ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/libev/ev.pod

Comparing libev/ev.pod (file contents):
Revision 1.27 by root, Wed Nov 14 05:02:07 2007 UTC vs.
Revision 1.49 by root, Tue Nov 27 08:20:42 2007 UTC

…		…
45		45
46	Libev represents time as a single floating point number, representing the	46	Libev represents time as a single floating point number, representing the
47	(fractional) number of seconds since the (POSIX) epoch (somewhere near	47	(fractional) number of seconds since the (POSIX) epoch (somewhere near
48	the beginning of 1970, details are complicated, don't ask). This type is	48	the beginning of 1970, details are complicated, don't ask). This type is
49	called C<ev_tstamp>, which is what you should use too. It usually aliases	49	called C<ev_tstamp>, which is what you should use too. It usually aliases
50	to the double type in C.	50	to the C<double> type in C, and when you need to do any calculations on
		51	it, you should treat it as such.
		52
51		53
52	=head1 GLOBAL FUNCTIONS	54	=head1 GLOBAL FUNCTIONS
53		55
54	These functions can be called anytime, even before initialising the	56	These functions can be called anytime, even before initialising the
55	library in any way.	57	library in any way.
…		…
75	Usually, it's a good idea to terminate if the major versions mismatch,	77	Usually, it's a good idea to terminate if the major versions mismatch,
76	as this indicates an incompatible change. Minor versions are usually	78	as this indicates an incompatible change. Minor versions are usually
77	compatible to older versions, so a larger minor version alone is usually	79	compatible to older versions, so a larger minor version alone is usually
78	not a problem.	80	not a problem.
79		81
		82	Example: make sure we haven't accidentally been linked against the wrong
		83	version:
		84
		85	assert (("libev version mismatch",
		86	ev_version_major () == EV_VERSION_MAJOR
		87	&& ev_version_minor () >= EV_VERSION_MINOR));
		88
		89	=item unsigned int ev_supported_backends ()
		90
		91	Return the set of all backends (i.e. their corresponding C<EV_BACKEND_*>
		92	value) compiled into this binary of libev (independent of their
		93	availability on the system you are running on). See C<ev_default_loop> for
		94	a description of the set values.
		95
		96	Example: make sure we have the epoll method, because yeah this is cool and
		97	a must have and can we have a torrent of it please!!!11
		98
		99	assert (("sorry, no epoll, no sex",
		100	ev_supported_backends () & EVBACKEND_EPOLL));
		101
		102	=item unsigned int ev_recommended_backends ()
		103
		104	Return the set of all backends compiled into this binary of libev and also
		105	recommended for this platform. This set is often smaller than the one
		106	returned by C<ev_supported_backends>, as for example kqueue is broken on
		107	most BSDs and will not be autodetected unless you explicitly request it
		108	(assuming you know what you are doing). This is the set of backends that
		109	libev will probe for if you specify no backends explicitly.
		110
		111	=item unsigned int ev_embeddable_backends ()
		112
		113	Returns the set of backends that are embeddable in other event loops. This
		114	is the theoretical, all-platform, value. To find which backends
		115	might be supported on the current system, you would need to look at
		116	C<ev_embeddable_backends () & ev_supported_backends ()>, likewise for
		117	recommended ones.
		118
		119	See the description of C<ev_embed> watchers for more info.
		120
80	=item ev_set_allocator (void *(*cb)(void *ptr, long size))	121	=item ev_set_allocator (void *(*cb)(void *ptr, long size))
81		122
82	Sets the allocation function to use (the prototype is similar to the	123	Sets the allocation function to use (the prototype is similar to the
83	realloc C function, the semantics are identical). It is used to allocate	124	realloc C function, the semantics are identical). It is used to allocate
84	and free memory (no surprises here). If it returns zero when memory	125	and free memory (no surprises here). If it returns zero when memory
…		…
86	destructive action. The default is your system realloc function.	127	destructive action. The default is your system realloc function.
87		128
88	You could override this function in high-availability programs to, say,	129	You could override this function in high-availability programs to, say,
89	free some memory if it cannot allocate memory, to use a special allocator,	130	free some memory if it cannot allocate memory, to use a special allocator,
90	or even to sleep a while and retry until some memory is available.	131	or even to sleep a while and retry until some memory is available.
		132
		133	Example: replace the libev allocator with one that waits a bit and then
		134	retries: better than mine).
		135
		136	static void *
		137	persistent_realloc (void *ptr, long size)
		138	{
		139	for (;;)
		140	{
		141	void *newptr = realloc (ptr, size);
		142
		143	if (newptr)
		144	return newptr;
		145
		146	sleep (60);
		147	}
		148	}
		149
		150	...
		151	ev_set_allocator (persistent_realloc);
91		152
92	=item ev_set_syserr_cb (void (*cb)(const char *msg));	153	=item ev_set_syserr_cb (void (*cb)(const char *msg));
93		154
94	Set the callback function to call on a retryable syscall error (such	155	Set the callback function to call on a retryable syscall error (such
95	as failed select, poll, epoll_wait). The message is a printable string	156	as failed select, poll, epoll_wait). The message is a printable string
…		…
97	callback is set, then libev will expect it to remedy the sitution, no	158	callback is set, then libev will expect it to remedy the sitution, no
98	matter what, when it returns. That is, libev will generally retry the	159	matter what, when it returns. That is, libev will generally retry the
99	requested operation, or, if the condition doesn't go away, do bad stuff	160	requested operation, or, if the condition doesn't go away, do bad stuff
100	(such as abort).	161	(such as abort).
101		162
		163	Example: do the same thing as libev does internally:
		164
		165	static void
		166	fatal_error (const char *msg)
		167	{
		168	perror (msg);
		169	abort ();
		170	}
		171
		172	...
		173	ev_set_syserr_cb (fatal_error);
		174
102	=back	175	=back
103		176
104	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP	177	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP
105		178
106	An event loop is described by a C<struct ev_loop *>. The library knows two	179	An event loop is described by a C<struct ev_loop *>. The library knows two
…		…
119	=item struct ev_loop *ev_default_loop (unsigned int flags)	192	=item struct ev_loop *ev_default_loop (unsigned int flags)
120		193
121	This will initialise the default event loop if it hasn't been initialised	194	This will initialise the default event loop if it hasn't been initialised
122	yet and return it. If the default loop could not be initialised, returns	195	yet and return it. If the default loop could not be initialised, returns
123	false. If it already was initialised it simply returns it (and ignores the	196	false. If it already was initialised it simply returns it (and ignores the
124	flags).	197	flags. If that is troubling you, check C<ev_backend ()> afterwards).
125		198
126	If you don't know what event loop to use, use the one returned from this	199	If you don't know what event loop to use, use the one returned from this
127	function.	200	function.
128		201
129	The flags argument can be used to specify special behaviour or specific	202	The flags argument can be used to specify special behaviour or specific
130	backends to use, and is usually specified as 0 (or EVFLAG_AUTO).	203	backends to use, and is usually specified as C<0> (or C<EVFLAG_AUTO>).
131		204
132	It supports the following flags:	205	The following flags are supported:
133		206
134	=over 4	207	=over 4
135		208
136	=item C<EVFLAG_AUTO>	209	=item C<EVFLAG_AUTO>
137		210
…		…
145	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will	218	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will
146	override the flags completely if it is found in the environment. This is	219	override the flags completely if it is found in the environment. This is
147	useful to try out specific backends to test their performance, or to work	220	useful to try out specific backends to test their performance, or to work
148	around bugs.	221	around bugs.
149		222
150	=item C<EVMETHOD_SELECT> (portable select backend)	223	=item C<EVBACKEND_SELECT> (value 1, portable select backend)
151		224
		225	This is your standard select(2) backend. Not I<completely> standard, as
		226	libev tries to roll its own fd_set with no limits on the number of fds,
		227	but if that fails, expect a fairly low limit on the number of fds when
		228	using this backend. It doesn't scale too well (O(highest_fd)), but its usually
		229	the fastest backend for a low number of fds.
		230
152	=item C<EVMETHOD_POLL> (poll backend, available everywhere except on windows)	231	=item C<EVBACKEND_POLL> (value 2, poll backend, available everywhere except on windows)
153		232
154	=item C<EVMETHOD_EPOLL> (linux only)	233	And this is your standard poll(2) backend. It's more complicated than
		234	select, but handles sparse fds better and has no artificial limit on the
		235	number of fds you can use (except it will slow down considerably with a
		236	lot of inactive fds). It scales similarly to select, i.e. O(total_fds).
155		237
156	=item C<EVMETHOD_KQUEUE> (some bsds only)	238	=item C<EVBACKEND_EPOLL> (value 4, Linux)
157		239
158	=item C<EVMETHOD_DEVPOLL> (solaris 8 only)	240	For few fds, this backend is a bit little slower than poll and select,
		241	but it scales phenomenally better. While poll and select usually scale like
		242	O(total_fds) where n is the total number of fds (or the highest fd), epoll scales
		243	either O(1) or O(active_fds).
159		244
160	=item C<EVMETHOD_PORT> (solaris 10 only)	245	While stopping and starting an I/O watcher in the same iteration will
		246	result in some caching, there is still a syscall per such incident
		247	(because the fd could point to a different file description now), so its
		248	best to avoid that. Also, dup()ed file descriptors might not work very
		249	well if you register events for both fds.
		250
		251	Please note that epoll sometimes generates spurious notifications, so you
		252	need to use non-blocking I/O or other means to avoid blocking when no data
		253	(or space) is available.
		254
		255	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)
		256
		257	Kqueue deserves special mention, as at the time of this writing, it
		258	was broken on all BSDs except NetBSD (usually it doesn't work with
		259	anything but sockets and pipes, except on Darwin, where of course its
		260	completely useless). For this reason its not being "autodetected"
		261	unless you explicitly specify it explicitly in the flags (i.e. using
		262	C<EVBACKEND_KQUEUE>).
		263
		264	It scales in the same way as the epoll backend, but the interface to the
		265	kernel is more efficient (which says nothing about its actual speed, of
		266	course). While starting and stopping an I/O watcher does not cause an
		267	extra syscall as with epoll, it still adds up to four event changes per
		268	incident, so its best to avoid that.
		269
		270	=item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8)
		271
		272	This is not implemented yet (and might never be).
		273
		274	=item C<EVBACKEND_PORT> (value 32, Solaris 10)
		275
		276	This uses the Solaris 10 port mechanism. As with everything on Solaris,
		277	it's really slow, but it still scales very well (O(active_fds)).
		278
		279	Please note that solaris ports can result in a lot of spurious
		280	notifications, so you need to use non-blocking I/O or other means to avoid
		281	blocking when no data (or space) is available.
		282
		283	=item C<EVBACKEND_ALL>
		284
		285	Try all backends (even potentially broken ones that wouldn't be tried
		286	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as
		287	C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>.
		288
		289	=back
161		290
162	If one or more of these are ored into the flags value, then only these	291	If one or more of these are ored into the flags value, then only these
163	backends will be tried (in the reverse order as given here). If one are	292	backends will be tried (in the reverse order as given here). If none are
164	specified, any backend will do.	293	specified, most compiled-in backend will be tried, usually in reverse
		294	order of their flag values :)
165		295
166	=back	296	The most typical usage is like this:
		297
		298	if (!ev_default_loop (0))
		299	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
		300
		301	Restrict libev to the select and poll backends, and do not allow
		302	environment settings to be taken into account:
		303
		304	ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);
		305
		306	Use whatever libev has to offer, but make sure that kqueue is used if
		307	available (warning, breaks stuff, best use only with your own private
		308	event loop and only if you know the OS supports your types of fds):
		309
		310	ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);
167		311
168	=item struct ev_loop *ev_loop_new (unsigned int flags)	312	=item struct ev_loop *ev_loop_new (unsigned int flags)
169		313
170	Similar to C<ev_default_loop>, but always creates a new event loop that is	314	Similar to C<ev_default_loop>, but always creates a new event loop that is
171	always distinct from the default loop. Unlike the default loop, it cannot	315	always distinct from the default loop. Unlike the default loop, it cannot
172	handle signal and child watchers, and attempts to do so will be greeted by	316	handle signal and child watchers, and attempts to do so will be greeted by
173	undefined behaviour (or a failed assertion if assertions are enabled).	317	undefined behaviour (or a failed assertion if assertions are enabled).
174		318
		319	Example: try to create a event loop that uses epoll and nothing else.
		320
		321	struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);
		322	if (!epoller)
		323	fatal ("no epoll found here, maybe it hides under your chair");
		324
175	=item ev_default_destroy ()	325	=item ev_default_destroy ()
176		326
177	Destroys the default loop again (frees all memory and kernel state	327	Destroys the default loop again (frees all memory and kernel state
178	etc.). This stops all registered event watchers (by not touching them in	328	etc.). None of the active event watchers will be stopped in the normal
179	any way whatsoever, although you cannot rely on this :).	329	sense, so e.g. C<ev_is_active> might still return true. It is your
		330	responsibility to either stop all watchers cleanly yoursef I<before>
		331	calling this function, or cope with the fact afterwards (which is usually
		332	the easiest thing, youc na just ignore the watchers and/or C<free ()> them
		333	for example).
180		334
181	=item ev_loop_destroy (loop)	335	=item ev_loop_destroy (loop)
182		336
183	Like C<ev_default_destroy>, but destroys an event loop created by an	337	Like C<ev_default_destroy>, but destroys an event loop created by an
184	earlier call to C<ev_loop_new>.	338	earlier call to C<ev_loop_new>.
…		…
188	This function reinitialises the kernel state for backends that have	342	This function reinitialises the kernel state for backends that have
189	one. Despite the name, you can call it anytime, but it makes most sense	343	one. Despite the name, you can call it anytime, but it makes most sense
190	after forking, in either the parent or child process (or both, but that	344	after forking, in either the parent or child process (or both, but that
191	again makes little sense).	345	again makes little sense).
192		346
193	You I<must> call this function after forking if and only if you want to	347	You I<must> call this function in the child process after forking if and
194	use the event library in both processes. If you just fork+exec, you don't	348	only if you want to use the event library in both processes. If you just
195	have to call it.	349	fork+exec, you don't have to call it.
196		350
197	The function itself is quite fast and it's usually not a problem to call	351	The function itself is quite fast and it's usually not a problem to call
198	it just in case after a fork. To make this easy, the function will fit in	352	it just in case after a fork. To make this easy, the function will fit in
199	quite nicely into a call to C<pthread_atfork>:	353	quite nicely into a call to C<pthread_atfork>:
200		354
201	pthread_atfork (0, 0, ev_default_fork);	355	pthread_atfork (0, 0, ev_default_fork);
202		356
		357	At the moment, C<EVBACKEND_SELECT> and C<EVBACKEND_POLL> are safe to use
		358	without calling this function, so if you force one of those backends you
		359	do not need to care.
		360
203	=item ev_loop_fork (loop)	361	=item ev_loop_fork (loop)
204		362
205	Like C<ev_default_fork>, but acts on an event loop created by	363	Like C<ev_default_fork>, but acts on an event loop created by
206	C<ev_loop_new>. Yes, you have to call this on every allocated event loop	364	C<ev_loop_new>. Yes, you have to call this on every allocated event loop
207	after fork, and how you do this is entirely your own problem.	365	after fork, and how you do this is entirely your own problem.
208		366
209	=item unsigned int ev_method (loop)	367	=item unsigned int ev_backend (loop)
210		368
211	Returns one of the C<EVMETHOD_*> flags indicating the event backend in	369	Returns one of the C<EVBACKEND_*> flags indicating the event backend in
212	use.	370	use.
213		371
214	=item ev_tstamp ev_now (loop)	372	=item ev_tstamp ev_now (loop)
215		373
216	Returns the current "event loop time", which is the time the event loop	374	Returns the current "event loop time", which is the time the event loop
217	got events and started processing them. This timestamp does not change	375	received events and started processing them. This timestamp does not
218	as long as callbacks are being processed, and this is also the base time	376	change as long as callbacks are being processed, and this is also the base
219	used for relative timers. You can treat it as the timestamp of the event	377	time used for relative timers. You can treat it as the timestamp of the
220	occuring (or more correctly, the mainloop finding out about it).	378	event occuring (or more correctly, libev finding out about it).
221		379
222	=item ev_loop (loop, int flags)	380	=item ev_loop (loop, int flags)
223		381
224	Finally, this is it, the event handler. This function usually is called	382	Finally, this is it, the event handler. This function usually is called
225	after you initialised all your watchers and you want to start handling	383	after you initialised all your watchers and you want to start handling
226	events.	384	events.
227		385
228	If the flags argument is specified as 0, it will not return until either	386	If the flags argument is specified as C<0>, it will not return until
229	no event watchers are active anymore or C<ev_unloop> was called.	387	either no event watchers are active anymore or C<ev_unloop> was called.
		388
		389	Please note that an explicit C<ev_unloop> is usually better than
		390	relying on all watchers to be stopped when deciding when a program has
		391	finished (especially in interactive programs), but having a program that
		392	automatically loops as long as it has to and no longer by virtue of
		393	relying on its watchers stopping correctly is a thing of beauty.
230		394
231	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle	395	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle
232	those events and any outstanding ones, but will not block your process in	396	those events and any outstanding ones, but will not block your process in
233	case there are no events and will return after one iteration of the loop.	397	case there are no events and will return after one iteration of the loop.
234		398
235	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if	399	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if
236	neccessary) and will handle those and any outstanding ones. It will block	400	neccessary) and will handle those and any outstanding ones. It will block
237	your process until at least one new event arrives, and will return after	401	your process until at least one new event arrives, and will return after
238	one iteration of the loop.	402	one iteration of the loop. This is useful if you are waiting for some
		403	external event in conjunction with something not expressible using other
		404	libev watchers. However, a pair of C<ev_prepare>/C<ev_check> watchers is
		405	usually a better approach for this kind of thing.
239		406
240	This flags value could be used to implement alternative looping
241	constructs, but the C<prepare> and C<check> watchers provide a better and
242	more generic mechanism.
243
244	Here are the gory details of what ev_loop does:	407	Here are the gory details of what C<ev_loop> does:
245		408
246	1. If there are no active watchers (reference count is zero), return.	409	* If there are no active watchers (reference count is zero), return.
247	2. Queue and immediately call all prepare watchers.	410	- Queue prepare watchers and then call all outstanding watchers.
248	3. If we have been forked, recreate the kernel state.	411	- If we have been forked, recreate the kernel state.
249	4. Update the kernel state with all outstanding changes.	412	- Update the kernel state with all outstanding changes.
250	5. Update the "event loop time".	413	- Update the "event loop time".
251	6. Calculate for how long to block.	414	- Calculate for how long to block.
252	7. Block the process, waiting for events.	415	- Block the process, waiting for any events.
		416	- Queue all outstanding I/O (fd) events.
253	8. Update the "event loop time" and do time jump handling.	417	- Update the "event loop time" and do time jump handling.
254	9. Queue all outstanding timers.	418	- Queue all outstanding timers.
255	10. Queue all outstanding periodics.	419	- Queue all outstanding periodics.
256	11. If no events are pending now, queue all idle watchers.	420	- If no events are pending now, queue all idle watchers.
257	12. Queue all check watchers.	421	- Queue all check watchers.
258	13. Call all queued watchers in reverse order (i.e. check watchers first).	422	- Call all queued watchers in reverse order (i.e. check watchers first).
		423	Signals and child watchers are implemented as I/O watchers, and will
		424	be handled here by queueing them when their watcher gets executed.
