[ViewVC] Diff of: cvs/libev/ev.pod

Comparing libev/ev.pod (file contents):
Revision 1.179 by root, Sat Sep 13 19:14:21 2008 UTC vs.
Revision 1.194 by root, Mon Oct 20 16:08:36 2008 UTC

…		…
214	C<ev_embeddable_backends () & ev_supported_backends ()>, likewise for	214	C<ev_embeddable_backends () & ev_supported_backends ()>, likewise for
215	recommended ones.	215	recommended ones.
216		216
217	See the description of C<ev_embed> watchers for more info.	217	See the description of C<ev_embed> watchers for more info.
218		218
219	=item ev_set_allocator (void (cb)(void *ptr, long size))	219	=item ev_set_allocator (void (cb)(void *ptr, long size)) [NOT REENTRANT]
220		220
221	Sets the allocation function to use (the prototype is similar - the	221	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	222	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	223	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	224	when memory needs to be allocated (C<size != 0>), the library might abort
…		…
250	}	250	}
251		251
252	...	252	...
253	ev_set_allocator (persistent_realloc);	253	ev_set_allocator (persistent_realloc);
254		254
255	=item ev_set_syserr_cb (void (cb)(const char msg));	255	=item ev_set_syserr_cb (void (cb)(const char msg)); [NOT REENTRANT]
256		256
257	Set the callback function to call on a retryable system call error (such	257	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	258	as failed select, poll, epoll_wait). The message is a printable string
259	indicating the system call or subsystem causing the problem. If this	259	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	260	callback is set, then libev will expect it to remedy the situation, no
…		…
396	Please note that epoll sometimes generates spurious notifications, so you	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	397	need to use non-blocking I/O or other means to avoid blocking when no data
398	(or space) is available.	398	(or space) is available.
399		399
400	Best performance from this backend is achieved by not unregistering all	400	Best performance from this backend is achieved by not unregistering all
401	watchers for a file descriptor until it has been closed, if possible, i.e.	401	watchers for a file descriptor until it has been closed, if possible,
402	keep at least one watcher active per fd at all times.	402	i.e. keep at least one watcher active per fd at all times. Stopping and
		403	starting a watcher (without re-setting it) also usually doesn't cause
		404	extra overhead.
403		405
404	While nominally embeddable in other event loops, this feature is broken in	406	While nominally embeddable in other event loops, this feature is broken in
405	all kernel versions tested so far.	407	all kernel versions tested so far.
406		408
407	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as	409	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
408	C<EVBACKEND_POLL>.	410	C<EVBACKEND_POLL>.
409		411
410	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)	412	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)
411		413
412	Kqueue deserves special mention, as at the time of this writing, it	414	Kqueue deserves special mention, as at the time of this writing, it was
413	was broken on all BSDs except NetBSD (usually it doesn't work reliably	415	broken on all BSDs except NetBSD (usually it doesn't work reliably with
414	with anything but sockets and pipes, except on Darwin, where of course	416	anything but sockets and pipes, except on Darwin, where of course it's
415	it's completely useless). For this reason it's not being "auto-detected"	417	completely useless). For this reason it's not being "auto-detected" unless
416	unless you explicitly specify it explicitly in the flags (i.e. using	418	you explicitly specify it in the flags (i.e. using C<EVBACKEND_KQUEUE>) or
417	C<EVBACKEND_KQUEUE>) or libev was compiled on a known-to-be-good (-enough)	419	libev was compiled on a known-to-be-good (-enough) system like NetBSD.
418	system like NetBSD.
419		420
420	You still can embed kqueue into a normal poll or select backend and use it	421	You still can embed kqueue into a normal poll or select backend and use it
421	only for sockets (after having made sure that sockets work with kqueue on	422	only for sockets (after having made sure that sockets work with kqueue on
422	the target platform). See C<ev_embed> watchers for more info.	423	the target platform). See C<ev_embed> watchers for more info.
423		424
424	It scales in the same way as the epoll backend, but the interface to the	425	It scales in the same way as the epoll backend, but the interface to the
425	kernel is more efficient (which says nothing about its actual speed, of	426	kernel is more efficient (which says nothing about its actual speed, of
426	course). While stopping, setting and starting an I/O watcher does never	427	course). While stopping, setting and starting an I/O watcher does never
427	cause an extra system call as with C<EVBACKEND_EPOLL>, it still adds up to	428	cause an extra system call as with C<EVBACKEND_EPOLL>, it still adds up to
428	two event changes per incident, support for C<fork ()> is very bad and it	429	two event changes per incident. Support for C<fork ()> is very bad and it
429	drops fds silently in similarly hard-to-detect cases.	430	drops fds silently in similarly hard-to-detect cases.
430		431
431	This backend usually performs well under most conditions.	432	This backend usually performs well under most conditions.
432		433
433	While nominally embeddable in other event loops, this doesn't work	434	While nominally embeddable in other event loops, this doesn't work
434	everywhere, so you might need to test for this. And since it is broken	435	everywhere, so you might need to test for this. And since it is broken
435	almost everywhere, you should only use it when you have a lot of sockets	436	almost everywhere, you should only use it when you have a lot of sockets
436	(for which it usually works), by embedding it into another event loop	437	(for which it usually works), by embedding it into another event loop
437	(e.g. C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>) and using it only for	438	(e.g. C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>) and, did I mention it,
438	sockets.	439	using it only for sockets.
439		440
440	This backend maps C<EV_READ> into an C<EVFILT_READ> kevent with	441	This backend maps C<EV_READ> into an C<EVFILT_READ> kevent with
441	C<NOTE_EOF>, and C<EV_WRITE> into an C<EVFILT_WRITE> kevent with	442	C<NOTE_EOF>, and C<EV_WRITE> into an C<EVFILT_WRITE> kevent with
442	C<NOTE_EOF>.	443	C<NOTE_EOF>.
443		444
…		…
460	While this backend scales well, it requires one system call per active	461	While this backend scales well, it requires one system call per active
461	file descriptor per loop iteration. For small and medium numbers of file	462	file descriptor per loop iteration. For small and medium numbers of file
462	descriptors a "slow" C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> backend	463	descriptors a "slow" C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> backend
463	might perform better.	464	might perform better.
464		465
465	On the positive side, ignoring the spurious readiness notifications, this	466	On the positive side, with the exception of the spurious readiness
466	backend actually performed to specification in all tests and is fully	467	notifications, this backend actually performed fully to specification
467	embeddable, which is a rare feat among the OS-specific backends.	468	in all tests and is fully embeddable, which is a rare feat among the
		469	OS-specific backends.
468		470
469	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as	471	This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
470	C<EVBACKEND_POLL>.	472	C<EVBACKEND_POLL>.
471		473
472	=item C<EVBACKEND_ALL>	474	=item C<EVBACKEND_ALL>
…		…
481		483
482	If one or more of these are or'ed into the flags value, then only these	484	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	485	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.	486	specified, all backends in C<ev_recommended_backends ()> will be tried.
485		487
486	The most typical usage is like this:	488	Example: This is the most typical usage.
487		489
488	if (!ev_default_loop (0))	490	if (!ev_default_loop (0))
489	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");	491	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
490		492
491	Restrict libev to the select and poll backends, and do not allow	493	Example: Restrict libev to the select and poll backends, and do not allow
492	environment settings to be taken into account:	494	environment settings to be taken into account:
493		495
494	ev_default_loop (EVBACKEND_POLL \| EVBACKEND_SELECT \| EVFLAG_NOENV);	496	ev_default_loop (EVBACKEND_POLL \| EVBACKEND_SELECT \| EVFLAG_NOENV);
495		497
496	Use whatever libev has to offer, but make sure that kqueue is used if	498	Example: Use whatever libev has to offer, but make sure that kqueue is
497	available (warning, breaks stuff, best use only with your own private	499	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):	500	private event loop and only if you know the OS supports your types of
		501	fds):
499		502
500	ev_default_loop (ev_recommended_backends () \| EVBACKEND_KQUEUE);	503	ev_default_loop (ev_recommended_backends () \| EVBACKEND_KQUEUE);
501		504
502	=item struct ev_loop *ev_loop_new (unsigned int flags)	505	=item struct ev_loop *ev_loop_new (unsigned int flags)
503		506
…		…
561		564
562	=item ev_loop_fork (loop)	565	=item ev_loop_fork (loop)
563		566
564	Like C<ev_default_fork>, but acts on an event loop created by	567	Like C<ev_default_fork>, but acts on an event loop created by
565	C<ev_loop_new>. Yes, you have to call this on every allocated event loop	568	C<ev_loop_new>. Yes, you have to call this on every allocated event loop
566	after fork, and how you do this is entirely your own problem.	569	after fork that you want to re-use in the child, and how you do this is
		570	entirely your own problem.
567		571
568	=item int ev_is_default_loop (loop)	572	=item int ev_is_default_loop (loop)
569		573
570	Returns true when the given loop actually is the default loop, false otherwise.	574	Returns true when the given loop is, in fact, the default loop, and false
		575	otherwise.
571		576
572	=item unsigned int ev_loop_count (loop)	577	=item unsigned int ev_loop_count (loop)
573		578
574	Returns the count of loop iterations for the loop, which is identical to	579	Returns the count of loop iterations for the loop, which is identical to
575	the number of times libev did poll for new events. It starts at C<0> and	580	the number of times libev did poll for new events. It starts at C<0> and
…		…
613	If the flags argument is specified as C<0>, it will not return until	618	If the flags argument is specified as C<0>, it will not return until
614	either no event watchers are active anymore or C<ev_unloop> was called.	619	either no event watchers are active anymore or C<ev_unloop> was called.
615		620
616	Please note that an explicit C<ev_unloop> is usually better than	621	Please note that an explicit C<ev_unloop> is usually better than
617	relying on all watchers to be stopped when deciding when a program has	622	relying on all watchers to be stopped when deciding when a program has
618	finished (especially in interactive programs), but having a program that	623	finished (especially in interactive programs), but having a program
619	automatically loops as long as it has to and no longer by virtue of	624	that automatically loops as long as it has to and no longer by virtue
620	relying on its watchers stopping correctly is a thing of beauty.	625	of relying on its watchers stopping correctly, that is truly a thing of
		626	beauty.
