5) { throw new Exception('Too many redirects (' . $followDepth . ')'); } $contentLength = 0; $fileLength = 0; if (!empty($requestBody) && is_array($requestBody)) { $requestBody = self::buildQuery($requestBody); } // Piwik services behave like a proxy, so we should act like one. $xff = 'X-Forwarded-For: ' . (isset($_SERVER['HTTP_X_FORWARDED_FOR']) && !empty($_SERVER['HTTP_X_FORWARDED_FOR']) ? $_SERVER['HTTP_X_FORWARDED_FOR'] . ',' : '') . IP::getIpFromHeader(); if (empty($userAgent)) { $userAgent = self::getUserAgent(); } $via = 'Via: ' . (isset($_SERVER['HTTP_VIA']) && !empty($_SERVER['HTTP_VIA']) ? $_SERVER['HTTP_VIA'] . ', ' : '') . Version::VERSION . ' ' . ($userAgent ? " ($userAgent)" : ''); // range header $rangeHeader = ''; if (!empty($byteRange)) { $rangeHeader = 'Range: bytes=' . $byteRange[0] . '-' . $byteRange[1] . "\r\n"; } list($proxyHost, $proxyPort, $proxyUser, $proxyPassword) = self::getProxyConfiguration($aUrl); $aUrl = trim($aUrl); // other result data $status = null; $headers = array(); $httpAuthIsUsed = !empty($httpUsername) || !empty($httpPassword); if ($method == 'socket') { if (!self::isSocketEnabled()) { // can be triggered in tests throw new Exception("HTTP socket support is not enabled (php function fsockopen is not available) "); } // initialization $url = @parse_url($aUrl); if ($url === false || !isset($url['scheme'])) { throw new Exception('Malformed URL: ' . $aUrl); } if ($url['scheme'] != 'http' && $url['scheme'] != 'https') { throw new Exception('Invalid protocol/scheme: ' . $url['scheme']); } $host = $url['host']; $port = isset($url['port']) ? $url['port'] : ('https' == $url['scheme'] ? 443 : 80); $path = isset($url['path']) ? $url['path'] : '/'; if (isset($url['query'])) { $path .= '?' . $url['query']; } $errno = null; $errstr = null; if ((!empty($proxyHost) && !empty($proxyPort)) || !empty($byteRange) ) { $httpVer = '1.1'; } else { $httpVer = '1.0'; } $proxyAuth = null; if (!empty($proxyHost) && !empty($proxyPort)) { $connectHost = $proxyHost; $connectPort = $proxyPort; if (!empty($proxyUser) && !empty($proxyPassword)) { $proxyAuth = 'Proxy-Authorization: Basic ' . base64_encode("$proxyUser:$proxyPassword") . "\r\n"; } $requestHeader = "$httpMethod $aUrl HTTP/$httpVer\r\n"; } else { $connectHost = $host; $connectPort = $port; $requestHeader = "$httpMethod $path HTTP/$httpVer\r\n"; if ('https' == $url['scheme']) { $connectHost = 'ssl://' . $connectHost; } } // connection attempt if (($fsock = @fsockopen($connectHost, $connectPort, $errno, $errstr, $timeout)) === false || !is_resource($fsock)) { if (is_resource($file)) { @fclose($file); } throw new Exception("Error while connecting to: $host. Please try again later. $errstr"); } $httpAuth = ''; if ($httpAuthIsUsed) { $httpAuth = 'Authorization: Basic ' . base64_encode($httpUsername.':'.$httpPassword) . "\r\n"; } // send HTTP request header $requestHeader .= "Host: $host" . ($port != 80 && ('https' == $url['scheme'] && $port != 443) ? ':' . $port : '') . "\r\n" . ($httpAuth ? $httpAuth : '') . ($proxyAuth ? $proxyAuth : '') . 'User-Agent: ' . $userAgent . "\r\n" . ($acceptLanguage ? $acceptLanguage . "\r\n" : '') . $xff . "\r\n" . $via . "\r\n" . $rangeHeader . (!empty($additionalHeaders) ? implode("\r\n", $additionalHeaders) . "\r\n" : '') . "Connection: close\r\n"; fwrite($fsock, $requestHeader); if (strtolower($httpMethod) === 'post' && !empty($requestBody)) { fwrite($fsock, self::buildHeadersForPost($requestBody)); fwrite($fsock, "\r\n"); fwrite($fsock, $requestBody); } else { fwrite($fsock, "\r\n"); } $streamMetaData = array('timed_out' => false); @stream_set_blocking($fsock, true); if (function_exists('stream_set_timeout')) { @stream_set_timeout($fsock, $timeout); } elseif (function_exists('socket_set_timeout')) { @socket_set_timeout($fsock, $timeout); } // process header $status = null; while (!feof($fsock)) { $line = fgets($fsock, 4096); $streamMetaData = @stream_get_meta_data($fsock); if ($streamMetaData['timed_out']) { if (is_resource($file)) { @fclose($file); } @fclose($fsock); throw new Exception('Timed out waiting for server response'); } // a blank line marks the end of the server response header if (rtrim($line, "\r\n") == '') { break; } // parse first line of server response header if (!$status) { // expect first line to be HTTP response status line, e.g., HTTP/1.1 200 OK if (!preg_match('~^HTTP/(\d\.\d)\s+(\d+)(\s*.*)?~', $line, $m)) { if (is_resource($file)) { @fclose($file); } @fclose($fsock); throw new Exception('Expected server response code. Got ' . rtrim($line, "\r\n")); } $status = (integer)$m[2]; // Informational 1xx or Client Error 4xx if ($status < 200 || $status >= 400) { if (is_resource($file)) { @fclose($file); } @fclose($fsock); if (!$getExtendedInfo) { return false; } else { return array('status' => $status); } } continue; } // handle redirect if (preg_match('/^Location:\s*(.+)/', rtrim($line, "\r\n"), $m)) { if (is_resource($file)) { @fclose($file); } @fclose($fsock); // Successful 2xx vs Redirect 3xx if ($status < 300) { throw new Exception('Unexpected redirect to Location: ' . rtrim($line) . ' for status code ' . $status); } return self::sendHttpRequestBy( $method, trim($m[1]), $timeout, $userAgent, $destinationPath, $file, $followDepth + 1, $acceptLanguage, $acceptInvalidSslCertificate = false, $byteRange, $getExtendedInfo, $httpMethod, $httpUsername, $httpPassword, $requestBody, $additionalHeaders ); } // save expected content length for later verification if (preg_match('/^Content-Length:\s*(\d+)/', $line, $m)) { $contentLength = (integer)$m[1]; } self::parseHeaderLine($headers, $line); } if (feof($fsock) && $httpMethod != 'HEAD' ) { throw new Exception('Unexpected end of transmission'); } // process content/body $response = ''; while (!feof($fsock)) { $line = fread($fsock, 8192); $streamMetaData = @stream_get_meta_data($fsock); if ($streamMetaData['timed_out']) { if (is_resource($file)) { @fclose($file); } @fclose($fsock); throw new Exception('Timed out waiting for server response'); } $fileLength += strlen($line); if (is_resource($file)) { // save to file fwrite($file, $line); } else { // concatenate to response string $response .= $line; } } // determine success or failure @fclose(@$fsock); } elseif ($method == 'fopen') { $response = false; // we make sure the request takes less than a few seconds to fail // we create a stream_context (works in php >= 5.2.1) // we also set the socket_timeout (for php < 5.2.1) $default_socket_timeout = @ini_get('default_socket_timeout'); @ini_set('default_socket_timeout', $timeout); $httpAuth = ''; if ($httpAuthIsUsed) { $httpAuth = 'Authorization: Basic ' . base64_encode($httpUsername.':'.$httpPassword) . "\r\n"; } $ctx = null; if (function_exists('stream_context_create')) { $stream_options = array( 'http' => array( 'header' => 'User-Agent: ' . $userAgent . "\r\n" . ($httpAuth ? $httpAuth : '') . ($acceptLanguage ? $acceptLanguage . "\r\n" : '') . $xff . "\r\n" . $via . "\r\n" . (!empty($additionalHeaders) ? implode("\r\n", $additionalHeaders) . "\r\n" : '') . $rangeHeader, 'max_redirects' => 5, // PHP 5.1.0 'timeout' => $timeout, // PHP 5.2.1 ) ); if (!empty($proxyHost) && !empty($proxyPort)) { $stream_options['http']['proxy'] = 'tcp://' . $proxyHost . ':' . $proxyPort; $stream_options['http']['request_fulluri'] = true; // required by squid proxy if (!empty($proxyUser) && !empty($proxyPassword)) { $stream_options['http']['header'] .= 'Proxy-Authorization: Basic ' . base64_encode("$proxyUser:$proxyPassword") . "\r\n"; } } if (strtolower($httpMethod) === 'post' && !empty($requestBody)) { $postHeader = self::buildHeadersForPost($requestBody); $postHeader .= "\r\n"; $stream_options['http']['method'] = 'POST'; $stream_options['http']['header'] .= $postHeader; $stream_options['http']['content'] = $requestBody; } $ctx = stream_context_create($stream_options); } // save to file if (is_resource($file)) { if (!($handle = fopen($aUrl, 'rb', false, $ctx))) { throw new Exception("Unable to open $aUrl"); } while (!feof($handle)) { $response = fread($handle, 8192); $fileLength += strlen($response); fwrite($file, $response); } fclose($handle); } else { $response = @file_get_contents($aUrl, 0, $ctx); // try to get http status code from response headers if (isset($http_response_header) && preg_match('~^HTTP/(\d\.\d)\s+(\d+)(\s*.*)?~', implode("\n", $http_response_header), $m)) { $status = (int)$m[2]; } if (!$status && $response === false) { $error = error_get_last(); throw new \Exception($error['message']); } $fileLength = strlen($response); } // restore the socket_timeout value if (!empty($default_socket_timeout)) { @ini_set('default_socket_timeout', $default_socket_timeout); } } elseif ($method == 'curl') { if (!self::isCurlEnabled()) { // can be triggered in tests throw new Exception("CURL is not enabled in php.ini, but is being used."); } $ch = @curl_init(); if (!empty($proxyHost) && !empty($proxyPort)) { @curl_setopt($ch, CURLOPT_PROXY, $proxyHost . ':' . $proxyPort); if (!empty($proxyUser) && !empty($proxyPassword)) { // PROXYAUTH defaults to BASIC @curl_setopt($ch, CURLOPT_PROXYUSERPWD, $proxyUser . ':' . $proxyPassword); } } $curl_options = array( // internal to ext/curl CURLOPT_BINARYTRANSFER => is_resource($file), // curl options (sorted oldest to newest) CURLOPT_URL => $aUrl, CURLOPT_USERAGENT => $userAgent, CURLOPT_HTTPHEADER => array_merge(array( $xff, $via, $rangeHeader, $acceptLanguage ), $additionalHeaders), // only get header info if not saving directly to file CURLOPT_HEADER => is_resource($file) ? false : true, CURLOPT_CONNECTTIMEOUT => $timeout, CURLOPT_TIMEOUT => $timeout, ); // Case core:archive command is triggering archiving on https:// and the certificate is not valid if ($acceptInvalidSslCertificate) { $curl_options += array( CURLOPT_SSL_VERIFYHOST => false, CURLOPT_SSL_VERIFYPEER => false, ); } @curl_setopt($ch, CURLOPT_CUSTOMREQUEST, $httpMethod); if ($httpMethod == 'HEAD') { @curl_setopt($ch, CURLOPT_NOBODY, true); } if (strtolower($httpMethod) === 'post' && !empty($requestBody)) { curl_setopt($ch, CURLOPT_POST, 1); curl_setopt($ch, CURLOPT_POSTFIELDS, $requestBody); } if (!empty($httpUsername) && !empty($httpPassword)) { $curl_options += array( CURLOPT_USERPWD => $httpUsername . ':' . $httpPassword, ); } @curl_setopt_array($ch, $curl_options); self::configCurlCertificate($ch); /* * as of php 5.2.0, CURLOPT_FOLLOWLOCATION can't be set if * in safe_mode or open_basedir is set */ if ((string)ini_get('safe_mode') == '' && ini_get('open_basedir') == '') { $curl_options = array( // curl options (sorted oldest to newest) CURLOPT_FOLLOWLOCATION => true, CURLOPT_MAXREDIRS => 5, ); @curl_setopt_array($ch, $curl_options); } if (is_resource($file)) { // write output directly to file @curl_setopt($ch, CURLOPT_FILE, $file); } else { // internal to ext/curl @curl_setopt($ch, CURLOPT_RETURNTRANSFER, true); } ob_start(); $response = @curl_exec($ch); ob_end_clean(); if ($response === true) { $response = ''; } elseif ($response === false) { $errstr = curl_error($ch); if ($errstr != '') { throw new Exception('curl_exec: ' . $errstr . '. Hostname requested was: ' . UrlHelper::getHostFromUrl($aUrl)); } $response = ''; } else { $header = ''; // redirects are included in the output html, so we look for the last line that starts w/ HTTP/... // to split the response while (substr($response, 0, 5) == "HTTP/") { $split = explode("\r\n\r\n", $response, 2); if(count($split) == 2) { list($header, $response) = $split; } else { $response = ''; $header = $split; } } foreach (explode("\r\n", $header) as $line) { self::parseHeaderLine($headers, $line); } } $contentLength = @curl_getinfo($ch, CURLINFO_CONTENT_LENGTH_DOWNLOAD); $fileLength = is_resource($file) ? @curl_getinfo($ch, CURLINFO_SIZE_DOWNLOAD) : strlen($response); $status = @curl_getinfo($ch, CURLINFO_HTTP_CODE); @curl_close($ch); unset($ch); } else { throw new Exception('Invalid request method: ' . $method); } if (is_resource($file)) { fflush($file); @fclose($file); $fileSize = filesize($destinationPath); if ((($contentLength > 0) && ($fileLength != $contentLength)) || ($fileSize != $fileLength) ) { throw new Exception('File size error: ' . $destinationPath . '; expected ' . $contentLength . ' bytes; received ' . $fileLength . ' bytes; saved ' . $fileSize . ' bytes to file'); } return true; } if (!$getExtendedInfo) { return trim($response); } else { return array( 'status' => $status, 'headers' => $headers, 'data' => $response ); } } public static function buildQuery($params) { return http_build_query($params, '', '&'); } private static function buildHeadersForPost($requestBody) { $postHeader = "Content-Type: application/x-www-form-urlencoded\r\n"; $postHeader .= "Content-Length: " . strlen($requestBody) . "\r\n"; return $postHeader; } /** * Downloads the next chunk of a specific file. The next chunk's byte range * is determined by the existing file's size and the expected file size, which * is stored in the option table before starting a download. The expected * file size is obtained through a `HEAD` HTTP request. * * _Note: this function uses the **Range** HTTP header to accomplish downloading in * parts. Not every server supports this header._ * * The proper use of this function is to call it once per request. The browser * should continue to send requests to Piwik which will in turn call this method * until the file has completely downloaded. In this way, the user can be informed * of a download's progress. * * **Example Usage** * * ``` * // browser JavaScript * var downloadFile = function (isStart) { * var ajax = new ajaxHelper(); * ajax.addParams({ * module: 'MyPlugin', * action: 'myAction', * isStart: isStart ? 1 : 0 * }, 'post'); * ajax.setCallback(function (response) { * var progress = response.progress * // ...update progress... * * downloadFile(false); * }); * ajax.send(); * } * * downloadFile(true); * ``` * * ``` * // PHP controller action * public function myAction() * { * $outputPath = PIWIK_INCLUDE_PATH . '/tmp/averybigfile.zip'; * $isStart = Common::getRequestVar('isStart', 1, 'int'); * Http::downloadChunk("http://bigfiles.com/averybigfile.zip", $outputPath, $isStart == 1); * } * ``` * * @param string $url The url to download from. * @param string $outputPath The path to the file to save/append to. * @param bool $isContinuation `true` if this is the continuation of a download, * or if we're starting a fresh one. * @throws Exception if the file already exists and we're starting a new download, * if we're trying to continue a download that never started * @return array * @api */ public static function downloadChunk($url, $outputPath, $isContinuation) { // make sure file doesn't already exist if we're starting a new download if (!$isContinuation && file_exists($outputPath) ) { throw new Exception( Piwik::translate('General_DownloadFail_FileExists', "'" . $outputPath . "'") . ' ' . Piwik::translate('General_DownloadPleaseRemoveExisting')); } // if we're starting a download, get the expected file size & save as an option $downloadOption = $outputPath . '_expectedDownloadSize'; if (!$isContinuation) { $expectedFileSizeResult = Http::sendHttpRequest( $url, $timeout = 300, $userAgent = null, $destinationPath = null, $followDepth = 0, $acceptLanguage = false, $byteRange = false, $getExtendedInfo = true, $httpMethod = 'HEAD' ); $expectedFileSize = 0; if (isset($expectedFileSizeResult['headers']['Content-Length'])) { $expectedFileSize = (int)$expectedFileSizeResult['headers']['Content-Length']; } if ($expectedFileSize == 0) { Log::info("HEAD request for '%s' failed, got following: %s", $url, print_r($expectedFileSizeResult, true)); throw new Exception(Piwik::translate('General_DownloadFail_HttpRequestFail')); } Option::set($downloadOption, $expectedFileSize); } else { $expectedFileSize = (int)Option::get($downloadOption); if ($expectedFileSize === false) { // sanity check throw new Exception("Trying to continue a download that never started?! That's not supposed to happen..."); } } // if existing file is already big enough, then fail so we don't accidentally overwrite // existing DB $existingSize = file_exists($outputPath) ? filesize($outputPath) : 0; if ($existingSize >= $expectedFileSize) { throw new Exception( Piwik::translate('General_DownloadFail_FileExistsContinue', "'" . $outputPath . "'") . ' ' . Piwik::translate('General_DownloadPleaseRemoveExisting')); } // download a chunk of the file $result = Http::sendHttpRequest( $url, $timeout = 300, $userAgent = null, $destinationPath = null, $followDepth = 0, $acceptLanguage = false, $byteRange = array($existingSize, min($existingSize + 1024 * 1024 - 1, $expectedFileSize)), $getExtendedInfo = true ); if ($result === false || $result['status'] < 200 || $result['status'] > 299 ) { $result['data'] = self::truncateStr($result['data'], 1024); Log::info("Failed to download range '%s-%s' of file from url '%s'. Got result: %s", $byteRange[0], $byteRange[1], $url, print_r($result, true)); throw new Exception(Piwik::translate('General_DownloadFail_HttpRequestFail')); } // write chunk to file $f = fopen($outputPath, 'ab'); fwrite($f, $result['data']); fclose($f); clearstatcache($clear_realpath_cache = true, $outputPath); return array( 'current_size' => filesize($outputPath), 'expected_file_size' => $expectedFileSize, ); } /** * Will configure CURL handle $ch * to use local list of Certificate Authorities, */ public static function configCurlCertificate(&$ch) { $general = Config::getInstance()->General; if (!empty($general['custom_cacert_pem'])) { $cacertPath = $general['custom_cacert_pem']; } else { $cacertPath = PIWIK_INCLUDE_PATH . '/core/DataFiles/cacert.pem'; } @curl_setopt($ch, CURLOPT_CAINFO, $cacertPath); } public static function getUserAgent() { return !empty($_SERVER['HTTP_USER_AGENT']) ? $_SERVER['HTTP_USER_AGENT'] : 'Piwik/' . Version::VERSION; } /** * Fetches a file located at `$url` and saves it to `$destinationPath`. * * @param string $url The URL of the file to download. * @param string $destinationPath The path to download the file to. * @param int $tries (deprecated) * @param int $timeout The amount of seconds to wait before aborting the HTTP request. * @throws Exception if the response cannot be saved to `$destinationPath`, if the HTTP response cannot be sent, * if there are more than 5 redirects or if the request times out. * @return bool `true` on success, throws Exception on failure * @api */ public static function fetchRemoteFile($url, $destinationPath = null, $tries = 0, $timeout = 10) { @ignore_user_abort(true); SettingsServer::setMaxExecutionTime(0); return self::sendHttpRequest($url, $timeout, 'Update', $destinationPath); } /** * Utility function, parses an HTTP header line into key/value & sets header * array with them. * * @param array $headers * @param string $line */ private static function parseHeaderLine(&$headers, $line) { $parts = explode(':', $line, 2); if (count($parts) == 1) { return; } list($name, $value) = $parts; $headers[trim($name)] = trim($value); } /** * Utility function that truncates a string to an arbitrary limit. * * @param string $str The string to truncate. * @param int $limit The maximum length of the truncated string. * @return string */ private static function truncateStr($str, $limit) { if (strlen($str) > $limit) { return substr($str, 0, $limit) . '...'; } return $str; } /** * Returns the If-Modified-Since HTTP header if it can be found. If it cannot be * found, an empty string is returned. * * @return string */ public static function getModifiedSinceHeader() { $modifiedSince = ''; if (isset($_SERVER['HTTP_IF_MODIFIED_SINCE'])) { $modifiedSince = $_SERVER['HTTP_IF_MODIFIED_SINCE']; // strip any trailing data appended to header if (false !== ($semicolonPos = strpos($modifiedSince, ';'))) { $modifiedSince = substr($modifiedSince, 0, $semicolonPos); } } return $modifiedSince; } /** * Returns Proxy to use for connecting via HTTP to given URL * * @param string $url * @return array */ private static function getProxyConfiguration($url) { $hostname = UrlHelper::getHostFromUrl($url); if (Url::isLocalHost($hostname)) { return array(null, null, null, null); } // proxy configuration $proxyHost = Config::getInstance()->proxy['host']; $proxyPort = Config::getInstance()->proxy['port']; $proxyUser = Config::getInstance()->proxy['username']; $proxyPassword = Config::getInstance()->proxy['password']; return array($proxyHost, $proxyPort, $proxyUser, $proxyPassword); } }