259	14. If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK	425	- If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK
260	was used, return, otherwise continue with step #1.	426	were used, return, otherwise continue with step *.
		427
		428	Example: queue some jobs and then loop until no events are outsanding
		429	anymore.
		430
		431	... queue jobs here, make sure they register event watchers as long
		432	... as they still have work to do (even an idle watcher will do..)
		433	ev_loop (my_loop, 0);
		434	... jobs done. yeah!
261		435
262	=item ev_unloop (loop, how)	436	=item ev_unloop (loop, how)
263		437
264	Can be used to make a call to C<ev_loop> return early (but only after it	438	Can be used to make a call to C<ev_loop> return early (but only after it
265	has processed all outstanding events). The C<how> argument must be either	439	has processed all outstanding events). The C<how> argument must be either
…		…
279	visible to the libev user and should not keep C<ev_loop> from exiting if	453	visible to the libev user and should not keep C<ev_loop> from exiting if
280	no event watchers registered by it are active. It is also an excellent	454	no event watchers registered by it are active. It is also an excellent
281	way to do this for generic recurring timers or from within third-party	455	way to do this for generic recurring timers or from within third-party
282	libraries. Just remember to I<unref after start> and I<ref before stop>.	456	libraries. Just remember to I<unref after start> and I<ref before stop>.
283		457
		458	Example: create a signal watcher, but keep it from keeping C<ev_loop>
		459	running when nothing else is active.
		460
		461	struct dv_signal exitsig;
		462	ev_signal_init (&exitsig, sig_cb, SIGINT);
		463	ev_signal_start (myloop, &exitsig);
		464	evf_unref (myloop);
		465
		466	Example: for some weird reason, unregister the above signal handler again.
		467
		468	ev_ref (myloop);
		469	ev_signal_stop (myloop, &exitsig);
		470
284	=back	471	=back
		472
285		473
286	=head1 ANATOMY OF A WATCHER	474	=head1 ANATOMY OF A WATCHER
287		475
288	A watcher is a structure that you create and register to record your	476	A watcher is a structure that you create and register to record your
289	interest in some event. For instance, if you want to wait for STDIN to	477	interest in some event. For instance, if you want to wait for STDIN to
…		…
322	*) >>), and you can stop watching for events at any time by calling the	510	*) >>), and you can stop watching for events at any time by calling the
323	corresponding stop function (C<< ev_<type>_stop (loop, watcher *) >>.	511	corresponding stop function (C<< ev_<type>_stop (loop, watcher *) >>.
324		512
325	As long as your watcher is active (has been started but not stopped) you	513	As long as your watcher is active (has been started but not stopped) you
326	must not touch the values stored in it. Most specifically you must never	514	must not touch the values stored in it. Most specifically you must never
327	reinitialise it or call its set method.	515	reinitialise it or call its C<set> macro.
328
329	You can check whether an event is active by calling the C<ev_is_active
330	(watcher *)> macro. To see whether an event is outstanding (but the
331	callback for it has not been called yet) you can use the C<ev_is_pending
332	(watcher *)> macro.
333		516
334	Each and every callback receives the event loop pointer as first, the	517	Each and every callback receives the event loop pointer as first, the
335	registered watcher structure as second, and a bitset of received events as	518	registered watcher structure as second, and a bitset of received events as
336	third argument.	519	third argument.
337		520
…		…
361	The signal specified in the C<ev_signal> watcher has been received by a thread.	544	The signal specified in the C<ev_signal> watcher has been received by a thread.
362		545
363	=item C<EV_CHILD>	546	=item C<EV_CHILD>
364		547
365	The pid specified in the C<ev_child> watcher has received a status change.	548	The pid specified in the C<ev_child> watcher has received a status change.
		549
		550	=item C<EV_STAT>
		551
		552	The path specified in the C<ev_stat> watcher changed its attributes somehow.
366		553
367	=item C<EV_IDLE>	554	=item C<EV_IDLE>
368		555
369	The C<ev_idle> watcher has determined that you have nothing better to do.	556	The C<ev_idle> watcher has determined that you have nothing better to do.
370		557
…		…
394	with the error from read() or write(). This will not work in multithreaded	581	with the error from read() or write(). This will not work in multithreaded
395	programs, though, so beware.	582	programs, though, so beware.
396		583
397	=back	584	=back
398		585
		586	=head2 GENERIC WATCHER FUNCTIONS
		587
		588	In the following description, C<TYPE> stands for the watcher type,
		589	e.g. C<timer> for C<ev_timer> watchers and C<io> for C<ev_io> watchers.
		590
		591	=over 4
		592
		593	=item C<ev_init> (ev_TYPE *watcher, callback)
		594
		595	This macro initialises the generic portion of a watcher. The contents
		596	of the watcher object can be arbitrary (so C<malloc> will do). Only
		597	the generic parts of the watcher are initialised, you I<need> to call
		598	the type-specific C<ev_TYPE_set> macro afterwards to initialise the
		599	type-specific parts. For each type there is also a C<ev_TYPE_init> macro
		600	which rolls both calls into one.
		601
		602	You can reinitialise a watcher at any time as long as it has been stopped
		603	(or never started) and there are no pending events outstanding.
		604
		605	The callback is always of type C<void (*)(ev_loop *loop, ev_TYPE *watcher,
		606	int revents)>.
		607
		608	=item C<ev_TYPE_set> (ev_TYPE *, [args])
		609
		610	This macro initialises the type-specific parts of a watcher. You need to
		611	call C<ev_init> at least once before you call this macro, but you can
		612	call C<ev_TYPE_set> any number of times. You must not, however, call this
		613	macro on a watcher that is active (it can be pending, however, which is a
		614	difference to the C<ev_init> macro).
		615
		616	Although some watcher types do not have type-specific arguments
		617	(e.g. C<ev_prepare>) you still need to call its C<set> macro.
		618
		619	=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])
		620
		621	This convinience macro rolls both C<ev_init> and C<ev_TYPE_set> macro
		622	calls into a single call. This is the most convinient method to initialise
		623	a watcher. The same limitations apply, of course.
		624
		625	=item C<ev_TYPE_start> (loop *, ev_TYPE *watcher)
		626
		627	Starts (activates) the given watcher. Only active watchers will receive
		628	events. If the watcher is already active nothing will happen.
		629
		630	=item C<ev_TYPE_stop> (loop *, ev_TYPE *watcher)
		631
		632	Stops the given watcher again (if active) and clears the pending
		633	status. It is possible that stopped watchers are pending (for example,
		634	non-repeating timers are being stopped when they become pending), but
		635	C<ev_TYPE_stop> ensures that the watcher is neither active nor pending. If
		636	you want to free or reuse the memory used by the watcher it is therefore a
		637	good idea to always call its C<ev_TYPE_stop> function.
		638
		639	=item bool ev_is_active (ev_TYPE *watcher)
		640
		641	Returns a true value iff the watcher is active (i.e. it has been started
		642	and not yet been stopped). As long as a watcher is active you must not modify
		643	it.
		644
		645	=item bool ev_is_pending (ev_TYPE *watcher)
		646
		647	Returns a true value iff the watcher is pending, (i.e. it has outstanding
		648	events but its callback has not yet been invoked). As long as a watcher
		649	is pending (but not active) you must not call an init function on it (but
		650	C<ev_TYPE_set> is safe) and you must make sure the watcher is available to
		651	libev (e.g. you cnanot C<free ()> it).
		652
		653	=item callback = ev_cb (ev_TYPE *watcher)
		654
		655	Returns the callback currently set on the watcher.
		656
		657	=item ev_cb_set (ev_TYPE *watcher, callback)
		658
		659	Change the callback. You can change the callback at virtually any time
		660	(modulo threads).
		661
		662	=back
		663
		664
399	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER	665	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER
400		666
401	Each watcher has, by default, a member C<void *data> that you can change	667	Each watcher has, by default, a member C<void *data> that you can change
402	and read at any time, libev will completely ignore it. This can be used	668	and read at any time, libev will completely ignore it. This can be used
403	to associate arbitrary data with your watcher. If you need more data and	669	to associate arbitrary data with your watcher. If you need more data and
…		…
427		693
428		694
429	=head1 WATCHER TYPES	695	=head1 WATCHER TYPES
430		696
431	This section describes each watcher in detail, but will not repeat	697	This section describes each watcher in detail, but will not repeat
432	information given in the last section.	698	information given in the last section. Any initialisation/set macros,
		699	functions and members specific to the watcher type are explained.
433		700
		701	Members are additionally marked with either I<[read-only]>, meaning that,
		702	while the watcher is active, you can look at the member and expect some
		703	sensible content, but you must not modify it (you can modify it while the
		704	watcher is stopped to your hearts content), or I<[read-write]>, which
		705	means you can expect it to have some sensible content while the watcher
		706	is active, but you can also modify it. Modifying it may not do something
		707	sensible or take immediate effect (or do anything at all), but libev will
		708	not crash or malfunction in any way.
		709
		710
434	=head2 C<ev_io> - is this file descriptor readable or writable	711	=head2 C<ev_io> - is this file descriptor readable or writable?
435		712
436	I/O watchers check whether a file descriptor is readable or writable	713	I/O watchers check whether a file descriptor is readable or writable
437	in each iteration of the event loop (This behaviour is called	714	in each iteration of the event loop, or, more precisely, when reading
438	level-triggering because you keep receiving events as long as the	715	would not block the process and writing would at least be able to write
439	condition persists. Remember you can stop the watcher if you don't want to	716	some data. This behaviour is called level-triggering because you keep
440	act on the event and neither want to receive future events).	717	receiving events as long as the condition persists. Remember you can stop
		718	the watcher if you don't want to act on the event and neither want to
		719	receive future events.
441		720
442	In general you can register as many read and/or write event watchers per	721	In general you can register as many read and/or write event watchers per
443	fd as you want (as long as you don't confuse yourself). Setting all file	722	fd as you want (as long as you don't confuse yourself). Setting all file
444	descriptors to non-blocking mode is also usually a good idea (but not	723	descriptors to non-blocking mode is also usually a good idea (but not
445	required if you know what you are doing).	724	required if you know what you are doing).
446		725
447	You have to be careful with dup'ed file descriptors, though. Some backends	726	You have to be careful with dup'ed file descriptors, though. Some backends
448	(the linux epoll backend is a notable example) cannot handle dup'ed file	727	(the linux epoll backend is a notable example) cannot handle dup'ed file
449	descriptors correctly if you register interest in two or more fds pointing	728	descriptors correctly if you register interest in two or more fds pointing
450	to the same underlying file/socket etc. description (that is, they share	729	to the same underlying file/socket/etc. description (that is, they share
451	the same underlying "file open").	730	the same underlying "file open").
452		731
453	If you must do this, then force the use of a known-to-be-good backend	732	If you must do this, then force the use of a known-to-be-good backend
454	(at the time of this writing, this includes only EVMETHOD_SELECT and	733	(at the time of this writing, this includes only C<EVBACKEND_SELECT> and
455	EVMETHOD_POLL).	734	C<EVBACKEND_POLL>).
		735
		736	Another thing you have to watch out for is that it is quite easy to
		737	receive "spurious" readyness notifications, that is your callback might
		738	be called with C<EV_READ> but a subsequent C<read>(2) will actually block
		739	because there is no data. Not only are some backends known to create a
		740	lot of those (for example solaris ports), it is very easy to get into
		741	this situation even with a relatively standard program structure. Thus
		742	it is best to always use non-blocking I/O: An extra C<read>(2) returning
		743	C<EAGAIN> is far preferable to a program hanging until some data arrives.
		744
		745	If you cannot run the fd in non-blocking mode (for example you should not
		746	play around with an Xlib connection), then you have to seperately re-test
		747	wether a file descriptor is really ready with a known-to-be good interface
		748	such as poll (fortunately in our Xlib example, Xlib already does this on
		749	its own, so its quite safe to use).
456		750
457	=over 4	751	=over 4
458		752
459	=item ev_io_init (ev_io *, callback, int fd, int events)	753	=item ev_io_init (ev_io *, callback, int fd, int events)
460		754
461	=item ev_io_set (ev_io *, int fd, int events)	755	=item ev_io_set (ev_io *, int fd, int events)
462		756
463	Configures an C<ev_io> watcher. The fd is the file descriptor to rceeive	757	Configures an C<ev_io> watcher. The C<fd> is the file descriptor to
464	events for and events is either C<EV_READ>, C<EV_WRITE> or C<EV_READ |	758	rceeive events for and events is either C<EV_READ>, C<EV_WRITE> or
465	EV_WRITE> to receive the given events.	759	C<EV_READ | EV_WRITE> to receive the given events.
		760
		761	=item int fd [read-only]
		762
		763	The file descriptor being watched.
		764
		765	=item int events [read-only]
		766
		767	The events being watched.
466		768
467	=back	769	=back
468		770
		771	Example: call C<stdin_readable_cb> when STDIN_FILENO has become, well
		772	readable, but only once. Since it is likely line-buffered, you could
		773	attempt to read a whole line in the callback:
		774
		775	static void
		776	stdin_readable_cb (struct ev_loop *loop, struct ev_io *w, int revents)
		777	{
		778	ev_io_stop (loop, w);
		779	.. read from stdin here (or from w->fd) and haqndle any I/O errors
		780	}
		781
		782	...
		783	struct ev_loop *loop = ev_default_init (0);
		784	struct ev_io stdin_readable;
		785	ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
		786	ev_io_start (loop, &stdin_readable);
		787	ev_loop (loop, 0);
		788
		789
469	=head2 C<ev_timer> - relative and optionally recurring timeouts	790	=head2 C<ev_timer> - relative and optionally repeating timeouts
470		791
471	Timer watchers are simple relative timers that generate an event after a	792	Timer watchers are simple relative timers that generate an event after a
472	given time, and optionally repeating in regular intervals after that.	793	given time, and optionally repeating in regular intervals after that.
473		794
474	The timers are based on real time, that is, if you register an event that	795	The timers are based on real time, that is, if you register an event that
475	times out after an hour and you reset your system clock to last years	796	times out after an hour and you reset your system clock to last years
476	time, it will still time out after (roughly) and hour. "Roughly" because	797	time, it will still time out after (roughly) and hour. "Roughly" because
477	detecting time jumps is hard, and soem inaccuracies are unavoidable (the	798	detecting time jumps is hard, and some inaccuracies are unavoidable (the
478	monotonic clock option helps a lot here).	799	monotonic clock option helps a lot here).
479		800
480	The relative timeouts are calculated relative to the C<ev_now ()>	801	The relative timeouts are calculated relative to the C<ev_now ()>
481	time. This is usually the right thing as this timestamp refers to the time	802	time. This is usually the right thing as this timestamp refers to the time
482	of the event triggering whatever timeout you are modifying/starting. If	803	of the event triggering whatever timeout you are modifying/starting. If
483	you suspect event processing to be delayed and you *need* to base the timeout	804	you suspect event processing to be delayed and you I<need> to base the timeout
484	on the current time, use something like this to adjust for this:	805	on the current time, use something like this to adjust for this:
485		806
486	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);	807	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);
		808
		809	The callback is guarenteed to be invoked only when its timeout has passed,
		810	but if multiple timers become ready during the same loop iteration then
		811	order of execution is undefined.
487		812
488	=over 4	813	=over 4
489		814
490	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)	815	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)
491		816
…		…
511		836
512	If the timer is repeating, either start it if necessary (with the repeat	837	If the timer is repeating, either start it if necessary (with the repeat
513	value), or reset the running timer to the repeat value.	838	value), or reset the running timer to the repeat value.
514		839
515	This sounds a bit complicated, but here is a useful and typical	840	This sounds a bit complicated, but here is a useful and typical
516	example: Imagine you have a tcp connection and you want a so-called idle	841	example: Imagine you have a tcp connection and you want a so-called
517	timeout, that is, you want to be called when there have been, say, 60	842	idle timeout, that is, you want to be called when there have been,
518	seconds of inactivity on the socket. The easiest way to do this is to	843	say, 60 seconds of inactivity on the socket. The easiest way to do
519	configure an C<ev_timer> with after=repeat=60 and calling ev_timer_again each	844	this is to configure an C<ev_timer> with C<after>=C<repeat>=C<60> and calling
520	time you successfully read or write some data. If you go into an idle	845	C<ev_timer_again> each time you successfully read or write some data. If
521	state where you do not expect data to travel on the socket, you can stop	846	you go into an idle state where you do not expect data to travel on the
522	the timer, and again will automatically restart it if need be.	847	socket, you can stop the timer, and again will automatically restart it if
		848	need be.
		849
		850	You can also ignore the C<after> value and C<ev_timer_start> altogether
		851	and only ever use the C<repeat> value:
		852
		853	ev_timer_init (timer, callback, 0., 5.);
		854	ev_timer_again (loop, timer);
		855	...
		856	timer->again = 17.;
		857	ev_timer_again (loop, timer);
		858	...
		859	timer->again = 10.;
		860	ev_timer_again (loop, timer);
		861
		862	This is more efficient then stopping/starting the timer eahc time you want
		863	to modify its timeout value.
		864
		865	=item ev_tstamp repeat [read-write]
		866
		867	The current C<repeat> value. Will be used each time the watcher times out
		868	or C<ev_timer_again> is called and determines the next timeout (if any),
		869	which is also when any modifications are taken into account.
523		870
524	=back	871	=back
525		872
		873	Example: create a timer that fires after 60 seconds.
		874
		875	static void
		876	one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
		877	{
		878	.. one minute over, w is actually stopped right here
		879	}
		880
		881	struct ev_timer mytimer;
		882	ev_timer_init (&mytimer, one_minute_cb, 60., 0.);
		883	ev_timer_start (loop, &mytimer);
		884
		885	Example: create a timeout timer that times out after 10 seconds of
		886	inactivity.
		887
		888	static void
		889	timeout_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
		890	{
		891	.. ten seconds without any activity
		892	}
		893
		894	struct ev_timer mytimer;
		895	ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */
		896	ev_timer_again (&mytimer); /* start timer */
		897	ev_loop (loop, 0);
		898
		899	// and in some piece of code that gets executed on any "activity":
		900	// reset the timeout to start ticking again at 10 seconds
		901	ev_timer_again (&mytimer);
		902
		903
526	=head2 C<ev_periodic> - to cron or not to cron	904	=head2 C<ev_periodic> - to cron or not to cron?
527		905
528	Periodic watchers are also timers of a kind, but they are very versatile	906	Periodic watchers are also timers of a kind, but they are very versatile
529	(and unfortunately a bit complex).	907	(and unfortunately a bit complex).
530		908
531	Unlike C<ev_timer>'s, they are not based on real time (or relative time)	909	Unlike C<ev_timer>'s, they are not based on real time (or relative time)
532	but on wallclock time (absolute time). You can tell a periodic watcher	910	but on wallclock time (absolute time). You can tell a periodic watcher
533	to trigger "at" some specific point in time. For example, if you tell a	911	to trigger "at" some specific point in time. For example, if you tell a
534	periodic watcher to trigger in 10 seconds (by specifiying e.g. c<ev_now ()	912	periodic watcher to trigger in 10 seconds (by specifiying e.g. C<ev_now ()
535	+ 10.>) and then reset your system clock to the last year, then it will	913	+ 10.>) and then reset your system clock to the last year, then it will
536	take a year to trigger the event (unlike an C<ev_timer>, which would trigger	914	take a year to trigger the event (unlike an C<ev_timer>, which would trigger
537	roughly 10 seconds later and of course not if you reset your system time	915	roughly 10 seconds later and of course not if you reset your system time
538	again).	916	again).
539		917
540	They can also be used to implement vastly more complex timers, such as	918	They can also be used to implement vastly more complex timers, such as
541	triggering an event on eahc midnight, local time.	919	triggering an event on eahc midnight, local time.
542		920
		921	As with timers, the callback is guarenteed to be invoked only when the
		922	time (C<at>) has been passed, but if multiple periodic timers become ready
		923	during the same loop iteration then order of execution is undefined.
		924
543	=over 4	925	=over 4
544		926
545	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)	927	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)
546		928
547	=item ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)	929	=item ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)
548		930
549	Lots of arguments, lets sort it out... There are basically three modes of	931	Lots of arguments, lets sort it out... There are basically three modes of
550	operation, and we will explain them from simplest to complex:	932	operation, and we will explain them from simplest to complex:
551
552		933
553	=over 4	934	=over 4
554		935
555	=item * absolute timer (interval = reschedule_cb = 0)	936	=item * absolute timer (interval = reschedule_cb = 0)
556		937
…		…
620	Simply stops and restarts the periodic watcher again. This is only useful	1001	Simply stops and restarts the periodic watcher again. This is only useful
621	when you changed some parameters or the reschedule callback would return	1002	when you changed some parameters or the reschedule callback would return
622	a different time than the last time it was called (e.g. in a crond like	1003	a different time than the last time it was called (e.g. in a crond like
623	program when the crontabs have changed).	1004	program when the crontabs have changed).
624		1005
		1006	=item ev_tstamp interval [read-write]
		1007
		1008	The current interval value. Can be modified any time, but changes only
		1009	take effect when the periodic timer fires or C<ev_periodic_again> is being
		1010	called.
		1011
		1012	=item ev_tstamp (*reschedule_cb)(struct ev_periodic *w, ev_tstamp now) [read-write]
		1013
		1014	The current reschedule callback, or C<0>, if this functionality is
		1015	switched off. Can be changed any time, but changes only take effect when
		1016	the periodic timer fires or C<ev_periodic_again> is being called.
		1017
625	=back	1018	=back
626		1019
		1020	Example: call a callback every hour, or, more precisely, whenever the
		1021	system clock is divisible by 3600. The callback invocation times have
		1022	potentially a lot of jittering, but good long-term stability.
		1023
		1024	static void
		1025	clock_cb (struct ev_loop *loop, struct ev_io *w, int revents)
		1026	{
		1027	... its now a full hour (UTC, or TAI or whatever your clock follows)
		1028	}
		1029
		1030	struct ev_periodic hourly_tick;
		1031	ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);
		1032	ev_periodic_start (loop, &hourly_tick);
		1033
		1034	Example: the same as above, but use a reschedule callback to do it:
		1035
		1036	#include <math.h>
		1037
		1038	static ev_tstamp
		1039	my_scheduler_cb (struct ev_periodic *w, ev_tstamp now)
		1040	{
		1041	return fmod (now, 3600.) + 3600.;
		1042	}
		1043
		1044	ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
		1045
		1046	Example: call a callback every hour, starting now:
		1047
		1048	struct ev_periodic hourly_tick;
		1049	ev_periodic_init (&hourly_tick, clock_cb,
		1050	fmod (ev_now (loop), 3600.), 3600., 0);
		1051	ev_periodic_start (loop, &hourly_tick);
		1052
		1053
627	=head2 C<ev_signal> - signal me when a signal gets signalled	1054	=head2 C<ev_signal> - signal me when a signal gets signalled!
628		1055
629	Signal watchers will trigger an event when the process receives a specific	1056	Signal watchers will trigger an event when the process receives a specific
630	signal one or more times. Even though signals are very asynchronous, libev	1057	signal one or more times. Even though signals are very asynchronous, libev
631	will try it's best to deliver signals synchronously, i.e. as part of the	1058	will try it's best to deliver signals synchronously, i.e. as part of the
632	normal event processing, like any other event.	1059	normal event processing, like any other event.
…		…
645	=item ev_signal_set (ev_signal *, int signum)	1072	=item ev_signal_set (ev_signal *, int signum)
646		1073
647	Configures the watcher to trigger on the given signal number (usually one	1074	Configures the watcher to trigger on the given signal number (usually one
648	of the C<SIGxxx> constants).	1075	of the C<SIGxxx> constants).
649		1076
		1077	=item int signum [read-only]
		1078
		1079	The signal the watcher watches out for.
		1080
650	=back	1081	=back
651		1082
		1083
652	=head2 C<ev_child> - wait for pid status changes	1084	=head2 C<ev_child> - watch out for process status changes
653		1085
654	Child watchers trigger when your process receives a SIGCHLD in response to	1086	Child watchers trigger when your process receives a SIGCHLD in response to
655	some child status changes (most typically when a child of yours dies).	1087	some child status changes (most typically when a child of yours dies).
656		1088
657	=over 4	1089	=over 4
…		…
665	at the C<rstatus> member of the C<ev_child> watcher structure to see	1097	at the C<rstatus> member of the C<ev_child> watcher structure to see
666	the status word (use the macros from C<sys/wait.h> and see your systems	1098	the status word (use the macros from C<sys/wait.h> and see your systems
667	C<waitpid> documentation). The C<rpid> member contains the pid of the	1099	C<waitpid> documentation). The C<rpid> member contains the pid of the
668	process causing the status change.	1100	process causing the status change.
669		1101
		1102	=item int pid [read-only]
		1103
		1104	The process id this watcher watches out for, or C<0>, meaning any process id.
		1105
		1106	=item int rpid [read-write]
		1107
		1108	The process id that detected a status change.
		1109
		1110	=item int rstatus [read-write]
		1111
		1112	The process exit/trace status caused by C<rpid> (see your systems
		1113	C<waitpid> and C<sys/wait.h> documentation for details).
		1114
670	=back	1115	=back
671		1116
		1117	Example: try to exit cleanly on SIGINT and SIGTERM.
		1118
		1119	static void
		1120	sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)
		1121	{
		1122	ev_unloop (loop, EVUNLOOP_ALL);
		1123	}
		1124
		1125	struct ev_signal signal_watcher;
		1126	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
		1127	ev_signal_start (loop, &sigint_cb);
		1128
		1129
		1130	=head2 C<ev_stat> - did the file attributes just change?
		1131
		1132	This watches a filesystem path for attribute changes. That is, it calls
		1133	C<stat> regularly (or when the OS says it changed) and sees if it changed
		1134	compared to the last time, invoking the callback if it did.
		1135
		1136	The path does not need to exist: changing from "path exists" to "path does
		1137	not exist" is a status change like any other. The condition "path does
		1138	not exist" is signified by the C<st_nlink> field being zero (which is
		1139	otherwise always forced to be at least one) and all the other fields of
		1140	the stat buffer having unspecified contents.
		1141
		1142	Since there is no standard to do this, the portable implementation simply
		1143	calls C<stat (2)> regulalry on the path to see if it changed somehow. You
		1144	can specify a recommended polling interval for this case. If you specify
		1145	a polling interval of C<0> (highly recommended!) then a I<suitable,
		1146	unspecified default> value will be used (which you can expect to be around
		1147	five seconds, although this might change dynamically). Libev will also
		1148	impose a minimum interval which is currently around C<0.1>, but thats
		1149	usually overkill.
		1150
		1151	This watcher type is not meant for massive numbers of stat watchers,
		1152	as even with OS-supported change notifications, this can be
		1153	resource-intensive.
		1154
		1155	At the time of this writing, no specific OS backends are implemented, but
		1156	if demand increases, at least a kqueue and inotify backend will be added.
		1157
		1158	=over 4
		1159
		1160	=item ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)
		1161
		1162	=item ev_stat_set (ev_stat *, const char *path, ev_tstamp interval)
		1163
		1164	Configures the watcher to wait for status changes of the given
		1165	C<path>. The C<interval> is a hint on how quickly a change is expected to
		1166	be detected and should normally be specified as C<0> to let libev choose
		1167	a suitable value. The memory pointed to by C<path> must point to the same
		1168	path for as long as the watcher is active.
		1169
		1170	The callback will be receive C<EV_STAT> when a change was detected,
		1171	relative to the attributes at the time the watcher was started (or the
		1172	last change was detected).
		1173
		1174	=item ev_stat_stat (ev_stat *)
		1175
		1176	Updates the stat buffer immediately with new values. If you change the
		1177	watched path in your callback, you could call this fucntion to avoid
		1178	detecting this change (while introducing a race condition). Can also be
		1179	useful simply to find out the new values.
		1180
		1181	=item ev_statdata attr [read-only]
		1182
		1183	The most-recently detected attributes of the file. Although the type is of
		1184	C<ev_statdata>, this is usually the (or one of the) C<struct stat> types
		1185	suitable for your system. If the C<st_nlink> member is C<0>, then there
		1186	was some error while C<stat>ing the file.
		1187
		1188	=item ev_statdata prev [read-only]
		1189
		1190	The previous attributes of the file. The callback gets invoked whenever
		1191	C<prev> != C<attr>.
		1192
		1193	=item ev_tstamp interval [read-only]
		1194
		1195	The specified interval.
		1196
		1197	=item const char *path [read-only]
		1198
		1199	The filesystem path that is being watched.
		1200
		1201	=back
		1202
		1203	Example: Watch C</etc/passwd> for attribute changes.
		1204
		1205	static void
		1206	passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)
		1207	{
		1208	/* /etc/passwd changed in some way */
		1209	if (w->attr.st_nlink)
		1210	{
		1211	printf ("passwd current size %ld\n", (long)w->attr.st_size);
		1212	printf ("passwd current atime %ld\n", (long)w->attr.st_mtime);
		1213	printf ("passwd current mtime %ld\n", (long)w->attr.st_mtime);
		1214	}
		1215	else
		1216	/* you shalt not abuse printf for puts */
		1217	puts ("wow, /etc/passwd is not there, expect problems. "
		1218	"if this is windows, they already arrived\n");
		1219	}
		1220
		1221	...
		1222	ev_stat passwd;
		1223
		1224	ev_stat_init (&passwd, passwd_cb, "/etc/passwd");
		1225	ev_stat_start (loop, &passwd);
		1226
		1227
672	=head2 C<ev_idle> - when you've got nothing better to do	1228	=head2 C<ev_idle> - when you've got nothing better to do...
673		1229
674	Idle watchers trigger events when there are no other events are pending	1230	Idle watchers trigger events when there are no other events are pending
675	(prepare, check and other idle watchers do not count). That is, as long	1231	(prepare, check and other idle watchers do not count). That is, as long
676	as your process is busy handling sockets or timeouts (or even signals,	1232	as your process is busy handling sockets or timeouts (or even signals,
677	imagine) it will not be triggered. But when your process is idle all idle	1233	imagine) it will not be triggered. But when your process is idle all idle
…		…
695	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,	1251	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,
696	believe me.	1252	believe me.
697		1253
698	=back	1254	=back
699		1255
		1256	Example: dynamically allocate an C<ev_idle>, start it, and in the
		1257	callback, free it. Alos, use no error checking, as usual.
		1258
		1259	static void
		1260	idle_cb (struct ev_loop *loop, struct ev_idle *w, int revents)
		1261	{
		1262	free (w);
		1263	// now do something you wanted to do when the program has
		1264	// no longer asnything immediate to do.
		1265	}
		1266
		1267	struct ev_idle *idle_watcher = malloc (sizeof (struct ev_idle));
		1268	ev_idle_init (idle_watcher, idle_cb);
		1269	ev_idle_start (loop, idle_cb);
		1270
		1271
700	=head2 C<ev_prepare> and C<ev_check> - customise your event loop	1272	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!
701		1273
702	Prepare and check watchers are usually (but not always) used in tandem:	1274	Prepare and check watchers are usually (but not always) used in tandem:
703	prepare watchers get invoked before the process blocks and check watchers	1275	prepare watchers get invoked before the process blocks and check watchers
704	afterwards.	1276	afterwards.
705		1277
		1278	You I<must not> call C<ev_loop> or similar functions that enter
		1279	the current event loop from either C<ev_prepare> or C<ev_check>
		1280	watchers. Other loops than the current one are fine, however. The
		1281	rationale behind this is that you do not need to check for recursion in
		1282	those watchers, i.e. the sequence will always be C<ev_prepare>, blocking,
		1283	C<ev_check> so if you have one watcher of each kind they will always be
		1284	called in pairs bracketing the blocking call.
		1285
706	Their main purpose is to integrate other event mechanisms into libev. This	1286	Their main purpose is to integrate other event mechanisms into libev and
707	could be used, for example, to track variable changes, implement your own	1287	their use is somewhat advanced. This could be used, for example, to track
708	watchers, integrate net-snmp or a coroutine library and lots more.	1288	variable changes, implement your own watchers, integrate net-snmp or a
		1289	coroutine library and lots more. They are also occasionally useful if
		1290	you cache some data and want to flush it before blocking (for example,
		1291	in X programs you might want to do an C<XFlush ()> in an C<ev_prepare>
		1292	watcher).
709		1293
710	This is done by examining in each prepare call which file descriptors need	1294	This is done by examining in each prepare call which file descriptors need
711	to be watched by the other library, registering C<ev_io> watchers for	1295	to be watched by the other library, registering C<ev_io> watchers for
712	them and starting an C<ev_timer> watcher for any timeouts (many libraries	1296	them and starting an C<ev_timer> watcher for any timeouts (many libraries
713	provide just this functionality). Then, in the check watcher you check for	1297	provide just this functionality). Then, in the check watcher you check for
…		…
735	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>	1319	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>
736	macros, but using them is utterly, utterly and completely pointless.	1320	macros, but using them is utterly, utterly and completely pointless.
737		1321
738	=back	1322	=back
739		1323
		1324	Example: To include a library such as adns, you would add IO watchers
		1325	and a timeout watcher in a prepare handler, as required by libadns, and
		1326	in a check watcher, destroy them and call into libadns. What follows is
		1327	pseudo-code only of course:
		1328
		1329	static ev_io iow [nfd];
		1330	static ev_timer tw;
		1331
		1332	static void
		1333	io_cb (ev_loop *loop, ev_io *w, int revents)
		1334	{
		1335	// set the relevant poll flags
		1336	// could also call adns_processreadable etc. here
		1337	struct pollfd *fd = (struct pollfd *)w->data;
		1338	if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;
		1339	if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;
		1340	}
		1341
		1342	// create io watchers for each fd and a timer before blocking
		1343	static void
		1344	adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents)
		1345	{
		1346	int timeout = 3600000;truct pollfd fds [nfd];
		1347	// actual code will need to loop here and realloc etc.
		1348	adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));
		1349
		1350	/* the callback is illegal, but won't be called as we stop during check */
		1351	ev_timer_init (&tw, 0, timeout * 1e-3);
		1352	ev_timer_start (loop, &tw);
		1353
		1354	// create on ev_io per pollfd
		1355	for (int i = 0; i < nfd; ++i)
		1356	{
		1357	ev_io_init (iow + i, io_cb, fds [i].fd,
		1358	((fds [i].events & POLLIN ? EV_READ : 0)
		1359	| (fds [i].events & POLLOUT ? EV_WRITE : 0)));
		1360
		1361	fds [i].revents = 0;
		1362	iow [i].data = fds + i;
		1363	ev_io_start (loop, iow + i);
		1364	}
		1365	}
		1366
		1367	// stop all watchers after blocking
		1368	static void
		1369	adns_check_cb (ev_loop *loop, ev_check *w, int revents)
		1370	{
		1371	ev_timer_stop (loop, &tw);
		1372
		1373	for (int i = 0; i < nfd; ++i)
		1374	ev_io_stop (loop, iow + i);
		1375
		1376	adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));
		1377	}
		1378
		1379
		1380	=head2 C<ev_embed> - when one backend isn't enough...
		1381
		1382	This is a rather advanced watcher type that lets you embed one event loop
		1383	into another (currently only C<ev_io> events are supported in the embedded
		1384	loop, other types of watchers might be handled in a delayed or incorrect
		1385	fashion and must not be used).
		1386
		1387	There are primarily two reasons you would want that: work around bugs and
		1388	prioritise I/O.
		1389
		1390	As an example for a bug workaround, the kqueue backend might only support
		1391	sockets on some platform, so it is unusable as generic backend, but you
		1392	still want to make use of it because you have many sockets and it scales
		1393	so nicely. In this case, you would create a kqueue-based loop and embed it
		1394	into your default loop (which might use e.g. poll). Overall operation will
		1395	be a bit slower because first libev has to poll and then call kevent, but
		1396	at least you can use both at what they are best.
		1397
		1398	As for prioritising I/O: rarely you have the case where some fds have
		1399	to be watched and handled very quickly (with low latency), and even
		1400	priorities and idle watchers might have too much overhead. In this case
		1401	you would put all the high priority stuff in one loop and all the rest in
		1402	a second one, and embed the second one in the first.
		1403
		1404	As long as the watcher is active, the callback will be invoked every time
		1405	there might be events pending in the embedded loop. The callback must then
		1406	call C<ev_embed_sweep (mainloop, watcher)> to make a single sweep and invoke
		1407	their callbacks (you could also start an idle watcher to give the embedded
		1408	loop strictly lower priority for example). You can also set the callback
		1409	to C<0>, in which case the embed watcher will automatically execute the
		1410	embedded loop sweep.
		1411
		1412	As long as the watcher is started it will automatically handle events. The
		1413	callback will be invoked whenever some events have been handled. You can
		1414	set the callback to C<0> to avoid having to specify one if you are not
		1415	interested in that.
		1416
		1417	Also, there have not currently been made special provisions for forking:
		1418	when you fork, you not only have to call C<ev_loop_fork> on both loops,
		1419	but you will also have to stop and restart any C<ev_embed> watchers
		1420	yourself.
		1421
		1422	Unfortunately, not all backends are embeddable, only the ones returned by
		1423	C<ev_embeddable_backends> are, which, unfortunately, does not include any
		1424	portable one.
		1425
		1426	So when you want to use this feature you will always have to be prepared
		1427	that you cannot get an embeddable loop. The recommended way to get around
		1428	this is to have a separate variables for your embeddable loop, try to
		1429	create it, and if that fails, use the normal loop for everything:
		1430
		1431	struct ev_loop *loop_hi = ev_default_init (0);
		1432	struct ev_loop *loop_lo = 0;
		1433	struct ev_embed embed;
		1434
		1435	// see if there is a chance of getting one that works
		1436	// (remember that a flags value of 0 means autodetection)
		1437	loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
		1438	? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
		1439	: 0;
		1440
		1441	// if we got one, then embed it, otherwise default to loop_hi
		1442	if (loop_lo)
		1443	{
		1444	ev_embed_init (&embed, 0, loop_lo);
		1445	ev_embed_start (loop_hi, &embed);
		1446	}
		1447	else
		1448	loop_lo = loop_hi;
		1449
		1450	=over 4
		1451
		1452	=item ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)
		1453
		1454	=item ev_embed_set (ev_embed *, callback, struct ev_loop *embedded_loop)
		1455
		1456	Configures the watcher to embed the given loop, which must be
		1457	embeddable. If the callback is C<0>, then C<ev_embed_sweep> will be
		1458	invoked automatically, otherwise it is the responsibility of the callback
		1459	to invoke it (it will continue to be called until the sweep has been done,
		1460	if you do not want thta, you need to temporarily stop the embed watcher).
		1461
		1462	=item ev_embed_sweep (loop, ev_embed *)
		1463
		1464	Make a single, non-blocking sweep over the embedded loop. This works
		1465	similarly to C<ev_loop (embedded_loop, EVLOOP_NONBLOCK)>, but in the most
		1466	apropriate way for embedded loops.
		1467
		1468	=item struct ev_loop *loop [read-only]
		1469
		1470	The embedded event loop.
		1471
		1472	=back
		1473
		1474
740	=head1 OTHER FUNCTIONS	1475	=head1 OTHER FUNCTIONS
741		1476
742	There are some other functions of possible interest. Described. Here. Now.	1477	There are some other functions of possible interest. Described. Here. Now.
743		1478
744	=over 4	1479	=over 4
…		…
773	/* stdin might have data for us, joy! */;	1508	/* stdin might have data for us, joy! */;
774	}	1509	}
775		1510
776	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);	1511	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
777		1512
778	=item ev_feed_event (loop, watcher, int events)	1513	=item ev_feed_event (ev_loop *, watcher *, int revents)
779		1514
780	Feeds the given event set into the event loop, as if the specified event	1515	Feeds the given event set into the event loop, as if the specified event
781	had happened for the specified watcher (which must be a pointer to an	1516	had happened for the specified watcher (which must be a pointer to an
782	initialised but not necessarily started event watcher).	1517	initialised but not necessarily started event watcher).
783		1518
784	=item ev_feed_fd_event (loop, int fd, int revents)	1519	=item ev_feed_fd_event (ev_loop *, int fd, int revents)
785		1520
786	Feed an event on the given fd, as if a file descriptor backend detected	1521	Feed an event on the given fd, as if a file descriptor backend detected
787	the given events it.	1522	the given events it.
788		1523
789	=item ev_feed_signal_event (loop, int signum)	1524	=item ev_feed_signal_event (ev_loop *loop, int signum)
790		1525
791	Feed an event as if the given signal occured (loop must be the default loop!).	1526	Feed an event as if the given signal occured (C<loop> must be the default
		1527	loop!).
792		1528
793	=back	1529	=back
		1530
794		1531
795	=head1 LIBEVENT EMULATION	1532	=head1 LIBEVENT EMULATION
796		1533
797	Libev offers a compatibility emulation layer for libevent. It cannot	1534	Libev offers a compatibility emulation layer for libevent. It cannot
798	emulate the internals of libevent, so here are some usage hints:	1535	emulate the internals of libevent, so here are some usage hints:
…		…
819		1556
820	=back	1557	=back
821		1558
822	=head1 C++ SUPPORT	1559	=head1 C++ SUPPORT
823		1560
824	TBD.	1561	Libev comes with some simplistic wrapper classes for C++ that mainly allow
		1562	you to use some convinience methods to start/stop watchers and also change
		1563	the callback model to a model using method callbacks on objects.
		1564
		1565	To use it,
		1566
		1567	#include <ev++.h>
		1568
		1569	(it is not installed by default). This automatically includes F<ev.h>
		1570	and puts all of its definitions (many of them macros) into the global
		1571	namespace. All C++ specific things are put into the C<ev> namespace.
		1572
		1573	It should support all the same embedding options as F<ev.h>, most notably
		1574	C<EV_MULTIPLICITY>.
		1575
		1576	Here is a list of things available in the C<ev> namespace:
		1577
		1578	=over 4
		1579
		1580	=item C<ev::READ>, C<ev::WRITE> etc.
		1581
		1582	These are just enum values with the same values as the C<EV_READ> etc.
		1583	macros from F<ev.h>.
		1584
		1585	=item C<ev::tstamp>, C<ev::now>
		1586
		1587	Aliases to the same types/functions as with the C<ev_> prefix.
		1588
		1589	=item C<ev::io>, C<ev::timer>, C<ev::periodic>, C<ev::idle>, C<ev::sig> etc.
		1590
		1591	For each C<ev_TYPE> watcher in F<ev.h> there is a corresponding class of
		1592	the same name in the C<ev> namespace, with the exception of C<ev_signal>
		1593	which is called C<ev::sig> to avoid clashes with the C<signal> macro
		1594	defines by many implementations.
		1595
		1596	All of those classes have these methods:
		1597
		1598	=over 4
		1599
		1600	=item ev::TYPE::TYPE (object *, object::method *)
		1601
		1602	=item ev::TYPE::TYPE (object *, object::method *, struct ev_loop *)
		1603
		1604	=item ev::TYPE::~TYPE
		1605
		1606	The constructor takes a pointer to an object and a method pointer to
		1607	the event handler callback to call in this class. The constructor calls
		1608	C<ev_init> for you, which means you have to call the C<set> method
		1609	before starting it. If you do not specify a loop then the constructor
		1610	automatically associates the default loop with this watcher.
		1611
		1612	The destructor automatically stops the watcher if it is active.
		1613
		1614	=item w->set (struct ev_loop *)
		1615
		1616	Associates a different C<struct ev_loop> with this watcher. You can only
		1617	do this when the watcher is inactive (and not pending either).
		1618
		1619	=item w->set ([args])
		1620
		1621	Basically the same as C<ev_TYPE_set>, with the same args. Must be
		1622	called at least once. Unlike the C counterpart, an active watcher gets
		1623	automatically stopped and restarted.
		1624
		1625	=item w->start ()
		1626
		1627	Starts the watcher. Note that there is no C<loop> argument as the
		1628	constructor already takes the loop.
		1629
		1630	=item w->stop ()
		1631
		1632	Stops the watcher if it is active. Again, no C<loop> argument.
		1633
		1634	=item w->again () C<ev::timer>, C<ev::periodic> only
		1635
		1636	For C<ev::timer> and C<ev::periodic>, this invokes the corresponding
		1637	C<ev_TYPE_again> function.
		1638
		1639	=item w->sweep () C<ev::embed> only
		1640
		1641	Invokes C<ev_embed_sweep>.
		1642
		1643	=item w->update () C<ev::stat> only
		1644
		1645	Invokes C<ev_stat_stat>.
		1646
		1647	=back
		1648
		1649	=back
		1650
		1651	Example: Define a class with an IO and idle watcher, start one of them in
		1652	the constructor.
		1653
		1654	class myclass
		1655	{
		1656	ev_io io; void io_cb (ev::io &w, int revents);
		1657	ev_idle idle void idle_cb (ev::idle &w, int revents);
		1658
		1659	myclass ();
		1660	}
		1661
		1662	myclass::myclass (int fd)
		1663	: io (this, &myclass::io_cb),
		1664	idle (this, &myclass::idle_cb)
		1665	{
		1666	io.start (fd, ev::READ);
		1667	}
		1668
		1669	=head1 EMBEDDING
		1670
		1671	Libev can (and often is) directly embedded into host
		1672	applications. Examples of applications that embed it include the Deliantra
		1673	Game Server, the EV perl module, the GNU Virtual Private Ethernet (gvpe)
		1674	and rxvt-unicode.
		1675
		1676	The goal is to enable you to just copy the neecssary files into your
		1677	source directory without having to change even a single line in them, so
		1678	you can easily upgrade by simply copying (or having a checked-out copy of
		1679	libev somewhere in your source tree).
		1680
		1681	=head2 FILESETS
		1682
		1683	Depending on what features you need you need to include one or more sets of files
		1684	in your app.
		1685
		1686	=head3 CORE EVENT LOOP
		1687
		1688	To include only the libev core (all the C<ev_*> functions), with manual
		1689	configuration (no autoconf):
		1690
		1691	#define EV_STANDALONE 1
		1692	#include "ev.c"
		1693
		1694	This will automatically include F<ev.h>, too, and should be done in a
		1695	single C source file only to provide the function implementations. To use
		1696	it, do the same for F<ev.h> in all files wishing to use this API (best
		1697	done by writing a wrapper around F<ev.h> that you can include instead and
		1698	where you can put other configuration options):
		1699
		1700	#define EV_STANDALONE 1
		1701	#include "ev.h"
		1702
		1703	Both header files and implementation files can be compiled with a C++
		1704	compiler (at least, thats a stated goal, and breakage will be treated
		1705	as a bug).
		1706
		1707	You need the following files in your source tree, or in a directory
		1708	in your include path (e.g. in libev/ when using -Ilibev):
		1709
		1710	ev.h
		1711	ev.c
		1712	ev_vars.h
		1713	ev_wrap.h
		1714
		1715	ev_win32.c required on win32 platforms only
		1716
		1717	ev_select.c only when select backend is enabled (which is by default)
		1718	ev_poll.c only when poll backend is enabled (disabled by default)
		1719	ev_epoll.c only when the epoll backend is enabled (disabled by default)
		1720	ev_kqueue.c only when the kqueue backend is enabled (disabled by default)
		1721	ev_port.c only when the solaris port backend is enabled (disabled by default)
		1722
		1723	F<ev.c> includes the backend files directly when enabled, so you only need
		1724	to compile this single file.
		1725
		1726	=head3 LIBEVENT COMPATIBILITY API
		1727
		1728	To include the libevent compatibility API, also include:
		1729
		1730	#include "event.c"
		1731
		1732	in the file including F<ev.c>, and:
		1733
		1734	#include "event.h"
		1735
		1736	in the files that want to use the libevent API. This also includes F<ev.h>.
		1737
		1738	You need the following additional files for this:
		1739
		1740	event.h
		1741	event.c
		1742
		1743	=head3 AUTOCONF SUPPORT
		1744
		1745	Instead of using C<EV_STANDALONE=1> and providing your config in
		1746	whatever way you want, you can also C<m4_include([libev.m4])> in your
		1747	F<configure.ac> and leave C<EV_STANDALONE> undefined. F<ev.c> will then
		1748	include F<config.h> and configure itself accordingly.
		1749
		1750	For this of course you need the m4 file:
		1751
		1752	libev.m4
		1753
		1754	=head2 PREPROCESSOR SYMBOLS/MACROS
		1755
		1756	Libev can be configured via a variety of preprocessor symbols you have to define
		1757	before including any of its files. The default is not to build for multiplicity
		1758	and only include the select backend.
