ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/libev/ev.pod

Comparing libev/ev.pod (file contents):
Revision 1.61 by root, Thu Nov 29 12:21:05 2007 UTC vs.
Revision 1.179 by root, Sat Sep 13 19:14:21 2008 UTC

…		…
2		2
3	libev - a high performance full-featured event loop written in C	3	libev - a high performance full-featured event loop written in C
4		4
5	=head1 SYNOPSIS	5	=head1 SYNOPSIS
6		6
7	#include <ev.h>	7	#include <ev.h>
8		8
9	=head1 EXAMPLE PROGRAM	9	=head2 EXAMPLE PROGRAM
10		10
		11	// a single header file is required
11	#include <ev.h>	12	#include <ev.h>
12		13
		14	// every watcher type has its own typedef'd struct
		15	// with the name ev_<type>
13	ev_io stdin_watcher;	16	ev_io stdin_watcher;
14	ev_timer timeout_watcher;	17	ev_timer timeout_watcher;
15		18
16	/* called when data readable on stdin */	19	// all watcher callbacks have a similar signature
		20	// this callback is called when data is readable on stdin
17	static void	21	static void
18	stdin_cb (EV_P_ struct ev_io *w, int revents)	22	stdin_cb (EV_P_ struct ev_io *w, int revents)
19	{	23	{
20	/* puts ("stdin ready"); */	24	puts ("stdin ready");
21	ev_io_stop (EV_A_ w); /* just a syntax example */	25	// for one-shot events, one must manually stop the watcher
22	ev_unloop (EV_A_ EVUNLOOP_ALL); /* leave all loop calls */	26	// with its corresponding stop function.
		27	ev_io_stop (EV_A_ w);
		28
		29	// this causes all nested ev_loop's to stop iterating
		30	ev_unloop (EV_A_ EVUNLOOP_ALL);
23	}	31	}
24		32
		33	// another callback, this time for a time-out
25	static void	34	static void
26	timeout_cb (EV_P_ struct ev_timer *w, int revents)	35	timeout_cb (EV_P_ struct ev_timer *w, int revents)
27	{	36	{
28	/* puts ("timeout"); */	37	puts ("timeout");
29	ev_unloop (EV_A_ EVUNLOOP_ONE); /* leave one loop call */	38	// this causes the innermost ev_loop to stop iterating
		39	ev_unloop (EV_A_ EVUNLOOP_ONE);
30	}	40	}
31		41
32	int	42	int
33	main (void)	43	main (void)
34	{	44	{
		45	// use the default event loop unless you have special needs
35	struct ev_loop *loop = ev_default_loop (0);	46	struct ev_loop *loop = ev_default_loop (0);
36		47
37	/* initialise an io watcher, then start it */	48	// initialise an io watcher, then start it
		49	// this one will watch for stdin to become readable
38	ev_io_init (&stdin_watcher, stdin_cb, /*STDIN_FILENO*/ 0, EV_READ);	50	ev_io_init (&stdin_watcher, stdin_cb, /*STDIN_FILENO*/ 0, EV_READ);
39	ev_io_start (loop, &stdin_watcher);	51	ev_io_start (loop, &stdin_watcher);
40		52
		53	// initialise a timer watcher, then start it
41	/* simple non-repeating 5.5 second timeout */	54	// simple non-repeating 5.5 second timeout
42	ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);	55	ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);
43	ev_timer_start (loop, &timeout_watcher);	56	ev_timer_start (loop, &timeout_watcher);
44		57
45	/* loop till timeout or data ready */	58	// now wait for events to arrive
46	ev_loop (loop, 0);	59	ev_loop (loop, 0);
47		60
		61	// unloop was called, so exit
48	return 0;	62	return 0;
49	}	63	}
50		64
51	=head1 DESCRIPTION	65	=head1 DESCRIPTION
52		66
		67	The newest version of this document is also available as an html-formatted
		68	web page you might find easier to navigate when reading it for the first
		69	time: L<http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod>.
		70
53	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
54	file descriptor being readable or a timeout occuring), and it will manage	72	file descriptor being readable or a timeout occurring), and it will manage
55	these event sources and provide your program with events.	73	these event sources and provide your program with events.
56		74
57	To do this, it must take more or less complete control over your process	75	To do this, it must take more or less complete control over your process
58	(or thread) by executing the I<event loop> handler, and will then	76	(or thread) by executing the I<event loop> handler, and will then
59	communicate events via a callback mechanism.	77	communicate events via a callback mechanism.
…		…
61	You register interest in certain events by registering so-called I<event	79	You register interest in certain events by registering so-called I<event
62	watchers>, which are relatively small C structures you initialise with the	80	watchers>, which are relatively small C structures you initialise with the
63	details of the event, and then hand it over to libev by I<starting> the	81	details of the event, and then hand it over to libev by I<starting> the
64	watcher.	82	watcher.
65		83
66	=head1 FEATURES	84	=head2 FEATURES
67		85
68	Libev supports C<select>, C<poll>, the Linux-specific C<epoll>, the	86	Libev supports C<select>, C<poll>, the Linux-specific C<epoll>, the
69	BSD-specific C<kqueue> and the Solaris-specific event port mechanisms	87	BSD-specific C<kqueue> and the Solaris-specific event port mechanisms
70	for file descriptor events (C<ev_io>), the Linux C<inotify> interface	88	for file descriptor events (C<ev_io>), the Linux C<inotify> interface
71	(for C<ev_stat>), relative timers (C<ev_timer>), absolute timers	89	(for C<ev_stat>), relative timers (C<ev_timer>), absolute timers
…		…
78		96
79	It also is quite fast (see this	97	It also is quite fast (see this
80	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
81	for example).	99	for example).
82		100
83	=head1 CONVENTIONS	101	=head2 CONVENTIONS
84		102
85	Libev is very configurable. In this manual the default configuration will	103	Libev is very configurable. In this manual the default (and most common)
86	be described, which supports multiple event loops. For more info about	104	configuration will be described, which supports multiple event loops. For
87	various configuration options please have a look at B<EMBED> section in	105	more info about various configuration options please have a look at
88	this manual. If libev was configured without support for multiple event	106	B<EMBED> section in this manual. If libev was configured without support
89	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
90	(which is always of type C<struct ev_loop *>) will not have this argument.	108	name C<loop> (which is always of type C<struct ev_loop *>) will not have
		109	this argument.
91		110
92	=head1 TIME REPRESENTATION	111	=head2 TIME REPRESENTATION
93		112
94	Libev represents time as a single floating point number, representing the	113	Libev represents time as a single floating point number, representing the
95	(fractional) number of seconds since the (POSIX) epoch (somewhere near	114	(fractional) number of seconds since the (POSIX) epoch (somewhere near
96	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
97	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
98	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
99	it, you should treat it as such.	118	it, you should treat it as some floating point value. Unlike the name
		119	component C<stamp> might indicate, it is also used for time differences
		120	throughout libev.
		121
		122	=head1 ERROR HANDLING
		123
		124	Libev knows three classes of errors: operating system errors, usage errors
		125	and internal errors (bugs).
		126
		127	When libev catches an operating system error it cannot handle (for example
		128	a system call indicating a condition libev cannot fix), it calls the callback
		129	set via C<ev_set_syserr_cb>, which is supposed to fix the problem or
		130	abort. The default is to print a diagnostic message and to call C<abort
		131	()>.
		132
		133	When libev detects a usage error such as a negative timer interval, then
		134	it will print a diagnostic message and abort (via the C<assert> mechanism,
		135	so C<NDEBUG> will disable this checking): these are programming errors in
		136	the libev caller and need to be fixed there.
		137
		138	Libev also has a few internal error-checking C<assert>ions, and also has
		139	extensive consistency checking code. These do not trigger under normal
		140	circumstances, as they indicate either a bug in libev or worse.
		141
100		142
101	=head1 GLOBAL FUNCTIONS	143	=head1 GLOBAL FUNCTIONS
102		144
103	These functions can be called anytime, even before initialising the	145	These functions can be called anytime, even before initialising the
104	library in any way.	146	library in any way.
…		…
109		151
110	Returns the current time as libev would use it. Please note that the	152	Returns the current time as libev would use it. Please note that the
111	C<ev_now> function is usually faster and also often returns the timestamp	153	C<ev_now> function is usually faster and also often returns the timestamp
112	you actually want to know.	154	you actually want to know.
113		155
		156	=item ev_sleep (ev_tstamp interval)
		157
		158	Sleep for the given interval: The current thread will be blocked until
		159	either it is interrupted or the given time interval has passed. Basically
		160	this is a sub-second-resolution C<sleep ()>.
		161
114	=item int ev_version_major ()	162	=item int ev_version_major ()
115		163
116	=item int ev_version_minor ()	164	=item int ev_version_minor ()
117		165
118	You can find out the major and minor version numbers of the library	166	You can find out the major and minor ABI version numbers of the library
119	you linked against by calling the functions C<ev_version_major> and	167	you linked against by calling the functions C<ev_version_major> and
120	C<ev_version_minor>. If you want, you can compare against the global	168	C<ev_version_minor>. If you want, you can compare against the global
121	symbols C<EV_VERSION_MAJOR> and C<EV_VERSION_MINOR>, which specify the	169	symbols C<EV_VERSION_MAJOR> and C<EV_VERSION_MINOR>, which specify the
122	version of the library your program was compiled against.	170	version of the library your program was compiled against.
123		171
		172	These version numbers refer to the ABI version of the library, not the
		173	release version.
		174
124	Usually, it's a good idea to terminate if the major versions mismatch,	175	Usually, it's a good idea to terminate if the major versions mismatch,
125	as this indicates an incompatible change. Minor versions are usually	176	as this indicates an incompatible change. Minor versions are usually
126	compatible to older versions, so a larger minor version alone is usually	177	compatible to older versions, so a larger minor version alone is usually
127	not a problem.	178	not a problem.
128		179
129	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
130	version.	181	version.
131		182
132	assert (("libev version mismatch",	183	assert (("libev version mismatch",
133	ev_version_major () == EV_VERSION_MAJOR	184	ev_version_major () == EV_VERSION_MAJOR
134	&& ev_version_minor () >= EV_VERSION_MINOR));	185	&& ev_version_minor () >= EV_VERSION_MINOR));
135		186
136	=item unsigned int ev_supported_backends ()	187	=item unsigned int ev_supported_backends ()
137		188
138	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_*>
139	value) compiled into this binary of libev (independent of their	190	value) compiled into this binary of libev (independent of their
…		…
141	a description of the set values.	192	a description of the set values.
142		193
143	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
144	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
145		196
146	assert (("sorry, no epoll, no sex",	197	assert (("sorry, no epoll, no sex",
147	ev_supported_backends () & EVBACKEND_EPOLL));	198	ev_supported_backends () & EVBACKEND_EPOLL));
148		199
149	=item unsigned int ev_recommended_backends ()	200	=item unsigned int ev_recommended_backends ()
150		201
151	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
152	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
153	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
154	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
155	(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
156	libev will probe for if you specify no backends explicitly.	207	libev will probe for if you specify no backends explicitly.
157		208
158	=item unsigned int ev_embeddable_backends ()	209	=item unsigned int ev_embeddable_backends ()
159		210
…		…
166	See the description of C<ev_embed> watchers for more info.	217	See the description of C<ev_embed> watchers for more info.
167		218
168	=item ev_set_allocator (void *(*cb)(void *ptr, long size))	219	=item ev_set_allocator (void *(*cb)(void *ptr, long size))
169		220
170	Sets the allocation function to use (the prototype is similar - the	221	Sets the allocation function to use (the prototype is similar - the
171	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
172	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
173	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
174	potentially destructive action. The default is your system realloc	225	or take some potentially destructive action.
175	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.
176		230
177	You could override this function in high-availability programs to, say,	231	You could override this function in high-availability programs to, say,
178	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,
179	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.
180		234
181	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
182	retries).	236	retries (example requires a standards-compliant C<realloc>).
183		237
184	static void *	238	static void *
185	persistent_realloc (void *ptr, size_t size)	239	persistent_realloc (void *ptr, size_t size)
186	{	240	{
187	for (;;)	241	for (;;)
…		…
198	...	252	...
199	ev_set_allocator (persistent_realloc);	253	ev_set_allocator (persistent_realloc);
200		254
201	=item ev_set_syserr_cb (void (*cb)(const char *msg));	255	=item ev_set_syserr_cb (void (*cb)(const char *msg));
202		256
203	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
204	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
205	indicating the system call or subsystem causing the problem. If this	259	indicating the system call or subsystem causing the problem. If this
206	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
207	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
208	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
209	(such as abort).	263	(such as abort).
210		264
211	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.
…		…
225	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP	279	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP
226		280
227	An event loop is described by a C<struct ev_loop *>. The library knows two	281	An event loop is described by a C<struct ev_loop *>. The library knows two
228	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
229	events, and dynamically created loops which do not.	283	events, and dynamically created loops which do not.
230
231	If you use threads, a common model is to run the default event loop
232	in your main thread (or in a separate thread) and for each thread you
233	create, you also create another event loop. Libev itself does no locking
234	whatsoever, so if you mix calls to the same event loop in different
235	threads, make sure you lock (this is usually a bad idea, though, even if
236	done correctly, because it's hideous and inefficient).
237		284
238	=over 4	285	=over 4
239		286
240	=item struct ev_loop *ev_default_loop (unsigned int flags)	287	=item struct ev_loop *ev_default_loop (unsigned int flags)
241		288
…		…
245	flags. If that is troubling you, check C<ev_backend ()> afterwards).	292	flags. If that is troubling you, check C<ev_backend ()> afterwards).
246		293
247	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
248	function.	295	function.
249		296
		297	Note that this function is I<not> thread-safe, so if you want to use it
		298	from multiple threads, you have to lock (note also that this is unlikely,
		299	as loops cannot bes hared easily between threads anyway).
		300
		301	The default loop is the only loop that can handle C<ev_signal> and
		302	C<ev_child> watchers, and to do this, it always registers a handler
		303	for C<SIGCHLD>. If this is a problem for your application you can either
		304	create a dynamic loop with C<ev_loop_new> that doesn't do that, or you
		305	can simply overwrite the C<SIGCHLD> signal handler I<after> calling
		306	C<ev_default_init>.
		307
250	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
251	backends to use, and is usually specified as C<0> (or C<EVFLAG_AUTO>).	309	backends to use, and is usually specified as C<0> (or C<EVFLAG_AUTO>).
252		310
253	The following flags are supported:	311	The following flags are supported:
254		312
…		…
259	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
260	thing, believe me).	318	thing, believe me).
261		319
262	=item C<EVFLAG_NOENV>	320	=item C<EVFLAG_NOENV>
263		321
264	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
265	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
266	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will	324	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will
267	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
268	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
269	around bugs.	327	around bugs.
270		328
		329	=item C<EVFLAG_FORKCHECK>
		330
		331	Instead of calling C<ev_default_fork> or C<ev_loop_fork> manually after
		332	a fork, you can also make libev check for a fork in each iteration by
		333	enabling this flag.
		334
		335	This works by calling C<getpid ()> on every iteration of the loop,
		336	and thus this might slow down your event loop if you do a lot of loop
		337	iterations and little real work, but is usually not noticeable (on my
		338	GNU/Linux system for example, C<getpid> is actually a simple 5-insn sequence
		339	without a system call and thus I<very> fast, but my GNU/Linux system also has
		340	C<pthread_atfork> which is even faster).
		341
		342	The big advantage of this flag is that you can forget about fork (and
		343	forget about forgetting to tell libev about forking) when you use this
		344	flag.
		345
		346	This flag setting cannot be overridden or specified in the C<LIBEV_FLAGS>
		347	environment variable.
		348
271	=item C<EVBACKEND_SELECT> (value 1, portable select backend)	349	=item C<EVBACKEND_SELECT> (value 1, portable select backend)
272		350
273	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
274	libev tries to roll its own fd_set with no limits on the number of fds,	352	libev tries to roll its own fd_set with no limits on the number of fds,
275	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
276	using this backend. It doesn't scale too well (O(highest_fd)), but its usually	354	using this backend. It doesn't scale too well (O(highest_fd)), but its
277	the fastest backend for a low number of fds.	355	usually the fastest backend for a low number of (low-numbered :) fds.
		356
		357	To get good performance out of this backend you need a high amount of
		358	parallelism (most of the file descriptors should be busy). If you are
		359	writing a server, you should C<accept ()> in a loop to accept as many
		360	connections as possible during one iteration. You might also want to have
		361	a look at C<ev_set_io_collect_interval ()> to increase the amount of
		362	readiness notifications you get per iteration.
		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).
278		367
279	=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)
280		369
281	And this is your standard poll(2) backend. It's more complicated than	370	And this is your standard poll(2) backend. It's more complicated
282	select, but handles sparse fds better and has no artificial limit on the	371	than select, but handles sparse fds better and has no artificial
283	number of fds you can use (except it will slow down considerably with a	372	limit on the number of fds you can use (except it will slow down
284	lot of inactive fds). It scales similarly to select, i.e. O(total_fds).	373	considerably with a lot of inactive fds). It scales similarly to select,
		374	i.e. O(total_fds). See the entry for C<EVBACKEND_SELECT>, above, for
		375	performance tips.
		376
		377	This backend maps C<EV_READ> to C<POLLIN | POLLERR | POLLHUP>, and
		378	C<EV_WRITE> to C<POLLOUT | POLLERR | POLLHUP>.
285		379
286	=item C<EVBACKEND_EPOLL> (value 4, Linux)	380	=item C<EVBACKEND_EPOLL> (value 4, Linux)
287		381
288	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,
289	but it scales phenomenally better. While poll and select usually scale like	383	but it scales phenomenally better. While poll and select usually scale
290	O(total_fds) where n is the total number of fds (or the highest fd), epoll scales	384	like O(total_fds) where n is the total number of fds (or the highest fd),
291	either O(1) or O(active_fds).	385	epoll scales either O(1) or O(active_fds). The epoll design has a number
		386	of shortcomings, such as silently dropping events in some hard-to-detect
		387	cases and requiring a system call per fd change, no fork support and bad
		388	support for dup.
292		389
293	While stopping and starting an I/O watcher in the same iteration will	390	While stopping, setting and starting an I/O watcher in the same iteration
294	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
295	(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
296	best to avoid that. Also, dup()ed file descriptors might not work very	393	best to avoid that. Also, C<dup ()>'ed file descriptors might not work
297	well if you register events for both fds.	394	very well if you register events for both fds.
298		395
299	Please note that epoll sometimes generates spurious notifications, so you	396	Please note that epoll sometimes generates spurious notifications, so you
300	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
301	(or space) is available.	398	(or space) is available.
302		399
		400	Best performance from this backend is achieved by not unregistering all
		401	watchers for a file descriptor until it has been closed, if possible, i.e.
		402	keep at least one watcher active per fd at all times.
		403
		404	While nominally embeddable in other event loops, this feature is broken in
		405	all kernel versions tested so far.
		406
		407	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
		408	C<EVBACKEND_POLL>.
		409
303	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)	410	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)
304		411
305	Kqueue deserves special mention, as at the time of this writing, it	412	Kqueue deserves special mention, as at the time of this writing, it
306	was broken on all BSDs except NetBSD (usually it doesn't work with	413	was broken on all BSDs except NetBSD (usually it doesn't work reliably
307	anything but sockets and pipes, except on Darwin, where of course its	414	with anything but sockets and pipes, except on Darwin, where of course
308	completely useless). For this reason its not being "autodetected"	415	it's completely useless). For this reason it's not being "auto-detected"
309	unless you explicitly specify it explicitly in the flags (i.e. using	416	unless you explicitly specify it explicitly in the flags (i.e. using
310	C<EVBACKEND_KQUEUE>).	417	C<EVBACKEND_KQUEUE>) or libev was compiled on a known-to-be-good (-enough)
		418	system like NetBSD.
		419
		420	You still can embed kqueue into a normal poll or select backend and use it
		421	only for sockets (after having made sure that sockets work with kqueue on
		422	the target platform). See C<ev_embed> watchers for more info.
311		423
312	It scales in the same way as the epoll backend, but the interface to the	424	It scales in the same way as the epoll backend, but the interface to the
313	kernel is more efficient (which says nothing about its actual speed, of	425	kernel is more efficient (which says nothing about its actual speed, of
314	course). While starting and stopping an I/O watcher does not cause an	426	course). While stopping, setting and starting an I/O watcher does never
315	extra syscall as with epoll, it still adds up to four event changes per	427	cause an extra system call as with C<EVBACKEND_EPOLL>, it still adds up to
316	incident, so its best to avoid that.	428	two event changes per incident, support for C<fork ()> is very bad and it
		429	drops fds silently in similarly hard-to-detect cases.
		430
		431	This backend usually performs well under most conditions.
		432
		433	While nominally embeddable in other event loops, this doesn't work
		434	everywhere, so you might need to test for this. And since it is broken
		435	almost everywhere, you should only use it when you have a lot of sockets
		436	(for which it usually works), by embedding it into another event loop
		437	(e.g. C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>) and using it only for
		438	sockets.
		439
		440	This backend maps C<EV_READ> into an C<EVFILT_READ> kevent with
		441	C<NOTE_EOF>, and C<EV_WRITE> into an C<EVFILT_WRITE> kevent with
		442	C<NOTE_EOF>.
317		443
318	=item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8)	444	=item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8)
319		445
320	This is not implemented yet (and might never be).	446	This is not implemented yet (and might never be, unless you send me an
		447	implementation). According to reports, C</dev/poll> only supports sockets
		448	and is not embeddable, which would limit the usefulness of this backend
		449	immensely.
321		450
322	=item C<EVBACKEND_PORT> (value 32, Solaris 10)	451	=item C<EVBACKEND_PORT> (value 32, Solaris 10)
323		452
324	This uses the Solaris 10 port mechanism. As with everything on Solaris,	453	This uses the Solaris 10 event port mechanism. As with everything on Solaris,
325	it's really slow, but it still scales very well (O(active_fds)).	454	it's really slow, but it still scales very well (O(active_fds)).
326		455
327	Please note that solaris ports can result in a lot of spurious	456	Please note that Solaris event ports can deliver a lot of spurious
328	notifications, so you need to use non-blocking I/O or other means to avoid	457	notifications, so you need to use non-blocking I/O or other means to avoid
329	blocking when no data (or space) is available.	458	blocking when no data (or space) is available.
		459
		460	While this backend scales well, it requires one system call per active
		461	file descriptor per loop iteration. For small and medium numbers of file
		462	descriptors a "slow" C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> backend
		463	might perform better.
		464
		465	On the positive side, ignoring the spurious readiness notifications, this
		466	backend actually performed to specification in all tests and is fully
		467	embeddable, which is a rare feat among the OS-specific backends.
		468
		469	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
		470	C<EVBACKEND_POLL>.
330		471
331	=item C<EVBACKEND_ALL>	472	=item C<EVBACKEND_ALL>
332		473
333	Try all backends (even potentially broken ones that wouldn't be tried	474	Try all backends (even potentially broken ones that wouldn't be tried
334	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as	475	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as
335	C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>.	476	C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>.
336		477
		478	It is definitely not recommended to use this flag.
		479
337	=back	480	=back
338		481
339	If one or more of these are ored into the flags value, then only these	482	If one or more of these are or'ed into the flags value, then only these
340	backends will be tried (in the reverse order as given here). If none are	483	backends will be tried (in the reverse order as listed here). If none are
341	specified, most compiled-in backend will be tried, usually in reverse	484	specified, all backends in C<ev_recommended_backends ()> will be tried.
