ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/libev/ev.pod

Comparing libev/ev.pod (file contents):
Revision 1.180 by root, Fri Sep 19 03:45:55 2008 UTC vs.
Revision 1.230 by root, Wed Apr 15 18:47:07 2009 UTC

…		…
9	=head2 EXAMPLE PROGRAM	9	=head2 EXAMPLE PROGRAM
10		10
11	// a single header file is required	11	// a single header file is required
12	#include <ev.h>	12	#include <ev.h>
13		13
		14	#include <stdio.h> // for puts
		15
14	// every watcher type has its own typedef'd struct	16	// every watcher type has its own typedef'd struct
15	// with the name ev_<type>	17	// with the name ev_TYPE
16	ev_io stdin_watcher;	18	ev_io stdin_watcher;
17	ev_timer timeout_watcher;	19	ev_timer timeout_watcher;
18		20
19	// all watcher callbacks have a similar signature	21	// all watcher callbacks have a similar signature
20	// this callback is called when data is readable on stdin	22	// this callback is called when data is readable on stdin
21	static void	23	static void
22	stdin_cb (EV_P_ struct ev_io *w, int revents)	24	stdin_cb (EV_P_ ev_io *w, int revents)
23	{	25	{
24	puts ("stdin ready");	26	puts ("stdin ready");
25	// for one-shot events, one must manually stop the watcher	27	// for one-shot events, one must manually stop the watcher
26	// with its corresponding stop function.	28	// with its corresponding stop function.
27	ev_io_stop (EV_A_ w);	29	ev_io_stop (EV_A_ w);
…		…
30	ev_unloop (EV_A_ EVUNLOOP_ALL);	32	ev_unloop (EV_A_ EVUNLOOP_ALL);
31	}	33	}
32		34
33	// another callback, this time for a time-out	35	// another callback, this time for a time-out
34	static void	36	static void
35	timeout_cb (EV_P_ struct ev_timer *w, int revents)	37	timeout_cb (EV_P_ ev_timer *w, int revents)
36	{	38	{
37	puts ("timeout");	39	puts ("timeout");
38	// this causes the innermost ev_loop to stop iterating	40	// this causes the innermost ev_loop to stop iterating
39	ev_unloop (EV_A_ EVUNLOOP_ONE);	41	ev_unloop (EV_A_ EVUNLOOP_ONE);
40	}	42	}
…		…
103	Libev is very configurable. In this manual the default (and most common)	105	Libev is very configurable. In this manual the default (and most common)
104	configuration will be described, which supports multiple event loops. For	106	configuration will be described, which supports multiple event loops. For
105	more info about various configuration options please have a look at	107	more info about various configuration options please have a look at
106	B<EMBED> section in this manual. If libev was configured without support	108	B<EMBED> section in this manual. If libev was configured without support
107	for multiple event loops, then all functions taking an initial argument of	109	for multiple event loops, then all functions taking an initial argument of
108	name C<loop> (which is always of type C<struct ev_loop *>) will not have	110	name C<loop> (which is always of type C<ev_loop *>) will not have
109	this argument.	111	this argument.
110		112
111	=head2 TIME REPRESENTATION	113	=head2 TIME REPRESENTATION
112		114
113	Libev represents time as a single floating point number, representing the	115	Libev represents time as a single floating point number, representing the
…		…
214	C<ev_embeddable_backends () & ev_supported_backends ()>, likewise for	216	C<ev_embeddable_backends () & ev_supported_backends ()>, likewise for
215	recommended ones.	217	recommended ones.
216		218
217	See the description of C<ev_embed> watchers for more info.	219	See the description of C<ev_embed> watchers for more info.
218		220
219	=item ev_set_allocator (void *(*cb)(void *ptr, long size))	221	=item ev_set_allocator (void *(*cb)(void *ptr, long size)) [NOT REENTRANT]
220		222
221	Sets the allocation function to use (the prototype is similar - the	223	Sets the allocation function to use (the prototype is similar - the
222	semantics are identical to the C<realloc> C89/SuS/POSIX function). It is	224	semantics are identical to the C<realloc> C89/SuS/POSIX function). It is
223	used to allocate and free memory (no surprises here). If it returns zero	225	used to allocate and free memory (no surprises here). If it returns zero
224	when memory needs to be allocated (C<size != 0>), the library might abort	226	when memory needs to be allocated (C<size != 0>), the library might abort
…		…
250	}	252	}
251		253
252	...	254	...
253	ev_set_allocator (persistent_realloc);	255	ev_set_allocator (persistent_realloc);
254		256
255	=item ev_set_syserr_cb (void (*cb)(const char *msg));	257	=item ev_set_syserr_cb (void (*cb)(const char *msg)); [NOT REENTRANT]
256		258
257	Set the callback function to call on a retryable system call error (such	259	Set the callback function to call on a retryable system call error (such
258	as failed select, poll, epoll_wait). The message is a printable string	260	as failed select, poll, epoll_wait). The message is a printable string
259	indicating the system call or subsystem causing the problem. If this	261	indicating the system call or subsystem causing the problem. If this
260	callback is set, then libev will expect it to remedy the situation, no	262	callback is set, then libev will expect it to remedy the situation, no
…		…
276		278
277	=back	279	=back
278		280
279	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP	281	=head1 FUNCTIONS CONTROLLING THE EVENT LOOP
280		282
281	An event loop is described by a C<struct ev_loop *>. The library knows two	283	An event loop is described by a C<struct ev_loop *> (the C<struct>
282	types of such loops, the I<default> loop, which supports signals and child	284	is I<not> optional in this case, as there is also an C<ev_loop>
283	events, and dynamically created loops which do not.	285	I<function>).
		286
		287	The library knows two types of such loops, the I<default> loop, which
		288	supports signals and child events, and dynamically created loops which do
		289	not.
284		290
285	=over 4	291	=over 4
286		292
287	=item struct ev_loop *ev_default_loop (unsigned int flags)	293	=item struct ev_loop *ev_default_loop (unsigned int flags)
288		294
…		…
294	If you don't know what event loop to use, use the one returned from this	300	If you don't know what event loop to use, use the one returned from this
295	function.	301	function.
296		302
297	Note that this function is I<not> thread-safe, so if you want to use it	303	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,	304	from multiple threads, you have to lock (note also that this is unlikely,
299	as loops cannot bes hared easily between threads anyway).	305	as loops cannot be shared easily between threads anyway).
300		306
301	The default loop is the only loop that can handle C<ev_signal> and	307	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	308	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	309	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	310	create a dynamic loop with C<ev_loop_new> that doesn't do that, or you
…		…
380	=item C<EVBACKEND_EPOLL> (value 4, Linux)	386	=item C<EVBACKEND_EPOLL> (value 4, Linux)
381		387
382	For few fds, this backend is a bit little slower than poll and select,	388	For few fds, this backend is a bit little slower than poll and select,
383	but it scales phenomenally better. While poll and select usually scale	389	but it scales phenomenally better. While poll and select usually scale
384	like O(total_fds) where n is the total number of fds (or the highest fd),	390	like O(total_fds) where n is the total number of fds (or the highest fd),
385	epoll scales either O(1) or O(active_fds). The epoll design has a number	391	epoll scales either O(1) or O(active_fds).
386	of shortcomings, such as silently dropping events in some hard-to-detect	392
387	cases and requiring a system call per fd change, no fork support and bad	393	The epoll mechanism deserves honorable mention as the most misdesigned
388	support for dup.	394	of the more advanced event mechanisms: mere annoyances include silently
		395	dropping file descriptors, requiring a system call per change per file
		396	descriptor (and unnecessary guessing of parameters), problems with dup and
		397	so on. The biggest issue is fork races, however - if a program forks then
		398	I<both> parent and child process have to recreate the epoll set, which can
		399	take considerable time (one syscall per file descriptor) and is of course
		400	hard to detect.
		401
		402	Epoll is also notoriously buggy - embedding epoll fds I<should> work, but
		403	of course I<doesn't>, and epoll just loves to report events for totally
		404	I<different> file descriptors (even already closed ones, so one cannot
		405	even remove them from the set) than registered in the set (especially
		406	on SMP systems). Libev tries to counter these spurious notifications by
		407	employing an additional generation counter and comparing that against the
		408	events to filter out spurious ones, recreating the set when required.
389		409
390	While stopping, setting and starting an I/O watcher in the same iteration	410	While stopping, setting and starting an I/O watcher in the same iteration
391	will result in some caching, there is still a system call per such incident	411	will result in some caching, there is still a system call per such
392	(because the fd could point to a different file description now), so its	412	incident (because the same I<file descriptor> could point to a different
393	best to avoid that. Also, C<dup ()>'ed file descriptors might not work	413	I<file description> now), so its best to avoid that. Also, C<dup ()>'ed
394	very well if you register events for both fds.	414	file descriptors might not work very well if you register events for both
395		415	file descriptors.
396	Please note that epoll sometimes generates spurious notifications, so you
397	need to use non-blocking I/O or other means to avoid blocking when no data
398	(or space) is available.
399		416
400	Best performance from this backend is achieved by not unregistering all	417	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.	418	watchers for a file descriptor until it has been closed, if possible,
402	keep at least one watcher active per fd at all times.	419	i.e. keep at least one watcher active per fd at all times. Stopping and
		420	starting a watcher (without re-setting it) also usually doesn't cause
		421	extra overhead. A fork can both result in spurious notifications as well
		422	as in libev having to destroy and recreate the epoll object, which can
		423	take considerable time and thus should be avoided.
		424
		425	All this means that, in practice, C<EVBACKEND_SELECT> can be as fast or
		426	faster than epoll for maybe up to a hundred file descriptors, depending on
		427	the usage. So sad.
403		428
404	While nominally embeddable in other event loops, this feature is broken in	429	While nominally embeddable in other event loops, this feature is broken in
405	all kernel versions tested so far.	430	all kernel versions tested so far.
406		431
407	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as	432	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
…		…
410	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)	435	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)
411		436
412	Kqueue deserves special mention, as at the time of this writing, it	437	Kqueue deserves special mention, as at the time of this writing, it
413	was broken on all BSDs except NetBSD (usually it doesn't work reliably	438	was broken on all BSDs except NetBSD (usually it doesn't work reliably
414	with anything but sockets and pipes, except on Darwin, where of course	439	with anything but sockets and pipes, except on Darwin, where of course
415	it's completely useless). For this reason it's not being "auto-detected"	440	it's completely useless). Unlike epoll, however, whose brokenness
		441	is by design, these kqueue bugs can (and eventually will) be fixed
		442	without API changes to existing programs. For this reason it's not being
416	unless you explicitly specify it explicitly in the flags (i.e. using	443	"auto-detected" unless you explicitly specify it in the flags (i.e. using
417	C<EVBACKEND_KQUEUE>) or libev was compiled on a known-to-be-good (-enough)	444	C<EVBACKEND_KQUEUE>) or libev was compiled on a known-to-be-good (-enough)
418	system like NetBSD.	445	system like NetBSD.
419		446
420	You still can embed kqueue into a normal poll or select backend and use it	447	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	448	only for sockets (after having made sure that sockets work with kqueue on
…		…
423		450
424	It scales in the same way as the epoll backend, but the interface to the	451	It scales in the same way as the epoll backend, but the interface to the
425	kernel is more efficient (which says nothing about its actual speed, of	452	kernel is more efficient (which says nothing about its actual speed, of
426	course). While stopping, setting and starting an I/O watcher does never	453	course). While stopping, setting and starting an I/O watcher does never
427	cause an extra system call as with C<EVBACKEND_EPOLL>, it still adds up to	454	cause an extra system call as with C<EVBACKEND_EPOLL>, it still adds up to
428	two event changes per incident, support for C<fork ()> is very bad and it	455	two event changes per incident. Support for C<fork ()> is very bad (but
429	drops fds silently in similarly hard-to-detect cases.	456	sane, unlike epoll) and it drops fds silently in similarly hard-to-detect
		457	cases
430		458
431	This backend usually performs well under most conditions.	459	This backend usually performs well under most conditions.
432		460
433	While nominally embeddable in other event loops, this doesn't work	461	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	462	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	463	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	464	(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	465	(e.g. C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> (but C<poll> is of course
438	sockets.	466	also broken on OS X)) and, did I mention it, using it only for sockets.
439		467
440	This backend maps C<EV_READ> into an C<EVFILT_READ> kevent with	468	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	469	C<NOTE_EOF>, and C<EV_WRITE> into an C<EVFILT_WRITE> kevent with
442	C<NOTE_EOF>.	470	C<NOTE_EOF>.
443		471
…		…
460	While this backend scales well, it requires one system call per active	488	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	489	file descriptor per loop iteration. For small and medium numbers of file
462	descriptors a "slow" C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> backend	490	descriptors a "slow" C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> backend
463	might perform better.	491	might perform better.
464		492
465	On the positive side, ignoring the spurious readiness notifications, this	493	On the positive side, with the exception of the spurious readiness
466	backend actually performed to specification in all tests and is fully	494	notifications, this backend actually performed fully to specification
467	embeddable, which is a rare feat among the OS-specific backends.	495	in all tests and is fully embeddable, which is a rare feat among the
		496	OS-specific backends (I vastly prefer correctness over speed hacks).
468		497
469	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as	498	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
470	C<EVBACKEND_POLL>.	499	C<EVBACKEND_POLL>.
471		500
472	=item C<EVBACKEND_ALL>	501	=item C<EVBACKEND_ALL>
…		…
481		510
482	If one or more of these are or'ed into the flags value, then only these	511	If one or more of these are or'ed into the flags value, then only these
483	backends will be tried (in the reverse order as listed here). If none are	512	backends will be tried (in the reverse order as listed here). If none are
484	specified, all backends in C<ev_recommended_backends ()> will be tried.	513	specified, all backends in C<ev_recommended_backends ()> will be tried.
485		514
486	The most typical usage is like this:	515	Example: This is the most typical usage.
487		516
488	if (!ev_default_loop (0))	517	if (!ev_default_loop (0))
489	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");	518	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
490		519
491	Restrict libev to the select and poll backends, and do not allow	520	Example: Restrict libev to the select and poll backends, and do not allow
492	environment settings to be taken into account:	521	environment settings to be taken into account:
493		522
494	ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);	523	ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);
495		524
496	Use whatever libev has to offer, but make sure that kqueue is used if	525	Example: Use whatever libev has to offer, but make sure that kqueue is
497	available (warning, breaks stuff, best use only with your own private	526	used if available (warning, breaks stuff, best use only with your own
498	event loop and only if you know the OS supports your types of fds):	527	private event loop and only if you know the OS supports your types of
		528	fds):
499		529
500	ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);	530	ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);
501		531
502	=item struct ev_loop *ev_loop_new (unsigned int flags)	532	=item struct ev_loop *ev_loop_new (unsigned int flags)
503		533
…		…
524	responsibility to either stop all watchers cleanly yourself I<before>	554	responsibility to either stop all watchers cleanly yourself I<before>
525	calling this function, or cope with the fact afterwards (which is usually	555	calling this function, or cope with the fact afterwards (which is usually
526	the easiest thing, you can just ignore the watchers and/or C<free ()> them	556	the easiest thing, you can just ignore the watchers and/or C<free ()> them
527	for example).	557	for example).
528		558
529	Note that certain global state, such as signal state, will not be freed by	559	Note that certain global state, such as signal state (and installed signal
530	this function, and related watchers (such as signal and child watchers)	560	handlers), will not be freed by this function, and related watchers (such
531	would need to be stopped manually.	561	as signal and child watchers) would need to be stopped manually.
532		562
533	In general it is not advisable to call this function except in the	563	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	564	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	565	pipe fds. If you need dynamically allocated loops it is better to use
536	C<ev_loop_new> and C<ev_loop_destroy>).	566	C<ev_loop_new> and C<ev_loop_destroy>).
…		…
561		591
562	=item ev_loop_fork (loop)	592	=item ev_loop_fork (loop)
563		593
564	Like C<ev_default_fork>, but acts on an event loop created by	594	Like C<ev_default_fork>, but acts on an event loop created by
565	C<ev_loop_new>. Yes, you have to call this on every allocated event loop	595	C<ev_loop_new>. Yes, you have to call this on every allocated event loop
566	after fork, and how you do this is entirely your own problem.	596	after fork that you want to re-use in the child, and how you do this is
		597	entirely your own problem.
567		598
568	=item int ev_is_default_loop (loop)	599	=item int ev_is_default_loop (loop)
569		600
570	Returns true when the given loop actually is the default loop, false otherwise.	601	Returns true when the given loop is, in fact, the default loop, and false
		602	otherwise.
571		603
572	=item unsigned int ev_loop_count (loop)	604	=item unsigned int ev_loop_count (loop)
573		605
574	Returns the count of loop iterations for the loop, which is identical to	606	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	607	the number of times libev did poll for new events. It starts at C<0> and
…		…
613	If the flags argument is specified as C<0>, it will not return until	645	If the flags argument is specified as C<0>, it will not return until
614	either no event watchers are active anymore or C<ev_unloop> was called.	646	either no event watchers are active anymore or C<ev_unloop> was called.
615		647
616	Please note that an explicit C<ev_unloop> is usually better than	648	Please note that an explicit C<ev_unloop> is usually better than
617	relying on all watchers to be stopped when deciding when a program has	649	relying on all watchers to be stopped when deciding when a program has
618	finished (especially in interactive programs), but having a program that	650	finished (especially in interactive programs), but having a program
619	automatically loops as long as it has to and no longer by virtue of	651	that automatically loops as long as it has to and no longer by virtue
620	relying on its watchers stopping correctly is a thing of beauty.	652	of relying on its watchers stopping correctly, that is truly a thing of
		653	beauty.
621		654
622	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle	655	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle
623	those events and any outstanding ones, but will not block your process in	656	those events and any already outstanding ones, but will not block your
624	case there are no events and will return after one iteration of the loop.	657	process in case there are no events and will return after one iteration of
		658	the loop.
625		659
626	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if	660	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if
627	necessary) and will handle those and any outstanding ones. It will block	661	necessary) and will handle those and any already outstanding ones. It
628	your process until at least one new event arrives, and will return after	662	will block your process until at least one new event arrives (which could
629	one iteration of the loop. This is useful if you are waiting for some	663	be an event internal to libev itself, so there is no guarantee that a
630	external event in conjunction with something not expressible using other	664	user-registered callback will be called), and will return after one
		665	iteration of the loop.
		666
		667	This is useful if you are waiting for some external event in conjunction
		668	with something not expressible using other libev watchers (i.e. "roll your
631	libev watchers. However, a pair of C<ev_prepare>/C<ev_check> watchers is	669	own C<ev_loop>"). However, a pair of C<ev_prepare>/C<ev_check> watchers is
632	usually a better approach for this kind of thing.	670	usually a better approach for this kind of thing.
633		671
634	Here are the gory details of what C<ev_loop> does:	672	Here are the gory details of what C<ev_loop> does:
635		673
636	- Before the first iteration, call any pending watchers.	674	- Before the first iteration, call any pending watchers.
