1 |
root |
1.1 |
=head1 NAME |
2 |
|
|
|
3 |
|
|
AnyEvent::HTTP - simple but non-blocking HTTP/HTTPS client |
4 |
|
|
|
5 |
|
|
=head1 SYNOPSIS |
6 |
|
|
|
7 |
|
|
use AnyEvent::HTTP; |
8 |
|
|
|
9 |
root |
1.17 |
http_get "http://www.nethype.de/", sub { print $_[1] }; |
10 |
|
|
|
11 |
|
|
# ... do something else here |
12 |
|
|
|
13 |
root |
1.1 |
=head1 DESCRIPTION |
14 |
|
|
|
15 |
|
|
This module is an L<AnyEvent> user, you need to make sure that you use and |
16 |
|
|
run a supported event loop. |
17 |
|
|
|
18 |
root |
1.11 |
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 |
|
|
|
33 |
root |
1.1 |
=head2 METHODS |
34 |
|
|
|
35 |
|
|
=over 4 |
36 |
|
|
|
37 |
|
|
=cut |
38 |
|
|
|
39 |
|
|
package AnyEvent::HTTP; |
40 |
|
|
|
41 |
|
|
use strict; |
42 |
|
|
no warnings; |
43 |
|
|
|
44 |
root |
1.41 |
use Errno (); |
45 |
root |
1.1 |
|
46 |
root |
1.51 |
use AnyEvent 5.0 (); |
47 |
root |
1.1 |
use AnyEvent::Util (); |
48 |
|
|
use AnyEvent::Handle (); |
49 |
|
|
|
50 |
|
|
use base Exporter::; |
51 |
|
|
|
52 |
root |
1.65 |
our $VERSION = '1.5'; |
53 |
root |
1.1 |
|
54 |
root |
1.17 |
our @EXPORT = qw(http_get http_post http_head http_request); |
55 |
root |
1.1 |
|
56 |
root |
1.40 |
our $USERAGENT = "Mozilla/5.0 (compatible; U; AnyEvent-HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)"; |
57 |
root |
1.3 |
our $MAX_RECURSE = 10; |
58 |
root |
1.2 |
our $MAX_PERSISTENT = 8; |
59 |
|
|
our $PERSISTENT_TIMEOUT = 2; |
60 |
|
|
our $TIMEOUT = 300; |
61 |
root |
1.1 |
|
62 |
|
|
# changing these is evil |
63 |
root |
1.43 |
our $MAX_PERSISTENT_PER_HOST = 0; |
64 |
root |
1.11 |
our $MAX_PER_HOST = 4; |
65 |
root |
1.1 |
|
66 |
root |
1.2 |
our $PROXY; |
67 |
root |
1.14 |
our $ACTIVE = 0; |
68 |
root |
1.2 |
|
69 |
root |
1.1 |
my %KA_COUNT; # number of open keep-alive connections per host |
70 |
root |
1.11 |
my %CO_SLOT; # number of open connections, and wait queue, per host |
71 |
root |
1.1 |
|
72 |
|
|
=item http_get $url, key => value..., $cb->($data, $headers) |
73 |
|
|
|
74 |
|
|
Executes an HTTP-GET request. See the http_request function for details on |
75 |
root |
1.29 |
additional parameters and the return value. |
76 |
root |
1.1 |
|
77 |
root |
1.5 |
=item http_head $url, key => value..., $cb->($data, $headers) |
78 |
|
|
|
79 |
root |
1.29 |
Executes an HTTP-HEAD request. See the http_request function for details |
80 |
|
|
on additional parameters and the return value. |
81 |
root |
1.5 |
|
82 |
|
|
=item http_post $url, $body, key => value..., $cb->($data, $headers) |
83 |
root |
1.3 |
|
84 |
root |
1.26 |
Executes an HTTP-POST request with a request body of C<$body>. See the |
85 |
root |
1.29 |
http_request function for details on additional parameters and the return |
86 |
|
|
value. |
87 |
root |
1.3 |
|
88 |
root |
1.1 |
=item http_request $method => $url, key => value..., $cb->($data, $headers) |
89 |
|
|
|
90 |
|
|
Executes a HTTP request of type C<$method> (e.g. C<GET>, C<POST>). The URL |
91 |
|
|
must be an absolute http or https URL. |
92 |
|
|
|
93 |
root |
1.29 |
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 |
root |
1.58 |
destroyed before the callback is called, the request will be cancelled. |
97 |
root |
1.29 |
|
98 |
root |
1.42 |
The callback will be called with the response body data as first argument |
99 |
root |
1.68 |
(or C<undef> if an error occured), and a hash-ref with response headers |
100 |
|
|
(and trailers) as second argument. |
101 |
root |
1.2 |
|
102 |
root |
1.7 |
All the headers in that hash are lowercased. In addition to the response |
103 |
root |
1.55 |
headers, the "pseudo-headers" (uppercase to avoid clashing with possible |
104 |
|
|
response headers) C<HTTPVersion>, C<Status> and C<Reason> contain the |
105 |
root |
1.64 |
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 |
root |
1.55 |
|
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 |
|
|
$headers->{Redirect}[1]{Redirect} >> will then contain the original |
121 |
|
|
response, and so on. |
122 |
root |
1.20 |
|
123 |
root |
1.32 |
If the server sends a header multiple times, then their contents will be |
124 |
|
|
joined together with a comma (C<,>), as per the HTTP spec. |
125 |
root |
1.2 |
|
126 |
|
|
If an internal error occurs, such as not being able to resolve a hostname, |
127 |
root |
1.77 |
then C<$data> will be C<undef>, C<< $headers->{Status} >> will be |
128 |
|
|
C<590>-C<599> and the C<Reason> pseudo-header will contain an error |
129 |
|
|
message. Currently the following status codes are used: |
130 |
|
|
|
131 |
|
|
=over 4 |
132 |
|
|
|
133 |
|
|
=item 595 - errors during connection etsbalishment, proxy handshake. |
134 |
|
|
|
135 |
|
|
=item 596 - errors during TLS negotiation, request sending and header processing. |
136 |
|
|
|
137 |
root |
1.78 |
=item 597 - errors during body receiving or processing. |
138 |
root |
1.77 |
|
139 |
root |
1.78 |
=item 598 - user aborted request via C<on_header> or C<on_body>. |
140 |
root |
1.77 |
|
141 |
|
|
=item 599 - other, usually nonretryable, errors (garbled URL etc.). |
142 |
|
|
|
143 |
|
|
=back |
144 |
root |
1.2 |
|
145 |
root |
1.6 |
A typical callback might look like this: |
146 |
|
|
|
147 |
|
|
sub { |
148 |
|
|
my ($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 |
|
|
} |
156 |
|
|
|
157 |
root |
1.1 |
Additional parameters are key-value pairs, and are fully optional. They |
158 |
|
|
include: |
159 |
|
|
|
160 |
|
|
=over 4 |
161 |
|
|
|
162 |
root |
1.3 |
=item recurse => $count (default: $MAX_RECURSE) |
163 |
root |
1.1 |
|
164 |
|
|
Whether to recurse requests or not, e.g. on redirects, authentication |
165 |
root |
1.3 |
retries and so on, and how often to do so. |
166 |
root |
1.1 |
|
167 |
|
|
=item headers => hashref |
168 |
|
|
|
169 |
root |
1.68 |
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 |
root |
1.69 |
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). |
174 |
root |
1.1 |
|
175 |
|
|
=item timeout => $seconds |
176 |
|
|
|
177 |
|
|
The time-out to use for various stages - each connect attempt will reset |
