ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/libev/ev.pod

Comparing libev/ev.pod (file contents):
Revision 1.66 by root, Mon Dec 3 13:41:25 2007 UTC vs.
Revision 1.117 by root, Wed Jan 9 04:15:39 2008 UTC

…		…
4		4
5	=head1 SYNOPSIS	5	=head1 SYNOPSIS
6		6
7	#include <ev.h>	7	#include <ev.h>
8		8
9	=head1 EXAMPLE PROGRAM	9	=head2 EXAMPLE PROGRAM
10		10
11	#include <ev.h>	11	#include <ev.h>
12		12
13	ev_io stdin_watcher;	13	ev_io stdin_watcher;
14	ev_timer timeout_watcher;	14	ev_timer timeout_watcher;
…		…
48	return 0;	48	return 0;
49	}	49	}
50		50
51	=head1 DESCRIPTION	51	=head1 DESCRIPTION
52		52
		53	The newest version of this document is also available as a html-formatted
		54	web page you might find easier to navigate when reading it for the first
		55	time: L<http://cvs.schmorp.de/libev/ev.html>.
		56
53	Libev is an event loop: you register interest in certain events (such as a	57	Libev is an event loop: you register interest in certain events (such as a
54	file descriptor being readable or a timeout occuring), and it will manage	58	file descriptor being readable or a timeout occurring), and it will manage
55	these event sources and provide your program with events.	59	these event sources and provide your program with events.
56		60
57	To do this, it must take more or less complete control over your process	61	To do this, it must take more or less complete control over your process
58	(or thread) by executing the I<event loop> handler, and will then	62	(or thread) by executing the I<event loop> handler, and will then
59	communicate events via a callback mechanism.	63	communicate events via a callback mechanism.
…		…
61	You register interest in certain events by registering so-called I<event	65	You register interest in certain events by registering so-called I<event
62	watchers>, which are relatively small C structures you initialise with the	66	watchers>, which are relatively small C structures you initialise with the
63	details of the event, and then hand it over to libev by I<starting> the	67	details of the event, and then hand it over to libev by I<starting> the
64	watcher.	68	watcher.
65		69
66	=head1 FEATURES	70	=head2 FEATURES
67		71
68	Libev supports C<select>, C<poll>, the Linux-specific C<epoll>, the	72	Libev supports C<select>, C<poll>, the Linux-specific C<epoll>, the
69	BSD-specific C<kqueue> and the Solaris-specific event port mechanisms	73	BSD-specific C<kqueue> and the Solaris-specific event port mechanisms
70	for file descriptor events (C<ev_io>), the Linux C<inotify> interface	74	for file descriptor events (C<ev_io>), the Linux C<inotify> interface
71	(for C<ev_stat>), relative timers (C<ev_timer>), absolute timers	75	(for C<ev_stat>), relative timers (C<ev_timer>), absolute timers
…		…
78		82
79	It also is quite fast (see this	83	It also is quite fast (see this
80	L<benchmark|http://libev.schmorp.de/bench.html> comparing it to libevent	84	L<benchmark|http://libev.schmorp.de/bench.html> comparing it to libevent
81	for example).	85	for example).
82		86
83	=head1 CONVENTIONS	87	=head2 CONVENTIONS
84		88
85	Libev is very configurable. In this manual the default configuration will	89	Libev is very configurable. In this manual the default configuration will
86	be described, which supports multiple event loops. For more info about	90	be described, which supports multiple event loops. For more info about
87	various configuration options please have a look at B<EMBED> section in	91	various configuration options please have a look at B<EMBED> section in
88	this manual. If libev was configured without support for multiple event	92	this manual. If libev was configured without support for multiple event
89	loops, then all functions taking an initial argument of name C<loop>	93	loops, then all functions taking an initial argument of name C<loop>
90	(which is always of type C<struct ev_loop *>) will not have this argument.	94	(which is always of type C<struct ev_loop *>) will not have this argument.
91		95
92	=head1 TIME REPRESENTATION	96	=head2 TIME REPRESENTATION
93		97
94	Libev represents time as a single floating point number, representing the	98	Libev represents time as a single floating point number, representing the
95	(fractional) number of seconds since the (POSIX) epoch (somewhere near	99	(fractional) number of seconds since the (POSIX) epoch (somewhere near
96	the beginning of 1970, details are complicated, don't ask). This type is	100	the beginning of 1970, details are complicated, don't ask). This type is
97	called C<ev_tstamp>, which is what you should use too. It usually aliases	101	called C<ev_tstamp>, which is what you should use too. It usually aliases
98	to the C<double> type in C, and when you need to do any calculations on	102	to the C<double> type in C, and when you need to do any calculations on
99	it, you should treat it as such.	103	it, you should treat it as some floatingpoint value. Unlike the name
		104	component C<stamp> might indicate, it is also used for time differences
		105	throughout libev.
100		106
101	=head1 GLOBAL FUNCTIONS	107	=head1 GLOBAL FUNCTIONS
102		108
103	These functions can be called anytime, even before initialising the	109	These functions can be called anytime, even before initialising the
104	library in any way.	110	library in any way.
…		…
109		115
110	Returns the current time as libev would use it. Please note that the	116	Returns the current time as libev would use it. Please note that the
111	C<ev_now> function is usually faster and also often returns the timestamp	117	C<ev_now> function is usually faster and also often returns the timestamp
112	you actually want to know.	118	you actually want to know.
113		119
		120	=item ev_sleep (ev_tstamp interval)
		121
		122	Sleep for the given interval: The current thread will be blocked until
		123	either it is interrupted or the given time interval has passed. Basically
		124	this is a subsecond-resolution C<sleep ()>.
		125
114	=item int ev_version_major ()	126	=item int ev_version_major ()
115		127
116	=item int ev_version_minor ()	128	=item int ev_version_minor ()
117		129
118	You can find out the major and minor version numbers of the library	130	You can find out the major and minor ABI version numbers of the library
119	you linked against by calling the functions C<ev_version_major> and	131	you linked against by calling the functions C<ev_version_major> and
120	C<ev_version_minor>. If you want, you can compare against the global	132	C<ev_version_minor>. If you want, you can compare against the global
121	symbols C<EV_VERSION_MAJOR> and C<EV_VERSION_MINOR>, which specify the	133	symbols C<EV_VERSION_MAJOR> and C<EV_VERSION_MINOR>, which specify the
122	version of the library your program was compiled against.	134	version of the library your program was compiled against.
123		135
		136	These version numbers refer to the ABI version of the library, not the
		137	release version.
		138
124	Usually, it's a good idea to terminate if the major versions mismatch,	139	Usually, it's a good idea to terminate if the major versions mismatch,
125	as this indicates an incompatible change. Minor versions are usually	140	as this indicates an incompatible change. Minor versions are usually
126	compatible to older versions, so a larger minor version alone is usually	141	compatible to older versions, so a larger minor version alone is usually
127	not a problem.	142	not a problem.
128		143
129	Example: Make sure we haven't accidentally been linked against the wrong	144	Example: Make sure we haven't accidentally been linked against the wrong
130	version.	145	version.
…		…
291	=item C<EVBACKEND_SELECT> (value 1, portable select backend)	306	=item C<EVBACKEND_SELECT> (value 1, portable select backend)
292		307
293	This is your standard select(2) backend. Not I<completely> standard, as	308	This is your standard select(2) backend. Not I<completely> standard, as
294	libev tries to roll its own fd_set with no limits on the number of fds,	309	libev tries to roll its own fd_set with no limits on the number of fds,
295	but if that fails, expect a fairly low limit on the number of fds when	310	but if that fails, expect a fairly low limit on the number of fds when
296	using this backend. It doesn't scale too well (O(highest_fd)), but its usually	311	using this backend. It doesn't scale too well (O(highest_fd)), but its
297	the fastest backend for a low number of fds.	312	usually the fastest backend for a low number of (low-numbered :) fds.
		313
		314	To get good performance out of this backend you need a high amount of
		315	parallelity (most of the file descriptors should be busy). If you are
		316	writing a server, you should C<accept ()> in a loop to accept as many
		317	connections as possible during one iteration. You might also want to have
		318	a look at C<ev_set_io_collect_interval ()> to increase the amount of
		319	readyness notifications you get per iteration.
298		320
299	=item C<EVBACKEND_POLL> (value 2, poll backend, available everywhere except on windows)	321	=item C<EVBACKEND_POLL> (value 2, poll backend, available everywhere except on windows)
300		322
301	And this is your standard poll(2) backend. It's more complicated than	323	And this is your standard poll(2) backend. It's more complicated
302	select, but handles sparse fds better and has no artificial limit on the	324	than select, but handles sparse fds better and has no artificial
303	number of fds you can use (except it will slow down considerably with a	325	limit on the number of fds you can use (except it will slow down
304	lot of inactive fds). It scales similarly to select, i.e. O(total_fds).	326	considerably with a lot of inactive fds). It scales similarly to select,
		327	i.e. O(total_fds). See the entry for C<EVBACKEND_SELECT>, above, for
		328	performance tips.
305		329
306	=item C<EVBACKEND_EPOLL> (value 4, Linux)	330	=item C<EVBACKEND_EPOLL> (value 4, Linux)
307		331
308	For few fds, this backend is a bit little slower than poll and select,	332	For few fds, this backend is a bit little slower than poll and select,
309	but it scales phenomenally better. While poll and select usually scale like	333	but it scales phenomenally better. While poll and select usually scale
310	O(total_fds) where n is the total number of fds (or the highest fd), epoll scales	334	like O(total_fds) where n is the total number of fds (or the highest fd),
311	either O(1) or O(active_fds).	335	epoll scales either O(1) or O(active_fds). The epoll design has a number
		336	of shortcomings, such as silently dropping events in some hard-to-detect
		337	cases and rewiring a syscall per fd change, no fork support and bad
		338	support for dup.
312		339
313	While stopping and starting an I/O watcher in the same iteration will	340	While stopping, setting and starting an I/O watcher in the same iteration
314	result in some caching, there is still a syscall per such incident	341	will result in some caching, there is still a syscall per such incident
315	(because the fd could point to a different file description now), so its	342	(because the fd could point to a different file description now), so its
316	best to avoid that. Also, dup()ed file descriptors might not work very	343	best to avoid that. Also, C<dup ()>'ed file descriptors might not work
317	well if you register events for both fds.	344	very well if you register events for both fds.
318		345
319	Please note that epoll sometimes generates spurious notifications, so you	346	Please note that epoll sometimes generates spurious notifications, so you
320	need to use non-blocking I/O or other means to avoid blocking when no data	347	need to use non-blocking I/O or other means to avoid blocking when no data
321	(or space) is available.	348	(or space) is available.
322		349
		350	Best performance from this backend is achieved by not unregistering all
		351	watchers for a file descriptor until it has been closed, if possible, i.e.
		352	keep at least one watcher active per fd at all times.
		353
		354	While nominally embeddeble in other event loops, this feature is broken in
		355	all kernel versions tested so far.
		356
323	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)	357	=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)
324		358
325	Kqueue deserves special mention, as at the time of this writing, it	359	Kqueue deserves special mention, as at the time of this writing, it
326	was broken on all BSDs except NetBSD (usually it doesn't work with	360	was broken on all BSDs except NetBSD (usually it doesn't work reliably
327	anything but sockets and pipes, except on Darwin, where of course its	361	with anything but sockets and pipes, except on Darwin, where of course
328	completely useless). For this reason its not being "autodetected"	362	it's completely useless). For this reason it's not being "autodetected"
329	unless you explicitly specify it explicitly in the flags (i.e. using	363	unless you explicitly specify it explicitly in the flags (i.e. using
330	C<EVBACKEND_KQUEUE>).	364	C<EVBACKEND_KQUEUE>) or libev was compiled on a known-to-be-good (-enough)
		365	system like NetBSD.
