X-Git-Url: https://git.mxchange.org/?a=blobdiff_plain;f=classes%2FFile_redirection.php;h=53c15bf8b213b95e43a78e937eee634151cfd973;hb=2e2519afee87009165c97026737f72634461e82b;hp=79052bf7d3e80ca6e5b9347ee2062e5e9cc8ad22;hpb=39598582d90eab2a8abaf2ca01bcad6cb09d2c73;p=quix0rs-gnu-social.git diff --git a/classes/File_redirection.php b/classes/File_redirection.php index 79052bf7d3..53c15bf8b2 100644 --- a/classes/File_redirection.php +++ b/classes/File_redirection.php @@ -47,31 +47,119 @@ class File_redirection extends Memcached_DataObject /* the code above is auto generated do not remove the tag below */ ###END_AUTOCODE - function _commonCurl($url, $redirs) { - $curlh = curl_init(); - curl_setopt($curlh, CURLOPT_URL, $url); - curl_setopt($curlh, CURLOPT_AUTOREFERER, true); // # setup referer header when folowing redirects - curl_setopt($curlh, CURLOPT_CONNECTTIMEOUT, 10); // # seconds to wait - curl_setopt($curlh, CURLOPT_MAXREDIRS, $redirs); // # max number of http redirections to follow - curl_setopt($curlh, CURLOPT_USERAGENT, USER_AGENT); - curl_setopt($curlh, CURLOPT_FOLLOWLOCATION, true); // Follow redirects - curl_setopt($curlh, CURLOPT_RETURNTRANSFER, true); - curl_setopt($curlh, CURLOPT_FILETIME, true); - curl_setopt($curlh, CURLOPT_HEADER, true); // Include header in output - return $curlh; + static function _commonHttp($url, $redirs) { + $request = new HTTPClient($url); + $request->setConfig(array( + 'connect_timeout' => 10, // # seconds to wait + 'max_redirs' => $redirs, // # max number of http redirections to follow + 'follow_redirects' => true, // Follow redirects + 'store_body' => false, // We won't need body content here. + )); + return $request; } - function _redirectWhere_imp($short_url, $redirs = 10, $protected = false) { + /** + * Check if this URL is a redirect and return redir info. + * + * Most code should call File_redirection::where instead, to check if we + * already know that redirection and avoid extra hits to the web. + * + * The URL is hit and any redirects are followed, up to 10 levels or until + * a protected URL is reached. + * + * @param string $in_url + * @return mixed one of: + * string - target URL, if this is a direct link or can't be followed + * array - redirect info if this is an *unknown* redirect: + * associative array with the following elements: + * code: HTTP status code + * redirects: count of redirects followed + * url: URL string of final target + * type (optional): MIME type from Content-Type header + * size (optional): byte size from Content-Length header + * time (optional): timestamp from Last-Modified header + */ + public function lookupWhere($short_url, $redirs = 10, $protected = false) { if ($redirs < 0) return false; + if(strpos($short_url,'://') === false){ + return $short_url; + } + try { + $request = self::_commonHttp($short_url, $redirs); + // Don't include body in output + $request->setMethod(HTTP_Request2::METHOD_HEAD); + $response = $request->send(); + + if (405 == $response->getStatus() || 204 == $response->getStatus()) { + // HTTP 405 Unsupported Method + // Server doesn't support HEAD method? Can this really happen? + // We'll try again as a GET and ignore the response data. + // + // HTTP 204 No Content + // YFrog sends 204 responses back for our HEAD checks, which + // seems like it may be a logic error in their servers. If + // we get a 204 back, re-run it as a GET... if there's really + // no content it'll be cheap. :) + $request = self::_commonHttp($short_url, $redirs); + $response = $request->send(); + } + } catch (Exception $e) { + // Invalid URL or failure to reach server + common_log(LOG_ERR, "Error while following redirects for $short_url: " . $e->getMessage()); + return $short_url; + } + + if ($response->getRedirectCount() && File::isProtected($response->getUrl())) { + // Bump back up the redirect chain until we find a non-protected URL + return self::lookupWhere($short_url, $response->getRedirectCount() - 1, true); + } + + $ret = array('code' => $response->getStatus() + , 'redirects' => $response->getRedirectCount() + , 'url' => $response->getUrl()); + + $type = $response->getHeader('Content-Type'); + if ($type) $ret['type'] = $type; + if ($protected) $ret['protected'] = true; + $size = $response->getHeader('Content-Length'); // @fixme bytes? + if ($size) $ret['size'] = $size; + $time = $response->getHeader('Last-Modified'); + if ($time) $ret['time'] = strtotime($time); + return $ret; + } + + /** + * Check if this URL is a redirect and return redir info. + * If a File record is present for this URL, it is not considered a redirect. + * If a File_redirection record is present for this URL, the recorded target is returned. + * + * If no File or File_redirect record is present, the URL is hit and any + * redirects are followed, up to 10 levels or until a protected URL is + * reached. + * + * @param string $in_url + * @param boolean $discover true to attempt dereferencing the redirect if we don't know it already + * @return mixed one of: + * string - target URL, if this is a direct link or a known redirect + * array - redirect info if this is an *unknown* redirect: + * associative array with the following elements: + * code: HTTP status code + * redirects: count of redirects followed + * url: URL string of final target + * type (optional): MIME type from Content-Type header + * size (optional): byte size from Content-Length header + * time (optional): timestamp from Last-Modified header + */ + public function where($in_url, $discover=true) { // let's see if we know this... - $a = File::staticGet('url', $short_url); + $a = File::staticGet('url', $in_url); if (!empty($a)) { // this is a direct link to $a->url return $a->url; } else { - $b = File_redirection::staticGet('url', $short_url); + $b = File_redirection::staticGet('url', $in_url); if (!empty($b)) { // this is a redirect to $b->file_id $a = File::staticGet('id', $b->file_id); @@ -79,48 +167,34 @@ class File_redirection extends Memcached_DataObject } } - if(strpos($short_url,'://') === false){ - return $short_url; - } - $curlh = File_redirection::_commonCurl($short_url, $redirs); - // Don't include body in output - curl_setopt($curlh, CURLOPT_NOBODY, true); - curl_exec($curlh); - $info = curl_getinfo($curlh); - curl_close($curlh); - - if (405 == $info['http_code']) { - $curlh = File_redirection::_commonCurl($short_url, $redirs); - curl_exec($curlh); - $info = curl_getinfo($curlh); - curl_close($curlh); - } - - if (!empty($info['redirect_count']) && File::isProtected($info['url'])) { - return File_redirection::_redirectWhere_imp($short_url, $info['redirect_count'] - 1, true); + if ($discover) { + $ret = File_redirection::lookupWhere($in_url); + return $ret; + } else { + // No manual dereferencing; leave the unknown URL as is. + return $in_url; } - - $ret = array('code' => $info['http_code'] - , 'redirects' => $info['redirect_count'] - , 'url' => $info['url']); - - if (!empty($info['content_type'])) $ret['type'] = $info['content_type']; - if ($protected) $ret['protected'] = true; - if (!empty($info['download_content_length'])) $ret['size'] = $info['download_content_length']; - if (isset($info['filetime']) && ($info['filetime'] > 0)) $ret['time'] = $info['filetime']; - return $ret; } - function where($in_url) { - $ret = File_redirection::_redirectWhere_imp($in_url); - return $ret; - } - - function makeShort($long_url) { + /** + * Shorten a URL with the current user's configured shortening + * options, if applicable. + * + * If it cannot be shortened or the "short" URL is longer than the + * original, the original is returned. + * + * If the referenced item has not been seen before, embedding data + * may be saved. + * + * @param string $long_url + * @param User $user whose shortening options to use; defaults to the current web session user + * @return string + */ + function makeShort($long_url, $user=null) { $canon = File_redirection::_canonUrl($long_url); - $short_url = File_redirection::_userMakeShort($canon); + $short_url = File_redirection::_userMakeShort($canon, $user); // Did we get one? Is it shorter? if (!empty($short_url) && mb_strlen($short_url) < mb_strlen($long_url)) { @@ -130,18 +204,35 @@ class File_redirection extends Memcached_DataObject } } - function _userMakeShort($long_url) { - $short_url = common_shorten_url($long_url); + function _userMakeShort($long_url, User $user=null) { + $short_url = common_shorten_url($long_url, $user); if (!empty($short_url) && $short_url != $long_url) { $short_url = (string)$short_url; // store it $file = File::staticGet('url', $long_url); if (empty($file)) { + // Check if the target URL is itself a redirect... $redir_data = File_redirection::where($long_url); - $file = File::saveNew($redir_data, $long_url); - $file_id = $file->id; - if (!empty($redir_data['oembed']['json'])) { - File_oembed::saveNew($redir_data['oembed']['json'], $file_id); + if (is_array($redir_data)) { + // We haven't seen the target URL before. + // Save file and embedding data about it! + $file = File::saveNew($redir_data, $long_url); + $file_id = $file->id; + if (!empty($redir_data['oembed']['json'])) { + File_oembed::saveNew($redir_data['oembed']['json'], $file_id); + } + } else if (is_string($redir_data)) { + // The file is a known redirect target. + $file = File::staticGet('url', $redir_data); + if (empty($file)) { + // @fixme should we save a new one? + // this case was triggering sometimes for redirects + // with unresolvable targets; found while fixing + // "can't linkify" bugs with shortened links to + // SSL sites with cert issues. + return null; + } + $file_id = $file->id; } } else { $file_id = $file->id; @@ -158,6 +249,18 @@ class File_redirection extends Memcached_DataObject return null; } + /** + * Basic attempt to canonicalize a URL, cleaning up some standard variants + * such as funny syntax or a missing path. Used internally when cleaning + * up URLs for storage and following redirect chains. + * + * Note that despite being on File_redirect, this function DOES NOT perform + * any dereferencing of redirects. + * + * @param string $in_url input URL + * @param string $default_scheme if given a bare link; defaults to 'http://' + * @return string + */ function _canonUrl($in_url, $default_scheme = 'http://') { if (empty($in_url)) return false; $out_url = $in_url; @@ -204,4 +307,3 @@ class File_redirection extends Memcached_DataObject $file_redir->insert(); } } -