342	order of their flag values :)
343		485
344	The most typical usage is like this:	486	The most typical usage is like this:
345		487
346	if (!ev_default_loop (0))	488	if (!ev_default_loop (0))
347	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");	489	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
348		490
349	Restrict libev to the select and poll backends, and do not allow	491	Restrict libev to the select and poll backends, and do not allow
350	environment settings to be taken into account:	492	environment settings to be taken into account:
351		493
352	ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);	494	ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);
353		495
354	Use whatever libev has to offer, but make sure that kqueue is used if	496	Use whatever libev has to offer, but make sure that kqueue is used if
355	available (warning, breaks stuff, best use only with your own private	497	available (warning, breaks stuff, best use only with your own private
356	event loop and only if you know the OS supports your types of fds):	498	event loop and only if you know the OS supports your types of fds):
357		499
358	ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);	500	ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);
359		501
360	=item struct ev_loop *ev_loop_new (unsigned int flags)	502	=item struct ev_loop *ev_loop_new (unsigned int flags)
361		503
362	Similar to C<ev_default_loop>, but always creates a new event loop that is	504	Similar to C<ev_default_loop>, but always creates a new event loop that is
363	always distinct from the default loop. Unlike the default loop, it cannot	505	always distinct from the default loop. Unlike the default loop, it cannot
364	handle signal and child watchers, and attempts to do so will be greeted by	506	handle signal and child watchers, and attempts to do so will be greeted by
365	undefined behaviour (or a failed assertion if assertions are enabled).	507	undefined behaviour (or a failed assertion if assertions are enabled).
366		508
		509	Note that this function I<is> thread-safe, and the recommended way to use
		510	libev with threads is indeed to create one loop per thread, and using the
		511	default loop in the "main" or "initial" thread.
		512
367	Example: Try to create a event loop that uses epoll and nothing else.	513	Example: Try to create a event loop that uses epoll and nothing else.
368		514
369	struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);	515	struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);
370	if (!epoller)	516	if (!epoller)
371	fatal ("no epoll found here, maybe it hides under your chair");	517	fatal ("no epoll found here, maybe it hides under your chair");
372		518
373	=item ev_default_destroy ()	519	=item ev_default_destroy ()
374		520
375	Destroys the default loop again (frees all memory and kernel state	521	Destroys the default loop again (frees all memory and kernel state
376	etc.). None of the active event watchers will be stopped in the normal	522	etc.). None of the active event watchers will be stopped in the normal
377	sense, so e.g. C<ev_is_active> might still return true. It is your	523	sense, so e.g. C<ev_is_active> might still return true. It is your
378	responsibility to either stop all watchers cleanly yoursef I<before>	524	responsibility to either stop all watchers cleanly yourself I<before>
379	calling this function, or cope with the fact afterwards (which is usually	525	calling this function, or cope with the fact afterwards (which is usually
380	the easiest thing, youc na just ignore the watchers and/or C<free ()> them	526	the easiest thing, you can just ignore the watchers and/or C<free ()> them
381	for example).	527	for example).
		528
		529	Note that certain global state, such as signal state, will not be freed by
		530	this function, and related watchers (such as signal and child watchers)
		531	would need to be stopped manually.
		532
		533	In general it is not advisable to call this function except in the
		534	rare occasion where you really need to free e.g. the signal handling
		535	pipe fds. If you need dynamically allocated loops it is better to use
		536	C<ev_loop_new> and C<ev_loop_destroy>).
382		537
383	=item ev_loop_destroy (loop)	538	=item ev_loop_destroy (loop)
384		539
385	Like C<ev_default_destroy>, but destroys an event loop created by an	540	Like C<ev_default_destroy>, but destroys an event loop created by an
386	earlier call to C<ev_loop_new>.	541	earlier call to C<ev_loop_new>.
387		542
388	=item ev_default_fork ()	543	=item ev_default_fork ()
389		544
		545	This function sets a flag that causes subsequent C<ev_loop> iterations
390	This function reinitialises the kernel state for backends that have	546	to reinitialise the kernel state for backends that have one. Despite the
391	one. Despite the name, you can call it anytime, but it makes most sense	547	name, you can call it anytime, but it makes most sense after forking, in
392	after forking, in either the parent or child process (or both, but that	548	the child process (or both child and parent, but that again makes little
393	again makes little sense).	549	sense). You I<must> call it in the child before using any of the libev
		550	functions, and it will only take effect at the next C<ev_loop> iteration.
394		551
395	You I<must> call this function in the child process after forking if and	552	On the other hand, you only need to call this function in the child
396	only if you want to use the event library in both processes. If you just	553	process if and only if you want to use the event library in the child. If
397	fork+exec, you don't have to call it.	554	you just fork+exec, you don't have to call it at all.
398		555
399	The function itself is quite fast and it's usually not a problem to call	556	The function itself is quite fast and it's usually not a problem to call
400	it just in case after a fork. To make this easy, the function will fit in	557	it just in case after a fork. To make this easy, the function will fit in
401	quite nicely into a call to C<pthread_atfork>:	558	quite nicely into a call to C<pthread_atfork>:
402		559
403	pthread_atfork (0, 0, ev_default_fork);	560	pthread_atfork (0, 0, ev_default_fork);
404		561
405	At the moment, C<EVBACKEND_SELECT> and C<EVBACKEND_POLL> are safe to use
406	without calling this function, so if you force one of those backends you
407	do not need to care.
408
409	=item ev_loop_fork (loop)	562	=item ev_loop_fork (loop)
410		563
411	Like C<ev_default_fork>, but acts on an event loop created by	564	Like C<ev_default_fork>, but acts on an event loop created by
412	C<ev_loop_new>. Yes, you have to call this on every allocated event loop	565	C<ev_loop_new>. Yes, you have to call this on every allocated event loop
413	after fork, and how you do this is entirely your own problem.	566	after fork, and how you do this is entirely your own problem.
		567
		568	=item int ev_is_default_loop (loop)
		569
		570	Returns true when the given loop actually is the default loop, false otherwise.
		571
		572	=item unsigned int ev_loop_count (loop)
		573
		574	Returns the count of loop iterations for the loop, which is identical to
		575	the number of times libev did poll for new events. It starts at C<0> and
		576	happily wraps around with enough iterations.
		577
		578	This value can sometimes be useful as a generation counter of sorts (it
		579	"ticks" the number of loop iterations), as it roughly corresponds with
		580	C<ev_prepare> and C<ev_check> calls.
414		581
415	=item unsigned int ev_backend (loop)	582	=item unsigned int ev_backend (loop)
416		583
417	Returns one of the C<EVBACKEND_*> flags indicating the event backend in	584	Returns one of the C<EVBACKEND_*> flags indicating the event backend in
418	use.	585	use.
…		…
421		588
422	Returns the current "event loop time", which is the time the event loop	589	Returns the current "event loop time", which is the time the event loop
423	received events and started processing them. This timestamp does not	590	received events and started processing them. This timestamp does not
424	change as long as callbacks are being processed, and this is also the base	591	change as long as callbacks are being processed, and this is also the base
425	time used for relative timers. You can treat it as the timestamp of the	592	time used for relative timers. You can treat it as the timestamp of the
426	event occuring (or more correctly, libev finding out about it).	593	event occurring (or more correctly, libev finding out about it).
		594
		595	=item ev_now_update (loop)
		596
		597	Establishes the current time by querying the kernel, updating the time
		598	returned by C<ev_now ()> in the progress. This is a costly operation and
		599	is usually done automatically within C<ev_loop ()>.
		600
		601	This function is rarely useful, but when some event callback runs for a
		602	very long time without entering the event loop, updating libev's idea of
		603	the current time is a good idea.
		604
		605	See also "The special problem of time updates" in the C<ev_timer> section.
427		606
428	=item ev_loop (loop, int flags)	607	=item ev_loop (loop, int flags)
429		608
430	Finally, this is it, the event handler. This function usually is called	609	Finally, this is it, the event handler. This function usually is called
431	after you initialised all your watchers and you want to start handling	610	after you initialised all your watchers and you want to start handling
…		…
443	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle	622	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle
444	those events and any outstanding ones, but will not block your process in	623	those events and any outstanding ones, but will not block your process in
445	case there are no events and will return after one iteration of the loop.	624	case there are no events and will return after one iteration of the loop.
446		625
447	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if	626	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if
448	neccessary) and will handle those and any outstanding ones. It will block	627	necessary) and will handle those and any outstanding ones. It will block
449	your process until at least one new event arrives, and will return after	628	your process until at least one new event arrives, and will return after
450	one iteration of the loop. This is useful if you are waiting for some	629	one iteration of the loop. This is useful if you are waiting for some
451	external event in conjunction with something not expressible using other	630	external event in conjunction with something not expressible using other
452	libev watchers. However, a pair of C<ev_prepare>/C<ev_check> watchers is	631	libev watchers. However, a pair of C<ev_prepare>/C<ev_check> watchers is
453	usually a better approach for this kind of thing.	632	usually a better approach for this kind of thing.
454		633
455	Here are the gory details of what C<ev_loop> does:	634	Here are the gory details of what C<ev_loop> does:
456		635
457	* If there are no active watchers (reference count is zero), return.	636	- Before the first iteration, call any pending watchers.
458	- Queue prepare watchers and then call all outstanding watchers.	637	* If EVFLAG_FORKCHECK was used, check for a fork.
		638	- If a fork was detected (by any means), queue and call all fork watchers.
		639	- Queue and call all prepare watchers.
459	- If we have been forked, recreate the kernel state.	640	- If we have been forked, detach and recreate the kernel state
		641	as to not disturb the other process.
460	- Update the kernel state with all outstanding changes.	642	- Update the kernel state with all outstanding changes.
461	- Update the "event loop time".	643	- Update the "event loop time" (ev_now ()).
462	- Calculate for how long to block.	644	- Calculate for how long to sleep or block, if at all
		645	(active idle watchers, EVLOOP_NONBLOCK or not having
		646	any active watchers at all will result in not sleeping).
		647	- Sleep if the I/O and timer collect interval say so.
463	- Block the process, waiting for any events.	648	- Block the process, waiting for any events.
464	- Queue all outstanding I/O (fd) events.	649	- Queue all outstanding I/O (fd) events.
465	- Update the "event loop time" and do time jump handling.	650	- Update the "event loop time" (ev_now ()), and do time jump adjustments.
466	- Queue all outstanding timers.	651	- Queue all outstanding timers.
467	- Queue all outstanding periodics.	652	- Queue all outstanding periodics.
468	- If no events are pending now, queue all idle watchers.	653	- Unless any events are pending now, queue all idle watchers.
469	- Queue all check watchers.	654	- Queue all check watchers.
470	- Call all queued watchers in reverse order (i.e. check watchers first).	655	- Call all queued watchers in reverse order (i.e. check watchers first).
471	Signals and child watchers are implemented as I/O watchers, and will	656	Signals and child watchers are implemented as I/O watchers, and will
472	be handled here by queueing them when their watcher gets executed.	657	be handled here by queueing them when their watcher gets executed.
473	- If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK	658	- If ev_unloop has been called, or EVLOOP_ONESHOT or EVLOOP_NONBLOCK
474	were used, return, otherwise continue with step *.	659	were used, or there are no active watchers, return, otherwise
		660	continue with step *.
475		661
476	Example: Queue some jobs and then loop until no events are outsanding	662	Example: Queue some jobs and then loop until no events are outstanding
477	anymore.	663	anymore.
478		664
479	... queue jobs here, make sure they register event watchers as long	665	... queue jobs here, make sure they register event watchers as long
480	... as they still have work to do (even an idle watcher will do..)	666	... as they still have work to do (even an idle watcher will do..)
481	ev_loop (my_loop, 0);	667	ev_loop (my_loop, 0);
482	... jobs done. yeah!	668	... jobs done or somebody called unloop. yeah!
483		669
484	=item ev_unloop (loop, how)	670	=item ev_unloop (loop, how)
485		671
486	Can be used to make a call to C<ev_loop> return early (but only after it	672	Can be used to make a call to C<ev_loop> return early (but only after it
487	has processed all outstanding events). The C<how> argument must be either	673	has processed all outstanding events). The C<how> argument must be either
488	C<EVUNLOOP_ONE>, which will make the innermost C<ev_loop> call return, or	674	C<EVUNLOOP_ONE>, which will make the innermost C<ev_loop> call return, or
489	C<EVUNLOOP_ALL>, which will make all nested C<ev_loop> calls return.	675	C<EVUNLOOP_ALL>, which will make all nested C<ev_loop> calls return.
		676
		677	This "unloop state" will be cleared when entering C<ev_loop> again.
490		678
491	=item ev_ref (loop)	679	=item ev_ref (loop)
492		680
493	=item ev_unref (loop)	681	=item ev_unref (loop)
494		682
…		…
499	returning, ev_unref() after starting, and ev_ref() before stopping it. For	687	returning, ev_unref() after starting, and ev_ref() before stopping it. For
500	example, libev itself uses this for its internal signal pipe: It is not	688	example, libev itself uses this for its internal signal pipe: It is not
501	visible to the libev user and should not keep C<ev_loop> from exiting if	689	visible to the libev user and should not keep C<ev_loop> from exiting if
502	no event watchers registered by it are active. It is also an excellent	690	no event watchers registered by it are active. It is also an excellent
503	way to do this for generic recurring timers or from within third-party	691	way to do this for generic recurring timers or from within third-party
504	libraries. Just remember to I<unref after start> and I<ref before stop>.	692	libraries. Just remember to I<unref after start> and I<ref before stop>
		693	(but only if the watcher wasn't active before, or was active before,
		694	respectively).
505		695
506	Example: Create a signal watcher, but keep it from keeping C<ev_loop>	696	Example: Create a signal watcher, but keep it from keeping C<ev_loop>
507	running when nothing else is active.	697	running when nothing else is active.
508		698
509	struct ev_signal exitsig;	699	struct ev_signal exitsig;
510	ev_signal_init (&exitsig, sig_cb, SIGINT);	700	ev_signal_init (&exitsig, sig_cb, SIGINT);
511	ev_signal_start (loop, &exitsig);	701	ev_signal_start (loop, &exitsig);
512	evf_unref (loop);	702	evf_unref (loop);
513		703
514	Example: For some weird reason, unregister the above signal handler again.	704	Example: For some weird reason, unregister the above signal handler again.
515		705
516	ev_ref (loop);	706	ev_ref (loop);
517	ev_signal_stop (loop, &exitsig);	707	ev_signal_stop (loop, &exitsig);
		708
		709	=item ev_set_io_collect_interval (loop, ev_tstamp interval)
		710
		711	=item ev_set_timeout_collect_interval (loop, ev_tstamp interval)
		712
		713	These advanced functions influence the time that libev will spend waiting
		714	for events. Both time intervals are by default C<0>, meaning that libev
		715	will try to invoke timer/periodic callbacks and I/O callbacks with minimum
		716	latency.
		717
		718	Setting these to a higher value (the C<interval> I<must> be >= C<0>)
		719	allows libev to delay invocation of I/O and timer/periodic callbacks
		720	to increase efficiency of loop iterations (or to increase power-saving
		721	opportunities).
		722
		723	The background is that sometimes your program runs just fast enough to
		724	handle one (or very few) event(s) per loop iteration. While this makes
		725	the program responsive, it also wastes a lot of CPU time to poll for new
		726	events, especially with backends like C<select ()> which have a high
		727	overhead for the actual polling but can deliver many events at once.
		728
		729	By setting a higher I<io collect interval> you allow libev to spend more
		730	time collecting I/O events, so you can handle more events per iteration,
		731	at the cost of increasing latency. Timeouts (both C<ev_periodic> and
		732	C<ev_timer>) will be not affected. Setting this to a non-null value will
		733	introduce an additional C<ev_sleep ()> call into most loop iterations.
		734
		735	Likewise, by setting a higher I<timeout collect interval> you allow libev
		736	to spend more time collecting timeouts, at the expense of increased
		737	latency (the watcher callback will be called later). C<ev_io> watchers
		738	will not be affected. Setting this to a non-null value will not introduce
		739	any overhead in libev.
		740
		741	Many (busy) programs can usually benefit by setting the I/O collect
		742	interval to a value near C<0.1> or so, which is often enough for
		743	interactive servers (of course not for games), likewise for timeouts. It
		744	usually doesn't make much sense to set it to a lower value than C<0.01>,
		745	as this approaches the timing granularity of most systems.
		746
		747	Setting the I<timeout collect interval> can improve the opportunity for
		748	saving power, as the program will "bundle" timer callback invocations that
		749	are "near" in time together, by delaying some, thus reducing the number of
		750	times the process sleeps and wakes up again. Another useful technique to
		751	reduce iterations/wake-ups is to use C<ev_periodic> watchers and make sure
		752	they fire on, say, one-second boundaries only.
		753
		754	=item ev_loop_verify (loop)
		755
		756	This function only does something when C<EV_VERIFY> support has been
		757	compiled in. It tries to go through all internal structures and checks
		758	them for validity. If anything is found to be inconsistent, it will print
		759	an error message to standard error and call C<abort ()>.
		760
		761	This can be used to catch bugs inside libev itself: under normal
		762	circumstances, this function will never abort as of course libev keeps its
		763	data structures consistent.
518		764
519	=back	765	=back
520		766
521		767
522	=head1 ANATOMY OF A WATCHER	768	=head1 ANATOMY OF A WATCHER
523		769
524	A watcher is a structure that you create and register to record your	770	A watcher is a structure that you create and register to record your
525	interest in some event. For instance, if you want to wait for STDIN to	771	interest in some event. For instance, if you want to wait for STDIN to
526	become readable, you would create an C<ev_io> watcher for that:	772	become readable, you would create an C<ev_io> watcher for that:
527		773
528	static void my_cb (struct ev_loop *loop, struct ev_io *w, int revents)	774	static void my_cb (struct ev_loop *loop, struct ev_io *w, int revents)
529	{	775	{
530	ev_io_stop (w);	776	ev_io_stop (w);
531	ev_unloop (loop, EVUNLOOP_ALL);	777	ev_unloop (loop, EVUNLOOP_ALL);
532	}	778	}
533		779
534	struct ev_loop *loop = ev_default_loop (0);	780	struct ev_loop *loop = ev_default_loop (0);
535	struct ev_io stdin_watcher;	781	struct ev_io stdin_watcher;
536	ev_init (&stdin_watcher, my_cb);	782	ev_init (&stdin_watcher, my_cb);
537	ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);	783	ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);
538	ev_io_start (loop, &stdin_watcher);	784	ev_io_start (loop, &stdin_watcher);
539	ev_loop (loop, 0);	785	ev_loop (loop, 0);
540		786
541	As you can see, you are responsible for allocating the memory for your	787	As you can see, you are responsible for allocating the memory for your
542	watcher structures (and it is usually a bad idea to do this on the stack,	788	watcher structures (and it is usually a bad idea to do this on the stack,
543	although this can sometimes be quite valid).	789	although this can sometimes be quite valid).
544		790
545	Each watcher structure must be initialised by a call to C<ev_init	791	Each watcher structure must be initialised by a call to C<ev_init
546	(watcher *, callback)>, which expects a callback to be provided. This	792	(watcher *, callback)>, which expects a callback to be provided. This
547	callback gets invoked each time the event occurs (or, in the case of io	793	callback gets invoked each time the event occurs (or, in the case of I/O
548	watchers, each time the event loop detects that the file descriptor given	794	watchers, each time the event loop detects that the file descriptor given
549	is readable and/or writable).	795	is readable and/or writable).
550		796
551	Each watcher type has its own C<< ev_<type>_set (watcher *, ...) >> macro	797	Each watcher type has its own C<< ev_<type>_set (watcher *, ...) >> macro
552	with arguments specific to this watcher type. There is also a macro	798	with arguments specific to this watcher type. There is also a macro
…		…
622	=item C<EV_FORK>	868	=item C<EV_FORK>
623		869
624	The event loop has been resumed in the child process after fork (see	870	The event loop has been resumed in the child process after fork (see
625	C<ev_fork>).	871	C<ev_fork>).
626		872
		873	=item C<EV_ASYNC>
		874
		875	The given async watcher has been asynchronously notified (see C<ev_async>).
		876
627	=item C<EV_ERROR>	877	=item C<EV_ERROR>
628		878
629	An unspecified error has occured, the watcher has been stopped. This might	879	An unspecified error has occurred, the watcher has been stopped. This might
630	happen because the watcher could not be properly started because libev	880	happen because the watcher could not be properly started because libev
631	ran out of memory, a file descriptor was found to be closed or any other	881	ran out of memory, a file descriptor was found to be closed or any other
632	problem. You best act on it by reporting the problem and somehow coping	882	problem. You best act on it by reporting the problem and somehow coping
633	with the watcher being stopped.	883	with the watcher being stopped.
634		884
635	Libev will usually signal a few "dummy" events together with an error,	885	Libev will usually signal a few "dummy" events together with an error,
636	for example it might indicate that a fd is readable or writable, and if	886	for example it might indicate that a fd is readable or writable, and if
637	your callbacks is well-written it can just attempt the operation and cope	887	your callbacks is well-written it can just attempt the operation and cope
638	with the error from read() or write(). This will not work in multithreaded	888	with the error from read() or write(). This will not work in multi-threaded
639	programs, though, so beware.	889	programs, though, so beware.
640		890
641	=back	891	=back
642		892
643	=head2 GENERIC WATCHER FUNCTIONS	893	=head2 GENERIC WATCHER FUNCTIONS
…		…
673	Although some watcher types do not have type-specific arguments	923	Although some watcher types do not have type-specific arguments
674	(e.g. C<ev_prepare>) you still need to call its C<set> macro.	924	(e.g. C<ev_prepare>) you still need to call its C<set> macro.
675		925
676	=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])	926	=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])
677		927
678	This convinience macro rolls both C<ev_init> and C<ev_TYPE_set> macro	928	This convenience macro rolls both C<ev_init> and C<ev_TYPE_set> macro
679	calls into a single call. This is the most convinient method to initialise	929	calls into a single call. This is the most convenient method to initialise
680	a watcher. The same limitations apply, of course.	930	a watcher. The same limitations apply, of course.
681		931
682	=item C<ev_TYPE_start> (loop *, ev_TYPE *watcher)	932	=item C<ev_TYPE_start> (loop *, ev_TYPE *watcher)
683		933
684	Starts (activates) the given watcher. Only active watchers will receive	934	Starts (activates) the given watcher. Only active watchers will receive
…		…
702	=item bool ev_is_pending (ev_TYPE *watcher)	952	=item bool ev_is_pending (ev_TYPE *watcher)
703		953
704	Returns a true value iff the watcher is pending, (i.e. it has outstanding	954	Returns a true value iff the watcher is pending, (i.e. it has outstanding
705	events but its callback has not yet been invoked). As long as a watcher	955	events but its callback has not yet been invoked). As long as a watcher
706	is pending (but not active) you must not call an init function on it (but	956	is pending (but not active) you must not call an init function on it (but
707	C<ev_TYPE_set> is safe) and you must make sure the watcher is available to	957	C<ev_TYPE_set> is safe), you must not change its priority, and you must
708	libev (e.g. you cnanot C<free ()> it).	958	make sure the watcher is available to libev (e.g. you cannot C<free ()>
		959	it).
709		960
710	=item callback ev_cb (ev_TYPE *watcher)	961	=item callback ev_cb (ev_TYPE *watcher)
711		962
712	Returns the callback currently set on the watcher.	963	Returns the callback currently set on the watcher.
713		964
714	=item ev_cb_set (ev_TYPE *watcher, callback)	965	=item ev_cb_set (ev_TYPE *watcher, callback)
715		966
716	Change the callback. You can change the callback at virtually any time	967	Change the callback. You can change the callback at virtually any time
717	(modulo threads).	968	(modulo threads).
		969
		970	=item ev_set_priority (ev_TYPE *watcher, priority)
		971
		972	=item int ev_priority (ev_TYPE *watcher)
		973
		974	Set and query the priority of the watcher. The priority is a small
		975	integer between C<EV_MAXPRI> (default: C<2>) and C<EV_MINPRI>
		976	(default: C<-2>). Pending watchers with higher priority will be invoked
		977	before watchers with lower priority, but priority will not keep watchers
		978	from being executed (except for C<ev_idle> watchers).
		979
		980	This means that priorities are I<only> used for ordering callback
		981	invocation after new events have been received. This is useful, for
		982	example, to reduce latency after idling, or more often, to bind two
		983	watchers on the same event and make sure one is called first.
		984
		985	If you need to suppress invocation when higher priority events are pending
		986	you need to look at C<ev_idle> watchers, which provide this functionality.
		987
		988	You I<must not> change the priority of a watcher as long as it is active or
		989	pending.
		990
		991	The default priority used by watchers when no priority has been set is
		992	always C<0>, which is supposed to not be too high and not be too low :).
		993
		994	Setting a priority outside the range of C<EV_MINPRI> to C<EV_MAXPRI> is
		995	fine, as long as you do not mind that the priority value you query might
		996	or might not have been adjusted to be within valid range.
		997
		998	=item ev_invoke (loop, ev_TYPE *watcher, int revents)
		999
		1000	Invoke the C<watcher> with the given C<loop> and C<revents>. Neither
		1001	C<loop> nor C<revents> need to be valid as long as the watcher callback
		1002	can deal with that fact.
		1003
		1004	=item int ev_clear_pending (loop, ev_TYPE *watcher)
		1005
		1006	If the watcher is pending, this function returns clears its pending status
		1007	and returns its C<revents> bitset (as if its callback was invoked). If the
		1008	watcher isn't pending it does nothing and returns C<0>.
718		1009
719	=back	1010	=back
720		1011
721		1012
722	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER	1013	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER
…		…
726	to associate arbitrary data with your watcher. If you need more data and	1017	to associate arbitrary data with your watcher. If you need more data and
727	don't want to allocate memory and store a pointer to it in that data	1018	don't want to allocate memory and store a pointer to it in that data
728	member, you can also "subclass" the watcher type and provide your own	1019	member, you can also "subclass" the watcher type and provide your own
729	data:	1020	data:
730		1021
731	struct my_io	1022	struct my_io
732	{	1023	{
733	struct ev_io io;	1024	struct ev_io io;
734	int otherfd;	1025	int otherfd;
735	void *somedata;	1026	void *somedata;
736	struct whatever *mostinteresting;	1027	struct whatever *mostinteresting;
737	}	1028	};
		1029
		1030	...
		1031	struct my_io w;
		1032	ev_io_init (&w.io, my_cb, fd, EV_READ);
738		1033
739	And since your callback will be called with a pointer to the watcher, you	1034	And since your callback will be called with a pointer to the watcher, you
740	can cast it back to your own type:	1035	can cast it back to your own type:
741		1036
742	static void my_cb (struct ev_loop *loop, struct ev_io *w_, int revents)	1037	static void my_cb (struct ev_loop *loop, struct ev_io *w_, int revents)
743	{	1038	{
744	struct my_io *w = (struct my_io *)w_;	1039	struct my_io *w = (struct my_io *)w_;
745	...	1040	...
746	}	1041	}
747		1042
748	More interesting and less C-conformant ways of casting your callback type	1043	More interesting and less C-conformant ways of casting your callback type
749	instead have been omitted.	1044	instead have been omitted.
750		1045
751	Another common scenario is having some data structure with multiple	1046	Another common scenario is to use some data structure with multiple
752	watchers:	1047	embedded watchers:
753		1048
754	struct my_biggy	1049	struct my_biggy
755	{	1050	{
756	int some_data;	1051	int some_data;
757	ev_timer t1;	1052	ev_timer t1;
758	ev_timer t2;	1053	ev_timer t2;
759	}	1054	}
760		1055
761	In this case getting the pointer to C<my_biggy> is a bit more complicated,	1056	In this case getting the pointer to C<my_biggy> is a bit more
762	you need to use C<offsetof>:	1057	complicated: Either you store the address of your C<my_biggy> struct
		1058	in the C<data> member of the watcher, or you need to use some pointer
		1059	arithmetic using C<offsetof> inside your watchers:
763		1060
764	#include <stddef.h>	1061	#include <stddef.h>
765		1062
766	static void	1063	static void
767	t1_cb (EV_P_ struct ev_timer *w, int revents)	1064	t1_cb (EV_P_ struct ev_timer *w, int revents)
768	{	1065	{
769	struct my_biggy big = (struct my_biggy *	1066	struct my_biggy big = (struct my_biggy *
770	(((char *)w) - offsetof (struct my_biggy, t1));	1067	(((char *)w) - offsetof (struct my_biggy, t1));
771	}	1068	}
772		1069
773	static void	1070	static void
774	t2_cb (EV_P_ struct ev_timer *w, int revents)	1071	t2_cb (EV_P_ struct ev_timer *w, int revents)
775	{	1072	{
776	struct my_biggy big = (struct my_biggy *	1073	struct my_biggy big = (struct my_biggy *
777	(((char *)w) - offsetof (struct my_biggy, t2));	1074	(((char *)w) - offsetof (struct my_biggy, t2));
778	}	1075	}
779		1076
780		1077
781	=head1 WATCHER TYPES	1078	=head1 WATCHER TYPES
782		1079
783	This section describes each watcher in detail, but will not repeat	1080	This section describes each watcher in detail, but will not repeat
…		…
807	In general you can register as many read and/or write event watchers per	1104	In general you can register as many read and/or write event watchers per
808	fd as you want (as long as you don't confuse yourself). Setting all file	1105	fd as you want (as long as you don't confuse yourself). Setting all file
809	descriptors to non-blocking mode is also usually a good idea (but not	1106	descriptors to non-blocking mode is also usually a good idea (but not
810	required if you know what you are doing).	1107	required if you know what you are doing).
811		1108
812	You have to be careful with dup'ed file descriptors, though. Some backends
813	(the linux epoll backend is a notable example) cannot handle dup'ed file
814	descriptors correctly if you register interest in two or more fds pointing
815	to the same underlying file/socket/etc. description (that is, they share
816	the same underlying "file open").
817
818	If you must do this, then force the use of a known-to-be-good backend	1109	If you must do this, then force the use of a known-to-be-good backend
819	(at the time of this writing, this includes only C<EVBACKEND_SELECT> and	1110	(at the time of this writing, this includes only C<EVBACKEND_SELECT> and
820	C<EVBACKEND_POLL>).	1111	C<EVBACKEND_POLL>).
821		1112
822	Another thing you have to watch out for is that it is quite easy to	1113	Another thing you have to watch out for is that it is quite easy to
823	receive "spurious" readyness notifications, that is your callback might	1114	receive "spurious" readiness notifications, that is your callback might
824	be called with C<EV_READ> but a subsequent C<read>(2) will actually block	1115	be called with C<EV_READ> but a subsequent C<read>(2) will actually block
825	because there is no data. Not only are some backends known to create a	1116	because there is no data. Not only are some backends known to create a
826	lot of those (for example solaris ports), it is very easy to get into	1117	lot of those (for example Solaris ports), it is very easy to get into
827	this situation even with a relatively standard program structure. Thus	1118	this situation even with a relatively standard program structure. Thus
828	it is best to always use non-blocking I/O: An extra C<read>(2) returning	1119	it is best to always use non-blocking I/O: An extra C<read>(2) returning
829	C<EAGAIN> is far preferable to a program hanging until some data arrives.	1120	C<EAGAIN> is far preferable to a program hanging until some data arrives.
830		1121
831	If you cannot run the fd in non-blocking mode (for example you should not	1122	If you cannot run the fd in non-blocking mode (for example you should not
832	play around with an Xlib connection), then you have to seperately re-test	1123	play around with an Xlib connection), then you have to separately re-test
833	wether a file descriptor is really ready with a known-to-be good interface	1124	whether a file descriptor is really ready with a known-to-be good interface
834	such as poll (fortunately in our Xlib example, Xlib already does this on	1125	such as poll (fortunately in our Xlib example, Xlib already does this on
835	its own, so its quite safe to use).	1126	its own, so its quite safe to use).
836		1127
		1128	=head3 The special problem of disappearing file descriptors
		1129
		1130	Some backends (e.g. kqueue, epoll) need to be told about closing a file
		1131	descriptor (either by calling C<close> explicitly or by any other means,
		1132	such as C<dup>). The reason is that you register interest in some file
		1133	descriptor, but when it goes away, the operating system will silently drop
		1134	this interest. If another file descriptor with the same number then is
		1135	registered with libev, there is no efficient way to see that this is, in
		1136	fact, a different file descriptor.
		1137
		1138	To avoid having to explicitly tell libev about such cases, libev follows
		1139	the following policy: Each time C<ev_io_set> is being called, libev
		1140	will assume that this is potentially a new file descriptor, otherwise
		1141	it is assumed that the file descriptor stays the same. That means that
		1142	you I<have> to call C<ev_io_set> (or C<ev_io_init>) when you change the
		1143	descriptor even if the file descriptor number itself did not change.
		1144
		1145	This is how one would do it normally anyway, the important point is that
		1146	the libev application should not optimise around libev but should leave
		1147	optimisations to libev.
		1148
		1149	=head3 The special problem of dup'ed file descriptors
		1150
		1151	Some backends (e.g. epoll), cannot register events for file descriptors,
		1152	but only events for the underlying file descriptions. That means when you
		1153	have C<dup ()>'ed file descriptors or weirder constellations, and register
		1154	events for them, only one file descriptor might actually receive events.
		1155
		1156	There is no workaround possible except not registering events
		1157	for potentially C<dup ()>'ed file descriptors, or to resort to
		1158	C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>.
		1159
		1160	=head3 The special problem of fork
		1161
		1162	Some backends (epoll, kqueue) do not support C<fork ()> at all or exhibit
		1163	useless behaviour. Libev fully supports fork, but needs to be told about
		1164	it in the child.
		1165
		1166	To support fork in your programs, you either have to call
		1167	C<ev_default_fork ()> or C<ev_loop_fork ()> after a fork in the child,
		1168	enable C<EVFLAG_FORKCHECK>, or resort to C<EVBACKEND_SELECT> or
		1169	C<EVBACKEND_POLL>.
		1170
		1171	=head3 The special problem of SIGPIPE
		1172
		1173	While not really specific to libev, it is easy to forget about SIGPIPE:
		1174	when writing to a pipe whose other end has been closed, your program gets
		1175	send a SIGPIPE, which, by default, aborts your program. For most programs
		1176	this is sensible behaviour, for daemons, this is usually undesirable.
		1177
		1178	So when you encounter spurious, unexplained daemon exits, make sure you
		1179	ignore SIGPIPE (and maybe make sure you log the exit status of your daemon
		1180	somewhere, as that would have given you a big clue).
		1181
		1182
		1183	=head3 Watcher-Specific Functions
		1184
837	=over 4	1185	=over 4
838		1186
839	=item ev_io_init (ev_io *, callback, int fd, int events)	1187	=item ev_io_init (ev_io *, callback, int fd, int events)
840		1188
841	=item ev_io_set (ev_io *, int fd, int events)	1189	=item ev_io_set (ev_io *, int fd, int events)
842		1190
843	Configures an C<ev_io> watcher. The C<fd> is the file descriptor to	1191	Configures an C<ev_io> watcher. The C<fd> is the file descriptor to
844	rceeive events for and events is either C<EV_READ>, C<EV_WRITE> or	1192	receive events for and events is either C<EV_READ>, C<EV_WRITE> or
845	C<EV_READ | EV_WRITE> to receive the given events.	1193	C<EV_READ | EV_WRITE> to receive the given events.
846		1194
847	=item int fd [read-only]	1195	=item int fd [read-only]
848		1196
849	The file descriptor being watched.	1197	The file descriptor being watched.
…		…
851	=item int events [read-only]	1199	=item int events [read-only]
852		1200
853	The events being watched.	1201	The events being watched.
854		1202
855	=back	1203	=back
		1204
		1205	=head3 Examples
856		1206
857	Example: Call C<stdin_readable_cb> when STDIN_FILENO has become, well	1207	Example: Call C<stdin_readable_cb> when STDIN_FILENO has become, well
858	readable, but only once. Since it is likely line-buffered, you could	1208	readable, but only once. Since it is likely line-buffered, you could
859	attempt to read a whole line in the callback.	1209	attempt to read a whole line in the callback.
860		1210
861	static void	1211	static void
862	stdin_readable_cb (struct ev_loop *loop, struct ev_io *w, int revents)	1212	stdin_readable_cb (struct ev_loop *loop, struct ev_io *w, int revents)
863	{	1213	{
864	ev_io_stop (loop, w);	1214	ev_io_stop (loop, w);
865	.. read from stdin here (or from w->fd) and haqndle any I/O errors	1215	.. read from stdin here (or from w->fd) and haqndle any I/O errors
866	}	1216	}
867		1217
868	...	1218	...
869	struct ev_loop *loop = ev_default_init (0);	1219	struct ev_loop *loop = ev_default_init (0);
870	struct ev_io stdin_readable;	1220	struct ev_io stdin_readable;
871	ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);	1221	ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
872	ev_io_start (loop, &stdin_readable);	1222	ev_io_start (loop, &stdin_readable);
873	ev_loop (loop, 0);	1223	ev_loop (loop, 0);
874		1224
875		1225
876	=head2 C<ev_timer> - relative and optionally repeating timeouts	1226	=head2 C<ev_timer> - relative and optionally repeating timeouts
877		1227
878	Timer watchers are simple relative timers that generate an event after a	1228	Timer watchers are simple relative timers that generate an event after a
879	given time, and optionally repeating in regular intervals after that.	1229	given time, and optionally repeating in regular intervals after that.
880		1230
881	The timers are based on real time, that is, if you register an event that	1231	The timers are based on real time, that is, if you register an event that
882	times out after an hour and you reset your system clock to last years	1232	times out after an hour and you reset your system clock to January last
883	time, it will still time out after (roughly) and hour. "Roughly" because	1233	year, it will still time out after (roughly) and hour. "Roughly" because
884	detecting time jumps is hard, and some inaccuracies are unavoidable (the	1234	detecting time jumps is hard, and some inaccuracies are unavoidable (the
885	monotonic clock option helps a lot here).	1235	monotonic clock option helps a lot here).
		1236
		1237	The callback is guaranteed to be invoked only after its timeout has passed,
		1238	but if multiple timers become ready during the same loop iteration then
		1239	order of execution is undefined.
		1240
		1241	=head3 The special problem of time updates
		1242
		1243	Establishing the current time is a costly operation (it usually takes at
		1244	least two system calls): EV therefore updates its idea of the current
		1245	time only before and after C<ev_loop> polls for new events, which causes
		1246	a growing difference between C<ev_now ()> and C<ev_time ()> when handling
		1247	lots of events.
886		1248
887	The relative timeouts are calculated relative to the C<ev_now ()>	1249	The relative timeouts are calculated relative to the C<ev_now ()>
888	time. This is usually the right thing as this timestamp refers to the time	1250	time. This is usually the right thing as this timestamp refers to the time
889	of the event triggering whatever timeout you are modifying/starting. If	1251	of the event triggering whatever timeout you are modifying/starting. If
890	you suspect event processing to be delayed and you I<need> to base the timeout	1252	you suspect event processing to be delayed and you I<need> to base the
891	on the current time, use something like this to adjust for this:	1253	timeout on the current time, use something like this to adjust for this:
892		1254
893	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);	1255	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);
894		1256
895	The callback is guarenteed to be invoked only when its timeout has passed,	1257	If the event loop is suspended for a long time, you can also force an
896	but if multiple timers become ready during the same loop iteration then	1258	update of the time returned by C<ev_now ()> by calling C<ev_now_update
897	order of execution is undefined.	1259	()>.
		1260
		1261	=head3 Watcher-Specific Functions and Data Members
898		1262
899	=over 4	1263	=over 4
900		1264
901	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)	1265	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)
902		1266
903	=item ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)	1267	=item ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)
904		1268
905	Configure the timer to trigger after C<after> seconds. If C<repeat> is	1269	Configure the timer to trigger after C<after> seconds. If C<repeat>
906	C<0.>, then it will automatically be stopped. If it is positive, then the	1270	is C<0.>, then it will automatically be stopped once the timeout is
907	timer will automatically be configured to trigger again C<repeat> seconds	1271	reached. If it is positive, then the timer will automatically be
908	later, again, and again, until stopped manually.	1272	configured to trigger again C<repeat> seconds later, again, and again,
		1273	until stopped manually.
909		1274
910	The timer itself will do a best-effort at avoiding drift, that is, if you	1275	The timer itself will do a best-effort at avoiding drift, that is, if
911	configure a timer to trigger every 10 seconds, then it will trigger at	1276	you configure a timer to trigger every 10 seconds, then it will normally
912	exactly 10 second intervals. If, however, your program cannot keep up with	1277	trigger at exactly 10 second intervals. If, however, your program cannot
913	the timer (because it takes longer than those 10 seconds to do stuff) the	1278	keep up with the timer (because it takes longer than those 10 seconds to
914	timer will not fire more than once per event loop iteration.	1279	do stuff) the timer will not fire more than once per event loop iteration.
915		1280
916	=item ev_timer_again (loop)	1281	=item ev_timer_again (loop, ev_timer *)
917		1282
918	This will act as if the timer timed out and restart it again if it is	1283	This will act as if the timer timed out and restart it again if it is
919	repeating. The exact semantics are:	1284	repeating. The exact semantics are:
920		1285
921	If the timer is pending, its pending status is cleared.	1286	If the timer is pending, its pending status is cleared.
922		1287
923	If the timer is started but nonrepeating, stop it (as if it timed out).	1288	If the timer is started but non-repeating, stop it (as if it timed out).
924		1289
925	If the timer is repeating, either start it if necessary (with the	1290	If the timer is repeating, either start it if necessary (with the
926	C<repeat> value), or reset the running timer to the C<repeat> value.	1291	C<repeat> value), or reset the running timer to the C<repeat> value.
927		1292
928	This sounds a bit complicated, but here is a useful and typical	1293	This sounds a bit complicated, but here is a useful and typical
929	example: Imagine you have a tcp connection and you want a so-called idle	1294	example: Imagine you have a TCP connection and you want a so-called idle
930	timeout, that is, you want to be called when there have been, say, 60	1295	timeout, that is, you want to be called when there have been, say, 60
931	seconds of inactivity on the socket. The easiest way to do this is to	1296	seconds of inactivity on the socket. The easiest way to do this is to
932	configure an C<ev_timer> with a C<repeat> value of C<60> and then call	1297	configure an C<ev_timer> with a C<repeat> value of C<60> and then call
933	C<ev_timer_again> each time you successfully read or write some data. If	1298	C<ev_timer_again> each time you successfully read or write some data. If
934	you go into an idle state where you do not expect data to travel on the	1299	you go into an idle state where you do not expect data to travel on the
…		…
956	or C<ev_timer_again> is called and determines the next timeout (if any),	1321	or C<ev_timer_again> is called and determines the next timeout (if any),
957	which is also when any modifications are taken into account.	1322	which is also when any modifications are taken into account.
958		1323
959	=back	1324	=back
960		1325
		1326	=head3 Examples
		1327
961	Example: Create a timer that fires after 60 seconds.	1328	Example: Create a timer that fires after 60 seconds.
962		1329
963	static void	1330	static void
964	one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)	1331	one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
965	{	1332	{
966	.. one minute over, w is actually stopped right here	1333	.. one minute over, w is actually stopped right here
967	}	1334	}
968		1335
969	struct ev_timer mytimer;	1336	struct ev_timer mytimer;
970	ev_timer_init (&mytimer, one_minute_cb, 60., 0.);	1337	ev_timer_init (&mytimer, one_minute_cb, 60., 0.);
971	ev_timer_start (loop, &mytimer);	1338	ev_timer_start (loop, &mytimer);
972		1339
973	Example: Create a timeout timer that times out after 10 seconds of	1340	Example: Create a timeout timer that times out after 10 seconds of
974	inactivity.	1341	inactivity.
975		1342
976	static void	1343	static void
977	timeout_cb (struct ev_loop *loop, struct ev_timer *w, int revents)	1344	timeout_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
978	{	1345	{
979	.. ten seconds without any activity	1346	.. ten seconds without any activity
980	}	1347	}
981		1348
982	struct ev_timer mytimer;	1349	struct ev_timer mytimer;
983	ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */	1350	ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */
984	ev_timer_again (&mytimer); /* start timer */	1351	ev_timer_again (&mytimer); /* start timer */
985	ev_loop (loop, 0);	1352	ev_loop (loop, 0);
986		1353
987	// and in some piece of code that gets executed on any "activity":	1354	// and in some piece of code that gets executed on any "activity":
988	// reset the timeout to start ticking again at 10 seconds	1355	// reset the timeout to start ticking again at 10 seconds
989	ev_timer_again (&mytimer);	1356	ev_timer_again (&mytimer);
990		1357
991		1358
992	=head2 C<ev_periodic> - to cron or not to cron?	1359	=head2 C<ev_periodic> - to cron or not to cron?
993		1360
994	Periodic watchers are also timers of a kind, but they are very versatile	1361	Periodic watchers are also timers of a kind, but they are very versatile
995	(and unfortunately a bit complex).	1362	(and unfortunately a bit complex).
996		1363
997	Unlike C<ev_timer>'s, they are not based on real time (or relative time)	1364	Unlike C<ev_timer>'s, they are not based on real time (or relative time)
998	but on wallclock time (absolute time). You can tell a periodic watcher	1365	but on wall clock time (absolute time). You can tell a periodic watcher
999	to trigger "at" some specific point in time. For example, if you tell a	1366	to trigger after some specific point in time. For example, if you tell a
1000	periodic watcher to trigger in 10 seconds (by specifiying e.g. C<ev_now ()	1367	periodic watcher to trigger in 10 seconds (by specifying e.g. C<ev_now ()
1001	+ 10.>) and then reset your system clock to the last year, then it will	1368	+ 10.>, that is, an absolute time not a delay) and then reset your system
		1369	clock to January of the previous year, then it will take more than year
1002	take a year to trigger the event (unlike an C<ev_timer>, which would trigger	1370	to trigger the event (unlike an C<ev_timer>, which would still trigger
1003	roughly 10 seconds later and of course not if you reset your system time	1371	roughly 10 seconds later as it uses a relative timeout).
1004	again).
1005		1372
1006	They can also be used to implement vastly more complex timers, such as	1373	C<ev_periodic>s can also be used to implement vastly more complex timers,
1007	triggering an event on eahc midnight, local time.	1374	such as triggering an event on each "midnight, local time", or other
		1375	complicated, rules.
1008		1376
1009	As with timers, the callback is guarenteed to be invoked only when the	1377	As with timers, the callback is guaranteed to be invoked only when the
1010	time (C<at>) has been passed, but if multiple periodic timers become ready	1378	time (C<at>) has passed, but if multiple periodic timers become ready
1011	during the same loop iteration then order of execution is undefined.	1379	during the same loop iteration then order of execution is undefined.
		1380
		1381	=head3 Watcher-Specific Functions and Data Members
1012		1382
1013	=over 4	1383	=over 4
1014		1384
1015	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)	1385	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)
1016		1386
…		…
1019	Lots of arguments, lets sort it out... There are basically three modes of	1389	Lots of arguments, lets sort it out... There are basically three modes of
1020	operation, and we will explain them from simplest to complex:	1390	operation, and we will explain them from simplest to complex:
1021		1391
1022	=over 4	1392	=over 4
1023		1393
1024	=item * absolute timer (interval = reschedule_cb = 0)	1394	=item * absolute timer (at = time, interval = reschedule_cb = 0)
1025		1395
1026	In this configuration the watcher triggers an event at the wallclock time	1396	In this configuration the watcher triggers an event after the wall clock
1027	C<at> and doesn't repeat. It will not adjust when a time jump occurs,	1397	time C<at> has passed and doesn't repeat. It will not adjust when a time
1028	that is, if it is to be run at January 1st 2011 then it will run when the	1398	jump occurs, that is, if it is to be run at January 1st 2011 then it will
1029	system time reaches or surpasses this time.	1399	run when the system time reaches or surpasses this time.
1030		1400
1031	=item * non-repeating interval timer (interval > 0, reschedule_cb = 0)	1401	=item * repeating interval timer (at = offset, interval > 0, reschedule_cb = 0)
1032		1402
1033	In this mode the watcher will always be scheduled to time out at the next	1403	In this mode the watcher will always be scheduled to time out at the next
1034	C<at + N * interval> time (for some integer N) and then repeat, regardless	1404	C<at + N * interval> time (for some integer N, which can also be negative)
1035	of any time jumps.	1405	and then repeat, regardless of any time jumps.
1036		1406
1037	This can be used to create timers that do not drift with respect to system	1407	This can be used to create timers that do not drift with respect to system
1038	time:	1408	time, for example, here is a C<ev_periodic> that triggers each hour, on
		1409	the hour:
1039		1410
1040	ev_periodic_set (&periodic, 0., 3600., 0);	1411	ev_periodic_set (&periodic, 0., 3600., 0);
1041		1412
1042	This doesn't mean there will always be 3600 seconds in between triggers,	1413	This doesn't mean there will always be 3600 seconds in between triggers,
1043	but only that the the callback will be called when the system time shows a	1414	but only that the callback will be called when the system time shows a
1044	full hour (UTC), or more correctly, when the system time is evenly divisible	1415	full hour (UTC), or more correctly, when the system time is evenly divisible
1045	by 3600.	1416	by 3600.
1046		1417
1047	Another way to think about it (for the mathematically inclined) is that	1418	Another way to think about it (for the mathematically inclined) is that
1048	C<ev_periodic> will try to run the callback in this mode at the next possible	1419	C<ev_periodic> will try to run the callback in this mode at the next possible
1049	time where C<time = at (mod interval)>, regardless of any time jumps.	1420	time where C<time = at (mod interval)>, regardless of any time jumps.
1050		1421
		1422	For numerical stability it is preferable that the C<at> value is near
		1423	C<ev_now ()> (the current time), but there is no range requirement for
		1424	this value, and in fact is often specified as zero.
		1425
		1426	Note also that there is an upper limit to how often a timer can fire (CPU
		1427	speed for example), so if C<interval> is very small then timing stability
		1428	will of course deteriorate. Libev itself tries to be exact to be about one
		1429	millisecond (if the OS supports it and the machine is fast enough).
		1430
1051	=item * manual reschedule mode (reschedule_cb = callback)	1431	=item * manual reschedule mode (at and interval ignored, reschedule_cb = callback)
1052		1432
1053	In this mode the values for C<interval> and C<at> are both being	1433	In this mode the values for C<interval> and C<at> are both being
1054	ignored. Instead, each time the periodic watcher gets scheduled, the	1434	ignored. Instead, each time the periodic watcher gets scheduled, the
1055	reschedule callback will be called with the watcher as first, and the	1435	reschedule callback will be called with the watcher as first, and the
1056	current time as second argument.	1436	current time as second argument.
1057		1437
1058	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher,	1438	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher,
1059	ever, or make any event loop modifications>. If you need to stop it,	1439	ever, or make ANY event loop modifications whatsoever>.
1060	return C<now + 1e30> (or so, fudge fudge) and stop it afterwards (e.g. by
1061	starting a prepare watcher).
1062		1440
		1441	If you need to stop it, return C<now + 1e30> (or so, fudge fudge) and stop
		1442	it afterwards (e.g. by starting an C<ev_prepare> watcher, which is the
		1443	only event loop modification you are allowed to do).
		1444
1063	Its prototype is C<ev_tstamp (*reschedule_cb)(struct ev_periodic *w,	1445	The callback prototype is C<ev_tstamp (*reschedule_cb)(struct ev_periodic
1064	ev_tstamp now)>, e.g.:	1446	*w, ev_tstamp now)>, e.g.:
1065		1447
1066	static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)	1448	static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)
1067	{	1449	{
1068	return now + 60.;	1450	return now + 60.;
1069	}	1451	}
…		…
1071	It must return the next time to trigger, based on the passed time value	1453	It must return the next time to trigger, based on the passed time value
1072	(that is, the lowest time value larger than to the second argument). It	1454	(that is, the lowest time value larger than to the second argument). It
1073	will usually be called just before the callback will be triggered, but	1455	will usually be called just before the callback will be triggered, but
1074	might be called at other times, too.	1456	might be called at other times, too.
1075		1457
1076	NOTE: I<< This callback must always return a time that is later than the	1458	NOTE: I<< This callback must always return a time that is higher than or
1077	passed C<now> value >>. Not even C<now> itself will do, it I<must> be larger.	1459	equal to the passed C<now> value >>.
1078		1460
1079	This can be used to create very complex timers, such as a timer that	1461	This can be used to create very complex timers, such as a timer that
1080	triggers on each midnight, local time. To do this, you would calculate the	1462	triggers on "next midnight, local time". To do this, you would calculate the
1081	next midnight after C<now> and return the timestamp value for this. How	1463	next midnight after C<now> and return the timestamp value for this. How
1082	you do this is, again, up to you (but it is not trivial, which is the main	1464	you do this is, again, up to you (but it is not trivial, which is the main
1083	reason I omitted it as an example).	1465	reason I omitted it as an example).
1084		1466
1085	=back	1467	=back
…		…
1089	Simply stops and restarts the periodic watcher again. This is only useful	1471	Simply stops and restarts the periodic watcher again. This is only useful
1090	when you changed some parameters or the reschedule callback would return	1472	when you changed some parameters or the reschedule callback would return
1091	a different time than the last time it was called (e.g. in a crond like	1473	a different time than the last time it was called (e.g. in a crond like
1092	program when the crontabs have changed).	1474	program when the crontabs have changed).
1093		1475
		1476	=item ev_tstamp ev_periodic_at (ev_periodic *)
		1477
		1478	When active, returns the absolute time that the watcher is supposed to
		1479	trigger next.
		1480
		1481	=item ev_tstamp offset [read-write]
		1482
		1483	When repeating, this contains the offset value, otherwise this is the
		1484	absolute point in time (the C<at> value passed to C<ev_periodic_set>).
		1485
		1486	Can be modified any time, but changes only take effect when the periodic
		1487	timer fires or C<ev_periodic_again> is being called.
		1488
1094	=item ev_tstamp interval [read-write]	1489	=item ev_tstamp interval [read-write]
1095		1490
1096	The current interval value. Can be modified any time, but changes only	1491	The current interval value. Can be modified any time, but changes only
1097	take effect when the periodic timer fires or C<ev_periodic_again> is being	1492	take effect when the periodic timer fires or C<ev_periodic_again> is being
1098	called.	1493	called.
…		…
1103	switched off. Can be changed any time, but changes only take effect when	1498	switched off. Can be changed any time, but changes only take effect when
1104	the periodic timer fires or C<ev_periodic_again> is being called.	1499	the periodic timer fires or C<ev_periodic_again> is being called.
1105		1500
1106	=back	1501	=back
1107		1502
		1503	=head3 Examples
		1504
1108	Example: Call a callback every hour, or, more precisely, whenever the	1505	Example: Call a callback every hour, or, more precisely, whenever the
1109	system clock is divisible by 3600. The callback invocation times have	1506	system clock is divisible by 3600. The callback invocation times have
1110	potentially a lot of jittering, but good long-term stability.	1507	potentially a lot of jitter, but good long-term stability.
1111		1508
1112	static void	1509	static void
1113	clock_cb (struct ev_loop *loop, struct ev_io *w, int revents)	1510	clock_cb (struct ev_loop *loop, struct ev_io *w, int revents)
1114	{	1511	{
1115	... its now a full hour (UTC, or TAI or whatever your clock follows)	1512	... its now a full hour (UTC, or TAI or whatever your clock follows)
1116	}	1513	}
1117		1514
1118	struct ev_periodic hourly_tick;	1515	struct ev_periodic hourly_tick;
1119	ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);	1516	ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);
1120	ev_periodic_start (loop, &hourly_tick);	1517	ev_periodic_start (loop, &hourly_tick);
1121		1518
1122	Example: The same as above, but use a reschedule callback to do it:	1519	Example: The same as above, but use a reschedule callback to do it:
1123		1520
1124	#include <math.h>	1521	#include <math.h>
1125		1522
1126	static ev_tstamp	1523	static ev_tstamp
1127	my_scheduler_cb (struct ev_periodic *w, ev_tstamp now)	1524	my_scheduler_cb (struct ev_periodic *w, ev_tstamp now)
1128	{	1525	{
1129	return fmod (now, 3600.) + 3600.;	1526	return fmod (now, 3600.) + 3600.;
1130	}	1527	}
1131		1528
1132	ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);	1529	ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
1133		1530
1134	Example: Call a callback every hour, starting now:	1531	Example: Call a callback every hour, starting now:
1135		1532
1136	struct ev_periodic hourly_tick;	1533	struct ev_periodic hourly_tick;
1137	ev_periodic_init (&hourly_tick, clock_cb,	1534	ev_periodic_init (&hourly_tick, clock_cb,
1138	fmod (ev_now (loop), 3600.), 3600., 0);	1535	fmod (ev_now (loop), 3600.), 3600., 0);
1139	ev_periodic_start (loop, &hourly_tick);	1536	ev_periodic_start (loop, &hourly_tick);
1140		1537
1141		1538
1142	=head2 C<ev_signal> - signal me when a signal gets signalled!	1539	=head2 C<ev_signal> - signal me when a signal gets signalled!
1143		1540
1144	Signal watchers will trigger an event when the process receives a specific	1541	Signal watchers will trigger an event when the process receives a specific
…		…
1151	with the kernel (thus it coexists with your own signal handlers as long	1548	with the kernel (thus it coexists with your own signal handlers as long
1152	as you don't register any with libev). Similarly, when the last signal	1549	as you don't register any with libev). Similarly, when the last signal
1153	watcher for a signal is stopped libev will reset the signal handler to	1550	watcher for a signal is stopped libev will reset the signal handler to
1154	SIG_DFL (regardless of what it was set to before).	1551	SIG_DFL (regardless of what it was set to before).
1155		1552
		1553	If possible and supported, libev will install its handlers with
		1554	C<SA_RESTART> behaviour enabled, so system calls should not be unduly
		1555	interrupted. If you have a problem with system calls getting interrupted by
		1556	signals you can block all signals in an C<ev_check> watcher and unblock
		1557	them in an C<ev_prepare> watcher.
		1558
		1559	=head3 Watcher-Specific Functions and Data Members
		1560
1156	=over 4	1561	=over 4
1157		1562
1158	=item ev_signal_init (ev_signal *, callback, int signum)	1563	=item ev_signal_init (ev_signal *, callback, int signum)
1159		1564
1160	=item ev_signal_set (ev_signal *, int signum)	1565	=item ev_signal_set (ev_signal *, int signum)
…		…
1166		1571
1167	The signal the watcher watches out for.	1572	The signal the watcher watches out for.
1168		1573
1169	=back	1574	=back
1170		1575
		1576	=head3 Examples
		1577
		1578	Example: Try to exit cleanly on SIGINT and SIGTERM.
		1579
		1580	static void
		1581	sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)
		1582	{
		1583	ev_unloop (loop, EVUNLOOP_ALL);
		1584	}
		1585
		1586	struct ev_signal signal_watcher;
		1587	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
		1588	ev_signal_start (loop, &sigint_cb);
		1589
1171		1590
1172	=head2 C<ev_child> - watch out for process status changes	1591	=head2 C<ev_child> - watch out for process status changes
1173		1592
1174	Child watchers trigger when your process receives a SIGCHLD in response to	1593	Child watchers trigger when your process receives a SIGCHLD in response to
1175	some child status changes (most typically when a child of yours dies).	1594	some child status changes (most typically when a child of yours dies). It
		1595	is permissible to install a child watcher I<after> the child has been
		1596	forked (which implies it might have already exited), as long as the event
		1597	loop isn't entered (or is continued from a watcher).
		1598
		1599	Only the default event loop is capable of handling signals, and therefore
		1600	you can only register child watchers in the default event loop.
		1601
		1602	=head3 Process Interaction
		1603
		1604	Libev grabs C<SIGCHLD> as soon as the default event loop is
		1605	initialised. This is necessary to guarantee proper behaviour even if
		1606	the first child watcher is started after the child exits. The occurrence
		1607	of C<SIGCHLD> is recorded asynchronously, but child reaping is done
		1608	synchronously as part of the event loop processing. Libev always reaps all
		1609	children, even ones not watched.
		1610
		1611	=head3 Overriding the Built-In Processing
		1612
		1613	Libev offers no special support for overriding the built-in child
		1614	processing, but if your application collides with libev's default child
		1615	handler, you can override it easily by installing your own handler for
		1616	C<SIGCHLD> after initialising the default loop, and making sure the
		1617	default loop never gets destroyed. You are encouraged, however, to use an
		1618	event-based approach to child reaping and thus use libev's support for
		1619	that, so other libev users can use C<ev_child> watchers freely.
		1620
		1621	=head3 Stopping the Child Watcher
		1622
		1623	Currently, the child watcher never gets stopped, even when the
		1624	child terminates, so normally one needs to stop the watcher in the
		1625	callback. Future versions of libev might stop the watcher automatically
		1626	when a child exit is detected.
		1627
		1628	=head3 Watcher-Specific Functions and Data Members
1176		1629
1177	=over 4	1630	=over 4
1178		1631
1179	=item ev_child_init (ev_child *, callback, int pid)	1632	=item ev_child_init (ev_child *, callback, int pid, int trace)
1180		1633
1181	=item ev_child_set (ev_child *, int pid)	1634	=item ev_child_set (ev_child *, int pid, int trace)
1182		1635
1183	Configures the watcher to wait for status changes of process C<pid> (or	1636	Configures the watcher to wait for status changes of process C<pid> (or
1184	I<any> process if C<pid> is specified as C<0>). The callback can look	1637	I<any> process if C<pid> is specified as C<0>). The callback can look
1185	at the C<rstatus> member of the C<ev_child> watcher structure to see	1638	at the C<rstatus> member of the C<ev_child> watcher structure to see
1186	the status word (use the macros from C<sys/wait.h> and see your systems	1639	the status word (use the macros from C<sys/wait.h> and see your systems
1187	C<waitpid> documentation). The C<rpid> member contains the pid of the	1640	C<waitpid> documentation). The C<rpid> member contains the pid of the
1188	process causing the status change.	1641	process causing the status change. C<trace> must be either C<0> (only
		1642	activate the watcher when the process terminates) or C<1> (additionally
		1643	activate the watcher when the process is stopped or continued).
1189		1644
1190	=item int pid [read-only]	1645	=item int pid [read-only]
1191		1646
1192	The process id this watcher watches out for, or C<0>, meaning any process id.	1647	The process id this watcher watches out for, or C<0>, meaning any process id.
1193		1648
…		…
1200	The process exit/trace status caused by C<rpid> (see your systems	1655	The process exit/trace status caused by C<rpid> (see your systems
1201	C<waitpid> and C<sys/wait.h> documentation for details).	1656	C<waitpid> and C<sys/wait.h> documentation for details).
1202		1657
1203	=back	1658	=back
1204		1659
1205	Example: Try to exit cleanly on SIGINT and SIGTERM.	1660	=head3 Examples
1206		1661
		1662	Example: C<fork()> a new process and install a child handler to wait for
		1663	its completion.
		1664
		1665	ev_child cw;
		1666
1207	static void	1667	static void
1208	sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)	1668	child_cb (EV_P_ struct ev_child *w, int revents)
1209	{	1669	{
1210	ev_unloop (loop, EVUNLOOP_ALL);	1670	ev_child_stop (EV_A_ w);
		1671	printf ("process %d exited with status %x\n", w->rpid, w->rstatus);
1211	}	1672	}
1212		1673
1213	struct ev_signal signal_watcher;	1674	pid_t pid = fork ();
1214	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);	1675
1215	ev_signal_start (loop, &sigint_cb);	1676	if (pid < 0)
		1677	// error
		1678	else if (pid == 0)
		1679	{
		1680	// the forked child executes here
		1681	exit (1);
		1682	}
		1683	else
		1684	{
		1685	ev_child_init (&cw, child_cb, pid, 0);
		1686	ev_child_start (EV_DEFAULT_ &cw);
		1687	}
1216		1688
1217		1689
1218	=head2 C<ev_stat> - did the file attributes just change?	1690	=head2 C<ev_stat> - did the file attributes just change?
1219		1691
1220	This watches a filesystem path for attribute changes. That is, it calls	1692	This watches a file system path for attribute changes. That is, it calls
1221	C<stat> regularly (or when the OS says it changed) and sees if it changed	1693	C<stat> regularly (or when the OS says it changed) and sees if it changed
1222	compared to the last time, invoking the callback if it did.	1694	compared to the last time, invoking the callback if it did.
1223		1695
1224	The path does not need to exist: changing from "path exists" to "path does	1696	The path does not need to exist: changing from "path exists" to "path does
1225	not exist" is a status change like any other. The condition "path does	1697	not exist" is a status change like any other. The condition "path does
…		…
1243	as even with OS-supported change notifications, this can be	1715	as even with OS-supported change notifications, this can be
1244	resource-intensive.	1716	resource-intensive.
1245		1717
1246	At the time of this writing, only the Linux inotify interface is	1718	At the time of this writing, only the Linux inotify interface is
1247	implemented (implementing kqueue support is left as an exercise for the	1719	implemented (implementing kqueue support is left as an exercise for the
		1720	reader, note, however, that the author sees no way of implementing ev_stat
1248	reader). Inotify will be used to give hints only and should not change the	1721	semantics with kqueue). Inotify will be used to give hints only and should
1249	semantics of C<ev_stat> watchers, which means that libev sometimes needs	1722	not change the semantics of C<ev_stat> watchers, which means that libev
1250	to fall back to regular polling again even with inotify, but changes are	1723	sometimes needs to fall back to regular polling again even with inotify,
1251	usually detected immediately, and if the file exists there will be no	1724	but changes are usually detected immediately, and if the file exists there
1252	polling.	1725	will be no polling.
		1726
		1727	=head3 ABI Issues (Largefile Support)
		1728
		1729	Libev by default (unless the user overrides this) uses the default
		1730	compilation environment, which means that on systems with large file
		1731	support disabled by default, you get the 32 bit version of the stat
		1732	structure. When using the library from programs that change the ABI to
		1733	use 64 bit file offsets the programs will fail. In that case you have to
		1734	compile libev with the same flags to get binary compatibility. This is
		1735	obviously the case with any flags that change the ABI, but the problem is
		1736	most noticeably disabled with ev_stat and large file support.
		1737
		1738	The solution for this is to lobby your distribution maker to make large
		1739	file interfaces available by default (as e.g. FreeBSD does) and not
		1740	optional. Libev cannot simply switch on large file support because it has
		1741	to exchange stat structures with application programs compiled using the
		1742	default compilation environment.
		1743
		1744	=head3 Inotify
		1745
		1746	When C<inotify (7)> support has been compiled into libev (generally only
		1747	available on Linux) and present at runtime, it will be used to speed up
		1748	change detection where possible. The inotify descriptor will be created lazily
		1749	when the first C<ev_stat> watcher is being started.
		1750
		1751	Inotify presence does not change the semantics of C<ev_stat> watchers
		1752	except that changes might be detected earlier, and in some cases, to avoid
		1753	making regular C<stat> calls. Even in the presence of inotify support
		1754	there are many cases where libev has to resort to regular C<stat> polling.
		1755
		1756	(There is no support for kqueue, as apparently it cannot be used to
		1757	implement this functionality, due to the requirement of having a file
		1758	descriptor open on the object at all times).
		1759
		1760	=head3 The special problem of stat time resolution
		1761
		1762	The C<stat ()> system call only supports full-second resolution portably, and
		1763	even on systems where the resolution is higher, many file systems still
		1764	only support whole seconds.
		1765
		1766	That means that, if the time is the only thing that changes, you can
		1767	easily miss updates: on the first update, C<ev_stat> detects a change and
		1768	calls your callback, which does something. When there is another update
		1769	within the same second, C<ev_stat> will be unable to detect it as the stat
		1770	data does not change.
		1771
		1772	The solution to this is to delay acting on a change for slightly more
		1773	than a second (or till slightly after the next full second boundary), using
		1774	a roughly one-second-delay C<ev_timer> (e.g. C<ev_timer_set (w, 0., 1.02);
		1775	ev_timer_again (loop, w)>).
		1776
		1777	The C<.02> offset is added to work around small timing inconsistencies
		1778	of some operating systems (where the second counter of the current time
		1779	might be be delayed. One such system is the Linux kernel, where a call to
		1780	C<gettimeofday> might return a timestamp with a full second later than
		1781	a subsequent C<time> call - if the equivalent of C<time ()> is used to
		1782	update file times then there will be a small window where the kernel uses
		1783	the previous second to update file times but libev might already execute
		1784	the timer callback).
		1785
		1786	=head3 Watcher-Specific Functions and Data Members
1253		1787
1254	=over 4	1788	=over 4
1255		1789
1256	=item ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)	1790	=item ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)
1257		1791
…		…
1261	C<path>. The C<interval> is a hint on how quickly a change is expected to	1795	C<path>. The C<interval> is a hint on how quickly a change is expected to
1262	be detected and should normally be specified as C<0> to let libev choose	1796	be detected and should normally be specified as C<0> to let libev choose
1263	a suitable value. The memory pointed to by C<path> must point to the same	1797	a suitable value. The memory pointed to by C<path> must point to the same
1264	path for as long as the watcher is active.	1798	path for as long as the watcher is active.
1265		1799
1266	The callback will be receive C<EV_STAT> when a change was detected,	1800	The callback will receive C<EV_STAT> when a change was detected, relative
1267	relative to the attributes at the time the watcher was started (or the	1801	to the attributes at the time the watcher was started (or the last change
1268	last change was detected).	1802	was detected).
1269		1803
1270	=item ev_stat_stat (ev_stat *)	1804	=item ev_stat_stat (loop, ev_stat *)
1271		1805
1272	Updates the stat buffer immediately with new values. If you change the	1806	Updates the stat buffer immediately with new values. If you change the
1273	watched path in your callback, you could call this fucntion to avoid	1807	watched path in your callback, you could call this function to avoid
1274	detecting this change (while introducing a race condition). Can also be	1808	detecting this change (while introducing a race condition if you are not
1275	useful simply to find out the new values.	1809	the only one changing the path). Can also be useful simply to find out the
		1810	new values.
1276		1811
1277	=item ev_statdata attr [read-only]	1812	=item ev_statdata attr [read-only]
1278		1813
1279	The most-recently detected attributes of the file. Although the type is of	1814	The most-recently detected attributes of the file. Although the type is
1280	C<ev_statdata>, this is usually the (or one of the) C<struct stat> types	1815	C<ev_statdata>, this is usually the (or one of the) C<struct stat> types
1281	suitable for your system. If the C<st_nlink> member is C<0>, then there	1816	suitable for your system, but you can only rely on the POSIX-standardised
		1817	members to be present. If the C<st_nlink> member is C<0>, then there was
1282	was some error while C<stat>ing the file.	1818	some error while C<stat>ing the file.
1283		1819
1284	=item ev_statdata prev [read-only]	1820	=item ev_statdata prev [read-only]
1285		1821
1286	The previous attributes of the file. The callback gets invoked whenever	1822	The previous attributes of the file. The callback gets invoked whenever
1287	C<prev> != C<attr>.	1823	C<prev> != C<attr>, or, more precisely, one or more of these members
		1824	differ: C<st_dev>, C<st_ino>, C<st_mode>, C<st_nlink>, C<st_uid>,
		1825	C<st_gid>, C<st_rdev>, C<st_size>, C<st_atime>, C<st_mtime>, C<st_ctime>.
1288		1826
1289	=item ev_tstamp interval [read-only]	1827	=item ev_tstamp interval [read-only]
1290		1828
1291	The specified interval.	1829	The specified interval.
1292		1830
1293	=item const char *path [read-only]	1831	=item const char *path [read-only]
1294		1832
1295	The filesystem path that is being watched.	1833	The file system path that is being watched.
1296		1834
1297	=back	1835	=back
1298		1836
		1837	=head3 Examples
		1838
1299	Example: Watch C</etc/passwd> for attribute changes.	1839	Example: Watch C</etc/passwd> for attribute changes.
1300		1840
1301	static void	1841	static void
1302	passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)	1842	passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)
1303	{	1843	{
1304	/* /etc/passwd changed in some way */	1844	/* /etc/passwd changed in some way */
1305	if (w->attr.st_nlink)	1845	if (w->attr.st_nlink)
1306	{	1846	{
1307	printf ("passwd current size %ld\n", (long)w->attr.st_size);	1847	printf ("passwd current size %ld\n", (long)w->attr.st_size);
1308	printf ("passwd current atime %ld\n", (long)w->attr.st_mtime);	1848	printf ("passwd current atime %ld\n", (long)w->attr.st_mtime);
1309	printf ("passwd current mtime %ld\n", (long)w->attr.st_mtime);	1849	printf ("passwd current mtime %ld\n", (long)w->attr.st_mtime);
1310	}	1850	}
1311	else	1851	else
1312	/* you shalt not abuse printf for puts */	1852	/* you shalt not abuse printf for puts */
1313	puts ("wow, /etc/passwd is not there, expect problems. "	1853	puts ("wow, /etc/passwd is not there, expect problems. "
1314	"if this is windows, they already arrived\n");	1854	"if this is windows, they already arrived\n");
1315	}	1855	}
1316		1856
1317	...	1857	...
1318	ev_stat passwd;	1858	ev_stat passwd;
1319		1859
1320	ev_stat_init (&passwd, passwd_cb, "/etc/passwd");	1860	ev_stat_init (&passwd, passwd_cb, "/etc/passwd", 0.);
1321	ev_stat_start (loop, &passwd);	1861	ev_stat_start (loop, &passwd);
		1862
		1863	Example: Like above, but additionally use a one-second delay so we do not
		1864	miss updates (however, frequent updates will delay processing, too, so
		1865	one might do the work both on C<ev_stat> callback invocation I<and> on
		1866	C<ev_timer> callback invocation).
		1867
		1868	static ev_stat passwd;
		1869	static ev_timer timer;
		1870
		1871	static void
		1872	timer_cb (EV_P_ ev_timer *w, int revents)
		1873	{
		1874	ev_timer_stop (EV_A_ w);
		1875
		1876	/* now it's one second after the most recent passwd change */
		1877	}
		1878
		1879	static void
		1880	stat_cb (EV_P_ ev_stat *w, int revents)
		1881	{
		1882	/* reset the one-second timer */
		1883	ev_timer_again (EV_A_ &timer);
		1884	}
		1885
		1886	...
		1887	ev_stat_init (&passwd, stat_cb, "/etc/passwd", 0.);
		1888	ev_stat_start (loop, &passwd);
		1889	ev_timer_init (&timer, timer_cb, 0., 1.02);
1322		1890
1323		1891
1324	=head2 C<ev_idle> - when you've got nothing better to do...	1892	=head2 C<ev_idle> - when you've got nothing better to do...
1325		1893
1326	Idle watchers trigger events when there are no other events are pending	1894	Idle watchers trigger events when no other events of the same or higher
1327	(prepare, check and other idle watchers do not count). That is, as long	1895	priority are pending (prepare, check and other idle watchers do not
1328	as your process is busy handling sockets or timeouts (or even signals,	1896	count).
1329	imagine) it will not be triggered. But when your process is idle all idle	1897
1330	watchers are being called again and again, once per event loop iteration -	1898	That is, as long as your process is busy handling sockets or timeouts
		1899	(or even signals, imagine) of the same or higher priority it will not be
		1900	triggered. But when your process is idle (or only lower-priority watchers
		1901	are pending), the idle watchers are being called once per event loop
1331	until stopped, that is, or your process receives more events and becomes	1902	iteration - until stopped, that is, or your process receives more events
1332	busy.	1903	and becomes busy again with higher priority stuff.
1333		1904
1334	The most noteworthy effect is that as long as any idle watchers are	1905	The most noteworthy effect is that as long as any idle watchers are
1335	active, the process will not block when waiting for new events.	1906	active, the process will not block when waiting for new events.
1336		1907
1337	Apart from keeping your process non-blocking (which is a useful	1908	Apart from keeping your process non-blocking (which is a useful
1338	effect on its own sometimes), idle watchers are a good place to do	1909	effect on its own sometimes), idle watchers are a good place to do
1339	"pseudo-background processing", or delay processing stuff to after the	1910	"pseudo-background processing", or delay processing stuff to after the
1340	event loop has handled all outstanding events.	1911	event loop has handled all outstanding events.
1341		1912
		1913	=head3 Watcher-Specific Functions and Data Members
		1914
1342	=over 4	1915	=over 4
1343		1916
1344	=item ev_idle_init (ev_signal *, callback)	1917	=item ev_idle_init (ev_signal *, callback)
1345		1918
1346	Initialises and configures the idle watcher - it has no parameters of any	1919	Initialises and configures the idle watcher - it has no parameters of any
1347	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,	1920	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,
1348	believe me.	1921	believe me.
1349		1922
1350	=back	1923	=back
1351		1924
		1925	=head3 Examples
		1926
1352	Example: Dynamically allocate an C<ev_idle> watcher, start it, and in the	1927	Example: Dynamically allocate an C<ev_idle> watcher, start it, and in the
1353	callback, free it. Also, use no error checking, as usual.	1928	callback, free it. Also, use no error checking, as usual.
1354		1929
1355	static void	1930	static void
1356	idle_cb (struct ev_loop *loop, struct ev_idle *w, int revents)	1931	idle_cb (struct ev_loop *loop, struct ev_idle *w, int revents)
1357	{	1932	{
1358	free (w);	1933	free (w);
1359	// now do something you wanted to do when the program has	1934	// now do something you wanted to do when the program has
1360	// no longer asnything immediate to do.	1935	// no longer anything immediate to do.
1361	}	1936	}
1362		1937
1363	struct ev_idle *idle_watcher = malloc (sizeof (struct ev_idle));	1938	struct ev_idle *idle_watcher = malloc (sizeof (struct ev_idle));
1364	ev_idle_init (idle_watcher, idle_cb);	1939	ev_idle_init (idle_watcher, idle_cb);
1365	ev_idle_start (loop, idle_cb);	1940	ev_idle_start (loop, idle_cb);
1366		1941
1367		1942
1368	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!	1943	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!
1369		1944
1370	Prepare and check watchers are usually (but not always) used in tandem:	1945	Prepare and check watchers are usually (but not always) used in tandem:
…		…
1389		1964
1390	This is done by examining in each prepare call which file descriptors need	1965	This is done by examining in each prepare call which file descriptors need
1391	to be watched by the other library, registering C<ev_io> watchers for	1966	to be watched by the other library, registering C<ev_io> watchers for
1392	them and starting an C<ev_timer> watcher for any timeouts (many libraries	1967	them and starting an C<ev_timer> watcher for any timeouts (many libraries
1393	provide just this functionality). Then, in the check watcher you check for	1968	provide just this functionality). Then, in the check watcher you check for
1394	any events that occured (by checking the pending status of all watchers	1969	any events that occurred (by checking the pending status of all watchers
1395	and stopping them) and call back into the library. The I/O and timer	1970	and stopping them) and call back into the library. The I/O and timer
1396	callbacks will never actually be called (but must be valid nevertheless,	1971	callbacks will never actually be called (but must be valid nevertheless,
1397	because you never know, you know?).	1972	because you never know, you know?).
1398		1973
1399	As another example, the Perl Coro module uses these hooks to integrate	1974	As another example, the Perl Coro module uses these hooks to integrate
…		…
1403	with priority higher than or equal to the event loop and one coroutine	1978	with priority higher than or equal to the event loop and one coroutine
1404	of lower priority, but only once, using idle watchers to keep the event	1979	of lower priority, but only once, using idle watchers to keep the event
1405	loop from blocking if lower-priority coroutines are active, thus mapping	1980	loop from blocking if lower-priority coroutines are active, thus mapping
1406	low-priority coroutines to idle/background tasks).	1981	low-priority coroutines to idle/background tasks).
1407		1982
		1983	It is recommended to give C<ev_check> watchers highest (C<EV_MAXPRI>)
		1984	priority, to ensure that they are being run before any other watchers
		1985	after the poll. Also, C<ev_check> watchers (and C<ev_prepare> watchers,
		1986	too) should not activate ("feed") events into libev. While libev fully
		1987	supports this, they might get executed before other C<ev_check> watchers
		1988	did their job. As C<ev_check> watchers are often used to embed other
		1989	(non-libev) event loops those other event loops might be in an unusable
		1990	state until their C<ev_check> watcher ran (always remind yourself to
		1991	coexist peacefully with others).
		1992
		1993	=head3 Watcher-Specific Functions and Data Members
		1994
1408	=over 4	1995	=over 4
1409		1996
1410	=item ev_prepare_init (ev_prepare *, callback)	1997	=item ev_prepare_init (ev_prepare *, callback)
1411		1998
1412	=item ev_check_init (ev_check *, callback)	1999	=item ev_check_init (ev_check *, callback)
…		…
1415	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>	2002	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>
1416	macros, but using them is utterly, utterly and completely pointless.	2003	macros, but using them is utterly, utterly and completely pointless.
1417		2004
1418	=back	2005	=back
1419		2006
1420	Example: To include a library such as adns, you would add IO watchers	2007	=head3 Examples
1421	and a timeout watcher in a prepare handler, as required by libadns, and	2008
		2009	There are a number of principal ways to embed other event loops or modules
		2010	into libev. Here are some ideas on how to include libadns into libev
		2011	(there is a Perl module named C<EV::ADNS> that does this, which you could
		2012	use as a working example. Another Perl module named C<EV::Glib> embeds a
		2013	Glib main context into libev, and finally, C<Glib::EV> embeds EV into the
		2014	Glib event loop).
		2015
		2016	Method 1: Add IO watchers and a timeout watcher in a prepare handler,
1422	in a check watcher, destroy them and call into libadns. What follows is	2017	and in a check watcher, destroy them and call into libadns. What follows
1423	pseudo-code only of course:	2018	is pseudo-code only of course. This requires you to either use a low
		2019	priority for the check watcher or use C<ev_clear_pending> explicitly, as
		2020	the callbacks for the IO/timeout watchers might not have been called yet.
1424		2021
1425	static ev_io iow [nfd];	2022	static ev_io iow [nfd];
1426	static ev_timer tw;	2023	static ev_timer tw;
1427		2024
1428	static void	2025	static void
1429	io_cb (ev_loop *loop, ev_io *w, int revents)	2026	io_cb (ev_loop *loop, ev_io *w, int revents)
1430	{	2027	{
1431	// set the relevant poll flags
1432	// could also call adns_processreadable etc. here
1433	struct pollfd *fd = (struct pollfd *)w->data;
1434	if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;
1435	if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;
1436	}	2028	}
1437		2029
1438	// create io watchers for each fd and a timer before blocking	2030	// create io watchers for each fd and a timer before blocking
1439	static void	2031	static void
1440	adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents)	2032	adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents)
1441	{	2033	{
1442	int timeout = 3600000;truct pollfd fds [nfd];	2034	int timeout = 3600000;
		2035	struct pollfd fds [nfd];
1443	// actual code will need to loop here and realloc etc.	2036	// actual code will need to loop here and realloc etc.
1444	adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));	2037	adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));
1445		2038
1446	/* the callback is illegal, but won't be called as we stop during check */	2039	/* the callback is illegal, but won't be called as we stop during check */
1447	ev_timer_init (&tw, 0, timeout * 1e-3);	2040	ev_timer_init (&tw, 0, timeout * 1e-3);
1448	ev_timer_start (loop, &tw);	2041	ev_timer_start (loop, &tw);
1449		2042
1450	// create on ev_io per pollfd	2043	// create one ev_io per pollfd
1451	for (int i = 0; i < nfd; ++i)	2044	for (int i = 0; i < nfd; ++i)
1452	{	2045	{
1453	ev_io_init (iow + i, io_cb, fds [i].fd,	2046	ev_io_init (iow + i, io_cb, fds [i].fd,
1454	((fds [i].events & POLLIN ? EV_READ : 0)	2047	((fds [i].events & POLLIN ? EV_READ : 0)
1455	| (fds [i].events & POLLOUT ? EV_WRITE : 0)));	2048	| (fds [i].events & POLLOUT ? EV_WRITE : 0)));
1456		2049
1457	fds [i].revents = 0;	2050	fds [i].revents = 0;
1458	iow [i].data = fds + i;
1459	ev_io_start (loop, iow + i);	2051	ev_io_start (loop, iow + i);
1460	}	2052	}
1461	}	2053	}
1462		2054
1463	// stop all watchers after blocking	2055	// stop all watchers after blocking
1464	static void	2056	static void
1465	adns_check_cb (ev_loop *loop, ev_check *w, int revents)	2057	adns_check_cb (ev_loop *loop, ev_check *w, int revents)
1466	{	2058	{
1467	ev_timer_stop (loop, &tw);	2059	ev_timer_stop (loop, &tw);
1468		2060
1469	for (int i = 0; i < nfd; ++i)	2061	for (int i = 0; i < nfd; ++i)
		2062	{
		2063	// set the relevant poll flags
		2064	// could also call adns_processreadable etc. here
		2065	struct pollfd *fd = fds + i;
		2066	int revents = ev_clear_pending (iow + i);
		2067	if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;
		2068	if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;
		2069
		2070	// now stop the watcher
1470	ev_io_stop (loop, iow + i);	2071	ev_io_stop (loop, iow + i);
		2072	}
1471		2073
1472	adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));	2074	adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));
1473	}	2075	}
		2076
		2077	Method 2: This would be just like method 1, but you run C<adns_afterpoll>
		2078	in the prepare watcher and would dispose of the check watcher.
		2079
		2080	Method 3: If the module to be embedded supports explicit event
		2081	notification (libadns does), you can also make use of the actual watcher
		2082	callbacks, and only destroy/create the watchers in the prepare watcher.
		2083
		2084	static void
		2085	timer_cb (EV_P_ ev_timer *w, int revents)
		2086	{
		2087	adns_state ads = (adns_state)w->data;
		2088	update_now (EV_A);
		2089
		2090	adns_processtimeouts (ads, &tv_now);
		2091	}
		2092
		2093	static void
		2094	io_cb (EV_P_ ev_io *w, int revents)
		2095	{
		2096	adns_state ads = (adns_state)w->data;
		2097	update_now (EV_A);
		2098
		2099	if (revents & EV_READ ) adns_processreadable (ads, w->fd, &tv_now);
		2100	if (revents & EV_WRITE) adns_processwriteable (ads, w->fd, &tv_now);
		2101	}
		2102
		2103	// do not ever call adns_afterpoll
		2104
		2105	Method 4: Do not use a prepare or check watcher because the module you
		2106	want to embed is too inflexible to support it. Instead, you can override
		2107	their poll function. The drawback with this solution is that the main
		2108	loop is now no longer controllable by EV. The C<Glib::EV> module does
		2109	this.
		2110
		2111	static gint
		2112	event_poll_func (GPollFD *fds, guint nfds, gint timeout)
		2113	{
		2114	int got_events = 0;
		2115
		2116	for (n = 0; n < nfds; ++n)
		2117	// create/start io watcher that sets the relevant bits in fds[n] and increment got_events
		2118
		2119	if (timeout >= 0)
		2120	// create/start timer
		2121
		2122	// poll
		2123	ev_loop (EV_A_ 0);
		2124
		2125	// stop timer again
		2126	if (timeout >= 0)
		2127	ev_timer_stop (EV_A_ &to);
		2128
		2129	// stop io watchers again - their callbacks should have set
		2130	for (n = 0; n < nfds; ++n)
		2131	ev_io_stop (EV_A_ iow [n]);
		2132
		2133	return got_events;
		2134	}
1474		2135
1475		2136
1476	=head2 C<ev_embed> - when one backend isn't enough...	2137	=head2 C<ev_embed> - when one backend isn't enough...
1477		2138
1478	This is a rather advanced watcher type that lets you embed one event loop	2139	This is a rather advanced watcher type that lets you embed one event loop
…		…
1520	portable one.	2181	portable one.
1521		2182
1522	So when you want to use this feature you will always have to be prepared	2183	So when you want to use this feature you will always have to be prepared
1523	that you cannot get an embeddable loop. The recommended way to get around	2184	that you cannot get an embeddable loop. The recommended way to get around
1524	this is to have a separate variables for your embeddable loop, try to	2185	this is to have a separate variables for your embeddable loop, try to
1525	create it, and if that fails, use the normal loop for everything:	2186	create it, and if that fails, use the normal loop for everything.
1526		2187
1527	struct ev_loop *loop_hi = ev_default_init (0);	2188	=head3 Watcher-Specific Functions and Data Members
1528	struct ev_loop *loop_lo = 0;
1529	struct ev_embed embed;
1530
1531	// see if there is a chance of getting one that works
1532	// (remember that a flags value of 0 means autodetection)
1533	loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
1534	? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
1535	: 0;
1536
1537	// if we got one, then embed it, otherwise default to loop_hi
1538	if (loop_lo)
1539	{
1540	ev_embed_init (&embed, 0, loop_lo);
1541	ev_embed_start (loop_hi, &embed);
1542	}
1543	else
1544	loop_lo = loop_hi;
1545		2189
1546	=over 4	2190	=over 4
1547		2191
1548	=item ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)	2192	=item ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)
1549		2193
…		…
1551		2195
1552	Configures the watcher to embed the given loop, which must be	2196	Configures the watcher to embed the given loop, which must be
1553	embeddable. If the callback is C<0>, then C<ev_embed_sweep> will be	2197	embeddable. If the callback is C<0>, then C<ev_embed_sweep> will be
1554	invoked automatically, otherwise it is the responsibility of the callback	2198	invoked automatically, otherwise it is the responsibility of the callback
1555	to invoke it (it will continue to be called until the sweep has been done,	2199	to invoke it (it will continue to be called until the sweep has been done,
1556	if you do not want thta, you need to temporarily stop the embed watcher).	2200	if you do not want that, you need to temporarily stop the embed watcher).
1557		2201
1558	=item ev_embed_sweep (loop, ev_embed *)	2202	=item ev_embed_sweep (loop, ev_embed *)
1559		2203
1560	Make a single, non-blocking sweep over the embedded loop. This works	2204	Make a single, non-blocking sweep over the embedded loop. This works
1561	similarly to C<ev_loop (embedded_loop, EVLOOP_NONBLOCK)>, but in the most	2205	similarly to C<ev_loop (embedded_loop, EVLOOP_NONBLOCK)>, but in the most
1562	apropriate way for embedded loops.	2206	appropriate way for embedded loops.
1563		2207
1564	=item struct ev_loop *loop [read-only]	2208	=item struct ev_loop *other [read-only]
1565		2209
1566	The embedded event loop.	2210	The embedded event loop.
1567		2211
1568	=back	2212	=back
		2213
		2214	=head3 Examples
		2215
		2216	Example: Try to get an embeddable event loop and embed it into the default
		2217	event loop. If that is not possible, use the default loop. The default
		2218	loop is stored in C<loop_hi>, while the embeddable loop is stored in
		2219	C<loop_lo> (which is C<loop_hi> in the case no embeddable loop can be
		2220	used).
		2221
		2222	struct ev_loop *loop_hi = ev_default_init (0);
		2223	struct ev_loop *loop_lo = 0;
		2224	struct ev_embed embed;
		2225
		2226	// see if there is a chance of getting one that works
		2227	// (remember that a flags value of 0 means autodetection)
		2228	loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
		2229	? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
		2230	: 0;
		2231
		2232	// if we got one, then embed it, otherwise default to loop_hi
		2233	if (loop_lo)
		2234	{
		2235	ev_embed_init (&embed, 0, loop_lo);
		2236	ev_embed_start (loop_hi, &embed);
		2237	}
		2238	else
		2239	loop_lo = loop_hi;
		2240
		2241	Example: Check if kqueue is available but not recommended and create
		2242	a kqueue backend for use with sockets (which usually work with any
		2243	kqueue implementation). Store the kqueue/socket-only event loop in
		2244	C<loop_socket>. (One might optionally use C<EVFLAG_NOENV>, too).
		2245
		2246	struct ev_loop *loop = ev_default_init (0);
		2247	struct ev_loop *loop_socket = 0;
		2248	struct ev_embed embed;
		2249
		2250	if (ev_supported_backends () & ~ev_recommended_backends () & EVBACKEND_KQUEUE)
		2251	if ((loop_socket = ev_loop_new (EVBACKEND_KQUEUE))
		2252	{
		2253	ev_embed_init (&embed, 0, loop_socket);
		2254	ev_embed_start (loop, &embed);
		2255	}
		2256
		2257	if (!loop_socket)
		2258	loop_socket = loop;
		2259
		2260	// now use loop_socket for all sockets, and loop for everything else
1569		2261
1570		2262
1571	=head2 C<ev_fork> - the audacity to resume the event loop after a fork	2263	=head2 C<ev_fork> - the audacity to resume the event loop after a fork
1572		2264
1573	Fork watchers are called when a C<fork ()> was detected (usually because	2265	Fork watchers are called when a C<fork ()> was detected (usually because
…		…
1576	event loop blocks next and before C<ev_check> watchers are being called,	2268	event loop blocks next and before C<ev_check> watchers are being called,
1577	and only in the child after the fork. If whoever good citizen calling	2269	and only in the child after the fork. If whoever good citizen calling
1578	C<ev_default_fork> cheats and calls it in the wrong process, the fork	2270	C<ev_default_fork> cheats and calls it in the wrong process, the fork
1579	handlers will be invoked, too, of course.	2271	handlers will be invoked, too, of course.
1580		2272
		2273	=head3 Watcher-Specific Functions and Data Members
		2274
1581	=over 4	2275	=over 4
1582		2276
1583	=item ev_fork_init (ev_signal *, callback)	2277	=item ev_fork_init (ev_signal *, callback)
1584		2278
1585	Initialises and configures the fork watcher - it has no parameters of any	2279	Initialises and configures the fork watcher - it has no parameters of any
1586	kind. There is a C<ev_fork_set> macro, but using it is utterly pointless,	2280	kind. There is a C<ev_fork_set> macro, but using it is utterly pointless,
1587	believe me.	2281	believe me.
		2282
		2283	=back
		2284
		2285
		2286	=head2 C<ev_async> - how to wake up another event loop
		2287
		2288	In general, you cannot use an C<ev_loop> from multiple threads or other
		2289	asynchronous sources such as signal handlers (as opposed to multiple event
		2290	loops - those are of course safe to use in different threads).
		2291
		2292	Sometimes, however, you need to wake up another event loop you do not
		2293	control, for example because it belongs to another thread. This is what
		2294	C<ev_async> watchers do: as long as the C<ev_async> watcher is active, you
		2295	can signal it by calling C<ev_async_send>, which is thread- and signal
		2296	safe.
		2297
		2298	This functionality is very similar to C<ev_signal> watchers, as signals,
		2299	too, are asynchronous in nature, and signals, too, will be compressed
		2300	(i.e. the number of callback invocations may be less than the number of
		2301	C<ev_async_sent> calls).
		2302
		2303	Unlike C<ev_signal> watchers, C<ev_async> works with any event loop, not
		2304	just the default loop.
		2305
		2306	=head3 Queueing
		2307
		2308	C<ev_async> does not support queueing of data in any way. The reason
		2309	is that the author does not know of a simple (or any) algorithm for a
		2310	multiple-writer-single-reader queue that works in all cases and doesn't
		2311	need elaborate support such as pthreads.
		2312
		2313	That means that if you want to queue data, you have to provide your own
		2314	queue. But at least I can tell you would implement locking around your
		2315	queue:
		2316
		2317	=over 4
		2318
		2319	=item queueing from a signal handler context
		2320
		2321	To implement race-free queueing, you simply add to the queue in the signal
		2322	handler but you block the signal handler in the watcher callback. Here is an example that does that for
		2323	some fictitious SIGUSR1 handler:
		2324
		2325	static ev_async mysig;
		2326
		2327	static void
		2328	sigusr1_handler (void)
		2329	{
		2330	sometype data;
		2331
		2332	// no locking etc.
		2333	queue_put (data);
		2334	ev_async_send (EV_DEFAULT_ &mysig);
		2335	}
		2336
		2337	static void
		2338	mysig_cb (EV_P_ ev_async *w, int revents)
		2339	{
		2340	sometype data;
		2341	sigset_t block, prev;
		2342
		2343	sigemptyset (&block);
		2344	sigaddset (&block, SIGUSR1);
		2345	sigprocmask (SIG_BLOCK, &block, &prev);
		2346
		2347	while (queue_get (&data))
		2348	process (data);
		2349
		2350	if (sigismember (&prev, SIGUSR1)
		2351	sigprocmask (SIG_UNBLOCK, &block, 0);
		2352	}
		2353
		2354	(Note: pthreads in theory requires you to use C<pthread_setmask>
		2355	instead of C<sigprocmask> when you use threads, but libev doesn't do it
		2356	either...).
		2357
		2358	=item queueing from a thread context
		2359
		2360	The strategy for threads is different, as you cannot (easily) block
		2361	threads but you can easily preempt them, so to queue safely you need to
		2362	employ a traditional mutex lock, such as in this pthread example:
		2363
		2364	static ev_async mysig;
		2365	static pthread_mutex_t mymutex = PTHREAD_MUTEX_INITIALIZER;
		2366
		2367	static void
		2368	otherthread (void)
		2369	{
		2370	// only need to lock the actual queueing operation
		2371	pthread_mutex_lock (&mymutex);
		2372	queue_put (data);
		2373	pthread_mutex_unlock (&mymutex);
		2374
		2375	ev_async_send (EV_DEFAULT_ &mysig);
		2376	}
		2377
		2378	static void
		2379	mysig_cb (EV_P_ ev_async *w, int revents)
		2380	{
		2381	pthread_mutex_lock (&mymutex);
		2382
		2383	while (queue_get (&data))
		2384	process (data);
		2385
		2386	pthread_mutex_unlock (&mymutex);
		2387	}
		2388
		2389	=back
		2390
		2391
		2392	=head3 Watcher-Specific Functions and Data Members
		2393
		2394	=over 4
		2395
		2396	=item ev_async_init (ev_async *, callback)
		2397
		2398	Initialises and configures the async watcher - it has no parameters of any
		2399	kind. There is a C<ev_asynd_set> macro, but using it is utterly pointless,
		2400	believe me.
		2401
		2402	=item ev_async_send (loop, ev_async *)
		2403
		2404	Sends/signals/activates the given C<ev_async> watcher, that is, feeds
		2405	an C<EV_ASYNC> event on the watcher into the event loop. Unlike
		2406	C<ev_feed_event>, this call is safe to do in other threads, signal or
		2407	similar contexts (see the discussion of C<EV_ATOMIC_T> in the embedding
		2408	section below on what exactly this means).
		2409
		2410	This call incurs the overhead of a system call only once per loop iteration,
		2411	so while the overhead might be noticeable, it doesn't apply to repeated
		2412	calls to C<ev_async_send>.
		2413
		2414	=item bool = ev_async_pending (ev_async *)
		2415
		2416	Returns a non-zero value when C<ev_async_send> has been called on the
		2417	watcher but the event has not yet been processed (or even noted) by the
		2418	event loop.
		2419
		2420	C<ev_async_send> sets a flag in the watcher and wakes up the loop. When
		2421	the loop iterates next and checks for the watcher to have become active,
		2422	it will reset the flag again. C<ev_async_pending> can be used to very
		2423	quickly check whether invoking the loop might be a good idea.
		2424
		2425	Not that this does I<not> check whether the watcher itself is pending, only
		2426	whether it has been requested to make this watcher pending.
1588		2427
1589	=back	2428	=back
1590		2429
1591		2430
1592	=head1 OTHER FUNCTIONS	2431	=head1 OTHER FUNCTIONS
…		…
1603	or timeout without having to allocate/configure/start/stop/free one or	2442	or timeout without having to allocate/configure/start/stop/free one or
1604	more watchers yourself.	2443	more watchers yourself.
1605		2444
1606	If C<fd> is less than 0, then no I/O watcher will be started and events	2445	If C<fd> is less than 0, then no I/O watcher will be started and events
1607	is being ignored. Otherwise, an C<ev_io> watcher for the given C<fd> and	2446	is being ignored. Otherwise, an C<ev_io> watcher for the given C<fd> and
1608	C<events> set will be craeted and started.	2447	C<events> set will be created and started.
1609		2448
1610	If C<timeout> is less than 0, then no timeout watcher will be	2449	If C<timeout> is less than 0, then no timeout watcher will be
1611	started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and	2450	started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and
1612	repeat = 0) will be started. While C<0> is a valid timeout, it is of	2451	repeat = 0) will be started. While C<0> is a valid timeout, it is of
1613	dubious value.	2452	dubious value.
…		…
1615	The callback has the type C<void (*cb)(int revents, void *arg)> and gets	2454	The callback has the type C<void (*cb)(int revents, void *arg)> and gets
1616	passed an C<revents> set like normal event callbacks (a combination of	2455	passed an C<revents> set like normal event callbacks (a combination of
1617	C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMEOUT>) and the C<arg>	2456	C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMEOUT>) and the C<arg>
1618	value passed to C<ev_once>:	2457	value passed to C<ev_once>:
1619		2458
1620	static void stdin_ready (int revents, void *arg)	2459	static void stdin_ready (int revents, void *arg)
1621	{	2460	{
1622	if (revents & EV_TIMEOUT)	2461	if (revents & EV_TIMEOUT)
1623	/* doh, nothing entered */;	2462	/* doh, nothing entered */;
1624	else if (revents & EV_READ)	2463	else if (revents & EV_READ)
1625	/* stdin might have data for us, joy! */;	2464	/* stdin might have data for us, joy! */;
1626	}	2465	}
1627		2466
1628	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);	2467	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
1629		2468
1630	=item ev_feed_event (ev_loop *, watcher *, int revents)	2469	=item ev_feed_event (ev_loop *, watcher *, int revents)
1631		2470
1632	Feeds the given event set into the event loop, as if the specified event	2471	Feeds the given event set into the event loop, as if the specified event
1633	had happened for the specified watcher (which must be a pointer to an	2472	had happened for the specified watcher (which must be a pointer to an
…		…
1638	Feed an event on the given fd, as if a file descriptor backend detected	2477	Feed an event on the given fd, as if a file descriptor backend detected
1639	the given events it.	2478	the given events it.
1640		2479
1641	=item ev_feed_signal_event (ev_loop *loop, int signum)	2480	=item ev_feed_signal_event (ev_loop *loop, int signum)
1642		2481
1643	Feed an event as if the given signal occured (C<loop> must be the default	2482	Feed an event as if the given signal occurred (C<loop> must be the default
1644	loop!).	2483	loop!).
1645		2484
1646	=back	2485	=back
1647		2486
1648		2487
…		…
1664		2503
1665	=item * Priorities are not currently supported. Initialising priorities	2504	=item * Priorities are not currently supported. Initialising priorities
1666	will fail and all watchers will have the same priority, even though there	2505	will fail and all watchers will have the same priority, even though there
1667	is an ev_pri field.	2506	is an ev_pri field.
1668		2507
		2508	=item * In libevent, the last base created gets the signals, in libev, the
		2509	first base created (== the default loop) gets the signals.
		2510
1669	=item * Other members are not supported.	2511	=item * Other members are not supported.
1670		2512
1671	=item * The libev emulation is I<not> ABI compatible to libevent, you need	2513	=item * The libev emulation is I<not> ABI compatible to libevent, you need
1672	to use the libev header file and library.	2514	to use the libev header file and library.
1673		2515
1674	=back	2516	=back
1675		2517
1676	=head1 C++ SUPPORT	2518	=head1 C++ SUPPORT
1677		2519
1678	Libev comes with some simplistic wrapper classes for C++ that mainly allow	2520	Libev comes with some simplistic wrapper classes for C++ that mainly allow
1679	you to use some convinience methods to start/stop watchers and also change	2521	you to use some convenience methods to start/stop watchers and also change
1680	the callback model to a model using method callbacks on objects.	2522	the callback model to a model using method callbacks on objects.
1681		2523
1682	To use it,	2524	To use it,
1683		2525
1684	#include <ev++.h>	2526	#include <ev++.h>
1685		2527
1686	(it is not installed by default). This automatically includes F<ev.h>	2528	This automatically includes F<ev.h> and puts all of its definitions (many
1687	and puts all of its definitions (many of them macros) into the global	2529	of them macros) into the global namespace. All C++ specific things are
1688	namespace. All C++ specific things are put into the C<ev> namespace.	2530	put into the C<ev> namespace. It should support all the same embedding
		2531	options as F<ev.h>, most notably C<EV_MULTIPLICITY>.
1689		2532
1690	It should support all the same embedding options as F<ev.h>, most notably	2533	Care has been taken to keep the overhead low. The only data member the C++
1691	C<EV_MULTIPLICITY>.	2534	classes add (compared to plain C-style watchers) is the event loop pointer
		2535	that the watcher is associated with (or no additional members at all if
		2536	you disable C<EV_MULTIPLICITY> when embedding libev).
		2537
		2538	Currently, functions, and static and non-static member functions can be
		2539	used as callbacks. Other types should be easy to add as long as they only
		2540	need one additional pointer for context. If you need support for other
		2541	types of functors please contact the author (preferably after implementing
		2542	it).
1692		2543
1693	Here is a list of things available in the C<ev> namespace:	2544	Here is a list of things available in the C<ev> namespace:
1694		2545
1695	=over 4	2546	=over 4
1696		2547
…		…
1712		2563
1713	All of those classes have these methods:	2564	All of those classes have these methods:
1714		2565
1715	=over 4	2566	=over 4
1716		2567
1717	=item ev::TYPE::TYPE (object *, object::method *)	2568	=item ev::TYPE::TYPE ()
1718		2569
1719	=item ev::TYPE::TYPE (object *, object::method *, struct ev_loop *)	2570	=item ev::TYPE::TYPE (struct ev_loop *)
1720		2571
1721	=item ev::TYPE::~TYPE	2572	=item ev::TYPE::~TYPE
1722		2573
1723	The constructor takes a pointer to an object and a method pointer to	2574	The constructor (optionally) takes an event loop to associate the watcher
1724	the event handler callback to call in this class. The constructor calls	2575	with. If it is omitted, it will use C<EV_DEFAULT>.
1725	C<ev_init> for you, which means you have to call the C<set> method	2576
1726	before starting it. If you do not specify a loop then the constructor	2577	The constructor calls C<ev_init> for you, which means you have to call the
1727	automatically associates the default loop with this watcher.	2578	C<set> method before starting it.
		2579
		2580	It will not set a callback, however: You have to call the templated C<set>
		2581	method to set a callback before you can start the watcher.
		2582
		2583	(The reason why you have to use a method is a limitation in C++ which does
		2584	not allow explicit template arguments for constructors).
1728		2585
1729	The destructor automatically stops the watcher if it is active.	2586	The destructor automatically stops the watcher if it is active.
		2587
		2588	=item w->set<class, &class::method> (object *)
		2589
		2590	This method sets the callback method to call. The method has to have a
		2591	signature of C<void (*)(ev_TYPE &, int)>, it receives the watcher as
		2592	first argument and the C<revents> as second. The object must be given as
		2593	parameter and is stored in the C<data> member of the watcher.
		2594
		2595	This method synthesizes efficient thunking code to call your method from
		2596	the C callback that libev requires. If your compiler can inline your
		2597	callback (i.e. it is visible to it at the place of the C<set> call and
		2598	your compiler is good :), then the method will be fully inlined into the
		2599	thunking function, making it as fast as a direct C callback.
		2600
		2601	Example: simple class declaration and watcher initialisation
		2602
		2603	struct myclass
		2604	{
		2605	void io_cb (ev::io &w, int revents) { }
		2606	}
		2607
		2608	myclass obj;
		2609	ev::io iow;
		2610	iow.set <myclass, &myclass::io_cb> (&obj);
		2611
		2612	=item w->set<function> (void *data = 0)
		2613
		2614	Also sets a callback, but uses a static method or plain function as
		2615	callback. The optional C<data> argument will be stored in the watcher's
		2616	C<data> member and is free for you to use.
		2617
		2618	The prototype of the C<function> must be C<void (*)(ev::TYPE &w, int)>.
		2619
		2620	See the method-C<set> above for more details.
		2621
		2622	Example:
		2623
		2624	static void io_cb (ev::io &w, int revents) { }
		2625	iow.set <io_cb> ();
1730		2626
1731	=item w->set (struct ev_loop *)	2627	=item w->set (struct ev_loop *)
1732		2628
1733	Associates a different C<struct ev_loop> with this watcher. You can only	2629	Associates a different C<struct ev_loop> with this watcher. You can only
1734	do this when the watcher is inactive (and not pending either).	2630	do this when the watcher is inactive (and not pending either).
1735		2631
1736	=item w->set ([args])	2632	=item w->set ([arguments])
1737		2633
1738	Basically the same as C<ev_TYPE_set>, with the same args. Must be	2634	Basically the same as C<ev_TYPE_set>, with the same arguments. Must be
1739	called at least once. Unlike the C counterpart, an active watcher gets	2635	called at least once. Unlike the C counterpart, an active watcher gets
1740	automatically stopped and restarted.	2636	automatically stopped and restarted when reconfiguring it with this
		2637	method.
1741		2638
1742	=item w->start ()	2639	=item w->start ()
1743		2640
1744	Starts the watcher. Note that there is no C<loop> argument as the	2641	Starts the watcher. Note that there is no C<loop> argument, as the
1745	constructor already takes the loop.	2642	constructor already stores the event loop.
1746		2643
1747	=item w->stop ()	2644	=item w->stop ()
1748		2645
1749	Stops the watcher if it is active. Again, no C<loop> argument.	2646	Stops the watcher if it is active. Again, no C<loop> argument.
1750		2647
1751	=item w->again () C<ev::timer>, C<ev::periodic> only	2648	=item w->again () (C<ev::timer>, C<ev::periodic> only)
1752		2649
1753	For C<ev::timer> and C<ev::periodic>, this invokes the corresponding	2650	For C<ev::timer> and C<ev::periodic>, this invokes the corresponding
1754	C<ev_TYPE_again> function.	2651	C<ev_TYPE_again> function.
1755		2652
1756	=item w->sweep () C<ev::embed> only	2653	=item w->sweep () (C<ev::embed> only)
1757		2654
1758	Invokes C<ev_embed_sweep>.	2655	Invokes C<ev_embed_sweep>.
1759		2656
1760	=item w->update () C<ev::stat> only	2657	=item w->update () (C<ev::stat> only)
1761		2658
1762	Invokes C<ev_stat_stat>.	2659	Invokes C<ev_stat_stat>.
1763		2660
1764	=back	2661	=back
1765		2662
1766	=back	2663	=back
1767		2664
1768	Example: Define a class with an IO and idle watcher, start one of them in	2665	Example: Define a class with an IO and idle watcher, start one of them in
1769	the constructor.	2666	the constructor.
1770		2667
1771	class myclass	2668	class myclass
1772	{	2669	{
1773	ev_io io; void io_cb (ev::io &w, int revents);	2670	ev::io io; void io_cb (ev::io &w, int revents);
1774	ev_idle idle void idle_cb (ev::idle &w, int revents);	2671	ev:idle idle void idle_cb (ev::idle &w, int revents);
1775		2672
1776	myclass ();	2673	myclass (int fd)
1777	}	2674	{
		2675	io .set <myclass, &myclass::io_cb > (this);
		2676	idle.set <myclass, &myclass::idle_cb> (this);
1778		2677
1779	myclass::myclass (int fd)
1780	: io (this, &myclass::io_cb),
1781	idle (this, &myclass::idle_cb)
1782	{
1783	io.start (fd, ev::READ);	2678	io.start (fd, ev::READ);
		2679	}
1784	}	2680	};
		2681
		2682
		2683	=head1 OTHER LANGUAGE BINDINGS
		2684
		2685	Libev does not offer other language bindings itself, but bindings for a
		2686	number of languages exist in the form of third-party packages. If you know
		2687	any interesting language binding in addition to the ones listed here, drop
		2688	me a note.
		2689
		2690	=over 4
		2691
		2692	=item Perl
		2693
		2694	The EV module implements the full libev API and is actually used to test
		2695	libev. EV is developed together with libev. Apart from the EV core module,
		2696	there are additional modules that implement libev-compatible interfaces
		2697	to C<libadns> (C<EV::ADNS>), C<Net::SNMP> (C<Net::SNMP::EV>) and the
		2698	C<libglib> event core (C<Glib::EV> and C<EV::Glib>).
		2699
		2700	It can be found and installed via CPAN, its homepage is at
		2701	L<http://software.schmorp.de/pkg/EV>.
		2702
		2703	=item Python
		2704
		2705	Python bindings can be found at L<http://code.google.com/p/pyev/>. It
		2706	seems to be quite complete and well-documented. Note, however, that the
		2707	patch they require for libev is outright dangerous as it breaks the ABI
		2708	for everybody else, and therefore, should never be applied in an installed
		2709	libev (if python requires an incompatible ABI then it needs to embed
		2710	libev).
		2711
		2712	=item Ruby
		2713
		2714	Tony Arcieri has written a ruby extension that offers access to a subset
		2715	of the libev API and adds file handle abstractions, asynchronous DNS and
		2716	more on top of it. It can be found via gem servers. Its homepage is at
		2717	L<http://rev.rubyforge.org/>.
		2718
		2719	=item D
		2720
		2721	Leandro Lucarella has written a D language binding (F<ev.d>) for libev, to
		2722	be found at L<http://proj.llucax.com.ar/wiki/evd>.
		2723
		2724	=back
1785		2725
1786		2726
1787	=head1 MACRO MAGIC	2727	=head1 MACRO MAGIC
1788		2728
1789	Libev can be compiled with a variety of options, the most fundemantal is	2729	Libev can be compiled with a variety of options, the most fundamental
1790	C<EV_MULTIPLICITY>. This option determines wether (most) functions and	2730	of which is C<EV_MULTIPLICITY>. This option determines whether (most)
1791	callbacks have an initial C<struct ev_loop *> argument.	2731	functions and callbacks have an initial C<struct ev_loop *> argument.
1792		2732
1793	To make it easier to write programs that cope with either variant, the	2733	To make it easier to write programs that cope with either variant, the
1794	following macros are defined:	2734	following macros are defined:
1795		2735
1796	=over 4	2736	=over 4
…		…
1799		2739
1800	This provides the loop I<argument> for functions, if one is required ("ev	2740	This provides the loop I<argument> for functions, if one is required ("ev
1801	loop argument"). The C<EV_A> form is used when this is the sole argument,	2741	loop argument"). The C<EV_A> form is used when this is the sole argument,
1802	C<EV_A_> is used when other arguments are following. Example:	2742	C<EV_A_> is used when other arguments are following. Example:
1803		2743
1804	ev_unref (EV_A);	2744	ev_unref (EV_A);
1805	ev_timer_add (EV_A_ watcher);	2745	ev_timer_add (EV_A_ watcher);
1806	ev_loop (EV_A_ 0);	2746	ev_loop (EV_A_ 0);
1807		2747
1808	It assumes the variable C<loop> of type C<struct ev_loop *> is in scope,	2748	It assumes the variable C<loop> of type C<struct ev_loop *> is in scope,
1809	which is often provided by the following macro.	2749	which is often provided by the following macro.
1810		2750
1811	=item C<EV_P>, C<EV_P_>	2751	=item C<EV_P>, C<EV_P_>
1812		2752
1813	This provides the loop I<parameter> for functions, if one is required ("ev	2753	This provides the loop I<parameter> for functions, if one is required ("ev
1814	loop parameter"). The C<EV_P> form is used when this is the sole parameter,	2754	loop parameter"). The C<EV_P> form is used when this is the sole parameter,
1815	C<EV_P_> is used when other parameters are following. Example:	2755	C<EV_P_> is used when other parameters are following. Example:
1816		2756
1817	// this is how ev_unref is being declared	2757	// this is how ev_unref is being declared
1818	static void ev_unref (EV_P);	2758	static void ev_unref (EV_P);
1819		2759
1820	// this is how you can declare your typical callback	2760	// this is how you can declare your typical callback
1821	static void cb (EV_P_ ev_timer *w, int revents)	2761	static void cb (EV_P_ ev_timer *w, int revents)
1822		2762
1823	It declares a parameter C<loop> of type C<struct ev_loop *>, quite	2763	It declares a parameter C<loop> of type C<struct ev_loop *>, quite
1824	suitable for use with C<EV_A>.	2764	suitable for use with C<EV_A>.
1825		2765
1826	=item C<EV_DEFAULT>, C<EV_DEFAULT_>	2766	=item C<EV_DEFAULT>, C<EV_DEFAULT_>
1827		2767
1828	Similar to the other two macros, this gives you the value of the default	2768	Similar to the other two macros, this gives you the value of the default
1829	loop, if multiple loops are supported ("ev loop default").	2769	loop, if multiple loops are supported ("ev loop default").
1830		2770
		2771	=item C<EV_DEFAULT_UC>, C<EV_DEFAULT_UC_>
		2772
		2773	Usage identical to C<EV_DEFAULT> and C<EV_DEFAULT_>, but requires that the
		2774	default loop has been initialised (C<UC> == unchecked). Their behaviour
		2775	is undefined when the default loop has not been initialised by a previous
		2776	execution of C<EV_DEFAULT>, C<EV_DEFAULT_> or C<ev_default_init (...)>.
		2777
		2778	It is often prudent to use C<EV_DEFAULT> when initialising the first
		2779	watcher in a function but use C<EV_DEFAULT_UC> afterwards.
		2780
1831	=back	2781	=back
1832		2782
1833	Example: Declare and initialise a check watcher, working regardless of	2783	Example: Declare and initialise a check watcher, utilising the above
1834	wether multiple loops are supported or not.	2784	macros so it will work regardless of whether multiple loops are supported
		2785	or not.
1835		2786
1836	static void	2787	static void
1837	check_cb (EV_P_ ev_timer *w, int revents)	2788	check_cb (EV_P_ ev_timer *w, int revents)
1838	{	2789	{
1839	ev_check_stop (EV_A_ w);	2790	ev_check_stop (EV_A_ w);
1840	}	2791	}
1841		2792
1842	ev_check check;	2793	ev_check check;
1843	ev_check_init (&check, check_cb);	2794	ev_check_init (&check, check_cb);
1844	ev_check_start (EV_DEFAULT_ &check);	2795	ev_check_start (EV_DEFAULT_ &check);
1845	ev_loop (EV_DEFAULT_ 0);	2796	ev_loop (EV_DEFAULT_ 0);
1846
1847		2797
1848	=head1 EMBEDDING	2798	=head1 EMBEDDING
1849		2799
1850	Libev can (and often is) directly embedded into host	2800	Libev can (and often is) directly embedded into host
1851	applications. Examples of applications that embed it include the Deliantra	2801	applications. Examples of applications that embed it include the Deliantra
1852	Game Server, the EV perl module, the GNU Virtual Private Ethernet (gvpe)	2802	Game Server, the EV perl module, the GNU Virtual Private Ethernet (gvpe)
1853	and rxvt-unicode.	2803	and rxvt-unicode.
1854		2804
1855	The goal is to enable you to just copy the neecssary files into your	2805	The goal is to enable you to just copy the necessary files into your
1856	source directory without having to change even a single line in them, so	2806	source directory without having to change even a single line in them, so
1857	you can easily upgrade by simply copying (or having a checked-out copy of	2807	you can easily upgrade by simply copying (or having a checked-out copy of
1858	libev somewhere in your source tree).	2808	libev somewhere in your source tree).
1859		2809
1860	=head2 FILESETS	2810	=head2 FILESETS
1861		2811
1862	Depending on what features you need you need to include one or more sets of files	2812	Depending on what features you need you need to include one or more sets of files
1863	in your app.	2813	in your application.
1864		2814
1865	=head3 CORE EVENT LOOP	2815	=head3 CORE EVENT LOOP
1866		2816
1867	To include only the libev core (all the C<ev_*> functions), with manual	2817	To include only the libev core (all the C<ev_*> functions), with manual
1868	configuration (no autoconf):	2818	configuration (no autoconf):
1869		2819
1870	#define EV_STANDALONE 1	2820	#define EV_STANDALONE 1
1871	#include "ev.c"	2821	#include "ev.c"
1872		2822
1873	This will automatically include F<ev.h>, too, and should be done in a	2823	This will automatically include F<ev.h>, too, and should be done in a
1874	single C source file only to provide the function implementations. To use	2824	single C source file only to provide the function implementations. To use
1875	it, do the same for F<ev.h> in all files wishing to use this API (best	2825	it, do the same for F<ev.h> in all files wishing to use this API (best
1876	done by writing a wrapper around F<ev.h> that you can include instead and	2826	done by writing a wrapper around F<ev.h> that you can include instead and
1877	where you can put other configuration options):	2827	where you can put other configuration options):
1878		2828
1879	#define EV_STANDALONE 1	2829	#define EV_STANDALONE 1
1880	#include "ev.h"	2830	#include "ev.h"
1881		2831
1882	Both header files and implementation files can be compiled with a C++	2832	Both header files and implementation files can be compiled with a C++
1883	compiler (at least, thats a stated goal, and breakage will be treated	2833	compiler (at least, thats a stated goal, and breakage will be treated
1884	as a bug).	2834	as a bug).
1885		2835
1886	You need the following files in your source tree, or in a directory	2836	You need the following files in your source tree, or in a directory
1887	in your include path (e.g. in libev/ when using -Ilibev):	2837	in your include path (e.g. in libev/ when using -Ilibev):
1888		2838
1889	ev.h	2839	ev.h
1890	ev.c	2840	ev.c
1891	ev_vars.h	2841	ev_vars.h
1892	ev_wrap.h	2842	ev_wrap.h
1893		2843
1894	ev_win32.c required on win32 platforms only	2844	ev_win32.c required on win32 platforms only
1895		2845
1896	ev_select.c only when select backend is enabled (which is by default)	2846	ev_select.c only when select backend is enabled (which is enabled by default)
1897	ev_poll.c only when poll backend is enabled (disabled by default)	2847	ev_poll.c only when poll backend is enabled (disabled by default)
1898	ev_epoll.c only when the epoll backend is enabled (disabled by default)	2848	ev_epoll.c only when the epoll backend is enabled (disabled by default)
1899	ev_kqueue.c only when the kqueue backend is enabled (disabled by default)	2849	ev_kqueue.c only when the kqueue backend is enabled (disabled by default)
1900	ev_port.c only when the solaris port backend is enabled (disabled by default)	2850	ev_port.c only when the solaris port backend is enabled (disabled by default)
1901		2851
1902	F<ev.c> includes the backend files directly when enabled, so you only need	2852	F<ev.c> includes the backend files directly when enabled, so you only need
1903	to compile this single file.	2853	to compile this single file.
1904		2854
1905	=head3 LIBEVENT COMPATIBILITY API	2855	=head3 LIBEVENT COMPATIBILITY API
1906		2856
1907	To include the libevent compatibility API, also include:	2857	To include the libevent compatibility API, also include:
1908		2858
1909	#include "event.c"	2859	#include "event.c"
1910		2860
1911	in the file including F<ev.c>, and:	2861	in the file including F<ev.c>, and:
1912		2862
1913	#include "event.h"	2863	#include "event.h"
1914		2864
1915	in the files that want to use the libevent API. This also includes F<ev.h>.	2865	in the files that want to use the libevent API. This also includes F<ev.h>.
1916		2866
1917	You need the following additional files for this:	2867	You need the following additional files for this:
1918		2868
1919	event.h	2869	event.h
1920	event.c	2870	event.c
1921		2871
1922	=head3 AUTOCONF SUPPORT	2872	=head3 AUTOCONF SUPPORT
1923		2873
1924	Instead of using C<EV_STANDALONE=1> and providing your config in	2874	Instead of using C<EV_STANDALONE=1> and providing your configuration in
1925	whatever way you want, you can also C<m4_include([libev.m4])> in your	2875	whatever way you want, you can also C<m4_include([libev.m4])> in your
1926	F<configure.ac> and leave C<EV_STANDALONE> undefined. F<ev.c> will then	2876	F<configure.ac> and leave C<EV_STANDALONE> undefined. F<ev.c> will then
1927	include F<config.h> and configure itself accordingly.	2877	include F<config.h> and configure itself accordingly.
1928		2878
1929	For this of course you need the m4 file:	2879	For this of course you need the m4 file:
1930		2880
1931	libev.m4	2881	libev.m4
1932		2882
1933	=head2 PREPROCESSOR SYMBOLS/MACROS	2883	=head2 PREPROCESSOR SYMBOLS/MACROS
1934		2884
1935	Libev can be configured via a variety of preprocessor symbols you have to define	2885	Libev can be configured via a variety of preprocessor symbols you have to
1936	before including any of its files. The default is not to build for multiplicity	2886	define before including any of its files. The default in the absence of
1937	and only include the select backend.	2887	autoconf is noted for every option.
1938		2888
1939	=over 4	2889	=over 4
1940		2890
1941	=item EV_STANDALONE	2891	=item EV_STANDALONE
1942		2892
…		…
1947	F<event.h> that are not directly supported by the libev core alone.	2897	F<event.h> that are not directly supported by the libev core alone.
1948		2898
1949	=item EV_USE_MONOTONIC	2899	=item EV_USE_MONOTONIC
1950		2900
1951	If defined to be C<1>, libev will try to detect the availability of the	2901	If defined to be C<1>, libev will try to detect the availability of the
1952	monotonic clock option at both compiletime and runtime. Otherwise no use	2902	monotonic clock option at both compile time and runtime. Otherwise no use
1953	of the monotonic clock option will be attempted. If you enable this, you	2903	of the monotonic clock option will be attempted. If you enable this, you
1954	usually have to link against librt or something similar. Enabling it when	2904	usually have to link against librt or something similar. Enabling it when
1955	the functionality isn't available is safe, though, althoguh you have	2905	the functionality isn't available is safe, though, although you have
1956	to make sure you link against any libraries where the C<clock_gettime>	2906	to make sure you link against any libraries where the C<clock_gettime>
1957	function is hiding in (often F<-lrt>).	2907	function is hiding in (often F<-lrt>).
1958		2908
1959	=item EV_USE_REALTIME	2909	=item EV_USE_REALTIME
1960		2910
1961	If defined to be C<1>, libev will try to detect the availability of the	2911	If defined to be C<1>, libev will try to detect the availability of the
1962	realtime clock option at compiletime (and assume its availability at	2912	real-time clock option at compile time (and assume its availability at
1963	runtime if successful). Otherwise no use of the realtime clock option will	2913	runtime if successful). Otherwise no use of the real-time clock option will
1964	be attempted. This effectively replaces C<gettimeofday> by C<clock_get	2914	be attempted. This effectively replaces C<gettimeofday> by C<clock_get
1965	(CLOCK_REALTIME, ...)> and will not normally affect correctness. See tzhe note about libraries	2915	(CLOCK_REALTIME, ...)> and will not normally affect correctness. See the
1966	in the description of C<EV_USE_MONOTONIC>, though.	2916	note about libraries in the description of C<EV_USE_MONOTONIC>, though.
		2917
		2918	=item EV_USE_NANOSLEEP
		2919
		2920	If defined to be C<1>, libev will assume that C<nanosleep ()> is available
		2921	and will use it for delays. Otherwise it will use C<select ()>.
		2922
		2923	=item EV_USE_EVENTFD
		2924
		2925	If defined to be C<1>, then libev will assume that C<eventfd ()> is
		2926	available and will probe for kernel support at runtime. This will improve
		2927	C<ev_signal> and C<ev_async> performance and reduce resource consumption.
		2928	If undefined, it will be enabled if the headers indicate GNU/Linux + Glibc
		2929	2.7 or newer, otherwise disabled.
1967		2930
1968	=item EV_USE_SELECT	2931	=item EV_USE_SELECT
1969		2932
1970	If undefined or defined to be C<1>, libev will compile in support for the	2933	If undefined or defined to be C<1>, libev will compile in support for the
1971	C<select>(2) backend. No attempt at autodetection will be done: if no	2934	C<select>(2) backend. No attempt at auto-detection will be done: if no
1972	other method takes over, select will be it. Otherwise the select backend	2935	other method takes over, select will be it. Otherwise the select backend
1973	will not be compiled in.	2936	will not be compiled in.
1974		2937
1975	=item EV_SELECT_USE_FD_SET	2938	=item EV_SELECT_USE_FD_SET
1976		2939
1977	If defined to C<1>, then the select backend will use the system C<fd_set>	2940	If defined to C<1>, then the select backend will use the system C<fd_set>
1978	structure. This is useful if libev doesn't compile due to a missing	2941	structure. This is useful if libev doesn't compile due to a missing
1979	C<NFDBITS> or C<fd_mask> definition or it misguesses the bitset layout on	2942	C<NFDBITS> or C<fd_mask> definition or it mis-guesses the bitset layout on
1980	exotic systems. This usually limits the range of file descriptors to some	2943	exotic systems. This usually limits the range of file descriptors to some
1981	low limit such as 1024 or might have other limitations (winsocket only	2944	low limit such as 1024 or might have other limitations (winsocket only
1982	allows 64 sockets). The C<FD_SETSIZE> macro, set before compilation, might	2945	allows 64 sockets). The C<FD_SETSIZE> macro, set before compilation, might
1983	influence the size of the C<fd_set> used.	2946	influence the size of the C<fd_set> used.
1984		2947
…		…
1990	be used is the winsock select). This means that it will call	2953	be used is the winsock select). This means that it will call
1991	C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,	2954	C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,
1992	it is assumed that all these functions actually work on fds, even	2955	it is assumed that all these functions actually work on fds, even
1993	on win32. Should not be defined on non-win32 platforms.	2956	on win32. Should not be defined on non-win32 platforms.
1994		2957
		2958	=item EV_FD_TO_WIN32_HANDLE
		2959
		2960	If C<EV_SELECT_IS_WINSOCKET> is enabled, then libev needs a way to map
		2961	file descriptors to socket handles. When not defining this symbol (the
		2962	default), then libev will call C<_get_osfhandle>, which is usually
		2963	correct. In some cases, programs use their own file descriptor management,
		2964	in which case they can provide this function to map fds to socket handles.
		2965
1995	=item EV_USE_POLL	2966	=item EV_USE_POLL
1996		2967
1997	If defined to be C<1>, libev will compile in support for the C<poll>(2)	2968	If defined to be C<1>, libev will compile in support for the C<poll>(2)
1998	backend. Otherwise it will be enabled on non-win32 platforms. It	2969	backend. Otherwise it will be enabled on non-win32 platforms. It
1999	takes precedence over select.	2970	takes precedence over select.
2000		2971
2001	=item EV_USE_EPOLL	2972	=item EV_USE_EPOLL
2002		2973
2003	If defined to be C<1>, libev will compile in support for the Linux	2974	If defined to be C<1>, libev will compile in support for the Linux
2004	C<epoll>(7) backend. Its availability will be detected at runtime,	2975	C<epoll>(7) backend. Its availability will be detected at runtime,
2005	otherwise another method will be used as fallback. This is the	2976	otherwise another method will be used as fallback. This is the preferred
2006	preferred backend for GNU/Linux systems.	2977	backend for GNU/Linux systems. If undefined, it will be enabled if the
		2978	headers indicate GNU/Linux + Glibc 2.4 or newer, otherwise disabled.
2007		2979
2008	=item EV_USE_KQUEUE	2980	=item EV_USE_KQUEUE
2009		2981
2010	If defined to be C<1>, libev will compile in support for the BSD style	2982	If defined to be C<1>, libev will compile in support for the BSD style
2011	C<kqueue>(2) backend. Its actual availability will be detected at runtime,	2983	C<kqueue>(2) backend. Its actual availability will be detected at runtime,
…		…
2024	otherwise another method will be used as fallback. This is the preferred	2996	otherwise another method will be used as fallback. This is the preferred
2025	backend for Solaris 10 systems.	2997	backend for Solaris 10 systems.
2026		2998
2027	=item EV_USE_DEVPOLL	2999	=item EV_USE_DEVPOLL
2028		3000
2029	reserved for future expansion, works like the USE symbols above.	3001	Reserved for future expansion, works like the USE symbols above.
2030		3002
2031	=item EV_USE_INOTIFY	3003	=item EV_USE_INOTIFY
2032		3004
2033	If defined to be C<1>, libev will compile in support for the Linux inotify	3005	If defined to be C<1>, libev will compile in support for the Linux inotify
2034	interface to speed up C<ev_stat> watchers. Its actual availability will	3006	interface to speed up C<ev_stat> watchers. Its actual availability will
2035	be detected at runtime.	3007	be detected at runtime. If undefined, it will be enabled if the headers
		3008	indicate GNU/Linux + Glibc 2.4 or newer, otherwise disabled.
		3009
		3010	=item EV_ATOMIC_T
		3011
		3012	Libev requires an integer type (suitable for storing C<0> or C<1>) whose
		3013	access is atomic with respect to other threads or signal contexts. No such
		3014	type is easily found in the C language, so you can provide your own type
		3015	that you know is safe for your purposes. It is used both for signal handler "locking"
		3016	as well as for signal and thread safety in C<ev_async> watchers.
		3017
		3018	In the absence of this define, libev will use C<sig_atomic_t volatile>
		3019	(from F<signal.h>), which is usually good enough on most platforms.
2036		3020
2037	=item EV_H	3021	=item EV_H
2038		3022
2039	The name of the F<ev.h> header file used to include it. The default if	3023	The name of the F<ev.h> header file used to include it. The default if
2040	undefined is C<< <ev.h> >> in F<event.h> and C<"ev.h"> in F<ev.c>. This	3024	undefined is C<"ev.h"> in F<event.h>, F<ev.c> and F<ev++.h>. This can be
2041	can be used to virtually rename the F<ev.h> header file in case of conflicts.	3025	used to virtually rename the F<ev.h> header file in case of conflicts.
2042		3026
2043	=item EV_CONFIG_H	3027	=item EV_CONFIG_H
2044		3028
2045	If C<EV_STANDALONE> isn't C<1>, this variable can be used to override	3029	If C<EV_STANDALONE> isn't C<1>, this variable can be used to override
2046	F<ev.c>'s idea of where to find the F<config.h> file, similarly to	3030	F<ev.c>'s idea of where to find the F<config.h> file, similarly to
2047	C<EV_H>, above.	3031	C<EV_H>, above.
2048		3032
2049	=item EV_EVENT_H	3033	=item EV_EVENT_H
2050		3034
2051	Similarly to C<EV_H>, this macro can be used to override F<event.c>'s idea	3035	Similarly to C<EV_H>, this macro can be used to override F<event.c>'s idea
2052	of how the F<event.h> header can be found.	3036	of how the F<event.h> header can be found, the default is C<"event.h">.
2053		3037
2054	=item EV_PROTOTYPES	3038	=item EV_PROTOTYPES
2055		3039
2056	If defined to be C<0>, then F<ev.h> will not define any function	3040	If defined to be C<0>, then F<ev.h> will not define any function
2057	prototypes, but still define all the structs and other symbols. This is	3041	prototypes, but still define all the structs and other symbols. This is
…		…
2064	will have the C<struct ev_loop *> as first argument, and you can create	3048	will have the C<struct ev_loop *> as first argument, and you can create
2065	additional independent event loops. Otherwise there will be no support	3049	additional independent event loops. Otherwise there will be no support
2066	for multiple event loops and there is no first event loop pointer	3050	for multiple event loops and there is no first event loop pointer
2067	argument. Instead, all functions act on the single default loop.	3051	argument. Instead, all functions act on the single default loop.
2068		3052
		3053	=item EV_MINPRI
		3054
		3055	=item EV_MAXPRI
		3056
		3057	The range of allowed priorities. C<EV_MINPRI> must be smaller or equal to
		3058	C<EV_MAXPRI>, but otherwise there are no non-obvious limitations. You can
		3059	provide for more priorities by overriding those symbols (usually defined
		3060	to be C<-2> and C<2>, respectively).
		3061
		3062	When doing priority-based operations, libev usually has to linearly search
		3063	all the priorities, so having many of them (hundreds) uses a lot of space
		3064	and time, so using the defaults of five priorities (-2 .. +2) is usually
		3065	fine.
		3066
		3067	If your embedding application does not need any priorities, defining these both to
		3068	C<0> will save some memory and CPU.
		3069
2069	=item EV_PERIODIC_ENABLE	3070	=item EV_PERIODIC_ENABLE
2070		3071
2071	If undefined or defined to be C<1>, then periodic timers are supported. If	3072	If undefined or defined to be C<1>, then periodic timers are supported. If
2072	defined to be C<0>, then they are not. Disabling them saves a few kB of	3073	defined to be C<0>, then they are not. Disabling them saves a few kB of
2073	code.	3074	code.
2074		3075
		3076	=item EV_IDLE_ENABLE
		3077
		3078	If undefined or defined to be C<1>, then idle watchers are supported. If
		3079	defined to be C<0>, then they are not. Disabling them saves a few kB of
		3080	code.
		3081
2075	=item EV_EMBED_ENABLE	3082	=item EV_EMBED_ENABLE
2076		3083
2077	If undefined or defined to be C<1>, then embed watchers are supported. If	3084	If undefined or defined to be C<1>, then embed watchers are supported. If
2078	defined to be C<0>, then they are not.	3085	defined to be C<0>, then they are not.
2079		3086
…		…
2085	=item EV_FORK_ENABLE	3092	=item EV_FORK_ENABLE
2086		3093
2087	If undefined or defined to be C<1>, then fork watchers are supported. If	3094	If undefined or defined to be C<1>, then fork watchers are supported. If
2088	defined to be C<0>, then they are not.	3095	defined to be C<0>, then they are not.
2089		3096
		3097	=item EV_ASYNC_ENABLE
		3098
		3099	If undefined or defined to be C<1>, then async watchers are supported. If
		3100	defined to be C<0>, then they are not.
		3101
2090	=item EV_MINIMAL	3102	=item EV_MINIMAL
2091		3103
2092	If you need to shave off some kilobytes of code at the expense of some	3104	If you need to shave off some kilobytes of code at the expense of some
2093	speed, define this symbol to C<1>. Currently only used for gcc to override	3105	speed, define this symbol to C<1>. Currently this is used to override some
2094	some inlining decisions, saves roughly 30% codesize of amd64.	3106	inlining decisions, saves roughly 30% code size on amd64. It also selects a
		3107	much smaller 2-heap for timer management over the default 4-heap.
2095		3108
2096	=item EV_PID_HASHSIZE	3109	=item EV_PID_HASHSIZE
2097		3110
2098	C<ev_child> watchers use a small hash table to distribute workload by	3111	C<ev_child> watchers use a small hash table to distribute workload by
2099	pid. The default size is C<16> (or C<1> with C<EV_MINIMAL>), usually more	3112	pid. The default size is C<16> (or C<1> with C<EV_MINIMAL>), usually more
2100	than enough. If you need to manage thousands of children you might want to	3113	than enough. If you need to manage thousands of children you might want to
2101	increase this value (I<must> be a power of two).	3114	increase this value (I<must> be a power of two).
2102		3115
2103	=item EV_INOTIFY_HASHSIZE	3116	=item EV_INOTIFY_HASHSIZE
2104		3117
2105	C<ev_staz> watchers use a small hash table to distribute workload by	3118	C<ev_stat> watchers use a small hash table to distribute workload by
2106	inotify watch id. The default size is C<16> (or C<1> with C<EV_MINIMAL>),	3119	inotify watch id. The default size is C<16> (or C<1> with C<EV_MINIMAL>),
2107	usually more than enough. If you need to manage thousands of C<ev_stat>	3120	usually more than enough. If you need to manage thousands of C<ev_stat>
2108	watchers you might want to increase this value (I<must> be a power of	3121	watchers you might want to increase this value (I<must> be a power of
2109	two).	3122	two).
2110		3123
		3124	=item EV_USE_4HEAP
		3125
		3126	Heaps are not very cache-efficient. To improve the cache-efficiency of the
		3127	timer and periodics heap, libev uses a 4-heap when this symbol is defined
		3128	to C<1>. The 4-heap uses more complicated (longer) code but has
		3129	noticeably faster performance with many (thousands) of watchers.
		3130
		3131	The default is C<1> unless C<EV_MINIMAL> is set in which case it is C<0>
		3132	(disabled).
		3133
		3134	=item EV_HEAP_CACHE_AT
		3135
		3136	Heaps are not very cache-efficient. To improve the cache-efficiency of the
		3137	timer and periodics heap, libev can cache the timestamp (I<at>) within
		3138	the heap structure (selected by defining C<EV_HEAP_CACHE_AT> to C<1>),
		3139	which uses 8-12 bytes more per watcher and a few hundred bytes more code,
		3140	but avoids random read accesses on heap changes. This improves performance
		3141	noticeably with with many (hundreds) of watchers.
		3142
		3143	The default is C<1> unless C<EV_MINIMAL> is set in which case it is C<0>
		3144	(disabled).
		3145
		3146	=item EV_VERIFY
		3147
		3148	Controls how much internal verification (see C<ev_loop_verify ()>) will
		3149	be done: If set to C<0>, no internal verification code will be compiled
		3150	in. If set to C<1>, then verification code will be compiled in, but not
		3151	called. If set to C<2>, then the internal verification code will be
		3152	called once per loop, which can slow down libev. If set to C<3>, then the
		3153	verification code will be called very frequently, which will slow down
		3154	libev considerably.
		3155
		3156	The default is C<1>, unless C<EV_MINIMAL> is set, in which case it will be
		3157	C<0.>
		3158
2111	=item EV_COMMON	3159	=item EV_COMMON
2112		3160
2113	By default, all watchers have a C<void *data> member. By redefining	3161	By default, all watchers have a C<void *data> member. By redefining
2114	this macro to a something else you can include more and other types of	3162	this macro to a something else you can include more and other types of
2115	members. You have to define it each time you include one of the files,	3163	members. You have to define it each time you include one of the files,
2116	though, and it must be identical each time.	3164	though, and it must be identical each time.
2117		3165
2118	For example, the perl EV module uses something like this:	3166	For example, the perl EV module uses something like this:
2119		3167
2120	#define EV_COMMON \	3168	#define EV_COMMON \
2121	SV *self; /* contains this struct */ \	3169	SV *self; /* contains this struct */ \
2122	SV *cb_sv, *fh /* note no trailing ";" */	3170	SV *cb_sv, *fh /* note no trailing ";" */
2123		3171
2124	=item EV_CB_DECLARE (type)	3172	=item EV_CB_DECLARE (type)
2125		3173
2126	=item EV_CB_INVOKE (watcher, revents)	3174	=item EV_CB_INVOKE (watcher, revents)
2127		3175
2128	=item ev_set_cb (ev, cb)	3176	=item ev_set_cb (ev, cb)
2129		3177
2130	Can be used to change the callback member declaration in each watcher,	3178	Can be used to change the callback member declaration in each watcher,
2131	and the way callbacks are invoked and set. Must expand to a struct member	3179	and the way callbacks are invoked and set. Must expand to a struct member
2132	definition and a statement, respectively. See the F<ev.v> header file for	3180	definition and a statement, respectively. See the F<ev.h> header file for
2133	their default definitions. One possible use for overriding these is to	3181	their default definitions. One possible use for overriding these is to
2134	avoid the C<struct ev_loop *> as first argument in all cases, or to use	3182	avoid the C<struct ev_loop *> as first argument in all cases, or to use
2135	method calls instead of plain function calls in C++.	3183	method calls instead of plain function calls in C++.
		3184
		3185	=head2 EXPORTED API SYMBOLS
		3186
		3187	If you need to re-export the API (e.g. via a DLL) and you need a list of
		3188	exported symbols, you can use the provided F<Symbol.*> files which list
		3189	all public symbols, one per line:
		3190
		3191	Symbols.ev for libev proper
		3192	Symbols.event for the libevent emulation
		3193
		3194	This can also be used to rename all public symbols to avoid clashes with
		3195	multiple versions of libev linked together (which is obviously bad in
		3196	itself, but sometimes it is inconvenient to avoid this).
		3197
		3198	A sed command like this will create wrapper C<#define>'s that you need to
		3199	include before including F<ev.h>:
		3200
		3201	<Symbols.ev sed -e "s/.*/#define & myprefix_&/" >wrap.h
		3202
		3203	This would create a file F<wrap.h> which essentially looks like this:
		3204
		3205	#define ev_backend myprefix_ev_backend
		3206	#define ev_check_start myprefix_ev_check_start
		3207	#define ev_check_stop myprefix_ev_check_stop
		3208	...
2136		3209
2137	=head2 EXAMPLES	3210	=head2 EXAMPLES
2138		3211
2139	For a real-world example of a program the includes libev	3212	For a real-world example of a program the includes libev
2140	verbatim, you can have a look at the EV perl module	3213	verbatim, you can have a look at the EV perl module
…		…
2143	interface) and F<EV.xs> (implementation) files. Only the F<EV.xs> file	3216	interface) and F<EV.xs> (implementation) files. Only the F<EV.xs> file
2144	will be compiled. It is pretty complex because it provides its own header	3217	will be compiled. It is pretty complex because it provides its own header
2145	file.	3218	file.
2146		3219
2147	The usage in rxvt-unicode is simpler. It has a F<ev_cpp.h> header file	3220	The usage in rxvt-unicode is simpler. It has a F<ev_cpp.h> header file
2148	that everybody includes and which overrides some autoconf choices:	3221	that everybody includes and which overrides some configure choices:
2149		3222
		3223	#define EV_MINIMAL 1
2150	#define EV_USE_POLL 0	3224	#define EV_USE_POLL 0
2151	#define EV_MULTIPLICITY 0	3225	#define EV_MULTIPLICITY 0
2152	#define EV_PERIODICS 0	3226	#define EV_PERIODIC_ENABLE 0
		3227	#define EV_STAT_ENABLE 0
		3228	#define EV_FORK_ENABLE 0
2153	#define EV_CONFIG_H <config.h>	3229	#define EV_CONFIG_H <config.h>
		3230	#define EV_MINPRI 0
		3231	#define EV_MAXPRI 0
2154		3232
2155	#include "ev++.h"	3233	#include "ev++.h"
2156		3234
2157	And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:	3235	And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:
2158		3236
2159	#include "ev_cpp.h"	3237	#include "ev_cpp.h"
2160	#include "ev.c"	3238	#include "ev.c"
		3239
		3240
		3241	=head1 THREADS AND COROUTINES
		3242
		3243	=head2 THREADS
		3244
		3245	Libev itself is completely thread-safe, but it uses no locking. This
		3246	means that you can use as many loops as you want in parallel, as long as
		3247	only one thread ever calls into one libev function with the same loop
		3248	parameter.
		3249
		3250	Or put differently: calls with different loop parameters can be done in
		3251	parallel from multiple threads, calls with the same loop parameter must be
		3252	done serially (but can be done from different threads, as long as only one
		3253	thread ever is inside a call at any point in time, e.g. by using a mutex
		3254	per loop).
		3255
		3256	If you want to know which design (one loop, locking, or multiple loops
		3257	without or something else still) is best for your problem, then I cannot
		3258	help you. I can give some generic advice however:
		3259
		3260	=over 4
		3261
		3262	=item * most applications have a main thread: use the default libev loop
		3263	in that thread, or create a separate thread running only the default loop.
		3264
		3265	This helps integrating other libraries or software modules that use libev
		3266	themselves and don't care/know about threading.
		3267
		3268	=item * one loop per thread is usually a good model.
		3269
		3270	Doing this is almost never wrong, sometimes a better-performance model
		3271	exists, but it is always a good start.
		3272
		3273	=item * other models exist, such as the leader/follower pattern, where one
		3274	loop is handed through multiple threads in a kind of round-robin fashion.
		3275
		3276	Choosing a model is hard - look around, learn, know that usually you can do
		3277	better than you currently do :-)
		3278
		3279	=item * often you need to talk to some other thread which blocks in the
		3280	event loop - C<ev_async> watchers can be used to wake them up from other
		3281	threads safely (or from signal contexts...).
		3282
		3283	=back
		3284
		3285	=head2 COROUTINES
		3286
		3287	Libev is much more accommodating to coroutines ("cooperative threads"):
		3288	libev fully supports nesting calls to it's functions from different
		3289	coroutines (e.g. you can call C<ev_loop> on the same loop from two
		3290	different coroutines and switch freely between both coroutines running the
		3291	loop, as long as you don't confuse yourself). The only exception is that
		3292	you must not do this from C<ev_periodic> reschedule callbacks.
		3293
		3294	Care has been invested into making sure that libev does not keep local
		3295	state inside C<ev_loop>, and other calls do not usually allow coroutine
		3296	switches.
2161		3297
2162		3298
2163	=head1 COMPLEXITIES	3299	=head1 COMPLEXITIES
2164		3300
2165	In this section the complexities of (many of) the algorithms used inside	3301	In this section the complexities of (many of) the algorithms used inside
2166	libev will be explained. For complexity discussions about backends see the	3302	libev will be explained. For complexity discussions about backends see the
2167	documentation for C<ev_default_init>.	3303	documentation for C<ev_default_init>.
2168		3304
		3305	All of the following are about amortised time: If an array needs to be
		3306	extended, libev needs to realloc and move the whole array, but this
		3307	happens asymptotically never with higher number of elements, so O(1) might
		3308	mean it might do a lengthy realloc operation in rare cases, but on average
		3309	it is much faster and asymptotically approaches constant time.
		3310
2169	=over 4	3311	=over 4
2170		3312
2171	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)	3313	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)
2172		3314
		3315	This means that, when you have a watcher that triggers in one hour and
		3316	there are 100 watchers that would trigger before that then inserting will
		3317	have to skip roughly seven (C<ld 100>) of these watchers.
		3318
2173	=item Changing timer/periodic watchers (by autorepeat, again): O(log skipped_other_timers)	3319	=item Changing timer/periodic watchers (by autorepeat or calling again): O(log skipped_other_timers)
2174		3320
		3321	That means that changing a timer costs less than removing/adding them
		3322	as only the relative motion in the event queue has to be paid for.
		3323
2175	=item Starting io/check/prepare/idle/signal/child watchers: O(1)	3324	=item Starting io/check/prepare/idle/signal/child/fork/async watchers: O(1)
2176		3325
		3326	These just add the watcher into an array or at the head of a list.
		3327
2177	=item Stopping check/prepare/idle watchers: O(1)	3328	=item Stopping check/prepare/idle/fork/async watchers: O(1)
2178		3329
2179	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))	3330	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))
2180		3331
		3332	These watchers are stored in lists then need to be walked to find the
		3333	correct watcher to remove. The lists are usually short (you don't usually
		3334	have many watchers waiting for the same fd or signal).
		3335
2181	=item Finding the next timer per loop iteration: O(1)	3336	=item Finding the next timer in each loop iteration: O(1)
		3337
		3338	By virtue of using a binary or 4-heap, the next timer is always found at a
		3339	fixed position in the storage array.
2182		3340
2183	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)	3341	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)
2184		3342
2185	=item Activating one watcher: O(1)	3343	A change means an I/O watcher gets started or stopped, which requires
		3344	libev to recalculate its status (and possibly tell the kernel, depending
		3345	on backend and whether C<ev_io_set> was used).
		3346
		3347	=item Activating one watcher (putting it into the pending state): O(1)
		3348
		3349	=item Priority handling: O(number_of_priorities)
		3350
		3351	Priorities are implemented by allocating some space for each
		3352	priority. When doing priority-based operations, libev usually has to
		3353	linearly search all the priorities, but starting/stopping and activating
		3354	watchers becomes O(1) w.r.t. priority handling.
		3355
		3356	=item Sending an ev_async: O(1)
		3357
		3358	=item Processing ev_async_send: O(number_of_async_watchers)
		3359
		3360	=item Processing signals: O(max_signal_number)
		3361
		3362	Sending involves a system call I<iff> there were no other C<ev_async_send>
		3363	calls in the current loop iteration. Checking for async and signal events
		3364	involves iterating over all running async watchers or all signal numbers.
2186		3365
2187	=back	3366	=back
2188		3367
2189		3368
		3369	=head1 WIN32 PLATFORM LIMITATIONS AND WORKAROUNDS
		3370
		3371	Win32 doesn't support any of the standards (e.g. POSIX) that libev
		3372	requires, and its I/O model is fundamentally incompatible with the POSIX
		3373	model. Libev still offers limited functionality on this platform in
		3374	the form of the C<EVBACKEND_SELECT> backend, and only supports socket
		3375	descriptors. This only applies when using Win32 natively, not when using
		3376	e.g. cygwin.
		3377
		3378	Lifting these limitations would basically require the full
		3379	re-implementation of the I/O system. If you are into these kinds of
		3380	things, then note that glib does exactly that for you in a very portable
		3381	way (note also that glib is the slowest event library known to man).
		3382
		3383	There is no supported compilation method available on windows except
		3384	embedding it into other applications.
		3385
		3386	Not a libev limitation but worth mentioning: windows apparently doesn't
		3387	accept large writes: instead of resulting in a partial write, windows will
		3388	either accept everything or return C<ENOBUFS> if the buffer is too large,
		3389	so make sure you only write small amounts into your sockets (less than a
		3390	megabyte seems safe, but thsi apparently depends on the amount of memory
		3391	available).
		3392
		3393	Due to the many, low, and arbitrary limits on the win32 platform and
		3394	the abysmal performance of winsockets, using a large number of sockets
		3395	is not recommended (and not reasonable). If your program needs to use
		3396	more than a hundred or so sockets, then likely it needs to use a totally
		3397	different implementation for windows, as libev offers the POSIX readiness
		3398	notification model, which cannot be implemented efficiently on windows
		3399	(Microsoft monopoly games).
		3400
		3401	A typical way to use libev under windows is to embed it (see the embedding
		3402	section for details) and use the following F<evwrap.h> header file instead
		3403	of F<ev.h>:
		3404
		3405	#define EV_STANDALONE /* keeps ev from requiring config.h */
		3406	#define EV_SELECT_IS_WINSOCKET 1 /* configure libev for windows select */
		3407
		3408	#include "ev.h"
		3409
		3410	And compile the following F<evwrap.c> file into your project (make sure
		3411	you do I<not> compile the F<ev.c> or any other embedded soruce files!):
		3412
		3413	#include "evwrap.h"
		3414	#include "ev.c"
		3415
		3416	=over 4
		3417
		3418	=item The winsocket select function
		3419
		3420	The winsocket C<select> function doesn't follow POSIX in that it
		3421	requires socket I<handles> and not socket I<file descriptors> (it is
		3422	also extremely buggy). This makes select very inefficient, and also
		3423	requires a mapping from file descriptors to socket handles (the Microsoft
		3424	C runtime provides the function C<_open_osfhandle> for this). See the
		3425	discussion of the C<EV_SELECT_USE_FD_SET>, C<EV_SELECT_IS_WINSOCKET> and
		3426	C<EV_FD_TO_WIN32_HANDLE> preprocessor symbols for more info.
		3427
		3428	The configuration for a "naked" win32 using the Microsoft runtime
		3429	libraries and raw winsocket select is:
		3430
		3431	#define EV_USE_SELECT 1
		3432	#define EV_SELECT_IS_WINSOCKET 1 /* forces EV_SELECT_USE_FD_SET, too */
		3433
		3434	Note that winsockets handling of fd sets is O(n), so you can easily get a
		3435	complexity in the O(n²) range when using win32.
		3436
		3437	=item Limited number of file descriptors
		3438
		3439	Windows has numerous arbitrary (and low) limits on things.
		3440
		3441	Early versions of winsocket's select only supported waiting for a maximum
		3442	of C<64> handles (probably owning to the fact that all windows kernels
		3443	can only wait for C<64> things at the same time internally; Microsoft
		3444	recommends spawning a chain of threads and wait for 63 handles and the
		3445	previous thread in each. Great).
		3446
		3447	Newer versions support more handles, but you need to define C<FD_SETSIZE>
		3448	to some high number (e.g. C<2048>) before compiling the winsocket select
		3449	call (which might be in libev or elsewhere, for example, perl does its own
		3450	select emulation on windows).
		3451
		3452	Another limit is the number of file descriptors in the Microsoft runtime
		3453	libraries, which by default is C<64> (there must be a hidden I<64> fetish
		3454	or something like this inside Microsoft). You can increase this by calling
		3455	C<_setmaxstdio>, which can increase this limit to C<2048> (another
		3456	arbitrary limit), but is broken in many versions of the Microsoft runtime
		3457	libraries.
		3458
		3459	This might get you to about C<512> or C<2048> sockets (depending on
		3460	windows version and/or the phase of the moon). To get more, you need to
		3461	wrap all I/O functions and provide your own fd management, but the cost of
		3462	calling select (O(n²)) will likely make this unworkable.
		3463
		3464	=back
		3465
		3466
		3467	=head1 PORTABILITY REQUIREMENTS
		3468
		3469	In addition to a working ISO-C implementation, libev relies on a few
		3470	additional extensions:
		3471
		3472	=over 4
		3473
		3474	=item C<void (*)(ev_watcher_type *, int revents)> must have compatible
		3475	calling conventions regardless of C<ev_watcher_type *>.
		3476
		3477	Libev assumes not only that all watcher pointers have the same internal
		3478	structure (guaranteed by POSIX but not by ISO C for example), but it also
		3479	assumes that the same (machine) code can be used to call any watcher
		3480	callback: The watcher callbacks have different type signatures, but libev
		3481	calls them using an C<ev_watcher *> internally.
		3482
		3483	=item C<sig_atomic_t volatile> must be thread-atomic as well
		3484
		3485	The type C<sig_atomic_t volatile> (or whatever is defined as
		3486	C<EV_ATOMIC_T>) must be atomic w.r.t. accesses from different
		3487	threads. This is not part of the specification for C<sig_atomic_t>, but is
		3488	believed to be sufficiently portable.
		3489
		3490	=item C<sigprocmask> must work in a threaded environment
		3491
		3492	Libev uses C<sigprocmask> to temporarily block signals. This is not
		3493	allowed in a threaded program (C<pthread_sigmask> has to be used). Typical
		3494	pthread implementations will either allow C<sigprocmask> in the "main
		3495	thread" or will block signals process-wide, both behaviours would
		3496	be compatible with libev. Interaction between C<sigprocmask> and
		3497	C<pthread_sigmask> could complicate things, however.
		3498
		3499	The most portable way to handle signals is to block signals in all threads
		3500	except the initial one, and run the default loop in the initial thread as
		3501	well.
		3502
		3503	=item C<long> must be large enough for common memory allocation sizes
		3504
		3505	To improve portability and simplify using libev, libev uses C<long>
		3506	internally instead of C<size_t> when allocating its data structures. On
		3507	non-POSIX systems (Microsoft...) this might be unexpectedly low, but
		3508	is still at least 31 bits everywhere, which is enough for hundreds of
		3509	millions of watchers.
		3510
		3511	=item C<double> must hold a time value in seconds with enough accuracy
		3512
		3513	The type C<double> is used to represent timestamps. It is required to
		3514	have at least 51 bits of mantissa (and 9 bits of exponent), which is good
		3515	enough for at least into the year 4000. This requirement is fulfilled by
		3516	implementations implementing IEEE 754 (basically all existing ones).
		3517
		3518	=back
		3519
		3520	If you know of other additional requirements drop me a note.
		3521
		3522
		3523	=head1 COMPILER WARNINGS
		3524
		3525	Depending on your compiler and compiler settings, you might get no or a
		3526	lot of warnings when compiling libev code. Some people are apparently
		3527	scared by this.
		3528
		3529	However, these are unavoidable for many reasons. For one, each compiler
		3530	has different warnings, and each user has different tastes regarding
		3531	warning options. "Warn-free" code therefore cannot be a goal except when
		3532	targeting a specific compiler and compiler-version.
		3533
		3534	Another reason is that some compiler warnings require elaborate
		3535	workarounds, or other changes to the code that make it less clear and less
		3536	maintainable.
		3537
		3538	And of course, some compiler warnings are just plain stupid, or simply
		3539	wrong (because they don't actually warn about the condition their message
		3540	seems to warn about).
		3541
		3542	While libev is written to generate as few warnings as possible,
		3543	"warn-free" code is not a goal, and it is recommended not to build libev
		3544	with any compiler warnings enabled unless you are prepared to cope with
		3545	them (e.g. by ignoring them). Remember that warnings are just that:
		3546	warnings, not errors, or proof of bugs.
		3547
		3548
		3549	=head1 VALGRIND
		3550
		3551	Valgrind has a special section here because it is a popular tool that is
		3552	highly useful, but valgrind reports are very hard to interpret.
		3553
		3554	If you think you found a bug (memory leak, uninitialised data access etc.)
		3555	in libev, then check twice: If valgrind reports something like:
		3556
		3557	==2274== definitely lost: 0 bytes in 0 blocks.
		3558	==2274== possibly lost: 0 bytes in 0 blocks.
		3559	==2274== still reachable: 256 bytes in 1 blocks.
		3560
		3561	Then there is no memory leak. Similarly, under some circumstances,
		3562	valgrind might report kernel bugs as if it were a bug in libev, or it
		3563	might be confused (it is a very good tool, but only a tool).
		3564
		3565	If you are unsure about something, feel free to contact the mailing list
		3566	with the full valgrind report and an explanation on why you think this is
		3567	a bug in libev. However, don't be annoyed when you get a brisk "this is
		3568	no bug" answer and take the chance of learning how to interpret valgrind
		3569	properly.
		3570
		3571	If you need, for some reason, empty reports from valgrind for your project
		3572	I suggest using suppression lists.
		3573
		3574
2190	=head1 AUTHOR	3575	=head1 AUTHOR
2191		3576
2192	Marc Lehmann <libev@schmorp.de>.	3577	Marc Lehmann <libev@schmorp.de>.
2193		3578

Diff Legend

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