… | |
… | |
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 (); |
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 = '1.1'; |
53 | our $VERSION = '1.45'; |
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 $USERAGENT = "Mozilla/5.0 (compatible; AnyEvent::HTTP/$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" (uppercase to avoid clashing with possible |
|
|
105 | response headers) C<HTTPVersion>, C<Status> and C<Reason> contain the |
105 | contain the three parts of the HTTP Status-Line of the same name. The |
106 | three parts of the HTTP Status-Line of the same name. |
|
|
107 | |
106 | pseudo-header C<URL> contains the original URL (which can differ from the |
108 | The pseudo-header C<URL> contains the actual URL (which can differ from |
107 | requested URL when following redirects). |
109 | the requested URL when following redirects - for example, you might get |
|
|
110 | an error that your URL scheme is not supported even though your URL is a |
|
|
111 | valid http URL because it redirected to an ftp URL, in which case you can |
|
|
112 | look at the URL pseudo header). |
|
|
113 | |
|
|
114 | The pseudo-header C<Redirect> only exists when the request was a result |
|
|
115 | of an internal redirect. In that case it is an array reference with |
|
|
116 | the C<($data, $headers)> from the redirect response. Note that this |
|
|
117 | response could in turn be the result of a redirect itself, and C<< |
|
|
118 | $headers->{Redirect}[1]{Redirect} >> will then contain the original |
|
|
119 | response, and so on. |
108 | |
120 | |
109 | If the server sends a header multiple times, then their contents will be |
121 | If the server sends a header multiple times, then their contents will be |
110 | joined together with a comma (C<,>), as per the HTTP spec. |
122 | joined together with a comma (C<,>), as per the HTTP spec. |
111 | |
123 | |
112 | If an internal error occurs, such as not being able to resolve a hostname, |
124 | If an internal error occurs, such as not being able to resolve a hostname, |
113 | then C<$data> will be C<undef>, C<< $headers->{Status} >> will be C<599> |
125 | then C<$data> will be C<undef>, C<< $headers->{Status} >> will be C<59x> |
114 | and the C<Reason> pseudo-header will contain an error message. |
126 | (usually C<599>) and the C<Reason> pseudo-header will contain an error |
|
|
127 | message. |
115 | |
128 | |
116 | A typical callback might look like this: |
129 | A typical callback might look like this: |
117 | |
130 | |
118 | sub { |
131 | sub { |
119 | my ($body, $hdr) = @_; |
132 | my ($body, $hdr) = @_; |
… | |
… | |
137 | |
150 | |
138 | =item headers => hashref |
151 | =item headers => hashref |
139 | |
152 | |
140 | The request headers to use. Currently, C<http_request> may provide its |
153 | 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 |
154 | own C<Host:>, C<Content-Length:>, C<Connection:> and C<Cookie:> headers |
142 | and will provide defaults for C<User-Agent:> and C<Referer:>. |
155 | and will provide defaults for C<User-Agent:> and C<Referer:> (this can be |
|
|
156 | suppressed by using C<undef> for these headers in which case they won't be |
|
|
157 | sent at all). |
143 | |
158 | |
144 | =item timeout => $seconds |
159 | =item timeout => $seconds |
145 | |
160 | |
146 | The time-out to use for various stages - each connect attempt will reset |
161 | The time-out to use for various stages - each connect attempt will reset |
147 | the timeout, as will read or write activity. Default timeout is 5 minutes. |
162 | the timeout, as will read or write activity, i.e. this is not an overall |
|
|
163 | timeout. |
|
|
164 | |
|
|
165 | Default timeout is 5 minutes. |
148 | |
166 | |
149 | =item proxy => [$host, $port[, $scheme]] or undef |
167 | =item proxy => [$host, $port[, $scheme]] or undef |
150 | |
168 | |
151 | Use the given http proxy for all requests. If not specified, then the |
169 | Use the given http proxy for all requests. If not specified, then the |
152 | default proxy (as specified by C<$ENV{http_proxy}>) is used. |
170 | default proxy (as specified by C<$ENV{http_proxy}>) is used. |
153 | |
171 | |
154 | C<$scheme> must be either missing or C<http> for HTTP, or C<https> for |
172 | C<$scheme> must be either missing, C<http> for HTTP or C<https> for |
155 | HTTPS. |
173 | HTTPS. |
156 | |
174 | |
157 | =item body => $string |
175 | =item body => $string |
158 | |
176 | |
159 | The request body, usually empty. Will be-sent as-is (future versions of |
