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 1.3 by elmex, Sun Apr 27 19:00:33 2008 UTC vs.
Revision 1.149 by root, Thu Jul 16 03:48:33 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 = 4.82;
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 $hdl; $hdl = new AnyEvent::Handle
29 fh => \*STDIN,
30 on_error => sub {
31 warn "got error $_[2]\n";
32 $cv->send;
33 );
30 34
31 $ae_fh->on_eof (sub { $cv->broadcast }); 35 # send some request line
36 $hdl->push_write ("getinfo\015\012");
32 37
33 $ae_fh->readlines (sub { 38 # read the response line
39 $hdl->push_read (line => sub {
34 my ($ae_fh, @lines) = @_; 40 my ($hdl, $line) = @_;
35 for (@lines) { 41 warn "got line <$line>\n";
36 chomp; 42 $cv->send;
37 print "Line: $_"; 43 });
44
45 $cv->recv;
46
47=head1 DESCRIPTION
48
49This module is a helper module to make it easier to do event-based I/O on
50filehandles. For utility functions for doing non-blocking connects and accepts
51on sockets see L<AnyEvent::Util>.
52
53The L<AnyEvent::Intro> tutorial contains some well-documented
54AnyEvent::Handle examples.
55
56In the following, when the documentation refers to of "bytes" then this
57means characters. As sysread and syswrite are used for all I/O, their
58treatment of characters applies to this module as well.
59
60All callbacks will be invoked with the handle object as their first
61argument.
62
63=head1 METHODS
64
65=over 4
66
67=item $handle = B<new> AnyEvent::TLS fh => $filehandle, key => value...
68
69The constructor supports these arguments (all as C<< key => value >> pairs).
70
71=over 4
72
73=item fh => $filehandle [MANDATORY]
74
75The filehandle this L<AnyEvent::Handle> object will operate on.
76
77NOTE: The filehandle will be set to non-blocking mode (using
78C<AnyEvent::Util::fh_nonblocking>) by the constructor and needs to stay in
79that mode.
80
81=item on_eof => $cb->($handle)
82
83Set the callback to be called when an end-of-file condition is detected,
84i.e. in the case of a socket, when the other side has closed the
85connection cleanly.
86
87For sockets, this just means that the other side has stopped sending data,
88you can still try to write data, and, in fact, one can return from the EOF
89callback and continue writing data, as only the read part has been shut
90down.
91
92While not mandatory, it is I<highly> recommended to set an EOF callback,
93otherwise you might end up with a closed socket while you are still
94waiting for data.
95
96If an EOF condition has been detected but no C<on_eof> callback has been
97set, then a fatal error will be raised with C<$!> set to <0>.
98
99=item on_error => $cb->($handle, $fatal, $message)
100
101This is the error callback, which is called when, well, some error
102occured, such as not being able to resolve the hostname, failure to
103connect or a read error.
104
105Some errors are fatal (which is indicated by C<$fatal> being true). On
106fatal errors the handle object will be destroyed (by a call to C<< ->
107destroy >>) after invoking the error callback (which means you are free to
108examine the handle object). Examples of fatal errors are an EOF condition
109with active (but unsatisifable) read watchers (C<EPIPE>) or I/O errors.
110
111AnyEvent::Handle tries to find an appropriate error code for you to check
112against, but in some cases (TLS errors), this does not work well. It is
113recommended to always output the C<$message> argument in human-readable
114error messages (it's usually the same as C<"$!">).
115
116Non-fatal errors can be retried by simply returning, but it is recommended
117to simply ignore this parameter and instead abondon the handle object
118when this callback is invoked. Examples of non-fatal errors are timeouts
119C<ETIMEDOUT>) or badly-formatted data (C<EBADMSG>).
120
121On callback entrance, the value of C<$!> contains the operating system
122error code (or C<ENOSPC>, C<EPIPE>, C<ETIMEDOUT>, C<EBADMSG> or
123C<EPROTO>).
124
125While not mandatory, it is I<highly> recommended to set this callback, as
126you will not be notified of errors otherwise. The default simply calls
127C<croak>.
128
129=item on_read => $cb->($handle)
130
131This sets the default read callback, which is called when data arrives
132and no read request is in the queue (unlike read queue callbacks, this
133callback will only be called when at least one octet of data is in the
134read buffer).
135
136To access (and remove data from) the read buffer, use the C<< ->rbuf >>
137method or access the C<< $handle->{rbuf} >> member directly. Note that you
138must not enlarge or modify the read buffer, you can only remove data at
139the beginning from it.
140
141When an EOF condition is detected then AnyEvent::Handle will first try to
142feed all the remaining data to the queued callbacks and C<on_read> before
143calling the C<on_eof> callback. If no progress can be made, then a fatal
144error will be raised (with C<$!> set to C<EPIPE>).
145
146=item on_drain => $cb->($handle)
147
148This sets the callback that is called when the write buffer becomes empty
149(or when the callback is set and the buffer is empty already).
150
151To append to the write buffer, use the C<< ->push_write >> method.
152
153This callback is useful when you don't want to put all of your write data
154into the queue at once, for example, when you want to write the contents
155of some file to the socket you might not want to read the whole file into
156memory and push it into the queue, but instead only read more data from
157the file when the write queue becomes empty.
158
159=item timeout => $fractional_seconds
160
161If non-zero, then this enables an "inactivity" timeout: whenever this many
162seconds pass without a successful read or write on the underlying file
163handle, the C<on_timeout> callback will be invoked (and if that one is
164missing, a non-fatal C<ETIMEDOUT> error will be raised).
165
166Note that timeout processing is also active when you currently do not have
167any outstanding read or write requests: If you plan to keep the connection
168idle then you should disable the timout temporarily or ignore the timeout
169in the C<on_timeout> callback, in which case AnyEvent::Handle will simply
170restart the timeout.
171
172Zero (the default) disables this timeout.
173
174=item on_timeout => $cb->($handle)
175
176Called whenever the inactivity timeout passes. If you return from this
177callback, then the timeout will be reset as if some activity had happened,
178so this condition is not fatal in any way.
179
180=item rbuf_max => <bytes>
181
182If defined, then a fatal error will be raised (with C<$!> set to C<ENOSPC>)
183when the read buffer ever (strictly) exceeds this size. This is useful to
184avoid some forms of denial-of-service attacks.
185
186For example, a server accepting connections from untrusted sources should
187be configured to accept only so-and-so much data that it cannot act on
188(for example, when expecting a line, an attacker could send an unlimited
189amount of data without a callback ever being called as long as the line
190isn't finished).
191
192=item autocork => <boolean>
193
194When disabled (the default), then C<push_write> will try to immediately
195write the data to the handle, if possible. This avoids having to register
196a write watcher and wait for the next event loop iteration, but can
197be inefficient if you write multiple small chunks (on the wire, this
198disadvantage is usually avoided by your kernel's nagle algorithm, see
199C<no_delay>, but this option can save costly syscalls).
200
201When enabled, then writes will always be queued till the next event loop
202iteration. This is efficient when you do many small writes per iteration,
203but less efficient when you do a single write only per iteration (or when
204the write buffer often is full). It also increases write latency.
205
206=item no_delay => <boolean>
207
208When doing small writes on sockets, your operating system kernel might
209wait a bit for more data before actually sending it out. This is called
210the Nagle algorithm, and usually it is beneficial.
211
212In some situations you want as low a delay as possible, which can be
213accomplishd by setting this option to a true value.
214
215The default is your opertaing system's default behaviour (most likely
216enabled), this option explicitly enables or disables it, if possible.
217
218=item read_size => <bytes>
219
220The default read block size (the amount of bytes this module will
221try to read during each loop iteration, which affects memory
222requirements). Default: C<8192>.
223
224=item low_water_mark => <bytes>
225
226Sets the amount of bytes (default: C<0>) that make up an "empty" write
227buffer: If the write reaches this size or gets even samller it is
228considered empty.
229
230Sometimes it can be beneficial (for performance reasons) to add data to
231the write buffer before it is fully drained, but this is a rare case, as
232the operating system kernel usually buffers data as well, so the default
233is good in almost all cases.
234
235=item linger => <seconds>
236
237If non-zero (default: C<3600>), then the destructor of the
238AnyEvent::Handle object will check whether there is still outstanding
239write data and will install a watcher that will write this data to the
240socket. No errors will be reported (this mostly matches how the operating
241system treats outstanding data at socket close time).
242
243This will not work for partial TLS data that could not be encoded
244yet. This data will be lost. Calling the C<stoptls> method in time might
245help.
246
247=item peername => $string
248
249A string used to identify the remote site - usually the DNS hostname
250(I<not> IDN!) used to create the connection, rarely the IP address.
251
252Apart from being useful in error messages, this string is also used in TLS
253peername verification (see C<verify_peername> in L<AnyEvent::TLS>). This
254verification will be skipped when C<peername> is not specified or
255C<undef>.
256
257=item tls => "accept" | "connect" | Net::SSLeay::SSL object
258
259When this parameter is given, it enables TLS (SSL) mode, that means
260AnyEvent will start a TLS handshake as soon as the conenction has been
261established and will transparently encrypt/decrypt data afterwards.
262
263All TLS protocol errors will be signalled as C<EPROTO>, with an
264appropriate error message.
265
266TLS mode requires Net::SSLeay to be installed (it will be loaded
267automatically when you try to create a TLS handle): this module doesn't
268have a dependency on that module, so if your module requires it, you have
269to add the dependency yourself.
270
271Unlike TCP, TLS has a server and client side: for the TLS server side, use
272C<accept>, and for the TLS client side of a connection, use C<connect>
273mode.
274
275You can also provide your own TLS connection object, but you have
276to make sure that you call either C<Net::SSLeay::set_connect_state>
277or C<Net::SSLeay::set_accept_state> on it before you pass it to
278AnyEvent::Handle. Also, this module will take ownership of this connection
279object.
280
281At some future point, AnyEvent::Handle might switch to another TLS
282implementation, then the option to use your own session object will go
283away.
284
285B<IMPORTANT:> since Net::SSLeay "objects" are really only integers,
286passing in the wrong integer will lead to certain crash. This most often
287happens when one uses a stylish C<< tls => 1 >> and is surprised about the
288segmentation fault.
289
290See the C<< ->starttls >> method for when need to start TLS negotiation later.
291
292=item tls_ctx => $anyevent_tls
293
294Use the given C<AnyEvent::TLS> object to create the new TLS connection
295(unless a connection object was specified directly). If this parameter is
296missing, then AnyEvent::Handle will use C<AnyEvent::Handle::TLS_CTX>.
297
298Instead of an object, you can also specify a hash reference with C<< key
299=> value >> pairs. Those will be passed to L<AnyEvent::TLS> to create a
300new TLS context object.
301
302=item on_starttls => $cb->($handle, $success[, $error_message])
303
304This callback will be invoked when the TLS/SSL handshake has finished. If
305C<$success> is true, then the TLS handshake succeeded, otherwise it failed
306(C<on_stoptls> will not be called in this case).
307
308The session in C<< $handle->{tls} >> can still be examined in this
309callback, even when the handshake was not successful.
310
311TLS handshake failures will not cause C<on_error> to be invoked when this
312callback is in effect, instead, the error message will be passed to C<on_starttls>.
313
314Without this callback, handshake failures lead to C<on_error> being
315called, as normal.
316
317Note that you cannot call C<starttls> right again in this callback. If you
318need to do that, start an zero-second timer instead whose callback can
319then call C<< ->starttls >> again.
320
321=item on_stoptls => $cb->($handle)
322
323When a SSLv3/TLS shutdown/close notify/EOF is detected and this callback is
324set, then it will be invoked after freeing the TLS session. If it is not,
325then a TLS shutdown condition will be treated like a normal EOF condition
326on the handle.
327
328The session in C<< $handle->{tls} >> can still be examined in this
329callback.
330
331This callback will only be called on TLS shutdowns, not when the
332underlying handle signals EOF.
333
334=item json => JSON or JSON::XS object
335
336This is the json coder object used by the C<json> read and write types.
337
338If you don't supply it, then AnyEvent::Handle will create and use a
339suitable one (on demand), which will write and expect UTF-8 encoded JSON
340texts.
341
342Note that you are responsible to depend on the JSON module if you want to
343use this functionality, as AnyEvent does not have a dependency itself.
344
345=back
346
347=cut
348
349sub new {
350 my $class = shift;
351 my $self = bless { @_ }, $class;
352
353 $self->{fh} or Carp::croak "mandatory argument fh is missing";
354
355 AnyEvent::Util::fh_nonblocking $self->{fh}, 1;
356
357 $self->{_activity} = AnyEvent->now;
358 $self->_timeout;
359
360 $self->no_delay (delete $self->{no_delay}) if exists $self->{no_delay};
361
362 $self->starttls (delete $self->{tls}, delete $self->{tls_ctx})
363 if $self->{tls};
364
365 $self->on_drain (delete $self->{on_drain}) if $self->{on_drain};
366
367 $self->start_read
368 if $self->{on_read};
369
370 $self->{fh} && $self
371}
372
373#sub _shutdown {
374# my ($self) = @_;
375#
376# delete @$self{qw(_tw _rw _ww fh wbuf on_read _queue)};
377# $self->{_eof} = 1; # tell starttls et. al to stop trying
378#
379# &_freetls;
380#}
381
382sub _error {
383 my ($self, $errno, $fatal, $message) = @_;
384
385 $! = $errno;
386 $message ||= "$!";
387
388 if ($self->{on_error}) {
389 $self->{on_error}($self, $fatal, $message);
390 $self->destroy;
391 } elsif ($self->{fh}) {
392 $self->destroy;
393 Carp::croak "AnyEvent::Handle uncaught error: $message";
394 }
395}
396
397=item $fh = $handle->fh
398
399This method returns the file handle used to create the L<AnyEvent::Handle> object.
400
401=cut
402
403sub fh { $_[0]{fh} }
404
405=item $handle->on_error ($cb)
406
407Replace the current C<on_error> callback (see the C<on_error> constructor argument).
408
409=cut
410
411sub on_error {
412 $_[0]{on_error} = $_[1];
413}
414
415=item $handle->on_eof ($cb)
416
417Replace the current C<on_eof> callback (see the C<on_eof> constructor argument).
418
419=cut
420
421sub on_eof {
422 $_[0]{on_eof} = $_[1];
423}
424
425=item $handle->on_timeout ($cb)
426
427Replace the current C<on_timeout> callback, or disables the callback (but
428not the timeout) if C<$cb> = C<undef>. See the C<timeout> constructor
429argument and method.
430
431=cut
432
433sub on_timeout {
434 $_[0]{on_timeout} = $_[1];
435}
436
437=item $handle->autocork ($boolean)
438
439Enables or disables the current autocork behaviour (see C<autocork>
440constructor argument). Changes will only take effect on the next write.
441
442=cut
443
444sub autocork {
445 $_[0]{autocork} = $_[1];
446}
447
448=item $handle->no_delay ($boolean)
449
450Enables or disables the C<no_delay> setting (see constructor argument of
451the same name for details).
452
453=cut
454
455sub no_delay {
456 $_[0]{no_delay} = $_[1];
457
458 eval {
459 local $SIG{__DIE__};
460 setsockopt $_[0]{fh}, &Socket::IPPROTO_TCP, &Socket::TCP_NODELAY, int $_[1];
461 };
462}
463
464=item $handle->on_starttls ($cb)
465
466Replace the current C<on_starttls> callback (see the C<on_starttls> constructor argument).
467
468=cut
469
470sub on_starttls {
471 $_[0]{on_starttls} = $_[1];
472}
473
474=item $handle->on_stoptls ($cb)
475
476Replace the current C<on_stoptls> callback (see the C<on_stoptls> constructor argument).
477
478=cut
479
480sub on_starttls {
481 $_[0]{on_stoptls} = $_[1];
482}
483
484#############################################################################
485
486=item $handle->timeout ($seconds)
487
488Configures (or disables) the inactivity timeout.
489
490=cut
491
492sub timeout {
493 my ($self, $timeout) = @_;
494
495 $self->{timeout} = $timeout;
496 $self->_timeout;
497}
498
499# reset the timeout watcher, as neccessary
500# also check for time-outs
501sub _timeout {
502 my ($self) = @_;
503
504 if ($self->{timeout}) {
505 my $NOW = AnyEvent->now;
506
507 # when would the timeout trigger?
508 my $after = $self->{_activity} + $self->{timeout} - $NOW;
509
510 # now or in the past already?
511 if ($after <= 0) {
512 $self->{_activity} = $NOW;
513
514 if ($self->{on_timeout}) {
515 $self->{on_timeout}($self);
516 } else {
517 $self->_error (&Errno::ETIMEDOUT);
518 }
519
520 # callback could have changed timeout value, optimise
521 return unless $self->{timeout};
522
523 # calculate new after
524 $after = $self->{timeout};
525 }
526
527 Scalar::Util::weaken $self;
528 return unless $self; # ->error could have destroyed $self
529
530 $self->{_tw} ||= AnyEvent->timer (after => $after, cb => sub {
531 delete $self->{_tw};
532 $self->_timeout;
533 });
534 } else {
535 delete $self->{_tw};
536 }
537}
538
539#############################################################################
540
541=back
542
543=head2 WRITE QUEUE
544
545AnyEvent::Handle manages two queues per handle, one for writing and one
546for reading.
547
548The write queue is very simple: you can add data to its end, and
549AnyEvent::Handle will automatically try to get rid of it for you.
550
551When data could be written and the write buffer is shorter then the low
552water mark, the C<on_drain> callback will be invoked.
553
554=over 4
555
556=item $handle->on_drain ($cb)
557
558Sets the C<on_drain> callback or clears it (see the description of
559C<on_drain> in the constructor).
560
561=cut
562
563sub on_drain {
564 my ($self, $cb) = @_;
565
566 $self->{on_drain} = $cb;
567
568 $cb->($self)
569 if $cb && $self->{low_water_mark} >= (length $self->{wbuf}) + (length $self->{_tls_wbuf});
570}
571
572=item $handle->push_write ($data)
573
574Queues the given scalar to be written. You can push as much data as you
575want (only limited by the available memory), as C<AnyEvent::Handle>
576buffers it independently of the kernel.
577
578=cut
579
580sub _drain_wbuf {
581 my ($self) = @_;
582
583 if (!$self->{_ww} && length $self->{wbuf}) {
584
585 Scalar::Util::weaken $self;
586
587 my $cb = sub {
588 my $len = syswrite $self->{fh}, $self->{wbuf};
589
590 if (defined $len) {
591 substr $self->{wbuf}, 0, $len, "";
592
593 $self->{_activity} = AnyEvent->now;
594
595 $self->{on_drain}($self)
596 if $self->{low_water_mark} >= (length $self->{wbuf}) + (length $self->{_tls_wbuf})
597 && $self->{on_drain};
598
599 delete $self->{_ww} unless length $self->{wbuf};
600 } elsif ($! != EAGAIN && $! != EINTR && $! != WSAEWOULDBLOCK) {
601 $self->_error ($!, 1);
602 }
603 };
604
605 # try to write data immediately
606 $cb->() unless $self->{autocork};
607
608 # if still data left in wbuf, we need to poll
609 $self->{_ww} = AnyEvent->io (fh => $self->{fh}, poll => "w", cb => $cb)
610 if length $self->{wbuf};
611 };
612}
613
614our %WH;
615
616sub register_write_type($$) {
617 $WH{$_[0]} = $_[1];
618}
619
620sub push_write {
621 my $self = shift;
622
623 if (@_ > 1) {
624 my $type = shift;
625
626 @_ = ($WH{$type} or Carp::croak "unsupported type passed to AnyEvent::Handle::push_write")
627 ->($self, @_);
628 }
629
630 if ($self->{tls}) {
631 $self->{_tls_wbuf} .= $_[0];
632
633 &_dotls ($self);
634 } else {
635 $self->{wbuf} .= $_[0];
636 $self->_drain_wbuf;
637 }
638}
639
640=item $handle->push_write (type => @args)
641
642Instead of formatting your data yourself, you can also let this module do
643the job by specifying a type and type-specific arguments.
644
645Predefined types are (if you have ideas for additional types, feel free to
646drop by and tell us):
647
648=over 4
649
650=item netstring => $string
651
652Formats the given value as netstring
653(http://cr.yp.to/proto/netstrings.txt, this is not a recommendation to use them).
654
655=cut
656
657register_write_type netstring => sub {
658 my ($self, $string) = @_;
659
660 (length $string) . ":$string,"
661};
662
663=item packstring => $format, $data
664
665An octet string prefixed with an encoded length. The encoding C<$format>
666uses the same format as a Perl C<pack> format, but must specify a single
667integer only (only one of C<cCsSlLqQiInNvVjJw> is allowed, plus an
668optional C<!>, C<< < >> or C<< > >> modifier).
669
670=cut
671
672register_write_type packstring => sub {
673 my ($self, $format, $string) = @_;
674
675 pack "$format/a*", $string
676};
677
678=item json => $array_or_hashref
679
680Encodes the given hash or array reference into a JSON object. Unless you
681provide your own JSON object, this means it will be encoded to JSON text
682in UTF-8.
683
684JSON objects (and arrays) are self-delimiting, so you can write JSON at
685one end of a handle and read them at the other end without using any
686additional framing.
687
688The generated JSON text is guaranteed not to contain any newlines: While
689this module doesn't need delimiters after or between JSON texts to be
690able to read them, many other languages depend on that.
691
692A simple RPC protocol that interoperates easily with others is to send
693JSON arrays (or objects, although arrays are usually the better choice as
694they mimic how function argument passing works) and a newline after each
695JSON text:
696
697 $handle->push_write (json => ["method", "arg1", "arg2"]); # whatever
698 $handle->push_write ("\012");
699
700An AnyEvent::Handle receiver would simply use the C<json> read type and
701rely on the fact that the newline will be skipped as leading whitespace:
702
703 $handle->push_read (json => sub { my $array = $_[1]; ... });
704
705Other languages could read single lines terminated by a newline and pass
706this line into their JSON decoder of choice.
707
708=cut
709
710register_write_type json => sub {
711 my ($self, $ref) = @_;
712
713 require JSON;
714
715 $self->{json} ? $self->{json}->encode ($ref)
716 : JSON::encode_json ($ref)
717};
718
719=item storable => $reference
720
721Freezes the given reference using L<Storable> and writes it to the
722handle. Uses the C<nfreeze> format.
723
724=cut
725
726register_write_type storable => sub {
727 my ($self, $ref) = @_;
728
729 require Storable;
730
731 pack "w/a*", Storable::nfreeze ($ref)
732};
733
734=back
735
736=item $handle->push_shutdown
737
738Sometimes you know you want to close the socket after writing your data
739before it was actually written. One way to do that is to replace your
740C<on_drain> handler by a callback that shuts down the socket (and set
741C<low_water_mark> to C<0>). This method is a shorthand for just that, and
742replaces the C<on_drain> callback with:
743
744 sub { shutdown $_[0]{fh}, 1 } # for push_shutdown
745
746This simply shuts down the write side and signals an EOF condition to the
747the peer.
748
749You can rely on the normal read queue and C<on_eof> handling
750afterwards. This is the cleanest way to close a connection.
751
752=cut
753
754sub push_shutdown {
755 my ($self) = @_;
756
757 delete $self->{low_water_mark};
758 $self->on_drain (sub { shutdown $_[0]{fh}, 1 });
759}
760
761=item AnyEvent::Handle::register_write_type type => $coderef->($handle, @args)
762
763This function (not method) lets you add your own types to C<push_write>.
764Whenever the given C<type> is used, C<push_write> will invoke the code
765reference with the handle object and the remaining arguments.
766
767The code reference is supposed to return a single octet string that will
768be appended to the write buffer.
769
770Note that this is a function, and all types registered this way will be
771global, so try to use unique names.
772
773=cut
774
775#############################################################################
776
777=back
778
779=head2 READ QUEUE
780
781AnyEvent::Handle manages two queues per handle, one for writing and one
782for reading.
783
784The read queue is more complex than the write queue. It can be used in two
785ways, the "simple" way, using only C<on_read> and the "complex" way, using
786a queue.
787
788In the simple case, you just install an C<on_read> callback and whenever
789new data arrives, it will be called. You can then remove some data (if
790enough is there) from the read buffer (C<< $handle->rbuf >>). Or you cna
791leave the data there if you want to accumulate more (e.g. when only a
792partial message has been received so far).
793
794In the more complex case, you want to queue multiple callbacks. In this
795case, AnyEvent::Handle will call the first queued callback each time new
796data arrives (also the first time it is queued) and removes it when it has
797done its job (see C<push_read>, below).
798
799This way you can, for example, push three line-reads, followed by reading
800a chunk of data, and AnyEvent::Handle will execute them in order.
801
802Example 1: EPP protocol parser. EPP sends 4 byte length info, followed by
803the specified number of bytes which give an XML datagram.
804
805 # in the default state, expect some header bytes
806 $handle->on_read (sub {
807 # some data is here, now queue the length-header-read (4 octets)
808 shift->unshift_read (chunk => 4, sub {
809 # header arrived, decode
810 my $len = unpack "N", $_[1];
811
812 # now read the payload
813 shift->unshift_read (chunk => $len, sub {
814 my $xml = $_[1];
815 # handle xml
816 });
817 });
818 });
819
820Example 2: Implement a client for a protocol that replies either with "OK"
821and another line or "ERROR" for the first request that is sent, and 64
822bytes for the second request. Due to the availability of a queue, we can
823just pipeline sending both requests and manipulate the queue as necessary
824in the callbacks.
825
826When the first callback is called and sees an "OK" response, it will
827C<unshift> another line-read. This line-read will be queued I<before> the
82864-byte chunk callback.
829
830 # request one, returns either "OK + extra line" or "ERROR"
831 $handle->push_write ("request 1\015\012");
832
833 # we expect "ERROR" or "OK" as response, so push a line read
834 $handle->push_read (line => sub {
835 # if we got an "OK", we have to _prepend_ another line,
836 # so it will be read before the second request reads its 64 bytes
837 # which are already in the queue when this callback is called
838 # we don't do this in case we got an error
839 if ($_[1] eq "OK") {
840 $_[0]->unshift_read (line => sub {
841 my $response = $_[1];
842 ...
843 });
38 } 844 }
39 }); 845 });
40 846
41 # or use the constructor to pass the callback: 847 # request two, simply returns 64 octets
848 $handle->push_write ("request 2\015\012");
42 849
43 my $ae_fh2 = 850 # simply read 64 bytes, always
44 AnyEvent::Handle->new ( 851 $handle->push_read (chunk => 64, sub {
45 fh => \*STDIN, 852 my $response = $_[1];
46 on_eof => sub { 853 ...
47 $cv->broadcast; 854 });
48 }, 855
49 on_readline => sub { 856=over 4
50 my ($ae_fh, @lines) = @_; 857
51 for (@lines) { 858=cut
52 chomp; 859
53 print "Line: $_"; 860sub _drain_rbuf {
861 my ($self) = @_;
862
863 local $self->{_in_drain} = 1;
864
865 if (
866 defined $self->{rbuf_max}
867 && $self->{rbuf_max} < length $self->{rbuf}
868 ) {
869 $self->_error (&Errno::ENOSPC, 1), return;
870 }
871
872 while () {
873 # we need to use a separate tls read buffer, as we must not receive data while
874 # we are draining the buffer, and this can only happen with TLS.
875 $self->{rbuf} .= delete $self->{_tls_rbuf} if exists $self->{_tls_rbuf};
876
877 my $len = length $self->{rbuf};
878
879 if (my $cb = shift @{ $self->{_queue} }) {
880 unless ($cb->($self)) {
881 if ($self->{_eof}) {
882 # no progress can be made (not enough data and no data forthcoming)
883 $self->_error (&Errno::EPIPE, 1), return;
54 } 884 }
885
886 unshift @{ $self->{_queue} }, $cb;
887 last;
55 } 888 }
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}) { 889 } elsif ($self->{on_read}) {
121 $self->readlines ($self->{on_readline}); 890 last unless $len;
122 891
892 $self->{on_read}($self);
893
894 if (
895 $len == length $self->{rbuf} # if no data has been consumed
896 && !@{ $self->{_queue} } # and the queue is still empty
897 && $self->{on_read} # but we still have on_read
898 ) {
899 # no further data will arrive
900 # so no progress can be made
901 $self->_error (&Errno::EPIPE, 1), return
902 if $self->{_eof};
903
904 last; # more data might arrive
905 }
906 } else {
907 # read side becomes idle
908 delete $self->{_rw} unless $self->{tls};
909 last;
910 }
911 }
912
123 } elsif ($self->{on_eof}) { 913 if ($self->{_eof}) {
124 $self->on_eof ($self->{on_eof});
125
126 } elsif ($self->{on_error}) { 914 if ($self->{on_eof}) {
127 $self->on_eof ($self->{on_error}); 915 $self->{on_eof}($self)
916 } else {
917 $self->_error (0, 1, "Unexpected end-of-file");
918 }
128 } 919 }
129 920
130 return $self 921 # may need to restart read watcher
922 unless ($self->{_rw}) {
923 $self->start_read
924 if $self->{on_read} || @{ $self->{_queue} };
925 }
131} 926}
132 927
133=item B<fh> 928=item $handle->on_read ($cb)
134 929
135This method returns the filehandle of the L<AnyEvent::Handle> object. 930This replaces the currently set C<on_read> callback, or clears it (when
136 931the new callback is C<undef>). See the description of C<on_read> in the
137=cut 932constructor.
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.
148 933
149=cut 934=cut
150 935
151sub on_read { 936sub on_read {
152 my ($self, $cb) = @_; 937 my ($self, $cb) = @_;
938
153 $self->{on_read} = $cb; 939 $self->{on_read} = $cb;
940 $self->_drain_rbuf if $cb && !$self->{_in_drain};
941}
154 942
155 unless (defined $self->{on_read}) { 943=item $handle->rbuf
156 delete $self->{on_read_w}; 944
945Returns the read buffer (as a modifiable lvalue).
946
947You can access the read buffer directly as the C<< ->{rbuf} >>
948member, if you want. However, the only operation allowed on the
949read buffer (apart from looking at it) is removing data from its
950beginning. Otherwise modifying or appending to it is not allowed and will
951lead to hard-to-track-down bugs.
952
953NOTE: The read buffer should only be used or modified if the C<on_read>,
954C<push_read> or C<unshift_read> methods are used. The other read methods
955automatically manage the read buffer.
956
957=cut
958
959sub rbuf : lvalue {
960 $_[0]{rbuf}
961}
962
963=item $handle->push_read ($cb)
964
965=item $handle->unshift_read ($cb)
966
967Append the given callback to the end of the queue (C<push_read>) or
968prepend it (C<unshift_read>).
969
970The callback is called each time some additional read data arrives.
971
972It must check whether enough data is in the read buffer already.
973
974If not enough data is available, it must return the empty list or a false
975value, in which case it will be called repeatedly until enough data is
976available (or an error condition is detected).
977
978If enough data was available, then the callback must remove all data it is
979interested in (which can be none at all) and return a true value. After returning
980true, it will be removed from the queue.
981
982=cut
983
984our %RH;
985
986sub register_read_type($$) {
987 $RH{$_[0]} = $_[1];
988}
989
990sub push_read {
991 my $self = shift;
992 my $cb = pop;
993
994 if (@_) {
995 my $type = shift;
996
997 $cb = ($RH{$type} or Carp::croak "unsupported type passed to AnyEvent::Handle::push_read")
998 ->($self, $cb, @_);
999 }
1000
1001 push @{ $self->{_queue} }, $cb;
1002 $self->_drain_rbuf unless $self->{_in_drain};
1003}
1004
1005sub unshift_read {
1006 my $self = shift;
1007 my $cb = pop;
1008
1009 if (@_) {
1010 my $type = shift;
1011
1012 $cb = ($RH{$type} or Carp::croak "unsupported type passed to AnyEvent::Handle::unshift_read")
1013 ->($self, $cb, @_);
1014 }
1015
1016
1017 unshift @{ $self->{_queue} }, $cb;
1018 $self->_drain_rbuf unless $self->{_in_drain};
1019}
1020
1021=item $handle->push_read (type => @args, $cb)
1022
1023=item $handle->unshift_read (type => @args, $cb)
1024
1025Instead of providing a callback that parses the data itself you can chose
1026between a number of predefined parsing formats, for chunks of data, lines
1027etc.
1028
1029Predefined types are (if you have ideas for additional types, feel free to
1030drop by and tell us):
1031
1032=over 4
1033
1034=item chunk => $octets, $cb->($handle, $data)
1035
1036Invoke the callback only once C<$octets> bytes have been read. Pass the
1037data read to the callback. The callback will never be called with less
1038data.
1039
1040Example: read 2 bytes.
1041
1042 $handle->push_read (chunk => 2, sub {
1043 warn "yay ", unpack "H*", $_[1];
1044 });
1045
1046=cut
1047
1048register_read_type chunk => sub {
1049 my ($self, $cb, $len) = @_;
1050
1051 sub {
1052 $len <= length $_[0]{rbuf} or return;
1053 $cb->($_[0], substr $_[0]{rbuf}, 0, $len, "");
1054 1
1055 }
1056};
1057
1058=item line => [$eol, ]$cb->($handle, $line, $eol)
1059
1060The callback will be called only once a full line (including the end of
1061line marker, C<$eol>) has been read. This line (excluding the end of line
1062marker) will be passed to the callback as second argument (C<$line>), and
1063the end of line marker as the third argument (C<$eol>).
1064
1065The end of line marker, C<$eol>, can be either a string, in which case it
1066will be interpreted as a fixed record end marker, or it can be a regex
1067object (e.g. created by C<qr>), in which case it is interpreted as a
1068regular expression.
1069
1070The end of line marker argument C<$eol> is optional, if it is missing (NOT
1071undef), then C<qr|\015?\012|> is used (which is good for most internet
1072protocols).
1073
1074Partial lines at the end of the stream will never be returned, as they are
1075not marked by the end of line marker.
1076
1077=cut
1078
1079register_read_type line => sub {
1080 my ($self, $cb, $eol) = @_;
1081
1082 if (@_ < 3) {
1083 # this is more than twice as fast as the generic code below
1084 sub {
1085 $_[0]{rbuf} =~ s/^([^\015\012]*)(\015?\012)// or return;
1086
1087 $cb->($_[0], $1, $2);
1088 1
1089 }
1090 } else {
1091 $eol = quotemeta $eol unless ref $eol;
1092 $eol = qr|^(.*?)($eol)|s;
1093
1094 sub {
1095 $_[0]{rbuf} =~ s/$eol// or return;
1096
1097 $cb->($_[0], $1, $2);
1098 1
1099 }
1100 }
1101};
1102
1103=item regex => $accept[, $reject[, $skip], $cb->($handle, $data)
1104
1105Makes a regex match against the regex object C<$accept> and returns
1106everything up to and including the match.
1107
1108Example: read a single line terminated by '\n'.
1109
1110 $handle->push_read (regex => qr<\n>, sub { ... });
1111
1112If C<$reject> is given and not undef, then it determines when the data is
1113to be rejected: it is matched against the data when the C<$accept> regex
1114does not match and generates an C<EBADMSG> error when it matches. This is
1115useful to quickly reject wrong data (to avoid waiting for a timeout or a
1116receive buffer overflow).
1117
1118Example: expect a single decimal number followed by whitespace, reject
1119anything else (not the use of an anchor).
1120
1121 $handle->push_read (regex => qr<^[0-9]+\s>, qr<[^0-9]>, sub { ... });
1122
1123If C<$skip> is given and not C<undef>, then it will be matched against
1124the receive buffer when neither C<$accept> nor C<$reject> match,
1125and everything preceding and including the match will be accepted
1126unconditionally. This is useful to skip large amounts of data that you
1127know cannot be matched, so that the C<$accept> or C<$reject> regex do not
1128have to start matching from the beginning. This is purely an optimisation
1129and is usually worth only when you expect more than a few kilobytes.
1130
1131Example: expect a http header, which ends at C<\015\012\015\012>. Since we
1132expect the header to be very large (it isn't in practise, but...), we use
1133a skip regex to skip initial portions. The skip regex is tricky in that
1134it only accepts something not ending in either \015 or \012, as these are
1135required for the accept regex.
1136
1137 $handle->push_read (regex =>
1138 qr<\015\012\015\012>,
1139 undef, # no reject
1140 qr<^.*[^\015\012]>,
1141 sub { ... });
1142
1143=cut
1144
1145register_read_type regex => sub {
1146 my ($self, $cb, $accept, $reject, $skip) = @_;
1147
1148 my $data;
1149 my $rbuf = \$self->{rbuf};
1150
1151 sub {
1152 # accept
1153 if ($$rbuf =~ $accept) {
1154 $data .= substr $$rbuf, 0, $+[0], "";
1155 $cb->($self, $data);
157 return; 1156 return 1;
1157 }
1158
1159 # reject
1160 if ($reject && $$rbuf =~ $reject) {
1161 $self->_error (&Errno::EBADMSG);
1162 }
1163
1164 # skip
1165 if ($skip && $$rbuf =~ $skip) {
1166 $data .= substr $$rbuf, 0, $+[0], "";
1167 }
1168
1169 ()
158 } 1170 }
159 1171};
160 $self->{on_read_w} = 1172
161 AnyEvent->io (poll => 'r', fh => $self->{fh}, cb => sub { 1173=item netstring => $cb->($handle, $string)
162 #d# warn "READ:[$self->{read_size}] $self->{read_block_size} : ".length ($self->{rbuf})."\n"; 1174
163 my $rbuf_len = length $self->{rbuf}; 1175A netstring (http://cr.yp.to/proto/netstrings.txt, this is not an endorsement).
164 my $l; 1176
165 if (defined $self->{read_size}) { 1177Throws an error with C<$!> set to EBADMSG on format violations.
166 $l = sysread $self->{fh}, $self->{rbuf}, 1178
167 ($self->{read_size} - $rbuf_len), $rbuf_len; 1179=cut
168 } else { 1180
169 $l = sysread $self->{fh}, $self->{rbuf}, $self->{read_block_size}, $rbuf_len; 1181register_read_type netstring => sub {
1182 my ($self, $cb) = @_;
1183
1184 sub {
1185 unless ($_[0]{rbuf} =~ s/^(0|[1-9][0-9]*)://) {
1186 if ($_[0]{rbuf} =~ /[^0-9]/) {
1187 $self->_error (&Errno::EBADMSG);
170 } 1188 }
171 #d# warn "READL $l [$self->{rbuf}]\n"; 1189 return;
1190 }
172 1191
173 if (not defined $l) { 1192 my $len = $1;
174 return if $! == EAGAIN || $! == EINTR;
175 $self->{on_error}->($self) if $self->{on_error};
176 delete $self->{on_read_w};
177 1193
178 } elsif ($l == 0) { 1194 $self->unshift_read (chunk => $len, sub {
179 $self->{on_eof}->($self) if $self->{on_eof}; 1195 my $string = $_[1];
180 delete $self->{on_read_w}; 1196 $_[0]->unshift_read (chunk => 1, sub {
181 1197 if ($_[1] eq ",") {
1198 $cb->($_[0], $string);
182 } else { 1199 } else {
183 $self->{on_read}->($self); 1200 $self->_error (&Errno::EBADMSG);
1201 }
1202 });
1203 });
1204
1205 1
1206 }
1207};
1208
1209=item packstring => $format, $cb->($handle, $string)
1210
1211An octet string prefixed with an encoded length. The encoding C<$format>
1212uses the same format as a Perl C<pack> format, but must specify a single
1213integer only (only one of C<cCsSlLqQiInNvVjJw> is allowed, plus an
1214optional C<!>, C<< < >> or C<< > >> modifier).
1215
1216For example, DNS over TCP uses a prefix of C<n> (2 octet network order),
1217EPP uses a prefix of C<N> (4 octtes).
1218
1219Example: read a block of data prefixed by its length in BER-encoded
1220format (very efficient).
1221
1222 $handle->push_read (packstring => "w", sub {
1223 my ($handle, $data) = @_;
1224 });
1225
1226=cut
1227
1228register_read_type packstring => sub {
1229 my ($self, $cb, $format) = @_;
1230
1231 sub {
1232 # when we can use 5.10 we can use ".", but for 5.8 we use the re-pack method
1233 defined (my $len = eval { unpack $format, $_[0]{rbuf} })
1234 or return;
1235
1236 $format = length pack $format, $len;
1237
1238 # bypass unshift if we already have the remaining chunk
1239 if ($format + $len <= length $_[0]{rbuf}) {
1240 my $data = substr $_[0]{rbuf}, $format, $len;
1241 substr $_[0]{rbuf}, 0, $format + $len, "";
1242 $cb->($_[0], $data);
1243 } else {
1244 # remove prefix
1245 substr $_[0]{rbuf}, 0, $format, "";
1246
1247 # read remaining chunk
1248 $_[0]->unshift_read (chunk => $len, $cb);
1249 }
1250
1251 1
1252 }
1253};
1254
1255=item json => $cb->($handle, $hash_or_arrayref)
1256
1257Reads a JSON object or array, decodes it and passes it to the
1258callback. When a parse error occurs, an C<EBADMSG> error will be raised.
1259
1260If a C<json> object was passed to the constructor, then that will be used
1261for the final decode, otherwise it will create a JSON coder expecting UTF-8.
1262
1263This read type uses the incremental parser available with JSON version
12642.09 (and JSON::XS version 2.2) and above. You have to provide a
1265dependency on your own: this module will load the JSON module, but
1266AnyEvent does not depend on it itself.
1267
1268Since JSON texts are fully self-delimiting, the C<json> read and write
1269types are an ideal simple RPC protocol: just exchange JSON datagrams. See
1270the C<json> write type description, above, for an actual example.
1271
1272=cut
1273
1274register_read_type json => sub {
1275 my ($self, $cb) = @_;
1276
1277 my $json = $self->{json} ||=
1278 eval { require JSON::XS; JSON::XS->new->utf8 }
1279 || do { require JSON; JSON->new->utf8 };
1280
1281 my $data;
1282 my $rbuf = \$self->{rbuf};
1283
1284 sub {
1285 my $ref = eval { $json->incr_parse ($self->{rbuf}) };
1286
1287 if ($ref) {
1288 $self->{rbuf} = $json->incr_text;
1289 $json->incr_text = "";
1290 $cb->($self, $ref);
1291
1292 1
1293 } elsif ($@) {
1294 # error case
1295 $json->incr_skip;
1296
1297 $self->{rbuf} = $json->incr_text;
1298 $json->incr_text = "";
1299
1300 $self->_error (&Errno::EBADMSG);
1301
1302 ()
1303 } else {
1304 $self->{rbuf} = "";
1305
1306 ()
1307 }
1308 }
1309};
1310
1311=item storable => $cb->($handle, $ref)
1312
1313Deserialises a L<Storable> frozen representation as written by the
1314C<storable> write type (BER-encoded length prefix followed by nfreeze'd
1315data).
1316
1317Raises C<EBADMSG> error if the data could not be decoded.
1318
1319=cut
1320
1321register_read_type storable => sub {
1322 my ($self, $cb) = @_;
1323
1324 require Storable;
1325
1326 sub {
1327 # when we can use 5.10 we can use ".", but for 5.8 we use the re-pack method
1328 defined (my $len = eval { unpack "w", $_[0]{rbuf} })
1329 or return;
1330
1331 my $format = length pack "w", $len;
1332
1333 # bypass unshift if we already have the remaining chunk
1334 if ($format + $len <= length $_[0]{rbuf}) {
1335 my $data = substr $_[0]{rbuf}, $format, $len;
1336 substr $_[0]{rbuf}, 0, $format + $len, "";
1337 $cb->($_[0], Storable::thaw ($data));
1338 } else {
1339 # remove prefix
1340 substr $_[0]{rbuf}, 0, $format, "";
1341
1342 # read remaining chunk
1343 $_[0]->unshift_read (chunk => $len, sub {
1344 if (my $ref = eval { Storable::thaw ($_[1]) }) {
1345 $cb->($_[0], $ref);
1346 } else {
1347 $self->_error (&Errno::EBADMSG);
1348 }
1349 });
1350 }
1351
1352 1
1353 }
1354};
1355
1356=back
1357
1358=item AnyEvent::Handle::register_read_type type => $coderef->($handle, $cb, @args)
1359
1360This function (not method) lets you add your own types to C<push_read>.
1361
1362Whenever the given C<type> is used, C<push_read> will invoke the code
1363reference with the handle object, the callback and the remaining
1364arguments.
1365
1366The code reference is supposed to return a callback (usually a closure)
1367that works as a plain read callback (see C<< ->push_read ($cb) >>).
1368
1369It should invoke the passed callback when it is done reading (remember to
1370pass C<$handle> as first argument as all other callbacks do that).
1371
1372Note that this is a function, and all types registered this way will be
1373global, so try to use unique names.
1374
1375For examples, see the source of this module (F<perldoc -m AnyEvent::Handle>,
1376search for C<register_read_type>)).
1377
1378=item $handle->stop_read
1379
1380=item $handle->start_read
1381
1382In rare cases you actually do not want to read anything from the
1383socket. In this case you can call C<stop_read>. Neither C<on_read> nor
1384any queued callbacks will be executed then. To start reading again, call
1385C<start_read>.
1386
1387Note that AnyEvent::Handle will automatically C<start_read> for you when
1388you change the C<on_read> callback or push/unshift a read callback, and it
1389will automatically C<stop_read> for you when neither C<on_read> is set nor
1390there are any read requests in the queue.
1391
1392These methods will have no effect when in TLS mode (as TLS doesn't support
1393half-duplex connections).
1394
1395=cut
1396
1397sub stop_read {
1398 my ($self) = @_;
1399
1400 delete $self->{_rw} unless $self->{tls};
1401}
1402
1403sub start_read {
1404 my ($self) = @_;
1405
1406 unless ($self->{_rw} || $self->{_eof}) {
1407 Scalar::Util::weaken $self;
1408
1409 $self->{_rw} = AnyEvent->io (fh => $self->{fh}, poll => "r", cb => sub {
1410 my $rbuf = \($self->{tls} ? my $buf : $self->{rbuf});
1411 my $len = sysread $self->{fh}, $$rbuf, $self->{read_size} || 8192, length $$rbuf;
1412
1413 if ($len > 0) {
1414 $self->{_activity} = AnyEvent->now;
1415
1416 if ($self->{tls}) {
1417 Net::SSLeay::BIO_write ($self->{_rbio}, $$rbuf);
1418
1419 &_dotls ($self);
1420 } else {
1421 $self->_drain_rbuf unless $self->{_in_drain};
1422 }
1423
1424 } elsif (defined $len) {
1425 delete $self->{_rw};
1426 $self->{_eof} = 1;
1427 $self->_drain_rbuf unless $self->{_in_drain};
1428
1429 } elsif ($! != EAGAIN && $! != EINTR && $! != WSAEWOULDBLOCK) {
1430 return $self->_error ($!, 1);
184 } 1431 }
185 }); 1432 });
1433 }
186} 1434}
187 1435
188=item B<on_error ($callback)> 1436our $ERROR_SYSCALL;
1437our $ERROR_WANT_READ;
189 1438
190Whenever a read or write operation resulted in an error the C<$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 { 1439sub _tls_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, $cb) = @_; 1440 my ($self, $err) = @_;
240 1441
241 $self->{read_cb} = $cb; 1442 return $self->_error ($!, 1)
242 my $old_blk_size = $self->{read_block_size}; 1443 if $err == Net::SSLeay::ERROR_SYSCALL ();
243 $self->{read_block_size} = $len;
244 1444
245 $self->on_read (sub { 1445 my $err =Net::SSLeay::ERR_error_string (Net::SSLeay::ERR_get_error ());
246 #d# warn "OFOFO $len || ".length($_[0]->{rbuf})."||\n";
247 1446
248 if ($len == length $_[0]->{rbuf}) { 1447 # reduce error string to look less scary
249 $_[0]->{read_block_size} = $old_blk_size; 1448 $err =~ s/^error:[0-9a-fA-F]{8}:[^:]+:([^:]+):/\L$1: /;
250 $_[0]->on_read (undef); 1449
251 $_[0]->{read_cb}->($_[0], (substr $self->{rbuf}, 0, $len, '')); 1450 if ($self->{_on_starttls}) {
1451 (delete $self->{_on_starttls})->($self, undef, $err);
1452 &_freetls;
1453 } else {
1454 &_freetls;
1455 $self->_error (&Errno::EPROTO, 1, $err);
1456 }
1457}
1458
1459# poll the write BIO and send the data if applicable
1460# also decode read data if possible
1461# this is basiclaly our TLS state machine
1462# more efficient implementations are possible with openssl,
1463# but not with the buggy and incomplete Net::SSLeay.
1464sub _dotls {
1465 my ($self) = @_;
1466
1467 my $tmp;
1468
1469 if (length $self->{_tls_wbuf}) {
1470 while (($tmp = Net::SSLeay::write ($self->{tls}, $self->{_tls_wbuf})) > 0) {
1471 substr $self->{_tls_wbuf}, 0, $tmp, "";
252 } 1472 }
1473
1474 $tmp = Net::SSLeay::get_error ($self->{tls}, $tmp);
1475 return $self->_tls_error ($tmp)
1476 if $tmp != $ERROR_WANT_READ
1477 && ($tmp != $ERROR_SYSCALL || $!);
253 }); 1478 }
254}
255 1479
256=item B<readlines ($callback)> 1480 while (defined ($tmp = Net::SSLeay::read ($self->{tls}))) {
1481 unless (length $tmp) {
1482 $self->{_on_starttls}
1483 and (delete $self->{_on_starttls})->($self, undef, "EOF during handshake"); # ???
1484 &_freetls;
257 1485
258=item B<readlines ($sep, $callback)> 1486 if ($self->{on_stoptls}) {
259 1487 $self->{on_stoptls}($self);
260This method will read lines from the filehandle, seperated by C<$sep> or C<"\n"> 1488 return;
261if C<$sep> is not provided. C<$sep> will be used as part of a regex, so it can be 1489 } else {
262a regex itself and won't be quoted! 1490 # let's treat SSL-eof as we treat normal EOF
263 1491 delete $self->{_rw};
264The C<$callback> will be called when at least one
265line could be read. The first argument to the C<$callback> will be the L<AnyEvent::Handle>
266object itself and the rest of the arguments will be the read lines.
267
268NOTE: This method will override any callbacks installed via the C<on_read> method.
269
270=cut
271
272sub readlines {
273 my ($self, $NL, $cb) = @_;
274
275 if (ref $NL) {
276 $cb = $NL;
277 $NL = "\n";
278 }
279
280 $self->{on_readline} = $cb;
281
282 $self->on_read (sub {
283 my @lines;
284 push @lines, $1 while $_[0]->{rbuf} =~ s/(.*)$NL//;
285 $self->{on_readline}->($_[0], @lines);
286 });
287}
288
289=item B<write ($data)>
290
291=item B<write ($callback)>
292
293=item B<write ($data, $callback)>
294
295This method will write C<$data> to the filehandle and call the C<$callback>
296afterwards. If only C<$callback> is provided it will be called when the
297write buffer becomes empty the next time (or immediately if it already is empty).
298
299=cut
300
301sub write {
302 my ($self, $data, $cb) = @_;
303 if (ref $data) { $cb = $data; undef $data }
304 push @{$self->{write_bufs}}, [$data, $cb];
305 $self->_check_writer;
306}
307
308sub _check_writer {
309 my ($self) = @_;
310
311 if ($self->{write_w}) {
312 unless ($self->{write_cb}) {
313 while (@{$self->{write_bufs}} && not defined $self->{write_bufs}->[0]->[1]) {
314 my $wba = shift @{$self->{write_bufs}};
315 $self->{wbuf} .= $wba->[0]; 1492 $self->{_eof} = 1;
316 } 1493 }
317 } 1494 }
318 return; 1495
1496 $self->{_tls_rbuf} .= $tmp;
1497 $self->_drain_rbuf unless $self->{_in_drain};
1498 $self->{tls} or return; # tls session might have gone away in callback
1499 }
1500
1501 $tmp = Net::SSLeay::get_error ($self->{tls}, -1);
1502 return $self->_tls_error ($tmp)
1503 if $tmp != $ERROR_WANT_READ
1504 && ($tmp != $ERROR_SYSCALL || $!);
1505
1506 while (length ($tmp = Net::SSLeay::BIO_read ($self->{_wbio}))) {
1507 $self->{wbuf} .= $tmp;
1508 $self->_drain_wbuf;
1509 }
1510
1511 $self->{_on_starttls}
1512 and Net::SSLeay::state ($self->{tls}) == Net::SSLeay::ST_OK ()
1513 and (delete $self->{_on_starttls})->($self, 1, "TLS/SSL connection established");
1514}
1515
1516=item $handle->starttls ($tls[, $tls_ctx])
1517
1518Instead of starting TLS negotiation immediately when the AnyEvent::Handle
1519object is created, you can also do that at a later time by calling
1520C<starttls>.
1521
1522The first argument is the same as the C<tls> constructor argument (either
1523C<"connect">, C<"accept"> or an existing Net::SSLeay object).
1524
1525The second argument is the optional C<AnyEvent::TLS> object that is used
1526when AnyEvent::Handle has to create its own TLS connection object, or
1527a hash reference with C<< key => value >> pairs that will be used to
1528construct a new context.
1529
1530The TLS connection object will end up in C<< $handle->{tls} >>, the TLS
1531context in C<< $handle->{tls_ctx} >> after this call and can be used or
1532changed to your liking. Note that the handshake might have already started
1533when this function returns.
1534
1535If it an error to start a TLS handshake more than once per
1536AnyEvent::Handle object (this is due to bugs in OpenSSL).
1537
1538=cut
1539
1540our %TLS_CACHE; #TODO not yet documented, should we?
1541
1542sub starttls {
1543 my ($self, $ssl, $ctx) = @_;
1544
1545 require Net::SSLeay;
1546
1547 Carp::croak "it is an error to call starttls more than once on an AnyEvent::Handle object"
1548 if $self->{tls};
1549
1550 $ERROR_SYSCALL = Net::SSLeay::ERROR_SYSCALL ();
1551 $ERROR_WANT_READ = Net::SSLeay::ERROR_WANT_READ ();
1552
1553 $ctx ||= $self->{tls_ctx};
1554
1555 if ("HASH" eq ref $ctx) {
1556 require AnyEvent::TLS;
1557
1558 local $Carp::CarpLevel = 1; # skip ourselves when creating a new context
1559
1560 if ($ctx->{cache}) {
1561 my $key = $ctx+0;
1562 $ctx = $TLS_CACHE{$key} ||= new AnyEvent::TLS %$ctx;
1563 } else {
1564 $ctx = new AnyEvent::TLS %$ctx;
1565 }
1566 }
319 } 1567
1568 $self->{tls_ctx} = $ctx || TLS_CTX ();
1569 $self->{tls} = $ssl = $self->{tls_ctx}->_get_session ($ssl, $self, $self->{peername});
320 1570
321 my $wba = shift @{$self->{write_bufs}} 1571 # basically, this is deep magic (because SSL_read should have the same issues)
322 or return; 1572 # but the openssl maintainers basically said: "trust us, it just works".
1573 # (unfortunately, we have to hardcode constants because the abysmally misdesigned
1574 # and mismaintained ssleay-module doesn't even offer them).
1575 # http://www.mail-archive.com/openssl-dev@openssl.org/msg22420.html
1576 #
1577 # in short: this is a mess.
1578 #
1579 # note that we do not try to keep the length constant between writes as we are required to do.
1580 # we assume that most (but not all) of this insanity only applies to non-blocking cases,
1581 # and we drive openssl fully in blocking mode here. Or maybe we don't - openssl seems to
1582 # have identity issues in that area.
1583# Net::SSLeay::CTX_set_mode ($ssl,
1584# (eval { local $SIG{__DIE__}; Net::SSLeay::MODE_ENABLE_PARTIAL_WRITE () } || 1)
1585# | (eval { local $SIG{__DIE__}; Net::SSLeay::MODE_ACCEPT_MOVING_WRITE_BUFFER () } || 2));
1586 Net::SSLeay::CTX_set_mode ($ssl, 1|2);
323 1587
324 unless (defined $wba->[0]) { 1588 $self->{_rbio} = Net::SSLeay::BIO_new (Net::SSLeay::BIO_s_mem ());
325 $wba->[1]->($self) if $wba->[1]; 1589 $self->{_wbio} = Net::SSLeay::BIO_new (Net::SSLeay::BIO_s_mem ());
326 $self->_check_writer; 1590
327 return; 1591 Net::SSLeay::set_bio ($ssl, $self->{_rbio}, $self->{_wbio});
1592
1593 $self->{_on_starttls} = sub { $_[0]{on_starttls}(@_) }
1594 if $self->{on_starttls};
1595
1596 &_dotls; # need to trigger the initial handshake
1597 $self->start_read; # make sure we actually do read
1598}
1599
1600=item $handle->stoptls
1601
1602Shuts down the SSL connection - this makes a proper EOF handshake by
1603sending a close notify to the other side, but since OpenSSL doesn't
1604support non-blocking shut downs, it is not possible to re-use the stream
1605afterwards.
1606
1607=cut
1608
1609sub stoptls {
1610 my ($self) = @_;
1611
1612 if ($self->{tls}) {
1613 Net::SSLeay::shutdown ($self->{tls});
1614
1615 &_dotls;
1616
1617# # we don't give a shit. no, we do, but we can't. no...#d#
1618# # we, we... have to use openssl :/#d#
1619# &_freetls;#d#
1620 }
1621}
1622
1623sub _freetls {
1624 my ($self) = @_;
1625
1626 return unless $self->{tls};
1627
1628 $self->{tls_ctx}->_put_session (delete $self->{tls});
328 } 1629
1630 delete @$self{qw(_rbio _wbio _tls_wbuf _on_starttls)};
1631}
329 1632
330 $self->{wbuf} = $wba->[0]; 1633sub DESTROY {
331 $self->{write_cb} = $wba->[1]; 1634 my ($self) = @_;
332 1635
333 $self->{write_w} = 1636 &_freetls;
334 AnyEvent->io (poll => 'w', fh => $self->{fh}, cb => sub { 1637
1638 my $linger = exists $self->{linger} ? $self->{linger} : 3600;
1639
1640 if ($linger && length $self->{wbuf}) {
1641 my $fh = delete $self->{fh};
1642 my $wbuf = delete $self->{wbuf};
1643
1644 my @linger;
1645
1646 push @linger, AnyEvent->io (fh => $fh, poll => "w", cb => sub {
335 my $l = syswrite $self->{fh}, $self->{wbuf}, length $self->{wbuf}; 1647 my $len = syswrite $fh, $wbuf, length $wbuf;
336 1648
337 if (not defined $l) { 1649 if ($len > 0) {
338 return if $! == EAGAIN || $! == EINTR; 1650 substr $wbuf, 0, $len, "";
339 delete $self->{write_w};
340 $self->{on_error}->($self) if $self->{on_error};
341
342 } else { 1651 } else {
343 substr $self->{wbuf}, 0, $l, ''; 1652 @linger = (); # end
344
345 if (length ($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 } 1653 }
355 }); 1654 });
1655 push @linger, AnyEvent->timer (after => $linger, cb => sub {
1656 @linger = ();
1657 });
1658 }
1659}
1660
1661=item $handle->destroy
1662
1663Shuts down the handle object as much as possible - this call ensures that
1664no further callbacks will be invoked and as many resources as possible
1665will be freed. You must not call any methods on the object afterwards.
1666
1667Normally, you can just "forget" any references to an AnyEvent::Handle
1668object and it will simply shut down. This works in fatal error and EOF
1669callbacks, as well as code outside. It does I<NOT> work in a read or write
1670callback, so when you want to destroy the AnyEvent::Handle object from
1671within such an callback. You I<MUST> call C<< ->destroy >> explicitly in
1672that case.
1673
1674Destroying the handle object in this way has the advantage that callbacks
1675will be removed as well, so if those are the only reference holders (as
1676is common), then one doesn't need to do anything special to break any
1677reference cycles.
1678
1679The handle might still linger in the background and write out remaining
1680data, as specified by the C<linger> option, however.
1681
1682=cut
1683
1684sub destroy {
1685 my ($self) = @_;
1686
1687 $self->DESTROY;
1688 %$self = ();
1689}
1690
1691=item AnyEvent::Handle::TLS_CTX
1692
1693This function creates and returns the AnyEvent::TLS object used by default
1694for TLS mode.
1695
1696The context is created by calling L<AnyEvent::TLS> without any arguments.
1697
1698=cut
1699
1700our $TLS_CTX;
1701
1702sub TLS_CTX() {
1703 $TLS_CTX ||= do {
1704 require AnyEvent::TLS;
1705
1706 new AnyEvent::TLS
1707 }
356} 1708}
357 1709
358=back 1710=back
359 1711
1712
1713=head1 NONFREQUENTLY ASKED QUESTIONS
1714
1715=over 4
1716
1717=item I C<undef> the AnyEvent::Handle reference inside my callback and
1718still get further invocations!
1719
1720That's because AnyEvent::Handle keeps a reference to itself when handling
1721read or write callbacks.
1722
1723It is only safe to "forget" the reference inside EOF or error callbacks,
1724from within all other callbacks, you need to explicitly call the C<<
1725->destroy >> method.
1726
1727=item I get different callback invocations in TLS mode/Why can't I pause
1728reading?
1729
1730Unlike, say, TCP, TLS connections do not consist of two independent
1731communication channels, one for each direction. Or put differently. The
1732read and write directions are not independent of each other: you cannot
1733write data unless you are also prepared to read, and vice versa.
1734
1735This can mean than, in TLS mode, you might get C<on_error> or C<on_eof>
1736callback invocations when you are not expecting any read data - the reason
1737is that AnyEvent::Handle always reads in TLS mode.
1738
1739During the connection, you have to make sure that you always have a
1740non-empty read-queue, or an C<on_read> watcher. At the end of the
1741connection (or when you no longer want to use it) you can call the
1742C<destroy> method.
1743
1744=item How do I read data until the other side closes the connection?
1745
1746If you just want to read your data into a perl scalar, the easiest way
1747to achieve this is by setting an C<on_read> callback that does nothing,
1748clearing the C<on_eof> callback and in the C<on_error> callback, the data
1749will be in C<$_[0]{rbuf}>:
1750
1751 $handle->on_read (sub { });
1752 $handle->on_eof (undef);
1753 $handle->on_error (sub {
1754 my $data = delete $_[0]{rbuf};
1755 });
1756
1757The reason to use C<on_error> is that TCP connections, due to latencies
1758and packets loss, might get closed quite violently with an error, when in
1759fact, all data has been received.
1760
1761It is usually better to use acknowledgements when transferring data,
1762to make sure the other side hasn't just died and you got the data
1763intact. This is also one reason why so many internet protocols have an
1764explicit QUIT command.
1765
1766=item I don't want to destroy the handle too early - how do I wait until
1767all data has been written?
1768
1769After writing your last bits of data, set the C<on_drain> callback
1770and destroy the handle in there - with the default setting of
1771C<low_water_mark> this will be called precisely when all data has been
1772written to the socket:
1773
1774 $handle->push_write (...);
1775 $handle->on_drain (sub {
1776 warn "all data submitted to the kernel\n";
1777 undef $handle;
1778 });
1779
1780If you just want to queue some data and then signal EOF to the other side,
1781consider using C<< ->push_shutdown >> instead.
1782
1783=item I want to contact a TLS/SSL server, I don't care about security.
1784
1785If your TLS server is a pure TLS server (e.g. HTTPS) that only speaks TLS,
1786simply connect to it and then create the AnyEvent::Handle with the C<tls>
1787parameter:
1788
1789 tcp_connect $host, $port, sub {
1790 my ($fh) = @_;
1791
1792 my $handle = new AnyEvent::Handle
1793 fh => $fh,
1794 tls => "connect",
1795 on_error => sub { ... };
1796
1797 $handle->push_write (...);
1798 };
1799
1800=item I want to contact a TLS/SSL server, I do care about security.
1801
1802Then you should additionally enable certificate verification, including
1803peername verification, if the protocol you use supports it (see
1804L<AnyEvent::TLS>, C<verify_peername>).
1805
1806E.g. for HTTPS:
1807
1808 tcp_connect $host, $port, sub {
1809 my ($fh) = @_;
1810
1811 my $handle = new AnyEvent::Handle
1812 fh => $fh,
1813 peername => $host,
1814 tls => "connect",
1815 tls_ctx => { verify => 1, verify_peername => "https" },
1816 ...
1817
1818Note that you must specify the hostname you connected to (or whatever
1819"peername" the protocol needs) as the C<peername> argument, otherwise no
1820peername verification will be done.
1821
1822The above will use the system-dependent default set of trusted CA
1823certificates. If you want to check against a specific CA, add the
1824C<ca_file> (or C<ca_cert>) arguments to C<tls_ctx>:
1825
1826 tls_ctx => {
1827 verify => 1,
1828 verify_peername => "https",
1829 ca_file => "my-ca-cert.pem",
1830 },
1831
1832=item I want to create a TLS/SSL server, how do I do that?
1833
1834Well, you first need to get a server certificate and key. You have
1835three options: a) ask a CA (buy one, use cacert.org etc.) b) create a
1836self-signed certificate (cheap. check the search engine of your choice,
1837there are many tutorials on the net) or c) make your own CA (tinyca2 is a
1838nice program for that purpose).
1839
1840Then create a file with your private key (in PEM format, see
1841L<AnyEvent::TLS>), followed by the certificate (also in PEM format). The
1842file should then look like this:
1843
1844 -----BEGIN RSA PRIVATE KEY-----
1845 ...header data
1846 ... lots of base64'y-stuff
1847 -----END RSA PRIVATE KEY-----
1848
1849 -----BEGIN CERTIFICATE-----
1850 ... lots of base64'y-stuff
1851 -----END CERTIFICATE-----
1852
1853The important bits are the "PRIVATE KEY" and "CERTIFICATE" parts. Then
1854specify this file as C<cert_file>:
1855
1856 tcp_server undef, $port, sub {
1857 my ($fh) = @_;
1858
1859 my $handle = new AnyEvent::Handle
1860 fh => $fh,
1861 tls => "accept",
1862 tls_ctx => { cert_file => "my-server-keycert.pem" },
1863 ...
1864
1865When you have intermediate CA certificates that your clients might not
1866know about, just append them to the C<cert_file>.
1867
1868=back
1869
1870
1871=head1 SUBCLASSING AnyEvent::Handle
1872
1873In many cases, you might want to subclass AnyEvent::Handle.
1874
1875To make this easier, a given version of AnyEvent::Handle uses these
1876conventions:
1877
1878=over 4
1879
1880=item * all constructor arguments become object members.
1881
1882At least initially, when you pass a C<tls>-argument to the constructor it
1883will end up in C<< $handle->{tls} >>. Those members might be changed or
1884mutated later on (for example C<tls> will hold the TLS connection object).
1885
1886=item * other object member names are prefixed with an C<_>.
1887
1888All object members not explicitly documented (internal use) are prefixed
1889with an underscore character, so the remaining non-C<_>-namespace is free
1890for use for subclasses.
1891
1892=item * all members not documented here and not prefixed with an underscore
1893are free to use in subclasses.
1894
1895Of course, new versions of AnyEvent::Handle may introduce more "public"
1896member variables, but thats just life, at least it is documented.
1897
1898=back
1899
360=head1 AUTHOR 1900=head1 AUTHOR
361 1901
362Robin Redeker, C<< <elmex at ta-sa.org> >> 1902Robin Redeker C<< <elmex at ta-sa.org> >>, Marc Lehmann <schmorp@schmorp.de>.
363
364=head1 SUPPORT
365
366You can find documentation for this module with the perldoc command.
367
368 perldoc AnyEvent::Handle
369
370You can also look for information at:
371
372=head1 COPYRIGHT & LICENSE
373
374Copyright 2008 Robin Redeker, all rights reserved.
375
376This program is free software; you can redistribute it and/or modify it
377under the same terms as Perl itself.
378 1903
379=cut 1904=cut
380 1905
3811; # End of AnyEvent::Handle 19061; # End of AnyEvent::Handle

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines