3 * @file src/Util/Network.php
5 namespace Friendica\Util;
8 use Friendica\Core\Addon;
9 use Friendica\Core\L10n;
10 use Friendica\Core\System;
11 use Friendica\Core\Config;
12 use Friendica\Network\Probe;
13 use Friendica\Object\Image;
14 use Friendica\Util\XML;
23 * If binary flag is true, return binary results.
24 * Set the cookiejar argument to a string (e.g. "/tmp/friendica-cookies.txt")
25 * to preserve cookies from one request to the next.
27 * @param string $url URL to fetch
28 * @param boolean $binary default false
29 * TRUE if asked to return binary results (file download)
30 * @param integer $redirects The recursion counter for internal use - default 0
31 * @param integer $timeout Timeout in seconds, default system config value or 60 seconds
32 * @param string $accept_content supply Accept: header with 'accept_content' as the value
33 * @param string $cookiejar Path to cookie jar file
35 * @return string The fetched content
37 public static function fetchUrl($url, $binary = false, &$redirects = 0, $timeout = 0, $accept_content = null, $cookiejar = 0)
44 'accept_content'=>$accept_content,
45 'cookiejar'=>$cookiejar
53 * @brief fetches an URL.
55 * @param string $url URL to fetch
56 * @param boolean $binary default false
57 * TRUE if asked to return binary results (file download)
58 * @param int $redirects The recursion counter for internal use - default 0
59 * @param array $opts (optional parameters) assoziative array with:
60 * 'accept_content' => supply Accept: header with 'accept_content' as the value
61 * 'timeout' => int Timeout in seconds, default system config value or 60 seconds
62 * 'http_auth' => username:password
63 * 'novalidate' => do not validate SSL certs, default is to validate using our CA list
64 * 'nobody' => only return the header
65 * 'cookiejar' => path to cookie jar file
67 * @return array an assoziative array with:
68 * int 'return_code' => HTTP return code or 0 if timeout or failure
69 * boolean 'success' => boolean true (if HTTP 2xx result) or false
70 * string 'redirect_url' => in case of redirect, content was finally retrieved from this URL
71 * string 'header' => HTTP headers
72 * string 'body' => fetched content
74 public static function curl($url, $binary = false, &$redirects = 0, $opts = [])
76 $ret = ['return_code' => 0, 'success' => false, 'header' => '', 'info' => '', 'body' => ''];
78 $stamp1 = microtime(true);
82 if (self::isUrlBlocked($url)) {
83 logger('domain of ' . $url . ' is blocked', LOGGER_DATA);
87 $ch = @curl_init($url);
89 if (($redirects > 8) || (!$ch)) {
93 @curl_setopt($ch, CURLOPT_HEADER, true);
95 if (x($opts, "cookiejar")) {
96 curl_setopt($ch, CURLOPT_COOKIEJAR, $opts["cookiejar"]);
97 curl_setopt($ch, CURLOPT_COOKIEFILE, $opts["cookiejar"]);
100 // These settings aren't needed. We're following the location already.
101 // @curl_setopt($ch, CURLOPT_FOLLOWLOCATION, true);
102 // @curl_setopt($ch, CURLOPT_MAXREDIRS, 5);
104 if (x($opts, 'accept_content')) {
108 ['Accept: ' . $opts['accept_content']]
112 @curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
113 @curl_setopt($ch, CURLOPT_USERAGENT, $a->get_useragent());
115 $range = intval(Config::get('system', 'curl_range_bytes', 0));
118 @curl_setopt($ch, CURLOPT_RANGE, '0-' . $range);
121 // Without this setting it seems as if some webservers send compressed content
122 // This seems to confuse curl so that it shows this uncompressed.