177 | The request body, usually empty. Will be-sent as-is (future versions of |
… | |
… | |
165 | based on the original netscape specification. |
183 | based on the original netscape specification. |
166 | |
184 | |
167 | The C<$hash_ref> must be an (initially empty) hash reference which will |
185 | The C<$hash_ref> must be an (initially empty) hash reference which will |
168 | get updated automatically. It is possible to save the cookie_jar to |
186 | get updated automatically. It is possible to save the cookie_jar to |
169 | persistent storage with something like JSON or Storable, but this is not |
187 | persistent storage with something like JSON or Storable, but this is not |
170 | recommended, as expire times are currently being ignored. |
188 | recommended, as expiry times are currently being ignored. |
171 | |
189 | |
172 | Note that this cookie implementation is not of very high quality, nor |
190 | Note that this cookie implementation is not of very high quality, nor |
173 | meant to be complete. If you want complete cookie management you have to |
191 | meant to be complete. If you want complete cookie management you have to |
174 | do that on your own. C<cookie_jar> is meant as a quick fix to get some |
192 | do that on your own. C<cookie_jar> is meant as a quick fix to get some |
175 | cookie-using sites working. Cookies are a privacy disaster, do not use |
193 | cookie-using sites working. Cookies are a privacy disaster, do not use |
176 | them unless required to. |
194 | them unless required to. |
|
|
195 | |
|
|
196 | =item tls_ctx => $scheme | $tls_ctx |
|
|
197 | |
|
|
198 | Specifies the AnyEvent::TLS context to be used for https connections. This |
|
|
199 | parameter follows the same rules as the C<tls_ctx> parameter to |
|
|
200 | L<AnyEvent::Handle>, but additionally, the two strings C<low> or |
|
|
201 | C<high> can be specified, which give you a predefined low-security (no |
|
|
202 | verification, highest compatibility) and high-security (CA and common-name |
|
|
203 | verification) TLS context. |
|
|
204 | |
|
|
205 | The default for this option is C<low>, which could be interpreted as "give |
|
|
206 | me the page, no matter what". |
|
|
207 | |
|
|
208 | =item on_prepare => $callback->($fh) |
|
|
209 | |
|
|
210 | In rare cases you need to "tune" the socket before it is used to |
|
|
211 | connect (for exmaple, to bind it on a given IP address). This parameter |
|
|
212 | overrides the prepare callback passed to C<AnyEvent::Socket::tcp_connect> |
|
|
213 | and behaves exactly the same way (e.g. it has to provide a |
|
|
214 | timeout). See the description for the C<$prepare_cb> argument of |
|
|
215 | C<AnyEvent::Socket::tcp_connect> for details. |
|
|
216 | |
|
|
217 | =item on_header => $callback->($headers) |
|
|
218 | |
|
|
219 | When specified, this callback will be called with the header hash as soon |
|
|
220 | as headers have been successfully received from the remote server (not on |
|
|
221 | locally-generated errors). |
|
|
222 | |
|
|
223 | It has to return either true (in which case AnyEvent::HTTP will continue), |
|
|
224 | or false, in which case AnyEvent::HTTP will cancel the download (and call |
|
|
225 | the finish callback with an error code of C<598>). |
|
|
226 | |
|
|
227 | This callback is useful, among other things, to quickly reject unwanted |
|
|
228 | content, which, if it is supposed to be rare, can be faster than first |
|
|
229 | doing a C<HEAD> request. |
|
|
230 | |
|
|
231 | Example: cancel the request unless the content-type is "text/html". |
|
|
232 | |
|
|
233 | on_header => sub { |
|
|
234 | $_[0]{"content-type"} =~ /^text\/html\s*(?:;|$)/ |
|
|
235 | }, |
|
|
236 | |
|
|
237 | =item on_body => $callback->($partial_body, $headers) |
|
|
238 | |
|
|
239 | When specified, all body data will be passed to this callback instead of |
|
|
240 | to the completion callback. The completion callback will get the empty |
|
|
241 | string instead of the body data. |
|
|
242 | |
|
|
243 | It has to return either true (in which case AnyEvent::HTTP will continue), |
|
|
244 | or false, in which case AnyEvent::HTTP will cancel the download (and call |
|
|
245 | the completion callback with an error code of C<598>). |
|
|
246 | |
|
|
247 | This callback is useful when the data is too large to be held in memory |
|
|
248 | (so the callback writes it to a file) or when only some information should |
|
|
249 | be extracted, or when the body should be processed incrementally. |
|
|
250 | |
|
|
251 | It is usually preferred over doing your own body handling via |
|
|
252 | C<want_body_handle>, but in case of streaming APIs, where HTTP is |
|
|
253 | only used to create a connection, C<want_body_handle> is the better |
|
|
254 | alternative, as it allows you to install your own event handler, reducing |
|
|
255 | resource usage. |
|
|
256 | |
|
|
257 | =item want_body_handle => $enable |
|
|
258 | |
|
|
259 | When enabled (default is disabled), the behaviour of AnyEvent::HTTP |
|
|
260 | changes considerably: after parsing the headers, and instead of |
|
|
261 | downloading the body (if any), the completion callback will be |
|
|
262 | called. Instead of the C<$body> argument containing the body data, the |
|
|
263 | callback will receive the L<AnyEvent::Handle> object associated with the |
|
|
264 | connection. In error cases, C<undef> will be passed. When there is no body |
|
|
265 | (e.g. status C<304>), the empty string will be passed. |
|
|
266 | |
|
|
267 | The handle object might or might not be in TLS mode, might be connected to |
|
|
268 | a proxy, be a persistent connection etc., and configured in unspecified |
|
|
269 | ways. The user is responsible for this handle (it will not be used by this |
|
|
270 | module anymore). |
|
|
271 | |
|
|
272 | This is useful with some push-type services, where, after the initial |
|
|
273 | headers, an interactive protocol is used (typical example would be the |
|
|
274 | push-style twitter API which starts a JSON/XML stream). |
|
|
275 | |
|
|
276 | If you think you need this, first have a look at C<on_body>, to see if |
|
|
277 | that doesn't solve your problem in a better way. |
177 | |
278 | |
178 | =back |
279 | =back |
179 | |
280 | |
180 | Example: make a simple HTTP GET request for http://www.nethype.de/ |
281 | Example: make a simple HTTP GET request for http://www.nethype.de/ |
181 | |
282 | |
… | |
… | |
237 | push @{ $CO_SLOT{$_[0]}[1] }, $_[1]; |
338 | push @{ $CO_SLOT{$_[0]}[1] }, $_[1]; |
238 | |
339 | |
239 | _slot_schedule $_[0]; |
340 | _slot_schedule $_[0]; |
240 | } |
341 | } |
241 | |
342 | |
242 | our $qr_nl = qr<\015?\012>; |
343 | our $qr_nl = qr{\015?\012}; |
243 | our $qr_nlnl = qr<\015?\012\015?\012>; |
344 | our $qr_nlnl = qr{(?<![^\012])\015?\012}; |
|
|
345 | |
|
|
346 | our $TLS_CTX_LOW = { cache => 1, sslv2 => 1 }; |
|
|
347 | our $TLS_CTX_HIGH = { cache => 1, verify => 1, verify_peername => "https" }; |
244 | |
348 | |
245 | sub http_request($$@) { |
349 | sub http_request($$@) { |
246 | my $cb = pop; |
350 | my $cb = pop; |
247 | my ($method, $url, %arg) = @_; |
351 | my ($method, $url, %arg) = @_; |
248 | |
352 | |
249 | my %hdr; |
353 | my %hdr; |
|
|
354 | |
|
|
355 | $arg{tls_ctx} = $TLS_CTX_LOW if $arg{tls_ctx} eq "low" || !exists $arg{tls_ctx}; |
|
|
356 | $arg{tls_ctx} = $TLS_CTX_HIGH if $arg{tls_ctx} eq "high"; |
250 | |
357 | |
251 | $method = uc $method; |
358 | $method = uc $method; |
252 | |
359 | |
253 | if (my $hdr = $arg{headers}) { |
360 | if (my $hdr = $arg{headers}) { |
254 | while (my ($k, $v) = each %$hdr) { |
361 | while (my ($k, $v) = each %$hdr) { |
255 | $hdr{lc $k} = $v; |
362 | $hdr{lc $k} = $v; |
256 | } |
363 | } |
257 | } |
364 | } |
258 | |
365 | |
|
|
366 | # pseudo headers for all subsequent responses |
|
|
367 | my @pseudo = (URL => $url); |
|
|
368 | push @pseudo, Redirect => delete $arg{Redirect} if exists $arg{Redirect}; |
|
|
369 | |
259 | my $recurse = exists $arg{recurse} ? delete $arg{recurse} : $MAX_RECURSE; |
370 | my $recurse = exists $arg{recurse} ? delete $arg{recurse} : $MAX_RECURSE; |
260 | |
371 | |
261 | return $cb->(undef, { Status => 599, Reason => "recursion limit reached", URL => $url }) |
372 | return $cb->(undef, { Status => 599, Reason => "Too many redirections", @pseudo }) |
262 | if $recurse < 0; |
373 | if $recurse < 0; |
263 | |
374 | |
264 | my $proxy = $arg{proxy} || $PROXY; |
375 | my $proxy = $arg{proxy} || $PROXY; |
265 | my $timeout = $arg{timeout} || $TIMEOUT; |
376 | my $timeout = $arg{timeout} || $TIMEOUT; |
266 | |
377 | |
267 | $hdr{"user-agent"} ||= $USERAGENT; |
|
|
268 | |
|
|
269 | my ($uscheme, $uauthority, $upath, $query, $fragment) = |
378 | my ($uscheme, $uauthority, $upath, $query, $fragment) = |
270 | $url =~ m|(?:([^:/?#]+):)?(?://([^/?#]*))?([^?#]*)(?:\?([^#]*))?(?:#(.*))?|; |
379 | $url =~ m|(?:([^:/?#]+):)?(?://([^/?#]*))?([^?#]*)(?:\?([^#]*))?(?:#(.*))?|; |
271 | |
380 | |
272 | $uscheme = lc $uscheme; |
381 | $uscheme = lc $uscheme; |
273 | |
382 | |
274 | my $uport = $uscheme eq "http" ? 80 |
383 | my $uport = $uscheme eq "http" ? 80 |
275 | : $uscheme eq "https" ? 443 |
384 | : $uscheme eq "https" ? 443 |
276 | : return $cb->(undef, { Status => 599, Reason => "only http and https URL schemes supported", URL => $url }); |
385 | : return $cb->(undef, { Status => 599, Reason => "Only http and https URL schemes supported", @pseudo }); |
277 | |
|
|
278 | $hdr{referer} ||= "$uscheme://$uauthority$upath"; # leave out fragment and query string, just a heuristic |
|
|
279 | |
386 | |
280 | $uauthority =~ /^(?: .*\@ )? ([^\@:]+) (?: : (\d+) )?$/x |
387 | $uauthority =~ /^(?: .*\@ )? ([^\@:]+) (?: : (\d+) )?$/x |
281 | or return $cb->(undef, { Status => 599, Reason => "unparsable URL", URL => $url }); |
388 | or return $cb->(undef, { Status => 599, Reason => "Unparsable URL", @pseudo }); |
282 | |
389 | |
283 | my $uhost = $1; |
390 | my $uhost = $1; |
284 | $uport = $2 if defined $2; |
391 | $uport = $2 if defined $2; |
|
|
392 | |
|
|
393 | $hdr{host} = defined $2 ? "$uhost:$2" : "$uhost" |
|
|
394 | unless exists $hdr{host}; |
285 | |
395 | |
286 | $uhost =~ s/^\[(.*)\]$/$1/; |
396 | $uhost =~ s/^\[(.*)\]$/$1/; |
287 | $upath .= "?$query" if length $query; |
397 | $upath .= "?$query" if length $query; |
288 | |
398 | |
289 | $upath =~ s%^/?%/%; |
399 | $upath =~ s%^/?%/%; |
… | |
… | |
320 | } |
430 | } |
321 | |
431 | |
322 | my ($rhost, $rport, $rscheme, $rpath); # request host, port, path |
432 | my ($rhost, $rport, $rscheme, $rpath); # request host, port, path |
323 | |
433 | |
324 | if ($proxy) { |
434 | if ($proxy) { |
325 | ($rhost, $rport, $rscheme, $rpath) = (@$proxy, $url); |
435 | ($rpath, $rhost, $rport, $rscheme) = ($url, @$proxy); |
|
|
436 | |
|
|
437 | $rscheme = "http" unless defined $rscheme; |
326 | |
438 | |
327 | # don't support https requests over https-proxy transport, |
439 | # don't support https requests over https-proxy transport, |
328 | # can't be done with tls as spec'ed. |
440 | # can't be done with tls as spec'ed, unless you double-encrypt. |
329 | $rscheme = "http" if $uscheme eq "https" && $rscheme eq "https"; |
441 | $rscheme = "http" if $uscheme eq "https" && $rscheme eq "https"; |
330 | } else { |
442 | } else { |
331 | ($rhost, $rport, $rscheme, $rpath) = ($uhost, $uport, $uscheme, $upath); |
443 | ($rhost, $rport, $rscheme, $rpath) = ($uhost, $uport, $uscheme, $upath); |
332 | } |
444 | } |
333 | |
445 | |
334 | $hdr{host} = $uhost; |
446 | # leave out fragment and query string, just a heuristic |
|
|
447 | $hdr{referer} ||= "$uscheme://$uauthority$upath" unless exists $hdr{referer}; |
|
|
448 | $hdr{"user-agent"} ||= $USERAGENT unless exists $hdr{"user-agent"}; |
|
|
449 | |
335 | $hdr{"content-length"} = length $arg{body}; |
450 | $hdr{"content-length"} = length $arg{body} |
|
|
451 | if length $arg{body} || $method ne "GET"; |
336 | |
452 | |
337 | my %state = (connect_guard => 1); |
453 | my %state = (connect_guard => 1); |
338 | |
454 | |
339 | _get_slot $uhost, sub { |
455 | _get_slot $uhost, sub { |
340 | $state{slot_guard} = shift; |
456 | $state{slot_guard} = shift; |
341 | |
457 | |
342 | return unless $state{connect_guard}; |
458 | return unless $state{connect_guard}; |
343 | |
459 | |
344 | $state{connect_guard} = AnyEvent::Socket::tcp_connect $rhost, $rport, sub { |
460 | $state{connect_guard} = AnyEvent::Socket::tcp_connect $rhost, $rport, sub { |
345 | $state{fh} = shift |
461 | $state{fh} = shift |
|
|
462 | or do { |
|
|
463 | my $err = "$!"; |
|
|
464 | %state = (); |
346 | or return $cb->(undef, { Status => 599, Reason => "$!", URL => $url }); |
465 | return $cb->(undef, { Status => 599, Reason => $err, @pseudo }); |
|
|
466 | }; |
|
|
467 | |
347 | pop; # free memory, save a tree |
468 | pop; # free memory, save a tree |
348 | |
469 | |
349 | return unless delete $state{connect_guard}; |
470 | return unless delete $state{connect_guard}; |
350 | |
471 | |
351 | # get handle |
472 | # get handle |
352 | $state{handle} = new AnyEvent::Handle |
473 | $state{handle} = new AnyEvent::Handle |
353 | fh => $state{fh}, |
474 | fh => $state{fh}, |
354 | timeout => $timeout; |
475 | timeout => $timeout, |
|
|
476 | peername => $rhost, |
|
|
477 | tls_ctx => $arg{tls_ctx}; |
355 | |
478 | |
356 | # limit the number of persistent connections |
479 | # limit the number of persistent connections |
357 | # keepalive not yet supported |
480 | # keepalive not yet supported |
358 | if ($KA_COUNT{$_[1]} < $MAX_PERSISTENT_PER_HOST) { |
481 | if ($KA_COUNT{$_[1]} < $MAX_PERSISTENT_PER_HOST) { |
359 | ++$KA_COUNT{$_[1]}; |
482 | ++$KA_COUNT{$_[1]}; |
… | |
… | |
365 | delete $hdr{connection}; |
488 | delete $hdr{connection}; |
366 | } |
489 | } |
367 | |
490 | |
368 | # (re-)configure handle |
491 | # (re-)configure handle |
369 | $state{handle}->on_error (sub { |
492 | $state{handle}->on_error (sub { |
370 | my $errno = "$!"; |
|
|
371 | %state = (); |
493 | %state = (); |
372 | $cb->(undef, { Status => 599, Reason => $errno, URL => $url }); |
494 | $cb->(undef, { Status => 599, Reason => $_[2], @pseudo }); |
373 | }); |
495 | }); |
374 | $state{handle}->on_eof (sub { |
496 | $state{handle}->on_eof (sub { |
375 | %state = (); |
497 | %state = (); |
376 | $cb->(undef, { Status => 599, Reason => "unexpected end-of-file", URL => $url }); |
498 | $cb->(undef, { Status => 599, Reason => "Unexpected end-of-file", @pseudo }); |
377 | }); |
499 | }); |
378 | |
500 | |
379 | $state{handle}->starttls ("connect") if $rscheme eq "https"; |
501 | $state{handle}->starttls ("connect") if $rscheme eq "https"; |
380 | |
502 | |
381 | # handle actual, non-tunneled, request |
503 | # handle actual, non-tunneled, request |
… | |
… | |
383 | $state{handle}->starttls ("connect") if $uscheme eq "https" && !exists $state{handle}{tls}; |
505 | $state{handle}->starttls ("connect") if $uscheme eq "https" && !exists $state{handle}{tls}; |
384 | |
506 | |
385 | # send request |
507 | # send request |
386 | $state{handle}->push_write ( |
508 | $state{handle}->push_write ( |
387 | "$method $rpath HTTP/1.0\015\012" |
509 | "$method $rpath HTTP/1.0\015\012" |
388 | . (join "", map "\u$_: $hdr{$_}\015\012", keys %hdr) |
510 | . (join "", map "\u$_: $hdr{$_}\015\012", grep defined $hdr{$_}, keys %hdr) |
389 | . "\015\012" |
511 | . "\015\012" |
390 | . (delete $arg{body}) |
512 | . (delete $arg{body}) |
391 | ); |
513 | ); |
392 | |
514 | |
|
|
515 | # return if error occured during push_write() |
|
|
516 | return unless %state; |
|
|
517 | |
393 | %hdr = (); # reduce memory usage, save a kitten |
518 | %hdr = (); # reduce memory usage, save a kitten, also make it possible to re-use |
394 | |
519 | |
395 | # status line |
520 | # status line |
396 | $state{handle}->push_read (line => $qr_nl, sub { |
521 | $state{handle}->push_read (line => $qr_nl, sub { |
397 | $_[1] =~ /^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\015\012]*) )?/ix |
522 | $_[1] =~ /^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\015\012]*) )?/ix |
398 | or return (%state = (), $cb->(undef, { Status => 599, Reason => "invalid server response ($_[1])", URL => $url })); |
523 | or return (%state = (), $cb->(undef, { Status => 599, Reason => "Invalid server response ($_[1])", @pseudo })); |
399 | |
524 | |
400 | my %hdr = ( # response headers |
525 | push @pseudo, |
401 | HTTPVersion => ",$1", |
526 | HTTPVersion => $1, |
402 | Status => ",$2", |
527 | Status => $2, |
403 | Reason => ",$3", |
528 | Reason => $3, |
404 | URL => ",$url" |
|
|
405 | ); |
529 | ; |
406 | |
530 | |
407 | # headers, could be optimized a bit |
531 | # headers, could be optimized a bit |
408 | $state{handle}->unshift_read (line => $qr_nlnl, sub { |
532 | $state{handle}->unshift_read (line => $qr_nlnl, sub { |
409 | for ("$_[1]\012") { |
533 | for ("$_[1]") { |
410 | y/\015//d; # weed out any \015, as they show up in the weirdest of places. |
534 | y/\015//d; # weed out any \015, as they show up in the weirdest of places. |
411 | |
535 | |
412 | # we support spaces in field names, as lotus domino |
536 | # things seen, not parsed: |
413 | # creates them (actually spaces around seperators |
537 | # p3pP="NON CUR OTPi OUR NOR UNI" |
414 | # are strictly allowed in http, they are a security issue). |
538 | |
415 | $hdr{lc $1} .= ",$2" |
539 | $hdr{lc $1} .= ",$2" |
416 | while /\G |
540 | while /\G |
417 | ([^:\000-\037]+): |
541 | ([^:\000-\037]*): |
418 | [\011\040]* |
542 | [\011\040]* |
419 | ((?: [^\012]+ | \012[\011\040] )*) |
543 | ((?: [^\012]+ | \012[\011\040] )*) |
420 | \012 |
544 | \012 |
421 | /gxc; |
545 | /gxc; |
422 | |
546 | |
423 | /\G$/ |
547 | /\G$/ |
424 | or return (%state = (), $cb->(undef, { Status => 599, Reason => "garbled response headers", URL => $url })); |
548 | or return (%state = (), $cb->(undef, { Status => 599, Reason => "Garbled response headers", @pseudo })); |
425 | } |
549 | } |
426 | |
550 | |
|
|
551 | # remove the "," prefix we added to all headers above |
427 | substr $_, 0, 1, "" |
552 | substr $_, 0, 1, "" |
428 | for values %hdr; |
553 | for values %hdr; |
429 | |
554 | |
|
|
555 | # patch in all pseudo headers |
|
|
556 | %hdr = (%hdr, @pseudo); |
|
|
557 | |
|
|
558 | # redirect handling |
|
|
559 | # microsoft and other shitheads don't give a shit for following standards, |
|
|
560 | # try to support some common forms of broken Location headers. |
|
|
561 | if ($hdr{location} !~ /^(?: $ | [^:\/?\#]+ : )/x) { |
|
|
562 | $hdr{location} =~ s/^\.\/+//; |
|
|
563 | |
|
|
564 | my $url = "$rscheme://$uhost:$uport"; |
|
|
565 | |
|
|
566 | unless ($hdr{location} =~ s/^\///) { |
|
|
567 | $url .= $upath; |
|
|
568 | $url =~ s/\/[^\/]*$//; |
|
|
569 | } |
|
|
570 | |
|
|
571 | $hdr{location} = "$url/$hdr{location}"; |
|
|
572 | } |
|
|
573 | |
|
|
574 | my $redirect; |
|
|
575 | |
|
|
576 | if ($recurse) { |
|
|
577 | my $status = $hdr{Status}; |
|
|
578 | |
|
|
579 | if (($status == 301 || $status == 302) && $method ne "POST") { |
|
|
580 | # apparently, mozilla et al. just change POST to GET here |
|
|
581 | # more research is needed before we do the same |
|
|
582 | $redirect = 1; |
|
|
583 | } elsif ($status == 303) { |
|
|
584 | # even http/1.1 is unclear on how to mutate the method |
|
|
585 | $method = "GET" unless $method eq "HEAD"; |
|
|
586 | $redirect = 1; |
|
|
587 | } elsif ($status == 307 && $method =~ /^(?:GET|HEAD)$/) { |
|
|
588 | $redirect = 1; |
|
|
589 | } |
|
|
590 | } |
|
|
591 | |
430 | my $finish = sub { |
592 | my $finish = sub { |
431 | $state{handle}->destroy; |
593 | $state{handle}->destroy if $state{handle}; |
432 | %state = (); |
594 | %state = (); |
433 | |
595 | |
434 | # set-cookie processing |
596 | # set-cookie processing |
435 | if ($arg{cookie_jar}) { |
597 | if ($arg{cookie_jar}) { |
436 | for ($hdr{"set-cookie"}) { |
598 | for ($_[1]{"set-cookie"}) { |
437 | # parse NAME=VALUE |
599 | # parse NAME=VALUE |
438 | my @kv; |
600 | my @kv; |
439 | |
601 | |
440 | while (/\G\s* ([^=;,[:space:]]+) \s*=\s* (?: "((?:[^\\"]+|\\.)*)" | ([^=;,[:space:]]*) )/gcxs) { |
602 | while (/\G\s* ([^=;,[:space:]]+) \s*=\s* (?: "((?:[^\\"]+|\\.)*)" | ([^=;,[:space:]]*) )/gcxs) { |
441 | my $name = $1; |
603 | my $name = $1; |
… | |
… | |
479 | |
641 | |
480 | redo if /\G\s*,/gc; |
642 | redo if /\G\s*,/gc; |
481 | } |
643 | } |
482 | } |
644 | } |
483 | |
645 | |
484 | # microsoft and other shitheads don't give a shit for following standards, |
646 | if ($redirect && exists $hdr{location}) { |
485 | # try to support some common forms of broken Location headers. |
647 | # we ignore any errors, as it is very common to receive |
486 | if ($_[1]{location} !~ /^(?: $ | [^:\/?\#]+ : )/x) { |
648 | # Content-Length != 0 but no actual body |
487 | $_[1]{location} =~ s/^\.\/+//; |
649 | # we also access %hdr, as $_[1] might be an erro |
488 | |
650 | http_request ( |
489 | my $url = "$rscheme://$uhost:$uport"; |
651 | $method => $hdr{location}, |
490 | |
|
|
491 | unless ($_[1]{location} =~ s/^\///) { |
|
|
492 | $url .= $upath; |
|
|
493 | $url =~ s/\/[^\/]*$//; |
|
|
494 | } |
652 | %arg, |
495 | |
653 | recurse => $recurse - 1, |
496 | $_[1]{location} = "$url/$_[1]{location}"; |
654 | Redirect => \@_, |
497 | } |
655 | $cb); |
498 | |
|
|
499 | if ($_[1]{Status} =~ /^30[12]$/ && $recurse && $method ne "POST") { |
|
|
500 | # apparently, mozilla et al. just change POST to GET here |
|
|
501 | # more research is needed before we do the same |
|
|
502 | http_request ($method, $_[1]{location}, %arg, recurse => $recurse - 1, $cb); |
|
|
503 | } elsif ($_[1]{Status} == 303 && $recurse) { |
|
|
504 | # even http/1.1 is unclear on how to mutate the method |
|
|
505 | $method = "GET" unless $method eq "HEAD"; |
|
|
506 | http_request ($method => $_[1]{location}, %arg, recurse => $recurse - 1, $cb); |
|
|
507 | } elsif ($_[1]{Status} == 307 && $recurse && $method =~ /^(?:GET|HEAD)$/) { |
|
|
508 | http_request ($method => $_[1]{location}, %arg, recurse => $recurse - 1, $cb); |
|
|
509 | } else { |
656 | } else { |
510 | $cb->($_[0], $_[1]); |
657 | $cb->($_[0], $_[1]); |
511 | } |
658 | } |
512 | }; |
659 | }; |
513 | |
660 | |
514 | if ($hdr{Status} =~ /^(?:1..|204|304)$/ or $method eq "HEAD") { |
661 | my $len = $hdr{"content-length"}; |
|
|
662 | |
|
|
663 | if (!$redirect && $arg{on_header} && !$arg{on_header}(\%hdr)) { |
|
|
664 | $finish->(undef, { Status => 598, Reason => "Request cancelled by on_header", @pseudo }); |
|
|
665 | } elsif ( |
|
|
666 | $hdr{Status} =~ /^(?:1..|[23]04)$/ |
|
|
667 | or $method eq "HEAD" |
|
|
668 | or (defined $len && !$len) |
|
|
669 | ) { |
|
|
670 | # no body |
515 | $finish->(undef, \%hdr); |
671 | $finish->("", \%hdr); |
516 | } else { |
672 | } else { |
517 | if (exists $hdr{"content-length"}) { |
673 | # body handling, four different code paths |
518 | $_[0]->unshift_read (chunk => $hdr{"content-length"}, sub { |
674 | # for want_body_handle, on_body (2x), normal (2x) |
519 | # could cache persistent connection now |
675 | # we might read too much here, but it does not matter yet (no pers. connections) |
520 | if ($hdr{connection} =~ /\bkeep-alive\b/i) { |
676 | if (!$redirect && $arg{want_body_handle}) { |
521 | # but we don't, due to misdesigns, this is annoyingly complex |
677 | $_[0]->on_eof (undef); |
|
|
678 | $_[0]->on_error (undef); |
|
|
679 | $_[0]->on_read (undef); |
|
|
680 | |
|
|
681 | $finish->(delete $state{handle}, \%hdr); |
|
|
682 | |
|
|
683 | } elsif ($arg{on_body}) { |
|
|
684 | $_[0]->on_error (sub { $finish->(undef, { Status => 599, Reason => $_[2], @pseudo }) }); |
|
|
685 | if ($len) { |
|
|
686 | $_[0]->on_eof (undef); |
|
|
687 | $_[0]->on_read (sub { |
|
|
688 | $len -= length $_[0]{rbuf}; |
|
|
689 | |
|
|
690 | $arg{on_body}(delete $_[0]{rbuf}, \%hdr) |
|
|
691 | or $finish->(undef, { Status => 598, Reason => "Request cancelled by on_body", @pseudo }); |
|
|
692 | |
|
|
693 | $len > 0 |
|
|
694 | or $finish->("", \%hdr); |
522 | }; |
695 | }); |
523 | |
696 | } else { |
|
|
697 | $_[0]->on_eof (sub { |
524 | $finish->($_[1], \%hdr); |
698 | $finish->("", \%hdr); |
|
|
699 | }); |
|
|
700 | $_[0]->on_read (sub { |
|
|
701 | $arg{on_body}(delete $_[0]{rbuf}, \%hdr) |
|
|
702 | or $finish->(undef, { Status => 598, Reason => "Request cancelled by on_body", @pseudo }); |
|
|
703 | }); |
525 | }); |
704 | } |
526 | } else { |
705 | } else { |
527 | # too bad, need to read until we get an error or EOF, |
|
|
528 | # no way to detect winged data. |
|
|
529 | $_[0]->on_error (sub { |
|
|
530 | $finish->($_[0]{rbuf}, \%hdr); |
|
|
531 | }); |
|
|
532 | $_[0]->on_eof (undef); |
706 | $_[0]->on_eof (undef); |
|
|
707 | |
|
|
708 | if ($len) { |
|
|
709 | $_[0]->on_error (sub { $finish->(undef, { Status => 599, Reason => $_[2], @pseudo }) }); |
|
|
710 | $_[0]->on_read (sub { |
|
|
711 | $finish->((substr delete $_[0]{rbuf}, 0, $len, ""), \%hdr) |
|
|
712 | if $len <= length $_[0]{rbuf}; |
|
|
713 | }); |
|
|
714 | } else { |
|
|
715 | $_[0]->on_error (sub { |
|
|
716 | $! == Errno::EPIPE || !$! |
|
|
717 | ? $finish->(delete $_[0]{rbuf}, \%hdr) |
|
|
718 | : $finish->(undef, { Status => 599, Reason => $_[2], @pseudo }); |
|
|
719 | }); |
533 | $_[0]->on_read (sub { }); |
720 | $_[0]->on_read (sub { }); |
|
|
721 | } |
534 | } |
722 | } |
535 | } |
723 | } |
536 | }); |
724 | }); |
537 | }); |
725 | }); |
538 | }; |
726 | }; |
… | |
… | |
543 | |
731 | |
544 | # maybe re-use $uauthority with patched port? |
732 | # maybe re-use $uauthority with patched port? |
545 | $state{handle}->push_write ("CONNECT $uhost:$uport HTTP/1.0\015\012Host: $uhost\015\012\015\012"); |
733 | $state{handle}->push_write ("CONNECT $uhost:$uport HTTP/1.0\015\012Host: $uhost\015\012\015\012"); |
546 | $state{handle}->push_read (line => $qr_nlnl, sub { |
734 | $state{handle}->push_read (line => $qr_nlnl, sub { |
547 | $_[1] =~ /^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\015\012]*) )?/ix |
735 | $_[1] =~ /^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\015\012]*) )?/ix |
548 | or return (%state = (), $cb->(undef, { Status => 599, Reason => "invalid proxy connect response ($_[1])", URL => $url })); |
736 | or return (%state = (), $cb->(undef, { Status => 599, Reason => "Invalid proxy connect response ($_[1])", @pseudo })); |
549 | |
737 | |
550 | if ($2 == 200) { |
738 | if ($2 == 200) { |
551 | $rpath = $upath; |
739 | $rpath = $upath; |
552 | &$handle_actual_request; |
740 | &$handle_actual_request; |
553 | } else { |
741 | } else { |
554 | %state = (); |
742 | %state = (); |
555 | $cb->(undef, { Status => $2, Reason => $3, URL => $url }); |
743 | $cb->(undef, { Status => $2, Reason => $3, @pseudo }); |
556 | } |
744 | } |
557 | }); |
745 | }); |
558 | } else { |
746 | } else { |
559 | &$handle_actual_request; |
747 | &$handle_actual_request; |
560 | } |
748 | } |
561 | |
749 | |
562 | }, sub { |
750 | }, $arg{on_prepare} || sub { $timeout }; |
563 | $timeout |
|
|
564 | }; |
|
|
565 | }; |
751 | }; |
566 | |
752 | |
567 | defined wantarray && AnyEvent::Util::guard { %state = () } |
753 | defined wantarray && AnyEvent::Util::guard { %state = () } |
568 | } |
754 | } |
569 | |
755 | |
… | |
… | |
583 | &http_request |
769 | &http_request |
584 | } |
770 | } |
585 | |
771 | |
586 | =back |
772 | =back |
587 | |
773 | |
|
|
774 | =head2 DNS CACHING |
|
|
775 | |
|
|
776 | AnyEvent::HTTP uses the AnyEvent::Socket::tcp_connect function for |
|
|
777 | the actual connection, which in turn uses AnyEvent::DNS to resolve |
|
|
778 | hostnames. The latter is a simple stub resolver and does no caching |
|
|
779 | on its own. If you want DNS caching, you currently have to provide |
|
|
780 | your own default resolver (by storing a suitable resolver object in |
|
|
781 | C<$AnyEvent::DNS::RESOLVER>). |
|
|
782 | |
588 | =head2 GLOBAL FUNCTIONS AND VARIABLES |
783 | =head2 GLOBAL FUNCTIONS AND VARIABLES |
589 | |
784 | |
590 | =over 4 |
785 | =over 4 |
591 | |
786 | |
592 | =item AnyEvent::HTTP::set_proxy "proxy-url" |
787 | =item AnyEvent::HTTP::set_proxy "proxy-url" |
593 | |
788 | |
594 | Sets the default proxy server to use. The proxy-url must begin with a |
789 | Sets the default proxy server to use. The proxy-url must begin with a |
595 | string of the form C<http://host:port> (optionally C<https:...>). |
790 | string of the form C<http://host:port> (optionally C<https:...>), croaks |
|
|
791 | otherwise. |
|
|
792 | |
|
|
793 | To clear an already-set proxy, use C<undef>. |
596 | |
794 | |
597 | =item $AnyEvent::HTTP::MAX_RECURSE |
795 | =item $AnyEvent::HTTP::MAX_RECURSE |
598 | |
796 | |
599 | The default value for the C<recurse> request parameter (default: C<10>). |
797 | The default value for the C<recurse> request parameter (default: C<10>). |
600 | |
798 | |
601 | =item $AnyEvent::HTTP::USERAGENT |
799 | =item $AnyEvent::HTTP::USERAGENT |
602 | |
800 | |
603 | The default value for the C<User-Agent> header (the default is |
801 | The default value for the C<User-Agent> header (the default is |
604 | C<Mozilla/5.0 (compatible; AnyEvent::HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)>). |
802 | C<Mozilla/5.0 (compatible; U; AnyEvent-HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)>). |
605 | |
803 | |
606 | =item $AnyEvent::HTTP::MAX_PERSISTENT |
804 | =item $AnyEvent::HTTP::MAX_PER_HOST |
607 | |
805 | |
608 | The maximum number of persistent connections to keep open (default: 8). |
806 | The maximum number of concurrent connections to the same host (identified |
|
|
807 | by the hostname). If the limit is exceeded, then the additional requests |
|
|
808 | are queued until previous connections are closed. |
609 | |
809 | |
610 | Not implemented currently. |
810 | The default value for this is C<4>, and it is highly advisable to not |
611 | |
811 | increase it. |
612 | =item $AnyEvent::HTTP::PERSISTENT_TIMEOUT |
|
|
613 | |
|
|
614 | The maximum time to cache a persistent connection, in seconds (default: 2). |
|
|
615 | |
|
|
616 | Not implemented currently. |
|
|
617 | |
812 | |
618 | =item $AnyEvent::HTTP::ACTIVE |
813 | =item $AnyEvent::HTTP::ACTIVE |
619 | |
814 | |
620 | The number of active connections. This is not the number of currently |
815 | The number of active connections. This is not the number of currently |
621 | running requests, but the number of currently open and non-idle TCP |
816 | running requests, but the number of currently open and non-idle TCP |
… | |
… | |
624 | =back |
819 | =back |
625 | |
820 | |
626 | =cut |
821 | =cut |
627 | |
822 | |
628 | sub set_proxy($) { |
823 | sub set_proxy($) { |
|
|
824 | if (length $_[0]) { |
629 | $PROXY = [$2, $3 || 3128, $1] if $_[0] =~ m%^(https?):// ([^:/]+) (?: : (\d*) )?%ix; |
825 | $_[0] =~ m%^(https?):// ([^:/]+) (?: : (\d*) )?%ix |
|
|
826 | or Carp::croak "$_[0]: invalid proxy URL"; |
|
|
827 | $PROXY = [$2, $3 || 3128, $1] |
|
|
828 | } else { |
|
|
829 | undef $PROXY; |
|
|
830 | } |
630 | } |
831 | } |
631 | |
832 | |
632 | # initialise proxy from environment |
833 | # initialise proxy from environment |
|
|
834 | eval { |
633 | set_proxy $ENV{http_proxy}; |
835 | set_proxy $ENV{http_proxy}; |
|
|
836 | }; |
634 | |
837 | |
635 | =head1 SEE ALSO |
838 | =head1 SEE ALSO |
636 | |
839 | |
637 | L<AnyEvent>. |
840 | L<AnyEvent>. |
638 | |
841 | |
639 | =head1 AUTHOR |
842 | =head1 AUTHOR |
640 | |
843 | |
641 | Marc Lehmann <schmorp@schmorp.de> |
844 | Marc Lehmann <schmorp@schmorp.de> |
642 | http://home.schmorp.de/ |
845 | http://home.schmorp.de/ |
643 | |
846 | |
|
|
847 | With many thanks to Дмитрий Шалашов, who provided countless |
|
|
848 | testcases and bugreports. |
|
|
849 | |
644 | =cut |
850 | =cut |
645 | |
851 | |
646 | 1 |
852 | 1 |
647 | |
853 | |