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

Comparing libev/ev.pod (file contents):
Revision 1.231 by root, Wed Apr 15 19:35:53 2009 UTC vs.
Revision 1.270 by root, Fri Sep 18 21:02:12 2009 UTC

…		…
62		62
63	// unloop was called, so exit	63	// unloop was called, so exit
64	return 0;	64	return 0;
65	}	65	}
66		66
67	=head1 DESCRIPTION	67	=head1 ABOUT THIS DOCUMENT
		68
		69	This document documents the libev software package.
68		70
69	The newest version of this document is also available as an html-formatted	71	The newest version of this document is also available as an html-formatted
70	web page you might find easier to navigate when reading it for the first	72	web page you might find easier to navigate when reading it for the first
71	time: L<http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod>.	73	time: L<http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod>.
		74
		75	While this document tries to be as complete as possible in documenting
		76	libev, its usage and the rationale behind its design, it is not a tutorial
		77	on event-based programming, nor will it introduce event-based programming
		78	with libev.
		79
		80	Familarity with event based programming techniques in general is assumed
		81	throughout this document.
		82
		83	=head1 ABOUT LIBEV
72		84
73	Libev is an event loop: you register interest in certain events (such as a	85	Libev is an event loop: you register interest in certain events (such as a
74	file descriptor being readable or a timeout occurring), and it will manage	86	file descriptor being readable or a timeout occurring), and it will manage
75	these event sources and provide your program with events.	87	these event sources and provide your program with events.
76		88
…		…
86	=head2 FEATURES	98	=head2 FEATURES
87		99
88	Libev supports C<select>, C<poll>, the Linux-specific C<epoll>, the	100	Libev supports C<select>, C<poll>, the Linux-specific C<epoll>, the
89	BSD-specific C<kqueue> and the Solaris-specific event port mechanisms	101	BSD-specific C<kqueue> and the Solaris-specific event port mechanisms
90	for file descriptor events (C<ev_io>), the Linux C<inotify> interface	102	for file descriptor events (C<ev_io>), the Linux C<inotify> interface
91	(for C<ev_stat>), relative timers (C<ev_timer>), absolute timers	103	(for C<ev_stat>), Linux eventfd/signalfd (for faster and cleaner
92	with customised rescheduling (C<ev_periodic>), synchronous signals	104	inter-thread wakeup (C<ev_async>)/signal handling (C<ev_signal>)) relative
93	(C<ev_signal>), process status change events (C<ev_child>), and event	105	timers (C<ev_timer>), absolute timers with customised rescheduling
94	watchers dealing with the event loop mechanism itself (C<ev_idle>,	106	(C<ev_periodic>), synchronous signals (C<ev_signal>), process status
95	C<ev_embed>, C<ev_prepare> and C<ev_check> watchers) as well as	107	change events (C<ev_child>), and event watchers dealing with the event
96	file watchers (C<ev_stat>) and even limited support for fork events	108	loop mechanism itself (C<ev_idle>, C<ev_embed>, C<ev_prepare> and
97	(C<ev_fork>).	109	C<ev_check> watchers) as well as file watchers (C<ev_stat>) and even
		110	limited support for fork events (C<ev_fork>).
98		111
99	It also is quite fast (see this	112	It also is quite fast (see this
100	L<benchmark\|http://libev.schmorp.de/bench.html> comparing it to libevent	113	L<benchmark\|http://libev.schmorp.de/bench.html> comparing it to libevent
101	for example).	114	for example).
102		115
…		…
110	name C<loop> (which is always of type C<ev_loop *>) will not have	123	name C<loop> (which is always of type C<ev_loop *>) will not have
111	this argument.	124	this argument.
112		125
113	=head2 TIME REPRESENTATION	126	=head2 TIME REPRESENTATION
114		127
115	Libev represents time as a single floating point number, representing the	128	Libev represents time as a single floating point number, representing
116	(fractional) number of seconds since the (POSIX) epoch (somewhere near	129	the (fractional) number of seconds since the (POSIX) epoch (somewhere
117	the beginning of 1970, details are complicated, don't ask). This type is	130	near the beginning of 1970, details are complicated, don't ask). This
118	called C<ev_tstamp>, which is what you should use too. It usually aliases	131	type is called C<ev_tstamp>, which is what you should use too. It usually
119	to the C<double> type in C, and when you need to do any calculations on	132	aliases to the C<double> type in C. When you need to do any calculations
120	it, you should treat it as some floating point value. Unlike the name	133	on it, you should treat it as some floating point value. Unlike the name
121	component C<stamp> might indicate, it is also used for time differences	134	component C<stamp> might indicate, it is also used for time differences
122	throughout libev.	135	throughout libev.
123		136
124	=head1 ERROR HANDLING	137	=head1 ERROR HANDLING
125		138
…		…
349	forget about forgetting to tell libev about forking) when you use this	362	forget about forgetting to tell libev about forking) when you use this
350	flag.	363	flag.
351		364
352	This flag setting cannot be overridden or specified in the C<LIBEV_FLAGS>	365	This flag setting cannot be overridden or specified in the C<LIBEV_FLAGS>
353	environment variable.	366	environment variable.
		367
		368	=item C<EVFLAG_NOINOTIFY>
		369
		370	When this flag is specified, then libev will not attempt to use the
		371	I<inotify> API for it's C<ev_stat> watchers. Apart from debugging and
		372	testing, this flag can be useful to conserve inotify file descriptors, as
		373	otherwise each loop using C<ev_stat> watchers consumes one inotify handle.
		374
		375	=item C<EVFLAG_NOSIGFD>
		376
		377	When this flag is specified, then libev will not attempt to use the
		378	I<signalfd> API for it's C<ev_signal> (and C<ev_child>) watchers. This is
		379	probably only useful to work around any bugs in libev. Consequently, this
		380	flag might go away once the signalfd functionality is considered stable,
		381	so it's useful mostly in environment variables and not in program code.
354		382
355	=item C<EVBACKEND_SELECT> (value 1, portable select backend)	383	=item C<EVBACKEND_SELECT> (value 1, portable select backend)
356		384
357	This is your standard select(2) backend. Not I<completely> standard, as	385	This is your standard select(2) backend. Not I<completely> standard, as
358	libev tries to roll its own fd_set with no limits on the number of fds,	386	libev tries to roll its own fd_set with no limits on the number of fds,
…		…
506		534
507	It is definitely not recommended to use this flag.	535	It is definitely not recommended to use this flag.
508		536
509	=back	537	=back
510		538
511	If one or more of these are or'ed into the flags value, then only these	539	If one or more of the backend flags are or'ed into the flags value,
512	backends will be tried (in the reverse order as listed here). If none are	540	then only these backends will be tried (in the reverse order as listed
513	specified, all backends in C<ev_recommended_backends ()> will be tried.	541	here). If none are specified, all backends in C<ev_recommended_backends
		542	()> will be tried.
514		543
515	Example: This is the most typical usage.	544	Example: This is the most typical usage.
516		545
517	if (!ev_default_loop (0))	546	if (!ev_default_loop (0))
518	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");	547	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
…		…
561	as signal and child watchers) would need to be stopped manually.	590	as signal and child watchers) would need to be stopped manually.
562		591
563	In general it is not advisable to call this function except in the	592	In general it is not advisable to call this function except in the
564	rare occasion where you really need to free e.g. the signal handling	593	rare occasion where you really need to free e.g. the signal handling
565	pipe fds. If you need dynamically allocated loops it is better to use	594	pipe fds. If you need dynamically allocated loops it is better to use
566	C<ev_loop_new> and C<ev_loop_destroy>).	595	C<ev_loop_new> and C<ev_loop_destroy>.
567		596
568	=item ev_loop_destroy (loop)	597	=item ev_loop_destroy (loop)
569		598
570	Like C<ev_default_destroy>, but destroys an event loop created by an	599	Like C<ev_default_destroy>, but destroys an event loop created by an
571	earlier call to C<ev_loop_new>.	600	earlier call to C<ev_loop_new>.
…		…
609		638
610	This value can sometimes be useful as a generation counter of sorts (it	639	This value can sometimes be useful as a generation counter of sorts (it
611	"ticks" the number of loop iterations), as it roughly corresponds with	640	"ticks" the number of loop iterations), as it roughly corresponds with
612	C<ev_prepare> and C<ev_check> calls.	641	C<ev_prepare> and C<ev_check> calls.
613		642
		643	=item unsigned int ev_loop_depth (loop)
		644
		645	Returns the number of times C<ev_loop> was entered minus the number of
		646	times C<ev_loop> was exited, in other words, the recursion depth.
		647
		648	Outside C<ev_loop>, this number is zero. In a callback, this number is
		649	C<1>, unless C<ev_loop> was invoked recursively (or from another thread),
		650	in which case it is higher.
		651
		652	Leaving C<ev_loop> abnormally (setjmp/longjmp, cancelling the thread
		653	etc.), doesn't count as exit.
		654
614	=item unsigned int ev_backend (loop)	655	=item unsigned int ev_backend (loop)
615		656
616	Returns one of the C<EVBACKEND_*> flags indicating the event backend in	657	Returns one of the C<EVBACKEND_*> flags indicating the event backend in
617	use.	658	use.
618		659
…		…
632		673
633	This function is rarely useful, but when some event callback runs for a	674	This function is rarely useful, but when some event callback runs for a
634	very long time without entering the event loop, updating libev's idea of	675	very long time without entering the event loop, updating libev's idea of
635	the current time is a good idea.	676	the current time is a good idea.
636		677
637	See also "The special problem of time updates" in the C<ev_timer> section.	678	See also L<The special problem of time updates> in the C<ev_timer> section.
638		679
639	=item ev_suspend (loop)	680	=item ev_suspend (loop)
640		681
641	=item ev_resume (loop)	682	=item ev_resume (loop)
642		683
…		…
663	event loop time (see C<ev_now_update>).	704	event loop time (see C<ev_now_update>).
664		705
665	=item ev_loop (loop, int flags)	706	=item ev_loop (loop, int flags)
666		707
667	Finally, this is it, the event handler. This function usually is called	708	Finally, this is it, the event handler. This function usually is called
668	after you initialised all your watchers and you want to start handling	709	after you have initialised all your watchers and you want to start
669	events.	710	handling events.
670		711
671	If the flags argument is specified as C<0>, it will not return until	712	If the flags argument is specified as C<0>, it will not return until
672	either no event watchers are active anymore or C<ev_unloop> was called.	713	either no event watchers are active anymore or C<ev_unloop> was called.
673		714
674	Please note that an explicit C<ev_unloop> is usually better than	715	Please note that an explicit C<ev_unloop> is usually better than
…		…
799		840
800	By setting a higher I<io collect interval> you allow libev to spend more	841	By setting a higher I<io collect interval> you allow libev to spend more
801	time collecting I/O events, so you can handle more events per iteration,	842	time collecting I/O events, so you can handle more events per iteration,
802	at the cost of increasing latency. Timeouts (both C<ev_periodic> and	843	at the cost of increasing latency. Timeouts (both C<ev_periodic> and
803	C<ev_timer>) will be not affected. Setting this to a non-null value will	844	C<ev_timer>) will be not affected. Setting this to a non-null value will
804	introduce an additional C<ev_sleep ()> call into most loop iterations.	845	introduce an additional C<ev_sleep ()> call into most loop iterations. The
		846	sleep time ensures that libev will not poll for I/O events more often then
		847	once per this interval, on average.
805		848
806	Likewise, by setting a higher I<timeout collect interval> you allow libev	849	Likewise, by setting a higher I<timeout collect interval> you allow libev
807	to spend more time collecting timeouts, at the expense of increased	850	to spend more time collecting timeouts, at the expense of increased
808	latency/jitter/inexactness (the watcher callback will be called	851	latency/jitter/inexactness (the watcher callback will be called
809	later). C<ev_io> watchers will not be affected. Setting this to a non-null	852	later). C<ev_io> watchers will not be affected. Setting this to a non-null
…		…
811		854
812	Many (busy) programs can usually benefit by setting the I/O collect	855	Many (busy) programs can usually benefit by setting the I/O collect
813	interval to a value near C<0.1> or so, which is often enough for	856	interval to a value near C<0.1> or so, which is often enough for
814	interactive servers (of course not for games), likewise for timeouts. It	857	interactive servers (of course not for games), likewise for timeouts. It
815	usually doesn't make much sense to set it to a lower value than C<0.01>,	858	usually doesn't make much sense to set it to a lower value than C<0.01>,
816	as this approaches the timing granularity of most systems.	859	as this approaches the timing granularity of most systems. Note that if
		860	you do transactions with the outside world and you can't increase the
		861	parallelity, then this setting will limit your transaction rate (if you
		862	need to poll once per transaction and the I/O collect interval is 0.01,
		863	then you can't do more than 100 transations per second).
817		864
818	Setting the I<timeout collect interval> can improve the opportunity for	865	Setting the I<timeout collect interval> can improve the opportunity for
819	saving power, as the program will "bundle" timer callback invocations that	866	saving power, as the program will "bundle" timer callback invocations that
820	are "near" in time together, by delaying some, thus reducing the number of	867	are "near" in time together, by delaying some, thus reducing the number of
821	times the process sleeps and wakes up again. Another useful technique to	868	times the process sleeps and wakes up again. Another useful technique to
822	reduce iterations/wake-ups is to use C<ev_periodic> watchers and make sure	869	reduce iterations/wake-ups is to use C<ev_periodic> watchers and make sure
823	they fire on, say, one-second boundaries only.	870	they fire on, say, one-second boundaries only.
		871
		872	Example: we only need 0.1s timeout granularity, and we wish not to poll
		873	more often than 100 times per second:
		874
		875	ev_set_timeout_collect_interval (EV_DEFAULT_UC_ 0.1);
		876	ev_set_io_collect_interval (EV_DEFAULT_UC_ 0.01);
		877
		878	=item ev_invoke_pending (loop)
		879
		880	This call will simply invoke all pending watchers while resetting their
		881	pending state. Normally, C<ev_loop> does this automatically when required,
		882	but when overriding the invoke callback this call comes handy.
		883
		884	=item int ev_pending_count (loop)
		885
		886	Returns the number of pending watchers - zero indicates that no watchers
		887	are pending.
		888
		889	=item ev_set_invoke_pending_cb (loop, void (*invoke_pending_cb)(EV_P))
		890
		891	This overrides the invoke pending functionality of the loop: Instead of
		892	invoking all pending watchers when there are any, C<ev_loop> will call
		893	this callback instead. This is useful, for example, when you want to
		894	invoke the actual watchers inside another context (another thread etc.).
		895
		896	If you want to reset the callback, use C<ev_invoke_pending> as new
		897	callback.
		898
		899	=item ev_set_loop_release_cb (loop, void (release)(EV_P), void (acquire)(EV_P))
		900
		901	Sometimes you want to share the same loop between multiple threads. This
		902	can be done relatively simply by putting mutex_lock/unlock calls around
		903	each call to a libev function.
		904
		905	However, C<ev_loop> can run an indefinite time, so it is not feasible to
		906	wait for it to return. One way around this is to wake up the loop via
		907	C<ev_unloop> and C<av_async_send>, another way is to set these I<release>
		908	and I<acquire> callbacks on the loop.
		909
		910	When set, then C<release> will be called just before the thread is
		911	suspended waiting for new events, and C<acquire> is called just
		912	afterwards.
		913
		914	Ideally, C<release> will just call your mutex_unlock function, and
		915	C<acquire> will just call the mutex_lock function again.
		916
		917	While event loop modifications are allowed between invocations of
		918	C<release> and C<acquire> (that's their only purpose after all), no
		919	modifications done will affect the event loop, i.e. adding watchers will
		920	have no effect on the set of file descriptors being watched, or the time
		921	waited. USe an C<ev_async> watcher to wake up C<ev_loop> when you want it
		922	to take note of any changes you made.
		923
		924	In theory, threads executing C<ev_loop> will be async-cancel safe between
		925	invocations of C<release> and C<acquire>.
		926
		927	See also the locking example in the C<THREADS> section later in this
		928	document.
		929
		930	=item ev_set_userdata (loop, void *data)
		931
		932	=item ev_userdata (loop)
		933
		934	Set and retrieve a single C<void *> associated with a loop. When
		935	C<ev_set_userdata> has never been called, then C<ev_userdata> returns
		936	C<0.>
		937
		938	These two functions can be used to associate arbitrary data with a loop,
		939	and are intended solely for the C<invoke_pending_cb>, C<release> and
		940	C<acquire> callbacks described above, but of course can be (ab-)used for
		941	any other purpose as well.
