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 1.296 by root, Tue Nov 17 01:19:49 2009 UTC

1	=head1 NAME	1	=head1 NAME
2		2
3	AnyEvent - provide framework for multiple event loops	3	AnyEvent - the DBI of event loop programming
4		4
5	Event, Coro, Glib, Tk - various supported event loops	5	EV, Event, Glib, Tk, Perl, Event::Lib, Irssi, rxvt-unicode, IO::Async, Qt
		6	and POE are various supported event loops/environments.
6		7
7	=head1 SYNOPSIS	8	=head1 SYNOPSIS
8		9
9	use AnyEvent;	10	use AnyEvent;
10		11
		12	# 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 SUPPORT
		47
		48	There is a mailinglist for discussing all things AnyEvent, and an IRC
		49	channel, too.
		50
		51	See the AnyEvent project page at the B<Schmorpforge Ta-Sa Software
		52	Repository>, at L<http://anyevent.schmorp.de>, for more info.
		53
		54	=head1 WHY YOU SHOULD USE THIS MODULE (OR NOT)
		55
		56	Glib, POE, IO::Async, Event... CPAN offers event models by the dozen
		57	nowadays. So what is different about AnyEvent?
		58
		59	Executive Summary: AnyEvent is I<compatible>, AnyEvent is I<free of
		60	policy> and AnyEvent is I<small and efficient>.
		61
		62	First and foremost, I<AnyEvent is not an event model> itself, it only
		63	interfaces to whatever event model the main program happens to use, in a
		64	pragmatic way. For event models and certain classes of immortals alike,
		65	the statement "there can only be one" is a bitter reality: In general,
		66	only one event loop can be active at the same time in a process. AnyEvent
		67	cannot change this, but it can hide the differences between those event
		68	loops.
		69
		70	The goal of AnyEvent is to offer module authors the ability to do event
		71	programming (waiting for I/O or timer events) without subscribing to a
		72	religion, a way of living, and most importantly: without forcing your
		73	module users into the same thing by forcing them to use the same event
		74	model you use.
		75
		76	For modules like POE or IO::Async (which is a total misnomer as it is
		77	actually doing all I/O I<synchronously>...), using them in your module is
		78	like joining a cult: After you joined, you are dependent on them and you
		79	cannot use anything else, as they are simply incompatible to everything
		80	that isn't them. What's worse, all the potential users of your
		81	module are I<also> forced to use the same event loop you use.
		82
		83	AnyEvent is different: AnyEvent + POE works fine. AnyEvent + Glib works
		84	fine. AnyEvent + Tk works fine etc. etc. but none of these work together
		85	with the rest: POE + IO::Async? No go. Tk + Event? No go. Again: if
		86	your module uses one of those, every user of your module has to use it,
		87	too. But if your module uses AnyEvent, it works transparently with all
		88	event models it supports (including stuff like IO::Async, as long as those
		89	use one of the supported event loops. It is trivial to add new event loops
		90	to AnyEvent, too, so it is future-proof).
		91
		92	In addition to being free of having to use I<the one and only true event
		93	model>, AnyEvent also is free of bloat and policy: with POE or similar
		94	modules, you get an enormous amount of code and strict rules you have to
		95	follow. AnyEvent, on the other hand, is lean and up to the point, by only
		96	offering the functionality that is necessary, in as thin as a wrapper as
		97	technically possible.
		98
		99	Of course, AnyEvent comes with a big (and fully optional!) toolbox
		100	of useful functionality, such as an asynchronous DNS resolver, 100%
		101	non-blocking connects (even with TLS/SSL, IPv6 and on broken platforms
		102	such as Windows) and lots of real-world knowledge and workarounds for
		103	platform bugs and differences.
		104
		105	Now, if you I<do want> lots of policy (this can arguably be somewhat
		106	useful) and you want to force your users to use the one and only event
		107	model, you should I<not> use this module.
45		108
46	=head1 DESCRIPTION	109	=head1 DESCRIPTION
47		110
48	L<AnyEvent> provides an identical interface to multiple event loops. This	111	L<AnyEvent> provides an identical interface to multiple event loops. This
49	allows module authors to utilizy an event loop without forcing module	112	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	113	users to use the same event loop (as only a single event loop can coexist
51	peacefully at any one time).	114	peacefully at any one time).
52		115
53	The interface itself is vaguely similar but not identical to the Event	116	The interface itself is vaguely similar, but not identical to the L<Event>
54	module.	117	module.
55		118
56	On the first call of any method, the module tries to detect the currently	119	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	120	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	121	following modules is already loaded: L<EV>,
59	used. If none is found, the module tries to load these modules in the	122	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	123	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.	124	to load these modules (excluding Tk, Event::Lib, Qt and POE as the pure perl
		125	adaptor should always succeed) in the order given. The first one that can
		126	be successfully loaded will be used. If, after this, still none could be
		127	found, AnyEvent will fall back to a pure-perl event loop, which is not
		128	very efficient, but should work everywhere.
		129
		130	Because AnyEvent first checks for modules that are already loaded, loading
		131	an event model explicitly before first using AnyEvent will likely make
		132	that model the default. For example:
		133
		134	use Tk;
		135	use AnyEvent;
		136
		137	# .. AnyEvent will likely default to Tk
		138
		139	The I<likely> means that, if any module loads another event model and
		140	starts using it, all bets are off. Maybe you should tell their authors to
		141	use AnyEvent so their modules work together with others seamlessly...
		142
		143	The pure-perl implementation of AnyEvent is called
		144	C<AnyEvent::Impl::Perl>. Like other event modules you can load it
		145	explicitly and enjoy the high availability of that event loop :)
		146
		147	=head1 WATCHERS
		148
		149	AnyEvent has the central concept of a I<watcher>, which is an object that
		150	stores relevant data for each kind of event you are waiting for, such as
		151	the callback to call, the file handle to watch, etc.
		152
		153	These watchers are normal Perl objects with normal Perl lifetime. After
		154	creating a watcher it will immediately "watch" for events and invoke the
		155	callback when the event occurs (of course, only when the event model
		156	is in control).
		157
		158	Note that B<callbacks must not permanently change global variables>
		159	potentially in use by the event loop (such as C<$_> or C<$[>) and that B<<
		160	callbacks must not C<die> >>. The former is good programming practise in
		161	Perl and the latter stems from the fact that exception handling differs
		162	widely between event loops.
		163
		164	To disable the watcher you have to destroy it (e.g. by setting the
		165	variable you store it in to C<undef> or otherwise deleting all references
		166	to it).
		167
		168	All watchers are created by calling a method on the C<AnyEvent> class.
		169
		170	Many watchers either are used with "recursion" (repeating timers for
		171	example), or need to refer to their watcher object in other ways.
		172
		173	An any way to achieve that is this pattern:
		174
		175	my $w; $w = AnyEvent->type (arg => value ..., cb => sub {
		176	# you can use $w here, for example to undef it
		177	undef $w;
		178	});
		179
		180	Note that C<my $w; $w => combination. This is necessary because in Perl,
		181	my variables are only visible after the statement in which they are
		182	declared.
		183
		184	=head2 I/O WATCHERS
		185
		186	$w = AnyEvent->io (
		187	fh => <filehandle_or_fileno>,
		188	poll => <"r" or "w">,
		189	cb => <callback>,
		190	);
		191
		192	You can create an I/O watcher by calling the C<< AnyEvent->io >> method
		193	with the following mandatory key-value pairs as arguments:
		194
		195	C<fh> is the Perl I<file handle> (or a naked file descriptor) to watch
		196	for events (AnyEvent might or might not keep a reference to this file
		197	handle). Note that only file handles pointing to things for which
		198	non-blocking operation makes sense are allowed. This includes sockets,
		199	most character devices, pipes, fifos and so on, but not for example files
		200	or block devices.
		201
		202	C<poll> must be a string that is either C<r> or C<w>, which creates a
		203	watcher waiting for "r"eadable or "w"ritable events, respectively.
		204
		205	C<cb> is the callback to invoke each time the file handle becomes ready.
		206
		207	Although the callback might get passed parameters, their value and
		208	presence is undefined and you cannot rely on them. Portable AnyEvent
		209	callbacks cannot use arguments passed to I/O watcher callbacks.
		210
		211	The I/O watcher might use the underlying file descriptor or a copy of it.
		212	You must not close a file handle as long as any watcher is active on the
		213	underlying file descriptor.
		214
		215	Some event loops issue spurious readyness notifications, so you should
		216	always use non-blocking calls when reading/writing from/to your file
		217	handles.
		218
		219	Example: wait for readability of STDIN, then read a line and disable the
		220	watcher.
		221
		222	my $w; $w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {
		223	chomp (my $input = <STDIN>);
		224	warn "read: $input\n";
		225	undef $w;
		226	});
		227
		228	=head2 TIME WATCHERS
		229
		230	$w = AnyEvent->timer (after => <seconds>, cb => <callback>);
		231
		232	$w = AnyEvent->timer (
		233	after => <fractional_seconds>,
		234	interval => <fractional_seconds>,
		235	cb => <callback>,
		236	);
		237
		238	You can create a time watcher by calling the C<< AnyEvent->timer >>
		239	method with the following mandatory arguments:
		240
		241	C<after> specifies after how many seconds (fractional values are
		242	supported) the callback should be invoked. C<cb> is the callback to invoke
		243	in that case.
		244
		245	Although the callback might get passed parameters, their value and
		246	presence is undefined and you cannot rely on them. Portable AnyEvent
		247	callbacks cannot use arguments passed to time watcher callbacks.
		248
		249	The callback will normally be invoked once only. If you specify another
		250	parameter, C<interval>, as a strictly positive number (> 0), then the
		251	callback will be invoked regularly at that interval (in fractional
		252	seconds) after the first invocation. If C<interval> is specified with a
		253	false value, then it is treated as if it were missing.
		254
		255	The callback will be rescheduled before invoking the callback, but no
		256	attempt is done to avoid timer drift in most backends, so the interval is
		257	only approximate.
		258
		259	Example: fire an event after 7.7 seconds.
		260
		261	my $w = AnyEvent->timer (after => 7.7, cb => sub {
		262	warn "timeout\n";
		263	});
		264
		265	# to cancel the timer:
		266	undef $w;
		267
		268	Example 2: fire an event after 0.5 seconds, then roughly every second.
		269
		270	my $w = AnyEvent->timer (after => 0.5, interval => 1, cb => sub {
		271	warn "timeout\n";
		272	};
		273
		274	=head3 TIMING ISSUES
		275
		276	There are two ways to handle timers: based on real time (relative, "fire
		277	in 10 seconds") and based on wallclock time (absolute, "fire at 12
		278	o'clock").
		279
		280	While most event loops expect timers to specified in a relative way, they
		281	use absolute time internally. This makes a difference when your clock
		282	"jumps", for example, when ntp decides to set your clock backwards from
		283	the wrong date of 2014-01-01 to 2008-01-01, a watcher that is supposed to
		284	fire "after" a second might actually take six years to finally fire.
		285
		286	AnyEvent cannot compensate for this. The only event loop that is conscious
		287	about these issues is L<EV>, which offers both relative (ev_timer, based
		288	on true relative time) and absolute (ev_periodic, based on wallclock time)
		289	timers.
		290
		291	AnyEvent always prefers relative timers, if available, matching the
		292	AnyEvent API.
		293
		294	AnyEvent has two additional methods that return the "current time":
62		295
63	=over 4	296	=over 4
64		297
		298	=item AnyEvent->time
		299
		300	This returns the "current wallclock time" as a fractional number of
		301	seconds since the Epoch (the same thing as C<time> or C<Time::HiRes::time>
		302	return, and the result is guaranteed to be compatible with those).
		303
		304	It progresses independently of any event loop processing, i.e. each call
		305	will check the system clock, which usually gets updated frequently.
		306
		307	=item AnyEvent->now
		308
		309	This also returns the "current wallclock time", but unlike C<time>, above,
		310	this value might change only once per event loop iteration, depending on
		311	the event loop (most return the same time as C<time>, above). This is the
		312	time that AnyEvent's timers get scheduled against.
		313
		314	I<In almost all cases (in all cases if you don't care), this is the
		315	function to call when you want to know the current time.>
		316
		317	This function is also often faster then C<< AnyEvent->time >>, and
		318	thus the preferred method if you want some timestamp (for example,
		319	L<AnyEvent::Handle> uses this to update it's activity timeouts).
		320
		321	The rest of this section is only of relevance if you try to be very exact
		322	with your timing, you can skip it without bad conscience.
		323
		324	For a practical example of when these times differ, consider L<Event::Lib>
		325	and L<EV> and the following set-up:
		326
		327	The event loop is running and has just invoked one of your callback at
		328	time=500 (assume no other callbacks delay processing). In your callback,
		329	you wait a second by executing C<sleep 1> (blocking the process for a
		330	second) and then (at time=501) you create a relative timer that fires
		331	after three seconds.
		332
		333	With L<Event::Lib>, C<< AnyEvent->time >> and C<< AnyEvent->now >> will
		334	both return C<501>, because that is the current time, and the timer will
		335	be scheduled to fire at time=504 (C<501> + C<3>).
		336
		337	With L<EV>, C<< AnyEvent->time >> returns C<501> (as that is the current
		338	time), but C<< AnyEvent->now >> returns C<500>, as that is the time the
		339	last event processing phase started. With L<EV>, your timer gets scheduled
		340	to run at time=503 (C<500> + C<3>).
		341
		342	In one sense, L<Event::Lib> is more exact, as it uses the current time
		343	regardless of any delays introduced by event processing. However, most
		344	callbacks do not expect large delays in processing, so this causes a
		345	higher drift (and a lot more system calls to get the current time).
		346
		347	In another sense, L<EV> is more exact, as your timer will be scheduled at
		348	the same time, regardless of how long event processing actually took.
