… | |
… | |
3 | AnyEvent::HTTP - simple but non-blocking HTTP/HTTPS client |
3 | AnyEvent::HTTP - simple but non-blocking HTTP/HTTPS client |
4 | |
4 | |
5 | =head1 SYNOPSIS |
5 | =head1 SYNOPSIS |
6 | |
6 | |
7 | use AnyEvent::HTTP; |
7 | use AnyEvent::HTTP; |
|
|
8 | |
|
|
9 | http_get "http://www.nethype.de/", sub { print $_[1] }; |
|
|
10 | |
|
|
11 | # ... do something else here |
8 | |
12 | |
9 | =head1 DESCRIPTION |
13 | =head1 DESCRIPTION |
10 | |
14 | |
11 | 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 |
12 | run a supported event loop. |
16 | run a supported event loop. |
13 | |
17 | |
|
|
18 | This module implements a simple, stateless and non-blocking HTTP |
|
|
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 |
|
|
21 | automatically limits the number of connections to the values specified in |
|
|
22 | the RFC. |
|
|
23 | |
|
|
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 |
|
|
26 | possible as the user retains control over request and response headers. |
|
|
27 | |
|
|
28 | The caller is responsible for authentication management, cookies (if |
|
|
29 | the simplistic implementation in this module doesn't suffice), referer |
|
|
30 | and other high-level protocol details for which this module offers only |
|
|
31 | limited support. |
|
|
32 | |
14 | =head2 METHODS |
33 | =head2 METHODS |
15 | |
34 | |
16 | =over 4 |
35 | =over 4 |
17 | |
36 | |
18 | =cut |
37 | =cut |
… | |
… | |
20 | package AnyEvent::HTTP; |
39 | package AnyEvent::HTTP; |
21 | |
40 | |
22 | use strict; |
41 | use strict; |
23 | no warnings; |
42 | no warnings; |
24 | |
43 | |
25 | use Carp; |
44 | use Errno (); |
26 | |
45 | |
27 | use AnyEvent (); |
46 | use AnyEvent 5.0 (); |
28 | use AnyEvent::Util (); |
47 | use AnyEvent::Util (); |
29 | use AnyEvent::Socket (); |
|
|
30 | use AnyEvent::Handle (); |
48 | use AnyEvent::Handle (); |
31 | |
49 | |
32 | use base Exporter::; |
50 | use base Exporter::; |
33 | |
51 | |
34 | our $VERSION = '1.0'; |
52 | our $VERSION = '1.5'; |
35 | |
53 | |
36 | our @EXPORT = qw(http_get http_request); |
54 | our @EXPORT = qw(http_get http_post http_head http_request); |
37 | |
55 | |
38 | 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)"; |
39 | our $MAX_RECURSE = 10; |
57 | our $MAX_RECURSE = 10; |
40 | our $MAX_PERSISTENT = 8; |
58 | our $MAX_PERSISTENT = 8; |
41 | our $PERSISTENT_TIMEOUT = 2; |
59 | our $PERSISTENT_TIMEOUT = 2; |
42 | our $TIMEOUT = 300; |
60 | our $TIMEOUT = 300; |
43 | |
61 | |
44 | # changing these is evil |
62 | # changing these is evil |
45 | our $MAX_PERSISTENT_PER_HOST = 2; |
63 | our $MAX_PERSISTENT_PER_HOST = 0; |
46 | our $MAX_PER_HOST = 4; # not respected yet :( |
64 | our $MAX_PER_HOST = 4; |
47 | |
65 | |
48 | our $PROXY; |
66 | our $PROXY; |
|
|
67 | our $ACTIVE = 0; |
49 | |
68 | |
50 | my %KA_COUNT; # number of open keep-alive connections per host |
69 | my %KA_COUNT; # number of open keep-alive connections per host |
|
|
70 | my %CO_SLOT; # number of open connections, and wait queue, per host |
51 | |
71 | |
52 | =item http_get $url, key => value..., $cb->($data, $headers) |
72 | =item http_get $url, key => value..., $cb->($data, $headers) |
53 | |
73 | |
54 | Executes an HTTP-GET request. See the http_request function for details on |
74 | Executes an HTTP-GET request. See the http_request function for details on |
55 | additional parameters. |
75 | additional parameters and the return value. |
56 | |
76 | |
57 | =item http_head $url, key => value..., $cb->($data, $headers) |
77 | =item http_head $url, key => value..., $cb->($data, $headers) |
58 | |
78 | |
59 | Executes an HTTP-HEAD request. See the http_request function for details on |
79 | Executes an HTTP-HEAD request. See the http_request function for details |
60 | additional parameters. |
80 | on additional parameters and the return value. |
61 | |
81 | |
62 | =item http_post $url, $body, key => value..., $cb->($data, $headers) |
82 | =item http_post $url, $body, key => value..., $cb->($data, $headers) |
63 | |
83 | |
64 | Executes an HTTP-POST request with a requets body of C<$bod>. See the |
84 | Executes an HTTP-POST request with a request body of C<$body>. See the |
65 | http_request function for details on additional parameters. |
85 | http_request function for details on additional parameters and the return |
|
|
86 | value. |
66 | |
87 | |
67 | =item http_request $method => $url, key => value..., $cb->($data, $headers) |
88 | =item http_request $method => $url, key => value..., $cb->($data, $headers) |
68 | |
89 | |
69 | Executes a HTTP request of type C<$method> (e.g. C<GET>, C<POST>). The URL |
90 | Executes a HTTP request of type C<$method> (e.g. C<GET>, C<POST>). The URL |
70 | must be an absolute http or https URL. |
91 | must be an absolute http or https URL. |
71 | |
92 | |
|
|
93 | When called in void context, nothing is returned. In other contexts, |
|
|
94 | C<http_request> returns a "cancellation guard" - you have to keep the |
|
|
95 | object at least alive until the callback get called. If the object gets |
|
|
96 | destroyed before the callback is called, the request will be cancelled. |
|
|
97 | |
72 | 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 |
73 | (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 |
74 | response headers as second argument. |
100 | (and trailers) as second argument. |
75 | |
101 | |
76 | All the headers in that has are lowercased. In addition to the response |
102 | All the headers in that hash are lowercased. In addition to the response |
77 | headers, the three "pseudo-headers" C<HTTPVersion>, C<Status> and |
103 | headers, the "pseudo-headers" (uppercase to avoid clashing with possible |
78 | C<Reason> contain the three parts of the HTTP Status-Line of the same |
104 | response headers) C<HTTPVersion>, C<Status> and C<Reason> contain the |
79 | name. |
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 | |
|
|
110 | The pseudo-header C<URL> contains the actual URL (which can differ from |
|
|
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 | $headers->{Redirect}[1]{Redirect} >> will then contain the original |
|
|
121 | response, and so on. |
|
|
122 | |
|
|
123 | If the server sends a header multiple times, then their contents will be |
|
|
124 | joined together with a comma (C<,>), as per the HTTP spec. |
80 | |
125 | |
81 | If an internal error occurs, such as not being able to resolve a hostname, |
126 | If an internal error occurs, such as not being able to resolve a hostname, |
82 | then C<$data> will be C<undef>, C<< $headers->{Status} >> will be C<599> |
127 | then C<$data> will be C<undef>, C<< $headers->{Status} >> will be |
83 | and the C<Reason> pseudo-header will contain an error message. |
128 | C<590>-C<599> and the C<Reason> pseudo-header will contain an error |
|
|
129 | message. Currently the following status codes are used: |
|
|
130 | |
|
|
131 | =over 4 |
|
|
132 | |
|
|
133 | =item 595 - errors during connection etsbalishment, proxy handshake. |
|
|
134 | |
|
|
135 | =item 596 - errors during TLS negotiation, request sending and header processing. |
|
|
136 | |
|
|
137 | =item 597 - errors during body receiving or processing. |
|
|
138 | |
|
|
139 | =item 598 - user aborted request via C<on_header> or C<on_body>. |
|
|
140 | |
|
|
141 | =item 599 - other, usually nonretryable, errors (garbled URL etc.). |
|
|
142 | |
|
|
143 | =back |
|
|
144 | |
|
|
145 | A typical callback might look like this: |
|
|
146 | |
|
|
147 | sub { |
|
|
148 | my ($body, $hdr) = @_; |
|
|
149 | |
|
|
150 | if ($hdr->{Status} =~ /^2/) { |
|
|
151 | ... everything should be ok |
|
|
152 | } else { |
|
|
153 | print "error, $hdr->{Status} $hdr->{Reason}\n"; |
|
|
154 | } |
|
|
155 | } |
84 | |
156 | |
85 | Additional parameters are key-value pairs, and are fully optional. They |
157 | Additional parameters are key-value pairs, and are fully optional. They |
86 | include: |
158 | include: |
87 | |
159 | |
88 | =over 4 |
160 | =over 4 |
… | |
… | |
92 | Whether to recurse requests or not, e.g. on redirects, authentication |
164 | Whether to recurse requests or not, e.g. on redirects, authentication |
93 | retries and so on, and how often to do so. |
165 | retries and so on, and how often to do so. |
94 | |
166 | |
95 | =item headers => hashref |
167 | =item headers => hashref |
96 | |
168 | |
97 | The request headers to use. |
169 | The request headers to use. Currently, C<http_request> may provide its own |
|
|
170 | C<Host:>, C<Content-Length:>, C<Connection:> and C<Cookie:> headers and |
|
|
171 | will provide defaults at least for C<TE:>, C<Referer:> and C<User-Agent:> |
|
|
172 | (this can be suppressed by using C<undef> for these headers in which case |
|
|
173 | they won't be sent at all). |
98 | |
174 | |
99 | =item timeout => $seconds |
175 | =item timeout => $seconds |
100 | |
176 | |
101 | The time-out to use for various stages - each connect attempt will reset |
177 | The time-out to use for various stages - each connect attempt will reset |
102 | the timeout, as will read or write activity. Default timeout is 5 minutes. |
178 | the timeout, as will read or write activity, i.e. this is not an overall |
|
|
179 | timeout. |
|
|
180 | |
|
|
181 | Default timeout is 5 minutes. |
103 | |
182 | |
104 | =item proxy => [$host, $port[, $scheme]] or undef |
183 | =item proxy => [$host, $port[, $scheme]] or undef |
105 | |
184 | |
106 | Use the given http proxy for all requests. If not specified, then the |
185 | Use the given http proxy for all requests. If not specified, then the |
107 | default proxy (as specified by C<$ENV{http_proxy}>) is used. |
186 | default proxy (as specified by C<$ENV{http_proxy}>) is used. |
108 | |
187 | |
109 | C<$scheme> must be either missing or C<http> for HTTP, or C<https> for |
188 | C<$scheme> must be either missing, C<http> for HTTP or C<https> for |
110 | HTTPS. |
189 | HTTPS. |
111 | |
190 | |
112 | =item body => $string |
191 | =item body => $string |
113 | |
192 | |
114 | The request body, usually empty. Will be-sent as-is (future versions of |
193 | The request body, usually empty. Will be sent as-is (future versions of |
115 | this module might offer more options). |
194 | this module might offer more options). |
116 | |
195 | |
|
|
196 | =item cookie_jar => $hash_ref |
|
|
197 | |
|
|
198 | Passing this parameter enables (simplified) cookie-processing, loosely |
|
|
199 | based on the original netscape specification. |
|
|
200 | |
|
|
201 | The C<$hash_ref> must be an (initially empty) hash reference which will |
|
|
202 | get updated automatically. It is possible to save the cookie jar to |
|
|
203 | persistent storage with something like JSON or Storable, but this is not |
|
|
204 | recommended, as session-only cookies might survive longer than expected. |
|
|
205 | |
|
|
206 | Note that this cookie implementation is not meant to be complete. If |
|
|
207 | you want complete cookie management you have to do that on your |
|
|
208 | own. C<cookie_jar> is meant as a quick fix to get some cookie-using sites |
|
|
209 | working. Cookies are a privacy disaster, do not use them unless required |
|
|
210 | to. |
|
|
211 | |
|
|
212 | When cookie processing is enabled, the C<Cookie:> and C<Set-Cookie:> |
|
|
213 | headers will be set and handled by this module, otherwise they will be |
|
|
214 | left untouched. |
|
|
215 | |
|
|
216 | =item tls_ctx => $scheme | $tls_ctx |
|
|
217 | |
|
|
218 | Specifies the AnyEvent::TLS context to be used for https connections. This |
|
|
219 | parameter follows the same rules as the C<tls_ctx> parameter to |
|
|
220 | L<AnyEvent::Handle>, but additionally, the two strings C<low> or |
|
|
221 | C<high> can be specified, which give you a predefined low-security (no |
|
|
222 | verification, highest compatibility) and high-security (CA and common-name |
|
|
223 | verification) TLS context. |
|
|
224 | |
|
|
225 | The default for this option is C<low>, which could be interpreted as "give |
|
|
226 | me the page, no matter what". |
|
|
227 | |
|
|
228 | =item on_prepare => $callback->($fh) |
|
|
229 | |
|
|
230 | In rare cases you need to "tune" the socket before it is used to |
|
|
231 | connect (for exmaple, to bind it on a given IP address). This parameter |
|
|
232 | overrides the prepare callback passed to C<AnyEvent::Socket::tcp_connect> |
|
|
233 | and behaves exactly the same way (e.g. it has to provide a |
|
|
234 | timeout). See the description for the C<$prepare_cb> argument of |
|
|
235 | C<AnyEvent::Socket::tcp_connect> for details. |
|
|
236 | |
|
|
237 | =item tcp_connect => $callback->($host, $service, $connect_cb, $prepare_cb) |
|
|
238 | |
|
|
239 | In even rarer cases you want total control over how AnyEvent::HTTP |
|
|
240 | establishes connections. Normally it uses L<AnyEvent::Socket::tcp_connect> |
|
|
241 | to do this, but you can provide your own C<tcp_connect> function - |
|
|
242 | obviously, it has to follow the same calling conventions, except that it |
|
|
243 | may always return a connection guard object. |
|
|
244 | |
|
|
245 | There are probably lots of weird uses for this function, starting from |
|
|
246 | tracing the hosts C<http_request> actually tries to connect, to (inexact |
|
|
247 | but fast) host => IP address caching or even socks protocol support. |
|
|
248 | |
|
|
249 | =item on_header => $callback->($headers) |
|
|
250 | |
|
|
251 | When specified, this callback will be called with the header hash as soon |
|
|
252 | as headers have been successfully received from the remote server (not on |
|
|
253 | locally-generated errors). |
|
|
254 | |
|
|
255 | It has to return either true (in which case AnyEvent::HTTP will continue), |
|
|
256 | or false, in which case AnyEvent::HTTP will cancel the download (and call |
|
|
257 | the finish callback with an error code of C<598>). |
|
|
258 | |
|
|
259 | This callback is useful, among other things, to quickly reject unwanted |
|
|
260 | content, which, if it is supposed to be rare, can be faster than first |
|
|
261 | doing a C<HEAD> request. |
|
|
262 | |
|
|
263 | The downside is that cancelling the request makes it impossible to re-use |
|
|
264 | the connection. Also, the C<on_header> callback will not receive any |
|
|
265 | trailer (headers sent after the response body). |
|
|
266 | |
|
|
267 | Example: cancel the request unless the content-type is "text/html". |
|
|
268 | |
|
|
269 | on_header => sub { |
|
|
270 | $_[0]{"content-type"} =~ /^text\/html\s*(?:;|$)/ |
|
|
271 | }, |
|
|
272 | |
|
|
273 | =item on_body => $callback->($partial_body, $headers) |
|
|
274 | |
|
|
275 | When specified, all body data will be passed to this callback instead of |
|
|
276 | to the completion callback. The completion callback will get the empty |
|
|
277 | string instead of the body data. |
|
|
278 | |
|
|
279 | It has to return either true (in which case AnyEvent::HTTP will continue), |
|
|
280 | or false, in which case AnyEvent::HTTP will cancel the download (and call |
|
|
281 | the completion callback with an error code of C<598>). |
|
|
282 | |
|
|
283 | The downside to cancelling the request is that it makes it impossible to |
|
|
284 | re-use the connection. |
|
|
285 | |
|
|
286 | This callback is useful when the data is too large to be held in memory |
|
|
287 | (so the callback writes it to a file) or when only some information should |
|
|
288 | be extracted, or when the body should be processed incrementally. |
|
|
289 | |
|
|
290 | It is usually preferred over doing your own body handling via |
|
|
291 | C<want_body_handle>, but in case of streaming APIs, where HTTP is |
|
|
292 | only used to create a connection, C<want_body_handle> is the better |
|
|
293 | alternative, as it allows you to install your own event handler, reducing |
|
|
294 | resource usage. |
|
|
295 | |
|
|
296 | =item want_body_handle => $enable |
|
|
297 | |
|
|
298 | When enabled (default is disabled), the behaviour of AnyEvent::HTTP |
|
|
299 | changes considerably: after parsing the headers, and instead of |
|
|
300 | downloading the body (if any), the completion callback will be |
|
|
301 | called. Instead of the C<$body> argument containing the body data, the |
|
|
302 | callback will receive the L<AnyEvent::Handle> object associated with the |
|
|
303 | connection. In error cases, C<undef> will be passed. When there is no body |
|
|
304 | (e.g. status C<304>), the empty string will be passed. |
|
|
305 | |
|
|
306 | The handle object might or might not be in TLS mode, might be connected to |
|
|
307 | a proxy, be a persistent connection etc., and configured in unspecified |
|
|
308 | ways. The user is responsible for this handle (it will not be used by this |
|
|
309 | module anymore). |
|
|
310 | |
|
|
311 | This is useful with some push-type services, where, after the initial |
|
|
312 | headers, an interactive protocol is used (typical example would be the |
|
|
313 | push-style twitter API which starts a JSON/XML stream). |
|
|
314 | |
|
|
315 | If you think you need this, first have a look at C<on_body>, to see if |
|
|
316 | that doesn't solve your problem in a better way. |
|
|
317 | |
117 | =back |
318 | =back |
118 | |
319 | |
119 | =back |
320 | Example: do a simple HTTP GET request for http://www.nethype.de/ and print |
|
|
321 | the response body. |
|
|
322 | |
|
|
323 | http_request GET => "http://www.nethype.de/", sub { |
|
|
324 | my ($body, $hdr) = @_; |
|
|
325 | print "$body\n"; |
|
|
326 | }; |
|
|
327 | |
|
|
328 | Example: do a HTTP HEAD request on https://www.google.com/, use a |
|
|
329 | timeout of 30 seconds. |
|
|
330 | |
|
|
331 | http_request |
|
|
332 | GET => "https://www.google.com", |
|
|
333 | timeout => 30, |
|
|
334 | sub { |
|
|
335 | my ($body, $hdr) = @_; |
|
|
336 | use Data::Dumper; |
|
|
337 | print Dumper $hdr; |
|
|
338 | } |
|
|
339 | ; |
|
|
340 | |
|
|
341 | Example: do another simple HTTP GET request, but immediately try to |
|
|
342 | cancel it. |
|
|
343 | |
|
|
344 | my $request = http_request GET => "http://www.nethype.de/", sub { |
|
|
345 | my ($body, $hdr) = @_; |
|
|
346 | print "$body\n"; |
|
|
347 | }; |
|
|
348 | |
|
|
349 | undef $request; |
120 | |
350 | |
121 | =cut |
351 | =cut |
122 | |
352 | |
|
|
353 | sub _slot_schedule; |
|
|
354 | sub _slot_schedule($) { |
|
|
355 | my $host = shift; |
|
|
356 | |
|
|
357 | while ($CO_SLOT{$host}[0] < $MAX_PER_HOST) { |
|
|
358 | if (my $cb = shift @{ $CO_SLOT{$host}[1] }) { |
|
|
359 | # somebody wants that slot |
|
|
360 | ++$CO_SLOT{$host}[0]; |
|
|
361 | ++$ACTIVE; |
|
|
362 | |
|
|
363 | $cb->(AnyEvent::Util::guard { |
|
|
364 | --$ACTIVE; |
|
|
365 | --$CO_SLOT{$host}[0]; |
|
|
366 | _slot_schedule $host; |
|
|
367 | }); |
|
|
368 | } else { |
|
|
369 | # nobody wants the slot, maybe we can forget about it |
|
|
370 | delete $CO_SLOT{$host} unless $CO_SLOT{$host}[0]; |
|
|
371 | last; |
|
|
372 | } |
|
|
373 | } |
|
|
374 | } |
|
|
375 | |
|
|
376 | # wait for a free slot on host, call callback |
|
|
377 | sub _get_slot($$) { |
|
|
378 | push @{ $CO_SLOT{$_[0]}[1] }, $_[1]; |
|
|
379 | |
|
|
380 | _slot_schedule $_[0]; |
|
|
381 | } |
|
|
382 | |
|
|
383 | # extract cookies from jar |
|
|
384 | sub cookie_jar_extract($$$$) { |
|
|
385 | my ($jar, $uscheme, $uhost, $upath) = @_; |
|
|
386 | |
|
|
387 | %$jar = () if $jar->{version} != 1; |
|
|
388 | |
|
|
389 | my @cookies; |
|
|
390 | |
|
|
391 | while (my ($chost, $paths) = each %$jar) { |
|
|
392 | next unless ref $paths; |
|
|
393 | |
|
|
394 | if ($chost =~ /^\./) { |
|
|
395 | next unless $chost eq substr $uhost, -length $chost; |
|
|
396 | } elsif ($chost =~ /\./) { |
|
|
397 | next unless $chost eq $uhost; |
|
|
398 | } else { |
|
|
399 | next; |
|
|
400 | } |
|
|
401 | |
|
|
402 | while (my ($cpath, $cookies) = each %$paths) { |
|
|
403 | next unless $cpath eq substr $upath, 0, length $cpath; |
|
|
404 | |
|
|
405 | while (my ($cookie, $kv) = each %$cookies) { |
|
|
406 | next if $uscheme ne "https" && exists $kv->{secure}; |
|
|
407 | |
|
|
408 | if (exists $kv->{expires}) { |
|
|
409 | if (AE::now > parse_date ($kv->{expires})) { |
|
|
410 | delete $cookies->{$cookie}; |
|
|
411 | next; |
|
|
412 | } |
|
|
413 | } |
|
|
414 | |
|
|
415 | my $value = $kv->{value}; |
|
|
416 | |
|
|
417 | if ($value =~ /[=;,[:space:]]/) { |
|
|
418 | $value =~ s/([\\"])/\\$1/g; |
|
|
419 | $value = "\"$value\""; |
|
|
420 | } |
|
|
421 | |
|
|
422 | push @cookies, "$cookie=$value"; |
|
|
423 | } |
|
|
424 | } |
|
|
425 | } |
|
|
426 | |
|
|
427 | \@cookies |
|
|
428 | } |
|
|
429 | |
|
|
430 | # parse set_cookie header into jar |
|
|
431 | sub cookie_jar_set_cookie($$$) { |
|
|
432 | my ($jar, $set_cookie, $uhost) = @_; |
|
|
433 | |
|
|
434 | for ($set_cookie) { |
|
|
435 | # parse NAME=VALUE |
|
|
436 | my @kv; |
|
|
437 | |
|
|
438 | while ( |
|
|
439 | m{ |
|
|
440 | \G\s* |
|
|
441 | (?: |
|
|
442 | expires \s*=\s* ([A-Z][a-z][a-z],\ [^,;]+) |
|
|
443 | | ([^=;,[:space:]]+) \s*=\s* (?: "((?:[^\\"]+|\\.)*)" | ([^=;,[:space:]]*) ) |
|
|
444 | ) |
|
|
445 | }gcxsi |
|
|
446 | ) { |
|
|
447 | my $name = $2; |
|
|
448 | my $value = $4; |
|
|
449 | |
|
|
450 | unless (defined $name) { |
|
|
451 | # expires |
|
|
452 | $name = "expires"; |
|
|
453 | $value = $1; |
|
|
454 | } elsif (!defined $value) { |
|
|
455 | # quoted |
|
|
456 | $value = $3; |
|
|
457 | $value =~ s/\\(.)/$1/gs; |
|
|
458 | } |
|
|
459 | |
|
|
460 | push @kv, lc $name, $value; |
|
|
461 | |
|
|
462 | last unless /\G\s*;/gc; |
|
|
463 | } |
|
|
464 | |
|
|
465 | last unless @kv; |
|
|
466 | |
|
|
467 | my $name = shift @kv; |
|
|
468 | my %kv = (value => shift @kv, @kv); |
|
|
469 | |
|
|
470 | $kv{expires} ||= format_date (AE::now + $kv{"max-age"}) |
|
|
471 | if exists $kv{"max-age"}; |
|
|
472 | |
|
|
473 | my $cdom; |
|
|
474 | my $cpath = (delete $kv{path}) || "/"; |
|
|
475 | |
|
|
476 | if (exists $kv{domain}) { |
|
|
477 | $cdom = delete $kv{domain}; |
|
|
478 | |
|
|
479 | $cdom =~ s/^\.?/./; # make sure it starts with a "." |
|
|
480 | |
|
|
481 | next if $cdom =~ /\.$/; |
|
|
482 | |
|
|
483 | # this is not rfc-like and not netscape-like. go figure. |
|
|
484 | my $ndots = $cdom =~ y/.//; |
|
|
485 | next if $ndots < ($cdom =~ /\.[^.][^.]\.[^.][^.]$/ ? 3 : 2); |
|
|
486 | } else { |
|
|
487 | $cdom = $uhost; |
|
|
488 | } |
|
|
489 | |
|
|
490 | # store it |
|
|
491 | $jar->{version} = 1; |
|
|
492 | $jar->{$cdom}{$cpath}{$name} = \%kv; |
|
|
493 | |
|
|
494 | redo if /\G\s*,/gc; |
|
|
495 | } |
|
|
496 | } |
|
|
497 | |
|
|
498 | # continue to parse $_ for headers and place them into the arg |
|
|
499 | sub parse_hdr() { |
|
|
500 | my %hdr; |
|
|
501 | |
|
|
502 | # things seen, not parsed: |
|
|
503 | # p3pP="NON CUR OTPi OUR NOR UNI" |
|
|
504 | |
|
|
505 | $hdr{lc $1} .= ",$2" |
|
|
506 | while /\G |
|
|
507 | ([^:\000-\037]*): |
|
|
508 | [\011\040]* |
|
|
509 | ((?: [^\012]+ | \012[\011\040] )*) |
|
|
510 | \012 |
|
|
511 | /gxc; |
|
|
512 | |
|
|
513 | /\G$/ |
|
|
514 | or return; |
|
|
515 | |
|
|
516 | # remove the "," prefix we added to all headers above |
|
|
517 | substr $_, 0, 1, "" |
|
|
518 | for values %hdr; |
|
|
519 | |
|
|
520 | \%hdr |
|
|
521 | } |
|
|
522 | |
|
|
523 | our $qr_nlnl = qr{(?<![^\012])\015?\012}; |
|
|
524 | |
|
|
525 | our $TLS_CTX_LOW = { cache => 1, sslv2 => 1 }; |
|
|
526 | our $TLS_CTX_HIGH = { cache => 1, verify => 1, verify_peername => "https" }; |
|
|
527 | |
123 | sub http_request($$$;@) { |
528 | sub http_request($$@) { |
124 | my $cb = pop; |
529 | my $cb = pop; |
125 | my ($method, $url, %arg) = @_; |
530 | my ($method, $url, %arg) = @_; |
126 | |
531 | |
127 | my %hdr; |
532 | my %hdr; |
128 | |
533 | |
|
|
534 | $arg{tls_ctx} = $TLS_CTX_LOW if $arg{tls_ctx} eq "low" || !exists $arg{tls_ctx}; |
|
|
535 | $arg{tls_ctx} = $TLS_CTX_HIGH if $arg{tls_ctx} eq "high"; |
|
|
536 | |
129 | $method = uc $method; |
537 | $method = uc $method; |
130 | |
538 | |
131 | if (my $hdr = delete $arg{headers}) { |
539 | if (my $hdr = $arg{headers}) { |
132 | while (my ($k, $v) = each %$hdr) { |
540 | while (my ($k, $v) = each %$hdr) { |
133 | $hdr{lc $k} = $v; |
541 | $hdr{lc $k} = $v; |
134 | } |
542 | } |
135 | } |
543 | } |
136 | |
544 | |
|
|
545 | # pseudo headers for all subsequent responses |
|
|
546 | my @pseudo = (URL => $url); |
|
|
547 | push @pseudo, Redirect => delete $arg{Redirect} if exists $arg{Redirect}; |
|
|
548 | |
|
|
549 | my $recurse = exists $arg{recurse} ? delete $arg{recurse} : $MAX_RECURSE; |
|
|
550 | |
|
|
551 | return $cb->(undef, { @pseudo, Status => 599, Reason => "Too many redirections" }) |
|
|
552 | if $recurse < 0; |
|
|
553 | |
137 | my $proxy = $arg{proxy} || $PROXY; |
554 | my $proxy = $arg{proxy} || $PROXY; |
138 | my $timeout = $arg{timeout} || $TIMEOUT; |
555 | my $timeout = $arg{timeout} || $TIMEOUT; |
139 | my $recurse = exists $arg{recurse} ? $arg{recurse} : $MAX_RECURSE; |
|
|
140 | |
556 | |
141 | $hdr{"user-agent"} ||= $USERAGENT; |
557 | my ($uscheme, $uauthority, $upath, $query, $fragment) = |
|
|
558 | $url =~ m|(?:([^:/?#]+):)?(?://([^/?#]*))?([^?#]*)(?:(\?[^#]*))?(?:#(.*))?|; |
142 | |
559 | |
143 | my ($host, $port, $path, $scheme); |
560 | $uscheme = lc $uscheme; |
|
|
561 | |
|
|
562 | my $uport = $uscheme eq "http" ? 80 |
|
|
563 | : $uscheme eq "https" ? 443 |
|
|
564 | : return $cb->(undef, { @pseudo, Status => 599, Reason => "Only http and https URL schemes supported" }); |
|
|
565 | |
|
|
566 | $uauthority =~ /^(?: .*\@ )? ([^\@:]+) (?: : (\d+) )?$/x |
|
|
567 | or return $cb->(undef, { @pseudo, Status => 599, Reason => "Unparsable URL" }); |
|
|
568 | |
|
|
569 | my $uhost = $1; |
|
|
570 | $uport = $2 if defined $2; |
|
|
571 | |
|
|
572 | $hdr{host} = defined $2 ? "$uhost:$2" : "$uhost" |
|
|
573 | unless exists $hdr{host}; |
|
|
574 | |
|
|
575 | $uhost =~ s/^\[(.*)\]$/$1/; |
|
|
576 | $upath .= $query if length $query; |
|
|
577 | |
|
|
578 | $upath =~ s%^/?%/%; |
|
|
579 | |
|
|
580 | # cookie processing |
|
|
581 | if (my $jar = $arg{cookie_jar}) { |
|
|
582 | my $cookies = cookie_jar_extract $jar, $uscheme, $uhost, $upath; |
|
|
583 | |
|
|
584 | $hdr{cookie} = join "; ", @$cookies |
|
|
585 | if @$cookies; |
|
|
586 | } |
|
|
587 | |
|
|
588 | my ($rhost, $rport, $rscheme, $rpath); # request host, port, path |
144 | |
589 | |
145 | if ($proxy) { |
590 | if ($proxy) { |
146 | ($host, $port, $scheme) = @$proxy; |
591 | ($rpath, $rhost, $rport, $rscheme) = ($url, @$proxy); |
147 | $path = $url; |
592 | |
|
|
593 | $rscheme = "http" unless defined $rscheme; |
|
|
594 | |
|
|
595 | # don't support https requests over https-proxy transport, |
|
|
596 | # can't be done with tls as spec'ed, unless you double-encrypt. |
|
|
597 | $rscheme = "http" if $uscheme eq "https" && $rscheme eq "https"; |
148 | } else { |
598 | } else { |
149 | ($scheme, my $authority, $path, my $query, my $fragment) = |
599 | ($rhost, $rport, $rscheme, $rpath) = ($uhost, $uport, $uscheme, $upath); |
150 | $url =~ m|(?:([^:/?#]+):)?(?://([^/?#]*))?([^?#]*)(?:\?([^#]*))?(?:#(.*))?|; |
|
|
151 | |
|
|
152 | $port = $scheme eq "http" ? 80 |
|
|
153 | : $scheme eq "https" ? 443 |
|
|
154 | : croak "$url: only http and https URLs supported"; |
|
|
155 | |
|
|
156 | $authority =~ /^(?: .*\@ )? ([^\@:]+) (?: : (\d+) )?$/x |
|
|
157 | or croak "$authority: unparsable URL"; |
|
|
158 | |
|
|
159 | $host = $1; |
|
|
160 | $port = $2 if defined $2; |
|
|
161 | |
|
|
162 | $host =~ s/^\[(.*)\]$/$1/; |
|
|
163 | $path .= "?$query" if length $query; |
|
|
164 | |
|
|
165 | $path = "/" unless $path; |
|
|
166 | |
|
|
167 | $hdr{host} = $host = lc $host; |
|
|
168 | } |
600 | } |
169 | |
601 | |
170 | $scheme = lc $scheme; |
602 | # leave out fragment and query string, just a heuristic |
|
|
603 | $hdr{referer} = "$uscheme://$uauthority$upath" unless exists $hdr{referer}; |
|
|
604 | $hdr{"user-agent"} = $USERAGENT unless exists $hdr{"user-agent"}; |
171 | |
605 | |
172 | my %state; |
|
|
173 | |
|
|
174 | $state{body} = delete $arg{body}; |
|
|
175 | |
|
|
176 | $hdr{"content-length"} = length $state{body}; |
606 | $hdr{"content-length"} = length $arg{body} |
|
|
607 | if length $arg{body} || $method ne "GET"; |
177 | |
608 | |
178 | $state{connect_guard} = AnyEvent::Socket::tcp_connect $host, $port, sub { |
609 | $hdr{connection} = "close TE"; #1.1 |
|
|
610 | $hdr{te} = "trailers" unless exists $hdr{te}; #1.1 |
|
|
611 | |
|
|
612 | my %state = (connect_guard => 1); |
|
|
613 | |
|
|
614 | _get_slot $uhost, sub { |
|
|
615 | $state{slot_guard} = shift; |
|
|
616 | |
|
|
617 | return unless $state{connect_guard}; |
|
|
618 | |
|
|
619 | my $ae_error = 595; # connecting |
|
|
620 | |
|
|
621 | my $connect_cb = sub { |
179 | $state{fh} = shift |
622 | $state{fh} = shift |
180 | or return $cb->(undef, { Status => 599, Reason => "$!" }); |
|
|
181 | |
|
|
182 | delete $state{connect_guard}; # reduce memory usage, save a tree |
|
|
183 | |
|
|
184 | # get handle |
|
|
185 | $state{handle} = new AnyEvent::Handle |
|
|
186 | fh => $state{fh}, |
|
|
187 | ($scheme eq "https" ? (tls => "connect") : ()); |
|
|
188 | |
|
|
189 | # limit the number of persistent connections |
|
|
190 | if ($KA_COUNT{$_[1]} < $MAX_PERSISTENT_PER_HOST) { |
|
|
191 | ++$KA_COUNT{$_[1]}; |
|
|
192 | $state{handle}{ka_count_guard} = AnyEvent::Util::guard { --$KA_COUNT{$_[1]} }; |
|
|
193 | $hdr{connection} = "keep-alive"; |
|
|
194 | delete $hdr{connection}; # keep-alive not yet supported |
|
|
195 | } else { |
|
|
196 | delete $hdr{connection}; |
|
|
197 | } |
|
|
198 | |
|
|
199 | # (re-)configure handle |
|
|
200 | $state{handle}->timeout ($timeout); |
|
|
201 | $state{handle}->on_error (sub { |
|
|
202 | %state = (); |
|
|
203 | $cb->(undef, { Status => 599, Reason => "$!" }); |
|
|
204 | }); |
|
|
205 | $state{handle}->on_eof (sub { |
|
|
206 | %state = (); |
|
|
207 | $cb->(undef, { Status => 599, Reason => "unexpected end-of-file" }); |
|
|
208 | }); |
|
|
209 | |
|
|
210 | # send request |
|
|
211 | $state{handle}->push_write ( |
|
|
212 | "$method $path HTTP/1.0\015\012" |
|
|
213 | . (join "", map "$_: $hdr{$_}\015\012", keys %hdr) |
|
|
214 | . "\015\012" |
|
|
215 | . (delete $state{body}) |
|
|
216 | ); |
|
|
217 | |
|
|
218 | %hdr = (); # reduce memory usage, save a kitten |
|
|
219 | |
|
|
220 | # status line |
|
|
221 | $state{handle}->push_read (line => qr/\015?\012/, sub { |
|
|
222 | $_[1] =~ /^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) \s+ ([^\015\012]+)/ix |
|
|
223 | or return (%state = (), $cb->(undef, { Status => 599, Reason => "invalid server response ($_[1])" })); |
|
|
224 | |
|
|
225 | my %hdr = ( # response headers |
|
|
226 | HTTPVersion => ",$1", |
|
|
227 | Status => ",$2", |
|
|
228 | Reason => ",$3", |
|
|
229 | ); |
|
|
230 | |
|
|
231 | # headers, could be optimized a bit |
|
|
232 | $state{handle}->unshift_read (line => qr/\015?\012\015?\012/, sub { |
|
|
233 | for ("$_[1]\012") { |
|
|
234 | # we support spaces in field names, as lotus domino |
|
|
235 | # creates them. |
|
|
236 | $hdr{lc $1} .= ",$2" |
|
|
237 | while /\G |
|
|
238 | ([^:\000-\037]+): |
|
|
239 | [\011\040]* |
|
|
240 | ((?: [^\015\012]+ | \015?\012[\011\040] )*) |
|
|
241 | \015?\012 |
|
|
242 | /gxc; |
|
|
243 | |
|
|
244 | /\G$/ |
|
|
245 | or return $cb->(undef, { Status => 599, Reason => "garbled response headers" }); |
|
|
246 | } |
623 | or do { |
247 | |
624 | my $err = "$!"; |
248 | substr $_, 0, 1, "" |
|
|
249 | for values %hdr; |
|
|
250 | |
|
|
251 | if ($method eq "HEAD") { |
|
|
252 | %state = (); |
625 | %state = (); |
|
|
626 | return $cb->(undef, { @pseudo, Status => $ae_error, Reason => $err }); |
|
|
627 | }; |
|
|
628 | |
|
|
629 | return unless delete $state{connect_guard}; |
|
|
630 | |
|
|
631 | # get handle |
|
|
632 | $state{handle} = new AnyEvent::Handle |
|
|
633 | fh => $state{fh}, |
|
|
634 | peername => $rhost, |
|
|
635 | tls_ctx => $arg{tls_ctx}, |
|
|
636 | # these need to be reconfigured on keepalive handles |
|
|
637 | timeout => $timeout, |
|
|
638 | on_error => sub { |
|
|
639 | %state = (); |
|
|
640 | $cb->(undef, { @pseudo, Status => $ae_error, Reason => $_[2] }); |
|
|
641 | }, |
|
|
642 | on_eof => sub { |
|
|
643 | %state = (); |
|
|
644 | $cb->(undef, { @pseudo, Status => $ae_error, Reason => "Unexpected end-of-file" }); |
|
|
645 | }, |
|
|
646 | ; |
|
|
647 | |
|
|
648 | # limit the number of persistent connections |
|
|
649 | # keepalive not yet supported |
|
|
650 | # if ($KA_COUNT{$_[1]} < $MAX_PERSISTENT_PER_HOST) { |
|
|
651 | # ++$KA_COUNT{$_[1]}; |
|
|
652 | # $state{handle}{ka_count_guard} = AnyEvent::Util::guard { |
|
|
653 | # --$KA_COUNT{$_[1]} |
|
|
654 | # }; |
|
|
655 | # $hdr{connection} = "keep-alive"; |
|
|
656 | # } |
|
|
657 | |
|
|
658 | $state{handle}->starttls ("connect") if $rscheme eq "https"; |
|
|
659 | |
|
|
660 | # handle actual, non-tunneled, request |
|
|
661 | my $handle_actual_request = sub { |
|
|
662 | $ae_error = 596; # request phase |
|
|
663 | |
|
|
664 | $state{handle}->starttls ("connect") if $uscheme eq "https" && !exists $state{handle}{tls}; |
|
|
665 | |
|
|
666 | # send request |
|
|
667 | $state{handle}->push_write ( |
|
|
668 | "$method $rpath HTTP/1.1\015\012" |
|
|
669 | . (join "", map "\u$_: $hdr{$_}\015\012", grep defined $hdr{$_}, keys %hdr) |
|
|
670 | . "\015\012" |
|
|
671 | . (delete $arg{body}) |
|
|
672 | ); |
|
|
673 | |
|
|
674 | # return if error occured during push_write() |
|
|
675 | return unless %state; |
|
|
676 | |
|
|
677 | %hdr = (); # reduce memory usage, save a kitten, also make it possible to re-use |
|
|
678 | |
|
|
679 | # status line and headers |
|
|
680 | $state{read_response} = sub { |
|
|
681 | for ("$_[1]") { |
|
|
682 | y/\015//d; # weed out any \015, as they show up in the weirdest of places. |
|
|
683 | |
|
|
684 | /^HTTP\/0*([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\012]*) )? \012/gxci |
|
|
685 | or return (%state = (), $cb->(undef, { @pseudo, Status => 599, Reason => "Invalid server response" })); |
|
|
686 | |
|
|
687 | # 100 Continue handling |
|
|
688 | # should not happen as we don't send expect: 100-continue, |
|
|
689 | # but we handle it just in case. |
|
|
690 | # since we send the request body regardless, if we get an error |
|
|
691 | # we are out of-sync, which we currently do NOT handle correctly. |
|
|
692 | return $state{handle}->push_read (line => $qr_nlnl, $state{read_response}) |
|
|
693 | if $2 eq 100; |
|
|
694 | |
|
|
695 | push @pseudo, |
|
|
696 | HTTPVersion => $1, |
|
|
697 | Status => $2, |
|
|
698 | Reason => $3, |
|
|
699 | ; |
|
|
700 | |
|
|
701 | my $hdr = parse_hdr |
|
|
702 | or return (%state = (), $cb->(undef, { @pseudo, Status => 599, Reason => "Garbled response headers" })); |
|
|
703 | |
|
|
704 | %hdr = (%$hdr, @pseudo); |
|
|
705 | } |
|
|
706 | |
|
|
707 | # redirect handling |
|
|
708 | # microsoft and other shitheads don't give a shit for following standards, |
|
|
709 | # try to support some common forms of broken Location headers. |
|
|
710 | if ($hdr{location} !~ /^(?: $ | [^:\/?\#]+ : )/x) { |
|
|
711 | $hdr{location} =~ s/^\.\/+//; |
|
|
712 | |
|
|
713 | my $url = "$rscheme://$uhost:$uport"; |
|
|
714 | |
|
|
715 | unless ($hdr{location} =~ s/^\///) { |
|
|
716 | $url .= $upath; |
|
|
717 | $url =~ s/\/[^\/]*$//; |
|
|
718 | } |
|
|
719 | |
|
|
720 | $hdr{location} = "$url/$hdr{location}"; |
|
|
721 | } |
|
|
722 | |
|
|
723 | my $redirect; |
|
|
724 | |
|
|
725 | if ($recurse) { |
|
|
726 | my $status = $hdr{Status}; |
|
|
727 | |
|
|
728 | # industry standard is to redirect POST as GET for |
|
|
729 | # 301, 302 and 303, in contrast to HTTP/1.0 and 1.1. |
|
|
730 | # also, the UA should ask the user for 301 and 307 and POST, |
|
|
731 | # industry standard seems to be to simply follow. |
|
|
732 | # we go with the industry standard. |
|
|
733 | if ($status == 301 or $status == 302 or $status == 303) { |
|
|
734 | # HTTP/1.1 is unclear on how to mutate the method |
|
|
735 | $method = "GET" unless $method eq "HEAD"; |
|
|
736 | $redirect = 1; |
|
|
737 | } elsif ($status == 307) { |
|
|
738 | $redirect = 1; |
|
|
739 | } |
|
|
740 | } |
|
|
741 | |
|
|
742 | my $finish = sub { # ($data, $err_status, $err_reason[, $keepalive]) |
|
|
743 | my $may_keep_alive = $_[3]; |
|
|
744 | |
|
|
745 | $state{handle}->destroy if $state{handle}; |
|
|
746 | %state = (); |
|
|
747 | |
|
|
748 | if (defined $_[1]) { |
|
|
749 | $hdr{OrigStatus} = $hdr{Status}; $hdr{Status} = $_[1]; |
|
|
750 | $hdr{OrigReason} = $hdr{Reason}; $hdr{Reason} = $_[2]; |
|
|
751 | } |
|
|
752 | |
|
|
753 | # set-cookie processing |
|
|
754 | if ($arg{cookie_jar}) { |
|
|
755 | cookie_jar_set_cookie $arg{cookie_jar}, $hdr{"set-cookie"}, $uhost; |
|
|
756 | } |
|
|
757 | |
|
|
758 | if ($redirect && exists $hdr{location}) { |
|
|
759 | # we ignore any errors, as it is very common to receive |
|
|
760 | # Content-Length != 0 but no actual body |
|
|
761 | # we also access %hdr, as $_[1] might be an erro |
|
|
762 | http_request ( |
|
|
763 | $method => $hdr{location}, |
|
|
764 | %arg, |
|
|
765 | recurse => $recurse - 1, |
|
|
766 | Redirect => [$_[0], \%hdr], |
|
|
767 | $cb); |
|
|
768 | } else { |
253 | $cb->(undef, \%hdr); |
769 | $cb->($_[0], \%hdr); |
|
|
770 | } |
|
|
771 | }; |
|
|
772 | |
|
|
773 | $ae_error = 597; # body phase |
|
|
774 | |
|
|
775 | my $len = $hdr{"content-length"}; |
|
|
776 | |
|
|
777 | if (!$redirect && $arg{on_header} && !$arg{on_header}(\%hdr)) { |
|
|
778 | $finish->(undef, 598 => "Request cancelled by on_header"); |
|
|
779 | } elsif ( |
|
|
780 | $hdr{Status} =~ /^(?:1..|204|205|304)$/ |
|
|
781 | or $method eq "HEAD" |
|
|
782 | or (defined $len && !$len) |
|
|
783 | ) { |
|
|
784 | # no body |
|
|
785 | $finish->("", undef, undef, 1); |
254 | } else { |
786 | } else { |
255 | if (exists $hdr{"content-length"}) { |
787 | # body handling, many different code paths |
|
|
788 | # - no body expected |
|
|
789 | # - want_body_handle |
|
|
790 | # - te chunked |
|
|
791 | # - 2x length known (with or without on_body) |
|
|
792 | # - 2x length not known (with or without on_body) |
|
|
793 | if (!$redirect && $arg{want_body_handle}) { |
|
|
794 | $_[0]->on_eof (undef); |
|
|
795 | $_[0]->on_error (undef); |
|
|
796 | $_[0]->on_read (undef); |
|
|
797 | |
|
|
798 | $finish->(delete $state{handle}); |
|
|
799 | |
|
|
800 | } elsif ($hdr{"transfer-encoding"} =~ /\bchunked\b/i) { |
|
|
801 | my $cl = 0; |
|
|
802 | my $body = undef; |
|
|
803 | my $on_body = $arg{on_body} || sub { $body .= shift; 1 }; |
|
|
804 | |
|
|
805 | my $read_chunk; $read_chunk = sub { |
|
|
806 | $_[1] =~ /^([0-9a-fA-F]+)/ |
|
|
807 | or $finish->(undef, $ae_error => "Garbled chunked transfer encoding"); |
|
|
808 | |
|
|
809 | my $len = hex $1; |
|
|
810 | |
|
|
811 | if ($len) { |
|
|
812 | $cl += $len; |
|
|
813 | |
256 | $_[0]->unshift_read (chunk => $hdr{"content-length"}, sub { |
814 | $_[0]->push_read (chunk => $len, sub { |
257 | # could cache persistent connection now |
815 | $on_body->($_[1], \%hdr) |
258 | if ($hdr{connection} =~ /\bkeep-alive\b/i) { |
816 | or return $finish->(undef, 598 => "Request cancelled by on_body"); |
259 | # but we don't, due to misdesigns, this is annoyingly complex |
817 | |
|
|
818 | $_[0]->push_read (line => sub { |
|
|
819 | length $_[1] |
|
|
820 | and return $finish->(undef, $ae_error => "Garbled chunked transfer encoding"); |
|
|
821 | $_[0]->push_read (line => $read_chunk); |
|
|
822 | }); |
|
|
823 | }); |
|
|
824 | } else { |
|
|
825 | $hdr{"content-length"} ||= $cl; |
|
|
826 | |
|
|
827 | $_[0]->push_read (line => $qr_nlnl, sub { |
|
|
828 | if (length $_[1]) { |
|
|
829 | for ("$_[1]") { |
|
|
830 | y/\015//d; # weed out any \015, as they show up in the weirdest of places. |
|
|
831 | |
|
|
832 | my $hdr = parse_hdr |
|
|
833 | or return $finish->(undef, $ae_error => "Garbled response trailers"); |
|
|
834 | |
|
|
835 | %hdr = (%hdr, %$hdr); |
|
|
836 | } |
|
|
837 | } |
|
|
838 | |
|
|
839 | $finish->($body, undef, undef, 1); |
|
|
840 | }); |
|
|
841 | } |
260 | }; |
842 | }; |
261 | |
843 | |
|
|
844 | $_[0]->push_read (line => $read_chunk); |
|
|
845 | |
|
|
846 | } elsif ($arg{on_body}) { |
|
|
847 | if ($len) { |
|
|
848 | $_[0]->on_read (sub { |
|
|
849 | $len -= length $_[0]{rbuf}; |
|
|
850 | |
|
|
851 | $arg{on_body}(delete $_[0]{rbuf}, \%hdr) |
|
|
852 | or return $finish->(undef, 598 => "Request cancelled by on_body"); |
|
|
853 | |
|
|
854 | $len > 0 |
|
|
855 | or $finish->("", undef, undef, 1); |
|
|
856 | }); |
262 | %state = (); |
857 | } else { |
263 | $cb->($_[1], \%hdr); |
858 | $_[0]->on_eof (sub { |
|
|
859 | $finish->(""); |
|
|
860 | }); |
|
|
861 | $_[0]->on_read (sub { |
|
|
862 | $arg{on_body}(delete $_[0]{rbuf}, \%hdr) |
|
|
863 | or $finish->(undef, 598 => "Request cancelled by on_body"); |
|
|
864 | }); |
|
|
865 | } |
|
|
866 | } else { |
|
|
867 | $_[0]->on_eof (undef); |
|
|
868 | |
|
|
869 | if ($len) { |
|
|
870 | $_[0]->on_read (sub { |
|
|
871 | $finish->((substr delete $_[0]{rbuf}, 0, $len, ""), undef, undef, 1) |
|
|
872 | if $len <= length $_[0]{rbuf}; |
|
|
873 | }); |
|
|
874 | } else { |
|
|
875 | $_[0]->on_error (sub { |
|
|
876 | ($! == Errno::EPIPE || !$!) |
|
|
877 | ? $finish->(delete $_[0]{rbuf}) |
|
|
878 | : $finish->(undef, $ae_error => $_[2]); |
|
|
879 | }); |
|
|
880 | $_[0]->on_read (sub { }); |
|
|
881 | } |
264 | }); |
882 | } |
|
|
883 | } |
|
|
884 | }; |
|
|
885 | |
|
|
886 | $state{handle}->push_read (line => $qr_nlnl, $state{read_response}); |
|
|
887 | }; |
|
|
888 | |
|
|
889 | # now handle proxy-CONNECT method |
|
|
890 | if ($proxy && $uscheme eq "https") { |
|
|
891 | # oh dear, we have to wrap it into a connect request |
|
|
892 | |
|
|
893 | # maybe re-use $uauthority with patched port? |
|
|
894 | $state{handle}->push_write ("CONNECT $uhost:$uport HTTP/1.0\015\012Host: $uhost\015\012\015\012"); |
|
|
895 | $state{handle}->push_read (line => $qr_nlnl, sub { |
|
|
896 | $_[1] =~ /^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\015\012]*) )?/ix |
|
|
897 | or return (%state = (), $cb->(undef, { @pseudo, Status => 599, Reason => "Invalid proxy connect response ($_[1])" })); |
|
|
898 | |
|
|
899 | if ($2 == 200) { |
|
|
900 | $rpath = $upath; |
|
|
901 | &$handle_actual_request; |
265 | } else { |
902 | } else { |
266 | # too bad, need to read until we get an error or EOF, |
|
|
267 | # no way to detect winged data. |
|
|
268 | $_[0]->on_error (sub { |
|
|
269 | %state = (); |
903 | %state = (); |
270 | $cb->($_[0]{rbuf}, \%hdr); |
904 | $cb->(undef, { @pseudo, Status => $2, Reason => $3 }); |
271 | }); |
|
|
272 | $_[0]->on_eof (undef); |
|
|
273 | $_[0]->on_read (sub { }); |
|
|
274 | } |
905 | } |
275 | } |
906 | }); |
|
|
907 | } else { |
|
|
908 | &$handle_actual_request; |
276 | }); |
909 | } |
277 | }); |
910 | }; |
278 | }, sub { |
911 | |
279 | $timeout |
912 | my $tcp_connect = $arg{tcp_connect} |
|
|
913 | || do { require AnyEvent::Socket; \&AnyEvent::Socket::tcp_connect }; |
|
|
914 | |
|
|
915 | $state{connect_guard} = $tcp_connect->($rhost, $rport, $connect_cb, $arg{on_prepare} || sub { $timeout }); |
|
|
916 | |
280 | }; |
917 | }; |
281 | |
918 | |
282 | defined wantarray && AnyEvent::Util::guard { %state = () } |
919 | defined wantarray && AnyEvent::Util::guard { %state = () } |
283 | } |
920 | } |
284 | |
921 | |
285 | sub http_get($$;@) { |
922 | sub http_get($@) { |
286 | unshift @_, "GET"; |
923 | unshift @_, "GET"; |
287 | &http_request |
924 | &http_request |
288 | } |
925 | } |
289 | |
926 | |
290 | sub http_head($$;@) { |
927 | sub http_head($@) { |
291 | unshift @_, "HEAD"; |
928 | unshift @_, "HEAD"; |
292 | &http_request |
929 | &http_request |
293 | } |
930 | } |
294 | |
931 | |
295 | sub http_post($$$;@) { |
932 | sub http_post($$@) { |
|
|
933 | my $url = shift; |
296 | unshift @_, "POST", "body"; |
934 | unshift @_, "POST", $url, "body"; |
297 | &http_request |
935 | &http_request |
298 | } |
936 | } |
299 | |
937 | |
|
|
938 | =back |
|
|
939 | |
|
|
940 | =head2 DNS CACHING |
|
|
941 | |
|
|
942 | AnyEvent::HTTP uses the AnyEvent::Socket::tcp_connect function for |
|
|
943 | the actual connection, which in turn uses AnyEvent::DNS to resolve |
|
|
944 | hostnames. The latter is a simple stub resolver and does no caching |
|
|
945 | on its own. If you want DNS caching, you currently have to provide |
|
|
946 | your own default resolver (by storing a suitable resolver object in |
|
|
947 | C<$AnyEvent::DNS::RESOLVER>). |
|
|
948 | |
300 | =head2 GLOBAL FUNCTIONS AND VARIABLES |
949 | =head2 GLOBAL FUNCTIONS AND VARIABLES |
301 | |
950 | |
302 | =over 4 |
951 | =over 4 |
303 | |
952 | |
304 | =item AnyEvent::HTTP::set_proxy "proxy-url" |
953 | =item AnyEvent::HTTP::set_proxy "proxy-url" |
305 | |
954 | |
306 | Sets the default proxy server to use. The proxy-url must begin with a |
955 | Sets the default proxy server to use. The proxy-url must begin with a |
307 | string of the form C<http://host:port> (optionally C<https:...>). |
956 | string of the form C<http://host:port> (optionally C<https:...>), croaks |
|
|
957 | otherwise. |
|
|
958 | |
|
|
959 | To clear an already-set proxy, use C<undef>. |
|
|
960 | |
|
|
961 | =item $date = AnyEvent::HTTP::format_date $timestamp |
|
|
962 | |
|
|
963 | Takes a POSIX timestamp (seconds since the epoch) and formats it as a HTTP |
|
|
964 | Date (RFC 2616). |
|
|
965 | |
|
|
966 | =item $timestamp = AnyEvent::HTTP::parse_date $date |
|
|
967 | |
|
|
968 | Takes a HTTP Date (RFC 2616) or a Cookie date (netscape cookie spec) and |
|
|
969 | returns the corresponding POSIX timestamp, or C<undef> if the date cannot |
|
|
970 | be parsed. |
308 | |
971 | |
309 | =item $AnyEvent::HTTP::MAX_RECURSE |
972 | =item $AnyEvent::HTTP::MAX_RECURSE |
310 | |
973 | |
311 | The default value for the C<recurse> request parameter (default: C<10>). |
974 | The default value for the C<recurse> request parameter (default: C<10>). |
312 | |
975 | |
313 | =item $AnyEvent::HTTP::USERAGENT |
976 | =item $AnyEvent::HTTP::USERAGENT |
314 | |
977 | |
315 | The default value for the C<User-Agent> header (the default is |
978 | The default value for the C<User-Agent> header (the default is |
316 | C<Mozilla/5.0 (compatible; AnyEvent::HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)>). |
979 | C<Mozilla/5.0 (compatible; U; AnyEvent-HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)>). |
317 | |
980 | |
318 | =item $AnyEvent::HTTP::MAX_PERSISTENT |
981 | =item $AnyEvent::HTTP::MAX_PER_HOST |
319 | |
982 | |
320 | The maximum number of persistent connections to keep open (default: 8). |
983 | The maximum number of concurrent connections to the same host (identified |
|
|
984 | by the hostname). If the limit is exceeded, then the additional requests |
|
|
985 | are queued until previous connections are closed. |
321 | |
986 | |
322 | Not implemented currently. |
987 | The default value for this is C<4>, and it is highly advisable to not |
|
|
988 | increase it. |
323 | |
989 | |
324 | =item $AnyEvent::HTTP::PERSISTENT_TIMEOUT |
990 | =item $AnyEvent::HTTP::ACTIVE |
325 | |
991 | |
326 | The maximum time to cache a persistent connection, in seconds (default: 2). |
992 | The number of active connections. This is not the number of currently |
327 | |
993 | running requests, but the number of currently open and non-idle TCP |
328 | Not implemented currently. |
994 | connections. This number of can be useful for load-leveling. |
329 | |
995 | |
330 | =back |
996 | =back |
331 | |
997 | |
332 | =cut |
998 | =cut |
333 | |
999 | |
|
|
1000 | our @month = qw(Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec); |
|
|
1001 | our @weekday = qw(Sun Mon Tue Wed Thu Fri Sat); |
|
|
1002 | |
|
|
1003 | sub format_date($) { |
|
|
1004 | my ($time) = @_; |
|
|
1005 | |
|
|
1006 | # RFC 822/1123 format |
|
|
1007 | my ($S, $M, $H, $mday, $mon, $year, $wday, $yday, undef) = gmtime $time; |
|
|
1008 | |
|
|
1009 | sprintf "%s, %02d %s %04d %02d:%02d:%02d GMT", |
|
|
1010 | $weekday[$wday], $mday, $month[$mon], $year + 1900, |
|
|
1011 | $H, $M, $S; |
|
|
1012 | } |
|
|
1013 | |
|
|
1014 | sub parse_date($) { |
|
|
1015 | my ($date) = @_; |
|
|
1016 | |
|
|
1017 | my ($d, $m, $y, $H, $M, $S); |
|
|
1018 | |
|
|
1019 | if ($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$/) { |
|
|
1020 | # RFC 822/1123, required by RFC 2616 (with " ") |
|
|
1021 | # cookie dates (with "-") |
|
|
1022 | |
|
|
1023 | ($d, $m, $y, $H, $M, $S) = ($1, $2, $3, $4, $5, $6); |
|
|
1024 | |
|
|
1025 | } elsif ($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$/) { |
|
|
1026 | # RFC 850 |
|
|
1027 | ($d, $m, $y, $H, $M, $S) = ($1, $2, $3 < 69 ? $3 + 2000 : $3 + 1900, $4, $5, $6); |
|
|
1028 | |
|
|
1029 | } elsif ($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])$/) { |
|
|
1030 | # ISO C's asctime |
|
|
1031 | ($d, $m, $y, $H, $M, $S) = ($2, $1, $6, $3, $4, $5); |
|
|
1032 | } |
|
|
1033 | # other formats fail in the loop below |
|
|
1034 | |
|
|
1035 | for (0..11) { |
|
|
1036 | if ($m eq $month[$_]) { |
|
|
1037 | require Time::Local; |
|
|
1038 | return Time::Local::timegm ($S, $M, $H, $d, $_, $y); |
|
|
1039 | } |
|
|
1040 | } |
|
|
1041 | |
|
|
1042 | undef |
|
|
1043 | } |
|
|
1044 | |
334 | sub set_proxy($) { |
1045 | sub set_proxy($) { |
|
|
1046 | if (length $_[0]) { |
335 | $PROXY = [$2, $3 || 3128, $1] if $_[0] =~ m%^(https?):// ([^:/]+) (?: : (\d*) )?%ix; |
1047 | $_[0] =~ m%^(https?):// ([^:/]+) (?: : (\d*) )?%ix |
|
|
1048 | or Carp::croak "$_[0]: invalid proxy URL"; |
|
|
1049 | $PROXY = [$2, $3 || 3128, $1] |
|
|
1050 | } else { |
|
|
1051 | undef $PROXY; |
|
|
1052 | } |
336 | } |
1053 | } |
337 | |
1054 | |
338 | # initialise proxy from environment |
1055 | # initialise proxy from environment |
|
|
1056 | eval { |
339 | set_proxy $ENV{http_proxy}; |
1057 | set_proxy $ENV{http_proxy}; |
|
|
1058 | }; |
|
|
1059 | |
|
|
1060 | =head2 SOCKS PROXIES |
|
|
1061 | |
|
|
1062 | Socks proxies are not directly supported by AnyEvent::HTTP. You can |
|
|
1063 | compile your perl to support socks, or use an external program such as |
|
|
1064 | F<socksify> (dante) or F<tsocks> to make your program use a socks proxy |
|
|
1065 | transparently. |
|
|
1066 | |
|
|
1067 | Alternatively, for AnyEvent::HTTP only, you can use your own |
|
|
1068 | C<tcp_connect> function that does the proxy handshake - here is an example |
|
|
1069 | that works with socks4a proxies: |
|
|
1070 | |
|
|
1071 | use Errno; |
|
|
1072 | use AnyEvent::Util; |
|
|
1073 | use AnyEvent::Socket; |
|
|
1074 | use AnyEvent::Handle; |
|
|
1075 | |
|
|
1076 | # host, port and username of/for your socks4a proxy |
|
|
1077 | my $socks_host = "10.0.0.23"; |
|
|
1078 | my $socks_port = 9050; |
|
|
1079 | my $socks_user = ""; |
|
|
1080 | |
|
|
1081 | sub socks4a_connect { |
|
|
1082 | my ($host, $port, $connect_cb, $prepare_cb) = @_; |
|
|
1083 | |
|
|
1084 | my $hdl = new AnyEvent::Handle |
|
|
1085 | connect => [$socks_host, $socks_port], |
|
|
1086 | on_prepare => sub { $prepare_cb->($_[0]{fh}) }, |
|
|
1087 | on_error => sub { $connect_cb->() }, |
|
|
1088 | ; |
|
|
1089 | |
|
|
1090 | $hdl->push_write (pack "CCnNZ*Z*", 4, 1, $port, 1, $socks_user, $host); |
|
|
1091 | |
|
|
1092 | $hdl->push_read (chunk => 8, sub { |
|
|
1093 | my ($hdl, $chunk) = @_; |
|
|
1094 | my ($status, $port, $ipn) = unpack "xCna4", $chunk; |
|
|
1095 | |
|
|
1096 | if ($status == 0x5a) { |
|
|
1097 | $connect_cb->($hdl->{fh}, (format_address $ipn) . ":$port"); |
|
|
1098 | } else { |
|
|
1099 | $! = Errno::ENXIO; $connect_cb->(); |
|
|
1100 | } |
|
|
1101 | }); |
|
|
1102 | |
|
|
1103 | $hdl |
|
|
1104 | } |
|
|
1105 | |
|
|
1106 | Use C<socks4a_connect> instead of C<tcp_connect> when doing C<http_request>s, |
|
|
1107 | possibly after switching off other proxy types: |
|
|
1108 | |
|
|
1109 | AnyEvent::HTTP::set_proxy undef; # usually you do not want other proxies |
|
|
1110 | |
|
|
1111 | http_get 'http://www.google.com', tcp_connect => \&socks4a_connect, sub { |
|
|
1112 | my ($data, $headers) = @_; |
|
|
1113 | ... |
|
|
1114 | }; |
340 | |
1115 | |
341 | =head1 SEE ALSO |
1116 | =head1 SEE ALSO |
342 | |
1117 | |
343 | L<AnyEvent>. |
1118 | L<AnyEvent>. |
344 | |
1119 | |
345 | =head1 AUTHOR |
1120 | =head1 AUTHOR |
346 | |
1121 | |
347 | Marc Lehmann <schmorp@schmorp.de> |
1122 | Marc Lehmann <schmorp@schmorp.de> |
348 | http://home.schmorp.de/ |
1123 | http://home.schmorp.de/ |
|
|
1124 | |
|
|
1125 | With many thanks to Дмитрий Шалашов, who provided countless |
|
|
1126 | testcases and bugreports. |
349 | |
1127 | |
350 | =cut |
1128 | =cut |
351 | |
1129 | |
352 | 1 |
1130 | 1 |
353 | |
1131 | |