[ViewVC] Diff of: cvs/EV/EV.pm

Comparing EV/EV.pm (file contents):
Revision 1.32 by root, Thu Nov 8 17:02:10 2007 UTC vs.
Revision 1.128 by root, Thu Oct 21 02:46:59 2010 UTC

…		…
2		2
3	EV - perl interface to libev, a high performance full-featured event loop	3	EV - perl interface to libev, a high performance full-featured event loop
4		4
5	=head1 SYNOPSIS	5	=head1 SYNOPSIS
6		6
7	use EV;	7	use EV;
		8
		9	# TIMERS
		10
		11	my $w = EV::timer 2, 0, sub {
		12	warn "is called after 2s";
		13	};
		14
		15	my $w = EV::timer 2, 2, sub {
		16	warn "is called roughly every 2s (repeat = 2)";
		17	};
		18
		19	undef $w; # destroy event watcher again
		20
		21	my $w = EV::periodic 0, 60, 0, sub {
		22	warn "is called every minute, on the minute, exactly";
		23	};
		24
		25	# IO
		26
		27	my $w = EV::io *STDIN, EV::READ, sub {
		28	my ($w, $revents) = @_; # all callbacks receive the watcher and event mask
		29	warn "stdin is readable, you entered: ", <STDIN>;
		30	};
		31
		32	# SIGNALS
		33
		34	my $w = EV::signal 'QUIT', sub {
		35	warn "sigquit received\n";
		36	};
		37
		38	# CHILD/PID STATUS CHANGES
8		39
9	# TIMERS	40	my $w = EV::child 666, 0, sub {
		41	my ($w, $revents) = @_;
		42	my $status = $w->rstatus;
		43	};
10		44
11	my $w = EV::timer 2, 0, sub {	45	# STAT CHANGES
12	warn "is called after 2s";	46	my $w = EV::stat "/etc/passwd", 10, sub {
13	};
14
15	my $w = EV::timer 2, 1, sub {
16	warn "is called roughly every 2s (repeat = 1)";
17	};
18
19	undef $w; # destroy event watcher again
20
21	my $w = EV::periodic 0, 60, 0, sub {
22	warn "is called every minute, on the minute, exactly";
23	};
24
25	# IO
26
27	my $w = EV::io *STDIN, EV::READ, sub {
28	my ($w, $revents) = @_; # all callbacks get the watcher object and event mask
29	warn "stdin is readable, you entered: ", <STDIN>;
30	};
31
32	# SIGNALS
33
34	my $w = EV::signal 'QUIT', sub {
35	warn "sigquit received\n";
36	};
37
38	my $w = EV::signal 3, sub {
39	warn "sigquit received (this is GNU/Linux, right?)\n";
40	};
41
42	# CHILD/PID STATUS CHANGES
43
44	my $w = EV::child 666, sub {
45	my ($w, $revents) = @_;	47	my ($w, $revents) = @_;
46	# my $pid = $w->rpid;	48	warn $w->path, " has changed somehow.\n";
47	my $status = $w->rstatus;
48	};	49	};
49		50
50	# MAINLOOP	51	# MAINLOOP
51	EV::loop; # loop until EV::loop_done is called	52	EV::loop; # loop until EV::unloop is called or all watchers stop
52	EV::loop EV::LOOP_ONESHOT; # block until at least one event could be handled	53	EV::loop EV::LOOP_ONESHOT; # block until at least one event could be handled
53	EV::loop EV::LOOP_NONBLOCK; # try to handle same events, but do not block	54	EV::loop EV::LOOP_NONBLOCK; # try to handle same events, but do not block
		55
		56	=head1 BEFORE YOU START USING THIS MODULE
		57
		58	If you only need timer, I/O, signal, child and idle watchers and not the
		59	advanced functionality of this module, consider using L<AnyEvent> instead,
		60	specifically the simplified API described in L<AE>.
		61
		62	When used with EV as backend, the L<AE> API is as fast as the native L<EV>
		63	API, but your programs/modules will still run with many other event loops.
54		64
55	=head1 DESCRIPTION	65	=head1 DESCRIPTION
56		66
57	This module provides an interface to libev	67	This module provides an interface to libev
58	(L<http://software.schmorp.de/pkg/libev.html>).	68	(L<http://software.schmorp.de/pkg/libev.html>). While the documentation
		69	below is comprehensive, one might also consult the documentation of
		70	libev itself (L<http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod> or
		71	F<perldoc EV::libev>) for more subtle details on watcher semantics or some
		72	discussion on the available backends, or how to force a specific backend
		73	with C<LIBEV_FLAGS>, or just about in any case because it has much more
		74	detailed information.
		75
		76	This module is very fast and scalable. It is actually so fast that you
		77	can use it through the L<AnyEvent> module, stay portable to other event
		78	loops (if you don't rely on any watcher types not available through it)
		79	and still be faster than with any other event loop currently supported in
		80	Perl.
		81
		82	=head2 MODULE EXPORTS
		83
		84	This module does not export any symbols.
59		85
60	=cut	86	=cut
61		87
62	package EV;	88	package EV;
63		89
64	use strict;	90	use common::sense;
65		91
66	BEGIN {	92	BEGIN {
67	our $VERSION = '0.6';	93	our $VERSION = '4.00';
68	use XSLoader;	94	use XSLoader;
69	XSLoader::load "EV", $VERSION;	95	XSLoader::load "EV", $VERSION;
70	}	96	}
71		97
72	@EV::Io::ISA =	98	@EV::IO::ISA =
73	@EV::Timer::ISA =	99	@EV::Timer::ISA =
74	@EV::Periodic::ISA =	100	@EV::Periodic::ISA =
75	@EV::Signal::ISA =	101	@EV::Signal::ISA =
		102	@EV::Child::ISA =
		103	@EV::Stat::ISA =
76	@EV::Idle::ISA =	104	@EV::Idle::ISA =
77	@EV::Prepare::ISA =	105	@EV::Prepare::ISA =
78	@EV::Check::ISA =	106	@EV::Check::ISA =
79	@EV::Child::ISA = "EV::Watcher";	107	@EV::Embed::ISA =
		108	@EV::Fork::ISA =
		109	@EV::Async::ISA =
		110	"EV::Watcher";
		111
		112	@EV::Loop::Default::ISA = "EV::Loop";
		113
		114	=head1 EVENT LOOPS
		115
		116	EV supports multiple event loops: There is a single "default event loop"
		117	that can handle everything including signals and child watchers, and any
		118	number of "dynamic event loops" that can use different backends (with
		119	various limitations), but no child and signal watchers.
		120
		121	You do not have to do anything to create the default event loop: When
		122	the module is loaded a suitable backend is selected on the premise of
		123	selecting a working backend (which for example rules out kqueue on most
		124	BSDs). Modules should, unless they have "special needs" always use the
		125	default loop as this is fastest (perl-wise), best supported by other
		126	modules (e.g. AnyEvent or Coro) and most portable event loop.
		127
		128	For specific programs you can create additional event loops dynamically.
		129
		130	If you want to take advantage of kqueue (which often works properly for
		131	sockets only) even though the default loop doesn't enable it, you can
		132	I<embed> a kqueue loop into the default loop: running the default loop
		133	will then also service the kqueue loop to some extent. See the example in
		134	the section about embed watchers for an example on how to achieve that.
		135
		136	=over 4
		137
		138	=item $loop = new EV::Loop [$flags]
		139
		140	Create a new event loop as per the specified flags. Please refer to
		141	the C<ev_loop_new ()> function description in the libev documentation
		142	(L<http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod#GLOBAL_FUNCTIONS>,
		143	or locally-installed as F<EV::libev> manpage) for more info.
		144
		145	The loop will automatically be destroyed when it is no longer referenced
		146	by any watcher and the loop object goes out of scope.
		147
		148	If you are not embedding the loop, then Using C<EV::FLAG_FORKCHECK>
		149	is recommended, as only the default event loop is protected by this
		150	module. If you I<are> embedding this loop in the default loop, this is not
		151	necessary, as C<EV::embed> automatically does the right thing on fork.
		152
		153	=item $loop->loop_fork
		154
		155	Must be called after a fork in the child, before entering or continuing
		156	the event loop. An alternative is to use C<EV::FLAG_FORKCHECK> which calls
		157	this function automatically, at some performance loss (refer to the libev
		158	documentation).
		159
		160	=item $loop->loop_verify
		161
		162	Calls C<ev_verify> to make internal consistency checks (for debugging
		163	libev) and abort the program if any data structures were found to be
		164	corrupted.
		165
		166	=item $loop = EV::default_loop [$flags]
		167
		168	Return the default loop (which is a singleton object). Since this module
		169	already creates the default loop with default flags, specifying flags here
		170	will not have any effect unless you destroy the default loop first, which
		171	isn't supported. So in short: don't do it, and if you break it, you get to
		172	keep the pieces.
		173
		174	=back
		175
80		176
81	=head1 BASIC INTERFACE	177	=head1 BASIC INTERFACE
82		178
83	=over 4	179	=over 4
84		180
85	=item $EV::DIED	181	=item $EV::DIED
86		182
87	Must contain a reference to a function that is called when a callback	183	Must contain a reference to a function that is called when a callback
88	throws an exception (with $@ containing thr error). The default prints an	184	throws an exception (with $@ containing the error). The default prints an
89	informative message and continues.	185	informative message and continues.
90		186
91	If this callback throws an exception it will be silently ignored.	187	If this callback throws an exception it will be silently ignored.
92		188
		189	=item $flags = EV::supported_backends
		190
		191	=item $flags = EV::recommended_backends
		192
		193	=item $flags = EV::embeddable_backends
		194
		195	Returns the set (see C<EV::BACKEND_*> flags) of backends supported by this
		196	instance of EV, the set of recommended backends (supposed to be good) for
		197	this platform and the set of embeddable backends (see EMBED WATCHERS).
		198
		199	=item EV::sleep $seconds
		200
		201	Block the process for the given number of (fractional) seconds.
		202
93	=item $time = EV::time	203	=item $time = EV::time
94		204
95	Returns the current time in (fractional) seconds since the epoch.	205	Returns the current time in (fractional) seconds since the epoch.
96		206
97	=item $time = EV::now	207	=item $time = EV::now
98		208
		209	=item $time = $loop->now
		210
99	Returns the time the last event loop iteration has been started. This	211	Returns the time the last event loop iteration has been started. This
100	is the time that (relative) timers are based on, and refering to it is	212	is the time that (relative) timers are based on, and referring to it is
101	usually faster then calling EV::time.	213	usually faster then calling EV::time.
102		214
103	=item $method = EV::ev_method	215	=item EV::now_update
104		216
		217	=item $loop->now_update
		218
		219	Establishes the current time by querying the kernel, updating the time
		220	returned by C<EV::now> in the progress. This is a costly operation and
		221	is usually done automatically within C<EV::loop>.
		222
		223	This function is rarely useful, but when some event callback runs for a
		224	very long time without entering the event loop, updating libev's idea of
		225	the current time is a good idea.
		226
		227	=item EV::suspend
		228
		229	=item $loop->suspend
		230
		231	=item EV::resume
		232
		233	=item $loop->resume
		234
		235	These two functions suspend and resume a loop, for use when the loop is
		236	not used for a while and timeouts should not be processed.
		237
		238	A typical use case would be an interactive program such as a game: When
		239	the user presses C<^Z> to suspend the game and resumes it an hour later it
		240	would be best to handle timeouts as if no time had actually passed while
		241	the program was suspended. This can be achieved by calling C<suspend>
		242	in your C<SIGTSTP> handler, sending yourself a C<SIGSTOP> and calling
		243	C<resume> directly afterwards to resume timer processing.
		244
		245	Effectively, all C<timer> watchers will be delayed by the time spend
		246	between C<suspend> and C<resume>, and all C<periodic> watchers
		247	will be rescheduled (that is, they will lose any events that would have
		248	occured while suspended).
		249
		250	After calling C<suspend> you B<must not> call I<any> function on the given
		251	loop other than C<resume>, and you B<must not> call C<resume>
		252	without a previous call to C<suspend>.
		253
		254	Calling C<suspend>/C<resume> has the side effect of updating the event
		255	loop time (see C<now_update>).
		256
		257	=item $backend = EV::backend
		258
		259	=item $backend = $loop->backend
		260
105	Returns an integer describing the backend used by libev (EV::METHOD_SELECT	261	Returns an integer describing the backend used by libev (EV::BACKEND_SELECT
106	or EV::METHOD_EPOLL).	262	or EV::BACKEND_EPOLL).
107		263
108	=item EV::loop [$flags]	264	=item EV::loop [$flags]
109		265
		266	=item $loop->loop ([$flags])
		267
110	Begin checking for events and calling callbacks. It returns when a	268	Begin checking for events and calling callbacks. It returns when a
111	callback calls EV::loop_done.	269	callback calls EV::unloop.
112		270
113	The $flags argument can be one of the following:	271	The $flags argument can be one of the following:
114		272
115	0 as above	273	0 as above
116	EV::LOOP_ONESHOT block at most once (wait, but do not loop)	274	EV::LOOP_ONESHOT block at most once (wait, but do not loop)
117	EV::LOOP_NONBLOCK do not block at all (fetch/handle events but do not wait)	275	EV::LOOP_NONBLOCK do not block at all (fetch/handle events but do not wait)
118		276
119	=item EV::loop_done [$how]	277	=item EV::unloop [$how]
120		278
		279	=item $loop->unloop ([$how])
		280
121	When called with no arguments or an argument of 1, makes the innermost	281	When called with no arguments or an argument of EV::UNLOOP_ONE, makes the
122	call to EV::loop return.	282	innermost call to EV::loop return.
123		283
124	When called with an agrument of 2, all calls to EV::loop will return as	284	When called with an argument of EV::UNLOOP_ALL, all calls to EV::loop will return as
125	fast as possible.	285	fast as possible.
126		286
127	=back	287	=item $count = EV::loop_count
128		288
129	=head2 WATCHER	289	=item $count = $loop->loop_count
		290
		291	Return the number of times the event loop has polled for new
		292	events. Sometimes useful as a generation counter.
		293
		294	=item EV::once $fh_or_undef, $events, $timeout, $cb->($revents)
		295
		296	=item $loop->once ($fh_or_undef, $events, $timeout, $cb->($revents))
		297
		298	This function rolls together an I/O and a timer watcher for a single
		299	one-shot event without the need for managing a watcher object.
		300
		301	If C<$fh_or_undef> is a filehandle or file descriptor, then C<$events>
		302	must be a bitset containing either C<EV::READ>, C<EV::WRITE> or C<EV::READ
		303	\| EV::WRITE>, indicating the type of I/O event you want to wait for. If
		304	you do not want to wait for some I/O event, specify C<undef> for
		305	C<$fh_or_undef> and C<0> for C<$events>).
		306
		307	If timeout is C<undef> or negative, then there will be no
		308	timeout. Otherwise a EV::timer with this value will be started.
		309
		310	When an error occurs or either the timeout or I/O watcher triggers, then
		311	the callback will be called with the received event set (in general
		312	you can expect it to be a combination of C<EV::ERROR>, C<EV::READ>,
		313	C<EV::WRITE> and C<EV::TIMER>).
		314
		315	EV::once doesn't return anything: the watchers stay active till either
		316	of them triggers, then they will be stopped and freed, and the callback
		317	invoked.
		318
		319	=item EV::feed_fd_event ($fd, $revents)
		320
		321	=item $loop->feed_fd_event ($fd, $revents)
		322
		323	Feed an event on a file descriptor into EV. EV will react to this call as
		324	if the readyness notifications specified by C<$revents> (a combination of
		325	C<EV::READ> and C<EV::WRITE>) happened on the file descriptor C<$fd>.
		326
		327	=item EV::feed_signal_event ($signal)
		328
		329	Feed a signal event into EV. EV will react to this call as if the signal
		330	specified by C<$signal> had occured.
		331
		332	=item EV::set_io_collect_interval $time
		333
		334	=item $loop->set_io_collect_interval ($time)
		335
		336	=item EV::set_timeout_collect_interval $time
		337
		338	=item $loop->set_timeout_collect_interval ($time)
		339
		340	These advanced functions set the minimum block interval when polling for I/O events and the minimum
		341	wait interval for timer events. See the libev documentation at
		342	L<http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod#FUNCTIONS_CONTROLLING_THE_EVENT_LOOP>
		343	(locally installed as F<EV::libev>) for a more detailed discussion.
		344
		345	=item $count = EV::pending_count
		346
		347	=item $count = $loop->pending_count
		348
		349	Returns the number of currently pending watchers.
		350
		351	=item EV::invoke_pending
		352
		353	=item $loop->invoke_pending
		354
		355	Invoke all currently pending watchers.
		356
		357	=back
		358
		359
		360	=head1 WATCHER OBJECTS
130		361
131	A watcher is an object that gets created to record your interest in some	362	A watcher is an object that gets created to record your interest in some
132	event. For instance, if you want to wait for STDIN to become readable, you	363	event. For instance, if you want to wait for STDIN to become readable, you
133	would create an EV::io watcher for that:	364	would create an EV::io watcher for that:
134		365
135	my $watcher = EV::io *STDIN, EV::READ, sub {	366	my $watcher = EV::io *STDIN, EV::READ, sub {
136	my ($watcher, $revents) = @_;	367	my ($watcher, $revents) = @_;
137	warn "yeah, STDIN should not be readable without blocking!\n"	368	warn "yeah, STDIN should now be readable without blocking!\n"
138	};	369	};
139		370
140	All watchers can be active (waiting for events) or inactive (paused). Only	371	All watchers can be active (waiting for events) or inactive (paused). Only
141	active watchers will have their callbacks invoked. All callbacks will be	372	active watchers will have their callbacks invoked. All callbacks will be
142	called with at least two arguments: the watcher and a bitmask of received	373	called with at least two arguments: the watcher and a bitmask of received
143	events.	374	events.
144		375
145	Each watcher type has its associated bit in revents, so you can use the	376	Each watcher type has its associated bit in revents, so you can use the
146	same callback for multiple watchers. The event mask is named after the	377	same callback for multiple watchers. The event mask is named after the
147	type, i..e. EV::child sets EV::CHILD, EV::prepare sets EV::PREPARE,	378	type, i.e. EV::child sets EV::CHILD, EV::prepare sets EV::PREPARE,
148	EV::periodic sets EV::PERIODIC and so on, with the exception of IO events	379	EV::periodic sets EV::PERIODIC and so on, with the exception of I/O events
149	(which can set both EV::READ and EV::WRITE bits), and EV::timer (which	380	(which can set both EV::READ and EV::WRITE bits).
150	uses EV::TIMEOUT).
151		381
152	In the rare case where one wants to create a watcher but not start it at	382	In the rare case where one wants to create a watcher but not start it at
153	the same time, each constructor has a variant with a trailing C<_ns> in	383	the same time, each constructor has a variant with a trailing C<_ns> in
154	its name, e.g. EV::io has a non-starting variant EV::io_ns and so on.	384	its name, e.g. EV::io has a non-starting variant EV::io_ns and so on.
155		385
…		…
159		389
160	Also, all methods changing some aspect of a watcher (->set, ->priority,	390	Also, all methods changing some aspect of a watcher (->set, ->priority,
161	->fh and so on) automatically stop and start it again if it is active,	391	->fh and so on) automatically stop and start it again if it is active,
162	which means pending events get lost.	392	which means pending events get lost.
163		393
164	=head2 WATCHER TYPES	394	=head2 COMMON WATCHER METHODS
165		395
166	Now lets move to the existing watcher types and asociated methods.	396	This section lists methods common to all watchers.
167
168	The following methods are available for all watchers. Then followes a
169	description of each watcher constructor (EV::io, EV::timer, EV::periodic,
170	EV::signal, EV::child, EV::idle, EV::prepare and EV::check), followed by
171	any type-specific methods (if any).
172		397
173	=over 4	398	=over 4
174		399
175	=item $w->start	400	=item $w->start
176		401
…		…
180		405
181	=item $w->stop	406	=item $w->stop
182		407
183	Stop a watcher if it is active. Also clear any pending events (events that	408	Stop a watcher if it is active. Also clear any pending events (events that
184	have been received but that didn't yet result in a callback invocation),	409	have been received but that didn't yet result in a callback invocation),
185	regardless of wether the watcher was active or not.	410	regardless of whether the watcher was active or not.
186		411
187	=item $bool = $w->is_active	412	=item $bool = $w->is_active
188		413
189	Returns true if the watcher is active, false otherwise.	414	Returns true if the watcher is active, false otherwise.
190		415
…		…
215	watchers with higher priority will be invoked first. The valid range of	440	watchers with higher priority will be invoked first. The valid range of
216	priorities lies between EV::MAXPRI (default 2) and EV::MINPRI (default	441	priorities lies between EV::MAXPRI (default 2) and EV::MINPRI (default
217	-2). If the priority is outside this range it will automatically be	442	-2). If the priority is outside this range it will automatically be
218	normalised to the nearest valid priority.	443	normalised to the nearest valid priority.
219		444
220	The default priority of any newly-created weatcher is 0.	445	The default priority of any newly-created watcher is 0.
221		446
		447	Note that the priority semantics have not yet been fleshed out and are
		448	subject to almost certain change.
		449
222	=item $w->trigger ($revents)	450	=item $w->invoke ($revents)
223		451
224	Call the callback now with the given event mask.	452	Call the callback now with the given event mask.
225		453
		454	=item $w->feed_event ($revents)
		455
		456	Feed some events on this watcher into EV. EV will react to this call as if
		457	the watcher had received the given C<$revents> mask.
		458
		459	=item $revents = $w->clear_pending
		460
		461	If the watcher is pending, this function clears its pending status and
		462	returns its C<$revents> bitset (as if its callback was invoked). If the
		463	watcher isn't pending it does nothing and returns C<0>.
		464
		465	=item $previous_state = $w->keepalive ($bool)
		466
		467	Normally, C<EV::loop> will return when there are no active watchers
		468	(which is a "deadlock" because no progress can be made anymore). This is
		469	convenient because it allows you to start your watchers (and your jobs),
		470	call C<EV::loop> once and when it returns you know that all your jobs are
		471	finished (or they forgot to register some watchers for their task :).
		472
		473	Sometimes, however, this gets in your way, for example when the module
		474	that calls C<EV::loop> (usually the main program) is not the same module
		475	as a long-living watcher (for example a DNS client module written by
		476	somebody else even). Then you might want any outstanding requests to be
		477	handled, but you would not want to keep C<EV::loop> from returning just
		478	because you happen to have this long-running UDP port watcher.
		479
		480	In this case you can clear the keepalive status, which means that even
		481	though your watcher is active, it won't keep C<EV::loop> from returning.
		482
		483	The initial value for keepalive is true (enabled), and you can change it
		484	any time.
		485
		486	Example: Register an I/O watcher for some UDP socket but do not keep the
		487	event loop from running just because of that watcher.
		488
		489	my $udp_socket = ...
		490	my $udp_watcher = EV::io $udp_socket, EV::READ, sub { ... };
		491	$udp_watcher->keepalive (0);
		492
		493	=item $loop = $w->loop
		494
		495	Return the loop that this watcher is attached to.
		496
		497	=back
		498
		499
		500	=head1 WATCHER TYPES
		501
		502	Each of the following subsections describes a single watcher type.
		503
		504	=head3 I/O WATCHERS - is this file descriptor readable or writable?
		505
		506	=over 4
226		507
227	=item $w = EV::io $fileno_or_fh, $eventmask, $callback	508	=item $w = EV::io $fileno_or_fh, $eventmask, $callback
228		509
229	=item $w = EV::io_ns $fileno_or_fh, $eventmask, $callback	510	=item $w = EV::io_ns $fileno_or_fh, $eventmask, $callback
230		511
		512	=item $w = $loop->io ($fileno_or_fh, $eventmask, $callback)
		513
		514	=item $w = $loop->io_ns ($fileno_or_fh, $eventmask, $callback)
		515
231	As long as the returned watcher object is alive, call the C<$callback>	516	As long as the returned watcher object is alive, call the C<$callback>
232	when the events specified in C<$eventmask>.	517	when at least one of events specified in C<$eventmask> occurs.
233		518
234	The $eventmask can be one or more of these constants ORed together:	519	The $eventmask can be one or more of these constants ORed together:
235		520
236	EV::READ wait until read() wouldn't block anymore	521	EV::READ wait until read() wouldn't block anymore
237	EV::WRITE wait until write() wouldn't block anymore	522	EV::WRITE wait until write() wouldn't block anymore
…		…
253		538
254	=item $old_eventmask = $w->events ($new_eventmask)	539	=item $old_eventmask = $w->events ($new_eventmask)
255		540
256	Returns the previously set event mask and optionally set a new one.	541	Returns the previously set event mask and optionally set a new one.
257		542
		543	=back
		544
		545
		546	=head3 TIMER WATCHERS - relative and optionally repeating timeouts
		547
		548	=over 4
258		549
259	=item $w = EV::timer $after, $repeat, $callback	550	=item $w = EV::timer $after, $repeat, $callback
260		551
261	=item $w = EV::timer_ns $after, $repeat, $callback	552	=item $w = EV::timer_ns $after, $repeat, $callback
262		553
263	Calls the callback after C<$after> seconds. If C<$repeat> is non-zero,	554	=item $w = $loop->timer ($after, $repeat, $callback)
264	the timer will be restarted (with the $repeat value as $after) after the	555
265	callback returns.	556	=item $w = $loop->timer_ns ($after, $repeat, $callback)
		557
		558	Calls the callback after C<$after> seconds (which may be fractional). If
		559	C<$repeat> is non-zero, the timer will be restarted (with the $repeat
		560	value as $after) after the callback returns.
266		561
267	This means that the callback would be called roughly after C<$after>	562	This means that the callback would be called roughly after C<$after>
268	seconds, and then every C<$repeat> seconds. "Roughly" because the time of	563	seconds, and then every C<$repeat> seconds. The timer does his best not
269	callback processing is not taken into account, so the timer will slowly	564	to drift, but it will not invoke the timer more often then once per event
270	drift. If that isn't acceptable, look at EV::periodic.	565	loop iteration, and might drift in other cases. If that isn't acceptable,
		566	look at EV::periodic, which can provide long-term stable timers.
271		567
272	The timer is based on a monotonic clock, that is if somebody is sitting	568	The timer is based on a monotonic clock, that is, if somebody is sitting
273	in front of the machine while the timer is running and changes the system	569	in front of the machine while the timer is running and changes the system
274	clock, the timer will nevertheless run (roughly) the same time.	570	clock, the timer will nevertheless run (roughly) the same time.
275		571
276	The C<timer_ns> variant doesn't start (activate) the newly created watcher.	572	The C<timer_ns> variant doesn't start (activate) the newly created watcher.
277		573
278	=item $w->set ($after, $repeat)	574	=item $w->set ($after, $repeat)
279		575
280	Reconfigures the watcher, see the constructor above for details. Can be at	576	Reconfigures the watcher, see the constructor above for details. Can be called at
281	any time.	577	any time.
282		578
283	=item $w->again	579	=item $w->again
284		580
285	Similar to the C<start> method, but has special semantics for repeating timers:	581	Similar to the C<start> method, but has special semantics for repeating timers:
		582
		583	If the timer is active and non-repeating, it will be stopped.
286		584
287	If the timer is active and repeating, reset the timeout to occur	585	If the timer is active and repeating, reset the timeout to occur
288	C<$repeat> seconds after now.	586	C<$repeat> seconds after now.
289		587
290	If the timer is active and non-repeating, it will be stopped.
291
292	If the timer is in active and repeating, start it.	588	If the timer is inactive and repeating, start it using the repeat value.
293		589
294	Otherwise do nothing.	590	Otherwise do nothing.
295		591
296	This behaviour is useful when you have a timeout for some IO	592	This behaviour is useful when you have a timeout for some IO
297	operation. You create a timer object with the same value for C<$after> and	593	operation. You create a timer object with the same value for C<$after> and
298	C<$repeat>, and then, in the read/write watcher, run the C<again> method	594	C<$repeat>, and then, in the read/write watcher, run the C<again> method
299	on the timeout.	595	on the timeout.
300		596
		597	=back
		598
		599
		600	=head3 PERIODIC WATCHERS - to cron or not to cron?
		601
		602	=over 4
301		603
302	=item $w = EV::periodic $at, $interval, $reschedule_cb, $callback	604	=item $w = EV::periodic $at, $interval, $reschedule_cb, $callback
303		605
304	=item $w = EV::periodic_ns $at, $interval, $reschedule_cb, $callback	606	=item $w = EV::periodic_ns $at, $interval, $reschedule_cb, $callback
		607
		608	=item $w = $loop->periodic ($at, $interval, $reschedule_cb, $callback)
		609
		610	=item $w = $loop->periodic_ns ($at, $interval, $reschedule_cb, $callback)
305		611
306	Similar to EV::timer, but is not based on relative timeouts but on	612	Similar to EV::timer, but is not based on relative timeouts but on
307	absolute times. Apart from creating "simple" timers that trigger "at" the	613	absolute times. Apart from creating "simple" timers that trigger "at" the
308	specified time, it can also be used for non-drifting absolute timers and	614	specified time, it can also be used for non-drifting absolute timers and
309	more complex, cron-like, setups that are not adversely affected by time	615	more complex, cron-like, setups that are not adversely affected by time
…		…
319	This time simply fires at the wallclock time C<$at> and doesn't repeat. It	625	This time simply fires at the wallclock time C<$at> and doesn't repeat. It
320	will not adjust when a time jump occurs, that is, if it is to be run	626	will not adjust when a time jump occurs, that is, if it is to be run
321	at January 1st 2011 then it will run when the system time reaches or	627	at January 1st 2011 then it will run when the system time reaches or
322	surpasses this time.	628	surpasses this time.
323		629
324	=item * non-repeating interval timer ($interval > 0, $reschedule_cb = 0)	630	=item * repeating interval timer ($interval > 0, $reschedule_cb = 0)
325		631
326	In this mode the watcher will always be scheduled to time out at the	632	In this mode the watcher will always be scheduled to time out at the
327	next C<$at + N * $interval> time (for some integer N) and then repeat,	633	next C<$at + N * $interval> time (for some integer N) and then repeat,
328	regardless of any time jumps.	634	regardless of any time jumps.
329		635
…		…
341	possible time where C<$time = $at (mod $interval)>, regardless of any time	647	possible time where C<$time = $at (mod $interval)>, regardless of any time
342	jumps.	648	jumps.
343		649
344	=item * manual reschedule mode ($reschedule_cb = coderef)	650	=item * manual reschedule mode ($reschedule_cb = coderef)
345		651
346	In this mode $interval and $at are both being ignored. Instead, each time	652	In this mode $interval and $at are both being ignored. Instead, each
347	the periodic watcher gets scheduled, the first callback ($reschedule_cb)	653	time the periodic watcher gets scheduled, the reschedule callback
348	will be called with the watcher as first, and the current time as second	654	($reschedule_cb) will be called with the watcher as first, and the current
349	argument.	655	time as second argument.
350		656
351	I<This callback MUST NOT stop or destroy this or any other periodic	657	I<This callback MUST NOT stop or destroy this or any other periodic
352	watcher, ever>. If you need to stop it, return 1e30 and stop it	658	watcher, ever, and MUST NOT call any event loop functions or methods>. If
353	afterwards.	659	you need to stop it, return 1e30 and stop it afterwards. You may create
		660	and start a C<EV::prepare> watcher for this task.
354		661
355	It must return the next time to trigger, based on the passed time value	662	It must return the next time to trigger, based on the passed time value
356	(that is, the lowest time value larger than to the second argument). It	663	(that is, the lowest time value larger than or equal to to the second
357	will usually be called just before the callback will be triggered, but	664	argument). It will usually be called just before the callback will be
358	might be called at other times, too.	665	triggered, but might be called at other times, too.
359		666
360	This can be used to create very complex timers, such as a timer that	667	This can be used to create very complex timers, such as a timer that
361	triggers on each midnight, local time (actually 24 hours after the last	668	triggers on each midnight, local time (actually 24 hours after the last
362	midnight, to keep the example simple. If you know a way to do it correctly	669	midnight, to keep the example simple. If you know a way to do it correctly
363	in about the same space (without requiring elaborate modules), drop me a	670	in about the same space (without requiring elaborate modules), drop me a
…		…
377		684
378	The C<periodic_ns> variant doesn't start (activate) the newly created watcher.	685	The C<periodic_ns> variant doesn't start (activate) the newly created watcher.
379		686
380	=item $w->set ($at, $interval, $reschedule_cb)	687	=item $w->set ($at, $interval, $reschedule_cb)
381		688
382	Reconfigures the watcher, see the constructor above for details. Can be at	689	Reconfigures the watcher, see the constructor above for details. Can be called at
383	any time.	690	any time.
384		691
385	=item $w->again	692	=item $w->again
386		693
387	Simply stops and starts the watcher again.	694	Simply stops and starts the watcher again.
388		695
		696	=item $time = $w->at
		697
		698	Return the time that the watcher is expected to trigger next.
		699
		700	=back
		701
		702
		703	=head3 SIGNAL WATCHERS - signal me when a signal gets signalled!
		704
		705	=over 4
389		706
390	=item $w = EV::signal $signal, $callback	707	=item $w = EV::signal $signal, $callback
391		708
392	=item $w = EV::signal_ns $signal, $callback	709	=item $w = EV::signal_ns $signal, $callback
393		710
		711	=item $w = $loop->signal ($signal, $callback)
		712
		713	=item $w = $loop->signal_ns ($signal, $callback)
		714
394	Call the callback when $signal is received (the signal can be specified	715	Call the callback when $signal is received (the signal can be specified by
395	by number or by name, just as with kill or %SIG).	716	number or by name, just as with C<kill> or C<%SIG>).
		717
		718	Only one event loop can grab a given signal - attempting to grab the same
		719	signal from two EV loops will crash the program immediately or cause data
		720	corruption.
396		721
397	EV will grab the signal for the process (the kernel only allows one	722	EV will grab the signal for the process (the kernel only allows one
398	component to receive a signal at a time) when you start a signal watcher,	723	component to receive a signal at a time) when you start a signal watcher,
399	and removes it again when you stop it. Perl does the same when you	724	and removes it again when you stop it. Perl does the same when you
400	add/remove callbacks to %SIG, so watch out.	725	add/remove callbacks to C<%SIG>, so watch out.
401		726
402	You can have as many signal watchers per signal as you want.	727	You can have as many signal watchers per signal as you want.
403		728
404	The C<signal_ns> variant doesn't start (activate) the newly created watcher.	729	The C<signal_ns> variant doesn't start (activate) the newly created watcher.
405		730
406	=item $w->set ($signal)	731	=item $w->set ($signal)
407		732
408	Reconfigures the watcher, see the constructor above for details. Can be at	733	Reconfigures the watcher, see the constructor above for details. Can be
409	any time.	734	called at any time.
410		735
411	=item $current_signum = $w->signal	736	=item $current_signum = $w->signal
412		737
413	=item $old_signum = $w->signal ($new_signal)	738	=item $old_signum = $w->signal ($new_signal)
414		739
415	Returns the previously set signal (always as a number not name) and	740	Returns the previously set signal (always as a number not name) and
416	optionally set a new one.	741	optionally set a new one.
417		742
		743	=back
418		744
		745
		746	=head3 CHILD WATCHERS - watch out for process status changes
		747
		748	=over 4
		749
419	=item $w = EV::child $pid, $callback	750	=item $w = EV::child $pid, $trace, $callback
420		751
421	=item $w = EV::child_ns $pid, $callback	752	=item $w = EV::child_ns $pid, $trace, $callback
		753
		754	=item $w = $loop->child ($pid, $trace, $callback)
		755
		756	=item $w = $loop->child_ns ($pid, $trace, $callback)
422		757
423	Call the callback when a status change for pid C<$pid> (or any pid	758	Call the callback when a status change for pid C<$pid> (or any pid
424	if C<$pid> is 0) has been received. More precisely: when the process	759	if C<$pid> is 0) has been received (a status change happens when the
		760	process terminates or is killed, or, when trace is true, additionally when
		761	it is stopped or continued). More precisely: when the process receives
425	receives a SIGCHLD, EV will fetch the outstanding exit/wait status for all	762	a C<SIGCHLD>, EV will fetch the outstanding exit/wait status for all
426	changed/zombie children and call the callback.	763	changed/zombie children and call the callback.
427		764
428	You can access both status and pid by using the C<rstatus> and C<rpid>	765	It is valid (and fully supported) to install a child watcher after a child
429	methods on the watcher object.	766	has exited but before the event loop has started its next iteration (for
		767	example, first you C<fork>, then the new child process might exit, and
		768	only then do you install a child watcher in the parent for the new pid).
430		769
		770	You can access both exit (or tracing) status and pid by using the
		771	C<rstatus> and C<rpid> methods on the watcher object.
		772
431	You can have as many pid watchers per pid as you want.	773	You can have as many pid watchers per pid as you want, they will all be
		774	called.
432		775
433	The C<child_ns> variant doesn't start (activate) the newly created watcher.	776	The C<child_ns> variant doesn't start (activate) the newly created watcher.
434		777
435	=item $w->set ($pid)	778	=item $w->set ($pid, $trace)
436		779
437	Reconfigures the watcher, see the constructor above for details. Can be at	780	Reconfigures the watcher, see the constructor above for details. Can be called at
438	any time.	781	any time.
439		782
440	=item $current_pid = $w->pid	783	=item $current_pid = $w->pid
441
442	=item $old_pid = $w->pid ($new_pid)
443		784
444	Returns the previously set process id and optionally set a new one.	785	Returns the previously set process id and optionally set a new one.
445		786
446	=item $exit_status = $w->rstatus	787	=item $exit_status = $w->rstatus
447		788
…		…
451	=item $pid = $w->rpid	792	=item $pid = $w->rpid
452		793
453	Return the pid of the awaited child (useful when you have installed a	794	Return the pid of the awaited child (useful when you have installed a
454	watcher for all pids).	795	watcher for all pids).
455		796
		797	=back
		798
		799
		800	=head3 STAT WATCHERS - did the file attributes just change?
		801
		802	=over 4
		803
		804	=item $w = EV::stat $path, $interval, $callback
		805
		806	=item $w = EV::stat_ns $path, $interval, $callback
		807
		808	=item $w = $loop->stat ($path, $interval, $callback)
		809
		810	=item $w = $loop->stat_ns ($path, $interval, $callback)
		811
		812	Call the callback when a file status change has been detected on
		813	C<$path>. The C<$path> does not need to exist, changing from "path exists"
		814	to "path does not exist" is a status change like any other.
		815
		816	The C<$interval> is a recommended polling interval for systems where
		817	OS-supported change notifications don't exist or are not supported. If
		818	you use C<0> then an unspecified default is used (which is highly
		819	recommended!), which is to be expected to be around five seconds usually.
		820
		821	This watcher type is not meant for massive numbers of stat watchers,
		822	as even with OS-supported change notifications, this can be
		823	resource-intensive.
		824
		825	The C<stat_ns> variant doesn't start (activate) the newly created watcher.
		826
		827	=item ... = $w->stat
		828
		829	This call is very similar to the perl C<stat> built-in: It stats (using
		830	C<lstat>) the path specified in the watcher and sets perls stat cache (as
		831	well as EV's idea of the current stat values) to the values found.
		832
		833	In scalar context, a boolean is return indicating success or failure of
		834	the stat. In list context, the same 13-value list as with stat is returned
		835	(except that the blksize and blocks fields are not reliable).
		836
		837	In the case of an error, errno is set to C<ENOENT> (regardless of the
		838	actual error value) and the C<nlink> value is forced to zero (if the stat
		839	was successful then nlink is guaranteed to be non-zero).
		840
		841	See also the next two entries for more info.
		842
		843	=item ... = $w->attr
		844
		845	Just like C<< $w->stat >>, but without the initial stat'ing: this returns
		846	the values most recently detected by EV. See the next entry for more info.
		847
		848	=item ... = $w->prev
		849
		850	Just like C<< $w->stat >>, but without the initial stat'ing: this returns
		851	the previous set of values, before the change.
		852
		853	That is, when the watcher callback is invoked, C<< $w->prev >> will be set
		854	to the values found I<before> a change was detected, while C<< $w->attr >>
		855	returns the values found leading to the change detection. The difference (if any)
		856	between C<prev> and C<attr> is what triggered the callback.
		857
		858	If you did something to the filesystem object and do not want to trigger
		859	yet another change, you can call C<stat> to update EV's idea of what the
		860	current attributes are.
		861
		862	=item $w->set ($path, $interval)
		863
		864	Reconfigures the watcher, see the constructor above for details. Can be
		865	called at any time.
		866
		867	=item $current_path = $w->path
		868
		869	=item $old_path = $w->path ($new_path)
		870
		871	Returns the previously set path and optionally set a new one.
		872
		873	=item $current_interval = $w->interval
		874
		875	=item $old_interval = $w->interval ($new_interval)
		876
		877	Returns the previously set interval and optionally set a new one. Can be
		878	used to query the actual interval used.
		879
		880	=back
		881
		882
		883	=head3 IDLE WATCHERS - when you've got nothing better to do...
		884
		885	=over 4
456		886
457	=item $w = EV::idle $callback	887	=item $w = EV::idle $callback
458		888
459	=item $w = EV::idle_ns $callback	889	=item $w = EV::idle_ns $callback
460		890
461	Call the callback when there are no pending io, timer/periodic, signal or	891	=item $w = $loop->idle ($callback)
462	child events, i.e. when the process is idle.	892
		893	=item $w = $loop->idle_ns ($callback)
		894
		895	Call the callback when there are no other pending watchers of the same or
		896	higher priority (excluding check, prepare and other idle watchers of the
		897	same or lower priority, of course). They are called idle watchers because
		898	when the watcher is the highest priority pending event in the process, the
		899	process is considered to be idle at that priority.
		900
		901	If you want a watcher that is only ever called when I<no> other events are
		902	outstanding you have to set the priority to C<EV::MINPRI>.
463		903
464	The process will not block as long as any idle watchers are active, and	904	The process will not block as long as any idle watchers are active, and
465	they will be called repeatedly until stopped.	905	they will be called repeatedly until stopped.
466		906
		907	For example, if you have idle watchers at priority C<0> and C<1>, and
		908	an I/O watcher at priority C<0>, then the idle watcher at priority C<1>
		909	and the I/O watcher will always run when ready. Only when the idle watcher
		910	at priority C<1> is stopped and the I/O watcher at priority C<0> is not
		911	pending with the C<0>-priority idle watcher be invoked.
		912
467	The C<idle_ns> variant doesn't start (activate) the newly created watcher.	913	The C<idle_ns> variant doesn't start (activate) the newly created watcher.
468		914
		915	=back
		916
		917
		918	=head3 PREPARE WATCHERS - customise your event loop!
		919
		920	=over 4
469		921
470	=item $w = EV::prepare $callback	922	=item $w = EV::prepare $callback
471		923
472	=item $w = EV::prepare_ns $callback	924	=item $w = EV::prepare_ns $callback
		925
		926	=item $w = $loop->prepare ($callback)
		927
		928	=item $w = $loop->prepare_ns ($callback)
473		929
474	Call the callback just before the process would block. You can still	930	Call the callback just before the process would block. You can still
475	create/modify any watchers at this point.	931	create/modify any watchers at this point.
476		932
477	See the EV::check watcher, below, for explanations and an example.	933	See the EV::check watcher, below, for explanations and an example.
478		934
479	The C<prepare_ns> variant doesn't start (activate) the newly created watcher.	935	The C<prepare_ns> variant doesn't start (activate) the newly created watcher.
480		936
		937	=back
		938
		939
		940	=head3 CHECK WATCHERS - customise your event loop even more!
		941
		942	=over 4
481		943
482	=item $w = EV::check $callback	944	=item $w = EV::check $callback
483		945
484	=item $w = EV::check_ns $callback	946	=item $w = EV::check_ns $callback
		947
		948	=item $w = $loop->check ($callback)
		949
		950	=item $w = $loop->check_ns ($callback)
485		951
486	Call the callback just after the process wakes up again (after it has	952	Call the callback just after the process wakes up again (after it has
487	gathered events), but before any other callbacks have been invoked.	953	gathered events), but before any other callbacks have been invoked.
488		954
489	This is used to integrate other event-based software into the EV	955	This can be used to integrate other event-based software into the EV
490	mainloop: You register a prepare callback and in there, you create io and	956	mainloop: You register a prepare callback and in there, you create io and
491	timer watchers as required by the other software. Here is a real-world	957	timer watchers as required by the other software. Here is a real-world
492	example of integrating Net::SNMP (with some details left out):	958	example of integrating Net::SNMP (with some details left out):
493		959
494	our @snmp_watcher;	960	our @snmp_watcher;
…		…
497	# do nothing unless active	963	# do nothing unless active
498	$dispatcher->{_event_queue_h}	964	$dispatcher->{_event_queue_h}
499	or return;	965	or return;
500		966
501	# make the dispatcher handle any outstanding stuff	967	# make the dispatcher handle any outstanding stuff
		968	... not shown
502		969
503	# create an IO watcher for each and every socket	970	# create an I/O watcher for each and every socket
504	@snmp_watcher = (	971	@snmp_watcher = (
505	(map { EV::io $_, EV::READ, sub { } }	972	(map { EV::io $_, EV::READ, sub { } }
506	keys %{ $dispatcher->{_descriptors} }),	973	keys %{ $dispatcher->{_descriptors} }),
		974
		975	EV::timer +($event->[Net::SNMP::Dispatcher::_ACTIVE]
		976	? $event->[Net::SNMP::Dispatcher::_TIME] - EV::now : 0),
		977	0, sub { },
507	);	978	);
508
509	# if there are any timeouts, also create a timer
510	push @snmp_watcher, EV::timer $event->[Net::SNMP::Dispatcher::_TIME] - EV::now, 0, sub { }
511	if $event->[Net::SNMP::Dispatcher::_ACTIVE];
512	};	979	};
513		980
514	The callbacks are irrelevant, the only purpose of those watchers is	981	The callbacks are irrelevant (and are not even being called), the
515	to wake up the process as soon as one of those events occurs (socket	982	only purpose of those watchers is to wake up the process as soon as
516	readable, or timer timed out). The corresponding EV::check watcher will then	983	one of those events occurs (socket readable, or timer timed out). The
517	clean up:	984	corresponding EV::check watcher will then clean up:
518		985
519	our $snmp_check = EV::check sub {	986	our $snmp_check = EV::check sub {
520	# destroy all watchers	987	# destroy all watchers
521	@snmp_watcher = ();	988	@snmp_watcher = ();
522		989
523	# make the dispatcher handle any new stuff	990	# make the dispatcher handle any new stuff
		991	... not shown
524	};	992	};
525		993
526	The callbacks of the created watchers will not be called as the watchers	994	The callbacks of the created watchers will not be called as the watchers
527	are destroyed before this cna happen (remember EV::check gets called	995	are destroyed before this can happen (remember EV::check gets called
528	first).	996	first).
529		997
530	The C<check_ns> variant doesn't start (activate) the newly created watcher.	998	The C<check_ns> variant doesn't start (activate) the newly created watcher.
531		999
532	=back	1000	=item EV::CHECK constant issues
533		1001
		1002	Like all other watcher types, there is a bitmask constant for use in
		1003	C<$revents> and other places. The C<EV::CHECK> is special as it has
		1004	the same name as the C<CHECK> sub called by Perl. This doesn't cause
		1005	big issues on newer perls (beginning with 5.8.9), but it means thatthe
		1006	constant must be I<inlined>, i.e. runtime calls will not work. That means
		1007	that as long as you always C<use EV> and then C<EV::CHECK> you are on the
		1008	safe side.
		1009
		1010	=back
		1011
		1012
		1013	=head3 FORK WATCHERS - the audacity to resume the event loop after a fork
		1014
		1015	Fork watchers are called when a C<fork ()> was detected. The invocation
		1016	is done before the event loop blocks next and before C<check> watchers
		1017	are being called, and only in the child after the fork.
		1018
		1019	=over 4
		1020
		1021	=item $w = EV::fork $callback
		1022
		1023	=item $w = EV::fork_ns $callback
		1024
		1025	=item $w = $loop->fork ($callback)
		1026
		1027	=item $w = $loop->fork_ns ($callback)
		1028
		1029	Call the callback before the event loop is resumed in the child process
		1030	after a fork.
		1031
		1032	The C<fork_ns> variant doesn't start (activate) the newly created watcher.
		1033
		1034	=back
		1035
		1036
		1037	=head3 EMBED WATCHERS - when one backend isn't enough...
		1038
		1039	This is a rather advanced watcher type that lets you embed one event loop
		1040	into another (currently only IO events are supported in the embedded
		1041	loop, other types of watchers might be handled in a delayed or incorrect
		1042	fashion and must not be used).
		1043
		1044	See the libev documentation at
		1045	L<http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod#code_ev_embed_code_when_one_backend_>
		1046	(locally installed as F<EV::libev>) for more details.
		1047
		1048	In short, this watcher is most useful on BSD systems without working
		1049	kqueue to still be able to handle a large number of sockets:
		1050
		1051	my $socket_loop;
		1052
		1053	# check wether we use SELECT or POLL _and_ KQUEUE is supported
		1054	if (
		1055	(EV::backend & (EV::BACKEND_POLL \| EV::BACKEND_SELECT))
		1056	&& (EV::supported_backends & EV::embeddable_backends & EV::BACKEND_KQUEUE)
		1057	) {
		1058	# use kqueue for sockets
		1059	$socket_loop = new EV::Loop EV::BACKEND_KQUEUE \| EV::FLAG_NOENV;
		1060	}
		1061
		1062	# use the default loop otherwise
		1063	$socket_loop \|\|= EV::default_loop;
		1064
		1065	=over 4
		1066
		1067	=item $w = EV::embed $otherloop[, $callback]
		1068
		1069	=item $w = EV::embed_ns $otherloop[, $callback]
		1070
		1071	=item $w = $loop->embed ($otherloop[, $callback])
		1072
		1073	=item $w = $loop->embed_ns ($otherloop[, $callback])
		1074
		1075	Call the callback when the embedded event loop (C<$otherloop>) has any
		1076	I/O activity. The C<$callback> is optional: if it is missing, then the
		1077	embedded event loop will be managed automatically (which is recommended),
		1078	otherwise you have to invoke C<sweep> yourself.
		1079
		1080	The C<embed_ns> variant doesn't start (activate) the newly created watcher.
		1081
		1082	=back
		1083
		1084	=head3 ASYNC WATCHERS - how to wake up another event loop
		1085
		1086	Async watchers are provided by EV, but have little use in perl directly,
		1087	as perl neither supports threads running in parallel nor direct access to
		1088	signal handlers or other contexts where they could be of value.
		1089
		1090	It is, however, possible to use them from the XS level.
		1091
		1092	Please see the libev documentation for further details.
		1093
		1094	=over 4
		1095
		1096	=item $w = EV::async $callback
		1097
		1098	=item $w = EV::async_ns $callback
		1099
		1100	=item $w->send
		1101
		1102	=item $bool = $w->async_pending
		1103
		1104	=back
		1105
		1106
		1107	=head1 PERL SIGNALS
		1108
		1109	While Perl signal handling (C<%SIG>) is not affected by EV, the behaviour
		1110	with EV is as the same as any other C library: Perl-signals will only be
		1111	handled when Perl runs, which means your signal handler might be invoked
		1112	only the next time an event callback is invoked.
		1113
		1114	The solution is to use EV signal watchers (see C<EV::signal>), which will
		1115	ensure proper operations with regards to other event watchers.
		1116
		1117	If you cannot do this for whatever reason, you can also force a watcher
		1118	to be called on every event loop iteration by installing a C<EV::check>
		1119	watcher:
		1120
		1121	my $async_check = EV::check sub { };
		1122
		1123	This ensures that perl gets into control for a short time to handle any
		1124	pending signals, and also ensures (slightly) slower overall operation.
		1125
534	=head1 THREADS	1126	=head1 ITHREADS
535		1127
536	Threads are not supported by this in any way. Perl pseudo-threads is evil	1128	Ithreads are not supported by this module in any way. Perl pseudo-threads
537	stuff and must die.	1129	is evil stuff and must die. Real threads as provided by Coro are fully
		1130	supported (and enhanced support is available via L<Coro::EV>).
		1131
		1132	=head1 FORK
		1133
		1134	Most of the "improved" event delivering mechanisms of modern operating
		1135	systems have quite a few problems with fork(2) (to put it bluntly: it is
		1136	not supported and usually destructive). Libev makes it possible to work
		1137	around this by having a function that recreates the kernel state after
		1138	fork in the child.
		1139
		1140	On non-win32 platforms, this module requires the pthread_atfork
		1141	functionality to do this automatically for you. This function is quite
		1142	buggy on most BSDs, though, so YMMV. The overhead for this is quite
		1143	negligible, because everything the function currently does is set a flag
		1144	that is checked only when the event loop gets used the next time, so when
		1145	you do fork but not use EV, the overhead is minimal.
		1146
		1147	On win32, there is no notion of fork so all this doesn't apply, of course.
538		1148
539	=cut	1149	=cut
540		1150
541	our $DIED = sub {	1151	our $DIED = sub {
542	warn "EV: error in callback (ignoring): $@";	1152	warn "EV: error in callback (ignoring): $@";
543	};	1153	};
544		1154
545	default_loop	1155	default_loop
546	or die 'EV: cannot initialise libev backend. bad $ENV{LIBEV_METHODS}?';	1156	or die 'EV: cannot initialise libev backend. bad $ENV{LIBEV_FLAGS}?';
547
548	push @AnyEvent::REGISTRY, [EV => "EV::AnyEvent"];
549		1157
550	1;	1158	1;
551		1159
552	=head1 SEE ALSO	1160	=head1 SEE ALSO
553		1161
554	L<EV::DNS>, L<EV::AnyEvent>.	1162	L<EV::ADNS> (asynchronous DNS), L<Glib::EV> (makes Glib/Gtk2 use EV as
		1163	event loop), L<EV::Glib> (embed Glib into EV), L<Coro::EV> (efficient
		1164	coroutines with EV), L<Net::SNMP::EV> (asynchronous SNMP), L<AnyEvent> for
		1165	event-loop agnostic and portable event driven programming.
555		1166
556	=head1 AUTHOR	1167	=head1 AUTHOR
557		1168
558	Marc Lehmann <schmorp@schmorp.de>	1169	Marc Lehmann <schmorp@schmorp.de>
559	http://home.schmorp.de/	1170	http://home.schmorp.de/
560		1171
561	=cut	1172	=cut
562		1173

Diff Legend

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

Comparing EV/EV.pm (file contents): Revision 1.32 by root, Thu Nov 8 17:02:10 2007 UTC vs. Revision 1.128 by root, Thu Oct 21 02:46:59 2010 UTC

Diff Legend

Comparing EV/EV.pm (file contents):
Revision 1.32 by root, Thu Nov 8 17:02:10 2007 UTC vs.
Revision 1.128 by root, Thu Oct 21 02:46:59 2010 UTC