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 1.17 by root, Sat May 24 04:17:45 2008 UTC vs.
Revision 1.118 by root, Thu Feb 12 17:33:38 2009 UTC

1package AnyEvent::Handle; 1package AnyEvent::Handle;
2 2
3no warnings; 3no warnings;
4use strict; 4use strict qw(subs vars);
5 5
6use AnyEvent (); 6use AnyEvent ();
7use AnyEvent::Util (); 7use AnyEvent::Util qw(WSAEWOULDBLOCK);
8use Scalar::Util (); 8use Scalar::Util ();
9use Carp (); 9use Carp ();
10use Fcntl (); 10use Fcntl ();
11use Errno qw/EAGAIN EINTR/; 11use Errno qw(EAGAIN EINTR);
12 12
13=head1 NAME 13=head1 NAME
14 14
15AnyEvent::Handle - non-blocking I/O on filehandles via AnyEvent 15AnyEvent::Handle - non-blocking I/O on file handles via AnyEvent
16 16
17This module is experimental.
18
19=cut 17=cut
20 18
21our $VERSION = '0.04'; 19our $VERSION = 4.34;
22 20
23=head1 SYNOPSIS 21=head1 SYNOPSIS
24 22
25 use AnyEvent; 23 use AnyEvent;
26 use AnyEvent::Handle; 24 use AnyEvent::Handle;
27 25
28 my $cv = AnyEvent->condvar; 26 my $cv = AnyEvent->condvar;
29 27
30 my $ae_fh = AnyEvent::Handle->new (fh => \*STDIN); 28 my $handle =
31
32 #TODO
33
34 # or use the constructor to pass the callback:
35
36 my $ae_fh2 =
37 AnyEvent::Handle->new ( 29 AnyEvent::Handle->new (
38 fh => \*STDIN, 30 fh => \*STDIN,
39 on_eof => sub { 31 on_eof => sub {
40 $cv->broadcast; 32 $cv->send;
41 }, 33 },
42 #TODO
43 ); 34 );
44 35
45 $cv->wait; 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;
46 47
47=head1 DESCRIPTION 48=head1 DESCRIPTION
48 49
49This module is a helper module to make it easier to do event-based I/O on 50This 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 51filehandles. For utility functions for doing non-blocking connects and accepts
51on sockets see L<AnyEvent::Util>. 52on sockets see L<AnyEvent::Util>.
52 53
54The L<AnyEvent::Intro> tutorial contains some well-documented
55AnyEvent::Handle examples.
56
53In the following, when the documentation refers to of "bytes" then this 57In the following, when the documentation refers to of "bytes" then this
54means characters. As sysread and syswrite are used for all I/O, their 58means characters. As sysread and syswrite are used for all I/O, their
55treatment of characters applies to this module as well. 59treatment of characters applies to this module as well.
56 60
57All callbacks will be invoked with the handle object as their first 61All callbacks will be invoked with the handle object as their first
69 73
70=item fh => $filehandle [MANDATORY] 74=item fh => $filehandle [MANDATORY]
71 75
72The filehandle this L<AnyEvent::Handle> object will operate on. 76The filehandle this L<AnyEvent::Handle> object will operate on.
73 77
74NOTE: The filehandle will be set to non-blocking (using 78NOTE: The filehandle will be set to non-blocking mode (using
75AnyEvent::Util::fh_nonblocking). 79C<AnyEvent::Util::fh_nonblocking>) by the constructor and needs to stay in
80that mode.
76 81
77=item on_eof => $cb->($self) 82=item on_eof => $cb->($handle)
78 83
79Set the callback to be called on EOF. 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.
80 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
81While not mandatory, it is highly recommended to set an eof callback, 93While not mandatory, it is I<highly> recommended to set an EOF callback,
82otherwise you might end up with a closed socket while you are still 94otherwise you might end up with a closed socket while you are still
83waiting for data. 95waiting for data.
84 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
85=item on_error => $cb->($self) 100=item on_error => $cb->($handle, $fatal)
86 101
87This is the fatal error callback, that is called when, well, a fatal error 102This is the error callback, which is called when, well, some error
88ocurs, such as not being able to resolve the hostname, failure to connect 103occured, such as not being able to resolve the hostname, failure to
89or a read error. 104connect or a read error.
90 105
91The object will not be in a usable state when this callback has been 106Some errors are fatal (which is indicated by C<$fatal> being true). On
92called. 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>).
93 116
94On callback entrance, the value of C<$!> contains the operating system 117On callback entrance, the value of C<$!> contains the operating system
95error (or C<ENOSPC> or C<EPIPE>). 118error (or C<ENOSPC>, C<EPIPE>, C<ETIMEDOUT> or C<EBADMSG>).
96 119
97While not mandatory, it is I<highly> recommended to set this callback, as 120While not mandatory, it is I<highly> recommended to set this callback, as
98you will not be notified of errors otherwise. The default simply calls 121you will not be notified of errors otherwise. The default simply calls
99die. 122C<croak>.
100 123
101=item on_read => $cb->($self) 124=item on_read => $cb->($handle)
102 125
103This sets the default read callback, which is called when data arrives 126This sets the default read callback, which is called when data arrives
104and no read request is in the queue. 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).
105 130
106To access (and remove data from) the read buffer, use the C<< ->rbuf >> 131To access (and remove data from) the read buffer, use the C<< ->rbuf >>
107method or acces sthe C<$self->{rbuf}> member directly. 132method or access the C<$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.
108 135
109When an EOF condition is detected then AnyEvent::Handle will first try to 136When an EOF condition is detected then AnyEvent::Handle will first try to
110feed all the remaining data to the queued callbacks and C<on_read> before 137feed all the remaining data to the queued callbacks and C<on_read> before
111calling the C<on_eof> callback. If no progress can be made, then a fatal 138calling the C<on_eof> callback. If no progress can be made, then a fatal
112error will be raised (with C<$!> set to C<EPIPE>). 139error will be raised (with C<$!> set to C<EPIPE>).
113 140
114=item on_drain => $cb->() 141=item on_drain => $cb->($handle)
115 142
116This sets the callback that is called when the write buffer becomes empty 143This sets the callback that is called when the write buffer becomes empty
117(or when the callback is set and the buffer is empty already). 144(or when the callback is set and the buffer is empty already).
118 145
119To append to the write buffer, use the C<< ->push_write >> method. 146To append to the write buffer, use the C<< ->push_write >> method.
120 147
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
121=item rbuf_max => <bytes> 175=item rbuf_max => <bytes>
122 176
123If defined, then a fatal error will be raised (with C<$!> set to C<ENOSPC>) 177If defined, then a fatal error will be raised (with C<$!> set to C<ENOSPC>)
124when the read buffer ever (strictly) exceeds this size. This is useful to 178when the read buffer ever (strictly) exceeds this size. This is useful to
125avoid denial-of-service attacks. 179avoid some forms of denial-of-service attacks.
126 180
127For example, a server accepting connections from untrusted sources should 181For example, a server accepting connections from untrusted sources should
128be configured to accept only so-and-so much data that it cannot act on 182be configured to accept only so-and-so much data that it cannot act on
129(for example, when expecting a line, an attacker could send an unlimited 183(for example, when expecting a line, an attacker could send an unlimited
130amount of data without a callback ever being called as long as the line 184amount of data without a callback ever being called as long as the line
131isn't finished). 185isn't finished).
132 186
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
133=item read_size => <bytes> 213=item read_size => <bytes>
134 214
135The default read block size (the amount of bytes this module will try to read 215The default read block size (the amount of bytes this module will
136on each [loop iteration). Default: C<4096>. 216try to read during each loop iteration, which affects memory
217requirements). Default: C<8192>.
137 218
138=item low_water_mark => <bytes> 219=item low_water_mark => <bytes>
139 220
140Sets the amount of bytes (default: C<0>) that make up an "empty" write 221Sets the amount of bytes (default: C<0>) that make up an "empty" write
141buffer: If the write reaches this size or gets even samller it is 222buffer: If the write reaches this size or gets even samller it is
142considered empty. 223considered empty.
143 224
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
144=back 286=back
145 287
146=cut 288=cut
147 289
148sub new { 290sub new {
152 294
153 $self->{fh} or Carp::croak "mandatory argument fh is missing"; 295 $self->{fh} or Carp::croak "mandatory argument fh is missing";
154 296
155 AnyEvent::Util::fh_nonblocking $self->{fh}, 1; 297 AnyEvent::Util::fh_nonblocking $self->{fh}, 1;
156 298
157 $self->on_eof (delete $self->{on_eof} ) if $self->{on_eof}; 299 $self->starttls (delete $self->{tls}, delete $self->{tls_ctx})
158 $self->on_error (delete $self->{on_error}) if $self->{on_error}; 300 if $self->{tls};
301
302 $self->{_activity} = AnyEvent->now;
303 $self->_timeout;
304
159 $self->on_drain (delete $self->{on_drain}) if $self->{on_drain}; 305 $self->on_drain (delete $self->{on_drain}) if exists $self->{on_drain};
160 $self->on_read (delete $self->{on_read} ) if $self->{on_read}; 306 $self->no_delay (delete $self->{no_delay}) if exists $self->{no_delay};
161 307
162 $self->start_read; 308 $self->start_read
309 if $self->{on_read};
163 310
164 $self 311 $self
165} 312}
166 313
167sub _shutdown { 314sub _shutdown {
168 my ($self) = @_; 315 my ($self) = @_;
169 316
317 delete $self->{_tw};
170 delete $self->{rw}; 318 delete $self->{_rw};
171 delete $self->{ww}; 319 delete $self->{_ww};
172 delete $self->{fh}; 320 delete $self->{fh};
173}
174 321
322 &_freetls;
323
324 delete $self->{on_read};
325 delete $self->{_queue};
326}
327
175sub error { 328sub _error {
176 my ($self) = @_; 329 my ($self, $errno, $fatal) = @_;
177 330
178 {
179 local $!;
180 $self->_shutdown; 331 $self->_shutdown
181 } 332 if $fatal;
333
334 $! = $errno;
182 335
183 if ($self->{on_error}) { 336 if ($self->{on_error}) {
184 $self->{on_error}($self); 337 $self->{on_error}($self, $fatal);
185 } else { 338 } elsif ($self->{fh}) {
186 die "AnyEvent::Handle uncaught fatal error: $!"; 339 Carp::croak "AnyEvent::Handle uncaught error: $!";
187 } 340 }
188} 341}
189 342
190=item $fh = $handle->fh 343=item $fh = $handle->fh
191 344
192This method returns the filehandle of the L<AnyEvent::Handle> object. 345This method returns the file handle used to create the L<AnyEvent::Handle> object.
193 346
194=cut 347=cut
195 348
196sub fh { $_[0]->{fh} } 349sub fh { $_[0]{fh} }
197 350
198=item $handle->on_error ($cb) 351=item $handle->on_error ($cb)
199 352
200Replace the current C<on_error> callback (see the C<on_error> constructor argument). 353Replace the current C<on_error> callback (see the C<on_error> constructor argument).
201 354
213 366
214sub on_eof { 367sub on_eof {
215 $_[0]{on_eof} = $_[1]; 368 $_[0]{on_eof} = $_[1];
216} 369}
217 370
371=item $handle->on_timeout ($cb)
372
373Replace the current C<on_timeout> callback, or disables the callback (but
374not the timeout) if C<$cb> = C<undef>. See the C<timeout> constructor
375argument and method.
376
377=cut
378
379sub on_timeout {
380 $_[0]{on_timeout} = $_[1];
381}
382
383=item $handle->autocork ($boolean)
384
385Enables or disables the current autocork behaviour (see C<autocork>
386constructor argument). Changes will only take effect on the next write.
387
388=cut
389
390sub autocork {
391 $_[0]{autocork} = $_[1];
392}
393
394=item $handle->no_delay ($boolean)
395
396Enables or disables the C<no_delay> setting (see constructor argument of
397the same name for details).
398
399=cut
400
401sub no_delay {
402 $_[0]{no_delay} = $_[1];
403
404 eval {
405 local $SIG{__DIE__};
406 setsockopt $_[0]{fh}, &Socket::IPPROTO_TCP, &Socket::TCP_NODELAY, int $_[1];
407 };
408}
409
410#############################################################################
411
412=item $handle->timeout ($seconds)
413
414Configures (or disables) the inactivity timeout.
415
416=cut
417
418sub timeout {
419 my ($self, $timeout) = @_;
420
421 $self->{timeout} = $timeout;
422 $self->_timeout;
423}
424
425# reset the timeout watcher, as neccessary
426# also check for time-outs
427sub _timeout {
428 my ($self) = @_;
429
430 if ($self->{timeout}) {
431 my $NOW = AnyEvent->now;
432
433 # when would the timeout trigger?
434 my $after = $self->{_activity} + $self->{timeout} - $NOW;
435
436 # now or in the past already?
437 if ($after <= 0) {
438 $self->{_activity} = $NOW;
439
440 if ($self->{on_timeout}) {
441 $self->{on_timeout}($self);
442 } else {
443 $self->_error (&Errno::ETIMEDOUT);
444 }
445
446 # callback could have changed timeout value, optimise
447 return unless $self->{timeout};
448
449 # calculate new after
450 $after = $self->{timeout};
451 }
452
453 Scalar::Util::weaken $self;
454 return unless $self; # ->error could have destroyed $self
455
456 $self->{_tw} ||= AnyEvent->timer (after => $after, cb => sub {
457 delete $self->{_tw};
458 $self->_timeout;
459 });
460 } else {
461 delete $self->{_tw};
462 }
463}
464
218############################################################################# 465#############################################################################
219 466
220=back 467=back
221 468
222=head2 WRITE QUEUE 469=head2 WRITE QUEUE
225for reading. 472for reading.
226 473
227The write queue is very simple: you can add data to its end, and 474The write queue is very simple: you can add data to its end, and
228AnyEvent::Handle will automatically try to get rid of it for you. 475AnyEvent::Handle will automatically try to get rid of it for you.
229 476
230When data could be writtena nd the write buffer is shorter then the low 477When data could be written and the write buffer is shorter then the low
231water mark, the C<on_drain> callback will be invoked. 478water mark, the C<on_drain> callback will be invoked.
232 479
233=over 4 480=over 4
234 481
235=item $handle->on_drain ($cb) 482=item $handle->on_drain ($cb)
243 my ($self, $cb) = @_; 490 my ($self, $cb) = @_;
244 491
245 $self->{on_drain} = $cb; 492 $self->{on_drain} = $cb;
246 493
247 $cb->($self) 494 $cb->($self)
248 if $cb && $self->{low_water_mark} >= length $self->{wbuf}; 495 if $cb && $self->{low_water_mark} >= (length $self->{wbuf}) + (length $self->{_tls_wbuf});
249} 496}
250 497
251=item $handle->push_write ($data) 498=item $handle->push_write ($data)
252 499
253Queues the given scalar to be written. You can push as much data as you 500Queues the given scalar to be written. You can push as much data as you
257=cut 504=cut
258 505
259sub _drain_wbuf { 506sub _drain_wbuf {
260 my ($self) = @_; 507 my ($self) = @_;
261 508
262 unless ($self->{ww}) { 509 if (!$self->{_ww} && length $self->{wbuf}) {
510
263 Scalar::Util::weaken $self; 511 Scalar::Util::weaken $self;
512
264 my $cb = sub { 513 my $cb = sub {
265 my $len = syswrite $self->{fh}, $self->{wbuf}; 514 my $len = syswrite $self->{fh}, $self->{wbuf};
266 515
267 if ($len > 0) { 516 if ($len >= 0) {
268 substr $self->{wbuf}, 0, $len, ""; 517 substr $self->{wbuf}, 0, $len, "";
269 518
519 $self->{_activity} = AnyEvent->now;
520
270 $self->{on_drain}($self) 521 $self->{on_drain}($self)
271 if $self->{low_water_mark} >= length $self->{wbuf} 522 if $self->{low_water_mark} >= (length $self->{wbuf}) + (length $self->{_tls_wbuf})
272 && $self->{on_drain}; 523 && $self->{on_drain};
273 524
274 delete $self->{ww} unless length $self->{wbuf}; 525 delete $self->{_ww} unless length $self->{wbuf};
275 } elsif ($! != EAGAIN && $! != EINTR) { 526 } elsif ($! != EAGAIN && $! != EINTR && $! != WSAEWOULDBLOCK) {
276 $self->error; 527 $self->_error ($!, 1);
277 } 528 }
278 }; 529 };
279 530
531 # try to write data immediately
532 $cb->() unless $self->{autocork};
533
534 # if still data left in wbuf, we need to poll
280 $self->{ww} = AnyEvent->io (fh => $self->{fh}, poll => "w", cb => $cb); 535 $self->{_ww} = AnyEvent->io (fh => $self->{fh}, poll => "w", cb => $cb)
281 536 if length $self->{wbuf};
282 $cb->($self);
283 }; 537 };
538}
539
540our %WH;
541
542sub register_write_type($$) {
543 $WH{$_[0]} = $_[1];
284} 544}
285 545
286sub push_write { 546sub push_write {
287 my $self = shift; 547 my $self = shift;
288 548
549 if (@_ > 1) {
550 my $type = shift;
551
552 @_ = ($WH{$type} or Carp::croak "unsupported type passed to AnyEvent::Handle::push_write")
553 ->($self, @_);
554 }
555
289 if ($self->{filter_w}) { 556 if ($self->{tls}) {
290 $self->{filter_w}->(\$_[0]); 557 $self->{_tls_wbuf} .= $_[0];
558
559 &_dotls ($self);
291 } else { 560 } else {
292 $self->{wbuf} .= $_[0]; 561 $self->{wbuf} .= $_[0];
293 $self->_drain_wbuf; 562 $self->_drain_wbuf;
294 } 563 }
295} 564}
296 565
566=item $handle->push_write (type => @args)
567
568Instead of formatting your data yourself, you can also let this module do
569the job by specifying a type and type-specific arguments.
570
571Predefined types are (if you have ideas for additional types, feel free to
572drop by and tell us):
573
574=over 4
575
576=item netstring => $string
577
578Formats the given value as netstring
579(http://cr.yp.to/proto/netstrings.txt, this is not a recommendation to use them).
580
581=cut
582
583register_write_type netstring => sub {
584 my ($self, $string) = @_;
585
586 (length $string) . ":$string,"
587};
588
589=item packstring => $format, $data
590
591An octet string prefixed with an encoded length. The encoding C<$format>
592uses the same format as a Perl C<pack> format, but must specify a single
593integer only (only one of C<cCsSlLqQiInNvVjJw> is allowed, plus an
594optional C<!>, C<< < >> or C<< > >> modifier).
595
596=cut
597
598register_write_type packstring => sub {
599 my ($self, $format, $string) = @_;
600
601 pack "$format/a*", $string
602};
603
604=item json => $array_or_hashref
605
606Encodes the given hash or array reference into a JSON object. Unless you
607provide your own JSON object, this means it will be encoded to JSON text
608in UTF-8.
609
610JSON objects (and arrays) are self-delimiting, so you can write JSON at
611one end of a handle and read them at the other end without using any
612additional framing.
613
614The generated JSON text is guaranteed not to contain any newlines: While
615this module doesn't need delimiters after or between JSON texts to be
616able to read them, many other languages depend on that.
617
618A simple RPC protocol that interoperates easily with others is to send
619JSON arrays (or objects, although arrays are usually the better choice as
620they mimic how function argument passing works) and a newline after each
621JSON text:
622
623 $handle->push_write (json => ["method", "arg1", "arg2"]); # whatever
624 $handle->push_write ("\012");
625
626An AnyEvent::Handle receiver would simply use the C<json> read type and
627rely on the fact that the newline will be skipped as leading whitespace:
628
629 $handle->push_read (json => sub { my $array = $_[1]; ... });
630
631Other languages could read single lines terminated by a newline and pass
632this line into their JSON decoder of choice.
633
634=cut
635
636register_write_type json => sub {
637 my ($self, $ref) = @_;
638
639 require JSON;
640
641 $self->{json} ? $self->{json}->encode ($ref)
642 : JSON::encode_json ($ref)
643};
644
645=item storable => $reference
646
647Freezes the given reference using L<Storable> and writes it to the
648handle. Uses the C<nfreeze> format.
649
650=cut
651
652register_write_type storable => sub {
653 my ($self, $ref) = @_;
654
655 require Storable;
656
657 pack "w/a*", Storable::nfreeze ($ref)
658};
659
660=back
661
662=item AnyEvent::Handle::register_write_type type => $coderef->($handle, @args)
663
664This function (not method) lets you add your own types to C<push_write>.
665Whenever the given C<type> is used, C<push_write> will invoke the code
666reference with the handle object and the remaining arguments.
667
668The code reference is supposed to return a single octet string that will
669be appended to the write buffer.
670
671Note that this is a function, and all types registered this way will be
672global, so try to use unique names.
673
674=cut
675
297############################################################################# 676#############################################################################
298 677
299=back 678=back
300 679
301=head2 READ QUEUE 680=head2 READ QUEUE
307ways, the "simple" way, using only C<on_read> and the "complex" way, using 686ways, the "simple" way, using only C<on_read> and the "complex" way, using
308a queue. 687a queue.
309 688
310In the simple case, you just install an C<on_read> callback and whenever 689In the simple case, you just install an C<on_read> callback and whenever
311new data arrives, it will be called. You can then remove some data (if 690new data arrives, it will be called. You can then remove some data (if
312enough is there) from the read buffer (C<< $handle->rbuf >>) if you want 691enough is there) from the read buffer (C<< $handle->rbuf >>). Or you cna
313or not. 692leave the data there if you want to accumulate more (e.g. when only a
693partial message has been received so far).
314 694
315In the more complex case, you want to queue multiple callbacks. In this 695In the more complex case, you want to queue multiple callbacks. In this
316case, AnyEvent::Handle will call the first queued callback each time new 696case, AnyEvent::Handle will call the first queued callback each time new
317data arrives and removes it when it has done its job (see C<push_read>, 697data arrives (also the first time it is queued) and removes it when it has
318below). 698done its job (see C<push_read>, below).
319 699
320This way you can, for example, push three line-reads, followed by reading 700This way you can, for example, push three line-reads, followed by reading
321a chunk of data, and AnyEvent::Handle will execute them in order. 701a chunk of data, and AnyEvent::Handle will execute them in order.
322 702
323Example 1: EPP protocol parser. EPP sends 4 byte length info, followed by 703Example 1: EPP protocol parser. EPP sends 4 byte length info, followed by
324the specified number of bytes which give an XML datagram. 704the specified number of bytes which give an XML datagram.
325 705
326 # in the default state, expect some header bytes 706 # in the default state, expect some header bytes
327 $handle->on_read (sub { 707 $handle->on_read (sub {
328 # some data is here, now queue the length-header-read (4 octets) 708 # some data is here, now queue the length-header-read (4 octets)
329 shift->unshift_read_chunk (4, sub { 709 shift->unshift_read (chunk => 4, sub {
330 # header arrived, decode 710 # header arrived, decode
331 my $len = unpack "N", $_[1]; 711 my $len = unpack "N", $_[1];
332 712
333 # now read the payload 713 # now read the payload
334 shift->unshift_read_chunk ($len, sub { 714 shift->unshift_read (chunk => $len, sub {
335 my $xml = $_[1]; 715 my $xml = $_[1];
336 # handle xml 716 # handle xml
337 }); 717 });
338 }); 718 });
339 }); 719 });
340 720
341Example 2: Implement a client for a protocol that replies either with 721Example 2: Implement a client for a protocol that replies either with "OK"
342"OK" and another line or "ERROR" for one request, and 64 bytes for the 722and another line or "ERROR" for the first request that is sent, and 64
343second request. Due tot he availability of a full queue, we can just 723bytes for the second request. Due to the availability of a queue, we can
344pipeline sending both requests and manipulate the queue as necessary in 724just pipeline sending both requests and manipulate the queue as necessary
345the callbacks: 725in the callbacks.
346 726
347 # request one 727When the first callback is called and sees an "OK" response, it will
728C<unshift> another line-read. This line-read will be queued I<before> the
72964-byte chunk callback.
730
731 # request one, returns either "OK + extra line" or "ERROR"
348 $handle->push_write ("request 1\015\012"); 732 $handle->push_write ("request 1\015\012");
349 733
350 # we expect "ERROR" or "OK" as response, so push a line read 734 # we expect "ERROR" or "OK" as response, so push a line read
351 $handle->push_read_line (sub { 735 $handle->push_read (line => sub {
352 # if we got an "OK", we have to _prepend_ another line, 736 # if we got an "OK", we have to _prepend_ another line,
353 # so it will be read before the second request reads its 64 bytes 737 # so it will be read before the second request reads its 64 bytes
354 # which are already in the queue when this callback is called 738 # which are already in the queue when this callback is called
355 # we don't do this in case we got an error 739 # we don't do this in case we got an error
356 if ($_[1] eq "OK") { 740 if ($_[1] eq "OK") {
357 $_[0]->unshift_read_line (sub { 741 $_[0]->unshift_read (line => sub {
358 my $response = $_[1]; 742 my $response = $_[1];
359 ... 743 ...
360 }); 744 });
361 } 745 }
362 }); 746 });
363 747
364 # request two 748 # request two, simply returns 64 octets
365 $handle->push_write ("request 2\015\012"); 749 $handle->push_write ("request 2\015\012");
366 750
367 # simply read 64 bytes, always 751 # simply read 64 bytes, always
368 $handle->push_read_chunk (64, sub { 752 $handle->push_read (chunk => 64, sub {
369 my $response = $_[1]; 753 my $response = $_[1];
370 ... 754 ...
371 }); 755 });
372 756
373=over 4 757=over 4
374 758
375=cut 759=cut
376 760
377sub _drain_rbuf { 761sub _drain_rbuf {
378 my ($self) = @_; 762 my ($self) = @_;
763
764 local $self->{_in_drain} = 1;
379 765
380 if ( 766 if (
381 defined $self->{rbuf_max} 767 defined $self->{rbuf_max}
382 && $self->{rbuf_max} < length $self->{rbuf} 768 && $self->{rbuf_max} < length $self->{rbuf}
383 ) { 769 ) {
384 $! = &Errno::ENOSPC; return $self->error; 770 $self->_error (&Errno::ENOSPC, 1), return;
385 } 771 }
386 772
387 return if $self->{in_drain}; 773 while () {
388 local $self->{in_drain} = 1; 774 # we need to use a separate tls read buffer, as we must not receive data while
775 # we are draining the buffer, and this can only happen with TLS.
776 $self->{rbuf} .= delete $self->{_tls_rbuf} if exists $self->{_tls_rbuf};
389 777
390 while (my $len = length $self->{rbuf}) { 778 my $len = length $self->{rbuf};
391 no strict 'refs'; 779
392 if (my $cb = shift @{ $self->{queue} }) { 780 if (my $cb = shift @{ $self->{_queue} }) {
393 if (!$cb->($self)) { 781 unless ($cb->($self)) {
394 if ($self->{eof}) { 782 if ($self->{_eof}) {
395 # no progress can be made (not enough data and no data forthcoming) 783 # no progress can be made (not enough data and no data forthcoming)
396 $! = &Errno::EPIPE; return $self->error; 784 $self->_error (&Errno::EPIPE, 1), return;
397 } 785 }
398 786
399 unshift @{ $self->{queue} }, $cb; 787 unshift @{ $self->{_queue} }, $cb;
400 return; 788 last;
401 } 789 }
402 } elsif ($self->{on_read}) { 790 } elsif ($self->{on_read}) {
791 last unless $len;
792
403 $self->{on_read}($self); 793 $self->{on_read}($self);
404 794
405 if ( 795 if (
406 $self->{eof} # if no further data will arrive
407 && $len == length $self->{rbuf} # and no data has been consumed 796 $len == length $self->{rbuf} # if no data has been consumed
408 && !@{ $self->{queue} } # and the queue is still empty 797 && !@{ $self->{_queue} } # and the queue is still empty
409 && $self->{on_read} # and we still want to read data 798 && $self->{on_read} # but we still have on_read
410 ) { 799 ) {
800 # no further data will arrive
411 # then no progress can be made 801 # so no progress can be made
412 $! = &Errno::EPIPE; return $self->error; 802 $self->_error (&Errno::EPIPE, 1), return
803 if $self->{_eof};
804
805 last; # more data might arrive
413 } 806 }
414 } else { 807 } else {
415 # read side becomes idle 808 # read side becomes idle
416 delete $self->{rw}; 809 delete $self->{_rw} unless $self->{tls};
417 return; 810 last;
418 } 811 }
419 } 812 }
420 813
421 if ($self->{eof}) { 814 if ($self->{_eof}) {
422 $self->_shutdown; 815 if ($self->{on_eof}) {
423 $self->{on_eof}($self) 816 $self->{on_eof}($self)
424 if $self->{on_eof}; 817 } else {
818 $self->_error (0, 1);
819 }
820 }
821
822 # may need to restart read watcher
823 unless ($self->{_rw}) {
824 $self->start_read
825 if $self->{on_read} || @{ $self->{_queue} };
425 } 826 }
426} 827}
427 828
428=item $handle->on_read ($cb) 829=item $handle->on_read ($cb)
429 830
435 836
436sub on_read { 837sub on_read {
437 my ($self, $cb) = @_; 838 my ($self, $cb) = @_;
438 839
439 $self->{on_read} = $cb; 840 $self->{on_read} = $cb;
841 $self->_drain_rbuf if $cb && !$self->{_in_drain};
440} 842}
441 843
442=item $handle->rbuf 844=item $handle->rbuf
443 845
444Returns the read buffer (as a modifiable lvalue). 846Returns the read buffer (as a modifiable lvalue).
445 847
446You can access the read buffer directly as the C<< ->{rbuf} >> member, if 848You can access the read buffer directly as the C<< ->{rbuf} >>
447you want. 849member, if you want. However, the only operation allowed on the
850read buffer (apart from looking at it) is removing data from its
851beginning. Otherwise modifying or appending to it is not allowed and will
852lead to hard-to-track-down bugs.
448 853
449NOTE: The read buffer should only be used or modified if the C<on_read>, 854NOTE: The read buffer should only be used or modified if the C<on_read>,
450C<push_read> or C<unshift_read> methods are used. The other read methods 855C<push_read> or C<unshift_read> methods are used. The other read methods
451automatically manage the read buffer. 856automatically manage the read buffer.
452 857
463Append the given callback to the end of the queue (C<push_read>) or 868Append the given callback to the end of the queue (C<push_read>) or
464prepend it (C<unshift_read>). 869prepend it (C<unshift_read>).
465 870
466The callback is called each time some additional read data arrives. 871The callback is called each time some additional read data arrives.
467 872
468It must check wether enough data is in the read buffer already. 873It must check whether enough data is in the read buffer already.
469 874
470If not enough data is available, it must return the empty list or a false 875If not enough data is available, it must return the empty list or a false
471value, in which case it will be called repeatedly until enough data is 876value, in which case it will be called repeatedly until enough data is
472available (or an error condition is detected). 877available (or an error condition is detected).
473 878
475interested in (which can be none at all) and return a true value. After returning 880interested in (which can be none at all) and return a true value. After returning
476true, it will be removed from the queue. 881true, it will be removed from the queue.
477 882
478=cut 883=cut
479 884
885our %RH;
886
887sub register_read_type($$) {
888 $RH{$_[0]} = $_[1];
889}
890
480sub push_read { 891sub push_read {
481 my ($self, $cb) = @_; 892 my $self = shift;
893 my $cb = pop;
482 894
895 if (@_) {
896 my $type = shift;
897
898 $cb = ($RH{$type} or Carp::croak "unsupported type passed to AnyEvent::Handle::push_read")
899 ->($self, $cb, @_);
900 }
901
483 push @{ $self->{queue} }, $cb; 902 push @{ $self->{_queue} }, $cb;
484 $self->_drain_rbuf; 903 $self->_drain_rbuf unless $self->{_in_drain};
485} 904}
486 905
487sub unshift_read { 906sub unshift_read {
488 my ($self, $cb) = @_; 907 my $self = shift;
908 my $cb = pop;
489 909
910 if (@_) {
911 my $type = shift;
912
913 $cb = ($RH{$type} or Carp::croak "unsupported type passed to AnyEvent::Handle::unshift_read")
914 ->($self, $cb, @_);
915 }
916
917
490 push @{ $self->{queue} }, $cb; 918 unshift @{ $self->{_queue} }, $cb;
491 $self->_drain_rbuf; 919 $self->_drain_rbuf unless $self->{_in_drain};
492} 920}
493 921
494=item $handle->push_read_chunk ($len, $cb->($self, $data)) 922=item $handle->push_read (type => @args, $cb)
495 923
496=item $handle->unshift_read_chunk ($len, $cb->($self, $data)) 924=item $handle->unshift_read (type => @args, $cb)
497 925
498Append the given callback to the end of the queue (C<push_read_chunk>) or 926Instead of providing a callback that parses the data itself you can chose
499prepend it (C<unshift_read_chunk>). 927between a number of predefined parsing formats, for chunks of data, lines
928etc.
500 929
501The callback will be called only once C<$len> bytes have been read, and 930Predefined types are (if you have ideas for additional types, feel free to
502these C<$len> bytes will be passed to the callback. 931drop by and tell us):
503 932
504=cut 933=over 4
505 934
506sub _read_chunk($$) { 935=item chunk => $octets, $cb->($handle, $data)
936
937Invoke the callback only once C<$octets> bytes have been read. Pass the
938data read to the callback. The callback will never be called with less
939data.
940
941Example: read 2 bytes.
942
943 $handle->push_read (chunk => 2, sub {
944 warn "yay ", unpack "H*", $_[1];
945 });
946
947=cut
948
949register_read_type chunk => sub {
507 my ($self, $len, $cb) = @_; 950 my ($self, $cb, $len) = @_;
508 951
509 sub { 952 sub {
510 $len <= length $_[0]{rbuf} or return; 953 $len <= length $_[0]{rbuf} or return;
511 $cb->($_[0], substr $_[0]{rbuf}, 0, $len, ""); 954 $cb->($_[0], substr $_[0]{rbuf}, 0, $len, "");
512 1 955 1
513 } 956 }
514} 957};
515 958
516sub push_read_chunk { 959=item line => [$eol, ]$cb->($handle, $line, $eol)
517 $_[0]->push_read (&_read_chunk);
518}
519
520
521sub unshift_read_chunk {
522 $_[0]->unshift_read (&_read_chunk);
523}
524
525=item $handle->push_read_line ([$eol, ]$cb->($self, $line, $eol))
526
527=item $handle->unshift_read_line ([$eol, ]$cb->($self, $line, $eol))
528
529Append the given callback to the end of the queue (C<push_read_line>) or
530prepend it (C<unshift_read_line>).
531 960
532The callback will be called only once a full line (including the end of 961The callback will be called only once a full line (including the end of
533line marker, C<$eol>) has been read. This line (excluding the end of line 962line marker, C<$eol>) has been read. This line (excluding the end of line
534marker) will be passed to the callback as second argument (C<$line>), and 963marker) will be passed to the callback as second argument (C<$line>), and
535the end of line marker as the third argument (C<$eol>). 964the end of line marker as the third argument (C<$eol>).
546Partial lines at the end of the stream will never be returned, as they are 975Partial lines at the end of the stream will never be returned, as they are
547not marked by the end of line marker. 976not marked by the end of line marker.
548 977
549=cut 978=cut
550 979
551sub _read_line($$) { 980register_read_type line => sub {
552 my $self = shift; 981 my ($self, $cb, $eol) = @_;
553 my $cb = pop;
554 my $eol = @_ ? shift : qr|(\015?\012)|;
555 my $pos;
556 982
983 if (@_ < 3) {
984 # this is more than twice as fast as the generic code below
985 sub {
986 $_[0]{rbuf} =~ s/^([^\015\012]*)(\015?\012)// or return;
987
988 $cb->($_[0], $1, $2);
989 1
990 }
991 } else {
557 $eol = quotemeta $eol unless ref $eol; 992 $eol = quotemeta $eol unless ref $eol;
558 $eol = qr|^(.*?)($eol)|s; 993 $eol = qr|^(.*?)($eol)|s;
994
995 sub {
996 $_[0]{rbuf} =~ s/$eol// or return;
997
998 $cb->($_[0], $1, $2);
999 1
1000 }
1001 }
1002};
1003
1004=item regex => $accept[, $reject[, $skip], $cb->($handle, $data)
1005
1006Makes a regex match against the regex object C<$accept> and returns
1007everything up to and including the match.
1008
1009Example: read a single line terminated by '\n'.
1010
1011 $handle->push_read (regex => qr<\n>, sub { ... });
1012
1013If C<$reject> is given and not undef, then it determines when the data is
1014to be rejected: it is matched against the data when the C<$accept> regex
1015does not match and generates an C<EBADMSG> error when it matches. This is
1016useful to quickly reject wrong data (to avoid waiting for a timeout or a
1017receive buffer overflow).
1018
1019Example: expect a single decimal number followed by whitespace, reject
1020anything else (not the use of an anchor).
1021
1022 $handle->push_read (regex => qr<^[0-9]+\s>, qr<[^0-9]>, sub { ... });
1023
1024If C<$skip> is given and not C<undef>, then it will be matched against
1025the receive buffer when neither C<$accept> nor C<$reject> match,
1026and everything preceding and including the match will be accepted
1027unconditionally. This is useful to skip large amounts of data that you
1028know cannot be matched, so that the C<$accept> or C<$reject> regex do not
1029have to start matching from the beginning. This is purely an optimisation
1030and is usually worth only when you expect more than a few kilobytes.
1031
1032Example: expect a http header, which ends at C<\015\012\015\012>. Since we
1033expect the header to be very large (it isn't in practise, but...), we use
1034a skip regex to skip initial portions. The skip regex is tricky in that
1035it only accepts something not ending in either \015 or \012, as these are
1036required for the accept regex.
1037
1038 $handle->push_read (regex =>
1039 qr<\015\012\015\012>,
1040 undef, # no reject
1041 qr<^.*[^\015\012]>,
1042 sub { ... });
1043
1044=cut
1045
1046register_read_type regex => sub {
1047 my ($self, $cb, $accept, $reject, $skip) = @_;
1048
1049 my $data;
1050 my $rbuf = \$self->{rbuf};
559 1051
560 sub { 1052 sub {
561 $_[0]{rbuf} =~ s/$eol// or return; 1053 # accept
1054 if ($$rbuf =~ $accept) {
1055 $data .= substr $$rbuf, 0, $+[0], "";
1056 $cb->($self, $data);
1057 return 1;
1058 }
1059
1060 # reject
1061 if ($reject && $$rbuf =~ $reject) {
1062 $self->_error (&Errno::EBADMSG);
1063 }
562 1064
563 $cb->($_[0], $1, $2); 1065 # skip
1066 if ($skip && $$rbuf =~ $skip) {
1067 $data .= substr $$rbuf, 0, $+[0], "";
1068 }
1069
1070 ()
1071 }
1072};
1073
1074=item netstring => $cb->($handle, $string)
1075
1076A netstring (http://cr.yp.to/proto/netstrings.txt, this is not an endorsement).
1077
1078Throws an error with C<$!> set to EBADMSG on format violations.
1079
1080=cut
1081
1082register_read_type netstring => sub {
1083 my ($self, $cb) = @_;
1084
1085 sub {
1086 unless ($_[0]{rbuf} =~ s/^(0|[1-9][0-9]*)://) {
1087 if ($_[0]{rbuf} =~ /[^0-9]/) {
1088 $self->_error (&Errno::EBADMSG);
1089 }
1090 return;
1091 }
1092
1093 my $len = $1;
1094
1095 $self->unshift_read (chunk => $len, sub {
1096 my $string = $_[1];
1097 $_[0]->unshift_read (chunk => 1, sub {
1098 if ($_[1] eq ",") {
1099 $cb->($_[0], $string);
1100 } else {
1101 $self->_error (&Errno::EBADMSG);
1102 }
1103 });
1104 });
1105
564 1 1106 1
565 } 1107 }
566} 1108};
567 1109
568sub push_read_line { 1110=item packstring => $format, $cb->($handle, $string)
569 $_[0]->push_read (&_read_line);
570}
571 1111
572sub unshift_read_line { 1112An octet string prefixed with an encoded length. The encoding C<$format>
573 $_[0]->unshift_read (&_read_line); 1113uses the same format as a Perl C<pack> format, but must specify a single
574} 1114integer only (only one of C<cCsSlLqQiInNvVjJw> is allowed, plus an
1115optional C<!>, C<< < >> or C<< > >> modifier).
1116
1117For example, DNS over TCP uses a prefix of C<n> (2 octet network order),
1118EPP uses a prefix of C<N> (4 octtes).
1119
1120Example: read a block of data prefixed by its length in BER-encoded
1121format (very efficient).
1122
1123 $handle->push_read (packstring => "w", sub {
1124 my ($handle, $data) = @_;
1125 });
1126
1127=cut
1128
1129register_read_type packstring => sub {
1130 my ($self, $cb, $format) = @_;
1131
1132 sub {
1133 # when we can use 5.10 we can use ".", but for 5.8 we use the re-pack method
1134 defined (my $len = eval { unpack $format, $_[0]{rbuf} })
1135 or return;
1136
1137 $format = length pack $format, $len;
1138
1139 # bypass unshift if we already have the remaining chunk
1140 if ($format + $len <= length $_[0]{rbuf}) {
1141 my $data = substr $_[0]{rbuf}, $format, $len;
1142 substr $_[0]{rbuf}, 0, $format + $len, "";
1143 $cb->($_[0], $data);
1144 } else {
1145 # remove prefix
1146 substr $_[0]{rbuf}, 0, $format, "";
1147
1148 # read remaining chunk
1149 $_[0]->unshift_read (chunk => $len, $cb);
1150 }
1151
1152 1
1153 }
1154};
1155
1156=item json => $cb->($handle, $hash_or_arrayref)
1157
1158Reads a JSON object or array, decodes it and passes it to the
1159callback. When a parse error occurs, an C<EBADMSG> error will be raised.
1160
1161If a C<json> object was passed to the constructor, then that will be used
1162for the final decode, otherwise it will create a JSON coder expecting UTF-8.
1163
1164This read type uses the incremental parser available with JSON version
11652.09 (and JSON::XS version 2.2) and above. You have to provide a
1166dependency on your own: this module will load the JSON module, but
1167AnyEvent does not depend on it itself.
1168
1169Since JSON texts are fully self-delimiting, the C<json> read and write
1170types are an ideal simple RPC protocol: just exchange JSON datagrams. See
1171the C<json> write type description, above, for an actual example.
1172
1173=cut
1174
1175register_read_type json => sub {
1176 my ($self, $cb) = @_;
1177
1178 require JSON;
1179
1180 my $data;
1181 my $rbuf = \$self->{rbuf};
1182
1183 my $json = $self->{json} ||= JSON->new->utf8;
1184
1185 sub {
1186 my $ref = eval { $json->incr_parse ($self->{rbuf}) };
1187
1188 if ($ref) {
1189 $self->{rbuf} = $json->incr_text;
1190 $json->incr_text = "";
1191 $cb->($self, $ref);
1192
1193 1
1194 } elsif ($@) {
1195 # error case
1196 $json->incr_skip;
1197
1198 $self->{rbuf} = $json->incr_text;
1199 $json->incr_text = "";
1200
1201 $self->_error (&Errno::EBADMSG);
1202
1203 ()
1204 } else {
1205 $self->{rbuf} = "";
1206
1207 ()
1208 }
1209 }
1210};
1211
1212=item storable => $cb->($handle, $ref)
1213
1214Deserialises a L<Storable> frozen representation as written by the
1215C<storable> write type (BER-encoded length prefix followed by nfreeze'd
1216data).
1217
1218Raises C<EBADMSG> error if the data could not be decoded.
1219
1220=cut
1221
1222register_read_type storable => sub {
1223 my ($self, $cb) = @_;
1224
1225 require Storable;
1226
1227 sub {
1228 # when we can use 5.10 we can use ".", but for 5.8 we use the re-pack method
1229 defined (my $len = eval { unpack "w", $_[0]{rbuf} })
1230 or return;
1231
1232 my $format = length pack "w", $len;
1233
1234 # bypass unshift if we already have the remaining chunk
1235 if ($format + $len <= length $_[0]{rbuf}) {
1236 my $data = substr $_[0]{rbuf}, $format, $len;
1237 substr $_[0]{rbuf}, 0, $format + $len, "";
1238 $cb->($_[0], Storable::thaw ($data));
1239 } else {
1240 # remove prefix
1241 substr $_[0]{rbuf}, 0, $format, "";
1242
1243 # read remaining chunk
1244 $_[0]->unshift_read (chunk => $len, sub {
1245 if (my $ref = eval { Storable::thaw ($_[1]) }) {
1246 $cb->($_[0], $ref);
1247 } else {
1248 $self->_error (&Errno::EBADMSG);
1249 }
1250 });
1251 }
1252
1253 1
1254 }
1255};
1256
1257=back
1258
1259=item AnyEvent::Handle::register_read_type type => $coderef->($handle, $cb, @args)
1260
1261This function (not method) lets you add your own types to C<push_read>.
1262
1263Whenever the given C<type> is used, C<push_read> will invoke the code
1264reference with the handle object, the callback and the remaining
1265arguments.
1266
1267The code reference is supposed to return a callback (usually a closure)
1268that works as a plain read callback (see C<< ->push_read ($cb) >>).
1269
1270It should invoke the passed callback when it is done reading (remember to
1271pass C<$handle> as first argument as all other callbacks do that).
1272
1273Note that this is a function, and all types registered this way will be
1274global, so try to use unique names.
1275
1276For examples, see the source of this module (F<perldoc -m AnyEvent::Handle>,
1277search for C<register_read_type>)).
575 1278
576=item $handle->stop_read 1279=item $handle->stop_read
577 1280
578=item $handle->start_read 1281=item $handle->start_read
579 1282
580In rare cases you actually do not want to read anything form the 1283In rare cases you actually do not want to read anything from the
581socket. In this case you can call C<stop_read>. Neither C<on_read> no 1284socket. In this case you can call C<stop_read>. Neither C<on_read> nor
582any queued callbacks will be executed then. To start readign again, call 1285any queued callbacks will be executed then. To start reading again, call
583C<start_read>. 1286C<start_read>.
1287
1288Note that AnyEvent::Handle will automatically C<start_read> for you when
1289you change the C<on_read> callback or push/unshift a read callback, and it
1290will automatically C<stop_read> for you when neither C<on_read> is set nor
1291there are any read requests in the queue.
1292
1293These methods will have no effect when in TLS mode (as TLS doesn't support
1294half-duplex connections).
584 1295
585=cut 1296=cut
586 1297
587sub stop_read { 1298sub stop_read {
588 my ($self) = @_; 1299 my ($self) = @_;
589 1300
590 delete $self->{rw}; 1301 delete $self->{_rw} unless $self->{tls};
591} 1302}
592 1303
593sub start_read { 1304sub start_read {
594 my ($self) = @_; 1305 my ($self) = @_;
595 1306
596 unless ($self->{rw} || $self->{eof}) { 1307 unless ($self->{_rw} || $self->{_eof}) {
597 Scalar::Util::weaken $self; 1308 Scalar::Util::weaken $self;
598 1309
599 $self->{rw} = AnyEvent->io (fh => $self->{fh}, poll => "r", cb => sub { 1310 $self->{_rw} = AnyEvent->io (fh => $self->{fh}, poll => "r", cb => sub {
600 my $rbuf = $self->{filter_r} ? \my $buf : \$self->{rbuf}; 1311 my $rbuf = \($self->{tls} ? my $buf : $self->{rbuf});
601 my $len = sysread $self->{fh}, $$rbuf, $self->{read_size} || 8192, length $$rbuf; 1312 my $len = sysread $self->{fh}, $$rbuf, $self->{read_size} || 8192, length $$rbuf;
602 1313
603 if ($len > 0) { 1314 if ($len > 0) {
604 $self->{filter_r} 1315 $self->{_activity} = AnyEvent->now;
605 ? $self->{filter_r}->($rbuf) 1316
606 : $self->_drain_rbuf; 1317 if ($self->{tls}) {
1318 Net::SSLeay::BIO_write ($self->{_rbio}, $$rbuf);
1319
1320 &_dotls ($self);
1321 } else {
1322 $self->_drain_rbuf unless $self->{_in_drain};
1323 }
607 1324
608 } elsif (defined $len) { 1325 } elsif (defined $len) {
609 delete $self->{rw}; 1326 delete $self->{_rw};
610 $self->{eof} = 1; 1327 $self->{_eof} = 1;
611 $self->_drain_rbuf; 1328 $self->_drain_rbuf unless $self->{_in_drain};
612 1329
613 } elsif ($! != EAGAIN && $! != EINTR) { 1330 } elsif ($! != EAGAIN && $! != EINTR && $! != WSAEWOULDBLOCK) {
614 return $self->error; 1331 return $self->_error ($!, 1);
615 } 1332 }
616 }); 1333 });
617 } 1334 }
618} 1335}
619 1336
1337# poll the write BIO and send the data if applicable
1338sub _dotls {
1339 my ($self) = @_;
1340
1341 my $tmp;
1342
1343 if (length $self->{_tls_wbuf}) {
1344 while (($tmp = Net::SSLeay::write ($self->{tls}, $self->{_tls_wbuf})) > 0) {
1345 substr $self->{_tls_wbuf}, 0, $tmp, "";
1346 }
1347 }
1348
1349 while (defined ($tmp = Net::SSLeay::read ($self->{tls}))) {
1350 unless (length $tmp) {
1351 # let's treat SSL-eof as we treat normal EOF
1352 delete $self->{_rw};
1353 $self->{_eof} = 1;
1354 &_freetls;
1355 }
1356
1357 $self->{_tls_rbuf} .= $tmp;
1358 $self->_drain_rbuf unless $self->{_in_drain};
1359 $self->{tls} or return; # tls session might have gone away in callback
1360 }
1361
1362 $tmp = Net::SSLeay::get_error ($self->{tls}, -1);
1363
1364 if ($tmp != Net::SSLeay::ERROR_WANT_READ ()) {
1365 if ($tmp == Net::SSLeay::ERROR_SYSCALL ()) {
1366 return $self->_error ($!, 1);
1367 } elsif ($tmp == Net::SSLeay::ERROR_SSL ()) {
1368 return $self->_error (&Errno::EIO, 1);
1369 }
1370
1371 # all other errors are fine for our purposes
1372 }
1373
1374 while (length ($tmp = Net::SSLeay::BIO_read ($self->{_wbio}))) {
1375 $self->{wbuf} .= $tmp;
1376 $self->_drain_wbuf;
1377 }
1378}
1379
1380=item $handle->starttls ($tls[, $tls_ctx])
1381
1382Instead of starting TLS negotiation immediately when the AnyEvent::Handle
1383object is created, you can also do that at a later time by calling
1384C<starttls>.
1385
1386The first argument is the same as the C<tls> constructor argument (either
1387C<"connect">, C<"accept"> or an existing Net::SSLeay object).
1388
1389The second argument is the optional C<Net::SSLeay::CTX> object that is
1390used when AnyEvent::Handle has to create its own TLS connection object.
1391
1392The TLS connection object will end up in C<< $handle->{tls} >> after this
1393call and can be used or changed to your liking. Note that the handshake
1394might have already started when this function returns.
1395
1396If it an error to start a TLS handshake more than once per
1397AnyEvent::Handle object (this is due to bugs in OpenSSL).
1398
1399=cut
1400
1401sub starttls {
1402 my ($self, $ssl, $ctx) = @_;
1403
1404 require Net::SSLeay;
1405
1406 Carp::croak "it is an error to call starttls more than once on an AnyEvent::Handle object"
1407 if $self->{tls};
1408
1409 if ($ssl eq "accept") {
1410 $ssl = Net::SSLeay::new ($ctx || TLS_CTX ());
1411 Net::SSLeay::set_accept_state ($ssl);
1412 } elsif ($ssl eq "connect") {
1413 $ssl = Net::SSLeay::new ($ctx || TLS_CTX ());
1414 Net::SSLeay::set_connect_state ($ssl);
1415 }
1416
1417 $self->{tls} = $ssl;
1418
1419 # basically, this is deep magic (because SSL_read should have the same issues)
1420 # but the openssl maintainers basically said: "trust us, it just works".
1421 # (unfortunately, we have to hardcode constants because the abysmally misdesigned
1422 # and mismaintained ssleay-module doesn't even offer them).
1423 # http://www.mail-archive.com/openssl-dev@openssl.org/msg22420.html
1424 #
1425 # in short: this is a mess.
1426 #
1427 # note that we do not try to keep the length constant between writes as we are required to do.
1428 # we assume that most (but not all) of this insanity only applies to non-blocking cases,
1429 # and we drive openssl fully in blocking mode here. Or maybe we don't - openssl seems to
1430 # have identity issues in that area.
1431 Net::SSLeay::CTX_set_mode ($self->{tls},
1432 (eval { local $SIG{__DIE__}; Net::SSLeay::MODE_ENABLE_PARTIAL_WRITE () } || 1)
1433 | (eval { local $SIG{__DIE__}; Net::SSLeay::MODE_ACCEPT_MOVING_WRITE_BUFFER () } || 2));
1434
1435 $self->{_rbio} = Net::SSLeay::BIO_new (Net::SSLeay::BIO_s_mem ());
1436 $self->{_wbio} = Net::SSLeay::BIO_new (Net::SSLeay::BIO_s_mem ());
1437
1438 Net::SSLeay::set_bio ($ssl, $self->{_rbio}, $self->{_wbio});
1439
1440 &_dotls; # need to trigger the initial handshake
1441 $self->start_read; # make sure we actually do read
1442}
1443
1444=item $handle->stoptls
1445
1446Shuts down the SSL connection - this makes a proper EOF handshake by
1447sending a close notify to the other side, but since OpenSSL doesn't
1448support non-blocking shut downs, it is not possible to re-use the stream
1449afterwards.
1450
1451=cut
1452
1453sub stoptls {
1454 my ($self) = @_;
1455
1456 if ($self->{tls}) {
1457 Net::SSLeay::shutdown ($self->{tls});
1458
1459 &_dotls;
1460
1461 # we don't give a shit. no, we do, but we can't. no...
1462 # we, we... have to use openssl :/
1463 &_freetls;
1464 }
1465}
1466
1467sub _freetls {
1468 my ($self) = @_;
1469
1470 return unless $self->{tls};
1471
1472 Net::SSLeay::free (delete $self->{tls});
1473
1474 delete @$self{qw(_rbio _wbio _tls_wbuf)};
1475}
1476
1477sub DESTROY {
1478 my $self = shift;
1479
1480 &_freetls;
1481
1482 my $linger = exists $self->{linger} ? $self->{linger} : 3600;
1483
1484 if ($linger && length $self->{wbuf}) {
1485 my $fh = delete $self->{fh};
1486 my $wbuf = delete $self->{wbuf};
1487
1488 my @linger;
1489
1490 push @linger, AnyEvent->io (fh => $fh, poll => "w", cb => sub {
1491 my $len = syswrite $fh, $wbuf, length $wbuf;
1492
1493 if ($len > 0) {
1494 substr $wbuf, 0, $len, "";
1495 } else {
1496 @linger = (); # end
1497 }
1498 });
1499 push @linger, AnyEvent->timer (after => $linger, cb => sub {
1500 @linger = ();
1501 });
1502 }
1503}
1504
1505=item $handle->destroy
1506
1507Shuts down the handle object as much as possible - this call ensures that
1508no further callbacks will be invoked and resources will be freed as much
1509as possible. You must not call any methods on the object afterwards.
1510
1511Normally, you can just "forget" any references to an AnyEvent::Handle
1512object and it will simply shut down. This works in fatal error and EOF
1513callbacks, as well as code outside. It does I<NOT> work in a read or write
1514callback, so when you want to destroy the AnyEvent::Handle object from
1515within such an callback. You I<MUST> call C<< ->destroy >> explicitly in
1516that case.
1517
1518The handle might still linger in the background and write out remaining
1519data, as specified by the C<linger> option, however.
1520
1521=cut
1522
1523sub destroy {
1524 my ($self) = @_;
1525
1526 $self->DESTROY;
1527 %$self = ();
1528}
1529
1530=item AnyEvent::Handle::TLS_CTX
1531
1532This function creates and returns the Net::SSLeay::CTX object used by
1533default for TLS mode.
1534
1535The context is created like this:
1536
1537 Net::SSLeay::load_error_strings;
1538 Net::SSLeay::SSLeay_add_ssl_algorithms;
1539 Net::SSLeay::randomize;
1540
1541 my $CTX = Net::SSLeay::CTX_new;
1542
1543 Net::SSLeay::CTX_set_options $CTX, Net::SSLeay::OP_ALL
1544
1545=cut
1546
1547our $TLS_CTX;
1548
1549sub TLS_CTX() {
1550 $TLS_CTX || do {
1551 require Net::SSLeay;
1552
1553 Net::SSLeay::load_error_strings ();
1554 Net::SSLeay::SSLeay_add_ssl_algorithms ();
1555 Net::SSLeay::randomize ();
1556
1557 $TLS_CTX = Net::SSLeay::CTX_new ();
1558
1559 Net::SSLeay::CTX_set_options ($TLS_CTX, Net::SSLeay::OP_ALL ());
1560
1561 $TLS_CTX
1562 }
1563}
1564
620=back 1565=back
621 1566
1567
1568=head1 NONFREQUENTLY ASKED QUESTIONS
1569
1570=over 4
1571
1572=item I C<undef> the AnyEvent::Handle reference inside my callback and
1573still get further invocations!
1574
1575That's because AnyEvent::Handle keeps a reference to itself when handling
1576read or write callbacks.
1577
1578It is only safe to "forget" the reference inside EOF or error callbacks,
1579from within all other callbacks, you need to explicitly call the C<<
1580->destroy >> method.
1581
1582=item I get different callback invocations in TLS mode/Why can't I pause
1583reading?
1584
1585Unlike, say, TCP, TLS connections do not consist of two independent
1586communication channels, one for each direction. Or put differently. The
1587read and write directions are not independent of each other: you cannot
1588write data unless you are also prepared to read, and vice versa.
1589
1590This can mean than, in TLS mode, you might get C<on_error> or C<on_eof>
1591callback invocations when you are not expecting any read data - the reason
1592is that AnyEvent::Handle always reads in TLS mode.
1593
1594During the connection, you have to make sure that you always have a
1595non-empty read-queue, or an C<on_read> watcher. At the end of the
1596connection (or when you no longer want to use it) you can call the
1597C<destroy> method.
1598
1599=item How do I read data until the other side closes the connection?
1600
1601If you just want to read your data into a perl scalar, the easiest way
1602to achieve this is by setting an C<on_read> callback that does nothing,
1603clearing the C<on_eof> callback and in the C<on_error> callback, the data
1604will be in C<$_[0]{rbuf}>:
1605
1606 $handle->on_read (sub { });
1607 $handle->on_eof (undef);
1608 $handle->on_error (sub {
1609 my $data = delete $_[0]{rbuf};
1610 undef $handle;
1611 });
1612
1613The reason to use C<on_error> is that TCP connections, due to latencies
1614and packets loss, might get closed quite violently with an error, when in
1615fact, all data has been received.
1616
1617It is usually better to use acknowledgements when transferring data,
1618to make sure the other side hasn't just died and you got the data
1619intact. This is also one reason why so many internet protocols have an
1620explicit QUIT command.
1621
1622=item I don't want to destroy the handle too early - how do I wait until
1623all data has been written?
1624
1625After writing your last bits of data, set the C<on_drain> callback
1626and destroy the handle in there - with the default setting of
1627C<low_water_mark> this will be called precisely when all data has been
1628written to the socket:
1629
1630 $handle->push_write (...);
1631 $handle->on_drain (sub {
1632 warn "all data submitted to the kernel\n";
1633 undef $handle;
1634 });
1635
1636=back
1637
1638
1639=head1 SUBCLASSING AnyEvent::Handle
1640
1641In many cases, you might want to subclass AnyEvent::Handle.
1642
1643To make this easier, a given version of AnyEvent::Handle uses these
1644conventions:
1645
1646=over 4
1647
1648=item * all constructor arguments become object members.
1649
1650At least initially, when you pass a C<tls>-argument to the constructor it
1651will end up in C<< $handle->{tls} >>. Those members might be changed or
1652mutated later on (for example C<tls> will hold the TLS connection object).
1653
1654=item * other object member names are prefixed with an C<_>.
1655
1656All object members not explicitly documented (internal use) are prefixed
1657with an underscore character, so the remaining non-C<_>-namespace is free
1658for use for subclasses.
1659
1660=item * all members not documented here and not prefixed with an underscore
1661are free to use in subclasses.
1662
1663Of course, new versions of AnyEvent::Handle may introduce more "public"
1664member variables, but thats just life, at least it is documented.
1665
1666=back
1667
622=head1 AUTHOR 1668=head1 AUTHOR
623 1669
624Robin Redeker C<< <elmex at ta-sa.org> >>, Marc Lehmann <schmorp@schmorp.de>. 1670Robin Redeker C<< <elmex at ta-sa.org> >>, Marc Lehmann <schmorp@schmorp.de>.
625 1671
626=cut 1672=cut

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines