ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/AnyEvent/lib/AnyEvent.pm

Comparing AnyEvent/lib/AnyEvent.pm (file contents):
Revision 1.40 by root, Mon Apr 7 19:00:55 2008 UTC vs.
Revision 1.402 by root, Mon Apr 9 02:25:48 2012 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	FLTK 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	# if you prefer function calls, look at the AE manpage for
		13	# an alternative API.
		14
		15	# file handle or descriptor readable
11	my $w = AnyEvent->io (fh => $fh, poll => "r|w", cb => sub {	16	my $w = AnyEvent->io (fh => $fh, poll => "r", cb => sub { ... });
		17
		18	# one-shot or repeating timers
		19	my $w = AnyEvent->timer (after => $seconds, cb => sub { ... });
		20	my $w = AnyEvent->timer (after => $seconds, interval => $seconds, cb => ...);
		21
		22	print AnyEvent->now; # prints current event loop time
		23	print AnyEvent->time; # think Time::HiRes::time or simply CORE::time.
		24
		25	# POSIX signal
		26	my $w = AnyEvent->signal (signal => "TERM", cb => sub { ... });
		27
		28	# child process exit
		29	my $w = AnyEvent->child (pid => $pid, cb => sub {
		30	my ($pid, $status) = @_;
12	...	31	...
13	});	32	});
14		33
15	my $w = AnyEvent->timer (after => $seconds, cb => sub {	34	# called when event loop idle (if applicable)
16	...	35	my $w = AnyEvent->idle (cb => sub { ... });
17	});
18		36
19	my $w = AnyEvent->condvar; # stores wether a condition was flagged	37	my $w = AnyEvent->condvar; # stores whether a condition was flagged
		38	$w->send; # wake up current and all future recv's
20	$w->wait; # enters "main loop" till $condvar gets ->broadcast	39	$w->recv; # enters "main loop" till $condvar gets ->send
21	$w->broadcast; # wake up current and all future wait's	40	# use a condvar in callback mode:
		41	$w->cb (sub { $_[0]->recv });
		42
		43	=head1 INTRODUCTION/TUTORIAL
		44
		45	This manpage is mainly a reference manual. If you are interested
		46	in a tutorial or some gentle introduction, have a look at the
		47	L<AnyEvent::Intro> manpage.
		48
		49	=head1 SUPPORT
		50
		51	An FAQ document is available as L<AnyEvent::FAQ>.
		52
		53	There also is a mailinglist for discussing all things AnyEvent, and an IRC
		54	channel, too.
		55
		56	See the AnyEvent project page at the B<Schmorpforge Ta-Sa Software
		57	Repository>, at L<http://anyevent.schmorp.de>, for more info.
		58
		59	=head1 WHY YOU SHOULD USE THIS MODULE (OR NOT)
		60
		61	Glib, POE, IO::Async, Event... CPAN offers event models by the dozen
		62	nowadays. So what is different about AnyEvent?
		63
		64	Executive Summary: AnyEvent is I<compatible>, AnyEvent is I<free of
		65	policy> and AnyEvent is I<small and efficient>.
		66
		67	First and foremost, I<AnyEvent is not an event model> itself, it only
		68	interfaces to whatever event model the main program happens to use, in a
		69	pragmatic way. For event models and certain classes of immortals alike,
		70	the statement "there can only be one" is a bitter reality: In general,
		71	only one event loop can be active at the same time in a process. AnyEvent
		72	cannot change this, but it can hide the differences between those event
		73	loops.
		74
		75	The goal of AnyEvent is to offer module authors the ability to do event
		76	programming (waiting for I/O or timer events) without subscribing to a
		77	religion, a way of living, and most importantly: without forcing your
		78	module users into the same thing by forcing them to use the same event
		79	model you use.
		80
		81	For modules like POE or IO::Async (which is a total misnomer as it is
		82	actually doing all I/O I<synchronously>...), using them in your module is
		83	like joining a cult: After you join, you are dependent on them and you
		84	cannot use anything else, as they are simply incompatible to everything
		85	that isn't them. What's worse, all the potential users of your
		86	module are I<also> forced to use the same event loop you use.
		87
		88	AnyEvent is different: AnyEvent + POE works fine. AnyEvent + Glib works
		89	fine. AnyEvent + Tk works fine etc. etc. but none of these work together
		90	with the rest: POE + EV? No go. Tk + Event? No go. Again: if your module
		91	uses one of those, every user of your module has to use it, too. But if
		92	your module uses AnyEvent, it works transparently with all event models it
		93	supports (including stuff like IO::Async, as long as those use one of the
		94	supported event loops. It is easy to add new event loops to AnyEvent, too,
		95	so it is future-proof).
		96
		97	In addition to being free of having to use I<the one and only true event
		98	model>, AnyEvent also is free of bloat and policy: with POE or similar
		99	modules, you get an enormous amount of code and strict rules you have to
		100	follow. AnyEvent, on the other hand, is lean and to the point, by only
		101	offering the functionality that is necessary, in as thin as a wrapper as
		102	technically possible.
		103
		104	Of course, AnyEvent comes with a big (and fully optional!) toolbox
		105	of useful functionality, such as an asynchronous DNS resolver, 100%
		106	non-blocking connects (even with TLS/SSL, IPv6 and on broken platforms
		107	such as Windows) and lots of real-world knowledge and workarounds for
		108	platform bugs and differences.
		109
		110	Now, if you I<do want> lots of policy (this can arguably be somewhat
		111	useful) and you want to force your users to use the one and only event
		112	model, you should I<not> use this module.
22		113
23	=head1 DESCRIPTION	114	=head1 DESCRIPTION
24		115
25	L<AnyEvent> provides an identical interface to multiple event loops. This	116	L<AnyEvent> provides a uniform interface to various event loops. This
26	allows module authors to utilise an event loop without forcing module	117	allows module authors to use event loop functionality without forcing
27	users to use the same event loop (as only a single event loop can coexist	118	module users to use a specific event loop implementation (since more
28	peacefully at any one time).	119	than one event loop cannot coexist peacefully).
29		120
30	The interface itself is vaguely similar but not identical to the Event	121	The interface itself is vaguely similar, but not identical to the L<Event>
31	module.	122	module.
32		123
33	On the first call of any method, the module tries to detect the currently	124	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	125	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	126	following modules is already loaded: L<EV>, L<AnyEvent::Loop>,
36	used. If none is found, the module tries to load these modules in the	127	L<Event>, L<Glib>, L<Tk>, L<Event::Lib>, L<Qt>, L<POE>. The first one
37	order given. The first one that could be successfully loaded will be	128	found is used. If none are detected, the module tries to load the first
38	used. If still none could be found, AnyEvent will fall back to a pure-perl	129	four modules in the order given; but note that if L<EV> is not
39	event loop, which is also not very efficient.	130	available, the pure-perl L<AnyEvent::Loop> should always work, so
		131	the other two are not normally tried.
40		132
41	Because AnyEvent first checks for modules that are already loaded, loading	133	Because AnyEvent first checks for modules that are already loaded, loading
42	an Event model explicitly before first using AnyEvent will likely make	134	an event model explicitly before first using AnyEvent will likely make
43	that model the default. For example:	135	that model the default. For example:
44		136
45	use Tk;	137	use Tk;
46	use AnyEvent;	138	use AnyEvent;
47		139
48	# .. AnyEvent will likely default to Tk	140	# .. AnyEvent will likely default to Tk
49		141
		142	The I<likely> means that, if any module loads another event model and
		143	starts using it, all bets are off - this case should be very rare though,
		144	as very few modules hardcode event loops without announcing this very
		145	loudly.
		146
50	The pure-perl implementation of AnyEvent is called	147	The pure-perl implementation of AnyEvent is called C<AnyEvent::Loop>. Like
51	C<AnyEvent::Impl::Perl>. Like other event modules you can load it	148	other event modules you can load it explicitly and enjoy the high
52	explicitly.	149	availability of that event loop :)
53		150
54	=head1 WATCHERS	151	=head1 WATCHERS
55		152
56	AnyEvent has the central concept of a I<watcher>, which is an object that	153	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	154	stores relevant data for each kind of event you are waiting for, such as
58	the callback to call, the filehandle to watch, etc.	155	the callback to call, the file handle to watch, etc.
59		156
60	These watchers are normal Perl objects with normal Perl lifetime. After	157	These watchers are normal Perl objects with normal Perl lifetime. After
61	creating a watcher it will immediately "watch" for events and invoke	158	creating a watcher it will immediately "watch" for events and invoke the
		159	callback when the event occurs (of course, only when the event model
		160	is in control).
		161
		162	Note that B<callbacks must not permanently change global variables>
		163	potentially in use by the event loop (such as C<$_> or C<$[>) and that B<<
		164	callbacks must not C<die> >>. The former is good programming practice in
		165	Perl and the latter stems from the fact that exception handling differs
		166	widely between event loops.
		167
62	the callback. To disable the watcher you have to destroy it (e.g. by	168	To disable a 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	169	variable you store it in to C<undef> or otherwise deleting all references
64	references to it).	170	to it).
65		171
66	All watchers are created by calling a method on the C<AnyEvent> class.	172	All watchers are created by calling a method on the C<AnyEvent> class.
67		173
		174	Many watchers either are used with "recursion" (repeating timers for
		175	example), or need to refer to their watcher object in other ways.
		176
		177	One way to achieve that is this pattern:
		178
		179	my $w; $w = AnyEvent->type (arg => value ..., cb => sub {
		180	# you can use $w here, for example to undef it
		181	undef $w;
		182	});
		183
		184	Note that C<my $w; $w => combination. This is necessary because in Perl,
		185	my variables are only visible after the statement in which they are
		186	declared.
		187
68	=head2 IO WATCHERS	188	=head2 I/O WATCHERS
69		189
		190	$w = AnyEvent->io (
		191	fh => <filehandle_or_fileno>,
		192	poll => <"r" or "w">,
		193	cb => <callback>,
		194	);
		195
70	You can create I/O watcher by calling the C<< AnyEvent->io >> method with	196	You can create an I/O watcher by calling the C<< AnyEvent->io >> method
71	the following mandatory arguments:	197	with the following mandatory key-value pairs as arguments:
72		198
73	C<fh> the Perl I<filehandle> (not filedescriptor) to watch for	199	C<fh> is the Perl I<file handle> (or a naked file descriptor) to watch
		200	for events (AnyEvent might or might not keep a reference to this file
		201	handle). Note that only file handles pointing to things for which
		202	non-blocking operation makes sense are allowed. This includes sockets,
		203	most character devices, pipes, fifos and so on, but not for example files
		204	or block devices.
		205
74	events. C<poll> must be a string that is either C<r> or C<w>, that creates	206	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	207	watcher waiting for "r"eadable or "w"ritable events, respectively.
76	to invoke everytime the filehandle becomes ready.
77		208
78	Only one io watcher per C<fh> and C<poll> combination is allowed (i.e. on	209	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		210
82	Filehandles will be kept alive, so as long as the watcher exists, the	211	Although the callback might get passed parameters, their value and
83	filehandle exists, too.	212	presence is undefined and you cannot rely on them. Portable AnyEvent
		213	callbacks cannot use arguments passed to I/O watcher callbacks.
84		214
85	Example:	215	The I/O watcher might use the underlying file descriptor or a copy of it.
		216	You must not close a file handle as long as any watcher is active on the
		217	underlying file descriptor.
86		218
		219	Some event loops issue spurious readiness notifications, so you should
		220	always use non-blocking calls when reading/writing from/to your file
		221	handles.
		222
87	# wait for readability of STDIN, then read a line and disable the watcher	223	Example: wait for readability of STDIN, then read a line and disable the
		224	watcher.
		225
88	my $w; $w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {	226	my $w; $w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {
89	chomp (my $input = <STDIN>);	227	chomp (my $input = <STDIN>);
90	warn "read: $input\n";	228	warn "read: $input\n";
91	undef $w;	229	undef $w;
92	});	230	});
93		231
94	=head2 TIME WATCHERS	232	=head2 TIME WATCHERS
95		233
		234	$w = AnyEvent->timer (after => <seconds>, cb => <callback>);
		235
		236	$w = AnyEvent->timer (
		237	after => <fractional_seconds>,
		238	interval => <fractional_seconds>,
		239	cb => <callback>,
		240	);
		241
96	You can create a time watcher by calling the C<< AnyEvent->timer >>	242	You can create a time watcher by calling the C<< AnyEvent->timer >>
97	method with the following mandatory arguments:	243	method with the following mandatory arguments:
98		244
99	C<after> after how many seconds (fractions are supported) should the timer	245	C<after> specifies after how many seconds (fractional values are
100	activate. C<cb> the callback to invoke.	246	supported) the callback should be invoked. C<cb> is the callback to invoke
		247	in that case.
101		248
102	The timer callback will be invoked at most once: if you want a repeating	249	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	250	presence is undefined and you cannot rely on them. Portable AnyEvent
104	and Glib).	251	callbacks cannot use arguments passed to time watcher callbacks.
105		252
106	Example:	253	The callback will normally be invoked only once. If you specify another
		254	parameter, C<interval>, as a strictly positive number (> 0), then the
		255	callback will be invoked regularly at that interval (in fractional
		256	seconds) after the first invocation. If C<interval> is specified with a
		257	false value, then it is treated as if it were not specified at all.
107		258
		259	The callback will be rescheduled before invoking the callback, but no
		260	attempt is made to avoid timer drift in most backends, so the interval is
		261	only approximate.
		262
108	# fire an event after 7.7 seconds	263	Example: fire an event after 7.7 seconds.
		264
109	my $w = AnyEvent->timer (after => 7.7, cb => sub {	265	my $w = AnyEvent->timer (after => 7.7, cb => sub {
110	warn "timeout\n";	266	warn "timeout\n";
111	});	267	});
112		268
113	# to cancel the timer:	269	# to cancel the timer:
114	undef $w;	270	undef $w;
115		271
116	=head2 CONDITION WATCHERS	272	Example 2: fire an event after 0.5 seconds, then roughly every second.
117		273
118	Condition watchers can be created by calling the C<< AnyEvent->condvar >>	274	my $w = AnyEvent->timer (after => 0.5, interval => 1, cb => sub {
119	method without any arguments.	275	warn "timeout\n";
		276	};
120		277
121	A condition watcher watches for a condition - precisely that the C<<	278	=head3 TIMING ISSUES
122	->broadcast >> method has been called.
123		279
124	The watcher has only two methods:	280	There are two ways to handle timers: based on real time (relative, "fire
		281	in 10 seconds") and based on wallclock time (absolute, "fire at 12
		282	o'clock").
		283
		284	While most event loops expect timers to specified in a relative way, they
		285	use absolute time internally. This makes a difference when your clock
		286	"jumps", for example, when ntp decides to set your clock backwards from
		287	the wrong date of 2014-01-01 to 2008-01-01, a watcher that is supposed to
		288	fire "after a second" might actually take six years to finally fire.
		289
		290	AnyEvent cannot compensate for this. The only event loop that is conscious
		291	of these issues is L<EV>, which offers both relative (ev_timer, based
		292	on true relative time) and absolute (ev_periodic, based on wallclock time)
		293	timers.
		294
		295	AnyEvent always prefers relative timers, if available, matching the
		296	AnyEvent API.
		297
		298	AnyEvent has two additional methods that return the "current time":
125		299
126	=over 4	300	=over 4
127		301
128	=item $cv->wait	302	=item AnyEvent->time
129		303
130	Wait (blocking if necessary) until the C<< ->broadcast >> method has been	304	This returns the "current wallclock time" as a fractional number of
131	called on c<$cv>, while servicing other watchers normally.	305	seconds since the Epoch (the same thing as C<time> or C<Time::HiRes::time>
		306	return, and the result is guaranteed to be compatible with those).
132		307
133	Not all event models support a blocking wait - some die in that case, so	308	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	309	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		310
140	You can only wait once on a condition - additional calls will return	311	=item AnyEvent->now
141	immediately.
142		312
143	=item $cv->broadcast	313	This also returns the "current wallclock time", but unlike C<time>, above,
		314	this value might change only once per event loop iteration, depending on
		315	the event loop (most return the same time as C<time>, above). This is the
		316	time that AnyEvent's timers get scheduled against.
144		317
145	Flag the condition as ready - a running C<< ->wait >> and all further	318	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	319	function to call when you want to know the current time.>
147	is waiting the broadcast will be remembered..
148		320
149	Example:	321	This function is also often faster then C<< AnyEvent->time >>, and
		322	thus the preferred method if you want some timestamp (for example,
		323	L<AnyEvent::Handle> uses this to update its activity timeouts).
150		324
151	# wait till the result is ready	325	The rest of this section is only of relevance if you try to be very exact
152	my $result_ready = AnyEvent->condvar;	326	with your timing; you can skip it without a bad conscience.
153		327
154	# do something such as adding a timer	328	For a practical example of when these times differ, consider L<Event::Lib>
155	# or socket watcher the calls $result_ready->broadcast	329	and L<EV> and the following set-up:
156	# when the "result" is ready.
157		330
158	$result_ready->wait;	331	The event loop is running and has just invoked one of your callbacks at
		332	time=500 (assume no other callbacks delay processing). In your callback,
		333	you wait a second by executing C<sleep 1> (blocking the process for a
		334	second) and then (at time=501) you create a relative timer that fires
		335	after three seconds.
		336
		337	With L<Event::Lib>, C<< AnyEvent->time >> and C<< AnyEvent->now >> will
		338	both return C<501>, because that is the current time, and the timer will
		339	be scheduled to fire at time=504 (C<501> + C<3>).
		340
		341	With L<EV>, C<< AnyEvent->time >> returns C<501> (as that is the current
		342	time), but C<< AnyEvent->now >> returns C<500>, as that is the time the
		343	last event processing phase started. With L<EV>, your timer gets scheduled
		344	to run at time=503 (C<500> + C<3>).
		345
		346	In one sense, L<Event::Lib> is more exact, as it uses the current time
		347	regardless of any delays introduced by event processing. However, most
		348	callbacks do not expect large delays in processing, so this causes a
		349	higher drift (and a lot more system calls to get the current time).
		350
		351	In another sense, L<EV> is more exact, as your timer will be scheduled at
		352	the same time, regardless of how long event processing actually took.
		353
		354	In either case, if you care (and in most cases, you don't), then you
		355	can get whatever behaviour you want with any event loop, by taking the
		356	difference between C<< AnyEvent->time >> and C<< AnyEvent->now >> into
		357	account.
		358
		359	=item AnyEvent->now_update
		360
		361	Some event loops (such as L<EV> or L<AnyEvent::Loop>) cache the current
		362	time for each loop iteration (see the discussion of L<< AnyEvent->now >>,
		363	above).
		364
		365	When a callback runs for a long time (or when the process sleeps), then
		366	this "current" time will differ substantially from the real time, which
		367	might affect timers and time-outs.
		368
		369	When this is the case, you can call this method, which will update the
		370	event loop's idea of "current time".
		371
		372	A typical example would be a script in a web server (e.g. C<mod_perl>) -
		373	when mod_perl executes the script, then the event loop will have the wrong
		374	idea about the "current time" (being potentially far in the past, when the
		375	script ran the last time). In that case you should arrange a call to C<<
		376	AnyEvent->now_update >> each time the web server process wakes up again
		377	(e.g. at the start of your script, or in a handler).
		378
		379	Note that updating the time I<might> cause some events to be handled.
159		380
160	=back	381	=back
161		382
162	=head2 SIGNAL WATCHERS	383	=head2 SIGNAL WATCHERS
163		384
		385	$w = AnyEvent->signal (signal => <uppercase_signal_name>, cb => <callback>);
		386
164	You can listen for signals using a signal watcher, C<signal> is the signal	387	You can watch for signals using a signal watcher, C<signal> is the signal
165	I<name> without any C<SIG> prefix. Multiple signals events can be clumped	388	I<name> in uppercase and without any C<SIG> prefix, C<cb> is the Perl
166	together into one callback invocation, and callback invocation might or	389	callback to be invoked whenever a signal occurs.
167	might not be asynchronous.
168		390
169	These watchers might use C<%SIG>, so programs overwriting those signals	391	Although the callback might get passed parameters, their value and
170	directly will likely not work correctly.	392	presence is undefined and you cannot rely on them. Portable AnyEvent
		393	callbacks cannot use arguments passed to signal watcher callbacks.
		394
		395	Multiple signal occurrences can be clumped together into one callback
		396	invocation, and callback invocation will be synchronous. Synchronous means
		397	that it might take a while until the signal gets handled by the process,
		398	but it is guaranteed not to interrupt any other callbacks.
		399
		400	The main advantage of using these watchers is that you can share a signal
		401	between multiple watchers, and AnyEvent will ensure that signals will not
		402	interrupt your program at bad times.
		403
		404	This watcher might use C<%SIG> (depending on the event loop used),
		405	so programs overwriting those signals directly will likely not work
		406	correctly.
171		407
172	Example: exit on SIGINT	408	Example: exit on SIGINT
173		409
174	my $w = AnyEvent->signal (signal => "INT", cb => sub { exit 1 });	410	my $w = AnyEvent->signal (signal => "INT", cb => sub { exit 1 });
175		411
		412	=head3 Restart Behaviour
		413
		414	While restart behaviour is up to the event loop implementation, most will
		415	not restart syscalls (that includes L<Async::Interrupt> and AnyEvent's
		416	pure perl implementation).
		417
		418	=head3 Safe/Unsafe Signals
		419
		420	Perl signals can be either "safe" (synchronous to opcode handling)
		421	or "unsafe" (asynchronous) - the former might delay signal delivery
		422	indefinitely, the latter might corrupt your memory.
		423
		424	AnyEvent signal handlers are, in addition, synchronous to the event loop,
		425	i.e. they will not interrupt your running perl program but will only be
		426	called as part of the normal event handling (just like timer, I/O etc.
		427	callbacks, too).
		428
		429	=head3 Signal Races, Delays and Workarounds
		430
		431	Many event loops (e.g. Glib, Tk, Qt, IO::Async) do not support
		432	attaching callbacks to signals in a generic way, which is a pity,
		433	as you cannot do race-free signal handling in perl, requiring
		434	C libraries for this. AnyEvent will try to do its best, which
		435	means in some cases, signals will be delayed. The maximum time
		436	a signal might be delayed is 10 seconds by default, but can
		437	be overriden via C<$ENV{PERL_ANYEVENT_MAX_SIGNAL_LATENCY}> or
		438	C<$AnyEvent::MAX_SIGNAL_LATENCY> - see the L<ENVIRONMENT VARIABLES>
		439	section for details.
		440
		441	All these problems can be avoided by installing the optional
		442	L<Async::Interrupt> module, which works with most event loops. It will not
		443	work with inherently broken event loops such as L<Event> or L<Event::Lib>
		444	(and not with L<POE> currently). For those, you just have to suffer the
		445	delays.
		446
176	=head2 CHILD PROCESS WATCHERS	447	=head2 CHILD PROCESS WATCHERS
177		448
178	You can also listen for the status of a child process specified by the	449	$w = AnyEvent->child (pid => <process id>, cb => <callback>);
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		450
184	Example: wait for pid 1333	451	You can also watch for a child process exit and catch its exit status.
185		452
186	my $w = AnyEvent->child (pid => 1333, cb => sub { warn "exit status $?" });	453	The child process is specified by the C<pid> argument (on some backends,
		454	using C<0> watches for any child process exit, on others this will
		455	croak). The watcher will be triggered only when the child process has
		456	finished and an exit status is available, not on any trace events
		457	(stopped/continued).
187		458
188	=head1 GLOBALS	459	The callback will be called with the pid and exit status (as returned by
		460	waitpid), so unlike other watcher types, you I<can> rely on child watcher
		461	callback arguments.
		462
		463	This watcher type works by installing a signal handler for C<SIGCHLD>,
		464	and since it cannot be shared, nothing else should use SIGCHLD or reap
		465	random child processes (waiting for specific child processes, e.g. inside
		466	C<system>, is just fine).
		467
		468	There is a slight catch to child watchers, however: you usually start them
		469	I<after> the child process was created, and this means the process could
		470	have exited already (and no SIGCHLD will be sent anymore).
		471
		472	Not all event models handle this correctly (neither POE nor IO::Async do,
		473	see their AnyEvent::Impl manpages for details), but even for event models
		474	that I<do> handle this correctly, they usually need to be loaded before
		475	the process exits (i.e. before you fork in the first place). AnyEvent's
		476	pure perl event loop handles all cases correctly regardless of when you
		477	start the watcher.
		478
		479	This means you cannot create a child watcher as the very first
		480	thing in an AnyEvent program, you I<have> to create at least one
		481	watcher before you C<fork> the child (alternatively, you can call
		482	C<AnyEvent::detect>).
		483
		484	As most event loops do not support waiting for child events, they will be
		485	emulated by AnyEvent in most cases, in which case the latency and race
		486	problems mentioned in the description of signal watchers apply.
		487
		488	Example: fork a process and wait for it
		489
		490	my $done = AnyEvent->condvar;
		491
		492	my $pid = fork or exit 5;
		493
		494	my $w = AnyEvent->child (
		495	pid => $pid,
		496	cb => sub {
		497	my ($pid, $status) = @_;
		498	warn "pid $pid exited with status $status";
		499	$done->send;
		500	},
		501	);
		502
		503	# do something else, then wait for process exit
		504	$done->recv;
		505
		506	=head2 IDLE WATCHERS
		507
		508	$w = AnyEvent->idle (cb => <callback>);
		509
		510	This will repeatedly invoke the callback after the process becomes idle,
		511	until either the watcher is destroyed or new events have been detected.
		512
		513	Idle watchers are useful when there is a need to do something, but it
		514	is not so important (or wise) to do it instantly. The callback will be
		515	invoked only when there is "nothing better to do", which is usually
		516	defined as "all outstanding events have been handled and no new events
		517	have been detected". That means that idle watchers ideally get invoked
		518	when the event loop has just polled for new events but none have been
		519	detected. Instead of blocking to wait for more events, the idle watchers
		520	will be invoked.
		521
		522	Unfortunately, most event loops do not really support idle watchers (only
		523	EV, Event and Glib do it in a usable fashion) - for the rest, AnyEvent
		524	will simply call the callback "from time to time".
		525
		526	Example: read lines from STDIN, but only process them when the
		527	program is otherwise idle:
		528
		529	my @lines; # read data
		530	my $idle_w;
		531	my $io_w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {
		532	push @lines, scalar <STDIN>;
		533
		534	# start an idle watcher, if not already done
		535	$idle_w ||= AnyEvent->idle (cb => sub {
		536	# handle only one line, when there are lines left
		537	if (my $line = shift @lines) {
		538	print "handled when idle: $line";
		539	} else {
		540	# otherwise disable the idle watcher again
		541	undef $idle_w;
		542	}
		543	});
		544	});
		545
		546	=head2 CONDITION VARIABLES
		547
		548	$cv = AnyEvent->condvar;
		549
		550	$cv->send (<list>);
		551	my @res = $cv->recv;
		552
		553	If you are familiar with some event loops you will know that all of them
		554	require you to run some blocking "loop", "run" or similar function that
		555	will actively watch for new events and call your callbacks.
		556
		557	AnyEvent is slightly different: it expects somebody else to run the event
		558	loop and will only block when necessary (usually when told by the user).
		559
		560	The tool to do that is called a "condition variable", so called because
		561	they represent a condition that must become true.
		562
		563	Now is probably a good time to look at the examples further below.
		564
		565	Condition variables can be created by calling the C<< AnyEvent->condvar
		566	>> method, usually without arguments. The only argument pair allowed is
		567	C<cb>, which specifies a callback to be called when the condition variable
		568	becomes true, with the condition variable as the first argument (but not
		569	the results).
		570
		571	After creation, the condition variable is "false" until it becomes "true"
		572	by calling the C<send> method (or calling the condition variable as if it
		573	were a callback, read about the caveats in the description for the C<<
		574	->send >> method).
		575
		576	Since condition variables are the most complex part of the AnyEvent API, here are
		577	some different mental models of what they are - pick the ones you can connect to:
189		578
190	=over 4	579	=over 4
191		580
		581	=item * Condition variables are like callbacks - you can call them (and pass them instead
		582	of callbacks). Unlike callbacks however, you can also wait for them to be called.
		583
		584	=item * Condition variables are signals - one side can emit or send them,
		585	the other side can wait for them, or install a handler that is called when
		586	the signal fires.
		587
		588	=item * Condition variables are like "Merge Points" - points in your program
		589	where you merge multiple independent results/control flows into one.
		590
		591	=item * Condition variables represent a transaction - functions that start
		592	some kind of transaction can return them, leaving the caller the choice
		593	between waiting in a blocking fashion, or setting a callback.
		594
		595	=item * Condition variables represent future values, or promises to deliver
		596	some result, long before the result is available.
		597
		598	=back
		599
		600	Condition variables are very useful to signal that something has finished,
		601	for example, if you write a module that does asynchronous http requests,
		602	then a condition variable would be the ideal candidate to signal the
		603	availability of results. The user can either act when the callback is
		604	called or can synchronously C<< ->recv >> for the results.
		605
		606	You can also use them to simulate traditional event loops - for example,
		607	you can block your main program until an event occurs - for example, you
		608	could C<< ->recv >> in your main program until the user clicks the Quit
		609	button of your app, which would C<< ->send >> the "quit" event.
		610
		611	Note that condition variables recurse into the event loop - if you have
		612	two pieces of code that call C<< ->recv >> in a round-robin fashion, you
		613	lose. Therefore, condition variables are good to export to your caller, but
		614	you should avoid making a blocking wait yourself, at least in callbacks,
		615	as this asks for trouble.
		616
		617	Condition variables are represented by hash refs in perl, and the keys
		618	used by AnyEvent itself are all named C<_ae_XXX> to make subclassing
		619	easy (it is often useful to build your own transaction class on top of
		620	AnyEvent). To subclass, use C<AnyEvent::CondVar> as base class and call
		621	its C<new> method in your own C<new> method.
		622
		623	There are two "sides" to a condition variable - the "producer side" which
		624	eventually calls C<< -> send >>, and the "consumer side", which waits
		625	for the send to occur.
		626
		627	Example: wait for a timer.
		628
		629	# condition: "wait till the timer is fired"
		630	my $timer_fired = AnyEvent->condvar;
		631
		632	# create the timer - we could wait for, say
		633	# a handle becomign ready, or even an
		634	# AnyEvent::HTTP request to finish, but
		635	# in this case, we simply use a timer:
		636	my $w = AnyEvent->timer (
		637	after => 1,
		638	cb => sub { $timer_fired->send },
		639	);
		640
		641	# this "blocks" (while handling events) till the callback
		642	# calls ->send
		643	$timer_fired->recv;
		644
		645	Example: wait for a timer, but take advantage of the fact that condition
		646	variables are also callable directly.
		647
		648	my $done = AnyEvent->condvar;
		649	my $delay = AnyEvent->timer (after => 5, cb => $done);
		650	$done->recv;
		651
		652	Example: Imagine an API that returns a condvar and doesn't support
		653	callbacks. This is how you make a synchronous call, for example from
		654	the main program:
		655
		656	use AnyEvent::CouchDB;
		657
		658	...
		659
		660	my @info = $couchdb->info->recv;
		661
		662	And this is how you would just set a callback to be called whenever the
		663	results are available:
		664
		665	$couchdb->info->cb (sub {
		666	my @info = $_[0]->recv;
		667	});
		668
		669	=head3 METHODS FOR PRODUCERS
		670
		671	These methods should only be used by the producing side, i.e. the
		672	code/module that eventually sends the signal. Note that it is also
		673	the producer side which creates the condvar in most cases, but it isn't
		674	uncommon for the consumer to create it as well.
		675
		676	=over 4
		677
		678	=item $cv->send (...)
		679
		680	Flag the condition as ready - a running C<< ->recv >> and all further
		681	calls to C<recv> will (eventually) return after this method has been
		682	called. If nobody is waiting the send will be remembered.
		683
		684	If a callback has been set on the condition variable, it is called
		685	immediately from within send.
		686
		687	Any arguments passed to the C<send> call will be returned by all
		688	future C<< ->recv >> calls.
		689
		690	Condition variables are overloaded so one can call them directly (as if
		691	they were a code reference). Calling them directly is the same as calling
		692	C<send>.
		693
		694	=item $cv->croak ($error)
		695
		696	Similar to send, but causes all calls to C<< ->recv >> to invoke
		697	C<Carp::croak> with the given error message/object/scalar.
		698
		699	This can be used to signal any errors to the condition variable
		700	user/consumer. Doing it this way instead of calling C<croak> directly
		701	delays the error detection, but has the overwhelming advantage that it
		702	diagnoses the error at the place where the result is expected, and not
		703	deep in some event callback with no connection to the actual code causing
		704	the problem.
		705
		706	=item $cv->begin ([group callback])
		707
		708	=item $cv->end
		709
		710	These two methods can be used to combine many transactions/events into
		711	one. For example, a function that pings many hosts in parallel might want
		712	to use a condition variable for the whole process.
		713
		714	Every call to C<< ->begin >> will increment a counter, and every call to
		715	C<< ->end >> will decrement it. If the counter reaches C<0> in C<< ->end
		716	>>, the (last) callback passed to C<begin> will be executed, passing the
		717	condvar as first argument. That callback is I<supposed> to call C<< ->send
		718	>>, but that is not required. If no group callback was set, C<send> will
		719	be called without any arguments.
		720
		721	You can think of C<< $cv->send >> giving you an OR condition (one call
		722	sends), while C<< $cv->begin >> and C<< $cv->end >> giving you an AND
		723	condition (all C<begin> calls must be C<end>'ed before the condvar sends).
		724
		725	Let's start with a simple example: you have two I/O watchers (for example,
		726	STDOUT and STDERR for a program), and you want to wait for both streams to
		727	close before activating a condvar:
		728
		729	my $cv = AnyEvent->condvar;
		730
		731	$cv->begin; # first watcher
		732	my $w1 = AnyEvent->io (fh => $fh1, cb => sub {
		733	defined sysread $fh1, my $buf, 4096
		734	or $cv->end;
		735	});
		736
		737	$cv->begin; # second watcher
		738	my $w2 = AnyEvent->io (fh => $fh2, cb => sub {
		739	defined sysread $fh2, my $buf, 4096
		740	or $cv->end;
		741	});
		742
		743	$cv->recv;
		744
		745	This works because for every event source (EOF on file handle), there is
		746	one call to C<begin>, so the condvar waits for all calls to C<end> before
		747	sending.
		748
		749	The ping example mentioned above is slightly more complicated, as the
		750	there are results to be passwd back, and the number of tasks that are
		751	begun can potentially be zero:
		752
		753	my $cv = AnyEvent->condvar;
		754
		755	my %result;
		756	$cv->begin (sub { shift->send (\%result) });
		757
		758	for my $host (@list_of_hosts) {
		759	$cv->begin;
		760	ping_host_then_call_callback $host, sub {
		761	$result{$host} = ...;
		762	$cv->end;
		763	};
		764	}
		765
		766	$cv->end;
		767
		768	This code fragment supposedly pings a number of hosts and calls
		769	C<send> after results for all then have have been gathered - in any
		770	order. To achieve this, the code issues a call to C<begin> when it starts
		771	each ping request and calls C<end> when it has received some result for
		772	it. Since C<begin> and C<end> only maintain a counter, the order in which
		773	results arrive is not relevant.
		774
		775	There is an additional bracketing call to C<begin> and C<end> outside the
		776	loop, which serves two important purposes: first, it sets the callback
		777	to be called once the counter reaches C<0>, and second, it ensures that
		778	C<send> is called even when C<no> hosts are being pinged (the loop
		779	doesn't execute once).
		780
		781	This is the general pattern when you "fan out" into multiple (but
		782	potentially zero) subrequests: use an outer C<begin>/C<end> pair to set
		783	the callback and ensure C<end> is called at least once, and then, for each
		784	subrequest you start, call C<begin> and for each subrequest you finish,
		785	call C<end>.
		786
		787	=back
		788
		789	=head3 METHODS FOR CONSUMERS
		790
		791	These methods should only be used by the consuming side, i.e. the
		792	code awaits the condition.
		793
		794	=over 4
		795
		796	=item $cv->recv
		797
		798	Wait (blocking if necessary) until the C<< ->send >> or C<< ->croak
		799	>> methods have been called on C<$cv>, while servicing other watchers
		800	normally.
		801
		802	You can only wait once on a condition - additional calls are valid but
		803	will return immediately.
		804
		805	If an error condition has been set by calling C<< ->croak >>, then this
		806	function will call C<croak>.
		807
		808	In list context, all parameters passed to C<send> will be returned,
		809	in scalar context only the first one will be returned.
		810
		811	Note that doing a blocking wait in a callback is not supported by any
		812	event loop, that is, recursive invocation of a blocking C<< ->recv
		813	>> is not allowed, and the C<recv> call will C<croak> if such a
		814	condition is detected. This condition can be slightly loosened by using
		815	L<Coro::AnyEvent>, which allows you to do a blocking C<< ->recv >> from
		816	any thread that doesn't run the event loop itself.
		817
		818	Not all event models support a blocking wait - some die in that case
		819	(programs might want to do that to stay interactive), so I<if you are
		820	using this from a module, never require a blocking wait>. Instead, let the
		821	caller decide whether the call will block or not (for example, by coupling
		822	condition variables with some kind of request results and supporting
		823	callbacks so the caller knows that getting the result will not block,
		824	while still supporting blocking waits if the caller so desires).
		825
		826	You can ensure that C<< ->recv >> never blocks by setting a callback and
		827	only calling C<< ->recv >> from within that callback (or at a later
		828	time). This will work even when the event loop does not support blocking
		829	waits otherwise.
		830
		831	=item $bool = $cv->ready
		832
		833	Returns true when the condition is "true", i.e. whether C<send> or
		834	C<croak> have been called.
		835
		836	=item $cb = $cv->cb ($cb->($cv))
		837
		838	This is a mutator function that returns the callback set and optionally
		839	replaces it before doing so.
		840
		841	The callback will be called when the condition becomes "true", i.e. when
		842	C<send> or C<croak> are called, with the only argument being the
		843	condition variable itself. If the condition is already true, the
		844	callback is called immediately when it is set. Calling C<recv> inside
		845	the callback or at any later time is guaranteed not to block.
		846
		847	=back
		848
		849	=head1 SUPPORTED EVENT LOOPS/BACKENDS
		850
		851	The available backend classes are (every class has its own manpage):
		852
		853	=over 4
		854
		855	=item Backends that are autoprobed when no other event loop can be found.
		856
		857	EV is the preferred backend when no other event loop seems to be in
		858	use. If EV is not installed, then AnyEvent will fall back to its own
		859	pure-perl implementation, which is available everywhere as it comes with
		860	AnyEvent itself.
		861
		862	AnyEvent::Impl::EV based on EV (interface to libev, best choice).
		863	AnyEvent::Impl::Perl pure-perl AnyEvent::Loop, fast and portable.
		864
		865	=item Backends that are transparently being picked up when they are used.
		866
		867	These will be used if they are already loaded when the first watcher
		868	is created, in which case it is assumed that the application is using
		869	them. This means that AnyEvent will automatically pick the right backend
		870	when the main program loads an event module before anything starts to
		871	create watchers. Nothing special needs to be done by the main program.
		872
		873	AnyEvent::Impl::Event based on Event, very stable, few glitches.
		874	AnyEvent::Impl::Glib based on Glib, slow but very stable.
		875	AnyEvent::Impl::Tk based on Tk, very broken.
		876	AnyEvent::Impl::EventLib based on Event::Lib, leaks memory and worse.
		877	AnyEvent::Impl::POE based on POE, very slow, some limitations.
		878	AnyEvent::Impl::Irssi used when running within irssi.
		879	AnyEvent::Impl::IOAsync based on IO::Async.
		880	AnyEvent::Impl::Cocoa based on Cocoa::EventLoop.
		881	AnyEvent::Impl::FLTK based on FLTK (fltk 2 binding).
		882
		883	=item Backends with special needs.
		884
		885	Qt requires the Qt::Application to be instantiated first, but will
		886	otherwise be picked up automatically. As long as the main program
		887	instantiates the application before any AnyEvent watchers are created,
		888	everything should just work.
		889
		890	AnyEvent::Impl::Qt based on Qt.
		891
		892	=item Event loops that are indirectly supported via other backends.
		893
		894	Some event loops can be supported via other modules:
		895
		896	There is no direct support for WxWidgets (L<Wx>) or L<Prima>.
		897
		898	B<WxWidgets> has no support for watching file handles. However, you can
		899	use WxWidgets through the POE adaptor, as POE has a Wx backend that simply
		900	polls 20 times per second, which was considered to be too horrible to even
		901	consider for AnyEvent.
		902
		903	B<Prima> is not supported as nobody seems to be using it, but it has a POE
		904	backend, so it can be supported through POE.
		905
		906	AnyEvent knows about both L<Prima> and L<Wx>, however, and will try to
		907	load L<POE> when detecting them, in the hope that POE will pick them up,
		908	in which case everything will be automatic.
		909
		910	=back
		911
		912	=head1 GLOBAL VARIABLES AND FUNCTIONS
		913
		914	These are not normally required to use AnyEvent, but can be useful to
		915	write AnyEvent extension modules.
		916
		917	=over 4
		918
192	=item $AnyEvent::MODEL	919	=item $AnyEvent::MODEL
193		920
194	Contains C<undef> until the first watcher is being created. Then it	921	Contains C<undef> until the first watcher is being created, before the
		922	backend has been autodetected.
		923
195	contains the event model that is being used, which is the name of the	924	Afterwards it contains the event model that is being used, which is the
196	Perl class implementing the model. This class is usually one of the	925	name of the Perl class implementing the model. This class is usually one
197	C<AnyEvent::Impl:xxx> modules, but can be any other class in the case	926	of the C<AnyEvent::Impl::xxx> modules, but can be any other class in the
198	AnyEvent has been extended at runtime (e.g. in I<rxvt-unicode>).	927	case AnyEvent has been extended at runtime (e.g. in I<rxvt-unicode> it
199		928	will be C<urxvt::anyevent>).
200	The known classes so far are:
201
202	AnyEvent::Impl::CoroEV based on Coro::EV, best choice.
203	AnyEvent::Impl::EV based on EV (an interface to libev, also best choice).
204	AnyEvent::Impl::CoroEvent based on Coro::Event, second best choice.
205	AnyEvent::Impl::Event based on Event, also second best choice :)
206	AnyEvent::Impl::Glib based on Glib, second-best choice.
207	AnyEvent::Impl::Tk based on Tk, very bad choice.
208	AnyEvent::Impl::Perl pure-perl implementation, inefficient.
209		929
210	=item AnyEvent::detect	930	=item AnyEvent::detect
211		931
212	Returns C<$AnyEvent::MODEL>, forcing autodetection of the event model if	932	Returns C<$AnyEvent::MODEL>, forcing autodetection of the event model
213	necessary. You should only call this function right before you would have	933	if necessary. You should only call this function right before you would
214	created an AnyEvent watcher anyway, that is, very late at runtime.	934	have created an AnyEvent watcher anyway, that is, as late as possible at
		935	runtime, and not e.g. during initialisation of your module.
		936
		937	The effect of calling this function is as if a watcher had been created
		938	(specifically, actions that happen "when the first watcher is created"
		939	happen when calling detetc as well).
		940
		941	If you need to do some initialisation before AnyEvent watchers are
		942	created, use C<post_detect>.
		943
		944	=item $guard = AnyEvent::post_detect { BLOCK }
		945
		946	Arranges for the code block to be executed as soon as the event model is
		947	autodetected (or immediately if that has already happened).
		948
		949	The block will be executed I<after> the actual backend has been detected
		950	(C<$AnyEvent::MODEL> is set), but I<before> any watchers have been
		951	created, so it is possible to e.g. patch C<@AnyEvent::ISA> or do
		952	other initialisations - see the sources of L<AnyEvent::Strict> or
		953	L<AnyEvent::AIO> to see how this is used.
		954
		955	The most common usage is to create some global watchers, without forcing
		956	event module detection too early, for example, L<AnyEvent::AIO> creates
		957	and installs the global L<IO::AIO> watcher in a C<post_detect> block to
		958	avoid autodetecting the event module at load time.
		959
		960	If called in scalar or list context, then it creates and returns an object
		961	that automatically removes the callback again when it is destroyed (or
		962	C<undef> when the hook was immediately executed). See L<AnyEvent::AIO> for
		963	a case where this is useful.
		964
		965	Example: Create a watcher for the IO::AIO module and store it in
		966	C<$WATCHER>, but do so only do so after the event loop is initialised.
		967
		968	our WATCHER;
		969
		970	my $guard = AnyEvent::post_detect {
		971	$WATCHER = AnyEvent->io (fh => IO::AIO::poll_fileno, poll => 'r', cb => \&IO::AIO::poll_cb);
		972	};
		973
		974	# the ||= is important in case post_detect immediately runs the block,
		975	# as to not clobber the newly-created watcher. assigning both watcher and
		976	# post_detect guard to the same variable has the advantage of users being
		977	# able to just C<undef $WATCHER> if the watcher causes them grief.
		978
		979	$WATCHER ||= $guard;
		980
		981	=item @AnyEvent::post_detect
		982
		983	If there are any code references in this array (you can C<push> to it
		984	before or after loading AnyEvent), then they will be called directly
		985	after the event loop has been chosen.
		986
		987	You should check C<$AnyEvent::MODEL> before adding to this array, though:
		988	if it is defined then the event loop has already been detected, and the
		989	array will be ignored.
		990
		991	Best use C<AnyEvent::post_detect { BLOCK }> when your application allows
		992	it, as it takes care of these details.
		993
		994	This variable is mainly useful for modules that can do something useful
		995	when AnyEvent is used and thus want to know when it is initialised, but do
		996	not need to even load it by default. This array provides the means to hook
		997	into AnyEvent passively, without loading it.
		998
		999	Example: To load Coro::AnyEvent whenever Coro and AnyEvent are used
		1000	together, you could put this into Coro (this is the actual code used by
		1001	Coro to accomplish this):
		1002
		1003	if (defined $AnyEvent::MODEL) {
		1004	# AnyEvent already initialised, so load Coro::AnyEvent
		1005	require Coro::AnyEvent;
		1006	} else {
		1007	# AnyEvent not yet initialised, so make sure to load Coro::AnyEvent
		1008	# as soon as it is
		1009	push @AnyEvent::post_detect, sub { require Coro::AnyEvent };
		1010	}
		1011
		1012	=item AnyEvent::postpone { BLOCK }
		1013
		1014	Arranges for the block to be executed as soon as possible, but not before
		1015	the call itself returns. In practise, the block will be executed just
		1016	before the event loop polls for new events, or shortly afterwards.
		1017
		1018	This function never returns anything (to make the C<return postpone { ...
		1019	}> idiom more useful.
		1020
		1021	To understand the usefulness of this function, consider a function that
		1022	asynchronously does something for you and returns some transaction
		1023	object or guard to let you cancel the operation. For example,
		1024	C<AnyEvent::Socket::tcp_connect>:
		1025
		1026	# start a conenction attempt unless one is active
		1027	$self->{connect_guard} ||= AnyEvent::Socket::tcp_connect "www.example.net", 80, sub {
		1028	delete $self->{connect_guard};
		1029	...
		1030	};
		1031
		1032	Imagine that this function could instantly call the callback, for
		1033	example, because it detects an obvious error such as a negative port
		1034	number. Invoking the callback before the function returns causes problems
		1035	however: the callback will be called and will try to delete the guard
		1036	object. But since the function hasn't returned yet, there is nothing to
		1037	delete. When the function eventually returns it will assign the guard
		1038	object to C<< $self->{connect_guard} >>, where it will likely never be
		1039	deleted, so the program thinks it is still trying to connect.
		1040
		1041	This is where C<AnyEvent::postpone> should be used. Instead of calling the
		1042	callback directly on error:
		1043
		1044	$cb->(undef), return # signal error to callback, BAD!
		1045	if $some_error_condition;
		1046
		1047	It should use C<postpone>:
		1048
		1049	AnyEvent::postpone { $cb->(undef) }, return # signal error to callback, later
		1050	if $some_error_condition;
		1051
		1052	=item AnyEvent::log $level, $msg[, @args]
		1053
		1054	Log the given C<$msg> at the given C<$level>.
		1055
		1056	If L<AnyEvent::Log> is not loaded then this function makes a simple test
		1057	to see whether the message will be logged. If the test succeeds it will
		1058	load AnyEvent::Log and call C<AnyEvent::Log::log> - consequently, look at
		1059	the L<AnyEvent::Log> documentation for details.
		1060
		1061	If the test fails it will simply return. Right now this happens when a
		1062	numerical loglevel is used and it is larger than the level specified via
		1063	C<$ENV{PERL_ANYEVENT_VERBOSE}>.
		1064
		1065	If you want to sprinkle loads of logging calls around your code, consider
		1066	creating a logger callback with the C<AnyEvent::Log::logger> function,
		1067	which can reduce typing, codesize and can reduce the logging overhead
		1068	enourmously.
215		1069
216	=back	1070	=back
217		1071
218	=head1 WHAT TO DO IN A MODULE	1072	=head1 WHAT TO DO IN A MODULE
219		1073
220	As a module author, you should "use AnyEvent" and call AnyEvent methods	1074	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.	1075	freely, but you should not load a specific event module or rely on it.
222		1076
223	Be careful when you create watchers in the module body - Anyevent will	1077	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	1078	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	1079	by calling AnyEvent in your module body you force the user of your module
226	to load the event module first.	1080	to load the event module first.
227		1081
		1082	Never call C<< ->recv >> on a condition variable unless you I<know> that
		1083	the C<< ->send >> method has been called on it already. This is
		1084	because it will stall the whole program, and the whole point of using
		1085	events is to stay interactive.
		1086
		1087	It is fine, however, to call C<< ->recv >> when the user of your module
		1088	requests it (i.e. if you create a http request object ad have a method
		1089	called C<results> that returns the results, it may call C<< ->recv >>
		1090	freely, as the user of your module knows what she is doing. Always).
		1091
228	=head1 WHAT TO DO IN THE MAIN PROGRAM	1092	=head1 WHAT TO DO IN THE MAIN PROGRAM
229		1093
230	There will always be a single main program - the only place that should	1094	There will always be a single main program - the only place that should
231	dictate which event model to use.	1095	dictate which event model to use.
232		1096
233	If it doesn't care, it can just "use AnyEvent" and use it itself, or not	1097	If the program is not event-based, it need not do anything special, even
234	do anything special and let AnyEvent decide which implementation to chose.	1098	when it depends on a module that uses an AnyEvent. If the program itself
		1099	uses AnyEvent, but does not care which event loop is used, all it needs
		1100	to do is C<use AnyEvent>. In either case, AnyEvent will choose the best
		1101	available loop implementation.
235		1102
236	If the main program relies on a specific event model (for example, in Gtk2	1103	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	1104	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	1105	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	1106	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	1107	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	1108	decide on the event model to use as soon as it creates watchers, and it
242	correct one yourself.	1109	might choose the wrong one unless you load the correct one yourself.
243		1110
244	You can chose to use a rather inefficient pure-perl implementation by	1111	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	1112	C<AnyEvent::Loop> module, which gives you similar behaviour
246	generally better.	1113	everywhere, but letting AnyEvent chose the model is generally better.
		1114
		1115	=head2 MAINLOOP EMULATION
		1116
		1117	Sometimes (often for short test scripts, or even standalone programs who
		1118	only want to use AnyEvent), you do not want to run a specific event loop.
		1119
		1120	In that case, you can use a condition variable like this:
		1121
		1122	AnyEvent->condvar->recv;
		1123
		1124	This has the effect of entering the event loop and looping forever.
		1125
		1126	Note that usually your program has some exit condition, in which case
		1127	it is better to use the "traditional" approach of storing a condition
		1128	variable somewhere, waiting for it, and sending it when the program should
		1129	exit cleanly.
		1130
		1131
		1132	=head1 OTHER MODULES
		1133
		1134	The following is a non-exhaustive list of additional modules that use
		1135	AnyEvent as a client and can therefore be mixed easily with other
		1136	AnyEvent modules and other event loops in the same program. Some of the
		1137	modules come as part of AnyEvent, the others are available via CPAN (see
		1138	L<http://search.cpan.org/search?m=module&q=anyevent%3A%3A*> for
		1139	a longer non-exhaustive list), and the list is heavily biased towards
		1140	modules of the AnyEvent author himself :)
		1141
		1142	=over 4
		1143
		1144	=item L<AnyEvent::Util>
		1145
		1146	Contains various utility functions that replace often-used blocking
		1147	functions such as C<inet_aton> with event/callback-based versions.
		1148
		1149	=item L<AnyEvent::Socket>
		1150
		1151	Provides various utility functions for (internet protocol) sockets,
		1152	addresses and name resolution. Also functions to create non-blocking tcp
		1153	connections or tcp servers, with IPv6 and SRV record support and more.
		1154
		1155	=item L<AnyEvent::Handle>
		1156
		1157	Provide read and write buffers, manages watchers for reads and writes,
		1158	supports raw and formatted I/O, I/O queued and fully transparent and
		1159	non-blocking SSL/TLS (via L<AnyEvent::TLS>).
		1160
		1161	=item L<AnyEvent::DNS>
		1162
		1163	Provides rich asynchronous DNS resolver capabilities.
		1164
		1165	=item L<AnyEvent::HTTP>, L<AnyEvent::IRC>, L<AnyEvent::XMPP>, L<AnyEvent::GPSD>, L<AnyEvent::IGS>, L<AnyEvent::FCP>
		1166
		1167	Implement event-based interfaces to the protocols of the same name (for
		1168	the curious, IGS is the International Go Server and FCP is the Freenet
		1169	Client Protocol).
		1170
		1171	=item L<AnyEvent::AIO>
		1172
		1173	Truly asynchronous (as opposed to non-blocking) I/O, should be in the
		1174	toolbox of every event programmer. AnyEvent::AIO transparently fuses
		1175	L<IO::AIO> and AnyEvent together, giving AnyEvent access to event-based
		1176	file I/O, and much more.
		1177
		1178	=item L<AnyEvent::Filesys::Notify>
		1179
		1180	AnyEvent is good for non-blocking stuff, but it can't detect file or
		1181	path changes (e.g. "watch this directory for new files", "watch this
		1182	file for changes"). The L<AnyEvent::Filesys::Notify> module promises to
		1183	do just that in a portbale fashion, supporting inotify on GNU/Linux and
		1184	some weird, without doubt broken, stuff on OS X to monitor files. It can
		1185	fall back to blocking scans at regular intervals transparently on other
		1186	platforms, so it's about as portable as it gets.
		1187
		1188	(I haven't used it myself, but I haven't heard anybody complaining about
		1189	it yet).
		1190
		1191	=item L<AnyEvent::DBI>
		1192
		1193	Executes L<DBI> requests asynchronously in a proxy process for you,
		1194	notifying you in an event-based way when the operation is finished.
		1195
		1196	=item L<AnyEvent::HTTPD>
		1197
		1198	A simple embedded webserver.
		1199
		1200	=item L<AnyEvent::FastPing>
		1201
		1202	The fastest ping in the west.
		1203
		1204	=item L<Coro>
		1205
		1206	Has special support for AnyEvent via L<Coro::AnyEvent>, which allows you
		1207	to simply invert the flow control - don't call us, we will call you:
		1208
		1209	async {
		1210	Coro::AnyEvent::sleep 5; # creates a 5s timer and waits for it
		1211	print "5 seconds later!\n";
		1212
		1213	Coro::AnyEvent::readable *STDIN; # uses an I/O watcher
		1214	my $line = <STDIN>; # works for ttys
		1215
		1216	AnyEvent::HTTP::http_get "url", Coro::rouse_cb;
		1217	my ($body, $hdr) = Coro::rouse_wait;
		1218	};
		1219
		1220	=back
247		1221
248	=cut	1222	=cut
249		1223
250	package AnyEvent;	1224	package AnyEvent;
251		1225
252	no warnings;	1226	# basically a tuned-down version of common::sense
253	use strict;	1227	sub common_sense {
		1228	# from common:.sense 3.5
		1229	local $^W;
		1230	${^WARNING_BITS} ^= ${^WARNING_BITS} ^ "\x3c\x3f\x33\x00\x0f\xf0\x0f\xc0\xf0\xfc\x33\x00";
		1231	# use strict vars subs - NO UTF-8, as Util.pm doesn't like this atm. (uts46data.pl)
		1232	$^H |= 0x00000600;
		1233	}
254		1234
		1235	BEGIN { AnyEvent::common_sense }
		1236
255	use Carp;	1237	use Carp ();
256		1238
257	our $VERSION = '2.9';	1239	our $VERSION = '6.14';
258	our $MODEL;	1240	our $MODEL;
259
260	our $AUTOLOAD;
261	our @ISA;	1241	our @ISA;
262
263	our $verbose = $ENV{PERL_ANYEVENT_VERBOSE}*1;
264
265	our @REGISTRY;	1242	our @REGISTRY;
		1243	our $VERBOSE;
		1244	our %PROTOCOL; # (ipv4|ipv6) => (1|2), higher numbers are preferred
		1245	our $MAX_SIGNAL_LATENCY = $ENV{PERL_ANYEVENT_MAX_SIGNAL_LATENCY} || 10; # executes after the BEGIN block below (tainting!)
266		1246
		1247	BEGIN {
		1248	require "AnyEvent/constants.pl";
		1249
		1250	eval "sub TAINT (){" . (${^TAINT}*1) . "}";
		1251
		1252	delete @ENV{grep /^PERL_ANYEVENT_/, keys %ENV}
		1253	if ${^TAINT};
		1254
		1255	$ENV{"PERL_ANYEVENT_$_"} = $ENV{"AE_$_"}
		1256	for grep s/^AE_// && !exists $ENV{"PERL_ANYEVENT_$_"}, keys %ENV;
		1257
		1258	@ENV{grep /^PERL_ANYEVENT_/, keys %ENV} = ()
		1259	if ${^TAINT};
		1260
		1261	# $ENV{PERL_ANYEVENT_xxx} now valid
		1262
		1263	$VERBOSE = length $ENV{PERL_ANYEVENT_VERBOSE} ? $ENV{PERL_ANYEVENT_VERBOSE}*1 : 4;
		1264
		1265	my $idx;
		1266	$PROTOCOL{$_} = ++$idx
		1267	for reverse split /\s*,\s*/,
		1268	$ENV{PERL_ANYEVENT_PROTOCOLS} || "ipv4,ipv6";
		1269	}
		1270
		1271	our @post_detect;
		1272
		1273	sub post_detect(&) {
		1274	my ($cb) = @_;
		1275
		1276	push @post_detect, $cb;
		1277
		1278	defined wantarray
		1279	? bless \$cb, "AnyEvent::Util::postdetect"
		1280	: ()
		1281	}
		1282
		1283	sub AnyEvent::Util::postdetect::DESTROY {
		1284	@post_detect = grep $_ != ${$_[0]}, @post_detect;
		1285	}
		1286
		1287	our $POSTPONE_W;
		1288	our @POSTPONE;
		1289
		1290	sub _postpone_exec {
		1291	undef $POSTPONE_W;
		1292
		1293	&{ shift @POSTPONE }
		1294	while @POSTPONE;
		1295	}
		1296
		1297	sub postpone(&) {
		1298	push @POSTPONE, shift;
		1299
		1300	$POSTPONE_W ||= AE::timer (0, 0, \&_postpone_exec);
		1301
		1302	()
		1303	}
		1304
		1305	sub log($$;@) {
		1306	# only load the big bloated module when we actually are about to log something
		1307	if ($_[0] <= ($VERBOSE || 1)) { # also catches non-numeric levels(!) and fatal
		1308	local ($!, $@);
		1309	require AnyEvent::Log; # among other things, sets $VERBOSE to 9
		1310	# AnyEvent::Log overwrites this function
		1311	goto &log;
		1312	}
		1313
		1314	0 # not logged
		1315	}
		1316
		1317	sub _logger($;$) {
		1318	my ($level, $renabled) = @_;
		1319
		1320	$$renabled = $level <= $VERBOSE;
		1321
		1322	my $logger = [(caller)[0], $level, $renabled];
		1323
		1324	$AnyEvent::Log::LOGGER{$logger+0} = $logger;
		1325
		1326	# return unless defined wantarray;
		1327	#
		1328	# require AnyEvent::Util;
		1329	# my $guard = AnyEvent::Util::guard (sub {
		1330	# # "clean up"
		1331	# delete $LOGGER{$logger+0};
		1332	# });
		1333	#
		1334	# sub {
		1335	# return 0 unless $$renabled;
		1336	#
		1337	# $guard if 0; # keep guard alive, but don't cause runtime overhead
		1338	# require AnyEvent::Log unless $AnyEvent::Log::VERSION;
		1339	# package AnyEvent::Log;
		1340	# _log ($logger->[0], $level, @_) # logger->[0] has been converted at load time
		1341	# }
		1342	}
		1343
		1344	if (length $ENV{PERL_ANYEVENT_LOG}) {
		1345	require AnyEvent::Log; # AnyEvent::Log does the thing for us
		1346	}
		1347
267	my @models = (	1348	our @models = (
268	[Coro::EV:: => AnyEvent::Impl::CoroEV::],
269	[EV:: => AnyEvent::Impl::EV::],	1349	[EV:: => AnyEvent::Impl::EV::],
270	[Coro::Event:: => AnyEvent::Impl::CoroEvent::],	1350	[AnyEvent::Loop:: => AnyEvent::Impl::Perl::],
		1351	# everything below here will not (normally) be autoprobed
		1352	# as the pure perl backend should work everywhere
		1353	# and is usually faster
		1354	[Irssi:: => AnyEvent::Impl::Irssi::], # Irssi has a bogus "Event" package, so msut be near the top
271	[Event:: => AnyEvent::Impl::Event::],	1355	[Event:: => AnyEvent::Impl::Event::], # slow, stable
272	[Glib:: => AnyEvent::Impl::Glib::],	1356	[Glib:: => AnyEvent::Impl::Glib::], # becomes extremely slow with many watchers
		1357	# everything below here should not be autoloaded
		1358	[Event::Lib:: => AnyEvent::Impl::EventLib::], # too buggy
		1359	[Tk:: => AnyEvent::Impl::Tk::], # crashes with many handles
		1360	[Qt:: => AnyEvent::Impl::Qt::], # requires special main program
		1361	[POE::Kernel:: => AnyEvent::Impl::POE::], # lasciate ogni speranza
273	[Tk:: => AnyEvent::Impl::Tk::],	1362	[Wx:: => AnyEvent::Impl::POE::],
274	[AnyEvent::Impl::Perl:: => AnyEvent::Impl::Perl::],	1363	[Prima:: => AnyEvent::Impl::POE::],
		1364	[IO::Async::Loop:: => AnyEvent::Impl::IOAsync::], # a bitch to autodetect
		1365	[Cocoa::EventLoop:: => AnyEvent::Impl::Cocoa::],
		1366	[FLTK:: => AnyEvent::Impl::FLTK::],
275	);	1367	);
276		1368
277	our %method = map +($_ => 1), qw(io timer condvar broadcast wait signal one_event DESTROY);	1369	our @isa_hook;
		1370
		1371	sub _isa_set {
		1372	my @pkg = ("AnyEvent", (map $_->[0], grep defined, @isa_hook), $MODEL);
		1373
		1374	@{"$pkg[$_-1]::ISA"} = $pkg[$_]
		1375	for 1 .. $#pkg;
		1376
		1377	grep $_ && $_->[1], @isa_hook
		1378	and AE::_reset ();
		1379	}
		1380
		1381	# used for hooking AnyEvent::Strict and AnyEvent::Debug::Wrap into the class hierarchy
		1382	sub _isa_hook($$;$) {
		1383	my ($i, $pkg, $reset_ae) = @_;
		1384
		1385	$isa_hook[$i] = $pkg ? [$pkg, $reset_ae] : undef;
		1386
		1387	_isa_set;
		1388	}
		1389
		1390	# all autoloaded methods reserve the complete glob, not just the method slot.
		1391	# due to bugs in perls method cache implementation.
		1392	our @methods = qw(io timer time now now_update signal child idle condvar);
278		1393
279	sub detect() {	1394	sub detect() {
		1395	return $MODEL if $MODEL; # some programs keep references to detect
		1396
		1397	# IO::Async::Loop::AnyEvent is extremely evil, refuse to work with it
		1398	# the author knows about the problems and what it does to AnyEvent as a whole
		1399	# (and the ability of others to use AnyEvent), but simply wants to abuse AnyEvent
		1400	# anyway.
		1401	AnyEvent::log fatal => "IO::Async::Loop::AnyEvent detected - that module is broken by\n"
		1402	. "design, abuses internals and breaks AnyEvent - will not continue."
		1403	if exists $INC{"IO/Async/Loop/AnyEvent.pm"};
		1404
		1405	local $!; # for good measure
		1406	local $SIG{__DIE__}; # we use eval
		1407
		1408	# free some memory
		1409	*detect = sub () { $MODEL };
		1410	# undef &func doesn't correctly update the method cache. grmbl.
		1411	# so we delete the whole glob. grmbl.
		1412	# otoh, perl doesn't let me undef an active usb, but it lets me free
		1413	# a glob with an active sub. hrm. i hope it works, but perl is
		1414	# usually buggy in this department. sigh.
		1415	delete @{"AnyEvent::"}{@methods};
		1416	undef @methods;
		1417
		1418	if ($ENV{PERL_ANYEVENT_MODEL} =~ /^([a-zA-Z0-9:]+)$/) {
		1419	my $model = $1;
		1420	$model = "AnyEvent::Impl::$model" unless $model =~ s/::$//;
		1421	if (eval "require $model") {
		1422	AnyEvent::log 7 => "Loaded model '$model' (forced by \$ENV{PERL_ANYEVENT_MODEL}), using it.";
		1423	$MODEL = $model;
		1424	} else {
		1425	AnyEvent::log 4 => "Unable to load model '$model' (from \$ENV{PERL_ANYEVENT_MODEL}):\n$@";
		1426	}
		1427	}
		1428
		1429	# check for already loaded models
280	unless ($MODEL) {	1430	unless ($MODEL) {
281	no strict 'refs';
282
283	# check for already loaded models
284	for (@REGISTRY, @models) {	1431	for (@REGISTRY, @models) {
285	my ($package, $model) = @$_;	1432	my ($package, $model) = @$_;
286	if (${"$package\::VERSION"} > 0) {	1433	if (${"$package\::VERSION"} > 0) {
287	if (eval "require $model") {	1434	if (eval "require $model") {
		1435	AnyEvent::log 7 => "Autodetected model '$model', using it.";
288	$MODEL = $model;	1436	$MODEL = $model;
289	warn "AnyEvent: found model '$model', using it.\n" if $verbose > 1;	1437	last;
		1438	} else {
		1439	AnyEvent::log 8 => "Detected event loop $package, but cannot load '$model', skipping: $@";
		1440	}
		1441	}
		1442	}
		1443
		1444	unless ($MODEL) {
		1445	# try to autoload a model
		1446	for (@REGISTRY, @models) {
		1447	my ($package, $model) = @$_;
		1448	if (
		1449	eval "require $package"
		1450	and ${"$package\::VERSION"} > 0
		1451	and eval "require $model"
		1452	) {
		1453	AnyEvent::log 7 => "Autoloaded model '$model', using it.";
		1454	$MODEL = $model;
290	last;	1455	last;
291	}	1456	}
292	}	1457	}
		1458
		1459	$MODEL
		1460	or AnyEvent::log fatal => "Backend autodetection failed - did you properly install AnyEvent?";
293	}	1461	}
		1462	}
294		1463
295	unless ($MODEL) {	1464	# free memory only needed for probing
296	# try to load a model	1465	undef @models;
		1466	undef @REGISTRY;
297		1467
298	for (@REGISTRY, @models) {	1468	push @{"$MODEL\::ISA"}, "AnyEvent::Base";
299	my ($package, $model) = @$_;	1469
300	if (eval "require $package"	1470	# now nuke some methods that are overridden by the backend.
301	and ${"$package\::VERSION"} > 0	1471	# SUPER usage is not allowed in these.
302	and eval "require $model") {	1472	for (qw(time signal child idle)) {
303	$MODEL = $model;	1473	undef &{"AnyEvent::Base::$_"}
304	warn "AnyEvent: autoprobed and loaded model '$model', using it.\n" if $verbose > 1;	1474	if defined &{"$MODEL\::$_"};
		1475	}
		1476
		1477	_isa_set;
		1478
		1479	# we're officially open!
		1480
		1481	if ($ENV{PERL_ANYEVENT_STRICT}) {
		1482	require AnyEvent::Strict;
		1483	}
		1484
		1485	if ($ENV{PERL_ANYEVENT_DEBUG_WRAP}) {
		1486	require AnyEvent::Debug;
		1487	AnyEvent::Debug::wrap ($ENV{PERL_ANYEVENT_DEBUG_WRAP});
		1488	}
		1489
		1490	if (length $ENV{PERL_ANYEVENT_DEBUG_SHELL}) {
		1491	require AnyEvent::Socket;
		1492	require AnyEvent::Debug;
		1493
		1494	my $shell = $ENV{PERL_ANYEVENT_DEBUG_SHELL};
		1495	$shell =~ s/\$\$/$$/g;
		1496
		1497	my ($host, $service) = AnyEvent::Socket::parse_hostport ($shell);
		1498	$AnyEvent::Debug::SHELL = AnyEvent::Debug::shell ($host, $service);
		1499	}
		1500
		1501	# now the anyevent environment is set up as the user told us to, so
		1502	# call the actual user code - post detects
		1503
		1504	(shift @post_detect)->() while @post_detect;
		1505	undef @post_detect;
		1506
		1507	*post_detect = sub(&) {
		1508	shift->();
		1509
		1510	undef
		1511	};
		1512
		1513	$MODEL
		1514	}
		1515
		1516	for my $name (@methods) {
		1517	*$name = sub {
		1518	detect;
		1519	# we use goto because
		1520	# a) it makes the thunk more transparent
		1521	# b) it allows us to delete the thunk later
		1522	goto &{ UNIVERSAL::can AnyEvent => "SUPER::$name" }
		1523	};
		1524	}
		1525
		1526	# utility function to dup a filehandle. this is used by many backends
		1527	# to support binding more than one watcher per filehandle (they usually
		1528	# allow only one watcher per fd, so we dup it to get a different one).
		1529	sub _dupfh($$;$$) {
		1530	my ($poll, $fh, $r, $w) = @_;
		1531
		1532	# cygwin requires the fh mode to be matching, unix doesn't
		1533	my ($rw, $mode) = $poll eq "r" ? ($r, "<&") : ($w, ">&");
		1534
		1535	open my $fh2, $mode, $fh
		1536	or die "AnyEvent->io: cannot dup() filehandle in mode '$poll': $!,";
		1537
		1538	# we assume CLOEXEC is already set by perl in all important cases
		1539
		1540	($fh2, $rw)
		1541	}
		1542
		1543	=head1 SIMPLIFIED AE API
		1544
		1545	Starting with version 5.0, AnyEvent officially supports a second, much
		1546	simpler, API that is designed to reduce the calling, typing and memory
		1547	overhead by using function call syntax and a fixed number of parameters.
		1548
		1549	See the L<AE> manpage for details.
		1550
		1551	=cut
		1552
		1553	package AE;
		1554
		1555	our $VERSION = $AnyEvent::VERSION;
		1556
		1557	sub _reset() {
		1558	eval q{
		1559	# fall back to the main API by default - backends and AnyEvent::Base
		1560	# implementations can overwrite these.
		1561
		1562	sub io($$$) {
		1563	AnyEvent->io (fh => $_[0], poll => $_[1] ? "w" : "r", cb => $_[2])
		1564	}
		1565
		1566	sub timer($$$) {
		1567	AnyEvent->timer (after => $_[0], interval => $_[1], cb => $_[2])
		1568	}
		1569
		1570	sub signal($$) {
		1571	AnyEvent->signal (signal => $_[0], cb => $_[1])
		1572	}
		1573
		1574	sub child($$) {
		1575	AnyEvent->child (pid => $_[0], cb => $_[1])
		1576	}
		1577
		1578	sub idle($) {
		1579	AnyEvent->idle (cb => $_[0]);
		1580	}
		1581
		1582	sub cv(;&) {
		1583	AnyEvent->condvar (@_ ? (cb => $_[0]) : ())
		1584	}
		1585
		1586	sub now() {
		1587	AnyEvent->now
		1588	}
		1589
		1590	sub now_update() {
		1591	AnyEvent->now_update
		1592	}
		1593
		1594	sub time() {
		1595	AnyEvent->time
		1596	}
		1597
		1598	*postpone = \&AnyEvent::postpone;
		1599	*log = \&AnyEvent::log;
		1600	};
		1601	die if $@;
		1602	}
		1603
		1604	BEGIN { _reset }
		1605
		1606	package AnyEvent::Base;
		1607
		1608	# default implementations for many methods
		1609
		1610	sub time {
		1611	eval q{ # poor man's autoloading {}
		1612	# probe for availability of Time::HiRes
		1613	if (eval "use Time::HiRes (); Time::HiRes::time (); 1") {
		1614	*time = sub { Time::HiRes::time () };
		1615	*AE::time = \& Time::HiRes::time ;
		1616	*now = \&time;
		1617	AnyEvent::log 8 => "using Time::HiRes for sub-second timing accuracy.";
		1618	# if (eval "use POSIX (); (POSIX::times())...
		1619	} else {
		1620	*time = sub { CORE::time };
		1621	*AE::time = sub (){ CORE::time };
		1622	*now = \&time;
		1623	AnyEvent::log 3 => "Using built-in time(), no sub-second resolution!";
		1624	}
		1625	};
		1626	die if $@;
		1627
		1628	&time
		1629	}
		1630
		1631	*now = \&time;
		1632	sub now_update { }
		1633
		1634	sub _poll {
		1635	Carp::croak "$AnyEvent::MODEL does not support blocking waits. Caught";
		1636	}
		1637
		1638	# default implementation for ->condvar
		1639	# in fact, the default should not be overwritten
		1640
		1641	sub condvar {
		1642	eval q{ # poor man's autoloading {}
		1643	*condvar = sub {
		1644	bless { @_ == 3 ? (_ae_cb => $_[2]) : () }, "AnyEvent::CondVar"
		1645	};
		1646
		1647	*AE::cv = sub (;&) {
		1648	bless { @_ ? (_ae_cb => shift) : () }, "AnyEvent::CondVar"
		1649	};
		1650	};
		1651	die if $@;
		1652
		1653	&condvar
		1654	}
		1655
		1656	# default implementation for ->signal
		1657
		1658	our $HAVE_ASYNC_INTERRUPT;
		1659
		1660	sub _have_async_interrupt() {
		1661	$HAVE_ASYNC_INTERRUPT = 1*(!$ENV{PERL_ANYEVENT_AVOID_ASYNC_INTERRUPT}
		1662	&& eval "use Async::Interrupt 1.02 (); 1")
		1663	unless defined $HAVE_ASYNC_INTERRUPT;
		1664
		1665	$HAVE_ASYNC_INTERRUPT
		1666	}
		1667
		1668	our ($SIGPIPE_R, $SIGPIPE_W, %SIG_CB, %SIG_EV, $SIG_IO);
		1669	our (%SIG_ASY, %SIG_ASY_W);
		1670	our ($SIG_COUNT, $SIG_TW);
		1671
		1672	# install a dummy wakeup watcher to reduce signal catching latency
		1673	# used by Impls
		1674	sub _sig_add() {
		1675	unless ($SIG_COUNT++) {
		1676	# try to align timer on a full-second boundary, if possible
		1677	my $NOW = AE::now;
		1678
		1679	$SIG_TW = AE::timer
		1680	$MAX_SIGNAL_LATENCY - ($NOW - int $NOW),
		1681	$MAX_SIGNAL_LATENCY,
		1682	sub { } # just for the PERL_ASYNC_CHECK
		1683	;
		1684	}
		1685	}
		1686
		1687	sub _sig_del {
		1688	undef $SIG_TW
		1689	unless --$SIG_COUNT;
		1690	}
		1691
		1692	our $_sig_name_init; $_sig_name_init = sub {
		1693	eval q{ # poor man's autoloading {}
		1694	undef $_sig_name_init;
		1695
		1696	if (_have_async_interrupt) {
		1697	*sig2num = \&Async::Interrupt::sig2num;
		1698	*sig2name = \&Async::Interrupt::sig2name;
		1699	} else {
		1700	require Config;
		1701
		1702	my %signame2num;
		1703	@signame2num{ split ' ', $Config::Config{sig_name} }
		1704	= split ' ', $Config::Config{sig_num};
		1705
		1706	my @signum2name;
		1707	@signum2name[values %signame2num] = keys %signame2num;
		1708
		1709	*sig2num = sub($) {
		1710	$_[0] > 0 ? shift : $signame2num{+shift}
		1711	};
		1712	*sig2name = sub ($) {
		1713	$_[0] > 0 ? $signum2name[+shift] : shift
		1714	};
		1715	}
		1716	};
		1717	die if $@;
		1718	};
		1719
		1720	sub sig2num ($) { &$_sig_name_init; &sig2num }
		1721	sub sig2name($) { &$_sig_name_init; &sig2name }
		1722
		1723	sub signal {
		1724	eval q{ # poor man's autoloading {}
		1725	# probe for availability of Async::Interrupt
		1726	if (_have_async_interrupt) {
		1727	AnyEvent::log 8 => "Using Async::Interrupt for race-free signal handling.";
		1728
		1729	$SIGPIPE_R = new Async::Interrupt::EventPipe;
		1730	$SIG_IO = AE::io $SIGPIPE_R->fileno, 0, \&_signal_exec;
		1731
		1732	} else {
		1733	AnyEvent::log 8 => "Using emulated perl signal handling with latency timer.";
		1734
		1735	if (AnyEvent::WIN32) {
		1736	require AnyEvent::Util;
		1737
		1738	($SIGPIPE_R, $SIGPIPE_W) = AnyEvent::Util::portable_pipe ();
		1739	AnyEvent::Util::fh_nonblocking ($SIGPIPE_R, 1) if $SIGPIPE_R;
		1740	AnyEvent::Util::fh_nonblocking ($SIGPIPE_W, 1) if $SIGPIPE_W; # just in case
		1741	} else {
		1742	pipe $SIGPIPE_R, $SIGPIPE_W;
		1743	fcntl $SIGPIPE_R, AnyEvent::F_SETFL, AnyEvent::O_NONBLOCK if $SIGPIPE_R;
		1744	fcntl $SIGPIPE_W, AnyEvent::F_SETFL, AnyEvent::O_NONBLOCK if $SIGPIPE_W; # just in case
		1745
		1746	# not strictly required, as $^F is normally 2, but let's make sure...
		1747	fcntl $SIGPIPE_R, AnyEvent::F_SETFD, AnyEvent::FD_CLOEXEC;
		1748	fcntl $SIGPIPE_W, AnyEvent::F_SETFD, AnyEvent::FD_CLOEXEC;
		1749	}
		1750
		1751	$SIGPIPE_R
		1752	or Carp::croak "AnyEvent: unable to create a signal reporting pipe: $!\n";
		1753
		1754	$SIG_IO = AE::io $SIGPIPE_R, 0, \&_signal_exec;
		1755	}
		1756
		1757	*signal = $HAVE_ASYNC_INTERRUPT
		1758	? sub {
		1759	my (undef, %arg) = @_;
		1760
		1761	# async::interrupt
		1762	my $signal = sig2num $arg{signal};
		1763	$SIG_CB{$signal}{$arg{cb}} = $arg{cb};
		1764
		1765	$SIG_ASY{$signal} ||= new Async::Interrupt
		1766	cb => sub { undef $SIG_EV{$signal} },
		1767	signal => $signal,
		1768	pipe => [$SIGPIPE_R->filenos],
		1769	pipe_autodrain => 0,
305	last;	1770	;
		1771
		1772	bless [$signal, $arg{cb}], "AnyEvent::Base::signal"
		1773	}
		1774	: sub {
		1775	my (undef, %arg) = @_;
		1776
		1777	# pure perl
		1778	my $signal = sig2name $arg{signal};
		1779	$SIG_CB{$signal}{$arg{cb}} = $arg{cb};
		1780
		1781	$SIG{$signal} ||= sub {
		1782	local $!;
		1783	syswrite $SIGPIPE_W, "\x00", 1 unless %SIG_EV;
		1784	undef $SIG_EV{$signal};
		1785	};
		1786
		1787	# can't do signal processing without introducing races in pure perl,
		1788	# so limit the signal latency.
		1789	_sig_add;
		1790
		1791	bless [$signal, $arg{cb}], "AnyEvent::Base::signal"
		1792	}
		1793	;
		1794
		1795	*AnyEvent::Base::signal::DESTROY = sub {
		1796	my ($signal, $cb) = @{$_[0]};
		1797
		1798	_sig_del;
		1799
		1800	delete $SIG_CB{$signal}{$cb};
		1801
		1802	$HAVE_ASYNC_INTERRUPT
		1803	? delete $SIG_ASY{$signal}
		1804	: # delete doesn't work with older perls - they then
		1805	# print weird messages, or just unconditionally exit
		1806	# instead of getting the default action.
		1807	undef $SIG{$signal}
		1808	unless keys %{ $SIG_CB{$signal} };
		1809	};
		1810
		1811	*_signal_exec = sub {
		1812	$HAVE_ASYNC_INTERRUPT
		1813	? $SIGPIPE_R->drain
		1814	: sysread $SIGPIPE_R, (my $dummy), 9;
		1815
		1816	while (%SIG_EV) {
		1817	for (keys %SIG_EV) {
		1818	delete $SIG_EV{$_};
		1819	&$_ for values %{ $SIG_CB{$_} || {} };
306	}	1820	}
307	}	1821	}
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	}	1822	};
312
313	unshift @ISA, $MODEL;
314	push @{"$MODEL\::ISA"}, "AnyEvent::Base";
315	}
316
317	$MODEL
318	}
319
320	sub AUTOLOAD {
321	(my $func = $AUTOLOAD) =~ s/.*://;
322
323	$method{$func}
324	or croak "$func: not a valid method for AnyEvent objects";
325
326	detect unless $MODEL;
327
328	my $class = shift;
329	$class->$func (@_);
330	}
331
332	package AnyEvent::Base;
333
334	# default implementation for ->condvar, ->wait, ->broadcast
335
336	sub condvar {
337	bless \my $flag, "AnyEvent::Base::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	}
347
348	# default implementation for ->signal
349
350	our %SIG_CB;
351
352	sub signal {
353	my (undef, %arg) = @_;
354
355	my $signal = uc $arg{signal}
356	or Carp::croak "required option 'signal' is missing";
357
358	$SIG_CB{$signal}{$arg{cb}} = $arg{cb};
359	$SIG{$signal} ||= sub {
360	$_->() for values %{ $SIG_CB{$signal} || {} };
361	};	1823	};
		1824	die if $@;
362		1825
363	bless [$signal, $arg{cb}], "AnyEvent::Base::Signal"	1826	&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	}	1827	}
373		1828
374	# default implementation for ->child	1829	# default implementation for ->child
375		1830
376	our %PID_CB;	1831	our %PID_CB;
377	our $CHLD_W;	1832	our $CHLD_W;
378	our $CHLD_DELAY_W;	1833	our $CHLD_DELAY_W;
379	our $PID_IDLE;
380	our $WNOHANG;
381		1834
382	sub _child_wait {	1835	# used by many Impl's
383	while (0 < (my $pid = waitpid -1, $WNOHANG)) {	1836	sub _emit_childstatus($$) {
		1837	my (undef, $rpid, $rstatus) = @_;
		1838
		1839	$_->($rpid, $rstatus)
384	$_->($pid, $?) for (values %{ $PID_CB{$pid} || {} }),	1840	for values %{ $PID_CB{$rpid} || {} },
385	(values %{ $PID_CB{0} || {} });	1841	values %{ $PID_CB{0} || {} };
		1842	}
		1843
		1844	sub child {
		1845	eval q{ # poor man's autoloading {}
		1846	*_sigchld = sub {
		1847	my $pid;
		1848
		1849	AnyEvent->_emit_childstatus ($pid, $?)
		1850	while ($pid = waitpid -1, WNOHANG) > 0;
		1851	};
		1852
		1853	*child = sub {
		1854	my (undef, %arg) = @_;
		1855
		1856	my $pid = $arg{pid};
		1857	my $cb = $arg{cb};
		1858
		1859	$PID_CB{$pid}{$cb+0} = $cb;
		1860
		1861	unless ($CHLD_W) {
		1862	$CHLD_W = AE::signal CHLD => \&_sigchld;
		1863	# child could be a zombie already, so make at least one round
		1864	&_sigchld;
		1865	}
		1866
		1867	bless [$pid, $cb+0], "AnyEvent::Base::child"
		1868	};
		1869
		1870	*AnyEvent::Base::child::DESTROY = sub {
		1871	my ($pid, $icb) = @{$_[0]};
		1872
		1873	delete $PID_CB{$pid}{$icb};
		1874	delete $PID_CB{$pid} unless keys %{ $PID_CB{$pid} };
		1875
		1876	undef $CHLD_W unless keys %PID_CB;
		1877	};
		1878	};
		1879	die if $@;
		1880
		1881	&child
		1882	}
		1883
		1884	# idle emulation is done by simply using a timer, regardless
		1885	# of whether the process is idle or not, and not letting
		1886	# the callback use more than 50% of the time.
		1887	sub idle {
		1888	eval q{ # poor man's autoloading {}
		1889	*idle = sub {
		1890	my (undef, %arg) = @_;
		1891
		1892	my ($cb, $w, $rcb) = $arg{cb};
		1893
		1894	$rcb = sub {
		1895	if ($cb) {
		1896	$w = AE::time;
		1897	&$cb;
		1898	$w = AE::time - $w;
		1899
		1900	# never use more then 50% of the time for the idle watcher,
		1901	# within some limits
		1902	$w = 0.0001 if $w < 0.0001;
		1903	$w = 5 if $w > 5;
		1904
		1905	$w = AE::timer $w, 0, $rcb;
		1906	} else {
		1907	# clean up...
		1908	undef $w;
		1909	undef $rcb;
		1910	}
		1911	};
		1912
		1913	$w = AE::timer 0.05, 0, $rcb;
		1914
		1915	bless \\$cb, "AnyEvent::Base::idle"
		1916	};
		1917
		1918	*AnyEvent::Base::idle::DESTROY = sub {
		1919	undef $${$_[0]};
		1920	};
		1921	};
		1922	die if $@;
		1923
		1924	&idle
		1925	}
		1926
		1927	package AnyEvent::CondVar;
		1928
		1929	our @ISA = AnyEvent::CondVar::Base::;
		1930
		1931	# only to be used for subclassing
		1932	sub new {
		1933	my $class = shift;
		1934	bless AnyEvent->condvar (@_), $class
		1935	}
		1936
		1937	package AnyEvent::CondVar::Base;
		1938
		1939	#use overload
		1940	# '&{}' => sub { my $self = shift; sub { $self->send (@_) } },
		1941	# fallback => 1;
		1942
		1943	# save 300+ kilobytes by dirtily hardcoding overloading
		1944	${"AnyEvent::CondVar::Base::OVERLOAD"}{dummy}++; # Register with magic by touching.
		1945	*{'AnyEvent::CondVar::Base::()'} = sub { }; # "Make it findable via fetchmethod."
		1946	*{'AnyEvent::CondVar::Base::(&{}'} = sub { my $self = shift; sub { $self->send (@_) } }; # &{}
		1947	${'AnyEvent::CondVar::Base::()'} = 1; # fallback
		1948
		1949	our $WAITING;
		1950
		1951	sub _send {
		1952	# nop
		1953	}
		1954
		1955	sub _wait {
		1956	AnyEvent->_poll until $_[0]{_ae_sent};
		1957	}
		1958
		1959	sub send {
		1960	my $cv = shift;
		1961	$cv->{_ae_sent} = [@_];
		1962	(delete $cv->{_ae_cb})->($cv) if $cv->{_ae_cb};
		1963	$cv->_send;
		1964	}
		1965
		1966	sub croak {
		1967	$_[0]{_ae_croak} = $_[1];
		1968	$_[0]->send;
		1969	}
		1970
		1971	sub ready {
		1972	$_[0]{_ae_sent}
		1973	}
		1974
		1975	sub recv {
		1976	unless ($_[0]{_ae_sent}) {
		1977	$WAITING
		1978	and Carp::croak "AnyEvent::CondVar: recursive blocking wait attempted";
		1979
		1980	local $WAITING = 1;
		1981	$_[0]->_wait;
386	}	1982	}
387		1983
388	undef $PID_IDLE;	1984	$_[0]{_ae_croak}
389	}	1985	and Carp::croak $_[0]{_ae_croak};
390		1986
391	sub _sigchld {	1987	wantarray
392	# make sure we deliver these changes "synchronous" with the event loop.	1988	? @{ $_[0]{_ae_sent} }
393	$CHLD_DELAY_W ||= AnyEvent->timer (after => 0, cb => sub {	1989	: $_[0]{_ae_sent}[0]
394	undef $CHLD_DELAY_W;
395	&_child_wait;
396	});
397	}	1990	}
398		1991
399	sub child {	1992	sub cb {
400	my (undef, %arg) = @_;	1993	my $cv = shift;
401		1994
402	defined (my $pid = $arg{pid} + 0)	1995	@_
403	or Carp::croak "required option 'pid' is missing";	1996	and $cv->{_ae_cb} = shift
		1997	and $cv->{_ae_sent}
		1998	and (delete $cv->{_ae_cb})->($cv);
404		1999
405	$PID_CB{$pid}{$arg{cb}} = $arg{cb};	2000	$cv->{_ae_cb}
406
407	unless ($WNOHANG) {
408	$WNOHANG = eval { require POSIX; &POSIX::WNOHANG } || 1;
409	}
410
411	unless ($CHLD_W) {
412	$CHLD_W = AnyEvent->signal (signal => 'CHLD', cb => \&_sigchld);
413	# child could be a zombie already, so make at least one round
414	&_sigchld;
415	}
416
417	bless [$pid, $arg{cb}], "AnyEvent::Base::Child"
418	}	2001	}
419		2002
420	sub AnyEvent::Base::Child::DESTROY {	2003	sub begin {
421	my ($pid, $cb) = @{$_[0]};	2004	++$_[0]{_ae_counter};
422		2005	$_[0]{_ae_end_cb} = $_[1] if @_ > 1;
423	delete $PID_CB{$pid}{$cb};
424	delete $PID_CB{$pid} unless keys %{ $PID_CB{$pid} };
425
426	undef $CHLD_W unless keys %PID_CB;
427	}	2006	}
		2007
		2008	sub end {
		2009	return if --$_[0]{_ae_counter};
		2010	&{ $_[0]{_ae_end_cb} || sub { $_[0]->send } };
		2011	}
		2012
		2013	# undocumented/compatibility with pre-3.4
		2014	*broadcast = \&send;
		2015	*wait = \&recv;
		2016
		2017	=head1 ERROR AND EXCEPTION HANDLING
		2018
		2019	In general, AnyEvent does not do any error handling - it relies on the
		2020	caller to do that if required. The L<AnyEvent::Strict> module (see also
		2021	the C<PERL_ANYEVENT_STRICT> environment variable, below) provides strict
		2022	checking of all AnyEvent methods, however, which is highly useful during
		2023	development.
		2024
		2025	As for exception handling (i.e. runtime errors and exceptions thrown while
		2026	executing a callback), this is not only highly event-loop specific, but
		2027	also not in any way wrapped by this module, as this is the job of the main
		2028	program.
		2029
		2030	The pure perl event loop simply re-throws the exception (usually
		2031	within C<< condvar->recv >>), the L<Event> and L<EV> modules call C<<
		2032	$Event/EV::DIED->() >>, L<Glib> uses C<< install_exception_handler >> and
		2033	so on.
		2034
		2035	=head1 ENVIRONMENT VARIABLES
		2036
		2037	AnyEvent supports a number of environment variables that tune the
		2038	runtime behaviour. They are usually evaluated when AnyEvent is
		2039	loaded, initialised, or a submodule that uses them is loaded. Many of
		2040	them also cause AnyEvent to load additional modules - for example,
		2041	C<PERL_ANYEVENT_DEBUG_WRAP> causes the L<AnyEvent::Debug> module to be
		2042	loaded.
		2043
		2044	All the environment variables documented here start with
		2045	C<PERL_ANYEVENT_>, which is what AnyEvent considers its own
		2046	namespace. Other modules are encouraged (but by no means required) to use
		2047	C<PERL_ANYEVENT_SUBMODULE> if they have registered the AnyEvent::Submodule
		2048	namespace on CPAN, for any submodule. For example, L<AnyEvent::HTTP> could
		2049	be expected to use C<PERL_ANYEVENT_HTTP_PROXY> (it should not access env
		2050	variables starting with C<AE_>, see below).
		2051
		2052	All variables can also be set via the C<AE_> prefix, that is, instead
		2053	of setting C<PERL_ANYEVENT_VERBOSE> you can also set C<AE_VERBOSE>. In
		2054	case there is a clash btween anyevent and another program that uses
		2055	C<AE_something> you can set the corresponding C<PERL_ANYEVENT_something>
		2056	variable to the empty string, as those variables take precedence.
		2057
		2058	When AnyEvent is first loaded, it copies all C<AE_xxx> env variables
		2059	to their C<PERL_ANYEVENT_xxx> counterpart unless that variable already
		2060	exists. If taint mode is on, then AnyEvent will remove I<all> environment
		2061	variables starting with C<PERL_ANYEVENT_> from C<%ENV> (or replace them
		2062	with C<undef> or the empty string, if the corresaponding C<AE_> variable
		2063	is set).
		2064
		2065	The exact algorithm is currently:
		2066
		2067	1. if taint mode enabled, delete all PERL_ANYEVENT_xyz variables from %ENV
		2068	2. copy over AE_xyz to PERL_ANYEVENT_xyz unless the latter alraedy exists
		2069	3. if taint mode enabled, set all PERL_ANYEVENT_xyz variables to undef.
		2070
		2071	This ensures that child processes will not see the C<AE_> variables.
		2072
		2073	The following environment variables are currently known to AnyEvent:
		2074
		2075	=over 4
		2076
		2077	=item C<PERL_ANYEVENT_VERBOSE>
		2078
		2079	By default, AnyEvent will log messages with loglevel C<4> (C<error>) or
		2080	higher (see L<AnyEvent::Log>). You can set this environment variable to a
		2081	numerical loglevel to make AnyEvent more (or less) talkative.
		2082
		2083	If you want to do more than just set the global logging level
		2084	you should have a look at C<PERL_ANYEVENT_LOG>, which allows much more
		2085	complex specifications.
		2086
		2087	When set to C<0> (C<off>), then no messages whatsoever will be logged with
		2088	everything else at defaults.
		2089
		2090	When set to C<5> or higher (C<warn>), AnyEvent warns about unexpected
		2091	conditions, such as not being able to load the event model specified by
		2092	C<PERL_ANYEVENT_MODEL>, or a guard callback throwing an exception - this
		2093	is the minimum recommended level for use during development.
		2094
		2095	When set to C<7> or higher (info), AnyEvent reports which event model it
		2096	chooses.
		2097
		2098	When set to C<8> or higher (debug), then AnyEvent will report extra
		2099	information on which optional modules it loads and how it implements
		2100	certain features.
		2101
		2102	=item C<PERL_ANYEVENT_LOG>
		2103
		2104	Accepts rather complex logging specifications. For example, you could log
		2105	all C<debug> messages of some module to stderr, warnings and above to
		2106	stderr, and errors and above to syslog, with:
		2107
		2108	PERL_ANYEVENT_LOG=Some::Module=debug,+log:filter=warn,+%syslog:%syslog=error,syslog
		2109
		2110	For the rather extensive details, see L<AnyEvent::Log>.
		2111
		2112	This variable is evaluated when AnyEvent (or L<AnyEvent::Log>) is loaded,
		2113	so will take effect even before AnyEvent has initialised itself.
		2114
		2115	Note that specifying this environment variable causes the L<AnyEvent::Log>
		2116	module to be loaded, while C<PERL_ANYEVENT_VERBOSE> does not, so only
		2117	using the latter saves a few hundred kB of memory unless a module
		2118	explicitly needs the extra features of AnyEvent::Log.
		2119
		2120	=item C<PERL_ANYEVENT_STRICT>
		2121
		2122	AnyEvent does not do much argument checking by default, as thorough
		2123	argument checking is very costly. Setting this variable to a true value
		2124	will cause AnyEvent to load C<AnyEvent::Strict> and then to thoroughly
		2125	check the arguments passed to most method calls. If it finds any problems,
		2126	it will croak.
		2127
		2128	In other words, enables "strict" mode.
		2129
		2130	Unlike C<use strict> (or its modern cousin, C<< use L<common::sense>
		2131	>>, it is definitely recommended to keep it off in production. Keeping
		2132	C<PERL_ANYEVENT_STRICT=1> in your environment while developing programs
		2133	can be very useful, however.
		2134
		2135	=item C<PERL_ANYEVENT_DEBUG_SHELL>
		2136
		2137	If this env variable is nonempty, then its contents will be interpreted by
		2138	C<AnyEvent::Socket::parse_hostport> and C<AnyEvent::Debug::shell> (after
		2139	replacing every occurance of C<$$> by the process pid). The shell object
		2140	is saved in C<$AnyEvent::Debug::SHELL>.
		2141
		2142	This happens when the first watcher is created.
		2143
		2144	For example, to bind a debug shell on a unix domain socket in
		2145	F<< /tmp/debug<pid>.sock >>, you could use this:
		2146
		2147	PERL_ANYEVENT_DEBUG_SHELL=/tmp/debug\$\$.sock perlprog
		2148	# connect with e.g.: socat readline /tmp/debug123.sock
		2149
		2150	Or to bind to tcp port 4545 on localhost:
		2151
		2152	PERL_ANYEVENT_DEBUG_SHELL=127.0.0.1:4545 perlprog
		2153	# connect with e.g.: telnet localhost 4545
		2154
		2155	Note that creating sockets in F</tmp> or on localhost is very unsafe on
		2156	multiuser systems.
		2157
		2158	=item C<PERL_ANYEVENT_DEBUG_WRAP>
		2159
		2160	Can be set to C<0>, C<1> or C<2> and enables wrapping of all watchers for
		2161	debugging purposes. See C<AnyEvent::Debug::wrap> for details.
		2162
		2163	=item C<PERL_ANYEVENT_MODEL>
		2164
		2165	This can be used to specify the event model to be used by AnyEvent, before
		2166	auto detection and -probing kicks in.
		2167
		2168	It normally is a string consisting entirely of ASCII letters (e.g. C<EV>
		2169	or C<IOAsync>). The string C<AnyEvent::Impl::> gets prepended and the
		2170	resulting module name is loaded and - if the load was successful - used as
		2171	event model backend. If it fails to load then AnyEvent will proceed with
		2172	auto detection and -probing.
		2173
		2174	If the string ends with C<::> instead (e.g. C<AnyEvent::Impl::EV::>) then
		2175	nothing gets prepended and the module name is used as-is (hint: C<::> at
		2176	the end of a string designates a module name and quotes it appropriately).
		2177
		2178	For example, to force the pure perl model (L<AnyEvent::Loop::Perl>) you
		2179	could start your program like this:
		2180
		2181	PERL_ANYEVENT_MODEL=Perl perl ...
		2182
		2183	=item C<PERL_ANYEVENT_IO_MODEL>
		2184
		2185	The current file I/O model - see L<AnyEvent::IO> for more info.
		2186
		2187	At the moment, only C<Perl> (small, pure-perl, synchronous) and
		2188	C<IOAIO> (truly asynchronous) are supported. The default is C<IOAIO> if
		2189	L<AnyEvent::AIO> can be loaded, otherwise it is C<Perl>.
		2190
		2191	=item C<PERL_ANYEVENT_PROTOCOLS>
		2192
		2193	Used by both L<AnyEvent::DNS> and L<AnyEvent::Socket> to determine preferences
		2194	for IPv4 or IPv6. The default is unspecified (and might change, or be the result
		2195	of auto probing).
		2196
		2197	Must be set to a comma-separated list of protocols or address families,
		2198	current supported: C<ipv4> and C<ipv6>. Only protocols mentioned will be
		2199	used, and preference will be given to protocols mentioned earlier in the
		2200	list.
		2201
		2202	This variable can effectively be used for denial-of-service attacks
		2203	against local programs (e.g. when setuid), although the impact is likely
		2204	small, as the program has to handle conenction and other failures anyways.
		2205
		2206	Examples: C<PERL_ANYEVENT_PROTOCOLS=ipv4,ipv6> - prefer IPv4 over IPv6,
		2207	but support both and try to use both. C<PERL_ANYEVENT_PROTOCOLS=ipv4>
		2208	- only support IPv4, never try to resolve or contact IPv6
		2209	addresses. C<PERL_ANYEVENT_PROTOCOLS=ipv6,ipv4> support either IPv4 or
		2210	IPv6, but prefer IPv6 over IPv4.
		2211
		2212	=item C<PERL_ANYEVENT_HOSTS>
		2213
		2214	This variable, if specified, overrides the F</etc/hosts> file used by
		2215	L<AnyEvent::Socket>C<::resolve_sockaddr>, i.e. hosts aliases will be read
		2216	from that file instead.
		2217
		2218	=item C<PERL_ANYEVENT_EDNS0>
		2219
		2220	Used by L<AnyEvent::DNS> to decide whether to use the EDNS0 extension for
		2221	DNS. This extension is generally useful to reduce DNS traffic, especially
		2222	when DNSSEC is involved, but some (broken) firewalls drop such DNS
		2223	packets, which is why it is off by default.
		2224
		2225	Setting this variable to C<1> will cause L<AnyEvent::DNS> to announce
		2226	EDNS0 in its DNS requests.
		2227
		2228	=item C<PERL_ANYEVENT_MAX_FORKS>
		2229
		2230	The maximum number of child processes that C<AnyEvent::Util::fork_call>
		2231	will create in parallel.
		2232
		2233	=item C<PERL_ANYEVENT_MAX_OUTSTANDING_DNS>
		2234
		2235	The default value for the C<max_outstanding> parameter for the default DNS
		2236	resolver - this is the maximum number of parallel DNS requests that are
		2237	sent to the DNS server.
		2238
		2239	=item C<PERL_ANYEVENT_MAX_SIGNAL_LATENCY>
		2240
		2241	Perl has inherently racy signal handling (you can basically choose between
		2242	losing signals and memory corruption) - pure perl event loops (including
		2243	C<AnyEvent::Loop>, when C<Async::Interrupt> isn't available) therefore
		2244	have to poll regularly to avoid losing signals.
		2245
		2246	Some event loops are racy, but don't poll regularly, and some event loops
		2247	are written in C but are still racy. For those event loops, AnyEvent
		2248	installs a timer that regularly wakes up the event loop.
		2249
		2250	By default, the interval for this timer is C<10> seconds, but you can
		2251	override this delay with this environment variable (or by setting
		2252	the C<$AnyEvent::MAX_SIGNAL_LATENCY> variable before creating signal
		2253	watchers).
		2254
		2255	Lower values increase CPU (and energy) usage, higher values can introduce
		2256	long delays when reaping children or waiting for signals.
		2257
		2258	The L<AnyEvent::Async> module, if available, will be used to avoid this
		2259	polling (with most event loops).
		2260
		2261	=item C<PERL_ANYEVENT_RESOLV_CONF>
		2262
		2263	The absolute path to a F<resolv.conf>-style file to use instead of
		2264	F</etc/resolv.conf> (or the OS-specific configuration) in the default
		2265	resolver, or the empty string to select the default configuration.
		2266
		2267	=item C<PERL_ANYEVENT_CA_FILE>, C<PERL_ANYEVENT_CA_PATH>.
		2268
		2269	When neither C<ca_file> nor C<ca_path> was specified during
		2270	L<AnyEvent::TLS> context creation, and either of these environment
		2271	variables are nonempty, they will be used to specify CA certificate
		2272	locations instead of a system-dependent default.
		2273
		2274	=item C<PERL_ANYEVENT_AVOID_GUARD> and C<PERL_ANYEVENT_AVOID_ASYNC_INTERRUPT>
		2275
		2276	When these are set to C<1>, then the respective modules are not
		2277	loaded. Mostly good for testing AnyEvent itself.
		2278
		2279	=back
428		2280
429	=head1 SUPPLYING YOUR OWN EVENT MODEL INTERFACE	2281	=head1 SUPPLYING YOUR OWN EVENT MODEL INTERFACE
		2282
		2283	This is an advanced topic that you do not normally need to use AnyEvent in
		2284	a module. This section is only of use to event loop authors who want to
		2285	provide AnyEvent compatibility.
430		2286
431	If you need to support another event library which isn't directly	2287	If you need to support another event library which isn't directly
432	supported by AnyEvent, you can supply your own interface to it by	2288	supported by AnyEvent, you can supply your own interface to it by
433	pushing, before the first watcher gets created, the package name of	2289	pushing, before the first watcher gets created, the package name of
434	the event module and the package name of the interface to use onto	2290	the event module and the package name of the interface to use onto
435	C<@AnyEvent::REGISTRY>. You can do that before and even without loading	2291	C<@AnyEvent::REGISTRY>. You can do that before and even without loading
436	AnyEvent.	2292	AnyEvent, so it is reasonably cheap.
437		2293
438	Example:	2294	Example:
439		2295
440	push @AnyEvent::REGISTRY, [urxvt => urxvt::anyevent::];	2296	push @AnyEvent::REGISTRY, [urxvt => urxvt::anyevent::];
441		2297
442	This tells AnyEvent to (literally) use the C<urxvt::anyevent::>	2298	This tells AnyEvent to (literally) use the C<urxvt::anyevent::>
443	package/class when it finds the C<urxvt> package/module is loaded. When	2299	package/class when it finds the C<urxvt> package/module is already loaded.
		2300
444	AnyEvent is loaded and asked to find a suitable event model, it will	2301	When AnyEvent is loaded and asked to find a suitable event model, it
445	first check for the presence of urxvt.	2302	will first check for the presence of urxvt by trying to C<use> the
		2303	C<urxvt::anyevent> module.
446		2304
447	The class should provide implementations for all watcher types (see	2305	The class should provide implementations for all watcher types. See
448	L<AnyEvent::Impl::Event> (source code), L<AnyEvent::Impl::Glib>	2306	L<AnyEvent::Impl::EV> (source code), L<AnyEvent::Impl::Glib> (Source code)
449	(Source code) and so on for actual examples, use C<perldoc -m	2307	and so on for actual examples. Use C<perldoc -m AnyEvent::Impl::Glib> to
450	AnyEvent::Impl::Glib> to see the sources).	2308	see the sources.
451		2309
		2310	If you don't provide C<signal> and C<child> watchers than AnyEvent will
		2311	provide suitable (hopefully) replacements.
		2312
452	The above isn't fictitious, the I<rxvt-unicode> (a.k.a. urxvt)	2313	The above example isn't fictitious, the I<rxvt-unicode> (a.k.a. urxvt)
453	uses the above line as-is. An interface isn't included in AnyEvent	2314	terminal emulator uses the above line as-is. An interface isn't included
454	because it doesn't make sense outside the embedded interpreter inside	2315	in AnyEvent because it doesn't make sense outside the embedded interpreter
455	I<rxvt-unicode>, and it is updated and maintained as part of the	2316	inside I<rxvt-unicode>, and it is updated and maintained as part of the
456	I<rxvt-unicode> distribution.	2317	I<rxvt-unicode> distribution.
457		2318
458	I<rxvt-unicode> also cheats a bit by not providing blocking access to	2319	I<rxvt-unicode> also cheats a bit by not providing blocking access to
459	condition variables: code blocking while waiting for a condition will	2320	condition variables: code blocking while waiting for a condition will
460	C<die>. This still works with most modules/usages, and blocking calls must	2321	C<die>. This still works with most modules/usages, and blocking calls must
461	not be in an interactive application, so it makes sense.	2322	not be done in an interactive application, so it makes sense.
462		2323
463	=head1 ENVIRONMENT VARIABLES
464
465	The following environment variables are used by this module:
466
467	C<PERL_ANYEVENT_VERBOSE> when set to C<2> or higher, reports which event
468	model gets used.
469
470	=head1 EXAMPLE	2324	=head1 EXAMPLE PROGRAM
471		2325
472	The following program uses an io watcher to read data from stdin, a timer	2326	The following program uses an I/O watcher to read data from STDIN, a timer
473	to display a message once per second, and a condvar to exit the program	2327	to display a message once per second, and a condition variable to quit the
474	when the user enters quit:	2328	program when the user enters quit:
475		2329
476	use AnyEvent;	2330	use AnyEvent;
477		2331
478	my $cv = AnyEvent->condvar;	2332	my $cv = AnyEvent->condvar;
479		2333
480	my $io_watcher = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {	2334	my $io_watcher = AnyEvent->io (
		2335	fh => \*STDIN,
		2336	poll => 'r',
		2337	cb => sub {
481	warn "io event <$_[0]>\n"; # will always output <r>	2338	warn "io event <$_[0]>\n"; # will always output <r>
482	chomp (my $input = <STDIN>); # read a line	2339	chomp (my $input = <STDIN>); # read a line
483	warn "read: $input\n"; # output what has been read	2340	warn "read: $input\n"; # output what has been read
484	$cv->broadcast if $input =~ /^q/i; # quit program if /^q/i	2341	$cv->send if $input =~ /^q/i; # quit program if /^q/i
		2342	},
		2343	);
		2344
		2345	my $time_watcher = AnyEvent->timer (after => 1, interval => 1, cb => sub {
		2346	warn "timeout\n"; # print 'timeout' at most every second
485	});	2347	});
486		2348
487	my $time_watcher; # can only be used once
488
489	sub new_timer {
490	$timer = AnyEvent->timer (after => 1, cb => sub {
491	warn "timeout\n"; # print 'timeout' about every second
492	&new_timer; # and restart the time
493	});
494	}
495
496	new_timer; # create first timer
497
498	$cv->wait; # wait until user enters /^q/i	2349	$cv->recv; # wait until user enters /^q/i
499		2350
500	=head1 REAL-WORLD EXAMPLE	2351	=head1 REAL-WORLD EXAMPLE
501		2352
502	Consider the L<Net::FCP> module. It features (among others) the following	2353	Consider the L<Net::FCP> module. It features (among others) the following
503	API calls, which are to freenet what HTTP GET requests are to http:	2354	API calls, which are to freenet what HTTP GET requests are to http:
…		…
553	syswrite $txn->{fh}, $txn->{request}	2404	syswrite $txn->{fh}, $txn->{request}
554	or die "connection or write error";	2405	or die "connection or write error";
555	$txn->{w} = AnyEvent->io (fh => $txn->{fh}, poll => 'r', cb => sub { $txn->fh_ready_r });	2406	$txn->{w} = AnyEvent->io (fh => $txn->{fh}, poll => 'r', cb => sub { $txn->fh_ready_r });
556		2407
557	Again, C<fh_ready_r> waits till all data has arrived, and then stores the	2408	Again, C<fh_ready_r> waits till all data has arrived, and then stores the
558	result and signals any possible waiters that the request ahs finished:	2409	result and signals any possible waiters that the request has finished:
559		2410
560	sysread $txn->{fh}, $txn->{buf}, length $txn->{$buf};	2411	sysread $txn->{fh}, $txn->{buf}, length $txn->{$buf};
561		2412
562	if (end-of-file or data complete) {	2413	if (end-of-file or data complete) {
563	$txn->{result} = $txn->{buf};	2414	$txn->{result} = $txn->{buf};
564	$txn->{finished}->broadcast;	2415	$txn->{finished}->send;
565	$txb->{cb}->($txn) of $txn->{cb}; # also call callback	2416	$txb->{cb}->($txn) of $txn->{cb}; # also call callback
566	}	2417	}
567		2418
568	The C<result> method, finally, just waits for the finished signal (if the	2419	The C<result> method, finally, just waits for the finished signal (if the
569	request was already finished, it doesn't wait, of course, and returns the	2420	request was already finished, it doesn't wait, of course, and returns the
570	data:	2421	data:
571		2422
572	$txn->{finished}->wait;	2423	$txn->{finished}->recv;
573	return $txn->{result};	2424	return $txn->{result};
574		2425
575	The actual code goes further and collects all errors (C<die>s, exceptions)	2426	The actual code goes further and collects all errors (C<die>s, exceptions)
576	that occured during request processing. The C<result> method detects	2427	that occurred during request processing. The C<result> method detects
577	wether an exception as thrown (it is stored inside the $txn object)	2428	whether an exception as thrown (it is stored inside the $txn object)
578	and just throws the exception, which means connection errors and other	2429	and just throws the exception, which means connection errors and other
579	problems get reported tot he code that tries to use the result, not in a	2430	problems get reported to the code that tries to use the result, not in a
580	random callback.	2431	random callback.
581		2432
582	All of this enables the following usage styles:	2433	All of this enables the following usage styles:
583		2434
584	1. Blocking:	2435	1. Blocking:
585		2436
586	my $data = $fcp->client_get ($url);	2437	my $data = $fcp->client_get ($url);
587		2438
588	2. Blocking, but parallelizing:	2439	2. Blocking, but running in parallel:
589		2440
590	my @datas = map $_->result,	2441	my @datas = map $_->result,
591	map $fcp->txn_client_get ($_),	2442	map $fcp->txn_client_get ($_),
592	@urls;	2443	@urls;
593		2444
594	Both blocking examples work without the module user having to know	2445	Both blocking examples work without the module user having to know
595	anything about events.	2446	anything about events.
596		2447
597	3a. Event-based in a main program, using any support Event module:	2448	3a. Event-based in a main program, using any supported event module:
598		2449
599	use Event;	2450	use EV;
600		2451
601	$fcp->txn_client_get ($url)->cb (sub {	2452	$fcp->txn_client_get ($url)->cb (sub {
602	my $txn = shift;	2453	my $txn = shift;
603	my $data = $txn->result;	2454	my $data = $txn->result;
604	...	2455	...
605	});	2456	});
606		2457
607	Event::loop;	2458	EV::loop;
608		2459
609	3b. The module user could use AnyEvent, too:	2460	3b. The module user could use AnyEvent, too:
610		2461
611	use AnyEvent;	2462	use AnyEvent;
612		2463
613	my $quit = AnyEvent->condvar;	2464	my $quit = AnyEvent->condvar;
614		2465
615	$fcp->txn_client_get ($url)->cb (sub {	2466	$fcp->txn_client_get ($url)->cb (sub {
616	...	2467	...
617	$quit->broadcast;	2468	$quit->send;
618	});	2469	});
619		2470
620	$quit->wait;	2471	$quit->recv;
		2472
		2473
		2474	=head1 BENCHMARKS
		2475
		2476	To give you an idea of the performance and overheads that AnyEvent adds
		2477	over the event loops themselves and to give you an impression of the speed
		2478	of various event loops I prepared some benchmarks.
		2479
		2480	=head2 BENCHMARKING ANYEVENT OVERHEAD
		2481
		2482	Here is a benchmark of various supported event models used natively and
		2483	through AnyEvent. The benchmark creates a lot of timers (with a zero
		2484	timeout) and I/O watchers (watching STDOUT, a pty, to become writable,
		2485	which it is), lets them fire exactly once and destroys them again.
		2486
		2487	Source code for this benchmark is found as F<eg/bench> in the AnyEvent
		2488	distribution. It uses the L<AE> interface, which makes a real difference
		2489	for the EV and Perl backends only.
		2490
		2491	=head3 Explanation of the columns
		2492
		2493	I<watcher> is the number of event watchers created/destroyed. Since
		2494	different event models feature vastly different performances, each event
		2495	loop was given a number of watchers so that overall runtime is acceptable
		2496	and similar between tested event loop (and keep them from crashing): Glib
		2497	would probably take thousands of years if asked to process the same number
		2498	of watchers as EV in this benchmark.
		2499
		2500	I<bytes> is the number of bytes (as measured by the resident set size,
		2501	RSS) consumed by each watcher. This method of measuring captures both C
		2502	and Perl-based overheads.
		2503
		2504	I<create> is the time, in microseconds (millionths of seconds), that it
		2505	takes to create a single watcher. The callback is a closure shared between
		2506	all watchers, to avoid adding memory overhead. That means closure creation
		2507	and memory usage is not included in the figures.
		2508
		2509	I<invoke> is the time, in microseconds, used to invoke a simple
		2510	callback. The callback simply counts down a Perl variable and after it was
		2511	invoked "watcher" times, it would C<< ->send >> a condvar once to
		2512	signal the end of this phase.
		2513
		2514	I<destroy> is the time, in microseconds, that it takes to destroy a single
		2515	watcher.
		2516
		2517	=head3 Results
		2518
		2519	name watchers bytes create invoke destroy comment
		2520	EV/EV 100000 223 0.47 0.43 0.27 EV native interface
		2521	EV/Any 100000 223 0.48 0.42 0.26 EV + AnyEvent watchers
		2522	Coro::EV/Any 100000 223 0.47 0.42 0.26 coroutines + Coro::Signal
		2523	Perl/Any 100000 431 2.70 0.74 0.92 pure perl implementation
		2524	Event/Event 16000 516 31.16 31.84 0.82 Event native interface
		2525	Event/Any 16000 1203 42.61 34.79 1.80 Event + AnyEvent watchers
		2526	IOAsync/Any 16000 1911 41.92 27.45 16.81 via IO::Async::Loop::IO_Poll
		2527	IOAsync/Any 16000 1726 40.69 26.37 15.25 via IO::Async::Loop::Epoll
		2528	Glib/Any 16000 1118 89.00 12.57 51.17 quadratic behaviour
		2529	Tk/Any 2000 1346 20.96 10.75 8.00 SEGV with >> 2000 watchers
		2530	POE/Any 2000 6951 108.97 795.32 14.24 via POE::Loop::Event
		2531	POE/Any 2000 6648 94.79 774.40 575.51 via POE::Loop::Select
		2532
		2533	=head3 Discussion
		2534
		2535	The benchmark does I<not> measure scalability of the event loop very
		2536	well. For example, a select-based event loop (such as the pure perl one)
		2537	can never compete with an event loop that uses epoll when the number of
		2538	file descriptors grows high. In this benchmark, all events become ready at
		2539	the same time, so select/poll-based implementations get an unnatural speed
		2540	boost.
		2541
		2542	Also, note that the number of watchers usually has a nonlinear effect on
		2543	overall speed, that is, creating twice as many watchers doesn't take twice
		2544	the time - usually it takes longer. This puts event loops tested with a
		2545	higher number of watchers at a disadvantage.
		2546
		2547	To put the range of results into perspective, consider that on the
		2548	benchmark machine, handling an event takes roughly 1600 CPU cycles with
		2549	EV, 3100 CPU cycles with AnyEvent's pure perl loop and almost 3000000 CPU
		2550	cycles with POE.
		2551
		2552	C<EV> is the sole leader regarding speed and memory use, which are both
		2553	maximal/minimal, respectively. When using the L<AE> API there is zero
		2554	overhead (when going through the AnyEvent API create is about 5-6 times
		2555	slower, with other times being equal, so still uses far less memory than
		2556	any other event loop and is still faster than Event natively).
		2557
		2558	The pure perl implementation is hit in a few sweet spots (both the
		2559	constant timeout and the use of a single fd hit optimisations in the perl
		2560	interpreter and the backend itself). Nevertheless this shows that it
		2561	adds very little overhead in itself. Like any select-based backend its
		2562	performance becomes really bad with lots of file descriptors (and few of
		2563	them active), of course, but this was not subject of this benchmark.
		2564
		2565	The C<Event> module has a relatively high setup and callback invocation
		2566	cost, but overall scores in on the third place.
		2567
		2568	C<IO::Async> performs admirably well, about on par with C<Event>, even
		2569	when using its pure perl backend.
		2570
		2571	C<Glib>'s memory usage is quite a bit higher, but it features a
		2572	faster callback invocation and overall ends up in the same class as
		2573	C<Event>. However, Glib scales extremely badly, doubling the number of
		2574	watchers increases the processing time by more than a factor of four,
		2575	making it completely unusable when using larger numbers of watchers
		2576	(note that only a single file descriptor was used in the benchmark, so
		2577	inefficiencies of C<poll> do not account for this).
		2578
		2579	The C<Tk> adaptor works relatively well. The fact that it crashes with
		2580	more than 2000 watchers is a big setback, however, as correctness takes
		2581	precedence over speed. Nevertheless, its performance is surprising, as the
		2582	file descriptor is dup()ed for each watcher. This shows that the dup()
		2583	employed by some adaptors is not a big performance issue (it does incur a
		2584	hidden memory cost inside the kernel which is not reflected in the figures
		2585	above).
		2586
		2587	C<POE>, regardless of underlying event loop (whether using its pure perl
		2588	select-based backend or the Event module, the POE-EV backend couldn't
		2589	be tested because it wasn't working) shows abysmal performance and
		2590	memory usage with AnyEvent: Watchers use almost 30 times as much memory
		2591	as EV watchers, and 10 times as much memory as Event (the high memory
		2592	requirements are caused by requiring a session for each watcher). Watcher
		2593	invocation speed is almost 900 times slower than with AnyEvent's pure perl
		2594	implementation.
		2595
		2596	The design of the POE adaptor class in AnyEvent can not really account
		2597	for the performance issues, though, as session creation overhead is
		2598	small compared to execution of the state machine, which is coded pretty
		2599	optimally within L<AnyEvent::Impl::POE> (and while everybody agrees that
		2600	using multiple sessions is not a good approach, especially regarding
		2601	memory usage, even the author of POE could not come up with a faster
		2602	design).
		2603
		2604	=head3 Summary
		2605
		2606	=over 4
		2607
		2608	=item * Using EV through AnyEvent is faster than any other event loop
		2609	(even when used without AnyEvent), but most event loops have acceptable
		2610	performance with or without AnyEvent.
		2611
		2612	=item * The overhead AnyEvent adds is usually much smaller than the overhead of
		2613	the actual event loop, only with extremely fast event loops such as EV
		2614	does AnyEvent add significant overhead.
		2615
		2616	=item * You should avoid POE like the plague if you want performance or
		2617	reasonable memory usage.
		2618
		2619	=back
		2620
		2621	=head2 BENCHMARKING THE LARGE SERVER CASE
		2622
		2623	This benchmark actually benchmarks the event loop itself. It works by
		2624	creating a number of "servers": each server consists of a socket pair, a
		2625	timeout watcher that gets reset on activity (but never fires), and an I/O
		2626	watcher waiting for input on one side of the socket. Each time the socket
		2627	watcher reads a byte it will write that byte to a random other "server".
		2628
		2629	The effect is that there will be a lot of I/O watchers, only part of which
		2630	are active at any one point (so there is a constant number of active
		2631	fds for each loop iteration, but which fds these are is random). The
		2632	timeout is reset each time something is read because that reflects how
		2633	most timeouts work (and puts extra pressure on the event loops).
		2634
		2635	In this benchmark, we use 10000 socket pairs (20000 sockets), of which 100
		2636	(1%) are active. This mirrors the activity of large servers with many
		2637	connections, most of which are idle at any one point in time.
		2638
		2639	Source code for this benchmark is found as F<eg/bench2> in the AnyEvent
		2640	distribution. It uses the L<AE> interface, which makes a real difference
		2641	for the EV and Perl backends only.
		2642
		2643	=head3 Explanation of the columns
		2644
		2645	I<sockets> is the number of sockets, and twice the number of "servers" (as
		2646	each server has a read and write socket end).
		2647
		2648	I<create> is the time it takes to create a socket pair (which is
		2649	nontrivial) and two watchers: an I/O watcher and a timeout watcher.
		2650
		2651	I<request>, the most important value, is the time it takes to handle a
		2652	single "request", that is, reading the token from the pipe and forwarding
		2653	it to another server. This includes deleting the old timeout and creating
		2654	a new one that moves the timeout into the future.
		2655
		2656	=head3 Results
		2657
		2658	name sockets create request
		2659	EV 20000 62.66 7.99
		2660	Perl 20000 68.32 32.64
		2661	IOAsync 20000 174.06 101.15 epoll
		2662	IOAsync 20000 174.67 610.84 poll
		2663	Event 20000 202.69 242.91
		2664	Glib 20000 557.01 1689.52
		2665	POE 20000 341.54 12086.32 uses POE::Loop::Event
		2666
		2667	=head3 Discussion
		2668
		2669	This benchmark I<does> measure scalability and overall performance of the
		2670	particular event loop.
		2671
		2672	EV is again fastest. Since it is using epoll on my system, the setup time
		2673	is relatively high, though.
		2674
		2675	Perl surprisingly comes second. It is much faster than the C-based event
		2676	loops Event and Glib.
		2677
		2678	IO::Async performs very well when using its epoll backend, and still quite
		2679	good compared to Glib when using its pure perl backend.
		2680
		2681	Event suffers from high setup time as well (look at its code and you will
		2682	understand why). Callback invocation also has a high overhead compared to
		2683	the C<< $_->() for .. >>-style loop that the Perl event loop uses. Event
		2684	uses select or poll in basically all documented configurations.
		2685
		2686	Glib is hit hard by its quadratic behaviour w.r.t. many watchers. It
		2687	clearly fails to perform with many filehandles or in busy servers.
		2688
		2689	POE is still completely out of the picture, taking over 1000 times as long
		2690	as EV, and over 100 times as long as the Perl implementation, even though
		2691	it uses a C-based event loop in this case.
		2692
		2693	=head3 Summary
		2694
		2695	=over 4
		2696
		2697	=item * The pure perl implementation performs extremely well.
		2698
		2699	=item * Avoid Glib or POE in large projects where performance matters.
		2700
		2701	=back
		2702
		2703	=head2 BENCHMARKING SMALL SERVERS
		2704
		2705	While event loops should scale (and select-based ones do not...) even to
		2706	large servers, most programs we (or I :) actually write have only a few
		2707	I/O watchers.
		2708
		2709	In this benchmark, I use the same benchmark program as in the large server
		2710	case, but it uses only eight "servers", of which three are active at any
		2711	one time. This should reflect performance for a small server relatively
		2712	well.
		2713
		2714	The columns are identical to the previous table.
		2715
		2716	=head3 Results
		2717
		2718	name sockets create request
		2719	EV 16 20.00 6.54
		2720	Perl 16 25.75 12.62
		2721	Event 16 81.27 35.86
		2722	Glib 16 32.63 15.48
		2723	POE 16 261.87 276.28 uses POE::Loop::Event
		2724
		2725	=head3 Discussion
		2726
		2727	The benchmark tries to test the performance of a typical small
		2728	server. While knowing how various event loops perform is interesting, keep
		2729	in mind that their overhead in this case is usually not as important, due
		2730	to the small absolute number of watchers (that is, you need efficiency and
		2731	speed most when you have lots of watchers, not when you only have a few of
		2732	them).
		2733
		2734	EV is again fastest.
		2735
		2736	Perl again comes second. It is noticeably faster than the C-based event
		2737	loops Event and Glib, although the difference is too small to really
		2738	matter.
		2739
		2740	POE also performs much better in this case, but is is still far behind the
		2741	others.
		2742
		2743	=head3 Summary
		2744
		2745	=over 4
		2746
		2747	=item * C-based event loops perform very well with small number of
		2748	watchers, as the management overhead dominates.
		2749
		2750	=back
		2751
		2752	=head2 THE IO::Lambda BENCHMARK
		2753
		2754	Recently I was told about the benchmark in the IO::Lambda manpage, which
		2755	could be misinterpreted to make AnyEvent look bad. In fact, the benchmark
		2756	simply compares IO::Lambda with POE, and IO::Lambda looks better (which
		2757	shouldn't come as a surprise to anybody). As such, the benchmark is
		2758	fine, and mostly shows that the AnyEvent backend from IO::Lambda isn't
		2759	very optimal. But how would AnyEvent compare when used without the extra
		2760	baggage? To explore this, I wrote the equivalent benchmark for AnyEvent.
		2761
		2762	The benchmark itself creates an echo-server, and then, for 500 times,
		2763	connects to the echo server, sends a line, waits for the reply, and then
		2764	creates the next connection. This is a rather bad benchmark, as it doesn't
		2765	test the efficiency of the framework or much non-blocking I/O, but it is a
		2766	benchmark nevertheless.
		2767
		2768	name runtime
		2769	Lambda/select 0.330 sec
		2770	+ optimized 0.122 sec
		2771	Lambda/AnyEvent 0.327 sec
		2772	+ optimized 0.138 sec
		2773	Raw sockets/select 0.077 sec
		2774	POE/select, components 0.662 sec
		2775	POE/select, raw sockets 0.226 sec
		2776	POE/select, optimized 0.404 sec
		2777
		2778	AnyEvent/select/nb 0.085 sec
		2779	AnyEvent/EV/nb 0.068 sec
		2780	+state machine 0.134 sec
		2781
		2782	The benchmark is also a bit unfair (my fault): the IO::Lambda/POE
		2783	benchmarks actually make blocking connects and use 100% blocking I/O,
		2784	defeating the purpose of an event-based solution. All of the newly
		2785	written AnyEvent benchmarks use 100% non-blocking connects (using
		2786	AnyEvent::Socket::tcp_connect and the asynchronous pure perl DNS
		2787	resolver), so AnyEvent is at a disadvantage here, as non-blocking connects
		2788	generally require a lot more bookkeeping and event handling than blocking
		2789	connects (which involve a single syscall only).
		2790
		2791	The last AnyEvent benchmark additionally uses L<AnyEvent::Handle>, which
		2792	offers similar expressive power as POE and IO::Lambda, using conventional
		2793	Perl syntax. This means that both the echo server and the client are 100%
		2794	non-blocking, further placing it at a disadvantage.
		2795
		2796	As you can see, the AnyEvent + EV combination even beats the
		2797	hand-optimised "raw sockets benchmark", while AnyEvent + its pure perl
		2798	backend easily beats IO::Lambda and POE.
		2799
		2800	And even the 100% non-blocking version written using the high-level (and
		2801	slow :) L<AnyEvent::Handle> abstraction beats both POE and IO::Lambda
		2802	higher level ("unoptimised") abstractions by a large margin, even though
		2803	it does all of DNS, tcp-connect and socket I/O in a non-blocking way.
		2804
		2805	The two AnyEvent benchmarks programs can be found as F<eg/ae0.pl> and
		2806	F<eg/ae2.pl> in the AnyEvent distribution, the remaining benchmarks are
		2807	part of the IO::Lambda distribution and were used without any changes.
		2808
		2809
		2810	=head1 SIGNALS
		2811
		2812	AnyEvent currently installs handlers for these signals:
		2813
		2814	=over 4
		2815
		2816	=item SIGCHLD
		2817
		2818	A handler for C<SIGCHLD> is installed by AnyEvent's child watcher
		2819	emulation for event loops that do not support them natively. Also, some
		2820	event loops install a similar handler.
		2821
		2822	Additionally, when AnyEvent is loaded and SIGCHLD is set to IGNORE, then
		2823	AnyEvent will reset it to default, to avoid losing child exit statuses.
		2824
		2825	=item SIGPIPE
		2826
		2827	A no-op handler is installed for C<SIGPIPE> when C<$SIG{PIPE}> is C<undef>
		2828	when AnyEvent gets loaded.
		2829
		2830	The rationale for this is that AnyEvent users usually do not really depend
		2831	on SIGPIPE delivery (which is purely an optimisation for shell use, or
		2832	badly-written programs), but C<SIGPIPE> can cause spurious and rare
		2833	program exits as a lot of people do not expect C<SIGPIPE> when writing to
		2834	some random socket.
		2835
		2836	The rationale for installing a no-op handler as opposed to ignoring it is
		2837	that this way, the handler will be restored to defaults on exec.
		2838
		2839	Feel free to install your own handler, or reset it to defaults.
		2840
		2841	=back
		2842
		2843	=cut
		2844
		2845	undef $SIG{CHLD}
		2846	if $SIG{CHLD} eq 'IGNORE';
		2847
		2848	$SIG{PIPE} = sub { }
		2849	unless defined $SIG{PIPE};
		2850
		2851	=head1 RECOMMENDED/OPTIONAL MODULES
		2852
		2853	One of AnyEvent's main goals is to be 100% Pure-Perl(tm): only perl (and
		2854	its built-in modules) are required to use it.
		2855
		2856	That does not mean that AnyEvent won't take advantage of some additional
		2857	modules if they are installed.
		2858
		2859	This section explains which additional modules will be used, and how they
		2860	affect AnyEvent's operation.
		2861
		2862	=over 4
		2863
		2864	=item L<Async::Interrupt>
		2865
		2866	This slightly arcane module is used to implement fast signal handling: To
		2867	my knowledge, there is no way to do completely race-free and quick
		2868	signal handling in pure perl. To ensure that signals still get
		2869	delivered, AnyEvent will start an interval timer to wake up perl (and
		2870	catch the signals) with some delay (default is 10 seconds, look for
		2871	C<$AnyEvent::MAX_SIGNAL_LATENCY>).
		2872
		2873	If this module is available, then it will be used to implement signal
		2874	catching, which means that signals will not be delayed, and the event loop
		2875	will not be interrupted regularly, which is more efficient (and good for
		2876	battery life on laptops).
		2877
		2878	This affects not just the pure-perl event loop, but also other event loops
		2879	that have no signal handling on their own (e.g. Glib, Tk, Qt).
		2880
		2881	Some event loops (POE, Event, Event::Lib) offer signal watchers natively,
		2882	and either employ their own workarounds (POE) or use AnyEvent's workaround
		2883	(using C<$AnyEvent::MAX_SIGNAL_LATENCY>). Installing L<Async::Interrupt>
		2884	does nothing for those backends.
		2885
		2886	=item L<EV>
		2887
		2888	This module isn't really "optional", as it is simply one of the backend
		2889	event loops that AnyEvent can use. However, it is simply the best event
		2890	loop available in terms of features, speed and stability: It supports
		2891	the AnyEvent API optimally, implements all the watcher types in XS, does
		2892	automatic timer adjustments even when no monotonic clock is available,
		2893	can take avdantage of advanced kernel interfaces such as C<epoll> and
		2894	C<kqueue>, and is the fastest backend I<by far>. You can even embed
		2895	L<Glib>/L<Gtk2> in it (or vice versa, see L<EV::Glib> and L<Glib::EV>).
		2896
		2897	If you only use backends that rely on another event loop (e.g. C<Tk>),
		2898	then this module will do nothing for you.
		2899
		2900	=item L<Guard>
		2901
		2902	The guard module, when used, will be used to implement
		2903	C<AnyEvent::Util::guard>. This speeds up guards considerably (and uses a
		2904	lot less memory), but otherwise doesn't affect guard operation much. It is
		2905	purely used for performance.
		2906
		2907	=item L<JSON> and L<JSON::XS>
		2908
		2909	One of these modules is required when you want to read or write JSON data
		2910	via L<AnyEvent::Handle>. L<JSON> is also written in pure-perl, but can take
		2911	advantage of the ultra-high-speed L<JSON::XS> module when it is installed.
		2912
		2913	=item L<Net::SSLeay>
		2914
		2915	Implementing TLS/SSL in Perl is certainly interesting, but not very
		2916	worthwhile: If this module is installed, then L<AnyEvent::Handle> (with
		2917	the help of L<AnyEvent::TLS>), gains the ability to do TLS/SSL.
		2918
		2919	=item L<Time::HiRes>
		2920
		2921	This module is part of perl since release 5.008. It will be used when the
		2922	chosen event library does not come with a timing source of its own. The
		2923	pure-perl event loop (L<AnyEvent::Loop>) will additionally load it to
		2924	try to use a monotonic clock for timing stability.
		2925
		2926	=back
		2927
		2928
		2929	=head1 FORK
		2930
		2931	Most event libraries are not fork-safe. The ones who are usually are
		2932	because they rely on inefficient but fork-safe C<select> or C<poll> calls
		2933	- higher performance APIs such as BSD's kqueue or the dreaded Linux epoll
		2934	are usually badly thought-out hacks that are incompatible with fork in
		2935	one way or another. Only L<EV> is fully fork-aware and ensures that you
		2936	continue event-processing in both parent and child (or both, if you know
		2937	what you are doing).
		2938
		2939	This means that, in general, you cannot fork and do event processing in
		2940	the child if the event library was initialised before the fork (which
		2941	usually happens when the first AnyEvent watcher is created, or the library
		2942	is loaded).
		2943
		2944	If you have to fork, you must either do so I<before> creating your first
		2945	watcher OR you must not use AnyEvent at all in the child OR you must do
		2946	something completely out of the scope of AnyEvent.
		2947
		2948	The problem of doing event processing in the parent I<and> the child
		2949	is much more complicated: even for backends that I<are> fork-aware or
		2950	fork-safe, their behaviour is not usually what you want: fork clones all
		2951	watchers, that means all timers, I/O watchers etc. are active in both
		2952	parent and child, which is almost never what you want. USing C<exec>
		2953	to start worker children from some kind of manage rprocess is usually
		2954	preferred, because it is much easier and cleaner, at the expense of having
		2955	to have another binary.
		2956
		2957
		2958	=head1 SECURITY CONSIDERATIONS
		2959
		2960	AnyEvent can be forced to load any event model via
		2961	$ENV{PERL_ANYEVENT_MODEL}. While this cannot (to my knowledge) be used to
		2962	execute arbitrary code or directly gain access, it can easily be used to
		2963	make the program hang or malfunction in subtle ways, as AnyEvent watchers
		2964	will not be active when the program uses a different event model than
		2965	specified in the variable.
		2966
		2967	You can make AnyEvent completely ignore this variable by deleting it
		2968	before the first watcher gets created, e.g. with a C<BEGIN> block:
		2969
		2970	BEGIN { delete $ENV{PERL_ANYEVENT_MODEL} }
		2971
		2972	use AnyEvent;
		2973
		2974	Similar considerations apply to $ENV{PERL_ANYEVENT_VERBOSE}, as that can
		2975	be used to probe what backend is used and gain other information (which is
		2976	probably even less useful to an attacker than PERL_ANYEVENT_MODEL), and
		2977	$ENV{PERL_ANYEVENT_STRICT}.
		2978
		2979	Note that AnyEvent will remove I<all> environment variables starting with
		2980	C<PERL_ANYEVENT_> from C<%ENV> when it is loaded while taint mode is
		2981	enabled.
		2982
		2983
		2984	=head1 BUGS
		2985
		2986	Perl 5.8 has numerous memleaks that sometimes hit this module and are hard
		2987	to work around. If you suffer from memleaks, first upgrade to Perl 5.10
		2988	and check wether the leaks still show up. (Perl 5.10.0 has other annoying
		2989	memleaks, such as leaking on C<map> and C<grep> but it is usually not as
		2990	pronounced).
		2991
621		2992
622	=head1 SEE ALSO	2993	=head1 SEE ALSO
623		2994
624	Event modules: L<Coro::Event>, L<Coro>, L<Event>, L<Glib::Event>, L<Glib>.	2995	Tutorial/Introduction: L<AnyEvent::Intro>.
625		2996
626	Implementations: L<AnyEvent::Impl::Coro>, L<AnyEvent::Impl::Event>, L<AnyEvent::Impl::Glib>, L<AnyEvent::Impl::Tk>.	2997	FAQ: L<AnyEvent::FAQ>.
627		2998
628	Nontrivial usage example: L<Net::FCP>.	2999	Utility functions: L<AnyEvent::Util> (misc. grab-bag), L<AnyEvent::Log>
		3000	(simply logging).
629		3001
630	=head1	3002	Development/Debugging: L<AnyEvent::Strict> (stricter checking),
		3003	L<AnyEvent::Debug> (interactive shell, watcher tracing).
		3004
		3005	Supported event modules: L<AnyEvent::Loop>, L<EV>, L<EV::Glib>,
		3006	L<Glib::EV>, L<Event>, L<Glib::Event>, L<Glib>, L<Tk>, L<Event::Lib>,
		3007	L<Qt>, L<POE>, L<FLTK>.
		3008
		3009	Implementations: L<AnyEvent::Impl::EV>, L<AnyEvent::Impl::Event>,
		3010	L<AnyEvent::Impl::Glib>, L<AnyEvent::Impl::Tk>, L<AnyEvent::Impl::Perl>,
		3011	L<AnyEvent::Impl::EventLib>, L<AnyEvent::Impl::Qt>,
		3012	L<AnyEvent::Impl::POE>, L<AnyEvent::Impl::IOAsync>, L<Anyevent::Impl::Irssi>,
		3013	L<AnyEvent::Impl::FLTK>.
		3014
		3015	Non-blocking handles, pipes, stream sockets, TCP clients and
		3016	servers: L<AnyEvent::Handle>, L<AnyEvent::Socket>, L<AnyEvent::TLS>.
		3017
		3018	Asynchronous File I/O: L<AnyEvent::IO>.
		3019
		3020	Asynchronous DNS: L<AnyEvent::DNS>.
		3021
		3022	Thread support: L<Coro>, L<Coro::AnyEvent>, L<Coro::EV>, L<Coro::Event>.
		3023
		3024	Nontrivial usage examples: L<AnyEvent::GPSD>, L<AnyEvent::IRC>,
		3025	L<AnyEvent::HTTP>.
		3026
		3027
		3028	=head1 AUTHOR
		3029
		3030	Marc Lehmann <schmorp@schmorp.de>
		3031	http://anyevent.schmorp.de
631		3032
632	=cut	3033	=cut
633		3034
634	1	3035	1
635		3036

Diff Legend

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