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 1.159 by root, Fri Jul 24 12:35:58 2009 UTC vs.
Revision 1.228 by root, Mon Feb 6 00:17:26 2012 UTC

1package AnyEvent::Handle;
2
3use Scalar::Util ();
4use Carp ();
5use Errno qw(EAGAIN EINTR);
6
7use AnyEvent (); BEGIN { AnyEvent::common_sense }
8use AnyEvent::Util qw(WSAEWOULDBLOCK);
9
10=head1 NAME 1=head1 NAME
11 2
12AnyEvent::Handle - non-blocking I/O on file handles via AnyEvent 3AnyEvent::Handle - non-blocking I/O on streaming handles via AnyEvent
13
14=cut
15
16our $VERSION = 4.86;
17 4
18=head1 SYNOPSIS 5=head1 SYNOPSIS
19 6
20 use AnyEvent; 7 use AnyEvent;
21 use AnyEvent::Handle; 8 use AnyEvent::Handle;
24 11
25 my $hdl; $hdl = new AnyEvent::Handle 12 my $hdl; $hdl = new AnyEvent::Handle
26 fh => \*STDIN, 13 fh => \*STDIN,
27 on_error => sub { 14 on_error => sub {
28 my ($hdl, $fatal, $msg) = @_; 15 my ($hdl, $fatal, $msg) = @_;
29 warn "got error $msg\n"; 16 AE::log error => "got error $msg\n";
30 $hdl->destroy; 17 $hdl->destroy;
31 $cv->send; 18 $cv->send;
32 ); 19 };
33 20
34 # send some request line 21 # send some request line
35 $hdl->push_write ("getinfo\015\012"); 22 $hdl->push_write ("getinfo\015\012");
36 23
37 # read the response line 24 # read the response line
38 $hdl->push_read (line => sub { 25 $hdl->push_read (line => sub {
39 my ($hdl, $line) = @_; 26 my ($hdl, $line) = @_;
40 warn "got line <$line>\n"; 27 say "got line <$line>";
41 $cv->send; 28 $cv->send;
42 }); 29 });
43 30
44 $cv->recv; 31 $cv->recv;
45 32
46=head1 DESCRIPTION 33=head1 DESCRIPTION
47 34
48This 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
49filehandles. 36stream-based filehandles (sockets, pipes, and other stream things).
50 37
51The L<AnyEvent::Intro> tutorial contains some well-documented 38The L<AnyEvent::Intro> tutorial contains some well-documented
52AnyEvent::Handle examples. 39AnyEvent::Handle examples.
53 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.
57 44
58At the very minimum, you should specify C<fh> or C<connect>, and the 45At the very minimum, you should specify C<fh> or C<connect>, and the
59C<on_error> callback. 46C<on_error> callback.
60 47
61All callbacks will be invoked with the handle object as their first 48All callbacks will be invoked with the handle object as their first
62argument. 49argument.
63 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
64=head1 METHODS 82=head1 METHODS
65 83
66=over 4 84=over 4
67 85
68=item $handle = B<new> AnyEvent::TLS fh => $filehandle, key => value... 86=item $handle = B<new> AnyEvent::Handle fh => $filehandle, key => value...
69 87
70The constructor supports these arguments (all as C<< key => value >> pairs). 88The constructor supports these arguments (all as C<< key => value >> pairs).
71 89
72=over 4 90=over 4
73 91
84C<AnyEvent::Socket::tcp_connect>. The C<$host> additionally becomes the 102C<AnyEvent::Socket::tcp_connect>. The C<$host> additionally becomes the
85default C<peername>. 103default C<peername>.
86 104
87You have to specify either this parameter, or C<fh>, above. 105You have to specify either this parameter, or C<fh>, above.
88 106
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
89When this parameter is specified, then the C<on_prepare>, 110When this parameter is specified, then the C<on_prepare>,
90C<on_connect_error> and C<on_connect> callbacks will be called under the 111C<on_connect_error> and C<on_connect> callbacks will be called under the
91appropriate circumstances: 112appropriate circumstances:
92 113
93=over 4 114=over 4
94 115
95=item on_prepare => $cb->($handle) 116=item on_prepare => $cb->($handle)
96 117
97This (rarely used) callback is called before a new connection is 118This (rarely used) callback is called before a new connection is
98attempted, but after the file handle has been created. It could be used to 119attempted, but after the file handle has been created (you can access that
120file handle via C<< $handle->{fh} >>). It could be used to prepare the
99prepare the file handle with parameters required for the actual connect 121file handle with parameters required for the actual connect (as opposed to
100(as opposed to settings that can be changed when the connection is already 122settings that can be changed when the connection is already established).
101established). 123
124The return value of this callback should be the connect timeout value in
125seconds (or C<0>, or C<undef>, or the empty list, to indicate that the
126default timeout is to be used).
102 127
103=item on_connect => $cb->($handle, $host, $port, $retry->()) 128=item on_connect => $cb->($handle, $host, $port, $retry->())
104 129
105This callback is called when a connection has been successfully established. 130This callback is called when a connection has been successfully established.
106 131
107The actual numeric host and port (the socket peername) are passed as 132The peer's numeric host and port (the socket peername) are passed as
108parameters, together with a retry callback. 133parameters, together with a retry callback. At the time it is called the
134read and write queues, EOF status, TLS status and similar properties of
135the handle will have been reset.
109 136
137It is not allowed to use the read or write queues while the handle object
138is connecting.
139
110When, for some reason, the handle is not acceptable, then calling 140If, for some reason, the handle is not acceptable, calling C<$retry> will
111C<$retry> will continue with the next conenction target (in case of 141continue with the next connection target (in case of multi-homed hosts or
112multi-homed hosts or SRV records there can be multiple connection 142SRV records there can be multiple connection endpoints). The C<$retry>
113endpoints). When it is called then the read and write queues, eof status, 143callback can be invoked after the connect callback returns, i.e. one can
114tls status and similar properties of the handle are being reset. 144start a handshake and then decide to retry with the next host if the
145handshake fails.
115 146
116In most cases, ignoring the C<$retry> parameter is the way to go. 147In most cases, you should ignore the C<$retry> parameter.
117 148
118=item on_connect_error => $cb->($handle, $message) 149=item on_connect_error => $cb->($handle, $message)
119 150
120This callback is called when the conenction could not be 151This callback is called when the connection could not be
121established. C<$!> will contain the relevant error code, and C<$message> a 152established. C<$!> will contain the relevant error code, and C<$message> a
122message describing it (usually the same as C<"$!">). 153message describing it (usually the same as C<"$!">).
123 154
124If this callback isn't specified, then C<on_error> will be called with a 155If this callback isn't specified, then C<on_error> will be called with a
125fatal error instead. 156fatal error instead.
128 159
129=item on_error => $cb->($handle, $fatal, $message) 160=item on_error => $cb->($handle, $fatal, $message)
130 161
131This is the error callback, which is called when, well, some error 162This is the error callback, which is called when, well, some error
132occured, such as not being able to resolve the hostname, failure to 163occured, such as not being able to resolve the hostname, failure to
133connect or a read error. 164connect, or a read error.
134 165
135Some errors are fatal (which is indicated by C<$fatal> being true). On 166Some errors are fatal (which is indicated by C<$fatal> being true). On
136fatal errors the handle object will be destroyed (by a call to C<< -> 167fatal errors the handle object will be destroyed (by a call to C<< ->
137destroy >>) after invoking the error callback (which means you are free to 168destroy >>) after invoking the error callback (which means you are free to
138examine the handle object). Examples of fatal errors are an EOF condition 169examine the handle object). Examples of fatal errors are an EOF condition
139with active (but unsatisifable) read watchers (C<EPIPE>) or I/O errors. In 170with active (but unsatisfiable) read watchers (C<EPIPE>) or I/O errors. In
140cases where the other side can close the connection at their will it is 171cases where the other side can close the connection at will, it is
141often easiest to not report C<EPIPE> errors in this callback. 172often easiest to not report C<EPIPE> errors in this callback.
142 173
143AnyEvent::Handle tries to find an appropriate error code for you to check 174AnyEvent::Handle tries to find an appropriate error code for you to check
144against, but in some cases (TLS errors), this does not work well. It is 175against, but in some cases (TLS errors), this does not work well. It is
145recommended to always output the C<$message> argument in human-readable 176recommended to always output the C<$message> argument in human-readable
146error messages (it's usually the same as C<"$!">). 177error messages (it's usually the same as C<"$!">).
147 178
148Non-fatal errors can be retried by simply returning, but it is recommended 179Non-fatal errors can be retried by returning, but it is recommended
149to simply ignore this parameter and instead abondon the handle object 180to simply ignore this parameter and instead abondon the handle object
150when this callback is invoked. Examples of non-fatal errors are timeouts 181when this callback is invoked. Examples of non-fatal errors are timeouts
151C<ETIMEDOUT>) or badly-formatted data (C<EBADMSG>). 182C<ETIMEDOUT>) or badly-formatted data (C<EBADMSG>).
152 183
153On callback entrance, the value of C<$!> contains the operating system 184On entry to the callback, the value of C<$!> contains the operating
154error code (or C<ENOSPC>, C<EPIPE>, C<ETIMEDOUT>, C<EBADMSG> or 185system error code (or C<ENOSPC>, C<EPIPE>, C<ETIMEDOUT>, C<EBADMSG> or
155C<EPROTO>). 186C<EPROTO>).
156 187
157While not mandatory, it is I<highly> recommended to set this callback, as 188While not mandatory, it is I<highly> recommended to set this callback, as
158you will not be notified of errors otherwise. The default simply calls 189you will not be notified of errors otherwise. The default just calls
159C<croak>. 190C<croak>.
160 191
161=item on_read => $cb->($handle) 192=item on_read => $cb->($handle)
162 193
163This sets the default read callback, which is called when data arrives 194This sets the default read callback, which is called when data arrives
168To access (and remove data from) the read buffer, use the C<< ->rbuf >> 199To access (and remove data from) the read buffer, use the C<< ->rbuf >>
169method or access the C<< $handle->{rbuf} >> member directly. Note that you 200method or access the C<< $handle->{rbuf} >> member directly. Note that you
170must not enlarge or modify the read buffer, you can only remove data at 201must not enlarge or modify the read buffer, you can only remove data at
171the beginning from it. 202the beginning from it.
172 203
204You can also call C<< ->push_read (...) >> or any other function that
205modifies the read queue. Or do both. Or ...
206
173When an EOF condition is detected then AnyEvent::Handle will first try to 207When an EOF condition is detected, AnyEvent::Handle will first try to
174feed all the remaining data to the queued callbacks and C<on_read> before 208feed all the remaining data to the queued callbacks and C<on_read> before
175calling the C<on_eof> callback. If no progress can be made, then a fatal 209calling the C<on_eof> callback. If no progress can be made, then a fatal
176error will be raised (with C<$!> set to C<EPIPE>). 210error will be raised (with C<$!> set to C<EPIPE>).
177 211
178Note that, unlike requests in the read queue, an C<on_read> callback 212Note that, unlike requests in the read queue, an C<on_read> callback
197set, then a fatal error will be raised with C<$!> set to <0>. 231set, then a fatal error will be raised with C<$!> set to <0>.
198 232
199=item on_drain => $cb->($handle) 233=item on_drain => $cb->($handle)
200 234
201This sets the callback that is called when the write buffer becomes empty 235This sets the callback that is called when the write buffer becomes empty
202(or when the callback is set and the buffer is empty already). 236(or immediately if the buffer is empty already).
203 237
204To append to the write buffer, use the C<< ->push_write >> method. 238To append to the write buffer, use the C<< ->push_write >> method.
205 239
206This callback is useful when you don't want to put all of your write data 240This callback is useful when you don't want to put all of your write data
207into the queue at once, for example, when you want to write the contents 241into the queue at once, for example, when you want to write the contents
209memory and push it into the queue, but instead only read more data from 243memory and push it into the queue, but instead only read more data from
210the file when the write queue becomes empty. 244the file when the write queue becomes empty.
211 245
212=item timeout => $fractional_seconds 246=item timeout => $fractional_seconds
213 247
248=item rtimeout => $fractional_seconds
249
250=item wtimeout => $fractional_seconds
251
214If non-zero, then this enables an "inactivity" timeout: whenever this many 252If non-zero, then these enables an "inactivity" timeout: whenever this
215seconds pass without a successful read or write on the underlying file 253many seconds pass without a successful read or write on the underlying
216handle, the C<on_timeout> callback will be invoked (and if that one is 254file handle (or a call to C<timeout_reset>), the C<on_timeout> callback
217missing, a non-fatal C<ETIMEDOUT> error will be raised). 255will be invoked (and if that one is missing, a non-fatal C<ETIMEDOUT>
256error will be raised).
218 257
258There are three variants of the timeouts that work independently of each
259other, for both read and write (triggered when nothing was read I<OR>
260written), just read (triggered when nothing was read), and just write:
261C<timeout>, C<rtimeout> and C<wtimeout>, with corresponding callbacks
262C<on_timeout>, C<on_rtimeout> and C<on_wtimeout>, and reset functions
263C<timeout_reset>, C<rtimeout_reset>, and C<wtimeout_reset>.
264
219Note that timeout processing is also active when you currently do not have 265Note that timeout processing is active even when you do not have any
220any outstanding read or write requests: If you plan to keep the connection 266outstanding read or write requests: If you plan to keep the connection
221idle then you should disable the timout temporarily or ignore the timeout 267idle then you should disable the timeout temporarily or ignore the
222in the C<on_timeout> callback, in which case AnyEvent::Handle will simply 268timeout in the corresponding C<on_timeout> callback, in which case
223restart the timeout. 269AnyEvent::Handle will simply restart the timeout.
224 270
225Zero (the default) disables this timeout. 271Zero (the default) disables the corresponding timeout.
226 272
227=item on_timeout => $cb->($handle) 273=item on_timeout => $cb->($handle)
274
275=item on_rtimeout => $cb->($handle)
276
277=item on_wtimeout => $cb->($handle)
228 278
229Called whenever the inactivity timeout passes. If you return from this 279Called whenever the inactivity timeout passes. If you return from this
230callback, then the timeout will be reset as if some activity had happened, 280callback, then the timeout will be reset as if some activity had happened,
231so this condition is not fatal in any way. 281so this condition is not fatal in any way.
232 282
240be configured to accept only so-and-so much data that it cannot act on 290be configured to accept only so-and-so much data that it cannot act on
241(for example, when expecting a line, an attacker could send an unlimited 291(for example, when expecting a line, an attacker could send an unlimited
242amount of data without a callback ever being called as long as the line 292amount of data without a callback ever being called as long as the line
243isn't finished). 293isn't finished).
244 294
295=item wbuf_max => <bytes>
296
297If defined, then a fatal error will be raised (with C<$!> set to C<ENOSPC>)
298when the write buffer ever (strictly) exceeds this size. This is useful to
299avoid some forms of denial-of-service attacks.
300
301Although the units of this parameter is bytes, this is the I<raw> number
302of bytes not yet accepted by the kernel. This can make a difference when
303you e.g. use TLS, as TLS typically makes your write data larger (but it
304can also make it smaller due to compression).
305
306As an example of when this limit is useful, take a chat server that sends
307chat messages to a client. If the client does not read those in a timely
308manner then the send buffer in the server would grow unbounded.
309
245=item autocork => <boolean> 310=item autocork => <boolean>
246 311
247When disabled (the default), then C<push_write> will try to immediately 312When disabled (the default), C<push_write> will try to immediately
248write the data to the handle, if possible. This avoids having to register 313write the data to the handle if possible. This avoids having to register
249a write watcher and wait for the next event loop iteration, but can 314a write watcher and wait for the next event loop iteration, but can
250be inefficient if you write multiple small chunks (on the wire, this 315be inefficient if you write multiple small chunks (on the wire, this
251disadvantage is usually avoided by your kernel's nagle algorithm, see 316disadvantage is usually avoided by your kernel's nagle algorithm, see
252C<no_delay>, but this option can save costly syscalls). 317C<no_delay>, but this option can save costly syscalls).
253 318
254When enabled, then writes will always be queued till the next event loop 319When enabled, writes will always be queued till the next event loop
255iteration. This is efficient when you do many small writes per iteration, 320iteration. This is efficient when you do many small writes per iteration,
256but less efficient when you do a single write only per iteration (or when 321but less efficient when you do a single write only per iteration (or when
257the write buffer often is full). It also increases write latency. 322the write buffer often is full). It also increases write latency.
258 323
259=item no_delay => <boolean> 324=item no_delay => <boolean>
263the Nagle algorithm, and usually it is beneficial. 328the Nagle algorithm, and usually it is beneficial.
264 329
265In some situations you want as low a delay as possible, which can be 330In some situations you want as low a delay as possible, which can be
266accomplishd by setting this option to a true value. 331accomplishd by setting this option to a true value.
267 332
268The default is your opertaing system's default behaviour (most likely 333The default is your operating system's default behaviour (most likely
269enabled), this option explicitly enables or disables it, if possible. 334enabled). This option explicitly enables or disables it, if possible.
335
336=item keepalive => <boolean>
337
338Enables (default disable) the SO_KEEPALIVE option on the stream socket:
339normally, TCP connections have no time-out once established, so TCP
340connections, once established, can stay alive forever even when the other
341side has long gone. TCP keepalives are a cheap way to take down long-lived
342TCP connections when the other side becomes unreachable. While the default
343is OS-dependent, TCP keepalives usually kick in after around two hours,
344and, if the other side doesn't reply, take down the TCP connection some 10
345to 15 minutes later.
346
347It is harmless to specify this option for file handles that do not support
348keepalives, and enabling it on connections that are potentially long-lived
349is usually a good idea.
350
351=item oobinline => <boolean>
352
353BSD majorly fucked up the implementation of TCP urgent data. The result
354is that almost no OS implements TCP according to the specs, and every OS
355implements it slightly differently.
356
357If you want to handle TCP urgent data, then setting this flag (the default
358is enabled) gives you the most portable way of getting urgent data, by
359putting it into the stream.
360
361Since BSD emulation of OOB data on top of TCP's urgent data can have
362security implications, AnyEvent::Handle sets this flag automatically
363unless explicitly specified. Note that setting this flag after
364establishing a connection I<may> be a bit too late (data loss could
365already have occured on BSD systems), but at least it will protect you
366from most attacks.
270 367
271=item read_size => <bytes> 368=item read_size => <bytes>
272 369
273The default read block size (the amount of bytes this module will 370The initial read block size, the number of bytes this module will try
274try to read during each loop iteration, which affects memory 371to read during each loop iteration. Each handle object will consume
275requirements). Default: C<8192>. 372at least this amount of memory for the read buffer as well, so when
373handling many connections watch out for memory requirements). See also
374C<max_read_size>. Default: C<2048>.
375
376=item max_read_size => <bytes>
377
378The maximum read buffer size used by the dynamic adjustment
379algorithm: Each time AnyEvent::Handle can read C<read_size> bytes in
380one go it will double C<read_size> up to the maximum given by this
381option. Default: C<131072> or C<read_size>, whichever is higher.
276 382
277=item low_water_mark => <bytes> 383=item low_water_mark => <bytes>
278 384
279Sets the amount of bytes (default: C<0>) that make up an "empty" write 385Sets the number of bytes (default: C<0>) that make up an "empty" write
280buffer: If the write reaches this size or gets even samller it is 386buffer: If the buffer reaches this size or gets even samller it is
281considered empty. 387considered empty.
282 388
283Sometimes it can be beneficial (for performance reasons) to add data to 389Sometimes it can be beneficial (for performance reasons) to add data to
284the write buffer before it is fully drained, but this is a rare case, as 390the write buffer before it is fully drained, but this is a rare case, as
285the operating system kernel usually buffers data as well, so the default 391the operating system kernel usually buffers data as well, so the default
286is good in almost all cases. 392is good in almost all cases.
287 393
288=item linger => <seconds> 394=item linger => <seconds>
289 395
290If non-zero (default: C<3600>), then the destructor of the 396If this is non-zero (default: C<3600>), the destructor of the
291AnyEvent::Handle object will check whether there is still outstanding 397AnyEvent::Handle object will check whether there is still outstanding
292write data and will install a watcher that will write this data to the 398write data and will install a watcher that will write this data to the
293socket. No errors will be reported (this mostly matches how the operating 399socket. No errors will be reported (this mostly matches how the operating
294system treats outstanding data at socket close time). 400system treats outstanding data at socket close time).
295 401
302A string used to identify the remote site - usually the DNS hostname 408A string used to identify the remote site - usually the DNS hostname
303(I<not> IDN!) used to create the connection, rarely the IP address. 409(I<not> IDN!) used to create the connection, rarely the IP address.
304 410
305Apart from being useful in error messages, this string is also used in TLS 411Apart from being useful in error messages, this string is also used in TLS
306peername verification (see C<verify_peername> in L<AnyEvent::TLS>). This 412peername verification (see C<verify_peername> in L<AnyEvent::TLS>). This
307verification will be skipped when C<peername> is not specified or 413verification will be skipped when C<peername> is not specified or is
308C<undef>. 414C<undef>.
309 415
310=item tls => "accept" | "connect" | Net::SSLeay::SSL object 416=item tls => "accept" | "connect" | Net::SSLeay::SSL object
311 417
312When this parameter is given, it enables TLS (SSL) mode, that means 418When this parameter is given, it enables TLS (SSL) mode, that means
313AnyEvent will start a TLS handshake as soon as the conenction has been 419AnyEvent will start a TLS handshake as soon as the connection has been
314established and will transparently encrypt/decrypt data afterwards. 420established and will transparently encrypt/decrypt data afterwards.
315 421
316All TLS protocol errors will be signalled as C<EPROTO>, with an 422All TLS protocol errors will be signalled as C<EPROTO>, with an
317appropriate error message. 423appropriate error message.
318 424
338B<IMPORTANT:> since Net::SSLeay "objects" are really only integers, 444B<IMPORTANT:> since Net::SSLeay "objects" are really only integers,
339passing in the wrong integer will lead to certain crash. This most often 445passing in the wrong integer will lead to certain crash. This most often
340happens when one uses a stylish C<< tls => 1 >> and is surprised about the 446happens when one uses a stylish C<< tls => 1 >> and is surprised about the
341segmentation fault. 447segmentation fault.
342 448
343See the C<< ->starttls >> method for when need to start TLS negotiation later. 449Use the C<< ->starttls >> method if you need to start TLS negotiation later.
344 450
345=item tls_ctx => $anyevent_tls 451=item tls_ctx => $anyevent_tls
346 452
347Use the given C<AnyEvent::TLS> object to create the new TLS connection 453Use the given C<AnyEvent::TLS> object to create the new TLS connection
348(unless a connection object was specified directly). If this parameter is 454(unless a connection object was specified directly). If this
349missing, then AnyEvent::Handle will use C<AnyEvent::Handle::TLS_CTX>. 455parameter is missing (or C<undef>), then AnyEvent::Handle will use
456C<AnyEvent::Handle::TLS_CTX>.
350 457
351Instead of an object, you can also specify a hash reference with C<< key 458Instead of an object, you can also specify a hash reference with C<< key
352=> value >> pairs. Those will be passed to L<AnyEvent::TLS> to create a 459=> value >> pairs. Those will be passed to L<AnyEvent::TLS> to create a
353new TLS context object. 460new TLS context object.
354 461
363 470
364TLS handshake failures will not cause C<on_error> to be invoked when this 471TLS handshake failures will not cause C<on_error> to be invoked when this
365callback is in effect, instead, the error message will be passed to C<on_starttls>. 472callback is in effect, instead, the error message will be passed to C<on_starttls>.
