… | |
… | |
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 (); |
|
|
49 | use AnyEvent::Handle (); |
48 | use AnyEvent::Handle (); |
50 | |
49 | |
51 | use base Exporter::; |
50 | use base Exporter::; |
52 | |
51 | |
53 | our $VERSION = '1.12'; |
52 | our $VERSION = '1.5'; |
54 | |
53 | |
55 | our @EXPORT = qw(http_get http_post http_head http_request); |
54 | our @EXPORT = qw(http_get http_post http_head http_request); |
56 | |
55 | |
57 | our $USERAGENT = "Mozilla/5.0 (compatible; AnyEvent::HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)"; |
56 | our $USERAGENT = "Mozilla/5.0 (compatible; U; AnyEvent-HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)"; |
58 | our $MAX_RECURSE = 10; |
57 | our $MAX_RECURSE = 10; |
59 | our $MAX_PERSISTENT = 8; |
58 | our $MAX_PERSISTENT = 8; |
60 | our $PERSISTENT_TIMEOUT = 2; |
59 | our $PERSISTENT_TIMEOUT = 2; |
61 | our $TIMEOUT = 300; |
60 | our $TIMEOUT = 300; |
62 | |
61 | |
63 | # changing these is evil |
62 | # changing these is evil |
64 | our $MAX_PERSISTENT_PER_HOST = 2; |
63 | our $MAX_PERSISTENT_PER_HOST = 0; |
65 | our $MAX_PER_HOST = 4; |
64 | our $MAX_PER_HOST = 4; |
66 | |
65 | |
67 | our $PROXY; |
66 | our $PROXY; |
68 | our $ACTIVE = 0; |
67 | our $ACTIVE = 0; |
69 | |
68 | |
… | |
… | |
92 | must be an absolute http or https URL. |
91 | must be an absolute http or https URL. |
93 | |
92 | |
94 | When called in void context, nothing is returned. In other contexts, |
93 | When called in void context, nothing is returned. In other contexts, |
95 | C<http_request> returns a "cancellation guard" - you have to keep the |
94 | 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 |
95 | 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. |
96 | destroyed before the callback is called, the request will be cancelled. |
98 | |
97 | |
99 | The callback will be called with the response data as first argument |
98 | 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 |
99 | (or C<undef> if an error occured), and a hash-ref with response headers |
101 | response headers as second argument. |
100 | (and trailers) as second argument. |
102 | |
101 | |
103 | All the headers in that hash are lowercased. In addition to the response |
102 | 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> |
103 | headers, the "pseudo-headers" (uppercase to avoid clashing with possible |
|
|
104 | 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 |
105 | three parts of the HTTP Status-Line of the same name. If an error occurs |
|
|
106 | during the body phase of a request, then the original C<Status> and |
|
|
107 | C<Reason> values from the header are available as C<OrigStatus> and |
|
|
108 | C<OrigReason>. |
|
|
109 | |
106 | pseudo-header C<URL> contains the original URL (which can differ from the |
110 | The pseudo-header C<URL> contains the actual URL (which can differ from |
107 | requested URL when following redirects). |
111 | the requested URL when following redirects - for example, you might get |
|
|
112 | an error that your URL scheme is not supported even though your URL is a |
|
|
113 | valid http URL because it redirected to an ftp URL, in which case you can |
|
|
114 | look at the URL pseudo header). |
|
|
115 | |
|
|
116 | The pseudo-header C<Redirect> only exists when the request was a result |
|
|
117 | of an internal redirect. In that case it is an array reference with |
|
|
118 | the C<($data, $headers)> from the redirect response. Note that this |
|
|
119 | response could in turn be the result of a redirect itself, and C<< |
|
|
120 | $headers->{Redirect}[1]{Redirect} >> will then contain the original |
|
|
121 | response, and so on. |
108 | |
122 | |
109 | If the server sends a header multiple times, then their contents will be |
123 | If the server sends a header multiple times, then their contents will be |
110 | joined together with a comma (C<,>), as per the HTTP spec. |
124 | joined together with a comma (C<,>), as per the HTTP spec. |
111 | |
125 | |
112 | If an internal error occurs, such as not being able to resolve a hostname, |
126 | If an internal error occurs, such as not being able to resolve a hostname, |
113 | then C<$data> will be C<undef>, C<< $headers->{Status} >> will be C<599> |
127 | then C<$data> will be C<undef>, C<< $headers->{Status} >> will be C<59x> |
114 | and the C<Reason> pseudo-header will contain an error message. |
128 | (usually C<599>) and the C<Reason> pseudo-header will contain an error |
|
|
129 | message. |
115 | |
130 | |
116 | A typical callback might look like this: |
131 | A typical callback might look like this: |
117 | |
132 | |
118 | sub { |
133 | sub { |
119 | my ($body, $hdr) = @_; |
134 | my ($body, $hdr) = @_; |
… | |
… | |
135 | Whether to recurse requests or not, e.g. on redirects, authentication |
150 | Whether to recurse requests or not, e.g. on redirects, authentication |
136 | retries and so on, and how often to do so. |
151 | retries and so on, and how often to do so. |
137 | |
152 | |
138 | =item headers => hashref |
153 | =item headers => hashref |
139 | |
154 | |
140 | The request headers to use. Currently, C<http_request> may provide its |
155 | The request headers to use. Currently, C<http_request> may provide its own |
141 | own C<Host:>, C<Content-Length:>, C<Connection:> and C<Cookie:> headers |
156 | C<Host:>, C<Content-Length:>, C<Connection:> and C<Cookie:> headers and |
142 | and will provide defaults for C<User-Agent:> and C<Referer:>. |
157 | will provide defaults at least for C<TE:>, C<Referer:> and C<User-Agent:> |
|
|
158 | (this can be suppressed by using C<undef> for these headers in which case |
|
|
159 | they won't be sent at all). |
143 | |
160 | |
144 | =item timeout => $seconds |
161 | =item timeout => $seconds |
145 | |
162 | |
146 | The time-out to use for various stages - each connect attempt will reset |
163 | The time-out to use for various stages - each connect attempt will reset |
147 | the timeout, as will read or write activity. Default timeout is 5 minutes. |
164 | the timeout, as will read or write activity, i.e. this is not an overall |
|
|
165 | timeout. |
|
|
166 | |
|
|
167 | Default timeout is 5 minutes. |
148 | |
168 | |
149 | =item proxy => [$host, $port[, $scheme]] or undef |
169 | =item proxy => [$host, $port[, $scheme]] or undef |
150 | |
170 | |
151 | Use the given http proxy for all requests. If not specified, then the |
171 | Use the given http proxy for all requests. If not specified, then the |
152 | default proxy (as specified by C<$ENV{http_proxy}>) is used. |
172 | default proxy (as specified by C<$ENV{http_proxy}>) is used. |
153 | |
173 | |
154 | C<$scheme> must be either missing or C<http> for HTTP, or C<https> for |
174 | C<$scheme> must be either missing, C<http> for HTTP or C<https> for |
155 | HTTPS. |
175 | HTTPS. |
156 | |
176 | |
157 | =item body => $string |
177 | =item body => $string |
158 | |
178 | |
159 | The request body, usually empty. Will be-sent as-is (future versions of |
179 | The request body, usually empty. Will be sent as-is (future versions of |
160 | this module might offer more options). |
180 | this module might offer more options). |
161 | |
181 | |
162 | =item cookie_jar => $hash_ref |
182 | =item cookie_jar => $hash_ref |
163 | |
183 | |
164 | Passing this parameter enables (simplified) cookie-processing, loosely |
184 | Passing this parameter enables (simplified) cookie-processing, loosely |
165 | based on the original netscape specification. |
185 | based on the original netscape specification. |
166 | |
186 | |
167 | The C<$hash_ref> must be an (initially empty) hash reference which will |
187 | 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 |
188 | 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 |
189 | persistent storage with something like JSON or Storable, but this is not |
170 | recommended, as expire times are currently being ignored. |
190 | recommended, as session-only cookies might survive longer than expected. |
171 | |
191 | |
172 | Note that this cookie implementation is not of very high quality, nor |
192 | Note that this cookie implementation is not meant to be complete. If |
173 | meant to be complete. If you want complete cookie management you have to |
193 | you want complete cookie management you have to do that on your |
174 | do that on your own. C<cookie_jar> is meant as a quick fix to get some |
194 | own. C<cookie_jar> is meant as a quick fix to get some cookie-using sites |
175 | cookie-using sites working. Cookies are a privacy disaster, do not use |
195 | working. Cookies are a privacy disaster, do not use them unless required |
176 | them unless required to. |
196 | to. |
|
|
197 | |
|
|
198 | When cookie processing is enabled, the C<Cookie:> and C<Set-Cookie:> |
|
|
199 | headers will be set and handled by this module, otherwise they will be |
|
|
200 | left untouched. |
|
|
201 | |
|
|
202 | =item tls_ctx => $scheme | $tls_ctx |
|
|
203 | |
|
|
204 | Specifies the AnyEvent::TLS context to be used for https connections. This |
|
|
205 | parameter follows the same rules as the C<tls_ctx> parameter to |
|
|
206 | L<AnyEvent::Handle>, but additionally, the two strings C<low> or |
|
|
207 | C<high> can be specified, which give you a predefined low-security (no |
|
|
208 | verification, highest compatibility) and high-security (CA and common-name |
|
|
209 | verification) TLS context. |
|
|
210 | |
|
|
211 | The default for this option is C<low>, which could be interpreted as "give |
|
|
212 | me the page, no matter what". |
|
|
213 | |
|
|
214 | =item on_prepare => $callback->($fh) |
|
|
215 | |
|
|
216 | In rare cases you need to "tune" the socket before it is used to |
|
|
217 | connect (for exmaple, to bind it on a given IP address). This parameter |
|
|
218 | overrides the prepare callback passed to C<AnyEvent::Socket::tcp_connect> |
|
|
219 | and behaves exactly the same way (e.g. it has to provide a |
|
|
220 | timeout). See the description for the C<$prepare_cb> argument of |
|
|
221 | C<AnyEvent::Socket::tcp_connect> for details. |
|
|
222 | |
|
|
223 | =item tcp_connect => $callback->($host, $service, $connect_cb, $prepare_cb) |
|
|
224 | |
|
|
225 | In even rarer cases you want total control over how AnyEvent::HTTP |
|
|
226 | establishes connections. Normally it uses L<AnyEvent::Socket::tcp_connect> |
|
|
227 | to do this, but you can provide your own C<tcp_connect> function - |
|
|
228 | obviously, it has to follow the same calling conventions, except that it |
|
|
229 | may always return a connection guard object. |
|
|
230 | |
|
|
231 | There are probably lots of weird uses for this function, starting from |
|
|
232 | tracing the hosts C<http_request> actually tries to connect, to (inexact |
|
|
233 | but fast) host => IP address caching or even socks protocol support. |
|
|
234 | |
|
|
235 | =item on_header => $callback->($headers) |
|
|
236 | |
|
|
237 | When specified, this callback will be called with the header hash as soon |
|
|
238 | as headers have been successfully received from the remote server (not on |
|
|
239 | locally-generated errors). |
|
|
240 | |
|
|
241 | It has to return either true (in which case AnyEvent::HTTP will continue), |
|
|
242 | or false, in which case AnyEvent::HTTP will cancel the download (and call |
|
|
243 | the finish callback with an error code of C<598>). |
|
|
244 | |
|
|
245 | This callback is useful, among other things, to quickly reject unwanted |
|
|
246 | content, which, if it is supposed to be rare, can be faster than first |
|
|
247 | doing a C<HEAD> request. |
|
|
248 | |
|
|
249 | The downside is that cancelling the request makes it impossible to re-use |
|
|
250 | the connection. Also, the C<on_header> callback will not receive any |
|
|
251 | trailer (headers sent after the response body). |
|
|
252 | |
|
|
253 | Example: cancel the request unless the content-type is "text/html". |
|
|
254 | |
|
|
255 | on_header => sub { |
|
|
256 | $_[0]{"content-type"} =~ /^text\/html\s*(?:;|$)/ |
|
|
257 | }, |
|
|
258 | |
|
|
259 | =item on_body => $callback->($partial_body, $headers) |
|
|
260 | |
|
|
261 | When specified, all body data will be passed to this callback instead of |
|
|
262 | to the completion callback. The completion callback will get the empty |
|
|
263 | string instead of the body data. |
|
|
264 | |
|
|
265 | It has to return either true (in which case AnyEvent::HTTP will continue), |
|
|
266 | or false, in which case AnyEvent::HTTP will cancel the download (and call |
|
|
267 | the completion callback with an error code of C<598>). |
|
|
268 | |
|
|
269 | The downside to cancelling the request is that it makes it impossible to |
|
|
270 | re-use the connection. |
|
|
271 | |
|
|
272 | This callback is useful when the data is too large to be held in memory |
|
|
273 | (so the callback writes it to a file) or when only some information should |
|
|
274 | be extracted, or when the body should be processed incrementally. |
|
|
275 | |
|
|
276 | It is usually preferred over doing your own body handling via |
|
|
277 | C<want_body_handle>, but in case of streaming APIs, where HTTP is |
|
|
278 | only used to create a connection, C<want_body_handle> is the better |
|
|
279 | alternative, as it allows you to install your own event handler, reducing |
|
|
280 | resource usage. |
|
|
281 | |
|
|
282 | =item want_body_handle => $enable |
|
|
283 | |
|
|
284 | When enabled (default is disabled), the behaviour of AnyEvent::HTTP |
|
|
285 | changes considerably: after parsing the headers, and instead of |
|
|
286 | downloading the body (if any), the completion callback will be |
|
|
287 | called. Instead of the C<$body> argument containing the body data, the |
|
|
288 | callback will receive the L<AnyEvent::Handle> object associated with the |
|
|
289 | connection. In error cases, C<undef> will be passed. When there is no body |
|
|
290 | (e.g. status C<304>), the empty string will be passed. |
|
|
291 | |
|
|
292 | The handle object might or might not be in TLS mode, might be connected to |
|
|
293 | a proxy, be a persistent connection etc., and configured in unspecified |
|
|
294 | ways. The user is responsible for this handle (it will not be used by this |
|
|
295 | module anymore). |
|
|
296 | |
|
|
297 | This is useful with some push-type services, where, after the initial |
|
|
298 | headers, an interactive protocol is used (typical example would be the |
|
|
299 | push-style twitter API which starts a JSON/XML stream). |
|
|
300 | |
|
|
301 | If you think you need this, first have a look at C<on_body>, to see if |
|
|
302 | that doesn't solve your problem in a better way. |
177 | |
303 | |
178 | =back |
304 | =back |
179 | |
305 | |
180 | Example: make a simple HTTP GET request for http://www.nethype.de/ |
306 | Example: do a simple HTTP GET request for http://www.nethype.de/ and print |
|
|
307 | the response body. |
181 | |
308 | |
182 | http_request GET => "http://www.nethype.de/", sub { |
309 | http_request GET => "http://www.nethype.de/", sub { |
183 | my ($body, $hdr) = @_; |
310 | my ($body, $hdr) = @_; |
184 | print "$body\n"; |
311 | print "$body\n"; |
185 | }; |
312 | }; |
186 | |
313 | |
187 | Example: make a HTTP HEAD request on https://www.google.com/, use a |
314 | Example: do a HTTP HEAD request on https://www.google.com/, use a |
188 | timeout of 30 seconds. |
315 | timeout of 30 seconds. |
189 | |
316 | |
190 | http_request |
317 | http_request |
191 | GET => "https://www.google.com", |
318 | GET => "https://www.google.com", |
192 | timeout => 30, |
319 | timeout => 30, |
… | |
… | |
195 | use Data::Dumper; |
322 | use Data::Dumper; |
196 | print Dumper $hdr; |
323 | print Dumper $hdr; |
197 | } |
324 | } |
198 | ; |
325 | ; |
199 | |
326 | |
200 | Example: make another simple HTTP GET request, but immediately try to |
327 | Example: do another simple HTTP GET request, but immediately try to |
201 | cancel it. |
328 | cancel it. |
202 | |
329 | |
203 | my $request = http_request GET => "http://www.nethype.de/", sub { |
330 | my $request = http_request GET => "http://www.nethype.de/", sub { |
204 | my ($body, $hdr) = @_; |
331 | my ($body, $hdr) = @_; |
205 | print "$body\n"; |
332 | print "$body\n"; |
… | |
… | |
237 | push @{ $CO_SLOT{$_[0]}[1] }, $_[1]; |
364 | push @{ $CO_SLOT{$_[0]}[1] }, $_[1]; |
238 | |
365 | |
239 | _slot_schedule $_[0]; |
366 | _slot_schedule $_[0]; |
240 | } |
367 | } |
241 | |
368 | |
242 | our $qr_nl = qr<\015?\012>; |
369 | # continue to parse $_ for headers and place them into the arg |
243 | our $qr_nlnl = qr<\015?\012\015?\012>; |
370 | sub parse_hdr() { |
|
|
371 | my %hdr; |
|
|
372 | |
|
|
373 | # things seen, not parsed: |
|
|
374 | # p3pP="NON CUR OTPi OUR NOR UNI" |
|
|
375 | |
|
|
376 | $hdr{lc $1} .= ",$2" |
|
|
377 | while /\G |
|
|
378 | ([^:\000-\037]*): |
|
|
379 | [\011\040]* |
|
|
380 | ((?: [^\012]+ | \012[\011\040] )*) |
|
|
381 | \012 |
|
|
382 | /gxc; |
|
|
383 | |
|
|
384 | /\G$/ |
|
|
385 | or return; |
|
|
386 | |
|
|
387 | # remove the "," prefix we added to all headers above |
|
|
388 | substr $_, 0, 1, "" |
|
|
389 | for values %hdr; |
|
|
390 | |
|
|
391 | \%hdr |
|
|
392 | } |
|
|
393 | |
|
|
394 | our $qr_nlnl = qr{(?<![^\012])\015?\012}; |
|
|
395 | |
|
|
396 | our $TLS_CTX_LOW = { cache => 1, sslv2 => 1 }; |
|
|
397 | our $TLS_CTX_HIGH = { cache => 1, verify => 1, verify_peername => "https" }; |
244 | |
398 | |
245 | sub http_request($$@) { |
399 | sub http_request($$@) { |
246 | my $cb = pop; |
400 | my $cb = pop; |
247 | my ($method, $url, %arg) = @_; |
401 | my ($method, $url, %arg) = @_; |
248 | |
402 | |
249 | my %hdr; |
403 | my %hdr; |
|
|
404 | |
|
|
405 | $arg{tls_ctx} = $TLS_CTX_LOW if $arg{tls_ctx} eq "low" || !exists $arg{tls_ctx}; |
|
|
406 | $arg{tls_ctx} = $TLS_CTX_HIGH if $arg{tls_ctx} eq "high"; |
250 | |
407 | |
251 | $method = uc $method; |
408 | $method = uc $method; |
252 | |
409 | |
253 | if (my $hdr = $arg{headers}) { |
410 | if (my $hdr = $arg{headers}) { |
254 | while (my ($k, $v) = each %$hdr) { |
411 | while (my ($k, $v) = each %$hdr) { |
255 | $hdr{lc $k} = $v; |
412 | $hdr{lc $k} = $v; |
256 | } |
413 | } |
257 | } |
414 | } |
258 | |
415 | |
|
|
416 | # pseudo headers for all subsequent responses |
|
|
417 | my @pseudo = (URL => $url); |
|
|
418 | push @pseudo, Redirect => delete $arg{Redirect} if exists $arg{Redirect}; |
|
|
419 | |
259 | my $recurse = exists $arg{recurse} ? delete $arg{recurse} : $MAX_RECURSE; |
420 | my $recurse = exists $arg{recurse} ? delete $arg{recurse} : $MAX_RECURSE; |
260 | |
421 | |
261 | return $cb->(undef, { Status => 599, Reason => "recursion limit reached", URL => $url }) |
422 | return $cb->(undef, { @pseudo, Status => 599, Reason => "Too many redirections" }) |
262 | if $recurse < 0; |
423 | if $recurse < 0; |
263 | |
424 | |
264 | my $proxy = $arg{proxy} || $PROXY; |
425 | my $proxy = $arg{proxy} || $PROXY; |
265 | my $timeout = $arg{timeout} || $TIMEOUT; |
426 | my $timeout = $arg{timeout} || $TIMEOUT; |
266 | |
427 | |
267 | $hdr{"user-agent"} ||= $USERAGENT; |
|
|
268 | |
|
|
269 | my ($uscheme, $uauthority, $upath, $query, $fragment) = |
428 | my ($uscheme, $uauthority, $upath, $query, $fragment) = |
270 | $url =~ m|(?:([^:/?#]+):)?(?://([^/?#]*))?([^?#]*)(?:\?([^#]*))?(?:#(.*))?|; |
429 | $url =~ m|(?:([^:/?#]+):)?(?://([^/?#]*))?([^?#]*)(?:(\?[^#]*))?(?:#(.*))?|; |
271 | |
430 | |
272 | $uscheme = lc $uscheme; |
431 | $uscheme = lc $uscheme; |
273 | |
432 | |
274 | my $uport = $uscheme eq "http" ? 80 |
433 | my $uport = $uscheme eq "http" ? 80 |
275 | : $uscheme eq "https" ? 443 |
434 | : $uscheme eq "https" ? 443 |
276 | : return $cb->(undef, { Status => 599, Reason => "only http and https URL schemes supported", URL => $url }); |
435 | : return $cb->(undef, { @pseudo, Status => 599, Reason => "Only http and https URL schemes supported" }); |
277 | |
|
|
278 | $hdr{referer} ||= "$uscheme://$uauthority$upath"; # leave out fragment and query string, just a heuristic |
|
|
279 | |
436 | |
280 | $uauthority =~ /^(?: .*\@ )? ([^\@:]+) (?: : (\d+) )?$/x |
437 | $uauthority =~ /^(?: .*\@ )? ([^\@:]+) (?: : (\d+) )?$/x |
281 | or return $cb->(undef, { Status => 599, Reason => "unparsable URL", URL => $url }); |
438 | or return $cb->(undef, { @pseudo, Status => 599, Reason => "Unparsable URL" }); |
282 | |
439 | |
283 | my $uhost = $1; |
440 | my $uhost = $1; |
284 | $uport = $2 if defined $2; |
441 | $uport = $2 if defined $2; |
285 | |
442 | |
|
|
443 | $hdr{host} = defined $2 ? "$uhost:$2" : "$uhost" |
|
|
444 | unless exists $hdr{host}; |
|
|
445 | |
286 | $uhost =~ s/^\[(.*)\]$/$1/; |
446 | $uhost =~ s/^\[(.*)\]$/$1/; |
287 | $upath .= "?$query" if length $query; |
447 | $upath .= $query if length $query; |
288 | |
448 | |
289 | $upath =~ s%^/?%/%; |
449 | $upath =~ s%^/?%/%; |
290 | |
450 | |
291 | # cookie processing |
451 | # cookie processing |
292 | if (my $jar = $arg{cookie_jar}) { |
452 | if (my $jar = $arg{cookie_jar}) { |
293 | %$jar = () if $jar->{version} != 1; |
453 | %$jar = () if $jar->{version} != 1; |
294 | |
454 | |
295 | my @cookie; |
455 | my @cookie; |
296 | |
456 | |
297 | while (my ($chost, $v) = each %$jar) { |
457 | while (my ($chost, $paths) = each %$jar) { |
298 | if ($chost =~ /^\./) { |
458 | if ($chost =~ /^\./) { |
299 | next unless $chost eq substr $uhost, -length $chost; |
459 | next unless $chost eq substr $uhost, -length $chost; |
300 | } elsif ($chost =~ /\./) { |
460 | } elsif ($chost =~ /\./) { |
301 | next unless $chost eq $uhost; |
461 | next unless $chost eq $uhost; |
302 | } else { |
462 | } else { |
303 | next; |
463 | next; |
304 | } |
464 | } |
305 | |
465 | |
306 | while (my ($cpath, $v) = each %$v) { |
466 | while (my ($cpath, $cookies) = each %$paths) { |
307 | next unless $cpath eq substr $upath, 0, length $cpath; |
467 | next unless $cpath eq substr $upath, 0, length $cpath; |
308 | |
468 | |
309 | while (my ($k, $v) = each %$v) { |
469 | while (my ($cookie, $kv) = each %$cookies) { |
310 | next if $uscheme ne "https" && exists $v->{secure}; |
470 | next if $uscheme ne "https" && exists $kv->{secure}; |
|
|
471 | |
|
|
472 | if (exists $kv->{expires}) { |
|
|
473 | if (AE::now > parse_date ($kv->{expires})) { |
|
|
474 | delete $cookies->{$cookie}; |
|
|
475 | next; |
|
|
476 | } |
|
|
477 | } |
|
|
478 | |
311 | my $value = $v->{value}; |
479 | my $value = $kv->{value}; |
312 | $value =~ s/([\\"])/\\$1/g; |
480 | $value =~ s/([\\"])/\\$1/g; |
313 | push @cookie, "$k=\"$value\""; |
481 | push @cookie, "$cookie=\"$value\""; |
314 | } |
482 | } |
315 | } |
483 | } |
316 | } |
484 | } |
317 | |
485 | |
318 | $hdr{cookie} = join "; ", @cookie |
486 | $hdr{cookie} = join "; ", @cookie |
… | |
… | |
322 | my ($rhost, $rport, $rscheme, $rpath); # request host, port, path |
490 | my ($rhost, $rport, $rscheme, $rpath); # request host, port, path |
323 | |
491 | |
324 | if ($proxy) { |
492 | if ($proxy) { |
325 | ($rpath, $rhost, $rport, $rscheme) = ($url, @$proxy); |
493 | ($rpath, $rhost, $rport, $rscheme) = ($url, @$proxy); |
326 | |
494 | |
|
|
495 | $rscheme = "http" unless defined $rscheme; |
|
|
496 | |
327 | # don't support https requests over https-proxy transport, |
497 | # don't support https requests over https-proxy transport, |
328 | # can't be done with tls as spec'ed, unless you double-encrypt. |
498 | # can't be done with tls as spec'ed, unless you double-encrypt. |
329 | $rscheme = "http" if $uscheme eq "https" && $rscheme eq "https"; |
499 | $rscheme = "http" if $uscheme eq "https" && $rscheme eq "https"; |
330 | } else { |
500 | } else { |
331 | ($rhost, $rport, $rscheme, $rpath) = ($uhost, $uport, $uscheme, $upath); |
501 | ($rhost, $rport, $rscheme, $rpath) = ($uhost, $uport, $uscheme, $upath); |
332 | } |
502 | } |
333 | |
503 | |
334 | $hdr{host} = $uhost; |
504 | # leave out fragment and query string, just a heuristic |
|
|
505 | $hdr{referer} = "$uscheme://$uauthority$upath" unless exists $hdr{referer}; |
|
|
506 | $hdr{"user-agent"} = $USERAGENT unless exists $hdr{"user-agent"}; |
|
|
507 | |
335 | $hdr{"content-length"} = length $arg{body}; |
508 | $hdr{"content-length"} = length $arg{body} |
|
|
509 | if length $arg{body} || $method ne "GET"; |
|
|
510 | |
|
|
511 | $hdr{connection} = "close TE"; #1.1 |
|
|
512 | $hdr{te} = "trailers" unless exists $hdr{te}; #1.1 |
336 | |
513 | |
337 | my %state = (connect_guard => 1); |
514 | my %state = (connect_guard => 1); |
338 | |
515 | |
339 | _get_slot $uhost, sub { |
516 | _get_slot $uhost, sub { |
340 | $state{slot_guard} = shift; |
517 | $state{slot_guard} = shift; |
341 | |
518 | |
342 | return unless $state{connect_guard}; |
519 | return unless $state{connect_guard}; |
343 | |
520 | |
344 | $state{connect_guard} = AnyEvent::Socket::tcp_connect $rhost, $rport, sub { |
521 | my $connect_cb = sub { |
345 | $state{fh} = shift |
522 | $state{fh} = shift |
|
|
523 | or do { |
|
|
524 | my $err = "$!"; |
|
|
525 | %state = (); |
346 | or return $cb->(undef, { Status => 599, Reason => "$!", URL => $url }); |
526 | return $cb->(undef, { @pseudo, Status => 599, Reason => $err }); |
|
|
527 | }; |
|
|
528 | |
347 | pop; # free memory, save a tree |
529 | pop; # free memory, save a tree |
348 | |
530 | |
349 | return unless delete $state{connect_guard}; |
531 | return unless delete $state{connect_guard}; |
350 | |
532 | |
351 | # get handle |
533 | # get handle |
352 | $state{handle} = new AnyEvent::Handle |
534 | $state{handle} = new AnyEvent::Handle |
353 | fh => $state{fh}, |
535 | fh => $state{fh}, |
|
|
536 | peername => $rhost, |
|
|
537 | tls_ctx => $arg{tls_ctx}, |
|
|
538 | # these need to be reconfigured on keepalive handles |
354 | timeout => $timeout; |
539 | timeout => $timeout, |
|
|
540 | on_error => sub { |
|
|
541 | %state = (); |
|
|
542 | $cb->(undef, { @pseudo, Status => 599, Reason => $_[2] }); |
|
|
543 | }, |
|
|
544 | on_eof => sub { |
|
|
545 | %state = (); |
|
|
546 | $cb->(undef, { @pseudo, Status => 599, Reason => "Unexpected end-of-file" }); |
|
|
547 | }, |
|
|
548 | ; |
355 | |
549 | |
356 | # limit the number of persistent connections |
550 | # limit the number of persistent connections |
357 | # keepalive not yet supported |
551 | # keepalive not yet supported |
358 | if ($KA_COUNT{$_[1]} < $MAX_PERSISTENT_PER_HOST) { |
552 | # if ($KA_COUNT{$_[1]} < $MAX_PERSISTENT_PER_HOST) { |
359 | ++$KA_COUNT{$_[1]}; |
553 | # ++$KA_COUNT{$_[1]}; |
360 | $state{handle}{ka_count_guard} = AnyEvent::Util::guard { |
554 | # $state{handle}{ka_count_guard} = AnyEvent::Util::guard { |
361 | --$KA_COUNT{$_[1]} |
555 | # --$KA_COUNT{$_[1]} |
362 | }; |
556 | # }; |
363 | $hdr{connection} = "keep-alive"; |
557 | # $hdr{connection} = "keep-alive"; |
364 | } else { |
|
|
365 | delete $hdr{connection}; |
|
|
366 | } |
558 | # } |
367 | |
|
|
368 | # (re-)configure handle |
|
|
369 | $state{handle}->on_error (sub { |
|
|
370 | my $errno = "$!"; |
|
|
371 | %state = (); |
|
|
372 | $cb->(undef, { Status => 599, Reason => $errno, URL => $url }); |
|
|
373 | }); |
|
|
374 | $state{handle}->on_eof (sub { |
|
|
375 | %state = (); |
|
|
376 | $cb->(undef, { Status => 599, Reason => "unexpected end-of-file", URL => $url }); |
|
|
377 | }); |
|
|
378 | |
559 | |
379 | $state{handle}->starttls ("connect") if $rscheme eq "https"; |
560 | $state{handle}->starttls ("connect") if $rscheme eq "https"; |
380 | |
561 | |
381 | # handle actual, non-tunneled, request |
562 | # handle actual, non-tunneled, request |
382 | my $handle_actual_request = sub { |
563 | my $handle_actual_request = sub { |
383 | $state{handle}->starttls ("connect") if $uscheme eq "https" && !exists $state{handle}{tls}; |
564 | $state{handle}->starttls ("connect") if $uscheme eq "https" && !exists $state{handle}{tls}; |
384 | |
565 | |
385 | # send request |
566 | # send request |
386 | $state{handle}->push_write ( |
567 | $state{handle}->push_write ( |
387 | "$method $rpath HTTP/1.0\015\012" |
568 | "$method $rpath HTTP/1.1\015\012" |
388 | . (join "", map "\u$_: $hdr{$_}\015\012", keys %hdr) |
569 | . (join "", map "\u$_: $hdr{$_}\015\012", grep defined $hdr{$_}, keys %hdr) |
389 | . "\015\012" |
570 | . "\015\012" |
390 | . (delete $arg{body}) |
571 | . (delete $arg{body}) |
391 | ); |
572 | ); |
392 | |
573 | |
|
|
574 | # return if error occured during push_write() |
|
|
575 | return unless %state; |
|
|
576 | |
393 | %hdr = (); # reduce memory usage, save a kitten |
577 | %hdr = (); # reduce memory usage, save a kitten, also make it possible to re-use |
394 | |
578 | |
395 | # status line |
579 | # status line and headers |
396 | $state{handle}->push_read (line => $qr_nl, sub { |
580 | $state{read_response} = sub { |
397 | $_[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 })); |
|
|
399 | |
|
|
400 | my %hdr = ( # response headers |
|
|
401 | HTTPVersion => ",$1", |
|
|
402 | Status => ",$2", |
|
|
403 | Reason => ",$3", |
|
|
404 | URL => ",$url" |
|
|
405 | ); |
|
|
406 | |
|
|
407 | # headers, could be optimized a bit |
|
|
408 | $state{handle}->unshift_read (line => $qr_nlnl, sub { |
|
|
409 | for ("$_[1]\012") { |
581 | for ("$_[1]") { |
410 | y/\015//d; # weed out any \015, as they show up in the weirdest of places. |
582 | y/\015//d; # weed out any \015, as they show up in the weirdest of places. |
411 | |
583 | |
412 | # we support spaces in field names, as lotus domino |
584 | /^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\012]*) )? \012/igxc |
413 | # creates them (actually spaces around seperators |
585 | or return (%state = (), $cb->(undef, { @pseudo, Status => 599, Reason => "Invalid server response" })); |
414 | # are strictly allowed in http, they are a security issue). |
|
|
415 | $hdr{lc $1} .= ",$2" |
|
|
416 | while /\G |
|
|
417 | ([^:\000-\037]+): |
|
|
418 | [\011\040]* |
|
|
419 | ((?: [^\012]+ | \012[\011\040] )*) |
|
|
420 | \012 |
|
|
421 | /gxc; |
|
|
422 | |
586 | |
|
|
587 | # 100 Continue handling |
|
|
588 | # should not happen as we don't send expect: 100-continue, |
|
|
589 | # but we handle it just in case. |
|
|
590 | # since we send the request body regardless, if we get an error |
|
|
591 | # we are out of-sync, which we currently do NOT handle correctly. |
|
|
592 | return $state{handle}->push_read (line => $qr_nlnl, $state{read_response}) |
|
|
593 | if $2 eq 100; |
|
|
594 | |
|
|
595 | push @pseudo, |
|
|
596 | HTTPVersion => $1, |
|
|
597 | Status => $2, |
|
|
598 | Reason => $3, |
423 | /\G$/ |
599 | ; |
|
|
600 | |
|
|
601 | my $hdr = parse_hdr |
424 | or return (%state = (), $cb->(undef, { Status => 599, Reason => "garbled response headers", URL => $url })); |
602 | or return (%state = (), $cb->(undef, { @pseudo, Status => 599, Reason => "Garbled response headers" })); |
|
|
603 | |
|
|
604 | %hdr = (%$hdr, @pseudo); |
|
|
605 | } |
|
|
606 | |
|
|
607 | # redirect handling |
|
|
608 | # microsoft and other shitheads don't give a shit for following standards, |
|
|
609 | # try to support some common forms of broken Location headers. |
|
|
610 | if ($hdr{location} !~ /^(?: $ | [^:\/?\#]+ : )/x) { |
|
|
611 | $hdr{location} =~ s/^\.\/+//; |
|
|
612 | |
|
|
613 | my $url = "$rscheme://$uhost:$uport"; |
|
|
614 | |
|
|
615 | unless ($hdr{location} =~ s/^\///) { |
|
|
616 | $url .= $upath; |
|
|
617 | $url =~ s/\/[^\/]*$//; |
425 | } |
618 | } |
426 | |
619 | |
427 | substr $_, 0, 1, "" |
620 | $hdr{location} = "$url/$hdr{location}"; |
428 | for values %hdr; |
621 | } |
429 | |
622 | |
430 | my $finish = sub { |
623 | my $redirect; |
|
|
624 | |
|
|
625 | if ($recurse) { |
|
|
626 | my $status = $hdr{Status}; |
|
|
627 | |
|
|
628 | # industry standard is to redirect POST as GET for |
|
|
629 | # 301, 302 and 303, in contrast to http/1.0 and 1.1. |
|
|
630 | # also, the UA should ask the user for 301 and 307 and POST, |
|
|
631 | # industry standard seems to be to simply follow. |
|
|
632 | # we go with the industry standard. |
|
|
633 | if ($status == 301 or $status == 302 or $status == 303) { |
|
|
634 | # HTTP/1.1 is unclear on how to mutate the method |
|
|
635 | $method = "GET" unless $method eq "HEAD"; |
|
|
636 | $redirect = 1; |
|
|
637 | } elsif ($status == 307) { |
|
|
638 | $redirect = 1; |
|
|
639 | } |
|
|
640 | } |
|
|
641 | |
|
|
642 | my $finish = sub { # ($data, $err_status, $err_reason[, $keepalive]) |
|
|
643 | my $keepalive = pop; |
|
|
644 | |
431 | $state{handle}->destroy; |
645 | $state{handle}->destroy if $state{handle}; |
432 | %state = (); |
646 | %state = (); |
433 | |
647 | |
|
|
648 | if (defined $_[1]) { |
|
|
649 | $hdr{OrigStatus} = $hdr{Status}; $hdr{Status} = $_[1]; |
|
|
650 | $hdr{OrigReason} = $hdr{Reason}; $hdr{Reason} = $_[2]; |
|
|
651 | } |
|
|
652 | |
434 | # set-cookie processing |
653 | # set-cookie processing |
435 | if ($arg{cookie_jar}) { |
654 | if ($arg{cookie_jar}) { |
436 | for ($hdr{"set-cookie"}) { |
655 | for ($hdr{"set-cookie"}) { |
437 | # parse NAME=VALUE |
656 | # parse NAME=VALUE |
438 | my @kv; |
657 | my @kv; |
439 | |
658 | |
|
|
659 | while ( |
|
|
660 | m{ |
|
|
661 | \G\s* |
|
|
662 | (?: |
|
|
663 | expires \s*=\s* ([A-Z][a-z][a-z],\ [^,;]+) |
440 | while (/\G\s* ([^=;,[:space:]]+) \s*=\s* (?: "((?:[^\\"]+|\\.)*)" | ([^=;,[:space:]]*) )/gcxs) { |
664 | | ([^=;,[:space:]]+) \s*=\s* (?: "((?:[^\\"]+|\\.)*)" | ([^=;,[:space:]]*) ) |
|
|
665 | ) |
|
|
666 | }gcxsi |
|
|
667 | ) { |
441 | my $name = $1; |
668 | my $name = $2; |
|
|
669 | my $value = $4; |
|
|
670 | |
|
|
671 | unless (defined $name) { |
|
|
672 | # expires |
|
|
673 | $name = "expires"; |
|
|
674 | $value = $1; |
|
|
675 | } elsif (!defined $value) { |
|
|
676 | # quoted |
442 | my $value = $3; |
677 | $value = $3; |
443 | |
|
|
444 | unless ($value) { |
|
|
445 | $value = $2; |
|
|
446 | $value =~ s/\\(.)/$1/gs; |
678 | $value =~ s/\\(.)/$1/gs; |
|
|
679 | } |
|
|
680 | |
|
|
681 | push @kv, lc $name, $value; |
|
|
682 | |
|
|
683 | last unless /\G\s*;/gc; |
|
|
684 | } |
|
|
685 | |
|
|
686 | last unless @kv; |
|
|
687 | |
|
|
688 | my $name = shift @kv; |
|
|
689 | my %kv = (value => shift @kv, @kv); |
|
|
690 | |
|
|
691 | $kv{expires} ||= format_date (AE::now + $kv{"max-age"}) |
|
|
692 | if exists $kv{"max-age"}; |
|
|
693 | |
|
|
694 | my $cdom; |
|
|
695 | my $cpath = (delete $kv{path}) || "/"; |
|
|
696 | |
|
|
697 | if (exists $kv{domain}) { |
|
|
698 | $cdom = delete $kv{domain}; |
|
|
699 | |
|
|
700 | $cdom =~ s/^\.?/./; # make sure it starts with a "." |
|
|
701 | |
|
|
702 | next if $cdom =~ /\.$/; |
|
|
703 | |
|
|
704 | # this is not rfc-like and not netscape-like. go figure. |
|
|
705 | my $ndots = $cdom =~ y/.//; |
|
|
706 | next if $ndots < ($cdom =~ /\.[^.][^.]\.[^.][^.]$/ ? 3 : 2); |
|
|
707 | } else { |
|
|
708 | $cdom = $uhost; |
|
|
709 | } |
|
|
710 | |
|
|
711 | # store it |
|
|
712 | $arg{cookie_jar}{version} = 1; |
|
|
713 | $arg{cookie_jar}{$cdom}{$cpath}{$name} = \%kv; |
|
|
714 | |
|
|
715 | redo if /\G\s*,/gc; |
|
|
716 | } |
|
|
717 | } |
|
|
718 | |
|
|
719 | if ($redirect && exists $hdr{location}) { |
|
|
720 | # we ignore any errors, as it is very common to receive |
|
|
721 | # Content-Length != 0 but no actual body |
|
|
722 | # we also access %hdr, as $_[1] might be an erro |
|
|
723 | http_request ( |
|
|
724 | $method => $hdr{location}, |
|
|
725 | %arg, |
|
|
726 | recurse => $recurse - 1, |
|
|
727 | Redirect => [$_[0], \%hdr], |
|
|
728 | $cb); |
|
|
729 | } else { |
|
|
730 | $cb->($_[0], \%hdr); |
|
|
731 | } |
|
|
732 | }; |
|
|
733 | |
|
|
734 | my $len = $hdr{"content-length"}; |
|
|
735 | |
|
|
736 | if (!$redirect && $arg{on_header} && !$arg{on_header}(\%hdr)) { |
|
|
737 | $finish->(undef, 598 => "Request cancelled by on_header"); |
|
|
738 | } elsif ( |
|
|
739 | $hdr{Status} =~ /^(?:1..|204|205|304)$/ |
|
|
740 | or $method eq "HEAD" |
|
|
741 | or (defined $len && !$len) |
|
|
742 | ) { |
|
|
743 | # no body |
|
|
744 | $finish->("", undef, undef, 1); |
|
|
745 | } else { |
|
|
746 | # body handling, many different code paths |
|
|
747 | # - no body expected |
|
|
748 | # - want_body_handle |
|
|
749 | # - te chunked |
|
|
750 | # - 2x length known (with or without on_body) |
|
|
751 | # - 2x length not known (with or without on_body) |
|
|
752 | if (!$redirect && $arg{want_body_handle}) { |
|
|
753 | $_[0]->on_eof (undef); |
|
|
754 | $_[0]->on_error (undef); |
|
|
755 | $_[0]->on_read (undef); |
|
|
756 | |
|
|
757 | $finish->(delete $state{handle}); |
|
|
758 | |
|
|
759 | } elsif ($hdr{"transfer-encoding"} =~ /\bchunked\b/i) { |
|
|
760 | my $cl = 0; |
|
|
761 | my $body = undef; |
|
|
762 | my $on_body = $arg{on_body} || sub { $body .= shift; 1 }; |
|
|
763 | |
|
|
764 | $_[0]->on_error (sub { $finish->(undef, 599 => $_[2]) }); |
|
|
765 | |
|
|
766 | my $read_chunk; $read_chunk = sub { |
|
|
767 | $_[1] =~ /^([0-9a-fA-F]+)/ |
|
|
768 | or $finish->(undef, 599 => "Garbled chunked transfer encoding"); |
|
|
769 | |
|
|
770 | my $len = hex $1; |
|
|
771 | |
|
|
772 | if ($len) { |
|
|
773 | $cl += $len; |
|
|
774 | |
|
|
775 | $_[0]->push_read (chunk => $len, sub { |
|
|
776 | $on_body->($_[1], \%hdr) |
|
|
777 | or return $finish->(undef, 598 => "Request cancelled by on_body"); |
|
|
778 | |
|
|
779 | $_[0]->push_read (line => sub { |
|
|
780 | length $_[1] |
|
|
781 | and return $finish->(undef, 599 => "Garbled chunked transfer encoding"); |
|
|
782 | $_[0]->push_read (line => $read_chunk); |
|
|
783 | }); |
|
|
784 | }); |
|
|
785 | } else { |
|
|
786 | $hdr{"content-length"} ||= $cl; |
|
|
787 | |
|
|
788 | $_[0]->push_read (line => $qr_nlnl, sub { |
|
|
789 | if (length $_[1]) { |
|
|
790 | for ("$_[1]") { |
|
|
791 | y/\015//d; # weed out any \015, as they show up in the weirdest of places. |
|
|
792 | |
|
|
793 | my $hdr = parse_hdr |
|
|
794 | or return $finish->(undef, 599 => "Garbled response trailers"); |
|
|
795 | |
|
|
796 | %hdr = (%hdr, %$hdr); |
|
|
797 | } |
447 | } |
798 | } |
448 | |
799 | |
449 | push @kv, $name => $value; |
800 | $finish->($body, undef, undef, 1); |
450 | |
|
|
451 | last unless /\G\s*;/gc; |
|
|
452 | } |
801 | }); |
453 | |
|
|
454 | last unless @kv; |
|
|
455 | |
|
|
456 | my $name = shift @kv; |
|
|
457 | my %kv = (value => shift @kv, @kv); |
|
|
458 | |
|
|
459 | my $cdom; |
|
|
460 | my $cpath = (delete $kv{path}) || "/"; |
|
|
461 | |
|
|
462 | if (exists $kv{domain}) { |
|
|
463 | $cdom = delete $kv{domain}; |
|
|
464 | |
|
|
465 | $cdom =~ s/^\.?/./; # make sure it starts with a "." |
|
|
466 | |
|
|
467 | next if $cdom =~ /\.$/; |
|
|
468 | |
|
|
469 | # this is not rfc-like and not netscape-like. go figure. |
|
|
470 | my $ndots = $cdom =~ y/.//; |
|
|
471 | next if $ndots < ($cdom =~ /\.[^.][^.]\.[^.][^.]$/ ? 3 : 2); |
|
|
472 | } else { |
|
|
473 | $cdom = $uhost; |
|
|
474 | } |
|
|
475 | |
|
|
476 | # store it |
|
|
477 | $arg{cookie_jar}{version} = 1; |
|
|
478 | $arg{cookie_jar}{$cdom}{$cpath}{$name} = \%kv; |
|
|
479 | |
|
|
480 | redo if /\G\s*,/gc; |
|
|
481 | } |
802 | } |
482 | } |
803 | }; |
483 | |
804 | |
484 | # microsoft and other shitheads don't give a shit for following standards, |
805 | $_[0]->push_read (line => $read_chunk); |
485 | # try to support some common forms of broken Location headers. |
|
|
486 | if ($_[1]{location} !~ /^(?: $ | [^:\/?\#]+ : )/x) { |
|
|
487 | $_[1]{location} =~ s/^\.\/+//; |
|
|
488 | |
806 | |
489 | my $url = "$rscheme://$uhost:$uport"; |
807 | } elsif ($arg{on_body}) { |
|
|
808 | $_[0]->on_error (sub { $finish->(undef, 599 => $_[2]) }); |
490 | |
809 | |
491 | unless ($_[1]{location} =~ s/^\///) { |
|
|
492 | $url .= $upath; |
|
|
493 | $url =~ s/\/[^\/]*$//; |
|
|
494 | } |
|
|
495 | |
|
|
496 | $_[1]{location} = "$url/$_[1]{location}"; |
|
|
497 | } |
|
|
498 | |
|
|
499 | if ($_[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, $cb); |
|
|
509 | } else { |
810 | if ($len) { |
510 | $cb->($_[0], $_[1]); |
811 | $_[0]->on_read (sub { |
511 | } |
812 | $len -= length $_[0]{rbuf}; |
512 | }; |
|
|
513 | |
813 | |
514 | if ($hdr{Status} =~ /^(?:1..|204|304)$/ or $method eq "HEAD") { |
814 | $arg{on_body}(delete $_[0]{rbuf}, \%hdr) |
515 | $finish->(undef, \%hdr); |
815 | or return $finish->(undef, 598 => "Request cancelled by on_body"); |
516 | } else { |
816 | |
517 | if (exists $hdr{"content-length"}) { |
|
|
518 | $_[0]->unshift_read (chunk => $hdr{"content-length"}, sub { |
|
|
519 | # could cache persistent connection now |
|
|
520 | if ($hdr{connection} =~ /\bkeep-alive\b/i) { |
|
|
521 | # but we don't, due to misdesigns, this is annoyingly complex |
|
|
522 | }; |
817 | $len > 0 |
523 | |
818 | or $finish->("", undef, undef, 1); |
524 | $finish->($_[1], \%hdr); |
|
|
525 | }); |
819 | }); |
526 | } else { |
820 | } else { |
527 | # too bad, need to read until we get an error or EOF, |
821 | $_[0]->on_eof (sub { |
528 | # no way to detect winged data. |
822 | $finish->(""); |
|
|
823 | }); |
|
|
824 | $_[0]->on_read (sub { |
|
|
825 | $arg{on_body}(delete $_[0]{rbuf}, \%hdr) |
|
|
826 | or $finish->(undef, 598 => "Request cancelled by on_body"); |
|
|
827 | }); |
|
|
828 | } |
|
|
829 | } else { |
|
|
830 | $_[0]->on_eof (undef); |
|
|
831 | |
|
|
832 | if ($len) { |
|
|
833 | $_[0]->on_error (sub { $finish->(undef, 599 => $_[2]) }); |
|
|
834 | $_[0]->on_read (sub { |
|
|
835 | $finish->((substr delete $_[0]{rbuf}, 0, $len, ""), undef, undef, 1) |
|
|
836 | if $len <= length $_[0]{rbuf}; |
|
|
837 | }); |
|
|
838 | } else { |
529 | $_[0]->on_error (sub { |
839 | $_[0]->on_error (sub { |
530 | # delete ought to be more efficient, as we would have to make |
840 | ($! == Errno::EPIPE || !$!) |
531 | # a copy otherwise as $_[0] gets destroyed. |
|
|
532 | $finish->(delete $_[0]{rbuf}, \%hdr); |
841 | ? $finish->(delete $_[0]{rbuf}) |
|
|
842 | : $finish->(undef, 599 => $_[2]); |
533 | }); |
843 | }); |
534 | $_[0]->on_eof (undef); |
|
|
535 | $_[0]->on_read (sub { }); |
844 | $_[0]->on_read (sub { }); |
536 | } |
845 | } |
537 | } |
846 | } |
538 | }); |
847 | } |
539 | }); |
848 | }; |
|
|
849 | |
|
|
850 | $state{handle}->push_read (line => $qr_nlnl, $state{read_response}); |
540 | }; |
851 | }; |
541 | |
852 | |
542 | # now handle proxy-CONNECT method |
853 | # now handle proxy-CONNECT method |
543 | if ($proxy && $uscheme eq "https") { |
854 | if ($proxy && $uscheme eq "https") { |
544 | # oh dear, we have to wrap it into a connect request |
855 | # oh dear, we have to wrap it into a connect request |
545 | |
856 | |
546 | # maybe re-use $uauthority with patched port? |
857 | # maybe re-use $uauthority with patched port? |
547 | $state{handle}->push_write ("CONNECT $uhost:$uport HTTP/1.0\015\012Host: $uhost\015\012\015\012"); |
858 | $state{handle}->push_write ("CONNECT $uhost:$uport HTTP/1.0\015\012Host: $uhost\015\012\015\012"); |
548 | $state{handle}->push_read (line => $qr_nlnl, sub { |
859 | $state{handle}->push_read (line => $qr_nlnl, sub { |
549 | $_[1] =~ /^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\015\012]*) )?/ix |
860 | $_[1] =~ /^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\015\012]*) )?/ix |
550 | or return (%state = (), $cb->(undef, { Status => 599, Reason => "invalid proxy connect response ($_[1])", URL => $url })); |
861 | or return (%state = (), $cb->(undef, { @pseudo, Status => 599, Reason => "Invalid proxy connect response ($_[1])" })); |
551 | |
862 | |
552 | if ($2 == 200) { |
863 | if ($2 == 200) { |
553 | $rpath = $upath; |
864 | $rpath = $upath; |
554 | &$handle_actual_request; |
865 | &$handle_actual_request; |
555 | } else { |
866 | } else { |
556 | %state = (); |
867 | %state = (); |
557 | $cb->(undef, { Status => $2, Reason => $3, URL => $url }); |
868 | $cb->(undef, { @pseudo, Status => $2, Reason => $3 }); |
558 | } |
869 | } |
559 | }); |
870 | }); |
560 | } else { |
871 | } else { |
561 | &$handle_actual_request; |
872 | &$handle_actual_request; |
562 | } |
873 | } |
563 | |
|
|
564 | }, sub { |
|
|
565 | $timeout |
|
|
566 | }; |
874 | }; |
|
|
875 | |
|
|
876 | my $tcp_connect = $arg{tcp_connect} |
|
|
877 | || do { require AnyEvent::Socket; \&AnyEvent::Socket::tcp_connect }; |
|
|
878 | |
|
|
879 | $state{connect_guard} = $tcp_connect->($rhost, $rport, $connect_cb, $arg{on_prepare} || sub { $timeout }); |
|
|
880 | |
567 | }; |
881 | }; |
568 | |
882 | |
569 | defined wantarray && AnyEvent::Util::guard { %state = () } |
883 | defined wantarray && AnyEvent::Util::guard { %state = () } |
570 | } |
884 | } |
571 | |
885 | |
… | |
… | |
585 | &http_request |
899 | &http_request |
586 | } |
900 | } |
587 | |
901 | |
588 | =back |
902 | =back |
589 | |
903 | |
|
|
904 | =head2 DNS CACHING |
|
|
905 | |
|
|
906 | AnyEvent::HTTP uses the AnyEvent::Socket::tcp_connect function for |
|
|
907 | the actual connection, which in turn uses AnyEvent::DNS to resolve |
|
|
908 | hostnames. The latter is a simple stub resolver and does no caching |
|
|
909 | on its own. If you want DNS caching, you currently have to provide |
|
|
910 | your own default resolver (by storing a suitable resolver object in |
|
|
911 | C<$AnyEvent::DNS::RESOLVER>). |
|
|
912 | |
590 | =head2 GLOBAL FUNCTIONS AND VARIABLES |
913 | =head2 GLOBAL FUNCTIONS AND VARIABLES |
591 | |
914 | |
592 | =over 4 |
915 | =over 4 |
593 | |
916 | |
594 | =item AnyEvent::HTTP::set_proxy "proxy-url" |
917 | =item AnyEvent::HTTP::set_proxy "proxy-url" |
595 | |
918 | |
596 | Sets the default proxy server to use. The proxy-url must begin with a |
919 | Sets the default proxy server to use. The proxy-url must begin with a |
597 | string of the form C<http://host:port> (optionally C<https:...>). |
920 | string of the form C<http://host:port> (optionally C<https:...>), croaks |
|
|
921 | otherwise. |
|
|
922 | |
|
|
923 | To clear an already-set proxy, use C<undef>. |
|
|
924 | |
|
|
925 | =item $date = AnyEvent::HTTP::format_date $timestamp |
|
|
926 | |
|
|
927 | Takes a POSIX timestamp (seconds since the epoch) and formats it as a HTTP |
|
|
928 | Date (RFC 2616). |
|
|
929 | |
|
|
930 | =item $timestamp = AnyEvent::HTTP::parse_date $date |
|
|
931 | |
|
|
932 | Takes a HTTP Date (RFC 2616) or a Cookie date (netscape cookie spec) and |
|
|
933 | returns the corresponding POSIX timestamp, or C<undef> if the date cannot |
|
|
934 | be parsed. |
598 | |
935 | |
599 | =item $AnyEvent::HTTP::MAX_RECURSE |
936 | =item $AnyEvent::HTTP::MAX_RECURSE |
600 | |
937 | |
601 | The default value for the C<recurse> request parameter (default: C<10>). |
938 | The default value for the C<recurse> request parameter (default: C<10>). |
602 | |
939 | |
603 | =item $AnyEvent::HTTP::USERAGENT |
940 | =item $AnyEvent::HTTP::USERAGENT |
604 | |
941 | |
605 | The default value for the C<User-Agent> header (the default is |
942 | The default value for the C<User-Agent> header (the default is |
606 | C<Mozilla/5.0 (compatible; AnyEvent::HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)>). |
943 | C<Mozilla/5.0 (compatible; U; AnyEvent-HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)>). |
607 | |
944 | |
608 | =item $AnyEvent::HTTP::MAX_PERSISTENT |
945 | =item $AnyEvent::HTTP::MAX_PER_HOST |
609 | |
946 | |
610 | The maximum number of persistent connections to keep open (default: 8). |
947 | The maximum number of concurrent connections to the same host (identified |
|
|
948 | by the hostname). If the limit is exceeded, then the additional requests |
|
|
949 | are queued until previous connections are closed. |
611 | |
950 | |
612 | Not implemented currently. |
951 | The default value for this is C<4>, and it is highly advisable to not |
613 | |
952 | increase it. |
614 | =item $AnyEvent::HTTP::PERSISTENT_TIMEOUT |
|
|
615 | |
|
|
616 | The maximum time to cache a persistent connection, in seconds (default: 2). |
|
|
617 | |
|
|
618 | Not implemented currently. |
|
|
619 | |
953 | |
620 | =item $AnyEvent::HTTP::ACTIVE |
954 | =item $AnyEvent::HTTP::ACTIVE |
621 | |
955 | |
622 | The number of active connections. This is not the number of currently |
956 | The number of active connections. This is not the number of currently |
623 | running requests, but the number of currently open and non-idle TCP |
957 | running requests, but the number of currently open and non-idle TCP |
… | |
… | |
625 | |
959 | |
626 | =back |
960 | =back |
627 | |
961 | |
628 | =cut |
962 | =cut |
629 | |
963 | |
|
|
964 | our @month = qw(Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec); |
|
|
965 | our @weekday = qw(Sun Mon Tue Wed Thu Fri Sat); |
|
|
966 | |
|
|
967 | sub format_date($) { |
|
|
968 | my ($time) = @_; |
|
|
969 | |
|
|
970 | # RFC 822/1123 format |
|
|
971 | my ($S, $M, $H, $mday, $mon, $year, $wday, $yday, undef) = gmtime $time; |
|
|
972 | |
|
|
973 | sprintf "%s, %02d %s %04d %02d:%02d:%02d GMT", |
|
|
974 | $weekday[$wday], $mday, $month[$mon], $year + 1900, |
|
|
975 | $H, $M, $S; |
|
|
976 | } |
|
|
977 | |
|
|
978 | sub parse_date($) { |
|
|
979 | my ($date) = @_; |
|
|
980 | |
|
|
981 | my ($d, $m, $y, $H, $M, $S); |
|
|
982 | |
|
|
983 | if ($date =~ /^[A-Z][a-z][a-z], ([0-9][0-9])[\- ]([A-Z][a-z][a-z])[\- ]([0-9][0-9][0-9][0-9]) ([0-9][0-9]):([0-9][0-9]):([0-9][0-9]) GMT$/) { |
|
|
984 | # RFC 822/1123, required by RFC 2616 (with " ") |
|
|
985 | # cookie dates (with "-") |
|
|
986 | |
|
|
987 | ($d, $m, $y, $H, $M, $S) = ($1, $2, $3, $4, $5, $6); |
|
|
988 | |
|
|
989 | } elsif ($date =~ /^[A-Z][a-z]+, ([0-9][0-9])-([A-Z][a-z][a-z])-([0-9][0-9]) ([0-9][0-9]):([0-9][0-9]):([0-9][0-9]) GMT$/) { |
|
|
990 | # RFC 850 |
|
|
991 | ($d, $m, $y, $H, $M, $S) = ($1, $2, $3 < 69 ? $3 + 2000 : $3 + 1900, $4, $5, $6); |
|
|
992 | |
|
|
993 | } elsif ($date =~ /^[A-Z][a-z][a-z] ([A-Z][a-z][a-z]) ([0-9 ][0-9]) ([0-9][0-9]):([0-9][0-9]):([0-9][0-9]) ([0-9][0-9][0-9][0-9])$/) { |
|
|
994 | # ISO C's asctime |
|
|
995 | ($d, $m, $y, $H, $M, $S) = ($2, $1, $6, $3, $4, $5); |
|
|
996 | } |
|
|
997 | # other formats fail in the loop below |
|
|
998 | |
|
|
999 | for (0..11) { |
|
|
1000 | if ($m eq $month[$_]) { |
|
|
1001 | require Time::Local; |
|
|
1002 | return Time::Local::timegm ($S, $M, $H, $d, $_, $y); |
|
|
1003 | } |
|
|
1004 | } |
|
|
1005 | |
|
|
1006 | undef |
|
|
1007 | } |
|
|
1008 | |
630 | sub set_proxy($) { |
1009 | sub set_proxy($) { |
|
|
1010 | if (length $_[0]) { |
631 | $PROXY = [$2, $3 || 3128, $1] if $_[0] =~ m%^(https?):// ([^:/]+) (?: : (\d*) )?%ix; |
1011 | $_[0] =~ m%^(https?):// ([^:/]+) (?: : (\d*) )?%ix |
|
|
1012 | or Carp::croak "$_[0]: invalid proxy URL"; |
|
|
1013 | $PROXY = [$2, $3 || 3128, $1] |
|
|
1014 | } else { |
|
|
1015 | undef $PROXY; |
|
|
1016 | } |
632 | } |
1017 | } |
633 | |
1018 | |
634 | # initialise proxy from environment |
1019 | # initialise proxy from environment |
|
|
1020 | eval { |
635 | set_proxy $ENV{http_proxy}; |
1021 | set_proxy $ENV{http_proxy}; |
|
|
1022 | }; |
|
|
1023 | |
|
|
1024 | =head2 SOCKS PROXIES |
|
|
1025 | |
|
|
1026 | Socks proxies are not directly supported by AnyEvent::HTTP. You can |
|
|
1027 | compile your perl to support socks, or use an external program such as |
|
|
1028 | F<socksify> (dante) or F<tsocks> to make your program use a socks proxy |
|
|
1029 | transparently. |
|
|
1030 | |
|
|
1031 | Alternatively, for AnyEvent::HTTP only, you can use your own |
|
|
1032 | C<tcp_connect> function that does the proxy handshake - here is an example |
|
|
1033 | that works with socks4a proxies: |
|
|
1034 | |
|
|
1035 | use Errno; |
|
|
1036 | use AnyEvent::Util; |
|
|
1037 | use AnyEvent::Socket; |
|
|
1038 | use AnyEvent::Handle; |
|
|
1039 | |
|
|
1040 | # host, port and username of/for your socks4a proxy |
|
|
1041 | my $socks_host = "10.0.0.23"; |
|
|
1042 | my $socks_port = 9050; |
|
|
1043 | my $socks_user = ""; |
|
|
1044 | |
|
|
1045 | sub socks4a_connect { |
|
|
1046 | my ($host, $port, $connect_cb, $prepare_cb) = @_; |
|
|
1047 | |
|
|
1048 | my $hdl = new AnyEvent::Handle |
|
|
1049 | connect => [$socks_host, $socks_port], |
|
|
1050 | on_prepare => sub { $prepare_cb->($_[0]{fh}) }, |
|
|
1051 | on_error => sub { $connect_cb->() }, |
|
|
1052 | ; |
|
|
1053 | |
|
|
1054 | $hdl->push_write (pack "CCnNZ*Z*", 4, 1, $port, 1, $socks_user, $host); |
|
|
1055 | |
|
|
1056 | $hdl->push_read (chunk => 8, sub { |
|
|
1057 | my ($hdl, $chunk) = @_; |
|
|
1058 | my ($status, $port, $ipn) = unpack "xCna4", $chunk; |
|
|
1059 | |
|
|
1060 | if ($status == 0x5a) { |
|
|
1061 | $connect_cb->($hdl->{fh}, (format_address $ipn) . ":$port"); |
|
|
1062 | } else { |
|
|
1063 | $! = Errno::ENXIO; $connect_cb->(); |
|
|
1064 | } |
|
|
1065 | }); |
|
|
1066 | |
|
|
1067 | $hdl |
|
|
1068 | } |
|
|
1069 | |
|
|
1070 | Use C<socks4a_connect> instead of C<tcp_connect> when doing C<http_request>s, |
|
|
1071 | possibly after switching off other proxy types: |
|
|
1072 | |
|
|
1073 | AnyEvent::HTTP::set_proxy undef; # usually you do not want other proxies |
|
|
1074 | |
|
|
1075 | http_get 'http://www.google.com', tcp_connect => \&socks4a_connect, sub { |
|
|
1076 | my ($data, $headers) = @_; |
|
|
1077 | ... |
|
|
1078 | }; |
636 | |
1079 | |
637 | =head1 SEE ALSO |
1080 | =head1 SEE ALSO |
638 | |
1081 | |
639 | L<AnyEvent>. |
1082 | L<AnyEvent>. |
640 | |
1083 | |