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

Comparing libev/ev.pod (file contents):
Revision 1.68 by root, Fri Dec 7 18:09:43 2007 UTC vs.
Revision 1.107 by root, Mon Dec 24 04:34:00 2007 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.
350		407
351	=item C<EVBACKEND_ALL>	408	=item C<EVBACKEND_ALL>
352		409
353	Try all backends (even potentially broken ones that wouldn't be tried	410	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	411	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as
355	C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>.	412	C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>.
		413
		414	It is definitely not recommended to use this flag.
356		415
357	=back	416	=back
358		417
359	If one or more of these are ored into the flags value, then only these	418	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	419	backends will be tried (in the reverse order as given here). If none are
…		…
395	Destroys the default loop again (frees all memory and kernel state	454	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	455	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	456	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>	457	responsibility to either stop all watchers cleanly yoursef I<before>
399	calling this function, or cope with the fact afterwards (which is usually	458	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	459	the easiest thing, you can just ignore the watchers and/or C<free ()> them
401	for example).	460	for example).
		461
		462	Note that certain global state, such as signal state, will not be freed by
		463	this function, and related watchers (such as signal and child watchers)
		464	would need to be stopped manually.
		465
		466	In general it is not advisable to call this function except in the
		467	rare occasion where you really need to free e.g. the signal handling
		468	pipe fds. If you need dynamically allocated loops it is better to use
		469	C<ev_loop_new> and C<ev_loop_destroy>).
402		470
403	=item ev_loop_destroy (loop)	471	=item ev_loop_destroy (loop)
404		472
405	Like C<ev_default_destroy>, but destroys an event loop created by an	473	Like C<ev_default_destroy>, but destroys an event loop created by an
406	earlier call to C<ev_loop_new>.	474	earlier call to C<ev_loop_new>.
…		…
451		519
452	Returns the current "event loop time", which is the time the event loop	520	Returns the current "event loop time", which is the time the event loop
453	received events and started processing them. This timestamp does not	521	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	522	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	523	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).	524	event occurring (or more correctly, libev finding out about it).
457		525
458	=item ev_loop (loop, int flags)	526	=item ev_loop (loop, int flags)
459		527
460	Finally, this is it, the event handler. This function usually is called	528	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	529	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	550	libev watchers. However, a pair of C<ev_prepare>/C<ev_check> watchers is
483	usually a better approach for this kind of thing.	551	usually a better approach for this kind of thing.
484		552
485	Here are the gory details of what C<ev_loop> does:	553	Here are the gory details of what C<ev_loop> does:
486		554
		555	- Before the first iteration, call any pending watchers.
487	* If there are no active watchers (reference count is zero), return.	556	* If there are no active watchers (reference count is zero), return.
488	- Queue prepare watchers and then call all outstanding watchers.	557	- Queue all prepare watchers and then call all outstanding watchers.
489	- If we have been forked, recreate the kernel state.	558	- If we have been forked, recreate the kernel state.
490	- Update the kernel state with all outstanding changes.	559	- Update the kernel state with all outstanding changes.
491	- Update the "event loop time".	560	- Update the "event loop time".
492	- Calculate for how long to block.	561	- Calculate for how long to block.
493	- Block the process, waiting for any events.	562	- Block the process, waiting for any events.
…		…
544	Example: For some weird reason, unregister the above signal handler again.	613	Example: For some weird reason, unregister the above signal handler again.
545		614
546	ev_ref (loop);	615	ev_ref (loop);
547	ev_signal_stop (loop, &exitsig);	616	ev_signal_stop (loop, &exitsig);
548		617
		618	=item ev_set_io_collect_interval (loop, ev_tstamp interval)
		619
		620	=item ev_set_timeout_collect_interval (loop, ev_tstamp interval)
		621
		622	These advanced functions influence the time that libev will spend waiting
		623	for events. Both are by default C<0>, meaning that libev will try to
		624	invoke timer/periodic callbacks and I/O callbacks with minimum latency.
		625
		626	Setting these to a higher value (the C<interval> I<must> be >= C<0>)
		627	allows libev to delay invocation of I/O and timer/periodic callbacks to
		628	increase efficiency of loop iterations.
		629
		630	The background is that sometimes your program runs just fast enough to
		631	handle one (or very few) event(s) per loop iteration. While this makes
		632	the program responsive, it also wastes a lot of CPU time to poll for new
		633	events, especially with backends like C<select ()> which have a high
		634	overhead for the actual polling but can deliver many events at once.
		635
		636	By setting a higher I<io collect interval> you allow libev to spend more
		637	time collecting I/O events, so you can handle more events per iteration,
		638	at the cost of increasing latency. Timeouts (both C<ev_periodic> and
		639	C<ev_timer>) will be not affected. Setting this to a non-null value will
		640	introduce an additional C<ev_sleep ()> call into most loop iterations.
		641
		642	Likewise, by setting a higher I<timeout collect interval> you allow libev
		643	to spend more time collecting timeouts, at the expense of increased
		644	latency (the watcher callback will be called later). C<ev_io> watchers
		645	will not be affected. Setting this to a non-null value will not introduce
		646	any overhead in libev.
		647
		648	Many (busy) programs can usually benefit by setting the io collect
		649	interval to a value near C<0.1> or so, which is often enough for
		650	interactive servers (of course not for games), likewise for timeouts. It
		651	usually doesn't make much sense to set it to a lower value than C<0.01>,
		652	as this approsaches the timing granularity of most systems.
		653
549	=back	654	=back
550		655
551		656
552	=head1 ANATOMY OF A WATCHER	657	=head1 ANATOMY OF A WATCHER
553		658
…		…
732	=item bool ev_is_pending (ev_TYPE *watcher)	837	=item bool ev_is_pending (ev_TYPE *watcher)
733		838
734	Returns a true value iff the watcher is pending, (i.e. it has outstanding	839	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	840	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	841	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	842	C<ev_TYPE_set> is safe), you must not change its priority, and you must
738	libev (e.g. you cnanot C<free ()> it).	843	make sure the watcher is available to libev (e.g. you cannot C<free ()>
		844	it).
739		845
740	=item callback ev_cb (ev_TYPE *watcher)	846	=item callback ev_cb (ev_TYPE *watcher)
741		847
742	Returns the callback currently set on the watcher.	848	Returns the callback currently set on the watcher.
743		849
…		…
762	watchers on the same event and make sure one is called first.	868	watchers on the same event and make sure one is called first.
763		869
764	If you need to suppress invocation when higher priority events are pending	870	If you need to suppress invocation when higher priority events are pending
765	you need to look at C<ev_idle> watchers, which provide this functionality.	871	you need to look at C<ev_idle> watchers, which provide this functionality.
766		872
		873	You I<must not> change the priority of a watcher as long as it is active or
		874	pending.
		875
767	The default priority used by watchers when no priority has been set is	876	The default priority used by watchers when no priority has been set is
768	always C<0>, which is supposed to not be too high and not be too low :).	877	always C<0>, which is supposed to not be too high and not be too low :).
