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 1.227 by root, Mon Jul 6 23:42:24 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 and POE are various supported
		6	event loops.
6		7
7	=head1 SYNOPSIS	8	=head1 SYNOPSIS
8		9
9	use AnyEvent;	10	use AnyEvent;
10		11
		12	# file descriptor readable
11	my $w = AnyEvent->io (fh => ..., poll => "[rw]+", cb => sub {	13	my $w = AnyEvent->io (fh => $fh, poll => "r", cb => sub { ... });
12	my ($poll_got) = @_;	14
		15	# one-shot or repeating timers
		16	my $w = AnyEvent->timer (after => $seconds, cb => sub { ... });
		17	my $w = AnyEvent->timer (after => $seconds, interval => $seconds, cb => ...
		18
		19	print AnyEvent->now; # prints current event loop time
		20	print AnyEvent->time; # think Time::HiRes::time or simply CORE::time.
		21
		22	# POSIX signal
		23	my $w = AnyEvent->signal (signal => "TERM", cb => sub { ... });
		24
		25	# child process exit
		26	my $w = AnyEvent->child (pid => $pid, cb => sub {
		27	my ($pid, $status) = @_;
13	...	28	...
14	});	29	});
15		30
16	* only one io watcher per $fh and $poll type is allowed (i.e. on a socket	31	# called when event loop idle (if applicable)
17	you can have one r + one w or one rw watcher, not any more (limitation by	32	my $w = AnyEvent->idle (cb => sub { ... });
18	Tk).
19		33
20	* the C<$poll_got> passed to the handler needs to be checked by looking	34	my $w = AnyEvent->condvar; # stores whether a condition was flagged
21	for single characters (e.g. with a regex), as it can contain more event	35	$w->send; # wake up current and all future recv's
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	36	$w->recv; # enters "main loop" till $condvar gets ->send
39	$w->broadcast; # wake up current and all future wait's	37	# use a condvar in callback mode:
		38	$w->cb (sub { $_[0]->recv });
40		39
41	* condvars are used to give blocking behaviour when neccessary. Create	40	=head1 INTRODUCTION/TUTORIAL
42	a condvar for any "request" or "event" your module might create, C<<	41
43	->broadcast >> it when the event happens and provide a function that calls	42	This manpage is mainly a reference manual. If you are interested
44	C<< ->wait >> for it. See the examples below.	43	in a tutorial or some gentle introduction, have a look at the
		44	L<AnyEvent::Intro> manpage.
		45
		46	=head1 WHY YOU SHOULD USE THIS MODULE (OR NOT)
		47
		48	Glib, POE, IO::Async, Event... CPAN offers event models by the dozen
		49	nowadays. So what is different about AnyEvent?
		50
		51	Executive Summary: AnyEvent is I<compatible>, AnyEvent is I<free of
		52	policy> and AnyEvent is I<small and efficient>.
		53
		54	First and foremost, I<AnyEvent is not an event model> itself, it only
		55	interfaces to whatever event model the main program happens to use, in a
		56	pragmatic way. For event models and certain classes of immortals alike,
		57	the statement "there can only be one" is a bitter reality: In general,
		58	only one event loop can be active at the same time in a process. AnyEvent
		59	cannot change this, but it can hide the differences between those event
		60	loops.
		61
		62	The goal of AnyEvent is to offer module authors the ability to do event
		63	programming (waiting for I/O or timer events) without subscribing to a
		64	religion, a way of living, and most importantly: without forcing your
		65	module users into the same thing by forcing them to use the same event
		66	model you use.
		67
		68	For modules like POE or IO::Async (which is a total misnomer as it is
		69	actually doing all I/O I<synchronously>...), using them in your module is
		70	like joining a cult: After you joined, you are dependent on them and you
		71	cannot use anything else, as they are simply incompatible to everything
		72	that isn't them. What's worse, all the potential users of your
		73	module are I<also> forced to use the same event loop you use.
		74
		75	AnyEvent is different: AnyEvent + POE works fine. AnyEvent + Glib works
		76	fine. AnyEvent + Tk works fine etc. etc. but none of these work together
		77	with the rest: POE + IO::Async? No go. Tk + Event? No go. Again: if
		78	your module uses one of those, every user of your module has to use it,
		79	too. But if your module uses AnyEvent, it works transparently with all
		80	event models it supports (including stuff like IO::Async, as long as those
		81	use one of the supported event loops. It is trivial to add new event loops
		82	to AnyEvent, too, so it is future-proof).
		83
		84	In addition to being free of having to use I<the one and only true event
		85	model>, AnyEvent also is free of bloat and policy: with POE or similar
		86	modules, you get an enormous amount of code and strict rules you have to
		87	follow. AnyEvent, on the other hand, is lean and up to the point, by only
		88	offering the functionality that is necessary, in as thin as a wrapper as
		89	technically possible.
		90
		91	Of course, AnyEvent comes with a big (and fully optional!) toolbox
		92	of useful functionality, such as an asynchronous DNS resolver, 100%
		93	non-blocking connects (even with TLS/SSL, IPv6 and on broken platforms
		94	such as Windows) and lots of real-world knowledge and workarounds for
		95	platform bugs and differences.
		96
		97	Now, if you I<do want> lots of policy (this can arguably be somewhat
		98	useful) and you want to force your users to use the one and only event
		99	model, you should I<not> use this module.
45		100
46	=head1 DESCRIPTION	101	=head1 DESCRIPTION
47		102
48	L<AnyEvent> provides an identical interface to multiple event loops. This	103	L<AnyEvent> provides an identical interface to multiple event loops. This
49	allows module authors to utilizy an event loop without forcing module	104	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	105	users to use the same event loop (as only a single event loop can coexist
51	peacefully at any one time).	106	peacefully at any one time).
52		107
53	The interface itself is vaguely similar but not identical to the Event	108	The interface itself is vaguely similar, but not identical to the L<Event>
54	module.	109	module.
55		110
56	On the first call of any method, the module tries to detect the currently	111	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	112	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	113	following modules is already loaded: L<EV>,
59	used. If none is found, the module tries to load these modules in the	114	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	115	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.	116	to load these modules (excluding Tk, Event::Lib, Qt and POE as the pure perl
		117	adaptor should always succeed) in the order given. The first one that can
		118	be successfully loaded will be used. If, after this, still none could be
		119	found, AnyEvent will fall back to a pure-perl event loop, which is not
		120	very efficient, but should work everywhere.
		121
		122	Because AnyEvent first checks for modules that are already loaded, loading
		123	an event model explicitly before first using AnyEvent will likely make
		124	that model the default. For example:
		125
		126	use Tk;
		127	use AnyEvent;
		128
		129	# .. AnyEvent will likely default to Tk
		130
		131	The I<likely> means that, if any module loads another event model and
		132	starts using it, all bets are off. Maybe you should tell their authors to
		133	use AnyEvent so their modules work together with others seamlessly...
		134
		135	The pure-perl implementation of AnyEvent is called
		136	C<AnyEvent::Impl::Perl>. Like other event modules you can load it
		137	explicitly and enjoy the high availability of that event loop :)
		138
		139	=head1 WATCHERS
		140
		141	AnyEvent has the central concept of a I<watcher>, which is an object that
		142	stores relevant data for each kind of event you are waiting for, such as
		143	the callback to call, the file handle to watch, etc.
		144
		145	These watchers are normal Perl objects with normal Perl lifetime. After
		146	creating a watcher it will immediately "watch" for events and invoke the
		147	callback when the event occurs (of course, only when the event model
		148	is in control).
		149
		150	Note that B<callbacks must not permanently change global variables>
		151	potentially in use by the event loop (such as C<$_> or C<$[>) and that B<<
		152	callbacks must not C<die> >>. The former is good programming practise in
		153	Perl and the latter stems from the fact that exception handling differs
		154	widely between event loops.
		155
		156	To disable the watcher you have to destroy it (e.g. by setting the
		157	variable you store it in to C<undef> or otherwise deleting all references
		158	to it).
		159
		160	All watchers are created by calling a method on the C<AnyEvent> class.
		161
		162	Many watchers either are used with "recursion" (repeating timers for
		163	example), or need to refer to their watcher object in other ways.
		164
		165	An any way to achieve that is this pattern:
		166
		167	my $w; $w = AnyEvent->type (arg => value ..., cb => sub {
		168	# you can use $w here, for example to undef it
		169	undef $w;
		170	});
		171
		172	Note that C<my $w; $w => combination. This is necessary because in Perl,
		173	my variables are only visible after the statement in which they are
		174	declared.
		175
		176	=head2 I/O WATCHERS
		177
		178	You can create an I/O watcher by calling the C<< AnyEvent->io >> method
		179	with the following mandatory key-value pairs as arguments:
		180
		181	C<fh> is the Perl I<file handle> (I<not> file descriptor) to watch
		182	for events (AnyEvent might or might not keep a reference to this file
		183	handle). Note that only file handles pointing to things for which
		184	non-blocking operation makes sense are allowed. This includes sockets,
		185	most character devices, pipes, fifos and so on, but not for example files
		186	or block devices.
		187
		188	C<poll> must be a string that is either C<r> or C<w>, which creates a
		189	watcher waiting for "r"eadable or "w"ritable events, respectively.
		190
		191	C<cb> is the callback to invoke each time the file handle becomes ready.
		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 I/O watcher callbacks.
		196
		197	The I/O watcher might use the underlying file descriptor or a copy of it.
		198	You must not close a file handle as long as any watcher is active on the
		199	underlying file descriptor.
		200
		201	Some event loops issue spurious readyness notifications, so you should
		202	always use non-blocking calls when reading/writing from/to your file
		203	handles.
		204
		205	Example: wait for readability of STDIN, then read a line and disable the
		206	watcher.
		207
		208	my $w; $w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {
		209	chomp (my $input = <STDIN>);
		210	warn "read: $input\n";
		211	undef $w;
		212	});
		213
		214	=head2 TIME WATCHERS
		215
		216	You can create a time watcher by calling the C<< AnyEvent->timer >>
		217	method with the following mandatory arguments:
		218
		219	C<after> specifies after how many seconds (fractional values are
		220	supported) the callback should be invoked. C<cb> is the callback to invoke
		221	in that case.
		222
		223	Although the callback might get passed parameters, their value and
		224	presence is undefined and you cannot rely on them. Portable AnyEvent
		225	callbacks cannot use arguments passed to time watcher callbacks.
		226
		227	The callback will normally be invoked once only. If you specify another
		228	parameter, C<interval>, as a strictly positive number (> 0), then the
		229	callback will be invoked regularly at that interval (in fractional
		230	seconds) after the first invocation. If C<interval> is specified with a
		231	false value, then it is treated as if it were missing.
		232
		233	The callback will be rescheduled before invoking the callback, but no
		234	attempt is done to avoid timer drift in most backends, so the interval is
		235	only approximate.
		236
		237	Example: fire an event after 7.7 seconds.
		238
		239	my $w = AnyEvent->timer (after => 7.7, cb => sub {
		240	warn "timeout\n";
		241	});
		242
		243	# to cancel the timer:
		244	undef $w;
		245
		246	Example 2: fire an event after 0.5 seconds, then roughly every second.
		247
		248	my $w = AnyEvent->timer (after => 0.5, interval => 1, cb => sub {
		249	warn "timeout\n";
		250	};
		251
		252	=head3 TIMING ISSUES
		253
		254	There are two ways to handle timers: based on real time (relative, "fire
		255	in 10 seconds") and based on wallclock time (absolute, "fire at 12
		256	o'clock").
		257
		258	While most event loops expect timers to specified in a relative way, they
		259	use absolute time internally. This makes a difference when your clock
		260	"jumps", for example, when ntp decides to set your clock backwards from
		261	the wrong date of 2014-01-01 to 2008-01-01, a watcher that is supposed to
		262	fire "after" a second might actually take six years to finally fire.
		263
		264	AnyEvent cannot compensate for this. The only event loop that is conscious
		265	about these issues is L<EV>, which offers both relative (ev_timer, based
		266	on true relative time) and absolute (ev_periodic, based on wallclock time)
		267	timers.
		268
		269	AnyEvent always prefers relative timers, if available, matching the
		270	AnyEvent API.
		271
		272	AnyEvent has two additional methods that return the "current time":
62		273
63	=over 4	274	=over 4
64		275
		276	=item AnyEvent->time
		277
		278	This returns the "current wallclock time" as a fractional number of
		279	seconds since the Epoch (the same thing as C<time> or C<Time::HiRes::time>
		280	return, and the result is guaranteed to be compatible with those).
		281
		282	It progresses independently of any event loop processing, i.e. each call
		283	will check the system clock, which usually gets updated frequently.
		284
		285	=item AnyEvent->now
		286
		287	This also returns the "current wallclock time", but unlike C<time>, above,
		288	this value might change only once per event loop iteration, depending on
		289	the event loop (most return the same time as C<time>, above). This is the
		290	time that AnyEvent's timers get scheduled against.
		291
		292	I<In almost all cases (in all cases if you don't care), this is the
		293	function to call when you want to know the current time.>
		294
		295	This function is also often faster then C<< AnyEvent->time >>, and
		296	thus the preferred method if you want some timestamp (for example,
		297	L<AnyEvent::Handle> uses this to update it's activity timeouts).
		298
		299	The rest of this section is only of relevance if you try to be very exact
		300	with your timing, you can skip it without bad conscience.
		301
		302	For a practical example of when these times differ, consider L<Event::Lib>
		303	and L<EV> and the following set-up:
		304
		305	The event loop is running and has just invoked one of your callback at
		306	time=500 (assume no other callbacks delay processing). In your callback,
		307	you wait a second by executing C<sleep 1> (blocking the process for a
		308	second) and then (at time=501) you create a relative timer that fires
		309	after three seconds.
		310
		311	With L<Event::Lib>, C<< AnyEvent->time >> and C<< AnyEvent->now >> will
		312	both return C<501>, because that is the current time, and the timer will
		313	be scheduled to fire at time=504 (C<501> + C<3>).
		314
		315	With L<EV>, C<< AnyEvent->time >> returns C<501> (as that is the current
		316	time), but C<< AnyEvent->now >> returns C<500>, as that is the time the
		317	last event processing phase started. With L<EV>, your timer gets scheduled
		318	to run at time=503 (C<500> + C<3>).
		319
		320	In one sense, L<Event::Lib> is more exact, as it uses the current time
		321	regardless of any delays introduced by event processing. However, most
		322	callbacks do not expect large delays in processing, so this causes a
		323	higher drift (and a lot more system calls to get the current time).
		324
		325	In another sense, L<EV> is more exact, as your timer will be scheduled at
		326	the same time, regardless of how long event processing actually took.
		327
		328	In either case, if you care (and in most cases, you don't), then you
		329	can get whatever behaviour you want with any event loop, by taking the
		330	difference between C<< AnyEvent->time >> and C<< AnyEvent->now >> into
		331	account.
		332
		333	=item AnyEvent->now_update
		334
		335	Some event loops (such as L<EV> or L<AnyEvent::Impl::Perl>) cache
		336	the current time for each loop iteration (see the discussion of L<<
		337	AnyEvent->now >>, above).
		338
		339	When a callback runs for a long time (or when the process sleeps), then
		340	this "current" time will differ substantially from the real time, which
		341	might affect timers and time-outs.
		342
		343	When this is the case, you can call this method, which will update the
		344	event loop's idea of "current time".
		345
		346	Note that updating the time I<might> cause some events to be handled.
		347
		348	=back
		349
		350	=head2 SIGNAL WATCHERS
		351
		352	You can watch for signals using a signal watcher, C<signal> is the signal
		353	I<name> in uppercase and without any C<SIG> prefix, C<cb> is the Perl
		354	callback to be invoked whenever a signal occurs.
		355
		356	Although the callback might get passed parameters, their value and
		357	presence is undefined and you cannot rely on them. Portable AnyEvent
		358	callbacks cannot use arguments passed to signal watcher callbacks.
		359
		360	Multiple signal occurrences can be clumped together into one callback
		361	invocation, and callback invocation will be synchronous. Synchronous means
		362	that it might take a while until the signal gets handled by the process,
		363	but it is guaranteed not to interrupt any other callbacks.
		364
		365	The main advantage of using these watchers is that you can share a signal
		366	between multiple watchers.
		367
		368	This watcher might use C<%SIG>, so programs overwriting those signals
		369	directly will likely not work correctly.
		370
		371	Example: exit on SIGINT
		372
		373	my $w = AnyEvent->signal (signal => "INT", cb => sub { exit 1 });
		374
		375	=head2 CHILD PROCESS WATCHERS
		376
		377	You can also watch on a child process exit and catch its exit status.
		378
		379	The child process is specified by the C<pid> argument (if set to C<0>, it
		380	watches for any child process exit). The watcher will triggered only when
		381	the child process has finished and an exit status is available, not on
		382	any trace events (stopped/continued).
		383
		384	The callback will be called with the pid and exit status (as returned by
		385	waitpid), so unlike other watcher types, you I<can> rely on child watcher
		386	callback arguments.
		387
		388	This watcher type works by installing a signal handler for C<SIGCHLD>,
		389	and since it cannot be shared, nothing else should use SIGCHLD or reap
		390	random child processes (waiting for specific child processes, e.g. inside
		391	C<system>, is just fine).
		392
		393	There is a slight catch to child watchers, however: you usually start them
		394	I<after> the child process was created, and this means the process could
		395	have exited already (and no SIGCHLD will be sent anymore).
		396
		397	Not all event models handle this correctly (neither POE nor IO::Async do,
		398	see their AnyEvent::Impl manpages for details), but even for event models
		399	that I<do> handle this correctly, they usually need to be loaded before
		400	the process exits (i.e. before you fork in the first place). AnyEvent's
		401	pure perl event loop handles all cases correctly regardless of when you
		402	start the watcher.
		403
		404	This means you cannot create a child watcher as the very first
		405	thing in an AnyEvent program, you I<have> to create at least one
		406	watcher before you C<fork> the child (alternatively, you can call
		407	C<AnyEvent::detect>).
		408
		409	Example: fork a process and wait for it
		410
		411	my $done = AnyEvent->condvar;
		412
		413	my $pid = fork or exit 5;
		414
		415	my $w = AnyEvent->child (
		416	pid => $pid,
		417	cb => sub {
		418	my ($pid, $status) = @_;
		419	warn "pid $pid exited with status $status";
		420	$done->send;
		421	},
		422	);
		423
		424	# do something else, then wait for process exit
		425	$done->recv;
		426
		427	=head2 IDLE WATCHERS
		428
		429	Sometimes there is a need to do something, but it is not so important
		430	to do it instantly, but only when there is nothing better to do. This
		431	"nothing better to do" is usually defined to be "no other events need
		432	attention by the event loop".
		433
		434	Idle watchers ideally get invoked when the event loop has nothing
		435	better to do, just before it would block the process to wait for new
		436	events. Instead of blocking, the idle watcher is invoked.
		437
		438	Most event loops unfortunately do not really support idle watchers (only
		439	EV, Event and Glib do it in a usable fashion) - for the rest, AnyEvent
		440	will simply call the callback "from time to time".
		441
		442	Example: read lines from STDIN, but only process them when the
		443	program is otherwise idle:
		444
		445	my @lines; # read data
		446	my $idle_w;
		447	my $io_w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {
		448	push @lines, scalar <STDIN>;
		449
		450	# start an idle watcher, if not already done
		451	$idle_w ||= AnyEvent->idle (cb => sub {
		452	# handle only one line, when there are lines left
		453	if (my $line = shift @lines) {
		454	print "handled when idle: $line";
		455	} else {
		456	# otherwise disable the idle watcher again
		457	undef $idle_w;
		458	}
		459	});
		460	});
		461
		462	=head2 CONDITION VARIABLES
		463
		464	If you are familiar with some event loops you will know that all of them
		465	require you to run some blocking "loop", "run" or similar function that
		466	will actively watch for new events and call your callbacks.
		467
		468	AnyEvent is different, it expects somebody else to run the event loop and
		469	will only block when necessary (usually when told by the user).
		470
		471	The instrument to do that is called a "condition variable", so called
		472	because they represent a condition that must become true.
		473
		474	Condition variables can be created by calling the C<< AnyEvent->condvar
		475	>> method, usually without arguments. The only argument pair allowed is
		476
		477	C<cb>, which specifies a callback to be called when the condition variable
		478	becomes true, with the condition variable as the first argument (but not
		479	the results).
		480
		481	After creation, the condition variable is "false" until it becomes "true"
		482	by calling the C<send> method (or calling the condition variable as if it
		483	were a callback, read about the caveats in the description for the C<<
		484	->send >> method).
		485
		486	Condition variables are similar to callbacks, except that you can
		487	optionally wait for them. They can also be called merge points - points
		488	in time where multiple outstanding events have been processed. And yet
		489	another way to call them is transactions - each condition variable can be
		490	used to represent a transaction, which finishes at some point and delivers
		491	a result.
		492
		493	Condition variables are very useful to signal that something has finished,
		494	for example, if you write a module that does asynchronous http requests,
		495	then a condition variable would be the ideal candidate to signal the
		496	availability of results. The user can either act when the callback is
		497	called or can synchronously C<< ->recv >> for the results.
		498
		499	You can also use them to simulate traditional event loops - for example,
		500	you can block your main program until an event occurs - for example, you
		501	could C<< ->recv >> in your main program until the user clicks the Quit
		502	button of your app, which would C<< ->send >> the "quit" event.
		503
		504	Note that condition variables recurse into the event loop - if you have
		505	two pieces of code that call C<< ->recv >> in a round-robin fashion, you
		506	lose. Therefore, condition variables are good to export to your caller, but
		507	you should avoid making a blocking wait yourself, at least in callbacks,
		508	as this asks for trouble.
		509
		510	Condition variables are represented by hash refs in perl, and the keys
		511	used by AnyEvent itself are all named C<_ae_XXX> to make subclassing
		512	easy (it is often useful to build your own transaction class on top of
		513	AnyEvent). To subclass, use C<AnyEvent::CondVar> as base class and call
		514	it's C<new> method in your own C<new> method.
		515
		516	There are two "sides" to a condition variable - the "producer side" which
		517	eventually calls C<< -> send >>, and the "consumer side", which waits
		518	for the send to occur.
		519
		520	Example: wait for a timer.
		521
		522	# wait till the result is ready
		523	my $result_ready = AnyEvent->condvar;
		524
		525	# do something such as adding a timer
		526	# or socket watcher the calls $result_ready->send
		527	# when the "result" is ready.
		528	# in this case, we simply use a timer:
		529	my $w = AnyEvent->timer (
		530	after => 1,
		531	cb => sub { $result_ready->send },
		532	);
		533
		534	# this "blocks" (while handling events) till the callback
		535	# calls send
		536	$result_ready->recv;
		537
		538	Example: wait for a timer, but take advantage of the fact that
		539	condition variables are also code references.
		540
		541	my $done = AnyEvent->condvar;
		542	my $delay = AnyEvent->timer (after => 5, cb => $done);
		543	$done->recv;
		544
		545	Example: Imagine an API that returns a condvar and doesn't support
		546	callbacks. This is how you make a synchronous call, for example from
		547	the main program:
		548
		549	use AnyEvent::CouchDB;
		550
		551	...
		552
		553	my @info = $couchdb->info->recv;
		554
		555	And this is how you would just ste a callback to be called whenever the
		556	results are available:
		557
		558	$couchdb->info->cb (sub {
		559	my @info = $_[0]->recv;
		560	});
		561
		562	=head3 METHODS FOR PRODUCERS
		563
		564	These methods should only be used by the producing side, i.e. the
		565	code/module that eventually sends the signal. Note that it is also
		566	the producer side which creates the condvar in most cases, but it isn't
		567	uncommon for the consumer to create it as well.
		568
		569	=over 4
		570
		571	=item $cv->send (...)
		572
		573	Flag the condition as ready - a running C<< ->recv >> and all further
		574	calls to C<recv> will (eventually) return after this method has been
		575	called. If nobody is waiting the send will be remembered.
		576
		577	If a callback has been set on the condition variable, it is called
		578	immediately from within send.
		579
		580	Any arguments passed to the C<send> call will be returned by all
		581	future C<< ->recv >> calls.
		582
		583	Condition variables are overloaded so one can call them directly
		584	(as a code reference). Calling them directly is the same as calling
		585	C<send>. Note, however, that many C-based event loops do not handle
		586	overloading, so as tempting as it may be, passing a condition variable
		587	instead of a callback does not work. Both the pure perl and EV loops
		588	support overloading, however, as well as all functions that use perl to
		589	invoke a callback (as in L<AnyEvent::Socket> and L<AnyEvent::DNS> for
		590	example).
		591
		592	=item $cv->croak ($error)
		593
		594	Similar to send, but causes all call's to C<< ->recv >> to invoke
		595	C<Carp::croak> with the given error message/object/scalar.
		596
		597	This can be used to signal any errors to the condition variable
		598	user/consumer.
		599
		600	=item $cv->begin ([group callback])
		601
		602	=item $cv->end
		603
		604	These two methods can be used to combine many transactions/events into
		605	one. For example, a function that pings many hosts in parallel might want
		606	to use a condition variable for the whole process.
		607
		608	Every call to C<< ->begin >> will increment a counter, and every call to
		609	C<< ->end >> will decrement it. If the counter reaches C<0> in C<< ->end
		610	>>, the (last) callback passed to C<begin> will be executed. That callback
		611	is I<supposed> to call C<< ->send >>, but that is not required. If no
		612	callback was set, C<send> will be called without any arguments.
		613
		614	You can think of C<< $cv->send >> giving you an OR condition (one call
		615	sends), while C<< $cv->begin >> and C<< $cv->end >> giving you an AND
		616	condition (all C<begin> calls must be C<end>'ed before the condvar sends).
		617
		618	Let's start with a simple example: you have two I/O watchers (for example,
		619	STDOUT and STDERR for a program), and you want to wait for both streams to
		620	close before activating a condvar:
		621
		622	my $cv = AnyEvent->condvar;
		623
		624	$cv->begin; # first watcher
		625	my $w1 = AnyEvent->io (fh => $fh1, cb => sub {
		626	defined sysread $fh1, my $buf, 4096
		627	or $cv->end;
		628	});
		629
		630	$cv->begin; # second watcher
		631	my $w2 = AnyEvent->io (fh => $fh2, cb => sub {
		632	defined sysread $fh2, my $buf, 4096
		633	or $cv->end;
		634	});
		635
		636	$cv->recv;
		637
		638	This works because for every event source (EOF on file handle), there is
		639	one call to C<begin>, so the condvar waits for all calls to C<end> before
		640	sending.
		641
		642	The ping example mentioned above is slightly more complicated, as the
		643	there are results to be passwd back, and the number of tasks that are
		644	begung can potentially be zero:
		645
		646	my $cv = AnyEvent->condvar;
		647
		648	my %result;
		649	$cv->begin (sub { $cv->send (\%result) });
		650
		651	for my $host (@list_of_hosts) {
		652	$cv->begin;
		653	ping_host_then_call_callback $host, sub {
		654	$result{$host} = ...;
		655	$cv->end;
		656	};
		657	}
		658
		659	$cv->end;
		660
		661	This code fragment supposedly pings a number of hosts and calls
		662	C<send> after results for all then have have been gathered - in any
		663	order. To achieve this, the code issues a call to C<begin> when it starts
		664	each ping request and calls C<end> when it has received some result for
		665	it. Since C<begin> and C<end> only maintain a counter, the order in which
		666	results arrive is not relevant.
		667
		668	There is an additional bracketing call to C<begin> and C<end> outside the
		669	loop, which serves two important purposes: first, it sets the callback
		670	to be called once the counter reaches C<0>, and second, it ensures that
		671	C<send> is called even when C<no> hosts are being pinged (the loop
		672	doesn't execute once).
		673
		674	This is the general pattern when you "fan out" into multiple (but
		675	potentially none) subrequests: use an outer C<begin>/C<end> pair to set
		676	the callback and ensure C<end> is called at least once, and then, for each
		677	subrequest you start, call C<begin> and for each subrequest you finish,
		678	call C<end>.
		679
		680	=back
		681
		682	=head3 METHODS FOR CONSUMERS
		683
		684	These methods should only be used by the consuming side, i.e. the
		685	code awaits the condition.
		686
		687	=over 4
		688
		689	=item $cv->recv
		690
		691	Wait (blocking if necessary) until the C<< ->send >> or C<< ->croak
		692	>> methods have been called on c<$cv>, while servicing other watchers
		693	normally.
		694
		695	You can only wait once on a condition - additional calls are valid but
		696	will return immediately.
		697
		698	If an error condition has been set by calling C<< ->croak >>, then this
		699	function will call C<croak>.
		700
		701	In list context, all parameters passed to C<send> will be returned,
		702	in scalar context only the first one will be returned.
		703
		704	Not all event models support a blocking wait - some die in that case
		705	(programs might want to do that to stay interactive), so I<if you are
		706	using this from a module, never require a blocking wait>, but let the
		707	caller decide whether the call will block or not (for example, by coupling
		708	condition variables with some kind of request results and supporting
		709	callbacks so the caller knows that getting the result will not block,
		710	while still supporting blocking waits if the caller so desires).
		711
		712	Another reason I<never> to C<< ->recv >> in a module is that you cannot
		713	sensibly have two C<< ->recv >>'s in parallel, as that would require
		714	multiple interpreters or coroutines/threads, none of which C<AnyEvent>
		715	can supply.
		716
		717	The L<Coro> module, however, I<can> and I<does> supply coroutines and, in
		718	fact, L<Coro::AnyEvent> replaces AnyEvent's condvars by coroutine-safe
		719	versions and also integrates coroutines into AnyEvent, making blocking
		720	C<< ->recv >> calls perfectly safe as long as they are done from another
		721	coroutine (one that doesn't run the event loop).
		722
		723	You can ensure that C<< -recv >> never blocks by setting a callback and
		724	only calling C<< ->recv >> from within that callback (or at a later
		725	time). This will work even when the event loop does not support blocking
		726	waits otherwise.
		727
		728	=item $bool = $cv->ready
		729
		730	Returns true when the condition is "true", i.e. whether C<send> or
		731	C<croak> have been called.
		732
		733	=item $cb = $cv->cb ($cb->($cv))
		734
		735	This is a mutator function that returns the callback set and optionally
		736	replaces it before doing so.
		737
		738	The callback will be called when the condition becomes "true", i.e. when
		739	C<send> or C<croak> are called, with the only argument being the condition
		740	variable itself. Calling C<recv> inside the callback or at any later time
		741	is guaranteed not to block.
		742
		743	=back
		744
		745	=head1 GLOBAL VARIABLES AND FUNCTIONS
		746
		747	=over 4
		748
		749	=item $AnyEvent::MODEL
		750
		751	Contains C<undef> until the first watcher is being created. Then it
		752	contains the event model that is being used, which is the name of the
		753	Perl class implementing the model. This class is usually one of the
		754	C<AnyEvent::Impl:xxx> modules, but can be any other class in the case
		755	AnyEvent has been extended at runtime (e.g. in I<rxvt-unicode>).
		756
		757	The known classes so far are:
		758
		759	AnyEvent::Impl::EV based on EV (an interface to libev, best choice).
		760	AnyEvent::Impl::Event based on Event, second best choice.
		761	AnyEvent::Impl::Perl pure-perl implementation, fast and portable.
		762	AnyEvent::Impl::Glib based on Glib, third-best choice.
		763	AnyEvent::Impl::Tk based on Tk, very bad choice.
		764	AnyEvent::Impl::Qt based on Qt, cannot be autoprobed (see its docs).
		765	AnyEvent::Impl::EventLib based on Event::Lib, leaks memory and worse.
		766	AnyEvent::Impl::POE based on POE, not generic enough for full support.
		767
		768	# warning, support for IO::Async is only partial, as it is too broken
		769	# and limited toe ven support the AnyEvent API. See AnyEvent::Impl::Async.
		770	AnyEvent::Impl::IOAsync based on IO::Async, cannot be autoprobed (see its docs).
		771
		772	There is no support for WxWidgets, as WxWidgets has no support for
		773	watching file handles. However, you can use WxWidgets through the
		774	POE Adaptor, as POE has a Wx backend that simply polls 20 times per
		775	second, which was considered to be too horrible to even consider for
		776	AnyEvent. Likewise, other POE backends can be used by AnyEvent by using
		777	it's adaptor.
		778
		779	AnyEvent knows about L<Prima> and L<Wx> and will try to use L<POE> when
		780	autodetecting them.
		781
		782	=item AnyEvent::detect
		783
		784	Returns C<$AnyEvent::MODEL>, forcing autodetection of the event model
		785	if necessary. You should only call this function right before you would
		786	have created an AnyEvent watcher anyway, that is, as late as possible at
		787	runtime.
		788
		789	=item $guard = AnyEvent::post_detect { BLOCK }
		790
		791	Arranges for the code block to be executed as soon as the event model is
		792	autodetected (or immediately if this has already happened).
		793
		794	If called in scalar or list context, then it creates and returns an object
		795	that automatically removes the callback again when it is destroyed. See
		796	L<Coro::BDB> for a case where this is useful.
		797
		798	=item @AnyEvent::post_detect
		799
		800	If there are any code references in this array (you can C<push> to it
		801	before or after loading AnyEvent), then they will called directly after
		802	the event loop has been chosen.
		803
		804	You should check C<$AnyEvent::MODEL> before adding to this array, though:
		805	if it contains a true value then the event loop has already been detected,
		806	and the array will be ignored.
		807
		808	Best use C<AnyEvent::post_detect { BLOCK }> instead.
		809
		810	=back
		811
		812	=head1 WHAT TO DO IN A MODULE
		813
		814	As a module author, you should C<use AnyEvent> and call AnyEvent methods
		815	freely, but you should not load a specific event module or rely on it.
		816
		817	Be careful when you create watchers in the module body - AnyEvent will
		818	decide which event module to use as soon as the first method is called, so
		819	by calling AnyEvent in your module body you force the user of your module
		820	to load the event module first.
		821
		822	Never call C<< ->recv >> on a condition variable unless you I<know> that
		823	the C<< ->send >> method has been called on it already. This is
		824	because it will stall the whole program, and the whole point of using
		825	events is to stay interactive.
		826
		827	It is fine, however, to call C<< ->recv >> when the user of your module
		828	requests it (i.e. if you create a http request object ad have a method
		829	called C<results> that returns the results, it should call C<< ->recv >>
		830	freely, as the user of your module knows what she is doing. always).
		831
		832	=head1 WHAT TO DO IN THE MAIN PROGRAM
		833
		834	There will always be a single main program - the only place that should
		835	dictate which event model to use.
		836
		837	If it doesn't care, it can just "use AnyEvent" and use it itself, or not
		838	do anything special (it does not need to be event-based) and let AnyEvent
		839	decide which implementation to chose if some module relies on it.
		840
		841	If the main program relies on a specific event model - for example, in
		842	Gtk2 programs you have to rely on the Glib module - you should load the
		843	event module before loading AnyEvent or any module that uses it: generally
		844	speaking, you should load it as early as possible. The reason is that
		845	modules might create watchers when they are loaded, and AnyEvent will
		846	decide on the event model to use as soon as it creates watchers, and it
		847	might chose the wrong one unless you load the correct one yourself.
		848
		849	You can chose to use a pure-perl implementation by loading the
		850	C<AnyEvent::Impl::Perl> module, which gives you similar behaviour
		851	everywhere, but letting AnyEvent chose the model is generally better.
		852
		853	=head2 MAINLOOP EMULATION
		854
		855	Sometimes (often for short test scripts, or even standalone programs who
		856	only want to use AnyEvent), you do not want to run a specific event loop.
		857
		858	In that case, you can use a condition variable like this:
		859
		860	AnyEvent->condvar->recv;
		861
		862	This has the effect of entering the event loop and looping forever.
		863
		864	Note that usually your program has some exit condition, in which case
		865	it is better to use the "traditional" approach of storing a condition
		866	variable somewhere, waiting for it, and sending it when the program should
		867	exit cleanly.
		868
		869
		870	=head1 OTHER MODULES
		871
		872	The following is a non-exhaustive list of additional modules that use
		873	AnyEvent and can therefore be mixed easily with other AnyEvent modules
		874	in the same program. Some of the modules come with AnyEvent, some are
		875	available via CPAN.
		876
		877	=over 4
		878
		879	=item L<AnyEvent::Util>
		880
		881	Contains various utility functions that replace often-used but blocking
		882	functions such as C<inet_aton> by event-/callback-based versions.
		883
		884	=item L<AnyEvent::Socket>
		885
		886	Provides various utility functions for (internet protocol) sockets,
		887	addresses and name resolution. Also functions to create non-blocking tcp
		888	connections or tcp servers, with IPv6 and SRV record support and more.
		889
		890	=item L<AnyEvent::Handle>
		891
		892	Provide read and write buffers, manages watchers for reads and writes,
		893	supports raw and formatted I/O, I/O queued and fully transparent and
		894	non-blocking SSL/TLS.
		895
		896	=item L<AnyEvent::DNS>
		897
		898	Provides rich asynchronous DNS resolver capabilities.
		899
		900	=item L<AnyEvent::HTTP>
		901
		902	A simple-to-use HTTP library that is capable of making a lot of concurrent
		903	HTTP requests.
		904
		905	=item L<AnyEvent::HTTPD>
		906
		907	Provides a simple web application server framework.
		908
		909	=item L<AnyEvent::FastPing>
		910
		911	The fastest ping in the west.
		912
		913	=item L<AnyEvent::DBI>
		914
		915	Executes L<DBI> requests asynchronously in a proxy process.
		916
		917	=item L<AnyEvent::AIO>
		918
		919	Truly asynchronous I/O, should be in the toolbox of every event
		920	programmer. AnyEvent::AIO transparently fuses L<IO::AIO> and AnyEvent
		921	together.
		922
		923	=item L<AnyEvent::BDB>
		924
		925	Truly asynchronous Berkeley DB access. AnyEvent::BDB transparently fuses
		926	L<BDB> and AnyEvent together.
		927
		928	=item L<AnyEvent::GPSD>
		929
		930	A non-blocking interface to gpsd, a daemon delivering GPS information.
		931
		932	=item L<AnyEvent::IGS>
		933
		934	A non-blocking interface to the Internet Go Server protocol (used by
		935	L<App::IGS>).
		936
		937	=item L<AnyEvent::IRC>
		938
		939	AnyEvent based IRC client module family (replacing the older Net::IRC3).
		940
		941	=item L<Net::XMPP2>
		942
		943	AnyEvent based XMPP (Jabber protocol) module family.
		944
		945	=item L<Net::FCP>
		946
		947	AnyEvent-based implementation of the Freenet Client Protocol, birthplace
		948	of AnyEvent.
		949
		950	=item L<Event::ExecFlow>
		951
		952	High level API for event-based execution flow control.
		953
		954	=item L<Coro>
		955
		956	Has special support for AnyEvent via L<Coro::AnyEvent>.
		957
		958	=item L<IO::Lambda>
		959
		960	The lambda approach to I/O - don't ask, look there. Can use AnyEvent.
		961
		962	=back
		963
65	=cut	964	=cut
66		965
67	package AnyEvent;	966	package AnyEvent;
68		967
69	no warnings;	968	no warnings;
70	use strict 'vars';	969	use strict qw(vars subs);
		970
71	use Carp;	971	use Carp;
72		972
73	our $VERSION = '1.02';	973	our $VERSION = 4.8;
74	our $MODEL;	974	our $MODEL;
75		975
76	our $AUTOLOAD;	976	our $AUTOLOAD;
77	our @ISA;	977	our @ISA;
78		978
		979	our @REGISTRY;
		980
		981	our $WIN32;
		982
		983	BEGIN {
		984	eval "sub WIN32(){ " . (($^O =~ /mswin32/i)*1) ." }";
		985	eval "sub TAINT(){ " . (${^TAINT}*1) . " }";
		986
		987	delete @ENV{grep /^PERL_ANYEVENT_/, keys %ENV}
		988	if ${^TAINT};
		989	}
		990
79	our $verbose = $ENV{PERL_ANYEVENT_VERBOSE}*1;	991	our $verbose = $ENV{PERL_ANYEVENT_VERBOSE}*1;
80		992
81	our @REGISTRY;	993	our %PROTOCOL; # (ipv4|ipv6) => (1|2), higher numbers are preferred
		994
		995	{
		996	my $idx;
		997	$PROTOCOL{$_} = ++$idx
		998	for reverse split /\s*,\s*/,
		999	$ENV{PERL_ANYEVENT_PROTOCOLS} || "ipv4,ipv6";
		1000	}
82		1001
83	my @models = (	1002	my @models = (
84	[Coro::Event:: => AnyEvent::Impl::Coro::],	1003	[EV:: => AnyEvent::Impl::EV::],
85	[Event:: => AnyEvent::Impl::Event::],	1004	[Event:: => AnyEvent::Impl::Event::],
86	[Glib:: => AnyEvent::Impl::Glib::],	1005	[AnyEvent::Impl::Perl:: => AnyEvent::Impl::Perl::],
		1006	# everything below here will not be autoprobed
		1007	# as the pureperl backend should work everywhere
		1008	# and is usually faster
		1009	[Tk:: => AnyEvent::Impl::Tk::], # crashes with many handles
		1010	[Glib:: => AnyEvent::Impl::Glib::], # becomes extremely slow with many watchers
		1011	[Event::Lib:: => AnyEvent::Impl::EventLib::], # too buggy
		1012	[Qt:: => AnyEvent::Impl::Qt::], # requires special main program
		1013	[POE::Kernel:: => AnyEvent::Impl::POE::], # lasciate ogni speranza
87	[Tk:: => AnyEvent::Impl::Tk::],	1014	[Wx:: => AnyEvent::Impl::POE::],
		1015	[Prima:: => AnyEvent::Impl::POE::],
		1016	# IO::Async is just too broken - we would need workaorunds for its
		1017	# byzantine signal and broken child handling, among others.
		1018	# IO::Async is rather hard to detect, as it doesn't have any
		1019	# obvious default class.
		1020	# [IO::Async:: => AnyEvent::Impl::IOAsync::], # requires special main program
		1021	# [IO::Async::Loop:: => AnyEvent::Impl::IOAsync::], # requires special main program
		1022	# [IO::Async::Notifier:: => AnyEvent::Impl::IOAsync::], # requires special main program
88	);	1023	);
89		1024
90	our %method = map +($_ => 1), qw(io timer condvar broadcast wait cancel DESTROY);	1025	our %method = map +($_ => 1),
		1026	qw(io timer time now now_update signal child idle condvar one_event DESTROY);
91		1027
92	sub AUTOLOAD {	1028	our @post_detect;
93	$AUTOLOAD =~ s/.*://;
94		1029
95	$method{$AUTOLOAD}	1030	sub post_detect(&) {
96	or croak "$AUTOLOAD: not a valid method for AnyEvent objects";	1031	my ($cb) = @_;
97		1032
		1033	if ($MODEL) {
		1034	$cb->();
		1035
		1036	1
		1037	} else {
		1038	push @post_detect, $cb;
		1039
		1040	defined wantarray
		1041	? bless \$cb, "AnyEvent::Util::postdetect"
		1042	: ()
		1043	}
		1044	}
		1045
		1046	sub AnyEvent::Util::postdetect::DESTROY {
		1047	@post_detect = grep $_ != ${$_[0]}, @post_detect;
		1048	}
		1049
		1050	sub detect() {
98	unless ($MODEL) {	1051	unless ($MODEL) {
		1052	no strict 'refs';
		1053	local $SIG{__DIE__};
		1054
		1055	if ($ENV{PERL_ANYEVENT_MODEL} =~ /^([a-zA-Z]+)$/) {
		1056	my $model = "AnyEvent::Impl::$1";
		1057	if (eval "require $model") {
		1058	$MODEL = $model;
		1059	warn "AnyEvent: loaded model '$model' (forced by \$PERL_ANYEVENT_MODEL), using it.\n" if $verbose > 1;
		1060	} else {
		1061	warn "AnyEvent: unable to load model '$model' (from \$PERL_ANYEVENT_MODEL):\n$@" if $verbose;
		1062	}
		1063	}
		1064
99	# check for already loaded models	1065	# check for already loaded models
		1066	unless ($MODEL) {
100	for (@REGISTRY, @models) {	1067	for (@REGISTRY, @models) {
101	my ($package, $model) = @$_;	1068	my ($package, $model) = @$_;
102	if (${"$package\::VERSION"} > 0) {	1069	if (${"$package\::VERSION"} > 0) {
103	if (eval "require $model") {	1070	if (eval "require $model") {
104	$MODEL = $model;	1071	$MODEL = $model;
105	warn "AnyEvent: found model '$model', using it.\n" if $verbose > 1;	1072	warn "AnyEvent: autodetected model '$model', using it.\n" if $verbose > 1;
106	last;	1073	last;
		1074	}
107	}	1075	}
108	}	1076	}
		1077
		1078	unless ($MODEL) {
		1079	# try to load a model
		1080
		1081	for (@REGISTRY, @models) {
		1082	my ($package, $model) = @$_;
		1083	if (eval "require $package"
		1084	and ${"$package\::VERSION"} > 0
		1085	and eval "require $model") {
		1086	$MODEL = $model;
		1087	warn "AnyEvent: autoprobed model '$model', using it.\n" if $verbose > 1;
		1088	last;
		1089	}
		1090	}
		1091
		1092	$MODEL
		1093	or die "No event module selected for AnyEvent and autodetect failed. Install any one of these modules: EV, Event or Glib.\n";
		1094	}
109	}	1095	}
110		1096
111	unless ($MODEL) {	1097	push @{"$MODEL\::ISA"}, "AnyEvent::Base";
112	# try to load a model
113		1098
114	for (@REGISTRY, @models) {	1099	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		1100
		1101	require AnyEvent::Strict if $ENV{PERL_ANYEVENT_STRICT};
		1102
		1103	(shift @post_detect)->() while @post_detect;
		1104	}
		1105
123	$MODEL	1106	$MODEL
124	or die "No event module selected for AnyEvent and autodetect failed. Install any one of these modules: Coro, Event, Glib or Tk.";	1107	}
		1108
		1109	sub AUTOLOAD {
		1110	(my $func = $AUTOLOAD) =~ s/.*://;
		1111
		1112	$method{$func}
		1113	or croak "$func: not a valid method for AnyEvent objects";
		1114
		1115	detect unless $MODEL;
		1116
		1117	my $class = shift;
		1118	$class->$func (@_);
		1119	}
		1120
		1121	# utility function to dup a filehandle. this is used by many backends
		1122	# to support binding more than one watcher per filehandle (they usually
		1123	# allow only one watcher per fd, so we dup it to get a different one).
		1124	sub _dupfh($$;$$) {
		1125	my ($poll, $fh, $r, $w) = @_;
		1126
		1127	# cygwin requires the fh mode to be matching, unix doesn't
		1128	my ($rw, $mode) = $poll eq "r" ? ($r, "<")
		1129	: $poll eq "w" ? ($w, ">")
		1130	: Carp::croak "AnyEvent->io requires poll set to either 'r' or 'w'";
		1131
		1132	open my $fh2, "$mode&" . fileno $fh
		1133	or die "cannot dup() filehandle: $!,";
		1134
		1135	# we assume CLOEXEC is already set by perl in all important cases
		1136
		1137	($fh2, $rw)
		1138	}
		1139
		1140	package AnyEvent::Base;
		1141
		1142	# default implementations for many methods
		1143
		1144	BEGIN {
		1145	if (eval "use Time::HiRes (); Time::HiRes::time (); 1") {
		1146	*_time = \&Time::HiRes::time;
		1147	# if (eval "use POSIX (); (POSIX::times())...
		1148	} else {
		1149	*_time = sub { time }; # epic fail
		1150	}
		1151	}
		1152
		1153	sub time { _time }
		1154	sub now { _time }
		1155	sub now_update { }
		1156
		1157	# default implementation for ->condvar
		1158
		1159	sub condvar {
		1160	bless { @_ == 3 ? (_ae_cb => $_[2]) : () }, "AnyEvent::CondVar"
		1161	}
		1162
		1163	# default implementation for ->signal
		1164
		1165	our ($SIGPIPE_R, $SIGPIPE_W, %SIG_CB, %SIG_EV, $SIG_IO);
		1166
		1167	sub _signal_exec {
		1168	sysread $SIGPIPE_R, my $dummy, 4;
		1169
		1170	while (%SIG_EV) {
		1171	for (keys %SIG_EV) {
		1172	delete $SIG_EV{$_};
		1173	$_->() for values %{ $SIG_CB{$_} || {} };
125	}	1174	}
126	}	1175	}
		1176	}
127		1177
128	@ISA = $MODEL;	1178	sub signal {
		1179	my (undef, %arg) = @_;
129		1180
		1181	unless ($SIGPIPE_R) {
		1182	require Fcntl;
		1183
		1184	if (AnyEvent::WIN32) {
		1185	require AnyEvent::Util;
		1186
		1187	($SIGPIPE_R, $SIGPIPE_W) = AnyEvent::Util::portable_pipe ();
		1188	AnyEvent::Util::fh_nonblocking ($SIGPIPE_R) if $SIGPIPE_R;
		1189	AnyEvent::Util::fh_nonblocking ($SIGPIPE_W) if $SIGPIPE_W; # just in case
		1190	} else {
		1191	pipe $SIGPIPE_R, $SIGPIPE_W;
		1192	fcntl $SIGPIPE_R, &Fcntl::F_SETFL, &Fcntl::O_NONBLOCK if $SIGPIPE_R;
		1193	fcntl $SIGPIPE_W, &Fcntl::F_SETFL, &Fcntl::O_NONBLOCK if $SIGPIPE_W; # just in case
		1194
		1195	# not strictly required, as $^F is normally 2, but let's make sure...
		1196	fcntl $SIGPIPE_R, &Fcntl::F_SETFD, &Fcntl::FD_CLOEXEC;
		1197	fcntl $SIGPIPE_W, &Fcntl::F_SETFD, &Fcntl::FD_CLOEXEC;
		1198	}
		1199
		1200	$SIGPIPE_R
		1201	or Carp::croak "AnyEvent: unable to create a signal reporting pipe: $!\n";
		1202
		1203	$SIG_IO = AnyEvent->io (fh => $SIGPIPE_R, poll => "r", cb => \&_signal_exec);
		1204	}
		1205
		1206	my $signal = uc $arg{signal}
		1207	or Carp::croak "required option 'signal' is missing";
		1208
		1209	$SIG_CB{$signal}{$arg{cb}} = $arg{cb};
		1210	$SIG{$signal} ||= sub {
		1211	local $!;
		1212	syswrite $SIGPIPE_W, "\x00", 1 unless %SIG_EV;
		1213	undef $SIG_EV{$signal};
		1214	};
		1215
		1216	bless [$signal, $arg{cb}], "AnyEvent::Base::signal"
		1217	}
		1218
		1219	sub AnyEvent::Base::signal::DESTROY {
		1220	my ($signal, $cb) = @{$_[0]};
		1221
		1222	delete $SIG_CB{$signal}{$cb};
		1223
		1224	# delete doesn't work with older perls - they then
		1225	# print weird messages, or just unconditionally exit
		1226	# instead of getting the default action.
		1227	undef $SIG{$signal} unless keys %{ $SIG_CB{$signal} };
		1228	}
		1229
		1230	# default implementation for ->child
		1231
		1232	our %PID_CB;
		1233	our $CHLD_W;
		1234	our $CHLD_DELAY_W;
		1235	our $WNOHANG;
		1236
		1237	sub _sigchld {
		1238	while (0 < (my $pid = waitpid -1, $WNOHANG)) {
		1239	$_->($pid, $?) for (values %{ $PID_CB{$pid} || {} }),
		1240	(values %{ $PID_CB{0} || {} });
		1241	}
		1242	}
		1243
		1244	sub child {
		1245	my (undef, %arg) = @_;
		1246
		1247	defined (my $pid = $arg{pid} + 0)
		1248	or Carp::croak "required option 'pid' is missing";
		1249
		1250	$PID_CB{$pid}{$arg{cb}} = $arg{cb};
		1251
		1252	$WNOHANG ||= eval { local $SIG{__DIE__}; require POSIX; &POSIX::WNOHANG } || 1;
		1253
		1254	unless ($CHLD_W) {
		1255	$CHLD_W = AnyEvent->signal (signal => 'CHLD', cb => \&_sigchld);
		1256	# child could be a zombie already, so make at least one round
		1257	&_sigchld;
		1258	}
		1259
		1260	bless [$pid, $arg{cb}], "AnyEvent::Base::child"
		1261	}
		1262
		1263	sub AnyEvent::Base::child::DESTROY {
		1264	my ($pid, $cb) = @{$_[0]};
		1265
		1266	delete $PID_CB{$pid}{$cb};
		1267	delete $PID_CB{$pid} unless keys %{ $PID_CB{$pid} };
		1268
		1269	undef $CHLD_W unless keys %PID_CB;
		1270	}
		1271
		1272	# idle emulation is done by simply using a timer, regardless
		1273	# of whether the process is idle or not, and not letting
		1274	# the callback use more than 50% of the time.
		1275	sub idle {
		1276	my (undef, %arg) = @_;
		1277
		1278	my ($cb, $w, $rcb) = $arg{cb};
		1279
		1280	$rcb = sub {
		1281	if ($cb) {
		1282	$w = _time;
		1283	&$cb;
		1284	$w = _time - $w;
		1285
		1286	# never use more then 50% of the time for the idle watcher,
		1287	# within some limits
		1288	$w = 0.0001 if $w < 0.0001;
		1289	$w = 5 if $w > 5;
		1290
		1291	$w = AnyEvent->timer (after => $w, cb => $rcb);
		1292	} else {
		1293	# clean up...
		1294	undef $w;
		1295	undef $rcb;
		1296	}
		1297	};
		1298
		1299	$w = AnyEvent->timer (after => 0.05, cb => $rcb);
		1300
		1301	bless \\$cb, "AnyEvent::Base::idle"
		1302	}
		1303
		1304	sub AnyEvent::Base::idle::DESTROY {
		1305	undef $${$_[0]};
		1306	}
		1307
		1308	package AnyEvent::CondVar;
		1309
		1310	our @ISA = AnyEvent::CondVar::Base::;
		1311
		1312	package AnyEvent::CondVar::Base;
		1313
		1314	use overload
		1315	'&{}' => sub { my $self = shift; sub { $self->send (@_) } },
		1316	fallback => 1;
		1317
		1318	sub _send {
		1319	# nop
		1320	}
		1321
		1322	sub send {
130	my $class = shift;	1323	my $cv = shift;
131	$class->$AUTOLOAD (@_);	1324	$cv->{_ae_sent} = [@_];
		1325	(delete $cv->{_ae_cb})->($cv) if $cv->{_ae_cb};
		1326	$cv->_send;
132	}	1327	}
		1328
		1329	sub croak {
		1330	$_[0]{_ae_croak} = $_[1];
		1331	$_[0]->send;
		1332	}
		1333
		1334	sub ready {
		1335	$_[0]{_ae_sent}
		1336	}
		1337
		1338	sub _wait {
		1339	AnyEvent->one_event while !$_[0]{_ae_sent};
		1340	}
		1341
		1342	sub recv {
		1343	$_[0]->_wait;
		1344
		1345	Carp::croak $_[0]{_ae_croak} if $_[0]{_ae_croak};
		1346	wantarray ? @{ $_[0]{_ae_sent} } : $_[0]{_ae_sent}[0]
		1347	}
		1348
		1349	sub cb {
		1350	$_[0]{_ae_cb} = $_[1] if @_ > 1;
		1351	$_[0]{_ae_cb}
		1352	}
		1353
		1354	sub begin {
		1355	++$_[0]{_ae_counter};
		1356	$_[0]{_ae_end_cb} = $_[1] if @_ > 1;
		1357	}
		1358
		1359	sub end {
		1360	return if --$_[0]{_ae_counter};
		1361	&{ $_[0]{_ae_end_cb} || sub { $_[0]->send } };
		1362	}
		1363
		1364	# undocumented/compatibility with pre-3.4
		1365	*broadcast = \&send;
		1366	*wait = \&_wait;
		1367
		1368	=head1 ERROR AND EXCEPTION HANDLING
		1369
		1370	In general, AnyEvent does not do any error handling - it relies on the
		1371	caller to do that if required. The L<AnyEvent::Strict> module (see also
		1372	the C<PERL_ANYEVENT_STRICT> environment variable, below) provides strict
		1373	checking of all AnyEvent methods, however, which is highly useful during
		1374	development.
		1375
		1376	As for exception handling (i.e. runtime errors and exceptions thrown while
		1377	executing a callback), this is not only highly event-loop specific, but
		1378	also not in any way wrapped by this module, as this is the job of the main
		1379	program.
		1380
		1381	The pure perl event loop simply re-throws the exception (usually
		1382	within C<< condvar->recv >>), the L<Event> and L<EV> modules call C<<
		1383	$Event/EV::DIED->() >>, L<Glib> uses C<< install_exception_handler >> and
		1384	so on.
		1385
		1386	=head1 ENVIRONMENT VARIABLES
		1387
		1388	The following environment variables are used by this module or its
		1389	submodules.
		1390
		1391	Note that AnyEvent will remove I<all> environment variables starting with
		1392	C<PERL_ANYEVENT_> from C<%ENV> when it is loaded while taint mode is
		1393	enabled.
		1394
		1395	=over 4
		1396
		1397	=item C<PERL_ANYEVENT_VERBOSE>
		1398
		1399	By default, AnyEvent will be completely silent except in fatal
		1400	conditions. You can set this environment variable to make AnyEvent more
		1401	talkative.
		1402
		1403	When set to C<1> or higher, causes AnyEvent to warn about unexpected
		1404	conditions, such as not being able to load the event model specified by
		1405	C<PERL_ANYEVENT_MODEL>.
		1406
		1407	When set to C<2> or higher, cause AnyEvent to report to STDERR which event
		1408	model it chooses.
		1409
		1410	=item C<PERL_ANYEVENT_STRICT>
		1411
		1412	AnyEvent does not do much argument checking by default, as thorough
		1413	argument checking is very costly. Setting this variable to a true value
		1414	will cause AnyEvent to load C<AnyEvent::Strict> and then to thoroughly
		1415	check the arguments passed to most method calls. If it finds any problems,
		1416	it will croak.
		1417
		1418	In other words, enables "strict" mode.
		1419
		1420	Unlike C<use strict>, it is definitely recommended to keep it off in
		1421	production. Keeping C<PERL_ANYEVENT_STRICT=1> in your environment while
		1422	developing programs can be very useful, however.
		1423
		1424	=item C<PERL_ANYEVENT_MODEL>
		1425
		1426	This can be used to specify the event model to be used by AnyEvent, before
		1427	auto detection and -probing kicks in. It must be a string consisting
		1428	entirely of ASCII letters. The string C<AnyEvent::Impl::> gets prepended
		1429	and the resulting module name is loaded and if the load was successful,
		1430	used as event model. If it fails to load AnyEvent will proceed with
		1431	auto detection and -probing.
		1432
		1433	This functionality might change in future versions.
		1434
		1435	For example, to force the pure perl model (L<AnyEvent::Impl::Perl>) you
		1436	could start your program like this:
		1437
		1438	PERL_ANYEVENT_MODEL=Perl perl ...
		1439
		1440	=item C<PERL_ANYEVENT_PROTOCOLS>
		1441
		1442	Used by both L<AnyEvent::DNS> and L<AnyEvent::Socket> to determine preferences
		1443	for IPv4 or IPv6. The default is unspecified (and might change, or be the result
		1444	of auto probing).
		1445
		1446	Must be set to a comma-separated list of protocols or address families,
		1447	current supported: C<ipv4> and C<ipv6>. Only protocols mentioned will be
		1448	used, and preference will be given to protocols mentioned earlier in the
		1449	list.
		1450
		1451	This variable can effectively be used for denial-of-service attacks
		1452	against local programs (e.g. when setuid), although the impact is likely
		1453	small, as the program has to handle conenction and other failures anyways.
		1454
		1455	Examples: C<PERL_ANYEVENT_PROTOCOLS=ipv4,ipv6> - prefer IPv4 over IPv6,
		1456	but support both and try to use both. C<PERL_ANYEVENT_PROTOCOLS=ipv4>
		1457	- only support IPv4, never try to resolve or contact IPv6
		1458	addresses. C<PERL_ANYEVENT_PROTOCOLS=ipv6,ipv4> support either IPv4 or
		1459	IPv6, but prefer IPv6 over IPv4.
		1460
		1461	=item C<PERL_ANYEVENT_EDNS0>
		1462
		1463	Used by L<AnyEvent::DNS> to decide whether to use the EDNS0 extension
		1464	for DNS. This extension is generally useful to reduce DNS traffic, but
		1465	some (broken) firewalls drop such DNS packets, which is why it is off by
		1466	default.
		1467
		1468	Setting this variable to C<1> will cause L<AnyEvent::DNS> to announce
		1469	EDNS0 in its DNS requests.
		1470
		1471	=item C<PERL_ANYEVENT_MAX_FORKS>
		1472
		1473	The maximum number of child processes that C<AnyEvent::Util::fork_call>
		1474	will create in parallel.
		1475
		1476	=item C<PERL_ANYEVENT_MAX_OUTSTANDING_DNS>
		1477
		1478	The default value for the C<max_outstanding> parameter for the default DNS
		1479	resolver - this is the maximum number of parallel DNS requests that are
		1480	sent to the DNS server.
		1481
		1482	=item C<PERL_ANYEVENT_RESOLV_CONF>
		1483
		1484	The file to use instead of F</etc/resolv.conf> (or OS-specific
		1485	configuration) in the default resolver. When set to the empty string, no
		1486	default config will be used.
		1487
		1488	=item C<PERL_ANYEVENT_CA_FILE>, C<PERL_ANYEVENT_CA_PATH>.
		1489
		1490	When neither C<ca_file> nor C<ca_path> was specified during
		1491	L<AnyEvent::TLS> context creation, and either of these environment
		1492	variables exist, they will be used to specify CA certificate locations
		1493	instead of a system-dependent default.
133		1494
134	=back	1495	=back
135		1496
136	=head1 SUPPLYING YOUR OWN EVENT MODEL INTERFACE	1497	=head1 SUPPLYING YOUR OWN EVENT MODEL INTERFACE
		1498
		1499	This is an advanced topic that you do not normally need to use AnyEvent in
		1500	a module. This section is only of use to event loop authors who want to
		1501	provide AnyEvent compatibility.
137		1502
138	If you need to support another event library which isn't directly	1503	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	1504	supported by AnyEvent, you can supply your own interface to it by
140	pushing, before the first watch gets created, the package name of	1505	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	1506	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	1507	C<@AnyEvent::REGISTRY>. You can do that before and even without loading
143	AnyEvent.	1508	AnyEvent, so it is reasonably cheap.
144		1509
145	Example:	1510	Example:
146		1511
147	push @AnyEvent::REGISTRY, [urxvt => urxvt::anyevent::];	1512	push @AnyEvent::REGISTRY, [urxvt => urxvt::anyevent::];
148		1513
149	This tells AnyEvent to (literally) use the C<urxvt::anyevent::> module	1514	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	1515	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		1516
		1517	When AnyEvent is loaded and asked to find a suitable event model, it
		1518	will first check for the presence of urxvt by trying to C<use> the
		1519	C<urxvt::anyevent> module.
		1520
		1521	The class should provide implementations for all watcher types. See
		1522	L<AnyEvent::Impl::EV> (source code), L<AnyEvent::Impl::Glib> (Source code)
		1523	and so on for actual examples. Use C<perldoc -m AnyEvent::Impl::Glib> to
		1524	see the sources.
		1525
		1526	If you don't provide C<signal> and C<child> watchers than AnyEvent will
		1527	provide suitable (hopefully) replacements.
		1528
154	The above isn't fictitious, the I<rxvt-unicode> (a.k.a. urxvt) uses	1529	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	1530	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	1531	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	1532	inside I<rxvt-unicode>, and it is updated and maintained as part of the
158	I<rxvt-unicode> distribution.	1533	I<rxvt-unicode> distribution.
159		1534
160	=head1 ENVIRONMENT VARIABLES	1535	I<rxvt-unicode> also cheats a bit by not providing blocking access to
		1536	condition variables: code blocking while waiting for a condition will
		1537	C<die>. This still works with most modules/usages, and blocking calls must
		1538	not be done in an interactive application, so it makes sense.
161		1539
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	1540	=head1 EXAMPLE PROGRAM
168		1541
169	The following program uses an io watcher to read data from stdin, a timer	1542	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	1543	to display a message once per second, and a condition variable to quit the
171	when the user enters quit:	1544	program when the user enters quit:
172		1545
173	use AnyEvent;	1546	use AnyEvent;
174		1547
175	my $cv = AnyEvent->condvar;	1548	my $cv = AnyEvent->condvar;
176		1549
177	my $io_watcher = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {	1550	my $io_watcher = AnyEvent->io (
		1551	fh => \*STDIN,
		1552	poll => 'r',
		1553	cb => sub {
178	warn "io event <$_[0]>\n"; # will always output <r>	1554	warn "io event <$_[0]>\n"; # will always output <r>
179	chomp (my $input = <STDIN>); # read a line	1555	chomp (my $input = <STDIN>); # read a line
180	warn "read: $input\n"; # output what has been read	1556	warn "read: $input\n"; # output what has been read
181	$cv->broadcast if $input =~ /^q/i; # quit program if /^q/i	1557	$cv->send if $input =~ /^q/i; # quit program if /^q/i
		1558	},
182	});	1559	);
183		1560
184	my $time_watcher; # can only be used once	1561	my $time_watcher; # can only be used once
185		1562
186	sub new_timer {	1563	sub new_timer {
187	$timer = AnyEvent->timer (after => 1, cb => sub {	1564	$timer = AnyEvent->timer (after => 1, cb => sub {
…		…
190	});	1567	});
191	}	1568	}
192		1569
193	new_timer; # create first timer	1570	new_timer; # create first timer
194		1571
195	$cv->wait; # wait until user enters /^q/i	1572	$cv->recv; # wait until user enters /^q/i
196		1573
197	=head1 REAL-WORLD EXAMPLE	1574	=head1 REAL-WORLD EXAMPLE
198		1575
199	Consider the L<Net::FCP> module. It features (among others) the following	1576	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:	1577	API calls, which are to freenet what HTTP GET requests are to http:
…		…
250	syswrite $txn->{fh}, $txn->{request}	1627	syswrite $txn->{fh}, $txn->{request}
251	or die "connection or write error";	1628	or die "connection or write error";
252	$txn->{w} = AnyEvent->io (fh => $txn->{fh}, poll => 'r', cb => sub { $txn->fh_ready_r });	1629	$txn->{w} = AnyEvent->io (fh => $txn->{fh}, poll => 'r', cb => sub { $txn->fh_ready_r });
253		1630
254	Again, C<fh_ready_r> waits till all data has arrived, and then stores the	1631	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:	1632	result and signals any possible waiters that the request has finished:
256		1633
257	sysread $txn->{fh}, $txn->{buf}, length $txn->{$buf};	1634	sysread $txn->{fh}, $txn->{buf}, length $txn->{$buf};
258		1635
259	if (end-of-file or data complete) {	1636	if (end-of-file or data complete) {
260	$txn->{result} = $txn->{buf};	1637	$txn->{result} = $txn->{buf};
261	$txn->{finished}->broadcast;	1638	$txn->{finished}->send;
262	$txb->{cb}->($txn) of $txn->{cb}; # also call callback	1639	$txb->{cb}->($txn) of $txn->{cb}; # also call callback
263	}	1640	}
264		1641
265	The C<result> method, finally, just waits for the finished signal (if the	1642	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	1643	request was already finished, it doesn't wait, of course, and returns the
267	data:	1644	data:
268		1645
269	$txn->{finished}->wait;	1646	$txn->{finished}->recv;
270	return $txn->{result};	1647	return $txn->{result};
271		1648
272	The actual code goes further and collects all errors (C<die>s, exceptions)	1649	The actual code goes further and collects all errors (C<die>s, exceptions)
273	that occured during request processing. The C<result> method detects	1650	that occurred during request processing. The C<result> method detects
274	wether an exception as thrown (it is stored inside the $txn object)	1651	whether an exception as thrown (it is stored inside the $txn object)
275	and just throws the exception, which means connection errors and other	1652	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	1653	problems get reported tot he code that tries to use the result, not in a
277	random callback.	1654	random callback.
278		1655
279	All of this enables the following usage styles:	1656	All of this enables the following usage styles:
280		1657
281	1. Blocking:	1658	1. Blocking:
282		1659
283	my $data = $fcp->client_get ($url);	1660	my $data = $fcp->client_get ($url);
284		1661
285	2. Blocking, but parallelizing:	1662	2. Blocking, but running in parallel:
286		1663
287	my @datas = map $_->result,	1664	my @datas = map $_->result,
288	map $fcp->txn_client_get ($_),	1665	map $fcp->txn_client_get ($_),
289	@urls;	1666	@urls;
290		1667
291	Both blocking examples work without the module user having to know	1668	Both blocking examples work without the module user having to know
292	anything about events.	1669	anything about events.
293		1670
294	3a. Event-based in a main program, using any support Event module:	1671	3a. Event-based in a main program, using any supported event module:
295		1672
296	use Event;	1673	use EV;
297		1674
298	$fcp->txn_client_get ($url)->cb (sub {	1675	$fcp->txn_client_get ($url)->cb (sub {
299	my $txn = shift;	1676	my $txn = shift;
300	my $data = $txn->result;	1677	my $data = $txn->result;
301	...	1678	...
302	});	1679	});
303		1680
304	Event::loop;	1681	EV::loop;
305		1682
306	3b. The module user could use AnyEvent, too:	1683	3b. The module user could use AnyEvent, too:
307		1684
308	use AnyEvent;	1685	use AnyEvent;
309		1686
310	my $quit = AnyEvent->condvar;	1687	my $quit = AnyEvent->condvar;
311		1688
312	$fcp->txn_client_get ($url)->cb (sub {	1689	$fcp->txn_client_get ($url)->cb (sub {
313	...	1690	...
314	$quit->broadcast;	1691	$quit->send;
315	});	1692	});
316		1693
317	$quit->wait;	1694	$quit->recv;
		1695
		1696
		1697	=head1 BENCHMARKS
		1698
		1699	To give you an idea of the performance and overheads that AnyEvent adds
		1700	over the event loops themselves and to give you an impression of the speed
		1701	of various event loops I prepared some benchmarks.
		1702
		1703	=head2 BENCHMARKING ANYEVENT OVERHEAD
		1704
		1705	Here is a benchmark of various supported event models used natively and
		1706	through AnyEvent. The benchmark creates a lot of timers (with a zero
		1707	timeout) and I/O watchers (watching STDOUT, a pty, to become writable,
		1708	which it is), lets them fire exactly once and destroys them again.
		1709
		1710	Source code for this benchmark is found as F<eg/bench> in the AnyEvent
		1711	distribution.
		1712
		1713	=head3 Explanation of the columns
		1714
		1715	I<watcher> is the number of event watchers created/destroyed. Since
		1716	different event models feature vastly different performances, each event
		1717	loop was given a number of watchers so that overall runtime is acceptable
		1718	and similar between tested event loop (and keep them from crashing): Glib
		1719	would probably take thousands of years if asked to process the same number
		1720	of watchers as EV in this benchmark.
		1721
		1722	I<bytes> is the number of bytes (as measured by the resident set size,
		1723	RSS) consumed by each watcher. This method of measuring captures both C
		1724	and Perl-based overheads.
		1725
		1726	I<create> is the time, in microseconds (millionths of seconds), that it
		1727	takes to create a single watcher. The callback is a closure shared between
		1728	all watchers, to avoid adding memory overhead. That means closure creation
		1729	and memory usage is not included in the figures.
		1730
		1731	I<invoke> is the time, in microseconds, used to invoke a simple
		1732	callback. The callback simply counts down a Perl variable and after it was
		1733	invoked "watcher" times, it would C<< ->send >> a condvar once to
		1734	signal the end of this phase.
		1735
		1736	I<destroy> is the time, in microseconds, that it takes to destroy a single
		1737	watcher.
		1738
		1739	=head3 Results
		1740
		1741	name watchers bytes create invoke destroy comment
		1742	EV/EV 400000 224 0.47 0.35 0.27 EV native interface
		1743	EV/Any 100000 224 2.88 0.34 0.27 EV + AnyEvent watchers
		1744	CoroEV/Any 100000 224 2.85 0.35 0.28 coroutines + Coro::Signal
		1745	Perl/Any 100000 452 4.13 0.73 0.95 pure perl implementation
		1746	Event/Event 16000 517 32.20 31.80 0.81 Event native interface
		1747	Event/Any 16000 590 35.85 31.55 1.06 Event + AnyEvent watchers
		1748	IOAsync/Any 16000 989 38.10 32.77 11.13 via IO::Async::Loop::IO_Poll
		1749	IOAsync/Any 16000 990 37.59 29.50 10.61 via IO::Async::Loop::Epoll
		1750	Glib/Any 16000 1357 102.33 12.31 51.00 quadratic behaviour
		1751	Tk/Any 2000 1860 27.20 66.31 14.00 SEGV with >> 2000 watchers
		1752	POE/Event 2000 6328 109.99 751.67 14.02 via POE::Loop::Event
		1753	POE/Select 2000 6027 94.54 809.13 579.80 via POE::Loop::Select
		1754
		1755	=head3 Discussion
		1756
		1757	The benchmark does I<not> measure scalability of the event loop very
		1758	well. For example, a select-based event loop (such as the pure perl one)
		1759	can never compete with an event loop that uses epoll when the number of
		1760	file descriptors grows high. In this benchmark, all events become ready at
		1761	the same time, so select/poll-based implementations get an unnatural speed
		1762	boost.
		1763
		1764	Also, note that the number of watchers usually has a nonlinear effect on
		1765	overall speed, that is, creating twice as many watchers doesn't take twice
		1766	the time - usually it takes longer. This puts event loops tested with a
		1767	higher number of watchers at a disadvantage.
		1768
		1769	To put the range of results into perspective, consider that on the
		1770	benchmark machine, handling an event takes roughly 1600 CPU cycles with
		1771	EV, 3100 CPU cycles with AnyEvent's pure perl loop and almost 3000000 CPU
		1772	cycles with POE.
		1773
		1774	C<EV> is the sole leader regarding speed and memory use, which are both
		1775	maximal/minimal, respectively. Even when going through AnyEvent, it uses
		1776	far less memory than any other event loop and is still faster than Event
		1777	natively.
		1778
		1779	The pure perl implementation is hit in a few sweet spots (both the
		1780	constant timeout and the use of a single fd hit optimisations in the perl
		1781	interpreter and the backend itself). Nevertheless this shows that it
		1782	adds very little overhead in itself. Like any select-based backend its
		1783	performance becomes really bad with lots of file descriptors (and few of
		1784	them active), of course, but this was not subject of this benchmark.
		1785
		1786	The C<Event> module has a relatively high setup and callback invocation
		1787	cost, but overall scores in on the third place.
		1788
		1789	C<IO::Async> performs admirably well, about on par with C<Event>, even
		1790	when using its pure perl backend.
		1791
		1792	C<Glib>'s memory usage is quite a bit higher, but it features a
		1793	faster callback invocation and overall ends up in the same class as
		1794	C<Event>. However, Glib scales extremely badly, doubling the number of
		1795	watchers increases the processing time by more than a factor of four,
		1796	making it completely unusable when using larger numbers of watchers
		1797	(note that only a single file descriptor was used in the benchmark, so
		1798	inefficiencies of C<poll> do not account for this).
		1799
		1800	The C<Tk> adaptor works relatively well. The fact that it crashes with
		1801	more than 2000 watchers is a big setback, however, as correctness takes
		1802	precedence over speed. Nevertheless, its performance is surprising, as the
		1803	file descriptor is dup()ed for each watcher. This shows that the dup()
		1804	employed by some adaptors is not a big performance issue (it does incur a
		1805	hidden memory cost inside the kernel which is not reflected in the figures
		1806	above).
		1807
		1808	C<POE>, regardless of underlying event loop (whether using its pure perl
		1809	select-based backend or the Event module, the POE-EV backend couldn't
		1810	be tested because it wasn't working) shows abysmal performance and
		1811	memory usage with AnyEvent: Watchers use almost 30 times as much memory
		1812	as EV watchers, and 10 times as much memory as Event (the high memory
		1813	requirements are caused by requiring a session for each watcher). Watcher
		1814	invocation speed is almost 900 times slower than with AnyEvent's pure perl
		1815	implementation.
		1816
		1817	The design of the POE adaptor class in AnyEvent can not really account
		1818	for the performance issues, though, as session creation overhead is
		1819	small compared to execution of the state machine, which is coded pretty
		1820	optimally within L<AnyEvent::Impl::POE> (and while everybody agrees that
		1821	using multiple sessions is not a good approach, especially regarding
		1822	memory usage, even the author of POE could not come up with a faster
		1823	design).
		1824
		1825	=head3 Summary
		1826
		1827	=over 4
		1828
		1829	=item * Using EV through AnyEvent is faster than any other event loop
		1830	(even when used without AnyEvent), but most event loops have acceptable
		1831	performance with or without AnyEvent.
		1832
		1833	=item * The overhead AnyEvent adds is usually much smaller than the overhead of
		1834	the actual event loop, only with extremely fast event loops such as EV
		1835	adds AnyEvent significant overhead.
		1836
		1837	=item * You should avoid POE like the plague if you want performance or
		1838	reasonable memory usage.
		1839
		1840	=back
		1841
		1842	=head2 BENCHMARKING THE LARGE SERVER CASE
		1843
		1844	This benchmark actually benchmarks the event loop itself. It works by
		1845	creating a number of "servers": each server consists of a socket pair, a
		1846	timeout watcher that gets reset on activity (but never fires), and an I/O
		1847	watcher waiting for input on one side of the socket. Each time the socket
		1848	watcher reads a byte it will write that byte to a random other "server".
		1849
		1850	The effect is that there will be a lot of I/O watchers, only part of which
		1851	are active at any one point (so there is a constant number of active
		1852	fds for each loop iteration, but which fds these are is random). The
		1853	timeout is reset each time something is read because that reflects how
		1854	most timeouts work (and puts extra pressure on the event loops).
		1855
		1856	In this benchmark, we use 10000 socket pairs (20000 sockets), of which 100
		1857	(1%) are active. This mirrors the activity of large servers with many
		1858	connections, most of which are idle at any one point in time.
		1859
		1860	Source code for this benchmark is found as F<eg/bench2> in the AnyEvent
		1861	distribution.
		1862
		1863	=head3 Explanation of the columns
		1864
		1865	I<sockets> is the number of sockets, and twice the number of "servers" (as
		1866	each server has a read and write socket end).
		1867
		1868	I<create> is the time it takes to create a socket pair (which is
		1869	nontrivial) and two watchers: an I/O watcher and a timeout watcher.
		1870
		1871	I<request>, the most important value, is the time it takes to handle a
		1872	single "request", that is, reading the token from the pipe and forwarding
		1873	it to another server. This includes deleting the old timeout and creating
		1874	a new one that moves the timeout into the future.
		1875
		1876	=head3 Results
		1877
		1878	name sockets create request
		1879	EV 20000 69.01 11.16
		1880	Perl 20000 73.32 35.87
		1881	IOAsync 20000 157.00 98.14 epoll
		1882	IOAsync 20000 159.31 616.06 poll
		1883	Event 20000 212.62 257.32
		1884	Glib 20000 651.16 1896.30
		1885	POE 20000 349.67 12317.24 uses POE::Loop::Event
		1886
		1887	=head3 Discussion
		1888
		1889	This benchmark I<does> measure scalability and overall performance of the
		1890	particular event loop.
		1891
		1892	EV is again fastest. Since it is using epoll on my system, the setup time
		1893	is relatively high, though.
		1894
		1895	Perl surprisingly comes second. It is much faster than the C-based event
		1896	loops Event and Glib.
		1897
		1898	IO::Async performs very well when using its epoll backend, and still quite
		1899	good compared to Glib when using its pure perl backend.
		1900
		1901	Event suffers from high setup time as well (look at its code and you will
		1902	understand why). Callback invocation also has a high overhead compared to
		1903	the C<< $_->() for .. >>-style loop that the Perl event loop uses. Event
		1904	uses select or poll in basically all documented configurations.
		1905
		1906	Glib is hit hard by its quadratic behaviour w.r.t. many watchers. It
		1907	clearly fails to perform with many filehandles or in busy servers.
		1908
		1909	POE is still completely out of the picture, taking over 1000 times as long
		1910	as EV, and over 100 times as long as the Perl implementation, even though
		1911	it uses a C-based event loop in this case.
		1912
		1913	=head3 Summary
		1914
		1915	=over 4
		1916
		1917	=item * The pure perl implementation performs extremely well.
		1918
		1919	=item * Avoid Glib or POE in large projects where performance matters.
		1920
		1921	=back
		1922
		1923	=head2 BENCHMARKING SMALL SERVERS
		1924
		1925	While event loops should scale (and select-based ones do not...) even to
		1926	large servers, most programs we (or I :) actually write have only a few
		1927	I/O watchers.
		1928
		1929	In this benchmark, I use the same benchmark program as in the large server
		1930	case, but it uses only eight "servers", of which three are active at any
		1931	one time. This should reflect performance for a small server relatively
		1932	well.
		1933
		1934	The columns are identical to the previous table.
		1935
		1936	=head3 Results
		1937
		1938	name sockets create request
		1939	EV 16 20.00 6.54
		1940	Perl 16 25.75 12.62
		1941	Event 16 81.27 35.86
		1942	Glib 16 32.63 15.48
		1943	POE 16 261.87 276.28 uses POE::Loop::Event
		1944
		1945	=head3 Discussion
		1946
		1947	The benchmark tries to test the performance of a typical small
		1948	server. While knowing how various event loops perform is interesting, keep
		1949	in mind that their overhead in this case is usually not as important, due
		1950	to the small absolute number of watchers (that is, you need efficiency and
		1951	speed most when you have lots of watchers, not when you only have a few of
		1952	them).
		1953
		1954	EV is again fastest.
		1955
		1956	Perl again comes second. It is noticeably faster than the C-based event
		1957	loops Event and Glib, although the difference is too small to really
		1958	matter.
		1959
		1960	POE also performs much better in this case, but is is still far behind the
		1961	others.
		1962
		1963	=head3 Summary
		1964
		1965	=over 4
		1966
		1967	=item * C-based event loops perform very well with small number of
		1968	watchers, as the management overhead dominates.
		1969
		1970	=back
		1971
		1972	=head2 THE IO::Lambda BENCHMARK
		1973
		1974	Recently I was told about the benchmark in the IO::Lambda manpage, which
		1975	could be misinterpreted to make AnyEvent look bad. In fact, the benchmark
		1976	simply compares IO::Lambda with POE, and IO::Lambda looks better (which
		1977	shouldn't come as a surprise to anybody). As such, the benchmark is
		1978	fine, and mostly shows that the AnyEvent backend from IO::Lambda isn't
		1979	very optimal. But how would AnyEvent compare when used without the extra
		1980	baggage? To explore this, I wrote the equivalent benchmark for AnyEvent.
		1981
		1982	The benchmark itself creates an echo-server, and then, for 500 times,
		1983	connects to the echo server, sends a line, waits for the reply, and then
		1984	creates the next connection. This is a rather bad benchmark, as it doesn't
		1985	test the efficiency of the framework or much non-blocking I/O, but it is a
		1986	benchmark nevertheless.
		1987
		1988	name runtime
		1989	Lambda/select 0.330 sec
		1990	+ optimized 0.122 sec
		1991	Lambda/AnyEvent 0.327 sec
		1992	+ optimized 0.138 sec
		1993	Raw sockets/select 0.077 sec
		1994	POE/select, components 0.662 sec
		1995	POE/select, raw sockets 0.226 sec
		1996	POE/select, optimized 0.404 sec
		1997
		1998	AnyEvent/select/nb 0.085 sec
		1999	AnyEvent/EV/nb 0.068 sec
		2000	+state machine 0.134 sec
		2001
		2002	The benchmark is also a bit unfair (my fault): the IO::Lambda/POE
		2003	benchmarks actually make blocking connects and use 100% blocking I/O,
		2004	defeating the purpose of an event-based solution. All of the newly
		2005	written AnyEvent benchmarks use 100% non-blocking connects (using
		2006	AnyEvent::Socket::tcp_connect and the asynchronous pure perl DNS
		2007	resolver), so AnyEvent is at a disadvantage here, as non-blocking connects
		2008	generally require a lot more bookkeeping and event handling than blocking
		2009	connects (which involve a single syscall only).
		2010
		2011	The last AnyEvent benchmark additionally uses L<AnyEvent::Handle>, which
		2012	offers similar expressive power as POE and IO::Lambda, using conventional
		2013	Perl syntax. This means that both the echo server and the client are 100%
		2014	non-blocking, further placing it at a disadvantage.
		2015
		2016	As you can see, the AnyEvent + EV combination even beats the
		2017	hand-optimised "raw sockets benchmark", while AnyEvent + its pure perl
		2018	backend easily beats IO::Lambda and POE.
		2019
		2020	And even the 100% non-blocking version written using the high-level (and
		2021	slow :) L<AnyEvent::Handle> abstraction beats both POE and IO::Lambda by a
		2022	large margin, even though it does all of DNS, tcp-connect and socket I/O
		2023	in a non-blocking way.
		2024
		2025	The two AnyEvent benchmarks programs can be found as F<eg/ae0.pl> and
		2026	F<eg/ae2.pl> in the AnyEvent distribution, the remaining benchmarks are
		2027	part of the IO::lambda distribution and were used without any changes.
		2028
		2029
		2030	=head1 SIGNALS
		2031
		2032	AnyEvent currently installs handlers for these signals:
		2033
		2034	=over 4
		2035
		2036	=item SIGCHLD
		2037
		2038	A handler for C<SIGCHLD> is installed by AnyEvent's child watcher
		2039	emulation for event loops that do not support them natively. Also, some
		2040	event loops install a similar handler.
		2041
		2042	If, when AnyEvent is loaded, SIGCHLD is set to IGNORE, then AnyEvent will
		2043	reset it to default, to avoid losing child exit statuses.
		2044
		2045	=item SIGPIPE
		2046
		2047	A no-op handler is installed for C<SIGPIPE> when C<$SIG{PIPE}> is C<undef>
		2048	when AnyEvent gets loaded.
		2049
		2050	The rationale for this is that AnyEvent users usually do not really depend
		2051	on SIGPIPE delivery (which is purely an optimisation for shell use, or
		2052	badly-written programs), but C<SIGPIPE> can cause spurious and rare
		2053	program exits as a lot of people do not expect C<SIGPIPE> when writing to
		2054	some random socket.
		2055
		2056	The rationale for installing a no-op handler as opposed to ignoring it is
		2057	that this way, the handler will be restored to defaults on exec.
		2058
		2059	Feel free to install your own handler, or reset it to defaults.
		2060
		2061	=back
		2062
		2063	=cut
		2064
		2065	undef $SIG{CHLD}
		2066	if $SIG{CHLD} eq 'IGNORE';
		2067
		2068	$SIG{PIPE} = sub { }
		2069	unless defined $SIG{PIPE};
		2070
		2071	=head1 FORK
		2072
		2073	Most event libraries are not fork-safe. The ones who are usually are
		2074	because they rely on inefficient but fork-safe C<select> or C<poll>
		2075	calls. Only L<EV> is fully fork-aware.
		2076
		2077	If you have to fork, you must either do so I<before> creating your first
		2078	watcher OR you must not use AnyEvent at all in the child.
		2079
		2080
		2081	=head1 SECURITY CONSIDERATIONS
		2082
		2083	AnyEvent can be forced to load any event model via
		2084	$ENV{PERL_ANYEVENT_MODEL}. While this cannot (to my knowledge) be used to
		2085	execute arbitrary code or directly gain access, it can easily be used to
		2086	make the program hang or malfunction in subtle ways, as AnyEvent watchers
		2087	will not be active when the program uses a different event model than
		2088	specified in the variable.
		2089
		2090	You can make AnyEvent completely ignore this variable by deleting it
		2091	before the first watcher gets created, e.g. with a C<BEGIN> block:
		2092
		2093	BEGIN { delete $ENV{PERL_ANYEVENT_MODEL} }
		2094
		2095	use AnyEvent;
		2096
		2097	Similar considerations apply to $ENV{PERL_ANYEVENT_VERBOSE}, as that can
		2098	be used to probe what backend is used and gain other information (which is
		2099	probably even less useful to an attacker than PERL_ANYEVENT_MODEL), and
		2100	$ENV{PERL_ANYEVENT_STRICT}.
		2101
		2102	Note that AnyEvent will remove I<all> environment variables starting with
		2103	C<PERL_ANYEVENT_> from C<%ENV> when it is loaded while taint mode is
		2104	enabled.
		2105
		2106
		2107	=head1 BUGS
		2108
		2109	Perl 5.8 has numerous memleaks that sometimes hit this module and are hard
		2110	to work around. If you suffer from memleaks, first upgrade to Perl 5.10
		2111	and check wether the leaks still show up. (Perl 5.10.0 has other annoying
		2112	memleaks, such as leaking on C<map> and C<grep> but it is usually not as
		2113	pronounced).
		2114
318		2115
319	=head1 SEE ALSO	2116	=head1 SEE ALSO
320		2117
321	Event modules: L<Coro::Event>, L<Coro>, L<Event>, L<Glib::Event>, L<Glib>.	2118	Utility functions: L<AnyEvent::Util>.
322		2119
323	Implementations: L<AnyEvent::Impl::Coro>, L<AnyEvent::Impl::Event>, L<AnyEvent::Impl::Glib>, L<AnyEvent::Impl::Tk>.	2120	Event modules: L<EV>, L<EV::Glib>, L<Glib::EV>, L<Event>, L<Glib::Event>,
		2121	L<Glib>, L<Tk>, L<Event::Lib>, L<Qt>, L<POE>.
324		2122
325	Nontrivial usage example: L<Net::FCP>.	2123	Implementations: L<AnyEvent::Impl::EV>, L<AnyEvent::Impl::Event>,
		2124	L<AnyEvent::Impl::Glib>, L<AnyEvent::Impl::Tk>, L<AnyEvent::Impl::Perl>,
		2125	L<AnyEvent::Impl::EventLib>, L<AnyEvent::Impl::Qt>,
		2126	L<AnyEvent::Impl::POE>.
326		2127
327	=head1	2128	Non-blocking file handles, sockets, TCP clients and
		2129	servers: L<AnyEvent::Handle>, L<AnyEvent::Socket>.
		2130
		2131	Asynchronous DNS: L<AnyEvent::DNS>.
		2132
		2133	Coroutine support: L<Coro>, L<Coro::AnyEvent>, L<Coro::EV>, L<Coro::Event>,
		2134
		2135	Nontrivial usage examples: L<Net::FCP>, L<Net::XMPP2>, L<AnyEvent::DNS>.
		2136
		2137
		2138	=head1 AUTHOR
		2139
		2140	Marc Lehmann <schmorp@schmorp.de>
		2141	http://home.schmorp.de/
328		2142
329	=cut	2143	=cut
330		2144
331	1	2145	1
332		2146

Diff Legend

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