824		942
825	=item ev_loop_verify (loop)	943	=item ev_loop_verify (loop)
826		944
827	This function only does something when C<EV_VERIFY> support has been	945	This function only does something when C<EV_VERIFY> support has been
828	compiled in, which is the default for non-minimal builds. It tries to go	946	compiled in, which is the default for non-minimal builds. It tries to go
…		…
1083	integer between C<EV_MAXPRI> (default: C<2>) and C<EV_MINPRI>	1201	integer between C<EV_MAXPRI> (default: C<2>) and C<EV_MINPRI>
1084	(default: C<-2>). Pending watchers with higher priority will be invoked	1202	(default: C<-2>). Pending watchers with higher priority will be invoked
1085	before watchers with lower priority, but priority will not keep watchers	1203	before watchers with lower priority, but priority will not keep watchers
1086	from being executed (except for C<ev_idle> watchers).	1204	from being executed (except for C<ev_idle> watchers).
1087		1205
1088	This means that priorities are I<only> used for ordering callback
1089	invocation after new events have been received. This is useful, for
1090	example, to reduce latency after idling, or more often, to bind two
1091	watchers on the same event and make sure one is called first.
1092
1093	If you need to suppress invocation when higher priority events are pending	1206	If you need to suppress invocation when higher priority events are pending
1094	you need to look at C<ev_idle> watchers, which provide this functionality.	1207	you need to look at C<ev_idle> watchers, which provide this functionality.
1095		1208
1096	You I<must not> change the priority of a watcher as long as it is active or	1209	You I<must not> change the priority of a watcher as long as it is active or
1097	pending.	1210	pending.
1098
1099	The default priority used by watchers when no priority has been set is
1100	always C<0>, which is supposed to not be too high and not be too low :).
1101		1211
1102	Setting a priority outside the range of C<EV_MINPRI> to C<EV_MAXPRI> is	1212	Setting a priority outside the range of C<EV_MINPRI> to C<EV_MAXPRI> is
1103	fine, as long as you do not mind that the priority value you query might	1213	fine, as long as you do not mind that the priority value you query might
1104	or might not have been clamped to the valid range.	1214	or might not have been clamped to the valid range.
		1215
		1216	The default priority used by watchers when no priority has been set is
		1217	always C<0>, which is supposed to not be too high and not be too low :).
		1218
		1219	See L<WATCHER PRIORITY MODELS>, below, for a more thorough treatment of
		1220	priorities.
1105		1221
1106	=item ev_invoke (loop, ev_TYPE *watcher, int revents)	1222	=item ev_invoke (loop, ev_TYPE *watcher, int revents)
1107		1223
1108	Invoke the C<watcher> with the given C<loop> and C<revents>. Neither	1224	Invoke the C<watcher> with the given C<loop> and C<revents>. Neither
1109	C<loop> nor C<revents> need to be valid as long as the watcher callback	1225	C<loop> nor C<revents> need to be valid as long as the watcher callback
…		…
1174	#include <stddef.h>	1290	#include <stddef.h>
1175		1291
1176	static void	1292	static void
1177	t1_cb (EV_P_ ev_timer *w, int revents)	1293	t1_cb (EV_P_ ev_timer *w, int revents)
1178	{	1294	{
1179	struct my_biggy big = (struct my_biggy *	1295	struct my_biggy big = (struct my_biggy *)
1180	(((char *)w) - offsetof (struct my_biggy, t1));	1296	(((char *)w) - offsetof (struct my_biggy, t1));
1181	}	1297	}
1182		1298
1183	static void	1299	static void
1184	t2_cb (EV_P_ ev_timer *w, int revents)	1300	t2_cb (EV_P_ ev_timer *w, int revents)
1185	{	1301	{
1186	struct my_biggy big = (struct my_biggy *	1302	struct my_biggy big = (struct my_biggy *)
1187	(((char *)w) - offsetof (struct my_biggy, t2));	1303	(((char *)w) - offsetof (struct my_biggy, t2));
1188	}	1304	}
		1305
		1306	=head2 WATCHER PRIORITY MODELS
		1307
		1308	Many event loops support I<watcher priorities>, which are usually small
		1309	integers that influence the ordering of event callback invocation
		1310	between watchers in some way, all else being equal.
		1311
		1312	In libev, Watcher priorities can be set using C<ev_set_priority>. See its
		1313	description for the more technical details such as the actual priority
		1314	range.
		1315
		1316	There are two common ways how these these priorities are being interpreted
		1317	by event loops:
		1318
		1319	In the more common lock-out model, higher priorities "lock out" invocation
		1320	of lower priority watchers, which means as long as higher priority
		1321	watchers receive events, lower priority watchers are not being invoked.
		1322
		1323	The less common only-for-ordering model uses priorities solely to order
		1324	callback invocation within a single event loop iteration: Higher priority
		1325	watchers are invoked before lower priority ones, but they all get invoked
		1326	before polling for new events.
		1327
		1328	Libev uses the second (only-for-ordering) model for all its watchers
		1329	except for idle watchers (which use the lock-out model).
		1330
		1331	The rationale behind this is that implementing the lock-out model for
		1332	watchers is not well supported by most kernel interfaces, and most event
		1333	libraries will just poll for the same events again and again as long as
		1334	their callbacks have not been executed, which is very inefficient in the
		1335	common case of one high-priority watcher locking out a mass of lower
		1336	priority ones.
		1337
		1338	Static (ordering) priorities are most useful when you have two or more
		1339	watchers handling the same resource: a typical usage example is having an
		1340	C<ev_io> watcher to receive data, and an associated C<ev_timer> to handle
		1341	timeouts. Under load, data might be received while the program handles
		1342	other jobs, but since timers normally get invoked first, the timeout
		1343	handler will be executed before checking for data. In that case, giving
		1344	the timer a lower priority than the I/O watcher ensures that I/O will be
		1345	handled first even under adverse conditions (which is usually, but not
		1346	always, what you want).
		1347
		1348	Since idle watchers use the "lock-out" model, meaning that idle watchers
		1349	will only be executed when no same or higher priority watchers have
		1350	received events, they can be used to implement the "lock-out" model when
		1351	required.
		1352
		1353	For example, to emulate how many other event libraries handle priorities,
		1354	you can associate an C<ev_idle> watcher to each such watcher, and in
		1355	the normal watcher callback, you just start the idle watcher. The real
		1356	processing is done in the idle watcher callback. This causes libev to
		1357	continously poll and process kernel event data for the watcher, but when
		1358	the lock-out case is known to be rare (which in turn is rare :), this is
		1359	workable.
		1360
		1361	Usually, however, the lock-out model implemented that way will perform
		1362	miserably under the type of load it was designed to handle. In that case,
		1363	it might be preferable to stop the real watcher before starting the
		1364	idle watcher, so the kernel will not have to process the event in case
		1365	the actual processing will be delayed for considerable time.
		1366
		1367	Here is an example of an I/O watcher that should run at a strictly lower
		1368	priority than the default, and which should only process data when no
		1369	other events are pending:
		1370
		1371	ev_idle idle; // actual processing watcher
		1372	ev_io io; // actual event watcher
		1373
		1374	static void
		1375	io_cb (EV_P_ ev_io *w, int revents)
		1376	{
		1377	// stop the I/O watcher, we received the event, but
		1378	// are not yet ready to handle it.
		1379	ev_io_stop (EV_A_ w);
		1380
		1381	// start the idle watcher to ahndle the actual event.
		1382	// it will not be executed as long as other watchers
		1383	// with the default priority are receiving events.
		1384	ev_idle_start (EV_A_ &idle);
		1385	}
		1386
		1387	static void
		1388	idle_cb (EV_P_ ev_idle *w, int revents)
		1389	{
		1390	// actual processing
		1391	read (STDIN_FILENO, ...);
		1392
		1393	// have to start the I/O watcher again, as
		1394	// we have handled the event
		1395	ev_io_start (EV_P_ &io);
		1396	}
		1397
		1398	// initialisation
		1399	ev_idle_init (&idle, idle_cb);
		1400	ev_io_init (&io, io_cb, STDIN_FILENO, EV_READ);
		1401	ev_io_start (EV_DEFAULT_ &io);
		1402
		1403	In the "real" world, it might also be beneficial to start a timer, so that
		1404	low-priority connections can not be locked out forever under load. This
		1405	enables your program to keep a lower latency for important connections
		1406	during short periods of high load, while not completely locking out less
		1407	important ones.
1189		1408
1190		1409
1191	=head1 WATCHER TYPES	1410	=head1 WATCHER TYPES
1192		1411
1193	This section describes each watcher in detail, but will not repeat	1412	This section describes each watcher in detail, but will not repeat
…		…
1219	descriptors to non-blocking mode is also usually a good idea (but not	1438	descriptors to non-blocking mode is also usually a good idea (but not
1220	required if you know what you are doing).	1439	required if you know what you are doing).
1221		1440
1222	If you cannot use non-blocking mode, then force the use of a	1441	If you cannot use non-blocking mode, then force the use of a
1223	known-to-be-good backend (at the time of this writing, this includes only	1442	known-to-be-good backend (at the time of this writing, this includes only
1224	C<EVBACKEND_SELECT> and C<EVBACKEND_POLL>).	1443	C<EVBACKEND_SELECT> and C<EVBACKEND_POLL>). The same applies to file
		1444	descriptors for which non-blocking operation makes no sense (such as
		1445	files) - libev doesn't guarentee any specific behaviour in that case.
1225		1446
1226	Another thing you have to watch out for is that it is quite easy to	1447	Another thing you have to watch out for is that it is quite easy to
1227	receive "spurious" readiness notifications, that is your callback might	1448	receive "spurious" readiness notifications, that is your callback might
1228	be called with C<EV_READ> but a subsequent C<read>(2) will actually block	1449	be called with C<EV_READ> but a subsequent C<read>(2) will actually block
1229	because there is no data. Not only are some backends known to create a	1450	because there is no data. Not only are some backends known to create a
…		…
1350	year, it will still time out after (roughly) one hour. "Roughly" because	1571	year, it will still time out after (roughly) one hour. "Roughly" because
1351	detecting time jumps is hard, and some inaccuracies are unavoidable (the	1572	detecting time jumps is hard, and some inaccuracies are unavoidable (the
1352	monotonic clock option helps a lot here).	1573	monotonic clock option helps a lot here).
1353		1574
1354	The callback is guaranteed to be invoked only I<after> its timeout has	1575	The callback is guaranteed to be invoked only I<after> its timeout has
1355	passed. If multiple timers become ready during the same loop iteration	1576	passed (not I<at>, so on systems with very low-resolution clocks this
1356	then the ones with earlier time-out values are invoked before ones with	1577	might introduce a small delay). If multiple timers become ready during the
1357	later time-out values (but this is no longer true when a callback calls	1578	same loop iteration then the ones with earlier time-out values are invoked
1358	C<ev_loop> recursively).	1579	before ones of the same priority with later time-out values (but this is
		1580	no longer true when a callback calls C<ev_loop> recursively).
1359		1581
1360	=head3 Be smart about timeouts	1582	=head3 Be smart about timeouts
1361		1583
1362	Many real-world problems involve some kind of timeout, usually for error	1584	Many real-world problems involve some kind of timeout, usually for error
1363	recovery. A typical example is an HTTP request - if the other side hangs,	1585	recovery. A typical example is an HTTP request - if the other side hangs,
…		…
1407	C<after> argument to C<ev_timer_set>, and only ever use the C<repeat>	1629	C<after> argument to C<ev_timer_set>, and only ever use the C<repeat>
1408	member and C<ev_timer_again>.	1630	member and C<ev_timer_again>.
1409		1631
1410	At start:	1632	At start:
1411		1633
1412	ev_timer_init (timer, callback);	1634	ev_init (timer, callback);
1413	timer->repeat = 60.;	1635	timer->repeat = 60.;
1414	ev_timer_again (loop, timer);	1636	ev_timer_again (loop, timer);
1415		1637
1416	Each time there is some activity:	1638	Each time there is some activity:
1417		1639
…		…
1479		1701
1480	To start the timer, simply initialise the watcher and set C<last_activity>	1702	To start the timer, simply initialise the watcher and set C<last_activity>
1481	to the current time (meaning we just have some activity :), then call the	1703	to the current time (meaning we just have some activity :), then call the
1482	callback, which will "do the right thing" and start the timer:	1704	callback, which will "do the right thing" and start the timer:
1483		1705
1484	ev_timer_init (timer, callback);	1706	ev_init (timer, callback);
1485	last_activity = ev_now (loop);	1707	last_activity = ev_now (loop);
1486	callback (loop, timer, EV_TIMEOUT);	1708	callback (loop, timer, EV_TIMEOUT);
1487		1709
1488	And when there is some activity, simply store the current time in	1710	And when there is some activity, simply store the current time in
1489	C<last_activity>, no libev calls at all:	1711	C<last_activity>, no libev calls at all:
…		…
1550		1772
1551	If the event loop is suspended for a long time, you can also force an	1773	If the event loop is suspended for a long time, you can also force an
1552	update of the time returned by C<ev_now ()> by calling C<ev_now_update	1774	update of the time returned by C<ev_now ()> by calling C<ev_now_update
1553	()>.	1775	()>.
1554		1776
		1777	=head3 The special problems of suspended animation
		1778
		1779	When you leave the server world it is quite customary to hit machines that
		1780	can suspend/hibernate - what happens to the clocks during such a suspend?
		1781
		1782	Some quick tests made with a Linux 2.6.28 indicate that a suspend freezes
		1783	all processes, while the clocks (C<times>, C<CLOCK_MONOTONIC>) continue
		1784	to run until the system is suspended, but they will not advance while the
		1785	system is suspended. That means, on resume, it will be as if the program
		1786	was frozen for a few seconds, but the suspend time will not be counted
		1787	towards C<ev_timer> when a monotonic clock source is used. The real time
		1788	clock advanced as expected, but if it is used as sole clocksource, then a
		1789	long suspend would be detected as a time jump by libev, and timers would
		1790	be adjusted accordingly.
		1791
		1792	I would not be surprised to see different behaviour in different between
		1793	operating systems, OS versions or even different hardware.
		1794
		1795	The other form of suspend (job control, or sending a SIGSTOP) will see a
		1796	time jump in the monotonic clocks and the realtime clock. If the program
		1797	is suspended for a very long time, and monotonic clock sources are in use,
		1798	then you can expect C<ev_timer>s to expire as the full suspension time
		1799	will be counted towards the timers. When no monotonic clock source is in
		1800	use, then libev will again assume a timejump and adjust accordingly.
		1801
		1802	It might be beneficial for this latter case to call C<ev_suspend>
		1803	and C<ev_resume> in code that handles C<SIGTSTP>, to at least get
		1804	deterministic behaviour in this case (you can do nothing against
		1805	C<SIGSTOP>).
		1806
1555	=head3 Watcher-Specific Functions and Data Members	1807	=head3 Watcher-Specific Functions and Data Members
1556		1808
1557	=over 4	1809	=over 4
1558		1810
1559	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)	1811	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)
…		…
1582	If the timer is started but non-repeating, stop it (as if it timed out).	1834	If the timer is started but non-repeating, stop it (as if it timed out).
1583		1835
1584	If the timer is repeating, either start it if necessary (with the	1836	If the timer is repeating, either start it if necessary (with the
1585	C<repeat> value), or reset the running timer to the C<repeat> value.	1837	C<repeat> value), or reset the running timer to the C<repeat> value.
1586		1838
1587	This sounds a bit complicated, see "Be smart about timeouts", above, for a	1839	This sounds a bit complicated, see L<Be smart about timeouts>, above, for a
1588	usage example.	1840	usage example.
		1841
		1842	=item ev_timer_remaining (loop, ev_timer *)
		1843
		1844	Returns the remaining time until a timer fires. If the timer is active,
		1845	then this time is relative to the current event loop time, otherwise it's
		1846	the timeout value currently configured.
		1847
		1848	That is, after an C<ev_timer_set (w, 5, 7)>, C<ev_timer_remaining> returns
		1849	C<5>. When the timer is started and one second passes, C<ev_timer_remain>
		1850	will return C<4>. When the timer expires and is restarted, it will return
		1851	roughly C<7> (likely slightly less as callback invocation takes some time,
		1852	too), and so on.
1589		1853
1590	=item ev_tstamp repeat [read-write]	1854	=item ev_tstamp repeat [read-write]
1591		1855
1592	The current C<repeat> value. Will be used each time the watcher times out	1856	The current C<repeat> value. Will be used each time the watcher times out
1593	or C<ev_timer_again> is called, and determines the next timeout (if any),	1857	or C<ev_timer_again> is called, and determines the next timeout (if any),
…		…
1829	Signal watchers will trigger an event when the process receives a specific	2093	Signal watchers will trigger an event when the process receives a specific
1830	signal one or more times. Even though signals are very asynchronous, libev	2094	signal one or more times. Even though signals are very asynchronous, libev
1831	will try it's best to deliver signals synchronously, i.e. as part of the	2095	will try it's best to deliver signals synchronously, i.e. as part of the
1832	normal event processing, like any other event.	2096	normal event processing, like any other event.
1833		2097
1834	If you want signals asynchronously, just use C<sigaction> as you would	2098	If you want signals to be delivered truly asynchronously, just use
1835	do without libev and forget about sharing the signal. You can even use	2099	C<sigaction> as you would do without libev and forget about sharing
1836	C<ev_async> from a signal handler to synchronously wake up an event loop.	2100	the signal. You can even use C<ev_async> from a signal handler to
		2101	synchronously wake up an event loop.
1837		2102
1838	You can configure as many watchers as you like per signal. Only when the	2103	You can configure as many watchers as you like for the same signal, but
		2104	only within the same loop, i.e. you can watch for C<SIGINT> in your
		2105	default loop and for C<SIGIO> in another loop, but you cannot watch for
		2106	C<SIGINT> in both the default loop and another loop at the same time. At
		2107	the moment, C<SIGCHLD> is permanently tied to the default loop.
		2108
1839	first watcher gets started will libev actually register a signal handler	2109	When the first watcher gets started will libev actually register something
1840	with the kernel (thus it coexists with your own signal handlers as long as	2110	with the kernel (thus it coexists with your own signal handlers as long as
1841	you don't register any with libev for the same signal). Similarly, when	2111	you don't register any with libev for the same signal).
1842	the last signal watcher for a signal is stopped, libev will reset the
1843	signal handler to SIG_DFL (regardless of what it was set to before).
1844		2112
1845	If possible and supported, libev will install its handlers with	2113	If possible and supported, libev will install its handlers with
1846	C<SA_RESTART> behaviour enabled, so system calls should not be unduly	2114	C<SA_RESTART> (or equivalent) behaviour enabled, so system calls should
1847	interrupted. If you have a problem with system calls getting interrupted by	2115	not be unduly interrupted. If you have a problem with system calls getting
1848	signals you can block all signals in an C<ev_check> watcher and unblock	2116	interrupted by signals you can block all signals in an C<ev_check> watcher
1849	them in an C<ev_prepare> watcher.	2117	and unblock them in an C<ev_prepare> watcher.
		2118
		2119	=head3 The special problem of inheritance over execve
		2120
		2121	Both the signal mask (C<sigprocmask>) and the signal disposition
		2122	(C<sigaction>) are unspecified after starting a signal watcher (and after
		2123	stopping it again), that is, libev might or might not block the signal,
		2124	and might or might not set or restore the installed signal handler.
		2125
		2126	While this does not matter for the signal disposition (libev never
		2127	sets signals to C<SIG_IGN>, so handlers will be reset to C<SIG_DFL> on
		2128	C<execve>), this matters for the signal mask: many programs do not expect
		2129	certain signals to be blocked.
		2130
		2131	This means that before calling C<exec> (from the child) you should reset
		2132	the signal mask to whatever "default" you expect (all clear is a good
		2133	choice usually).
		2134
		2135	The simplest way to ensure that the signal mask is reset in the child is
		2136	to install a fork handler with C<pthread_atfork> that resets it. That will
		2137	catch fork calls done by libraries (such as the libc) as well.
		2138
		2139	In current versions of libev, you can also ensure that the signal mask is
		2140	not blocking any signals (except temporarily, so thread users watch out)
		2141	by specifying the C<EVFLAG_NOSIGNALFD> when creating the event loop. This
		2142	is not guaranteed for future versions, however.
1850		2143
1851	=head3 Watcher-Specific Functions and Data Members	2144	=head3 Watcher-Specific Functions and Data Members
1852		2145
1853	=over 4	2146	=over 4
1854		2147
…		…
1886	some child status changes (most typically when a child of yours dies or	2179	some child status changes (most typically when a child of yours dies or
1887	exits). It is permissible to install a child watcher I<after> the child	2180	exits). It is permissible to install a child watcher I<after> the child
1888	has been forked (which implies it might have already exited), as long	2181	has been forked (which implies it might have already exited), as long
1889	as the event loop isn't entered (or is continued from a watcher), i.e.,	2182	as the event loop isn't entered (or is continued from a watcher), i.e.,
1890	forking and then immediately registering a watcher for the child is fine,	2183	forking and then immediately registering a watcher for the child is fine,
1891	but forking and registering a watcher a few event loop iterations later is	2184	but forking and registering a watcher a few event loop iterations later or
1892	not.	2185	in the next callback invocation is not.
1893		2186
1894	Only the default event loop is capable of handling signals, and therefore	2187	Only the default event loop is capable of handling signals, and therefore
1895	you can only register child watchers in the default event loop.	2188	you can only register child watchers in the default event loop.
1896		2189
		2190	Due to some design glitches inside libev, child watchers will always be
		2191	handled at maximum priority (their priority is set to C<EV_MAXPRI> by
		2192	libev)
		2193
1897	=head3 Process Interaction	2194	=head3 Process Interaction
1898		2195
1899	Libev grabs C<SIGCHLD> as soon as the default event loop is	2196	Libev grabs C<SIGCHLD> as soon as the default event loop is
1900	initialised. This is necessary to guarantee proper behaviour even if	2197	initialised. This is necessary to guarantee proper behaviour even if the
1901	the first child watcher is started after the child exits. The occurrence	2198	first child watcher is started after the child exits. The occurrence
1902	of C<SIGCHLD> is recorded asynchronously, but child reaping is done	2199	of C<SIGCHLD> is recorded asynchronously, but child reaping is done
1903	synchronously as part of the event loop processing. Libev always reaps all	2200	synchronously as part of the event loop processing. Libev always reaps all
1904	children, even ones not watched.	2201	children, even ones not watched.
1905		2202
1906	=head3 Overriding the Built-In Processing	2203	=head3 Overriding the Built-In Processing
…		…
1916	=head3 Stopping the Child Watcher	2213	=head3 Stopping the Child Watcher
1917		2214
1918	Currently, the child watcher never gets stopped, even when the	2215	Currently, the child watcher never gets stopped, even when the
1919	child terminates, so normally one needs to stop the watcher in the	2216	child terminates, so normally one needs to stop the watcher in the
1920	callback. Future versions of libev might stop the watcher automatically	2217	callback. Future versions of libev might stop the watcher automatically
1921	when a child exit is detected.	2218	when a child exit is detected (calling C<ev_child_stop> twice is not a
		2219	problem).
1922		2220
1923	=head3 Watcher-Specific Functions and Data Members	2221	=head3 Watcher-Specific Functions and Data Members
1924		2222
1925	=over 4	2223	=over 4
1926		2224
…		…
2252	// no longer anything immediate to do.	2550	// no longer anything immediate to do.
2253	}	2551	}
2254		2552
2255	ev_idle *idle_watcher = malloc (sizeof (ev_idle));	2553	ev_idle *idle_watcher = malloc (sizeof (ev_idle));
2256	ev_idle_init (idle_watcher, idle_cb);	2554	ev_idle_init (idle_watcher, idle_cb);
2257	ev_idle_start (loop, idle_cb);	2555	ev_idle_start (loop, idle_watcher);
2258		2556
2259		2557
2260	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!	2558	=head2 C<ev_prepare> and C<ev_check> - customise your event loop!
2261		2559
2262	Prepare and check watchers are usually (but not always) used in pairs:	2560	Prepare and check watchers are usually (but not always) used in pairs:
…		…
2355	struct pollfd fds [nfd];	2653	struct pollfd fds [nfd];
2356	// actual code will need to loop here and realloc etc.	2654	// actual code will need to loop here and realloc etc.
2357	adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));	2655	adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));
2358		2656
2359	/* the callback is illegal, but won't be called as we stop during check */	2657	/* the callback is illegal, but won't be called as we stop during check */
2360	ev_timer_init (&tw, 0, timeout * 1e-3);	2658	ev_timer_init (&tw, 0, timeout * 1e-3, 0.);
2361	ev_timer_start (loop, &tw);	2659	ev_timer_start (loop, &tw);
2362		2660
2363	// create one ev_io per pollfd	2661	// create one ev_io per pollfd
2364	for (int i = 0; i < nfd; ++i)	2662	for (int i = 0; i < nfd; ++i)
2365	{	2663	{
…		…
2595	event loop blocks next and before C<ev_check> watchers are being called,	2893	event loop blocks next and before C<ev_check> watchers are being called,
2596	and only in the child after the fork. If whoever good citizen calling	2894	and only in the child after the fork. If whoever good citizen calling
2597	C<ev_default_fork> cheats and calls it in the wrong process, the fork	2895	C<ev_default_fork> cheats and calls it in the wrong process, the fork
2598	handlers will be invoked, too, of course.	2896	handlers will be invoked, too, of course.
2599		2897
		2898	=head3 The special problem of life after fork - how is it possible?
		2899
		2900	Most uses of C<fork()> consist of forking, then some simple calls to ste
		2901	up/change the process environment, followed by a call to C<exec()>. This
		2902	sequence should be handled by libev without any problems.
		2903
		2904	This changes when the application actually wants to do event handling
		2905	in the child, or both parent in child, in effect "continuing" after the
		2906	fork.
		2907
		2908	The default mode of operation (for libev, with application help to detect
		2909	forks) is to duplicate all the state in the child, as would be expected
		2910	when I<either> the parent I<or> the child process continues.
		2911
		2912	When both processes want to continue using libev, then this is usually the
		2913	wrong result. In that case, usually one process (typically the parent) is
		2914	supposed to continue with all watchers in place as before, while the other
		2915	process typically wants to start fresh, i.e. without any active watchers.
		2916
		2917	The cleanest and most efficient way to achieve that with libev is to
		2918	simply create a new event loop, which of course will be "empty", and
		2919	use that for new watchers. This has the advantage of not touching more
		2920	memory than necessary, and thus avoiding the copy-on-write, and the
		2921	disadvantage of having to use multiple event loops (which do not support
		2922	signal watchers).
		2923
		2924	When this is not possible, or you want to use the default loop for
		2925	other reasons, then in the process that wants to start "fresh", call
		2926	C<ev_default_destroy ()> followed by C<ev_default_loop (...)>. Destroying
		2927	the default loop will "orphan" (not stop) all registered watchers, so you
		2928	have to be careful not to execute code that modifies those watchers. Note
		2929	also that in that case, you have to re-register any signal watchers.
		2930
2600	=head3 Watcher-Specific Functions and Data Members	2931	=head3 Watcher-Specific Functions and Data Members
2601		2932
2602	=over 4	2933	=over 4
2603		2934
2604	=item ev_fork_init (ev_signal *, callback)	2935	=item ev_fork_init (ev_signal *, callback)
…		…
3096	=item Ocaml	3427	=item Ocaml
3097		3428
3098	Erkki Seppala has written Ocaml bindings for libev, to be found at	3429	Erkki Seppala has written Ocaml bindings for libev, to be found at
3099	L<http://modeemi.cs.tut.fi/~flux/software/ocaml-ev/>.	3430	L<http://modeemi.cs.tut.fi/~flux/software/ocaml-ev/>.
3100		3431
		3432	=item Lua
		3433
		3434	Brian Maher has written a partial interface to libev
		3435	for lua (only C<ev_io> and C<ev_timer>), to be found at
		3436	L<http://github.com/brimworks/lua-ev>.
		3437
3101	=back	3438	=back
3102		3439
3103		3440
3104	=head1 MACRO MAGIC	3441	=head1 MACRO MAGIC
3105		3442
…		…
3271	keeps libev from including F<config.h>, and it also defines dummy	3608	keeps libev from including F<config.h>, and it also defines dummy
3272	implementations for some libevent functions (such as logging, which is not	3609	implementations for some libevent functions (such as logging, which is not
3273	supported). It will also not define any of the structs usually found in	3610	supported). It will also not define any of the structs usually found in
3274	F<event.h> that are not directly supported by the libev core alone.	3611	F<event.h> that are not directly supported by the libev core alone.
3275		3612
3276	In stanbdalone mode, libev will still try to automatically deduce the	3613	In standalone mode, libev will still try to automatically deduce the
3277	configuration, but has to be more conservative.	3614	configuration, but has to be more conservative.
3278		3615
3279	=item EV_USE_MONOTONIC	3616	=item EV_USE_MONOTONIC
3280		3617
3281	If defined to be C<1>, libev will try to detect the availability of the	3618	If defined to be C<1>, libev will try to detect the availability of the
…		…
3346	be used is the winsock select). This means that it will call	3683	be used is the winsock select). This means that it will call
3347	C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,	3684	C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,
3348	it is assumed that all these functions actually work on fds, even	3685	it is assumed that all these functions actually work on fds, even
3349	on win32. Should not be defined on non-win32 platforms.	3686	on win32. Should not be defined on non-win32 platforms.
3350		3687
3351	=item EV_FD_TO_WIN32_HANDLE	3688	=item EV_FD_TO_WIN32_HANDLE(fd)
3352		3689
3353	If C<EV_SELECT_IS_WINSOCKET> is enabled, then libev needs a way to map	3690	If C<EV_SELECT_IS_WINSOCKET> is enabled, then libev needs a way to map
3354	file descriptors to socket handles. When not defining this symbol (the	3691	file descriptors to socket handles. When not defining this symbol (the
3355	default), then libev will call C<_get_osfhandle>, which is usually	3692	default), then libev will call C<_get_osfhandle>, which is usually
3356	correct. In some cases, programs use their own file descriptor management,	3693	correct. In some cases, programs use their own file descriptor management,
3357	in which case they can provide this function to map fds to socket handles.	3694	in which case they can provide this function to map fds to socket handles.
		3695
		3696	=item EV_WIN32_HANDLE_TO_FD(handle)
		3697
		3698	If C<EV_SELECT_IS_WINSOCKET> then libev maps handles to file descriptors
		3699	using the standard C<_open_osfhandle> function. For programs implementing
		3700	their own fd to handle mapping, overwriting this function makes it easier
		3701	to do so. This can be done by defining this macro to an appropriate value.
		3702
		3703	=item EV_WIN32_CLOSE_FD(fd)
		3704
		3705	If programs implement their own fd to handle mapping on win32, then this
		3706	macro can be used to override the C<close> function, useful to unregister
		3707	file descriptors again. Note that the replacement function has to close
		3708	the underlying OS handle.
3358		3709
3359	=item EV_USE_POLL	3710	=item EV_USE_POLL
3360		3711
3361	If defined to be C<1>, libev will compile in support for the C<poll>(2)	3712	If defined to be C<1>, libev will compile in support for the C<poll>(2)
3362	backend. Otherwise it will be enabled on non-win32 platforms. It	3713	backend. Otherwise it will be enabled on non-win32 platforms. It
…		…
3494	defined to be C<0>, then they are not.	3845	defined to be C<0>, then they are not.
3495		3846
3496	=item EV_MINIMAL	3847	=item EV_MINIMAL
3497		3848
3498	If you need to shave off some kilobytes of code at the expense of some	3849	If you need to shave off some kilobytes of code at the expense of some
3499	speed, define this symbol to C<1>. Currently this is used to override some	3850	speed (but with the full API), define this symbol to C<1>. Currently this
3500	inlining decisions, saves roughly 30% code size on amd64. It also selects a	3851	is used to override some inlining decisions, saves roughly 30% code size
3501	much smaller 2-heap for timer management over the default 4-heap.	3852	on amd64. It also selects a much smaller 2-heap for timer management over
		3853	the default 4-heap.
		3854
		3855	You can save even more by disabling watcher types you do not need
		3856	and setting C<EV_MAXPRI> == C<EV_MINPRI>. Also, disabling C<assert>
		3857	(C<-DNDEBUG>) will usually reduce code size a lot.
		3858
		3859	Defining C<EV_MINIMAL> to C<2> will additionally reduce the core API to
		3860	provide a bare-bones event library. See C<ev.h> for details on what parts
		3861	of the API are still available, and do not complain if this subset changes
		3862	over time.
		3863
		3864	=item EV_NSIG
		3865
		3866	The highest supported signal number, +1 (or, the number of
		3867	signals): Normally, libev tries to deduce the maximum number of signals
		3868	automatically, but sometimes this fails, in which case it can be
		3869	specified. Also, using a lower number than detected (C<32> should be
		3870	good for about any system in existance) can save some memory, as libev
		3871	statically allocates some 12-24 bytes per signal number.
3502		3872
3503	=item EV_PID_HASHSIZE	3873	=item EV_PID_HASHSIZE
3504		3874
3505	C<ev_child> watchers use a small hash table to distribute workload by	3875	C<ev_child> watchers use a small hash table to distribute workload by
3506	pid. The default size is C<16> (or C<1> with C<EV_MINIMAL>), usually more	3876	pid. The default size is C<16> (or C<1> with C<EV_MINIMAL>), usually more
…		…
3692	default loop and triggering an C<ev_async> watcher from the default loop	4062	default loop and triggering an C<ev_async> watcher from the default loop
3693	watcher callback into the event loop interested in the signal.	4063	watcher callback into the event loop interested in the signal.
3694		4064
3695	=back	4065	=back
3696		4066
		4067	=head4 THREAD LOCKING EXAMPLE
		4068
		4069	Here is a fictitious example of how to run an event loop in a different
		4070	thread than where callbacks are being invoked and watchers are
		4071	created/added/removed.
		4072
		4073	For a real-world example, see the C<EV::Loop::Async> perl module,
		4074	which uses exactly this technique (which is suited for many high-level
		4075	languages).
		4076
		4077	The example uses a pthread mutex to protect the loop data, a condition
		4078	variable to wait for callback invocations, an async watcher to notify the
		4079	event loop thread and an unspecified mechanism to wake up the main thread.
		4080
		4081	First, you need to associate some data with the event loop:
		4082
		4083	typedef struct {
		4084	mutex_t lock; /* global loop lock */
		4085	ev_async async_w;
		4086	thread_t tid;
		4087	cond_t invoke_cv;
		4088	} userdata;
		4089
		4090	void prepare_loop (EV_P)
		4091	{
		4092	// for simplicity, we use a static userdata struct.
		4093	static userdata u;
		4094
		4095	ev_async_init (&u->async_w, async_cb);
		4096	ev_async_start (EV_A_ &u->async_w);
		4097
		4098	pthread_mutex_init (&u->lock, 0);
		4099	pthread_cond_init (&u->invoke_cv, 0);
		4100
		4101	// now associate this with the loop
		4102	ev_set_userdata (EV_A_ u);
		4103	ev_set_invoke_pending_cb (EV_A_ l_invoke);
		4104	ev_set_loop_release_cb (EV_A_ l_release, l_acquire);
		4105
		4106	// then create the thread running ev_loop
		4107	pthread_create (&u->tid, 0, l_run, EV_A);
		4108	}
		4109
		4110	The callback for the C<ev_async> watcher does nothing: the watcher is used
		4111	solely to wake up the event loop so it takes notice of any new watchers
		4112	that might have been added:
		4113
		4114	static void
		4115	async_cb (EV_P_ ev_async *w, int revents)
		4116	{
		4117	// just used for the side effects
		4118	}
		4119
		4120	The C<l_release> and C<l_acquire> callbacks simply unlock/lock the mutex
		4121	protecting the loop data, respectively.
		4122
		4123	static void
		4124	l_release (EV_P)
		4125	{
		4126	userdata *u = ev_userdata (EV_A);
		4127	pthread_mutex_unlock (&u->lock);
		4128	}
		4129
		4130	static void
		4131	l_acquire (EV_P)
		4132	{
		4133	userdata *u = ev_userdata (EV_A);
		4134	pthread_mutex_lock (&u->lock);
		4135	}
		4136
		4137	The event loop thread first acquires the mutex, and then jumps straight
		4138	into C<ev_loop>:
		4139
		4140	void *
		4141	l_run (void *thr_arg)
		4142	{
		4143	struct ev_loop loop = (struct ev_loop )thr_arg;
		4144
		4145	l_acquire (EV_A);
		4146	pthread_setcanceltype (PTHREAD_CANCEL_ASYNCHRONOUS, 0);
		4147	ev_loop (EV_A_ 0);
		4148	l_release (EV_A);
		4149
		4150	return 0;
		4151	}
		4152
		4153	Instead of invoking all pending watchers, the C<l_invoke> callback will
		4154	signal the main thread via some unspecified mechanism (signals? pipe
		4155	writes? C<Async::Interrupt>?) and then waits until all pending watchers
		4156	have been called (in a while loop because a) spurious wakeups are possible
		4157	and b) skipping inter-thread-communication when there are no pending
		4158	watchers is very beneficial):
		4159
		4160	static void
		4161	l_invoke (EV_P)
		4162	{
		4163	userdata *u = ev_userdata (EV_A);
		4164
		4165	while (ev_pending_count (EV_A))
		4166	{
		4167	wake_up_other_thread_in_some_magic_or_not_so_magic_way ();
		4168	pthread_cond_wait (&u->invoke_cv, &u->lock);
		4169	}
		4170	}
		4171
		4172	Now, whenever the main thread gets told to invoke pending watchers, it
		4173	will grab the lock, call C<ev_invoke_pending> and then signal the loop
		4174	thread to continue:
		4175
		4176	static void
		4177	real_invoke_pending (EV_P)
		4178	{
		4179	userdata *u = ev_userdata (EV_A);
		4180
		4181	pthread_mutex_lock (&u->lock);
		4182	ev_invoke_pending (EV_A);
		4183	pthread_cond_signal (&u->invoke_cv);
		4184	pthread_mutex_unlock (&u->lock);
		4185	}
		4186
		4187	Whenever you want to start/stop a watcher or do other modifications to an
		4188	event loop, you will now have to lock:
		4189
		4190	ev_timer timeout_watcher;
		4191	userdata *u = ev_userdata (EV_A);
		4192
		4193	ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);
		4194
		4195	pthread_mutex_lock (&u->lock);
		4196	ev_timer_start (EV_A_ &timeout_watcher);
		4197	ev_async_send (EV_A_ &u->async_w);
		4198	pthread_mutex_unlock (&u->lock);
		4199
		4200	Note that sending the C<ev_async> watcher is required because otherwise
		4201	an event loop currently blocking in the kernel will have no knowledge
		4202	about the newly added timer. By waking up the loop it will pick up any new
		4203	watchers in the next event loop iteration.
		4204
3697	=head3 COROUTINES	4205	=head3 COROUTINES
3698		4206
3699	Libev is very accommodating to coroutines ("cooperative threads"):	4207	Libev is very accommodating to coroutines ("cooperative threads"):
3700	libev fully supports nesting calls to its functions from different	4208	libev fully supports nesting calls to its functions from different
3701	coroutines (e.g. you can call C<ev_loop> on the same loop from two	4209	coroutines (e.g. you can call C<ev_loop> on the same loop from two
3702	different coroutines, and switch freely between both coroutines running the	4210	different coroutines, and switch freely between both coroutines running
3703	loop, as long as you don't confuse yourself). The only exception is that	4211	the loop, as long as you don't confuse yourself). The only exception is
3704	you must not do this from C<ev_periodic> reschedule callbacks.	4212	that you must not do this from C<ev_periodic> reschedule callbacks.
3705		4213
3706	Care has been taken to ensure that libev does not keep local state inside	4214	Care has been taken to ensure that libev does not keep local state inside
3707	C<ev_loop>, and other calls do not usually allow for coroutine switches as	4215	C<ev_loop>, and other calls do not usually allow for coroutine switches as
3708	they do not call any callbacks.	4216	they do not call any callbacks.
3709		4217
…		…
3786	way (note also that glib is the slowest event library known to man).	4294	way (note also that glib is the slowest event library known to man).
3787		4295
3788	There is no supported compilation method available on windows except	4296	There is no supported compilation method available on windows except
3789	embedding it into other applications.	4297	embedding it into other applications.
3790		4298
		4299	Sensible signal handling is officially unsupported by Microsoft - libev
		4300	tries its best, but under most conditions, signals will simply not work.
		4301
3791	Not a libev limitation but worth mentioning: windows apparently doesn't	4302	Not a libev limitation but worth mentioning: windows apparently doesn't
3792	accept large writes: instead of resulting in a partial write, windows will	4303	accept large writes: instead of resulting in a partial write, windows will
3793	either accept everything or return C<ENOBUFS> if the buffer is too large,	4304	either accept everything or return C<ENOBUFS> if the buffer is too large,
3794	so make sure you only write small amounts into your sockets (less than a	4305	so make sure you only write small amounts into your sockets (less than a
3795	megabyte seems safe, but this apparently depends on the amount of memory	4306	megabyte seems safe, but this apparently depends on the amount of memory
…		…
3799	the abysmal performance of winsockets, using a large number of sockets	4310	the abysmal performance of winsockets, using a large number of sockets
3800	is not recommended (and not reasonable). If your program needs to use	4311	is not recommended (and not reasonable). If your program needs to use
3801	more than a hundred or so sockets, then likely it needs to use a totally	4312	more than a hundred or so sockets, then likely it needs to use a totally
3802	different implementation for windows, as libev offers the POSIX readiness	4313	different implementation for windows, as libev offers the POSIX readiness
3803	notification model, which cannot be implemented efficiently on windows	4314	notification model, which cannot be implemented efficiently on windows
3804	(Microsoft monopoly games).	4315	(due to Microsoft monopoly games).
3805		4316
3806	A typical way to use libev under windows is to embed it (see the embedding	4317	A typical way to use libev under windows is to embed it (see the embedding
3807	section for details) and use the following F<evwrap.h> header file instead	4318	section for details) and use the following F<evwrap.h> header file instead
3808	of F<ev.h>:	4319	of F<ev.h>:
3809		4320
…		…
3845		4356
3846	Early versions of winsocket's select only supported waiting for a maximum	4357	Early versions of winsocket's select only supported waiting for a maximum
3847	of C<64> handles (probably owning to the fact that all windows kernels	4358	of C<64> handles (probably owning to the fact that all windows kernels
3848	can only wait for C<64> things at the same time internally; Microsoft	4359	can only wait for C<64> things at the same time internally; Microsoft
3849	recommends spawning a chain of threads and wait for 63 handles and the	4360	recommends spawning a chain of threads and wait for 63 handles and the
3850	previous thread in each. Great).	4361	previous thread in each. Sounds great!).
3851		4362
3852	Newer versions support more handles, but you need to define C<FD_SETSIZE>	4363	Newer versions support more handles, but you need to define C<FD_SETSIZE>
3853	to some high number (e.g. C<2048>) before compiling the winsocket select	4364	to some high number (e.g. C<2048>) before compiling the winsocket select
3854	call (which might be in libev or elsewhere, for example, perl does its own	4365	call (which might be in libev or elsewhere, for example, perl and many
3855	select emulation on windows).	4366	other interpreters do their own select emulation on windows).
3856		4367
3857	Another limit is the number of file descriptors in the Microsoft runtime	4368	Another limit is the number of file descriptors in the Microsoft runtime
3858	libraries, which by default is C<64> (there must be a hidden I<64> fetish	4369	libraries, which by default is C<64> (there must be a hidden I<64>
3859	or something like this inside Microsoft). You can increase this by calling	4370	fetish or something like this inside Microsoft). You can increase this
3860	C<_setmaxstdio>, which can increase this limit to C<2048> (another	4371	by calling C<_setmaxstdio>, which can increase this limit to C<2048>
3861	arbitrary limit), but is broken in many versions of the Microsoft runtime	4372	(another arbitrary limit), but is broken in many versions of the Microsoft
3862	libraries.
3863
3864	This might get you to about C<512> or C<2048> sockets (depending on	4373	runtime libraries. This might get you to about C<512> or C<2048> sockets
3865	windows version and/or the phase of the moon). To get more, you need to	4374	(depending on windows version and/or the phase of the moon). To get more,
3866	wrap all I/O functions and provide your own fd management, but the cost of	4375	you need to wrap all I/O functions and provide your own fd management, but
3867	calling select (O(n²)) will likely make this unworkable.	4376	the cost of calling select (O(n²)) will likely make this unworkable.
3868		4377
3869	=back	4378	=back
3870		4379
3871	=head2 PORTABILITY REQUIREMENTS	4380	=head2 PORTABILITY REQUIREMENTS
3872		4381
…		…
3915	=item C<double> must hold a time value in seconds with enough accuracy	4424	=item C<double> must hold a time value in seconds with enough accuracy
3916		4425
3917	The type C<double> is used to represent timestamps. It is required to	4426	The type C<double> is used to represent timestamps. It is required to
3918	have at least 51 bits of mantissa (and 9 bits of exponent), which is good	4427	have at least 51 bits of mantissa (and 9 bits of exponent), which is good
3919	enough for at least into the year 4000. This requirement is fulfilled by	4428	enough for at least into the year 4000. This requirement is fulfilled by
3920	implementations implementing IEEE 754 (basically all existing ones).	4429	implementations implementing IEEE 754, which is basically all existing
		4430	ones. With IEEE 754 doubles, you get microsecond accuracy until at least
		4431	2200.
3921		4432
3922	=back	4433	=back
3923		4434
3924	If you know of other additional requirements drop me a note.	4435	If you know of other additional requirements drop me a note.
3925		4436
…		…
3993	involves iterating over all running async watchers or all signal numbers.	4504	involves iterating over all running async watchers or all signal numbers.
3994		4505
3995	=back	4506	=back
3996		4507
3997		4508
		4509	=head1 GLOSSARY
		4510
		4511	=over 4
		4512
		4513	=item active
		4514
		4515	A watcher is active as long as it has been started (has been attached to
		4516	an event loop) but not yet stopped (disassociated from the event loop).
		4517
		4518	=item application
		4519
		4520	In this document, an application is whatever is using libev.
		4521
		4522	=item callback
		4523
		4524	The address of a function that is called when some event has been
		4525	detected. Callbacks are being passed the event loop, the watcher that
		4526	received the event, and the actual event bitset.
		4527
		4528	=item callback invocation
		4529
		4530	The act of calling the callback associated with a watcher.
		4531
		4532	=item event
		4533
		4534	A change of state of some external event, such as data now being available
		4535	for reading on a file descriptor, time having passed or simply not having
		4536	any other events happening anymore.
		4537
		4538	In libev, events are represented as single bits (such as C<EV_READ> or
		4539	C<EV_TIMEOUT>).
		4540
		4541	=item event library
		4542
		4543	A software package implementing an event model and loop.
		4544
		4545	=item event loop
		4546
		4547	An entity that handles and processes external events and converts them
		4548	into callback invocations.
		4549
		4550	=item event model
		4551
		4552	The model used to describe how an event loop handles and processes
		4553	watchers and events.
		4554
		4555	=item pending
		4556
		4557	A watcher is pending as soon as the corresponding event has been detected,
		4558	and stops being pending as soon as the watcher will be invoked or its
		4559	pending status is explicitly cleared by the application.
		4560
		4561	A watcher can be pending, but not active. Stopping a watcher also clears
		4562	its pending status.
		4563
		4564	=item real time
		4565
		4566	The physical time that is observed. It is apparently strictly monotonic :)
		4567
		4568	=item wall-clock time
		4569
		4570	The time and date as shown on clocks. Unlike real time, it can actually
		4571	be wrong and jump forwards and backwards, e.g. when the you adjust your
		4572	clock.
		4573
		4574	=item watcher
		4575
		4576	A data structure that describes interest in certain events. Watchers need
		4577	to be started (attached to an event loop) before they can receive events.
		4578
		4579	=item watcher invocation
		4580
		4581	The act of calling the callback associated with a watcher.
		4582
		4583	=back
		4584
3998	=head1 AUTHOR	4585	=head1 AUTHOR
3999		4586
4000	Marc Lehmann <libev@schmorp.de>, with repeated corrections by Mikael Magnusson.	4587	Marc Lehmann <libev@schmorp.de>, with repeated corrections by Mikael Magnusson.
4001		4588

Diff Legend

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

Comparing libev/ev.pod (file contents): Revision 1.231 by root, Wed Apr 15 19:35:53 2009 UTC vs. Revision 1.270 by root, Fri Sep 18 21:02:12 2009 UTC

Diff Legend

Comparing libev/ev.pod (file contents):
Revision 1.231 by root, Wed Apr 15 19:35:53 2009 UTC vs.
Revision 1.270 by root, Fri Sep 18 21:02:12 2009 UTC