		349
		350	In either case, if you care (and in most cases, you don't), then you
		351	can get whatever behaviour you want with any event loop, by taking the
		352	difference between C<< AnyEvent->time >> and C<< AnyEvent->now >> into
		353	account.
		354
		355	=item AnyEvent->now_update
		356
		357	Some event loops (such as L<EV> or L<AnyEvent::Impl::Perl>) cache
		358	the current time for each loop iteration (see the discussion of L<<
		359	AnyEvent->now >>, above).
		360
		361	When a callback runs for a long time (or when the process sleeps), then
		362	this "current" time will differ substantially from the real time, which
		363	might affect timers and time-outs.
		364
		365	When this is the case, you can call this method, which will update the
		366	event loop's idea of "current time".
		367
		368	A typical example would be a script in a web server (e.g. C<mod_perl>) -
		369	when mod_perl executes the script, then the event loop will have the wrong
		370	idea about the "current time" (being potentially far in the past, when the
		371	script ran the last time). In that case you should arrange a call to C<<
		372	AnyEvent->now_update >> each time the web server process wakes up again
		373	(e.g. at the start of your script, or in a handler).
		374
		375	Note that updating the time I<might> cause some events to be handled.
		376
		377	=back
		378
		379	=head2 SIGNAL WATCHERS
		380
		381	$w = AnyEvent->signal (signal => <uppercase_signal_name>, cb => <callback>);
		382
		383	You can watch for signals using a signal watcher, C<signal> is the signal
		384	I<name> in uppercase and without any C<SIG> prefix, C<cb> is the Perl
		385	callback to be invoked whenever a signal occurs.
		386
		387	Although the callback might get passed parameters, their value and
		388	presence is undefined and you cannot rely on them. Portable AnyEvent
		389	callbacks cannot use arguments passed to signal watcher callbacks.
		390
		391	Multiple signal occurrences can be clumped together into one callback
		392	invocation, and callback invocation will be synchronous. Synchronous means
		393	that it might take a while until the signal gets handled by the process,
		394	but it is guaranteed not to interrupt any other callbacks.
		395
		396	The main advantage of using these watchers is that you can share a signal
		397	between multiple watchers, and AnyEvent will ensure that signals will not
		398	interrupt your program at bad times.
		399
		400	This watcher might use C<%SIG> (depending on the event loop used),
		401	so programs overwriting those signals directly will likely not work
		402	correctly.
		403
		404	Example: exit on SIGINT
		405
		406	my $w = AnyEvent->signal (signal => "INT", cb => sub { exit 1 });
		407
		408	=head3 Signal Races, Delays and Workarounds
		409
		410	Many event loops (e.g. Glib, Tk, Qt, IO::Async) do not support attaching
		411	callbacks to signals in a generic way, which is a pity, as you cannot
		412	do race-free signal handling in perl, requiring C libraries for
		413	this. AnyEvent will try to do it's best, which means in some cases,
		414	signals will be delayed. The maximum time a signal might be delayed is
		415	specified in C<$AnyEvent::MAX_SIGNAL_LATENCY> (default: 10 seconds). This
		416	variable can be changed only before the first signal watcher is created,
		417	and should be left alone otherwise. This variable determines how often
		418	AnyEvent polls for signals (in case a wake-up was missed). Higher values
		419	will cause fewer spurious wake-ups, which is better for power and CPU
		420	saving.
		421
		422	All these problems can be avoided by installing the optional
		423	L<Async::Interrupt> module, which works with most event loops. It will not
		424	work with inherently broken event loops such as L<Event> or L<Event::Lib>
		425	(and not with L<POE> currently, as POE does it's own workaround with
		426	one-second latency). For those, you just have to suffer the delays.
		427
		428	=head2 CHILD PROCESS WATCHERS
		429
		430	$w = AnyEvent->child (pid => <process id>, cb => <callback>);
		431
		432	You can also watch on a child process exit and catch its exit status.
		433
		434	The child process is specified by the C<pid> argument (one some backends,
		435	using C<0> watches for any child process exit, on others this will
		436	croak). The watcher will be triggered only when the child process has
		437	finished and an exit status is available, not on any trace events
		438	(stopped/continued).
		439
		440	The callback will be called with the pid and exit status (as returned by
		441	waitpid), so unlike other watcher types, you I<can> rely on child watcher
		442	callback arguments.
		443
		444	This watcher type works by installing a signal handler for C<SIGCHLD>,
		445	and since it cannot be shared, nothing else should use SIGCHLD or reap
		446	random child processes (waiting for specific child processes, e.g. inside
		447	C<system>, is just fine).
		448
		449	There is a slight catch to child watchers, however: you usually start them
		450	I<after> the child process was created, and this means the process could
		451	have exited already (and no SIGCHLD will be sent anymore).
		452
		453	Not all event models handle this correctly (neither POE nor IO::Async do,
		454	see their AnyEvent::Impl manpages for details), but even for event models
		455	that I<do> handle this correctly, they usually need to be loaded before
		456	the process exits (i.e. before you fork in the first place). AnyEvent's
		457	pure perl event loop handles all cases correctly regardless of when you
		458	start the watcher.
		459
		460	This means you cannot create a child watcher as the very first
		461	thing in an AnyEvent program, you I<have> to create at least one
		462	watcher before you C<fork> the child (alternatively, you can call
		463	C<AnyEvent::detect>).
		464
		465	As most event loops do not support waiting for child events, they will be
		466	emulated by AnyEvent in most cases, in which the latency and race problems
		467	mentioned in the description of signal watchers apply.
		468
		469	Example: fork a process and wait for it
		470
		471	my $done = AnyEvent->condvar;
		472
		473	my $pid = fork or exit 5;
		474
		475	my $w = AnyEvent->child (
		476	pid => $pid,
		477	cb => sub {
		478	my ($pid, $status) = @_;
		479	warn "pid $pid exited with status $status";
		480	$done->send;
		481	},
		482	);
		483
		484	# do something else, then wait for process exit
		485	$done->recv;
		486
		487	=head2 IDLE WATCHERS
		488
		489	$w = AnyEvent->idle (cb => <callback>);
		490
		491	Sometimes there is a need to do something, but it is not so important
		492	to do it instantly, but only when there is nothing better to do. This
		493	"nothing better to do" is usually defined to be "no other events need
		494	attention by the event loop".
		495
		496	Idle watchers ideally get invoked when the event loop has nothing
		497	better to do, just before it would block the process to wait for new
		498	events. Instead of blocking, the idle watcher is invoked.
		499
		500	Most event loops unfortunately do not really support idle watchers (only
		501	EV, Event and Glib do it in a usable fashion) - for the rest, AnyEvent
		502	will simply call the callback "from time to time".
		503
		504	Example: read lines from STDIN, but only process them when the
		505	program is otherwise idle:
		506
		507	my @lines; # read data
		508	my $idle_w;
		509	my $io_w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {
		510	push @lines, scalar <STDIN>;
		511
		512	# start an idle watcher, if not already done
		513	$idle_w ||= AnyEvent->idle (cb => sub {
		514	# handle only one line, when there are lines left
		515	if (my $line = shift @lines) {
		516	print "handled when idle: $line";
		517	} else {
		518	# otherwise disable the idle watcher again
		519	undef $idle_w;
		520	}
		521	});
		522	});
		523
		524	=head2 CONDITION VARIABLES
		525
		526	$cv = AnyEvent->condvar;
		527
		528	$cv->send (<list>);
		529	my @res = $cv->recv;
		530
		531	If you are familiar with some event loops you will know that all of them
		532	require you to run some blocking "loop", "run" or similar function that
		533	will actively watch for new events and call your callbacks.
		534
		535	AnyEvent is slightly different: it expects somebody else to run the event
		536	loop and will only block when necessary (usually when told by the user).
		537
		538	The instrument to do that is called a "condition variable", so called
		539	because they represent a condition that must become true.
		540
		541	Now is probably a good time to look at the examples further below.
		542
		543	Condition variables can be created by calling the C<< AnyEvent->condvar
		544	>> method, usually without arguments. The only argument pair allowed is
		545	C<cb>, which specifies a callback to be called when the condition variable
		546	becomes true, with the condition variable as the first argument (but not
		547	the results).
		548
		549	After creation, the condition variable is "false" until it becomes "true"
		550	by calling the C<send> method (or calling the condition variable as if it
		551	were a callback, read about the caveats in the description for the C<<
		552	->send >> method).
		553
		554	Condition variables are similar to callbacks, except that you can
		555	optionally wait for them. They can also be called merge points - points
		556	in time where multiple outstanding events have been processed. And yet
		557	another way to call them is transactions - each condition variable can be
		558	used to represent a transaction, which finishes at some point and delivers
		559	a result. And yet some people know them as "futures" - a promise to
		560	compute/deliver something that you can wait for.
		561
		562	Condition variables are very useful to signal that something has finished,
		563	for example, if you write a module that does asynchronous http requests,
		564	then a condition variable would be the ideal candidate to signal the
		565	availability of results. The user can either act when the callback is
		566	called or can synchronously C<< ->recv >> for the results.
		567
		568	You can also use them to simulate traditional event loops - for example,
		569	you can block your main program until an event occurs - for example, you
		570	could C<< ->recv >> in your main program until the user clicks the Quit
		571	button of your app, which would C<< ->send >> the "quit" event.
		572
		573	Note that condition variables recurse into the event loop - if you have
		574	two pieces of code that call C<< ->recv >> in a round-robin fashion, you
		575	lose. Therefore, condition variables are good to export to your caller, but
		576	you should avoid making a blocking wait yourself, at least in callbacks,
		577	as this asks for trouble.
		578
		579	Condition variables are represented by hash refs in perl, and the keys
		580	used by AnyEvent itself are all named C<_ae_XXX> to make subclassing
		581	easy (it is often useful to build your own transaction class on top of
		582	AnyEvent). To subclass, use C<AnyEvent::CondVar> as base class and call
		583	it's C<new> method in your own C<new> method.
		584
		585	There are two "sides" to a condition variable - the "producer side" which
		586	eventually calls C<< -> send >>, and the "consumer side", which waits
		587	for the send to occur.
		588
		589	Example: wait for a timer.
		590
		591	# wait till the result is ready
		592	my $result_ready = AnyEvent->condvar;
		593
		594	# do something such as adding a timer
		595	# or socket watcher the calls $result_ready->send
		596	# when the "result" is ready.
		597	# in this case, we simply use a timer:
		598	my $w = AnyEvent->timer (
		599	after => 1,
		600	cb => sub { $result_ready->send },
		601	);
		602
		603	# this "blocks" (while handling events) till the callback
		604	# calls ->send
		605	$result_ready->recv;
		606
		607	Example: wait for a timer, but take advantage of the fact that condition
		608	variables are also callable directly.
		609
		610	my $done = AnyEvent->condvar;
		611	my $delay = AnyEvent->timer (after => 5, cb => $done);
		612	$done->recv;
		613
		614	Example: Imagine an API that returns a condvar and doesn't support
		615	callbacks. This is how you make a synchronous call, for example from
		616	the main program:
		617
		618	use AnyEvent::CouchDB;
		619
		620	...
		621
		622	my @info = $couchdb->info->recv;
		623
		624	And this is how you would just set a callback to be called whenever the
		625	results are available:
		626
		627	$couchdb->info->cb (sub {
		628	my @info = $_[0]->recv;
		629	});
		630
		631	=head3 METHODS FOR PRODUCERS
		632
		633	These methods should only be used by the producing side, i.e. the
		634	code/module that eventually sends the signal. Note that it is also
		635	the producer side which creates the condvar in most cases, but it isn't
		636	uncommon for the consumer to create it as well.
		637
		638	=over 4
		639
		640	=item $cv->send (...)
		641
		642	Flag the condition as ready - a running C<< ->recv >> and all further
		643	calls to C<recv> will (eventually) return after this method has been
		644	called. If nobody is waiting the send will be remembered.
		645
		646	If a callback has been set on the condition variable, it is called
		647	immediately from within send.
		648
		649	Any arguments passed to the C<send> call will be returned by all
		650	future C<< ->recv >> calls.
		651
		652	Condition variables are overloaded so one can call them directly (as if
		653	they were a code reference). Calling them directly is the same as calling
		654	C<send>.
		655
		656	=item $cv->croak ($error)
		657
		658	Similar to send, but causes all call's to C<< ->recv >> to invoke
		659	C<Carp::croak> with the given error message/object/scalar.
		660
		661	This can be used to signal any errors to the condition variable
		662	user/consumer. Doing it this way instead of calling C<croak> directly
		663	delays the error detetcion, but has the overwhelmign advantage that it
		664	diagnoses the error at the place where the result is expected, and not
		665	deep in some event clalback without connection to the actual code causing
		666	the problem.
		667
		668	=item $cv->begin ([group callback])
		669
		670	=item $cv->end
		671
		672	These two methods can be used to combine many transactions/events into
		673	one. For example, a function that pings many hosts in parallel might want
		674	to use a condition variable for the whole process.
		675
		676	Every call to C<< ->begin >> will increment a counter, and every call to
		677	C<< ->end >> will decrement it. If the counter reaches C<0> in C<< ->end
		678	>>, the (last) callback passed to C<begin> will be executed, passing the
		679	condvar as first argument. That callback is I<supposed> to call C<< ->send
		680	>>, but that is not required. If no group callback was set, C<send> will
		681	be called without any arguments.
		682
		683	You can think of C<< $cv->send >> giving you an OR condition (one call
		684	sends), while C<< $cv->begin >> and C<< $cv->end >> giving you an AND
		685	condition (all C<begin> calls must be C<end>'ed before the condvar sends).
		686
		687	Let's start with a simple example: you have two I/O watchers (for example,
		688	STDOUT and STDERR for a program), and you want to wait for both streams to
		689	close before activating a condvar:
		690
		691	my $cv = AnyEvent->condvar;
		692
		693	$cv->begin; # first watcher
		694	my $w1 = AnyEvent->io (fh => $fh1, cb => sub {
		695	defined sysread $fh1, my $buf, 4096
		696	or $cv->end;
		697	});
		698
		699	$cv->begin; # second watcher
		700	my $w2 = AnyEvent->io (fh => $fh2, cb => sub {
		701	defined sysread $fh2, my $buf, 4096
		702	or $cv->end;
		703	});
		704
		705	$cv->recv;
		706
		707	This works because for every event source (EOF on file handle), there is
		708	one call to C<begin>, so the condvar waits for all calls to C<end> before
		709	sending.
		710
		711	The ping example mentioned above is slightly more complicated, as the
		712	there are results to be passwd back, and the number of tasks that are
		713	begung can potentially be zero:
		714
		715	my $cv = AnyEvent->condvar;
		716
		717	my %result;
		718	$cv->begin (sub { shift->send (\%result) });
		719
		720	for my $host (@list_of_hosts) {
		721	$cv->begin;
		722	ping_host_then_call_callback $host, sub {
		723	$result{$host} = ...;
		724	$cv->end;
		725	};
		726	}
		727
		728	$cv->end;
		729
		730	This code fragment supposedly pings a number of hosts and calls
		731	C<send> after results for all then have have been gathered - in any
		732	order. To achieve this, the code issues a call to C<begin> when it starts
		733	each ping request and calls C<end> when it has received some result for
		734	it. Since C<begin> and C<end> only maintain a counter, the order in which
		735	results arrive is not relevant.
		736
		737	There is an additional bracketing call to C<begin> and C<end> outside the
		738	loop, which serves two important purposes: first, it sets the callback
		739	to be called once the counter reaches C<0>, and second, it ensures that
		740	C<send> is called even when C<no> hosts are being pinged (the loop
		741	doesn't execute once).
		742
		743	This is the general pattern when you "fan out" into multiple (but
		744	potentially none) subrequests: use an outer C<begin>/C<end> pair to set
		745	the callback and ensure C<end> is called at least once, and then, for each
		746	subrequest you start, call C<begin> and for each subrequest you finish,
		747	call C<end>.
		748
		749	=back
		750
		751	=head3 METHODS FOR CONSUMERS
		752
		753	These methods should only be used by the consuming side, i.e. the
		754	code awaits the condition.
		755
		756	=over 4
		757
		758	=item $cv->recv
		759
		760	Wait (blocking if necessary) until the C<< ->send >> or C<< ->croak
		761	>> methods have been called on c<$cv>, while servicing other watchers
		762	normally.
		763
		764	You can only wait once on a condition - additional calls are valid but
		765	will return immediately.
		766
		767	If an error condition has been set by calling C<< ->croak >>, then this
		768	function will call C<croak>.
		769
		770	In list context, all parameters passed to C<send> will be returned,
		771	in scalar context only the first one will be returned.
		772
		773	Note that doing a blocking wait in a callback is not supported by any
		774	event loop, that is, recursive invocation of a blocking C<< ->recv
		775	>> is not allowed, and the C<recv> call will C<croak> if such a
		776	condition is detected. This condition can be slightly loosened by using
		777	L<Coro::AnyEvent>, which allows you to do a blocking C<< ->recv >> from
		778	any thread that doesn't run the event loop itself.
		779
		780	Not all event models support a blocking wait - some die in that case
		781	(programs might want to do that to stay interactive), so I<if you are
		782	using this from a module, never require a blocking wait>. Instead, let the
		783	caller decide whether the call will block or not (for example, by coupling
		784	condition variables with some kind of request results and supporting
		785	callbacks so the caller knows that getting the result will not block,
		786	while still supporting blocking waits if the caller so desires).
		787
		788	You can ensure that C<< -recv >> never blocks by setting a callback and
		789	only calling C<< ->recv >> from within that callback (or at a later
		790	time). This will work even when the event loop does not support blocking
		791	waits otherwise.
		792
		793	=item $bool = $cv->ready
		794
		795	Returns true when the condition is "true", i.e. whether C<send> or
		796	C<croak> have been called.
		797
		798	=item $cb = $cv->cb ($cb->($cv))
		799
		800	This is a mutator function that returns the callback set and optionally
		801	replaces it before doing so.
		802
		803	The callback will be called when the condition becomes (or already was)
		804	"true", i.e. when C<send> or C<croak> are called (or were called), with
		805	the only argument being the condition variable itself. Calling C<recv>
		806	inside the callback or at any later time is guaranteed not to block.
		807
		808	=back
		809
		810	=head1 SUPPORTED EVENT LOOPS/BACKENDS
		811
		812	The available backend classes are (every class has its own manpage):
		813
		814	=over 4
		815
		816	=item Backends that are autoprobed when no other event loop can be found.
		817
		818	EV is the preferred backend when no other event loop seems to be in
		819	use. If EV is not installed, then AnyEvent will fall back to its own
		820	pure-perl implementation, which is available everywhere as it comes with
		821	AnyEvent itself.
		822
		823	AnyEvent::Impl::EV based on EV (interface to libev, best choice).
		824	AnyEvent::Impl::Perl pure-perl implementation, fast and portable.
		825
		826	=item Backends that are transparently being picked up when they are used.
		827
		828	These will be used when they are currently loaded when the first watcher
		829	is created, in which case it is assumed that the application is using
		830	them. This means that AnyEvent will automatically pick the right backend
		831	when the main program loads an event module before anything starts to
		832	create watchers. Nothing special needs to be done by the main program.
		833
		834	AnyEvent::Impl::Event based on Event, very stable, few glitches.
		835	AnyEvent::Impl::Glib based on Glib, slow but very stable.
		836	AnyEvent::Impl::Tk based on Tk, very broken.
		837	AnyEvent::Impl::EventLib based on Event::Lib, leaks memory and worse.
		838	AnyEvent::Impl::POE based on POE, very slow, some limitations.
		839	AnyEvent::Impl::Irssi used when running within irssi.
		840
		841	=item Backends with special needs.
		842
		843	Qt requires the Qt::Application to be instantiated first, but will
		844	otherwise be picked up automatically. As long as the main program
		845	instantiates the application before any AnyEvent watchers are created,
		846	everything should just work.
		847
		848	AnyEvent::Impl::Qt based on Qt.
		849
		850	Support for IO::Async can only be partial, as it is too broken and
		851	architecturally limited to even support the AnyEvent API. It also
		852	is the only event loop that needs the loop to be set explicitly, so
		853	it can only be used by a main program knowing about AnyEvent. See
		854	L<AnyEvent::Impl::Async> for the gory details.
		855
		856	AnyEvent::Impl::IOAsync based on IO::Async, cannot be autoprobed.
		857
		858	=item Event loops that are indirectly supported via other backends.
		859
		860	Some event loops can be supported via other modules:
		861
		862	There is no direct support for WxWidgets (L<Wx>) or L<Prima>.
		863
		864	B<WxWidgets> has no support for watching file handles. However, you can
		865	use WxWidgets through the POE adaptor, as POE has a Wx backend that simply
		866	polls 20 times per second, which was considered to be too horrible to even
		867	consider for AnyEvent.
		868
		869	B<Prima> is not supported as nobody seems to be using it, but it has a POE
		870	backend, so it can be supported through POE.
		871
		872	AnyEvent knows about both L<Prima> and L<Wx>, however, and will try to
		873	load L<POE> when detecting them, in the hope that POE will pick them up,
		874	in which case everything will be automatic.
		875
		876	=back
		877
		878	=head1 GLOBAL VARIABLES AND FUNCTIONS
		879
		880	These are not normally required to use AnyEvent, but can be useful to
		881	write AnyEvent extension modules.
		882
		883	=over 4
		884
		885	=item $AnyEvent::MODEL
		886
		887	Contains C<undef> until the first watcher is being created, before the
		888	backend has been autodetected.
		889
		890	Afterwards it contains the event model that is being used, which is the
		891	name of the Perl class implementing the model. This class is usually one
		892	of the C<AnyEvent::Impl:xxx> modules, but can be any other class in the
		893	case AnyEvent has been extended at runtime (e.g. in I<rxvt-unicode> it
		894	will be C<urxvt::anyevent>).
		895
		896	=item AnyEvent::detect
		897
		898	Returns C<$AnyEvent::MODEL>, forcing autodetection of the event model
		899	if necessary. You should only call this function right before you would
		900	have created an AnyEvent watcher anyway, that is, as late as possible at
		901	runtime, and not e.g. while initialising of your module.
		902
		903	If you need to do some initialisation before AnyEvent watchers are
		904	created, use C<post_detect>.
		905
		906	=item $guard = AnyEvent::post_detect { BLOCK }
		907
		908	Arranges for the code block to be executed as soon as the event model is
		909	autodetected (or immediately if this has already happened).
		910
		911	The block will be executed I<after> the actual backend has been detected
		912	(C<$AnyEvent::MODEL> is set), but I<before> any watchers have been
		913	created, so it is possible to e.g. patch C<@AnyEvent::ISA> or do
		914	other initialisations - see the sources of L<AnyEvent::Strict> or
		915	L<AnyEvent::AIO> to see how this is used.
		916
		917	The most common usage is to create some global watchers, without forcing
		918	event module detection too early, for example, L<AnyEvent::AIO> creates
		919	and installs the global L<IO::AIO> watcher in a C<post_detect> block to
		920	avoid autodetecting the event module at load time.
		921
		922	If called in scalar or list context, then it creates and returns an object
		923	that automatically removes the callback again when it is destroyed (or
		924	C<undef> when the hook was immediately executed). See L<AnyEvent::AIO> for
		925	a case where this is useful.
		926
		927	Example: Create a watcher for the IO::AIO module and store it in
		928	C<$WATCHER>. Only do so after the event loop is initialised, though.
		929
		930	our WATCHER;
		931
		932	my $guard = AnyEvent::post_detect {
		933	$WATCHER = AnyEvent->io (fh => IO::AIO::poll_fileno, poll => 'r', cb => \&IO::AIO::poll_cb);
		934	};
		935
		936	# the ||= is important in case post_detect immediately runs the block,
		937	# as to not clobber the newly-created watcher. assigning both watcher and
		938	# post_detect guard to the same variable has the advantage of users being
		939	# able to just C<undef $WATCHER> if the watcher causes them grief.
		940
		941	$WATCHER ||= $guard;
		942
		943	=item @AnyEvent::post_detect
		944
		945	If there are any code references in this array (you can C<push> to it
		946	before or after loading AnyEvent), then they will called directly after
		947	the event loop has been chosen.
		948
		949	You should check C<$AnyEvent::MODEL> before adding to this array, though:
		950	if it is defined then the event loop has already been detected, and the
		951	array will be ignored.
		952
		953	Best use C<AnyEvent::post_detect { BLOCK }> when your application allows
		954	it,as it takes care of these details.
		955
		956	This variable is mainly useful for modules that can do something useful
		957	when AnyEvent is used and thus want to know when it is initialised, but do
		958	not need to even load it by default. This array provides the means to hook
		959	into AnyEvent passively, without loading it.
		960
		961	=back
		962
		963	=head1 WHAT TO DO IN A MODULE
		964
		965	As a module author, you should C<use AnyEvent> and call AnyEvent methods
		966	freely, but you should not load a specific event module or rely on it.
		967
		968	Be careful when you create watchers in the module body - AnyEvent will
		969	decide which event module to use as soon as the first method is called, so
		970	by calling AnyEvent in your module body you force the user of your module
		971	to load the event module first.
		972
		973	Never call C<< ->recv >> on a condition variable unless you I<know> that
		974	the C<< ->send >> method has been called on it already. This is
		975	because it will stall the whole program, and the whole point of using
		976	events is to stay interactive.
		977
		978	It is fine, however, to call C<< ->recv >> when the user of your module
		979	requests it (i.e. if you create a http request object ad have a method
		980	called C<results> that returns the results, it should call C<< ->recv >>
		981	freely, as the user of your module knows what she is doing. always).
		982
		983	=head1 WHAT TO DO IN THE MAIN PROGRAM
		984
		985	There will always be a single main program - the only place that should
		986	dictate which event model to use.
		987
		988	If it doesn't care, it can just "use AnyEvent" and use it itself, or not
		989	do anything special (it does not need to be event-based) and let AnyEvent
		990	decide which implementation to chose if some module relies on it.
		991
		992	If the main program relies on a specific event model - for example, in
		993	Gtk2 programs you have to rely on the Glib module - you should load the
		994	event module before loading AnyEvent or any module that uses it: generally
		995	speaking, you should load it as early as possible. The reason is that
		996	modules might create watchers when they are loaded, and AnyEvent will
		997	decide on the event model to use as soon as it creates watchers, and it
		998	might chose the wrong one unless you load the correct one yourself.
		999
		1000	You can chose to use a pure-perl implementation by loading the
		1001	C<AnyEvent::Impl::Perl> module, which gives you similar behaviour
		1002	everywhere, but letting AnyEvent chose the model is generally better.
		1003
		1004	=head2 MAINLOOP EMULATION
		1005
		1006	Sometimes (often for short test scripts, or even standalone programs who
		1007	only want to use AnyEvent), you do not want to run a specific event loop.
		1008
		1009	In that case, you can use a condition variable like this:
		1010
		1011	AnyEvent->condvar->recv;
		1012
		1013	This has the effect of entering the event loop and looping forever.
		1014
		1015	Note that usually your program has some exit condition, in which case
		1016	it is better to use the "traditional" approach of storing a condition
		1017	variable somewhere, waiting for it, and sending it when the program should
		1018	exit cleanly.
		1019
		1020
		1021	=head1 OTHER MODULES
		1022
		1023	The following is a non-exhaustive list of additional modules that use
		1024	AnyEvent as a client and can therefore be mixed easily with other AnyEvent
		1025	modules and other event loops in the same program. Some of the modules
		1026	come with AnyEvent, most are available via CPAN.
		1027
		1028	=over 4
		1029
		1030	=item L<AnyEvent::Util>
		1031
		1032	Contains various utility functions that replace often-used but blocking
		1033	functions such as C<inet_aton> by event-/callback-based versions.
		1034
		1035	=item L<AnyEvent::Socket>
		1036
		1037	Provides various utility functions for (internet protocol) sockets,
		1038	addresses and name resolution. Also functions to create non-blocking tcp
		1039	connections or tcp servers, with IPv6 and SRV record support and more.
		1040
		1041	=item L<AnyEvent::Handle>
		1042
		1043	Provide read and write buffers, manages watchers for reads and writes,
		1044	supports raw and formatted I/O, I/O queued and fully transparent and
		1045	non-blocking SSL/TLS (via L<AnyEvent::TLS>.
		1046
		1047	=item L<AnyEvent::DNS>
		1048
		1049	Provides rich asynchronous DNS resolver capabilities.
		1050
		1051	=item L<AnyEvent::HTTP>
		1052
		1053	A simple-to-use HTTP library that is capable of making a lot of concurrent
		1054	HTTP requests.
		1055
		1056	=item L<AnyEvent::HTTPD>
		1057
		1058	Provides a simple web application server framework.
		1059
		1060	=item L<AnyEvent::FastPing>
		1061
		1062	The fastest ping in the west.
		1063
		1064	=item L<AnyEvent::DBI>
		1065
		1066	Executes L<DBI> requests asynchronously in a proxy process.
		1067
		1068	=item L<AnyEvent::AIO>
		1069
		1070	Truly asynchronous I/O, should be in the toolbox of every event
		1071	programmer. AnyEvent::AIO transparently fuses L<IO::AIO> and AnyEvent
		1072	together.
		1073
		1074	=item L<AnyEvent::BDB>
		1075
		1076	Truly asynchronous Berkeley DB access. AnyEvent::BDB transparently fuses
		1077	L<BDB> and AnyEvent together.
		1078
		1079	=item L<AnyEvent::GPSD>
		1080
		1081	A non-blocking interface to gpsd, a daemon delivering GPS information.
		1082
		1083	=item L<AnyEvent::IRC>
		1084
		1085	AnyEvent based IRC client module family (replacing the older Net::IRC3).
		1086
		1087	=item L<AnyEvent::XMPP>
		1088
		1089	AnyEvent based XMPP (Jabber protocol) module family (replacing the older
		1090	Net::XMPP2>.
		1091
		1092	=item L<AnyEvent::IGS>
		1093
		1094	A non-blocking interface to the Internet Go Server protocol (used by
		1095	L<App::IGS>).
		1096
		1097	=item L<Net::FCP>
		1098
		1099	AnyEvent-based implementation of the Freenet Client Protocol, birthplace
		1100	of AnyEvent.
		1101
		1102	=item L<Event::ExecFlow>
		1103
		1104	High level API for event-based execution flow control.
		1105
		1106	=item L<Coro>
		1107
		1108	Has special support for AnyEvent via L<Coro::AnyEvent>.
		1109
		1110	=back
		1111
65	=cut	1112	=cut
66		1113
67	package AnyEvent;	1114	package AnyEvent;
68		1115
69	no warnings;	1116	# basically a tuned-down version of common::sense
70	use strict 'vars';	1117	sub common_sense {
		1118	# from common:.sense 1.0
		1119	${^WARNING_BITS} = "\xfc\x3f\xf3\x00\x0f\xf3\xcf\xc0\xf3\xfc\x33\x03";
		1120	# use strict vars subs
		1121	$^H |= 0x00000600;
		1122	}
		1123
		1124	BEGIN { AnyEvent::common_sense }
		1125
71	use Carp;	1126	use Carp ();
72		1127
73	our $VERSION = '1.02';	1128	our $VERSION = '5.202';
74	our $MODEL;	1129	our $MODEL;
75		1130
76	our $AUTOLOAD;	1131	our $AUTOLOAD;
77	our @ISA;	1132	our @ISA;
78		1133
79	our $verbose = $ENV{PERL_ANYEVENT_VERBOSE}*1;
80
81	our @REGISTRY;	1134	our @REGISTRY;
82		1135
		1136	our $WIN32;
		1137
		1138	our $VERBOSE;
		1139
		1140	BEGIN {
		1141	eval "sub WIN32(){ " . (($^O =~ /mswin32/i)*1) ." }";
		1142	eval "sub TAINT(){ " . (${^TAINT}*1) . " }";
		1143
		1144	delete @ENV{grep /^PERL_ANYEVENT_/, keys %ENV}
		1145	if ${^TAINT};
		1146
		1147	$VERBOSE = $ENV{PERL_ANYEVENT_VERBOSE}*1;
		1148
		1149	}
		1150
		1151	our $MAX_SIGNAL_LATENCY = 10;
		1152
		1153	our %PROTOCOL; # (ipv4|ipv6) => (1|2), higher numbers are preferred
		1154
		1155	{
		1156	my $idx;
		1157	$PROTOCOL{$_} = ++$idx
		1158	for reverse split /\s*,\s*/,
		1159	$ENV{PERL_ANYEVENT_PROTOCOLS} || "ipv4,ipv6";
		1160	}
		1161
83	my @models = (	1162	my @models = (
84	[Coro::Event:: => AnyEvent::Impl::Coro::],	1163	[EV:: => AnyEvent::Impl::EV:: , 1],
		1164	[AnyEvent::Impl::Perl:: => AnyEvent::Impl::Perl:: , 1],
		1165	# everything below here will not (normally) be autoprobed
		1166	# as the pureperl backend should work everywhere
		1167	# and is usually faster
85	[Event:: => AnyEvent::Impl::Event::],	1168	[Event:: => AnyEvent::Impl::Event::, 1],
86	[Glib:: => AnyEvent::Impl::Glib::],	1169	[Glib:: => AnyEvent::Impl::Glib:: , 1], # becomes extremely slow with many watchers
		1170	[Event::Lib:: => AnyEvent::Impl::EventLib::], # too buggy
		1171	[Irssi:: => AnyEvent::Impl::Irssi::], # Irssi has a bogus "Event" package
		1172	[Tk:: => AnyEvent::Impl::Tk::], # crashes with many handles
		1173	[Qt:: => AnyEvent::Impl::Qt::], # requires special main program
		1174	[POE::Kernel:: => AnyEvent::Impl::POE::], # lasciate ogni speranza
87	[Tk:: => AnyEvent::Impl::Tk::],	1175	[Wx:: => AnyEvent::Impl::POE::],
		1176	[Prima:: => AnyEvent::Impl::POE::],
		1177	# IO::Async is just too broken - we would need workarounds for its
		1178	# byzantine signal and broken child handling, among others.
		1179	# IO::Async is rather hard to detect, as it doesn't have any
		1180	# obvious default class.
		1181	[IO::Async:: => AnyEvent::Impl::IOAsync::], # requires special main program
		1182	[IO::Async::Loop:: => AnyEvent::Impl::IOAsync::], # requires special main program
		1183	[IO::Async::Notifier:: => AnyEvent::Impl::IOAsync::], # requires special main program
		1184	[AnyEvent::Impl::IOAsync:: => AnyEvent::Impl::IOAsync::], # requires special main program
88	);	1185	);
89		1186
90	our %method = map +($_ => 1), qw(io timer condvar broadcast wait cancel DESTROY);	1187	our %method = map +($_ => 1),
		1188	qw(io timer time now now_update signal child idle condvar one_event DESTROY);
91		1189
92	sub AUTOLOAD {	1190	our @post_detect;
93	$AUTOLOAD =~ s/.*://;
94		1191
95	$method{$AUTOLOAD}	1192	sub post_detect(&) {
96	or croak "$AUTOLOAD: not a valid method for AnyEvent objects";	1193	my ($cb) = @_;
97		1194
		1195	if ($MODEL) {
		1196	$cb->();
		1197
		1198	undef
		1199	} else {
		1200	push @post_detect, $cb;
		1201
		1202	defined wantarray
		1203	? bless \$cb, "AnyEvent::Util::postdetect"
		1204	: ()
		1205	}
		1206	}
		1207
		1208	sub AnyEvent::Util::postdetect::DESTROY {
		1209	@post_detect = grep $_ != ${$_[0]}, @post_detect;
		1210	}
		1211
		1212	sub detect() {
98	unless ($MODEL) {	1213	unless ($MODEL) {
		1214	local $SIG{__DIE__};
		1215
		1216	if ($ENV{PERL_ANYEVENT_MODEL} =~ /^([a-zA-Z]+)$/) {
		1217	my $model = "AnyEvent::Impl::$1";
		1218	if (eval "require $model") {
		1219	$MODEL = $model;
		1220	warn "AnyEvent: loaded model '$model' (forced by \$ENV{PERL_ANYEVENT_MODEL}), using it.\n" if $VERBOSE >= 2;
		1221	} else {
		1222	warn "AnyEvent: unable to load model '$model' (from \$ENV{PERL_ANYEVENT_MODEL}):\n$@" if $VERBOSE;
		1223	}
		1224	}
		1225
99	# check for already loaded models	1226	# check for already loaded models
		1227	unless ($MODEL) {
100	for (@REGISTRY, @models) {	1228	for (@REGISTRY, @models) {
101	my ($package, $model) = @$_;	1229	my ($package, $model) = @$_;
102	if (${"$package\::VERSION"} > 0) {	1230	if (${"$package\::VERSION"} > 0) {
103	if (eval "require $model") {	1231	if (eval "require $model") {
104	$MODEL = $model;	1232	$MODEL = $model;
105	warn "AnyEvent: found model '$model', using it.\n" if $verbose > 1;	1233	warn "AnyEvent: autodetected model '$model', using it.\n" if $VERBOSE >= 2;
106	last;	1234	last;
		1235	}
107	}	1236	}
108	}	1237	}
		1238
		1239	unless ($MODEL) {
		1240	# try to autoload a model
		1241	for (@REGISTRY, @models) {
		1242	my ($package, $model, $autoload) = @$_;
		1243	if (
		1244	$autoload
		1245	and eval "require $package"
		1246	and ${"$package\::VERSION"} > 0
		1247	and eval "require $model"
		1248	) {
		1249	$MODEL = $model;
		1250	warn "AnyEvent: autoloaded model '$model', using it.\n" if $VERBOSE >= 2;
		1251	last;
		1252	}
		1253	}
		1254
		1255	$MODEL
		1256	or die "No event module selected for AnyEvent and autodetect failed. Install any one of these modules: EV, Event or Glib.\n";
		1257	}
109	}	1258	}
110		1259
111	unless ($MODEL) {	1260	push @{"$MODEL\::ISA"}, "AnyEvent::Base";
112	# try to load a model
113		1261
114	for (@REGISTRY, @models) {	1262	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		1263
		1264	require AnyEvent::Strict if $ENV{PERL_ANYEVENT_STRICT};
		1265
		1266	(shift @post_detect)->() while @post_detect;
		1267	}
		1268
123	$MODEL	1269	$MODEL
124	or die "No event module selected for AnyEvent and autodetect failed. Install any one of these modules: Coro, Event, Glib or Tk.";	1270	}
		1271
		1272	sub AUTOLOAD {
		1273	(my $func = $AUTOLOAD) =~ s/.*://;
		1274
		1275	$method{$func}
		1276	or Carp::croak "$func: not a valid method for AnyEvent objects";
		1277
		1278	detect unless $MODEL;
		1279
		1280	my $class = shift;
		1281	$class->$func (@_);
		1282	}
		1283
		1284	# utility function to dup a filehandle. this is used by many backends
		1285	# to support binding more than one watcher per filehandle (they usually
		1286	# allow only one watcher per fd, so we dup it to get a different one).
		1287	sub _dupfh($$;$$) {
		1288	my ($poll, $fh, $r, $w) = @_;
		1289
		1290	# cygwin requires the fh mode to be matching, unix doesn't
		1291	my ($rw, $mode) = $poll eq "r" ? ($r, "<&") : ($w, ">&");
		1292
		1293	open my $fh2, $mode, $fh
		1294	or die "AnyEvent->io: cannot dup() filehandle in mode '$poll': $!,";
		1295
		1296	# we assume CLOEXEC is already set by perl in all important cases
		1297
		1298	($fh2, $rw)
		1299	}
		1300
		1301	=head1 SIMPLIFIED AE API
		1302
		1303	Starting with version 5.0, AnyEvent officially supports a second, much
		1304	simpler, API that is designed to reduce the calling, typing and memory
		1305	overhead.
		1306
		1307	See the L<AE> manpage for details.
		1308
		1309	=cut
		1310
		1311	package AE;
		1312
		1313	our $VERSION = $AnyEvent::VERSION;
		1314
		1315	sub io($$$) {
		1316	AnyEvent->io (fh => $_[0], poll => $_[1] ? "w" : "r", cb => $_[2])
		1317	}
		1318
		1319	sub timer($$$) {
		1320	AnyEvent->timer (after => $_[0], interval => $_[1], cb => $_[2])
		1321	}
		1322
		1323	sub signal($$) {
		1324	AnyEvent->signal (signal => $_[0], cb => $_[1])
		1325	}
		1326
		1327	sub child($$) {
		1328	AnyEvent->child (pid => $_[0], cb => $_[1])
		1329	}
		1330
		1331	sub idle($) {
		1332	AnyEvent->idle (cb => $_[0])
		1333	}
		1334
		1335	sub cv(;&) {
		1336	AnyEvent->condvar (@_ ? (cb => $_[0]) : ())
		1337	}
		1338
		1339	sub now() {
		1340	AnyEvent->now
		1341	}
		1342
		1343	sub now_update() {
		1344	AnyEvent->now_update
		1345	}
		1346
		1347	sub time() {
		1348	AnyEvent->time
		1349	}
		1350
		1351	package AnyEvent::Base;
		1352
		1353	# default implementations for many methods
		1354
		1355	sub _time() {
		1356	# probe for availability of Time::HiRes
		1357	if (eval "use Time::HiRes (); Time::HiRes::time (); 1") {
		1358	warn "AnyEvent: using Time::HiRes for sub-second timing accuracy.\n" if $VERBOSE >= 8;
		1359	*_time = \&Time::HiRes::time;
		1360	# if (eval "use POSIX (); (POSIX::times())...
		1361	} else {
		1362	warn "AnyEvent: using built-in time(), WARNING, no sub-second resolution!\n" if $VERBOSE;
		1363	*_time = sub { time }; # epic fail
		1364	}
		1365
		1366	&_time
		1367	}
		1368
		1369	sub time { _time }
		1370	sub now { _time }
		1371	sub now_update { }
		1372
		1373	# default implementation for ->condvar
		1374
		1375	sub condvar {
		1376	bless { @_ == 3 ? (_ae_cb => $_[2]) : () }, "AnyEvent::CondVar"
		1377	}
		1378
		1379	# default implementation for ->signal
		1380
		1381	our $HAVE_ASYNC_INTERRUPT;
		1382
		1383	sub _have_async_interrupt() {
		1384	$HAVE_ASYNC_INTERRUPT = 1*(!$ENV{PERL_ANYEVENT_AVOID_ASYNC_INTERRUPT}
		1385	&& eval "use Async::Interrupt 1.02 (); 1")
		1386	unless defined $HAVE_ASYNC_INTERRUPT;
		1387
		1388	$HAVE_ASYNC_INTERRUPT
		1389	}
		1390
		1391	our ($SIGPIPE_R, $SIGPIPE_W, %SIG_CB, %SIG_EV, $SIG_IO);
		1392	our (%SIG_ASY, %SIG_ASY_W);
		1393	our ($SIG_COUNT, $SIG_TW);
		1394
		1395	sub _signal_exec {
		1396	$HAVE_ASYNC_INTERRUPT
		1397	? $SIGPIPE_R->drain
		1398	: sysread $SIGPIPE_R, (my $dummy), 9;
		1399
		1400	while (%SIG_EV) {
		1401	for (keys %SIG_EV) {
		1402	delete $SIG_EV{$_};
		1403	$_->() for values %{ $SIG_CB{$_} || {} };
125	}	1404	}
126	}	1405	}
		1406	}
127		1407
128	@ISA = $MODEL;	1408	# install a dummy wakeup watcher to reduce signal catching latency
		1409	sub _sig_add() {
		1410	unless ($SIG_COUNT++) {
		1411	# try to align timer on a full-second boundary, if possible
		1412	my $NOW = AE::now;
129		1413
		1414	$SIG_TW = AE::timer
		1415	$MAX_SIGNAL_LATENCY - ($NOW - int $NOW),
		1416	$MAX_SIGNAL_LATENCY,
		1417	sub { } # just for the PERL_ASYNC_CHECK
		1418	;
		1419	}
		1420	}
		1421
		1422	sub _sig_del {
		1423	undef $SIG_TW
		1424	unless --$SIG_COUNT;
		1425	}
		1426
		1427	our $_sig_name_init; $_sig_name_init = sub {
		1428	eval q{ # poor man's autoloading
		1429	undef $_sig_name_init;
		1430
		1431	if (_have_async_interrupt) {
		1432	*sig2num = \&Async::Interrupt::sig2num;
		1433	*sig2name = \&Async::Interrupt::sig2name;
		1434	} else {
		1435	require Config;
		1436
		1437	my %signame2num;
		1438	@signame2num{ split ' ', $Config::Config{sig_name} }
		1439	= split ' ', $Config::Config{sig_num};
		1440
		1441	my @signum2name;
		1442	@signum2name[values %signame2num] = keys %signame2num;
		1443
		1444	*sig2num = sub($) {
		1445	$_[0] > 0 ? shift : $signame2num{+shift}
		1446	};
		1447	*sig2name = sub ($) {
		1448	$_[0] > 0 ? $signum2name[+shift] : shift
		1449	};
		1450	}
		1451	};
		1452	die if $@;
		1453	};
		1454
		1455	sub sig2num ($) { &$_sig_name_init; &sig2num }
		1456	sub sig2name($) { &$_sig_name_init; &sig2name }
		1457
		1458	sub signal {
		1459	eval q{ # poor man's autoloading {}
		1460	# probe for availability of Async::Interrupt
		1461	if (_have_async_interrupt) {
		1462	warn "AnyEvent: using Async::Interrupt for race-free signal handling.\n" if $VERBOSE >= 8;
		1463
		1464	$SIGPIPE_R = new Async::Interrupt::EventPipe;
		1465	$SIG_IO = AE::io $SIGPIPE_R->fileno, 0, \&_signal_exec;
		1466
		1467	} else {
		1468	warn "AnyEvent: using emulated perl signal handling with latency timer.\n" if $VERBOSE >= 8;
		1469
		1470	require Fcntl;
		1471
		1472	if (AnyEvent::WIN32) {
		1473	require AnyEvent::Util;
		1474
		1475	($SIGPIPE_R, $SIGPIPE_W) = AnyEvent::Util::portable_pipe ();
		1476	AnyEvent::Util::fh_nonblocking ($SIGPIPE_R, 1) if $SIGPIPE_R;
		1477	AnyEvent::Util::fh_nonblocking ($SIGPIPE_W, 1) if $SIGPIPE_W; # just in case
		1478	} else {
		1479	pipe $SIGPIPE_R, $SIGPIPE_W;
		1480	fcntl $SIGPIPE_R, &Fcntl::F_SETFL, &Fcntl::O_NONBLOCK if $SIGPIPE_R;
		1481	fcntl $SIGPIPE_W, &Fcntl::F_SETFL, &Fcntl::O_NONBLOCK if $SIGPIPE_W; # just in case
		1482
		1483	# not strictly required, as $^F is normally 2, but let's make sure...
		1484	fcntl $SIGPIPE_R, &Fcntl::F_SETFD, &Fcntl::FD_CLOEXEC;
		1485	fcntl $SIGPIPE_W, &Fcntl::F_SETFD, &Fcntl::FD_CLOEXEC;
		1486	}
		1487
		1488	$SIGPIPE_R
		1489	or Carp::croak "AnyEvent: unable to create a signal reporting pipe: $!\n";
		1490
		1491	$SIG_IO = AE::io $SIGPIPE_R, 0, \&_signal_exec;
		1492	}
		1493
		1494	*signal = sub {
		1495	my (undef, %arg) = @_;
		1496
		1497	my $signal = uc $arg{signal}
		1498	or Carp::croak "required option 'signal' is missing";
		1499
		1500	if ($HAVE_ASYNC_INTERRUPT) {
		1501	# async::interrupt
		1502
		1503	$signal = sig2num $signal;
		1504	$SIG_CB{$signal}{$arg{cb}} = $arg{cb};
		1505
		1506	$SIG_ASY{$signal} ||= new Async::Interrupt
		1507	cb => sub { undef $SIG_EV{$signal} },
		1508	signal => $signal,
		1509	pipe => [$SIGPIPE_R->filenos],
		1510	pipe_autodrain => 0,
		1511	;
		1512
		1513	} else {
		1514	# pure perl
		1515
		1516	# AE::Util has been loaded in signal
		1517	$signal = sig2name $signal;
		1518	$SIG_CB{$signal}{$arg{cb}} = $arg{cb};
		1519
		1520	$SIG{$signal} ||= sub {
		1521	local $!;
		1522	syswrite $SIGPIPE_W, "\x00", 1 unless %SIG_EV;
		1523	undef $SIG_EV{$signal};
		1524	};
		1525
		1526	# can't do signal processing without introducing races in pure perl,
		1527	# so limit the signal latency.
		1528	_sig_add;
		1529	}
		1530
		1531	bless [$signal, $arg{cb}], "AnyEvent::Base::signal"
		1532	};
		1533
		1534	*AnyEvent::Base::signal::DESTROY = sub {
		1535	my ($signal, $cb) = @{$_[0]};
		1536
		1537	_sig_del;
		1538
		1539	delete $SIG_CB{$signal}{$cb};
		1540
		1541	$HAVE_ASYNC_INTERRUPT
		1542	? delete $SIG_ASY{$signal}
		1543	: # delete doesn't work with older perls - they then
		1544	# print weird messages, or just unconditionally exit
		1545	# instead of getting the default action.
		1546	undef $SIG{$signal}
		1547	unless keys %{ $SIG_CB{$signal} };
		1548	};
		1549	};
		1550	die if $@;
		1551	&signal
		1552	}
		1553
		1554	# default implementation for ->child
		1555
		1556	our %PID_CB;
		1557	our $CHLD_W;
		1558	our $CHLD_DELAY_W;
		1559	our $WNOHANG;
		1560
		1561	sub _emit_childstatus($$) {
		1562	my (undef, $rpid, $rstatus) = @_;
		1563
		1564	$_->($rpid, $rstatus)
		1565	for values %{ $PID_CB{$rpid} || {} },
		1566	values %{ $PID_CB{0} || {} };
		1567	}
		1568
		1569	sub _sigchld {
		1570	my $pid;
		1571
		1572	AnyEvent->_emit_childstatus ($pid, $?)
		1573	while ($pid = waitpid -1, $WNOHANG) > 0;
		1574	}
		1575
		1576	sub child {
		1577	my (undef, %arg) = @_;
		1578
		1579	defined (my $pid = $arg{pid} + 0)
		1580	or Carp::croak "required option 'pid' is missing";
		1581
		1582	$PID_CB{$pid}{$arg{cb}} = $arg{cb};
		1583
		1584	# WNOHANG is almost cetrainly 1 everywhere
		1585	$WNOHANG ||= $^O =~ /^(?:openbsd|netbsd|linux|freebsd|cygwin|MSWin32)$/
		1586	? 1
		1587	: eval { local $SIG{__DIE__}; require POSIX; &POSIX::WNOHANG } || 1;
		1588
		1589	unless ($CHLD_W) {
		1590	$CHLD_W = AE::signal CHLD => \&_sigchld;
		1591	# child could be a zombie already, so make at least one round
		1592	&_sigchld;
		1593	}
		1594
		1595	bless [$pid, $arg{cb}], "AnyEvent::Base::child"
		1596	}
		1597
		1598	sub AnyEvent::Base::child::DESTROY {
		1599	my ($pid, $cb) = @{$_[0]};
		1600
		1601	delete $PID_CB{$pid}{$cb};
		1602	delete $PID_CB{$pid} unless keys %{ $PID_CB{$pid} };
		1603
		1604	undef $CHLD_W unless keys %PID_CB;
		1605	}
		1606
		1607	# idle emulation is done by simply using a timer, regardless
		1608	# of whether the process is idle or not, and not letting
		1609	# the callback use more than 50% of the time.
		1610	sub idle {
		1611	my (undef, %arg) = @_;
		1612
		1613	my ($cb, $w, $rcb) = $arg{cb};
		1614
		1615	$rcb = sub {
		1616	if ($cb) {
		1617	$w = _time;
		1618	&$cb;
		1619	$w = _time - $w;
		1620
		1621	# never use more then 50% of the time for the idle watcher,
		1622	# within some limits
		1623	$w = 0.0001 if $w < 0.0001;
		1624	$w = 5 if $w > 5;
		1625
		1626	$w = AE::timer $w, 0, $rcb;
		1627	} else {
		1628	# clean up...
		1629	undef $w;
		1630	undef $rcb;
		1631	}
		1632	};
		1633
		1634	$w = AE::timer 0.05, 0, $rcb;
		1635
		1636	bless \\$cb, "AnyEvent::Base::idle"
		1637	}
		1638
		1639	sub AnyEvent::Base::idle::DESTROY {
		1640	undef $${$_[0]};
		1641	}
		1642
		1643	package AnyEvent::CondVar;
		1644
		1645	our @ISA = AnyEvent::CondVar::Base::;
		1646
		1647	package AnyEvent::CondVar::Base;
		1648
		1649	#use overload
		1650	# '&{}' => sub { my $self = shift; sub { $self->send (@_) } },
		1651	# fallback => 1;
		1652
		1653	# save 300+ kilobytes by dirtily hardcoding overloading
		1654	${"AnyEvent::CondVar::Base::OVERLOAD"}{dummy}++; # Register with magic by touching.
		1655	*{'AnyEvent::CondVar::Base::()'} = sub { }; # "Make it findable via fetchmethod."
		1656	*{'AnyEvent::CondVar::Base::(&{}'} = sub { my $self = shift; sub { $self->send (@_) } }; # &{}
		1657	${'AnyEvent::CondVar::Base::()'} = 1; # fallback
		1658
		1659	our $WAITING;
		1660
		1661	sub _send {
		1662	# nop
		1663	}
		1664
		1665	sub send {
130	my $class = shift;	1666	my $cv = shift;
131	$class->$AUTOLOAD (@_);	1667	$cv->{_ae_sent} = [@_];
		1668	(delete $cv->{_ae_cb})->($cv) if $cv->{_ae_cb};
		1669	$cv->_send;
132	}	1670	}
		1671
		1672	sub croak {
		1673	$_[0]{_ae_croak} = $_[1];
		1674	$_[0]->send;
		1675	}
		1676
		1677	sub ready {
		1678	$_[0]{_ae_sent}
		1679	}
		1680
		1681	sub _wait {
		1682	$WAITING
		1683	and !$_[0]{_ae_sent}
		1684	and Carp::croak "AnyEvent::CondVar: recursive blocking wait detected";
		1685
		1686	local $WAITING = 1;
		1687	AnyEvent->one_event while !$_[0]{_ae_sent};
		1688	}
		1689
		1690	sub recv {
		1691	$_[0]->_wait;
		1692
		1693	Carp::croak $_[0]{_ae_croak} if $_[0]{_ae_croak};
		1694	wantarray ? @{ $_[0]{_ae_sent} } : $_[0]{_ae_sent}[0]
		1695	}
		1696
		1697	sub cb {
		1698	my $cv = shift;
		1699
		1700	@_
		1701	and $cv->{_ae_cb} = shift
		1702	and $cv->{_ae_sent}
		1703	and (delete $cv->{_ae_cb})->($cv);
		1704
		1705	$cv->{_ae_cb}
		1706	}
		1707
		1708	sub begin {
		1709	++$_[0]{_ae_counter};
		1710	$_[0]{_ae_end_cb} = $_[1] if @_ > 1;
		1711	}
		1712
		1713	sub end {
		1714	return if --$_[0]{_ae_counter};
		1715	&{ $_[0]{_ae_end_cb} || sub { $_[0]->send } };
		1716	}
		1717
		1718	# undocumented/compatibility with pre-3.4
		1719	*broadcast = \&send;
		1720	*wait = \&_wait;
		1721
		1722	=head1 ERROR AND EXCEPTION HANDLING
		1723
		1724	In general, AnyEvent does not do any error handling - it relies on the
		1725	caller to do that if required. The L<AnyEvent::Strict> module (see also
		1726	the C<PERL_ANYEVENT_STRICT> environment variable, below) provides strict
		1727	checking of all AnyEvent methods, however, which is highly useful during
		1728	development.
		1729
		1730	As for exception handling (i.e. runtime errors and exceptions thrown while
		1731	executing a callback), this is not only highly event-loop specific, but
		1732	also not in any way wrapped by this module, as this is the job of the main
		1733	program.
		1734
		1735	The pure perl event loop simply re-throws the exception (usually
		1736	within C<< condvar->recv >>), the L<Event> and L<EV> modules call C<<
		1737	$Event/EV::DIED->() >>, L<Glib> uses C<< install_exception_handler >> and
		1738	so on.
		1739
		1740	=head1 ENVIRONMENT VARIABLES
		1741
		1742	The following environment variables are used by this module or its
		1743	submodules.
		1744
		1745	Note that AnyEvent will remove I<all> environment variables starting with
		1746	C<PERL_ANYEVENT_> from C<%ENV> when it is loaded while taint mode is
		1747	enabled.
		1748
		1749	=over 4
		1750
		1751	=item C<PERL_ANYEVENT_VERBOSE>
		1752
		1753	By default, AnyEvent will be completely silent except in fatal
		1754	conditions. You can set this environment variable to make AnyEvent more
		1755	talkative.
		1756
		1757	When set to C<1> or higher, causes AnyEvent to warn about unexpected
		1758	conditions, such as not being able to load the event model specified by
		1759	C<PERL_ANYEVENT_MODEL>.
		1760
		1761	When set to C<2> or higher, cause AnyEvent to report to STDERR which event
		1762	model it chooses.
		1763
		1764	When set to C<8> or higher, then AnyEvent will report extra information on
		1765	which optional modules it loads and how it implements certain features.
		1766
		1767	=item C<PERL_ANYEVENT_STRICT>
		1768
		1769	AnyEvent does not do much argument checking by default, as thorough
		1770	argument checking is very costly. Setting this variable to a true value
		1771	will cause AnyEvent to load C<AnyEvent::Strict> and then to thoroughly
		1772	check the arguments passed to most method calls. If it finds any problems,
		1773	it will croak.
		1774
		1775	In other words, enables "strict" mode.
		1776
		1777	Unlike C<use strict> (or it's modern cousin, C<< use L<common::sense>
		1778	>>, it is definitely recommended to keep it off in production. Keeping
		1779	C<PERL_ANYEVENT_STRICT=1> in your environment while developing programs
		1780	can be very useful, however.
		1781
		1782	=item C<PERL_ANYEVENT_MODEL>
		1783
		1784	This can be used to specify the event model to be used by AnyEvent, before
		1785	auto detection and -probing kicks in. It must be a string consisting
		1786	entirely of ASCII letters. The string C<AnyEvent::Impl::> gets prepended
		1787	and the resulting module name is loaded and if the load was successful,
		1788	used as event model. If it fails to load AnyEvent will proceed with
		1789	auto detection and -probing.
		1790
		1791	This functionality might change in future versions.
		1792
		1793	For example, to force the pure perl model (L<AnyEvent::Impl::Perl>) you
		1794	could start your program like this:
		1795
		1796	PERL_ANYEVENT_MODEL=Perl perl ...
		1797
		1798	=item C<PERL_ANYEVENT_PROTOCOLS>
		1799
		1800	Used by both L<AnyEvent::DNS> and L<AnyEvent::Socket> to determine preferences
		1801	for IPv4 or IPv6. The default is unspecified (and might change, or be the result
		1802	of auto probing).
		1803
		1804	Must be set to a comma-separated list of protocols or address families,
		1805	current supported: C<ipv4> and C<ipv6>. Only protocols mentioned will be
		1806	used, and preference will be given to protocols mentioned earlier in the
		1807	list.
		1808
		1809	This variable can effectively be used for denial-of-service attacks
		1810	against local programs (e.g. when setuid), although the impact is likely
		1811	small, as the program has to handle conenction and other failures anyways.
		1812
		1813	Examples: C<PERL_ANYEVENT_PROTOCOLS=ipv4,ipv6> - prefer IPv4 over IPv6,
		1814	but support both and try to use both. C<PERL_ANYEVENT_PROTOCOLS=ipv4>
		1815	- only support IPv4, never try to resolve or contact IPv6
		1816	addresses. C<PERL_ANYEVENT_PROTOCOLS=ipv6,ipv4> support either IPv4 or
		1817	IPv6, but prefer IPv6 over IPv4.
		1818
		1819	=item C<PERL_ANYEVENT_EDNS0>
		1820
		1821	Used by L<AnyEvent::DNS> to decide whether to use the EDNS0 extension
		1822	for DNS. This extension is generally useful to reduce DNS traffic, but
		1823	some (broken) firewalls drop such DNS packets, which is why it is off by
		1824	default.
		1825
		1826	Setting this variable to C<1> will cause L<AnyEvent::DNS> to announce
		1827	EDNS0 in its DNS requests.
		1828
		1829	=item C<PERL_ANYEVENT_MAX_FORKS>
		1830
		1831	The maximum number of child processes that C<AnyEvent::Util::fork_call>
		1832	will create in parallel.
		1833
		1834	=item C<PERL_ANYEVENT_MAX_OUTSTANDING_DNS>
		1835
		1836	The default value for the C<max_outstanding> parameter for the default DNS
		1837	resolver - this is the maximum number of parallel DNS requests that are
		1838	sent to the DNS server.
		1839
		1840	=item C<PERL_ANYEVENT_RESOLV_CONF>
		1841
		1842	The file to use instead of F</etc/resolv.conf> (or OS-specific
		1843	configuration) in the default resolver. When set to the empty string, no
		1844	default config will be used.
		1845
		1846	=item C<PERL_ANYEVENT_CA_FILE>, C<PERL_ANYEVENT_CA_PATH>.
		1847
		1848	When neither C<ca_file> nor C<ca_path> was specified during
		1849	L<AnyEvent::TLS> context creation, and either of these environment
		1850	variables exist, they will be used to specify CA certificate locations
		1851	instead of a system-dependent default.
		1852
		1853	=item C<PERL_ANYEVENT_AVOID_GUARD> and C<PERL_ANYEVENT_AVOID_ASYNC_INTERRUPT>
		1854
		1855	When these are set to C<1>, then the respective modules are not
		1856	loaded. Mostly good for testing AnyEvent itself.
133		1857
134	=back	1858	=back
135		1859
136	=head1 SUPPLYING YOUR OWN EVENT MODEL INTERFACE	1860	=head1 SUPPLYING YOUR OWN EVENT MODEL INTERFACE
		1861
		1862	This is an advanced topic that you do not normally need to use AnyEvent in
		1863	a module. This section is only of use to event loop authors who want to
		1864	provide AnyEvent compatibility.
137		1865
138	If you need to support another event library which isn't directly	1866	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	1867	supported by AnyEvent, you can supply your own interface to it by
140	pushing, before the first watch gets created, the package name of	1868	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	1869	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	1870	C<@AnyEvent::REGISTRY>. You can do that before and even without loading
143	AnyEvent.	1871	AnyEvent, so it is reasonably cheap.
144		1872
145	Example:	1873	Example:
146		1874
147	push @AnyEvent::REGISTRY, [urxvt => urxvt::anyevent::];	1875	push @AnyEvent::REGISTRY, [urxvt => urxvt::anyevent::];
148		1876
149	This tells AnyEvent to (literally) use the C<urxvt::anyevent::> module	1877	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	1878	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		1879
		1880	When AnyEvent is loaded and asked to find a suitable event model, it
		1881	will first check for the presence of urxvt by trying to C<use> the
		1882	C<urxvt::anyevent> module.
		1883
		1884	The class should provide implementations for all watcher types. See
		1885	L<AnyEvent::Impl::EV> (source code), L<AnyEvent::Impl::Glib> (Source code)
		1886	and so on for actual examples. Use C<perldoc -m AnyEvent::Impl::Glib> to
		1887	see the sources.
		1888
		1889	If you don't provide C<signal> and C<child> watchers than AnyEvent will
		1890	provide suitable (hopefully) replacements.
		1891
154	The above isn't fictitious, the I<rxvt-unicode> (a.k.a. urxvt) uses	1892	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	1893	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	1894	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	1895	inside I<rxvt-unicode>, and it is updated and maintained as part of the
158	I<rxvt-unicode> distribution.	1896	I<rxvt-unicode> distribution.
159		1897
160	=head1 ENVIRONMENT VARIABLES	1898	I<rxvt-unicode> also cheats a bit by not providing blocking access to
		1899	condition variables: code blocking while waiting for a condition will
		1900	C<die>. This still works with most modules/usages, and blocking calls must
		1901	not be done in an interactive application, so it makes sense.
161		1902
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	1903	=head1 EXAMPLE PROGRAM
168		1904
169	The following program uses an io watcher to read data from stdin, a timer	1905	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	1906	to display a message once per second, and a condition variable to quit the
171	when the user enters quit:	1907	program when the user enters quit:
172		1908
173	use AnyEvent;	1909	use AnyEvent;
174		1910
175	my $cv = AnyEvent->condvar;	1911	my $cv = AnyEvent->condvar;
176		1912
177	my $io_watcher = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {	1913	my $io_watcher = AnyEvent->io (
		1914	fh => \*STDIN,
		1915	poll => 'r',
		1916	cb => sub {
178	warn "io event <$_[0]>\n"; # will always output <r>	1917	warn "io event <$_[0]>\n"; # will always output <r>
179	chomp (my $input = <STDIN>); # read a line	1918	chomp (my $input = <STDIN>); # read a line
180	warn "read: $input\n"; # output what has been read	1919	warn "read: $input\n"; # output what has been read
181	$cv->broadcast if $input =~ /^q/i; # quit program if /^q/i	1920	$cv->send if $input =~ /^q/i; # quit program if /^q/i
		1921	},
		1922	);
		1923
		1924	my $time_watcher = AnyEvent->timer (after => 1, interval => 1, cb => sub {
		1925	warn "timeout\n"; # print 'timeout' at most every second
182	});	1926	});
183		1927
184	my $time_watcher; # can only be used once
185
186	sub new_timer {
187	$timer = AnyEvent->timer (after => 1, cb => sub {
188	warn "timeout\n"; # print 'timeout' about every second
189	&new_timer; # and restart the time
190	});
191	}
192
193	new_timer; # create first timer
194
195	$cv->wait; # wait until user enters /^q/i	1928	$cv->recv; # wait until user enters /^q/i
196		1929
197	=head1 REAL-WORLD EXAMPLE	1930	=head1 REAL-WORLD EXAMPLE
198		1931
199	Consider the L<Net::FCP> module. It features (among others) the following	1932	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:	1933	API calls, which are to freenet what HTTP GET requests are to http:
…		…
250	syswrite $txn->{fh}, $txn->{request}	1983	syswrite $txn->{fh}, $txn->{request}
251	or die "connection or write error";	1984	or die "connection or write error";
252	$txn->{w} = AnyEvent->io (fh => $txn->{fh}, poll => 'r', cb => sub { $txn->fh_ready_r });	1985	$txn->{w} = AnyEvent->io (fh => $txn->{fh}, poll => 'r', cb => sub { $txn->fh_ready_r });
253		1986
254	Again, C<fh_ready_r> waits till all data has arrived, and then stores the	1987	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:	1988	result and signals any possible waiters that the request has finished:
256		1989
257	sysread $txn->{fh}, $txn->{buf}, length $txn->{$buf};	1990	sysread $txn->{fh}, $txn->{buf}, length $txn->{$buf};
258		1991
259	if (end-of-file or data complete) {	1992	if (end-of-file or data complete) {
260	$txn->{result} = $txn->{buf};	1993	$txn->{result} = $txn->{buf};
261	$txn->{finished}->broadcast;	1994	$txn->{finished}->send;
262	$txb->{cb}->($txn) of $txn->{cb}; # also call callback	1995	$txb->{cb}->($txn) of $txn->{cb}; # also call callback
263	}	1996	}
264		1997
265	The C<result> method, finally, just waits for the finished signal (if the	1998	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	1999	request was already finished, it doesn't wait, of course, and returns the
267	data:	2000	data:
268		2001
269	$txn->{finished}->wait;	2002	$txn->{finished}->recv;
270	return $txn->{result};	2003	return $txn->{result};
271		2004
272	The actual code goes further and collects all errors (C<die>s, exceptions)	2005	The actual code goes further and collects all errors (C<die>s, exceptions)
273	that occured during request processing. The C<result> method detects	2006	that occurred during request processing. The C<result> method detects
274	wether an exception as thrown (it is stored inside the $txn object)	2007	whether an exception as thrown (it is stored inside the $txn object)
275	and just throws the exception, which means connection errors and other	2008	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	2009	problems get reported tot he code that tries to use the result, not in a
277	random callback.	2010	random callback.
278		2011
279	All of this enables the following usage styles:	2012	All of this enables the following usage styles:
280		2013
281	1. Blocking:	2014	1. Blocking:
282		2015
283	my $data = $fcp->client_get ($url);	2016	my $data = $fcp->client_get ($url);
284		2017
285	2. Blocking, but parallelizing:	2018	2. Blocking, but running in parallel:
286		2019
287	my @datas = map $_->result,	2020	my @datas = map $_->result,
288	map $fcp->txn_client_get ($_),	2021	map $fcp->txn_client_get ($_),
289	@urls;	2022	@urls;
290		2023
291	Both blocking examples work without the module user having to know	2024	Both blocking examples work without the module user having to know
292	anything about events.	2025	anything about events.
293		2026
294	3a. Event-based in a main program, using any support Event module:	2027	3a. Event-based in a main program, using any supported event module:
295		2028
296	use Event;	2029	use EV;
297		2030
298	$fcp->txn_client_get ($url)->cb (sub {	2031	$fcp->txn_client_get ($url)->cb (sub {
299	my $txn = shift;	2032	my $txn = shift;
300	my $data = $txn->result;	2033	my $data = $txn->result;
301	...	2034	...
302	});	2035	});
303		2036
304	Event::loop;	2037	EV::loop;
305		2038
306	3b. The module user could use AnyEvent, too:	2039	3b. The module user could use AnyEvent, too:
307		2040
308	use AnyEvent;	2041	use AnyEvent;
309		2042
310	my $quit = AnyEvent->condvar;	2043	my $quit = AnyEvent->condvar;
311		2044
312	$fcp->txn_client_get ($url)->cb (sub {	2045	$fcp->txn_client_get ($url)->cb (sub {
313	...	2046	...
314	$quit->broadcast;	2047	$quit->send;
315	});	2048	});
316		2049
317	$quit->wait;	2050	$quit->recv;
		2051
		2052
		2053	=head1 BENCHMARKS
		2054
		2055	To give you an idea of the performance and overheads that AnyEvent adds
		2056	over the event loops themselves and to give you an impression of the speed
		2057	of various event loops I prepared some benchmarks.
		2058
		2059	=head2 BENCHMARKING ANYEVENT OVERHEAD
		2060
		2061	Here is a benchmark of various supported event models used natively and
		2062	through AnyEvent. The benchmark creates a lot of timers (with a zero
		2063	timeout) and I/O watchers (watching STDOUT, a pty, to become writable,
		2064	which it is), lets them fire exactly once and destroys them again.
		2065
		2066	Source code for this benchmark is found as F<eg/bench> in the AnyEvent
		2067	distribution. It uses the L<AE> interface, which makes a real difference
		2068	for the EV and Perl backends only.
		2069
		2070	=head3 Explanation of the columns
		2071
		2072	I<watcher> is the number of event watchers created/destroyed. Since
		2073	different event models feature vastly different performances, each event
		2074	loop was given a number of watchers so that overall runtime is acceptable
		2075	and similar between tested event loop (and keep them from crashing): Glib
		2076	would probably take thousands of years if asked to process the same number
		2077	of watchers as EV in this benchmark.
		2078
		2079	I<bytes> is the number of bytes (as measured by the resident set size,
		2080	RSS) consumed by each watcher. This method of measuring captures both C
		2081	and Perl-based overheads.
		2082
		2083	I<create> is the time, in microseconds (millionths of seconds), that it
		2084	takes to create a single watcher. The callback is a closure shared between
		2085	all watchers, to avoid adding memory overhead. That means closure creation
		2086	and memory usage is not included in the figures.
		2087
		2088	I<invoke> is the time, in microseconds, used to invoke a simple
		2089	callback. The callback simply counts down a Perl variable and after it was
		2090	invoked "watcher" times, it would C<< ->send >> a condvar once to
		2091	signal the end of this phase.
		2092
		2093	I<destroy> is the time, in microseconds, that it takes to destroy a single
		2094	watcher.
		2095
		2096	=head3 Results
		2097
		2098	name watchers bytes create invoke destroy comment
		2099	EV/EV 100000 223 0.47 0.43 0.27 EV native interface
		2100	EV/Any 100000 223 0.48 0.42 0.26 EV + AnyEvent watchers
		2101	Coro::EV/Any 100000 223 0.47 0.42 0.26 coroutines + Coro::Signal
		2102	Perl/Any 100000 431 2.70 0.74 0.92 pure perl implementation
		2103	Event/Event 16000 516 31.16 31.84 0.82 Event native interface
		2104	Event/Any 16000 1203 42.61 34.79 1.80 Event + AnyEvent watchers
		2105	IOAsync/Any 16000 1911 41.92 27.45 16.81 via IO::Async::Loop::IO_Poll
		2106	IOAsync/Any 16000 1726 40.69 26.37 15.25 via IO::Async::Loop::Epoll
		2107	Glib/Any 16000 1118 89.00 12.57 51.17 quadratic behaviour
		2108	Tk/Any 2000 1346 20.96 10.75 8.00 SEGV with >> 2000 watchers
		2109	POE/Any 2000 6951 108.97 795.32 14.24 via POE::Loop::Event
		2110	POE/Any 2000 6648 94.79 774.40 575.51 via POE::Loop::Select
		2111
		2112	=head3 Discussion
		2113
		2114	The benchmark does I<not> measure scalability of the event loop very
		2115	well. For example, a select-based event loop (such as the pure perl one)
		2116	can never compete with an event loop that uses epoll when the number of
		2117	file descriptors grows high. In this benchmark, all events become ready at
		2118	the same time, so select/poll-based implementations get an unnatural speed
		2119	boost.
		2120
		2121	Also, note that the number of watchers usually has a nonlinear effect on
		2122	overall speed, that is, creating twice as many watchers doesn't take twice
		2123	the time - usually it takes longer. This puts event loops tested with a
		2124	higher number of watchers at a disadvantage.
		2125
		2126	To put the range of results into perspective, consider that on the
		2127	benchmark machine, handling an event takes roughly 1600 CPU cycles with
		2128	EV, 3100 CPU cycles with AnyEvent's pure perl loop and almost 3000000 CPU
		2129	cycles with POE.
		2130
		2131	C<EV> is the sole leader regarding speed and memory use, which are both
		2132	maximal/minimal, respectively. When using the L<AE> API there is zero
		2133	overhead (when going through the AnyEvent API create is about 5-6 times
		2134	slower, with other times being equal, so still uses far less memory than
		2135	any other event loop and is still faster than Event natively).
		2136
		2137	The pure perl implementation is hit in a few sweet spots (both the
		2138	constant timeout and the use of a single fd hit optimisations in the perl
		2139	interpreter and the backend itself). Nevertheless this shows that it
		2140	adds very little overhead in itself. Like any select-based backend its
		2141	performance becomes really bad with lots of file descriptors (and few of
		2142	them active), of course, but this was not subject of this benchmark.
		2143
		2144	The C<Event> module has a relatively high setup and callback invocation
		2145	cost, but overall scores in on the third place.
		2146
		2147	C<IO::Async> performs admirably well, about on par with C<Event>, even
		2148	when using its pure perl backend.
		2149
		2150	C<Glib>'s memory usage is quite a bit higher, but it features a
		2151	faster callback invocation and overall ends up in the same class as
		2152	C<Event>. However, Glib scales extremely badly, doubling the number of
		2153	watchers increases the processing time by more than a factor of four,
		2154	making it completely unusable when using larger numbers of watchers
		2155	(note that only a single file descriptor was used in the benchmark, so
		2156	inefficiencies of C<poll> do not account for this).
		2157
		2158	The C<Tk> adaptor works relatively well. The fact that it crashes with
		2159	more than 2000 watchers is a big setback, however, as correctness takes
		2160	precedence over speed. Nevertheless, its performance is surprising, as the
		2161	file descriptor is dup()ed for each watcher. This shows that the dup()
		2162	employed by some adaptors is not a big performance issue (it does incur a
		2163	hidden memory cost inside the kernel which is not reflected in the figures
		2164	above).
		2165
		2166	C<POE>, regardless of underlying event loop (whether using its pure perl
		2167	select-based backend or the Event module, the POE-EV backend couldn't
		2168	be tested because it wasn't working) shows abysmal performance and
		2169	memory usage with AnyEvent: Watchers use almost 30 times as much memory
		2170	as EV watchers, and 10 times as much memory as Event (the high memory
		2171	requirements are caused by requiring a session for each watcher). Watcher
		2172	invocation speed is almost 900 times slower than with AnyEvent's pure perl
		2173	implementation.
		2174
		2175	The design of the POE adaptor class in AnyEvent can not really account
		2176	for the performance issues, though, as session creation overhead is
		2177	small compared to execution of the state machine, which is coded pretty
		2178	optimally within L<AnyEvent::Impl::POE> (and while everybody agrees that
		2179	using multiple sessions is not a good approach, especially regarding
		2180	memory usage, even the author of POE could not come up with a faster
		2181	design).
		2182
		2183	=head3 Summary
		2184
		2185	=over 4
		2186
		2187	=item * Using EV through AnyEvent is faster than any other event loop
		2188	(even when used without AnyEvent), but most event loops have acceptable
		2189	performance with or without AnyEvent.
		2190
		2191	=item * The overhead AnyEvent adds is usually much smaller than the overhead of
		2192	the actual event loop, only with extremely fast event loops such as EV
		2193	adds AnyEvent significant overhead.
		2194
		2195	=item * You should avoid POE like the plague if you want performance or
		2196	reasonable memory usage.
		2197
		2198	=back
		2199
		2200	=head2 BENCHMARKING THE LARGE SERVER CASE
		2201
		2202	This benchmark actually benchmarks the event loop itself. It works by
		2203	creating a number of "servers": each server consists of a socket pair, a
		2204	timeout watcher that gets reset on activity (but never fires), and an I/O
		2205	watcher waiting for input on one side of the socket. Each time the socket
		2206	watcher reads a byte it will write that byte to a random other "server".
		2207
		2208	The effect is that there will be a lot of I/O watchers, only part of which
		2209	are active at any one point (so there is a constant number of active
		2210	fds for each loop iteration, but which fds these are is random). The
		2211	timeout is reset each time something is read because that reflects how
		2212	most timeouts work (and puts extra pressure on the event loops).
		2213
		2214	In this benchmark, we use 10000 socket pairs (20000 sockets), of which 100
		2215	(1%) are active. This mirrors the activity of large servers with many
		2216	connections, most of which are idle at any one point in time.
		2217
		2218	Source code for this benchmark is found as F<eg/bench2> in the AnyEvent
		2219	distribution. It uses the L<AE> interface, which makes a real difference
		2220	for the EV and Perl backends only.
		2221
		2222	=head3 Explanation of the columns
		2223
		2224	I<sockets> is the number of sockets, and twice the number of "servers" (as
		2225	each server has a read and write socket end).
		2226
		2227	I<create> is the time it takes to create a socket pair (which is
		2228	nontrivial) and two watchers: an I/O watcher and a timeout watcher.
		2229
		2230	I<request>, the most important value, is the time it takes to handle a
		2231	single "request", that is, reading the token from the pipe and forwarding
		2232	it to another server. This includes deleting the old timeout and creating
		2233	a new one that moves the timeout into the future.
		2234
		2235	=head3 Results
		2236
		2237	name sockets create request
		2238	EV 20000 62.66 7.99
		2239	Perl 20000 68.32 32.64
		2240	IOAsync 20000 174.06 101.15 epoll
		2241	IOAsync 20000 174.67 610.84 poll
		2242	Event 20000 202.69 242.91
		2243	Glib 20000 557.01 1689.52
		2244	POE 20000 341.54 12086.32 uses POE::Loop::Event
		2245
		2246	=head3 Discussion
		2247
		2248	This benchmark I<does> measure scalability and overall performance of the
		2249	particular event loop.
		2250
		2251	EV is again fastest. Since it is using epoll on my system, the setup time
		2252	is relatively high, though.
		2253
		2254	Perl surprisingly comes second. It is much faster than the C-based event
		2255	loops Event and Glib.
		2256
		2257	IO::Async performs very well when using its epoll backend, and still quite
		2258	good compared to Glib when using its pure perl backend.
		2259
		2260	Event suffers from high setup time as well (look at its code and you will
		2261	understand why). Callback invocation also has a high overhead compared to
		2262	the C<< $_->() for .. >>-style loop that the Perl event loop uses. Event
		2263	uses select or poll in basically all documented configurations.
		2264
		2265	Glib is hit hard by its quadratic behaviour w.r.t. many watchers. It
		2266	clearly fails to perform with many filehandles or in busy servers.
		2267
		2268	POE is still completely out of the picture, taking over 1000 times as long
		2269	as EV, and over 100 times as long as the Perl implementation, even though
		2270	it uses a C-based event loop in this case.
		2271
		2272	=head3 Summary
		2273
		2274	=over 4
		2275
		2276	=item * The pure perl implementation performs extremely well.
		2277
		2278	=item * Avoid Glib or POE in large projects where performance matters.
		2279
		2280	=back
		2281
		2282	=head2 BENCHMARKING SMALL SERVERS
		2283
		2284	While event loops should scale (and select-based ones do not...) even to
		2285	large servers, most programs we (or I :) actually write have only a few
		2286	I/O watchers.
		2287
		2288	In this benchmark, I use the same benchmark program as in the large server
		2289	case, but it uses only eight "servers", of which three are active at any
		2290	one time. This should reflect performance for a small server relatively
		2291	well.
		2292
		2293	The columns are identical to the previous table.
		2294
		2295	=head3 Results
		2296
		2297	name sockets create request
		2298	EV 16 20.00 6.54
		2299	Perl 16 25.75 12.62
		2300	Event 16 81.27 35.86
		2301	Glib 16 32.63 15.48
		2302	POE 16 261.87 276.28 uses POE::Loop::Event
		2303
		2304	=head3 Discussion
		2305
		2306	The benchmark tries to test the performance of a typical small
		2307	server. While knowing how various event loops perform is interesting, keep
		2308	in mind that their overhead in this case is usually not as important, due
		2309	to the small absolute number of watchers (that is, you need efficiency and
		2310	speed most when you have lots of watchers, not when you only have a few of
		2311	them).
		2312
		2313	EV is again fastest.
		2314
		2315	Perl again comes second. It is noticeably faster than the C-based event
		2316	loops Event and Glib, although the difference is too small to really
		2317	matter.
		2318
		2319	POE also performs much better in this case, but is is still far behind the
		2320	others.
		2321
		2322	=head3 Summary
		2323
		2324	=over 4
		2325
		2326	=item * C-based event loops perform very well with small number of
		2327	watchers, as the management overhead dominates.
		2328
		2329	=back
		2330
		2331	=head2 THE IO::Lambda BENCHMARK
		2332
		2333	Recently I was told about the benchmark in the IO::Lambda manpage, which
		2334	could be misinterpreted to make AnyEvent look bad. In fact, the benchmark
		2335	simply compares IO::Lambda with POE, and IO::Lambda looks better (which
		2336	shouldn't come as a surprise to anybody). As such, the benchmark is
		2337	fine, and mostly shows that the AnyEvent backend from IO::Lambda isn't
		2338	very optimal. But how would AnyEvent compare when used without the extra
		2339	baggage? To explore this, I wrote the equivalent benchmark for AnyEvent.
		2340
		2341	The benchmark itself creates an echo-server, and then, for 500 times,
		2342	connects to the echo server, sends a line, waits for the reply, and then
		2343	creates the next connection. This is a rather bad benchmark, as it doesn't
		2344	test the efficiency of the framework or much non-blocking I/O, but it is a
		2345	benchmark nevertheless.
		2346
		2347	name runtime
		2348	Lambda/select 0.330 sec
		2349	+ optimized 0.122 sec
		2350	Lambda/AnyEvent 0.327 sec
		2351	+ optimized 0.138 sec
		2352	Raw sockets/select 0.077 sec
		2353	POE/select, components 0.662 sec
		2354	POE/select, raw sockets 0.226 sec
		2355	POE/select, optimized 0.404 sec
		2356
		2357	AnyEvent/select/nb 0.085 sec
		2358	AnyEvent/EV/nb 0.068 sec
		2359	+state machine 0.134 sec
		2360
		2361	The benchmark is also a bit unfair (my fault): the IO::Lambda/POE
		2362	benchmarks actually make blocking connects and use 100% blocking I/O,
		2363	defeating the purpose of an event-based solution. All of the newly
		2364	written AnyEvent benchmarks use 100% non-blocking connects (using
		2365	AnyEvent::Socket::tcp_connect and the asynchronous pure perl DNS
		2366	resolver), so AnyEvent is at a disadvantage here, as non-blocking connects
		2367	generally require a lot more bookkeeping and event handling than blocking
		2368	connects (which involve a single syscall only).
		2369
		2370	The last AnyEvent benchmark additionally uses L<AnyEvent::Handle>, which
		2371	offers similar expressive power as POE and IO::Lambda, using conventional
		2372	Perl syntax. This means that both the echo server and the client are 100%
		2373	non-blocking, further placing it at a disadvantage.
		2374
		2375	As you can see, the AnyEvent + EV combination even beats the
		2376	hand-optimised "raw sockets benchmark", while AnyEvent + its pure perl
		2377	backend easily beats IO::Lambda and POE.
		2378
		2379	And even the 100% non-blocking version written using the high-level (and
		2380	slow :) L<AnyEvent::Handle> abstraction beats both POE and IO::Lambda
		2381	higher level ("unoptimised") abstractions by a large margin, even though
		2382	it does all of DNS, tcp-connect and socket I/O in a non-blocking way.
		2383
		2384	The two AnyEvent benchmarks programs can be found as F<eg/ae0.pl> and
		2385	F<eg/ae2.pl> in the AnyEvent distribution, the remaining benchmarks are
		2386	part of the IO::Lambda distribution and were used without any changes.
		2387
		2388
		2389	=head1 SIGNALS
		2390
		2391	AnyEvent currently installs handlers for these signals:
		2392
		2393	=over 4
		2394
		2395	=item SIGCHLD
		2396
		2397	A handler for C<SIGCHLD> is installed by AnyEvent's child watcher
		2398	emulation for event loops that do not support them natively. Also, some
		2399	event loops install a similar handler.
		2400
		2401	Additionally, when AnyEvent is loaded and SIGCHLD is set to IGNORE, then
		2402	AnyEvent will reset it to default, to avoid losing child exit statuses.
		2403
		2404	=item SIGPIPE
		2405
		2406	A no-op handler is installed for C<SIGPIPE> when C<$SIG{PIPE}> is C<undef>
		2407	when AnyEvent gets loaded.
		2408
		2409	The rationale for this is that AnyEvent users usually do not really depend
		2410	on SIGPIPE delivery (which is purely an optimisation for shell use, or
		2411	badly-written programs), but C<SIGPIPE> can cause spurious and rare
		2412	program exits as a lot of people do not expect C<SIGPIPE> when writing to
		2413	some random socket.
		2414
		2415	The rationale for installing a no-op handler as opposed to ignoring it is
		2416	that this way, the handler will be restored to defaults on exec.
		2417
		2418	Feel free to install your own handler, or reset it to defaults.
		2419
		2420	=back
		2421
		2422	=cut
		2423
		2424	undef $SIG{CHLD}
		2425	if $SIG{CHLD} eq 'IGNORE';
		2426
		2427	$SIG{PIPE} = sub { }
		2428	unless defined $SIG{PIPE};
		2429
		2430	=head1 RECOMMENDED/OPTIONAL MODULES
		2431
		2432	One of AnyEvent's main goals is to be 100% Pure-Perl(tm): only perl (and
		2433	it's built-in modules) are required to use it.
		2434
		2435	That does not mean that AnyEvent won't take advantage of some additional
		2436	modules if they are installed.
		2437
		2438	This section epxlains which additional modules will be used, and how they
		2439	affect AnyEvent's operetion.
		2440
		2441	=over 4
		2442
		2443	=item L<Async::Interrupt>
		2444
		2445	This slightly arcane module is used to implement fast signal handling: To
		2446	my knowledge, there is no way to do completely race-free and quick
		2447	signal handling in pure perl. To ensure that signals still get
		2448	delivered, AnyEvent will start an interval timer to wake up perl (and
		2449	catch the signals) with some delay (default is 10 seconds, look for
		2450	C<$AnyEvent::MAX_SIGNAL_LATENCY>).
		2451
		2452	If this module is available, then it will be used to implement signal
		2453	catching, which means that signals will not be delayed, and the event loop
		2454	will not be interrupted regularly, which is more efficient (And good for
		2455	battery life on laptops).
		2456
		2457	This affects not just the pure-perl event loop, but also other event loops
		2458	that have no signal handling on their own (e.g. Glib, Tk, Qt).
		2459
		2460	Some event loops (POE, Event, Event::Lib) offer signal watchers natively,
		2461	and either employ their own workarounds (POE) or use AnyEvent's workaround
		2462	(using C<$AnyEvent::MAX_SIGNAL_LATENCY>). Installing L<Async::Interrupt>
		2463	does nothing for those backends.
		2464
		2465	=item L<EV>
		2466
		2467	This module isn't really "optional", as it is simply one of the backend
		2468	event loops that AnyEvent can use. However, it is simply the best event
		2469	loop available in terms of features, speed and stability: It supports
		2470	the AnyEvent API optimally, implements all the watcher types in XS, does
		2471	automatic timer adjustments even when no monotonic clock is available,
		2472	can take avdantage of advanced kernel interfaces such as C<epoll> and
		2473	C<kqueue>, and is the fastest backend I<by far>. You can even embed
		2474	L<Glib>/L<Gtk2> in it (or vice versa, see L<EV::Glib> and L<Glib::EV>).
		2475
		2476	=item L<Guard>
		2477
		2478	The guard module, when used, will be used to implement
		2479	C<AnyEvent::Util::guard>. This speeds up guards considerably (and uses a
		2480	lot less memory), but otherwise doesn't affect guard operation much. It is
		2481	purely used for performance.
		2482
		2483	=item L<JSON> and L<JSON::XS>
		2484
		2485	One of these modules is required when you want to read or write JSON data
		2486	via L<AnyEvent::Handle>. It is also written in pure-perl, but can take
		2487	advantage of the ultra-high-speed L<JSON::XS> module when it is installed.
		2488
		2489	In fact, L<AnyEvent::Handle> will use L<JSON::XS> by default if it is
		2490	installed.
		2491
		2492	=item L<Net::SSLeay>
		2493
		2494	Implementing TLS/SSL in Perl is certainly interesting, but not very
		2495	worthwhile: If this module is installed, then L<AnyEvent::Handle> (with
		2496	the help of L<AnyEvent::TLS>), gains the ability to do TLS/SSL.
		2497
		2498	=item L<Time::HiRes>
		2499
		2500	This module is part of perl since release 5.008. It will be used when the
		2501	chosen event library does not come with a timing source on it's own. The
		2502	pure-perl event loop (L<AnyEvent::Impl::Perl>) will additionally use it to
		2503	try to use a monotonic clock for timing stability.
		2504
		2505	=back
		2506
		2507
		2508	=head1 FORK
		2509
		2510	Most event libraries are not fork-safe. The ones who are usually are
		2511	because they rely on inefficient but fork-safe C<select> or C<poll>
		2512	calls. Only L<EV> is fully fork-aware.
		2513
		2514	If you have to fork, you must either do so I<before> creating your first
		2515	watcher OR you must not use AnyEvent at all in the child OR you must do
		2516	something completely out of the scope of AnyEvent.
		2517
		2518
		2519	=head1 SECURITY CONSIDERATIONS
		2520
		2521	AnyEvent can be forced to load any event model via
		2522	$ENV{PERL_ANYEVENT_MODEL}. While this cannot (to my knowledge) be used to
		2523	execute arbitrary code or directly gain access, it can easily be used to
		2524	make the program hang or malfunction in subtle ways, as AnyEvent watchers
		2525	will not be active when the program uses a different event model than
		2526	specified in the variable.
		2527
		2528	You can make AnyEvent completely ignore this variable by deleting it
		2529	before the first watcher gets created, e.g. with a C<BEGIN> block:
		2530
		2531	BEGIN { delete $ENV{PERL_ANYEVENT_MODEL} }
		2532
		2533	use AnyEvent;
		2534
		2535	Similar considerations apply to $ENV{PERL_ANYEVENT_VERBOSE}, as that can
		2536	be used to probe what backend is used and gain other information (which is
		2537	probably even less useful to an attacker than PERL_ANYEVENT_MODEL), and
		2538	$ENV{PERL_ANYEVENT_STRICT}.
		2539
		2540	Note that AnyEvent will remove I<all> environment variables starting with
		2541	C<PERL_ANYEVENT_> from C<%ENV> when it is loaded while taint mode is
		2542	enabled.
		2543
		2544
		2545	=head1 BUGS
		2546
		2547	Perl 5.8 has numerous memleaks that sometimes hit this module and are hard
		2548	to work around. If you suffer from memleaks, first upgrade to Perl 5.10
		2549	and check wether the leaks still show up. (Perl 5.10.0 has other annoying
		2550	memleaks, such as leaking on C<map> and C<grep> but it is usually not as
		2551	pronounced).
		2552
318		2553
319	=head1 SEE ALSO	2554	=head1 SEE ALSO
320		2555
321	Event modules: L<Coro::Event>, L<Coro>, L<Event>, L<Glib::Event>, L<Glib>.	2556	Utility functions: L<AnyEvent::Util>.
322		2557
323	Implementations: L<AnyEvent::Impl::Coro>, L<AnyEvent::Impl::Event>, L<AnyEvent::Impl::Glib>, L<AnyEvent::Impl::Tk>.	2558	Event modules: L<EV>, L<EV::Glib>, L<Glib::EV>, L<Event>, L<Glib::Event>,
		2559	L<Glib>, L<Tk>, L<Event::Lib>, L<Qt>, L<POE>.
324		2560
325	Nontrivial usage example: L<Net::FCP>.	2561	Implementations: L<AnyEvent::Impl::EV>, L<AnyEvent::Impl::Event>,
		2562	L<AnyEvent::Impl::Glib>, L<AnyEvent::Impl::Tk>, L<AnyEvent::Impl::Perl>,
		2563	L<AnyEvent::Impl::EventLib>, L<AnyEvent::Impl::Qt>,
		2564	L<AnyEvent::Impl::POE>, L<AnyEvent::Impl::IOAsync>, L<Anyevent::Impl::Irssi>.
326		2565
327	=head1	2566	Non-blocking file handles, sockets, TCP clients and
		2567	servers: L<AnyEvent::Handle>, L<AnyEvent::Socket>, L<AnyEvent::TLS>.
		2568
		2569	Asynchronous DNS: L<AnyEvent::DNS>.
		2570
		2571	Coroutine support: L<Coro>, L<Coro::AnyEvent>, L<Coro::EV>,
		2572	L<Coro::Event>,
		2573
		2574	Nontrivial usage examples: L<AnyEvent::GPSD>, L<AnyEvent::XMPP>,
		2575	L<AnyEvent::HTTP>.
		2576
		2577
		2578	=head1 AUTHOR
		2579
		2580	Marc Lehmann <schmorp@schmorp.de>
		2581	http://home.schmorp.de/
328		2582
329	=cut	2583	=cut
330		2584
331	1	2585	1
332		2586

Diff Legend

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