621		627
622	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle	628	A flags value of C<EVLOOP_NONBLOCK> will look for new events, will handle
623	those events and any outstanding ones, but will not block your process in	629	those events and any already outstanding ones, but will not block your
624	case there are no events and will return after one iteration of the loop.	630	process in case there are no events and will return after one iteration of
		631	the loop.
625		632
626	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if	633	A flags value of C<EVLOOP_ONESHOT> will look for new events (waiting if
627	necessary) and will handle those and any outstanding ones. It will block	634	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	635	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	636	be an event internal to libev itself, so there is no guarentee that a
630	external event in conjunction with something not expressible using other	637	user-registered callback will be called), and will return after one
		638	iteration of the loop.
		639
		640	This is useful if you are waiting for some external event in conjunction
		641	with something not expressible using other libev watchers (i.e. "roll your
631	libev watchers. However, a pair of C<ev_prepare>/C<ev_check> watchers is	642	own C<ev_loop>"). However, a pair of C<ev_prepare>/C<ev_check> watchers is
632	usually a better approach for this kind of thing.	643	usually a better approach for this kind of thing.
633		644
634	Here are the gory details of what C<ev_loop> does:	645	Here are the gory details of what C<ev_loop> does:
635		646
636	- Before the first iteration, call any pending watchers.	647	- Before the first iteration, call any pending watchers.
…		…
646	any active watchers at all will result in not sleeping).	657	any active watchers at all will result in not sleeping).
647	- Sleep if the I/O and timer collect interval say so.	658	- Sleep if the I/O and timer collect interval say so.
648	- Block the process, waiting for any events.	659	- Block the process, waiting for any events.
649	- Queue all outstanding I/O (fd) events.	660	- Queue all outstanding I/O (fd) events.
650	- Update the "event loop time" (ev_now ()), and do time jump adjustments.	661	- Update the "event loop time" (ev_now ()), and do time jump adjustments.
651	- Queue all outstanding timers.	662	- Queue all expired timers.
652	- Queue all outstanding periodics.	663	- Queue all expired periodics.
653	- Unless any events are pending now, queue all idle watchers.	664	- Unless any events are pending now, queue all idle watchers.
654	- Queue all check watchers.	665	- Queue all check watchers.
655	- Call all queued watchers in reverse order (i.e. check watchers first).	666	- Call all queued watchers in reverse order (i.e. check watchers first).
656	Signals and child watchers are implemented as I/O watchers, and will	667	Signals and child watchers are implemented as I/O watchers, and will
657	be handled here by queueing them when their watcher gets executed.	668	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	685	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.	686	C<EVUNLOOP_ALL>, which will make all nested C<ev_loop> calls return.
676		687
677	This "unloop state" will be cleared when entering C<ev_loop> again.	688	This "unloop state" will be cleared when entering C<ev_loop> again.
678		689
		690	It is safe to call C<ev_unloop> from otuside any C<ev_loop> calls.
		691
679	=item ev_ref (loop)	692	=item ev_ref (loop)
680		693
681	=item ev_unref (loop)	694	=item ev_unref (loop)
682		695
683	Ref/unref can be used to add or remove a reference count on the event	696	Ref/unref can be used to add or remove a reference count on the event
684	loop: Every watcher keeps one reference, and as long as the reference	697	loop: Every watcher keeps one reference, and as long as the reference
685	count is nonzero, C<ev_loop> will not return on its own. If you have	698	count is nonzero, C<ev_loop> will not return on its own.
		699
686	a watcher you never unregister that should not keep C<ev_loop> from	700	If you have a watcher you never unregister that should not keep C<ev_loop>
687	returning, ev_unref() after starting, and ev_ref() before stopping it. For	701	from returning, call ev_unref() after starting, and ev_ref() before
		702	stopping it.
		703
688	example, libev itself uses this for its internal signal pipe: It is not	704	As an example, libev itself uses this for its internal signal pipe: It is
689	visible to the libev user and should not keep C<ev_loop> from exiting if	705	not visible to the libev user and should not keep C<ev_loop> from exiting
690	no event watchers registered by it are active. It is also an excellent	706	if no event watchers registered by it are active. It is also an excellent
691	way to do this for generic recurring timers or from within third-party	707	way to do this for generic recurring timers or from within third-party
692	libraries. Just remember to I<unref after start> and I<ref before stop>	708	libraries. Just remember to I<unref after start> and I<ref before stop>
693	(but only if the watcher wasn't active before, or was active before,	709	(but only if the watcher wasn't active before, or was active before,
694	respectively).	710	respectively).
695		711
…		…
718	Setting these to a higher value (the C<interval> I<must> be >= C<0>)	734	Setting these to a higher value (the C<interval> I<must> be >= C<0>)
719	allows libev to delay invocation of I/O and timer/periodic callbacks	735	allows libev to delay invocation of I/O and timer/periodic callbacks
720	to increase efficiency of loop iterations (or to increase power-saving	736	to increase efficiency of loop iterations (or to increase power-saving
721	opportunities).	737	opportunities).
722		738
723	The background is that sometimes your program runs just fast enough to	739	The idea is that sometimes your program runs just fast enough to handle
724	handle one (or very few) event(s) per loop iteration. While this makes	740	one (or very few) event(s) per loop iteration. While this makes the
725	the program responsive, it also wastes a lot of CPU time to poll for new	741	program responsive, it also wastes a lot of CPU time to poll for new
726	events, especially with backends like C<select ()> which have a high	742	events, especially with backends like C<select ()> which have a high
727	overhead for the actual polling but can deliver many events at once.	743	overhead for the actual polling but can deliver many events at once.
728		744
729	By setting a higher I<io collect interval> you allow libev to spend more	745	By setting a higher I<io collect interval> you allow libev to spend more
730	time collecting I/O events, so you can handle more events per iteration,	746	time collecting I/O events, so you can handle more events per iteration,
…		…
732	C<ev_timer>) will be not affected. Setting this to a non-null value will	748	C<ev_timer>) will be not affected. Setting this to a non-null value will
733	introduce an additional C<ev_sleep ()> call into most loop iterations.	749	introduce an additional C<ev_sleep ()> call into most loop iterations.
734		750
735	Likewise, by setting a higher I<timeout collect interval> you allow libev	751	Likewise, by setting a higher I<timeout collect interval> you allow libev
736	to spend more time collecting timeouts, at the expense of increased	752	to spend more time collecting timeouts, at the expense of increased
737	latency (the watcher callback will be called later). C<ev_io> watchers	753	latency/jitter/inexactness (the watcher callback will be called
738	will not be affected. Setting this to a non-null value will not introduce	754	later). C<ev_io> watchers will not be affected. Setting this to a non-null
739	any overhead in libev.	755	value will not introduce any overhead in libev.
740		756
741	Many (busy) programs can usually benefit by setting the I/O collect	757	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	758	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	759	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>,	760	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.	768	they fire on, say, one-second boundaries only.
753		769
754	=item ev_loop_verify (loop)	770	=item ev_loop_verify (loop)
755		771
756	This function only does something when C<EV_VERIFY> support has been	772	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	773	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	774	through all internal structures and checks them for validity. If anything
759	an error message to standard error and call C<abort ()>.	775	is found to be inconsistent, it will print an error message to standard
		776	error and call C<abort ()>.
760		777
761	This can be used to catch bugs inside libev itself: under normal	778	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	779	circumstances, this function will never abort as of course libev keeps its
763	data structures consistent.	780	data structures consistent.
764		781
…		…
880	happen because the watcher could not be properly started because libev	897	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	898	ran out of memory, a file descriptor was found to be closed or any other
882	problem. You best act on it by reporting the problem and somehow coping	899	problem. You best act on it by reporting the problem and somehow coping
883	with the watcher being stopped.	900	with the watcher being stopped.
884		901
885	Libev will usually signal a few "dummy" events together with an error,	902	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	903	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	904	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	905	the error from read() or write(). This will not work in multi-threaded
889	programs, though, so beware.	906	programs, though, as the fd could already be closed and reused for another
		907	thing, so beware.
890		908
891	=back	909	=back
892		910
893	=head2 GENERIC WATCHER FUNCTIONS	911	=head2 GENERIC WATCHER FUNCTIONS
894		912
…		…
910	(or never started) and there are no pending events outstanding.	928	(or never started) and there are no pending events outstanding.
911		929
912	The callback is always of type C<void ()(ev_loop loop, ev_TYPE *watcher,	930	The callback is always of type C<void ()(ev_loop loop, ev_TYPE *watcher,
913	int revents)>.	931	int revents)>.
914		932
		933	Example: Initialise an C<ev_io> watcher in two steps.
		934
		935	ev_io w;
		936	ev_init (&w, my_cb);
		937	ev_io_set (&w, STDIN_FILENO, EV_READ);
		938
915	=item C<ev_TYPE_set> (ev_TYPE *, [args])	939	=item C<ev_TYPE_set> (ev_TYPE *, [args])
916		940
917	This macro initialises the type-specific parts of a watcher. You need to	941	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	942	call C<ev_init> at least once before you call this macro, but you can
919	call C<ev_TYPE_set> any number of times. You must not, however, call this	943	call C<ev_TYPE_set> any number of times. You must not, however, call this
…		…
921	difference to the C<ev_init> macro).	945	difference to the C<ev_init> macro).
922		946
923	Although some watcher types do not have type-specific arguments	947	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.	948	(e.g. C<ev_prepare>) you still need to call its C<set> macro.
925		949
		950	See C<ev_init>, above, for an example.
		951
926	=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])	952	=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])
927		953
928	This convenience macro rolls both C<ev_init> and C<ev_TYPE_set> macro	954	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	955	calls into a single call. This is the most convenient method to initialise
930	a watcher. The same limitations apply, of course.	956	a watcher. The same limitations apply, of course.
931		957
		958	Example: Initialise and set an C<ev_io> watcher in one step.
		959
		960	ev_io_init (&w, my_cb, STDIN_FILENO, EV_READ);
		961
932	=item C<ev_TYPE_start> (loop , ev_TYPE watcher)	962	=item C<ev_TYPE_start> (loop , ev_TYPE watcher)
933		963
934	Starts (activates) the given watcher. Only active watchers will receive	964	Starts (activates) the given watcher. Only active watchers will receive
935	events. If the watcher is already active nothing will happen.	965	events. If the watcher is already active nothing will happen.
		966
		967	Example: Start the C<ev_io> watcher that is being abused as example in this
		968	whole section.
		969
		970	ev_io_start (EV_DEFAULT_UC, &w);
936		971
937	=item C<ev_TYPE_stop> (loop , ev_TYPE watcher)	972	=item C<ev_TYPE_stop> (loop , ev_TYPE watcher)
938		973
939	Stops the given watcher again (if active) and clears the pending	974	Stops the given watcher again (if active) and clears the pending
940	status. It is possible that stopped watchers are pending (for example,	975	status. It is possible that stopped watchers are pending (for example,
…		…
997		1032
998	=item ev_invoke (loop, ev_TYPE *watcher, int revents)	1033	=item ev_invoke (loop, ev_TYPE *watcher, int revents)
999		1034
1000	Invoke the C<watcher> with the given C<loop> and C<revents>. Neither	1035	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	1036	C<loop> nor C<revents> need to be valid as long as the watcher callback
1002	can deal with that fact.	1037	can deal with that fact, as both are simply passed through to the
		1038	callback.
1003		1039
1004	=item int ev_clear_pending (loop, ev_TYPE *watcher)	1040	=item int ev_clear_pending (loop, ev_TYPE *watcher)
1005		1041
1006	If the watcher is pending, this function returns clears its pending status	1042	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	1043	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>.	1044	watcher isn't pending it does nothing and returns C<0>.
1009		1045
		1046	Sometimes it can be useful to "poll" a watcher instead of waiting for its
		1047	callback to be invoked, which can be accomplished with this function.
		1048
1010	=back	1049	=back
1011		1050
1012		1051
1013	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER	1052	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER
1014		1053
1015	Each watcher has, by default, a member C<void *data> that you can change	1054	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	1055	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	1056	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	1057	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	1058	member, you can also "subclass" the watcher type and provide your own
1020	data:	1059	data:
1021		1060
…		…
1053	ev_timer t2;	1092	ev_timer t2;
1054	}	1093	}
1055		1094
1056	In this case getting the pointer to C<my_biggy> is a bit more	1095	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	1096	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	1097	in the C<data> member of the watcher (for woozies), or you need to use
1059	arithmetic using C<offsetof> inside your watchers:	1098	some pointer arithmetic using C<offsetof> inside your watchers (for real
		1099	programmers):
1060		1100
1061	#include <stddef.h>	1101	#include <stddef.h>
1062		1102
1063	static void	1103	static void
1064	t1_cb (EV_P_ struct ev_timer *w, int revents)	1104	t1_cb (EV_P_ struct ev_timer *w, int revents)
…		…
1104	In general you can register as many read and/or write event watchers per	1144	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	1145	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	1146	descriptors to non-blocking mode is also usually a good idea (but not
1107	required if you know what you are doing).	1147	required if you know what you are doing).
1108		1148
1109	If you must do this, then force the use of a known-to-be-good backend	1149	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	1150	known-to-be-good backend (at the time of this writing, this includes only
1111	C<EVBACKEND_POLL>).	1151	C<EVBACKEND_SELECT> and C<EVBACKEND_POLL>).
1112		1152
1113	Another thing you have to watch out for is that it is quite easy to	1153	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	1154	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	1155	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	1156	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	1157	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	1158	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	1159	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.	1160	C<EAGAIN> is far preferable to a program hanging until some data arrives.
1121		1161
1122	If you cannot run the fd in non-blocking mode (for example you should not	1162	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	1163	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	1164	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	1165	interface such as poll (fortunately in our Xlib example, Xlib already
1126	its own, so its quite safe to use).	1166	does this on its own, so its quite safe to use). Some people additionally
		1167	use C<SIGALRM> and an interval timer, just to be sure you won't block
		1168	indefinitely.
		1169
		1170	But really, best use non-blocking mode.
1127		1171
1128	=head3 The special problem of disappearing file descriptors	1172	=head3 The special problem of disappearing file descriptors
1129		1173
1130	Some backends (e.g. kqueue, epoll) need to be told about closing a file	1174	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,	1175	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	1176	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	1177	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	1178	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	1179	registered with libev, there is no efficient way to see that this is, in
1136	fact, a different file descriptor.	1180	fact, a different file descriptor.
1137		1181
…		…
1168	enable C<EVFLAG_FORKCHECK>, or resort to C<EVBACKEND_SELECT> or	1212	enable C<EVFLAG_FORKCHECK>, or resort to C<EVBACKEND_SELECT> or
1169	C<EVBACKEND_POLL>.	1213	C<EVBACKEND_POLL>.
1170		1214
1171	=head3 The special problem of SIGPIPE	1215	=head3 The special problem of SIGPIPE
1172		1216
1173	While not really specific to libev, it is easy to forget about SIGPIPE:	1217	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	1218	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	1219	sent a SIGPIPE, which, by default, aborts your program. For most programs
1176	this is sensible behaviour, for daemons, this is usually undesirable.	1220	this is sensible behaviour, for daemons, this is usually undesirable.
1177		1221
1178	So when you encounter spurious, unexplained daemon exits, make sure you	1222	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	1223	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).	1224	somewhere, as that would have given you a big clue).
…		…
1187	=item ev_io_init (ev_io *, callback, int fd, int events)	1231	=item ev_io_init (ev_io *, callback, int fd, int events)
1188		1232
1189	=item ev_io_set (ev_io *, int fd, int events)	1233	=item ev_io_set (ev_io *, int fd, int events)
1190		1234
1191	Configures an C<ev_io> watcher. The C<fd> is the file descriptor to	1235	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	1236	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.	1237	C<EV_READ \| EV_WRITE>, to express the desire to receive the given events.
1194		1238
1195	=item int fd [read-only]	1239	=item int fd [read-only]
1196		1240
1197	The file descriptor being watched.	1241	The file descriptor being watched.
1198		1242
…		…
1210		1254
1211	static void	1255	static void
1212	stdin_readable_cb (struct ev_loop loop, struct ev_io w, int revents)	1256	stdin_readable_cb (struct ev_loop loop, struct ev_io w, int revents)
1213	{	1257	{
1214	ev_io_stop (loop, w);	1258	ev_io_stop (loop, w);
1215	.. read from stdin here (or from w->fd) and haqndle any I/O errors	1259	.. read from stdin here (or from w->fd) and handle any I/O errors
1216	}	1260	}
1217		1261
1218	...	1262	...
1219	struct ev_loop *loop = ev_default_init (0);	1263	struct ev_loop *loop = ev_default_init (0);
1220	struct ev_io stdin_readable;	1264	struct ev_io stdin_readable;
…		…
1228	Timer watchers are simple relative timers that generate an event after a	1272	Timer watchers are simple relative timers that generate an event after a
1229	given time, and optionally repeating in regular intervals after that.	1273	given time, and optionally repeating in regular intervals after that.
1230		1274
1231	The timers are based on real time, that is, if you register an event that	1275	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	1276	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	1277	year, it will still time out after (roughly) one hour. "Roughly" because
1234	detecting time jumps is hard, and some inaccuracies are unavoidable (the	1278	detecting time jumps is hard, and some inaccuracies are unavoidable (the
1235	monotonic clock option helps a lot here).	1279	monotonic clock option helps a lot here).
1236		1280
1237	The callback is guaranteed to be invoked only after its timeout has passed,	1281	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	1282	passed, but if multiple timers become ready during the same loop iteration
1239	order of execution is undefined.	1283	then order of execution is undefined.
1240		1284
1241	=head3 The special problem of time updates	1285	=head3 The special problem of time updates
1242		1286
1243	Establishing the current time is a costly operation (it usually takes at	1287	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	1288	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	1289	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	1290	growing difference between C<ev_now ()> and C<ev_time ()> when handling
1247	lots of events.	1291	lots of events in one iteration.
1248		1292
1249	The relative timeouts are calculated relative to the C<ev_now ()>	1293	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	1294	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	1295	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	1296	you suspect event processing to be delayed and you I<need> to base the
…		…
1313	ev_timer_again (loop, timer);	1357	ev_timer_again (loop, timer);
1314		1358
1315	This is more slightly efficient then stopping/starting the timer each time	1359	This is more slightly efficient then stopping/starting the timer each time
1316	you want to modify its timeout value.	1360	you want to modify its timeout value.
1317		1361
		1362	Note, however, that it is often even more efficient to remember the
		1363	time of the last activity and let the timer time-out naturally. In the
		1364	callback, you then check whether the time-out is real, or, if there was
		1365	some activity, you reschedule the watcher to time-out in "last_activity +
		1366	timeout - ev_now ()" seconds.
		1367
1318	=item ev_tstamp repeat [read-write]	1368	=item ev_tstamp repeat [read-write]
1319		1369
1320	The current C<repeat> value. Will be used each time the watcher times out	1370	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),	1371	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.	1372	which is also when any modifications are taken into account.
1323		1373
1324	=back	1374	=back
1325		1375
1326	=head3 Examples	1376	=head3 Examples
…		…
1370	to trigger the event (unlike an C<ev_timer>, which would still trigger	1420	to trigger the event (unlike an C<ev_timer>, which would still trigger
1371	roughly 10 seconds later as it uses a relative timeout).	1421	roughly 10 seconds later as it uses a relative timeout).
1372		1422
1373	C<ev_periodic>s can also be used to implement vastly more complex timers,	1423	C<ev_periodic>s can also be used to implement vastly more complex timers,
1374	such as triggering an event on each "midnight, local time", or other	1424	such as triggering an event on each "midnight, local time", or other
1375	complicated, rules.	1425	complicated rules.
1376		1426
1377	As with timers, the callback is guaranteed to be invoked only when the	1427	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	1428	time (C<at>) has passed, but if multiple periodic timers become ready
1379	during the same loop iteration then order of execution is undefined.	1429	during the same loop iteration, then order of execution is undefined.
1380		1430
1381	=head3 Watcher-Specific Functions and Data Members	1431	=head3 Watcher-Specific Functions and Data Members
1382		1432
1383	=over 4	1433	=over 4
1384		1434
1385	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)	1435	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)
1386		1436
1387	=item ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)	1437	=item ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)
1388		1438
1389	Lots of arguments, lets sort it out... There are basically three modes of	1439	Lots of arguments, lets sort it out... There are basically three modes of
1390	operation, and we will explain them from simplest to complex:	1440	operation, and we will explain them from simplest to most complex:
1391		1441
1392	=over 4	1442	=over 4
1393		1443
1394	=item * absolute timer (at = time, interval = reschedule_cb = 0)	1444	=item * absolute timer (at = time, interval = reschedule_cb = 0)
1395		1445
1396	In this configuration the watcher triggers an event after the wall clock	1446	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	1447	time C<at> has passed. It will not repeat and will not adjust when a time
1398	jump occurs, that is, if it is to be run at January 1st 2011 then it will	1448	jump occurs, that is, if it is to be run at January 1st 2011 then it will
1399	run when the system time reaches or surpasses this time.	1449	only run when the system clock reaches or surpasses this time.
1400		1450
1401	=item * repeating interval timer (at = offset, interval > 0, reschedule_cb = 0)	1451	=item * repeating interval timer (at = offset, interval > 0, reschedule_cb = 0)
1402		1452
1403	In this mode the watcher will always be scheduled to time out at the next	1453	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)	1454	C<at + N * interval> time (for some integer N, which can also be negative)
1405	and then repeat, regardless of any time jumps.	1455	and then repeat, regardless of any time jumps.
1406		1456
1407	This can be used to create timers that do not drift with respect to system	1457	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	1458	system clock, for example, here is a C<ev_periodic> that triggers each
1409	the hour:	1459	hour, on the hour:
1410		1460
1411	ev_periodic_set (&periodic, 0., 3600., 0);	1461	ev_periodic_set (&periodic, 0., 3600., 0);
1412		1462
1413	This doesn't mean there will always be 3600 seconds in between triggers,	1463	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	1464	but only that the callback will be called when the system time shows a
…		…
1501	=back	1551	=back
1502		1552
1503	=head3 Examples	1553	=head3 Examples
1504		1554
1505	Example: Call a callback every hour, or, more precisely, whenever the	1555	Example: Call a callback every hour, or, more precisely, whenever the
1506	system clock is divisible by 3600. The callback invocation times have	1556	system time is divisible by 3600. The callback invocation times have
1507	potentially a lot of jitter, but good long-term stability.	1557	potentially a lot of jitter, but good long-term stability.
1508		1558
1509	static void	1559	static void
1510	clock_cb (struct ev_loop loop, struct ev_io w, int revents)	1560	clock_cb (struct ev_loop loop, struct ev_io w, int revents)
1511	{	1561	{
…		…
1521	#include <math.h>	1571	#include <math.h>
1522		1572
1523	static ev_tstamp	1573	static ev_tstamp
1524	my_scheduler_cb (struct ev_periodic *w, ev_tstamp now)	1574	my_scheduler_cb (struct ev_periodic *w, ev_tstamp now)
1525	{	1575	{
1526	return fmod (now, 3600.) + 3600.;	1576	return now + (3600. - fmod (now, 3600.));
1527	}	1577	}
1528		1578
1529	ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);	1579	ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
1530		1580
1531	Example: Call a callback every hour, starting now:	1581	Example: Call a callback every hour, starting now:
…		…
1541	Signal watchers will trigger an event when the process receives a specific	1591	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	1592	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	1593	will try it's best to deliver signals synchronously, i.e. as part of the
1544	normal event processing, like any other event.	1594	normal event processing, like any other event.
1545		1595
		1596	If you want signals asynchronously, just use C<sigaction> as you would
		1597	do without libev and forget about sharing the signal. You can even use
		1598	C<ev_async> from a signal handler to synchronously wake up an event loop.
		1599
1546	You can configure as many watchers as you like per signal. Only when the	1600	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	1601	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	1602	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	1603	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	1604	the last signal watcher for a signal is stopped, libev will reset the
1551	SIG_DFL (regardless of what it was set to before).	1605	signal handler to SIG_DFL (regardless of what it was set to before).
1552		1606
1553	If possible and supported, libev will install its handlers with	1607	If possible and supported, libev will install its handlers with
1554	C<SA_RESTART> behaviour enabled, so system calls should not be unduly	1608	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	1609	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	1610	signals you can block all signals in an C<ev_check> watcher and unblock
…		…
1573		1627
1574	=back	1628	=back
1575		1629
1576	=head3 Examples	1630	=head3 Examples
1577		1631
1578	Example: Try to exit cleanly on SIGINT and SIGTERM.	1632	Example: Try to exit cleanly on SIGINT.
1579		1633
1580	static void	1634	static void
1581	sigint_cb (struct ev_loop loop, struct ev_signal w, int revents)	1635	sigint_cb (struct ev_loop loop, struct ev_signal w, int revents)
1582	{	1636	{
1583	ev_unloop (loop, EVUNLOOP_ALL);	1637	ev_unloop (loop, EVUNLOOP_ALL);
1584	}	1638	}
1585		1639
1586	struct ev_signal signal_watcher;	1640	struct ev_signal signal_watcher;
1587	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);	1641	ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
1588	ev_signal_start (loop, &sigint_cb);	1642	ev_signal_start (loop, &signal_watcher);
1589		1643
1590		1644
1591	=head2 C<ev_child> - watch out for process status changes	1645	=head2 C<ev_child> - watch out for process status changes
1592		1646
1593	Child watchers trigger when your process receives a SIGCHLD in response to	1647	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	1648	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	1649	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	1650	has been forked (which implies it might have already exited), as long
1597	loop isn't entered (or is continued from a watcher).	1651	as the event loop isn't entered (or is continued from a watcher), i.e.,
		1652	forking and then immediately registering a watcher for the child is fine,
		1653	but forking and registering a watcher a few event loop iterations later is
		1654	not.
1598		1655
1599	Only the default event loop is capable of handling signals, and therefore	1656	Only the default event loop is capable of handling signals, and therefore
1600	you can only register child watchers in the default event loop.	1657	you can only register child watchers in the default event loop.
1601		1658
1602	=head3 Process Interaction	1659	=head3 Process Interaction
…		…
1700	the stat buffer having unspecified contents.	1757	the stat buffer having unspecified contents.
1701		1758
1702	The path I<should> be absolute and I<must not> end in a slash. If it is	1759	The path I<should> be absolute and I<must not> end in a slash. If it is
1703	relative and your working directory changes, the behaviour is undefined.	1760	relative and your working directory changes, the behaviour is undefined.
1704		1761
1705	Since there is no standard to do this, the portable implementation simply	1762	Since there is no standard kernel interface to do this, the portable
1706	calls C<stat (2)> regularly on the path to see if it changed somehow. You	1763	implementation simply calls C<stat (2)> regularly on the path to see if
1707	can specify a recommended polling interval for this case. If you specify	1764	it changed somehow. You can specify a recommended polling interval for
1708	a polling interval of C<0> (highly recommended!) then a I<suitable,	1765	this case. If you specify a polling interval of C<0> (highly recommended!)
1709	unspecified default> value will be used (which you can expect to be around	1766	then a I<suitable, unspecified default> value will be used (which
1710	five seconds, although this might change dynamically). Libev will also	1767	you can expect to be around five seconds, although this might change
1711	impose a minimum interval which is currently around C<0.1>, but thats	1768	dynamically). Libev will also impose a minimum interval which is currently
1712	usually overkill.	1769	around C<0.1>, but thats usually overkill.
1713		1770
1714	This watcher type is not meant for massive numbers of stat watchers,	1771	This watcher type is not meant for massive numbers of stat watchers,
1715	as even with OS-supported change notifications, this can be	1772	as even with OS-supported change notifications, this can be
1716	resource-intensive.	1773	resource-intensive.
1717		1774
1718	At the time of this writing, only the Linux inotify interface is	1775	At the time of this writing, the only OS-specific interface implemented
1719	implemented (implementing kqueue support is left as an exercise for the	1776	is the Linux inotify interface (implementing kqueue support is left as
1720	reader, note, however, that the author sees no way of implementing ev_stat	1777	an exercise for the reader. Note, however, that the author sees no way
1721	semantics with kqueue). Inotify will be used to give hints only and should	1778	of implementing C<ev_stat> semantics with kqueue).
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		1779
1727	=head3 ABI Issues (Largefile Support)	1780	=head3 ABI Issues (Largefile Support)
1728		1781
1729	Libev by default (unless the user overrides this) uses the default	1782	Libev by default (unless the user overrides this) uses the default
1730	compilation environment, which means that on systems with large file	1783	compilation environment, which means that on systems with large file
…		…
1739	file interfaces available by default (as e.g. FreeBSD does) and not	1792	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	1793	optional. Libev cannot simply switch on large file support because it has
1741	to exchange stat structures with application programs compiled using the	1794	to exchange stat structures with application programs compiled using the
1742	default compilation environment.	1795	default compilation environment.
1743		1796
1744	=head3 Inotify	1797	=head3 Inotify and Kqueue
1745		1798
1746	When C<inotify (7)> support has been compiled into libev (generally only	1799	When C<inotify (7)> support has been compiled into libev (generally only
1747	available on Linux) and present at runtime, it will be used to speed up	1800	available with Linux) and present at runtime, it will be used to speed up
1748	change detection where possible. The inotify descriptor will be created lazily	1801	change detection where possible. The inotify descriptor will be created lazily
1749	when the first C<ev_stat> watcher is being started.	1802	when the first C<ev_stat> watcher is being started.
1750		1803
1751	Inotify presence does not change the semantics of C<ev_stat> watchers	1804	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	1805	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	1806	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.	1807	there are many cases where libev has to resort to regular C<stat> polling,
		1808	but as long as the path exists, libev usually gets away without polling.
1755		1809
1756	(There is no support for kqueue, as apparently it cannot be used to	1810	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	1811	implement this functionality, due to the requirement of having a file
1758	descriptor open on the object at all times).	1812	descriptor open on the object at all times, and detecting renames, unlinks
		1813	etc. is difficult.
1759		1814
1760	=head3 The special problem of stat time resolution	1815	=head3 The special problem of stat time resolution
1761		1816
1762	The C<stat ()> system call only supports full-second resolution portably, and	1817	The C<stat ()> system call only supports full-second resolution portably, and
1763	even on systems where the resolution is higher, many file systems still	1818	even on systems where the resolution is higher, most file systems still
1764	only support whole seconds.	1819	only support whole seconds.
1765		1820
1766	That means that, if the time is the only thing that changes, you can	1821	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	1822	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	1823	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	1824	within the same second, C<ev_stat> will be unable to detect unless the
1770	data does not change.	1825	stat data does change in other ways (e.g. file size).
1771		1826
1772	The solution to this is to delay acting on a change for slightly more	1827	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	1828	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);	1829	a roughly one-second-delay C<ev_timer> (e.g. C<ev_timer_set (w, 0., 1.02);
1775	ev_timer_again (loop, w)>).	1830	ev_timer_again (loop, w)>).
…		…
1795	C<path>. The C<interval> is a hint on how quickly a change is expected to	1850	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	1851	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	1852	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.	1853	path for as long as the watcher is active.
1799		1854
1800	The callback will receive C<EV_STAT> when a change was detected, relative	1855	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	1856	relative to the attributes at the time the watcher was started (or the
1802	was detected).	1857	last change was detected).
1803		1858
1804	=item ev_stat_stat (loop, ev_stat *)	1859	=item ev_stat_stat (loop, ev_stat *)
1805		1860
1806	Updates the stat buffer immediately with new values. If you change the	1861	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	1862	watched path in your callback, you could call this function to avoid
…		…
1890		1945
1891		1946
1892	=head2 C<ev_idle> - when you've got nothing better to do...	1947	=head2 C<ev_idle> - when you've got nothing better to do...
1893		1948
1894	Idle watchers trigger events when no other events of the same or higher	1949	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	1950	priority are pending (prepare, check and other idle watchers do not count
1896	count).	1951	as receiving "events").
1897		1952
1898	That is, as long as your process is busy handling sockets or timeouts	1953	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	1954	(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	1955	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	1956	are pending), the idle watchers are being called once per event loop
…		…
1940	ev_idle_start (loop, idle_cb);	1995	ev_idle_start (loop, idle_cb);
1941		1996
1942		1997
1943	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!	1998	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!
1944		1999
1945	Prepare and check watchers are usually (but not always) used in tandem:	2000	Prepare and check watchers are usually (but not always) used in pairs:
1946	prepare watchers get invoked before the process blocks and check watchers	2001	prepare watchers get invoked before the process blocks and check watchers
1947	afterwards.	2002	afterwards.
1948		2003
1949	You I<must not> call C<ev_loop> or similar functions that enter	2004	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>	2005	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,	2008	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	2009	C<ev_check> so if you have one watcher of each kind they will always be
1955	called in pairs bracketing the blocking call.	2010	called in pairs bracketing the blocking call.
1956		2011
1957	Their main purpose is to integrate other event mechanisms into libev and	2012	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	2013	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	2014	variable changes, implement your own watchers, integrate net-snmp or a
1960	coroutine library and lots more. They are also occasionally useful if	2015	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,	2016	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>	2017	in X programs you might want to do an C<XFlush ()> in an C<ev_prepare>
1963	watcher).	2018	watcher).
1964		2019
1965	This is done by examining in each prepare call which file descriptors need	2020	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	2021	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	2022	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	2023	libraries provide exactly this functionality). Then, in the check watcher,
1969	any events that occurred (by checking the pending status of all watchers	2024	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	2025	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,	2026	I/O and timer callbacks will never actually be called (but must be valid
1972	because you never know, you know?).	2027	nevertheless, because you never know, you know?).
1973		2028
1974	As another example, the Perl Coro module uses these hooks to integrate	2029	As another example, the Perl Coro module uses these hooks to integrate
1975	coroutines into libev programs, by yielding to other active coroutines	2030	coroutines into libev programs, by yielding to other active coroutines
1976	during each prepare and only letting the process block if no coroutines	2031	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	2032	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	2035	loop from blocking if lower-priority coroutines are active, thus mapping
1981	low-priority coroutines to idle/background tasks).	2036	low-priority coroutines to idle/background tasks).
1982		2037
1983	It is recommended to give C<ev_check> watchers highest (C<EV_MAXPRI>)	2038	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	2039	priority, to ensure that they are being run before any other watchers
		2040	after the poll (this doesn't matter for C<ev_prepare> watchers).
		2041
1985	after the poll. Also, C<ev_check> watchers (and C<ev_prepare> watchers,	2042	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	2043	activate ("feed") events into libev. While libev fully supports this, they
1987	supports this, they might get executed before other C<ev_check> watchers	2044	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	2045	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	2046	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	2047	C<ev_check> watcher ran (always remind yourself to coexist peacefully with
1991	coexist peacefully with others).	2048	others).
1992		2049
1993	=head3 Watcher-Specific Functions and Data Members	2050	=head3 Watcher-Specific Functions and Data Members
1994		2051
1995	=over 4	2052	=over 4
1996		2053
…		…
1998		2055
1999	=item ev_check_init (ev_check *, callback)	2056	=item ev_check_init (ev_check *, callback)
2000		2057
2001	Initialises and configures the prepare or check watcher - they have no	2058	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>	2059	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.	2060	macros, but using them is utterly, utterly, utterly and completely
		2061	pointless.
2004		2062
2005	=back	2063	=back
2006		2064
2007	=head3 Examples	2065	=head3 Examples
2008		2066
…		…
2101	}	2159	}
2102		2160
2103	// do not ever call adns_afterpoll	2161	// do not ever call adns_afterpoll
2104		2162
2105	Method 4: Do not use a prepare or check watcher because the module you	2163	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	2164	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	2165	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	2166	main loop is now no longer controllable by EV. The C<Glib::EV> module uses
2109	this.	2167	this approach, effectively embedding EV as a client into the horrible
		2168	libglib event loop.
2110		2169
2111	static gint	2170	static gint
2112	event_poll_func (GPollFD *fds, guint nfds, gint timeout)	2171	event_poll_func (GPollFD *fds, guint nfds, gint timeout)
2113	{	2172	{
2114	int got_events = 0;	2173	int got_events = 0;
…		…
2145	prioritise I/O.	2204	prioritise I/O.
2146		2205
2147	As an example for a bug workaround, the kqueue backend might only support	2206	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	2207	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	2208	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	2209	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	2210	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	2211	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.	2212	C<kevent>, but at least you can use both mechanisms for what they are
		2213	best: C<kqueue> for scalable sockets and C<poll> if you want it to work :)
2154		2214
2155	As for prioritising I/O: rarely you have the case where some fds have	2215	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	2216	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	2217	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	2218	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.	2219	the rest in a second one, and embed the second one in the first.
2160		2220
2161	As long as the watcher is active, the callback will be invoked every time	2221	As long as the watcher is active, the callback will be invoked every time
2162	there might be events pending in the embedded loop. The callback must then	2222	there might be events pending in the embedded loop. The callback must then
2163	call C<ev_embed_sweep (mainloop, watcher)> to make a single sweep and invoke	2223	call C<ev_embed_sweep (mainloop, watcher)> to make a single sweep and invoke
2164	their callbacks (you could also start an idle watcher to give the embedded	2224	their callbacks (you could also start an idle watcher to give the embedded
…		…
2172	interested in that.	2232	interested in that.
2173		2233
2174	Also, there have not currently been made special provisions for forking:	2234	Also, there have not currently been made special provisions for forking:
2175	when you fork, you not only have to call C<ev_loop_fork> on both loops,	2235	when you fork, you not only have to call C<ev_loop_fork> on both loops,
2176	but you will also have to stop and restart any C<ev_embed> watchers	2236	but you will also have to stop and restart any C<ev_embed> watchers
2177	yourself.	2237	yourself - but you can use a fork watcher to handle this automatically,
		2238	and future versions of libev might do just that.
2178		2239
2179	Unfortunately, not all backends are embeddable, only the ones returned by	2240	Unfortunately, not all backends are embeddable: only the ones returned by
2180	C<ev_embeddable_backends> are, which, unfortunately, does not include any	2241	C<ev_embeddable_backends> are, which, unfortunately, does not include any
2181	portable one.	2242	portable one.
2182		2243
2183	So when you want to use this feature you will always have to be prepared	2244	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	2245	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	2246	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.	2247	create it, and if that fails, use the normal loop for everything.
		2248
		2249	=head3 C<ev_embed> and fork
		2250
		2251	While the C<ev_embed> watcher is running, forks in the embedding loop will
		2252	automatically be applied to the embedded loop as well, so no special
		2253	fork handling is required in that case. When the watcher is not running,
		2254	however, it is still the task of the libev user to call C<ev_loop_fork ()>
		2255	as applicable.
2187		2256
2188	=head3 Watcher-Specific Functions and Data Members	2257	=head3 Watcher-Specific Functions and Data Members
2189		2258
2190	=over 4	2259	=over 4
2191		2260
…		…
2309	is that the author does not know of a simple (or any) algorithm for a	2378	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	2379	multiple-writer-single-reader queue that works in all cases and doesn't
2311	need elaborate support such as pthreads.	2380	need elaborate support such as pthreads.
2312		2381
2313	That means that if you want to queue data, you have to provide your own	2382	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	2383	queue. But at least I can tell you how to implement locking around your
2315	queue:	2384	queue:
2316		2385
2317	=over 4	2386	=over 4
2318		2387
2319	=item queueing from a signal handler context	2388	=item queueing from a signal handler context
2320		2389
2321	To implement race-free queueing, you simply add to the queue in the signal	2390	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	2391	handler but you block the signal handler in the watcher callback. Here is
2323	some fictitious SIGUSR1 handler:	2392	an example that does that for some fictitious SIGUSR1 handler:
2324		2393
2325	static ev_async mysig;	2394	static ev_async mysig;
2326		2395
2327	static void	2396	static void
2328	sigusr1_handler (void)	2397	sigusr1_handler (void)
…		…
2395		2464
2396	=item ev_async_init (ev_async *, callback)	2465	=item ev_async_init (ev_async *, callback)
2397		2466
2398	Initialises and configures the async watcher - it has no parameters of any	2467	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,	2468	kind. There is a C<ev_asynd_set> macro, but using it is utterly pointless,
2400	believe me.	2469	trust me.
2401		2470
2402	=item ev_async_send (loop, ev_async *)	2471	=item ev_async_send (loop, ev_async *)
2403		2472
2404	Sends/signals/activates the given C<ev_async> watcher, that is, feeds	2473	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	2474	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	2475	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	2476	similar contexts (see the discussion of C<EV_ATOMIC_T> in the embedding
2408	section below on what exactly this means).	2477	section below on what exactly this means).
2409		2478
2410	This call incurs the overhead of a system call only once per loop iteration,	2479	This call incurs the overhead of a system call only once per loop iteration,
2411	so while the overhead might be noticeable, it doesn't apply to repeated	2480	so while the overhead might be noticeable, it doesn't apply to repeated
…		…
2435	=over 4	2504	=over 4
2436		2505
2437	=item ev_once (loop, int fd, int events, ev_tstamp timeout, callback)	2506	=item ev_once (loop, int fd, int events, ev_tstamp timeout, callback)
2438		2507
2439	This function combines a simple timer and an I/O watcher, calls your	2508	This function combines a simple timer and an I/O watcher, calls your
2440	callback on whichever event happens first and automatically stop both	2509	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	2510	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	2511	or timeout without having to allocate/configure/start/stop/free one or
2443	more watchers yourself.	2512	more watchers yourself.
2444		2513
2445	If C<fd> is less than 0, then no I/O watcher will be started and events	2514	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	2515	C<events> argument is being ignored. Otherwise, an C<ev_io> watcher for
2447	C<events> set will be created and started.	2516	the given C<fd> and C<events> set will be created and started.
2448		2517
2449	If C<timeout> is less than 0, then no timeout watcher will be	2518	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	2519	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	2520	repeat = 0) will be started. C<0> is a valid timeout.
2452	dubious value.
2453		2521
2454	The callback has the type C<void (cb)(int revents, void arg)> and gets	2522	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	2523	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>	2524	C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMEOUT>) and the C<arg>
2457	value passed to C<ev_once>:	2525	value passed to C<ev_once>. Note that it is possible to receive I<both>
		2526	a timeout and an io event at the same time - you probably should give io
		2527	events precedence.
		2528
		2529	Example: wait up to ten seconds for data to appear on STDIN_FILENO.
2458		2530
2459	static void stdin_ready (int revents, void *arg)	2531	static void stdin_ready (int revents, void *arg)
2460	{	2532	{
		2533	if (revents & EV_READ)
		2534	/* stdin might have data for us, joy! */;
2461	if (revents & EV_TIMEOUT)	2535	else if (revents & EV_TIMEOUT)
2462	/* doh, nothing entered */;	2536	/* doh, nothing entered */;
2463	else if (revents & EV_READ)
2464	/* stdin might have data for us, joy! */;
2465	}	2537	}
2466		2538
2467	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);	2539	ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
2468		2540
2469	=item ev_feed_event (ev_loop , watcher , int revents)	2541	=item ev_feed_event (ev_loop , watcher , int revents)
…		…
2617		2689
2618	The prototype of the C<function> must be C<void (*)(ev::TYPE &w, int)>.	2690	The prototype of the C<function> must be C<void (*)(ev::TYPE &w, int)>.
2619		2691
2620	See the method-C<set> above for more details.	2692	See the method-C<set> above for more details.
2621		2693
2622	Example:	2694	Example: Use a plain function as callback.
2623		2695
2624	static void io_cb (ev::io &w, int revents) { }	2696	static void io_cb (ev::io &w, int revents) { }
2625	iow.set <io_cb> ();	2697	iow.set <io_cb> ();
2626		2698
2627	=item w->set (struct ev_loop *)	2699	=item w->set (struct ev_loop *)
…		…
2665	Example: Define a class with an IO and idle watcher, start one of them in	2737	Example: Define a class with an IO and idle watcher, start one of them in
2666	the constructor.	2738	the constructor.
2667		2739
2668	class myclass	2740	class myclass
2669	{	2741	{
2670	ev::io io; void io_cb (ev::io &w, int revents);	2742	ev::io io ; void io_cb (ev::io &w, int revents);
2671	ev:idle idle void idle_cb (ev::idle &w, int revents);	2743	ev::idle idle; void idle_cb (ev::idle &w, int revents);
2672		2744
2673	myclass (int fd)	2745	myclass (int fd)
2674	{	2746	{
2675	io .set <myclass, &myclass::io_cb > (this);	2747	io .set <myclass, &myclass::io_cb > (this);
2676	idle.set <myclass, &myclass::idle_cb> (this);	2748	idle.set <myclass, &myclass::idle_cb> (this);
…		…
2692	=item Perl	2764	=item Perl
2693		2765
2694	The EV module implements the full libev API and is actually used to test	2766	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,	2767	libev. EV is developed together with libev. Apart from the EV core module,
2696	there are additional modules that implement libev-compatible interfaces	2768	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	2769	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>).	2770	C<Net::SNMP> (C<Net::SNMP::EV>) and the C<libglib> event core (C<Glib::EV>
		2771	and C<EV::Glib>).
2699		2772
2700	It can be found and installed via CPAN, its homepage is at	2773	It can be found and installed via CPAN, its homepage is at
2701	L<http://software.schmorp.de/pkg/EV>.	2774	L<http://software.schmorp.de/pkg/EV>.
2702		2775
2703	=item Python	2776	=item Python
…		…
2882		2955
2883	=head2 PREPROCESSOR SYMBOLS/MACROS	2956	=head2 PREPROCESSOR SYMBOLS/MACROS
2884		2957
2885	Libev can be configured via a variety of preprocessor symbols you have to	2958	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	2959	define before including any of its files. The default in the absence of
2887	autoconf is noted for every option.	2960	autoconf is documented for every option.
2888		2961
2889	=over 4	2962	=over 4
2890		2963
2891	=item EV_STANDALONE	2964	=item EV_STANDALONE
2892		2965
…		…
3062	When doing priority-based operations, libev usually has to linearly search	3135	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	3136	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	3137	and time, so using the defaults of five priorities (-2 .. +2) is usually
3065	fine.	3138	fine.
3066		3139
3067	If your embedding application does not need any priorities, defining these both to	3140	If your embedding application does not need any priorities, defining these
3068	C<0> will save some memory and CPU.	3141	both to C<0> will save some memory and CPU.
3069		3142
3070	=item EV_PERIODIC_ENABLE	3143	=item EV_PERIODIC_ENABLE
3071		3144
3072	If undefined or defined to be C<1>, then periodic timers are supported. If	3145	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	3146	defined to be C<0>, then they are not. Disabling them saves a few kB of
…		…
3080	code.	3153	code.
3081		3154
3082	=item EV_EMBED_ENABLE	3155	=item EV_EMBED_ENABLE
3083		3156
3084	If undefined or defined to be C<1>, then embed watchers are supported. If	3157	If undefined or defined to be C<1>, then embed watchers are supported. If
3085	defined to be C<0>, then they are not.	3158	defined to be C<0>, then they are not. Embed watchers rely on most other
		3159	watcher types, which therefore must not be disabled.
3086		3160
3087	=item EV_STAT_ENABLE	3161	=item EV_STAT_ENABLE
3088		3162
3089	If undefined or defined to be C<1>, then stat watchers are supported. If	3163	If undefined or defined to be C<1>, then stat watchers are supported. If
3090	defined to be C<0>, then they are not.	3164	defined to be C<0>, then they are not.
…		…
3122	two).	3196	two).
3123		3197
3124	=item EV_USE_4HEAP	3198	=item EV_USE_4HEAP
3125		3199
3126	Heaps are not very cache-efficient. To improve the cache-efficiency of the	3200	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	3201	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	3202	to C<1>. The 4-heap uses more complicated (longer) code but has noticeably
3129	noticeably faster performance with many (thousands) of watchers.	3203	faster performance with many (thousands) of watchers.
3130		3204
3131	The default is C<1> unless C<EV_MINIMAL> is set in which case it is C<0>	3205	The default is C<1> unless C<EV_MINIMAL> is set in which case it is C<0>
3132	(disabled).	3206	(disabled).
3133		3207
3134	=item EV_HEAP_CACHE_AT	3208	=item EV_HEAP_CACHE_AT
3135		3209
3136	Heaps are not very cache-efficient. To improve the cache-efficiency of the	3210	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	3211	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>),	3212	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,	3213	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	3214	but avoids random read accesses on heap changes. This improves performance
3141	noticeably with with many (hundreds) of watchers.	3215	noticeably with many (hundreds) of watchers.
3142		3216
3143	The default is C<1> unless C<EV_MINIMAL> is set in which case it is C<0>	3217	The default is C<1> unless C<EV_MINIMAL> is set in which case it is C<0>
3144	(disabled).	3218	(disabled).
3145		3219
3146	=item EV_VERIFY	3220	=item EV_VERIFY
…		…
3152	called once per loop, which can slow down libev. If set to C<3>, then the	3226	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	3227	verification code will be called very frequently, which will slow down
3154	libev considerably.	3228	libev considerably.
3155		3229
3156	The default is C<1>, unless C<EV_MINIMAL> is set, in which case it will be	3230	The default is C<1>, unless C<EV_MINIMAL> is set, in which case it will be
3157	C<0.>	3231	C<0>.
3158		3232
3159	=item EV_COMMON	3233	=item EV_COMMON
3160		3234
3161	By default, all watchers have a C<void *data> member. By redefining	3235	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	3236	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	3253	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	3254	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	3255	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	3256	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++.	3257	method calls instead of plain function calls in C++.
		3258
		3259	=back
3184		3260
3185	=head2 EXPORTED API SYMBOLS	3261	=head2 EXPORTED API SYMBOLS
3186		3262
3187	If you need to re-export the API (e.g. via a DLL) and you need a list of	3263	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	3264	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:	3311	And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:
3236		3312
3237	#include "ev_cpp.h"	3313	#include "ev_cpp.h"
3238	#include "ev.c"	3314	#include "ev.c"
3239		3315
		3316	=head1 INTERACTION WITH OTHER PROGRAMS OR LIBRARIES
3240		3317
3241	=head1 THREADS AND COROUTINES	3318	=head2 THREADS AND COROUTINES
3242		3319
3243	=head2 THREADS	3320	=head3 THREADS
3244		3321
3245	Libev itself is completely thread-safe, but it uses no locking. This	3322	All libev functions are reentrant and thread-safe unless explicitly
		3323	documented otherwise, but libev implements no locking itself. This means
3246	means that you can use as many loops as you want in parallel, as long as	3324	that you can use as many loops as you want in parallel, as long as there
3247	only one thread ever calls into one libev function with the same loop	3325	are no concurrent calls into any libev function with the same loop
3248	parameter.	3326	parameter (C<ev_default_*> calls have an implicit default loop parameter,
		3327	of course): libev guarantees that different event loops share no data
		3328	structures that need any locking.
3249		3329
3250	Or put differently: calls with different loop parameters can be done in	3330	Or to put it differently: calls with different loop parameters can be done
3251	parallel from multiple threads, calls with the same loop parameter must be	3331	concurrently from multiple threads, calls with the same loop parameter
3252	done serially (but can be done from different threads, as long as only one	3332	must be done serially (but can be done from different threads, as long as
3253	thread ever is inside a call at any point in time, e.g. by using a mutex	3333	only one thread ever is inside a call at any point in time, e.g. by using
3254	per loop).	3334	a mutex per loop).
		3335
		3336	Specifically to support threads (and signal handlers), libev implements
		3337	so-called C<ev_async> watchers, which allow some limited form of
		3338	concurrency on the same event loop, namely waking it up "from the
		3339	outside".
3255		3340
3256	If you want to know which design (one loop, locking, or multiple loops	3341	If you want to know which design (one loop, locking, or multiple loops
3257	without or something else still) is best for your problem, then I cannot	3342	without or something else still) is best for your problem, then I cannot
3258	help you. I can give some generic advice however:	3343	help you, but here is some generic advice:
3259		3344
3260	=over 4	3345	=over 4
3261		3346
3262	=item * most applications have a main thread: use the default libev loop	3347	=item * most applications have a main thread: use the default libev loop
3263	in that thread, or create a separate thread running only the default loop.	3348	in that thread, or create a separate thread running only the default loop.
…		…
3275		3360
3276	Choosing a model is hard - look around, learn, know that usually you can do	3361	Choosing a model is hard - look around, learn, know that usually you can do
3277	better than you currently do :-)	3362	better than you currently do :-)
3278		3363
3279	=item * often you need to talk to some other thread which blocks in the	3364	=item * often you need to talk to some other thread which blocks in the
		3365	event loop.
		3366
3280	event loop - C<ev_async> watchers can be used to wake them up from other	3367	C<ev_async> watchers can be used to wake them up from other threads safely
3281	threads safely (or from signal contexts...).	3368	(or from signal contexts...).
		3369
		3370	An example use would be to communicate signals or other events that only
		3371	work in the default loop by registering the signal watcher with the
		3372	default loop and triggering an C<ev_async> watcher from the default loop
		3373	watcher callback into the event loop interested in the signal.
3282		3374
3283	=back	3375	=back
3284		3376
3285	=head2 COROUTINES	3377	=head3 COROUTINES
3286		3378
3287	Libev is much more accommodating to coroutines ("cooperative threads"):	3379	Libev is very accommodating to coroutines ("cooperative threads"):
3288	libev fully supports nesting calls to it's functions from different	3380	libev fully supports nesting calls to its functions from different
3289	coroutines (e.g. you can call C<ev_loop> on the same loop from two	3381	coroutines (e.g. you can call C<ev_loop> on the same loop from two
3290	different coroutines and switch freely between both coroutines running the	3382	different coroutines, and switch freely between both coroutines running the
3291	loop, as long as you don't confuse yourself). The only exception is that	3383	loop, as long as you don't confuse yourself). The only exception is that
3292	you must not do this from C<ev_periodic> reschedule callbacks.	3384	you must not do this from C<ev_periodic> reschedule callbacks.
3293		3385
3294	Care has been invested into making sure that libev does not keep local	3386	Care has been taken to ensure that libev does not keep local state inside
3295	state inside C<ev_loop>, and other calls do not usually allow coroutine	3387	C<ev_loop>, and other calls do not usually allow for coroutine switches as
3296	switches.	3388	they do not clal any callbacks.
3297		3389
		3390	=head2 COMPILER WARNINGS
3298		3391
3299	=head1 COMPLEXITIES	3392	Depending on your compiler and compiler settings, you might get no or a
		3393	lot of warnings when compiling libev code. Some people are apparently
		3394	scared by this.
3300		3395
3301	In this section the complexities of (many of) the algorithms used inside	3396	However, these are unavoidable for many reasons. For one, each compiler
3302	libev will be explained. For complexity discussions about backends see the	3397	has different warnings, and each user has different tastes regarding
3303	documentation for C<ev_default_init>.	3398	warning options. "Warn-free" code therefore cannot be a goal except when
		3399	targeting a specific compiler and compiler-version.
3304		3400
3305	All of the following are about amortised time: If an array needs to be	3401	Another reason is that some compiler warnings require elaborate
3306	extended, libev needs to realloc and move the whole array, but this	3402	workarounds, or other changes to the code that make it less clear and less
3307	happens asymptotically never with higher number of elements, so O(1) might	3403	maintainable.
3308	mean it might do a lengthy realloc operation in rare cases, but on average
3309	it is much faster and asymptotically approaches constant time.
3310		3404
3311	=over 4	3405	And of course, some compiler warnings are just plain stupid, or simply
		3406	wrong (because they don't actually warn about the condition their message
		3407	seems to warn about). For example, certain older gcc versions had some
		3408	warnings that resulted an extreme number of false positives. These have
		3409	been fixed, but some people still insist on making code warn-free with
		3410	such buggy versions.
3312		3411
3313	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)	3412	While libev is written to generate as few warnings as possible,
		3413	"warn-free" code is not a goal, and it is recommended not to build libev
		3414	with any compiler warnings enabled unless you are prepared to cope with
		3415	them (e.g. by ignoring them). Remember that warnings are just that:
		3416	warnings, not errors, or proof of bugs.
3314		3417
3315	This means that, when you have a watcher that triggers in one hour and
3316	there are 100 watchers that would trigger before that then inserting will
3317	have to skip roughly seven (C<ld 100>) of these watchers.
3318		3418
3319	=item Changing timer/periodic watchers (by autorepeat or calling again): O(log skipped_other_timers)	3419	=head2 VALGRIND
3320		3420
3321	That means that changing a timer costs less than removing/adding them	3421	Valgrind has a special section here because it is a popular tool that is
3322	as only the relative motion in the event queue has to be paid for.	3422	highly useful. Unfortunately, valgrind reports are very hard to interpret.
3323		3423
3324	=item Starting io/check/prepare/idle/signal/child/fork/async watchers: O(1)	3424	If you think you found a bug (memory leak, uninitialised data access etc.)
		3425	in libev, then check twice: If valgrind reports something like:
3325		3426
3326	These just add the watcher into an array or at the head of a list.	3427	==2274== definitely lost: 0 bytes in 0 blocks.
		3428	==2274== possibly lost: 0 bytes in 0 blocks.
		3429	==2274== still reachable: 256 bytes in 1 blocks.
3327		3430
3328	=item Stopping check/prepare/idle/fork/async watchers: O(1)	3431	Then there is no memory leak, just as memory accounted to global variables
		3432	is not a memleak - the memory is still being refernced, and didn't leak.
3329		3433
3330	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))	3434	Similarly, under some circumstances, valgrind might report kernel bugs
		3435	as if it were a bug in libev (e.g. in realloc or in the poll backend,
		3436	although an acceptable workaround has been found here), or it might be
		3437	confused.
3331		3438
3332	These watchers are stored in lists then need to be walked to find the	3439	Keep in mind that valgrind is a very good tool, but only a tool. Don't
3333	correct watcher to remove. The lists are usually short (you don't usually	3440	make it into some kind of religion.
3334	have many watchers waiting for the same fd or signal).
3335		3441
3336	=item Finding the next timer in each loop iteration: O(1)	3442	If you are unsure about something, feel free to contact the mailing list
		3443	with the full valgrind report and an explanation on why you think this
		3444	is a bug in libev (best check the archives, too :). However, don't be
		3445	annoyed when you get a brisk "this is no bug" answer and take the chance
		3446	of learning how to interpret valgrind properly.
3337		3447
3338	By virtue of using a binary or 4-heap, the next timer is always found at a	3448	If you need, for some reason, empty reports from valgrind for your project
3339	fixed position in the storage array.	3449	I suggest using suppression lists.
3340		3450
3341	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)
3342		3451
3343	A change means an I/O watcher gets started or stopped, which requires	3452	=head1 PORTABILITY NOTES
3344	libev to recalculate its status (and possibly tell the kernel, depending
3345	on backend and whether C<ev_io_set> was used).
3346		3453
3347	=item Activating one watcher (putting it into the pending state): O(1)
3348
3349	=item Priority handling: O(number_of_priorities)
3350
3351	Priorities are implemented by allocating some space for each
3352	priority. When doing priority-based operations, libev usually has to
3353	linearly search all the priorities, but starting/stopping and activating
3354	watchers becomes O(1) w.r.t. priority handling.
3355
3356	=item Sending an ev_async: O(1)
3357
3358	=item Processing ev_async_send: O(number_of_async_watchers)
3359
3360	=item Processing signals: O(max_signal_number)
3361
3362	Sending involves a system call I<iff> there were no other C<ev_async_send>
3363	calls in the current loop iteration. Checking for async and signal events
3364	involves iterating over all running async watchers or all signal numbers.
3365
3366	=back
3367
3368
3369	=head1 WIN32 PLATFORM LIMITATIONS AND WORKAROUNDS	3454	=head2 WIN32 PLATFORM LIMITATIONS AND WORKAROUNDS
3370		3455
3371	Win32 doesn't support any of the standards (e.g. POSIX) that libev	3456	Win32 doesn't support any of the standards (e.g. POSIX) that libev
3372	requires, and its I/O model is fundamentally incompatible with the POSIX	3457	requires, and its I/O model is fundamentally incompatible with the POSIX
3373	model. Libev still offers limited functionality on this platform in	3458	model. Libev still offers limited functionality on this platform in
3374	the form of the C<EVBACKEND_SELECT> backend, and only supports socket	3459	the form of the C<EVBACKEND_SELECT> backend, and only supports socket
…		…
3385		3470
3386	Not a libev limitation but worth mentioning: windows apparently doesn't	3471	Not a libev limitation but worth mentioning: windows apparently doesn't
3387	accept large writes: instead of resulting in a partial write, windows will	3472	accept large writes: instead of resulting in a partial write, windows will
3388	either accept everything or return C<ENOBUFS> if the buffer is too large,	3473	either accept everything or return C<ENOBUFS> if the buffer is too large,
3389	so make sure you only write small amounts into your sockets (less than a	3474	so make sure you only write small amounts into your sockets (less than a
3390	megabyte seems safe, but thsi apparently depends on the amount of memory	3475	megabyte seems safe, but this apparently depends on the amount of memory
3391	available).	3476	available).
3392		3477
3393	Due to the many, low, and arbitrary limits on the win32 platform and	3478	Due to the many, low, and arbitrary limits on the win32 platform and
3394	the abysmal performance of winsockets, using a large number of sockets	3479	the abysmal performance of winsockets, using a large number of sockets
3395	is not recommended (and not reasonable). If your program needs to use	3480	is not recommended (and not reasonable). If your program needs to use
…		…
3406	#define EV_SELECT_IS_WINSOCKET 1 /* configure libev for windows select */	3491	#define EV_SELECT_IS_WINSOCKET 1 /* configure libev for windows select */
3407		3492
3408	#include "ev.h"	3493	#include "ev.h"
3409		3494
3410	And compile the following F<evwrap.c> file into your project (make sure	3495	And compile the following F<evwrap.c> file into your project (make sure
3411	you do I<not> compile the F<ev.c> or any other embedded soruce files!):	3496	you do I<not> compile the F<ev.c> or any other embedded source files!):
3412		3497
3413	#include "evwrap.h"	3498	#include "evwrap.h"
3414	#include "ev.c"	3499	#include "ev.c"
3415		3500
3416	=over 4	3501	=over 4
…		…
3461	wrap all I/O functions and provide your own fd management, but the cost of	3546	wrap all I/O functions and provide your own fd management, but the cost of
3462	calling select (O(n²)) will likely make this unworkable.	3547	calling select (O(n²)) will likely make this unworkable.
3463		3548
3464	=back	3549	=back
3465		3550
3466
3467	=head1 PORTABILITY REQUIREMENTS	3551	=head2 PORTABILITY REQUIREMENTS
3468		3552
3469	In addition to a working ISO-C implementation, libev relies on a few	3553	In addition to a working ISO-C implementation and of course the
3470	additional extensions:	3554	backend-specific APIs, libev relies on a few additional extensions:
3471		3555
3472	=over 4	3556	=over 4
3473		3557
3474	=item C<void ()(ev_watcher_type , int revents)> must have compatible	3558	=item C<void ()(ev_watcher_type , int revents)> must have compatible
3475	calling conventions regardless of C<ev_watcher_type *>.	3559	calling conventions regardless of C<ev_watcher_type *>.
…		…
3481	calls them using an C<ev_watcher *> internally.	3565	calls them using an C<ev_watcher *> internally.
3482		3566
3483	=item C<sig_atomic_t volatile> must be thread-atomic as well	3567	=item C<sig_atomic_t volatile> must be thread-atomic as well
3484		3568
3485	The type C<sig_atomic_t volatile> (or whatever is defined as	3569	The type C<sig_atomic_t volatile> (or whatever is defined as
3486	C<EV_ATOMIC_T>) must be atomic w.r.t. accesses from different	3570	C<EV_ATOMIC_T>) must be atomic with respect to accesses from different
3487	threads. This is not part of the specification for C<sig_atomic_t>, but is	3571	threads. This is not part of the specification for C<sig_atomic_t>, but is
3488	believed to be sufficiently portable.	3572	believed to be sufficiently portable.
3489		3573
3490	=item C<sigprocmask> must work in a threaded environment	3574	=item C<sigprocmask> must work in a threaded environment
3491		3575
…		…
3500	except the initial one, and run the default loop in the initial thread as	3584	except the initial one, and run the default loop in the initial thread as
3501	well.	3585	well.
3502		3586
3503	=item C<long> must be large enough for common memory allocation sizes	3587	=item C<long> must be large enough for common memory allocation sizes
3504		3588
3505	To improve portability and simplify using libev, libev uses C<long>	3589	To improve portability and simplify its API, libev uses C<long> internally
3506	internally instead of C<size_t> when allocating its data structures. On	3590	instead of C<size_t> when allocating its data structures. On non-POSIX
3507	non-POSIX systems (Microsoft...) this might be unexpectedly low, but	3591	systems (Microsoft...) this might be unexpectedly low, but is still at
3508	is still at least 31 bits everywhere, which is enough for hundreds of	3592	least 31 bits everywhere, which is enough for hundreds of millions of
3509	millions of watchers.	3593	watchers.
3510		3594
3511	=item C<double> must hold a time value in seconds with enough accuracy	3595	=item C<double> must hold a time value in seconds with enough accuracy
3512		3596
3513	The type C<double> is used to represent timestamps. It is required to	3597	The type C<double> is used to represent timestamps. It is required to
3514	have at least 51 bits of mantissa (and 9 bits of exponent), which is good	3598	have at least 51 bits of mantissa (and 9 bits of exponent), which is good
…		…
3518	=back	3602	=back
3519		3603
3520	If you know of other additional requirements drop me a note.	3604	If you know of other additional requirements drop me a note.
3521		3605
3522		3606
3523	=head1 COMPILER WARNINGS	3607	=head1 ALGORITHMIC COMPLEXITIES
3524		3608
3525	Depending on your compiler and compiler settings, you might get no or a	3609	In this section the complexities of (many of) the algorithms used inside
3526	lot of warnings when compiling libev code. Some people are apparently	3610	libev will be documented. For complexity discussions about backends see
3527	scared by this.	3611	the documentation for C<ev_default_init>.
3528		3612
3529	However, these are unavoidable for many reasons. For one, each compiler	3613	All of the following are about amortised time: If an array needs to be
3530	has different warnings, and each user has different tastes regarding	3614	extended, libev needs to realloc and move the whole array, but this
3531	warning options. "Warn-free" code therefore cannot be a goal except when	3615	happens asymptotically rarer with higher number of elements, so O(1) might
3532	targeting a specific compiler and compiler-version.	3616	mean that libev does a lengthy realloc operation in rare cases, but on
		3617	average it is much faster and asymptotically approaches constant time.
3533		3618
3534	Another reason is that some compiler warnings require elaborate	3619	=over 4
3535	workarounds, or other changes to the code that make it less clear and less
3536	maintainable.
3537		3620
3538	And of course, some compiler warnings are just plain stupid, or simply	3621	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)
3539	wrong (because they don't actually warn about the condition their message
3540	seems to warn about).
3541		3622
3542	While libev is written to generate as few warnings as possible,	3623	This means that, when you have a watcher that triggers in one hour and
3543	"warn-free" code is not a goal, and it is recommended not to build libev	3624	there are 100 watchers that would trigger before that, then inserting will
3544	with any compiler warnings enabled unless you are prepared to cope with	3625	have to skip roughly seven (C<ld 100>) of these watchers.
3545	them (e.g. by ignoring them). Remember that warnings are just that:
3546	warnings, not errors, or proof of bugs.
3547		3626
		3627	=item Changing timer/periodic watchers (by autorepeat or calling again): O(log skipped_other_timers)
3548		3628
3549	=head1 VALGRIND	3629	That means that changing a timer costs less than removing/adding them,
		3630	as only the relative motion in the event queue has to be paid for.
3550		3631
3551	Valgrind has a special section here because it is a popular tool that is	3632	=item Starting io/check/prepare/idle/signal/child/fork/async watchers: O(1)
3552	highly useful, but valgrind reports are very hard to interpret.
3553		3633
3554	If you think you found a bug (memory leak, uninitialised data access etc.)	3634	These just add the watcher into an array or at the head of a list.
3555	in libev, then check twice: If valgrind reports something like:
3556		3635
3557	==2274== definitely lost: 0 bytes in 0 blocks.	3636	=item Stopping check/prepare/idle/fork/async watchers: O(1)
3558	==2274== possibly lost: 0 bytes in 0 blocks.
3559	==2274== still reachable: 256 bytes in 1 blocks.
3560		3637
3561	Then there is no memory leak. Similarly, under some circumstances,	3638	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))
3562	valgrind might report kernel bugs as if it were a bug in libev, or it
3563	might be confused (it is a very good tool, but only a tool).
3564		3639
3565	If you are unsure about something, feel free to contact the mailing list	3640	These watchers are stored in lists, so they need to be walked to find the
3566	with the full valgrind report and an explanation on why you think this is	3641	correct watcher to remove. The lists are usually short (you don't usually
3567	a bug in libev. However, don't be annoyed when you get a brisk "this is	3642	have many watchers waiting for the same fd or signal: one is typical, two
3568	no bug" answer and take the chance of learning how to interpret valgrind	3643	is rare).
3569	properly.
3570		3644
3571	If you need, for some reason, empty reports from valgrind for your project	3645	=item Finding the next timer in each loop iteration: O(1)
3572	I suggest using suppression lists.	3646
		3647	By virtue of using a binary or 4-heap, the next timer is always found at a
		3648	fixed position in the storage array.
		3649
		3650	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)
		3651
		3652	A change means an I/O watcher gets started or stopped, which requires
		3653	libev to recalculate its status (and possibly tell the kernel, depending
		3654	on backend and whether C<ev_io_set> was used).
		3655
		3656	=item Activating one watcher (putting it into the pending state): O(1)
		3657
		3658	=item Priority handling: O(number_of_priorities)
		3659
		3660	Priorities are implemented by allocating some space for each
		3661	priority. When doing priority-based operations, libev usually has to
		3662	linearly search all the priorities, but starting/stopping and activating
		3663	watchers becomes O(1) with respect to priority handling.
		3664
		3665	=item Sending an ev_async: O(1)
		3666
		3667	=item Processing ev_async_send: O(number_of_async_watchers)
		3668
		3669	=item Processing signals: O(max_signal_number)
		3670
		3671	Sending involves a system call I<iff> there were no other C<ev_async_send>
		3672	calls in the current loop iteration. Checking for async and signal events
		3673	involves iterating over all running async watchers or all signal numbers.
		3674
		3675	=back
3573		3676
3574		3677
3575	=head1 AUTHOR	3678	=head1 AUTHOR
3576		3679
3577	Marc Lehmann <libev@schmorp.de>.	3680	Marc Lehmann <libev@schmorp.de>.

Diff Legend

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

Comparing libev/ev.pod (file contents): Revision 1.179 by root, Sat Sep 13 19:14:21 2008 UTC vs. Revision 1.194 by root, Mon Oct 20 16:08:36 2008 UTC

Diff Legend

Comparing libev/ev.pod (file contents):
Revision 1.179 by root, Sat Sep 13 19:14:21 2008 UTC vs.
Revision 1.194 by root, Mon Oct 20 16:08:36 2008 UTC