		1759
		1760	=over 4
		1761
		1762	=item EV_STANDALONE
		1763
		1764	Must always be C<1> if you do not use autoconf configuration, which
		1765	keeps libev from including F<config.h>, and it also defines dummy
		1766	implementations for some libevent functions (such as logging, which is not
		1767	supported). It will also not define any of the structs usually found in
		1768	F<event.h> that are not directly supported by the libev core alone.
		1769
		1770	=item EV_USE_MONOTONIC
		1771
		1772	If defined to be C<1>, libev will try to detect the availability of the
		1773	monotonic clock option at both compiletime and runtime. Otherwise no use
		1774	of the monotonic clock option will be attempted. If you enable this, you
		1775	usually have to link against librt or something similar. Enabling it when
		1776	the functionality isn't available is safe, though, althoguh you have
		1777	to make sure you link against any libraries where the C<clock_gettime>
		1778	function is hiding in (often F<-lrt>).
		1779
		1780	=item EV_USE_REALTIME
		1781
		1782	If defined to be C<1>, libev will try to detect the availability of the
		1783	realtime clock option at compiletime (and assume its availability at
		1784	runtime if successful). Otherwise no use of the realtime clock option will
		1785	be attempted. This effectively replaces C<gettimeofday> by C<clock_get
		1786	(CLOCK_REALTIME, ...)> and will not normally affect correctness. See tzhe note about libraries
		1787	in the description of C<EV_USE_MONOTONIC>, though.
		1788
		1789	=item EV_USE_SELECT
		1790
		1791	If undefined or defined to be C<1>, libev will compile in support for the
		1792	C<select>(2) backend. No attempt at autodetection will be done: if no
		1793	other method takes over, select will be it. Otherwise the select backend
		1794	will not be compiled in.
		1795
		1796	=item EV_SELECT_USE_FD_SET
		1797
		1798	If defined to C<1>, then the select backend will use the system C<fd_set>
		1799	structure. This is useful if libev doesn't compile due to a missing
		1800	C<NFDBITS> or C<fd_mask> definition or it misguesses the bitset layout on
		1801	exotic systems. This usually limits the range of file descriptors to some
		1802	low limit such as 1024 or might have other limitations (winsocket only
		1803	allows 64 sockets). The C<FD_SETSIZE> macro, set before compilation, might
		1804	influence the size of the C<fd_set> used.
		1805
		1806	=item EV_SELECT_IS_WINSOCKET
		1807
		1808	When defined to C<1>, the select backend will assume that
		1809	select/socket/connect etc. don't understand file descriptors but
		1810	wants osf handles on win32 (this is the case when the select to
		1811	be used is the winsock select). This means that it will call
		1812	C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,
		1813	it is assumed that all these functions actually work on fds, even
		1814	on win32. Should not be defined on non-win32 platforms.
		1815
		1816	=item EV_USE_POLL
		1817
		1818	If defined to be C<1>, libev will compile in support for the C<poll>(2)
		1819	backend. Otherwise it will be enabled on non-win32 platforms. It
		1820	takes precedence over select.
		1821
		1822	=item EV_USE_EPOLL
		1823
		1824	If defined to be C<1>, libev will compile in support for the Linux
		1825	C<epoll>(7) backend. Its availability will be detected at runtime,
		1826	otherwise another method will be used as fallback. This is the
		1827	preferred backend for GNU/Linux systems.
		1828
		1829	=item EV_USE_KQUEUE
		1830
		1831	If defined to be C<1>, libev will compile in support for the BSD style
		1832	C<kqueue>(2) backend. Its actual availability will be detected at runtime,
		1833	otherwise another method will be used as fallback. This is the preferred
		1834	backend for BSD and BSD-like systems, although on most BSDs kqueue only
		1835	supports some types of fds correctly (the only platform we found that
		1836	supports ptys for example was NetBSD), so kqueue might be compiled in, but
		1837	not be used unless explicitly requested. The best way to use it is to find
		1838	out whether kqueue supports your type of fd properly and use an embedded
		1839	kqueue loop.
		1840
		1841	=item EV_USE_PORT
		1842
		1843	If defined to be C<1>, libev will compile in support for the Solaris
		1844	10 port style backend. Its availability will be detected at runtime,
		1845	otherwise another method will be used as fallback. This is the preferred
		1846	backend for Solaris 10 systems.
		1847
		1848	=item EV_USE_DEVPOLL
		1849
		1850	reserved for future expansion, works like the USE symbols above.
		1851
		1852	=item EV_H
		1853
		1854	The name of the F<ev.h> header file used to include it. The default if
		1855	undefined is C<< <ev.h> >> in F<event.h> and C<"ev.h"> in F<ev.c>. This
		1856	can be used to virtually rename the F<ev.h> header file in case of conflicts.
		1857
		1858	=item EV_CONFIG_H
		1859
		1860	If C<EV_STANDALONE> isn't C<1>, this variable can be used to override
		1861	F<ev.c>'s idea of where to find the F<config.h> file, similarly to
		1862	C<EV_H>, above.
		1863
		1864	=item EV_EVENT_H
		1865
		1866	Similarly to C<EV_H>, this macro can be used to override F<event.c>'s idea
		1867	of how the F<event.h> header can be found.
		1868
		1869	=item EV_PROTOTYPES
		1870
		1871	If defined to be C<0>, then F<ev.h> will not define any function
		1872	prototypes, but still define all the structs and other symbols. This is
		1873	occasionally useful if you want to provide your own wrapper functions
		1874	around libev functions.
		1875
		1876	=item EV_MULTIPLICITY
		1877
		1878	If undefined or defined to C<1>, then all event-loop-specific functions
		1879	will have the C<struct ev_loop *> as first argument, and you can create
		1880	additional independent event loops. Otherwise there will be no support
		1881	for multiple event loops and there is no first event loop pointer
		1882	argument. Instead, all functions act on the single default loop.
		1883
		1884	=item EV_PERIODIC_ENABLE
		1885
		1886	If undefined or defined to be C<1>, then periodic timers are supported. If
		1887	defined to be C<0>, then they are not. Disabling them saves a few kB of
		1888	code.
		1889
		1890	=item EV_EMBED_ENABLE
		1891
		1892	If undefined or defined to be C<1>, then embed watchers are supported. If
		1893	defined to be C<0>, then they are not.
		1894
		1895	=item EV_STAT_ENABLE
		1896
		1897	If undefined or defined to be C<1>, then stat watchers are supported. If
		1898	defined to be C<0>, then they are not.
		1899
		1900	=item EV_MINIMAL
		1901
		1902	If you need to shave off some kilobytes of code at the expense of some
		1903	speed, define this symbol to C<1>. Currently only used for gcc to override
		1904	some inlining decisions, saves roughly 30% codesize of amd64.
		1905
		1906	=item EV_COMMON
		1907
		1908	By default, all watchers have a C<void *data> member. By redefining
		1909	this macro to a something else you can include more and other types of
		1910	members. You have to define it each time you include one of the files,
		1911	though, and it must be identical each time.
		1912
		1913	For example, the perl EV module uses something like this:
		1914
		1915	#define EV_COMMON \
		1916	SV *self; /* contains this struct */ \
		1917	SV *cb_sv, *fh /* note no trailing ";" */
		1918
		1919	=item EV_CB_DECLARE (type)
		1920
		1921	=item EV_CB_INVOKE (watcher, revents)
		1922
		1923	=item ev_set_cb (ev, cb)
		1924
		1925	Can be used to change the callback member declaration in each watcher,
		1926	and the way callbacks are invoked and set. Must expand to a struct member
		1927	definition and a statement, respectively. See the F<ev.v> header file for
		1928	their default definitions. One possible use for overriding these is to
		1929	avoid the C<struct ev_loop *> as first argument in all cases, or to use
		1930	method calls instead of plain function calls in C++.
		1931
		1932	=head2 EXAMPLES
		1933
		1934	For a real-world example of a program the includes libev
		1935	verbatim, you can have a look at the EV perl module
		1936	(L<http://software.schmorp.de/pkg/EV.html>). It has the libev files in
		1937	the F<libev/> subdirectory and includes them in the F<EV/EVAPI.h> (public
		1938	interface) and F<EV.xs> (implementation) files. Only the F<EV.xs> file
		1939	will be compiled. It is pretty complex because it provides its own header
		1940	file.
		1941
		1942	The usage in rxvt-unicode is simpler. It has a F<ev_cpp.h> header file
		1943	that everybody includes and which overrides some autoconf choices:
		1944
		1945	#define EV_USE_POLL 0
		1946	#define EV_MULTIPLICITY 0
		1947	#define EV_PERIODICS 0
		1948	#define EV_CONFIG_H <config.h>
		1949
		1950	#include "ev++.h"
		1951
		1952	And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:
		1953
		1954	#include "ev_cpp.h"
		1955	#include "ev.c"
		1956
		1957
		1958	=head1 COMPLEXITIES
		1959
		1960	In this section the complexities of (many of) the algorithms used inside
		1961	libev will be explained. For complexity discussions about backends see the
		1962	documentation for C<ev_default_init>.
		1963
		1964	=over 4
		1965
		1966	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)
		1967
		1968	=item Changing timer/periodic watchers (by autorepeat, again): O(log skipped_other_timers)
		1969
		1970	=item Starting io/check/prepare/idle/signal/child watchers: O(1)
		1971
		1972	=item Stopping check/prepare/idle watchers: O(1)
		1973
		1974	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % 16))
		1975
		1976	=item Finding the next timer per loop iteration: O(1)
		1977
		1978	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)
		1979
		1980	=item Activating one watcher: O(1)
		1981
		1982	=back
		1983
825		1984
826	=head1 AUTHOR	1985	=head1 AUTHOR
827		1986
828	Marc Lehmann <libev@schmorp.de>.	1987	Marc Lehmann <libev@schmorp.de>.
829		1988

Diff Legend

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