769		878
770	Setting a priority outside the range of C<EV_MINPRI> to C<EV_MAXPRI> is	879	Setting a priority outside the range of C<EV_MINPRI> to C<EV_MAXPRI> is
771	fine, as long as you do not mind that the priority value you query might	880	fine, as long as you do not mind that the priority value you query might
772	or might not have been adjusted to be within valid range.	881	or might not have been adjusted to be within valid range.
		882
		883	=item ev_invoke (loop, ev_TYPE *watcher, int revents)
		884
		885	Invoke the C<watcher> with the given C<loop> and C<revents>. Neither
		886	C<loop> nor C<revents> need to be valid as long as the watcher callback
		887	can deal with that fact.
		888
		889	=item int ev_clear_pending (loop, ev_TYPE *watcher)
		890
		891	If the watcher is pending, this function returns clears its pending status
		892	and returns its C<revents> bitset (as if its callback was invoked). If the
		893	watcher isn't pending it does nothing and returns C<0>.
773		894
774	=back	895	=back
775		896
776		897
777	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER	898	=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER
…		…
887	play around with an Xlib connection), then you have to seperately re-test	1008	play around with an Xlib connection), then you have to seperately re-test
888	whether a file descriptor is really ready with a known-to-be good interface	1009	whether a file descriptor is really ready with a known-to-be good interface
889	such as poll (fortunately in our Xlib example, Xlib already does this on	1010	such as poll (fortunately in our Xlib example, Xlib already does this on
890	its own, so its quite safe to use).	1011	its own, so its quite safe to use).
891		1012
		1013	=head3 The special problem of disappearing file descriptors
		1014
		1015	Some backends (e.g. kqueue, epoll) need to be told about closing a file
		1016	descriptor (either by calling C<close> explicitly or by any other means,
		1017	such as C<dup>). The reason is that you register interest in some file
		1018	descriptor, but when it goes away, the operating system will silently drop
		1019	this interest. If another file descriptor with the same number then is
		1020	registered with libev, there is no efficient way to see that this is, in
		1021	fact, a different file descriptor.
		1022
		1023	To avoid having to explicitly tell libev about such cases, libev follows
		1024	the following policy: Each time C<ev_io_set> is being called, libev
		1025	will assume that this is potentially a new file descriptor, otherwise
		1026	it is assumed that the file descriptor stays the same. That means that
		1027	you I<have> to call C<ev_io_set> (or C<ev_io_init>) when you change the
		1028	descriptor even if the file descriptor number itself did not change.
		1029
		1030	This is how one would do it normally anyway, the important point is that
		1031	the libev application should not optimise around libev but should leave
		1032	optimisations to libev.
		1033
		1034	=head3 The special problem of dup'ed file descriptors
		1035
		1036	Some backends (e.g. epoll), cannot register events for file descriptors,
		1037	but only events for the underlying file descriptions. That means when you
		1038	have C<dup ()>'ed file descriptors and register events for them, only one
		1039	file descriptor might actually receive events.
		1040
		1041	There is no workaround possible except not registering events
		1042	for potentially C<dup ()>'ed file descriptors, or to resort to
		1043	C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>.
		1044
		1045	=head3 The special problem of fork
		1046
		1047	Some backends (epoll, kqueue) do not support C<fork ()> at all or exhibit
		1048	useless behaviour. Libev fully supports fork, but needs to be told about
		1049	it in the child.
		1050
		1051	To support fork in your programs, you either have to call
		1052	C<ev_default_fork ()> or C<ev_loop_fork ()> after a fork in the child,
		1053	enable C<EVFLAG_FORKCHECK>, or resort to C<EVBACKEND_SELECT> or
		1054	C<EVBACKEND_POLL>.
		1055
		1056
		1057	=head3 Watcher-Specific Functions
		1058
892	=over 4	1059	=over 4
893		1060
894	=item ev_io_init (ev_io *, callback, int fd, int events)	1061	=item ev_io_init (ev_io *, callback, int fd, int events)
895		1062
896	=item ev_io_set (ev_io *, int fd, int events)	1063	=item ev_io_set (ev_io *, int fd, int events)
…		…
948	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);	1115	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);
949		1116
950	The callback is guarenteed to be invoked only when its timeout has passed,	1117	The callback is guarenteed to be invoked only when its timeout has passed,
951	but if multiple timers become ready during the same loop iteration then	1118	but if multiple timers become ready during the same loop iteration then
952	order of execution is undefined.	1119	order of execution is undefined.
		1120
		1121	=head3 Watcher-Specific Functions and Data Members
953		1122
954	=over 4	1123	=over 4
955		1124
956	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)	1125	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)
957		1126
…		…
1053	but on wallclock time (absolute time). You can tell a periodic watcher	1222	but on wallclock time (absolute time). You can tell a periodic watcher
1054	to trigger "at" some specific point in time. For example, if you tell a	1223	to trigger "at" some specific point in time. For example, if you tell a
1055	periodic watcher to trigger in 10 seconds (by specifiying e.g. C<ev_now ()	1224	periodic watcher to trigger in 10 seconds (by specifiying e.g. C<ev_now ()
1056	+ 10.>) and then reset your system clock to the last year, then it will	1225	+ 10.>) and then reset your system clock to the last year, then it will
1057	take a year to trigger the event (unlike an C<ev_timer>, which would trigger	1226	take a year to trigger the event (unlike an C<ev_timer>, which would trigger
1058	roughly 10 seconds later and of course not if you reset your system time	1227	roughly 10 seconds later).
1059	again).
1060		1228
1061	They can also be used to implement vastly more complex timers, such as	1229	They can also be used to implement vastly more complex timers, such as
1062	triggering an event on eahc midnight, local time.	1230	triggering an event on each midnight, local time or other, complicated,
		1231	rules.
1063		1232
1064	As with timers, the callback is guarenteed to be invoked only when the	1233	As with timers, the callback is guarenteed to be invoked only when the
1065	time (C<at>) has been passed, but if multiple periodic timers become ready	1234	time (C<at>) has been passed, but if multiple periodic timers become ready
1066	during the same loop iteration then order of execution is undefined.	1235	during the same loop iteration then order of execution is undefined.
1067		1236
		1237	=head3 Watcher-Specific Functions and Data Members
		1238
1068	=over 4	1239	=over 4
1069		1240
1070	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)	1241	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)
1071		1242
1072	=item ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)	1243	=item ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)
…		…
1074	Lots of arguments, lets sort it out... There are basically three modes of	1245	Lots of arguments, lets sort it out... There are basically three modes of
1075	operation, and we will explain them from simplest to complex:	1246	operation, and we will explain them from simplest to complex:
1076		1247
1077	=over 4	1248	=over 4
1078		1249
1079	=item * absolute timer (interval = reschedule_cb = 0)	1250	=item * absolute timer (at = time, interval = reschedule_cb = 0)
1080		1251
1081	In this configuration the watcher triggers an event at the wallclock time	1252	In this configuration the watcher triggers an event at the wallclock time
1082	C<at> and doesn't repeat. It will not adjust when a time jump occurs,	1253	C<at> and doesn't repeat. It will not adjust when a time jump occurs,
1083	that is, if it is to be run at January 1st 2011 then it will run when the	1254	that is, if it is to be run at January 1st 2011 then it will run when the
1084	system time reaches or surpasses this time.	1255	system time reaches or surpasses this time.
1085		1256
1086	=item * non-repeating interval timer (interval > 0, reschedule_cb = 0)	1257	=item * non-repeating interval timer (at = offset, interval > 0, reschedule_cb = 0)
1087		1258
1088	In this mode the watcher will always be scheduled to time out at the next	1259	In this mode the watcher will always be scheduled to time out at the next
1089	C<at + N * interval> time (for some integer N) and then repeat, regardless	1260	C<at + N * interval> time (for some integer N, which can also be negative)
1090	of any time jumps.	1261	and then repeat, regardless of any time jumps.
1091		1262
1092	This can be used to create timers that do not drift with respect to system	1263	This can be used to create timers that do not drift with respect to system
1093	time:	1264	time:
1094		1265
1095	ev_periodic_set (&periodic, 0., 3600., 0);	1266	ev_periodic_set (&periodic, 0., 3600., 0);
…		…
1101		1272
1102	Another way to think about it (for the mathematically inclined) is that	1273	Another way to think about it (for the mathematically inclined) is that
1103	C<ev_periodic> will try to run the callback in this mode at the next possible	1274	C<ev_periodic> will try to run the callback in this mode at the next possible
1104	time where C<time = at (mod interval)>, regardless of any time jumps.	1275	time where C<time = at (mod interval)>, regardless of any time jumps.
1105		1276
		1277	For numerical stability it is preferable that the C<at> value is near
		1278	C<ev_now ()> (the current time), but there is no range requirement for
		1279	this value.
		1280
1106	=item * manual reschedule mode (reschedule_cb = callback)	1281	=item * manual reschedule mode (at and interval ignored, reschedule_cb = callback)
1107		1282
1108	In this mode the values for C<interval> and C<at> are both being	1283	In this mode the values for C<interval> and C<at> are both being
1109	ignored. Instead, each time the periodic watcher gets scheduled, the	1284	ignored. Instead, each time the periodic watcher gets scheduled, the
1110	reschedule callback will be called with the watcher as first, and the	1285	reschedule callback will be called with the watcher as first, and the
1111	current time as second argument.	1286	current time as second argument.
1112		1287
1113	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher,	1288	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher,
1114	ever, or make any event loop modifications>. If you need to stop it,	1289	ever, or make any event loop modifications>. If you need to stop it,
1115	return C<now + 1e30> (or so, fudge fudge) and stop it afterwards (e.g. by	1290	return C<now + 1e30> (or so, fudge fudge) and stop it afterwards (e.g. by
1116	starting a prepare watcher).	1291	starting an C<ev_prepare> watcher, which is legal).
1117		1292
1118	Its prototype is C<ev_tstamp (reschedule_cb)(struct ev_periodic w,	1293	Its prototype is C<ev_tstamp (reschedule_cb)(struct ev_periodic w,
1119	ev_tstamp now)>, e.g.:	1294	ev_tstamp now)>, e.g.:
1120		1295
1121	static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)	1296	static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)
…		…
1144	Simply stops and restarts the periodic watcher again. This is only useful	1319	Simply stops and restarts the periodic watcher again. This is only useful
1145	when you changed some parameters or the reschedule callback would return	1320	when you changed some parameters or the reschedule callback would return
1146	a different time than the last time it was called (e.g. in a crond like	1321	a different time than the last time it was called (e.g. in a crond like
1147	program when the crontabs have changed).	1322	program when the crontabs have changed).
1148		1323
		1324	=item ev_tstamp offset [read-write]
		1325
		1326	When repeating, this contains the offset value, otherwise this is the
		1327	absolute point in time (the C<at> value passed to C<ev_periodic_set>).
		1328
		1329	Can be modified any time, but changes only take effect when the periodic
		1330	timer fires or C<ev_periodic_again> is being called.
		1331
1149	=item ev_tstamp interval [read-write]	1332	=item ev_tstamp interval [read-write]
1150		1333
1151	The current interval value. Can be modified any time, but changes only	1334	The current interval value. Can be modified any time, but changes only
1152	take effect when the periodic timer fires or C<ev_periodic_again> is being	1335	take effect when the periodic timer fires or C<ev_periodic_again> is being
1153	called.	1336	called.
…		…
1155	=item ev_tstamp (reschedule_cb)(struct ev_periodic w, ev_tstamp now) [read-write]	1338	=item ev_tstamp (reschedule_cb)(struct ev_periodic w, ev_tstamp now) [read-write]
1156		1339
1157	The current reschedule callback, or C<0>, if this functionality is	1340	The current reschedule callback, or C<0>, if this functionality is
1158	switched off. Can be changed any time, but changes only take effect when	1341	switched off. Can be changed any time, but changes only take effect when
1159	the periodic timer fires or C<ev_periodic_again> is being called.	1342	the periodic timer fires or C<ev_periodic_again> is being called.
		1343
		1344	=item ev_tstamp at [read-only]
		1345
		1346	When active, contains the absolute time that the watcher is supposed to
		1347	trigger next.
1160		1348
1161	=back	1349	=back
1162		1350
1163	Example: Call a callback every hour, or, more precisely, whenever the	1351	Example: Call a callback every hour, or, more precisely, whenever the
1164	system clock is divisible by 3600. The callback invocation times have	1352	system clock is divisible by 3600. The callback invocation times have
…		…
1206	with the kernel (thus it coexists with your own signal handlers as long	1394	with the kernel (thus it coexists with your own signal handlers as long
1207	as you don't register any with libev). Similarly, when the last signal	1395	as you don't register any with libev). Similarly, when the last signal
1208	watcher for a signal is stopped libev will reset the signal handler to	1396	watcher for a signal is stopped libev will reset the signal handler to
1209	SIG_DFL (regardless of what it was set to before).	1397	SIG_DFL (regardless of what it was set to before).
1210		1398
		1399	=head3 Watcher-Specific Functions and Data Members
		1400
1211	=over 4	1401	=over 4
1212		1402
1213	=item ev_signal_init (ev_signal *, callback, int signum)	1403	=item ev_signal_init (ev_signal *, callback, int signum)
1214		1404
1215	=item ev_signal_set (ev_signal *, int signum)	1405	=item ev_signal_set (ev_signal *, int signum)
…		…
1226		1416
1227	=head2 C<ev_child> - watch out for process status changes	1417	=head2 C<ev_child> - watch out for process status changes
1228		1418
1229	Child watchers trigger when your process receives a SIGCHLD in response to	1419	Child watchers trigger when your process receives a SIGCHLD in response to
1230	some child status changes (most typically when a child of yours dies).	1420	some child status changes (most typically when a child of yours dies).
		1421
		1422	=head3 Watcher-Specific Functions and Data Members
1231		1423
1232	=over 4	1424	=over 4
1233		1425
1234	=item ev_child_init (ev_child *, callback, int pid)	1426	=item ev_child_init (ev_child *, callback, int pid)
1235		1427
…		…
1303	reader). Inotify will be used to give hints only and should not change the	1495	reader). Inotify will be used to give hints only and should not change the
1304	semantics of C<ev_stat> watchers, which means that libev sometimes needs	1496	semantics of C<ev_stat> watchers, which means that libev sometimes needs
1305	to fall back to regular polling again even with inotify, but changes are	1497	to fall back to regular polling again even with inotify, but changes are
1306	usually detected immediately, and if the file exists there will be no	1498	usually detected immediately, and if the file exists there will be no
1307	polling.	1499	polling.
		1500
		1501	=head3 The special problem of stat time resolution
		1502
		1503	The C<stat ()> syscall only supports full-second resolution portably, and
		1504	even on systems where the resolution is higher, many filesystems still
		1505	only support whole seconds.
		1506
		1507	That means that, if the time is the only thing that changes, you might
		1508	miss updates: on the first update, C<ev_stat> detects a change and calls
		1509	your callback, which does something. When there is another update within
		1510	the same second, C<ev_stat> will be unable to detect it.
		1511
		1512	The solution to this is to delay acting on a change for a second (or till
		1513	the next second boundary), using a roughly one-second delay C<ev_timer>
		1514	(C<ev_timer_set (w, 0., 1.01); ev_timer_again (loop, w)>). The C<.01>
		1515	is added to work around small timing inconsistencies of some operating
		1516	systems.
		1517
		1518	=head3 Watcher-Specific Functions and Data Members
1308		1519
1309	=over 4	1520	=over 4
1310		1521
1311	=item ev_stat_init (ev_stat , callback, const char path, ev_tstamp interval)	1522	=item ev_stat_init (ev_stat , callback, const char path, ev_tstamp interval)
1312		1523
…		…
1370	}	1581	}
1371		1582
1372	...	1583	...
1373	ev_stat passwd;	1584	ev_stat passwd;
1374		1585
1375	ev_stat_init (&passwd, passwd_cb, "/etc/passwd");	1586	ev_stat_init (&passwd, passwd_cb, "/etc/passwd", 0.);
1376	ev_stat_start (loop, &passwd);	1587	ev_stat_start (loop, &passwd);
		1588
		1589	Example: Like above, but additionally use a one-second delay so we do not
		1590	miss updates (however, frequent updates will delay processing, too, so
		1591	one might do the work both on C<ev_stat> callback invocation I<and> on
		1592	C<ev_timer> callback invocation).
		1593
		1594	static ev_stat passwd;
		1595	static ev_timer timer;
		1596
		1597	static void
		1598	timer_cb (EV_P_ ev_timer *w, int revents)
		1599	{
		1600	ev_timer_stop (EV_A_ w);
		1601
		1602	/* now it's one second after the most recent passwd change */
		1603	}
		1604
		1605	static void
		1606	stat_cb (EV_P_ ev_stat *w, int revents)
		1607	{
		1608	/* reset the one-second timer */
		1609	ev_timer_again (EV_A_ &timer);
		1610	}
		1611
		1612	...
		1613	ev_stat_init (&passwd, stat_cb, "/etc/passwd", 0.);
		1614	ev_stat_start (loop, &passwd);
		1615	ev_timer_init (&timer, timer_cb, 0., 1.01);
1377		1616
1378		1617
1379	=head2 C<ev_idle> - when you've got nothing better to do...	1618	=head2 C<ev_idle> - when you've got nothing better to do...
1380		1619
1381	Idle watchers trigger events when no other events of the same or higher	1620	Idle watchers trigger events when no other events of the same or higher
…		…
1394		1633
1395	Apart from keeping your process non-blocking (which is a useful	1634	Apart from keeping your process non-blocking (which is a useful
1396	effect on its own sometimes), idle watchers are a good place to do	1635	effect on its own sometimes), idle watchers are a good place to do
1397	"pseudo-background processing", or delay processing stuff to after the	1636	"pseudo-background processing", or delay processing stuff to after the
1398	event loop has handled all outstanding events.	1637	event loop has handled all outstanding events.
		1638
		1639	=head3 Watcher-Specific Functions and Data Members
1399		1640
1400	=over 4	1641	=over 4
1401		1642
1402	=item ev_idle_init (ev_signal *, callback)	1643	=item ev_idle_init (ev_signal *, callback)
1403		1644
…		…
1461	with priority higher than or equal to the event loop and one coroutine	1702	with priority higher than or equal to the event loop and one coroutine
1462	of lower priority, but only once, using idle watchers to keep the event	1703	of lower priority, but only once, using idle watchers to keep the event
1463	loop from blocking if lower-priority coroutines are active, thus mapping	1704	loop from blocking if lower-priority coroutines are active, thus mapping
1464	low-priority coroutines to idle/background tasks).	1705	low-priority coroutines to idle/background tasks).
1465		1706
		1707	It is recommended to give C<ev_check> watchers highest (C<EV_MAXPRI>)
		1708	priority, to ensure that they are being run before any other watchers
		1709	after the poll. Also, C<ev_check> watchers (and C<ev_prepare> watchers,
		1710	too) should not activate ("feed") events into libev. While libev fully
		1711	supports this, they will be called before other C<ev_check> watchers
		1712	did their job. As C<ev_check> watchers are often used to embed other
		1713	(non-libev) event loops those other event loops might be in an unusable
		1714	state until their C<ev_check> watcher ran (always remind yourself to
		1715	coexist peacefully with others).
		1716
		1717	=head3 Watcher-Specific Functions and Data Members
		1718
1466	=over 4	1719	=over 4
1467		1720
1468	=item ev_prepare_init (ev_prepare *, callback)	1721	=item ev_prepare_init (ev_prepare *, callback)
1469		1722
1470	=item ev_check_init (ev_check *, callback)	1723	=item ev_check_init (ev_check *, callback)
…		…
1473	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>	1726	parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>
1474	macros, but using them is utterly, utterly and completely pointless.	1727	macros, but using them is utterly, utterly and completely pointless.
1475		1728
1476	=back	1729	=back
1477		1730
1478	Example: To include a library such as adns, you would add IO watchers	1731	There are a number of principal ways to embed other event loops or modules
1479	and a timeout watcher in a prepare handler, as required by libadns, and	1732	into libev. Here are some ideas on how to include libadns into libev
		1733	(there is a Perl module named C<EV::ADNS> that does this, which you could
		1734	use for an actually working example. Another Perl module named C<EV::Glib>
		1735	embeds a Glib main context into libev, and finally, C<Glib::EV> embeds EV
		1736	into the Glib event loop).
		1737
		1738	Method 1: Add IO watchers and a timeout watcher in a prepare handler,
1480	in a check watcher, destroy them and call into libadns. What follows is	1739	and in a check watcher, destroy them and call into libadns. What follows
1481	pseudo-code only of course:	1740	is pseudo-code only of course. This requires you to either use a low
		1741	priority for the check watcher or use C<ev_clear_pending> explicitly, as
		1742	the callbacks for the IO/timeout watchers might not have been called yet.
1482		1743
1483	static ev_io iow [nfd];	1744	static ev_io iow [nfd];
1484	static ev_timer tw;	1745	static ev_timer tw;
1485		1746
1486	static void	1747	static void
1487	io_cb (ev_loop loop, ev_io w, int revents)	1748	io_cb (ev_loop loop, ev_io w, int revents)
1488	{	1749	{
1489	// set the relevant poll flags
1490	// could also call adns_processreadable etc. here
1491	struct pollfd fd = (struct pollfd )w->data;
1492	if (revents & EV_READ ) fd->revents \|= fd->events & POLLIN;
1493	if (revents & EV_WRITE) fd->revents \|= fd->events & POLLOUT;
1494	}	1750	}
1495		1751
1496	// create io watchers for each fd and a timer before blocking	1752	// create io watchers for each fd and a timer before blocking
1497	static void	1753	static void
1498	adns_prepare_cb (ev_loop loop, ev_prepare w, int revents)	1754	adns_prepare_cb (ev_loop loop, ev_prepare w, int revents)
…		…
1504		1760
1505	/* the callback is illegal, but won't be called as we stop during check */	1761	/* the callback is illegal, but won't be called as we stop during check */
1506	ev_timer_init (&tw, 0, timeout * 1e-3);	1762	ev_timer_init (&tw, 0, timeout * 1e-3);
1507	ev_timer_start (loop, &tw);	1763	ev_timer_start (loop, &tw);
1508		1764
1509	// create on ev_io per pollfd	1765	// create one ev_io per pollfd
1510	for (int i = 0; i < nfd; ++i)	1766	for (int i = 0; i < nfd; ++i)
1511	{	1767	{
1512	ev_io_init (iow + i, io_cb, fds [i].fd,	1768	ev_io_init (iow + i, io_cb, fds [i].fd,
1513	((fds [i].events & POLLIN ? EV_READ : 0)	1769	((fds [i].events & POLLIN ? EV_READ : 0)
1514	\| (fds [i].events & POLLOUT ? EV_WRITE : 0)));	1770	\| (fds [i].events & POLLOUT ? EV_WRITE : 0)));
1515		1771
1516	fds [i].revents = 0;	1772	fds [i].revents = 0;
1517	iow [i].data = fds + i;
1518	ev_io_start (loop, iow + i);	1773	ev_io_start (loop, iow + i);
1519	}	1774	}
1520	}	1775	}
1521		1776
1522	// stop all watchers after blocking	1777	// stop all watchers after blocking
…		…
1524	adns_check_cb (ev_loop loop, ev_check w, int revents)	1779	adns_check_cb (ev_loop loop, ev_check w, int revents)
1525	{	1780	{
1526	ev_timer_stop (loop, &tw);	1781	ev_timer_stop (loop, &tw);
1527		1782
1528	for (int i = 0; i < nfd; ++i)	1783	for (int i = 0; i < nfd; ++i)
		1784	{
		1785	// set the relevant poll flags
		1786	// could also call adns_processreadable etc. here
		1787	struct pollfd *fd = fds + i;
		1788	int revents = ev_clear_pending (iow + i);
		1789	if (revents & EV_READ ) fd->revents \|= fd->events & POLLIN;
		1790	if (revents & EV_WRITE) fd->revents \|= fd->events & POLLOUT;
		1791
		1792	// now stop the watcher
1529	ev_io_stop (loop, iow + i);	1793	ev_io_stop (loop, iow + i);
		1794	}
1530		1795
1531	adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));	1796	adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));
		1797	}
		1798
		1799	Method 2: This would be just like method 1, but you run C<adns_afterpoll>
		1800	in the prepare watcher and would dispose of the check watcher.
		1801
		1802	Method 3: If the module to be embedded supports explicit event
		1803	notification (adns does), you can also make use of the actual watcher
		1804	callbacks, and only destroy/create the watchers in the prepare watcher.
		1805
		1806	static void
		1807	timer_cb (EV_P_ ev_timer *w, int revents)
		1808	{
		1809	adns_state ads = (adns_state)w->data;
		1810	update_now (EV_A);
		1811
		1812	adns_processtimeouts (ads, &tv_now);
		1813	}
		1814
		1815	static void
		1816	io_cb (EV_P_ ev_io *w, int revents)
		1817	{
		1818	adns_state ads = (adns_state)w->data;
		1819	update_now (EV_A);
		1820
		1821	if (revents & EV_READ ) adns_processreadable (ads, w->fd, &tv_now);
		1822	if (revents & EV_WRITE) adns_processwriteable (ads, w->fd, &tv_now);
		1823	}
		1824
		1825	// do not ever call adns_afterpoll
		1826
		1827	Method 4: Do not use a prepare or check watcher because the module you
		1828	want to embed is too inflexible to support it. Instead, youc na override
		1829	their poll function. The drawback with this solution is that the main
		1830	loop is now no longer controllable by EV. The C<Glib::EV> module does
		1831	this.
		1832
		1833	static gint
		1834	event_poll_func (GPollFD *fds, guint nfds, gint timeout)
		1835	{
		1836	int got_events = 0;
		1837
		1838	for (n = 0; n < nfds; ++n)
		1839	// create/start io watcher that sets the relevant bits in fds[n] and increment got_events
		1840
		1841	if (timeout >= 0)
		1842	// create/start timer
		1843
		1844	// poll
		1845	ev_loop (EV_A_ 0);
		1846
		1847	// stop timer again
		1848	if (timeout >= 0)
		1849	ev_timer_stop (EV_A_ &to);
		1850
		1851	// stop io watchers again - their callbacks should have set
		1852	for (n = 0; n < nfds; ++n)
		1853	ev_io_stop (EV_A_ iow [n]);
		1854
		1855	return got_events;
1532	}	1856	}
1533		1857
1534		1858
1535	=head2 C<ev_embed> - when one backend isn't enough...	1859	=head2 C<ev_embed> - when one backend isn't enough...
1536		1860
…		…
1600	ev_embed_start (loop_hi, &embed);	1924	ev_embed_start (loop_hi, &embed);
1601	}	1925	}
1602	else	1926	else
1603	loop_lo = loop_hi;	1927	loop_lo = loop_hi;
1604		1928
		1929	=head3 Watcher-Specific Functions and Data Members
		1930
1605	=over 4	1931	=over 4
1606		1932
1607	=item ev_embed_init (ev_embed , callback, struct ev_loop embedded_loop)	1933	=item ev_embed_init (ev_embed , callback, struct ev_loop embedded_loop)
1608		1934
1609	=item ev_embed_set (ev_embed , callback, struct ev_loop embedded_loop)	1935	=item ev_embed_set (ev_embed , callback, struct ev_loop embedded_loop)
…		…
1618		1944
1619	Make a single, non-blocking sweep over the embedded loop. This works	1945	Make a single, non-blocking sweep over the embedded loop. This works
1620	similarly to C<ev_loop (embedded_loop, EVLOOP_NONBLOCK)>, but in the most	1946	similarly to C<ev_loop (embedded_loop, EVLOOP_NONBLOCK)>, but in the most
1621	apropriate way for embedded loops.	1947	apropriate way for embedded loops.
1622		1948
1623	=item struct ev_loop *loop [read-only]	1949	=item struct ev_loop *other [read-only]
1624		1950
1625	The embedded event loop.	1951	The embedded event loop.
1626		1952
1627	=back	1953	=back
1628		1954
…		…
1635	event loop blocks next and before C<ev_check> watchers are being called,	1961	event loop blocks next and before C<ev_check> watchers are being called,
1636	and only in the child after the fork. If whoever good citizen calling	1962	and only in the child after the fork. If whoever good citizen calling
1637	C<ev_default_fork> cheats and calls it in the wrong process, the fork	1963	C<ev_default_fork> cheats and calls it in the wrong process, the fork
1638	handlers will be invoked, too, of course.	1964	handlers will be invoked, too, of course.
1639		1965
		1966	=head3 Watcher-Specific Functions and Data Members
		1967
1640	=over 4	1968	=over 4
1641		1969
1642	=item ev_fork_init (ev_signal *, callback)	1970	=item ev_fork_init (ev_signal *, callback)
1643		1971
1644	Initialises and configures the fork watcher - it has no parameters of any	1972	Initialises and configures the fork watcher - it has no parameters of any
…		…
1740		2068
1741	To use it,	2069	To use it,
1742		2070
1743	#include <ev++.h>	2071	#include <ev++.h>
1744		2072
1745	(it is not installed by default). This automatically includes F<ev.h>	2073	This automatically includes F<ev.h> and puts all of its definitions (many
1746	and puts all of its definitions (many of them macros) into the global	2074	of them macros) into the global namespace. All C++ specific things are
1747	namespace. All C++ specific things are put into the C<ev> namespace.	2075	put into the C<ev> namespace. It should support all the same embedding
		2076	options as F<ev.h>, most notably C<EV_MULTIPLICITY>.
1748		2077
1749	It should support all the same embedding options as F<ev.h>, most notably	2078	Care has been taken to keep the overhead low. The only data member the C++
1750	C<EV_MULTIPLICITY>.	2079	classes add (compared to plain C-style watchers) is the event loop pointer
		2080	that the watcher is associated with (or no additional members at all if
		2081	you disable C<EV_MULTIPLICITY> when embedding libev).
		2082
		2083	Currently, functions, and static and non-static member functions can be
		2084	used as callbacks. Other types should be easy to add as long as they only
		2085	need one additional pointer for context. If you need support for other
		2086	types of functors please contact the author (preferably after implementing
		2087	it).
1751		2088
1752	Here is a list of things available in the C<ev> namespace:	2089	Here is a list of things available in the C<ev> namespace:
1753		2090
1754	=over 4	2091	=over 4
1755		2092
…		…
1771		2108
1772	All of those classes have these methods:	2109	All of those classes have these methods:
1773		2110
1774	=over 4	2111	=over 4
1775		2112
1776	=item ev::TYPE::TYPE (object , object::method )	2113	=item ev::TYPE::TYPE ()
1777		2114
1778	=item ev::TYPE::TYPE (object , object::method , struct ev_loop *)	2115	=item ev::TYPE::TYPE (struct ev_loop *)
1779		2116
1780	=item ev::TYPE::~TYPE	2117	=item ev::TYPE::~TYPE
1781		2118
1782	The constructor takes a pointer to an object and a method pointer to	2119	The constructor (optionally) takes an event loop to associate the watcher
1783	the event handler callback to call in this class. The constructor calls	2120	with. If it is omitted, it will use C<EV_DEFAULT>.
1784	C<ev_init> for you, which means you have to call the C<set> method	2121
1785	before starting it. If you do not specify a loop then the constructor	2122	The constructor calls C<ev_init> for you, which means you have to call the
1786	automatically associates the default loop with this watcher.	2123	C<set> method before starting it.
		2124
		2125	It will not set a callback, however: You have to call the templated C<set>
		2126	method to set a callback before you can start the watcher.
		2127
		2128	(The reason why you have to use a method is a limitation in C++ which does
		2129	not allow explicit template arguments for constructors).
1787		2130
1788	The destructor automatically stops the watcher if it is active.	2131	The destructor automatically stops the watcher if it is active.
		2132
		2133	=item w->set<class, &class::method> (object *)
		2134
		2135	This method sets the callback method to call. The method has to have a
		2136	signature of C<void (*)(ev_TYPE &, int)>, it receives the watcher as
		2137	first argument and the C<revents> as second. The object must be given as
		2138	parameter and is stored in the C<data> member of the watcher.
		2139
		2140	This method synthesizes efficient thunking code to call your method from
		2141	the C callback that libev requires. If your compiler can inline your
		2142	callback (i.e. it is visible to it at the place of the C<set> call and
		2143	your compiler is good :), then the method will be fully inlined into the
		2144	thunking function, making it as fast as a direct C callback.
		2145
		2146	Example: simple class declaration and watcher initialisation
		2147
		2148	struct myclass
		2149	{
		2150	void io_cb (ev::io &w, int revents) { }
		2151	}
		2152
		2153	myclass obj;
		2154	ev::io iow;
		2155	iow.set <myclass, &myclass::io_cb> (&obj);
		2156
		2157	=item w->set<function> (void *data = 0)
		2158
		2159	Also sets a callback, but uses a static method or plain function as
		2160	callback. The optional C<data> argument will be stored in the watcher's
		2161	C<data> member and is free for you to use.
		2162
		2163	The prototype of the C<function> must be C<void (*)(ev::TYPE &w, int)>.
		2164
		2165	See the method-C<set> above for more details.
		2166
		2167	Example:
		2168
		2169	static void io_cb (ev::io &w, int revents) { }
		2170	iow.set <io_cb> ();
1789		2171
1790	=item w->set (struct ev_loop *)	2172	=item w->set (struct ev_loop *)
1791		2173
1792	Associates a different C<struct ev_loop> with this watcher. You can only	2174	Associates a different C<struct ev_loop> with this watcher. You can only
1793	do this when the watcher is inactive (and not pending either).	2175	do this when the watcher is inactive (and not pending either).
1794		2176
1795	=item w->set ([args])	2177	=item w->set ([args])
1796		2178
1797	Basically the same as C<ev_TYPE_set>, with the same args. Must be	2179	Basically the same as C<ev_TYPE_set>, with the same args. Must be
1798	called at least once. Unlike the C counterpart, an active watcher gets	2180	called at least once. Unlike the C counterpart, an active watcher gets
1799	automatically stopped and restarted.	2181	automatically stopped and restarted when reconfiguring it with this
		2182	method.
1800		2183
1801	=item w->start ()	2184	=item w->start ()
1802		2185
1803	Starts the watcher. Note that there is no C<loop> argument as the	2186	Starts the watcher. Note that there is no C<loop> argument, as the
1804	constructor already takes the loop.	2187	constructor already stores the event loop.
1805		2188
1806	=item w->stop ()	2189	=item w->stop ()
1807		2190
1808	Stops the watcher if it is active. Again, no C<loop> argument.	2191	Stops the watcher if it is active. Again, no C<loop> argument.
1809		2192
1810	=item w->again () C<ev::timer>, C<ev::periodic> only	2193	=item w->again () (C<ev::timer>, C<ev::periodic> only)
1811		2194
1812	For C<ev::timer> and C<ev::periodic>, this invokes the corresponding	2195	For C<ev::timer> and C<ev::periodic>, this invokes the corresponding
1813	C<ev_TYPE_again> function.	2196	C<ev_TYPE_again> function.
1814		2197
1815	=item w->sweep () C<ev::embed> only	2198	=item w->sweep () (C<ev::embed> only)
1816		2199
1817	Invokes C<ev_embed_sweep>.	2200	Invokes C<ev_embed_sweep>.
1818		2201
1819	=item w->update () C<ev::stat> only	2202	=item w->update () (C<ev::stat> only)
1820		2203
1821	Invokes C<ev_stat_stat>.	2204	Invokes C<ev_stat_stat>.
1822		2205
1823	=back	2206	=back
1824		2207
…		…
1834		2217
1835	myclass ();	2218	myclass ();
1836	}	2219	}
1837		2220
1838	myclass::myclass (int fd)	2221	myclass::myclass (int fd)
1839	: io (this, &myclass::io_cb),
1840	idle (this, &myclass::idle_cb)
1841	{	2222	{
		2223	io .set <myclass, &myclass::io_cb > (this);
		2224	idle.set <myclass, &myclass::idle_cb> (this);
		2225
1842	io.start (fd, ev::READ);	2226	io.start (fd, ev::READ);
1843	}	2227	}
1844		2228
1845		2229
1846	=head1 MACRO MAGIC	2230	=head1 MACRO MAGIC
1847		2231
1848	Libev can be compiled with a variety of options, the most fundemantal is	2232	Libev can be compiled with a variety of options, the most fundamantal
1849	C<EV_MULTIPLICITY>. This option determines whether (most) functions and	2233	of which is C<EV_MULTIPLICITY>. This option determines whether (most)
1850	callbacks have an initial C<struct ev_loop *> argument.	2234	functions and callbacks have an initial C<struct ev_loop *> argument.
1851		2235
1852	To make it easier to write programs that cope with either variant, the	2236	To make it easier to write programs that cope with either variant, the
1853	following macros are defined:	2237	following macros are defined:
1854		2238
1855	=over 4	2239	=over 4
…		…
1909	Libev can (and often is) directly embedded into host	2293	Libev can (and often is) directly embedded into host
1910	applications. Examples of applications that embed it include the Deliantra	2294	applications. Examples of applications that embed it include the Deliantra
1911	Game Server, the EV perl module, the GNU Virtual Private Ethernet (gvpe)	2295	Game Server, the EV perl module, the GNU Virtual Private Ethernet (gvpe)
1912	and rxvt-unicode.	2296	and rxvt-unicode.
1913		2297
1914	The goal is to enable you to just copy the neecssary files into your	2298	The goal is to enable you to just copy the necessary files into your
1915	source directory without having to change even a single line in them, so	2299	source directory without having to change even a single line in them, so
1916	you can easily upgrade by simply copying (or having a checked-out copy of	2300	you can easily upgrade by simply copying (or having a checked-out copy of
1917	libev somewhere in your source tree).	2301	libev somewhere in your source tree).
1918		2302
1919	=head2 FILESETS	2303	=head2 FILESETS
…		…
2009		2393
2010	If defined to be C<1>, libev will try to detect the availability of the	2394	If defined to be C<1>, libev will try to detect the availability of the
2011	monotonic clock option at both compiletime and runtime. Otherwise no use	2395	monotonic clock option at both compiletime and runtime. Otherwise no use
2012	of the monotonic clock option will be attempted. If you enable this, you	2396	of the monotonic clock option will be attempted. If you enable this, you
2013	usually have to link against librt or something similar. Enabling it when	2397	usually have to link against librt or something similar. Enabling it when
2014	the functionality isn't available is safe, though, althoguh you have	2398	the functionality isn't available is safe, though, although you have
2015	to make sure you link against any libraries where the C<clock_gettime>	2399	to make sure you link against any libraries where the C<clock_gettime>
2016	function is hiding in (often F<-lrt>).	2400	function is hiding in (often F<-lrt>).
2017		2401
2018	=item EV_USE_REALTIME	2402	=item EV_USE_REALTIME
2019		2403
2020	If defined to be C<1>, libev will try to detect the availability of the	2404	If defined to be C<1>, libev will try to detect the availability of the
2021	realtime clock option at compiletime (and assume its availability at	2405	realtime clock option at compiletime (and assume its availability at
2022	runtime if successful). Otherwise no use of the realtime clock option will	2406	runtime if successful). Otherwise no use of the realtime clock option will
2023	be attempted. This effectively replaces C<gettimeofday> by C<clock_get	2407	be attempted. This effectively replaces C<gettimeofday> by C<clock_get
2024	(CLOCK_REALTIME, ...)> and will not normally affect correctness. See tzhe note about libraries	2408	(CLOCK_REALTIME, ...)> and will not normally affect correctness. See the
2025	in the description of C<EV_USE_MONOTONIC>, though.	2409	note about libraries in the description of C<EV_USE_MONOTONIC>, though.
		2410
		2411	=item EV_USE_NANOSLEEP
		2412
		2413	If defined to be C<1>, libev will assume that C<nanosleep ()> is available
		2414	and will use it for delays. Otherwise it will use C<select ()>.
2026		2415
2027	=item EV_USE_SELECT	2416	=item EV_USE_SELECT
2028		2417
2029	If undefined or defined to be C<1>, libev will compile in support for the	2418	If undefined or defined to be C<1>, libev will compile in support for the
2030	C<select>(2) backend. No attempt at autodetection will be done: if no	2419	C<select>(2) backend. No attempt at autodetection will be done: if no
…		…
2123	will have the C<struct ev_loop *> as first argument, and you can create	2512	will have the C<struct ev_loop *> as first argument, and you can create
2124	additional independent event loops. Otherwise there will be no support	2513	additional independent event loops. Otherwise there will be no support
2125	for multiple event loops and there is no first event loop pointer	2514	for multiple event loops and there is no first event loop pointer
2126	argument. Instead, all functions act on the single default loop.	2515	argument. Instead, all functions act on the single default loop.
2127		2516
		2517	=item EV_MINPRI
		2518
		2519	=item EV_MAXPRI
		2520
		2521	The range of allowed priorities. C<EV_MINPRI> must be smaller or equal to
		2522	C<EV_MAXPRI>, but otherwise there are no non-obvious limitations. You can
		2523	provide for more priorities by overriding those symbols (usually defined
		2524	to be C<-2> and C<2>, respectively).
		2525
		2526	When doing priority-based operations, libev usually has to linearly search
		2527	all the priorities, so having many of them (hundreds) uses a lot of space
		2528	and time, so using the defaults of five priorities (-2 .. +2) is usually
		2529	fine.
		2530
		2531	If your embedding app does not need any priorities, defining these both to
		2532	C<0> will save some memory and cpu.
		2533
2128	=item EV_PERIODIC_ENABLE	2534	=item EV_PERIODIC_ENABLE
2129		2535
2130	If undefined or defined to be C<1>, then periodic timers are supported. If	2536	If undefined or defined to be C<1>, then periodic timers are supported. If
2131	defined to be C<0>, then they are not. Disabling them saves a few kB of	2537	defined to be C<0>, then they are not. Disabling them saves a few kB of
2132	code.	2538	code.
…		…
2165	than enough. If you need to manage thousands of children you might want to	2571	than enough. If you need to manage thousands of children you might want to
2166	increase this value (I<must> be a power of two).	2572	increase this value (I<must> be a power of two).
2167		2573
2168	=item EV_INOTIFY_HASHSIZE	2574	=item EV_INOTIFY_HASHSIZE
2169		2575
2170	C<ev_staz> watchers use a small hash table to distribute workload by	2576	C<ev_stat> watchers use a small hash table to distribute workload by
2171	inotify watch id. The default size is C<16> (or C<1> with C<EV_MINIMAL>),	2577	inotify watch id. The default size is C<16> (or C<1> with C<EV_MINIMAL>),
2172	usually more than enough. If you need to manage thousands of C<ev_stat>	2578	usually more than enough. If you need to manage thousands of C<ev_stat>
2173	watchers you might want to increase this value (I<must> be a power of	2579	watchers you might want to increase this value (I<must> be a power of
2174	two).	2580	two).
2175		2581
…		…
2192		2598
2193	=item ev_set_cb (ev, cb)	2599	=item ev_set_cb (ev, cb)
2194		2600
2195	Can be used to change the callback member declaration in each watcher,	2601	Can be used to change the callback member declaration in each watcher,
2196	and the way callbacks are invoked and set. Must expand to a struct member	2602	and the way callbacks are invoked and set. Must expand to a struct member
2197	definition and a statement, respectively. See the F<ev.v> header file for	2603	definition and a statement, respectively. See the F<ev.h> header file for
2198	their default definitions. One possible use for overriding these is to	2604	their default definitions. One possible use for overriding these is to
2199	avoid the C<struct ev_loop *> as first argument in all cases, or to use	2605	avoid the C<struct ev_loop *> as first argument in all cases, or to use
2200	method calls instead of plain function calls in C++.	2606	method calls instead of plain function calls in C++.
		2607
		2608	=head2 EXPORTED API SYMBOLS
		2609
		2610	If you need to re-export the API (e.g. via a dll) and you need a list of
		2611	exported symbols, you can use the provided F<Symbol.*> files which list
		2612	all public symbols, one per line:
		2613
		2614	Symbols.ev for libev proper
		2615	Symbols.event for the libevent emulation
		2616
		2617	This can also be used to rename all public symbols to avoid clashes with
		2618	multiple versions of libev linked together (which is obviously bad in
		2619	itself, but sometimes it is inconvinient to avoid this).
		2620
		2621	A sed command like this will create wrapper C<#define>'s that you need to
		2622	include before including F<ev.h>:
		2623
		2624	<Symbols.ev sed -e "s/.*/#define & myprefix_&/" >wrap.h
		2625
		2626	This would create a file F<wrap.h> which essentially looks like this:
		2627
		2628	#define ev_backend myprefix_ev_backend
		2629	#define ev_check_start myprefix_ev_check_start
		2630	#define ev_check_stop myprefix_ev_check_stop
		2631	...
2201		2632
2202	=head2 EXAMPLES	2633	=head2 EXAMPLES
2203		2634
2204	For a real-world example of a program the includes libev	2635	For a real-world example of a program the includes libev
2205	verbatim, you can have a look at the EV perl module	2636	verbatim, you can have a look at the EV perl module
…		…
2234		2665
2235	In this section the complexities of (many of) the algorithms used inside	2666	In this section the complexities of (many of) the algorithms used inside
2236	libev will be explained. For complexity discussions about backends see the	2667	libev will be explained. For complexity discussions about backends see the
2237	documentation for C<ev_default_init>.	2668	documentation for C<ev_default_init>.
2238		2669
		2670	All of the following are about amortised time: If an array needs to be
		2671	extended, libev needs to realloc and move the whole array, but this
		2672	happens asymptotically never with higher number of elements, so O(1) might
		2673	mean it might do a lengthy realloc operation in rare cases, but on average
		2674	it is much faster and asymptotically approaches constant time.
		2675
2239	=over 4	2676	=over 4
2240		2677
2241	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)	2678	=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)
2242		2679
		2680	This means that, when you have a watcher that triggers in one hour and
		2681	there are 100 watchers that would trigger before that then inserting will
		2682	have to skip roughly seven (C<ld 100>) of these watchers.
		2683
2243	=item Changing timer/periodic watchers (by autorepeat, again): O(log skipped_other_timers)	2684	=item Changing timer/periodic watchers (by autorepeat or calling again): O(log skipped_other_timers)
		2685
		2686	That means that changing a timer costs less than removing/adding them
		2687	as only the relative motion in the event queue has to be paid for.
2244		2688
2245	=item Starting io/check/prepare/idle/signal/child watchers: O(1)	2689	=item Starting io/check/prepare/idle/signal/child watchers: O(1)
2246		2690
		2691	These just add the watcher into an array or at the head of a list.
		2692
2247	=item Stopping check/prepare/idle watchers: O(1)	2693	=item Stopping check/prepare/idle watchers: O(1)
2248		2694
2249	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))	2695	=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))
2250		2696
		2697	These watchers are stored in lists then need to be walked to find the
		2698	correct watcher to remove. The lists are usually short (you don't usually
		2699	have many watchers waiting for the same fd or signal).
		2700
2251	=item Finding the next timer per loop iteration: O(1)	2701	=item Finding the next timer in each loop iteration: O(1)
		2702
		2703	By virtue of using a binary heap, the next timer is always found at the
		2704	beginning of the storage array.
2252		2705
2253	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)	2706	=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)
2254		2707
2255	=item Activating one watcher: O(1)	2708	A change means an I/O watcher gets started or stopped, which requires
		2709	libev to recalculate its status (and possibly tell the kernel, depending
		2710	on backend and wether C<ev_io_set> was used).
		2711
		2712	=item Activating one watcher (putting it into the pending state): O(1)
		2713
		2714	=item Priority handling: O(number_of_priorities)
		2715
		2716	Priorities are implemented by allocating some space for each
		2717	priority. When doing priority-based operations, libev usually has to
		2718	linearly search all the priorities, but starting/stopping and activating
		2719	watchers becomes O(1) w.r.t. prioritiy handling.
2256		2720
2257	=back	2721	=back
2258		2722
2259		2723
2260	=head1 AUTHOR	2724	=head1 AUTHOR

Diff Legend

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

Comparing libev/ev.pod (file contents): Revision 1.68 by root, Fri Dec 7 18:09:43 2007 UTC vs. Revision 1.107 by root, Mon Dec 24 04:34:00 2007 UTC

Diff Legend

Comparing libev/ev.pod (file contents):
Revision 1.68 by root, Fri Dec 7 18:09:43 2007 UTC vs.
Revision 1.107 by root, Mon Dec 24 04:34:00 2007 UTC