3 * StatusNet, the distributed open-source microblogging tool
5 * Utility for doing HTTP-related things
9 * LICENCE: This program is free software: you can redistribute it and/or modify
10 * it under the terms of the GNU Affero General Public License as published by
11 * the Free Software Foundation, either version 3 of the License, or
12 * (at your option) any later version.
14 * This program is distributed in the hope that it will be useful,
15 * but WITHOUT ANY WARRANTY; without even the implied warranty of
16 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
17 * GNU Affero General Public License for more details.
19 * You should have received a copy of the GNU Affero General Public License
20 * along with this program. If not, see <http://www.gnu.org/licenses/>.
24 * @author Evan Prodromou <evan@status.net>
25 * @copyright 2009 StatusNet, Inc.
26 * @license http://www.fsf.org/licensing/licenses/agpl-3.0.html GNU Affero General Public License version 3.0
27 * @link http://status.net/
30 if (!defined('GNUSOCIAL')) { exit(1); }
33 * Useful structure for HTTP responses
35 * We make HTTP calls in several places, and we have several different
36 * ways of doing them. This class hides the specifics of what underlying
37 * library (curl or PHP-HTTP or whatever) that's used.
39 * This extends the HTTP_Request2_Response class with methods to get info
40 * about any followed redirects.
42 * Originally used the name 'HTTPResponse' to match earlier code, but
43 * this conflicts with a class in in the PECL HTTP extension.
47 * @author Evan Prodromou <evan@status.net>
48 * @author Brion Vibber <brion@status.net>
49 * @license http://www.fsf.org/licensing/licenses/agpl-3.0.html GNU Affero General Public License version 3.0
50 * @link http://status.net/
52 class GNUsocial_HTTPResponse extends HTTP_Request2_Response
54 function __construct(HTTP_Request2_Response $response, $url, $redirects=0)
56 foreach (get_object_vars($response) as $key => $val) {
59 $this->url = strval($url);
60 $this->redirectCount = intval($redirects);
64 * Get the count of redirects that have been followed, if any.
67 function getRedirectCount()
69 return $this->redirectCount;
73 * Gets the target URL, before any redirects. Use getEffectiveUrl() for final target.
82 * Check if the response is OK, generally a 200 or other 2xx status code.
87 $status = $this->getStatus();
88 return ($status >= 200 && $status < 300);
93 * Utility class for doing HTTP client stuff
95 * We make HTTP calls in several places, and we have several different
96 * ways of doing them. This class hides the specifics of what underlying
97 * library (curl or PHP-HTTP or whatever) that's used.
99 * This extends the PEAR HTTP_Request2 package:
100 * - sends StatusNet-specific User-Agent header
101 * - 'follow_redirects' config option, defaulting on
102 * - 'max_redirs' config option, defaulting to 10
103 * - extended response class adds getRedirectCount() and getUrl() methods
104 * - get() and post() convenience functions return body content directly
108 * @author Evan Prodromou <evan@status.net>
109 * @author Brion Vibber <brion@status.net>
110 * @license http://www.fsf.org/licensing/licenses/agpl-3.0.html GNU Affero General Public License version 3.0
111 * @link http://status.net/
114 class HTTPClient extends HTTP_Request2
117 function __construct($url=null, $method=self::METHOD_GET, $config=array())
119 if (is_int(common_config('http', 'timeout'))) {
120 // Reasonably you shouldn't set http/timeout to 0 because of
121 // malicious remote servers that can cause infinitely long
122 // responses... But the default in HTTP_Request2 is 0 for
123 // some reason and should probably be considered a valid value.
124 $this->config['timeout'] = common_config('http', 'timeout');
126 common_log(LOG_ERR, 'config option http/timeout is not an integer value: '._ve(common_config('http', 'timeout')));
128 if (!empty(common_config('http', 'connect_timeout'))) {
129 $this->config['connect_timeout'] = common_config('http', 'connect_timeout');
131 $this->config['max_redirs'] = 10;
132 $this->config['follow_redirects'] = true;
134 // We've had some issues with keepalive breaking with
135 // HEAD requests, such as to youtube which seems to be
136 // emitting chunked encoding info for an empty body
137 // instead of not emitting anything. This may be a
138 // bug on YouTube's end, but the upstream libray
139 // ought to be investigated to see if we can handle
140 // it gracefully in that case as well.
141 $this->config['protocol_version'] = '1.0';
143 // Default state of OpenSSL seems to have no trusted
144 // SSL certificate authorities, which breaks hostname
145 // verification and means we have a hard time communicating
146 // with other sites' HTTPS interfaces.
148 // Turn off verification unless we've configured a CA bundle.
149 if (common_config('http', 'ssl_cafile')) {
150 $this->config['ssl_cafile'] = common_config('http', 'ssl_cafile');
152 $this->config['ssl_verify_peer'] = false;
155 // This means "verify the cert hostname against what we connect to", it does not
156 // imply CA trust or anything like that. Just the hostname.
157 $this->config['ssl_verify_host'] = common_config('http', 'ssl_verify_host');
159 if (common_config('http', 'curl') && extension_loaded('curl')) {
160 $this->config['adapter'] = 'HTTP_Request2_Adapter_Curl';
163 foreach (array('host', 'port', 'user', 'password', 'auth_scheme') as $cf) {
165 $v = common_config('http', $k);
167 $this->config[$k] = $v;
171 parent::__construct($url, $method, $config);
172 $this->setHeader('User-Agent', self::userAgent());
176 * Convenience/back-compat instantiator
179 public static function start()
181 return new HTTPClient();
185 * Quick static function to GET a URL
187 public static function quickGet($url, $accept=null, array $params=array(), array $headers=array())
189 if (!empty($params)) {
190 $params = http_build_query($params, null, '&');
191 if (strpos($url, '?') === false) {
192 $url .= '?' . $params;
194 $url .= '&' . $params;
198 $client = new HTTPClient();
199 if (!is_null($accept)) {
200 $client->setHeader('Accept', $accept);
202 $response = $client->get($url, $headers);
203 if (!$response->isOk()) {
204 // TRANS: Exception. %s is the URL we tried to GET.
205 throw new Exception(sprintf(_m('Could not GET URL %s.'), $url), $response->getStatus());
207 return $response->getBody();
210 public static function quickGetJson($url, $params=array())
212 $data = json_decode(self::quickGet($url, null, $params));
213 if (is_null($data)) {
214 common_debug('Could not decode JSON data from URL: '.$url);
215 throw new ServerException('Could not decode JSON data from URL');
221 * If you want an Accept header, put it in $headers
223 public static function quickHead($url, array $params=array(), array $headers=array())
225 if (!empty($params)) {
226 $params = http_build_query($params, null, '&');
227 if (strpos($url, '?') === false) {
228 $url .= '?' . $params;
230 $url .= '&' . $params;
234 $client = new HTTPClient();
235 $response = $client->head($url, $headers);
236 if (!$response->isOk()) {
237 // TRANS: Exception. %s is the URL we tried to GET.
238 throw new Exception(sprintf(_m('Could not GET URL %s.'), $url), $response->getStatus());
240 return $response->getHeader();
244 * Convenience function to run a GET request.
246 * @return GNUsocial_HTTPResponse
247 * @throws HTTP_Request2_Exception
249 public function get($url, $headers=array())
251 return $this->doRequest($url, self::METHOD_GET, $headers);
255 * Convenience function to run a HEAD request.
257 * NOTE: Will probably turn into a GET request if you let it follow redirects!
258 * That option is only there to be flexible and may be removed in the future!
260 * @return GNUsocial_HTTPResponse
261 * @throws HTTP_Request2_Exception
263 public function head($url, $headers=array(), $follow_redirects=false)
265 // Save the configured value for follow_redirects
266 $old_follow = $this->config['follow_redirects'];
268 // Temporarily (possibly) override the follow_redirects setting
269 $this->config['follow_redirects'] = $follow_redirects;
270 return $this->doRequest($url, self::METHOD_HEAD, $headers);
271 } catch (Exception $e) {
272 // Let the exception go on its merry way.
275 // reset to the old value
276 $this->config['follow_redirects'] = $old_follow;
278 //we've either returned or thrown exception here
282 * Convenience function to POST form data.
285 * @param array $headers optional associative array of HTTP headers
286 * @param array $data optional associative array or blob of form data to submit
287 * @return GNUsocial_HTTPResponse
288 * @throws HTTP_Request2_Exception
290 public function post($url, $headers=array(), $data=array())
293 $this->addPostParameter($data);
295 return $this->doRequest($url, self::METHOD_POST, $headers);
299 * @param string $url The URL including possible querystring
300 * @param string $method The HTTP method to use
301 * @param array $headers List of already formatted strings
302 * (not an associative array, to allow
303 * multiple same-named headers)
305 * @return GNUsocial_HTTPResponse
306 * @throws HTTP_Request2_Exception
308 protected function doRequest($url, $method, array $headers=array())
312 // Workaround for HTTP_Request2 not setting up SNI in socket contexts;
313 // This fixes cert validation for SSL virtual hosts using SNI.
314 // Requires PHP 5.3.2 or later and OpenSSL with SNI support.
315 if ($this->url->getScheme() == 'https' && defined('OPENSSL_TLSEXT_SERVER_NAME')) {
316 $this->config['ssl_SNI_enabled'] = true;
317 $this->config['ssl_SNI_server_name'] = $this->url->getHost();
320 $this->setMethod($method);
321 foreach ($headers as $header) {
322 $this->setHeader($header);
324 $response = $this->send();
325 if (is_null($response)) {
326 // TRANS: Failed to retrieve a remote web resource, %s is the target URL.
327 throw new NoHttpResponseException($url);
332 protected function log($level, $detail) {
333 $method = $this->getMethod();
334 $url = $this->getUrl();
335 common_log($level, __CLASS__ . ": HTTP $method $url - $detail");
339 * Pulls up GNU Social's customized user-agent string, so services
340 * we hit can track down the responsible software.
344 static public function userAgent()
346 return GNUSOCIAL_ENGINE . '/' . GNUSOCIAL_VERSION
347 . ' (' . GNUSOCIAL_CODENAME . ')';
351 * Actually performs the HTTP request and returns a
352 * GNUsocial_HTTPResponse object with response body and header info.
354 * Wraps around parent send() to add logging and redirection processing.
356 * @return GNUsocial_HTTPResponse
357 * @throw HTTP_Request2_Exception
359 public function send()
361 $maxRedirs = intval($this->config['max_redirs']);
362 if (empty($this->config['max_redirs'])) {
366 $redirUrls = array();
369 $response = parent::send();
370 } catch (Exception $e) {
371 $this->log(LOG_ERR, $e->getMessage());
374 $code = $response->getStatus();
375 $effectiveUrl = $response->getEffectiveUrl();
376 $redirUrls[] = $effectiveUrl;
377 $response->redirUrls = $redirUrls;
378 if ($code >= 200 && $code < 300) {
379 $reason = $response->getReasonPhrase();
380 $this->log(LOG_INFO, "$code $reason");
381 } elseif ($code >= 300 && $code < 400) {
382 $url = $this->getUrl();
383 $target = $response->getHeader('Location');
385 if (++$redirs >= $maxRedirs) {
386 common_log(LOG_ERR, __CLASS__ . ": Too many redirects: skipping $code redirect from $url to $target");
390 $this->setUrl($target);
391 $this->setHeader('Referer', $url);
392 common_log(LOG_INFO, __CLASS__ . ": Following $code redirect from $url to $target");
394 } catch (HTTP_Request2_Exception $e) {
395 common_log(LOG_ERR, __CLASS__ . ": Invalid $code redirect from $url to $target");
398 $reason = $response->getReasonPhrase();
399 $this->log(LOG_ERR, "$code $reason");
402 } while ($maxRedirs);
403 return new GNUsocial_HTTPResponse($response, $this->getUrl(), $redirs);