[ViewVC] Diff of: cvs/AnyEvent/README

Comparing AnyEvent/README (file contents):
Revision 1.18 by root, Thu Apr 24 09:13:26 2008 UTC vs.
Revision 1.29 by root, Tue Jul 29 10:20:33 2008 UTC

…		…
1	NAME	1	NAME
2	AnyEvent - provide framework for multiple event loops	2	AnyEvent - provide framework for multiple event loops
3		3
4	EV, Event, Coro::EV, Coro::Event, Glib, Tk, Perl, Event::Lib, Qt -	4	EV, Event, Glib, Tk, Perl, Event::Lib, Qt, POE - various supported event
5	various supported event loops	5	loops
6		6
7	SYNOPSIS	7	SYNOPSIS
8	use AnyEvent;	8	use AnyEvent;
9		9
10	my $w = AnyEvent->io (fh => $fh, poll => "r\|w", cb => sub {	10	my $w = AnyEvent->io (fh => $fh, poll => "r\|w", cb => sub { ... });
		11
		12	my $w = AnyEvent->timer (after => $seconds, cb => sub { ... });
		13	my $w = AnyEvent->timer (after => $seconds, interval => $seconds, cb => ...
		14
		15	print AnyEvent->now; # prints current event loop time
		16	print AnyEvent->time; # think Time::HiRes::time or simply CORE::time.
		17
		18	my $w = AnyEvent->signal (signal => "TERM", cb => sub { ... });
		19
		20	my $w = AnyEvent->child (pid => $pid, cb => sub {
		21	my ($pid, $status) = @_;
11	...	22	...
12	});	23	});
13		24
14	my $w = AnyEvent->timer (after => $seconds, cb => sub {
15	...
16	});
17
18	my $w = AnyEvent->condvar; # stores whether a condition was flagged	25	my $w = AnyEvent->condvar; # stores whether a condition was flagged
		26	$w->send; # wake up current and all future recv's
19	$w->wait; # enters "main loop" till $condvar gets ->broadcast	27	$w->recv; # enters "main loop" till $condvar gets ->send
20	$w->broadcast; # wake up current and all future wait's	28	# use a condvar in callback mode:
		29	$w->cb (sub { $_[0]->recv });
		30
		31	INTRODUCTION/TUTORIAL
		32	This manpage is mainly a reference manual. If you are interested in a
		33	tutorial or some gentle introduction, have a look at the AnyEvent::Intro
		34	manpage.
21		35
22	WHY YOU SHOULD USE THIS MODULE (OR NOT)	36	WHY YOU SHOULD USE THIS MODULE (OR NOT)
23	Glib, POE, IO::Async, Event... CPAN offers event models by the dozen	37	Glib, POE, IO::Async, Event... CPAN offers event models by the dozen
24	nowadays. So what is different about AnyEvent?	38	nowadays. So what is different about AnyEvent?
25		39
26	Executive Summary: AnyEvent is compatible, AnyEvent is *free of	40	Executive Summary: AnyEvent is compatible, AnyEvent is *free of
27	policy* and AnyEvent is small and efficient.	41	policy* and AnyEvent is small and efficient.
28		42
29	First and foremost, AnyEvent is not an event model itself, it only	43	First and foremost, AnyEvent is not an event model itself, it only
30	interfaces to whatever event model the main program happens to use in a	44	interfaces to whatever event model the main program happens to use, in a
31	pragmatic way. For event models and certain classes of immortals alike,	45	pragmatic way. For event models and certain classes of immortals alike,
32	the statement "there can only be one" is a bitter reality: In general,	46	the statement "there can only be one" is a bitter reality: In general,
33	only one event loop can be active at the same time in a process.	47	only one event loop can be active at the same time in a process.
34	AnyEvent helps hiding the differences between those event loops.	48	AnyEvent cannot change this, but it can hide the differences between
		49	those event loops.
35		50
36	The goal of AnyEvent is to offer module authors the ability to do event	51	The goal of AnyEvent is to offer module authors the ability to do event
37	programming (waiting for I/O or timer events) without subscribing to a	52	programming (waiting for I/O or timer events) without subscribing to a
38	religion, a way of living, and most importantly: without forcing your	53	religion, a way of living, and most importantly: without forcing your
39	module users into the same thing by forcing them to use the same event	54	module users into the same thing by forcing them to use the same event
40	model you use.	55	model you use.
41		56
42	For modules like POE or IO::Async (which is a total misnomer as it is	57	For modules like POE or IO::Async (which is a total misnomer as it is
43	actually doing all I/O synchronously...), using them in your module is	58	actually doing all I/O synchronously...), using them in your module is
44	like joining a cult: After you joined, you are dependent on them and you	59	like joining a cult: After you joined, you are dependent on them and you
45	cannot use anything else, as it is simply incompatible to everything	60	cannot use anything else, as they are simply incompatible to everything
46	that isn't itself. What's worse, all the potential users of your module	61	that isn't them. What's worse, all the potential users of your module
47	are also forced to use the same event loop you use.	62	are also forced to use the same event loop you use.
48		63
49	AnyEvent is different: AnyEvent + POE works fine. AnyEvent + Glib works	64	AnyEvent is different: AnyEvent + POE works fine. AnyEvent + Glib works
50	fine. AnyEvent + Tk works fine etc. etc. but none of these work together	65	fine. AnyEvent + Tk works fine etc. etc. but none of these work together
51	with the rest: POE + IO::Async? no go. Tk + Event? no go. Again: if your	66	with the rest: POE + IO::Async? No go. Tk + Event? No go. Again: if your
52	module uses one of those, every user of your module has to use it, too.	67	module uses one of those, every user of your module has to use it, too.
53	But if your module uses AnyEvent, it works transparently with all event	68	But if your module uses AnyEvent, it works transparently with all event
54	models it supports (including stuff like POE and IO::Async, as long as	69	models it supports (including stuff like IO::Async, as long as those use
55	those use one of the supported event loops. It is trivial to add new	70	one of the supported event loops. It is trivial to add new event loops
56	event loops to AnyEvent, too, so it is future-proof).	71	to AnyEvent, too, so it is future-proof).
57		72
58	In addition to being free of having to use *the one and only true event	73	In addition to being free of having to use *the one and only true event
59	model*, AnyEvent also is free of bloat and policy: with POE or similar	74	model*, AnyEvent also is free of bloat and policy: with POE or similar
60	modules, you get an enourmous amount of code and strict rules you have	75	modules, you get an enormous amount of code and strict rules you have to
61	to follow. AnyEvent, on the other hand, is lean and up to the point, by	76	follow. AnyEvent, on the other hand, is lean and up to the point, by
62	only offering the functionality that is necessary, in as thin as a	77	only offering the functionality that is necessary, in as thin as a
63	wrapper as technically possible.	78	wrapper as technically possible.
64		79
		80	Of course, AnyEvent comes with a big (and fully optional!) toolbox of
		81	useful functionality, such as an asynchronous DNS resolver, 100%
		82	non-blocking connects (even with TLS/SSL, IPv6 and on broken platforms
		83	such as Windows) and lots of real-world knowledge and workarounds for
		84	platform bugs and differences.
		85
65	Of course, if you want lots of policy (this can arguably be somewhat	86	Now, if you do want lots of policy (this can arguably be somewhat
66	useful) and you want to force your users to use the one and only event	87	useful) and you want to force your users to use the one and only event
67	model, you should not use this module.	88	model, you should not use this module.
68		89
69	DESCRIPTION	90	DESCRIPTION
70	AnyEvent provides an identical interface to multiple event loops. This	91	AnyEvent provides an identical interface to multiple event loops. This
75	The interface itself is vaguely similar, but not identical to the Event	96	The interface itself is vaguely similar, but not identical to the Event
76	module.	97	module.
77		98
78	During the first call of any watcher-creation method, the module tries	99	During the first call of any watcher-creation method, the module tries
79	to detect the currently loaded event loop by probing whether one of the	100	to detect the currently loaded event loop by probing whether one of the
80	following modules is already loaded: Coro::EV, Coro::Event, EV, Event,	101	following modules is already loaded: EV, Event, Glib,
81	Glib, Tk, Event::Lib, Qt. The first one found is used. If none are	102	AnyEvent::Impl::Perl, Tk, Event::Lib, Qt, POE. The first one found is
82	found, the module tries to load these modules (excluding Event::Lib and	103	used. If none are found, the module tries to load these modules
83	Qt) in the order given. The first one that can be successfully loaded	104	(excluding Tk, Event::Lib, Qt and POE as the pure perl adaptor should
84	will be used. If, after this, still none could be found, AnyEvent will	105	always succeed) in the order given. The first one that can be
85	fall back to a pure-perl event loop, which is not very efficient, but	106	successfully loaded will be used. If, after this, still none could be
86	should work everywhere.	107	found, AnyEvent will fall back to a pure-perl event loop, which is not
		108	very efficient, but should work everywhere.
87		109
88	Because AnyEvent first checks for modules that are already loaded,	110	Because AnyEvent first checks for modules that are already loaded,
89	loading an event model explicitly before first using AnyEvent will	111	loading an event model explicitly before first using AnyEvent will
90	likely make that model the default. For example:	112	likely make that model the default. For example:
91		113
…		…
98	starts using it, all bets are off. Maybe you should tell their authors	120	starts using it, all bets are off. Maybe you should tell their authors
99	to use AnyEvent so their modules work together with others seamlessly...	121	to use AnyEvent so their modules work together with others seamlessly...
100		122
101	The pure-perl implementation of AnyEvent is called	123	The pure-perl implementation of AnyEvent is called
102	"AnyEvent::Impl::Perl". Like other event modules you can load it	124	"AnyEvent::Impl::Perl". Like other event modules you can load it
103	explicitly.	125	explicitly and enjoy the high availability of that event loop :)
104		126
105	WATCHERS	127	WATCHERS
106	AnyEvent has the central concept of a watcher, which is an object that	128	AnyEvent has the central concept of a watcher, which is an object that
107	stores relevant data for each kind of event you are waiting for, such as	129	stores relevant data for each kind of event you are waiting for, such as
108	the callback to call, the filehandle to watch, etc.	130	the callback to call, the file handle to watch, etc.
109		131
110	These watchers are normal Perl objects with normal Perl lifetime. After	132	These watchers are normal Perl objects with normal Perl lifetime. After
111	creating a watcher it will immediately "watch" for events and invoke the	133	creating a watcher it will immediately "watch" for events and invoke the
112	callback when the event occurs (of course, only when the event model is	134	callback when the event occurs (of course, only when the event model is
113	in control).	135	in control).
…		…
121	Many watchers either are used with "recursion" (repeating timers for	143	Many watchers either are used with "recursion" (repeating timers for
122	example), or need to refer to their watcher object in other ways.	144	example), or need to refer to their watcher object in other ways.
123		145
124	An any way to achieve that is this pattern:	146	An any way to achieve that is this pattern:
125		147
126	my $w; $w = AnyEvent->type (arg => value ..., cb => sub {	148	my $w; $w = AnyEvent->type (arg => value ..., cb => sub {
127	# you can use $w here, for example to undef it	149	# you can use $w here, for example to undef it
128	undef $w;	150	undef $w;
129	});	151	});
130		152
131	Note that "my $w; $w =" combination. This is necessary because in Perl,	153	Note that "my $w; $w =" combination. This is necessary because in Perl,
132	my variables are only visible after the statement in which they are	154	my variables are only visible after the statement in which they are
133	declared.	155	declared.
134		156
135	IO WATCHERS	157	I/O WATCHERS
136	You can create an I/O watcher by calling the "AnyEvent->io" method with	158	You can create an I/O watcher by calling the "AnyEvent->io" method with
137	the following mandatory key-value pairs as arguments:	159	the following mandatory key-value pairs as arguments:
138		160
139	"fh" the Perl file handle (not file descriptor) to watch for events.	161	"fh" the Perl file handle (not file descriptor) to watch for events
		162	(AnyEvent might or might not keep a reference to this file handle).
140	"poll" must be a string that is either "r" or "w", which creates a	163	"poll" must be a string that is either "r" or "w", which creates a
141	watcher waiting for "r"eadable or "w"ritable events, respectively. "cb"	164	watcher waiting for "r"eadable or "w"ritable events, respectively. "cb"
142	is the callback to invoke each time the file handle becomes ready.	165	is the callback to invoke each time the file handle becomes ready.
143		166
144	As long as the I/O watcher exists it will keep the file descriptor or a	167	Although the callback might get passed parameters, their value and
145	copy of it alive/open.	168	presence is undefined and you cannot rely on them. Portable AnyEvent
		169	callbacks cannot use arguments passed to I/O watcher callbacks.
146		170
		171	The I/O watcher might use the underlying file descriptor or a copy of
147	It is not allowed to close a file handle as long as any watcher is	172	it. You must not close a file handle as long as any watcher is active on
148	active on the underlying file descriptor.	173	the underlying file descriptor.
149		174
150	Some event loops issue spurious readyness notifications, so you should	175	Some event loops issue spurious readyness notifications, so you should
151	always use non-blocking calls when reading/writing from/to your file	176	always use non-blocking calls when reading/writing from/to your file
152	handles.	177	handles.
153		178
154	Example:
155
156	# wait for readability of STDIN, then read a line and disable the watcher	179	Example: wait for readability of STDIN, then read a line and disable the
		180	watcher.
		181
157	my $w; $w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {	182	my $w; $w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {
158	chomp (my $input = <STDIN>);	183	chomp (my $input = <STDIN>);
159	warn "read: $input\n";	184	warn "read: $input\n";
160	undef $w;	185	undef $w;
161	});	186	});
…		…
163	TIME WATCHERS	188	TIME WATCHERS
164	You can create a time watcher by calling the "AnyEvent->timer" method	189	You can create a time watcher by calling the "AnyEvent->timer" method
165	with the following mandatory arguments:	190	with the following mandatory arguments:
166		191
167	"after" specifies after how many seconds (fractional values are	192	"after" specifies after how many seconds (fractional values are
168	supported) should the timer activate. "cb" the callback to invoke in	193	supported) the callback should be invoked. "cb" is the callback to
169	that case.	194	invoke in that case.
170		195
171	The timer callback will be invoked at most once: if you want a repeating	196	Although the callback might get passed parameters, their value and
172	timer you have to create a new watcher (this is a limitation by both Tk	197	presence is undefined and you cannot rely on them. Portable AnyEvent
173	and Glib).	198	callbacks cannot use arguments passed to time watcher callbacks.
174		199
175	Example:	200	The callback will normally be invoked once only. If you specify another
		201	parameter, "interval", as a strictly positive number (> 0), then the
		202	callback will be invoked regularly at that interval (in fractional
		203	seconds) after the first invocation. If "interval" is specified with a
		204	false value, then it is treated as if it were missing.
176		205
		206	The callback will be rescheduled before invoking the callback, but no
		207	attempt is done to avoid timer drift in most backends, so the interval
		208	is only approximate.
		209
177	# fire an event after 7.7 seconds	210	Example: fire an event after 7.7 seconds.
		211
178	my $w = AnyEvent->timer (after => 7.7, cb => sub {	212	my $w = AnyEvent->timer (after => 7.7, cb => sub {
179	warn "timeout\n";	213	warn "timeout\n";
180	});	214	});
181		215
182	# to cancel the timer:	216	# to cancel the timer:
183	undef $w;	217	undef $w;
184		218
185	Example 2:
186
187	# fire an event after 0.5 seconds, then roughly every second	219	Example 2: fire an event after 0.5 seconds, then roughly every second.
188	my $w;
189		220
190	my $cb = sub {
191	# cancel the old timer while creating a new one
192	$w = AnyEvent->timer (after => 1, cb => $cb);	221	my $w = AnyEvent->timer (after => 0.5, interval => 1, cb => sub {
		222	warn "timeout\n";
193	};	223	};
194
195	# start the "loop" by creating the first watcher
196	$w = AnyEvent->timer (after => 0.5, cb => $cb);
197		224
198	TIMING ISSUES	225	TIMING ISSUES
199	There are two ways to handle timers: based on real time (relative, "fire	226	There are two ways to handle timers: based on real time (relative, "fire
200	in 10 seconds") and based on wallclock time (absolute, "fire at 12	227	in 10 seconds") and based on wallclock time (absolute, "fire at 12
201	o'clock").	228	o'clock").
…		…
213	on wallclock time) timers.	240	on wallclock time) timers.
214		241
215	AnyEvent always prefers relative timers, if available, matching the	242	AnyEvent always prefers relative timers, if available, matching the
216	AnyEvent API.	243	AnyEvent API.
217		244
		245	AnyEvent has two additional methods that return the "current time":
		246
		247	AnyEvent->time
		248	This returns the "current wallclock time" as a fractional number of
		249	seconds since the Epoch (the same thing as "time" or
		250	"Time::HiRes::time" return, and the result is guaranteed to be
		251	compatible with those).
		252
		253	It progresses independently of any event loop processing, i.e. each
		254	call will check the system clock, which usually gets updated
		255	frequently.
		256
		257	AnyEvent->now
		258	This also returns the "current wallclock time", but unlike "time",
		259	above, this value might change only once per event loop iteration,
		260	depending on the event loop (most return the same time as "time",
		261	above). This is the time that AnyEvent's timers get scheduled
		262	against.
		263
		264	*In almost all cases (in all cases if you don't care), this is the
		265	function to call when you want to know the current time.*
		266
		267	This function is also often faster then "AnyEvent->time", and thus
		268	the preferred method if you want some timestamp (for example,
		269	AnyEvent::Handle uses this to update it's activity timeouts).
		270
		271	The rest of this section is only of relevance if you try to be very
		272	exact with your timing, you can skip it without bad conscience.
		273
		274	For a practical example of when these times differ, consider
		275	Event::Lib and EV and the following set-up:
		276
		277	The event loop is running and has just invoked one of your callback
		278	at time=500 (assume no other callbacks delay processing). In your
		279	callback, you wait a second by executing "sleep 1" (blocking the
		280	process for a second) and then (at time=501) you create a relative
		281	timer that fires after three seconds.
		282
		283	With Event::Lib, "AnyEvent->time" and "AnyEvent->now" will both
		284	return 501, because that is the current time, and the timer will be
		285	scheduled to fire at time=504 (501 + 3).
		286
		287	With EV, "AnyEvent->time" returns 501 (as that is the current time),
		288	but "AnyEvent->now" returns 500, as that is the time the last event
		289	processing phase started. With EV, your timer gets scheduled to run
		290	at time=503 (500 + 3).
		291
		292	In one sense, Event::Lib is more exact, as it uses the current time
		293	regardless of any delays introduced by event processing. However,
		294	most callbacks do not expect large delays in processing, so this
		295	causes a higher drift (and a lot more system calls to get the
		296	current time).
		297
		298	In another sense, EV is more exact, as your timer will be scheduled
		299	at the same time, regardless of how long event processing actually
		300	took.
		301
		302	In either case, if you care (and in most cases, you don't), then you
		303	can get whatever behaviour you want with any event loop, by taking
		304	the difference between "AnyEvent->time" and "AnyEvent->now" into
		305	account.
		306
218	SIGNAL WATCHERS	307	SIGNAL WATCHERS
219	You can watch for signals using a signal watcher, "signal" is the signal	308	You can watch for signals using a signal watcher, "signal" is the signal
220	name without any "SIG" prefix, "cb" is the Perl callback to be invoked	309	name in uppercase and without any "SIG" prefix, "cb" is the Perl
221	whenever a signal occurs.	310	callback to be invoked whenever a signal occurs.
222		311
		312	Although the callback might get passed parameters, their value and
		313	presence is undefined and you cannot rely on them. Portable AnyEvent
		314	callbacks cannot use arguments passed to signal watcher callbacks.
		315
223	Multiple signal occurances can be clumped together into one callback	316	Multiple signal occurrences can be clumped together into one callback
224	invocation, and callback invocation will be synchronous. synchronous	317	invocation, and callback invocation will be synchronous. Synchronous
225	means that it might take a while until the signal gets handled by the	318	means that it might take a while until the signal gets handled by the
226	process, but it is guarenteed not to interrupt any other callbacks.	319	process, but it is guaranteed not to interrupt any other callbacks.
227		320
228	The main advantage of using these watchers is that you can share a	321	The main advantage of using these watchers is that you can share a
229	signal between multiple watchers.	322	signal between multiple watchers.
230		323
231	This watcher might use %SIG, so programs overwriting those signals	324	This watcher might use %SIG, so programs overwriting those signals
…		…
240		333
241	The child process is specified by the "pid" argument (if set to 0, it	334	The child process is specified by the "pid" argument (if set to 0, it
242	watches for any child process exit). The watcher will trigger as often	335	watches for any child process exit). The watcher will trigger as often
243	as status change for the child are received. This works by installing a	336	as status change for the child are received. This works by installing a
244	signal handler for "SIGCHLD". The callback will be called with the pid	337	signal handler for "SIGCHLD". The callback will be called with the pid
245	and exit status (as returned by waitpid).	338	and exit status (as returned by waitpid), so unlike other watcher types,
		339	you can rely on child watcher callback arguments.
246		340
247	Example: wait for pid 1333	341	There is a slight catch to child watchers, however: you usually start
		342	them after the child process was created, and this means the process
		343	could have exited already (and no SIGCHLD will be sent anymore).
248		344
		345	Not all event models handle this correctly (POE doesn't), but even for
		346	event models that do handle this correctly, they usually need to be
		347	loaded before the process exits (i.e. before you fork in the first
		348	place).
		349
		350	This means you cannot create a child watcher as the very first thing in
		351	an AnyEvent program, you have to create at least one watcher before
		352	you "fork" the child (alternatively, you can call "AnyEvent::detect").
		353
		354	Example: fork a process and wait for it
		355
		356	my $done = AnyEvent->condvar;
		357
		358	my $pid = fork or exit 5;
		359
249	my $w = AnyEvent->child (	360	my $w = AnyEvent->child (
250	pid => 1333,	361	pid => $pid,
251	cb => sub {	362	cb => sub {
252	my ($pid, $status) = @_;	363	my ($pid, $status) = @_;
253	warn "pid $pid exited with status $status";	364	warn "pid $pid exited with status $status";
		365	$done->send;
254	},	366	},
255	);	367	);
		368
		369	# do something else, then wait for process exit
		370	$done->recv;
256		371
257	CONDITION VARIABLES	372	CONDITION VARIABLES
		373	If you are familiar with some event loops you will know that all of them
		374	require you to run some blocking "loop", "run" or similar function that
		375	will actively watch for new events and call your callbacks.
		376
		377	AnyEvent is different, it expects somebody else to run the event loop
		378	and will only block when necessary (usually when told by the user).
		379
		380	The instrument to do that is called a "condition variable", so called
		381	because they represent a condition that must become true.
		382
258	Condition variables can be created by calling the "AnyEvent->condvar"	383	Condition variables can be created by calling the "AnyEvent->condvar"
259	method without any arguments.	384	method, usually without arguments. The only argument pair allowed is
260		385
261	A condition variable waits for a condition - precisely that the	386	"cb", which specifies a callback to be called when the condition
262	"->broadcast" method has been called.	387	variable becomes true, with the condition variable as the first argument
		388	(but not the results).
263		389
264	They are very useful to signal that a condition has been fulfilled, for	390	After creation, the condition variable is "false" until it becomes
		391	"true" by calling the "send" method (or calling the condition variable
		392	as if it were a callback, read about the caveats in the description for
		393	the "->send" method).
		394
		395	Condition variables are similar to callbacks, except that you can
		396	optionally wait for them. They can also be called merge points - points
		397	in time where multiple outstanding events have been processed. And yet
		398	another way to call them is transactions - each condition variable can
		399	be used to represent a transaction, which finishes at some point and
		400	delivers a result.
		401
		402	Condition variables are very useful to signal that something has
265	example, if you write a module that does asynchronous http requests,	403	finished, for example, if you write a module that does asynchronous http
266	then a condition variable would be the ideal candidate to signal the	404	requests, then a condition variable would be the ideal candidate to
267	availability of results.	405	signal the availability of results. The user can either act when the
		406	callback is called or can synchronously "->recv" for the results.
268		407
269	You can also use condition variables to block your main program until an	408	You can also use them to simulate traditional event loops - for example,
270	event occurs - for example, you could "->wait" in your main program	409	you can block your main program until an event occurs - for example, you
271	until the user clicks the Quit button in your app, which would	410	could "->recv" in your main program until the user clicks the Quit
272	"->broadcast" the "quit" event.	411	button of your app, which would "->send" the "quit" event.
273		412
274	Note that condition variables recurse into the event loop - if you have	413	Note that condition variables recurse into the event loop - if you have
275	two pirces of code that call "->wait" in a round-robbin fashion, you	414	two pieces of code that call "->recv" in a round-robin fashion, you
276	lose. Therefore, condition variables are good to export to your caller,	415	lose. Therefore, condition variables are good to export to your caller,
277	but you should avoid making a blocking wait yourself, at least in	416	but you should avoid making a blocking wait yourself, at least in
278	callbacks, as this asks for trouble.	417	callbacks, as this asks for trouble.
279		418
280	This object has two methods:	419	Condition variables are represented by hash refs in perl, and the keys
		420	used by AnyEvent itself are all named "_ae_XXX" to make subclassing easy
		421	(it is often useful to build your own transaction class on top of
		422	AnyEvent). To subclass, use "AnyEvent::CondVar" as base class and call
		423	it's "new" method in your own "new" method.
281		424
282	$cv->wait	425	There are two "sides" to a condition variable - the "producer side"
		426	which eventually calls "-> send", and the "consumer side", which waits
		427	for the send to occur.
		428
		429	Example: wait for a timer.
		430
		431	# wait till the result is ready
		432	my $result_ready = AnyEvent->condvar;
		433
		434	# do something such as adding a timer
		435	# or socket watcher the calls $result_ready->send
		436	# when the "result" is ready.
		437	# in this case, we simply use a timer:
		438	my $w = AnyEvent->timer (
		439	after => 1,
		440	cb => sub { $result_ready->send },
		441	);
		442
		443	# this "blocks" (while handling events) till the callback
		444	# calls send
		445	$result_ready->recv;
		446
		447	Example: wait for a timer, but take advantage of the fact that condition
		448	variables are also code references.
		449
		450	my $done = AnyEvent->condvar;
		451	my $delay = AnyEvent->timer (after => 5, cb => $done);
		452	$done->recv;
		453
		454	Example: Imagine an API that returns a condvar and doesn't support
		455	callbacks. This is how you make a synchronous call, for example from the
		456	main program:
		457
		458	use AnyEvent::CouchDB;
		459
		460	...
		461
		462	my @info = $couchdb->info->recv;
		463
		464	And this is how you would just ste a callback to be called whenever the
		465	results are available:
		466
		467	$couchdb->info->cb (sub {
		468	my @info = $_[0]->recv;
		469	});
		470
		471	METHODS FOR PRODUCERS
		472	These methods should only be used by the producing side, i.e. the
		473	code/module that eventually sends the signal. Note that it is also the
		474	producer side which creates the condvar in most cases, but it isn't
		475	uncommon for the consumer to create it as well.
		476
		477	$cv->send (...)
		478	Flag the condition as ready - a running "->recv" and all further
		479	calls to "recv" will (eventually) return after this method has been
		480	called. If nobody is waiting the send will be remembered.
		481
		482	If a callback has been set on the condition variable, it is called
		483	immediately from within send.
		484
		485	Any arguments passed to the "send" call will be returned by all
		486	future "->recv" calls.
		487
		488	Condition variables are overloaded so one can call them directly (as
		489	a code reference). Calling them directly is the same as calling
		490	"send". Note, however, that many C-based event loops do not handle
		491	overloading, so as tempting as it may be, passing a condition
		492	variable instead of a callback does not work. Both the pure perl and
		493	EV loops support overloading, however, as well as all functions that
		494	use perl to invoke a callback (as in AnyEvent::Socket and
		495	AnyEvent::DNS for example).
		496
		497	$cv->croak ($error)
		498	Similar to send, but causes all call's to "->recv" to invoke
		499	"Carp::croak" with the given error message/object/scalar.
		500
		501	This can be used to signal any errors to the condition variable
		502	user/consumer.
		503
		504	$cv->begin ([group callback])
		505	$cv->end
		506	These two methods are EXPERIMENTAL and MIGHT CHANGE.
		507
		508	These two methods can be used to combine many transactions/events
		509	into one. For example, a function that pings many hosts in parallel
		510	might want to use a condition variable for the whole process.
		511
		512	Every call to "->begin" will increment a counter, and every call to
		513	"->end" will decrement it. If the counter reaches 0 in "->end", the
		514	(last) callback passed to "begin" will be executed. That callback is
		515	supposed to call "->send", but that is not required. If no
		516	callback was set, "send" will be called without any arguments.
		517
		518	Let's clarify this with the ping example:
		519
		520	my $cv = AnyEvent->condvar;
		521
		522	my %result;
		523	$cv->begin (sub { $cv->send (\%result) });
		524
		525	for my $host (@list_of_hosts) {
		526	$cv->begin;
		527	ping_host_then_call_callback $host, sub {
		528	$result{$host} = ...;
		529	$cv->end;
		530	};
		531	}
		532
		533	$cv->end;
		534
		535	This code fragment supposedly pings a number of hosts and calls
		536	"send" after results for all then have have been gathered - in any
		537	order. To achieve this, the code issues a call to "begin" when it
		538	starts each ping request and calls "end" when it has received some
		539	result for it. Since "begin" and "end" only maintain a counter, the
		540	order in which results arrive is not relevant.
		541
		542	There is an additional bracketing call to "begin" and "end" outside
		543	the loop, which serves two important purposes: first, it sets the
		544	callback to be called once the counter reaches 0, and second, it
		545	ensures that "send" is called even when "no" hosts are being pinged
		546	(the loop doesn't execute once).
		547
		548	This is the general pattern when you "fan out" into multiple
		549	subrequests: use an outer "begin"/"end" pair to set the callback and
		550	ensure "end" is called at least once, and then, for each subrequest
		551	you start, call "begin" and for each subrequest you finish, call
		552	"end".
		553
		554	METHODS FOR CONSUMERS
		555	These methods should only be used by the consuming side, i.e. the code
		556	awaits the condition.
		557
		558	$cv->recv
283	Wait (blocking if necessary) until the "->broadcast" method has been	559	Wait (blocking if necessary) until the "->send" or "->croak" methods
284	called on c<$cv>, while servicing other watchers normally.	560	have been called on c<$cv>, while servicing other watchers normally.
285		561
286	You can only wait once on a condition - additional calls will return	562	You can only wait once on a condition - additional calls are valid
287	immediately.	563	but will return immediately.
		564
		565	If an error condition has been set by calling "->croak", then this
		566	function will call "croak".
		567
		568	In list context, all parameters passed to "send" will be returned,
		569	in scalar context only the first one will be returned.
288		570
289	Not all event models support a blocking wait - some die in that case	571	Not all event models support a blocking wait - some die in that case
290	(programs might want to do that to stay interactive), so *if you are	572	(programs might want to do that to stay interactive), so *if you are
291	using this from a module, never require a blocking wait*, but let	573	using this from a module, never require a blocking wait*, but let
292	the caller decide whether the call will block or not (for example,	574	the caller decide whether the call will block or not (for example,
293	by coupling condition variables with some kind of request results	575	by coupling condition variables with some kind of request results
294	and supporting callbacks so the caller knows that getting the result	576	and supporting callbacks so the caller knows that getting the result
295	will not block, while still suppporting blocking waits if the caller	577	will not block, while still supporting blocking waits if the caller
296	so desires).	578	so desires).
297		579
298	Another reason never to "->wait" in a module is that you cannot	580	Another reason never to "->recv" in a module is that you cannot
299	sensibly have two "->wait"'s in parallel, as that would require	581	sensibly have two "->recv"'s in parallel, as that would require
300	multiple interpreters or coroutines/threads, none of which	582	multiple interpreters or coroutines/threads, none of which
301	"AnyEvent" can supply (the coroutine-aware backends	583	"AnyEvent" can supply.
302	AnyEvent::Impl::CoroEV and AnyEvent::Impl::CoroEvent explicitly
303	support concurrent "->wait"'s from different coroutines, however).
304		584
305	$cv->broadcast	585	The Coro module, however, can and does supply coroutines and, in
306	Flag the condition as ready - a running "->wait" and all further	586	fact, Coro::AnyEvent replaces AnyEvent's condvars by coroutine-safe
307	calls to "wait" will (eventually) return after this method has been	587	versions and also integrates coroutines into AnyEvent, making
308	called. If nobody is waiting the broadcast will be remembered..	588	blocking "->recv" calls perfectly safe as long as they are done from
		589	another coroutine (one that doesn't run the event loop).
309		590
310	Example:	591	You can ensure that "-recv" never blocks by setting a callback and
		592	only calling "->recv" from within that callback (or at a later
		593	time). This will work even when the event loop does not support
		594	blocking waits otherwise.
311		595
312	# wait till the result is ready	596	$bool = $cv->ready
313	my $result_ready = AnyEvent->condvar;	597	Returns true when the condition is "true", i.e. whether "send" or
		598	"croak" have been called.
314		599
315	# do something such as adding a timer	600	$cb = $cv->cb ($cb->($cv))
316	# or socket watcher the calls $result_ready->broadcast	601	This is a mutator function that returns the callback set and
317	# when the "result" is ready.	602	optionally replaces it before doing so.
318	# in this case, we simply use a timer:
319	my $w = AnyEvent->timer (
320	after => 1,
321	cb => sub { $result_ready->broadcast },
322	);
323		603
324	# this "blocks" (while handling events) till the watcher	604	The callback will be called when the condition becomes "true", i.e.
325	# calls broadcast	605	when "send" or "croak" are called, with the only argument being the
326	$result_ready->wait;	606	condition variable itself. Calling "recv" inside the callback or at
		607	any later time is guaranteed not to block.
327		608
328	GLOBAL VARIABLES AND FUNCTIONS	609	GLOBAL VARIABLES AND FUNCTIONS
329	$AnyEvent::MODEL	610	$AnyEvent::MODEL
330	Contains "undef" until the first watcher is being created. Then it	611	Contains "undef" until the first watcher is being created. Then it
331	contains the event model that is being used, which is the name of	612	contains the event model that is being used, which is the name of
…		…
333	the "AnyEvent::Impl:xxx" modules, but can be any other class in the	614	the "AnyEvent::Impl:xxx" modules, but can be any other class in the
334	case AnyEvent has been extended at runtime (e.g. in rxvt-unicode).	615	case AnyEvent has been extended at runtime (e.g. in rxvt-unicode).
335		616
336	The known classes so far are:	617	The known classes so far are:
337		618
338	AnyEvent::Impl::CoroEV based on Coro::EV, best choice.
339	AnyEvent::Impl::CoroEvent based on Coro::Event, second best choice.
340	AnyEvent::Impl::EV based on EV (an interface to libev, best choice).	619	AnyEvent::Impl::EV based on EV (an interface to libev, best choice).
341	AnyEvent::Impl::Event based on Event, second best choice.	620	AnyEvent::Impl::Event based on Event, second best choice.
		621	AnyEvent::Impl::Perl pure-perl implementation, fast and portable.
342	AnyEvent::Impl::Glib based on Glib, third-best choice.	622	AnyEvent::Impl::Glib based on Glib, third-best choice.
343	AnyEvent::Impl::Tk based on Tk, very bad choice.	623	AnyEvent::Impl::Tk based on Tk, very bad choice.
344	AnyEvent::Impl::Perl pure-perl implementation, inefficient but portable.
345	AnyEvent::Impl::Qt based on Qt, cannot be autoprobed (see its docs).	624	AnyEvent::Impl::Qt based on Qt, cannot be autoprobed (see its docs).
346	AnyEvent::Impl::EventLib based on Event::Lib, leaks memory and worse.	625	AnyEvent::Impl::EventLib based on Event::Lib, leaks memory and worse.
		626	AnyEvent::Impl::POE based on POE, not generic enough for full support.
		627
		628	There is no support for WxWidgets, as WxWidgets has no support for
		629	watching file handles. However, you can use WxWidgets through the
		630	POE Adaptor, as POE has a Wx backend that simply polls 20 times per
		631	second, which was considered to be too horrible to even consider for
		632	AnyEvent. Likewise, other POE backends can be used by AnyEvent by
		633	using it's adaptor.
		634
		635	AnyEvent knows about Prima and Wx and will try to use POE when
		636	autodetecting them.
347		637
348	AnyEvent::detect	638	AnyEvent::detect
349	Returns $AnyEvent::MODEL, forcing autodetection of the event model	639	Returns $AnyEvent::MODEL, forcing autodetection of the event model
350	if necessary. You should only call this function right before you	640	if necessary. You should only call this function right before you
351	would have created an AnyEvent watcher anyway, that is, as late as	641	would have created an AnyEvent watcher anyway, that is, as late as
352	possible at runtime.	642	possible at runtime.
353		643
		644	$guard = AnyEvent::post_detect { BLOCK }
		645	Arranges for the code block to be executed as soon as the event
		646	model is autodetected (or immediately if this has already happened).
		647
		648	If called in scalar or list context, then it creates and returns an
		649	object that automatically removes the callback again when it is
		650	destroyed. See Coro::BDB for a case where this is useful.
		651
		652	@AnyEvent::post_detect
		653	If there are any code references in this array (you can "push" to it
		654	before or after loading AnyEvent), then they will called directly
		655	after the event loop has been chosen.
		656
		657	You should check $AnyEvent::MODEL before adding to this array,
		658	though: if it contains a true value then the event loop has already
		659	been detected, and the array will be ignored.
		660
		661	Best use "AnyEvent::post_detect { BLOCK }" instead.
		662
354	WHAT TO DO IN A MODULE	663	WHAT TO DO IN A MODULE
355	As a module author, you should "use AnyEvent" and call AnyEvent methods	664	As a module author, you should "use AnyEvent" and call AnyEvent methods
356	freely, but you should not load a specific event module or rely on it.	665	freely, but you should not load a specific event module or rely on it.
357		666
358	Be careful when you create watchers in the module body - AnyEvent will	667	Be careful when you create watchers in the module body - AnyEvent will
359	decide which event module to use as soon as the first method is called,	668	decide which event module to use as soon as the first method is called,
360	so by calling AnyEvent in your module body you force the user of your	669	so by calling AnyEvent in your module body you force the user of your
361	module to load the event module first.	670	module to load the event module first.
362		671
363	Never call "->wait" on a condition variable unless you know that the	672	Never call "->recv" on a condition variable unless you know that the
364	"->broadcast" method has been called on it already. This is because it	673	"->send" method has been called on it already. This is because it will
365	will stall the whole program, and the whole point of using events is to	674	stall the whole program, and the whole point of using events is to stay
366	stay interactive.	675	interactive.
367		676
368	It is fine, however, to call "->wait" when the user of your module	677	It is fine, however, to call "->recv" when the user of your module
369	requests it (i.e. if you create a http request object ad have a method	678	requests it (i.e. if you create a http request object ad have a method
370	called "results" that returns the results, it should call "->wait"	679	called "results" that returns the results, it should call "->recv"
371	freely, as the user of your module knows what she is doing. always).	680	freely, as the user of your module knows what she is doing. always).
372		681
373	WHAT TO DO IN THE MAIN PROGRAM	682	WHAT TO DO IN THE MAIN PROGRAM
374	There will always be a single main program - the only place that should	683	There will always be a single main program - the only place that should
375	dictate which event model to use.	684	dictate which event model to use.
…		…
377	If it doesn't care, it can just "use AnyEvent" and use it itself, or not	686	If it doesn't care, it can just "use AnyEvent" and use it itself, or not
378	do anything special (it does not need to be event-based) and let	687	do anything special (it does not need to be event-based) and let
379	AnyEvent decide which implementation to chose if some module relies on	688	AnyEvent decide which implementation to chose if some module relies on
380	it.	689	it.
381		690
382	If the main program relies on a specific event model. For example, in	691	If the main program relies on a specific event model - for example, in
383	Gtk2 programs you have to rely on the Glib module. You should load the	692	Gtk2 programs you have to rely on the Glib module - you should load the
384	event module before loading AnyEvent or any module that uses it:	693	event module before loading AnyEvent or any module that uses it:
385	generally speaking, you should load it as early as possible. The reason	694	generally speaking, you should load it as early as possible. The reason
386	is that modules might create watchers when they are loaded, and AnyEvent	695	is that modules might create watchers when they are loaded, and AnyEvent
387	will decide on the event model to use as soon as it creates watchers,	696	will decide on the event model to use as soon as it creates watchers,
388	and it might chose the wrong one unless you load the correct one	697	and it might chose the wrong one unless you load the correct one
389	yourself.	698	yourself.
390		699
391	You can chose to use a rather inefficient pure-perl implementation by	700	You can chose to use a pure-perl implementation by loading the
392	loading the "AnyEvent::Impl::Perl" module, which gives you similar	701	"AnyEvent::Impl::Perl" module, which gives you similar behaviour
393	behaviour everywhere, but letting AnyEvent chose is generally better.	702	everywhere, but letting AnyEvent chose the model is generally better.
		703
		704	MAINLOOP EMULATION
		705	Sometimes (often for short test scripts, or even standalone programs who
		706	only want to use AnyEvent), you do not want to run a specific event
		707	loop.
		708
		709	In that case, you can use a condition variable like this:
		710
		711	AnyEvent->condvar->recv;
		712
		713	This has the effect of entering the event loop and looping forever.
		714
		715	Note that usually your program has some exit condition, in which case it
		716	is better to use the "traditional" approach of storing a condition
		717	variable somewhere, waiting for it, and sending it when the program
		718	should exit cleanly.
		719
		720	OTHER MODULES
		721	The following is a non-exhaustive list of additional modules that use
		722	AnyEvent and can therefore be mixed easily with other AnyEvent modules
		723	in the same program. Some of the modules come with AnyEvent, some are
		724	available via CPAN.
		725
		726	AnyEvent::Util
		727	Contains various utility functions that replace often-used but
		728	blocking functions such as "inet_aton" by event-/callback-based
		729	versions.
		730
		731	AnyEvent::Socket
		732	Provides various utility functions for (internet protocol) sockets,
		733	addresses and name resolution. Also functions to create non-blocking
		734	tcp connections or tcp servers, with IPv6 and SRV record support and
		735	more.
		736
		737	AnyEvent::Handle
		738	Provide read and write buffers, manages watchers for reads and
		739	writes, supports raw and formatted I/O, I/O queued and fully
		740	transparent and non-blocking SSL/TLS.
		741
		742	AnyEvent::DNS
		743	Provides rich asynchronous DNS resolver capabilities.
		744
		745	AnyEvent::HTTP
		746	A simple-to-use HTTP library that is capable of making a lot of
		747	concurrent HTTP requests.
		748
		749	AnyEvent::HTTPD
		750	Provides a simple web application server framework.
		751
		752	AnyEvent::FastPing
		753	The fastest ping in the west.
		754
		755	AnyEvent::DBI
		756	Executes DBI requests asynchronously in a proxy process.
		757
		758	AnyEvent::AIO
		759	Truly asynchronous I/O, should be in the toolbox of every event
		760	programmer. AnyEvent::AIO transparently fuses IO::AIO and AnyEvent
		761	together.
		762
		763	AnyEvent::BDB
		764	Truly asynchronous Berkeley DB access. AnyEvent::BDB transparently
		765	fuses BDB and AnyEvent together.
		766
		767	AnyEvent::GPSD
		768	A non-blocking interface to gpsd, a daemon delivering GPS
		769	information.
		770
		771	AnyEvent::IGS
		772	A non-blocking interface to the Internet Go Server protocol (used by
		773	App::IGS).
		774
		775	Net::IRC3
		776	AnyEvent based IRC client module family.
		777
		778	Net::XMPP2
		779	AnyEvent based XMPP (Jabber protocol) module family.
		780
		781	Net::FCP
		782	AnyEvent-based implementation of the Freenet Client Protocol,
		783	birthplace of AnyEvent.
		784
		785	Event::ExecFlow
		786	High level API for event-based execution flow control.
		787
		788	Coro
		789	Has special support for AnyEvent via Coro::AnyEvent.
		790
		791	IO::Lambda
		792	The lambda approach to I/O - don't ask, look there. Can use
		793	AnyEvent.
394		794
395	SUPPLYING YOUR OWN EVENT MODEL INTERFACE	795	SUPPLYING YOUR OWN EVENT MODEL INTERFACE
396	This is an advanced topic that you do not normally need to use AnyEvent	796	This is an advanced topic that you do not normally need to use AnyEvent
397	in a module. This section is only of use to event loop authors who want	797	in a module. This section is only of use to event loop authors who want
398	to provide AnyEvent compatibility.	798	to provide AnyEvent compatibility.
…		…
437		837
438	ENVIRONMENT VARIABLES	838	ENVIRONMENT VARIABLES
439	The following environment variables are used by this module:	839	The following environment variables are used by this module:
440		840
441	"PERL_ANYEVENT_VERBOSE"	841	"PERL_ANYEVENT_VERBOSE"
		842	By default, AnyEvent will be completely silent except in fatal
		843	conditions. You can set this environment variable to make AnyEvent
		844	more talkative.
		845
		846	When set to 1 or higher, causes AnyEvent to warn about unexpected
		847	conditions, such as not being able to load the event model specified
		848	by "PERL_ANYEVENT_MODEL".
		849
442	When set to 2 or higher, cause AnyEvent to report to STDERR which	850	When set to 2 or higher, cause AnyEvent to report to STDERR which
443	event model it chooses.	851	event model it chooses.
444		852
		853	"PERL_ANYEVENT_STRICT"
		854	AnyEvent does not do much argument checking by default, as thorough
		855	argument checking is very costly. Setting this variable to a true
		856	value will cause AnyEvent to load "AnyEvent::Strict" and then to
		857	thoroughly check the arguments passed to most method calls. If it
		858	finds any problems it will croak.
		859
		860	In other words, enables "strict" mode.
		861
		862	Unlike "use strict" it is definitely recommended ot keep it off in
		863	production.
		864
445	"PERL_ANYEVENT_MODEL"	865	"PERL_ANYEVENT_MODEL"
446	This can be used to specify the event model to be used by AnyEvent,	866	This can be used to specify the event model to be used by AnyEvent,
447	before autodetection and -probing kicks in. It must be a string	867	before auto detection and -probing kicks in. It must be a string
448	consisting entirely of ASCII letters. The string "AnyEvent::Impl::"	868	consisting entirely of ASCII letters. The string "AnyEvent::Impl::"
449	gets prepended and the resulting module name is loaded and if the	869	gets prepended and the resulting module name is loaded and if the
450	load was successful, used as event model. If it fails to load	870	load was successful, used as event model. If it fails to load
451	AnyEvent will proceed with autodetection and -probing.	871	AnyEvent will proceed with auto detection and -probing.
452		872
453	This functionality might change in future versions.	873	This functionality might change in future versions.
454		874
455	For example, to force the pure perl model (AnyEvent::Impl::Perl) you	875	For example, to force the pure perl model (AnyEvent::Impl::Perl) you
456	could start your program like this:	876	could start your program like this:
457		877
458	PERL_ANYEVENT_MODEL=Perl perl ...	878	PERL_ANYEVENT_MODEL=Perl perl ...
		879
		880	"PERL_ANYEVENT_PROTOCOLS"
		881	Used by both AnyEvent::DNS and AnyEvent::Socket to determine
		882	preferences for IPv4 or IPv6. The default is unspecified (and might
		883	change, or be the result of auto probing).
		884
		885	Must be set to a comma-separated list of protocols or address
		886	families, current supported: "ipv4" and "ipv6". Only protocols
		887	mentioned will be used, and preference will be given to protocols
		888	mentioned earlier in the list.
		889
		890	This variable can effectively be used for denial-of-service attacks
		891	against local programs (e.g. when setuid), although the impact is
		892	likely small, as the program has to handle connection errors
		893	already-
		894
		895	Examples: "PERL_ANYEVENT_PROTOCOLS=ipv4,ipv6" - prefer IPv4 over
		896	IPv6, but support both and try to use both.
		897	"PERL_ANYEVENT_PROTOCOLS=ipv4" - only support IPv4, never try to
		898	resolve or contact IPv6 addresses.
		899	"PERL_ANYEVENT_PROTOCOLS=ipv6,ipv4" support either IPv4 or IPv6, but
		900	prefer IPv6 over IPv4.
		901
		902	"PERL_ANYEVENT_EDNS0"
		903	Used by AnyEvent::DNS to decide whether to use the EDNS0 extension
		904	for DNS. This extension is generally useful to reduce DNS traffic,
		905	but some (broken) firewalls drop such DNS packets, which is why it
		906	is off by default.
		907
		908	Setting this variable to 1 will cause AnyEvent::DNS to announce
		909	EDNS0 in its DNS requests.
		910
		911	"PERL_ANYEVENT_MAX_FORKS"
		912	The maximum number of child processes that
		913	"AnyEvent::Util::fork_call" will create in parallel.
459		914
460	EXAMPLE PROGRAM	915	EXAMPLE PROGRAM
461	The following program uses an IO watcher to read data from STDIN, a	916	The following program uses an I/O watcher to read data from STDIN, a
462	timer to display a message once per second, and a condition variable to	917	timer to display a message once per second, and a condition variable to
463	quit the program when the user enters quit:	918	quit the program when the user enters quit:
464		919
465	use AnyEvent;	920	use AnyEvent;
466		921
…		…
471	poll => 'r',	926	poll => 'r',
472	cb => sub {	927	cb => sub {
473	warn "io event <$_[0]>\n"; # will always output <r>	928	warn "io event <$_[0]>\n"; # will always output <r>
474	chomp (my $input = <STDIN>); # read a line	929	chomp (my $input = <STDIN>); # read a line
475	warn "read: $input\n"; # output what has been read	930	warn "read: $input\n"; # output what has been read
476	$cv->broadcast if $input =~ /^q/i; # quit program if /^q/i	931	$cv->send if $input =~ /^q/i; # quit program if /^q/i
477	},	932	},
478	);	933	);
479		934
480	my $time_watcher; # can only be used once	935	my $time_watcher; # can only be used once
481		936
…		…
486	});	941	});
487	}	942	}
488		943
489	new_timer; # create first timer	944	new_timer; # create first timer
490		945
491	$cv->wait; # wait until user enters /^q/i	946	$cv->recv; # wait until user enters /^q/i
492		947
493	REAL-WORLD EXAMPLE	948	REAL-WORLD EXAMPLE
494	Consider the Net::FCP module. It features (among others) the following	949	Consider the Net::FCP module. It features (among others) the following
495	API calls, which are to freenet what HTTP GET requests are to http:	950	API calls, which are to freenet what HTTP GET requests are to http:
496		951
…		…
545	syswrite $txn->{fh}, $txn->{request}	1000	syswrite $txn->{fh}, $txn->{request}
546	or die "connection or write error";	1001	or die "connection or write error";
547	$txn->{w} = AnyEvent->io (fh => $txn->{fh}, poll => 'r', cb => sub { $txn->fh_ready_r });	1002	$txn->{w} = AnyEvent->io (fh => $txn->{fh}, poll => 'r', cb => sub { $txn->fh_ready_r });
548		1003
549	Again, "fh_ready_r" waits till all data has arrived, and then stores the	1004	Again, "fh_ready_r" waits till all data has arrived, and then stores the
550	result and signals any possible waiters that the request ahs finished:	1005	result and signals any possible waiters that the request has finished:
551		1006
552	sysread $txn->{fh}, $txn->{buf}, length $txn->{$buf};	1007	sysread $txn->{fh}, $txn->{buf}, length $txn->{$buf};
553		1008
554	if (end-of-file or data complete) {	1009	if (end-of-file or data complete) {
555	$txn->{result} = $txn->{buf};	1010	$txn->{result} = $txn->{buf};
556	$txn->{finished}->broadcast;	1011	$txn->{finished}->send;
557	$txb->{cb}->($txn) of $txn->{cb}; # also call callback	1012	$txb->{cb}->($txn) of $txn->{cb}; # also call callback
558	}	1013	}
559		1014
560	The "result" method, finally, just waits for the finished signal (if the	1015	The "result" method, finally, just waits for the finished signal (if the
561	request was already finished, it doesn't wait, of course, and returns	1016	request was already finished, it doesn't wait, of course, and returns
562	the data:	1017	the data:
563		1018
564	$txn->{finished}->wait;	1019	$txn->{finished}->recv;
565	return $txn->{result};	1020	return $txn->{result};
566		1021
567	The actual code goes further and collects all errors ("die"s,	1022	The actual code goes further and collects all errors ("die"s,
568	exceptions) that occured during request processing. The "result" method	1023	exceptions) that occurred during request processing. The "result" method
569	detects whether an exception as thrown (it is stored inside the $txn	1024	detects whether an exception as thrown (it is stored inside the $txn
570	object) and just throws the exception, which means connection errors and	1025	object) and just throws the exception, which means connection errors and
571	other problems get reported tot he code that tries to use the result,	1026	other problems get reported tot he code that tries to use the result,
572	not in a random callback.	1027	not in a random callback.
573		1028
…		…
604		1059
605	my $quit = AnyEvent->condvar;	1060	my $quit = AnyEvent->condvar;
606		1061
607	$fcp->txn_client_get ($url)->cb (sub {	1062	$fcp->txn_client_get ($url)->cb (sub {
608	...	1063	...
609	$quit->broadcast;	1064	$quit->send;
610	});	1065	});
611		1066
612	$quit->wait;	1067	$quit->recv;
		1068
		1069	BENCHMARKS
		1070	To give you an idea of the performance and overheads that AnyEvent adds
		1071	over the event loops themselves and to give you an impression of the
		1072	speed of various event loops I prepared some benchmarks.
		1073
		1074	BENCHMARKING ANYEVENT OVERHEAD
		1075	Here is a benchmark of various supported event models used natively and
		1076	through AnyEvent. The benchmark creates a lot of timers (with a zero
		1077	timeout) and I/O watchers (watching STDOUT, a pty, to become writable,
		1078	which it is), lets them fire exactly once and destroys them again.
		1079
		1080	Source code for this benchmark is found as eg/bench in the AnyEvent
		1081	distribution.
		1082
		1083	Explanation of the columns
		1084	watcher is the number of event watchers created/destroyed. Since
		1085	different event models feature vastly different performances, each event
		1086	loop was given a number of watchers so that overall runtime is
		1087	acceptable and similar between tested event loop (and keep them from
		1088	crashing): Glib would probably take thousands of years if asked to
		1089	process the same number of watchers as EV in this benchmark.
		1090
		1091	bytes is the number of bytes (as measured by the resident set size,
		1092	RSS) consumed by each watcher. This method of measuring captures both C
		1093	and Perl-based overheads.
		1094
		1095	create is the time, in microseconds (millionths of seconds), that it
		1096	takes to create a single watcher. The callback is a closure shared
		1097	between all watchers, to avoid adding memory overhead. That means
		1098	closure creation and memory usage is not included in the figures.
		1099
		1100	invoke is the time, in microseconds, used to invoke a simple callback.
		1101	The callback simply counts down a Perl variable and after it was invoked
		1102	"watcher" times, it would "->send" a condvar once to signal the end of
		1103	this phase.
		1104
		1105	destroy is the time, in microseconds, that it takes to destroy a
		1106	single watcher.
		1107
		1108	Results
		1109	name watchers bytes create invoke destroy comment
		1110	EV/EV 400000 244 0.56 0.46 0.31 EV native interface
		1111	EV/Any 100000 244 2.50 0.46 0.29 EV + AnyEvent watchers
		1112	CoroEV/Any 100000 244 2.49 0.44 0.29 coroutines + Coro::Signal
		1113	Perl/Any 100000 513 4.92 0.87 1.12 pure perl implementation
		1114	Event/Event 16000 516 31.88 31.30 0.85 Event native interface
		1115	Event/Any 16000 590 35.75 31.42 1.08 Event + AnyEvent watchers
		1116	Glib/Any 16000 1357 98.22 12.41 54.00 quadratic behaviour
		1117	Tk/Any 2000 1860 26.97 67.98 14.00 SEGV with >> 2000 watchers
		1118	POE/Event 2000 6644 108.64 736.02 14.73 via POE::Loop::Event
		1119	POE/Select 2000 6343 94.13 809.12 565.96 via POE::Loop::Select
		1120
		1121	Discussion
		1122	The benchmark does not measure scalability of the event loop very
		1123	well. For example, a select-based event loop (such as the pure perl one)
		1124	can never compete with an event loop that uses epoll when the number of
		1125	file descriptors grows high. In this benchmark, all events become ready
		1126	at the same time, so select/poll-based implementations get an unnatural
		1127	speed boost.
		1128
		1129	Also, note that the number of watchers usually has a nonlinear effect on
		1130	overall speed, that is, creating twice as many watchers doesn't take
		1131	twice the time - usually it takes longer. This puts event loops tested
		1132	with a higher number of watchers at a disadvantage.
		1133
		1134	To put the range of results into perspective, consider that on the
		1135	benchmark machine, handling an event takes roughly 1600 CPU cycles with
		1136	EV, 3100 CPU cycles with AnyEvent's pure perl loop and almost 3000000
		1137	CPU cycles with POE.
		1138
		1139	"EV" is the sole leader regarding speed and memory use, which are both
		1140	maximal/minimal, respectively. Even when going through AnyEvent, it uses
		1141	far less memory than any other event loop and is still faster than Event
		1142	natively.
		1143
		1144	The pure perl implementation is hit in a few sweet spots (both the
		1145	constant timeout and the use of a single fd hit optimisations in the
		1146	perl interpreter and the backend itself). Nevertheless this shows that
		1147	it adds very little overhead in itself. Like any select-based backend
		1148	its performance becomes really bad with lots of file descriptors (and
		1149	few of them active), of course, but this was not subject of this
		1150	benchmark.
		1151
		1152	The "Event" module has a relatively high setup and callback invocation
		1153	cost, but overall scores in on the third place.
		1154
		1155	"Glib"'s memory usage is quite a bit higher, but it features a faster
		1156	callback invocation and overall ends up in the same class as "Event".
		1157	However, Glib scales extremely badly, doubling the number of watchers
		1158	increases the processing time by more than a factor of four, making it
		1159	completely unusable when using larger numbers of watchers (note that
		1160	only a single file descriptor was used in the benchmark, so
		1161	inefficiencies of "poll" do not account for this).
		1162
		1163	The "Tk" adaptor works relatively well. The fact that it crashes with
		1164	more than 2000 watchers is a big setback, however, as correctness takes
		1165	precedence over speed. Nevertheless, its performance is surprising, as
		1166	the file descriptor is dup()ed for each watcher. This shows that the
		1167	dup() employed by some adaptors is not a big performance issue (it does
		1168	incur a hidden memory cost inside the kernel which is not reflected in
		1169	the figures above).
		1170
		1171	"POE", regardless of underlying event loop (whether using its pure perl
		1172	select-based backend or the Event module, the POE-EV backend couldn't be
		1173	tested because it wasn't working) shows abysmal performance and memory
		1174	usage with AnyEvent: Watchers use almost 30 times as much memory as EV
		1175	watchers, and 10 times as much memory as Event (the high memory
		1176	requirements are caused by requiring a session for each watcher).
		1177	Watcher invocation speed is almost 900 times slower than with AnyEvent's
		1178	pure perl implementation.
		1179
		1180	The design of the POE adaptor class in AnyEvent can not really account
		1181	for the performance issues, though, as session creation overhead is
		1182	small compared to execution of the state machine, which is coded pretty
		1183	optimally within AnyEvent::Impl::POE (and while everybody agrees that
		1184	using multiple sessions is not a good approach, especially regarding
		1185	memory usage, even the author of POE could not come up with a faster
		1186	design).
		1187
		1188	Summary
		1189	* Using EV through AnyEvent is faster than any other event loop (even
		1190	when used without AnyEvent), but most event loops have acceptable
		1191	performance with or without AnyEvent.
		1192
		1193	* The overhead AnyEvent adds is usually much smaller than the overhead
		1194	of the actual event loop, only with extremely fast event loops such
		1195	as EV adds AnyEvent significant overhead.
		1196
		1197	* You should avoid POE like the plague if you want performance or
		1198	reasonable memory usage.
		1199
		1200	BENCHMARKING THE LARGE SERVER CASE
		1201	This benchmark actually benchmarks the event loop itself. It works by
		1202	creating a number of "servers": each server consists of a socket pair, a
		1203	timeout watcher that gets reset on activity (but never fires), and an
		1204	I/O watcher waiting for input on one side of the socket. Each time the
		1205	socket watcher reads a byte it will write that byte to a random other
		1206	"server".
		1207
		1208	The effect is that there will be a lot of I/O watchers, only part of
		1209	which are active at any one point (so there is a constant number of
		1210	active fds for each loop iteration, but which fds these are is random).
		1211	The timeout is reset each time something is read because that reflects
		1212	how most timeouts work (and puts extra pressure on the event loops).
		1213
		1214	In this benchmark, we use 10000 socket pairs (20000 sockets), of which
		1215	100 (1%) are active. This mirrors the activity of large servers with
		1216	many connections, most of which are idle at any one point in time.
		1217
		1218	Source code for this benchmark is found as eg/bench2 in the AnyEvent
		1219	distribution.
		1220
		1221	Explanation of the columns
		1222	sockets is the number of sockets, and twice the number of "servers"
		1223	(as each server has a read and write socket end).
		1224
		1225	create is the time it takes to create a socket pair (which is
		1226	nontrivial) and two watchers: an I/O watcher and a timeout watcher.
		1227
		1228	request, the most important value, is the time it takes to handle a
		1229	single "request", that is, reading the token from the pipe and
		1230	forwarding it to another server. This includes deleting the old timeout
		1231	and creating a new one that moves the timeout into the future.
		1232
		1233	Results
		1234	name sockets create request
		1235	EV 20000 69.01 11.16
		1236	Perl 20000 73.32 35.87
		1237	Event 20000 212.62 257.32
		1238	Glib 20000 651.16 1896.30
		1239	POE 20000 349.67 12317.24 uses POE::Loop::Event
		1240
		1241	Discussion
		1242	This benchmark does measure scalability and overall performance of the
		1243	particular event loop.
		1244
		1245	EV is again fastest. Since it is using epoll on my system, the setup
		1246	time is relatively high, though.
		1247
		1248	Perl surprisingly comes second. It is much faster than the C-based event
		1249	loops Event and Glib.
		1250
		1251	Event suffers from high setup time as well (look at its code and you
		1252	will understand why). Callback invocation also has a high overhead
		1253	compared to the "$_->() for .."-style loop that the Perl event loop
		1254	uses. Event uses select or poll in basically all documented
		1255	configurations.
		1256
		1257	Glib is hit hard by its quadratic behaviour w.r.t. many watchers. It
		1258	clearly fails to perform with many filehandles or in busy servers.
		1259
		1260	POE is still completely out of the picture, taking over 1000 times as
		1261	long as EV, and over 100 times as long as the Perl implementation, even
		1262	though it uses a C-based event loop in this case.
		1263
		1264	Summary
		1265	* The pure perl implementation performs extremely well.
		1266
		1267	* Avoid Glib or POE in large projects where performance matters.
		1268
		1269	BENCHMARKING SMALL SERVERS
		1270	While event loops should scale (and select-based ones do not...) even to
		1271	large servers, most programs we (or I :) actually write have only a few
		1272	I/O watchers.
		1273
		1274	In this benchmark, I use the same benchmark program as in the large
		1275	server case, but it uses only eight "servers", of which three are active
		1276	at any one time. This should reflect performance for a small server
		1277	relatively well.
		1278
		1279	The columns are identical to the previous table.
		1280
		1281	Results
		1282	name sockets create request
		1283	EV 16 20.00 6.54
		1284	Perl 16 25.75 12.62
		1285	Event 16 81.27 35.86
		1286	Glib 16 32.63 15.48
		1287	POE 16 261.87 276.28 uses POE::Loop::Event
		1288
		1289	Discussion
		1290	The benchmark tries to test the performance of a typical small server.
		1291	While knowing how various event loops perform is interesting, keep in
		1292	mind that their overhead in this case is usually not as important, due
		1293	to the small absolute number of watchers (that is, you need efficiency
		1294	and speed most when you have lots of watchers, not when you only have a
		1295	few of them).
		1296
		1297	EV is again fastest.
		1298
		1299	Perl again comes second. It is noticeably faster than the C-based event
		1300	loops Event and Glib, although the difference is too small to really
		1301	matter.
		1302
		1303	POE also performs much better in this case, but is is still far behind
		1304	the others.
		1305
		1306	Summary
		1307	* C-based event loops perform very well with small number of watchers,
		1308	as the management overhead dominates.
613		1309
614	FORK	1310	FORK
615	Most event libraries are not fork-safe. The ones who are usually are	1311	Most event libraries are not fork-safe. The ones who are usually are
616	because they are so inefficient. Only EV is fully fork-aware.	1312	because they rely on inefficient but fork-safe "select" or "poll" calls.
		1313	Only EV is fully fork-aware.
617		1314
618	If you have to fork, you must either do so before creating your first	1315	If you have to fork, you must either do so before creating your first
619	watcher OR you must not use AnyEvent at all in the child.	1316	watcher OR you must not use AnyEvent at all in the child.
620		1317
621	SECURITY CONSIDERATIONS	1318	SECURITY CONSIDERATIONS
…		…
627	model than specified in the variable.	1324	model than specified in the variable.
628		1325
629	You can make AnyEvent completely ignore this variable by deleting it	1326	You can make AnyEvent completely ignore this variable by deleting it
630	before the first watcher gets created, e.g. with a "BEGIN" block:	1327	before the first watcher gets created, e.g. with a "BEGIN" block:
631		1328
632	BEGIN { delete $ENV{PERL_ANYEVENT_MODEL} }	1329	BEGIN { delete $ENV{PERL_ANYEVENT_MODEL} }
633		1330
634	use AnyEvent;	1331	use AnyEvent;
		1332
		1333	Similar considerations apply to $ENV{PERL_ANYEVENT_VERBOSE}, as that can
		1334	be used to probe what backend is used and gain other information (which
		1335	is probably even less useful to an attacker than PERL_ANYEVENT_MODEL),
		1336	and $ENV{PERL_ANYEGENT_STRICT}.
		1337
		1338	BUGS
		1339	Perl 5.8 has numerous memleaks that sometimes hit this module and are
		1340	hard to work around. If you suffer from memleaks, first upgrade to Perl
		1341	5.10 and check wether the leaks still show up. (Perl 5.10.0 has other
		1342	annoying mamleaks, such as leaking on "map" and "grep" but it is usually
		1343	not as pronounced).
635		1344
636	SEE ALSO	1345	SEE ALSO
637	Event modules: Coro::EV, EV, EV::Glib, Glib::EV, Coro::Event, Event,	1346	Utility functions: AnyEvent::Util.
638	Glib::Event, Glib, Coro, Tk, Event::Lib, Qt.
639		1347
640	Implementations: AnyEvent::Impl::CoroEV, AnyEvent::Impl::EV,	1348	Event modules: EV, EV::Glib, Glib::EV, Event, Glib::Event, Glib, Tk,
641	AnyEvent::Impl::CoroEvent, AnyEvent::Impl::Event, AnyEvent::Impl::Glib,	1349	Event::Lib, Qt, POE.
642	AnyEvent::Impl::Tk, AnyEvent::Impl::Perl, AnyEvent::Impl::EventLib,
643	AnyEvent::Impl::Qt.
644		1350
		1351	Implementations: AnyEvent::Impl::EV, AnyEvent::Impl::Event,
		1352	AnyEvent::Impl::Glib, AnyEvent::Impl::Tk, AnyEvent::Impl::Perl,
		1353	AnyEvent::Impl::EventLib, AnyEvent::Impl::Qt, AnyEvent::Impl::POE.
		1354
		1355	Non-blocking file handles, sockets, TCP clients and servers:
		1356	AnyEvent::Handle, AnyEvent::Socket.
		1357
		1358	Asynchronous DNS: AnyEvent::DNS.
		1359
		1360	Coroutine support: Coro, Coro::AnyEvent, Coro::EV, Coro::Event,
		1361
645	Nontrivial usage examples: Net::FCP, Net::XMPP2.	1362	Nontrivial usage examples: Net::FCP, Net::XMPP2, AnyEvent::DNS.
646		1363
647	AUTHOR	1364	AUTHOR
648	Marc Lehmann <schmorp@schmorp.de>	1365	Marc Lehmann <schmorp@schmorp.de>
649	http://home.schmorp.de/	1366	http://home.schmorp.de/
650		1367

Diff Legend

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

Comparing AnyEvent/README (file contents): Revision 1.18 by root, Thu Apr 24 09:13:26 2008 UTC vs. Revision 1.29 by root, Tue Jul 29 10:20:33 2008 UTC

Diff Legend

Comparing AnyEvent/README (file contents):
Revision 1.18 by root, Thu Apr 24 09:13:26 2008 UTC vs.
Revision 1.29 by root, Tue Jul 29 10:20:33 2008 UTC