…		…
646	any active watchers at all will result in not sleeping).	684	any active watchers at all will result in not sleeping).
647	- Sleep if the I/O and timer collect interval say so.	685	- Sleep if the I/O and timer collect interval say so.
648	- Block the process, waiting for any events.	686	- Block the process, waiting for any events.
649	- Queue all outstanding I/O (fd) events.	687	- Queue all outstanding I/O (fd) events.
650	- Update the "event loop time" (ev_now ()), and do time jump adjustments.	688	- Update the "event loop time" (ev_now ()), and do time jump adjustments.
651	- Queue all outstanding timers.	689	- Queue all expired timers.
652	- Queue all outstanding periodics.	690	- Queue all expired periodics.
653	- Unless any events are pending now, queue all idle watchers.	691	- Unless any events are pending now, queue all idle watchers.
654	- Queue all check watchers.	692	- Queue all check watchers.
655	- Call all queued watchers in reverse order (i.e. check watchers first).	693	- Call all queued watchers in reverse order (i.e. check watchers first).
656	Signals and child watchers are implemented as I/O watchers, and will	694	Signals and child watchers are implemented as I/O watchers, and will
657	be handled here by queueing them when their watcher gets executed.	695	be handled here by queueing them when their watcher gets executed.
…		…
674	C<EVUNLOOP_ONE>, which will make the innermost C<ev_loop> call return, or	712	C<EVUNLOOP_ONE>, which will make the innermost C<ev_loop> call return, or
675	C<EVUNLOOP_ALL>, which will make all nested C<ev_loop> calls return.	713	C<EVUNLOOP_ALL>, which will make all nested C<ev_loop> calls return.
676		714
677	This "unloop state" will be cleared when entering C<ev_loop> again.	715	This "unloop state" will be cleared when entering C<ev_loop> again.
678		716
		717	It is safe to call C<ev_unloop> from otuside any C<ev_loop> calls.
		718
679	=item ev_ref (loop)	719	=item ev_ref (loop)
680		720
681	=item ev_unref (loop)	721	=item ev_unref (loop)
682		722
683	Ref/unref can be used to add or remove a reference count on the event	723	Ref/unref can be used to add or remove a reference count on the event
684	loop: Every watcher keeps one reference, and as long as the reference	724	loop: Every watcher keeps one reference, and as long as the reference
685	count is nonzero, C<ev_loop> will not return on its own. If you have	725	count is nonzero, C<ev_loop> will not return on its own.
		726
686	a watcher you never unregister that should not keep C<ev_loop> from	727	If you have a watcher you never unregister that should not keep C<ev_loop>
687	returning, ev_unref() after starting, and ev_ref() before stopping it. For	728	from returning, call ev_unref() after starting, and ev_ref() before
		729	stopping it.
		730
688	example, libev itself uses this for its internal signal pipe: It is not	731	As an example, libev itself uses this for its internal signal pipe: It
689	visible to the libev user and should not keep C<ev_loop> from exiting if	732	is not visible to the libev user and should not keep C<ev_loop> from
690	no event watchers registered by it are active. It is also an excellent	733	exiting if no event watchers registered by it are active. It is also an
691	way to do this for generic recurring timers or from within third-party	734	excellent way to do this for generic recurring timers or from within
692	libraries. Just remember to I<unref after start> and I<ref before stop>	735	third-party libraries. Just remember to I<unref after start> and I<ref
693	(but only if the watcher wasn't active before, or was active before,	736	before stop> (but only if the watcher wasn't active before, or was active
694	respectively).	737	before, respectively. Note also that libev might stop watchers itself
		738	(e.g. non-repeating timers) in which case you have to C<ev_ref>
		739	in the callback).
695		740
696	Example: Create a signal watcher, but keep it from keeping C<ev_loop>	741	Example: Create a signal watcher, but keep it from keeping C<ev_loop>
697	running when nothing else is active.	742	running when nothing else is active.
698		743
699	struct ev_signal exitsig;	744	ev_signal exitsig;
700	ev_signal_init (&exitsig, sig_cb, SIGINT);	745	ev_signal_init (&exitsig, sig_cb, SIGINT);
701	ev_signal_start (loop, &exitsig);	746	ev_signal_start (loop, &exitsig);
702	evf_unref (loop);	747	evf_unref (loop);
703		748
704	Example: For some weird reason, unregister the above signal handler again.	749	Example: For some weird reason, unregister the above signal handler again.
…		…
718	Setting these to a higher value (the C<interval> I<must> be >= C<0>)	763	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	764	allows libev to delay invocation of I/O and timer/periodic callbacks
720	to increase efficiency of loop iterations (or to increase power-saving	765	to increase efficiency of loop iterations (or to increase power-saving
721	opportunities).	766	opportunities).
722		767
723	The background is that sometimes your program runs just fast enough to	768	The idea is that sometimes your program runs just fast enough to handle
724	handle one (or very few) event(s) per loop iteration. While this makes	769	one (or very few) event(s) per loop iteration. While this makes the
725	the program responsive, it also wastes a lot of CPU time to poll for new	770	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	771	events, especially with backends like C<select ()> which have a high
727	overhead for the actual polling but can deliver many events at once.	772	overhead for the actual polling but can deliver many events at once.
728		773
729	By setting a higher I<io collect interval> you allow libev to spend more	774	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,	775	time collecting I/O events, so you can handle more events per iteration,
…		…
732	C<ev_timer>) will be not affected. Setting this to a non-null value will	777	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.	778	introduce an additional C<ev_sleep ()> call into most loop iterations.
734		779
735	Likewise, by setting a higher I<timeout collect interval> you allow libev	780	Likewise, by setting a higher I<timeout collect interval> you allow libev
736	to spend more time collecting timeouts, at the expense of increased	781	to spend more time collecting timeouts, at the expense of increased
737	latency (the watcher callback will be called later). C<ev_io> watchers	782	latency/jitter/inexactness (the watcher callback will be called
738	will not be affected. Setting this to a non-null value will not introduce	783	later). C<ev_io> watchers will not be affected. Setting this to a non-null
739	any overhead in libev.	784	value will not introduce any overhead in libev.
740		785
741	Many (busy) programs can usually benefit by setting the I/O collect	786	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	787	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	788	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>,	789	usually doesn't make much sense to set it to a lower value than C<0.01>,
…		…
752	they fire on, say, one-second boundaries only.	797	they fire on, say, one-second boundaries only.
753		798
754	=item ev_loop_verify (loop)	799	=item ev_loop_verify (loop)
755		800
756	This function only does something when C<EV_VERIFY> support has been	801	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	802	compiled in, which is the default for non-minimal builds. It tries to go
758	them for validity. If anything is found to be inconsistent, it will print	803	through all internal structures and checks them for validity. If anything
759	an error message to standard error and call C<abort ()>.	804	is found to be inconsistent, it will print an error message to standard
		805	error and call C<abort ()>.
760		806
761	This can be used to catch bugs inside libev itself: under normal	807	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	808	circumstances, this function will never abort as of course libev keeps its
763	data structures consistent.	809	data structures consistent.
764		810
765	=back	811	=back
766		812
767		813
768	=head1 ANATOMY OF A WATCHER	814	=head1 ANATOMY OF A WATCHER
769		815
		816	In the following description, uppercase C<TYPE> in names stands for the
		817	watcher type, e.g. C<ev_TYPE_start> can mean C<ev_timer_start> for timer
		818	watchers and C<ev_io_start> for I/O watchers.
		819
770	A watcher is a structure that you create and register to record your	820	A watcher is a structure that you create and register to record your
771	interest in some event. For instance, if you want to wait for STDIN to	821	interest in some event. For instance, if you want to wait for STDIN to
772	become readable, you would create an C<ev_io> watcher for that:	822	become readable, you would create an C<ev_io> watcher for that:
773		823
774	static void my_cb (struct ev_loop *loop, struct ev_io *w, int revents)	824	static void my_cb (struct ev_loop *loop, ev_io *w, int revents)
775	{	825	{
776	ev_io_stop (w);	826	ev_io_stop (w);
777	ev_unloop (loop, EVUNLOOP_ALL);	827	ev_unloop (loop, EVUNLOOP_ALL);
778	}	828	}
779		829
780	struct ev_loop *loop = ev_default_loop (0);	830	struct ev_loop *loop = ev_default_loop (0);
		831
781	struct ev_io stdin_watcher;	832	ev_io stdin_watcher;
		833
782	ev_init (&stdin_watcher, my_cb);	834	ev_init (&stdin_watcher, my_cb);
783	ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);	835	ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);
784	ev_io_start (loop, &stdin_watcher);	836	ev_io_start (loop, &stdin_watcher);
		837
785	ev_loop (loop, 0);	838	ev_loop (loop, 0);
786		839
787	As you can see, you are responsible for allocating the memory for your	840	As you can see, you are responsible for allocating the memory for your
788	watcher structures (and it is usually a bad idea to do this on the stack,	841	watcher structures (and it is I<usually> a bad idea to do this on the
789	although this can sometimes be quite valid).	842	stack).
		843
		844	Each watcher has an associated watcher structure (called C<struct ev_TYPE>
		845	or simply C<ev_TYPE>, as typedefs are provided for all watcher structs).
790		846
791	Each watcher structure must be initialised by a call to C<ev_init	847	Each watcher structure must be initialised by a call to C<ev_init
792	(watcher *, callback)>, which expects a callback to be provided. This	848	(watcher *, callback)>, which expects a callback to be provided. This
793	callback gets invoked each time the event occurs (or, in the case of I/O	849	callback gets invoked each time the event occurs (or, in the case of I/O
794	watchers, each time the event loop detects that the file descriptor given	850	watchers, each time the event loop detects that the file descriptor given
795	is readable and/or writable).	851	is readable and/or writable).
796		852
797	Each watcher type has its own C<< ev_<type>_set (watcher *, ...) >> macro	853	Each watcher type further has its own C<< ev_TYPE_set (watcher *, ...) >>
798	with arguments specific to this watcher type. There is also a macro	854	macro to configure it, with arguments specific to the watcher type. There
799	to combine initialisation and setting in one call: C<< ev_<type>_init	855	is also a macro to combine initialisation and setting in one call: C<<
800	(watcher *, callback, ...) >>.	856	ev_TYPE_init (watcher *, callback, ...) >>.
801		857
802	To make the watcher actually watch out for events, you have to start it	858	To make the watcher actually watch out for events, you have to start it
803	with a watcher-specific start function (C<< ev_<type>_start (loop, watcher	859	with a watcher-specific start function (C<< ev_TYPE_start (loop, watcher
804	*) >>), and you can stop watching for events at any time by calling the	860	*) >>), and you can stop watching for events at any time by calling the
805	corresponding stop function (C<< ev_<type>_stop (loop, watcher *) >>.	861	corresponding stop function (C<< ev_TYPE_stop (loop, watcher *) >>.
806		862
807	As long as your watcher is active (has been started but not stopped) you	863	As long as your watcher is active (has been started but not stopped) you
808	must not touch the values stored in it. Most specifically you must never	864	must not touch the values stored in it. Most specifically you must never
809	reinitialise it or call its C<set> macro.	865	reinitialise it or call its C<ev_TYPE_set> macro.
810		866
811	Each and every callback receives the event loop pointer as first, the	867	Each and every callback receives the event loop pointer as first, the
812	registered watcher structure as second, and a bitset of received events as	868	registered watcher structure as second, and a bitset of received events as
813	third argument.	869	third argument.
814		870
…		…
872		928
873	=item C<EV_ASYNC>	929	=item C<EV_ASYNC>
874		930
875	The given async watcher has been asynchronously notified (see C<ev_async>).	931	The given async watcher has been asynchronously notified (see C<ev_async>).
876		932
		933	=item C<EV_CUSTOM>
		934
		935	Not ever sent (or otherwise used) by libev itself, but can be freely used
		936	by libev users to signal watchers (e.g. via C<ev_feed_event>).
		937
877	=item C<EV_ERROR>	938	=item C<EV_ERROR>
878		939
879	An unspecified error has occurred, the watcher has been stopped. This might	940	An unspecified error has occurred, the watcher has been stopped. This might
880	happen because the watcher could not be properly started because libev	941	happen because the watcher could not be properly started because libev
881	ran out of memory, a file descriptor was found to be closed or any other	942	ran out of memory, a file descriptor was found to be closed or any other
		943	problem. Libev considers these application bugs.
		944
882	problem. You best act on it by reporting the problem and somehow coping	945	You best act on it by reporting the problem and somehow coping with the
883	with the watcher being stopped.	946	watcher being stopped. Note that well-written programs should not receive
		947	an error ever, so when your watcher receives it, this usually indicates a
		948	bug in your program.
884		949
885	Libev will usually signal a few "dummy" events together with an error,	950	Libev will usually signal a few "dummy" events together with an error, for
886	for example it might indicate that a fd is readable or writable, and if	951	example it might indicate that a fd is readable or writable, and if your
887	your callbacks is well-written it can just attempt the operation and cope	952	callbacks is well-written it can just attempt the operation and cope with
888	with the error from read() or write(). This will not work in multi-threaded	953	the error from read() or write(). This will not work in multi-threaded
889	programs, though, so beware.	954	programs, though, as the fd could already be closed and reused for another
		955	thing, so beware.
890		956
891	=back	957	=back
892		958
893	=head2 GENERIC WATCHER FUNCTIONS	959	=head2 GENERIC WATCHER FUNCTIONS
894
895	In the following description, C<TYPE> stands for the watcher type,
896	e.g. C<timer> for C<ev_timer> watchers and C<io> for C<ev_io> watchers.
897		960
898	=over 4	961	=over 4
899		962
900	=item C<ev_init> (ev_TYPE *watcher, callback)	963	=item C<ev_init> (ev_TYPE *watcher, callback)
901		964
…		…
907	which rolls both calls into one.	970	which rolls both calls into one.
908		971
909	You can reinitialise a watcher at any time as long as it has been stopped	972	You can reinitialise a watcher at any time as long as it has been stopped
910	(or never started) and there are no pending events outstanding.	973	(or never started) and there are no pending events outstanding.
911		974
912	The callback is always of type C<void (*)(ev_loop *loop, ev_TYPE *watcher,	975	The callback is always of type C<void (*)(struct ev_loop *loop, ev_TYPE *watcher,
913	int revents)>.	976	int revents)>.
		977
		978	Example: Initialise an C<ev_io> watcher in two steps.
		979
		980	ev_io w;
		981	ev_init (&w, my_cb);
		982	ev_io_set (&w, STDIN_FILENO, EV_READ);
914		983
915	=item C<ev_TYPE_set> (ev_TYPE *, [args])	984	=item C<ev_TYPE_set> (ev_TYPE *, [args])
916		985
917	This macro initialises the type-specific parts of a watcher. You need to	986	This macro initialises the type-specific parts of a watcher. You need to
918	call C<ev_init> at least once before you call this macro, but you can	987	call C<ev_init> at least once before you call this macro, but you can
…		…
921	difference to the C<ev_init> macro).	990	difference to the C<ev_init> macro).
922		991
923	Although some watcher types do not have type-specific arguments	992	Although some watcher types do not have type-specific arguments
924	(e.g. C<ev_prepare>) you still need to call its C<set> macro.	993	(e.g. C<ev_prepare>) you still need to call its C<set> macro.
925		994
		995	See C<ev_init>, above, for an example.
		996
926	=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])	997	=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])
927		998
928	This convenience macro rolls both C<ev_init> and C<ev_TYPE_set> macro	999	This convenience macro rolls both C<ev_init> and C<ev_TYPE_set> macro
929	calls into a single call. This is the most convenient method to initialise	1000	calls into a single call. This is the most convenient method to initialise
930	a watcher. The same limitations apply, of course.	1001	a watcher. The same limitations apply, of course.
931		1002
		1003	Example: Initialise and set an C<ev_io> watcher in one step.
		1004
		1005	ev_io_init (&w, my_cb, STDIN_FILENO, EV_READ);
		1006
932	=item C<ev_TYPE_start> (loop *, ev_TYPE *watcher)	1007	=item C<ev_TYPE_start> (loop *, ev_TYPE *watcher)
933		1008
934	Starts (activates) the given watcher. Only active watchers will receive	1009	Starts (activates) the given watcher. Only active watchers will receive
935	events. If the watcher is already active nothing will happen.	1010	events. If the watcher is already active nothing will happen.
936		1011
		1012	Example: Start the C<ev_io> watcher that is being abused as example in this
		1013	whole section.
		1014
		1015	ev_io_start (EV_DEFAULT_UC, &w);
		1016
937	=item C<ev_TYPE_stop> (loop *, ev_TYPE *watcher)	1017	=item C<ev_TYPE_stop> (loop *, ev_TYPE *watcher)
938		1018
939	Stops the given watcher again (if active) and clears the pending	1019	Stops the given watcher if active, and clears the pending status (whether
		1020	the watcher was active or not).
		1021
940	status. It is possible that stopped watchers are pending (for example,	1022	It is possible that stopped watchers are pending - for example,
941	non-repeating timers are being stopped when they become pending), but	1023	non-repeating timers are being stopped when they become pending - but
942	C<ev_TYPE_stop> ensures that the watcher is neither active nor pending. If	1024	calling C<ev_TYPE_stop> ensures that the watcher is neither active nor
943	you want to free or reuse the memory used by the watcher it is therefore a	1025	pending. If you want to free or reuse the memory used by the watcher it is
944	good idea to always call its C<ev_TYPE_stop> function.	1026	therefore a good idea to always call its C<ev_TYPE_stop> function.
945		1027
946	=item bool ev_is_active (ev_TYPE *watcher)	1028	=item bool ev_is_active (ev_TYPE *watcher)
947		1029
948	Returns a true value iff the watcher is active (i.e. it has been started	1030	Returns a true value iff the watcher is active (i.e. it has been started
949	and not yet been stopped). As long as a watcher is active you must not modify	1031	and not yet been stopped). As long as a watcher is active you must not modify
…		…
991	The default priority used by watchers when no priority has been set is	1073	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 :).	1074	always C<0>, which is supposed to not be too high and not be too low :).
993		1075
994	Setting a priority outside the range of C<EV_MINPRI> to C<EV_MAXPRI> is	1076	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	1077	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.	1078	or might not have been clamped to the valid range.
997		1079
998	=item ev_invoke (loop, ev_TYPE *watcher, int revents)	1080	=item ev_invoke (loop, ev_TYPE *watcher, int revents)
999		1081
1000	Invoke the C<watcher> with the given C<loop> and C<revents>. Neither	1082	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	1083	C<loop> nor C<revents> need to be valid as long as the watcher callback
1002	can deal with that fact.	1084	can deal with that fact, as both are simply passed through to the
		1085	callback.
1003		1086
1004	=item int ev_clear_pending (loop, ev_TYPE *watcher)	1087	=item int ev_clear_pending (loop, ev_TYPE *watcher)
1005		1088
1006	If the watcher is pending, this function returns clears its pending status	1089	If the watcher is pending, this function clears its pending status and
1007	and returns its C<revents> bitset (as if its callback was invoked). If the	1090	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>.	1091	watcher isn't pending it does nothing and returns C<0>.
1009		1092
		1093	Sometimes it can be useful to "poll" a watcher instead of waiting for its
		1094	callback to be invoked, which can be accomplished with this function.
		1095
1010	=back	1096	=back
1011		1097
1012		1098
1013	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER	1099	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER
1014		1100
1015	Each watcher has, by default, a member C<void *data> that you can change	1101	Each watcher has, by default, a member C<void *data> that you can change
1016	and read at any time, libev will completely ignore it. This can be used	1102	and read at any time: libev will completely ignore it. This can be used
1017	to associate arbitrary data with your watcher. If you need more data and	1103	to associate arbitrary data with your watcher. If you need more data and
1018	don't want to allocate memory and store a pointer to it in that data	1104	don't want to allocate memory and store a pointer to it in that data
1019	member, you can also "subclass" the watcher type and provide your own	1105	member, you can also "subclass" the watcher type and provide your own
1020	data:	1106	data:
1021		1107
1022	struct my_io	1108	struct my_io
1023	{	1109	{
1024	struct ev_io io;	1110	ev_io io;
1025	int otherfd;	1111	int otherfd;
1026	void *somedata;	1112	void *somedata;
1027	struct whatever *mostinteresting;	1113	struct whatever *mostinteresting;
1028	};	1114	};
1029		1115
…		…
1032	ev_io_init (&w.io, my_cb, fd, EV_READ);	1118	ev_io_init (&w.io, my_cb, fd, EV_READ);
1033		1119
1034	And since your callback will be called with a pointer to the watcher, you	1120	And since your callback will be called with a pointer to the watcher, you
1035	can cast it back to your own type:	1121	can cast it back to your own type:
1036		1122
1037	static void my_cb (struct ev_loop *loop, struct ev_io *w_, int revents)	1123	static void my_cb (struct ev_loop *loop, ev_io *w_, int revents)
1038	{	1124	{
1039	struct my_io *w = (struct my_io *)w_;	1125	struct my_io *w = (struct my_io *)w_;
1040	...	1126	...
1041	}	1127	}
1042		1128
…		…
1053	ev_timer t2;	1139	ev_timer t2;
1054	}	1140	}
1055		1141
1056	In this case getting the pointer to C<my_biggy> is a bit more	1142	In this case getting the pointer to C<my_biggy> is a bit more
1057	complicated: Either you store the address of your C<my_biggy> struct	1143	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	1144	in the C<data> member of the watcher (for woozies), or you need to use
1059	arithmetic using C<offsetof> inside your watchers:	1145	some pointer arithmetic using C<offsetof> inside your watchers (for real
		1146	programmers):
1060		1147
1061	#include <stddef.h>	1148	#include <stddef.h>
1062		1149
1063	static void	1150	static void
1064	t1_cb (EV_P_ struct ev_timer *w, int revents)	1151	t1_cb (EV_P_ ev_timer *w, int revents)
1065	{	1152	{
1066	struct my_biggy big = (struct my_biggy *	1153	struct my_biggy big = (struct my_biggy *
1067	(((char *)w) - offsetof (struct my_biggy, t1));	1154	(((char *)w) - offsetof (struct my_biggy, t1));
1068	}	1155	}
1069		1156
1070	static void	1157	static void
1071	t2_cb (EV_P_ struct ev_timer *w, int revents)	1158	t2_cb (EV_P_ ev_timer *w, int revents)
1072	{	1159	{
1073	struct my_biggy big = (struct my_biggy *	1160	struct my_biggy big = (struct my_biggy *
1074	(((char *)w) - offsetof (struct my_biggy, t2));	1161	(((char *)w) - offsetof (struct my_biggy, t2));
1075	}	1162	}
1076		1163
…		…
1104	In general you can register as many read and/or write event watchers per	1191	In general you can register as many read and/or write event watchers per
1105	fd as you want (as long as you don't confuse yourself). Setting all file	1192	fd as you want (as long as you don't confuse yourself). Setting all file
1106	descriptors to non-blocking mode is also usually a good idea (but not	1193	descriptors to non-blocking mode is also usually a good idea (but not
1107	required if you know what you are doing).	1194	required if you know what you are doing).
1108		1195
1109	If you must do this, then force the use of a known-to-be-good backend	1196	If you cannot use non-blocking mode, then force the use of a
1110	(at the time of this writing, this includes only C<EVBACKEND_SELECT> and	1197	known-to-be-good backend (at the time of this writing, this includes only
1111	C<EVBACKEND_POLL>).	1198	C<EVBACKEND_SELECT> and C<EVBACKEND_POLL>).
1112		1199
1113	Another thing you have to watch out for is that it is quite easy to	1200	Another thing you have to watch out for is that it is quite easy to
1114	receive "spurious" readiness notifications, that is your callback might	1201	receive "spurious" readiness notifications, that is your callback might
1115	be called with C<EV_READ> but a subsequent C<read>(2) will actually block	1202	be called with C<EV_READ> but a subsequent C<read>(2) will actually block
1116	because there is no data. Not only are some backends known to create a	1203	because there is no data. Not only are some backends known to create a
1117	lot of those (for example Solaris ports), it is very easy to get into	1204	lot of those (for example Solaris ports), it is very easy to get into
1118	this situation even with a relatively standard program structure. Thus	1205	this situation even with a relatively standard program structure. Thus
1119	it is best to always use non-blocking I/O: An extra C<read>(2) returning	1206	it is best to always use non-blocking I/O: An extra C<read>(2) returning
1120	C<EAGAIN> is far preferable to a program hanging until some data arrives.	1207	C<EAGAIN> is far preferable to a program hanging until some data arrives.
1121		1208
1122	If you cannot run the fd in non-blocking mode (for example you should not	1209	If you cannot run the fd in non-blocking mode (for example you should
1123	play around with an Xlib connection), then you have to separately re-test	1210	not play around with an Xlib connection), then you have to separately
1124	whether a file descriptor is really ready with a known-to-be good interface	1211	re-test whether a file descriptor is really ready with a known-to-be good
1125	such as poll (fortunately in our Xlib example, Xlib already does this on	1212	interface such as poll (fortunately in our Xlib example, Xlib already
1126	its own, so its quite safe to use).	1213	does this on its own, so its quite safe to use). Some people additionally
		1214	use C<SIGALRM> and an interval timer, just to be sure you won't block
		1215	indefinitely.
		1216
		1217	But really, best use non-blocking mode.
1127		1218
1128	=head3 The special problem of disappearing file descriptors	1219	=head3 The special problem of disappearing file descriptors
1129		1220
1130	Some backends (e.g. kqueue, epoll) need to be told about closing a file	1221	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,	1222	descriptor (either due to calling C<close> explicitly or any other means,
1132	such as C<dup>). The reason is that you register interest in some file	1223	such as C<dup2>). The reason is that you register interest in some file
1133	descriptor, but when it goes away, the operating system will silently drop	1224	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	1225	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	1226	registered with libev, there is no efficient way to see that this is, in
1136	fact, a different file descriptor.	1227	fact, a different file descriptor.
1137		1228
…		…
1168	enable C<EVFLAG_FORKCHECK>, or resort to C<EVBACKEND_SELECT> or	1259	enable C<EVFLAG_FORKCHECK>, or resort to C<EVBACKEND_SELECT> or
1169	C<EVBACKEND_POLL>.	1260	C<EVBACKEND_POLL>.
1170		1261
1171	=head3 The special problem of SIGPIPE	1262	=head3 The special problem of SIGPIPE
1172		1263
1173	While not really specific to libev, it is easy to forget about SIGPIPE:	1264	While not really specific to libev, it is easy to forget about C<SIGPIPE>:
1174	when writing to a pipe whose other end has been closed, your program gets	1265	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	1266	sent a SIGPIPE, which, by default, aborts your program. For most programs
1176	this is sensible behaviour, for daemons, this is usually undesirable.	1267	this is sensible behaviour, for daemons, this is usually undesirable.
1177		1268
1178	So when you encounter spurious, unexplained daemon exits, make sure you	1269	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	1270	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).	1271	somewhere, as that would have given you a big clue).
…		…
1187	=item ev_io_init (ev_io *, callback, int fd, int events)	1278	=item ev_io_init (ev_io *, callback, int fd, int events)
1188		1279
1189	=item ev_io_set (ev_io *, int fd, int events)	1280	=item ev_io_set (ev_io *, int fd, int events)
1190		1281
1191	Configures an C<ev_io> watcher. The C<fd> is the file descriptor to	1282	Configures an C<ev_io> watcher. The C<fd> is the file descriptor to
1192	receive events for and events is either C<EV_READ>, C<EV_WRITE> or	1283	receive events for and C<events> is either C<EV_READ>, C<EV_WRITE> or
1193	C<EV_READ | EV_WRITE> to receive the given events.	1284	C<EV_READ | EV_WRITE>, to express the desire to receive the given events.
1194		1285
1195	=item int fd [read-only]	1286	=item int fd [read-only]
1196		1287
1197	The file descriptor being watched.	1288	The file descriptor being watched.
1198		1289
…		…
1207	Example: Call C<stdin_readable_cb> when STDIN_FILENO has become, well	1298	Example: Call C<stdin_readable_cb> when STDIN_FILENO has become, well
1208	readable, but only once. Since it is likely line-buffered, you could	1299	readable, but only once. Since it is likely line-buffered, you could
1209	attempt to read a whole line in the callback.	1300	attempt to read a whole line in the callback.
1210		1301
1211	static void	1302	static void
1212	stdin_readable_cb (struct ev_loop *loop, struct ev_io *w, int revents)	1303	stdin_readable_cb (struct ev_loop *loop, ev_io *w, int revents)
1213	{	1304	{
1214	ev_io_stop (loop, w);	1305	ev_io_stop (loop, w);
1215	.. read from stdin here (or from w->fd) and haqndle any I/O errors	1306	.. read from stdin here (or from w->fd) and handle any I/O errors
1216	}	1307	}
1217		1308
1218	...	1309	...
1219	struct ev_loop *loop = ev_default_init (0);	1310	struct ev_loop *loop = ev_default_init (0);
1220	struct ev_io stdin_readable;	1311	ev_io stdin_readable;
1221	ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);	1312	ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
1222	ev_io_start (loop, &stdin_readable);	1313	ev_io_start (loop, &stdin_readable);
1223	ev_loop (loop, 0);	1314	ev_loop (loop, 0);
1224		1315
1225		1316
…		…
1228	Timer watchers are simple relative timers that generate an event after a	1319	Timer watchers are simple relative timers that generate an event after a
1229	given time, and optionally repeating in regular intervals after that.	1320	given time, and optionally repeating in regular intervals after that.
1230		1321
1231	The timers are based on real time, that is, if you register an event that	1322	The timers are based on real time, that is, if you register an event that
1232	times out after an hour and you reset your system clock to January last	1323	times out after an hour and you reset your system clock to January last
1233	year, it will still time out after (roughly) and hour. "Roughly" because	1324	year, it will still time out after (roughly) one hour. "Roughly" because
1234	detecting time jumps is hard, and some inaccuracies are unavoidable (the	1325	detecting time jumps is hard, and some inaccuracies are unavoidable (the
1235	monotonic clock option helps a lot here).	1326	monotonic clock option helps a lot here).
1236		1327
1237	The callback is guaranteed to be invoked only after its timeout has passed,	1328	The callback is guaranteed to be invoked only I<after> its timeout has
1238	but if multiple timers become ready during the same loop iteration then	1329	passed. If multiple timers become ready during the same loop iteration
1239	order of execution is undefined.	1330	then the ones with earlier time-out values are invoked before ones with
		1331	later time-out values (but this is no longer true when a callback calls
		1332	C<ev_loop> recursively).
		1333
		1334	=head3 Be smart about timeouts
		1335
		1336	Many real-world problems involve some kind of timeout, usually for error
		1337	recovery. A typical example is an HTTP request - if the other side hangs,
		1338	you want to raise some error after a while.
		1339
		1340	What follows are some ways to handle this problem, from obvious and
		1341	inefficient to smart and efficient.
		1342
		1343	In the following, a 60 second activity timeout is assumed - a timeout that
		1344	gets reset to 60 seconds each time there is activity (e.g. each time some
		1345	data or other life sign was received).
		1346
		1347	=over 4
		1348
		1349	=item 1. Use a timer and stop, reinitialise and start it on activity.
		1350
		1351	This is the most obvious, but not the most simple way: In the beginning,
		1352	start the watcher:
		1353
		1354	ev_timer_init (timer, callback, 60., 0.);
		1355	ev_timer_start (loop, timer);
		1356
		1357	Then, each time there is some activity, C<ev_timer_stop> it, initialise it
		1358	and start it again:
		1359
		1360	ev_timer_stop (loop, timer);
		1361	ev_timer_set (timer, 60., 0.);
		1362	ev_timer_start (loop, timer);
		1363
		1364	This is relatively simple to implement, but means that each time there is
		1365	some activity, libev will first have to remove the timer from its internal
		1366	data structure and then add it again. Libev tries to be fast, but it's
		1367	still not a constant-time operation.
		1368
		1369	=item 2. Use a timer and re-start it with C<ev_timer_again> inactivity.
		1370
		1371	This is the easiest way, and involves using C<ev_timer_again> instead of
		1372	C<ev_timer_start>.
		1373
		1374	To implement this, configure an C<ev_timer> with a C<repeat> value
		1375	of C<60> and then call C<ev_timer_again> at start and each time you
		1376	successfully read or write some data. If you go into an idle state where
		1377	you do not expect data to travel on the socket, you can C<ev_timer_stop>
		1378	the timer, and C<ev_timer_again> will automatically restart it if need be.
		1379
		1380	That means you can ignore both the C<ev_timer_start> function and the
		1381	C<after> argument to C<ev_timer_set>, and only ever use the C<repeat>
		1382	member and C<ev_timer_again>.
		1383
		1384	At start:
		1385
		1386	ev_timer_init (timer, callback);
		1387	timer->repeat = 60.;
		1388	ev_timer_again (loop, timer);
		1389
		1390	Each time there is some activity:
		1391
		1392	ev_timer_again (loop, timer);
		1393
		1394	It is even possible to change the time-out on the fly, regardless of
		1395	whether the watcher is active or not:
		1396
		1397	timer->repeat = 30.;
		1398	ev_timer_again (loop, timer);
		1399
		1400	This is slightly more efficient then stopping/starting the timer each time
		1401	you want to modify its timeout value, as libev does not have to completely
		1402	remove and re-insert the timer from/into its internal data structure.
		1403
		1404	It is, however, even simpler than the "obvious" way to do it.
		1405
		1406	=item 3. Let the timer time out, but then re-arm it as required.
		1407
		1408	This method is more tricky, but usually most efficient: Most timeouts are
		1409	relatively long compared to the intervals between other activity - in
		1410	our example, within 60 seconds, there are usually many I/O events with
		1411	associated activity resets.
		1412
		1413	In this case, it would be more efficient to leave the C<ev_timer> alone,
		1414	but remember the time of last activity, and check for a real timeout only
		1415	within the callback:
		1416
		1417	ev_tstamp last_activity; // time of last activity
		1418
		1419	static void
		1420	callback (EV_P_ ev_timer *w, int revents)
		1421	{
		1422	ev_tstamp now = ev_now (EV_A);
		1423	ev_tstamp timeout = last_activity + 60.;
		1424
		1425	// if last_activity + 60. is older than now, we did time out
		1426	if (timeout < now)
		1427	{
		1428	// timeout occured, take action
		1429	}
		1430	else
		1431	{
		1432	// callback was invoked, but there was some activity, re-arm
		1433	// the watcher to fire in last_activity + 60, which is
		1434	// guaranteed to be in the future, so "again" is positive:
		1435	w->repeat = timeout - now;
		1436	ev_timer_again (EV_A_ w);
		1437	}
		1438	}
		1439
		1440	To summarise the callback: first calculate the real timeout (defined
		1441	as "60 seconds after the last activity"), then check if that time has
		1442	been reached, which means something I<did>, in fact, time out. Otherwise
		1443	the callback was invoked too early (C<timeout> is in the future), so
		1444	re-schedule the timer to fire at that future time, to see if maybe we have
		1445	a timeout then.
		1446
		1447	Note how C<ev_timer_again> is used, taking advantage of the
		1448	C<ev_timer_again> optimisation when the timer is already running.
		1449
		1450	This scheme causes more callback invocations (about one every 60 seconds
		1451	minus half the average time between activity), but virtually no calls to
		1452	libev to change the timeout.
		1453
		1454	To start the timer, simply initialise the watcher and set C<last_activity>
		1455	to the current time (meaning we just have some activity :), then call the
		1456	callback, which will "do the right thing" and start the timer:
		1457
		1458	ev_timer_init (timer, callback);
		1459	last_activity = ev_now (loop);
		1460	callback (loop, timer, EV_TIMEOUT);
		1461
		1462	And when there is some activity, simply store the current time in
		1463	C<last_activity>, no libev calls at all:
		1464
		1465	last_actiivty = ev_now (loop);
		1466
		1467	This technique is slightly more complex, but in most cases where the
		1468	time-out is unlikely to be triggered, much more efficient.
		1469
		1470	Changing the timeout is trivial as well (if it isn't hard-coded in the
		1471	callback :) - just change the timeout and invoke the callback, which will
		1472	fix things for you.
		1473
		1474	=item 4. Wee, just use a double-linked list for your timeouts.
		1475
		1476	If there is not one request, but many thousands (millions...), all
		1477	employing some kind of timeout with the same timeout value, then one can
		1478	do even better:
		1479
		1480	When starting the timeout, calculate the timeout value and put the timeout
		1481	at the I<end> of the list.
		1482
		1483	Then use an C<ev_timer> to fire when the timeout at the I<beginning> of
		1484	the list is expected to fire (for example, using the technique #3).
		1485
		1486	When there is some activity, remove the timer from the list, recalculate
		1487	the timeout, append it to the end of the list again, and make sure to
		1488	update the C<ev_timer> if it was taken from the beginning of the list.
		1489
		1490	This way, one can manage an unlimited number of timeouts in O(1) time for
		1491	starting, stopping and updating the timers, at the expense of a major
		1492	complication, and having to use a constant timeout. The constant timeout
		1493	ensures that the list stays sorted.
		1494
		1495	=back
		1496
		1497	So which method the best?
		1498
		1499	Method #2 is a simple no-brain-required solution that is adequate in most
		1500	situations. Method #3 requires a bit more thinking, but handles many cases
		1501	better, and isn't very complicated either. In most case, choosing either
		1502	one is fine, with #3 being better in typical situations.
		1503
		1504	Method #1 is almost always a bad idea, and buys you nothing. Method #4 is
		1505	rather complicated, but extremely efficient, something that really pays
		1506	off after the first million or so of active timers, i.e. it's usually
		1507	overkill :)
1240		1508
1241	=head3 The special problem of time updates	1509	=head3 The special problem of time updates
1242		1510
1243	Establishing the current time is a costly operation (it usually takes at	1511	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	1512	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	1513	time only before and after C<ev_loop> collects new events, which causes a
1246	a growing difference between C<ev_now ()> and C<ev_time ()> when handling	1514	growing difference between C<ev_now ()> and C<ev_time ()> when handling
1247	lots of events.	1515	lots of events in one iteration.
1248		1516
1249	The relative timeouts are calculated relative to the C<ev_now ()>	1517	The relative timeouts are calculated relative to the C<ev_now ()>
1250	time. This is usually the right thing as this timestamp refers to the time	1518	time. This is usually the right thing as this timestamp refers to the time
1251	of the event triggering whatever timeout you are modifying/starting. If	1519	of the event triggering whatever timeout you are modifying/starting. If
1252	you suspect event processing to be delayed and you I<need> to base the	1520	you suspect event processing to be delayed and you I<need> to base the
…		…
1288	If the timer is started but non-repeating, stop it (as if it timed out).	1556	If the timer is started but non-repeating, stop it (as if it timed out).
1289		1557
1290	If the timer is repeating, either start it if necessary (with the	1558	If the timer is repeating, either start it if necessary (with the
1291	C<repeat> value), or reset the running timer to the C<repeat> value.	1559	C<repeat> value), or reset the running timer to the C<repeat> value.
1292		1560
1293	This sounds a bit complicated, but here is a useful and typical	1561	This sounds a bit complicated, see "Be smart about timeouts", above, for a
1294	example: Imagine you have a TCP connection and you want a so-called idle	1562	usage example.
1295	timeout, that is, you want to be called when there have been, say, 60
1296	seconds of inactivity on the socket. The easiest way to do this is to
1297	configure an C<ev_timer> with a C<repeat> value of C<60> and then call
1298	C<ev_timer_again> each time you successfully read or write some data. If
1299	you go into an idle state where you do not expect data to travel on the
1300	socket, you can C<ev_timer_stop> the timer, and C<ev_timer_again> will
1301	automatically restart it if need be.
1302
1303	That means you can ignore the C<after> value and C<ev_timer_start>
1304	altogether and only ever use the C<repeat> value and C<ev_timer_again>:
1305
1306	ev_timer_init (timer, callback, 0., 5.);
1307	ev_timer_again (loop, timer);
1308	...
1309	timer->again = 17.;
1310	ev_timer_again (loop, timer);
1311	...
1312	timer->again = 10.;
1313	ev_timer_again (loop, timer);
1314
1315	This is more slightly efficient then stopping/starting the timer each time
1316	you want to modify its timeout value.
1317		1563
1318	=item ev_tstamp repeat [read-write]	1564	=item ev_tstamp repeat [read-write]
1319		1565
1320	The current C<repeat> value. Will be used each time the watcher times out	1566	The current C<repeat> value. Will be used each time the watcher times out
1321	or C<ev_timer_again> is called and determines the next timeout (if any),	1567	or C<ev_timer_again> is called, and determines the next timeout (if any),
1322	which is also when any modifications are taken into account.	1568	which is also when any modifications are taken into account.
1323		1569
1324	=back	1570	=back
1325		1571
1326	=head3 Examples	1572	=head3 Examples
1327		1573
1328	Example: Create a timer that fires after 60 seconds.	1574	Example: Create a timer that fires after 60 seconds.
1329		1575
1330	static void	1576	static void
1331	one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)	1577	one_minute_cb (struct ev_loop *loop, ev_timer *w, int revents)
1332	{	1578	{
1333	.. one minute over, w is actually stopped right here	1579	.. one minute over, w is actually stopped right here
1334	}	1580	}
1335		1581
1336	struct ev_timer mytimer;	1582	ev_timer mytimer;
1337	ev_timer_init (&mytimer, one_minute_cb, 60., 0.);	1583	ev_timer_init (&mytimer, one_minute_cb, 60., 0.);
1338	ev_timer_start (loop, &mytimer);	1584	ev_timer_start (loop, &mytimer);
1339		1585
1340	Example: Create a timeout timer that times out after 10 seconds of	1586	Example: Create a timeout timer that times out after 10 seconds of
1341	inactivity.	1587	inactivity.
1342		1588
1343	static void	1589	static void
1344	timeout_cb (struct ev_loop *loop, struct ev_timer *w, int revents)	1590	timeout_cb (struct ev_loop *loop, ev_timer *w, int revents)
1345	{	1591	{
1346	.. ten seconds without any activity	1592	.. ten seconds without any activity
1347	}	1593	}
1348		1594
1349	struct ev_timer mytimer;	1595	ev_timer mytimer;
1350	ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */	1596	ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */
1351	ev_timer_again (&mytimer); /* start timer */	1597	ev_timer_again (&mytimer); /* start timer */
1352	ev_loop (loop, 0);	1598	ev_loop (loop, 0);
1353		1599
1354	// and in some piece of code that gets executed on any "activity":	1600	// and in some piece of code that gets executed on any "activity":
…		…
1359	=head2 C<ev_periodic> - to cron or not to cron?	1605	=head2 C<ev_periodic> - to cron or not to cron?
1360		1606
1361	Periodic watchers are also timers of a kind, but they are very versatile	1607	Periodic watchers are also timers of a kind, but they are very versatile
1362	(and unfortunately a bit complex).	1608	(and unfortunately a bit complex).
1363		1609
1364	Unlike C<ev_timer>'s, they are not based on real time (or relative time)	1610	Unlike C<ev_timer>, periodic watchers are not based on real time (or
1365	but on wall clock time (absolute time). You can tell a periodic watcher	1611	relative time, the physical time that passes) but on wall clock time
1366	to trigger after some specific point in time. For example, if you tell a	1612	(absolute time, the thing you can read on your calender or clock). The
1367	periodic watcher to trigger in 10 seconds (by specifying e.g. C<ev_now ()	1613	difference is that wall clock time can run faster or slower than real
1368	+ 10.>, that is, an absolute time not a delay) and then reset your system	1614	time, and time jumps are not uncommon (e.g. when you adjust your
1369	clock to January of the previous year, then it will take more than year	1615	wrist-watch).
1370	to trigger the event (unlike an C<ev_timer>, which would still trigger
1371	roughly 10 seconds later as it uses a relative timeout).
1372		1616
		1617	You can tell a periodic watcher to trigger after some specific point
		1618	in time: for example, if you tell a periodic watcher to trigger "in 10
		1619	seconds" (by specifying e.g. C<ev_now () + 10.>, that is, an absolute time
		1620	not a delay) and then reset your system clock to January of the previous
		1621	year, then it will take a year or more to trigger the event (unlike an
		1622	C<ev_timer>, which would still trigger roughly 10 seconds after starting
		1623	it, as it uses a relative timeout).
		1624
1373	C<ev_periodic>s can also be used to implement vastly more complex timers,	1625	C<ev_periodic> watchers can also be used to implement vastly more complex
1374	such as triggering an event on each "midnight, local time", or other	1626	timers, such as triggering an event on each "midnight, local time", or
1375	complicated, rules.	1627	other complicated rules. This cannot be done with C<ev_timer> watchers, as
		1628	those cannot react to time jumps.
1376		1629
1377	As with timers, the callback is guaranteed to be invoked only when the	1630	As with timers, the callback is guaranteed to be invoked only when the
1378	time (C<at>) has passed, but if multiple periodic timers become ready	1631	point in time where it is supposed to trigger has passed. If multiple
1379	during the same loop iteration then order of execution is undefined.	1632	timers become ready during the same loop iteration then the ones with
		1633	earlier time-out values are invoked before ones with later time-out values
		1634	(but this is no longer true when a callback calls C<ev_loop> recursively).
1380		1635
1381	=head3 Watcher-Specific Functions and Data Members	1636	=head3 Watcher-Specific Functions and Data Members
1382		1637
1383	=over 4	1638	=over 4
1384		1639
1385	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)	1640	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp offset, ev_tstamp interval, reschedule_cb)
1386		1641
1387	=item ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)	1642	=item ev_periodic_set (ev_periodic *, ev_tstamp offset, ev_tstamp interval, reschedule_cb)
1388		1643
1389	Lots of arguments, lets sort it out... There are basically three modes of	1644	Lots of arguments, let's sort it out... There are basically three modes of
1390	operation, and we will explain them from simplest to complex:	1645	operation, and we will explain them from simplest to most complex:
1391		1646
1392	=over 4	1647	=over 4
1393		1648
1394	=item * absolute timer (at = time, interval = reschedule_cb = 0)	1649	=item * absolute timer (offset = absolute time, interval = 0, reschedule_cb = 0)
1395		1650
1396	In this configuration the watcher triggers an event after the wall clock	1651	In this configuration the watcher triggers an event after the wall clock
1397	time C<at> has passed and doesn't repeat. It will not adjust when a time	1652	time C<offset> has passed. It will not repeat and will not adjust when a
1398	jump occurs, that is, if it is to be run at January 1st 2011 then it will	1653	time jump occurs, that is, if it is to be run at January 1st 2011 then it
1399	run when the system time reaches or surpasses this time.	1654	will be stopped and invoked when the system clock reaches or surpasses
		1655	this point in time.
1400		1656
1401	=item * repeating interval timer (at = offset, interval > 0, reschedule_cb = 0)	1657	=item * repeating interval timer (offset = offset within interval, interval > 0, reschedule_cb = 0)
1402		1658
1403	In this mode the watcher will always be scheduled to time out at the next	1659	In this mode the watcher will always be scheduled to time out at the next
1404	C<at + N * interval> time (for some integer N, which can also be negative)	1660	C<offset + N * interval> time (for some integer N, which can also be
1405	and then repeat, regardless of any time jumps.	1661	negative) and then repeat, regardless of any time jumps. The C<offset>
		1662	argument is merely an offset into the C<interval> periods.
1406		1663
1407	This can be used to create timers that do not drift with respect to system	1664	This can be used to create timers that do not drift with respect to the
1408	time, for example, here is a C<ev_periodic> that triggers each hour, on	1665	system clock, for example, here is an C<ev_periodic> that triggers each
1409	the hour:	1666	hour, on the hour (with respect to UTC):
1410		1667
1411	ev_periodic_set (&periodic, 0., 3600., 0);	1668	ev_periodic_set (&periodic, 0., 3600., 0);
1412		1669
1413	This doesn't mean there will always be 3600 seconds in between triggers,	1670	This doesn't mean there will always be 3600 seconds in between triggers,
1414	but only that the callback will be called when the system time shows a	1671	but only that the callback will be called when the system time shows a
1415	full hour (UTC), or more correctly, when the system time is evenly divisible	1672	full hour (UTC), or more correctly, when the system time is evenly divisible
1416	by 3600.	1673	by 3600.
1417		1674
1418	Another way to think about it (for the mathematically inclined) is that	1675	Another way to think about it (for the mathematically inclined) is that
1419	C<ev_periodic> will try to run the callback in this mode at the next possible	1676	C<ev_periodic> will try to run the callback in this mode at the next possible
1420	time where C<time = at (mod interval)>, regardless of any time jumps.	1677	time where C<time = offset (mod interval)>, regardless of any time jumps.
1421		1678
1422	For numerical stability it is preferable that the C<at> value is near	1679	For numerical stability it is preferable that the C<offset> value is near
1423	C<ev_now ()> (the current time), but there is no range requirement for	1680	C<ev_now ()> (the current time), but there is no range requirement for
1424	this value, and in fact is often specified as zero.	1681	this value, and in fact is often specified as zero.
1425		1682
1426	Note also that there is an upper limit to how often a timer can fire (CPU	1683	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	1684	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	1685	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).	1686	millisecond (if the OS supports it and the machine is fast enough).
1430		1687
1431	=item * manual reschedule mode (at and interval ignored, reschedule_cb = callback)	1688	=item * manual reschedule mode (offset ignored, interval ignored, reschedule_cb = callback)
1432		1689
1433	In this mode the values for C<interval> and C<at> are both being	1690	In this mode the values for C<interval> and C<offset> are both being
1434	ignored. Instead, each time the periodic watcher gets scheduled, the	1691	ignored. Instead, each time the periodic watcher gets scheduled, the
1435	reschedule callback will be called with the watcher as first, and the	1692	reschedule callback will be called with the watcher as first, and the
1436	current time as second argument.	1693	current time as second argument.
1437		1694
1438	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher,	1695	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher, ever,
1439	ever, or make ANY event loop modifications whatsoever>.	1696	or make ANY other event loop modifications whatsoever, unless explicitly
		1697	allowed by documentation here>.
1440		1698
1441	If you need to stop it, return C<now + 1e30> (or so, fudge fudge) and stop	1699	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	1700	it afterwards (e.g. by starting an C<ev_prepare> watcher, which is the
1443	only event loop modification you are allowed to do).	1701	only event loop modification you are allowed to do).
1444		1702
1445	The callback prototype is C<ev_tstamp (*reschedule_cb)(struct ev_periodic	1703	The callback prototype is C<ev_tstamp (*reschedule_cb)(ev_periodic
1446	*w, ev_tstamp now)>, e.g.:	1704	*w, ev_tstamp now)>, e.g.:
1447		1705
		1706	static ev_tstamp
1448	static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)	1707	my_rescheduler (ev_periodic *w, ev_tstamp now)
1449	{	1708	{
1450	return now + 60.;	1709	return now + 60.;
1451	}	1710	}
1452		1711
1453	It must return the next time to trigger, based on the passed time value	1712	It must return the next time to trigger, based on the passed time value
…		…
1473	a different time than the last time it was called (e.g. in a crond like	1732	a different time than the last time it was called (e.g. in a crond like
1474	program when the crontabs have changed).	1733	program when the crontabs have changed).
1475		1734
1476	=item ev_tstamp ev_periodic_at (ev_periodic *)	1735	=item ev_tstamp ev_periodic_at (ev_periodic *)
1477		1736
1478	When active, returns the absolute time that the watcher is supposed to	1737	When active, returns the absolute time that the watcher is supposed
1479	trigger next.	1738	to trigger next. This is not the same as the C<offset> argument to
		1739	C<ev_periodic_set>, but indeed works even in interval and manual
		1740	rescheduling modes.
1480		1741
1481	=item ev_tstamp offset [read-write]	1742	=item ev_tstamp offset [read-write]
1482		1743
1483	When repeating, this contains the offset value, otherwise this is the	1744	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>).	1745	absolute point in time (the C<offset> value passed to C<ev_periodic_set>,
		1746	although libev might modify this value for better numerical stability).
1485		1747
1486	Can be modified any time, but changes only take effect when the periodic	1748	Can be modified any time, but changes only take effect when the periodic
1487	timer fires or C<ev_periodic_again> is being called.	1749	timer fires or C<ev_periodic_again> is being called.
1488		1750
1489	=item ev_tstamp interval [read-write]	1751	=item ev_tstamp interval [read-write]
1490		1752
1491	The current interval value. Can be modified any time, but changes only	1753	The current interval value. Can be modified any time, but changes only
1492	take effect when the periodic timer fires or C<ev_periodic_again> is being	1754	take effect when the periodic timer fires or C<ev_periodic_again> is being
1493	called.	1755	called.
1494		1756
1495	=item ev_tstamp (*reschedule_cb)(struct ev_periodic *w, ev_tstamp now) [read-write]	1757	=item ev_tstamp (*reschedule_cb)(ev_periodic *w, ev_tstamp now) [read-write]
1496		1758
1497	The current reschedule callback, or C<0>, if this functionality is	1759	The current reschedule callback, or C<0>, if this functionality is
1498	switched off. Can be changed any time, but changes only take effect when	1760	switched off. Can be changed any time, but changes only take effect when
1499	the periodic timer fires or C<ev_periodic_again> is being called.	1761	the periodic timer fires or C<ev_periodic_again> is being called.
1500		1762
1501	=back	1763	=back
1502		1764
1503	=head3 Examples	1765	=head3 Examples
1504		1766
1505	Example: Call a callback every hour, or, more precisely, whenever the	1767	Example: Call a callback every hour, or, more precisely, whenever the
1506	system clock is divisible by 3600. The callback invocation times have	1768	system time is divisible by 3600. The callback invocation times have
1507	potentially a lot of jitter, but good long-term stability.	1769	potentially a lot of jitter, but good long-term stability.
1508		1770
1509	static void	1771	static void
1510	clock_cb (struct ev_loop *loop, struct ev_io *w, int revents)	1772	clock_cb (struct ev_loop *loop, ev_io *w, int revents)
1511	{	1773	{
1512	... its now a full hour (UTC, or TAI or whatever your clock follows)	1774	... its now a full hour (UTC, or TAI or whatever your clock follows)
1513	}	1775	}
1514		1776
1515	struct ev_periodic hourly_tick;	1777	ev_periodic hourly_tick;
1516	ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);	1778	ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);
1517	ev_periodic_start (loop, &hourly_tick);	1779	ev_periodic_start (loop, &hourly_tick);
1518		1780
1519	Example: The same as above, but use a reschedule callback to do it:	1781	Example: The same as above, but use a reschedule callback to do it:
1520		1782
1521	#include <math.h>	1783	#include <math.h>
1522		1784
1523	static ev_tstamp	1785	static ev_tstamp
1524	my_scheduler_cb (struct ev_periodic *w, ev_tstamp now)	1786	my_scheduler_cb (ev_periodic *w, ev_tstamp now)
1525	{	1787	{
1526	return fmod (now, 3600.) + 3600.;	1788	return now + (3600. - fmod (now, 3600.));
1527	}	1789	}
1528		1790
1529	ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);	1791	ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
1530		1792
1531	Example: Call a callback every hour, starting now:	1793	Example: Call a callback every hour, starting now:
1532		1794
1533	struct ev_periodic hourly_tick;	1795	ev_periodic hourly_tick;
1534	ev_periodic_init (&hourly_tick, clock_cb,	1796	ev_periodic_init (&hourly_tick, clock_cb,
1535	fmod (ev_now (loop), 3600.), 3600., 0);	1797	fmod (ev_now (loop), 3600.), 3600., 0);
1536	ev_periodic_start (loop, &hourly_tick);	1798	ev_periodic_start (loop, &hourly_tick);
1537		1799
1538		1800
…		…
1541	Signal watchers will trigger an event when the process receives a specific	1803	Signal watchers will trigger an event when the process receives a specific
1542	signal one or more times. Even though signals are very asynchronous, libev	1804	signal one or more times. Even though signals are very asynchronous, libev
1543	will try it's best to deliver signals synchronously, i.e. as part of the	1805	will try it's best to deliver signals synchronously, i.e. as part of the
1544	normal event processing, like any other event.	1806	normal event processing, like any other event.
1545		1807
		1808	If you want signals asynchronously, just use C<sigaction> as you would
		1809	do without libev and forget about sharing the signal. You can even use
		1810	C<ev_async> from a signal handler to synchronously wake up an event loop.
		1811
1546	You can configure as many watchers as you like per signal. Only when the	1812	You can configure as many watchers as you like per signal. Only when the
1547	first watcher gets started will libev actually register a signal watcher	1813	first watcher gets started will libev actually register a signal handler
1548	with the kernel (thus it coexists with your own signal handlers as long	1814	with the kernel (thus it coexists with your own signal handlers as long as
1549	as you don't register any with libev). Similarly, when the last signal	1815	you don't register any with libev for the same signal). Similarly, when
1550	watcher for a signal is stopped libev will reset the signal handler to	1816	the last signal watcher for a signal is stopped, libev will reset the
1551	SIG_DFL (regardless of what it was set to before).	1817	signal handler to SIG_DFL (regardless of what it was set to before).
1552		1818
1553	If possible and supported, libev will install its handlers with	1819	If possible and supported, libev will install its handlers with
1554	C<SA_RESTART> behaviour enabled, so system calls should not be unduly	1820	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	1821	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	1822	signals you can block all signals in an C<ev_check> watcher and unblock
…		…
1573		1839
1574	=back	1840	=back
1575		1841
1576	=head3 Examples	1842	=head3 Examples
1577		1843
1578	Example: Try to exit cleanly on SIGINT and SIGTERM.	1844	Example: Try to exit cleanly on SIGINT.
1579		1845
1580	static void	1846	static void
1581	sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)	1847	sigint_cb (struct ev_loop *loop, ev_signal *w, int revents)
1582	{	1848	{
1583	ev_unloop (loop, EVUNLOOP_ALL);	1849	ev_unloop (loop, EVUNLOOP_ALL);
1584	}	1850	}
1585		1851
1586	struct ev_signal signal_watcher;	1852	ev_signal signal_watcher;
1587	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);	1853	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
1588	ev_signal_start (loop, &sigint_cb);	1854	ev_signal_start (loop, &signal_watcher);
1589		1855
1590		1856
1591	=head2 C<ev_child> - watch out for process status changes	1857	=head2 C<ev_child> - watch out for process status changes
1592		1858
1593	Child watchers trigger when your process receives a SIGCHLD in response to	1859	Child watchers trigger when your process receives a SIGCHLD in response to
1594	some child status changes (most typically when a child of yours dies). It	1860	some child status changes (most typically when a child of yours dies or
1595	is permissible to install a child watcher I<after> the child has been	1861	exits). It is permissible to install a child watcher I<after> the child
1596	forked (which implies it might have already exited), as long as the event	1862	has been forked (which implies it might have already exited), as long
1597	loop isn't entered (or is continued from a watcher).	1863	as the event loop isn't entered (or is continued from a watcher), i.e.,
		1864	forking and then immediately registering a watcher for the child is fine,
		1865	but forking and registering a watcher a few event loop iterations later is
		1866	not.
1598		1867
1599	Only the default event loop is capable of handling signals, and therefore	1868	Only the default event loop is capable of handling signals, and therefore
1600	you can only register child watchers in the default event loop.	1869	you can only register child watchers in the default event loop.
1601		1870
1602	=head3 Process Interaction	1871	=head3 Process Interaction
…		…
1663	its completion.	1932	its completion.
1664		1933
1665	ev_child cw;	1934	ev_child cw;
1666		1935
1667	static void	1936	static void
1668	child_cb (EV_P_ struct ev_child *w, int revents)	1937	child_cb (EV_P_ ev_child *w, int revents)
1669	{	1938	{
1670	ev_child_stop (EV_A_ w);	1939	ev_child_stop (EV_A_ w);
1671	printf ("process %d exited with status %x\n", w->rpid, w->rstatus);	1940	printf ("process %d exited with status %x\n", w->rpid, w->rstatus);
1672	}	1941	}
1673		1942
…		…
1688		1957
1689		1958
1690	=head2 C<ev_stat> - did the file attributes just change?	1959	=head2 C<ev_stat> - did the file attributes just change?
1691		1960
1692	This watches a file system path for attribute changes. That is, it calls	1961	This watches a file system path for attribute changes. That is, it calls
1693	C<stat> regularly (or when the OS says it changed) and sees if it changed	1962	C<stat> on that path in regular intervals (or when the OS says it changed)
1694	compared to the last time, invoking the callback if it did.	1963	and sees if it changed compared to the last time, invoking the callback if
		1964	it did.
1695		1965
1696	The path does not need to exist: changing from "path exists" to "path does	1966	The path does not need to exist: changing from "path exists" to "path does
1697	not exist" is a status change like any other. The condition "path does	1967	not exist" is a status change like any other. The condition "path does not
1698	not exist" is signified by the C<st_nlink> field being zero (which is	1968	exist" (or more correctly "path cannot be stat'ed") is signified by the
1699	otherwise always forced to be at least one) and all the other fields of	1969	C<st_nlink> field being zero (which is otherwise always forced to be at
1700	the stat buffer having unspecified contents.	1970	least one) and all the other fields of the stat buffer having unspecified
		1971	contents.
1701		1972
1702	The path I<should> be absolute and I<must not> end in a slash. If it is	1973	The path I<must not> end in a slash or contain special components such as
		1974	C<.> or C<..>. The path I<should> be absolute: If it is relative and
1703	relative and your working directory changes, the behaviour is undefined.	1975	your working directory changes, then the behaviour is undefined.
1704		1976
1705	Since there is no standard to do this, the portable implementation simply	1977	Since there is no portable change notification interface available, the
1706	calls C<stat (2)> regularly on the path to see if it changed somehow. You	1978	portable implementation simply calls C<stat(2)> regularly on the path
1707	can specify a recommended polling interval for this case. If you specify	1979	to see if it changed somehow. You can specify a recommended polling
1708	a polling interval of C<0> (highly recommended!) then a I<suitable,	1980	interval for this case. If you specify a polling interval of C<0> (highly
1709	unspecified default> value will be used (which you can expect to be around	1981	recommended!) then a I<suitable, unspecified default> value will be used
1710	five seconds, although this might change dynamically). Libev will also	1982	(which you can expect to be around five seconds, although this might
1711	impose a minimum interval which is currently around C<0.1>, but thats	1983	change dynamically). Libev will also impose a minimum interval which is
1712	usually overkill.	1984	currently around C<0.1>, but that's usually overkill.
1713		1985
1714	This watcher type is not meant for massive numbers of stat watchers,	1986	This watcher type is not meant for massive numbers of stat watchers,
1715	as even with OS-supported change notifications, this can be	1987	as even with OS-supported change notifications, this can be
1716	resource-intensive.	1988	resource-intensive.
1717		1989
1718	At the time of this writing, only the Linux inotify interface is	1990	At the time of this writing, the only OS-specific interface implemented
1719	implemented (implementing kqueue support is left as an exercise for the	1991	is the Linux inotify interface (implementing kqueue support is left as an
1720	reader, note, however, that the author sees no way of implementing ev_stat	1992	exercise for the reader. Note, however, that the author sees no way of
1721	semantics with kqueue). Inotify will be used to give hints only and should	1993	implementing C<ev_stat> semantics with kqueue, except as a hint).
1722	not change the semantics of C<ev_stat> watchers, which means that libev
1723	sometimes needs to fall back to regular polling again even with inotify,
1724	but changes are usually detected immediately, and if the file exists there
1725	will be no polling.
1726		1994
1727	=head3 ABI Issues (Largefile Support)	1995	=head3 ABI Issues (Largefile Support)
1728		1996
1729	Libev by default (unless the user overrides this) uses the default	1997	Libev by default (unless the user overrides this) uses the default
1730	compilation environment, which means that on systems with large file	1998	compilation environment, which means that on systems with large file
1731	support disabled by default, you get the 32 bit version of the stat	1999	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	2000	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	2001	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	2002	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	2003	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.	2004	most noticeably displayed with ev_stat and large file support.
1737		2005
1738	The solution for this is to lobby your distribution maker to make large	2006	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	2007	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	2008	optional. Libev cannot simply switch on large file support because it has
1741	to exchange stat structures with application programs compiled using the	2009	to exchange stat structures with application programs compiled using the
1742	default compilation environment.	2010	default compilation environment.
1743		2011
1744	=head3 Inotify	2012	=head3 Inotify and Kqueue
1745		2013
1746	When C<inotify (7)> support has been compiled into libev (generally only	2014	When C<inotify (7)> support has been compiled into libev and present at
1747	available on Linux) and present at runtime, it will be used to speed up	2015	runtime, it will be used to speed up change detection where possible. The
1748	change detection where possible. The inotify descriptor will be created lazily	2016	inotify descriptor will be created lazily when the first C<ev_stat>
1749	when the first C<ev_stat> watcher is being started.	2017	watcher is being started.
1750		2018
1751	Inotify presence does not change the semantics of C<ev_stat> watchers	2019	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	2020	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	2021	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.	2022	there are many cases where libev has to resort to regular C<stat> polling,
		2023	but as long as kernel 2.6.25 or newer is used (2.6.24 and older have too
		2024	many bugs), the path exists (i.e. stat succeeds), and the path resides on
		2025	a local filesystem (libev currently assumes only ext2/3, jfs, reiserfs and
		2026	xfs are fully working) libev usually gets away without polling.
1755		2027
1756	(There is no support for kqueue, as apparently it cannot be used to	2028	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	2029	implement this functionality, due to the requirement of having a file
1758	descriptor open on the object at all times).	2030	descriptor open on the object at all times, and detecting renames, unlinks
		2031	etc. is difficult.
		2032
		2033	=head3 C<stat ()> is a synchronous operation
		2034
		2035	Libev doesn't normally do any kind of I/O itself, and so is not blocking
		2036	the process. The exception are C<ev_stat> watchers - those call C<stat
		2037	()>, which is a synchronous operation.
		2038
		2039	For local paths, this usually doesn't matter: unless the system is very
		2040	busy or the intervals between stat's are large, a stat call will be fast,
		2041	as the path data is usually in memory already (except when starting the
		2042	watcher).
		2043
		2044	For networked file systems, calling C<stat ()> can block an indefinite
		2045	time due to network issues, and even under good conditions, a stat call
		2046	often takes multiple milliseconds.
		2047
		2048	Therefore, it is best to avoid using C<ev_stat> watchers on networked
		2049	paths, although this is fully supported by libev.
1759		2050
1760	=head3 The special problem of stat time resolution	2051	=head3 The special problem of stat time resolution
1761		2052
1762	The C<stat ()> system call only supports full-second resolution portably, and	2053	The C<stat ()> system call only supports full-second resolution portably,
1763	even on systems where the resolution is higher, many file systems still	2054	and even on systems where the resolution is higher, most file systems
1764	only support whole seconds.	2055	still only support whole seconds.
1765		2056
1766	That means that, if the time is the only thing that changes, you can	2057	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	2058	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	2059	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	2060	within the same second, C<ev_stat> will be unable to detect unless the
1770	data does not change.	2061	stat data does change in other ways (e.g. file size).
1771		2062
1772	The solution to this is to delay acting on a change for slightly more	2063	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	2064	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);	2065	a roughly one-second-delay C<ev_timer> (e.g. C<ev_timer_set (w, 0., 1.02);
1775	ev_timer_again (loop, w)>).	2066	ev_timer_again (loop, w)>).
…		…
1795	C<path>. The C<interval> is a hint on how quickly a change is expected to	2086	C<path>. The C<interval> is a hint on how quickly a change is expected to
1796	be detected and should normally be specified as C<0> to let libev choose	2087	be detected and should normally be specified as C<0> to let libev choose
1797	a suitable value. The memory pointed to by C<path> must point to the same	2088	a suitable value. The memory pointed to by C<path> must point to the same
1798	path for as long as the watcher is active.	2089	path for as long as the watcher is active.
1799		2090
1800	The callback will receive C<EV_STAT> when a change was detected, relative	2091	The callback will receive an C<EV_STAT> event when a change was detected,
1801	to the attributes at the time the watcher was started (or the last change	2092	relative to the attributes at the time the watcher was started (or the
1802	was detected).	2093	last change was detected).
1803		2094
1804	=item ev_stat_stat (loop, ev_stat *)	2095	=item ev_stat_stat (loop, ev_stat *)
1805		2096
1806	Updates the stat buffer immediately with new values. If you change the	2097	Updates the stat buffer immediately with new values. If you change the
1807	watched path in your callback, you could call this function to avoid	2098	watched path in your callback, you could call this function to avoid
…		…
1890		2181
1891		2182
1892	=head2 C<ev_idle> - when you've got nothing better to do...	2183	=head2 C<ev_idle> - when you've got nothing better to do...
1893		2184
1894	Idle watchers trigger events when no other events of the same or higher	2185	Idle watchers trigger events when no other events of the same or higher
1895	priority are pending (prepare, check and other idle watchers do not	2186	priority are pending (prepare, check and other idle watchers do not count
1896	count).	2187	as receiving "events").
1897		2188
1898	That is, as long as your process is busy handling sockets or timeouts	2189	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	2190	(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	2191	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	2192	are pending), the idle watchers are being called once per event loop
…		…
1912		2203
1913	=head3 Watcher-Specific Functions and Data Members	2204	=head3 Watcher-Specific Functions and Data Members
1914		2205
1915	=over 4	2206	=over 4
1916		2207
1917	=item ev_idle_init (ev_signal *, callback)	2208	=item ev_idle_init (ev_idle *, callback)
1918		2209
1919	Initialises and configures the idle watcher - it has no parameters of any	2210	Initialises and configures the idle watcher - it has no parameters of any
1920	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,	2211	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,
1921	believe me.	2212	believe me.
1922		2213
…		…
1926		2217
1927	Example: Dynamically allocate an C<ev_idle> watcher, start it, and in the	2218	Example: Dynamically allocate an C<ev_idle> watcher, start it, and in the
1928	callback, free it. Also, use no error checking, as usual.	2219	callback, free it. Also, use no error checking, as usual.
1929		2220
1930	static void	2221	static void
1931	idle_cb (struct ev_loop *loop, struct ev_idle *w, int revents)	2222	idle_cb (struct ev_loop *loop, ev_idle *w, int revents)
1932	{	2223	{
1933	free (w);	2224	free (w);
1934	// now do something you wanted to do when the program has	2225	// now do something you wanted to do when the program has
1935	// no longer anything immediate to do.	2226	// no longer anything immediate to do.
1936	}	2227	}
1937		2228
1938	struct ev_idle *idle_watcher = malloc (sizeof (struct ev_idle));	2229	ev_idle *idle_watcher = malloc (sizeof (ev_idle));
1939	ev_idle_init (idle_watcher, idle_cb);	2230	ev_idle_init (idle_watcher, idle_cb);
1940	ev_idle_start (loop, idle_cb);	2231	ev_idle_start (loop, idle_cb);
1941		2232
1942		2233
1943	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!	2234	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!
1944		2235
1945	Prepare and check watchers are usually (but not always) used in tandem:	2236	Prepare and check watchers are usually (but not always) used in pairs:
1946	prepare watchers get invoked before the process blocks and check watchers	2237	prepare watchers get invoked before the process blocks and check watchers
1947	afterwards.	2238	afterwards.
1948		2239
1949	You I<must not> call C<ev_loop> or similar functions that enter	2240	You I<must not> call C<ev_loop> or similar functions that enter
1950	the current event loop from either C<ev_prepare> or C<ev_check>	2241	the current event loop from either C<ev_prepare> or C<ev_check>
…		…
1953	those watchers, i.e. the sequence will always be C<ev_prepare>, blocking,	2244	those watchers, i.e. the sequence will always be C<ev_prepare>, blocking,
1954	C<ev_check> so if you have one watcher of each kind they will always be	2245	C<ev_check> so if you have one watcher of each kind they will always be
1955	called in pairs bracketing the blocking call.	2246	called in pairs bracketing the blocking call.
1956		2247
1957	Their main purpose is to integrate other event mechanisms into libev and	2248	Their main purpose is to integrate other event mechanisms into libev and
1958	their use is somewhat advanced. This could be used, for example, to track	2249	their use is somewhat advanced. They could be used, for example, to track
1959	variable changes, implement your own watchers, integrate net-snmp or a	2250	variable changes, implement your own watchers, integrate net-snmp or a
1960	coroutine library and lots more. They are also occasionally useful if	2251	coroutine library and lots more. They are also occasionally useful if
1961	you cache some data and want to flush it before blocking (for example,	2252	you cache some data and want to flush it before blocking (for example,
1962	in X programs you might want to do an C<XFlush ()> in an C<ev_prepare>	2253	in X programs you might want to do an C<XFlush ()> in an C<ev_prepare>
1963	watcher).	2254	watcher).
1964		2255
1965	This is done by examining in each prepare call which file descriptors need	2256	This is done by examining in each prepare call which file descriptors
1966	to be watched by the other library, registering C<ev_io> watchers for	2257	need to be watched by the other library, registering C<ev_io> watchers
1967	them and starting an C<ev_timer> watcher for any timeouts (many libraries	2258	for them and starting an C<ev_timer> watcher for any timeouts (many
1968	provide just this functionality). Then, in the check watcher you check for	2259	libraries provide exactly this functionality). Then, in the check watcher,
1969	any events that occurred (by checking the pending status of all watchers	2260	you check for any events that occurred (by checking the pending status
1970	and stopping them) and call back into the library. The I/O and timer	2261	of all watchers and stopping them) and call back into the library. The
1971	callbacks will never actually be called (but must be valid nevertheless,	2262	I/O and timer callbacks will never actually be called (but must be valid
1972	because you never know, you know?).	2263	nevertheless, because you never know, you know?).
1973		2264
1974	As another example, the Perl Coro module uses these hooks to integrate	2265	As another example, the Perl Coro module uses these hooks to integrate
1975	coroutines into libev programs, by yielding to other active coroutines	2266	coroutines into libev programs, by yielding to other active coroutines
1976	during each prepare and only letting the process block if no coroutines	2267	during each prepare and only letting the process block if no coroutines
1977	are ready to run (it's actually more complicated: it only runs coroutines	2268	are ready to run (it's actually more complicated: it only runs coroutines
…		…
1980	loop from blocking if lower-priority coroutines are active, thus mapping	2271	loop from blocking if lower-priority coroutines are active, thus mapping
1981	low-priority coroutines to idle/background tasks).	2272	low-priority coroutines to idle/background tasks).
1982		2273
1983	It is recommended to give C<ev_check> watchers highest (C<EV_MAXPRI>)	2274	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	2275	priority, to ensure that they are being run before any other watchers
		2276	after the poll (this doesn't matter for C<ev_prepare> watchers).
		2277
1985	after the poll. Also, C<ev_check> watchers (and C<ev_prepare> watchers,	2278	Also, C<ev_check> watchers (and C<ev_prepare> watchers, too) should not
1986	too) should not activate ("feed") events into libev. While libev fully	2279	activate ("feed") events into libev. While libev fully supports this, they
1987	supports this, they might get executed before other C<ev_check> watchers	2280	might get executed before other C<ev_check> watchers did their job. As
1988	did their job. As C<ev_check> watchers are often used to embed other	2281	C<ev_check> watchers are often used to embed other (non-libev) event
1989	(non-libev) event loops those other event loops might be in an unusable	2282	loops those other event loops might be in an unusable state until their
1990	state until their C<ev_check> watcher ran (always remind yourself to	2283	C<ev_check> watcher ran (always remind yourself to coexist peacefully with
1991	coexist peacefully with others).	2284	others).
1992		2285
1993	=head3 Watcher-Specific Functions and Data Members	2286	=head3 Watcher-Specific Functions and Data Members
1994		2287
1995	=over 4	2288	=over 4
1996		2289
…		…
1998		2291
1999	=item ev_check_init (ev_check *, callback)	2292	=item ev_check_init (ev_check *, callback)
2000		2293
2001	Initialises and configures the prepare or check watcher - they have no	2294	Initialises and configures the prepare or check watcher - they have no
2002	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>	2295	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>
2003	macros, but using them is utterly, utterly and completely pointless.	2296	macros, but using them is utterly, utterly, utterly and completely
		2297	pointless.
2004		2298
2005	=back	2299	=back
2006		2300
2007	=head3 Examples	2301	=head3 Examples
2008		2302
…		…
2021		2315
2022	static ev_io iow [nfd];	2316	static ev_io iow [nfd];
2023	static ev_timer tw;	2317	static ev_timer tw;
2024		2318
2025	static void	2319	static void
2026	io_cb (ev_loop *loop, ev_io *w, int revents)	2320	io_cb (struct ev_loop *loop, ev_io *w, int revents)
2027	{	2321	{
2028	}	2322	}
2029		2323
2030	// create io watchers for each fd and a timer before blocking	2324	// create io watchers for each fd and a timer before blocking
2031	static void	2325	static void
2032	adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents)	2326	adns_prepare_cb (struct ev_loop *loop, ev_prepare *w, int revents)
2033	{	2327	{
2034	int timeout = 3600000;	2328	int timeout = 3600000;
2035	struct pollfd fds [nfd];	2329	struct pollfd fds [nfd];
2036	// actual code will need to loop here and realloc etc.	2330	// actual code will need to loop here and realloc etc.
2037	adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));	2331	adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));
…		…
2052	}	2346	}
2053	}	2347	}
2054		2348
2055	// stop all watchers after blocking	2349	// stop all watchers after blocking
2056	static void	2350	static void
2057	adns_check_cb (ev_loop *loop, ev_check *w, int revents)	2351	adns_check_cb (struct ev_loop *loop, ev_check *w, int revents)
2058	{	2352	{
2059	ev_timer_stop (loop, &tw);	2353	ev_timer_stop (loop, &tw);
2060		2354
2061	for (int i = 0; i < nfd; ++i)	2355	for (int i = 0; i < nfd; ++i)
2062	{	2356	{
…		…
2101	}	2395	}
2102		2396
2103	// do not ever call adns_afterpoll	2397	// do not ever call adns_afterpoll
2104		2398
2105	Method 4: Do not use a prepare or check watcher because the module you	2399	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	2400	want to embed is not flexible enough to support it. Instead, you can
2107	their poll function. The drawback with this solution is that the main	2401	override their poll function. The drawback with this solution is that the
2108	loop is now no longer controllable by EV. The C<Glib::EV> module does	2402	main loop is now no longer controllable by EV. The C<Glib::EV> module uses
2109	this.	2403	this approach, effectively embedding EV as a client into the horrible
		2404	libglib event loop.
2110		2405
2111	static gint	2406	static gint
2112	event_poll_func (GPollFD *fds, guint nfds, gint timeout)	2407	event_poll_func (GPollFD *fds, guint nfds, gint timeout)
2113	{	2408	{
2114	int got_events = 0;	2409	int got_events = 0;
…		…
2145	prioritise I/O.	2440	prioritise I/O.
2146		2441
2147	As an example for a bug workaround, the kqueue backend might only support	2442	As an example for a bug workaround, the kqueue backend might only support
2148	sockets on some platform, so it is unusable as generic backend, but you	2443	sockets on some platform, so it is unusable as generic backend, but you
2149	still want to make use of it because you have many sockets and it scales	2444	still want to make use of it because you have many sockets and it scales
2150	so nicely. In this case, you would create a kqueue-based loop and embed it	2445	so nicely. In this case, you would create a kqueue-based loop and embed
2151	into your default loop (which might use e.g. poll). Overall operation will	2446	it into your default loop (which might use e.g. poll). Overall operation
2152	be a bit slower because first libev has to poll and then call kevent, but	2447	will be a bit slower because first libev has to call C<poll> and then
2153	at least you can use both at what they are best.	2448	C<kevent>, but at least you can use both mechanisms for what they are
		2449	best: C<kqueue> for scalable sockets and C<poll> if you want it to work :)
2154		2450
2155	As for prioritising I/O: rarely you have the case where some fds have	2451	As for prioritising I/O: under rare circumstances you have the case where
2156	to be watched and handled very quickly (with low latency), and even	2452	some fds have to be watched and handled very quickly (with low latency),
2157	priorities and idle watchers might have too much overhead. In this case	2453	and even priorities and idle watchers might have too much overhead. In
2158	you would put all the high priority stuff in one loop and all the rest in	2454	this case you would put all the high priority stuff in one loop and all
2159	a second one, and embed the second one in the first.	2455	the rest in a second one, and embed the second one in the first.
2160		2456
2161	As long as the watcher is active, the callback will be invoked every time	2457	As long as the watcher is active, the callback will be invoked every
2162	there might be events pending in the embedded loop. The callback must then	2458	time there might be events pending in the embedded loop. The callback
2163	call C<ev_embed_sweep (mainloop, watcher)> to make a single sweep and invoke	2459	must then call C<ev_embed_sweep (mainloop, watcher)> to make a single
2164	their callbacks (you could also start an idle watcher to give the embedded	2460	sweep and invoke their callbacks (the callback doesn't need to invoke the
2165	loop strictly lower priority for example). You can also set the callback	2461	C<ev_embed_sweep> function directly, it could also start an idle watcher
2166	to C<0>, in which case the embed watcher will automatically execute the	2462	to give the embedded loop strictly lower priority for example).
2167	embedded loop sweep.
2168		2463
2169	As long as the watcher is started it will automatically handle events. The	2464	You can also set the callback to C<0>, in which case the embed watcher
2170	callback will be invoked whenever some events have been handled. You can	2465	will automatically execute the embedded loop sweep whenever necessary.
2171	set the callback to C<0> to avoid having to specify one if you are not
2172	interested in that.
2173		2466
2174	Also, there have not currently been made special provisions for forking:	2467	Fork detection will be handled transparently while the C<ev_embed> watcher
2175	when you fork, you not only have to call C<ev_loop_fork> on both loops,	2468	is active, i.e., the embedded loop will automatically be forked when the
2176	but you will also have to stop and restart any C<ev_embed> watchers	2469	embedding loop forks. In other cases, the user is responsible for calling
2177	yourself.	2470	C<ev_loop_fork> on the embedded loop.
2178		2471
2179	Unfortunately, not all backends are embeddable, only the ones returned by	2472	Unfortunately, not all backends are embeddable: only the ones returned by
2180	C<ev_embeddable_backends> are, which, unfortunately, does not include any	2473	C<ev_embeddable_backends> are, which, unfortunately, does not include any
2181	portable one.	2474	portable one.
2182		2475
2183	So when you want to use this feature you will always have to be prepared	2476	So when you want to use this feature you will always have to be prepared
2184	that you cannot get an embeddable loop. The recommended way to get around	2477	that you cannot get an embeddable loop. The recommended way to get around
2185	this is to have a separate variables for your embeddable loop, try to	2478	this is to have a separate variables for your embeddable loop, try to
2186	create it, and if that fails, use the normal loop for everything.	2479	create it, and if that fails, use the normal loop for everything.
		2480
		2481	=head3 C<ev_embed> and fork
		2482
		2483	While the C<ev_embed> watcher is running, forks in the embedding loop will
		2484	automatically be applied to the embedded loop as well, so no special
		2485	fork handling is required in that case. When the watcher is not running,
		2486	however, it is still the task of the libev user to call C<ev_loop_fork ()>
		2487	as applicable.
2187		2488
2188	=head3 Watcher-Specific Functions and Data Members	2489	=head3 Watcher-Specific Functions and Data Members
2189		2490
2190	=over 4	2491	=over 4
2191		2492
…		…
2219	C<loop_lo> (which is C<loop_hi> in the case no embeddable loop can be	2520	C<loop_lo> (which is C<loop_hi> in the case no embeddable loop can be
2220	used).	2521	used).
2221		2522
2222	struct ev_loop *loop_hi = ev_default_init (0);	2523	struct ev_loop *loop_hi = ev_default_init (0);
2223	struct ev_loop *loop_lo = 0;	2524	struct ev_loop *loop_lo = 0;
2224	struct ev_embed embed;	2525	ev_embed embed;
2225		2526
2226	// see if there is a chance of getting one that works	2527	// see if there is a chance of getting one that works
2227	// (remember that a flags value of 0 means autodetection)	2528	// (remember that a flags value of 0 means autodetection)
2228	loop_lo = ev_embeddable_backends () & ev_recommended_backends ()	2529	loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
2229	? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())	2530	? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
…		…
2243	kqueue implementation). Store the kqueue/socket-only event loop in	2544	kqueue implementation). Store the kqueue/socket-only event loop in
2244	C<loop_socket>. (One might optionally use C<EVFLAG_NOENV>, too).	2545	C<loop_socket>. (One might optionally use C<EVFLAG_NOENV>, too).
2245		2546
2246	struct ev_loop *loop = ev_default_init (0);	2547	struct ev_loop *loop = ev_default_init (0);
2247	struct ev_loop *loop_socket = 0;	2548	struct ev_loop *loop_socket = 0;
2248	struct ev_embed embed;	2549	ev_embed embed;
2249		2550
2250	if (ev_supported_backends () & ~ev_recommended_backends () & EVBACKEND_KQUEUE)	2551	if (ev_supported_backends () & ~ev_recommended_backends () & EVBACKEND_KQUEUE)
2251	if ((loop_socket = ev_loop_new (EVBACKEND_KQUEUE))	2552	if ((loop_socket = ev_loop_new (EVBACKEND_KQUEUE))
2252	{	2553	{
2253	ev_embed_init (&embed, 0, loop_socket);	2554	ev_embed_init (&embed, 0, loop_socket);
…		…
2309	is that the author does not know of a simple (or any) algorithm for a	2610	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	2611	multiple-writer-single-reader queue that works in all cases and doesn't
2311	need elaborate support such as pthreads.	2612	need elaborate support such as pthreads.
2312		2613
2313	That means that if you want to queue data, you have to provide your own	2614	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	2615	queue. But at least I can tell you how to implement locking around your
2315	queue:	2616	queue:
2316		2617
2317	=over 4	2618	=over 4
2318		2619
2319	=item queueing from a signal handler context	2620	=item queueing from a signal handler context
2320		2621
2321	To implement race-free queueing, you simply add to the queue in the signal	2622	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	2623	handler but you block the signal handler in the watcher callback. Here is
2323	some fictitious SIGUSR1 handler:	2624	an example that does that for some fictitious SIGUSR1 handler:
2324		2625
2325	static ev_async mysig;	2626	static ev_async mysig;
2326		2627
2327	static void	2628	static void
2328	sigusr1_handler (void)	2629	sigusr1_handler (void)
…		…
2394	=over 4	2695	=over 4
2395		2696
2396	=item ev_async_init (ev_async *, callback)	2697	=item ev_async_init (ev_async *, callback)
2397		2698
2398	Initialises and configures the async watcher - it has no parameters of any	2699	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,	2700	kind. There is a C<ev_async_set> macro, but using it is utterly pointless,
2400	believe me.	2701	trust me.
2401		2702
2402	=item ev_async_send (loop, ev_async *)	2703	=item ev_async_send (loop, ev_async *)
2403		2704
2404	Sends/signals/activates the given C<ev_async> watcher, that is, feeds	2705	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	2706	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	2707	C<ev_feed_event>, this call is safe to do from other threads, signal or
2407	similar contexts (see the discussion of C<EV_ATOMIC_T> in the embedding	2708	similar contexts (see the discussion of C<EV_ATOMIC_T> in the embedding
2408	section below on what exactly this means).	2709	section below on what exactly this means).
2409		2710
		2711	Note that, as with other watchers in libev, multiple events might get
		2712	compressed into a single callback invocation (another way to look at this
		2713	is that C<ev_async> watchers are level-triggered, set on C<ev_async_send>,
		2714	reset when the event loop detects that).
		2715
2410	This call incurs the overhead of a system call only once per loop iteration,	2716	This call incurs the overhead of a system call only once per event loop
2411	so while the overhead might be noticeable, it doesn't apply to repeated	2717	iteration, so while the overhead might be noticeable, it doesn't apply to
2412	calls to C<ev_async_send>.	2718	repeated calls to C<ev_async_send> for the same event loop.
2413		2719
2414	=item bool = ev_async_pending (ev_async *)	2720	=item bool = ev_async_pending (ev_async *)
2415		2721
2416	Returns a non-zero value when C<ev_async_send> has been called on the	2722	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	2723	watcher but the event has not yet been processed (or even noted) by the
…		…
2420	C<ev_async_send> sets a flag in the watcher and wakes up the loop. When	2726	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,	2727	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	2728	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.	2729	quickly check whether invoking the loop might be a good idea.
2424		2730
2425	Not that this does I<not> check whether the watcher itself is pending, only	2731	Not that this does I<not> check whether the watcher itself is pending,
2426	whether it has been requested to make this watcher pending.	2732	only whether it has been requested to make this watcher pending: there
		2733	is a time window between the event loop checking and resetting the async
		2734	notification, and the callback being invoked.
2427		2735
2428	=back	2736	=back
2429		2737
2430		2738
2431	=head1 OTHER FUNCTIONS	2739	=head1 OTHER FUNCTIONS
…		…
2435	=over 4	2743	=over 4
2436		2744
2437	=item ev_once (loop, int fd, int events, ev_tstamp timeout, callback)	2745	=item ev_once (loop, int fd, int events, ev_tstamp timeout, callback)
2438		2746
2439	This function combines a simple timer and an I/O watcher, calls your	2747	This function combines a simple timer and an I/O watcher, calls your
2440	callback on whichever event happens first and automatically stop both	2748	callback on whichever event happens first and automatically stops both
2441	watchers. This is useful if you want to wait for a single event on an fd	2749	watchers. This is useful if you want to wait for a single event on an fd
2442	or timeout without having to allocate/configure/start/stop/free one or	2750	or timeout without having to allocate/configure/start/stop/free one or
2443	more watchers yourself.	2751	more watchers yourself.
2444		2752
2445	If C<fd> is less than 0, then no I/O watcher will be started and events	2753	If C<fd> is less than 0, then no I/O watcher will be started and the
2446	is being ignored. Otherwise, an C<ev_io> watcher for the given C<fd> and	2754	C<events> argument is being ignored. Otherwise, an C<ev_io> watcher for
2447	C<events> set will be created and started.	2755	the given C<fd> and C<events> set will be created and started.
2448		2756
2449	If C<timeout> is less than 0, then no timeout watcher will be	2757	If C<timeout> is less than 0, then no timeout watcher will be
2450	started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and	2758	started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and
2451	repeat = 0) will be started. While C<0> is a valid timeout, it is of	2759	repeat = 0) will be started. C<0> is a valid timeout.
2452	dubious value.
2453		2760
2454	The callback has the type C<void (*cb)(int revents, void *arg)> and gets	2761	The callback has the type C<void (*cb)(int revents, void *arg)> and gets
2455	passed an C<revents> set like normal event callbacks (a combination of	2762	passed an C<revents> set like normal event callbacks (a combination of
2456	C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMEOUT>) and the C<arg>	2763	C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMEOUT>) and the C<arg>
2457	value passed to C<ev_once>:	2764	value passed to C<ev_once>. Note that it is possible to receive I<both>
		2765	a timeout and an io event at the same time - you probably should give io
		2766	events precedence.
		2767
		2768	Example: wait up to ten seconds for data to appear on STDIN_FILENO.
2458		2769
2459	static void stdin_ready (int revents, void *arg)	2770	static void stdin_ready (int revents, void *arg)
2460	{	2771	{
		2772	if (revents & EV_READ)
		2773	/* stdin might have data for us, joy! */;
2461	if (revents & EV_TIMEOUT)	2774	else if (revents & EV_TIMEOUT)
2462	/* doh, nothing entered */;	2775	/* doh, nothing entered */;
2463	else if (revents & EV_READ)
2464	/* stdin might have data for us, joy! */;
2465	}	2776	}
2466		2777
2467	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);	2778	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
2468		2779
2469	=item ev_feed_event (ev_loop *, watcher *, int revents)	2780	=item ev_feed_event (struct ev_loop *, watcher *, int revents)
2470		2781
2471	Feeds the given event set into the event loop, as if the specified event	2782	Feeds the given event set into the event loop, as if the specified event
2472	had happened for the specified watcher (which must be a pointer to an	2783	had happened for the specified watcher (which must be a pointer to an
2473	initialised but not necessarily started event watcher).	2784	initialised but not necessarily started event watcher).
2474		2785
2475	=item ev_feed_fd_event (ev_loop *, int fd, int revents)	2786	=item ev_feed_fd_event (struct ev_loop *, int fd, int revents)
2476		2787
2477	Feed an event on the given fd, as if a file descriptor backend detected	2788	Feed an event on the given fd, as if a file descriptor backend detected
2478	the given events it.	2789	the given events it.
2479		2790
2480	=item ev_feed_signal_event (ev_loop *loop, int signum)	2791	=item ev_feed_signal_event (struct ev_loop *loop, int signum)
2481		2792
2482	Feed an event as if the given signal occurred (C<loop> must be the default	2793	Feed an event as if the given signal occurred (C<loop> must be the default
2483	loop!).	2794	loop!).
2484		2795
2485	=back	2796	=back
…		…
2607		2918
2608	myclass obj;	2919	myclass obj;
2609	ev::io iow;	2920	ev::io iow;
2610	iow.set <myclass, &myclass::io_cb> (&obj);	2921	iow.set <myclass, &myclass::io_cb> (&obj);
2611		2922
		2923	=item w->set (object *)
		2924
		2925	This is an B<experimental> feature that might go away in a future version.
		2926
		2927	This is a variation of a method callback - leaving out the method to call
		2928	will default the method to C<operator ()>, which makes it possible to use
		2929	functor objects without having to manually specify the C<operator ()> all
		2930	the time. Incidentally, you can then also leave out the template argument
		2931	list.
		2932
		2933	The C<operator ()> method prototype must be C<void operator ()(watcher &w,
		2934	int revents)>.
		2935
		2936	See the method-C<set> above for more details.
		2937
		2938	Example: use a functor object as callback.
		2939
		2940	struct myfunctor
		2941	{
		2942	void operator() (ev::io &w, int revents)
		2943	{
		2944	...
		2945	}
		2946	}
		2947
		2948	myfunctor f;
		2949
		2950	ev::io w;
		2951	w.set (&f);
		2952
2612	=item w->set<function> (void *data = 0)	2953	=item w->set<function> (void *data = 0)
2613		2954
2614	Also sets a callback, but uses a static method or plain function as	2955	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	2956	callback. The optional C<data> argument will be stored in the watcher's
2616	C<data> member and is free for you to use.	2957	C<data> member and is free for you to use.
2617		2958
2618	The prototype of the C<function> must be C<void (*)(ev::TYPE &w, int)>.	2959	The prototype of the C<function> must be C<void (*)(ev::TYPE &w, int)>.
2619		2960
2620	See the method-C<set> above for more details.	2961	See the method-C<set> above for more details.
2621		2962
2622	Example:	2963	Example: Use a plain function as callback.
2623		2964
2624	static void io_cb (ev::io &w, int revents) { }	2965	static void io_cb (ev::io &w, int revents) { }
2625	iow.set <io_cb> ();	2966	iow.set <io_cb> ();
2626		2967
2627	=item w->set (struct ev_loop *)	2968	=item w->set (struct ev_loop *)
…		…
2665	Example: Define a class with an IO and idle watcher, start one of them in	3006	Example: Define a class with an IO and idle watcher, start one of them in
2666	the constructor.	3007	the constructor.
2667		3008
2668	class myclass	3009	class myclass
2669	{	3010	{
2670	ev::io io; void io_cb (ev::io &w, int revents);	3011	ev::io io ; void io_cb (ev::io &w, int revents);
2671	ev:idle idle void idle_cb (ev::idle &w, int revents);	3012	ev::idle idle; void idle_cb (ev::idle &w, int revents);
2672		3013
2673	myclass (int fd)	3014	myclass (int fd)
2674	{	3015	{
2675	io .set <myclass, &myclass::io_cb > (this);	3016	io .set <myclass, &myclass::io_cb > (this);
2676	idle.set <myclass, &myclass::idle_cb> (this);	3017	idle.set <myclass, &myclass::idle_cb> (this);
…		…
2692	=item Perl	3033	=item Perl
2693		3034
2694	The EV module implements the full libev API and is actually used to test	3035	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,	3036	libev. EV is developed together with libev. Apart from the EV core module,
2696	there are additional modules that implement libev-compatible interfaces	3037	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	3038	to C<libadns> (C<EV::ADNS>, but C<AnyEvent::DNS> is preferred nowadays),
2698	C<libglib> event core (C<Glib::EV> and C<EV::Glib>).	3039	C<Net::SNMP> (C<Net::SNMP::EV>) and the C<libglib> event core (C<Glib::EV>
		3040	and C<EV::Glib>).
2699		3041
2700	It can be found and installed via CPAN, its homepage is at	3042	It can be found and installed via CPAN, its homepage is at
2701	L<http://software.schmorp.de/pkg/EV>.	3043	L<http://software.schmorp.de/pkg/EV>.
2702		3044
2703	=item Python	3045	=item Python
2704		3046
2705	Python bindings can be found at L<http://code.google.com/p/pyev/>. It	3047	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	3048	seems to be quite complete and well-documented.
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		3049
2712	=item Ruby	3050	=item Ruby
2713		3051
2714	Tony Arcieri has written a ruby extension that offers access to a subset	3052	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	3053	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	3054	more on top of it. It can be found via gem servers. Its homepage is at
2717	L<http://rev.rubyforge.org/>.	3055	L<http://rev.rubyforge.org/>.
2718		3056
		3057	Roger Pack reports that using the link order C<-lws2_32 -lmsvcrt-ruby-190>
		3058	makes rev work even on mingw.
		3059
		3060	=item Haskell
		3061
		3062	A haskell binding to libev is available at
		3063	L<http://hackage.haskell.org/cgi-bin/hackage-scripts/package/hlibev>.
		3064
2719	=item D	3065	=item D
2720		3066
2721	Leandro Lucarella has written a D language binding (F<ev.d>) for libev, to	3067	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>.	3068	be found at L<http://proj.llucax.com.ar/wiki/evd>.
		3069
		3070	=item Ocaml
		3071
		3072	Erkki Seppala has written Ocaml bindings for libev, to be found at
		3073	L<http://modeemi.cs.tut.fi/~flux/software/ocaml-ev/>.
2723		3074
2724	=back	3075	=back
2725		3076
2726		3077
2727	=head1 MACRO MAGIC	3078	=head1 MACRO MAGIC
…		…
2828		3179
2829	#define EV_STANDALONE 1	3180	#define EV_STANDALONE 1
2830	#include "ev.h"	3181	#include "ev.h"
2831		3182
2832	Both header files and implementation files can be compiled with a C++	3183	Both header files and implementation files can be compiled with a C++
2833	compiler (at least, thats a stated goal, and breakage will be treated	3184	compiler (at least, that's a stated goal, and breakage will be treated
2834	as a bug).	3185	as a bug).
2835		3186
2836	You need the following files in your source tree, or in a directory	3187	You need the following files in your source tree, or in a directory
2837	in your include path (e.g. in libev/ when using -Ilibev):	3188	in your include path (e.g. in libev/ when using -Ilibev):
2838		3189
…		…
2882		3233
2883	=head2 PREPROCESSOR SYMBOLS/MACROS	3234	=head2 PREPROCESSOR SYMBOLS/MACROS
2884		3235
2885	Libev can be configured via a variety of preprocessor symbols you have to	3236	Libev can be configured via a variety of preprocessor symbols you have to
2886	define before including any of its files. The default in the absence of	3237	define before including any of its files. The default in the absence of
2887	autoconf is noted for every option.	3238	autoconf is documented for every option.
2888		3239
2889	=over 4	3240	=over 4
2890		3241
2891	=item EV_STANDALONE	3242	=item EV_STANDALONE
2892		3243
…		…
2894	keeps libev from including F<config.h>, and it also defines dummy	3245	keeps libev from including F<config.h>, and it also defines dummy
2895	implementations for some libevent functions (such as logging, which is not	3246	implementations for some libevent functions (such as logging, which is not
2896	supported). It will also not define any of the structs usually found in	3247	supported). It will also not define any of the structs usually found in
2897	F<event.h> that are not directly supported by the libev core alone.	3248	F<event.h> that are not directly supported by the libev core alone.
2898		3249
		3250	In stanbdalone mode, libev will still try to automatically deduce the
		3251	configuration, but has to be more conservative.
		3252
2899	=item EV_USE_MONOTONIC	3253	=item EV_USE_MONOTONIC
2900		3254
2901	If defined to be C<1>, libev will try to detect the availability of the	3255	If defined to be C<1>, libev will try to detect the availability of the
2902	monotonic clock option at both compile time and runtime. Otherwise no use	3256	monotonic clock option at both compile time and runtime. Otherwise no
2903	of the monotonic clock option will be attempted. If you enable this, you	3257	use of the monotonic clock option will be attempted. If you enable this,
2904	usually have to link against librt or something similar. Enabling it when	3258	you usually have to link against librt or something similar. Enabling it
2905	the functionality isn't available is safe, though, although you have	3259	when the functionality isn't available is safe, though, although you have
2906	to make sure you link against any libraries where the C<clock_gettime>	3260	to make sure you link against any libraries where the C<clock_gettime>
2907	function is hiding in (often F<-lrt>).	3261	function is hiding in (often F<-lrt>). See also C<EV_USE_CLOCK_SYSCALL>.
2908		3262
2909	=item EV_USE_REALTIME	3263	=item EV_USE_REALTIME
2910		3264
2911	If defined to be C<1>, libev will try to detect the availability of the	3265	If defined to be C<1>, libev will try to detect the availability of the
2912	real-time clock option at compile time (and assume its availability at	3266	real-time clock option at compile time (and assume its availability
2913	runtime if successful). Otherwise no use of the real-time clock option will	3267	at runtime if successful). Otherwise no use of the real-time clock
2914	be attempted. This effectively replaces C<gettimeofday> by C<clock_get	3268	option will be attempted. This effectively replaces C<gettimeofday>
2915	(CLOCK_REALTIME, ...)> and will not normally affect correctness. See the	3269	by C<clock_get (CLOCK_REALTIME, ...)> and will not normally affect
2916	note about libraries in the description of C<EV_USE_MONOTONIC>, though.	3270	correctness. See the note about libraries in the description of
		3271	C<EV_USE_MONOTONIC>, though. Defaults to the opposite value of
		3272	C<EV_USE_CLOCK_SYSCALL>.
		3273
		3274	=item EV_USE_CLOCK_SYSCALL
		3275
		3276	If defined to be C<1>, libev will try to use a direct syscall instead
		3277	of calling the system-provided C<clock_gettime> function. This option
		3278	exists because on GNU/Linux, C<clock_gettime> is in C<librt>, but C<librt>
		3279	unconditionally pulls in C<libpthread>, slowing down single-threaded
		3280	programs needlessly. Using a direct syscall is slightly slower (in
		3281	theory), because no optimised vdso implementation can be used, but avoids
		3282	the pthread dependency. Defaults to C<1> on GNU/Linux with glibc 2.x or
		3283	higher, as it simplifies linking (no need for C<-lrt>).
2917		3284
2918	=item EV_USE_NANOSLEEP	3285	=item EV_USE_NANOSLEEP
2919		3286
2920	If defined to be C<1>, libev will assume that C<nanosleep ()> is available	3287	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 ()>.	3288	and will use it for delays. Otherwise it will use C<select ()>.
…		…
2937		3304
2938	=item EV_SELECT_USE_FD_SET	3305	=item EV_SELECT_USE_FD_SET
2939		3306
2940	If defined to C<1>, then the select backend will use the system C<fd_set>	3307	If defined to C<1>, then the select backend will use the system C<fd_set>
2941	structure. This is useful if libev doesn't compile due to a missing	3308	structure. This is useful if libev doesn't compile due to a missing
2942	C<NFDBITS> or C<fd_mask> definition or it mis-guesses the bitset layout on	3309	C<NFDBITS> or C<fd_mask> definition or it mis-guesses the bitset layout
2943	exotic systems. This usually limits the range of file descriptors to some	3310	on exotic systems. This usually limits the range of file descriptors to
2944	low limit such as 1024 or might have other limitations (winsocket only	3311	some low limit such as 1024 or might have other limitations (winsocket
2945	allows 64 sockets). The C<FD_SETSIZE> macro, set before compilation, might	3312	only allows 64 sockets). The C<FD_SETSIZE> macro, set before compilation,
2946	influence the size of the C<fd_set> used.	3313	configures the maximum size of the C<fd_set>.
2947		3314
2948	=item EV_SELECT_IS_WINSOCKET	3315	=item EV_SELECT_IS_WINSOCKET
2949		3316
2950	When defined to C<1>, the select backend will assume that	3317	When defined to C<1>, the select backend will assume that
2951	select/socket/connect etc. don't understand file descriptors but	3318	select/socket/connect etc. don't understand file descriptors but
…		…
3062	When doing priority-based operations, libev usually has to linearly search	3429	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	3430	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	3431	and time, so using the defaults of five priorities (-2 .. +2) is usually
3065	fine.	3432	fine.
3066		3433
3067	If your embedding application does not need any priorities, defining these both to	3434	If your embedding application does not need any priorities, defining these
3068	C<0> will save some memory and CPU.	3435	both to C<0> will save some memory and CPU.
3069		3436
3070	=item EV_PERIODIC_ENABLE	3437	=item EV_PERIODIC_ENABLE
3071		3438
3072	If undefined or defined to be C<1>, then periodic timers are supported. If	3439	If undefined or defined to be C<1>, then periodic timers are supported. If
3073	defined to be C<0>, then they are not. Disabling them saves a few kB of	3440	defined to be C<0>, then they are not. Disabling them saves a few kB of
…		…
3080	code.	3447	code.
3081		3448
3082	=item EV_EMBED_ENABLE	3449	=item EV_EMBED_ENABLE
3083		3450
3084	If undefined or defined to be C<1>, then embed watchers are supported. If	3451	If undefined or defined to be C<1>, then embed watchers are supported. If
3085	defined to be C<0>, then they are not.	3452	defined to be C<0>, then they are not. Embed watchers rely on most other
		3453	watcher types, which therefore must not be disabled.
3086		3454
3087	=item EV_STAT_ENABLE	3455	=item EV_STAT_ENABLE
3088		3456
3089	If undefined or defined to be C<1>, then stat watchers are supported. If	3457	If undefined or defined to be C<1>, then stat watchers are supported. If
3090	defined to be C<0>, then they are not.	3458	defined to be C<0>, then they are not.
…		…
3122	two).	3490	two).
3123		3491
3124	=item EV_USE_4HEAP	3492	=item EV_USE_4HEAP
3125		3493
3126	Heaps are not very cache-efficient. To improve the cache-efficiency of the	3494	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	3495	timer and periodics heaps, libev uses a 4-heap when this symbol is defined
3128	to C<1>. The 4-heap uses more complicated (longer) code but has	3496	to C<1>. The 4-heap uses more complicated (longer) code but has noticeably
3129	noticeably faster performance with many (thousands) of watchers.	3497	faster performance with many (thousands) of watchers.
3130		3498
3131	The default is C<1> unless C<EV_MINIMAL> is set in which case it is C<0>	3499	The default is C<1> unless C<EV_MINIMAL> is set in which case it is C<0>
3132	(disabled).	3500	(disabled).
3133		3501
3134	=item EV_HEAP_CACHE_AT	3502	=item EV_HEAP_CACHE_AT
3135		3503
3136	Heaps are not very cache-efficient. To improve the cache-efficiency of the	3504	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	3505	timer and periodics heaps, libev can cache the timestamp (I<at>) within
3138	the heap structure (selected by defining C<EV_HEAP_CACHE_AT> to C<1>),	3506	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,	3507	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	3508	but avoids random read accesses on heap changes. This improves performance
3141	noticeably with with many (hundreds) of watchers.	3509	noticeably with many (hundreds) of watchers.
3142		3510
3143	The default is C<1> unless C<EV_MINIMAL> is set in which case it is C<0>	3511	The default is C<1> unless C<EV_MINIMAL> is set in which case it is C<0>
3144	(disabled).	3512	(disabled).
3145		3513
3146	=item EV_VERIFY	3514	=item EV_VERIFY
…		…
3152	called once per loop, which can slow down libev. If set to C<3>, then the	3520	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	3521	verification code will be called very frequently, which will slow down
3154	libev considerably.	3522	libev considerably.
3155		3523
3156	The default is C<1>, unless C<EV_MINIMAL> is set, in which case it will be	3524	The default is C<1>, unless C<EV_MINIMAL> is set, in which case it will be
3157	C<0.>	3525	C<0>.
3158		3526
3159	=item EV_COMMON	3527	=item EV_COMMON
3160		3528
3161	By default, all watchers have a C<void *data> member. By redefining	3529	By default, all watchers have a C<void *data> member. By redefining
3162	this macro to a something else you can include more and other types of	3530	this macro to a something else you can include more and other types of
…		…
3179	and the way callbacks are invoked and set. Must expand to a struct member	3547	and the way callbacks are invoked and set. Must expand to a struct member
3180	definition and a statement, respectively. See the F<ev.h> header file for	3548	definition and a statement, respectively. See the F<ev.h> header file for
3181	their default definitions. One possible use for overriding these is to	3549	their default definitions. One possible use for overriding these is to
3182	avoid the C<struct ev_loop *> as first argument in all cases, or to use	3550	avoid the C<struct ev_loop *> as first argument in all cases, or to use
3183	method calls instead of plain function calls in C++.	3551	method calls instead of plain function calls in C++.
		3552
		3553	=back
3184		3554
3185	=head2 EXPORTED API SYMBOLS	3555	=head2 EXPORTED API SYMBOLS
3186		3556
3187	If you need to re-export the API (e.g. via a DLL) and you need a list of	3557	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	3558	exported symbols, you can use the provided F<Symbol.*> files which list
…		…
3235	And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:	3605	And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:
3236		3606
3237	#include "ev_cpp.h"	3607	#include "ev_cpp.h"
3238	#include "ev.c"	3608	#include "ev.c"
3239		3609
		3610	=head1 INTERACTION WITH OTHER PROGRAMS OR LIBRARIES
3240		3611
3241	=head1 THREADS AND COROUTINES	3612	=head2 THREADS AND COROUTINES
3242		3613
3243	=head2 THREADS	3614	=head3 THREADS
3244		3615
3245	Libev itself is thread-safe (unless the opposite is specifically	3616	All libev functions are reentrant and thread-safe unless explicitly
3246	documented for a function), but it uses no locking itself. This means that	3617	documented otherwise, but libev implements no locking itself. This means
3247	you can use as many loops as you want in parallel, as long as only one	3618	that you can use as many loops as you want in parallel, as long as there
3248	thread ever calls into one libev function with the same loop parameter:	3619	are no concurrent calls into any libev function with the same loop
		3620	parameter (C<ev_default_*> calls have an implicit default loop parameter,
3249	libev guarentees that different event loops share no data structures that	3621	of course): libev guarantees that different event loops share no data
3250	need locking.	3622	structures that need any locking.
3251		3623
3252	Or to put it differently: calls with different loop parameters can be done	3624	Or to put it differently: calls with different loop parameters can be done
3253	concurrently from multiple threads, calls with the same loop parameter	3625	concurrently from multiple threads, calls with the same loop parameter
3254	must be done serially (but can be done from different threads, as long as	3626	must be done serially (but can be done from different threads, as long as
3255	only one thread ever is inside a call at any point in time, e.g. by using	3627	only one thread ever is inside a call at any point in time, e.g. by using
3256	a mutex per loop).	3628	a mutex per loop).
3257		3629
3258	Specifically to support threads (and signal handlers), libev implements	3630	Specifically to support threads (and signal handlers), libev implements
3259	so-called C<ev_async> watchers, which allow some limited form of	3631	so-called C<ev_async> watchers, which allow some limited form of
3260	concurrency on the same event loop.	3632	concurrency on the same event loop, namely waking it up "from the
		3633	outside".
3261		3634
3262	If you want to know which design (one loop, locking, or multiple loops	3635	If you want to know which design (one loop, locking, or multiple loops
3263	without or something else still) is best for your problem, then I cannot	3636	without or something else still) is best for your problem, then I cannot
3264	help you. I can give some generic advice however:	3637	help you, but here is some generic advice:
3265		3638
3266	=over 4	3639	=over 4
3267		3640
3268	=item * most applications have a main thread: use the default libev loop	3641	=item * most applications have a main thread: use the default libev loop
3269	in that thread, or create a separate thread running only the default loop.	3642	in that thread, or create a separate thread running only the default loop.
…		…
3281		3654
3282	Choosing a model is hard - look around, learn, know that usually you can do	3655	Choosing a model is hard - look around, learn, know that usually you can do
3283	better than you currently do :-)	3656	better than you currently do :-)
3284		3657
3285	=item * often you need to talk to some other thread which blocks in the	3658	=item * often you need to talk to some other thread which blocks in the
		3659	event loop.
		3660
3286	event loop - C<ev_async> watchers can be used to wake them up from other	3661	C<ev_async> watchers can be used to wake them up from other threads safely
3287	threads safely (or from signal contexts...).	3662	(or from signal contexts...).
3288		3663
3289	=item * some watcher types are only supported in the default loop - use	3664	An example use would be to communicate signals or other events that only
3290	C<ev_async> watchers to tell your other loops about any such events.	3665	work in the default loop by registering the signal watcher with the
		3666	default loop and triggering an C<ev_async> watcher from the default loop
		3667	watcher callback into the event loop interested in the signal.
3291		3668
3292	=back	3669	=back
3293		3670
3294	=head2 COROUTINES	3671	=head3 COROUTINES
3295		3672
3296	Libev is much more accommodating to coroutines ("cooperative threads"):	3673	Libev is very accommodating to coroutines ("cooperative threads"):
3297	libev fully supports nesting calls to it's functions from different	3674	libev fully supports nesting calls to its functions from different
3298	coroutines (e.g. you can call C<ev_loop> on the same loop from two	3675	coroutines (e.g. you can call C<ev_loop> on the same loop from two
3299	different coroutines and switch freely between both coroutines running the	3676	different coroutines, and switch freely between both coroutines running the
3300	loop, as long as you don't confuse yourself). The only exception is that	3677	loop, as long as you don't confuse yourself). The only exception is that
3301	you must not do this from C<ev_periodic> reschedule callbacks.	3678	you must not do this from C<ev_periodic> reschedule callbacks.
3302		3679
3303	Care has been invested into making sure that libev does not keep local	3680	Care has been taken to ensure that libev does not keep local state inside
3304	state inside C<ev_loop>, and other calls do not usually allow coroutine	3681	C<ev_loop>, and other calls do not usually allow for coroutine switches as
3305	switches.	3682	they do not call any callbacks.
3306		3683
		3684	=head2 COMPILER WARNINGS
3307		3685
3308	=head1 COMPLEXITIES	3686	Depending on your compiler and compiler settings, you might get no or a
		3687	lot of warnings when compiling libev code. Some people are apparently
		3688	scared by this.
3309		3689
3310	In this section the complexities of (many of) the algorithms used inside	3690	However, these are unavoidable for many reasons. For one, each compiler
3311	libev will be explained. For complexity discussions about backends see the	3691	has different warnings, and each user has different tastes regarding
3312	documentation for C<ev_default_init>.	3692	warning options. "Warn-free" code therefore cannot be a goal except when
		3693	targeting a specific compiler and compiler-version.
3313		3694
3314	All of the following are about amortised time: If an array needs to be	3695	Another reason is that some compiler warnings require elaborate
3315	extended, libev needs to realloc and move the whole array, but this	3696	workarounds, or other changes to the code that make it less clear and less
3316	happens asymptotically never with higher number of elements, so O(1) might	3697	maintainable.
3317	mean it might do a lengthy realloc operation in rare cases, but on average
3318	it is much faster and asymptotically approaches constant time.
3319		3698
3320	=over 4	3699	And of course, some compiler warnings are just plain stupid, or simply
		3700	wrong (because they don't actually warn about the condition their message
		3701	seems to warn about). For example, certain older gcc versions had some
		3702	warnings that resulted an extreme number of false positives. These have
		3703	been fixed, but some people still insist on making code warn-free with
		3704	such buggy versions.
3321		3705
3322	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)	3706	While libev is written to generate as few warnings as possible,
		3707	"warn-free" code is not a goal, and it is recommended not to build libev
		3708	with any compiler warnings enabled unless you are prepared to cope with
		3709	them (e.g. by ignoring them). Remember that warnings are just that:
		3710	warnings, not errors, or proof of bugs.
3323		3711
3324	This means that, when you have a watcher that triggers in one hour and
3325	there are 100 watchers that would trigger before that then inserting will
3326	have to skip roughly seven (C<ld 100>) of these watchers.
3327		3712
3328	=item Changing timer/periodic watchers (by autorepeat or calling again): O(log skipped_other_timers)	3713	=head2 VALGRIND
3329		3714
3330	That means that changing a timer costs less than removing/adding them	3715	Valgrind has a special section here because it is a popular tool that is
3331	as only the relative motion in the event queue has to be paid for.	3716	highly useful. Unfortunately, valgrind reports are very hard to interpret.
3332		3717
3333	=item Starting io/check/prepare/idle/signal/child/fork/async watchers: O(1)	3718	If you think you found a bug (memory leak, uninitialised data access etc.)
		3719	in libev, then check twice: If valgrind reports something like:
3334		3720
3335	These just add the watcher into an array or at the head of a list.	3721	==2274== definitely lost: 0 bytes in 0 blocks.
		3722	==2274== possibly lost: 0 bytes in 0 blocks.
		3723	==2274== still reachable: 256 bytes in 1 blocks.
3336		3724
3337	=item Stopping check/prepare/idle/fork/async watchers: O(1)	3725	Then there is no memory leak, just as memory accounted to global variables
		3726	is not a memleak - the memory is still being referenced, and didn't leak.
3338		3727
3339	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))	3728	Similarly, under some circumstances, valgrind might report kernel bugs
		3729	as if it were a bug in libev (e.g. in realloc or in the poll backend,
		3730	although an acceptable workaround has been found here), or it might be
		3731	confused.
3340		3732
3341	These watchers are stored in lists then need to be walked to find the	3733	Keep in mind that valgrind is a very good tool, but only a tool. Don't
3342	correct watcher to remove. The lists are usually short (you don't usually	3734	make it into some kind of religion.
3343	have many watchers waiting for the same fd or signal).
3344		3735
3345	=item Finding the next timer in each loop iteration: O(1)	3736	If you are unsure about something, feel free to contact the mailing list
		3737	with the full valgrind report and an explanation on why you think this
		3738	is a bug in libev (best check the archives, too :). However, don't be
		3739	annoyed when you get a brisk "this is no bug" answer and take the chance
		3740	of learning how to interpret valgrind properly.
3346		3741
3347	By virtue of using a binary or 4-heap, the next timer is always found at a	3742	If you need, for some reason, empty reports from valgrind for your project
3348	fixed position in the storage array.	3743	I suggest using suppression lists.
3349		3744
3350	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)
3351		3745
3352	A change means an I/O watcher gets started or stopped, which requires	3746	=head1 PORTABILITY NOTES
3353	libev to recalculate its status (and possibly tell the kernel, depending
3354	on backend and whether C<ev_io_set> was used).
3355		3747
3356	=item Activating one watcher (putting it into the pending state): O(1)
3357
3358	=item Priority handling: O(number_of_priorities)
3359
3360	Priorities are implemented by allocating some space for each
3361	priority. When doing priority-based operations, libev usually has to
3362	linearly search all the priorities, but starting/stopping and activating
3363	watchers becomes O(1) w.r.t. priority handling.
3364
3365	=item Sending an ev_async: O(1)
3366
3367	=item Processing ev_async_send: O(number_of_async_watchers)
3368
3369	=item Processing signals: O(max_signal_number)
3370
3371	Sending involves a system call I<iff> there were no other C<ev_async_send>
3372	calls in the current loop iteration. Checking for async and signal events
3373	involves iterating over all running async watchers or all signal numbers.
3374
3375	=back
3376
3377
3378	=head1 WIN32 PLATFORM LIMITATIONS AND WORKAROUNDS	3748	=head2 WIN32 PLATFORM LIMITATIONS AND WORKAROUNDS
3379		3749
3380	Win32 doesn't support any of the standards (e.g. POSIX) that libev	3750	Win32 doesn't support any of the standards (e.g. POSIX) that libev
3381	requires, and its I/O model is fundamentally incompatible with the POSIX	3751	requires, and its I/O model is fundamentally incompatible with the POSIX
3382	model. Libev still offers limited functionality on this platform in	3752	model. Libev still offers limited functionality on this platform in
3383	the form of the C<EVBACKEND_SELECT> backend, and only supports socket	3753	the form of the C<EVBACKEND_SELECT> backend, and only supports socket
…		…
3394		3764
3395	Not a libev limitation but worth mentioning: windows apparently doesn't	3765	Not a libev limitation but worth mentioning: windows apparently doesn't
3396	accept large writes: instead of resulting in a partial write, windows will	3766	accept large writes: instead of resulting in a partial write, windows will
3397	either accept everything or return C<ENOBUFS> if the buffer is too large,	3767	either accept everything or return C<ENOBUFS> if the buffer is too large,
3398	so make sure you only write small amounts into your sockets (less than a	3768	so make sure you only write small amounts into your sockets (less than a
3399	megabyte seems safe, but thsi apparently depends on the amount of memory	3769	megabyte seems safe, but this apparently depends on the amount of memory
3400	available).	3770	available).
3401		3771
3402	Due to the many, low, and arbitrary limits on the win32 platform and	3772	Due to the many, low, and arbitrary limits on the win32 platform and
3403	the abysmal performance of winsockets, using a large number of sockets	3773	the abysmal performance of winsockets, using a large number of sockets
3404	is not recommended (and not reasonable). If your program needs to use	3774	is not recommended (and not reasonable). If your program needs to use
…		…
3415	#define EV_SELECT_IS_WINSOCKET 1 /* configure libev for windows select */	3785	#define EV_SELECT_IS_WINSOCKET 1 /* configure libev for windows select */
3416		3786
3417	#include "ev.h"	3787	#include "ev.h"
3418		3788
3419	And compile the following F<evwrap.c> file into your project (make sure	3789	And compile the following F<evwrap.c> file into your project (make sure
3420	you do I<not> compile the F<ev.c> or any other embedded soruce files!):	3790	you do I<not> compile the F<ev.c> or any other embedded source files!):
3421		3791
3422	#include "evwrap.h"	3792	#include "evwrap.h"
3423	#include "ev.c"	3793	#include "ev.c"
3424		3794
3425	=over 4	3795	=over 4
…		…
3470	wrap all I/O functions and provide your own fd management, but the cost of	3840	wrap all I/O functions and provide your own fd management, but the cost of
3471	calling select (O(n²)) will likely make this unworkable.	3841	calling select (O(n²)) will likely make this unworkable.
3472		3842
3473	=back	3843	=back
3474		3844
3475
3476	=head1 PORTABILITY REQUIREMENTS	3845	=head2 PORTABILITY REQUIREMENTS
3477		3846
3478	In addition to a working ISO-C implementation, libev relies on a few	3847	In addition to a working ISO-C implementation and of course the
3479	additional extensions:	3848	backend-specific APIs, libev relies on a few additional extensions:
3480		3849
3481	=over 4	3850	=over 4
3482		3851
3483	=item C<void (*)(ev_watcher_type *, int revents)> must have compatible	3852	=item C<void (*)(ev_watcher_type *, int revents)> must have compatible
3484	calling conventions regardless of C<ev_watcher_type *>.	3853	calling conventions regardless of C<ev_watcher_type *>.
…		…
3490	calls them using an C<ev_watcher *> internally.	3859	calls them using an C<ev_watcher *> internally.
3491		3860
3492	=item C<sig_atomic_t volatile> must be thread-atomic as well	3861	=item C<sig_atomic_t volatile> must be thread-atomic as well
3493		3862
3494	The type C<sig_atomic_t volatile> (or whatever is defined as	3863	The type C<sig_atomic_t volatile> (or whatever is defined as
3495	C<EV_ATOMIC_T>) must be atomic w.r.t. accesses from different	3864	C<EV_ATOMIC_T>) must be atomic with respect to accesses from different
3496	threads. This is not part of the specification for C<sig_atomic_t>, but is	3865	threads. This is not part of the specification for C<sig_atomic_t>, but is
3497	believed to be sufficiently portable.	3866	believed to be sufficiently portable.
3498		3867
3499	=item C<sigprocmask> must work in a threaded environment	3868	=item C<sigprocmask> must work in a threaded environment
3500		3869
…		…
3509	except the initial one, and run the default loop in the initial thread as	3878	except the initial one, and run the default loop in the initial thread as
3510	well.	3879	well.
3511		3880
3512	=item C<long> must be large enough for common memory allocation sizes	3881	=item C<long> must be large enough for common memory allocation sizes
3513		3882
3514	To improve portability and simplify using libev, libev uses C<long>	3883	To improve portability and simplify its API, libev uses C<long> internally
3515	internally instead of C<size_t> when allocating its data structures. On	3884	instead of C<size_t> when allocating its data structures. On non-POSIX
3516	non-POSIX systems (Microsoft...) this might be unexpectedly low, but	3885	systems (Microsoft...) this might be unexpectedly low, but is still at
3517	is still at least 31 bits everywhere, which is enough for hundreds of	3886	least 31 bits everywhere, which is enough for hundreds of millions of
3518	millions of watchers.	3887	watchers.
3519		3888
3520	=item C<double> must hold a time value in seconds with enough accuracy	3889	=item C<double> must hold a time value in seconds with enough accuracy
3521		3890
3522	The type C<double> is used to represent timestamps. It is required to	3891	The type C<double> is used to represent timestamps. It is required to
3523	have at least 51 bits of mantissa (and 9 bits of exponent), which is good	3892	have at least 51 bits of mantissa (and 9 bits of exponent), which is good
…		…
3527	=back	3896	=back
3528		3897
3529	If you know of other additional requirements drop me a note.	3898	If you know of other additional requirements drop me a note.
3530		3899
3531		3900
3532	=head1 COMPILER WARNINGS	3901	=head1 ALGORITHMIC COMPLEXITIES
3533		3902
3534	Depending on your compiler and compiler settings, you might get no or a	3903	In this section the complexities of (many of) the algorithms used inside
3535	lot of warnings when compiling libev code. Some people are apparently	3904	libev will be documented. For complexity discussions about backends see
3536	scared by this.	3905	the documentation for C<ev_default_init>.
3537		3906
3538	However, these are unavoidable for many reasons. For one, each compiler	3907	All of the following are about amortised time: If an array needs to be
3539	has different warnings, and each user has different tastes regarding	3908	extended, libev needs to realloc and move the whole array, but this
3540	warning options. "Warn-free" code therefore cannot be a goal except when	3909	happens asymptotically rarer with higher number of elements, so O(1) might
3541	targeting a specific compiler and compiler-version.	3910	mean that libev does a lengthy realloc operation in rare cases, but on
		3911	average it is much faster and asymptotically approaches constant time.
3542		3912
3543	Another reason is that some compiler warnings require elaborate	3913	=over 4
3544	workarounds, or other changes to the code that make it less clear and less
3545	maintainable.
3546		3914
3547	And of course, some compiler warnings are just plain stupid, or simply	3915	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)
3548	wrong (because they don't actually warn about the condition their message
3549	seems to warn about).
3550		3916
3551	While libev is written to generate as few warnings as possible,	3917	This means that, when you have a watcher that triggers in one hour and
3552	"warn-free" code is not a goal, and it is recommended not to build libev	3918	there are 100 watchers that would trigger before that, then inserting will
3553	with any compiler warnings enabled unless you are prepared to cope with	3919	have to skip roughly seven (C<ld 100>) of these watchers.
3554	them (e.g. by ignoring them). Remember that warnings are just that:
3555	warnings, not errors, or proof of bugs.
3556		3920
		3921	=item Changing timer/periodic watchers (by autorepeat or calling again): O(log skipped_other_timers)
3557		3922
3558	=head1 VALGRIND	3923	That means that changing a timer costs less than removing/adding them,
		3924	as only the relative motion in the event queue has to be paid for.
3559		3925
3560	Valgrind has a special section here because it is a popular tool that is	3926	=item Starting io/check/prepare/idle/signal/child/fork/async watchers: O(1)
3561	highly useful, but valgrind reports are very hard to interpret.
3562		3927
3563	If you think you found a bug (memory leak, uninitialised data access etc.)	3928	These just add the watcher into an array or at the head of a list.
3564	in libev, then check twice: If valgrind reports something like:
3565		3929
3566	==2274== definitely lost: 0 bytes in 0 blocks.	3930	=item Stopping check/prepare/idle/fork/async watchers: O(1)
3567	==2274== possibly lost: 0 bytes in 0 blocks.
3568	==2274== still reachable: 256 bytes in 1 blocks.
3569		3931
3570	Then there is no memory leak. Similarly, under some circumstances,	3932	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))
3571	valgrind might report kernel bugs as if it were a bug in libev, or it
3572	might be confused (it is a very good tool, but only a tool).
3573		3933
3574	If you are unsure about something, feel free to contact the mailing list	3934	These watchers are stored in lists, so they need to be walked to find the
3575	with the full valgrind report and an explanation on why you think this is	3935	correct watcher to remove. The lists are usually short (you don't usually
3576	a bug in libev. However, don't be annoyed when you get a brisk "this is	3936	have many watchers waiting for the same fd or signal: one is typical, two
3577	no bug" answer and take the chance of learning how to interpret valgrind	3937	is rare).
3578	properly.
3579		3938
3580	If you need, for some reason, empty reports from valgrind for your project	3939	=item Finding the next timer in each loop iteration: O(1)
3581	I suggest using suppression lists.	3940
		3941	By virtue of using a binary or 4-heap, the next timer is always found at a
		3942	fixed position in the storage array.
		3943
		3944	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)
		3945
		3946	A change means an I/O watcher gets started or stopped, which requires
		3947	libev to recalculate its status (and possibly tell the kernel, depending
		3948	on backend and whether C<ev_io_set> was used).
		3949
		3950	=item Activating one watcher (putting it into the pending state): O(1)
		3951
		3952	=item Priority handling: O(number_of_priorities)
		3953
		3954	Priorities are implemented by allocating some space for each
		3955	priority. When doing priority-based operations, libev usually has to
		3956	linearly search all the priorities, but starting/stopping and activating
		3957	watchers becomes O(1) with respect to priority handling.
		3958
		3959	=item Sending an ev_async: O(1)
		3960
		3961	=item Processing ev_async_send: O(number_of_async_watchers)
		3962
		3963	=item Processing signals: O(max_signal_number)
		3964
		3965	Sending involves a system call I<iff> there were no other C<ev_async_send>
		3966	calls in the current loop iteration. Checking for async and signal events
		3967	involves iterating over all running async watchers or all signal numbers.
		3968
		3969	=back
3582		3970
3583		3971
3584	=head1 AUTHOR	3972	=head1 AUTHOR
3585		3973
3586	Marc Lehmann <libev@schmorp.de>.	3974	Marc Lehmann <libev@schmorp.de>, with repeated corrections by Mikael Magnusson.
3587		3975

Diff Legend

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