1 |
=head1 NAME |
2 |
|
3 |
AnyEvent - provide framework for multiple event loops |
4 |
|
5 |
Event, Coro, Glib, Tk, Perl - various supported event loops |
6 |
|
7 |
=head1 SYNOPSIS |
8 |
|
9 |
use AnyEvent; |
10 |
|
11 |
my $w = AnyEvent->io (fh => $fh, poll => "r|w", cb => sub { |
12 |
... |
13 |
}); |
14 |
|
15 |
my $w = AnyEvent->timer (after => $seconds, cb => sub { |
16 |
... |
17 |
}); |
18 |
|
19 |
my $w = AnyEvent->condvar; # stores wether a condition was flagged |
20 |
$w->wait; # enters "main loop" till $condvar gets ->broadcast |
21 |
$w->broadcast; # wake up current and all future wait's |
22 |
|
23 |
=head1 DESCRIPTION |
24 |
|
25 |
L<AnyEvent> provides an identical interface to multiple event loops. This |
26 |
allows module authors to utilise an event loop without forcing module |
27 |
users to use the same event loop (as only a single event loop can coexist |
28 |
peacefully at any one time). |
29 |
|
30 |
The interface itself is vaguely similar but not identical to the Event |
31 |
module. |
32 |
|
33 |
On the first call of any method, the module tries to detect the currently |
34 |
loaded event loop by probing wether any of the following modules is |
35 |
loaded: L<Coro::Event>, L<Event>, L<Glib>, L<Tk>. The first one found is |
36 |
used. If none is found, the module tries to load these modules in the |
37 |
order given. The first one that could be successfully loaded will be |
38 |
used. If still none could be found, AnyEvent will fall back to a pure-perl |
39 |
event loop, which is also not very efficient. |
40 |
|
41 |
Because AnyEvent first checks for modules that are already loaded, loading |
42 |
an Event model explicitly before first using AnyEvent will likely make |
43 |
that model the default. For example: |
44 |
|
45 |
use Tk; |
46 |
use AnyEvent; |
47 |
|
48 |
# .. AnyEvent will likely default to Tk |
49 |
|
50 |
The pure-perl implementation of AnyEvent is called |
51 |
C<AnyEvent::Impl::Perl>. Like other event modules you can load it |
52 |
explicitly. |
53 |
|
54 |
=head1 WATCHERS |
55 |
|
56 |
AnyEvent has the central concept of a I<watcher>, which is an object that |
57 |
stores relevant data for each kind of event you are waiting for, such as |
58 |
the callback to call, the filehandle to watch, etc. |
59 |
|
60 |
These watchers are normal Perl objects with normal Perl lifetime. After |
61 |
creating a watcher it will immediately "watch" for events and invoke |
62 |
the callback. To disable the watcher you have to destroy it (e.g. by |
63 |
setting the variable that stores it to C<undef> or otherwise deleting all |
64 |
references to it). |
65 |
|
66 |
All watchers are created by calling a method on the C<AnyEvent> class. |
67 |
|
68 |
=head2 IO WATCHERS |
69 |
|
70 |
You can create I/O watcher by calling the C<< AnyEvent->io >> method with |
71 |
the following mandatory arguments: |
72 |
|
73 |
C<fh> the Perl I<filehandle> (not filedescriptor) to watch for |
74 |
events. C<poll> must be a string that is either C<r> or C<w>, that creates |
75 |
a watcher waiting for "r"eadable or "w"ritable events. C<cb> teh callback |
76 |
to invoke everytime the filehandle becomes ready. |
77 |
|
78 |
Only one io watcher per C<fh> and C<poll> combination is allowed (i.e. on |
79 |
a socket you can have one r + one w, not any more (limitation comes from |
80 |
Tk - if you are sure you are not using Tk this limitation is gone). |
81 |
|
82 |
Filehandles will be kept alive, so as long as the watcher exists, the |
83 |
filehandle exists, too. |
84 |
|
85 |
Example: |
86 |
|
87 |
# wait for readability of STDIN, then read a line and disable the watcher |
88 |
my $w; $w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub { |
89 |
chomp (my $input = <STDIN>); |
90 |
warn "read: $input\n"; |
91 |
undef $w; |
92 |
}); |
93 |
|
94 |
=head2 TIMER WATCHERS |
95 |
|
96 |
You can create a timer watcher by calling the C<< AnyEvent->timer >> |
97 |
method with the following mandatory arguments: |
98 |
|
99 |
C<after> after how many seconds (fractions are supported) should the timer |
100 |
activate. C<cb> the callback to invoke. |
101 |
|
102 |
The timer callback will be invoked at most once: if you want a repeating |
103 |
timer you have to create a new watcher (this is a limitation by both Tk |
104 |
and Glib). |
105 |
|
106 |
Example: |
107 |
|
108 |
# fire an event after 7.7 seconds |
109 |
my $w = AnyEvent->timer (after => 7.7, cb => sub { |
110 |
warn "timeout\n"; |
111 |
}); |
112 |
|
113 |
# to cancel the timer: |
114 |
undef $w |
115 |
|
116 |
=head2 CONDITION WATCHERS |
117 |
|
118 |
Condition watchers can be created by calling the C<< AnyEvent->condvar >> |
119 |
method without any arguments. |
120 |
|
121 |
A condition watcher watches for a condition - precisely that the C<< |
122 |
->broadcast >> method has been called. |
123 |
|
124 |
The watcher has only two methods: |
125 |
|
126 |
=over 4 |
127 |
|
128 |
=item $cv->wait |
129 |
|
130 |
Wait (blocking if necessary) until the C<< ->broadcast >> method has been |
131 |
called on c<$cv>, while servicing other watchers normally. |
132 |
|
133 |
Not all event models support a blocking wait - some die in that case, so |
134 |
if you are using this from a module, never require a blocking wait, but |
135 |
let the caller decide wether the call will block or not (for example, |
136 |
by coupling condition variables with some kind of request results and |
137 |
supporting callbacks so the caller knows that getting the result will not |
138 |
block, while still suppporting blockign waits if the caller so desires). |
139 |
|
140 |
You can only wait once on a condition - additional calls will return |
141 |
immediately. |
142 |
|
143 |
=item $cv->broadcast |
144 |
|
145 |
Flag the condition as ready - a running C<< ->wait >> and all further |
146 |
calls to C<wait> will return after this method has been called. If nobody |
147 |
is waiting the broadcast will be remembered.. |
148 |
|
149 |
Example: |
150 |
|
151 |
# wait till the result is ready |
152 |
my $result_ready = AnyEvent->condvar; |
153 |
|
154 |
# do something such as adding a timer |
155 |
# or socket watcher the calls $result_ready->broadcast |
156 |
# when the "result" is ready. |
157 |
|
158 |
$result_ready->wait; |
159 |
|
160 |
=back |
161 |
|
162 |
=head1 GLOBALS |
163 |
|
164 |
=over 4 |
165 |
|
166 |
=item $AnyEvent::MODEL |
167 |
|
168 |
Contains C<undef> until the first watcher is being created. Then it |
169 |
contains the event model that is being used, which is the name of the |
170 |
Perl class implementing the model. This class is usually one of the |
171 |
C<AnyEvent::Impl:xxx> modules, but can be any other class in the case |
172 |
AnyEvent has been extended at runtime (e.g. in I<rxvt-unicode>). |
173 |
|
174 |
The known classes so far are: |
175 |
|
176 |
AnyEvent::Impl::Coro based on Coro::Event, best choise. |
177 |
AnyEvent::Impl::Event based on Event, also best choice :) |
178 |
AnyEvent::Impl::Glib based on Glib, second-best choice. |
179 |
AnyEvent::Impl::Tk based on Tk, very bad choice. |
180 |
AnyEvent::Impl::Perl pure-perl implementation, inefficient. |
181 |
|
182 |
=back |
183 |
|
184 |
=head1 WHAT TO DO IN A MODULE |
185 |
|
186 |
As a module author, you should "use AnyEvent" and call AnyEvent methods |
187 |
freely, but you should not load a specific event module or rely on it. |
188 |
|
189 |
Be careful when you create watchers in the module body - Anyevent will |
190 |
decide which event module to use as soon as the first method is called, so |
191 |
by calling AnyEvent in your module body you force the user of your module |
192 |
to load the event module first. |
193 |
|
194 |
=head1 WHAT TO DO IN THE MAIN PROGRAM |
195 |
|
196 |
There will always be a single main program - the only place that should |
197 |
dictate which event model to use. |
198 |
|
199 |
If it doesn't care, it can just "use AnyEvent" and use it itself, or not |
200 |
do anything special and let AnyEvent decide which implementation to chose. |
201 |
|
202 |
If the main program relies on a specific event model (for example, in Gtk2 |
203 |
programs you have to rely on either Glib or Glib::Event), you should load |
204 |
it before loading AnyEvent or any module that uses it, generally, as early |
205 |
as possible. The reason is that modules might create watchers when they |
206 |
are loaded, and AnyEvent will decide on the event model to use as soon as |
207 |
it creates watchers, and it might chose the wrong one unless you load the |
208 |
correct one yourself. |
209 |
|
210 |
You can chose to use a rather inefficient pure-perl implementation by |
211 |
loading the C<AnyEvent::Impl::Perl> module, but letting AnyEvent chose is |
212 |
generally better. |
213 |
|
214 |
=cut |
215 |
|
216 |
package AnyEvent; |
217 |
|
218 |
no warnings; |
219 |
use strict 'vars'; |
220 |
use Carp; |
221 |
|
222 |
our $VERSION = '2.1'; |
223 |
our $MODEL; |
224 |
|
225 |
our $AUTOLOAD; |
226 |
our @ISA; |
227 |
|
228 |
our $verbose = $ENV{PERL_ANYEVENT_VERBOSE}*1; |
229 |
|
230 |
our @REGISTRY; |
231 |
|
232 |
my @models = ( |
233 |
[Coro::Event:: => AnyEvent::Impl::Coro::], |
234 |
[Event:: => AnyEvent::Impl::Event::], |
235 |
[Glib:: => AnyEvent::Impl::Glib::], |
236 |
[Tk:: => AnyEvent::Impl::Tk::], |
237 |
[AnyEvent::Impl::Perl:: => AnyEvent::Impl::Perl::], |
238 |
); |
239 |
|
240 |
our %method = map +($_ => 1), qw(io timer condvar broadcast wait DESTROY); |
241 |
|
242 |
sub AUTOLOAD { |
243 |
$AUTOLOAD =~ s/.*://; |
244 |
|
245 |
$method{$AUTOLOAD} |
246 |
or croak "$AUTOLOAD: not a valid method for AnyEvent objects"; |
247 |
|
248 |
unless ($MODEL) { |
249 |
# check for already loaded models |
250 |
for (@REGISTRY, @models) { |
251 |
my ($package, $model) = @$_; |
252 |
if (${"$package\::VERSION"} > 0) { |
253 |
if (eval "require $model") { |
254 |
$MODEL = $model; |
255 |
warn "AnyEvent: found model '$model', using it.\n" if $verbose > 1; |
256 |
last; |
257 |
} |
258 |
} |
259 |
} |
260 |
|
261 |
unless ($MODEL) { |
262 |
# try to load a model |
263 |
|
264 |
for (@REGISTRY, @models) { |
265 |
my ($package, $model) = @$_; |
266 |
if (eval "require $model") { |
267 |
$MODEL = $model; |
268 |
warn "AnyEvent: autoprobed and loaded model '$model', using it.\n" if $verbose > 1; |
269 |
last; |
270 |
} |
271 |
} |
272 |
|
273 |
$MODEL |
274 |
or die "No event module selected for AnyEvent and autodetect failed. Install any one of these modules: Coro, Event, Glib or Tk."; |
275 |
} |
276 |
} |
277 |
|
278 |
@ISA = $MODEL; |
279 |
|
280 |
my $class = shift; |
281 |
$class->$AUTOLOAD (@_); |
282 |
} |
283 |
|
284 |
=head1 SUPPLYING YOUR OWN EVENT MODEL INTERFACE |
285 |
|
286 |
If you need to support another event library which isn't directly |
287 |
supported by AnyEvent, you can supply your own interface to it by |
288 |
pushing, before the first watcher gets created, the package name of |
289 |
the event module and the package name of the interface to use onto |
290 |
C<@AnyEvent::REGISTRY>. You can do that before and even without loading |
291 |
AnyEvent. |
292 |
|
293 |
Example: |
294 |
|
295 |
push @AnyEvent::REGISTRY, [urxvt => urxvt::anyevent::]; |
296 |
|
297 |
This tells AnyEvent to (literally) use the C<urxvt::anyevent::> |
298 |
package/class when it finds the C<urxvt> package/module is loaded. When |
299 |
AnyEvent is loaded and asked to find a suitable event model, it will |
300 |
first check for the presence of urxvt. |
301 |
|
302 |
The class should prove implementations for all watcher types (see |
303 |
L<AnyEvent::Impl::Event> (source code), L<AnyEvent::Impl::Glib> |
304 |
(Source code) and so on for actual examples, use C<perldoc -m |
305 |
AnyEvent::Impl::Glib> to see the sources). |
306 |
|
307 |
The above isn't fictitious, the I<rxvt-unicode> (a.k.a. urxvt) |
308 |
uses the above line as-is. An interface isn't included in AnyEvent |
309 |
because it doesn't make sense outside the embedded interpreter inside |
310 |
I<rxvt-unicode>, and it is updated and maintained as part of the |
311 |
I<rxvt-unicode> distribution. |
312 |
|
313 |
I<rxvt-unicode> also cheats a bit by not providing blocking access to |
314 |
condition variables: code blocking while waiting for a condition will |
315 |
C<die>. This still works with most modules/usages, and blocking calls must |
316 |
not be in an interactive appliation, so it makes sense. |
317 |
|
318 |
=head1 ENVIRONMENT VARIABLES |
319 |
|
320 |
The following environment variables are used by this module: |
321 |
|
322 |
C<PERL_ANYEVENT_VERBOSE> when set to C<2> or higher, reports which event |
323 |
model gets used. |
324 |
|
325 |
=head1 EXAMPLE |
326 |
|
327 |
The following program uses an io watcher to read data from stdin, a timer |
328 |
to display a message once per second, and a condvar to exit the program |
329 |
when the user enters quit: |
330 |
|
331 |
use AnyEvent; |
332 |
|
333 |
my $cv = AnyEvent->condvar; |
334 |
|
335 |
my $io_watcher = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub { |
336 |
warn "io event <$_[0]>\n"; # will always output <r> |
337 |
chomp (my $input = <STDIN>); # read a line |
338 |
warn "read: $input\n"; # output what has been read |
339 |
$cv->broadcast if $input =~ /^q/i; # quit program if /^q/i |
340 |
}); |
341 |
|
342 |
my $time_watcher; # can only be used once |
343 |
|
344 |
sub new_timer { |
345 |
$timer = AnyEvent->timer (after => 1, cb => sub { |
346 |
warn "timeout\n"; # print 'timeout' about every second |
347 |
&new_timer; # and restart the time |
348 |
}); |
349 |
} |
350 |
|
351 |
new_timer; # create first timer |
352 |
|
353 |
$cv->wait; # wait until user enters /^q/i |
354 |
|
355 |
=head1 REAL-WORLD EXAMPLE |
356 |
|
357 |
Consider the L<Net::FCP> module. It features (among others) the following |
358 |
API calls, which are to freenet what HTTP GET requests are to http: |
359 |
|
360 |
my $data = $fcp->client_get ($url); # blocks |
361 |
|
362 |
my $transaction = $fcp->txn_client_get ($url); # does not block |
363 |
$transaction->cb ( sub { ... } ); # set optional result callback |
364 |
my $data = $transaction->result; # possibly blocks |
365 |
|
366 |
The C<client_get> method works like C<LWP::Simple::get>: it requests the |
367 |
given URL and waits till the data has arrived. It is defined to be: |
368 |
|
369 |
sub client_get { $_[0]->txn_client_get ($_[1])->result } |
370 |
|
371 |
And in fact is automatically generated. This is the blocking API of |
372 |
L<Net::FCP>, and it works as simple as in any other, similar, module. |
373 |
|
374 |
More complicated is C<txn_client_get>: It only creates a transaction |
375 |
(completion, result, ...) object and initiates the transaction. |
376 |
|
377 |
my $txn = bless { }, Net::FCP::Txn::; |
378 |
|
379 |
It also creates a condition variable that is used to signal the completion |
380 |
of the request: |
381 |
|
382 |
$txn->{finished} = AnyAvent->condvar; |
383 |
|
384 |
It then creates a socket in non-blocking mode. |
385 |
|
386 |
socket $txn->{fh}, ...; |
387 |
fcntl $txn->{fh}, F_SETFL, O_NONBLOCK; |
388 |
connect $txn->{fh}, ... |
389 |
and !$!{EWOULDBLOCK} |
390 |
and !$!{EINPROGRESS} |
391 |
and Carp::croak "unable to connect: $!\n"; |
392 |
|
393 |
Then it creates a write-watcher which gets called whenever an error occurs |
394 |
or the connection succeeds: |
395 |
|
396 |
$txn->{w} = AnyEvent->io (fh => $txn->{fh}, poll => 'w', cb => sub { $txn->fh_ready_w }); |
397 |
|
398 |
And returns this transaction object. The C<fh_ready_w> callback gets |
399 |
called as soon as the event loop detects that the socket is ready for |
400 |
writing. |
401 |
|
402 |
The C<fh_ready_w> method makes the socket blocking again, writes the |
403 |
request data and replaces the watcher by a read watcher (waiting for reply |
404 |
data). The actual code is more complicated, but that doesn't matter for |
405 |
this example: |
406 |
|
407 |
fcntl $txn->{fh}, F_SETFL, 0; |
408 |
syswrite $txn->{fh}, $txn->{request} |
409 |
or die "connection or write error"; |
410 |
$txn->{w} = AnyEvent->io (fh => $txn->{fh}, poll => 'r', cb => sub { $txn->fh_ready_r }); |
411 |
|
412 |
Again, C<fh_ready_r> waits till all data has arrived, and then stores the |
413 |
result and signals any possible waiters that the request ahs finished: |
414 |
|
415 |
sysread $txn->{fh}, $txn->{buf}, length $txn->{$buf}; |
416 |
|
417 |
if (end-of-file or data complete) { |
418 |
$txn->{result} = $txn->{buf}; |
419 |
$txn->{finished}->broadcast; |
420 |
$txb->{cb}->($txn) of $txn->{cb}; # also call callback |
421 |
} |
422 |
|
423 |
The C<result> method, finally, just waits for the finished signal (if the |
424 |
request was already finished, it doesn't wait, of course, and returns the |
425 |
data: |
426 |
|
427 |
$txn->{finished}->wait; |
428 |
return $txn->{result}; |
429 |
|
430 |
The actual code goes further and collects all errors (C<die>s, exceptions) |
431 |
that occured during request processing. The C<result> method detects |
432 |
wether an exception as thrown (it is stored inside the $txn object) |
433 |
and just throws the exception, which means connection errors and other |
434 |
problems get reported tot he code that tries to use the result, not in a |
435 |
random callback. |
436 |
|
437 |
All of this enables the following usage styles: |
438 |
|
439 |
1. Blocking: |
440 |
|
441 |
my $data = $fcp->client_get ($url); |
442 |
|
443 |
2. Blocking, but parallelizing: |
444 |
|
445 |
my @datas = map $_->result, |
446 |
map $fcp->txn_client_get ($_), |
447 |
@urls; |
448 |
|
449 |
Both blocking examples work without the module user having to know |
450 |
anything about events. |
451 |
|
452 |
3a. Event-based in a main program, using any support Event module: |
453 |
|
454 |
use Event; |
455 |
|
456 |
$fcp->txn_client_get ($url)->cb (sub { |
457 |
my $txn = shift; |
458 |
my $data = $txn->result; |
459 |
... |
460 |
}); |
461 |
|
462 |
Event::loop; |
463 |
|
464 |
3b. The module user could use AnyEvent, too: |
465 |
|
466 |
use AnyEvent; |
467 |
|
468 |
my $quit = AnyEvent->condvar; |
469 |
|
470 |
$fcp->txn_client_get ($url)->cb (sub { |
471 |
... |
472 |
$quit->broadcast; |
473 |
}); |
474 |
|
475 |
$quit->wait; |
476 |
|
477 |
=head1 SEE ALSO |
478 |
|
479 |
Event modules: L<Coro::Event>, L<Coro>, L<Event>, L<Glib::Event>, L<Glib>. |
480 |
|
481 |
Implementations: L<AnyEvent::Impl::Coro>, L<AnyEvent::Impl::Event>, L<AnyEvent::Impl::Glib>, L<AnyEvent::Impl::Tk>. |
482 |
|
483 |
Nontrivial usage example: L<Net::FCP>. |
484 |
|
485 |
=head1 |
486 |
|
487 |
=cut |
488 |
|
489 |
1 |
490 |
|