ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/AnyEvent/lib/AnyEvent.pm

Comparing AnyEvent/lib/AnyEvent.pm (file contents):
Revision 1.36 by root, Fri Nov 16 09:13:11 2007 UTC vs.
Revision 1.277 by root, Sun Aug 9 13:27:23 2009 UTC

1	=head1 NAME	1	=head1 NAME
2		2
3	AnyEvent - provide framework for multiple event loops	3	AnyEvent - the DBI of event loop programming
4		4
5	Event, Coro, Glib, Tk, Perl - various supported event loops	5	EV, Event, Glib, Tk, Perl, Event::Lib, Irssi, rxvt-unicode, IO::Async, Qt
		6	and POE are various supported event loops/environments.
6		7
7	=head1 SYNOPSIS	8	=head1 SYNOPSIS
8		9
9	use AnyEvent;	10	use AnyEvent;
10		11
		12	# file descriptor readable
11	my $w = AnyEvent->io (fh => $fh, poll => "r|w", cb => sub {	13	my $w = AnyEvent->io (fh => $fh, poll => "r", cb => sub { ... });
		14
		15	# one-shot or repeating timers
		16	my $w = AnyEvent->timer (after => $seconds, cb => sub { ... });
		17	my $w = AnyEvent->timer (after => $seconds, interval => $seconds, cb => ...
		18
		19	print AnyEvent->now; # prints current event loop time
		20	print AnyEvent->time; # think Time::HiRes::time or simply CORE::time.
		21
		22	# POSIX signal
		23	my $w = AnyEvent->signal (signal => "TERM", cb => sub { ... });
		24
		25	# child process exit
		26	my $w = AnyEvent->child (pid => $pid, cb => sub {
		27	my ($pid, $status) = @_;
12	...	28	...
13	});	29	});
14		30
15	my $w = AnyEvent->timer (after => $seconds, cb => sub {	31	# called when event loop idle (if applicable)
16	...	32	my $w = AnyEvent->idle (cb => sub { ... });
17	});
18		33
19	my $w = AnyEvent->condvar; # stores wether a condition was flagged	34	my $w = AnyEvent->condvar; # stores whether a condition was flagged
		35	$w->send; # wake up current and all future recv's
20	$w->wait; # enters "main loop" till $condvar gets ->broadcast	36	$w->recv; # enters "main loop" till $condvar gets ->send
21	$w->broadcast; # wake up current and all future wait's	37	# use a condvar in callback mode:
		38	$w->cb (sub { $_[0]->recv });
		39
		40	=head1 INTRODUCTION/TUTORIAL
		41
		42	This manpage is mainly a reference manual. If you are interested
		43	in a tutorial or some gentle introduction, have a look at the
		44	L<AnyEvent::Intro> manpage.
		45
		46	=head1 SUPPORT
		47
		48	There is a mailinglist for discussing all things AnyEvent, and an IRC
		49	channel, too.
		50
		51	See the AnyEvent project page at the B<Schmorpforge Ta-Sa Software
		52	Repository>, at L<http://anyevent.schmorp.de>, for more info.
		53
		54	=head1 WHY YOU SHOULD USE THIS MODULE (OR NOT)
		55
		56	Glib, POE, IO::Async, Event... CPAN offers event models by the dozen
		57	nowadays. So what is different about AnyEvent?
		58
		59	Executive Summary: AnyEvent is I<compatible>, AnyEvent is I<free of
		60	policy> and AnyEvent is I<small and efficient>.
		61
		62	First and foremost, I<AnyEvent is not an event model> itself, it only
		63	interfaces to whatever event model the main program happens to use, in a
		64	pragmatic way. For event models and certain classes of immortals alike,
		65	the statement "there can only be one" is a bitter reality: In general,
		66	only one event loop can be active at the same time in a process. AnyEvent
		67	cannot change this, but it can hide the differences between those event
		68	loops.
		69
		70	The goal of AnyEvent is to offer module authors the ability to do event
		71	programming (waiting for I/O or timer events) without subscribing to a
		72	religion, a way of living, and most importantly: without forcing your
		73	module users into the same thing by forcing them to use the same event
		74	model you use.
		75
		76	For modules like POE or IO::Async (which is a total misnomer as it is
		77	actually doing all I/O I<synchronously>...), using them in your module is
		78	like joining a cult: After you joined, you are dependent on them and you
		79	cannot use anything else, as they are simply incompatible to everything
		80	that isn't them. What's worse, all the potential users of your
		81	module are I<also> forced to use the same event loop you use.
		82
		83	AnyEvent is different: AnyEvent + POE works fine. AnyEvent + Glib works
		84	fine. AnyEvent + Tk works fine etc. etc. but none of these work together
		85	with the rest: POE + IO::Async? No go. Tk + Event? No go. Again: if
		86	your module uses one of those, every user of your module has to use it,
		87	too. But if your module uses AnyEvent, it works transparently with all
		88	event models it supports (including stuff like IO::Async, as long as those
		89	use one of the supported event loops. It is trivial to add new event loops
		90	to AnyEvent, too, so it is future-proof).
		91
		92	In addition to being free of having to use I<the one and only true event
		93	model>, AnyEvent also is free of bloat and policy: with POE or similar
		94	modules, you get an enormous amount of code and strict rules you have to
		95	follow. AnyEvent, on the other hand, is lean and up to the point, by only
		96	offering the functionality that is necessary, in as thin as a wrapper as
		97	technically possible.
		98
		99	Of course, AnyEvent comes with a big (and fully optional!) toolbox
		100	of useful functionality, such as an asynchronous DNS resolver, 100%
		101	non-blocking connects (even with TLS/SSL, IPv6 and on broken platforms
		102	such as Windows) and lots of real-world knowledge and workarounds for
		103	platform bugs and differences.
		104
		105	Now, if you I<do want> lots of policy (this can arguably be somewhat
		106	useful) and you want to force your users to use the one and only event
		107	model, you should I<not> use this module.
22		108
23	=head1 DESCRIPTION	109	=head1 DESCRIPTION
24		110
25	L<AnyEvent> provides an identical interface to multiple event loops. This	111	L<AnyEvent> provides an identical interface to multiple event loops. This
26	allows module authors to utilise an event loop without forcing module	112	allows module authors to utilise an event loop without forcing module
27	users to use the same event loop (as only a single event loop can coexist	113	users to use the same event loop (as only a single event loop can coexist
28	peacefully at any one time).	114	peacefully at any one time).
29		115
30	The interface itself is vaguely similar but not identical to the Event	116	The interface itself is vaguely similar, but not identical to the L<Event>
31	module.	117	module.
32		118
33	On the first call of any method, the module tries to detect the currently	119	During the first call of any watcher-creation method, the module tries
34	loaded event loop by probing wether any of the following modules is	120	to detect the currently loaded event loop by probing whether one of the
35	loaded: L<Coro::Event>, L<Event>, L<Glib>, L<Tk>. The first one found is	121	following modules is already loaded: L<EV>,
36	used. If none is found, the module tries to load these modules in the	122	L<Event>, L<Glib>, L<AnyEvent::Impl::Perl>, L<Tk>, L<Event::Lib>, L<Qt>,
37	order given. The first one that could be successfully loaded will be	123	L<POE>. The first one found is used. If none are found, the module tries
38	used. If still none could be found, AnyEvent will fall back to a pure-perl	124	to load these modules (excluding Tk, Event::Lib, Qt and POE as the pure perl
39	event loop, which is also not very efficient.	125	adaptor should always succeed) in the order given. The first one that can
		126	be successfully loaded will be used. If, after this, still none could be
		127	found, AnyEvent will fall back to a pure-perl event loop, which is not
		128	very efficient, but should work everywhere.
40		129
41	Because AnyEvent first checks for modules that are already loaded, loading	130	Because AnyEvent first checks for modules that are already loaded, loading
42	an Event model explicitly before first using AnyEvent will likely make	131	an event model explicitly before first using AnyEvent will likely make
43	that model the default. For example:	132	that model the default. For example:
44		133
45	use Tk;	134	use Tk;
46	use AnyEvent;	135	use AnyEvent;
47		136
48	# .. AnyEvent will likely default to Tk	137	# .. AnyEvent will likely default to Tk
49		138
		139	The I<likely> means that, if any module loads another event model and
		140	starts using it, all bets are off. Maybe you should tell their authors to
		141	use AnyEvent so their modules work together with others seamlessly...
		142
50	The pure-perl implementation of AnyEvent is called	143	The pure-perl implementation of AnyEvent is called
51	C<AnyEvent::Impl::Perl>. Like other event modules you can load it	144	C<AnyEvent::Impl::Perl>. Like other event modules you can load it
52	explicitly.	145	explicitly and enjoy the high availability of that event loop :)
53		146
54	=head1 WATCHERS	147	=head1 WATCHERS
55		148
56	AnyEvent has the central concept of a I<watcher>, which is an object that	149	AnyEvent has the central concept of a I<watcher>, which is an object that
57	stores relevant data for each kind of event you are waiting for, such as	150	stores relevant data for each kind of event you are waiting for, such as
58	the callback to call, the filehandle to watch, etc.	151	the callback to call, the file handle to watch, etc.
59		152
60	These watchers are normal Perl objects with normal Perl lifetime. After	153	These watchers are normal Perl objects with normal Perl lifetime. After
61	creating a watcher it will immediately "watch" for events and invoke	154	creating a watcher it will immediately "watch" for events and invoke the
		155	callback when the event occurs (of course, only when the event model
		156	is in control).
		157
		158	Note that B<callbacks must not permanently change global variables>
		159	potentially in use by the event loop (such as C<$_> or C<$[>) and that B<<
		160	callbacks must not C<die> >>. The former is good programming practise in
		161	Perl and the latter stems from the fact that exception handling differs
		162	widely between event loops.
		163
62	the callback. To disable the watcher you have to destroy it (e.g. by	164	To disable the watcher you have to destroy it (e.g. by setting the
63	setting the variable that stores it to C<undef> or otherwise deleting all	165	variable you store it in to C<undef> or otherwise deleting all references
64	references to it).	166	to it).
65		167
66	All watchers are created by calling a method on the C<AnyEvent> class.	168	All watchers are created by calling a method on the C<AnyEvent> class.
67		169
		170	Many watchers either are used with "recursion" (repeating timers for
		171	example), or need to refer to their watcher object in other ways.
		172
		173	An any way to achieve that is this pattern:
		174
		175	my $w; $w = AnyEvent->type (arg => value ..., cb => sub {
		176	# you can use $w here, for example to undef it
		177	undef $w;
		178	});
		179
		180	Note that C<my $w; $w => combination. This is necessary because in Perl,
		181	my variables are only visible after the statement in which they are
		182	declared.
		183
68	=head2 IO WATCHERS	184	=head2 I/O WATCHERS
69		185
		186	$w = AnyEvent->io (
		187	fh => <filehandle_or_fileno>,
		188	poll => <"r" or "w">,
		189	cb => <callback>,
		190	);
		191
70	You can create I/O watcher by calling the C<< AnyEvent->io >> method with	192	You can create an I/O watcher by calling the C<< AnyEvent->io >> method
71	the following mandatory arguments:	193	with the following mandatory key-value pairs as arguments:
72		194
73	C<fh> the Perl I<filehandle> (not filedescriptor) to watch for	195	C<fh> is the Perl I<file handle> (or a naked file descriptor) to watch
		196	for events (AnyEvent might or might not keep a reference to this file
		197	handle). Note that only file handles pointing to things for which
		198	non-blocking operation makes sense are allowed. This includes sockets,
		199	most character devices, pipes, fifos and so on, but not for example files
		200	or block devices.
		201
74	events. C<poll> must be a string that is either C<r> or C<w>, that creates	202	C<poll> must be a string that is either C<r> or C<w>, which creates a
75	a watcher waiting for "r"eadable or "w"ritable events. C<cb> the callback	203	watcher waiting for "r"eadable or "w"ritable events, respectively.
76	to invoke everytime the filehandle becomes ready.
77		204
78	Only one io watcher per C<fh> and C<poll> combination is allowed (i.e. on	205	C<cb> is the callback to invoke each time the file handle becomes ready.
79	a socket you can have one r + one w, not any more (limitation comes from
80	Tk - if you are sure you are not using Tk this limitation is gone).
81		206
82	Filehandles will be kept alive, so as long as the watcher exists, the	207	Although the callback might get passed parameters, their value and
83	filehandle exists, too.	208	presence is undefined and you cannot rely on them. Portable AnyEvent
		209	callbacks cannot use arguments passed to I/O watcher callbacks.
84		210
85	Example:	211	The I/O watcher might use the underlying file descriptor or a copy of it.
		212	You must not close a file handle as long as any watcher is active on the
		213	underlying file descriptor.
86		214
		215	Some event loops issue spurious readyness notifications, so you should
		216	always use non-blocking calls when reading/writing from/to your file
		217	handles.
		218
87	# wait for readability of STDIN, then read a line and disable the watcher	219	Example: wait for readability of STDIN, then read a line and disable the
		220	watcher.
		221
88	my $w; $w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {	222	my $w; $w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {
89	chomp (my $input = <STDIN>);	223	chomp (my $input = <STDIN>);
90	warn "read: $input\n";	224	warn "read: $input\n";
91	undef $w;	225	undef $w;
92	});	226	});
93		227
94	=head2 TIME WATCHERS	228	=head2 TIME WATCHERS
95		229
		230	$w = AnyEvent->timer (after => <seconds>, cb => <callback>);
		231
		232	$w = AnyEvent->timer (
		233	after => <fractional_seconds>,
		234	interval => <fractional_seconds>,
		235	cb => <callback>,
		236	);
		237
96	You can create a time watcher by calling the C<< AnyEvent->timer >>	238	You can create a time watcher by calling the C<< AnyEvent->timer >>
97	method with the following mandatory arguments:	239	method with the following mandatory arguments:
98		240
99	C<after> after how many seconds (fractions are supported) should the timer	241	C<after> specifies after how many seconds (fractional values are
100	activate. C<cb> the callback to invoke.	242	supported) the callback should be invoked. C<cb> is the callback to invoke
		243	in that case.
101		244
102	The timer callback will be invoked at most once: if you want a repeating	245	Although the callback might get passed parameters, their value and
103	timer you have to create a new watcher (this is a limitation by both Tk	246	presence is undefined and you cannot rely on them. Portable AnyEvent
104	and Glib).	247	callbacks cannot use arguments passed to time watcher callbacks.
105		248
106	Example:	249	The callback will normally be invoked once only. If you specify another
		250	parameter, C<interval>, as a strictly positive number (> 0), then the
		251	callback will be invoked regularly at that interval (in fractional
		252	seconds) after the first invocation. If C<interval> is specified with a
		253	false value, then it is treated as if it were missing.
107		254
		255	The callback will be rescheduled before invoking the callback, but no
		256	attempt is done to avoid timer drift in most backends, so the interval is
		257	only approximate.
		258
108	# fire an event after 7.7 seconds	259	Example: fire an event after 7.7 seconds.
		260
109	my $w = AnyEvent->timer (after => 7.7, cb => sub {	261	my $w = AnyEvent->timer (after => 7.7, cb => sub {
110	warn "timeout\n";	262	warn "timeout\n";
111	});	263	});
112		264
113	# to cancel the timer:	265	# to cancel the timer:
114	undef $w	266	undef $w;
115		267
116	=head2 CONDITION WATCHERS	268	Example 2: fire an event after 0.5 seconds, then roughly every second.
117		269
118	Condition watchers can be created by calling the C<< AnyEvent->condvar >>	270	my $w = AnyEvent->timer (after => 0.5, interval => 1, cb => sub {
119	method without any arguments.	271	warn "timeout\n";
		272	};
120		273
121	A condition watcher watches for a condition - precisely that the C<<	274	=head3 TIMING ISSUES
122	->broadcast >> method has been called.
123		275
124	The watcher has only two methods:	276	There are two ways to handle timers: based on real time (relative, "fire
		277	in 10 seconds") and based on wallclock time (absolute, "fire at 12
		278	o'clock").
		279
		280	While most event loops expect timers to specified in a relative way, they
		281	use absolute time internally. This makes a difference when your clock
		282	"jumps", for example, when ntp decides to set your clock backwards from
		283	the wrong date of 2014-01-01 to 2008-01-01, a watcher that is supposed to
		284	fire "after" a second might actually take six years to finally fire.
		285
		286	AnyEvent cannot compensate for this. The only event loop that is conscious
		287	about these issues is L<EV>, which offers both relative (ev_timer, based
		288	on true relative time) and absolute (ev_periodic, based on wallclock time)
		289	timers.
		290
		291	AnyEvent always prefers relative timers, if available, matching the
		292	AnyEvent API.
		293
		294	AnyEvent has two additional methods that return the "current time":
125		295
126	=over 4	296	=over 4
127		297
128	=item $cv->wait	298	=item AnyEvent->time
129		299
130	Wait (blocking if necessary) until the C<< ->broadcast >> method has been	300	This returns the "current wallclock time" as a fractional number of
131	called on c<$cv>, while servicing other watchers normally.	301	seconds since the Epoch (the same thing as C<time> or C<Time::HiRes::time>
		302	return, and the result is guaranteed to be compatible with those).
132		303
133	Not all event models support a blocking wait - some die in that case, so	304	It progresses independently of any event loop processing, i.e. each call
134	if you are using this from a module, never require a blocking wait, but	305	will check the system clock, which usually gets updated frequently.
135	let the caller decide wether the call will block or not (for example,
136	by coupling condition variables with some kind of request results and
137	supporting callbacks so the caller knows that getting the result will not
138	block, while still suppporting blockign waits if the caller so desires).
139		306
140	You can only wait once on a condition - additional calls will return	307	=item AnyEvent->now
141	immediately.
142		308
143	=item $cv->broadcast	309	This also returns the "current wallclock time", but unlike C<time>, above,
		310	this value might change only once per event loop iteration, depending on
		311	the event loop (most return the same time as C<time>, above). This is the
		312	time that AnyEvent's timers get scheduled against.
144		313
145	Flag the condition as ready - a running C<< ->wait >> and all further	314	I<In almost all cases (in all cases if you don't care), this is the
146	calls to C<wait> will return after this method has been called. If nobody	315	function to call when you want to know the current time.>
147	is waiting the broadcast will be remembered..
148		316
149	Example:	317	This function is also often faster then C<< AnyEvent->time >>, and
		318	thus the preferred method if you want some timestamp (for example,
		319	L<AnyEvent::Handle> uses this to update it's activity timeouts).
		320
		321	The rest of this section is only of relevance if you try to be very exact
		322	with your timing, you can skip it without bad conscience.
		323
		324	For a practical example of when these times differ, consider L<Event::Lib>
		325	and L<EV> and the following set-up:
		326
		327	The event loop is running and has just invoked one of your callback at
		328	time=500 (assume no other callbacks delay processing). In your callback,
		329	you wait a second by executing C<sleep 1> (blocking the process for a
		330	second) and then (at time=501) you create a relative timer that fires
		331	after three seconds.
		332
		333	With L<Event::Lib>, C<< AnyEvent->time >> and C<< AnyEvent->now >> will
		334	both return C<501>, because that is the current time, and the timer will
		335	be scheduled to fire at time=504 (C<501> + C<3>).
		336
		337	With L<EV>, C<< AnyEvent->time >> returns C<501> (as that is the current
		338	time), but C<< AnyEvent->now >> returns C<500>, as that is the time the
		339	last event processing phase started. With L<EV>, your timer gets scheduled
		340	to run at time=503 (C<500> + C<3>).
		341
		342	In one sense, L<Event::Lib> is more exact, as it uses the current time
		343	regardless of any delays introduced by event processing. However, most
		344	callbacks do not expect large delays in processing, so this causes a
		345	higher drift (and a lot more system calls to get the current time).
		346
		347	In another sense, L<EV> is more exact, as your timer will be scheduled at
		348	the same time, regardless of how long event processing actually took.
		349
		350	In either case, if you care (and in most cases, you don't), then you
		351	can get whatever behaviour you want with any event loop, by taking the
		352	difference between C<< AnyEvent->time >> and C<< AnyEvent->now >> into
		353	account.
		354
		355	=item AnyEvent->now_update
		356
		357	Some event loops (such as L<EV> or L<AnyEvent::Impl::Perl>) cache
		358	the current time for each loop iteration (see the discussion of L<<
		359	AnyEvent->now >>, above).
		360
		361	When a callback runs for a long time (or when the process sleeps), then
		362	this "current" time will differ substantially from the real time, which
		363	might affect timers and time-outs.
		364
		365	When this is the case, you can call this method, which will update the
		366	event loop's idea of "current time".
		367
		368	Note that updating the time I<might> cause some events to be handled.
		369
		370	=back
		371
		372	=head2 SIGNAL WATCHERS
		373
		374	$w = AnyEvent->signal (signal => <uppercase_signal_name>, cb => <callback>);
		375
		376	You can watch for signals using a signal watcher, C<signal> is the signal
		377	I<name> in uppercase and without any C<SIG> prefix, C<cb> is the Perl
		378	callback to be invoked whenever a signal occurs.
		379
		380	Although the callback might get passed parameters, their value and
		381	presence is undefined and you cannot rely on them. Portable AnyEvent
		382	callbacks cannot use arguments passed to signal watcher callbacks.
		383
		384	Multiple signal occurrences can be clumped together into one callback
		385	invocation, and callback invocation will be synchronous. Synchronous means
		386	that it might take a while until the signal gets handled by the process,
		387	but it is guaranteed not to interrupt any other callbacks.
		388
		389	The main advantage of using these watchers is that you can share a signal
		390	between multiple watchers, and AnyEvent will ensure that signals will not
		391	interrupt your program at bad times.
		392
		393	This watcher might use C<%SIG> (depending on the event loop used),
		394	so programs overwriting those signals directly will likely not work
		395	correctly.
		396
		397	Example: exit on SIGINT
		398
		399	my $w = AnyEvent->signal (signal => "INT", cb => sub { exit 1 });
		400
		401	=head3 Signal Races, Delays and Workarounds
		402
		403	Many event loops (e.g. Glib, Tk, Qt, IO::Async) do not support attaching
		404	callbacks to signals in a generic way, which is a pity, as you cannot
		405	do race-free signal handling in perl, requiring C libraries for
		406	this. AnyEvent will try to do it's best, which means in some cases,
		407	signals will be delayed. The maximum time a signal might be delayed is
		408	specified in C<$AnyEvent::MAX_SIGNAL_LATENCY> (default: 10 seconds). This
		409	variable can be changed only before the first signal watcher is created,
		410	and should be left alone otherwise. This variable determines how often
		411	AnyEvent polls for signals (in case a wake-up was missed). Higher values
		412	will cause fewer spurious wake-ups, which is better for power and CPU
		413	saving.
		414
		415	All these problems can be avoided by installing the optional
		416	L<Async::Interrupt> module, which works with most event loops. It will not
		417	work with inherently broken event loops such as L<Event> or L<Event::Lib>
		418	(and not with L<POE> currently, as POE does it's own workaround with
		419	one-second latency). For those, you just have to suffer the delays.
		420
		421	=head2 CHILD PROCESS WATCHERS
		422
		423	$w = AnyEvent->child (pid => <process id>, cb => <callback>);
		424
		425	You can also watch on a child process exit and catch its exit status.
		426
		427	The child process is specified by the C<pid> argument (one some backends,
		428	using C<0> watches for any child process exit, on others this will
		429	croak). The watcher will be triggered only when the child process has
		430	finished and an exit status is available, not on any trace events
		431	(stopped/continued).
		432
		433	The callback will be called with the pid and exit status (as returned by
		434	waitpid), so unlike other watcher types, you I<can> rely on child watcher
		435	callback arguments.
		436
		437	This watcher type works by installing a signal handler for C<SIGCHLD>,
		438	and since it cannot be shared, nothing else should use SIGCHLD or reap
		439	random child processes (waiting for specific child processes, e.g. inside
		440	C<system>, is just fine).
		441
		442	There is a slight catch to child watchers, however: you usually start them
		443	I<after> the child process was created, and this means the process could
		444	have exited already (and no SIGCHLD will be sent anymore).
		445
		446	Not all event models handle this correctly (neither POE nor IO::Async do,
		447	see their AnyEvent::Impl manpages for details), but even for event models
		448	that I<do> handle this correctly, they usually need to be loaded before
		449	the process exits (i.e. before you fork in the first place). AnyEvent's
		450	pure perl event loop handles all cases correctly regardless of when you
		451	start the watcher.
		452
		453	This means you cannot create a child watcher as the very first
		454	thing in an AnyEvent program, you I<have> to create at least one
		455	watcher before you C<fork> the child (alternatively, you can call
		456	C<AnyEvent::detect>).
		457
		458	As most event loops do not support waiting for child events, they will be
		459	emulated by AnyEvent in most cases, in which the latency and race problems
		460	mentioned in the description of signal watchers apply.
		461
		462	Example: fork a process and wait for it
		463
		464	my $done = AnyEvent->condvar;
		465
		466	my $pid = fork or exit 5;
		467
		468	my $w = AnyEvent->child (
		469	pid => $pid,
		470	cb => sub {
		471	my ($pid, $status) = @_;
		472	warn "pid $pid exited with status $status";
		473	$done->send;
		474	},
		475	);
		476
		477	# do something else, then wait for process exit
		478	$done->recv;
		479
		480	=head2 IDLE WATCHERS
		481
		482	$w = AnyEvent->idle (cb => <callback>);
		483
		484	Sometimes there is a need to do something, but it is not so important
		485	to do it instantly, but only when there is nothing better to do. This
		486	"nothing better to do" is usually defined to be "no other events need
		487	attention by the event loop".
		488
		489	Idle watchers ideally get invoked when the event loop has nothing
		490	better to do, just before it would block the process to wait for new
		491	events. Instead of blocking, the idle watcher is invoked.
		492
		493	Most event loops unfortunately do not really support idle watchers (only
		494	EV, Event and Glib do it in a usable fashion) - for the rest, AnyEvent
		495	will simply call the callback "from time to time".
		496
		497	Example: read lines from STDIN, but only process them when the
		498	program is otherwise idle:
		499
		500	my @lines; # read data
		501	my $idle_w;
		502	my $io_w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {
		503	push @lines, scalar <STDIN>;
		504
		505	# start an idle watcher, if not already done
		506	$idle_w ||= AnyEvent->idle (cb => sub {
		507	# handle only one line, when there are lines left
		508	if (my $line = shift @lines) {
		509	print "handled when idle: $line";
		510	} else {
		511	# otherwise disable the idle watcher again
		512	undef $idle_w;
		513	}
		514	});
		515	});
		516
		517	=head2 CONDITION VARIABLES
		518
		519	$cv = AnyEvent->condvar;
		520
		521	$cv->send (<list>);
		522	my @res = $cv->recv;
		523
		524	If you are familiar with some event loops you will know that all of them
		525	require you to run some blocking "loop", "run" or similar function that
		526	will actively watch for new events and call your callbacks.
		527
		528	AnyEvent is slightly different: it expects somebody else to run the event
		529	loop and will only block when necessary (usually when told by the user).
		530
		531	The instrument to do that is called a "condition variable", so called
		532	because they represent a condition that must become true.
		533
		534	Now is probably a good time to look at the examples further below.
		535
		536	Condition variables can be created by calling the C<< AnyEvent->condvar
		537	>> method, usually without arguments. The only argument pair allowed is
		538	C<cb>, which specifies a callback to be called when the condition variable
		539	becomes true, with the condition variable as the first argument (but not
		540	the results).
		541
		542	After creation, the condition variable is "false" until it becomes "true"
		543	by calling the C<send> method (or calling the condition variable as if it
		544	were a callback, read about the caveats in the description for the C<<
		545	->send >> method).
		546
		547	Condition variables are similar to callbacks, except that you can
		548	optionally wait for them. They can also be called merge points - points
		549	in time where multiple outstanding events have been processed. And yet
		550	another way to call them is transactions - each condition variable can be
		551	used to represent a transaction, which finishes at some point and delivers
		552	a result. And yet some people know them as "futures" - a promise to
		553	compute/deliver something that you can wait for.
		554
		555	Condition variables are very useful to signal that something has finished,
		556	for example, if you write a module that does asynchronous http requests,
		557	then a condition variable would be the ideal candidate to signal the
		558	availability of results. The user can either act when the callback is
		559	called or can synchronously C<< ->recv >> for the results.
		560
		561	You can also use them to simulate traditional event loops - for example,
		562	you can block your main program until an event occurs - for example, you
		563	could C<< ->recv >> in your main program until the user clicks the Quit
		564	button of your app, which would C<< ->send >> the "quit" event.
		565
		566	Note that condition variables recurse into the event loop - if you have
		567	two pieces of code that call C<< ->recv >> in a round-robin fashion, you
		568	lose. Therefore, condition variables are good to export to your caller, but
		569	you should avoid making a blocking wait yourself, at least in callbacks,
		570	as this asks for trouble.
		571
		572	Condition variables are represented by hash refs in perl, and the keys
		573	used by AnyEvent itself are all named C<_ae_XXX> to make subclassing
		574	easy (it is often useful to build your own transaction class on top of
		575	AnyEvent). To subclass, use C<AnyEvent::CondVar> as base class and call
		576	it's C<new> method in your own C<new> method.
		577
		578	There are two "sides" to a condition variable - the "producer side" which
		579	eventually calls C<< -> send >>, and the "consumer side", which waits
		580	for the send to occur.
		581
		582	Example: wait for a timer.
150		583
151	# wait till the result is ready	584	# wait till the result is ready
152	my $result_ready = AnyEvent->condvar;	585	my $result_ready = AnyEvent->condvar;
153		586
154	# do something such as adding a timer	587	# do something such as adding a timer
155	# or socket watcher the calls $result_ready->broadcast	588	# or socket watcher the calls $result_ready->send
156	# when the "result" is ready.	589	# when the "result" is ready.
		590	# in this case, we simply use a timer:
		591	my $w = AnyEvent->timer (
		592	after => 1,
		593	cb => sub { $result_ready->send },
		594	);
157		595
		596	# this "blocks" (while handling events) till the callback
		597	# calls -<send
158	$result_ready->wait;	598	$result_ready->recv;
		599
		600	Example: wait for a timer, but take advantage of the fact that condition
		601	variables are also callable directly.
		602
		603	my $done = AnyEvent->condvar;
		604	my $delay = AnyEvent->timer (after => 5, cb => $done);
		605	$done->recv;
		606
		607	Example: Imagine an API that returns a condvar and doesn't support
		608	callbacks. This is how you make a synchronous call, for example from
		609	the main program:
		610
		611	use AnyEvent::CouchDB;
		612
		613	...
		614
		615	my @info = $couchdb->info->recv;
		616
		617	And this is how you would just set a callback to be called whenever the
		618	results are available:
		619
		620	$couchdb->info->cb (sub {
		621	my @info = $_[0]->recv;
		622	});
		623
		624	=head3 METHODS FOR PRODUCERS
		625
		626	These methods should only be used by the producing side, i.e. the
		627	code/module that eventually sends the signal. Note that it is also
		628	the producer side which creates the condvar in most cases, but it isn't
		629	uncommon for the consumer to create it as well.
		630
		631	=over 4
		632
		633	=item $cv->send (...)
		634
		635	Flag the condition as ready - a running C<< ->recv >> and all further
		636	calls to C<recv> will (eventually) return after this method has been
		637	called. If nobody is waiting the send will be remembered.
		638
		639	If a callback has been set on the condition variable, it is called
		640	immediately from within send.
		641
		642	Any arguments passed to the C<send> call will be returned by all
		643	future C<< ->recv >> calls.
		644
		645	Condition variables are overloaded so one can call them directly (as if
		646	they were a code reference). Calling them directly is the same as calling
		647	C<send>.
		648
		649	=item $cv->croak ($error)
		650
		651	Similar to send, but causes all call's to C<< ->recv >> to invoke
		652	C<Carp::croak> with the given error message/object/scalar.
		653
		654	This can be used to signal any errors to the condition variable
		655	user/consumer. Doing it this way instead of calling C<croak> directly
		656	delays the error detetcion, but has the overwhelmign advantage that it
		657	diagnoses the error at the place where the result is expected, and not
		658	deep in some event clalback without connection to the actual code causing
		659	the problem.
		660
		661	=item $cv->begin ([group callback])
		662
		663	=item $cv->end
		664
		665	These two methods can be used to combine many transactions/events into
		666	one. For example, a function that pings many hosts in parallel might want
		667	to use a condition variable for the whole process.
		668
		669	Every call to C<< ->begin >> will increment a counter, and every call to
		670	C<< ->end >> will decrement it. If the counter reaches C<0> in C<< ->end
		671	>>, the (last) callback passed to C<begin> will be executed. That callback
		672	is I<supposed> to call C<< ->send >>, but that is not required. If no
		673	callback was set, C<send> will be called without any arguments.
		674
		675	You can think of C<< $cv->send >> giving you an OR condition (one call
		676	sends), while C<< $cv->begin >> and C<< $cv->end >> giving you an AND
		677	condition (all C<begin> calls must be C<end>'ed before the condvar sends).
		678
		679	Let's start with a simple example: you have two I/O watchers (for example,
		680	STDOUT and STDERR for a program), and you want to wait for both streams to
		681	close before activating a condvar:
		682
		683	my $cv = AnyEvent->condvar;
		684
		685	$cv->begin; # first watcher
		686	my $w1 = AnyEvent->io (fh => $fh1, cb => sub {
		687	defined sysread $fh1, my $buf, 4096
		688	or $cv->end;
		689	});
		690
		691	$cv->begin; # second watcher
		692	my $w2 = AnyEvent->io (fh => $fh2, cb => sub {
		693	defined sysread $fh2, my $buf, 4096
		694	or $cv->end;
		695	});
		696
		697	$cv->recv;
		698
		699	This works because for every event source (EOF on file handle), there is
		700	one call to C<begin>, so the condvar waits for all calls to C<end> before
		701	sending.
		702
		703	The ping example mentioned above is slightly more complicated, as the
		704	there are results to be passwd back, and the number of tasks that are
		705	begung can potentially be zero:
		706
		707	my $cv = AnyEvent->condvar;
		708
		709	my %result;
		710	$cv->begin (sub { $cv->send (\%result) });
		711
		712	for my $host (@list_of_hosts) {
		713	$cv->begin;
		714	ping_host_then_call_callback $host, sub {
		715	$result{$host} = ...;
		716	$cv->end;
		717	};
		718	}
		719
		720	$cv->end;
		721
		722	This code fragment supposedly pings a number of hosts and calls
		723	C<send> after results for all then have have been gathered - in any
		724	order. To achieve this, the code issues a call to C<begin> when it starts
		725	each ping request and calls C<end> when it has received some result for
		726	it. Since C<begin> and C<end> only maintain a counter, the order in which
		727	results arrive is not relevant.
		728
		729	There is an additional bracketing call to C<begin> and C<end> outside the
		730	loop, which serves two important purposes: first, it sets the callback
		731	to be called once the counter reaches C<0>, and second, it ensures that
		732	C<send> is called even when C<no> hosts are being pinged (the loop
		733	doesn't execute once).
		734
		735	This is the general pattern when you "fan out" into multiple (but
		736	potentially none) subrequests: use an outer C<begin>/C<end> pair to set
		737	the callback and ensure C<end> is called at least once, and then, for each
		738	subrequest you start, call C<begin> and for each subrequest you finish,
		739	call C<end>.
159		740
160	=back	741	=back
161		742
162	=head2 SIGNAL WATCHERS	743	=head3 METHODS FOR CONSUMERS
163		744
164	You can listen for signals using a signal watcher, C<signal> is the signal	745	These methods should only be used by the consuming side, i.e. the
165	I<name> without any C<SIG> prefix. Multiple signals events can be clumped	746	code awaits the condition.
166	together into one callback invocation, and callback invocation might or
167	might not be asynchronous.
168
169	These watchers might use C<%SIG>, so programs overwriting those signals
170	directly will likely not work correctly.
171
172	Example: exit on SIGINT
173
174	my $w = AnyEvent->signal (signal => "INT", cb => sub { exit 1 });
175
176	=head2 CHILD PROCESS WATCHERS
177
178	You can also listen for the status of a child process specified by the
179	C<pid> argument (or any child if the pid argument is 0). The watcher will
180	trigger as often as status change for the child are received. This works
181	by installing a signal handler for C<SIGCHLD>. The callback will be called with
182	the pid and exit status (as returned by waitpid).
183
184	Example: wait for pid 1333
185
186	my $w = AnyEvent->child (pid => 1333, cb => sub { warn "exit status $?" });
187
188	=head1 GLOBALS
189		747
190	=over 4	748	=over 4
191		749
		750	=item $cv->recv
		751
		752	Wait (blocking if necessary) until the C<< ->send >> or C<< ->croak
		753	>> methods have been called on c<$cv>, while servicing other watchers
		754	normally.
		755
		756	You can only wait once on a condition - additional calls are valid but
		757	will return immediately.
		758
		759	If an error condition has been set by calling C<< ->croak >>, then this
		760	function will call C<croak>.
		761
		762	In list context, all parameters passed to C<send> will be returned,
		763	in scalar context only the first one will be returned.
		764
		765	Note that doing a blocking wait in a callback is not supported by any
		766	event loop, that is, recursive invocation of a blocking C<< ->recv
		767	>> is not allowed, and the C<recv> call will C<croak> if such a
		768	condition is detected. This condition can be slightly loosened by using
		769	L<Coro::AnyEvent>, which allows you to do a blocking C<< ->recv >> from
		770	any thread that doesn't run the event loop itself.
		771
		772	Not all event models support a blocking wait - some die in that case
		773	(programs might want to do that to stay interactive), so I<if you are
		774	using this from a module, never require a blocking wait>. Instead, let the
		775	caller decide whether the call will block or not (for example, by coupling
		776	condition variables with some kind of request results and supporting
		777	callbacks so the caller knows that getting the result will not block,
		778	while still supporting blocking waits if the caller so desires).
		779
		780	You can ensure that C<< -recv >> never blocks by setting a callback and
		781	only calling C<< ->recv >> from within that callback (or at a later
		782	time). This will work even when the event loop does not support blocking
		783	waits otherwise.
		784
		785	=item $bool = $cv->ready
		786
		787	Returns true when the condition is "true", i.e. whether C<send> or
		788	C<croak> have been called.
		789
		790	=item $cb = $cv->cb ($cb->($cv))
		791
		792	This is a mutator function that returns the callback set and optionally
		793	replaces it before doing so.
		794
		795	The callback will be called when the condition becomes (or already was)
		796	"true", i.e. when C<send> or C<croak> are called (or were called), with
		797	the only argument being the condition variable itself. Calling C<recv>
		798	inside the callback or at any later time is guaranteed not to block.
		799
		800	=back
		801
		802	=head1 SUPPORTED EVENT LOOPS/BACKENDS
		803
		804	The available backend classes are (every class has its own manpage):
		805
		806	=over 4
		807
		808	=item Backends that are autoprobed when no other event loop can be found.
		809
		810	EV is the preferred backend when no other event loop seems to be in
		811	use. If EV is not installed, then AnyEvent will fall back to its own
		812	pure-perl implementation, which is available everywhere as it comes with
		813	AnyEvent itself.
		814
		815	AnyEvent::Impl::EV based on EV (interface to libev, best choice).
		816	AnyEvent::Impl::Perl pure-perl implementation, fast and portable.
		817
		818	=item Backends that are transparently being picked up when they are used.
		819
		820	These will be used when they are currently loaded when the first watcher
		821	is created, in which case it is assumed that the application is using
		822	them. This means that AnyEvent will automatically pick the right backend
		823	when the main program loads an event module before anything starts to
		824	create watchers. Nothing special needs to be done by the main program.
		825
		826	AnyEvent::Impl::Event based on Event, very stable, few glitches.
		827	AnyEvent::Impl::Glib based on Glib, slow but very stable.
		828	AnyEvent::Impl::Tk based on Tk, very broken.
		829	AnyEvent::Impl::EventLib based on Event::Lib, leaks memory and worse.
		830	AnyEvent::Impl::POE based on POE, very slow, some limitations.
		831	AnyEvent::Impl::Irssi used when running within irssi.
		832
		833	=item Backends with special needs.
		834
		835	Qt requires the Qt::Application to be instantiated first, but will
		836	otherwise be picked up automatically. As long as the main program
		837	instantiates the application before any AnyEvent watchers are created,
		838	everything should just work.
		839
		840	AnyEvent::Impl::Qt based on Qt.
		841
		842	Support for IO::Async can only be partial, as it is too broken and
		843	architecturally limited to even support the AnyEvent API. It also
		844	is the only event loop that needs the loop to be set explicitly, so
		845	it can only be used by a main program knowing about AnyEvent. See
		846	L<AnyEvent::Impl::Async> for the gory details.
		847
		848	AnyEvent::Impl::IOAsync based on IO::Async, cannot be autoprobed.
		849
		850	=item Event loops that are indirectly supported via other backends.
		851
		852	Some event loops can be supported via other modules:
		853
		854	There is no direct support for WxWidgets (L<Wx>) or L<Prima>.
		855
		856	B<WxWidgets> has no support for watching file handles. However, you can
		857	use WxWidgets through the POE adaptor, as POE has a Wx backend that simply
		858	polls 20 times per second, which was considered to be too horrible to even
		859	consider for AnyEvent.
		860
		861	B<Prima> is not supported as nobody seems to be using it, but it has a POE
		862	backend, so it can be supported through POE.
		863
		864	AnyEvent knows about both L<Prima> and L<Wx>, however, and will try to
		865	load L<POE> when detecting them, in the hope that POE will pick them up,
		866	in which case everything will be automatic.
		867
		868	=back
		869
		870	=head1 GLOBAL VARIABLES AND FUNCTIONS
		871
		872	These are not normally required to use AnyEvent, but can be useful to
		873	write AnyEvent extension modules.
		874
		875	=over 4
		876
192	=item $AnyEvent::MODEL	877	=item $AnyEvent::MODEL
193		878
194	Contains C<undef> until the first watcher is being created. Then it	879	Contains C<undef> until the first watcher is being created, before the
		880	backend has been autodetected.
		881
195	contains the event model that is being used, which is the name of the	882	Afterwards it contains the event model that is being used, which is the
196	Perl class implementing the model. This class is usually one of the	883	name of the Perl class implementing the model. This class is usually one
197	C<AnyEvent::Impl:xxx> modules, but can be any other class in the case	884	of the C<AnyEvent::Impl:xxx> modules, but can be any other class in the
198	AnyEvent has been extended at runtime (e.g. in I<rxvt-unicode>).	885	case AnyEvent has been extended at runtime (e.g. in I<rxvt-unicode> it
199		886	will be C<urxvt::anyevent>).
200	The known classes so far are:
201
202	AnyEvent::Impl::CoroEV based on Coro::EV, best choice.
203	AnyEvent::Impl::EV based on EV (an interface to libev, also best choice).
204	AnyEvent::Impl::Coro based on Coro::Event, second best choice.
205	AnyEvent::Impl::Event based on Event, also second best choice :)
206	AnyEvent::Impl::Glib based on Glib, second-best choice.
207	AnyEvent::Impl::Tk based on Tk, very bad choice.
208	AnyEvent::Impl::Perl pure-perl implementation, inefficient.
209		887
210	=item AnyEvent::detect	888	=item AnyEvent::detect
211		889
212	Returns C<$AnyEvent::MODEL>, forcing autodetection of the event model if	890	Returns C<$AnyEvent::MODEL>, forcing autodetection of the event model
213	necessary. You should only call this function right before you would have	891	if necessary. You should only call this function right before you would
214	created an AnyEvent watcher anyway, that is, very late at runtime.	892	have created an AnyEvent watcher anyway, that is, as late as possible at
		893	runtime, and not e.g. while initialising of your module.
		894
		895	If you need to do some initialisation before AnyEvent watchers are
		896	created, use C<post_detect>.
		897
		898	=item $guard = AnyEvent::post_detect { BLOCK }
		899
		900	Arranges for the code block to be executed as soon as the event model is
		901	autodetected (or immediately if this has already happened).
		902
		903	The block will be executed I<after> the actual backend has been detected
		904	(C<$AnyEvent::MODEL> is set), but I<before> any watchers have been
		905	created, so it is possible to e.g. patch C<@AnyEvent::ISA> or do
		906	other initialisations - see the sources of L<AnyEvent::Strict> or
		907	L<AnyEvent::AIO> to see how this is used.
		908
		909	The most common usage is to create some global watchers, without forcing
		910	event module detection too early, for example, L<AnyEvent::AIO> creates
		911	and installs the global L<IO::AIO> watcher in a C<post_detect> block to
		912	avoid autodetecting the event module at load time.
		913
		914	If called in scalar or list context, then it creates and returns an object
		915	that automatically removes the callback again when it is destroyed (or
		916	C<undef> when the hook was immediately executed). See L<AnyEvent::AIO> for
		917	a case where this is useful.
		918
		919	Example: Create a watcher for the IO::AIO module and store it in
		920	C<$WATCHER>. Only do so after the event loop is initialised, though.
		921
		922	our WATCHER;
		923
		924	my $guard = AnyEvent::post_detect {
		925	$WATCHER = AnyEvent->io (fh => IO::AIO::poll_fileno, poll => 'r', cb => \&IO::AIO::poll_cb);
		926	};
		927
		928	# the ||= is important in case post_detect immediately runs the block,
		929	# as to not clobber the newly-created watcher. assigning both watcher and
		930	# post_detect guard to the same variable has the advantage of users being
		931	# able to just C<undef $WATCHER> if the watcher causes them grief.
		932
		933	$WATCHER ||= $guard;
		934
		935	=item @AnyEvent::post_detect
		936
		937	If there are any code references in this array (you can C<push> to it
		938	before or after loading AnyEvent), then they will called directly after
		939	the event loop has been chosen.
		940
		941	You should check C<$AnyEvent::MODEL> before adding to this array, though:
		942	if it is defined then the event loop has already been detected, and the
		943	array will be ignored.
		944
		945	Best use C<AnyEvent::post_detect { BLOCK }> when your application allows
		946	it,as it takes care of these details.
		947
		948	This variable is mainly useful for modules that can do something useful
		949	when AnyEvent is used and thus want to know when it is initialised, but do
		950	not need to even load it by default. This array provides the means to hook
		951	into AnyEvent passively, without loading it.
215		952
216	=back	953	=back
217		954
218	=head1 WHAT TO DO IN A MODULE	955	=head1 WHAT TO DO IN A MODULE
219		956
220	As a module author, you should "use AnyEvent" and call AnyEvent methods	957	As a module author, you should C<use AnyEvent> and call AnyEvent methods
221	freely, but you should not load a specific event module or rely on it.	958	freely, but you should not load a specific event module or rely on it.
222		959
223	Be careful when you create watchers in the module body - Anyevent will	960	Be careful when you create watchers in the module body - AnyEvent will
224	decide which event module to use as soon as the first method is called, so	961	decide which event module to use as soon as the first method is called, so
225	by calling AnyEvent in your module body you force the user of your module	962	by calling AnyEvent in your module body you force the user of your module
226	to load the event module first.	963	to load the event module first.
227		964
		965	Never call C<< ->recv >> on a condition variable unless you I<know> that
		966	the C<< ->send >> method has been called on it already. This is
		967	because it will stall the whole program, and the whole point of using
		968	events is to stay interactive.
		969
		970	It is fine, however, to call C<< ->recv >> when the user of your module
		971	requests it (i.e. if you create a http request object ad have a method
		972	called C<results> that returns the results, it should call C<< ->recv >>
		973	freely, as the user of your module knows what she is doing. always).
		974
228	=head1 WHAT TO DO IN THE MAIN PROGRAM	975	=head1 WHAT TO DO IN THE MAIN PROGRAM
229		976
230	There will always be a single main program - the only place that should	977	There will always be a single main program - the only place that should
231	dictate which event model to use.	978	dictate which event model to use.
232		979
233	If it doesn't care, it can just "use AnyEvent" and use it itself, or not	980	If it doesn't care, it can just "use AnyEvent" and use it itself, or not
234	do anything special and let AnyEvent decide which implementation to chose.	981	do anything special (it does not need to be event-based) and let AnyEvent
		982	decide which implementation to chose if some module relies on it.
235		983
236	If the main program relies on a specific event model (for example, in Gtk2	984	If the main program relies on a specific event model - for example, in
237	programs you have to rely on either Glib or Glib::Event), you should load	985	Gtk2 programs you have to rely on the Glib module - you should load the
238	it before loading AnyEvent or any module that uses it, generally, as early	986	event module before loading AnyEvent or any module that uses it: generally
239	as possible. The reason is that modules might create watchers when they	987	speaking, you should load it as early as possible. The reason is that
240	are loaded, and AnyEvent will decide on the event model to use as soon as	988	modules might create watchers when they are loaded, and AnyEvent will
241	it creates watchers, and it might chose the wrong one unless you load the	989	decide on the event model to use as soon as it creates watchers, and it
242	correct one yourself.	990	might chose the wrong one unless you load the correct one yourself.
243		991
244	You can chose to use a rather inefficient pure-perl implementation by	992	You can chose to use a pure-perl implementation by loading the
245	loading the C<AnyEvent::Impl::Perl> module, but letting AnyEvent chose is	993	C<AnyEvent::Impl::Perl> module, which gives you similar behaviour
246	generally better.	994	everywhere, but letting AnyEvent chose the model is generally better.
		995
		996	=head2 MAINLOOP EMULATION
		997
		998	Sometimes (often for short test scripts, or even standalone programs who
		999	only want to use AnyEvent), you do not want to run a specific event loop.
		1000
		1001	In that case, you can use a condition variable like this:
		1002
		1003	AnyEvent->condvar->recv;
		1004
		1005	This has the effect of entering the event loop and looping forever.
		1006
		1007	Note that usually your program has some exit condition, in which case
		1008	it is better to use the "traditional" approach of storing a condition
		1009	variable somewhere, waiting for it, and sending it when the program should
		1010	exit cleanly.
		1011
		1012
		1013	=head1 OTHER MODULES
		1014
		1015	The following is a non-exhaustive list of additional modules that use
		1016	AnyEvent as a client and can therefore be mixed easily with other AnyEvent
		1017	modules and other event loops in the same program. Some of the modules
		1018	come with AnyEvent, most are available via CPAN.
		1019
		1020	=over 4
		1021
		1022	=item L<AnyEvent::Util>
		1023
		1024	Contains various utility functions that replace often-used but blocking
		1025	functions such as C<inet_aton> by event-/callback-based versions.
		1026
		1027	=item L<AnyEvent::Socket>
		1028
		1029	Provides various utility functions for (internet protocol) sockets,
		1030	addresses and name resolution. Also functions to create non-blocking tcp
		1031	connections or tcp servers, with IPv6 and SRV record support and more.
		1032
		1033	=item L<AnyEvent::Handle>
		1034
		1035	Provide read and write buffers, manages watchers for reads and writes,
		1036	supports raw and formatted I/O, I/O queued and fully transparent and
		1037	non-blocking SSL/TLS (via L<AnyEvent::TLS>.
		1038
		1039	=item L<AnyEvent::DNS>
		1040
		1041	Provides rich asynchronous DNS resolver capabilities.
		1042
		1043	=item L<AnyEvent::HTTP>
		1044
		1045	A simple-to-use HTTP library that is capable of making a lot of concurrent
		1046	HTTP requests.
		1047
		1048	=item L<AnyEvent::HTTPD>
		1049
		1050	Provides a simple web application server framework.
		1051
		1052	=item L<AnyEvent::FastPing>
		1053
		1054	The fastest ping in the west.
		1055
		1056	=item L<AnyEvent::DBI>
		1057
		1058	Executes L<DBI> requests asynchronously in a proxy process.
		1059
		1060	=item L<AnyEvent::AIO>
		1061
		1062	Truly asynchronous I/O, should be in the toolbox of every event
		1063	programmer. AnyEvent::AIO transparently fuses L<IO::AIO> and AnyEvent
		1064	together.
		1065
		1066	=item L<AnyEvent::BDB>
		1067
		1068	Truly asynchronous Berkeley DB access. AnyEvent::BDB transparently fuses
		1069	L<BDB> and AnyEvent together.
		1070
		1071	=item L<AnyEvent::GPSD>
		1072
		1073	A non-blocking interface to gpsd, a daemon delivering GPS information.
		1074
		1075	=item L<AnyEvent::IRC>
		1076
		1077	AnyEvent based IRC client module family (replacing the older Net::IRC3).
		1078
		1079	=item L<AnyEvent::XMPP>
		1080
		1081	AnyEvent based XMPP (Jabber protocol) module family (replacing the older
		1082	Net::XMPP2>.
		1083
		1084	=item L<AnyEvent::IGS>
		1085
		1086	A non-blocking interface to the Internet Go Server protocol (used by
		1087	L<App::IGS>).
		1088
		1089	=item L<Net::FCP>
		1090
		1091	AnyEvent-based implementation of the Freenet Client Protocol, birthplace
		1092	of AnyEvent.
		1093
		1094	=item L<Event::ExecFlow>
		1095
		1096	High level API for event-based execution flow control.
		1097
		1098	=item L<Coro>
		1099
		1100	Has special support for AnyEvent via L<Coro::AnyEvent>.
		1101
		1102	=back
247		1103
248	=cut	1104	=cut
249		1105
250	package AnyEvent;	1106	package AnyEvent;
251		1107
		1108	# basically a tuned-down version of common::sense
		1109	sub common_sense {
252	no warnings;	1110	# no warnings
253	use strict;	1111	${^WARNING_BITS} ^= ${^WARNING_BITS};
		1112	# use strict vars subs
		1113	$^H |= 0x00000600;
		1114	}
254		1115
		1116	BEGIN { AnyEvent::common_sense }
		1117
255	use Carp;	1118	use Carp ();
256		1119
257	our $VERSION = '2.6';	1120	our $VERSION = 4.92;
258	our $MODEL;	1121	our $MODEL;
259		1122
260	our $AUTOLOAD;	1123	our $AUTOLOAD;
261	our @ISA;	1124	our @ISA;
262		1125
263	our $verbose = $ENV{PERL_ANYEVENT_VERBOSE}*1;
264
265	our @REGISTRY;	1126	our @REGISTRY;
266		1127
		1128	our $WIN32;
		1129
		1130	our $VERBOSE;
		1131
		1132	BEGIN {
		1133	eval "sub WIN32(){ " . (($^O =~ /mswin32/i)*1) ." }";
		1134	eval "sub TAINT(){ " . (${^TAINT}*1) . " }";
		1135
		1136	delete @ENV{grep /^PERL_ANYEVENT_/, keys %ENV}
		1137	if ${^TAINT};
		1138
		1139	$VERBOSE = $ENV{PERL_ANYEVENT_VERBOSE}*1;
		1140
		1141	}
		1142
		1143	our $MAX_SIGNAL_LATENCY = 10;
		1144
		1145	our %PROTOCOL; # (ipv4|ipv6) => (1|2), higher numbers are preferred
		1146
		1147	{
		1148	my $idx;
		1149	$PROTOCOL{$_} = ++$idx
		1150	for reverse split /\s*,\s*/,
		1151	$ENV{PERL_ANYEVENT_PROTOCOLS} || "ipv4,ipv6";
		1152	}
		1153
267	my @models = (	1154	my @models = (
268	[Coro::EV:: => AnyEvent::Impl::CoroEV::],
269	[EV:: => AnyEvent::Impl::EV::],	1155	[EV:: => AnyEvent::Impl::EV:: , 1],
270	[Coro::Event:: => AnyEvent::Impl::Coro::],	1156	[AnyEvent::Impl::Perl:: => AnyEvent::Impl::Perl:: , 1],
		1157	# everything below here will not (normally) be autoprobed
		1158	# as the pureperl backend should work everywhere
		1159	# and is usually faster
271	[Event:: => AnyEvent::Impl::Event::],	1160	[Event:: => AnyEvent::Impl::Event::, 1],
272	[Glib:: => AnyEvent::Impl::Glib::],	1161	[Glib:: => AnyEvent::Impl::Glib:: , 1], # becomes extremely slow with many watchers
		1162	[Event::Lib:: => AnyEvent::Impl::EventLib::], # too buggy
		1163	[Irssi:: => AnyEvent::Impl::Irssi::], # Irssi has a bogus "Event" package
		1164	[Tk:: => AnyEvent::Impl::Tk::], # crashes with many handles
		1165	[Qt:: => AnyEvent::Impl::Qt::], # requires special main program
		1166	[POE::Kernel:: => AnyEvent::Impl::POE::], # lasciate ogni speranza
273	[Tk:: => AnyEvent::Impl::Tk::],	1167	[Wx:: => AnyEvent::Impl::POE::],
274	[AnyEvent::Impl::Perl:: => AnyEvent::Impl::Perl::],	1168	[Prima:: => AnyEvent::Impl::POE::],
		1169	# IO::Async is just too broken - we would need workarounds for its
		1170	# byzantine signal and broken child handling, among others.
		1171	# IO::Async is rather hard to detect, as it doesn't have any
		1172	# obvious default class.
		1173	[IO::Async:: => AnyEvent::Impl::IOAsync::], # requires special main program
		1174	[IO::Async::Loop:: => AnyEvent::Impl::IOAsync::], # requires special main program
		1175	[IO::Async::Notifier:: => AnyEvent::Impl::IOAsync::], # requires special main program
		1176	[AnyEvent::Impl::IOAsync:: => AnyEvent::Impl::IOAsync::], # requires special main program
275	);	1177	);
276		1178
277	our %method = map +($_ => 1), qw(io timer condvar broadcast wait signal one_event DESTROY);	1179	our %method = map +($_ => 1),
		1180	qw(io timer time now now_update signal child idle condvar one_event DESTROY);
		1181
		1182	our @post_detect;
		1183
		1184	sub post_detect(&) {
		1185	my ($cb) = @_;
		1186
		1187	if ($MODEL) {
		1188	$cb->();
		1189
		1190	undef
		1191	} else {
		1192	push @post_detect, $cb;
		1193
		1194	defined wantarray
		1195	? bless \$cb, "AnyEvent::Util::postdetect"
		1196	: ()
		1197	}
		1198	}
		1199
		1200	sub AnyEvent::Util::postdetect::DESTROY {
		1201	@post_detect = grep $_ != ${$_[0]}, @post_detect;
		1202	}
278		1203
279	sub detect() {	1204	sub detect() {
280	unless ($MODEL) {	1205	unless ($MODEL) {
281	no strict 'refs';	1206	local $SIG{__DIE__};
		1207
		1208	if ($ENV{PERL_ANYEVENT_MODEL} =~ /^([a-zA-Z]+)$/) {
		1209	my $model = "AnyEvent::Impl::$1";
		1210	if (eval "require $model") {
		1211	$MODEL = $model;
		1212	warn "AnyEvent: loaded model '$model' (forced by \$ENV{PERL_ANYEVENT_MODEL}), using it.\n" if $VERBOSE >= 2;
		1213	} else {
		1214	warn "AnyEvent: unable to load model '$model' (from \$ENV{PERL_ANYEVENT_MODEL}):\n$@" if $VERBOSE;
		1215	}
		1216	}
282		1217
283	# check for already loaded models	1218	# check for already loaded models
		1219	unless ($MODEL) {
284	for (@REGISTRY, @models) {	1220	for (@REGISTRY, @models) {
285	my ($package, $model) = @$_;	1221	my ($package, $model) = @$_;
286	if (${"$package\::VERSION"} > 0) {	1222	if (${"$package\::VERSION"} > 0) {
287	if (eval "require $model") {	1223	if (eval "require $model") {
288	$MODEL = $model;	1224	$MODEL = $model;
289	warn "AnyEvent: found model '$model', using it.\n" if $verbose > 1;	1225	warn "AnyEvent: autodetected model '$model', using it.\n" if $VERBOSE >= 2;
290	last;	1226	last;
		1227	}
291	}	1228	}
292	}	1229	}
		1230
		1231	unless ($MODEL) {
		1232	# try to autoload a model
		1233	for (@REGISTRY, @models) {
		1234	my ($package, $model, $autoload) = @$_;
		1235	if (
		1236	$autoload
		1237	and eval "require $package"
		1238	and ${"$package\::VERSION"} > 0
		1239	and eval "require $model"
		1240	) {
		1241	$MODEL = $model;
		1242	warn "AnyEvent: autoloaded model '$model', using it.\n" if $VERBOSE >= 2;
		1243	last;
		1244	}
		1245	}
		1246
		1247	$MODEL
		1248	or die "No event module selected for AnyEvent and autodetect failed. Install any one of these modules: EV, Event or Glib.\n";
		1249	}
293	}	1250	}
294		1251
295	unless ($MODEL) {	1252	push @{"$MODEL\::ISA"}, "AnyEvent::Base";
296	# try to load a model
297
298	for (@REGISTRY, @models) {
299	my ($package, $model) = @$_;
300	if (eval "require $package"
301	and ${"$package\::VERSION"} > 0
302	and eval "require $model") {
303	$MODEL = $model;
304	warn "AnyEvent: autoprobed and loaded model '$model', using it.\n" if $verbose > 1;
305	last;
306	}
307	}
308
309	$MODEL
310	or die "No event module selected for AnyEvent and autodetect failed. Install any one of these modules: EV (or Coro+EV), Event (or Coro+Event), Glib or Tk.";
311	}
312		1253
313	unshift @ISA, $MODEL;	1254	unshift @ISA, $MODEL;
314	push @{"$MODEL\::ISA"}, "AnyEvent::Base";	1255
		1256	require AnyEvent::Strict if $ENV{PERL_ANYEVENT_STRICT};
		1257
		1258	(shift @post_detect)->() while @post_detect;
315	}	1259	}
316		1260
317	$MODEL	1261	$MODEL
318	}	1262	}
319		1263
320	sub AUTOLOAD {	1264	sub AUTOLOAD {
321	(my $func = $AUTOLOAD) =~ s/.*://;	1265	(my $func = $AUTOLOAD) =~ s/.*://;
322		1266
323	$method{$func}	1267	$method{$func}
324	or croak "$func: not a valid method for AnyEvent objects";	1268	or Carp::croak "$func: not a valid method for AnyEvent objects";
325		1269
326	detect unless $MODEL;	1270	detect unless $MODEL;
327		1271
328	my $class = shift;	1272	my $class = shift;
329	$class->$func (@_);	1273	$class->$func (@_);
330	}	1274	}
331		1275
		1276	# utility function to dup a filehandle. this is used by many backends
		1277	# to support binding more than one watcher per filehandle (they usually
		1278	# allow only one watcher per fd, so we dup it to get a different one).
		1279	sub _dupfh($$;$$) {
		1280	my ($poll, $fh, $r, $w) = @_;
		1281
		1282	# cygwin requires the fh mode to be matching, unix doesn't
		1283	my ($rw, $mode) = $poll eq "r" ? ($r, "<&") : ($w, ">&");
		1284
		1285	open my $fh2, $mode, $fh
		1286	or die "AnyEvent->io: cannot dup() filehandle in mode '$poll': $!,";
		1287
		1288	# we assume CLOEXEC is already set by perl in all important cases
		1289
		1290	($fh2, $rw)
		1291	}
		1292
		1293	#############################################################################
		1294	# "new" API, currently only emulation of it
		1295	#############################################################################
		1296
		1297	package AE;
		1298
		1299	our $VERSION = $AnyEvent::VERSION;
		1300
		1301	sub io($$$) {
		1302	AnyEvent->io (fh => $_[0], poll => $_[1] ? "w" : "r", cb => $_[2])
		1303	}
		1304
		1305	sub timer($$$) {
		1306	AnyEvent->timer (after => $_[0], interval => $_[1], cb => $_[2])
		1307	}
		1308
		1309	sub signal($$) {
		1310	AnyEvent->signal (signal => $_[0], cb => $_[1])
		1311	}
		1312
		1313	sub child($$) {
		1314	AnyEvent->child (pid => $_[0], cb => $_[1])
		1315	}
		1316
		1317	sub idle($) {
		1318	AnyEvent->idle (cb => $_[0])
		1319	}
		1320
		1321	sub cv(;&) {
		1322	AnyEvent->condvar (@_ ? (cb => $_[0]) : ())
		1323	}
		1324
		1325	sub now() {
		1326	AnyEvent->now
		1327	}
		1328
		1329	sub now_update() {
		1330	AnyEvent->now_update
		1331	}
		1332
		1333	sub time() {
		1334	AnyEvent->time
		1335	}
		1336
332	package AnyEvent::Base;	1337	package AnyEvent::Base;
333		1338
		1339	# default implementations for many methods
		1340
		1341	sub _time {
		1342	# probe for availability of Time::HiRes
		1343	if (eval "use Time::HiRes (); Time::HiRes::time (); 1") {
		1344	warn "AnyEvent: using Time::HiRes for sub-second timing accuracy.\n" if $VERBOSE >= 8;
		1345	*_time = \&Time::HiRes::time;
		1346	# if (eval "use POSIX (); (POSIX::times())...
		1347	} else {
		1348	warn "AnyEvent: using built-in time(), WARNING, no sub-second resolution!\n" if $VERBOSE;
		1349	*_time = sub { time }; # epic fail
		1350	}
		1351
		1352	&_time
		1353	}
		1354
		1355	sub time { _time }
		1356	sub now { _time }
		1357	sub now_update { }
		1358
334	# default implementation for ->condvar, ->wait, ->broadcast	1359	# default implementation for ->condvar
335		1360
336	sub condvar {	1361	sub condvar {
337	bless \my $flag, "AnyEvent::Base::CondVar"	1362	bless { @_ == 3 ? (_ae_cb => $_[2]) : () }, "AnyEvent::CondVar"
338	}
339
340	sub AnyEvent::Base::CondVar::broadcast {
341	${$_[0]}++;
342	}
343
344	sub AnyEvent::Base::CondVar::wait {
345	AnyEvent->one_event while !${$_[0]};
346	}	1363	}
347		1364
348	# default implementation for ->signal	1365	# default implementation for ->signal
349		1366
350	our %SIG_CB;	1367	our $HAVE_ASYNC_INTERRUPT;
		1368
		1369	sub _have_async_interrupt() {
		1370	$HAVE_ASYNC_INTERRUPT = 1*(!$ENV{PERL_ANYEVENT_AVOID_ASYNC_INTERRUPT}
		1371	&& eval "use Async::Interrupt 1.0 (); 1")
		1372	unless defined $HAVE_ASYNC_INTERRUPT;
		1373
		1374	$HAVE_ASYNC_INTERRUPT
		1375	}
		1376
		1377	our ($SIGPIPE_R, $SIGPIPE_W, %SIG_CB, %SIG_EV, $SIG_IO);
		1378	our (%SIG_ASY, %SIG_ASY_W);
		1379	our ($SIG_COUNT, $SIG_TW);
		1380
		1381	sub _signal_exec {
		1382	$HAVE_ASYNC_INTERRUPT
		1383	? $SIGPIPE_R->drain
		1384	: sysread $SIGPIPE_R, my $dummy, 9;
		1385
		1386	while (%SIG_EV) {
		1387	for (keys %SIG_EV) {
		1388	delete $SIG_EV{$_};
		1389	$_->() for values %{ $SIG_CB{$_} || {} };
		1390	}
		1391	}
		1392	}
		1393
		1394	# install a dummy wakeup watcher to reduce signal catching latency
		1395	sub _sig_add() {
		1396	unless ($SIG_COUNT++) {
		1397	# try to align timer on a full-second boundary, if possible
		1398	my $NOW = AE::now;
		1399
		1400	$SIG_TW = AE::timer
		1401	$MAX_SIGNAL_LATENCY - ($NOW - int $NOW),
		1402	$MAX_SIGNAL_LATENCY,
		1403	sub { } # just for the PERL_ASYNC_CHECK
		1404	;
		1405	}
		1406	}
		1407
		1408	sub _sig_del {
		1409	undef $SIG_TW
		1410	unless --$SIG_COUNT;
		1411	}
		1412
		1413	our $_sig_name_init; $_sig_name_init = sub {
		1414	eval q{ # poor man's autoloading
		1415	undef $_sig_name_init;
		1416
		1417	if (_have_async_interrupt) {
		1418	*sig2num = \&Async::Interrupt::sig2num;
		1419	*sig2name = \&Async::Interrupt::sig2name;
		1420	} else {
		1421	require Config;
		1422
		1423	my %signame2num;
		1424	@signame2num{ split ' ', $Config::Config{sig_name} }
		1425	= split ' ', $Config::Config{sig_num};
		1426
		1427	my @signum2name;
		1428	@signum2name[values %signame2num] = keys %signame2num;
		1429
		1430	*sig2num = sub($) {
		1431	$_[0] > 0 ? shift : $signame2num{+shift}
		1432	};
		1433	*sig2name = sub ($) {
		1434	$_[0] > 0 ? $signum2name[+shift] : shift
		1435	};
		1436	}
		1437	};
		1438	die if $@;
		1439	};
		1440
		1441	sub sig2num ($) { &$_sig_name_init; &sig2num }
		1442	sub sig2name($) { &$_sig_name_init; &sig2name }
351		1443
352	sub signal {	1444	sub signal {
		1445	eval q{ # poor man's autoloading {}
		1446	# probe for availability of Async::Interrupt
		1447	if (_have_async_interrupt) {
		1448	warn "AnyEvent: using Async::Interrupt for race-free signal handling.\n" if $VERBOSE >= 8;
		1449
		1450	$SIGPIPE_R = new Async::Interrupt::EventPipe;
		1451	$SIG_IO = AE::io $SIGPIPE_R->fileno, 0, \&_signal_exec;
		1452
		1453	} else {
		1454	warn "AnyEvent: using emulated perl signal handling with latency timer.\n" if $VERBOSE >= 8;
		1455
		1456	require Fcntl;
		1457
		1458	if (AnyEvent::WIN32) {
		1459	require AnyEvent::Util;
		1460
		1461	($SIGPIPE_R, $SIGPIPE_W) = AnyEvent::Util::portable_pipe ();
		1462	AnyEvent::Util::fh_nonblocking ($SIGPIPE_R, 1) if $SIGPIPE_R;
		1463	AnyEvent::Util::fh_nonblocking ($SIGPIPE_W, 1) if $SIGPIPE_W; # just in case
		1464	} else {
		1465	pipe $SIGPIPE_R, $SIGPIPE_W;
		1466	fcntl $SIGPIPE_R, &Fcntl::F_SETFL, &Fcntl::O_NONBLOCK if $SIGPIPE_R;
		1467	fcntl $SIGPIPE_W, &Fcntl::F_SETFL, &Fcntl::O_NONBLOCK if $SIGPIPE_W; # just in case
		1468
		1469	# not strictly required, as $^F is normally 2, but let's make sure...
		1470	fcntl $SIGPIPE_R, &Fcntl::F_SETFD, &Fcntl::FD_CLOEXEC;
		1471	fcntl $SIGPIPE_W, &Fcntl::F_SETFD, &Fcntl::FD_CLOEXEC;
		1472	}
		1473
		1474	$SIGPIPE_R
		1475	or Carp::croak "AnyEvent: unable to create a signal reporting pipe: $!\n";
		1476
		1477	$SIG_IO = AE::io $SIGPIPE_R, 0, \&_signal_exec;
		1478	}
		1479
		1480	*signal = sub {
353	my (undef, %arg) = @_;	1481	my (undef, %arg) = @_;
354		1482
355	my $signal = uc $arg{signal}	1483	my $signal = uc $arg{signal}
356	or Carp::croak "required option 'signal' is missing";	1484	or Carp::croak "required option 'signal' is missing";
357		1485
		1486	if ($HAVE_ASYNC_INTERRUPT) {
		1487	# async::interrupt
		1488
		1489	$signal = sig2num $signal;
358	$SIG_CB{$signal}{$arg{cb}} = $arg{cb};	1490	$SIG_CB{$signal}{$arg{cb}} = $arg{cb};
		1491
		1492	$SIG_ASY{$signal} ||= new Async::Interrupt
		1493	cb => sub { undef $SIG_EV{$signal} },
		1494	signal => $signal,
		1495	pipe => [$SIGPIPE_R->filenos],
		1496	pipe_autodrain => 0,
		1497	;
		1498
		1499	} else {
		1500	# pure perl
		1501
		1502	# AE::Util has been loaded in signal
		1503	$signal = sig2name $signal;
		1504	$SIG_CB{$signal}{$arg{cb}} = $arg{cb};
		1505
359	$SIG{$signal} ||= sub {	1506	$SIG{$signal} ||= sub {
360	$_->() for values %{ $SIG_CB{$signal} || {} };	1507	local $!;
		1508	syswrite $SIGPIPE_W, "\x00", 1 unless %SIG_EV;
		1509	undef $SIG_EV{$signal};
		1510	};
		1511
		1512	# can't do signal processing without introducing races in pure perl,
		1513	# so limit the signal latency.
		1514	_sig_add;
		1515	}
		1516
		1517	bless [$signal, $arg{cb}], "AnyEvent::Base::signal"
		1518	};
		1519
		1520	*AnyEvent::Base::signal::DESTROY = sub {
		1521	my ($signal, $cb) = @{$_[0]};
		1522
		1523	_sig_del;
		1524
		1525	delete $SIG_CB{$signal}{$cb};
		1526
		1527	$HAVE_ASYNC_INTERRUPT
		1528	? delete $SIG_ASY{$signal}
		1529	: # delete doesn't work with older perls - they then
		1530	# print weird messages, or just unconditionally exit
		1531	# instead of getting the default action.
		1532	undef $SIG{$signal}
		1533	unless keys %{ $SIG_CB{$signal} };
		1534	};
361	};	1535	};
362		1536	die if $@;
363	bless [$signal, $arg{cb}], "AnyEvent::Base::Signal"	1537	&signal
364	}
365
366	sub AnyEvent::Base::Signal::DESTROY {
367	my ($signal, $cb) = @{$_[0]};
368
369	delete $SIG_CB{$signal}{$cb};
370
371	$SIG{$signal} = 'DEFAULT' unless keys %{ $SIG_CB{$signal} };
372	}	1538	}
373		1539
374	# default implementation for ->child	1540	# default implementation for ->child
375		1541
376	our %PID_CB;	1542	our %PID_CB;
377	our $CHLD_W;	1543	our $CHLD_W;
378	our $PID_IDLE;	1544	our $CHLD_DELAY_W;
379	our $WNOHANG;	1545	our $WNOHANG;
380		1546
381	sub _child_wait {	1547	sub _emit_childstatus($$) {
382	while (0 <= (my $pid = waitpid -1, $WNOHANG)) {	1548	my (undef, $rpid, $rstatus) = @_;
		1549
		1550	$_->($rpid, $rstatus)
383	$_->($pid, $?) for (values %{ $PID_CB{$pid} || {} }),	1551	for values %{ $PID_CB{$rpid} || {} },
384	(values %{ $PID_CB{0} || {} });	1552	values %{ $PID_CB{0} || {} };
385	}	1553	}
386		1554
387	undef $PID_IDLE;	1555	sub _sigchld {
		1556	my $pid;
		1557
		1558	AnyEvent->_emit_childstatus ($pid, $?)
		1559	while ($pid = waitpid -1, $WNOHANG) > 0;
388	}	1560	}
389		1561
390	sub child {	1562	sub child {
391	my (undef, %arg) = @_;	1563	my (undef, %arg) = @_;
392		1564
393	defined (my $pid = $arg{pid} + 0)	1565	defined (my $pid = $arg{pid} + 0)
394	or Carp::croak "required option 'pid' is missing";	1566	or Carp::croak "required option 'pid' is missing";
395		1567
396	$PID_CB{$pid}{$arg{cb}} = $arg{cb};	1568	$PID_CB{$pid}{$arg{cb}} = $arg{cb};
397		1569
398	unless ($WNOHANG) {	1570	# WNOHANG is almost cetrainly 1 everywhere
399	$WNOHANG = eval { require POSIX; &POSIX::WNOHANG } || 1;	1571	$WNOHANG ||= $^O =~ /^(?:openbsd|netbsd|linux|freebsd|cygwin|MSWin32)$/
		1572	? 1
		1573	: eval { local $SIG{__DIE__}; require POSIX; &POSIX::WNOHANG } || 1;
		1574
		1575	unless ($CHLD_W) {
		1576	$CHLD_W = AE::signal CHLD => \&_sigchld;
		1577	# child could be a zombie already, so make at least one round
		1578	&_sigchld;
400	}	1579	}
401		1580
402	unless ($CHLD_W) {
403	$CHLD_W = AnyEvent->signal (signal => 'CHLD', cb => \&_child_wait);
404	# child could be a zombie already
405	$PID_IDLE ||= AnyEvent->timer (after => 0, cb => \&_child_wait);
406	}
407
408	bless [$pid, $arg{cb}], "AnyEvent::Base::Child"	1581	bless [$pid, $arg{cb}], "AnyEvent::Base::child"
409	}	1582	}
410		1583
411	sub AnyEvent::Base::Child::DESTROY {	1584	sub AnyEvent::Base::child::DESTROY {
412	my ($pid, $cb) = @{$_[0]};	1585	my ($pid, $cb) = @{$_[0]};
413		1586
414	delete $PID_CB{$pid}{$cb};	1587	delete $PID_CB{$pid}{$cb};
415	delete $PID_CB{$pid} unless keys %{ $PID_CB{$pid} };	1588	delete $PID_CB{$pid} unless keys %{ $PID_CB{$pid} };
416		1589
417	undef $CHLD_W unless keys %PID_CB;	1590	undef $CHLD_W unless keys %PID_CB;
418	}	1591	}
419		1592
		1593	# idle emulation is done by simply using a timer, regardless
		1594	# of whether the process is idle or not, and not letting
		1595	# the callback use more than 50% of the time.
		1596	sub idle {
		1597	my (undef, %arg) = @_;
		1598
		1599	my ($cb, $w, $rcb) = $arg{cb};
		1600
		1601	$rcb = sub {
		1602	if ($cb) {
		1603	$w = _time;
		1604	&$cb;
		1605	$w = _time - $w;
		1606
		1607	# never use more then 50% of the time for the idle watcher,
		1608	# within some limits
		1609	$w = 0.0001 if $w < 0.0001;
		1610	$w = 5 if $w > 5;
		1611
		1612	$w = AE::timer $w, 0, $rcb;
		1613	} else {
		1614	# clean up...
		1615	undef $w;
		1616	undef $rcb;
		1617	}
		1618	};
		1619
		1620	$w = AE::timer 0.05, 0, $rcb;
		1621
		1622	bless \\$cb, "AnyEvent::Base::idle"
		1623	}
		1624
		1625	sub AnyEvent::Base::idle::DESTROY {
		1626	undef $${$_[0]};
		1627	}
		1628
		1629	package AnyEvent::CondVar;
		1630
		1631	our @ISA = AnyEvent::CondVar::Base::;
		1632
		1633	package AnyEvent::CondVar::Base;
		1634
		1635	#use overload
		1636	# '&{}' => sub { my $self = shift; sub { $self->send (@_) } },
		1637	# fallback => 1;
		1638
		1639	# save 300+ kilobytes by dirtily hardcoding overloading
		1640	${"AnyEvent::CondVar::Base::OVERLOAD"}{dummy}++; # Register with magic by touching.
		1641	*{'AnyEvent::CondVar::Base::()'} = sub { }; # "Make it findable via fetchmethod."
		1642	*{'AnyEvent::CondVar::Base::(&{}'} = sub { my $self = shift; sub { $self->send (@_) } }; # &{}
		1643	${'AnyEvent::CondVar::Base::()'} = 1; # fallback
		1644
		1645	our $WAITING;
		1646
		1647	sub _send {
		1648	# nop
		1649	}
		1650
		1651	sub send {
		1652	my $cv = shift;
		1653	$cv->{_ae_sent} = [@_];
		1654	(delete $cv->{_ae_cb})->($cv) if $cv->{_ae_cb};
		1655	$cv->_send;
		1656	}
		1657
		1658	sub croak {
		1659	$_[0]{_ae_croak} = $_[1];
		1660	$_[0]->send;
		1661	}
		1662
		1663	sub ready {
		1664	$_[0]{_ae_sent}
		1665	}
		1666
		1667	sub _wait {
		1668	$WAITING
		1669	and !$_[0]{_ae_sent}
		1670	and Carp::croak "AnyEvent::CondVar: recursive blocking wait detected";
		1671
		1672	local $WAITING = 1;
		1673	AnyEvent->one_event while !$_[0]{_ae_sent};
		1674	}
		1675
		1676	sub recv {
		1677	$_[0]->_wait;
		1678
		1679	Carp::croak $_[0]{_ae_croak} if $_[0]{_ae_croak};
		1680	wantarray ? @{ $_[0]{_ae_sent} } : $_[0]{_ae_sent}[0]
		1681	}
		1682
		1683	sub cb {
		1684	my $cv = shift;
		1685
		1686	@_
		1687	and $cv->{_ae_cb} = shift
		1688	and $cv->{_ae_sent}
		1689	and (delete $cv->{_ae_cb})->($cv);
		1690
		1691	$cv->{_ae_cb}
		1692	}
		1693
		1694	sub begin {
		1695	++$_[0]{_ae_counter};
		1696	$_[0]{_ae_end_cb} = $_[1] if @_ > 1;
		1697	}
		1698
		1699	sub end {
		1700	return if --$_[0]{_ae_counter};
		1701	&{ $_[0]{_ae_end_cb} || sub { $_[0]->send } };
		1702	}
		1703
		1704	# undocumented/compatibility with pre-3.4
		1705	*broadcast = \&send;
		1706	*wait = \&_wait;
		1707
		1708	=head1 ERROR AND EXCEPTION HANDLING
		1709
		1710	In general, AnyEvent does not do any error handling - it relies on the
		1711	caller to do that if required. The L<AnyEvent::Strict> module (see also
		1712	the C<PERL_ANYEVENT_STRICT> environment variable, below) provides strict
		1713	checking of all AnyEvent methods, however, which is highly useful during
		1714	development.
		1715
		1716	As for exception handling (i.e. runtime errors and exceptions thrown while
		1717	executing a callback), this is not only highly event-loop specific, but
		1718	also not in any way wrapped by this module, as this is the job of the main
		1719	program.
		1720
		1721	The pure perl event loop simply re-throws the exception (usually
		1722	within C<< condvar->recv >>), the L<Event> and L<EV> modules call C<<
		1723	$Event/EV::DIED->() >>, L<Glib> uses C<< install_exception_handler >> and
		1724	so on.
		1725
		1726	=head1 ENVIRONMENT VARIABLES
		1727
		1728	The following environment variables are used by this module or its
		1729	submodules.
		1730
		1731	Note that AnyEvent will remove I<all> environment variables starting with
		1732	C<PERL_ANYEVENT_> from C<%ENV> when it is loaded while taint mode is
		1733	enabled.
		1734
		1735	=over 4
		1736
		1737	=item C<PERL_ANYEVENT_VERBOSE>
		1738
		1739	By default, AnyEvent will be completely silent except in fatal
		1740	conditions. You can set this environment variable to make AnyEvent more
		1741	talkative.
		1742
		1743	When set to C<1> or higher, causes AnyEvent to warn about unexpected
		1744	conditions, such as not being able to load the event model specified by
		1745	C<PERL_ANYEVENT_MODEL>.
		1746
		1747	When set to C<2> or higher, cause AnyEvent to report to STDERR which event
		1748	model it chooses.
		1749
		1750	When set to C<8> or higher, then AnyEvent will report extra information on
		1751	which optional modules it loads and how it implements certain features.
		1752
		1753	=item C<PERL_ANYEVENT_STRICT>
		1754
		1755	AnyEvent does not do much argument checking by default, as thorough
		1756	argument checking is very costly. Setting this variable to a true value
		1757	will cause AnyEvent to load C<AnyEvent::Strict> and then to thoroughly
		1758	check the arguments passed to most method calls. If it finds any problems,
		1759	it will croak.
		1760
		1761	In other words, enables "strict" mode.
		1762
		1763	Unlike C<use strict> (or it's modern cousin, C<< use L<common::sense>
		1764	>>, it is definitely recommended to keep it off in production. Keeping
		1765	C<PERL_ANYEVENT_STRICT=1> in your environment while developing programs
		1766	can be very useful, however.
		1767
		1768	=item C<PERL_ANYEVENT_MODEL>
		1769
		1770	This can be used to specify the event model to be used by AnyEvent, before
		1771	auto detection and -probing kicks in. It must be a string consisting
		1772	entirely of ASCII letters. The string C<AnyEvent::Impl::> gets prepended
		1773	and the resulting module name is loaded and if the load was successful,
		1774	used as event model. If it fails to load AnyEvent will proceed with
		1775	auto detection and -probing.
		1776
		1777	This functionality might change in future versions.
		1778
		1779	For example, to force the pure perl model (L<AnyEvent::Impl::Perl>) you
		1780	could start your program like this:
		1781
		1782	PERL_ANYEVENT_MODEL=Perl perl ...
		1783
		1784	=item C<PERL_ANYEVENT_PROTOCOLS>
		1785
		1786	Used by both L<AnyEvent::DNS> and L<AnyEvent::Socket> to determine preferences
		1787	for IPv4 or IPv6. The default is unspecified (and might change, or be the result
		1788	of auto probing).
		1789
		1790	Must be set to a comma-separated list of protocols or address families,
		1791	current supported: C<ipv4> and C<ipv6>. Only protocols mentioned will be
		1792	used, and preference will be given to protocols mentioned earlier in the
		1793	list.
		1794
		1795	This variable can effectively be used for denial-of-service attacks
		1796	against local programs (e.g. when setuid), although the impact is likely
		1797	small, as the program has to handle conenction and other failures anyways.
		1798
		1799	Examples: C<PERL_ANYEVENT_PROTOCOLS=ipv4,ipv6> - prefer IPv4 over IPv6,
		1800	but support both and try to use both. C<PERL_ANYEVENT_PROTOCOLS=ipv4>
		1801	- only support IPv4, never try to resolve or contact IPv6
		1802	addresses. C<PERL_ANYEVENT_PROTOCOLS=ipv6,ipv4> support either IPv4 or
		1803	IPv6, but prefer IPv6 over IPv4.
		1804
		1805	=item C<PERL_ANYEVENT_EDNS0>
		1806
		1807	Used by L<AnyEvent::DNS> to decide whether to use the EDNS0 extension
		1808	for DNS. This extension is generally useful to reduce DNS traffic, but
		1809	some (broken) firewalls drop such DNS packets, which is why it is off by
		1810	default.
		1811
		1812	Setting this variable to C<1> will cause L<AnyEvent::DNS> to announce
		1813	EDNS0 in its DNS requests.
		1814
		1815	=item C<PERL_ANYEVENT_MAX_FORKS>
		1816
		1817	The maximum number of child processes that C<AnyEvent::Util::fork_call>
		1818	will create in parallel.
		1819
		1820	=item C<PERL_ANYEVENT_MAX_OUTSTANDING_DNS>
		1821
		1822	The default value for the C<max_outstanding> parameter for the default DNS
		1823	resolver - this is the maximum number of parallel DNS requests that are
		1824	sent to the DNS server.
		1825
		1826	=item C<PERL_ANYEVENT_RESOLV_CONF>
		1827
		1828	The file to use instead of F</etc/resolv.conf> (or OS-specific
		1829	configuration) in the default resolver. When set to the empty string, no
		1830	default config will be used.
		1831
		1832	=item C<PERL_ANYEVENT_CA_FILE>, C<PERL_ANYEVENT_CA_PATH>.
		1833
		1834	When neither C<ca_file> nor C<ca_path> was specified during
		1835	L<AnyEvent::TLS> context creation, and either of these environment
		1836	variables exist, they will be used to specify CA certificate locations
		1837	instead of a system-dependent default.
		1838
		1839	=item C<PERL_ANYEVENT_AVOID_GUARD> and C<PERL_ANYEVENT_AVOID_ASYNC_INTERRUPT>
		1840
		1841	When these are set to C<1>, then the respective modules are not
		1842	loaded. Mostly good for testing AnyEvent itself.
		1843
		1844	=back
		1845
420	=head1 SUPPLYING YOUR OWN EVENT MODEL INTERFACE	1846	=head1 SUPPLYING YOUR OWN EVENT MODEL INTERFACE
		1847
		1848	This is an advanced topic that you do not normally need to use AnyEvent in
		1849	a module. This section is only of use to event loop authors who want to
		1850	provide AnyEvent compatibility.
421		1851
422	If you need to support another event library which isn't directly	1852	If you need to support another event library which isn't directly
423	supported by AnyEvent, you can supply your own interface to it by	1853	supported by AnyEvent, you can supply your own interface to it by
424	pushing, before the first watcher gets created, the package name of	1854	pushing, before the first watcher gets created, the package name of
425	the event module and the package name of the interface to use onto	1855	the event module and the package name of the interface to use onto
426	C<@AnyEvent::REGISTRY>. You can do that before and even without loading	1856	C<@AnyEvent::REGISTRY>. You can do that before and even without loading
427	AnyEvent.	1857	AnyEvent, so it is reasonably cheap.
428		1858
429	Example:	1859	Example:
430		1860
431	push @AnyEvent::REGISTRY, [urxvt => urxvt::anyevent::];	1861	push @AnyEvent::REGISTRY, [urxvt => urxvt::anyevent::];
432		1862
433	This tells AnyEvent to (literally) use the C<urxvt::anyevent::>	1863	This tells AnyEvent to (literally) use the C<urxvt::anyevent::>
434	package/class when it finds the C<urxvt> package/module is loaded. When	1864	package/class when it finds the C<urxvt> package/module is already loaded.
		1865
435	AnyEvent is loaded and asked to find a suitable event model, it will	1866	When AnyEvent is loaded and asked to find a suitable event model, it
436	first check for the presence of urxvt.	1867	will first check for the presence of urxvt by trying to C<use> the
		1868	C<urxvt::anyevent> module.
437		1869
438	The class should provide implementations for all watcher types (see	1870	The class should provide implementations for all watcher types. See
439	L<AnyEvent::Impl::Event> (source code), L<AnyEvent::Impl::Glib>	1871	L<AnyEvent::Impl::EV> (source code), L<AnyEvent::Impl::Glib> (Source code)
440	(Source code) and so on for actual examples, use C<perldoc -m	1872	and so on for actual examples. Use C<perldoc -m AnyEvent::Impl::Glib> to
441	AnyEvent::Impl::Glib> to see the sources).	1873	see the sources.
442		1874
		1875	If you don't provide C<signal> and C<child> watchers than AnyEvent will
		1876	provide suitable (hopefully) replacements.
		1877
443	The above isn't fictitious, the I<rxvt-unicode> (a.k.a. urxvt)	1878	The above example isn't fictitious, the I<rxvt-unicode> (a.k.a. urxvt)
444	uses the above line as-is. An interface isn't included in AnyEvent	1879	terminal emulator uses the above line as-is. An interface isn't included
445	because it doesn't make sense outside the embedded interpreter inside	1880	in AnyEvent because it doesn't make sense outside the embedded interpreter
446	I<rxvt-unicode>, and it is updated and maintained as part of the	1881	inside I<rxvt-unicode>, and it is updated and maintained as part of the
447	I<rxvt-unicode> distribution.	1882	I<rxvt-unicode> distribution.
448		1883
449	I<rxvt-unicode> also cheats a bit by not providing blocking access to	1884	I<rxvt-unicode> also cheats a bit by not providing blocking access to
450	condition variables: code blocking while waiting for a condition will	1885	condition variables: code blocking while waiting for a condition will
451	C<die>. This still works with most modules/usages, and blocking calls must	1886	C<die>. This still works with most modules/usages, and blocking calls must
452	not be in an interactive application, so it makes sense.	1887	not be done in an interactive application, so it makes sense.
453		1888
454	=head1 ENVIRONMENT VARIABLES
455
456	The following environment variables are used by this module:
457
458	C<PERL_ANYEVENT_VERBOSE> when set to C<2> or higher, reports which event
459	model gets used.
460
461	=head1 EXAMPLE	1889	=head1 EXAMPLE PROGRAM
462		1890
463	The following program uses an io watcher to read data from stdin, a timer	1891	The following program uses an I/O watcher to read data from STDIN, a timer
464	to display a message once per second, and a condvar to exit the program	1892	to display a message once per second, and a condition variable to quit the
465	when the user enters quit:	1893	program when the user enters quit:
466		1894
467	use AnyEvent;	1895	use AnyEvent;
468		1896
469	my $cv = AnyEvent->condvar;	1897	my $cv = AnyEvent->condvar;
470		1898
471	my $io_watcher = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {	1899	my $io_watcher = AnyEvent->io (
		1900	fh => \*STDIN,
		1901	poll => 'r',
		1902	cb => sub {
472	warn "io event <$_[0]>\n"; # will always output <r>	1903	warn "io event <$_[0]>\n"; # will always output <r>
473	chomp (my $input = <STDIN>); # read a line	1904	chomp (my $input = <STDIN>); # read a line
474	warn "read: $input\n"; # output what has been read	1905	warn "read: $input\n"; # output what has been read
475	$cv->broadcast if $input =~ /^q/i; # quit program if /^q/i	1906	$cv->send if $input =~ /^q/i; # quit program if /^q/i
		1907	},
476	});	1908	);
477		1909
478	my $time_watcher; # can only be used once	1910	my $time_watcher; # can only be used once
479		1911
480	sub new_timer {	1912	sub new_timer {
481	$timer = AnyEvent->timer (after => 1, cb => sub {	1913	$timer = AnyEvent->timer (after => 1, cb => sub {
…		…
484	});	1916	});
485	}	1917	}
486		1918
487	new_timer; # create first timer	1919	new_timer; # create first timer
488		1920
489	$cv->wait; # wait until user enters /^q/i	1921	$cv->recv; # wait until user enters /^q/i
490		1922
491	=head1 REAL-WORLD EXAMPLE	1923	=head1 REAL-WORLD EXAMPLE
492		1924
493	Consider the L<Net::FCP> module. It features (among others) the following	1925	Consider the L<Net::FCP> module. It features (among others) the following
494	API calls, which are to freenet what HTTP GET requests are to http:	1926	API calls, which are to freenet what HTTP GET requests are to http:
…		…
544	syswrite $txn->{fh}, $txn->{request}	1976	syswrite $txn->{fh}, $txn->{request}
545	or die "connection or write error";	1977	or die "connection or write error";
546	$txn->{w} = AnyEvent->io (fh => $txn->{fh}, poll => 'r', cb => sub { $txn->fh_ready_r });	1978	$txn->{w} = AnyEvent->io (fh => $txn->{fh}, poll => 'r', cb => sub { $txn->fh_ready_r });
547		1979
548	Again, C<fh_ready_r> waits till all data has arrived, and then stores the	1980	Again, C<fh_ready_r> waits till all data has arrived, and then stores the
549	result and signals any possible waiters that the request ahs finished:	1981	result and signals any possible waiters that the request has finished:
550		1982
551	sysread $txn->{fh}, $txn->{buf}, length $txn->{$buf};	1983	sysread $txn->{fh}, $txn->{buf}, length $txn->{$buf};
552		1984
553	if (end-of-file or data complete) {	1985	if (end-of-file or data complete) {
554	$txn->{result} = $txn->{buf};	1986	$txn->{result} = $txn->{buf};
555	$txn->{finished}->broadcast;	1987	$txn->{finished}->send;
556	$txb->{cb}->($txn) of $txn->{cb}; # also call callback	1988	$txb->{cb}->($txn) of $txn->{cb}; # also call callback
557	}	1989	}
558		1990
559	The C<result> method, finally, just waits for the finished signal (if the	1991	The C<result> method, finally, just waits for the finished signal (if the
560	request was already finished, it doesn't wait, of course, and returns the	1992	request was already finished, it doesn't wait, of course, and returns the
561	data:	1993	data:
562		1994
563	$txn->{finished}->wait;	1995	$txn->{finished}->recv;
564	return $txn->{result};	1996	return $txn->{result};
565		1997
566	The actual code goes further and collects all errors (C<die>s, exceptions)	1998	The actual code goes further and collects all errors (C<die>s, exceptions)
567	that occured during request processing. The C<result> method detects	1999	that occurred during request processing. The C<result> method detects
568	wether an exception as thrown (it is stored inside the $txn object)	2000	whether an exception as thrown (it is stored inside the $txn object)
569	and just throws the exception, which means connection errors and other	2001	and just throws the exception, which means connection errors and other
570	problems get reported tot he code that tries to use the result, not in a	2002	problems get reported tot he code that tries to use the result, not in a
571	random callback.	2003	random callback.
572		2004
573	All of this enables the following usage styles:	2005	All of this enables the following usage styles:
574		2006
575	1. Blocking:	2007	1. Blocking:
576		2008
577	my $data = $fcp->client_get ($url);	2009	my $data = $fcp->client_get ($url);
578		2010
579	2. Blocking, but parallelizing:	2011	2. Blocking, but running in parallel:
580		2012
581	my @datas = map $_->result,	2013	my @datas = map $_->result,
582	map $fcp->txn_client_get ($_),	2014	map $fcp->txn_client_get ($_),
583	@urls;	2015	@urls;
584		2016
585	Both blocking examples work without the module user having to know	2017	Both blocking examples work without the module user having to know
586	anything about events.	2018	anything about events.
587		2019
588	3a. Event-based in a main program, using any support Event module:	2020	3a. Event-based in a main program, using any supported event module:
589		2021
590	use Event;	2022	use EV;
591		2023
592	$fcp->txn_client_get ($url)->cb (sub {	2024	$fcp->txn_client_get ($url)->cb (sub {
593	my $txn = shift;	2025	my $txn = shift;
594	my $data = $txn->result;	2026	my $data = $txn->result;
595	...	2027	...
596	});	2028	});
597		2029
598	Event::loop;	2030	EV::loop;
599		2031
600	3b. The module user could use AnyEvent, too:	2032	3b. The module user could use AnyEvent, too:
601		2033
602	use AnyEvent;	2034	use AnyEvent;
603		2035
604	my $quit = AnyEvent->condvar;	2036	my $quit = AnyEvent->condvar;
605		2037
606	$fcp->txn_client_get ($url)->cb (sub {	2038	$fcp->txn_client_get ($url)->cb (sub {
607	...	2039	...
608	$quit->broadcast;	2040	$quit->send;
609	});	2041	});
610		2042
611	$quit->wait;	2043	$quit->recv;
		2044
		2045
		2046	=head1 BENCHMARKS
		2047
		2048	To give you an idea of the performance and overheads that AnyEvent adds
		2049	over the event loops themselves and to give you an impression of the speed
		2050	of various event loops I prepared some benchmarks.
		2051
		2052	=head2 BENCHMARKING ANYEVENT OVERHEAD
		2053
		2054	Here is a benchmark of various supported event models used natively and
		2055	through AnyEvent. The benchmark creates a lot of timers (with a zero
		2056	timeout) and I/O watchers (watching STDOUT, a pty, to become writable,
		2057	which it is), lets them fire exactly once and destroys them again.
		2058
		2059	Source code for this benchmark is found as F<eg/bench> in the AnyEvent
		2060	distribution.
		2061
		2062	=head3 Explanation of the columns
		2063
		2064	I<watcher> is the number of event watchers created/destroyed. Since
		2065	different event models feature vastly different performances, each event
		2066	loop was given a number of watchers so that overall runtime is acceptable
		2067	and similar between tested event loop (and keep them from crashing): Glib
		2068	would probably take thousands of years if asked to process the same number
		2069	of watchers as EV in this benchmark.
		2070
		2071	I<bytes> is the number of bytes (as measured by the resident set size,
		2072	RSS) consumed by each watcher. This method of measuring captures both C
		2073	and Perl-based overheads.
		2074
		2075	I<create> is the time, in microseconds (millionths of seconds), that it
		2076	takes to create a single watcher. The callback is a closure shared between
		2077	all watchers, to avoid adding memory overhead. That means closure creation
		2078	and memory usage is not included in the figures.
		2079
		2080	I<invoke> is the time, in microseconds, used to invoke a simple
		2081	callback. The callback simply counts down a Perl variable and after it was
		2082	invoked "watcher" times, it would C<< ->send >> a condvar once to
		2083	signal the end of this phase.
		2084
		2085	I<destroy> is the time, in microseconds, that it takes to destroy a single
		2086	watcher.
		2087
		2088	=head3 Results
		2089
		2090	name watchers bytes create invoke destroy comment
		2091	EV/EV 400000 224 0.47 0.35 0.27 EV native interface
		2092	EV/Any 100000 224 2.88 0.34 0.27 EV + AnyEvent watchers
		2093	CoroEV/Any 100000 224 2.85 0.35 0.28 coroutines + Coro::Signal
		2094	Perl/Any 100000 452 4.13 0.73 0.95 pure perl implementation
		2095	Event/Event 16000 517 32.20 31.80 0.81 Event native interface
		2096	Event/Any 16000 590 35.85 31.55 1.06 Event + AnyEvent watchers
		2097	IOAsync/Any 16000 989 38.10 32.77 11.13 via IO::Async::Loop::IO_Poll
		2098	IOAsync/Any 16000 990 37.59 29.50 10.61 via IO::Async::Loop::Epoll
		2099	Glib/Any 16000 1357 102.33 12.31 51.00 quadratic behaviour
		2100	Tk/Any 2000 1860 27.20 66.31 14.00 SEGV with >> 2000 watchers
		2101	POE/Event 2000 6328 109.99 751.67 14.02 via POE::Loop::Event
		2102	POE/Select 2000 6027 94.54 809.13 579.80 via POE::Loop::Select
		2103
		2104	=head3 Discussion
		2105
		2106	The benchmark does I<not> measure scalability of the event loop very
		2107	well. For example, a select-based event loop (such as the pure perl one)
		2108	can never compete with an event loop that uses epoll when the number of
		2109	file descriptors grows high. In this benchmark, all events become ready at
		2110	the same time, so select/poll-based implementations get an unnatural speed
		2111	boost.
		2112
		2113	Also, note that the number of watchers usually has a nonlinear effect on
		2114	overall speed, that is, creating twice as many watchers doesn't take twice
		2115	the time - usually it takes longer. This puts event loops tested with a
		2116	higher number of watchers at a disadvantage.
		2117
		2118	To put the range of results into perspective, consider that on the
		2119	benchmark machine, handling an event takes roughly 1600 CPU cycles with
		2120	EV, 3100 CPU cycles with AnyEvent's pure perl loop and almost 3000000 CPU
		2121	cycles with POE.
		2122
		2123	C<EV> is the sole leader regarding speed and memory use, which are both
		2124	maximal/minimal, respectively. Even when going through AnyEvent, it uses
		2125	far less memory than any other event loop and is still faster than Event
		2126	natively.
		2127
		2128	The pure perl implementation is hit in a few sweet spots (both the
		2129	constant timeout and the use of a single fd hit optimisations in the perl
		2130	interpreter and the backend itself). Nevertheless this shows that it
		2131	adds very little overhead in itself. Like any select-based backend its
		2132	performance becomes really bad with lots of file descriptors (and few of
		2133	them active), of course, but this was not subject of this benchmark.
		2134
		2135	The C<Event> module has a relatively high setup and callback invocation
		2136	cost, but overall scores in on the third place.
		2137
		2138	C<IO::Async> performs admirably well, about on par with C<Event>, even
		2139	when using its pure perl backend.
		2140
		2141	C<Glib>'s memory usage is quite a bit higher, but it features a
		2142	faster callback invocation and overall ends up in the same class as
		2143	C<Event>. However, Glib scales extremely badly, doubling the number of
		2144	watchers increases the processing time by more than a factor of four,
		2145	making it completely unusable when using larger numbers of watchers
		2146	(note that only a single file descriptor was used in the benchmark, so
		2147	inefficiencies of C<poll> do not account for this).
		2148
		2149	The C<Tk> adaptor works relatively well. The fact that it crashes with
		2150	more than 2000 watchers is a big setback, however, as correctness takes
		2151	precedence over speed. Nevertheless, its performance is surprising, as the
		2152	file descriptor is dup()ed for each watcher. This shows that the dup()
		2153	employed by some adaptors is not a big performance issue (it does incur a
		2154	hidden memory cost inside the kernel which is not reflected in the figures
		2155	above).
		2156
		2157	C<POE>, regardless of underlying event loop (whether using its pure perl
		2158	select-based backend or the Event module, the POE-EV backend couldn't
		2159	be tested because it wasn't working) shows abysmal performance and
		2160	memory usage with AnyEvent: Watchers use almost 30 times as much memory
		2161	as EV watchers, and 10 times as much memory as Event (the high memory
		2162	requirements are caused by requiring a session for each watcher). Watcher
		2163	invocation speed is almost 900 times slower than with AnyEvent's pure perl
		2164	implementation.
		2165
		2166	The design of the POE adaptor class in AnyEvent can not really account
		2167	for the performance issues, though, as session creation overhead is
		2168	small compared to execution of the state machine, which is coded pretty
		2169	optimally within L<AnyEvent::Impl::POE> (and while everybody agrees that
		2170	using multiple sessions is not a good approach, especially regarding
		2171	memory usage, even the author of POE could not come up with a faster
		2172	design).
		2173
		2174	=head3 Summary
		2175
		2176	=over 4
		2177
		2178	=item * Using EV through AnyEvent is faster than any other event loop
		2179	(even when used without AnyEvent), but most event loops have acceptable
		2180	performance with or without AnyEvent.
		2181
		2182	=item * The overhead AnyEvent adds is usually much smaller than the overhead of
		2183	the actual event loop, only with extremely fast event loops such as EV
		2184	adds AnyEvent significant overhead.
		2185
		2186	=item * You should avoid POE like the plague if you want performance or
		2187	reasonable memory usage.
		2188
		2189	=back
		2190
		2191	=head2 BENCHMARKING THE LARGE SERVER CASE
		2192
		2193	This benchmark actually benchmarks the event loop itself. It works by
		2194	creating a number of "servers": each server consists of a socket pair, a
		2195	timeout watcher that gets reset on activity (but never fires), and an I/O
		2196	watcher waiting for input on one side of the socket. Each time the socket
		2197	watcher reads a byte it will write that byte to a random other "server".
		2198
		2199	The effect is that there will be a lot of I/O watchers, only part of which
		2200	are active at any one point (so there is a constant number of active
		2201	fds for each loop iteration, but which fds these are is random). The
		2202	timeout is reset each time something is read because that reflects how
		2203	most timeouts work (and puts extra pressure on the event loops).
		2204
		2205	In this benchmark, we use 10000 socket pairs (20000 sockets), of which 100
		2206	(1%) are active. This mirrors the activity of large servers with many
		2207	connections, most of which are idle at any one point in time.
		2208
		2209	Source code for this benchmark is found as F<eg/bench2> in the AnyEvent
		2210	distribution.
		2211
		2212	=head3 Explanation of the columns
		2213
		2214	I<sockets> is the number of sockets, and twice the number of "servers" (as
		2215	each server has a read and write socket end).
		2216
		2217	I<create> is the time it takes to create a socket pair (which is
		2218	nontrivial) and two watchers: an I/O watcher and a timeout watcher.
		2219
		2220	I<request>, the most important value, is the time it takes to handle a
		2221	single "request", that is, reading the token from the pipe and forwarding
		2222	it to another server. This includes deleting the old timeout and creating
		2223	a new one that moves the timeout into the future.
		2224
		2225	=head3 Results
		2226
		2227	name sockets create request
		2228	EV 20000 69.01 11.16
		2229	Perl 20000 73.32 35.87
		2230	IOAsync 20000 157.00 98.14 epoll
		2231	IOAsync 20000 159.31 616.06 poll
		2232	Event 20000 212.62 257.32
		2233	Glib 20000 651.16 1896.30
		2234	POE 20000 349.67 12317.24 uses POE::Loop::Event
		2235
		2236	=head3 Discussion
		2237
		2238	This benchmark I<does> measure scalability and overall performance of the
		2239	particular event loop.
		2240
		2241	EV is again fastest. Since it is using epoll on my system, the setup time
		2242	is relatively high, though.
		2243
		2244	Perl surprisingly comes second. It is much faster than the C-based event
		2245	loops Event and Glib.
		2246
		2247	IO::Async performs very well when using its epoll backend, and still quite
		2248	good compared to Glib when using its pure perl backend.
		2249
		2250	Event suffers from high setup time as well (look at its code and you will
		2251	understand why). Callback invocation also has a high overhead compared to
		2252	the C<< $_->() for .. >>-style loop that the Perl event loop uses. Event
		2253	uses select or poll in basically all documented configurations.
		2254
		2255	Glib is hit hard by its quadratic behaviour w.r.t. many watchers. It
		2256	clearly fails to perform with many filehandles or in busy servers.
		2257
		2258	POE is still completely out of the picture, taking over 1000 times as long
		2259	as EV, and over 100 times as long as the Perl implementation, even though
		2260	it uses a C-based event loop in this case.
		2261
		2262	=head3 Summary
		2263
		2264	=over 4
		2265
		2266	=item * The pure perl implementation performs extremely well.
		2267
		2268	=item * Avoid Glib or POE in large projects where performance matters.
		2269
		2270	=back
		2271
		2272	=head2 BENCHMARKING SMALL SERVERS
		2273
		2274	While event loops should scale (and select-based ones do not...) even to
		2275	large servers, most programs we (or I :) actually write have only a few
		2276	I/O watchers.
		2277
		2278	In this benchmark, I use the same benchmark program as in the large server
		2279	case, but it uses only eight "servers", of which three are active at any
		2280	one time. This should reflect performance for a small server relatively
		2281	well.
		2282
		2283	The columns are identical to the previous table.
		2284
		2285	=head3 Results
		2286
		2287	name sockets create request
		2288	EV 16 20.00 6.54
		2289	Perl 16 25.75 12.62
		2290	Event 16 81.27 35.86
		2291	Glib 16 32.63 15.48
		2292	POE 16 261.87 276.28 uses POE::Loop::Event
		2293
		2294	=head3 Discussion
		2295
		2296	The benchmark tries to test the performance of a typical small
		2297	server. While knowing how various event loops perform is interesting, keep
		2298	in mind that their overhead in this case is usually not as important, due
		2299	to the small absolute number of watchers (that is, you need efficiency and
		2300	speed most when you have lots of watchers, not when you only have a few of
		2301	them).
		2302
		2303	EV is again fastest.
		2304
		2305	Perl again comes second. It is noticeably faster than the C-based event
		2306	loops Event and Glib, although the difference is too small to really
		2307	matter.
		2308
		2309	POE also performs much better in this case, but is is still far behind the
		2310	others.
		2311
		2312	=head3 Summary
		2313
		2314	=over 4
		2315
		2316	=item * C-based event loops perform very well with small number of
		2317	watchers, as the management overhead dominates.
		2318
		2319	=back
		2320
		2321	=head2 THE IO::Lambda BENCHMARK
		2322
		2323	Recently I was told about the benchmark in the IO::Lambda manpage, which
		2324	could be misinterpreted to make AnyEvent look bad. In fact, the benchmark
		2325	simply compares IO::Lambda with POE, and IO::Lambda looks better (which
		2326	shouldn't come as a surprise to anybody). As such, the benchmark is
		2327	fine, and mostly shows that the AnyEvent backend from IO::Lambda isn't
		2328	very optimal. But how would AnyEvent compare when used without the extra
		2329	baggage? To explore this, I wrote the equivalent benchmark for AnyEvent.
		2330
		2331	The benchmark itself creates an echo-server, and then, for 500 times,
		2332	connects to the echo server, sends a line, waits for the reply, and then
		2333	creates the next connection. This is a rather bad benchmark, as it doesn't
		2334	test the efficiency of the framework or much non-blocking I/O, but it is a
		2335	benchmark nevertheless.
		2336
		2337	name runtime
		2338	Lambda/select 0.330 sec
		2339	+ optimized 0.122 sec
		2340	Lambda/AnyEvent 0.327 sec
		2341	+ optimized 0.138 sec
		2342	Raw sockets/select 0.077 sec
		2343	POE/select, components 0.662 sec
		2344	POE/select, raw sockets 0.226 sec
		2345	POE/select, optimized 0.404 sec
		2346
		2347	AnyEvent/select/nb 0.085 sec
		2348	AnyEvent/EV/nb 0.068 sec
		2349	+state machine 0.134 sec
		2350
		2351	The benchmark is also a bit unfair (my fault): the IO::Lambda/POE
		2352	benchmarks actually make blocking connects and use 100% blocking I/O,
		2353	defeating the purpose of an event-based solution. All of the newly
		2354	written AnyEvent benchmarks use 100% non-blocking connects (using
		2355	AnyEvent::Socket::tcp_connect and the asynchronous pure perl DNS
		2356	resolver), so AnyEvent is at a disadvantage here, as non-blocking connects
		2357	generally require a lot more bookkeeping and event handling than blocking
		2358	connects (which involve a single syscall only).
		2359
		2360	The last AnyEvent benchmark additionally uses L<AnyEvent::Handle>, which
		2361	offers similar expressive power as POE and IO::Lambda, using conventional
		2362	Perl syntax. This means that both the echo server and the client are 100%
		2363	non-blocking, further placing it at a disadvantage.
		2364
		2365	As you can see, the AnyEvent + EV combination even beats the
		2366	hand-optimised "raw sockets benchmark", while AnyEvent + its pure perl
		2367	backend easily beats IO::Lambda and POE.
		2368
		2369	And even the 100% non-blocking version written using the high-level (and
		2370	slow :) L<AnyEvent::Handle> abstraction beats both POE and IO::Lambda by a
		2371	large margin, even though it does all of DNS, tcp-connect and socket I/O
		2372	in a non-blocking way.
		2373
		2374	The two AnyEvent benchmarks programs can be found as F<eg/ae0.pl> and
		2375	F<eg/ae2.pl> in the AnyEvent distribution, the remaining benchmarks are
		2376	part of the IO::lambda distribution and were used without any changes.
		2377
		2378
		2379	=head1 SIGNALS
		2380
		2381	AnyEvent currently installs handlers for these signals:
		2382
		2383	=over 4
		2384
		2385	=item SIGCHLD
		2386
		2387	A handler for C<SIGCHLD> is installed by AnyEvent's child watcher
		2388	emulation for event loops that do not support them natively. Also, some
		2389	event loops install a similar handler.
		2390
		2391	Additionally, when AnyEvent is loaded and SIGCHLD is set to IGNORE, then
		2392	AnyEvent will reset it to default, to avoid losing child exit statuses.
		2393
		2394	=item SIGPIPE
		2395
		2396	A no-op handler is installed for C<SIGPIPE> when C<$SIG{PIPE}> is C<undef>
		2397	when AnyEvent gets loaded.
		2398
		2399	The rationale for this is that AnyEvent users usually do not really depend
		2400	on SIGPIPE delivery (which is purely an optimisation for shell use, or
		2401	badly-written programs), but C<SIGPIPE> can cause spurious and rare
		2402	program exits as a lot of people do not expect C<SIGPIPE> when writing to
		2403	some random socket.
		2404
		2405	The rationale for installing a no-op handler as opposed to ignoring it is
		2406	that this way, the handler will be restored to defaults on exec.
		2407
		2408	Feel free to install your own handler, or reset it to defaults.
		2409
		2410	=back
		2411
		2412	=cut
		2413
		2414	undef $SIG{CHLD}
		2415	if $SIG{CHLD} eq 'IGNORE';
		2416
		2417	$SIG{PIPE} = sub { }
		2418	unless defined $SIG{PIPE};
		2419
		2420	=head1 RECOMMENDED/OPTIONAL MODULES
		2421
		2422	One of AnyEvent's main goals is to be 100% Pure-Perl(tm): only perl (and
		2423	it's built-in modules) are required to use it.
		2424
		2425	That does not mean that AnyEvent won't take advantage of some additional
		2426	modules if they are installed.
		2427
		2428	This section epxlains which additional modules will be used, and how they
		2429	affect AnyEvent's operetion.
		2430
		2431	=over 4
		2432
		2433	=item L<Async::Interrupt>
		2434
		2435	This slightly arcane module is used to implement fast signal handling: To
		2436	my knowledge, there is no way to do completely race-free and quick
		2437	signal handling in pure perl. To ensure that signals still get
		2438	delivered, AnyEvent will start an interval timer to wake up perl (and
		2439	catch the signals) with some delay (default is 10 seconds, look for
		2440	C<$AnyEvent::MAX_SIGNAL_LATENCY>).
		2441
		2442	If this module is available, then it will be used to implement signal
		2443	catching, which means that signals will not be delayed, and the event loop
		2444	will not be interrupted regularly, which is more efficient (And good for
		2445	battery life on laptops).
		2446
		2447	This affects not just the pure-perl event loop, but also other event loops
		2448	that have no signal handling on their own (e.g. Glib, Tk, Qt).
		2449
		2450	Some event loops (POE, Event, Event::Lib) offer signal watchers natively,
		2451	and either employ their own workarounds (POE) or use AnyEvent's workaround
		2452	(using C<$AnyEvent::MAX_SIGNAL_LATENCY>). Installing L<Async::Interrupt>
		2453	does nothing for those backends.
		2454
		2455	=item L<EV>
		2456
		2457	This module isn't really "optional", as it is simply one of the backend
		2458	event loops that AnyEvent can use. However, it is simply the best event
		2459	loop available in terms of features, speed and stability: It supports
		2460	the AnyEvent API optimally, implements all the watcher types in XS, does
		2461	automatic timer adjustments even when no monotonic clock is available,
		2462	can take avdantage of advanced kernel interfaces such as C<epoll> and
		2463	C<kqueue>, and is the fastest backend I<by far>. You can even embed
		2464	L<Glib>/L<Gtk2> in it (or vice versa, see L<EV::Glib> and L<Glib::EV>).
		2465
		2466	=item L<Guard>
		2467
		2468	The guard module, when used, will be used to implement
		2469	C<AnyEvent::Util::guard>. This speeds up guards considerably (and uses a
		2470	lot less memory), but otherwise doesn't affect guard operation much. It is
		2471	purely used for performance.
		2472
		2473	=item L<JSON> and L<JSON::XS>
		2474
		2475	This module is required when you want to read or write JSON data via
		2476	L<AnyEvent::Handle>. It is also written in pure-perl, but can take
		2477	advantage of the ultra-high-speed L<JSON::XS> module when it is installed.
		2478
		2479	In fact, L<AnyEvent::Handle> will use L<JSON::XS> by default if it is
		2480	installed.
		2481
		2482	=item L<Net::SSLeay>
		2483
		2484	Implementing TLS/SSL in Perl is certainly interesting, but not very
		2485	worthwhile: If this module is installed, then L<AnyEvent::Handle> (with
		2486	the help of L<AnyEvent::TLS>), gains the ability to do TLS/SSL.
		2487
		2488	=item L<Time::HiRes>
		2489
		2490	This module is part of perl since release 5.008. It will be used when the
		2491	chosen event library does not come with a timing source on it's own. The
		2492	pure-perl event loop (L<AnyEvent::Impl::Perl>) will additionally use it to
		2493	try to use a monotonic clock for timing stability.
		2494
		2495	=back
		2496
		2497
		2498	=head1 FORK
		2499
		2500	Most event libraries are not fork-safe. The ones who are usually are
		2501	because they rely on inefficient but fork-safe C<select> or C<poll>
		2502	calls. Only L<EV> is fully fork-aware.
		2503
		2504	If you have to fork, you must either do so I<before> creating your first
		2505	watcher OR you must not use AnyEvent at all in the child OR you must do
		2506	something completely out of the scope of AnyEvent.
		2507
		2508
		2509	=head1 SECURITY CONSIDERATIONS
		2510
		2511	AnyEvent can be forced to load any event model via
		2512	$ENV{PERL_ANYEVENT_MODEL}. While this cannot (to my knowledge) be used to
		2513	execute arbitrary code or directly gain access, it can easily be used to
		2514	make the program hang or malfunction in subtle ways, as AnyEvent watchers
		2515	will not be active when the program uses a different event model than
		2516	specified in the variable.
		2517
		2518	You can make AnyEvent completely ignore this variable by deleting it
		2519	before the first watcher gets created, e.g. with a C<BEGIN> block:
		2520
		2521	BEGIN { delete $ENV{PERL_ANYEVENT_MODEL} }
		2522
		2523	use AnyEvent;
		2524
		2525	Similar considerations apply to $ENV{PERL_ANYEVENT_VERBOSE}, as that can
		2526	be used to probe what backend is used and gain other information (which is
		2527	probably even less useful to an attacker than PERL_ANYEVENT_MODEL), and
		2528	$ENV{PERL_ANYEVENT_STRICT}.
		2529
		2530	Note that AnyEvent will remove I<all> environment variables starting with
		2531	C<PERL_ANYEVENT_> from C<%ENV> when it is loaded while taint mode is
		2532	enabled.
		2533
		2534
		2535	=head1 BUGS
		2536
		2537	Perl 5.8 has numerous memleaks that sometimes hit this module and are hard
		2538	to work around. If you suffer from memleaks, first upgrade to Perl 5.10
		2539	and check wether the leaks still show up. (Perl 5.10.0 has other annoying
		2540	memleaks, such as leaking on C<map> and C<grep> but it is usually not as
		2541	pronounced).
		2542
612		2543
613	=head1 SEE ALSO	2544	=head1 SEE ALSO
614		2545
615	Event modules: L<Coro::Event>, L<Coro>, L<Event>, L<Glib::Event>, L<Glib>.	2546	Utility functions: L<AnyEvent::Util>.
616		2547
617	Implementations: L<AnyEvent::Impl::Coro>, L<AnyEvent::Impl::Event>, L<AnyEvent::Impl::Glib>, L<AnyEvent::Impl::Tk>.	2548	Event modules: L<EV>, L<EV::Glib>, L<Glib::EV>, L<Event>, L<Glib::Event>,
		2549	L<Glib>, L<Tk>, L<Event::Lib>, L<Qt>, L<POE>.
618		2550
619	Nontrivial usage example: L<Net::FCP>.	2551	Implementations: L<AnyEvent::Impl::EV>, L<AnyEvent::Impl::Event>,
		2552	L<AnyEvent::Impl::Glib>, L<AnyEvent::Impl::Tk>, L<AnyEvent::Impl::Perl>,
		2553	L<AnyEvent::Impl::EventLib>, L<AnyEvent::Impl::Qt>,
		2554	L<AnyEvent::Impl::POE>, L<AnyEvent::Impl::IOAsync>, L<Anyevent::Impl::Irssi>.
620		2555
621	=head1	2556	Non-blocking file handles, sockets, TCP clients and
		2557	servers: L<AnyEvent::Handle>, L<AnyEvent::Socket>, L<AnyEvent::TLS>.
		2558
		2559	Asynchronous DNS: L<AnyEvent::DNS>.
		2560
		2561	Coroutine support: L<Coro>, L<Coro::AnyEvent>, L<Coro::EV>,
		2562	L<Coro::Event>,
		2563
		2564	Nontrivial usage examples: L<AnyEvent::GPSD>, L<AnyEvent::XMPP>,
		2565	L<AnyEvent::HTTP>.
		2566
		2567
		2568	=head1 AUTHOR
		2569
		2570	Marc Lehmann <schmorp@schmorp.de>
		2571	http://home.schmorp.de/
622		2572
623	=cut	2573	=cut
624		2574
625	1	2575	1
626		2576

Diff Legend

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