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 1.53 by root, Mon Jun 2 09:12:14 2008 UTC vs.
Revision 1.240 by root, Tue Dec 17 16:43:15 2013 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 = 4.1;
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 => $filehandle [C<fh> or C<connect> MANDATORY]
72 93
73The filehandle this L<AnyEvent::Handle> object will operate on. 94The filehandle this L<AnyEvent::Handle> object will operate on.
74
75NOTE: The filehandle will be set to non-blocking (using 95NOTE: The filehandle will be set to non-blocking mode (using
76AnyEvent::Util::fh_nonblocking). 96C<AnyEvent::Util::fh_nonblocking>) by the constructor and needs to stay in
97that mode.
77 98
99=item connect => [$host, $service] [C<fh> or C<connect> MANDATORY]
100
101Try to connect to the specified host and service (port), using
102C<AnyEvent::Socket::tcp_connect>. The C<$host> additionally becomes the
103default C<peername>.
104
105You have to specify either this parameter, or C<fh>, above.
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
78=item on_eof => $cb->($handle) 116=item on_prepare => $cb->($handle)
79 117
80Set the callback to be called when an end-of-file condition is detcted, 118This (rarely used) callback is called before a new connection is
81i.e. in the case of a socket, when the other side has closed the 119attempted, but after the file handle has been created (you can access that
82connection cleanly. 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).
83 123
84While not mandatory, it is highly recommended to set an eof callback, 124The return value of this callback should be the connect timeout value in
85otherwise you might end up with a closed socket while you are still 125seconds (or C<0>, or C<undef>, or the empty list, to indicate that the
86waiting for data. 126default timeout is to be used).
87 127
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
88=item on_error => $cb->($handle, $fatal) 160=item on_error => $cb->($handle, $fatal, $message)
89 161
90This is the error callback, which is called when, well, some error 162This is the error callback, which is called when, well, some error
91occured, such as not being able to resolve the hostname, failure to 163occured, such as not being able to resolve the hostname, failure to
92connect or a read error. 164connect, or a read error.
93 165
94Some errors are fatal (which is indicated by C<$fatal> being true). On 166Some errors are fatal (which is indicated by C<$fatal> being true). On
95fatal errors the handle object will be shut down and will not be 167fatal errors the handle object will be destroyed (by a call to C<< ->
168destroy >>) after invoking the error callback (which means you are free to
169examine the handle object). Examples of fatal errors are an EOF condition
170with active (but unsatisfiable) read watchers (C<EPIPE>) or I/O errors. In
171cases where the other side can close the connection at will, it is
172often easiest to not report C<EPIPE> errors in this callback.
173
174AnyEvent::Handle tries to find an appropriate error code for you to check
175against, but in some cases (TLS errors), this does not work well.
176
177If you report the error to the user, it is recommended to always output
178the C<$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
96usable. Non-fatal errors can be retried by simply returning, but it is 185Non-fatal errors can be retried by returning, but it is recommended
97recommended to simply ignore this parameter and instead abondon the handle 186to simply ignore this parameter and instead abondon the handle object
98object when this callback is invoked. 187when this callback is invoked. Examples of non-fatal errors are timeouts
188C<ETIMEDOUT>) or badly-formatted data (C<EBADMSG>).
99 189
100On callback entrance, the value of C<$!> contains the operating system 190On entry to the callback, the value of C<$!> contains the operating
101error (or C<ENOSPC>, C<EPIPE>, C<ETIMEDOUT> or C<EBADMSG>). 191system error code (or C<ENOSPC>, C<EPIPE>, C<ETIMEDOUT>, C<EBADMSG> or
192C<EPROTO>).
102 193
103While 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
104you will not be notified of errors otherwise. The default simply calls 195you will not be notified of errors otherwise. The default just calls
105C<croak>. 196C<croak>.
106 197
107=item on_read => $cb->($handle) 198=item on_read => $cb->($handle)
108 199
109This sets the default read callback, which is called when data arrives 200This sets the default read callback, which is called when data arrives
110and 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).
111 204
112To 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 >>
113method or access the C<$handle->{rbuf}> member directly. 206method or access the C<< $handle->{rbuf} >> member directly. Note that you
207must not enlarge or modify the read buffer, you can only remove data at
208the beginning from it.
114 209
210You can also call C<< ->push_read (...) >> or any other function that
211modifies the read queue. Or do both. Or ...
212
115When an EOF condition is detected then AnyEvent::Handle will first try to 213When an EOF condition is detected, AnyEvent::Handle will first try to
116feed all the remaining data to the queued callbacks and C<on_read> before 214feed all the remaining data to the queued callbacks and C<on_read> before
117calling the C<on_eof> callback. If no progress can be made, then a fatal 215calling the C<on_eof> callback. If no progress can be made, then a fatal
118error will be raised (with C<$!> set to C<EPIPE>). 216error will be raised (with C<$!> set to C<EPIPE>).
119 217
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
120=item on_drain => $cb->($handle) 239=item on_drain => $cb->($handle)
121 240
122This 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
123(or when the callback is set and the buffer is empty already). 242empty (and immediately when the handle object is created).
124 243
125To append to the write buffer, use the C<< ->push_write >> method. 244To append to the write buffer, use the C<< ->push_write >> method.
126 245
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
127=item timeout => $fractional_seconds 252=item timeout => $fractional_seconds
128 253
254=item rtimeout => $fractional_seconds
255
256=item wtimeout => $fractional_seconds
257
129If non-zero, then this enables an "inactivity" timeout: whenever this many 258If non-zero, then these enables an "inactivity" timeout: whenever this
130seconds pass without a successful read or write on the underlying file 259many seconds pass without a successful read or write on the underlying
131handle, the C<on_timeout> callback will be invoked (and if that one is 260file handle (or a call to C<timeout_reset>), the C<on_timeout> callback
132missing, an C<ETIMEDOUT> error will be raised). 261will be invoked (and if that one is missing, a non-fatal C<ETIMEDOUT>
262error will be raised).
133 263
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
134Note that timeout processing is also active when you currently do not have 271Note that timeout processing is active even when you do not have any
135any outstanding read or write requests: If you plan to keep the connection 272outstanding read or write requests: If you plan to keep the connection
136idle then you should disable the timout temporarily or ignore the timeout 273idle then you should disable the timeout temporarily or ignore the
137in the C<on_timeout> callback. 274timeout in the corresponding C<on_timeout> callback, in which case
275AnyEvent::Handle will simply restart the timeout.
138 276
139Zero (the default) disables this timeout. 277Zero (the default) disables the corresponding timeout.
140 278
141=item on_timeout => $cb->($handle) 279=item on_timeout => $cb->($handle)
280
281=item on_rtimeout => $cb->($handle)
282
283=item on_wtimeout => $cb->($handle)
142 284
143Called whenever the inactivity timeout passes. If you return from this 285Called whenever the inactivity timeout passes. If you return from this
144callback, then the timeout will be reset as if some activity had happened, 286callback, then the timeout will be reset as if some activity had happened,
145so this condition is not fatal in any way. 287so this condition is not fatal in any way.
146 288
147=item rbuf_max => <bytes> 289=item rbuf_max => <bytes>
148 290
149If 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>)
150when 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
151avoid denial-of-service attacks. 293avoid some forms of denial-of-service attacks.
152 294
153For example, a server accepting connections from untrusted sources should 295For example, a server accepting connections from untrusted sources should
154be 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
155(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
156amount 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
157isn't finished). 299isn't finished).
158 300
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
159=item read_size => <bytes> 374=item read_size => <bytes>
160 375
161The 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
162during each (loop iteration). Default: C<8192>. 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.
163 388
164=item low_water_mark => <bytes> 389=item low_water_mark => <bytes>
165 390
166Sets 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
167buffer: 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
168considered empty. 393considered empty.
169 394
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
170=item tls => "accept" | "connect" | Net::SSLeay::SSL object 422=item tls => "accept" | "connect" | Net::SSLeay::SSL object
171 423
172When this parameter is given, it enables TLS (SSL) mode, that means it 424When this parameter is given, it enables TLS (SSL) mode, that means
173will start making tls handshake and will transparently encrypt/decrypt 425AnyEvent will start a TLS handshake as soon as the connection has been
174data. 426established and will transparently encrypt/decrypt data afterwards.
427
428All TLS protocol errors will be signalled as C<EPROTO>, with an
429appropriate error message.
175 430
176TLS mode requires Net::SSLeay to be installed (it will be loaded 431TLS mode requires Net::SSLeay to be installed (it will be loaded
177automatically 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. If Net::SSLeay cannot be loaded or is too
435old, you get an C<EPROTO> error.
178 436
179For the TLS server side, use C<accept>, and for the TLS client side of a 437Unlike TCP, TLS has a server and client side: for the TLS server side, use
180connection, use C<connect> mode. 438C<accept>, and for the TLS client side of a connection, use C<connect>
439mode.
181 440
182You can also provide your own TLS connection object, but you have 441You can also provide your own TLS connection object, but you have
183to make sure that you call either C<Net::SSLeay::set_connect_state> 442to make sure that you call either C<Net::SSLeay::set_connect_state>
184or C<Net::SSLeay::set_accept_state> on it before you pass it to 443or C<Net::SSLeay::set_accept_state> on it before you pass it to
185AnyEvent::Handle. 444AnyEvent::Handle. Also, this module will take ownership of this connection
445object.
186 446
447At some future point, AnyEvent::Handle might switch to another TLS
448implementation, then the option to use your own session object will go
449away.
450
451B<IMPORTANT:> since Net::SSLeay "objects" are really only integers,
452passing in the wrong integer will lead to certain crash. This most often
453happens when one uses a stylish C<< tls => 1 >> and is surprised about the
454segmentation fault.
455
187See the C<starttls> method if you need to start TLs negotiation later. 456Use the C<< ->starttls >> method if you need to start TLS negotiation later.
188 457
189=item tls_ctx => $ssl_ctx 458=item tls_ctx => $anyevent_tls
190 459
191Use the given Net::SSLeay::CTX object to create the new TLS connection 460Use the given C<AnyEvent::TLS> object to create the new TLS connection
192(unless a connection object was specified directly). If this parameter is 461(unless a connection object was specified directly). If this
193missing, then AnyEvent::Handle will use C<AnyEvent::Handle::TLS_CTX>. 462parameter is missing (or C<undef>), then AnyEvent::Handle will use
463C<AnyEvent::Handle::TLS_CTX>.
194 464
465Instead of an object, you can also specify a hash reference with C<< key
466=> value >> pairs. Those will be passed to L<AnyEvent::TLS> to create a
467new TLS context object.
468
469=item on_starttls => $cb->($handle, $success[, $error_message])
470
471This callback will be invoked when the TLS/SSL handshake has finished. If
472C<$success> is true, then the TLS handshake succeeded, otherwise it failed
473(C<on_stoptls> will not be called in this case).
474
475The session in C<< $handle->{tls} >> can still be examined in this
476callback, even when the handshake was not successful.
477
478TLS handshake failures will not cause C<on_error> to be invoked when this
479callback is in effect, instead, the error message will be passed to C<on_starttls>.
480
481Without this callback, handshake failures lead to C<on_error> being
482called as usual.
483
484Note that you cannot just call C<starttls> again in this callback. If you
485need to do that, start an zero-second timer instead whose callback can
486then call C<< ->starttls >> again.
487
488=item on_stoptls => $cb->($handle)
489
490When a SSLv3/TLS shutdown/close notify/EOF is detected and this callback is
491set, then it will be invoked after freeing the TLS session. If it is not,
492then a TLS shutdown condition will be treated like a normal EOF condition
493on the handle.
494
495The session in C<< $handle->{tls} >> can still be examined in this
496callback.
497
498This callback will only be called on TLS shutdowns, not when the
499underlying handle signals EOF.
500
195=item json => JSON or JSON::XS object 501=item json => L<JSON>, L<JSON::PP> or L<JSON::XS> object
196 502
197This is the json coder object used by the C<json> read and write types. 503This is the json coder object used by the C<json> read and write types.
198 504
199If you don't supply it, then AnyEvent::Handle will create and use a 505If you don't supply it, then AnyEvent::Handle will create and use a
200suitable one, which will write and expect UTF-8 encoded JSON texts. 506suitable one (on demand), which will write and expect UTF-8 encoded JSON
507texts.
201 508
509=item cbor => L<CBOR::XS> object
510
511This is the cbor coder object used by the C<cbor> read and write types.
512
513If you don't supply it, then AnyEvent::Handle will create and use a
514suitable one (on demand), which will write CBOR without using extensions,
515if possible. texts.
516
202Note that you are responsible to depend on the JSON module if you want to 517Note that you are responsible to depend on the L<CBOR::XS> module if you
203use this functionality, as AnyEvent does not have a dependency itself. 518want to use this functionality, as AnyEvent does not have a dependency on
204 519it itself.
205=item filter_r => $cb
206
207=item filter_w => $cb
208
209These exist, but are undocumented at this time.
210 520
211=back 521=back
212 522
213=cut 523=cut
214 524
215sub new { 525sub new {
216 my $class = shift; 526 my $class = shift;
217
218 my $self = bless { @_ }, $class; 527 my $self = bless { @_ }, $class;
219 528
220 $self->{fh} or Carp::croak "mandatory argument fh is missing"; 529 if ($self->{fh}) {
530 $self->_start;
531 return unless $self->{fh}; # could be gone by now
532
533 } elsif ($self->{connect}) {
534 require AnyEvent::Socket;
535
536 $self->{peername} = $self->{connect}[0]
537 unless exists $self->{peername};
538
539 $self->{_skip_drain_rbuf} = 1;
540
541 {
542 Scalar::Util::weaken (my $self = $self);
543
544 $self->{_connect} =
545 AnyEvent::Socket::tcp_connect (
546 $self->{connect}[0],
547 $self->{connect}[1],
548 sub {
549 my ($fh, $host, $port, $retry) = @_;
550
551 delete $self->{_connect}; # no longer needed
552
553 if ($fh) {
554 $self->{fh} = $fh;
555
556 delete $self->{_skip_drain_rbuf};
557 $self->_start;
558
559 $self->{on_connect}
560 and $self->{on_connect}($self, $host, $port, sub {
561 delete @$self{qw(fh _tw _rtw _wtw _ww _rw _eof _queue rbuf _wbuf tls _tls_rbuf _tls_wbuf)};
562 $self->{_skip_drain_rbuf} = 1;
563 &$retry;
564 });
565
566 } else {
567 if ($self->{on_connect_error}) {
568 $self->{on_connect_error}($self, "$!");
569 $self->destroy if $self;
570 } else {
571 $self->_error ($!, 1);
572 }
573 }
574 },
575 sub {
576 local $self->{fh} = $_[0];
577
578 $self->{on_prepare}
579 ? $self->{on_prepare}->($self)
580 : ()
581 }
582 );
583 }
584
585 } else {
586 Carp::croak "AnyEvent::Handle: either an existing fh or the connect parameter must be specified";
587 }
588
589 $self
590}
591
592sub _start {
593 my ($self) = @_;
594
595 # too many clueless people try to use udp and similar sockets
596 # with AnyEvent::Handle, do them a favour.
597 my $type = getsockopt $self->{fh}, Socket::SOL_SOCKET (), Socket::SO_TYPE ();
598 Carp::croak "AnyEvent::Handle: only stream sockets supported, anything else will NOT work!"
599 if Socket::SOCK_STREAM () != (unpack "I", $type) && defined $type;
221 600
222 AnyEvent::Util::fh_nonblocking $self->{fh}, 1; 601 AnyEvent::Util::fh_nonblocking $self->{fh}, 1;
223 602
224 if ($self->{tls}) { 603 $self->{_activity} =
225 require Net::SSLeay; 604 $self->{_ractivity} =
605 $self->{_wactivity} = AE::now;
606
607 $self->{read_size} ||= 2048;
608 $self->{max_read_size} = $self->{read_size}
609 if $self->{read_size} > ($self->{max_read_size} || MAX_READ_SIZE);
610
611 $self->timeout (delete $self->{timeout} ) if $self->{timeout};
612 $self->rtimeout (delete $self->{rtimeout} ) if $self->{rtimeout};
613 $self->wtimeout (delete $self->{wtimeout} ) if $self->{wtimeout};
614
615 $self->no_delay (delete $self->{no_delay} ) if exists $self->{no_delay} && $self->{no_delay};
616 $self->keepalive (delete $self->{keepalive}) if exists $self->{keepalive} && $self->{keepalive};
617
618 $self->oobinline (exists $self->{oobinline} ? delete $self->{oobinline} : 1);
619
226 $self->starttls (delete $self->{tls}, delete $self->{tls_ctx}); 620 $self->starttls (delete $self->{tls}, delete $self->{tls_ctx})
227 } 621 if $self->{tls};
228 622
229# $self->on_eof (delete $self->{on_eof} ) if $self->{on_eof}; # nop
230# $self->on_error (delete $self->{on_error}) if $self->{on_error}; # nop
231# $self->on_read (delete $self->{on_read} ) if $self->{on_read}; # nop
232 $self->on_drain (delete $self->{on_drain}) if $self->{on_drain}; 623 $self->on_drain (delete $self->{on_drain} ) if $self->{on_drain};
233 624
234 $self->{_activity} = AnyEvent->now;
235 $self->_timeout;
236
237 $self->start_read; 625 $self->start_read
626 if $self->{on_read} || @{ $self->{_queue} };
238 627
239 $self 628 $self->_drain_wbuf;
240}
241
242sub _shutdown {
243 my ($self) = @_;
244
245 delete $self->{_tw};
246 delete $self->{_rw};
247 delete $self->{_ww};
248 delete $self->{fh};
249
250 $self->stoptls;
251} 629}
252 630
253sub _error { 631sub _error {
254 my ($self, $errno, $fatal) = @_; 632 my ($self, $errno, $fatal, $message) = @_;
255
256 $self->_shutdown
257 if $fatal;
258 633
259 $! = $errno; 634 $! = $errno;
635 $message ||= "$!";
260 636
261 if ($self->{on_error}) { 637 if ($self->{on_error}) {
262 $self->{on_error}($self, $fatal); 638 $self->{on_error}($self, $fatal, $message);
263 } else { 639 $self->destroy if $fatal;
640 } elsif ($self->{fh} || $self->{connect}) {
641 $self->destroy;
264 Carp::croak "AnyEvent::Handle uncaught error: $!"; 642 Carp::croak "AnyEvent::Handle uncaught error: $message";
265 } 643 }
266} 644}
267 645
268=item $fh = $handle->fh 646=item $fh = $handle->fh
269 647
270This method returns the file handle of the L<AnyEvent::Handle> object. 648This method returns the file handle used to create the L<AnyEvent::Handle> object.
271 649
272=cut 650=cut
273 651
274sub fh { $_[0]{fh} } 652sub fh { $_[0]{fh} }
275 653
293 $_[0]{on_eof} = $_[1]; 671 $_[0]{on_eof} = $_[1];
294} 672}
295 673
296=item $handle->on_timeout ($cb) 674=item $handle->on_timeout ($cb)
297 675
298Replace the current C<on_timeout> callback, or disables the callback 676=item $handle->on_rtimeout ($cb)
299(but not the timeout) if C<$cb> = C<undef>. See C<timeout> constructor
300argument.
301 677
302=cut 678=item $handle->on_wtimeout ($cb)
303 679
680Replace the current C<on_timeout>, C<on_rtimeout> or C<on_wtimeout>
681callback, or disables the callback (but not the timeout) if C<$cb> =
682C<undef>. See the C<timeout> constructor argument and method.
683
684=cut
685
686# see below
687
688=item $handle->autocork ($boolean)
689
690Enables or disables the current autocork behaviour (see C<autocork>
691constructor argument). Changes will only take effect on the next write.
692
693=cut
694
695sub autocork {
696 $_[0]{autocork} = $_[1];
697}
698
699=item $handle->no_delay ($boolean)
700
701Enables or disables the C<no_delay> setting (see constructor argument of
702the same name for details).
703
704=cut
705
706sub no_delay {
707 $_[0]{no_delay} = $_[1];
708
709 setsockopt $_[0]{fh}, Socket::IPPROTO_TCP (), Socket::TCP_NODELAY (), int $_[1]
710 if $_[0]{fh};
711}
712
713=item $handle->keepalive ($boolean)
714
715Enables or disables the C<keepalive> setting (see constructor argument of
716the same name for details).
717
718=cut
719
720sub keepalive {
721 $_[0]{keepalive} = $_[1];
722
723 eval {
724 local $SIG{__DIE__};
725 setsockopt $_[0]{fh}, Socket::SOL_SOCKET (), Socket::SO_KEEPALIVE (), int $_[1]
726 if $_[0]{fh};
727 };
728}
729
730=item $handle->oobinline ($boolean)
731
732Enables or disables the C<oobinline> setting (see constructor argument of
733the same name for details).
734
735=cut
736
737sub oobinline {
738 $_[0]{oobinline} = $_[1];
739
740 eval {
741 local $SIG{__DIE__};
742 setsockopt $_[0]{fh}, Socket::SOL_SOCKET (), Socket::SO_OOBINLINE (), int $_[1]
743 if $_[0]{fh};
744 };
745}
746
747=item $handle->keepalive ($boolean)
748
749Enables or disables the C<keepalive> setting (see constructor argument of
750the same name for details).
751
752=cut
753
754sub keepalive {
755 $_[0]{keepalive} = $_[1];
756
757 eval {
758 local $SIG{__DIE__};
759 setsockopt $_[0]{fh}, Socket::SOL_SOCKET (), Socket::SO_KEEPALIVE (), int $_[1]
760 if $_[0]{fh};
761 };
762}
763
764=item $handle->on_starttls ($cb)
765
766Replace the current C<on_starttls> callback (see the C<on_starttls> constructor argument).
767
768=cut
769
770sub on_starttls {
771 $_[0]{on_starttls} = $_[1];
772}
773
774=item $handle->on_stoptls ($cb)
775
776Replace the current C<on_stoptls> callback (see the C<on_stoptls> constructor argument).
777
778=cut
779
304sub on_timeout { 780sub on_stoptls {
305 $_[0]{on_timeout} = $_[1]; 781 $_[0]{on_stoptls} = $_[1];
782}
783
784=item $handle->rbuf_max ($max_octets)
785
786Configures the C<rbuf_max> setting (C<undef> disables it).
787
788=item $handle->wbuf_max ($max_octets)
789
790Configures the C<wbuf_max> setting (C<undef> disables it).
791
792=cut
793
794sub rbuf_max {
795 $_[0]{rbuf_max} = $_[1];
796}
797
798sub wbuf_max {
799 $_[0]{wbuf_max} = $_[1];
306} 800}
307 801
308############################################################################# 802#############################################################################
309 803
310=item $handle->timeout ($seconds) 804=item $handle->timeout ($seconds)
311 805
806=item $handle->rtimeout ($seconds)
807
808=item $handle->wtimeout ($seconds)
809
312Configures (or disables) the inactivity timeout. 810Configures (or disables) the inactivity timeout.
313 811
314=cut 812The timeout will be checked instantly, so this method might destroy the
813handle before it returns.
315 814
316sub timeout { 815=item $handle->timeout_reset
816
817=item $handle->rtimeout_reset
818
819=item $handle->wtimeout_reset
820
821Reset the activity timeout, as if data was received or sent.
822
823These methods are cheap to call.
824
825=cut
826
827for my $dir ("", "r", "w") {
828 my $timeout = "${dir}timeout";
829 my $tw = "_${dir}tw";
830 my $on_timeout = "on_${dir}timeout";
831 my $activity = "_${dir}activity";
832 my $cb;
833
834 *$on_timeout = sub {
835 $_[0]{$on_timeout} = $_[1];
836 };
837
838 *$timeout = sub {
317 my ($self, $timeout) = @_; 839 my ($self, $new_value) = @_;
318 840
841 $new_value >= 0
842 or Carp::croak "AnyEvent::Handle->$timeout called with negative timeout ($new_value), caught";
843
319 $self->{timeout} = $timeout; 844 $self->{$timeout} = $new_value;
320 $self->_timeout; 845 delete $self->{$tw}; &$cb;
321} 846 };
322 847
848 *{"${dir}timeout_reset"} = sub {
849 $_[0]{$activity} = AE::now;
850 };
851
852 # main workhorse:
323# reset the timeout watcher, as neccessary 853 # reset the timeout watcher, as neccessary
324# also check for time-outs 854 # also check for time-outs
325sub _timeout { 855 $cb = sub {
326 my ($self) = @_; 856 my ($self) = @_;
327 857
328 if ($self->{timeout}) { 858 if ($self->{$timeout} && $self->{fh}) {
329 my $NOW = AnyEvent->now; 859 my $NOW = AE::now;
330 860
331 # when would the timeout trigger? 861 # when would the timeout trigger?
332 my $after = $self->{_activity} + $self->{timeout} - $NOW; 862 my $after = $self->{$activity} + $self->{$timeout} - $NOW;
333 863
334 # now or in the past already? 864 # now or in the past already?
335 if ($after <= 0) { 865 if ($after <= 0) {
336 $self->{_activity} = $NOW; 866 $self->{$activity} = $NOW;
337 867
338 if ($self->{on_timeout}) { 868 if ($self->{$on_timeout}) {
339 $self->{on_timeout}($self); 869 $self->{$on_timeout}($self);
340 } else { 870 } else {
341 $self->_error (&Errno::ETIMEDOUT); 871 $self->_error (Errno::ETIMEDOUT);
872 }
873
874 # callback could have changed timeout value, optimise
875 return unless $self->{$timeout};
876
877 # calculate new after
878 $after = $self->{$timeout};
342 } 879 }
343 880
344 # callbakx could have changed timeout value, optimise 881 Scalar::Util::weaken $self;
345 return unless $self->{timeout}; 882 return unless $self; # ->error could have destroyed $self
346 883
347 # calculate new after 884 $self->{$tw} ||= AE::timer $after, 0, sub {
348 $after = $self->{timeout}; 885 delete $self->{$tw};
886 $cb->($self);
887 };
888 } else {
889 delete $self->{$tw};
349 } 890 }
350
351 Scalar::Util::weaken $self;
352
353 $self->{_tw} ||= AnyEvent->timer (after => $after, cb => sub {
354 delete $self->{_tw};
355 $self->_timeout;
356 });
357 } else {
358 delete $self->{_tw};
359 } 891 }
360} 892}
361 893
362############################################################################# 894#############################################################################
363 895
370 902
371The write queue is very simple: you can add data to its end, and 903The write queue is very simple: you can add data to its end, and
372AnyEvent::Handle will automatically try to get rid of it for you. 904AnyEvent::Handle will automatically try to get rid of it for you.
373 905
374When data could be written and the write buffer is shorter then the low 906When data could be written and the write buffer is shorter then the low
375water mark, the C<on_drain> callback will be invoked. 907water mark, the C<on_drain> callback will be invoked once.
376 908
377=over 4 909=over 4
378 910
379=item $handle->on_drain ($cb) 911=item $handle->on_drain ($cb)
380 912
381Sets the C<on_drain> callback or clears it (see the description of 913Sets the C<on_drain> callback or clears it (see the description of
382C<on_drain> in the constructor). 914C<on_drain> in the constructor).
383 915
916This method may invoke callbacks (and therefore the handle might be
917destroyed after it returns).
918
384=cut 919=cut
385 920
386sub on_drain { 921sub on_drain {
387 my ($self, $cb) = @_; 922 my ($self, $cb) = @_;
388 923
389 $self->{on_drain} = $cb; 924 $self->{on_drain} = $cb;
390 925
391 $cb->($self) 926 $cb->($self)
392 if $cb && $self->{low_water_mark} >= length $self->{wbuf}; 927 if $cb && $self->{low_water_mark} >= (length $self->{wbuf}) + (length $self->{_tls_wbuf});
393} 928}
394 929
395=item $handle->push_write ($data) 930=item $handle->push_write ($data)
396 931
397Queues the given scalar to be written. You can push as much data as you 932Queues the given scalar to be written. You can push as much data as
398want (only limited by the available memory), as C<AnyEvent::Handle> 933you want (only limited by the available memory and C<wbuf_max>), as
399buffers it independently of the kernel. 934C<AnyEvent::Handle> buffers it independently of the kernel.
935
936This method may invoke callbacks (and therefore the handle might be
937destroyed after it returns).
400 938
401=cut 939=cut
402 940
403sub _drain_wbuf { 941sub _drain_wbuf {
404 my ($self) = @_; 942 my ($self) = @_;
408 Scalar::Util::weaken $self; 946 Scalar::Util::weaken $self;
409 947
410 my $cb = sub { 948 my $cb = sub {
411 my $len = syswrite $self->{fh}, $self->{wbuf}; 949 my $len = syswrite $self->{fh}, $self->{wbuf};
412 950
413 if ($len >= 0) { 951 if (defined $len) {
414 substr $self->{wbuf}, 0, $len, ""; 952 substr $self->{wbuf}, 0, $len, "";
415 953
416 $self->{_activity} = AnyEvent->now; 954 $self->{_activity} = $self->{_wactivity} = AE::now;
417 955
418 $self->{on_drain}($self) 956 $self->{on_drain}($self)
419 if $self->{low_water_mark} >= length $self->{wbuf} 957 if $self->{low_water_mark} >= (length $self->{wbuf}) + (length $self->{_tls_wbuf})
420 && $self->{on_drain}; 958 && $self->{on_drain};
421 959
422 delete $self->{_ww} unless length $self->{wbuf}; 960 delete $self->{_ww} unless length $self->{wbuf};
423 } elsif ($! != EAGAIN && $! != EINTR && $! != WSAEWOULDBLOCK) { 961 } elsif ($! != EAGAIN && $! != EINTR && $! != WSAEWOULDBLOCK) {
424 $self->_error ($!, 1); 962 $self->_error ($!, 1);
425 } 963 }
426 }; 964 };
427 965
428 # try to write data immediately 966 # try to write data immediately
429 $cb->(); 967 $cb->() unless $self->{autocork};
430 968
431 # if still data left in wbuf, we need to poll 969 # if still data left in wbuf, we need to poll
432 $self->{_ww} = AnyEvent->io (fh => $self->{fh}, poll => "w", cb => $cb) 970 $self->{_ww} = AE::io $self->{fh}, 1, $cb
433 if length $self->{wbuf}; 971 if length $self->{wbuf};
972
973 if (
974 defined $self->{wbuf_max}
975 && $self->{wbuf_max} < length $self->{wbuf}
976 ) {
977 $self->_error (Errno::ENOSPC, 1), return;
978 }
434 }; 979 };
435} 980}
436 981
437our %WH; 982our %WH;
438 983
984# deprecated
439sub register_write_type($$) { 985sub register_write_type($$) {
440 $WH{$_[0]} = $_[1]; 986 $WH{$_[0]} = $_[1];
441} 987}
442 988
443sub push_write { 989sub push_write {
444 my $self = shift; 990 my $self = shift;
445 991
446 if (@_ > 1) { 992 if (@_ > 1) {
447 my $type = shift; 993 my $type = shift;
448 994
995 @_ = ($WH{$type} ||= _load_func "$type\::anyevent_write_type"
449 @_ = ($WH{$type} or Carp::croak "unsupported type passed to AnyEvent::Handle::push_write") 996 or Carp::croak "unsupported/unloadable type '$type' passed to AnyEvent::Handle::push_write")
450 ->($self, @_); 997 ->($self, @_);
451 } 998 }
452 999
1000 # we downgrade here to avoid hard-to-track-down bugs,
1001 # and diagnose the problem earlier and better.
1002
453 if ($self->{filter_w}) { 1003 if ($self->{tls}) {
454 $self->{filter_w}($self, \$_[0]); 1004 utf8::downgrade $self->{_tls_wbuf} .= $_[0];
1005 &_dotls ($self) if $self->{fh};
455 } else { 1006 } else {
456 $self->{wbuf} .= $_[0]; 1007 utf8::downgrade $self->{wbuf} .= $_[0];
457 $self->_drain_wbuf; 1008 $self->_drain_wbuf if $self->{fh};
458 } 1009 }
459} 1010}
460 1011
461=item $handle->push_write (type => @args) 1012=item $handle->push_write (type => @args)
462 1013
463Instead of formatting your data yourself, you can also let this module do 1014Instead of formatting your data yourself, you can also let this module
464the job by specifying a type and type-specific arguments. 1015do the job by specifying a type and type-specific arguments. You
1016can also specify the (fully qualified) name of a package, in which
1017case AnyEvent tries to load the package and then expects to find the
1018C<anyevent_write_type> function inside (see "custom write types", below).
465 1019
466Predefined types are (if you have ideas for additional types, feel free to 1020Predefined types are (if you have ideas for additional types, feel free to
467drop by and tell us): 1021drop by and tell us):
468 1022
469=over 4 1023=over 4
476=cut 1030=cut
477 1031
478register_write_type netstring => sub { 1032register_write_type netstring => sub {
479 my ($self, $string) = @_; 1033 my ($self, $string) = @_;
480 1034
481 sprintf "%d:%s,", (length $string), $string 1035 (length $string) . ":$string,"
1036};
1037
1038=item packstring => $format, $data
1039
1040An octet string prefixed with an encoded length. The encoding C<$format>
1041uses the same format as a Perl C<pack> format, but must specify a single
1042integer only (only one of C<cCsSlLqQiInNvVjJw> is allowed, plus an
1043optional C<!>, C<< < >> or C<< > >> modifier).
1044
1045=cut
1046
1047register_write_type packstring => sub {
1048 my ($self, $format, $string) = @_;
1049
1050 pack "$format/a*", $string
482}; 1051};
483 1052
484=item json => $array_or_hashref 1053=item json => $array_or_hashref
485 1054
486Encodes the given hash or array reference into a JSON object. Unless you 1055Encodes the given hash or array reference into a JSON object. Unless you
493 1062
494The generated JSON text is guaranteed not to contain any newlines: While 1063The generated JSON text is guaranteed not to contain any newlines: While
495this module doesn't need delimiters after or between JSON texts to be 1064this module doesn't need delimiters after or between JSON texts to be
496able to read them, many other languages depend on that. 1065able to read them, many other languages depend on that.
497 1066
498A simple RPC protocol that interoperates easily with others is to send 1067A simple RPC protocol that interoperates easily with other languages is
499JSON arrays (or objects, although arrays are usually the better choice as 1068to send JSON arrays (or objects, although arrays are usually the better
500they mimic how function argument passing works) and a newline after each 1069choice as they mimic how function argument passing works) and a newline
501JSON text: 1070after each JSON text:
502 1071
503 $handle->push_write (json => ["method", "arg1", "arg2"]); # whatever 1072 $handle->push_write (json => ["method", "arg1", "arg2"]); # whatever
504 $handle->push_write ("\012"); 1073 $handle->push_write ("\012");
505 1074
506An AnyEvent::Handle receiver would simply use the C<json> read type and 1075An AnyEvent::Handle receiver would simply use the C<json> read type and
509 $handle->push_read (json => sub { my $array = $_[1]; ... }); 1078 $handle->push_read (json => sub { my $array = $_[1]; ... });
510 1079
511Other languages could read single lines terminated by a newline and pass 1080Other languages could read single lines terminated by a newline and pass
512this line into their JSON decoder of choice. 1081this line into their JSON decoder of choice.
513 1082
1083=item cbor => $perl_scalar
1084
1085Encodes the given scalar into a CBOR value. Unless you provide your own
1086L<CBOR::XS> object, this means it will be encoded to a CBOR string not
1087using any extensions, if possible.
1088
1089CBOR values are self-delimiting, so you can write CBOR at one end of
1090a handle and read them at the other end without using any additional
1091framing.
1092
1093A simple nd very very fast RPC protocol that interoperates with
1094other languages is to send CBOR and receive CBOR values (arrays are
1095recommended):
1096
1097 $handle->push_write (cbor => ["method", "arg1", "arg2"]); # whatever
1098
1099An AnyEvent::Handle receiver would simply use the C<cbor> read type:
1100
1101 $handle->push_read (cbor => sub { my $array = $_[1]; ... });
1102
514=cut 1103=cut
1104
1105sub json_coder() {
1106 eval { require JSON::XS; JSON::XS->new->utf8 }
1107 || do { require JSON::PP; JSON::PP->new->utf8 }
1108}
515 1109
516register_write_type json => sub { 1110register_write_type json => sub {
517 my ($self, $ref) = @_; 1111 my ($self, $ref) = @_;
518 1112
519 require JSON; 1113 ($self->{json} ||= json_coder)
520 1114 ->encode ($ref)
521 $self->{json} ? $self->{json}->encode ($ref)
522 : JSON::encode_json ($ref)
523}; 1115};
524 1116
1117sub cbor_coder() {
1118 require CBOR::XS;
1119 CBOR::XS->new
1120}
1121
1122register_write_type cbor => sub {
1123 my ($self, $scalar) = @_;
1124
1125 ($self->{cbor} ||= cbor_coder)
1126 ->encode ($scalar)
1127};
1128
1129=item storable => $reference
1130
1131Freezes the given reference using L<Storable> and writes it to the
1132handle. Uses the C<nfreeze> format.
1133
1134=cut
1135
1136register_write_type storable => sub {
1137 my ($self, $ref) = @_;
1138
1139 require Storable unless $Storable::VERSION;
1140
1141 pack "w/a*", Storable::nfreeze ($ref)
1142};
1143
525=back 1144=back
526 1145
527=item AnyEvent::Handle::register_write_type type => $coderef->($handle, @args) 1146=item $handle->push_shutdown
528 1147
529This function (not method) lets you add your own types to C<push_write>. 1148Sometimes you know you want to close the socket after writing your data
1149before it was actually written. One way to do that is to replace your
1150C<on_drain> handler by a callback that shuts down the socket (and set
1151C<low_water_mark> to C<0>). This method is a shorthand for just that, and
1152replaces the C<on_drain> callback with:
1153
1154 sub { shutdown $_[0]{fh}, 1 }
1155
1156This simply shuts down the write side and signals an EOF condition to the
1157the peer.
1158
1159You can rely on the normal read queue and C<on_eof> handling
1160afterwards. This is the cleanest way to close a connection.
1161
1162This method may invoke callbacks (and therefore the handle might be
1163destroyed after it returns).
1164
1165=cut
1166
1167sub push_shutdown {
1168 my ($self) = @_;
1169
1170 delete $self->{low_water_mark};
1171 $self->on_drain (sub { shutdown $_[0]{fh}, 1 });
1172}
1173
1174=item custom write types - Package::anyevent_write_type $handle, @args
1175
1176Instead of one of the predefined types, you can also specify the name of
1177a package. AnyEvent will try to load the package and then expects to find
1178a function named C<anyevent_write_type> inside. If it isn't found, it
1179progressively tries to load the parent package until it either finds the
1180function (good) or runs out of packages (bad).
1181
530Whenever the given C<type> is used, C<push_write> will invoke the code 1182Whenever the given C<type> is used, C<push_write> will the function with
531reference with the handle object and the remaining arguments. 1183the handle object and the remaining arguments.
532 1184
533The code reference is supposed to return a single octet string that will 1185The function is supposed to return a single octet string that will be
534be appended to the write buffer. 1186appended to the write buffer, so you can mentally treat this function as a
1187"arguments to on-the-wire-format" converter.
535 1188
536Note that this is a function, and all types registered this way will be 1189Example: implement a custom write type C<join> that joins the remaining
537global, so try to use unique names. 1190arguments using the first one.
1191
1192 $handle->push_write (My::Type => " ", 1,2,3);
1193
1194 # uses the following package, which can be defined in the "My::Type" or in
1195 # the "My" modules to be auto-loaded, or just about anywhere when the
1196 # My::Type::anyevent_write_type is defined before invoking it.
1197
1198 package My::Type;
1199
1200 sub anyevent_write_type {
1201 my ($handle, $delim, @args) = @_;
1202
1203 join $delim, @args
1204 }
538 1205
539=cut 1206=cut
540 1207
541############################################################################# 1208#############################################################################
542 1209
551ways, the "simple" way, using only C<on_read> and the "complex" way, using 1218ways, the "simple" way, using only C<on_read> and the "complex" way, using
552a queue. 1219a queue.
553 1220
554In the simple case, you just install an C<on_read> callback and whenever 1221In the simple case, you just install an C<on_read> callback and whenever
555new data arrives, it will be called. You can then remove some data (if 1222new data arrives, it will be called. You can then remove some data (if
556enough is there) from the read buffer (C<< $handle->rbuf >>) if you want 1223enough is there) from the read buffer (C<< $handle->rbuf >>). Or you can
557or not. 1224leave the data there if you want to accumulate more (e.g. when only a
1225partial message has been received so far), or change the read queue with
1226e.g. C<push_read>.
558 1227
559In the more complex case, you want to queue multiple callbacks. In this 1228In the more complex case, you want to queue multiple callbacks. In this
560case, AnyEvent::Handle will call the first queued callback each time new 1229case, AnyEvent::Handle will call the first queued callback each time new
561data arrives and removes it when it has done its job (see C<push_read>, 1230data arrives (also the first time it is queued) and remove it when it has
562below). 1231done its job (see C<push_read>, below).
563 1232
564This way you can, for example, push three line-reads, followed by reading 1233This way you can, for example, push three line-reads, followed by reading
565a chunk of data, and AnyEvent::Handle will execute them in order. 1234a chunk of data, and AnyEvent::Handle will execute them in order.
566 1235
567Example 1: EPP protocol parser. EPP sends 4 byte length info, followed by 1236Example 1: EPP protocol parser. EPP sends 4 byte length info, followed by
580 # handle xml 1249 # handle xml
581 }); 1250 });
582 }); 1251 });
583 }); 1252 });
584 1253
585Example 2: Implement a client for a protocol that replies either with 1254Example 2: Implement a client for a protocol that replies either with "OK"
586"OK" and another line or "ERROR" for one request, and 64 bytes for the 1255and another line or "ERROR" for the first request that is sent, and 64
587second request. Due tot he availability of a full queue, we can just 1256bytes for the second request. Due to the availability of a queue, we can
588pipeline sending both requests and manipulate the queue as necessary in 1257just pipeline sending both requests and manipulate the queue as necessary
589the callbacks: 1258in the callbacks.
590 1259
591 # request one 1260When the first callback is called and sees an "OK" response, it will
1261C<unshift> another line-read. This line-read will be queued I<before> the
126264-byte chunk callback.
1263
1264 # request one, returns either "OK + extra line" or "ERROR"
592 $handle->push_write ("request 1\015\012"); 1265 $handle->push_write ("request 1\015\012");
593 1266
594 # we expect "ERROR" or "OK" as response, so push a line read 1267 # we expect "ERROR" or "OK" as response, so push a line read
595 $handle->push_read (line => sub { 1268 $handle->push_read (line => sub {
596 # if we got an "OK", we have to _prepend_ another line, 1269 # if we got an "OK", we have to _prepend_ another line,
603 ... 1276 ...
604 }); 1277 });
605 } 1278 }
606 }); 1279 });
607 1280
608 # request two 1281 # request two, simply returns 64 octets
609 $handle->push_write ("request 2\015\012"); 1282 $handle->push_write ("request 2\015\012");
610 1283
611 # simply read 64 bytes, always 1284 # simply read 64 bytes, always
612 $handle->push_read (chunk => 64, sub { 1285 $handle->push_read (chunk => 64, sub {
613 my $response = $_[1]; 1286 my $response = $_[1];
619=cut 1292=cut
620 1293
621sub _drain_rbuf { 1294sub _drain_rbuf {
622 my ($self) = @_; 1295 my ($self) = @_;
623 1296
1297 # avoid recursion
1298 return if $self->{_skip_drain_rbuf};
1299 local $self->{_skip_drain_rbuf} = 1;
1300
1301 while () {
1302 # we need to use a separate tls read buffer, as we must not receive data while
1303 # we are draining the buffer, and this can only happen with TLS.
1304 $self->{rbuf} .= delete $self->{_tls_rbuf}
1305 if exists $self->{_tls_rbuf};
1306
1307 my $len = length $self->{rbuf};
1308
1309 if (my $cb = shift @{ $self->{_queue} }) {
1310 unless ($cb->($self)) {
1311 # no progress can be made
1312 # (not enough data and no data forthcoming)
1313 $self->_error (Errno::EPIPE, 1), return
1314 if $self->{_eof};
1315
1316 unshift @{ $self->{_queue} }, $cb;
1317 last;
1318 }
1319 } elsif ($self->{on_read}) {
1320 last unless $len;
1321
1322 $self->{on_read}($self);
1323
1324 if (
1325 $len == length $self->{rbuf} # if no data has been consumed
1326 && !@{ $self->{_queue} } # and the queue is still empty
1327 && $self->{on_read} # but we still have on_read
1328 ) {
1329 # no further data will arrive
1330 # so no progress can be made
1331 $self->_error (Errno::EPIPE, 1), return
1332 if $self->{_eof};
1333
1334 last; # more data might arrive
1335 }
1336 } else {
1337 # read side becomes idle
1338 delete $self->{_rw} unless $self->{tls};
1339 last;
1340 }
1341 }
1342
1343 if ($self->{_eof}) {
1344 $self->{on_eof}
1345 ? $self->{on_eof}($self)
1346 : $self->_error (0, 1, "Unexpected end-of-file");
1347
1348 return;
1349 }
1350
624 if ( 1351 if (
625 defined $self->{rbuf_max} 1352 defined $self->{rbuf_max}
626 && $self->{rbuf_max} < length $self->{rbuf} 1353 && $self->{rbuf_max} < length $self->{rbuf}
627 ) { 1354 ) {
628 return $self->_error (&Errno::ENOSPC, 1); 1355 $self->_error (Errno::ENOSPC, 1), return;
629 } 1356 }
630 1357
631 return if $self->{in_drain}; 1358 # may need to restart read watcher
632 local $self->{in_drain} = 1; 1359 unless ($self->{_rw}) {
633 1360 $self->start_read
634 while (my $len = length $self->{rbuf}) { 1361 if $self->{on_read} || @{ $self->{_queue} };
635 no strict 'refs';
636 if (my $cb = shift @{ $self->{_queue} }) {
637 unless ($cb->($self)) {
638 if ($self->{_eof}) {
639 # no progress can be made (not enough data and no data forthcoming)
640 return $self->_error (&Errno::EPIPE, 1);
641 }
642
643 unshift @{ $self->{_queue} }, $cb;
644 return;
645 }
646 } elsif ($self->{on_read}) {
647 $self->{on_read}($self);
648
649 if (
650 $self->{_eof} # if no further data will arrive
651 && $len == length $self->{rbuf} # and no data has been consumed
652 && !@{ $self->{_queue} } # and the queue is still empty
653 && $self->{on_read} # and we still want to read data
654 ) {
655 # then no progress can be made
656 return $self->_error (&Errno::EPIPE, 1);
657 }
658 } else {
659 # read side becomes idle
660 delete $self->{_rw};
661 return;
662 }
663 } 1362 }
664
665 $self->{on_eof}($self)
666 if $self->{_eof} && $self->{on_eof};
667} 1363}
668 1364
669=item $handle->on_read ($cb) 1365=item $handle->on_read ($cb)
670 1366
671This replaces the currently set C<on_read> callback, or clears it (when 1367This replaces the currently set C<on_read> callback, or clears it (when
672the new callback is C<undef>). See the description of C<on_read> in the 1368the new callback is C<undef>). See the description of C<on_read> in the
673constructor. 1369constructor.
674 1370
1371This method may invoke callbacks (and therefore the handle might be
1372destroyed after it returns).
1373
675=cut 1374=cut
676 1375
677sub on_read { 1376sub on_read {
678 my ($self, $cb) = @_; 1377 my ($self, $cb) = @_;
679 1378
680 $self->{on_read} = $cb; 1379 $self->{on_read} = $cb;
1380 $self->_drain_rbuf if $cb;
681} 1381}
682 1382
683=item $handle->rbuf 1383=item $handle->rbuf
684 1384
685Returns the read buffer (as a modifiable lvalue). 1385Returns the read buffer (as a modifiable lvalue). You can also access the
1386read buffer directly as the C<< ->{rbuf} >> member, if you want (this is
1387much faster, and no less clean).
686 1388
687You can access the read buffer directly as the C<< ->{rbuf} >> member, if 1389The only operation allowed on the read buffer (apart from looking at it)
688you want. 1390is removing data from its beginning. Otherwise modifying or appending to
1391it is not allowed and will lead to hard-to-track-down bugs.
689 1392
690NOTE: The read buffer should only be used or modified if the C<on_read>, 1393NOTE: The read buffer should only be used or modified in the C<on_read>
691C<push_read> or C<unshift_read> methods are used. The other read methods 1394callback or when C<push_read> or C<unshift_read> are used with a single
692automatically manage the read buffer. 1395callback (i.e. untyped). Typed C<push_read> and C<unshift_read> methods
1396will manage the read buffer on their own.
693 1397
694=cut 1398=cut
695 1399
696sub rbuf : lvalue { 1400sub rbuf : lvalue {
697 $_[0]{rbuf} 1401 $_[0]{rbuf}
714 1418
715If enough data was available, then the callback must remove all data it is 1419If enough data was available, then the callback must remove all data it is
716interested in (which can be none at all) and return a true value. After returning 1420interested in (which can be none at all) and return a true value. After returning
717true, it will be removed from the queue. 1421true, it will be removed from the queue.
718 1422
1423These methods may invoke callbacks (and therefore the handle might be
1424destroyed after it returns).
1425
719=cut 1426=cut
720 1427
721our %RH; 1428our %RH;
722 1429
723sub register_read_type($$) { 1430sub register_read_type($$) {
729 my $cb = pop; 1436 my $cb = pop;
730 1437
731 if (@_) { 1438 if (@_) {
732 my $type = shift; 1439 my $type = shift;
733 1440
1441 $cb = ($RH{$type} ||= _load_func "$type\::anyevent_read_type"
734 $cb = ($RH{$type} or Carp::croak "unsupported type passed to AnyEvent::Handle::push_read") 1442 or Carp::croak "unsupported/unloadable type '$type' passed to AnyEvent::Handle::push_read")
735 ->($self, $cb, @_); 1443 ->($self, $cb, @_);
736 } 1444 }
737 1445
738 push @{ $self->{_queue} }, $cb; 1446 push @{ $self->{_queue} }, $cb;
739 $self->_drain_rbuf; 1447 $self->_drain_rbuf;
744 my $cb = pop; 1452 my $cb = pop;
745 1453
746 if (@_) { 1454 if (@_) {
747 my $type = shift; 1455 my $type = shift;
748 1456
1457 $cb = ($RH{$type} ||= _load_func "$type\::anyevent_read_type"
749 $cb = ($RH{$type} or Carp::croak "unsupported type passed to AnyEvent::Handle::unshift_read") 1458 or Carp::croak "unsupported/unloadable type '$type' passed to AnyEvent::Handle::unshift_read")
750 ->($self, $cb, @_); 1459 ->($self, $cb, @_);
751 } 1460 }
752
753 1461
754 unshift @{ $self->{_queue} }, $cb; 1462 unshift @{ $self->{_queue} }, $cb;
755 $self->_drain_rbuf; 1463 $self->_drain_rbuf;
756} 1464}
757 1465
759 1467
760=item $handle->unshift_read (type => @args, $cb) 1468=item $handle->unshift_read (type => @args, $cb)
761 1469
762Instead of providing a callback that parses the data itself you can chose 1470Instead of providing a callback that parses the data itself you can chose
763between a number of predefined parsing formats, for chunks of data, lines 1471between a number of predefined parsing formats, for chunks of data, lines
764etc. 1472etc. You can also specify the (fully qualified) name of a package, in
1473which case AnyEvent tries to load the package and then expects to find the
1474C<anyevent_read_type> function inside (see "custom read types", below).
765 1475
766Predefined types are (if you have ideas for additional types, feel free to 1476Predefined types are (if you have ideas for additional types, feel free to
767drop by and tell us): 1477drop by and tell us):
768 1478
769=over 4 1479=over 4
775data. 1485data.
776 1486
777Example: read 2 bytes. 1487Example: read 2 bytes.
778 1488
779 $handle->push_read (chunk => 2, sub { 1489 $handle->push_read (chunk => 2, sub {
780 warn "yay ", unpack "H*", $_[1]; 1490 say "yay " . unpack "H*", $_[1];
781 }); 1491 });
782 1492
783=cut 1493=cut
784 1494
785register_read_type chunk => sub { 1495register_read_type chunk => sub {
790 $cb->($_[0], substr $_[0]{rbuf}, 0, $len, ""); 1500 $cb->($_[0], substr $_[0]{rbuf}, 0, $len, "");
791 1 1501 1
792 } 1502 }
793}; 1503};
794 1504
795# compatibility with older API
796sub push_read_chunk {
797 $_[0]->push_read (chunk => $_[1], $_[2]);
798}
799
800sub unshift_read_chunk {
801 $_[0]->unshift_read (chunk => $_[1], $_[2]);
802}
803
804=item line => [$eol, ]$cb->($handle, $line, $eol) 1505=item line => [$eol, ]$cb->($handle, $line, $eol)
805 1506
806The callback will be called only once a full line (including the end of 1507The callback will be called only once a full line (including the end of
807line marker, C<$eol>) has been read. This line (excluding the end of line 1508line marker, C<$eol>) has been read. This line (excluding the end of line
808marker) will be passed to the callback as second argument (C<$line>), and 1509marker) will be passed to the callback as second argument (C<$line>), and
823=cut 1524=cut
824 1525
825register_read_type line => sub { 1526register_read_type line => sub {
826 my ($self, $cb, $eol) = @_; 1527 my ($self, $cb, $eol) = @_;
827 1528
828 $eol = qr|(\015?\012)| if @_ < 3; 1529 if (@_ < 3) {
829 $eol = quotemeta $eol unless ref $eol; 1530 # this is faster then the generic code below
830 $eol = qr|^(.*?)($eol)|s;
831
832 sub { 1531 sub {
833 $_[0]{rbuf} =~ s/$eol// or return; 1532 (my $pos = index $_[0]{rbuf}, "\012") >= 0
1533 or return;
834 1534
1535 (my $str = substr $_[0]{rbuf}, 0, $pos + 1, "") =~ s/(\015?\012)\Z// or die;
835 $cb->($_[0], $1, $2); 1536 $cb->($_[0], $str, "$1");
836 1
837 }
838};
839
840# compatibility with older API
841sub push_read_line {
842 my $self = shift;
843 $self->push_read (line => @_);
844}
845
846sub unshift_read_line {
847 my $self = shift;
848 $self->unshift_read (line => @_);
849}
850
851=item netstring => $cb->($handle, $string)
852
853A netstring (http://cr.yp.to/proto/netstrings.txt, this is not an endorsement).
854
855Throws an error with C<$!> set to EBADMSG on format violations.
856
857=cut
858
859register_read_type netstring => sub {
860 my ($self, $cb) = @_;
861
862 sub {
863 unless ($_[0]{rbuf} =~ s/^(0|[1-9][0-9]*)://) {
864 if ($_[0]{rbuf} =~ /[^0-9]/) {
865 $self->_error (&Errno::EBADMSG);
866 } 1537 1
867 return;
868 } 1538 }
1539 } else {
1540 $eol = quotemeta $eol unless ref $eol;
1541 $eol = qr|^(.*?)($eol)|s;
869 1542
870 my $len = $1; 1543 sub {
1544 $_[0]{rbuf} =~ s/$eol// or return;
871 1545
872 $self->unshift_read (chunk => $len, sub { 1546 $cb->($_[0], "$1", "$2");
873 my $string = $_[1];
874 $_[0]->unshift_read (chunk => 1, sub {
875 if ($_[1] eq ",") {
876 $cb->($_[0], $string);
877 } else {
878 $self->_error (&Errno::EBADMSG);
879 }
880 }); 1547 1
881 }); 1548 }
882
883 1
884 } 1549 }
885}; 1550};
886 1551
887=item regex => $accept[, $reject[, $skip], $cb->($handle, $data) 1552=item regex => $accept[, $reject[, $skip], $cb->($handle, $data)
888 1553
908the receive buffer when neither C<$accept> nor C<$reject> match, 1573the receive buffer when neither C<$accept> nor C<$reject> match,
909and everything preceding and including the match will be accepted 1574and everything preceding and including the match will be accepted
910unconditionally. This is useful to skip large amounts of data that you 1575unconditionally. This is useful to skip large amounts of data that you
911know cannot be matched, so that the C<$accept> or C<$reject> regex do not 1576know cannot be matched, so that the C<$accept> or C<$reject> regex do not
912have to start matching from the beginning. This is purely an optimisation 1577have to start matching from the beginning. This is purely an optimisation
913and is usually worth only when you expect more than a few kilobytes. 1578and is usually worth it only when you expect more than a few kilobytes.
914 1579
915Example: expect a http header, which ends at C<\015\012\015\012>. Since we 1580Example: expect a http header, which ends at C<\015\012\015\012>. Since we
916expect the header to be very large (it isn't in practise, but...), we use 1581expect the header to be very large (it isn't in practice, but...), we use
917a skip regex to skip initial portions. The skip regex is tricky in that 1582a skip regex to skip initial portions. The skip regex is tricky in that
918it only accepts something not ending in either \015 or \012, as these are 1583it only accepts something not ending in either \015 or \012, as these are
919required for the accept regex. 1584required for the accept regex.
920 1585
921 $handle->push_read (regex => 1586 $handle->push_read (regex =>
934 1599
935 sub { 1600 sub {
936 # accept 1601 # accept
937 if ($$rbuf =~ $accept) { 1602 if ($$rbuf =~ $accept) {
938 $data .= substr $$rbuf, 0, $+[0], ""; 1603 $data .= substr $$rbuf, 0, $+[0], "";
939 $cb->($self, $data); 1604 $cb->($_[0], $data);
940 return 1; 1605 return 1;
941 } 1606 }
942 1607
943 # reject 1608 # reject
944 if ($reject && $$rbuf =~ $reject) { 1609 if ($reject && $$rbuf =~ $reject) {
945 $self->_error (&Errno::EBADMSG); 1610 $_[0]->_error (Errno::EBADMSG);
946 } 1611 }
947 1612
948 # skip 1613 # skip
949 if ($skip && $$rbuf =~ $skip) { 1614 if ($skip && $$rbuf =~ $skip) {
950 $data .= substr $$rbuf, 0, $+[0], ""; 1615 $data .= substr $$rbuf, 0, $+[0], "";
952 1617
953 () 1618 ()
954 } 1619 }
955}; 1620};
956 1621
1622=item netstring => $cb->($handle, $string)
1623
1624A netstring (http://cr.yp.to/proto/netstrings.txt, this is not an endorsement).
1625
1626Throws an error with C<$!> set to EBADMSG on format violations.
1627
1628=cut
1629
1630register_read_type netstring => sub {
1631 my ($self, $cb) = @_;
1632
1633 sub {
1634 unless ($_[0]{rbuf} =~ s/^(0|[1-9][0-9]*)://) {
1635 if ($_[0]{rbuf} =~ /[^0-9]/) {
1636 $_[0]->_error (Errno::EBADMSG);
1637 }
1638 return;
1639 }
1640
1641 my $len = $1;
1642
1643 $_[0]->unshift_read (chunk => $len, sub {
1644 my $string = $_[1];
1645 $_[0]->unshift_read (chunk => 1, sub {
1646 if ($_[1] eq ",") {
1647 $cb->($_[0], $string);
1648 } else {
1649 $_[0]->_error (Errno::EBADMSG);
1650 }
1651 });
1652 });
1653
1654 1
1655 }
1656};
1657
1658=item packstring => $format, $cb->($handle, $string)
1659
1660An octet string prefixed with an encoded length. The encoding C<$format>
1661uses the same format as a Perl C<pack> format, but must specify a single
1662integer only (only one of C<cCsSlLqQiInNvVjJw> is allowed, plus an
1663optional C<!>, C<< < >> or C<< > >> modifier).
1664
1665For example, DNS over TCP uses a prefix of C<n> (2 octet network order),
1666EPP uses a prefix of C<N> (4 octtes).
1667
1668Example: read a block of data prefixed by its length in BER-encoded
1669format (very efficient).
1670
1671 $handle->push_read (packstring => "w", sub {
1672 my ($handle, $data) = @_;
1673 });
1674
1675=cut
1676
1677register_read_type packstring => sub {
1678 my ($self, $cb, $format) = @_;
1679
1680 sub {
1681 # when we can use 5.10 we can use ".", but for 5.8 we use the re-pack method
1682 defined (my $len = eval { unpack $format, $_[0]{rbuf} })
1683 or return;
1684
1685 $format = length pack $format, $len;
1686
1687 # bypass unshift if we already have the remaining chunk
1688 if ($format + $len <= length $_[0]{rbuf}) {
1689 my $data = substr $_[0]{rbuf}, $format, $len;
1690 substr $_[0]{rbuf}, 0, $format + $len, "";
1691 $cb->($_[0], $data);
1692 } else {
1693 # remove prefix
1694 substr $_[0]{rbuf}, 0, $format, "";
1695
1696 # read remaining chunk
1697 $_[0]->unshift_read (chunk => $len, $cb);
1698 }
1699
1700 1
1701 }
1702};
1703
957=item json => $cb->($handle, $hash_or_arrayref) 1704=item json => $cb->($handle, $hash_or_arrayref)
958 1705
959Reads a JSON object or array, decodes it and passes it to the callback. 1706Reads a JSON object or array, decodes it and passes it to the
1707callback. When a parse error occurs, an C<EBADMSG> error will be raised.
960 1708
961If a C<json> object was passed to the constructor, then that will be used 1709If a C<json> object was passed to the constructor, then that will be
962for the final decode, otherwise it will create a JSON coder expecting UTF-8. 1710used for the final decode, otherwise it will create a L<JSON::XS> or
1711L<JSON::PP> coder object expecting UTF-8.
963 1712
964This read type uses the incremental parser available with JSON version 1713This read type uses the incremental parser available with JSON version
9652.09 (and JSON::XS version 2.2) and above. You have to provide a 17142.09 (and JSON::XS version 2.2) and above.
966dependency on your own: this module will load the JSON module, but
967AnyEvent does not depend on it itself.
968 1715
969Since JSON texts are fully self-delimiting, the C<json> read and write 1716Since JSON texts are fully self-delimiting, the C<json> read and write
970types are an ideal simple RPC protocol: just exchange JSON datagrams. See 1717types are an ideal simple RPC protocol: just exchange JSON datagrams. See
971the C<json> write type description, above, for an actual example. 1718the C<json> write type description, above, for an actual example.
972 1719
973=cut 1720=cut
974 1721
975register_read_type json => sub { 1722register_read_type json => sub {
976 my ($self, $cb, $accept, $reject, $skip) = @_; 1723 my ($self, $cb) = @_;
977 1724
978 require JSON; 1725 my $json = $self->{json} ||= json_coder;
979 1726
980 my $data; 1727 my $data;
981 my $rbuf = \$self->{rbuf};
982
983 my $json = $self->{json} ||= JSON->new->utf8;
984 1728
985 sub { 1729 sub {
986 my $ref = $json->incr_parse ($self->{rbuf}); 1730 my $ref = eval { $json->incr_parse ($_[0]{rbuf}) };
987 1731
988 if ($ref) { 1732 if ($ref) {
989 $self->{rbuf} = $json->incr_text; 1733 $_[0]{rbuf} = $json->incr_text;
990 $json->incr_text = ""; 1734 $json->incr_text = "";
991 $cb->($self, $ref); 1735 $cb->($_[0], $ref);
992 1736
993 1 1737 1
1738 } elsif ($@) {
1739 # error case
1740 $json->incr_skip;
1741
1742 $_[0]{rbuf} = $json->incr_text;
1743 $json->incr_text = "";
1744
1745 $_[0]->_error (Errno::EBADMSG);
1746
1747 ()
994 } else { 1748 } else {
995 $self->{rbuf} = ""; 1749 $_[0]{rbuf} = "";
1750
996 () 1751 ()
997 } 1752 }
998 } 1753 }
999}; 1754};
1000 1755
1756=item cbor => $cb->($handle, $scalar)
1757
1758Reads a CBOR value, decodes it and passes it to the callback. When a parse
1759error occurs, an C<EBADMSG> error will be raised.
1760
1761If a L<CBOR::XS> object was passed to the constructor, then that will be
1762used for the final decode, otherwise it will create a CBOR coder without
1763enabling any options.
1764
1765You have to provide a dependency to L<CBOR::XS> on your own: this module
1766will load the L<CBOR::XS> module, but AnyEvent does not depend on it
1767itself.
1768
1769Since CBOR values are fully self-delimiting, the C<cbor> read and write
1770types are an ideal simple RPC protocol: just exchange CBOR datagrams. See
1771the C<cbor> write type description, above, for an actual example.
1772
1773=cut
1774
1775register_read_type cbor => sub {
1776 my ($self, $cb) = @_;
1777
1778 my $cbor = $self->{cbor} ||= cbor_coder;
1779
1780 my $data;
1781
1782 sub {
1783 my (@value) = eval { $cbor->incr_parse ($_[0]{rbuf}) };
1784
1785 if (@value) {
1786 $cb->($_[0], @value);
1787
1788 1
1789 } elsif ($@) {
1790 # error case
1791 $cbor->incr_reset;
1792
1793 $_[0]->_error (Errno::EBADMSG);
1794
1795 ()
1796 } else {
1797 ()
1798 }
1799 }
1800};
1801
1802=item storable => $cb->($handle, $ref)
1803
1804Deserialises a L<Storable> frozen representation as written by the
1805C<storable> write type (BER-encoded length prefix followed by nfreeze'd
1806data).
1807
1808Raises C<EBADMSG> error if the data could not be decoded.
1809
1810=cut
1811
1812register_read_type storable => sub {
1813 my ($self, $cb) = @_;
1814
1815 require Storable unless $Storable::VERSION;
1816
1817 sub {
1818 # when we can use 5.10 we can use ".", but for 5.8 we use the re-pack method
1819 defined (my $len = eval { unpack "w", $_[0]{rbuf} })
1820 or return;
1821
1822 my $format = length pack "w", $len;
1823
1824 # bypass unshift if we already have the remaining chunk
1825 if ($format + $len <= length $_[0]{rbuf}) {
1826 my $data = substr $_[0]{rbuf}, $format, $len;
1827 substr $_[0]{rbuf}, 0, $format + $len, "";
1828
1829 eval { $cb->($_[0], Storable::thaw ($data)); 1 }
1830 or return $_[0]->_error (Errno::EBADMSG);
1831 } else {
1832 # remove prefix
1833 substr $_[0]{rbuf}, 0, $format, "";
1834
1835 # read remaining chunk
1836 $_[0]->unshift_read (chunk => $len, sub {
1837 eval { $cb->($_[0], Storable::thaw ($_[1])); 1 }
1838 or $_[0]->_error (Errno::EBADMSG);
1839 });
1840 }
1841
1842 1
1843 }
1844};
1845
1846=item tls_detect => $cb->($handle, $detect, $major, $minor)
1847
1848Checks the input stream for a valid SSL or TLS handshake TLSPaintext
1849record without consuming anything. Only SSL version 3 or higher
1850is handled, up to the fictituous protocol 4.x (but both SSL3+ and
1851SSL2-compatible framing is supported).
1852
1853If it detects that the input data is likely TLS, it calls the callback
1854with a true value for C<$detect> and the (on-wire) TLS version as second
1855and third argument (C<$major> is C<3>, and C<$minor> is 0..3 for SSL
18563.0, TLS 1.0, 1.1 and 1.2, respectively). If it detects the input to
1857be definitely not TLS, it calls the callback with a false value for
1858C<$detect>.
1859
1860The callback could use this information to decide whether or not to start
1861TLS negotiation.
1862
1863In all cases the data read so far is passed to the following read
1864handlers.
1865
1866Usually you want to use the C<tls_autostart> read type instead.
1867
1868If you want to design a protocol that works in the presence of TLS
1869dtection, make sure that any non-TLS data doesn't start with the octet 22
1870(ASCII SYN, 16 hex) or 128-255 (i.e. highest bit set). The checks this
1871read type does are a bit more strict, but might losen in the future to
1872accomodate protocol changes.
1873
1874This read type does not rely on L<AnyEvent::TLS> (and thus, not on
1875L<Net::SSLeay>).
1876
1877=item tls_autostart => $tls[, $tls_ctx]
1878
1879Tries to detect a valid SSL or TLS handshake. If one is detected, it tries
1880to start tls by calling C<starttls> with the given arguments.
1881
1882In practise, C<$tls> must be C<accept>, or a Net::SSLeay context that has
1883been configured to accept, as servers do not normally send a handshake on
1884their own and ths cannot be detected in this way.
1885
1886See C<tls_detect> above for more details.
1887
1888Example: give the client a chance to start TLS before accepting a text
1889line.
1890
1891 $hdl->push_read (tls_detect => "accept");
1892 $hdl->push_read (line => sub {
1893 print "received ", ($_[0]{tls} ? "encrypted" : "cleartext"), " <$_[1]>\n";
1894 });
1895
1896=cut
1897
1898register_read_type tls_detect => sub {
1899 my ($self, $cb) = @_;
1900
1901 sub {
1902 # this regex matches a full or partial tls record
1903 if (
1904 # ssl3+: type(22=handshake) major(=3) minor(any) length_hi
1905 $self->{rbuf} =~ /^(?:\z| \x16 (\z| [\x03\x04] (?:\z| . (?:\z| [\x00-\x40] ))))/xs
1906 # ssl2 comapatible: len_hi len_lo type(1) major minor dummy(forlength)
1907 or $self->{rbuf} =~ /^(?:\z| [\x80-\xff] (?:\z| . (?:\z| \x01 (\z| [\x03\x04] (?:\z| . (?:\z| . ))))))/xs
1908 ) {
1909 return if 3 != length $1; # partial match, can't decide yet
1910
1911 # full match, valid TLS record
1912 my ($major, $minor) = unpack "CC", $1;
1913 $cb->($self, "accept", $major + $minor * 0.1);
1914 } else {
1915 # mismatch == guaranteed not TLS
1916 $cb->($self, undef);
1917 }
1918
1919 1
1920 }
1921};
1922
1923register_read_type tls_autostart => sub {
1924 my ($self, @tls) = @_;
1925
1926 $RH{tls_detect}($self, sub {
1927 return unless $_[1];
1928 $_[0]->starttls (@tls);
1929 })
1930};
1931
1001=back 1932=back
1002 1933
1003=item AnyEvent::Handle::register_read_type type => $coderef->($handle, $cb, @args) 1934=item custom read types - Package::anyevent_read_type $handle, $cb, @args
1004 1935
1005This function (not method) lets you add your own types to C<push_read>. 1936Instead of one of the predefined types, you can also specify the name
1937of a package. AnyEvent will try to load the package and then expects to
1938find a function named C<anyevent_read_type> inside. If it isn't found, it
1939progressively tries to load the parent package until it either finds the
1940function (good) or runs out of packages (bad).
1006 1941
1007Whenever the given C<type> is used, C<push_read> will invoke the code 1942Whenever this type is used, C<push_read> will invoke the function with the
1008reference with the handle object, the callback and the remaining 1943handle object, the original callback and the remaining arguments.
1009arguments.
1010 1944
1011The code reference is supposed to return a callback (usually a closure) 1945The function is supposed to return a callback (usually a closure) that
1012that works as a plain read callback (see C<< ->push_read ($cb) >>). 1946works as a plain read callback (see C<< ->push_read ($cb) >>), so you can
1947mentally treat the function as a "configurable read type to read callback"
1948converter.
1013 1949
1014It should invoke the passed callback when it is done reading (remember to 1950It should invoke the original callback when it is done reading (remember
1015pass C<$handle> as first argument as all other callbacks do that). 1951to pass C<$handle> as first argument as all other callbacks do that,
1952although there is no strict requirement on this).
1016 1953
1017Note that this is a function, and all types registered this way will be
1018global, so try to use unique names.
1019
1020For examples, see the source of this module (F<perldoc -m AnyEvent::Handle>, 1954For examples, see the source of this module (F<perldoc -m
1021search for C<register_read_type>)). 1955AnyEvent::Handle>, search for C<register_read_type>)).
1022 1956
1023=item $handle->stop_read 1957=item $handle->stop_read
1024 1958
1025=item $handle->start_read 1959=item $handle->start_read
1026 1960
1027In rare cases you actually do not want to read anything from the 1961In rare cases you actually do not want to read anything from the
1028socket. In this case you can call C<stop_read>. Neither C<on_read> no 1962socket. In this case you can call C<stop_read>. Neither C<on_read> nor
1029any queued callbacks will be executed then. To start reading again, call 1963any queued callbacks will be executed then. To start reading again, call
1030C<start_read>. 1964C<start_read>.
1031 1965
1966Note that AnyEvent::Handle will automatically C<start_read> for you when
1967you change the C<on_read> callback or push/unshift a read callback, and it
1968will automatically C<stop_read> for you when neither C<on_read> is set nor
1969there are any read requests in the queue.
1970
1971In older versions of this module (<= 5.3), these methods had no effect,
1972as TLS does not support half-duplex connections. In current versions they
1973work as expected, as this behaviour is required to avoid certain resource
1974attacks, where the program would be forced to read (and buffer) arbitrary
1975amounts of data before being able to send some data. The drawback is that
1976some readings of the the SSL/TLS specifications basically require this
1977attack to be working, as SSL/TLS implementations might stall sending data
1978during a rehandshake.
1979
1980As a guideline, during the initial handshake, you should not stop reading,
1981and as a client, it might cause problems, depending on your application.
1982
1032=cut 1983=cut
1033 1984
1034sub stop_read { 1985sub stop_read {
1035 my ($self) = @_; 1986 my ($self) = @_;
1036 1987
1038} 1989}
1039 1990
1040sub start_read { 1991sub start_read {
1041 my ($self) = @_; 1992 my ($self) = @_;
1042 1993
1043 unless ($self->{_rw} || $self->{_eof}) { 1994 unless ($self->{_rw} || $self->{_eof} || !$self->{fh}) {
1044 Scalar::Util::weaken $self; 1995 Scalar::Util::weaken $self;
1045 1996
1046 $self->{_rw} = AnyEvent->io (fh => $self->{fh}, poll => "r", cb => sub { 1997 $self->{_rw} = AE::io $self->{fh}, 0, sub {
1047 my $rbuf = $self->{filter_r} ? \my $buf : \$self->{rbuf}; 1998 my $rbuf = \($self->{tls} ? my $buf : $self->{rbuf});
1048 my $len = sysread $self->{fh}, $$rbuf, $self->{read_size} || 8192, length $$rbuf; 1999 my $len = sysread $self->{fh}, $$rbuf, $self->{read_size}, length $$rbuf;
1049 2000
1050 if ($len > 0) { 2001 if ($len > 0) {
1051 $self->{_activity} = AnyEvent->now; 2002 $self->{_activity} = $self->{_ractivity} = AE::now;
1052 2003
1053 $self->{filter_r} 2004 if ($self->{tls}) {
1054 ? $self->{filter_r}($self, $rbuf) 2005 Net::SSLeay::BIO_write ($self->{_rbio}, $$rbuf);
2006
2007 &_dotls ($self);
2008 } else {
1055 : $self->_drain_rbuf; 2009 $self->_drain_rbuf;
2010 }
2011
2012 if ($len == $self->{read_size}) {
2013 $self->{read_size} *= 2;
2014 $self->{read_size} = $self->{max_read_size} || MAX_READ_SIZE
2015 if $self->{read_size} > ($self->{max_read_size} || MAX_READ_SIZE);
2016 }
1056 2017
1057 } elsif (defined $len) { 2018 } elsif (defined $len) {
1058 delete $self->{_rw}; 2019 delete $self->{_rw};
1059 $self->{_eof} = 1; 2020 $self->{_eof} = 1;
1060 $self->_drain_rbuf; 2021 $self->_drain_rbuf;
1061 2022
1062 } elsif ($! != EAGAIN && $! != EINTR && $! != WSAEWOULDBLOCK) { 2023 } elsif ($! != EAGAIN && $! != EINTR && $! != WSAEWOULDBLOCK) {
1063 return $self->_error ($!, 1); 2024 return $self->_error ($!, 1);
1064 } 2025 }
1065 }); 2026 };
1066 } 2027 }
1067} 2028}
1068 2029
2030our $ERROR_SYSCALL;
2031our $ERROR_WANT_READ;
2032
2033sub _tls_error {
2034 my ($self, $err) = @_;
2035
2036 return $self->_error ($!, 1)
2037 if $err == Net::SSLeay::ERROR_SYSCALL ();
2038
2039 my $err = Net::SSLeay::ERR_error_string (Net::SSLeay::ERR_get_error ());
2040
2041 # reduce error string to look less scary
2042 $err =~ s/^error:[0-9a-fA-F]{8}:[^:]+:([^:]+):/\L$1: /;
2043
2044 if ($self->{_on_starttls}) {
2045 (delete $self->{_on_starttls})->($self, undef, $err);
2046 &_freetls;
2047 } else {
2048 &_freetls;
2049 $self->_error (Errno::EPROTO, 1, $err);
2050 }
2051}
2052
2053# poll the write BIO and send the data if applicable
2054# also decode read data if possible
2055# this is basiclaly our TLS state machine
2056# more efficient implementations are possible with openssl,
2057# but not with the buggy and incomplete Net::SSLeay.
1069sub _dotls { 2058sub _dotls {
1070 my ($self) = @_; 2059 my ($self) = @_;
1071 2060
2061 my $tmp;
2062
1072 if (length $self->{_tls_wbuf}) { 2063 while (length $self->{_tls_wbuf}) {
1073 while ((my $len = Net::SSLeay::write ($self->{tls}, $self->{_tls_wbuf})) > 0) { 2064 if (($tmp = Net::SSLeay::write ($self->{tls}, $self->{_tls_wbuf})) <= 0) {
1074 substr $self->{_tls_wbuf}, 0, $len, ""; 2065 $tmp = Net::SSLeay::get_error ($self->{tls}, $tmp);
2066
2067 return $self->_tls_error ($tmp)
2068 if $tmp != $ERROR_WANT_READ
2069 && ($tmp != $ERROR_SYSCALL || $!);
2070
2071 last;
1075 } 2072 }
1076 }
1077 2073
2074 substr $self->{_tls_wbuf}, 0, $tmp, "";
2075 }
2076
2077 while (defined ($tmp = Net::SSLeay::read ($self->{tls}))) {
2078 unless (length $tmp) {
2079 $self->{_on_starttls}
2080 and (delete $self->{_on_starttls})->($self, undef, "EOF during handshake"); # ???
2081 &_freetls;
2082
2083 if ($self->{on_stoptls}) {
2084 $self->{on_stoptls}($self);
2085 return;
2086 } else {
2087 # let's treat SSL-eof as we treat normal EOF
2088 delete $self->{_rw};
2089 $self->{_eof} = 1;
2090 }
2091 }
2092
2093 $self->{_tls_rbuf} .= $tmp;
2094 $self->_drain_rbuf;
2095 $self->{tls} or return; # tls session might have gone away in callback
2096 }
2097
2098 $tmp = Net::SSLeay::get_error ($self->{tls}, -1); # -1 is not neccessarily correct, but Net::SSLeay doesn't tell us
2099 return $self->_tls_error ($tmp)
2100 if $tmp != $ERROR_WANT_READ
2101 && ($tmp != $ERROR_SYSCALL || $!);
2102
1078 if (defined (my $buf = Net::SSLeay::BIO_read ($self->{_wbio}))) { 2103 while (length ($tmp = Net::SSLeay::BIO_read ($self->{_wbio}))) {
1079 $self->{wbuf} .= $buf; 2104 $self->{wbuf} .= $tmp;
1080 $self->_drain_wbuf; 2105 $self->_drain_wbuf;
2106 $self->{tls} or return; # tls session might have gone away in callback
1081 } 2107 }
1082 2108
1083 while (defined (my $buf = Net::SSLeay::read ($self->{tls}))) { 2109 $self->{_on_starttls}
1084 $self->{rbuf} .= $buf; 2110 and Net::SSLeay::state ($self->{tls}) == Net::SSLeay::ST_OK ()
1085 $self->_drain_rbuf; 2111 and (delete $self->{_on_starttls})->($self, 1, "TLS/SSL connection established");
1086 }
1087
1088 my $err = Net::SSLeay::get_error ($self->{tls}, -1);
1089
1090 if ($err!= Net::SSLeay::ERROR_WANT_READ ()) {
1091 if ($err == Net::SSLeay::ERROR_SYSCALL ()) {
1092 return $self->_error ($!, 1);
1093 } elsif ($err == Net::SSLeay::ERROR_SSL ()) {
1094 return $self->_error (&Errno::EIO, 1);
1095 }
1096
1097 # all others are fine for our purposes
1098 }
1099} 2112}
1100 2113
1101=item $handle->starttls ($tls[, $tls_ctx]) 2114=item $handle->starttls ($tls[, $tls_ctx])
1102 2115
1103Instead of starting TLS negotiation immediately when the AnyEvent::Handle 2116Instead of starting TLS negotiation immediately when the AnyEvent::Handle
1104object is created, you can also do that at a later time by calling 2117object is created, you can also do that at a later time by calling
1105C<starttls>. 2118C<starttls>. See the C<tls> constructor argument for general info.
2119
2120Starting TLS is currently an asynchronous operation - when you push some
2121write data and then call C<< ->starttls >> then TLS negotiation will start
2122immediately, after which the queued write data is then sent. This might
2123change in future versions, so best make sure you have no outstanding write
2124data when calling this method.
1106 2125
1107The first argument is the same as the C<tls> constructor argument (either 2126The first argument is the same as the C<tls> constructor argument (either
1108C<"connect">, C<"accept"> or an existing Net::SSLeay object). 2127C<"connect">, C<"accept"> or an existing Net::SSLeay object).
1109 2128
1110The second argument is the optional C<Net::SSLeay::CTX> object that is 2129The second argument is the optional C<AnyEvent::TLS> object that is used
1111used when AnyEvent::Handle has to create its own TLS connection object. 2130when AnyEvent::Handle has to create its own TLS connection object, or
2131a hash reference with C<< key => value >> pairs that will be used to
2132construct a new context.
1112 2133
1113The TLS connection object will end up in C<< $handle->{tls} >> after this 2134The TLS connection object will end up in C<< $handle->{tls} >>, the TLS
1114call and can be used or changed to your liking. Note that the handshake 2135context in C<< $handle->{tls_ctx} >> after this call and can be used or
1115might have already started when this function returns. 2136changed to your liking. Note that the handshake might have already started
2137when this function returns.
1116 2138
2139Due to bugs in OpenSSL, it might or might not be possible to do multiple
2140handshakes on the same stream. It is best to not attempt to use the
2141stream after stopping TLS.
2142
2143This method may invoke callbacks (and therefore the handle might be
2144destroyed after it returns).
2145
1117=cut 2146=cut
2147
2148our %TLS_CACHE; #TODO not yet documented, should we?
1118 2149
1119sub starttls { 2150sub starttls {
1120 my ($self, $ssl, $ctx) = @_; 2151 my ($self, $tls, $ctx) = @_;
1121 2152
1122 $self->stoptls; 2153 Carp::croak "It is an error to call starttls on an AnyEvent::Handle object while TLS is already active, caught"
2154 if $self->{tls};
1123 2155
1124 if ($ssl eq "accept") { 2156 unless (defined $AnyEvent::TLS::VERSION) {
1125 $ssl = Net::SSLeay::new ($ctx || TLS_CTX ()); 2157 eval {
1126 Net::SSLeay::set_accept_state ($ssl); 2158 require Net::SSLeay;
1127 } elsif ($ssl eq "connect") { 2159 require AnyEvent::TLS;
1128 $ssl = Net::SSLeay::new ($ctx || TLS_CTX ()); 2160 1
1129 Net::SSLeay::set_connect_state ($ssl); 2161 } or return $self->_error (Errno::EPROTO, 1, "TLS support not available on this system");
2162 }
2163
2164 $self->{tls} = $tls;
2165 $self->{tls_ctx} = $ctx if @_ > 2;
2166
2167 return unless $self->{fh};
2168
2169 $ERROR_SYSCALL = Net::SSLeay::ERROR_SYSCALL ();
2170 $ERROR_WANT_READ = Net::SSLeay::ERROR_WANT_READ ();
2171
2172 $tls = delete $self->{tls};
2173 $ctx = $self->{tls_ctx};
2174
2175 local $Carp::CarpLevel = 1; # skip ourselves when creating a new context or session
2176
2177 if ("HASH" eq ref $ctx) {
2178 if ($ctx->{cache}) {
2179 my $key = $ctx+0;
2180 $ctx = $TLS_CACHE{$key} ||= new AnyEvent::TLS %$ctx;
2181 } else {
2182 $ctx = new AnyEvent::TLS %$ctx;
2183 }
2184 }
1130 } 2185
1131 2186 $self->{tls_ctx} = $ctx || TLS_CTX ();
1132 $self->{tls} = $ssl; 2187 $self->{tls} = $tls = $self->{tls_ctx}->_get_session ($tls, $self, $self->{peername});
1133 2188
1134 # basically, this is deep magic (because SSL_read should have the same issues) 2189 # basically, this is deep magic (because SSL_read should have the same issues)
1135 # but the openssl maintainers basically said: "trust us, it just works". 2190 # but the openssl maintainers basically said: "trust us, it just works".
1136 # (unfortunately, we have to hardcode constants because the abysmally misdesigned 2191 # (unfortunately, we have to hardcode constants because the abysmally misdesigned
1137 # and mismaintained ssleay-module doesn't even offer them). 2192 # and mismaintained ssleay-module doesn't even offer them).
1138 # http://www.mail-archive.com/openssl-dev@openssl.org/msg22420.html 2193 # http://www.mail-archive.com/openssl-dev@openssl.org/msg22420.html
2194 #
2195 # in short: this is a mess.
2196 #
2197 # note that we do not try to keep the length constant between writes as we are required to do.
2198 # we assume that most (but not all) of this insanity only applies to non-blocking cases,
2199 # and we drive openssl fully in blocking mode here. Or maybe we don't - openssl seems to
2200 # have identity issues in that area.
1139 Net::SSLeay::CTX_set_mode ($self->{tls}, 2201# Net::SSLeay::CTX_set_mode ($ssl,
1140 (eval { local $SIG{__DIE__}; Net::SSLeay::MODE_ENABLE_PARTIAL_WRITE () } || 1) 2202# (eval { local $SIG{__DIE__}; Net::SSLeay::MODE_ENABLE_PARTIAL_WRITE () } || 1)
1141 | (eval { local $SIG{__DIE__}; Net::SSLeay::MODE_ACCEPT_MOVING_WRITE_BUFFER () } || 2)); 2203# | (eval { local $SIG{__DIE__}; Net::SSLeay::MODE_ACCEPT_MOVING_WRITE_BUFFER () } || 2));
2204 Net::SSLeay::CTX_set_mode ($tls, 1|2);
1142 2205
1143 $self->{_rbio} = Net::SSLeay::BIO_new (Net::SSLeay::BIO_s_mem ()); 2206 $self->{_rbio} = Net::SSLeay::BIO_new (Net::SSLeay::BIO_s_mem ());
1144 $self->{_wbio} = Net::SSLeay::BIO_new (Net::SSLeay::BIO_s_mem ()); 2207 $self->{_wbio} = Net::SSLeay::BIO_new (Net::SSLeay::BIO_s_mem ());
1145 2208
2209 Net::SSLeay::BIO_write ($self->{_rbio}, $self->{rbuf});
2210 $self->{rbuf} = "";
2211
1146 Net::SSLeay::set_bio ($ssl, $self->{_rbio}, $self->{_wbio}); 2212 Net::SSLeay::set_bio ($tls, $self->{_rbio}, $self->{_wbio});
1147 2213
1148 $self->{filter_w} = sub { 2214 $self->{_on_starttls} = sub { $_[0]{on_starttls}(@_) }
1149 $_[0]{_tls_wbuf} .= ${$_[1]}; 2215 if $self->{on_starttls};
1150 &_dotls; 2216
1151 }; 2217 &_dotls; # need to trigger the initial handshake
1152 $self->{filter_r} = sub { 2218 $self->start_read; # make sure we actually do read
1153 Net::SSLeay::BIO_write ($_[0]{_rbio}, ${$_[1]});
1154 &_dotls;
1155 };
1156} 2219}
1157 2220
1158=item $handle->stoptls 2221=item $handle->stoptls
1159 2222
1160Destroys the SSL connection, if any. Partial read or write data will be 2223Shuts down the SSL connection - this makes a proper EOF handshake by
1161lost. 2224sending a close notify to the other side, but since OpenSSL doesn't
2225support non-blocking shut downs, it is not guaranteed that you can re-use
2226the stream afterwards.
2227
2228This method may invoke callbacks (and therefore the handle might be
2229destroyed after it returns).
1162 2230
1163=cut 2231=cut
1164 2232
1165sub stoptls { 2233sub stoptls {
1166 my ($self) = @_; 2234 my ($self) = @_;
1167 2235
1168 Net::SSLeay::free (delete $self->{tls}) if $self->{tls}; 2236 if ($self->{tls} && $self->{fh}) {
2237 Net::SSLeay::shutdown ($self->{tls});
1169 2238
1170 delete $self->{_rbio}; 2239 &_dotls;
1171 delete $self->{_wbio}; 2240
1172 delete $self->{_tls_wbuf}; 2241# # we don't give a shit. no, we do, but we can't. no...#d#
1173 delete $self->{filter_r}; 2242# # we, we... have to use openssl :/#d#
1174 delete $self->{filter_w}; 2243# &_freetls;#d#
2244 }
1175} 2245}
2246
2247sub _freetls {
2248 my ($self) = @_;
2249
2250 return unless $self->{tls};
2251
2252 $self->{tls_ctx}->_put_session (delete $self->{tls})
2253 if $self->{tls} > 0;
2254
2255 delete @$self{qw(_rbio _wbio _tls_wbuf _on_starttls)};
2256}
2257
2258=item $handle->resettls
2259
2260This rarely-used method simply resets and TLS state on the handle, usually
2261causing data loss.
2262
2263One case where it may be useful is when you want to skip over the data in
2264the stream but you are not interested in interpreting it, so data loss is
2265no concern.
2266
2267=cut
2268
2269*resettls = \&_freetls;
1176 2270
1177sub DESTROY { 2271sub DESTROY {
1178 my $self = shift; 2272 my ($self) = @_;
1179 2273
1180 $self->stoptls; 2274 &_freetls;
2275
2276 my $linger = exists $self->{linger} ? $self->{linger} : 3600;
2277
2278 if ($linger && length $self->{wbuf} && $self->{fh}) {
2279 my $fh = delete $self->{fh};
2280 my $wbuf = delete $self->{wbuf};
2281
2282 my @linger;
2283
2284 push @linger, AE::io $fh, 1, sub {
2285 my $len = syswrite $fh, $wbuf, length $wbuf;
2286
2287 if ($len > 0) {
2288 substr $wbuf, 0, $len, "";
2289 } elsif (defined $len || ($! != EAGAIN && $! != EINTR && $! != WSAEWOULDBLOCK)) {
2290 @linger = (); # end
2291 }
2292 };
2293 push @linger, AE::timer $linger, 0, sub {
2294 @linger = ();
2295 };
2296 }
1181} 2297}
2298
2299=item $handle->destroy
2300
2301Shuts down the handle object as much as possible - this call ensures that
2302no further callbacks will be invoked and as many resources as possible
2303will be freed. Any method you will call on the handle object after
2304destroying it in this way will be silently ignored (and it will return the
2305empty list).
2306
2307Normally, you can just "forget" any references to an AnyEvent::Handle
2308object and it will simply shut down. This works in fatal error and EOF
2309callbacks, as well as code outside. It does I<NOT> work in a read or write
2310callback, so when you want to destroy the AnyEvent::Handle object from
2311within such an callback. You I<MUST> call C<< ->destroy >> explicitly in
2312that case.
2313
2314Destroying the handle object in this way has the advantage that callbacks
2315will be removed as well, so if those are the only reference holders (as
2316is common), then one doesn't need to do anything special to break any
2317reference cycles.
2318
2319The handle might still linger in the background and write out remaining
2320data, as specified by the C<linger> option, however.
2321
2322=cut
2323
2324sub destroy {
2325 my ($self) = @_;
2326
2327 $self->DESTROY;
2328 %$self = ();
2329 bless $self, "AnyEvent::Handle::destroyed";
2330}
2331
2332sub AnyEvent::Handle::destroyed::AUTOLOAD {
2333 #nop
2334}
2335
2336=item $handle->destroyed
2337
2338Returns false as long as the handle hasn't been destroyed by a call to C<<
2339->destroy >>, true otherwise.
2340
2341Can be useful to decide whether the handle is still valid after some
2342callback possibly destroyed the handle. For example, C<< ->push_write >>,
2343C<< ->starttls >> and other methods can call user callbacks, which in turn
2344can destroy the handle, so work can be avoided by checking sometimes:
2345
2346 $hdl->starttls ("accept");
2347 return if $hdl->destroyed;
2348 $hdl->push_write (...
2349
2350Note that the call to C<push_write> will silently be ignored if the handle
2351has been destroyed, so often you can just ignore the possibility of the
2352handle being destroyed.
2353
2354=cut
2355
2356sub destroyed { 0 }
2357sub AnyEvent::Handle::destroyed::destroyed { 1 }
1182 2358
1183=item AnyEvent::Handle::TLS_CTX 2359=item AnyEvent::Handle::TLS_CTX
1184 2360
1185This function creates and returns the Net::SSLeay::CTX object used by 2361This function creates and returns the AnyEvent::TLS object used by default
1186default for TLS mode. 2362for TLS mode.
1187 2363
1188The context is created like this: 2364The context is created by calling L<AnyEvent::TLS> without any arguments.
1189
1190 Net::SSLeay::load_error_strings;
1191 Net::SSLeay::SSLeay_add_ssl_algorithms;
1192 Net::SSLeay::randomize;
1193
1194 my $CTX = Net::SSLeay::CTX_new;
1195
1196 Net::SSLeay::CTX_set_options $CTX, Net::SSLeay::OP_ALL
1197 2365
1198=cut 2366=cut
1199 2367
1200our $TLS_CTX; 2368our $TLS_CTX;
1201 2369
1202sub TLS_CTX() { 2370sub TLS_CTX() {
1203 $TLS_CTX || do { 2371 $TLS_CTX ||= do {
1204 require Net::SSLeay; 2372 require AnyEvent::TLS;
1205 2373
1206 Net::SSLeay::load_error_strings (); 2374 new AnyEvent::TLS
1207 Net::SSLeay::SSLeay_add_ssl_algorithms (); 2375 }
1208 Net::SSLeay::randomize (); 2376}
1209 2377
1210 $TLS_CTX = Net::SSLeay::CTX_new (); 2378=back
1211 2379
1212 Net::SSLeay::CTX_set_options ($TLS_CTX, Net::SSLeay::OP_ALL ());
1213 2380
1214 $TLS_CTX 2381=head1 NONFREQUENTLY ASKED QUESTIONS
2382
2383=over 4
2384
2385=item I C<undef> the AnyEvent::Handle reference inside my callback and
2386still get further invocations!
2387
2388That's because AnyEvent::Handle keeps a reference to itself when handling
2389read or write callbacks.
2390
2391It is only safe to "forget" the reference inside EOF or error callbacks,
2392from within all other callbacks, you need to explicitly call the C<<
2393->destroy >> method.
2394
2395=item Why is my C<on_eof> callback never called?
2396
2397Probably because your C<on_error> callback is being called instead: When
2398you have outstanding requests in your read queue, then an EOF is
2399considered an error as you clearly expected some data.
2400
2401To avoid this, make sure you have an empty read queue whenever your handle
2402is supposed to be "idle" (i.e. connection closes are O.K.). You can set
2403an C<on_read> handler that simply pushes the first read requests in the
2404queue.
2405
2406See also the next question, which explains this in a bit more detail.
2407
2408=item How can I serve requests in a loop?
2409
2410Most protocols consist of some setup phase (authentication for example)
2411followed by a request handling phase, where the server waits for requests
2412and handles them, in a loop.
2413
2414There are two important variants: The first (traditional, better) variant
2415handles requests until the server gets some QUIT command, causing it to
2416close the connection first (highly desirable for a busy TCP server). A
2417client dropping the connection is an error, which means this variant can
2418detect an unexpected detection close.
2419
2420To handle this case, always make sure you have a non-empty read queue, by
2421pushing the "read request start" handler on it:
2422
2423 # we assume a request starts with a single line
2424 my @start_request; @start_request = (line => sub {
2425 my ($hdl, $line) = @_;
2426
2427 ... handle request
2428
2429 # push next request read, possibly from a nested callback
2430 $hdl->push_read (@start_request);
2431 });
2432
2433 # auth done, now go into request handling loop
2434 # now push the first @start_request
2435 $hdl->push_read (@start_request);
2436
2437By always having an outstanding C<push_read>, the handle always expects
2438some data and raises the C<EPIPE> error when the connction is dropped
2439unexpectedly.
2440
2441The second variant is a protocol where the client can drop the connection
2442at any time. For TCP, this means that the server machine may run out of
2443sockets easier, and in general, it means you cannot distinguish a protocl
2444failure/client crash from a normal connection close. Nevertheless, these
2445kinds of protocols are common (and sometimes even the best solution to the
2446problem).
2447
2448Having an outstanding read request at all times is possible if you ignore
2449C<EPIPE> errors, but this doesn't help with when the client drops the
2450connection during a request, which would still be an error.
2451
2452A better solution is to push the initial request read in an C<on_read>
2453callback. This avoids an error, as when the server doesn't expect data
2454(i.e. is idly waiting for the next request, an EOF will not raise an
2455error, but simply result in an C<on_eof> callback. It is also a bit slower
2456and simpler:
2457
2458 # auth done, now go into request handling loop
2459 $hdl->on_read (sub {
2460 my ($hdl) = @_;
2461
2462 # called each time we receive data but the read queue is empty
2463 # simply start read the request
2464
2465 $hdl->push_read (line => sub {
2466 my ($hdl, $line) = @_;
2467
2468 ... handle request
2469
2470 # do nothing special when the request has been handled, just
2471 # let the request queue go empty.
2472 });
2473 });
2474
2475=item I get different callback invocations in TLS mode/Why can't I pause
2476reading?
2477
2478Unlike, say, TCP, TLS connections do not consist of two independent
2479communication channels, one for each direction. Or put differently, the
2480read and write directions are not independent of each other: you cannot
2481write data unless you are also prepared to read, and vice versa.
2482
2483This means that, in TLS mode, you might get C<on_error> or C<on_eof>
2484callback invocations when you are not expecting any read data - the reason
2485is that AnyEvent::Handle always reads in TLS mode.
2486
2487During the connection, you have to make sure that you always have a
2488non-empty read-queue, or an C<on_read> watcher. At the end of the
2489connection (or when you no longer want to use it) you can call the
2490C<destroy> method.
2491
2492=item How do I read data until the other side closes the connection?
2493
2494If you just want to read your data into a perl scalar, the easiest way
2495to achieve this is by setting an C<on_read> callback that does nothing,
2496clearing the C<on_eof> callback and in the C<on_error> callback, the data
2497will be in C<$_[0]{rbuf}>:
2498
2499 $handle->on_read (sub { });
2500 $handle->on_eof (undef);
2501 $handle->on_error (sub {
2502 my $data = delete $_[0]{rbuf};
2503 });
2504
2505Note that this example removes the C<rbuf> member from the handle object,
2506which is not normally allowed by the API. It is expressly permitted in
2507this case only, as the handle object needs to be destroyed afterwards.
2508
2509The reason to use C<on_error> is that TCP connections, due to latencies
2510and packets loss, might get closed quite violently with an error, when in
2511fact all data has been received.
2512
2513It is usually better to use acknowledgements when transferring data,
2514to make sure the other side hasn't just died and you got the data
2515intact. This is also one reason why so many internet protocols have an
2516explicit QUIT command.
2517
2518=item I don't want to destroy the handle too early - how do I wait until
2519all data has been written?
2520
2521After writing your last bits of data, set the C<on_drain> callback
2522and destroy the handle in there - with the default setting of
2523C<low_water_mark> this will be called precisely when all data has been
2524written to the socket:
2525
2526 $handle->push_write (...);
2527 $handle->on_drain (sub {
2528 AE::log debug => "All data submitted to the kernel.";
2529 undef $handle;
2530 });
2531
2532If you just want to queue some data and then signal EOF to the other side,
2533consider using C<< ->push_shutdown >> instead.
2534
2535=item I want to contact a TLS/SSL server, I don't care about security.
2536
2537If your TLS server is a pure TLS server (e.g. HTTPS) that only speaks TLS,
2538connect to it and then create the AnyEvent::Handle with the C<tls>
2539parameter:
2540
2541 tcp_connect $host, $port, sub {
2542 my ($fh) = @_;
2543
2544 my $handle = new AnyEvent::Handle
2545 fh => $fh,
2546 tls => "connect",
2547 on_error => sub { ... };
2548
2549 $handle->push_write (...);
1215 } 2550 };
1216} 2551
2552=item I want to contact a TLS/SSL server, I do care about security.
2553
2554Then you should additionally enable certificate verification, including
2555peername verification, if the protocol you use supports it (see
2556L<AnyEvent::TLS>, C<verify_peername>).
2557
2558E.g. for HTTPS:
2559
2560 tcp_connect $host, $port, sub {
2561 my ($fh) = @_;
2562
2563 my $handle = new AnyEvent::Handle
2564 fh => $fh,
2565 peername => $host,
2566 tls => "connect",
2567 tls_ctx => { verify => 1, verify_peername => "https" },
2568 ...
2569
2570Note that you must specify the hostname you connected to (or whatever
2571"peername" the protocol needs) as the C<peername> argument, otherwise no
2572peername verification will be done.
2573
2574The above will use the system-dependent default set of trusted CA
2575certificates. If you want to check against a specific CA, add the
2576C<ca_file> (or C<ca_cert>) arguments to C<tls_ctx>:
2577
2578 tls_ctx => {
2579 verify => 1,
2580 verify_peername => "https",
2581 ca_file => "my-ca-cert.pem",
2582 },
2583
2584=item I want to create a TLS/SSL server, how do I do that?
2585
2586Well, you first need to get a server certificate and key. You have
2587three options: a) ask a CA (buy one, use cacert.org etc.) b) create a
2588self-signed certificate (cheap. check the search engine of your choice,
2589there are many tutorials on the net) or c) make your own CA (tinyca2 is a
2590nice program for that purpose).
2591
2592Then create a file with your private key (in PEM format, see
2593L<AnyEvent::TLS>), followed by the certificate (also in PEM format). The
2594file should then look like this:
2595
2596 -----BEGIN RSA PRIVATE KEY-----
2597 ...header data
2598 ... lots of base64'y-stuff
2599 -----END RSA PRIVATE KEY-----
2600
2601 -----BEGIN CERTIFICATE-----
2602 ... lots of base64'y-stuff
2603 -----END CERTIFICATE-----
2604
2605The important bits are the "PRIVATE KEY" and "CERTIFICATE" parts. Then
2606specify this file as C<cert_file>:
2607
2608 tcp_server undef, $port, sub {
2609 my ($fh) = @_;
2610
2611 my $handle = new AnyEvent::Handle
2612 fh => $fh,
2613 tls => "accept",
2614 tls_ctx => { cert_file => "my-server-keycert.pem" },
2615 ...
2616
2617When you have intermediate CA certificates that your clients might not
2618know about, just append them to the C<cert_file>.
1217 2619
1218=back 2620=back
1219 2621
1220=head1 SUBCLASSING AnyEvent::Handle 2622=head1 SUBCLASSING AnyEvent::Handle
1221 2623
1227=over 4 2629=over 4
1228 2630
1229=item * all constructor arguments become object members. 2631=item * all constructor arguments become object members.
1230 2632
1231At least initially, when you pass a C<tls>-argument to the constructor it 2633At least initially, when you pass a C<tls>-argument to the constructor it
1232will end up in C<< $handle->{tls} >>. Those members might be changes or 2634will end up in C<< $handle->{tls} >>. Those members might be changed or
1233mutated later on (for example C<tls> will hold the TLS connection object). 2635mutated later on (for example C<tls> will hold the TLS connection object).
1234 2636
1235=item * other object member names are prefixed with an C<_>. 2637=item * other object member names are prefixed with an C<_>.
1236 2638
1237All object members not explicitly documented (internal use) are prefixed 2639All object members not explicitly documented (internal use) are prefixed
1240 2642
1241=item * all members not documented here and not prefixed with an underscore 2643=item * all members not documented here and not prefixed with an underscore
1242are free to use in subclasses. 2644are free to use in subclasses.
1243 2645
1244Of course, new versions of AnyEvent::Handle may introduce more "public" 2646Of course, new versions of AnyEvent::Handle may introduce more "public"
1245member variables, but thats just life, at least it is documented. 2647member variables, but that's just life. At least it is documented.
1246 2648
1247=back 2649=back
1248 2650
1249=head1 AUTHOR 2651=head1 AUTHOR
1250 2652
1251Robin Redeker C<< <elmex at ta-sa.org> >>, Marc Lehmann <schmorp@schmorp.de>. 2653Robin Redeker C<< <elmex at ta-sa.org> >>, Marc Lehmann <schmorp@schmorp.de>.
1252 2654
1253=cut 2655=cut
1254 2656
12551; # End of AnyEvent::Handle 26571
2658

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines