ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/AnyEvent/lib/AnyEvent.pm

Comparing AnyEvent/lib/AnyEvent.pm (file contents):
Revision 1.100 by elmex, Sun Apr 27 19:15:43 2008 UTC vs.
Revision 1.328 by root, Tue Jun 8 10:06:15 2010 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	EV, Event, Coro::EV, Coro::Event, Glib, Tk, Perl, Event::Lib, Qt, POE - 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	# if you prefer function calls, look at the AE manpage for
		13	# an alternative API.
		14
		15	# file handle or descriptor readable
11	my $w = AnyEvent->io (fh => $fh, poll => "r|w", cb => sub {	16	my $w = AnyEvent->io (fh => $fh, poll => "r", cb => sub { ... });
		17
		18	# one-shot or repeating timers
		19	my $w = AnyEvent->timer (after => $seconds, cb => sub { ... });
		20	my $w = AnyEvent->timer (after => $seconds, interval => $seconds, cb => ...
		21
		22	print AnyEvent->now; # prints current event loop time
		23	print AnyEvent->time; # think Time::HiRes::time or simply CORE::time.
		24
		25	# POSIX signal
		26	my $w = AnyEvent->signal (signal => "TERM", cb => sub { ... });
		27
		28	# child process exit
		29	my $w = AnyEvent->child (pid => $pid, cb => sub {
		30	my ($pid, $status) = @_;
12	...	31	...
13	});	32	});
14		33
15	my $w = AnyEvent->timer (after => $seconds, cb => sub {	34	# called when event loop idle (if applicable)
16	...	35	my $w = AnyEvent->idle (cb => sub { ... });
17	});
18		36
19	my $w = AnyEvent->condvar; # stores whether a condition was flagged	37	my $w = AnyEvent->condvar; # stores whether a condition was flagged
		38	$w->send; # wake up current and all future recv's
20	$w->wait; # enters "main loop" till $condvar gets ->broadcast	39	$w->recv; # enters "main loop" till $condvar gets ->send
21	$w->broadcast; # wake up current and all future wait's	40	# use a condvar in callback mode:
		41	$w->cb (sub { $_[0]->recv });
		42
		43	=head1 INTRODUCTION/TUTORIAL
		44
		45	This manpage is mainly a reference manual. If you are interested
		46	in a tutorial or some gentle introduction, have a look at the
		47	L<AnyEvent::Intro> manpage.
		48
		49	=head1 SUPPORT
		50
		51	There is a mailinglist for discussing all things AnyEvent, and an IRC
		52	channel, too.
		53
		54	See the AnyEvent project page at the B<Schmorpforge Ta-Sa Software
		55	Repository>, at L<http://anyevent.schmorp.de>, for more info.
22		56
23	=head1 WHY YOU SHOULD USE THIS MODULE (OR NOT)	57	=head1 WHY YOU SHOULD USE THIS MODULE (OR NOT)
24		58
25	Glib, POE, IO::Async, Event... CPAN offers event models by the dozen	59	Glib, POE, IO::Async, Event... CPAN offers event models by the dozen
26	nowadays. So what is different about AnyEvent?	60	nowadays. So what is different about AnyEvent?
27		61
28	Executive Summary: AnyEvent is I<compatible>, AnyEvent is I<free of	62	Executive Summary: AnyEvent is I<compatible>, AnyEvent is I<free of
29	policy> and AnyEvent is I<small and efficient>.	63	policy> and AnyEvent is I<small and efficient>.
30		64
31	First and foremost, I<AnyEvent is not an event model> itself, it only	65	First and foremost, I<AnyEvent is not an event model> itself, it only
32	interfaces to whatever event model the main program happens to use in a	66	interfaces to whatever event model the main program happens to use, in a
33	pragmatic way. For event models and certain classes of immortals alike,	67	pragmatic way. For event models and certain classes of immortals alike,
34	the statement "there can only be one" is a bitter reality: In general,	68	the statement "there can only be one" is a bitter reality: In general,
35	only one event loop can be active at the same time in a process. AnyEvent	69	only one event loop can be active at the same time in a process. AnyEvent
36	helps hiding the differences between those event loops.	70	cannot change this, but it can hide the differences between those event
		71	loops.
37		72
38	The goal of AnyEvent is to offer module authors the ability to do event	73	The goal of AnyEvent is to offer module authors the ability to do event
39	programming (waiting for I/O or timer events) without subscribing to a	74	programming (waiting for I/O or timer events) without subscribing to a
40	religion, a way of living, and most importantly: without forcing your	75	religion, a way of living, and most importantly: without forcing your
41	module users into the same thing by forcing them to use the same event	76	module users into the same thing by forcing them to use the same event
42	model you use.	77	model you use.
43		78
44	For modules like POE or IO::Async (which is a total misnomer as it is	79	For modules like POE or IO::Async (which is a total misnomer as it is
45	actually doing all I/O I<synchronously>...), using them in your module is	80	actually doing all I/O I<synchronously>...), using them in your module is
46	like joining a cult: After you joined, you are dependent on them and you	81	like joining a cult: After you joined, you are dependent on them and you
47	cannot use anything else, as it is simply incompatible to everything that	82	cannot use anything else, as they are simply incompatible to everything
48	isn't itself. What's worse, all the potential users of your module are	83	that isn't them. What's worse, all the potential users of your
49	I<also> forced to use the same event loop you use.	84	module are I<also> forced to use the same event loop you use.
50		85
51	AnyEvent is different: AnyEvent + POE works fine. AnyEvent + Glib works	86	AnyEvent is different: AnyEvent + POE works fine. AnyEvent + Glib works
52	fine. AnyEvent + Tk works fine etc. etc. but none of these work together	87	fine. AnyEvent + Tk works fine etc. etc. but none of these work together
53	with the rest: POE + IO::Async? no go. Tk + Event? no go. Again: if	88	with the rest: POE + IO::Async? No go. Tk + Event? No go. Again: if
54	your module uses one of those, every user of your module has to use it,	89	your module uses one of those, every user of your module has to use it,
55	too. But if your module uses AnyEvent, it works transparently with all	90	too. But if your module uses AnyEvent, it works transparently with all
56	event models it supports (including stuff like POE and IO::Async, as long	91	event models it supports (including stuff like IO::Async, as long as those
57	as those use one of the supported event loops. It is trivial to add new	92	use one of the supported event loops. It is trivial to add new event loops
58	event loops to AnyEvent, too, so it is future-proof).	93	to AnyEvent, too, so it is future-proof).
59		94
60	In addition to being free of having to use I<the one and only true event	95	In addition to being free of having to use I<the one and only true event
61	model>, AnyEvent also is free of bloat and policy: with POE or similar	96	model>, AnyEvent also is free of bloat and policy: with POE or similar
62	modules, you get an enourmous amount of code and strict rules you have to	97	modules, you get an enormous amount of code and strict rules you have to
63	follow. AnyEvent, on the other hand, is lean and up to the point, by only	98	follow. AnyEvent, on the other hand, is lean and up to the point, by only
64	offering the functionality that is necessary, in as thin as a wrapper as	99	offering the functionality that is necessary, in as thin as a wrapper as
65	technically possible.	100	technically possible.
66		101
		102	Of course, AnyEvent comes with a big (and fully optional!) toolbox
		103	of useful functionality, such as an asynchronous DNS resolver, 100%
		104	non-blocking connects (even with TLS/SSL, IPv6 and on broken platforms
		105	such as Windows) and lots of real-world knowledge and workarounds for
		106	platform bugs and differences.
		107
67	Of course, if you want lots of policy (this can arguably be somewhat	108	Now, if you I<do want> lots of policy (this can arguably be somewhat
68	useful) and you want to force your users to use the one and only event	109	useful) and you want to force your users to use the one and only event
69	model, you should I<not> use this module.	110	model, you should I<not> use this module.
70
71	#TODO#
72
73	Net::IRC3
74	AnyEvent::HTTPD
75	AnyEvent::DNS
76	IO::AnyEvent
77	Net::FPing
78	Net::XMPP2
79	Coro
80
81	AnyEvent::IRC
82	AnyEvent::HTTPD
83	AnyEvent::DNS
84	AnyEvent::Handle
85	AnyEvent::Socket
86	AnyEvent::FPing
87	AnyEvent::XMPP
88	AnyEvent::SNMP
89	Coro
90		111
91	=head1 DESCRIPTION	112	=head1 DESCRIPTION
92		113
93	L<AnyEvent> provides an identical interface to multiple event loops. This	114	L<AnyEvent> provides an identical interface to multiple event loops. This
94	allows module authors to utilise an event loop without forcing module	115	allows module authors to utilise an event loop without forcing module
…		…
98	The interface itself is vaguely similar, but not identical to the L<Event>	119	The interface itself is vaguely similar, but not identical to the L<Event>
99	module.	120	module.
100		121
101	During the first call of any watcher-creation method, the module tries	122	During the first call of any watcher-creation method, the module tries
102	to detect the currently loaded event loop by probing whether one of the	123	to detect the currently loaded event loop by probing whether one of the
103	following modules is already loaded: L<Coro::EV>, L<Coro::Event>, L<EV>,	124	following modules is already loaded: L<EV>,
104	L<Event>, L<Glib>, L<AnyEvent::Impl::Perl>, L<Tk>, L<Event::Lib>, L<Qt>,	125	L<Event>, L<Glib>, L<AnyEvent::Impl::Perl>, L<Tk>, L<Event::Lib>, L<Qt>,
105	L<POE>. The first one found is used. If none are found, the module tries	126	L<POE>. The first one found is used. If none are found, the module tries
106	to load these modules (excluding Tk, Event::Lib, Qt and POE as the pure perl	127	to load these modules (excluding Tk, Event::Lib, Qt and POE as the pure perl
107	adaptor should always succeed) in the order given. The first one that can	128	adaptor should always succeed) in the order given. The first one that can
108	be successfully loaded will be used. If, after this, still none could be	129	be successfully loaded will be used. If, after this, still none could be
…		…
122	starts using it, all bets are off. Maybe you should tell their authors to	143	starts using it, all bets are off. Maybe you should tell their authors to
123	use AnyEvent so their modules work together with others seamlessly...	144	use AnyEvent so their modules work together with others seamlessly...
124		145
125	The pure-perl implementation of AnyEvent is called	146	The pure-perl implementation of AnyEvent is called
126	C<AnyEvent::Impl::Perl>. Like other event modules you can load it	147	C<AnyEvent::Impl::Perl>. Like other event modules you can load it
127	explicitly.	148	explicitly and enjoy the high availability of that event loop :)
128		149
129	=head1 WATCHERS	150	=head1 WATCHERS
130		151
131	AnyEvent has the central concept of a I<watcher>, which is an object that	152	AnyEvent has the central concept of a I<watcher>, which is an object that
132	stores relevant data for each kind of event you are waiting for, such as	153	stores relevant data for each kind of event you are waiting for, such as
133	the callback to call, the filehandle to watch, etc.	154	the callback to call, the file handle to watch, etc.
134		155
135	These watchers are normal Perl objects with normal Perl lifetime. After	156	These watchers are normal Perl objects with normal Perl lifetime. After
136	creating a watcher it will immediately "watch" for events and invoke the	157	creating a watcher it will immediately "watch" for events and invoke the
137	callback when the event occurs (of course, only when the event model	158	callback when the event occurs (of course, only when the event model
138	is in control).	159	is in control).
139		160
		161	Note that B<callbacks must not permanently change global variables>
		162	potentially in use by the event loop (such as C<$_> or C<$[>) and that B<<
		163	callbacks must not C<die> >>. The former is good programming practise in
		164	Perl and the latter stems from the fact that exception handling differs
		165	widely between event loops.
		166
140	To disable the watcher you have to destroy it (e.g. by setting the	167	To disable the watcher you have to destroy it (e.g. by setting the
141	variable you store it in to C<undef> or otherwise deleting all references	168	variable you store it in to C<undef> or otherwise deleting all references
142	to it).	169	to it).
143		170
144	All watchers are created by calling a method on the C<AnyEvent> class.	171	All watchers are created by calling a method on the C<AnyEvent> class.
…		…
146	Many watchers either are used with "recursion" (repeating timers for	173	Many watchers either are used with "recursion" (repeating timers for
147	example), or need to refer to their watcher object in other ways.	174	example), or need to refer to their watcher object in other ways.
148		175
149	An any way to achieve that is this pattern:	176	An any way to achieve that is this pattern:
150		177
151	my $w; $w = AnyEvent->type (arg => value ..., cb => sub {	178	my $w; $w = AnyEvent->type (arg => value ..., cb => sub {
152	# you can use $w here, for example to undef it	179	# you can use $w here, for example to undef it
153	undef $w;	180	undef $w;
154	});	181	});
155		182
156	Note that C<my $w; $w => combination. This is necessary because in Perl,	183	Note that C<my $w; $w => combination. This is necessary because in Perl,
157	my variables are only visible after the statement in which they are	184	my variables are only visible after the statement in which they are
158	declared.	185	declared.
159		186
160	=head2 I/O WATCHERS	187	=head2 I/O WATCHERS
161		188
		189	$w = AnyEvent->io (
		190	fh => <filehandle_or_fileno>,
		191	poll => <"r" or "w">,
		192	cb => <callback>,
		193	);
		194
162	You can create an I/O watcher by calling the C<< AnyEvent->io >> method	195	You can create an I/O watcher by calling the C<< AnyEvent->io >> method
163	with the following mandatory key-value pairs as arguments:	196	with the following mandatory key-value pairs as arguments:
164		197
165	C<fh> the Perl I<file handle> (I<not> file descriptor) to watch	198	C<fh> is the Perl I<file handle> (or a naked file descriptor) to watch
		199	for events (AnyEvent might or might not keep a reference to this file
		200	handle). Note that only file handles pointing to things for which
		201	non-blocking operation makes sense are allowed. This includes sockets,
		202	most character devices, pipes, fifos and so on, but not for example files
		203	or block devices.
		204
166	for events. C<poll> must be a string that is either C<r> or C<w>,	205	C<poll> must be a string that is either C<r> or C<w>, which creates a
167	which creates a watcher waiting for "r"eadable or "w"ritable events,	206	watcher waiting for "r"eadable or "w"ritable events, respectively.
		207
168	respectively. C<cb> is the callback to invoke each time the file handle	208	C<cb> is the callback to invoke each time the file handle becomes ready.
169	becomes ready.
170		209
171	Although the callback might get passed parameters, their value and	210	Although the callback might get passed parameters, their value and
172	presence is undefined and you cannot rely on them. Portable AnyEvent	211	presence is undefined and you cannot rely on them. Portable AnyEvent
173	callbacks cannot use arguments passed to I/O watcher callbacks.	212	callbacks cannot use arguments passed to I/O watcher callbacks.
174		213
…		…
178		217
179	Some event loops issue spurious readyness notifications, so you should	218	Some event loops issue spurious readyness notifications, so you should
180	always use non-blocking calls when reading/writing from/to your file	219	always use non-blocking calls when reading/writing from/to your file
181	handles.	220	handles.
182		221
183	Example:
184
185	# wait for readability of STDIN, then read a line and disable the watcher	222	Example: wait for readability of STDIN, then read a line and disable the
		223	watcher.
		224
186	my $w; $w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {	225	my $w; $w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {
187	chomp (my $input = <STDIN>);	226	chomp (my $input = <STDIN>);
188	warn "read: $input\n";	227	warn "read: $input\n";
189	undef $w;	228	undef $w;
190	});	229	});
191		230
192	=head2 TIME WATCHERS	231	=head2 TIME WATCHERS
193		232
		233	$w = AnyEvent->timer (after => <seconds>, cb => <callback>);
		234
		235	$w = AnyEvent->timer (
		236	after => <fractional_seconds>,
		237	interval => <fractional_seconds>,
		238	cb => <callback>,
		239	);
		240
194	You can create a time watcher by calling the C<< AnyEvent->timer >>	241	You can create a time watcher by calling the C<< AnyEvent->timer >>
195	method with the following mandatory arguments:	242	method with the following mandatory arguments:
196		243
197	C<after> specifies after how many seconds (fractional values are	244	C<after> specifies after how many seconds (fractional values are
198	supported) the callback should be invoked. C<cb> is the callback to invoke	245	supported) the callback should be invoked. C<cb> is the callback to invoke
…		…
200		247
201	Although the callback might get passed parameters, their value and	248	Although the callback might get passed parameters, their value and
202	presence is undefined and you cannot rely on them. Portable AnyEvent	249	presence is undefined and you cannot rely on them. Portable AnyEvent
203	callbacks cannot use arguments passed to time watcher callbacks.	250	callbacks cannot use arguments passed to time watcher callbacks.
204		251
205	The timer callback will be invoked at most once: if you want a repeating	252	The callback will normally be invoked once only. If you specify another
206	timer you have to create a new watcher (this is a limitation by both Tk	253	parameter, C<interval>, as a strictly positive number (> 0), then the
207	and Glib).	254	callback will be invoked regularly at that interval (in fractional
		255	seconds) after the first invocation. If C<interval> is specified with a
		256	false value, then it is treated as if it were missing.
208		257
209	Example:	258	The callback will be rescheduled before invoking the callback, but no
		259	attempt is done to avoid timer drift in most backends, so the interval is
		260	only approximate.
210		261
211	# fire an event after 7.7 seconds	262	Example: fire an event after 7.7 seconds.
		263
212	my $w = AnyEvent->timer (after => 7.7, cb => sub {	264	my $w = AnyEvent->timer (after => 7.7, cb => sub {
213	warn "timeout\n";	265	warn "timeout\n";
214	});	266	});
215		267
216	# to cancel the timer:	268	# to cancel the timer:
217	undef $w;	269	undef $w;
218		270
219	Example 2:
220
221	# fire an event after 0.5 seconds, then roughly every second	271	Example 2: fire an event after 0.5 seconds, then roughly every second.
222	my $w;
223		272
224	my $cb = sub {
225	# cancel the old timer while creating a new one
226	$w = AnyEvent->timer (after => 1, cb => $cb);	273	my $w = AnyEvent->timer (after => 0.5, interval => 1, cb => sub {
		274	warn "timeout\n";
227	};	275	};
228
229	# start the "loop" by creating the first watcher
230	$w = AnyEvent->timer (after => 0.5, cb => $cb);
231		276
232	=head3 TIMING ISSUES	277	=head3 TIMING ISSUES
233		278
234	There are two ways to handle timers: based on real time (relative, "fire	279	There are two ways to handle timers: based on real time (relative, "fire
235	in 10 seconds") and based on wallclock time (absolute, "fire at 12	280	in 10 seconds") and based on wallclock time (absolute, "fire at 12
…		…
247	timers.	292	timers.
248		293
249	AnyEvent always prefers relative timers, if available, matching the	294	AnyEvent always prefers relative timers, if available, matching the
250	AnyEvent API.	295	AnyEvent API.
251		296
		297	AnyEvent has two additional methods that return the "current time":
		298
		299	=over 4
		300
		301	=item AnyEvent->time
		302
		303	This returns the "current wallclock time" as a fractional number of
		304	seconds since the Epoch (the same thing as C<time> or C<Time::HiRes::time>
		305	return, and the result is guaranteed to be compatible with those).
		306
		307	It progresses independently of any event loop processing, i.e. each call
		308	will check the system clock, which usually gets updated frequently.
		309
		310	=item AnyEvent->now
		311
		312	This also returns the "current wallclock time", but unlike C<time>, above,
		313	this value might change only once per event loop iteration, depending on
		314	the event loop (most return the same time as C<time>, above). This is the
		315	time that AnyEvent's timers get scheduled against.
		316
		317	I<In almost all cases (in all cases if you don't care), this is the
		318	function to call when you want to know the current time.>
		319
		320	This function is also often faster then C<< AnyEvent->time >>, and
		321	thus the preferred method if you want some timestamp (for example,
		322	L<AnyEvent::Handle> uses this to update it's activity timeouts).
		323
		324	The rest of this section is only of relevance if you try to be very exact
		325	with your timing, you can skip it without bad conscience.
		326
		327	For a practical example of when these times differ, consider L<Event::Lib>
		328	and L<EV> and the following set-up:
		329
		330	The event loop is running and has just invoked one of your callback at
		331	time=500 (assume no other callbacks delay processing). In your callback,
		332	you wait a second by executing C<sleep 1> (blocking the process for a
		333	second) and then (at time=501) you create a relative timer that fires
		334	after three seconds.
		335
		336	With L<Event::Lib>, C<< AnyEvent->time >> and C<< AnyEvent->now >> will
		337	both return C<501>, because that is the current time, and the timer will
		338	be scheduled to fire at time=504 (C<501> + C<3>).
		339
		340	With L<EV>, C<< AnyEvent->time >> returns C<501> (as that is the current
		341	time), but C<< AnyEvent->now >> returns C<500>, as that is the time the
		342	last event processing phase started. With L<EV>, your timer gets scheduled
		343	to run at time=503 (C<500> + C<3>).
		344
		345	In one sense, L<Event::Lib> is more exact, as it uses the current time
		346	regardless of any delays introduced by event processing. However, most
		347	callbacks do not expect large delays in processing, so this causes a
		348	higher drift (and a lot more system calls to get the current time).
		349
		350	In another sense, L<EV> is more exact, as your timer will be scheduled at
		351	the same time, regardless of how long event processing actually took.
		352
		353	In either case, if you care (and in most cases, you don't), then you
		354	can get whatever behaviour you want with any event loop, by taking the
		355	difference between C<< AnyEvent->time >> and C<< AnyEvent->now >> into
		356	account.
		357
		358	=item AnyEvent->now_update
		359
		360	Some event loops (such as L<EV> or L<AnyEvent::Impl::Perl>) cache
		361	the current time for each loop iteration (see the discussion of L<<
		362	AnyEvent->now >>, above).
		363
		364	When a callback runs for a long time (or when the process sleeps), then
		365	this "current" time will differ substantially from the real time, which
		366	might affect timers and time-outs.
		367
		368	When this is the case, you can call this method, which will update the
		369	event loop's idea of "current time".
		370
		371	A typical example would be a script in a web server (e.g. C<mod_perl>) -
		372	when mod_perl executes the script, then the event loop will have the wrong
		373	idea about the "current time" (being potentially far in the past, when the
		374	script ran the last time). In that case you should arrange a call to C<<
		375	AnyEvent->now_update >> each time the web server process wakes up again
		376	(e.g. at the start of your script, or in a handler).
		377
		378	Note that updating the time I<might> cause some events to be handled.
		379
		380	=back
		381
252	=head2 SIGNAL WATCHERS	382	=head2 SIGNAL WATCHERS
253		383
		384	$w = AnyEvent->signal (signal => <uppercase_signal_name>, cb => <callback>);
		385
254	You can watch for signals using a signal watcher, C<signal> is the signal	386	You can watch for signals using a signal watcher, C<signal> is the signal
255	I<name> without any C<SIG> prefix, C<cb> is the Perl callback to	387	I<name> in uppercase and without any C<SIG> prefix, C<cb> is the Perl
256	be invoked whenever a signal occurs.	388	callback to be invoked whenever a signal occurs.
257		389
258	Although the callback might get passed parameters, their value and	390	Although the callback might get passed parameters, their value and
259	presence is undefined and you cannot rely on them. Portable AnyEvent	391	presence is undefined and you cannot rely on them. Portable AnyEvent
260	callbacks cannot use arguments passed to signal watcher callbacks.	392	callbacks cannot use arguments passed to signal watcher callbacks.
261		393
262	Multiple signal occurances can be clumped together into one callback	394	Multiple signal occurrences can be clumped together into one callback
263	invocation, and callback invocation will be synchronous. synchronous means	395	invocation, and callback invocation will be synchronous. Synchronous means
264	that it might take a while until the signal gets handled by the process,	396	that it might take a while until the signal gets handled by the process,
265	but it is guarenteed not to interrupt any other callbacks.	397	but it is guaranteed not to interrupt any other callbacks.
266		398
267	The main advantage of using these watchers is that you can share a signal	399	The main advantage of using these watchers is that you can share a signal
268	between multiple watchers.	400	between multiple watchers, and AnyEvent will ensure that signals will not
		401	interrupt your program at bad times.
269		402
270	This watcher might use C<%SIG>, so programs overwriting those signals	403	This watcher might use C<%SIG> (depending on the event loop used),
271	directly will likely not work correctly.	404	so programs overwriting those signals directly will likely not work
		405	correctly.
272		406
273	Example: exit on SIGINT	407	Example: exit on SIGINT
274		408
275	my $w = AnyEvent->signal (signal => "INT", cb => sub { exit 1 });	409	my $w = AnyEvent->signal (signal => "INT", cb => sub { exit 1 });
276		410
		411	=head3 Restart Behaviour
		412
		413	While restart behaviour is up to the event loop implementation, most will
		414	not restart syscalls (that includes L<Async::Interrupt> and AnyEvent's
		415	pure perl implementation).
		416
		417	=head3 Safe/Unsafe Signals
		418
		419	Perl signals can be either "safe" (synchronous to opcode handling) or
		420	"unsafe" (asynchronous) - the former might get delayed indefinitely, the
		421	latter might corrupt your memory.
		422
		423	AnyEvent signal handlers are, in addition, synchronous to the event loop,
		424	i.e. they will not interrupt your running perl program but will only be
		425	called as part of the normal event handling (just like timer, I/O etc.
		426	callbacks, too).
		427
		428	=head3 Signal Races, Delays and Workarounds
		429
		430	Many event loops (e.g. Glib, Tk, Qt, IO::Async) do not support attaching
		431	callbacks to signals in a generic way, which is a pity, as you cannot
		432	do race-free signal handling in perl, requiring C libraries for
		433	this. AnyEvent will try to do it's best, which means in some cases,
		434	signals will be delayed. The maximum time a signal might be delayed is
		435	specified in C<$AnyEvent::MAX_SIGNAL_LATENCY> (default: 10 seconds). This
		436	variable can be changed only before the first signal watcher is created,
		437	and should be left alone otherwise. This variable determines how often
		438	AnyEvent polls for signals (in case a wake-up was missed). Higher values
		439	will cause fewer spurious wake-ups, which is better for power and CPU
		440	saving.
		441
		442	All these problems can be avoided by installing the optional
		443	L<Async::Interrupt> module, which works with most event loops. It will not
		444	work with inherently broken event loops such as L<Event> or L<Event::Lib>
		445	(and not with L<POE> currently, as POE does it's own workaround with
		446	one-second latency). For those, you just have to suffer the delays.
		447
277	=head2 CHILD PROCESS WATCHERS	448	=head2 CHILD PROCESS WATCHERS
278		449
		450	$w = AnyEvent->child (pid => <process id>, cb => <callback>);
		451
279	You can also watch on a child process exit and catch its exit status.	452	You can also watch on a child process exit and catch its exit status.
280		453
281	The child process is specified by the C<pid> argument (if set to C<0>, it	454	The child process is specified by the C<pid> argument (one some backends,
282	watches for any child process exit). The watcher will trigger as often	455	using C<0> watches for any child process exit, on others this will
283	as status change for the child are received. This works by installing a	456	croak). The watcher will be triggered only when the child process has
284	signal handler for C<SIGCHLD>. The callback will be called with the pid	457	finished and an exit status is available, not on any trace events
285	and exit status (as returned by waitpid), so unlike other watcher types,	458	(stopped/continued).
286	you I<can> rely on child watcher callback arguments.	459
		460	The callback will be called with the pid and exit status (as returned by
		461	waitpid), so unlike other watcher types, you I<can> rely on child watcher
		462	callback arguments.
		463
		464	This watcher type works by installing a signal handler for C<SIGCHLD>,
		465	and since it cannot be shared, nothing else should use SIGCHLD or reap
		466	random child processes (waiting for specific child processes, e.g. inside
		467	C<system>, is just fine).
287		468
288	There is a slight catch to child watchers, however: you usually start them	469	There is a slight catch to child watchers, however: you usually start them
289	I<after> the child process was created, and this means the process could	470	I<after> the child process was created, and this means the process could
290	have exited already (and no SIGCHLD will be sent anymore).	471	have exited already (and no SIGCHLD will be sent anymore).
291		472
292	Not all event models handle this correctly (POE doesn't), but even for	473	Not all event models handle this correctly (neither POE nor IO::Async do,
		474	see their AnyEvent::Impl manpages for details), but even for event models
293	event models that I<do> handle this correctly, they usually need to be	475	that I<do> handle this correctly, they usually need to be loaded before
294	loaded before the process exits (i.e. before you fork in the first place).	476	the process exits (i.e. before you fork in the first place). AnyEvent's
		477	pure perl event loop handles all cases correctly regardless of when you
		478	start the watcher.
295		479
296	This means you cannot create a child watcher as the very first thing in an	480	This means you cannot create a child watcher as the very first
297	AnyEvent program, you I<have> to create at least one watcher before you	481	thing in an AnyEvent program, you I<have> to create at least one
298	C<fork> the child (alternatively, you can call C<AnyEvent::detect>).	482	watcher before you C<fork> the child (alternatively, you can call
		483	C<AnyEvent::detect>).
		484
		485	As most event loops do not support waiting for child events, they will be
		486	emulated by AnyEvent in most cases, in which the latency and race problems
		487	mentioned in the description of signal watchers apply.
299		488
300	Example: fork a process and wait for it	489	Example: fork a process and wait for it
301		490
302	my $done = AnyEvent->condvar;	491	my $done = AnyEvent->condvar;
303		492
304	AnyEvent::detect; # force event module to be initialised
305
306	my $pid = fork or exit 5;	493	my $pid = fork or exit 5;
307		494
308	my $w = AnyEvent->child (	495	my $w = AnyEvent->child (
309	pid => $pid,	496	pid => $pid,
310	cb => sub {	497	cb => sub {
311	my ($pid, $status) = @_;	498	my ($pid, $status) = @_;
312	warn "pid $pid exited with status $status";	499	warn "pid $pid exited with status $status";
313	$done->broadcast;	500	$done->send;
314	},	501	},
315	);	502	);
316		503
317	# do something else, then wait for process exit	504	# do something else, then wait for process exit
318	$done->wait;	505	$done->recv;
		506
		507	=head2 IDLE WATCHERS
		508
		509	$w = AnyEvent->idle (cb => <callback>);
		510
		511	Repeatedly invoke the callback after the process becomes idle, until
		512	either the watcher is destroyed or new events have been detected.
		513
		514	Idle watchers are useful when there is a need to do something, but it
		515	is not so important (or wise) to do it instantly. The callback will be
		516	invoked only when there is "nothing better to do", which is usually
		517	defined as "all outstanding events have been handled and no new events
		518	have been detected". That means that idle watchers ideally get invoked
		519	when the event loop has just polled for new events but none have been
		520	detected. Instead of blocking to wait for more events, the idle watchers
		521	will be invoked.
		522
		523	Unfortunately, most event loops do not really support idle watchers (only
		524	EV, Event and Glib do it in a usable fashion) - for the rest, AnyEvent
		525	will simply call the callback "from time to time".
		526
		527	Example: read lines from STDIN, but only process them when the
		528	program is otherwise idle:
		529
		530	my @lines; # read data
		531	my $idle_w;
		532	my $io_w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {
		533	push @lines, scalar <STDIN>;
		534
		535	# start an idle watcher, if not already done
		536	$idle_w ||= AnyEvent->idle (cb => sub {
		537	# handle only one line, when there are lines left
		538	if (my $line = shift @lines) {
		539	print "handled when idle: $line";
		540	} else {
		541	# otherwise disable the idle watcher again
		542	undef $idle_w;
		543	}
		544	});
		545	});
319		546
320	=head2 CONDITION VARIABLES	547	=head2 CONDITION VARIABLES
321		548
		549	$cv = AnyEvent->condvar;
		550
		551	$cv->send (<list>);
		552	my @res = $cv->recv;
		553
		554	If you are familiar with some event loops you will know that all of them
		555	require you to run some blocking "loop", "run" or similar function that
		556	will actively watch for new events and call your callbacks.
		557
		558	AnyEvent is slightly different: it expects somebody else to run the event
		559	loop and will only block when necessary (usually when told by the user).
		560
		561	The tool to do that is called a "condition variable", so called because
		562	they represent a condition that must become true.
		563
		564	Now is probably a good time to look at the examples further below.
		565
322	Condition variables can be created by calling the C<< AnyEvent->condvar >>	566	Condition variables can be created by calling the C<< AnyEvent->condvar
323	method without any arguments.	567	>> method, usually without arguments. The only argument pair allowed is
		568	C<cb>, which specifies a callback to be called when the condition variable
		569	becomes true, with the condition variable as the first argument (but not
		570	the results).
324		571
325	A condition variable waits for a condition - precisely that the C<<	572	After creation, the condition variable is "false" until it becomes "true"
326	->broadcast >> method has been called.	573	by calling the C<send> method (or calling the condition variable as if it
		574	were a callback, read about the caveats in the description for the C<<
		575	->send >> method).
327		576
328	They are very useful to signal that a condition has been fulfilled, for	577	Since condition variables are the most complex part of the AnyEvent API, here are
		578	some different mental models of what they are - pick the ones you can connect to:
		579
		580	=over 4
		581
		582	=item * Condition variables are like callbacks - you can call them (and pass them instead
		583	of callbacks). Unlike callbacks however, you can also wait for them to be called.
		584
		585	=item * Condition variables are signals - one side can emit or send them,
		586	the other side can wait for them, or install a handler that is called when
		587	the signal fires.
		588
		589	=item * Condition variables are like "Merge Points" - points in your program
		590	where you merge multiple independent results/control flows into one.
		591
		592	=item * Condition variables represent a transaction - function that start
		593	some kind of transaction can return them, leaving the caller the choice
		594	between waiting in a blocking fashion, or setting a callback.
		595
		596	=item * Condition variables represent future values, or promises to deliver
		597	some result, long before the result is available.
		598
		599	=back
		600
		601	Condition variables are very useful to signal that something has finished,
329	example, if you write a module that does asynchronous http requests,	602	for example, if you write a module that does asynchronous http requests,
330	then a condition variable would be the ideal candidate to signal the	603	then a condition variable would be the ideal candidate to signal the
331	availability of results.	604	availability of results. The user can either act when the callback is
		605	called or can synchronously C<< ->recv >> for the results.
332		606
333	You can also use condition variables to block your main program until	607	You can also use them to simulate traditional event loops - for example,
334	an event occurs - for example, you could C<< ->wait >> in your main	608	you can block your main program until an event occurs - for example, you
335	program until the user clicks the Quit button in your app, which would C<<	609	could C<< ->recv >> in your main program until the user clicks the Quit
336	->broadcast >> the "quit" event.	610	button of your app, which would C<< ->send >> the "quit" event.
337		611
338	Note that condition variables recurse into the event loop - if you have	612	Note that condition variables recurse into the event loop - if you have
339	two pirces of code that call C<< ->wait >> in a round-robbin fashion, you	613	two pieces of code that call C<< ->recv >> in a round-robin fashion, you
340	lose. Therefore, condition variables are good to export to your caller, but	614	lose. Therefore, condition variables are good to export to your caller, but
341	you should avoid making a blocking wait yourself, at least in callbacks,	615	you should avoid making a blocking wait yourself, at least in callbacks,
342	as this asks for trouble.	616	as this asks for trouble.
343		617
344	This object has two methods:	618	Condition variables are represented by hash refs in perl, and the keys
		619	used by AnyEvent itself are all named C<_ae_XXX> to make subclassing
		620	easy (it is often useful to build your own transaction class on top of
		621	AnyEvent). To subclass, use C<AnyEvent::CondVar> as base class and call
		622	it's C<new> method in your own C<new> method.
		623
		624	There are two "sides" to a condition variable - the "producer side" which
		625	eventually calls C<< -> send >>, and the "consumer side", which waits
		626	for the send to occur.
		627
		628	Example: wait for a timer.
		629
		630	# condition: "wait till the timer is fired"
		631	my $timer_fired = AnyEvent->condvar;
		632
		633	# create the timer - we could wait for, say
		634	# a handle becomign ready, or even an
		635	# AnyEvent::HTTP request to finish, but
		636	# in this case, we simply use a timer:
		637	my $w = AnyEvent->timer (
		638	after => 1,
		639	cb => sub { $timer_fired->send },
		640	);
		641
		642	# this "blocks" (while handling events) till the callback
		643	# calls ->send
		644	$timer_fired->recv;
		645
		646	Example: wait for a timer, but take advantage of the fact that condition
		647	variables are also callable directly.
		648
		649	my $done = AnyEvent->condvar;
		650	my $delay = AnyEvent->timer (after => 5, cb => $done);
		651	$done->recv;
		652
		653	Example: Imagine an API that returns a condvar and doesn't support
		654	callbacks. This is how you make a synchronous call, for example from
		655	the main program:
		656
		657	use AnyEvent::CouchDB;
		658
		659	...
		660
		661	my @info = $couchdb->info->recv;
		662
		663	And this is how you would just set a callback to be called whenever the
		664	results are available:
		665
		666	$couchdb->info->cb (sub {
		667	my @info = $_[0]->recv;
		668	});
		669
		670	=head3 METHODS FOR PRODUCERS
		671
		672	These methods should only be used by the producing side, i.e. the
		673	code/module that eventually sends the signal. Note that it is also
		674	the producer side which creates the condvar in most cases, but it isn't
		675	uncommon for the consumer to create it as well.
345		676
346	=over 4	677	=over 4
347		678
		679	=item $cv->send (...)
		680
		681	Flag the condition as ready - a running C<< ->recv >> and all further
		682	calls to C<recv> will (eventually) return after this method has been
		683	called. If nobody is waiting the send will be remembered.
		684
		685	If a callback has been set on the condition variable, it is called
		686	immediately from within send.
		687
		688	Any arguments passed to the C<send> call will be returned by all
		689	future C<< ->recv >> calls.
		690
		691	Condition variables are overloaded so one can call them directly (as if
		692	they were a code reference). Calling them directly is the same as calling
		693	C<send>.
		694
		695	=item $cv->croak ($error)
		696
		697	Similar to send, but causes all call's to C<< ->recv >> to invoke
		698	C<Carp::croak> with the given error message/object/scalar.
		699
		700	This can be used to signal any errors to the condition variable
		701	user/consumer. Doing it this way instead of calling C<croak> directly
		702	delays the error detetcion, but has the overwhelmign advantage that it
		703	diagnoses the error at the place where the result is expected, and not
		704	deep in some event clalback without connection to the actual code causing
		705	the problem.
		706
		707	=item $cv->begin ([group callback])
		708
348	=item $cv->wait	709	=item $cv->end
349		710
350	Wait (blocking if necessary) until the C<< ->broadcast >> method has been	711	These two methods can be used to combine many transactions/events into
		712	one. For example, a function that pings many hosts in parallel might want
		713	to use a condition variable for the whole process.
		714
		715	Every call to C<< ->begin >> will increment a counter, and every call to
		716	C<< ->end >> will decrement it. If the counter reaches C<0> in C<< ->end
		717	>>, the (last) callback passed to C<begin> will be executed, passing the
		718	condvar as first argument. That callback is I<supposed> to call C<< ->send
		719	>>, but that is not required. If no group callback was set, C<send> will
		720	be called without any arguments.
		721
		722	You can think of C<< $cv->send >> giving you an OR condition (one call
		723	sends), while C<< $cv->begin >> and C<< $cv->end >> giving you an AND
		724	condition (all C<begin> calls must be C<end>'ed before the condvar sends).
		725
		726	Let's start with a simple example: you have two I/O watchers (for example,
		727	STDOUT and STDERR for a program), and you want to wait for both streams to
		728	close before activating a condvar:
		729
		730	my $cv = AnyEvent->condvar;
		731
		732	$cv->begin; # first watcher
		733	my $w1 = AnyEvent->io (fh => $fh1, cb => sub {
		734	defined sysread $fh1, my $buf, 4096
		735	or $cv->end;
		736	});
		737
		738	$cv->begin; # second watcher
		739	my $w2 = AnyEvent->io (fh => $fh2, cb => sub {
		740	defined sysread $fh2, my $buf, 4096
		741	or $cv->end;
		742	});
		743
		744	$cv->recv;
		745
		746	This works because for every event source (EOF on file handle), there is
		747	one call to C<begin>, so the condvar waits for all calls to C<end> before
		748	sending.
		749
		750	The ping example mentioned above is slightly more complicated, as the
		751	there are results to be passwd back, and the number of tasks that are
		752	begung can potentially be zero:
		753
		754	my $cv = AnyEvent->condvar;
		755
		756	my %result;
		757	$cv->begin (sub { shift->send (\%result) });
		758
		759	for my $host (@list_of_hosts) {
		760	$cv->begin;
		761	ping_host_then_call_callback $host, sub {
		762	$result{$host} = ...;
		763	$cv->end;
		764	};
		765	}
		766
		767	$cv->end;
		768
		769	This code fragment supposedly pings a number of hosts and calls
		770	C<send> after results for all then have have been gathered - in any
		771	order. To achieve this, the code issues a call to C<begin> when it starts
		772	each ping request and calls C<end> when it has received some result for
		773	it. Since C<begin> and C<end> only maintain a counter, the order in which
		774	results arrive is not relevant.
		775
		776	There is an additional bracketing call to C<begin> and C<end> outside the
		777	loop, which serves two important purposes: first, it sets the callback
		778	to be called once the counter reaches C<0>, and second, it ensures that
		779	C<send> is called even when C<no> hosts are being pinged (the loop
		780	doesn't execute once).
		781
		782	This is the general pattern when you "fan out" into multiple (but
		783	potentially none) subrequests: use an outer C<begin>/C<end> pair to set
		784	the callback and ensure C<end> is called at least once, and then, for each
		785	subrequest you start, call C<begin> and for each subrequest you finish,
		786	call C<end>.
		787
		788	=back
		789
		790	=head3 METHODS FOR CONSUMERS
		791
		792	These methods should only be used by the consuming side, i.e. the
		793	code awaits the condition.
		794
		795	=over 4
		796
		797	=item $cv->recv
		798
		799	Wait (blocking if necessary) until the C<< ->send >> or C<< ->croak
351	called on c<$cv>, while servicing other watchers normally.	800	>> methods have been called on c<$cv>, while servicing other watchers
		801	normally.
352		802
353	You can only wait once on a condition - additional calls will return	803	You can only wait once on a condition - additional calls are valid but
354	immediately.	804	will return immediately.
		805
		806	If an error condition has been set by calling C<< ->croak >>, then this
		807	function will call C<croak>.
		808
		809	In list context, all parameters passed to C<send> will be returned,
		810	in scalar context only the first one will be returned.
		811
		812	Note that doing a blocking wait in a callback is not supported by any
		813	event loop, that is, recursive invocation of a blocking C<< ->recv
		814	>> is not allowed, and the C<recv> call will C<croak> if such a
		815	condition is detected. This condition can be slightly loosened by using
		816	L<Coro::AnyEvent>, which allows you to do a blocking C<< ->recv >> from
		817	any thread that doesn't run the event loop itself.
355		818
356	Not all event models support a blocking wait - some die in that case	819	Not all event models support a blocking wait - some die in that case
357	(programs might want to do that to stay interactive), so I<if you are	820	(programs might want to do that to stay interactive), so I<if you are
358	using this from a module, never require a blocking wait>, but let the	821	using this from a module, never require a blocking wait>. Instead, let the
359	caller decide whether the call will block or not (for example, by coupling	822	caller decide whether the call will block or not (for example, by coupling
360	condition variables with some kind of request results and supporting	823	condition variables with some kind of request results and supporting
361	callbacks so the caller knows that getting the result will not block,	824	callbacks so the caller knows that getting the result will not block,
362	while still suppporting blocking waits if the caller so desires).	825	while still supporting blocking waits if the caller so desires).
363		826
364	Another reason I<never> to C<< ->wait >> in a module is that you cannot	827	You can ensure that C<< -recv >> never blocks by setting a callback and
365	sensibly have two C<< ->wait >>'s in parallel, as that would require	828	only calling C<< ->recv >> from within that callback (or at a later
366	multiple interpreters or coroutines/threads, none of which C<AnyEvent>	829	time). This will work even when the event loop does not support blocking
367	can supply (the coroutine-aware backends L<AnyEvent::Impl::CoroEV> and	830	waits otherwise.
368	L<AnyEvent::Impl::CoroEvent> explicitly support concurrent C<< ->wait >>'s
369	from different coroutines, however).
370		831
371	=item $cv->broadcast	832	=item $bool = $cv->ready
372		833
373	Flag the condition as ready - a running C<< ->wait >> and all further	834	Returns true when the condition is "true", i.e. whether C<send> or
374	calls to C<wait> will (eventually) return after this method has been	835	C<croak> have been called.
375	called. If nobody is waiting the broadcast will be remembered..	836
		837	=item $cb = $cv->cb ($cb->($cv))
		838
		839	This is a mutator function that returns the callback set and optionally
		840	replaces it before doing so.
		841
		842	The callback will be called when the condition becomes (or already was)
		843	"true", i.e. when C<send> or C<croak> are called (or were called), with
		844	the only argument being the condition variable itself. Calling C<recv>
		845	inside the callback or at any later time is guaranteed not to block.
376		846
377	=back	847	=back
378		848
379	Example:	849	=head1 SUPPORTED EVENT LOOPS/BACKENDS
380		850
381	# wait till the result is ready	851	The available backend classes are (every class has its own manpage):
382	my $result_ready = AnyEvent->condvar;
383		852
384	# do something such as adding a timer	853	=over 4
385	# or socket watcher the calls $result_ready->broadcast
386	# when the "result" is ready.
387	# in this case, we simply use a timer:
388	my $w = AnyEvent->timer (
389	after => 1,
390	cb => sub { $result_ready->broadcast },
391	);
392		854
393	# this "blocks" (while handling events) till the watcher	855	=item Backends that are autoprobed when no other event loop can be found.
394	# calls broadcast	856
395	$result_ready->wait;	857	EV is the preferred backend when no other event loop seems to be in
		858	use. If EV is not installed, then AnyEvent will fall back to its own
		859	pure-perl implementation, which is available everywhere as it comes with
		860	AnyEvent itself.
		861
		862	AnyEvent::Impl::EV based on EV (interface to libev, best choice).
		863	AnyEvent::Impl::Perl pure-perl implementation, fast and portable.
		864
		865	=item Backends that are transparently being picked up when they are used.
		866
		867	These will be used when they are currently loaded when the first watcher
		868	is created, in which case it is assumed that the application is using
		869	them. This means that AnyEvent will automatically pick the right backend
		870	when the main program loads an event module before anything starts to
		871	create watchers. Nothing special needs to be done by the main program.
		872
		873	AnyEvent::Impl::Event based on Event, very stable, few glitches.
		874	AnyEvent::Impl::Glib based on Glib, slow but very stable.
		875	AnyEvent::Impl::Tk based on Tk, very broken.
		876	AnyEvent::Impl::EventLib based on Event::Lib, leaks memory and worse.
		877	AnyEvent::Impl::POE based on POE, very slow, some limitations.
		878	AnyEvent::Impl::Irssi used when running within irssi.
		879
		880	=item Backends with special needs.
		881
		882	Qt requires the Qt::Application to be instantiated first, but will
		883	otherwise be picked up automatically. As long as the main program
		884	instantiates the application before any AnyEvent watchers are created,
		885	everything should just work.
		886
		887	AnyEvent::Impl::Qt based on Qt.
		888
		889	Support for IO::Async can only be partial, as it is too broken and
		890	architecturally limited to even support the AnyEvent API. It also
		891	is the only event loop that needs the loop to be set explicitly, so
		892	it can only be used by a main program knowing about AnyEvent. See
		893	L<AnyEvent::Impl::Async> for the gory details.
		894
		895	AnyEvent::Impl::IOAsync based on IO::Async, cannot be autoprobed.
		896
		897	=item Event loops that are indirectly supported via other backends.
		898
		899	Some event loops can be supported via other modules:
		900
		901	There is no direct support for WxWidgets (L<Wx>) or L<Prima>.
		902
		903	B<WxWidgets> has no support for watching file handles. However, you can
		904	use WxWidgets through the POE adaptor, as POE has a Wx backend that simply
		905	polls 20 times per second, which was considered to be too horrible to even
		906	consider for AnyEvent.
		907
		908	B<Prima> is not supported as nobody seems to be using it, but it has a POE
		909	backend, so it can be supported through POE.
		910
		911	AnyEvent knows about both L<Prima> and L<Wx>, however, and will try to
		912	load L<POE> when detecting them, in the hope that POE will pick them up,
		913	in which case everything will be automatic.
		914
		915	=back
396		916
397	=head1 GLOBAL VARIABLES AND FUNCTIONS	917	=head1 GLOBAL VARIABLES AND FUNCTIONS
398		918
		919	These are not normally required to use AnyEvent, but can be useful to
		920	write AnyEvent extension modules.
		921
399	=over 4	922	=over 4
400		923
401	=item $AnyEvent::MODEL	924	=item $AnyEvent::MODEL
402		925
403	Contains C<undef> until the first watcher is being created. Then it	926	Contains C<undef> until the first watcher is being created, before the
		927	backend has been autodetected.
		928
404	contains the event model that is being used, which is the name of the	929	Afterwards it contains the event model that is being used, which is the
405	Perl class implementing the model. This class is usually one of the	930	name of the Perl class implementing the model. This class is usually one
406	C<AnyEvent::Impl:xxx> modules, but can be any other class in the case	931	of the C<AnyEvent::Impl:xxx> modules, but can be any other class in the
407	AnyEvent has been extended at runtime (e.g. in I<rxvt-unicode>).	932	case AnyEvent has been extended at runtime (e.g. in I<rxvt-unicode> it
408		933	will be C<urxvt::anyevent>).
409	The known classes so far are:
410
411	AnyEvent::Impl::CoroEV based on Coro::EV, best choice.
412	AnyEvent::Impl::CoroEvent based on Coro::Event, second best choice.
413	AnyEvent::Impl::EV based on EV (an interface to libev, best choice).
414	AnyEvent::Impl::Event based on Event, second best choice.
415	AnyEvent::Impl::Glib based on Glib, third-best choice.
416	AnyEvent::Impl::Perl pure-perl implementation, inefficient but portable.
417	AnyEvent::Impl::Tk based on Tk, very bad choice.
418	AnyEvent::Impl::Qt based on Qt, cannot be autoprobed (see its docs).
419	AnyEvent::Impl::EventLib based on Event::Lib, leaks memory and worse.
420	AnyEvent::Impl::POE based on POE, not generic enough for full support.
421
422	There is no support for WxWidgets, as WxWidgets has no support for
423	watching file handles. However, you can use WxWidgets through the
424	POE Adaptor, as POE has a Wx backend that simply polls 20 times per
425	second, which was considered to be too horrible to even consider for
426	AnyEvent. Likewise, other POE backends can be used by AnyEvent by using
427	it's adaptor.
428
429	AnyEvent knows about L<Prima> and L<Wx> and will try to use L<POE> when
430	autodetecting them.
431		934
432	=item AnyEvent::detect	935	=item AnyEvent::detect
433		936
434	Returns C<$AnyEvent::MODEL>, forcing autodetection of the event model	937	Returns C<$AnyEvent::MODEL>, forcing autodetection of the event model
435	if necessary. You should only call this function right before you would	938	if necessary. You should only call this function right before you would
436	have created an AnyEvent watcher anyway, that is, as late as possible at	939	have created an AnyEvent watcher anyway, that is, as late as possible at
437	runtime.	940	runtime, and not e.g. while initialising of your module.
		941
		942	If you need to do some initialisation before AnyEvent watchers are
		943	created, use C<post_detect>.
		944
		945	=item $guard = AnyEvent::post_detect { BLOCK }
		946
		947	Arranges for the code block to be executed as soon as the event model is
		948	autodetected (or immediately if this has already happened).
		949
		950	The block will be executed I<after> the actual backend has been detected
		951	(C<$AnyEvent::MODEL> is set), but I<before> any watchers have been
		952	created, so it is possible to e.g. patch C<@AnyEvent::ISA> or do
		953	other initialisations - see the sources of L<AnyEvent::Strict> or
		954	L<AnyEvent::AIO> to see how this is used.
		955
		956	The most common usage is to create some global watchers, without forcing
		957	event module detection too early, for example, L<AnyEvent::AIO> creates
		958	and installs the global L<IO::AIO> watcher in a C<post_detect> block to
		959	avoid autodetecting the event module at load time.
		960
		961	If called in scalar or list context, then it creates and returns an object
		962	that automatically removes the callback again when it is destroyed (or
		963	C<undef> when the hook was immediately executed). See L<AnyEvent::AIO> for
		964	a case where this is useful.
		965
		966	Example: Create a watcher for the IO::AIO module and store it in
		967	C<$WATCHER>. Only do so after the event loop is initialised, though.
		968
		969	our WATCHER;
		970
		971	my $guard = AnyEvent::post_detect {
		972	$WATCHER = AnyEvent->io (fh => IO::AIO::poll_fileno, poll => 'r', cb => \&IO::AIO::poll_cb);
		973	};
		974
		975	# the ||= is important in case post_detect immediately runs the block,
		976	# as to not clobber the newly-created watcher. assigning both watcher and
		977	# post_detect guard to the same variable has the advantage of users being
		978	# able to just C<undef $WATCHER> if the watcher causes them grief.
		979
		980	$WATCHER ||= $guard;
		981
		982	=item @AnyEvent::post_detect
		983
		984	If there are any code references in this array (you can C<push> to it
		985	before or after loading AnyEvent), then they will called directly after
		986	the event loop has been chosen.
		987
		988	You should check C<$AnyEvent::MODEL> before adding to this array, though:
		989	if it is defined then the event loop has already been detected, and the
		990	array will be ignored.
		991
		992	Best use C<AnyEvent::post_detect { BLOCK }> when your application allows
		993	it, as it takes care of these details.
		994
		995	This variable is mainly useful for modules that can do something useful
		996	when AnyEvent is used and thus want to know when it is initialised, but do
		997	not need to even load it by default. This array provides the means to hook
		998	into AnyEvent passively, without loading it.
		999
		1000	Example: To load Coro::AnyEvent whenever Coro and AnyEvent are used
		1001	together, you could put this into Coro (this is the actual code used by
		1002	Coro to accomplish this):
		1003
		1004	if (defined $AnyEvent::MODEL) {
		1005	# AnyEvent already initialised, so load Coro::AnyEvent
		1006	require Coro::AnyEvent;
		1007	} else {
		1008	# AnyEvent not yet initialised, so make sure to load Coro::AnyEvent
		1009	# as soon as it is
		1010	push @AnyEvent::post_detect, sub { require Coro::AnyEvent };
		1011	}
438		1012
439	=back	1013	=back
440		1014
441	=head1 WHAT TO DO IN A MODULE	1015	=head1 WHAT TO DO IN A MODULE
442		1016
…		…
446	Be careful when you create watchers in the module body - AnyEvent will	1020	Be careful when you create watchers in the module body - AnyEvent will
447	decide which event module to use as soon as the first method is called, so	1021	decide which event module to use as soon as the first method is called, so
448	by calling AnyEvent in your module body you force the user of your module	1022	by calling AnyEvent in your module body you force the user of your module
449	to load the event module first.	1023	to load the event module first.
450		1024
451	Never call C<< ->wait >> on a condition variable unless you I<know> that	1025	Never call C<< ->recv >> on a condition variable unless you I<know> that
452	the C<< ->broadcast >> method has been called on it already. This is	1026	the C<< ->send >> method has been called on it already. This is
453	because it will stall the whole program, and the whole point of using	1027	because it will stall the whole program, and the whole point of using
454	events is to stay interactive.	1028	events is to stay interactive.
455		1029
456	It is fine, however, to call C<< ->wait >> when the user of your module	1030	It is fine, however, to call C<< ->recv >> when the user of your module
457	requests it (i.e. if you create a http request object ad have a method	1031	requests it (i.e. if you create a http request object ad have a method
458	called C<results> that returns the results, it should call C<< ->wait >>	1032	called C<results> that returns the results, it should call C<< ->recv >>
459	freely, as the user of your module knows what she is doing. always).	1033	freely, as the user of your module knows what she is doing. always).
460		1034
461	=head1 WHAT TO DO IN THE MAIN PROGRAM	1035	=head1 WHAT TO DO IN THE MAIN PROGRAM
462		1036
463	There will always be a single main program - the only place that should	1037	There will always be a single main program - the only place that should
…		…
465		1039
466	If it doesn't care, it can just "use AnyEvent" and use it itself, or not	1040	If it doesn't care, it can just "use AnyEvent" and use it itself, or not
467	do anything special (it does not need to be event-based) and let AnyEvent	1041	do anything special (it does not need to be event-based) and let AnyEvent
468	decide which implementation to chose if some module relies on it.	1042	decide which implementation to chose if some module relies on it.
469		1043
470	If the main program relies on a specific event model. For example, in	1044	If the main program relies on a specific event model - for example, in
471	Gtk2 programs you have to rely on the Glib module. You should load the	1045	Gtk2 programs you have to rely on the Glib module - you should load the
472	event module before loading AnyEvent or any module that uses it: generally	1046	event module before loading AnyEvent or any module that uses it: generally
473	speaking, you should load it as early as possible. The reason is that	1047	speaking, you should load it as early as possible. The reason is that
474	modules might create watchers when they are loaded, and AnyEvent will	1048	modules might create watchers when they are loaded, and AnyEvent will
475	decide on the event model to use as soon as it creates watchers, and it	1049	decide on the event model to use as soon as it creates watchers, and it
476	might chose the wrong one unless you load the correct one yourself.	1050	might chose the wrong one unless you load the correct one yourself.
477		1051
478	You can chose to use a rather inefficient pure-perl implementation by	1052	You can chose to use a pure-perl implementation by loading the
479	loading the C<AnyEvent::Impl::Perl> module, which gives you similar	1053	C<AnyEvent::Impl::Perl> module, which gives you similar behaviour
480	behaviour everywhere, but letting AnyEvent chose is generally better.	1054	everywhere, but letting AnyEvent chose the model is generally better.
		1055
		1056	=head2 MAINLOOP EMULATION
		1057
		1058	Sometimes (often for short test scripts, or even standalone programs who
		1059	only want to use AnyEvent), you do not want to run a specific event loop.
		1060
		1061	In that case, you can use a condition variable like this:
		1062
		1063	AnyEvent->condvar->recv;
		1064
		1065	This has the effect of entering the event loop and looping forever.
		1066
		1067	Note that usually your program has some exit condition, in which case
		1068	it is better to use the "traditional" approach of storing a condition
		1069	variable somewhere, waiting for it, and sending it when the program should
		1070	exit cleanly.
		1071
481		1072
482	=head1 OTHER MODULES	1073	=head1 OTHER MODULES
483		1074
484	L<AnyEvent> itself comes with useful utility modules:	1075	The following is a non-exhaustive list of additional modules that use
485		1076	AnyEvent as a client and can therefore be mixed easily with other AnyEvent
486	To make it easier to do non-blocking IO the modules L<AnyEvent::Handle>	1077	modules and other event loops in the same program. Some of the modules
487	and L<AnyEvent::Socket> are provided. L<AnyEvent::Handle> provides	1078	come as part of AnyEvent, the others are available via CPAN.
488	read and write buffers and manages watchers for reads and writes.
489	L<AnyEvent::Socket> provides means to do non-blocking connects.
490
491	Aside from those there are these modules that support AnyEvent (and use it
492	for non-blocking IO):
493		1079
494	=over 4	1080	=over 4
495		1081
		1082	=item L<AnyEvent::Util>
		1083
		1084	Contains various utility functions that replace often-used but blocking
		1085	functions such as C<inet_aton> by event-/callback-based versions.
		1086
		1087	=item L<AnyEvent::Socket>
		1088
		1089	Provides various utility functions for (internet protocol) sockets,
		1090	addresses and name resolution. Also functions to create non-blocking tcp
		1091	connections or tcp servers, with IPv6 and SRV record support and more.
		1092
		1093	=item L<AnyEvent::Handle>
		1094
		1095	Provide read and write buffers, manages watchers for reads and writes,
		1096	supports raw and formatted I/O, I/O queued and fully transparent and
		1097	non-blocking SSL/TLS (via L<AnyEvent::TLS>.
		1098
		1099	=item L<AnyEvent::DNS>
		1100
		1101	Provides rich asynchronous DNS resolver capabilities.
		1102
		1103	=item L<AnyEvent::HTTP>, L<AnyEvent::IRC>, L<AnyEvent::XMPP>, L<AnyEvent::GPSD>, L<AnyEvent::IGS>, L<AnyEvent::FCP>
		1104
		1105	Implement event-based interfaces to the protocols of the same name (for
		1106	the curious, IGS is the International Go Server and FCP is the Freenet
		1107	Client Protocol).
		1108
		1109	=item L<AnyEvent::Handle::UDP>
		1110
		1111	Here be danger!
		1112
		1113	As Pauli would put it, "Not only is it not right, it's not even wrong!" -
		1114	there are so many things wrong with AnyEvent::Handle::UDP, most notably
		1115	it's use of a stream-based API with a protocol that isn't streamable, that
		1116	the only way to improve it is to delete it.
		1117
		1118	It features data corruption (but typically only under load) and general
		1119	confusion. On top, the author is not only clueless about UDP but also
		1120	fact-resistant - some gems of his understanding: "connect doesn't work
		1121	with UDP", "UDP packets are not IP packets", "UDP only has datagrams, not
		1122	packets", "I don't need to implement proper error checking as UDP doesn't
		1123	support error checking" and so on - he doesn't even understand what's
		1124	wrong with his module when it is explained to him.
		1125
		1126	=item L<AnyEvent::DBI>
		1127
		1128	Executes L<DBI> requests asynchronously in a proxy process for you,
		1129	notifying you in an event-bnased way when the operation is finished.
		1130
		1131	=item L<AnyEvent::AIO>
		1132
		1133	Truly asynchronous (as opposed to non-blocking) I/O, should be in the
		1134	toolbox of every event programmer. AnyEvent::AIO transparently fuses
		1135	L<IO::AIO> and AnyEvent together, giving AnyEvent access to event-based
		1136	file I/O, and much more.
		1137
		1138	=item L<AnyEvent::HTTPD>
		1139
		1140	A simple embedded webserver.
		1141
496	=item L<AnyEvent::FastPing>	1142	=item L<AnyEvent::FastPing>
497		1143
498	=item L<Net::IRC3>	1144	The fastest ping in the west.
499		1145
500	=item L<Net::XMPP2>	1146	=item L<Coro>
		1147
		1148	Has special support for AnyEvent via L<Coro::AnyEvent>.
501		1149
502	=back	1150	=back
503		1151
504	=cut	1152	=cut
505		1153
506	package AnyEvent;	1154	package AnyEvent;
507		1155
508	no warnings;	1156	# basically a tuned-down version of common::sense
509	use strict;	1157	sub common_sense {
		1158	# from common:.sense 1.0
		1159	${^WARNING_BITS} = "\xfc\x3f\x33\x00\x0f\xf3\xcf\xc0\xf3\xfc\x33\x00";
		1160	# use strict vars subs - NO UTF-8, as Util.pm doesn't like this atm. (uts46data.pl)
		1161	$^H |= 0x00000600;
		1162	}
510		1163
		1164	BEGIN { AnyEvent::common_sense }
		1165
511	use Carp;	1166	use Carp ();
512		1167
513	our $VERSION = '3.3';	1168	our $VERSION = '5.271';
514	our $MODEL;	1169	our $MODEL;
515		1170
516	our $AUTOLOAD;	1171	our $AUTOLOAD;
517	our @ISA;	1172	our @ISA;
518		1173
519	our $verbose = $ENV{PERL_ANYEVENT_VERBOSE}*1;
520
521	our @REGISTRY;	1174	our @REGISTRY;
522		1175
		1176	our $VERBOSE;
		1177
		1178	BEGIN {
		1179	require "AnyEvent/constants.pl";
		1180
		1181	eval "sub TAINT (){" . (${^TAINT}*1) . "}";
		1182
		1183	delete @ENV{grep /^PERL_ANYEVENT_/, keys %ENV}
		1184	if ${^TAINT};
		1185
		1186	$VERBOSE = $ENV{PERL_ANYEVENT_VERBOSE}*1;
		1187
		1188	}
		1189
		1190	our $MAX_SIGNAL_LATENCY = 10;
		1191
		1192	our %PROTOCOL; # (ipv4|ipv6) => (1|2), higher numbers are preferred
		1193
		1194	{
		1195	my $idx;
		1196	$PROTOCOL{$_} = ++$idx
		1197	for reverse split /\s*,\s*/,
		1198	$ENV{PERL_ANYEVENT_PROTOCOLS} || "ipv4,ipv6";
		1199	}
		1200
523	my @models = (	1201	my @models = (
524	[Coro::EV:: => AnyEvent::Impl::CoroEV::],
525	[Coro::Event:: => AnyEvent::Impl::CoroEvent::],
526	[EV:: => AnyEvent::Impl::EV::],	1202	[EV:: => AnyEvent::Impl::EV:: , 1],
		1203	[AnyEvent::Impl::Perl:: => AnyEvent::Impl::Perl:: , 1],
		1204	# everything below here will not (normally) be autoprobed
		1205	# as the pureperl backend should work everywhere
		1206	# and is usually faster
527	[Event:: => AnyEvent::Impl::Event::],	1207	[Event:: => AnyEvent::Impl::Event::, 1],
528	[Glib:: => AnyEvent::Impl::Glib::],	1208	[Glib:: => AnyEvent::Impl::Glib:: , 1], # becomes extremely slow with many watchers
		1209	[Event::Lib:: => AnyEvent::Impl::EventLib::], # too buggy
		1210	[Irssi:: => AnyEvent::Impl::Irssi::], # Irssi has a bogus "Event" package
529	[Tk:: => AnyEvent::Impl::Tk::],	1211	[Tk:: => AnyEvent::Impl::Tk::], # crashes with many handles
		1212	[Qt:: => AnyEvent::Impl::Qt::], # requires special main program
		1213	[POE::Kernel:: => AnyEvent::Impl::POE::], # lasciate ogni speranza
530	[Wx:: => AnyEvent::Impl::POE::],	1214	[Wx:: => AnyEvent::Impl::POE::],
531	[Prima:: => AnyEvent::Impl::POE::],	1215	[Prima:: => AnyEvent::Impl::POE::],
532	[AnyEvent::Impl::Perl:: => AnyEvent::Impl::Perl::],	1216	# IO::Async is just too broken - we would need workarounds for its
533	# everything below here will not be autoprobed as the pureperl backend should work everywhere	1217	# byzantine signal and broken child handling, among others.
534	[Event::Lib:: => AnyEvent::Impl::EventLib::], # too buggy	1218	# IO::Async is rather hard to detect, as it doesn't have any
		1219	# obvious default class.
535	[Qt:: => AnyEvent::Impl::Qt::], # requires special main program	1220	[IO::Async:: => AnyEvent::Impl::IOAsync::], # requires special main program
536	[POE::Kernel:: => AnyEvent::Impl::POE::], # lasciate ogni speranza	1221	[IO::Async::Loop:: => AnyEvent::Impl::IOAsync::], # requires special main program
		1222	[IO::Async::Notifier:: => AnyEvent::Impl::IOAsync::], # requires special main program
		1223	[AnyEvent::Impl::IOAsync:: => AnyEvent::Impl::IOAsync::], # requires special main program
537	);	1224	);
538		1225
539	our %method = map +($_ => 1), qw(io timer signal child condvar broadcast wait one_event DESTROY);	1226	our %method = map +($_ => 1),
		1227	qw(io timer time now now_update signal child idle condvar one_event DESTROY);
		1228
		1229	our @post_detect;
		1230
		1231	sub post_detect(&) {
		1232	my ($cb) = @_;
		1233
		1234	push @post_detect, $cb;
		1235
		1236	defined wantarray
		1237	? bless \$cb, "AnyEvent::Util::postdetect"
		1238	: ()
		1239	}
		1240
		1241	sub AnyEvent::Util::postdetect::DESTROY {
		1242	@post_detect = grep $_ != ${$_[0]}, @post_detect;
		1243	}
540		1244
541	sub detect() {	1245	sub detect() {
		1246	# free some memory
		1247	*detect = sub () { $MODEL };
		1248
		1249	local $!; # for good measure
		1250	local $SIG{__DIE__};
		1251
		1252	if ($ENV{PERL_ANYEVENT_MODEL} =~ /^([a-zA-Z]+)$/) {
		1253	my $model = "AnyEvent::Impl::$1";
		1254	if (eval "require $model") {
		1255	$MODEL = $model;
		1256	warn "AnyEvent: loaded model '$model' (forced by \$ENV{PERL_ANYEVENT_MODEL}), using it.\n" if $VERBOSE >= 2;
		1257	} else {
		1258	warn "AnyEvent: unable to load model '$model' (from \$ENV{PERL_ANYEVENT_MODEL}):\n$@" if $VERBOSE;
		1259	}
		1260	}
		1261
		1262	# check for already loaded models
542	unless ($MODEL) {	1263	unless ($MODEL) {
543	no strict 'refs';	1264	for (@REGISTRY, @models) {
544		1265	my ($package, $model) = @$_;
545	if ($ENV{PERL_ANYEVENT_MODEL} =~ /^([a-zA-Z]+)$/) {	1266	if (${"$package\::VERSION"} > 0) {
546	my $model = "AnyEvent::Impl::$1";
547	if (eval "require $model") {	1267	if (eval "require $model") {
548	$MODEL = $model;	1268	$MODEL = $model;
549	warn "AnyEvent: loaded model '$model' (forced by \$PERL_ANYEVENT_MODEL), using it.\n" if $verbose > 1;	1269	warn "AnyEvent: autodetected model '$model', using it.\n" if $VERBOSE >= 2;
550	} else {	1270	last;
551	warn "AnyEvent: unable to load model '$model' (from \$PERL_ANYEVENT_MODEL):\n$@" if $verbose;	1271	}
552	}	1272	}
553	}	1273	}
554		1274
555	# check for already loaded models
556	unless ($MODEL) {	1275	unless ($MODEL) {
		1276	# try to autoload a model
557	for (@REGISTRY, @models) {	1277	for (@REGISTRY, @models) {
558	my ($package, $model) = @$_;	1278	my ($package, $model, $autoload) = @$_;
		1279	if (
		1280	$autoload
		1281	and eval "require $package"
559	if (${"$package\::VERSION"} > 0) {	1282	and ${"$package\::VERSION"} > 0
560	if (eval "require $model") {	1283	and eval "require $model"
		1284	) {
561	$MODEL = $model;	1285	$MODEL = $model;
562	warn "AnyEvent: autodetected model '$model', using it.\n" if $verbose > 1;	1286	warn "AnyEvent: autoloaded model '$model', using it.\n" if $VERBOSE >= 2;
563	last;	1287	last;
564	}
565	}	1288	}
566	}	1289	}
567		1290
568	unless ($MODEL) {
569	# try to load a model
570
571	for (@REGISTRY, @models) {
572	my ($package, $model) = @$_;
573	if (eval "require $package"
574	and ${"$package\::VERSION"} > 0
575	and eval "require $model") {
576	$MODEL = $model;
577	warn "AnyEvent: autoprobed model '$model', using it.\n" if $verbose > 1;
578	last;
579	}
580	}
581
582	$MODEL	1291	$MODEL
583	or die "No event module selected for AnyEvent and autodetect failed. Install any one of these modules: EV (or Coro+EV), Event (or Coro+Event) or Glib.";	1292	or die "No event module selected for AnyEvent and autodetect failed. Install any one of these modules: EV, Event or Glib.\n";
584	}
585	}	1293	}
586
587	unshift @ISA, $MODEL;
588	push @{"$MODEL\::ISA"}, "AnyEvent::Base";
589	}	1294	}
		1295
		1296	@models = (); # free probe data
		1297
		1298	push @{"$MODEL\::ISA"}, "AnyEvent::Base";
		1299	unshift @ISA, $MODEL;
		1300
		1301	# now nuke some methods that are overriden by the backend.
		1302	# SUPER is not allowed.
		1303	for (qw(time signal child idle)) {
		1304	undef &{"AnyEvent::Base::$_"}
		1305	if defined &{"$MODEL\::$_"};
		1306	}
		1307
		1308	require AnyEvent::Strict if $ENV{PERL_ANYEVENT_STRICT};
		1309
		1310	(shift @post_detect)->() while @post_detect;
		1311
		1312	*post_detect = sub(&) {
		1313	shift->();
		1314
		1315	undef
		1316	};
590		1317
591	$MODEL	1318	$MODEL
592	}	1319	}
593		1320
594	sub AUTOLOAD {	1321	sub AUTOLOAD {
595	(my $func = $AUTOLOAD) =~ s/.*://;	1322	(my $func = $AUTOLOAD) =~ s/.*://;
596		1323
597	$method{$func}	1324	$method{$func}
598	or croak "$func: not a valid method for AnyEvent objects";	1325	or Carp::croak "$func: not a valid AnyEvent class method";
599		1326
600	detect unless $MODEL;	1327	detect;
601		1328
602	my $class = shift;	1329	my $class = shift;
603	$class->$func (@_);	1330	$class->$func (@_);
604	}	1331	}
605		1332
		1333	# utility function to dup a filehandle. this is used by many backends
		1334	# to support binding more than one watcher per filehandle (they usually
		1335	# allow only one watcher per fd, so we dup it to get a different one).
		1336	sub _dupfh($$;$$) {
		1337	my ($poll, $fh, $r, $w) = @_;
		1338
		1339	# cygwin requires the fh mode to be matching, unix doesn't
		1340	my ($rw, $mode) = $poll eq "r" ? ($r, "<&") : ($w, ">&");
		1341
		1342	open my $fh2, $mode, $fh
		1343	or die "AnyEvent->io: cannot dup() filehandle in mode '$poll': $!,";
		1344
		1345	# we assume CLOEXEC is already set by perl in all important cases
		1346
		1347	($fh2, $rw)
		1348	}
		1349
		1350	=head1 SIMPLIFIED AE API
		1351
		1352	Starting with version 5.0, AnyEvent officially supports a second, much
		1353	simpler, API that is designed to reduce the calling, typing and memory
		1354	overhead by using function call syntax and a fixed number of parameters.
		1355
		1356	See the L<AE> manpage for details.
		1357
		1358	=cut
		1359
		1360	package AE;
		1361
		1362	our $VERSION = $AnyEvent::VERSION;
		1363
		1364	# fall back to the main API by default - backends and AnyEvent::Base
		1365	# implementations can overwrite these.
		1366
		1367	sub io($$$) {
		1368	AnyEvent->io (fh => $_[0], poll => $_[1] ? "w" : "r", cb => $_[2])
		1369	}
		1370
		1371	sub timer($$$) {
		1372	AnyEvent->timer (after => $_[0], interval => $_[1], cb => $_[2])
		1373	}
		1374
		1375	sub signal($$) {
		1376	AnyEvent->signal (signal => $_[0], cb => $_[1])
		1377	}
		1378
		1379	sub child($$) {
		1380	AnyEvent->child (pid => $_[0], cb => $_[1])
		1381	}
		1382
		1383	sub idle($) {
		1384	AnyEvent->idle (cb => $_[0])
		1385	}
		1386
		1387	sub cv(;&) {
		1388	AnyEvent->condvar (@_ ? (cb => $_[0]) : ())
		1389	}
		1390
		1391	sub now() {
		1392	AnyEvent->now
		1393	}
		1394
		1395	sub now_update() {
		1396	AnyEvent->now_update
		1397	}
		1398
		1399	sub time() {
		1400	AnyEvent->time
		1401	}
		1402
606	package AnyEvent::Base;	1403	package AnyEvent::Base;
607		1404
		1405	# default implementations for many methods
		1406
		1407	sub time {
		1408	eval q{ # poor man's autoloading {}
		1409	# probe for availability of Time::HiRes
		1410	if (eval "use Time::HiRes (); Time::HiRes::time (); 1") {
		1411	warn "AnyEvent: using Time::HiRes for sub-second timing accuracy.\n" if $VERBOSE >= 8;
		1412	*AE::time = \&Time::HiRes::time;
		1413	# if (eval "use POSIX (); (POSIX::times())...
		1414	} else {
		1415	warn "AnyEvent: using built-in time(), WARNING, no sub-second resolution!\n" if $VERBOSE;
		1416	*AE::time = sub (){ time }; # epic fail
		1417	}
		1418
		1419	*time = sub { AE::time }; # different prototypes
		1420	};
		1421	die if $@;
		1422
		1423	&time
		1424	}
		1425
		1426	*now = \&time;
		1427
		1428	sub now_update { }
		1429
608	# default implementation for ->condvar, ->wait, ->broadcast	1430	# default implementation for ->condvar
609		1431
610	sub condvar {	1432	sub condvar {
611	bless \my $flag, "AnyEvent::Base::CondVar"	1433	eval q{ # poor man's autoloading {}
612	}	1434	*condvar = sub {
		1435	bless { @_ == 3 ? (_ae_cb => $_[2]) : () }, "AnyEvent::CondVar"
		1436	};
613		1437
614	sub AnyEvent::Base::CondVar::broadcast {	1438	*AE::cv = sub (;&) {
615	${$_[0]}++;	1439	bless { @_ ? (_ae_cb => shift) : () }, "AnyEvent::CondVar"
616	}	1440	};
		1441	};
		1442	die if $@;
617		1443
618	sub AnyEvent::Base::CondVar::wait {	1444	&condvar
619	AnyEvent->one_event while !${$_[0]};
620	}	1445	}
621		1446
622	# default implementation for ->signal	1447	# default implementation for ->signal
623		1448
624	our %SIG_CB;	1449	our $HAVE_ASYNC_INTERRUPT;
		1450
		1451	sub _have_async_interrupt() {
		1452	$HAVE_ASYNC_INTERRUPT = 1*(!$ENV{PERL_ANYEVENT_AVOID_ASYNC_INTERRUPT}
		1453	&& eval "use Async::Interrupt 1.02 (); 1")
		1454	unless defined $HAVE_ASYNC_INTERRUPT;
		1455
		1456	$HAVE_ASYNC_INTERRUPT
		1457	}
		1458
		1459	our ($SIGPIPE_R, $SIGPIPE_W, %SIG_CB, %SIG_EV, $SIG_IO);
		1460	our (%SIG_ASY, %SIG_ASY_W);
		1461	our ($SIG_COUNT, $SIG_TW);
		1462
		1463	# install a dummy wakeup watcher to reduce signal catching latency
		1464	# used by Impls
		1465	sub _sig_add() {
		1466	unless ($SIG_COUNT++) {
		1467	# try to align timer on a full-second boundary, if possible
		1468	my $NOW = AE::now;
		1469
		1470	$SIG_TW = AE::timer
		1471	$MAX_SIGNAL_LATENCY - ($NOW - int $NOW),
		1472	$MAX_SIGNAL_LATENCY,
		1473	sub { } # just for the PERL_ASYNC_CHECK
		1474	;
		1475	}
		1476	}
		1477
		1478	sub _sig_del {
		1479	undef $SIG_TW
		1480	unless --$SIG_COUNT;
		1481	}
		1482
		1483	our $_sig_name_init; $_sig_name_init = sub {
		1484	eval q{ # poor man's autoloading {}
		1485	undef $_sig_name_init;
		1486
		1487	if (_have_async_interrupt) {
		1488	*sig2num = \&Async::Interrupt::sig2num;
		1489	*sig2name = \&Async::Interrupt::sig2name;
		1490	} else {
		1491	require Config;
		1492
		1493	my %signame2num;
		1494	@signame2num{ split ' ', $Config::Config{sig_name} }
		1495	= split ' ', $Config::Config{sig_num};
		1496
		1497	my @signum2name;
		1498	@signum2name[values %signame2num] = keys %signame2num;
		1499
		1500	*sig2num = sub($) {
		1501	$_[0] > 0 ? shift : $signame2num{+shift}
		1502	};
		1503	*sig2name = sub ($) {
		1504	$_[0] > 0 ? $signum2name[+shift] : shift
		1505	};
		1506	}
		1507	};
		1508	die if $@;
		1509	};
		1510
		1511	sub sig2num ($) { &$_sig_name_init; &sig2num }
		1512	sub sig2name($) { &$_sig_name_init; &sig2name }
625		1513
626	sub signal {	1514	sub signal {
		1515	eval q{ # poor man's autoloading {}
		1516	# probe for availability of Async::Interrupt
		1517	if (_have_async_interrupt) {
		1518	warn "AnyEvent: using Async::Interrupt for race-free signal handling.\n" if $VERBOSE >= 8;
		1519
		1520	$SIGPIPE_R = new Async::Interrupt::EventPipe;
		1521	$SIG_IO = AE::io $SIGPIPE_R->fileno, 0, \&_signal_exec;
		1522
		1523	} else {
		1524	warn "AnyEvent: using emulated perl signal handling with latency timer.\n" if $VERBOSE >= 8;
		1525
		1526	if (AnyEvent::WIN32) {
		1527	require AnyEvent::Util;
		1528
		1529	($SIGPIPE_R, $SIGPIPE_W) = AnyEvent::Util::portable_pipe ();
		1530	AnyEvent::Util::fh_nonblocking ($SIGPIPE_R, 1) if $SIGPIPE_R;
		1531	AnyEvent::Util::fh_nonblocking ($SIGPIPE_W, 1) if $SIGPIPE_W; # just in case
		1532	} else {
		1533	pipe $SIGPIPE_R, $SIGPIPE_W;
		1534	fcntl $SIGPIPE_R, AnyEvent::F_SETFL, AnyEvent::O_NONBLOCK if $SIGPIPE_R;
		1535	fcntl $SIGPIPE_W, AnyEvent::F_SETFL, AnyEvent::O_NONBLOCK if $SIGPIPE_W; # just in case
		1536
		1537	# not strictly required, as $^F is normally 2, but let's make sure...
		1538	fcntl $SIGPIPE_R, AnyEvent::F_SETFD, AnyEvent::FD_CLOEXEC;
		1539	fcntl $SIGPIPE_W, AnyEvent::F_SETFD, AnyEvent::FD_CLOEXEC;
		1540	}
		1541
		1542	$SIGPIPE_R
		1543	or Carp::croak "AnyEvent: unable to create a signal reporting pipe: $!\n";
		1544
		1545	$SIG_IO = AE::io $SIGPIPE_R, 0, \&_signal_exec;
		1546	}
		1547
		1548	*signal = $HAVE_ASYNC_INTERRUPT
		1549	? sub {
627	my (undef, %arg) = @_;	1550	my (undef, %arg) = @_;
628		1551
		1552	# async::interrupt
629	my $signal = uc $arg{signal}	1553	my $signal = sig2num $arg{signal};
630	or Carp::croak "required option 'signal' is missing";
631
632	$SIG_CB{$signal}{$arg{cb}} = $arg{cb};	1554	$SIG_CB{$signal}{$arg{cb}} = $arg{cb};
		1555
		1556	$SIG_ASY{$signal} ||= new Async::Interrupt
		1557	cb => sub { undef $SIG_EV{$signal} },
		1558	signal => $signal,
		1559	pipe => [$SIGPIPE_R->filenos],
		1560	pipe_autodrain => 0,
		1561	;
		1562
		1563	bless [$signal, $arg{cb}], "AnyEvent::Base::signal"
		1564	}
		1565	: sub {
		1566	my (undef, %arg) = @_;
		1567
		1568	# pure perl
		1569	my $signal = sig2name $arg{signal};
		1570	$SIG_CB{$signal}{$arg{cb}} = $arg{cb};
		1571
633	$SIG{$signal} ||= sub {	1572	$SIG{$signal} ||= sub {
		1573	local $!;
		1574	syswrite $SIGPIPE_W, "\x00", 1 unless %SIG_EV;
		1575	undef $SIG_EV{$signal};
		1576	};
		1577
		1578	# can't do signal processing without introducing races in pure perl,
		1579	# so limit the signal latency.
		1580	_sig_add;
		1581
		1582	bless [$signal, $arg{cb}], "AnyEvent::Base::signal"
		1583	}
		1584	;
		1585
		1586	*AnyEvent::Base::signal::DESTROY = sub {
		1587	my ($signal, $cb) = @{$_[0]};
		1588
		1589	_sig_del;
		1590
		1591	delete $SIG_CB{$signal}{$cb};
		1592
		1593	$HAVE_ASYNC_INTERRUPT
		1594	? delete $SIG_ASY{$signal}
		1595	: # delete doesn't work with older perls - they then
		1596	# print weird messages, or just unconditionally exit
		1597	# instead of getting the default action.
		1598	undef $SIG{$signal}
		1599	unless keys %{ $SIG_CB{$signal} };
		1600	};
		1601
		1602	*_signal_exec = sub {
		1603	$HAVE_ASYNC_INTERRUPT
		1604	? $SIGPIPE_R->drain
		1605	: sysread $SIGPIPE_R, (my $dummy), 9;
		1606
		1607	while (%SIG_EV) {
		1608	for (keys %SIG_EV) {
		1609	delete $SIG_EV{$_};
634	$_->() for values %{ $SIG_CB{$signal} || {} };	1610	$_->() for values %{ $SIG_CB{$_} || {} };
		1611	}
		1612	}
		1613	};
635	};	1614	};
		1615	die if $@;
636		1616
637	bless [$signal, $arg{cb}], "AnyEvent::Base::Signal"	1617	&signal
638	}
639
640	sub AnyEvent::Base::Signal::DESTROY {
641	my ($signal, $cb) = @{$_[0]};
642
643	delete $SIG_CB{$signal}{$cb};
644
645	$SIG{$signal} = 'DEFAULT' unless keys %{ $SIG_CB{$signal} };
646	}	1618	}
647		1619
648	# default implementation for ->child	1620	# default implementation for ->child
649		1621
650	our %PID_CB;	1622	our %PID_CB;
651	our $CHLD_W;	1623	our $CHLD_W;
652	our $CHLD_DELAY_W;	1624	our $CHLD_DELAY_W;
653	our $PID_IDLE;
654	our $WNOHANG;	1625	our $WNOHANG;
655		1626
656	sub _child_wait {	1627	# used by many Impl's
657	while (0 < (my $pid = waitpid -1, $WNOHANG)) {	1628	sub _emit_childstatus($$) {
		1629	my (undef, $rpid, $rstatus) = @_;
		1630
		1631	$_->($rpid, $rstatus)
658	$_->($pid, $?) for (values %{ $PID_CB{$pid} || {} }),	1632	for values %{ $PID_CB{$rpid} || {} },
659	(values %{ $PID_CB{0} || {} });	1633	values %{ $PID_CB{0} || {} };
660	}
661
662	undef $PID_IDLE;
663	}
664
665	sub _sigchld {
666	# make sure we deliver these changes "synchronous" with the event loop.
667	$CHLD_DELAY_W ||= AnyEvent->timer (after => 0, cb => sub {
668	undef $CHLD_DELAY_W;
669	&_child_wait;
670	});
671	}	1634	}
672		1635
673	sub child {	1636	sub child {
		1637	eval q{ # poor man's autoloading {}
		1638	*_sigchld = sub {
		1639	my $pid;
		1640
		1641	AnyEvent->_emit_childstatus ($pid, $?)
		1642	while ($pid = waitpid -1, $WNOHANG) > 0;
		1643	};
		1644
		1645	*child = sub {
674	my (undef, %arg) = @_;	1646	my (undef, %arg) = @_;
675		1647
676	defined (my $pid = $arg{pid} + 0)	1648	defined (my $pid = $arg{pid} + 0)
677	or Carp::croak "required option 'pid' is missing";	1649	or Carp::croak "required option 'pid' is missing";
678		1650
679	$PID_CB{$pid}{$arg{cb}} = $arg{cb};	1651	$PID_CB{$pid}{$arg{cb}} = $arg{cb};
680		1652
681	unless ($WNOHANG) {	1653	# WNOHANG is almost cetrainly 1 everywhere
682	$WNOHANG = eval { require POSIX; &POSIX::WNOHANG } || 1;	1654	$WNOHANG ||= $^O =~ /^(?:openbsd|netbsd|linux|freebsd|cygwin|MSWin32)$/
683	}	1655	? 1
		1656	: eval { local $SIG{__DIE__}; require POSIX; &POSIX::WNOHANG } || 1;
684		1657
685	unless ($CHLD_W) {	1658	unless ($CHLD_W) {
686	$CHLD_W = AnyEvent->signal (signal => 'CHLD', cb => \&_sigchld);	1659	$CHLD_W = AE::signal CHLD => \&_sigchld;
687	# child could be a zombie already, so make at least one round	1660	# child could be a zombie already, so make at least one round
688	&_sigchld;	1661	&_sigchld;
689	}	1662	}
690		1663
691	bless [$pid, $arg{cb}], "AnyEvent::Base::Child"	1664	bless [$pid, $arg{cb}], "AnyEvent::Base::child"
692	}	1665	};
693		1666
694	sub AnyEvent::Base::Child::DESTROY {	1667	*AnyEvent::Base::child::DESTROY = sub {
695	my ($pid, $cb) = @{$_[0]};	1668	my ($pid, $cb) = @{$_[0]};
696		1669
697	delete $PID_CB{$pid}{$cb};	1670	delete $PID_CB{$pid}{$cb};
698	delete $PID_CB{$pid} unless keys %{ $PID_CB{$pid} };	1671	delete $PID_CB{$pid} unless keys %{ $PID_CB{$pid} };
699		1672
700	undef $CHLD_W unless keys %PID_CB;	1673	undef $CHLD_W unless keys %PID_CB;
		1674	};
		1675	};
		1676	die if $@;
		1677
		1678	&child
701	}	1679	}
		1680
		1681	# idle emulation is done by simply using a timer, regardless
		1682	# of whether the process is idle or not, and not letting
		1683	# the callback use more than 50% of the time.
		1684	sub idle {
		1685	eval q{ # poor man's autoloading {}
		1686	*idle = sub {
		1687	my (undef, %arg) = @_;
		1688
		1689	my ($cb, $w, $rcb) = $arg{cb};
		1690
		1691	$rcb = sub {
		1692	if ($cb) {
		1693	$w = _time;
		1694	&$cb;
		1695	$w = _time - $w;
		1696
		1697	# never use more then 50% of the time for the idle watcher,
		1698	# within some limits
		1699	$w = 0.0001 if $w < 0.0001;
		1700	$w = 5 if $w > 5;
		1701
		1702	$w = AE::timer $w, 0, $rcb;
		1703	} else {
		1704	# clean up...
		1705	undef $w;
		1706	undef $rcb;
		1707	}
		1708	};
		1709
		1710	$w = AE::timer 0.05, 0, $rcb;
		1711
		1712	bless \\$cb, "AnyEvent::Base::idle"
		1713	};
		1714
		1715	*AnyEvent::Base::idle::DESTROY = sub {
		1716	undef $${$_[0]};
		1717	};
		1718	};
		1719	die if $@;
		1720
		1721	&idle
		1722	}
		1723
		1724	package AnyEvent::CondVar;
		1725
		1726	our @ISA = AnyEvent::CondVar::Base::;
		1727
		1728	package AnyEvent::CondVar::Base;
		1729
		1730	#use overload
		1731	# '&{}' => sub { my $self = shift; sub { $self->send (@_) } },
		1732	# fallback => 1;
		1733
		1734	# save 300+ kilobytes by dirtily hardcoding overloading
		1735	${"AnyEvent::CondVar::Base::OVERLOAD"}{dummy}++; # Register with magic by touching.
		1736	*{'AnyEvent::CondVar::Base::()'} = sub { }; # "Make it findable via fetchmethod."
		1737	*{'AnyEvent::CondVar::Base::(&{}'} = sub { my $self = shift; sub { $self->send (@_) } }; # &{}
		1738	${'AnyEvent::CondVar::Base::()'} = 1; # fallback
		1739
		1740	our $WAITING;
		1741
		1742	sub _send {
		1743	# nop
		1744	}
		1745
		1746	sub send {
		1747	my $cv = shift;
		1748	$cv->{_ae_sent} = [@_];
		1749	(delete $cv->{_ae_cb})->($cv) if $cv->{_ae_cb};
		1750	$cv->_send;
		1751	}
		1752
		1753	sub croak {
		1754	$_[0]{_ae_croak} = $_[1];
		1755	$_[0]->send;
		1756	}
		1757
		1758	sub ready {
		1759	$_[0]{_ae_sent}
		1760	}
		1761
		1762	sub _wait {
		1763	$WAITING
		1764	and !$_[0]{_ae_sent}
		1765	and Carp::croak "AnyEvent::CondVar: recursive blocking wait detected";
		1766
		1767	local $WAITING = 1;
		1768	AnyEvent->one_event while !$_[0]{_ae_sent};
		1769	}
		1770
		1771	sub recv {
		1772	$_[0]->_wait;
		1773
		1774	Carp::croak $_[0]{_ae_croak} if $_[0]{_ae_croak};
		1775	wantarray ? @{ $_[0]{_ae_sent} } : $_[0]{_ae_sent}[0]
		1776	}
		1777
		1778	sub cb {
		1779	my $cv = shift;
		1780
		1781	@_
		1782	and $cv->{_ae_cb} = shift
		1783	and $cv->{_ae_sent}
		1784	and (delete $cv->{_ae_cb})->($cv);
		1785
		1786	$cv->{_ae_cb}
		1787	}
		1788
		1789	sub begin {
		1790	++$_[0]{_ae_counter};
		1791	$_[0]{_ae_end_cb} = $_[1] if @_ > 1;
		1792	}
		1793
		1794	sub end {
		1795	return if --$_[0]{_ae_counter};
		1796	&{ $_[0]{_ae_end_cb} || sub { $_[0]->send } };
		1797	}
		1798
		1799	# undocumented/compatibility with pre-3.4
		1800	*broadcast = \&send;
		1801	*wait = \&_wait;
		1802
		1803	=head1 ERROR AND EXCEPTION HANDLING
		1804
		1805	In general, AnyEvent does not do any error handling - it relies on the
		1806	caller to do that if required. The L<AnyEvent::Strict> module (see also
		1807	the C<PERL_ANYEVENT_STRICT> environment variable, below) provides strict
		1808	checking of all AnyEvent methods, however, which is highly useful during
		1809	development.
		1810
		1811	As for exception handling (i.e. runtime errors and exceptions thrown while
		1812	executing a callback), this is not only highly event-loop specific, but
		1813	also not in any way wrapped by this module, as this is the job of the main
		1814	program.
		1815
		1816	The pure perl event loop simply re-throws the exception (usually
		1817	within C<< condvar->recv >>), the L<Event> and L<EV> modules call C<<
		1818	$Event/EV::DIED->() >>, L<Glib> uses C<< install_exception_handler >> and
		1819	so on.
		1820
		1821	=head1 ENVIRONMENT VARIABLES
		1822
		1823	The following environment variables are used by this module or its
		1824	submodules.
		1825
		1826	Note that AnyEvent will remove I<all> environment variables starting with
		1827	C<PERL_ANYEVENT_> from C<%ENV> when it is loaded while taint mode is
		1828	enabled.
		1829
		1830	=over 4
		1831
		1832	=item C<PERL_ANYEVENT_VERBOSE>
		1833
		1834	By default, AnyEvent will be completely silent except in fatal
		1835	conditions. You can set this environment variable to make AnyEvent more
		1836	talkative.
		1837
		1838	When set to C<1> or higher, causes AnyEvent to warn about unexpected
		1839	conditions, such as not being able to load the event model specified by
		1840	C<PERL_ANYEVENT_MODEL>.
		1841
		1842	When set to C<2> or higher, cause AnyEvent to report to STDERR which event
		1843	model it chooses.
		1844
		1845	When set to C<8> or higher, then AnyEvent will report extra information on
		1846	which optional modules it loads and how it implements certain features.
		1847
		1848	=item C<PERL_ANYEVENT_STRICT>
		1849
		1850	AnyEvent does not do much argument checking by default, as thorough
		1851	argument checking is very costly. Setting this variable to a true value
		1852	will cause AnyEvent to load C<AnyEvent::Strict> and then to thoroughly
		1853	check the arguments passed to most method calls. If it finds any problems,
		1854	it will croak.
		1855
		1856	In other words, enables "strict" mode.
		1857
		1858	Unlike C<use strict> (or it's modern cousin, C<< use L<common::sense>
		1859	>>, it is definitely recommended to keep it off in production. Keeping
		1860	C<PERL_ANYEVENT_STRICT=1> in your environment while developing programs
		1861	can be very useful, however.
		1862
		1863	=item C<PERL_ANYEVENT_MODEL>
		1864
		1865	This can be used to specify the event model to be used by AnyEvent, before
		1866	auto detection and -probing kicks in. It must be a string consisting
		1867	entirely of ASCII letters. The string C<AnyEvent::Impl::> gets prepended
		1868	and the resulting module name is loaded and if the load was successful,
		1869	used as event model. If it fails to load AnyEvent will proceed with
		1870	auto detection and -probing.
		1871
		1872	This functionality might change in future versions.
		1873
		1874	For example, to force the pure perl model (L<AnyEvent::Impl::Perl>) you
		1875	could start your program like this:
		1876
		1877	PERL_ANYEVENT_MODEL=Perl perl ...
		1878
		1879	=item C<PERL_ANYEVENT_PROTOCOLS>
		1880
		1881	Used by both L<AnyEvent::DNS> and L<AnyEvent::Socket> to determine preferences
		1882	for IPv4 or IPv6. The default is unspecified (and might change, or be the result
		1883	of auto probing).
		1884
		1885	Must be set to a comma-separated list of protocols or address families,
		1886	current supported: C<ipv4> and C<ipv6>. Only protocols mentioned will be
		1887	used, and preference will be given to protocols mentioned earlier in the
		1888	list.
		1889
		1890	This variable can effectively be used for denial-of-service attacks
		1891	against local programs (e.g. when setuid), although the impact is likely
		1892	small, as the program has to handle conenction and other failures anyways.
		1893
		1894	Examples: C<PERL_ANYEVENT_PROTOCOLS=ipv4,ipv6> - prefer IPv4 over IPv6,
		1895	but support both and try to use both. C<PERL_ANYEVENT_PROTOCOLS=ipv4>
		1896	- only support IPv4, never try to resolve or contact IPv6
		1897	addresses. C<PERL_ANYEVENT_PROTOCOLS=ipv6,ipv4> support either IPv4 or
		1898	IPv6, but prefer IPv6 over IPv4.
		1899
		1900	=item C<PERL_ANYEVENT_EDNS0>
		1901
		1902	Used by L<AnyEvent::DNS> to decide whether to use the EDNS0 extension
		1903	for DNS. This extension is generally useful to reduce DNS traffic, but
		1904	some (broken) firewalls drop such DNS packets, which is why it is off by
		1905	default.
		1906
		1907	Setting this variable to C<1> will cause L<AnyEvent::DNS> to announce
		1908	EDNS0 in its DNS requests.
		1909
		1910	=item C<PERL_ANYEVENT_MAX_FORKS>
		1911
		1912	The maximum number of child processes that C<AnyEvent::Util::fork_call>
		1913	will create in parallel.
		1914
		1915	=item C<PERL_ANYEVENT_MAX_OUTSTANDING_DNS>
		1916
		1917	The default value for the C<max_outstanding> parameter for the default DNS
		1918	resolver - this is the maximum number of parallel DNS requests that are
		1919	sent to the DNS server.
		1920
		1921	=item C<PERL_ANYEVENT_RESOLV_CONF>
		1922
		1923	The file to use instead of F</etc/resolv.conf> (or OS-specific
		1924	configuration) in the default resolver. When set to the empty string, no
		1925	default config will be used.
		1926
		1927	=item C<PERL_ANYEVENT_CA_FILE>, C<PERL_ANYEVENT_CA_PATH>.
		1928
		1929	When neither C<ca_file> nor C<ca_path> was specified during
		1930	L<AnyEvent::TLS> context creation, and either of these environment
		1931	variables exist, they will be used to specify CA certificate locations
		1932	instead of a system-dependent default.
		1933
		1934	=item C<PERL_ANYEVENT_AVOID_GUARD> and C<PERL_ANYEVENT_AVOID_ASYNC_INTERRUPT>
		1935
		1936	When these are set to C<1>, then the respective modules are not
		1937	loaded. Mostly good for testing AnyEvent itself.
		1938
		1939	=back
702		1940
703	=head1 SUPPLYING YOUR OWN EVENT MODEL INTERFACE	1941	=head1 SUPPLYING YOUR OWN EVENT MODEL INTERFACE
704		1942
705	This is an advanced topic that you do not normally need to use AnyEvent in	1943	This is an advanced topic that you do not normally need to use AnyEvent in
706	a module. This section is only of use to event loop authors who want to	1944	a module. This section is only of use to event loop authors who want to
…		…
740		1978
741	I<rxvt-unicode> also cheats a bit by not providing blocking access to	1979	I<rxvt-unicode> also cheats a bit by not providing blocking access to
742	condition variables: code blocking while waiting for a condition will	1980	condition variables: code blocking while waiting for a condition will
743	C<die>. This still works with most modules/usages, and blocking calls must	1981	C<die>. This still works with most modules/usages, and blocking calls must
744	not be done in an interactive application, so it makes sense.	1982	not be done in an interactive application, so it makes sense.
745
746	=head1 ENVIRONMENT VARIABLES
747
748	The following environment variables are used by this module:
749
750	=over 4
751
752	=item C<PERL_ANYEVENT_VERBOSE>
753
754	By default, AnyEvent will be completely silent except in fatal
755	conditions. You can set this environment variable to make AnyEvent more
756	talkative.
757
758	When set to C<1> or higher, causes AnyEvent to warn about unexpected
759	conditions, such as not being able to load the event model specified by
760	C<PERL_ANYEVENT_MODEL>.
761
762	When set to C<2> or higher, cause AnyEvent to report to STDERR which event
763	model it chooses.
764
765	=item C<PERL_ANYEVENT_MODEL>
766
767	This can be used to specify the event model to be used by AnyEvent, before
768	autodetection and -probing kicks in. It must be a string consisting
769	entirely of ASCII letters. The string C<AnyEvent::Impl::> gets prepended
770	and the resulting module name is loaded and if the load was successful,
771	used as event model. If it fails to load AnyEvent will proceed with
772	autodetection and -probing.
773
774	This functionality might change in future versions.
775
776	For example, to force the pure perl model (L<AnyEvent::Impl::Perl>) you
777	could start your program like this:
778
779	PERL_ANYEVENT_MODEL=Perl perl ...
780
781	=back
782		1983
783	=head1 EXAMPLE PROGRAM	1984	=head1 EXAMPLE PROGRAM
784		1985
785	The following program uses an I/O watcher to read data from STDIN, a timer	1986	The following program uses an I/O watcher to read data from STDIN, a timer
786	to display a message once per second, and a condition variable to quit the	1987	to display a message once per second, and a condition variable to quit the
…		…
795	poll => 'r',	1996	poll => 'r',
796	cb => sub {	1997	cb => sub {
797	warn "io event <$_[0]>\n"; # will always output <r>	1998	warn "io event <$_[0]>\n"; # will always output <r>
798	chomp (my $input = <STDIN>); # read a line	1999	chomp (my $input = <STDIN>); # read a line
799	warn "read: $input\n"; # output what has been read	2000	warn "read: $input\n"; # output what has been read
800	$cv->broadcast if $input =~ /^q/i; # quit program if /^q/i	2001	$cv->send if $input =~ /^q/i; # quit program if /^q/i
801	},	2002	},
802	);	2003	);
803		2004
804	my $time_watcher; # can only be used once
805
806	sub new_timer {
807	$timer = AnyEvent->timer (after => 1, cb => sub {	2005	my $time_watcher = AnyEvent->timer (after => 1, interval => 1, cb => sub {
808	warn "timeout\n"; # print 'timeout' about every second	2006	warn "timeout\n"; # print 'timeout' at most every second
809	&new_timer; # and restart the time
810	});	2007	});
811	}
812		2008
813	new_timer; # create first timer
814
815	$cv->wait; # wait until user enters /^q/i	2009	$cv->recv; # wait until user enters /^q/i
816		2010
817	=head1 REAL-WORLD EXAMPLE	2011	=head1 REAL-WORLD EXAMPLE
818		2012
819	Consider the L<Net::FCP> module. It features (among others) the following	2013	Consider the L<Net::FCP> module. It features (among others) the following
820	API calls, which are to freenet what HTTP GET requests are to http:	2014	API calls, which are to freenet what HTTP GET requests are to http:
…		…
870	syswrite $txn->{fh}, $txn->{request}	2064	syswrite $txn->{fh}, $txn->{request}
871	or die "connection or write error";	2065	or die "connection or write error";
872	$txn->{w} = AnyEvent->io (fh => $txn->{fh}, poll => 'r', cb => sub { $txn->fh_ready_r });	2066	$txn->{w} = AnyEvent->io (fh => $txn->{fh}, poll => 'r', cb => sub { $txn->fh_ready_r });
873		2067
874	Again, C<fh_ready_r> waits till all data has arrived, and then stores the	2068	Again, C<fh_ready_r> waits till all data has arrived, and then stores the
875	result and signals any possible waiters that the request ahs finished:	2069	result and signals any possible waiters that the request has finished:
876		2070
877	sysread $txn->{fh}, $txn->{buf}, length $txn->{$buf};	2071	sysread $txn->{fh}, $txn->{buf}, length $txn->{$buf};
878		2072
879	if (end-of-file or data complete) {	2073	if (end-of-file or data complete) {
880	$txn->{result} = $txn->{buf};	2074	$txn->{result} = $txn->{buf};
881	$txn->{finished}->broadcast;	2075	$txn->{finished}->send;
882	$txb->{cb}->($txn) of $txn->{cb}; # also call callback	2076	$txb->{cb}->($txn) of $txn->{cb}; # also call callback
883	}	2077	}
884		2078
885	The C<result> method, finally, just waits for the finished signal (if the	2079	The C<result> method, finally, just waits for the finished signal (if the
886	request was already finished, it doesn't wait, of course, and returns the	2080	request was already finished, it doesn't wait, of course, and returns the
887	data:	2081	data:
888		2082
889	$txn->{finished}->wait;	2083	$txn->{finished}->recv;
890	return $txn->{result};	2084	return $txn->{result};
891		2085
892	The actual code goes further and collects all errors (C<die>s, exceptions)	2086	The actual code goes further and collects all errors (C<die>s, exceptions)
893	that occured during request processing. The C<result> method detects	2087	that occurred during request processing. The C<result> method detects
894	whether an exception as thrown (it is stored inside the $txn object)	2088	whether an exception as thrown (it is stored inside the $txn object)
895	and just throws the exception, which means connection errors and other	2089	and just throws the exception, which means connection errors and other
896	problems get reported tot he code that tries to use the result, not in a	2090	problems get reported to the code that tries to use the result, not in a
897	random callback.	2091	random callback.
898		2092
899	All of this enables the following usage styles:	2093	All of this enables the following usage styles:
900		2094
901	1. Blocking:	2095	1. Blocking:
…		…
929		2123
930	my $quit = AnyEvent->condvar;	2124	my $quit = AnyEvent->condvar;
931		2125
932	$fcp->txn_client_get ($url)->cb (sub {	2126	$fcp->txn_client_get ($url)->cb (sub {
933	...	2127	...
934	$quit->broadcast;	2128	$quit->send;
935	});	2129	});
936		2130
937	$quit->wait;	2131	$quit->recv;
938		2132
939		2133
940	=head1 BENCHMARKS	2134	=head1 BENCHMARKS
941		2135
942	To give you an idea of the performance and overheads that AnyEvent adds	2136	To give you an idea of the performance and overheads that AnyEvent adds
…		…
944	of various event loops I prepared some benchmarks.	2138	of various event loops I prepared some benchmarks.
945		2139
946	=head2 BENCHMARKING ANYEVENT OVERHEAD	2140	=head2 BENCHMARKING ANYEVENT OVERHEAD
947		2141
948	Here is a benchmark of various supported event models used natively and	2142	Here is a benchmark of various supported event models used natively and
949	through anyevent. The benchmark creates a lot of timers (with a zero	2143	through AnyEvent. The benchmark creates a lot of timers (with a zero
950	timeout) and I/O watchers (watching STDOUT, a pty, to become writable,	2144	timeout) and I/O watchers (watching STDOUT, a pty, to become writable,
951	which it is), lets them fire exactly once and destroys them again.	2145	which it is), lets them fire exactly once and destroys them again.
952		2146
953	Source code for this benchmark is found as F<eg/bench> in the AnyEvent	2147	Source code for this benchmark is found as F<eg/bench> in the AnyEvent
954	distribution.	2148	distribution. It uses the L<AE> interface, which makes a real difference
		2149	for the EV and Perl backends only.
955		2150
956	=head3 Explanation of the columns	2151	=head3 Explanation of the columns
957		2152
958	I<watcher> is the number of event watchers created/destroyed. Since	2153	I<watcher> is the number of event watchers created/destroyed. Since
959	different event models feature vastly different performances, each event	2154	different event models feature vastly different performances, each event
…		…
971	all watchers, to avoid adding memory overhead. That means closure creation	2166	all watchers, to avoid adding memory overhead. That means closure creation
972	and memory usage is not included in the figures.	2167	and memory usage is not included in the figures.
973		2168
974	I<invoke> is the time, in microseconds, used to invoke a simple	2169	I<invoke> is the time, in microseconds, used to invoke a simple
975	callback. The callback simply counts down a Perl variable and after it was	2170	callback. The callback simply counts down a Perl variable and after it was
976	invoked "watcher" times, it would C<< ->broadcast >> a condvar once to	2171	invoked "watcher" times, it would C<< ->send >> a condvar once to
977	signal the end of this phase.	2172	signal the end of this phase.
978		2173
979	I<destroy> is the time, in microseconds, that it takes to destroy a single	2174	I<destroy> is the time, in microseconds, that it takes to destroy a single
980	watcher.	2175	watcher.
981		2176
982	=head3 Results	2177	=head3 Results
983		2178
984	name watchers bytes create invoke destroy comment	2179	name watchers bytes create invoke destroy comment
985	EV/EV 400000 244 0.56 0.46 0.31 EV native interface	2180	EV/EV 100000 223 0.47 0.43 0.27 EV native interface
986	EV/Any 100000 244 2.50 0.46 0.29 EV + AnyEvent watchers	2181	EV/Any 100000 223 0.48 0.42 0.26 EV + AnyEvent watchers
987	CoroEV/Any 100000 244 2.49 0.44 0.29 coroutines + Coro::Signal	2182	Coro::EV/Any 100000 223 0.47 0.42 0.26 coroutines + Coro::Signal
988	Perl/Any 100000 513 4.92 0.87 1.12 pure perl implementation	2183	Perl/Any 100000 431 2.70 0.74 0.92 pure perl implementation
989	Event/Event 16000 516 31.88 31.30 0.85 Event native interface	2184	Event/Event 16000 516 31.16 31.84 0.82 Event native interface
990	Event/Any 16000 590 35.75 31.42 1.08 Event + AnyEvent watchers	2185	Event/Any 16000 1203 42.61 34.79 1.80 Event + AnyEvent watchers
		2186	IOAsync/Any 16000 1911 41.92 27.45 16.81 via IO::Async::Loop::IO_Poll
		2187	IOAsync/Any 16000 1726 40.69 26.37 15.25 via IO::Async::Loop::Epoll
991	Glib/Any 16000 1357 98.22 12.41 54.00 quadratic behaviour	2188	Glib/Any 16000 1118 89.00 12.57 51.17 quadratic behaviour
992	Tk/Any 2000 1860 26.97 67.98 14.00 SEGV with >> 2000 watchers	2189	Tk/Any 2000 1346 20.96 10.75 8.00 SEGV with >> 2000 watchers
993	POE/Event 2000 6644 108.64 736.02 14.73 via POE::Loop::Event	2190	POE/Any 2000 6951 108.97 795.32 14.24 via POE::Loop::Event
994	POE/Select 2000 6343 94.13 809.12 565.96 via POE::Loop::Select	2191	POE/Any 2000 6648 94.79 774.40 575.51 via POE::Loop::Select
995		2192
996	=head3 Discussion	2193	=head3 Discussion
997		2194
998	The benchmark does I<not> measure scalability of the event loop very	2195	The benchmark does I<not> measure scalability of the event loop very
999	well. For example, a select-based event loop (such as the pure perl one)	2196	well. For example, a select-based event loop (such as the pure perl one)
…		…
1011	benchmark machine, handling an event takes roughly 1600 CPU cycles with	2208	benchmark machine, handling an event takes roughly 1600 CPU cycles with
1012	EV, 3100 CPU cycles with AnyEvent's pure perl loop and almost 3000000 CPU	2209	EV, 3100 CPU cycles with AnyEvent's pure perl loop and almost 3000000 CPU
1013	cycles with POE.	2210	cycles with POE.
1014		2211
1015	C<EV> is the sole leader regarding speed and memory use, which are both	2212	C<EV> is the sole leader regarding speed and memory use, which are both
1016	maximal/minimal, respectively. Even when going through AnyEvent, it uses	2213	maximal/minimal, respectively. When using the L<AE> API there is zero
		2214	overhead (when going through the AnyEvent API create is about 5-6 times
		2215	slower, with other times being equal, so still uses far less memory than
1017	far less memory than any other event loop and is still faster than Event	2216	any other event loop and is still faster than Event natively).
1018	natively.
1019		2217
1020	The pure perl implementation is hit in a few sweet spots (both the	2218	The pure perl implementation is hit in a few sweet spots (both the
1021	constant timeout and the use of a single fd hit optimisations in the perl	2219	constant timeout and the use of a single fd hit optimisations in the perl
1022	interpreter and the backend itself). Nevertheless this shows that it	2220	interpreter and the backend itself). Nevertheless this shows that it
1023	adds very little overhead in itself. Like any select-based backend its	2221	adds very little overhead in itself. Like any select-based backend its
1024	performance becomes really bad with lots of file descriptors (and few of	2222	performance becomes really bad with lots of file descriptors (and few of
1025	them active), of course, but this was not subject of this benchmark.	2223	them active), of course, but this was not subject of this benchmark.
1026		2224
1027	The C<Event> module has a relatively high setup and callback invocation	2225	The C<Event> module has a relatively high setup and callback invocation
1028	cost, but overall scores in on the third place.	2226	cost, but overall scores in on the third place.
		2227
		2228	C<IO::Async> performs admirably well, about on par with C<Event>, even
		2229	when using its pure perl backend.
1029		2230
1030	C<Glib>'s memory usage is quite a bit higher, but it features a	2231	C<Glib>'s memory usage is quite a bit higher, but it features a
1031	faster callback invocation and overall ends up in the same class as	2232	faster callback invocation and overall ends up in the same class as
1032	C<Event>. However, Glib scales extremely badly, doubling the number of	2233	C<Event>. However, Glib scales extremely badly, doubling the number of
1033	watchers increases the processing time by more than a factor of four,	2234	watchers increases the processing time by more than a factor of four,
…		…
1041	file descriptor is dup()ed for each watcher. This shows that the dup()	2242	file descriptor is dup()ed for each watcher. This shows that the dup()
1042	employed by some adaptors is not a big performance issue (it does incur a	2243	employed by some adaptors is not a big performance issue (it does incur a
1043	hidden memory cost inside the kernel which is not reflected in the figures	2244	hidden memory cost inside the kernel which is not reflected in the figures
1044	above).	2245	above).
1045		2246
1046	C<POE>, regardless of underlying event loop (whether using its pure	2247	C<POE>, regardless of underlying event loop (whether using its pure perl
1047	perl select-based backend or the Event module, the POE-EV backend	2248	select-based backend or the Event module, the POE-EV backend couldn't
1048	couldn't be tested because it wasn't working) shows abysmal performance	2249	be tested because it wasn't working) shows abysmal performance and
1049	and memory usage: Watchers use almost 30 times as much memory as	2250	memory usage with AnyEvent: Watchers use almost 30 times as much memory
1050	EV watchers, and 10 times as much memory as Event (the high memory	2251	as EV watchers, and 10 times as much memory as Event (the high memory
1051	requirements are caused by requiring a session for each watcher). Watcher	2252	requirements are caused by requiring a session for each watcher). Watcher
1052	invocation speed is almost 900 times slower than with AnyEvent's pure perl	2253	invocation speed is almost 900 times slower than with AnyEvent's pure perl
		2254	implementation.
		2255
1053	implementation. The design of the POE adaptor class in AnyEvent can not	2256	The design of the POE adaptor class in AnyEvent can not really account
1054	really account for this, as session creation overhead is small compared	2257	for the performance issues, though, as session creation overhead is
1055	to execution of the state machine, which is coded pretty optimally within	2258	small compared to execution of the state machine, which is coded pretty
1056	L<AnyEvent::Impl::POE>. POE simply seems to be abysmally slow.	2259	optimally within L<AnyEvent::Impl::POE> (and while everybody agrees that
		2260	using multiple sessions is not a good approach, especially regarding
		2261	memory usage, even the author of POE could not come up with a faster
		2262	design).
1057		2263
1058	=head3 Summary	2264	=head3 Summary
1059		2265
1060	=over 4	2266	=over 4
1061		2267
…		…
1072		2278
1073	=back	2279	=back
1074		2280
1075	=head2 BENCHMARKING THE LARGE SERVER CASE	2281	=head2 BENCHMARKING THE LARGE SERVER CASE
1076		2282
1077	This benchmark atcually benchmarks the event loop itself. It works by	2283	This benchmark actually benchmarks the event loop itself. It works by
1078	creating a number of "servers": each server consists of a socketpair, a	2284	creating a number of "servers": each server consists of a socket pair, a
1079	timeout watcher that gets reset on activity (but never fires), and an I/O	2285	timeout watcher that gets reset on activity (but never fires), and an I/O
1080	watcher waiting for input on one side of the socket. Each time the socket	2286	watcher waiting for input on one side of the socket. Each time the socket
1081	watcher reads a byte it will write that byte to a random other "server".	2287	watcher reads a byte it will write that byte to a random other "server".
1082		2288
1083	The effect is that there will be a lot of I/O watchers, only part of which	2289	The effect is that there will be a lot of I/O watchers, only part of which
1084	are active at any one point (so there is a constant number of active	2290	are active at any one point (so there is a constant number of active
1085	fds for each loop iterstaion, but which fds these are is random). The	2291	fds for each loop iteration, but which fds these are is random). The
1086	timeout is reset each time something is read because that reflects how	2292	timeout is reset each time something is read because that reflects how
1087	most timeouts work (and puts extra pressure on the event loops).	2293	most timeouts work (and puts extra pressure on the event loops).
1088		2294
1089	In this benchmark, we use 10000 socketpairs (20000 sockets), of which 100	2295	In this benchmark, we use 10000 socket pairs (20000 sockets), of which 100
1090	(1%) are active. This mirrors the activity of large servers with many	2296	(1%) are active. This mirrors the activity of large servers with many
1091	connections, most of which are idle at any one point in time.	2297	connections, most of which are idle at any one point in time.
1092		2298
1093	Source code for this benchmark is found as F<eg/bench2> in the AnyEvent	2299	Source code for this benchmark is found as F<eg/bench2> in the AnyEvent
1094	distribution.	2300	distribution. It uses the L<AE> interface, which makes a real difference
		2301	for the EV and Perl backends only.
1095		2302
1096	=head3 Explanation of the columns	2303	=head3 Explanation of the columns
1097		2304
1098	I<sockets> is the number of sockets, and twice the number of "servers" (as	2305	I<sockets> is the number of sockets, and twice the number of "servers" (as
1099	each server has a read and write socket end).	2306	each server has a read and write socket end).
1100		2307
1101	I<create> is the time it takes to create a socketpair (which is	2308	I<create> is the time it takes to create a socket pair (which is
1102	nontrivial) and two watchers: an I/O watcher and a timeout watcher.	2309	nontrivial) and two watchers: an I/O watcher and a timeout watcher.
1103		2310
1104	I<request>, the most important value, is the time it takes to handle a	2311	I<request>, the most important value, is the time it takes to handle a
1105	single "request", that is, reading the token from the pipe and forwarding	2312	single "request", that is, reading the token from the pipe and forwarding
1106	it to another server. This includes deleting the old timeout and creating	2313	it to another server. This includes deleting the old timeout and creating
1107	a new one that moves the timeout into the future.	2314	a new one that moves the timeout into the future.
1108		2315
1109	=head3 Results	2316	=head3 Results
1110		2317
1111	name sockets create request	2318	name sockets create request
1112	EV 20000 69.01 11.16	2319	EV 20000 62.66 7.99
1113	Perl 20000 73.32 35.87	2320	Perl 20000 68.32 32.64
1114	Event 20000 212.62 257.32	2321	IOAsync 20000 174.06 101.15 epoll
1115	Glib 20000 651.16 1896.30	2322	IOAsync 20000 174.67 610.84 poll
		2323	Event 20000 202.69 242.91
		2324	Glib 20000 557.01 1689.52
1116	POE 20000 349.67 12317.24 uses POE::Loop::Event	2325	POE 20000 341.54 12086.32 uses POE::Loop::Event
1117		2326
1118	=head3 Discussion	2327	=head3 Discussion
1119		2328
1120	This benchmark I<does> measure scalability and overall performance of the	2329	This benchmark I<does> measure scalability and overall performance of the
1121	particular event loop.	2330	particular event loop.
…		…
1123	EV is again fastest. Since it is using epoll on my system, the setup time	2332	EV is again fastest. Since it is using epoll on my system, the setup time
1124	is relatively high, though.	2333	is relatively high, though.
1125		2334
1126	Perl surprisingly comes second. It is much faster than the C-based event	2335	Perl surprisingly comes second. It is much faster than the C-based event
1127	loops Event and Glib.	2336	loops Event and Glib.
		2337
		2338	IO::Async performs very well when using its epoll backend, and still quite
		2339	good compared to Glib when using its pure perl backend.
1128		2340
1129	Event suffers from high setup time as well (look at its code and you will	2341	Event suffers from high setup time as well (look at its code and you will
1130	understand why). Callback invocation also has a high overhead compared to	2342	understand why). Callback invocation also has a high overhead compared to
1131	the C<< $_->() for .. >>-style loop that the Perl event loop uses. Event	2343	the C<< $_->() for .. >>-style loop that the Perl event loop uses. Event
1132	uses select or poll in basically all documented configurations.	2344	uses select or poll in basically all documented configurations.
…		…
1140		2352
1141	=head3 Summary	2353	=head3 Summary
1142		2354
1143	=over 4	2355	=over 4
1144		2356
1145	=item * The pure perl implementation performs extremely well, considering	2357	=item * The pure perl implementation performs extremely well.
1146	that it uses select.
1147		2358
1148	=item * Avoid Glib or POE in large projects where performance matters.	2359	=item * Avoid Glib or POE in large projects where performance matters.
1149		2360
1150	=back	2361	=back
1151		2362
…		…
1180	speed most when you have lots of watchers, not when you only have a few of	2391	speed most when you have lots of watchers, not when you only have a few of
1181	them).	2392	them).
1182		2393
1183	EV is again fastest.	2394	EV is again fastest.
1184		2395
1185	The C-based event loops Event and Glib come in second this time, as the	2396	Perl again comes second. It is noticeably faster than the C-based event
1186	overhead of running an iteration is much smaller in C than in Perl (little	2397	loops Event and Glib, although the difference is too small to really
1187	code to execute in the inner loop, and perl's function calling overhead is	2398	matter.
1188	high, and updating all the data structures is costly).
1189
1190	The pure perl event loop is much slower, but still competitive.
1191		2399
1192	POE also performs much better in this case, but is is still far behind the	2400	POE also performs much better in this case, but is is still far behind the
1193	others.	2401	others.
1194		2402
1195	=head3 Summary	2403	=head3 Summary
…		…
1199	=item * C-based event loops perform very well with small number of	2407	=item * C-based event loops perform very well with small number of
1200	watchers, as the management overhead dominates.	2408	watchers, as the management overhead dominates.
1201		2409
1202	=back	2410	=back
1203		2411
		2412	=head2 THE IO::Lambda BENCHMARK
		2413
		2414	Recently I was told about the benchmark in the IO::Lambda manpage, which
		2415	could be misinterpreted to make AnyEvent look bad. In fact, the benchmark
		2416	simply compares IO::Lambda with POE, and IO::Lambda looks better (which
		2417	shouldn't come as a surprise to anybody). As such, the benchmark is
		2418	fine, and mostly shows that the AnyEvent backend from IO::Lambda isn't
		2419	very optimal. But how would AnyEvent compare when used without the extra
		2420	baggage? To explore this, I wrote the equivalent benchmark for AnyEvent.
		2421
		2422	The benchmark itself creates an echo-server, and then, for 500 times,
		2423	connects to the echo server, sends a line, waits for the reply, and then
		2424	creates the next connection. This is a rather bad benchmark, as it doesn't
		2425	test the efficiency of the framework or much non-blocking I/O, but it is a
		2426	benchmark nevertheless.
		2427
		2428	name runtime
		2429	Lambda/select 0.330 sec
		2430	+ optimized 0.122 sec
		2431	Lambda/AnyEvent 0.327 sec
		2432	+ optimized 0.138 sec
		2433	Raw sockets/select 0.077 sec
		2434	POE/select, components 0.662 sec
		2435	POE/select, raw sockets 0.226 sec
		2436	POE/select, optimized 0.404 sec
		2437
		2438	AnyEvent/select/nb 0.085 sec
		2439	AnyEvent/EV/nb 0.068 sec
		2440	+state machine 0.134 sec
		2441
		2442	The benchmark is also a bit unfair (my fault): the IO::Lambda/POE
		2443	benchmarks actually make blocking connects and use 100% blocking I/O,
		2444	defeating the purpose of an event-based solution. All of the newly
		2445	written AnyEvent benchmarks use 100% non-blocking connects (using
		2446	AnyEvent::Socket::tcp_connect and the asynchronous pure perl DNS
		2447	resolver), so AnyEvent is at a disadvantage here, as non-blocking connects
		2448	generally require a lot more bookkeeping and event handling than blocking
		2449	connects (which involve a single syscall only).
		2450
		2451	The last AnyEvent benchmark additionally uses L<AnyEvent::Handle>, which
		2452	offers similar expressive power as POE and IO::Lambda, using conventional
		2453	Perl syntax. This means that both the echo server and the client are 100%
		2454	non-blocking, further placing it at a disadvantage.
		2455
		2456	As you can see, the AnyEvent + EV combination even beats the
		2457	hand-optimised "raw sockets benchmark", while AnyEvent + its pure perl
		2458	backend easily beats IO::Lambda and POE.
		2459
		2460	And even the 100% non-blocking version written using the high-level (and
		2461	slow :) L<AnyEvent::Handle> abstraction beats both POE and IO::Lambda
		2462	higher level ("unoptimised") abstractions by a large margin, even though
		2463	it does all of DNS, tcp-connect and socket I/O in a non-blocking way.
		2464
		2465	The two AnyEvent benchmarks programs can be found as F<eg/ae0.pl> and
		2466	F<eg/ae2.pl> in the AnyEvent distribution, the remaining benchmarks are
		2467	part of the IO::Lambda distribution and were used without any changes.
		2468
		2469
		2470	=head1 SIGNALS
		2471
		2472	AnyEvent currently installs handlers for these signals:
		2473
		2474	=over 4
		2475
		2476	=item SIGCHLD
		2477
		2478	A handler for C<SIGCHLD> is installed by AnyEvent's child watcher
		2479	emulation for event loops that do not support them natively. Also, some
		2480	event loops install a similar handler.
		2481
		2482	Additionally, when AnyEvent is loaded and SIGCHLD is set to IGNORE, then
		2483	AnyEvent will reset it to default, to avoid losing child exit statuses.
		2484
		2485	=item SIGPIPE
		2486
		2487	A no-op handler is installed for C<SIGPIPE> when C<$SIG{PIPE}> is C<undef>
		2488	when AnyEvent gets loaded.
		2489
		2490	The rationale for this is that AnyEvent users usually do not really depend
		2491	on SIGPIPE delivery (which is purely an optimisation for shell use, or
		2492	badly-written programs), but C<SIGPIPE> can cause spurious and rare
		2493	program exits as a lot of people do not expect C<SIGPIPE> when writing to
		2494	some random socket.
		2495
		2496	The rationale for installing a no-op handler as opposed to ignoring it is
		2497	that this way, the handler will be restored to defaults on exec.
		2498
		2499	Feel free to install your own handler, or reset it to defaults.
		2500
		2501	=back
		2502
		2503	=cut
		2504
		2505	undef $SIG{CHLD}
		2506	if $SIG{CHLD} eq 'IGNORE';
		2507
		2508	$SIG{PIPE} = sub { }
		2509	unless defined $SIG{PIPE};
		2510
		2511	=head1 RECOMMENDED/OPTIONAL MODULES
		2512
		2513	One of AnyEvent's main goals is to be 100% Pure-Perl(tm): only perl (and
		2514	it's built-in modules) are required to use it.
		2515
		2516	That does not mean that AnyEvent won't take advantage of some additional
		2517	modules if they are installed.
		2518
		2519	This section explains which additional modules will be used, and how they
		2520	affect AnyEvent's operation.
		2521
		2522	=over 4
		2523
		2524	=item L<Async::Interrupt>
		2525
		2526	This slightly arcane module is used to implement fast signal handling: To
		2527	my knowledge, there is no way to do completely race-free and quick
		2528	signal handling in pure perl. To ensure that signals still get
		2529	delivered, AnyEvent will start an interval timer to wake up perl (and
		2530	catch the signals) with some delay (default is 10 seconds, look for
		2531	C<$AnyEvent::MAX_SIGNAL_LATENCY>).
		2532
		2533	If this module is available, then it will be used to implement signal
		2534	catching, which means that signals will not be delayed, and the event loop
		2535	will not be interrupted regularly, which is more efficient (and good for
		2536	battery life on laptops).
		2537
		2538	This affects not just the pure-perl event loop, but also other event loops
		2539	that have no signal handling on their own (e.g. Glib, Tk, Qt).
		2540
		2541	Some event loops (POE, Event, Event::Lib) offer signal watchers natively,
		2542	and either employ their own workarounds (POE) or use AnyEvent's workaround
		2543	(using C<$AnyEvent::MAX_SIGNAL_LATENCY>). Installing L<Async::Interrupt>
		2544	does nothing for those backends.
		2545
		2546	=item L<EV>
		2547
		2548	This module isn't really "optional", as it is simply one of the backend
		2549	event loops that AnyEvent can use. However, it is simply the best event
		2550	loop available in terms of features, speed and stability: It supports
		2551	the AnyEvent API optimally, implements all the watcher types in XS, does
		2552	automatic timer adjustments even when no monotonic clock is available,
		2553	can take avdantage of advanced kernel interfaces such as C<epoll> and
		2554	C<kqueue>, and is the fastest backend I<by far>. You can even embed
		2555	L<Glib>/L<Gtk2> in it (or vice versa, see L<EV::Glib> and L<Glib::EV>).
		2556
		2557	If you only use backends that rely on another event loop (e.g. C<Tk>),
		2558	then this module will do nothing for you.
		2559
		2560	=item L<Guard>
		2561
		2562	The guard module, when used, will be used to implement
		2563	C<AnyEvent::Util::guard>. This speeds up guards considerably (and uses a
		2564	lot less memory), but otherwise doesn't affect guard operation much. It is
		2565	purely used for performance.
		2566
		2567	=item L<JSON> and L<JSON::XS>
		2568
		2569	One of these modules is required when you want to read or write JSON data
		2570	via L<AnyEvent::Handle>. L<JSON> is also written in pure-perl, but can take
		2571	advantage of the ultra-high-speed L<JSON::XS> module when it is installed.
		2572
		2573	=item L<Net::SSLeay>
		2574
		2575	Implementing TLS/SSL in Perl is certainly interesting, but not very
		2576	worthwhile: If this module is installed, then L<AnyEvent::Handle> (with
		2577	the help of L<AnyEvent::TLS>), gains the ability to do TLS/SSL.
		2578
		2579	=item L<Time::HiRes>
		2580
		2581	This module is part of perl since release 5.008. It will be used when the
		2582	chosen event library does not come with a timing source on it's own. The
		2583	pure-perl event loop (L<AnyEvent::Impl::Perl>) will additionally use it to
		2584	try to use a monotonic clock for timing stability.
		2585
		2586	=back
		2587
1204		2588
1205	=head1 FORK	2589	=head1 FORK
1206		2590
1207	Most event libraries are not fork-safe. The ones who are usually are	2591	Most event libraries are not fork-safe. The ones who are usually are
1208	because they are so inefficient. Only L<EV> is fully fork-aware.	2592	because they rely on inefficient but fork-safe C<select> or C<poll> calls
		2593	- higher performance APIs such as BSD's kqueue or the dreaded Linux epoll
		2594	are usually badly thought-out hacks that are incompatible with fork in
		2595	one way or another. Only L<EV> is fully fork-aware and ensures that you
		2596	continue event-processing in both parent and child (or both, if you know
		2597	what you are doing).
		2598
		2599	This means that, in general, you cannot fork and do event processing in
		2600	the child if the event library was initialised before the fork (which
		2601	usually happens when the first AnyEvent watcher is created, or the library
		2602	is loaded).
1209		2603
1210	If you have to fork, you must either do so I<before> creating your first	2604	If you have to fork, you must either do so I<before> creating your first
1211	watcher OR you must not use AnyEvent at all in the child.	2605	watcher OR you must not use AnyEvent at all in the child OR you must do
		2606	something completely out of the scope of AnyEvent.
		2607
		2608	The problem of doing event processing in the parent I<and> the child
		2609	is much more complicated: even for backends that I<are> fork-aware or
		2610	fork-safe, their behaviour is not usually what you want: fork clones all
		2611	watchers, that means all timers, I/O watchers etc. are active in both
		2612	parent and child, which is almost never what you want. USing C<exec>
		2613	to start worker children from some kind of manage rprocess is usually
		2614	preferred, because it is much easier and cleaner, at the expense of having
		2615	to have another binary.
1212		2616
1213		2617
1214	=head1 SECURITY CONSIDERATIONS	2618	=head1 SECURITY CONSIDERATIONS
1215		2619
1216	AnyEvent can be forced to load any event model via	2620	AnyEvent can be forced to load any event model via
…		…
1221	specified in the variable.	2625	specified in the variable.
1222		2626
1223	You can make AnyEvent completely ignore this variable by deleting it	2627	You can make AnyEvent completely ignore this variable by deleting it
1224	before the first watcher gets created, e.g. with a C<BEGIN> block:	2628	before the first watcher gets created, e.g. with a C<BEGIN> block:
1225		2629
1226	BEGIN { delete $ENV{PERL_ANYEVENT_MODEL} }	2630	BEGIN { delete $ENV{PERL_ANYEVENT_MODEL} }
1227		2631
1228	use AnyEvent;	2632	use AnyEvent;
		2633
		2634	Similar considerations apply to $ENV{PERL_ANYEVENT_VERBOSE}, as that can
		2635	be used to probe what backend is used and gain other information (which is
		2636	probably even less useful to an attacker than PERL_ANYEVENT_MODEL), and
		2637	$ENV{PERL_ANYEVENT_STRICT}.
		2638
		2639	Note that AnyEvent will remove I<all> environment variables starting with
		2640	C<PERL_ANYEVENT_> from C<%ENV> when it is loaded while taint mode is
		2641	enabled.
		2642
		2643
		2644	=head1 BUGS
		2645
		2646	Perl 5.8 has numerous memleaks that sometimes hit this module and are hard
		2647	to work around. If you suffer from memleaks, first upgrade to Perl 5.10
		2648	and check wether the leaks still show up. (Perl 5.10.0 has other annoying
		2649	memleaks, such as leaking on C<map> and C<grep> but it is usually not as
		2650	pronounced).
1229		2651
1230		2652
1231	=head1 SEE ALSO	2653	=head1 SEE ALSO
1232		2654
1233	Event modules: L<Coro::EV>, L<EV>, L<EV::Glib>, L<Glib::EV>,	2655	Utility functions: L<AnyEvent::Util>.
1234	L<Coro::Event>, L<Event>, L<Glib::Event>, L<Glib>, L<Coro>, L<Tk>,	2656
		2657	Event modules: L<EV>, L<EV::Glib>, L<Glib::EV>, L<Event>, L<Glib::Event>,
1235	L<Event::Lib>, L<Qt>, L<POE>.	2658	L<Glib>, L<Tk>, L<Event::Lib>, L<Qt>, L<POE>.
1236		2659
1237	Implementations: L<AnyEvent::Impl::CoroEV>, L<AnyEvent::Impl::EV>,	2660	Implementations: L<AnyEvent::Impl::EV>, L<AnyEvent::Impl::Event>,
		2661	L<AnyEvent::Impl::Glib>, L<AnyEvent::Impl::Tk>, L<AnyEvent::Impl::Perl>,
		2662	L<AnyEvent::Impl::EventLib>, L<AnyEvent::Impl::Qt>,
1238	L<AnyEvent::Impl::CoroEvent>, L<AnyEvent::Impl::Event>, L<AnyEvent::Impl::Glib>,	2663	L<AnyEvent::Impl::POE>, L<AnyEvent::Impl::IOAsync>, L<Anyevent::Impl::Irssi>.
1239	L<AnyEvent::Impl::Tk>, L<AnyEvent::Impl::Perl>, L<AnyEvent::Impl::EventLib>,
1240	L<AnyEvent::Impl::Qt>, L<AnyEvent::Impl::POE>.
1241		2664
		2665	Non-blocking file handles, sockets, TCP clients and
		2666	servers: L<AnyEvent::Handle>, L<AnyEvent::Socket>, L<AnyEvent::TLS>.
		2667
		2668	Asynchronous DNS: L<AnyEvent::DNS>.
		2669
		2670	Coroutine support: L<Coro>, L<Coro::AnyEvent>, L<Coro::EV>,
		2671	L<Coro::Event>,
		2672
1242	Nontrivial usage examples: L<Net::FCP>, L<Net::XMPP2>.	2673	Nontrivial usage examples: L<AnyEvent::GPSD>, L<AnyEvent::XMPP>,
		2674	L<AnyEvent::HTTP>.
1243		2675
1244		2676
1245	=head1 AUTHOR	2677	=head1 AUTHOR
1246		2678
1247	Marc Lehmann <schmorp@schmorp.de>	2679	Marc Lehmann <schmorp@schmorp.de>
1248	http://home.schmorp.de/	2680	http://home.schmorp.de/
1249		2681
1250	=cut	2682	=cut
1251		2683
1252	1	2684	1
1253		2685

Diff Legend

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