1 |
package AnyEvent::Handle; |
2 |
|
3 |
no warnings; |
4 |
use strict qw(subs vars); |
5 |
|
6 |
use AnyEvent (); |
7 |
use AnyEvent::Util qw(WSAEWOULDBLOCK); |
8 |
use Scalar::Util (); |
9 |
use Carp (); |
10 |
use Fcntl (); |
11 |
use Errno qw(EAGAIN EINTR); |
12 |
|
13 |
=head1 NAME |
14 |
|
15 |
AnyEvent::Handle - non-blocking I/O on file handles via AnyEvent |
16 |
|
17 |
=cut |
18 |
|
19 |
our $VERSION = 4.232; |
20 |
|
21 |
=head1 SYNOPSIS |
22 |
|
23 |
use AnyEvent; |
24 |
use AnyEvent::Handle; |
25 |
|
26 |
my $cv = AnyEvent->condvar; |
27 |
|
28 |
my $handle = |
29 |
AnyEvent::Handle->new ( |
30 |
fh => \*STDIN, |
31 |
on_eof => sub { |
32 |
$cv->broadcast; |
33 |
}, |
34 |
); |
35 |
|
36 |
# send some request line |
37 |
$handle->push_write ("getinfo\015\012"); |
38 |
|
39 |
# read the response line |
40 |
$handle->push_read (line => sub { |
41 |
my ($handle, $line) = @_; |
42 |
warn "read line <$line>\n"; |
43 |
$cv->send; |
44 |
}); |
45 |
|
46 |
$cv->recv; |
47 |
|
48 |
=head1 DESCRIPTION |
49 |
|
50 |
This module is a helper module to make it easier to do event-based I/O on |
51 |
filehandles. For utility functions for doing non-blocking connects and accepts |
52 |
on sockets see L<AnyEvent::Util>. |
53 |
|
54 |
The L<AnyEvent::Intro> tutorial contains some well-documented |
55 |
AnyEvent::Handle examples. |
56 |
|
57 |
In the following, when the documentation refers to of "bytes" then this |
58 |
means characters. As sysread and syswrite are used for all I/O, their |
59 |
treatment of characters applies to this module as well. |
60 |
|
61 |
All callbacks will be invoked with the handle object as their first |
62 |
argument. |
63 |
|
64 |
=head1 METHODS |
65 |
|
66 |
=over 4 |
67 |
|
68 |
=item B<new (%args)> |
69 |
|
70 |
The constructor supports these arguments (all as key => value pairs). |
71 |
|
72 |
=over 4 |
73 |
|
74 |
=item fh => $filehandle [MANDATORY] |
75 |
|
76 |
The filehandle this L<AnyEvent::Handle> object will operate on. |
77 |
|
78 |
NOTE: The filehandle will be set to non-blocking mode (using |
79 |
C<AnyEvent::Util::fh_nonblocking>) by the constructor and needs to stay in |
80 |
that mode. |
81 |
|
82 |
=item on_eof => $cb->($handle) |
83 |
|
84 |
Set the callback to be called when an end-of-file condition is detected, |
85 |
i.e. in the case of a socket, when the other side has closed the |
86 |
connection cleanly. |
87 |
|
88 |
For sockets, this just means that the other side has stopped sending data, |
89 |
you can still try to write data, and, in fact, one can return from the eof |
90 |
callback and continue writing data, as only the read part has been shut |
91 |
down. |
92 |
|
93 |
While not mandatory, it is I<highly> recommended to set an eof callback, |
94 |
otherwise you might end up with a closed socket while you are still |
95 |
waiting for data. |
96 |
|
97 |
If an EOF condition has been detected but no C<on_eof> callback has been |
98 |
set, then a fatal error will be raised with C<$!> set to <0>. |
99 |
|
100 |
=item on_error => $cb->($handle, $fatal) |
101 |
|
102 |
This is the error callback, which is called when, well, some error |
103 |
occured, such as not being able to resolve the hostname, failure to |
104 |
connect or a read error. |
105 |
|
106 |
Some errors are fatal (which is indicated by C<$fatal> being true). On |
107 |
fatal errors the handle object will be shut down and will not be usable |
108 |
(but you are free to look at the current C< ->rbuf >). Examples of fatal |
109 |
errors are an EOF condition with active (but unsatisifable) read watchers |
110 |
(C<EPIPE>) or I/O errors. |
111 |
|
112 |
Non-fatal errors can be retried by simply returning, but it is recommended |
113 |
to simply ignore this parameter and instead abondon the handle object |
114 |
when this callback is invoked. Examples of non-fatal errors are timeouts |
115 |
C<ETIMEDOUT>) or badly-formatted data (C<EBADMSG>). |
116 |
|
117 |
On callback entrance, the value of C<$!> contains the operating system |
118 |
error (or C<ENOSPC>, C<EPIPE>, C<ETIMEDOUT> or C<EBADMSG>). |
119 |
|
120 |
While not mandatory, it is I<highly> recommended to set this callback, as |
121 |
you will not be notified of errors otherwise. The default simply calls |
122 |
C<croak>. |
123 |
|
124 |
=item on_read => $cb->($handle) |
125 |
|
126 |
This sets the default read callback, which is called when data arrives |
127 |
and no read request is in the queue (unlike read queue callbacks, this |
128 |
callback will only be called when at least one octet of data is in the |
129 |
read buffer). |
130 |
|
131 |
To access (and remove data from) the read buffer, use the C<< ->rbuf >> |
132 |
method or access the C<$handle->{rbuf}> member directly. |
133 |
|
134 |
When an EOF condition is detected then AnyEvent::Handle will first try to |
135 |
feed all the remaining data to the queued callbacks and C<on_read> before |
136 |
calling the C<on_eof> callback. If no progress can be made, then a fatal |
137 |
error will be raised (with C<$!> set to C<EPIPE>). |
138 |
|
139 |
=item on_drain => $cb->($handle) |
140 |
|
141 |
This sets the callback that is called when the write buffer becomes empty |
142 |
(or when the callback is set and the buffer is empty already). |
143 |
|
144 |
To append to the write buffer, use the C<< ->push_write >> method. |
145 |
|
146 |
This callback is useful when you don't want to put all of your write data |
147 |
into the queue at once, for example, when you want to write the contents |
148 |
of some file to the socket you might not want to read the whole file into |
149 |
memory and push it into the queue, but instead only read more data from |
150 |
the file when the write queue becomes empty. |
151 |
|
152 |
=item timeout => $fractional_seconds |
153 |
|
154 |
If non-zero, then this enables an "inactivity" timeout: whenever this many |
155 |
seconds pass without a successful read or write on the underlying file |
156 |
handle, the C<on_timeout> callback will be invoked (and if that one is |
157 |
missing, an C<ETIMEDOUT> error will be raised). |
158 |
|
159 |
Note that timeout processing is also active when you currently do not have |
160 |
any outstanding read or write requests: If you plan to keep the connection |
161 |
idle then you should disable the timout temporarily or ignore the timeout |
162 |
in the C<on_timeout> callback. |
163 |
|
164 |
Zero (the default) disables this timeout. |
165 |
|
166 |
=item on_timeout => $cb->($handle) |
167 |
|
168 |
Called whenever the inactivity timeout passes. If you return from this |
169 |
callback, then the timeout will be reset as if some activity had happened, |
170 |
so this condition is not fatal in any way. |
171 |
|
172 |
=item rbuf_max => <bytes> |
173 |
|
174 |
If defined, then a fatal error will be raised (with C<$!> set to C<ENOSPC>) |
175 |
when the read buffer ever (strictly) exceeds this size. This is useful to |
176 |
avoid denial-of-service attacks. |
177 |
|
178 |
For example, a server accepting connections from untrusted sources should |
179 |
be configured to accept only so-and-so much data that it cannot act on |
180 |
(for example, when expecting a line, an attacker could send an unlimited |
181 |
amount of data without a callback ever being called as long as the line |
182 |
isn't finished). |
183 |
|
184 |
=item autocork => <boolean> |
185 |
|
186 |
When disabled (the default), then C<push_write> will try to immediately |
187 |
write the data to the handle if possible. This avoids having to register |
188 |
a write watcher and wait for the next event loop iteration, but can be |
189 |
inefficient if you write multiple small chunks (this disadvantage is |
190 |
usually avoided by your kernel's nagle algorithm, see C<low_delay>). |
191 |
|
192 |
When enabled, then writes will always be queued till the next event loop |
193 |
iteration. This is efficient when you do many small writes per iteration, |
194 |
but less efficient when you do a single write only. |
195 |
|
196 |
=item no_delay => <boolean> |
197 |
|
198 |
When doing small writes on sockets, your operating system kernel might |
199 |
wait a bit for more data before actually sending it out. This is called |
200 |
the Nagle algorithm, and usually it is beneficial. |
201 |
|
202 |
In some situations you want as low a delay as possible, which cna be |
203 |
accomplishd by setting this option to true. |
204 |
|
205 |
The default is your opertaing system's default behaviour, this option |
206 |
explicitly enables or disables it, if possible. |
207 |
|
208 |
=item read_size => <bytes> |
209 |
|
210 |
The default read block size (the amount of bytes this module will try to read |
211 |
during each (loop iteration). Default: C<8192>. |
212 |
|
213 |
=item low_water_mark => <bytes> |
214 |
|
215 |
Sets the amount of bytes (default: C<0>) that make up an "empty" write |
216 |
buffer: If the write reaches this size or gets even samller it is |
217 |
considered empty. |
218 |
|
219 |
=item linger => <seconds> |
220 |
|
221 |
If non-zero (default: C<3600>), then the destructor of the |
222 |
AnyEvent::Handle object will check wether there is still outstanding write |
223 |
data and will install a watcher that will write out this data. No errors |
224 |
will be reported (this mostly matches how the operating system treats |
225 |
outstanding data at socket close time). |
226 |
|
227 |
This will not work for partial TLS data that could not yet been |
228 |
encoded. This data will be lost. |
229 |
|
230 |
=item tls => "accept" | "connect" | Net::SSLeay::SSL object |
231 |
|
232 |
When this parameter is given, it enables TLS (SSL) mode, that means |
233 |
AnyEvent will start a TLS handshake and will transparently encrypt/decrypt |
234 |
data. |
235 |
|
236 |
TLS mode requires Net::SSLeay to be installed (it will be loaded |
237 |
automatically when you try to create a TLS handle). |
238 |
|
239 |
Unlike TCP, TLS has a server and client side: for the TLS server side, use |
240 |
C<accept>, and for the TLS client side of a connection, use C<connect> |
241 |
mode. |
242 |
|
243 |
You can also provide your own TLS connection object, but you have |
244 |
to make sure that you call either C<Net::SSLeay::set_connect_state> |
245 |
or C<Net::SSLeay::set_accept_state> on it before you pass it to |
246 |
AnyEvent::Handle. |
247 |
|
248 |
See the C<starttls> method for when need to start TLS negotiation later. |
249 |
|
250 |
=item tls_ctx => $ssl_ctx |
251 |
|
252 |
Use the given Net::SSLeay::CTX object to create the new TLS connection |
253 |
(unless a connection object was specified directly). If this parameter is |
254 |
missing, then AnyEvent::Handle will use C<AnyEvent::Handle::TLS_CTX>. |
255 |
|
256 |
=item json => JSON or JSON::XS object |
257 |
|
258 |
This is the json coder object used by the C<json> read and write types. |
259 |
|
260 |
If you don't supply it, then AnyEvent::Handle will create and use a |
261 |
suitable one (on demand), which will write and expect UTF-8 encoded JSON |
262 |
texts. |
263 |
|
264 |
Note that you are responsible to depend on the JSON module if you want to |
265 |
use this functionality, as AnyEvent does not have a dependency itself. |
266 |
|
267 |
=item filter_r => $cb |
268 |
|
269 |
=item filter_w => $cb |
270 |
|
271 |
These exist, but are undocumented at this time. |
272 |
|
273 |
=back |
274 |
|
275 |
=cut |
276 |
|
277 |
sub new { |
278 |
my $class = shift; |
279 |
|
280 |
my $self = bless { @_ }, $class; |
281 |
|
282 |
$self->{fh} or Carp::croak "mandatory argument fh is missing"; |
283 |
|
284 |
AnyEvent::Util::fh_nonblocking $self->{fh}, 1; |
285 |
|
286 |
if ($self->{tls}) { |
287 |
require Net::SSLeay; |
288 |
$self->starttls (delete $self->{tls}, delete $self->{tls_ctx}); |
289 |
} |
290 |
|
291 |
$self->{_activity} = AnyEvent->now; |
292 |
$self->_timeout; |
293 |
|
294 |
$self->on_drain (delete $self->{on_drain}) if exists $self->{on_drain}; |
295 |
$self->no_delay (delete $self->{no_delay}) if exists $self->{no_delay}; |
296 |
|
297 |
$self->start_read |
298 |
if $self->{on_read}; |
299 |
|
300 |
$self |
301 |
} |
302 |
|
303 |
sub _shutdown { |
304 |
my ($self) = @_; |
305 |
|
306 |
delete $self->{_tw}; |
307 |
delete $self->{_rw}; |
308 |
delete $self->{_ww}; |
309 |
delete $self->{fh}; |
310 |
|
311 |
$self->stoptls; |
312 |
|
313 |
delete $self->{on_read}; |
314 |
delete $self->{_queue}; |
315 |
} |
316 |
|
317 |
sub _error { |
318 |
my ($self, $errno, $fatal) = @_; |
319 |
|
320 |
$self->_shutdown |
321 |
if $fatal; |
322 |
|
323 |
$! = $errno; |
324 |
|
325 |
if ($self->{on_error}) { |
326 |
$self->{on_error}($self, $fatal); |
327 |
} else { |
328 |
Carp::croak "AnyEvent::Handle uncaught error: $!"; |
329 |
} |
330 |
} |
331 |
|
332 |
=item $fh = $handle->fh |
333 |
|
334 |
This method returns the file handle of the L<AnyEvent::Handle> object. |
335 |
|
336 |
=cut |
337 |
|
338 |
sub fh { $_[0]{fh} } |
339 |
|
340 |
=item $handle->on_error ($cb) |
341 |
|
342 |
Replace the current C<on_error> callback (see the C<on_error> constructor argument). |
343 |
|
344 |
=cut |
345 |
|
346 |
sub on_error { |
347 |
$_[0]{on_error} = $_[1]; |
348 |
} |
349 |
|
350 |
=item $handle->on_eof ($cb) |
351 |
|
352 |
Replace the current C<on_eof> callback (see the C<on_eof> constructor argument). |
353 |
|
354 |
=cut |
355 |
|
356 |
sub on_eof { |
357 |
$_[0]{on_eof} = $_[1]; |
358 |
} |
359 |
|
360 |
=item $handle->on_timeout ($cb) |
361 |
|
362 |
Replace the current C<on_timeout> callback, or disables the callback |
363 |
(but not the timeout) if C<$cb> = C<undef>. See C<timeout> constructor |
364 |
argument. |
365 |
|
366 |
=cut |
367 |
|
368 |
sub on_timeout { |
369 |
$_[0]{on_timeout} = $_[1]; |
370 |
} |
371 |
|
372 |
=item $handle->autocork ($boolean) |
373 |
|
374 |
Enables or disables the current autocork behaviour (see C<autocork> |
375 |
constructor argument). |
376 |
|
377 |
=cut |
378 |
|
379 |
=item $handle->no_delay ($boolean) |
380 |
|
381 |
Enables or disables the C<no_delay> setting (see constructor argument of |
382 |
the same name for details). |
383 |
|
384 |
=cut |
385 |
|
386 |
sub no_delay { |
387 |
$_[0]{no_delay} = $_[1]; |
388 |
|
389 |
eval { |
390 |
local $SIG{__DIE__}; |
391 |
setsockopt $_[0]{fh}, &Socket::IPPROTO_TCP, &Socket::TCP_NODELAY, int $_[1]; |
392 |
}; |
393 |
} |
394 |
|
395 |
############################################################################# |
396 |
|
397 |
=item $handle->timeout ($seconds) |
398 |
|
399 |
Configures (or disables) the inactivity timeout. |
400 |
|
401 |
=cut |
402 |
|
403 |
sub timeout { |
404 |
my ($self, $timeout) = @_; |
405 |
|
406 |
$self->{timeout} = $timeout; |
407 |
$self->_timeout; |
408 |
} |
409 |
|
410 |
# reset the timeout watcher, as neccessary |
411 |
# also check for time-outs |
412 |
sub _timeout { |
413 |
my ($self) = @_; |
414 |
|
415 |
if ($self->{timeout}) { |
416 |
my $NOW = AnyEvent->now; |
417 |
|
418 |
# when would the timeout trigger? |
419 |
my $after = $self->{_activity} + $self->{timeout} - $NOW; |
420 |
|
421 |
# now or in the past already? |
422 |
if ($after <= 0) { |
423 |
$self->{_activity} = $NOW; |
424 |
|
425 |
if ($self->{on_timeout}) { |
426 |
$self->{on_timeout}($self); |
427 |
} else { |
428 |
$self->_error (&Errno::ETIMEDOUT); |
429 |
} |
430 |
|
431 |
# callback could have changed timeout value, optimise |
432 |
return unless $self->{timeout}; |
433 |
|
434 |
# calculate new after |
435 |
$after = $self->{timeout}; |
436 |
} |
437 |
|
438 |
Scalar::Util::weaken $self; |
439 |
return unless $self; # ->error could have destroyed $self |
440 |
|
441 |
$self->{_tw} ||= AnyEvent->timer (after => $after, cb => sub { |
442 |
delete $self->{_tw}; |
443 |
$self->_timeout; |
444 |
}); |
445 |
} else { |
446 |
delete $self->{_tw}; |
447 |
} |
448 |
} |
449 |
|
450 |
############################################################################# |
451 |
|
452 |
=back |
453 |
|
454 |
=head2 WRITE QUEUE |
455 |
|
456 |
AnyEvent::Handle manages two queues per handle, one for writing and one |
457 |
for reading. |
458 |
|
459 |
The write queue is very simple: you can add data to its end, and |
460 |
AnyEvent::Handle will automatically try to get rid of it for you. |
461 |
|
462 |
When data could be written and the write buffer is shorter then the low |
463 |
water mark, the C<on_drain> callback will be invoked. |
464 |
|
465 |
=over 4 |
466 |
|
467 |
=item $handle->on_drain ($cb) |
468 |
|
469 |
Sets the C<on_drain> callback or clears it (see the description of |
470 |
C<on_drain> in the constructor). |
471 |
|
472 |
=cut |
473 |
|
474 |
sub on_drain { |
475 |
my ($self, $cb) = @_; |
476 |
|
477 |
$self->{on_drain} = $cb; |
478 |
|
479 |
$cb->($self) |
480 |
if $cb && $self->{low_water_mark} >= length $self->{wbuf}; |
481 |
} |
482 |
|
483 |
=item $handle->push_write ($data) |
484 |
|
485 |
Queues the given scalar to be written. You can push as much data as you |
486 |
want (only limited by the available memory), as C<AnyEvent::Handle> |
487 |
buffers it independently of the kernel. |
488 |
|
489 |
=cut |
490 |
|
491 |
sub _drain_wbuf { |
492 |
my ($self) = @_; |
493 |
|
494 |
if (!$self->{_ww} && length $self->{wbuf}) { |
495 |
|
496 |
Scalar::Util::weaken $self; |
497 |
|
498 |
my $cb = sub { |
499 |
my $len = syswrite $self->{fh}, $self->{wbuf}; |
500 |
|
501 |
if ($len >= 0) { |
502 |
substr $self->{wbuf}, 0, $len, ""; |
503 |
|
504 |
$self->{_activity} = AnyEvent->now; |
505 |
|
506 |
$self->{on_drain}($self) |
507 |
if $self->{low_water_mark} >= length $self->{wbuf} |
508 |
&& $self->{on_drain}; |
509 |
|
510 |
delete $self->{_ww} unless length $self->{wbuf}; |
511 |
} elsif ($! != EAGAIN && $! != EINTR && $! != WSAEWOULDBLOCK) { |
512 |
$self->_error ($!, 1); |
513 |
} |
514 |
}; |
515 |
|
516 |
# try to write data immediately |
517 |
$cb->() unless $self->{autocork}; |
518 |
|
519 |
# if still data left in wbuf, we need to poll |
520 |
$self->{_ww} = AnyEvent->io (fh => $self->{fh}, poll => "w", cb => $cb) |
521 |
if length $self->{wbuf}; |
522 |
}; |
523 |
} |
524 |
|
525 |
our %WH; |
526 |
|
527 |
sub register_write_type($$) { |
528 |
$WH{$_[0]} = $_[1]; |
529 |
} |
530 |
|
531 |
sub push_write { |
532 |
my $self = shift; |
533 |
|
534 |
if (@_ > 1) { |
535 |
my $type = shift; |
536 |
|
537 |
@_ = ($WH{$type} or Carp::croak "unsupported type passed to AnyEvent::Handle::push_write") |
538 |
->($self, @_); |
539 |
} |
540 |
|
541 |
if ($self->{filter_w}) { |
542 |
$self->{filter_w}($self, \$_[0]); |
543 |
} else { |
544 |
$self->{wbuf} .= $_[0]; |
545 |
$self->_drain_wbuf; |
546 |
} |
547 |
} |
548 |
|
549 |
=item $handle->push_write (type => @args) |
550 |
|
551 |
Instead of formatting your data yourself, you can also let this module do |
552 |
the job by specifying a type and type-specific arguments. |
553 |
|
554 |
Predefined types are (if you have ideas for additional types, feel free to |
555 |
drop by and tell us): |
556 |
|
557 |
=over 4 |
558 |
|
559 |
=item netstring => $string |
560 |
|
561 |
Formats the given value as netstring |
562 |
(http://cr.yp.to/proto/netstrings.txt, this is not a recommendation to use them). |
563 |
|
564 |
=cut |
565 |
|
566 |
register_write_type netstring => sub { |
567 |
my ($self, $string) = @_; |
568 |
|
569 |
sprintf "%d:%s,", (length $string), $string |
570 |
}; |
571 |
|
572 |
=item packstring => $format, $data |
573 |
|
574 |
An octet string prefixed with an encoded length. The encoding C<$format> |
575 |
uses the same format as a Perl C<pack> format, but must specify a single |
576 |
integer only (only one of C<cCsSlLqQiInNvVjJw> is allowed, plus an |
577 |
optional C<!>, C<< < >> or C<< > >> modifier). |
578 |
|
579 |
=cut |
580 |
|
581 |
register_write_type packstring => sub { |
582 |
my ($self, $format, $string) = @_; |
583 |
|
584 |
pack "$format/a*", $string |
585 |
}; |
586 |
|
587 |
=item json => $array_or_hashref |
588 |
|
589 |
Encodes the given hash or array reference into a JSON object. Unless you |
590 |
provide your own JSON object, this means it will be encoded to JSON text |
591 |
in UTF-8. |
592 |
|
593 |
JSON objects (and arrays) are self-delimiting, so you can write JSON at |
594 |
one end of a handle and read them at the other end without using any |
595 |
additional framing. |
596 |
|
597 |
The generated JSON text is guaranteed not to contain any newlines: While |
598 |
this module doesn't need delimiters after or between JSON texts to be |
599 |
able to read them, many other languages depend on that. |
600 |
|
601 |
A simple RPC protocol that interoperates easily with others is to send |
602 |
JSON arrays (or objects, although arrays are usually the better choice as |
603 |
they mimic how function argument passing works) and a newline after each |
604 |
JSON text: |
605 |
|
606 |
$handle->push_write (json => ["method", "arg1", "arg2"]); # whatever |
607 |
$handle->push_write ("\012"); |
608 |
|
609 |
An AnyEvent::Handle receiver would simply use the C<json> read type and |
610 |
rely on the fact that the newline will be skipped as leading whitespace: |
611 |
|
612 |
$handle->push_read (json => sub { my $array = $_[1]; ... }); |
613 |
|
614 |
Other languages could read single lines terminated by a newline and pass |
615 |
this line into their JSON decoder of choice. |
616 |
|
617 |
=cut |
618 |
|
619 |
register_write_type json => sub { |
620 |
my ($self, $ref) = @_; |
621 |
|
622 |
require JSON; |
623 |
|
624 |
$self->{json} ? $self->{json}->encode ($ref) |
625 |
: JSON::encode_json ($ref) |
626 |
}; |
627 |
|
628 |
=item storable => $reference |
629 |
|
630 |
Freezes the given reference using L<Storable> and writes it to the |
631 |
handle. Uses the C<nfreeze> format. |
632 |
|
633 |
=cut |
634 |
|
635 |
register_write_type storable => sub { |
636 |
my ($self, $ref) = @_; |
637 |
|
638 |
require Storable; |
639 |
|
640 |
pack "w/a*", Storable::nfreeze ($ref) |
641 |
}; |
642 |
|
643 |
=back |
644 |
|
645 |
=item AnyEvent::Handle::register_write_type type => $coderef->($handle, @args) |
646 |
|
647 |
This function (not method) lets you add your own types to C<push_write>. |
648 |
Whenever the given C<type> is used, C<push_write> will invoke the code |
649 |
reference with the handle object and the remaining arguments. |
650 |
|
651 |
The code reference is supposed to return a single octet string that will |
652 |
be appended to the write buffer. |
653 |
|
654 |
Note that this is a function, and all types registered this way will be |
655 |
global, so try to use unique names. |
656 |
|
657 |
=cut |
658 |
|
659 |
############################################################################# |
660 |
|
661 |
=back |
662 |
|
663 |
=head2 READ QUEUE |
664 |
|
665 |
AnyEvent::Handle manages two queues per handle, one for writing and one |
666 |
for reading. |
667 |
|
668 |
The read queue is more complex than the write queue. It can be used in two |
669 |
ways, the "simple" way, using only C<on_read> and the "complex" way, using |
670 |
a queue. |
671 |
|
672 |
In the simple case, you just install an C<on_read> callback and whenever |
673 |
new data arrives, it will be called. You can then remove some data (if |
674 |
enough is there) from the read buffer (C<< $handle->rbuf >>). Or you cna |
675 |
leave the data there if you want to accumulate more (e.g. when only a |
676 |
partial message has been received so far). |
677 |
|
678 |
In the more complex case, you want to queue multiple callbacks. In this |
679 |
case, AnyEvent::Handle will call the first queued callback each time new |
680 |
data arrives (also the first time it is queued) and removes it when it has |
681 |
done its job (see C<push_read>, below). |
682 |
|
683 |
This way you can, for example, push three line-reads, followed by reading |
684 |
a chunk of data, and AnyEvent::Handle will execute them in order. |
685 |
|
686 |
Example 1: EPP protocol parser. EPP sends 4 byte length info, followed by |
687 |
the specified number of bytes which give an XML datagram. |
688 |
|
689 |
# in the default state, expect some header bytes |
690 |
$handle->on_read (sub { |
691 |
# some data is here, now queue the length-header-read (4 octets) |
692 |
shift->unshift_read (chunk => 4, sub { |
693 |
# header arrived, decode |
694 |
my $len = unpack "N", $_[1]; |
695 |
|
696 |
# now read the payload |
697 |
shift->unshift_read (chunk => $len, sub { |
698 |
my $xml = $_[1]; |
699 |
# handle xml |
700 |
}); |
701 |
}); |
702 |
}); |
703 |
|
704 |
Example 2: Implement a client for a protocol that replies either with "OK" |
705 |
and another line or "ERROR" for the first request that is sent, and 64 |
706 |
bytes for the second request. Due to the availability of a queue, we can |
707 |
just pipeline sending both requests and manipulate the queue as necessary |
708 |
in the callbacks. |
709 |
|
710 |
When the first callback is called and sees an "OK" response, it will |
711 |
C<unshift> another line-read. This line-read will be queued I<before> the |
712 |
64-byte chunk callback. |
713 |
|
714 |
# request one, returns either "OK + extra line" or "ERROR" |
715 |
$handle->push_write ("request 1\015\012"); |
716 |
|
717 |
# we expect "ERROR" or "OK" as response, so push a line read |
718 |
$handle->push_read (line => sub { |
719 |
# if we got an "OK", we have to _prepend_ another line, |
720 |
# so it will be read before the second request reads its 64 bytes |
721 |
# which are already in the queue when this callback is called |
722 |
# we don't do this in case we got an error |
723 |
if ($_[1] eq "OK") { |
724 |
$_[0]->unshift_read (line => sub { |
725 |
my $response = $_[1]; |
726 |
... |
727 |
}); |
728 |
} |
729 |
}); |
730 |
|
731 |
# request two, simply returns 64 octets |
732 |
$handle->push_write ("request 2\015\012"); |
733 |
|
734 |
# simply read 64 bytes, always |
735 |
$handle->push_read (chunk => 64, sub { |
736 |
my $response = $_[1]; |
737 |
... |
738 |
}); |
739 |
|
740 |
=over 4 |
741 |
|
742 |
=cut |
743 |
|
744 |
sub _drain_rbuf { |
745 |
my ($self) = @_; |
746 |
|
747 |
local $self->{_in_drain} = 1; |
748 |
|
749 |
if ( |
750 |
defined $self->{rbuf_max} |
751 |
&& $self->{rbuf_max} < length $self->{rbuf} |
752 |
) { |
753 |
$self->_error (&Errno::ENOSPC, 1), return; |
754 |
} |
755 |
|
756 |
while () { |
757 |
my $len = length $self->{rbuf}; |
758 |
|
759 |
if (my $cb = shift @{ $self->{_queue} }) { |
760 |
unless ($cb->($self)) { |
761 |
if ($self->{_eof}) { |
762 |
# no progress can be made (not enough data and no data forthcoming) |
763 |
$self->_error (&Errno::EPIPE, 1), return; |
764 |
} |
765 |
|
766 |
unshift @{ $self->{_queue} }, $cb; |
767 |
last; |
768 |
} |
769 |
} elsif ($self->{on_read}) { |
770 |
last unless $len; |
771 |
|
772 |
$self->{on_read}($self); |
773 |
|
774 |
if ( |
775 |
$len == length $self->{rbuf} # if no data has been consumed |
776 |
&& !@{ $self->{_queue} } # and the queue is still empty |
777 |
&& $self->{on_read} # but we still have on_read |
778 |
) { |
779 |
# no further data will arrive |
780 |
# so no progress can be made |
781 |
$self->_error (&Errno::EPIPE, 1), return |
782 |
if $self->{_eof}; |
783 |
|
784 |
last; # more data might arrive |
785 |
} |
786 |
} else { |
787 |
# read side becomes idle |
788 |
delete $self->{_rw}; |
789 |
last; |
790 |
} |
791 |
} |
792 |
|
793 |
if ($self->{_eof}) { |
794 |
if ($self->{on_eof}) { |
795 |
$self->{on_eof}($self) |
796 |
} else { |
797 |
$self->_error (0, 1); |
798 |
} |
799 |
} |
800 |
|
801 |
# may need to restart read watcher |
802 |
unless ($self->{_rw}) { |
803 |
$self->start_read |
804 |
if $self->{on_read} || @{ $self->{_queue} }; |
805 |
} |
806 |
} |
807 |
|
808 |
=item $handle->on_read ($cb) |
809 |
|
810 |
This replaces the currently set C<on_read> callback, or clears it (when |
811 |
the new callback is C<undef>). See the description of C<on_read> in the |
812 |
constructor. |
813 |
|
814 |
=cut |
815 |
|
816 |
sub on_read { |
817 |
my ($self, $cb) = @_; |
818 |
|
819 |
$self->{on_read} = $cb; |
820 |
$self->_drain_rbuf if $cb && !$self->{_in_drain}; |
821 |
} |
822 |
|
823 |
=item $handle->rbuf |
824 |
|
825 |
Returns the read buffer (as a modifiable lvalue). |
826 |
|
827 |
You can access the read buffer directly as the C<< ->{rbuf} >> member, if |
828 |
you want. |
829 |
|
830 |
NOTE: The read buffer should only be used or modified if the C<on_read>, |
831 |
C<push_read> or C<unshift_read> methods are used. The other read methods |
832 |
automatically manage the read buffer. |
833 |
|
834 |
=cut |
835 |
|
836 |
sub rbuf : lvalue { |
837 |
$_[0]{rbuf} |
838 |
} |
839 |
|
840 |
=item $handle->push_read ($cb) |
841 |
|
842 |
=item $handle->unshift_read ($cb) |
843 |
|
844 |
Append the given callback to the end of the queue (C<push_read>) or |
845 |
prepend it (C<unshift_read>). |
846 |
|
847 |
The callback is called each time some additional read data arrives. |
848 |
|
849 |
It must check whether enough data is in the read buffer already. |
850 |
|
851 |
If not enough data is available, it must return the empty list or a false |
852 |
value, in which case it will be called repeatedly until enough data is |
853 |
available (or an error condition is detected). |
854 |
|
855 |
If enough data was available, then the callback must remove all data it is |
856 |
interested in (which can be none at all) and return a true value. After returning |
857 |
true, it will be removed from the queue. |
858 |
|
859 |
=cut |
860 |
|
861 |
our %RH; |
862 |
|
863 |
sub register_read_type($$) { |
864 |
$RH{$_[0]} = $_[1]; |
865 |
} |
866 |
|
867 |
sub push_read { |
868 |
my $self = shift; |
869 |
my $cb = pop; |
870 |
|
871 |
if (@_) { |
872 |
my $type = shift; |
873 |
|
874 |
$cb = ($RH{$type} or Carp::croak "unsupported type passed to AnyEvent::Handle::push_read") |
875 |
->($self, $cb, @_); |
876 |
} |
877 |
|
878 |
push @{ $self->{_queue} }, $cb; |
879 |
$self->_drain_rbuf unless $self->{_in_drain}; |
880 |
} |
881 |
|
882 |
sub unshift_read { |
883 |
my $self = shift; |
884 |
my $cb = pop; |
885 |
|
886 |
if (@_) { |
887 |
my $type = shift; |
888 |
|
889 |
$cb = ($RH{$type} or Carp::croak "unsupported type passed to AnyEvent::Handle::unshift_read") |
890 |
->($self, $cb, @_); |
891 |
} |
892 |
|
893 |
|
894 |
unshift @{ $self->{_queue} }, $cb; |
895 |
$self->_drain_rbuf unless $self->{_in_drain}; |
896 |
} |
897 |
|
898 |
=item $handle->push_read (type => @args, $cb) |
899 |
|
900 |
=item $handle->unshift_read (type => @args, $cb) |
901 |
|
902 |
Instead of providing a callback that parses the data itself you can chose |
903 |
between a number of predefined parsing formats, for chunks of data, lines |
904 |
etc. |
905 |
|
906 |
Predefined types are (if you have ideas for additional types, feel free to |
907 |
drop by and tell us): |
908 |
|
909 |
=over 4 |
910 |
|
911 |
=item chunk => $octets, $cb->($handle, $data) |
912 |
|
913 |
Invoke the callback only once C<$octets> bytes have been read. Pass the |
914 |
data read to the callback. The callback will never be called with less |
915 |
data. |
916 |
|
917 |
Example: read 2 bytes. |
918 |
|
919 |
$handle->push_read (chunk => 2, sub { |
920 |
warn "yay ", unpack "H*", $_[1]; |
921 |
}); |
922 |
|
923 |
=cut |
924 |
|
925 |
register_read_type chunk => sub { |
926 |
my ($self, $cb, $len) = @_; |
927 |
|
928 |
sub { |
929 |
$len <= length $_[0]{rbuf} or return; |
930 |
$cb->($_[0], substr $_[0]{rbuf}, 0, $len, ""); |
931 |
1 |
932 |
} |
933 |
}; |
934 |
|
935 |
=item line => [$eol, ]$cb->($handle, $line, $eol) |
936 |
|
937 |
The callback will be called only once a full line (including the end of |
938 |
line marker, C<$eol>) has been read. This line (excluding the end of line |
939 |
marker) will be passed to the callback as second argument (C<$line>), and |
940 |
the end of line marker as the third argument (C<$eol>). |
941 |
|
942 |
The end of line marker, C<$eol>, can be either a string, in which case it |
943 |
will be interpreted as a fixed record end marker, or it can be a regex |
944 |
object (e.g. created by C<qr>), in which case it is interpreted as a |
945 |
regular expression. |
946 |
|
947 |
The end of line marker argument C<$eol> is optional, if it is missing (NOT |
948 |
undef), then C<qr|\015?\012|> is used (which is good for most internet |
949 |
protocols). |
950 |
|
951 |
Partial lines at the end of the stream will never be returned, as they are |
952 |
not marked by the end of line marker. |
953 |
|
954 |
=cut |
955 |
|
956 |
register_read_type line => sub { |
957 |
my ($self, $cb, $eol) = @_; |
958 |
|
959 |
if (@_ < 3) { |
960 |
# this is more than twice as fast as the generic code below |
961 |
sub { |
962 |
$_[0]{rbuf} =~ s/^([^\015\012]*)(\015?\012)// or return; |
963 |
|
964 |
$cb->($_[0], $1, $2); |
965 |
1 |
966 |
} |
967 |
} else { |
968 |
$eol = quotemeta $eol unless ref $eol; |
969 |
$eol = qr|^(.*?)($eol)|s; |
970 |
|
971 |
sub { |
972 |
$_[0]{rbuf} =~ s/$eol// or return; |
973 |
|
974 |
$cb->($_[0], $1, $2); |
975 |
1 |
976 |
} |
977 |
} |
978 |
}; |
979 |
|
980 |
=item regex => $accept[, $reject[, $skip], $cb->($handle, $data) |
981 |
|
982 |
Makes a regex match against the regex object C<$accept> and returns |
983 |
everything up to and including the match. |
984 |
|
985 |
Example: read a single line terminated by '\n'. |
986 |
|
987 |
$handle->push_read (regex => qr<\n>, sub { ... }); |
988 |
|
989 |
If C<$reject> is given and not undef, then it determines when the data is |
990 |
to be rejected: it is matched against the data when the C<$accept> regex |
991 |
does not match and generates an C<EBADMSG> error when it matches. This is |
992 |
useful to quickly reject wrong data (to avoid waiting for a timeout or a |
993 |
receive buffer overflow). |
994 |
|
995 |
Example: expect a single decimal number followed by whitespace, reject |
996 |
anything else (not the use of an anchor). |
997 |
|
998 |
$handle->push_read (regex => qr<^[0-9]+\s>, qr<[^0-9]>, sub { ... }); |
999 |
|
1000 |
If C<$skip> is given and not C<undef>, then it will be matched against |
1001 |
the receive buffer when neither C<$accept> nor C<$reject> match, |
1002 |
and everything preceding and including the match will be accepted |
1003 |
unconditionally. This is useful to skip large amounts of data that you |
1004 |
know cannot be matched, so that the C<$accept> or C<$reject> regex do not |
1005 |
have to start matching from the beginning. This is purely an optimisation |
1006 |
and is usually worth only when you expect more than a few kilobytes. |
1007 |
|
1008 |
Example: expect a http header, which ends at C<\015\012\015\012>. Since we |
1009 |
expect the header to be very large (it isn't in practise, but...), we use |
1010 |
a skip regex to skip initial portions. The skip regex is tricky in that |
1011 |
it only accepts something not ending in either \015 or \012, as these are |
1012 |
required for the accept regex. |
1013 |
|
1014 |
$handle->push_read (regex => |
1015 |
qr<\015\012\015\012>, |
1016 |
undef, # no reject |
1017 |
qr<^.*[^\015\012]>, |
1018 |
sub { ... }); |
1019 |
|
1020 |
=cut |
1021 |
|
1022 |
register_read_type regex => sub { |
1023 |
my ($self, $cb, $accept, $reject, $skip) = @_; |
1024 |
|
1025 |
my $data; |
1026 |
my $rbuf = \$self->{rbuf}; |
1027 |
|
1028 |
sub { |
1029 |
# accept |
1030 |
if ($$rbuf =~ $accept) { |
1031 |
$data .= substr $$rbuf, 0, $+[0], ""; |
1032 |
$cb->($self, $data); |
1033 |
return 1; |
1034 |
} |
1035 |
|
1036 |
# reject |
1037 |
if ($reject && $$rbuf =~ $reject) { |
1038 |
$self->_error (&Errno::EBADMSG); |
1039 |
} |
1040 |
|
1041 |
# skip |
1042 |
if ($skip && $$rbuf =~ $skip) { |
1043 |
$data .= substr $$rbuf, 0, $+[0], ""; |
1044 |
} |
1045 |
|
1046 |
() |
1047 |
} |
1048 |
}; |
1049 |
|
1050 |
=item netstring => $cb->($handle, $string) |
1051 |
|
1052 |
A netstring (http://cr.yp.to/proto/netstrings.txt, this is not an endorsement). |
1053 |
|
1054 |
Throws an error with C<$!> set to EBADMSG on format violations. |
1055 |
|
1056 |
=cut |
1057 |
|
1058 |
register_read_type netstring => sub { |
1059 |
my ($self, $cb) = @_; |
1060 |
|
1061 |
sub { |
1062 |
unless ($_[0]{rbuf} =~ s/^(0|[1-9][0-9]*)://) { |
1063 |
if ($_[0]{rbuf} =~ /[^0-9]/) { |
1064 |
$self->_error (&Errno::EBADMSG); |
1065 |
} |
1066 |
return; |
1067 |
} |
1068 |
|
1069 |
my $len = $1; |
1070 |
|
1071 |
$self->unshift_read (chunk => $len, sub { |
1072 |
my $string = $_[1]; |
1073 |
$_[0]->unshift_read (chunk => 1, sub { |
1074 |
if ($_[1] eq ",") { |
1075 |
$cb->($_[0], $string); |
1076 |
} else { |
1077 |
$self->_error (&Errno::EBADMSG); |
1078 |
} |
1079 |
}); |
1080 |
}); |
1081 |
|
1082 |
1 |
1083 |
} |
1084 |
}; |
1085 |
|
1086 |
=item packstring => $format, $cb->($handle, $string) |
1087 |
|
1088 |
An octet string prefixed with an encoded length. The encoding C<$format> |
1089 |
uses the same format as a Perl C<pack> format, but must specify a single |
1090 |
integer only (only one of C<cCsSlLqQiInNvVjJw> is allowed, plus an |
1091 |
optional C<!>, C<< < >> or C<< > >> modifier). |
1092 |
|
1093 |
DNS over TCP uses a prefix of C<n>, EPP uses a prefix of C<N>. |
1094 |
|
1095 |
Example: read a block of data prefixed by its length in BER-encoded |
1096 |
format (very efficient). |
1097 |
|
1098 |
$handle->push_read (packstring => "w", sub { |
1099 |
my ($handle, $data) = @_; |
1100 |
}); |
1101 |
|
1102 |
=cut |
1103 |
|
1104 |
register_read_type packstring => sub { |
1105 |
my ($self, $cb, $format) = @_; |
1106 |
|
1107 |
sub { |
1108 |
# when we can use 5.10 we can use ".", but for 5.8 we use the re-pack method |
1109 |
defined (my $len = eval { unpack $format, $_[0]{rbuf} }) |
1110 |
or return; |
1111 |
|
1112 |
$format = length pack $format, $len; |
1113 |
|
1114 |
# bypass unshift if we already have the remaining chunk |
1115 |
if ($format + $len <= length $_[0]{rbuf}) { |
1116 |
my $data = substr $_[0]{rbuf}, $format, $len; |
1117 |
substr $_[0]{rbuf}, 0, $format + $len, ""; |
1118 |
$cb->($_[0], $data); |
1119 |
} else { |
1120 |
# remove prefix |
1121 |
substr $_[0]{rbuf}, 0, $format, ""; |
1122 |
|
1123 |
# read remaining chunk |
1124 |
$_[0]->unshift_read (chunk => $len, $cb); |
1125 |
} |
1126 |
|
1127 |
1 |
1128 |
} |
1129 |
}; |
1130 |
|
1131 |
=item json => $cb->($handle, $hash_or_arrayref) |
1132 |
|
1133 |
Reads a JSON object or array, decodes it and passes it to the callback. |
1134 |
|
1135 |
If a C<json> object was passed to the constructor, then that will be used |
1136 |
for the final decode, otherwise it will create a JSON coder expecting UTF-8. |
1137 |
|
1138 |
This read type uses the incremental parser available with JSON version |
1139 |
2.09 (and JSON::XS version 2.2) and above. You have to provide a |
1140 |
dependency on your own: this module will load the JSON module, but |
1141 |
AnyEvent does not depend on it itself. |
1142 |
|
1143 |
Since JSON texts are fully self-delimiting, the C<json> read and write |
1144 |
types are an ideal simple RPC protocol: just exchange JSON datagrams. See |
1145 |
the C<json> write type description, above, for an actual example. |
1146 |
|
1147 |
=cut |
1148 |
|
1149 |
register_read_type json => sub { |
1150 |
my ($self, $cb) = @_; |
1151 |
|
1152 |
require JSON; |
1153 |
|
1154 |
my $data; |
1155 |
my $rbuf = \$self->{rbuf}; |
1156 |
|
1157 |
my $json = $self->{json} ||= JSON->new->utf8; |
1158 |
|
1159 |
sub { |
1160 |
my $ref = $json->incr_parse ($self->{rbuf}); |
1161 |
|
1162 |
if ($ref) { |
1163 |
$self->{rbuf} = $json->incr_text; |
1164 |
$json->incr_text = ""; |
1165 |
$cb->($self, $ref); |
1166 |
|
1167 |
1 |
1168 |
} else { |
1169 |
$self->{rbuf} = ""; |
1170 |
() |
1171 |
} |
1172 |
} |
1173 |
}; |
1174 |
|
1175 |
=item storable => $cb->($handle, $ref) |
1176 |
|
1177 |
Deserialises a L<Storable> frozen representation as written by the |
1178 |
C<storable> write type (BER-encoded length prefix followed by nfreeze'd |
1179 |
data). |
1180 |
|
1181 |
Raises C<EBADMSG> error if the data could not be decoded. |
1182 |
|
1183 |
=cut |
1184 |
|
1185 |
register_read_type storable => sub { |
1186 |
my ($self, $cb) = @_; |
1187 |
|
1188 |
require Storable; |
1189 |
|
1190 |
sub { |
1191 |
# when we can use 5.10 we can use ".", but for 5.8 we use the re-pack method |
1192 |
defined (my $len = eval { unpack "w", $_[0]{rbuf} }) |
1193 |
or return; |
1194 |
|
1195 |
my $format = length pack "w", $len; |
1196 |
|
1197 |
# bypass unshift if we already have the remaining chunk |
1198 |
if ($format + $len <= length $_[0]{rbuf}) { |
1199 |
my $data = substr $_[0]{rbuf}, $format, $len; |
1200 |
substr $_[0]{rbuf}, 0, $format + $len, ""; |
1201 |
$cb->($_[0], Storable::thaw ($data)); |
1202 |
} else { |
1203 |
# remove prefix |
1204 |
substr $_[0]{rbuf}, 0, $format, ""; |
1205 |
|
1206 |
# read remaining chunk |
1207 |
$_[0]->unshift_read (chunk => $len, sub { |
1208 |
if (my $ref = eval { Storable::thaw ($_[1]) }) { |
1209 |
$cb->($_[0], $ref); |
1210 |
} else { |
1211 |
$self->_error (&Errno::EBADMSG); |
1212 |
} |
1213 |
}); |
1214 |
} |
1215 |
|
1216 |
1 |
1217 |
} |
1218 |
}; |
1219 |
|
1220 |
=back |
1221 |
|
1222 |
=item AnyEvent::Handle::register_read_type type => $coderef->($handle, $cb, @args) |
1223 |
|
1224 |
This function (not method) lets you add your own types to C<push_read>. |
1225 |
|
1226 |
Whenever the given C<type> is used, C<push_read> will invoke the code |
1227 |
reference with the handle object, the callback and the remaining |
1228 |
arguments. |
1229 |
|
1230 |
The code reference is supposed to return a callback (usually a closure) |
1231 |
that works as a plain read callback (see C<< ->push_read ($cb) >>). |
1232 |
|
1233 |
It should invoke the passed callback when it is done reading (remember to |
1234 |
pass C<$handle> as first argument as all other callbacks do that). |
1235 |
|
1236 |
Note that this is a function, and all types registered this way will be |
1237 |
global, so try to use unique names. |
1238 |
|
1239 |
For examples, see the source of this module (F<perldoc -m AnyEvent::Handle>, |
1240 |
search for C<register_read_type>)). |
1241 |
|
1242 |
=item $handle->stop_read |
1243 |
|
1244 |
=item $handle->start_read |
1245 |
|
1246 |
In rare cases you actually do not want to read anything from the |
1247 |
socket. In this case you can call C<stop_read>. Neither C<on_read> nor |
1248 |
any queued callbacks will be executed then. To start reading again, call |
1249 |
C<start_read>. |
1250 |
|
1251 |
Note that AnyEvent::Handle will automatically C<start_read> for you when |
1252 |
you change the C<on_read> callback or push/unshift a read callback, and it |
1253 |
will automatically C<stop_read> for you when neither C<on_read> is set nor |
1254 |
there are any read requests in the queue. |
1255 |
|
1256 |
=cut |
1257 |
|
1258 |
sub stop_read { |
1259 |
my ($self) = @_; |
1260 |
|
1261 |
delete $self->{_rw}; |
1262 |
} |
1263 |
|
1264 |
sub start_read { |
1265 |
my ($self) = @_; |
1266 |
|
1267 |
unless ($self->{_rw} || $self->{_eof}) { |
1268 |
Scalar::Util::weaken $self; |
1269 |
|
1270 |
$self->{_rw} = AnyEvent->io (fh => $self->{fh}, poll => "r", cb => sub { |
1271 |
my $rbuf = $self->{filter_r} ? \my $buf : \$self->{rbuf}; |
1272 |
my $len = sysread $self->{fh}, $$rbuf, $self->{read_size} || 8192, length $$rbuf; |
1273 |
|
1274 |
if ($len > 0) { |
1275 |
$self->{_activity} = AnyEvent->now; |
1276 |
|
1277 |
$self->{filter_r} |
1278 |
? $self->{filter_r}($self, $rbuf) |
1279 |
: $self->{_in_drain} || $self->_drain_rbuf; |
1280 |
|
1281 |
} elsif (defined $len) { |
1282 |
delete $self->{_rw}; |
1283 |
$self->{_eof} = 1; |
1284 |
$self->_drain_rbuf unless $self->{_in_drain}; |
1285 |
|
1286 |
} elsif ($! != EAGAIN && $! != EINTR && $! != WSAEWOULDBLOCK) { |
1287 |
return $self->_error ($!, 1); |
1288 |
} |
1289 |
}); |
1290 |
} |
1291 |
} |
1292 |
|
1293 |
sub _dotls { |
1294 |
my ($self) = @_; |
1295 |
|
1296 |
my $buf; |
1297 |
|
1298 |
if (length $self->{_tls_wbuf}) { |
1299 |
while ((my $len = Net::SSLeay::write ($self->{tls}, $self->{_tls_wbuf})) > 0) { |
1300 |
substr $self->{_tls_wbuf}, 0, $len, ""; |
1301 |
} |
1302 |
} |
1303 |
|
1304 |
if (length ($buf = Net::SSLeay::BIO_read ($self->{_wbio}))) { |
1305 |
$self->{wbuf} .= $buf; |
1306 |
$self->_drain_wbuf; |
1307 |
} |
1308 |
|
1309 |
while (defined ($buf = Net::SSLeay::read ($self->{tls}))) { |
1310 |
if (length $buf) { |
1311 |
$self->{rbuf} .= $buf; |
1312 |
$self->_drain_rbuf unless $self->{_in_drain}; |
1313 |
} else { |
1314 |
# let's treat SSL-eof as we treat normal EOF |
1315 |
$self->{_eof} = 1; |
1316 |
$self->_shutdown; |
1317 |
return; |
1318 |
} |
1319 |
} |
1320 |
|
1321 |
my $err = Net::SSLeay::get_error ($self->{tls}, -1); |
1322 |
|
1323 |
if ($err!= Net::SSLeay::ERROR_WANT_READ ()) { |
1324 |
if ($err == Net::SSLeay::ERROR_SYSCALL ()) { |
1325 |
return $self->_error ($!, 1); |
1326 |
} elsif ($err == Net::SSLeay::ERROR_SSL ()) { |
1327 |
return $self->_error (&Errno::EIO, 1); |
1328 |
} |
1329 |
|
1330 |
# all others are fine for our purposes |
1331 |
} |
1332 |
} |
1333 |
|
1334 |
=item $handle->starttls ($tls[, $tls_ctx]) |
1335 |
|
1336 |
Instead of starting TLS negotiation immediately when the AnyEvent::Handle |
1337 |
object is created, you can also do that at a later time by calling |
1338 |
C<starttls>. |
1339 |
|
1340 |
The first argument is the same as the C<tls> constructor argument (either |
1341 |
C<"connect">, C<"accept"> or an existing Net::SSLeay object). |
1342 |
|
1343 |
The second argument is the optional C<Net::SSLeay::CTX> object that is |
1344 |
used when AnyEvent::Handle has to create its own TLS connection object. |
1345 |
|
1346 |
The TLS connection object will end up in C<< $handle->{tls} >> after this |
1347 |
call and can be used or changed to your liking. Note that the handshake |
1348 |
might have already started when this function returns. |
1349 |
|
1350 |
=cut |
1351 |
|
1352 |
sub starttls { |
1353 |
my ($self, $ssl, $ctx) = @_; |
1354 |
|
1355 |
$self->stoptls; |
1356 |
|
1357 |
if ($ssl eq "accept") { |
1358 |
$ssl = Net::SSLeay::new ($ctx || TLS_CTX ()); |
1359 |
Net::SSLeay::set_accept_state ($ssl); |
1360 |
} elsif ($ssl eq "connect") { |
1361 |
$ssl = Net::SSLeay::new ($ctx || TLS_CTX ()); |
1362 |
Net::SSLeay::set_connect_state ($ssl); |
1363 |
} |
1364 |
|
1365 |
$self->{tls} = $ssl; |
1366 |
|
1367 |
# basically, this is deep magic (because SSL_read should have the same issues) |
1368 |
# but the openssl maintainers basically said: "trust us, it just works". |
1369 |
# (unfortunately, we have to hardcode constants because the abysmally misdesigned |
1370 |
# and mismaintained ssleay-module doesn't even offer them). |
1371 |
# http://www.mail-archive.com/openssl-dev@openssl.org/msg22420.html |
1372 |
Net::SSLeay::CTX_set_mode ($self->{tls}, |
1373 |
(eval { local $SIG{__DIE__}; Net::SSLeay::MODE_ENABLE_PARTIAL_WRITE () } || 1) |
1374 |
| (eval { local $SIG{__DIE__}; Net::SSLeay::MODE_ACCEPT_MOVING_WRITE_BUFFER () } || 2)); |
1375 |
|
1376 |
$self->{_rbio} = Net::SSLeay::BIO_new (Net::SSLeay::BIO_s_mem ()); |
1377 |
$self->{_wbio} = Net::SSLeay::BIO_new (Net::SSLeay::BIO_s_mem ()); |
1378 |
|
1379 |
Net::SSLeay::set_bio ($ssl, $self->{_rbio}, $self->{_wbio}); |
1380 |
|
1381 |
$self->{filter_w} = sub { |
1382 |
$_[0]{_tls_wbuf} .= ${$_[1]}; |
1383 |
&_dotls; |
1384 |
}; |
1385 |
$self->{filter_r} = sub { |
1386 |
Net::SSLeay::BIO_write ($_[0]{_rbio}, ${$_[1]}); |
1387 |
&_dotls; |
1388 |
}; |
1389 |
} |
1390 |
|
1391 |
=item $handle->stoptls |
1392 |
|
1393 |
Destroys the SSL connection, if any. Partial read or write data will be |
1394 |
lost. |
1395 |
|
1396 |
=cut |
1397 |
|
1398 |
sub stoptls { |
1399 |
my ($self) = @_; |
1400 |
|
1401 |
Net::SSLeay::free (delete $self->{tls}) if $self->{tls}; |
1402 |
|
1403 |
delete $self->{_rbio}; |
1404 |
delete $self->{_wbio}; |
1405 |
delete $self->{_tls_wbuf}; |
1406 |
delete $self->{filter_r}; |
1407 |
delete $self->{filter_w}; |
1408 |
} |
1409 |
|
1410 |
sub DESTROY { |
1411 |
my $self = shift; |
1412 |
|
1413 |
$self->stoptls; |
1414 |
|
1415 |
my $linger = exists $self->{linger} ? $self->{linger} : 3600; |
1416 |
|
1417 |
if ($linger && length $self->{wbuf}) { |
1418 |
my $fh = delete $self->{fh}; |
1419 |
my $wbuf = delete $self->{wbuf}; |
1420 |
|
1421 |
my @linger; |
1422 |
|
1423 |
push @linger, AnyEvent->io (fh => $fh, poll => "w", cb => sub { |
1424 |
my $len = syswrite $fh, $wbuf, length $wbuf; |
1425 |
|
1426 |
if ($len > 0) { |
1427 |
substr $wbuf, 0, $len, ""; |
1428 |
} else { |
1429 |
@linger = (); # end |
1430 |
} |
1431 |
}); |
1432 |
push @linger, AnyEvent->timer (after => $linger, cb => sub { |
1433 |
@linger = (); |
1434 |
}); |
1435 |
} |
1436 |
} |
1437 |
|
1438 |
=item AnyEvent::Handle::TLS_CTX |
1439 |
|
1440 |
This function creates and returns the Net::SSLeay::CTX object used by |
1441 |
default for TLS mode. |
1442 |
|
1443 |
The context is created like this: |
1444 |
|
1445 |
Net::SSLeay::load_error_strings; |
1446 |
Net::SSLeay::SSLeay_add_ssl_algorithms; |
1447 |
Net::SSLeay::randomize; |
1448 |
|
1449 |
my $CTX = Net::SSLeay::CTX_new; |
1450 |
|
1451 |
Net::SSLeay::CTX_set_options $CTX, Net::SSLeay::OP_ALL |
1452 |
|
1453 |
=cut |
1454 |
|
1455 |
our $TLS_CTX; |
1456 |
|
1457 |
sub TLS_CTX() { |
1458 |
$TLS_CTX || do { |
1459 |
require Net::SSLeay; |
1460 |
|
1461 |
Net::SSLeay::load_error_strings (); |
1462 |
Net::SSLeay::SSLeay_add_ssl_algorithms (); |
1463 |
Net::SSLeay::randomize (); |
1464 |
|
1465 |
$TLS_CTX = Net::SSLeay::CTX_new (); |
1466 |
|
1467 |
Net::SSLeay::CTX_set_options ($TLS_CTX, Net::SSLeay::OP_ALL ()); |
1468 |
|
1469 |
$TLS_CTX |
1470 |
} |
1471 |
} |
1472 |
|
1473 |
=back |
1474 |
|
1475 |
=head1 SUBCLASSING AnyEvent::Handle |
1476 |
|
1477 |
In many cases, you might want to subclass AnyEvent::Handle. |
1478 |
|
1479 |
To make this easier, a given version of AnyEvent::Handle uses these |
1480 |
conventions: |
1481 |
|
1482 |
=over 4 |
1483 |
|
1484 |
=item * all constructor arguments become object members. |
1485 |
|
1486 |
At least initially, when you pass a C<tls>-argument to the constructor it |
1487 |
will end up in C<< $handle->{tls} >>. Those members might be changed or |
1488 |
mutated later on (for example C<tls> will hold the TLS connection object). |
1489 |
|
1490 |
=item * other object member names are prefixed with an C<_>. |
1491 |
|
1492 |
All object members not explicitly documented (internal use) are prefixed |
1493 |
with an underscore character, so the remaining non-C<_>-namespace is free |
1494 |
for use for subclasses. |
1495 |
|
1496 |
=item * all members not documented here and not prefixed with an underscore |
1497 |
are free to use in subclasses. |
1498 |
|
1499 |
Of course, new versions of AnyEvent::Handle may introduce more "public" |
1500 |
member variables, but thats just life, at least it is documented. |
1501 |
|
1502 |
=back |
1503 |
|
1504 |
=head1 AUTHOR |
1505 |
|
1506 |
Robin Redeker C<< <elmex at ta-sa.org> >>, Marc Lehmann <schmorp@schmorp.de>. |
1507 |
|
1508 |
=cut |
1509 |
|
1510 |
1; # End of AnyEvent::Handle |