ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/AnyEvent/lib/AnyEvent.pm

Comparing AnyEvent/lib/AnyEvent.pm (file contents):
Revision 1.28 by root, Sat Oct 27 15:10:09 2007 UTC vs.
Revision 1.148 by root, Sat May 31 00:40:16 2008 UTC

1	=head1 NAME	1	=head1 => NAME
2		2
3	AnyEvent - provide framework for multiple event loops	3	AnyEvent - provide framework for multiple event loops
4		4
5	Event, Coro, Glib, Tk, Perl - various supported event loops	5	EV, Event, Glib, Tk, Perl, Event::Lib, Qt, POE - various supported event loops
6		6
7	=head1 SYNOPSIS	7	=head1 SYNOPSIS
8		8
9	use AnyEvent;	9	use AnyEvent;
10		10
…		…
14		14
15	my $w = AnyEvent->timer (after => $seconds, cb => sub {	15	my $w = AnyEvent->timer (after => $seconds, cb => sub {
16	...	16	...
17	});	17	});
18		18
19	my $w = AnyEvent->condvar; # stores wether a condition was flagged	19	my $w = AnyEvent->condvar; # stores whether a condition was flagged
		20	$w->send; # wake up current and all future recv's
20	$w->wait; # enters "main loop" till $condvar gets ->broadcast	21	$w->recv; # enters "main loop" till $condvar gets ->send
21	$w->broadcast; # wake up current and all future wait's	22
		23	=head1 INTRODUCTION/TUTORIAL
		24
		25	This manpage is mainly a reference manual. If you are interested
		26	in a tutorial or some gentle introduction, have a look at the
		27	L<AnyEvent::Intro> manpage.
		28
		29	=head1 WHY YOU SHOULD USE THIS MODULE (OR NOT)
		30
		31	Glib, POE, IO::Async, Event... CPAN offers event models by the dozen
		32	nowadays. So what is different about AnyEvent?
		33
		34	Executive Summary: AnyEvent is I<compatible>, AnyEvent is I<free of
		35	policy> and AnyEvent is I<small and efficient>.
		36
		37	First and foremost, I<AnyEvent is not an event model> itself, it only
		38	interfaces to whatever event model the main program happens to use in a
		39	pragmatic way. For event models and certain classes of immortals alike,
		40	the statement "there can only be one" is a bitter reality: In general,
		41	only one event loop can be active at the same time in a process. AnyEvent
		42	helps hiding the differences between those event loops.
		43
		44	The goal of AnyEvent is to offer module authors the ability to do event
		45	programming (waiting for I/O or timer events) without subscribing to a
		46	religion, a way of living, and most importantly: without forcing your
		47	module users into the same thing by forcing them to use the same event
		48	model you use.
		49
		50	For modules like POE or IO::Async (which is a total misnomer as it is
		51	actually doing all I/O I<synchronously>...), using them in your module is
		52	like joining a cult: After you joined, you are dependent on them and you
		53	cannot use anything else, as it is simply incompatible to everything that
		54	isn't itself. What's worse, all the potential users of your module are
		55	I<also> forced to use the same event loop you use.
		56
		57	AnyEvent is different: AnyEvent + POE works fine. AnyEvent + Glib works
		58	fine. AnyEvent + Tk works fine etc. etc. but none of these work together
		59	with the rest: POE + IO::Async? No go. Tk + Event? No go. Again: if
		60	your module uses one of those, every user of your module has to use it,
		61	too. But if your module uses AnyEvent, it works transparently with all
		62	event models it supports (including stuff like POE and IO::Async, as long
		63	as those use one of the supported event loops. It is trivial to add new
		64	event loops to AnyEvent, too, so it is future-proof).
		65
		66	In addition to being free of having to use I<the one and only true event
		67	model>, AnyEvent also is free of bloat and policy: with POE or similar
		68	modules, you get an enormous amount of code and strict rules you have to
		69	follow. AnyEvent, on the other hand, is lean and up to the point, by only
		70	offering the functionality that is necessary, in as thin as a wrapper as
		71	technically possible.
		72
		73	Of course, AnyEvent comes with a big (and fully optional!) toolbox
		74	of useful functionality, such as an asynchronous DNS resolver, 100%
		75	non-blocking connects (even with TLS/SSL, IPv6 and on broken platforms
		76	such as Windows) and lots of real-world knowledge and workarounds for
		77	platform bugs and differences.
		78
		79	Now, if you I<do want> lots of policy (this can arguably be somewhat
		80	useful) and you want to force your users to use the one and only event
		81	model, you should I<not> use this module.
22		82
23	=head1 DESCRIPTION	83	=head1 DESCRIPTION
24		84
25	L<AnyEvent> provides an identical interface to multiple event loops. This	85	L<AnyEvent> provides an identical interface to multiple event loops. This
26	allows module authors to utilise an event loop without forcing module	86	allows module authors to utilise an event loop without forcing module
27	users to use the same event loop (as only a single event loop can coexist	87	users to use the same event loop (as only a single event loop can coexist
28	peacefully at any one time).	88	peacefully at any one time).
29		89
30	The interface itself is vaguely similar but not identical to the Event	90	The interface itself is vaguely similar, but not identical to the L<Event>
31	module.	91	module.
32		92
33	On the first call of any method, the module tries to detect the currently	93	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	94	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	95	following modules is already loaded: L<EV>,
36	used. If none is found, the module tries to load these modules in the	96	L<Event>, L<Glib>, L<AnyEvent::Impl::Perl>, L<Tk>, L<Event::Lib>, L<Qt>,
37	order given. The first one that could be successfully loaded will be	97	L<POE>. The first one found is used. If none are found, the module tries
38	used. If still none could be found, AnyEvent will fall back to a pure-perl	98	to load these modules (excluding Tk, Event::Lib, Qt and POE as the pure perl
39	event loop, which is also not very efficient.	99	adaptor should always succeed) in the order given. The first one that can
		100	be successfully loaded will be used. If, after this, still none could be
		101	found, AnyEvent will fall back to a pure-perl event loop, which is not
		102	very efficient, but should work everywhere.
40		103
41	Because AnyEvent first checks for modules that are already loaded, loading	104	Because AnyEvent first checks for modules that are already loaded, loading
42	an Event model explicitly before first using AnyEvent will likely make	105	an event model explicitly before first using AnyEvent will likely make
43	that model the default. For example:	106	that model the default. For example:
44		107
45	use Tk;	108	use Tk;
46	use AnyEvent;	109	use AnyEvent;
47		110
48	# .. AnyEvent will likely default to Tk	111	# .. AnyEvent will likely default to Tk
49		112
		113	The I<likely> means that, if any module loads another event model and
		114	starts using it, all bets are off. Maybe you should tell their authors to
		115	use AnyEvent so their modules work together with others seamlessly...
		116
50	The pure-perl implementation of AnyEvent is called	117	The pure-perl implementation of AnyEvent is called
51	C<AnyEvent::Impl::Perl>. Like other event modules you can load it	118	C<AnyEvent::Impl::Perl>. Like other event modules you can load it
52	explicitly.	119	explicitly and enjoy the high availability of that event loop :)
53		120
54	=head1 WATCHERS	121	=head1 WATCHERS
55		122
56	AnyEvent has the central concept of a I<watcher>, which is an object that	123	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	124	stores relevant data for each kind of event you are waiting for, such as
58	the callback to call, the filehandle to watch, etc.	125	the callback to call, the file handle to watch, etc.
59		126
60	These watchers are normal Perl objects with normal Perl lifetime. After	127	These watchers are normal Perl objects with normal Perl lifetime. After
61	creating a watcher it will immediately "watch" for events and invoke	128	creating a watcher it will immediately "watch" for events and invoke the
		129	callback when the event occurs (of course, only when the event model
		130	is in control).
		131
62	the callback. To disable the watcher you have to destroy it (e.g. by	132	To disable the watcher you have to destroy it (e.g. by setting the
63	setting the variable that stores it to C<undef> or otherwise deleting all	133	variable you store it in to C<undef> or otherwise deleting all references
64	references to it).	134	to it).
65		135
66	All watchers are created by calling a method on the C<AnyEvent> class.	136	All watchers are created by calling a method on the C<AnyEvent> class.
67		137
		138	Many watchers either are used with "recursion" (repeating timers for
		139	example), or need to refer to their watcher object in other ways.
		140
		141	An any way to achieve that is this pattern:
		142
		143	my $w; $w = AnyEvent->type (arg => value ..., cb => sub {
		144	# you can use $w here, for example to undef it
		145	undef $w;
		146	});
		147
		148	Note that C<my $w; $w => combination. This is necessary because in Perl,
		149	my variables are only visible after the statement in which they are
		150	declared.
		151
68	=head2 IO WATCHERS	152	=head2 I/O WATCHERS
69		153
70	You can create I/O watcher by calling the C<< AnyEvent->io >> method with	154	You can create an I/O watcher by calling the C<< AnyEvent->io >> method
71	the following mandatory arguments:	155	with the following mandatory key-value pairs as arguments:
72		156
73	C<fh> the Perl I<filehandle> (not filedescriptor) to watch for	157	C<fh> the Perl I<file handle> (I<not> file descriptor) to watch
74	events. C<poll> must be a string that is either C<r> or C<w>, that creates	158	for events. C<poll> must be a string that is either C<r> or C<w>,
75	a watcher waiting for "r"eadable or "w"ritable events. C<cb> teh callback	159	which creates a watcher waiting for "r"eadable or "w"ritable events,
76	to invoke everytime the filehandle becomes ready.	160	respectively. C<cb> is the callback to invoke each time the file handle
		161	becomes ready.
77		162
78	Only one io watcher per C<fh> and C<poll> combination is allowed (i.e. on	163	Although the callback might get passed parameters, their value and
79	a socket you can have one r + one w, not any more (limitation comes from	164	presence is undefined and you cannot rely on them. Portable AnyEvent
80	Tk - if you are sure you are not using Tk this limitation is gone).	165	callbacks cannot use arguments passed to I/O watcher callbacks.
81		166
82	Filehandles will be kept alive, so as long as the watcher exists, the	167	The I/O watcher might use the underlying file descriptor or a copy of it.
83	filehandle exists, too.	168	You must not close a file handle as long as any watcher is active on the
		169	underlying file descriptor.
		170
		171	Some event loops issue spurious readyness notifications, so you should
		172	always use non-blocking calls when reading/writing from/to your file
		173	handles.
84		174
85	Example:	175	Example:
86		176
87	# wait for readability of STDIN, then read a line and disable the watcher	177	# wait for readability of STDIN, then read a line and disable the watcher
88	my $w; $w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {	178	my $w; $w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {
…		…
94	=head2 TIME WATCHERS	184	=head2 TIME WATCHERS
95		185
96	You can create a time watcher by calling the C<< AnyEvent->timer >>	186	You can create a time watcher by calling the C<< AnyEvent->timer >>
97	method with the following mandatory arguments:	187	method with the following mandatory arguments:
98		188
99	C<after> after how many seconds (fractions are supported) should the timer	189	C<after> specifies after how many seconds (fractional values are
100	activate. C<cb> the callback to invoke.	190	supported) the callback should be invoked. C<cb> is the callback to invoke
		191	in that case.
		192
		193	Although the callback might get passed parameters, their value and
		194	presence is undefined and you cannot rely on them. Portable AnyEvent
		195	callbacks cannot use arguments passed to time watcher callbacks.
101		196
102	The timer callback will be invoked at most once: if you want a repeating	197	The timer callback will be invoked at most once: if you want a repeating
103	timer you have to create a new watcher (this is a limitation by both Tk	198	timer you have to create a new watcher (this is a limitation by both Tk
104	and Glib).	199	and Glib).
105		200
…		…
109	my $w = AnyEvent->timer (after => 7.7, cb => sub {	204	my $w = AnyEvent->timer (after => 7.7, cb => sub {
110	warn "timeout\n";	205	warn "timeout\n";
111	});	206	});
112		207
113	# to cancel the timer:	208	# to cancel the timer:
114	undef $w	209	undef $w;
115		210
116	=head2 CONDITION WATCHERS	211	Example 2:
117		212
118	Condition watchers can be created by calling the C<< AnyEvent->condvar >>	213	# fire an event after 0.5 seconds, then roughly every second
119	method without any arguments.	214	my $w;
120		215
121	A condition watcher watches for a condition - precisely that the C<<	216	my $cb = sub {
122	->broadcast >> method has been called.	217	# cancel the old timer while creating a new one
		218	$w = AnyEvent->timer (after => 1, cb => $cb);
		219	};
123		220
124	The watcher has only two methods:	221	# start the "loop" by creating the first watcher
		222	$w = AnyEvent->timer (after => 0.5, cb => $cb);
		223
		224	=head3 TIMING ISSUES
		225
		226	There are two ways to handle timers: based on real time (relative, "fire
		227	in 10 seconds") and based on wallclock time (absolute, "fire at 12
		228	o'clock").
		229
		230	While most event loops expect timers to specified in a relative way, they
		231	use absolute time internally. This makes a difference when your clock
		232	"jumps", for example, when ntp decides to set your clock backwards from
		233	the wrong date of 2014-01-01 to 2008-01-01, a watcher that is supposed to
		234	fire "after" a second might actually take six years to finally fire.
		235
		236	AnyEvent cannot compensate for this. The only event loop that is conscious
		237	about these issues is L<EV>, which offers both relative (ev_timer, based
		238	on true relative time) and absolute (ev_periodic, based on wallclock time)
		239	timers.
		240
		241	AnyEvent always prefers relative timers, if available, matching the
		242	AnyEvent API.
		243
		244	AnyEvent has two additional methods that return the "current time":
125		245
126	=over 4	246	=over 4
127		247
128	=item $cv->wait	248	=item AnyEvent->time
129		249
130	Wait (blocking if necessary) until the C<< ->broadcast >> method has been	250	This returns the "current wallclock time" as a fractional number of
131	called on c<$cv>, while servicing other watchers normally.	251	seconds since the Epoch (the same thing as C<time> or C<Time::HiRes::time>
		252	return, and the result is guaranteed to be compatible with those).
132		253
133	Not all event models support a blocking wait - some die in that case, so	254	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	255	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		256
140	You can only wait once on a condition - additional calls will return	257	=item AnyEvent->now
141	immediately.
142		258
143	=item $cv->broadcast	259	This also returns the "current wallclock time", but unlike C<time>, above,
		260	this value might change only once per event loop iteration, depending on
		261	the event loop (most return the same time as C<time>, above). This is the
		262	time that AnyEvent's timers get scheduled against.
144		263
145	Flag the condition as ready - a running C<< ->wait >> and all further	264	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	265	function to call when you want to know the current time.>
147	is waiting the broadcast will be remembered..
148		266
149	Example:	267	This function is also often faster then C<< AnyEvent->time >>, and
		268	thus the preferred method if you want some timestamp (for example,
		269	L<AnyEvent::Handle> uses this to update it's activity timeouts).
		270
		271	The rest of this section is only of relevance if you try to be very exact
		272	with your timing, you can skip it without bad conscience.
		273
		274	For a practical example of when these times differ, consider L<Event::Lib>
		275	and L<EV> and the following set-up:
		276
		277	The event loop is running and has just invoked one of your callback at
		278	time=500 (assume no other callbacks delay processing). In your callback,
		279	you wait a second by executing C<sleep 1> (blocking the process for a
		280	second) and then (at time=501) you create a relative timer that fires
		281	after three seconds.
		282
		283	With L<Event::Lib>, C<< AnyEvent->time >> and C<< AnyEvent->now >> will
		284	both return C<501>, because that is the current time, and the timer will
		285	be scheduled to fire at time=504 (C<501> + C<3>).
		286
		287	With L<EV>, C<< AnyEvent->time >> returns C<501> (as that is the current
		288	time), but C<< AnyEvent->now >> returns C<500>, as that is the time the
		289	last event processing phase started. With L<EV>, your timer gets scheduled
		290	to run at time=503 (C<500> + C<3>).
		291
		292	In one sense, L<Event::Lib> is more exact, as it uses the current time
		293	regardless of any delays introduced by event processing. However, most
		294	callbacks do not expect large delays in processing, so this causes a
		295	higher drift (and a lot more system calls to get the current time).
		296
		297	In another sense, L<EV> is more exact, as your timer will be scheduled at
		298	the same time, regardless of how long event processing actually took.
		299
		300	In either case, if you care (and in most cases, you don't), then you
		301	can get whatever behaviour you want with any event loop, by taking the
		302	difference between C<< AnyEvent->time >> and C<< AnyEvent->now >> into
		303	account.
		304
		305	=back
		306
		307	=head2 SIGNAL WATCHERS
		308
		309	You can watch for signals using a signal watcher, C<signal> is the signal
		310	I<name> without any C<SIG> prefix, C<cb> is the Perl callback to
		311	be invoked whenever a signal occurs.
		312
		313	Although the callback might get passed parameters, their value and
		314	presence is undefined and you cannot rely on them. Portable AnyEvent
		315	callbacks cannot use arguments passed to signal watcher callbacks.
		316
		317	Multiple signal occurrences can be clumped together into one callback
		318	invocation, and callback invocation will be synchronous. Synchronous means
		319	that it might take a while until the signal gets handled by the process,
		320	but it is guaranteed not to interrupt any other callbacks.
		321
		322	The main advantage of using these watchers is that you can share a signal
		323	between multiple watchers.
		324
		325	This watcher might use C<%SIG>, so programs overwriting those signals
		326	directly will likely not work correctly.
		327
		328	Example: exit on SIGINT
		329
		330	my $w = AnyEvent->signal (signal => "INT", cb => sub { exit 1 });
		331
		332	=head2 CHILD PROCESS WATCHERS
		333
		334	You can also watch on a child process exit and catch its exit status.
		335
		336	The child process is specified by the C<pid> argument (if set to C<0>, it
		337	watches for any child process exit). The watcher will trigger as often
		338	as status change for the child are received. This works by installing a
		339	signal handler for C<SIGCHLD>. The callback will be called with the pid
		340	and exit status (as returned by waitpid), so unlike other watcher types,
		341	you I<can> rely on child watcher callback arguments.
		342
		343	There is a slight catch to child watchers, however: you usually start them
		344	I<after> the child process was created, and this means the process could
		345	have exited already (and no SIGCHLD will be sent anymore).
		346
		347	Not all event models handle this correctly (POE doesn't), but even for
		348	event models that I<do> handle this correctly, they usually need to be
		349	loaded before the process exits (i.e. before you fork in the first place).
		350
		351	This means you cannot create a child watcher as the very first thing in an
		352	AnyEvent program, you I<have> to create at least one watcher before you
		353	C<fork> the child (alternatively, you can call C<AnyEvent::detect>).
		354
		355	Example: fork a process and wait for it
		356
		357	my $done = AnyEvent->condvar;
		358
		359	my $pid = fork or exit 5;
		360
		361	my $w = AnyEvent->child (
		362	pid => $pid,
		363	cb => sub {
		364	my ($pid, $status) = @_;
		365	warn "pid $pid exited with status $status";
		366	$done->send;
		367	},
		368	);
		369
		370	# do something else, then wait for process exit
		371	$done->recv;
		372
		373	=head2 CONDITION VARIABLES
		374
		375	If you are familiar with some event loops you will know that all of them
		376	require you to run some blocking "loop", "run" or similar function that
		377	will actively watch for new events and call your callbacks.
		378
		379	AnyEvent is different, it expects somebody else to run the event loop and
		380	will only block when necessary (usually when told by the user).
		381
		382	The instrument to do that is called a "condition variable", so called
		383	because they represent a condition that must become true.
		384
		385	Condition variables can be created by calling the C<< AnyEvent->condvar
		386	>> method, usually without arguments. The only argument pair allowed is
		387	C<cb>, which specifies a callback to be called when the condition variable
		388	becomes true.
		389
		390	After creation, the condition variable is "false" until it becomes "true"
		391	by calling the C<send> method (or calling the condition variable as if it
		392	were a callback, read about the caveats in the description for the C<<
		393	->send >> method).
		394
		395	Condition variables are similar to callbacks, except that you can
		396	optionally wait for them. They can also be called merge points - points
		397	in time where multiple outstanding events have been processed. And yet
		398	another way to call them is transactions - each condition variable can be
		399	used to represent a transaction, which finishes at some point and delivers
		400	a result.
		401
		402	Condition variables are very useful to signal that something has finished,
		403	for example, if you write a module that does asynchronous http requests,
		404	then a condition variable would be the ideal candidate to signal the
		405	availability of results. The user can either act when the callback is
		406	called or can synchronously C<< ->recv >> for the results.
		407
		408	You can also use them to simulate traditional event loops - for example,
		409	you can block your main program until an event occurs - for example, you
		410	could C<< ->recv >> in your main program until the user clicks the Quit
		411	button of your app, which would C<< ->send >> the "quit" event.
		412
		413	Note that condition variables recurse into the event loop - if you have
		414	two pieces of code that call C<< ->recv >> in a round-robin fashion, you
		415	lose. Therefore, condition variables are good to export to your caller, but
		416	you should avoid making a blocking wait yourself, at least in callbacks,
		417	as this asks for trouble.
		418
		419	Condition variables are represented by hash refs in perl, and the keys
		420	used by AnyEvent itself are all named C<_ae_XXX> to make subclassing
		421	easy (it is often useful to build your own transaction class on top of
		422	AnyEvent). To subclass, use C<AnyEvent::CondVar> as base class and call
		423	it's C<new> method in your own C<new> method.
		424
		425	There are two "sides" to a condition variable - the "producer side" which
		426	eventually calls C<< -> send >>, and the "consumer side", which waits
		427	for the send to occur.
		428
		429	Example: wait for a timer.
150		430
151	# wait till the result is ready	431	# wait till the result is ready
152	my $result_ready = AnyEvent->condvar;	432	my $result_ready = AnyEvent->condvar;
153		433
154	# do something such as adding a timer	434	# do something such as adding a timer
155	# or socket watcher the calls $result_ready->broadcast	435	# or socket watcher the calls $result_ready->send
156	# when the "result" is ready.	436	# when the "result" is ready.
		437	# in this case, we simply use a timer:
		438	my $w = AnyEvent->timer (
		439	after => 1,
		440	cb => sub { $result_ready->send },
		441	);
157		442
		443	# this "blocks" (while handling events) till the callback
		444	# calls send
158	$result_ready->wait;	445	$result_ready->recv;
		446
		447	Example: wait for a timer, but take advantage of the fact that
		448	condition variables are also code references.
		449
		450	my $done = AnyEvent->condvar;
		451	my $delay = AnyEvent->timer (after => 5, cb => $done);
		452	$done->recv;
		453
		454	=head3 METHODS FOR PRODUCERS
		455
		456	These methods should only be used by the producing side, i.e. the
		457	code/module that eventually sends the signal. Note that it is also
		458	the producer side which creates the condvar in most cases, but it isn't
		459	uncommon for the consumer to create it as well.
		460
		461	=over 4
		462
		463	=item $cv->send (...)
		464
		465	Flag the condition as ready - a running C<< ->recv >> and all further
		466	calls to C<recv> will (eventually) return after this method has been
		467	called. If nobody is waiting the send will be remembered.
		468
		469	If a callback has been set on the condition variable, it is called
		470	immediately from within send.
		471
		472	Any arguments passed to the C<send> call will be returned by all
		473	future C<< ->recv >> calls.
		474
		475	Condition variables are overloaded so one can call them directly
		476	(as a code reference). Calling them directly is the same as calling
		477	C<send>. Note, however, that many C-based event loops do not handle
		478	overloading, so as tempting as it may be, passing a condition variable
		479	instead of a callback does not work. Both the pure perl and EV loops
		480	support overloading, however, as well as all functions that use perl to
		481	invoke a callback (as in L<AnyEvent::Socket> and L<AnyEvent::DNS> for
		482	example).
		483
		484	=item $cv->croak ($error)
		485
		486	Similar to send, but causes all call's to C<< ->recv >> to invoke
		487	C<Carp::croak> with the given error message/object/scalar.
		488
		489	This can be used to signal any errors to the condition variable
		490	user/consumer.
		491
		492	=item $cv->begin ([group callback])
		493
		494	=item $cv->end
		495
		496	These two methods are EXPERIMENTAL and MIGHT CHANGE.
		497
		498	These two methods can be used to combine many transactions/events into
		499	one. For example, a function that pings many hosts in parallel might want
		500	to use a condition variable for the whole process.
		501
		502	Every call to C<< ->begin >> will increment a counter, and every call to
		503	C<< ->end >> will decrement it. If the counter reaches C<0> in C<< ->end
		504	>>, the (last) callback passed to C<begin> will be executed. That callback
		505	is I<supposed> to call C<< ->send >>, but that is not required. If no
		506	callback was set, C<send> will be called without any arguments.
		507
		508	Let's clarify this with the ping example:
		509
		510	my $cv = AnyEvent->condvar;
		511
		512	my %result;
		513	$cv->begin (sub { $cv->send (\%result) });
		514
		515	for my $host (@list_of_hosts) {
		516	$cv->begin;
		517	ping_host_then_call_callback $host, sub {
		518	$result{$host} = ...;
		519	$cv->end;
		520	};
		521	}
		522
		523	$cv->end;
		524
		525	This code fragment supposedly pings a number of hosts and calls
		526	C<send> after results for all then have have been gathered - in any
		527	order. To achieve this, the code issues a call to C<begin> when it starts
		528	each ping request and calls C<end> when it has received some result for
		529	it. Since C<begin> and C<end> only maintain a counter, the order in which
		530	results arrive is not relevant.
		531
		532	There is an additional bracketing call to C<begin> and C<end> outside the
		533	loop, which serves two important purposes: first, it sets the callback
		534	to be called once the counter reaches C<0>, and second, it ensures that
		535	C<send> is called even when C<no> hosts are being pinged (the loop
		536	doesn't execute once).
		537
		538	This is the general pattern when you "fan out" into multiple subrequests:
		539	use an outer C<begin>/C<end> pair to set the callback and ensure C<end>
		540	is called at least once, and then, for each subrequest you start, call
		541	C<begin> and for each subrequest you finish, call C<end>.
159		542
160	=back	543	=back
161		544
162	=head2 SIGNAL WATCHERS	545	=head3 METHODS FOR CONSUMERS
163		546
164	You can listen for signals using a signal watcher, C<signal> is the signal	547	These methods should only be used by the consuming side, i.e. the
165	I<name> without any C<SIG> prefix. Multiple signals events can be clumped	548	code awaits the condition.
166	together into one callback invocation, and callback invocation might or
167	might not be asynchronous.
168		549
169	These watchers might use C<%SIG>, so programs overwriting those signals	550	=over 4
170	directly will likely not work correctly.
171		551
172	Example: exit on SIGINT	552	=item $cv->recv
173		553
174	my $w = AnyEvent->signal (signal => "INT", cb => sub { exit 1 });	554	Wait (blocking if necessary) until the C<< ->send >> or C<< ->croak
		555	>> methods have been called on c<$cv>, while servicing other watchers
		556	normally.
175		557
176	=head2 CHILD PROCESS WATCHERS	558	You can only wait once on a condition - additional calls are valid but
		559	will return immediately.
177		560
178	You can also listen for the status of a child process specified by the	561	If an error condition has been set by calling C<< ->croak >>, then this
179	C<pid> argument. The watcher will only trigger once. This works by	562	function will call C<croak>.
180	installing a signal handler for C<SIGCHLD>.
181		563
182	Example: wait for pid 1333	564	In list context, all parameters passed to C<send> will be returned,
		565	in scalar context only the first one will be returned.
183		566
184	my $w = AnyEvent->child (pid => 1333, cb => sub { warn "exit status $?" });	567	Not all event models support a blocking wait - some die in that case
		568	(programs might want to do that to stay interactive), so I<if you are
		569	using this from a module, never require a blocking wait>, but let the
		570	caller decide whether the call will block or not (for example, by coupling
		571	condition variables with some kind of request results and supporting
		572	callbacks so the caller knows that getting the result will not block,
		573	while still supporting blocking waits if the caller so desires).
185		574
186	=head1 GLOBALS	575	Another reason I<never> to C<< ->recv >> in a module is that you cannot
		576	sensibly have two C<< ->recv >>'s in parallel, as that would require
		577	multiple interpreters or coroutines/threads, none of which C<AnyEvent>
		578	can supply.
		579
		580	The L<Coro> module, however, I<can> and I<does> supply coroutines and, in
		581	fact, L<Coro::AnyEvent> replaces AnyEvent's condvars by coroutine-safe
		582	versions and also integrates coroutines into AnyEvent, making blocking
		583	C<< ->recv >> calls perfectly safe as long as they are done from another
		584	coroutine (one that doesn't run the event loop).
		585
		586	You can ensure that C<< -recv >> never blocks by setting a callback and
		587	only calling C<< ->recv >> from within that callback (or at a later
		588	time). This will work even when the event loop does not support blocking
		589	waits otherwise.
		590
		591	=item $bool = $cv->ready
		592
		593	Returns true when the condition is "true", i.e. whether C<send> or
		594	C<croak> have been called.
		595
		596	=item $cb = $cv->cb ([new callback])
		597
		598	This is a mutator function that returns the callback set and optionally
		599	replaces it before doing so.
		600
		601	The callback will be called when the condition becomes "true", i.e. when
		602	C<send> or C<croak> are called. Calling C<recv> inside the callback
		603	or at any later time is guaranteed not to block.
		604
		605	=back
		606
		607	=head1 GLOBAL VARIABLES AND FUNCTIONS
187		608
188	=over 4	609	=over 4
189		610
190	=item $AnyEvent::MODEL	611	=item $AnyEvent::MODEL
191		612
…		…
195	C<AnyEvent::Impl:xxx> modules, but can be any other class in the case	616	C<AnyEvent::Impl:xxx> modules, but can be any other class in the case
196	AnyEvent has been extended at runtime (e.g. in I<rxvt-unicode>).	617	AnyEvent has been extended at runtime (e.g. in I<rxvt-unicode>).
197		618
198	The known classes so far are:	619	The known classes so far are:
199		620
200	AnyEvent::Impl::Coro based on Coro::Event, best choice.
201	EV::AnyEvent based on EV (an interface to libevent)	621	AnyEvent::Impl::EV based on EV (an interface to libev, best choice).
202	AnyEvent::Impl::Event based on Event, also best choice :)	622	AnyEvent::Impl::Event based on Event, second best choice.
		623	AnyEvent::Impl::Perl pure-perl implementation, fast and portable.
203	AnyEvent::Impl::Glib based on Glib, second-best choice.	624	AnyEvent::Impl::Glib based on Glib, third-best choice.
204	AnyEvent::Impl::Tk based on Tk, very bad choice.	625	AnyEvent::Impl::Tk based on Tk, very bad choice.
205	AnyEvent::Impl::Perl pure-perl implementation, inefficient.	626	AnyEvent::Impl::Qt based on Qt, cannot be autoprobed (see its docs).
		627	AnyEvent::Impl::EventLib based on Event::Lib, leaks memory and worse.
		628	AnyEvent::Impl::POE based on POE, not generic enough for full support.
		629
		630	There is no support for WxWidgets, as WxWidgets has no support for
		631	watching file handles. However, you can use WxWidgets through the
		632	POE Adaptor, as POE has a Wx backend that simply polls 20 times per
		633	second, which was considered to be too horrible to even consider for
		634	AnyEvent. Likewise, other POE backends can be used by AnyEvent by using
		635	it's adaptor.
		636
		637	AnyEvent knows about L<Prima> and L<Wx> and will try to use L<POE> when
		638	autodetecting them.
206		639
207	=item AnyEvent::detect	640	=item AnyEvent::detect
208		641
209	Returns C<$AnyEvent::MODEL>, forcing autodetection of the event model if	642	Returns C<$AnyEvent::MODEL>, forcing autodetection of the event model
210	necessary. You should only call this function right before you would have	643	if necessary. You should only call this function right before you would
211	created an AnyEvent watcher anyway, that is, very late at runtime.	644	have created an AnyEvent watcher anyway, that is, as late as possible at
		645	runtime.
		646
		647	=item $guard = AnyEvent::post_detect { BLOCK }
		648
		649	Arranges for the code block to be executed as soon as the event model is
		650	autodetected (or immediately if this has already happened).
		651
		652	If called in scalar or list context, then it creates and returns an object
		653	that automatically removes the callback again when it is destroyed. See
		654	L<Coro::BDB> for a case where this is useful.
		655
		656	=item @AnyEvent::post_detect
		657
		658	If there are any code references in this array (you can C<push> to it
		659	before or after loading AnyEvent), then they will called directly after
		660	the event loop has been chosen.
		661
		662	You should check C<$AnyEvent::MODEL> before adding to this array, though:
		663	if it contains a true value then the event loop has already been detected,
		664	and the array will be ignored.
		665
		666	Best use C<AnyEvent::post_detect { BLOCK }> instead.
212		667
213	=back	668	=back
214		669
215	=head1 WHAT TO DO IN A MODULE	670	=head1 WHAT TO DO IN A MODULE
216		671
217	As a module author, you should "use AnyEvent" and call AnyEvent methods	672	As a module author, you should C<use AnyEvent> and call AnyEvent methods
218	freely, but you should not load a specific event module or rely on it.	673	freely, but you should not load a specific event module or rely on it.
219		674
220	Be careful when you create watchers in the module body - Anyevent will	675	Be careful when you create watchers in the module body - AnyEvent will
221	decide which event module to use as soon as the first method is called, so	676	decide which event module to use as soon as the first method is called, so
222	by calling AnyEvent in your module body you force the user of your module	677	by calling AnyEvent in your module body you force the user of your module
223	to load the event module first.	678	to load the event module first.
224		679
		680	Never call C<< ->recv >> on a condition variable unless you I<know> that
		681	the C<< ->send >> method has been called on it already. This is
		682	because it will stall the whole program, and the whole point of using
		683	events is to stay interactive.
		684
		685	It is fine, however, to call C<< ->recv >> when the user of your module
		686	requests it (i.e. if you create a http request object ad have a method
		687	called C<results> that returns the results, it should call C<< ->recv >>
		688	freely, as the user of your module knows what she is doing. always).
		689
225	=head1 WHAT TO DO IN THE MAIN PROGRAM	690	=head1 WHAT TO DO IN THE MAIN PROGRAM
226		691
227	There will always be a single main program - the only place that should	692	There will always be a single main program - the only place that should
228	dictate which event model to use.	693	dictate which event model to use.
229		694
230	If it doesn't care, it can just "use AnyEvent" and use it itself, or not	695	If it doesn't care, it can just "use AnyEvent" and use it itself, or not
231	do anything special and let AnyEvent decide which implementation to chose.	696	do anything special (it does not need to be event-based) and let AnyEvent
		697	decide which implementation to chose if some module relies on it.
232		698
233	If the main program relies on a specific event model (for example, in Gtk2	699	If the main program relies on a specific event model - for example, in
234	programs you have to rely on either Glib or Glib::Event), you should load	700	Gtk2 programs you have to rely on the Glib module - you should load the
235	it before loading AnyEvent or any module that uses it, generally, as early	701	event module before loading AnyEvent or any module that uses it: generally
236	as possible. The reason is that modules might create watchers when they	702	speaking, you should load it as early as possible. The reason is that
237	are loaded, and AnyEvent will decide on the event model to use as soon as	703	modules might create watchers when they are loaded, and AnyEvent will
238	it creates watchers, and it might chose the wrong one unless you load the	704	decide on the event model to use as soon as it creates watchers, and it
239	correct one yourself.	705	might chose the wrong one unless you load the correct one yourself.
240		706
241	You can chose to use a rather inefficient pure-perl implementation by	707	You can chose to use a pure-perl implementation by loading the
242	loading the C<AnyEvent::Impl::Perl> module, but letting AnyEvent chose is	708	C<AnyEvent::Impl::Perl> module, which gives you similar behaviour
243	generally better.	709	everywhere, but letting AnyEvent chose the model is generally better.
		710
		711	=head2 MAINLOOP EMULATION
		712
		713	Sometimes (often for short test scripts, or even standalone programs who
		714	only want to use AnyEvent), you do not want to run a specific event loop.
		715
		716	In that case, you can use a condition variable like this:
		717
		718	AnyEvent->condvar->recv;
		719
		720	This has the effect of entering the event loop and looping forever.
		721
		722	Note that usually your program has some exit condition, in which case
		723	it is better to use the "traditional" approach of storing a condition
		724	variable somewhere, waiting for it, and sending it when the program should
		725	exit cleanly.
		726
		727
		728	=head1 OTHER MODULES
		729
		730	The following is a non-exhaustive list of additional modules that use
		731	AnyEvent and can therefore be mixed easily with other AnyEvent modules
		732	in the same program. Some of the modules come with AnyEvent, some are
		733	available via CPAN.
		734
		735	=over 4
		736
		737	=item L<AnyEvent::Util>
		738
		739	Contains various utility functions that replace often-used but blocking
		740	functions such as C<inet_aton> by event-/callback-based versions.
		741
		742	=item L<AnyEvent::Handle>
		743
		744	Provide read and write buffers and manages watchers for reads and writes.
		745
		746	=item L<AnyEvent::Socket>
		747
		748	Provides various utility functions for (internet protocol) sockets,
		749	addresses and name resolution. Also functions to create non-blocking tcp
		750	connections or tcp servers, with IPv6 and SRV record support and more.
		751
		752	=item L<AnyEvent::DNS>
		753
		754	Provides rich asynchronous DNS resolver capabilities.
		755
		756	=item L<AnyEvent::HTTPD>
		757
		758	Provides a simple web application server framework.
		759
		760	=item L<AnyEvent::FastPing>
		761
		762	The fastest ping in the west.
		763
		764	=item L<Net::IRC3>
		765
		766	AnyEvent based IRC client module family.
		767
		768	=item L<Net::XMPP2>
		769
		770	AnyEvent based XMPP (Jabber protocol) module family.
		771
		772	=item L<Net::FCP>
		773
		774	AnyEvent-based implementation of the Freenet Client Protocol, birthplace
		775	of AnyEvent.
		776
		777	=item L<Event::ExecFlow>
		778
		779	High level API for event-based execution flow control.
		780
		781	=item L<Coro>
		782
		783	Has special support for AnyEvent via L<Coro::AnyEvent>.
		784
		785	=item L<AnyEvent::AIO>, L<IO::AIO>
		786
		787	Truly asynchronous I/O, should be in the toolbox of every event
		788	programmer. AnyEvent::AIO transparently fuses IO::AIO and AnyEvent
		789	together.
		790
		791	=item L<AnyEvent::BDB>, L<BDB>
		792
		793	Truly asynchronous Berkeley DB access. AnyEvent::AIO transparently fuses
		794	IO::AIO and AnyEvent together.
		795
		796	=item L<IO::Lambda>
		797
		798	The lambda approach to I/O - don't ask, look there. Can use AnyEvent.
		799
		800	=back
244		801
245	=cut	802	=cut
246		803
247	package AnyEvent;	804	package AnyEvent;
248		805
249	no warnings;	806	no warnings;
250	use strict;	807	use strict;
251		808
252	use Carp;	809	use Carp;
253		810
254	our $VERSION = '2.55';	811	our $VERSION = 4.11;
255	our $MODEL;	812	our $MODEL;
256		813
257	our $AUTOLOAD;	814	our $AUTOLOAD;
258	our @ISA;	815	our @ISA;
259		816
		817	our @REGISTRY;
		818
		819	our $WIN32;
		820
		821	BEGIN {
		822	my $win32 = ! ! ($^O =~ /mswin32/i);
		823	eval "sub WIN32(){ $win32 }";
		824	}
		825
260	our $verbose = $ENV{PERL_ANYEVENT_VERBOSE}*1;	826	our $verbose = $ENV{PERL_ANYEVENT_VERBOSE}*1;
261		827
262	our @REGISTRY;	828	our %PROTOCOL; # (ipv4|ipv6) => (1|2), higher numbers are preferred
		829
		830	{
		831	my $idx;
		832	$PROTOCOL{$_} = ++$idx
		833	for reverse split /\s*,\s*/,
		834	$ENV{PERL_ANYEVENT_PROTOCOLS} || "ipv4,ipv6";
		835	}
263		836
264	my @models = (	837	my @models = (
265	[Coro::Event:: => AnyEvent::Impl::Coro::],
266	[EV:: => EV::AnyEvent::],	838	[EV:: => AnyEvent::Impl::EV::],
267	[Event:: => AnyEvent::Impl::Event::],	839	[Event:: => AnyEvent::Impl::Event::],
268	[Glib:: => AnyEvent::Impl::Glib::],
269	[Tk:: => AnyEvent::Impl::Tk::],
270	[AnyEvent::Impl::Perl:: => AnyEvent::Impl::Perl::],	840	[AnyEvent::Impl::Perl:: => AnyEvent::Impl::Perl::],
		841	# everything below here will not be autoprobed
		842	# as the pureperl backend should work everywhere
		843	# and is usually faster
		844	[Tk:: => AnyEvent::Impl::Tk::], # crashes with many handles
		845	[Glib:: => AnyEvent::Impl::Glib::], # becomes extremely slow with many watchers
		846	[Event::Lib:: => AnyEvent::Impl::EventLib::], # too buggy
		847	[Qt:: => AnyEvent::Impl::Qt::], # requires special main program
		848	[POE::Kernel:: => AnyEvent::Impl::POE::], # lasciate ogni speranza
		849	[Wx:: => AnyEvent::Impl::POE::],
		850	[Prima:: => AnyEvent::Impl::POE::],
271	);	851	);
272		852
273	our %method = map +($_ => 1), qw(io timer condvar broadcast wait signal one_event DESTROY);	853	our %method = map +($_ => 1), qw(io timer time now signal child condvar one_event DESTROY);
		854
		855	our @post_detect;
		856
		857	sub post_detect(&) {
		858	my ($cb) = @_;
		859
		860	if ($MODEL) {
		861	$cb->();
		862
		863	1
		864	} else {
		865	push @post_detect, $cb;
		866
		867	defined wantarray
		868	? bless \$cb, "AnyEvent::Util::PostDetect"
		869	: ()
		870	}
		871	}
		872
		873	sub AnyEvent::Util::PostDetect::DESTROY {
		874	@post_detect = grep $_ != ${$_[0]}, @post_detect;
		875	}
274		876
275	sub detect() {	877	sub detect() {
276	unless ($MODEL) {	878	unless ($MODEL) {
277	no strict 'refs';	879	no strict 'refs';
		880	local $SIG{__DIE__};
		881
		882	if ($ENV{PERL_ANYEVENT_MODEL} =~ /^([a-zA-Z]+)$/) {
		883	my $model = "AnyEvent::Impl::$1";
		884	if (eval "require $model") {
		885	$MODEL = $model;
		886	warn "AnyEvent: loaded model '$model' (forced by \$PERL_ANYEVENT_MODEL), using it.\n" if $verbose > 1;
		887	} else {
		888	warn "AnyEvent: unable to load model '$model' (from \$PERL_ANYEVENT_MODEL):\n$@" if $verbose;
		889	}
		890	}
278		891
279	# check for already loaded models	892	# check for already loaded models
		893	unless ($MODEL) {
280	for (@REGISTRY, @models) {	894	for (@REGISTRY, @models) {
281	my ($package, $model) = @$_;	895	my ($package, $model) = @$_;
282	if (${"$package\::VERSION"} > 0) {	896	if (${"$package\::VERSION"} > 0) {
283	if (eval "require $model") {	897	if (eval "require $model") {
284	$MODEL = $model;	898	$MODEL = $model;
285	warn "AnyEvent: found model '$model', using it.\n" if $verbose > 1;	899	warn "AnyEvent: autodetected model '$model', using it.\n" if $verbose > 1;
286	last;	900	last;
		901	}
287	}	902	}
288	}	903	}
289	}
290		904
291	unless ($MODEL) {	905	unless ($MODEL) {
292	# try to load a model	906	# try to load a model
293		907
294	for (@REGISTRY, @models) {	908	for (@REGISTRY, @models) {
295	my ($package, $model) = @$_;	909	my ($package, $model) = @$_;
296	if (eval "require $package"	910	if (eval "require $package"
297	and ${"$package\::VERSION"} > 0	911	and ${"$package\::VERSION"} > 0
298	and eval "require $model") {	912	and eval "require $model") {
299	$MODEL = $model;	913	$MODEL = $model;
300	warn "AnyEvent: autoprobed and loaded model '$model', using it.\n" if $verbose > 1;	914	warn "AnyEvent: autoprobed model '$model', using it.\n" if $verbose > 1;
301	last;	915	last;
		916	}
302	}	917	}
		918
		919	$MODEL
		920	or die "No event module selected for AnyEvent and autodetect failed. Install any one of these modules: EV, Event or Glib.";
303	}	921	}
304
305	$MODEL
306	or die "No event module selected for AnyEvent and autodetect failed. Install any one of these modules: Event (or Coro+Event), Glib or Tk.";
307	}	922	}
308		923
309	unshift @ISA, $MODEL;	924	unshift @ISA, $MODEL;
310	push @{"$MODEL\::ISA"}, "AnyEvent::Base";	925	push @{"$MODEL\::ISA"}, "AnyEvent::Base";
		926
		927	(shift @post_detect)->() while @post_detect;
311	}	928	}
312		929
313	$MODEL	930	$MODEL
314	}	931	}
315		932
…		…
325	$class->$func (@_);	942	$class->$func (@_);
326	}	943	}
327		944
328	package AnyEvent::Base;	945	package AnyEvent::Base;
329		946
		947	# default implementation for now and time
		948
		949	use Time::HiRes ();
		950
		951	sub time { Time::HiRes::time }
		952	sub now { Time::HiRes::time }
		953
330	# default implementation for ->condvar, ->wait, ->broadcast	954	# default implementation for ->condvar
331		955
332	sub condvar {	956	sub condvar {
333	bless \my $flag, "AnyEvent::Base::CondVar"	957	bless { @_ == 3 ? (_ae_cb => $_[2]) : () }, AnyEvent::CondVar::
334	}
335
336	sub AnyEvent::Base::CondVar::broadcast {
337	${$_[0]}++;
338	}
339
340	sub AnyEvent::Base::CondVar::wait {
341	AnyEvent->one_event while !${$_[0]};
342	}	958	}
343		959
344	# default implementation for ->signal	960	# default implementation for ->signal
345		961
346	our %SIG_CB;	962	our %SIG_CB;
…		…
369		985
370	# default implementation for ->child	986	# default implementation for ->child
371		987
372	our %PID_CB;	988	our %PID_CB;
373	our $CHLD_W;	989	our $CHLD_W;
		990	our $CHLD_DELAY_W;
374	our $PID_IDLE;	991	our $PID_IDLE;
375	our $WNOHANG;	992	our $WNOHANG;
376		993
377	sub _child_wait {	994	sub _child_wait {
378	while (0 < (my $pid = waitpid -1, $WNOHANG)) {	995	while (0 < (my $pid = waitpid -1, $WNOHANG)) {
379	$_->() for values %{ (delete $PID_CB{$pid}) || {} };	996	$_->($pid, $?) for (values %{ $PID_CB{$pid} || {} }),
		997	(values %{ $PID_CB{0} || {} });
380	}	998	}
381		999
382	undef $PID_IDLE;	1000	undef $PID_IDLE;
		1001	}
		1002
		1003	sub _sigchld {
		1004	# make sure we deliver these changes "synchronous" with the event loop.
		1005	$CHLD_DELAY_W ||= AnyEvent->timer (after => 0, cb => sub {
		1006	undef $CHLD_DELAY_W;
		1007	&_child_wait;
		1008	});
383	}	1009	}
384		1010
385	sub child {	1011	sub child {
386	my (undef, %arg) = @_;	1012	my (undef, %arg) = @_;
387		1013
388	my $pid = uc $arg{pid}	1014	defined (my $pid = $arg{pid} + 0)
389	or Carp::croak "required option 'pid' is missing";	1015	or Carp::croak "required option 'pid' is missing";
390		1016
391	$PID_CB{$pid}{$arg{cb}} = $arg{cb};	1017	$PID_CB{$pid}{$arg{cb}} = $arg{cb};
392		1018
393	unless ($WNOHANG) {	1019	unless ($WNOHANG) {
394	$WNOHANG = eval { require POSIX; &POSIX::WNOHANG } || 1;	1020	$WNOHANG = eval { local $SIG{__DIE__}; require POSIX; &POSIX::WNOHANG } || 1;
395	}	1021	}
396		1022
397	unless ($CHLD_W) {	1023	unless ($CHLD_W) {
398	$CHLD_W = AnyEvent->signal (signal => 'CHLD', cb => \&_child_wait);	1024	$CHLD_W = AnyEvent->signal (signal => 'CHLD', cb => \&_sigchld);
399	# child could be a zombie already	1025	# child could be a zombie already, so make at least one round
400	$PID_IDLE ||= AnyEvent->timer (after => 0, cb => \&_child_wait);	1026	&_sigchld;
401	}	1027	}
402		1028
403	bless [$pid, $arg{cb}], "AnyEvent::Base::Child"	1029	bless [$pid, $arg{cb}], "AnyEvent::Base::Child"
404	}	1030	}
405		1031
…		…
410	delete $PID_CB{$pid} unless keys %{ $PID_CB{$pid} };	1036	delete $PID_CB{$pid} unless keys %{ $PID_CB{$pid} };
411		1037
412	undef $CHLD_W unless keys %PID_CB;	1038	undef $CHLD_W unless keys %PID_CB;
413	}	1039	}
414		1040
		1041	package AnyEvent::CondVar;
		1042
		1043	our @ISA = AnyEvent::CondVar::Base::;
		1044
		1045	package AnyEvent::CondVar::Base;
		1046
		1047	use overload
		1048	'&{}' => sub { my $self = shift; sub { $self->send (@_) } },
		1049	fallback => 1;
		1050
		1051	sub _send {
		1052	# nop
		1053	}
		1054
		1055	sub send {
		1056	my $cv = shift;
		1057	$cv->{_ae_sent} = [@_];
		1058	(delete $cv->{_ae_cb})->($cv) if $cv->{_ae_cb};
		1059	$cv->_send;
		1060	}
		1061
		1062	sub croak {
		1063	$_[0]{_ae_croak} = $_[1];
		1064	$_[0]->send;
		1065	}
		1066
		1067	sub ready {
		1068	$_[0]{_ae_sent}
		1069	}
		1070
		1071	sub _wait {
		1072	AnyEvent->one_event while !$_[0]{_ae_sent};
		1073	}
		1074
		1075	sub recv {
		1076	$_[0]->_wait;
		1077
		1078	Carp::croak $_[0]{_ae_croak} if $_[0]{_ae_croak};
		1079	wantarray ? @{ $_[0]{_ae_sent} } : $_[0]{_ae_sent}[0]
		1080	}
		1081
		1082	sub cb {
		1083	$_[0]{_ae_cb} = $_[1] if @_ > 1;
		1084	$_[0]{_ae_cb}
		1085	}
		1086
		1087	sub begin {
		1088	++$_[0]{_ae_counter};
		1089	$_[0]{_ae_end_cb} = $_[1] if @_ > 1;
		1090	}
		1091
		1092	sub end {
		1093	return if --$_[0]{_ae_counter};
		1094	&{ $_[0]{_ae_end_cb} || sub { $_[0]->send } };
		1095	}
		1096
		1097	# undocumented/compatibility with pre-3.4
		1098	*broadcast = \&send;
		1099	*wait = \&_wait;
		1100
415	=head1 SUPPLYING YOUR OWN EVENT MODEL INTERFACE	1101	=head1 SUPPLYING YOUR OWN EVENT MODEL INTERFACE
		1102
		1103	This is an advanced topic that you do not normally need to use AnyEvent in
		1104	a module. This section is only of use to event loop authors who want to
		1105	provide AnyEvent compatibility.
416		1106
417	If you need to support another event library which isn't directly	1107	If you need to support another event library which isn't directly
418	supported by AnyEvent, you can supply your own interface to it by	1108	supported by AnyEvent, you can supply your own interface to it by
419	pushing, before the first watcher gets created, the package name of	1109	pushing, before the first watcher gets created, the package name of
420	the event module and the package name of the interface to use onto	1110	the event module and the package name of the interface to use onto
421	C<@AnyEvent::REGISTRY>. You can do that before and even without loading	1111	C<@AnyEvent::REGISTRY>. You can do that before and even without loading
422	AnyEvent.	1112	AnyEvent, so it is reasonably cheap.
423		1113
424	Example:	1114	Example:
425		1115
426	push @AnyEvent::REGISTRY, [urxvt => urxvt::anyevent::];	1116	push @AnyEvent::REGISTRY, [urxvt => urxvt::anyevent::];
427		1117
428	This tells AnyEvent to (literally) use the C<urxvt::anyevent::>	1118	This tells AnyEvent to (literally) use the C<urxvt::anyevent::>
429	package/class when it finds the C<urxvt> package/module is loaded. When	1119	package/class when it finds the C<urxvt> package/module is already loaded.
		1120
430	AnyEvent is loaded and asked to find a suitable event model, it will	1121	When AnyEvent is loaded and asked to find a suitable event model, it
431	first check for the presence of urxvt.	1122	will first check for the presence of urxvt by trying to C<use> the
		1123	C<urxvt::anyevent> module.
432		1124
433	The class should provide implementations for all watcher types (see	1125	The class should provide implementations for all watcher types. See
434	L<AnyEvent::Impl::Event> (source code), L<AnyEvent::Impl::Glib>	1126	L<AnyEvent::Impl::EV> (source code), L<AnyEvent::Impl::Glib> (Source code)
435	(Source code) and so on for actual examples, use C<perldoc -m	1127	and so on for actual examples. Use C<perldoc -m AnyEvent::Impl::Glib> to
436	AnyEvent::Impl::Glib> to see the sources).	1128	see the sources.
437		1129
		1130	If you don't provide C<signal> and C<child> watchers than AnyEvent will
		1131	provide suitable (hopefully) replacements.
		1132
438	The above isn't fictitious, the I<rxvt-unicode> (a.k.a. urxvt)	1133	The above example isn't fictitious, the I<rxvt-unicode> (a.k.a. urxvt)
439	uses the above line as-is. An interface isn't included in AnyEvent	1134	terminal emulator uses the above line as-is. An interface isn't included
440	because it doesn't make sense outside the embedded interpreter inside	1135	in AnyEvent because it doesn't make sense outside the embedded interpreter
441	I<rxvt-unicode>, and it is updated and maintained as part of the	1136	inside I<rxvt-unicode>, and it is updated and maintained as part of the
442	I<rxvt-unicode> distribution.	1137	I<rxvt-unicode> distribution.
443		1138
444	I<rxvt-unicode> also cheats a bit by not providing blocking access to	1139	I<rxvt-unicode> also cheats a bit by not providing blocking access to
445	condition variables: code blocking while waiting for a condition will	1140	condition variables: code blocking while waiting for a condition will
446	C<die>. This still works with most modules/usages, and blocking calls must	1141	C<die>. This still works with most modules/usages, and blocking calls must
447	not be in an interactive application, so it makes sense.	1142	not be done in an interactive application, so it makes sense.
448		1143
449	=head1 ENVIRONMENT VARIABLES	1144	=head1 ENVIRONMENT VARIABLES
450		1145
451	The following environment variables are used by this module:	1146	The following environment variables are used by this module:
452		1147
453	C<PERL_ANYEVENT_VERBOSE> when set to C<2> or higher, reports which event	1148	=over 4
454	model gets used.
455		1149
		1150	=item C<PERL_ANYEVENT_VERBOSE>
		1151
		1152	By default, AnyEvent will be completely silent except in fatal
		1153	conditions. You can set this environment variable to make AnyEvent more
		1154	talkative.
		1155
		1156	When set to C<1> or higher, causes AnyEvent to warn about unexpected
		1157	conditions, such as not being able to load the event model specified by
		1158	C<PERL_ANYEVENT_MODEL>.
		1159
		1160	When set to C<2> or higher, cause AnyEvent to report to STDERR which event
		1161	model it chooses.
		1162
		1163	=item C<PERL_ANYEVENT_MODEL>
		1164
		1165	This can be used to specify the event model to be used by AnyEvent, before
		1166	auto detection and -probing kicks in. It must be a string consisting
		1167	entirely of ASCII letters. The string C<AnyEvent::Impl::> gets prepended
		1168	and the resulting module name is loaded and if the load was successful,
		1169	used as event model. If it fails to load AnyEvent will proceed with
		1170	auto detection and -probing.
		1171
		1172	This functionality might change in future versions.
		1173
		1174	For example, to force the pure perl model (L<AnyEvent::Impl::Perl>) you
		1175	could start your program like this:
		1176
		1177	PERL_ANYEVENT_MODEL=Perl perl ...
		1178
		1179	=item C<PERL_ANYEVENT_PROTOCOLS>
		1180
		1181	Used by both L<AnyEvent::DNS> and L<AnyEvent::Socket> to determine preferences
		1182	for IPv4 or IPv6. The default is unspecified (and might change, or be the result
		1183	of auto probing).
		1184
		1185	Must be set to a comma-separated list of protocols or address families,
		1186	current supported: C<ipv4> and C<ipv6>. Only protocols mentioned will be
		1187	used, and preference will be given to protocols mentioned earlier in the
		1188	list.
		1189
		1190	This variable can effectively be used for denial-of-service attacks
		1191	against local programs (e.g. when setuid), although the impact is likely
		1192	small, as the program has to handle connection errors already-
		1193
		1194	Examples: C<PERL_ANYEVENT_PROTOCOLS=ipv4,ipv6> - prefer IPv4 over IPv6,
		1195	but support both and try to use both. C<PERL_ANYEVENT_PROTOCOLS=ipv4>
		1196	- only support IPv4, never try to resolve or contact IPv6
		1197	addresses. C<PERL_ANYEVENT_PROTOCOLS=ipv6,ipv4> support either IPv4 or
		1198	IPv6, but prefer IPv6 over IPv4.
		1199
		1200	=item C<PERL_ANYEVENT_EDNS0>
		1201
		1202	Used by L<AnyEvent::DNS> to decide whether to use the EDNS0 extension
		1203	for DNS. This extension is generally useful to reduce DNS traffic, but
		1204	some (broken) firewalls drop such DNS packets, which is why it is off by
		1205	default.
		1206
		1207	Setting this variable to C<1> will cause L<AnyEvent::DNS> to announce
		1208	EDNS0 in its DNS requests.
		1209
		1210	=item C<PERL_ANYEVENT_MAX_FORKS>
		1211
		1212	The maximum number of child processes that C<AnyEvent::Util::fork_call>
		1213	will create in parallel.
		1214
		1215	=back
		1216
456	=head1 EXAMPLE	1217	=head1 EXAMPLE PROGRAM
457		1218
458	The following program uses an io watcher to read data from stdin, a timer	1219	The following program uses an I/O watcher to read data from STDIN, a timer
459	to display a message once per second, and a condvar to exit the program	1220	to display a message once per second, and a condition variable to quit the
460	when the user enters quit:	1221	program when the user enters quit:
461		1222
462	use AnyEvent;	1223	use AnyEvent;
463		1224
464	my $cv = AnyEvent->condvar;	1225	my $cv = AnyEvent->condvar;
465		1226
466	my $io_watcher = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {	1227	my $io_watcher = AnyEvent->io (
		1228	fh => \*STDIN,
		1229	poll => 'r',
		1230	cb => sub {
467	warn "io event <$_[0]>\n"; # will always output <r>	1231	warn "io event <$_[0]>\n"; # will always output <r>
468	chomp (my $input = <STDIN>); # read a line	1232	chomp (my $input = <STDIN>); # read a line
469	warn "read: $input\n"; # output what has been read	1233	warn "read: $input\n"; # output what has been read
470	$cv->broadcast if $input =~ /^q/i; # quit program if /^q/i	1234	$cv->send if $input =~ /^q/i; # quit program if /^q/i
		1235	},
471	});	1236	);
472		1237
473	my $time_watcher; # can only be used once	1238	my $time_watcher; # can only be used once
474		1239
475	sub new_timer {	1240	sub new_timer {
476	$timer = AnyEvent->timer (after => 1, cb => sub {	1241	$timer = AnyEvent->timer (after => 1, cb => sub {
…		…
479	});	1244	});
480	}	1245	}
481		1246
482	new_timer; # create first timer	1247	new_timer; # create first timer
483		1248
484	$cv->wait; # wait until user enters /^q/i	1249	$cv->recv; # wait until user enters /^q/i
485		1250
486	=head1 REAL-WORLD EXAMPLE	1251	=head1 REAL-WORLD EXAMPLE
487		1252
488	Consider the L<Net::FCP> module. It features (among others) the following	1253	Consider the L<Net::FCP> module. It features (among others) the following
489	API calls, which are to freenet what HTTP GET requests are to http:	1254	API calls, which are to freenet what HTTP GET requests are to http:
…		…
539	syswrite $txn->{fh}, $txn->{request}	1304	syswrite $txn->{fh}, $txn->{request}
540	or die "connection or write error";	1305	or die "connection or write error";
541	$txn->{w} = AnyEvent->io (fh => $txn->{fh}, poll => 'r', cb => sub { $txn->fh_ready_r });	1306	$txn->{w} = AnyEvent->io (fh => $txn->{fh}, poll => 'r', cb => sub { $txn->fh_ready_r });
542		1307
543	Again, C<fh_ready_r> waits till all data has arrived, and then stores the	1308	Again, C<fh_ready_r> waits till all data has arrived, and then stores the
544	result and signals any possible waiters that the request ahs finished:	1309	result and signals any possible waiters that the request has finished:
545		1310
546	sysread $txn->{fh}, $txn->{buf}, length $txn->{$buf};	1311	sysread $txn->{fh}, $txn->{buf}, length $txn->{$buf};
547		1312
548	if (end-of-file or data complete) {	1313	if (end-of-file or data complete) {
549	$txn->{result} = $txn->{buf};	1314	$txn->{result} = $txn->{buf};
550	$txn->{finished}->broadcast;	1315	$txn->{finished}->send;
551	$txb->{cb}->($txn) of $txn->{cb}; # also call callback	1316	$txb->{cb}->($txn) of $txn->{cb}; # also call callback
552	}	1317	}
553		1318
554	The C<result> method, finally, just waits for the finished signal (if the	1319	The C<result> method, finally, just waits for the finished signal (if the
555	request was already finished, it doesn't wait, of course, and returns the	1320	request was already finished, it doesn't wait, of course, and returns the
556	data:	1321	data:
557		1322
558	$txn->{finished}->wait;	1323	$txn->{finished}->recv;
559	return $txn->{result};	1324	return $txn->{result};
560		1325
561	The actual code goes further and collects all errors (C<die>s, exceptions)	1326	The actual code goes further and collects all errors (C<die>s, exceptions)
562	that occured during request processing. The C<result> method detects	1327	that occurred during request processing. The C<result> method detects
563	wether an exception as thrown (it is stored inside the $txn object)	1328	whether an exception as thrown (it is stored inside the $txn object)
564	and just throws the exception, which means connection errors and other	1329	and just throws the exception, which means connection errors and other
565	problems get reported tot he code that tries to use the result, not in a	1330	problems get reported tot he code that tries to use the result, not in a
566	random callback.	1331	random callback.
567		1332
568	All of this enables the following usage styles:	1333	All of this enables the following usage styles:
569		1334
570	1. Blocking:	1335	1. Blocking:
571		1336
572	my $data = $fcp->client_get ($url);	1337	my $data = $fcp->client_get ($url);
573		1338
574	2. Blocking, but parallelizing:	1339	2. Blocking, but running in parallel:
575		1340
576	my @datas = map $_->result,	1341	my @datas = map $_->result,
577	map $fcp->txn_client_get ($_),	1342	map $fcp->txn_client_get ($_),
578	@urls;	1343	@urls;
579		1344
580	Both blocking examples work without the module user having to know	1345	Both blocking examples work without the module user having to know
581	anything about events.	1346	anything about events.
582		1347
583	3a. Event-based in a main program, using any support Event module:	1348	3a. Event-based in a main program, using any supported event module:
584		1349
585	use Event;	1350	use EV;
586		1351
587	$fcp->txn_client_get ($url)->cb (sub {	1352	$fcp->txn_client_get ($url)->cb (sub {
588	my $txn = shift;	1353	my $txn = shift;
589	my $data = $txn->result;	1354	my $data = $txn->result;
590	...	1355	...
591	});	1356	});
592		1357
593	Event::loop;	1358	EV::loop;
594		1359
595	3b. The module user could use AnyEvent, too:	1360	3b. The module user could use AnyEvent, too:
596		1361
597	use AnyEvent;	1362	use AnyEvent;
598		1363
599	my $quit = AnyEvent->condvar;	1364	my $quit = AnyEvent->condvar;
600		1365
601	$fcp->txn_client_get ($url)->cb (sub {	1366	$fcp->txn_client_get ($url)->cb (sub {
602	...	1367	...
603	$quit->broadcast;	1368	$quit->send;
604	});	1369	});
605		1370
606	$quit->wait;	1371	$quit->recv;
		1372
		1373
		1374	=head1 BENCHMARKS
		1375
		1376	To give you an idea of the performance and overheads that AnyEvent adds
		1377	over the event loops themselves and to give you an impression of the speed
		1378	of various event loops I prepared some benchmarks.
		1379
		1380	=head2 BENCHMARKING ANYEVENT OVERHEAD
		1381
		1382	Here is a benchmark of various supported event models used natively and
		1383	through AnyEvent. The benchmark creates a lot of timers (with a zero
		1384	timeout) and I/O watchers (watching STDOUT, a pty, to become writable,
		1385	which it is), lets them fire exactly once and destroys them again.
		1386
		1387	Source code for this benchmark is found as F<eg/bench> in the AnyEvent
		1388	distribution.
		1389
		1390	=head3 Explanation of the columns
		1391
		1392	I<watcher> is the number of event watchers created/destroyed. Since
		1393	different event models feature vastly different performances, each event
		1394	loop was given a number of watchers so that overall runtime is acceptable
		1395	and similar between tested event loop (and keep them from crashing): Glib
		1396	would probably take thousands of years if asked to process the same number
		1397	of watchers as EV in this benchmark.
		1398
		1399	I<bytes> is the number of bytes (as measured by the resident set size,
		1400	RSS) consumed by each watcher. This method of measuring captures both C
		1401	and Perl-based overheads.
		1402
		1403	I<create> is the time, in microseconds (millionths of seconds), that it
		1404	takes to create a single watcher. The callback is a closure shared between
		1405	all watchers, to avoid adding memory overhead. That means closure creation
		1406	and memory usage is not included in the figures.
		1407
		1408	I<invoke> is the time, in microseconds, used to invoke a simple
		1409	callback. The callback simply counts down a Perl variable and after it was
		1410	invoked "watcher" times, it would C<< ->send >> a condvar once to
		1411	signal the end of this phase.
		1412
		1413	I<destroy> is the time, in microseconds, that it takes to destroy a single
		1414	watcher.
		1415
		1416	=head3 Results
		1417
		1418	name watchers bytes create invoke destroy comment
		1419	EV/EV 400000 244 0.56 0.46 0.31 EV native interface
		1420	EV/Any 100000 244 2.50 0.46 0.29 EV + AnyEvent watchers
		1421	CoroEV/Any 100000 244 2.49 0.44 0.29 coroutines + Coro::Signal
		1422	Perl/Any 100000 513 4.92 0.87 1.12 pure perl implementation
		1423	Event/Event 16000 516 31.88 31.30 0.85 Event native interface
		1424	Event/Any 16000 590 35.75 31.42 1.08 Event + AnyEvent watchers
		1425	Glib/Any 16000 1357 98.22 12.41 54.00 quadratic behaviour
		1426	Tk/Any 2000 1860 26.97 67.98 14.00 SEGV with >> 2000 watchers
		1427	POE/Event 2000 6644 108.64 736.02 14.73 via POE::Loop::Event
		1428	POE/Select 2000 6343 94.13 809.12 565.96 via POE::Loop::Select
		1429
		1430	=head3 Discussion
		1431
		1432	The benchmark does I<not> measure scalability of the event loop very
		1433	well. For example, a select-based event loop (such as the pure perl one)
		1434	can never compete with an event loop that uses epoll when the number of
		1435	file descriptors grows high. In this benchmark, all events become ready at
		1436	the same time, so select/poll-based implementations get an unnatural speed
		1437	boost.
		1438
		1439	Also, note that the number of watchers usually has a nonlinear effect on
		1440	overall speed, that is, creating twice as many watchers doesn't take twice
		1441	the time - usually it takes longer. This puts event loops tested with a
		1442	higher number of watchers at a disadvantage.
		1443
		1444	To put the range of results into perspective, consider that on the
		1445	benchmark machine, handling an event takes roughly 1600 CPU cycles with
		1446	EV, 3100 CPU cycles with AnyEvent's pure perl loop and almost 3000000 CPU
		1447	cycles with POE.
		1448
		1449	C<EV> is the sole leader regarding speed and memory use, which are both
		1450	maximal/minimal, respectively. Even when going through AnyEvent, it uses
		1451	far less memory than any other event loop and is still faster than Event
		1452	natively.
		1453
		1454	The pure perl implementation is hit in a few sweet spots (both the
		1455	constant timeout and the use of a single fd hit optimisations in the perl
		1456	interpreter and the backend itself). Nevertheless this shows that it
		1457	adds very little overhead in itself. Like any select-based backend its
		1458	performance becomes really bad with lots of file descriptors (and few of
		1459	them active), of course, but this was not subject of this benchmark.
		1460
		1461	The C<Event> module has a relatively high setup and callback invocation
		1462	cost, but overall scores in on the third place.
		1463
		1464	C<Glib>'s memory usage is quite a bit higher, but it features a
		1465	faster callback invocation and overall ends up in the same class as
		1466	C<Event>. However, Glib scales extremely badly, doubling the number of
		1467	watchers increases the processing time by more than a factor of four,
		1468	making it completely unusable when using larger numbers of watchers
		1469	(note that only a single file descriptor was used in the benchmark, so
		1470	inefficiencies of C<poll> do not account for this).
		1471
		1472	The C<Tk> adaptor works relatively well. The fact that it crashes with
		1473	more than 2000 watchers is a big setback, however, as correctness takes
		1474	precedence over speed. Nevertheless, its performance is surprising, as the
		1475	file descriptor is dup()ed for each watcher. This shows that the dup()
		1476	employed by some adaptors is not a big performance issue (it does incur a
		1477	hidden memory cost inside the kernel which is not reflected in the figures
		1478	above).
		1479
		1480	C<POE>, regardless of underlying event loop (whether using its pure perl
		1481	select-based backend or the Event module, the POE-EV backend couldn't
		1482	be tested because it wasn't working) shows abysmal performance and
		1483	memory usage with AnyEvent: Watchers use almost 30 times as much memory
		1484	as EV watchers, and 10 times as much memory as Event (the high memory
		1485	requirements are caused by requiring a session for each watcher). Watcher
		1486	invocation speed is almost 900 times slower than with AnyEvent's pure perl
		1487	implementation.
		1488
		1489	The design of the POE adaptor class in AnyEvent can not really account
		1490	for the performance issues, though, as session creation overhead is
		1491	small compared to execution of the state machine, which is coded pretty
		1492	optimally within L<AnyEvent::Impl::POE> (and while everybody agrees that
		1493	using multiple sessions is not a good approach, especially regarding
		1494	memory usage, even the author of POE could not come up with a faster
		1495	design).
		1496
		1497	=head3 Summary
		1498
		1499	=over 4
		1500
		1501	=item * Using EV through AnyEvent is faster than any other event loop
		1502	(even when used without AnyEvent), but most event loops have acceptable
		1503	performance with or without AnyEvent.
		1504
		1505	=item * The overhead AnyEvent adds is usually much smaller than the overhead of
		1506	the actual event loop, only with extremely fast event loops such as EV
		1507	adds AnyEvent significant overhead.
		1508
		1509	=item * You should avoid POE like the plague if you want performance or
		1510	reasonable memory usage.
		1511
		1512	=back
		1513
		1514	=head2 BENCHMARKING THE LARGE SERVER CASE
		1515
		1516	This benchmark actually benchmarks the event loop itself. It works by
		1517	creating a number of "servers": each server consists of a socket pair, a
		1518	timeout watcher that gets reset on activity (but never fires), and an I/O
		1519	watcher waiting for input on one side of the socket. Each time the socket
		1520	watcher reads a byte it will write that byte to a random other "server".
		1521
		1522	The effect is that there will be a lot of I/O watchers, only part of which
		1523	are active at any one point (so there is a constant number of active
		1524	fds for each loop iteration, but which fds these are is random). The
		1525	timeout is reset each time something is read because that reflects how
		1526	most timeouts work (and puts extra pressure on the event loops).
		1527
		1528	In this benchmark, we use 10000 socket pairs (20000 sockets), of which 100
		1529	(1%) are active. This mirrors the activity of large servers with many
		1530	connections, most of which are idle at any one point in time.
		1531
		1532	Source code for this benchmark is found as F<eg/bench2> in the AnyEvent
		1533	distribution.
		1534
		1535	=head3 Explanation of the columns
		1536
		1537	I<sockets> is the number of sockets, and twice the number of "servers" (as
		1538	each server has a read and write socket end).
		1539
		1540	I<create> is the time it takes to create a socket pair (which is
		1541	nontrivial) and two watchers: an I/O watcher and a timeout watcher.
		1542
		1543	I<request>, the most important value, is the time it takes to handle a
		1544	single "request", that is, reading the token from the pipe and forwarding
		1545	it to another server. This includes deleting the old timeout and creating
		1546	a new one that moves the timeout into the future.
		1547
		1548	=head3 Results
		1549
		1550	name sockets create request
		1551	EV 20000 69.01 11.16
		1552	Perl 20000 73.32 35.87
		1553	Event 20000 212.62 257.32
		1554	Glib 20000 651.16 1896.30
		1555	POE 20000 349.67 12317.24 uses POE::Loop::Event
		1556
		1557	=head3 Discussion
		1558
		1559	This benchmark I<does> measure scalability and overall performance of the
		1560	particular event loop.
		1561
		1562	EV is again fastest. Since it is using epoll on my system, the setup time
		1563	is relatively high, though.
		1564
		1565	Perl surprisingly comes second. It is much faster than the C-based event
		1566	loops Event and Glib.
		1567
		1568	Event suffers from high setup time as well (look at its code and you will
		1569	understand why). Callback invocation also has a high overhead compared to
		1570	the C<< $_->() for .. >>-style loop that the Perl event loop uses. Event
		1571	uses select or poll in basically all documented configurations.
		1572
		1573	Glib is hit hard by its quadratic behaviour w.r.t. many watchers. It
		1574	clearly fails to perform with many filehandles or in busy servers.
		1575
		1576	POE is still completely out of the picture, taking over 1000 times as long
		1577	as EV, and over 100 times as long as the Perl implementation, even though
		1578	it uses a C-based event loop in this case.
		1579
		1580	=head3 Summary
		1581
		1582	=over 4
		1583
		1584	=item * The pure perl implementation performs extremely well.
		1585
		1586	=item * Avoid Glib or POE in large projects where performance matters.
		1587
		1588	=back
		1589
		1590	=head2 BENCHMARKING SMALL SERVERS
		1591
		1592	While event loops should scale (and select-based ones do not...) even to
		1593	large servers, most programs we (or I :) actually write have only a few
		1594	I/O watchers.
		1595
		1596	In this benchmark, I use the same benchmark program as in the large server
		1597	case, but it uses only eight "servers", of which three are active at any
		1598	one time. This should reflect performance for a small server relatively
		1599	well.
		1600
		1601	The columns are identical to the previous table.
		1602
		1603	=head3 Results
		1604
		1605	name sockets create request
		1606	EV 16 20.00 6.54
		1607	Perl 16 25.75 12.62
		1608	Event 16 81.27 35.86
		1609	Glib 16 32.63 15.48
		1610	POE 16 261.87 276.28 uses POE::Loop::Event
		1611
		1612	=head3 Discussion
		1613
		1614	The benchmark tries to test the performance of a typical small
		1615	server. While knowing how various event loops perform is interesting, keep
		1616	in mind that their overhead in this case is usually not as important, due
		1617	to the small absolute number of watchers (that is, you need efficiency and
		1618	speed most when you have lots of watchers, not when you only have a few of
		1619	them).
		1620
		1621	EV is again fastest.
		1622
		1623	Perl again comes second. It is noticeably faster than the C-based event
		1624	loops Event and Glib, although the difference is too small to really
		1625	matter.
		1626
		1627	POE also performs much better in this case, but is is still far behind the
		1628	others.
		1629
		1630	=head3 Summary
		1631
		1632	=over 4
		1633
		1634	=item * C-based event loops perform very well with small number of
		1635	watchers, as the management overhead dominates.
		1636
		1637	=back
		1638
		1639
		1640	=head1 FORK
		1641
		1642	Most event libraries are not fork-safe. The ones who are usually are
		1643	because they rely on inefficient but fork-safe C<select> or C<poll>
		1644	calls. Only L<EV> is fully fork-aware.
		1645
		1646	If you have to fork, you must either do so I<before> creating your first
		1647	watcher OR you must not use AnyEvent at all in the child.
		1648
		1649
		1650	=head1 SECURITY CONSIDERATIONS
		1651
		1652	AnyEvent can be forced to load any event model via
		1653	$ENV{PERL_ANYEVENT_MODEL}. While this cannot (to my knowledge) be used to
		1654	execute arbitrary code or directly gain access, it can easily be used to
		1655	make the program hang or malfunction in subtle ways, as AnyEvent watchers
		1656	will not be active when the program uses a different event model than
		1657	specified in the variable.
		1658
		1659	You can make AnyEvent completely ignore this variable by deleting it
		1660	before the first watcher gets created, e.g. with a C<BEGIN> block:
		1661
		1662	BEGIN { delete $ENV{PERL_ANYEVENT_MODEL} }
		1663
		1664	use AnyEvent;
		1665
		1666	Similar considerations apply to $ENV{PERL_ANYEVENT_VERBOSE}, as that can
		1667	be used to probe what backend is used and gain other information (which is
		1668	probably even less useful to an attacker than PERL_ANYEVENT_MODEL).
		1669
607		1670
608	=head1 SEE ALSO	1671	=head1 SEE ALSO
609		1672
610	Event modules: L<Coro::Event>, L<Coro>, L<Event>, L<Glib::Event>, L<Glib>.	1673	Utility functions: L<AnyEvent::Util>.
611		1674
612	Implementations: L<AnyEvent::Impl::Coro>, L<AnyEvent::Impl::Event>, L<AnyEvent::Impl::Glib>, L<AnyEvent::Impl::Tk>.	1675	Event modules: L<EV>, L<EV::Glib>, L<Glib::EV>, L<Event>, L<Glib::Event>,
		1676	L<Glib>, L<Tk>, L<Event::Lib>, L<Qt>, L<POE>.
613		1677
614	Nontrivial usage example: L<Net::FCP>.	1678	Implementations: L<AnyEvent::Impl::EV>, L<AnyEvent::Impl::Event>,
		1679	L<AnyEvent::Impl::Glib>, L<AnyEvent::Impl::Tk>, L<AnyEvent::Impl::Perl>,
		1680	L<AnyEvent::Impl::EventLib>, L<AnyEvent::Impl::Qt>,
		1681	L<AnyEvent::Impl::POE>.
615		1682
616	=head1	1683	Non-blocking file handles, sockets, TCP clients and
		1684	servers: L<AnyEvent::Handle>, L<AnyEvent::Socket>.
		1685
		1686	Asynchronous DNS: L<AnyEvent::DNS>.
		1687
		1688	Coroutine support: L<Coro>, L<Coro::AnyEvent>, L<Coro::EV>, L<Coro::Event>,
		1689
		1690	Nontrivial usage examples: L<Net::FCP>, L<Net::XMPP2>, L<AnyEvent::DNS>.
		1691
		1692
		1693	=head1 AUTHOR
		1694
		1695	Marc Lehmann <schmorp@schmorp.de>
		1696	http://home.schmorp.de/
617		1697
618	=cut	1698	=cut
619		1699
620	1	1700	1
621		1701

Diff Legend

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