366 473
367Without this callback, handshake failures lead to C<on_error> being 474Without this callback, handshake failures lead to C<on_error> being
368called, as normal. 475called as usual.
369 476
370Note that you cannot call C<starttls> right again in this callback. If you 477Note that you cannot just call C<starttls> again in this callback. If you
371need to do that, start an zero-second timer instead whose callback can 478need to do that, start an zero-second timer instead whose callback can
372then call C<< ->starttls >> again. 479then call C<< ->starttls >> again.
373 480
374=item on_stoptls => $cb->($handle) 481=item on_stoptls => $cb->($handle)
375 482
423 $self->{connect}[0], 530 $self->{connect}[0],
424 $self->{connect}[1], 531 $self->{connect}[1],
425 sub { 532 sub {
426 my ($fh, $host, $port, $retry) = @_; 533 my ($fh, $host, $port, $retry) = @_;
427 534
535 delete $self->{_connect}; # no longer needed
536
428 if ($fh) { 537 if ($fh) {
429 $self->{fh} = $fh; 538 $self->{fh} = $fh;
430 539
431 delete $self->{_skip_drain_rbuf}; 540 delete $self->{_skip_drain_rbuf};
432 $self->_start; 541 $self->_start;
433 542
434 $self->{on_connect} 543 $self->{on_connect}
435 and $self->{on_connect}($self, $host, $port, sub { 544 and $self->{on_connect}($self, $host, $port, sub {
436 delete @$self{qw(fh _tw _ww _rw _eof _queue rbuf _wbuf tls _tls_rbuf _tls_wbuf)}; 545 delete @$self{qw(fh _tw _rtw _wtw _ww _rw _eof _queue rbuf _wbuf tls _tls_rbuf _tls_wbuf)};
437 $self->{_skip_drain_rbuf} = 1; 546 $self->{_skip_drain_rbuf} = 1;
438 &$retry; 547 &$retry;
439 }); 548 });
440 549
441 } else { 550 } else {
442 if ($self->{on_connect_error}) { 551 if ($self->{on_connect_error}) {
443 $self->{on_connect_error}($self, "$!"); 552 $self->{on_connect_error}($self, "$!");
444 $self->destroy; 553 $self->destroy if $self;
445 } else { 554 } else {
446 $self->fatal ($!, 1); 555 $self->_error ($!, 1);
447 } 556 }
448 } 557 }
449 }, 558 },
450 sub { 559 sub {
451 local $self->{fh} = $_[0]; 560 local $self->{fh} = $_[0];
452 561
562 $self->{on_prepare}
453 $self->{on_prepare}->($self) 563 ? $self->{on_prepare}->($self)
454 if $self->{on_prepare}; 564 : ()
455 } 565 }
456 ); 566 );
457 } 567 }
458 568
459 } else { 569 } else {
464} 574}
465 575
466sub _start { 576sub _start {
467 my ($self) = @_; 577 my ($self) = @_;
468 578
579 # too many clueless people try to use udp and similar sockets
580 # with AnyEvent::Handle, do them a favour.
581 my $type = getsockopt $self->{fh}, Socket::SOL_SOCKET (), Socket::SO_TYPE ();
582 Carp::croak "AnyEvent::Handle: only stream sockets supported, anything else will NOT work!"
583 if Socket::SOCK_STREAM () != (unpack "I", $type) && defined $type;
584
469 AnyEvent::Util::fh_nonblocking $self->{fh}, 1; 585 AnyEvent::Util::fh_nonblocking $self->{fh}, 1;
470 586
587 $self->{_activity} =
588 $self->{_ractivity} =
471 $self->{_activity} = AnyEvent->now; 589 $self->{_wactivity} = AE::now;
472 $self->_timeout;
473 590
591 $self->{read_size} ||= 2048;
592 $self->{max_read_size} = $self->{read_size}
593 if $self->{read_size} > ($self->{max_read_size} || MAX_READ_SIZE);
594
595 $self->timeout (delete $self->{timeout} ) if $self->{timeout};
596 $self->rtimeout (delete $self->{rtimeout} ) if $self->{rtimeout};
597 $self->wtimeout (delete $self->{wtimeout} ) if $self->{wtimeout};
598
474 $self->no_delay (delete $self->{no_delay}) if exists $self->{no_delay}; 599 $self->no_delay (delete $self->{no_delay} ) if exists $self->{no_delay} && $self->{no_delay};
600 $self->keepalive (delete $self->{keepalive}) if exists $self->{keepalive} && $self->{keepalive};
475 601
602 $self->oobinline (exists $self->{oobinline} ? delete $self->{oobinline} : 1);
603
476 $self->starttls (delete $self->{tls}, delete $self->{tls_ctx}) 604 $self->starttls (delete $self->{tls}, delete $self->{tls_ctx})
477 if $self->{tls}; 605 if $self->{tls};
478 606
479 $self->on_drain (delete $self->{on_drain}) if $self->{on_drain}; 607 $self->on_drain (delete $self->{on_drain} ) if $self->{on_drain};
480 608
481 $self->start_read 609 $self->start_read
482 if $self->{on_read} || @{ $self->{_queue} }; 610 if $self->{on_read} || @{ $self->{_queue} };
483}
484 611
485#sub _shutdown { 612 $self->_drain_wbuf;
486# my ($self) = @_; 613}
487#
488# delete @$self{qw(_tw _rw _ww fh wbuf on_read _queue)};
489# $self->{_eof} = 1; # tell starttls et. al to stop trying
490#
491# &_freetls;
492#}
493 614
494sub _error { 615sub _error {
495 my ($self, $errno, $fatal, $message) = @_; 616 my ($self, $errno, $fatal, $message) = @_;
496 617
497 $! = $errno; 618 $! = $errno;
498 $message ||= "$!"; 619 $message ||= "$!";
499 620
500 if ($self->{on_error}) { 621 if ($self->{on_error}) {
501 $self->{on_error}($self, $fatal, $message); 622 $self->{on_error}($self, $fatal, $message);
502 $self->destroy if $fatal; 623 $self->destroy if $fatal;
503 } elsif ($self->{fh}) { 624 } elsif ($self->{fh} || $self->{connect}) {
504 $self->destroy; 625 $self->destroy;
505 Carp::croak "AnyEvent::Handle uncaught error: $message"; 626 Carp::croak "AnyEvent::Handle uncaught error: $message";
506 } 627 }
507} 628}
508 629
534 $_[0]{on_eof} = $_[1]; 655 $_[0]{on_eof} = $_[1];
535} 656}
536 657
537=item $handle->on_timeout ($cb) 658=item $handle->on_timeout ($cb)
538 659
539Replace the current C<on_timeout> callback, or disables the callback (but 660=item $handle->on_rtimeout ($cb)
540not the timeout) if C<$cb> = C<undef>. See the C<timeout> constructor
541argument and method.
542 661
543=cut 662=item $handle->on_wtimeout ($cb)
544 663
545sub on_timeout { 664Replace the current C<on_timeout>, C<on_rtimeout> or C<on_wtimeout>
546 $_[0]{on_timeout} = $_[1]; 665callback, or disables the callback (but not the timeout) if C<$cb> =
547} 666C<undef>. See the C<timeout> constructor argument and method.
667
668=cut
669
670# see below
548 671
549=item $handle->autocork ($boolean) 672=item $handle->autocork ($boolean)
550 673
551Enables or disables the current autocork behaviour (see C<autocork> 674Enables or disables the current autocork behaviour (see C<autocork>
552constructor argument). Changes will only take effect on the next write. 675constructor argument). Changes will only take effect on the next write.
565=cut 688=cut
566 689
567sub no_delay { 690sub no_delay {
568 $_[0]{no_delay} = $_[1]; 691 $_[0]{no_delay} = $_[1];
569 692
693 setsockopt $_[0]{fh}, Socket::IPPROTO_TCP (), Socket::TCP_NODELAY (), int $_[1]
694 if $_[0]{fh};
695}
696
697=item $handle->keepalive ($boolean)
698
699Enables or disables the C<keepalive> setting (see constructor argument of
700the same name for details).
701
702=cut
703
704sub keepalive {
705 $_[0]{keepalive} = $_[1];
706
570 eval { 707 eval {
571 local $SIG{__DIE__}; 708 local $SIG{__DIE__};
572 setsockopt $_[0]{fh}, &Socket::IPPROTO_TCP, &Socket::TCP_NODELAY, int $_[1] 709 setsockopt $_[0]{fh}, Socket::SOL_SOCKET (), Socket::SO_KEEPALIVE (), int $_[1]
573 if $_[0]{fh}; 710 if $_[0]{fh};
574 }; 711 };
575} 712}
576 713
714=item $handle->oobinline ($boolean)
715
716Enables or disables the C<oobinline> setting (see constructor argument of
717the same name for details).
718
719=cut
720
721sub oobinline {
722 $_[0]{oobinline} = $_[1];
723
724 eval {
725 local $SIG{__DIE__};
726 setsockopt $_[0]{fh}, Socket::SOL_SOCKET (), Socket::SO_OOBINLINE (), int $_[1]
727 if $_[0]{fh};
728 };
729}
730
731=item $handle->keepalive ($boolean)
732
733Enables or disables the C<keepalive> setting (see constructor argument of
734the same name for details).
735
736=cut
737
738sub keepalive {
739 $_[0]{keepalive} = $_[1];
740
741 eval {
742 local $SIG{__DIE__};
743 setsockopt $_[0]{fh}, Socket::SOL_SOCKET (), Socket::SO_KEEPALIVE (), int $_[1]
744 if $_[0]{fh};
745 };
746}
747
577=item $handle->on_starttls ($cb) 748=item $handle->on_starttls ($cb)
578 749
579Replace the current C<on_starttls> callback (see the C<on_starttls> constructor argument). 750Replace the current C<on_starttls> callback (see the C<on_starttls> constructor argument).
580 751
581=cut 752=cut
588 759
589Replace the current C<on_stoptls> callback (see the C<on_stoptls> constructor argument). 760Replace the current C<on_stoptls> callback (see the C<on_stoptls> constructor argument).
590 761
591=cut 762=cut
592 763
593sub on_starttls { 764sub on_stoptls {
594 $_[0]{on_stoptls} = $_[1]; 765 $_[0]{on_stoptls} = $_[1];
595} 766}
596 767
768=item $handle->rbuf_max ($max_octets)
769
770Configures the C<rbuf_max> setting (C<undef> disables it).
771
772=item $handle->wbuf_max ($max_octets)
773
774Configures the C<wbuf_max> setting (C<undef> disables it).
775
776=cut
777
778sub rbuf_max {
779 $_[0]{rbuf_max} = $_[1];
780}
781
782sub wbuf_max {
783 $_[0]{wbuf_max} = $_[1];
784}
785
597############################################################################# 786#############################################################################
598 787
599=item $handle->timeout ($seconds) 788=item $handle->timeout ($seconds)
600 789
790=item $handle->rtimeout ($seconds)
791
792=item $handle->wtimeout ($seconds)
793
601Configures (or disables) the inactivity timeout. 794Configures (or disables) the inactivity timeout.
602 795
603=cut 796The timeout will be checked instantly, so this method might destroy the
797handle before it returns.
604 798
605sub timeout { 799=item $handle->timeout_reset
800
801=item $handle->rtimeout_reset
802
803=item $handle->wtimeout_reset
804
805Reset the activity timeout, as if data was received or sent.
806
807These methods are cheap to call.
808
809=cut
810
811for my $dir ("", "r", "w") {
812 my $timeout = "${dir}timeout";
813 my $tw = "_${dir}tw";
814 my $on_timeout = "on_${dir}timeout";
815 my $activity = "_${dir}activity";
816 my $cb;
817
818 *$on_timeout = sub {
819 $_[0]{$on_timeout} = $_[1];
820 };
821
822 *$timeout = sub {
606 my ($self, $timeout) = @_; 823 my ($self, $new_value) = @_;
607 824
825 $new_value >= 0
826 or Carp::croak "AnyEvent::Handle->$timeout called with negative timeout ($new_value), caught";
827
608 $self->{timeout} = $timeout; 828 $self->{$timeout} = $new_value;
609 $self->_timeout; 829 delete $self->{$tw}; &$cb;
610} 830 };
611 831
832 *{"${dir}timeout_reset"} = sub {
833 $_[0]{$activity} = AE::now;
834 };
835
836 # main workhorse:
612# reset the timeout watcher, as neccessary 837 # reset the timeout watcher, as neccessary
613# also check for time-outs 838 # also check for time-outs
614sub _timeout { 839 $cb = sub {
615 my ($self) = @_; 840 my ($self) = @_;
616 841
617 if ($self->{timeout} && $self->{fh}) { 842 if ($self->{$timeout} && $self->{fh}) {
618 my $NOW = AnyEvent->now; 843 my $NOW = AE::now;
619 844
620 # when would the timeout trigger? 845 # when would the timeout trigger?
621 my $after = $self->{_activity} + $self->{timeout} - $NOW; 846 my $after = $self->{$activity} + $self->{$timeout} - $NOW;
622 847
623 # now or in the past already? 848 # now or in the past already?
624 if ($after <= 0) { 849 if ($after <= 0) {
625 $self->{_activity} = $NOW; 850 $self->{$activity} = $NOW;
626 851
627 if ($self->{on_timeout}) { 852 if ($self->{$on_timeout}) {
628 $self->{on_timeout}($self); 853 $self->{$on_timeout}($self);
629 } else { 854 } else {
630 $self->_error (Errno::ETIMEDOUT); 855 $self->_error (Errno::ETIMEDOUT);
856 }
857
858 # callback could have changed timeout value, optimise
859 return unless $self->{$timeout};
860
861 # calculate new after
862 $after = $self->{$timeout};
631 } 863 }
632 864
633 # callback could have changed timeout value, optimise 865 Scalar::Util::weaken $self;
634 return unless $self->{timeout}; 866 return unless $self; # ->error could have destroyed $self
635 867
636 # calculate new after 868 $self->{$tw} ||= AE::timer $after, 0, sub {
637 $after = $self->{timeout}; 869 delete $self->{$tw};
870 $cb->($self);
871 };
872 } else {
873 delete $self->{$tw};
638 } 874 }
639
640 Scalar::Util::weaken $self;
641 return unless $self; # ->error could have destroyed $self
642
643 $self->{_tw} ||= AnyEvent->timer (after => $after, cb => sub {
644 delete $self->{_tw};
645 $self->_timeout;
646 });
647 } else {
648 delete $self->{_tw};
649 } 875 }
650} 876}
651 877
652############################################################################# 878#############################################################################
653 879
669=item $handle->on_drain ($cb) 895=item $handle->on_drain ($cb)
670 896
671Sets the C<on_drain> callback or clears it (see the description of 897Sets the C<on_drain> callback or clears it (see the description of
672C<on_drain> in the constructor). 898C<on_drain> in the constructor).
673 899
900This method may invoke callbacks (and therefore the handle might be
901destroyed after it returns).
902
674=cut 903=cut
675 904
676sub on_drain { 905sub on_drain {
677 my ($self, $cb) = @_; 906 my ($self, $cb) = @_;
678 907
682 if $cb && $self->{low_water_mark} >= (length $self->{wbuf}) + (length $self->{_tls_wbuf}); 911 if $cb && $self->{low_water_mark} >= (length $self->{wbuf}) + (length $self->{_tls_wbuf});
683} 912}
684 913
685=item $handle->push_write ($data) 914=item $handle->push_write ($data)
686 915
687Queues the given scalar to be written. You can push as much data as you 916Queues the given scalar to be written. You can push as much data as
688want (only limited by the available memory), as C<AnyEvent::Handle> 917you want (only limited by the available memory and C<wbuf_max>), as
689buffers it independently of the kernel. 918C<AnyEvent::Handle> buffers it independently of the kernel.
919
920This method may invoke callbacks (and therefore the handle might be
921destroyed after it returns).
690 922
691=cut 923=cut
692 924
693sub _drain_wbuf { 925sub _drain_wbuf {
694 my ($self) = @_; 926 my ($self) = @_;
701 my $len = syswrite $self->{fh}, $self->{wbuf}; 933 my $len = syswrite $self->{fh}, $self->{wbuf};
702 934
703 if (defined $len) { 935 if (defined $len) {
704 substr $self->{wbuf}, 0, $len, ""; 936 substr $self->{wbuf}, 0, $len, "";
705 937
706 $self->{_activity} = AnyEvent->now; 938 $self->{_activity} = $self->{_wactivity} = AE::now;
707 939
708 $self->{on_drain}($self) 940 $self->{on_drain}($self)
709 if $self->{low_water_mark} >= (length $self->{wbuf}) + (length $self->{_tls_wbuf}) 941 if $self->{low_water_mark} >= (length $self->{wbuf}) + (length $self->{_tls_wbuf})
710 && $self->{on_drain}; 942 && $self->{on_drain};
711 943
717 949
718 # try to write data immediately 950 # try to write data immediately
719 $cb->() unless $self->{autocork}; 951 $cb->() unless $self->{autocork};
720 952
721 # if still data left in wbuf, we need to poll 953 # if still data left in wbuf, we need to poll
722 $self->{_ww} = AnyEvent->io (fh => $self->{fh}, poll => "w", cb => $cb) 954 $self->{_ww} = AE::io $self->{fh}, 1, $cb
723 if length $self->{wbuf}; 955 if length $self->{wbuf};
956
957 if (
958 defined $self->{wbuf_max}
959 && $self->{wbuf_max} < length $self->{wbuf}
960 ) {
961 $self->_error (Errno::ENOSPC, 1), return;
962 }
724 }; 963 };
725} 964}
726 965
727our %WH; 966our %WH;
728 967
968# deprecated
729sub register_write_type($$) { 969sub register_write_type($$) {
730 $WH{$_[0]} = $_[1]; 970 $WH{$_[0]} = $_[1];
731} 971}
732 972
733sub push_write { 973sub push_write {
734 my $self = shift; 974 my $self = shift;
735 975
736 if (@_ > 1) { 976 if (@_ > 1) {
737 my $type = shift; 977 my $type = shift;
738 978
979 @_ = ($WH{$type} ||= _load_func "$type\::anyevent_write_type"
739 @_ = ($WH{$type} or Carp::croak "unsupported type passed to AnyEvent::Handle::push_write") 980 or Carp::croak "unsupported/unloadable type '$type' passed to AnyEvent::Handle::push_write")
740 ->($self, @_); 981 ->($self, @_);
741 } 982 }
742 983
984 # we downgrade here to avoid hard-to-track-down bugs,
985 # and diagnose the problem earlier and better.
986
743 if ($self->{tls}) { 987 if ($self->{tls}) {
744 $self->{_tls_wbuf} .= $_[0]; 988 utf8::downgrade $self->{_tls_wbuf} .= $_[0];
745 989 &_dotls ($self) if $self->{fh};
746 &_dotls ($self);
747 } else { 990 } else {
748 $self->{wbuf} .= $_[0]; 991 utf8::downgrade $self->{wbuf} .= $_[0];
749 $self->_drain_wbuf if $self->{fh}; 992 $self->_drain_wbuf if $self->{fh};
750 } 993 }
751} 994}
752 995
753=item $handle->push_write (type => @args) 996=item $handle->push_write (type => @args)
754 997
755Instead of formatting your data yourself, you can also let this module do 998Instead of formatting your data yourself, you can also let this module
756the job by specifying a type and type-specific arguments. 999do the job by specifying a type and type-specific arguments. You
1000can also specify the (fully qualified) name of a package, in which
1001case AnyEvent tries to load the package and then expects to find the
1002C<anyevent_write_type> function inside (see "custom write types", below).
757 1003
758Predefined types are (if you have ideas for additional types, feel free to 1004Predefined types are (if you have ideas for additional types, feel free to
759drop by and tell us): 1005drop by and tell us):
760 1006
761=over 4 1007=over 4
818Other languages could read single lines terminated by a newline and pass 1064Other languages could read single lines terminated by a newline and pass
819this line into their JSON decoder of choice. 1065this line into their JSON decoder of choice.
820 1066
821=cut 1067=cut
822 1068
1069sub json_coder() {
1070 eval { require JSON::XS; JSON::XS->new->utf8 }
1071 || do { require JSON; JSON->new->utf8 }
1072}
1073
823register_write_type json => sub { 1074register_write_type json => sub {
824 my ($self, $ref) = @_; 1075 my ($self, $ref) = @_;
825 1076
826 require JSON; 1077 my $json = $self->{json} ||= json_coder;
827 1078
828 $self->{json} ? $self->{json}->encode ($ref) 1079 $json->encode ($ref)
829 : JSON::encode_json ($ref)
830}; 1080};
831 1081
832=item storable => $reference 1082=item storable => $reference
833 1083
834Freezes the given reference using L<Storable> and writes it to the 1084Freezes the given reference using L<Storable> and writes it to the
837=cut 1087=cut
838 1088
839register_write_type storable => sub { 1089register_write_type storable => sub {
840 my ($self, $ref) = @_; 1090 my ($self, $ref) = @_;
841 1091
842 require Storable; 1092 require Storable unless $Storable::VERSION;
843 1093
844 pack "w/a*", Storable::nfreeze ($ref) 1094 pack "w/a*", Storable::nfreeze ($ref)
845}; 1095};
846 1096
847=back 1097=back
852before it was actually written. One way to do that is to replace your 1102before it was actually written. One way to do that is to replace your
853C<on_drain> handler by a callback that shuts down the socket (and set 1103C<on_drain> handler by a callback that shuts down the socket (and set
854C<low_water_mark> to C<0>). This method is a shorthand for just that, and 1104C<low_water_mark> to C<0>). This method is a shorthand for just that, and
855replaces the C<on_drain> callback with: 1105replaces the C<on_drain> callback with:
856 1106
857 sub { shutdown $_[0]{fh}, 1 } # for push_shutdown 1107 sub { shutdown $_[0]{fh}, 1 }
858 1108
859This simply shuts down the write side and signals an EOF condition to the 1109This simply shuts down the write side and signals an EOF condition to the
860the peer. 1110the peer.
861 1111
862You can rely on the normal read queue and C<on_eof> handling 1112You can rely on the normal read queue and C<on_eof> handling
863afterwards. This is the cleanest way to close a connection. 1113afterwards. This is the cleanest way to close a connection.
864 1114
1115This method may invoke callbacks (and therefore the handle might be
1116destroyed after it returns).
1117
865=cut 1118=cut
866 1119
867sub push_shutdown { 1120sub push_shutdown {
868 my ($self) = @_; 1121 my ($self) = @_;
869 1122
870 delete $self->{low_water_mark}; 1123 delete $self->{low_water_mark};
871 $self->on_drain (sub { shutdown $_[0]{fh}, 1 }); 1124 $self->on_drain (sub { shutdown $_[0]{fh}, 1 });
872} 1125}
873 1126
874=item AnyEvent::Handle::register_write_type type => $coderef->($handle, @args) 1127=item custom write types - Package::anyevent_write_type $handle, @args
875 1128
876This function (not method) lets you add your own types to C<push_write>. 1129Instead of one of the predefined types, you can also specify the name of
1130a package. AnyEvent will try to load the package and then expects to find
1131a function named C<anyevent_write_type> inside. If it isn't found, it
1132progressively tries to load the parent package until it either finds the
1133function (good) or runs out of packages (bad).
1134
877Whenever the given C<type> is used, C<push_write> will invoke the code 1135Whenever the given C<type> is used, C<push_write> will the function with
878reference with the handle object and the remaining arguments. 1136the handle object and the remaining arguments.
879 1137
880The code reference is supposed to return a single octet string that will 1138The function is supposed to return a single octet string that will be
881be appended to the write buffer. 1139appended to the write buffer, so you can mentally treat this function as a
1140"arguments to on-the-wire-format" converter.
882 1141
883Note that this is a function, and all types registered this way will be 1142Example: implement a custom write type C<join> that joins the remaining
884global, so try to use unique names. 1143arguments using the first one.
1144
1145 $handle->push_write (My::Type => " ", 1,2,3);
1146
1147 # uses the following package, which can be defined in the "My::Type" or in
1148 # the "My" modules to be auto-loaded, or just about anywhere when the
1149 # My::Type::anyevent_write_type is defined before invoking it.
1150
1151 package My::Type;
1152
1153 sub anyevent_write_type {
1154 my ($handle, $delim, @args) = @_;
1155
1156 join $delim, @args
1157 }
885 1158
886=cut 1159=cut
887 1160
888############################################################################# 1161#############################################################################
889 1162
898ways, the "simple" way, using only C<on_read> and the "complex" way, using 1171ways, the "simple" way, using only C<on_read> and the "complex" way, using
899a queue. 1172a queue.
900 1173
901In the simple case, you just install an C<on_read> callback and whenever 1174In the simple case, you just install an C<on_read> callback and whenever
902new data arrives, it will be called. You can then remove some data (if 1175new data arrives, it will be called. You can then remove some data (if
903enough is there) from the read buffer (C<< $handle->rbuf >>). Or you cna 1176enough is there) from the read buffer (C<< $handle->rbuf >>). Or you can
904leave the data there if you want to accumulate more (e.g. when only a 1177leave the data there if you want to accumulate more (e.g. when only a
905partial message has been received so far). 1178partial message has been received so far), or change the read queue with
1179e.g. C<push_read>.
906 1180
907In the more complex case, you want to queue multiple callbacks. In this 1181In the more complex case, you want to queue multiple callbacks. In this
908case, AnyEvent::Handle will call the first queued callback each time new 1182case, AnyEvent::Handle will call the first queued callback each time new
909data arrives (also the first time it is queued) and removes it when it has 1183data arrives (also the first time it is queued) and remove it when it has
910done its job (see C<push_read>, below). 1184done its job (see C<push_read>, below).
911 1185
912This way you can, for example, push three line-reads, followed by reading 1186This way you can, for example, push three line-reads, followed by reading
913a chunk of data, and AnyEvent::Handle will execute them in order. 1187a chunk of data, and AnyEvent::Handle will execute them in order.
914 1188
972 1246
973sub _drain_rbuf { 1247sub _drain_rbuf {
974 my ($self) = @_; 1248 my ($self) = @_;
975 1249
976 # avoid recursion 1250 # avoid recursion
977 return if exists $self->{_skip_drain_rbuf}; 1251 return if $self->{_skip_drain_rbuf};
978 local $self->{_skip_drain_rbuf} = 1; 1252 local $self->{_skip_drain_rbuf} = 1;
979
980 if (
981 defined $self->{rbuf_max}
982 && $self->{rbuf_max} < length $self->{rbuf}
983 ) {
984 $self->_error (Errno::ENOSPC, 1), return;
985 }
986 1253
987 while () { 1254 while () {
988 # we need to use a separate tls read buffer, as we must not receive data while 1255 # we need to use a separate tls read buffer, as we must not receive data while
989 # we are draining the buffer, and this can only happen with TLS. 1256 # we are draining the buffer, and this can only happen with TLS.
990 $self->{rbuf} .= delete $self->{_tls_rbuf} if exists $self->{_tls_rbuf}; 1257 $self->{rbuf} .= delete $self->{_tls_rbuf}
1258 if exists $self->{_tls_rbuf};
991 1259
992 my $len = length $self->{rbuf}; 1260 my $len = length $self->{rbuf};
993 1261
994 if (my $cb = shift @{ $self->{_queue} }) { 1262 if (my $cb = shift @{ $self->{_queue} }) {
995 unless ($cb->($self)) { 1263 unless ($cb->($self)) {
996 if ($self->{_eof}) { 1264 # no progress can be made
997 # no progress can be made (not enough data and no data forthcoming) 1265 # (not enough data and no data forthcoming)
998 $self->_error (Errno::EPIPE, 1), return; 1266 $self->_error (Errno::EPIPE, 1), return
999 } 1267 if $self->{_eof};
1000 1268
1001 unshift @{ $self->{_queue} }, $cb; 1269 unshift @{ $self->{_queue} }, $cb;
1002 last; 1270 last;
1003 } 1271 }
1004 } elsif ($self->{on_read}) { 1272 } elsif ($self->{on_read}) {
1024 last; 1292 last;
1025 } 1293 }
1026 } 1294 }
1027 1295
1028 if ($self->{_eof}) { 1296 if ($self->{_eof}) {
1029 if ($self->{on_eof}) { 1297 $self->{on_eof}
1030 $self->{on_eof}($self) 1298 ? $self->{on_eof}($self)
1031 } else {
1032 $self->_error (0, 1, "Unexpected end-of-file"); 1299 : $self->_error (0, 1, "Unexpected end-of-file");
1033 } 1300
1301 return;
1302 }
1303
1304 if (
1305 defined $self->{rbuf_max}
1306 && $self->{rbuf_max} < length $self->{rbuf}
1307 ) {
1308 $self->_error (Errno::ENOSPC, 1), return;
1034 } 1309 }
1035 1310
1036 # may need to restart read watcher 1311 # may need to restart read watcher
1037 unless ($self->{_rw}) { 1312 unless ($self->{_rw}) {
1038 $self->start_read 1313 $self->start_read
1044 1319
1045This replaces the currently set C<on_read> callback, or clears it (when 1320This replaces the currently set C<on_read> callback, or clears it (when
1046the new callback is C<undef>). See the description of C<on_read> in the 1321the new callback is C<undef>). See the description of C<on_read> in the
1047constructor. 1322constructor.
1048 1323
1324This method may invoke callbacks (and therefore the handle might be
1325destroyed after it returns).
1326
1049=cut 1327=cut
1050 1328
1051sub on_read { 1329sub on_read {
1052 my ($self, $cb) = @_; 1330 my ($self, $cb) = @_;
1053 1331
1055 $self->_drain_rbuf if $cb; 1333 $self->_drain_rbuf if $cb;
1056} 1334}
1057 1335
1058=item $handle->rbuf 1336=item $handle->rbuf
1059 1337
1060Returns the read buffer (as a modifiable lvalue). 1338Returns the read buffer (as a modifiable lvalue). You can also access the
1339read buffer directly as the C<< ->{rbuf} >> member, if you want (this is
1340much faster, and no less clean).
1061 1341
1062You can access the read buffer directly as the C<< ->{rbuf} >> 1342The only operation allowed on the read buffer (apart from looking at it)
1063member, if you want. However, the only operation allowed on the 1343is removing data from its beginning. Otherwise modifying or appending to
1064read buffer (apart from looking at it) is removing data from its 1344it is not allowed and will lead to hard-to-track-down bugs.
1065beginning. Otherwise modifying or appending to it is not allowed and will
1066lead to hard-to-track-down bugs.
1067 1345
1068NOTE: The read buffer should only be used or modified if the C<on_read>, 1346NOTE: The read buffer should only be used or modified in the C<on_read>
1069C<push_read> or C<unshift_read> methods are used. The other read methods 1347callback or when C<push_read> or C<unshift_read> are used with a single
1070automatically manage the read buffer. 1348callback (i.e. untyped). Typed C<push_read> and C<unshift_read> methods
1349will manage the read buffer on their own.
1071 1350
1072=cut 1351=cut
1073 1352
1074sub rbuf : lvalue { 1353sub rbuf : lvalue {
1075 $_[0]{rbuf} 1354 $_[0]{rbuf}
1092 1371
1093If enough data was available, then the callback must remove all data it is 1372If enough data was available, then the callback must remove all data it is
1094interested in (which can be none at all) and return a true value. After returning 1373interested in (which can be none at all) and return a true value. After returning
1095true, it will be removed from the queue. 1374true, it will be removed from the queue.
1096 1375
1376These methods may invoke callbacks (and therefore the handle might be
1377destroyed after it returns).
1378
1097=cut 1379=cut
1098 1380
1099our %RH; 1381our %RH;
1100 1382
1101sub register_read_type($$) { 1383sub register_read_type($$) {
1107 my $cb = pop; 1389 my $cb = pop;
1108 1390
1109 if (@_) { 1391 if (@_) {
1110 my $type = shift; 1392 my $type = shift;
1111 1393
1394 $cb = ($RH{$type} ||= _load_func "$type\::anyevent_read_type"
1112 $cb = ($RH{$type} or Carp::croak "unsupported type passed to AnyEvent::Handle::push_read") 1395 or Carp::croak "unsupported/unloadable type '$type' passed to AnyEvent::Handle::push_read")
1113 ->($self, $cb, @_); 1396 ->($self, $cb, @_);
1114 } 1397 }
1115 1398
1116 push @{ $self->{_queue} }, $cb; 1399 push @{ $self->{_queue} }, $cb;
1117 $self->_drain_rbuf; 1400 $self->_drain_rbuf;
1122 my $cb = pop; 1405 my $cb = pop;
1123 1406
1124 if (@_) { 1407 if (@_) {
1125 my $type = shift; 1408 my $type = shift;
1126 1409
1410 $cb = ($RH{$type} ||= _load_func "$type\::anyevent_read_type"
1127 $cb = ($RH{$type} or Carp::croak "unsupported type passed to AnyEvent::Handle::unshift_read") 1411 or Carp::croak "unsupported/unloadable type '$type' passed to AnyEvent::Handle::unshift_read")
1128 ->($self, $cb, @_); 1412 ->($self, $cb, @_);
1129 } 1413 }
1130
1131 1414
1132 unshift @{ $self->{_queue} }, $cb; 1415 unshift @{ $self->{_queue} }, $cb;
1133 $self->_drain_rbuf; 1416 $self->_drain_rbuf;
1134} 1417}
1135 1418
1137 1420
1138=item $handle->unshift_read (type => @args, $cb) 1421=item $handle->unshift_read (type => @args, $cb)
1139 1422
1140Instead of providing a callback that parses the data itself you can chose 1423Instead of providing a callback that parses the data itself you can chose
1141between a number of predefined parsing formats, for chunks of data, lines 1424between a number of predefined parsing formats, for chunks of data, lines
1142etc. 1425etc. You can also specify the (fully qualified) name of a package, in
1426which case AnyEvent tries to load the package and then expects to find the
1427C<anyevent_read_type> function inside (see "custom read types", below).
1143 1428
1144Predefined types are (if you have ideas for additional types, feel free to 1429Predefined types are (if you have ideas for additional types, feel free to
1145drop by and tell us): 1430drop by and tell us):
1146 1431
1147=over 4 1432=over 4
1153data. 1438data.
1154 1439
1155Example: read 2 bytes. 1440Example: read 2 bytes.
1156 1441
1157 $handle->push_read (chunk => 2, sub { 1442 $handle->push_read (chunk => 2, sub {
1158 warn "yay ", unpack "H*", $_[1]; 1443 say "yay " . unpack "H*", $_[1];
1159 }); 1444 });
1160 1445
1161=cut 1446=cut
1162 1447
1163register_read_type chunk => sub { 1448register_read_type chunk => sub {
1197 if (@_ < 3) { 1482 if (@_ < 3) {
1198 # this is more than twice as fast as the generic code below 1483 # this is more than twice as fast as the generic code below
1199 sub { 1484 sub {
1200 $_[0]{rbuf} =~ s/^([^\015\012]*)(\015?\012)// or return; 1485 $_[0]{rbuf} =~ s/^([^\015\012]*)(\015?\012)// or return;
1201 1486
1202 $cb->($_[0], $1, $2); 1487 $cb->($_[0], "$1", "$2");
1203 1 1488 1
1204 } 1489 }
1205 } else { 1490 } else {
1206 $eol = quotemeta $eol unless ref $eol; 1491 $eol = quotemeta $eol unless ref $eol;
1207 $eol = qr|^(.*?)($eol)|s; 1492 $eol = qr|^(.*?)($eol)|s;
1208 1493
1209 sub { 1494 sub {
1210 $_[0]{rbuf} =~ s/$eol// or return; 1495 $_[0]{rbuf} =~ s/$eol// or return;
1211 1496
1212 $cb->($_[0], $1, $2); 1497 $cb->($_[0], "$1", "$2");
1213 1 1498 1
1214 } 1499 }
1215 } 1500 }
1216}; 1501};
1217 1502
1239the receive buffer when neither C<$accept> nor C<$reject> match, 1524the receive buffer when neither C<$accept> nor C<$reject> match,
1240and everything preceding and including the match will be accepted 1525and everything preceding and including the match will be accepted
1241unconditionally. This is useful to skip large amounts of data that you 1526unconditionally. This is useful to skip large amounts of data that you
1242know cannot be matched, so that the C<$accept> or C<$reject> regex do not 1527know cannot be matched, so that the C<$accept> or C<$reject> regex do not
1243have to start matching from the beginning. This is purely an optimisation 1528have to start matching from the beginning. This is purely an optimisation
1244and is usually worth only when you expect more than a few kilobytes. 1529and is usually worth it only when you expect more than a few kilobytes.
1245 1530
1246Example: expect a http header, which ends at C<\015\012\015\012>. Since we 1531Example: expect a http header, which ends at C<\015\012\015\012>. Since we
1247expect the header to be very large (it isn't in practise, but...), we use 1532expect the header to be very large (it isn't in practice, but...), we use
1248a skip regex to skip initial portions. The skip regex is tricky in that 1533a skip regex to skip initial portions. The skip regex is tricky in that
1249it only accepts something not ending in either \015 or \012, as these are 1534it only accepts something not ending in either \015 or \012, as these are
1250required for the accept regex. 1535required for the accept regex.
1251 1536
1252 $handle->push_read (regex => 1537 $handle->push_read (regex =>
1265 1550
1266 sub { 1551 sub {
1267 # accept 1552 # accept
1268 if ($$rbuf =~ $accept) { 1553 if ($$rbuf =~ $accept) {
1269 $data .= substr $$rbuf, 0, $+[0], ""; 1554 $data .= substr $$rbuf, 0, $+[0], "";
1270 $cb->($self, $data); 1555 $cb->($_[0], $data);
1271 return 1; 1556 return 1;
1272 } 1557 }
1273 1558
1274 # reject 1559 # reject
1275 if ($reject && $$rbuf =~ $reject) { 1560 if ($reject && $$rbuf =~ $reject) {
1276 $self->_error (Errno::EBADMSG); 1561 $_[0]->_error (Errno::EBADMSG);
1277 } 1562 }
1278 1563
1279 # skip 1564 # skip
1280 if ($skip && $$rbuf =~ $skip) { 1565 if ($skip && $$rbuf =~ $skip) {
1281 $data .= substr $$rbuf, 0, $+[0], ""; 1566 $data .= substr $$rbuf, 0, $+[0], "";
1297 my ($self, $cb) = @_; 1582 my ($self, $cb) = @_;
1298 1583
1299 sub { 1584 sub {
1300 unless ($_[0]{rbuf} =~ s/^(0|[1-9][0-9]*)://) { 1585 unless ($_[0]{rbuf} =~ s/^(0|[1-9][0-9]*)://) {
1301 if ($_[0]{rbuf} =~ /[^0-9]/) { 1586 if ($_[0]{rbuf} =~ /[^0-9]/) {
1302 $self->_error (Errno::EBADMSG); 1587 $_[0]->_error (Errno::EBADMSG);
1303 } 1588 }
1304 return; 1589 return;
1305 } 1590 }
1306 1591
1307 my $len = $1; 1592 my $len = $1;
1308 1593
1309 $self->unshift_read (chunk => $len, sub { 1594 $_[0]->unshift_read (chunk => $len, sub {
1310 my $string = $_[1]; 1595 my $string = $_[1];
1311 $_[0]->unshift_read (chunk => 1, sub { 1596 $_[0]->unshift_read (chunk => 1, sub {
1312 if ($_[1] eq ",") { 1597 if ($_[1] eq ",") {
1313 $cb->($_[0], $string); 1598 $cb->($_[0], $string);
1314 } else { 1599 } else {
1315 $self->_error (Errno::EBADMSG); 1600 $_[0]->_error (Errno::EBADMSG);
1316 } 1601 }
1317 }); 1602 });
1318 }); 1603 });
1319 1604
1320 1 1605 1
1387=cut 1672=cut
1388 1673
1389register_read_type json => sub { 1674register_read_type json => sub {
1390 my ($self, $cb) = @_; 1675 my ($self, $cb) = @_;
1391 1676
1392 my $json = $self->{json} ||= 1677 my $json = $self->{json} ||= json_coder;
1393 eval { require JSON::XS; JSON::XS->new->utf8 }
1394 || do { require JSON; JSON->new->utf8 };
1395 1678
1396 my $data; 1679 my $data;
1397 my $rbuf = \$self->{rbuf}; 1680 my $rbuf = \$self->{rbuf};
1398 1681
1399 sub { 1682 sub {
1400 my $ref = eval { $json->incr_parse ($self->{rbuf}) }; 1683 my $ref = eval { $json->incr_parse ($_[0]{rbuf}) };
1401 1684
1402 if ($ref) { 1685 if ($ref) {
1403 $self->{rbuf} = $json->incr_text; 1686 $_[0]{rbuf} = $json->incr_text;
1404 $json->incr_text = ""; 1687 $json->incr_text = "";
1405 $cb->($self, $ref); 1688 $cb->($_[0], $ref);
1406 1689
1407 1 1690 1
1408 } elsif ($@) { 1691 } elsif ($@) {
1409 # error case 1692 # error case
1410 $json->incr_skip; 1693 $json->incr_skip;
1411 1694
1412 $self->{rbuf} = $json->incr_text; 1695 $_[0]{rbuf} = $json->incr_text;
1413 $json->incr_text = ""; 1696 $json->incr_text = "";
1414 1697
1415 $self->_error (Errno::EBADMSG); 1698 $_[0]->_error (Errno::EBADMSG);
1416 1699
1417 () 1700 ()
1418 } else { 1701 } else {
1419 $self->{rbuf} = ""; 1702 $_[0]{rbuf} = "";
1420 1703
1421 () 1704 ()
1422 } 1705 }
1423 } 1706 }
1424}; 1707};
1434=cut 1717=cut
1435 1718
1436register_read_type storable => sub { 1719register_read_type storable => sub {
1437 my ($self, $cb) = @_; 1720 my ($self, $cb) = @_;
1438 1721
1439 require Storable; 1722 require Storable unless $Storable::VERSION;
1440 1723
1441 sub { 1724 sub {
1442 # when we can use 5.10 we can use ".", but for 5.8 we use the re-pack method 1725 # when we can use 5.10 we can use ".", but for 5.8 we use the re-pack method
1443 defined (my $len = eval { unpack "w", $_[0]{rbuf} }) 1726 defined (my $len = eval { unpack "w", $_[0]{rbuf} })
1444 or return; 1727 or return;
1457 # read remaining chunk 1740 # read remaining chunk
1458 $_[0]->unshift_read (chunk => $len, sub { 1741 $_[0]->unshift_read (chunk => $len, sub {
1459 if (my $ref = eval { Storable::thaw ($_[1]) }) { 1742 if (my $ref = eval { Storable::thaw ($_[1]) }) {
1460 $cb->($_[0], $ref); 1743 $cb->($_[0], $ref);
1461 } else { 1744 } else {
1462 $self->_error (Errno::EBADMSG); 1745 $_[0]->_error (Errno::EBADMSG);
1463 } 1746 }
1464 }); 1747 });
1465 } 1748 }
1466 1749
1467 1 1750 1
1468 } 1751 }
1469}; 1752};
1470 1753
1471=back 1754=back
1472 1755
1473=item AnyEvent::Handle::register_read_type type => $coderef->($handle, $cb, @args) 1756=item custom read types - Package::anyevent_read_type $handle, $cb, @args
1474 1757
1475This function (not method) lets you add your own types to C<push_read>. 1758Instead of one of the predefined types, you can also specify the name
1759of a package. AnyEvent will try to load the package and then expects to
1760find a function named C<anyevent_read_type> inside. If it isn't found, it
1761progressively tries to load the parent package until it either finds the
1762function (good) or runs out of packages (bad).
1476 1763
1477Whenever the given C<type> is used, C<push_read> will invoke the code 1764Whenever this type is used, C<push_read> will invoke the function with the
1478reference with the handle object, the callback and the remaining 1765handle object, the original callback and the remaining arguments.
1479arguments.
1480 1766
1481The code reference is supposed to return a callback (usually a closure) 1767The function is supposed to return a callback (usually a closure) that
1482that works as a plain read callback (see C<< ->push_read ($cb) >>). 1768works as a plain read callback (see C<< ->push_read ($cb) >>), so you can
1769mentally treat the function as a "configurable read type to read callback"
1770converter.
1483 1771
1484It should invoke the passed callback when it is done reading (remember to 1772It should invoke the original callback when it is done reading (remember
1485pass C<$handle> as first argument as all other callbacks do that). 1773to pass C<$handle> as first argument as all other callbacks do that,
1774although there is no strict requirement on this).
1486 1775
1487Note that this is a function, and all types registered this way will be
1488global, so try to use unique names.
1489
1490For examples, see the source of this module (F<perldoc -m AnyEvent::Handle>, 1776For examples, see the source of this module (F<perldoc -m
1491search for C<register_read_type>)). 1777AnyEvent::Handle>, search for C<register_read_type>)).
1492 1778
1493=item $handle->stop_read 1779=item $handle->stop_read
1494 1780
1495=item $handle->start_read 1781=item $handle->start_read
1496 1782
1502Note that AnyEvent::Handle will automatically C<start_read> for you when 1788Note that AnyEvent::Handle will automatically C<start_read> for you when
1503you change the C<on_read> callback or push/unshift a read callback, and it 1789you change the C<on_read> callback or push/unshift a read callback, and it
1504will automatically C<stop_read> for you when neither C<on_read> is set nor 1790will automatically C<stop_read> for you when neither C<on_read> is set nor
1505there are any read requests in the queue. 1791there are any read requests in the queue.
1506 1792
1507These methods will have no effect when in TLS mode (as TLS doesn't support 1793In older versions of this module (<= 5.3), these methods had no effect,
1508half-duplex connections). 1794as TLS does not support half-duplex connections. In current versions they
1795work as expected, as this behaviour is required to avoid certain resource
1796attacks, where the program would be forced to read (and buffer) arbitrary
1797amounts of data before being able to send some data. The drawback is that
1798some readings of the the SSL/TLS specifications basically require this
1799attack to be working, as SSL/TLS implementations might stall sending data
1800during a rehandshake.
1801
1802As a guideline, during the initial handshake, you should not stop reading,
1803and as a client, it might cause problems, depending on your application.
1509 1804
1510=cut 1805=cut
1511 1806
1512sub stop_read { 1807sub stop_read {
1513 my ($self) = @_; 1808 my ($self) = @_;
1514 1809
1515 delete $self->{_rw} unless $self->{tls}; 1810 delete $self->{_rw};
1516} 1811}
1517 1812
1518sub start_read { 1813sub start_read {
1519 my ($self) = @_; 1814 my ($self) = @_;
1520 1815
1521 unless ($self->{_rw} || $self->{_eof}) { 1816 unless ($self->{_rw} || $self->{_eof} || !$self->{fh}) {
1522 Scalar::Util::weaken $self; 1817 Scalar::Util::weaken $self;
1523 1818
1524 $self->{_rw} = AnyEvent->io (fh => $self->{fh}, poll => "r", cb => sub { 1819 $self->{_rw} = AE::io $self->{fh}, 0, sub {
1525 my $rbuf = \($self->{tls} ? my $buf : $self->{rbuf}); 1820 my $rbuf = \($self->{tls} ? my $buf : $self->{rbuf});
1526 my $len = sysread $self->{fh}, $$rbuf, $self->{read_size} || 8192, length $$rbuf; 1821 my $len = sysread $self->{fh}, $$rbuf, $self->{read_size}, length $$rbuf;
1527 1822
1528 if ($len > 0) { 1823 if ($len > 0) {
1529 $self->{_activity} = AnyEvent->now; 1824 $self->{_activity} = $self->{_ractivity} = AE::now;
1530 1825
1531 if ($self->{tls}) { 1826 if ($self->{tls}) {
1532 Net::SSLeay::BIO_write ($self->{_rbio}, $$rbuf); 1827 Net::SSLeay::BIO_write ($self->{_rbio}, $$rbuf);
1533 1828
1534 &_dotls ($self); 1829 &_dotls ($self);
1535 } else { 1830 } else {
1536 $self->_drain_rbuf; 1831 $self->_drain_rbuf;
1537 } 1832 }
1538 1833
1834 if ($len == $self->{read_size}) {
1835 $self->{read_size} *= 2;
1836 $self->{read_size} = $self->{max_read_size} || MAX_READ_SIZE
1837 if $self->{read_size} > ($self->{max_read_size} || MAX_READ_SIZE);
1838 }
1839
1539 } elsif (defined $len) { 1840 } elsif (defined $len) {
1540 delete $self->{_rw}; 1841 delete $self->{_rw};
1541 $self->{_eof} = 1; 1842 $self->{_eof} = 1;
1542 $self->_drain_rbuf; 1843 $self->_drain_rbuf;
1543 1844
1544 } elsif ($! != EAGAIN && $! != EINTR && $! != WSAEWOULDBLOCK) { 1845 } elsif ($! != EAGAIN && $! != EINTR && $! != WSAEWOULDBLOCK) {
1545 return $self->_error ($!, 1); 1846 return $self->_error ($!, 1);
1546 } 1847 }
1547 }); 1848 };
1548 } 1849 }
1549} 1850}
1550 1851
1551our $ERROR_SYSCALL; 1852our $ERROR_SYSCALL;
1552our $ERROR_WANT_READ; 1853our $ERROR_WANT_READ;
1619 && ($tmp != $ERROR_SYSCALL || $!); 1920 && ($tmp != $ERROR_SYSCALL || $!);
1620 1921
1621 while (length ($tmp = Net::SSLeay::BIO_read ($self->{_wbio}))) { 1922 while (length ($tmp = Net::SSLeay::BIO_read ($self->{_wbio}))) {
1622 $self->{wbuf} .= $tmp; 1923 $self->{wbuf} .= $tmp;
1623 $self->_drain_wbuf; 1924 $self->_drain_wbuf;
1925 $self->{tls} or return; # tls session might have gone away in callback
1624 } 1926 }
1625 1927
1626 $self->{_on_starttls} 1928 $self->{_on_starttls}
1627 and Net::SSLeay::state ($self->{tls}) == Net::SSLeay::ST_OK () 1929 and Net::SSLeay::state ($self->{tls}) == Net::SSLeay::ST_OK ()
1628 and (delete $self->{_on_starttls})->($self, 1, "TLS/SSL connection established"); 1930 and (delete $self->{_on_starttls})->($self, 1, "TLS/SSL connection established");
1649The TLS connection object will end up in C<< $handle->{tls} >>, the TLS 1951The TLS connection object will end up in C<< $handle->{tls} >>, the TLS
1650context in C<< $handle->{tls_ctx} >> after this call and can be used or 1952context in C<< $handle->{tls_ctx} >> after this call and can be used or
1651changed to your liking. Note that the handshake might have already started 1953changed to your liking. Note that the handshake might have already started
1652when this function returns. 1954when this function returns.
1653 1955
1654If it an error to start a TLS handshake more than once per 1956Due to bugs in OpenSSL, it might or might not be possible to do multiple
1655AnyEvent::Handle object (this is due to bugs in OpenSSL). 1957handshakes on the same stream. It is best to not attempt to use the
1958stream after stopping TLS.
1959
1960This method may invoke callbacks (and therefore the handle might be
1961destroyed after it returns).
1656 1962
1657=cut 1963=cut
1658 1964
1659our %TLS_CACHE; #TODO not yet documented, should we? 1965our %TLS_CACHE; #TODO not yet documented, should we?
1660 1966
1661sub starttls { 1967sub starttls {
1662 my ($self, $ssl, $ctx) = @_; 1968 my ($self, $tls, $ctx) = @_;
1969
1970 Carp::croak "It is an error to call starttls on an AnyEvent::Handle object while TLS is already active, caught"
1971 if $self->{tls};
1972
1973 $self->{tls} = $tls;
1974 $self->{tls_ctx} = $ctx if @_ > 2;
1975
1976 return unless $self->{fh};
1663 1977
1664 require Net::SSLeay; 1978 require Net::SSLeay;
1665
1666 Carp::croak "it is an error to call starttls more than once on an AnyEvent::Handle object"
1667 if $self->{tls};
1668 1979
1669 $ERROR_SYSCALL = Net::SSLeay::ERROR_SYSCALL (); 1980 $ERROR_SYSCALL = Net::SSLeay::ERROR_SYSCALL ();
1670 $ERROR_WANT_READ = Net::SSLeay::ERROR_WANT_READ (); 1981 $ERROR_WANT_READ = Net::SSLeay::ERROR_WANT_READ ();
1671 1982
1983 $tls = delete $self->{tls};
1672 $ctx ||= $self->{tls_ctx}; 1984 $ctx = $self->{tls_ctx};
1673 1985
1674 local $Carp::CarpLevel = 1; # skip ourselves when creating a new context or session 1986 local $Carp::CarpLevel = 1; # skip ourselves when creating a new context or session
1675 1987
1676 if ("HASH" eq ref $ctx) { 1988 if ("HASH" eq ref $ctx) {
1677 require AnyEvent::TLS; 1989 require AnyEvent::TLS;
1683 $ctx = new AnyEvent::TLS %$ctx; 1995 $ctx = new AnyEvent::TLS %$ctx;
1684 } 1996 }
1685 } 1997 }
1686 1998
1687 $self->{tls_ctx} = $ctx || TLS_CTX (); 1999 $self->{tls_ctx} = $ctx || TLS_CTX ();
1688 $self->{tls} = $ssl = $self->{tls_ctx}->_get_session ($ssl, $self, $self->{peername}); 2000 $self->{tls} = $tls = $self->{tls_ctx}->_get_session ($tls, $self, $self->{peername});
1689 2001
1690 # basically, this is deep magic (because SSL_read should have the same issues) 2002 # basically, this is deep magic (because SSL_read should have the same issues)
1691 # but the openssl maintainers basically said: "trust us, it just works". 2003 # but the openssl maintainers basically said: "trust us, it just works".
1692 # (unfortunately, we have to hardcode constants because the abysmally misdesigned 2004 # (unfortunately, we have to hardcode constants because the abysmally misdesigned
1693 # and mismaintained ssleay-module doesn't even offer them). 2005 # and mismaintained ssleay-module doesn't even offer them).
1700 # and we drive openssl fully in blocking mode here. Or maybe we don't - openssl seems to 2012 # and we drive openssl fully in blocking mode here. Or maybe we don't - openssl seems to
1701 # have identity issues in that area. 2013 # have identity issues in that area.
1702# Net::SSLeay::CTX_set_mode ($ssl, 2014# Net::SSLeay::CTX_set_mode ($ssl,
1703# (eval { local $SIG{__DIE__}; Net::SSLeay::MODE_ENABLE_PARTIAL_WRITE () } || 1) 2015# (eval { local $SIG{__DIE__}; Net::SSLeay::MODE_ENABLE_PARTIAL_WRITE () } || 1)
1704# | (eval { local $SIG{__DIE__}; Net::SSLeay::MODE_ACCEPT_MOVING_WRITE_BUFFER () } || 2)); 2016# | (eval { local $SIG{__DIE__}; Net::SSLeay::MODE_ACCEPT_MOVING_WRITE_BUFFER () } || 2));
1705 Net::SSLeay::CTX_set_mode ($ssl, 1|2); 2017 Net::SSLeay::CTX_set_mode ($tls, 1|2);
1706 2018
1707 $self->{_rbio} = Net::SSLeay::BIO_new (Net::SSLeay::BIO_s_mem ()); 2019 $self->{_rbio} = Net::SSLeay::BIO_new (Net::SSLeay::BIO_s_mem ());
1708 $self->{_wbio} = Net::SSLeay::BIO_new (Net::SSLeay::BIO_s_mem ()); 2020 $self->{_wbio} = Net::SSLeay::BIO_new (Net::SSLeay::BIO_s_mem ());
1709 2021
2022 Net::SSLeay::BIO_write ($self->{_rbio}, $self->{rbuf});
2023 $self->{rbuf} = "";
2024
1710 Net::SSLeay::set_bio ($ssl, $self->{_rbio}, $self->{_wbio}); 2025 Net::SSLeay::set_bio ($tls, $self->{_rbio}, $self->{_wbio});
1711 2026
1712 $self->{_on_starttls} = sub { $_[0]{on_starttls}(@_) } 2027 $self->{_on_starttls} = sub { $_[0]{on_starttls}(@_) }
1713 if $self->{on_starttls}; 2028 if $self->{on_starttls};
1714 2029
1715 &_dotls; # need to trigger the initial handshake 2030 &_dotls; # need to trigger the initial handshake
1718 2033
1719=item $handle->stoptls 2034=item $handle->stoptls
1720 2035
1721Shuts down the SSL connection - this makes a proper EOF handshake by 2036Shuts down the SSL connection - this makes a proper EOF handshake by
1722sending a close notify to the other side, but since OpenSSL doesn't 2037sending a close notify to the other side, but since OpenSSL doesn't
1723support non-blocking shut downs, it is not possible to re-use the stream 2038support non-blocking shut downs, it is not guaranteed that you can re-use
1724afterwards. 2039the stream afterwards.
2040
2041This method may invoke callbacks (and therefore the handle might be
2042destroyed after it returns).
1725 2043
1726=cut 2044=cut
1727 2045
1728sub stoptls { 2046sub stoptls {
1729 my ($self) = @_; 2047 my ($self) = @_;
1730 2048
1731 if ($self->{tls}) { 2049 if ($self->{tls} && $self->{fh}) {
1732 Net::SSLeay::shutdown ($self->{tls}); 2050 Net::SSLeay::shutdown ($self->{tls});
1733 2051
1734 &_dotls; 2052 &_dotls;
1735 2053
1736# # we don't give a shit. no, we do, but we can't. no...#d# 2054# # we don't give a shit. no, we do, but we can't. no...#d#
1742sub _freetls { 2060sub _freetls {
1743 my ($self) = @_; 2061 my ($self) = @_;
1744 2062
1745 return unless $self->{tls}; 2063 return unless $self->{tls};
1746 2064
1747 $self->{tls_ctx}->_put_session (delete $self->{tls}); 2065 $self->{tls_ctx}->_put_session (delete $self->{tls})
2066 if $self->{tls} > 0;
1748 2067
1749 delete @$self{qw(_rbio _wbio _tls_wbuf _on_starttls)}; 2068 delete @$self{qw(_rbio _wbio _tls_wbuf _on_starttls)};
1750} 2069}
2070
2071=item $handle->resettls
2072
2073This rarely-used method simply resets and TLS state on the handle, usually
2074causing data loss.
2075
2076One case where it may be useful is when you want to skip over the data in
2077the stream but you are not interested in interpreting it, so data loss is
2078no concern.
2079
2080=cut
2081
2082*resettls = \&_freetls;
1751 2083
1752sub DESTROY { 2084sub DESTROY {
1753 my ($self) = @_; 2085 my ($self) = @_;
1754 2086
1755 &_freetls; 2087 &_freetls;
1760 my $fh = delete $self->{fh}; 2092 my $fh = delete $self->{fh};
1761 my $wbuf = delete $self->{wbuf}; 2093 my $wbuf = delete $self->{wbuf};
1762 2094
1763 my @linger; 2095 my @linger;
1764 2096
1765 push @linger, AnyEvent->io (fh => $fh, poll => "w", cb => sub { 2097 push @linger, AE::io $fh, 1, sub {
1766 my $len = syswrite $fh, $wbuf, length $wbuf; 2098 my $len = syswrite $fh, $wbuf, length $wbuf;
1767 2099
1768 if ($len > 0) { 2100 if ($len > 0) {
1769 substr $wbuf, 0, $len, ""; 2101 substr $wbuf, 0, $len, "";
1770 } else { 2102 } elsif (defined $len || ($! != EAGAIN && $! != EINTR && $! != WSAEWOULDBLOCK)) {
1771 @linger = (); # end 2103 @linger = (); # end
1772 } 2104 }
1773 }); 2105 };
1774 push @linger, AnyEvent->timer (after => $linger, cb => sub { 2106 push @linger, AE::timer $linger, 0, sub {
1775 @linger = (); 2107 @linger = ();
1776 }); 2108 };
1777 } 2109 }
1778} 2110}
1779 2111
1780=item $handle->destroy 2112=item $handle->destroy
1781 2113
1782Shuts down the handle object as much as possible - this call ensures that 2114Shuts down the handle object as much as possible - this call ensures that
1783no further callbacks will be invoked and as many resources as possible 2115no further callbacks will be invoked and as many resources as possible
1784will be freed. You must not call any methods on the object afterwards. 2116will be freed. Any method you will call on the handle object after
2117destroying it in this way will be silently ignored (and it will return the
2118empty list).
1785 2119
1786Normally, you can just "forget" any references to an AnyEvent::Handle 2120Normally, you can just "forget" any references to an AnyEvent::Handle
1787object and it will simply shut down. This works in fatal error and EOF 2121object and it will simply shut down. This works in fatal error and EOF
1788callbacks, as well as code outside. It does I<NOT> work in a read or write 2122callbacks, as well as code outside. It does I<NOT> work in a read or write
1789callback, so when you want to destroy the AnyEvent::Handle object from 2123callback, so when you want to destroy the AnyEvent::Handle object from
1803sub destroy { 2137sub destroy {
1804 my ($self) = @_; 2138 my ($self) = @_;
1805 2139
1806 $self->DESTROY; 2140 $self->DESTROY;
1807 %$self = (); 2141 %$self = ();
2142 bless $self, "AnyEvent::Handle::destroyed";
1808} 2143}
2144
2145sub AnyEvent::Handle::destroyed::AUTOLOAD {
2146 #nop
2147}
2148
2149=item $handle->destroyed
2150
2151Returns false as long as the handle hasn't been destroyed by a call to C<<
2152->destroy >>, true otherwise.
2153
2154Can be useful to decide whether the handle is still valid after some
2155callback possibly destroyed the handle. For example, C<< ->push_write >>,
2156C<< ->starttls >> and other methods can call user callbacks, which in turn
2157can destroy the handle, so work can be avoided by checking sometimes:
2158
2159 $hdl->starttls ("accept");
2160 return if $hdl->destroyed;
2161 $hdl->push_write (...
2162
2163Note that the call to C<push_write> will silently be ignored if the handle
2164has been destroyed, so often you can just ignore the possibility of the
2165handle being destroyed.
2166
2167=cut
2168
2169sub destroyed { 0 }
2170sub AnyEvent::Handle::destroyed::destroyed { 1 }
1809 2171
1810=item AnyEvent::Handle::TLS_CTX 2172=item AnyEvent::Handle::TLS_CTX
1811 2173
1812This function creates and returns the AnyEvent::TLS object used by default 2174This function creates and returns the AnyEvent::TLS object used by default
1813for TLS mode. 2175for TLS mode.
1841 2203
1842It is only safe to "forget" the reference inside EOF or error callbacks, 2204It is only safe to "forget" the reference inside EOF or error callbacks,
1843from within all other callbacks, you need to explicitly call the C<< 2205from within all other callbacks, you need to explicitly call the C<<
1844->destroy >> method. 2206->destroy >> method.
1845 2207
2208=item Why is my C<on_eof> callback never called?
2209
2210Probably because your C<on_error> callback is being called instead: When
2211you have outstanding requests in your read queue, then an EOF is
2212considered an error as you clearly expected some data.
2213
2214To avoid this, make sure you have an empty read queue whenever your handle
2215is supposed to be "idle" (i.e. connection closes are O.K.). You can set
2216an C<on_read> handler that simply pushes the first read requests in the
2217queue.
2218
2219See also the next question, which explains this in a bit more detail.
2220
2221=item How can I serve requests in a loop?
2222
2223Most protocols consist of some setup phase (authentication for example)
2224followed by a request handling phase, where the server waits for requests
2225and handles them, in a loop.
2226
2227There are two important variants: The first (traditional, better) variant
2228handles requests until the server gets some QUIT command, causing it to
2229close the connection first (highly desirable for a busy TCP server). A
2230client dropping the connection is an error, which means this variant can
2231detect an unexpected detection close.
2232
2233To handle this case, always make sure you have a on-empty read queue, by
2234pushing the "read request start" handler on it:
2235
2236 # we assume a request starts with a single line
2237 my @start_request; @start_request = (line => sub {
2238 my ($hdl, $line) = @_;
2239
2240 ... handle request
2241
2242 # push next request read, possibly from a nested callback
2243 $hdl->push_read (@start_request);
2244 });
2245
2246 # auth done, now go into request handling loop
2247 # now push the first @start_request
2248 $hdl->push_read (@start_request);
2249
2250By always having an outstanding C<push_read>, the handle always expects
2251some data and raises the C<EPIPE> error when the connction is dropped
2252unexpectedly.
2253
2254The second variant is a protocol where the client can drop the connection
2255at any time. For TCP, this means that the server machine may run out of
2256sockets easier, and in general, it means you cannot distinguish a protocl
2257failure/client crash from a normal connection close. Nevertheless, these
2258kinds of protocols are common (and sometimes even the best solution to the
2259problem).
2260
2261Having an outstanding read request at all times is possible if you ignore
2262C<EPIPE> errors, but this doesn't help with when the client drops the
2263connection during a request, which would still be an error.
2264
2265A better solution is to push the initial request read in an C<on_read>
2266callback. This avoids an error, as when the server doesn't expect data
2267(i.e. is idly waiting for the next request, an EOF will not raise an
2268error, but simply result in an C<on_eof> callback. It is also a bit slower
2269and simpler:
2270
2271 # auth done, now go into request handling loop
2272 $hdl->on_read (sub {
2273 my ($hdl) = @_;
2274
2275 # called each time we receive data but the read queue is empty
2276 # simply start read the request
2277
2278 $hdl->push_read (line => sub {
2279 my ($hdl, $line) = @_;
2280
2281 ... handle request
2282
2283 # do nothing special when the request has been handled, just
2284 # let the request queue go empty.
2285 });
2286 });
2287
1846=item I get different callback invocations in TLS mode/Why can't I pause 2288=item I get different callback invocations in TLS mode/Why can't I pause
1847reading? 2289reading?
1848 2290
1849Unlike, say, TCP, TLS connections do not consist of two independent 2291Unlike, say, TCP, TLS connections do not consist of two independent
1850communication channels, one for each direction. Or put differently. The 2292communication channels, one for each direction. Or put differently, the
1851read and write directions are not independent of each other: you cannot 2293read and write directions are not independent of each other: you cannot
1852write data unless you are also prepared to read, and vice versa. 2294write data unless you are also prepared to read, and vice versa.
1853 2295
1854This can mean than, in TLS mode, you might get C<on_error> or C<on_eof> 2296This means that, in TLS mode, you might get C<on_error> or C<on_eof>
1855callback invocations when you are not expecting any read data - the reason 2297callback invocations when you are not expecting any read data - the reason
1856is that AnyEvent::Handle always reads in TLS mode. 2298is that AnyEvent::Handle always reads in TLS mode.
1857 2299
1858During the connection, you have to make sure that you always have a 2300During the connection, you have to make sure that you always have a
1859non-empty read-queue, or an C<on_read> watcher. At the end of the 2301non-empty read-queue, or an C<on_read> watcher. At the end of the
1871 $handle->on_eof (undef); 2313 $handle->on_eof (undef);
1872 $handle->on_error (sub { 2314 $handle->on_error (sub {
1873 my $data = delete $_[0]{rbuf}; 2315 my $data = delete $_[0]{rbuf};
1874 }); 2316 });
1875 2317
2318Note that this example removes the C<rbuf> member from the handle object,
2319which is not normally allowed by the API. It is expressly permitted in
2320this case only, as the handle object needs to be destroyed afterwards.
2321
1876The reason to use C<on_error> is that TCP connections, due to latencies 2322The reason to use C<on_error> is that TCP connections, due to latencies
1877and packets loss, might get closed quite violently with an error, when in 2323and packets loss, might get closed quite violently with an error, when in
1878fact, all data has been received. 2324fact all data has been received.
1879 2325
1880It is usually better to use acknowledgements when transferring data, 2326It is usually better to use acknowledgements when transferring data,
1881to make sure the other side hasn't just died and you got the data 2327to make sure the other side hasn't just died and you got the data
1882intact. This is also one reason why so many internet protocols have an 2328intact. This is also one reason why so many internet protocols have an
1883explicit QUIT command. 2329explicit QUIT command.
1890C<low_water_mark> this will be called precisely when all data has been 2336C<low_water_mark> this will be called precisely when all data has been
1891written to the socket: 2337written to the socket:
1892 2338
1893 $handle->push_write (...); 2339 $handle->push_write (...);
1894 $handle->on_drain (sub { 2340 $handle->on_drain (sub {
1895 warn "all data submitted to the kernel\n"; 2341 AE::log debug => "all data submitted to the kernel\n";
1896 undef $handle; 2342 undef $handle;
1897 }); 2343 });
1898 2344
1899If you just want to queue some data and then signal EOF to the other side, 2345If you just want to queue some data and then signal EOF to the other side,
1900consider using C<< ->push_shutdown >> instead. 2346consider using C<< ->push_shutdown >> instead.
1901 2347
1902=item I want to contact a TLS/SSL server, I don't care about security. 2348=item I want to contact a TLS/SSL server, I don't care about security.
1903 2349
1904If your TLS server is a pure TLS server (e.g. HTTPS) that only speaks TLS, 2350If your TLS server is a pure TLS server (e.g. HTTPS) that only speaks TLS,
1905simply connect to it and then create the AnyEvent::Handle with the C<tls> 2351connect to it and then create the AnyEvent::Handle with the C<tls>
1906parameter: 2352parameter:
1907 2353
1908 tcp_connect $host, $port, sub { 2354 tcp_connect $host, $port, sub {
1909 my ($fh) = @_; 2355 my ($fh) = @_;
1910 2356
2010 2456
2011=item * all members not documented here and not prefixed with an underscore 2457=item * all members not documented here and not prefixed with an underscore
2012are free to use in subclasses. 2458are free to use in subclasses.
2013 2459
2014Of course, new versions of AnyEvent::Handle may introduce more "public" 2460Of course, new versions of AnyEvent::Handle may introduce more "public"
2015member variables, but thats just life, at least it is documented. 2461member variables, but that's just life. At least it is documented.
2016 2462
2017=back 2463=back
2018 2464
2019=head1 AUTHOR 2465=head1 AUTHOR
2020 2466

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines