… | |
… | |
39 | package AnyEvent::HTTP; |
39 | package AnyEvent::HTTP; |
40 | |
40 | |
41 | use strict; |
41 | use strict; |
42 | no warnings; |
42 | no warnings; |
43 | |
43 | |
44 | use Carp; |
44 | use Errno (); |
45 | |
45 | |
46 | use AnyEvent 4.452 (); |
46 | use AnyEvent 5.0 (); |
47 | use AnyEvent::Util (); |
47 | use AnyEvent::Util (); |
48 | use AnyEvent::Socket (); |
48 | use AnyEvent::Socket (); |
49 | use AnyEvent::Handle (); |
49 | use AnyEvent::Handle (); |
50 | |
50 | |
51 | use base Exporter::; |
51 | use base Exporter::; |
52 | |
52 | |
53 | our $VERSION = '1.12'; |
53 | our $VERSION = '1.44'; |
54 | |
54 | |
55 | our @EXPORT = qw(http_get http_post http_head http_request); |
55 | our @EXPORT = qw(http_get http_post http_head http_request); |
56 | |
56 | |
57 | our $USERAGENT = "Mozilla/5.0 (compatible; U; AnyEvent-HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)"; |
57 | our $USERAGENT = "Mozilla/5.0 (compatible; U; AnyEvent-HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)"; |
58 | our $MAX_RECURSE = 10; |
58 | our $MAX_RECURSE = 10; |
59 | our $MAX_PERSISTENT = 8; |
59 | our $MAX_PERSISTENT = 8; |
60 | our $PERSISTENT_TIMEOUT = 2; |
60 | our $PERSISTENT_TIMEOUT = 2; |
61 | our $TIMEOUT = 300; |
61 | our $TIMEOUT = 300; |
62 | |
62 | |
63 | # changing these is evil |
63 | # changing these is evil |
64 | our $MAX_PERSISTENT_PER_HOST = 2; |
64 | our $MAX_PERSISTENT_PER_HOST = 0; |
65 | our $MAX_PER_HOST = 4; |
65 | our $MAX_PER_HOST = 4; |
66 | |
66 | |
67 | our $PROXY; |
67 | our $PROXY; |
68 | our $ACTIVE = 0; |
68 | our $ACTIVE = 0; |
69 | |
69 | |
… | |
… | |
94 | When called in void context, nothing is returned. In other contexts, |
94 | When called in void context, nothing is returned. In other contexts, |
95 | C<http_request> returns a "cancellation guard" - you have to keep the |
95 | C<http_request> returns a "cancellation guard" - you have to keep the |
96 | object at least alive until the callback get called. If the object gets |
96 | object at least alive until the callback get called. If the object gets |
97 | destroyed before the callbakc is called, the request will be cancelled. |
97 | destroyed before the callbakc is called, the request will be cancelled. |
98 | |
98 | |
99 | The callback will be called with the response data as first argument |
99 | The callback will be called with the response body data as first argument |
100 | (or C<undef> if it wasn't available due to errors), and a hash-ref with |
100 | (or C<undef> if an error occured), and a hash-ref with response headers as |
101 | response headers as second argument. |
101 | second argument. |
102 | |
102 | |
103 | All the headers in that hash are lowercased. In addition to the response |
103 | All the headers in that hash are lowercased. In addition to the response |
104 | headers, the "pseudo-headers" C<HTTPVersion>, C<Status> and C<Reason> |
104 | headers, the "pseudo-headers" C<HTTPVersion>, C<Status> and C<Reason> |
105 | contain the three parts of the HTTP Status-Line of the same name. The |
105 | contain the three parts of the HTTP Status-Line of the same name. The |
106 | pseudo-header C<URL> contains the original URL (which can differ from the |
106 | pseudo-header C<URL> contains the original URL (which can differ from the |
… | |
… | |
108 | |
108 | |
109 | If the server sends a header multiple times, then their contents will be |
109 | If the server sends a header multiple times, then their contents will be |
110 | joined together with a comma (C<,>), as per the HTTP spec. |
110 | joined together with a comma (C<,>), as per the HTTP spec. |
111 | |
111 | |
112 | If an internal error occurs, such as not being able to resolve a hostname, |
112 | If an internal error occurs, such as not being able to resolve a hostname, |
113 | then C<$data> will be C<undef>, C<< $headers->{Status} >> will be C<599> |
113 | then C<$data> will be C<undef>, C<< $headers->{Status} >> will be C<59x> |
114 | and the C<Reason> pseudo-header will contain an error message. |
114 | (usually C<599>) and the C<Reason> pseudo-header will contain an error |
|
|
115 | message. |
115 | |
116 | |
116 | A typical callback might look like this: |
117 | A typical callback might look like this: |
117 | |
118 | |
118 | sub { |
119 | sub { |
119 | my ($body, $hdr) = @_; |
120 | my ($body, $hdr) = @_; |
… | |
… | |
137 | |
138 | |
138 | =item headers => hashref |
139 | =item headers => hashref |
139 | |
140 | |
140 | The request headers to use. Currently, C<http_request> may provide its |
141 | The request headers to use. Currently, C<http_request> may provide its |
141 | own C<Host:>, C<Content-Length:>, C<Connection:> and C<Cookie:> headers |
142 | own C<Host:>, C<Content-Length:>, C<Connection:> and C<Cookie:> headers |
142 | and will provide defaults for C<User-Agent:> and C<Referer:>. |
143 | and will provide defaults for C<User-Agent:> and C<Referer:> (this can be |
|
|
144 | suppressed by using C<undef> for these headers in which case they won't be |
|
|
145 | sent at all). |
143 | |
146 | |
144 | =item timeout => $seconds |
147 | =item timeout => $seconds |
145 | |
148 | |
146 | The time-out to use for various stages - each connect attempt will reset |
149 | The time-out to use for various stages - each connect attempt will reset |
147 | the timeout, as will read or write activity. Default timeout is 5 minutes. |
150 | the timeout, as will read or write activity, i.e. this is not an overall |
|
|
151 | timeout. |
|
|
152 | |
|
|
153 | Default timeout is 5 minutes. |
148 | |
154 | |
149 | =item proxy => [$host, $port[, $scheme]] or undef |
155 | =item proxy => [$host, $port[, $scheme]] or undef |
150 | |
156 | |
151 | Use the given http proxy for all requests. If not specified, then the |
157 | Use the given http proxy for all requests. If not specified, then the |
152 | default proxy (as specified by C<$ENV{http_proxy}>) is used. |
158 | default proxy (as specified by C<$ENV{http_proxy}>) is used. |
153 | |
159 | |
154 | C<$scheme> must be either missing or C<http> for HTTP, or C<https> for |
160 | C<$scheme> must be either missing, C<http> for HTTP or C<https> for |
155 | HTTPS. |
161 | HTTPS. |
156 | |
162 | |
157 | =item body => $string |
163 | =item body => $string |
158 | |
164 | |
159 | The request body, usually empty. Will be-sent as-is (future versions of |
165 | The request body, usually empty. Will be-sent as-is (future versions of |
… | |
… | |
184 | verification, highest compatibility) and high-security (CA and common-name |
190 | verification, highest compatibility) and high-security (CA and common-name |
185 | verification) TLS context. |
191 | verification) TLS context. |
186 | |
192 | |
187 | The default for this option is C<low>, which could be interpreted as "give |
193 | The default for this option is C<low>, which could be interpreted as "give |
188 | me the page, no matter what". |
194 | me the page, no matter what". |
|
|
195 | |
|
|
196 | =item on_prepare => $callback->($fh) |
|
|
197 | |
|
|
198 | In rare cases you need to "tune" the socket before it is used to |
|
|
199 | connect (for exmaple, to bind it on a given IP address). This parameter |
|
|
200 | overrides the prepare callback passed to C<AnyEvent::Socket::tcp_connect> |
|
|
201 | and behaves exactly the same way (e.g. it has to provide a |
|
|
202 | timeout). See the description for the C<$prepare_cb> argument of |
|
|
203 | C<AnyEvent::Socket::tcp_connect> for details. |
|
|
204 | |
|
|
205 | =item on_header => $callback->($headers) |
|
|
206 | |
|
|
207 | When specified, this callback will be called with the header hash as soon |
|
|
208 | as headers have been successfully received from the remote server (not on |
|
|
209 | locally-generated errors). |
|
|
210 | |
|
|
211 | It has to return either true (in which case AnyEvent::HTTP will continue), |
|
|
212 | or false, in which case AnyEvent::HTTP will cancel the download (and call |
|
|
213 | the finish callback with an error code of C<598>). |
|
|
214 | |
|
|
215 | This callback is useful, among other things, to quickly reject unwanted |
|
|
216 | content, which, if it is supposed to be rare, can be faster than first |
|
|
217 | doing a C<HEAD> request. |
|
|
218 | |
|
|
219 | Example: cancel the request unless the content-type is "text/html". |
|
|
220 | |
|
|
221 | on_header => sub { |
|
|
222 | $_[0]{"content-type"} =~ /^text\/html\s*(?:;|$)/ |
|
|
223 | }, |
|
|
224 | |
|
|
225 | =item on_body => $callback->($partial_body, $headers) |
|
|
226 | |
|
|
227 | When specified, all body data will be passed to this callback instead of |
|
|
228 | to the completion callback. The completion callback will get the empty |
|
|
229 | string instead of the body data. |
|
|
230 | |
|
|
231 | It has to return either true (in which case AnyEvent::HTTP will continue), |
|
|
232 | or false, in which case AnyEvent::HTTP will cancel the download (and call |
|
|
233 | the completion callback with an error code of C<598>). |
|
|
234 | |
|
|
235 | This callback is useful when the data is too large to be held in memory |
|
|
236 | (so the callback writes it to a file) or when only some information should |
|
|
237 | be extracted, or when the body should be processed incrementally. |
|
|
238 | |
|
|
239 | It is usually preferred over doing your own body handling via |
|
|
240 | C<want_body_handle>, but in case of streaming APIs, where HTTP is |
|
|
241 | only used to create a connection, C<want_body_handle> is the better |
|
|
242 | alternative, as it allows you to install your own event handler, reducing |
|
|
243 | resource usage. |
|
|
244 | |
|
|
245 | =item want_body_handle => $enable |
|
|
246 | |
|
|
247 | When enabled (default is disabled), the behaviour of AnyEvent::HTTP |
|
|
248 | changes considerably: after parsing the headers, and instead of |
|
|
249 | downloading the body (if any), the completion callback will be |
|
|
250 | called. Instead of the C<$body> argument containing the body data, the |
|
|
251 | callback will receive the L<AnyEvent::Handle> object associated with the |
|
|
252 | connection. In error cases, C<undef> will be passed. When there is no body |
|
|
253 | (e.g. status C<304>), the empty string will be passed. |
|
|
254 | |
|
|
255 | The handle object might or might not be in TLS mode, might be connected to |
|
|
256 | a proxy, be a persistent connection etc., and configured in unspecified |
|
|
257 | ways. The user is responsible for this handle (it will not be used by this |
|
|
258 | module anymore). |
|
|
259 | |
|
|
260 | This is useful with some push-type services, where, after the initial |
|
|
261 | headers, an interactive protocol is used (typical example would be the |
|
|
262 | push-style twitter API which starts a JSON/XML stream). |
|
|
263 | |
|
|
264 | If you think you need this, first have a look at C<on_body>, to see if |
|
|
265 | that doesn't solve your problem in a better way. |
189 | |
266 | |
190 | =back |
267 | =back |
191 | |
268 | |
192 | Example: make a simple HTTP GET request for http://www.nethype.de/ |
269 | Example: make a simple HTTP GET request for http://www.nethype.de/ |
193 | |
270 | |
… | |
… | |
249 | push @{ $CO_SLOT{$_[0]}[1] }, $_[1]; |
326 | push @{ $CO_SLOT{$_[0]}[1] }, $_[1]; |
250 | |
327 | |
251 | _slot_schedule $_[0]; |
328 | _slot_schedule $_[0]; |
252 | } |
329 | } |
253 | |
330 | |
254 | our $qr_nl = qr<\015?\012>; |
331 | our $qr_nl = qr{\015?\012}; |
255 | our $qr_nlnl = qr<\015?\012\015?\012>; |
332 | our $qr_nlnl = qr{(?<![^\012])\015?\012}; |
256 | |
333 | |
257 | our $TLS_CTX_LOW = { cache => 1, dh => undef, sslv2 => 1 }; |
334 | our $TLS_CTX_LOW = { cache => 1, sslv2 => 1 }; |
258 | our $TLS_CTX_HIGH = { cache => 1, verify => 1, verify_cn => "https", dh => "skip4096" }; |
335 | our $TLS_CTX_HIGH = { cache => 1, verify => 1, verify_peername => "https" }; |
259 | |
336 | |
260 | sub http_request($$@) { |
337 | sub http_request($$@) { |
261 | my $cb = pop; |
338 | my $cb = pop; |
262 | my ($method, $url, %arg) = @_; |
339 | my ($method, $url, %arg) = @_; |
263 | |
340 | |
… | |
… | |
280 | if $recurse < 0; |
357 | if $recurse < 0; |
281 | |
358 | |
282 | my $proxy = $arg{proxy} || $PROXY; |
359 | my $proxy = $arg{proxy} || $PROXY; |
283 | my $timeout = $arg{timeout} || $TIMEOUT; |
360 | my $timeout = $arg{timeout} || $TIMEOUT; |
284 | |
361 | |
285 | $hdr{"user-agent"} ||= $USERAGENT; |
|
|
286 | |
|
|
287 | my ($uscheme, $uauthority, $upath, $query, $fragment) = |
362 | my ($uscheme, $uauthority, $upath, $query, $fragment) = |
288 | $url =~ m|(?:([^:/?#]+):)?(?://([^/?#]*))?([^?#]*)(?:\?([^#]*))?(?:#(.*))?|; |
363 | $url =~ m|(?:([^:/?#]+):)?(?://([^/?#]*))?([^?#]*)(?:\?([^#]*))?(?:#(.*))?|; |
289 | |
364 | |
290 | $uscheme = lc $uscheme; |
365 | $uscheme = lc $uscheme; |
291 | |
366 | |
292 | my $uport = $uscheme eq "http" ? 80 |
367 | my $uport = $uscheme eq "http" ? 80 |
293 | : $uscheme eq "https" ? 443 |
368 | : $uscheme eq "https" ? 443 |
294 | : return $cb->(undef, { Status => 599, Reason => "Only http and https URL schemes supported (not '$uscheme')", URL => $url }); |
369 | : return $cb->(undef, { Status => 599, Reason => "Only http and https URL schemes supported", URL => $url }); |
295 | |
|
|
296 | $hdr{referer} ||= "$uscheme://$uauthority$upath"; # leave out fragment and query string, just a heuristic |
|
|
297 | |
370 | |
298 | $uauthority =~ /^(?: .*\@ )? ([^\@:]+) (?: : (\d+) )?$/x |
371 | $uauthority =~ /^(?: .*\@ )? ([^\@:]+) (?: : (\d+) )?$/x |
299 | or return $cb->(undef, { Status => 599, Reason => "Unparsable URL", URL => $url }); |
372 | or return $cb->(undef, { Status => 599, Reason => "Unparsable URL", URL => $url }); |
300 | |
373 | |
301 | my $uhost = $1; |
374 | my $uhost = $1; |
302 | $uport = $2 if defined $2; |
375 | $uport = $2 if defined $2; |
|
|
376 | |
|
|
377 | $hdr{host} = defined $2 ? "$uhost:$2" : "$uhost" |
|
|
378 | unless exists $hdr{host}; |
303 | |
379 | |
304 | $uhost =~ s/^\[(.*)\]$/$1/; |
380 | $uhost =~ s/^\[(.*)\]$/$1/; |
305 | $upath .= "?$query" if length $query; |
381 | $upath .= "?$query" if length $query; |
306 | |
382 | |
307 | $upath =~ s%^/?%/%; |
383 | $upath =~ s%^/?%/%; |
… | |
… | |
340 | my ($rhost, $rport, $rscheme, $rpath); # request host, port, path |
416 | my ($rhost, $rport, $rscheme, $rpath); # request host, port, path |
341 | |
417 | |
342 | if ($proxy) { |
418 | if ($proxy) { |
343 | ($rpath, $rhost, $rport, $rscheme) = ($url, @$proxy); |
419 | ($rpath, $rhost, $rport, $rscheme) = ($url, @$proxy); |
344 | |
420 | |
|
|
421 | $rscheme = "http" unless defined $rscheme; |
|
|
422 | |
345 | # don't support https requests over https-proxy transport, |
423 | # don't support https requests over https-proxy transport, |
346 | # can't be done with tls as spec'ed, unless you double-encrypt. |
424 | # can't be done with tls as spec'ed, unless you double-encrypt. |
347 | $rscheme = "http" if $uscheme eq "https" && $rscheme eq "https"; |
425 | $rscheme = "http" if $uscheme eq "https" && $rscheme eq "https"; |
348 | } else { |
426 | } else { |
349 | ($rhost, $rport, $rscheme, $rpath) = ($uhost, $uport, $uscheme, $upath); |
427 | ($rhost, $rport, $rscheme, $rpath) = ($uhost, $uport, $uscheme, $upath); |
350 | } |
428 | } |
351 | |
429 | |
352 | $hdr{host} = $uhost; |
430 | # leave out fragment and query string, just a heuristic |
|
|
431 | $hdr{referer} ||= "$uscheme://$uauthority$upath" unless exists $hdr{referer}; |
|
|
432 | $hdr{"user-agent"} ||= $USERAGENT unless exists $hdr{"user-agent"}; |
|
|
433 | |
353 | $hdr{"content-length"} = length $arg{body}; |
434 | $hdr{"content-length"} = length $arg{body} |
|
|
435 | if length $arg{body} || $method ne "GET"; |
354 | |
436 | |
355 | my %state = (connect_guard => 1); |
437 | my %state = (connect_guard => 1); |
356 | |
438 | |
357 | _get_slot $uhost, sub { |
439 | _get_slot $uhost, sub { |
358 | $state{slot_guard} = shift; |
440 | $state{slot_guard} = shift; |
359 | |
441 | |
360 | return unless $state{connect_guard}; |
442 | return unless $state{connect_guard}; |
361 | |
443 | |
362 | $state{connect_guard} = AnyEvent::Socket::tcp_connect $rhost, $rport, sub { |
444 | $state{connect_guard} = AnyEvent::Socket::tcp_connect $rhost, $rport, sub { |
363 | $state{fh} = shift |
445 | $state{fh} = shift |
|
|
446 | or do { |
|
|
447 | my $err = "$!"; |
|
|
448 | %state = (); |
364 | or return (%state = (), $cb->(undef, { Status => 599, Reason => "$!", URL => $url })); |
449 | return $cb->(undef, { Status => 599, Reason => $err, URL => $url }); |
|
|
450 | }; |
|
|
451 | |
365 | pop; # free memory, save a tree |
452 | pop; # free memory, save a tree |
366 | |
453 | |
367 | return unless delete $state{connect_guard}; |
454 | return unless delete $state{connect_guard}; |
368 | |
455 | |
369 | # get handle |
456 | # get handle |
… | |
… | |
402 | $state{handle}->starttls ("connect") if $uscheme eq "https" && !exists $state{handle}{tls}; |
489 | $state{handle}->starttls ("connect") if $uscheme eq "https" && !exists $state{handle}{tls}; |
403 | |
490 | |
404 | # send request |
491 | # send request |
405 | $state{handle}->push_write ( |
492 | $state{handle}->push_write ( |
406 | "$method $rpath HTTP/1.0\015\012" |
493 | "$method $rpath HTTP/1.0\015\012" |
407 | . (join "", map "\u$_: $hdr{$_}\015\012", keys %hdr) |
494 | . (join "", map "\u$_: $hdr{$_}\015\012", grep defined $hdr{$_}, keys %hdr) |
408 | . "\015\012" |
495 | . "\015\012" |
409 | . (delete $arg{body}) |
496 | . (delete $arg{body}) |
410 | ); |
497 | ); |
|
|
498 | |
|
|
499 | # return if error occured during push_write() |
|
|
500 | return unless %state; |
411 | |
501 | |
412 | %hdr = (); # reduce memory usage, save a kitten |
502 | %hdr = (); # reduce memory usage, save a kitten |
413 | |
503 | |
414 | # status line |
504 | # status line |
415 | $state{handle}->push_read (line => $qr_nl, sub { |
505 | $state{handle}->push_read (line => $qr_nl, sub { |
… | |
… | |
423 | URL => ",$url" |
513 | URL => ",$url" |
424 | ); |
514 | ); |
425 | |
515 | |
426 | # headers, could be optimized a bit |
516 | # headers, could be optimized a bit |
427 | $state{handle}->unshift_read (line => $qr_nlnl, sub { |
517 | $state{handle}->unshift_read (line => $qr_nlnl, sub { |
428 | for ("$_[1]\012") { |
518 | for ("$_[1]") { |
429 | y/\015//d; # weed out any \015, as they show up in the weirdest of places. |
519 | y/\015//d; # weed out any \015, as they show up in the weirdest of places. |
430 | |
520 | |
431 | # things seen, not parsed: |
521 | # things seen, not parsed: |
432 | # p3pP="NON CUR OTPi OUR NOR UNI" |
522 | # p3pP="NON CUR OTPi OUR NOR UNI" |
433 | |
523 | |
434 | $hdr{lc $1} .= ",$2" |
524 | $hdr{lc $1} .= ",$2" |
435 | while /\G |
525 | while /\G |
436 | ([^:\000-\037]+): |
526 | ([^:\000-\037]*): |
437 | [\011\040]* |
527 | [\011\040]* |
438 | ((?: [^\012]+ | \012[\011\040] )*) |
528 | ((?: [^\012]+ | \012[\011\040] )*) |
439 | \012 |
529 | \012 |
440 | /gxc; |
530 | /gxc; |
441 | |
531 | |
… | |
… | |
444 | } |
534 | } |
445 | |
535 | |
446 | substr $_, 0, 1, "" |
536 | substr $_, 0, 1, "" |
447 | for values %hdr; |
537 | for values %hdr; |
448 | |
538 | |
|
|
539 | # redirect handling |
|
|
540 | # microsoft and other shitheads don't give a shit for following standards, |
|
|
541 | # try to support some common forms of broken Location headers. |
|
|
542 | if ($hdr{location} !~ /^(?: $ | [^:\/?\#]+ : )/x) { |
|
|
543 | $hdr{location} =~ s/^\.\/+//; |
|
|
544 | |
|
|
545 | my $url = "$rscheme://$uhost:$uport"; |
|
|
546 | |
|
|
547 | unless ($hdr{location} =~ s/^\///) { |
|
|
548 | $url .= $upath; |
|
|
549 | $url =~ s/\/[^\/]*$//; |
|
|
550 | } |
|
|
551 | |
|
|
552 | $hdr{location} = "$url/$hdr{location}"; |
|
|
553 | } |
|
|
554 | |
|
|
555 | my $redirect; |
|
|
556 | |
|
|
557 | if ($recurse) { |
|
|
558 | if ($hdr{Status} =~ /^30[12]$/ && $method ne "POST") { |
|
|
559 | # apparently, mozilla et al. just change POST to GET here |
|
|
560 | # more research is needed before we do the same |
|
|
561 | $redirect = 1; |
|
|
562 | } elsif ($hdr{Status} == 303) { |
|
|
563 | # even http/1.1 is unclear on how to mutate the method |
|
|
564 | $method = "GET" unless $method eq "HEAD"; |
|
|
565 | $redirect = 1; |
|
|
566 | } elsif ($hdr{Status} == 307 && $method =~ /^(?:GET|HEAD)$/) { |
|
|
567 | $redirect = 1; |
|
|
568 | } |
|
|
569 | } |
|
|
570 | |
449 | my $finish = sub { |
571 | my $finish = sub { |
450 | $state{handle}->destroy; |
572 | $state{handle}->destroy if $state{handle}; |
451 | %state = (); |
573 | %state = (); |
452 | |
574 | |
453 | # set-cookie processing |
575 | # set-cookie processing |
454 | if ($arg{cookie_jar}) { |
576 | if ($arg{cookie_jar}) { |
455 | for ($hdr{"set-cookie"}) { |
577 | for ($_[1]{"set-cookie"}) { |
456 | # parse NAME=VALUE |
578 | # parse NAME=VALUE |
457 | my @kv; |
579 | my @kv; |
458 | |
580 | |
459 | while (/\G\s* ([^=;,[:space:]]+) \s*=\s* (?: "((?:[^\\"]+|\\.)*)" | ([^=;,[:space:]]*) )/gcxs) { |
581 | while (/\G\s* ([^=;,[:space:]]+) \s*=\s* (?: "((?:[^\\"]+|\\.)*)" | ([^=;,[:space:]]*) )/gcxs) { |
460 | my $name = $1; |
582 | my $name = $1; |
… | |
… | |
498 | |
620 | |
499 | redo if /\G\s*,/gc; |
621 | redo if /\G\s*,/gc; |
500 | } |
622 | } |
501 | } |
623 | } |
502 | |
624 | |
503 | # microsoft and other shitheads don't give a shit for following standards, |
625 | if ($redirect && exists $hdr{location}) { |
504 | # try to support some common forms of broken Location headers. |
626 | # we ignore any errors, as it is very common to receive |
505 | if ($_[1]{location} !~ /^(?: $ | [^:\/?\#]+ : )/x) { |
627 | # Content-Length != 0 but no actual body |
506 | $_[1]{location} =~ s/^\.\/+//; |
628 | # we also access %hdr, as $_[1] might be an erro |
507 | |
|
|
508 | my $url = "$rscheme://$uhost:$uport"; |
|
|
509 | |
|
|
510 | unless ($_[1]{location} =~ s/^\///) { |
|
|
511 | $url .= $upath; |
|
|
512 | $url =~ s/\/[^\/]*$//; |
|
|
513 | } |
|
|
514 | |
|
|
515 | $_[1]{location} = "$url/$_[1]{location}"; |
|
|
516 | } |
|
|
517 | |
|
|
518 | if ($_[1]{Status} =~ /^30[12]$/ && $recurse && $method ne "POST") { |
|
|
519 | # apparently, mozilla et al. just change POST to GET here |
|
|
520 | # more research is needed before we do the same |
|
|
521 | http_request ($method => $_[1]{location}, %arg, recurse => $recurse - 1, $cb); |
629 | http_request ($method => $hdr{location}, %arg, recurse => $recurse - 1, $cb); |
522 | } elsif ($_[1]{Status} == 303 && $recurse) { |
|
|
523 | # even http/1.1 is unclear on how to mutate the method |
|
|
524 | $method = "GET" unless $method eq "HEAD"; |
|
|
525 | http_request ($method => $_[1]{location}, %arg, recurse => $recurse - 1, $cb); |
|
|
526 | } elsif ($_[1]{Status} == 307 && $recurse && $method =~ /^(?:GET|HEAD)$/) { |
|
|
527 | http_request ($method => $_[1]{location}, %arg, recurse => $recurse - 1, $cb); |
|
|
528 | } else { |
630 | } else { |
529 | $cb->($_[0], $_[1]); |
631 | $cb->($_[0], $_[1]); |
530 | } |
632 | } |
531 | }; |
633 | }; |
532 | |
634 | |
533 | if ($hdr{Status} =~ /^(?:1..|204|304)$/ or $method eq "HEAD") { |
635 | my $len = $hdr{"content-length"}; |
|
|
636 | |
|
|
637 | if (!$redirect && $arg{on_header} && !$arg{on_header}(\%hdr)) { |
|
|
638 | $finish->(undef, { Status => 598, Reason => "Request cancelled by on_header", URL => $url }); |
|
|
639 | } elsif ( |
|
|
640 | $hdr{Status} =~ /^(?:1..|[23]04)$/ |
|
|
641 | or $method eq "HEAD" |
|
|
642 | or (defined $len && !$len) |
|
|
643 | ) { |
|
|
644 | # no body |
534 | $finish->(undef, \%hdr); |
645 | $finish->("", \%hdr); |
535 | } else { |
646 | } else { |
536 | if (exists $hdr{"content-length"}) { |
647 | # body handling, four different code paths |
537 | $_[0]->unshift_read (chunk => $hdr{"content-length"}, sub { |
648 | # for want_body_handle, on_body (2x), normal (2x) |
538 | # could cache persistent connection now |
649 | # we might read too much here, but it does not matter yet (no pers. connections) |
539 | if ($hdr{connection} =~ /\bkeep-alive\b/i) { |
650 | if (!$redirect && $arg{want_body_handle}) { |
540 | # but we don't, due to misdesigns, this is annoyingly complex |
651 | $_[0]->on_eof (undef); |
|
|
652 | $_[0]->on_error (undef); |
|
|
653 | $_[0]->on_read (undef); |
|
|
654 | |
|
|
655 | $finish->(delete $state{handle}, \%hdr); |
|
|
656 | |
|
|
657 | } elsif ($arg{on_body}) { |
|
|
658 | $_[0]->on_error (sub { $finish->(undef, { Status => 599, Reason => $_[2], URL => $url }) }); |
|
|
659 | if ($len) { |
|
|
660 | $_[0]->on_eof (undef); |
|
|
661 | $_[0]->on_read (sub { |
|
|
662 | $len -= length $_[0]{rbuf}; |
|
|
663 | |
|
|
664 | $arg{on_body}(delete $_[0]{rbuf}, \%hdr) |
|
|
665 | or $finish->(undef, { Status => 598, Reason => "Request cancelled by on_body", URL => $url }); |
|
|
666 | |
|
|
667 | $len > 0 |
|
|
668 | or $finish->("", \%hdr); |
541 | }; |
669 | }); |
542 | |
670 | } else { |
|
|
671 | $_[0]->on_eof (sub { |
543 | $finish->($_[1], \%hdr); |
672 | $finish->("", \%hdr); |
|
|
673 | }); |
|
|
674 | $_[0]->on_read (sub { |
|
|
675 | $arg{on_body}(delete $_[0]{rbuf}, \%hdr) |
|
|
676 | or $finish->(undef, { Status => 598, Reason => "Request cancelled by on_body", URL => $url }); |
|
|
677 | }); |
544 | }); |
678 | } |
545 | } else { |
679 | } else { |
546 | # too bad, need to read until we get an error or EOF, |
|
|
547 | # no way to detect winged data. |
|
|
548 | $_[0]->on_error (sub { |
|
|
549 | # delete ought to be more efficient, as we would have to make |
|
|
550 | # a copy otherwise as $_[0] gets destroyed. |
|
|
551 | $finish->(delete $_[0]{rbuf}, \%hdr); |
|
|
552 | }); |
|
|
553 | $_[0]->on_eof (undef); |
680 | $_[0]->on_eof (undef); |
|
|
681 | |
|
|
682 | if ($len) { |
|
|
683 | $_[0]->on_error (sub { $finish->(undef, { Status => 599, Reason => $_[2], URL => $url }) }); |
|
|
684 | $_[0]->on_read (sub { |
|
|
685 | $finish->((substr delete $_[0]{rbuf}, 0, $len, ""), \%hdr) |
|
|
686 | if $len <= length $_[0]{rbuf}; |
|
|
687 | }); |
|
|
688 | } else { |
|
|
689 | $_[0]->on_error (sub { |
|
|
690 | $! == Errno::EPIPE || !$! |
|
|
691 | ? $finish->(delete $_[0]{rbuf}, \%hdr) |
|
|
692 | : $finish->(undef, { Status => 599, Reason => $_[2], URL => $url }); |
|
|
693 | }); |
554 | $_[0]->on_read (sub { }); |
694 | $_[0]->on_read (sub { }); |
|
|
695 | } |
555 | } |
696 | } |
556 | } |
697 | } |
557 | }); |
698 | }); |
558 | }); |
699 | }); |
559 | }; |
700 | }; |
… | |
… | |
578 | }); |
719 | }); |
579 | } else { |
720 | } else { |
580 | &$handle_actual_request; |
721 | &$handle_actual_request; |
581 | } |
722 | } |
582 | |
723 | |
583 | }, sub { |
724 | }, $arg{on_prepare} || sub { $timeout }; |
584 | $timeout |
|
|
585 | }; |
|
|
586 | }; |
725 | }; |
587 | |
726 | |
588 | defined wantarray && AnyEvent::Util::guard { %state = () } |
727 | defined wantarray && AnyEvent::Util::guard { %state = () } |
589 | } |
728 | } |
590 | |
729 | |
… | |
… | |
611 | =over 4 |
750 | =over 4 |
612 | |
751 | |
613 | =item AnyEvent::HTTP::set_proxy "proxy-url" |
752 | =item AnyEvent::HTTP::set_proxy "proxy-url" |
614 | |
753 | |
615 | Sets the default proxy server to use. The proxy-url must begin with a |
754 | Sets the default proxy server to use. The proxy-url must begin with a |
616 | string of the form C<http://host:port> (optionally C<https:...>). |
755 | string of the form C<http://host:port> (optionally C<https:...>), croaks |
|
|
756 | otherwise. |
|
|
757 | |
|
|
758 | To clear an already-set proxy, use C<undef>. |
617 | |
759 | |
618 | =item $AnyEvent::HTTP::MAX_RECURSE |
760 | =item $AnyEvent::HTTP::MAX_RECURSE |
619 | |
761 | |
620 | The default value for the C<recurse> request parameter (default: C<10>). |
762 | The default value for the C<recurse> request parameter (default: C<10>). |
621 | |
763 | |
622 | =item $AnyEvent::HTTP::USERAGENT |
764 | =item $AnyEvent::HTTP::USERAGENT |
623 | |
765 | |
624 | The default value for the C<User-Agent> header (the default is |
766 | The default value for the C<User-Agent> header (the default is |
625 | C<Mozilla/5.0 (compatible; U; AnyEvent-HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)>). |
767 | C<Mozilla/5.0 (compatible; U; AnyEvent-HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)>). |
626 | |
768 | |
627 | =item $AnyEvent::HTTP::MAX_PERSISTENT |
769 | =item $AnyEvent::HTTP::MAX_PER_HOST |
628 | |
770 | |
629 | The maximum number of persistent connections to keep open (default: 8). |
771 | The maximum number of concurrent connections to the same host (identified |
|
|
772 | by the hostname). If the limit is exceeded, then the additional requests |
|
|
773 | are queued until previous connections are closed. |
630 | |
774 | |
631 | Not implemented currently. |
775 | The default value for this is C<4>, and it is highly advisable to not |
632 | |
776 | increase it. |
633 | =item $AnyEvent::HTTP::PERSISTENT_TIMEOUT |
|
|
634 | |
|
|
635 | The maximum time to cache a persistent connection, in seconds (default: 2). |
|
|
636 | |
|
|
637 | Not implemented currently. |
|
|
638 | |
777 | |
639 | =item $AnyEvent::HTTP::ACTIVE |
778 | =item $AnyEvent::HTTP::ACTIVE |
640 | |
779 | |
641 | The number of active connections. This is not the number of currently |
780 | The number of active connections. This is not the number of currently |
642 | running requests, but the number of currently open and non-idle TCP |
781 | running requests, but the number of currently open and non-idle TCP |
… | |
… | |
645 | =back |
784 | =back |
646 | |
785 | |
647 | =cut |
786 | =cut |
648 | |
787 | |
649 | sub set_proxy($) { |
788 | sub set_proxy($) { |
|
|
789 | if (length $_[0]) { |
650 | $PROXY = [$2, $3 || 3128, $1] if $_[0] =~ m%^(https?):// ([^:/]+) (?: : (\d*) )?%ix; |
790 | $_[0] =~ m%^(https?):// ([^:/]+) (?: : (\d*) )?%ix |
|
|
791 | or Carp::croak "$_[0]: invalid proxy URL"; |
|
|
792 | $PROXY = [$2, $3 || 3128, $1] |
|
|
793 | } else { |
|
|
794 | undef $PROXY; |
|
|
795 | } |
651 | } |
796 | } |
652 | |
797 | |
653 | # initialise proxy from environment |
798 | # initialise proxy from environment |
|
|
799 | eval { |
654 | set_proxy $ENV{http_proxy}; |
800 | set_proxy $ENV{http_proxy}; |
|
|
801 | }; |
655 | |
802 | |
656 | =head1 SEE ALSO |
803 | =head1 SEE ALSO |
657 | |
804 | |
658 | L<AnyEvent>. |
805 | L<AnyEvent>. |
659 | |
806 | |