ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/AnyEvent/lib/AnyEvent.pm

Comparing AnyEvent/lib/AnyEvent.pm (file contents):
Revision 1.10 by root, Fri Jan 13 13:16:15 2006 UTC vs.
Revision 1.206 by root, Mon Apr 20 14:34:18 2009 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 - 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
11	my $w = AnyEvent->io (fh => ..., poll => "[rw]+", cb => sub {	11	my $w = AnyEvent->io (fh => $fh, poll => "r|w", cb => sub { ... });
12	my ($poll_got) = @_;	12
		13	my $w = AnyEvent->timer (after => $seconds, cb => sub { ... });
		14	my $w = AnyEvent->timer (after => $seconds, interval => $seconds, cb => ...
		15
		16	print AnyEvent->now; # prints current event loop time
		17	print AnyEvent->time; # think Time::HiRes::time or simply CORE::time.
		18
		19	my $w = AnyEvent->signal (signal => "TERM", cb => sub { ... });
		20
		21	my $w = AnyEvent->child (pid => $pid, cb => sub {
		22	my ($pid, $status) = @_;
13	...	23	...
14	});	24	});
15		25
16	* only one io watcher per $fh and $poll type is allowed (i.e. on a socket	26	my $w = AnyEvent->condvar; # stores whether a condition was flagged
17	you can have one r + one w or one rw watcher, not any more (limitation by	27	$w->send; # wake up current and all future recv's
18	Tk).
19
20	* the C<$poll_got> passed to the handler needs to be checked by looking
21	for single characters (e.g. with a regex), as it can contain more event
22	types than were requested (e.g. a 'w' watcher might generate 'rw' events,
23	limitation by Glib).
24
25	* AnyEvent will keep filehandles alive, so as long as the watcher exists,
26	the filehandle exists.
27
28	my $w = AnyEvent->timer (after => $seconds, cb => sub {
29	...
30	});
31
32	* io and time watchers get canceled whenever $w is destroyed, so keep a copy
33
34	* timers can only be used once and must be recreated for repeated
35	operation (limitation by Glib and Tk).
36
37	my $w = AnyEvent->condvar; # kind of main loop replacement
38	$w->wait; # enters main loop till $condvar gets ->broadcast	28	$w->recv; # enters "main loop" till $condvar gets ->send
39	$w->broadcast; # wake up current and all future wait's	29	# use a condvar in callback mode:
		30	$w->cb (sub { $_[0]->recv });
40		31
41	* condvars are used to give blocking behaviour when neccessary. Create	32	=head1 INTRODUCTION/TUTORIAL
42	a condvar for any "request" or "event" your module might create, C<<	33
43	->broadcast >> it when the event happens and provide a function that calls	34	This manpage is mainly a reference manual. If you are interested
44	C<< ->wait >> for it. See the examples below.	35	in a tutorial or some gentle introduction, have a look at the
		36	L<AnyEvent::Intro> manpage.
		37
		38	=head1 WHY YOU SHOULD USE THIS MODULE (OR NOT)
		39
		40	Glib, POE, IO::Async, Event... CPAN offers event models by the dozen
		41	nowadays. So what is different about AnyEvent?
		42
		43	Executive Summary: AnyEvent is I<compatible>, AnyEvent is I<free of
		44	policy> and AnyEvent is I<small and efficient>.
		45
		46	First and foremost, I<AnyEvent is not an event model> itself, it only
		47	interfaces to whatever event model the main program happens to use, in a
		48	pragmatic way. For event models and certain classes of immortals alike,
		49	the statement "there can only be one" is a bitter reality: In general,
		50	only one event loop can be active at the same time in a process. AnyEvent
		51	cannot change this, but it can hide the differences between those event
		52	loops.
		53
		54	The goal of AnyEvent is to offer module authors the ability to do event
		55	programming (waiting for I/O or timer events) without subscribing to a
		56	religion, a way of living, and most importantly: without forcing your
		57	module users into the same thing by forcing them to use the same event
		58	model you use.
		59
		60	For modules like POE or IO::Async (which is a total misnomer as it is
		61	actually doing all I/O I<synchronously>...), using them in your module is
		62	like joining a cult: After you joined, you are dependent on them and you
		63	cannot use anything else, as they are simply incompatible to everything
		64	that isn't them. What's worse, all the potential users of your
		65	module are I<also> forced to use the same event loop you use.
		66
		67	AnyEvent is different: AnyEvent + POE works fine. AnyEvent + Glib works
		68	fine. AnyEvent + Tk works fine etc. etc. but none of these work together
		69	with the rest: POE + IO::Async? No go. Tk + Event? No go. Again: if
		70	your module uses one of those, every user of your module has to use it,
		71	too. But if your module uses AnyEvent, it works transparently with all
		72	event models it supports (including stuff like IO::Async, as long as those
		73	use one of the supported event loops. It is trivial to add new event loops
		74	to AnyEvent, too, so it is future-proof).
		75
		76	In addition to being free of having to use I<the one and only true event
		77	model>, AnyEvent also is free of bloat and policy: with POE or similar
		78	modules, you get an enormous amount of code and strict rules you have to
		79	follow. AnyEvent, on the other hand, is lean and up to the point, by only
		80	offering the functionality that is necessary, in as thin as a wrapper as
		81	technically possible.
		82
		83	Of course, AnyEvent comes with a big (and fully optional!) toolbox
		84	of useful functionality, such as an asynchronous DNS resolver, 100%
		85	non-blocking connects (even with TLS/SSL, IPv6 and on broken platforms
		86	such as Windows) and lots of real-world knowledge and workarounds for
		87	platform bugs and differences.
		88
		89	Now, if you I<do want> lots of policy (this can arguably be somewhat
		90	useful) and you want to force your users to use the one and only event
		91	model, you should I<not> use this module.
45		92
46	=head1 DESCRIPTION	93	=head1 DESCRIPTION
47		94
48	L<AnyEvent> provides an identical interface to multiple event loops. This	95	L<AnyEvent> provides an identical interface to multiple event loops. This
49	allows module authors to utilizy an event loop without forcing module	96	allows module authors to utilise an event loop without forcing module
50	users to use the same event loop (as only a single event loop can coexist	97	users to use the same event loop (as only a single event loop can coexist
51	peacefully at any one time).	98	peacefully at any one time).
52		99
53	The interface itself is vaguely similar but not identical to the Event	100	The interface itself is vaguely similar, but not identical to the L<Event>
54	module.	101	module.
55		102
56	On the first call of any method, the module tries to detect the currently	103	During the first call of any watcher-creation method, the module tries
57	loaded event loop by probing wether any of the following modules is	104	to detect the currently loaded event loop by probing whether one of the
58	loaded: L<Coro::Event>, L<Event>, L<Glib>, L<Tk>. The first one found is	105	following modules is already loaded: L<EV>,
59	used. If none is found, the module tries to load these modules in the	106	L<Event>, L<Glib>, L<AnyEvent::Impl::Perl>, L<Tk>, L<Event::Lib>, L<Qt>,
60	order given. The first one that could be successfully loaded will be	107	L<POE>. The first one found is used. If none are found, the module tries
61	used. If still none could be found, it will issue an error.	108	to load these modules (excluding Tk, Event::Lib, Qt and POE as the pure perl
		109	adaptor should always succeed) in the order given. The first one that can
		110	be successfully loaded will be used. If, after this, still none could be
		111	found, AnyEvent will fall back to a pure-perl event loop, which is not
		112	very efficient, but should work everywhere.
		113
		114	Because AnyEvent first checks for modules that are already loaded, loading
		115	an event model explicitly before first using AnyEvent will likely make
		116	that model the default. For example:
		117
		118	use Tk;
		119	use AnyEvent;
		120
		121	# .. AnyEvent will likely default to Tk
		122
		123	The I<likely> means that, if any module loads another event model and
		124	starts using it, all bets are off. Maybe you should tell their authors to
		125	use AnyEvent so their modules work together with others seamlessly...
		126
		127	The pure-perl implementation of AnyEvent is called
		128	C<AnyEvent::Impl::Perl>. Like other event modules you can load it
		129	explicitly and enjoy the high availability of that event loop :)
		130
		131	=head1 WATCHERS
		132
		133	AnyEvent has the central concept of a I<watcher>, which is an object that
		134	stores relevant data for each kind of event you are waiting for, such as
		135	the callback to call, the file handle to watch, etc.
		136
		137	These watchers are normal Perl objects with normal Perl lifetime. After
		138	creating a watcher it will immediately "watch" for events and invoke the
		139	callback when the event occurs (of course, only when the event model
		140	is in control).
		141
		142	Note that B<callbacks must not permanently change global variables>
		143	potentially in use by the event loop (such as C<$_> or C<$[>) and that B<<
		144	callbacks must not C<die> >>. The former is good programming practise in
		145	Perl and the latter stems from the fact that exception handling differs
		146	widely between event loops.
		147
		148	To disable the watcher you have to destroy it (e.g. by setting the
		149	variable you store it in to C<undef> or otherwise deleting all references
		150	to it).
		151
		152	All watchers are created by calling a method on the C<AnyEvent> class.
		153
		154	Many watchers either are used with "recursion" (repeating timers for
		155	example), or need to refer to their watcher object in other ways.
		156
		157	An any way to achieve that is this pattern:
		158
		159	my $w; $w = AnyEvent->type (arg => value ..., cb => sub {
		160	# you can use $w here, for example to undef it
		161	undef $w;
		162	});
		163
		164	Note that C<my $w; $w => combination. This is necessary because in Perl,
		165	my variables are only visible after the statement in which they are
		166	declared.
		167
		168	=head2 I/O WATCHERS
		169
		170	You can create an I/O watcher by calling the C<< AnyEvent->io >> method
		171	with the following mandatory key-value pairs as arguments:
		172
		173	C<fh> is the Perl I<file handle> (I<not> file descriptor) to watch
		174	for events (AnyEvent might or might not keep a reference to this file
		175	handle). Note that only file handles pointing to things for which
		176	non-blocking operation makes sense are allowed. This includes sockets,
		177	most character devices, pipes, fifos and so on, but not for example files
		178	or block devices.
		179
		180	C<poll> must be a string that is either C<r> or C<w>, which creates a
		181	watcher waiting for "r"eadable or "w"ritable events, respectively.
		182
		183	C<cb> is the callback to invoke each time the file handle becomes ready.
		184
		185	Although the callback might get passed parameters, their value and
		186	presence is undefined and you cannot rely on them. Portable AnyEvent
		187	callbacks cannot use arguments passed to I/O watcher callbacks.
		188
		189	The I/O watcher might use the underlying file descriptor or a copy of it.
		190	You must not close a file handle as long as any watcher is active on the
		191	underlying file descriptor.
		192
		193	Some event loops issue spurious readyness notifications, so you should
		194	always use non-blocking calls when reading/writing from/to your file
		195	handles.
		196
		197	Example: wait for readability of STDIN, then read a line and disable the
		198	watcher.
		199
		200	my $w; $w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {
		201	chomp (my $input = <STDIN>);
		202	warn "read: $input\n";
		203	undef $w;
		204	});
		205
		206	=head2 TIME WATCHERS
		207
		208	You can create a time watcher by calling the C<< AnyEvent->timer >>
		209	method with the following mandatory arguments:
		210
		211	C<after> specifies after how many seconds (fractional values are
		212	supported) the callback should be invoked. C<cb> is the callback to invoke
		213	in that case.
		214
		215	Although the callback might get passed parameters, their value and
		216	presence is undefined and you cannot rely on them. Portable AnyEvent
		217	callbacks cannot use arguments passed to time watcher callbacks.
		218
		219	The callback will normally be invoked once only. If you specify another
		220	parameter, C<interval>, as a strictly positive number (> 0), then the
		221	callback will be invoked regularly at that interval (in fractional
		222	seconds) after the first invocation. If C<interval> is specified with a
		223	false value, then it is treated as if it were missing.
		224
		225	The callback will be rescheduled before invoking the callback, but no
		226	attempt is done to avoid timer drift in most backends, so the interval is
		227	only approximate.
		228
		229	Example: fire an event after 7.7 seconds.
		230
		231	my $w = AnyEvent->timer (after => 7.7, cb => sub {
		232	warn "timeout\n";
		233	});
		234
		235	# to cancel the timer:
		236	undef $w;
		237
		238	Example 2: fire an event after 0.5 seconds, then roughly every second.
		239
		240	my $w = AnyEvent->timer (after => 0.5, interval => 1, cb => sub {
		241	warn "timeout\n";
		242	};
		243
		244	=head3 TIMING ISSUES
		245
		246	There are two ways to handle timers: based on real time (relative, "fire
		247	in 10 seconds") and based on wallclock time (absolute, "fire at 12
		248	o'clock").
		249
		250	While most event loops expect timers to specified in a relative way, they
		251	use absolute time internally. This makes a difference when your clock
		252	"jumps", for example, when ntp decides to set your clock backwards from
		253	the wrong date of 2014-01-01 to 2008-01-01, a watcher that is supposed to
		254	fire "after" a second might actually take six years to finally fire.
		255
		256	AnyEvent cannot compensate for this. The only event loop that is conscious
		257	about these issues is L<EV>, which offers both relative (ev_timer, based
		258	on true relative time) and absolute (ev_periodic, based on wallclock time)
		259	timers.
		260
		261	AnyEvent always prefers relative timers, if available, matching the
		262	AnyEvent API.
		263
		264	AnyEvent has two additional methods that return the "current time":
62		265
63	=over 4	266	=over 4
64		267
		268	=item AnyEvent->time
		269
		270	This returns the "current wallclock time" as a fractional number of
		271	seconds since the Epoch (the same thing as C<time> or C<Time::HiRes::time>
		272	return, and the result is guaranteed to be compatible with those).
		273
		274	It progresses independently of any event loop processing, i.e. each call
		275	will check the system clock, which usually gets updated frequently.
		276
		277	=item AnyEvent->now
		278
		279	This also returns the "current wallclock time", but unlike C<time>, above,
		280	this value might change only once per event loop iteration, depending on
		281	the event loop (most return the same time as C<time>, above). This is the
		282	time that AnyEvent's timers get scheduled against.
		283
		284	I<In almost all cases (in all cases if you don't care), this is the
		285	function to call when you want to know the current time.>
		286
		287	This function is also often faster then C<< AnyEvent->time >>, and
		288	thus the preferred method if you want some timestamp (for example,
		289	L<AnyEvent::Handle> uses this to update it's activity timeouts).
		290
		291	The rest of this section is only of relevance if you try to be very exact
		292	with your timing, you can skip it without bad conscience.
		293
		294	For a practical example of when these times differ, consider L<Event::Lib>
		295	and L<EV> and the following set-up:
		296
		297	The event loop is running and has just invoked one of your callback at
		298	time=500 (assume no other callbacks delay processing). In your callback,
		299	you wait a second by executing C<sleep 1> (blocking the process for a
		300	second) and then (at time=501) you create a relative timer that fires
		301	after three seconds.
		302
		303	With L<Event::Lib>, C<< AnyEvent->time >> and C<< AnyEvent->now >> will
		304	both return C<501>, because that is the current time, and the timer will
		305	be scheduled to fire at time=504 (C<501> + C<3>).
		306
		307	With L<EV>, C<< AnyEvent->time >> returns C<501> (as that is the current
		308	time), but C<< AnyEvent->now >> returns C<500>, as that is the time the
		309	last event processing phase started. With L<EV>, your timer gets scheduled
		310	to run at time=503 (C<500> + C<3>).
		311
		312	In one sense, L<Event::Lib> is more exact, as it uses the current time
		313	regardless of any delays introduced by event processing. However, most
		314	callbacks do not expect large delays in processing, so this causes a
		315	higher drift (and a lot more system calls to get the current time).
		316
		317	In another sense, L<EV> is more exact, as your timer will be scheduled at
		318	the same time, regardless of how long event processing actually took.
		319
		320	In either case, if you care (and in most cases, you don't), then you
		321	can get whatever behaviour you want with any event loop, by taking the
		322	difference between C<< AnyEvent->time >> and C<< AnyEvent->now >> into
		323	account.
		324
		325	=item AnyEvent->now_update
		326
		327	Some event loops (such as L<EV> or L<AnyEvent::Impl::Perl>) cache
		328	the current time for each loop iteration (see the discussion of L<<
		329	AnyEvent->now >>, above).
		330
		331	When a callback runs for a long time (or when the process sleeps), then
		332	this "current" time will differ substantially from the real time, which
		333	might affect timers and time-outs.
		334
		335	When this is the case, you can call this method, which will update the
		336	event loop's idea of "current time".
		337
		338	Note that updating the time I<might> cause some events to be handled.
		339
		340	=back
		341
		342	=head2 SIGNAL WATCHERS
		343
		344	You can watch for signals using a signal watcher, C<signal> is the signal
		345	I<name> in uppercase and without any C<SIG> prefix, C<cb> is the Perl
		346	callback to be invoked whenever a signal occurs.
		347
		348	Although the callback might get passed parameters, their value and
		349	presence is undefined and you cannot rely on them. Portable AnyEvent
		350	callbacks cannot use arguments passed to signal watcher callbacks.
		351
		352	Multiple signal occurrences can be clumped together into one callback
		353	invocation, and callback invocation will be synchronous. Synchronous means
		354	that it might take a while until the signal gets handled by the process,
		355	but it is guaranteed not to interrupt any other callbacks.
		356
		357	The main advantage of using these watchers is that you can share a signal
		358	between multiple watchers.
		359
		360	This watcher might use C<%SIG>, so programs overwriting those signals
		361	directly will likely not work correctly.
		362
		363	Example: exit on SIGINT
		364
		365	my $w = AnyEvent->signal (signal => "INT", cb => sub { exit 1 });
		366
		367	=head2 CHILD PROCESS WATCHERS
		368
		369	You can also watch on a child process exit and catch its exit status.
		370
		371	The child process is specified by the C<pid> argument (if set to C<0>, it
		372	watches for any child process exit). The watcher will triggered only when
		373	the child process has finished and an exit status is available, not on
		374	any trace events (stopped/continued).
		375
		376	The callback will be called with the pid and exit status (as returned by
		377	waitpid), so unlike other watcher types, you I<can> rely on child watcher
		378	callback arguments.
		379
		380	This watcher type works by installing a signal handler for C<SIGCHLD>,
		381	and since it cannot be shared, nothing else should use SIGCHLD or reap
		382	random child processes (waiting for specific child processes, e.g. inside
		383	C<system>, is just fine).
		384
		385	There is a slight catch to child watchers, however: you usually start them
		386	I<after> the child process was created, and this means the process could
		387	have exited already (and no SIGCHLD will be sent anymore).
		388
		389	Not all event models handle this correctly (POE doesn't), but even for
		390	event models that I<do> handle this correctly, they usually need to be
		391	loaded before the process exits (i.e. before you fork in the first place).
		392
		393	This means you cannot create a child watcher as the very first thing in an
		394	AnyEvent program, you I<have> to create at least one watcher before you
		395	C<fork> the child (alternatively, you can call C<AnyEvent::detect>).
		396
		397	Example: fork a process and wait for it
		398
		399	my $done = AnyEvent->condvar;
		400
		401	my $pid = fork or exit 5;
		402
		403	my $w = AnyEvent->child (
		404	pid => $pid,
		405	cb => sub {
		406	my ($pid, $status) = @_;
		407	warn "pid $pid exited with status $status";
		408	$done->send;
		409	},
		410	);
		411
		412	# do something else, then wait for process exit
		413	$done->recv;
		414
		415	=head2 CONDITION VARIABLES
		416
		417	If you are familiar with some event loops you will know that all of them
		418	require you to run some blocking "loop", "run" or similar function that
		419	will actively watch for new events and call your callbacks.
		420
		421	AnyEvent is different, it expects somebody else to run the event loop and
		422	will only block when necessary (usually when told by the user).
		423
		424	The instrument to do that is called a "condition variable", so called
		425	because they represent a condition that must become true.
		426
		427	Condition variables can be created by calling the C<< AnyEvent->condvar
		428	>> method, usually without arguments. The only argument pair allowed is
		429
		430	C<cb>, which specifies a callback to be called when the condition variable
		431	becomes true, with the condition variable as the first argument (but not
		432	the results).
		433
		434	After creation, the condition variable is "false" until it becomes "true"
		435	by calling the C<send> method (or calling the condition variable as if it
		436	were a callback, read about the caveats in the description for the C<<
		437	->send >> method).
		438
		439	Condition variables are similar to callbacks, except that you can
		440	optionally wait for them. They can also be called merge points - points
		441	in time where multiple outstanding events have been processed. And yet
		442	another way to call them is transactions - each condition variable can be
		443	used to represent a transaction, which finishes at some point and delivers
		444	a result.
		445
		446	Condition variables are very useful to signal that something has finished,
		447	for example, if you write a module that does asynchronous http requests,
		448	then a condition variable would be the ideal candidate to signal the
		449	availability of results. The user can either act when the callback is
		450	called or can synchronously C<< ->recv >> for the results.
		451
		452	You can also use them to simulate traditional event loops - for example,
		453	you can block your main program until an event occurs - for example, you
		454	could C<< ->recv >> in your main program until the user clicks the Quit
		455	button of your app, which would C<< ->send >> the "quit" event.
		456
		457	Note that condition variables recurse into the event loop - if you have
		458	two pieces of code that call C<< ->recv >> in a round-robin fashion, you
		459	lose. Therefore, condition variables are good to export to your caller, but
		460	you should avoid making a blocking wait yourself, at least in callbacks,
		461	as this asks for trouble.
		462
		463	Condition variables are represented by hash refs in perl, and the keys
		464	used by AnyEvent itself are all named C<_ae_XXX> to make subclassing
		465	easy (it is often useful to build your own transaction class on top of
		466	AnyEvent). To subclass, use C<AnyEvent::CondVar> as base class and call
		467	it's C<new> method in your own C<new> method.
		468
		469	There are two "sides" to a condition variable - the "producer side" which
		470	eventually calls C<< -> send >>, and the "consumer side", which waits
		471	for the send to occur.
		472
		473	Example: wait for a timer.
		474
		475	# wait till the result is ready
		476	my $result_ready = AnyEvent->condvar;
		477
		478	# do something such as adding a timer
		479	# or socket watcher the calls $result_ready->send
		480	# when the "result" is ready.
		481	# in this case, we simply use a timer:
		482	my $w = AnyEvent->timer (
		483	after => 1,
		484	cb => sub { $result_ready->send },
		485	);
		486
		487	# this "blocks" (while handling events) till the callback
		488	# calls send
		489	$result_ready->recv;
		490
		491	Example: wait for a timer, but take advantage of the fact that
		492	condition variables are also code references.
		493
		494	my $done = AnyEvent->condvar;
		495	my $delay = AnyEvent->timer (after => 5, cb => $done);
		496	$done->recv;
		497
		498	Example: Imagine an API that returns a condvar and doesn't support
		499	callbacks. This is how you make a synchronous call, for example from
		500	the main program:
		501
		502	use AnyEvent::CouchDB;
		503
		504	...
		505
		506	my @info = $couchdb->info->recv;
		507
		508	And this is how you would just ste a callback to be called whenever the
		509	results are available:
		510
		511	$couchdb->info->cb (sub {
		512	my @info = $_[0]->recv;
		513	});
		514
		515	=head3 METHODS FOR PRODUCERS
		516
		517	These methods should only be used by the producing side, i.e. the
		518	code/module that eventually sends the signal. Note that it is also
		519	the producer side which creates the condvar in most cases, but it isn't
		520	uncommon for the consumer to create it as well.
		521
		522	=over 4
		523
		524	=item $cv->send (...)
		525
		526	Flag the condition as ready - a running C<< ->recv >> and all further
		527	calls to C<recv> will (eventually) return after this method has been
		528	called. If nobody is waiting the send will be remembered.
		529
		530	If a callback has been set on the condition variable, it is called
		531	immediately from within send.
		532
		533	Any arguments passed to the C<send> call will be returned by all
		534	future C<< ->recv >> calls.
		535
		536	Condition variables are overloaded so one can call them directly
		537	(as a code reference). Calling them directly is the same as calling
		538	C<send>. Note, however, that many C-based event loops do not handle
		539	overloading, so as tempting as it may be, passing a condition variable
		540	instead of a callback does not work. Both the pure perl and EV loops
		541	support overloading, however, as well as all functions that use perl to
		542	invoke a callback (as in L<AnyEvent::Socket> and L<AnyEvent::DNS> for
		543	example).
		544
		545	=item $cv->croak ($error)
		546
		547	Similar to send, but causes all call's to C<< ->recv >> to invoke
		548	C<Carp::croak> with the given error message/object/scalar.
		549
		550	This can be used to signal any errors to the condition variable
		551	user/consumer.
		552
		553	=item $cv->begin ([group callback])
		554
		555	=item $cv->end
		556
		557	These two methods are EXPERIMENTAL and MIGHT CHANGE.
		558
		559	These two methods can be used to combine many transactions/events into
		560	one. For example, a function that pings many hosts in parallel might want
		561	to use a condition variable for the whole process.
		562
		563	Every call to C<< ->begin >> will increment a counter, and every call to
		564	C<< ->end >> will decrement it. If the counter reaches C<0> in C<< ->end
		565	>>, the (last) callback passed to C<begin> will be executed. That callback
		566	is I<supposed> to call C<< ->send >>, but that is not required. If no
		567	callback was set, C<send> will be called without any arguments.
		568
		569	Let's clarify this with the ping example:
		570
		571	my $cv = AnyEvent->condvar;
		572
		573	my %result;
		574	$cv->begin (sub { $cv->send (\%result) });
		575
		576	for my $host (@list_of_hosts) {
		577	$cv->begin;
		578	ping_host_then_call_callback $host, sub {
		579	$result{$host} = ...;
		580	$cv->end;
		581	};
		582	}
		583
		584	$cv->end;
		585
		586	This code fragment supposedly pings a number of hosts and calls
		587	C<send> after results for all then have have been gathered - in any
		588	order. To achieve this, the code issues a call to C<begin> when it starts
		589	each ping request and calls C<end> when it has received some result for
		590	it. Since C<begin> and C<end> only maintain a counter, the order in which
		591	results arrive is not relevant.
		592
		593	There is an additional bracketing call to C<begin> and C<end> outside the
		594	loop, which serves two important purposes: first, it sets the callback
		595	to be called once the counter reaches C<0>, and second, it ensures that
		596	C<send> is called even when C<no> hosts are being pinged (the loop
		597	doesn't execute once).
		598
		599	This is the general pattern when you "fan out" into multiple subrequests:
		600	use an outer C<begin>/C<end> pair to set the callback and ensure C<end>
		601	is called at least once, and then, for each subrequest you start, call
		602	C<begin> and for each subrequest you finish, call C<end>.
		603
		604	=back
		605
		606	=head3 METHODS FOR CONSUMERS
		607
		608	These methods should only be used by the consuming side, i.e. the
		609	code awaits the condition.
		610
		611	=over 4
		612
		613	=item $cv->recv
		614
		615	Wait (blocking if necessary) until the C<< ->send >> or C<< ->croak
		616	>> methods have been called on c<$cv>, while servicing other watchers
		617	normally.
		618
		619	You can only wait once on a condition - additional calls are valid but
		620	will return immediately.
		621
		622	If an error condition has been set by calling C<< ->croak >>, then this
		623	function will call C<croak>.
		624
		625	In list context, all parameters passed to C<send> will be returned,
		626	in scalar context only the first one will be returned.
		627
		628	Not all event models support a blocking wait - some die in that case
		629	(programs might want to do that to stay interactive), so I<if you are
		630	using this from a module, never require a blocking wait>, but let the
		631	caller decide whether the call will block or not (for example, by coupling
		632	condition variables with some kind of request results and supporting
		633	callbacks so the caller knows that getting the result will not block,
		634	while still supporting blocking waits if the caller so desires).
		635
		636	Another reason I<never> to C<< ->recv >> in a module is that you cannot
		637	sensibly have two C<< ->recv >>'s in parallel, as that would require
		638	multiple interpreters or coroutines/threads, none of which C<AnyEvent>
		639	can supply.
		640
		641	The L<Coro> module, however, I<can> and I<does> supply coroutines and, in
		642	fact, L<Coro::AnyEvent> replaces AnyEvent's condvars by coroutine-safe
		643	versions and also integrates coroutines into AnyEvent, making blocking
		644	C<< ->recv >> calls perfectly safe as long as they are done from another
		645	coroutine (one that doesn't run the event loop).
		646
		647	You can ensure that C<< -recv >> never blocks by setting a callback and
		648	only calling C<< ->recv >> from within that callback (or at a later
		649	time). This will work even when the event loop does not support blocking
		650	waits otherwise.
		651
		652	=item $bool = $cv->ready
		653
		654	Returns true when the condition is "true", i.e. whether C<send> or
		655	C<croak> have been called.
		656
		657	=item $cb = $cv->cb ($cb->($cv))
		658
		659	This is a mutator function that returns the callback set and optionally
		660	replaces it before doing so.
		661
		662	The callback will be called when the condition becomes "true", i.e. when
		663	C<send> or C<croak> are called, with the only argument being the condition
		664	variable itself. Calling C<recv> inside the callback or at any later time
		665	is guaranteed not to block.
		666
		667	=back
		668
		669	=head1 GLOBAL VARIABLES AND FUNCTIONS
		670
		671	=over 4
		672
		673	=item $AnyEvent::MODEL
		674
		675	Contains C<undef> until the first watcher is being created. Then it
		676	contains the event model that is being used, which is the name of the
		677	Perl class implementing the model. This class is usually one of the
		678	C<AnyEvent::Impl:xxx> modules, but can be any other class in the case
		679	AnyEvent has been extended at runtime (e.g. in I<rxvt-unicode>).
		680
		681	The known classes so far are:
		682
		683	AnyEvent::Impl::EV based on EV (an interface to libev, best choice).
		684	AnyEvent::Impl::Event based on Event, second best choice.
		685	AnyEvent::Impl::Perl pure-perl implementation, fast and portable.
		686	AnyEvent::Impl::Glib based on Glib, third-best choice.
		687	AnyEvent::Impl::Tk based on Tk, very bad choice.
		688	AnyEvent::Impl::Qt based on Qt, cannot be autoprobed (see its docs).
		689	AnyEvent::Impl::EventLib based on Event::Lib, leaks memory and worse.
		690	AnyEvent::Impl::POE based on POE, not generic enough for full support.
		691
		692	There is no support for WxWidgets, as WxWidgets has no support for
		693	watching file handles. However, you can use WxWidgets through the
		694	POE Adaptor, as POE has a Wx backend that simply polls 20 times per
		695	second, which was considered to be too horrible to even consider for
		696	AnyEvent. Likewise, other POE backends can be used by AnyEvent by using
		697	it's adaptor.
		698
		699	AnyEvent knows about L<Prima> and L<Wx> and will try to use L<POE> when
		700	autodetecting them.
		701
		702	=item AnyEvent::detect
		703
		704	Returns C<$AnyEvent::MODEL>, forcing autodetection of the event model
		705	if necessary. You should only call this function right before you would
		706	have created an AnyEvent watcher anyway, that is, as late as possible at
		707	runtime.
		708
		709	=item $guard = AnyEvent::post_detect { BLOCK }
		710
		711	Arranges for the code block to be executed as soon as the event model is
		712	autodetected (or immediately if this has already happened).
		713
		714	If called in scalar or list context, then it creates and returns an object
		715	that automatically removes the callback again when it is destroyed. See
		716	L<Coro::BDB> for a case where this is useful.
		717
		718	=item @AnyEvent::post_detect
		719
		720	If there are any code references in this array (you can C<push> to it
		721	before or after loading AnyEvent), then they will called directly after
		722	the event loop has been chosen.
		723
		724	You should check C<$AnyEvent::MODEL> before adding to this array, though:
		725	if it contains a true value then the event loop has already been detected,
		726	and the array will be ignored.
		727
		728	Best use C<AnyEvent::post_detect { BLOCK }> instead.
		729
		730	=back
		731
		732	=head1 WHAT TO DO IN A MODULE
		733
		734	As a module author, you should C<use AnyEvent> and call AnyEvent methods
		735	freely, but you should not load a specific event module or rely on it.
		736
		737	Be careful when you create watchers in the module body - AnyEvent will
		738	decide which event module to use as soon as the first method is called, so
		739	by calling AnyEvent in your module body you force the user of your module
		740	to load the event module first.
		741
		742	Never call C<< ->recv >> on a condition variable unless you I<know> that
		743	the C<< ->send >> method has been called on it already. This is
		744	because it will stall the whole program, and the whole point of using
		745	events is to stay interactive.
		746
		747	It is fine, however, to call C<< ->recv >> when the user of your module
		748	requests it (i.e. if you create a http request object ad have a method
		749	called C<results> that returns the results, it should call C<< ->recv >>
		750	freely, as the user of your module knows what she is doing. always).
		751
		752	=head1 WHAT TO DO IN THE MAIN PROGRAM
		753
		754	There will always be a single main program - the only place that should
		755	dictate which event model to use.
		756
		757	If it doesn't care, it can just "use AnyEvent" and use it itself, or not
		758	do anything special (it does not need to be event-based) and let AnyEvent
		759	decide which implementation to chose if some module relies on it.
		760
		761	If the main program relies on a specific event model - for example, in
		762	Gtk2 programs you have to rely on the Glib module - you should load the
		763	event module before loading AnyEvent or any module that uses it: generally
		764	speaking, you should load it as early as possible. The reason is that
		765	modules might create watchers when they are loaded, and AnyEvent will
		766	decide on the event model to use as soon as it creates watchers, and it
		767	might chose the wrong one unless you load the correct one yourself.
		768
		769	You can chose to use a pure-perl implementation by loading the
		770	C<AnyEvent::Impl::Perl> module, which gives you similar behaviour
		771	everywhere, but letting AnyEvent chose the model is generally better.
		772
		773	=head2 MAINLOOP EMULATION
		774
		775	Sometimes (often for short test scripts, or even standalone programs who
		776	only want to use AnyEvent), you do not want to run a specific event loop.
		777
		778	In that case, you can use a condition variable like this:
		779
		780	AnyEvent->condvar->recv;
		781
		782	This has the effect of entering the event loop and looping forever.
		783
		784	Note that usually your program has some exit condition, in which case
		785	it is better to use the "traditional" approach of storing a condition
		786	variable somewhere, waiting for it, and sending it when the program should
		787	exit cleanly.
		788
		789
		790	=head1 OTHER MODULES
		791
		792	The following is a non-exhaustive list of additional modules that use
		793	AnyEvent and can therefore be mixed easily with other AnyEvent modules
		794	in the same program. Some of the modules come with AnyEvent, some are
		795	available via CPAN.
		796
		797	=over 4
		798
		799	=item L<AnyEvent::Util>
		800
		801	Contains various utility functions that replace often-used but blocking
		802	functions such as C<inet_aton> by event-/callback-based versions.
		803
		804	=item L<AnyEvent::Socket>
		805
		806	Provides various utility functions for (internet protocol) sockets,
		807	addresses and name resolution. Also functions to create non-blocking tcp
		808	connections or tcp servers, with IPv6 and SRV record support and more.
		809
		810	=item L<AnyEvent::Handle>
		811
		812	Provide read and write buffers, manages watchers for reads and writes,
		813	supports raw and formatted I/O, I/O queued and fully transparent and
		814	non-blocking SSL/TLS.
		815
		816	=item L<AnyEvent::DNS>
		817
		818	Provides rich asynchronous DNS resolver capabilities.
		819
		820	=item L<AnyEvent::HTTP>
		821
		822	A simple-to-use HTTP library that is capable of making a lot of concurrent
		823	HTTP requests.
		824
		825	=item L<AnyEvent::HTTPD>
		826
		827	Provides a simple web application server framework.
		828
		829	=item L<AnyEvent::FastPing>
		830
		831	The fastest ping in the west.
		832
		833	=item L<AnyEvent::DBI>
		834
		835	Executes L<DBI> requests asynchronously in a proxy process.
		836
		837	=item L<AnyEvent::AIO>
		838
		839	Truly asynchronous I/O, should be in the toolbox of every event
		840	programmer. AnyEvent::AIO transparently fuses L<IO::AIO> and AnyEvent
		841	together.
		842
		843	=item L<AnyEvent::BDB>
		844
		845	Truly asynchronous Berkeley DB access. AnyEvent::BDB transparently fuses
		846	L<BDB> and AnyEvent together.
		847
		848	=item L<AnyEvent::GPSD>
		849
		850	A non-blocking interface to gpsd, a daemon delivering GPS information.
		851
		852	=item L<AnyEvent::IGS>
		853
		854	A non-blocking interface to the Internet Go Server protocol (used by
		855	L<App::IGS>).
		856
		857	=item L<AnyEvent::IRC>
		858
		859	AnyEvent based IRC client module family (replacing the older Net::IRC3).
		860
		861	=item L<Net::XMPP2>
		862
		863	AnyEvent based XMPP (Jabber protocol) module family.
		864
		865	=item L<Net::FCP>
		866
		867	AnyEvent-based implementation of the Freenet Client Protocol, birthplace
		868	of AnyEvent.
		869
		870	=item L<Event::ExecFlow>
		871
		872	High level API for event-based execution flow control.
		873
		874	=item L<Coro>
		875
		876	Has special support for AnyEvent via L<Coro::AnyEvent>.
		877
		878	=item L<IO::Lambda>
		879
		880	The lambda approach to I/O - don't ask, look there. Can use AnyEvent.
		881
		882	=back
		883
65	=cut	884	=cut
66		885
67	package AnyEvent;	886	package AnyEvent;
68		887
69	no warnings;	888	no warnings;
70	use strict 'vars';	889	use strict qw(vars subs);
		890
71	use Carp;	891	use Carp;
72		892
73	our $VERSION = '1.02';	893	our $VERSION = 4.352;
74	our $MODEL;	894	our $MODEL;
75		895
76	our $AUTOLOAD;	896	our $AUTOLOAD;
77	our @ISA;	897	our @ISA;
78		898
		899	our @REGISTRY;
		900
		901	our $WIN32;
		902
		903	BEGIN {
		904	my $win32 = ! ! ($^O =~ /mswin32/i);
		905	eval "sub WIN32(){ $win32 }";
		906	}
		907
79	our $verbose = $ENV{PERL_ANYEVENT_VERBOSE}*1;	908	our $verbose = $ENV{PERL_ANYEVENT_VERBOSE}*1;
80		909
81	our @REGISTRY;	910	our %PROTOCOL; # (ipv4|ipv6) => (1|2), higher numbers are preferred
		911
		912	{
		913	my $idx;
		914	$PROTOCOL{$_} = ++$idx
		915	for reverse split /\s*,\s*/,
		916	$ENV{PERL_ANYEVENT_PROTOCOLS} || "ipv4,ipv6";
		917	}
82		918
83	my @models = (	919	my @models = (
84	[Coro::Event:: => AnyEvent::Impl::Coro::],	920	[EV:: => AnyEvent::Impl::EV::],
85	[Event:: => AnyEvent::Impl::Event::],	921	[Event:: => AnyEvent::Impl::Event::],
86	[Glib:: => AnyEvent::Impl::Glib::],	922	[AnyEvent::Impl::Perl:: => AnyEvent::Impl::Perl::],
		923	# everything below here will not be autoprobed
		924	# as the pureperl backend should work everywhere
		925	# and is usually faster
		926	[Tk:: => AnyEvent::Impl::Tk::], # crashes with many handles
		927	[Glib:: => AnyEvent::Impl::Glib::], # becomes extremely slow with many watchers
		928	[Event::Lib:: => AnyEvent::Impl::EventLib::], # too buggy
		929	[Qt:: => AnyEvent::Impl::Qt::], # requires special main program
		930	[POE::Kernel:: => AnyEvent::Impl::POE::], # lasciate ogni speranza
87	[Tk:: => AnyEvent::Impl::Tk::],	931	[Wx:: => AnyEvent::Impl::POE::],
		932	[Prima:: => AnyEvent::Impl::POE::],
88	);	933	);
89		934
90	our %method = map +($_ => 1), qw(io timer condvar broadcast wait cancel DESTROY);	935	our %method = map +($_ => 1),
		936	qw(io timer time now now_update signal child condvar one_event DESTROY);
91		937
92	sub AUTOLOAD {	938	our @post_detect;
93	$AUTOLOAD =~ s/.*://;
94		939
95	$method{$AUTOLOAD}	940	sub post_detect(&) {
96	or croak "$AUTOLOAD: not a valid method for AnyEvent objects";	941	my ($cb) = @_;
97		942
		943	if ($MODEL) {
		944	$cb->();
		945
		946	1
		947	} else {
		948	push @post_detect, $cb;
		949
		950	defined wantarray
		951	? bless \$cb, "AnyEvent::Util::PostDetect"
		952	: ()
		953	}
		954	}
		955
		956	sub AnyEvent::Util::PostDetect::DESTROY {
		957	@post_detect = grep $_ != ${$_[0]}, @post_detect;
		958	}
		959
		960	sub detect() {
98	unless ($MODEL) {	961	unless ($MODEL) {
		962	no strict 'refs';
		963	local $SIG{__DIE__};
		964
		965	if ($ENV{PERL_ANYEVENT_MODEL} =~ /^([a-zA-Z]+)$/) {
		966	my $model = "AnyEvent::Impl::$1";
		967	if (eval "require $model") {
		968	$MODEL = $model;
		969	warn "AnyEvent: loaded model '$model' (forced by \$PERL_ANYEVENT_MODEL), using it.\n" if $verbose > 1;
		970	} else {
		971	warn "AnyEvent: unable to load model '$model' (from \$PERL_ANYEVENT_MODEL):\n$@" if $verbose;
		972	}
		973	}
		974
99	# check for already loaded models	975	# check for already loaded models
		976	unless ($MODEL) {
100	for (@REGISTRY, @models) {	977	for (@REGISTRY, @models) {
101	my ($package, $model) = @$_;	978	my ($package, $model) = @$_;
102	if (${"$package\::VERSION"} > 0) {	979	if (${"$package\::VERSION"} > 0) {
103	if (eval "require $model") {	980	if (eval "require $model") {
104	$MODEL = $model;	981	$MODEL = $model;
105	warn "AnyEvent: found model '$model', using it.\n" if $verbose > 1;	982	warn "AnyEvent: autodetected model '$model', using it.\n" if $verbose > 1;
106	last;	983	last;
		984	}
107	}	985	}
108	}	986	}
		987
		988	unless ($MODEL) {
		989	# try to load a model
		990
		991	for (@REGISTRY, @models) {
		992	my ($package, $model) = @$_;
		993	if (eval "require $package"
		994	and ${"$package\::VERSION"} > 0
		995	and eval "require $model") {
		996	$MODEL = $model;
		997	warn "AnyEvent: autoprobed model '$model', using it.\n" if $verbose > 1;
		998	last;
		999	}
		1000	}
		1001
		1002	$MODEL
		1003	or die "No event module selected for AnyEvent and autodetect failed. Install any one of these modules: EV, Event or Glib.\n";
		1004	}
109	}	1005	}
110		1006
111	unless ($MODEL) {	1007	push @{"$MODEL\::ISA"}, "AnyEvent::Base";
112	# try to load a model
113		1008
114	for (@REGISTRY, @models) {	1009	unshift @ISA, $MODEL;
115	my ($package, $model) = @$_;
116	if (eval "require $model") {
117	$MODEL = $model;
118	warn "AnyEvent: autoprobed and loaded model '$model', using it.\n" if $verbose > 1;
119	last;
120	}
121	}
122		1010
		1011	require AnyEvent::Strict if $ENV{PERL_ANYEVENT_STRICT};
		1012
		1013	(shift @post_detect)->() while @post_detect;
		1014	}
		1015
123	$MODEL	1016	$MODEL
124	or die "No event module selected for AnyEvent and autodetect failed. Install any one of these modules: Coro, Event, Glib or Tk.";	1017	}
		1018
		1019	sub AUTOLOAD {
		1020	(my $func = $AUTOLOAD) =~ s/.*://;
		1021
		1022	$method{$func}
		1023	or croak "$func: not a valid method for AnyEvent objects";
		1024
		1025	detect unless $MODEL;
		1026
		1027	my $class = shift;
		1028	$class->$func (@_);
		1029	}
		1030
		1031	# utility function to dup a filehandle. this is used by many backends
		1032	# to support binding more than one watcher per filehandle (they usually
		1033	# allow only one watcher per fd, so we dup it to get a different one).
		1034	sub _dupfh($$$$) {
		1035	my ($poll, $fh, $r, $w) = @_;
		1036
		1037	# cygwin requires the fh mode to be matching, unix doesn't
		1038	my ($rw, $mode) = $poll eq "r" ? ($r, "<")
		1039	: $poll eq "w" ? ($w, ">")
		1040	: Carp::croak "AnyEvent->io requires poll set to either 'r' or 'w'";
		1041
		1042	open my $fh2, "$mode&" . fileno $fh
		1043	or die "cannot dup() filehandle: $!,";
		1044
		1045	# we assume CLOEXEC is already set by perl in all important cases
		1046
		1047	($fh2, $rw)
		1048	}
		1049
		1050	package AnyEvent::Base;
		1051
		1052	# default implementations for many methods
		1053
		1054	BEGIN {
		1055	if (eval "use Time::HiRes (); time (); 1") {
		1056	*_time = \&Time::HiRes::time;
		1057	# if (eval "use POSIX (); (POSIX::times())...
		1058	} else {
		1059	*_time = sub { time }; # epic fail
		1060	}
		1061	}
		1062
		1063	sub time { _time }
		1064	sub now { _time }
		1065	sub now_update { }
		1066
		1067	# default implementation for ->condvar
		1068
		1069	sub condvar {
		1070	bless { @_ == 3 ? (_ae_cb => $_[2]) : () }, AnyEvent::CondVar::
		1071	}
		1072
		1073	# default implementation for ->signal
		1074
		1075	our ($SIGPIPE_R, $SIGPIPE_W, %SIG_CB, %SIG_EV, $SIG_IO);
		1076
		1077	sub _signal_exec {
		1078	sysread $SIGPIPE_R, my $dummy, 4;
		1079
		1080	while (%SIG_EV) {
		1081	for (keys %SIG_EV) {
		1082	delete $SIG_EV{$_};
		1083	$_->() for values %{ $SIG_CB{$_} || {} };
125	}	1084	}
126	}	1085	}
		1086	}
127		1087
128	@ISA = $MODEL;	1088	sub signal {
		1089	my (undef, %arg) = @_;
129		1090
		1091	unless ($SIGPIPE_R) {
		1092	require Fcntl;
		1093
		1094	if (AnyEvent::WIN32) {
		1095	require AnyEvent::Util;
		1096
		1097	($SIGPIPE_R, $SIGPIPE_W) = AnyEvent::Util::portable_pipe ();
		1098	AnyEvent::Util::fh_nonblocking ($SIGPIPE_R) if $SIGPIPE_R;
		1099	AnyEvent::Util::fh_nonblocking ($SIGPIPE_W) if $SIGPIPE_W; # just in case
		1100	} else {
		1101	pipe $SIGPIPE_R, $SIGPIPE_W;
		1102	fcntl $SIGPIPE_R, &Fcntl::F_SETFL, &Fcntl::O_NONBLOCK if $SIGPIPE_R;
		1103	fcntl $SIGPIPE_W, &Fcntl::F_SETFL, &Fcntl::O_NONBLOCK if $SIGPIPE_W; # just in case
		1104	}
		1105
		1106	$SIGPIPE_R
		1107	or Carp::croak "AnyEvent: unable to create a signal reporting pipe: $!\n";
		1108
		1109	# not strictly required, as $^F is normally 2, but let's make sure...
		1110	fcntl $SIGPIPE_R, &Fcntl::F_SETFD, &Fcntl::FD_CLOEXEC;
		1111	fcntl $SIGPIPE_W, &Fcntl::F_SETFD, &Fcntl::FD_CLOEXEC;
		1112
		1113	$SIG_IO = AnyEvent->io (fh => $SIGPIPE_R, poll => "r", cb => \&_signal_exec);
		1114	}
		1115
		1116	my $signal = uc $arg{signal}
		1117	or Carp::croak "required option 'signal' is missing";
		1118
		1119	$SIG_CB{$signal}{$arg{cb}} = $arg{cb};
		1120	$SIG{$signal} ||= sub {
		1121	local $!;
		1122	syswrite $SIGPIPE_W, "\x00", 1 unless %SIG_EV;
		1123	undef $SIG_EV{$signal};
		1124	};
		1125
		1126	bless [$signal, $arg{cb}], "AnyEvent::Base::Signal"
		1127	}
		1128
		1129	sub AnyEvent::Base::Signal::DESTROY {
		1130	my ($signal, $cb) = @{$_[0]};
		1131
		1132	delete $SIG_CB{$signal}{$cb};
		1133
		1134	delete $SIG{$signal} unless keys %{ $SIG_CB{$signal} };
		1135	}
		1136
		1137	# default implementation for ->child
		1138
		1139	our %PID_CB;
		1140	our $CHLD_W;
		1141	our $CHLD_DELAY_W;
		1142	our $PID_IDLE;
		1143	our $WNOHANG;
		1144
		1145	sub _child_wait {
		1146	while (0 < (my $pid = waitpid -1, $WNOHANG)) {
		1147	$_->($pid, $?) for (values %{ $PID_CB{$pid} || {} }),
		1148	(values %{ $PID_CB{0} || {} });
		1149	}
		1150
		1151	undef $PID_IDLE;
		1152	}
		1153
		1154	sub _sigchld {
		1155	# make sure we deliver these changes "synchronous" with the event loop.
		1156	$CHLD_DELAY_W ||= AnyEvent->timer (after => 0, cb => sub {
		1157	undef $CHLD_DELAY_W;
		1158	&_child_wait;
		1159	});
		1160	}
		1161
		1162	sub child {
		1163	my (undef, %arg) = @_;
		1164
		1165	defined (my $pid = $arg{pid} + 0)
		1166	or Carp::croak "required option 'pid' is missing";
		1167
		1168	$PID_CB{$pid}{$arg{cb}} = $arg{cb};
		1169
		1170	unless ($WNOHANG) {
		1171	$WNOHANG = eval { local $SIG{__DIE__}; require POSIX; &POSIX::WNOHANG } || 1;
		1172	}
		1173
		1174	unless ($CHLD_W) {
		1175	$CHLD_W = AnyEvent->signal (signal => 'CHLD', cb => \&_sigchld);
		1176	# child could be a zombie already, so make at least one round
		1177	&_sigchld;
		1178	}
		1179
		1180	bless [$pid, $arg{cb}], "AnyEvent::Base::Child"
		1181	}
		1182
		1183	sub AnyEvent::Base::Child::DESTROY {
		1184	my ($pid, $cb) = @{$_[0]};
		1185
		1186	delete $PID_CB{$pid}{$cb};
		1187	delete $PID_CB{$pid} unless keys %{ $PID_CB{$pid} };
		1188
		1189	undef $CHLD_W unless keys %PID_CB;
		1190	}
		1191
		1192	package AnyEvent::CondVar;
		1193
		1194	our @ISA = AnyEvent::CondVar::Base::;
		1195
		1196	package AnyEvent::CondVar::Base;
		1197
		1198	use overload
		1199	'&{}' => sub { my $self = shift; sub { $self->send (@_) } },
		1200	fallback => 1;
		1201
		1202	sub _send {
		1203	# nop
		1204	}
		1205
		1206	sub send {
130	my $class = shift;	1207	my $cv = shift;
131	$class->$AUTOLOAD (@_);	1208	$cv->{_ae_sent} = [@_];
		1209	(delete $cv->{_ae_cb})->($cv) if $cv->{_ae_cb};
		1210	$cv->_send;
132	}	1211	}
		1212
		1213	sub croak {
		1214	$_[0]{_ae_croak} = $_[1];
		1215	$_[0]->send;
		1216	}
		1217
		1218	sub ready {
		1219	$_[0]{_ae_sent}
		1220	}
		1221
		1222	sub _wait {
		1223	AnyEvent->one_event while !$_[0]{_ae_sent};
		1224	}
		1225
		1226	sub recv {
		1227	$_[0]->_wait;
		1228
		1229	Carp::croak $_[0]{_ae_croak} if $_[0]{_ae_croak};
		1230	wantarray ? @{ $_[0]{_ae_sent} } : $_[0]{_ae_sent}[0]
		1231	}
		1232
		1233	sub cb {
		1234	$_[0]{_ae_cb} = $_[1] if @_ > 1;
		1235	$_[0]{_ae_cb}
		1236	}
		1237
		1238	sub begin {
		1239	++$_[0]{_ae_counter};
		1240	$_[0]{_ae_end_cb} = $_[1] if @_ > 1;
		1241	}
		1242
		1243	sub end {
		1244	return if --$_[0]{_ae_counter};
		1245	&{ $_[0]{_ae_end_cb} || sub { $_[0]->send } };
		1246	}
		1247
		1248	# undocumented/compatibility with pre-3.4
		1249	*broadcast = \&send;
		1250	*wait = \&_wait;
		1251
		1252	=head1 ERROR AND EXCEPTION HANDLING
		1253
		1254	In general, AnyEvent does not do any error handling - it relies on the
		1255	caller to do that if required. The L<AnyEvent::Strict> module (see also
		1256	the C<PERL_ANYEVENT_STRICT> environment variable, below) provides strict
		1257	checking of all AnyEvent methods, however, which is highly useful during
		1258	development.
		1259
		1260	As for exception handling (i.e. runtime errors and exceptions thrown while
		1261	executing a callback), this is not only highly event-loop specific, but
		1262	also not in any way wrapped by this module, as this is the job of the main
		1263	program.
		1264
		1265	The pure perl event loop simply re-throws the exception (usually
		1266	within C<< condvar->recv >>), the L<Event> and L<EV> modules call C<<
		1267	$Event/EV::DIED->() >>, L<Glib> uses C<< install_exception_handler >> and
		1268	so on.
		1269
		1270	=head1 ENVIRONMENT VARIABLES
		1271
		1272	The following environment variables are used by this module or its
		1273	submodules:
		1274
		1275	=over 4
		1276
		1277	=item C<PERL_ANYEVENT_VERBOSE>
		1278
		1279	By default, AnyEvent will be completely silent except in fatal
		1280	conditions. You can set this environment variable to make AnyEvent more
		1281	talkative.
		1282
		1283	When set to C<1> or higher, causes AnyEvent to warn about unexpected
		1284	conditions, such as not being able to load the event model specified by
		1285	C<PERL_ANYEVENT_MODEL>.
		1286
		1287	When set to C<2> or higher, cause AnyEvent to report to STDERR which event
		1288	model it chooses.
		1289
		1290	=item C<PERL_ANYEVENT_STRICT>
		1291
		1292	AnyEvent does not do much argument checking by default, as thorough
		1293	argument checking is very costly. Setting this variable to a true value
		1294	will cause AnyEvent to load C<AnyEvent::Strict> and then to thoroughly
		1295	check the arguments passed to most method calls. If it finds any problems
		1296	it will croak.
		1297
		1298	In other words, enables "strict" mode.
		1299
		1300	Unlike C<use strict>, it is definitely recommended ot keep it off in
		1301	production. Keeping C<PERL_ANYEVENT_STRICT=1> in your environment while
		1302	developing programs can be very useful, however.
		1303
		1304	=item C<PERL_ANYEVENT_MODEL>
		1305
		1306	This can be used to specify the event model to be used by AnyEvent, before
		1307	auto detection and -probing kicks in. It must be a string consisting
		1308	entirely of ASCII letters. The string C<AnyEvent::Impl::> gets prepended
		1309	and the resulting module name is loaded and if the load was successful,
		1310	used as event model. If it fails to load AnyEvent will proceed with
		1311	auto detection and -probing.
		1312
		1313	This functionality might change in future versions.
		1314
		1315	For example, to force the pure perl model (L<AnyEvent::Impl::Perl>) you
		1316	could start your program like this:
		1317
		1318	PERL_ANYEVENT_MODEL=Perl perl ...
		1319
		1320	=item C<PERL_ANYEVENT_PROTOCOLS>
		1321
		1322	Used by both L<AnyEvent::DNS> and L<AnyEvent::Socket> to determine preferences
		1323	for IPv4 or IPv6. The default is unspecified (and might change, or be the result
		1324	of auto probing).
		1325
		1326	Must be set to a comma-separated list of protocols or address families,
		1327	current supported: C<ipv4> and C<ipv6>. Only protocols mentioned will be
		1328	used, and preference will be given to protocols mentioned earlier in the
		1329	list.
		1330
		1331	This variable can effectively be used for denial-of-service attacks
		1332	against local programs (e.g. when setuid), although the impact is likely
		1333	small, as the program has to handle conenction and other failures anyways.
		1334
		1335	Examples: C<PERL_ANYEVENT_PROTOCOLS=ipv4,ipv6> - prefer IPv4 over IPv6,
		1336	but support both and try to use both. C<PERL_ANYEVENT_PROTOCOLS=ipv4>
		1337	- only support IPv4, never try to resolve or contact IPv6
		1338	addresses. C<PERL_ANYEVENT_PROTOCOLS=ipv6,ipv4> support either IPv4 or
		1339	IPv6, but prefer IPv6 over IPv4.
		1340
		1341	=item C<PERL_ANYEVENT_EDNS0>
		1342
		1343	Used by L<AnyEvent::DNS> to decide whether to use the EDNS0 extension
		1344	for DNS. This extension is generally useful to reduce DNS traffic, but
		1345	some (broken) firewalls drop such DNS packets, which is why it is off by
		1346	default.
		1347
		1348	Setting this variable to C<1> will cause L<AnyEvent::DNS> to announce
		1349	EDNS0 in its DNS requests.
		1350
		1351	=item C<PERL_ANYEVENT_MAX_FORKS>
		1352
		1353	The maximum number of child processes that C<AnyEvent::Util::fork_call>
		1354	will create in parallel.
133		1355
134	=back	1356	=back
135		1357
136	=head1 SUPPLYING YOUR OWN EVENT MODEL INTERFACE	1358	=head1 SUPPLYING YOUR OWN EVENT MODEL INTERFACE
		1359
		1360	This is an advanced topic that you do not normally need to use AnyEvent in
		1361	a module. This section is only of use to event loop authors who want to
		1362	provide AnyEvent compatibility.
137		1363
138	If you need to support another event library which isn't directly	1364	If you need to support another event library which isn't directly
139	supported by AnyEvent, you can supply your own interface to it by	1365	supported by AnyEvent, you can supply your own interface to it by
140	pushing, before the first watch gets created, the package name of	1366	pushing, before the first watcher gets created, the package name of
141	the event module and the package name of the interface to use onto	1367	the event module and the package name of the interface to use onto
142	C<@AnyEvent::REGISTRY>. You can do that before and even without loading	1368	C<@AnyEvent::REGISTRY>. You can do that before and even without loading
143	AnyEvent.	1369	AnyEvent, so it is reasonably cheap.
144		1370
145	Example:	1371	Example:
146		1372
147	push @AnyEvent::REGISTRY, [urxvt => urxvt::anyevent::];	1373	push @AnyEvent::REGISTRY, [urxvt => urxvt::anyevent::];
148		1374
149	This tells AnyEvent to (literally) use the C<urxvt::anyevent::> module	1375	This tells AnyEvent to (literally) use the C<urxvt::anyevent::>
150	when it finds the C<urxvt> module is loaded. When AnyEvent is loaded and	1376	package/class when it finds the C<urxvt> package/module is already loaded.
151	requested to find a suitable event model, it will first check for the
152	urxvt module.
153		1377
		1378	When AnyEvent is loaded and asked to find a suitable event model, it
		1379	will first check for the presence of urxvt by trying to C<use> the
		1380	C<urxvt::anyevent> module.
		1381
		1382	The class should provide implementations for all watcher types. See
		1383	L<AnyEvent::Impl::EV> (source code), L<AnyEvent::Impl::Glib> (Source code)
		1384	and so on for actual examples. Use C<perldoc -m AnyEvent::Impl::Glib> to
		1385	see the sources.
		1386
		1387	If you don't provide C<signal> and C<child> watchers than AnyEvent will
		1388	provide suitable (hopefully) replacements.
		1389
154	The above isn't fictitious, the I<rxvt-unicode> (a.k.a. urxvt) uses	1390	The above example isn't fictitious, the I<rxvt-unicode> (a.k.a. urxvt)
155	the above line exactly. An interface isn't included in AnyEvent	1391	terminal emulator uses the above line as-is. An interface isn't included
156	because it doesn't make sense outside the embedded interpreter inside	1392	in AnyEvent because it doesn't make sense outside the embedded interpreter
157	I<rxvt-unicode>, and it is updated and maintained as part of the	1393	inside I<rxvt-unicode>, and it is updated and maintained as part of the
158	I<rxvt-unicode> distribution.	1394	I<rxvt-unicode> distribution.
159		1395
160	=head1 ENVIRONMENT VARIABLES	1396	I<rxvt-unicode> also cheats a bit by not providing blocking access to
		1397	condition variables: code blocking while waiting for a condition will
		1398	C<die>. This still works with most modules/usages, and blocking calls must
		1399	not be done in an interactive application, so it makes sense.
161		1400
162	The following environment variables are used by this module:
163
164	C<PERL_ANYEVENT_VERBOSE> when set to C<2> or higher, reports which event
165	model gets used.
166
167	=head1 EXAMPLE	1401	=head1 EXAMPLE PROGRAM
168		1402
169	The following program uses an io watcher to read data from stdin, a timer	1403	The following program uses an I/O watcher to read data from STDIN, a timer
170	to display a message once per second, and a condvar to exit the program	1404	to display a message once per second, and a condition variable to quit the
171	when the user enters quit:	1405	program when the user enters quit:
172		1406
173	use AnyEvent;	1407	use AnyEvent;
174		1408
175	my $cv = AnyEvent->condvar;	1409	my $cv = AnyEvent->condvar;
176		1410
177	my $io_watcher = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {	1411	my $io_watcher = AnyEvent->io (
		1412	fh => \*STDIN,
		1413	poll => 'r',
		1414	cb => sub {
178	warn "io event <$_[0]>\n"; # will always output <r>	1415	warn "io event <$_[0]>\n"; # will always output <r>
179	chomp (my $input = <STDIN>); # read a line	1416	chomp (my $input = <STDIN>); # read a line
180	warn "read: $input\n"; # output what has been read	1417	warn "read: $input\n"; # output what has been read
181	$cv->broadcast if $input =~ /^q/i; # quit program if /^q/i	1418	$cv->send if $input =~ /^q/i; # quit program if /^q/i
		1419	},
182	});	1420	);
183		1421
184	my $time_watcher; # can only be used once	1422	my $time_watcher; # can only be used once
185		1423
186	sub new_timer {	1424	sub new_timer {
187	$timer = AnyEvent->timer (after => 1, cb => sub {	1425	$timer = AnyEvent->timer (after => 1, cb => sub {
…		…
190	});	1428	});
191	}	1429	}
192		1430
193	new_timer; # create first timer	1431	new_timer; # create first timer
194		1432
195	$cv->wait; # wait until user enters /^q/i	1433	$cv->recv; # wait until user enters /^q/i
196		1434
197	=head1 REAL-WORLD EXAMPLE	1435	=head1 REAL-WORLD EXAMPLE
198		1436
199	Consider the L<Net::FCP> module. It features (among others) the following	1437	Consider the L<Net::FCP> module. It features (among others) the following
200	API calls, which are to freenet what HTTP GET requests are to http:	1438	API calls, which are to freenet what HTTP GET requests are to http:
…		…
250	syswrite $txn->{fh}, $txn->{request}	1488	syswrite $txn->{fh}, $txn->{request}
251	or die "connection or write error";	1489	or die "connection or write error";
252	$txn->{w} = AnyEvent->io (fh => $txn->{fh}, poll => 'r', cb => sub { $txn->fh_ready_r });	1490	$txn->{w} = AnyEvent->io (fh => $txn->{fh}, poll => 'r', cb => sub { $txn->fh_ready_r });
253		1491
254	Again, C<fh_ready_r> waits till all data has arrived, and then stores the	1492	Again, C<fh_ready_r> waits till all data has arrived, and then stores the
255	result and signals any possible waiters that the request ahs finished:	1493	result and signals any possible waiters that the request has finished:
256		1494
257	sysread $txn->{fh}, $txn->{buf}, length $txn->{$buf};	1495	sysread $txn->{fh}, $txn->{buf}, length $txn->{$buf};
258		1496
259	if (end-of-file or data complete) {	1497	if (end-of-file or data complete) {
260	$txn->{result} = $txn->{buf};	1498	$txn->{result} = $txn->{buf};
261	$txn->{finished}->broadcast;	1499	$txn->{finished}->send;
262	$txb->{cb}->($txn) of $txn->{cb}; # also call callback	1500	$txb->{cb}->($txn) of $txn->{cb}; # also call callback
263	}	1501	}
264		1502
265	The C<result> method, finally, just waits for the finished signal (if the	1503	The C<result> method, finally, just waits for the finished signal (if the
266	request was already finished, it doesn't wait, of course, and returns the	1504	request was already finished, it doesn't wait, of course, and returns the
267	data:	1505	data:
268		1506
269	$txn->{finished}->wait;	1507	$txn->{finished}->recv;
270	return $txn->{result};	1508	return $txn->{result};
271		1509
272	The actual code goes further and collects all errors (C<die>s, exceptions)	1510	The actual code goes further and collects all errors (C<die>s, exceptions)
273	that occured during request processing. The C<result> method detects	1511	that occurred during request processing. The C<result> method detects
274	wether an exception as thrown (it is stored inside the $txn object)	1512	whether an exception as thrown (it is stored inside the $txn object)
275	and just throws the exception, which means connection errors and other	1513	and just throws the exception, which means connection errors and other
276	problems get reported tot he code that tries to use the result, not in a	1514	problems get reported tot he code that tries to use the result, not in a
277	random callback.	1515	random callback.
278		1516
279	All of this enables the following usage styles:	1517	All of this enables the following usage styles:
280		1518
281	1. Blocking:	1519	1. Blocking:
282		1520
283	my $data = $fcp->client_get ($url);	1521	my $data = $fcp->client_get ($url);
284		1522
285	2. Blocking, but parallelizing:	1523	2. Blocking, but running in parallel:
286		1524
287	my @datas = map $_->result,	1525	my @datas = map $_->result,
288	map $fcp->txn_client_get ($_),	1526	map $fcp->txn_client_get ($_),
289	@urls;	1527	@urls;
290		1528
291	Both blocking examples work without the module user having to know	1529	Both blocking examples work without the module user having to know
292	anything about events.	1530	anything about events.
293		1531
294	3a. Event-based in a main program, using any support Event module:	1532	3a. Event-based in a main program, using any supported event module:
295		1533
296	use Event;	1534	use EV;
297		1535
298	$fcp->txn_client_get ($url)->cb (sub {	1536	$fcp->txn_client_get ($url)->cb (sub {
299	my $txn = shift;	1537	my $txn = shift;
300	my $data = $txn->result;	1538	my $data = $txn->result;
301	...	1539	...
302	});	1540	});
303		1541
304	Event::loop;	1542	EV::loop;
305		1543
306	3b. The module user could use AnyEvent, too:	1544	3b. The module user could use AnyEvent, too:
307		1545
308	use AnyEvent;	1546	use AnyEvent;
309		1547
310	my $quit = AnyEvent->condvar;	1548	my $quit = AnyEvent->condvar;
311		1549
312	$fcp->txn_client_get ($url)->cb (sub {	1550	$fcp->txn_client_get ($url)->cb (sub {
313	...	1551	...
314	$quit->broadcast;	1552	$quit->send;
315	});	1553	});
316		1554
317	$quit->wait;	1555	$quit->recv;
		1556
		1557
		1558	=head1 BENCHMARKS
		1559
		1560	To give you an idea of the performance and overheads that AnyEvent adds
		1561	over the event loops themselves and to give you an impression of the speed
		1562	of various event loops I prepared some benchmarks.
		1563
		1564	=head2 BENCHMARKING ANYEVENT OVERHEAD
		1565
		1566	Here is a benchmark of various supported event models used natively and
		1567	through AnyEvent. The benchmark creates a lot of timers (with a zero
		1568	timeout) and I/O watchers (watching STDOUT, a pty, to become writable,
		1569	which it is), lets them fire exactly once and destroys them again.
		1570
		1571	Source code for this benchmark is found as F<eg/bench> in the AnyEvent
		1572	distribution.
		1573
		1574	=head3 Explanation of the columns
		1575
		1576	I<watcher> is the number of event watchers created/destroyed. Since
		1577	different event models feature vastly different performances, each event
		1578	loop was given a number of watchers so that overall runtime is acceptable
		1579	and similar between tested event loop (and keep them from crashing): Glib
		1580	would probably take thousands of years if asked to process the same number
		1581	of watchers as EV in this benchmark.
		1582
		1583	I<bytes> is the number of bytes (as measured by the resident set size,
		1584	RSS) consumed by each watcher. This method of measuring captures both C
		1585	and Perl-based overheads.
		1586
		1587	I<create> is the time, in microseconds (millionths of seconds), that it
		1588	takes to create a single watcher. The callback is a closure shared between
		1589	all watchers, to avoid adding memory overhead. That means closure creation
		1590	and memory usage is not included in the figures.
		1591
		1592	I<invoke> is the time, in microseconds, used to invoke a simple
		1593	callback. The callback simply counts down a Perl variable and after it was
		1594	invoked "watcher" times, it would C<< ->send >> a condvar once to
		1595	signal the end of this phase.
		1596
		1597	I<destroy> is the time, in microseconds, that it takes to destroy a single
		1598	watcher.
		1599
		1600	=head3 Results
		1601
		1602	name watchers bytes create invoke destroy comment
		1603	EV/EV 400000 224 0.47 0.35 0.27 EV native interface
		1604	EV/Any 100000 224 2.88 0.34 0.27 EV + AnyEvent watchers
		1605	CoroEV/Any 100000 224 2.85 0.35 0.28 coroutines + Coro::Signal
		1606	Perl/Any 100000 452 4.13 0.73 0.95 pure perl implementation
		1607	Event/Event 16000 517 32.20 31.80 0.81 Event native interface
		1608	Event/Any 16000 590 35.85 31.55 1.06 Event + AnyEvent watchers
		1609	Glib/Any 16000 1357 102.33 12.31 51.00 quadratic behaviour
		1610	Tk/Any 2000 1860 27.20 66.31 14.00 SEGV with >> 2000 watchers
		1611	POE/Event 2000 6328 109.99 751.67 14.02 via POE::Loop::Event
		1612	POE/Select 2000 6027 94.54 809.13 579.80 via POE::Loop::Select
		1613
		1614	=head3 Discussion
		1615
		1616	The benchmark does I<not> measure scalability of the event loop very
		1617	well. For example, a select-based event loop (such as the pure perl one)
		1618	can never compete with an event loop that uses epoll when the number of
		1619	file descriptors grows high. In this benchmark, all events become ready at
		1620	the same time, so select/poll-based implementations get an unnatural speed
		1621	boost.
		1622
		1623	Also, note that the number of watchers usually has a nonlinear effect on
		1624	overall speed, that is, creating twice as many watchers doesn't take twice
		1625	the time - usually it takes longer. This puts event loops tested with a
		1626	higher number of watchers at a disadvantage.
		1627
		1628	To put the range of results into perspective, consider that on the
		1629	benchmark machine, handling an event takes roughly 1600 CPU cycles with
		1630	EV, 3100 CPU cycles with AnyEvent's pure perl loop and almost 3000000 CPU
		1631	cycles with POE.
		1632
		1633	C<EV> is the sole leader regarding speed and memory use, which are both
		1634	maximal/minimal, respectively. Even when going through AnyEvent, it uses
		1635	far less memory than any other event loop and is still faster than Event
		1636	natively.
		1637
		1638	The pure perl implementation is hit in a few sweet spots (both the
		1639	constant timeout and the use of a single fd hit optimisations in the perl
		1640	interpreter and the backend itself). Nevertheless this shows that it
		1641	adds very little overhead in itself. Like any select-based backend its
		1642	performance becomes really bad with lots of file descriptors (and few of
		1643	them active), of course, but this was not subject of this benchmark.
		1644
		1645	The C<Event> module has a relatively high setup and callback invocation
		1646	cost, but overall scores in on the third place.
		1647
		1648	C<Glib>'s memory usage is quite a bit higher, but it features a
		1649	faster callback invocation and overall ends up in the same class as
		1650	C<Event>. However, Glib scales extremely badly, doubling the number of
		1651	watchers increases the processing time by more than a factor of four,
		1652	making it completely unusable when using larger numbers of watchers
		1653	(note that only a single file descriptor was used in the benchmark, so
		1654	inefficiencies of C<poll> do not account for this).
		1655
		1656	The C<Tk> adaptor works relatively well. The fact that it crashes with
		1657	more than 2000 watchers is a big setback, however, as correctness takes
		1658	precedence over speed. Nevertheless, its performance is surprising, as the
		1659	file descriptor is dup()ed for each watcher. This shows that the dup()
		1660	employed by some adaptors is not a big performance issue (it does incur a
		1661	hidden memory cost inside the kernel which is not reflected in the figures
		1662	above).
		1663
		1664	C<POE>, regardless of underlying event loop (whether using its pure perl
		1665	select-based backend or the Event module, the POE-EV backend couldn't
		1666	be tested because it wasn't working) shows abysmal performance and
		1667	memory usage with AnyEvent: Watchers use almost 30 times as much memory
		1668	as EV watchers, and 10 times as much memory as Event (the high memory
		1669	requirements are caused by requiring a session for each watcher). Watcher
		1670	invocation speed is almost 900 times slower than with AnyEvent's pure perl
		1671	implementation.
		1672
		1673	The design of the POE adaptor class in AnyEvent can not really account
		1674	for the performance issues, though, as session creation overhead is
		1675	small compared to execution of the state machine, which is coded pretty
		1676	optimally within L<AnyEvent::Impl::POE> (and while everybody agrees that
		1677	using multiple sessions is not a good approach, especially regarding
		1678	memory usage, even the author of POE could not come up with a faster
		1679	design).
		1680
		1681	=head3 Summary
		1682
		1683	=over 4
		1684
		1685	=item * Using EV through AnyEvent is faster than any other event loop
		1686	(even when used without AnyEvent), but most event loops have acceptable
		1687	performance with or without AnyEvent.
		1688
		1689	=item * The overhead AnyEvent adds is usually much smaller than the overhead of
		1690	the actual event loop, only with extremely fast event loops such as EV
		1691	adds AnyEvent significant overhead.
		1692
		1693	=item * You should avoid POE like the plague if you want performance or
		1694	reasonable memory usage.
		1695
		1696	=back
		1697
		1698	=head2 BENCHMARKING THE LARGE SERVER CASE
		1699
		1700	This benchmark actually benchmarks the event loop itself. It works by
		1701	creating a number of "servers": each server consists of a socket pair, a
		1702	timeout watcher that gets reset on activity (but never fires), and an I/O
		1703	watcher waiting for input on one side of the socket. Each time the socket
		1704	watcher reads a byte it will write that byte to a random other "server".
		1705
		1706	The effect is that there will be a lot of I/O watchers, only part of which
		1707	are active at any one point (so there is a constant number of active
		1708	fds for each loop iteration, but which fds these are is random). The
		1709	timeout is reset each time something is read because that reflects how
		1710	most timeouts work (and puts extra pressure on the event loops).
		1711
		1712	In this benchmark, we use 10000 socket pairs (20000 sockets), of which 100
		1713	(1%) are active. This mirrors the activity of large servers with many
		1714	connections, most of which are idle at any one point in time.
		1715
		1716	Source code for this benchmark is found as F<eg/bench2> in the AnyEvent
		1717	distribution.
		1718
		1719	=head3 Explanation of the columns
		1720
		1721	I<sockets> is the number of sockets, and twice the number of "servers" (as
		1722	each server has a read and write socket end).
		1723
		1724	I<create> is the time it takes to create a socket pair (which is
		1725	nontrivial) and two watchers: an I/O watcher and a timeout watcher.
		1726
		1727	I<request>, the most important value, is the time it takes to handle a
		1728	single "request", that is, reading the token from the pipe and forwarding
		1729	it to another server. This includes deleting the old timeout and creating
		1730	a new one that moves the timeout into the future.
		1731
		1732	=head3 Results
		1733
		1734	name sockets create request
		1735	EV 20000 69.01 11.16
		1736	Perl 20000 73.32 35.87
		1737	Event 20000 212.62 257.32
		1738	Glib 20000 651.16 1896.30
		1739	POE 20000 349.67 12317.24 uses POE::Loop::Event
		1740
		1741	=head3 Discussion
		1742
		1743	This benchmark I<does> measure scalability and overall performance of the
		1744	particular event loop.
		1745
		1746	EV is again fastest. Since it is using epoll on my system, the setup time
		1747	is relatively high, though.
		1748
		1749	Perl surprisingly comes second. It is much faster than the C-based event
		1750	loops Event and Glib.
		1751
		1752	Event suffers from high setup time as well (look at its code and you will
		1753	understand why). Callback invocation also has a high overhead compared to
		1754	the C<< $_->() for .. >>-style loop that the Perl event loop uses. Event
		1755	uses select or poll in basically all documented configurations.
		1756
		1757	Glib is hit hard by its quadratic behaviour w.r.t. many watchers. It
		1758	clearly fails to perform with many filehandles or in busy servers.
		1759
		1760	POE is still completely out of the picture, taking over 1000 times as long
		1761	as EV, and over 100 times as long as the Perl implementation, even though
		1762	it uses a C-based event loop in this case.
		1763
		1764	=head3 Summary
		1765
		1766	=over 4
		1767
		1768	=item * The pure perl implementation performs extremely well.
		1769
		1770	=item * Avoid Glib or POE in large projects where performance matters.
		1771
		1772	=back
		1773
		1774	=head2 BENCHMARKING SMALL SERVERS
		1775
		1776	While event loops should scale (and select-based ones do not...) even to
		1777	large servers, most programs we (or I :) actually write have only a few
		1778	I/O watchers.
		1779
		1780	In this benchmark, I use the same benchmark program as in the large server
		1781	case, but it uses only eight "servers", of which three are active at any
		1782	one time. This should reflect performance for a small server relatively
		1783	well.
		1784
		1785	The columns are identical to the previous table.
		1786
		1787	=head3 Results
		1788
		1789	name sockets create request
		1790	EV 16 20.00 6.54
		1791	Perl 16 25.75 12.62
		1792	Event 16 81.27 35.86
		1793	Glib 16 32.63 15.48
		1794	POE 16 261.87 276.28 uses POE::Loop::Event
		1795
		1796	=head3 Discussion
		1797
		1798	The benchmark tries to test the performance of a typical small
		1799	server. While knowing how various event loops perform is interesting, keep
		1800	in mind that their overhead in this case is usually not as important, due
		1801	to the small absolute number of watchers (that is, you need efficiency and
		1802	speed most when you have lots of watchers, not when you only have a few of
		1803	them).
		1804
		1805	EV is again fastest.
		1806
		1807	Perl again comes second. It is noticeably faster than the C-based event
		1808	loops Event and Glib, although the difference is too small to really
		1809	matter.
		1810
		1811	POE also performs much better in this case, but is is still far behind the
		1812	others.
		1813
		1814	=head3 Summary
		1815
		1816	=over 4
		1817
		1818	=item * C-based event loops perform very well with small number of
		1819	watchers, as the management overhead dominates.
		1820
		1821	=back
		1822
		1823
		1824	=head1 SIGNALS
		1825
		1826	AnyEvent currently installs handlers for these signals:
		1827
		1828	=over 4
		1829
		1830	=item SIGCHLD
		1831
		1832	A handler for C<SIGCHLD> is installed by AnyEvent's child watcher
		1833	emulation for event loops that do not support them natively. Also, some
		1834	event loops install a similar handler.
		1835
		1836	=item SIGPIPE
		1837
		1838	A no-op handler is installed for C<SIGPIPE> when C<$SIG{PIPE}> is C<undef>
		1839	when AnyEvent gets loaded.
		1840
		1841	The rationale for this is that AnyEvent users usually do not really depend
		1842	on SIGPIPE delivery (which is purely an optimisation for shell use, or
		1843	badly-written programs), but C<SIGPIPE> can cause spurious and rare
		1844	program exits as a lot of people do not expect C<SIGPIPE> when writing to
		1845	some random socket.
		1846
		1847	The rationale for installing a no-op handler as opposed to ignoring it is
		1848	that this way, the handler will be restored to defaults on exec.
		1849
		1850	Feel free to install your own handler, or reset it to defaults.
		1851
		1852	=back
		1853
		1854	=cut
		1855
		1856	$SIG{PIPE} = sub { }
		1857	unless defined $SIG{PIPE};
		1858
		1859
		1860	=head1 FORK
		1861
		1862	Most event libraries are not fork-safe. The ones who are usually are
		1863	because they rely on inefficient but fork-safe C<select> or C<poll>
		1864	calls. Only L<EV> is fully fork-aware.
		1865
		1866	If you have to fork, you must either do so I<before> creating your first
		1867	watcher OR you must not use AnyEvent at all in the child.
		1868
		1869
		1870	=head1 SECURITY CONSIDERATIONS
		1871
		1872	AnyEvent can be forced to load any event model via
		1873	$ENV{PERL_ANYEVENT_MODEL}. While this cannot (to my knowledge) be used to
		1874	execute arbitrary code or directly gain access, it can easily be used to
		1875	make the program hang or malfunction in subtle ways, as AnyEvent watchers
		1876	will not be active when the program uses a different event model than
		1877	specified in the variable.
		1878
		1879	You can make AnyEvent completely ignore this variable by deleting it
		1880	before the first watcher gets created, e.g. with a C<BEGIN> block:
		1881
		1882	BEGIN { delete $ENV{PERL_ANYEVENT_MODEL} }
		1883
		1884	use AnyEvent;
		1885
		1886	Similar considerations apply to $ENV{PERL_ANYEVENT_VERBOSE}, as that can
		1887	be used to probe what backend is used and gain other information (which is
		1888	probably even less useful to an attacker than PERL_ANYEVENT_MODEL), and
		1889	$ENV{PERL_ANYEGENT_STRICT}.
		1890
		1891
		1892	=head1 BUGS
		1893
		1894	Perl 5.8 has numerous memleaks that sometimes hit this module and are hard
		1895	to work around. If you suffer from memleaks, first upgrade to Perl 5.10
		1896	and check wether the leaks still show up. (Perl 5.10.0 has other annoying
		1897	memleaks, such as leaking on C<map> and C<grep> but it is usually not as
		1898	pronounced).
		1899
318		1900
319	=head1 SEE ALSO	1901	=head1 SEE ALSO
320		1902
321	Event modules: L<Coro::Event>, L<Coro>, L<Event>, L<Glib::Event>, L<Glib>.	1903	Utility functions: L<AnyEvent::Util>.
322		1904
323	Implementations: L<AnyEvent::Impl::Coro>, L<AnyEvent::Impl::Event>, L<AnyEvent::Impl::Glib>, L<AnyEvent::Impl::Tk>.	1905	Event modules: L<EV>, L<EV::Glib>, L<Glib::EV>, L<Event>, L<Glib::Event>,
		1906	L<Glib>, L<Tk>, L<Event::Lib>, L<Qt>, L<POE>.
324		1907
325	Nontrivial usage example: L<Net::FCP>.	1908	Implementations: L<AnyEvent::Impl::EV>, L<AnyEvent::Impl::Event>,
		1909	L<AnyEvent::Impl::Glib>, L<AnyEvent::Impl::Tk>, L<AnyEvent::Impl::Perl>,
		1910	L<AnyEvent::Impl::EventLib>, L<AnyEvent::Impl::Qt>,
		1911	L<AnyEvent::Impl::POE>.
326		1912
327	=head1	1913	Non-blocking file handles, sockets, TCP clients and
		1914	servers: L<AnyEvent::Handle>, L<AnyEvent::Socket>.
		1915
		1916	Asynchronous DNS: L<AnyEvent::DNS>.
		1917
		1918	Coroutine support: L<Coro>, L<Coro::AnyEvent>, L<Coro::EV>, L<Coro::Event>,
		1919
		1920	Nontrivial usage examples: L<Net::FCP>, L<Net::XMPP2>, L<AnyEvent::DNS>.
		1921
		1922
		1923	=head1 AUTHOR
		1924
		1925	Marc Lehmann <schmorp@schmorp.de>
		1926	http://home.schmorp.de/
328		1927
329	=cut	1928	=cut
330		1929
331	1	1930	1
332		1931

Diff Legend

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