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 1.11 by root, Sun May 11 17:54:13 2008 UTC vs.
Revision 1.134 by root, Fri Jul 3 00:09:04 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
17=cut 17=cut
18 18
19our $VERSION = '0.02'; 19our $VERSION = 4.45;
20 20
21=head1 SYNOPSIS 21=head1 SYNOPSIS
22 22
23 use AnyEvent; 23 use AnyEvent;
24 use AnyEvent::Handle; 24 use AnyEvent::Handle;
25 25
26 my $cv = AnyEvent->condvar; 26 my $cv = AnyEvent->condvar;
27 27
28 my $ae_fh = AnyEvent::Handle->new (fh => \*STDIN); 28 my $handle =
29
30 #TODO
31
32 # or use the constructor to pass the callback:
33
34 my $ae_fh2 =
35 AnyEvent::Handle->new ( 29 AnyEvent::Handle->new (
36 fh => \*STDIN, 30 fh => \*STDIN,
37 on_eof => sub { 31 on_eof => sub {
38 $cv->broadcast; 32 $cv->send;
39 }, 33 },
40 #TODO
41 ); 34 );
42 35
43 $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;
44 47
45=head1 DESCRIPTION 48=head1 DESCRIPTION
46 49
47This 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
48filehandles (and sockets, see L<AnyEvent::Socket> for an easy way to make 51filehandles. For utility functions for doing non-blocking connects and accepts
49non-blocking resolves and connects). 52on sockets see L<AnyEvent::Util>.
53
54The L<AnyEvent::Intro> tutorial contains some well-documented
55AnyEvent::Handle examples.
50 56
51In the following, when the documentation refers to of "bytes" then this 57In the following, when the documentation refers to of "bytes" then this
52means 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
53treatment of characters applies to this module as well. 59treatment of characters applies to this module as well.
54 60
57 63
58=head1 METHODS 64=head1 METHODS
59 65
60=over 4 66=over 4
61 67
62=item B<new (%args)> 68=item $handle = B<new> AnyEvent::TLS fh => $filehandle, key => value...
63 69
64The constructor supports these arguments (all as key => value pairs). 70The constructor supports these arguments (all as C<< key => value >> pairs).
65 71
66=over 4 72=over 4
67 73
68=item fh => $filehandle [MANDATORY] 74=item fh => $filehandle [MANDATORY]
69 75
70The filehandle this L<AnyEvent::Handle> object will operate on. 76The filehandle this L<AnyEvent::Handle> object will operate on.
71 77
72NOTE: The filehandle will be set to non-blocking (using 78NOTE: The filehandle will be set to non-blocking mode (using
73AnyEvent::Util::fh_nonblocking). 79C<AnyEvent::Util::fh_nonblocking>) by the constructor and needs to stay in
80that mode.
74 81
75=item on_eof => $cb->($self) [MANDATORY]
76
77Set the callback to be called on EOF.
78
79=item on_error => $cb->($self) 82=item on_eof => $cb->($handle)
80 83
84Set the callback to be called when an end-of-file condition is detected,
85i.e. in the case of a socket, when the other side has closed the
86connection cleanly.
87
88For sockets, this just means that the other side has stopped sending data,
89you can still try to write data, and, in fact, one can return from the EOF
90callback and continue writing data, as only the read part has been shut
91down.
92
93While not mandatory, it is I<highly> recommended to set an EOF callback,
94otherwise you might end up with a closed socket while you are still
95waiting for data.
96
97If an EOF condition has been detected but no C<on_eof> callback has been
98set, then a fatal error will be raised with C<$!> set to <0>.
99
100=item on_error => $cb->($handle, $fatal, $message)
101
81This is the fatal error callback, that is called when, well, a fatal error 102This is the error callback, which is called when, well, some error
82ocurs, such as not being able to resolve the hostname, failure to connect 103occured, such as not being able to resolve the hostname, failure to
83or a read error. 104connect or a read error.
84 105
85The object will not be in a usable state when this callback has been 106Some errors are fatal (which is indicated by C<$fatal> being true). On
86called. 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
112AnyEvent::Handle tries to find an appropriate error code for you to check
113against, but in some cases (TLS errors), this does not work well. It is
114recommended to always output the C<$message> argument in human-readable
115error messages (it's usually the same as C<"$!">).
116
117Non-fatal errors can be retried by simply returning, but it is recommended
118to simply ignore this parameter and instead abondon the handle object
119when this callback is invoked. Examples of non-fatal errors are timeouts
120C<ETIMEDOUT>) or badly-formatted data (C<EBADMSG>).
87 121
88On callback entrance, the value of C<$!> contains the operating system 122On callback entrance, the value of C<$!> contains the operating system
89error (or C<ENOSPC> or C<EPIPE>). 123error code (or C<ENOSPC>, C<EPIPE>, C<ETIMEDOUT>, C<EBADMSG> or
124C<EPROTO>).
90 125
91While not mandatory, it is I<highly> recommended to set this callback, as 126While not mandatory, it is I<highly> recommended to set this callback, as
92you will not be notified of errors otherwise. The default simply calls 127you will not be notified of errors otherwise. The default simply calls
93die. 128C<croak>.
94 129
95=item on_read => $cb->($self) 130=item on_read => $cb->($handle)
96 131
97This sets the default read callback, which is called when data arrives 132This sets the default read callback, which is called when data arrives
98and no read request is in the queue. 133and no read request is in the queue (unlike read queue callbacks, this
134callback will only be called when at least one octet of data is in the
135read buffer).
99 136
100To access (and remove data from) the read buffer, use the C<< ->rbuf >> 137To access (and remove data from) the read buffer, use the C<< ->rbuf >>
101method or acces sthe C<$self->{rbuf}> member directly. 138method or access the C<$handle->{rbuf}> member directly. Note that you
139must not enlarge or modify the read buffer, you can only remove data at
140the beginning from it.
102 141
103When an EOF condition is detected then AnyEvent::Handle will first try to 142When an EOF condition is detected then AnyEvent::Handle will first try to
104feed all the remaining data to the queued callbacks and C<on_read> before 143feed all the remaining data to the queued callbacks and C<on_read> before
105calling the C<on_eof> callback. If no progress can be made, then a fatal 144calling the C<on_eof> callback. If no progress can be made, then a fatal
106error will be raised (with C<$!> set to C<EPIPE>). 145error will be raised (with C<$!> set to C<EPIPE>).
107 146
108=item on_drain => $cb->() 147=item on_drain => $cb->($handle)
109 148
110This sets the callback that is called when the write buffer becomes empty 149This sets the callback that is called when the write buffer becomes empty
111(or when the callback is set and the buffer is empty already). 150(or when the callback is set and the buffer is empty already).
112 151
113To append to the write buffer, use the C<< ->push_write >> method. 152To append to the write buffer, use the C<< ->push_write >> method.
114 153
154This callback is useful when you don't want to put all of your write data
155into the queue at once, for example, when you want to write the contents
156of some file to the socket you might not want to read the whole file into
157memory and push it into the queue, but instead only read more data from
158the file when the write queue becomes empty.
159
160=item timeout => $fractional_seconds
161
162If non-zero, then this enables an "inactivity" timeout: whenever this many
163seconds pass without a successful read or write on the underlying file
164handle, the C<on_timeout> callback will be invoked (and if that one is
165missing, a non-fatal C<ETIMEDOUT> error will be raised).
166
167Note that timeout processing is also active when you currently do not have
168any outstanding read or write requests: If you plan to keep the connection
169idle then you should disable the timout temporarily or ignore the timeout
170in the C<on_timeout> callback, in which case AnyEvent::Handle will simply
171restart the timeout.
172
173Zero (the default) disables this timeout.
174
175=item on_timeout => $cb->($handle)
176
177Called whenever the inactivity timeout passes. If you return from this
178callback, then the timeout will be reset as if some activity had happened,
179so this condition is not fatal in any way.
180
115=item rbuf_max => <bytes> 181=item rbuf_max => <bytes>
116 182
117If defined, then a fatal error will be raised (with C<$!> set to C<ENOSPC>) 183If defined, then a fatal error will be raised (with C<$!> set to C<ENOSPC>)
118when the read buffer ever (strictly) exceeds this size. This is useful to 184when the read buffer ever (strictly) exceeds this size. This is useful to
119avoid denial-of-service attacks. 185avoid some forms of denial-of-service attacks.
120 186
121For example, a server accepting connections from untrusted sources should 187For example, a server accepting connections from untrusted sources should
122be configured to accept only so-and-so much data that it cannot act on 188be configured to accept only so-and-so much data that it cannot act on
123(for example, when expecting a line, an attacker could send an unlimited 189(for example, when expecting a line, an attacker could send an unlimited
124amount of data without a callback ever being called as long as the line 190amount of data without a callback ever being called as long as the line
125isn't finished). 191isn't finished).
126 192
193=item autocork => <boolean>
194
195When disabled (the default), then C<push_write> will try to immediately
196write the data to the handle, if possible. This avoids having to register
197a write watcher and wait for the next event loop iteration, but can
198be inefficient if you write multiple small chunks (on the wire, this
199disadvantage is usually avoided by your kernel's nagle algorithm, see
200C<no_delay>, but this option can save costly syscalls).
201
202When enabled, then writes will always be queued till the next event loop
203iteration. This is efficient when you do many small writes per iteration,
204but less efficient when you do a single write only per iteration (or when
205the write buffer often is full). It also increases write latency.
206
207=item no_delay => <boolean>
208
209When doing small writes on sockets, your operating system kernel might
210wait a bit for more data before actually sending it out. This is called
211the Nagle algorithm, and usually it is beneficial.
212
213In some situations you want as low a delay as possible, which can be
214accomplishd by setting this option to a true value.
215
216The default is your opertaing system's default behaviour (most likely
217enabled), this option explicitly enables or disables it, if possible.
218
127=item read_size => <bytes> 219=item read_size => <bytes>
128 220
129The default read block size (the amount of bytes this module will try to read 221The default read block size (the amount of bytes this module will
130on each [loop iteration). Default: C<4096>. 222try to read during each loop iteration, which affects memory
223requirements). Default: C<8192>.
131 224
132=item low_water_mark => <bytes> 225=item low_water_mark => <bytes>
133 226
134Sets the amount of bytes (default: C<0>) that make up an "empty" write 227Sets the amount of bytes (default: C<0>) that make up an "empty" write
135buffer: If the write reaches this size or gets even samller it is 228buffer: If the write reaches this size or gets even samller it is
136considered empty. 229considered empty.
137 230
231Sometimes it can be beneficial (for performance reasons) to add data to
232the write buffer before it is fully drained, but this is a rare case, as
233the operating system kernel usually buffers data as well, so the default
234is good in almost all cases.
235
236=item linger => <seconds>
237
238If non-zero (default: C<3600>), then the destructor of the
239AnyEvent::Handle object will check whether there is still outstanding
240write data and will install a watcher that will write this data to the
241socket. No errors will be reported (this mostly matches how the operating
242system treats outstanding data at socket close time).
243
244This will not work for partial TLS data that could not be encoded
245yet. This data will be lost. Calling the C<stoptls> method in time might
246help.
247
248=item peername => $string
249
250A string used to identify the remote site - usually the DNS hostname
251(I<not> IDN!) used to create the connection, rarely the IP address.
252
253Apart from being useful in error messages, this string is also used in TLS
254common name verification (see C<verify_cn> in L<AnyEvent::TLS>).
255
256=item tls => "accept" | "connect" | Net::SSLeay::SSL object
257
258When this parameter is given, it enables TLS (SSL) mode, that means
259AnyEvent will start a TLS handshake as soon as the conenction has been
260established and will transparently encrypt/decrypt data afterwards.
261
262All TLS protocol errors will be signalled as C<EPROTO>, with an
263appropriate error message.
264
265TLS mode requires Net::SSLeay to be installed (it will be loaded
266automatically when you try to create a TLS handle): this module doesn't
267have a dependency on that module, so if your module requires it, you have
268to add the dependency yourself.
269
270Unlike TCP, TLS has a server and client side: for the TLS server side, use
271C<accept>, and for the TLS client side of a connection, use C<connect>
272mode.
273
274You can also provide your own TLS connection object, but you have
275to make sure that you call either C<Net::SSLeay::set_connect_state>
276or C<Net::SSLeay::set_accept_state> on it before you pass it to
277AnyEvent::Handle. Also, this module will take ownership of this connection
278object.
279
280At some future point, AnyEvent::Handle might switch to another TLS
281implementation, then the option to use your own session object will go
282away.
283
284B<IMPORTANT:> since Net::SSLeay "objects" are really only integers,
285passing in the wrong integer will lead to certain crash. This most often
286happens when one uses a stylish C<< tls => 1 >> and is surprised about the
287segmentation fault.
288
289See the C<< ->starttls >> method for when need to start TLS negotiation later.
290
291=item tls_ctx => $anyevent_tls
292
293Use the given C<AnyEvent::TLS> object to create the new TLS connection
294(unless a connection object was specified directly). If this parameter is
295missing, then AnyEvent::Handle will use C<AnyEvent::Handle::TLS_CTX>.
296
297Instead of an object, you can also specify a hash reference with C<< key
298=> value >> pairs. Those will be passed to L<AnyEvent::TLS> to create a
299new TLS context object.
300
301=item json => JSON or JSON::XS object
302
303This is the json coder object used by the C<json> read and write types.
304
305If you don't supply it, then AnyEvent::Handle will create and use a
306suitable one (on demand), which will write and expect UTF-8 encoded JSON
307texts.
308
309Note that you are responsible to depend on the JSON module if you want to
310use this functionality, as AnyEvent does not have a dependency itself.
311
138=back 312=back
139 313
140=cut 314=cut
141 315
142sub new { 316sub new {
143 my $class = shift; 317 my $class = shift;
144
145 my $self = bless { @_ }, $class; 318 my $self = bless { @_ }, $class;
146 319
147 $self->{fh} or Carp::croak "mandatory argument fh is missing"; 320 $self->{fh} or Carp::croak "mandatory argument fh is missing";
148 321
149 AnyEvent::Util::fh_nonblocking $self->{fh}, 1; 322 AnyEvent::Util::fh_nonblocking $self->{fh}, 1;
150 323
151 $self->on_eof ((delete $self->{on_eof} ) or Carp::croak "mandatory argument on_eof is missing"); 324 $self->{_activity} = AnyEvent->now;
325 $self->_timeout;
152 326
153 $self->on_error (delete $self->{on_error}) if $self->{on_error}; 327 $self->no_delay (delete $self->{no_delay}) if exists $self->{no_delay};
328
329 $self->starttls (delete $self->{tls}, delete $self->{tls_ctx})
330 if $self->{tls};
331
154 $self->on_drain (delete $self->{on_drain}) if $self->{on_drain}; 332 $self->on_drain (delete $self->{on_drain}) if exists $self->{on_drain};
155 $self->on_read (delete $self->{on_read} ) if $self->{on_read};
156 333
157 $self->start_read; 334 $self->start_read
335 if $self->{on_read};
158 336
159 $self 337 $self->{fh} && $self
160} 338}
161 339
162sub _shutdown { 340sub _shutdown {
163 my ($self) = @_; 341 my ($self) = @_;
164 342
165 delete $self->{rw}; 343 delete @$self{qw(_tw _rw _ww fh wbuf on_read _queue)};
166 delete $self->{ww}; 344 $self->{_eof} = 1; # tell starttls et. al to stop trying
167 delete $self->{fh};
168}
169 345
346 &_freetls;
347}
348
170sub error { 349sub _error {
171 my ($self) = @_; 350 my ($self, $errno, $fatal, $message) = @_;
172 351
173 {
174 local $!;
175 $self->_shutdown; 352 $self->_shutdown
176 } 353 if $fatal;
354
355 $! = $errno;
356 $message ||= "$!";
177 357
178 if ($self->{on_error}) { 358 if ($self->{on_error}) {
179 $self->{on_error}($self); 359 $self->{on_error}($self, $fatal, $message);
180 } else { 360 } elsif ($self->{fh}) {
181 die "AnyEvent::Handle uncaught fatal error: $!"; 361 Carp::croak "AnyEvent::Handle uncaught error: $message";
182 } 362 }
183} 363}
184 364
185=item $fh = $handle->fh 365=item $fh = $handle->fh
186 366
187This method returns the filehandle of the L<AnyEvent::Handle> object. 367This method returns the file handle used to create the L<AnyEvent::Handle> object.
188 368
189=cut 369=cut
190 370
191sub fh { $_[0]->{fh} } 371sub fh { $_[0]{fh} }
192 372
193=item $handle->on_error ($cb) 373=item $handle->on_error ($cb)
194 374
195Replace the current C<on_error> callback (see the C<on_error> constructor argument). 375Replace the current C<on_error> callback (see the C<on_error> constructor argument).
196 376
208 388
209sub on_eof { 389sub on_eof {
210 $_[0]{on_eof} = $_[1]; 390 $_[0]{on_eof} = $_[1];
211} 391}
212 392
393=item $handle->on_timeout ($cb)
394
395Replace the current C<on_timeout> callback, or disables the callback (but
396not the timeout) if C<$cb> = C<undef>. See the C<timeout> constructor
397argument and method.
398
399=cut
400
401sub on_timeout {
402 $_[0]{on_timeout} = $_[1];
403}
404
405=item $handle->autocork ($boolean)
406
407Enables or disables the current autocork behaviour (see C<autocork>
408constructor argument). Changes will only take effect on the next write.
409
410=cut
411
412sub autocork {
413 $_[0]{autocork} = $_[1];
414}
415
416=item $handle->no_delay ($boolean)
417
418Enables or disables the C<no_delay> setting (see constructor argument of
419the same name for details).
420
421=cut
422
423sub no_delay {
424 $_[0]{no_delay} = $_[1];
425
426 eval {
427 local $SIG{__DIE__};
428 setsockopt $_[0]{fh}, &Socket::IPPROTO_TCP, &Socket::TCP_NODELAY, int $_[1];
429 };
430}
431
432#############################################################################
433
434=item $handle->timeout ($seconds)
435
436Configures (or disables) the inactivity timeout.
437
438=cut
439
440sub timeout {
441 my ($self, $timeout) = @_;
442
443 $self->{timeout} = $timeout;
444 $self->_timeout;
445}
446
447# reset the timeout watcher, as neccessary
448# also check for time-outs
449sub _timeout {
450 my ($self) = @_;
451
452 if ($self->{timeout}) {
453 my $NOW = AnyEvent->now;
454
455 # when would the timeout trigger?
456 my $after = $self->{_activity} + $self->{timeout} - $NOW;
457
458 # now or in the past already?
459 if ($after <= 0) {
460 $self->{_activity} = $NOW;
461
462 if ($self->{on_timeout}) {
463 $self->{on_timeout}($self);
464 } else {
465 $self->_error (&Errno::ETIMEDOUT);
466 }
467
468 # callback could have changed timeout value, optimise
469 return unless $self->{timeout};
470
471 # calculate new after
472 $after = $self->{timeout};
473 }
474
475 Scalar::Util::weaken $self;
476 return unless $self; # ->error could have destroyed $self
477
478 $self->{_tw} ||= AnyEvent->timer (after => $after, cb => sub {
479 delete $self->{_tw};
480 $self->_timeout;
481 });
482 } else {
483 delete $self->{_tw};
484 }
485}
486
213############################################################################# 487#############################################################################
214 488
215=back 489=back
216 490
217=head2 WRITE QUEUE 491=head2 WRITE QUEUE
220for reading. 494for reading.
221 495
222The write queue is very simple: you can add data to its end, and 496The write queue is very simple: you can add data to its end, and
223AnyEvent::Handle will automatically try to get rid of it for you. 497AnyEvent::Handle will automatically try to get rid of it for you.
224 498
225When data could be writtena nd the write buffer is shorter then the low 499When data could be written and the write buffer is shorter then the low
226water mark, the C<on_drain> callback will be invoked. 500water mark, the C<on_drain> callback will be invoked.
227 501
228=over 4 502=over 4
229 503
230=item $handle->on_drain ($cb) 504=item $handle->on_drain ($cb)
238 my ($self, $cb) = @_; 512 my ($self, $cb) = @_;
239 513
240 $self->{on_drain} = $cb; 514 $self->{on_drain} = $cb;
241 515
242 $cb->($self) 516 $cb->($self)
243 if $cb && $self->{low_water_mark} >= length $self->{wbuf}; 517 if $cb && $self->{low_water_mark} >= (length $self->{wbuf}) + (length $self->{_tls_wbuf});
244} 518}
245 519
246=item $handle->push_write ($data) 520=item $handle->push_write ($data)
247 521
248Queues the given scalar to be written. You can push as much data as you 522Queues the given scalar to be written. You can push as much data as you
249want (only limited by the available memory), as C<AnyEvent::Handle> 523want (only limited by the available memory), as C<AnyEvent::Handle>
250buffers it independently of the kernel. 524buffers it independently of the kernel.
251 525
252=cut 526=cut
253 527
254sub push_write { 528sub _drain_wbuf {
255 my ($self, $data) = @_; 529 my ($self) = @_;
256 530
257 $self->{wbuf} .= $data; 531 if (!$self->{_ww} && length $self->{wbuf}) {
258 532
259 unless ($self->{ww}) {
260 Scalar::Util::weaken $self; 533 Scalar::Util::weaken $self;
534
261 my $cb = sub { 535 my $cb = sub {
262 my $len = syswrite $self->{fh}, $self->{wbuf}; 536 my $len = syswrite $self->{fh}, $self->{wbuf};
263 537
264 if ($len > 0) { 538 if ($len >= 0) {
265 substr $self->{wbuf}, 0, $len, ""; 539 substr $self->{wbuf}, 0, $len, "";
266 540
541 $self->{_activity} = AnyEvent->now;
267 542
268 $self->{on_drain}($self) 543 $self->{on_drain}($self)
269 if $self->{low_water_mark} >= length $self->{wbuf} 544 if $self->{low_water_mark} >= (length $self->{wbuf}) + (length $self->{_tls_wbuf})
270 && $self->{on_drain}; 545 && $self->{on_drain};
271 546
272 delete $self->{ww} unless length $self->{wbuf}; 547 delete $self->{_ww} unless length $self->{wbuf};
273 } elsif ($! != EAGAIN && $! != EINTR) { 548 } elsif ($! != EAGAIN && $! != EINTR && $! != WSAEWOULDBLOCK) {
274 $self->error; 549 $self->_error ($!, 1);
275 } 550 }
276 }; 551 };
277 552
553 # try to write data immediately
554 $cb->() unless $self->{autocork};
555
556 # if still data left in wbuf, we need to poll
278 $self->{ww} = AnyEvent->io (fh => $self->{fh}, poll => "w", cb => $cb); 557 $self->{_ww} = AnyEvent->io (fh => $self->{fh}, poll => "w", cb => $cb)
279 558 if length $self->{wbuf};
280 $cb->($self);
281 }; 559 };
282} 560}
561
562our %WH;
563
564sub register_write_type($$) {
565 $WH{$_[0]} = $_[1];
566}
567
568sub push_write {
569 my $self = shift;
570
571 if (@_ > 1) {
572 my $type = shift;
573
574 @_ = ($WH{$type} or Carp::croak "unsupported type passed to AnyEvent::Handle::push_write")
575 ->($self, @_);
576 }
577
578 if ($self->{tls}) {
579 $self->{_tls_wbuf} .= $_[0];
580
581 &_dotls ($self);
582 } else {
583 $self->{wbuf} .= $_[0];
584 $self->_drain_wbuf;
585 }
586}
587
588=item $handle->push_write (type => @args)
589
590Instead of formatting your data yourself, you can also let this module do
591the job by specifying a type and type-specific arguments.
592
593Predefined types are (if you have ideas for additional types, feel free to
594drop by and tell us):
595
596=over 4
597
598=item netstring => $string
599
600Formats the given value as netstring
601(http://cr.yp.to/proto/netstrings.txt, this is not a recommendation to use them).
602
603=cut
604
605register_write_type netstring => sub {
606 my ($self, $string) = @_;
607
608 (length $string) . ":$string,"
609};
610
611=item packstring => $format, $data
612
613An octet string prefixed with an encoded length. The encoding C<$format>
614uses the same format as a Perl C<pack> format, but must specify a single
615integer only (only one of C<cCsSlLqQiInNvVjJw> is allowed, plus an
616optional C<!>, C<< < >> or C<< > >> modifier).
617
618=cut
619
620register_write_type packstring => sub {
621 my ($self, $format, $string) = @_;
622
623 pack "$format/a*", $string
624};
625
626=item json => $array_or_hashref
627
628Encodes the given hash or array reference into a JSON object. Unless you
629provide your own JSON object, this means it will be encoded to JSON text
630in UTF-8.
631
632JSON objects (and arrays) are self-delimiting, so you can write JSON at
633one end of a handle and read them at the other end without using any
634additional framing.
635
636The generated JSON text is guaranteed not to contain any newlines: While
637this module doesn't need delimiters after or between JSON texts to be
638able to read them, many other languages depend on that.
639
640A simple RPC protocol that interoperates easily with others is to send
641JSON arrays (or objects, although arrays are usually the better choice as
642they mimic how function argument passing works) and a newline after each
643JSON text:
644
645 $handle->push_write (json => ["method", "arg1", "arg2"]); # whatever
646 $handle->push_write ("\012");
647
648An AnyEvent::Handle receiver would simply use the C<json> read type and
649rely on the fact that the newline will be skipped as leading whitespace:
650
651 $handle->push_read (json => sub { my $array = $_[1]; ... });
652
653Other languages could read single lines terminated by a newline and pass
654this line into their JSON decoder of choice.
655
656=cut
657
658register_write_type json => sub {
659 my ($self, $ref) = @_;
660
661 require JSON;
662
663 $self->{json} ? $self->{json}->encode ($ref)
664 : JSON::encode_json ($ref)
665};
666
667=item storable => $reference
668
669Freezes the given reference using L<Storable> and writes it to the
670handle. Uses the C<nfreeze> format.
671
672=cut
673
674register_write_type storable => sub {
675 my ($self, $ref) = @_;
676
677 require Storable;
678
679 pack "w/a*", Storable::nfreeze ($ref)
680};
681
682=back
683
684=item $handle->push_shutdown
685
686Sometimes you know you want to close the socket after writing your data
687before it was actually written. One way to do that is to replace your
688C<on_drain> handler by a callback that shuts down the socket. This method
689is a shorthand for just that, and replaces the C<on_drain> callback with:
690
691 sub { shutdown $_[0]{fh}, 1 } # for push_shutdown
692
693This simply shuts down the write side and signals an EOF condition to the
694the peer.
695
696You can rely on the normal read queue and C<on_eof> handling
697afterwards. This is the cleanest way to close a connection.
698
699=cut
700
701sub push_shutdown {
702 $_[0]->{on_drain} = sub { shutdown $_[0]{fh}, 1 };
703}
704
705=item AnyEvent::Handle::register_write_type type => $coderef->($handle, @args)
706
707This function (not method) lets you add your own types to C<push_write>.
708Whenever the given C<type> is used, C<push_write> will invoke the code
709reference with the handle object and the remaining arguments.
710
711The code reference is supposed to return a single octet string that will
712be appended to the write buffer.
713
714Note that this is a function, and all types registered this way will be
715global, so try to use unique names.
716
717=cut
283 718
284############################################################################# 719#############################################################################
285 720
286=back 721=back
287 722
294ways, the "simple" way, using only C<on_read> and the "complex" way, using 729ways, the "simple" way, using only C<on_read> and the "complex" way, using
295a queue. 730a queue.
296 731
297In the simple case, you just install an C<on_read> callback and whenever 732In the simple case, you just install an C<on_read> callback and whenever
298new data arrives, it will be called. You can then remove some data (if 733new data arrives, it will be called. You can then remove some data (if
299enough is there) from the read buffer (C<< $handle->rbuf >>) if you want 734enough is there) from the read buffer (C<< $handle->rbuf >>). Or you cna
300or not. 735leave the data there if you want to accumulate more (e.g. when only a
736partial message has been received so far).
301 737
302In the more complex case, you want to queue multiple callbacks. In this 738In the more complex case, you want to queue multiple callbacks. In this
303case, AnyEvent::Handle will call the first queued callback each time new 739case, AnyEvent::Handle will call the first queued callback each time new
304data arrives and removes it when it has done its job (see C<push_read>, 740data arrives (also the first time it is queued) and removes it when it has
305below). 741done its job (see C<push_read>, below).
306 742
307This way you can, for example, push three line-reads, followed by reading 743This way you can, for example, push three line-reads, followed by reading
308a chunk of data, and AnyEvent::Handle will execute them in order. 744a chunk of data, and AnyEvent::Handle will execute them in order.
309 745
310Example 1: EPP protocol parser. EPP sends 4 byte length info, followed by 746Example 1: EPP protocol parser. EPP sends 4 byte length info, followed by
311the specified number of bytes which give an XML datagram. 747the specified number of bytes which give an XML datagram.
312 748
313 # in the default state, expect some header bytes 749 # in the default state, expect some header bytes
314 $handle->on_read (sub { 750 $handle->on_read (sub {
315 # some data is here, now queue the length-header-read (4 octets) 751 # some data is here, now queue the length-header-read (4 octets)
316 shift->unshift_read_chunk (4, sub { 752 shift->unshift_read (chunk => 4, sub {
317 # header arrived, decode 753 # header arrived, decode
318 my $len = unpack "N", $_[1]; 754 my $len = unpack "N", $_[1];
319 755
320 # now read the payload 756 # now read the payload
321 shift->unshift_read_chunk ($len, sub { 757 shift->unshift_read (chunk => $len, sub {
322 my $xml = $_[1]; 758 my $xml = $_[1];
323 # handle xml 759 # handle xml
324 }); 760 });
325 }); 761 });
326 }); 762 });
327 763
328Example 2: Implement a client for a protocol that replies either with 764Example 2: Implement a client for a protocol that replies either with "OK"
329"OK" and another line or "ERROR" for one request, and 64 bytes for the 765and another line or "ERROR" for the first request that is sent, and 64
330second request. Due tot he availability of a full queue, we can just 766bytes for the second request. Due to the availability of a queue, we can
331pipeline sending both requests and manipulate the queue as necessary in 767just pipeline sending both requests and manipulate the queue as necessary
332the callbacks: 768in the callbacks.
333 769
334 # request one 770When the first callback is called and sees an "OK" response, it will
771C<unshift> another line-read. This line-read will be queued I<before> the
77264-byte chunk callback.
773
774 # request one, returns either "OK + extra line" or "ERROR"
335 $handle->push_write ("request 1\015\012"); 775 $handle->push_write ("request 1\015\012");
336 776
337 # we expect "ERROR" or "OK" as response, so push a line read 777 # we expect "ERROR" or "OK" as response, so push a line read
338 $handle->push_read_line (sub { 778 $handle->push_read (line => sub {
339 # if we got an "OK", we have to _prepend_ another line, 779 # if we got an "OK", we have to _prepend_ another line,
340 # so it will be read before the second request reads its 64 bytes 780 # so it will be read before the second request reads its 64 bytes
341 # which are already in the queue when this callback is called 781 # which are already in the queue when this callback is called
342 # we don't do this in case we got an error 782 # we don't do this in case we got an error
343 if ($_[1] eq "OK") { 783 if ($_[1] eq "OK") {
344 $_[0]->unshift_read_line (sub { 784 $_[0]->unshift_read (line => sub {
345 my $response = $_[1]; 785 my $response = $_[1];
346 ... 786 ...
347 }); 787 });
348 } 788 }
349 }); 789 });
350 790
351 # request two 791 # request two, simply returns 64 octets
352 $handle->push_write ("request 2\015\012"); 792 $handle->push_write ("request 2\015\012");
353 793
354 # simply read 64 bytes, always 794 # simply read 64 bytes, always
355 $handle->push_read_chunk (64, sub { 795 $handle->push_read (chunk => 64, sub {
356 my $response = $_[1]; 796 my $response = $_[1];
357 ... 797 ...
358 }); 798 });
359 799
360=over 4 800=over 4
362=cut 802=cut
363 803
364sub _drain_rbuf { 804sub _drain_rbuf {
365 my ($self) = @_; 805 my ($self) = @_;
366 806
367 return if $self->{in_drain};
368 local $self->{in_drain} = 1; 807 local $self->{_in_drain} = 1;
369 808
809 if (
810 defined $self->{rbuf_max}
811 && $self->{rbuf_max} < length $self->{rbuf}
812 ) {
813 $self->_error (&Errno::ENOSPC, 1), return;
814 }
815
816 while () {
817 # we need to use a separate tls read buffer, as we must not receive data while
818 # we are draining the buffer, and this can only happen with TLS.
819 $self->{rbuf} .= delete $self->{_tls_rbuf} if exists $self->{_tls_rbuf};
820
370 while (my $len = length $self->{rbuf}) { 821 my $len = length $self->{rbuf};
371 no strict 'refs'; 822
372 if (my $cb = shift @{ $self->{queue} }) { 823 if (my $cb = shift @{ $self->{_queue} }) {
373 if (!$cb->($self)) { 824 unless ($cb->($self)) {
374 if ($self->{eof}) { 825 if ($self->{_eof}) {
375 # no progress can be made (not enough data and no data forthcoming) 826 # no progress can be made (not enough data and no data forthcoming)
376 $! = &Errno::EPIPE; return $self->error; 827 $self->_error (&Errno::EPIPE, 1), return;
377 } 828 }
378 829
379 unshift @{ $self->{queue} }, $cb; 830 unshift @{ $self->{_queue} }, $cb;
380 return; 831 last;
381 } 832 }
382 } elsif ($self->{on_read}) { 833 } elsif ($self->{on_read}) {
834 last unless $len;
835
383 $self->{on_read}($self); 836 $self->{on_read}($self);
384 837
385 if ( 838 if (
386 $self->{eof} # if no further data will arrive
387 && $len == length $self->{rbuf} # and no data has been consumed 839 $len == length $self->{rbuf} # if no data has been consumed
388 && !@{ $self->{queue} } # and the queue is still empty 840 && !@{ $self->{_queue} } # and the queue is still empty
389 && $self->{on_read} # and we still want to read data 841 && $self->{on_read} # but we still have on_read
390 ) { 842 ) {
843 # no further data will arrive
391 # then no progress can be made 844 # so no progress can be made
392 $! = &Errno::EPIPE; return $self->error; 845 $self->_error (&Errno::EPIPE, 1), return
846 if $self->{_eof};
847
848 last; # more data might arrive
393 } 849 }
394 } else { 850 } else {
395 # read side becomes idle 851 # read side becomes idle
396 delete $self->{rw}; 852 delete $self->{_rw} unless $self->{tls};
397 return; 853 last;
398 } 854 }
399 } 855 }
400 856
401 if ($self->{eof}) { 857 if ($self->{_eof}) {
402 $self->_shutdown; 858 if ($self->{on_eof}) {
403 $self->{on_eof}($self); 859 $self->{on_eof}($self)
860 } else {
861 $self->_error (0, 1);
862 }
863 }
864
865 # may need to restart read watcher
866 unless ($self->{_rw}) {
867 $self->start_read
868 if $self->{on_read} || @{ $self->{_queue} };
404 } 869 }
405} 870}
406 871
407=item $handle->on_read ($cb) 872=item $handle->on_read ($cb)
408 873
414 879
415sub on_read { 880sub on_read {
416 my ($self, $cb) = @_; 881 my ($self, $cb) = @_;
417 882
418 $self->{on_read} = $cb; 883 $self->{on_read} = $cb;
884 $self->_drain_rbuf if $cb && !$self->{_in_drain};
419} 885}
420 886
421=item $handle->rbuf 887=item $handle->rbuf
422 888
423Returns the read buffer (as a modifiable lvalue). 889Returns the read buffer (as a modifiable lvalue).
424 890
425You can access the read buffer directly as the C<< ->{rbuf} >> member, if 891You can access the read buffer directly as the C<< ->{rbuf} >>
426you want. 892member, if you want. However, the only operation allowed on the
893read buffer (apart from looking at it) is removing data from its
894beginning. Otherwise modifying or appending to it is not allowed and will
895lead to hard-to-track-down bugs.
427 896
428NOTE: The read buffer should only be used or modified if the C<on_read>, 897NOTE: The read buffer should only be used or modified if the C<on_read>,
429C<push_read> or C<unshift_read> methods are used. The other read methods 898C<push_read> or C<unshift_read> methods are used. The other read methods
430automatically manage the read buffer. 899automatically manage the read buffer.
431 900
442Append the given callback to the end of the queue (C<push_read>) or 911Append the given callback to the end of the queue (C<push_read>) or
443prepend it (C<unshift_read>). 912prepend it (C<unshift_read>).
444 913
445The callback is called each time some additional read data arrives. 914The callback is called each time some additional read data arrives.
446 915
447It must check wether enough data is in the read buffer already. 916It must check whether enough data is in the read buffer already.
448 917
449If not enough data is available, it must return the empty list or a false 918If not enough data is available, it must return the empty list or a false
450value, in which case it will be called repeatedly until enough data is 919value, in which case it will be called repeatedly until enough data is
451available (or an error condition is detected). 920available (or an error condition is detected).
452 921
454interested in (which can be none at all) and return a true value. After returning 923interested in (which can be none at all) and return a true value. After returning
455true, it will be removed from the queue. 924true, it will be removed from the queue.
456 925
457=cut 926=cut
458 927
928our %RH;
929
930sub register_read_type($$) {
931 $RH{$_[0]} = $_[1];
932}
933
459sub push_read { 934sub push_read {
460 my ($self, $cb) = @_; 935 my $self = shift;
936 my $cb = pop;
461 937
938 if (@_) {
939 my $type = shift;
940
941 $cb = ($RH{$type} or Carp::croak "unsupported type passed to AnyEvent::Handle::push_read")
942 ->($self, $cb, @_);
943 }
944
462 push @{ $self->{queue} }, $cb; 945 push @{ $self->{_queue} }, $cb;
463 $self->_drain_rbuf; 946 $self->_drain_rbuf unless $self->{_in_drain};
464} 947}
465 948
466sub unshift_read { 949sub unshift_read {
467 my ($self, $cb) = @_; 950 my $self = shift;
951 my $cb = pop;
468 952
953 if (@_) {
954 my $type = shift;
955
956 $cb = ($RH{$type} or Carp::croak "unsupported type passed to AnyEvent::Handle::unshift_read")
957 ->($self, $cb, @_);
958 }
959
960
469 push @{ $self->{queue} }, $cb; 961 unshift @{ $self->{_queue} }, $cb;
470 $self->_drain_rbuf; 962 $self->_drain_rbuf unless $self->{_in_drain};
471} 963}
472 964
473=item $handle->push_read_chunk ($len, $cb->($self, $data)) 965=item $handle->push_read (type => @args, $cb)
474 966
475=item $handle->unshift_read_chunk ($len, $cb->($self, $data)) 967=item $handle->unshift_read (type => @args, $cb)
476 968
477Append the given callback to the end of the queue (C<push_read_chunk>) or 969Instead of providing a callback that parses the data itself you can chose
478prepend it (C<unshift_read_chunk>). 970between a number of predefined parsing formats, for chunks of data, lines
971etc.
479 972
480The callback will be called only once C<$len> bytes have been read, and 973Predefined types are (if you have ideas for additional types, feel free to
481these C<$len> bytes will be passed to the callback. 974drop by and tell us):
482 975
483=cut 976=over 4
484 977
485sub _read_chunk($$) { 978=item chunk => $octets, $cb->($handle, $data)
979
980Invoke the callback only once C<$octets> bytes have been read. Pass the
981data read to the callback. The callback will never be called with less
982data.
983
984Example: read 2 bytes.
985
986 $handle->push_read (chunk => 2, sub {
987 warn "yay ", unpack "H*", $_[1];
988 });
989
990=cut
991
992register_read_type chunk => sub {
486 my ($self, $len, $cb) = @_; 993 my ($self, $cb, $len) = @_;
487 994
488 sub { 995 sub {
489 $len <= length $_[0]{rbuf} or return; 996 $len <= length $_[0]{rbuf} or return;
490 $cb->($self, $_[0], substr $_[0]{rbuf}, 0, $len, ""); 997 $cb->($_[0], substr $_[0]{rbuf}, 0, $len, "");
491 1 998 1
492 } 999 }
493} 1000};
494 1001
495sub push_read_chunk { 1002=item line => [$eol, ]$cb->($handle, $line, $eol)
496 $_[0]->push_read (&_read_chunk);
497}
498
499
500sub unshift_read_chunk {
501 $_[0]->unshift_read (&_read_chunk);
502}
503
504=item $handle->push_read_line ([$eol, ]$cb->($self, $line, $eol))
505
506=item $handle->unshift_read_line ([$eol, ]$cb->($self, $line, $eol))
507
508Append the given callback to the end of the queue (C<push_read_line>) or
509prepend it (C<unshift_read_line>).
510 1003
511The callback will be called only once a full line (including the end of 1004The callback will be called only once a full line (including the end of
512line marker, C<$eol>) has been read. This line (excluding the end of line 1005line marker, C<$eol>) has been read. This line (excluding the end of line
513marker) will be passed to the callback as second argument (C<$line>), and 1006marker) will be passed to the callback as second argument (C<$line>), and
514the end of line marker as the third argument (C<$eol>). 1007the end of line marker as the third argument (C<$eol>).
525Partial lines at the end of the stream will never be returned, as they are 1018Partial lines at the end of the stream will never be returned, as they are
526not marked by the end of line marker. 1019not marked by the end of line marker.
527 1020
528=cut 1021=cut
529 1022
530sub _read_line($$) { 1023register_read_type line => sub {
531 my $self = shift; 1024 my ($self, $cb, $eol) = @_;
532 my $cb = pop;
533 my $eol = @_ ? shift : qr|(\015?\012)|;
534 my $pos;
535 1025
1026 if (@_ < 3) {
1027 # this is more than twice as fast as the generic code below
1028 sub {
1029 $_[0]{rbuf} =~ s/^([^\015\012]*)(\015?\012)// or return;
1030
1031 $cb->($_[0], $1, $2);
1032 1
1033 }
1034 } else {
536 $eol = qr|(\Q$eol\E)| unless ref $eol; 1035 $eol = quotemeta $eol unless ref $eol;
537 $eol = qr|^(.*?)($eol)|; 1036 $eol = qr|^(.*?)($eol)|s;
1037
1038 sub {
1039 $_[0]{rbuf} =~ s/$eol// or return;
1040
1041 $cb->($_[0], $1, $2);
1042 1
1043 }
1044 }
1045};
1046
1047=item regex => $accept[, $reject[, $skip], $cb->($handle, $data)
1048
1049Makes a regex match against the regex object C<$accept> and returns
1050everything up to and including the match.
1051
1052Example: read a single line terminated by '\n'.
1053
1054 $handle->push_read (regex => qr<\n>, sub { ... });
1055
1056If C<$reject> is given and not undef, then it determines when the data is
1057to be rejected: it is matched against the data when the C<$accept> regex
1058does not match and generates an C<EBADMSG> error when it matches. This is
1059useful to quickly reject wrong data (to avoid waiting for a timeout or a
1060receive buffer overflow).
1061
1062Example: expect a single decimal number followed by whitespace, reject
1063anything else (not the use of an anchor).
1064
1065 $handle->push_read (regex => qr<^[0-9]+\s>, qr<[^0-9]>, sub { ... });
1066
1067If C<$skip> is given and not C<undef>, then it will be matched against
1068the receive buffer when neither C<$accept> nor C<$reject> match,
1069and everything preceding and including the match will be accepted
1070unconditionally. This is useful to skip large amounts of data that you
1071know cannot be matched, so that the C<$accept> or C<$reject> regex do not
1072have to start matching from the beginning. This is purely an optimisation
1073and is usually worth only when you expect more than a few kilobytes.
1074
1075Example: expect a http header, which ends at C<\015\012\015\012>. Since we
1076expect the header to be very large (it isn't in practise, but...), we use
1077a skip regex to skip initial portions. The skip regex is tricky in that
1078it only accepts something not ending in either \015 or \012, as these are
1079required for the accept regex.
1080
1081 $handle->push_read (regex =>
1082 qr<\015\012\015\012>,
1083 undef, # no reject
1084 qr<^.*[^\015\012]>,
1085 sub { ... });
1086
1087=cut
1088
1089register_read_type regex => sub {
1090 my ($self, $cb, $accept, $reject, $skip) = @_;
1091
1092 my $data;
1093 my $rbuf = \$self->{rbuf};
538 1094
539 sub { 1095 sub {
540 $_[0]{rbuf} =~ s/$eol// or return; 1096 # accept
541 1097 if ($$rbuf =~ $accept) {
1098 $data .= substr $$rbuf, 0, $+[0], "";
542 $cb->($self, $1, $2); 1099 $cb->($self, $data);
1100 return 1;
1101 }
1102
1103 # reject
1104 if ($reject && $$rbuf =~ $reject) {
1105 $self->_error (&Errno::EBADMSG);
1106 }
1107
1108 # skip
1109 if ($skip && $$rbuf =~ $skip) {
1110 $data .= substr $$rbuf, 0, $+[0], "";
1111 }
1112
1113 ()
1114 }
1115};
1116
1117=item netstring => $cb->($handle, $string)
1118
1119A netstring (http://cr.yp.to/proto/netstrings.txt, this is not an endorsement).
1120
1121Throws an error with C<$!> set to EBADMSG on format violations.
1122
1123=cut
1124
1125register_read_type netstring => sub {
1126 my ($self, $cb) = @_;
1127
1128 sub {
1129 unless ($_[0]{rbuf} =~ s/^(0|[1-9][0-9]*)://) {
1130 if ($_[0]{rbuf} =~ /[^0-9]/) {
1131 $self->_error (&Errno::EBADMSG);
1132 }
1133 return;
1134 }
1135
1136 my $len = $1;
1137
1138 $self->unshift_read (chunk => $len, sub {
1139 my $string = $_[1];
1140 $_[0]->unshift_read (chunk => 1, sub {
1141 if ($_[1] eq ",") {
1142 $cb->($_[0], $string);
1143 } else {
1144 $self->_error (&Errno::EBADMSG);
1145 }
1146 });
1147 });
1148
543 1 1149 1
544 } 1150 }
545} 1151};
546 1152
547sub push_read_line { 1153=item packstring => $format, $cb->($handle, $string)
548 $_[0]->push_read (&_read_line);
549}
550 1154
551sub unshift_read_line { 1155An octet string prefixed with an encoded length. The encoding C<$format>
552 $_[0]->unshift_read (&_read_line); 1156uses the same format as a Perl C<pack> format, but must specify a single
553} 1157integer only (only one of C<cCsSlLqQiInNvVjJw> is allowed, plus an
1158optional C<!>, C<< < >> or C<< > >> modifier).
1159
1160For example, DNS over TCP uses a prefix of C<n> (2 octet network order),
1161EPP uses a prefix of C<N> (4 octtes).
1162
1163Example: read a block of data prefixed by its length in BER-encoded
1164format (very efficient).
1165
1166 $handle->push_read (packstring => "w", sub {
1167 my ($handle, $data) = @_;
1168 });
1169
1170=cut
1171
1172register_read_type packstring => sub {
1173 my ($self, $cb, $format) = @_;
1174
1175 sub {
1176 # when we can use 5.10 we can use ".", but for 5.8 we use the re-pack method
1177 defined (my $len = eval { unpack $format, $_[0]{rbuf} })
1178 or return;
1179
1180 $format = length pack $format, $len;
1181
1182 # bypass unshift if we already have the remaining chunk
1183 if ($format + $len <= length $_[0]{rbuf}) {
1184 my $data = substr $_[0]{rbuf}, $format, $len;
1185 substr $_[0]{rbuf}, 0, $format + $len, "";
1186 $cb->($_[0], $data);
1187 } else {
1188 # remove prefix
1189 substr $_[0]{rbuf}, 0, $format, "";
1190
1191 # read remaining chunk
1192 $_[0]->unshift_read (chunk => $len, $cb);
1193 }
1194
1195 1
1196 }
1197};
1198
1199=item json => $cb->($handle, $hash_or_arrayref)
1200
1201Reads a JSON object or array, decodes it and passes it to the
1202callback. When a parse error occurs, an C<EBADMSG> error will be raised.
1203
1204If a C<json> object was passed to the constructor, then that will be used
1205for the final decode, otherwise it will create a JSON coder expecting UTF-8.
1206
1207This read type uses the incremental parser available with JSON version
12082.09 (and JSON::XS version 2.2) and above. You have to provide a
1209dependency on your own: this module will load the JSON module, but
1210AnyEvent does not depend on it itself.
1211
1212Since JSON texts are fully self-delimiting, the C<json> read and write
1213types are an ideal simple RPC protocol: just exchange JSON datagrams. See
1214the C<json> write type description, above, for an actual example.
1215
1216=cut
1217
1218register_read_type json => sub {
1219 my ($self, $cb) = @_;
1220
1221 require JSON;
1222
1223 my $data;
1224 my $rbuf = \$self->{rbuf};
1225
1226 my $json = $self->{json} ||= JSON->new->utf8;
1227
1228 sub {
1229 my $ref = eval { $json->incr_parse ($self->{rbuf}) };
1230
1231 if ($ref) {
1232 $self->{rbuf} = $json->incr_text;
1233 $json->incr_text = "";
1234 $cb->($self, $ref);
1235
1236 1
1237 } elsif ($@) {
1238 # error case
1239 $json->incr_skip;
1240
1241 $self->{rbuf} = $json->incr_text;
1242 $json->incr_text = "";
1243
1244 $self->_error (&Errno::EBADMSG);
1245
1246 ()
1247 } else {
1248 $self->{rbuf} = "";
1249
1250 ()
1251 }
1252 }
1253};
1254
1255=item storable => $cb->($handle, $ref)
1256
1257Deserialises a L<Storable> frozen representation as written by the
1258C<storable> write type (BER-encoded length prefix followed by nfreeze'd
1259data).
1260
1261Raises C<EBADMSG> error if the data could not be decoded.
1262
1263=cut
1264
1265register_read_type storable => sub {
1266 my ($self, $cb) = @_;
1267
1268 require Storable;
1269
1270 sub {
1271 # when we can use 5.10 we can use ".", but for 5.8 we use the re-pack method
1272 defined (my $len = eval { unpack "w", $_[0]{rbuf} })
1273 or return;
1274
1275 my $format = length pack "w", $len;
1276
1277 # bypass unshift if we already have the remaining chunk
1278 if ($format + $len <= length $_[0]{rbuf}) {
1279 my $data = substr $_[0]{rbuf}, $format, $len;
1280 substr $_[0]{rbuf}, 0, $format + $len, "";
1281 $cb->($_[0], Storable::thaw ($data));
1282 } else {
1283 # remove prefix
1284 substr $_[0]{rbuf}, 0, $format, "";
1285
1286 # read remaining chunk
1287 $_[0]->unshift_read (chunk => $len, sub {
1288 if (my $ref = eval { Storable::thaw ($_[1]) }) {
1289 $cb->($_[0], $ref);
1290 } else {
1291 $self->_error (&Errno::EBADMSG);
1292 }
1293 });
1294 }
1295
1296 1
1297 }
1298};
1299
1300=back
1301
1302=item AnyEvent::Handle::register_read_type type => $coderef->($handle, $cb, @args)
1303
1304This function (not method) lets you add your own types to C<push_read>.
1305
1306Whenever the given C<type> is used, C<push_read> will invoke the code
1307reference with the handle object, the callback and the remaining
1308arguments.
1309
1310The code reference is supposed to return a callback (usually a closure)
1311that works as a plain read callback (see C<< ->push_read ($cb) >>).
1312
1313It should invoke the passed callback when it is done reading (remember to
1314pass C<$handle> as first argument as all other callbacks do that).
1315
1316Note that this is a function, and all types registered this way will be
1317global, so try to use unique names.
1318
1319For examples, see the source of this module (F<perldoc -m AnyEvent::Handle>,
1320search for C<register_read_type>)).
554 1321
555=item $handle->stop_read 1322=item $handle->stop_read
556 1323
557=item $handle->start_read 1324=item $handle->start_read
558 1325
559In rare cases you actually do not want to read anything form the 1326In rare cases you actually do not want to read anything from the
560socket. In this case you can call C<stop_read>. Neither C<on_read> no 1327socket. In this case you can call C<stop_read>. Neither C<on_read> nor
561any queued callbacks will be executed then. To start readign again, call 1328any queued callbacks will be executed then. To start reading again, call
562C<start_read>. 1329C<start_read>.
1330
1331Note that AnyEvent::Handle will automatically C<start_read> for you when
1332you change the C<on_read> callback or push/unshift a read callback, and it
1333will automatically C<stop_read> for you when neither C<on_read> is set nor
1334there are any read requests in the queue.
1335
1336These methods will have no effect when in TLS mode (as TLS doesn't support
1337half-duplex connections).
563 1338
564=cut 1339=cut
565 1340
566sub stop_read { 1341sub stop_read {
567 my ($self) = @_; 1342 my ($self) = @_;
568 1343
569 delete $self->{rw}; 1344 delete $self->{_rw} unless $self->{tls};
570} 1345}
571 1346
572sub start_read { 1347sub start_read {
573 my ($self) = @_; 1348 my ($self) = @_;
574 1349
575 unless ($self->{rw} || $self->{eof}) { 1350 unless ($self->{_rw} || $self->{_eof}) {
576 Scalar::Util::weaken $self; 1351 Scalar::Util::weaken $self;
577 1352
578 $self->{rw} = AnyEvent->io (fh => $self->{fh}, poll => "r", cb => sub { 1353 $self->{_rw} = AnyEvent->io (fh => $self->{fh}, poll => "r", cb => sub {
1354 my $rbuf = \($self->{tls} ? my $buf : $self->{rbuf});
579 my $len = sysread $self->{fh}, $self->{rbuf}, $self->{read_size} || 8192, length $self->{rbuf}; 1355 my $len = sysread $self->{fh}, $$rbuf, $self->{read_size} || 8192, length $$rbuf;
580 1356
581 if ($len > 0) { 1357 if ($len > 0) {
582 if (defined $self->{rbuf_max}) { 1358 $self->{_activity} = AnyEvent->now;
583 if ($self->{rbuf_max} < length $self->{rbuf}) { 1359
584 $! = &Errno::ENOSPC; return $self->error; 1360 if ($self->{tls}) {
585 } 1361 Net::SSLeay::BIO_write ($self->{_rbio}, $$rbuf);
1362
1363 &_dotls ($self);
1364 } else {
1365 $self->_drain_rbuf unless $self->{_in_drain};
586 } 1366 }
587 1367
588 } elsif (defined $len) { 1368 } elsif (defined $len) {
589 $self->{eof} = 1;
590 delete $self->{rw}; 1369 delete $self->{_rw};
1370 $self->{_eof} = 1;
1371 $self->_drain_rbuf unless $self->{_in_drain};
591 1372
592 } elsif ($! != EAGAIN && $! != EINTR) { 1373 } elsif ($! != EAGAIN && $! != EINTR && $! != WSAEWOULDBLOCK) {
593 return $self->error; 1374 return $self->_error ($!, 1);
594 } 1375 }
595
596 $self->_drain_rbuf;
597 }); 1376 });
598 } 1377 }
599} 1378}
600 1379
1380our $ERROR_SYSCALL;
1381our $ERROR_WANT_READ;
1382our $ERROR_ZERO_RETURN;
1383
1384sub _tls_error {
1385 my ($self, $err) = @_;
1386 warn "$err,$!\n";#d#
1387
1388 return $self->_error ($!, 1)
1389 if $err == Net::SSLeay::ERROR_SYSCALL ();
1390
1391 $self->_error (&Errno::EPROTO, 1,
1392 Net::SSLeay::ERR_error_string (Net::SSLeay::ERR_get_error ()));
1393}
1394
1395# poll the write BIO and send the data if applicable
1396# also decode read data if possible
1397# this is basiclaly our TLS state machine
1398# more efficient implementations are possible with openssl,
1399# but not with the buggy and incomplete Net::SSLeay.
1400sub _dotls {
1401 my ($self) = @_;
1402
1403 my $tmp;
1404
1405 if (length $self->{_tls_wbuf}) {
1406 while (($tmp = Net::SSLeay::write ($self->{tls}, $self->{_tls_wbuf})) > 0) {
1407 substr $self->{_tls_wbuf}, 0, $tmp, "";
1408 }
1409
1410 $tmp = Net::SSLeay::get_error ($self->{tls}, $tmp);
1411 return $self->_tls_error ($tmp)
1412 if $tmp != $ERROR_WANT_READ
1413 && ($tmp != $ERROR_SYSCALL || $!)
1414 && $tmp != $ERROR_ZERO_RETURN;
1415 }
1416
1417 while (defined ($tmp = Net::SSLeay::read ($self->{tls}))) {
1418 unless (length $tmp) {
1419 # let's treat SSL-eof as we treat normal EOF
1420 delete $self->{_rw};
1421 $self->{_eof} = 1;
1422 &_freetls;
1423 }
1424
1425 $self->{_tls_rbuf} .= $tmp;
1426 $self->_drain_rbuf unless $self->{_in_drain};
1427 $self->{tls} or return; # tls session might have gone away in callback
1428 }
1429
1430 $tmp = Net::SSLeay::get_error ($self->{tls}, -1);
1431 return $self->_tls_error ($tmp)
1432 if $tmp != $ERROR_WANT_READ
1433 && ($tmp != $ERROR_SYSCALL || $!)
1434 && $tmp != $ERROR_ZERO_RETURN;
1435
1436 while (length ($tmp = Net::SSLeay::BIO_read ($self->{_wbio}))) {
1437 $self->{wbuf} .= $tmp;
1438 $self->_drain_wbuf;
1439 }
1440}
1441
1442=item $handle->starttls ($tls[, $tls_ctx])
1443
1444Instead of starting TLS negotiation immediately when the AnyEvent::Handle
1445object is created, you can also do that at a later time by calling
1446C<starttls>.
1447
1448The first argument is the same as the C<tls> constructor argument (either
1449C<"connect">, C<"accept"> or an existing Net::SSLeay object).
1450
1451The second argument is the optional C<AnyEvent::TLS> object that is used
1452when AnyEvent::Handle has to create its own TLS connection object, or
1453a hash reference with C<< key => value >> pairs that will be used to
1454construct a new context.
1455
1456The TLS connection object will end up in C<< $handle->{tls} >>, the TLS
1457context in C<< $handle->{tls_ctx} >> after this call and can be used or
1458changed to your liking. Note that the handshake might have already started
1459when this function returns.
1460
1461If it an error to start a TLS handshake more than once per
1462AnyEvent::Handle object (this is due to bugs in OpenSSL).
1463
1464=cut
1465
1466sub starttls {
1467 my ($self, $ssl, $ctx) = @_;
1468
1469 require Net::SSLeay;
1470
1471 Carp::croak "it is an error to call starttls more than once on an AnyEvent::Handle object"
1472 if $self->{tls};
1473
1474 $ERROR_SYSCALL = Net::SSLeay::ERROR_SYSCALL ();
1475 $ERROR_WANT_READ = Net::SSLeay::ERROR_WANT_READ ();
1476 $ERROR_ZERO_RETURN = Net::SSLeay::ERROR_ZERO_RETURN ();
1477
1478 $ctx ||= $self->{tls_ctx};
1479
1480 if ("HASH" eq ref $ctx) {
1481 require AnyEvent::TLS;
1482
1483 local $Carp::CarpLevel = 1; # skip ourselves when creating a new context
1484 $ctx = new AnyEvent::TLS %$ctx;
1485 }
1486
1487 $self->{tls_ctx} = $ctx || TLS_CTX ();
1488 $self->{tls} = $ssl = $self->{tls_ctx}->_get_session ($ssl, $self, $self->{peername});
1489
1490 # basically, this is deep magic (because SSL_read should have the same issues)
1491 # but the openssl maintainers basically said: "trust us, it just works".
1492 # (unfortunately, we have to hardcode constants because the abysmally misdesigned
1493 # and mismaintained ssleay-module doesn't even offer them).
1494 # http://www.mail-archive.com/openssl-dev@openssl.org/msg22420.html
1495 #
1496 # in short: this is a mess.
1497 #
1498 # note that we do not try to keep the length constant between writes as we are required to do.
1499 # we assume that most (but not all) of this insanity only applies to non-blocking cases,
1500 # and we drive openssl fully in blocking mode here. Or maybe we don't - openssl seems to
1501 # have identity issues in that area.
1502# Net::SSLeay::CTX_set_mode ($ssl,
1503# (eval { local $SIG{__DIE__}; Net::SSLeay::MODE_ENABLE_PARTIAL_WRITE () } || 1)
1504# | (eval { local $SIG{__DIE__}; Net::SSLeay::MODE_ACCEPT_MOVING_WRITE_BUFFER () } || 2));
1505 Net::SSLeay::CTX_set_mode ($ssl, 1|2);
1506
1507 $self->{_rbio} = Net::SSLeay::BIO_new (Net::SSLeay::BIO_s_mem ());
1508 $self->{_wbio} = Net::SSLeay::BIO_new (Net::SSLeay::BIO_s_mem ());
1509
1510 Net::SSLeay::set_bio ($ssl, $self->{_rbio}, $self->{_wbio});
1511
1512 &_dotls; # need to trigger the initial handshake
1513 $self->start_read; # make sure we actually do read
1514}
1515
1516=item $handle->stoptls
1517
1518Shuts down the SSL connection - this makes a proper EOF handshake by
1519sending a close notify to the other side, but since OpenSSL doesn't
1520support non-blocking shut downs, it is not possible to re-use the stream
1521afterwards.
1522
1523=cut
1524
1525sub stoptls {
1526 my ($self) = @_;
1527
1528 if ($self->{tls}) {
1529 Net::SSLeay::shutdown ($self->{tls});
1530
1531 &_dotls;
1532
1533 # we don't give a shit. no, we do, but we can't. no...
1534 # we, we... have to use openssl :/
1535 &_freetls;
1536 }
1537}
1538
1539sub _freetls {
1540 my ($self) = @_;
1541
1542 return unless $self->{tls};
1543
1544 $self->{tls_ctx}->_put_session (delete $self->{tls});
1545
1546 delete @$self{qw(_rbio _wbio _tls_wbuf)};
1547}
1548
1549sub DESTROY {
1550 my ($self) = @_;
1551
1552 &_freetls;
1553
1554 my $linger = exists $self->{linger} ? $self->{linger} : 3600;
1555
1556 if ($linger && length $self->{wbuf}) {
1557 my $fh = delete $self->{fh};
1558 my $wbuf = delete $self->{wbuf};
1559
1560 my @linger;
1561
1562 push @linger, AnyEvent->io (fh => $fh, poll => "w", cb => sub {
1563 my $len = syswrite $fh, $wbuf, length $wbuf;
1564
1565 if ($len > 0) {
1566 substr $wbuf, 0, $len, "";
1567 } else {
1568 @linger = (); # end
1569 }
1570 });
1571 push @linger, AnyEvent->timer (after => $linger, cb => sub {
1572 @linger = ();
1573 });
1574 }
1575}
1576
1577=item $handle->destroy
1578
1579Shuts down the handle object as much as possible - this call ensures that
1580no further callbacks will be invoked and resources will be freed as much
1581as possible. You must not call any methods on the object afterwards.
1582
1583Normally, you can just "forget" any references to an AnyEvent::Handle
1584object and it will simply shut down. This works in fatal error and EOF
1585callbacks, as well as code outside. It does I<NOT> work in a read or write
1586callback, so when you want to destroy the AnyEvent::Handle object from
1587within such an callback. You I<MUST> call C<< ->destroy >> explicitly in
1588that case.
1589
1590The handle might still linger in the background and write out remaining
1591data, as specified by the C<linger> option, however.
1592
1593=cut
1594
1595sub destroy {
1596 my ($self) = @_;
1597
1598 $self->DESTROY;
1599 %$self = ();
1600}
1601
1602=item AnyEvent::Handle::TLS_CTX
1603
1604This function creates and returns the AnyEvent::TLS object used by default
1605for TLS mode.
1606
1607The context is created by calling L<AnyEvent::TLS> without any arguments.
1608
1609=cut
1610
1611our $TLS_CTX;
1612
1613sub TLS_CTX() {
1614 $TLS_CTX ||= do {
1615 require AnyEvent::TLS;
1616
1617 new AnyEvent::TLS
1618 }
1619}
1620
601=back 1621=back
602 1622
1623
1624=head1 NONFREQUENTLY ASKED QUESTIONS
1625
1626=over 4
1627
1628=item I C<undef> the AnyEvent::Handle reference inside my callback and
1629still get further invocations!
1630
1631That's because AnyEvent::Handle keeps a reference to itself when handling
1632read or write callbacks.
1633
1634It is only safe to "forget" the reference inside EOF or error callbacks,
1635from within all other callbacks, you need to explicitly call the C<<
1636->destroy >> method.
1637
1638=item I get different callback invocations in TLS mode/Why can't I pause
1639reading?
1640
1641Unlike, say, TCP, TLS connections do not consist of two independent
1642communication channels, one for each direction. Or put differently. The
1643read and write directions are not independent of each other: you cannot
1644write data unless you are also prepared to read, and vice versa.
1645
1646This can mean than, in TLS mode, you might get C<on_error> or C<on_eof>
1647callback invocations when you are not expecting any read data - the reason
1648is that AnyEvent::Handle always reads in TLS mode.
1649
1650During the connection, you have to make sure that you always have a
1651non-empty read-queue, or an C<on_read> watcher. At the end of the
1652connection (or when you no longer want to use it) you can call the
1653C<destroy> method.
1654
1655=item How do I read data until the other side closes the connection?
1656
1657If you just want to read your data into a perl scalar, the easiest way
1658to achieve this is by setting an C<on_read> callback that does nothing,
1659clearing the C<on_eof> callback and in the C<on_error> callback, the data
1660will be in C<$_[0]{rbuf}>:
1661
1662 $handle->on_read (sub { });
1663 $handle->on_eof (undef);
1664 $handle->on_error (sub {
1665 my $data = delete $_[0]{rbuf};
1666 undef $handle;
1667 });
1668
1669The reason to use C<on_error> is that TCP connections, due to latencies
1670and packets loss, might get closed quite violently with an error, when in
1671fact, all data has been received.
1672
1673It is usually better to use acknowledgements when transferring data,
1674to make sure the other side hasn't just died and you got the data
1675intact. This is also one reason why so many internet protocols have an
1676explicit QUIT command.
1677
1678=item I don't want to destroy the handle too early - how do I wait until
1679all data has been written?
1680
1681After writing your last bits of data, set the C<on_drain> callback
1682and destroy the handle in there - with the default setting of
1683C<low_water_mark> this will be called precisely when all data has been
1684written to the socket:
1685
1686 $handle->push_write (...);
1687 $handle->on_drain (sub {
1688 warn "all data submitted to the kernel\n";
1689 undef $handle;
1690 });
1691
1692=back
1693
1694
1695=head1 SUBCLASSING AnyEvent::Handle
1696
1697In many cases, you might want to subclass AnyEvent::Handle.
1698
1699To make this easier, a given version of AnyEvent::Handle uses these
1700conventions:
1701
1702=over 4
1703
1704=item * all constructor arguments become object members.
1705
1706At least initially, when you pass a C<tls>-argument to the constructor it
1707will end up in C<< $handle->{tls} >>. Those members might be changed or
1708mutated later on (for example C<tls> will hold the TLS connection object).
1709
1710=item * other object member names are prefixed with an C<_>.
1711
1712All object members not explicitly documented (internal use) are prefixed
1713with an underscore character, so the remaining non-C<_>-namespace is free
1714for use for subclasses.
1715
1716=item * all members not documented here and not prefixed with an underscore
1717are free to use in subclasses.
1718
1719Of course, new versions of AnyEvent::Handle may introduce more "public"
1720member variables, but thats just life, at least it is documented.
1721
1722=back
1723
603=head1 AUTHOR 1724=head1 AUTHOR
604 1725
605Robin Redeker C<< <elmex at ta-sa.org> >>, Marc Lehmann <schmorp@schmorp.de>. 1726Robin Redeker C<< <elmex at ta-sa.org> >>, Marc Lehmann <schmorp@schmorp.de>.
606 1727
607=cut 1728=cut

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines