X-Git-Url: https://git.mxchange.org/?a=blobdiff_plain;f=src%2FUtil%2FNetwork.php;h=af4d18b4ab037301ea3bc5692519cb8334977dae;hb=7026dd37db61e6672d6f9c6ac2deb5046be1c7a2;hp=6c7fd731cbf711f096efab0990edcf88d5bf9329;hpb=2c56d2f3360c08e312e5c167261af8e5d4b87af4;p=friendica.git diff --git a/src/Util/Network.php b/src/Util/Network.php index 6c7fd731cb..af4d18b4ab 100644 --- a/src/Util/Network.php +++ b/src/Util/Network.php @@ -1,6 +1,6 @@ getBody(); - } - - /** - * Curl wrapper with array of return values. - * - * Inner workings and parameters are the same as @ref fetchUrl but returns an array with - * all the information collected during the fetch. - * - * @param string $url URL to fetch - * @param bool $binary default false - * TRUE if asked to return binary results (file download) - * @param int $timeout Timeout in seconds, default system config value or 60 seconds - * @param string $accept_content supply Accept: header with 'accept_content' as the value - * @param string $cookiejar Path to cookie jar file - * @param int $redirects The recursion counter for internal use - default 0 - * - * @return CurlResult With all relevant information, 'body' contains the actual fetched content. - * @throws \Friendica\Network\HTTPException\InternalServerErrorException - */ - public static function fetchUrlFull(string $url, bool $binary = false, int $timeout = 0, string $accept_content = '', string $cookiejar = '', int &$redirects = 0) - { - return self::curl( - $url, - $binary, - [ - 'timeout' => $timeout, - 'accept_content' => $accept_content, - 'cookiejar' => $cookiejar - ], - $redirects - ); - } - - /** - * fetches an URL. - * - * @param string $url URL to fetch - * @param bool $binary default false - * TRUE if asked to return binary results (file download) - * @param array $opts (optional parameters) assoziative array with: - * 'accept_content' => supply Accept: header with 'accept_content' as the value - * 'timeout' => int Timeout in seconds, default system config value or 60 seconds - * 'http_auth' => username:password - * 'novalidate' => do not validate SSL certs, default is to validate using our CA list - * 'nobody' => only return the header - * 'cookiejar' => path to cookie jar file - * 'header' => header array - * @param int $redirects The recursion counter for internal use - default 0 - * - * @return CurlResult - * @throws \Friendica\Network\HTTPException\InternalServerErrorException - */ - public static function curl(string $url, bool $binary = false, array $opts = [], int &$redirects = 0) - { - $stamp1 = microtime(true); - - $a = DI::app(); - - if (strlen($url) > 1000) { - Logger::log('URL is longer than 1000 characters. Callstack: ' . System::callstack(20), Logger::DEBUG); - return CurlResult::createErrorCurl(substr($url, 0, 200)); - } - - $parts2 = []; - $parts = parse_url($url); - $path_parts = explode('/', $parts['path'] ?? ''); - foreach ($path_parts as $part) { - if (strlen($part) <> mb_strlen($part)) { - $parts2[] = rawurlencode($part); - } else { - $parts2[] = $part; - } - } - $parts['path'] = implode('/', $parts2); - $url = self::unparseURL($parts); - - if (self::isUrlBlocked($url)) { - Logger::log('domain of ' . $url . ' is blocked', Logger::DATA); - return CurlResult::createErrorCurl($url); - } - - $ch = @curl_init($url); - - if (($redirects > 8) || (!$ch)) { - return CurlResult::createErrorCurl($url); - } - - @curl_setopt($ch, CURLOPT_HEADER, true); - - if (!empty($opts['cookiejar'])) { - curl_setopt($ch, CURLOPT_COOKIEJAR, $opts["cookiejar"]); - curl_setopt($ch, CURLOPT_COOKIEFILE, $opts["cookiejar"]); - } - - // These settings aren't needed. We're following the location already. - // @curl_setopt($ch, CURLOPT_FOLLOWLOCATION, true); - // @curl_setopt($ch, CURLOPT_MAXREDIRS, 5); - - if (!empty($opts['accept_content'])) { - curl_setopt( - $ch, - CURLOPT_HTTPHEADER, - ['Accept: ' . $opts['accept_content']] - ); - } - - if (!empty($opts['header'])) { - curl_setopt($ch, CURLOPT_HTTPHEADER, $opts['header']); - } - - @curl_setopt($ch, CURLOPT_RETURNTRANSFER, true); - @curl_setopt($ch, CURLOPT_USERAGENT, $a->getUserAgent()); - - $range = intval(DI::config()->get('system', 'curl_range_bytes', 0)); - - if ($range > 0) { - @curl_setopt($ch, CURLOPT_RANGE, '0-' . $range); - } - - // Without this setting it seems as if some webservers send compressed content - // This seems to confuse curl so that it shows this uncompressed. - /// @todo We could possibly set this value to "gzip" or something similar - curl_setopt($ch, CURLOPT_ENCODING, ''); - - if (!empty($opts['headers'])) { - @curl_setopt($ch, CURLOPT_HTTPHEADER, $opts['headers']); - } - - if (!empty($opts['nobody'])) { - @curl_setopt($ch, CURLOPT_NOBODY, $opts['nobody']); - } - - if (!empty($opts['timeout'])) { - @curl_setopt($ch, CURLOPT_TIMEOUT, $opts['timeout']); - } else { - $curl_time = DI::config()->get('system', 'curl_timeout', 60); - @curl_setopt($ch, CURLOPT_TIMEOUT, intval($curl_time)); - } - - // by default we will allow self-signed certs - // but you can override this - - $check_cert = DI::config()->get('system', 'verifyssl'); - @curl_setopt($ch, CURLOPT_SSL_VERIFYPEER, (($check_cert) ? true : false)); - - if ($check_cert) { - @curl_setopt($ch, CURLOPT_SSL_VERIFYHOST, 2); - } - - $proxy = DI::config()->get('system', 'proxy'); - - if (strlen($proxy)) { - @curl_setopt($ch, CURLOPT_HTTPPROXYTUNNEL, 1); - @curl_setopt($ch, CURLOPT_PROXY, $proxy); - $proxyuser = @DI::config()->get('system', 'proxyuser'); - - if (strlen($proxyuser)) { - @curl_setopt($ch, CURLOPT_PROXYUSERPWD, $proxyuser); - } - } - - if (DI::config()->get('system', 'ipv4_resolve', false)) { - curl_setopt($ch, CURLOPT_IPRESOLVE, CURL_IPRESOLVE_V4); - } - - if ($binary) { - @curl_setopt($ch, CURLOPT_BINARYTRANSFER, 1); - } - - // don't let curl abort the entire application - // if it throws any errors. - - $s = @curl_exec($ch); - $curl_info = @curl_getinfo($ch); - - // Special treatment for HTTP Code 416 - // See https://developer.mozilla.org/en-US/docs/Web/HTTP/Status/416 - if (($curl_info['http_code'] == 416) && ($range > 0)) { - @curl_setopt($ch, CURLOPT_RANGE, ''); - $s = @curl_exec($ch); - $curl_info = @curl_getinfo($ch); - } - - $curlResponse = new CurlResult($url, $s, $curl_info, curl_errno($ch), curl_error($ch)); - - if ($curlResponse->isRedirectUrl()) { - $redirects++; - Logger::log('curl: redirect ' . $url . ' to ' . $curlResponse->getRedirectUrl()); - @curl_close($ch); - return self::curl($curlResponse->getRedirectUrl(), $binary, $opts, $redirects); - } - - @curl_close($ch); - - DI::profiler()->saveTimestamp($stamp1, 'network', System::callstack()); - - return $curlResponse; - } - - /** - * Send POST request to $url - * - * @param string $url URL to post - * @param mixed $params array of POST variables - * @param array $headers HTTP headers - * @param int $redirects Recursion counter for internal use - default = 0 - * @param int $timeout The timeout in seconds, default system config value or 60 seconds - * - * @return CurlResult The content - * @throws \Friendica\Network\HTTPException\InternalServerErrorException - */ - public static function post(string $url, $params, array $headers = [], int $timeout = 0, int &$redirects = 0) - { - $stamp1 = microtime(true); - - if (self::isUrlBlocked($url)) { - Logger::log('post_url: domain of ' . $url . ' is blocked', Logger::DATA); - return CurlResult::createErrorCurl($url); - } - - $a = DI::app(); - $ch = curl_init($url); - - if (($redirects > 8) || (!$ch)) { - return CurlResult::createErrorCurl($url); - } - - Logger::log('post_url: start ' . $url, Logger::DATA); - - curl_setopt($ch, CURLOPT_HEADER, true); - curl_setopt($ch, CURLOPT_RETURNTRANSFER, true); - curl_setopt($ch, CURLOPT_POST, 1); - curl_setopt($ch, CURLOPT_POSTFIELDS, $params); - curl_setopt($ch, CURLOPT_USERAGENT, $a->getUserAgent()); - - if (DI::config()->get('system', 'ipv4_resolve', false)) { - curl_setopt($ch, CURLOPT_IPRESOLVE, CURL_IPRESOLVE_V4); - } - - if (intval($timeout)) { - curl_setopt($ch, CURLOPT_TIMEOUT, $timeout); - } else { - $curl_time = DI::config()->get('system', 'curl_timeout', 60); - curl_setopt($ch, CURLOPT_TIMEOUT, intval($curl_time)); - } - - if (!empty($headers)) { - curl_setopt($ch, CURLOPT_HTTPHEADER, $headers); - } - - $check_cert = DI::config()->get('system', 'verifyssl'); - curl_setopt($ch, CURLOPT_SSL_VERIFYPEER, (($check_cert) ? true : false)); - - if ($check_cert) { - @curl_setopt($ch, CURLOPT_SSL_VERIFYHOST, 2); - } - - $proxy = DI::config()->get('system', 'proxy'); - - if (strlen($proxy)) { - curl_setopt($ch, CURLOPT_HTTPPROXYTUNNEL, 1); - curl_setopt($ch, CURLOPT_PROXY, $proxy); - $proxyuser = DI::config()->get('system', 'proxyuser'); - if (strlen($proxyuser)) { - curl_setopt($ch, CURLOPT_PROXYUSERPWD, $proxyuser); - } - } - - // don't let curl abort the entire application - // if it throws any errors. - - $s = @curl_exec($ch); - - $curl_info = curl_getinfo($ch); - - $curlResponse = new CurlResult($url, $s, $curl_info, curl_errno($ch), curl_error($ch)); - - if ($curlResponse->isRedirectUrl()) { - $redirects++; - Logger::log('post_url: redirect ' . $url . ' to ' . $curlResponse->getRedirectUrl()); - curl_close($ch); - return self::post($curlResponse->getRedirectUrl(), $params, $headers, $redirects, $timeout); - } - - curl_close($ch); - - DI::profiler()->saveTimestamp($stamp1, 'network', System::callstack()); - - // Very old versions of Lighttpd don't like the "Expect" header, so we remove it when needed - if ($curlResponse->getReturnCode() == 417) { - $redirects++; - - if (empty($headers)) { - $headers = ['Expect:']; - } else { - if (!in_array('Expect:', $headers)) { - array_push($headers, 'Expect:'); - } - } - Logger::info('Server responds with 417, applying workaround', ['url' => $url]); - return self::post($url, $params, $headers, $redirects, $timeout); - } - - Logger::log('post_url: end ' . $url, Logger::DATA); - - return $curlResponse; - } /** * Return raw post data from a post request @@ -510,6 +179,35 @@ class Network return false; } + /** + * Checks if the provided url is on the list of domains where redirects are blocked. + * Returns true if it is or malformed URL, false if not. + * + * @param string $url The url to check the domain from + * + * @return boolean + */ + public static function isRedirectBlocked(string $url) + { + $host = @parse_url($url, PHP_URL_HOST); + if (!$host) { + return false; + } + + $no_redirect_list = DI::config()->get('system', 'no_redirect_list', []); + if (!$no_redirect_list) { + return false; + } + + foreach ($no_redirect_list as $no_redirect) { + if (fnmatch(strtolower($no_redirect), strtolower($host))) { + return true; + } + } + + return false; + } + /** * Check if email address is allowed to register here. * @@ -569,10 +267,10 @@ class Network Hook::callAll('avatar_lookup', $avatar); if (! $avatar['success']) { - $avatar['url'] = DI::baseUrl() . '/images/person-300.jpg'; + $avatar['url'] = DI::baseUrl() . Contact::DEFAULT_AVATAR_PHOTO; } - Logger::log('Avatar: ' . $avatar['email'] . ' ' . $avatar['url'], Logger::DEBUG); + Logger::info('Avatar: ' . $avatar['email'] . ' ' . $avatar['url']); return $avatar['url']; } @@ -626,123 +324,23 @@ class Network } /** - * Returns the original URL of the provided URL - * - * This function strips tracking query params and follows redirections, either - * through HTTP code or meta refresh tags. Stops after 10 redirections. - * - * @todo Remove the $fetchbody parameter that generates an extraneous HEAD request + * Add a missing base path (scheme and host) to a given url * - * @see ParseUrl::getSiteinfo - * - * @param string $url A user-submitted URL - * @param int $depth The current redirection recursion level (internal) - * @param bool $fetchbody Wether to fetch the body or not after the HEAD requests - * @return string A canonical URL - * @throws \Friendica\Network\HTTPException\InternalServerErrorException + * @param string $url + * @param string $basepath + * @return string url */ - public static function finalUrl(string $url, int $depth = 1, bool $fetchbody = false) + public static function addBasePath(string $url, string $basepath) { - $a = DI::app(); - - $url = self::stripTrackingQueryParams($url); - - if ($depth > 10) { - return $url; - } - - $url = trim($url, "'"); - - $stamp1 = microtime(true); - - $ch = curl_init(); - curl_setopt($ch, CURLOPT_URL, $url); - curl_setopt($ch, CURLOPT_HEADER, 1); - curl_setopt($ch, CURLOPT_NOBODY, 1); - curl_setopt($ch, CURLOPT_TIMEOUT, 10); - curl_setopt($ch, CURLOPT_RETURNTRANSFER, true); - curl_setopt($ch, CURLOPT_USERAGENT, $a->getUserAgent()); - - curl_exec($ch); - $curl_info = @curl_getinfo($ch); - $http_code = $curl_info['http_code']; - curl_close($ch); - - DI::profiler()->saveTimestamp($stamp1, "network", System::callstack()); - - if ($http_code == 0) { - return $url; - } - - if (in_array($http_code, ['301', '302'])) { - if (!empty($curl_info['redirect_url'])) { - return self::finalUrl($curl_info['redirect_url'], ++$depth, $fetchbody); - } elseif (!empty($curl_info['location'])) { - return self::finalUrl($curl_info['location'], ++$depth, $fetchbody); - } - } - - // Check for redirects in the meta elements of the body if there are no redirects in the header. - if (!$fetchbody) { - return(self::finalUrl($url, ++$depth, true)); - } - - // if the file is too large then exit - if ($curl_info["download_content_length"] > 1000000) { - return $url; - } - - // if it isn't a HTML file then exit - if (!empty($curl_info["content_type"]) && !strstr(strtolower($curl_info["content_type"]), "html")) { + if (!empty(parse_url($url, PHP_URL_SCHEME)) || empty(parse_url($basepath, PHP_URL_SCHEME)) || empty($url) || empty(parse_url($url))) { return $url; } - $stamp1 = microtime(true); - - $ch = curl_init(); - curl_setopt($ch, CURLOPT_URL, $url); - curl_setopt($ch, CURLOPT_HEADER, 0); - curl_setopt($ch, CURLOPT_NOBODY, 0); - curl_setopt($ch, CURLOPT_TIMEOUT, 10); - curl_setopt($ch, CURLOPT_RETURNTRANSFER, true); - curl_setopt($ch, CURLOPT_USERAGENT, $a->getUserAgent()); - - $body = curl_exec($ch); - curl_close($ch); - - DI::profiler()->saveTimestamp($stamp1, "network", System::callstack()); - - if (trim($body) == "") { - return $url; - } + $base = ['scheme' => parse_url($basepath, PHP_URL_SCHEME), + 'host' => parse_url($basepath, PHP_URL_HOST)]; - // Check for redirect in meta elements - $doc = new DOMDocument(); - @$doc->loadHTML($body); - - $xpath = new DomXPath($doc); - - $list = $xpath->query("//meta[@content]"); - foreach ($list as $node) { - $attr = []; - if ($node->attributes->length) { - foreach ($node->attributes as $attribute) { - $attr[$attribute->name] = $attribute->value; - } - } - - if (@$attr["http-equiv"] == 'refresh') { - $path = $attr["content"]; - $pathinfo = explode(";", $path); - foreach ($pathinfo as $value) { - if (substr(strtolower($value), 0, 4) == "url=") { - return self::finalUrl(substr($value, 4), ++$depth); - } - } - } - } - - return $url; + $parts = array_merge($base, parse_url('/' . ltrim($url, '/'))); + return self::unparseURL($parts); } /** @@ -838,7 +436,8 @@ class Network * * @param array $parsed URL parts * - * @return string The glued URL + * @return string The glued URL. + * @deprecated since version 2021.12, use GuzzleHttp\Psr7\Uri::fromParts($parts) instead */ public static function unparseURL(array $parsed) { @@ -864,6 +463,29 @@ class Network (strlen($fragment) ? "#".$fragment : ''); } + /** + * Convert an URI to an IDN compatible URI + * + * @param string $uri + * @return string + */ + public static function convertToIdn(string $uri): string + { + $parts = parse_url($uri); + if (!empty($parts['scheme']) && !empty($parts['host'])) { + $parts['host'] = idn_to_ascii($parts['host']); + $uri = Uri::fromParts($parts); + } else { + $parts = explode('@', $uri); + if (count($parts) == 2) { + $uri = $parts[0] . '@' . idn_to_ascii($parts[1]); + } else { + $uri = idn_to_ascii($uri); + } + } + + return $uri; + } /** * Switch the scheme of an url between http and https @@ -910,4 +532,68 @@ class Network return self::unparseURL($parsed); } + + /** + * Generates ETag and Last-Modified response headers and checks them against + * If-None-Match and If-Modified-Since request headers if present. + * + * Blocking function, sends 304 headers and exits if check passes. + * + * @param string $etag The page etag + * @param string $last_modified The page last modification UTC date + * @throws \Exception + */ + public static function checkEtagModified(string $etag, string $last_modified) + { + $last_modified = DateTimeFormat::utc($last_modified, 'D, d M Y H:i:s') . ' GMT'; + + /** + * @see http://www.w3.org/Protocols/rfc2616/rfc2616-sec14.html#sec14.26 + */ + $if_none_match = filter_input(INPUT_SERVER, 'HTTP_IF_NONE_MATCH'); + $if_modified_since = filter_input(INPUT_SERVER, 'HTTP_IF_MODIFIED_SINCE'); + $flag_not_modified = null; + if ($if_none_match) { + $result = []; + preg_match('/^(?:W\/")?([^"]+)"?$/i', $etag, $result); + $etagTrimmed = $result[1]; + // Lazy exact ETag match, could check weak/strong ETags + $flag_not_modified = $if_none_match == '*' || strpos($if_none_match, $etagTrimmed) !== false; + } + + if ($if_modified_since && (!$if_none_match || $flag_not_modified)) { + // Lazy exact Last-Modified match, could check If-Modified-Since validity + $flag_not_modified = $if_modified_since == $last_modified; + } + + header('Etag: ' . $etag); + header('Last-Modified: ' . $last_modified); + + if ($flag_not_modified) { + throw new NotModifiedException(); + } + } + + /** + * Check if the given URL is a local link + * + * @param string $url + * @return bool + */ + public static function isLocalLink(string $url) + { + return (strpos(Strings::normaliseLink($url), Strings::normaliseLink(DI::baseUrl())) !== false); + } + + /** + * Check if the given URL is a valid HTTP/HTTPS URL + * + * @param string $url + * @return bool + */ + public static function isValidHttpUrl(string $url) + { + $scheme = parse_url($url, PHP_URL_SCHEME); + return !empty($scheme) && in_array($scheme, ['http', 'https']) && parse_url($url, PHP_URL_HOST); + } }