… | |
… | |
15 | This module is an L<AnyEvent> user, you need to make sure that you use and |
15 | This module is an L<AnyEvent> user, you need to make sure that you use and |
16 | run a supported event loop. |
16 | run a supported event loop. |
17 | |
17 | |
18 | This module implements a simple, stateless and non-blocking HTTP |
18 | This module implements a simple, stateless and non-blocking HTTP |
19 | client. It supports GET, POST and other request methods, cookies and more, |
19 | client. It supports GET, POST and other request methods, cookies and more, |
20 | all on a very low level. It can follow redirects supports proxies and |
20 | all on a very low level. It can follow redirects, supports proxies, and |
21 | automatically limits the number of connections to the values specified in |
21 | automatically limits the number of connections to the values specified in |
22 | the RFC. |
22 | the RFC. |
23 | |
23 | |
24 | It should generally be a "good client" that is enough for most HTTP |
24 | It should generally be a "good client" that is enough for most HTTP |
25 | tasks. Simple tasks should be simple, but complex tasks should still be |
25 | tasks. Simple tasks should be simple, but complex tasks should still be |
… | |
… | |
36 | |
36 | |
37 | =cut |
37 | =cut |
38 | |
38 | |
39 | package AnyEvent::HTTP; |
39 | package AnyEvent::HTTP; |
40 | |
40 | |
41 | use strict; |
41 | use common::sense; |
42 | no warnings; |
|
|
43 | |
42 | |
44 | use Errno (); |
43 | use Errno (); |
45 | |
44 | |
46 | use AnyEvent 4.8 (); |
45 | use AnyEvent 5.0 (); |
47 | use AnyEvent::Util (); |
46 | use AnyEvent::Util (); |
48 | use AnyEvent::Socket (); |
|
|
49 | use AnyEvent::Handle (); |
47 | use AnyEvent::Handle (); |
50 | |
48 | |
51 | use base Exporter::; |
49 | use base Exporter::; |
52 | |
50 | |
53 | our $VERSION = '1.42'; |
51 | our $VERSION = '2.15'; |
54 | |
52 | |
55 | our @EXPORT = qw(http_get http_post http_head http_request); |
53 | our @EXPORT = qw(http_get http_post http_head http_request); |
56 | |
54 | |
57 | our $USERAGENT = "Mozilla/5.0 (compatible; U; AnyEvent-HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)"; |
55 | our $USERAGENT = "Mozilla/5.0 (compatible; U; AnyEvent-HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)"; |
58 | our $MAX_RECURSE = 10; |
56 | our $MAX_RECURSE = 10; |
59 | our $MAX_PERSISTENT = 8; |
|
|
60 | our $PERSISTENT_TIMEOUT = 2; |
57 | our $PERSISTENT_TIMEOUT = 3; |
61 | our $TIMEOUT = 300; |
58 | our $TIMEOUT = 300; |
62 | |
59 | our $MAX_PER_HOST = 4; # changing this is evil |
63 | # changing these is evil |
|
|
64 | our $MAX_PERSISTENT_PER_HOST = 0; |
|
|
65 | our $MAX_PER_HOST = 4; |
|
|
66 | |
60 | |
67 | our $PROXY; |
61 | our $PROXY; |
68 | our $ACTIVE = 0; |
62 | our $ACTIVE = 0; |
69 | |
63 | |
70 | my %KA_COUNT; # number of open keep-alive connections per host |
64 | my %KA_CACHE; # indexed by uhost currently, points to [$handle...] array |
71 | my %CO_SLOT; # number of open connections, and wait queue, per host |
65 | my %CO_SLOT; # number of open connections, and wait queue, per host |
72 | |
66 | |
73 | =item http_get $url, key => value..., $cb->($data, $headers) |
67 | =item http_get $url, key => value..., $cb->($data, $headers) |
74 | |
68 | |
75 | Executes an HTTP-GET request. See the http_request function for details on |
69 | Executes an HTTP-GET request. See the http_request function for details on |
… | |
… | |
92 | must be an absolute http or https URL. |
86 | must be an absolute http or https URL. |
93 | |
87 | |
94 | When called in void context, nothing is returned. In other contexts, |
88 | When called in void context, nothing is returned. In other contexts, |
95 | C<http_request> returns a "cancellation guard" - you have to keep the |
89 | 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 |
90 | 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. |
91 | destroyed before the callback is called, the request will be cancelled. |
98 | |
92 | |
99 | The callback will be called with the response body data as first argument |
93 | The callback will be called with the response body data as first argument |
100 | (or C<undef> if an error occured), and a hash-ref with response headers as |
94 | (or C<undef> if an error occured), and a hash-ref with response headers |
101 | second argument. |
95 | (and trailers) as second argument. |
102 | |
96 | |
103 | All the headers in that hash are lowercased. In addition to the response |
97 | 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> |
98 | headers, the "pseudo-headers" (uppercase to avoid clashing with possible |
|
|
99 | 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 |
100 | three parts of the HTTP Status-Line of the same name. If an error occurs |
|
|
101 | during the body phase of a request, then the original C<Status> and |
|
|
102 | C<Reason> values from the header are available as C<OrigStatus> and |
|
|
103 | C<OrigReason>. |
|
|
104 | |
106 | pseudo-header C<URL> contains the original URL (which can differ from the |
105 | The pseudo-header C<URL> contains the actual URL (which can differ from |
107 | requested URL when following redirects). |
106 | the requested URL when following redirects - for example, you might get |
|
|
107 | an error that your URL scheme is not supported even though your URL is a |
|
|
108 | valid http URL because it redirected to an ftp URL, in which case you can |
|
|
109 | look at the URL pseudo header). |
|
|
110 | |
|
|
111 | The pseudo-header C<Redirect> only exists when the request was a result |
|
|
112 | of an internal redirect. In that case it is an array reference with |
|
|
113 | the C<($data, $headers)> from the redirect response. Note that this |
|
|
114 | response could in turn be the result of a redirect itself, and C<< |
|
|
115 | $headers->{Redirect}[1]{Redirect} >> will then contain the original |
|
|
116 | response, and so on. |
108 | |
117 | |
109 | If the server sends a header multiple times, then their contents will be |
118 | If the server sends a header multiple times, then their contents will be |
110 | joined together with a comma (C<,>), as per the HTTP spec. |
119 | joined together with a comma (C<,>), as per the HTTP spec. |
111 | |
120 | |
112 | If an internal error occurs, such as not being able to resolve a hostname, |
121 | If an internal error occurs, such as not being able to resolve a hostname, |
113 | then C<$data> will be C<undef>, C<< $headers->{Status} >> will be C<59x> |
122 | then C<$data> will be C<undef>, C<< $headers->{Status} >> will be |
114 | (usually C<599>) and the C<Reason> pseudo-header will contain an error |
123 | C<590>-C<599> and the C<Reason> pseudo-header will contain an error |
115 | message. |
124 | message. Currently the following status codes are used: |
|
|
125 | |
|
|
126 | =over 4 |
|
|
127 | |
|
|
128 | =item 595 - errors during connection establishment, proxy handshake. |
|
|
129 | |
|
|
130 | =item 596 - errors during TLS negotiation, request sending and header processing. |
|
|
131 | |
|
|
132 | =item 597 - errors during body receiving or processing. |
|
|
133 | |
|
|
134 | =item 598 - user aborted request via C<on_header> or C<on_body>. |
|
|
135 | |
|
|
136 | =item 599 - other, usually nonretryable, errors (garbled URL etc.). |
|
|
137 | |
|
|
138 | =back |
116 | |
139 | |
117 | A typical callback might look like this: |
140 | A typical callback might look like this: |
118 | |
141 | |
119 | sub { |
142 | sub { |
120 | my ($body, $hdr) = @_; |
143 | my ($body, $hdr) = @_; |
… | |
… | |
131 | |
154 | |
132 | =over 4 |
155 | =over 4 |
133 | |
156 | |
134 | =item recurse => $count (default: $MAX_RECURSE) |
157 | =item recurse => $count (default: $MAX_RECURSE) |
135 | |
158 | |
136 | Whether to recurse requests or not, e.g. on redirects, authentication |
159 | Whether to recurse requests or not, e.g. on redirects, authentication and |
137 | retries and so on, and how often to do so. |
160 | other retries and so on, and how often to do so. |
138 | |
161 | |
139 | =item headers => hashref |
162 | =item headers => hashref |
140 | |
163 | |
141 | The request headers to use. Currently, C<http_request> may provide its |
164 | The request headers to use. Currently, C<http_request> may provide its own |
142 | own C<Host:>, C<Content-Length:>, C<Connection:> and C<Cookie:> headers |
165 | C<Host:>, C<Content-Length:>, C<Connection:> and C<Cookie:> headers and |
143 | and will provide defaults for C<User-Agent:> and C<Referer:> (this can be |
166 | will provide defaults at least for C<TE:>, C<Referer:> and C<User-Agent:> |
144 | suppressed by using C<undef> for these headers in which case they won't be |
167 | (this can be suppressed by using C<undef> for these headers in which case |
145 | sent at all). |
168 | they won't be sent at all). |
|
|
169 | |
|
|
170 | You really should provide your own C<User-Agent:> header value that is |
|
|
171 | appropriate for your program - I wouldn't be surprised if the default |
|
|
172 | AnyEvent string gets blocked by webservers sooner or later. |
|
|
173 | |
|
|
174 | Also, make sure that your headers names and values do not contain any |
|
|
175 | embedded newlines. |
146 | |
176 | |
147 | =item timeout => $seconds |
177 | =item timeout => $seconds |
148 | |
178 | |
149 | The time-out to use for various stages - each connect attempt will reset |
179 | The time-out to use for various stages - each connect attempt will reset |
150 | the timeout, as will read or write activity. Default timeout is 5 minutes. |
180 | the timeout, as will read or write activity, i.e. this is not an overall |
|
|
181 | timeout. |
|
|
182 | |
|
|
183 | Default timeout is 5 minutes. |
151 | |
184 | |
152 | =item proxy => [$host, $port[, $scheme]] or undef |
185 | =item proxy => [$host, $port[, $scheme]] or undef |
153 | |
186 | |
154 | Use the given http proxy for all requests. If not specified, then the |
187 | Use the given http proxy for all requests, or no proxy if C<undef> is |
155 | default proxy (as specified by C<$ENV{http_proxy}>) is used. |
188 | used. |
156 | |
189 | |
157 | C<$scheme> must be either missing, C<http> for HTTP or C<https> for |
190 | C<$scheme> must be either missing or must be C<http> for HTTP. |
158 | HTTPS. |
191 | |
|
|
192 | If not specified, then the default proxy is used (see |
|
|
193 | C<AnyEvent::HTTP::set_proxy>). |
159 | |
194 | |
160 | =item body => $string |
195 | =item body => $string |
161 | |
196 | |
162 | The request body, usually empty. Will be-sent as-is (future versions of |
197 | The request body, usually empty. Will be sent as-is (future versions of |
163 | this module might offer more options). |
198 | this module might offer more options). |
164 | |
199 | |
165 | =item cookie_jar => $hash_ref |
200 | =item cookie_jar => $hash_ref |
166 | |
201 | |
167 | Passing this parameter enables (simplified) cookie-processing, loosely |
202 | Passing this parameter enables (simplified) cookie-processing, loosely |
168 | based on the original netscape specification. |
203 | based on the original netscape specification. |
169 | |
204 | |
170 | The C<$hash_ref> must be an (initially empty) hash reference which will |
205 | The C<$hash_ref> must be an (initially empty) hash reference which |
171 | get updated automatically. It is possible to save the cookie_jar to |
206 | will get updated automatically. It is possible to save the cookie jar |
172 | persistent storage with something like JSON or Storable, but this is not |
207 | to persistent storage with something like JSON or Storable - see the |
173 | recommended, as expiry times are currently being ignored. |
208 | C<AnyEvent::HTTP::cookie_jar_expire> function if you wish to remove |
|
|
209 | expired or session-only cookies, and also for documentation on the format |
|
|
210 | of the cookie jar. |
174 | |
211 | |
175 | Note that this cookie implementation is not of very high quality, nor |
212 | Note that this cookie implementation is not meant to be complete. If |
176 | meant to be complete. If you want complete cookie management you have to |
213 | you want complete cookie management you have to do that on your |
177 | do that on your own. C<cookie_jar> is meant as a quick fix to get some |
214 | own. C<cookie_jar> is meant as a quick fix to get most cookie-using sites |
178 | cookie-using sites working. Cookies are a privacy disaster, do not use |
215 | working. Cookies are a privacy disaster, do not use them unless required |
179 | them unless required to. |
216 | to. |
|
|
217 | |
|
|
218 | When cookie processing is enabled, the C<Cookie:> and C<Set-Cookie:> |
|
|
219 | headers will be set and handled by this module, otherwise they will be |
|
|
220 | left untouched. |
180 | |
221 | |
181 | =item tls_ctx => $scheme | $tls_ctx |
222 | =item tls_ctx => $scheme | $tls_ctx |
182 | |
223 | |
183 | Specifies the AnyEvent::TLS context to be used for https connections. This |
224 | Specifies the AnyEvent::TLS context to be used for https connections. This |
184 | parameter follows the same rules as the C<tls_ctx> parameter to |
225 | parameter follows the same rules as the C<tls_ctx> parameter to |
… | |
… | |
188 | verification) TLS context. |
229 | verification) TLS context. |
189 | |
230 | |
190 | The default for this option is C<low>, which could be interpreted as "give |
231 | The default for this option is C<low>, which could be interpreted as "give |
191 | me the page, no matter what". |
232 | me the page, no matter what". |
192 | |
233 | |
|
|
234 | See also the C<sessionid> parameter. |
|
|
235 | |
|
|
236 | =item session => $string |
|
|
237 | |
|
|
238 | The module might reuse connections to the same host internally. Sometimes |
|
|
239 | (e.g. when using TLS), you do not want to reuse connections from other |
|
|
240 | sessions. This can be achieved by setting this parameter to some unique |
|
|
241 | ID (such as the address of an object storing your state data, or the TLS |
|
|
242 | context) - only connections using the same unique ID will be reused. |
|
|
243 | |
|
|
244 | =item on_prepare => $callback->($fh) |
|
|
245 | |
|
|
246 | In rare cases you need to "tune" the socket before it is used to |
|
|
247 | connect (for exmaple, to bind it on a given IP address). This parameter |
|
|
248 | overrides the prepare callback passed to C<AnyEvent::Socket::tcp_connect> |
|
|
249 | and behaves exactly the same way (e.g. it has to provide a |
|
|
250 | timeout). See the description for the C<$prepare_cb> argument of |
|
|
251 | C<AnyEvent::Socket::tcp_connect> for details. |
|
|
252 | |
|
|
253 | =item tcp_connect => $callback->($host, $service, $connect_cb, $prepare_cb) |
|
|
254 | |
|
|
255 | In even rarer cases you want total control over how AnyEvent::HTTP |
|
|
256 | establishes connections. Normally it uses L<AnyEvent::Socket::tcp_connect> |
|
|
257 | to do this, but you can provide your own C<tcp_connect> function - |
|
|
258 | obviously, it has to follow the same calling conventions, except that it |
|
|
259 | may always return a connection guard object. |
|
|
260 | |
|
|
261 | There are probably lots of weird uses for this function, starting from |
|
|
262 | tracing the hosts C<http_request> actually tries to connect, to (inexact |
|
|
263 | but fast) host => IP address caching or even socks protocol support. |
|
|
264 | |
193 | =item on_header => $callback->($headers) |
265 | =item on_header => $callback->($headers) |
194 | |
266 | |
195 | When specified, this callback will be called with the header hash as soon |
267 | When specified, this callback will be called with the header hash as soon |
196 | as headers have been successfully received from the remote server (not on |
268 | as headers have been successfully received from the remote server (not on |
197 | locally-generated errors). |
269 | locally-generated errors). |
… | |
… | |
202 | |
274 | |
203 | This callback is useful, among other things, to quickly reject unwanted |
275 | This callback is useful, among other things, to quickly reject unwanted |
204 | content, which, if it is supposed to be rare, can be faster than first |
276 | content, which, if it is supposed to be rare, can be faster than first |
205 | doing a C<HEAD> request. |
277 | doing a C<HEAD> request. |
206 | |
278 | |
|
|
279 | The downside is that cancelling the request makes it impossible to re-use |
|
|
280 | the connection. Also, the C<on_header> callback will not receive any |
|
|
281 | trailer (headers sent after the response body). |
|
|
282 | |
207 | Example: cancel the request unless the content-type is "text/html". |
283 | Example: cancel the request unless the content-type is "text/html". |
208 | |
284 | |
209 | on_header => sub { |
285 | on_header => sub { |
210 | $_[0]{"content-type"} =~ /^text\/html\s*(?:;|$)/ |
286 | $_[0]{"content-type"} =~ /^text\/html\s*(?:;|$)/ |
211 | }, |
287 | }, |
… | |
… | |
217 | string instead of the body data. |
293 | string instead of the body data. |
218 | |
294 | |
219 | It has to return either true (in which case AnyEvent::HTTP will continue), |
295 | It has to return either true (in which case AnyEvent::HTTP will continue), |
220 | or false, in which case AnyEvent::HTTP will cancel the download (and call |
296 | or false, in which case AnyEvent::HTTP will cancel the download (and call |
221 | the completion callback with an error code of C<598>). |
297 | the completion callback with an error code of C<598>). |
|
|
298 | |
|
|
299 | The downside to cancelling the request is that it makes it impossible to |
|
|
300 | re-use the connection. |
222 | |
301 | |
223 | This callback is useful when the data is too large to be held in memory |
302 | This callback is useful when the data is too large to be held in memory |
224 | (so the callback writes it to a file) or when only some information should |
303 | (so the callback writes it to a file) or when only some information should |
225 | be extracted, or when the body should be processed incrementally. |
304 | be extracted, or when the body should be processed incrementally. |
226 | |
305 | |
… | |
… | |
238 | called. Instead of the C<$body> argument containing the body data, the |
317 | called. Instead of the C<$body> argument containing the body data, the |
239 | callback will receive the L<AnyEvent::Handle> object associated with the |
318 | callback will receive the L<AnyEvent::Handle> object associated with the |
240 | connection. In error cases, C<undef> will be passed. When there is no body |
319 | connection. In error cases, C<undef> will be passed. When there is no body |
241 | (e.g. status C<304>), the empty string will be passed. |
320 | (e.g. status C<304>), the empty string will be passed. |
242 | |
321 | |
243 | The handle object might or might not be in TLS mode, might be connected to |
322 | The handle object might or might not be in TLS mode, might be connected |
244 | a proxy, be a persistent connection etc., and configured in unspecified |
323 | to a proxy, be a persistent connection, use chunked transfer encoding |
245 | ways. The user is responsible for this handle (it will not be used by this |
324 | etc., and configured in unspecified ways. The user is responsible for this |
246 | module anymore). |
325 | handle (it will not be used by this module anymore). |
247 | |
326 | |
248 | This is useful with some push-type services, where, after the initial |
327 | This is useful with some push-type services, where, after the initial |
249 | headers, an interactive protocol is used (typical example would be the |
328 | headers, an interactive protocol is used (typical example would be the |
250 | push-style twitter API which starts a JSON/XML stream). |
329 | push-style twitter API which starts a JSON/XML stream). |
251 | |
330 | |
252 | If you think you need this, first have a look at C<on_body>, to see if |
331 | If you think you need this, first have a look at C<on_body>, to see if |
253 | that doesn't solve your problem in a better way. |
332 | that doesn't solve your problem in a better way. |
254 | |
333 | |
|
|
334 | =item persistent => $boolean |
|
|
335 | |
|
|
336 | Try to create/reuse a persistent connection. When this flag is set |
|
|
337 | (default: true for idempotent requests, false for all others), then |
|
|
338 | C<http_request> tries to re-use an existing (previously-created) |
|
|
339 | persistent connection to the host and, failing that, tries to create a new |
|
|
340 | one. |
|
|
341 | |
|
|
342 | Requests failing in certain ways will be automatically retried once, which |
|
|
343 | is dangerous for non-idempotent requests, which is why it defaults to off |
|
|
344 | for them. The reason for this is because the bozos who designed HTTP/1.1 |
|
|
345 | made it impossible to distinguish between a fatal error and a normal |
|
|
346 | connection timeout, so you never know whether there was a problem with |
|
|
347 | your request or not. |
|
|
348 | |
|
|
349 | When reusing an existent connection, many parameters (such as TLS context) |
|
|
350 | will be ignored. See the C<session> parameter for a workaround. |
|
|
351 | |
|
|
352 | =item keepalive => $boolean |
|
|
353 | |
|
|
354 | Only used when C<persistent> is also true. This parameter decides whether |
|
|
355 | C<http_request> tries to handshake a HTTP/1.0-style keep-alive connection |
|
|
356 | (as opposed to only a HTTP/1.1 persistent connection). |
|
|
357 | |
|
|
358 | The default is true, except when using a proxy, in which case it defaults |
|
|
359 | to false, as HTTP/1.0 proxies cannot support this in a meaningful way. |
|
|
360 | |
|
|
361 | =item handle_params => { key => value ... } |
|
|
362 | |
|
|
363 | The key-value pairs in this hash will be passed to any L<AnyEvent::Handle> |
|
|
364 | constructor that is called - not all requests will create a handle, and |
|
|
365 | sometimes more than one is created, so this parameter is only good for |
|
|
366 | setting hints. |
|
|
367 | |
|
|
368 | Example: set the maximum read size to 4096, to potentially conserve memory |
|
|
369 | at the cost of speed. |
|
|
370 | |
|
|
371 | handle_params => { |
|
|
372 | max_read_size => 4096, |
|
|
373 | }, |
|
|
374 | |
255 | =back |
375 | =back |
256 | |
376 | |
257 | Example: make a simple HTTP GET request for http://www.nethype.de/ |
377 | Example: do a simple HTTP GET request for http://www.nethype.de/ and print |
|
|
378 | the response body. |
258 | |
379 | |
259 | http_request GET => "http://www.nethype.de/", sub { |
380 | http_request GET => "http://www.nethype.de/", sub { |
260 | my ($body, $hdr) = @_; |
381 | my ($body, $hdr) = @_; |
261 | print "$body\n"; |
382 | print "$body\n"; |
262 | }; |
383 | }; |
263 | |
384 | |
264 | Example: make a HTTP HEAD request on https://www.google.com/, use a |
385 | Example: do a HTTP HEAD request on https://www.google.com/, use a |
265 | timeout of 30 seconds. |
386 | timeout of 30 seconds. |
266 | |
387 | |
267 | http_request |
388 | http_request |
268 | GET => "https://www.google.com", |
389 | HEAD => "https://www.google.com", |
|
|
390 | headers => { "user-agent" => "MySearchClient 1.0" }, |
269 | timeout => 30, |
391 | timeout => 30, |
270 | sub { |
392 | sub { |
271 | my ($body, $hdr) = @_; |
393 | my ($body, $hdr) = @_; |
272 | use Data::Dumper; |
394 | use Data::Dumper; |
273 | print Dumper $hdr; |
395 | print Dumper $hdr; |
274 | } |
396 | } |
275 | ; |
397 | ; |
276 | |
398 | |
277 | Example: make another simple HTTP GET request, but immediately try to |
399 | Example: do another simple HTTP GET request, but immediately try to |
278 | cancel it. |
400 | cancel it. |
279 | |
401 | |
280 | my $request = http_request GET => "http://www.nethype.de/", sub { |
402 | my $request = http_request GET => "http://www.nethype.de/", sub { |
281 | my ($body, $hdr) = @_; |
403 | my ($body, $hdr) = @_; |
282 | print "$body\n"; |
404 | print "$body\n"; |
283 | }; |
405 | }; |
284 | |
406 | |
285 | undef $request; |
407 | undef $request; |
286 | |
408 | |
287 | =cut |
409 | =cut |
|
|
410 | |
|
|
411 | ############################################################################# |
|
|
412 | # wait queue/slots |
288 | |
413 | |
289 | sub _slot_schedule; |
414 | sub _slot_schedule; |
290 | sub _slot_schedule($) { |
415 | sub _slot_schedule($) { |
291 | my $host = shift; |
416 | my $host = shift; |
292 | |
417 | |
… | |
… | |
314 | push @{ $CO_SLOT{$_[0]}[1] }, $_[1]; |
439 | push @{ $CO_SLOT{$_[0]}[1] }, $_[1]; |
315 | |
440 | |
316 | _slot_schedule $_[0]; |
441 | _slot_schedule $_[0]; |
317 | } |
442 | } |
318 | |
443 | |
319 | our $qr_nl = qr{\015?\012}; |
444 | ############################################################################# |
|
|
445 | # cookie handling |
|
|
446 | |
|
|
447 | # expire cookies |
|
|
448 | sub cookie_jar_expire($;$) { |
|
|
449 | my ($jar, $session_end) = @_; |
|
|
450 | |
|
|
451 | %$jar = () if $jar->{version} != 1; |
|
|
452 | |
|
|
453 | my $anow = AE::now; |
|
|
454 | |
|
|
455 | while (my ($chost, $paths) = each %$jar) { |
|
|
456 | next unless ref $paths; |
|
|
457 | |
|
|
458 | while (my ($cpath, $cookies) = each %$paths) { |
|
|
459 | while (my ($cookie, $kv) = each %$cookies) { |
|
|
460 | if (exists $kv->{_expires}) { |
|
|
461 | delete $cookies->{$cookie} |
|
|
462 | if $anow > $kv->{_expires}; |
|
|
463 | } elsif ($session_end) { |
|
|
464 | delete $cookies->{$cookie}; |
|
|
465 | } |
|
|
466 | } |
|
|
467 | |
|
|
468 | delete $paths->{$cpath} |
|
|
469 | unless %$cookies; |
|
|
470 | } |
|
|
471 | |
|
|
472 | delete $jar->{$chost} |
|
|
473 | unless %$paths; |
|
|
474 | } |
|
|
475 | } |
|
|
476 | |
|
|
477 | # extract cookies from jar |
|
|
478 | sub cookie_jar_extract($$$$) { |
|
|
479 | my ($jar, $scheme, $host, $path) = @_; |
|
|
480 | |
|
|
481 | %$jar = () if $jar->{version} != 1; |
|
|
482 | |
|
|
483 | my @cookies; |
|
|
484 | |
|
|
485 | while (my ($chost, $paths) = each %$jar) { |
|
|
486 | next unless ref $paths; |
|
|
487 | |
|
|
488 | if ($chost =~ /^\./) { |
|
|
489 | next unless $chost eq substr $host, -length $chost; |
|
|
490 | } elsif ($chost =~ /\./) { |
|
|
491 | next unless $chost eq $host; |
|
|
492 | } else { |
|
|
493 | next; |
|
|
494 | } |
|
|
495 | |
|
|
496 | while (my ($cpath, $cookies) = each %$paths) { |
|
|
497 | next unless $cpath eq substr $path, 0, length $cpath; |
|
|
498 | |
|
|
499 | while (my ($cookie, $kv) = each %$cookies) { |
|
|
500 | next if $scheme ne "https" && exists $kv->{secure}; |
|
|
501 | |
|
|
502 | if (exists $kv->{_expires} and AE::now > $kv->{_expires}) { |
|
|
503 | delete $cookies->{$cookie}; |
|
|
504 | next; |
|
|
505 | } |
|
|
506 | |
|
|
507 | my $value = $kv->{value}; |
|
|
508 | |
|
|
509 | if ($value =~ /[=;,[:space:]]/) { |
|
|
510 | $value =~ s/([\\"])/\\$1/g; |
|
|
511 | $value = "\"$value\""; |
|
|
512 | } |
|
|
513 | |
|
|
514 | push @cookies, "$cookie=$value"; |
|
|
515 | } |
|
|
516 | } |
|
|
517 | } |
|
|
518 | |
|
|
519 | \@cookies |
|
|
520 | } |
|
|
521 | |
|
|
522 | # parse set_cookie header into jar |
|
|
523 | sub cookie_jar_set_cookie($$$$) { |
|
|
524 | my ($jar, $set_cookie, $host, $date) = @_; |
|
|
525 | |
|
|
526 | my $anow = int AE::now; |
|
|
527 | my $snow; # server-now |
|
|
528 | |
|
|
529 | for ($set_cookie) { |
|
|
530 | # parse NAME=VALUE |
|
|
531 | my @kv; |
|
|
532 | |
|
|
533 | # expires is not http-compliant in the original cookie-spec, |
|
|
534 | # we support the official date format and some extensions |
|
|
535 | while ( |
|
|
536 | m{ |
|
|
537 | \G\s* |
|
|
538 | (?: |
|
|
539 | expires \s*=\s* ([A-Z][a-z][a-z]+,\ [^,;]+) |
|
|
540 | | ([^=;,[:space:]]+) (?: \s*=\s* (?: "((?:[^\\"]+|\\.)*)" | ([^;,[:space:]]*) ) )? |
|
|
541 | ) |
|
|
542 | }gcxsi |
|
|
543 | ) { |
|
|
544 | my $name = $2; |
|
|
545 | my $value = $4; |
|
|
546 | |
|
|
547 | if (defined $1) { |
|
|
548 | # expires |
|
|
549 | $name = "expires"; |
|
|
550 | $value = $1; |
|
|
551 | } elsif (defined $3) { |
|
|
552 | # quoted |
|
|
553 | $value = $3; |
|
|
554 | $value =~ s/\\(.)/$1/gs; |
|
|
555 | } |
|
|
556 | |
|
|
557 | push @kv, @kv ? lc $name : $name, $value; |
|
|
558 | |
|
|
559 | last unless /\G\s*;/gc; |
|
|
560 | } |
|
|
561 | |
|
|
562 | last unless @kv; |
|
|
563 | |
|
|
564 | my $name = shift @kv; |
|
|
565 | my %kv = (value => shift @kv, @kv); |
|
|
566 | |
|
|
567 | if (exists $kv{"max-age"}) { |
|
|
568 | $kv{_expires} = $anow + delete $kv{"max-age"}; |
|
|
569 | } elsif (exists $kv{expires}) { |
|
|
570 | $snow ||= parse_date ($date) || $anow; |
|
|
571 | $kv{_expires} = $anow + (parse_date (delete $kv{expires}) - $snow); |
|
|
572 | } else { |
|
|
573 | delete $kv{_expires}; |
|
|
574 | } |
|
|
575 | |
|
|
576 | my $cdom; |
|
|
577 | my $cpath = (delete $kv{path}) || "/"; |
|
|
578 | |
|
|
579 | if (exists $kv{domain}) { |
|
|
580 | $cdom = delete $kv{domain}; |
|
|
581 | |
|
|
582 | $cdom =~ s/^\.?/./; # make sure it starts with a "." |
|
|
583 | |
|
|
584 | next if $cdom =~ /\.$/; |
|
|
585 | |
|
|
586 | # this is not rfc-like and not netscape-like. go figure. |
|
|
587 | my $ndots = $cdom =~ y/.//; |
|
|
588 | next if $ndots < ($cdom =~ /\.[^.][^.]\.[^.][^.]$/ ? 3 : 2); |
|
|
589 | } else { |
|
|
590 | $cdom = $host; |
|
|
591 | } |
|
|
592 | |
|
|
593 | # store it |
|
|
594 | $jar->{version} = 1; |
|
|
595 | $jar->{lc $cdom}{$cpath}{$name} = \%kv; |
|
|
596 | |
|
|
597 | redo if /\G\s*,/gc; |
|
|
598 | } |
|
|
599 | } |
|
|
600 | |
|
|
601 | ############################################################################# |
|
|
602 | # keepalive/persistent connection cache |
|
|
603 | |
|
|
604 | # fetch a connection from the keepalive cache |
|
|
605 | sub ka_fetch($) { |
|
|
606 | my $ka_key = shift; |
|
|
607 | |
|
|
608 | my $hdl = pop @{ $KA_CACHE{$ka_key} }; # currently we reuse the MOST RECENTLY USED connection |
|
|
609 | delete $KA_CACHE{$ka_key} |
|
|
610 | unless @{ $KA_CACHE{$ka_key} }; |
|
|
611 | |
|
|
612 | $hdl |
|
|
613 | } |
|
|
614 | |
|
|
615 | sub ka_store($$) { |
|
|
616 | my ($ka_key, $hdl) = @_; |
|
|
617 | |
|
|
618 | my $kaa = $KA_CACHE{$ka_key} ||= []; |
|
|
619 | |
|
|
620 | my $destroy = sub { |
|
|
621 | my @ka = grep $_ != $hdl, @{ $KA_CACHE{$ka_key} }; |
|
|
622 | |
|
|
623 | $hdl->destroy; |
|
|
624 | |
|
|
625 | @ka |
|
|
626 | ? $KA_CACHE{$ka_key} = \@ka |
|
|
627 | : delete $KA_CACHE{$ka_key}; |
|
|
628 | }; |
|
|
629 | |
|
|
630 | # on error etc., destroy |
|
|
631 | $hdl->on_error ($destroy); |
|
|
632 | $hdl->on_eof ($destroy); |
|
|
633 | $hdl->on_read ($destroy); |
|
|
634 | $hdl->timeout ($PERSISTENT_TIMEOUT); |
|
|
635 | |
|
|
636 | push @$kaa, $hdl; |
|
|
637 | shift @$kaa while @$kaa > $MAX_PER_HOST; |
|
|
638 | } |
|
|
639 | |
|
|
640 | ############################################################################# |
|
|
641 | # utilities |
|
|
642 | |
|
|
643 | # continue to parse $_ for headers and place them into the arg |
|
|
644 | sub _parse_hdr() { |
|
|
645 | my %hdr; |
|
|
646 | |
|
|
647 | # things seen, not parsed: |
|
|
648 | # p3pP="NON CUR OTPi OUR NOR UNI" |
|
|
649 | |
|
|
650 | $hdr{lc $1} .= ",$2" |
|
|
651 | while /\G |
|
|
652 | ([^:\000-\037]*): |
|
|
653 | [\011\040]* |
|
|
654 | ((?: [^\012]+ | \012[\011\040] )*) |
|
|
655 | \012 |
|
|
656 | /gxc; |
|
|
657 | |
|
|
658 | /\G$/ |
|
|
659 | or return; |
|
|
660 | |
|
|
661 | # remove the "," prefix we added to all headers above |
|
|
662 | substr $_, 0, 1, "" |
|
|
663 | for values %hdr; |
|
|
664 | |
|
|
665 | \%hdr |
|
|
666 | } |
|
|
667 | |
|
|
668 | ############################################################################# |
|
|
669 | # http_get |
|
|
670 | |
320 | our $qr_nlnl = qr{(?<![^\012])\015?\012}; |
671 | our $qr_nlnl = qr{(?<![^\012])\015?\012}; |
321 | |
672 | |
322 | our $TLS_CTX_LOW = { cache => 1, sslv2 => 1 }; |
673 | our $TLS_CTX_LOW = { cache => 1, sslv2 => 1 }; |
323 | our $TLS_CTX_HIGH = { cache => 1, verify => 1, verify_peername => "https" }; |
674 | our $TLS_CTX_HIGH = { cache => 1, verify => 1, verify_peername => "https" }; |
|
|
675 | |
|
|
676 | # maybe it should just become a normal object :/ |
|
|
677 | |
|
|
678 | sub _destroy_state(\%) { |
|
|
679 | my ($state) = @_; |
|
|
680 | |
|
|
681 | $state->{handle}->destroy if $state->{handle}; |
|
|
682 | %$state = (); |
|
|
683 | } |
|
|
684 | |
|
|
685 | sub _error(\%$$) { |
|
|
686 | my ($state, $cb, $hdr) = @_; |
|
|
687 | |
|
|
688 | &_destroy_state ($state); |
|
|
689 | |
|
|
690 | $cb->(undef, $hdr); |
|
|
691 | () |
|
|
692 | } |
324 | |
693 | |
325 | sub http_request($$@) { |
694 | sub http_request($$@) { |
326 | my $cb = pop; |
695 | my $cb = pop; |
327 | my ($method, $url, %arg) = @_; |
696 | my ($method, $url, %arg) = @_; |
328 | |
697 | |
… | |
… | |
337 | while (my ($k, $v) = each %$hdr) { |
706 | while (my ($k, $v) = each %$hdr) { |
338 | $hdr{lc $k} = $v; |
707 | $hdr{lc $k} = $v; |
339 | } |
708 | } |
340 | } |
709 | } |
341 | |
710 | |
|
|
711 | # pseudo headers for all subsequent responses |
|
|
712 | my @pseudo = (URL => $url); |
|
|
713 | push @pseudo, Redirect => delete $arg{Redirect} if exists $arg{Redirect}; |
|
|
714 | |
342 | my $recurse = exists $arg{recurse} ? delete $arg{recurse} : $MAX_RECURSE; |
715 | my $recurse = exists $arg{recurse} ? delete $arg{recurse} : $MAX_RECURSE; |
343 | |
716 | |
344 | return $cb->(undef, { Status => 599, Reason => "Too many redirections", URL => $url }) |
717 | return $cb->(undef, { @pseudo, Status => 599, Reason => "Too many redirections" }) |
345 | if $recurse < 0; |
718 | if $recurse < 0; |
346 | |
719 | |
347 | my $proxy = $arg{proxy} || $PROXY; |
720 | my $proxy = exists $arg{proxy} ? $arg{proxy} : $PROXY; |
348 | my $timeout = $arg{timeout} || $TIMEOUT; |
721 | my $timeout = $arg{timeout} || $TIMEOUT; |
349 | |
722 | |
350 | my ($uscheme, $uauthority, $upath, $query, $fragment) = |
723 | my ($uscheme, $uauthority, $upath, $query, undef) = # ignore fragment |
351 | $url =~ m|(?:([^:/?#]+):)?(?://([^/?#]*))?([^?#]*)(?:\?([^#]*))?(?:#(.*))?|; |
724 | $url =~ m|^([^:]+):(?://([^/?#]*))?([^?#]*)(?:(\?[^#]*))?(?:#(.*))?$|; |
352 | |
725 | |
353 | $uscheme = lc $uscheme; |
726 | $uscheme = lc $uscheme; |
354 | |
727 | |
355 | my $uport = $uscheme eq "http" ? 80 |
728 | my $uport = $uscheme eq "http" ? 80 |
356 | : $uscheme eq "https" ? 443 |
729 | : $uscheme eq "https" ? 443 |
357 | : return $cb->(undef, { Status => 599, Reason => "Only http and https URL schemes supported", URL => $url }); |
730 | : return $cb->(undef, { @pseudo, Status => 599, Reason => "Only http and https URL schemes supported" }); |
358 | |
731 | |
359 | $uauthority =~ /^(?: .*\@ )? ([^\@:]+) (?: : (\d+) )?$/x |
732 | $uauthority =~ /^(?: .*\@ )? ([^\@:]+) (?: : (\d+) )?$/x |
360 | or return $cb->(undef, { Status => 599, Reason => "Unparsable URL", URL => $url }); |
733 | or return $cb->(undef, { @pseudo, Status => 599, Reason => "Unparsable URL" }); |
361 | |
734 | |
362 | my $uhost = $1; |
735 | my $uhost = lc $1; |
363 | $uport = $2 if defined $2; |
736 | $uport = $2 if defined $2; |
364 | |
737 | |
365 | $hdr{host} = defined $2 ? "$uhost:$2" : "$uhost"; |
738 | $hdr{host} = defined $2 ? "$uhost:$2" : "$uhost" |
|
|
739 | unless exists $hdr{host}; |
366 | |
740 | |
367 | $uhost =~ s/^\[(.*)\]$/$1/; |
741 | $uhost =~ s/^\[(.*)\]$/$1/; |
368 | $upath .= "?$query" if length $query; |
742 | $upath .= $query if length $query; |
369 | |
743 | |
370 | $upath =~ s%^/?%/%; |
744 | $upath =~ s%^/?%/%; |
371 | |
745 | |
372 | # cookie processing |
746 | # cookie processing |
373 | if (my $jar = $arg{cookie_jar}) { |
747 | if (my $jar = $arg{cookie_jar}) { |
374 | %$jar = () if $jar->{version} != 1; |
748 | my $cookies = cookie_jar_extract $jar, $uscheme, $uhost, $upath; |
375 | |
749 | |
376 | my @cookie; |
|
|
377 | |
|
|
378 | while (my ($chost, $v) = each %$jar) { |
|
|
379 | if ($chost =~ /^\./) { |
|
|
380 | next unless $chost eq substr $uhost, -length $chost; |
|
|
381 | } elsif ($chost =~ /\./) { |
|
|
382 | next unless $chost eq $uhost; |
|
|
383 | } else { |
|
|
384 | next; |
|
|
385 | } |
|
|
386 | |
|
|
387 | while (my ($cpath, $v) = each %$v) { |
|
|
388 | next unless $cpath eq substr $upath, 0, length $cpath; |
|
|
389 | |
|
|
390 | while (my ($k, $v) = each %$v) { |
|
|
391 | next if $uscheme ne "https" && exists $v->{secure}; |
|
|
392 | my $value = $v->{value}; |
|
|
393 | $value =~ s/([\\"])/\\$1/g; |
|
|
394 | push @cookie, "$k=\"$value\""; |
|
|
395 | } |
|
|
396 | } |
|
|
397 | } |
|
|
398 | |
|
|
399 | $hdr{cookie} = join "; ", @cookie |
750 | $hdr{cookie} = join "; ", @$cookies |
400 | if @cookie; |
751 | if @$cookies; |
401 | } |
752 | } |
402 | |
753 | |
403 | my ($rhost, $rport, $rscheme, $rpath); # request host, port, path |
754 | my ($rhost, $rport, $rscheme, $rpath); # request host, port, path |
404 | |
755 | |
405 | if ($proxy) { |
756 | if ($proxy) { |
… | |
… | |
408 | $rscheme = "http" unless defined $rscheme; |
759 | $rscheme = "http" unless defined $rscheme; |
409 | |
760 | |
410 | # don't support https requests over https-proxy transport, |
761 | # don't support https requests over https-proxy transport, |
411 | # can't be done with tls as spec'ed, unless you double-encrypt. |
762 | # can't be done with tls as spec'ed, unless you double-encrypt. |
412 | $rscheme = "http" if $uscheme eq "https" && $rscheme eq "https"; |
763 | $rscheme = "http" if $uscheme eq "https" && $rscheme eq "https"; |
|
|
764 | |
|
|
765 | $rhost = lc $rhost; |
|
|
766 | $rscheme = lc $rscheme; |
413 | } else { |
767 | } else { |
414 | ($rhost, $rport, $rscheme, $rpath) = ($uhost, $uport, $uscheme, $upath); |
768 | ($rhost, $rport, $rscheme, $rpath) = ($uhost, $uport, $uscheme, $upath); |
415 | } |
769 | } |
416 | |
770 | |
417 | # leave out fragment and query string, just a heuristic |
771 | # leave out fragment and query string, just a heuristic |
418 | $hdr{referer} ||= "$uscheme://$uauthority$upath" unless exists $hdr{referer}; |
772 | $hdr{referer} = "$uscheme://$uauthority$upath" unless exists $hdr{referer}; |
419 | $hdr{"user-agent"} ||= $USERAGENT unless exists $hdr{"user-agent"}; |
773 | $hdr{"user-agent"} = $USERAGENT unless exists $hdr{"user-agent"}; |
420 | |
774 | |
421 | $hdr{"content-length"} = length $arg{body}; |
775 | $hdr{"content-length"} = length $arg{body} |
|
|
776 | if length $arg{body} || $method ne "GET"; |
|
|
777 | |
|
|
778 | my $idempotent = $method =~ /^(?:GET|HEAD|PUT|DELETE|OPTIONS|TRACE)$/; |
|
|
779 | |
|
|
780 | # default value for keepalive is true iff the request is for an idempotent method |
|
|
781 | my $persistent = exists $arg{persistent} ? !!$arg{persistent} : $idempotent; |
|
|
782 | my $keepalive = exists $arg{keepalive} ? !!$arg{keepalive} : !$proxy; |
|
|
783 | my $was_persistent; # true if this is actually a recycled connection |
|
|
784 | |
|
|
785 | # the key to use in the keepalive cache |
|
|
786 | my $ka_key = "$uscheme\x00$uhost\x00$uport\x00$arg{sessionid}"; |
|
|
787 | |
|
|
788 | $hdr{connection} = ($persistent ? $keepalive ? "keep-alive " : "" : "close ") . "Te"; #1.1 |
|
|
789 | $hdr{te} = "trailers" unless exists $hdr{te}; #1.1 |
422 | |
790 | |
423 | my %state = (connect_guard => 1); |
791 | my %state = (connect_guard => 1); |
|
|
792 | |
|
|
793 | my $ae_error = 595; # connecting |
|
|
794 | |
|
|
795 | # handle actual, non-tunneled, request |
|
|
796 | my $handle_actual_request = sub { |
|
|
797 | $ae_error = 596; # request phase |
|
|
798 | |
|
|
799 | my $hdl = $state{handle}; |
|
|
800 | |
|
|
801 | $hdl->starttls ("connect") if $uscheme eq "https" && !exists $hdl->{tls}; |
|
|
802 | |
|
|
803 | # send request |
|
|
804 | $hdl->push_write ( |
|
|
805 | "$method $rpath HTTP/1.1\015\012" |
|
|
806 | . (join "", map "\u$_: $hdr{$_}\015\012", grep defined $hdr{$_}, keys %hdr) |
|
|
807 | . "\015\012" |
|
|
808 | . (delete $arg{body}) |
|
|
809 | ); |
|
|
810 | |
|
|
811 | # return if error occured during push_write() |
|
|
812 | return unless %state; |
|
|
813 | |
|
|
814 | # reduce memory usage, save a kitten, also re-use it for the response headers. |
|
|
815 | %hdr = (); |
|
|
816 | |
|
|
817 | # status line and headers |
|
|
818 | $state{read_response} = sub { |
|
|
819 | return unless %state; |
|
|
820 | |
|
|
821 | for ("$_[1]") { |
|
|
822 | y/\015//d; # weed out any \015, as they show up in the weirdest of places. |
|
|
823 | |
|
|
824 | /^HTTP\/0*([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\012]*) )? \012/gxci |
|
|
825 | or return _error %state, $cb, { @pseudo, Status => 599, Reason => "Invalid server response" }; |
|
|
826 | |
|
|
827 | # 100 Continue handling |
|
|
828 | # should not happen as we don't send expect: 100-continue, |
|
|
829 | # but we handle it just in case. |
|
|
830 | # since we send the request body regardless, if we get an error |
|
|
831 | # we are out of-sync, which we currently do NOT handle correctly. |
|
|
832 | return $state{handle}->push_read (line => $qr_nlnl, $state{read_response}) |
|
|
833 | if $2 eq 100; |
|
|
834 | |
|
|
835 | push @pseudo, |
|
|
836 | HTTPVersion => $1, |
|
|
837 | Status => $2, |
|
|
838 | Reason => $3, |
|
|
839 | ; |
|
|
840 | |
|
|
841 | my $hdr = _parse_hdr |
|
|
842 | or return _error %state, $cb, { @pseudo, Status => 599, Reason => "Garbled response headers" }; |
|
|
843 | |
|
|
844 | %hdr = (%$hdr, @pseudo); |
|
|
845 | } |
|
|
846 | |
|
|
847 | # redirect handling |
|
|
848 | # microsoft and other shitheads don't give a shit for following standards, |
|
|
849 | # try to support some common forms of broken Location headers. |
|
|
850 | if ($hdr{location} !~ /^(?: $ | [^:\/?\#]+ : )/x) { |
|
|
851 | $hdr{location} =~ s/^\.\/+//; |
|
|
852 | |
|
|
853 | my $url = "$rscheme://$uhost:$uport"; |
|
|
854 | |
|
|
855 | unless ($hdr{location} =~ s/^\///) { |
|
|
856 | $url .= $upath; |
|
|
857 | $url =~ s/\/[^\/]*$//; |
|
|
858 | } |
|
|
859 | |
|
|
860 | $hdr{location} = "$url/$hdr{location}"; |
|
|
861 | } |
|
|
862 | |
|
|
863 | my $redirect; |
|
|
864 | |
|
|
865 | if ($recurse) { |
|
|
866 | my $status = $hdr{Status}; |
|
|
867 | |
|
|
868 | # industry standard is to redirect POST as GET for |
|
|
869 | # 301, 302 and 303, in contrast to HTTP/1.0 and 1.1. |
|
|
870 | # also, the UA should ask the user for 301 and 307 and POST, |
|
|
871 | # industry standard seems to be to simply follow. |
|
|
872 | # we go with the industry standard. |
|
|
873 | if ($status == 301 or $status == 302 or $status == 303) { |
|
|
874 | # HTTP/1.1 is unclear on how to mutate the method |
|
|
875 | $method = "GET" unless $method eq "HEAD"; |
|
|
876 | $redirect = 1; |
|
|
877 | } elsif ($status == 307) { |
|
|
878 | $redirect = 1; |
|
|
879 | } |
|
|
880 | } |
|
|
881 | |
|
|
882 | my $finish = sub { # ($data, $err_status, $err_reason[, $persistent]) |
|
|
883 | if ($state{handle}) { |
|
|
884 | # handle keepalive |
|
|
885 | if ( |
|
|
886 | $persistent |
|
|
887 | && $_[3] |
|
|
888 | && ($hdr{HTTPVersion} < 1.1 |
|
|
889 | ? $hdr{connection} =~ /\bkeep-?alive\b/i |
|
|
890 | : $hdr{connection} !~ /\bclose\b/i) |
|
|
891 | ) { |
|
|
892 | ka_store $ka_key, delete $state{handle}; |
|
|
893 | } else { |
|
|
894 | # no keepalive, destroy the handle |
|
|
895 | $state{handle}->destroy; |
|
|
896 | } |
|
|
897 | } |
|
|
898 | |
|
|
899 | %state = (); |
|
|
900 | |
|
|
901 | if (defined $_[1]) { |
|
|
902 | $hdr{OrigStatus} = $hdr{Status}; $hdr{Status} = $_[1]; |
|
|
903 | $hdr{OrigReason} = $hdr{Reason}; $hdr{Reason} = $_[2]; |
|
|
904 | } |
|
|
905 | |
|
|
906 | # set-cookie processing |
|
|
907 | if ($arg{cookie_jar}) { |
|
|
908 | cookie_jar_set_cookie $arg{cookie_jar}, $hdr{"set-cookie"}, $uhost, $hdr{date}; |
|
|
909 | } |
|
|
910 | |
|
|
911 | if ($redirect && exists $hdr{location}) { |
|
|
912 | # we ignore any errors, as it is very common to receive |
|
|
913 | # Content-Length != 0 but no actual body |
|
|
914 | # we also access %hdr, as $_[1] might be an erro |
|
|
915 | $state{recurse} = |
|
|
916 | http_request ( |
|
|
917 | $method => $hdr{location}, |
|
|
918 | %arg, |
|
|
919 | recurse => $recurse - 1, |
|
|
920 | Redirect => [$_[0], \%hdr], |
|
|
921 | sub { |
|
|
922 | %state = (); |
|
|
923 | &$cb |
|
|
924 | }, |
|
|
925 | ); |
|
|
926 | } else { |
|
|
927 | $cb->($_[0], \%hdr); |
|
|
928 | } |
|
|
929 | }; |
|
|
930 | |
|
|
931 | $ae_error = 597; # body phase |
|
|
932 | |
|
|
933 | my $chunked = $hdr{"transfer-encoding"} =~ /\bchunked\b/i; # not quite correct... |
|
|
934 | |
|
|
935 | my $len = $chunked ? undef : $hdr{"content-length"}; |
|
|
936 | |
|
|
937 | # body handling, many different code paths |
|
|
938 | # - no body expected |
|
|
939 | # - want_body_handle |
|
|
940 | # - te chunked |
|
|
941 | # - 2x length known (with or without on_body) |
|
|
942 | # - 2x length not known (with or without on_body) |
|
|
943 | if (!$redirect && $arg{on_header} && !$arg{on_header}(\%hdr)) { |
|
|
944 | $finish->(undef, 598 => "Request cancelled by on_header"); |
|
|
945 | } elsif ( |
|
|
946 | $hdr{Status} =~ /^(?:1..|204|205|304)$/ |
|
|
947 | or $method eq "HEAD" |
|
|
948 | or (defined $len && $len == 0) # == 0, not !, because "0 " is true |
|
|
949 | ) { |
|
|
950 | # no body |
|
|
951 | $finish->("", undef, undef, 1); |
|
|
952 | |
|
|
953 | } elsif (!$redirect && $arg{want_body_handle}) { |
|
|
954 | $_[0]->on_eof (undef); |
|
|
955 | $_[0]->on_error (undef); |
|
|
956 | $_[0]->on_read (undef); |
|
|
957 | |
|
|
958 | $finish->(delete $state{handle}); |
|
|
959 | |
|
|
960 | } elsif ($chunked) { |
|
|
961 | my $cl = 0; |
|
|
962 | my $body = ""; |
|
|
963 | my $on_body = $arg{on_body} || sub { $body .= shift; 1 }; |
|
|
964 | |
|
|
965 | $state{read_chunk} = sub { |
|
|
966 | $_[1] =~ /^([0-9a-fA-F]+)/ |
|
|
967 | or return $finish->(undef, $ae_error => "Garbled chunked transfer encoding"); |
|
|
968 | |
|
|
969 | my $len = hex $1; |
|
|
970 | |
|
|
971 | if ($len) { |
|
|
972 | $cl += $len; |
|
|
973 | |
|
|
974 | $_[0]->push_read (chunk => $len, sub { |
|
|
975 | $on_body->($_[1], \%hdr) |
|
|
976 | or return $finish->(undef, 598 => "Request cancelled by on_body"); |
|
|
977 | |
|
|
978 | $_[0]->push_read (line => sub { |
|
|
979 | length $_[1] |
|
|
980 | and return $finish->(undef, $ae_error => "Garbled chunked transfer encoding"); |
|
|
981 | $_[0]->push_read (line => $state{read_chunk}); |
|
|
982 | }); |
|
|
983 | }); |
|
|
984 | } else { |
|
|
985 | $hdr{"content-length"} ||= $cl; |
|
|
986 | |
|
|
987 | $_[0]->push_read (line => $qr_nlnl, sub { |
|
|
988 | if (length $_[1]) { |
|
|
989 | for ("$_[1]") { |
|
|
990 | y/\015//d; # weed out any \015, as they show up in the weirdest of places. |
|
|
991 | |
|
|
992 | my $hdr = _parse_hdr |
|
|
993 | or return $finish->(undef, $ae_error => "Garbled response trailers"); |
|
|
994 | |
|
|
995 | %hdr = (%hdr, %$hdr); |
|
|
996 | } |
|
|
997 | } |
|
|
998 | |
|
|
999 | $finish->($body, undef, undef, 1); |
|
|
1000 | }); |
|
|
1001 | } |
|
|
1002 | }; |
|
|
1003 | |
|
|
1004 | $_[0]->push_read (line => $state{read_chunk}); |
|
|
1005 | |
|
|
1006 | } elsif ($arg{on_body}) { |
|
|
1007 | if (defined $len) { |
|
|
1008 | $_[0]->on_read (sub { |
|
|
1009 | $len -= length $_[0]{rbuf}; |
|
|
1010 | |
|
|
1011 | $arg{on_body}(delete $_[0]{rbuf}, \%hdr) |
|
|
1012 | or return $finish->(undef, 598 => "Request cancelled by on_body"); |
|
|
1013 | |
|
|
1014 | $len > 0 |
|
|
1015 | or $finish->("", undef, undef, 1); |
|
|
1016 | }); |
|
|
1017 | } else { |
|
|
1018 | $_[0]->on_eof (sub { |
|
|
1019 | $finish->(""); |
|
|
1020 | }); |
|
|
1021 | $_[0]->on_read (sub { |
|
|
1022 | $arg{on_body}(delete $_[0]{rbuf}, \%hdr) |
|
|
1023 | or $finish->(undef, 598 => "Request cancelled by on_body"); |
|
|
1024 | }); |
|
|
1025 | } |
|
|
1026 | } else { |
|
|
1027 | $_[0]->on_eof (undef); |
|
|
1028 | |
|
|
1029 | if (defined $len) { |
|
|
1030 | $_[0]->on_read (sub { |
|
|
1031 | $finish->((substr delete $_[0]{rbuf}, 0, $len, ""), undef, undef, 1) |
|
|
1032 | if $len <= length $_[0]{rbuf}; |
|
|
1033 | }); |
|
|
1034 | } else { |
|
|
1035 | $_[0]->on_error (sub { |
|
|
1036 | ($! == Errno::EPIPE || !$!) |
|
|
1037 | ? $finish->(delete $_[0]{rbuf}) |
|
|
1038 | : $finish->(undef, $ae_error => $_[2]); |
|
|
1039 | }); |
|
|
1040 | $_[0]->on_read (sub { }); |
|
|
1041 | } |
|
|
1042 | } |
|
|
1043 | }; |
|
|
1044 | |
|
|
1045 | # if keepalive is enabled, then the server closing the connection |
|
|
1046 | # before a response can happen legally - we retry on idempotent methods. |
|
|
1047 | if ($was_persistent && $idempotent) { |
|
|
1048 | my $old_eof = $hdl->{on_eof}; |
|
|
1049 | $hdl->{on_eof} = sub { |
|
|
1050 | _destroy_state %state; |
|
|
1051 | |
|
|
1052 | %state = (); |
|
|
1053 | $state{recurse} = |
|
|
1054 | http_request ( |
|
|
1055 | $method => $url, |
|
|
1056 | %arg, |
|
|
1057 | recurse => $recurse - 1, |
|
|
1058 | keepalive => 0, |
|
|
1059 | sub { |
|
|
1060 | %state = (); |
|
|
1061 | &$cb |
|
|
1062 | } |
|
|
1063 | ); |
|
|
1064 | }; |
|
|
1065 | $hdl->on_read (sub { |
|
|
1066 | return unless %state; |
|
|
1067 | |
|
|
1068 | # as soon as we receive something, a connection close |
|
|
1069 | # once more becomes a hard error |
|
|
1070 | $hdl->{on_eof} = $old_eof; |
|
|
1071 | $hdl->push_read (line => $qr_nlnl, $state{read_response}); |
|
|
1072 | }); |
|
|
1073 | } else { |
|
|
1074 | $hdl->push_read (line => $qr_nlnl, $state{read_response}); |
|
|
1075 | } |
|
|
1076 | }; |
|
|
1077 | |
|
|
1078 | my $prepare_handle = sub { |
|
|
1079 | my ($hdl) = $state{handle}; |
|
|
1080 | |
|
|
1081 | $hdl->on_error (sub { |
|
|
1082 | _error %state, $cb, { @pseudo, Status => $ae_error, Reason => $_[2] }; |
|
|
1083 | }); |
|
|
1084 | $hdl->on_eof (sub { |
|
|
1085 | _error %state, $cb, { @pseudo, Status => $ae_error, Reason => "Unexpected end-of-file" }; |
|
|
1086 | }); |
|
|
1087 | $hdl->timeout_reset; |
|
|
1088 | $hdl->timeout ($timeout); |
|
|
1089 | }; |
|
|
1090 | |
|
|
1091 | # connected to proxy (or origin server) |
|
|
1092 | my $connect_cb = sub { |
|
|
1093 | my $fh = shift |
|
|
1094 | or return _error %state, $cb, { @pseudo, Status => $ae_error, Reason => "$!" }; |
|
|
1095 | |
|
|
1096 | return unless delete $state{connect_guard}; |
|
|
1097 | |
|
|
1098 | # get handle |
|
|
1099 | $state{handle} = new AnyEvent::Handle |
|
|
1100 | %{ $arg{handle_params} }, |
|
|
1101 | fh => $fh, |
|
|
1102 | peername => $uhost, |
|
|
1103 | tls_ctx => $arg{tls_ctx}, |
|
|
1104 | ; |
|
|
1105 | |
|
|
1106 | $prepare_handle->(); |
|
|
1107 | |
|
|
1108 | #$state{handle}->starttls ("connect") if $rscheme eq "https"; |
|
|
1109 | |
|
|
1110 | # now handle proxy-CONNECT method |
|
|
1111 | if ($proxy && $uscheme eq "https") { |
|
|
1112 | # oh dear, we have to wrap it into a connect request |
|
|
1113 | |
|
|
1114 | # maybe re-use $uauthority with patched port? |
|
|
1115 | $state{handle}->push_write ("CONNECT $uhost:$uport HTTP/1.0\015\012\015\012"); |
|
|
1116 | $state{handle}->push_read (line => $qr_nlnl, sub { |
|
|
1117 | $_[1] =~ /^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\015\012]*) )?/ix |
|
|
1118 | or return _error %state, $cb, { @pseudo, Status => 599, Reason => "Invalid proxy connect response ($_[1])" }; |
|
|
1119 | |
|
|
1120 | if ($2 == 200) { |
|
|
1121 | $rpath = $upath; |
|
|
1122 | $handle_actual_request->(); |
|
|
1123 | } else { |
|
|
1124 | _error %state, $cb, { @pseudo, Status => $2, Reason => $3 }; |
|
|
1125 | } |
|
|
1126 | }); |
|
|
1127 | } else { |
|
|
1128 | $handle_actual_request->(); |
|
|
1129 | } |
|
|
1130 | }; |
424 | |
1131 | |
425 | _get_slot $uhost, sub { |
1132 | _get_slot $uhost, sub { |
426 | $state{slot_guard} = shift; |
1133 | $state{slot_guard} = shift; |
427 | |
1134 | |
428 | return unless $state{connect_guard}; |
1135 | return unless $state{connect_guard}; |
429 | |
1136 | |
430 | $state{connect_guard} = AnyEvent::Socket::tcp_connect $rhost, $rport, sub { |
1137 | # try to use an existing keepalive connection, but only if we, ourselves, plan |
431 | $state{fh} = shift |
1138 | # on a keepalive request (in theory, this should be a separate config option). |
432 | or do { |
1139 | if ($persistent && $KA_CACHE{$ka_key}) { |
433 | my $err = "$!"; |
1140 | $was_persistent = 1; |
434 | %state = (); |
|
|
435 | return $cb->(undef, { Status => 599, Reason => $err, URL => $url }); |
|
|
436 | }; |
|
|
437 | |
1141 | |
438 | pop; # free memory, save a tree |
1142 | $state{handle} = ka_fetch $ka_key; |
|
|
1143 | $state{handle}->destroyed |
|
|
1144 | and die "AnyEvent::HTTP: unexpectedly got a destructed handle (1), please report.";#d# |
|
|
1145 | $prepare_handle->(); |
|
|
1146 | $state{handle}->destroyed |
|
|
1147 | and die "AnyEvent::HTTP: unexpectedly got a destructed handle (2), please report.";#d# |
|
|
1148 | $handle_actual_request->(); |
439 | |
1149 | |
440 | return unless delete $state{connect_guard}; |
|
|
441 | |
|
|
442 | # get handle |
|
|
443 | $state{handle} = new AnyEvent::Handle |
|
|
444 | fh => $state{fh}, |
|
|
445 | timeout => $timeout, |
|
|
446 | peername => $rhost, |
|
|
447 | tls_ctx => $arg{tls_ctx}; |
|
|
448 | |
|
|
449 | # limit the number of persistent connections |
|
|
450 | # keepalive not yet supported |
|
|
451 | if ($KA_COUNT{$_[1]} < $MAX_PERSISTENT_PER_HOST) { |
|
|
452 | ++$KA_COUNT{$_[1]}; |
|
|
453 | $state{handle}{ka_count_guard} = AnyEvent::Util::guard { |
|
|
454 | --$KA_COUNT{$_[1]} |
|
|
455 | }; |
|
|
456 | $hdr{connection} = "keep-alive"; |
|
|
457 | } else { |
1150 | } else { |
458 | delete $hdr{connection}; |
1151 | my $tcp_connect = $arg{tcp_connect} |
459 | } |
1152 | || do { require AnyEvent::Socket; \&AnyEvent::Socket::tcp_connect }; |
460 | |
1153 | |
461 | # (re-)configure handle |
1154 | $state{connect_guard} = $tcp_connect->($rhost, $rport, $connect_cb, $arg{on_prepare} || sub { $timeout }); |
462 | $state{handle}->on_error (sub { |
|
|
463 | %state = (); |
|
|
464 | $cb->(undef, { Status => 599, Reason => $_[2], URL => $url }); |
|
|
465 | }); |
|
|
466 | $state{handle}->on_eof (sub { |
|
|
467 | %state = (); |
|
|
468 | $cb->(undef, { Status => 599, Reason => "Unexpected end-of-file", URL => $url }); |
|
|
469 | }); |
|
|
470 | |
|
|
471 | $state{handle}->starttls ("connect") if $rscheme eq "https"; |
|
|
472 | |
|
|
473 | # handle actual, non-tunneled, request |
|
|
474 | my $handle_actual_request = sub { |
|
|
475 | $state{handle}->starttls ("connect") if $uscheme eq "https" && !exists $state{handle}{tls}; |
|
|
476 | |
|
|
477 | # send request |
|
|
478 | $state{handle}->push_write ( |
|
|
479 | "$method $rpath HTTP/1.0\015\012" |
|
|
480 | . (join "", map "\u$_: $hdr{$_}\015\012", grep defined $hdr{$_}, keys %hdr) |
|
|
481 | . "\015\012" |
|
|
482 | . (delete $arg{body}) |
|
|
483 | ); |
|
|
484 | |
|
|
485 | %hdr = (); # reduce memory usage, save a kitten |
|
|
486 | |
|
|
487 | # status line |
|
|
488 | $state{handle}->push_read (line => $qr_nl, sub { |
|
|
489 | $_[1] =~ /^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\015\012]*) )?/ix |
|
|
490 | or return (%state = (), $cb->(undef, { Status => 599, Reason => "Invalid server response ($_[1])", URL => $url })); |
|
|
491 | |
|
|
492 | my %hdr = ( # response headers |
|
|
493 | HTTPVersion => ",$1", |
|
|
494 | Status => ",$2", |
|
|
495 | Reason => ",$3", |
|
|
496 | URL => ",$url" |
|
|
497 | ); |
|
|
498 | |
|
|
499 | # headers, could be optimized a bit |
|
|
500 | $state{handle}->unshift_read (line => $qr_nlnl, sub { |
|
|
501 | for ("$_[1]") { |
|
|
502 | y/\015//d; # weed out any \015, as they show up in the weirdest of places. |
|
|
503 | |
|
|
504 | # things seen, not parsed: |
|
|
505 | # p3pP="NON CUR OTPi OUR NOR UNI" |
|
|
506 | |
|
|
507 | $hdr{lc $1} .= ",$2" |
|
|
508 | while /\G |
|
|
509 | ([^:\000-\037]*): |
|
|
510 | [\011\040]* |
|
|
511 | ((?: [^\012]+ | \012[\011\040] )*) |
|
|
512 | \012 |
|
|
513 | /gxc; |
|
|
514 | |
|
|
515 | /\G$/ |
|
|
516 | or return (%state = (), $cb->(undef, { Status => 599, Reason => "Garbled response headers", URL => $url })); |
|
|
517 | } |
|
|
518 | |
|
|
519 | substr $_, 0, 1, "" |
|
|
520 | for values %hdr; |
|
|
521 | |
|
|
522 | # redirect handling |
|
|
523 | # microsoft and other shitheads don't give a shit for following standards, |
|
|
524 | # try to support some common forms of broken Location headers. |
|
|
525 | if ($hdr{location} !~ /^(?: $ | [^:\/?\#]+ : )/x) { |
|
|
526 | $hdr{location} =~ s/^\.\/+//; |
|
|
527 | |
|
|
528 | my $url = "$rscheme://$uhost:$uport"; |
|
|
529 | |
|
|
530 | unless ($hdr{location} =~ s/^\///) { |
|
|
531 | $url .= $upath; |
|
|
532 | $url =~ s/\/[^\/]*$//; |
|
|
533 | } |
|
|
534 | |
|
|
535 | $hdr{location} = "$url/$hdr{location}"; |
|
|
536 | } |
|
|
537 | |
|
|
538 | my $redirect; |
|
|
539 | |
|
|
540 | if ($recurse) { |
|
|
541 | if ($hdr{Status} =~ /^30[12]$/ && $method ne "POST") { |
|
|
542 | # apparently, mozilla et al. just change POST to GET here |
|
|
543 | # more research is needed before we do the same |
|
|
544 | $redirect = 1; |
|
|
545 | } elsif ($hdr{Status} == 303) { |
|
|
546 | # even http/1.1 is unclear on how to mutate the method |
|
|
547 | $method = "GET" unless $method eq "HEAD"; |
|
|
548 | $redirect = 1; |
|
|
549 | } elsif ($hdr{Status} == 307 && $method =~ /^(?:GET|HEAD)$/) { |
|
|
550 | $redirect = 1; |
|
|
551 | } |
|
|
552 | } |
|
|
553 | |
|
|
554 | my $finish = sub { |
|
|
555 | $state{handle}->destroy if $state{handle}; |
|
|
556 | %state = (); |
|
|
557 | |
|
|
558 | # set-cookie processing |
|
|
559 | if ($arg{cookie_jar}) { |
|
|
560 | for ($_[1]{"set-cookie"}) { |
|
|
561 | # parse NAME=VALUE |
|
|
562 | my @kv; |
|
|
563 | |
|
|
564 | while (/\G\s* ([^=;,[:space:]]+) \s*=\s* (?: "((?:[^\\"]+|\\.)*)" | ([^=;,[:space:]]*) )/gcxs) { |
|
|
565 | my $name = $1; |
|
|
566 | my $value = $3; |
|
|
567 | |
|
|
568 | unless ($value) { |
|
|
569 | $value = $2; |
|
|
570 | $value =~ s/\\(.)/$1/gs; |
|
|
571 | } |
|
|
572 | |
|
|
573 | push @kv, $name => $value; |
|
|
574 | |
|
|
575 | last unless /\G\s*;/gc; |
|
|
576 | } |
|
|
577 | |
|
|
578 | last unless @kv; |
|
|
579 | |
|
|
580 | my $name = shift @kv; |
|
|
581 | my %kv = (value => shift @kv, @kv); |
|
|
582 | |
|
|
583 | my $cdom; |
|
|
584 | my $cpath = (delete $kv{path}) || "/"; |
|
|
585 | |
|
|
586 | if (exists $kv{domain}) { |
|
|
587 | $cdom = delete $kv{domain}; |
|
|
588 | |
1155 | } |
589 | $cdom =~ s/^\.?/./; # make sure it starts with a "." |
|
|
590 | |
|
|
591 | next if $cdom =~ /\.$/; |
|
|
592 | |
|
|
593 | # this is not rfc-like and not netscape-like. go figure. |
|
|
594 | my $ndots = $cdom =~ y/.//; |
|
|
595 | next if $ndots < ($cdom =~ /\.[^.][^.]\.[^.][^.]$/ ? 3 : 2); |
|
|
596 | } else { |
|
|
597 | $cdom = $uhost; |
|
|
598 | } |
|
|
599 | |
|
|
600 | # store it |
|
|
601 | $arg{cookie_jar}{version} = 1; |
|
|
602 | $arg{cookie_jar}{$cdom}{$cpath}{$name} = \%kv; |
|
|
603 | |
|
|
604 | redo if /\G\s*,/gc; |
|
|
605 | } |
|
|
606 | } |
|
|
607 | |
|
|
608 | if ($redirect && exists $hdr{location}) { |
|
|
609 | # we ignore any errors, as it is very common to receive |
|
|
610 | # Content-Length != 0 but no actual body |
|
|
611 | # we also access %hdr, as $_[1] might be an erro |
|
|
612 | http_request ($method => $hdr{location}, %arg, recurse => $recurse - 1, $cb); |
|
|
613 | } else { |
|
|
614 | $cb->($_[0], $_[1]); |
|
|
615 | } |
|
|
616 | }; |
|
|
617 | |
|
|
618 | my $len = $hdr{"content-length"}; |
|
|
619 | |
|
|
620 | if (!$redirect && $arg{on_header} && !$arg{on_header}(\%hdr)) { |
|
|
621 | $finish->(undef, { Status => 598, Reason => "Request cancelled by on_header", URL => $url }); |
|
|
622 | } elsif ( |
|
|
623 | $hdr{Status} =~ /^(?:1..|[23]04)$/ |
|
|
624 | or $method eq "HEAD" |
|
|
625 | or (defined $len && !$len) |
|
|
626 | ) { |
|
|
627 | # no body |
|
|
628 | $finish->("", \%hdr); |
|
|
629 | } else { |
|
|
630 | # body handling, four different code paths |
|
|
631 | # for want_body_handle, on_body (2x), normal (2x) |
|
|
632 | # we might read too much here, but it does not matter yet (no pers. connections) |
|
|
633 | if (!$redirect && $arg{want_body_handle}) { |
|
|
634 | $_[0]->on_eof (undef); |
|
|
635 | $_[0]->on_error (undef); |
|
|
636 | $_[0]->on_read (undef); |
|
|
637 | |
|
|
638 | $finish->(delete $state{handle}, \%hdr); |
|
|
639 | |
|
|
640 | } elsif ($arg{on_body}) { |
|
|
641 | $_[0]->on_error (sub { $finish->(undef, { Status => 599, Reason => $_[2], URL => $url }) }); |
|
|
642 | if ($len) { |
|
|
643 | $_[0]->on_eof (undef); |
|
|
644 | $_[0]->on_read (sub { |
|
|
645 | $len -= length $_[0]{rbuf}; |
|
|
646 | |
|
|
647 | $arg{on_body}(delete $_[0]{rbuf}, \%hdr) |
|
|
648 | or $finish->(undef, { Status => 598, Reason => "Request cancelled by on_body", URL => $url }); |
|
|
649 | |
|
|
650 | $len > 0 |
|
|
651 | or $finish->("", \%hdr); |
|
|
652 | }); |
|
|
653 | } else { |
|
|
654 | $_[0]->on_eof (sub { |
|
|
655 | $finish->("", \%hdr); |
|
|
656 | }); |
|
|
657 | $_[0]->on_read (sub { |
|
|
658 | $arg{on_body}(delete $_[0]{rbuf}, \%hdr) |
|
|
659 | or $finish->(undef, { Status => 598, Reason => "Request cancelled by on_body", URL => $url }); |
|
|
660 | }); |
|
|
661 | } |
|
|
662 | } else { |
|
|
663 | $_[0]->on_eof (undef); |
|
|
664 | |
|
|
665 | if ($len) { |
|
|
666 | $_[0]->on_error (sub { $finish->(undef, { Status => 599, Reason => $_[2], URL => $url }) }); |
|
|
667 | $_[0]->on_read (sub { |
|
|
668 | $finish->((substr delete $_[0]{rbuf}, 0, $len, ""), \%hdr) |
|
|
669 | if $len <= length $_[0]{rbuf}; |
|
|
670 | }); |
|
|
671 | } else { |
|
|
672 | $_[0]->on_error (sub { |
|
|
673 | $! == Errno::EPIPE || !$! |
|
|
674 | ? $finish->(delete $_[0]{rbuf}, \%hdr) |
|
|
675 | : $finish->(undef, { Status => 599, Reason => $_[2], URL => $url }); |
|
|
676 | }); |
|
|
677 | $_[0]->on_read (sub { }); |
|
|
678 | } |
|
|
679 | } |
|
|
680 | } |
|
|
681 | }); |
|
|
682 | }); |
|
|
683 | }; |
|
|
684 | |
|
|
685 | # now handle proxy-CONNECT method |
|
|
686 | if ($proxy && $uscheme eq "https") { |
|
|
687 | # oh dear, we have to wrap it into a connect request |
|
|
688 | |
|
|
689 | # maybe re-use $uauthority with patched port? |
|
|
690 | $state{handle}->push_write ("CONNECT $uhost:$uport HTTP/1.0\015\012Host: $uhost\015\012\015\012"); |
|
|
691 | $state{handle}->push_read (line => $qr_nlnl, sub { |
|
|
692 | $_[1] =~ /^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\015\012]*) )?/ix |
|
|
693 | or return (%state = (), $cb->(undef, { Status => 599, Reason => "Invalid proxy connect response ($_[1])", URL => $url })); |
|
|
694 | |
|
|
695 | if ($2 == 200) { |
|
|
696 | $rpath = $upath; |
|
|
697 | &$handle_actual_request; |
|
|
698 | } else { |
|
|
699 | %state = (); |
|
|
700 | $cb->(undef, { Status => $2, Reason => $3, URL => $url }); |
|
|
701 | } |
|
|
702 | }); |
|
|
703 | } else { |
|
|
704 | &$handle_actual_request; |
|
|
705 | } |
|
|
706 | |
|
|
707 | }, $arg{on_prepare} || sub { $timeout }; |
|
|
708 | }; |
1156 | }; |
709 | |
1157 | |
710 | defined wantarray && AnyEvent::Util::guard { %state = () } |
1158 | defined wantarray && AnyEvent::Util::guard { _destroy_state %state } |
711 | } |
1159 | } |
712 | |
1160 | |
713 | sub http_get($@) { |
1161 | sub http_get($@) { |
714 | unshift @_, "GET"; |
1162 | unshift @_, "GET"; |
715 | &http_request |
1163 | &http_request |
… | |
… | |
726 | &http_request |
1174 | &http_request |
727 | } |
1175 | } |
728 | |
1176 | |
729 | =back |
1177 | =back |
730 | |
1178 | |
|
|
1179 | =head2 DNS CACHING |
|
|
1180 | |
|
|
1181 | AnyEvent::HTTP uses the AnyEvent::Socket::tcp_connect function for |
|
|
1182 | the actual connection, which in turn uses AnyEvent::DNS to resolve |
|
|
1183 | hostnames. The latter is a simple stub resolver and does no caching |
|
|
1184 | on its own. If you want DNS caching, you currently have to provide |
|
|
1185 | your own default resolver (by storing a suitable resolver object in |
|
|
1186 | C<$AnyEvent::DNS::RESOLVER>) or your own C<tcp_connect> callback. |
|
|
1187 | |
731 | =head2 GLOBAL FUNCTIONS AND VARIABLES |
1188 | =head2 GLOBAL FUNCTIONS AND VARIABLES |
732 | |
1189 | |
733 | =over 4 |
1190 | =over 4 |
734 | |
1191 | |
735 | =item AnyEvent::HTTP::set_proxy "proxy-url" |
1192 | =item AnyEvent::HTTP::set_proxy "proxy-url" |
736 | |
1193 | |
737 | Sets the default proxy server to use. The proxy-url must begin with a |
1194 | Sets the default proxy server to use. The proxy-url must begin with a |
738 | string of the form C<http://host:port> (optionally C<https:...>). |
1195 | string of the form C<http://host:port>, croaks otherwise. |
|
|
1196 | |
|
|
1197 | To clear an already-set proxy, use C<undef>. |
|
|
1198 | |
|
|
1199 | When AnyEvent::HTTP is loaded for the first time it will query the |
|
|
1200 | default proxy from the operating system, currently by looking at |
|
|
1201 | C<$ENV{http_proxy>}. |
|
|
1202 | |
|
|
1203 | =item AnyEvent::HTTP::cookie_jar_expire $jar[, $session_end] |
|
|
1204 | |
|
|
1205 | Remove all cookies from the cookie jar that have been expired. If |
|
|
1206 | C<$session_end> is given and true, then additionally remove all session |
|
|
1207 | cookies. |
|
|
1208 | |
|
|
1209 | You should call this function (with a true C<$session_end>) before you |
|
|
1210 | save cookies to disk, and you should call this function after loading them |
|
|
1211 | again. If you have a long-running program you can additonally call this |
|
|
1212 | function from time to time. |
|
|
1213 | |
|
|
1214 | A cookie jar is initially an empty hash-reference that is managed by this |
|
|
1215 | module. It's format is subject to change, but currently it is like this: |
|
|
1216 | |
|
|
1217 | The key C<version> has to contain C<1>, otherwise the hash gets |
|
|
1218 | emptied. All other keys are hostnames or IP addresses pointing to |
|
|
1219 | hash-references. The key for these inner hash references is the |
|
|
1220 | server path for which this cookie is meant, and the values are again |
|
|
1221 | hash-references. The keys of those hash-references is the cookie name, and |
|
|
1222 | the value, you guessed it, is another hash-reference, this time with the |
|
|
1223 | key-value pairs from the cookie, except for C<expires> and C<max-age>, |
|
|
1224 | which have been replaced by a C<_expires> key that contains the cookie |
|
|
1225 | expiry timestamp. |
|
|
1226 | |
|
|
1227 | Here is an example of a cookie jar with a single cookie, so you have a |
|
|
1228 | chance of understanding the above paragraph: |
|
|
1229 | |
|
|
1230 | { |
|
|
1231 | version => 1, |
|
|
1232 | "10.0.0.1" => { |
|
|
1233 | "/" => { |
|
|
1234 | "mythweb_id" => { |
|
|
1235 | _expires => 1293917923, |
|
|
1236 | value => "ooRung9dThee3ooyXooM1Ohm", |
|
|
1237 | }, |
|
|
1238 | }, |
|
|
1239 | }, |
|
|
1240 | } |
|
|
1241 | |
|
|
1242 | =item $date = AnyEvent::HTTP::format_date $timestamp |
|
|
1243 | |
|
|
1244 | Takes a POSIX timestamp (seconds since the epoch) and formats it as a HTTP |
|
|
1245 | Date (RFC 2616). |
|
|
1246 | |
|
|
1247 | =item $timestamp = AnyEvent::HTTP::parse_date $date |
|
|
1248 | |
|
|
1249 | Takes a HTTP Date (RFC 2616) or a Cookie date (netscape cookie spec) or a |
|
|
1250 | bunch of minor variations of those, and returns the corresponding POSIX |
|
|
1251 | timestamp, or C<undef> if the date cannot be parsed. |
739 | |
1252 | |
740 | =item $AnyEvent::HTTP::MAX_RECURSE |
1253 | =item $AnyEvent::HTTP::MAX_RECURSE |
741 | |
1254 | |
742 | The default value for the C<recurse> request parameter (default: C<10>). |
1255 | The default value for the C<recurse> request parameter (default: C<10>). |
|
|
1256 | |
|
|
1257 | =item $AnyEvent::HTTP::TIMEOUT |
|
|
1258 | |
|
|
1259 | The default timeout for connection operations (default: C<300>). |
743 | |
1260 | |
744 | =item $AnyEvent::HTTP::USERAGENT |
1261 | =item $AnyEvent::HTTP::USERAGENT |
745 | |
1262 | |
746 | The default value for the C<User-Agent> header (the default is |
1263 | The default value for the C<User-Agent> header (the default is |
747 | C<Mozilla/5.0 (compatible; U; AnyEvent-HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)>). |
1264 | C<Mozilla/5.0 (compatible; U; AnyEvent-HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)>). |
748 | |
1265 | |
749 | =item $AnyEvent::HTTP::MAX_PER_HOST |
1266 | =item $AnyEvent::HTTP::MAX_PER_HOST |
750 | |
1267 | |
751 | The maximum number of concurrent connections to the same host (identified |
1268 | The maximum number of concurrent connections to the same host (identified |
752 | by the hostname). If the limit is exceeded, then the additional requests |
1269 | by the hostname). If the limit is exceeded, then the additional requests |
753 | are queued until previous connections are closed. |
1270 | are queued until previous connections are closed. Both persistent and |
|
|
1271 | non-persistent connections are counted in this limit. |
754 | |
1272 | |
755 | The default value for this is C<4>, and it is highly advisable to not |
1273 | The default value for this is C<4>, and it is highly advisable to not |
756 | increase it. |
1274 | increase it much. |
|
|
1275 | |
|
|
1276 | For comparison: the RFC's recommend 4 non-persistent or 2 persistent |
|
|
1277 | connections, older browsers used 2, newers (such as firefox 3) typically |
|
|
1278 | use 6, and Opera uses 8 because like, they have the fastest browser and |
|
|
1279 | give a shit for everybody else on the planet. |
|
|
1280 | |
|
|
1281 | =item $AnyEvent::HTTP::PERSISTENT_TIMEOUT |
|
|
1282 | |
|
|
1283 | The time after which idle persistent conenctions get closed by |
|
|
1284 | AnyEvent::HTTP (default: C<3>). |
757 | |
1285 | |
758 | =item $AnyEvent::HTTP::ACTIVE |
1286 | =item $AnyEvent::HTTP::ACTIVE |
759 | |
1287 | |
760 | The number of active connections. This is not the number of currently |
1288 | The number of active connections. This is not the number of currently |
761 | running requests, but the number of currently open and non-idle TCP |
1289 | running requests, but the number of currently open and non-idle TCP |
762 | connections. This number of can be useful for load-leveling. |
1290 | connections. This number can be useful for load-leveling. |
763 | |
1291 | |
764 | =back |
1292 | =back |
765 | |
1293 | |
766 | =cut |
1294 | =cut |
767 | |
1295 | |
|
|
1296 | our @month = qw(Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec); |
|
|
1297 | our @weekday = qw(Sun Mon Tue Wed Thu Fri Sat); |
|
|
1298 | |
|
|
1299 | sub format_date($) { |
|
|
1300 | my ($time) = @_; |
|
|
1301 | |
|
|
1302 | # RFC 822/1123 format |
|
|
1303 | my ($S, $M, $H, $mday, $mon, $year, $wday, $yday, undef) = gmtime $time; |
|
|
1304 | |
|
|
1305 | sprintf "%s, %02d %s %04d %02d:%02d:%02d GMT", |
|
|
1306 | $weekday[$wday], $mday, $month[$mon], $year + 1900, |
|
|
1307 | $H, $M, $S; |
|
|
1308 | } |
|
|
1309 | |
|
|
1310 | sub parse_date($) { |
|
|
1311 | my ($date) = @_; |
|
|
1312 | |
|
|
1313 | my ($d, $m, $y, $H, $M, $S); |
|
|
1314 | |
|
|
1315 | if ($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$/) { |
|
|
1316 | # RFC 822/1123, required by RFC 2616 (with " ") |
|
|
1317 | # cookie dates (with "-") |
|
|
1318 | |
|
|
1319 | ($d, $m, $y, $H, $M, $S) = ($1, $2, $3, $4, $5, $6); |
|
|
1320 | |
|
|
1321 | } elsif ($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]?) GMT$/) { |
|
|
1322 | # RFC 850 |
|
|
1323 | ($d, $m, $y, $H, $M, $S) = ($1, $2, $3 < 69 ? $3 + 2000 : $3 + 1900, $4, $5, $6); |
|
|
1324 | |
|
|
1325 | } elsif ($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])$/) { |
|
|
1326 | # ISO C's asctime |
|
|
1327 | ($d, $m, $y, $H, $M, $S) = ($2, $1, $6, $3, $4, $5); |
|
|
1328 | } |
|
|
1329 | # other formats fail in the loop below |
|
|
1330 | |
|
|
1331 | for (0..11) { |
|
|
1332 | if ($m eq $month[$_]) { |
|
|
1333 | require Time::Local; |
|
|
1334 | return eval { Time::Local::timegm ($S, $M, $H, $d, $_, $y) }; |
|
|
1335 | } |
|
|
1336 | } |
|
|
1337 | |
|
|
1338 | undef |
|
|
1339 | } |
|
|
1340 | |
768 | sub set_proxy($) { |
1341 | sub set_proxy($) { |
769 | $PROXY = [$2, $3 || 3128, $1] if $_[0] =~ m%^(https?):// ([^:/]+) (?: : (\d*) )?%ix; |
1342 | if (length $_[0]) { |
|
|
1343 | $_[0] =~ m%^(http):// ([^:/]+) (?: : (\d*) )?%ix |
|
|
1344 | or Carp::croak "$_[0]: invalid proxy URL"; |
|
|
1345 | $PROXY = [$2, $3 || 3128, $1] |
|
|
1346 | } else { |
|
|
1347 | undef $PROXY; |
|
|
1348 | } |
770 | } |
1349 | } |
771 | |
1350 | |
772 | # initialise proxy from environment |
1351 | # initialise proxy from environment |
|
|
1352 | eval { |
773 | set_proxy $ENV{http_proxy}; |
1353 | set_proxy $ENV{http_proxy}; |
|
|
1354 | }; |
|
|
1355 | |
|
|
1356 | =head2 SHOWCASE |
|
|
1357 | |
|
|
1358 | This section contaisn some more elaborate "real-world" examples or code |
|
|
1359 | snippets. |
|
|
1360 | |
|
|
1361 | =head2 HTTP/1.1 FILE DOWNLOAD |
|
|
1362 | |
|
|
1363 | Downloading files with HTTP can be quite tricky, especially when something |
|
|
1364 | goes wrong and you want to resume. |
|
|
1365 | |
|
|
1366 | Here is a function that initiates and resumes a download. It uses the |
|
|
1367 | last modified time to check for file content changes, and works with many |
|
|
1368 | HTTP/1.0 servers as well, and usually falls back to a complete re-download |
|
|
1369 | on older servers. |
|
|
1370 | |
|
|
1371 | It calls the completion callback with either C<undef>, which means a |
|
|
1372 | nonretryable error occured, C<0> when the download was partial and should |
|
|
1373 | be retried, and C<1> if it was successful. |
|
|
1374 | |
|
|
1375 | use AnyEvent::HTTP; |
|
|
1376 | |
|
|
1377 | sub download($$$) { |
|
|
1378 | my ($url, $file, $cb) = @_; |
|
|
1379 | |
|
|
1380 | open my $fh, "+<", $file |
|
|
1381 | or die "$file: $!"; |
|
|
1382 | |
|
|
1383 | my %hdr; |
|
|
1384 | my $ofs = 0; |
|
|
1385 | |
|
|
1386 | warn stat $fh; |
|
|
1387 | warn -s _; |
|
|
1388 | if (stat $fh and -s _) { |
|
|
1389 | $ofs = -s _; |
|
|
1390 | warn "-s is ", $ofs; |
|
|
1391 | $hdr{"if-unmodified-since"} = AnyEvent::HTTP::format_date +(stat _)[9]; |
|
|
1392 | $hdr{"range"} = "bytes=$ofs-"; |
|
|
1393 | } |
|
|
1394 | |
|
|
1395 | http_get $url, |
|
|
1396 | headers => \%hdr, |
|
|
1397 | on_header => sub { |
|
|
1398 | my ($hdr) = @_; |
|
|
1399 | |
|
|
1400 | if ($hdr->{Status} == 200 && $ofs) { |
|
|
1401 | # resume failed |
|
|
1402 | truncate $fh, $ofs = 0; |
|
|
1403 | } |
|
|
1404 | |
|
|
1405 | sysseek $fh, $ofs, 0; |
|
|
1406 | |
|
|
1407 | 1 |
|
|
1408 | }, |
|
|
1409 | on_body => sub { |
|
|
1410 | my ($data, $hdr) = @_; |
|
|
1411 | |
|
|
1412 | if ($hdr->{Status} =~ /^2/) { |
|
|
1413 | length $data == syswrite $fh, $data |
|
|
1414 | or return; # abort on write errors |
|
|
1415 | } |
|
|
1416 | |
|
|
1417 | 1 |
|
|
1418 | }, |
|
|
1419 | sub { |
|
|
1420 | my (undef, $hdr) = @_; |
|
|
1421 | |
|
|
1422 | my $status = $hdr->{Status}; |
|
|
1423 | |
|
|
1424 | if (my $time = AnyEvent::HTTP::parse_date $hdr->{"last-modified"}) { |
|
|
1425 | utime $fh, $time, $time; |
|
|
1426 | } |
|
|
1427 | |
|
|
1428 | if ($status == 200 || $status == 206 || $status == 416) { |
|
|
1429 | # download ok || resume ok || file already fully downloaded |
|
|
1430 | $cb->(1, $hdr); |
|
|
1431 | |
|
|
1432 | } elsif ($status == 412) { |
|
|
1433 | # file has changed while resuming, delete and retry |
|
|
1434 | unlink $file; |
|
|
1435 | $cb->(0, $hdr); |
|
|
1436 | |
|
|
1437 | } elsif ($status == 500 or $status == 503 or $status =~ /^59/) { |
|
|
1438 | # retry later |
|
|
1439 | $cb->(0, $hdr); |
|
|
1440 | |
|
|
1441 | } else { |
|
|
1442 | $cb->(undef, $hdr); |
|
|
1443 | } |
|
|
1444 | } |
|
|
1445 | ; |
|
|
1446 | } |
|
|
1447 | |
|
|
1448 | download "http://server/somelargefile", "/tmp/somelargefile", sub { |
|
|
1449 | if ($_[0]) { |
|
|
1450 | print "OK!\n"; |
|
|
1451 | } elsif (defined $_[0]) { |
|
|
1452 | print "please retry later\n"; |
|
|
1453 | } else { |
|
|
1454 | print "ERROR\n"; |
|
|
1455 | } |
|
|
1456 | }; |
|
|
1457 | |
|
|
1458 | =head3 SOCKS PROXIES |
|
|
1459 | |
|
|
1460 | Socks proxies are not directly supported by AnyEvent::HTTP. You can |
|
|
1461 | compile your perl to support socks, or use an external program such as |
|
|
1462 | F<socksify> (dante) or F<tsocks> to make your program use a socks proxy |
|
|
1463 | transparently. |
|
|
1464 | |
|
|
1465 | Alternatively, for AnyEvent::HTTP only, you can use your own |
|
|
1466 | C<tcp_connect> function that does the proxy handshake - here is an example |
|
|
1467 | that works with socks4a proxies: |
|
|
1468 | |
|
|
1469 | use Errno; |
|
|
1470 | use AnyEvent::Util; |
|
|
1471 | use AnyEvent::Socket; |
|
|
1472 | use AnyEvent::Handle; |
|
|
1473 | |
|
|
1474 | # host, port and username of/for your socks4a proxy |
|
|
1475 | my $socks_host = "10.0.0.23"; |
|
|
1476 | my $socks_port = 9050; |
|
|
1477 | my $socks_user = ""; |
|
|
1478 | |
|
|
1479 | sub socks4a_connect { |
|
|
1480 | my ($host, $port, $connect_cb, $prepare_cb) = @_; |
|
|
1481 | |
|
|
1482 | my $hdl = new AnyEvent::Handle |
|
|
1483 | connect => [$socks_host, $socks_port], |
|
|
1484 | on_prepare => sub { $prepare_cb->($_[0]{fh}) }, |
|
|
1485 | on_error => sub { $connect_cb->() }, |
|
|
1486 | ; |
|
|
1487 | |
|
|
1488 | $hdl->push_write (pack "CCnNZ*Z*", 4, 1, $port, 1, $socks_user, $host); |
|
|
1489 | |
|
|
1490 | $hdl->push_read (chunk => 8, sub { |
|
|
1491 | my ($hdl, $chunk) = @_; |
|
|
1492 | my ($status, $port, $ipn) = unpack "xCna4", $chunk; |
|
|
1493 | |
|
|
1494 | if ($status == 0x5a) { |
|
|
1495 | $connect_cb->($hdl->{fh}, (format_address $ipn) . ":$port"); |
|
|
1496 | } else { |
|
|
1497 | $! = Errno::ENXIO; $connect_cb->(); |
|
|
1498 | } |
|
|
1499 | }); |
|
|
1500 | |
|
|
1501 | $hdl |
|
|
1502 | } |
|
|
1503 | |
|
|
1504 | Use C<socks4a_connect> instead of C<tcp_connect> when doing C<http_request>s, |
|
|
1505 | possibly after switching off other proxy types: |
|
|
1506 | |
|
|
1507 | AnyEvent::HTTP::set_proxy undef; # usually you do not want other proxies |
|
|
1508 | |
|
|
1509 | http_get 'http://www.google.com', tcp_connect => \&socks4a_connect, sub { |
|
|
1510 | my ($data, $headers) = @_; |
|
|
1511 | ... |
|
|
1512 | }; |
774 | |
1513 | |
775 | =head1 SEE ALSO |
1514 | =head1 SEE ALSO |
776 | |
1515 | |
777 | L<AnyEvent>. |
1516 | L<AnyEvent>. |
778 | |
1517 | |