		366
		367	You still can embed kqueue into a normal poll or select backend and use it
		368	only for sockets (after having made sure that sockets work with kqueue on
		369	the target platform). See C<ev_embed> watchers for more info.
331		370
332	It scales in the same way as the epoll backend, but the interface to the	371	It scales in the same way as the epoll backend, but the interface to the
333	kernel is more efficient (which says nothing about its actual speed, of	372	kernel is more efficient (which says nothing about its actual speed, of
334	course). While starting and stopping an I/O watcher does not cause an	373	course). While stopping, setting and starting an I/O watcher does never
335	extra syscall as with epoll, it still adds up to four event changes per	374	cause an extra syscall as with C<EVBACKEND_EPOLL>, it still adds up to
336	incident, so its best to avoid that.	375	two event changes per incident, support for C<fork ()> is very bad and it
		376	drops fds silently in similarly hard-to-detect cases.
		377
		378	This backend usually performs well under most conditions.
		379
		380	While nominally embeddable in other event loops, this doesn't work
		381	everywhere, so you might need to test for this. And since it is broken
		382	almost everywhere, you should only use it when you have a lot of sockets
		383	(for which it usually works), by embedding it into another event loop
		384	(e.g. C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>) and using it only for
		385	sockets.
337		386
338	=item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8)	387	=item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8)
339		388
340	This is not implemented yet (and might never be).	389	This is not implemented yet (and might never be, unless you send me an
		390	implementation). According to reports, C</dev/poll> only supports sockets
		391	and is not embeddable, which would limit the usefulness of this backend
		392	immensely.
341		393
342	=item C<EVBACKEND_PORT> (value 32, Solaris 10)	394	=item C<EVBACKEND_PORT> (value 32, Solaris 10)
343		395
344	This uses the Solaris 10 port mechanism. As with everything on Solaris,	396	This uses the Solaris 10 event port mechanism. As with everything on Solaris,
345	it's really slow, but it still scales very well (O(active_fds)).	397	it's really slow, but it still scales very well (O(active_fds)).
346		398
347	Please note that solaris ports can result in a lot of spurious	399	Please note that solaris event ports can deliver a lot of spurious
348	notifications, so you need to use non-blocking I/O or other means to avoid	400	notifications, so you need to use non-blocking I/O or other means to avoid
349	blocking when no data (or space) is available.	401	blocking when no data (or space) is available.
		402
		403	While this backend scales well, it requires one system call per active
		404	file descriptor per loop iteration. For small and medium numbers of file
		405	descriptors a "slow" C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> backend
		406	might perform better.
		407
		408	On the positive side, ignoring the spurious readyness notifications, this
		409	backend actually performed to specification in all tests and is fully
		410	embeddable, which is a rare feat among the OS-specific backends.
350		411
351	=item C<EVBACKEND_ALL>	412	=item C<EVBACKEND_ALL>
352		413
353	Try all backends (even potentially broken ones that wouldn't be tried	414	Try all backends (even potentially broken ones that wouldn't be tried
354	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as	415	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as
355	C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>.	416	C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>.
356		417
		418	It is definitely not recommended to use this flag.
		419
357	=back	420	=back
358		421
359	If one or more of these are ored into the flags value, then only these	422	If one or more of these are ored into the flags value, then only these
360	backends will be tried (in the reverse order as given here). If none are	423	backends will be tried (in the reverse order as listed here). If none are
361	specified, most compiled-in backend will be tried, usually in reverse	424	specified, all backends in C<ev_recommended_backends ()> will be tried.
362	order of their flag values :)
363		425
364	The most typical usage is like this:	426	The most typical usage is like this:
365		427
366	if (!ev_default_loop (0))	428	if (!ev_default_loop (0))
367	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");	429	fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
…		…
395	Destroys the default loop again (frees all memory and kernel state	457	Destroys the default loop again (frees all memory and kernel state
396	etc.). None of the active event watchers will be stopped in the normal	458	etc.). None of the active event watchers will be stopped in the normal
397	sense, so e.g. C<ev_is_active> might still return true. It is your	459	sense, so e.g. C<ev_is_active> might still return true. It is your
398	responsibility to either stop all watchers cleanly yoursef I<before>	460	responsibility to either stop all watchers cleanly yoursef I<before>
399	calling this function, or cope with the fact afterwards (which is usually	461	calling this function, or cope with the fact afterwards (which is usually
400	the easiest thing, youc na just ignore the watchers and/or C<free ()> them	462	the easiest thing, you can just ignore the watchers and/or C<free ()> them
401	for example).	463	for example).
		464
		465	Note that certain global state, such as signal state, will not be freed by
		466	this function, and related watchers (such as signal and child watchers)
		467	would need to be stopped manually.
		468
		469	In general it is not advisable to call this function except in the
		470	rare occasion where you really need to free e.g. the signal handling
		471	pipe fds. If you need dynamically allocated loops it is better to use
		472	C<ev_loop_new> and C<ev_loop_destroy>).
402		473
403	=item ev_loop_destroy (loop)	474	=item ev_loop_destroy (loop)
404		475
405	Like C<ev_default_destroy>, but destroys an event loop created by an	476	Like C<ev_default_destroy>, but destroys an event loop created by an
406	earlier call to C<ev_loop_new>.	477	earlier call to C<ev_loop_new>.
…		…
451		522
452	Returns the current "event loop time", which is the time the event loop	523	Returns the current "event loop time", which is the time the event loop
453	received events and started processing them. This timestamp does not	524	received events and started processing them. This timestamp does not
454	change as long as callbacks are being processed, and this is also the base	525	change as long as callbacks are being processed, and this is also the base
455	time used for relative timers. You can treat it as the timestamp of the	526	time used for relative timers. You can treat it as the timestamp of the
456	event occuring (or more correctly, libev finding out about it).	527	event occurring (or more correctly, libev finding out about it).
457		528
458	=item ev_loop (loop, int flags)	529	=item ev_loop (loop, int flags)
459		530
460	Finally, this is it, the event handler. This function usually is called	531	Finally, this is it, the event handler. This function usually is called
461	after you initialised all your watchers and you want to start handling	532	after you initialised all your watchers and you want to start handling
…		…
482	libev watchers. However, a pair of C<ev_prepare>/C<ev_check> watchers is	553	libev watchers. However, a pair of C<ev_prepare>/C<ev_check> watchers is
483	usually a better approach for this kind of thing.	554	usually a better approach for this kind of thing.
484		555
485	Here are the gory details of what C<ev_loop> does:	556	Here are the gory details of what C<ev_loop> does:
486		557
487	* If there are no active watchers (reference count is zero), return.	558	- Before the first iteration, call any pending watchers.
488	- Queue prepare watchers and then call all outstanding watchers.	559	* If EVFLAG_FORKCHECK was used, check for a fork.
		560	- If a fork was detected, queue and call all fork watchers.
		561	- Queue and call all prepare watchers.
489	- If we have been forked, recreate the kernel state.	562	- If we have been forked, recreate the kernel state.
490	- Update the kernel state with all outstanding changes.	563	- Update the kernel state with all outstanding changes.
491	- Update the "event loop time".	564	- Update the "event loop time".
492	- Calculate for how long to block.	565	- Calculate for how long to sleep or block, if at all
		566	(active idle watchers, EVLOOP_NONBLOCK or not having
		567	any active watchers at all will result in not sleeping).
		568	- Sleep if the I/O and timer collect interval say so.
493	- Block the process, waiting for any events.	569	- Block the process, waiting for any events.
494	- Queue all outstanding I/O (fd) events.	570	- Queue all outstanding I/O (fd) events.
495	- Update the "event loop time" and do time jump handling.	571	- Update the "event loop time" and do time jump handling.
496	- Queue all outstanding timers.	572	- Queue all outstanding timers.
497	- Queue all outstanding periodics.	573	- Queue all outstanding periodics.
498	- If no events are pending now, queue all idle watchers.	574	- If no events are pending now, queue all idle watchers.
499	- Queue all check watchers.	575	- Queue all check watchers.
500	- Call all queued watchers in reverse order (i.e. check watchers first).	576	- Call all queued watchers in reverse order (i.e. check watchers first).
501	Signals and child watchers are implemented as I/O watchers, and will	577	Signals and child watchers are implemented as I/O watchers, and will
502	be handled here by queueing them when their watcher gets executed.	578	be handled here by queueing them when their watcher gets executed.
503	- If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK	579	- If ev_unloop has been called, or EVLOOP_ONESHOT or EVLOOP_NONBLOCK
504	were used, return, otherwise continue with step *.	580	were used, or there are no active watchers, return, otherwise
		581	continue with step *.
505		582
506	Example: Queue some jobs and then loop until no events are outsanding	583	Example: Queue some jobs and then loop until no events are outstanding
507	anymore.	584	anymore.
508		585
509	... queue jobs here, make sure they register event watchers as long	586	... queue jobs here, make sure they register event watchers as long
510	... as they still have work to do (even an idle watcher will do..)	587	... as they still have work to do (even an idle watcher will do..)
511	ev_loop (my_loop, 0);	588	ev_loop (my_loop, 0);
…		…
515		592
516	Can be used to make a call to C<ev_loop> return early (but only after it	593	Can be used to make a call to C<ev_loop> return early (but only after it
517	has processed all outstanding events). The C<how> argument must be either	594	has processed all outstanding events). The C<how> argument must be either
518	C<EVUNLOOP_ONE>, which will make the innermost C<ev_loop> call return, or	595	C<EVUNLOOP_ONE>, which will make the innermost C<ev_loop> call return, or
519	C<EVUNLOOP_ALL>, which will make all nested C<ev_loop> calls return.	596	C<EVUNLOOP_ALL>, which will make all nested C<ev_loop> calls return.
		597
		598	This "unloop state" will be cleared when entering C<ev_loop> again.
520		599
521	=item ev_ref (loop)	600	=item ev_ref (loop)
522		601
523	=item ev_unref (loop)	602	=item ev_unref (loop)
524		603
…		…
529	returning, ev_unref() after starting, and ev_ref() before stopping it. For	608	returning, ev_unref() after starting, and ev_ref() before stopping it. For
530	example, libev itself uses this for its internal signal pipe: It is not	609	example, libev itself uses this for its internal signal pipe: It is not
531	visible to the libev user and should not keep C<ev_loop> from exiting if	610	visible to the libev user and should not keep C<ev_loop> from exiting if
532	no event watchers registered by it are active. It is also an excellent	611	no event watchers registered by it are active. It is also an excellent
533	way to do this for generic recurring timers or from within third-party	612	way to do this for generic recurring timers or from within third-party
534	libraries. Just remember to I<unref after start> and I<ref before stop>.	613	libraries. Just remember to I<unref after start> and I<ref before stop>
		614	(but only if the watcher wasn't active before, or was active before,
		615	respectively).
535		616
536	Example: Create a signal watcher, but keep it from keeping C<ev_loop>	617	Example: Create a signal watcher, but keep it from keeping C<ev_loop>
537	running when nothing else is active.	618	running when nothing else is active.
538		619
539	struct ev_signal exitsig;	620	struct ev_signal exitsig;
…		…
543		624
544	Example: For some weird reason, unregister the above signal handler again.	625	Example: For some weird reason, unregister the above signal handler again.
