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 1.42 by root, Tue May 27 06:23:15 2008 UTC vs.
Revision 1.233 by root, Thu Apr 5 06:14:10 2012 UTC

1package AnyEvent::Handle;
2
3no warnings;
4use strict;
5
6use AnyEvent ();
7use AnyEvent::Util qw(WSAEWOULDBLOCK);
8use Scalar::Util ();
9use Carp ();
10use Fcntl ();
11use Errno qw/EAGAIN EINTR/;
12
13=head1 NAME 1=head1 NAME
14 2
15AnyEvent::Handle - non-blocking I/O on file handles via AnyEvent 3AnyEvent::Handle - non-blocking I/O on streaming handles via AnyEvent
16
17=cut
18
19our $VERSION = '0.04';
20 4
21=head1 SYNOPSIS 5=head1 SYNOPSIS
22 6
23 use AnyEvent; 7 use AnyEvent;
24 use AnyEvent::Handle; 8 use AnyEvent::Handle;
25 9
26 my $cv = AnyEvent->condvar; 10 my $cv = AnyEvent->condvar;
27 11
28 my $handle = 12 my $hdl; $hdl = new AnyEvent::Handle
29 AnyEvent::Handle->new (
30 fh => \*STDIN, 13 fh => \*STDIN,
31 on_eof => sub { 14 on_error => sub {
32 $cv->broadcast; 15 my ($hdl, $fatal, $msg) = @_;
33 }, 16 AE::log error => $msg;
17 $hdl->destroy;
18 $cv->send;
34 ); 19 };
35 20
36 # send some request line 21 # send some request line
37 $handle->push_write ("getinfo\015\012"); 22 $hdl->push_write ("getinfo\015\012");
38 23
39 # read the response line 24 # read the response line
40 $handle->push_read (line => sub { 25 $hdl->push_read (line => sub {
41 my ($handle, $line) = @_; 26 my ($hdl, $line) = @_;
42 warn "read line <$line>\n"; 27 say "got line <$line>";
43 $cv->send; 28 $cv->send;
44 }); 29 });
45 30
46 $cv->recv; 31 $cv->recv;
47 32
48=head1 DESCRIPTION 33=head1 DESCRIPTION
49 34
50This module is a helper module to make it easier to do event-based I/O on 35This is a helper module to make it easier to do event-based I/O on
51filehandles. For utility functions for doing non-blocking connects and accepts 36stream-based filehandles (sockets, pipes, and other stream things).
52on sockets see L<AnyEvent::Util>.
53 37
38The L<AnyEvent::Intro> tutorial contains some well-documented
39AnyEvent::Handle examples.
40
54In the following, when the documentation refers to of "bytes" then this 41In the following, where the documentation refers to "bytes", it means
55means characters. As sysread and syswrite are used for all I/O, their 42characters. As sysread and syswrite are used for all I/O, their
56treatment of characters applies to this module as well. 43treatment of characters applies to this module as well.
44
45At the very minimum, you should specify C<fh> or C<connect>, and the
46C<on_error> callback.
57 47
58All callbacks will be invoked with the handle object as their first 48All callbacks will be invoked with the handle object as their first
59argument. 49argument.
60 50
51=cut
52
53package AnyEvent::Handle;
54
55use Scalar::Util ();
56use List::Util ();
57use Carp ();
58use Errno qw(EAGAIN EINTR);
59
60use AnyEvent (); BEGIN { AnyEvent::common_sense }
61use AnyEvent::Util qw(WSAEWOULDBLOCK);
62
63our $VERSION = $AnyEvent::VERSION;
64
65sub _load_func($) {
66 my $func = $_[0];
67
68 unless (defined &$func) {
69 my $pkg = $func;
70 do {
71 $pkg =~ s/::[^:]+$//
72 or return;
73 eval "require $pkg";
74 } until defined &$func;
75 }
76
77 \&$func
78}
79
80sub MAX_READ_SIZE() { 131072 }
81
61=head1 METHODS 82=head1 METHODS
62 83
63=over 4 84=over 4
64 85
65=item B<new (%args)> 86=item $handle = B<new> AnyEvent::Handle fh => $filehandle, key => value...
66 87
67The constructor supports these arguments (all as key => value pairs). 88The constructor supports these arguments (all as C<< key => value >> pairs).
68 89
69=over 4 90=over 4
70 91
71=item fh => $filehandle [MANDATORY] 92=item fh => $filehandle [C<fh> or C<connect> MANDATORY]
72 93
73The filehandle this L<AnyEvent::Handle> object will operate on. 94The filehandle this L<AnyEvent::Handle> object will operate on.
74
75NOTE: The filehandle will be set to non-blocking (using 95NOTE: The filehandle will be set to non-blocking mode (using
76AnyEvent::Util::fh_nonblocking). 96C<AnyEvent::Util::fh_nonblocking>) by the constructor and needs to stay in
97that mode.
77 98
78=item on_eof => $cb->($handle) 99=item connect => [$host, $service] [C<fh> or C<connect> MANDATORY]
79 100
80Set the callback to be called on EOF. 101Try to connect to the specified host and service (port), using
102C<AnyEvent::Socket::tcp_connect>. The C<$host> additionally becomes the
103default C<peername>.
81 104
82While not mandatory, it is highly recommended to set an eof callback, 105You have to specify either this parameter, or C<fh>, above.
83otherwise you might end up with a closed socket while you are still
84waiting for data.
85 106
107It is possible to push requests on the read and write queues, and modify
108properties of the stream, even while AnyEvent::Handle is connecting.
109
110When this parameter is specified, then the C<on_prepare>,
111C<on_connect_error> and C<on_connect> callbacks will be called under the
112appropriate circumstances:
113
114=over 4
115
86=item on_error => $cb->($handle) 116=item on_prepare => $cb->($handle)
87 117
118This (rarely used) callback is called before a new connection is
119attempted, but after the file handle has been created (you can access that
120file handle via C<< $handle->{fh} >>). It could be used to prepare the
121file handle with parameters required for the actual connect (as opposed to
122settings that can be changed when the connection is already established).
123
124The return value of this callback should be the connect timeout value in
125seconds (or C<0>, or C<undef>, or the empty list, to indicate that the
126default timeout is to be used).
127
128=item on_connect => $cb->($handle, $host, $port, $retry->())
129
130This callback is called when a connection has been successfully established.
131
132The peer's numeric host and port (the socket peername) are passed as
133parameters, together with a retry callback. At the time it is called the
134read and write queues, EOF status, TLS status and similar properties of
135the handle will have been reset.
136
137It is not allowed to use the read or write queues while the handle object
138is connecting.
139
140If, for some reason, the handle is not acceptable, calling C<$retry> will
141continue with the next connection target (in case of multi-homed hosts or
142SRV records there can be multiple connection endpoints). The C<$retry>
143callback can be invoked after the connect callback returns, i.e. one can
144start a handshake and then decide to retry with the next host if the
145handshake fails.
146
147In most cases, you should ignore the C<$retry> parameter.
148
149=item on_connect_error => $cb->($handle, $message)
150
151This callback is called when the connection could not be
152established. C<$!> will contain the relevant error code, and C<$message> a
153message describing it (usually the same as C<"$!">).
154
155If this callback isn't specified, then C<on_error> will be called with a
156fatal error instead.
157
158=back
159
160=item on_error => $cb->($handle, $fatal, $message)
161
88This is the fatal error callback, that is called when, well, a fatal error 162This is the error callback, which is called when, well, some error
89occurs, such as not being able to resolve the hostname, failure to connect 163occured, such as not being able to resolve the hostname, failure to
90or a read error. 164connect, or a read error.
91 165
92The object will not be in a usable state when this callback has been 166Some errors are fatal (which is indicated by C<$fatal> being true). On
93called. 167fatal errors the handle object will be destroyed (by a call to C<< ->
168destroy >>) after invoking the error callback (which means you are free to
169examine the handle object). Examples of fatal errors are an EOF condition
170with active (but unsatisfiable) read watchers (C<EPIPE>) or I/O errors. In
171cases where the other side can close the connection at will, it is
172often easiest to not report C<EPIPE> errors in this callback.
94 173
174AnyEvent::Handle tries to find an appropriate error code for you to check
175against, but in some cases (TLS errors), this does not work well.
176
177If you report the error to the user, it is recommended to always output
178the C<$message> argument in human-readable error messages (you don't need
179to report C<"$!"> if you report C<$message>).
180
181If you want to react programmatically to the error, then looking at C<$!>
182and comparing it against some of the documented C<Errno> values is usually
183better than looking at the C<$message>.
184
185Non-fatal errors can be retried by returning, but it is recommended
186to simply ignore this parameter and instead abondon the handle object
187when this callback is invoked. Examples of non-fatal errors are timeouts
188C<ETIMEDOUT>) or badly-formatted data (C<EBADMSG>).
189
95On callback entrance, the value of C<$!> contains the operating system 190On entry to the callback, the value of C<$!> contains the operating
96error (or C<ENOSPC>, C<EPIPE> or C<EBADMSG>). 191system error code (or C<ENOSPC>, C<EPIPE>, C<ETIMEDOUT>, C<EBADMSG> or
97 192C<EPROTO>).
98The callback should throw an exception. If it returns, then
99AnyEvent::Handle will C<croak> for you.
100 193
101While not mandatory, it is I<highly> recommended to set this callback, as 194While not mandatory, it is I<highly> recommended to set this callback, as
102you will not be notified of errors otherwise. The default simply calls 195you will not be notified of errors otherwise. The default just calls
103die. 196C<croak>.
104 197
105=item on_read => $cb->($handle) 198=item on_read => $cb->($handle)
106 199
107This sets the default read callback, which is called when data arrives 200This sets the default read callback, which is called when data arrives
108and no read request is in the queue. 201and no read request is in the queue (unlike read queue callbacks, this
202callback will only be called when at least one octet of data is in the
203read buffer).
109 204
110To access (and remove data from) the read buffer, use the C<< ->rbuf >> 205To access (and remove data from) the read buffer, use the C<< ->rbuf >>
111method or access the C<$handle->{rbuf}> member directly. 206method or access the C<< $handle->{rbuf} >> member directly. Note that you
207must not enlarge or modify the read buffer, you can only remove data at
208the beginning from it.
112 209
210You can also call C<< ->push_read (...) >> or any other function that
211modifies the read queue. Or do both. Or ...
212
113When an EOF condition is detected then AnyEvent::Handle will first try to 213When an EOF condition is detected, AnyEvent::Handle will first try to
114feed all the remaining data to the queued callbacks and C<on_read> before 214feed all the remaining data to the queued callbacks and C<on_read> before
115calling the C<on_eof> callback. If no progress can be made, then a fatal 215calling the C<on_eof> callback. If no progress can be made, then a fatal
116error will be raised (with C<$!> set to C<EPIPE>). 216error will be raised (with C<$!> set to C<EPIPE>).
117 217
218Note that, unlike requests in the read queue, an C<on_read> callback
219doesn't mean you I<require> some data: if there is an EOF and there
220are outstanding read requests then an error will be flagged. With an
221C<on_read> callback, the C<on_eof> callback will be invoked.
222
223=item on_eof => $cb->($handle)
224
225Set the callback to be called when an end-of-file condition is detected,
226i.e. in the case of a socket, when the other side has closed the
227connection cleanly, and there are no outstanding read requests in the
228queue (if there are read requests, then an EOF counts as an unexpected
229connection close and will be flagged as an error).
230
231For sockets, this just means that the other side has stopped sending data,
232you can still try to write data, and, in fact, one can return from the EOF
233callback and continue writing data, as only the read part has been shut
234down.
235
236If an EOF condition has been detected but no C<on_eof> callback has been
237set, then a fatal error will be raised with C<$!> set to <0>.
238
118=item on_drain => $cb->($handle) 239=item on_drain => $cb->($handle)
119 240
120This sets the callback that is called when the write buffer becomes empty 241This sets the callback that is called once when the write buffer becomes
121(or when the callback is set and the buffer is empty already). 242empty (and immediately when the handle object is created).
122 243
123To append to the write buffer, use the C<< ->push_write >> method. 244To append to the write buffer, use the C<< ->push_write >> method.
245
246This callback is useful when you don't want to put all of your write data
247into the queue at once, for example, when you want to write the contents
248of some file to the socket you might not want to read the whole file into
249memory and push it into the queue, but instead only read more data from
250the file when the write queue becomes empty.
251
252=item timeout => $fractional_seconds
253
254=item rtimeout => $fractional_seconds
255
256=item wtimeout => $fractional_seconds
257
258If non-zero, then these enables an "inactivity" timeout: whenever this
259many seconds pass without a successful read or write on the underlying
260file handle (or a call to C<timeout_reset>), the C<on_timeout> callback
261will be invoked (and if that one is missing, a non-fatal C<ETIMEDOUT>
262error will be raised).
263
264There are three variants of the timeouts that work independently of each
265other, for both read and write (triggered when nothing was read I<OR>
266written), just read (triggered when nothing was read), and just write:
267C<timeout>, C<rtimeout> and C<wtimeout>, with corresponding callbacks
268C<on_timeout>, C<on_rtimeout> and C<on_wtimeout>, and reset functions
269C<timeout_reset>, C<rtimeout_reset>, and C<wtimeout_reset>.
270
271Note that timeout processing is active even when you do not have any
272outstanding read or write requests: If you plan to keep the connection
273idle then you should disable the timeout temporarily or ignore the
274timeout in the corresponding C<on_timeout> callback, in which case
275AnyEvent::Handle will simply restart the timeout.
276
277Zero (the default) disables the corresponding timeout.
278
279=item on_timeout => $cb->($handle)
280
281=item on_rtimeout => $cb->($handle)
282
283=item on_wtimeout => $cb->($handle)
284
285Called whenever the inactivity timeout passes. If you return from this
286callback, then the timeout will be reset as if some activity had happened,
287so this condition is not fatal in any way.
124 288
125=item rbuf_max => <bytes> 289=item rbuf_max => <bytes>
126 290
127If defined, then a fatal error will be raised (with C<$!> set to C<ENOSPC>) 291If defined, then a fatal error will be raised (with C<$!> set to C<ENOSPC>)
128when the read buffer ever (strictly) exceeds this size. This is useful to 292when the read buffer ever (strictly) exceeds this size. This is useful to
129avoid denial-of-service attacks. 293avoid some forms of denial-of-service attacks.
130 294
131For example, a server accepting connections from untrusted sources should 295For example, a server accepting connections from untrusted sources should
132be configured to accept only so-and-so much data that it cannot act on 296be configured to accept only so-and-so much data that it cannot act on
133(for example, when expecting a line, an attacker could send an unlimited 297(for example, when expecting a line, an attacker could send an unlimited
134amount of data without a callback ever being called as long as the line 298amount of data without a callback ever being called as long as the line
135isn't finished). 299isn't finished).
136 300
301=item wbuf_max => <bytes>
302
303If defined, then a fatal error will be raised (with C<$!> set to C<ENOSPC>)
304when the write buffer ever (strictly) exceeds this size. This is useful to
305avoid some forms of denial-of-service attacks.
306
307Although the units of this parameter is bytes, this is the I<raw> number
308of bytes not yet accepted by the kernel. This can make a difference when
309you e.g. use TLS, as TLS typically makes your write data larger (but it
310can also make it smaller due to compression).
311
312As an example of when this limit is useful, take a chat server that sends
313chat messages to a client. If the client does not read those in a timely
314manner then the send buffer in the server would grow unbounded.
315
316=item autocork => <boolean>
317
318When disabled (the default), C<push_write> will try to immediately
319write the data to the handle if possible. This avoids having to register
320a write watcher and wait for the next event loop iteration, but can
321be inefficient if you write multiple small chunks (on the wire, this
322disadvantage is usually avoided by your kernel's nagle algorithm, see
323C<no_delay>, but this option can save costly syscalls).
324
325When enabled, writes will always be queued till the next event loop
326iteration. This is efficient when you do many small writes per iteration,
327but less efficient when you do a single write only per iteration (or when
328the write buffer often is full). It also increases write latency.
329
330=item no_delay => <boolean>
331
332When doing small writes on sockets, your operating system kernel might
333wait a bit for more data before actually sending it out. This is called
334the Nagle algorithm, and usually it is beneficial.
335
336In some situations you want as low a delay as possible, which can be
337accomplishd by setting this option to a true value.
338
339The default is your operating system's default behaviour (most likely
340enabled). This option explicitly enables or disables it, if possible.
341
342=item keepalive => <boolean>
343
344Enables (default disable) the SO_KEEPALIVE option on the stream socket:
345normally, TCP connections have no time-out once established, so TCP
346connections, once established, can stay alive forever even when the other
347side has long gone. TCP keepalives are a cheap way to take down long-lived
348TCP connections when the other side becomes unreachable. While the default
349is OS-dependent, TCP keepalives usually kick in after around two hours,
350and, if the other side doesn't reply, take down the TCP connection some 10
351to 15 minutes later.
352
353It is harmless to specify this option for file handles that do not support
354keepalives, and enabling it on connections that are potentially long-lived
355is usually a good idea.
356
357=item oobinline => <boolean>
358
359BSD majorly fucked up the implementation of TCP urgent data. The result
360is that almost no OS implements TCP according to the specs, and every OS
361implements it slightly differently.
362
363If you want to handle TCP urgent data, then setting this flag (the default
364is enabled) gives you the most portable way of getting urgent data, by
365putting it into the stream.
366
367Since BSD emulation of OOB data on top of TCP's urgent data can have
368security implications, AnyEvent::Handle sets this flag automatically
369unless explicitly specified. Note that setting this flag after
370establishing a connection I<may> be a bit too late (data loss could
371already have occured on BSD systems), but at least it will protect you
372from most attacks.
373
137=item read_size => <bytes> 374=item read_size => <bytes>
138 375
139The default read block size (the amount of bytes this module will try to read 376The initial read block size, the number of bytes this module will try
140on each [loop iteration). Default: C<4096>. 377to read during each loop iteration. Each handle object will consume
378at least this amount of memory for the read buffer as well, so when
379handling many connections watch out for memory requirements). See also
380C<max_read_size>. Default: C<2048>.
381
382=item max_read_size => <bytes>
383
384The maximum read buffer size used by the dynamic adjustment
385algorithm: Each time AnyEvent::Handle can read C<read_size> bytes in
386one go it will double C<read_size> up to the maximum given by this
387option. Default: C<131072> or C<read_size>, whichever is higher.
141 388
142=item low_water_mark => <bytes> 389=item low_water_mark => <bytes>
143 390
144Sets the amount of bytes (default: C<0>) that make up an "empty" write 391Sets the number of bytes (default: C<0>) that make up an "empty" write
145buffer: If the write reaches this size or gets even samller it is 392buffer: If the buffer reaches this size or gets even samller it is
146considered empty. 393considered empty.
147 394
395Sometimes it can be beneficial (for performance reasons) to add data to
396the write buffer before it is fully drained, but this is a rare case, as
397the operating system kernel usually buffers data as well, so the default
398is good in almost all cases.
399
400=item linger => <seconds>
401
402If this is non-zero (default: C<3600>), the destructor of the
403AnyEvent::Handle object will check whether there is still outstanding
404write data and will install a watcher that will write this data to the
405socket. No errors will be reported (this mostly matches how the operating
406system treats outstanding data at socket close time).
407
408This will not work for partial TLS data that could not be encoded
409yet. This data will be lost. Calling the C<stoptls> method in time might
410help.
411
412=item peername => $string
413
414A string used to identify the remote site - usually the DNS hostname
415(I<not> IDN!) used to create the connection, rarely the IP address.
416
417Apart from being useful in error messages, this string is also used in TLS
418peername verification (see C<verify_peername> in L<AnyEvent::TLS>). This
419verification will be skipped when C<peername> is not specified or is
420C<undef>.
421
148=item tls => "accept" | "connect" | Net::SSLeay::SSL object 422=item tls => "accept" | "connect" | Net::SSLeay::SSL object
149 423
150When this parameter is given, it enables TLS (SSL) mode, that means it 424When this parameter is given, it enables TLS (SSL) mode, that means
151will start making tls handshake and will transparently encrypt/decrypt 425AnyEvent will start a TLS handshake as soon as the connection has been
152data. 426established and will transparently encrypt/decrypt data afterwards.
427
428All TLS protocol errors will be signalled as C<EPROTO>, with an
429appropriate error message.
153 430
154TLS mode requires Net::SSLeay to be installed (it will be loaded 431TLS mode requires Net::SSLeay to be installed (it will be loaded
155automatically when you try to create a TLS handle). 432automatically when you try to create a TLS handle): this module doesn't
433have a dependency on that module, so if your module requires it, you have
434to add the dependency yourself.
156 435
157For the TLS server side, use C<accept>, and for the TLS client side of a 436Unlike TCP, TLS has a server and client side: for the TLS server side, use
158connection, use C<connect> mode. 437C<accept>, and for the TLS client side of a connection, use C<connect>
438mode.
159 439
160You can also provide your own TLS connection object, but you have 440You can also provide your own TLS connection object, but you have
161to make sure that you call either C<Net::SSLeay::set_connect_state> 441to make sure that you call either C<Net::SSLeay::set_connect_state>
162or C<Net::SSLeay::set_accept_state> on it before you pass it to 442or C<Net::SSLeay::set_accept_state> on it before you pass it to
163AnyEvent::Handle. 443AnyEvent::Handle. Also, this module will take ownership of this connection
444object.
164 445
446At some future point, AnyEvent::Handle might switch to another TLS
447implementation, then the option to use your own session object will go
448away.
449
450B<IMPORTANT:> since Net::SSLeay "objects" are really only integers,
451passing in the wrong integer will lead to certain crash. This most often
452happens when one uses a stylish C<< tls => 1 >> and is surprised about the
453segmentation fault.
454
165See the C<starttls> method if you need to start TLs negotiation later. 455Use the C<< ->starttls >> method if you need to start TLS negotiation later.
166 456
167=item tls_ctx => $ssl_ctx 457=item tls_ctx => $anyevent_tls
168 458
169Use the given Net::SSLeay::CTX object to create the new TLS connection 459Use the given C<AnyEvent::TLS> object to create the new TLS connection
170(unless a connection object was specified directly). If this parameter is 460(unless a connection object was specified directly). If this
171missing, then AnyEvent::Handle will use C<AnyEvent::Handle::TLS_CTX>. 461parameter is missing (or C<undef>), then AnyEvent::Handle will use
462C<AnyEvent::Handle::TLS_CTX>.
463
464Instead of an object, you can also specify a hash reference with C<< key
465=> value >> pairs. Those will be passed to L<AnyEvent::TLS> to create a
466new TLS context object.
467
468=item on_starttls => $cb->($handle, $success[, $error_message])
469
470This callback will be invoked when the TLS/SSL handshake has finished. If
471C<$success> is true, then the TLS handshake succeeded, otherwise it failed
472(C<on_stoptls> will not be called in this case).
473
474The session in C<< $handle->{tls} >> can still be examined in this
475callback, even when the handshake was not successful.
476
477TLS handshake failures will not cause C<on_error> to be invoked when this
478callback is in effect, instead, the error message will be passed to C<on_starttls>.
479
480Without this callback, handshake failures lead to C<on_error> being
481called as usual.
482
483Note that you cannot just call C<starttls> again in this callback. If you
484need to do that, start an zero-second timer instead whose callback can
485then call C<< ->starttls >> again.
486
487=item on_stoptls => $cb->($handle)
488
489When a SSLv3/TLS shutdown/close notify/EOF is detected and this callback is
490set, then it will be invoked after freeing the TLS session. If it is not,
491then a TLS shutdown condition will be treated like a normal EOF condition
492on the handle.
493
494The session in C<< $handle->{tls} >> can still be examined in this
495callback.
496
497This callback will only be called on TLS shutdowns, not when the
498underlying handle signals EOF.
172 499
173=item json => JSON or JSON::XS object 500=item json => JSON or JSON::XS object
174 501
175This is the json coder object used by the C<json> read and write types. 502This is the json coder object used by the C<json> read and write types.
176 503
177If you don't supply it, then AnyEvent::Handle will create and use a 504If you don't supply it, then AnyEvent::Handle will create and use a
178suitable one, which will write and expect UTF-8 encoded JSON texts. 505suitable one (on demand), which will write and expect UTF-8 encoded JSON
506texts.
179 507
180Note that you are responsible to depend on the JSON module if you want to 508Note that you are responsible to depend on the JSON module if you want to
181use this functionality, as AnyEvent does not have a dependency itself. 509use this functionality, as AnyEvent does not have a dependency itself.
182 510
183=item filter_r => $cb
184
185=item filter_w => $cb
186
187These exist, but are undocumented at this time.
188
189=back 511=back
190 512
191=cut 513=cut
192 514
193sub new { 515sub new {
194 my $class = shift; 516 my $class = shift;
195
196 my $self = bless { @_ }, $class; 517 my $self = bless { @_ }, $class;
197 518
198 $self->{fh} or Carp::croak "mandatory argument fh is missing"; 519 if ($self->{fh}) {
520 $self->_start;
521 return unless $self->{fh}; # could be gone by now
522
523 } elsif ($self->{connect}) {
524 require AnyEvent::Socket;
525
526 $self->{peername} = $self->{connect}[0]
527 unless exists $self->{peername};
528
529 $self->{_skip_drain_rbuf} = 1;
530
531 {
532 Scalar::Util::weaken (my $self = $self);
533
534 $self->{_connect} =
535 AnyEvent::Socket::tcp_connect (
536 $self->{connect}[0],
537 $self->{connect}[1],
538 sub {
539 my ($fh, $host, $port, $retry) = @_;
540
541 delete $self->{_connect}; # no longer needed
542
543 if ($fh) {
544 $self->{fh} = $fh;
545
546 delete $self->{_skip_drain_rbuf};
547 $self->_start;
548
549 $self->{on_connect}
550 and $self->{on_connect}($self, $host, $port, sub {
551 delete @$self{qw(fh _tw _rtw _wtw _ww _rw _eof _queue rbuf _wbuf tls _tls_rbuf _tls_wbuf)};
552 $self->{_skip_drain_rbuf} = 1;
553 &$retry;
554 });
555
556 } else {
557 if ($self->{on_connect_error}) {
558 $self->{on_connect_error}($self, "$!");
559 $self->destroy if $self;
560 } else {
561 $self->_error ($!, 1);
562 }
563 }
564 },
565 sub {
566 local $self->{fh} = $_[0];
567
568 $self->{on_prepare}
569 ? $self->{on_prepare}->($self)
570 : ()
571 }
572 );
573 }
574
575 } else {
576 Carp::croak "AnyEvent::Handle: either an existing fh or the connect parameter must be specified";
577 }
578
579 $self
580}
581
582sub _start {
583 my ($self) = @_;
584
585 # too many clueless people try to use udp and similar sockets
586 # with AnyEvent::Handle, do them a favour.
587 my $type = getsockopt $self->{fh}, Socket::SOL_SOCKET (), Socket::SO_TYPE ();
588 Carp::croak "AnyEvent::Handle: only stream sockets supported, anything else will NOT work!"
589 if Socket::SOCK_STREAM () != (unpack "I", $type) && defined $type;
199 590
200 AnyEvent::Util::fh_nonblocking $self->{fh}, 1; 591 AnyEvent::Util::fh_nonblocking $self->{fh}, 1;
201 592
202 if ($self->{tls}) { 593 $self->{_activity} =
203 require Net::SSLeay; 594 $self->{_ractivity} =
595 $self->{_wactivity} = AE::now;
596
597 $self->{read_size} ||= 2048;
598 $self->{max_read_size} = $self->{read_size}
599 if $self->{read_size} > ($self->{max_read_size} || MAX_READ_SIZE);
600
601 $self->timeout (delete $self->{timeout} ) if $self->{timeout};
602 $self->rtimeout (delete $self->{rtimeout} ) if $self->{rtimeout};
603 $self->wtimeout (delete $self->{wtimeout} ) if $self->{wtimeout};
604
605 $self->no_delay (delete $self->{no_delay} ) if exists $self->{no_delay} && $self->{no_delay};
606 $self->keepalive (delete $self->{keepalive}) if exists $self->{keepalive} && $self->{keepalive};
607
608 $self->oobinline (exists $self->{oobinline} ? delete $self->{oobinline} : 1);
609
204 $self->starttls (delete $self->{tls}, delete $self->{tls_ctx}); 610 $self->starttls (delete $self->{tls}, delete $self->{tls_ctx})
205 } 611 if $self->{tls};
206 612
207 $self->on_eof (delete $self->{on_eof} ) if $self->{on_eof};
208 $self->on_error (delete $self->{on_error}) if $self->{on_error};
209 $self->on_drain (delete $self->{on_drain}) if $self->{on_drain}; 613 $self->on_drain (delete $self->{on_drain} ) if $self->{on_drain};
210 $self->on_read (delete $self->{on_read} ) if $self->{on_read};
211 614
212 $self->start_read; 615 $self->start_read
616 if $self->{on_read} || @{ $self->{_queue} };
213 617
214 $self 618 $self->_drain_wbuf;
215} 619}
216 620
217sub _shutdown {
218 my ($self) = @_;
219
220 delete $self->{_rw};
221 delete $self->{_ww};
222 delete $self->{fh};
223}
224
225sub error { 621sub _error {
226 my ($self) = @_; 622 my ($self, $errno, $fatal, $message) = @_;
227 623
228 { 624 $! = $errno;
229 local $!; 625 $message ||= "$!";
230 $self->_shutdown;
231 }
232 626
233 $self->{on_error}($self)
234 if $self->{on_error}; 627 if ($self->{on_error}) {
235 628 $self->{on_error}($self, $fatal, $message);
629 $self->destroy if $fatal;
630 } elsif ($self->{fh} || $self->{connect}) {
631 $self->destroy;
236 Carp::croak "AnyEvent::Handle uncaught fatal error: $!"; 632 Carp::croak "AnyEvent::Handle uncaught error: $message";
633 }
237} 634}
238 635
239=item $fh = $handle->fh 636=item $fh = $handle->fh
240 637
241This method returns the file handle of the L<AnyEvent::Handle> object. 638This method returns the file handle used to create the L<AnyEvent::Handle> object.
242 639
243=cut 640=cut
244 641
245sub fh { $_[0]{fh} } 642sub fh { $_[0]{fh} }
246 643
262 659
263sub on_eof { 660sub on_eof {
264 $_[0]{on_eof} = $_[1]; 661 $_[0]{on_eof} = $_[1];
265} 662}
266 663
664=item $handle->on_timeout ($cb)
665
666=item $handle->on_rtimeout ($cb)
667
668=item $handle->on_wtimeout ($cb)
669
670Replace the current C<on_timeout>, C<on_rtimeout> or C<on_wtimeout>
671callback, or disables the callback (but not the timeout) if C<$cb> =
672C<undef>. See the C<timeout> constructor argument and method.
673
674=cut
675
676# see below
677
678=item $handle->autocork ($boolean)
679
680Enables or disables the current autocork behaviour (see C<autocork>
681constructor argument). Changes will only take effect on the next write.
682
683=cut
684
685sub autocork {
686 $_[0]{autocork} = $_[1];
687}
688
689=item $handle->no_delay ($boolean)
690
691Enables or disables the C<no_delay> setting (see constructor argument of
692the same name for details).
693
694=cut
695
696sub no_delay {
697 $_[0]{no_delay} = $_[1];
698
699 setsockopt $_[0]{fh}, Socket::IPPROTO_TCP (), Socket::TCP_NODELAY (), int $_[1]
700 if $_[0]{fh};
701}
702
703=item $handle->keepalive ($boolean)
704
705Enables or disables the C<keepalive> setting (see constructor argument of
706the same name for details).
707
708=cut
709
710sub keepalive {
711 $_[0]{keepalive} = $_[1];
712
713 eval {
714 local $SIG{__DIE__};
715 setsockopt $_[0]{fh}, Socket::SOL_SOCKET (), Socket::SO_KEEPALIVE (), int $_[1]
716 if $_[0]{fh};
717 };
718}
719
720=item $handle->oobinline ($boolean)
721
722Enables or disables the C<oobinline> setting (see constructor argument of
723the same name for details).
724
725=cut
726
727sub oobinline {
728 $_[0]{oobinline} = $_[1];
729
730 eval {
731 local $SIG{__DIE__};
732 setsockopt $_[0]{fh}, Socket::SOL_SOCKET (), Socket::SO_OOBINLINE (), int $_[1]
733 if $_[0]{fh};
734 };
735}
736
737=item $handle->keepalive ($boolean)
738
739Enables or disables the C<keepalive> setting (see constructor argument of
740the same name for details).
741
742=cut
743
744sub keepalive {
745 $_[0]{keepalive} = $_[1];
746
747 eval {
748 local $SIG{__DIE__};
749 setsockopt $_[0]{fh}, Socket::SOL_SOCKET (), Socket::SO_KEEPALIVE (), int $_[1]
750 if $_[0]{fh};
751 };
752}
753
754=item $handle->on_starttls ($cb)
755
756Replace the current C<on_starttls> callback (see the C<on_starttls> constructor argument).
757
758=cut
759
760sub on_starttls {
761 $_[0]{on_starttls} = $_[1];
762}
763
764=item $handle->on_stoptls ($cb)
765
766Replace the current C<on_stoptls> callback (see the C<on_stoptls> constructor argument).
767
768=cut
769
770sub on_stoptls {
771 $_[0]{on_stoptls} = $_[1];
772}
773
774=item $handle->rbuf_max ($max_octets)
775
776Configures the C<rbuf_max> setting (C<undef> disables it).
777
778=item $handle->wbuf_max ($max_octets)
779
780Configures the C<wbuf_max> setting (C<undef> disables it).
781
782=cut
783
784sub rbuf_max {
785 $_[0]{rbuf_max} = $_[1];
786}
787
788sub wbuf_max {
789 $_[0]{wbuf_max} = $_[1];
790}
791
792#############################################################################
793
794=item $handle->timeout ($seconds)
795
796=item $handle->rtimeout ($seconds)
797
798=item $handle->wtimeout ($seconds)
799
800Configures (or disables) the inactivity timeout.
801
802The timeout will be checked instantly, so this method might destroy the
803handle before it returns.
804
805=item $handle->timeout_reset
806
807=item $handle->rtimeout_reset
808
809=item $handle->wtimeout_reset
810
811Reset the activity timeout, as if data was received or sent.
812
813These methods are cheap to call.
814
815=cut
816
817for my $dir ("", "r", "w") {
818 my $timeout = "${dir}timeout";
819 my $tw = "_${dir}tw";
820 my $on_timeout = "on_${dir}timeout";
821 my $activity = "_${dir}activity";
822 my $cb;
823
824 *$on_timeout = sub {
825 $_[0]{$on_timeout} = $_[1];
826 };
827
828 *$timeout = sub {
829 my ($self, $new_value) = @_;
830
831 $new_value >= 0
832 or Carp::croak "AnyEvent::Handle->$timeout called with negative timeout ($new_value), caught";
833
834 $self->{$timeout} = $new_value;
835 delete $self->{$tw}; &$cb;
836 };
837
838 *{"${dir}timeout_reset"} = sub {
839 $_[0]{$activity} = AE::now;
840 };
841
842 # main workhorse:
843 # reset the timeout watcher, as neccessary
844 # also check for time-outs
845 $cb = sub {
846 my ($self) = @_;
847
848 if ($self->{$timeout} && $self->{fh}) {
849 my $NOW = AE::now;
850
851 # when would the timeout trigger?
852 my $after = $self->{$activity} + $self->{$timeout} - $NOW;
853
854 # now or in the past already?
855 if ($after <= 0) {
856 $self->{$activity} = $NOW;
857
858 if ($self->{$on_timeout}) {
859 $self->{$on_timeout}($self);
860 } else {
861 $self->_error (Errno::ETIMEDOUT);
862 }
863
864 # callback could have changed timeout value, optimise
865 return unless $self->{$timeout};
866
867 # calculate new after
868 $after = $self->{$timeout};
869 }
870
871 Scalar::Util::weaken $self;
872 return unless $self; # ->error could have destroyed $self
873
874 $self->{$tw} ||= AE::timer $after, 0, sub {
875 delete $self->{$tw};
876 $cb->($self);
877 };
878 } else {
879 delete $self->{$tw};
880 }
881 }
882}
883
267############################################################################# 884#############################################################################
268 885
269=back 886=back
270 887
271=head2 WRITE QUEUE 888=head2 WRITE QUEUE
275 892
276The write queue is very simple: you can add data to its end, and 893The write queue is very simple: you can add data to its end, and
277AnyEvent::Handle will automatically try to get rid of it for you. 894AnyEvent::Handle will automatically try to get rid of it for you.
278 895
279When data could be written and the write buffer is shorter then the low 896When data could be written and the write buffer is shorter then the low
280water mark, the C<on_drain> callback will be invoked. 897water mark, the C<on_drain> callback will be invoked once.
281 898
282=over 4 899=over 4
283 900
284=item $handle->on_drain ($cb) 901=item $handle->on_drain ($cb)
285 902
286Sets the C<on_drain> callback or clears it (see the description of 903Sets the C<on_drain> callback or clears it (see the description of
287C<on_drain> in the constructor). 904C<on_drain> in the constructor).
288 905
906This method may invoke callbacks (and therefore the handle might be
907destroyed after it returns).
908
289=cut 909=cut
290 910
291sub on_drain { 911sub on_drain {
292 my ($self, $cb) = @_; 912 my ($self, $cb) = @_;
293 913
294 $self->{on_drain} = $cb; 914 $self->{on_drain} = $cb;
295 915
296 $cb->($self) 916 $cb->($self)
297 if $cb && $self->{low_water_mark} >= length $self->{wbuf}; 917 if $cb && $self->{low_water_mark} >= (length $self->{wbuf}) + (length $self->{_tls_wbuf});
298} 918}
299 919
300=item $handle->push_write ($data) 920=item $handle->push_write ($data)
301 921
302Queues the given scalar to be written. You can push as much data as you 922Queues the given scalar to be written. You can push as much data as
303want (only limited by the available memory), as C<AnyEvent::Handle> 923you want (only limited by the available memory and C<wbuf_max>), as
304buffers it independently of the kernel. 924C<AnyEvent::Handle> buffers it independently of the kernel.
925
926This method may invoke callbacks (and therefore the handle might be
927destroyed after it returns).
305 928
306=cut 929=cut
307 930
308sub _drain_wbuf { 931sub _drain_wbuf {
309 my ($self) = @_; 932 my ($self) = @_;
313 Scalar::Util::weaken $self; 936 Scalar::Util::weaken $self;
314 937
315 my $cb = sub { 938 my $cb = sub {
316 my $len = syswrite $self->{fh}, $self->{wbuf}; 939 my $len = syswrite $self->{fh}, $self->{wbuf};
317 940
318 if ($len >= 0) { 941 if (defined $len) {
319 substr $self->{wbuf}, 0, $len, ""; 942 substr $self->{wbuf}, 0, $len, "";
320 943
944 $self->{_activity} = $self->{_wactivity} = AE::now;
945
321 $self->{on_drain}($self) 946 $self->{on_drain}($self)
322 if $self->{low_water_mark} >= length $self->{wbuf} 947 if $self->{low_water_mark} >= (length $self->{wbuf}) + (length $self->{_tls_wbuf})
323 && $self->{on_drain}; 948 && $self->{on_drain};
324 949
325 delete $self->{_ww} unless length $self->{wbuf}; 950 delete $self->{_ww} unless length $self->{wbuf};
326 } elsif ($! != EAGAIN && $! != EINTR && $! != WSAEWOULDBLOCK) { 951 } elsif ($! != EAGAIN && $! != EINTR && $! != WSAEWOULDBLOCK) {
327 $self->error; 952 $self->_error ($!, 1);
328 } 953 }
329 }; 954 };
330 955
331 # try to write data immediately 956 # try to write data immediately
332 $cb->(); 957 $cb->() unless $self->{autocork};
333 958
334 # if still data left in wbuf, we need to poll 959 # if still data left in wbuf, we need to poll
335 $self->{_ww} = AnyEvent->io (fh => $self->{fh}, poll => "w", cb => $cb) 960 $self->{_ww} = AE::io $self->{fh}, 1, $cb
336 if length $self->{wbuf}; 961 if length $self->{wbuf};
962
963 if (
964 defined $self->{wbuf_max}
965 && $self->{wbuf_max} < length $self->{wbuf}
966 ) {
967 $self->_error (Errno::ENOSPC, 1), return;
968 }
337 }; 969 };
338} 970}
339 971
340our %WH; 972our %WH;
341 973
974# deprecated
342sub register_write_type($$) { 975sub register_write_type($$) {
343 $WH{$_[0]} = $_[1]; 976 $WH{$_[0]} = $_[1];
344} 977}
345 978
346sub push_write { 979sub push_write {
347 my $self = shift; 980 my $self = shift;
348 981
349 if (@_ > 1) { 982 if (@_ > 1) {
350 my $type = shift; 983 my $type = shift;
351 984
985 @_ = ($WH{$type} ||= _load_func "$type\::anyevent_write_type"
352 @_ = ($WH{$type} or Carp::croak "unsupported type passed to AnyEvent::Handle::push_write") 986 or Carp::croak "unsupported/unloadable type '$type' passed to AnyEvent::Handle::push_write")
353 ->($self, @_); 987 ->($self, @_);
354 } 988 }
355 989
990 # we downgrade here to avoid hard-to-track-down bugs,
991 # and diagnose the problem earlier and better.
992
356 if ($self->{filter_w}) { 993 if ($self->{tls}) {
357 $self->{filter_w}->($self, \$_[0]); 994 utf8::downgrade $self->{_tls_wbuf} .= $_[0];
995 &_dotls ($self) if $self->{fh};
358 } else { 996 } else {
359 $self->{wbuf} .= $_[0]; 997 utf8::downgrade $self->{wbuf} .= $_[0];
360 $self->_drain_wbuf; 998 $self->_drain_wbuf if $self->{fh};
361 } 999 }
362} 1000}
363 1001
364=item $handle->push_write (type => @args) 1002=item $handle->push_write (type => @args)
365 1003
366=item $handle->unshift_write (type => @args)
367
368Instead of formatting your data yourself, you can also let this module do 1004Instead of formatting your data yourself, you can also let this module
369the job by specifying a type and type-specific arguments. 1005do the job by specifying a type and type-specific arguments. You
1006can also specify the (fully qualified) name of a package, in which
1007case AnyEvent tries to load the package and then expects to find the
1008C<anyevent_write_type> function inside (see "custom write types", below).
370 1009
371Predefined types are (if you have ideas for additional types, feel free to 1010Predefined types are (if you have ideas for additional types, feel free to
372drop by and tell us): 1011drop by and tell us):
373 1012
374=over 4 1013=over 4
376=item netstring => $string 1015=item netstring => $string
377 1016
378Formats the given value as netstring 1017Formats the given value as netstring
379(http://cr.yp.to/proto/netstrings.txt, this is not a recommendation to use them). 1018(http://cr.yp.to/proto/netstrings.txt, this is not a recommendation to use them).
380 1019
381=back
382
383=cut 1020=cut
384 1021
385register_write_type netstring => sub { 1022register_write_type netstring => sub {
386 my ($self, $string) = @_; 1023 my ($self, $string) = @_;
387 1024
388 sprintf "%d:%s,", (length $string), $string 1025 (length $string) . ":$string,"
1026};
1027
1028=item packstring => $format, $data
1029
1030An octet string prefixed with an encoded length. The encoding C<$format>
1031uses the same format as a Perl C<pack> format, but must specify a single
1032integer only (only one of C<cCsSlLqQiInNvVjJw> is allowed, plus an
1033optional C<!>, C<< < >> or C<< > >> modifier).
1034
1035=cut
1036
1037register_write_type packstring => sub {
1038 my ($self, $format, $string) = @_;
1039
1040 pack "$format/a*", $string
389}; 1041};
390 1042
391=item json => $array_or_hashref 1043=item json => $array_or_hashref
392 1044
393Encodes the given hash or array reference into a JSON object. Unless you 1045Encodes the given hash or array reference into a JSON object. Unless you
418Other languages could read single lines terminated by a newline and pass 1070Other languages could read single lines terminated by a newline and pass
419this line into their JSON decoder of choice. 1071this line into their JSON decoder of choice.
420 1072
421=cut 1073=cut
422 1074
1075sub json_coder() {
1076 eval { require JSON::XS; JSON::XS->new->utf8 }
1077 || do { require JSON; JSON->new->utf8 }
1078}
1079
423register_write_type json => sub { 1080register_write_type json => sub {
424 my ($self, $ref) = @_; 1081 my ($self, $ref) = @_;
425 1082
426 require JSON; 1083 my $json = $self->{json} ||= json_coder;
427 1084
428 $self->{json} ? $self->{json}->encode ($ref) 1085 $json->encode ($ref)
429 : JSON::encode_json ($ref)
430}; 1086};
431 1087
432=item AnyEvent::Handle::register_write_type type => $coderef->($handle, @args) 1088=item storable => $reference
433 1089
434This function (not method) lets you add your own types to C<push_write>. 1090Freezes the given reference using L<Storable> and writes it to the
1091handle. Uses the C<nfreeze> format.
1092
1093=cut
1094
1095register_write_type storable => sub {
1096 my ($self, $ref) = @_;
1097
1098 require Storable unless $Storable::VERSION;
1099
1100 pack "w/a*", Storable::nfreeze ($ref)
1101};
1102
1103=back
1104
1105=item $handle->push_shutdown
1106
1107Sometimes you know you want to close the socket after writing your data
1108before it was actually written. One way to do that is to replace your
1109C<on_drain> handler by a callback that shuts down the socket (and set
1110C<low_water_mark> to C<0>). This method is a shorthand for just that, and
1111replaces the C<on_drain> callback with:
1112
1113 sub { shutdown $_[0]{fh}, 1 }
1114
1115This simply shuts down the write side and signals an EOF condition to the
1116the peer.
1117
1118You can rely on the normal read queue and C<on_eof> handling
1119afterwards. This is the cleanest way to close a connection.
1120
1121This method may invoke callbacks (and therefore the handle might be
1122destroyed after it returns).
1123
1124=cut
1125
1126sub push_shutdown {
1127 my ($self) = @_;
1128
1129 delete $self->{low_water_mark};
1130 $self->on_drain (sub { shutdown $_[0]{fh}, 1 });
1131}
1132
1133=item custom write types - Package::anyevent_write_type $handle, @args
1134
1135Instead of one of the predefined types, you can also specify the name of
1136a package. AnyEvent will try to load the package and then expects to find
1137a function named C<anyevent_write_type> inside. If it isn't found, it
1138progressively tries to load the parent package until it either finds the
1139function (good) or runs out of packages (bad).
1140
435Whenever the given C<type> is used, C<push_write> will invoke the code 1141Whenever the given C<type> is used, C<push_write> will the function with
436reference with the handle object and the remaining arguments. 1142the handle object and the remaining arguments.
437 1143
438The code reference is supposed to return a single octet string that will 1144The function is supposed to return a single octet string that will be
439be appended to the write buffer. 1145appended to the write buffer, so you can mentally treat this function as a
1146"arguments to on-the-wire-format" converter.
440 1147
441Note that this is a function, and all types registered this way will be 1148Example: implement a custom write type C<join> that joins the remaining
442global, so try to use unique names. 1149arguments using the first one.
1150
1151 $handle->push_write (My::Type => " ", 1,2,3);
1152
1153 # uses the following package, which can be defined in the "My::Type" or in
1154 # the "My" modules to be auto-loaded, or just about anywhere when the
1155 # My::Type::anyevent_write_type is defined before invoking it.
1156
1157 package My::Type;
1158
1159 sub anyevent_write_type {
1160 my ($handle, $delim, @args) = @_;
1161
1162 join $delim, @args
1163 }
443 1164
444=cut 1165=cut
445 1166
446############################################################################# 1167#############################################################################
447 1168
456ways, the "simple" way, using only C<on_read> and the "complex" way, using 1177ways, the "simple" way, using only C<on_read> and the "complex" way, using
457a queue. 1178a queue.
458 1179
459In the simple case, you just install an C<on_read> callback and whenever 1180In the simple case, you just install an C<on_read> callback and whenever
460new data arrives, it will be called. You can then remove some data (if 1181new data arrives, it will be called. You can then remove some data (if
461enough is there) from the read buffer (C<< $handle->rbuf >>) if you want 1182enough is there) from the read buffer (C<< $handle->rbuf >>). Or you can
462or not. 1183leave the data there if you want to accumulate more (e.g. when only a
1184partial message has been received so far), or change the read queue with
1185e.g. C<push_read>.
463 1186
464In the more complex case, you want to queue multiple callbacks. In this 1187In the more complex case, you want to queue multiple callbacks. In this
465case, AnyEvent::Handle will call the first queued callback each time new 1188case, AnyEvent::Handle will call the first queued callback each time new
466data arrives and removes it when it has done its job (see C<push_read>, 1189data arrives (also the first time it is queued) and remove it when it has
467below). 1190done its job (see C<push_read>, below).
468 1191
469This way you can, for example, push three line-reads, followed by reading 1192This way you can, for example, push three line-reads, followed by reading
470a chunk of data, and AnyEvent::Handle will execute them in order. 1193a chunk of data, and AnyEvent::Handle will execute them in order.
471 1194
472Example 1: EPP protocol parser. EPP sends 4 byte length info, followed by 1195Example 1: EPP protocol parser. EPP sends 4 byte length info, followed by
473the specified number of bytes which give an XML datagram. 1196the specified number of bytes which give an XML datagram.
474 1197
475 # in the default state, expect some header bytes 1198 # in the default state, expect some header bytes
476 $handle->on_read (sub { 1199 $handle->on_read (sub {
477 # some data is here, now queue the length-header-read (4 octets) 1200 # some data is here, now queue the length-header-read (4 octets)
478 shift->unshift_read_chunk (4, sub { 1201 shift->unshift_read (chunk => 4, sub {
479 # header arrived, decode 1202 # header arrived, decode
480 my $len = unpack "N", $_[1]; 1203 my $len = unpack "N", $_[1];
481 1204
482 # now read the payload 1205 # now read the payload
483 shift->unshift_read_chunk ($len, sub { 1206 shift->unshift_read (chunk => $len, sub {
484 my $xml = $_[1]; 1207 my $xml = $_[1];
485 # handle xml 1208 # handle xml
486 }); 1209 });
487 }); 1210 });
488 }); 1211 });
489 1212
490Example 2: Implement a client for a protocol that replies either with 1213Example 2: Implement a client for a protocol that replies either with "OK"
491"OK" and another line or "ERROR" for one request, and 64 bytes for the 1214and another line or "ERROR" for the first request that is sent, and 64
492second request. Due tot he availability of a full queue, we can just 1215bytes for the second request. Due to the availability of a queue, we can
493pipeline sending both requests and manipulate the queue as necessary in 1216just pipeline sending both requests and manipulate the queue as necessary
494the callbacks: 1217in the callbacks.
495 1218
496 # request one 1219When the first callback is called and sees an "OK" response, it will
1220C<unshift> another line-read. This line-read will be queued I<before> the
122164-byte chunk callback.
1222
1223 # request one, returns either "OK + extra line" or "ERROR"
497 $handle->push_write ("request 1\015\012"); 1224 $handle->push_write ("request 1\015\012");
498 1225
499 # we expect "ERROR" or "OK" as response, so push a line read 1226 # we expect "ERROR" or "OK" as response, so push a line read
500 $handle->push_read_line (sub { 1227 $handle->push_read (line => sub {
501 # if we got an "OK", we have to _prepend_ another line, 1228 # if we got an "OK", we have to _prepend_ another line,
502 # so it will be read before the second request reads its 64 bytes 1229 # so it will be read before the second request reads its 64 bytes
503 # which are already in the queue when this callback is called 1230 # which are already in the queue when this callback is called
504 # we don't do this in case we got an error 1231 # we don't do this in case we got an error
505 if ($_[1] eq "OK") { 1232 if ($_[1] eq "OK") {
506 $_[0]->unshift_read_line (sub { 1233 $_[0]->unshift_read (line => sub {
507 my $response = $_[1]; 1234 my $response = $_[1];
508 ... 1235 ...
509 }); 1236 });
510 } 1237 }
511 }); 1238 });
512 1239
513 # request two 1240 # request two, simply returns 64 octets
514 $handle->push_write ("request 2\015\012"); 1241 $handle->push_write ("request 2\015\012");
515 1242
516 # simply read 64 bytes, always 1243 # simply read 64 bytes, always
517 $handle->push_read_chunk (64, sub { 1244 $handle->push_read (chunk => 64, sub {
518 my $response = $_[1]; 1245 my $response = $_[1];
519 ... 1246 ...
520 }); 1247 });
521 1248
522=over 4 1249=over 4
523 1250
524=cut 1251=cut
525 1252
526sub _drain_rbuf { 1253sub _drain_rbuf {
527 my ($self) = @_; 1254 my ($self) = @_;
1255
1256 # avoid recursion
1257 return if $self->{_skip_drain_rbuf};
1258 local $self->{_skip_drain_rbuf} = 1;
1259
1260 while () {
1261 # we need to use a separate tls read buffer, as we must not receive data while
1262 # we are draining the buffer, and this can only happen with TLS.
1263 $self->{rbuf} .= delete $self->{_tls_rbuf}
1264 if exists $self->{_tls_rbuf};
1265
1266 my $len = length $self->{rbuf};
1267
1268 if (my $cb = shift @{ $self->{_queue} }) {
1269 unless ($cb->($self)) {
1270 # no progress can be made
1271 # (not enough data and no data forthcoming)
1272 $self->_error (Errno::EPIPE, 1), return
1273 if $self->{_eof};
1274
1275 unshift @{ $self->{_queue} }, $cb;
1276 last;
1277 }
1278 } elsif ($self->{on_read}) {
1279 last unless $len;
1280
1281 $self->{on_read}($self);
1282
1283 if (
1284 $len == length $self->{rbuf} # if no data has been consumed
1285 && !@{ $self->{_queue} } # and the queue is still empty
1286 && $self->{on_read} # but we still have on_read
1287 ) {
1288 # no further data will arrive
1289 # so no progress can be made
1290 $self->_error (Errno::EPIPE, 1), return
1291 if $self->{_eof};
1292
1293 last; # more data might arrive
1294 }
1295 } else {
1296 # read side becomes idle
1297 delete $self->{_rw} unless $self->{tls};
1298 last;
1299 }
1300 }
1301
1302 if ($self->{_eof}) {
1303 $self->{on_eof}
1304 ? $self->{on_eof}($self)
1305 : $self->_error (0, 1, "Unexpected end-of-file");
1306
1307 return;
1308 }
528 1309
529 if ( 1310 if (
530 defined $self->{rbuf_max} 1311 defined $self->{rbuf_max}
531 && $self->{rbuf_max} < length $self->{rbuf} 1312 && $self->{rbuf_max} < length $self->{rbuf}
532 ) { 1313 ) {
533 $! = &Errno::ENOSPC; 1314 $self->_error (Errno::ENOSPC, 1), return;
534 $self->error;
535 } 1315 }
536 1316
537 return if $self->{in_drain}; 1317 # may need to restart read watcher
538 local $self->{in_drain} = 1; 1318 unless ($self->{_rw}) {
539 1319 $self->start_read
540 while (my $len = length $self->{rbuf}) { 1320 if $self->{on_read} || @{ $self->{_queue} };
541 no strict 'refs';
542 if (my $cb = shift @{ $self->{_queue} }) {
543 unless ($cb->($self)) {
544 if ($self->{_eof}) {
545 # no progress can be made (not enough data and no data forthcoming)
546 $! = &Errno::EPIPE;
547 $self->error;
548 }
549
550 unshift @{ $self->{_queue} }, $cb;
551 return;
552 }
553 } elsif ($self->{on_read}) {
554 $self->{on_read}($self);
555
556 if (
557 $self->{_eof} # if no further data will arrive
558 && $len == length $self->{rbuf} # and no data has been consumed
559 && !@{ $self->{_queue} } # and the queue is still empty
560 && $self->{on_read} # and we still want to read data
561 ) {
562 # then no progress can be made
563 $! = &Errno::EPIPE;
564 $self->error;
565 }
566 } else {
567 # read side becomes idle
568 delete $self->{_rw};
569 return;
570 }
571 }
572
573 if ($self->{_eof}) {
574 $self->_shutdown;
575 $self->{on_eof}($self)
576 if $self->{on_eof};
577 } 1321 }
578} 1322}
579 1323
580=item $handle->on_read ($cb) 1324=item $handle->on_read ($cb)
581 1325
582This replaces the currently set C<on_read> callback, or clears it (when 1326This replaces the currently set C<on_read> callback, or clears it (when
583the new callback is C<undef>). See the description of C<on_read> in the 1327the new callback is C<undef>). See the description of C<on_read> in the
584constructor. 1328constructor.
585 1329
1330This method may invoke callbacks (and therefore the handle might be
1331destroyed after it returns).
1332
586=cut 1333=cut
587 1334
588sub on_read { 1335sub on_read {
589 my ($self, $cb) = @_; 1336 my ($self, $cb) = @_;
590 1337
591 $self->{on_read} = $cb; 1338 $self->{on_read} = $cb;
1339 $self->_drain_rbuf if $cb;
592} 1340}
593 1341
594=item $handle->rbuf 1342=item $handle->rbuf
595 1343
596Returns the read buffer (as a modifiable lvalue). 1344Returns the read buffer (as a modifiable lvalue). You can also access the
1345read buffer directly as the C<< ->{rbuf} >> member, if you want (this is
1346much faster, and no less clean).
597 1347
598You can access the read buffer directly as the C<< ->{rbuf} >> member, if 1348The only operation allowed on the read buffer (apart from looking at it)
599you want. 1349is removing data from its beginning. Otherwise modifying or appending to
1350it is not allowed and will lead to hard-to-track-down bugs.
600 1351
601NOTE: The read buffer should only be used or modified if the C<on_read>, 1352NOTE: The read buffer should only be used or modified in the C<on_read>
602C<push_read> or C<unshift_read> methods are used. The other read methods 1353callback or when C<push_read> or C<unshift_read> are used with a single
603automatically manage the read buffer. 1354callback (i.e. untyped). Typed C<push_read> and C<unshift_read> methods
1355will manage the read buffer on their own.
604 1356
605=cut 1357=cut
606 1358
607sub rbuf : lvalue { 1359sub rbuf : lvalue {
608 $_[0]{rbuf} 1360 $_[0]{rbuf}
625 1377
626If enough data was available, then the callback must remove all data it is 1378If enough data was available, then the callback must remove all data it is
627interested in (which can be none at all) and return a true value. After returning 1379interested in (which can be none at all) and return a true value. After returning
628true, it will be removed from the queue. 1380true, it will be removed from the queue.
629 1381
1382These methods may invoke callbacks (and therefore the handle might be
1383destroyed after it returns).
1384
630=cut 1385=cut
631 1386
632our %RH; 1387our %RH;
633 1388
634sub register_read_type($$) { 1389sub register_read_type($$) {
640 my $cb = pop; 1395 my $cb = pop;
641 1396
642 if (@_) { 1397 if (@_) {
643 my $type = shift; 1398 my $type = shift;
644 1399
1400 $cb = ($RH{$type} ||= _load_func "$type\::anyevent_read_type"
645 $cb = ($RH{$type} or Carp::croak "unsupported type passed to AnyEvent::Handle::push_read") 1401 or Carp::croak "unsupported/unloadable type '$type' passed to AnyEvent::Handle::push_read")
646 ->($self, $cb, @_); 1402 ->($self, $cb, @_);
647 } 1403 }
648 1404
649 push @{ $self->{_queue} }, $cb; 1405 push @{ $self->{_queue} }, $cb;
650 $self->_drain_rbuf; 1406 $self->_drain_rbuf;
655 my $cb = pop; 1411 my $cb = pop;
656 1412
657 if (@_) { 1413 if (@_) {
658 my $type = shift; 1414 my $type = shift;
659 1415
1416 $cb = ($RH{$type} ||= _load_func "$type\::anyevent_read_type"
660 $cb = ($RH{$type} or Carp::croak "unsupported type passed to AnyEvent::Handle::unshift_read") 1417 or Carp::croak "unsupported/unloadable type '$type' passed to AnyEvent::Handle::unshift_read")
661 ->($self, $cb, @_); 1418 ->($self, $cb, @_);
662 } 1419 }
663
664 1420
665 unshift @{ $self->{_queue} }, $cb; 1421 unshift @{ $self->{_queue} }, $cb;
666 $self->_drain_rbuf; 1422 $self->_drain_rbuf;
667} 1423}
668 1424
670 1426
671=item $handle->unshift_read (type => @args, $cb) 1427=item $handle->unshift_read (type => @args, $cb)
672 1428
673Instead of providing a callback that parses the data itself you can chose 1429Instead of providing a callback that parses the data itself you can chose
674between a number of predefined parsing formats, for chunks of data, lines 1430between a number of predefined parsing formats, for chunks of data, lines
675etc. 1431etc. You can also specify the (fully qualified) name of a package, in
1432which case AnyEvent tries to load the package and then expects to find the
1433C<anyevent_read_type> function inside (see "custom read types", below).
676 1434
677Predefined types are (if you have ideas for additional types, feel free to 1435Predefined types are (if you have ideas for additional types, feel free to
678drop by and tell us): 1436drop by and tell us):
679 1437
680=over 4 1438=over 4
686data. 1444data.
687 1445
688Example: read 2 bytes. 1446Example: read 2 bytes.
689 1447
690 $handle->push_read (chunk => 2, sub { 1448 $handle->push_read (chunk => 2, sub {
691 warn "yay ", unpack "H*", $_[1]; 1449 say "yay " . unpack "H*", $_[1];
692 }); 1450 });
693 1451
694=cut 1452=cut
695 1453
696register_read_type chunk => sub { 1454register_read_type chunk => sub {
701 $cb->($_[0], substr $_[0]{rbuf}, 0, $len, ""); 1459 $cb->($_[0], substr $_[0]{rbuf}, 0, $len, "");
702 1 1460 1
703 } 1461 }
704}; 1462};
705 1463
706# compatibility with older API
707sub push_read_chunk {
708 $_[0]->push_read (chunk => $_[1], $_[2]);
709}
710
711sub unshift_read_chunk {
712 $_[0]->unshift_read (chunk => $_[1], $_[2]);
713}
714
715=item line => [$eol, ]$cb->($handle, $line, $eol) 1464=item line => [$eol, ]$cb->($handle, $line, $eol)
716 1465
717The callback will be called only once a full line (including the end of 1466The callback will be called only once a full line (including the end of
718line marker, C<$eol>) has been read. This line (excluding the end of line 1467line marker, C<$eol>) has been read. This line (excluding the end of line
719marker) will be passed to the callback as second argument (C<$line>), and 1468marker) will be passed to the callback as second argument (C<$line>), and
734=cut 1483=cut
735 1484
736register_read_type line => sub { 1485register_read_type line => sub {
737 my ($self, $cb, $eol) = @_; 1486 my ($self, $cb, $eol) = @_;
738 1487
739 $eol = qr|(\015?\012)| if @_ < 3; 1488 if (@_ < 3) {
740 $eol = quotemeta $eol unless ref $eol; 1489 # this is more than twice as fast as the generic code below
741 $eol = qr|^(.*?)($eol)|s;
742
743 sub { 1490 sub {
744 $_[0]{rbuf} =~ s/$eol// or return; 1491 $_[0]{rbuf} =~ s/^([^\015\012]*)(\015?\012)// or return;
745 1492
746 $cb->($_[0], $1, $2); 1493 $cb->($_[0], "$1", "$2");
747 1
748 }
749};
750
751# compatibility with older API
752sub push_read_line {
753 my $self = shift;
754 $self->push_read (line => @_);
755}
756
757sub unshift_read_line {
758 my $self = shift;
759 $self->unshift_read (line => @_);
760}
761
762=item netstring => $cb->($handle, $string)
763
764A netstring (http://cr.yp.to/proto/netstrings.txt, this is not an endorsement).
765
766Throws an error with C<$!> set to EBADMSG on format violations.
767
768=cut
769
770register_read_type netstring => sub {
771 my ($self, $cb) = @_;
772
773 sub {
774 unless ($_[0]{rbuf} =~ s/^(0|[1-9][0-9]*)://) {
775 if ($_[0]{rbuf} =~ /[^0-9]/) {
776 $! = &Errno::EBADMSG;
777 $self->error;
778 } 1494 1
779 return;
780 } 1495 }
1496 } else {
1497 $eol = quotemeta $eol unless ref $eol;
1498 $eol = qr|^(.*?)($eol)|s;
781 1499
782 my $len = $1; 1500 sub {
1501 $_[0]{rbuf} =~ s/$eol// or return;
783 1502
784 $self->unshift_read (chunk => $len, sub { 1503 $cb->($_[0], "$1", "$2");
785 my $string = $_[1];
786 $_[0]->unshift_read (chunk => 1, sub {
787 if ($_[1] eq ",") {
788 $cb->($_[0], $string);
789 } else {
790 $! = &Errno::EBADMSG;
791 $self->error;
792 }
793 }); 1504 1
794 }); 1505 }
795
796 1
797 } 1506 }
798}; 1507};
799 1508
800=item regex => $accept[, $reject[, $skip], $cb->($handle, $data) 1509=item regex => $accept[, $reject[, $skip], $cb->($handle, $data)
801 1510
821the receive buffer when neither C<$accept> nor C<$reject> match, 1530the receive buffer when neither C<$accept> nor C<$reject> match,
822and everything preceding and including the match will be accepted 1531and everything preceding and including the match will be accepted
823unconditionally. This is useful to skip large amounts of data that you 1532unconditionally. This is useful to skip large amounts of data that you
824know cannot be matched, so that the C<$accept> or C<$reject> regex do not 1533know cannot be matched, so that the C<$accept> or C<$reject> regex do not
825have to start matching from the beginning. This is purely an optimisation 1534have to start matching from the beginning. This is purely an optimisation
826and is usually worth only when you expect more than a few kilobytes. 1535and is usually worth it only when you expect more than a few kilobytes.
827 1536
828Example: expect a http header, which ends at C<\015\012\015\012>. Since we 1537Example: expect a http header, which ends at C<\015\012\015\012>. Since we
829expect the header to be very large (it isn't in practise, but...), we use 1538expect the header to be very large (it isn't in practice, but...), we use
830a skip regex to skip initial portions. The skip regex is tricky in that 1539a skip regex to skip initial portions. The skip regex is tricky in that
831it only accepts something not ending in either \015 or \012, as these are 1540it only accepts something not ending in either \015 or \012, as these are
832required for the accept regex. 1541required for the accept regex.
833 1542
834 $handle->push_read (regex => 1543 $handle->push_read (regex =>
847 1556
848 sub { 1557 sub {
849 # accept 1558 # accept
850 if ($$rbuf =~ $accept) { 1559 if ($$rbuf =~ $accept) {
851 $data .= substr $$rbuf, 0, $+[0], ""; 1560 $data .= substr $$rbuf, 0, $+[0], "";
852 $cb->($self, $data); 1561 $cb->($_[0], $data);
853 return 1; 1562 return 1;
854 } 1563 }
855 1564
856 # reject 1565 # reject
857 if ($reject && $$rbuf =~ $reject) { 1566 if ($reject && $$rbuf =~ $reject) {
858 $! = &Errno::EBADMSG; 1567 $_[0]->_error (Errno::EBADMSG);
859 $self->error;
860 } 1568 }
861 1569
862 # skip 1570 # skip
863 if ($skip && $$rbuf =~ $skip) { 1571 if ($skip && $$rbuf =~ $skip) {
864 $data .= substr $$rbuf, 0, $+[0], ""; 1572 $data .= substr $$rbuf, 0, $+[0], "";
866 1574
867 () 1575 ()
868 } 1576 }
869}; 1577};
870 1578
1579=item netstring => $cb->($handle, $string)
1580
1581A netstring (http://cr.yp.to/proto/netstrings.txt, this is not an endorsement).
1582
1583Throws an error with C<$!> set to EBADMSG on format violations.
1584
1585=cut
1586
1587register_read_type netstring => sub {
1588 my ($self, $cb) = @_;
1589
1590 sub {
1591 unless ($_[0]{rbuf} =~ s/^(0|[1-9][0-9]*)://) {
1592 if ($_[0]{rbuf} =~ /[^0-9]/) {
1593 $_[0]->_error (Errno::EBADMSG);
1594 }
1595 return;
1596 }
1597
1598 my $len = $1;
1599
1600 $_[0]->unshift_read (chunk => $len, sub {
1601 my $string = $_[1];
1602 $_[0]->unshift_read (chunk => 1, sub {
1603 if ($_[1] eq ",") {
1604 $cb->($_[0], $string);
1605 } else {
1606 $_[0]->_error (Errno::EBADMSG);
1607 }
1608 });
1609 });
1610
1611 1
1612 }
1613};
1614
1615=item packstring => $format, $cb->($handle, $string)
1616
1617An octet string prefixed with an encoded length. The encoding C<$format>
1618uses the same format as a Perl C<pack> format, but must specify a single
1619integer only (only one of C<cCsSlLqQiInNvVjJw> is allowed, plus an
1620optional C<!>, C<< < >> or C<< > >> modifier).
1621
1622For example, DNS over TCP uses a prefix of C<n> (2 octet network order),
1623EPP uses a prefix of C<N> (4 octtes).
1624
1625Example: read a block of data prefixed by its length in BER-encoded
1626format (very efficient).
1627
1628 $handle->push_read (packstring => "w", sub {
1629 my ($handle, $data) = @_;
1630 });
1631
1632=cut
1633
1634register_read_type packstring => sub {
1635 my ($self, $cb, $format) = @_;
1636
1637 sub {
1638 # when we can use 5.10 we can use ".", but for 5.8 we use the re-pack method
1639 defined (my $len = eval { unpack $format, $_[0]{rbuf} })
1640 or return;
1641
1642 $format = length pack $format, $len;
1643
1644 # bypass unshift if we already have the remaining chunk
1645 if ($format + $len <= length $_[0]{rbuf}) {
1646 my $data = substr $_[0]{rbuf}, $format, $len;
1647 substr $_[0]{rbuf}, 0, $format + $len, "";
1648 $cb->($_[0], $data);
1649 } else {
1650 # remove prefix
1651 substr $_[0]{rbuf}, 0, $format, "";
1652
1653 # read remaining chunk
1654 $_[0]->unshift_read (chunk => $len, $cb);
1655 }
1656
1657 1
1658 }
1659};
1660
871=item json => $cb->($handle, $hash_or_arrayref) 1661=item json => $cb->($handle, $hash_or_arrayref)
872 1662
873Reads a JSON object or array, decodes it and passes it to the callback. 1663Reads a JSON object or array, decodes it and passes it to the
1664callback. When a parse error occurs, an C<EBADMSG> error will be raised.
874 1665
875If a C<json> object was passed to the constructor, then that will be used 1666If a C<json> object was passed to the constructor, then that will be used
876for the final decode, otherwise it will create a JSON coder expecting UTF-8. 1667for the final decode, otherwise it will create a JSON coder expecting UTF-8.
877 1668
878This read type uses the incremental parser available with JSON version 1669This read type uses the incremental parser available with JSON version
885the C<json> write type description, above, for an actual example. 1676the C<json> write type description, above, for an actual example.
886 1677
887=cut 1678=cut
888 1679
889register_read_type json => sub { 1680register_read_type json => sub {
890 my ($self, $cb, $accept, $reject, $skip) = @_; 1681 my ($self, $cb) = @_;
891 1682
892 require JSON; 1683 my $json = $self->{json} ||= json_coder;
893 1684
894 my $data; 1685 my $data;
895 my $rbuf = \$self->{rbuf}; 1686 my $rbuf = \$self->{rbuf};
896 1687
897 my $json = $self->{json} ||= JSON->new->utf8;
898
899 sub { 1688 sub {
900 my $ref = $json->incr_parse ($self->{rbuf}); 1689 my $ref = eval { $json->incr_parse ($_[0]{rbuf}) };
901 1690
902 if ($ref) { 1691 if ($ref) {
903 $self->{rbuf} = $json->incr_text; 1692 $_[0]{rbuf} = $json->incr_text;
904 $json->incr_text = ""; 1693 $json->incr_text = "";
905 $cb->($self, $ref); 1694 $cb->($_[0], $ref);
906 1695
907 1 1696 1
1697 } elsif ($@) {
1698 # error case
1699 $json->incr_skip;
1700
1701 $_[0]{rbuf} = $json->incr_text;
1702 $json->incr_text = "";
1703
1704 $_[0]->_error (Errno::EBADMSG);
1705
1706 ()
908 } else { 1707 } else {
909 $self->{rbuf} = ""; 1708 $_[0]{rbuf} = "";
1709
910 () 1710 ()
911 } 1711 }
912 } 1712 }
913}; 1713};
914 1714
1715=item storable => $cb->($handle, $ref)
1716
1717Deserialises a L<Storable> frozen representation as written by the
1718C<storable> write type (BER-encoded length prefix followed by nfreeze'd
1719data).
1720
1721Raises C<EBADMSG> error if the data could not be decoded.
1722
1723=cut
1724
1725register_read_type storable => sub {
1726 my ($self, $cb) = @_;
1727
1728 require Storable unless $Storable::VERSION;
1729
1730 sub {
1731 # when we can use 5.10 we can use ".", but for 5.8 we use the re-pack method
1732 defined (my $len = eval { unpack "w", $_[0]{rbuf} })
1733 or return;
1734
1735 my $format = length pack "w", $len;
1736
1737 # bypass unshift if we already have the remaining chunk
1738 if ($format + $len <= length $_[0]{rbuf}) {
1739 my $data = substr $_[0]{rbuf}, $format, $len;
1740 substr $_[0]{rbuf}, 0, $format + $len, "";
1741
1742 eval { $cb->($_[0], Storable::thaw ($data)); 1 }
1743 or return $_[0]->_error (Errno::EBADMSG);
1744 } else {
1745 # remove prefix
1746 substr $_[0]{rbuf}, 0, $format, "";
1747
1748 # read remaining chunk
1749 $_[0]->unshift_read (chunk => $len, sub {
1750 eval { $cb->($_[0], Storable::thaw ($_[1])); 1 }
1751 or $_[0]->_error (Errno::EBADMSG);
1752 });
1753 }
1754
1755 1
1756 }
1757};
1758
915=back 1759=back
916 1760
917=item AnyEvent::Handle::register_read_type type => $coderef->($handle, $cb, @args) 1761=item custom read types - Package::anyevent_read_type $handle, $cb, @args
918 1762
919This function (not method) lets you add your own types to C<push_read>. 1763Instead of one of the predefined types, you can also specify the name
1764of a package. AnyEvent will try to load the package and then expects to
1765find a function named C<anyevent_read_type> inside. If it isn't found, it
1766progressively tries to load the parent package until it either finds the
1767function (good) or runs out of packages (bad).
920 1768
921Whenever the given C<type> is used, C<push_read> will invoke the code 1769Whenever this type is used, C<push_read> will invoke the function with the
922reference with the handle object, the callback and the remaining 1770handle object, the original callback and the remaining arguments.
923arguments.
924 1771
925The code reference is supposed to return a callback (usually a closure) 1772The function is supposed to return a callback (usually a closure) that
926that works as a plain read callback (see C<< ->push_read ($cb) >>). 1773works as a plain read callback (see C<< ->push_read ($cb) >>), so you can
1774mentally treat the function as a "configurable read type to read callback"
1775converter.
927 1776
928It should invoke the passed callback when it is done reading (remember to 1777It should invoke the original callback when it is done reading (remember
929pass C<$handle> as first argument as all other callbacks do that). 1778to pass C<$handle> as first argument as all other callbacks do that,
1779although there is no strict requirement on this).
930 1780
931Note that this is a function, and all types registered this way will be
932global, so try to use unique names.
933
934For examples, see the source of this module (F<perldoc -m AnyEvent::Handle>, 1781For examples, see the source of this module (F<perldoc -m
935search for C<register_read_type>)). 1782AnyEvent::Handle>, search for C<register_read_type>)).
936 1783
937=item $handle->stop_read 1784=item $handle->stop_read
938 1785
939=item $handle->start_read 1786=item $handle->start_read
940 1787
941In rare cases you actually do not want to read anything from the 1788In rare cases you actually do not want to read anything from the
942socket. In this case you can call C<stop_read>. Neither C<on_read> no 1789socket. In this case you can call C<stop_read>. Neither C<on_read> nor
943any queued callbacks will be executed then. To start reading again, call 1790any queued callbacks will be executed then. To start reading again, call
944C<start_read>. 1791C<start_read>.
945 1792
1793Note that AnyEvent::Handle will automatically C<start_read> for you when
1794you change the C<on_read> callback or push/unshift a read callback, and it
1795will automatically C<stop_read> for you when neither C<on_read> is set nor
1796there are any read requests in the queue.
1797
1798In older versions of this module (<= 5.3), these methods had no effect,
1799as TLS does not support half-duplex connections. In current versions they
1800work as expected, as this behaviour is required to avoid certain resource
1801attacks, where the program would be forced to read (and buffer) arbitrary
1802amounts of data before being able to send some data. The drawback is that
1803some readings of the the SSL/TLS specifications basically require this
1804attack to be working, as SSL/TLS implementations might stall sending data
1805during a rehandshake.
1806
1807As a guideline, during the initial handshake, you should not stop reading,
1808and as a client, it might cause problems, depending on your application.
1809
946=cut 1810=cut
947 1811
948sub stop_read { 1812sub stop_read {
949 my ($self) = @_; 1813 my ($self) = @_;
950 1814
952} 1816}
953 1817
954sub start_read { 1818sub start_read {
955 my ($self) = @_; 1819 my ($self) = @_;
956 1820
957 unless ($self->{_rw} || $self->{_eof}) { 1821 unless ($self->{_rw} || $self->{_eof} || !$self->{fh}) {
958 Scalar::Util::weaken $self; 1822 Scalar::Util::weaken $self;
959 1823
960 $self->{_rw} = AnyEvent->io (fh => $self->{fh}, poll => "r", cb => sub { 1824 $self->{_rw} = AE::io $self->{fh}, 0, sub {
961 my $rbuf = $self->{filter_r} ? \my $buf : \$self->{rbuf}; 1825 my $rbuf = \($self->{tls} ? my $buf : $self->{rbuf});
962 my $len = sysread $self->{fh}, $$rbuf, $self->{read_size} || 8192, length $$rbuf; 1826 my $len = sysread $self->{fh}, $$rbuf, $self->{read_size}, length $$rbuf;
963 1827
964 if ($len > 0) { 1828 if ($len > 0) {
965 $self->{filter_r} 1829 $self->{_activity} = $self->{_ractivity} = AE::now;
966 ? $self->{filter_r}->($self, $rbuf) 1830
1831 if ($self->{tls}) {
1832 Net::SSLeay::BIO_write ($self->{_rbio}, $$rbuf);
1833
1834 &_dotls ($self);
1835 } else {
967 : $self->_drain_rbuf; 1836 $self->_drain_rbuf;
1837 }
1838
1839 if ($len == $self->{read_size}) {
1840 $self->{read_size} *= 2;
1841 $self->{read_size} = $self->{max_read_size} || MAX_READ_SIZE
1842 if $self->{read_size} > ($self->{max_read_size} || MAX_READ_SIZE);
1843 }
968 1844
969 } elsif (defined $len) { 1845 } elsif (defined $len) {
970 delete $self->{_rw}; 1846 delete $self->{_rw};
971 $self->{_eof} = 1; 1847 $self->{_eof} = 1;
972 $self->_drain_rbuf; 1848 $self->_drain_rbuf;
973 1849
974 } elsif ($! != EAGAIN && $! != EINTR && $! != WSAEWOULDBLOCK) { 1850 } elsif ($! != EAGAIN && $! != EINTR && $! != WSAEWOULDBLOCK) {
975 return $self->error; 1851 return $self->_error ($!, 1);
976 } 1852 }
977 }); 1853 };
978 } 1854 }
979} 1855}
980 1856
1857our $ERROR_SYSCALL;
1858our $ERROR_WANT_READ;
1859
1860sub _tls_error {
1861 my ($self, $err) = @_;
1862
1863 return $self->_error ($!, 1)
1864 if $err == Net::SSLeay::ERROR_SYSCALL ();
1865
1866 my $err = Net::SSLeay::ERR_error_string (Net::SSLeay::ERR_get_error ());
1867
1868 # reduce error string to look less scary
1869 $err =~ s/^error:[0-9a-fA-F]{8}:[^:]+:([^:]+):/\L$1: /;
1870
1871 if ($self->{_on_starttls}) {
1872 (delete $self->{_on_starttls})->($self, undef, $err);
1873 &_freetls;
1874 } else {
1875 &_freetls;
1876 $self->_error (Errno::EPROTO, 1, $err);
1877 }
1878}
1879
1880# poll the write BIO and send the data if applicable
1881# also decode read data if possible
1882# this is basiclaly our TLS state machine
1883# more efficient implementations are possible with openssl,
1884# but not with the buggy and incomplete Net::SSLeay.
981sub _dotls { 1885sub _dotls {
982 my ($self) = @_; 1886 my ($self) = @_;
983 1887
1888 my $tmp;
1889
984 if (length $self->{_tls_wbuf}) { 1890 if (length $self->{_tls_wbuf}) {
985 while ((my $len = Net::SSLeay::write ($self->{tls}, $self->{_tls_wbuf})) > 0) { 1891 while (($tmp = Net::SSLeay::write ($self->{tls}, $self->{_tls_wbuf})) > 0) {
986 substr $self->{_tls_wbuf}, 0, $len, ""; 1892 substr $self->{_tls_wbuf}, 0, $tmp, "";
987 } 1893 }
988 }
989 1894
1895 $tmp = Net::SSLeay::get_error ($self->{tls}, $tmp);
1896 return $self->_tls_error ($tmp)
1897 if $tmp != $ERROR_WANT_READ
1898 && ($tmp != $ERROR_SYSCALL || $!);
1899 }
1900
1901 while (defined ($tmp = Net::SSLeay::read ($self->{tls}))) {
1902 unless (length $tmp) {
1903 $self->{_on_starttls}
1904 and (delete $self->{_on_starttls})->($self, undef, "EOF during handshake"); # ???
1905 &_freetls;
1906
1907 if ($self->{on_stoptls}) {
1908 $self->{on_stoptls}($self);
1909 return;
1910 } else {
1911 # let's treat SSL-eof as we treat normal EOF
1912 delete $self->{_rw};
1913 $self->{_eof} = 1;
1914 }
1915 }
1916
1917 $self->{_tls_rbuf} .= $tmp;
1918 $self->_drain_rbuf;
1919 $self->{tls} or return; # tls session might have gone away in callback
1920 }
1921
1922 $tmp = Net::SSLeay::get_error ($self->{tls}, -1);
1923 return $self->_tls_error ($tmp)
1924 if $tmp != $ERROR_WANT_READ
1925 && ($tmp != $ERROR_SYSCALL || $!);
1926
990 if (defined (my $buf = Net::SSLeay::BIO_read ($self->{_wbio}))) { 1927 while (length ($tmp = Net::SSLeay::BIO_read ($self->{_wbio}))) {
991 $self->{wbuf} .= $buf; 1928 $self->{wbuf} .= $tmp;
992 $self->_drain_wbuf; 1929 $self->_drain_wbuf;
1930 $self->{tls} or return; # tls session might have gone away in callback
993 } 1931 }
994 1932
995 while (defined (my $buf = Net::SSLeay::read ($self->{tls}))) { 1933 $self->{_on_starttls}
996 $self->{rbuf} .= $buf; 1934 and Net::SSLeay::state ($self->{tls}) == Net::SSLeay::ST_OK ()
997 $self->_drain_rbuf; 1935 and (delete $self->{_on_starttls})->($self, 1, "TLS/SSL connection established");
998 }
999
1000 my $err = Net::SSLeay::get_error ($self->{tls}, -1);
1001
1002 if ($err!= Net::SSLeay::ERROR_WANT_READ ()) {
1003 if ($err == Net::SSLeay::ERROR_SYSCALL ()) {
1004 $self->error;
1005 } elsif ($err == Net::SSLeay::ERROR_SSL ()) {
1006 $! = &Errno::EIO;
1007 $self->error;
1008 }
1009
1010 # all others are fine for our purposes
1011 }
1012} 1936}
1013 1937
1014=item $handle->starttls ($tls[, $tls_ctx]) 1938=item $handle->starttls ($tls[, $tls_ctx])
1015 1939
1016Instead of starting TLS negotiation immediately when the AnyEvent::Handle 1940Instead of starting TLS negotiation immediately when the AnyEvent::Handle
1017object is created, you can also do that at a later time by calling 1941object is created, you can also do that at a later time by calling
1018C<starttls>. 1942C<starttls>.
1019 1943
1944Starting TLS is currently an asynchronous operation - when you push some
1945write data and then call C<< ->starttls >> then TLS negotiation will start
1946immediately, after which the queued write data is then sent.
1947
1020The first argument is the same as the C<tls> constructor argument (either 1948The first argument is the same as the C<tls> constructor argument (either
1021C<"connect">, C<"accept"> or an existing Net::SSLeay object). 1949C<"connect">, C<"accept"> or an existing Net::SSLeay object).
1022 1950
1023The second argument is the optional C<Net::SSLeay::CTX> object that is 1951The second argument is the optional C<AnyEvent::TLS> object that is used
1024used when AnyEvent::Handle has to create its own TLS connection object. 1952when AnyEvent::Handle has to create its own TLS connection object, or
1953a hash reference with C<< key => value >> pairs that will be used to
1954construct a new context.
1025 1955
1026The TLS connection object will end up in C<< $handle->{tls} >> after this 1956The TLS connection object will end up in C<< $handle->{tls} >>, the TLS
1027call and can be used or changed to your liking. Note that the handshake 1957context in C<< $handle->{tls_ctx} >> after this call and can be used or
1028might have already started when this function returns. 1958changed to your liking. Note that the handshake might have already started
1959when this function returns.
1029 1960
1030=cut 1961Due to bugs in OpenSSL, it might or might not be possible to do multiple
1962handshakes on the same stream. It is best to not attempt to use the
1963stream after stopping TLS.
1031 1964
1032# TODO: maybe document... 1965This method may invoke callbacks (and therefore the handle might be
1966destroyed after it returns).
1967
1968=cut
1969
1970our %TLS_CACHE; #TODO not yet documented, should we?
1971
1033sub starttls { 1972sub starttls {
1034 my ($self, $ssl, $ctx) = @_; 1973 my ($self, $tls, $ctx) = @_;
1035 1974
1036 $self->stoptls; 1975 Carp::croak "It is an error to call starttls on an AnyEvent::Handle object while TLS is already active, caught"
1976 if $self->{tls};
1037 1977
1038 if ($ssl eq "accept") { 1978 $self->{tls} = $tls;
1039 $ssl = Net::SSLeay::new ($ctx || TLS_CTX ()); 1979 $self->{tls_ctx} = $ctx if @_ > 2;
1040 Net::SSLeay::set_accept_state ($ssl); 1980
1041 } elsif ($ssl eq "connect") { 1981 return unless $self->{fh};
1042 $ssl = Net::SSLeay::new ($ctx || TLS_CTX ()); 1982
1043 Net::SSLeay::set_connect_state ($ssl); 1983 require Net::SSLeay;
1984
1985 $ERROR_SYSCALL = Net::SSLeay::ERROR_SYSCALL ();
1986 $ERROR_WANT_READ = Net::SSLeay::ERROR_WANT_READ ();
1987
1988 $tls = delete $self->{tls};
1989 $ctx = $self->{tls_ctx};
1990
1991 local $Carp::CarpLevel = 1; # skip ourselves when creating a new context or session
1992
1993 if ("HASH" eq ref $ctx) {
1994 require AnyEvent::TLS;
1995
1996 if ($ctx->{cache}) {
1997 my $key = $ctx+0;
1998 $ctx = $TLS_CACHE{$key} ||= new AnyEvent::TLS %$ctx;
1999 } else {
2000 $ctx = new AnyEvent::TLS %$ctx;
2001 }
2002 }
1044 } 2003
1045 2004 $self->{tls_ctx} = $ctx || TLS_CTX ();
1046 $self->{tls} = $ssl; 2005 $self->{tls} = $tls = $self->{tls_ctx}->_get_session ($tls, $self, $self->{peername});
1047 2006
1048 # basically, this is deep magic (because SSL_read should have the same issues) 2007 # basically, this is deep magic (because SSL_read should have the same issues)
1049 # but the openssl maintainers basically said: "trust us, it just works". 2008 # but the openssl maintainers basically said: "trust us, it just works".
1050 # (unfortunately, we have to hardcode constants because the abysmally misdesigned 2009 # (unfortunately, we have to hardcode constants because the abysmally misdesigned
1051 # and mismaintained ssleay-module doesn't even offer them). 2010 # and mismaintained ssleay-module doesn't even offer them).
1052 # http://www.mail-archive.com/openssl-dev@openssl.org/msg22420.html 2011 # http://www.mail-archive.com/openssl-dev@openssl.org/msg22420.html
2012 #
2013 # in short: this is a mess.
2014 #
2015 # note that we do not try to keep the length constant between writes as we are required to do.
2016 # we assume that most (but not all) of this insanity only applies to non-blocking cases,
2017 # and we drive openssl fully in blocking mode here. Or maybe we don't - openssl seems to
2018 # have identity issues in that area.
1053 Net::SSLeay::CTX_set_mode ($self->{tls}, 2019# Net::SSLeay::CTX_set_mode ($ssl,
1054 (eval { local $SIG{__DIE__}; Net::SSLeay::MODE_ENABLE_PARTIAL_WRITE () } || 1) 2020# (eval { local $SIG{__DIE__}; Net::SSLeay::MODE_ENABLE_PARTIAL_WRITE () } || 1)
1055 | (eval { local $SIG{__DIE__}; Net::SSLeay::MODE_ACCEPT_MOVING_WRITE_BUFFER () } || 2)); 2021# | (eval { local $SIG{__DIE__}; Net::SSLeay::MODE_ACCEPT_MOVING_WRITE_BUFFER () } || 2));
2022 Net::SSLeay::CTX_set_mode ($tls, 1|2);
1056 2023
1057 $self->{_rbio} = Net::SSLeay::BIO_new (Net::SSLeay::BIO_s_mem ()); 2024 $self->{_rbio} = Net::SSLeay::BIO_new (Net::SSLeay::BIO_s_mem ());
1058 $self->{_wbio} = Net::SSLeay::BIO_new (Net::SSLeay::BIO_s_mem ()); 2025 $self->{_wbio} = Net::SSLeay::BIO_new (Net::SSLeay::BIO_s_mem ());
1059 2026
2027 Net::SSLeay::BIO_write ($self->{_rbio}, $self->{rbuf});
2028 $self->{rbuf} = "";
2029
1060 Net::SSLeay::set_bio ($ssl, $self->{_rbio}, $self->{_wbio}); 2030 Net::SSLeay::set_bio ($tls, $self->{_rbio}, $self->{_wbio});
1061 2031
1062 $self->{filter_w} = sub { 2032 $self->{_on_starttls} = sub { $_[0]{on_starttls}(@_) }
1063 $_[0]{_tls_wbuf} .= ${$_[1]}; 2033 if $self->{on_starttls};
1064 &_dotls; 2034
1065 }; 2035 &_dotls; # need to trigger the initial handshake
1066 $self->{filter_r} = sub { 2036 $self->start_read; # make sure we actually do read
1067 Net::SSLeay::BIO_write ($_[0]{_rbio}, ${$_[1]});
1068 &_dotls;
1069 };
1070} 2037}
1071 2038
1072=item $handle->stoptls 2039=item $handle->stoptls
1073 2040
1074Destroys the SSL connection, if any. Partial read or write data will be 2041Shuts down the SSL connection - this makes a proper EOF handshake by
1075lost. 2042sending a close notify to the other side, but since OpenSSL doesn't
2043support non-blocking shut downs, it is not guaranteed that you can re-use
2044the stream afterwards.
2045
2046This method may invoke callbacks (and therefore the handle might be
2047destroyed after it returns).
1076 2048
1077=cut 2049=cut
1078 2050
1079sub stoptls { 2051sub stoptls {
1080 my ($self) = @_; 2052 my ($self) = @_;
1081 2053
1082 Net::SSLeay::free (delete $self->{tls}) if $self->{tls}; 2054 if ($self->{tls} && $self->{fh}) {
2055 Net::SSLeay::shutdown ($self->{tls});
1083 2056
1084 delete $self->{_rbio}; 2057 &_dotls;
1085 delete $self->{_wbio}; 2058
1086 delete $self->{_tls_wbuf}; 2059# # we don't give a shit. no, we do, but we can't. no...#d#
1087 delete $self->{filter_r}; 2060# # we, we... have to use openssl :/#d#
1088 delete $self->{filter_w}; 2061# &_freetls;#d#
2062 }
1089} 2063}
2064
2065sub _freetls {
2066 my ($self) = @_;
2067
2068 return unless $self->{tls};
2069
2070 $self->{tls_ctx}->_put_session (delete $self->{tls})
2071 if $self->{tls} > 0;
2072
2073 delete @$self{qw(_rbio _wbio _tls_wbuf _on_starttls)};
2074}
2075
2076=item $handle->resettls
2077
2078This rarely-used method simply resets and TLS state on the handle, usually
2079causing data loss.
2080
2081One case where it may be useful is when you want to skip over the data in
2082the stream but you are not interested in interpreting it, so data loss is
2083no concern.
2084
2085=cut
2086
2087*resettls = \&_freetls;
1090 2088
1091sub DESTROY { 2089sub DESTROY {
1092 my $self = shift; 2090 my ($self) = @_;
1093 2091
1094 $self->stoptls; 2092 &_freetls;
2093
2094 my $linger = exists $self->{linger} ? $self->{linger} : 3600;
2095
2096 if ($linger && length $self->{wbuf} && $self->{fh}) {
2097 my $fh = delete $self->{fh};
2098 my $wbuf = delete $self->{wbuf};
2099
2100 my @linger;
2101
2102 push @linger, AE::io $fh, 1, sub {
2103 my $len = syswrite $fh, $wbuf, length $wbuf;
2104
2105 if ($len > 0) {
2106 substr $wbuf, 0, $len, "";
2107 } elsif (defined $len || ($! != EAGAIN && $! != EINTR && $! != WSAEWOULDBLOCK)) {
2108 @linger = (); # end
2109 }
2110 };
2111 push @linger, AE::timer $linger, 0, sub {
2112 @linger = ();
2113 };
2114 }
1095} 2115}
2116
2117=item $handle->destroy
2118
2119Shuts down the handle object as much as possible - this call ensures that
2120no further callbacks will be invoked and as many resources as possible
2121will be freed. Any method you will call on the handle object after
2122destroying it in this way will be silently ignored (and it will return the
2123empty list).
2124
2125Normally, you can just "forget" any references to an AnyEvent::Handle
2126object and it will simply shut down. This works in fatal error and EOF
2127callbacks, as well as code outside. It does I<NOT> work in a read or write
2128callback, so when you want to destroy the AnyEvent::Handle object from
2129within such an callback. You I<MUST> call C<< ->destroy >> explicitly in
2130that case.
2131
2132Destroying the handle object in this way has the advantage that callbacks
2133will be removed as well, so if those are the only reference holders (as
2134is common), then one doesn't need to do anything special to break any
2135reference cycles.
2136
2137The handle might still linger in the background and write out remaining
2138data, as specified by the C<linger> option, however.
2139
2140=cut
2141
2142sub destroy {
2143 my ($self) = @_;
2144
2145 $self->DESTROY;
2146 %$self = ();
2147 bless $self, "AnyEvent::Handle::destroyed";
2148}
2149
2150sub AnyEvent::Handle::destroyed::AUTOLOAD {
2151 #nop
2152}
2153
2154=item $handle->destroyed
2155
2156Returns false as long as the handle hasn't been destroyed by a call to C<<
2157->destroy >>, true otherwise.
2158
2159Can be useful to decide whether the handle is still valid after some
2160callback possibly destroyed the handle. For example, C<< ->push_write >>,
2161C<< ->starttls >> and other methods can call user callbacks, which in turn
2162can destroy the handle, so work can be avoided by checking sometimes:
2163
2164 $hdl->starttls ("accept");
2165 return if $hdl->destroyed;
2166 $hdl->push_write (...
2167
2168Note that the call to C<push_write> will silently be ignored if the handle
2169has been destroyed, so often you can just ignore the possibility of the
2170handle being destroyed.
2171
2172=cut
2173
2174sub destroyed { 0 }
2175sub AnyEvent::Handle::destroyed::destroyed { 1 }
1096 2176
1097=item AnyEvent::Handle::TLS_CTX 2177=item AnyEvent::Handle::TLS_CTX
1098 2178
1099This function creates and returns the Net::SSLeay::CTX object used by 2179This function creates and returns the AnyEvent::TLS object used by default
1100default for TLS mode. 2180for TLS mode.
1101 2181
1102The context is created like this: 2182The context is created by calling L<AnyEvent::TLS> without any arguments.
1103
1104 Net::SSLeay::load_error_strings;
1105 Net::SSLeay::SSLeay_add_ssl_algorithms;
1106 Net::SSLeay::randomize;
1107
1108 my $CTX = Net::SSLeay::CTX_new;
1109
1110 Net::SSLeay::CTX_set_options $CTX, Net::SSLeay::OP_ALL
1111 2183
1112=cut 2184=cut
1113 2185
1114our $TLS_CTX; 2186our $TLS_CTX;
1115 2187
1116sub TLS_CTX() { 2188sub TLS_CTX() {
1117 $TLS_CTX || do { 2189 $TLS_CTX ||= do {
1118 require Net::SSLeay; 2190 require AnyEvent::TLS;
1119 2191
1120 Net::SSLeay::load_error_strings (); 2192 new AnyEvent::TLS
1121 Net::SSLeay::SSLeay_add_ssl_algorithms (); 2193 }
1122 Net::SSLeay::randomize (); 2194}
1123 2195
1124 $TLS_CTX = Net::SSLeay::CTX_new (); 2196=back
1125 2197
1126 Net::SSLeay::CTX_set_options ($TLS_CTX, Net::SSLeay::OP_ALL ());
1127 2198
1128 $TLS_CTX 2199=head1 NONFREQUENTLY ASKED QUESTIONS
2200
2201=over 4
2202
2203=item I C<undef> the AnyEvent::Handle reference inside my callback and
2204still get further invocations!
2205
2206That's because AnyEvent::Handle keeps a reference to itself when handling
2207read or write callbacks.
2208
2209It is only safe to "forget" the reference inside EOF or error callbacks,
2210from within all other callbacks, you need to explicitly call the C<<
2211->destroy >> method.
2212
2213=item Why is my C<on_eof> callback never called?
2214
2215Probably because your C<on_error> callback is being called instead: When
2216you have outstanding requests in your read queue, then an EOF is
2217considered an error as you clearly expected some data.
2218
2219To avoid this, make sure you have an empty read queue whenever your handle
2220is supposed to be "idle" (i.e. connection closes are O.K.). You can set
2221an C<on_read> handler that simply pushes the first read requests in the
2222queue.
2223
2224See also the next question, which explains this in a bit more detail.
2225
2226=item How can I serve requests in a loop?
2227
2228Most protocols consist of some setup phase (authentication for example)
2229followed by a request handling phase, where the server waits for requests
2230and handles them, in a loop.
2231
2232There are two important variants: The first (traditional, better) variant
2233handles requests until the server gets some QUIT command, causing it to
2234close the connection first (highly desirable for a busy TCP server). A
2235client dropping the connection is an error, which means this variant can
2236detect an unexpected detection close.
2237
2238To handle this case, always make sure you have a on-empty read queue, by
2239pushing the "read request start" handler on it:
2240
2241 # we assume a request starts with a single line
2242 my @start_request; @start_request = (line => sub {
2243 my ($hdl, $line) = @_;
2244
2245 ... handle request
2246
2247 # push next request read, possibly from a nested callback
2248 $hdl->push_read (@start_request);
2249 });
2250
2251 # auth done, now go into request handling loop
2252 # now push the first @start_request
2253 $hdl->push_read (@start_request);
2254
2255By always having an outstanding C<push_read>, the handle always expects
2256some data and raises the C<EPIPE> error when the connction is dropped
2257unexpectedly.
2258
2259The second variant is a protocol where the client can drop the connection
2260at any time. For TCP, this means that the server machine may run out of
2261sockets easier, and in general, it means you cannot distinguish a protocl
2262failure/client crash from a normal connection close. Nevertheless, these
2263kinds of protocols are common (and sometimes even the best solution to the
2264problem).
2265
2266Having an outstanding read request at all times is possible if you ignore
2267C<EPIPE> errors, but this doesn't help with when the client drops the
2268connection during a request, which would still be an error.
2269
2270A better solution is to push the initial request read in an C<on_read>
2271callback. This avoids an error, as when the server doesn't expect data
2272(i.e. is idly waiting for the next request, an EOF will not raise an
2273error, but simply result in an C<on_eof> callback. It is also a bit slower
2274and simpler:
2275
2276 # auth done, now go into request handling loop
2277 $hdl->on_read (sub {
2278 my ($hdl) = @_;
2279
2280 # called each time we receive data but the read queue is empty
2281 # simply start read the request
2282
2283 $hdl->push_read (line => sub {
2284 my ($hdl, $line) = @_;
2285
2286 ... handle request
2287
2288 # do nothing special when the request has been handled, just
2289 # let the request queue go empty.
2290 });
2291 });
2292
2293=item I get different callback invocations in TLS mode/Why can't I pause
2294reading?
2295
2296Unlike, say, TCP, TLS connections do not consist of two independent
2297communication channels, one for each direction. Or put differently, the
2298read and write directions are not independent of each other: you cannot
2299write data unless you are also prepared to read, and vice versa.
2300
2301This means that, in TLS mode, you might get C<on_error> or C<on_eof>
2302callback invocations when you are not expecting any read data - the reason
2303is that AnyEvent::Handle always reads in TLS mode.
2304
2305During the connection, you have to make sure that you always have a
2306non-empty read-queue, or an C<on_read> watcher. At the end of the
2307connection (or when you no longer want to use it) you can call the
2308C<destroy> method.
2309
2310=item How do I read data until the other side closes the connection?
2311
2312If you just want to read your data into a perl scalar, the easiest way
2313to achieve this is by setting an C<on_read> callback that does nothing,
2314clearing the C<on_eof> callback and in the C<on_error> callback, the data
2315will be in C<$_[0]{rbuf}>:
2316
2317 $handle->on_read (sub { });
2318 $handle->on_eof (undef);
2319 $handle->on_error (sub {
2320 my $data = delete $_[0]{rbuf};
2321 });
2322
2323Note that this example removes the C<rbuf> member from the handle object,
2324which is not normally allowed by the API. It is expressly permitted in
2325this case only, as the handle object needs to be destroyed afterwards.
2326
2327The reason to use C<on_error> is that TCP connections, due to latencies
2328and packets loss, might get closed quite violently with an error, when in
2329fact all data has been received.
2330
2331It is usually better to use acknowledgements when transferring data,
2332to make sure the other side hasn't just died and you got the data
2333intact. This is also one reason why so many internet protocols have an
2334explicit QUIT command.
2335
2336=item I don't want to destroy the handle too early - how do I wait until
2337all data has been written?
2338
2339After writing your last bits of data, set the C<on_drain> callback
2340and destroy the handle in there - with the default setting of
2341C<low_water_mark> this will be called precisely when all data has been
2342written to the socket:
2343
2344 $handle->push_write (...);
2345 $handle->on_drain (sub {
2346 AE::log debug => "All data submitted to the kernel.";
2347 undef $handle;
2348 });
2349
2350If you just want to queue some data and then signal EOF to the other side,
2351consider using C<< ->push_shutdown >> instead.
2352
2353=item I want to contact a TLS/SSL server, I don't care about security.
2354
2355If your TLS server is a pure TLS server (e.g. HTTPS) that only speaks TLS,
2356connect to it and then create the AnyEvent::Handle with the C<tls>
2357parameter:
2358
2359 tcp_connect $host, $port, sub {
2360 my ($fh) = @_;
2361
2362 my $handle = new AnyEvent::Handle
2363 fh => $fh,
2364 tls => "connect",
2365 on_error => sub { ... };
2366
2367 $handle->push_write (...);
1129 } 2368 };
1130} 2369
2370=item I want to contact a TLS/SSL server, I do care about security.
2371
2372Then you should additionally enable certificate verification, including
2373peername verification, if the protocol you use supports it (see
2374L<AnyEvent::TLS>, C<verify_peername>).
2375
2376E.g. for HTTPS:
2377
2378 tcp_connect $host, $port, sub {
2379 my ($fh) = @_;
2380
2381 my $handle = new AnyEvent::Handle
2382 fh => $fh,
2383 peername => $host,
2384 tls => "connect",
2385 tls_ctx => { verify => 1, verify_peername => "https" },
2386 ...
2387
2388Note that you must specify the hostname you connected to (or whatever
2389"peername" the protocol needs) as the C<peername> argument, otherwise no
2390peername verification will be done.
2391
2392The above will use the system-dependent default set of trusted CA
2393certificates. If you want to check against a specific CA, add the
2394C<ca_file> (or C<ca_cert>) arguments to C<tls_ctx>:
2395
2396 tls_ctx => {
2397 verify => 1,
2398 verify_peername => "https",
2399 ca_file => "my-ca-cert.pem",
2400 },
2401
2402=item I want to create a TLS/SSL server, how do I do that?
2403
2404Well, you first need to get a server certificate and key. You have
2405three options: a) ask a CA (buy one, use cacert.org etc.) b) create a
2406self-signed certificate (cheap. check the search engine of your choice,
2407there are many tutorials on the net) or c) make your own CA (tinyca2 is a
2408nice program for that purpose).
2409
2410Then create a file with your private key (in PEM format, see
2411L<AnyEvent::TLS>), followed by the certificate (also in PEM format). The
2412file should then look like this:
2413
2414 -----BEGIN RSA PRIVATE KEY-----
2415 ...header data
2416 ... lots of base64'y-stuff
2417 -----END RSA PRIVATE KEY-----
2418
2419 -----BEGIN CERTIFICATE-----
2420 ... lots of base64'y-stuff
2421 -----END CERTIFICATE-----
2422
2423The important bits are the "PRIVATE KEY" and "CERTIFICATE" parts. Then
2424specify this file as C<cert_file>:
2425
2426 tcp_server undef, $port, sub {
2427 my ($fh) = @_;
2428
2429 my $handle = new AnyEvent::Handle
2430 fh => $fh,
2431 tls => "accept",
2432 tls_ctx => { cert_file => "my-server-keycert.pem" },
2433 ...
2434
2435When you have intermediate CA certificates that your clients might not
2436know about, just append them to the C<cert_file>.
1131 2437
1132=back 2438=back
1133 2439
1134=head1 SUBCLASSING AnyEvent::Handle 2440=head1 SUBCLASSING AnyEvent::Handle
1135 2441
1141=over 4 2447=over 4
1142 2448
1143=item * all constructor arguments become object members. 2449=item * all constructor arguments become object members.
1144 2450
1145At least initially, when you pass a C<tls>-argument to the constructor it 2451At least initially, when you pass a C<tls>-argument to the constructor it
1146will end up in C<< $handle->{tls} >>. Those members might be changes or 2452will end up in C<< $handle->{tls} >>. Those members might be changed or
1147mutated later on (for example C<tls> will hold the TLS connection object). 2453mutated later on (for example C<tls> will hold the TLS connection object).
1148 2454
1149=item * other object member names are prefixed with an C<_>. 2455=item * other object member names are prefixed with an C<_>.
1150 2456
1151All object members not explicitly documented (internal use) are prefixed 2457All object members not explicitly documented (internal use) are prefixed
1154 2460
1155=item * all members not documented here and not prefixed with an underscore 2461=item * all members not documented here and not prefixed with an underscore
1156are free to use in subclasses. 2462are free to use in subclasses.
1157 2463
1158Of course, new versions of AnyEvent::Handle may introduce more "public" 2464Of course, new versions of AnyEvent::Handle may introduce more "public"
1159member variables, but thats just life, at least it is documented. 2465member variables, but that's just life. At least it is documented.
1160 2466
1161=back 2467=back
1162 2468
1163=head1 AUTHOR 2469=head1 AUTHOR
1164 2470
1165Robin Redeker C<< <elmex at ta-sa.org> >>, Marc Lehmann <schmorp@schmorp.de>. 2471Robin Redeker C<< <elmex at ta-sa.org> >>, Marc Lehmann <schmorp@schmorp.de>.
1166 2472
1167=cut 2473=cut
1168 2474
11691; # End of AnyEvent::Handle 24751
2476

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines