ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/AnyEvent/lib/AnyEvent/Handle.pm
Revision: 1.16
Committed: Fri May 23 05:16:57 2008 UTC (16 years ago) by root
Branch: MAIN
Changes since 1.15: +8 -4 lines
Log Message:
implement vc support

File Contents

# Content
1 package AnyEvent::Handle;
2
3 no warnings;
4 use strict;
5
6 use AnyEvent ();
7 use AnyEvent::Util ();
8 use Scalar::Util ();
9 use Carp ();
10 use Fcntl ();
11 use Errno qw/EAGAIN EINTR/;
12
13 =head1 NAME
14
15 AnyEvent::Handle - non-blocking I/O on filehandles via AnyEvent
16
17 This module is experimental.
18
19 =cut
20
21 our $VERSION = '0.04';
22
23 =head1 SYNOPSIS
24
25 use AnyEvent;
26 use AnyEvent::Handle;
27
28 my $cv = AnyEvent->condvar;
29
30 my $ae_fh = AnyEvent::Handle->new (fh => \*STDIN);
31
32 #TODO
33
34 # or use the constructor to pass the callback:
35
36 my $ae_fh2 =
37 AnyEvent::Handle->new (
38 fh => \*STDIN,
39 on_eof => sub {
40 $cv->broadcast;
41 },
42 #TODO
43 );
44
45 $cv->wait;
46
47 =head1 DESCRIPTION
48
49 This module is a helper module to make it easier to do event-based I/O on
50 filehandles. For utility functions for doing non-blocking connects and accepts
51 on sockets see L<AnyEvent::Util>.
52
53 In the following, when the documentation refers to of "bytes" then this
54 means characters. As sysread and syswrite are used for all I/O, their
55 treatment of characters applies to this module as well.
56
57 All callbacks will be invoked with the handle object as their first
58 argument.
59
60 =head1 METHODS
61
62 =over 4
63
64 =item B<new (%args)>
65
66 The constructor supports these arguments (all as key => value pairs).
67
68 =over 4
69
70 =item fh => $filehandle [MANDATORY]
71
72 The filehandle this L<AnyEvent::Handle> object will operate on.
73
74 NOTE: The filehandle will be set to non-blocking (using
75 AnyEvent::Util::fh_nonblocking).
76
77 =item on_eof => $cb->($self)
78
79 Set the callback to be called on EOF.
80
81 While not mandatory, it is highly recommended to set an eof callback,
82 otherwise you might end up with a closed socket while you are still
83 waiting for data.
84
85 =item on_error => $cb->($self)
86
87 This is the fatal error callback, that is called when, well, a fatal error
88 ocurs, such as not being able to resolve the hostname, failure to connect
89 or a read error.
90
91 The object will not be in a usable state when this callback has been
92 called.
93
94 On callback entrance, the value of C<$!> contains the operating system
95 error (or C<ENOSPC> or C<EPIPE>).
96
97 While not mandatory, it is I<highly> recommended to set this callback, as
98 you will not be notified of errors otherwise. The default simply calls
99 die.
100
101 =item on_read => $cb->($self)
102
103 This sets the default read callback, which is called when data arrives
104 and no read request is in the queue.
105
106 To access (and remove data from) the read buffer, use the C<< ->rbuf >>
107 method or acces sthe C<$self->{rbuf}> member directly.
108
109 When an EOF condition is detected then AnyEvent::Handle will first try to
110 feed all the remaining data to the queued callbacks and C<on_read> before
111 calling the C<on_eof> callback. If no progress can be made, then a fatal
112 error will be raised (with C<$!> set to C<EPIPE>).
113
114 =item on_drain => $cb->()
115
116 This sets the callback that is called when the write buffer becomes empty
117 (or when the callback is set and the buffer is empty already).
118
119 To append to the write buffer, use the C<< ->push_write >> method.
120
121 =item rbuf_max => <bytes>
122
123 If defined, then a fatal error will be raised (with C<$!> set to C<ENOSPC>)
124 when the read buffer ever (strictly) exceeds this size. This is useful to
125 avoid denial-of-service attacks.
126
127 For example, a server accepting connections from untrusted sources should
128 be configured to accept only so-and-so much data that it cannot act on
129 (for example, when expecting a line, an attacker could send an unlimited
130 amount of data without a callback ever being called as long as the line
131 isn't finished).
132
133 =item read_size => <bytes>
134
135 The default read block size (the amount of bytes this module will try to read
136 on each [loop iteration). Default: C<4096>.
137
138 =item low_water_mark => <bytes>
139
140 Sets the amount of bytes (default: C<0>) that make up an "empty" write
141 buffer: If the write reaches this size or gets even samller it is
142 considered empty.
143
144 =back
145
146 =cut
147
148 sub new {
149 my $class = shift;
150
151 my $self = bless { @_ }, $class;
152
153 $self->{fh} or Carp::croak "mandatory argument fh is missing";
154
155 AnyEvent::Util::fh_nonblocking $self->{fh}, 1;
156
157 $self->on_eof (delete $self->{on_eof} ) if $self->{on_eof};
158 $self->on_error (delete $self->{on_error}) if $self->{on_error};
159 $self->on_drain (delete $self->{on_drain}) if $self->{on_drain};
160 $self->on_read (delete $self->{on_read} ) if $self->{on_read};
161
162 $self->start_read;
163
164 $self
165 }
166
167 sub _shutdown {
168 my ($self) = @_;
169
170 delete $self->{rw};
171 delete $self->{ww};
172 delete $self->{fh};
173 }
174
175 sub error {
176 my ($self) = @_;
177
178 {
179 local $!;
180 $self->_shutdown;
181 }
182
183 if ($self->{on_error}) {
184 $self->{on_error}($self);
185 } else {
186 die "AnyEvent::Handle uncaught fatal error: $!";
187 }
188 }
189
190 =item $fh = $handle->fh
191
192 This method returns the filehandle of the L<AnyEvent::Handle> object.
193
194 =cut
195
196 sub fh { $_[0]->{fh} }
197
198 =item $handle->on_error ($cb)
199
200 Replace the current C<on_error> callback (see the C<on_error> constructor argument).
201
202 =cut
203
204 sub on_error {
205 $_[0]{on_error} = $_[1];
206 }
207
208 =item $handle->on_eof ($cb)
209
210 Replace the current C<on_eof> callback (see the C<on_eof> constructor argument).
211
212 =cut
213
214 sub on_eof {
215 $_[0]{on_eof} = $_[1];
216 }
217
218 #############################################################################
219
220 =back
221
222 =head2 WRITE QUEUE
223
224 AnyEvent::Handle manages two queues per handle, one for writing and one
225 for reading.
226
227 The write queue is very simple: you can add data to its end, and
228 AnyEvent::Handle will automatically try to get rid of it for you.
229
230 When data could be writtena nd the write buffer is shorter then the low
231 water mark, the C<on_drain> callback will be invoked.
232
233 =over 4
234
235 =item $handle->on_drain ($cb)
236
237 Sets the C<on_drain> callback or clears it (see the description of
238 C<on_drain> in the constructor).
239
240 =cut
241
242 sub on_drain {
243 my ($self, $cb) = @_;
244
245 $self->{on_drain} = $cb;
246
247 $cb->($self)
248 if $cb && $self->{low_water_mark} >= length $self->{wbuf};
249 }
250
251 =item $handle->push_write ($data)
252
253 Queues the given scalar to be written. You can push as much data as you
254 want (only limited by the available memory), as C<AnyEvent::Handle>
255 buffers it independently of the kernel.
256
257 =cut
258
259 sub push_write {
260 my ($self, $data) = @_;
261
262 $self->{wbuf} .= $data;
263
264 unless ($self->{ww}) {
265 Scalar::Util::weaken $self;
266 my $cb = sub {
267 my $len = syswrite $self->{fh}, $self->{wbuf};
268
269 if ($len > 0) {
270 substr $self->{wbuf}, 0, $len, "";
271
272
273 $self->{on_drain}($self)
274 if $self->{low_water_mark} >= length $self->{wbuf}
275 && $self->{on_drain};
276
277 delete $self->{ww} unless length $self->{wbuf};
278 } elsif ($! != EAGAIN && $! != EINTR) {
279 $self->error;
280 }
281 };
282
283 $self->{ww} = AnyEvent->io (fh => $self->{fh}, poll => "w", cb => $cb);
284
285 $cb->($self);
286 };
287 }
288
289 #############################################################################
290
291 =back
292
293 =head2 READ QUEUE
294
295 AnyEvent::Handle manages two queues per handle, one for writing and one
296 for reading.
297
298 The read queue is more complex than the write queue. It can be used in two
299 ways, the "simple" way, using only C<on_read> and the "complex" way, using
300 a queue.
301
302 In the simple case, you just install an C<on_read> callback and whenever
303 new data arrives, it will be called. You can then remove some data (if
304 enough is there) from the read buffer (C<< $handle->rbuf >>) if you want
305 or not.
306
307 In the more complex case, you want to queue multiple callbacks. In this
308 case, AnyEvent::Handle will call the first queued callback each time new
309 data arrives and removes it when it has done its job (see C<push_read>,
310 below).
311
312 This way you can, for example, push three line-reads, followed by reading
313 a chunk of data, and AnyEvent::Handle will execute them in order.
314
315 Example 1: EPP protocol parser. EPP sends 4 byte length info, followed by
316 the specified number of bytes which give an XML datagram.
317
318 # in the default state, expect some header bytes
319 $handle->on_read (sub {
320 # some data is here, now queue the length-header-read (4 octets)
321 shift->unshift_read_chunk (4, sub {
322 # header arrived, decode
323 my $len = unpack "N", $_[1];
324
325 # now read the payload
326 shift->unshift_read_chunk ($len, sub {
327 my $xml = $_[1];
328 # handle xml
329 });
330 });
331 });
332
333 Example 2: Implement a client for a protocol that replies either with
334 "OK" and another line or "ERROR" for one request, and 64 bytes for the
335 second request. Due tot he availability of a full queue, we can just
336 pipeline sending both requests and manipulate the queue as necessary in
337 the callbacks:
338
339 # request one
340 $handle->push_write ("request 1\015\012");
341
342 # we expect "ERROR" or "OK" as response, so push a line read
343 $handle->push_read_line (sub {
344 # if we got an "OK", we have to _prepend_ another line,
345 # so it will be read before the second request reads its 64 bytes
346 # which are already in the queue when this callback is called
347 # we don't do this in case we got an error
348 if ($_[1] eq "OK") {
349 $_[0]->unshift_read_line (sub {
350 my $response = $_[1];
351 ...
352 });
353 }
354 });
355
356 # request two
357 $handle->push_write ("request 2\015\012");
358
359 # simply read 64 bytes, always
360 $handle->push_read_chunk (64, sub {
361 my $response = $_[1];
362 ...
363 });
364
365 =over 4
366
367 =cut
368
369 sub _drain_rbuf {
370 my ($self) = @_;
371
372 return if $self->{in_drain};
373 local $self->{in_drain} = 1;
374
375 while (my $len = length $self->{rbuf}) {
376 no strict 'refs';
377 if (my $cb = shift @{ $self->{queue} }) {
378 if (!$cb->($self)) {
379 if ($self->{eof}) {
380 # no progress can be made (not enough data and no data forthcoming)
381 $! = &Errno::EPIPE; return $self->error;
382 }
383
384 unshift @{ $self->{queue} }, $cb;
385 return;
386 }
387 } elsif ($self->{on_read}) {
388 $self->{on_read}($self);
389
390 if (
391 $self->{eof} # if no further data will arrive
392 && $len == length $self->{rbuf} # and no data has been consumed
393 && !@{ $self->{queue} } # and the queue is still empty
394 && $self->{on_read} # and we still want to read data
395 ) {
396 # then no progress can be made
397 $! = &Errno::EPIPE; return $self->error;
398 }
399 } else {
400 # read side becomes idle
401 delete $self->{rw};
402 return;
403 }
404 }
405
406 if ($self->{eof}) {
407 $self->_shutdown;
408 $self->{on_eof}($self)
409 if $self->{on_eof};
410 }
411 }
412
413 =item $handle->on_read ($cb)
414
415 This replaces the currently set C<on_read> callback, or clears it (when
416 the new callback is C<undef>). See the description of C<on_read> in the
417 constructor.
418
419 =cut
420
421 sub on_read {
422 my ($self, $cb) = @_;
423
424 $self->{on_read} = $cb;
425 }
426
427 =item $handle->rbuf
428
429 Returns the read buffer (as a modifiable lvalue).
430
431 You can access the read buffer directly as the C<< ->{rbuf} >> member, if
432 you want.
433
434 NOTE: The read buffer should only be used or modified if the C<on_read>,
435 C<push_read> or C<unshift_read> methods are used. The other read methods
436 automatically manage the read buffer.
437
438 =cut
439
440 sub rbuf : lvalue {
441 $_[0]{rbuf}
442 }
443
444 =item $handle->push_read ($cb)
445
446 =item $handle->unshift_read ($cb)
447
448 Append the given callback to the end of the queue (C<push_read>) or
449 prepend it (C<unshift_read>).
450
451 The callback is called each time some additional read data arrives.
452
453 It must check wether enough data is in the read buffer already.
454
455 If not enough data is available, it must return the empty list or a false
456 value, in which case it will be called repeatedly until enough data is
457 available (or an error condition is detected).
458
459 If enough data was available, then the callback must remove all data it is
460 interested in (which can be none at all) and return a true value. After returning
461 true, it will be removed from the queue.
462
463 =cut
464
465 sub push_read {
466 my ($self, $cb) = @_;
467
468 push @{ $self->{queue} }, $cb;
469 $self->_drain_rbuf;
470 }
471
472 sub unshift_read {
473 my ($self, $cb) = @_;
474
475 push @{ $self->{queue} }, $cb;
476 $self->_drain_rbuf;
477 }
478
479 =item $handle->push_read_chunk ($len, $cb->($self, $data))
480
481 =item $handle->unshift_read_chunk ($len, $cb->($self, $data))
482
483 Append the given callback to the end of the queue (C<push_read_chunk>) or
484 prepend it (C<unshift_read_chunk>).
485
486 The callback will be called only once C<$len> bytes have been read, and
487 these C<$len> bytes will be passed to the callback.
488
489 =cut
490
491 sub _read_chunk($$) {
492 my ($self, $len, $cb) = @_;
493
494 sub {
495 $len <= length $_[0]{rbuf} or return;
496 $cb->($_[0], substr $_[0]{rbuf}, 0, $len, "");
497 1
498 }
499 }
500
501 sub push_read_chunk {
502 $_[0]->push_read (&_read_chunk);
503 }
504
505
506 sub unshift_read_chunk {
507 $_[0]->unshift_read (&_read_chunk);
508 }
509
510 =item $handle->push_read_line ([$eol, ]$cb->($self, $line, $eol))
511
512 =item $handle->unshift_read_line ([$eol, ]$cb->($self, $line, $eol))
513
514 Append the given callback to the end of the queue (C<push_read_line>) or
515 prepend it (C<unshift_read_line>).
516
517 The callback will be called only once a full line (including the end of
518 line marker, C<$eol>) has been read. This line (excluding the end of line
519 marker) will be passed to the callback as second argument (C<$line>), and
520 the end of line marker as the third argument (C<$eol>).
521
522 The end of line marker, C<$eol>, can be either a string, in which case it
523 will be interpreted as a fixed record end marker, or it can be a regex
524 object (e.g. created by C<qr>), in which case it is interpreted as a
525 regular expression.
526
527 The end of line marker argument C<$eol> is optional, if it is missing (NOT
528 undef), then C<qr|\015?\012|> is used (which is good for most internet
529 protocols).
530
531 Partial lines at the end of the stream will never be returned, as they are
532 not marked by the end of line marker.
533
534 =cut
535
536 sub _read_line($$) {
537 my $self = shift;
538 my $cb = pop;
539 my $eol = @_ ? shift : qr|(\015?\012)|;
540 my $pos;
541
542 $eol = quotemeta $eol unless ref $eol;
543 $eol = qr|^(.*?)($eol)|s;
544
545 sub {
546 $_[0]{rbuf} =~ s/$eol// or return;
547
548 $cb->($_[0], $1, $2);
549 1
550 }
551 }
552
553 sub push_read_line {
554 $_[0]->push_read (&_read_line);
555 }
556
557 sub unshift_read_line {
558 $_[0]->unshift_read (&_read_line);
559 }
560
561 =item $handle->stop_read
562
563 =item $handle->start_read
564
565 In rare cases you actually do not want to read anything form the
566 socket. In this case you can call C<stop_read>. Neither C<on_read> no
567 any queued callbacks will be executed then. To start readign again, call
568 C<start_read>.
569
570 =cut
571
572 sub stop_read {
573 my ($self) = @_;
574
575 delete $self->{rw};
576 }
577
578 sub start_read {
579 my ($self) = @_;
580
581 unless ($self->{rw} || $self->{eof}) {
582 Scalar::Util::weaken $self;
583
584 $self->{rw} = AnyEvent->io (fh => $self->{fh}, poll => "r", cb => sub {
585 my $len = sysread $self->{fh}, $self->{rbuf}, $self->{read_size} || 8192, length $self->{rbuf};
586
587 if ($len > 0) {
588 if (defined $self->{rbuf_max}) {
589 if ($self->{rbuf_max} < length $self->{rbuf}) {
590 $! = &Errno::ENOSPC; return $self->error;
591 }
592 }
593
594 } elsif (defined $len) {
595 $self->{eof} = 1;
596 delete $self->{rw};
597
598 } elsif ($! != EAGAIN && $! != EINTR) {
599 return $self->error;
600 }
601
602 $self->_drain_rbuf;
603 });
604 }
605 }
606
607 =back
608
609 =head1 AUTHOR
610
611 Robin Redeker C<< <elmex at ta-sa.org> >>, Marc Lehmann <schmorp@schmorp.de>.
612
613 =cut
614
615 1; # End of AnyEvent::Handle