545		626
546	ev_ref (loop);	627	ev_ref (loop);
547	ev_signal_stop (loop, &exitsig);	628	ev_signal_stop (loop, &exitsig);
		629
		630	=item ev_set_io_collect_interval (loop, ev_tstamp interval)
		631
		632	=item ev_set_timeout_collect_interval (loop, ev_tstamp interval)
		633
		634	These advanced functions influence the time that libev will spend waiting
		635	for events. Both are by default C<0>, meaning that libev will try to
		636	invoke timer/periodic callbacks and I/O callbacks with minimum latency.
		637
		638	Setting these to a higher value (the C<interval> I<must> be >= C<0>)
		639	allows libev to delay invocation of I/O and timer/periodic callbacks to
		640	increase efficiency of loop iterations.
		641
		642	The background is that sometimes your program runs just fast enough to
		643	handle one (or very few) event(s) per loop iteration. While this makes
		644	the program responsive, it also wastes a lot of CPU time to poll for new
		645	events, especially with backends like C<select ()> which have a high
		646	overhead for the actual polling but can deliver many events at once.
		647
		648	By setting a higher I<io collect interval> you allow libev to spend more
		649	time collecting I/O events, so you can handle more events per iteration,
		650	at the cost of increasing latency. Timeouts (both C<ev_periodic> and
		651	C<ev_timer>) will be not affected. Setting this to a non-null value will
		652	introduce an additional C<ev_sleep ()> call into most loop iterations.
		653
		654	Likewise, by setting a higher I<timeout collect interval> you allow libev
		655	to spend more time collecting timeouts, at the expense of increased
		656	latency (the watcher callback will be called later). C<ev_io> watchers
		657	will not be affected. Setting this to a non-null value will not introduce
		658	any overhead in libev.
		659
		660	Many (busy) programs can usually benefit by setting the io collect
		661	interval to a value near C<0.1> or so, which is often enough for
		662	interactive servers (of course not for games), likewise for timeouts. It
		663	usually doesn't make much sense to set it to a lower value than C<0.01>,
		664	as this approsaches the timing granularity of most systems.
548		665
549	=back	666	=back
550		667
551		668
552	=head1 ANATOMY OF A WATCHER	669	=head1 ANATOMY OF A WATCHER
…		…
732	=item bool ev_is_pending (ev_TYPE *watcher)	849	=item bool ev_is_pending (ev_TYPE *watcher)
733		850
734	Returns a true value iff the watcher is pending, (i.e. it has outstanding	851	Returns a true value iff the watcher is pending, (i.e. it has outstanding
735	events but its callback has not yet been invoked). As long as a watcher	852	events but its callback has not yet been invoked). As long as a watcher
736	is pending (but not active) you must not call an init function on it (but	853	is pending (but not active) you must not call an init function on it (but
737	C<ev_TYPE_set> is safe) and you must make sure the watcher is available to	854	C<ev_TYPE_set> is safe), you must not change its priority, and you must
738	libev (e.g. you cnanot C<free ()> it).	855	make sure the watcher is available to libev (e.g. you cannot C<free ()>
		856	it).
739		857
740	=item callback ev_cb (ev_TYPE *watcher)	858	=item callback ev_cb (ev_TYPE *watcher)
741		859
742	Returns the callback currently set on the watcher.	860	Returns the callback currently set on the watcher.
743		861
744	=item ev_cb_set (ev_TYPE *watcher, callback)	862	=item ev_cb_set (ev_TYPE *watcher, callback)
745		863
746	Change the callback. You can change the callback at virtually any time	864	Change the callback. You can change the callback at virtually any time
747	(modulo threads).	865	(modulo threads).
		866
		867	=item ev_set_priority (ev_TYPE *watcher, priority)
		868
		869	=item int ev_priority (ev_TYPE *watcher)
		870
		871	Set and query the priority of the watcher. The priority is a small
		872	integer between C<EV_MAXPRI> (default: C<2>) and C<EV_MINPRI>
		873	(default: C<-2>). Pending watchers with higher priority will be invoked
		874	before watchers with lower priority, but priority will not keep watchers
		875	from being executed (except for C<ev_idle> watchers).
		876
		877	This means that priorities are I<only> used for ordering callback
		878	invocation after new events have been received. This is useful, for
		879	example, to reduce latency after idling, or more often, to bind two
		880	watchers on the same event and make sure one is called first.
		881
		882	If you need to suppress invocation when higher priority events are pending
		883	you need to look at C<ev_idle> watchers, which provide this functionality.
		884
		885	You I<must not> change the priority of a watcher as long as it is active or
		886	pending.
		887
		888	The default priority used by watchers when no priority has been set is
		889	always C<0>, which is supposed to not be too high and not be too low :).
		890
		891	Setting a priority outside the range of C<EV_MINPRI> to C<EV_MAXPRI> is
		892	fine, as long as you do not mind that the priority value you query might
		893	or might not have been adjusted to be within valid range.
		894
		895	=item ev_invoke (loop, ev_TYPE *watcher, int revents)
		896
		897	Invoke the C<watcher> with the given C<loop> and C<revents>. Neither
		898	C<loop> nor C<revents> need to be valid as long as the watcher callback
		899	can deal with that fact.
		900
		901	=item int ev_clear_pending (loop, ev_TYPE *watcher)
		902
		903	If the watcher is pending, this function returns clears its pending status
		904	and returns its C<revents> bitset (as if its callback was invoked). If the
		905	watcher isn't pending it does nothing and returns C<0>.
748		906
749	=back	907	=back
750		908
751		909
752	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER	910	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER
…		…
837	In general you can register as many read and/or write event watchers per	995	In general you can register as many read and/or write event watchers per
838	fd as you want (as long as you don't confuse yourself). Setting all file	996	fd as you want (as long as you don't confuse yourself). Setting all file
839	descriptors to non-blocking mode is also usually a good idea (but not	997	descriptors to non-blocking mode is also usually a good idea (but not
840	required if you know what you are doing).	998	required if you know what you are doing).
841		999
842	You have to be careful with dup'ed file descriptors, though. Some backends
843	(the linux epoll backend is a notable example) cannot handle dup'ed file
844	descriptors correctly if you register interest in two or more fds pointing
845	to the same underlying file/socket/etc. description (that is, they share
846	the same underlying "file open").
847
848	If you must do this, then force the use of a known-to-be-good backend	1000	If you must do this, then force the use of a known-to-be-good backend
849	(at the time of this writing, this includes only C<EVBACKEND_SELECT> and	1001	(at the time of this writing, this includes only C<EVBACKEND_SELECT> and
850	C<EVBACKEND_POLL>).	1002	C<EVBACKEND_POLL>).
851		1003
852	Another thing you have to watch out for is that it is quite easy to	1004	Another thing you have to watch out for is that it is quite easy to
…		…
858	it is best to always use non-blocking I/O: An extra C<read>(2) returning	1010	it is best to always use non-blocking I/O: An extra C<read>(2) returning
859	C<EAGAIN> is far preferable to a program hanging until some data arrives.	1011	C<EAGAIN> is far preferable to a program hanging until some data arrives.
860		1012
861	If you cannot run the fd in non-blocking mode (for example you should not	1013	If you cannot run the fd in non-blocking mode (for example you should not
862	play around with an Xlib connection), then you have to seperately re-test	1014	play around with an Xlib connection), then you have to seperately re-test
863	wether a file descriptor is really ready with a known-to-be good interface	1015	whether a file descriptor is really ready with a known-to-be good interface
864	such as poll (fortunately in our Xlib example, Xlib already does this on	1016	such as poll (fortunately in our Xlib example, Xlib already does this on
865	its own, so its quite safe to use).	1017	its own, so its quite safe to use).
		1018
		1019	=head3 The special problem of disappearing file descriptors
		1020
		1021	Some backends (e.g. kqueue, epoll) need to be told about closing a file
		1022	descriptor (either by calling C<close> explicitly or by any other means,
		1023	such as C<dup>). The reason is that you register interest in some file
		1024	descriptor, but when it goes away, the operating system will silently drop
		1025	this interest. If another file descriptor with the same number then is
		1026	registered with libev, there is no efficient way to see that this is, in
		1027	fact, a different file descriptor.
		1028
		1029	To avoid having to explicitly tell libev about such cases, libev follows
		1030	the following policy: Each time C<ev_io_set> is being called, libev
		1031	will assume that this is potentially a new file descriptor, otherwise
		1032	it is assumed that the file descriptor stays the same. That means that
		1033	you I<have> to call C<ev_io_set> (or C<ev_io_init>) when you change the
		1034	descriptor even if the file descriptor number itself did not change.
		1035
		1036	This is how one would do it normally anyway, the important point is that
		1037	the libev application should not optimise around libev but should leave
		1038	optimisations to libev.
		1039
		1040	=head3 The special problem of dup'ed file descriptors
		1041
		1042	Some backends (e.g. epoll), cannot register events for file descriptors,
		1043	but only events for the underlying file descriptions. That means when you
		1044	have C<dup ()>'ed file descriptors or weirder constellations, and register
		1045	events for them, only one file descriptor might actually receive events.
		1046
		1047	There is no workaround possible except not registering events
		1048	for potentially C<dup ()>'ed file descriptors, or to resort to
		1049	C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>.
		1050
		1051	=head3 The special problem of fork
		1052
		1053	Some backends (epoll, kqueue) do not support C<fork ()> at all or exhibit
		1054	useless behaviour. Libev fully supports fork, but needs to be told about
		1055	it in the child.
		1056
		1057	To support fork in your programs, you either have to call
		1058	C<ev_default_fork ()> or C<ev_loop_fork ()> after a fork in the child,
		1059	enable C<EVFLAG_FORKCHECK>, or resort to C<EVBACKEND_SELECT> or
		1060	C<EVBACKEND_POLL>.
		1061
		1062
		1063	=head3 Watcher-Specific Functions
866		1064
867	=over 4	1065	=over 4
868		1066
869	=item ev_io_init (ev_io *, callback, int fd, int events)	1067	=item ev_io_init (ev_io *, callback, int fd, int events)
870		1068
…		…
881	=item int events [read-only]	1079	=item int events [read-only]
882		1080
883	The events being watched.	1081	The events being watched.
884		1082
885	=back	1083	=back
		1084
		1085	=head3 Examples
886		1086
887	Example: Call C<stdin_readable_cb> when STDIN_FILENO has become, well	1087	Example: Call C<stdin_readable_cb> when STDIN_FILENO has become, well
888	readable, but only once. Since it is likely line-buffered, you could	1088	readable, but only once. Since it is likely line-buffered, you could
889	attempt to read a whole line in the callback.	1089	attempt to read a whole line in the callback.
890		1090
…		…
923	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);	1123	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);
924		1124
925	The callback is guarenteed to be invoked only when its timeout has passed,	1125	The callback is guarenteed to be invoked only when its timeout has passed,
926	but if multiple timers become ready during the same loop iteration then	1126	but if multiple timers become ready during the same loop iteration then
927	order of execution is undefined.	1127	order of execution is undefined.
		1128
		1129	=head3 Watcher-Specific Functions and Data Members
928		1130
929	=over 4	1131	=over 4
930		1132
931	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)	1133	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)
932		1134
…		…
986	or C<ev_timer_again> is called and determines the next timeout (if any),	1188	or C<ev_timer_again> is called and determines the next timeout (if any),
987	which is also when any modifications are taken into account.	1189	which is also when any modifications are taken into account.
988		1190
989	=back	1191	=back
990		1192
		1193	=head3 Examples
		1194
991	Example: Create a timer that fires after 60 seconds.	1195	Example: Create a timer that fires after 60 seconds.
992		1196
993	static void	1197	static void
994	one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)	1198	one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
995	{	1199	{
…		…
1028	but on wallclock time (absolute time). You can tell a periodic watcher	1232	but on wallclock time (absolute time). You can tell a periodic watcher
1029	to trigger "at" some specific point in time. For example, if you tell a	1233	to trigger "at" some specific point in time. For example, if you tell a
1030	periodic watcher to trigger in 10 seconds (by specifiying e.g. C<ev_now ()	1234	periodic watcher to trigger in 10 seconds (by specifiying e.g. C<ev_now ()
1031	+ 10.>) and then reset your system clock to the last year, then it will	1235	+ 10.>) and then reset your system clock to the last year, then it will
1032	take a year to trigger the event (unlike an C<ev_timer>, which would trigger	1236	take a year to trigger the event (unlike an C<ev_timer>, which would trigger
1033	roughly 10 seconds later and of course not if you reset your system time	1237	roughly 10 seconds later).
1034	again).
1035		1238
1036	They can also be used to implement vastly more complex timers, such as	1239	They can also be used to implement vastly more complex timers, such as
1037	triggering an event on eahc midnight, local time.	1240	triggering an event on each midnight, local time or other, complicated,
		1241	rules.
1038		1242
1039	As with timers, the callback is guarenteed to be invoked only when the	1243	As with timers, the callback is guarenteed to be invoked only when the
1040	time (C<at>) has been passed, but if multiple periodic timers become ready	1244	time (C<at>) has been passed, but if multiple periodic timers become ready
1041	during the same loop iteration then order of execution is undefined.	1245	during the same loop iteration then order of execution is undefined.
1042		1246
		1247	=head3 Watcher-Specific Functions and Data Members
		1248
1043	=over 4	1249	=over 4
1044		1250
1045	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)	1251	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)
1046		1252
1047	=item ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)	1253	=item ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)
…		…
1049	Lots of arguments, lets sort it out... There are basically three modes of	1255	Lots of arguments, lets sort it out... There are basically three modes of
1050	operation, and we will explain them from simplest to complex:	1256	operation, and we will explain them from simplest to complex:
1051		1257
1052	=over 4	1258	=over 4
1053		1259
1054	=item * absolute timer (interval = reschedule_cb = 0)	1260	=item * absolute timer (at = time, interval = reschedule_cb = 0)
1055		1261
1056	In this configuration the watcher triggers an event at the wallclock time	1262	In this configuration the watcher triggers an event at the wallclock time
1057	C<at> and doesn't repeat. It will not adjust when a time jump occurs,	1263	C<at> and doesn't repeat. It will not adjust when a time jump occurs,
1058	that is, if it is to be run at January 1st 2011 then it will run when the	1264	that is, if it is to be run at January 1st 2011 then it will run when the
1059	system time reaches or surpasses this time.	1265	system time reaches or surpasses this time.
1060		1266
1061	=item * non-repeating interval timer (interval > 0, reschedule_cb = 0)	1267	=item * non-repeating interval timer (at = offset, interval > 0, reschedule_cb = 0)
1062		1268
1063	In this mode the watcher will always be scheduled to time out at the next	1269	In this mode the watcher will always be scheduled to time out at the next
1064	C<at + N * interval> time (for some integer N) and then repeat, regardless	1270	C<at + N * interval> time (for some integer N, which can also be negative)
1065	of any time jumps.	1271	and then repeat, regardless of any time jumps.
1066		1272
1067	This can be used to create timers that do not drift with respect to system	1273	This can be used to create timers that do not drift with respect to system
1068	time:	1274	time:
1069		1275
1070	ev_periodic_set (&periodic, 0., 3600., 0);	1276	ev_periodic_set (&periodic, 0., 3600., 0);
…		…
1076		1282
1077	Another way to think about it (for the mathematically inclined) is that	1283	Another way to think about it (for the mathematically inclined) is that
1078	C<ev_periodic> will try to run the callback in this mode at the next possible	1284	C<ev_periodic> will try to run the callback in this mode at the next possible
1079	time where C<time = at (mod interval)>, regardless of any time jumps.	1285	time where C<time = at (mod interval)>, regardless of any time jumps.
1080		1286
		1287	For numerical stability it is preferable that the C<at> value is near
		1288	C<ev_now ()> (the current time), but there is no range requirement for
		1289	this value.
		1290
1081	=item * manual reschedule mode (reschedule_cb = callback)	1291	=item * manual reschedule mode (at and interval ignored, reschedule_cb = callback)
1082		1292
1083	In this mode the values for C<interval> and C<at> are both being	1293	In this mode the values for C<interval> and C<at> are both being
1084	ignored. Instead, each time the periodic watcher gets scheduled, the	1294	ignored. Instead, each time the periodic watcher gets scheduled, the
1085	reschedule callback will be called with the watcher as first, and the	1295	reschedule callback will be called with the watcher as first, and the
1086	current time as second argument.	1296	current time as second argument.
1087		1297
1088	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher,	1298	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher,
1089	ever, or make any event loop modifications>. If you need to stop it,	1299	ever, or make any event loop modifications>. If you need to stop it,
1090	return C<now + 1e30> (or so, fudge fudge) and stop it afterwards (e.g. by	1300	return C<now + 1e30> (or so, fudge fudge) and stop it afterwards (e.g. by
1091	starting a prepare watcher).	1301	starting an C<ev_prepare> watcher, which is legal).
1092		1302
1093	Its prototype is C<ev_tstamp (*reschedule_cb)(struct ev_periodic *w,	1303	Its prototype is C<ev_tstamp (*reschedule_cb)(struct ev_periodic *w,
1094	ev_tstamp now)>, e.g.:	1304	ev_tstamp now)>, e.g.:
1095		1305
1096	static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)	1306	static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)
…		…
1119	Simply stops and restarts the periodic watcher again. This is only useful	1329	Simply stops and restarts the periodic watcher again. This is only useful
1120	when you changed some parameters or the reschedule callback would return	1330	when you changed some parameters or the reschedule callback would return
1121	a different time than the last time it was called (e.g. in a crond like	1331	a different time than the last time it was called (e.g. in a crond like
1122	program when the crontabs have changed).	1332	program when the crontabs have changed).
1123		1333
		1334	=item ev_tstamp offset [read-write]
		1335
		1336	When repeating, this contains the offset value, otherwise this is the
		1337	absolute point in time (the C<at> value passed to C<ev_periodic_set>).
		1338
		1339	Can be modified any time, but changes only take effect when the periodic
		1340	timer fires or C<ev_periodic_again> is being called.
		1341
1124	=item ev_tstamp interval [read-write]	1342	=item ev_tstamp interval [read-write]
1125		1343
1126	The current interval value. Can be modified any time, but changes only	1344	The current interval value. Can be modified any time, but changes only
1127	take effect when the periodic timer fires or C<ev_periodic_again> is being	1345	take effect when the periodic timer fires or C<ev_periodic_again> is being
1128	called.	1346	called.
…		…
1131		1349
1132	The current reschedule callback, or C<0>, if this functionality is	1350	The current reschedule callback, or C<0>, if this functionality is
1133	switched off. Can be changed any time, but changes only take effect when	1351	switched off. Can be changed any time, but changes only take effect when
1134	the periodic timer fires or C<ev_periodic_again> is being called.	1352	the periodic timer fires or C<ev_periodic_again> is being called.
1135		1353
		1354	=item ev_tstamp at [read-only]
		1355
		1356	When active, contains the absolute time that the watcher is supposed to
		1357	trigger next.
		1358
1136	=back	1359	=back
		1360
		1361	=head3 Examples
1137		1362
1138	Example: Call a callback every hour, or, more precisely, whenever the	1363	Example: Call a callback every hour, or, more precisely, whenever the
1139	system clock is divisible by 3600. The callback invocation times have	1364	system clock is divisible by 3600. The callback invocation times have
1140	potentially a lot of jittering, but good long-term stability.	1365	potentially a lot of jittering, but good long-term stability.
1141		1366
…		…
1181	with the kernel (thus it coexists with your own signal handlers as long	1406	with the kernel (thus it coexists with your own signal handlers as long
1182	as you don't register any with libev). Similarly, when the last signal	1407	as you don't register any with libev). Similarly, when the last signal
1183	watcher for a signal is stopped libev will reset the signal handler to	1408	watcher for a signal is stopped libev will reset the signal handler to
1184	SIG_DFL (regardless of what it was set to before).	1409	SIG_DFL (regardless of what it was set to before).
1185		1410
		1411	=head3 Watcher-Specific Functions and Data Members
		1412
1186	=over 4	1413	=over 4
1187		1414
1188	=item ev_signal_init (ev_signal *, callback, int signum)	1415	=item ev_signal_init (ev_signal *, callback, int signum)
1189		1416
1190	=item ev_signal_set (ev_signal *, int signum)	1417	=item ev_signal_set (ev_signal *, int signum)
…		…
1201		1428
1202	=head2 C<ev_child> - watch out for process status changes	1429	=head2 C<ev_child> - watch out for process status changes
1203		1430
1204	Child watchers trigger when your process receives a SIGCHLD in response to	1431	Child watchers trigger when your process receives a SIGCHLD in response to
1205	some child status changes (most typically when a child of yours dies).	1432	some child status changes (most typically when a child of yours dies).
		1433
		1434	=head3 Watcher-Specific Functions and Data Members
1206		1435
1207	=over 4	1436	=over 4
1208		1437
1209	=item ev_child_init (ev_child *, callback, int pid)	1438	=item ev_child_init (ev_child *, callback, int pid)
1210		1439
…		…
1229		1458
1230	The process exit/trace status caused by C<rpid> (see your systems	1459	The process exit/trace status caused by C<rpid> (see your systems
1231	C<waitpid> and C<sys/wait.h> documentation for details).	1460	C<waitpid> and C<sys/wait.h> documentation for details).
1232		1461
1233	=back	1462	=back
		1463
		1464	=head3 Examples
1234		1465
1235	Example: Try to exit cleanly on SIGINT and SIGTERM.	1466	Example: Try to exit cleanly on SIGINT and SIGTERM.
1236		1467
1237	static void	1468	static void
1238	sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)	1469	sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)
…		…
1279	semantics of C<ev_stat> watchers, which means that libev sometimes needs	1510	semantics of C<ev_stat> watchers, which means that libev sometimes needs
1280	to fall back to regular polling again even with inotify, but changes are	1511	to fall back to regular polling again even with inotify, but changes are
1281	usually detected immediately, and if the file exists there will be no	1512	usually detected immediately, and if the file exists there will be no
1282	polling.	1513	polling.
1283		1514
		1515	=head3 Inotify
		1516
		1517	When C<inotify (7)> support has been compiled into libev (generally only
		1518	available on Linux) and present at runtime, it will be used to speed up
		1519	change detection where possible. The inotify descriptor will be created lazily
		1520	when the first C<ev_stat> watcher is being started.
		1521
		1522	Inotify presense does not change the semantics of C<ev_stat> watchers
		1523	except that changes might be detected earlier, and in some cases, to avoid
		1524	making regular C<stat> calls. Even in the presense of inotify support
		1525	there are many cases where libev has to resort to regular C<stat> polling.
		1526
		1527	(There is no support for kqueue, as apparently it cannot be used to
		1528	implement this functionality, due to the requirement of having a file
		1529	descriptor open on the object at all times).
		1530
		1531	=head3 The special problem of stat time resolution
		1532
		1533	The C<stat ()> syscall only supports full-second resolution portably, and
		1534	even on systems where the resolution is higher, many filesystems still
		1535	only support whole seconds.
		1536
		1537	That means that, if the time is the only thing that changes, you might
		1538	miss updates: on the first update, C<ev_stat> detects a change and calls
		1539	your callback, which does something. When there is another update within
		1540	the same second, C<ev_stat> will be unable to detect it.
		1541
		1542	The solution to this is to delay acting on a change for a second (or till
		1543	the next second boundary), using a roughly one-second delay C<ev_timer>
		1544	(C<ev_timer_set (w, 0., 1.01); ev_timer_again (loop, w)>). The C<.01>
		1545	is added to work around small timing inconsistencies of some operating
		1546	systems.
		1547
		1548	=head3 Watcher-Specific Functions and Data Members
		1549
1284	=over 4	1550	=over 4
1285		1551
1286	=item ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)	1552	=item ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)
1287		1553
1288	=item ev_stat_set (ev_stat *, const char *path, ev_tstamp interval)	1554	=item ev_stat_set (ev_stat *, const char *path, ev_tstamp interval)
…		…
1323	=item const char *path [read-only]	1589	=item const char *path [read-only]
1324		1590
1325	The filesystem path that is being watched.	1591	The filesystem path that is being watched.
1326		1592
1327	=back	1593	=back
		1594
		1595	=head3 Examples
1328		1596
1329	Example: Watch C</etc/passwd> for attribute changes.	1597	Example: Watch C</etc/passwd> for attribute changes.
1330		1598
1331	static void	1599	static void
1332	passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)	1600	passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)
…		…
1345	}	1613	}
1346		1614
1347	...	1615	...
1348	ev_stat passwd;	1616	ev_stat passwd;
1349		1617
1350	ev_stat_init (&passwd, passwd_cb, "/etc/passwd");	1618	ev_stat_init (&passwd, passwd_cb, "/etc/passwd", 0.);
1351	ev_stat_start (loop, &passwd);	1619	ev_stat_start (loop, &passwd);
1352		1620
		1621	Example: Like above, but additionally use a one-second delay so we do not
		1622	miss updates (however, frequent updates will delay processing, too, so
		1623	one might do the work both on C<ev_stat> callback invocation I<and> on
		1624	C<ev_timer> callback invocation).
		1625
		1626	static ev_stat passwd;
		1627	static ev_timer timer;
		1628
		1629	static void
		1630	timer_cb (EV_P_ ev_timer *w, int revents)
		1631	{
		1632	ev_timer_stop (EV_A_ w);
		1633
		1634	/* now it's one second after the most recent passwd change */
		1635	}
		1636
		1637	static void
		1638	stat_cb (EV_P_ ev_stat *w, int revents)
		1639	{
		1640	/* reset the one-second timer */
		1641	ev_timer_again (EV_A_ &timer);
		1642	}
		1643
		1644	...
		1645	ev_stat_init (&passwd, stat_cb, "/etc/passwd", 0.);
		1646	ev_stat_start (loop, &passwd);
		1647	ev_timer_init (&timer, timer_cb, 0., 1.01);
		1648
1353		1649
1354	=head2 C<ev_idle> - when you've got nothing better to do...	1650	=head2 C<ev_idle> - when you've got nothing better to do...
1355		1651
1356	Idle watchers trigger events when there are no other events are pending	1652	Idle watchers trigger events when no other events of the same or higher
1357	(prepare, check and other idle watchers do not count). That is, as long	1653	priority are pending (prepare, check and other idle watchers do not
1358	as your process is busy handling sockets or timeouts (or even signals,	1654	count).
1359	imagine) it will not be triggered. But when your process is idle all idle	1655
1360	watchers are being called again and again, once per event loop iteration -	1656	That is, as long as your process is busy handling sockets or timeouts
		1657	(or even signals, imagine) of the same or higher priority it will not be
		1658	triggered. But when your process is idle (or only lower-priority watchers
		1659	are pending), the idle watchers are being called once per event loop
1361	until stopped, that is, or your process receives more events and becomes	1660	iteration - until stopped, that is, or your process receives more events
1362	busy.	1661	and becomes busy again with higher priority stuff.
1363		1662
1364	The most noteworthy effect is that as long as any idle watchers are	1663	The most noteworthy effect is that as long as any idle watchers are
1365	active, the process will not block when waiting for new events.	1664	active, the process will not block when waiting for new events.
1366		1665
1367	Apart from keeping your process non-blocking (which is a useful	1666	Apart from keeping your process non-blocking (which is a useful
1368	effect on its own sometimes), idle watchers are a good place to do	1667	effect on its own sometimes), idle watchers are a good place to do
1369	"pseudo-background processing", or delay processing stuff to after the	1668	"pseudo-background processing", or delay processing stuff to after the
1370	event loop has handled all outstanding events.	1669	event loop has handled all outstanding events.
1371		1670
		1671	=head3 Watcher-Specific Functions and Data Members
		1672
1372	=over 4	1673	=over 4
1373		1674
1374	=item ev_idle_init (ev_signal *, callback)	1675	=item ev_idle_init (ev_signal *, callback)
1375		1676
1376	Initialises and configures the idle watcher - it has no parameters of any	1677	Initialises and configures the idle watcher - it has no parameters of any
1377	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,	1678	kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,
1378	believe me.	1679	believe me.
1379		1680
1380	=back	1681	=back
		1682
		1683	=head3 Examples
1381		1684
1382	Example: Dynamically allocate an C<ev_idle> watcher, start it, and in the	1685	Example: Dynamically allocate an C<ev_idle> watcher, start it, and in the
1383	callback, free it. Also, use no error checking, as usual.	1686	callback, free it. Also, use no error checking, as usual.
1384		1687
1385	static void	1688	static void
…		…
1433	with priority higher than or equal to the event loop and one coroutine	1736	with priority higher than or equal to the event loop and one coroutine
1434	of lower priority, but only once, using idle watchers to keep the event	1737	of lower priority, but only once, using idle watchers to keep the event
1435	loop from blocking if lower-priority coroutines are active, thus mapping	1738	loop from blocking if lower-priority coroutines are active, thus mapping
1436	low-priority coroutines to idle/background tasks).	1739	low-priority coroutines to idle/background tasks).
1437		1740
		1741	It is recommended to give C<ev_check> watchers highest (C<EV_MAXPRI>)
		1742	priority, to ensure that they are being run before any other watchers
		1743	after the poll. Also, C<ev_check> watchers (and C<ev_prepare> watchers,
		1744	too) should not activate ("feed") events into libev. While libev fully
		1745	supports this, they will be called before other C<ev_check> watchers
		1746	did their job. As C<ev_check> watchers are often used to embed other
		1747	(non-libev) event loops those other event loops might be in an unusable
		1748	state until their C<ev_check> watcher ran (always remind yourself to
		1749	coexist peacefully with others).
		1750
		1751	=head3 Watcher-Specific Functions and Data Members
		1752
1438	=over 4	1753	=over 4
1439		1754
1440	=item ev_prepare_init (ev_prepare *, callback)	1755	=item ev_prepare_init (ev_prepare *, callback)
1441		1756
1442	=item ev_check_init (ev_check *, callback)	1757	=item ev_check_init (ev_check *, callback)
…		…
1445	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>	1760	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>
1446	macros, but using them is utterly, utterly and completely pointless.	1761	macros, but using them is utterly, utterly and completely pointless.
1447		1762
1448	=back	1763	=back
1449		1764
1450	Example: To include a library such as adns, you would add IO watchers	1765	=head3 Examples
1451	and a timeout watcher in a prepare handler, as required by libadns, and	1766
		1767	There are a number of principal ways to embed other event loops or modules
		1768	into libev. Here are some ideas on how to include libadns into libev
		1769	(there is a Perl module named C<EV::ADNS> that does this, which you could
		1770	use for an actually working example. Another Perl module named C<EV::Glib>
		1771	embeds a Glib main context into libev, and finally, C<Glib::EV> embeds EV
		1772	into the Glib event loop).
		1773
		1774	Method 1: Add IO watchers and a timeout watcher in a prepare handler,
1452	in a check watcher, destroy them and call into libadns. What follows is	1775	and in a check watcher, destroy them and call into libadns. What follows
1453	pseudo-code only of course:	1776	is pseudo-code only of course. This requires you to either use a low
		1777	priority for the check watcher or use C<ev_clear_pending> explicitly, as
		1778	the callbacks for the IO/timeout watchers might not have been called yet.
1454		1779
1455	static ev_io iow [nfd];	1780	static ev_io iow [nfd];
1456	static ev_timer tw;	1781	static ev_timer tw;
1457		1782
1458	static void	1783	static void
1459	io_cb (ev_loop *loop, ev_io *w, int revents)	1784	io_cb (ev_loop *loop, ev_io *w, int revents)
1460	{	1785	{
1461	// set the relevant poll flags
1462	// could also call adns_processreadable etc. here
1463	struct pollfd *fd = (struct pollfd *)w->data;
1464	if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;
1465	if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;
1466	}	1786	}
1467		1787
1468	// create io watchers for each fd and a timer before blocking	1788	// create io watchers for each fd and a timer before blocking
1469	static void	1789	static void
1470	adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents)	1790	adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents)
…		…
1476		1796
1477	/* the callback is illegal, but won't be called as we stop during check */	1797	/* the callback is illegal, but won't be called as we stop during check */
1478	ev_timer_init (&tw, 0, timeout * 1e-3);	1798	ev_timer_init (&tw, 0, timeout * 1e-3);
1479	ev_timer_start (loop, &tw);	1799	ev_timer_start (loop, &tw);
1480		1800
1481	// create on ev_io per pollfd	1801	// create one ev_io per pollfd
1482	for (int i = 0; i < nfd; ++i)	1802	for (int i = 0; i < nfd; ++i)
1483	{	1803	{
1484	ev_io_init (iow + i, io_cb, fds [i].fd,	1804	ev_io_init (iow + i, io_cb, fds [i].fd,
1485	((fds [i].events & POLLIN ? EV_READ : 0)	1805	((fds [i].events & POLLIN ? EV_READ : 0)
1486	| (fds [i].events & POLLOUT ? EV_WRITE : 0)));	1806	| (fds [i].events & POLLOUT ? EV_WRITE : 0)));
1487		1807
1488	fds [i].revents = 0;	1808	fds [i].revents = 0;
1489	iow [i].data = fds + i;
1490	ev_io_start (loop, iow + i);	1809	ev_io_start (loop, iow + i);
1491	}	1810	}
1492	}	1811	}
1493		1812
1494	// stop all watchers after blocking	1813	// stop all watchers after blocking
…		…
1496	adns_check_cb (ev_loop *loop, ev_check *w, int revents)	1815	adns_check_cb (ev_loop *loop, ev_check *w, int revents)
1497	{	1816	{
1498	ev_timer_stop (loop, &tw);	1817	ev_timer_stop (loop, &tw);
1499		1818
1500	for (int i = 0; i < nfd; ++i)	1819	for (int i = 0; i < nfd; ++i)
		1820	{
		1821	// set the relevant poll flags
		1822	// could also call adns_processreadable etc. here
		1823	struct pollfd *fd = fds + i;
		1824	int revents = ev_clear_pending (iow + i);
		1825	if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;
		1826	if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;
		1827
		1828	// now stop the watcher
1501	ev_io_stop (loop, iow + i);	1829	ev_io_stop (loop, iow + i);
		1830	}
1502		1831
1503	adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));	1832	adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));
		1833	}
		1834
		1835	Method 2: This would be just like method 1, but you run C<adns_afterpoll>
		1836	in the prepare watcher and would dispose of the check watcher.
		1837
		1838	Method 3: If the module to be embedded supports explicit event
		1839	notification (adns does), you can also make use of the actual watcher
		1840	callbacks, and only destroy/create the watchers in the prepare watcher.
		1841
		1842	static void
		1843	timer_cb (EV_P_ ev_timer *w, int revents)
		1844	{
		1845	adns_state ads = (adns_state)w->data;
		1846	update_now (EV_A);
		1847
		1848	adns_processtimeouts (ads, &tv_now);
		1849	}
		1850
		1851	static void
		1852	io_cb (EV_P_ ev_io *w, int revents)
		1853	{
		1854	adns_state ads = (adns_state)w->data;
		1855	update_now (EV_A);
		1856
		1857	if (revents & EV_READ ) adns_processreadable (ads, w->fd, &tv_now);
		1858	if (revents & EV_WRITE) adns_processwriteable (ads, w->fd, &tv_now);
		1859	}
		1860
		1861	// do not ever call adns_afterpoll
		1862
		1863	Method 4: Do not use a prepare or check watcher because the module you
		1864	want to embed is too inflexible to support it. Instead, youc na override
		1865	their poll function. The drawback with this solution is that the main
		1866	loop is now no longer controllable by EV. The C<Glib::EV> module does
		1867	this.
		1868
		1869	static gint
		1870	event_poll_func (GPollFD *fds, guint nfds, gint timeout)
		1871	{
		1872	int got_events = 0;
		1873
		1874	for (n = 0; n < nfds; ++n)
		1875	// create/start io watcher that sets the relevant bits in fds[n] and increment got_events
		1876
		1877	if (timeout >= 0)
		1878	// create/start timer
		1879
		1880	// poll
		1881	ev_loop (EV_A_ 0);
		1882
		1883	// stop timer again
		1884	if (timeout >= 0)
		1885	ev_timer_stop (EV_A_ &to);
		1886
		1887	// stop io watchers again - their callbacks should have set
		1888	for (n = 0; n < nfds; ++n)
		1889	ev_io_stop (EV_A_ iow [n]);
		1890
		1891	return got_events;
1504	}	1892	}
1505		1893
1506		1894
1507	=head2 C<ev_embed> - when one backend isn't enough...	1895	=head2 C<ev_embed> - when one backend isn't enough...
1508		1896
…		…
1551	portable one.	1939	portable one.
1552		1940
1553	So when you want to use this feature you will always have to be prepared	1941	So when you want to use this feature you will always have to be prepared
1554	that you cannot get an embeddable loop. The recommended way to get around	1942	that you cannot get an embeddable loop. The recommended way to get around
1555	this is to have a separate variables for your embeddable loop, try to	1943	this is to have a separate variables for your embeddable loop, try to
1556	create it, and if that fails, use the normal loop for everything:	1944	create it, and if that fails, use the normal loop for everything.
		1945
		1946	=head3 Watcher-Specific Functions and Data Members
		1947
		1948	=over 4
		1949
		1950	=item ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)
		1951
		1952	=item ev_embed_set (ev_embed *, callback, struct ev_loop *embedded_loop)
		1953
		1954	Configures the watcher to embed the given loop, which must be
		1955	embeddable. If the callback is C<0>, then C<ev_embed_sweep> will be
		1956	invoked automatically, otherwise it is the responsibility of the callback
		1957	to invoke it (it will continue to be called until the sweep has been done,
		1958	if you do not want thta, you need to temporarily stop the embed watcher).
		1959
		1960	=item ev_embed_sweep (loop, ev_embed *)
		1961
		1962	Make a single, non-blocking sweep over the embedded loop. This works
		1963	similarly to C<ev_loop (embedded_loop, EVLOOP_NONBLOCK)>, but in the most
		1964	apropriate way for embedded loops.
		1965
		1966	=item struct ev_loop *other [read-only]
		1967
		1968	The embedded event loop.
		1969
		1970	=back
		1971
		1972	=head3 Examples
		1973
		1974	Example: Try to get an embeddable event loop and embed it into the default
		1975	event loop. If that is not possible, use the default loop. The default
		1976	loop is stored in C<loop_hi>, while the mebeddable loop is stored in
		1977	C<loop_lo> (which is C<loop_hi> in the acse no embeddable loop can be
		1978	used).
1557		1979
1558	struct ev_loop *loop_hi = ev_default_init (0);	1980	struct ev_loop *loop_hi = ev_default_init (0);
1559	struct ev_loop *loop_lo = 0;	1981	struct ev_loop *loop_lo = 0;
1560	struct ev_embed embed;	1982	struct ev_embed embed;
1561		1983
…		…
1572	ev_embed_start (loop_hi, &embed);	1994	ev_embed_start (loop_hi, &embed);
1573	}	1995	}
1574	else	1996	else
1575	loop_lo = loop_hi;	1997	loop_lo = loop_hi;
1576		1998
1577	=over 4	1999	Example: Check if kqueue is available but not recommended and create
		2000	a kqueue backend for use with sockets (which usually work with any
		2001	kqueue implementation). Store the kqueue/socket-only event loop in
		2002	C<loop_socket>. (One might optionally use C<EVFLAG_NOENV>, too).
1578		2003
1579	=item ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)	2004	struct ev_loop *loop = ev_default_init (0);
		2005	struct ev_loop *loop_socket = 0;
		2006	struct ev_embed embed;
		2007
		2008	if (ev_supported_backends () & ~ev_recommended_backends () & EVBACKEND_KQUEUE)
		2009	if ((loop_socket = ev_loop_new (EVBACKEND_KQUEUE))
		2010	{
		2011	ev_embed_init (&embed, 0, loop_socket);
		2012	ev_embed_start (loop, &embed);
		2013	}
1580		2014
1581	=item ev_embed_set (ev_embed *, callback, struct ev_loop *embedded_loop)	2015	if (!loop_socket)
		2016	loop_socket = loop;
1582		2017
1583	Configures the watcher to embed the given loop, which must be	2018	// now use loop_socket for all sockets, and loop for everything else
1584	embeddable. If the callback is C<0>, then C<ev_embed_sweep> will be
1585	invoked automatically, otherwise it is the responsibility of the callback
1586	to invoke it (it will continue to be called until the sweep has been done,
1587	if you do not want thta, you need to temporarily stop the embed watcher).
1588
1589	=item ev_embed_sweep (loop, ev_embed *)
1590
1591	Make a single, non-blocking sweep over the embedded loop. This works
1592	similarly to C<ev_loop (embedded_loop, EVLOOP_NONBLOCK)>, but in the most
1593	apropriate way for embedded loops.
1594
1595	=item struct ev_loop *loop [read-only]
1596
1597	The embedded event loop.
1598
1599	=back
1600		2019
1601		2020
1602	=head2 C<ev_fork> - the audacity to resume the event loop after a fork	2021	=head2 C<ev_fork> - the audacity to resume the event loop after a fork
1603		2022
1604	Fork watchers are called when a C<fork ()> was detected (usually because	2023	Fork watchers are called when a C<fork ()> was detected (usually because
…		…
1607	event loop blocks next and before C<ev_check> watchers are being called,	2026	event loop blocks next and before C<ev_check> watchers are being called,
1608	and only in the child after the fork. If whoever good citizen calling	2027	and only in the child after the fork. If whoever good citizen calling
1609	C<ev_default_fork> cheats and calls it in the wrong process, the fork	2028	C<ev_default_fork> cheats and calls it in the wrong process, the fork
1610	handlers will be invoked, too, of course.	2029	handlers will be invoked, too, of course.
1611		2030
		2031	=head3 Watcher-Specific Functions and Data Members
		2032
1612	=over 4	2033	=over 4
1613		2034
1614	=item ev_fork_init (ev_signal *, callback)	2035	=item ev_fork_init (ev_signal *, callback)
1615		2036
1616	Initialises and configures the fork watcher - it has no parameters of any	2037	Initialises and configures the fork watcher - it has no parameters of any
…		…
1712		2133
1713	To use it,	2134	To use it,
1714		2135
1715	#include <ev++.h>	2136	#include <ev++.h>
1716		2137
1717	(it is not installed by default). This automatically includes F<ev.h>	2138	This automatically includes F<ev.h> and puts all of its definitions (many
1718	and puts all of its definitions (many of them macros) into the global	2139	of them macros) into the global namespace. All C++ specific things are
1719	namespace. All C++ specific things are put into the C<ev> namespace.	2140	put into the C<ev> namespace. It should support all the same embedding
		2141	options as F<ev.h>, most notably C<EV_MULTIPLICITY>.
1720		2142
1721	It should support all the same embedding options as F<ev.h>, most notably	2143	Care has been taken to keep the overhead low. The only data member the C++
1722	C<EV_MULTIPLICITY>.	2144	classes add (compared to plain C-style watchers) is the event loop pointer
		2145	that the watcher is associated with (or no additional members at all if
		2146	you disable C<EV_MULTIPLICITY> when embedding libev).
		2147
		2148	Currently, functions, and static and non-static member functions can be
		2149	used as callbacks. Other types should be easy to add as long as they only
		2150	need one additional pointer for context. If you need support for other
		2151	types of functors please contact the author (preferably after implementing
		2152	it).
1723		2153
1724	Here is a list of things available in the C<ev> namespace:	2154	Here is a list of things available in the C<ev> namespace:
1725		2155
1726	=over 4	2156	=over 4
1727		2157
…		…
1743		2173
1744	All of those classes have these methods:	2174	All of those classes have these methods:
1745		2175
1746	=over 4	2176	=over 4
1747		2177
1748	=item ev::TYPE::TYPE (object *, object::method *)	2178	=item ev::TYPE::TYPE ()
1749		2179
1750	=item ev::TYPE::TYPE (object *, object::method *, struct ev_loop *)	2180	=item ev::TYPE::TYPE (struct ev_loop *)
1751		2181
1752	=item ev::TYPE::~TYPE	2182	=item ev::TYPE::~TYPE
1753		2183
1754	The constructor takes a pointer to an object and a method pointer to	2184	The constructor (optionally) takes an event loop to associate the watcher
1755	the event handler callback to call in this class. The constructor calls	2185	with. If it is omitted, it will use C<EV_DEFAULT>.
1756	C<ev_init> for you, which means you have to call the C<set> method	2186
1757	before starting it. If you do not specify a loop then the constructor	2187	The constructor calls C<ev_init> for you, which means you have to call the
1758	automatically associates the default loop with this watcher.	2188	C<set> method before starting it.
		2189
		2190	It will not set a callback, however: You have to call the templated C<set>
		2191	method to set a callback before you can start the watcher.
		2192
		2193	(The reason why you have to use a method is a limitation in C++ which does
		2194	not allow explicit template arguments for constructors).
1759		2195
1760	The destructor automatically stops the watcher if it is active.	2196	The destructor automatically stops the watcher if it is active.
		2197
		2198	=item w->set<class, &class::method> (object *)
		2199
		2200	This method sets the callback method to call. The method has to have a
		2201	signature of C<void (*)(ev_TYPE &, int)>, it receives the watcher as
		2202	first argument and the C<revents> as second. The object must be given as
		2203	parameter and is stored in the C<data> member of the watcher.
		2204
		2205	This method synthesizes efficient thunking code to call your method from
		2206	the C callback that libev requires. If your compiler can inline your
		2207	callback (i.e. it is visible to it at the place of the C<set> call and
		2208	your compiler is good :), then the method will be fully inlined into the
		2209	thunking function, making it as fast as a direct C callback.
		2210
		2211	Example: simple class declaration and watcher initialisation
		2212
		2213	struct myclass
		2214	{
		2215	void io_cb (ev::io &w, int revents) { }
		2216	}
		2217
		2218	myclass obj;
		2219	ev::io iow;
		2220	iow.set <myclass, &myclass::io_cb> (&obj);
		2221
		2222	=item w->set<function> (void *data = 0)
		2223
		2224	Also sets a callback, but uses a static method or plain function as
		2225	callback. The optional C<data> argument will be stored in the watcher's
		2226	C<data> member and is free for you to use.
		2227
		2228	The prototype of the C<function> must be C<void (*)(ev::TYPE &w, int)>.
		2229
		2230	See the method-C<set> above for more details.
		2231
		2232	Example:
		2233
		2234	static void io_cb (ev::io &w, int revents) { }
		2235	iow.set <io_cb> ();
1761		2236
1762	=item w->set (struct ev_loop *)	2237	=item w->set (struct ev_loop *)
1763		2238
1764	Associates a different C<struct ev_loop> with this watcher. You can only	2239	Associates a different C<struct ev_loop> with this watcher. You can only
1765	do this when the watcher is inactive (and not pending either).	2240	do this when the watcher is inactive (and not pending either).
1766		2241
1767	=item w->set ([args])	2242	=item w->set ([args])
1768		2243
1769	Basically the same as C<ev_TYPE_set>, with the same args. Must be	2244	Basically the same as C<ev_TYPE_set>, with the same args. Must be
1770	called at least once. Unlike the C counterpart, an active watcher gets	2245	called at least once. Unlike the C counterpart, an active watcher gets
1771	automatically stopped and restarted.	2246	automatically stopped and restarted when reconfiguring it with this
		2247	method.
1772		2248
1773	=item w->start ()	2249	=item w->start ()
1774		2250
1775	Starts the watcher. Note that there is no C<loop> argument as the	2251	Starts the watcher. Note that there is no C<loop> argument, as the
1776	constructor already takes the loop.	2252	constructor already stores the event loop.
1777		2253
1778	=item w->stop ()	2254	=item w->stop ()
1779		2255
1780	Stops the watcher if it is active. Again, no C<loop> argument.	2256	Stops the watcher if it is active. Again, no C<loop> argument.
1781		2257
1782	=item w->again () C<ev::timer>, C<ev::periodic> only	2258	=item w->again () (C<ev::timer>, C<ev::periodic> only)
1783		2259
1784	For C<ev::timer> and C<ev::periodic>, this invokes the corresponding	2260	For C<ev::timer> and C<ev::periodic>, this invokes the corresponding
1785	C<ev_TYPE_again> function.	2261	C<ev_TYPE_again> function.
1786		2262
1787	=item w->sweep () C<ev::embed> only	2263	=item w->sweep () (C<ev::embed> only)
1788		2264
1789	Invokes C<ev_embed_sweep>.	2265	Invokes C<ev_embed_sweep>.
1790		2266
1791	=item w->update () C<ev::stat> only	2267	=item w->update () (C<ev::stat> only)
1792		2268
1793	Invokes C<ev_stat_stat>.	2269	Invokes C<ev_stat_stat>.
1794		2270
1795	=back	2271	=back
1796		2272
…		…
1806		2282
1807	myclass ();	2283	myclass ();
1808	}	2284	}
1809		2285
1810	myclass::myclass (int fd)	2286	myclass::myclass (int fd)
1811	: io (this, &myclass::io_cb),
1812	idle (this, &myclass::idle_cb)
1813	{	2287	{
		2288	io .set <myclass, &myclass::io_cb > (this);
		2289	idle.set <myclass, &myclass::idle_cb> (this);
		2290
1814	io.start (fd, ev::READ);	2291	io.start (fd, ev::READ);
1815	}	2292	}
1816		2293
1817		2294
1818	=head1 MACRO MAGIC	2295	=head1 MACRO MAGIC
1819		2296
1820	Libev can be compiled with a variety of options, the most fundemantal is	2297	Libev can be compiled with a variety of options, the most fundamantal
1821	C<EV_MULTIPLICITY>. This option determines wether (most) functions and	2298	of which is C<EV_MULTIPLICITY>. This option determines whether (most)
1822	callbacks have an initial C<struct ev_loop *> argument.	2299	functions and callbacks have an initial C<struct ev_loop *> argument.
1823		2300
1824	To make it easier to write programs that cope with either variant, the	2301	To make it easier to write programs that cope with either variant, the
1825	following macros are defined:	2302	following macros are defined:
1826		2303
1827	=over 4	2304	=over 4
…		…
1860	loop, if multiple loops are supported ("ev loop default").	2337	loop, if multiple loops are supported ("ev loop default").
1861		2338
1862	=back	2339	=back
1863		2340
1864	Example: Declare and initialise a check watcher, utilising the above	2341	Example: Declare and initialise a check watcher, utilising the above
1865	macros so it will work regardless of wether multiple loops are supported	2342	macros so it will work regardless of whether multiple loops are supported
1866	or not.	2343	or not.
1867		2344
1868	static void	2345	static void
1869	check_cb (EV_P_ ev_timer *w, int revents)	2346	check_cb (EV_P_ ev_timer *w, int revents)
1870	{	2347	{
…		…
1881	Libev can (and often is) directly embedded into host	2358	Libev can (and often is) directly embedded into host
1882	applications. Examples of applications that embed it include the Deliantra	2359	applications. Examples of applications that embed it include the Deliantra
1883	Game Server, the EV perl module, the GNU Virtual Private Ethernet (gvpe)	2360	Game Server, the EV perl module, the GNU Virtual Private Ethernet (gvpe)
1884	and rxvt-unicode.	2361	and rxvt-unicode.
1885		2362
1886	The goal is to enable you to just copy the neecssary files into your	2363	The goal is to enable you to just copy the necessary files into your
1887	source directory without having to change even a single line in them, so	2364	source directory without having to change even a single line in them, so
1888	you can easily upgrade by simply copying (or having a checked-out copy of	2365	you can easily upgrade by simply copying (or having a checked-out copy of
1889	libev somewhere in your source tree).	2366	libev somewhere in your source tree).
1890		2367
1891	=head2 FILESETS	2368	=head2 FILESETS
…		…
1981		2458
1982	If defined to be C<1>, libev will try to detect the availability of the	2459	If defined to be C<1>, libev will try to detect the availability of the
1983	monotonic clock option at both compiletime and runtime. Otherwise no use	2460	monotonic clock option at both compiletime and runtime. Otherwise no use
1984	of the monotonic clock option will be attempted. If you enable this, you	2461	of the monotonic clock option will be attempted. If you enable this, you
1985	usually have to link against librt or something similar. Enabling it when	2462	usually have to link against librt or something similar. Enabling it when
1986	the functionality isn't available is safe, though, althoguh you have	2463	the functionality isn't available is safe, though, although you have
1987	to make sure you link against any libraries where the C<clock_gettime>	2464	to make sure you link against any libraries where the C<clock_gettime>
1988	function is hiding in (often F<-lrt>).	2465	function is hiding in (often F<-lrt>).
1989		2466
1990	=item EV_USE_REALTIME	2467	=item EV_USE_REALTIME
1991		2468
1992	If defined to be C<1>, libev will try to detect the availability of the	2469	If defined to be C<1>, libev will try to detect the availability of the
1993	realtime clock option at compiletime (and assume its availability at	2470	realtime clock option at compiletime (and assume its availability at
1994	runtime if successful). Otherwise no use of the realtime clock option will	2471	runtime if successful). Otherwise no use of the realtime clock option will
1995	be attempted. This effectively replaces C<gettimeofday> by C<clock_get	2472	be attempted. This effectively replaces C<gettimeofday> by C<clock_get
1996	(CLOCK_REALTIME, ...)> and will not normally affect correctness. See tzhe note about libraries	2473	(CLOCK_REALTIME, ...)> and will not normally affect correctness. See the
1997	in the description of C<EV_USE_MONOTONIC>, though.	2474	note about libraries in the description of C<EV_USE_MONOTONIC>, though.
		2475
		2476	=item EV_USE_NANOSLEEP
		2477
		2478	If defined to be C<1>, libev will assume that C<nanosleep ()> is available
		2479	and will use it for delays. Otherwise it will use C<select ()>.
1998		2480
1999	=item EV_USE_SELECT	2481	=item EV_USE_SELECT
2000		2482
2001	If undefined or defined to be C<1>, libev will compile in support for the	2483	If undefined or defined to be C<1>, libev will compile in support for the
2002	C<select>(2) backend. No attempt at autodetection will be done: if no	2484	C<select>(2) backend. No attempt at autodetection will be done: if no
…		…
2020	wants osf handles on win32 (this is the case when the select to	2502	wants osf handles on win32 (this is the case when the select to
2021	be used is the winsock select). This means that it will call	2503	be used is the winsock select). This means that it will call
2022	C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,	2504	C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,
2023	it is assumed that all these functions actually work on fds, even	2505	it is assumed that all these functions actually work on fds, even
2024	on win32. Should not be defined on non-win32 platforms.	2506	on win32. Should not be defined on non-win32 platforms.
		2507
		2508	=item EV_FD_TO_WIN32_HANDLE
		2509
		2510	If C<EV_SELECT_IS_WINSOCKET> is enabled, then libev needs a way to map
		2511	file descriptors to socket handles. When not defining this symbol (the
		2512	default), then libev will call C<_get_osfhandle>, which is usually
		2513	correct. In some cases, programs use their own file descriptor management,
		2514	in which case they can provide this function to map fds to socket handles.
2025		2515
2026	=item EV_USE_POLL	2516	=item EV_USE_POLL
2027		2517
2028	If defined to be C<1>, libev will compile in support for the C<poll>(2)	2518	If defined to be C<1>, libev will compile in support for the C<poll>(2)
2029	backend. Otherwise it will be enabled on non-win32 platforms. It	2519	backend. Otherwise it will be enabled on non-win32 platforms. It
…		…
2066	be detected at runtime.	2556	be detected at runtime.
2067		2557
2068	=item EV_H	2558	=item EV_H
2069		2559
2070	The name of the F<ev.h> header file used to include it. The default if	2560	The name of the F<ev.h> header file used to include it. The default if
2071	undefined is C<< <ev.h> >> in F<event.h> and C<"ev.h"> in F<ev.c>. This	2561	undefined is C<"ev.h"> in F<event.h> and F<ev.c>. This can be used to
2072	can be used to virtually rename the F<ev.h> header file in case of conflicts.	2562	virtually rename the F<ev.h> header file in case of conflicts.
2073		2563
2074	=item EV_CONFIG_H	2564	=item EV_CONFIG_H
2075		2565
2076	If C<EV_STANDALONE> isn't C<1>, this variable can be used to override	2566	If C<EV_STANDALONE> isn't C<1>, this variable can be used to override
2077	F<ev.c>'s idea of where to find the F<config.h> file, similarly to	2567	F<ev.c>'s idea of where to find the F<config.h> file, similarly to
2078	C<EV_H>, above.	2568	C<EV_H>, above.
2079		2569
2080	=item EV_EVENT_H	2570	=item EV_EVENT_H
2081		2571
2082	Similarly to C<EV_H>, this macro can be used to override F<event.c>'s idea	2572	Similarly to C<EV_H>, this macro can be used to override F<event.c>'s idea
2083	of how the F<event.h> header can be found.	2573	of how the F<event.h> header can be found, the dfeault is C<"event.h">.
2084		2574
2085	=item EV_PROTOTYPES	2575	=item EV_PROTOTYPES
2086		2576
2087	If defined to be C<0>, then F<ev.h> will not define any function	2577	If defined to be C<0>, then F<ev.h> will not define any function
2088	prototypes, but still define all the structs and other symbols. This is	2578	prototypes, but still define all the structs and other symbols. This is
…		…
2095	will have the C<struct ev_loop *> as first argument, and you can create	2585	will have the C<struct ev_loop *> as first argument, and you can create
2096	additional independent event loops. Otherwise there will be no support	2586	additional independent event loops. Otherwise there will be no support
2097	for multiple event loops and there is no first event loop pointer	2587	for multiple event loops and there is no first event loop pointer
2098	argument. Instead, all functions act on the single default loop.	2588	argument. Instead, all functions act on the single default loop.
2099		2589
		2590	=item EV_MINPRI
		2591
		2592	=item EV_MAXPRI
		2593
		2594	The range of allowed priorities. C<EV_MINPRI> must be smaller or equal to
		2595	C<EV_MAXPRI>, but otherwise there are no non-obvious limitations. You can
		2596	provide for more priorities by overriding those symbols (usually defined
		2597	to be C<-2> and C<2>, respectively).
		2598
		2599	When doing priority-based operations, libev usually has to linearly search
		2600	all the priorities, so having many of them (hundreds) uses a lot of space
		2601	and time, so using the defaults of five priorities (-2 .. +2) is usually
		2602	fine.
		2603
		2604	If your embedding app does not need any priorities, defining these both to
		2605	C<0> will save some memory and cpu.
		2606
2100	=item EV_PERIODIC_ENABLE	2607	=item EV_PERIODIC_ENABLE
2101		2608
2102	If undefined or defined to be C<1>, then periodic timers are supported. If	2609	If undefined or defined to be C<1>, then periodic timers are supported. If
		2610	defined to be C<0>, then they are not. Disabling them saves a few kB of
		2611	code.
		2612
		2613	=item EV_IDLE_ENABLE
		2614
		2615	If undefined or defined to be C<1>, then idle watchers are supported. If
2103	defined to be C<0>, then they are not. Disabling them saves a few kB of	2616	defined to be C<0>, then they are not. Disabling them saves a few kB of
2104	code.	2617	code.
2105		2618
2106	=item EV_EMBED_ENABLE	2619	=item EV_EMBED_ENABLE
2107		2620
…		…
2131	than enough. If you need to manage thousands of children you might want to	2644	than enough. If you need to manage thousands of children you might want to
2132	increase this value (I<must> be a power of two).	2645	increase this value (I<must> be a power of two).
2133		2646
2134	=item EV_INOTIFY_HASHSIZE	2647	=item EV_INOTIFY_HASHSIZE
2135		2648
2136	C<ev_staz> watchers use a small hash table to distribute workload by	2649	C<ev_stat> watchers use a small hash table to distribute workload by
2137	inotify watch id. The default size is C<16> (or C<1> with C<EV_MINIMAL>),	2650	inotify watch id. The default size is C<16> (or C<1> with C<EV_MINIMAL>),
2138	usually more than enough. If you need to manage thousands of C<ev_stat>	2651	usually more than enough. If you need to manage thousands of C<ev_stat>
2139	watchers you might want to increase this value (I<must> be a power of	2652	watchers you might want to increase this value (I<must> be a power of
2140	two).	2653	two).
2141		2654
…		…
2158		2671
2159	=item ev_set_cb (ev, cb)	2672	=item ev_set_cb (ev, cb)
2160		2673
2161	Can be used to change the callback member declaration in each watcher,	2674	Can be used to change the callback member declaration in each watcher,
2162	and the way callbacks are invoked and set. Must expand to a struct member	2675	and the way callbacks are invoked and set. Must expand to a struct member
2163	definition and a statement, respectively. See the F<ev.v> header file for	2676	definition and a statement, respectively. See the F<ev.h> header file for
2164	their default definitions. One possible use for overriding these is to	2677	their default definitions. One possible use for overriding these is to
2165	avoid the C<struct ev_loop *> as first argument in all cases, or to use	2678	avoid the C<struct ev_loop *> as first argument in all cases, or to use
2166	method calls instead of plain function calls in C++.	2679	method calls instead of plain function calls in C++.
		2680
		2681	=head2 EXPORTED API SYMBOLS
		2682
		2683	If you need to re-export the API (e.g. via a dll) and you need a list of
		2684	exported symbols, you can use the provided F<Symbol.*> files which list
		2685	all public symbols, one per line:
		2686
		2687	Symbols.ev for libev proper
		2688	Symbols.event for the libevent emulation
		2689
		2690	This can also be used to rename all public symbols to avoid clashes with
		2691	multiple versions of libev linked together (which is obviously bad in
		2692	itself, but sometimes it is inconvinient to avoid this).
		2693
		2694	A sed command like this will create wrapper C<#define>'s that you need to
		2695	include before including F<ev.h>:
		2696
		2697	<Symbols.ev sed -e "s/.*/#define & myprefix_&/" >wrap.h
		2698
		2699	This would create a file F<wrap.h> which essentially looks like this:
		2700
		2701	#define ev_backend myprefix_ev_backend
		2702	#define ev_check_start myprefix_ev_check_start
		2703	#define ev_check_stop myprefix_ev_check_stop
		2704	...
2167		2705
2168	=head2 EXAMPLES	2706	=head2 EXAMPLES
2169		2707
2170	For a real-world example of a program the includes libev	2708	For a real-world example of a program the includes libev
2171	verbatim, you can have a look at the EV perl module	2709	verbatim, you can have a look at the EV perl module
…		…
2200		2738
2201	In this section the complexities of (many of) the algorithms used inside	2739	In this section the complexities of (many of) the algorithms used inside
2202	libev will be explained. For complexity discussions about backends see the	2740	libev will be explained. For complexity discussions about backends see the
2203	documentation for C<ev_default_init>.	2741	documentation for C<ev_default_init>.
2204		2742
		2743	All of the following are about amortised time: If an array needs to be
		2744	extended, libev needs to realloc and move the whole array, but this
		2745	happens asymptotically never with higher number of elements, so O(1) might
		2746	mean it might do a lengthy realloc operation in rare cases, but on average
		2747	it is much faster and asymptotically approaches constant time.
		2748
2205	=over 4	2749	=over 4
2206		2750
2207	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)	2751	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)
2208		2752
		2753	This means that, when you have a watcher that triggers in one hour and
		2754	there are 100 watchers that would trigger before that then inserting will
		2755	have to skip roughly seven (C<ld 100>) of these watchers.
		2756
2209	=item Changing timer/periodic watchers (by autorepeat, again): O(log skipped_other_timers)	2757	=item Changing timer/periodic watchers (by autorepeat or calling again): O(log skipped_other_timers)
		2758
		2759	That means that changing a timer costs less than removing/adding them
		2760	as only the relative motion in the event queue has to be paid for.
2210		2761
2211	=item Starting io/check/prepare/idle/signal/child watchers: O(1)	2762	=item Starting io/check/prepare/idle/signal/child watchers: O(1)
2212		2763
		2764	These just add the watcher into an array or at the head of a list.
		2765
2213	=item Stopping check/prepare/idle watchers: O(1)	2766	=item Stopping check/prepare/idle watchers: O(1)
2214		2767
2215	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))	2768	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))
2216		2769
		2770	These watchers are stored in lists then need to be walked to find the
		2771	correct watcher to remove. The lists are usually short (you don't usually
		2772	have many watchers waiting for the same fd or signal).
		2773
2217	=item Finding the next timer per loop iteration: O(1)	2774	=item Finding the next timer in each loop iteration: O(1)
		2775
		2776	By virtue of using a binary heap, the next timer is always found at the
		2777	beginning of the storage array.
2218		2778
2219	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)	2779	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)
2220		2780
2221	=item Activating one watcher: O(1)	2781	A change means an I/O watcher gets started or stopped, which requires
		2782	libev to recalculate its status (and possibly tell the kernel, depending
		2783	on backend and wether C<ev_io_set> was used).
		2784
		2785	=item Activating one watcher (putting it into the pending state): O(1)
		2786
		2787	=item Priority handling: O(number_of_priorities)
		2788
		2789	Priorities are implemented by allocating some space for each
		2790	priority. When doing priority-based operations, libev usually has to
		2791	linearly search all the priorities, but starting/stopping and activating
		2792	watchers becomes O(1) w.r.t. prioritiy handling.
2222		2793
2223	=back	2794	=back
2224		2795
2225		2796
		2797	=head1 Win32 platform limitations and workarounds
		2798
		2799	Win32 doesn't support any of the standards (e.g. POSIX) that libev
		2800	requires, and its I/O model is fundamentally incompatible with the POSIX
		2801	model. Libev still offers limited functionality on this platform in
		2802	the form of the C<EVBACKEND_SELECT> backend, and only supports socket
		2803	descriptors. This only applies when using Win32 natively, not when using
		2804	e.g. cygwin.
		2805
		2806	There is no supported compilation method available on windows except
		2807	embedding it into other applications.
		2808
		2809	Due to the many, low, and arbitrary limits on the win32 platform and the
		2810	abysmal performance of winsockets, using a large number of sockets is not
		2811	recommended (and not reasonable). If your program needs to use more than
		2812	a hundred or so sockets, then likely it needs to use a totally different
		2813	implementation for windows, as libev offers the POSIX model, which cannot
		2814	be implemented efficiently on windows (microsoft monopoly games).
		2815
		2816	=over 4
		2817
		2818	=item The winsocket select function
		2819
		2820	The winsocket C<select> function doesn't follow POSIX in that it requires
		2821	socket I<handles> and not socket I<file descriptors>. This makes select
		2822	very inefficient, and also requires a mapping from file descriptors
		2823	to socket handles. See the discussion of the C<EV_SELECT_USE_FD_SET>,
		2824	C<EV_SELECT_IS_WINSOCKET> and C<EV_FD_TO_WIN32_HANDLE> preprocessor
		2825	symbols for more info.
		2826
		2827	The configuration for a "naked" win32 using the microsoft runtime
		2828	libraries and raw winsocket select is:
		2829
		2830	#define EV_USE_SELECT 1
		2831	#define EV_SELECT_IS_WINSOCKET 1 /* forces EV_SELECT_USE_FD_SET, too */
		2832
		2833	Note that winsockets handling of fd sets is O(n), so you can easily get a
		2834	complexity in the O(n²) range when using win32.
		2835
		2836	=item Limited number of file descriptors
		2837
		2838	Windows has numerous arbitrary (and low) limits on things. Early versions
		2839	of winsocket's select only supported waiting for a max. of C<64> handles
		2840	(probably owning to the fact that all windows kernels can only wait for
		2841	C<64> things at the same time internally; microsoft recommends spawning a
		2842	chain of threads and wait for 63 handles and the previous thread in each).
		2843
		2844	Newer versions support more handles, but you need to define C<FD_SETSIZE>
		2845	to some high number (e.g. C<2048>) before compiling the winsocket select
		2846	call (which might be in libev or elsewhere, for example, perl does its own
		2847	select emulation on windows).
		2848
		2849	Another limit is the number of file descriptors in the microsoft runtime
		2850	libraries, which by default is C<64> (there must be a hidden I<64> fetish
		2851	or something like this inside microsoft). You can increase this by calling
		2852	C<_setmaxstdio>, which can increase this limit to C<2048> (another
		2853	arbitrary limit), but is broken in many versions of the microsoft runtime
		2854	libraries.
		2855
		2856	This might get you to about C<512> or C<2048> sockets (depending on
		2857	windows version and/or the phase of the moon). To get more, you need to
		2858	wrap all I/O functions and provide your own fd management, but the cost of
		2859	calling select (O(n²)) will likely make this unworkable.
		2860
		2861	=back
		2862
		2863
2226	=head1 AUTHOR	2864	=head1 AUTHOR
2227		2865
2228	Marc Lehmann <libev@schmorp.de>.	2866	Marc Lehmann <libev@schmorp.de>.
2229		2867

Diff Legend

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