ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/AnyEvent/lib/AnyEvent/Handle.pm
(Generate patch)

Comparing AnyEvent/lib/AnyEvent/Handle.pm (file contents):
Revision 1.2 by elmex, Sun Apr 27 17:27:34 2008 UTC vs.
Revision 1.130 by root, Mon Jun 29 21:00:32 2009 UTC

1package AnyEvent::Handle; 1package AnyEvent::Handle;
2 2
3use warnings; 3no warnings;
4use strict; 4use strict qw(subs vars);
5 5
6use AnyEvent; 6use AnyEvent ();
7use IO::Handle; 7use AnyEvent::Util qw(WSAEWOULDBLOCK);
8use Scalar::Util ();
9use Carp ();
10use Fcntl ();
8use Errno qw/EAGAIN EINTR/; 11use Errno qw(EAGAIN EINTR);
9 12
10=head1 NAME 13=head1 NAME
11 14
12AnyEvent::Handle - non-blocking I/O on filehandles via AnyEvent 15AnyEvent::Handle - non-blocking I/O on file handles via AnyEvent
13 16
14=head1 VERSION
15
16Version 0.01
17
18=cut 17=cut
19 18
20our $VERSION = '0.01'; 19our $VERSION = 4.45;
21 20
22=head1 SYNOPSIS 21=head1 SYNOPSIS
23 22
24 use AnyEvent; 23 use AnyEvent;
25 use AnyEvent::Handle; 24 use AnyEvent::Handle;
26 25
27 my $cv = AnyEvent->condvar; 26 my $cv = AnyEvent->condvar;
28 27
29 my $ae_fh = AnyEvent::Handle->new (fh => \*STDIN); 28 my $handle =
30
31 $ae_fh->on_eof (sub { $cv->broadcast });
32
33 $ae_fh->readlines (sub {
34 my ($ae_fh, @lines) = @_;
35 for (@lines) {
36 chomp;
37 print "Line: $_";
38 }
39 });
40
41 # or use the constructor to pass the callback:
42
43 my $ae_fh2 =
44 AnyEvent::Handle->new ( 29 AnyEvent::Handle->new (
45 fh => \*STDIN, 30 fh => \*STDIN,
46 on_eof => sub { 31 on_eof => sub {
47 $cv->broadcast; 32 $cv->send;
48 }, 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
50This module is a helper module to make it easier to do event-based I/O on
51filehandles. For utility functions for doing non-blocking connects and accepts
52on sockets see L<AnyEvent::Util>.
53
54The L<AnyEvent::Intro> tutorial contains some well-documented
55AnyEvent::Handle examples.
56
57In the following, when the documentation refers to of "bytes" then this
58means characters. As sysread and syswrite are used for all I/O, their
59treatment of characters applies to this module as well.
60
61All callbacks will be invoked with the handle object as their first
62argument.
63
64=head1 METHODS
65
66=over 4
67
68=item B<new (%args)>
69
70The constructor supports these arguments (all as key => value pairs).
71
72=over 4
73
74=item fh => $filehandle [MANDATORY]
75
76The filehandle this L<AnyEvent::Handle> object will operate on.
77
78NOTE: The filehandle will be set to non-blocking mode (using
79C<AnyEvent::Util::fh_nonblocking>) by the constructor and needs to stay in
80that mode.
81
82=item on_eof => $cb->($handle)
83
84Set the callback to be called when an end-of-file condition is detected,
85i.e. in the case of a socket, when the other side has closed the
86connection cleanly.
87
88For sockets, this just means that the other side has stopped sending data,
89you can still try to write data, and, in fact, one can return from the EOF
90callback and continue writing data, as only the read part has been shut
91down.
92
93While not mandatory, it is I<highly> recommended to set an EOF callback,
94otherwise you might end up with a closed socket while you are still
95waiting for data.
96
97If an EOF condition has been detected but no C<on_eof> callback has been
98set, then a fatal error will be raised with C<$!> set to <0>.
99
100=item on_error => $cb->($handle, $fatal)
101
102This is the error callback, which is called when, well, some error
103occured, such as not being able to resolve the hostname, failure to
104connect or a read error.
105
106Some errors are fatal (which is indicated by C<$fatal> being true). On
107fatal 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
109errors are an EOF condition with active (but unsatisifable) read watchers
110(C<EPIPE>) or I/O errors.
111
112Non-fatal errors can be retried by simply returning, but it is recommended
113to simply ignore this parameter and instead abondon the handle object
114when this callback is invoked. Examples of non-fatal errors are timeouts
115C<ETIMEDOUT>) or badly-formatted data (C<EBADMSG>).
116
117On callback entrance, the value of C<$!> contains the operating system
118error (or C<ENOSPC>, C<EPIPE>, C<ETIMEDOUT> or C<EBADMSG>).
119
120While not mandatory, it is I<highly> recommended to set this callback, as
121you will not be notified of errors otherwise. The default simply calls
122C<croak>.
123
124=item on_read => $cb->($handle)
125
126This sets the default read callback, which is called when data arrives
127and no read request is in the queue (unlike read queue callbacks, this
128callback will only be called when at least one octet of data is in the
129read buffer).
130
131To access (and remove data from) the read buffer, use the C<< ->rbuf >>
132method or access the C<$handle->{rbuf}> member directly. Note that you
133must not enlarge or modify the read buffer, you can only remove data at
134the beginning from it.
135
136When an EOF condition is detected then AnyEvent::Handle will first try to
137feed all the remaining data to the queued callbacks and C<on_read> before
138calling the C<on_eof> callback. If no progress can be made, then a fatal
139error will be raised (with C<$!> set to C<EPIPE>).
140
141=item on_drain => $cb->($handle)
142
143This sets the callback that is called when the write buffer becomes empty
144(or when the callback is set and the buffer is empty already).
145
146To append to the write buffer, use the C<< ->push_write >> method.
147
148This callback is useful when you don't want to put all of your write data
149into the queue at once, for example, when you want to write the contents
150of some file to the socket you might not want to read the whole file into
151memory and push it into the queue, but instead only read more data from
152the file when the write queue becomes empty.
153
154=item timeout => $fractional_seconds
155
156If non-zero, then this enables an "inactivity" timeout: whenever this many
157seconds pass without a successful read or write on the underlying file
158handle, the C<on_timeout> callback will be invoked (and if that one is
159missing, a non-fatal C<ETIMEDOUT> error will be raised).
160
161Note that timeout processing is also active when you currently do not have
162any outstanding read or write requests: If you plan to keep the connection
163idle then you should disable the timout temporarily or ignore the timeout
164in the C<on_timeout> callback, in which case AnyEvent::Handle will simply
165restart the timeout.
166
167Zero (the default) disables this timeout.
168
169=item on_timeout => $cb->($handle)
170
171Called whenever the inactivity timeout passes. If you return from this
172callback, then the timeout will be reset as if some activity had happened,
173so this condition is not fatal in any way.
174
175=item rbuf_max => <bytes>
176
177If defined, then a fatal error will be raised (with C<$!> set to C<ENOSPC>)
178when the read buffer ever (strictly) exceeds this size. This is useful to
179avoid some forms of denial-of-service attacks.
180
181For example, a server accepting connections from untrusted sources should
182be configured to accept only so-and-so much data that it cannot act on
183(for example, when expecting a line, an attacker could send an unlimited
184amount of data without a callback ever being called as long as the line
185isn't finished).
186
187=item autocork => <boolean>
188
189When disabled (the default), then C<push_write> will try to immediately
190write the data to the handle, if possible. This avoids having to register
191a write watcher and wait for the next event loop iteration, but can
192be inefficient if you write multiple small chunks (on the wire, this
193disadvantage is usually avoided by your kernel's nagle algorithm, see
194C<no_delay>, but this option can save costly syscalls).
195
196When enabled, then writes will always be queued till the next event loop
197iteration. This is efficient when you do many small writes per iteration,
198but less efficient when you do a single write only per iteration (or when
199the write buffer often is full). It also increases write latency.
200
201=item no_delay => <boolean>
202
203When doing small writes on sockets, your operating system kernel might
204wait a bit for more data before actually sending it out. This is called
205the Nagle algorithm, and usually it is beneficial.
206
207In some situations you want as low a delay as possible, which can be
208accomplishd by setting this option to a true value.
209
210The default is your opertaing system's default behaviour (most likely
211enabled), this option explicitly enables or disables it, if possible.
212
213=item read_size => <bytes>
214
215The default read block size (the amount of bytes this module will
216try to read during each loop iteration, which affects memory
217requirements). Default: C<8192>.
218
219=item low_water_mark => <bytes>
220
221Sets the amount of bytes (default: C<0>) that make up an "empty" write
222buffer: If the write reaches this size or gets even samller it is
223considered empty.
224
225Sometimes it can be beneficial (for performance reasons) to add data to
226the write buffer before it is fully drained, but this is a rare case, as
227the operating system kernel usually buffers data as well, so the default
228is good in almost all cases.
229
230=item linger => <seconds>
231
232If non-zero (default: C<3600>), then the destructor of the
233AnyEvent::Handle object will check whether there is still outstanding
234write data and will install a watcher that will write this data to the
235socket. No errors will be reported (this mostly matches how the operating
236system treats outstanding data at socket close time).
237
238This will not work for partial TLS data that could not be encoded
239yet. This data will be lost. Calling the C<stoptls> method in time might
240help.
241
242=item tls => "accept" | "connect" | Net::SSLeay::SSL object
243
244When this parameter is given, it enables TLS (SSL) mode, that means
245AnyEvent will start a TLS handshake as soon as the conenction has been
246established and will transparently encrypt/decrypt data afterwards.
247
248TLS mode requires Net::SSLeay to be installed (it will be loaded
249automatically when you try to create a TLS handle): this module doesn't
250have a dependency on that module, so if your module requires it, you have
251to add the dependency yourself.
252
253Unlike TCP, TLS has a server and client side: for the TLS server side, use
254C<accept>, and for the TLS client side of a connection, use C<connect>
255mode.
256
257You can also provide your own TLS connection object, but you have
258to make sure that you call either C<Net::SSLeay::set_connect_state>
259or C<Net::SSLeay::set_accept_state> on it before you pass it to
260AnyEvent::Handle.
261
262B<IMPORTANT:> since Net::SSLeay "objects" are really only integers,
263passing in the wrong integer will lead to certain crash. This most often
264happens when one uses a stylish C<< tls => 1 >> and is surprised about the
265segmentation fault.
266
267See the C<< ->starttls >> method for when need to start TLS negotiation later.
268
269=item tls_ctx => $ssl_ctx
270
271Use the given C<Net::SSLeay::CTX> object to create the new TLS connection
272(unless a connection object was specified directly). If this parameter is
273missing, then AnyEvent::Handle will use C<AnyEvent::Handle::TLS_CTX>.
274
275=item json => JSON or JSON::XS object
276
277This is the json coder object used by the C<json> read and write types.
278
279If you don't supply it, then AnyEvent::Handle will create and use a
280suitable one (on demand), which will write and expect UTF-8 encoded JSON
281texts.
282
283Note that you are responsible to depend on the JSON module if you want to
284use this functionality, as AnyEvent does not have a dependency itself.
285
286=back
287
288=cut
289
290sub new {
291 my $class = shift;
292
293 my $self = bless { @_ }, $class;
294
295 $self->{fh} or Carp::croak "mandatory argument fh is missing";
296
297 AnyEvent::Util::fh_nonblocking $self->{fh}, 1;
298
299 $self->starttls (delete $self->{tls}, delete $self->{tls_ctx})
300 if $self->{tls};
301
302 $self->{_activity} = AnyEvent->now;
303 $self->_timeout;
304
305 $self->on_drain (delete $self->{on_drain}) if exists $self->{on_drain};
306 $self->no_delay (delete $self->{no_delay}) if exists $self->{no_delay};
307
308 $self->start_read
309 if $self->{on_read};
310
311 $self
312}
313
314sub _shutdown {
315 my ($self) = @_;
316
317 delete @$self{qw(_tw _rw _ww fh rbuf wbuf on_read _queue)};
318
319 &_freetls;
320}
321
322sub _error {
323 my ($self, $errno, $fatal) = @_;
324
325 $self->_shutdown
326 if $fatal;
327
328 $! = $errno;
329
330 if ($self->{on_error}) {
331 $self->{on_error}($self, $fatal);
332 } elsif ($self->{fh}) {
333 Carp::croak "AnyEvent::Handle uncaught error: $!";
334 }
335}
336
337=item $fh = $handle->fh
338
339This method returns the file handle used to create the L<AnyEvent::Handle> object.
340
341=cut
342
343sub fh { $_[0]{fh} }
344
345=item $handle->on_error ($cb)
346
347Replace the current C<on_error> callback (see the C<on_error> constructor argument).
348
349=cut
350
351sub on_error {
352 $_[0]{on_error} = $_[1];
353}
354
355=item $handle->on_eof ($cb)
356
357Replace the current C<on_eof> callback (see the C<on_eof> constructor argument).
358
359=cut
360
361sub on_eof {
362 $_[0]{on_eof} = $_[1];
363}
364
365=item $handle->on_timeout ($cb)
366
367Replace the current C<on_timeout> callback, or disables the callback (but
368not the timeout) if C<$cb> = C<undef>. See the C<timeout> constructor
369argument and method.
370
371=cut
372
373sub on_timeout {
374 $_[0]{on_timeout} = $_[1];
375}
376
377=item $handle->autocork ($boolean)
378
379Enables or disables the current autocork behaviour (see C<autocork>
380constructor argument). Changes will only take effect on the next write.
381
382=cut
383
384sub autocork {
385 $_[0]{autocork} = $_[1];
386}
387
388=item $handle->no_delay ($boolean)
389
390Enables or disables the C<no_delay> setting (see constructor argument of
391the same name for details).
392
393=cut
394
395sub no_delay {
396 $_[0]{no_delay} = $_[1];
397
398 eval {
399 local $SIG{__DIE__};
400 setsockopt $_[0]{fh}, &Socket::IPPROTO_TCP, &Socket::TCP_NODELAY, int $_[1];
401 };
402}
403
404#############################################################################
405
406=item $handle->timeout ($seconds)
407
408Configures (or disables) the inactivity timeout.
409
410=cut
411
412sub timeout {
413 my ($self, $timeout) = @_;
414
415 $self->{timeout} = $timeout;
416 $self->_timeout;
417}
418
419# reset the timeout watcher, as neccessary
420# also check for time-outs
421sub _timeout {
422 my ($self) = @_;
423
424 if ($self->{timeout}) {
425 my $NOW = AnyEvent->now;
426
427 # when would the timeout trigger?
428 my $after = $self->{_activity} + $self->{timeout} - $NOW;
429
430 # now or in the past already?
431 if ($after <= 0) {
432 $self->{_activity} = $NOW;
433
434 if ($self->{on_timeout}) {
435 $self->{on_timeout}($self);
436 } else {
437 $self->_error (&Errno::ETIMEDOUT);
438 }
439
440 # callback could have changed timeout value, optimise
441 return unless $self->{timeout};
442
443 # calculate new after
444 $after = $self->{timeout};
445 }
446
447 Scalar::Util::weaken $self;
448 return unless $self; # ->error could have destroyed $self
449
450 $self->{_tw} ||= AnyEvent->timer (after => $after, cb => sub {
451 delete $self->{_tw};
452 $self->_timeout;
453 });
454 } else {
455 delete $self->{_tw};
456 }
457}
458
459#############################################################################
460
461=back
462
463=head2 WRITE QUEUE
464
465AnyEvent::Handle manages two queues per handle, one for writing and one
466for reading.
467
468The write queue is very simple: you can add data to its end, and
469AnyEvent::Handle will automatically try to get rid of it for you.
470
471When data could be written and the write buffer is shorter then the low
472water mark, the C<on_drain> callback will be invoked.
473
474=over 4
475
476=item $handle->on_drain ($cb)
477
478Sets the C<on_drain> callback or clears it (see the description of
479C<on_drain> in the constructor).
480
481=cut
482
483sub on_drain {
484 my ($self, $cb) = @_;
485
486 $self->{on_drain} = $cb;
487
488 $cb->($self)
489 if $cb && $self->{low_water_mark} >= (length $self->{wbuf}) + (length $self->{_tls_wbuf});
490}
491
492=item $handle->push_write ($data)
493
494Queues the given scalar to be written. You can push as much data as you
495want (only limited by the available memory), as C<AnyEvent::Handle>
496buffers it independently of the kernel.
497
498=cut
499
500sub _drain_wbuf {
501 my ($self) = @_;
502
503 if (!$self->{_ww} && length $self->{wbuf}) {
504
505 Scalar::Util::weaken $self;
506
507 my $cb = sub {
508 my $len = syswrite $self->{fh}, $self->{wbuf};
509
510 if ($len >= 0) {
511 substr $self->{wbuf}, 0, $len, "";
512
513 $self->{_activity} = AnyEvent->now;
514
515 $self->{on_drain}($self)
516 if $self->{low_water_mark} >= (length $self->{wbuf}) + (length $self->{_tls_wbuf})
517 && $self->{on_drain};
518
519 delete $self->{_ww} unless length $self->{wbuf};
520 } elsif ($! != EAGAIN && $! != EINTR && $! != WSAEWOULDBLOCK) {
521 $self->_error ($!, 1);
522 }
523 };
524
525 # try to write data immediately
526 $cb->() unless $self->{autocork};
527
528 # if still data left in wbuf, we need to poll
529 $self->{_ww} = AnyEvent->io (fh => $self->{fh}, poll => "w", cb => $cb)
530 if length $self->{wbuf};
531 };
532}
533
534our %WH;
535
536sub register_write_type($$) {
537 $WH{$_[0]} = $_[1];
538}
539
540sub push_write {
541 my $self = shift;
542
543 if (@_ > 1) {
544 my $type = shift;
545
546 @_ = ($WH{$type} or Carp::croak "unsupported type passed to AnyEvent::Handle::push_write")
547 ->($self, @_);
548 }
549
550 if ($self->{tls}) {
551 $self->{_tls_wbuf} .= $_[0];
552
553 &_dotls ($self);
554 } else {
555 $self->{wbuf} .= $_[0];
556 $self->_drain_wbuf;
557 }
558}
559
560=item $handle->push_write (type => @args)
561
562Instead of formatting your data yourself, you can also let this module do
563the job by specifying a type and type-specific arguments.
564
565Predefined types are (if you have ideas for additional types, feel free to
566drop by and tell us):
567
568=over 4
569
570=item netstring => $string
571
572Formats the given value as netstring
573(http://cr.yp.to/proto/netstrings.txt, this is not a recommendation to use them).
574
575=cut
576
577register_write_type netstring => sub {
578 my ($self, $string) = @_;
579
580 (length $string) . ":$string,"
581};
582
583=item packstring => $format, $data
584
585An octet string prefixed with an encoded length. The encoding C<$format>
586uses the same format as a Perl C<pack> format, but must specify a single
587integer only (only one of C<cCsSlLqQiInNvVjJw> is allowed, plus an
588optional C<!>, C<< < >> or C<< > >> modifier).
589
590=cut
591
592register_write_type packstring => sub {
593 my ($self, $format, $string) = @_;
594
595 pack "$format/a*", $string
596};
597
598=item json => $array_or_hashref
599
600Encodes the given hash or array reference into a JSON object. Unless you
601provide your own JSON object, this means it will be encoded to JSON text
602in UTF-8.
603
604JSON objects (and arrays) are self-delimiting, so you can write JSON at
605one end of a handle and read them at the other end without using any
606additional framing.
607
608The generated JSON text is guaranteed not to contain any newlines: While
609this module doesn't need delimiters after or between JSON texts to be
610able to read them, many other languages depend on that.
611
612A simple RPC protocol that interoperates easily with others is to send
613JSON arrays (or objects, although arrays are usually the better choice as
614they mimic how function argument passing works) and a newline after each
615JSON text:
616
617 $handle->push_write (json => ["method", "arg1", "arg2"]); # whatever
618 $handle->push_write ("\012");
619
620An AnyEvent::Handle receiver would simply use the C<json> read type and
621rely on the fact that the newline will be skipped as leading whitespace:
622
623 $handle->push_read (json => sub { my $array = $_[1]; ... });
624
625Other languages could read single lines terminated by a newline and pass
626this line into their JSON decoder of choice.
627
628=cut
629
630register_write_type json => sub {
631 my ($self, $ref) = @_;
632
633 require JSON;
634
635 $self->{json} ? $self->{json}->encode ($ref)
636 : JSON::encode_json ($ref)
637};
638
639=item storable => $reference
640
641Freezes the given reference using L<Storable> and writes it to the
642handle. Uses the C<nfreeze> format.
643
644=cut
645
646register_write_type storable => sub {
647 my ($self, $ref) = @_;
648
649 require Storable;
650
651 pack "w/a*", Storable::nfreeze ($ref)
652};
653
654=back
655
656=item AnyEvent::Handle::register_write_type type => $coderef->($handle, @args)
657
658This function (not method) lets you add your own types to C<push_write>.
659Whenever the given C<type> is used, C<push_write> will invoke the code
660reference with the handle object and the remaining arguments.
661
662The code reference is supposed to return a single octet string that will
663be appended to the write buffer.
664
665Note that this is a function, and all types registered this way will be
666global, so try to use unique names.
667
668=cut
669
670#############################################################################
671
672=back
673
674=head2 READ QUEUE
675
676AnyEvent::Handle manages two queues per handle, one for writing and one
677for reading.
678
679The read queue is more complex than the write queue. It can be used in two
680ways, the "simple" way, using only C<on_read> and the "complex" way, using
681a queue.
682
683In the simple case, you just install an C<on_read> callback and whenever
684new data arrives, it will be called. You can then remove some data (if
685enough is there) from the read buffer (C<< $handle->rbuf >>). Or you cna
686leave the data there if you want to accumulate more (e.g. when only a
687partial message has been received so far).
688
689In the more complex case, you want to queue multiple callbacks. In this
690case, AnyEvent::Handle will call the first queued callback each time new
691data arrives (also the first time it is queued) and removes it when it has
692done its job (see C<push_read>, below).
693
694This way you can, for example, push three line-reads, followed by reading
695a chunk of data, and AnyEvent::Handle will execute them in order.
696
697Example 1: EPP protocol parser. EPP sends 4 byte length info, followed by
698the specified number of bytes which give an XML datagram.
699
700 # in the default state, expect some header bytes
701 $handle->on_read (sub {
702 # some data is here, now queue the length-header-read (4 octets)
703 shift->unshift_read (chunk => 4, sub {
704 # header arrived, decode
705 my $len = unpack "N", $_[1];
706
707 # now read the payload
708 shift->unshift_read (chunk => $len, sub {
709 my $xml = $_[1];
710 # handle xml
711 });
712 });
713 });
714
715Example 2: Implement a client for a protocol that replies either with "OK"
716and another line or "ERROR" for the first request that is sent, and 64
717bytes for the second request. Due to the availability of a queue, we can
718just pipeline sending both requests and manipulate the queue as necessary
719in the callbacks.
720
721When the first callback is called and sees an "OK" response, it will
722C<unshift> another line-read. This line-read will be queued I<before> the
72364-byte chunk callback.
724
725 # request one, returns either "OK + extra line" or "ERROR"
726 $handle->push_write ("request 1\015\012");
727
728 # we expect "ERROR" or "OK" as response, so push a line read
729 $handle->push_read (line => sub {
730 # if we got an "OK", we have to _prepend_ another line,
731 # so it will be read before the second request reads its 64 bytes
732 # which are already in the queue when this callback is called
733 # we don't do this in case we got an error
734 if ($_[1] eq "OK") {
49 on_readline => sub { 735 $_[0]->unshift_read (line => sub {
50 my ($ae_fh, @lines) = @_; 736 my $response = $_[1];
51 for (@lines) { 737 ...
52 chomp; 738 });
53 print "Line: $_"; 739 }
740 });
741
742 # request two, simply returns 64 octets
743 $handle->push_write ("request 2\015\012");
744
745 # simply read 64 bytes, always
746 $handle->push_read (chunk => 64, sub {
747 my $response = $_[1];
748 ...
749 });
750
751=over 4
752
753=cut
754
755sub _drain_rbuf {
756 my ($self) = @_;
757
758 local $self->{_in_drain} = 1;
759
760 if (
761 defined $self->{rbuf_max}
762 && $self->{rbuf_max} < length $self->{rbuf}
763 ) {
764 $self->_error (&Errno::ENOSPC, 1), return;
765 }
766
767 while () {
768 # we need to use a separate tls read buffer, as we must not receive data while
769 # we are draining the buffer, and this can only happen with TLS.
770 $self->{rbuf} .= delete $self->{_tls_rbuf} if exists $self->{_tls_rbuf};
771
772 my $len = length $self->{rbuf};
773
774 if (my $cb = shift @{ $self->{_queue} }) {
775 unless ($cb->($self)) {
776 if ($self->{_eof}) {
777 # no progress can be made (not enough data and no data forthcoming)
778 $self->_error (&Errno::EPIPE, 1), return;
54 } 779 }
780
781 unshift @{ $self->{_queue} }, $cb;
782 last;
55 } 783 }
56 );
57
58 $cv->wait;
59
60=head1 DESCRIPTION
61
62This module is a helper module to make it easier to do non-blocking I/O
63on filehandles (and sockets, see L<AnyEvent::Socket>).
64
65The event loop is provided by L<AnyEvent>.
66
67=head1 METHODS
68
69=over 4
70
71=item B<new (%args)>
72
73The constructor has these arguments:
74
75=over 4
76
77=item fh => $filehandle
78
79The filehandle this L<AnyEvent::Handle> object will operate on.
80
81NOTE: The filehandle will be set to non-blocking.
82
83=item read_block_size => $size
84
85The default read block size use for reads via the C<on_read>
86method.
87
88=item on_read => $cb
89
90=item on_eof => $cb
91
92=item on_error => $cb
93
94These are shortcuts, that will call the corresponding method and set the callback to C<$cb>.
95
96=item on_readline => $cb
97
98The C<readlines> method is called with the default seperator and C<$cb> as callback
99for you.
100
101=back
102
103=cut
104
105sub new {
106 my $this = shift;
107 my $class = ref($this) || $this;
108 my $self = {
109 read_block_size => 4096,
110 rbuf => '',
111 @_
112 };
113 bless $self, $class;
114
115 $self->{fh}->blocking (0) if $self->{fh};
116
117 if ($self->{on_read}) {
118 $self->on_read ($self->{on_read});
119
120 } elsif ($self->{on_readline}) { 784 } elsif ($self->{on_read}) {
121 $self->readlines ($self->{on_readline}); 785 last unless $len;
122 786
787 $self->{on_read}($self);
788
789 if (
790 $len == length $self->{rbuf} # if no data has been consumed
791 && !@{ $self->{_queue} } # and the queue is still empty
792 && $self->{on_read} # but we still have on_read
793 ) {
794 # no further data will arrive
795 # so no progress can be made
796 $self->_error (&Errno::EPIPE, 1), return
797 if $self->{_eof};
798
799 last; # more data might arrive
800 }
801 } else {
802 # read side becomes idle
803 delete $self->{_rw} unless $self->{tls};
804 last;
805 }
806 }
807
123 } elsif ($self->{on_eof}) { 808 if ($self->{_eof}) {
124 $self->on_eof ($self->{on_eof});
125
126 } elsif ($self->{on_error}) { 809 if ($self->{on_eof}) {
127 $self->on_eof ($self->{on_error}); 810 $self->{on_eof}($self)
811 } else {
812 $self->_error (0, 1);
813 }
128 } 814 }
129 815
130 return $self 816 # may need to restart read watcher
817 unless ($self->{_rw}) {
818 $self->start_read
819 if $self->{on_read} || @{ $self->{_queue} };
820 }
131} 821}
132 822
133=item B<fh> 823=item $handle->on_read ($cb)
134 824
135This method returns the filehandle of the L<AnyEvent::Handle> object. 825This replaces the currently set C<on_read> callback, or clears it (when
136 826the new callback is C<undef>). See the description of C<on_read> in the
137=cut 827constructor.
138
139sub fh { $_[0]->{fh} }
140
141=item B<on_read ($callback)>
142
143This method installs a C<$callback> that will be called
144when new data arrived. You can access the read buffer via the C<rbuf>
145method (see below).
146
147The first argument of the C<$callback> will be the L<AnyEvent::Handle> object.
148 828
149=cut 829=cut
150 830
151sub on_read { 831sub on_read {
152 my ($self, $cb) = @_; 832 my ($self, $cb) = @_;
833
153 $self->{on_read} = $cb; 834 $self->{on_read} = $cb;
835 $self->_drain_rbuf if $cb && !$self->{_in_drain};
836}
154 837
155 unless (defined $self->{on_read}) { 838=item $handle->rbuf
156 delete $self->{on_read_w}; 839
840Returns the read buffer (as a modifiable lvalue).
841
842You can access the read buffer directly as the C<< ->{rbuf} >>
843member, if you want. However, the only operation allowed on the
844read buffer (apart from looking at it) is removing data from its
845beginning. Otherwise modifying or appending to it is not allowed and will
846lead to hard-to-track-down bugs.
847
848NOTE: The read buffer should only be used or modified if the C<on_read>,
849C<push_read> or C<unshift_read> methods are used. The other read methods
850automatically manage the read buffer.
851
852=cut
853
854sub rbuf : lvalue {
855 $_[0]{rbuf}
856}
857
858=item $handle->push_read ($cb)
859
860=item $handle->unshift_read ($cb)
861
862Append the given callback to the end of the queue (C<push_read>) or
863prepend it (C<unshift_read>).
864
865The callback is called each time some additional read data arrives.
866
867It must check whether enough data is in the read buffer already.
868
869If not enough data is available, it must return the empty list or a false
870value, in which case it will be called repeatedly until enough data is
871available (or an error condition is detected).
872
873If enough data was available, then the callback must remove all data it is
874interested in (which can be none at all) and return a true value. After returning
875true, it will be removed from the queue.
876
877=cut
878
879our %RH;
880
881sub register_read_type($$) {
882 $RH{$_[0]} = $_[1];
883}
884
885sub push_read {
886 my $self = shift;
887 my $cb = pop;
888
889 if (@_) {
890 my $type = shift;
891
892 $cb = ($RH{$type} or Carp::croak "unsupported type passed to AnyEvent::Handle::push_read")
893 ->($self, $cb, @_);
894 }
895
896 push @{ $self->{_queue} }, $cb;
897 $self->_drain_rbuf unless $self->{_in_drain};
898}
899
900sub unshift_read {
901 my $self = shift;
902 my $cb = pop;
903
904 if (@_) {
905 my $type = shift;
906
907 $cb = ($RH{$type} or Carp::croak "unsupported type passed to AnyEvent::Handle::unshift_read")
908 ->($self, $cb, @_);
909 }
910
911
912 unshift @{ $self->{_queue} }, $cb;
913 $self->_drain_rbuf unless $self->{_in_drain};
914}
915
916=item $handle->push_read (type => @args, $cb)
917
918=item $handle->unshift_read (type => @args, $cb)
919
920Instead of providing a callback that parses the data itself you can chose
921between a number of predefined parsing formats, for chunks of data, lines
922etc.
923
924Predefined types are (if you have ideas for additional types, feel free to
925drop by and tell us):
926
927=over 4
928
929=item chunk => $octets, $cb->($handle, $data)
930
931Invoke the callback only once C<$octets> bytes have been read. Pass the
932data read to the callback. The callback will never be called with less
933data.
934
935Example: read 2 bytes.
936
937 $handle->push_read (chunk => 2, sub {
938 warn "yay ", unpack "H*", $_[1];
939 });
940
941=cut
942
943register_read_type chunk => sub {
944 my ($self, $cb, $len) = @_;
945
946 sub {
947 $len <= length $_[0]{rbuf} or return;
948 $cb->($_[0], substr $_[0]{rbuf}, 0, $len, "");
949 1
950 }
951};
952
953=item line => [$eol, ]$cb->($handle, $line, $eol)
954
955The callback will be called only once a full line (including the end of
956line marker, C<$eol>) has been read. This line (excluding the end of line
957marker) will be passed to the callback as second argument (C<$line>), and
958the end of line marker as the third argument (C<$eol>).
959
960The end of line marker, C<$eol>, can be either a string, in which case it
961will be interpreted as a fixed record end marker, or it can be a regex
962object (e.g. created by C<qr>), in which case it is interpreted as a
963regular expression.
964
965The end of line marker argument C<$eol> is optional, if it is missing (NOT
966undef), then C<qr|\015?\012|> is used (which is good for most internet
967protocols).
968
969Partial lines at the end of the stream will never be returned, as they are
970not marked by the end of line marker.
971
972=cut
973
974register_read_type line => sub {
975 my ($self, $cb, $eol) = @_;
976
977 if (@_ < 3) {
978 # this is more than twice as fast as the generic code below
979 sub {
980 $_[0]{rbuf} =~ s/^([^\015\012]*)(\015?\012)// or return;
981
982 $cb->($_[0], $1, $2);
983 1
984 }
985 } else {
986 $eol = quotemeta $eol unless ref $eol;
987 $eol = qr|^(.*?)($eol)|s;
988
989 sub {
990 $_[0]{rbuf} =~ s/$eol// or return;
991
992 $cb->($_[0], $1, $2);
993 1
994 }
995 }
996};
997
998=item regex => $accept[, $reject[, $skip], $cb->($handle, $data)
999
1000Makes a regex match against the regex object C<$accept> and returns
1001everything up to and including the match.
1002
1003Example: read a single line terminated by '\n'.
1004
1005 $handle->push_read (regex => qr<\n>, sub { ... });
1006
1007If C<$reject> is given and not undef, then it determines when the data is
1008to be rejected: it is matched against the data when the C<$accept> regex
1009does not match and generates an C<EBADMSG> error when it matches. This is
1010useful to quickly reject wrong data (to avoid waiting for a timeout or a
1011receive buffer overflow).
1012
1013Example: expect a single decimal number followed by whitespace, reject
1014anything else (not the use of an anchor).
1015
1016 $handle->push_read (regex => qr<^[0-9]+\s>, qr<[^0-9]>, sub { ... });
1017
1018If C<$skip> is given and not C<undef>, then it will be matched against
1019the receive buffer when neither C<$accept> nor C<$reject> match,
1020and everything preceding and including the match will be accepted
1021unconditionally. This is useful to skip large amounts of data that you
1022know cannot be matched, so that the C<$accept> or C<$reject> regex do not
1023have to start matching from the beginning. This is purely an optimisation
1024and is usually worth only when you expect more than a few kilobytes.
1025
1026Example: expect a http header, which ends at C<\015\012\015\012>. Since we
1027expect the header to be very large (it isn't in practise, but...), we use
1028a skip regex to skip initial portions. The skip regex is tricky in that
1029it only accepts something not ending in either \015 or \012, as these are
1030required for the accept regex.
1031
1032 $handle->push_read (regex =>
1033 qr<\015\012\015\012>,
1034 undef, # no reject
1035 qr<^.*[^\015\012]>,
1036 sub { ... });
1037
1038=cut
1039
1040register_read_type regex => sub {
1041 my ($self, $cb, $accept, $reject, $skip) = @_;
1042
1043 my $data;
1044 my $rbuf = \$self->{rbuf};
1045
1046 sub {
1047 # accept
1048 if ($$rbuf =~ $accept) {
1049 $data .= substr $$rbuf, 0, $+[0], "";
1050 $cb->($self, $data);
157 return; 1051 return 1;
1052 }
1053
1054 # reject
1055 if ($reject && $$rbuf =~ $reject) {
1056 $self->_error (&Errno::EBADMSG);
1057 }
1058
1059 # skip
1060 if ($skip && $$rbuf =~ $skip) {
1061 $data .= substr $$rbuf, 0, $+[0], "";
1062 }
1063
1064 ()
158 } 1065 }
159 1066};
160 $self->{on_read_w} = 1067
161 AnyEvent->io (poll => 'r', fh => $self->{fh}, cb => sub { 1068=item netstring => $cb->($handle, $string)
162 #d# warn "READ:[$self->{read_size}] $self->{read_block_size} : ".length ($self->{rbuf})."\n"; 1069
163 my $rbuf_len = length $self->{rbuf}; 1070A netstring (http://cr.yp.to/proto/netstrings.txt, this is not an endorsement).
164 my $l; 1071
165 if (defined $self->{read_size}) { 1072Throws an error with C<$!> set to EBADMSG on format violations.
166 $l = sysread $self->{fh}, $self->{rbuf}, 1073
167 ($self->{read_size} - $rbuf_len), $rbuf_len; 1074=cut
168 } else { 1075
169 $l = sysread $self->{fh}, $self->{rbuf}, $self->{read_block_size}, $rbuf_len; 1076register_read_type netstring => sub {
1077 my ($self, $cb) = @_;
1078
1079 sub {
1080 unless ($_[0]{rbuf} =~ s/^(0|[1-9][0-9]*)://) {
1081 if ($_[0]{rbuf} =~ /[^0-9]/) {
1082 $self->_error (&Errno::EBADMSG);
170 } 1083 }
171 #d# warn "READL $l [$self->{rbuf}]\n"; 1084 return;
1085 }
172 1086
173 if (not defined $l) { 1087 my $len = $1;
174 return if $! == EAGAIN || $! == EINTR;
175 $self->{on_error}->($self) if $self->{on_error};
176 delete $self->{on_read_w};
177 1088
178 } elsif ($l == 0) { 1089 $self->unshift_read (chunk => $len, sub {
179 $self->{on_eof}->($self) if $self->{on_eof}; 1090 my $string = $_[1];
180 delete $self->{on_read_w}; 1091 $_[0]->unshift_read (chunk => 1, sub {
181 1092 if ($_[1] eq ",") {
1093 $cb->($_[0], $string);
182 } else { 1094 } else {
183 $self->{on_read}->($self); 1095 $self->_error (&Errno::EBADMSG);
1096 }
1097 });
1098 });
1099
1100 1
1101 }
1102};
1103
1104=item packstring => $format, $cb->($handle, $string)
1105
1106An octet string prefixed with an encoded length. The encoding C<$format>
1107uses the same format as a Perl C<pack> format, but must specify a single
1108integer only (only one of C<cCsSlLqQiInNvVjJw> is allowed, plus an
1109optional C<!>, C<< < >> or C<< > >> modifier).
1110
1111For example, DNS over TCP uses a prefix of C<n> (2 octet network order),
1112EPP uses a prefix of C<N> (4 octtes).
1113
1114Example: read a block of data prefixed by its length in BER-encoded
1115format (very efficient).
1116
1117 $handle->push_read (packstring => "w", sub {
1118 my ($handle, $data) = @_;
1119 });
1120
1121=cut
1122
1123register_read_type packstring => sub {
1124 my ($self, $cb, $format) = @_;
1125
1126 sub {
1127 # when we can use 5.10 we can use ".", but for 5.8 we use the re-pack method
1128 defined (my $len = eval { unpack $format, $_[0]{rbuf} })
1129 or return;
1130
1131 $format = length pack $format, $len;
1132
1133 # bypass unshift if we already have the remaining chunk
1134 if ($format + $len <= length $_[0]{rbuf}) {
1135 my $data = substr $_[0]{rbuf}, $format, $len;
1136 substr $_[0]{rbuf}, 0, $format + $len, "";
1137 $cb->($_[0], $data);
1138 } else {
1139 # remove prefix
1140 substr $_[0]{rbuf}, 0, $format, "";
1141
1142 # read remaining chunk
1143 $_[0]->unshift_read (chunk => $len, $cb);
1144 }
1145
1146 1
1147 }
1148};
1149
1150=item json => $cb->($handle, $hash_or_arrayref)
1151
1152Reads a JSON object or array, decodes it and passes it to the
1153callback. When a parse error occurs, an C<EBADMSG> error will be raised.
1154
1155If a C<json> object was passed to the constructor, then that will be used
1156for the final decode, otherwise it will create a JSON coder expecting UTF-8.
1157
1158This read type uses the incremental parser available with JSON version
11592.09 (and JSON::XS version 2.2) and above. You have to provide a
1160dependency on your own: this module will load the JSON module, but
1161AnyEvent does not depend on it itself.
1162
1163Since JSON texts are fully self-delimiting, the C<json> read and write
1164types are an ideal simple RPC protocol: just exchange JSON datagrams. See
1165the C<json> write type description, above, for an actual example.
1166
1167=cut
1168
1169register_read_type json => sub {
1170 my ($self, $cb) = @_;
1171
1172 require JSON;
1173
1174 my $data;
1175 my $rbuf = \$self->{rbuf};
1176
1177 my $json = $self->{json} ||= JSON->new->utf8;
1178
1179 sub {
1180 my $ref = eval { $json->incr_parse ($self->{rbuf}) };
1181
1182 if ($ref) {
1183 $self->{rbuf} = $json->incr_text;
1184 $json->incr_text = "";
1185 $cb->($self, $ref);
1186
1187 1
1188 } elsif ($@) {
1189 # error case
1190 $json->incr_skip;
1191
1192 $self->{rbuf} = $json->incr_text;
1193 $json->incr_text = "";
1194
1195 $self->_error (&Errno::EBADMSG);
1196
1197 ()
1198 } else {
1199 $self->{rbuf} = "";
1200
1201 ()
1202 }
1203 }
1204};
1205
1206=item storable => $cb->($handle, $ref)
1207
1208Deserialises a L<Storable> frozen representation as written by the
1209C<storable> write type (BER-encoded length prefix followed by nfreeze'd
1210data).
1211
1212Raises C<EBADMSG> error if the data could not be decoded.
1213
1214=cut
1215
1216register_read_type storable => sub {
1217 my ($self, $cb) = @_;
1218
1219 require Storable;
1220
1221 sub {
1222 # when we can use 5.10 we can use ".", but for 5.8 we use the re-pack method
1223 defined (my $len = eval { unpack "w", $_[0]{rbuf} })
1224 or return;
1225
1226 my $format = length pack "w", $len;
1227
1228 # bypass unshift if we already have the remaining chunk
1229 if ($format + $len <= length $_[0]{rbuf}) {
1230 my $data = substr $_[0]{rbuf}, $format, $len;
1231 substr $_[0]{rbuf}, 0, $format + $len, "";
1232 $cb->($_[0], Storable::thaw ($data));
1233 } else {
1234 # remove prefix
1235 substr $_[0]{rbuf}, 0, $format, "";
1236
1237 # read remaining chunk
1238 $_[0]->unshift_read (chunk => $len, sub {
1239 if (my $ref = eval { Storable::thaw ($_[1]) }) {
1240 $cb->($_[0], $ref);
1241 } else {
1242 $self->_error (&Errno::EBADMSG);
1243 }
1244 });
1245 }
1246
1247 1
1248 }
1249};
1250
1251=back
1252
1253=item AnyEvent::Handle::register_read_type type => $coderef->($handle, $cb, @args)
1254
1255This function (not method) lets you add your own types to C<push_read>.
1256
1257Whenever the given C<type> is used, C<push_read> will invoke the code
1258reference with the handle object, the callback and the remaining
1259arguments.
1260
1261The code reference is supposed to return a callback (usually a closure)
1262that works as a plain read callback (see C<< ->push_read ($cb) >>).
1263
1264It should invoke the passed callback when it is done reading (remember to
1265pass C<$handle> as first argument as all other callbacks do that).
1266
1267Note that this is a function, and all types registered this way will be
1268global, so try to use unique names.
1269
1270For examples, see the source of this module (F<perldoc -m AnyEvent::Handle>,
1271search for C<register_read_type>)).
1272
1273=item $handle->stop_read
1274
1275=item $handle->start_read
1276
1277In rare cases you actually do not want to read anything from the
1278socket. In this case you can call C<stop_read>. Neither C<on_read> nor
1279any queued callbacks will be executed then. To start reading again, call
1280C<start_read>.
1281
1282Note that AnyEvent::Handle will automatically C<start_read> for you when
1283you change the C<on_read> callback or push/unshift a read callback, and it
1284will automatically C<stop_read> for you when neither C<on_read> is set nor
1285there are any read requests in the queue.
1286
1287These methods will have no effect when in TLS mode (as TLS doesn't support
1288half-duplex connections).
1289
1290=cut
1291
1292sub stop_read {
1293 my ($self) = @_;
1294
1295 delete $self->{_rw} unless $self->{tls};
1296}
1297
1298sub start_read {
1299 my ($self) = @_;
1300
1301 unless ($self->{_rw} || $self->{_eof}) {
1302 Scalar::Util::weaken $self;
1303
1304 $self->{_rw} = AnyEvent->io (fh => $self->{fh}, poll => "r", cb => sub {
1305 my $rbuf = \($self->{tls} ? my $buf : $self->{rbuf});
1306 my $len = sysread $self->{fh}, $$rbuf, $self->{read_size} || 8192, length $$rbuf;
1307
1308 if ($len > 0) {
1309 $self->{_activity} = AnyEvent->now;
1310
1311 if ($self->{tls}) {
1312 Net::SSLeay::BIO_write ($self->{_rbio}, $$rbuf);
1313
1314 &_dotls ($self);
1315 } else {
1316 $self->_drain_rbuf unless $self->{_in_drain};
1317 }
1318
1319 } elsif (defined $len) {
1320 delete $self->{_rw};
1321 $self->{_eof} = 1;
1322 $self->_drain_rbuf unless $self->{_in_drain};
1323
1324 } elsif ($! != EAGAIN && $! != EINTR && $! != WSAEWOULDBLOCK) {
1325 return $self->_error ($!, 1);
184 } 1326 }
185 }); 1327 });
1328 }
186} 1329}
187 1330
188=item B<on_error ($callback)> 1331# poll the write BIO and send the data if applicable
189 1332sub _dotls {
190Whenever a read or write operation resulted in an error the C<$callback>
191will be called.
192
193The first argument of C<$callback> will be the L<AnyEvent::Handle> object itself.
194The error is given as errno in C<$!>.
195
196=cut
197
198sub on_error {
199 $_[0]->{on_error} = $_[1];
200}
201
202=item B<on_eof ($callback)>
203
204Installs the C<$callback> that will be called when the end of file is
205encountered in a read operation this C<$callback> will be called. The first
206argument will be the L<AnyEvent::Handle> object itself.
207
208=cut
209
210sub on_eof {
211 $_[0]->{on_eof} = $_[1];
212}
213
214=item B<rbuf>
215
216Returns a reference to the read buffer.
217
218NOTE: The read buffer should only be used or modified if the C<on_read>
219method is used directly. The C<read> and C<readlines> methods will provide
220the read data to their callbacks.
221
222=cut
223
224sub rbuf : lvalue {
225 $_[0]->{rbuf}
226}
227
228=item B<read ($len, $callback)>
229
230Will read exactly C<$len> bytes from the filehandle and call the C<$callback>
231if done so. The first argument to the C<$callback> will be the L<AnyEvent::Handle>
232object itself and the second argument the read data.
233
234NOTE: This method will override any callbacks installed via the C<on_read> method.
235
236=cut
237
238sub read {
239 my ($self, $len, $cb) = @_; 1333 my ($self) = @_;
240 1334
241 $self->{read_cb} = $cb; 1335 my $tmp;
242 my $old_blk_size = $self->{read_block_size};
243 $self->{read_block_size} = $len;
244 1336
245 $self->on_read (sub { 1337 if (length $self->{_tls_wbuf}) {
246 #d# warn "OFOFO $len || ".length($_[0]->{rbuf})."||\n"; 1338 while (($tmp = Net::SSLeay::write ($self->{tls}, $self->{_tls_wbuf})) > 0) {
247 1339 substr $self->{_tls_wbuf}, 0, $tmp, "";
248 if ($len == length $_[0]->{rbuf}) {
249 $_[0]->{read_block_size} = $old_blk_size;
250 $_[0]->on_read (undef);
251 $_[0]->{read_cb}->($_[0], (substr $self->{rbuf}, 0, $len, ''));
252 } 1340 }
253 }); 1341 }
254}
255 1342
256=item B<readlines ($callback)> 1343 while (defined ($tmp = Net::SSLeay::read ($self->{tls}))) {
1344 unless (length $tmp) {
1345 # let's treat SSL-eof as we treat normal EOF
1346 delete $self->{_rw};
1347 $self->{_eof} = 1;
1348 &_freetls;
1349 }
257 1350
258=item B<readlines ($sep, $callback)> 1351 $self->{_tls_rbuf} .= $tmp;
1352 $self->_drain_rbuf unless $self->{_in_drain};
1353 $self->{tls} or return; # tls session might have gone away in callback
1354 }
259 1355
260This method will read lines from the filehandle, seperated by C<$sep> or C<"\n"> 1356 $tmp = Net::SSLeay::get_error ($self->{tls}, -1);
261if C<$sep> is not provided. C<$sep> will be used as part of a regex, so it can be
262a regex itself and won't be quoted!
263 1357
264The C<$callback> will be called when at least one 1358 if ($tmp != Net::SSLeay::ERROR_WANT_READ ()) {
265line could be read. The first argument to the C<$callback> will be the L<AnyEvent::Handle> 1359 if ($tmp == Net::SSLeay::ERROR_SYSCALL ()) {
266object itself and the rest of the arguments will be the read lines. 1360 return $self->_error ($!, 1);
1361 } elsif ($tmp == Net::SSLeay::ERROR_SSL ()) {
1362 return $self->_error (&Errno::EIO, 1);
1363 }
267 1364
268NOTE: This method will override any callbacks installed via the C<on_read> method. 1365 # all other errors are fine for our purposes
1366 }
269 1367
270=cut 1368 while (length ($tmp = Net::SSLeay::BIO_read ($self->{_wbio}))) {
1369 $self->{wbuf} .= $tmp;
1370 $self->_drain_wbuf;
1371 }
1372}
271 1373
272sub readlines { 1374=item $handle->starttls ($tls[, $tls_ctx])
1375
1376Instead of starting TLS negotiation immediately when the AnyEvent::Handle
1377object is created, you can also do that at a later time by calling
1378C<starttls>.
1379
1380The first argument is the same as the C<tls> constructor argument (either
1381C<"connect">, C<"accept"> or an existing Net::SSLeay object).
1382
1383The second argument is the optional C<Net::SSLeay::CTX> object that is
1384used when AnyEvent::Handle has to create its own TLS connection object.
1385
1386The TLS connection object will end up in C<< $handle->{tls} >> after this
1387call and can be used or changed to your liking. Note that the handshake
1388might have already started when this function returns.
1389
1390If it an error to start a TLS handshake more than once per
1391AnyEvent::Handle object (this is due to bugs in OpenSSL).
1392
1393=cut
1394
1395sub starttls {
273 my ($self, $NL, $cb) = @_; 1396 my ($self, $ssl, $ctx) = @_;
274 1397
275 if (ref $NL) { 1398 require Net::SSLeay;
276 $cb = $NL; 1399
277 $NL = "\n"; 1400 Carp::croak "it is an error to call starttls more than once on an AnyEvent::Handle object"
1401 if $self->{tls};
278 } 1402
279 1403 if ($ssl eq "accept") {
280 $self->{on_readline} = $cb; 1404 $ssl = Net::SSLeay::new ($ctx || TLS_CTX ());
281 1405 Net::SSLeay::set_accept_state ($ssl);
282 $self->on_read (sub { 1406 } elsif ($ssl eq "connect") {
283 my @lines; 1407 $ssl = Net::SSLeay::new ($ctx || TLS_CTX ());
284 push @lines, $1 while $_[0]->{rbuf} =~ s/(.*)$NL//; 1408 Net::SSLeay::set_connect_state ($ssl);
285 $self->{on_readline}->($_[0], @lines);
286 }); 1409 }
287}
288 1410
289=item B<write ($data)> 1411 $self->{tls} = $ssl;
290 1412
291=item B<write ($callback)> 1413 # basically, this is deep magic (because SSL_read should have the same issues)
1414 # but the openssl maintainers basically said: "trust us, it just works".
1415 # (unfortunately, we have to hardcode constants because the abysmally misdesigned
1416 # and mismaintained ssleay-module doesn't even offer them).
1417 # http://www.mail-archive.com/openssl-dev@openssl.org/msg22420.html
1418 #
1419 # in short: this is a mess.
1420 #
1421 # note that we do not try to keep the length constant between writes as we are required to do.
1422 # we assume that most (but not all) of this insanity only applies to non-blocking cases,
1423 # and we drive openssl fully in blocking mode here. Or maybe we don't - openssl seems to
1424 # have identity issues in that area.
1425 Net::SSLeay::CTX_set_mode ($self->{tls},
1426 (eval { local $SIG{__DIE__}; Net::SSLeay::MODE_ENABLE_PARTIAL_WRITE () } || 1)
1427 | (eval { local $SIG{__DIE__}; Net::SSLeay::MODE_ACCEPT_MOVING_WRITE_BUFFER () } || 2));
292 1428
293=item B<write ($data, $callback)> 1429 $self->{_rbio} = Net::SSLeay::BIO_new (Net::SSLeay::BIO_s_mem ());
1430 $self->{_wbio} = Net::SSLeay::BIO_new (Net::SSLeay::BIO_s_mem ());
294 1431
295This method will write C<$data> to the filehandle and call the C<$callback> 1432 Net::SSLeay::set_bio ($ssl, $self->{_rbio}, $self->{_wbio});
296afterwards. If only C<$callback> is provided it will be called when the
297write buffer becomes empty the next time (or immediately if it already is empty).
298 1433
299=cut 1434 &_dotls; # need to trigger the initial handshake
300 1435 $self->start_read; # make sure we actually do read
301sub write {
302 my ($self, $data, $cb) = @_;
303 if (ref $data) { $cb = $data; undef $data }
304 push @{$self->{write_bufs}}, [$data, $cb];
305 $self->_check_writer;
306} 1436}
307 1437
308sub _check_writer { 1438=item $handle->stoptls
1439
1440Shuts down the SSL connection - this makes a proper EOF handshake by
1441sending a close notify to the other side, but since OpenSSL doesn't
1442support non-blocking shut downs, it is not possible to re-use the stream
1443afterwards.
1444
1445=cut
1446
1447sub stoptls {
309 my ($self) = @_; 1448 my ($self) = @_;
310 1449
311 if ($self->{write_w}) { 1450 if ($self->{tls}) {
312 unless ($self->{write_cb}) { 1451 Net::SSLeay::shutdown ($self->{tls});
313 while (@{$self->{write_bufs}} && not defined $self->{write_bufs}->[0]->[1]) { 1452
314 my $wba = shift @{$self->{write_bufs}}; 1453 &_dotls;
315 $self->{wbuf} .= $wba->[0]; 1454
316 } 1455 # we don't give a shit. no, we do, but we can't. no...
317 } 1456 # we, we... have to use openssl :/
318 return; 1457 &_freetls;
1458 }
1459}
1460
1461sub _freetls {
1462 my ($self) = @_;
1463
1464 return unless $self->{tls};
1465
1466 Net::SSLeay::free (delete $self->{tls});
319 } 1467
1468 delete @$self{qw(_rbio _wbio _tls_wbuf)};
1469}
320 1470
321 my $wba = shift @{$self->{write_bufs}} 1471sub DESTROY {
322 or return; 1472 my ($self) = @_;
323 1473
324 unless (defined $wba->[0]) { 1474 &_freetls;
325 $wba->[1]->($self) if $wba->[1];
326 $self->_check_writer;
327 return;
328 }
329 1475
330 $self->{wbuf} = $wba->[0]; 1476 my $linger = exists $self->{linger} ? $self->{linger} : 3600;
331 $self->{write_cb} = $wba->[1];
332 1477
333 $self->{write_w} = 1478 if ($linger && length $self->{wbuf}) {
334 AnyEvent->io (poll => 'w', fh => $self->{fh}, cb => sub { 1479 my $fh = delete $self->{fh};
1480 my $wbuf = delete $self->{wbuf};
1481
1482 my @linger;
1483
1484 push @linger, AnyEvent->io (fh => $fh, poll => "w", cb => sub {
335 my $l = syswrite $self->{fh}, $self->{wbuf}, length $self->{wbuf}; 1485 my $len = syswrite $fh, $wbuf, length $wbuf;
336 1486
337 if (not defined $l) { 1487 if ($len > 0) {
338 return if $! == EAGAIN || $! == EINTR; 1488 substr $wbuf, 0, $len, "";
339 delete $self->{write_w};
340 $self->{on_error}->($self) if $self->{on_error};
341
342 } else { 1489 } else {
343 substr $self->{wbuf}, 0, $l, ''; 1490 @linger = (); # end
344
345 if (length ($self->{wbuf}) == 0) {
346 $self->{write_cb}->($self) if $self->{write_cb};
347
348 delete $self->{write_w};
349 delete $self->{wbuf};
350 delete $self->{write_cb};
351
352 $self->_check_writer;
353 }
354 } 1491 }
355 }); 1492 });
1493 push @linger, AnyEvent->timer (after => $linger, cb => sub {
1494 @linger = ();
1495 });
1496 }
1497}
1498
1499=item $handle->destroy
1500
1501Shuts down the handle object as much as possible - this call ensures that
1502no further callbacks will be invoked and resources will be freed as much
1503as possible. You must not call any methods on the object afterwards.