123 /// @todo We could possibly set this value to "gzip" or something similar
124 curl_setopt($ch, CURLOPT_ENCODING, '');
126 if (x($opts, 'headers')) {
127 @curl_setopt($ch, CURLOPT_HTTPHEADER, $opts['headers']);
130 if (x($opts, 'nobody')) {
131 @curl_setopt($ch, CURLOPT_NOBODY, $opts['nobody']);
134 if (x($opts, 'timeout')) {
135 @curl_setopt($ch, CURLOPT_TIMEOUT, $opts['timeout']);
137 $curl_time = Config::get('system', 'curl_timeout', 60);
138 @curl_setopt($ch, CURLOPT_TIMEOUT, intval($curl_time));
141 // by default we will allow self-signed certs
142 // but you can override this
144 $check_cert = Config::get('system', 'verifyssl');
145 @curl_setopt($ch, CURLOPT_SSL_VERIFYPEER, (($check_cert) ? true : false));
148 @curl_setopt($ch, CURLOPT_SSL_VERIFYHOST, 2);
151 $proxy = Config::get('system', 'proxy');
153 if (strlen($proxy)) {
154 @curl_setopt($ch, CURLOPT_HTTPPROXYTUNNEL, 1);
155 @curl_setopt($ch, CURLOPT_PROXY, $proxy);
156 $proxyuser = @Config::get('system', 'proxyuser');
158 if (strlen($proxyuser)) {
159 @curl_setopt($ch, CURLOPT_PROXYUSERPWD, $proxyuser);
163 if (Config::get('system', 'ipv4_resolve', false)) {
164 curl_setopt($ch, CURLOPT_IPRESOLVE, CURL_IPRESOLVE_V4);
168 @curl_setopt($ch, CURLOPT_BINARYTRANSFER, 1);
171 $a->set_curl_code(0);
173 // don't let curl abort the entire application
174 // if it throws any errors.
176 $s = @curl_exec($ch);
177 $curl_info = @curl_getinfo($ch);
179 // Special treatment for HTTP Code 416
180 // See https://developer.mozilla.org/en-US/docs/Web/HTTP/Status/416
181 if (($curl_info['http_code'] == 416) && ($range > 0)) {
182 @curl_setopt($ch, CURLOPT_RANGE, '');
183 $s = @curl_exec($ch);
184 $curl_info = @curl_getinfo($ch);
187 if (curl_errno($ch) !== CURLE_OK) {
188 logger('error fetching ' . $url . ': ' . curl_error($ch), LOGGER_NORMAL);
191 $ret['errno'] = curl_errno($ch);
194 $ret['info'] = $curl_info;
196 $http_code = $curl_info['http_code'];
198 logger($url . ': ' . $http_code . " " . $s, LOGGER_DATA);
201 // Pull out multiple headers, e.g. proxy and continuation headers
202 // allow for HTTP/2.x without fixing code
204 while (preg_match('/^HTTP\/[1-2].+? [1-5][0-9][0-9]/', $base)) {
205 $chunk = substr($base, 0, strpos($base, "\r\n\r\n") + 4);
207 $base = substr($base, strlen($chunk));
210 $a->set_curl_code($http_code);
211 $a->set_curl_content_type($curl_info['content_type']);
212 $a->set_curl_headers($header);
214 if ($http_code == 301 || $http_code == 302 || $http_code == 303 || $http_code == 307) {
215 $new_location_info = @parse_url($curl_info['redirect_url']);
216 $old_location_info = @parse_url($curl_info['url']);
218 $newurl = $curl_info['redirect_url'];
220 if (($new_location_info['path'] == '') && ( $new_location_info['host'] != '')) {
221 $newurl = $new_location_info['scheme'] . '://' . $new_location_info['host'] . $old_location_info['path'];
226 if (preg_match('/(Location:|URI:)(.*?)\n/i', $header, $matches)) {
227 $newurl = trim(array_pop($matches));
229 if (strpos($newurl, '/') === 0) {
230 $newurl = $old_location_info["scheme"]."://".$old_location_info["host"].$newurl;
233 if (filter_var($newurl, FILTER_VALIDATE_URL)) {
236 return self::curl($newurl, $binary, $redirects, $opts);
240 $a->set_curl_code($http_code);
241 $a->set_curl_content_type($curl_info['content_type']);
243 $rc = intval($http_code);
244 $ret['return_code'] = $rc;
245 $ret['success'] = (($rc >= 200 && $rc <= 299) ? true : false);
246 $ret['redirect_url'] = $url;
248 if (!$ret['success']) {
249 $ret['error'] = curl_error($ch);
250 $ret['debug'] = $curl_info;
251 logger('error: '.$url.': '.$ret['return_code'].' - '.$ret['error'], LOGGER_DEBUG);
252 logger('debug: '.print_r($curl_info, true), LOGGER_DATA);
255 $ret['body'] = substr($s, strlen($header));
256 $ret['header'] = $header;
258 if (x($opts, 'debug')) {
259 $ret['debug'] = $curl_info;
264 $a->save_timestamp($stamp1, 'network');
270 * @brief Send POST request to $url
272 * @param string $url URL to post
273 * @param mixed $params array of POST variables
274 * @param string $headers HTTP headers
275 * @param integer $redirects Recursion counter for internal use - default = 0
276 * @param integer $timeout The timeout in seconds, default system config value or 60 seconds
278 * @return string The content
280 public static function post($url, $params, $headers = null, &$redirects = 0, $timeout = 0)
282 $stamp1 = microtime(true);
284 if (self::isUrlBlocked($url)) {
285 logger('post_url: domain of ' . $url . ' is blocked', LOGGER_DATA);
290 $ch = curl_init($url);
292 if (($redirects > 8) || (!$ch)) {
296 logger('post_url: start ' . $url, LOGGER_DATA);
298 curl_setopt($ch, CURLOPT_HEADER, true);
299 curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
300 curl_setopt($ch, CURLOPT_POST, 1);
301 curl_setopt($ch, CURLOPT_POSTFIELDS, $params);
302 curl_setopt($ch, CURLOPT_USERAGENT, $a->get_useragent());
304 if (Config::get('system', 'ipv4_resolve', false)) {
305 curl_setopt($ch, CURLOPT_IPRESOLVE, CURL_IPRESOLVE_V4);
308 if (intval($timeout)) {
309 curl_setopt($ch, CURLOPT_TIMEOUT, $timeout);
311 $curl_time = Config::get('system', 'curl_timeout', 60);
312 curl_setopt($ch, CURLOPT_TIMEOUT, intval($curl_time));
315 if (defined('LIGHTTPD')) {
316 if (!is_array($headers)) {
317 $headers = ['Expect:'];
319 if (!in_array('Expect:', $headers)) {
320 array_push($headers, 'Expect:');
326 curl_setopt($ch, CURLOPT_HTTPHEADER, $headers);
329 $check_cert = Config::get('system', 'verifyssl');
330 curl_setopt($ch, CURLOPT_SSL_VERIFYPEER, (($check_cert) ? true : false));
333 @curl_setopt($ch, CURLOPT_SSL_VERIFYHOST, 2);
336 $proxy = Config::get('system', 'proxy');
338 if (strlen($proxy)) {
339 curl_setopt($ch, CURLOPT_HTTPPROXYTUNNEL, 1);
340 curl_setopt($ch, CURLOPT_PROXY, $proxy);
341 $proxyuser = Config::get('system', 'proxyuser');
342 if (strlen($proxyuser)) {
343 curl_setopt($ch, CURLOPT_PROXYUSERPWD, $proxyuser);
347 $a->set_curl_code(0);
349 // don't let curl abort the entire application
350 // if it throws any errors.
352 $s = @curl_exec($ch);
355 $curl_info = curl_getinfo($ch);
356 $http_code = $curl_info['http_code'];
358 logger('post_url: result ' . $http_code . ' - ' . $url, LOGGER_DATA);
362 // Pull out multiple headers, e.g. proxy and continuation headers
363 // allow for HTTP/2.x without fixing code
365 while (preg_match('/^HTTP\/[1-2].+? [1-5][0-9][0-9]/', $base)) {
366 $chunk = substr($base, 0, strpos($base, "\r\n\r\n") + 4);
368 $base = substr($base, strlen($chunk));
371 if ($http_code == 301 || $http_code == 302 || $http_code == 303 || $http_code == 307) {
373 $new_location_info = @parse_url($curl_info['redirect_url']);
374 $old_location_info = @parse_url($curl_info['url']);
376 preg_match('/(Location:|URI:)(.*?)\n/', $header, $matches);
377 $newurl = trim(array_pop($matches));
379 if (strpos($newurl, '/') === 0) {
380 $newurl = $old_location_info["scheme"] . "://" . $old_location_info["host"] . $newurl;
383 if (filter_var($newurl, FILTER_VALIDATE_URL)) {
385 logger('post_url: redirect ' . $url . ' to ' . $newurl);
386 return self::post($newurl, $params, $headers, $redirects, $timeout);
390 $a->set_curl_code($http_code);
392 $body = substr($s, strlen($header));
394 $a->set_curl_headers($header);
398 $a->save_timestamp($stamp1, 'network');
400 logger('post_url: end ' . $url, LOGGER_DATA);
406 * @brief Check URL to see if it's real
408 * Take a URL from the wild, prepend http:// if necessary
409 * and check DNS to see if it's real (or check if is a valid IP address)
411 * @param string $url The URL to be validated
412 * @return string|boolean The actual working URL, false else
414 public static function isUrlValid($url)
416 if (Config::get('system', 'disable_url_validation')) {
420 // no naked subdomains (allow localhost for tests)
421 if (strpos($url, '.') === false && strpos($url, '/localhost/') === false) {
425 if (substr($url, 0, 4) != 'http') {
426 $url = 'http://' . $url;
429 /// @TODO Really suppress function outcomes? Why not find them + debug them?
430 $h = @parse_url($url);
432 if ((is_array($h)) && (@dns_get_record($h['host'], DNS_A + DNS_CNAME + DNS_PTR) || filter_var($h['host'], FILTER_VALIDATE_IP) )) {
440 * @brief Checks that email is an actual resolvable internet address
442 * @param string $addr The email address
443 * @return boolean True if it's a valid email address, false if it's not
445 public static function isEmailDomainValid($addr)
447 if (Config::get('system', 'disable_email_validation')) {
451 if (! strpos($addr, '@')) {
455 $h = substr($addr, strpos($addr, '@') + 1);
457 if (($h) && (dns_get_record($h, DNS_A + DNS_CNAME + DNS_PTR + DNS_MX) || filter_var($h, FILTER_VALIDATE_IP) )) {
464 * @brief Check if URL is allowed
466 * Check $url against our list of allowed sites,
467 * wildcards allowed. If allowed_sites is unset return true;
469 * @param string $url URL which get tested
470 * @return boolean True if url is allowed otherwise return false
472 public static function isUrlAllowed($url)
474 $h = @parse_url($url);
480 $str_allowed = Config::get('system', 'allowed_sites');
481 if (! $str_allowed) {
487 $host = strtolower($h['host']);
489 // always allow our own site
490 if ($host == strtolower($_SERVER['SERVER_NAME'])) {
494 $fnmatch = function_exists('fnmatch');
495 $allowed = explode(',', $str_allowed);
497 if (count($allowed)) {
498 foreach ($allowed as $a) {
499 $pat = strtolower(trim($a));
500 if (($fnmatch && fnmatch($pat, $host)) || ($pat == $host)) {
510 * Checks if the provided url domain is on the domain blocklist.
511 * Returns true if it is or malformed URL, false if not.
513 * @param string $url The url to check the domain from
517 public static function isUrlBlocked($url)
519 $h = @parse_url($url);
525 $domain_blocklist = Config::get('system', 'blocklist', []);
526 if (! $domain_blocklist) {
530 $host = strtolower($h['host']);
532 foreach ($domain_blocklist as $domain_block) {
533 if (strtolower($domain_block['domain']) == $host) {
542 * @brief Check if email address is allowed to register here.
544 * Compare against our list (wildcards allowed).
546 * @param string $email email address
547 * @return boolean False if not allowed, true if allowed
548 * or if allowed list is not configured
550 public static function isEmailDomainAllowed($email)
552 $domain = strtolower(substr($email, strpos($email, '@') + 1));
557 $str_allowed = Config::get('system', 'allowed_email', '');
558 if (!x($str_allowed)) {
562 $allowed = explode(',', $str_allowed);
564 return self::isDomainAllowed($domain, $allowed);
568 * Checks for the existence of a domain in a domain list
570 * @brief Checks for the existence of a domain in a domain list
571 * @param string $domain
572 * @param array $domain_list
575 public static function isDomainAllowed($domain, array $domain_list)
579 foreach ($domain_list as $item) {
580 $pat = strtolower(trim($item));
581 if (fnmatch($pat, $domain) || ($pat == $domain)) {
590 public static function lookupAvatarByEmail($email)
592 $avatar['size'] = 175;
593 $avatar['email'] = $email;
595 $avatar['success'] = false;
597 Addon::callHooks('avatar_lookup', $avatar);
599 if (! $avatar['success']) {
600 $avatar['url'] = System::baseUrl() . '/images/person-175.jpg';
603 logger('Avatar: ' . $avatar['email'] . ' ' . $avatar['url'], LOGGER_DEBUG);
604 return $avatar['url'];
608 * @brief Remove Google Analytics and other tracking platforms params from URL
610 * @param string $url Any user-submitted URL that may contain tracking params
611 * @return string The same URL stripped of tracking parameters
613 public static function stripTrackingQueryParams($url)
615 $urldata = parse_url($url);
616 if (is_string($urldata["query"])) {
617 $query = $urldata["query"];
618 parse_str($query, $querydata);
620 if (is_array($querydata)) {
621 foreach ($querydata as $param => $value) {
625 "utm_source", "utm_medium", "utm_term", "utm_content", "utm_campaign",
626 "wt_mc", "pk_campaign", "pk_kwd", "mc_cid", "mc_eid",
627 "fb_action_ids", "fb_action_types", "fb_ref",
629 "woo_campaign", "woo_source", "woo_medium", "woo_content", "woo_term"]
632 $pair = $param . "=" . urlencode($value);
633 $url = str_replace($pair, "", $url);
635 // Second try: if the url isn't encoded completely
636 $pair = $param . "=" . str_replace(" ", "+", $value);
637 $url = str_replace($pair, "", $url);
639 // Third try: Maybey the url isn't encoded at all
640 $pair = $param . "=" . $value;
641 $url = str_replace($pair, "", $url);
643 $url = str_replace(["?&", "&&"], ["?", ""], $url);
648 if (substr($url, -1, 1) == "?") {
649 $url = substr($url, 0, -1);
657 * @brief Returns the original URL of the provided URL
659 * This function strips tracking query params and follows redirections, either
660 * through HTTP code or meta refresh tags. Stops after 10 redirections.
662 * @todo Remove the $fetchbody parameter that generates an extraneous HEAD request
664 * @see ParseUrl::getSiteinfo
666 * @param string $url A user-submitted URL
667 * @param int $depth The current redirection recursion level (internal)
668 * @param bool $fetchbody Wether to fetch the body or not after the HEAD requests
669 * @return string A canonical URL
671 public static function finalUrl($url, $depth = 1, $fetchbody = false)
675 $url = self::stripTrackingQueryParams($url);
681 $url = trim($url, "'");
683 $stamp1 = microtime(true);
686 curl_setopt($ch, CURLOPT_URL, $url);
687 curl_setopt($ch, CURLOPT_HEADER, 1);
688 curl_setopt($ch, CURLOPT_NOBODY, 1);
689 curl_setopt($ch, CURLOPT_TIMEOUT, 10);
690 curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
691 curl_setopt($ch, CURLOPT_USERAGENT, $a->get_useragent());
694 $curl_info = @curl_getinfo($ch);
695 $http_code = $curl_info['http_code'];
698 $a->save_timestamp($stamp1, "network");
700 if ($http_code == 0) {
704 if ((($curl_info['http_code'] == "301") || ($curl_info['http_code'] == "302"))
705 && (($curl_info['redirect_url'] != "") || ($curl_info['location'] != ""))
707 if ($curl_info['redirect_url'] != "") {
708 return(self::finalUrl($curl_info['redirect_url'], ++$depth, $fetchbody));
710 return(self::finalUrl($curl_info['location'], ++$depth, $fetchbody));
714 // Check for redirects in the meta elements of the body if there are no redirects in the header.
716 return(self::finalUrl($url, ++$depth, true));
719 // if the file is too large then exit
720 if ($curl_info["download_content_length"] > 1000000) {
724 // if it isn't a HTML file then exit
725 if (($curl_info["content_type"] != "") && !strstr(strtolower($curl_info["content_type"]), "html")) {
729 $stamp1 = microtime(true);
732 curl_setopt($ch, CURLOPT_URL, $url);
733 curl_setopt($ch, CURLOPT_HEADER, 0);
734 curl_setopt($ch, CURLOPT_NOBODY, 0);
735 curl_setopt($ch, CURLOPT_TIMEOUT, 10);
736 curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
737 curl_setopt($ch, CURLOPT_USERAGENT, $a->get_useragent());
739 $body = curl_exec($ch);
742 $a->save_timestamp($stamp1, "network");
744 if (trim($body) == "") {
748 // Check for redirect in meta elements
749 $doc = new DOMDocument();
750 @$doc->loadHTML($body);
752 $xpath = new DomXPath($doc);
754 $list = $xpath->query("//meta[@content]");
755 foreach ($list as $node) {
757 if ($node->attributes->length) {
758 foreach ($node->attributes as $attribute) {
759 $attr[$attribute->name] = $attribute->value;
763 if (@$attr["http-equiv"] == 'refresh') {
764 $path = $attr["content"];
765 $pathinfo = explode(";", $path);
766 foreach ($pathinfo as $value) {
767 if (substr(strtolower($value), 0, 4) == "url=") {
768 return(self::finalUrl(substr($value, 4), ++$depth));
778 * @brief Find the matching part between two url
780 * @param string $url1
781 * @param string $url2
782 * @return string The matching part
784 public static function getUrlMatch($url1, $url2)
786 if (($url1 == "") || ($url2 == "")) {
790 $url1 = normalise_link($url1);
791 $url2 = normalise_link($url2);
793 $parts1 = parse_url($url1);
794 $parts2 = parse_url($url2);
796 if (!isset($parts1["host"]) || !isset($parts2["host"])) {
800 if ($parts1["scheme"] != $parts2["scheme"]) {
804 if ($parts1["host"] != $parts2["host"]) {
808 if ($parts1["port"] != $parts2["port"]) {
812 $match = $parts1["scheme"]."://".$parts1["host"];
814 if ($parts1["port"]) {
815 $match .= ":".$parts1["port"];
818 $pathparts1 = explode("/", $parts1["path"]);
819 $pathparts2 = explode("/", $parts2["path"]);
824 $path1 = $pathparts1[$i];
825 $path2 = $pathparts2[$i];
827 if ($path1 == $path2) {
830 } while (($path1 == $path2) && ($i++ <= count($pathparts1)));
834 return normalise_link($match);
838 * @brief Glue url parts together
840 * @param array $parsed URL parts
842 * @return string The glued URL
844 public static function unparseURL($parsed)
846 $get = function ($key) use ($parsed) {
847 return isset($parsed[$key]) ? $parsed[$key] : null;
850 $pass = $get('pass');
851 $user = $get('user');
852 $userinfo = $pass !== null ? "$user:$pass" : $user;
853 $port = $get('port');
854 $scheme = $get('scheme');
855 $query = $get('query');
856 $fragment = $get('fragment');
857 $authority = ($userinfo !== null ? $userinfo."@" : '') .
859 ($port ? ":$port" : '');
861 return (strlen($scheme) ? $scheme.":" : '') .
862 (strlen($authority) ? "//".$authority : '') .
864 (strlen($query) ? "?".$query : '') .
865 (strlen($fragment) ? "#".$fragment : '');