178 |
root |
1.51 |
the timeout, as will read or write activity, i.e. this is not an overall |
179 |
|
|
timeout. |
180 |
|
|
|
181 |
|
|
Default timeout is 5 minutes. |
182 |
root |
1.2 |
|
183 |
|
|
=item proxy => [$host, $port[, $scheme]] or undef |
184 |
|
|
|
185 |
|
|
Use the given http proxy for all requests. If not specified, then the |
186 |
|
|
default proxy (as specified by C<$ENV{http_proxy}>) is used. |
187 |
|
|
|
188 |
root |
1.47 |
C<$scheme> must be either missing, C<http> for HTTP or C<https> for |
189 |
root |
1.2 |
HTTPS. |
190 |
root |
1.1 |
|
191 |
root |
1.3 |
=item body => $string |
192 |
|
|
|
193 |
root |
1.68 |
The request body, usually empty. Will be sent as-is (future versions of |
194 |
root |
1.3 |
this module might offer more options). |
195 |
|
|
|
196 |
root |
1.10 |
=item cookie_jar => $hash_ref |
197 |
|
|
|
198 |
|
|
Passing this parameter enables (simplified) cookie-processing, loosely |
199 |
|
|
based on the original netscape specification. |
200 |
|
|
|
201 |
root |
1.80 |
The C<$hash_ref> must be an (initially empty) hash reference which |
202 |
|
|
will get updated automatically. It is possible to save the cookie jar |
203 |
|
|
to persistent storage with something like JSON or Storable - see the |
204 |
|
|
C<AnyEvent::HTTP::cookie_jar_expire> function if you wish to remove |
205 |
|
|
expired or session-only cookies, and also for documentation on the format |
206 |
|
|
of the cookie jar. |
207 |
root |
1.10 |
|
208 |
root |
1.70 |
Note that this cookie implementation is not meant to be complete. If |
209 |
|
|
you want complete cookie management you have to do that on your |
210 |
root |
1.80 |
own. C<cookie_jar> is meant as a quick fix to get most cookie-using sites |
211 |
root |
1.70 |
working. Cookies are a privacy disaster, do not use them unless required |
212 |
|
|
to. |
213 |
root |
1.10 |
|
214 |
root |
1.69 |
When cookie processing is enabled, the C<Cookie:> and C<Set-Cookie:> |
215 |
root |
1.70 |
headers will be set and handled by this module, otherwise they will be |
216 |
root |
1.69 |
left untouched. |
217 |
|
|
|
218 |
root |
1.40 |
=item tls_ctx => $scheme | $tls_ctx |
219 |
|
|
|
220 |
|
|
Specifies the AnyEvent::TLS context to be used for https connections. This |
221 |
|
|
parameter follows the same rules as the C<tls_ctx> parameter to |
222 |
|
|
L<AnyEvent::Handle>, but additionally, the two strings C<low> or |
223 |
|
|
C<high> can be specified, which give you a predefined low-security (no |
224 |
|
|
verification, highest compatibility) and high-security (CA and common-name |
225 |
|
|
verification) TLS context. |
226 |
|
|
|
227 |
|
|
The default for this option is C<low>, which could be interpreted as "give |
228 |
|
|
me the page, no matter what". |
229 |
|
|
|
230 |
root |
1.51 |
=item on_prepare => $callback->($fh) |
231 |
|
|
|
232 |
|
|
In rare cases you need to "tune" the socket before it is used to |
233 |
|
|
connect (for exmaple, to bind it on a given IP address). This parameter |
234 |
|
|
overrides the prepare callback passed to C<AnyEvent::Socket::tcp_connect> |
235 |
|
|
and behaves exactly the same way (e.g. it has to provide a |
236 |
|
|
timeout). See the description for the C<$prepare_cb> argument of |
237 |
|
|
C<AnyEvent::Socket::tcp_connect> for details. |
238 |
|
|
|
239 |
root |
1.59 |
=item tcp_connect => $callback->($host, $service, $connect_cb, $prepare_cb) |
240 |
|
|
|
241 |
|
|
In even rarer cases you want total control over how AnyEvent::HTTP |
242 |
|
|
establishes connections. Normally it uses L<AnyEvent::Socket::tcp_connect> |
243 |
|
|
to do this, but you can provide your own C<tcp_connect> function - |
244 |
root |
1.60 |
obviously, it has to follow the same calling conventions, except that it |
245 |
|
|
may always return a connection guard object. |
246 |
root |
1.59 |
|
247 |
|
|
There are probably lots of weird uses for this function, starting from |
248 |
|
|
tracing the hosts C<http_request> actually tries to connect, to (inexact |
249 |
|
|
but fast) host => IP address caching or even socks protocol support. |
250 |
|
|
|
251 |
root |
1.42 |
=item on_header => $callback->($headers) |
252 |
root |
1.41 |
|
253 |
|
|
When specified, this callback will be called with the header hash as soon |
254 |
|
|
as headers have been successfully received from the remote server (not on |
255 |
|
|
locally-generated errors). |
256 |
|
|
|
257 |
|
|
It has to return either true (in which case AnyEvent::HTTP will continue), |
258 |
|
|
or false, in which case AnyEvent::HTTP will cancel the download (and call |
259 |
|
|
the finish callback with an error code of C<598>). |
260 |
|
|
|
261 |
|
|
This callback is useful, among other things, to quickly reject unwanted |
262 |
|
|
content, which, if it is supposed to be rare, can be faster than first |
263 |
|
|
doing a C<HEAD> request. |
264 |
|
|
|
265 |
root |
1.68 |
The downside is that cancelling the request makes it impossible to re-use |
266 |
|
|
the connection. Also, the C<on_header> callback will not receive any |
267 |
|
|
trailer (headers sent after the response body). |
268 |
|
|
|
269 |
root |
1.42 |
Example: cancel the request unless the content-type is "text/html". |
270 |
root |
1.41 |
|
271 |
root |
1.42 |
on_header => sub { |
272 |
|
|
$_[0]{"content-type"} =~ /^text\/html\s*(?:;|$)/ |
273 |
|
|
}, |
274 |
root |
1.41 |
|
275 |
root |
1.42 |
=item on_body => $callback->($partial_body, $headers) |
276 |
root |
1.41 |
|
277 |
root |
1.42 |
When specified, all body data will be passed to this callback instead of |
278 |
|
|
to the completion callback. The completion callback will get the empty |
279 |
|
|
string instead of the body data. |
280 |
root |
1.41 |
|
281 |
root |
1.42 |
It has to return either true (in which case AnyEvent::HTTP will continue), |
282 |
|
|
or false, in which case AnyEvent::HTTP will cancel the download (and call |
283 |
|
|
the completion callback with an error code of C<598>). |
284 |
|
|
|
285 |
root |
1.68 |
The downside to cancelling the request is that it makes it impossible to |
286 |
|
|
re-use the connection. |
287 |
|
|
|
288 |
root |
1.42 |
This callback is useful when the data is too large to be held in memory |
289 |
|
|
(so the callback writes it to a file) or when only some information should |
290 |
|
|
be extracted, or when the body should be processed incrementally. |
291 |
root |
1.41 |
|
292 |
|
|
It is usually preferred over doing your own body handling via |
293 |
root |
1.45 |
C<want_body_handle>, but in case of streaming APIs, where HTTP is |
294 |
|
|
only used to create a connection, C<want_body_handle> is the better |
295 |
|
|
alternative, as it allows you to install your own event handler, reducing |
296 |
|
|
resource usage. |
297 |
root |
1.41 |
|
298 |
|
|
=item want_body_handle => $enable |
299 |
|
|
|
300 |
|
|
When enabled (default is disabled), the behaviour of AnyEvent::HTTP |
301 |
|
|
changes considerably: after parsing the headers, and instead of |
302 |
|
|
downloading the body (if any), the completion callback will be |
303 |
|
|
called. Instead of the C<$body> argument containing the body data, the |
304 |
|
|
callback will receive the L<AnyEvent::Handle> object associated with the |
305 |
|
|
connection. In error cases, C<undef> will be passed. When there is no body |
306 |
|
|
(e.g. status C<304>), the empty string will be passed. |
307 |
|
|
|
308 |
|
|
The handle object might or might not be in TLS mode, might be connected to |
309 |
|
|
a proxy, be a persistent connection etc., and configured in unspecified |
310 |
|
|
ways. The user is responsible for this handle (it will not be used by this |
311 |
|
|
module anymore). |
312 |
|
|
|
313 |
|
|
This is useful with some push-type services, where, after the initial |
314 |
|
|
headers, an interactive protocol is used (typical example would be the |
315 |
|
|
push-style twitter API which starts a JSON/XML stream). |
316 |
|
|
|
317 |
|
|
If you think you need this, first have a look at C<on_body>, to see if |
318 |
root |
1.45 |
that doesn't solve your problem in a better way. |
319 |
root |
1.41 |
|
320 |
root |
1.1 |
=back |
321 |
|
|
|
322 |
root |
1.68 |
Example: do a simple HTTP GET request for http://www.nethype.de/ and print |
323 |
|
|
the response body. |
324 |
root |
1.9 |
|
325 |
|
|
http_request GET => "http://www.nethype.de/", sub { |
326 |
|
|
my ($body, $hdr) = @_; |
327 |
|
|
print "$body\n"; |
328 |
|
|
}; |
329 |
|
|
|
330 |
root |
1.68 |
Example: do a HTTP HEAD request on https://www.google.com/, use a |
331 |
root |
1.9 |
timeout of 30 seconds. |
332 |
|
|
|
333 |
|
|
http_request |
334 |
|
|
GET => "https://www.google.com", |
335 |
|
|
timeout => 30, |
336 |
|
|
sub { |
337 |
|
|
my ($body, $hdr) = @_; |
338 |
|
|
use Data::Dumper; |
339 |
|
|
print Dumper $hdr; |
340 |
|
|
} |
341 |
|
|
; |
342 |
root |
1.1 |
|
343 |
root |
1.68 |
Example: do another simple HTTP GET request, but immediately try to |
344 |
root |
1.29 |
cancel it. |
345 |
|
|
|
346 |
|
|
my $request = http_request GET => "http://www.nethype.de/", sub { |
347 |
|
|
my ($body, $hdr) = @_; |
348 |
|
|
print "$body\n"; |
349 |
|
|
}; |
350 |
|
|
|
351 |
|
|
undef $request; |
352 |
|
|
|
353 |
root |
1.1 |
=cut |
354 |
|
|
|
355 |
root |
1.12 |
sub _slot_schedule; |
356 |
root |
1.11 |
sub _slot_schedule($) { |
357 |
|
|
my $host = shift; |
358 |
|
|
|
359 |
|
|
while ($CO_SLOT{$host}[0] < $MAX_PER_HOST) { |
360 |
|
|
if (my $cb = shift @{ $CO_SLOT{$host}[1] }) { |
361 |
root |
1.12 |
# somebody wants that slot |
362 |
root |
1.11 |
++$CO_SLOT{$host}[0]; |
363 |
root |
1.14 |
++$ACTIVE; |
364 |
root |
1.11 |
|
365 |
|
|
$cb->(AnyEvent::Util::guard { |
366 |
root |
1.14 |
--$ACTIVE; |
367 |
root |
1.11 |
--$CO_SLOT{$host}[0]; |
368 |
|
|
_slot_schedule $host; |
369 |
|
|
}); |
370 |
|
|
} else { |
371 |
|
|
# nobody wants the slot, maybe we can forget about it |
372 |
|
|
delete $CO_SLOT{$host} unless $CO_SLOT{$host}[0]; |
373 |
|
|
last; |
374 |
|
|
} |
375 |
|
|
} |
376 |
|
|
} |
377 |
|
|
|
378 |
|
|
# wait for a free slot on host, call callback |
379 |
|
|
sub _get_slot($$) { |
380 |
|
|
push @{ $CO_SLOT{$_[0]}[1] }, $_[1]; |
381 |
|
|
|
382 |
|
|
_slot_schedule $_[0]; |
383 |
|
|
} |
384 |
|
|
|
385 |
root |
1.80 |
############################################################################# |
386 |
|
|
|
387 |
|
|
# expire cookies |
388 |
|
|
sub cookie_jar_expire($;$) { |
389 |
|
|
my ($jar, $session_end) = @_; |
390 |
|
|
|
391 |
|
|
%$jar = () if $jar->{version} != 1; |
392 |
|
|
|
393 |
|
|
my $anow = AE::now; |
394 |
|
|
|
395 |
|
|
while (my ($chost, $paths) = each %$jar) { |
396 |
|
|
next unless ref $paths; |
397 |
|
|
|
398 |
|
|
while (my ($cpath, $cookies) = each %$paths) { |
399 |
|
|
while (my ($cookie, $kv) = each %$cookies) { |
400 |
|
|
if (exists $kv->{_expires}) { |
401 |
|
|
delete $cookies->{$cookie} |
402 |
|
|
if $anow > $kv->{_expires}; |
403 |
|
|
} elsif ($session_end) { |
404 |
|
|
delete $cookies->{$cookie}; |
405 |
|
|
} |
406 |
|
|
} |
407 |
|
|
|
408 |
|
|
delete $paths->{$cpath} |
409 |
|
|
unless %$cookies; |
410 |
|
|
} |
411 |
|
|
|
412 |
|
|
delete $jar->{$chost} |
413 |
|
|
unless %$paths; |
414 |
|
|
} |
415 |
|
|
} |
416 |
|
|
|
417 |
root |
1.72 |
# extract cookies from jar |
418 |
root |
1.71 |
sub cookie_jar_extract($$$$) { |
419 |
|
|
my ($jar, $uscheme, $uhost, $upath) = @_; |
420 |
|
|
|
421 |
root |
1.83 |
$uhost = lc $uhost; |
422 |
|
|
|
423 |
root |
1.71 |
%$jar = () if $jar->{version} != 1; |
424 |
|
|
|
425 |
|
|
my @cookies; |
426 |
|
|
|
427 |
|
|
while (my ($chost, $paths) = each %$jar) { |
428 |
|
|
next unless ref $paths; |
429 |
|
|
|
430 |
|
|
if ($chost =~ /^\./) { |
431 |
|
|
next unless $chost eq substr $uhost, -length $chost; |
432 |
|
|
} elsif ($chost =~ /\./) { |
433 |
|
|
next unless $chost eq $uhost; |
434 |
|
|
} else { |
435 |
|
|
next; |
436 |
|
|
} |
437 |
|
|
|
438 |
|
|
while (my ($cpath, $cookies) = each %$paths) { |
439 |
|
|
next unless $cpath eq substr $upath, 0, length $cpath; |
440 |
|
|
|
441 |
|
|
while (my ($cookie, $kv) = each %$cookies) { |
442 |
|
|
next if $uscheme ne "https" && exists $kv->{secure}; |
443 |
|
|
|
444 |
root |
1.80 |
if (exists $kv->{_expires} and AE::now > $kv->{_expires}) { |
445 |
|
|
delete $cookies->{$cookie}; |
446 |
|
|
next; |
447 |
root |
1.71 |
} |
448 |
|
|
|
449 |
|
|
my $value = $kv->{value}; |
450 |
|
|
|
451 |
|
|
if ($value =~ /[=;,[:space:]]/) { |
452 |
|
|
$value =~ s/([\\"])/\\$1/g; |
453 |
|
|
$value = "\"$value\""; |
454 |
|
|
} |
455 |
|
|
|
456 |
|
|
push @cookies, "$cookie=$value"; |
457 |
|
|
} |
458 |
|
|
} |
459 |
|
|
} |
460 |
|
|
|
461 |
|
|
\@cookies |
462 |
|
|
} |
463 |
|
|
|
464 |
root |
1.72 |
# parse set_cookie header into jar |
465 |
root |
1.80 |
sub cookie_jar_set_cookie($$$$) { |
466 |
|
|
my ($jar, $set_cookie, $uhost, $date) = @_; |
467 |
|
|
|
468 |
|
|
my $anow = int AE::now; |
469 |
|
|
my $snow; # server-now |
470 |
root |
1.72 |
|
471 |
|
|
for ($set_cookie) { |
472 |
|
|
# parse NAME=VALUE |
473 |
|
|
my @kv; |
474 |
|
|
|
475 |
root |
1.79 |
# expires is not http-compliant in the original cookie-spec, |
476 |
|
|
# we support the official date format and some extensions |
477 |
root |
1.72 |
while ( |
478 |
|
|
m{ |
479 |
|
|
\G\s* |
480 |
|
|
(?: |
481 |
root |
1.79 |
expires \s*=\s* ([A-Z][a-z][a-z]+,\ [^,;]+) |
482 |
root |
1.82 |
| ([^=;,[:space:]]+) (?: \s*=\s* (?: "((?:[^\\"]+|\\.)*)" | ([^=;,[:space:]]*) ) )? |
483 |
root |
1.72 |
) |
484 |
|
|
}gcxsi |
485 |
|
|
) { |
486 |
|
|
my $name = $2; |
487 |
|
|
my $value = $4; |
488 |
|
|
|
489 |
root |
1.82 |
if (defined $1) { |
490 |
root |
1.72 |
# expires |
491 |
|
|
$name = "expires"; |
492 |
|
|
$value = $1; |
493 |
root |
1.82 |
} elsif (defined $3) { |
494 |
root |
1.72 |
# quoted |
495 |
|
|
$value = $3; |
496 |
|
|
$value =~ s/\\(.)/$1/gs; |
497 |
|
|
} |
498 |
|
|
|
499 |
|
|
push @kv, lc $name, $value; |
500 |
|
|
|
501 |
|
|
last unless /\G\s*;/gc; |
502 |
|
|
} |
503 |
|
|
|
504 |
|
|
last unless @kv; |
505 |
|
|
|
506 |
|
|
my $name = shift @kv; |
507 |
|
|
my %kv = (value => shift @kv, @kv); |
508 |
|
|
|
509 |
root |
1.80 |
if (exists $kv{"max-age"}) { |
510 |
|
|
$kv{_expires} = $anow + delete $kv{"max-age"}; |
511 |
|
|
} elsif (exists $kv{expires}) { |
512 |
|
|
$snow ||= parse_date ($date) || $anow; |
513 |
|
|
$kv{_expires} = $anow + (parse_date (delete $kv{expires}) - $snow); |
514 |
|
|
} else { |
515 |
|
|
delete $kv{_expires}; |
516 |
|
|
} |
517 |
root |
1.72 |
|
518 |
|
|
my $cdom; |
519 |
|
|
my $cpath = (delete $kv{path}) || "/"; |
520 |
|
|
|
521 |
|
|
if (exists $kv{domain}) { |
522 |
|
|
$cdom = delete $kv{domain}; |
523 |
|
|
|
524 |
|
|
$cdom =~ s/^\.?/./; # make sure it starts with a "." |
525 |
|
|
|
526 |
|
|
next if $cdom =~ /\.$/; |
527 |
|
|
|
528 |
|
|
# this is not rfc-like and not netscape-like. go figure. |
529 |
|
|
my $ndots = $cdom =~ y/.//; |
530 |
|
|
next if $ndots < ($cdom =~ /\.[^.][^.]\.[^.][^.]$/ ? 3 : 2); |
531 |
|
|
} else { |
532 |
|
|
$cdom = $uhost; |
533 |
|
|
} |
534 |
|
|
|
535 |
|
|
# store it |
536 |
root |
1.73 |
$jar->{version} = 1; |
537 |
root |
1.83 |
$jar->{lc $cdom}{$cpath}{$name} = \%kv; |
538 |
root |
1.72 |
|
539 |
|
|
redo if /\G\s*,/gc; |
540 |
|
|
} |
541 |
|
|
} |
542 |
|
|
|
543 |
root |
1.66 |
# continue to parse $_ for headers and place them into the arg |
544 |
|
|
sub parse_hdr() { |
545 |
|
|
my %hdr; |
546 |
|
|
|
547 |
|
|
# things seen, not parsed: |
548 |
|
|
# p3pP="NON CUR OTPi OUR NOR UNI" |
549 |
|
|
|
550 |
|
|
$hdr{lc $1} .= ",$2" |
551 |
|
|
while /\G |
552 |
|
|
([^:\000-\037]*): |
553 |
|
|
[\011\040]* |
554 |
|
|
((?: [^\012]+ | \012[\011\040] )*) |
555 |
|
|
\012 |
556 |
|
|
/gxc; |
557 |
|
|
|
558 |
|
|
/\G$/ |
559 |
|
|
or return; |
560 |
|
|
|
561 |
|
|
# remove the "," prefix we added to all headers above |
562 |
|
|
substr $_, 0, 1, "" |
563 |
|
|
for values %hdr; |
564 |
|
|
|
565 |
|
|
\%hdr |
566 |
|
|
} |
567 |
|
|
|
568 |
root |
1.46 |
our $qr_nlnl = qr{(?<![^\012])\015?\012}; |
569 |
root |
1.34 |
|
570 |
root |
1.41 |
our $TLS_CTX_LOW = { cache => 1, sslv2 => 1 }; |
571 |
|
|
our $TLS_CTX_HIGH = { cache => 1, verify => 1, verify_peername => "https" }; |
572 |
root |
1.40 |
|
573 |
elmex |
1.15 |
sub http_request($$@) { |
574 |
root |
1.1 |
my $cb = pop; |
575 |
|
|
my ($method, $url, %arg) = @_; |
576 |
|
|
|
577 |
|
|
my %hdr; |
578 |
|
|
|
579 |
root |
1.40 |
$arg{tls_ctx} = $TLS_CTX_LOW if $arg{tls_ctx} eq "low" || !exists $arg{tls_ctx}; |
580 |
|
|
$arg{tls_ctx} = $TLS_CTX_HIGH if $arg{tls_ctx} eq "high"; |
581 |
|
|
|
582 |
root |
1.3 |
$method = uc $method; |
583 |
|
|
|
584 |
root |
1.8 |
if (my $hdr = $arg{headers}) { |
585 |
root |
1.1 |
while (my ($k, $v) = each %$hdr) { |
586 |
|
|
$hdr{lc $k} = $v; |
587 |
|
|
} |
588 |
|
|
} |
589 |
|
|
|
590 |
root |
1.55 |
# pseudo headers for all subsequent responses |
591 |
|
|
my @pseudo = (URL => $url); |
592 |
|
|
push @pseudo, Redirect => delete $arg{Redirect} if exists $arg{Redirect}; |
593 |
|
|
|
594 |
root |
1.23 |
my $recurse = exists $arg{recurse} ? delete $arg{recurse} : $MAX_RECURSE; |
595 |
root |
1.8 |
|
596 |
root |
1.64 |
return $cb->(undef, { @pseudo, Status => 599, Reason => "Too many redirections" }) |
597 |
root |
1.8 |
if $recurse < 0; |
598 |
|
|
|
599 |
root |
1.2 |
my $proxy = $arg{proxy} || $PROXY; |
600 |
root |
1.1 |
my $timeout = $arg{timeout} || $TIMEOUT; |
601 |
|
|
|
602 |
root |
1.31 |
my ($uscheme, $uauthority, $upath, $query, $fragment) = |
603 |
root |
1.56 |
$url =~ m|(?:([^:/?#]+):)?(?://([^/?#]*))?([^?#]*)(?:(\?[^#]*))?(?:#(.*))?|; |
604 |
root |
1.2 |
|
605 |
root |
1.31 |
$uscheme = lc $uscheme; |
606 |
root |
1.1 |
|
607 |
root |
1.31 |
my $uport = $uscheme eq "http" ? 80 |
608 |
|
|
: $uscheme eq "https" ? 443 |
609 |
root |
1.64 |
: return $cb->(undef, { @pseudo, Status => 599, Reason => "Only http and https URL schemes supported" }); |
610 |
root |
1.13 |
|
611 |
root |
1.31 |
$uauthority =~ /^(?: .*\@ )? ([^\@:]+) (?: : (\d+) )?$/x |
612 |
root |
1.64 |
or return $cb->(undef, { @pseudo, Status => 599, Reason => "Unparsable URL" }); |
613 |
root |
1.10 |
|
614 |
|
|
my $uhost = $1; |
615 |
|
|
$uport = $2 if defined $2; |
616 |
|
|
|
617 |
root |
1.53 |
$hdr{host} = defined $2 ? "$uhost:$2" : "$uhost" |
618 |
|
|
unless exists $hdr{host}; |
619 |
root |
1.43 |
|
620 |
root |
1.10 |
$uhost =~ s/^\[(.*)\]$/$1/; |
621 |
root |
1.56 |
$upath .= $query if length $query; |
622 |
root |
1.10 |
|
623 |
|
|
$upath =~ s%^/?%/%; |
624 |
|
|
|
625 |
|
|
# cookie processing |
626 |
|
|
if (my $jar = $arg{cookie_jar}) { |
627 |
root |
1.71 |
my $cookies = cookie_jar_extract $jar, $uscheme, $uhost, $upath; |
628 |
root |
1.70 |
|
629 |
root |
1.71 |
$hdr{cookie} = join "; ", @$cookies |
630 |
|
|
if @$cookies; |
631 |
root |
1.10 |
} |
632 |
root |
1.1 |
|
633 |
root |
1.31 |
my ($rhost, $rport, $rscheme, $rpath); # request host, port, path |
634 |
root |
1.2 |
|
635 |
root |
1.10 |
if ($proxy) { |
636 |
root |
1.38 |
($rpath, $rhost, $rport, $rscheme) = ($url, @$proxy); |
637 |
root |
1.31 |
|
638 |
root |
1.47 |
$rscheme = "http" unless defined $rscheme; |
639 |
|
|
|
640 |
root |
1.31 |
# don't support https requests over https-proxy transport, |
641 |
root |
1.38 |
# can't be done with tls as spec'ed, unless you double-encrypt. |
642 |
root |
1.31 |
$rscheme = "http" if $uscheme eq "https" && $rscheme eq "https"; |
643 |
root |
1.10 |
} else { |
644 |
root |
1.31 |
($rhost, $rport, $rscheme, $rpath) = ($uhost, $uport, $uscheme, $upath); |
645 |
root |
1.2 |
} |
646 |
|
|
|
647 |
root |
1.47 |
# leave out fragment and query string, just a heuristic |
648 |
root |
1.66 |
$hdr{referer} = "$uscheme://$uauthority$upath" unless exists $hdr{referer}; |
649 |
|
|
$hdr{"user-agent"} = $USERAGENT unless exists $hdr{"user-agent"}; |
650 |
root |
1.41 |
|
651 |
root |
1.53 |
$hdr{"content-length"} = length $arg{body} |
652 |
|
|
if length $arg{body} || $method ne "GET"; |
653 |
root |
1.1 |
|
654 |
root |
1.68 |
$hdr{connection} = "close TE"; #1.1 |
655 |
|
|
$hdr{te} = "trailers" unless exists $hdr{te}; #1.1 |
656 |
root |
1.66 |
|
657 |
root |
1.11 |
my %state = (connect_guard => 1); |
658 |
|
|
|
659 |
|
|
_get_slot $uhost, sub { |
660 |
|
|
$state{slot_guard} = shift; |
661 |
root |
1.1 |
|
662 |
root |
1.11 |
return unless $state{connect_guard}; |
663 |
root |
1.1 |
|
664 |
root |
1.77 |
my $ae_error = 595; # connecting |
665 |
|
|
|
666 |
root |
1.82 |
# handle actual, non-tunneled, request |
667 |
|
|
my $handle_actual_request = sub { |
668 |
|
|
$ae_error = 596; # request phase |
669 |
|
|
|
670 |
|
|
$state{handle}->starttls ("connect") if $uscheme eq "https" && !exists $state{handle}{tls}; |
671 |
|
|
|
672 |
|
|
# send request |
673 |
|
|
$state{handle}->push_write ( |
674 |
|
|
"$method $rpath HTTP/1.1\015\012" |
675 |
|
|
. (join "", map "\u$_: $hdr{$_}\015\012", grep defined $hdr{$_}, keys %hdr) |
676 |
|
|
. "\015\012" |
677 |
|
|
. (delete $arg{body}) |
678 |
|
|
); |
679 |
|
|
|
680 |
|
|
# return if error occured during push_write() |
681 |
|
|
return unless %state; |
682 |
|
|
|
683 |
|
|
%hdr = (); # reduce memory usage, save a kitten, also make it possible to re-use |
684 |
|
|
|
685 |
|
|
# status line and headers |
686 |
|
|
$state{read_response} = sub { |
687 |
|
|
for ("$_[1]") { |
688 |
|
|
y/\015//d; # weed out any \015, as they show up in the weirdest of places. |
689 |
|
|
|
690 |
|
|
/^HTTP\/0*([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\012]*) )? \012/gxci |
691 |
|
|
or return (%state = (), $cb->(undef, { @pseudo, Status => 599, Reason => "Invalid server response" })); |
692 |
|
|
|
693 |
|
|
# 100 Continue handling |
694 |
|
|
# should not happen as we don't send expect: 100-continue, |
695 |
|
|
# but we handle it just in case. |
696 |
|
|
# since we send the request body regardless, if we get an error |
697 |
|
|
# we are out of-sync, which we currently do NOT handle correctly. |
698 |
|
|
return $state{handle}->push_read (line => $qr_nlnl, $state{read_response}) |
699 |
|
|
if $2 eq 100; |
700 |
|
|
|
701 |
|
|
push @pseudo, |
702 |
|
|
HTTPVersion => $1, |
703 |
|
|
Status => $2, |
704 |
|
|
Reason => $3, |
705 |
|
|
; |
706 |
|
|
|
707 |
|
|
my $hdr = parse_hdr |
708 |
|
|
or return (%state = (), $cb->(undef, { @pseudo, Status => 599, Reason => "Garbled response headers" })); |
709 |
|
|
|
710 |
|
|
%hdr = (%$hdr, @pseudo); |
711 |
|
|
} |
712 |
|
|
|
713 |
|
|
# redirect handling |
714 |
|
|
# microsoft and other shitheads don't give a shit for following standards, |
715 |
|
|
# try to support some common forms of broken Location headers. |
716 |
|
|
if ($hdr{location} !~ /^(?: $ | [^:\/?\#]+ : )/x) { |
717 |
|
|
$hdr{location} =~ s/^\.\/+//; |
718 |
|
|
|
719 |
|
|
my $url = "$rscheme://$uhost:$uport"; |
720 |
|
|
|
721 |
|
|
unless ($hdr{location} =~ s/^\///) { |
722 |
|
|
$url .= $upath; |
723 |
|
|
$url =~ s/\/[^\/]*$//; |
724 |
|
|
} |
725 |
|
|
|
726 |
|
|
$hdr{location} = "$url/$hdr{location}"; |
727 |
|
|
} |
728 |
|
|
|
729 |
|
|
my $redirect; |
730 |
|
|
|
731 |
|
|
if ($recurse) { |
732 |
|
|
my $status = $hdr{Status}; |
733 |
|
|
|
734 |
|
|
# industry standard is to redirect POST as GET for |
735 |
|
|
# 301, 302 and 303, in contrast to HTTP/1.0 and 1.1. |
736 |
|
|
# also, the UA should ask the user for 301 and 307 and POST, |
737 |
|
|
# industry standard seems to be to simply follow. |
738 |
|
|
# we go with the industry standard. |
739 |
|
|
if ($status == 301 or $status == 302 or $status == 303) { |
740 |
|
|
# HTTP/1.1 is unclear on how to mutate the method |
741 |
|
|
$method = "GET" unless $method eq "HEAD"; |
742 |
|
|
$redirect = 1; |
743 |
|
|
} elsif ($status == 307) { |
744 |
|
|
$redirect = 1; |
745 |
|
|
} |
746 |
|
|
} |
747 |
|
|
|
748 |
|
|
my $finish = sub { # ($data, $err_status, $err_reason[, $keepalive]) |
749 |
|
|
my $may_keep_alive = $_[3]; |
750 |
|
|
|
751 |
|
|
$state{handle}->destroy if $state{handle}; |
752 |
|
|
%state = (); |
753 |
|
|
|
754 |
|
|
if (defined $_[1]) { |
755 |
|
|
$hdr{OrigStatus} = $hdr{Status}; $hdr{Status} = $_[1]; |
756 |
|
|
$hdr{OrigReason} = $hdr{Reason}; $hdr{Reason} = $_[2]; |
757 |
|
|
} |
758 |
|
|
|
759 |
|
|
# set-cookie processing |
760 |
|
|
if ($arg{cookie_jar}) { |
761 |
|
|
cookie_jar_set_cookie $arg{cookie_jar}, $hdr{"set-cookie"}, $uhost, $hdr{date}; |
762 |
|
|
} |
763 |
|
|
|
764 |
|
|
if ($redirect && exists $hdr{location}) { |
765 |
|
|
# we ignore any errors, as it is very common to receive |
766 |
|
|
# Content-Length != 0 but no actual body |
767 |
|
|
# we also access %hdr, as $_[1] might be an erro |
768 |
|
|
http_request ( |
769 |
|
|
$method => $hdr{location}, |
770 |
|
|
%arg, |
771 |
|
|
recurse => $recurse - 1, |
772 |
|
|
Redirect => [$_[0], \%hdr], |
773 |
|
|
$cb); |
774 |
|
|
} else { |
775 |
|
|
$cb->($_[0], \%hdr); |
776 |
|
|
} |
777 |
|
|
}; |
778 |
|
|
|
779 |
|
|
$ae_error = 597; # body phase |
780 |
|
|
|
781 |
|
|
my $len = $hdr{"content-length"}; |
782 |
root |
1.84 |
warn "no content $redirect x<$len>$hdr{Status}\n";#d# |
783 |
root |
1.82 |
|
784 |
root |
1.84 |
# body handling, many different code paths |
785 |
|
|
# - no body expected |
786 |
|
|
# - want_body_handle |
787 |
|
|
# - te chunked |
788 |
|
|
# - 2x length known (with or without on_body) |
789 |
|
|
# - 2x length not known (with or without on_body) |
790 |
root |
1.82 |
if (!$redirect && $arg{on_header} && !$arg{on_header}(\%hdr)) { |
791 |
|
|
$finish->(undef, 598 => "Request cancelled by on_header"); |
792 |
|
|
} elsif ( |
793 |
|
|
$hdr{Status} =~ /^(?:1..|204|205|304)$/ |
794 |
|
|
or $method eq "HEAD" |
795 |
root |
1.84 |
or (defined $len && $len == 0) # == 0, not !, because "0 " is true |
796 |
root |
1.82 |
) { |
797 |
|
|
# no body |
798 |
|
|
$finish->("", undef, undef, 1); |
799 |
|
|
|
800 |
root |
1.84 |
} elsif (!$redirect && $arg{want_body_handle}) { |
801 |
|
|
$_[0]->on_eof (undef); |
802 |
|
|
$_[0]->on_error (undef); |
803 |
|
|
$_[0]->on_read (undef); |
804 |
|
|
|
805 |
|
|
$finish->(delete $state{handle}); |
806 |
|
|
|
807 |
|
|
} elsif ($hdr{"transfer-encoding"} =~ /\bchunked\b/i) { |
808 |
|
|
my $cl = 0; |
809 |
|
|
my $body = undef; |
810 |
|
|
my $on_body = $arg{on_body} || sub { $body .= shift; 1 }; |
811 |
|
|
|
812 |
|
|
$state{read_chunk} = sub { |
813 |
|
|
$_[1] =~ /^([0-9a-fA-F]+)/ |
814 |
|
|
or $finish->(undef, $ae_error => "Garbled chunked transfer encoding"); |
815 |
|
|
|
816 |
|
|
my $len = hex $1; |
817 |
root |
1.82 |
|
818 |
root |
1.84 |
if ($len) { |
819 |
|
|
$cl += $len; |
820 |
root |
1.82 |
|
821 |
root |
1.84 |
$_[0]->push_read (chunk => $len, sub { |
822 |
|
|
$on_body->($_[1], \%hdr) |
823 |
|
|
or return $finish->(undef, 598 => "Request cancelled by on_body"); |
824 |
root |
1.82 |
|
825 |
root |
1.84 |
$_[0]->push_read (line => sub { |
826 |
|
|
length $_[1] |
827 |
|
|
and return $finish->(undef, $ae_error => "Garbled chunked transfer encoding"); |
828 |
|
|
$_[0]->push_read (line => $state{read_chunk}); |
829 |
root |
1.82 |
}); |
830 |
root |
1.84 |
}); |
831 |
|
|
} else { |
832 |
|
|
$hdr{"content-length"} ||= $cl; |
833 |
root |
1.82 |
|
834 |
root |
1.84 |
$_[0]->push_read (line => $qr_nlnl, sub { |
835 |
|
|
if (length $_[1]) { |
836 |
|
|
for ("$_[1]") { |
837 |
|
|
y/\015//d; # weed out any \015, as they show up in the weirdest of places. |
838 |
root |
1.82 |
|
839 |
root |
1.84 |
my $hdr = parse_hdr |
840 |
|
|
or return $finish->(undef, $ae_error => "Garbled response trailers"); |
841 |
root |
1.82 |
|
842 |
root |
1.84 |
%hdr = (%hdr, %$hdr); |
843 |
|
|
} |
844 |
|
|
} |
845 |
root |
1.82 |
|
846 |
root |
1.84 |
$finish->($body, undef, undef, 1); |
847 |
root |
1.82 |
}); |
848 |
|
|
} |
849 |
root |
1.84 |
}; |
850 |
|
|
|
851 |
|
|
$_[0]->push_read (line => $state{read_chunk}); |
852 |
|
|
|
853 |
|
|
} elsif ($arg{on_body}) { |
854 |
|
|
if (defined $len) { |
855 |
|
|
$_[0]->on_read (sub { |
856 |
|
|
$len -= length $_[0]{rbuf}; |
857 |
|
|
|
858 |
|
|
$arg{on_body}(delete $_[0]{rbuf}, \%hdr) |
859 |
|
|
or return $finish->(undef, 598 => "Request cancelled by on_body"); |
860 |
|
|
|
861 |
|
|
$len > 0 |
862 |
|
|
or $finish->("", undef, undef, 1); |
863 |
|
|
}); |
864 |
root |
1.82 |
} else { |
865 |
root |
1.84 |
$_[0]->on_eof (sub { |
866 |
|
|
$finish->(""); |
867 |
|
|
}); |
868 |
|
|
$_[0]->on_read (sub { |
869 |
|
|
$arg{on_body}(delete $_[0]{rbuf}, \%hdr) |
870 |
|
|
or $finish->(undef, 598 => "Request cancelled by on_body"); |
871 |
|
|
}); |
872 |
|
|
} |
873 |
|
|
} else { |
874 |
|
|
$_[0]->on_eof (undef); |
875 |
root |
1.82 |
|
876 |
root |
1.84 |
if (defined $len) { |
877 |
|
|
$_[0]->on_read (sub { |
878 |
|
|
$finish->((substr delete $_[0]{rbuf}, 0, $len, ""), undef, undef, 1) |
879 |
|
|
if $len <= length $_[0]{rbuf}; |
880 |
|
|
}); |
881 |
|
|
} else { |
882 |
|
|
$_[0]->on_error (sub { |
883 |
|
|
($! == Errno::EPIPE || !$!) |
884 |
|
|
? $finish->(delete $_[0]{rbuf}) |
885 |
|
|
: $finish->(undef, $ae_error => $_[2]); |
886 |
|
|
}); |
887 |
|
|
$_[0]->on_read (sub { }); |
888 |
root |
1.82 |
} |
889 |
|
|
} |
890 |
|
|
}; |
891 |
|
|
|
892 |
|
|
$state{handle}->push_read (line => $qr_nlnl, $state{read_response}); |
893 |
|
|
}; |
894 |
|
|
|
895 |
root |
1.64 |
my $connect_cb = sub { |
896 |
|
|
$state{fh} = shift |
897 |
|
|
or do { |
898 |
|
|
my $err = "$!"; |
899 |
|
|
%state = (); |
900 |
root |
1.77 |
return $cb->(undef, { @pseudo, Status => $ae_error, Reason => $err }); |
901 |
root |
1.64 |
}; |
902 |
root |
1.44 |
|
903 |
root |
1.64 |
return unless delete $state{connect_guard}; |
904 |
root |
1.11 |
|
905 |
root |
1.64 |
# get handle |
906 |
|
|
$state{handle} = new AnyEvent::Handle |
907 |
|
|
fh => $state{fh}, |
908 |
|
|
peername => $rhost, |
909 |
|
|
tls_ctx => $arg{tls_ctx}, |
910 |
|
|
# these need to be reconfigured on keepalive handles |
911 |
|
|
timeout => $timeout, |
912 |
|
|
on_error => sub { |
913 |
|
|
%state = (); |
914 |
root |
1.77 |
$cb->(undef, { @pseudo, Status => $ae_error, Reason => $_[2] }); |
915 |
root |
1.64 |
}, |
916 |
|
|
on_eof => sub { |
917 |
|
|
%state = (); |
918 |
root |
1.77 |
$cb->(undef, { @pseudo, Status => $ae_error, Reason => "Unexpected end-of-file" }); |
919 |
root |
1.64 |
}, |
920 |
|
|
; |
921 |
root |
1.11 |
|
922 |
root |
1.64 |
# limit the number of persistent connections |
923 |
|
|
# keepalive not yet supported |
924 |
root |
1.56 |
# if ($KA_COUNT{$_[1]} < $MAX_PERSISTENT_PER_HOST) { |
925 |
|
|
# ++$KA_COUNT{$_[1]}; |
926 |
|
|
# $state{handle}{ka_count_guard} = AnyEvent::Util::guard { |
927 |
|
|
# --$KA_COUNT{$_[1]} |
928 |
|
|
# }; |
929 |
|
|
# $hdr{connection} = "keep-alive"; |
930 |
|
|
# } |
931 |
root |
1.1 |
|
932 |
root |
1.64 |
$state{handle}->starttls ("connect") if $rscheme eq "https"; |
933 |
|
|
|
934 |
|
|
# now handle proxy-CONNECT method |
935 |
|
|
if ($proxy && $uscheme eq "https") { |
936 |
|
|
# oh dear, we have to wrap it into a connect request |
937 |
|
|
|
938 |
|
|
# maybe re-use $uauthority with patched port? |
939 |
|
|
$state{handle}->push_write ("CONNECT $uhost:$uport HTTP/1.0\015\012Host: $uhost\015\012\015\012"); |
940 |
|
|
$state{handle}->push_read (line => $qr_nlnl, sub { |
941 |
|
|
$_[1] =~ /^HTTP\/([0-9\.]+) \s+ ([0-9]{3}) (?: \s+ ([^\015\012]*) )?/ix |
942 |
|
|
or return (%state = (), $cb->(undef, { @pseudo, Status => 599, Reason => "Invalid proxy connect response ($_[1])" })); |
943 |
|
|
|
944 |
|
|
if ($2 == 200) { |
945 |
|
|
$rpath = $upath; |
946 |
root |
1.81 |
$handle_actual_request->(); |
947 |
root |
1.64 |
} else { |
948 |
|
|
%state = (); |
949 |
|
|
$cb->(undef, { @pseudo, Status => $2, Reason => $3 }); |
950 |
|
|
} |
951 |
|
|
}); |
952 |
|
|
} else { |
953 |
root |
1.81 |
$handle_actual_request->(); |
954 |
root |
1.64 |
} |
955 |
|
|
}; |
956 |
|
|
|
957 |
|
|
my $tcp_connect = $arg{tcp_connect} |
958 |
|
|
|| do { require AnyEvent::Socket; \&AnyEvent::Socket::tcp_connect }; |
959 |
root |
1.57 |
|
960 |
root |
1.64 |
$state{connect_guard} = $tcp_connect->($rhost, $rport, $connect_cb, $arg{on_prepare} || sub { $timeout }); |
961 |
root |
1.1 |
}; |
962 |
|
|
|
963 |
|
|
defined wantarray && AnyEvent::Util::guard { %state = () } |
964 |
|
|
} |
965 |
|
|
|
966 |
elmex |
1.15 |
sub http_get($@) { |
967 |
root |
1.1 |
unshift @_, "GET"; |
968 |
|
|
&http_request |
969 |
|
|
} |
970 |
|
|
|
971 |
elmex |
1.15 |
sub http_head($@) { |
972 |
root |
1.4 |
unshift @_, "HEAD"; |
973 |
|
|
&http_request |
974 |
|
|
} |
975 |
|
|
|
976 |
elmex |
1.15 |
sub http_post($$@) { |
977 |
root |
1.22 |
my $url = shift; |
978 |
|
|
unshift @_, "POST", $url, "body"; |
979 |
root |
1.3 |
&http_request |
980 |
|
|
} |
981 |
|
|
|
982 |
root |
1.9 |
=back |
983 |
|
|
|
984 |
root |
1.55 |
=head2 DNS CACHING |
985 |
|
|
|
986 |
|
|
AnyEvent::HTTP uses the AnyEvent::Socket::tcp_connect function for |
987 |
|
|
the actual connection, which in turn uses AnyEvent::DNS to resolve |
988 |
|
|
hostnames. The latter is a simple stub resolver and does no caching |
989 |
|
|
on its own. If you want DNS caching, you currently have to provide |
990 |
|
|
your own default resolver (by storing a suitable resolver object in |
991 |
|
|
C<$AnyEvent::DNS::RESOLVER>). |
992 |
|
|
|
993 |
root |
1.2 |
=head2 GLOBAL FUNCTIONS AND VARIABLES |
994 |
root |
1.1 |
|
995 |
|
|
=over 4 |
996 |
|
|
|
997 |
root |
1.2 |
=item AnyEvent::HTTP::set_proxy "proxy-url" |
998 |
|
|
|
999 |
|
|
Sets the default proxy server to use. The proxy-url must begin with a |
1000 |
root |
1.52 |
string of the form C<http://host:port> (optionally C<https:...>), croaks |
1001 |
|
|
otherwise. |
1002 |
|
|
|
1003 |
|
|
To clear an already-set proxy, use C<undef>. |
1004 |
root |
1.2 |
|
1005 |
root |
1.80 |
=item AnyEvent::HTTP::cookie_jar_expire $jar[, $session_end] |
1006 |
|
|
|
1007 |
|
|
Remove all cookies from the cookie jar that have been expired. If |
1008 |
|
|
C<$session_end> is given and true, then additionally remove all session |
1009 |
|
|
cookies. |
1010 |
|
|
|
1011 |
|
|
You should call this function (with a true C<$session_end>) before you |
1012 |
|
|
save cookies to disk, and you should call this function after loading them |
1013 |
|
|
again. If you have a long-running program you can additonally call this |
1014 |
|
|
function from time to time. |
1015 |
|
|
|
1016 |
|
|
A cookie jar is initially an empty hash-reference that is managed by this |
1017 |
|
|
module. It's format is subject to change, but currently it is like this: |
1018 |
|
|
|
1019 |
|
|
The key C<version> has to contain C<1>, otherwise the hash gets |
1020 |
|
|
emptied. All other keys are hostnames or IP addresses pointing to |
1021 |
|
|
hash-references. The key for these inner hash references is the |
1022 |
|
|
server path for which this cookie is meant, and the values are again |
1023 |
|
|
hash-references. The keys of those hash-references is the cookie name, and |
1024 |
|
|
the value, you guessed it, is another hash-reference, this time with the |
1025 |
|
|
key-value pairs from the cookie, except for C<expires> and C<max-age>, |
1026 |
|
|
which have been replaced by a C<_expires> key that contains the cookie |
1027 |
|
|
expiry timestamp. |
1028 |
|
|
|
1029 |
|
|
Here is an example of a cookie jar with a single cookie, so you have a |
1030 |
|
|
chance of understanding the above paragraph: |
1031 |
|
|
|
1032 |
|
|
{ |
1033 |
|
|
version => 1, |
1034 |
|
|
"10.0.0.1" => { |
1035 |
|
|
"/" => { |
1036 |
|
|
"mythweb_id" => { |
1037 |
|
|
_expires => 1293917923, |
1038 |
|
|
value => "ooRung9dThee3ooyXooM1Ohm", |
1039 |
|
|
}, |
1040 |
|
|
}, |
1041 |
|
|
}, |
1042 |
|
|
} |
1043 |
|
|
|
1044 |
root |
1.61 |
=item $date = AnyEvent::HTTP::format_date $timestamp |
1045 |
|
|
|
1046 |
|
|
Takes a POSIX timestamp (seconds since the epoch) and formats it as a HTTP |
1047 |
|
|
Date (RFC 2616). |
1048 |
|
|
|
1049 |
|
|
=item $timestamp = AnyEvent::HTTP::parse_date $date |
1050 |
|
|
|
1051 |
root |
1.79 |
Takes a HTTP Date (RFC 2616) or a Cookie date (netscape cookie spec) or a |
1052 |
|
|
bunch of minor variations of those, and returns the corresponding POSIX |
1053 |
|
|
timestamp, or C<undef> if the date cannot be parsed. |
1054 |
root |
1.61 |
|
1055 |
root |
1.3 |
=item $AnyEvent::HTTP::MAX_RECURSE |
1056 |
root |
1.1 |
|
1057 |
root |
1.3 |
The default value for the C<recurse> request parameter (default: C<10>). |
1058 |
root |
1.1 |
|
1059 |
|
|
=item $AnyEvent::HTTP::USERAGENT |
1060 |
|
|
|
1061 |
|
|
The default value for the C<User-Agent> header (the default is |
1062 |
root |
1.40 |
C<Mozilla/5.0 (compatible; U; AnyEvent-HTTP/$VERSION; +http://software.schmorp.de/pkg/AnyEvent)>). |
1063 |
root |
1.1 |
|
1064 |
root |
1.43 |
=item $AnyEvent::HTTP::MAX_PER_HOST |
1065 |
root |
1.1 |
|
1066 |
root |
1.47 |
The maximum number of concurrent connections to the same host (identified |
1067 |
root |
1.43 |
by the hostname). If the limit is exceeded, then the additional requests |
1068 |
|
|
are queued until previous connections are closed. |
1069 |
root |
1.1 |
|
1070 |
root |
1.43 |
The default value for this is C<4>, and it is highly advisable to not |
1071 |
|
|
increase it. |
1072 |
root |
1.3 |
|
1073 |
root |
1.14 |
=item $AnyEvent::HTTP::ACTIVE |
1074 |
|
|
|
1075 |
|
|
The number of active connections. This is not the number of currently |
1076 |
|
|
running requests, but the number of currently open and non-idle TCP |
1077 |
|
|
connections. This number of can be useful for load-leveling. |
1078 |
|
|
|
1079 |
root |
1.1 |
=back |
1080 |
|
|
|
1081 |
|
|
=cut |
1082 |
|
|
|
1083 |
root |
1.61 |
our @month = qw(Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec); |
1084 |
|
|
our @weekday = qw(Sun Mon Tue Wed Thu Fri Sat); |
1085 |
|
|
|
1086 |
|
|
sub format_date($) { |
1087 |
|
|
my ($time) = @_; |
1088 |
|
|
|
1089 |
|
|
# RFC 822/1123 format |
1090 |
|
|
my ($S, $M, $H, $mday, $mon, $year, $wday, $yday, undef) = gmtime $time; |
1091 |
|
|
|
1092 |
|
|
sprintf "%s, %02d %s %04d %02d:%02d:%02d GMT", |
1093 |
|
|
$weekday[$wday], $mday, $month[$mon], $year + 1900, |
1094 |
|
|
$H, $M, $S; |
1095 |
|
|
} |
1096 |
|
|
|
1097 |
|
|
sub parse_date($) { |
1098 |
|
|
my ($date) = @_; |
1099 |
|
|
|
1100 |
|
|
my ($d, $m, $y, $H, $M, $S); |
1101 |
|
|
|
1102 |
root |
1.79 |
if ($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$/) { |
1103 |
root |
1.70 |
# RFC 822/1123, required by RFC 2616 (with " ") |
1104 |
|
|
# cookie dates (with "-") |
1105 |
|
|
|
1106 |
root |
1.61 |
($d, $m, $y, $H, $M, $S) = ($1, $2, $3, $4, $5, $6); |
1107 |
|
|
|
1108 |
root |
1.79 |
} elsif ($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$/) { |
1109 |
root |
1.61 |
# RFC 850 |
1110 |
|
|
($d, $m, $y, $H, $M, $S) = ($1, $2, $3 < 69 ? $3 + 2000 : $3 + 1900, $4, $5, $6); |
1111 |
|
|
|
1112 |
root |
1.79 |
} elsif ($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])$/) { |
1113 |
root |
1.61 |
# ISO C's asctime |
1114 |
|
|
($d, $m, $y, $H, $M, $S) = ($2, $1, $6, $3, $4, $5); |
1115 |
|
|
} |
1116 |
|
|
# other formats fail in the loop below |
1117 |
|
|
|
1118 |
|
|
for (0..11) { |
1119 |
|
|
if ($m eq $month[$_]) { |
1120 |
|
|
require Time::Local; |
1121 |
|
|
return Time::Local::timegm ($S, $M, $H, $d, $_, $y); |
1122 |
|
|
} |
1123 |
|
|
} |
1124 |
|
|
|
1125 |
|
|
undef |
1126 |
|
|
} |
1127 |
|
|
|
1128 |
root |
1.2 |
sub set_proxy($) { |
1129 |
root |
1.52 |
if (length $_[0]) { |
1130 |
|
|
$_[0] =~ m%^(https?):// ([^:/]+) (?: : (\d*) )?%ix |
1131 |
|
|
or Carp::croak "$_[0]: invalid proxy URL"; |
1132 |
|
|
$PROXY = [$2, $3 || 3128, $1] |
1133 |
|
|
} else { |
1134 |
|
|
undef $PROXY; |
1135 |
|
|
} |
1136 |
root |
1.2 |
} |
1137 |
|
|
|
1138 |
|
|
# initialise proxy from environment |
1139 |
root |
1.52 |
eval { |
1140 |
|
|
set_proxy $ENV{http_proxy}; |
1141 |
|
|
}; |
1142 |
root |
1.2 |
|
1143 |
root |
1.60 |
=head2 SOCKS PROXIES |
1144 |
|
|
|
1145 |
|
|
Socks proxies are not directly supported by AnyEvent::HTTP. You can |
1146 |
|
|
compile your perl to support socks, or use an external program such as |
1147 |
|
|
F<socksify> (dante) or F<tsocks> to make your program use a socks proxy |
1148 |
|
|
transparently. |
1149 |
|
|
|
1150 |
|
|
Alternatively, for AnyEvent::HTTP only, you can use your own |
1151 |
|
|
C<tcp_connect> function that does the proxy handshake - here is an example |
1152 |
|
|
that works with socks4a proxies: |
1153 |
|
|
|
1154 |
|
|
use Errno; |
1155 |
|
|
use AnyEvent::Util; |
1156 |
|
|
use AnyEvent::Socket; |
1157 |
|
|
use AnyEvent::Handle; |
1158 |
|
|
|
1159 |
|
|
# host, port and username of/for your socks4a proxy |
1160 |
|
|
my $socks_host = "10.0.0.23"; |
1161 |
|
|
my $socks_port = 9050; |
1162 |
|
|
my $socks_user = ""; |
1163 |
|
|
|
1164 |
|
|
sub socks4a_connect { |
1165 |
|
|
my ($host, $port, $connect_cb, $prepare_cb) = @_; |
1166 |
|
|
|
1167 |
|
|
my $hdl = new AnyEvent::Handle |
1168 |
|
|
connect => [$socks_host, $socks_port], |
1169 |
|
|
on_prepare => sub { $prepare_cb->($_[0]{fh}) }, |
1170 |
|
|
on_error => sub { $connect_cb->() }, |
1171 |
|
|
; |
1172 |
|
|
|
1173 |
|
|
$hdl->push_write (pack "CCnNZ*Z*", 4, 1, $port, 1, $socks_user, $host); |
1174 |
|
|
|
1175 |
|
|
$hdl->push_read (chunk => 8, sub { |
1176 |
|
|
my ($hdl, $chunk) = @_; |
1177 |
|
|
my ($status, $port, $ipn) = unpack "xCna4", $chunk; |
1178 |
|
|
|
1179 |
|
|
if ($status == 0x5a) { |
1180 |
|
|
$connect_cb->($hdl->{fh}, (format_address $ipn) . ":$port"); |
1181 |
|
|
} else { |
1182 |
|
|
$! = Errno::ENXIO; $connect_cb->(); |
1183 |
|
|
} |
1184 |
|
|
}); |
1185 |
|
|
|
1186 |
|
|
$hdl |
1187 |
|
|
} |
1188 |
|
|
|
1189 |
|
|
Use C<socks4a_connect> instead of C<tcp_connect> when doing C<http_request>s, |
1190 |
|
|
possibly after switching off other proxy types: |
1191 |
|
|
|
1192 |
|
|
AnyEvent::HTTP::set_proxy undef; # usually you do not want other proxies |
1193 |
|
|
|
1194 |
|
|
http_get 'http://www.google.com', tcp_connect => \&socks4a_connect, sub { |
1195 |
|
|
my ($data, $headers) = @_; |
1196 |
|
|
... |
1197 |
|
|
}; |
1198 |
|
|
|
1199 |
root |
1.1 |
=head1 SEE ALSO |
1200 |
|
|
|
1201 |
|
|
L<AnyEvent>. |
1202 |
|
|
|
1203 |
|
|
=head1 AUTHOR |
1204 |
|
|
|
1205 |
root |
1.18 |
Marc Lehmann <schmorp@schmorp.de> |
1206 |
|
|
http://home.schmorp.de/ |
1207 |
root |
1.1 |
|
1208 |
root |
1.36 |
With many thanks to Дмитрий Шалашов, who provided countless |
1209 |
|
|
testcases and bugreports. |
1210 |
|
|
|
1211 |
root |
1.1 |
=cut |
1212 |
|
|
|
1213 |
|
|
1 |
1214 |
|
|
|