the	1448	is C<0.>, then it will automatically be stopped once the timeout is
1148	timer will automatically be configured to trigger again C<repeat> seconds	1449	reached. If it is positive, then the timer will automatically be
1149	later, again, and again, until stopped manually.	1450	configured to trigger again C<repeat> seconds later, again, and again,
		1451	until stopped manually.
1150		1452
1151	The timer itself will do a best-effort at avoiding drift, that is, if you	1453	The timer itself will do a best-effort at avoiding drift, that is, if
1152	configure a timer to trigger every 10 seconds, then it will trigger at	1454	you configure a timer to trigger every 10 seconds, then it will normally
1153	exactly 10 second intervals. If, however, your program cannot keep up with	1455	trigger at exactly 10 second intervals. If, however, your program cannot
1154	the timer (because it takes longer than those 10 seconds to do stuff) the	1456	keep up with the timer (because it takes longer than those 10 seconds to
1155	timer will not fire more than once per event loop iteration.	1457	do stuff) the timer will not fire more than once per event loop iteration.
1156		1458
1157	=item ev_timer_again (loop)	1459	=item ev_timer_again (loop, ev_timer *)
1158		1460
1159	This will act as if the timer timed out and restart it again if it is	1461	This will act as if the timer timed out and restart it again if it is
1160	repeating. The exact semantics are:	1462	repeating. The exact semantics are:
1161		1463
1162	If the timer is pending, its pending status is cleared.	1464	If the timer is pending, its pending status is cleared.
1163		1465
1164	If the timer is started but nonrepeating, stop it (as if it timed out).	1466	If the timer is started but non-repeating, stop it (as if it timed out).
1165		1467
1166	If the timer is repeating, either start it if necessary (with the	1468	If the timer is repeating, either start it if necessary (with the
1167	C<repeat> value), or reset the running timer to the C<repeat> value.	1469	C<repeat> value), or reset the running timer to the C<repeat> value.
1168		1470
1169	This sounds a bit complicated, but here is a useful and typical	1471	This sounds a bit complicated, see "Be smart about timeouts", above, for a
1170	example: Imagine you have a tcp connection and you want a so-called idle	1472	usage example.
1171	timeout, that is, you want to be called when there have been, say, 60
1172	seconds of inactivity on the socket. The easiest way to do this is to
1173	configure an C<ev_timer> with a C<repeat> value of C<60> and then call
1174	C<ev_timer_again> each time you successfully read or write some data. If
1175	you go into an idle state where you do not expect data to travel on the
1176	socket, you can C<ev_timer_stop> the timer, and C<ev_timer_again> will
1177	automatically restart it if need be.
1178
1179	That means you can ignore the C<after> value and C<ev_timer_start>
1180	altogether and only ever use the C<repeat> value and C<ev_timer_again>:
1181
1182	ev_timer_init (timer, callback, 0., 5.);
1183	ev_timer_again (loop, timer);
1184	...
1185	timer->again = 17.;
1186	ev_timer_again (loop, timer);
1187	...
1188	timer->again = 10.;
1189	ev_timer_again (loop, timer);
1190
1191	This is more slightly efficient then stopping/starting the timer each time
1192	you want to modify its timeout value.
1193		1473
1194	=item ev_tstamp repeat [read-write]	1474	=item ev_tstamp repeat [read-write]
1195		1475
1196	The current C<repeat> value. Will be used each time the watcher times out	1476	The current C<repeat> value. Will be used each time the watcher times out
1197	or C<ev_timer_again> is called and determines the next timeout (if any),	1477	or C<ev_timer_again> is called, and determines the next timeout (if any),
1198	which is also when any modifications are taken into account.	1478	which is also when any modifications are taken into account.
1199		1479
1200	=back	1480	=back
1201		1481
1202	=head3 Examples	1482	=head3 Examples
1203		1483
1204	Example: Create a timer that fires after 60 seconds.	1484	Example: Create a timer that fires after 60 seconds.
1205		1485
1206	static void	1486	static void
1207	one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)	1487	one_minute_cb (struct ev_loop *loop, ev_timer *w, int revents)
1208	{	1488	{
1209	.. one minute over, w is actually stopped right here	1489	.. one minute over, w is actually stopped right here
1210	}	1490	}
1211		1491
1212	struct ev_timer mytimer;	1492	ev_timer mytimer;
1213	ev_timer_init (&mytimer, one_minute_cb, 60., 0.);	1493	ev_timer_init (&mytimer, one_minute_cb, 60., 0.);
1214	ev_timer_start (loop, &mytimer);	1494	ev_timer_start (loop, &mytimer);
1215		1495
1216	Example: Create a timeout timer that times out after 10 seconds of	1496	Example: Create a timeout timer that times out after 10 seconds of
1217	inactivity.	1497	inactivity.
1218		1498
1219	static void	1499	static void
1220	timeout_cb (struct ev_loop *loop, struct ev_timer *w, int revents)	1500	timeout_cb (struct ev_loop *loop, ev_timer *w, int revents)
1221	{	1501	{
1222	.. ten seconds without any activity	1502	.. ten seconds without any activity
1223	}	1503	}
1224		1504
1225	struct ev_timer mytimer;	1505	ev_timer mytimer;
1226	ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */	1506	ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */
1227	ev_timer_again (&mytimer); /* start timer */	1507	ev_timer_again (&mytimer); /* start timer */
1228	ev_loop (loop, 0);	1508	ev_loop (loop, 0);
1229		1509
1230	// and in some piece of code that gets executed on any "activity":	1510	// and in some piece of code that gets executed on any "activity":
1231	// reset the timeout to start ticking again at 10 seconds	1511	// reset the timeout to start ticking again at 10 seconds
1232	ev_timer_again (&mytimer);	1512	ev_timer_again (&mytimer);
1233		1513
1234		1514
1235	=head2 C<ev_periodic> - to cron or not to cron?	1515	=head2 C<ev_periodic> - to cron or not to cron?
1236		1516
1237	Periodic watchers are also timers of a kind, but they are very versatile	1517	Periodic watchers are also timers of a kind, but they are very versatile
1238	(and unfortunately a bit complex).	1518	(and unfortunately a bit complex).
1239		1519
1240	Unlike C<ev_timer>'s, they are not based on real time (or relative time)	1520	Unlike C<ev_timer>'s, they are not based on real time (or relative time)
1241	but on wallclock time (absolute time). You can tell a periodic watcher	1521	but on wall clock time (absolute time). You can tell a periodic watcher
1242	to trigger "at" some specific point in time. For example, if you tell a	1522	to trigger after some specific point in time. For example, if you tell a
1243	periodic watcher to trigger in 10 seconds (by specifiying e.g. C<ev_now ()	1523	periodic watcher to trigger in 10 seconds (by specifying e.g. C<ev_now ()
1244	+ 10.>) and then reset your system clock to the last year, then it will	1524	+ 10.>, that is, an absolute time not a delay) and then reset your system
		1525	clock to January of the previous year, then it will take more than year
1245	take a year to trigger the event (unlike an C<ev_timer>, which would trigger	1526	to trigger the event (unlike an C<ev_timer>, which would still trigger
1246	roughly 10 seconds later).	1527	roughly 10 seconds later as it uses a relative timeout).
1247		1528
1248	They can also be used to implement vastly more complex timers, such as	1529	C<ev_periodic>s can also be used to implement vastly more complex timers,
1249	triggering an event on each midnight, local time or other, complicated,	1530	such as triggering an event on each "midnight, local time", or other
1250	rules.	1531	complicated rules.
1251		1532
1252	As with timers, the callback is guarenteed to be invoked only when the	1533	As with timers, the callback is guaranteed to be invoked only when the
1253	time (C<at>) has been passed, but if multiple periodic timers become ready	1534	time (C<at>) has passed, but if multiple periodic timers become ready
1254	during the same loop iteration then order of execution is undefined.	1535	during the same loop iteration, then order of execution is undefined.
1255		1536
1256	=head3 Watcher-Specific Functions and Data Members	1537	=head3 Watcher-Specific Functions and Data Members
1257		1538
1258	=over 4	1539	=over 4
1259		1540
1260	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)	1541	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)
1261		1542
1262	=item ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)	1543	=item ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)
1263		1544
1264	Lots of arguments, lets sort it out... There are basically three modes of	1545	Lots of arguments, lets sort it out... There are basically three modes of
1265	operation, and we will explain them from simplest to complex:	1546	operation, and we will explain them from simplest to most complex:
1266		1547
1267	=over 4	1548	=over 4
1268		1549
1269	=item * absolute timer (at = time, interval = reschedule_cb = 0)	1550	=item * absolute timer (at = time, interval = reschedule_cb = 0)
1270		1551
1271	In this configuration the watcher triggers an event at the wallclock time	1552	In this configuration the watcher triggers an event after the wall clock
1272	C<at> and doesn't repeat. It will not adjust when a time jump occurs,	1553	time C<at> has passed. It will not repeat and will not adjust when a time
1273	that is, if it is to be run at January 1st 2011 then it will run when the	1554	jump occurs, that is, if it is to be run at January 1st 2011 then it will
1274	system time reaches or surpasses this time.	1555	only run when the system clock reaches or surpasses this time.
1275		1556
1276	=item * non-repeating interval timer (at = offset, interval > 0, reschedule_cb = 0)	1557	=item * repeating interval timer (at = offset, interval > 0, reschedule_cb = 0)
1277		1558
1278	In this mode the watcher will always be scheduled to time out at the next	1559	In this mode the watcher will always be scheduled to time out at the next
1279	C<at + N * interval> time (for some integer N, which can also be negative)	1560	C<at + N * interval> time (for some integer N, which can also be negative)
1280	and then repeat, regardless of any time jumps.	1561	and then repeat, regardless of any time jumps.
1281		1562
1282	This can be used to create timers that do not drift with respect to system	1563	This can be used to create timers that do not drift with respect to the
1283	time:	1564	system clock, for example, here is a C<ev_periodic> that triggers each
		1565	hour, on the hour:
1284		1566
1285	ev_periodic_set (&periodic, 0., 3600., 0);	1567	ev_periodic_set (&periodic, 0., 3600., 0);
1286		1568
1287	This doesn't mean there will always be 3600 seconds in between triggers,	1569	This doesn't mean there will always be 3600 seconds in between triggers,
1288	but only that the the callback will be called when the system time shows a	1570	but only that the callback will be called when the system time shows a
1289	full hour (UTC), or more correctly, when the system time is evenly divisible	1571	full hour (UTC), or more correctly, when the system time is evenly divisible
1290	by 3600.	1572	by 3600.
1291		1573
1292	Another way to think about it (for the mathematically inclined) is that	1574	Another way to think about it (for the mathematically inclined) is that
1293	C<ev_periodic> will try to run the callback in this mode at the next possible	1575	C<ev_periodic> will try to run the callback in this mode at the next possible
1294	time where C<time = at (mod interval)>, regardless of any time jumps.	1576	time where C<time = at (mod interval)>, regardless of any time jumps.
1295		1577
1296	For numerical stability it is preferable that the C<at> value is near	1578	For numerical stability it is preferable that the C<at> value is near
1297	C<ev_now ()> (the current time), but there is no range requirement for	1579	C<ev_now ()> (the current time), but there is no range requirement for
1298	this value.	1580	this value, and in fact is often specified as zero.
		1581
		1582	Note also that there is an upper limit to how often a timer can fire (CPU
		1583	speed for example), so if C<interval> is very small then timing stability
		1584	will of course deteriorate. Libev itself tries to be exact to be about one
		1585	millisecond (if the OS supports it and the machine is fast enough).
1299		1586
1300	=item * manual reschedule mode (at and interval ignored, reschedule_cb = callback)	1587	=item * manual reschedule mode (at and interval ignored, reschedule_cb = callback)
1301		1588
1302	In this mode the values for C<interval> and C<at> are both being	1589	In this mode the values for C<interval> and C<at> are both being
1303	ignored. Instead, each time the periodic watcher gets scheduled, the	1590	ignored. Instead, each time the periodic watcher gets scheduled, the
1304	reschedule callback will be called with the watcher as first, and the	1591	reschedule callback will be called with the watcher as first, and the
1305	current time as second argument.	1592	current time as second argument.
1306		1593
1307	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher,	1594	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher,
1308	ever, or make any event loop modifications>. If you need to stop it,	1595	ever, or make ANY event loop modifications whatsoever>.
1309	return C<now + 1e30> (or so, fudge fudge) and stop it afterwards (e.g. by
1310	starting an C<ev_prepare> watcher, which is legal).
1311		1596
		1597	If you need to stop it, return C<now + 1e30> (or so, fudge fudge) and stop
		1598	it afterwards (e.g. by starting an C<ev_prepare> watcher, which is the
		1599	only event loop modification you are allowed to do).
		1600
1312	Its prototype is C<ev_tstamp (*reschedule_cb)(struct ev_periodic *w,	1601	The callback prototype is C<ev_tstamp (*reschedule_cb)(ev_periodic
1313	ev_tstamp now)>, e.g.:	1602	*w, ev_tstamp now)>, e.g.:
1314		1603
		1604	static ev_tstamp
1315	static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)	1605	my_rescheduler (ev_periodic *w, ev_tstamp now)
1316	{	1606	{
1317	return now + 60.;	1607	return now + 60.;
1318	}	1608	}
1319		1609
1320	It must return the next time to trigger, based on the passed time value	1610	It must return the next time to trigger, based on the passed time value
1321	(that is, the lowest time value larger than to the second argument). It	1611	(that is, the lowest time value larger than to the second argument). It
1322	will usually be called just before the callback will be triggered, but	1612	will usually be called just before the callback will be triggered, but
1323	might be called at other times, too.	1613	might be called at other times, too.
1324		1614
1325	NOTE: I<< This callback must always return a time that is later than the	1615	NOTE: I<< This callback must always return a time that is higher than or
1326	passed C<now> value >>. Not even C<now> itself will do, it I<must> be larger.	1616	equal to the passed C<now> value >>.
1327		1617
1328	This can be used to create very complex timers, such as a timer that	1618	This can be used to create very complex timers, such as a timer that
1329	triggers on each midnight, local time. To do this, you would calculate the	1619	triggers on "next midnight, local time". To do this, you would calculate the
1330	next midnight after C<now> and return the timestamp value for this. How	1620	next midnight after C<now> and return the timestamp value for this. How
1331	you do this is, again, up to you (but it is not trivial, which is the main	1621	you do this is, again, up to you (but it is not trivial, which is the main
1332	reason I omitted it as an example).	1622	reason I omitted it as an example).
1333		1623
1334	=back	1624	=back
…		…
1338	Simply stops and restarts the periodic watcher again. This is only useful	1628	Simply stops and restarts the periodic watcher again. This is only useful
1339	when you changed some parameters or the reschedule callback would return	1629	when you changed some parameters or the reschedule callback would return
1340	a different time than the last time it was called (e.g. in a crond like	1630	a different time than the last time it was called (e.g. in a crond like
1341	program when the crontabs have changed).	1631	program when the crontabs have changed).
1342		1632
		1633	=item ev_tstamp ev_periodic_at (ev_periodic *)
		1634
		1635	When active, returns the absolute time that the watcher is supposed to
		1636	trigger next.
		1637
1343	=item ev_tstamp offset [read-write]	1638	=item ev_tstamp offset [read-write]
1344		1639
1345	When repeating, this contains the offset value, otherwise this is the	1640	When repeating, this contains the offset value, otherwise this is the
1346	absolute point in time (the C<at> value passed to C<ev_periodic_set>).	1641	absolute point in time (the C<at> value passed to C<ev_periodic_set>).
1347		1642
…		…
1352		1647
1353	The current interval value. Can be modified any time, but changes only	1648	The current interval value. Can be modified any time, but changes only
1354	take effect when the periodic timer fires or C<ev_periodic_again> is being	1649	take effect when the periodic timer fires or C<ev_periodic_again> is being
1355	called.	1650	called.
1356		1651
1357	=item ev_tstamp (*reschedule_cb)(struct ev_periodic *w, ev_tstamp now) [read-write]	1652	=item ev_tstamp (*reschedule_cb)(ev_periodic *w, ev_tstamp now) [read-write]
1358		1653
1359	The current reschedule callback, or C<0>, if this functionality is	1654	The current reschedule callback, or C<0>, if this functionality is
1360	switched off. Can be changed any time, but changes only take effect when	1655	switched off. Can be changed any time, but changes only take effect when
1361	the periodic timer fires or C<ev_periodic_again> is being called.	1656	the periodic timer fires or C<ev_periodic_again> is being called.
1362		1657
1363	=item ev_tstamp at [read-only]
1364
1365	When active, contains the absolute time that the watcher is supposed to
1366	trigger next.
1367
1368	=back	1658	=back
1369		1659
1370	=head3 Examples	1660	=head3 Examples
1371		1661
1372	Example: Call a callback every hour, or, more precisely, whenever the	1662	Example: Call a callback every hour, or, more precisely, whenever the
1373	system clock is divisible by 3600. The callback invocation times have	1663	system time is divisible by 3600. The callback invocation times have
1374	potentially a lot of jittering, but good long-term stability.	1664	potentially a lot of jitter, but good long-term stability.
1375		1665
1376	static void	1666	static void
1377	clock_cb (struct ev_loop *loop, struct ev_io *w, int revents)	1667	clock_cb (struct ev_loop *loop, ev_io *w, int revents)
1378	{	1668	{
1379	... its now a full hour (UTC, or TAI or whatever your clock follows)	1669	... its now a full hour (UTC, or TAI or whatever your clock follows)
1380	}	1670	}
1381		1671
1382	struct ev_periodic hourly_tick;	1672	ev_periodic hourly_tick;
1383	ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);	1673	ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);
1384	ev_periodic_start (loop, &hourly_tick);	1674	ev_periodic_start (loop, &hourly_tick);
1385		1675
1386	Example: The same as above, but use a reschedule callback to do it:	1676	Example: The same as above, but use a reschedule callback to do it:
1387		1677
1388	#include <math.h>	1678	#include <math.h>
1389		1679
1390	static ev_tstamp	1680	static ev_tstamp
1391	my_scheduler_cb (struct ev_periodic *w, ev_tstamp now)	1681	my_scheduler_cb (ev_periodic *w, ev_tstamp now)
1392	{	1682	{
1393	return fmod (now, 3600.) + 3600.;	1683	return now + (3600. - fmod (now, 3600.));
1394	}	1684	}
1395		1685
1396	ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);	1686	ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
1397		1687
1398	Example: Call a callback every hour, starting now:	1688	Example: Call a callback every hour, starting now:
1399		1689
1400	struct ev_periodic hourly_tick;	1690	ev_periodic hourly_tick;
1401	ev_periodic_init (&hourly_tick, clock_cb,	1691	ev_periodic_init (&hourly_tick, clock_cb,
1402	fmod (ev_now (loop), 3600.), 3600., 0);	1692	fmod (ev_now (loop), 3600.), 3600., 0);
1403	ev_periodic_start (loop, &hourly_tick);	1693	ev_periodic_start (loop, &hourly_tick);
1404		1694
1405		1695
1406	=head2 C<ev_signal> - signal me when a signal gets signalled!	1696	=head2 C<ev_signal> - signal me when a signal gets signalled!
1407		1697
1408	Signal watchers will trigger an event when the process receives a specific	1698	Signal watchers will trigger an event when the process receives a specific
1409	signal one or more times. Even though signals are very asynchronous, libev	1699	signal one or more times. Even though signals are very asynchronous, libev
1410	will try it's best to deliver signals synchronously, i.e. as part of the	1700	will try it's best to deliver signals synchronously, i.e. as part of the
1411	normal event processing, like any other event.	1701	normal event processing, like any other event.
1412		1702
		1703	If you want signals asynchronously, just use C<sigaction> as you would
		1704	do without libev and forget about sharing the signal. You can even use
		1705	C<ev_async> from a signal handler to synchronously wake up an event loop.
		1706
1413	You can configure as many watchers as you like per signal. Only when the	1707	You can configure as many watchers as you like per signal. Only when the
1414	first watcher gets started will libev actually register a signal watcher	1708	first watcher gets started will libev actually register a signal handler
1415	with the kernel (thus it coexists with your own signal handlers as long	1709	with the kernel (thus it coexists with your own signal handlers as long as
1416	as you don't register any with libev). Similarly, when the last signal	1710	you don't register any with libev for the same signal). Similarly, when
1417	watcher for a signal is stopped libev will reset the signal handler to	1711	the last signal watcher for a signal is stopped, libev will reset the
1418	SIG_DFL (regardless of what it was set to before).	1712	signal handler to SIG_DFL (regardless of what it was set to before).
		1713
		1714	If possible and supported, libev will install its handlers with
		1715	C<SA_RESTART> behaviour enabled, so system calls should not be unduly
		1716	interrupted. If you have a problem with system calls getting interrupted by
		1717	signals you can block all signals in an C<ev_check> watcher and unblock
		1718	them in an C<ev_prepare> watcher.
1419		1719
1420	=head3 Watcher-Specific Functions and Data Members	1720	=head3 Watcher-Specific Functions and Data Members
1421		1721
1422	=over 4	1722	=over 4
1423		1723
…		…
1432		1732
1433	The signal the watcher watches out for.	1733	The signal the watcher watches out for.
1434		1734
1435	=back	1735	=back
1436		1736
		1737	=head3 Examples
		1738
		1739	Example: Try to exit cleanly on SIGINT.
		1740
		1741	static void
		1742	sigint_cb (struct ev_loop *loop, ev_signal *w, int revents)
		1743	{
		1744	ev_unloop (loop, EVUNLOOP_ALL);
		1745	}
		1746
		1747	ev_signal signal_watcher;
		1748	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
		1749	ev_signal_start (loop, &signal_watcher);
		1750
1437		1751
1438	=head2 C<ev_child> - watch out for process status changes	1752	=head2 C<ev_child> - watch out for process status changes
1439		1753
1440	Child watchers trigger when your process receives a SIGCHLD in response to	1754	Child watchers trigger when your process receives a SIGCHLD in response to
1441	some child status changes (most typically when a child of yours dies).	1755	some child status changes (most typically when a child of yours dies or
		1756	exits). It is permissible to install a child watcher I<after> the child
		1757	has been forked (which implies it might have already exited), as long
		1758	as the event loop isn't entered (or is continued from a watcher), i.e.,
		1759	forking and then immediately registering a watcher for the child is fine,
		1760	but forking and registering a watcher a few event loop iterations later is
		1761	not.
		1762
		1763	Only the default event loop is capable of handling signals, and therefore
		1764	you can only register child watchers in the default event loop.
		1765
		1766	=head3 Process Interaction
		1767
		1768	Libev grabs C<SIGCHLD> as soon as the default event loop is
		1769	initialised. This is necessary to guarantee proper behaviour even if
		1770	the first child watcher is started after the child exits. The occurrence
		1771	of C<SIGCHLD> is recorded asynchronously, but child reaping is done
		1772	synchronously as part of the event loop processing. Libev always reaps all
		1773	children, even ones not watched.
		1774
		1775	=head3 Overriding the Built-In Processing
		1776
		1777	Libev offers no special support for overriding the built-in child
		1778	processing, but if your application collides with libev's default child
		1779	handler, you can override it easily by installing your own handler for
		1780	C<SIGCHLD> after initialising the default loop, and making sure the
		1781	default loop never gets destroyed. You are encouraged, however, to use an
		1782	event-based approach to child reaping and thus use libev's support for
		1783	that, so other libev users can use C<ev_child> watchers freely.
		1784
		1785	=head3 Stopping the Child Watcher
		1786
		1787	Currently, the child watcher never gets stopped, even when the
		1788	child terminates, so normally one needs to stop the watcher in the
		1789	callback. Future versions of libev might stop the watcher automatically
		1790	when a child exit is detected.
1442		1791
1443	=head3 Watcher-Specific Functions and Data Members	1792	=head3 Watcher-Specific Functions and Data Members
1444		1793
1445	=over 4	1794	=over 4
1446		1795
…		…
1472		1821
1473	=back	1822	=back
1474		1823
1475	=head3 Examples	1824	=head3 Examples
1476		1825
1477	Example: Try to exit cleanly on SIGINT and SIGTERM.	1826	Example: C<fork()> a new process and install a child handler to wait for
		1827	its completion.
1478		1828
		1829	ev_child cw;
		1830
1479	static void	1831	static void
1480	sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)	1832	child_cb (EV_P_ ev_child *w, int revents)
1481	{	1833	{
1482	ev_unloop (loop, EVUNLOOP_ALL);	1834	ev_child_stop (EV_A_ w);
		1835	printf ("process %d exited with status %x\n", w->rpid, w->rstatus);
1483	}	1836	}
1484		1837
1485	struct ev_signal signal_watcher;	1838	pid_t pid = fork ();
1486	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);	1839
1487	ev_signal_start (loop, &sigint_cb);	1840	if (pid < 0)
		1841	// error
		1842	else if (pid == 0)
		1843	{
		1844	// the forked child executes here
		1845	exit (1);
		1846	}
		1847	else
		1848	{
		1849	ev_child_init (&cw, child_cb, pid, 0);
		1850	ev_child_start (EV_DEFAULT_ &cw);
		1851	}
1488		1852
1489		1853
1490	=head2 C<ev_stat> - did the file attributes just change?	1854	=head2 C<ev_stat> - did the file attributes just change?
1491		1855
1492	This watches a filesystem path for attribute changes. That is, it calls	1856	This watches a file system path for attribute changes. That is, it calls
1493	C<stat> regularly (or when the OS says it changed) and sees if it changed	1857	C<stat> regularly (or when the OS says it changed) and sees if it changed
1494	compared to the last time, invoking the callback if it did.	1858	compared to the last time, invoking the callback if it did.
1495		1859
1496	The path does not need to exist: changing from "path exists" to "path does	1860	The path does not need to exist: changing from "path exists" to "path does
1497	not exist" is a status change like any other. The condition "path does	1861	not exist" is a status change like any other. The condition "path does
…		…
1500	the stat buffer having unspecified contents.	1864	the stat buffer having unspecified contents.
1501		1865
1502	The path I<should> be absolute and I<must not> end in a slash. If it is	1866	The path I<should> be absolute and I<must not> end in a slash. If it is
1503	relative and your working directory changes, the behaviour is undefined.	1867	relative and your working directory changes, the behaviour is undefined.
1504		1868
1505	Since there is no standard to do this, the portable implementation simply	1869	Since there is no standard kernel interface to do this, the portable
1506	calls C<stat (2)> regularly on the path to see if it changed somehow. You	1870	implementation simply calls C<stat (2)> regularly on the path to see if
1507	can specify a recommended polling interval for this case. If you specify	1871	it changed somehow. You can specify a recommended polling interval for
1508	a polling interval of C<0> (highly recommended!) then a I<suitable,	1872	this case. If you specify a polling interval of C<0> (highly recommended!)
1509	unspecified default> value will be used (which you can expect to be around	1873	then a I<suitable, unspecified default> value will be used (which
1510	five seconds, although this might change dynamically). Libev will also	1874	you can expect to be around five seconds, although this might change
1511	impose a minimum interval which is currently around C<0.1>, but thats	1875	dynamically). Libev will also impose a minimum interval which is currently
1512	usually overkill.	1876	around C<0.1>, but thats usually overkill.
1513		1877
1514	This watcher type is not meant for massive numbers of stat watchers,	1878	This watcher type is not meant for massive numbers of stat watchers,
1515	as even with OS-supported change notifications, this can be	1879	as even with OS-supported change notifications, this can be
1516	resource-intensive.	1880	resource-intensive.
1517		1881
1518	At the time of this writing, only the Linux inotify interface is	1882	At the time of this writing, the only OS-specific interface implemented
1519	implemented (implementing kqueue support is left as an exercise for the	1883	is the Linux inotify interface (implementing kqueue support is left as
1520	reader). Inotify will be used to give hints only and should not change the	1884	an exercise for the reader. Note, however, that the author sees no way
1521	semantics of C<ev_stat> watchers, which means that libev sometimes needs	1885	of implementing C<ev_stat> semantics with kqueue).
1522	to fall back to regular polling again even with inotify, but changes are
1523	usually detected immediately, and if the file exists there will be no
1524	polling.
1525		1886
1526	=head3 Inotify	1887	=head3 ABI Issues (Largefile Support)
1527		1888
		1889	Libev by default (unless the user overrides this) uses the default
		1890	compilation environment, which means that on systems with large file
		1891	support disabled by default, you get the 32 bit version of the stat
		1892	structure. When using the library from programs that change the ABI to
		1893	use 64 bit file offsets the programs will fail. In that case you have to
		1894	compile libev with the same flags to get binary compatibility. This is
		1895	obviously the case with any flags that change the ABI, but the problem is
		1896	most noticeably disabled with ev_stat and large file support.
		1897
		1898	The solution for this is to lobby your distribution maker to make large
		1899	file interfaces available by default (as e.g. FreeBSD does) and not
		1900	optional. Libev cannot simply switch on large file support because it has
		1901	to exchange stat structures with application programs compiled using the
		1902	default compilation environment.
		1903
		1904	=head3 Inotify and Kqueue
		1905
1528	When C<inotify (7)> support has been compiled into libev (generally only	1906	When C<inotify (7)> support has been compiled into libev (generally
		1907	only available with Linux 2.6.25 or above due to bugs in earlier
1529	available on Linux) and present at runtime, it will be used to speed up	1908	implementations) and present at runtime, it will be used to speed up
1530	change detection where possible. The inotify descriptor will be created lazily	1909	change detection where possible. The inotify descriptor will be created
1531	when the first C<ev_stat> watcher is being started.	1910	lazily when the first C<ev_stat> watcher is being started.
1532		1911
1533	Inotify presense does not change the semantics of C<ev_stat> watchers	1912	Inotify presence does not change the semantics of C<ev_stat> watchers
1534	except that changes might be detected earlier, and in some cases, to avoid	1913	except that changes might be detected earlier, and in some cases, to avoid
1535	making regular C<stat> calls. Even in the presense of inotify support	1914	making regular C<stat> calls. Even in the presence of inotify support
1536	there are many cases where libev has to resort to regular C<stat> polling.	1915	there are many cases where libev has to resort to regular C<stat> polling,
		1916	but as long as the path exists, libev usually gets away without polling.
1537		1917
1538	(There is no support for kqueue, as apparently it cannot be used to	1918	There is no support for kqueue, as apparently it cannot be used to
1539	implement this functionality, due to the requirement of having a file	1919	implement this functionality, due to the requirement of having a file
1540	descriptor open on the object at all times).	1920	descriptor open on the object at all times, and detecting renames, unlinks
		1921	etc. is difficult.
1541		1922
1542	=head3 The special problem of stat time resolution	1923	=head3 The special problem of stat time resolution
1543		1924
1544	The C<stat ()> syscall only supports full-second resolution portably, and	1925	The C<stat ()> system call only supports full-second resolution portably, and
1545	even on systems where the resolution is higher, many filesystems still	1926	even on systems where the resolution is higher, most file systems still
1546	only support whole seconds.	1927	only support whole seconds.
1547		1928
1548	That means that, if the time is the only thing that changes, you might	1929	That means that, if the time is the only thing that changes, you can
1549	miss updates: on the first update, C<ev_stat> detects a change and calls	1930	easily miss updates: on the first update, C<ev_stat> detects a change and
1550	your callback, which does something. When there is another update within	1931	calls your callback, which does something. When there is another update
1551	the same second, C<ev_stat> will be unable to detect it.	1932	within the same second, C<ev_stat> will be unable to detect unless the
		1933	stat data does change in other ways (e.g. file size).
1552		1934
1553	The solution to this is to delay acting on a change for a second (or till	1935	The solution to this is to delay acting on a change for slightly more
1554	the next second boundary), using a roughly one-second delay C<ev_timer>	1936	than a second (or till slightly after the next full second boundary), using
1555	(C<ev_timer_set (w, 0., 1.01); ev_timer_again (loop, w)>). The C<.01>	1937	a roughly one-second-delay C<ev_timer> (e.g. C<ev_timer_set (w, 0., 1.02);
1556	is added to work around small timing inconsistencies of some operating	1938	ev_timer_again (loop, w)>).
1557	systems.	1939
		1940	The C<.02> offset is added to work around small timing inconsistencies
		1941	of some operating systems (where the second counter of the current time
		1942	might be be delayed. One such system is the Linux kernel, where a call to
		1943	C<gettimeofday> might return a timestamp with a full second later than
		1944	a subsequent C<time> call - if the equivalent of C<time ()> is used to
		1945	update file times then there will be a small window where the kernel uses
		1946	the previous second to update file times but libev might already execute
		1947	the timer callback).
1558		1948
1559	=head3 Watcher-Specific Functions and Data Members	1949	=head3 Watcher-Specific Functions and Data Members
1560		1950
1561	=over 4	1951	=over 4
1562		1952
…		…
1568	C<path>. The C<interval> is a hint on how quickly a change is expected to	1958	C<path>. The C<interval> is a hint on how quickly a change is expected to
1569	be detected and should normally be specified as C<0> to let libev choose	1959	be detected and should normally be specified as C<0> to let libev choose
1570	a suitable value. The memory pointed to by C<path> must point to the same	1960	a suitable value. The memory pointed to by C<path> must point to the same
1571	path for as long as the watcher is active.	1961	path for as long as the watcher is active.
1572		1962
1573	The callback will be receive C<EV_STAT> when a change was detected,	1963	The callback will receive an C<EV_STAT> event when a change was detected,
1574	relative to the attributes at the time the watcher was started (or the	1964	relative to the attributes at the time the watcher was started (or the
1575	last change was detected).	1965	last change was detected).
1576		1966
1577	=item ev_stat_stat (ev_stat *)	1967	=item ev_stat_stat (loop, ev_stat *)
1578		1968
1579	Updates the stat buffer immediately with new values. If you change the	1969	Updates the stat buffer immediately with new values. If you change the
1580	watched path in your callback, you could call this fucntion to avoid	1970	watched path in your callback, you could call this function to avoid
1581	detecting this change (while introducing a race condition). Can also be	1971	detecting this change (while introducing a race condition if you are not
1582	useful simply to find out the new values.	1972	the only one changing the path). Can also be useful simply to find out the
		1973	new values.
1583		1974
1584	=item ev_statdata attr [read-only]	1975	=item ev_statdata attr [read-only]
1585		1976
1586	The most-recently detected attributes of the file. Although the type is of	1977	The most-recently detected attributes of the file. Although the type is
1587	C<ev_statdata>, this is usually the (or one of the) C<struct stat> types	1978	C<ev_statdata>, this is usually the (or one of the) C<struct stat> types
1588	suitable for your system. If the C<st_nlink> member is C<0>, then there	1979	suitable for your system, but you can only rely on the POSIX-standardised
		1980	members to be present. If the C<st_nlink> member is C<0>, then there was
1589	was some error while C<stat>ing the file.	1981	some error while C<stat>ing the file.
1590		1982
1591	=item ev_statdata prev [read-only]	1983	=item ev_statdata prev [read-only]
1592		1984
1593	The previous attributes of the file. The callback gets invoked whenever	1985	The previous attributes of the file. The callback gets invoked whenever
1594	C<prev> != C<attr>.	1986	C<prev> != C<attr>, or, more precisely, one or more of these members
		1987	differ: C<st_dev>, C<st_ino>, C<st_mode>, C<st_nlink>, C<st_uid>,
		1988	C<st_gid>, C<st_rdev>, C<st_size>, C<st_atime>, C<st_mtime>, C<st_ctime>.
1595		1989
1596	=item ev_tstamp interval [read-only]	1990	=item ev_tstamp interval [read-only]
1597		1991
1598	The specified interval.	1992	The specified interval.
1599		1993
1600	=item const char *path [read-only]	1994	=item const char *path [read-only]
1601		1995
1602	The filesystem path that is being watched.	1996	The file system path that is being watched.
1603		1997
1604	=back	1998	=back
1605		1999
1606	=head3 Examples	2000	=head3 Examples
1607		2001
1608	Example: Watch C</etc/passwd> for attribute changes.	2002	Example: Watch C</etc/passwd> for attribute changes.
1609		2003
1610	static void	2004	static void
1611	passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)	2005	passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)
1612	{	2006	{
1613	/* /etc/passwd changed in some way */	2007	/* /etc/passwd changed in some way */
1614	if (w->attr.st_nlink)	2008	if (w->attr.st_nlink)
1615	{	2009	{
1616	printf ("passwd current size %ld\n", (long)w->attr.st_size);	2010	printf ("passwd current size %ld\n", (long)w->attr.st_size);
1617	printf ("passwd current atime %ld\n", (long)w->attr.st_mtime);	2011	printf ("passwd current atime %ld\n", (long)w->attr.st_mtime);
1618	printf ("passwd current mtime %ld\n", (long)w->attr.st_mtime);	2012	printf ("passwd current mtime %ld\n", (long)w->attr.st_mtime);
1619	}	2013	}
1620	else	2014	else
1621	/* you shalt not abuse printf for puts */	2015	/* you shalt not abuse printf for puts */
1622	puts ("wow, /etc/passwd is not there, expect problems. "	2016	puts ("wow, /etc/passwd is not there, expect problems. "
1623	"if this is windows, they already arrived\n");	2017	"if this is windows, they already arrived\n");
1624	}	2018	}
1625		2019
1626	...	2020	...
1627	ev_stat passwd;	2021	ev_stat passwd;
1628		2022
1629	ev_stat_init (&passwd, passwd_cb, "/etc/passwd", 0.);	2023	ev_stat_init (&passwd, passwd_cb, "/etc/passwd", 0.);
1630	ev_stat_start (loop, &passwd);	2024	ev_stat_start (loop, &passwd);
1631		2025
1632	Example: Like above, but additionally use a one-second delay so we do not	2026	Example: Like above, but additionally use a one-second delay so we do not
1633	miss updates (however, frequent updates will delay processing, too, so	2027	miss updates (however, frequent updates will delay processing, too, so
1634	one might do the work both on C<ev_stat> callback invocation I<and> on	2028	one might do the work both on C<ev_stat> callback invocation I<and> on
1635	C<ev_timer> callback invocation).	2029	C<ev_timer> callback invocation).
1636		2030
1637	static ev_stat passwd;	2031	static ev_stat passwd;
1638	static ev_timer timer;	2032	static ev_timer timer;
1639		2033
1640	static void	2034	static void
1641	timer_cb (EV_P_ ev_timer *w, int revents)	2035	timer_cb (EV_P_ ev_timer *w, int revents)
1642	{	2036	{
1643	ev_timer_stop (EV_A_ w);	2037	ev_timer_stop (EV_A_ w);
1644		2038
1645	/* now it's one second after the most recent passwd change */	2039	/* now it's one second after the most recent passwd change */
1646	}	2040	}
1647		2041
1648	static void	2042	static void
1649	stat_cb (EV_P_ ev_stat *w, int revents)	2043	stat_cb (EV_P_ ev_stat *w, int revents)
1650	{	2044	{
1651	/* reset the one-second timer */	2045	/* reset the one-second timer */
1652	ev_timer_again (EV_A_ &timer);	2046	ev_timer_again (EV_A_ &timer);
1653	}	2047	}
1654		2048
1655	...	2049	...
1656	ev_stat_init (&passwd, stat_cb, "/etc/passwd", 0.);	2050	ev_stat_init (&passwd, stat_cb, "/etc/passwd", 0.);
1657	ev_stat_start (loop, &passwd);	2051	ev_stat_start (loop, &passwd);
1658	ev_timer_init (&timer, timer_cb, 0., 1.01);	2052	ev_timer_init (&timer, timer_cb, 0., 1.02);
1659		2053
1660		2054
1661	=head2 C<ev_idle> - when you've got nothing better to do...	2055	=head2 C<ev_idle> - when you've got nothing better to do...
1662		2056
1663	Idle watchers trigger events when no other events of the same or higher	2057	Idle watchers trigger events when no other events of the same or higher
1664	priority are pending (prepare, check and other idle watchers do not	2058	priority are pending (prepare, check and other idle watchers do not count
1665	count).	2059	as receiving "events").
1666		2060
1667	That is, as long as your process is busy handling sockets or timeouts	2061	That is, as long as your process is busy handling sockets or timeouts
1668	(or even signals, imagine) of the same or higher priority it will not be	2062	(or even signals, imagine) of the same or higher priority it will not be
1669	triggered. But when your process is idle (or only lower-priority watchers	2063	triggered. But when your process is idle (or only lower-priority watchers
1670	are pending), the idle watchers are being called once per event loop	2064	are pending), the idle watchers are being called once per event loop
…		…
1694	=head3 Examples	2088	=head3 Examples
1695		2089
1696	Example: Dynamically allocate an C<ev_idle> watcher, start it, and in the	2090	Example: Dynamically allocate an C<ev_idle> watcher, start it, and in the
1697	callback, free it. Also, use no error checking, as usual.	2091	callback, free it. Also, use no error checking, as usual.
1698		2092
1699	static void	2093	static void
1700	idle_cb (struct ev_loop *loop, struct ev_idle *w, int revents)	2094	idle_cb (struct ev_loop *loop, ev_idle *w, int revents)
1701	{	2095	{
1702	free (w);	2096	free (w);
1703	// now do something you wanted to do when the program has	2097	// now do something you wanted to do when the program has
1704	// no longer anything immediate to do.	2098	// no longer anything immediate to do.
1705	}	2099	}
1706		2100
1707	struct ev_idle *idle_watcher = malloc (sizeof (struct ev_idle));	2101	ev_idle *idle_watcher = malloc (sizeof (ev_idle));
1708	ev_idle_init (idle_watcher, idle_cb);	2102	ev_idle_init (idle_watcher, idle_cb);
1709	ev_idle_start (loop, idle_cb);	2103	ev_idle_start (loop, idle_cb);
1710		2104
1711		2105
1712	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!	2106	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!
1713		2107
1714	Prepare and check watchers are usually (but not always) used in tandem:	2108	Prepare and check watchers are usually (but not always) used in pairs:
1715	prepare watchers get invoked before the process blocks and check watchers	2109	prepare watchers get invoked before the process blocks and check watchers
1716	afterwards.	2110	afterwards.
1717		2111
1718	You I<must not> call C<ev_loop> or similar functions that enter	2112	You I<must not> call C<ev_loop> or similar functions that enter
1719	the current event loop from either C<ev_prepare> or C<ev_check>	2113	the current event loop from either C<ev_prepare> or C<ev_check>
…		…
1722	those watchers, i.e. the sequence will always be C<ev_prepare>, blocking,	2116	those watchers, i.e. the sequence will always be C<ev_prepare>, blocking,
1723	C<ev_check> so if you have one watcher of each kind they will always be	2117	C<ev_check> so if you have one watcher of each kind they will always be
1724	called in pairs bracketing the blocking call.	2118	called in pairs bracketing the blocking call.
1725		2119
1726	Their main purpose is to integrate other event mechanisms into libev and	2120	Their main purpose is to integrate other event mechanisms into libev and
1727	their use is somewhat advanced. This could be used, for example, to track	2121	their use is somewhat advanced. They could be used, for example, to track
1728	variable changes, implement your own watchers, integrate net-snmp or a	2122	variable changes, implement your own watchers, integrate net-snmp or a
1729	coroutine library and lots more. They are also occasionally useful if	2123	coroutine library and lots more. They are also occasionally useful if
1730	you cache some data and want to flush it before blocking (for example,	2124	you cache some data and want to flush it before blocking (for example,
1731	in X programs you might want to do an C<XFlush ()> in an C<ev_prepare>	2125	in X programs you might want to do an C<XFlush ()> in an C<ev_prepare>
1732	watcher).	2126	watcher).
1733		2127
1734	This is done by examining in each prepare call which file descriptors need	2128	This is done by examining in each prepare call which file descriptors
1735	to be watched by the other library, registering C<ev_io> watchers for	2129	need to be watched by the other library, registering C<ev_io> watchers
1736	them and starting an C<ev_timer> watcher for any timeouts (many libraries	2130	for them and starting an C<ev_timer> watcher for any timeouts (many
1737	provide just this functionality). Then, in the check watcher you check for	2131	libraries provide exactly this functionality). Then, in the check watcher,
1738	any events that occured (by checking the pending status of all watchers	2132	you check for any events that occurred (by checking the pending status
1739	and stopping them) and call back into the library. The I/O and timer	2133	of all watchers and stopping them) and call back into the library. The
1740	callbacks will never actually be called (but must be valid nevertheless,	2134	I/O and timer callbacks will never actually be called (but must be valid
1741	because you never know, you know?).	2135	nevertheless, because you never know, you know?).
1742		2136
1743	As another example, the Perl Coro module uses these hooks to integrate	2137	As another example, the Perl Coro module uses these hooks to integrate
1744	coroutines into libev programs, by yielding to other active coroutines	2138	coroutines into libev programs, by yielding to other active coroutines
1745	during each prepare and only letting the process block if no coroutines	2139	during each prepare and only letting the process block if no coroutines
1746	are ready to run (it's actually more complicated: it only runs coroutines	2140	are ready to run (it's actually more complicated: it only runs coroutines
…		…
1749	loop from blocking if lower-priority coroutines are active, thus mapping	2143	loop from blocking if lower-priority coroutines are active, thus mapping
1750	low-priority coroutines to idle/background tasks).	2144	low-priority coroutines to idle/background tasks).
1751		2145
1752	It is recommended to give C<ev_check> watchers highest (C<EV_MAXPRI>)	2146	It is recommended to give C<ev_check> watchers highest (C<EV_MAXPRI>)
1753	priority, to ensure that they are being run before any other watchers	2147	priority, to ensure that they are being run before any other watchers
		2148	after the poll (this doesn't matter for C<ev_prepare> watchers).
		2149
1754	after the poll. Also, C<ev_check> watchers (and C<ev_prepare> watchers,	2150	Also, C<ev_check> watchers (and C<ev_prepare> watchers, too) should not
1755	too) should not activate ("feed") events into libev. While libev fully	2151	activate ("feed") events into libev. While libev fully supports this, they
1756	supports this, they will be called before other C<ev_check> watchers	2152	might get executed before other C<ev_check> watchers did their job. As
1757	did their job. As C<ev_check> watchers are often used to embed other	2153	C<ev_check> watchers are often used to embed other (non-libev) event
1758	(non-libev) event loops those other event loops might be in an unusable	2154	loops those other event loops might be in an unusable state until their
1759	state until their C<ev_check> watcher ran (always remind yourself to	2155	C<ev_check> watcher ran (always remind yourself to coexist peacefully with
1760	coexist peacefully with others).	2156	others).
1761		2157
1762	=head3 Watcher-Specific Functions and Data Members	2158	=head3 Watcher-Specific Functions and Data Members
1763		2159
1764	=over 4	2160	=over 4
1765		2161
…		…
1767		2163
1768	=item ev_check_init (ev_check *, callback)	2164	=item ev_check_init (ev_check *, callback)
1769		2165
1770	Initialises and configures the prepare or check watcher - they have no	2166	Initialises and configures the prepare or check watcher - they have no
1771	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>	2167	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>
1772	macros, but using them is utterly, utterly and completely pointless.	2168	macros, but using them is utterly, utterly, utterly and completely
		2169	pointless.
1773		2170
1774	=back	2171	=back
1775		2172
1776	=head3 Examples	2173	=head3 Examples
1777		2174
1778	There are a number of principal ways to embed other event loops or modules	2175	There are a number of principal ways to embed other event loops or modules
1779	into libev. Here are some ideas on how to include libadns into libev	2176	into libev. Here are some ideas on how to include libadns into libev
1780	(there is a Perl module named C<EV::ADNS> that does this, which you could	2177	(there is a Perl module named C<EV::ADNS> that does this, which you could
1781	use for an actually working example. Another Perl module named C<EV::Glib>	2178	use as a working example. Another Perl module named C<EV::Glib> embeds a
1782	embeds a Glib main context into libev, and finally, C<Glib::EV> embeds EV	2179	Glib main context into libev, and finally, C<Glib::EV> embeds EV into the
1783	into the Glib event loop).	2180	Glib event loop).
1784		2181
1785	Method 1: Add IO watchers and a timeout watcher in a prepare handler,	2182	Method 1: Add IO watchers and a timeout watcher in a prepare handler,
1786	and in a check watcher, destroy them and call into libadns. What follows	2183	and in a check watcher, destroy them and call into libadns. What follows
1787	is pseudo-code only of course. This requires you to either use a low	2184	is pseudo-code only of course. This requires you to either use a low
1788	priority for the check watcher or use C<ev_clear_pending> explicitly, as	2185	priority for the check watcher or use C<ev_clear_pending> explicitly, as
1789	the callbacks for the IO/timeout watchers might not have been called yet.	2186	the callbacks for the IO/timeout watchers might not have been called yet.
1790		2187
1791	static ev_io iow [nfd];	2188	static ev_io iow [nfd];
1792	static ev_timer tw;	2189	static ev_timer tw;
1793		2190
1794	static void	2191	static void
1795	io_cb (ev_loop *loop, ev_io *w, int revents)	2192	io_cb (struct ev_loop *loop, ev_io *w, int revents)
1796	{	2193	{
1797	}	2194	}
1798		2195
1799	// create io watchers for each fd and a timer before blocking	2196	// create io watchers for each fd and a timer before blocking
1800	static void	2197	static void
1801	adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents)	2198	adns_prepare_cb (struct ev_loop *loop, ev_prepare *w, int revents)
1802	{	2199	{
1803	int timeout = 3600000;	2200	int timeout = 3600000;
1804	struct pollfd fds [nfd];	2201	struct pollfd fds [nfd];
1805	// actual code will need to loop here and realloc etc.	2202	// actual code will need to loop here and realloc etc.
1806	adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));	2203	adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));
1807		2204
1808	/* the callback is illegal, but won't be called as we stop during check */	2205	/* the callback is illegal, but won't be called as we stop during check */
1809	ev_timer_init (&tw, 0, timeout * 1e-3);	2206	ev_timer_init (&tw, 0, timeout * 1e-3);
1810	ev_timer_start (loop, &tw);	2207	ev_timer_start (loop, &tw);
1811		2208
1812	// create one ev_io per pollfd	2209	// create one ev_io per pollfd
1813	for (int i = 0; i < nfd; ++i)	2210	for (int i = 0; i < nfd; ++i)
1814	{	2211	{
1815	ev_io_init (iow + i, io_cb, fds [i].fd,	2212	ev_io_init (iow + i, io_cb, fds [i].fd,
1816	((fds [i].events & POLLIN ? EV_READ : 0)	2213	((fds [i].events & POLLIN ? EV_READ : 0)
1817	| (fds [i].events & POLLOUT ? EV_WRITE : 0)));	2214	| (fds [i].events & POLLOUT ? EV_WRITE : 0)));
1818		2215
1819	fds [i].revents = 0;	2216	fds [i].revents = 0;
1820	ev_io_start (loop, iow + i);	2217	ev_io_start (loop, iow + i);
1821	}	2218	}
1822	}	2219	}
1823		2220
1824	// stop all watchers after blocking	2221	// stop all watchers after blocking
1825	static void	2222	static void
1826	adns_check_cb (ev_loop *loop, ev_check *w, int revents)	2223	adns_check_cb (struct ev_loop *loop, ev_check *w, int revents)
1827	{	2224	{
1828	ev_timer_stop (loop, &tw);	2225	ev_timer_stop (loop, &tw);
1829		2226
1830	for (int i = 0; i < nfd; ++i)	2227	for (int i = 0; i < nfd; ++i)
1831	{	2228	{
1832	// set the relevant poll flags	2229	// set the relevant poll flags
1833	// could also call adns_processreadable etc. here	2230	// could also call adns_processreadable etc. here
1834	struct pollfd *fd = fds + i;	2231	struct pollfd *fd = fds + i;
1835	int revents = ev_clear_pending (iow + i);	2232	int revents = ev_clear_pending (iow + i);
1836	if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;	2233	if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;
1837	if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;	2234	if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;
1838		2235
1839	// now stop the watcher	2236	// now stop the watcher
1840	ev_io_stop (loop, iow + i);	2237	ev_io_stop (loop, iow + i);
1841	}	2238	}
1842		2239
1843	adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));	2240	adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));
1844	}	2241	}
1845		2242
1846	Method 2: This would be just like method 1, but you run C<adns_afterpoll>	2243	Method 2: This would be just like method 1, but you run C<adns_afterpoll>
1847	in the prepare watcher and would dispose of the check watcher.	2244	in the prepare watcher and would dispose of the check watcher.
1848		2245
1849	Method 3: If the module to be embedded supports explicit event	2246	Method 3: If the module to be embedded supports explicit event
1850	notification (adns does), you can also make use of the actual watcher	2247	notification (libadns does), you can also make use of the actual watcher
1851	callbacks, and only destroy/create the watchers in the prepare watcher.	2248	callbacks, and only destroy/create the watchers in the prepare watcher.
1852		2249
1853	static void	2250	static void
1854	timer_cb (EV_P_ ev_timer *w, int revents)	2251	timer_cb (EV_P_ ev_timer *w, int revents)
1855	{	2252	{
1856	adns_state ads = (adns_state)w->data;	2253	adns_state ads = (adns_state)w->data;
1857	update_now (EV_A);	2254	update_now (EV_A);
1858		2255
1859	adns_processtimeouts (ads, &tv_now);	2256	adns_processtimeouts (ads, &tv_now);
1860	}	2257	}
1861		2258
1862	static void	2259	static void
1863	io_cb (EV_P_ ev_io *w, int revents)	2260	io_cb (EV_P_ ev_io *w, int revents)
1864	{	2261	{
1865	adns_state ads = (adns_state)w->data;	2262	adns_state ads = (adns_state)w->data;
1866	update_now (EV_A);	2263	update_now (EV_A);
1867		2264
1868	if (revents & EV_READ ) adns_processreadable (ads, w->fd, &tv_now);	2265	if (revents & EV_READ ) adns_processreadable (ads, w->fd, &tv_now);
1869	if (revents & EV_WRITE) adns_processwriteable (ads, w->fd, &tv_now);	2266	if (revents & EV_WRITE) adns_processwriteable (ads, w->fd, &tv_now);
1870	}	2267	}
1871		2268
1872	// do not ever call adns_afterpoll	2269	// do not ever call adns_afterpoll
1873		2270
1874	Method 4: Do not use a prepare or check watcher because the module you	2271	Method 4: Do not use a prepare or check watcher because the module you
1875	want to embed is too inflexible to support it. Instead, youc na override	2272	want to embed is not flexible enough to support it. Instead, you can
1876	their poll function. The drawback with this solution is that the main	2273	override their poll function. The drawback with this solution is that the
1877	loop is now no longer controllable by EV. The C<Glib::EV> module does	2274	main loop is now no longer controllable by EV. The C<Glib::EV> module uses
1878	this.	2275	this approach, effectively embedding EV as a client into the horrible
		2276	libglib event loop.
1879		2277
1880	static gint	2278	static gint
1881	event_poll_func (GPollFD *fds, guint nfds, gint timeout)	2279	event_poll_func (GPollFD *fds, guint nfds, gint timeout)
1882	{	2280	{
1883	int got_events = 0;	2281	int got_events = 0;
1884		2282
1885	for (n = 0; n < nfds; ++n)	2283	for (n = 0; n < nfds; ++n)
1886	// create/start io watcher that sets the relevant bits in fds[n] and increment got_events	2284	// create/start io watcher that sets the relevant bits in fds[n] and increment got_events
1887		2285
1888	if (timeout >= 0)	2286	if (timeout >= 0)
1889	// create/start timer	2287	// create/start timer
1890		2288
1891	// poll	2289	// poll
1892	ev_loop (EV_A_ 0);	2290	ev_loop (EV_A_ 0);
1893		2291
1894	// stop timer again	2292	// stop timer again
1895	if (timeout >= 0)	2293	if (timeout >= 0)
1896	ev_timer_stop (EV_A_ &to);	2294	ev_timer_stop (EV_A_ &to);
1897		2295
1898	// stop io watchers again - their callbacks should have set	2296	// stop io watchers again - their callbacks should have set
1899	for (n = 0; n < nfds; ++n)	2297	for (n = 0; n < nfds; ++n)
1900	ev_io_stop (EV_A_ iow [n]);	2298	ev_io_stop (EV_A_ iow [n]);
1901		2299
1902	return got_events;	2300	return got_events;
1903	}	2301	}
1904		2302
1905		2303
1906	=head2 C<ev_embed> - when one backend isn't enough...	2304	=head2 C<ev_embed> - when one backend isn't enough...
1907		2305
1908	This is a rather advanced watcher type that lets you embed one event loop	2306	This is a rather advanced watcher type that lets you embed one event loop
…		…
1914	prioritise I/O.	2312	prioritise I/O.
1915		2313
1916	As an example for a bug workaround, the kqueue backend might only support	2314	As an example for a bug workaround, the kqueue backend might only support
1917	sockets on some platform, so it is unusable as generic backend, but you	2315	sockets on some platform, so it is unusable as generic backend, but you
1918	still want to make use of it because you have many sockets and it scales	2316	still want to make use of it because you have many sockets and it scales
1919	so nicely. In this case, you would create a kqueue-based loop and embed it	2317	so nicely. In this case, you would create a kqueue-based loop and embed
1920	into your default loop (which might use e.g. poll). Overall operation will	2318	it into your default loop (which might use e.g. poll). Overall operation
1921	be a bit slower because first libev has to poll and then call kevent, but	2319	will be a bit slower because first libev has to call C<poll> and then
1922	at least you can use both at what they are best.	2320	C<kevent>, but at least you can use both mechanisms for what they are
		2321	best: C<kqueue> for scalable sockets and C<poll> if you want it to work :)
1923		2322
1924	As for prioritising I/O: rarely you have the case where some fds have	2323	As for prioritising I/O: under rare circumstances you have the case where
1925	to be watched and handled very quickly (with low latency), and even	2324	some fds have to be watched and handled very quickly (with low latency),
1926	priorities and idle watchers might have too much overhead. In this case	2325	and even priorities and idle watchers might have too much overhead. In
1927	you would put all the high priority stuff in one loop and all the rest in	2326	this case you would put all the high priority stuff in one loop and all
1928	a second one, and embed the second one in the first.	2327	the rest in a second one, and embed the second one in the first.
1929		2328
1930	As long as the watcher is active, the callback will be invoked every time	2329	As long as the watcher is active, the callback will be invoked every time
1931	there might be events pending in the embedded loop. The callback must then	2330	there might be events pending in the embedded loop. The callback must then
1932	call C<ev_embed_sweep (mainloop, watcher)> to make a single sweep and invoke	2331	call C<ev_embed_sweep (mainloop, watcher)> to make a single sweep and invoke
1933	their callbacks (you could also start an idle watcher to give the embedded	2332	their callbacks (you could also start an idle watcher to give the embedded
…		…
1941	interested in that.	2340	interested in that.
1942		2341
1943	Also, there have not currently been made special provisions for forking:	2342	Also, there have not currently been made special provisions for forking:
1944	when you fork, you not only have to call C<ev_loop_fork> on both loops,	2343	when you fork, you not only have to call C<ev_loop_fork> on both loops,
1945	but you will also have to stop and restart any C<ev_embed> watchers	2344	but you will also have to stop and restart any C<ev_embed> watchers
1946	yourself.	2345	yourself - but you can use a fork watcher to handle this automatically,
		2346	and future versions of libev might do just that.
1947		2347
1948	Unfortunately, not all backends are embeddable, only the ones returned by	2348	Unfortunately, not all backends are embeddable: only the ones returned by
1949	C<ev_embeddable_backends> are, which, unfortunately, does not include any	2349	C<ev_embeddable_backends> are, which, unfortunately, does not include any
1950	portable one.	2350	portable one.
1951		2351
1952	So when you want to use this feature you will always have to be prepared	2352	So when you want to use this feature you will always have to be prepared
1953	that you cannot get an embeddable loop. The recommended way to get around	2353	that you cannot get an embeddable loop. The recommended way to get around
1954	this is to have a separate variables for your embeddable loop, try to	2354	this is to have a separate variables for your embeddable loop, try to
1955	create it, and if that fails, use the normal loop for everything.	2355	create it, and if that fails, use the normal loop for everything.
1956		2356
		2357	=head3 C<ev_embed> and fork
		2358
		2359	While the C<ev_embed> watcher is running, forks in the embedding loop will
		2360	automatically be applied to the embedded loop as well, so no special
		2361	fork handling is required in that case. When the watcher is not running,
		2362	however, it is still the task of the libev user to call C<ev_loop_fork ()>
		2363	as applicable.
		2364
1957	=head3 Watcher-Specific Functions and Data Members	2365	=head3 Watcher-Specific Functions and Data Members
1958		2366
1959	=over 4	2367	=over 4
1960		2368
1961	=item ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)	2369	=item ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)
…		…
1964		2372
1965	Configures the watcher to embed the given loop, which must be	2373	Configures the watcher to embed the given loop, which must be
1966	embeddable. If the callback is C<0>, then C<ev_embed_sweep> will be	2374	embeddable. If the callback is C<0>, then C<ev_embed_sweep> will be
1967	invoked automatically, otherwise it is the responsibility of the callback	2375	invoked automatically, otherwise it is the responsibility of the callback
1968	to invoke it (it will continue to be called until the sweep has been done,	2376	to invoke it (it will continue to be called until the sweep has been done,
1969	if you do not want thta, you need to temporarily stop the embed watcher).	2377	if you do not want that, you need to temporarily stop the embed watcher).
1970		2378
1971	=item ev_embed_sweep (loop, ev_embed *)	2379	=item ev_embed_sweep (loop, ev_embed *)
1972		2380
1973	Make a single, non-blocking sweep over the embedded loop. This works	2381	Make a single, non-blocking sweep over the embedded loop. This works
1974	similarly to C<ev_loop (embedded_loop, EVLOOP_NONBLOCK)>, but in the most	2382	similarly to C<ev_loop (embedded_loop, EVLOOP_NONBLOCK)>, but in the most
1975	apropriate way for embedded loops.	2383	appropriate way for embedded loops.
1976		2384
1977	=item struct ev_loop *other [read-only]	2385	=item struct ev_loop *other [read-only]
1978		2386
1979	The embedded event loop.	2387	The embedded event loop.
1980		2388
…		…
1982		2390
1983	=head3 Examples	2391	=head3 Examples
1984		2392
1985	Example: Try to get an embeddable event loop and embed it into the default	2393	Example: Try to get an embeddable event loop and embed it into the default
1986	event loop. If that is not possible, use the default loop. The default	2394	event loop. If that is not possible, use the default loop. The default
1987	loop is stored in C<loop_hi>, while the mebeddable loop is stored in	2395	loop is stored in C<loop_hi>, while the embeddable loop is stored in
1988	C<loop_lo> (which is C<loop_hi> in the acse no embeddable loop can be	2396	C<loop_lo> (which is C<loop_hi> in the case no embeddable loop can be
1989	used).	2397	used).
1990		2398
1991	struct ev_loop *loop_hi = ev_default_init (0);	2399	struct ev_loop *loop_hi = ev_default_init (0);
1992	struct ev_loop *loop_lo = 0;	2400	struct ev_loop *loop_lo = 0;
1993	struct ev_embed embed;	2401	ev_embed embed;
1994		2402
1995	// see if there is a chance of getting one that works	2403	// see if there is a chance of getting one that works
1996	// (remember that a flags value of 0 means autodetection)	2404	// (remember that a flags value of 0 means autodetection)
1997	loop_lo = ev_embeddable_backends () & ev_recommended_backends ()	2405	loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
1998	? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())	2406	? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
1999	: 0;	2407	: 0;
2000		2408
2001	// if we got one, then embed it, otherwise default to loop_hi	2409	// if we got one, then embed it, otherwise default to loop_hi
2002	if (loop_lo)	2410	if (loop_lo)
2003	{	2411	{
2004	ev_embed_init (&embed, 0, loop_lo);	2412	ev_embed_init (&embed, 0, loop_lo);
2005	ev_embed_start (loop_hi, &embed);	2413	ev_embed_start (loop_hi, &embed);
2006	}	2414	}
2007	else	2415	else
2008	loop_lo = loop_hi;	2416	loop_lo = loop_hi;
2009		2417
2010	Example: Check if kqueue is available but not recommended and create	2418	Example: Check if kqueue is available but not recommended and create
2011	a kqueue backend for use with sockets (which usually work with any	2419	a kqueue backend for use with sockets (which usually work with any
2012	kqueue implementation). Store the kqueue/socket-only event loop in	2420	kqueue implementation). Store the kqueue/socket-only event loop in
2013	C<loop_socket>. (One might optionally use C<EVFLAG_NOENV>, too).	2421	C<loop_socket>. (One might optionally use C<EVFLAG_NOENV>, too).
2014		2422
2015	struct ev_loop *loop = ev_default_init (0);	2423	struct ev_loop *loop = ev_default_init (0);
2016	struct ev_loop *loop_socket = 0;	2424	struct ev_loop *loop_socket = 0;
2017	struct ev_embed embed;	2425	ev_embed embed;
2018		2426
2019	if (ev_supported_backends () & ~ev_recommended_backends () & EVBACKEND_KQUEUE)	2427	if (ev_supported_backends () & ~ev_recommended_backends () & EVBACKEND_KQUEUE)
2020	if ((loop_socket = ev_loop_new (EVBACKEND_KQUEUE))	2428	if ((loop_socket = ev_loop_new (EVBACKEND_KQUEUE))
2021	{	2429	{
2022	ev_embed_init (&embed, 0, loop_socket);	2430	ev_embed_init (&embed, 0, loop_socket);
2023	ev_embed_start (loop, &embed);	2431	ev_embed_start (loop, &embed);
2024	}	2432	}
2025		2433
2026	if (!loop_socket)	2434	if (!loop_socket)
2027	loop_socket = loop;	2435	loop_socket = loop;
2028		2436
2029	// now use loop_socket for all sockets, and loop for everything else	2437	// now use loop_socket for all sockets, and loop for everything else
2030		2438
2031		2439
2032	=head2 C<ev_fork> - the audacity to resume the event loop after a fork	2440	=head2 C<ev_fork> - the audacity to resume the event loop after a fork
2033		2441
2034	Fork watchers are called when a C<fork ()> was detected (usually because	2442	Fork watchers are called when a C<fork ()> was detected (usually because
…		…
2070	C<ev_async_sent> calls).	2478	C<ev_async_sent> calls).
2071		2479
2072	Unlike C<ev_signal> watchers, C<ev_async> works with any event loop, not	2480	Unlike C<ev_signal> watchers, C<ev_async> works with any event loop, not
2073	just the default loop.	2481	just the default loop.
2074		2482
		2483	=head3 Queueing
		2484
		2485	C<ev_async> does not support queueing of data in any way. The reason
		2486	is that the author does not know of a simple (or any) algorithm for a
		2487	multiple-writer-single-reader queue that works in all cases and doesn't
		2488	need elaborate support such as pthreads.
		2489
		2490	That means that if you want to queue data, you have to provide your own
		2491	queue. But at least I can tell you how to implement locking around your
		2492	queue:
		2493
		2494	=over 4
		2495
		2496	=item queueing from a signal handler context
		2497
		2498	To implement race-free queueing, you simply add to the queue in the signal
		2499	handler but you block the signal handler in the watcher callback. Here is
		2500	an example that does that for some fictitious SIGUSR1 handler:
		2501
		2502	static ev_async mysig;
		2503
		2504	static void
		2505	sigusr1_handler (void)
		2506	{
		2507	sometype data;
		2508
		2509	// no locking etc.
		2510	queue_put (data);
		2511	ev_async_send (EV_DEFAULT_ &mysig);
		2512	}
		2513
		2514	static void
		2515	mysig_cb (EV_P_ ev_async *w, int revents)
		2516	{
		2517	sometype data;
		2518	sigset_t block, prev;
		2519
		2520	sigemptyset (&block);
		2521	sigaddset (&block, SIGUSR1);
		2522	sigprocmask (SIG_BLOCK, &block, &prev);
		2523
		2524	while (queue_get (&data))
		2525	process (data);
		2526
		2527	if (sigismember (&prev, SIGUSR1)
		2528	sigprocmask (SIG_UNBLOCK, &block, 0);
		2529	}
		2530
		2531	(Note: pthreads in theory requires you to use C<pthread_setmask>
		2532	instead of C<sigprocmask> when you use threads, but libev doesn't do it
		2533	either...).
		2534
		2535	=item queueing from a thread context
		2536
		2537	The strategy for threads is different, as you cannot (easily) block
		2538	threads but you can easily preempt them, so to queue safely you need to
		2539	employ a traditional mutex lock, such as in this pthread example:
		2540
		2541	static ev_async mysig;
		2542	static pthread_mutex_t mymutex = PTHREAD_MUTEX_INITIALIZER;
		2543
		2544	static void
		2545	otherthread (void)
		2546	{
		2547	// only need to lock the actual queueing operation
		2548	pthread_mutex_lock (&mymutex);
		2549	queue_put (data);
		2550	pthread_mutex_unlock (&mymutex);
		2551
		2552	ev_async_send (EV_DEFAULT_ &mysig);
		2553	}
		2554
		2555	static void
		2556	mysig_cb (EV_P_ ev_async *w, int revents)
		2557	{
		2558	pthread_mutex_lock (&mymutex);
		2559
		2560	while (queue_get (&data))
		2561	process (data);
		2562
		2563	pthread_mutex_unlock (&mymutex);
		2564	}
		2565
		2566	=back
		2567
		2568
2075	=head3 Watcher-Specific Functions and Data Members	2569	=head3 Watcher-Specific Functions and Data Members
2076		2570
2077	=over 4	2571	=over 4
2078		2572
2079	=item ev_async_init (ev_async *, callback)	2573	=item ev_async_init (ev_async *, callback)
2080		2574
2081	Initialises and configures the async watcher - it has no parameters of any	2575	Initialises and configures the async watcher - it has no parameters of any
2082	kind. There is a C<ev_asynd_set> macro, but using it is utterly pointless,	2576	kind. There is a C<ev_asynd_set> macro, but using it is utterly pointless,
2083	believe me.	2577	trust me.
2084		2578
2085	=item ev_async_send (loop, ev_async *)	2579	=item ev_async_send (loop, ev_async *)
2086		2580
2087	Sends/signals/activates the given C<ev_async> watcher, that is, feeds	2581	Sends/signals/activates the given C<ev_async> watcher, that is, feeds
2088	an C<EV_ASYNC> event on the watcher into the event loop. Unlike	2582	an C<EV_ASYNC> event on the watcher into the event loop. Unlike
2089	C<ev_feed_event>, this call is safe to do in other threads, signal or	2583	C<ev_feed_event>, this call is safe to do from other threads, signal or
2090	similar contexts (see the dicusssion of C<EV_ATOMIC_T> in the embedding	2584	similar contexts (see the discussion of C<EV_ATOMIC_T> in the embedding
2091	section below on what exactly this means).	2585	section below on what exactly this means).
2092		2586
2093	This call incurs the overhead of a syscall only once per loop iteration,	2587	This call incurs the overhead of a system call only once per loop iteration,
2094	so while the overhead might be noticable, it doesn't apply to repeated	2588	so while the overhead might be noticeable, it doesn't apply to repeated
2095	calls to C<ev_async_send>.	2589	calls to C<ev_async_send>.
2096		2590
		2591	=item bool = ev_async_pending (ev_async *)
		2592
		2593	Returns a non-zero value when C<ev_async_send> has been called on the
		2594	watcher but the event has not yet been processed (or even noted) by the
		2595	event loop.
		2596
		2597	C<ev_async_send> sets a flag in the watcher and wakes up the loop. When
		2598	the loop iterates next and checks for the watcher to have become active,
		2599	it will reset the flag again. C<ev_async_pending> can be used to very
		2600	quickly check whether invoking the loop might be a good idea.
		2601
		2602	Not that this does I<not> check whether the watcher itself is pending, only
		2603	whether it has been requested to make this watcher pending.
		2604
2097	=back	2605	=back
2098		2606
2099		2607
2100	=head1 OTHER FUNCTIONS	2608	=head1 OTHER FUNCTIONS
2101		2609
…		…
2104	=over 4	2612	=over 4
2105		2613
2106	=item ev_once (loop, int fd, int events, ev_tstamp timeout, callback)	2614	=item ev_once (loop, int fd, int events, ev_tstamp timeout, callback)
2107		2615
2108	This function combines a simple timer and an I/O watcher, calls your	2616	This function combines a simple timer and an I/O watcher, calls your
2109	callback on whichever event happens first and automatically stop both	2617	callback on whichever event happens first and automatically stops both
2110	watchers. This is useful if you want to wait for a single event on an fd	2618	watchers. This is useful if you want to wait for a single event on an fd
2111	or timeout without having to allocate/configure/start/stop/free one or	2619	or timeout without having to allocate/configure/start/stop/free one or
2112	more watchers yourself.	2620	more watchers yourself.
2113		2621
2114	If C<fd> is less than 0, then no I/O watcher will be started and events	2622	If C<fd> is less than 0, then no I/O watcher will be started and the
2115	is being ignored. Otherwise, an C<ev_io> watcher for the given C<fd> and	2623	C<events> argument is being ignored. Otherwise, an C<ev_io> watcher for
2116	C<events> set will be craeted and started.	2624	the given C<fd> and C<events> set will be created and started.
2117		2625
2118	If C<timeout> is less than 0, then no timeout watcher will be	2626	If C<timeout> is less than 0, then no timeout watcher will be
2119	started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and	2627	started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and
2120	repeat = 0) will be started. While C<0> is a valid timeout, it is of	2628	repeat = 0) will be started. C<0> is a valid timeout.
2121	dubious value.
2122		2629
2123	The callback has the type C<void (*cb)(int revents, void *arg)> and gets	2630	The callback has the type C<void (*cb)(int revents, void *arg)> and gets
2124	passed an C<revents> set like normal event callbacks (a combination of	2631	passed an C<revents> set like normal event callbacks (a combination of
2125	C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMEOUT>) and the C<arg>	2632	C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMEOUT>) and the C<arg>
2126	value passed to C<ev_once>:	2633	value passed to C<ev_once>. Note that it is possible to receive I<both>
		2634	a timeout and an io event at the same time - you probably should give io
		2635	events precedence.
2127		2636
		2637	Example: wait up to ten seconds for data to appear on STDIN_FILENO.
		2638
2128	static void stdin_ready (int revents, void *arg)	2639	static void stdin_ready (int revents, void *arg)
2129	{	2640	{
2130	if (revents & EV_TIMEOUT)
2131	/* doh, nothing entered */;
2132	else if (revents & EV_READ)	2641	if (revents & EV_READ)
2133	/* stdin might have data for us, joy! */;	2642	/* stdin might have data for us, joy! */;
		2643	else if (revents & EV_TIMEOUT)
		2644	/* doh, nothing entered */;
2134	}	2645	}
2135		2646
2136	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);	2647	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
2137		2648
2138	=item ev_feed_event (ev_loop *, watcher *, int revents)	2649	=item ev_feed_event (struct ev_loop *, watcher *, int revents)
2139		2650
2140	Feeds the given event set into the event loop, as if the specified event	2651	Feeds the given event set into the event loop, as if the specified event
2141	had happened for the specified watcher (which must be a pointer to an	2652	had happened for the specified watcher (which must be a pointer to an
2142	initialised but not necessarily started event watcher).	2653	initialised but not necessarily started event watcher).
2143		2654
2144	=item ev_feed_fd_event (ev_loop *, int fd, int revents)	2655	=item ev_feed_fd_event (struct ev_loop *, int fd, int revents)
2145		2656
2146	Feed an event on the given fd, as if a file descriptor backend detected	2657	Feed an event on the given fd, as if a file descriptor backend detected
2147	the given events it.	2658	the given events it.
2148		2659
2149	=item ev_feed_signal_event (ev_loop *loop, int signum)	2660	=item ev_feed_signal_event (struct ev_loop *loop, int signum)
2150		2661
2151	Feed an event as if the given signal occured (C<loop> must be the default	2662	Feed an event as if the given signal occurred (C<loop> must be the default
2152	loop!).	2663	loop!).
2153		2664
2154	=back	2665	=back
2155		2666
2156		2667
…		…
2172		2683
2173	=item * Priorities are not currently supported. Initialising priorities	2684	=item * Priorities are not currently supported. Initialising priorities
2174	will fail and all watchers will have the same priority, even though there	2685	will fail and all watchers will have the same priority, even though there
2175	is an ev_pri field.	2686	is an ev_pri field.
2176		2687
		2688	=item * In libevent, the last base created gets the signals, in libev, the
		2689	first base created (== the default loop) gets the signals.
		2690
2177	=item * Other members are not supported.	2691	=item * Other members are not supported.
2178		2692
2179	=item * The libev emulation is I<not> ABI compatible to libevent, you need	2693	=item * The libev emulation is I<not> ABI compatible to libevent, you need
2180	to use the libev header file and library.	2694	to use the libev header file and library.
2181		2695
2182	=back	2696	=back
2183		2697
2184	=head1 C++ SUPPORT	2698	=head1 C++ SUPPORT
2185		2699
2186	Libev comes with some simplistic wrapper classes for C++ that mainly allow	2700	Libev comes with some simplistic wrapper classes for C++ that mainly allow
2187	you to use some convinience methods to start/stop watchers and also change	2701	you to use some convenience methods to start/stop watchers and also change
2188	the callback model to a model using method callbacks on objects.	2702	the callback model to a model using method callbacks on objects.
2189		2703
2190	To use it,	2704	To use it,
2191		2705
2192	#include <ev++.h>	2706	#include <ev++.h>
2193		2707
2194	This automatically includes F<ev.h> and puts all of its definitions (many	2708	This automatically includes F<ev.h> and puts all of its definitions (many
2195	of them macros) into the global namespace. All C++ specific things are	2709	of them macros) into the global namespace. All C++ specific things are
2196	put into the C<ev> namespace. It should support all the same embedding	2710	put into the C<ev> namespace. It should support all the same embedding
2197	options as F<ev.h>, most notably C<EV_MULTIPLICITY>.	2711	options as F<ev.h>, most notably C<EV_MULTIPLICITY>.
…		…
2264	your compiler is good :), then the method will be fully inlined into the	2778	your compiler is good :), then the method will be fully inlined into the
2265	thunking function, making it as fast as a direct C callback.	2779	thunking function, making it as fast as a direct C callback.
2266		2780
2267	Example: simple class declaration and watcher initialisation	2781	Example: simple class declaration and watcher initialisation
2268		2782
2269	struct myclass	2783	struct myclass
2270	{	2784	{
2271	void io_cb (ev::io &w, int revents) { }	2785	void io_cb (ev::io &w, int revents) { }
2272	}	2786	}
2273		2787
2274	myclass obj;	2788	myclass obj;
2275	ev::io iow;	2789	ev::io iow;
2276	iow.set <myclass, &myclass::io_cb> (&obj);	2790	iow.set <myclass, &myclass::io_cb> (&obj);
2277		2791
2278	=item w->set<function> (void *data = 0)	2792	=item w->set<function> (void *data = 0)
2279		2793
2280	Also sets a callback, but uses a static method or plain function as	2794	Also sets a callback, but uses a static method or plain function as
2281	callback. The optional C<data> argument will be stored in the watcher's	2795	callback. The optional C<data> argument will be stored in the watcher's
…		…
2283		2797
2284	The prototype of the C<function> must be C<void (*)(ev::TYPE &w, int)>.	2798	The prototype of the C<function> must be C<void (*)(ev::TYPE &w, int)>.
2285		2799
2286	See the method-C<set> above for more details.	2800	See the method-C<set> above for more details.
2287		2801
2288	Example:	2802	Example: Use a plain function as callback.
2289		2803
2290	static void io_cb (ev::io &w, int revents) { }	2804	static void io_cb (ev::io &w, int revents) { }
2291	iow.set <io_cb> ();	2805	iow.set <io_cb> ();
2292		2806
2293	=item w->set (struct ev_loop *)	2807	=item w->set (struct ev_loop *)
2294		2808
2295	Associates a different C<struct ev_loop> with this watcher. You can only	2809	Associates a different C<struct ev_loop> with this watcher. You can only
2296	do this when the watcher is inactive (and not pending either).	2810	do this when the watcher is inactive (and not pending either).
2297		2811
2298	=item w->set ([args])	2812	=item w->set ([arguments])
2299		2813
2300	Basically the same as C<ev_TYPE_set>, with the same args. Must be	2814	Basically the same as C<ev_TYPE_set>, with the same arguments. Must be
2301	called at least once. Unlike the C counterpart, an active watcher gets	2815	called at least once. Unlike the C counterpart, an active watcher gets
2302	automatically stopped and restarted when reconfiguring it with this	2816	automatically stopped and restarted when reconfiguring it with this
2303	method.	2817	method.
2304		2818
2305	=item w->start ()	2819	=item w->start ()
…		…
2329	=back	2843	=back
2330		2844
2331	Example: Define a class with an IO and idle watcher, start one of them in	2845	Example: Define a class with an IO and idle watcher, start one of them in
2332	the constructor.	2846	the constructor.
2333		2847
2334	class myclass	2848	class myclass
2335	{	2849	{
2336	ev::io io; void io_cb (ev::io &w, int revents);	2850	ev::io io ; void io_cb (ev::io &w, int revents);
2337	ev:idle idle void idle_cb (ev::idle &w, int revents);	2851	ev::idle idle; void idle_cb (ev::idle &w, int revents);
2338		2852
2339	myclass (int fd)	2853	myclass (int fd)
2340	{	2854	{
2341	io .set <myclass, &myclass::io_cb > (this);	2855	io .set <myclass, &myclass::io_cb > (this);
2342	idle.set <myclass, &myclass::idle_cb> (this);	2856	idle.set <myclass, &myclass::idle_cb> (this);
2343		2857
2344	io.start (fd, ev::READ);	2858	io.start (fd, ev::READ);
2345	}	2859	}
2346	};	2860	};
		2861
		2862
		2863	=head1 OTHER LANGUAGE BINDINGS
		2864
		2865	Libev does not offer other language bindings itself, but bindings for a
		2866	number of languages exist in the form of third-party packages. If you know
		2867	any interesting language binding in addition to the ones listed here, drop
		2868	me a note.
		2869
		2870	=over 4
		2871
		2872	=item Perl
		2873
		2874	The EV module implements the full libev API and is actually used to test
		2875	libev. EV is developed together with libev. Apart from the EV core module,
		2876	there are additional modules that implement libev-compatible interfaces
		2877	to C<libadns> (C<EV::ADNS>, but C<AnyEvent::DNS> is preferred nowadays),
		2878	C<Net::SNMP> (C<Net::SNMP::EV>) and the C<libglib> event core (C<Glib::EV>
		2879	and C<EV::Glib>).
		2880
		2881	It can be found and installed via CPAN, its homepage is at
		2882	L<http://software.schmorp.de/pkg/EV>.
		2883
		2884	=item Python
		2885
		2886	Python bindings can be found at L<http://code.google.com/p/pyev/>. It
		2887	seems to be quite complete and well-documented. Note, however, that the
		2888	patch they require for libev is outright dangerous as it breaks the ABI
		2889	for everybody else, and therefore, should never be applied in an installed
		2890	libev (if python requires an incompatible ABI then it needs to embed
		2891	libev).
		2892
		2893	=item Ruby
		2894
		2895	Tony Arcieri has written a ruby extension that offers access to a subset
		2896	of the libev API and adds file handle abstractions, asynchronous DNS and
		2897	more on top of it. It can be found via gem servers. Its homepage is at
		2898	L<http://rev.rubyforge.org/>.
		2899
		2900	=item D
		2901
		2902	Leandro Lucarella has written a D language binding (F<ev.d>) for libev, to
		2903	be found at L<http://proj.llucax.com.ar/wiki/evd>.
		2904
		2905	=back
2347		2906
2348		2907
2349	=head1 MACRO MAGIC	2908	=head1 MACRO MAGIC
2350		2909
2351	Libev can be compiled with a variety of options, the most fundamantal	2910	Libev can be compiled with a variety of options, the most fundamental
2352	of which is C<EV_MULTIPLICITY>. This option determines whether (most)	2911	of which is C<EV_MULTIPLICITY>. This option determines whether (most)
2353	functions and callbacks have an initial C<struct ev_loop *> argument.	2912	functions and callbacks have an initial C<struct ev_loop *> argument.
2354		2913
2355	To make it easier to write programs that cope with either variant, the	2914	To make it easier to write programs that cope with either variant, the
2356	following macros are defined:	2915	following macros are defined:
…		…
2361		2920
2362	This provides the loop I<argument> for functions, if one is required ("ev	2921	This provides the loop I<argument> for functions, if one is required ("ev
2363	loop argument"). The C<EV_A> form is used when this is the sole argument,	2922	loop argument"). The C<EV_A> form is used when this is the sole argument,
2364	C<EV_A_> is used when other arguments are following. Example:	2923	C<EV_A_> is used when other arguments are following. Example:
2365		2924
2366	ev_unref (EV_A);	2925	ev_unref (EV_A);
2367	ev_timer_add (EV_A_ watcher);	2926	ev_timer_add (EV_A_ watcher);
2368	ev_loop (EV_A_ 0);	2927	ev_loop (EV_A_ 0);
2369		2928
2370	It assumes the variable C<loop> of type C<struct ev_loop *> is in scope,	2929	It assumes the variable C<loop> of type C<struct ev_loop *> is in scope,
2371	which is often provided by the following macro.	2930	which is often provided by the following macro.
2372		2931
2373	=item C<EV_P>, C<EV_P_>	2932	=item C<EV_P>, C<EV_P_>
2374		2933
2375	This provides the loop I<parameter> for functions, if one is required ("ev	2934	This provides the loop I<parameter> for functions, if one is required ("ev
2376	loop parameter"). The C<EV_P> form is used when this is the sole parameter,	2935	loop parameter"). The C<EV_P> form is used when this is the sole parameter,
2377	C<EV_P_> is used when other parameters are following. Example:	2936	C<EV_P_> is used when other parameters are following. Example:
2378		2937
2379	// this is how ev_unref is being declared	2938	// this is how ev_unref is being declared
2380	static void ev_unref (EV_P);	2939	static void ev_unref (EV_P);
2381		2940
2382	// this is how you can declare your typical callback	2941	// this is how you can declare your typical callback
2383	static void cb (EV_P_ ev_timer *w, int revents)	2942	static void cb (EV_P_ ev_timer *w, int revents)
2384		2943
2385	It declares a parameter C<loop> of type C<struct ev_loop *>, quite	2944	It declares a parameter C<loop> of type C<struct ev_loop *>, quite
2386	suitable for use with C<EV_A>.	2945	suitable for use with C<EV_A>.
2387		2946
2388	=item C<EV_DEFAULT>, C<EV_DEFAULT_>	2947	=item C<EV_DEFAULT>, C<EV_DEFAULT_>
2389		2948
2390	Similar to the other two macros, this gives you the value of the default	2949	Similar to the other two macros, this gives you the value of the default
2391	loop, if multiple loops are supported ("ev loop default").	2950	loop, if multiple loops are supported ("ev loop default").
		2951
		2952	=item C<EV_DEFAULT_UC>, C<EV_DEFAULT_UC_>
		2953
		2954	Usage identical to C<EV_DEFAULT> and C<EV_DEFAULT_>, but requires that the
		2955	default loop has been initialised (C<UC> == unchecked). Their behaviour
		2956	is undefined when the default loop has not been initialised by a previous
		2957	execution of C<EV_DEFAULT>, C<EV_DEFAULT_> or C<ev_default_init (...)>.
		2958
		2959	It is often prudent to use C<EV_DEFAULT> when initialising the first
		2960	watcher in a function but use C<EV_DEFAULT_UC> afterwards.
2392		2961
2393	=back	2962	=back
2394		2963
2395	Example: Declare and initialise a check watcher, utilising the above	2964	Example: Declare and initialise a check watcher, utilising the above
2396	macros so it will work regardless of whether multiple loops are supported	2965	macros so it will work regardless of whether multiple loops are supported
2397	or not.	2966	or not.
2398		2967
2399	static void	2968	static void
2400	check_cb (EV_P_ ev_timer *w, int revents)	2969	check_cb (EV_P_ ev_timer *w, int revents)
2401	{	2970	{
2402	ev_check_stop (EV_A_ w);	2971	ev_check_stop (EV_A_ w);
2403	}	2972	}
2404		2973
2405	ev_check check;	2974	ev_check check;
2406	ev_check_init (&check, check_cb);	2975	ev_check_init (&check, check_cb);
2407	ev_check_start (EV_DEFAULT_ &check);	2976	ev_check_start (EV_DEFAULT_ &check);
2408	ev_loop (EV_DEFAULT_ 0);	2977	ev_loop (EV_DEFAULT_ 0);
2409		2978
2410	=head1 EMBEDDING	2979	=head1 EMBEDDING
2411		2980
2412	Libev can (and often is) directly embedded into host	2981	Libev can (and often is) directly embedded into host
2413	applications. Examples of applications that embed it include the Deliantra	2982	applications. Examples of applications that embed it include the Deliantra
…		…
2420	libev somewhere in your source tree).	2989	libev somewhere in your source tree).
2421		2990
2422	=head2 FILESETS	2991	=head2 FILESETS
2423		2992
2424	Depending on what features you need you need to include one or more sets of files	2993	Depending on what features you need you need to include one or more sets of files
2425	in your app.	2994	in your application.
2426		2995
2427	=head3 CORE EVENT LOOP	2996	=head3 CORE EVENT LOOP
2428		2997
2429	To include only the libev core (all the C<ev_*> functions), with manual	2998	To include only the libev core (all the C<ev_*> functions), with manual
2430	configuration (no autoconf):	2999	configuration (no autoconf):
2431		3000
2432	#define EV_STANDALONE 1	3001	#define EV_STANDALONE 1
2433	#include "ev.c"	3002	#include "ev.c"
2434		3003
2435	This will automatically include F<ev.h>, too, and should be done in a	3004	This will automatically include F<ev.h>, too, and should be done in a
2436	single C source file only to provide the function implementations. To use	3005	single C source file only to provide the function implementations. To use
2437	it, do the same for F<ev.h> in all files wishing to use this API (best	3006	it, do the same for F<ev.h> in all files wishing to use this API (best
2438	done by writing a wrapper around F<ev.h> that you can include instead and	3007	done by writing a wrapper around F<ev.h> that you can include instead and
2439	where you can put other configuration options):	3008	where you can put other configuration options):
2440		3009
2441	#define EV_STANDALONE 1	3010	#define EV_STANDALONE 1
2442	#include "ev.h"	3011	#include "ev.h"
2443		3012
2444	Both header files and implementation files can be compiled with a C++	3013	Both header files and implementation files can be compiled with a C++
2445	compiler (at least, thats a stated goal, and breakage will be treated	3014	compiler (at least, thats a stated goal, and breakage will be treated
2446	as a bug).	3015	as a bug).
2447		3016
2448	You need the following files in your source tree, or in a directory	3017	You need the following files in your source tree, or in a directory
2449	in your include path (e.g. in libev/ when using -Ilibev):	3018	in your include path (e.g. in libev/ when using -Ilibev):
2450		3019
2451	ev.h	3020	ev.h
2452	ev.c	3021	ev.c
2453	ev_vars.h	3022	ev_vars.h
2454	ev_wrap.h	3023	ev_wrap.h
2455		3024
2456	ev_win32.c required on win32 platforms only	3025	ev_win32.c required on win32 platforms only
2457		3026
2458	ev_select.c only when select backend is enabled (which is enabled by default)	3027	ev_select.c only when select backend is enabled (which is enabled by default)
2459	ev_poll.c only when poll backend is enabled (disabled by default)	3028	ev_poll.c only when poll backend is enabled (disabled by default)
2460	ev_epoll.c only when the epoll backend is enabled (disabled by default)	3029	ev_epoll.c only when the epoll backend is enabled (disabled by default)
2461	ev_kqueue.c only when the kqueue backend is enabled (disabled by default)	3030	ev_kqueue.c only when the kqueue backend is enabled (disabled by default)
2462	ev_port.c only when the solaris port backend is enabled (disabled by default)	3031	ev_port.c only when the solaris port backend is enabled (disabled by default)
2463		3032
2464	F<ev.c> includes the backend files directly when enabled, so you only need	3033	F<ev.c> includes the backend files directly when enabled, so you only need
2465	to compile this single file.	3034	to compile this single file.
2466		3035
2467	=head3 LIBEVENT COMPATIBILITY API	3036	=head3 LIBEVENT COMPATIBILITY API
2468		3037
2469	To include the libevent compatibility API, also include:	3038	To include the libevent compatibility API, also include:
2470		3039
2471	#include "event.c"	3040	#include "event.c"
2472		3041
2473	in the file including F<ev.c>, and:	3042	in the file including F<ev.c>, and:
2474		3043
2475	#include "event.h"	3044	#include "event.h"
2476		3045
2477	in the files that want to use the libevent API. This also includes F<ev.h>.	3046	in the files that want to use the libevent API. This also includes F<ev.h>.
2478		3047
2479	You need the following additional files for this:	3048	You need the following additional files for this:
2480		3049
2481	event.h	3050	event.h
2482	event.c	3051	event.c
2483		3052
2484	=head3 AUTOCONF SUPPORT	3053	=head3 AUTOCONF SUPPORT
2485		3054
2486	Instead of using C<EV_STANDALONE=1> and providing your config in	3055	Instead of using C<EV_STANDALONE=1> and providing your configuration in
2487	whatever way you want, you can also C<m4_include([libev.m4])> in your	3056	whatever way you want, you can also C<m4_include([libev.m4])> in your
2488	F<configure.ac> and leave C<EV_STANDALONE> undefined. F<ev.c> will then	3057	F<configure.ac> and leave C<EV_STANDALONE> undefined. F<ev.c> will then
2489	include F<config.h> and configure itself accordingly.	3058	include F<config.h> and configure itself accordingly.
2490		3059
2491	For this of course you need the m4 file:	3060	For this of course you need the m4 file:
2492		3061
2493	libev.m4	3062	libev.m4
2494		3063
2495	=head2 PREPROCESSOR SYMBOLS/MACROS	3064	=head2 PREPROCESSOR SYMBOLS/MACROS
2496		3065
2497	Libev can be configured via a variety of preprocessor symbols you have to define	3066	Libev can be configured via a variety of preprocessor symbols you have to
2498	before including any of its files. The default is not to build for multiplicity	3067	define before including any of its files. The default in the absence of
2499	and only include the select backend.	3068	autoconf is documented for every option.
2500		3069
2501	=over 4	3070	=over 4
2502		3071
2503	=item EV_STANDALONE	3072	=item EV_STANDALONE
2504		3073
…		…
2509	F<event.h> that are not directly supported by the libev core alone.	3078	F<event.h> that are not directly supported by the libev core alone.
2510		3079
2511	=item EV_USE_MONOTONIC	3080	=item EV_USE_MONOTONIC
2512		3081
2513	If defined to be C<1>, libev will try to detect the availability of the	3082	If defined to be C<1>, libev will try to detect the availability of the
2514	monotonic clock option at both compiletime and runtime. Otherwise no use	3083	monotonic clock option at both compile time and runtime. Otherwise no use
2515	of the monotonic clock option will be attempted. If you enable this, you	3084	of the monotonic clock option will be attempted. If you enable this, you
2516	usually have to link against librt or something similar. Enabling it when	3085	usually have to link against librt or something similar. Enabling it when
2517	the functionality isn't available is safe, though, although you have	3086	the functionality isn't available is safe, though, although you have
2518	to make sure you link against any libraries where the C<clock_gettime>	3087	to make sure you link against any libraries where the C<clock_gettime>
2519	function is hiding in (often F<-lrt>).	3088	function is hiding in (often F<-lrt>).
2520		3089
2521	=item EV_USE_REALTIME	3090	=item EV_USE_REALTIME
2522		3091
2523	If defined to be C<1>, libev will try to detect the availability of the	3092	If defined to be C<1>, libev will try to detect the availability of the
2524	realtime clock option at compiletime (and assume its availability at	3093	real-time clock option at compile time (and assume its availability at
2525	runtime if successful). Otherwise no use of the realtime clock option will	3094	runtime if successful). Otherwise no use of the real-time clock option will
2526	be attempted. This effectively replaces C<gettimeofday> by C<clock_get	3095	be attempted. This effectively replaces C<gettimeofday> by C<clock_get
2527	(CLOCK_REALTIME, ...)> and will not normally affect correctness. See the	3096	(CLOCK_REALTIME, ...)> and will not normally affect correctness. See the
2528	note about libraries in the description of C<EV_USE_MONOTONIC>, though.	3097	note about libraries in the description of C<EV_USE_MONOTONIC>, though.
2529		3098
2530	=item EV_USE_NANOSLEEP	3099	=item EV_USE_NANOSLEEP
2531		3100
2532	If defined to be C<1>, libev will assume that C<nanosleep ()> is available	3101	If defined to be C<1>, libev will assume that C<nanosleep ()> is available
2533	and will use it for delays. Otherwise it will use C<select ()>.	3102	and will use it for delays. Otherwise it will use C<select ()>.
2534		3103
		3104	=item EV_USE_EVENTFD
		3105
		3106	If defined to be C<1>, then libev will assume that C<eventfd ()> is
		3107	available and will probe for kernel support at runtime. This will improve
		3108	C<ev_signal> and C<ev_async> performance and reduce resource consumption.
		3109	If undefined, it will be enabled if the headers indicate GNU/Linux + Glibc
		3110	2.7 or newer, otherwise disabled.
		3111
2535	=item EV_USE_SELECT	3112	=item EV_USE_SELECT
2536		3113
2537	If undefined or defined to be C<1>, libev will compile in support for the	3114	If undefined or defined to be C<1>, libev will compile in support for the
2538	C<select>(2) backend. No attempt at autodetection will be done: if no	3115	C<select>(2) backend. No attempt at auto-detection will be done: if no
2539	other method takes over, select will be it. Otherwise the select backend	3116	other method takes over, select will be it. Otherwise the select backend
2540	will not be compiled in.	3117	will not be compiled in.
2541		3118
2542	=item EV_SELECT_USE_FD_SET	3119	=item EV_SELECT_USE_FD_SET
2543		3120
2544	If defined to C<1>, then the select backend will use the system C<fd_set>	3121	If defined to C<1>, then the select backend will use the system C<fd_set>
2545	structure. This is useful if libev doesn't compile due to a missing	3122	structure. This is useful if libev doesn't compile due to a missing
2546	C<NFDBITS> or C<fd_mask> definition or it misguesses the bitset layout on	3123	C<NFDBITS> or C<fd_mask> definition or it mis-guesses the bitset layout on
2547	exotic systems. This usually limits the range of file descriptors to some	3124	exotic systems. This usually limits the range of file descriptors to some
2548	low limit such as 1024 or might have other limitations (winsocket only	3125	low limit such as 1024 or might have other limitations (winsocket only
2549	allows 64 sockets). The C<FD_SETSIZE> macro, set before compilation, might	3126	allows 64 sockets). The C<FD_SETSIZE> macro, set before compilation, might
2550	influence the size of the C<fd_set> used.	3127	influence the size of the C<fd_set> used.
2551		3128
…		…
2575		3152
2576	=item EV_USE_EPOLL	3153	=item EV_USE_EPOLL
2577		3154
2578	If defined to be C<1>, libev will compile in support for the Linux	3155	If defined to be C<1>, libev will compile in support for the Linux
2579	C<epoll>(7) backend. Its availability will be detected at runtime,	3156	C<epoll>(7) backend. Its availability will be detected at runtime,
2580	otherwise another method will be used as fallback. This is the	3157	otherwise another method will be used as fallback. This is the preferred
2581	preferred backend for GNU/Linux systems.	3158	backend for GNU/Linux systems. If undefined, it will be enabled if the
		3159	headers indicate GNU/Linux + Glibc 2.4 or newer, otherwise disabled.
2582		3160
2583	=item EV_USE_KQUEUE	3161	=item EV_USE_KQUEUE
2584		3162
2585	If defined to be C<1>, libev will compile in support for the BSD style	3163	If defined to be C<1>, libev will compile in support for the BSD style
2586	C<kqueue>(2) backend. Its actual availability will be detected at runtime,	3164	C<kqueue>(2) backend. Its actual availability will be detected at runtime,
…		…
2599	otherwise another method will be used as fallback. This is the preferred	3177	otherwise another method will be used as fallback. This is the preferred
2600	backend for Solaris 10 systems.	3178	backend for Solaris 10 systems.
2601		3179
2602	=item EV_USE_DEVPOLL	3180	=item EV_USE_DEVPOLL
2603		3181
2604	reserved for future expansion, works like the USE symbols above.	3182	Reserved for future expansion, works like the USE symbols above.
2605		3183
2606	=item EV_USE_INOTIFY	3184	=item EV_USE_INOTIFY
2607		3185
2608	If defined to be C<1>, libev will compile in support for the Linux inotify	3186	If defined to be C<1>, libev will compile in support for the Linux inotify
2609	interface to speed up C<ev_stat> watchers. Its actual availability will	3187	interface to speed up C<ev_stat> watchers. Its actual availability will
2610	be detected at runtime.	3188	be detected at runtime. If undefined, it will be enabled if the headers
		3189	indicate GNU/Linux + Glibc 2.4 or newer, otherwise disabled.
		3190
		3191	=item EV_ATOMIC_T
		3192
		3193	Libev requires an integer type (suitable for storing C<0> or C<1>) whose
		3194	access is atomic with respect to other threads or signal contexts. No such
		3195	type is easily found in the C language, so you can provide your own type
		3196	that you know is safe for your purposes. It is used both for signal handler "locking"
		3197	as well as for signal and thread safety in C<ev_async> watchers.
		3198
		3199	In the absence of this define, libev will use C<sig_atomic_t volatile>
		3200	(from F<signal.h>), which is usually good enough on most platforms.
2611		3201
2612	=item EV_H	3202	=item EV_H
2613		3203
2614	The name of the F<ev.h> header file used to include it. The default if	3204	The name of the F<ev.h> header file used to include it. The default if
2615	undefined is C<"ev.h"> in F<event.h>, F<ev.c> and F<ev++.h>. This can be	3205	undefined is C<"ev.h"> in F<event.h>, F<ev.c> and F<ev++.h>. This can be
…		…
2653	When doing priority-based operations, libev usually has to linearly search	3243	When doing priority-based operations, libev usually has to linearly search
2654	all the priorities, so having many of them (hundreds) uses a lot of space	3244	all the priorities, so having many of them (hundreds) uses a lot of space
2655	and time, so using the defaults of five priorities (-2 .. +2) is usually	3245	and time, so using the defaults of five priorities (-2 .. +2) is usually
2656	fine.	3246	fine.
2657		3247
2658	If your embedding app does not need any priorities, defining these both to	3248	If your embedding application does not need any priorities, defining these
2659	C<0> will save some memory and cpu.	3249	both to C<0> will save some memory and CPU.
2660		3250
2661	=item EV_PERIODIC_ENABLE	3251	=item EV_PERIODIC_ENABLE
2662		3252
2663	If undefined or defined to be C<1>, then periodic timers are supported. If	3253	If undefined or defined to be C<1>, then periodic timers are supported. If
2664	defined to be C<0>, then they are not. Disabling them saves a few kB of	3254	defined to be C<0>, then they are not. Disabling them saves a few kB of
…		…
2671	code.	3261	code.
2672		3262
2673	=item EV_EMBED_ENABLE	3263	=item EV_EMBED_ENABLE
2674		3264
2675	If undefined or defined to be C<1>, then embed watchers are supported. If	3265	If undefined or defined to be C<1>, then embed watchers are supported. If
2676	defined to be C<0>, then they are not.	3266	defined to be C<0>, then they are not. Embed watchers rely on most other
		3267	watcher types, which therefore must not be disabled.
2677		3268
2678	=item EV_STAT_ENABLE	3269	=item EV_STAT_ENABLE
2679		3270
2680	If undefined or defined to be C<1>, then stat watchers are supported. If	3271	If undefined or defined to be C<1>, then stat watchers are supported. If
2681	defined to be C<0>, then they are not.	3272	defined to be C<0>, then they are not.
…		…
2683	=item EV_FORK_ENABLE	3274	=item EV_FORK_ENABLE
2684		3275
2685	If undefined or defined to be C<1>, then fork watchers are supported. If	3276	If undefined or defined to be C<1>, then fork watchers are supported. If
2686	defined to be C<0>, then they are not.	3277	defined to be C<0>, then they are not.
2687		3278
		3279	=item EV_ASYNC_ENABLE
		3280
		3281	If undefined or defined to be C<1>, then async watchers are supported. If
		3282	defined to be C<0>, then they are not.
		3283
2688	=item EV_MINIMAL	3284	=item EV_MINIMAL
2689		3285
2690	If you need to shave off some kilobytes of code at the expense of some	3286	If you need to shave off some kilobytes of code at the expense of some
2691	speed, define this symbol to C<1>. Currently only used for gcc to override	3287	speed, define this symbol to C<1>. Currently this is used to override some
2692	some inlining decisions, saves roughly 30% codesize of amd64.	3288	inlining decisions, saves roughly 30% code size on amd64. It also selects a
		3289	much smaller 2-heap for timer management over the default 4-heap.
2693		3290
2694	=item EV_PID_HASHSIZE	3291	=item EV_PID_HASHSIZE
2695		3292
2696	C<ev_child> watchers use a small hash table to distribute workload by	3293	C<ev_child> watchers use a small hash table to distribute workload by
2697	pid. The default size is C<16> (or C<1> with C<EV_MINIMAL>), usually more	3294	pid. The default size is C<16> (or C<1> with C<EV_MINIMAL>), usually more
…		…
2704	inotify watch id. The default size is C<16> (or C<1> with C<EV_MINIMAL>),	3301	inotify watch id. The default size is C<16> (or C<1> with C<EV_MINIMAL>),
2705	usually more than enough. If you need to manage thousands of C<ev_stat>	3302	usually more than enough. If you need to manage thousands of C<ev_stat>
2706	watchers you might want to increase this value (I<must> be a power of	3303	watchers you might want to increase this value (I<must> be a power of
2707	two).	3304	two).
2708		3305
		3306	=item EV_USE_4HEAP
		3307
		3308	Heaps are not very cache-efficient. To improve the cache-efficiency of the
		3309	timer and periodics heaps, libev uses a 4-heap when this symbol is defined
		3310	to C<1>. The 4-heap uses more complicated (longer) code but has noticeably
		3311	faster performance with many (thousands) of watchers.
		3312
		3313	The default is C<1> unless C<EV_MINIMAL> is set in which case it is C<0>
		3314	(disabled).
		3315
		3316	=item EV_HEAP_CACHE_AT
		3317
		3318	Heaps are not very cache-efficient. To improve the cache-efficiency of the
		3319	timer and periodics heaps, libev can cache the timestamp (I<at>) within
		3320	the heap structure (selected by defining C<EV_HEAP_CACHE_AT> to C<1>),
		3321	which uses 8-12 bytes more per watcher and a few hundred bytes more code,
		3322	but avoids random read accesses on heap changes. This improves performance
		3323	noticeably with many (hundreds) of watchers.
		3324
		3325	The default is C<1> unless C<EV_MINIMAL> is set in which case it is C<0>
		3326	(disabled).
		3327
		3328	=item EV_VERIFY
		3329
		3330	Controls how much internal verification (see C<ev_loop_verify ()>) will
		3331	be done: If set to C<0>, no internal verification code will be compiled
		3332	in. If set to C<1>, then verification code will be compiled in, but not
		3333	called. If set to C<2>, then the internal verification code will be
		3334	called once per loop, which can slow down libev. If set to C<3>, then the
		3335	verification code will be called very frequently, which will slow down
		3336	libev considerably.
		3337
		3338	The default is C<1>, unless C<EV_MINIMAL> is set, in which case it will be
		3339	C<0>.
		3340
2709	=item EV_COMMON	3341	=item EV_COMMON
2710		3342
2711	By default, all watchers have a C<void *data> member. By redefining	3343	By default, all watchers have a C<void *data> member. By redefining
2712	this macro to a something else you can include more and other types of	3344	this macro to a something else you can include more and other types of
2713	members. You have to define it each time you include one of the files,	3345	members. You have to define it each time you include one of the files,
2714	though, and it must be identical each time.	3346	though, and it must be identical each time.
2715		3347
2716	For example, the perl EV module uses something like this:	3348	For example, the perl EV module uses something like this:
2717		3349
2718	#define EV_COMMON \	3350	#define EV_COMMON \
2719	SV *self; /* contains this struct */ \	3351	SV *self; /* contains this struct */ \
2720	SV *cb_sv, *fh /* note no trailing ";" */	3352	SV *cb_sv, *fh /* note no trailing ";" */
2721		3353
2722	=item EV_CB_DECLARE (type)	3354	=item EV_CB_DECLARE (type)
2723		3355
2724	=item EV_CB_INVOKE (watcher, revents)	3356	=item EV_CB_INVOKE (watcher, revents)
2725		3357
…		…
2730	definition and a statement, respectively. See the F<ev.h> header file for	3362	definition and a statement, respectively. See the F<ev.h> header file for
2731	their default definitions. One possible use for overriding these is to	3363	their default definitions. One possible use for overriding these is to
2732	avoid the C<struct ev_loop *> as first argument in all cases, or to use	3364	avoid the C<struct ev_loop *> as first argument in all cases, or to use
2733	method calls instead of plain function calls in C++.	3365	method calls instead of plain function calls in C++.
2734		3366
		3367	=back
		3368
2735	=head2 EXPORTED API SYMBOLS	3369	=head2 EXPORTED API SYMBOLS
2736		3370
2737	If you need to re-export the API (e.g. via a dll) and you need a list of	3371	If you need to re-export the API (e.g. via a DLL) and you need a list of
2738	exported symbols, you can use the provided F<Symbol.*> files which list	3372	exported symbols, you can use the provided F<Symbol.*> files which list
2739	all public symbols, one per line:	3373	all public symbols, one per line:
2740		3374
2741	Symbols.ev for libev proper	3375	Symbols.ev for libev proper
2742	Symbols.event for the libevent emulation	3376	Symbols.event for the libevent emulation
2743		3377
2744	This can also be used to rename all public symbols to avoid clashes with	3378	This can also be used to rename all public symbols to avoid clashes with
2745	multiple versions of libev linked together (which is obviously bad in	3379	multiple versions of libev linked together (which is obviously bad in
2746	itself, but sometimes it is inconvinient to avoid this).	3380	itself, but sometimes it is inconvenient to avoid this).
2747		3381
2748	A sed command like this will create wrapper C<#define>'s that you need to	3382	A sed command like this will create wrapper C<#define>'s that you need to
2749	include before including F<ev.h>:	3383	include before including F<ev.h>:
2750		3384
2751	<Symbols.ev sed -e "s/.*/#define & myprefix_&/" >wrap.h	3385	<Symbols.ev sed -e "s/.*/#define & myprefix_&/" >wrap.h
…		…
2768	file.	3402	file.
2769		3403
2770	The usage in rxvt-unicode is simpler. It has a F<ev_cpp.h> header file	3404	The usage in rxvt-unicode is simpler. It has a F<ev_cpp.h> header file
2771	that everybody includes and which overrides some configure choices:	3405	that everybody includes and which overrides some configure choices:
2772		3406
2773	#define EV_MINIMAL 1	3407	#define EV_MINIMAL 1
2774	#define EV_USE_POLL 0	3408	#define EV_USE_POLL 0
2775	#define EV_MULTIPLICITY 0	3409	#define EV_MULTIPLICITY 0
2776	#define EV_PERIODIC_ENABLE 0	3410	#define EV_PERIODIC_ENABLE 0
2777	#define EV_STAT_ENABLE 0	3411	#define EV_STAT_ENABLE 0
2778	#define EV_FORK_ENABLE 0	3412	#define EV_FORK_ENABLE 0
2779	#define EV_CONFIG_H <config.h>	3413	#define EV_CONFIG_H <config.h>
2780	#define EV_MINPRI 0	3414	#define EV_MINPRI 0
2781	#define EV_MAXPRI 0	3415	#define EV_MAXPRI 0
2782		3416
2783	#include "ev++.h"	3417	#include "ev++.h"
2784		3418
2785	And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:	3419	And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:
2786		3420
2787	#include "ev_cpp.h"	3421	#include "ev_cpp.h"
2788	#include "ev.c"	3422	#include "ev.c"
2789		3423
		3424	=head1 INTERACTION WITH OTHER PROGRAMS OR LIBRARIES
2790		3425
2791	=head1 COMPLEXITIES	3426	=head2 THREADS AND COROUTINES
2792		3427
2793	In this section the complexities of (many of) the algorithms used inside	3428	=head3 THREADS
2794	libev will be explained. For complexity discussions about backends see the
2795	documentation for C<ev_default_init>.
2796		3429
2797	All of the following are about amortised time: If an array needs to be	3430	All libev functions are reentrant and thread-safe unless explicitly
2798	extended, libev needs to realloc and move the whole array, but this	3431	documented otherwise, but libev implements no locking itself. This means
2799	happens asymptotically never with higher number of elements, so O(1) might	3432	that you can use as many loops as you want in parallel, as long as there
2800	mean it might do a lengthy realloc operation in rare cases, but on average	3433	are no concurrent calls into any libev function with the same loop
2801	it is much faster and asymptotically approaches constant time.	3434	parameter (C<ev_default_*> calls have an implicit default loop parameter,
		3435	of course): libev guarantees that different event loops share no data
		3436	structures that need any locking.
		3437
		3438	Or to put it differently: calls with different loop parameters can be done
		3439	concurrently from multiple threads, calls with the same loop parameter
		3440	must be done serially (but can be done from different threads, as long as
		3441	only one thread ever is inside a call at any point in time, e.g. by using
		3442	a mutex per loop).
		3443
		3444	Specifically to support threads (and signal handlers), libev implements
		3445	so-called C<ev_async> watchers, which allow some limited form of
		3446	concurrency on the same event loop, namely waking it up "from the
		3447	outside".
		3448
		3449	If you want to know which design (one loop, locking, or multiple loops
		3450	without or something else still) is best for your problem, then I cannot
		3451	help you, but here is some generic advice:
2802		3452
2803	=over 4	3453	=over 4
2804		3454
2805	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)	3455	=item * most applications have a main thread: use the default libev loop
		3456	in that thread, or create a separate thread running only the default loop.
2806		3457
2807	This means that, when you have a watcher that triggers in one hour and	3458	This helps integrating other libraries or software modules that use libev
2808	there are 100 watchers that would trigger before that then inserting will	3459	themselves and don't care/know about threading.
2809	have to skip roughly seven (C<ld 100>) of these watchers.
2810		3460
2811	=item Changing timer/periodic watchers (by autorepeat or calling again): O(log skipped_other_timers)	3461	=item * one loop per thread is usually a good model.
2812		3462
2813	That means that changing a timer costs less than removing/adding them	3463	Doing this is almost never wrong, sometimes a better-performance model
2814	as only the relative motion in the event queue has to be paid for.	3464	exists, but it is always a good start.
2815		3465
2816	=item Starting io/check/prepare/idle/signal/child watchers: O(1)	3466	=item * other models exist, such as the leader/follower pattern, where one
		3467	loop is handed through multiple threads in a kind of round-robin fashion.
2817		3468
2818	These just add the watcher into an array or at the head of a list.	3469	Choosing a model is hard - look around, learn, know that usually you can do
		3470	better than you currently do :-)
2819		3471
2820	=item Stopping check/prepare/idle watchers: O(1)	3472	=item * often you need to talk to some other thread which blocks in the
		3473	event loop.
2821		3474
2822	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))	3475	C<ev_async> watchers can be used to wake them up from other threads safely
		3476	(or from signal contexts...).
2823		3477
2824	These watchers are stored in lists then need to be walked to find the	3478	An example use would be to communicate signals or other events that only
2825	correct watcher to remove. The lists are usually short (you don't usually	3479	work in the default loop by registering the signal watcher with the
2826	have many watchers waiting for the same fd or signal).	3480	default loop and triggering an C<ev_async> watcher from the default loop
2827		3481	watcher callback into the event loop interested in the signal.
2828	=item Finding the next timer in each loop iteration: O(1)
2829
2830	By virtue of using a binary heap, the next timer is always found at the
2831	beginning of the storage array.
2832
2833	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)
2834
2835	A change means an I/O watcher gets started or stopped, which requires
2836	libev to recalculate its status (and possibly tell the kernel, depending
2837	on backend and wether C<ev_io_set> was used).
2838
2839	=item Activating one watcher (putting it into the pending state): O(1)
2840
2841	=item Priority handling: O(number_of_priorities)
2842
2843	Priorities are implemented by allocating some space for each
2844	priority. When doing priority-based operations, libev usually has to
2845	linearly search all the priorities, but starting/stopping and activating
2846	watchers becomes O(1) w.r.t. prioritiy handling.
2847		3482
2848	=back	3483	=back
2849		3484
		3485	=head3 COROUTINES
2850		3486
2851	=head1 Win32 platform limitations and workarounds	3487	Libev is very accommodating to coroutines ("cooperative threads"):
		3488	libev fully supports nesting calls to its functions from different
		3489	coroutines (e.g. you can call C<ev_loop> on the same loop from two
		3490	different coroutines, and switch freely between both coroutines running the
		3491	loop, as long as you don't confuse yourself). The only exception is that
		3492	you must not do this from C<ev_periodic> reschedule callbacks.
		3493
		3494	Care has been taken to ensure that libev does not keep local state inside
		3495	C<ev_loop>, and other calls do not usually allow for coroutine switches as
		3496	they do not clal any callbacks.
		3497
		3498	=head2 COMPILER WARNINGS
		3499
		3500	Depending on your compiler and compiler settings, you might get no or a
		3501	lot of warnings when compiling libev code. Some people are apparently
		3502	scared by this.
		3503
		3504	However, these are unavoidable for many reasons. For one, each compiler
		3505	has different warnings, and each user has different tastes regarding
		3506	warning options. "Warn-free" code therefore cannot be a goal except when
		3507	targeting a specific compiler and compiler-version.
		3508
		3509	Another reason is that some compiler warnings require elaborate
		3510	workarounds, or other changes to the code that make it less clear and less
		3511	maintainable.
		3512
		3513	And of course, some compiler warnings are just plain stupid, or simply
		3514	wrong (because they don't actually warn about the condition their message
		3515	seems to warn about). For example, certain older gcc versions had some
		3516	warnings that resulted an extreme number of false positives. These have
		3517	been fixed, but some people still insist on making code warn-free with
		3518	such buggy versions.
		3519
		3520	While libev is written to generate as few warnings as possible,
		3521	"warn-free" code is not a goal, and it is recommended not to build libev
		3522	with any compiler warnings enabled unless you are prepared to cope with
		3523	them (e.g. by ignoring them). Remember that warnings are just that:
		3524	warnings, not errors, or proof of bugs.
		3525
		3526
		3527	=head2 VALGRIND
		3528
		3529	Valgrind has a special section here because it is a popular tool that is
		3530	highly useful. Unfortunately, valgrind reports are very hard to interpret.
		3531
		3532	If you think you found a bug (memory leak, uninitialised data access etc.)
		3533	in libev, then check twice: If valgrind reports something like:
		3534
		3535	==2274== definitely lost: 0 bytes in 0 blocks.
		3536	==2274== possibly lost: 0 bytes in 0 blocks.
		3537	==2274== still reachable: 256 bytes in 1 blocks.
		3538
		3539	Then there is no memory leak, just as memory accounted to global variables
		3540	is not a memleak - the memory is still being refernced, and didn't leak.
		3541
		3542	Similarly, under some circumstances, valgrind might report kernel bugs
		3543	as if it were a bug in libev (e.g. in realloc or in the poll backend,
		3544	although an acceptable workaround has been found here), or it might be
		3545	confused.
		3546
		3547	Keep in mind that valgrind is a very good tool, but only a tool. Don't
		3548	make it into some kind of religion.
		3549
		3550	If you are unsure about something, feel free to contact the mailing list
		3551	with the full valgrind report and an explanation on why you think this
		3552	is a bug in libev (best check the archives, too :). However, don't be
		3553	annoyed when you get a brisk "this is no bug" answer and take the chance
		3554	of learning how to interpret valgrind properly.
		3555
		3556	If you need, for some reason, empty reports from valgrind for your project
		3557	I suggest using suppression lists.
		3558
		3559
		3560	=head1 PORTABILITY NOTES
		3561
		3562	=head2 WIN32 PLATFORM LIMITATIONS AND WORKAROUNDS
2852		3563
2853	Win32 doesn't support any of the standards (e.g. POSIX) that libev	3564	Win32 doesn't support any of the standards (e.g. POSIX) that libev
2854	requires, and its I/O model is fundamentally incompatible with the POSIX	3565	requires, and its I/O model is fundamentally incompatible with the POSIX
2855	model. Libev still offers limited functionality on this platform in	3566	model. Libev still offers limited functionality on this platform in
2856	the form of the C<EVBACKEND_SELECT> backend, and only supports socket	3567	the form of the C<EVBACKEND_SELECT> backend, and only supports socket
2857	descriptors. This only applies when using Win32 natively, not when using	3568	descriptors. This only applies when using Win32 natively, not when using
2858	e.g. cygwin.	3569	e.g. cygwin.
2859		3570
		3571	Lifting these limitations would basically require the full
		3572	re-implementation of the I/O system. If you are into these kinds of
		3573	things, then note that glib does exactly that for you in a very portable
		3574	way (note also that glib is the slowest event library known to man).
		3575
2860	There is no supported compilation method available on windows except	3576	There is no supported compilation method available on windows except
2861	embedding it into other applications.	3577	embedding it into other applications.
2862		3578
		3579	Not a libev limitation but worth mentioning: windows apparently doesn't
		3580	accept large writes: instead of resulting in a partial write, windows will
		3581	either accept everything or return C<ENOBUFS> if the buffer is too large,
		3582	so make sure you only write small amounts into your sockets (less than a
		3583	megabyte seems safe, but this apparently depends on the amount of memory
		3584	available).
		3585
2863	Due to the many, low, and arbitrary limits on the win32 platform and the	3586	Due to the many, low, and arbitrary limits on the win32 platform and
2864	abysmal performance of winsockets, using a large number of sockets is not	3587	the abysmal performance of winsockets, using a large number of sockets
2865	recommended (and not reasonable). If your program needs to use more than	3588	is not recommended (and not reasonable). If your program needs to use
2866	a hundred or so sockets, then likely it needs to use a totally different	3589	more than a hundred or so sockets, then likely it needs to use a totally
2867	implementation for windows, as libev offers the POSIX model, which cannot	3590	different implementation for windows, as libev offers the POSIX readiness
2868	be implemented efficiently on windows (microsoft monopoly games).	3591	notification model, which cannot be implemented efficiently on windows
		3592	(Microsoft monopoly games).
		3593
		3594	A typical way to use libev under windows is to embed it (see the embedding
		3595	section for details) and use the following F<evwrap.h> header file instead
		3596	of F<ev.h>:
		3597
		3598	#define EV_STANDALONE /* keeps ev from requiring config.h */
		3599	#define EV_SELECT_IS_WINSOCKET 1 /* configure libev for windows select */
		3600
		3601	#include "ev.h"
		3602
		3603	And compile the following F<evwrap.c> file into your project (make sure
		3604	you do I<not> compile the F<ev.c> or any other embedded source files!):
		3605
		3606	#include "evwrap.h"
		3607	#include "ev.c"
2869		3608
2870	=over 4	3609	=over 4
2871		3610
2872	=item The winsocket select function	3611	=item The winsocket select function
2873		3612
2874	The winsocket C<select> function doesn't follow POSIX in that it requires	3613	The winsocket C<select> function doesn't follow POSIX in that it
2875	socket I<handles> and not socket I<file descriptors>. This makes select	3614	requires socket I<handles> and not socket I<file descriptors> (it is
2876	very inefficient, and also requires a mapping from file descriptors	3615	also extremely buggy). This makes select very inefficient, and also
2877	to socket handles. See the discussion of the C<EV_SELECT_USE_FD_SET>,	3616	requires a mapping from file descriptors to socket handles (the Microsoft
2878	C<EV_SELECT_IS_WINSOCKET> and C<EV_FD_TO_WIN32_HANDLE> preprocessor	3617	C runtime provides the function C<_open_osfhandle> for this). See the
2879	symbols for more info.	3618	discussion of the C<EV_SELECT_USE_FD_SET>, C<EV_SELECT_IS_WINSOCKET> and
		3619	C<EV_FD_TO_WIN32_HANDLE> preprocessor symbols for more info.
2880		3620
2881	The configuration for a "naked" win32 using the microsoft runtime	3621	The configuration for a "naked" win32 using the Microsoft runtime
2882	libraries and raw winsocket select is:	3622	libraries and raw winsocket select is:
2883		3623
2884	#define EV_USE_SELECT 1	3624	#define EV_USE_SELECT 1
2885	#define EV_SELECT_IS_WINSOCKET 1 /* forces EV_SELECT_USE_FD_SET, too */	3625	#define EV_SELECT_IS_WINSOCKET 1 /* forces EV_SELECT_USE_FD_SET, too */
2886		3626
2887	Note that winsockets handling of fd sets is O(n), so you can easily get a	3627	Note that winsockets handling of fd sets is O(n), so you can easily get a
2888	complexity in the O(n²) range when using win32.	3628	complexity in the O(n²) range when using win32.
2889		3629
2890	=item Limited number of file descriptors	3630	=item Limited number of file descriptors
2891		3631
2892	Windows has numerous arbitrary (and low) limits on things. Early versions	3632	Windows has numerous arbitrary (and low) limits on things.
2893	of winsocket's select only supported waiting for a max. of C<64> handles	3633
		3634	Early versions of winsocket's select only supported waiting for a maximum
2894	(probably owning to the fact that all windows kernels can only wait for	3635	of C<64> handles (probably owning to the fact that all windows kernels
2895	C<64> things at the same time internally; microsoft recommends spawning a	3636	can only wait for C<64> things at the same time internally; Microsoft
2896	chain of threads and wait for 63 handles and the previous thread in each).	3637	recommends spawning a chain of threads and wait for 63 handles and the
		3638	previous thread in each. Great).
2897		3639
2898	Newer versions support more handles, but you need to define C<FD_SETSIZE>	3640	Newer versions support more handles, but you need to define C<FD_SETSIZE>
2899	to some high number (e.g. C<2048>) before compiling the winsocket select	3641	to some high number (e.g. C<2048>) before compiling the winsocket select
2900	call (which might be in libev or elsewhere, for example, perl does its own	3642	call (which might be in libev or elsewhere, for example, perl does its own
2901	select emulation on windows).	3643	select emulation on windows).
2902		3644
2903	Another limit is the number of file descriptors in the microsoft runtime	3645	Another limit is the number of file descriptors in the Microsoft runtime
2904	libraries, which by default is C<64> (there must be a hidden I<64> fetish	3646	libraries, which by default is C<64> (there must be a hidden I<64> fetish
2905	or something like this inside microsoft). You can increase this by calling	3647	or something like this inside Microsoft). You can increase this by calling
2906	C<_setmaxstdio>, which can increase this limit to C<2048> (another	3648	C<_setmaxstdio>, which can increase this limit to C<2048> (another
2907	arbitrary limit), but is broken in many versions of the microsoft runtime	3649	arbitrary limit), but is broken in many versions of the Microsoft runtime
2908	libraries.	3650	libraries.
2909		3651
2910	This might get you to about C<512> or C<2048> sockets (depending on	3652	This might get you to about C<512> or C<2048> sockets (depending on
2911	windows version and/or the phase of the moon). To get more, you need to	3653	windows version and/or the phase of the moon). To get more, you need to
2912	wrap all I/O functions and provide your own fd management, but the cost of	3654	wrap all I/O functions and provide your own fd management, but the cost of
2913	calling select (O(n²)) will likely make this unworkable.	3655	calling select (O(n²)) will likely make this unworkable.
2914		3656
2915	=back	3657	=back
2916		3658
		3659	=head2 PORTABILITY REQUIREMENTS
		3660
		3661	In addition to a working ISO-C implementation and of course the
		3662	backend-specific APIs, libev relies on a few additional extensions:
		3663
		3664	=over 4
		3665
		3666	=item C<void (*)(ev_watcher_type *, int revents)> must have compatible
		3667	calling conventions regardless of C<ev_watcher_type *>.
		3668
		3669	Libev assumes not only that all watcher pointers have the same internal
		3670	structure (guaranteed by POSIX but not by ISO C for example), but it also
		3671	assumes that the same (machine) code can be used to call any watcher
		3672	callback: The watcher callbacks have different type signatures, but libev
		3673	calls them using an C<ev_watcher *> internally.
		3674
		3675	=item C<sig_atomic_t volatile> must be thread-atomic as well
		3676
		3677	The type C<sig_atomic_t volatile> (or whatever is defined as
		3678	C<EV_ATOMIC_T>) must be atomic with respect to accesses from different
		3679	threads. This is not part of the specification for C<sig_atomic_t>, but is
		3680	believed to be sufficiently portable.
		3681
		3682	=item C<sigprocmask> must work in a threaded environment
		3683
		3684	Libev uses C<sigprocmask> to temporarily block signals. This is not
		3685	allowed in a threaded program (C<pthread_sigmask> has to be used). Typical
		3686	pthread implementations will either allow C<sigprocmask> in the "main
		3687	thread" or will block signals process-wide, both behaviours would
		3688	be compatible with libev. Interaction between C<sigprocmask> and
		3689	C<pthread_sigmask> could complicate things, however.
		3690
		3691	The most portable way to handle signals is to block signals in all threads
		3692	except the initial one, and run the default loop in the initial thread as
		3693	well.
		3694
		3695	=item C<long> must be large enough for common memory allocation sizes
		3696
		3697	To improve portability and simplify its API, libev uses C<long> internally
		3698	instead of C<size_t> when allocating its data structures. On non-POSIX
		3699	systems (Microsoft...) this might be unexpectedly low, but is still at
		3700	least 31 bits everywhere, which is enough for hundreds of millions of
		3701	watchers.
		3702
		3703	=item C<double> must hold a time value in seconds with enough accuracy
		3704
		3705	The type C<double> is used to represent timestamps. It is required to
		3706	have at least 51 bits of mantissa (and 9 bits of exponent), which is good
		3707	enough for at least into the year 4000. This requirement is fulfilled by
		3708	implementations implementing IEEE 754 (basically all existing ones).
		3709
		3710	=back
		3711
		3712	If you know of other additional requirements drop me a note.
		3713
		3714
		3715	=head1 ALGORITHMIC COMPLEXITIES
		3716
		3717	In this section the complexities of (many of) the algorithms used inside
		3718	libev will be documented. For complexity discussions about backends see
		3719	the documentation for C<ev_default_init>.
		3720
		3721	All of the following are about amortised time: If an array needs to be
		3722	extended, libev needs to realloc and move the whole array, but this
		3723	happens asymptotically rarer with higher number of elements, so O(1) might
		3724	mean that libev does a lengthy realloc operation in rare cases, but on
		3725	average it is much faster and asymptotically approaches constant time.
		3726
		3727	=over 4
		3728
		3729	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)
		3730
		3731	This means that, when you have a watcher that triggers in one hour and
		3732	there are 100 watchers that would trigger before that, then inserting will
		3733	have to skip roughly seven (C<ld 100>) of these watchers.
		3734
		3735	=item Changing timer/periodic watchers (by autorepeat or calling again): O(log skipped_other_timers)
		3736
		3737	That means that changing a timer costs less than removing/adding them,
		3738	as only the relative motion in the event queue has to be paid for.
		3739
		3740	=item Starting io/check/prepare/idle/signal/child/fork/async watchers: O(1)
		3741
		3742	These just add the watcher into an array or at the head of a list.
		3743
		3744	=item Stopping check/prepare/idle/fork/async watchers: O(1)
		3745
		3746	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))
		3747
		3748	These watchers are stored in lists, so they need to be walked to find the
		3749	correct watcher to remove. The lists are usually short (you don't usually
		3750	have many watchers waiting for the same fd or signal: one is typical, two
		3751	is rare).
		3752
		3753	=item Finding the next timer in each loop iteration: O(1)
		3754
		3755	By virtue of using a binary or 4-heap, the next timer is always found at a
		3756	fixed position in the storage array.
		3757
		3758	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)
		3759
		3760	A change means an I/O watcher gets started or stopped, which requires
		3761	libev to recalculate its status (and possibly tell the kernel, depending
		3762	on backend and whether C<ev_io_set> was used).
		3763
		3764	=item Activating one watcher (putting it into the pending state): O(1)
		3765
		3766	=item Priority handling: O(number_of_priorities)
		3767
		3768	Priorities are implemented by allocating some space for each
		3769	priority. When doing priority-based operations, libev usually has to
		3770	linearly search all the priorities, but starting/stopping and activating
		3771	watchers becomes O(1) with respect to priority handling.
		3772
		3773	=item Sending an ev_async: O(1)
		3774
		3775	=item Processing ev_async_send: O(number_of_async_watchers)
		3776
		3777	=item Processing signals: O(max_signal_number)
		3778
		3779	Sending involves a system call I<iff> there were no other C<ev_async_send>
		3780	calls in the current loop iteration. Checking for async and signal events
		3781	involves iterating over all running async watchers or all signal numbers.
		3782
		3783	=back
		3784
2917		3785
2918	=head1 AUTHOR	3786	=head1 AUTHOR
2919		3787
2920	Marc Lehmann <libev@schmorp.de>.	3788	Marc Lehmann <libev@schmorp.de>.
2921		3789

Diff Legend

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