… | |
… | |
8 | |
8 | |
9 | =head1 DESCRIPTION |
9 | =head1 DESCRIPTION |
10 | |
10 | |
11 | This module is an L<AnyEvent> user, you need to make sure that you use and |
11 | This module is an L<AnyEvent> user, you need to make sure that you use and |
12 | run a supported event loop. |
12 | run a supported event loop. |
|
|
13 | |
|
|
14 | This module implements a simple, stateless and non-blocking HTTP |
|
|
15 | client. It supports GET, POST and other request methods, cookies and more, |
|
|
16 | all on a very low level. It can follow redirects supports proxies and |
|
|
17 | automatically limits the number of connections to the values specified in |
|
|
18 | the RFC. |
|
|
19 | |
|
|
20 | It should generally be a "good client" that is enough for most HTTP |
|
|
21 | tasks. Simple tasks should be simple, but complex tasks should still be |
|
|
22 | possible as the user retains control over request and response headers. |
|
|
23 | |
|
|
24 | The caller is responsible for authentication management, cookies (if |
|
|
25 | the simplistic implementation in this module doesn't suffice), referer |
|
|
26 | and other high-level protocol details for which this module offers only |
|
|
27 | limited support. |
13 | |
28 | |
14 | =head2 METHODS |
29 | =head2 METHODS |
15 | |
30 | |
16 | =over 4 |
31 | =over 4 |
17 | |
32 | |
… | |
… | |
41 | our $PERSISTENT_TIMEOUT = 2; |
56 | our $PERSISTENT_TIMEOUT = 2; |
42 | our $TIMEOUT = 300; |
57 | our $TIMEOUT = 300; |
43 | |
58 | |
44 | # changing these is evil |
59 | # changing these is evil |
45 | our $MAX_PERSISTENT_PER_HOST = 2; |
60 | our $MAX_PERSISTENT_PER_HOST = 2; |
46 | our $MAX_PER_HOST = 4; # not respected yet :( |
61 | our $MAX_PER_HOST = 4; |
47 | |
62 | |
48 | our $PROXY; |
63 | our $PROXY; |
49 | |
64 | |
50 | my %KA_COUNT; # number of open keep-alive connections per host |
65 | my %KA_COUNT; # number of open keep-alive connections per host |
|
|
66 | my %CO_SLOT; # number of open connections, and wait queue, per host |
51 | |
67 | |
52 | =item http_get $url, key => value..., $cb->($data, $headers) |
68 | =item http_get $url, key => value..., $cb->($data, $headers) |
53 | |
69 | |
54 | Executes an HTTP-GET request. See the http_request function for details on |
70 | Executes an HTTP-GET request. See the http_request function for details on |
55 | additional parameters. |
71 | additional parameters. |
… | |
… | |
164 | print Dumper $hdr; |
180 | print Dumper $hdr; |
165 | } |
181 | } |
166 | ; |
182 | ; |
167 | |
183 | |
168 | =cut |
184 | =cut |
|
|
185 | |
|
|
186 | sub _slot_schedule($) { |
|
|
187 | my $host = shift; |
|
|
188 | |
|
|
189 | while ($CO_SLOT{$host}[0] < $MAX_PER_HOST) { |
|
|
190 | if (my $cb = shift @{ $CO_SLOT{$host}[1] }) { |
|
|
191 | # somebody wnats that slot |
|
|
192 | ++$CO_SLOT{$host}[0]; |
|
|
193 | |
|
|
194 | $cb->(AnyEvent::Util::guard { |
|
|
195 | --$CO_SLOT{$host}[0]; |
|
|
196 | _slot_schedule $host; |
|
|
197 | }); |
|
|
198 | } else { |
|
|
199 | # nobody wants the slot, maybe we can forget about it |
|
|
200 | delete $CO_SLOT{$host} unless $CO_SLOT{$host}[0]; |
|
|
201 | warn "$host deleted" unless $CO_SLOT{$host}[0];#d# |
|
|
202 | last; |
|
|
203 | } |
|
|
204 | } |
|
|
205 | } |
|
|
206 | |
|
|
207 | # wait for a free slot on host, call callback |
|
|
208 | sub _get_slot($$) { |
|
|
209 | push @{ $CO_SLOT{$_[0]}[1] }, $_[1]; |
|
|
210 | |
|
|
211 | _slot_schedule $_[0]; |
|
|
212 | } |
169 | |
213 | |
170 | sub http_request($$$;@) { |
214 | sub http_request($$$;@) { |
171 | my $cb = pop; |
215 | my $cb = pop; |
172 | my ($method, $url, %arg) = @_; |
216 | my ($method, $url, %arg) = @_; |
173 | |
217 | |
… | |
… | |
245 | $hdr{host} = $uhost; |
289 | $hdr{host} = $uhost; |
246 | } |
290 | } |
247 | |
291 | |
248 | $hdr{"content-length"} = length $arg{body}; |
292 | $hdr{"content-length"} = length $arg{body}; |
249 | |
293 | |
250 | my %state; |
294 | my %state = (connect_guard => 1); |
251 | |
295 | |
|
|
296 | _get_slot $uhost, sub { |
|
|
297 | $state{slot_guard} = shift; |
|
|
298 | |
|
|
299 | return unless $state{connect_guard}; |
|
|
300 | |
252 | $state{connect_guard} = AnyEvent::Socket::tcp_connect $rhost, $rport, sub { |
301 | $state{connect_guard} = AnyEvent::Socket::tcp_connect $rhost, $rport, sub { |
253 | $state{fh} = shift |
302 | $state{fh} = shift |
254 | or return $cb->(undef, { Status => 599, Reason => "$!" }); |
303 | or return $cb->(undef, { Status => 599, Reason => "$!" }); |
255 | |
304 | |
256 | delete $state{connect_guard}; # reduce memory usage, save a tree |
305 | delete $state{connect_guard}; # reduce memory usage, save a tree |
257 | |
306 | |
258 | # get handle |
307 | # get handle |
259 | $state{handle} = new AnyEvent::Handle |
308 | $state{handle} = new AnyEvent::Handle |
260 | fh => $state{fh}, |
309 | fh => $state{fh}, |
261 | ($scheme eq "https" ? (tls => "connect") : ()); |
310 | ($scheme eq "https" ? (tls => "connect") : ()); |
262 | |
311 | |
263 | # limit the number of persistent connections |
312 | # limit the number of persistent connections |
264 | if ($KA_COUNT{$_[1]} < $MAX_PERSISTENT_PER_HOST) { |
313 | if ($KA_COUNT{$_[1]} < $MAX_PERSISTENT_PER_HOST) { |
265 | ++$KA_COUNT{$_[1]}; |
314 | ++$KA_COUNT{$_[1]}; |
266 | $state{handle}{ka_count_guard} = AnyEvent::Util::guard { --$KA_COUNT{$_[1]} }; |
315 | $state{handle}{ka_count_guard} = AnyEvent::Util::guard { --$KA_COUNT{$_[1]} }; |
267 | $hdr{connection} = "keep-alive"; |
316 | $hdr{connection} = "keep-alive"; |
268 | delete $hdr{connection}; # keep-alive not yet supported |
317 | delete $hdr{connection}; # keep-alive not yet supported |
269 | } else { |
318 | } else { |
270 | delete $hdr{connection}; |
319 | delete $hdr{connection}; |
271 | } |
320 | } |
272 | |
321 | |
273 | # (re-)configure handle |
322 | # (re-)configure handle |
274 | $state{handle}->timeout ($timeout); |
323 | $state{handle}->timeout ($timeout); |
275 | $state{handle}->on_error (sub { |
324 | $state{handle}->on_error (sub { |
276 | %state = (); |
325 | %state = (); |
277 | $cb->(undef, { Status => 599, Reason => "$!" }); |
326 | $cb->(undef, { Status => 599, Reason => "$!" }); |
278 | }); |
327 | }); |
279 | $state{handle}->on_eof (sub { |
328 | $state{handle}->on_eof (sub { |
280 | %state = (); |
329 | %state = (); |
281 | $cb->(undef, { Status => 599, Reason => "unexpected end-of-file" }); |
330 | $cb->(undef, { Status => 599, Reason => "unexpected end-of-file" }); |
282 | }); |
331 | }); |
283 | |
332 | |
284 | # send request |
333 | # send request |
285 | $state{handle}->push_write ( |
334 | $state{handle}->push_write ( |
286 | "$method $rpath HTTP/1.0\015\012" |
335 | "$method $rpath HTTP/1.0\015\012" |
287 | . (join "", map "$_: $hdr{$_}\015\012", keys %hdr) |
336 | . (join "", map "$_: $hdr{$_}\015\012", keys %hdr) |
288 | . "\015\012" |
337 | . "\015\012" |
289 | . (delete $arg{body}) |
338 | . (delete $arg{body}) |
290 | ); |
|
|
291 | |
|
|
292 | %hdr = (); # reduce memory usage, save a kitten |
|
|
293 | |
|
|
294 | # status line |
|
|
295 | $state{handle}->push_read (line => qr/\015?\012/, sub { |
|
|
296 | $_[1] =~ /^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) \s+ ([^\015\012]+)/ix |
|
|
297 | or return (%state = (), $cb->(undef, { Status => 599, Reason => "invalid server response ($_[1])" })); |
|
|
298 | |
|
|
299 | my %hdr = ( # response headers |
|
|
300 | HTTPVersion => "\x00$1", |
|
|
301 | Status => "\x00$2", |
|
|
302 | Reason => "\x00$3", |
|
|
303 | ); |
339 | ); |
304 | |
340 | |
|
|
341 | %hdr = (); # reduce memory usage, save a kitten |
|
|
342 | |
|
|
343 | # status line |
|
|
344 | $state{handle}->push_read (line => qr/\015?\012/, sub { |
|
|
345 | $_[1] =~ /^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) \s+ ([^\015\012]+)/ix |
|
|
346 | or return (%state = (), $cb->(undef, { Status => 599, Reason => "invalid server response ($_[1])" })); |
|
|
347 | |
|
|
348 | my %hdr = ( # response headers |
|
|
349 | HTTPVersion => "\x00$1", |
|
|
350 | Status => "\x00$2", |
|
|
351 | Reason => "\x00$3", |
|
|
352 | ); |
|
|
353 | |
305 | # headers, could be optimized a bit |
354 | # headers, could be optimized a bit |
306 | $state{handle}->unshift_read (line => qr/\015?\012\015?\012/, sub { |
355 | $state{handle}->unshift_read (line => qr/\015?\012\015?\012/, sub { |
307 | for ("$_[1]\012") { |
356 | for ("$_[1]\012") { |
308 | # we support spaces in field names, as lotus domino |
357 | # we support spaces in field names, as lotus domino |
309 | # creates them. |
358 | # creates them. |
310 | $hdr{lc $1} .= "\x00$2" |
359 | $hdr{lc $1} .= "\x00$2" |
311 | while /\G |
360 | while /\G |
312 | ([^:\000-\037]+): |
361 | ([^:\000-\037]+): |
313 | [\011\040]* |
362 | [\011\040]* |
314 | ((?: [^\015\012]+ | \015?\012[\011\040] )*) |
363 | ((?: [^\015\012]+ | \015?\012[\011\040] )*) |
315 | \015?\012 |
364 | \015?\012 |
316 | /gxc; |
365 | /gxc; |
317 | |
366 | |
318 | /\G$/ |
367 | /\G$/ |
319 | or return (%state = (), $cb->(undef, { Status => 599, Reason => "garbled response headers" })); |
368 | or return (%state = (), $cb->(undef, { Status => 599, Reason => "garbled response headers" })); |
320 | } |
369 | } |
321 | |
370 | |
322 | substr $_, 0, 1, "" |
371 | substr $_, 0, 1, "" |
323 | for values %hdr; |
372 | for values %hdr; |
324 | |
373 | |
325 | my $finish = sub { |
374 | my $finish = sub { |
326 | %state = (); |
375 | %state = (); |
327 | |
376 | |
328 | # set-cookie processing |
377 | # set-cookie processing |
329 | if ($arg{cookie_jar} && exists $hdr{"set-cookie"}) { |
378 | if ($arg{cookie_jar} && exists $hdr{"set-cookie"}) { |
330 | for (split /\x00/, $hdr{"set-cookie"}) { |
379 | for (split /\x00/, $hdr{"set-cookie"}) { |
331 | my ($cookie, @arg) = split /;\s*/; |
380 | my ($cookie, @arg) = split /;\s*/; |
332 | my ($name, $value) = split /=/, $cookie, 2; |
381 | my ($name, $value) = split /=/, $cookie, 2; |
333 | my %kv = (value => $value, map { split /=/, $_, 2 } @arg); |
382 | my %kv = (value => $value, map { split /=/, $_, 2 } @arg); |
334 | |
383 | |
335 | my $cdom = (delete $kv{domain}) || $uhost; |
384 | my $cdom = (delete $kv{domain}) || $uhost; |
336 | my $cpath = (delete $kv{path}) || "/"; |
385 | my $cpath = (delete $kv{path}) || "/"; |
337 | |
386 | |
338 | $cdom =~ s/^.?/./; # make sure it starts with a "." |
387 | $cdom =~ s/^.?/./; # make sure it starts with a "." |
339 | |
388 | |
|
|
389 | next if $cdom =~ /\.$/; |
|
|
390 | |
|
|
391 | # this is not rfc-like and not netscape-like. go figure. |
340 | my $ndots = $cdom =~ y/.//; |
392 | my $ndots = $cdom =~ y/.//; |
341 | next if $ndots < ($cdom =~ /[^.]{3}$/ ? 2 : 3); |
393 | next if $ndots < ($cdom =~ /\.[^.][^.]\.[^.][^.]$/ ? 3 : 2); |
342 | |
394 | |
343 | # store it |
395 | # store it |
344 | $arg{cookie_jar}{version} = 1; |
396 | $arg{cookie_jar}{version} = 1; |
345 | $arg{cookie_jar}{$cdom}{$cpath}{$name} = \%kv; |
397 | $arg{cookie_jar}{$cdom}{$cpath}{$name} = \%kv; |
|
|
398 | } |
|
|
399 | } |
|
|
400 | |
|
|
401 | if ($_[1]{Status} =~ /^x30[12]$/ && $recurse) { |
|
|
402 | # microsoft and other assholes don't give a shit for following standards, |
|
|
403 | # try to support a common form of broken Location header. |
|
|
404 | $_[1]{location} =~ s%^/%$scheme://$uhost:$uport/%; |
|
|
405 | |
|
|
406 | http_request ($method, $_[1]{location}, %arg, recurse => $recurse - 1, $cb); |
|
|
407 | } else { |
|
|
408 | $cb->($_[0], $_[1]); |
|
|
409 | } |
|
|
410 | }; |
|
|
411 | |
|
|
412 | if ($hdr{Status} =~ /^(?:1..|204|304)$/ or $method eq "HEAD") { |
|
|
413 | $finish->(undef, \%hdr); |
|
|
414 | } else { |
|
|
415 | if (exists $hdr{"content-length"}) { |
|
|
416 | $_[0]->unshift_read (chunk => $hdr{"content-length"}, sub { |
|
|
417 | # could cache persistent connection now |
|
|
418 | if ($hdr{connection} =~ /\bkeep-alive\b/i) { |
|
|
419 | # but we don't, due to misdesigns, this is annoyingly complex |
|
|
420 | }; |
|
|
421 | |
|
|
422 | $finish->($_[1], \%hdr); |
|
|
423 | }); |
|
|
424 | } else { |
|
|
425 | # too bad, need to read until we get an error or EOF, |
|
|
426 | # no way to detect winged data. |
|
|
427 | $_[0]->on_error (sub { |
|
|
428 | $finish->($_[0]{rbuf}, \%hdr); |
|
|
429 | }); |
|
|
430 | $_[0]->on_eof (undef); |
|
|
431 | $_[0]->on_read (sub { }); |
346 | } |
432 | } |
347 | } |
433 | } |
348 | |
|
|
349 | if ($_[1]{Status} =~ /^x30[12]$/ && $recurse) { |
|
|
350 | # microsoft and other assholes don't give a shit for following standards, |
|
|
351 | # try to support a common form of broken Location header. |
|
|
352 | $_[1]{location} =~ s%^/%$scheme://$uhost:$uport/%; |
|
|
353 | |
|
|
354 | http_request ($method, $_[1]{location}, %arg, recurse => $recurse - 1, $cb); |
|
|
355 | } else { |
|
|
356 | $cb->($_[0], $_[1]); |
|
|
357 | } |
|
|
358 | }; |
434 | }); |
359 | |
|
|
360 | if ($hdr{Status} =~ /^(?:1..|204|304)$/ or $method eq "HEAD") { |
|
|
361 | $finish->(undef, \%hdr); |
|
|
362 | } else { |
|
|
363 | if (exists $hdr{"content-length"}) { |
|
|
364 | $_[0]->unshift_read (chunk => $hdr{"content-length"}, sub { |
|
|
365 | # could cache persistent connection now |
|
|
366 | if ($hdr{connection} =~ /\bkeep-alive\b/i) { |
|
|
367 | # but we don't, due to misdesigns, this is annoyingly complex |
|
|
368 | }; |
|
|
369 | |
|
|
370 | $finish->($_[1], \%hdr); |
|
|
371 | }); |
|
|
372 | } else { |
|
|
373 | # too bad, need to read until we get an error or EOF, |
|
|
374 | # no way to detect winged data. |
|
|
375 | $_[0]->on_error (sub { |
|
|
376 | $finish->($_[0]{rbuf}, \%hdr); |
|
|
377 | }); |
|
|
378 | $_[0]->on_eof (undef); |
|
|
379 | $_[0]->on_read (sub { }); |
|
|
380 | } |
|
|
381 | } |
|
|
382 | }); |
435 | }); |
|
|
436 | }, sub { |
|
|
437 | $timeout |
383 | }); |
438 | }; |
384 | }, sub { |
|
|
385 | $timeout |
|
|
386 | }; |
439 | }; |
387 | |
440 | |
388 | defined wantarray && AnyEvent::Util::guard { %state = () } |
441 | defined wantarray && AnyEvent::Util::guard { %state = () } |
389 | } |
442 | } |
390 | |
443 | |