ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/AnyEvent/README

Comparing AnyEvent/README (file contents):
Revision 1.6 by root, Wed Nov 1 01:22:19 2006 UTC vs.
Revision 1.37 by root, Mon Apr 20 14:34:18 2009 UTC

1	NAME	1	NAME
2	AnyEvent - provide framework for multiple event loops	2	AnyEvent - provide framework for multiple event loops
3		3
4	Event, Coro, Glib, Tk, Perl - various supported event loops	4	EV, Event, Glib, Tk, Perl, Event::Lib, Qt, POE - various supported event
		5	loops
5		6
6	SYNOPSIS	7	SYNOPSIS
7	use AnyEvent;	8	use AnyEvent;
8		9
9	my $w = AnyEvent->io (fh => $fh, poll => "r|w", cb => sub {	10	my $w = AnyEvent->io (fh => $fh, poll => "r|w", cb => sub { ... });
		11
		12	my $w = AnyEvent->timer (after => $seconds, cb => sub { ... });
		13	my $w = AnyEvent->timer (after => $seconds, interval => $seconds, cb => ...
		14
		15	print AnyEvent->now; # prints current event loop time
		16	print AnyEvent->time; # think Time::HiRes::time or simply CORE::time.
		17
		18	my $w = AnyEvent->signal (signal => "TERM", cb => sub { ... });
		19
		20	my $w = AnyEvent->child (pid => $pid, cb => sub {
		21	my ($pid, $status) = @_;
10	...	22	...
11	});	23	});
12		24
13	my $w = AnyEvent->timer (after => $seconds, cb => sub {
14	...
15	});
16
17	my $w = AnyEvent->condvar; # stores wether a condition was flagged	25	my $w = AnyEvent->condvar; # stores whether a condition was flagged
		26	$w->send; # wake up current and all future recv's
18	$w->wait; # enters "main loop" till $condvar gets ->broadcast	27	$w->recv; # enters "main loop" till $condvar gets ->send
19	$w->broadcast; # wake up current and all future wait's	28	# use a condvar in callback mode:
		29	$w->cb (sub { $_[0]->recv });
		30
		31	INTRODUCTION/TUTORIAL
		32	This manpage is mainly a reference manual. If you are interested in a
		33	tutorial or some gentle introduction, have a look at the AnyEvent::Intro
		34	manpage.
		35
		36	WHY YOU SHOULD USE THIS MODULE (OR NOT)
		37	Glib, POE, IO::Async, Event... CPAN offers event models by the dozen
		38	nowadays. So what is different about AnyEvent?
		39
		40	Executive Summary: AnyEvent is *compatible*, AnyEvent is *free of
		41	policy* and AnyEvent is *small and efficient*.
		42
		43	First and foremost, *AnyEvent is not an event model* itself, it only
		44	interfaces to whatever event model the main program happens to use, in a
		45	pragmatic way. For event models and certain classes of immortals alike,
		46	the statement "there can only be one" is a bitter reality: In general,
		47	only one event loop can be active at the same time in a process.
		48	AnyEvent cannot change this, but it can hide the differences between
		49	those event loops.
		50
		51	The goal of AnyEvent is to offer module authors the ability to do event
		52	programming (waiting for I/O or timer events) without subscribing to a
		53	religion, a way of living, and most importantly: without forcing your
		54	module users into the same thing by forcing them to use the same event
		55	model you use.
		56
		57	For modules like POE or IO::Async (which is a total misnomer as it is
		58	actually doing all I/O *synchronously*...), using them in your module is
		59	like joining a cult: After you joined, you are dependent on them and you
		60	cannot use anything else, as they are simply incompatible to everything
		61	that isn't them. What's worse, all the potential users of your module
		62	are *also* forced to use the same event loop you use.
		63
		64	AnyEvent is different: AnyEvent + POE works fine. AnyEvent + Glib works
		65	fine. AnyEvent + Tk works fine etc. etc. but none of these work together
		66	with the rest: POE + IO::Async? No go. Tk + Event? No go. Again: if your
		67	module uses one of those, every user of your module has to use it, too.
		68	But if your module uses AnyEvent, it works transparently with all event
		69	models it supports (including stuff like IO::Async, as long as those use
		70	one of the supported event loops. It is trivial to add new event loops
		71	to AnyEvent, too, so it is future-proof).
		72
		73	In addition to being free of having to use *the one and only true event
		74	model*, AnyEvent also is free of bloat and policy: with POE or similar
		75	modules, you get an enormous amount of code and strict rules you have to
		76	follow. AnyEvent, on the other hand, is lean and up to the point, by
		77	only offering the functionality that is necessary, in as thin as a
		78	wrapper as technically possible.
		79
		80	Of course, AnyEvent comes with a big (and fully optional!) toolbox of
		81	useful functionality, such as an asynchronous DNS resolver, 100%
		82	non-blocking connects (even with TLS/SSL, IPv6 and on broken platforms
		83	such as Windows) and lots of real-world knowledge and workarounds for
		84	platform bugs and differences.
		85
		86	Now, if you *do want* lots of policy (this can arguably be somewhat
		87	useful) and you want to force your users to use the one and only event
		88	model, you should *not* use this module.
20		89
21	DESCRIPTION	90	DESCRIPTION
22	AnyEvent provides an identical interface to multiple event loops. This	91	AnyEvent provides an identical interface to multiple event loops. This
23	allows module authors to utilise an event loop without forcing module	92	allows module authors to utilise an event loop without forcing module
24	users to use the same event loop (as only a single event loop can	93	users to use the same event loop (as only a single event loop can
25	coexist peacefully at any one time).	94	coexist peacefully at any one time).
26		95
27	The interface itself is vaguely similar but not identical to the Event	96	The interface itself is vaguely similar, but not identical to the Event
28	module.	97	module.
29		98
30	On the first call of any method, the module tries to detect the	99	During the first call of any watcher-creation method, the module tries
31	currently loaded event loop by probing wether any of the following	100	to detect the currently loaded event loop by probing whether one of the
32	modules is loaded: Coro::Event, Event, Glib, Tk. The first one found is	101	following modules is already loaded: EV, Event, Glib,
		102	AnyEvent::Impl::Perl, Tk, Event::Lib, Qt, POE. The first one found is
33	used. If none is found, the module tries to load these modules in the	103	used. If none are found, the module tries to load these modules
34	order given. The first one that could be successfully loaded will be	104	(excluding Tk, Event::Lib, Qt and POE as the pure perl adaptor should
35	used. If still none could be found, AnyEvent will fall back to a	105	always succeed) in the order given. The first one that can be
36	pure-perl event loop, which is also not very efficient.	106	successfully loaded will be used. If, after this, still none could be
		107	found, AnyEvent will fall back to a pure-perl event loop, which is not
		108	very efficient, but should work everywhere.
37		109
38	Because AnyEvent first checks for modules that are already loaded,	110	Because AnyEvent first checks for modules that are already loaded,
39	loading an Event model explicitly before first using AnyEvent will	111	loading an event model explicitly before first using AnyEvent will
40	likely make that model the default. For example:	112	likely make that model the default. For example:
41		113
42	use Tk;	114	use Tk;
43	use AnyEvent;	115	use AnyEvent;
44		116
45	# .. AnyEvent will likely default to Tk	117	# .. AnyEvent will likely default to Tk
46		118
		119	The *likely* means that, if any module loads another event model and
		120	starts using it, all bets are off. Maybe you should tell their authors
		121	to use AnyEvent so their modules work together with others seamlessly...
		122
47	The pure-perl implementation of AnyEvent is called	123	The pure-perl implementation of AnyEvent is called
48	"AnyEvent::Impl::Perl". Like other event modules you can load it	124	"AnyEvent::Impl::Perl". Like other event modules you can load it
49	explicitly.	125	explicitly and enjoy the high availability of that event loop :)
50		126
51	WATCHERS	127	WATCHERS
52	AnyEvent has the central concept of a *watcher*, which is an object that	128	AnyEvent has the central concept of a *watcher*, which is an object that
53	stores relevant data for each kind of event you are waiting for, such as	129	stores relevant data for each kind of event you are waiting for, such as
54	the callback to call, the filehandle to watch, etc.	130	the callback to call, the file handle to watch, etc.
55		131
56	These watchers are normal Perl objects with normal Perl lifetime. After	132	These watchers are normal Perl objects with normal Perl lifetime. After
57	creating a watcher it will immediately "watch" for events and invoke the	133	creating a watcher it will immediately "watch" for events and invoke the
		134	callback when the event occurs (of course, only when the event model is
		135	in control).
		136
		137	Note that callbacks must not permanently change global variables
		138	potentially in use by the event loop (such as $_ or $[) and that
		139	callbacks must not "die". The former is good programming practise in
		140	Perl and the latter stems from the fact that exception handling differs
		141	widely between event loops.
		142
58	callback. To disable the watcher you have to destroy it (e.g. by setting	143	To disable the watcher you have to destroy it (e.g. by setting the
59	the variable that stores it to "undef" or otherwise deleting all	144	variable you store it in to "undef" or otherwise deleting all references
60	references to it).	145	to it).
61		146
62	All watchers are created by calling a method on the "AnyEvent" class.	147	All watchers are created by calling a method on the "AnyEvent" class.
63		148
		149	Many watchers either are used with "recursion" (repeating timers for
		150	example), or need to refer to their watcher object in other ways.
		151
		152	An any way to achieve that is this pattern:
		153
		154	my $w; $w = AnyEvent->type (arg => value ..., cb => sub {
		155	# you can use $w here, for example to undef it
		156	undef $w;
		157	});
		158
		159	Note that "my $w; $w =" combination. This is necessary because in Perl,
		160	my variables are only visible after the statement in which they are
		161	declared.
		162
64	IO WATCHERS	163	I/O WATCHERS
65	You can create I/O watcher by calling the "AnyEvent->io" method with the	164	You can create an I/O watcher by calling the "AnyEvent->io" method with
66	following mandatory arguments:	165	the following mandatory key-value pairs as arguments:
67		166
68	"fh" the Perl *filehandle* (not filedescriptor) to watch for events.	167	"fh" is the Perl *file handle* (*not* file descriptor) to watch for
		168	events (AnyEvent might or might not keep a reference to this file
		169	handle). Note that only file handles pointing to things for which
		170	non-blocking operation makes sense are allowed. This includes sockets,
		171	most character devices, pipes, fifos and so on, but not for example
		172	files or block devices.
		173
69	"poll" must be a string that is either "r" or "w", that creates a	174	"poll" must be a string that is either "r" or "w", which creates a
70	watcher waiting for "r"eadable or "w"ritable events. "cb" teh callback	175	watcher waiting for "r"eadable or "w"ritable events, respectively.
		176
71	to invoke everytime the filehandle becomes ready.	177	"cb" is the callback to invoke each time the file handle becomes ready.
72		178
73	Only one io watcher per "fh" and "poll" combination is allowed (i.e. on	179	Although the callback might get passed parameters, their value and
74	a socket you can have one r + one w, not any more (limitation comes from	180	presence is undefined and you cannot rely on them. Portable AnyEvent
75	Tk - if you are sure you are not using Tk this limitation is gone).	181	callbacks cannot use arguments passed to I/O watcher callbacks.
76		182
77	Filehandles will be kept alive, so as long as the watcher exists, the	183	The I/O watcher might use the underlying file descriptor or a copy of
78	filehandle exists, too.	184	it. You must not close a file handle as long as any watcher is active on
		185	the underlying file descriptor.
79		186
80	Example:	187	Some event loops issue spurious readyness notifications, so you should
		188	always use non-blocking calls when reading/writing from/to your file
		189	handles.
81		190
82	# wait for readability of STDIN, then read a line and disable the watcher	191	Example: wait for readability of STDIN, then read a line and disable the
		192	watcher.
		193
83	my $w; $w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {	194	my $w; $w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {
84	chomp (my $input = <STDIN>);	195	chomp (my $input = <STDIN>);
85	warn "read: $input\n";	196	warn "read: $input\n";
86	undef $w;	197	undef $w;
87	});	198	});
88		199
89	TIMER WATCHERS	200	TIME WATCHERS
90	You can create a timer watcher by calling the "AnyEvent->timer" method	201	You can create a time watcher by calling the "AnyEvent->timer" method
91	with the following mandatory arguments:	202	with the following mandatory arguments:
92		203
93	"after" after how many seconds (fractions are supported) should the	204	"after" specifies after how many seconds (fractional values are
94	timer activate. "cb" the callback to invoke.	205	supported) the callback should be invoked. "cb" is the callback to
		206	invoke in that case.
95		207
96	The timer callback will be invoked at most once: if you want a repeating	208	Although the callback might get passed parameters, their value and
97	timer you have to create a new watcher (this is a limitation by both Tk	209	presence is undefined and you cannot rely on them. Portable AnyEvent
98	and Glib).	210	callbacks cannot use arguments passed to time watcher callbacks.
99		211
100	Example:	212	The callback will normally be invoked once only. If you specify another
		213	parameter, "interval", as a strictly positive number (> 0), then the
		214	callback will be invoked regularly at that interval (in fractional
		215	seconds) after the first invocation. If "interval" is specified with a
		216	false value, then it is treated as if it were missing.
101		217
		218	The callback will be rescheduled before invoking the callback, but no
		219	attempt is done to avoid timer drift in most backends, so the interval
		220	is only approximate.
		221
102	# fire an event after 7.7 seconds	222	Example: fire an event after 7.7 seconds.
		223
103	my $w = AnyEvent->timer (after => 7.7, cb => sub {	224	my $w = AnyEvent->timer (after => 7.7, cb => sub {
104	warn "timeout\n";	225	warn "timeout\n";
105	});	226	});
106		227
107	# to cancel the timer:	228	# to cancel the timer:
108	undef $w	229	undef $w;
109		230
110	CONDITION WATCHERS	231	Example 2: fire an event after 0.5 seconds, then roughly every second.
		232
		233	my $w = AnyEvent->timer (after => 0.5, interval => 1, cb => sub {
		234	warn "timeout\n";
		235	};
		236
		237	TIMING ISSUES
		238	There are two ways to handle timers: based on real time (relative, "fire
		239	in 10 seconds") and based on wallclock time (absolute, "fire at 12
		240	o'clock").
		241
		242	While most event loops expect timers to specified in a relative way,
		243	they use absolute time internally. This makes a difference when your
		244	clock "jumps", for example, when ntp decides to set your clock backwards
		245	from the wrong date of 2014-01-01 to 2008-01-01, a watcher that is
		246	supposed to fire "after" a second might actually take six years to
		247	finally fire.
		248
		249	AnyEvent cannot compensate for this. The only event loop that is
		250	conscious about these issues is EV, which offers both relative
		251	(ev_timer, based on true relative time) and absolute (ev_periodic, based
		252	on wallclock time) timers.
		253
		254	AnyEvent always prefers relative timers, if available, matching the
		255	AnyEvent API.
		256
		257	AnyEvent has two additional methods that return the "current time":
		258
		259	AnyEvent->time
		260	This returns the "current wallclock time" as a fractional number of
		261	seconds since the Epoch (the same thing as "time" or
		262	"Time::HiRes::time" return, and the result is guaranteed to be
		263	compatible with those).
		264
		265	It progresses independently of any event loop processing, i.e. each
		266	call will check the system clock, which usually gets updated
		267	frequently.
		268
		269	AnyEvent->now
		270	This also returns the "current wallclock time", but unlike "time",
		271	above, this value might change only once per event loop iteration,
		272	depending on the event loop (most return the same time as "time",
		273	above). This is the time that AnyEvent's timers get scheduled
		274	against.
		275
		276	*In almost all cases (in all cases if you don't care), this is the
		277	function to call when you want to know the current time.*
		278
		279	This function is also often faster then "AnyEvent->time", and thus
		280	the preferred method if you want some timestamp (for example,
		281	AnyEvent::Handle uses this to update it's activity timeouts).
		282
		283	The rest of this section is only of relevance if you try to be very
		284	exact with your timing, you can skip it without bad conscience.
		285
		286	For a practical example of when these times differ, consider
		287	Event::Lib and EV and the following set-up:
		288
		289	The event loop is running and has just invoked one of your callback
		290	at time=500 (assume no other callbacks delay processing). In your
		291	callback, you wait a second by executing "sleep 1" (blocking the
		292	process for a second) and then (at time=501) you create a relative
		293	timer that fires after three seconds.
		294
		295	With Event::Lib, "AnyEvent->time" and "AnyEvent->now" will both
		296	return 501, because that is the current time, and the timer will be
		297	scheduled to fire at time=504 (501 + 3).
		298
		299	With EV, "AnyEvent->time" returns 501 (as that is the current time),
		300	but "AnyEvent->now" returns 500, as that is the time the last event
		301	processing phase started. With EV, your timer gets scheduled to run
		302	at time=503 (500 + 3).
		303
		304	In one sense, Event::Lib is more exact, as it uses the current time
		305	regardless of any delays introduced by event processing. However,
		306	most callbacks do not expect large delays in processing, so this
		307	causes a higher drift (and a lot more system calls to get the
		308	current time).
		309
		310	In another sense, EV is more exact, as your timer will be scheduled
		311	at the same time, regardless of how long event processing actually
		312	took.
		313
		314	In either case, if you care (and in most cases, you don't), then you
		315	can get whatever behaviour you want with any event loop, by taking
		316	the difference between "AnyEvent->time" and "AnyEvent->now" into
		317	account.
		318
		319	AnyEvent->now_update
		320	Some event loops (such as EV or AnyEvent::Impl::Perl) cache the
		321	current time for each loop iteration (see the discussion of
		322	AnyEvent->now, above).
		323
		324	When a callback runs for a long time (or when the process sleeps),
		325	then this "current" time will differ substantially from the real
		326	time, which might affect timers and time-outs.
		327
		328	When this is the case, you can call this method, which will update
		329	the event loop's idea of "current time".
		330
		331	Note that updating the time *might* cause some events to be handled.
		332
		333	SIGNAL WATCHERS
		334	You can watch for signals using a signal watcher, "signal" is the signal
		335	*name* in uppercase and without any "SIG" prefix, "cb" is the Perl
		336	callback to be invoked whenever a signal occurs.
		337
		338	Although the callback might get passed parameters, their value and
		339	presence is undefined and you cannot rely on them. Portable AnyEvent
		340	callbacks cannot use arguments passed to signal watcher callbacks.
		341
		342	Multiple signal occurrences can be clumped together into one callback
		343	invocation, and callback invocation will be synchronous. Synchronous
		344	means that it might take a while until the signal gets handled by the
		345	process, but it is guaranteed not to interrupt any other callbacks.
		346
		347	The main advantage of using these watchers is that you can share a
		348	signal between multiple watchers.
		349
		350	This watcher might use %SIG, so programs overwriting those signals
		351	directly will likely not work correctly.
		352
		353	Example: exit on SIGINT
		354
		355	my $w = AnyEvent->signal (signal => "INT", cb => sub { exit 1 });
		356
		357	CHILD PROCESS WATCHERS
		358	You can also watch on a child process exit and catch its exit status.
		359
		360	The child process is specified by the "pid" argument (if set to 0, it
		361	watches for any child process exit). The watcher will triggered only
		362	when the child process has finished and an exit status is available, not
		363	on any trace events (stopped/continued).
		364
		365	The callback will be called with the pid and exit status (as returned by
		366	waitpid), so unlike other watcher types, you *can* rely on child watcher
		367	callback arguments.
		368
		369	This watcher type works by installing a signal handler for "SIGCHLD",
		370	and since it cannot be shared, nothing else should use SIGCHLD or reap
		371	random child processes (waiting for specific child processes, e.g.
		372	inside "system", is just fine).
		373
		374	There is a slight catch to child watchers, however: you usually start
		375	them *after* the child process was created, and this means the process
		376	could have exited already (and no SIGCHLD will be sent anymore).
		377
		378	Not all event models handle this correctly (POE doesn't), but even for
		379	event models that *do* handle this correctly, they usually need to be
		380	loaded before the process exits (i.e. before you fork in the first
		381	place).
		382
		383	This means you cannot create a child watcher as the very first thing in
		384	an AnyEvent program, you *have* to create at least one watcher before
		385	you "fork" the child (alternatively, you can call "AnyEvent::detect").
		386
		387	Example: fork a process and wait for it
		388
		389	my $done = AnyEvent->condvar;
		390
		391	my $pid = fork or exit 5;
		392
		393	my $w = AnyEvent->child (
		394	pid => $pid,
		395	cb => sub {
		396	my ($pid, $status) = @_;
		397	warn "pid $pid exited with status $status";
		398	$done->send;
		399	},
		400	);
		401
		402	# do something else, then wait for process exit
		403	$done->recv;
		404
		405	CONDITION VARIABLES
		406	If you are familiar with some event loops you will know that all of them
		407	require you to run some blocking "loop", "run" or similar function that
		408	will actively watch for new events and call your callbacks.
		409
		410	AnyEvent is different, it expects somebody else to run the event loop
		411	and will only block when necessary (usually when told by the user).
		412
		413	The instrument to do that is called a "condition variable", so called
		414	because they represent a condition that must become true.
		415
111	Condition watchers can be created by calling the "AnyEvent->condvar"	416	Condition variables can be created by calling the "AnyEvent->condvar"
112	method without any arguments.	417	method, usually without arguments. The only argument pair allowed is
113		418
114	A condition watcher watches for a condition - precisely that the	419	"cb", which specifies a callback to be called when the condition
115	"->broadcast" method has been called.	420	variable becomes true, with the condition variable as the first argument
		421	(but not the results).
116		422
117	The watcher has only two methods:	423	After creation, the condition variable is "false" until it becomes
		424	"true" by calling the "send" method (or calling the condition variable
		425	as if it were a callback, read about the caveats in the description for
		426	the "->send" method).
118		427
119	$cv->wait	428	Condition variables are similar to callbacks, except that you can
		429	optionally wait for them. They can also be called merge points - points
		430	in time where multiple outstanding events have been processed. And yet
		431	another way to call them is transactions - each condition variable can
		432	be used to represent a transaction, which finishes at some point and
		433	delivers a result.
		434
		435	Condition variables are very useful to signal that something has
		436	finished, for example, if you write a module that does asynchronous http
		437	requests, then a condition variable would be the ideal candidate to
		438	signal the availability of results. The user can either act when the
		439	callback is called or can synchronously "->recv" for the results.
		440
		441	You can also use them to simulate traditional event loops - for example,
		442	you can block your main program until an event occurs - for example, you
		443	could "->recv" in your main program until the user clicks the Quit
		444	button of your app, which would "->send" the "quit" event.
		445
		446	Note that condition variables recurse into the event loop - if you have
		447	two pieces of code that call "->recv" in a round-robin fashion, you
		448	lose. Therefore, condition variables are good to export to your caller,
		449	but you should avoid making a blocking wait yourself, at least in
		450	callbacks, as this asks for trouble.
		451
		452	Condition variables are represented by hash refs in perl, and the keys
		453	used by AnyEvent itself are all named "_ae_XXX" to make subclassing easy
		454	(it is often useful to build your own transaction class on top of
		455	AnyEvent). To subclass, use "AnyEvent::CondVar" as base class and call
		456	it's "new" method in your own "new" method.
		457
		458	There are two "sides" to a condition variable - the "producer side"
		459	which eventually calls "-> send", and the "consumer side", which waits
		460	for the send to occur.
		461
		462	Example: wait for a timer.
		463
		464	# wait till the result is ready
		465	my $result_ready = AnyEvent->condvar;
		466
		467	# do something such as adding a timer
		468	# or socket watcher the calls $result_ready->send
		469	# when the "result" is ready.
		470	# in this case, we simply use a timer:
		471	my $w = AnyEvent->timer (
		472	after => 1,
		473	cb => sub { $result_ready->send },
		474	);
		475
		476	# this "blocks" (while handling events) till the callback
		477	# calls send
		478	$result_ready->recv;
		479
		480	Example: wait for a timer, but take advantage of the fact that condition
		481	variables are also code references.
		482
		483	my $done = AnyEvent->condvar;
		484	my $delay = AnyEvent->timer (after => 5, cb => $done);
		485	$done->recv;
		486
		487	Example: Imagine an API that returns a condvar and doesn't support
		488	callbacks. This is how you make a synchronous call, for example from the
		489	main program:
		490
		491	use AnyEvent::CouchDB;
		492
		493	...
		494
		495	my @info = $couchdb->info->recv;
		496
		497	And this is how you would just ste a callback to be called whenever the
		498	results are available:
		499
		500	$couchdb->info->cb (sub {
		501	my @info = $_[0]->recv;
		502	});
		503
		504	METHODS FOR PRODUCERS
		505	These methods should only be used by the producing side, i.e. the
		506	code/module that eventually sends the signal. Note that it is also the
		507	producer side which creates the condvar in most cases, but it isn't
		508	uncommon for the consumer to create it as well.
		509
		510	$cv->send (...)
		511	Flag the condition as ready - a running "->recv" and all further
		512	calls to "recv" will (eventually) return after this method has been
		513	called. If nobody is waiting the send will be remembered.
		514
		515	If a callback has been set on the condition variable, it is called
		516	immediately from within send.
		517
		518	Any arguments passed to the "send" call will be returned by all
		519	future "->recv" calls.
		520
		521	Condition variables are overloaded so one can call them directly (as
		522	a code reference). Calling them directly is the same as calling
		523	"send". Note, however, that many C-based event loops do not handle
		524	overloading, so as tempting as it may be, passing a condition
		525	variable instead of a callback does not work. Both the pure perl and
		526	EV loops support overloading, however, as well as all functions that
		527	use perl to invoke a callback (as in AnyEvent::Socket and
		528	AnyEvent::DNS for example).
		529
		530	$cv->croak ($error)
		531	Similar to send, but causes all call's to "->recv" to invoke
		532	"Carp::croak" with the given error message/object/scalar.
		533
		534	This can be used to signal any errors to the condition variable
		535	user/consumer.
		536
		537	$cv->begin ([group callback])
		538	$cv->end
		539	These two methods are EXPERIMENTAL and MIGHT CHANGE.
		540
		541	These two methods can be used to combine many transactions/events
		542	into one. For example, a function that pings many hosts in parallel
		543	might want to use a condition variable for the whole process.
		544
		545	Every call to "->begin" will increment a counter, and every call to
		546	"->end" will decrement it. If the counter reaches 0 in "->end", the
		547	(last) callback passed to "begin" will be executed. That callback is
		548	*supposed* to call "->send", but that is not required. If no
		549	callback was set, "send" will be called without any arguments.
		550
		551	Let's clarify this with the ping example:
		552
		553	my $cv = AnyEvent->condvar;
		554
		555	my %result;
		556	$cv->begin (sub { $cv->send (\%result) });
		557
		558	for my $host (@list_of_hosts) {
		559	$cv->begin;
		560	ping_host_then_call_callback $host, sub {
		561	$result{$host} = ...;
		562	$cv->end;
		563	};
		564	}
		565
		566	$cv->end;
		567
		568	This code fragment supposedly pings a number of hosts and calls
		569	"send" after results for all then have have been gathered - in any
		570	order. To achieve this, the code issues a call to "begin" when it
		571	starts each ping request and calls "end" when it has received some
		572	result for it. Since "begin" and "end" only maintain a counter, the
		573	order in which results arrive is not relevant.
		574
		575	There is an additional bracketing call to "begin" and "end" outside
		576	the loop, which serves two important purposes: first, it sets the
		577	callback to be called once the counter reaches 0, and second, it
		578	ensures that "send" is called even when "no" hosts are being pinged
		579	(the loop doesn't execute once).
		580
		581	This is the general pattern when you "fan out" into multiple
		582	subrequests: use an outer "begin"/"end" pair to set the callback and
		583	ensure "end" is called at least once, and then, for each subrequest
		584	you start, call "begin" and for each subrequest you finish, call
		585	"end".
		586
		587	METHODS FOR CONSUMERS
		588	These methods should only be used by the consuming side, i.e. the code
		589	awaits the condition.
		590
		591	$cv->recv
120	Wait (blocking if necessary) until the "->broadcast" method has been	592	Wait (blocking if necessary) until the "->send" or "->croak" methods
121	called on c<$cv>, while servicing other watchers normally.	593	have been called on c<$cv>, while servicing other watchers normally.
122		594
		595	You can only wait once on a condition - additional calls are valid
		596	but will return immediately.
		597
		598	If an error condition has been set by calling "->croak", then this
		599	function will call "croak".
		600
		601	In list context, all parameters passed to "send" will be returned,
		602	in scalar context only the first one will be returned.
		603
123	Not all event models support a blocking wait - some die in that	604	Not all event models support a blocking wait - some die in that case
124	case, so if you are using this from a module, never require a	605	(programs might want to do that to stay interactive), so *if you are
125	blocking wait, but let the caller decide wether the call will block	606	using this from a module, never require a blocking wait*, but let
126	or not (for example, by coupling condition variables with some kind	607	the caller decide whether the call will block or not (for example,
		608	by coupling condition variables with some kind of request results
127	of request results and supporting callbacks so the caller knows that	609	and supporting callbacks so the caller knows that getting the result
128	getting the result will not block, while still suppporting blockign	610	will not block, while still supporting blocking waits if the caller
129	waits if the caller so desires).	611	so desires).
130		612
131	You can only wait once on a condition - additional calls will return	613	Another reason *never* to "->recv" in a module is that you cannot
132	immediately.	614	sensibly have two "->recv"'s in parallel, as that would require
		615	multiple interpreters or coroutines/threads, none of which
		616	"AnyEvent" can supply.
133		617
134	$cv->broadcast	618	The Coro module, however, *can* and *does* supply coroutines and, in
135	Flag the condition as ready - a running "->wait" and all further	619	fact, Coro::AnyEvent replaces AnyEvent's condvars by coroutine-safe
136	calls to "wait" will return after this method has been called. If	620	versions and also integrates coroutines into AnyEvent, making
137	nobody is waiting the broadcast will be remembered..	621	blocking "->recv" calls perfectly safe as long as they are done from
		622	another coroutine (one that doesn't run the event loop).
138		623
139	Example:	624	You can ensure that "-recv" never blocks by setting a callback and
		625	only calling "->recv" from within that callback (or at a later
		626	time). This will work even when the event loop does not support
		627	blocking waits otherwise.
140		628
141	# wait till the result is ready	629	$bool = $cv->ready
142	my $result_ready = AnyEvent->condvar;	630	Returns true when the condition is "true", i.e. whether "send" or
		631	"croak" have been called.
143		632
144	# do something such as adding a timer	633	$cb = $cv->cb ($cb->($cv))
145	# or socket watcher the calls $result_ready->broadcast	634	This is a mutator function that returns the callback set and
146	# when the "result" is ready.	635	optionally replaces it before doing so.
147		636
148	$result_ready->wait;	637	The callback will be called when the condition becomes "true", i.e.
		638	when "send" or "croak" are called, with the only argument being the
		639	condition variable itself. Calling "recv" inside the callback or at
		640	any later time is guaranteed not to block.
		641
		642	GLOBAL VARIABLES AND FUNCTIONS
		643	$AnyEvent::MODEL
		644	Contains "undef" until the first watcher is being created. Then it
		645	contains the event model that is being used, which is the name of
		646	the Perl class implementing the model. This class is usually one of
		647	the "AnyEvent::Impl:xxx" modules, but can be any other class in the
		648	case AnyEvent has been extended at runtime (e.g. in *rxvt-unicode*).
		649
		650	The known classes so far are:
		651
		652	AnyEvent::Impl::EV based on EV (an interface to libev, best choice).
		653	AnyEvent::Impl::Event based on Event, second best choice.
		654	AnyEvent::Impl::Perl pure-perl implementation, fast and portable.
		655	AnyEvent::Impl::Glib based on Glib, third-best choice.
		656	AnyEvent::Impl::Tk based on Tk, very bad choice.
		657	AnyEvent::Impl::Qt based on Qt, cannot be autoprobed (see its docs).
		658	AnyEvent::Impl::EventLib based on Event::Lib, leaks memory and worse.
		659	AnyEvent::Impl::POE based on POE, not generic enough for full support.
		660
		661	There is no support for WxWidgets, as WxWidgets has no support for
		662	watching file handles. However, you can use WxWidgets through the
		663	POE Adaptor, as POE has a Wx backend that simply polls 20 times per
		664	second, which was considered to be too horrible to even consider for
		665	AnyEvent. Likewise, other POE backends can be used by AnyEvent by
		666	using it's adaptor.
		667
		668	AnyEvent knows about Prima and Wx and will try to use POE when
		669	autodetecting them.
		670
		671	AnyEvent::detect
		672	Returns $AnyEvent::MODEL, forcing autodetection of the event model
		673	if necessary. You should only call this function right before you
		674	would have created an AnyEvent watcher anyway, that is, as late as
		675	possible at runtime.
		676
		677	$guard = AnyEvent::post_detect { BLOCK }
		678	Arranges for the code block to be executed as soon as the event
		679	model is autodetected (or immediately if this has already happened).
		680
		681	If called in scalar or list context, then it creates and returns an
		682	object that automatically removes the callback again when it is
		683	destroyed. See Coro::BDB for a case where this is useful.
		684
		685	@AnyEvent::post_detect
		686	If there are any code references in this array (you can "push" to it
		687	before or after loading AnyEvent), then they will called directly
		688	after the event loop has been chosen.
		689
		690	You should check $AnyEvent::MODEL before adding to this array,
		691	though: if it contains a true value then the event loop has already
		692	been detected, and the array will be ignored.
		693
		694	Best use "AnyEvent::post_detect { BLOCK }" instead.
149		695
150	WHAT TO DO IN A MODULE	696	WHAT TO DO IN A MODULE
151	As a module author, you should "use AnyEvent" and call AnyEvent methods	697	As a module author, you should "use AnyEvent" and call AnyEvent methods
152	freely, but you should not load a specific event module or rely on it.	698	freely, but you should not load a specific event module or rely on it.
153		699
154	Be careful when you create watchers in the module body - Anyevent will	700	Be careful when you create watchers in the module body - AnyEvent will
155	decide which event module to use as soon as the first method is called,	701	decide which event module to use as soon as the first method is called,
156	so by calling AnyEvent in your module body you force the user of your	702	so by calling AnyEvent in your module body you force the user of your
157	module to load the event module first.	703	module to load the event module first.
158		704
		705	Never call "->recv" on a condition variable unless you *know* that the
		706	"->send" method has been called on it already. This is because it will
		707	stall the whole program, and the whole point of using events is to stay
		708	interactive.
		709
		710	It is fine, however, to call "->recv" when the user of your module
		711	requests it (i.e. if you create a http request object ad have a method
		712	called "results" that returns the results, it should call "->recv"
		713	freely, as the user of your module knows what she is doing. always).
		714
159	WHAT TO DO IN THE MAIN PROGRAM	715	WHAT TO DO IN THE MAIN PROGRAM
160	There will always be a single main program - the only place that should	716	There will always be a single main program - the only place that should
161	dictate which event model to use.	717	dictate which event model to use.
162		718
163	If it doesn't care, it can just "use AnyEvent" and use it itself, or not	719	If it doesn't care, it can just "use AnyEvent" and use it itself, or not
164	do anything special and let AnyEvent decide which implementation to	720	do anything special (it does not need to be event-based) and let
165	chose.	721	AnyEvent decide which implementation to chose if some module relies on
		722	it.
166		723
167	If the main program relies on a specific event model (for example, in	724	If the main program relies on a specific event model - for example, in
168	Gtk2 programs you have to rely on either Glib or Glib::Event), you	725	Gtk2 programs you have to rely on the Glib module - you should load the
169	should load it before loading AnyEvent or any module that uses it,	726	event module before loading AnyEvent or any module that uses it:
170	generally, as early as possible. The reason is that modules might create	727	generally speaking, you should load it as early as possible. The reason
171	watchers when they are loaded, and AnyEvent will decide on the event	728	is that modules might create watchers when they are loaded, and AnyEvent
172	model to use as soon as it creates watchers, and it might chose the	729	will decide on the event model to use as soon as it creates watchers,
173	wrong one unless you load the correct one yourself.	730	and it might chose the wrong one unless you load the correct one
		731	yourself.
174		732
175	You can chose to use a rather inefficient pure-perl implementation by	733	You can chose to use a pure-perl implementation by loading the
176	loading the "AnyEvent::Impl::Perl" module, but letting AnyEvent chose is	734	"AnyEvent::Impl::Perl" module, which gives you similar behaviour
177	generally better.	735	everywhere, but letting AnyEvent chose the model is generally better.
		736
		737	MAINLOOP EMULATION
		738	Sometimes (often for short test scripts, or even standalone programs who
		739	only want to use AnyEvent), you do not want to run a specific event
		740	loop.
		741
		742	In that case, you can use a condition variable like this:
		743
		744	AnyEvent->condvar->recv;
		745
		746	This has the effect of entering the event loop and looping forever.
		747
		748	Note that usually your program has some exit condition, in which case it
		749	is better to use the "traditional" approach of storing a condition
		750	variable somewhere, waiting for it, and sending it when the program
		751	should exit cleanly.
		752
		753	OTHER MODULES
		754	The following is a non-exhaustive list of additional modules that use
		755	AnyEvent and can therefore be mixed easily with other AnyEvent modules
		756	in the same program. Some of the modules come with AnyEvent, some are
		757	available via CPAN.
		758
		759	AnyEvent::Util
		760	Contains various utility functions that replace often-used but
		761	blocking functions such as "inet_aton" by event-/callback-based
		762	versions.
		763
		764	AnyEvent::Socket
		765	Provides various utility functions for (internet protocol) sockets,
		766	addresses and name resolution. Also functions to create non-blocking
		767	tcp connections or tcp servers, with IPv6 and SRV record support and
		768	more.
		769
		770	AnyEvent::Handle
		771	Provide read and write buffers, manages watchers for reads and
		772	writes, supports raw and formatted I/O, I/O queued and fully
		773	transparent and non-blocking SSL/TLS.
		774
		775	AnyEvent::DNS
		776	Provides rich asynchronous DNS resolver capabilities.
		777
		778	AnyEvent::HTTP
		779	A simple-to-use HTTP library that is capable of making a lot of
		780	concurrent HTTP requests.
		781
		782	AnyEvent::HTTPD
		783	Provides a simple web application server framework.
		784
		785	AnyEvent::FastPing
		786	The fastest ping in the west.
		787
		788	AnyEvent::DBI
		789	Executes DBI requests asynchronously in a proxy process.
		790
		791	AnyEvent::AIO
		792	Truly asynchronous I/O, should be in the toolbox of every event
		793	programmer. AnyEvent::AIO transparently fuses IO::AIO and AnyEvent
		794	together.
		795
		796	AnyEvent::BDB
		797	Truly asynchronous Berkeley DB access. AnyEvent::BDB transparently
		798	fuses BDB and AnyEvent together.
		799
		800	AnyEvent::GPSD
		801	A non-blocking interface to gpsd, a daemon delivering GPS
		802	information.
		803
		804	AnyEvent::IGS
		805	A non-blocking interface to the Internet Go Server protocol (used by
		806	App::IGS).
		807
		808	AnyEvent::IRC
		809	AnyEvent based IRC client module family (replacing the older
		810	Net::IRC3).
		811
		812	Net::XMPP2
		813	AnyEvent based XMPP (Jabber protocol) module family.
		814
		815	Net::FCP
		816	AnyEvent-based implementation of the Freenet Client Protocol,
		817	birthplace of AnyEvent.
		818
		819	Event::ExecFlow
		820	High level API for event-based execution flow control.
		821
		822	Coro
		823	Has special support for AnyEvent via Coro::AnyEvent.
		824
		825	IO::Lambda
		826	The lambda approach to I/O - don't ask, look there. Can use
		827	AnyEvent.
		828
		829	ERROR AND EXCEPTION HANDLING
		830	In general, AnyEvent does not do any error handling - it relies on the
		831	caller to do that if required. The AnyEvent::Strict module (see also the
		832	"PERL_ANYEVENT_STRICT" environment variable, below) provides strict
		833	checking of all AnyEvent methods, however, which is highly useful during
		834	development.
		835
		836	As for exception handling (i.e. runtime errors and exceptions thrown
		837	while executing a callback), this is not only highly event-loop
		838	specific, but also not in any way wrapped by this module, as this is the
		839	job of the main program.
		840
		841	The pure perl event loop simply re-throws the exception (usually within
		842	"condvar->recv"), the Event and EV modules call "$Event/EV::DIED->()",
		843	Glib uses "install_exception_handler" and so on.
		844
		845	ENVIRONMENT VARIABLES
		846	The following environment variables are used by this module or its
		847	submodules:
		848
		849	"PERL_ANYEVENT_VERBOSE"
		850	By default, AnyEvent will be completely silent except in fatal
		851	conditions. You can set this environment variable to make AnyEvent
		852	more talkative.
		853
		854	When set to 1 or higher, causes AnyEvent to warn about unexpected
		855	conditions, such as not being able to load the event model specified
		856	by "PERL_ANYEVENT_MODEL".
		857
		858	When set to 2 or higher, cause AnyEvent to report to STDERR which
		859	event model it chooses.
		860
		861	"PERL_ANYEVENT_STRICT"
		862	AnyEvent does not do much argument checking by default, as thorough
		863	argument checking is very costly. Setting this variable to a true
		864	value will cause AnyEvent to load "AnyEvent::Strict" and then to
		865	thoroughly check the arguments passed to most method calls. If it
		866	finds any problems it will croak.
		867
		868	In other words, enables "strict" mode.
		869
		870	Unlike "use strict", it is definitely recommended ot keep it off in
		871	production. Keeping "PERL_ANYEVENT_STRICT=1" in your environment
		872	while developing programs can be very useful, however.
		873
		874	"PERL_ANYEVENT_MODEL"
		875	This can be used to specify the event model to be used by AnyEvent,
		876	before auto detection and -probing kicks in. It must be a string
		877	consisting entirely of ASCII letters. The string "AnyEvent::Impl::"
		878	gets prepended and the resulting module name is loaded and if the
		879	load was successful, used as event model. If it fails to load
		880	AnyEvent will proceed with auto detection and -probing.
		881
		882	This functionality might change in future versions.
		883
		884	For example, to force the pure perl model (AnyEvent::Impl::Perl) you
		885	could start your program like this:
		886
		887	PERL_ANYEVENT_MODEL=Perl perl ...
		888
		889	"PERL_ANYEVENT_PROTOCOLS"
		890	Used by both AnyEvent::DNS and AnyEvent::Socket to determine
		891	preferences for IPv4 or IPv6. The default is unspecified (and might
		892	change, or be the result of auto probing).
		893
		894	Must be set to a comma-separated list of protocols or address
		895	families, current supported: "ipv4" and "ipv6". Only protocols
		896	mentioned will be used, and preference will be given to protocols
		897	mentioned earlier in the list.
		898
		899	This variable can effectively be used for denial-of-service attacks
		900	against local programs (e.g. when setuid), although the impact is
		901	likely small, as the program has to handle conenction and other
		902	failures anyways.
		903
		904	Examples: "PERL_ANYEVENT_PROTOCOLS=ipv4,ipv6" - prefer IPv4 over
		905	IPv6, but support both and try to use both.
		906	"PERL_ANYEVENT_PROTOCOLS=ipv4" - only support IPv4, never try to
		907	resolve or contact IPv6 addresses.
		908	"PERL_ANYEVENT_PROTOCOLS=ipv6,ipv4" support either IPv4 or IPv6, but
		909	prefer IPv6 over IPv4.
		910
		911	"PERL_ANYEVENT_EDNS0"
		912	Used by AnyEvent::DNS to decide whether to use the EDNS0 extension
		913	for DNS. This extension is generally useful to reduce DNS traffic,
		914	but some (broken) firewalls drop such DNS packets, which is why it
		915	is off by default.
		916
		917	Setting this variable to 1 will cause AnyEvent::DNS to announce
		918	EDNS0 in its DNS requests.
		919
		920	"PERL_ANYEVENT_MAX_FORKS"
		921	The maximum number of child processes that
		922	"AnyEvent::Util::fork_call" will create in parallel.
178		923
179	SUPPLYING YOUR OWN EVENT MODEL INTERFACE	924	SUPPLYING YOUR OWN EVENT MODEL INTERFACE
		925	This is an advanced topic that you do not normally need to use AnyEvent
		926	in a module. This section is only of use to event loop authors who want
		927	to provide AnyEvent compatibility.
		928
180	If you need to support another event library which isn't directly	929	If you need to support another event library which isn't directly
181	supported by AnyEvent, you can supply your own interface to it by	930	supported by AnyEvent, you can supply your own interface to it by
182	pushing, before the first watcher gets created, the package name of the	931	pushing, before the first watcher gets created, the package name of the
183	event module and the package name of the interface to use onto	932	event module and the package name of the interface to use onto
184	@AnyEvent::REGISTRY. You can do that before and even without loading	933	@AnyEvent::REGISTRY. You can do that before and even without loading
185	AnyEvent.	934	AnyEvent, so it is reasonably cheap.
186		935
187	Example:	936	Example:
188		937
189	push @AnyEvent::REGISTRY, [urxvt => urxvt::anyevent::];	938	push @AnyEvent::REGISTRY, [urxvt => urxvt::anyevent::];
190		939
191	This tells AnyEvent to (literally) use the "urxvt::anyevent::"	940	This tells AnyEvent to (literally) use the "urxvt::anyevent::"
192	package/class when it finds the "urxvt" package/module is loaded. When	941	package/class when it finds the "urxvt" package/module is already
		942	loaded.
		943
193	AnyEvent is loaded and asked to find a suitable event model, it will	944	When AnyEvent is loaded and asked to find a suitable event model, it
194	first check for the presence of urxvt.	945	will first check for the presence of urxvt by trying to "use" the
		946	"urxvt::anyevent" module.
195		947
196	The class should prove implementations for all watcher types (see	948	The class should provide implementations for all watcher types. See
197	AnyEvent::Impl::Event (source code), AnyEvent::Impl::Glib (Source code)	949	AnyEvent::Impl::EV (source code), AnyEvent::Impl::Glib (Source code) and
198	and so on for actual examples, use "perldoc -m AnyEvent::Impl::Glib" to	950	so on for actual examples. Use "perldoc -m AnyEvent::Impl::Glib" to see
199	see the sources).	951	the sources.
200		952
		953	If you don't provide "signal" and "child" watchers than AnyEvent will
		954	provide suitable (hopefully) replacements.
		955
201	The above isn't fictitious, the *rxvt-unicode* (a.k.a. urxvt) uses the	956	The above example isn't fictitious, the *rxvt-unicode* (a.k.a. urxvt)
202	above line as-is. An interface isn't included in AnyEvent because it	957	terminal emulator uses the above line as-is. An interface isn't included
203	doesn't make sense outside the embedded interpreter inside	958	in AnyEvent because it doesn't make sense outside the embedded
204	*rxvt-unicode*, and it is updated and maintained as part of the	959	interpreter inside *rxvt-unicode*, and it is updated and maintained as
205	*rxvt-unicode* distribution.	960	part of the *rxvt-unicode* distribution.
206		961
207	*rxvt-unicode* also cheats a bit by not providing blocking access to	962	*rxvt-unicode* also cheats a bit by not providing blocking access to
208	condition variables: code blocking while waiting for a condition will	963	condition variables: code blocking while waiting for a condition will
209	"die". This still works with most modules/usages, and blocking calls	964	"die". This still works with most modules/usages, and blocking calls
210	must not be in an interactive appliation, so it makes sense.	965	must not be done in an interactive application, so it makes sense.
211		966
212	ENVIRONMENT VARIABLES	967	EXAMPLE PROGRAM
213	The following environment variables are used by this module:
214
215	"PERL_ANYEVENT_VERBOSE" when set to 2 or higher, reports which event
216	model gets used.
217
218	EXAMPLE
219	The following program uses an io watcher to read data from stdin, a	968	The following program uses an I/O watcher to read data from STDIN, a
220	timer to display a message once per second, and a condvar to exit the	969	timer to display a message once per second, and a condition variable to
221	program when the user enters quit:	970	quit the program when the user enters quit:
222		971
223	use AnyEvent;	972	use AnyEvent;
224		973
225	my $cv = AnyEvent->condvar;	974	my $cv = AnyEvent->condvar;
226		975
227	my $io_watcher = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {	976	my $io_watcher = AnyEvent->io (
		977	fh => \*STDIN,
		978	poll => 'r',
		979	cb => sub {
228	warn "io event <$_[0]>\n"; # will always output <r>	980	warn "io event <$_[0]>\n"; # will always output <r>
229	chomp (my $input = <STDIN>); # read a line	981	chomp (my $input = <STDIN>); # read a line
230	warn "read: $input\n"; # output what has been read	982	warn "read: $input\n"; # output what has been read
231	$cv->broadcast if $input =~ /^q/i; # quit program if /^q/i	983	$cv->send if $input =~ /^q/i; # quit program if /^q/i
		984	},
232	});	985	);
233		986
234	my $time_watcher; # can only be used once	987	my $time_watcher; # can only be used once
235		988
236	sub new_timer {	989	sub new_timer {
237	$timer = AnyEvent->timer (after => 1, cb => sub {	990	$timer = AnyEvent->timer (after => 1, cb => sub {
…		…
240	});	993	});
241	}	994	}
242		995
243	new_timer; # create first timer	996	new_timer; # create first timer
244		997
245	$cv->wait; # wait until user enters /^q/i	998	$cv->recv; # wait until user enters /^q/i
246		999
247	REAL-WORLD EXAMPLE	1000	REAL-WORLD EXAMPLE
248	Consider the Net::FCP module. It features (among others) the following	1001	Consider the Net::FCP module. It features (among others) the following
249	API calls, which are to freenet what HTTP GET requests are to http:	1002	API calls, which are to freenet what HTTP GET requests are to http:
250		1003
…		…
299	syswrite $txn->{fh}, $txn->{request}	1052	syswrite $txn->{fh}, $txn->{request}
300	or die "connection or write error";	1053	or die "connection or write error";
301	$txn->{w} = AnyEvent->io (fh => $txn->{fh}, poll => 'r', cb => sub { $txn->fh_ready_r });	1054	$txn->{w} = AnyEvent->io (fh => $txn->{fh}, poll => 'r', cb => sub { $txn->fh_ready_r });
302		1055
303	Again, "fh_ready_r" waits till all data has arrived, and then stores the	1056	Again, "fh_ready_r" waits till all data has arrived, and then stores the
304	result and signals any possible waiters that the request ahs finished:	1057	result and signals any possible waiters that the request has finished:
305		1058
306	sysread $txn->{fh}, $txn->{buf}, length $txn->{$buf};	1059	sysread $txn->{fh}, $txn->{buf}, length $txn->{$buf};
307		1060
308	if (end-of-file or data complete) {	1061	if (end-of-file or data complete) {
309	$txn->{result} = $txn->{buf};	1062	$txn->{result} = $txn->{buf};
310	$txn->{finished}->broadcast;	1063	$txn->{finished}->send;
311	$txb->{cb}->($txn) of $txn->{cb}; # also call callback	1064	$txb->{cb}->($txn) of $txn->{cb}; # also call callback
312	}	1065	}
313		1066
314	The "result" method, finally, just waits for the finished signal (if the	1067	The "result" method, finally, just waits for the finished signal (if the
315	request was already finished, it doesn't wait, of course, and returns	1068	request was already finished, it doesn't wait, of course, and returns
316	the data:	1069	the data:
317		1070
318	$txn->{finished}->wait;	1071	$txn->{finished}->recv;
319	return $txn->{result};	1072	return $txn->{result};
320		1073
321	The actual code goes further and collects all errors ("die"s,	1074	The actual code goes further and collects all errors ("die"s,
322	exceptions) that occured during request processing. The "result" method	1075	exceptions) that occurred during request processing. The "result" method
323	detects wether an exception as thrown (it is stored inside the $txn	1076	detects whether an exception as thrown (it is stored inside the $txn
324	object) and just throws the exception, which means connection errors and	1077	object) and just throws the exception, which means connection errors and
325	other problems get reported tot he code that tries to use the result,	1078	other problems get reported tot he code that tries to use the result,
326	not in a random callback.	1079	not in a random callback.
327		1080
328	All of this enables the following usage styles:	1081	All of this enables the following usage styles:
329		1082
330	1. Blocking:	1083	1. Blocking:
331		1084
332	my $data = $fcp->client_get ($url);	1085	my $data = $fcp->client_get ($url);
333		1086
334	2. Blocking, but parallelizing:	1087	2. Blocking, but running in parallel:
335		1088
336	my @datas = map $_->result,	1089	my @datas = map $_->result,
337	map $fcp->txn_client_get ($_),	1090	map $fcp->txn_client_get ($_),
338	@urls;	1091	@urls;
339		1092
340	Both blocking examples work without the module user having to know	1093	Both blocking examples work without the module user having to know
341	anything about events.	1094	anything about events.
342		1095
343	3a. Event-based in a main program, using any support Event module:	1096	3a. Event-based in a main program, using any supported event module:
344		1097
345	use Event;	1098	use EV;
346		1099
347	$fcp->txn_client_get ($url)->cb (sub {	1100	$fcp->txn_client_get ($url)->cb (sub {
348	my $txn = shift;	1101	my $txn = shift;
349	my $data = $txn->result;	1102	my $data = $txn->result;
350	...	1103	...
351	});	1104	});
352		1105
353	Event::loop;	1106	EV::loop;
354		1107
355	3b. The module user could use AnyEvent, too:	1108	3b. The module user could use AnyEvent, too:
356		1109
357	use AnyEvent;	1110	use AnyEvent;
358		1111
359	my $quit = AnyEvent->condvar;	1112	my $quit = AnyEvent->condvar;
360		1113
361	$fcp->txn_client_get ($url)->cb (sub {	1114	$fcp->txn_client_get ($url)->cb (sub {
362	...	1115	...
363	$quit->broadcast;	1116	$quit->send;
364	});	1117	});
365		1118
366	$quit->wait;	1119	$quit->recv;
		1120
		1121	BENCHMARKS
		1122	To give you an idea of the performance and overheads that AnyEvent adds
		1123	over the event loops themselves and to give you an impression of the
		1124	speed of various event loops I prepared some benchmarks.
		1125
		1126	BENCHMARKING ANYEVENT OVERHEAD
		1127	Here is a benchmark of various supported event models used natively and
		1128	through AnyEvent. The benchmark creates a lot of timers (with a zero
		1129	timeout) and I/O watchers (watching STDOUT, a pty, to become writable,
		1130	which it is), lets them fire exactly once and destroys them again.
		1131
		1132	Source code for this benchmark is found as eg/bench in the AnyEvent
		1133	distribution.
		1134
		1135	Explanation of the columns
		1136	*watcher* is the number of event watchers created/destroyed. Since
		1137	different event models feature vastly different performances, each event
		1138	loop was given a number of watchers so that overall runtime is
		1139	acceptable and similar between tested event loop (and keep them from
		1140	crashing): Glib would probably take thousands of years if asked to
		1141	process the same number of watchers as EV in this benchmark.
		1142
		1143	*bytes* is the number of bytes (as measured by the resident set size,
		1144	RSS) consumed by each watcher. This method of measuring captures both C
		1145	and Perl-based overheads.
		1146
		1147	*create* is the time, in microseconds (millionths of seconds), that it
		1148	takes to create a single watcher. The callback is a closure shared
		1149	between all watchers, to avoid adding memory overhead. That means
		1150	closure creation and memory usage is not included in the figures.
		1151
		1152	*invoke* is the time, in microseconds, used to invoke a simple callback.
		1153	The callback simply counts down a Perl variable and after it was invoked
		1154	"watcher" times, it would "->send" a condvar once to signal the end of
		1155	this phase.
		1156
		1157	*destroy* is the time, in microseconds, that it takes to destroy a
		1158	single watcher.
		1159
		1160	Results
		1161	name watchers bytes create invoke destroy comment
		1162	EV/EV 400000 224 0.47 0.35 0.27 EV native interface
		1163	EV/Any 100000 224 2.88 0.34 0.27 EV + AnyEvent watchers
		1164	CoroEV/Any 100000 224 2.85 0.35 0.28 coroutines + Coro::Signal
		1165	Perl/Any 100000 452 4.13 0.73 0.95 pure perl implementation
		1166	Event/Event 16000 517 32.20 31.80 0.81 Event native interface
		1167	Event/Any 16000 590 35.85 31.55 1.06 Event + AnyEvent watchers
		1168	Glib/Any 16000 1357 102.33 12.31 51.00 quadratic behaviour
		1169	Tk/Any 2000 1860 27.20 66.31 14.00 SEGV with >> 2000 watchers
		1170	POE/Event 2000 6328 109.99 751.67 14.02 via POE::Loop::Event
		1171	POE/Select 2000 6027 94.54 809.13 579.80 via POE::Loop::Select
		1172
		1173	Discussion
		1174	The benchmark does *not* measure scalability of the event loop very
		1175	well. For example, a select-based event loop (such as the pure perl one)
		1176	can never compete with an event loop that uses epoll when the number of
		1177	file descriptors grows high. In this benchmark, all events become ready
		1178	at the same time, so select/poll-based implementations get an unnatural
		1179	speed boost.
		1180
		1181	Also, note that the number of watchers usually has a nonlinear effect on
		1182	overall speed, that is, creating twice as many watchers doesn't take
		1183	twice the time - usually it takes longer. This puts event loops tested
		1184	with a higher number of watchers at a disadvantage.
		1185
		1186	To put the range of results into perspective, consider that on the
		1187	benchmark machine, handling an event takes roughly 1600 CPU cycles with
		1188	EV, 3100 CPU cycles with AnyEvent's pure perl loop and almost 3000000
		1189	CPU cycles with POE.
		1190
		1191	"EV" is the sole leader regarding speed and memory use, which are both
		1192	maximal/minimal, respectively. Even when going through AnyEvent, it uses
		1193	far less memory than any other event loop and is still faster than Event
		1194	natively.
		1195
		1196	The pure perl implementation is hit in a few sweet spots (both the
		1197	constant timeout and the use of a single fd hit optimisations in the
		1198	perl interpreter and the backend itself). Nevertheless this shows that
		1199	it adds very little overhead in itself. Like any select-based backend
		1200	its performance becomes really bad with lots of file descriptors (and
		1201	few of them active), of course, but this was not subject of this
		1202	benchmark.
		1203
		1204	The "Event" module has a relatively high setup and callback invocation
		1205	cost, but overall scores in on the third place.
		1206
		1207	"Glib"'s memory usage is quite a bit higher, but it features a faster
		1208	callback invocation and overall ends up in the same class as "Event".
		1209	However, Glib scales extremely badly, doubling the number of watchers
		1210	increases the processing time by more than a factor of four, making it
		1211	completely unusable when using larger numbers of watchers (note that
		1212	only a single file descriptor was used in the benchmark, so
		1213	inefficiencies of "poll" do not account for this).
		1214
		1215	The "Tk" adaptor works relatively well. The fact that it crashes with
		1216	more than 2000 watchers is a big setback, however, as correctness takes
		1217	precedence over speed. Nevertheless, its performance is surprising, as
		1218	the file descriptor is dup()ed for each watcher. This shows that the
		1219	dup() employed by some adaptors is not a big performance issue (it does
		1220	incur a hidden memory cost inside the kernel which is not reflected in
		1221	the figures above).
		1222
		1223	"POE", regardless of underlying event loop (whether using its pure perl
		1224	select-based backend or the Event module, the POE-EV backend couldn't be
		1225	tested because it wasn't working) shows abysmal performance and memory
		1226	usage with AnyEvent: Watchers use almost 30 times as much memory as EV
		1227	watchers, and 10 times as much memory as Event (the high memory
		1228	requirements are caused by requiring a session for each watcher).
		1229	Watcher invocation speed is almost 900 times slower than with AnyEvent's
		1230	pure perl implementation.
		1231
		1232	The design of the POE adaptor class in AnyEvent can not really account
		1233	for the performance issues, though, as session creation overhead is
		1234	small compared to execution of the state machine, which is coded pretty
		1235	optimally within AnyEvent::Impl::POE (and while everybody agrees that
		1236	using multiple sessions is not a good approach, especially regarding
		1237	memory usage, even the author of POE could not come up with a faster
		1238	design).
		1239
		1240	Summary
		1241	* Using EV through AnyEvent is faster than any other event loop (even
		1242	when used without AnyEvent), but most event loops have acceptable
		1243	performance with or without AnyEvent.
		1244
		1245	* The overhead AnyEvent adds is usually much smaller than the overhead
		1246	of the actual event loop, only with extremely fast event loops such
		1247	as EV adds AnyEvent significant overhead.
		1248
		1249	* You should avoid POE like the plague if you want performance or
		1250	reasonable memory usage.
		1251
		1252	BENCHMARKING THE LARGE SERVER CASE
		1253	This benchmark actually benchmarks the event loop itself. It works by
		1254	creating a number of "servers": each server consists of a socket pair, a
		1255	timeout watcher that gets reset on activity (but never fires), and an
		1256	I/O watcher waiting for input on one side of the socket. Each time the
		1257	socket watcher reads a byte it will write that byte to a random other
		1258	"server".
		1259
		1260	The effect is that there will be a lot of I/O watchers, only part of
		1261	which are active at any one point (so there is a constant number of
		1262	active fds for each loop iteration, but which fds these are is random).
		1263	The timeout is reset each time something is read because that reflects
		1264	how most timeouts work (and puts extra pressure on the event loops).
		1265
		1266	In this benchmark, we use 10000 socket pairs (20000 sockets), of which
		1267	100 (1%) are active. This mirrors the activity of large servers with
		1268	many connections, most of which are idle at any one point in time.
		1269
		1270	Source code for this benchmark is found as eg/bench2 in the AnyEvent
		1271	distribution.
		1272
		1273	Explanation of the columns
		1274	*sockets* is the number of sockets, and twice the number of "servers"
		1275	(as each server has a read and write socket end).
		1276
		1277	*create* is the time it takes to create a socket pair (which is
		1278	nontrivial) and two watchers: an I/O watcher and a timeout watcher.
		1279
		1280	*request*, the most important value, is the time it takes to handle a
		1281	single "request", that is, reading the token from the pipe and
		1282	forwarding it to another server. This includes deleting the old timeout
		1283	and creating a new one that moves the timeout into the future.
		1284
		1285	Results
		1286	name sockets create request
		1287	EV 20000 69.01 11.16
		1288	Perl 20000 73.32 35.87
		1289	Event 20000 212.62 257.32
		1290	Glib 20000 651.16 1896.30
		1291	POE 20000 349.67 12317.24 uses POE::Loop::Event
		1292
		1293	Discussion
		1294	This benchmark *does* measure scalability and overall performance of the
		1295	particular event loop.
		1296
		1297	EV is again fastest. Since it is using epoll on my system, the setup
		1298	time is relatively high, though.
		1299
		1300	Perl surprisingly comes second. It is much faster than the C-based event
		1301	loops Event and Glib.
		1302
		1303	Event suffers from high setup time as well (look at its code and you
		1304	will understand why). Callback invocation also has a high overhead
		1305	compared to the "$_->() for .."-style loop that the Perl event loop
		1306	uses. Event uses select or poll in basically all documented
		1307	configurations.
		1308
		1309	Glib is hit hard by its quadratic behaviour w.r.t. many watchers. It
		1310	clearly fails to perform with many filehandles or in busy servers.
		1311
		1312	POE is still completely out of the picture, taking over 1000 times as
		1313	long as EV, and over 100 times as long as the Perl implementation, even
		1314	though it uses a C-based event loop in this case.
		1315
		1316	Summary
		1317	* The pure perl implementation performs extremely well.
		1318
		1319	* Avoid Glib or POE in large projects where performance matters.
		1320
		1321	BENCHMARKING SMALL SERVERS
		1322	While event loops should scale (and select-based ones do not...) even to
		1323	large servers, most programs we (or I :) actually write have only a few
		1324	I/O watchers.
		1325
		1326	In this benchmark, I use the same benchmark program as in the large
		1327	server case, but it uses only eight "servers", of which three are active
		1328	at any one time. This should reflect performance for a small server
		1329	relatively well.
		1330
		1331	The columns are identical to the previous table.
		1332
		1333	Results
		1334	name sockets create request
		1335	EV 16 20.00 6.54
		1336	Perl 16 25.75 12.62
		1337	Event 16 81.27 35.86
		1338	Glib 16 32.63 15.48
		1339	POE 16 261.87 276.28 uses POE::Loop::Event
		1340
		1341	Discussion
		1342	The benchmark tries to test the performance of a typical small server.
		1343	While knowing how various event loops perform is interesting, keep in
		1344	mind that their overhead in this case is usually not as important, due
		1345	to the small absolute number of watchers (that is, you need efficiency
		1346	and speed most when you have lots of watchers, not when you only have a
		1347	few of them).
		1348
		1349	EV is again fastest.
		1350
		1351	Perl again comes second. It is noticeably faster than the C-based event
		1352	loops Event and Glib, although the difference is too small to really
		1353	matter.
		1354
		1355	POE also performs much better in this case, but is is still far behind
		1356	the others.
		1357
		1358	Summary
		1359	* C-based event loops perform very well with small number of watchers,
		1360	as the management overhead dominates.
		1361
		1362	SIGNALS
		1363	AnyEvent currently installs handlers for these signals:
		1364
		1365	SIGCHLD
		1366	A handler for "SIGCHLD" is installed by AnyEvent's child watcher
		1367	emulation for event loops that do not support them natively. Also,
		1368	some event loops install a similar handler.
		1369
		1370	SIGPIPE
		1371	A no-op handler is installed for "SIGPIPE" when $SIG{PIPE} is
		1372	"undef" when AnyEvent gets loaded.
		1373
		1374	The rationale for this is that AnyEvent users usually do not really
		1375	depend on SIGPIPE delivery (which is purely an optimisation for
		1376	shell use, or badly-written programs), but "SIGPIPE" can cause
		1377	spurious and rare program exits as a lot of people do not expect
		1378	"SIGPIPE" when writing to some random socket.
		1379
		1380	The rationale for installing a no-op handler as opposed to ignoring
		1381	it is that this way, the handler will be restored to defaults on
		1382	exec.
		1383
		1384	Feel free to install your own handler, or reset it to defaults.
		1385
		1386	FORK
		1387	Most event libraries are not fork-safe. The ones who are usually are
		1388	because they rely on inefficient but fork-safe "select" or "poll" calls.
		1389	Only EV is fully fork-aware.
		1390
		1391	If you have to fork, you must either do so *before* creating your first
		1392	watcher OR you must not use AnyEvent at all in the child.
		1393
		1394	SECURITY CONSIDERATIONS
		1395	AnyEvent can be forced to load any event model via
		1396	$ENV{PERL_ANYEVENT_MODEL}. While this cannot (to my knowledge) be used
		1397	to execute arbitrary code or directly gain access, it can easily be used
		1398	to make the program hang or malfunction in subtle ways, as AnyEvent
		1399	watchers will not be active when the program uses a different event
		1400	model than specified in the variable.
		1401
		1402	You can make AnyEvent completely ignore this variable by deleting it
		1403	before the first watcher gets created, e.g. with a "BEGIN" block:
		1404
		1405	BEGIN { delete $ENV{PERL_ANYEVENT_MODEL} }
		1406
		1407	use AnyEvent;
		1408
		1409	Similar considerations apply to $ENV{PERL_ANYEVENT_VERBOSE}, as that can
		1410	be used to probe what backend is used and gain other information (which
		1411	is probably even less useful to an attacker than PERL_ANYEVENT_MODEL),
		1412	and $ENV{PERL_ANYEGENT_STRICT}.
		1413
		1414	BUGS
		1415	Perl 5.8 has numerous memleaks that sometimes hit this module and are
		1416	hard to work around. If you suffer from memleaks, first upgrade to Perl
		1417	5.10 and check wether the leaks still show up. (Perl 5.10.0 has other
		1418	annoying memleaks, such as leaking on "map" and "grep" but it is usually
		1419	not as pronounced).
367		1420
368	SEE ALSO	1421	SEE ALSO
369	Event modules: Coro::Event, Coro, Event, Glib::Event, Glib.	1422	Utility functions: AnyEvent::Util.
370		1423
		1424	Event modules: EV, EV::Glib, Glib::EV, Event, Glib::Event, Glib, Tk,
		1425	Event::Lib, Qt, POE.
		1426
371	Implementations: AnyEvent::Impl::Coro, AnyEvent::Impl::Event,	1427	Implementations: AnyEvent::Impl::EV, AnyEvent::Impl::Event,
372	AnyEvent::Impl::Glib, AnyEvent::Impl::Tk.	1428	AnyEvent::Impl::Glib, AnyEvent::Impl::Tk, AnyEvent::Impl::Perl,
		1429	AnyEvent::Impl::EventLib, AnyEvent::Impl::Qt, AnyEvent::Impl::POE.
373		1430
374	Nontrivial usage example: Net::FCP.	1431	Non-blocking file handles, sockets, TCP clients and servers:
		1432	AnyEvent::Handle, AnyEvent::Socket.
375		1433
		1434	Asynchronous DNS: AnyEvent::DNS.
376		1435
		1436	Coroutine support: Coro, Coro::AnyEvent, Coro::EV, Coro::Event,
		1437
		1438	Nontrivial usage examples: Net::FCP, Net::XMPP2, AnyEvent::DNS.
		1439
		1440	AUTHOR
		1441	Marc Lehmann <schmorp@schmorp.de>
		1442	http://home.schmorp.de/
		1443

Diff Legend

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