1504
1505Normally, you can just "forget" any references to an AnyEvent::Handle
1506object and it will simply shut down. This works in fatal error and EOF
1507callbacks, as well as code outside. It does I<NOT> work in a read or write
1508callback, so when you want to destroy the AnyEvent::Handle object from
1509within such an callback. You I<MUST> call C<< ->destroy >> explicitly in
1510that case.
1511
1512The handle might still linger in the background and write out remaining
1513data, as specified by the C<linger> option, however.
1514
1515=cut
1516
1517sub destroy {
1518 my ($self) = @_;
1519
1520 $self->DESTROY;
1521 %$self = ();
1522}
1523
1524=item AnyEvent::Handle::TLS_CTX
1525
1526This function creates and returns the Net::SSLeay::CTX object used by
1527default for TLS mode.
1528
1529The context is created like this:
1530
1531 Net::SSLeay::load_error_strings;
1532 Net::SSLeay::SSLeay_add_ssl_algorithms;
1533 Net::SSLeay::randomize;
1534
1535 my $CTX = Net::SSLeay::CTX_new;
1536
1537 Net::SSLeay::CTX_set_options $CTX, Net::SSLeay::OP_ALL
1538
1539=cut
1540
1541our $TLS_CTX;
1542
1543sub TLS_CTX() {
1544 $TLS_CTX || do {
1545 require Net::SSLeay;
1546
1547 Net::SSLeay::load_error_strings ();
1548 Net::SSLeay::SSLeay_add_ssl_algorithms ();
1549 Net::SSLeay::randomize ();
1550
1551 $TLS_CTX = Net::SSLeay::CTX_new ();
1552
1553 Net::SSLeay::CTX_set_options ($TLS_CTX, Net::SSLeay::OP_ALL ());
1554
1555 $TLS_CTX
1556 }
356} 1557}
357 1558
358=back 1559=back
359 1560
1561
1562=head1 NONFREQUENTLY ASKED QUESTIONS
1563
1564=over 4
1565
1566=item I C<undef> the AnyEvent::Handle reference inside my callback and
1567still get further invocations!
1568
1569That's because AnyEvent::Handle keeps a reference to itself when handling
1570read or write callbacks.
1571
1572It is only safe to "forget" the reference inside EOF or error callbacks,
1573from within all other callbacks, you need to explicitly call the C<<
1574->destroy >> method.
1575
1576=item I get different callback invocations in TLS mode/Why can't I pause
1577reading?
1578
1579Unlike, say, TCP, TLS connections do not consist of two independent
1580communication channels, one for each direction. Or put differently. The
1581read and write directions are not independent of each other: you cannot
1582write data unless you are also prepared to read, and vice versa.
1583
1584This can mean than, in TLS mode, you might get C<on_error> or C<on_eof>
1585callback invocations when you are not expecting any read data - the reason
1586is that AnyEvent::Handle always reads in TLS mode.
1587
1588During the connection, you have to make sure that you always have a
1589non-empty read-queue, or an C<on_read> watcher. At the end of the
1590connection (or when you no longer want to use it) you can call the
1591C<destroy> method.
1592
1593=item How do I read data until the other side closes the connection?
1594
1595If you just want to read your data into a perl scalar, the easiest way
1596to achieve this is by setting an C<on_read> callback that does nothing,
1597clearing the C<on_eof> callback and in the C<on_error> callback, the data
1598will be in C<$_[0]{rbuf}>:
1599
1600 $handle->on_read (sub { });
1601 $handle->on_eof (undef);
1602 $handle->on_error (sub {
1603 my $data = delete $_[0]{rbuf};
1604 undef $handle;
1605 });
1606
1607The reason to use C<on_error> is that TCP connections, due to latencies
1608and packets loss, might get closed quite violently with an error, when in
1609fact, all data has been received.
1610
1611It is usually better to use acknowledgements when transferring data,
1612to make sure the other side hasn't just died and you got the data
1613intact. This is also one reason why so many internet protocols have an
1614explicit QUIT command.
1615
1616=item I don't want to destroy the handle too early - how do I wait until
1617all data has been written?
1618
1619After writing your last bits of data, set the C<on_drain> callback
1620and destroy the handle in there - with the default setting of
1621C<low_water_mark> this will be called precisely when all data has been
1622written to the socket:
1623
1624 $handle->push_write (...);
1625 $handle->on_drain (sub {
1626 warn "all data submitted to the kernel\n";
1627 undef $handle;
1628 });
1629
1630=back
1631
1632
1633=head1 SUBCLASSING AnyEvent::Handle
1634
1635In many cases, you might want to subclass AnyEvent::Handle.
1636
1637To make this easier, a given version of AnyEvent::Handle uses these
1638conventions:
1639
1640=over 4
1641
1642=item * all constructor arguments become object members.
1643
1644At least initially, when you pass a C<tls>-argument to the constructor it
1645will end up in C<< $handle->{tls} >>. Those members might be changed or
1646mutated later on (for example C<tls> will hold the TLS connection object).
1647
1648=item * other object member names are prefixed with an C<_>.
1649
1650All object members not explicitly documented (internal use) are prefixed
1651with an underscore character, so the remaining non-C<_>-namespace is free
1652for use for subclasses.
1653
1654=item * all members not documented here and not prefixed with an underscore
1655are free to use in subclasses.
1656
1657Of course, new versions of AnyEvent::Handle may introduce more "public"
1658member variables, but thats just life, at least it is documented.
1659
1660=back
1661
360=head1 AUTHOR 1662=head1 AUTHOR
361 1663
362Robin Redeker, C<< <elmex at ta-sa.org> >> 1664Robin Redeker C<< <elmex at ta-sa.org> >>, Marc Lehmann <schmorp@schmorp.de>.
363
364=head1 BUGS
365
366Please report any bugs or feature requests to
367C<bug-io-anyevent at rt.cpan.org>, or through the web interface at
368L<http://rt.cpan.org/NoAuth/ReportBug.html?Queue=IO-AnyEvent>.
369I will be notified, and then you'll automatically be notified of progress on
370your bug as I make changes.
371
372=head1 SUPPORT
373
374You can find documentation for this module with the perldoc command.
375
376 perldoc AnyEvent::Handle
377
378You can also look for information at:
379
380=over 4
381
382=item * AnnoCPAN: Annotated CPAN documentation
383
384L<http://annocpan.org/dist/IO-AnyEvent>
385
386=item * CPAN Ratings
387
388L<http://cpanratings.perl.org/d/IO-AnyEvent>
389
390=item * RT: CPAN's request tracker
391
392L<http://rt.cpan.org/NoAuth/Bugs.html?Dist=IO-AnyEvent>
393
394=item * Search CPAN
395
396L<http://search.cpan.org/dist/IO-AnyEvent>
397
398=back
399
400=head1 ACKNOWLEDGEMENTS
401
402=head1 COPYRIGHT & LICENSE
403
404Copyright 2008 Robin Redeker, all rights reserved.
405
406This program is free software; you can redistribute it and/or modify it
407under the same terms as Perl itself.
408 1665
409=cut 1666=cut
410 1667
4111; # End of AnyEvent::Handle 16681; # End of AnyEvent::Handle

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines